Imunology

editOrial
On supplementary information
Nature Immunology is implementing new editorial guidelines for the supplementary material of research articles.
eginning with this issue, Nature Immunology is changing its guidelines for supplementary information by implementing a restriction on the number of supplementary figures that accompany research articles. Under these new guidelines, the number of supplementary figures should not exceed the number of main figures in each article. This change represents a natural response to feedback received during the past year from the research community. Immunologists, along with researchers in other areas of science, are unsatisfied with the present practice of adding an ever-growing amount of supplementary information to published papers. A change is needed in the way supplementary material is viewed. Authors often complain that experimental data requested during the review process and requiring substantial effort to generate, in terms of time and money, ends up as part of a large file of supplementary information that the non-aficionado might never access. Indeed, this complaint may be part of a more general dissatisfaction with the present tendency of the review process to always demand additional data, some only marginally related to the main story. For referees, a large amount of supplementary data makes it more difficult to find the information needed to evaluate manuscripts, adds considerably to the total workload and possibly increases the time required for review. In addition, readers complain that authors sometimes use the supplementary information section as a repository for unrelated and unpublished data in search of a home. Those are all justified concerns that are only part of a more complex picture. The size of the average research article has continued to grow over time, and as more data are included in one report, more controls and additional experimental clarifications must be presented along with the main results. We recognize the utility of having a repository for this additional information. Data that provide proof of principle of the utility of an assay or that demonstrate efficient deletion of a protein or silencing of an RNA message, experiments that exclude alternative possibilities, and gating strategies for complex flow cytometry or sorting procedures represent an important and informative part of the research that needs to be presented to the audience. They provide experts in the field and reviewers with valuable details on how the experiments were done and how conclusions and interpretations were derived from the data. In addition, certain types of data, such as movies, imaging files or large data sets, can be included only as supplementary information and thus will continue to have no restric-
tions. Consistent with that, we will allow resource articles, which are reports of large data sets, such as comprehensive analyses of proteins or mRNA in various cells or tissues, to have an unlimited number of associated supplementary items. However, the same guiding principle should be applied when deciding what to display in the supplementary information or main body of a manuscript; that is, data that are directly relevant to the main conclusions of the paper. As always, the main purpose of the review and editorial processes here at Nature Immunology is not only to report innovative and technically reliable results but also to produce a coherent, elegant and selfcontained body of work. We encourage authors to be selective and carefully review the content and extent of the supplementary material included in research articles at submission. As the review process might result in the addition of new data, the editors will work with the authors and reviewers to critically evaluate the scope and quality of the supplementary information. During the editing stage, we will continue to provide help and suggestions to ensure that this new material is sensibly integrated into the manuscript and that important information is easily available to our readers. As part of the same effort to streamline our papers and to make them more accessible to our readers, supplementary methods will no longer be included in the supplementary information. All research and resource papers will have a single methods section included with the main body of the article, but it will be published online only. To accommodate this change, we have increased the size of the methods section from the previous limit of 1,000 words to a limit of approximately 1,500 words. On a case-by-case basis, longer methods sections will be accepted at the editors discretion. To compensate for limitations in the length of the methods section, Nature Protocols offers the option of uploading protocols used for each manuscript to Protocol Exchange, an open online resource that allows researchers to share detailed protocols and experimental know-how. Protocols are linked from the article, are made freely available, are assigned digital object identifier numbers for ease of citation and are fully searchable at the Nature website. Less can be more, and by imposing restrictions on the number of supplementary figures and reorganizing the methods section, we hope to shape more concise and focused papers that will benefit our audience. As needs may change in the future, the guidelines for and format of our research articles will change as well, in an effort to accommodate the requirements of quickly evolving research.
npg
2012 Nature America, Inc. All rights reserved.
nature immunology volume 13 number 6 june 2012
519
C o M M E N TA RY
Immune to addiction: the ethical dimensions of vaccines against substance abuse

Michael J Young, Dominic A Sisti, Hila Rimon-Greenspan, Jason L Schwartz & Arthur L Caplan
Promising advances have been made in recent years for a unique class of immunotherapies that use vaccination to combat substance-use disorders. Although such vaccines are potentially useful for addictions, they raise a variety of ethical and social questions.
he appearance of anti-addiction immunoprophylaxis represents a paradigm shift both in vaccinology, which has been traditionally focused on infectious, somatic disease, and in addiction medicine, which traditionally has aimed to rehabilitate patients through a combination of cognitive and behavioral therapies, pharmacological treatments and maintenance programs. Interest has increased in the development of vaccines for the prophylaxis and treatment of many substance-use disorders. Vaccines to treat addiction to nicotine, methamphetamine and morphine, or phencyclidine are at various stages of preclinical or clinical investigation1. The ethical questions raised by the development and deployment of such vaccines are multifaceted and are pertinent to patients, clinicians, researchers and society (Box 1). As it is furthest along in trials and more is understood about its biology, the emerging vaccine used as therapy for addictioncocaine dependence (TA-CD) is the focus here as a paradigm case that elucidates many of the questions raised by other proposed immunological
Michael J. Young is in the Faculty of Philosophy, University of Cambridge, Cambridge, UK. Dominic A. Sisti and Hila Rimon-Greenspan are with The Scattergood Program for the Applied Ethics of Behavioral Healthcare, Department of Medical Ethics and Health Policy, University of Pennsylvania Perelman School of Medicine, Philadelphia, Pennsylvania, USA. Jason L. Schwartz is in the Department of Medical Ethics and Health Policy, University of Pennsylvania Perelman School of Medicine, Philadelphia, Pennsylvania, USA. Arthur L. Caplan is with the Division of Medical Ethics, NYU Langone Medical Center, New York, New York, USA. e-mail: mjy25@cam.ac.uk
interventions for substance abuse. Key issues posed by the availability of vaccines of this kind include their prophylactic use, therapeutic use and unintended socioeconomic consequences and the ethics of conducting research to bring such vaccines to market. Why vaccinate for substance dependence? The growing interest in vaccines of this kind has been motivated by the considerable effects of substance-use disorders on health and society and by the inadequacy of many existing treatments. Such issues are particularly notable in the epidemiology of cocaine use and dependence. According to the US Department of Justice, in 2008, approximately 5.3 million people 12 years of age or older reported cocaine use in the past year. Cocaine accounts for a larger percentage of visits to the emergency department than does any other illicit drug2. In addition to its own risks, cocaine is often surreptitiously cut with other unsafe substances before sale, which adds to the drugs threats to health. Cocaine use also adversely affects rates of crime and productivity at work. Despite the severity of this public-health problem, there is no medication approved by the US Food and Drug Administration for the treatment of cocaine dependence. Some products are being studied and used off-label for this purpose with limited success3. The disproportionately high relapse rates of patients treated with existing psychopharmacotherapies, coupled with the growing socioeconomic burdens of cocaine dependence4, have triggered recent interest in the exploration of new kinds of treatments. The TA-CD vaccine, which is now entering phase II multisite clinical trials, is produced by
the conjugation of a cocaine derivative (succinylnorcocaine) with an inactive cholera toxin as an adjuvant (Box 2). This immunogenic duo elicits cocaine-specific antibodies that bind to cocaine after intake, which prevents the drug from crossing into the central nervous system and blunts its psychoactive effects5. Unlike anti-cocaine pharmacotherapies, which act in the brain to modulate the effects of cocaine or mitigate withdrawal symptoms, the active cocaine vaccine triggers the production of antibodies that sequester cocaine in the blood to prevent it from reaching the brain5,6. TA-CD has the dual potential of treating and preventing cocaine dependence by recruiting the immune system to block cocaine from crossing the blood-brain barrier57. Other emerging vaccines for the treatment of addiction, including those that target addiction to nicotine and opiates, typically operate through analogous immunological mechanisms and have similar theoretical potential1. To treat or prevent? Clinical testing of the TA-CD vaccine has demonstrated its therapeutic promise. In early trials, the vaccine was well tolerated and demonstrated a dose-dependent potential to decrease likelihood of drug use in human and animal subjects5,6. Should research outcomes remain favorable for this product or others like it, the potential use of a vaccine of this kind would raise many ethical and social concerns that would require attention and deliberation. Chief among such questions is whether such a vaccine ought to be used to treat only those already diagnosed with dependence or whether it should also be used to prevent dependence in healthy people.
521
npg
c O m m e n ta r y
Box 1 Ethical questions in substance abuse vaccine development and implementation
Medicine & Public Health Should vaccines be used to control risky or destructive behaviors? Ought vaccines be used only to treat or also to prevent drug dependence? Is drug dependence sufficiently comparable to infectious disease to warrant similar public health policies? Who should be vaccinated? Which agencies or professionals should administer vaccines against substance abuse? Should vaccination be compulsory or optional? Is there a professional obligation to protect children from future addictions? How can clinical equipoise be maximized in trials? Would mandated vaccination trigger replacement-drug markets? Will the emergence of substance abuse vaccines lead to a lower perceived risk of substance use? What strategies should be implemented to address the social determinants of substance abuse? How can equitable access to the vaccine be maximized? Do the costs of the development and implementation of vaccines against substance abuse outweigh the benefits? Will high-risk people become the subject of more stigma? How might the use of vaccines for substance abuse affect a persons sense of personal responsibility in recovery? Would the prophylactic use of a vaccine against substance abuse undermine the right to future choices or options? Does the use of a vaccine against substance abuse undermine the ability to exercise what might be considered commendable self-restraint? How can therapeutic misconceptions be addressed? Because drug cravings outlast successful vaccination, to what extent should complementary strategies be used to treat drug cravings? Is immunity to addiction an enhancement or a treatment?
Society
Individual people
If a vaccine against cocaine addiction is shown to be safe and highly effective, proposals to include it among the vaccinations routinely administered or even mandated for certain populations are likely to attract considerable interest. For vaccines now licensed, such mandates exist for children in public schools or daycare facilities, college students, members of the military, healthcare professionals, immigrants and other groups. In the specific case of a vaccine against cocaine, mandates directed at specific populations, such as those eligible for parole or those who accept welfare, have already been suggested8. People on trial for drug-related offenses would also be a likely population for such a policy, which would create another potentially attractive alternative to incarceration. Exploring attitudes toward such proposals would probably illuminate underlying assumptions about the nature of addiction and the place of its treatment and prevention in the traditional parameters of public health. The complexity of the choices that would be presented by this novel type of vaccine is underscored by the fact that it is aimed at controlling behavior that is often deemed ethically fraught. Thus, the more fundamental question of whether such vaccines ought to be used prophylactically revolves around two distinct concerns: first, the ontology of substance
522
dependence; and second, the parameters of free will and autonomy. Is addiction infectious? Determining the appropriate use of such a vaccine demands clear understanding of the nature and nosology of substance dependence. How substance dependence is characterized and classified informs the appropriateness of strategies aimed to prevent or treat it. In recent years there has been a marked departure away from the traditional understanding of substance dependence, in which it is framed as a failure of self-control that can lead to addiction, to one in which it is framed as a multivariate neurobiological disease with genetic, anatomical, chemical and physiological determinants9,10. Researchers are beginning to identify a heritable profile of brain abnormalities that may predispose individual people and family members to addiction9. Whereas traditional models of substance dependence place the burden of recovery squarely on the patient and his or her ability to exert self-control, present models emphasize the need for medical interventions that treat addiction as a neurobiological disease9. Thus, some analysts find it appropriate to make drug dependence the focus of a large-scale preventative public-health strategy that includes mandated vaccination. For example, some pro-
ponents of this position have even advocated incorporating vaccines against addiction into compulsory state-mandated statutes for the immunization of 11- and 12-year-old schoolchildren11. The preemptive compulsory use of this sort of vaccine would be unprecedented, because the condition it aims to prevent is not infectious and the vaccine does not confer herd immunity12. In addition to the prevention of disease transmission, the benefit of herd immunity is, historically, a persuasive component of any justification for vaccination mandates. However, a case can be made for herd immunity for this type of vaccine. Although substance dependence is not itself an infectious disease, research has shown that cocaine dependence, for example, places a person at high risk of contracting infectious diseases, including tuberculosis and AIDS13. Moreover, social-contagion theorists have shown that one persons behavior is very likely to influence the behavior of others, which suggests a more infectious disease model for addictive behavior14. It may be appropriate to manage the prevention of addiction akin to the prevention of infectious disease, given such emerging knowledge of risk factors and social contagion. As resistance to the present vaccine mandates illustrates, any eventual policy may require
npg
volume 13 number 6 june 2012 nature immunology
c O m m e n ta r y
A likely argument against the preventative, compulsory use of such a vaccine in healthy people is that such a mandate would inappropriately violate a persons liberty to make victimless or autonomous choices about drug use. However, given the present understanding of addiction and peer-to-peer influence, it is arguable that cocaine use is not victimless and is in a sense contagious9,10,14. This prompts the question of why, given the biomedical model of addiction and identifiable risk factors, preventative strategies should be deemed inappropriate. Who should be vaccinated? The question of how a preventive vaccine of this kind ought to be used must be considered in tandem with the question of who ought to be vaccinated. In the case of the vaccine against cocaine, some parents may argue that it is their moral obligation to protect their children from developing cocaine dependence. Some might argue that targeting at-risk groups or people would be stigmatizing. Another likely opposing view would suggest that the vaccination of children or high-risk people against addiction constricts their rights to future choices and options18, a future in which they ought to be able to experience the effects of cocaine if they so please. The issue of allowing a person to choose to use addictive substances such as cocaine or restricting that possibility carries striking parallels to genetic enhancement. This area has prompted similar debates over biomedical interventions that aim for some phenotypic ideal, behavioral or otherwise19. Such issues bring to light underlying problems that surround how the concept of health should be considered and how this understanding ought to guide medical decision-making. It also requires assessment of whether immunity to addiction should be considered an enhancement or a treatment. The preemptive use of vaccines against addiction raises the further concern that those who are vaccinated will be stigmatized as highrisk pariahs20. The issue of stigmatization has figured prominently in ethical considerations pertaining to the use of potential vaccines against human immunodeficiency virus. It has also affected willingness to participate in clinical trials and to volunteer for vaccination21. It will be imperative to consider preemptively these unwanted outcomes, as well as strategies to attenuate them, in the construction of future research and in the use of the vaccine against cocaine and others like it. Although it is probable that a vaccine against cocaine would be used to aid only those with preexisting dependence, at least initially researchers have suggested that the vaccine could be administered to cocaine-dependent pregnant women to protect fetuses or neonates from the effects of cocaine abuse22. Others have discussed incorporating this vaccine into parole programs for those incarcerated for cocaine userelated infractions17. Requiring or providing incentives for vaccine administration in each of those scenarios would introduce a range of ethical complexities that highlight the difficulties inherent in making policies on the proper use and distribution of such a vaccine. Forcing treatment on either pregnant women or parolees would raise ethical concerns that offering the choice of treatment does not. Safety and therapeutic misconceptions Issues of safety and therapeutic misconceptions further complicate the ethics of vaccines against addiction. Clinical trials of the TA-CD vaccine have brought these issues into focus. Unlike many of the pharmacotherapies now being studied for the treatment of cocaine dependence, the leading candidate vaccine against cocaine does not attenuate cravings or the symptoms of withdrawal. Thus, users might attempt to use excessively high quan-
Vaccines against substance abuse raise many ethical and social questions.
opt-out options and exemptions. Greater attention to when, why and for whom such exemptions may be warranted is needed as part of public-health policy, both for vaccines against infectious disease as well as for potential vaccines against addiction. Parameters of autonomy The ethical principle of respect for autonomy emphasizes that a patients capacity for selfdetermination ought to be respected and that patients should be free from coercion. This principle invites the dilemma of whether such vaccines might in some way do harm by minimizing the extent to which people can make autonomous choices about drug use. However, in dealing with addicted patients, the parameters of this principle can be ambiguous, especially if temporary restrictions on a patients autonomy could create more autonomy in the long term15. The extent to which respect ought to be accorded to a patients choices is particularly unclear when the drug-seeking behavior of an addicted patient is considered biologically coerced10, akratic (a weakness of will) or compulsive16. The fundamental question motivating this debate is whether an addict is truly autonomous. This uncertainty gives rise to the associated issue of whether the principle of respect for autonomy extends to preferences that are the result of addiction. The principle of respect for autonomy may correspondingly suggest that people have the right to exercise the choice not to be addicted, rather than being made biologically immune to this possibility. The concern is that there is some virtue in the choice not to become addicted. For example, people vaccinated against a harmful drug such as cocaine would be deprived of the chance to show commendable self-restraint if they were simply immune to ever feeling the psychoactive effects of cocaine17.
npg
Chris Sharp
Box 2 The TA-CD vaccine

Vaccine vector Effector mechanism Proof of concept Early clinical trials conjugate of cocaine derivative and recombinant cholera toxin B cocaine-specific immunoglobulin G antibodies elicited through vaccination limited cocaine from reaching the brain by up to 80% in rats7 38% of users achieved sufficient serum antibody levels after vaccination to diminish cocaine use6; the cocaine-blockade in these users remained active for up to 2 months Booster vaccinations are needed to maintain sufficient serum concentrations of antibody5,6 no adverse physiological effects on the brain have been observed
523
c O m m e n ta r y
tities of cocaine to overwhelm the vaccines effects5. Patients associating the efficacy and function of this vaccine with that of other vaccines they have previously received may believe that after they are vaccinated, their cocaine dependency will simply dissolve. Instead, the present formulation of the vaccine does not confer sufficiently high serum concentrations of antibody until 2 weeks into treatment. It also requires regular readministration to maintain functional concentrations of antibody6. The need for frequent booster vaccinations diminishes the treatments cost-effectiveness and makes prophylactic use of the present version of the vaccine unlikely. Furthermore, there is considerable variability among individual immune responses to the vaccine5,6. Some vaccinated cocaine users never produce concentrations of antibody adequate to substantially forestall the effects of cocaine. As the desire to use the targeted substance endures, many will continue to attempt to use the substance at higher and higher doses or, in the case of polydrug users, begin to rely more heavily on other substances23. Unintended socioeconomic consequences The possible socioeconomic consequences of a vaccine such as the TA-CD vaccine add to the complexities that face its introduction and use. It is plausible that widespread prophylactic use of this vaccine could trigger a replacement drug market, such that addiction-prone people would simply turn to other substances if the option to use their substance of choice were eliminated through vaccination. As has been observed in trials of vaccines against human immunodeficiency virus, there might be a decrease in the perceived risk of, or an increased interest in, an unsafe behavior if a vaccine tailored to it were to become available, which would thereby increase the likelihood that people would try it21. Providing a quick method for the cessation of substance abuse might remove a major deterrent; that is, the possibility of developing an enduring addiction12. Although a vaccine such as the TA-CD vaccine may effectively address the biological determinants of addiction, it does not address the more pertinent social and environmental determinants of addiction. According to one study, substance use and abuse may partly reflect biological differences, but it is more likely that they can be explained largely by socioeconomic differences, cultural factors, and prejudice and discrimination, both institutional and individual.24 If a potential blockbuster vaccine for a substance-use disorder is released, the causal role of these social factors may come to be overlooked and left unaddressed. Moreover, those who do not choose to be vaccinated or to have their children vaccinated could find themselves the objects of even greater social stigma should they later become addicted25. The connotations of the term vaccine itself introduce the potential for widespread misconceptions about what the intervention is meant to do and the nature of the addiction. This underscores the importance of ensuring that the public is well informed about the limitations of such a vaccine. Conclusions A vaccine against cocaine addiction presents many social, legal and ethical issues. These issues are emblematic of those linked to the development and deployment of other vaccines that belong to the evolving category of immunotherapies that target substance-use disorders. A vaccine will not eliminate the need for other interventions. Complementary therapeutic measures, such as cognitive-behavioral therapy or maintenance programs, would still be needed to attenuate the cravings and drugseeking habits associated with substance dependence. Social programs need to be carefully crafted to address the underlying socioeconomic determinants of drug use and addiction. In investigating such vaccines, researchers must consider the principle of clinical equipoise, designing equitable and sensitive clinical studies that diminish undesirable outcomes through careful implementation, multilevel monitoring and continuous evaluation. Any anti-addiction vaccination strategy should be coupled with a comprehensive educational scheme to minimize therapeutic misconception and to maximize adherence to treatment schedules. The temptation to use a potential vaccine to solve the immense problem of abuse and addiction is enormous, but the ethical challenges that would accompany any proposal for widespread or mandatory use are daunting. The opprobrium and suspicion that vaccines often seem to attract portends intensified controversy when the targets are behavioral phenotypes. It is vital to address these issues now, lest some of the battles, misperceptions and fear-mongering that have dominated vaccine policy too much resurface in the ethically fraught area of addiction treatment.
ACKNOWLEDGMENTS We thank those who participated in early discussions of this topic, including S.D. Halpern, N. Jones and J. Powers; and K. Buckley for editorial support. Supported by the Thomas Scattergood Behavioral Health Foundation and the Scattergood Program for the Applied Ethics of Behavioral Healthcare at the University of Pennsylvania Department of Medical Ethics and Health Policy. COMPETING FINANCIAL INTERESTS The authors declare no competing financial interests.
1. Orson, F.m. et al. Ann. NY Acad. Sci. 1141, 257269 (2008). 2. U.S. Department of Health and Human Services. Substance abuse and mental Health Services administration, Drug abuse Warning network, 2009: national estimates of Drug-related emergency Department Visits. HHS Publication (Sma) 11-4659, DaWn Series D-35 (Substance abuse and mental Health Services administration, rockville, maryland, 2011). 3. Volkow, n.D. & Skolnick, P. Neuropsychopharmacology 37, 290292 (2012). 4. OBrien, m.S. & anthony, J.c. Neuropsychopharmacology 30, 10061018 (2005). 5. Haney, m., Gunderson, e.W., Jiang, H., collins, e.D. & Foltin, r.W. Biol. Psychiatry 67, 5965 (2010). 6. martell, B.a. et al. Arch. Gen. Psychiatry 66, 1116 1123 (2009). 7. carrera, m.r. et al. Nature 378, 727730 (1995). 8. Stevenson, D. Rutgers J. L. & Urb. Poly 3, 4, n. 170 (2005). 9. ersche, K.D. et al. Science 335, 601604 (2012). 10. Hyman, S.e. Am. J. Bioeth. 7, 811 (2007). 11. Osburn, a. Cleveland State Law Rev. 55, 159188 (2008). 12. ashcroft, r.e. & Franey, c. J. Med. Ethics 30, 341343 (2004). 13. Friedman, H., Pross, S. & Klein, t.W. Fems. Immunol. Med. Microbiol. 47, 330342 (2006). 14. christakis, n.a. & Fowler, J.H. Connected: The Surprising Power of Our Social Networks and How They Shape Our Lives, 1st edn. (Little, Brown and co./ Hachette Book Group, new york, 2009). 15. caplan, a. Addiction 103, 19191921 (2008). 16. yaffe, G. Philos. Public Aff. 30, 178221 (2001). 17. Hall, W. & carter, L. J. Med. Ethics 30, 337340 (2004). 18. Hasman, a. & Holm, S. J. Med. Ethics 30, 344345 (2004). 19. caplan, a. in Human Enhancement (eds. Savelescu, J. & Bostrum, n.) 199210 (Oxford University Press, Oxford, 2009). 20. newman, P.a., Daley, a., Halpenny, r. & Loutfy, m. Vaccine 26, 10911097 (2008). 21. newman, P.a., Duan, n., Kakinami, L. & roberts, K. Vaccine 26, 25282536 (2008). 22. Kosten, t. & Owens, S.m. Pharmacol. Ther. 108, 7685 (2005). 23. Kosten, t.r. et al. Vaccine 20, 11961204 (2002). 24. Buka, S.L. Public Health Rep. 117, S118S125 (2002). 25. Francis, a. Sociol. Health. Illn. 20, 116 (2012).
npg
524
volume 13 number 6 june 2012 nature immunology
nEWS AnD vIEWS
DOCKing innate to adaptive signaling for persistent antibody production

Markus Werner & Hassan Jumaa
Protection against recurrent infections requires the generation of memory B cells and persistent antibody production. An adaptor complex of DOCK8-MyD88-Pyk2 now links signaling via TLR9 to activation of the transcription factor STAT3 and the establishment of serological memory.
he adaptor DOCK8 has become a focus of immunological research since a mouse genetic screen identified DOCK8 as an essential element for the formation of long-lasting humoral immunity1. Loss-of-function mutations in the gene encoding DOCK8 are associated with combined immunodeficiency in humans2. In addition to greater susceptibility to a diversity of recurrent severe infections, DOCK8-deficient patients have higher concentrations of serum immunoglobulin E (IgE) and severe allergic symptoms2. In the present issue of Nature Immunology, Jabara et al. describe a role for DOCK8 as an adaptor that connects via Toll-like receptor 9 (TLR9) to the signaling cascade of the B cell antigen receptor (BCR)3. DOCK8 belongs to the DOCK180 family of atypical guanine nucleotideexchange factors, which characteristically contain two DOCKhomology regions, DHR-1 and DHR-2 (ref. 4). The guanine nucleotideexchange activity is located in the DHR-2 domain, which binds to and activates GTPases of the Rac-Rho family, whereas DHR-1 binds to phosphatidylinositol-3,4,5-trisphosphate (PtdIns(3,4,5)P3) and mediates the recruitment of DOCK proteins to the plasma membrane5. Thus, DOCK8 function requires the activity of phosphatidylinositol-3-OH kinase, the enzyme that generates PtdIns(3,4,5)P3. Interestingly, the immunodeficiency caused by the loss of DOCK8 resembles that caused by mutations in the gene encoding the
Markus Werner and Hassan Jumaa are with the Centre for Biological Signaling Studies, AlbertLudwigs University of Freiburg and Max-PlanckInstitute of Immunobiology and Epigenetics, Freiburg, Germany. e-mail: jumaa@immunbio.mpg.de
transcription factor STAT3 (ref. 6). However, DOCK8-deficient patients develop an autosomal recessive form of hyper-IgE syndrome, whereas the hyper-IgE syndrome induced by mutations in STAT3 is an autosomal dominant form6. Despite the obvious central role
of DOCK8 in the immune response and the increasing number of patients with mutations in DOCK8, the molecular mechanism of how DOCK8 is involved in the formation of long-lasting humoral immunity has remained obscure. Moreover, as mutations in DOCK8
BCR CD19 Ig- Ig- Cytoplasm

PI(3)K
IL-4R
npg
TLR9 activation by DNA
MyD88 Pyk2 DOCK8
Endosome
Src Syk STAT6
STAT3
Class switch to IgE
Survival and Ig production
Figure 1 DOCK8 connects TLR signaling to the activation of STAT3. Activation of TLR9 by its ligand CpG induces the formation and stabilization of a trimeric signaling complex consisting of MyD88, Pyk2 and DOCK8. This leads to activation of Src kinases that, together with Syk, phosphorylate and activate STAT3 (ref. 3), whose target genes are linked to the survival of B cells and the production of immunoglobulins. DOCK8 is recruited to the plasma membrane by binding via its DHR-1 domain to PtdIns(3,4,5)P3, the product of phosphatidylinositol-3-OH kinase (PI(3)K) activity5. Enhanced and/or uncontrolled switching to IgE is reported in the absence of DOCK8 or STAT3 function6, which suggests that signaling through DOCK8-STAT3 interferes, in an as-yet-unknown manner, with the activation of CSR to IgE dependent on the receptor for interleukin 4 (IL-4R). Ig- and Ig-, immunoglobulin -chain and -chain.
nature immunology volume 13 number 6 June 2012
525
nEWS AnD VIEWS

or STAT3 result in similar patterns of disease, the question arises of whether DOCK8 and STAT3 are components of the same signaling cascade. Jabara et al. now show that DOCK8 is part of a trimeric signaling complex that contains DOCK8, the adaptor MyD88 and the kinase Pyk2 (ref. 3). This complex activates STAT3 after the induction of TLR9 signaling by its ligand CpG. Those results suggest that the activation of TLR9 stabilizes a pre-existing trimeric complex and that such stabilization is required for the phosphorylation of DOCK8 by Pyk2, thereby creating binding sites for the Src-kinase Lyn, which then activates the tyrosine kinase Syk in a not-yet-well-understood manner. Such activation of Lyn downstream of DOCK8 integrates TLR-signals into the BCR signaling cascade and connects DOCK8 to the activation of STAT3, as both Lyn and Syk have been shown to phosphorylate and activate STAT3 (refs. 3,7,8). This newly discovered link between DOCK8 and STAT3 explains the similar pattern of disease in DOCK8-deficient patients and STAT3-defective patients who have defective long-lasting antibody responses while having higher concentrations of serum IgE6. Together with data already available, the present study3 suggests that the signaling cascade that involves DOCK8 and STAT3 regulates the generation or survival of long-lived memory B and antibody-secreting cells. Differentiation into memory B cells occurs in the secondary lymphoid organs in specialized structures known as germinal centers, in which mature B cells typically undergo classswitch recombination (CSR) from IgM to IgG, IgA or IgE and improve their recognition of antigen by somatic hypermutation (SHM) of their immunoglobulin genes9. SHM and CSR are mediated by the cytosine deaminase AID10, and in agreement with published data1, Jabara et al. show that DOCK8 is not required for initial BCR signaling events or the induction of AID3. In fact, SHM and initial IgG responses are induced in the absence of DOCK8 (ref. 1). That result suggests that DOCK8 is not essential for the generation of switched, IgG-expressing B cells but instead is needed for their long-term survival. The finding that the initial steps of BCR signaling, when antigen concentrations are probably high, are unaffected suggests that the DOCK8-mediated integration of innate receptor signaling into BCR signaling becomes crucial for the later survival of memory B cells when antigen concentrations decrease. However, although this DOCK8-mediated enhancement of BCR signaling might explain the role of DOCK8 in the survival of switched antigen-specific B cells, it raises the question how this leads to the higher IgE serum concentrations of the DOCK8-deficient patients. A simple answer to this question may be that DOCK8STAT3 signaling prevents switching to IgE. For the generation of different isotypes, CSR is regulated by various stimuli, of which the activation of STAT6 induced by the receptor for interleukin 4 is key for CSR to IgE11. The higher IgE concentrations noted in the absence of the function of DOCK8 or STAT3 suggest that activation of DOCK8-STAT3 signaling interferes, in an as-yet-unknown manner, with the activation of switching to IgE induced by the receptor for interleukin 4 during the generation of antibody responses to pathogens (Fig. 1). In this scenario, recurrent infections in DOCK8-deficient patients may lead to repeated switching to IgE, and the chronic formation of IgE-secreting cells could result in sustained IgE production in these patients. Although it is not addressed in this study3, it is possible that the adaptor function of DOCK8 may also integrate other receptors of innate immunity into BCR signaling. By showing that DOCK8 augments BCR signaling and results in STAT3 activation, Jabara et al. provide critical insights into how innate and adaptive immune signaling cascades are interconnected to ensure the long-lived survival of memory B cells and persistent antibody responses3. Thus, characterizing the molecular links that coordinate the activity of these signaling cascades is crucial for understanding the molecular basis of the formation of persistent immunity and the defects observed in patients with immunodeficiency.
COMPETING FINANCIAL INTERESTS The authors declare no competing financial interests.
1. Randall, K.L. et al. Nat. Immunol. 10, 12831291 (2009). 2. Zhang, Q. et al. N. Engl. J. Med. 361, 20462055 (2009). 3. Jabara, H.H. et al. Nat. Immunol. 13, 612620 (2012). 4. Cte, J.F. & Vuori, K. Trends Cell Biol. 17, 383393 (2007). 5. Cte, J.F., Motoyama, A.B., Bush, J.A. & Vuori, K. Nat. Cell Biol. 7, 797807 (2005). 6. Su, H.C. Curr. Opin. Allergy Clin. Immunol. 10, 515520 (2010). 7. Turkson, J. et al. Mol. Cell Biol. 18, 25452552 (1998). 8. Uckun, F.M., Qazi, S., Ma, H., Tuel-Ahlgren, L. & Ozer, Z. Proc. Natl. Acad. Sci. USA 107, 29022907 (2010). 9. McHeyzer-Williams, L.J. & McHeyzer-Williams, M.G. Annu. Rev. Immunol. 23, 487513 (2005). 10. Honjo, T., Kinoshita, K. & Muramatsu, M. Annu. Rev. Immunol. 20, 165196 (2002). 11. Geha, R.S., Jabara, H.H. & Brodeur, S.R. Nat. Rev. Immunol. 3, 721732 (2003).
npg
Dont mess with ERAAP!

Jonathan W Yewdell & Xiuju Lu
Immunosurveillance monitors subversion of the endoplasmic reticulum aminopeptidase ERAAP. ERAAP-deficient cells are killed by T cells that recognize nonclassical major histocompatibility complex class I Qa-1 molecules presenting peptides generated in the absence of ERAAP.
n a classic example of scientific myopia, the genes encoding the many nonclassical major histocompatibility complex (MHC) class I molecules expressed by vertebrates were once
Jonathan W. Yewdell and Xiuju Lu are with the Laboratory of viral Diseases, national Institute of Allergy and Infectious Diseases, Bethesda, Maryland, USA. e-mail: jyewdell@nih.gov
considered a junkyard, encoding molecules that could not pass muster as classical class I molecules, which function mainly to present self and foreign peptides to CD8+ T cells. Truth, however, eventually comes into focus. Junk genes encoding MHC molecules (like junk DNA) are anything but junk, as they encode critical participants in the recognition of viruses and tumors by natural killer (NK) cells, the presentation of lipid antigens, antibody trafficking, iron metabolism
and maternal-fetal tolerance1. In this issue of Nature Immunology, Nagarajan et al. add another important function to that list2. They report that the mouse nonclassical MHC class I molecule Qa-1 presents a peptide generated when the endoplasmic reticulum aminopeptidase ERAAP (ERAP) is inhibited, which allows a substantial subset of CD8+ T cells to monitor the activity of this enzyme dedicated to processing antigens presented by classical MHC class I.
526
volume 13 number 6 June 2012 nature immunology
nEWS AnD VIEWS

or STAT3 result in similar patterns of disease, the question arises of whether DOCK8 and STAT3 are components of the same signaling cascade. Jabara et al. now show that DOCK8 is part of a trimeric signaling complex that contains DOCK8, the adaptor MyD88 and the kinase Pyk2 (ref. 3). This complex activates STAT3 after the induction of TLR9 signaling by its ligand CpG. Those results suggest that the activation of TLR9 stabilizes a pre-existing trimeric complex and that such stabilization is required for the phosphorylation of DOCK8 by Pyk2, thereby creating binding sites for the Src-kinase Lyn, which then activates the tyrosine kinase Syk in a not-yet-well-understood manner. Such activation of Lyn downstream of DOCK8 integrates TLR-signals into the BCR signaling cascade and connects DOCK8 to the activation of STAT3, as both Lyn and Syk have been shown to phosphorylate and activate STAT3 (refs. 3,7,8). This newly discovered link between DOCK8 and STAT3 explains the similar pattern of disease in DOCK8-deficient patients and STAT3-defective patients who have defective long-lasting antibody responses while having higher concentrations of serum IgE6. Together with data already available, the present study3 suggests that the signaling cascade that involves DOCK8 and STAT3 regulates the generation or survival of long-lived memory B and antibody-secreting cells. Differentiation into memory B cells occurs in the secondary lymphoid organs in specialized structures known as germinal centers, in which mature B cells typically undergo classswitch recombination (CSR) from IgM to IgG, IgA or IgE and improve their recognition of antigen by somatic hypermutation (SHM) of their immunoglobulin genes9. SHM and CSR are mediated by the cytosine deaminase AID10, and in agreement with published data1, Jabara et al. show that DOCK8 is not required for initial BCR signaling events or the induction of AID3. In fact, SHM and initial IgG responses are induced in the absence of DOCK8 (ref. 1). That result suggests that DOCK8 is not essential for the generation of switched, IgG-expressing B cells but instead is needed for their long-term survival. The finding that the initial steps of BCR signaling, when antigen concentrations are probably high, are unaffected suggests that the DOCK8-mediated integration of innate receptor signaling into BCR signaling becomes crucial for the later survival of memory B cells when antigen concentrations decrease. However, although this DOCK8-mediated enhancement of BCR signaling might explain the role of DOCK8 in the survival of switched antigen-specific B cells, it raises the question how this leads to the higher IgE serum concentrations of the DOCK8-deficient patients. A simple answer to this question may be that DOCK8STAT3 signaling prevents switching to IgE. For the generation of different isotypes, CSR is regulated by various stimuli, of which the activation of STAT6 induced by the receptor for interleukin 4 is key for CSR to IgE11. The higher IgE concentrations noted in the absence of the function of DOCK8 or STAT3 suggest that activation of DOCK8-STAT3 signaling interferes, in an as-yet-unknown manner, with the activation of switching to IgE induced by the receptor for interleukin 4 during the generation of antibody responses to pathogens (Fig. 1). In this scenario, recurrent infections in DOCK8-deficient patients may lead to repeated switching to IgE, and the chronic formation of IgE-secreting cells could result in sustained IgE production in these patients. Although it is not addressed in this study3, it is possible that the adaptor function of DOCK8 may also integrate other receptors of innate immunity into BCR signaling. By showing that DOCK8 augments BCR signaling and results in STAT3 activation, Jabara et al. provide critical insights into how innate and adaptive immune signaling cascades are interconnected to ensure the long-lived survival of memory B cells and persistent antibody responses3. Thus, characterizing the molecular links that coordinate the activity of these signaling cascades is crucial for understanding the molecular basis of the formation of persistent immunity and the defects observed in patients with immunodeficiency.
1. Randall, K.L. et al. Nat. Immunol. 10, 12831291 (2009). 2. Zhang, Q. et al. N. Engl. J. Med. 361, 20462055 (2009). 3. Jabara, H.H. et al. Nat. Immunol. 13, 612620 (2012). 4. Cte, J.F. & Vuori, K. Trends Cell Biol. 17, 383393 (2007). 5. Cte, J.F., Motoyama, A.B., Bush, J.A. & Vuori, K. Nat. Cell Biol. 7, 797807 (2005). 6. Su, H.C. Curr. Opin. Allergy Clin. Immunol. 10, 515520 (2010). 7. Turkson, J. et al. Mol. Cell Biol. 18, 25452552 (1998). 8. Uckun, F.M., Qazi, S., Ma, H., Tuel-Ahlgren, L. & Ozer, Z. Proc. Natl. Acad. Sci. USA 107, 29022907 (2010). 9. McHeyzer-Williams, L.J. & McHeyzer-Williams, M.G. Annu. Rev. Immunol. 23, 487513 (2005). 10. Honjo, T., Kinoshita, K. & Muramatsu, M. Annu. Rev. Immunol. 20, 165196 (2002). 11. Geha, R.S., Jabara, H.H. & Brodeur, S.R. Nat. Rev. Immunol. 3, 721732 (2003).
npg
Dont mess with ERAAP!

Jonathan W Yewdell & Xiuju Lu
Immunosurveillance monitors subversion of the endoplasmic reticulum aminopeptidase ERAAP. ERAAP-deficient cells are killed by T cells that recognize nonclassical major histocompatibility complex class I Qa-1 molecules presenting peptides generated in the absence of ERAAP.
n a classic example of scientific myopia, the genes encoding the many nonclassical major histocompatibility complex (MHC) class I molecules expressed by vertebrates were once
Jonathan W. Yewdell and Xiuju Lu are with the Laboratory of viral Diseases, national Institute of Allergy and Infectious Diseases, Bethesda, Maryland, USA. e-mail: jyewdell@nih.gov
considered a junkyard, encoding molecules that could not pass muster as classical class I molecules, which function mainly to present self and foreign peptides to CD8+ T cells. Truth, however, eventually comes into focus. Junk genes encoding MHC molecules (like junk DNA) are anything but junk, as they encode critical participants in the recognition of viruses and tumors by natural killer (NK) cells, the presentation of lipid antigens, antibody trafficking, iron metabolism
and maternal-fetal tolerance1. In this issue of Nature Immunology, Nagarajan et al. add another important function to that list2. They report that the mouse nonclassical MHC class I molecule Qa-1 presents a peptide generated when the endoplasmic reticulum aminopeptidase ERAAP (ERAP) is inhibited, which allows a substantial subset of CD8+ T cells to monitor the activity of this enzyme dedicated to processing antigens presented by classical MHC class I.
526
nEWS AnD VIEWS

affinity and hence their stability as cell-surface complexes with MHC class I molecules14. The altered repertoire elicits a robust CD8+ T cell response when wild-type mice are immunized with ERAAP-deficient cells15. What do those T cells recognize? That seemingly esoteric question has yielded an interesting and important answer. Nagarjan et al. have isolated CD8+ T cells that recognize Qa-1 presenting a TAP-dependent ligand generated when ERAAP deficiency is induced genetically or chemically2. This recognition does not require the expression of classical MHC class I molecules, the sole source of Qdm. To identify antigenic peptides, the authors heroically screen 2,880 plasmid pools comprising a cDNA library derived from ERAAP-deficient splenic dendritic cells. They transfect each pool into Qa-1expressing L929 mouse fibroblasts and assess its ability to activate a Qa-1-restricted T cell hybridoma that recognizes ERAAP-deficient cells. They identify three positive pools, each containing a plasmid encoding the hypothetical protein Fam49b. They use a series of truncation plasmids to identify the nine-residue peptide FYAEATPML (FL9) located at positions 190198 in Fam49b and find that synthetic FL9 peptide activates T cells in a Qa-1-dependent manner at subnanomolar concentrations. The isolation of naturally processed peptides by HPLC shows that FL9 elutes together with the natural peptide, whose recovery requires the presence of Qa-1. To identify cells that recognize complexes of Qa-1 and FL9 (Qa-1FL9), Nagarjan et al. find that in contrast to the staining of NK cells by Qa-1bQdm tetramers, NK cells are not stained by Qa-1FL9 tetramers2. Instead, Qa-1FL9 tetramers stain ~20% of interferon-secreting CD8+ T cells induced by the stimulation of splenocytes with ERAAP-deficient cells, which demonstrates its immunodominant status but also indicates the recognition of other potential peptide ligands. Remarkably, the authors identify CD8+ T cells stained with Qa-1FL9, most expressing the memory markers CD44 and CD122, in naive C57BL/6 spleens at relatively high frequency. Indicative of their probable function, CD8+ T cells specific for Qa-1FL9 efficiently kill ERAAP-deficient cells or tumor cells treated with ERAAP inhibitors. Nagarajan et al. clearly show that immunosurveillance monitors ERAAP activity via the presentation by Qa-1 of at least one peptide whose presentation is eradicated by ERAAP. Although their findings remain to be extended to humans and other vertebrates, several aspects of this system support the proposal of its generality. First, Qa-1 and HLA-E demonstrate considerable functional and structural homology16 despite their limited amino acid homology. Second, the gene encoding Fam49
527
QFL-specific CD8+ T cell
No recognition Cell surface AA 190198 FAM49b
Recognition
Proteasome Golgi
Peptide precursors
TAP
ERAAP
Qa-1 Endoplasmic recticulum
2m
ERAAP-WT
ERAAP-KO
Figure 1 Model of ERAAP immunosurveillance. Like most MHC class I peptide ligands, a peptide of amino acids (AA) 190198 of Fam49b (Fam49b(190198), or an extended precursor) is generated in the cytosol and transported into the endoplasmic reticulum by TAP in ERAAP-sufficient (ERAAP-WT) cells (left). Cells that synthesize the MHC class I molecule Qa-1 do not present Fam49b(190198) when ERAAP is expressed, either because of its possible destruction by ERAAP in the endoplasmic reticulum, as shown here, or because of a less direct mechanism (for example, losing in a competition with ERAAP-generated peptides). In the absence of ERAAP (ERAAP-deficient (ERAAP-KO) cells; right), Fam49b(190198) is delivered to Qa-1 and the Qa-1FL9 complex is delivered to the cell surface, where it activates specific CD8+ T cells that kill the cell. This prevents pathogens or tumor cells from interfering with ERAAP function to avoid immunosurveillance. It might also be used for other purposes by the immune system, as memory T cells are present in young, specific pathogenfree mice. QFL, Qa-1FL9; 2m, 2-microglobulin.
npg
Qa-1 was discovered in 1978 as a lymphocyte marker, but its first distinct function was identified decades later, as a ligand for the NK cell receptor heterodimer CD94-NKG2 (also present on some T cells), on the basis of its presentation of a highly conserved peptide (Qdm3) present in the endoplasmic reticulum insertion sequence of two of the three classical class I molecules (H-2D and H-2L, but not H-2K)4. As the presentation of Qdm requires the entire classical MHC class I pathway (proteasomes, the transporter associated with antigen processing (TAP) and tapasin; Fig. 1), Qa-1 serves as a sentinel that directs NK cells to destroy potentially dangerous cells whose MHC class I pathway has been manipulated by pathogens or oncogenesis. Humans have a nearly exact functional homolog in HLA-E, which interacts with the homologous NK cell receptor by presenting a Qdm-like peptide from the signal sequence of many classical HLA class I molecules. HLA-E is expressed
by T cells, B cells, NK cells and endothelial cells and is upregulated by proinflammatory cytokines, which also induce the secretion of a soluble form of HLA-E that protects cells from lysis by NK cells5. Qa-1 also presents foreign and non-Qdm self peptides to CD8+ T cells6. In the case of activated autoreactive CD4+ cells, the presentation of self peptide leads to CD8+ T cellmediated suppression of autoimmunity7. Early investigation into the proteolytic processing of TAP-independent antigens delivered to the endoplasmic reticulum showed that cells robustly trim amino termini8. Biochemical studies led to the identification and cloning of interferon-responsive genes encoding ERAAP in mice and ERAP1 and ERAP2 in humans911. The first ascribed function elucidated for ERAAP involved the shedding of cytokine receptors12, and ERAAP polymorphism has been linked to human autoimmune diseases13. ERAAP-deficient mice present peptides with amino-terminal extensions that diminish their
nEWS AnD VIEWS

is highly conserved among vertebrates, and BLAST analysis shows that the FL9 peptide is completely conserved in vertebrates as diverse as monotremes, birds, reptiles, amphibians and fish. Nagarajan et al. suggest that ERAAP monitoring is used for eradicating tumor cells that downregulate ERAAP to avoid the presentation of immunogenic tumor-specific peptides and cells infected with human cytomegalovirus2, which encodes a microRNA that targets ERAP1 to diminish viral peptide presentation17. As is typical, this groundbreaking study raises many intriguing questions. Is FL9 actually degraded by ERAAP, or is it regulated by a less-direct mechanism? Are the genes that encode Qa-1, Fam49a and ERAAP regulated coordinately in various tissues to coordinate immunosurveillance? Given the involvement of Qa-1 in the suppression of autoimmune CD4+ T cells (and the link between ERAAP polymorphism and autoimmunity), the latter question is particularly relevant for various CD4+ T cell subsets. What are the other peptide ligands presented by Qa-1, and how do they increase the functionality of the system? Why are activated Qa-1FL9specific CD8+ T cells present in specific pathogenfree mice? Do they control commensal organisms? Are they involved in nontraditional functions, such as organ development or tissue remodeling, by eradicating unwanted (junk) cells that signal their status via ERAAP modulation? Could ERAAP inhibitors be used clinically to recruit Qa-1FL9specific CD8+ T cells to eliminate tumor cells or autoimmune T cells? Finally, in this era of increasing direction of research funding toward practical ends, this remarkable study provides a vivid reminder that curiosity-driven research into esoteric questions predictably leads to unpredictable translational applications. Now, that is classic!
1. Rodgers, J.R. & Cook, R.G. Nat. Rev. Immunol. 5, 459471 (2005). 2. nagarajan, n.A., Gonzalez, F. & Shastri, n. Nat. Immunol. 13, 579586 (2012). 3. Aldrich, C.J., Rodgers, J.R. & Rich, R.R. Immunogenetics 28, 334344 (1988). 4. van Hall, T., Oliveira, C.C., Joosten, S.A. & Ottenhoff, T.H. Microbes Infect. 12, 910918 (2010). 5. Coupel, S. et al. Blood 109, 28062814 (2007). 6. Soloski, M.J. & Metcalf, E.S. Microbes Infect. 3, 12491259 (2001). 7. Kim, H.J. & Cantor, H. Semin. Immunol. 23, 446452 (2011). 8. Snyder, H.L., Yewdell, J.W. & Bennink, J.R. J. Exp. Med. 180, 23892394 (1994). 9. Saric, T. et al. Nat. Immunol. 3, 11691176 (2002). 10. Serwold, T., Gonzalez, F., Kim, J., Jacob, R. & Shastri, n. Nature 419, 480483 (2002). 11. Saveanu, L. et al. Nat. Immunol. 6, 689697 (2005). 12. Cui, X. et al. J. Clin. Invest. 110, 515526 (2002). 13. Haroon, n. & Inman, R.D. Nat. Rev. Rheumatol. 6, 461467 (2010). 14. Blanchard, n. et al. J. Immunol. 184, 30333042 (2010). 15. Hammer, G.E., Gonzalez, F., James, E., nolla, H. & Shastri, n. Nat. Immunol. 8, 101108 (2007). 16. Zeng, L. et al. J. Immunol. 188, 302310 (2012). 17. Kim, S. et al. Nat. Immunol. 12, 984991 (2011).
Death diverted, but to what?

Gretta L Stritesky & Kristin A Hogquist
The role of the coreceptor CD28 in thymic clonal deletion has been controversial. New evidence suggests that CD28 deficiency impairs the clonal deletion of self-reactive T cells but also results in their developmental diversion to an anergic lineage that ends up in the gut.
t has been long known that T cells that express T cell antigen receptors (TCRs) with low to medium affinity for self peptides become positively selected in the thymus and those that express TCRs with high affinity undergo clonal deletion. Positive selection of T cells does not require the TCR coreceptor CD28, but the role of CD28 in clonal deletion has been controversial. Costimulation through CD28 can clearly enhance the death of thymocytes in vitro, and thymic antigen-presenting cells have high expression of B7-1 (CD80) and B7-2 (CD86), the ligands for CD28. However, CD28-deficient mice do not accumulate more self-reactive T cells in vivo, as shown in several different model systems1. Nonetheless, over the past decade it has become clear that highaffinity TCR interactions can sometimes lead to outcomes other than death. For example, strong TCR signals are required for the development of invariant natural killer T cells (iNKT
Gretta L. Stritesky and Kristin A. Hogquist are in the Department of Laboratory Medicine & Pathology and the Center for Immunology, University of Minnesota, Minneapolis, Minnesota, USA. e-mail: hogqu001@umn.edu
cells), CD4+ regulatory T cells (Treg cells) and the precursors of intraepithelial lymphocytes (IELs) of the gut, a process collectively known as agonist selection2 (Fig. 1a). Interestingly, costimulation via CD28 is known to be required for the positive selection of iNKT cells and Treg cells, as are other unique factors, such as the immunomodulatory receptor SLAM and its adaptor SAP for iNKT cells, and interleukin 2 (IL-2) for Treg cells. Together such observations suggest that costimulatory signals do not always promote death2. In a study published in this issue of Nature Immunology, Pobezinsky and colleagues revisit the role of CD28 in thymic selection through the analysis of mice genetically deficient in CD28 or deficient in both B7-1 and B7-2 (called B7-deficient mice here)3. They find that unlike iNKT cells or Treg cells, IELs expressing homodimers of CD8 (CD8+) do not require costimulation via CD28. Instead, they report that CD28- or B7-deficient mice have considerably more gut IELs and thymic IEL precursors, a process they call clonal diversion (Fig. 1b). By highlighting this alternative fate, these data help explain the controversy surrounding the role of costimulation via CD28 in clonal deletion.
The study by Pobezinsky and colleagues provides compelling evidence that autoreactive T cell clones that would have been deleted in normal mice can instead be diverted to a CD4CD8 double-negative (DN) T cell population in costimulation-deficient (CD28or B7-deficient) mice3. T cell clones expressing TCRs containing -chain variable region 5 (V5) or V11 are normally deleted from the conventional CD4+ or CD8+ T cell pool of BALB/c mice because of reactivity with endogenous superantigens Mtv-6, Mtv-8 and Mtv-9; as far as the thymus is concerned, these are seen as autoreactive cells. Notably, however, costimulation-deficient mice have a greater frequency of V5+ and V11+ T cell clones among TCR+ DN cells. Similarly, the thymus of mice overexpressing the prosurvival factors Bcl-2 or Mcl-1 shows enrichment for TCR+ DN cells expressing Mtv superantigen reactive V+ TCRs, which further supports the idea that such clones are normally deleted but are diverted into TCR+ DN cells when they cannot undergo apoptosis. To determine if such diverted cells are IEL precursors, the authors use adoptive-transfer studies. They find that TCR+ DN cells from
npg
528
nEWS AnD VIEWS

is highly conserved among vertebrates, and BLAST analysis shows that the FL9 peptide is completely conserved in vertebrates as diverse as monotremes, birds, reptiles, amphibians and fish. Nagarajan et al. suggest that ERAAP monitoring is used for eradicating tumor cells that downregulate ERAAP to avoid the presentation of immunogenic tumor-specific peptides and cells infected with human cytomegalovirus2, which encodes a microRNA that targets ERAP1 to diminish viral peptide presentation17. As is typical, this groundbreaking study raises many intriguing questions. Is FL9 actually degraded by ERAAP, or is it regulated by a less-direct mechanism? Are the genes that encode Qa-1, Fam49a and ERAAP regulated coordinately in various tissues to coordinate immunosurveillance? Given the involvement of Qa-1 in the suppression of autoimmune CD4+ T cells (and the link between ERAAP polymorphism and autoimmunity), the latter question is particularly relevant for various CD4+ T cell subsets. What are the other peptide ligands presented by Qa-1, and how do they increase the functionality of the system? Why are activated Qa-1FL9specific CD8+ T cells present in specific pathogenfree mice? Do they control commensal organisms? Are they involved in nontraditional functions, such as organ development or tissue remodeling, by eradicating unwanted (junk) cells that signal their status via ERAAP modulation? Could ERAAP inhibitors be used clinically to recruit Qa-1FL9specific CD8+ T cells to eliminate tumor cells or autoimmune T cells? Finally, in this era of increasing direction of research funding toward practical ends, this remarkable study provides a vivid reminder that curiosity-driven research into esoteric questions predictably leads to unpredictable translational applications. Now, that is classic!
1. Rodgers, J.R. & Cook, R.G. Nat. Rev. Immunol. 5, 459471 (2005). 2. nagarajan, n.A., Gonzalez, F. & Shastri, n. Nat. Immunol. 13, 579586 (2012). 3. Aldrich, C.J., Rodgers, J.R. & Rich, R.R. Immunogenetics 28, 334344 (1988). 4. van Hall, T., Oliveira, C.C., Joosten, S.A. & Ottenhoff, T.H. Microbes Infect. 12, 910918 (2010). 5. Coupel, S. et al. Blood 109, 28062814 (2007). 6. Soloski, M.J. & Metcalf, E.S. Microbes Infect. 3, 12491259 (2001). 7. Kim, H.J. & Cantor, H. Semin. Immunol. 23, 446452 (2011). 8. Snyder, H.L., Yewdell, J.W. & Bennink, J.R. J. Exp. Med. 180, 23892394 (1994). 9. Saric, T. et al. Nat. Immunol. 3, 11691176 (2002). 10. Serwold, T., Gonzalez, F., Kim, J., Jacob, R. & Shastri, n. Nature 419, 480483 (2002). 11. Saveanu, L. et al. Nat. Immunol. 6, 689697 (2005). 12. Cui, X. et al. J. Clin. Invest. 110, 515526 (2002). 13. Haroon, n. & Inman, R.D. Nat. Rev. Rheumatol. 6, 461467 (2010). 14. Blanchard, n. et al. J. Immunol. 184, 30333042 (2010). 15. Hammer, G.E., Gonzalez, F., James, E., nolla, H. & Shastri, n. Nat. Immunol. 8, 101108 (2007). 16. Zeng, L. et al. J. Immunol. 188, 302310 (2012). 17. Kim, S. et al. Nat. Immunol. 12, 984991 (2011).
Death diverted, but to what?

Gretta L Stritesky & Kristin A Hogquist
The role of the coreceptor CD28 in thymic clonal deletion has been controversial. New evidence suggests that CD28 deficiency impairs the clonal deletion of self-reactive T cells but also results in their developmental diversion to an anergic lineage that ends up in the gut.
t has been long known that T cells that express T cell antigen receptors (TCRs) with low to medium affinity for self peptides become positively selected in the thymus and those that express TCRs with high affinity undergo clonal deletion. Positive selection of T cells does not require the TCR coreceptor CD28, but the role of CD28 in clonal deletion has been controversial. Costimulation through CD28 can clearly enhance the death of thymocytes in vitro, and thymic antigen-presenting cells have high expression of B7-1 (CD80) and B7-2 (CD86), the ligands for CD28. However, CD28-deficient mice do not accumulate more self-reactive T cells in vivo, as shown in several different model systems1. Nonetheless, over the past decade it has become clear that highaffinity TCR interactions can sometimes lead to outcomes other than death. For example, strong TCR signals are required for the development of invariant natural killer T cells (iNKT
Gretta L. Stritesky and Kristin A. Hogquist are in the Department of Laboratory Medicine & Pathology and the Center for Immunology, University of Minnesota, Minneapolis, Minnesota, USA. e-mail: hogqu001@umn.edu
cells), CD4+ regulatory T cells (Treg cells) and the precursors of intraepithelial lymphocytes (IELs) of the gut, a process collectively known as agonist selection2 (Fig. 1a). Interestingly, costimulation via CD28 is known to be required for the positive selection of iNKT cells and Treg cells, as are other unique factors, such as the immunomodulatory receptor SLAM and its adaptor SAP for iNKT cells, and interleukin 2 (IL-2) for Treg cells. Together such observations suggest that costimulatory signals do not always promote death2. In a study published in this issue of Nature Immunology, Pobezinsky and colleagues revisit the role of CD28 in thymic selection through the analysis of mice genetically deficient in CD28 or deficient in both B7-1 and B7-2 (called B7-deficient mice here)3. They find that unlike iNKT cells or Treg cells, IELs expressing homodimers of CD8 (CD8+) do not require costimulation via CD28. Instead, they report that CD28- or B7-deficient mice have considerably more gut IELs and thymic IEL precursors, a process they call clonal diversion (Fig. 1b). By highlighting this alternative fate, these data help explain the controversy surrounding the role of costimulation via CD28 in clonal deletion.
The study by Pobezinsky and colleagues provides compelling evidence that autoreactive T cell clones that would have been deleted in normal mice can instead be diverted to a CD4CD8 double-negative (DN) T cell population in costimulation-deficient (CD28or B7-deficient) mice3. T cell clones expressing TCRs containing -chain variable region 5 (V5) or V11 are normally deleted from the conventional CD4+ or CD8+ T cell pool of BALB/c mice because of reactivity with endogenous superantigens Mtv-6, Mtv-8 and Mtv-9; as far as the thymus is concerned, these are seen as autoreactive cells. Notably, however, costimulation-deficient mice have a greater frequency of V5+ and V11+ T cell clones among TCR+ DN cells. Similarly, the thymus of mice overexpressing the prosurvival factors Bcl-2 or Mcl-1 shows enrichment for TCR+ DN cells expressing Mtv superantigen reactive V+ TCRs, which further supports the idea that such clones are normally deleted but are diverted into TCR+ DN cells when they cannot undergo apoptosis. To determine if such diverted cells are IEL precursors, the authors use adoptive-transfer studies. They find that TCR+ DN cells from
npg
528
nEWS AnD VIEWS a b

also confirmed here3. However, the frequency of such clones in the thymic IEL progenitor pool (TCR+ DN cells) is not as high as it is in the CD8+ IEL pool. This discrepancy may reflect the inadequacy of markers available at present in identifying true IEL precursors in the thymus. Indeed, the relationship between the TCR+ DN cells studied here and a previously identified thymic IEL precursor4 is unclear. Alternatively, perhaps not all CD8+ IELs arise from CD4+CD8+ thymocyte progenitors10. In the future it will be important to develop better means of identifying thymic IEL precursors, an effort that could be aided by the use of costimulation-deficient mice. It will also be important to further explore the expression of homing molecules, as the potential of thymic IEL precursors critically depends on their ability to traffic to the gut. Despite the accumulating wealth of knowledge about development and homeostasis of CD8+ IELs, it remains unclear what these autoreactive yet anergic cells do in the gut. Do they function to maintain intestinal health? Do they continue to recognize self antigen? Most CD8+ IELs express the activation marker CD69, but this is probably due to environmental factors, not antigen exposure11. CD8+ molecules interact with high affinity with thymus leukemia antigen expressed on many intestinal epithelial cells, but the precise function of this interaction remains poorly defined12. Although they are anergic by the criteria in the present study3 (after stimulation via the TCR, they fail to flux calcium, produce IL-2 or proliferate), IELs express granzyme B and are directly cytolytic ex vivo. They also have the potential to produce IL-10 and the growth factor KGF to diminish inflammation and promote epithelial integrity, respectively. Such characteristics of CD8+ IELs would suggest they may serve a regulatory role in the gut; however, further studies are needed to determine their precise function. Overall, these studies have provided compelling evidence that in the absence of costimulation via CD28, self-reactive T cells are not deleted but instead take an alternative route and become cells with the potential to develop into IELs. In the future it will be important to develop tools to help further the understanding of the development and function of CD8+ IELs. For example, it will be useful to study the IEL TCR repertoire to gain a better understanding of TCR specificity in the natural IEL population, and to develop TCR-transgenic models to characterize the development of CD8+ IELs. Other important objectives include identifying the key transcription factor(s) that maintain(s) the identity and survival of IELs. Finally, it will be important to develop methods
529
Costimulation-sufficient thymus
Costimulation-deficient thymus
i NKT CD28 SLAM & SAP DP Self-reactive CD28 IL-2 4SP int
Treg
i NKT
Treg
CD28?
DP Self-reactive
4SP int Clonal diversion
IELp (TCR+ DN)
IELp (TCR DN)

+
Gut IL-15
IELp CD8+ IEL CD8+ IEL
IL-15
IELp
Runx3
Figure 1 Outcomes of high-affinity TCR signaling in the thymus. (a) CD4+CD8+ double-positive (DP) thymocytes that express a self-reactive TCR have at least four potential fates. Cells with self lipid reactive V14-J18+ TCRs can develop into inKT cells through interactions that require signaling via SLAM and SAP, as well as CD28 (blue). T cells reactive to complexes of self peptide and MHC class II can develop into Foxp3+ Treg cells by a process that requires IL-2 and CD28 (green). However, the fate of most self-reactive T cells is clonal deletion or elimination through apoptosis (). IEL precursors (IELp) of the gut (pink) are thought to be strongly self-reactive, although the unique requirements for their development are not fully understood. 4SP int, CD4+ single-positive intermediate cell. (b) CD28 deficiency changes the fates noted above. In the absence of CD28-B7 interactions, fewer inKT and Treg cells develop in the thymus and there is impaired clonal deletion (dashed arrows). CD28 deficiency also results in the adoption of the IEL fate by more cells (thick arrow), a process called clonal diversion by Pobezinsky and colleagues3. Dn IEL precursors leave the thymus, home to the gut and differentiate into CD8-expressing IELs through the action of Runx3 and IL-15.
B7-deficient mice efficiently generate CD8+ IELs after adoptive transfer into hosts deficient in recombination-activating gene 2. Mice deficient in 2-microglobulin normally have few gut IELs; however, clones restricted to major histocompatibility complex class II (which can develop without 2-microglobulin) that are diverted to the TCR+ DN lineage in costimulation-deficient mice still accumulate in large numbers as gut IELs. Finally, the frequency of normally negatively selected, forbidden clones among B7-deficient TCR+ DN cells is identical to that of IELs, consistent with the idea that TCR+ DN cells are progenitors of CD8+ IELs4,5. Pobezinsky and colleagues further show that the CD8+ IELs generated by a developmental diversion pathway are functionally anergic, given their inability to mobilize calcium, produce IL-2 or proliferate after stimulation via the TCR3. These data suggest a potential mechanism for keeping self-reactive T cells of the gut in check. Once in the gut, DN IEL precursors are hypothesized to become CD8-expressing IELs through the inductive effect of IL-15. This was previously established through use of adoptive-transfer models in which DN
progenitors require IL-15 (ref. 6), CD122 (IL-2 receptor -chain)7 or the proto-oncoprotein c-Myc5 to develop into IELs. Here, Pobezinsky and colleagues suggest that this also requires the transcription factor Runx3 (ref. 3), a known regulator of CD8 expression8. Using in vitro culture, they show that stimulation via the TCR in the presence of IL-15 leads to the expression of Runx3 and CD8 but not of CD8. Furthermore, Runx3-deficient mice have fewer CD8+ IELs than wild-type mice have, even though they have similar numbers of TCR+ DN precursors. But what does the finding of clonal diversion in costimulation-deficient mice suggest for IEL development in normal mice? Does this pathway represent a developmental fate for clones that are strongly self-reactive but fail to encounter B7-expressing, antigen-presenting cells? Both thymic epithelial and dendritic cells express B7 ligands, although thymocytes themselves do not, and cortical epithelial cells have lower express of B7 ligands. Realistically, what are the chances that a T cell progenitor will recognize high-affinity self antigens but not B7? It has long been known that the CD8+ IEL pool contains forbidden clones9, a finding
npg
nEWS AnD VIEWS

to achieve selective depletion of CD8+ IELs to fully understand their function in the homeostasis of the immune system.
1. Walunas, T.L., Sperling, A.I., Khattri, R., Thompson, C.B. & Bluestone, J.A. J. Immunol. 156, 10061013 (1996). 2. Stritesky, G.L., Jameson, S.C. & Hogquist, K.A. Annu. Rev. Immunol. 30, 95114 (2012). 3. Pobezinsky, L.A. et al. Nat. Immunol. 13, 569578 (2012). 4. Gangadharan, D. et al. Immunity 25, 631641 (2006). 5. Jiang, W., Ferrero, I., Laurenti, E., Trumpp, A. & MacDonald, H.R. Blood 115, 44314438 (2010). 6. Kennedy, M.K. et al. J. Exp. Med. 191, 771780 (2000). 7. Suzuki, H., Duncan, G.S., Takimoto, H. & Mak, T.W. J. Exp. Med. 185, 499505 (1997). 8. Sato, T. et al. Immunity 22, 317328 (2005). 9. Rocha, B., Vassalli, P. & Guy-Grand, D. J. Exp. Med. 173, 483486 (1991). 10. Hayday, A. & Gibbons, D. Mucosal Immunol. 1, 172174 (2008). 11. Casey, K.A. et al. J. Immunol. doi:10.4049/jimmunol. 1200402 (13 April 2012). 12. Olivares-Villagmez, D. & Van Kaer, L. Immunol. Lett. 134, 16 (2010).
Tespa1: another gatekeeper for positive selection

Nicholas R J Gascoigne & Guo Fu
A previously unknown protein, Tespa1, that regulates the thymocyte positive-selection checkpoint has now been identified. The phenotype of Tespa1-deficient mice and the role of Tespa1 in thymocyte signaling are very similar to those of Themis-deficient mice and Themis itself, another recently described but unrelated protein.
developing thymocyte, once it has started to express its individuality via a correctly rearranged and expressed T cell antigen receptor (TCR), must make a decision based on how well that TCR interacts with its environment of major histocompatibility complex (MHC) molecules presenting self peptides. That decision will determine whether the individual cell will be positively selected to grow up to be a member of the naive T cell community, commuting to and from the suburbs of the secondary lymphoid organs, or if it will overdose on TCR stimulation and die young (negative selection). The TCR signal-transduction pathways that regulate this decision have been intensively studied for nearly 30 years, so it is rather surprising that important participants are still being discovered. In a paper published in this issue of Nature Immunology, Wang and co-workers describe just such a protein, Tespa1 (for thymocyte-expressed, positive selection associated 1)1. Tespa1 was found through the use of an online data-mining tool to identify expressed sequence tags with preferential expression in cells of the T lineage1. Tespa1 mRNA encodes a protein of ~52 kilodaltons with no known conserved domains. The amino-terminal ~240 amino acids are highly conserved among vertebrates, but the carboxy-terminal half of the protein (of placental mammals) is apparently not present in monotremes (platypus) or Xenopus1, although sequences from bony fish have an unrelated carboxy-terminal domain. There is a conserved potential Src homology 3 (SH3)
nicholas R.J. Gascoigne and Guo Fu are in the Department of Immunology and Microbial Science, The Scripps Research Institute, La Jolla, California, USA. e-mail: gascoigne@scripps.edu
npg
domainbinding site at Pro184 (position according to mouse sequence numbering) that is not shared with platypus or Xenopus and a proline-rich region with three potential SH3 domainbinding sites (Pro273, Pro277 and Pro280) that is also not present in the sequences from nonplacental mammals. The Tespa1 message has its highest expression in CD4+CD8+ double-positive (DP) thymocytes and lower expression in CD4CD8 double-negative thymocytes and CD4+ or CD8+ single-positive (SP) thymocytes and in mature T cells. Tespa1/ mice are deficient in the development of thymocytes from DP cells to SP cells. This description applies equally to another recently identified protein, Themis26, and in fact the phenotype of the Tespa1-deficient mice is very similar to that of Themis/ mice. The expression of Tespa1 differs from that of Themis in that Tespa1 is also present in cells of the B lineage and mast cells, although Tespa1-deficient mice do not have altered development of B cells or mast cells1. The effect of Tespa1 deletion on inhibiting the DP-to-SP transition has all of the hallmarks of a defect in positive selection. There are fewer than normal SP thymocytes, with the effect first noticeable at the stage at which expression of the activation marker CD69 and the TCR begins to increase in response to signaling through the TCR and downregulation of CD8 begins (CD4+CD8int). Expression of Tespa1 is about twice as high in CD4+ SP T cells and mature T cells as in their CD8+ equivalents, and the defect in the development of SP thymocytes seems to affect cells with transgenic expression of an MHC class IIrestricted TCR much more severely than it affects cells with transgenic expression of an MHC class Irestricted TCR1. Tespa1/ thymocytes do not seem to have a defect in negative selection.
The block in positive selection in Tespa1deficient mice is incomplete and results in fewer T cells in the periphery than that in wild-type mice. Those T cells undergo homeostatic proliferation in the periphery and thus have the phenotype of memory cells. They are less able than wild-type cells to respond to TCR stimulation, as measured by upregulation of CD69, proliferation and cytokine production. However, these studies do not address whether Tespa1 is involved in signaling in mature T cells, as the peripheral cells in Tespa1/ mice develop abnormally. Analysis of this point must await mice with conditional deletion of Tespa1 in mature T cells. Tespa1 is involved in the formation of the Lat signalosome as a result of TCR stimulation1, as is Themis7 (Fig. 1a). Both molecules interact with the TCR signaling components Grb2 and PLC-1 (refs. 15,7). The absence of Tespa1 results in less TCR stimulationinduced interaction of PLC-1 with Themis and Grb2, and of Lat with PLC-1, Themis and Grb2. Tespa1-deficient thymocytes seem to have less interaction between Themis and Grb2, a finding that is somewhat difficult to reconcile with the constitutive nature of this interaction. Tespa1 immunoprecipitates together with PLC-1 and Grb2 but not with Themis or Lat1. These data suggest that Tespa1 is involved at an early stage in the formation of the Lat signalosome and that its presence aids the recruitment of PLC-1 and Themis-Grb2, but the molecular mechanisms remain obscure. Mice with T lineagespecific deletion of Grb2 have a phenotype similar to that of mice deficient in Tespa1 or Themis8, which suggests that the functions of these molecules in thymocyte signaling are bound together. Tespa1-deficient DP thymocytes have considerable defects in activation of the
530
nEWS AnD VIEWS

to achieve selective depletion of CD8+ IELs to fully understand their function in the homeostasis of the immune system.
1. Walunas, T.L., Sperling, A.I., Khattri, R., Thompson, C.B. & Bluestone, J.A. J. Immunol. 156, 10061013 (1996). 2. Stritesky, G.L., Jameson, S.C. & Hogquist, K.A. Annu. Rev. Immunol. 30, 95114 (2012). 3. Pobezinsky, L.A. et al. Nat. Immunol. 13, 569578 (2012). 4. Gangadharan, D. et al. Immunity 25, 631641 (2006). 5. Jiang, W., Ferrero, I., Laurenti, E., Trumpp, A. & MacDonald, H.R. Blood 115, 44314438 (2010). 6. Kennedy, M.K. et al. J. Exp. Med. 191, 771780 (2000). 7. Suzuki, H., Duncan, G.S., Takimoto, H. & Mak, T.W. J. Exp. Med. 185, 499505 (1997). 8. Sato, T. et al. Immunity 22, 317328 (2005). 9. Rocha, B., Vassalli, P. & Guy-Grand, D. J. Exp. Med. 173, 483486 (1991). 10. Hayday, A. & Gibbons, D. Mucosal Immunol. 1, 172174 (2008). 11. Casey, K.A. et al. J. Immunol. doi:10.4049/jimmunol. 1200402 (13 April 2012). 12. Olivares-Villagmez, D. & Van Kaer, L. Immunol. Lett. 134, 16 (2010).
Tespa1: another gatekeeper for positive selection

Nicholas R J Gascoigne & Guo Fu
A previously unknown protein, Tespa1, that regulates the thymocyte positive-selection checkpoint has now been identified. The phenotype of Tespa1-deficient mice and the role of Tespa1 in thymocyte signaling are very similar to those of Themis-deficient mice and Themis itself, another recently described but unrelated protein.
developing thymocyte, once it has started to express its individuality via a correctly rearranged and expressed T cell antigen receptor (TCR), must make a decision based on how well that TCR interacts with its environment of major histocompatibility complex (MHC) molecules presenting self peptides. That decision will determine whether the individual cell will be positively selected to grow up to be a member of the naive T cell community, commuting to and from the suburbs of the secondary lymphoid organs, or if it will overdose on TCR stimulation and die young (negative selection). The TCR signal-transduction pathways that regulate this decision have been intensively studied for nearly 30 years, so it is rather surprising that important participants are still being discovered. In a paper published in this issue of Nature Immunology, Wang and co-workers describe just such a protein, Tespa1 (for thymocyte-expressed, positive selection associated 1)1. Tespa1 was found through the use of an online data-mining tool to identify expressed sequence tags with preferential expression in cells of the T lineage1. Tespa1 mRNA encodes a protein of ~52 kilodaltons with no known conserved domains. The amino-terminal ~240 amino acids are highly conserved among vertebrates, but the carboxy-terminal half of the protein (of placental mammals) is apparently not present in monotremes (platypus) or Xenopus1, although sequences from bony fish have an unrelated carboxy-terminal domain. There is a conserved potential Src homology 3 (SH3)
nicholas R.J. Gascoigne and Guo Fu are in the Department of Immunology and Microbial Science, The Scripps Research Institute, La Jolla, California, USA. e-mail: gascoigne@scripps.edu
npg
domainbinding site at Pro184 (position according to mouse sequence numbering) that is not shared with platypus or Xenopus and a proline-rich region with three potential SH3 domainbinding sites (Pro273, Pro277 and Pro280) that is also not present in the sequences from nonplacental mammals. The Tespa1 message has its highest expression in CD4+CD8+ double-positive (DP) thymocytes and lower expression in CD4CD8 double-negative thymocytes and CD4+ or CD8+ single-positive (SP) thymocytes and in mature T cells. Tespa1/ mice are deficient in the development of thymocytes from DP cells to SP cells. This description applies equally to another recently identified protein, Themis26, and in fact the phenotype of the Tespa1-deficient mice is very similar to that of Themis/ mice. The expression of Tespa1 differs from that of Themis in that Tespa1 is also present in cells of the B lineage and mast cells, although Tespa1-deficient mice do not have altered development of B cells or mast cells1. The effect of Tespa1 deletion on inhibiting the DP-to-SP transition has all of the hallmarks of a defect in positive selection. There are fewer than normal SP thymocytes, with the effect first noticeable at the stage at which expression of the activation marker CD69 and the TCR begins to increase in response to signaling through the TCR and downregulation of CD8 begins (CD4+CD8int). Expression of Tespa1 is about twice as high in CD4+ SP T cells and mature T cells as in their CD8+ equivalents, and the defect in the development of SP thymocytes seems to affect cells with transgenic expression of an MHC class IIrestricted TCR much more severely than it affects cells with transgenic expression of an MHC class Irestricted TCR1. Tespa1/ thymocytes do not seem to have a defect in negative selection.
The block in positive selection in Tespa1deficient mice is incomplete and results in fewer T cells in the periphery than that in wild-type mice. Those T cells undergo homeostatic proliferation in the periphery and thus have the phenotype of memory cells. They are less able than wild-type cells to respond to TCR stimulation, as measured by upregulation of CD69, proliferation and cytokine production. However, these studies do not address whether Tespa1 is involved in signaling in mature T cells, as the peripheral cells in Tespa1/ mice develop abnormally. Analysis of this point must await mice with conditional deletion of Tespa1 in mature T cells. Tespa1 is involved in the formation of the Lat signalosome as a result of TCR stimulation1, as is Themis7 (Fig. 1a). Both molecules interact with the TCR signaling components Grb2 and PLC-1 (refs. 15,7). The absence of Tespa1 results in less TCR stimulationinduced interaction of PLC-1 with Themis and Grb2, and of Lat with PLC-1, Themis and Grb2. Tespa1-deficient thymocytes seem to have less interaction between Themis and Grb2, a finding that is somewhat difficult to reconcile with the constitutive nature of this interaction. Tespa1 immunoprecipitates together with PLC-1 and Grb2 but not with Themis or Lat1. These data suggest that Tespa1 is involved at an early stage in the formation of the Lat signalosome and that its presence aids the recruitment of PLC-1 and Themis-Grb2, but the molecular mechanisms remain obscure. Mice with T lineagespecific deletion of Grb2 have a phenotype similar to that of mice deficient in Tespa1 or Themis8, which suggests that the functions of these molecules in thymocyte signaling are bound together. Tespa1-deficient DP thymocytes have considerable defects in activation of the
530
nEWS AnD VIEWS

a
PIP2 DAG InsP3 InsP3 PKC RasGRP InsP3 Tespa1 Lck Zap70 Lat Sos1 Themis PLC-1 Grb2
NF-B
InsP3 R
b
Position: Tespa1
P FF PPP
1252 0 100 200 300 400 500 521
Ssfa2
P FF
npg
Figure 1 Tespa1 interactions in thymocyte signaling. (a) Tespa1 seems to be involved in the initial formation of the Lat signalosome, as it aids the interactions among Lat, PLC-1, Themis and Grb2 (ref. 1). Loss of Tespa1 causes less signaling via calcium (Ca2+) and Erk. There is highly speculative potential involvement of Tespa1 (orange) in the interaction with the receptor for Ins(1,4,5)P 3 (InsP3R), given the sequence homology between Tespa1 and Ssfa2. DAG, diacylglycerol; PIP 2, phosphatidylinositol-4,5-bisphosphate; Greek letters (right), TCR chains; PKC, protein kinase C; RasGRP and Sos1, Ras guanine nucleotideexchange factor; Lck and Zap70, tyrosine kinase. (b) Region of homology (green box) in Tespa1 and Ssfa2, defined by alignment of Mus musculus Tespa1 and Ssfa2 sequences through the use of the position-specific iterative basic local alignment search tool PSI BLAST (from the national Center for Biotechnology Information); analysis statistics: P < 4 1036, 35% identities, 52% positives, 13% gaps. P, potential SH3 domainbinding sites; FF, di-phenylalanine sequence identified as being important for the interaction of Ssfa2 with the receptor for Ins(1,4,5)P3 (ref. 10).
Te sp a
Erk
Ca2+
mitogen-activated protein kinase Erk pathway and in the calcium-calcineurin pathway, which results in less activity of the transcription factors AP-1 and NFAT. The activity of the transcription factor NF-B, whose translocation to the nucleus requires activation of protein kinase C after stimulation via the TCR, is not affected. Thus, Tespa1 seems to act in the TCR signaling cascade in the formation of the Lat signalosome, and its absence results in deficient calcium and Erk signaling1 (Fig. 1a). Although Tespa1 is reported to have no paralogs1, we note that it has substantial homology to Ssfa2 (sperm-specific antigen 2; also known as Ki-Rasinduced actin-interacting protein (KRAP)). The region of homology extends over the amino-terminal ~300 amino acids of both proteins (Fig. 1b). This includes the region with the most homology in the vertebrate paralogs of Tespa1 and extends past the positions at which the platypus and Xenopus
Tespa1 sequences terminate (Fig. 1b). Ssfa2 is expressed mainly in the epidermis and pancreas, as well as in mast cells and macrophages. It is not found in the cells of the T lineage (according to the gene-annotation portal BioGPS). It is important in regulating the release of calcium from the endoplasmic reticulum, working through an interaction with the receptor for inositol-1,4,5-trisphosphate (Ins(1,4,5)P3)9. Interestingly, Ssfa2 and the receptor for Ins(1,4,5)P3 interact in particular regions of the endoplasmic reticulum that are near the plasma membrane in cell-cell contact regions in confluent epithelial cell cultures. The region of amino acids 200218 of Ssfa2 (in particular, Phe202 and Phe203) is important in the interaction with the Ins(1,4,5)P3 receptor10. This region is highly conserved in amino acids 185203 of Tespa1 and includes Phe187 and Phe188 of Tespa1. A potential SH3 domainbinding site is immediately upstream
of this region in both Tespa1 and Ssfa2. These similarities suggest the possibility that like Ssfa2, Tespa1 may have a role in the release of calcium from the endoplasmic reticulum involving an interaction with the receptor for Ins(1,4,5)P3. Thus, Tespa1 is important in the Lat signalosome, functioning alongside Themis and Grb2 in interacting with PLC-1 and in activating calcium and Erk signaling pathways. Although nothing is known of its specific mechanism, the relationship between the sequences of Tespa1 and Ssfa2 (Fig. 1b) suggests that the receptor for Ins(1,4,5)P3 and release of calcium from the endoplasmic reticulum may conceivably be involved in this (Fig. 1a). Signaling components that are expressed particularly in thymocytes at the positive-selection developmental checkpoint may be involved in the response to signals of relatively low affinity with carefully regulated responses to allow positive, agonist or negative selection11. Tespa1 and Themis may both fulfill this function, as suggested by their restricted organ-specific expression, by the dynamics of their expression at various stages of thymocyte development and by the developmental blockade in mice deficient in Tespa1 or Themis. However, an intriguing observation is that in both Tespa1/ thymocytes and Themis/ thymocytes, despite the clear involvement of Tespa1 and Themis in the Lat signalosome, the proximal TCR signaling cascade is essentially unaffected, followed by relatively mild1 or undetectable35 to at best very mild2 decreases in distal signaling defects (for example, calcium flux and phosphorylation of Erk). These signaling defects seem to be not matched by the strong positiveselection defects seen in both Tespa1/ mice and Themis/ mice. It is possible that these paradoxical results may be due to the use of antibody crosslinking as a surrogate reagent for stimulation of the thymocyte TCR. This delivers a far stronger signal than would a normal physiological positive-selecting ligand and indeed probably delivers a stronger signal than would the natural negative-selecting ligand. Given that thymocytes are more sensitive to stimulation via the TCR than are mature T cells and given their propensity to undergo apoptosis in response to strong stimulation, antibody crosslinking seems particularly inappropriate in this context. It is likely that the differences among Tespa1/, Themis/ and wild-type thymocytes in signal transduction that are relevant to positive selection may be masked by this overly strong stimulus. We propose that a more physiological stimulant, such as MHC class I or MHC class II tetramers on TCR-transgenic T cells, or very limited crosslinking with titrated antibody on
531
nEWS AnD VIEWS

polyclonal T cells12, should be used in future biochemical studies.
1. Wang, D. et al. Nat. Immunol. 13, 560568 (2012). 2. Fu, G. et al. Nat. Immunol. 10, 848856 (2009). 3. Johnson, A.L. et al. Nat. Immunol. 10, 831839 (2009). 4. Lesourne, R. et al. Nat. Immunol. 10, 840847 (2009). 5. Patrick, M.S. et al. Proc. Natl. Acad. Sci. USA 106, 1634516350 (2009). 6. Kakugawa, K. et al. Mol. Cell. Biol. 29, 51285135 (2009). 7. Brockmeyer, C. et al. J. Biol. Chem. 286, 75357547 (2011). 8. Jang, I.K. et al. Proc. Natl. Acad. Sci. USA 107, 1062010625 (2010). 9. Fujimoto, T. et al. Biochem. Biophys. Res. Commun. 408, 214217 (2011). 10. Fujimoto, T. et al. Biochem. Biophys. Res. Commun. 408, 282286 (2011). 11. Morris, G.P. & Allen, P.M. Nat. Immunol. 13, 121128 (2012). 12. Freedman, B.D., Liu, Q.H., Somersan, S., Kotlikoff, M.I. & Punt, J.A. J. Exp. Med. 190, 943952 (1999).
npg
532
reSearcH HiGHLiGHTS
Inflammasome jigsaw puzzles

Inflammasomes are multiprotein complexes that assemble in response to infection or tissue damage to activate caspase-1 and induce the secretion of interleukin 1 (IL-1) and IL-18. In Science, Shenoy et al. show that GBP5 is specifically required for assembly of the NLRP3 inflammasome in human and mouse macrophages in response to live bacteria and soluble agents but not in response to crystalline agents. GBP5 is a member of a family of guanylate-binding proteins with recently identified immunological functions, and its expression is optimally induced by stimulation with lipopolysaccharide plus interferon-. GBP5 directly interacts with NLRP3 and undergoes tetramerization in a GTPase-independent manner. The GBP5 tetramer promotes oligomerization of the adaptor ASC via its interaction with NLRP3, which suggests that this newly identified inflammasome component can coordinate the modular assembly requirements of the inflammasome. IV Science 27, 481485 (2012)
TSLP underlies the atopic march

Atopic dermatitis is a common allergic skin disorder that often precedes the development of allergy in distal tissues such as the lung. In Mucosal Immunology, Ziegler and colleagues describe how the epithelial cellassociated cytokine TSLP can drive this so-called atopic march from one tissue to another. Intradermal injection of mice with TSLP plus antigen results in asthma after subsequent lung challenge with that same antigen alone. TSLP expression in the lung is not required for the manifestation of asthma, but an antigen-specific memory CD4+ T cell response is essential for this. Similar results are obtained after inducible expression of TSLP by keratinocytes. IL-25 is another epithelial cellderived cytokine linked to allergy; however, it has no role in the dissemination of allergy from the skin to the lungs. Although mice can be made tolerant to asthma orally, this is possible only if done before skin sensitization, a finding with implications for the translation of such immunotherapy to patients. ZF Mucosal Immunol. 5, 342351 (2012)
TCR tuning Pathogen-tailored TH17 cells

Patients with genetic defects in cells of the TH17 subset of helper T cells suffer recurrent infections with Candida albicans and Staphylococcus aureus. In Nature, Zielinski et al. use whole-microbe pulsing of monocytes to stimulate naive and memory human CD4+ T cells to assess the polarizing conditions and effector characteristics induced by complex microbes. C. albicansspecific TH17 cells mainly produce both IL-17 and interferon-g (IFN-g), whereas S. aureusspecific TH17 cells mostly produce only IL-17 but can induce IL-10 after restimulation. IL-6, IL-23 and IL-1b contribute to the TH17 differentiation induced by each pathogen, but IL-1b is essential in C. albicansinduced TH17 differentiation. IL-1b is needed to counteract the inhibitory effect of IL-12, which is detected only in C. albicans cultures, to induce cells that produce both IL-17 and IFN-g and to inhibit IL-10 production. IL-1b has similar effects on memory T cells, which suggests that IL-1b primes cells not only in IV lymphoid organs but also in target tissues. Nature (1 April 2012) doi:10.1038/nature10957
The tuning of T cell antigen receptor (TCR) responsiveness allows immature thymocytes to be positively selected by complexes of self peptide and major histocompatibility complex molecules of lower affinity but prevents mature T cells from responding to similar complexes in the periphery. In the Proceedings of the National Academy of Science, Laufer and colleagues show the kinase Lck redistributes in tuned thymocytes to increase the threshold required for the activation of mature T cells. Tuning occurs in a major histocompatibility complexdependent manner after positive selection to downregulate TCR sensitivity. In untuned or preselection CD4+CD8+ double-positive thymocytes, more phosphorylated Lck is associated with the TCR CD3 chain. In contrast, more Lck associates with CD4 in tuned thymocytes, and these complexes localize in membrane-associated protein islands separate from CD3. This separation of protein islands in tuned cells increases the threshold for TCR activation and thus requires agonist peptides of higher affinity, which thereby lessens the potential for T cell autoreactivity. LAD Proc. Natl. Acad. Sci. USA (23 April 2012) doi:10.1073/ pnas.1119272109
npg
Building blocks of hematopoiesis

Specialized tissue microenvironments direct and maintain the development of hematopoietic stem cells, but the cellular and molecular interactions of this process are still being determined. In Cell, Calderon and Boehm use a modular approach to delineate the rules of hematopoiesis. Through reconstitution of the thymic epithelium of nude mice that lack hematopoietic function with various combinations of chemokines (CCL25 and CXCL12), the cytokine SCF and the Notch ligand DLL4, the thymic epithelium is made functional to support different progenitor populations. For example, CXCL12, DLL4 and SCF together support T cell progenitors, whereas CCL25, CXCL12 and SCF are biased toward B cells. In total, the authors generate 15 different combinations of factors, each with its own functionality. This study identifies the surprisingly minimalist rules that can govern hematopoiesis. ZF Cell 149, 159172 (2012)
Good microbes
Increasing evidence supports the proposal of a role for commensal microorganisms in shaping systemic immune responses. In Nature Medicine, Artis and colleagues show that commensal bacteria can suppress the switching of B cells to immunoglobulin E (IgE). Germ-free mice and mice treated with broad-spectrum antibiotics have higher concentrations of circulating IgE. In turn, higher IgE concentrations correlate with more basophils and T helper type 2 cells and skewing toward type 2 immune responses in these mice. IgE induces upregulation of the receptor for IL-3 (CD123) in bone marrow precursor cells, which results in more basophil progenitor cells. Inhibition of IgE production or blockade of the receptor FceRIa diminishes the basophilia of the antibiotic-treated mice. Intriguingly, B cellintrinsic expression of the adaptor MyD88 suppresses IgE production, basophilia and enhanced susceptibility to atopic disease. These findings may explain the greater incidence of atopic diseases as a result of more antibiotic use. LAD Nat. Med. 18, 538546 (2012)
533
Written by Laurie A. Dempsey, Zoltan Fehervari & Ioana Visan
Articles
Translational control of the activation of transcription factor NF-kB and production of type I interferon by phosphorylation of the translation factor eIF4E
Barbara Herdy1,11, Maritza Jaramillo1,11, Yuri V Svitkin1, Amy B Rosenfeld1, Mariko Kobayashi2, Derek Walsh3, Tommy Alain1, Polen Sean1, Nathaniel Robichaud1, Ivan Topisirovic1, Luc Furic4, Ryan J O Dowling1, Annie Sylvestre1, Liwei Rong5, Rodney Colina6, Mauro Costa-Mattioli7, Jrg H Fritz8, Martin Olivier9, Earl Brown10, Ian Mohr2 & Nahum Sonenberg1
Type I interferon is an integral component of the antiviral response, and its production is tightly controlled at the levels of transcription and translation. The eukaryotic translation-initiation factor eIF4E is a rate-limiting factor whose activity is regulated by phosphorylation of Ser209. Here we found that mice and fibroblasts in which eIF4E cannot be phosphorylated were less susceptible to virus infection. More production of type I interferon, resulting from less translation of Nfkbia mRNA (which encodes the inhibitor IkBa), largely explained this phenotype. The lower abundance of IkBa resulted in enhanced activity of the transcription factor NF-kB, which promoted the production of interferon-b (IFN-b). Thus, regulated phosphorylation of eIF4E has a key role in antiviral host defense by selectively controlling the translation of an mRNA that encodes a critical suppressor of the innate antiviral response. The host innate immune system is the first line of defense against invading pathogens, which encompasses viruses1. Type I interferon, which includes interferon- (IFN-) and IFN-, is a pivotal component of this system. Rapid synthesis and secretion of these cytokines is crucial for a potent antiviral and immunomodulatory response. The initial induction of type I interferon is dependent on the recognition of pathogens by pattern-recognition receptors, which survey the extracellular and intracellular milieu. DNA and RNA viruses are recognized by pattern-recognition receptors, including Toll-like receptors, present on the cell surface and in endosomes, as well as several cytoplasmic receptors2. The presence of a virus triggers a cascade of events that ultimately results in the activation of several transcription factors, including IRF3, IRF7, ATF2c-Jun and NF-B. Those factors, together with the transcription factor IRF1, the transcriptional coactivators CBP and p300 and the architectural protein HMGI(Y), form the IFN- enhanceosome, which activates transcription of the gene encoding IFN- (Ifnb1)3. The NF-B pathway is particularly crucial in the control of various arms of the innate immune response and is phylogenetically conserved from insects to vertebrates4. Under basal conditions, the short-lived inhibitor IB binds to NF-B in the cytoplasm, which prevents translocation of
1Department
NF-B to the nucleus and subsequent activation of transcription5. After infection with a virus or stimulation with tumor necrosis factor, type I interferon or the synthetic RNA duplex poly(I:C), IB is phosphorylated by IB kinases, which results in ubiquitination of IB and its degradation by the proteasome6. NF-B then translocates to the nucleus and facilitates transcription of NF-B-regulated genes, including Ifnb1 (ref. 7). The synthesis of most components of the type I interferon pathway, including regulators and interferon itself, requires stringent control, which is accomplished at transcriptional and translational levels8,9. Translational control enables the cell to instantly adjust to its environment by regulating the translation rate of select mRNAs. It is thus ideally suited for the rapid responses required for host defense against viruses, which must utilize the cellular translation machinery to produce viral proteins. Under most circumstances, translational control is exerted at the initiation step, at which the ribosome is recruited to the 5 end of an mRNA bearing the cap structure m7GpppN (where m7 indicates 7-methylguanosine, ppp indicates triphosphate and N is any nucleotide). The interaction between the ribosome and the mRNA is facilitated by the heterotrimeric eIF4F complex that consists of eIF4E, which directly binds the mRNA 5-cap structure;
npg
of Biochemistry and Goodman Cancer Research Centre, McGill University, Montreal, Quebec, Canada. 2Department of Microbiology and NYU Cancer Institute, New York University School of Medicine, New York, New York, USA. 3National Institute for Cellular Biotechnology, Dublin City University, Dublin, Ireland. 4Department of Anatomy and Developmental Biology, Monash University, Clayton, Australia. 5McGill AIDS Center, Lady Davis Institute of the Jewish General Hospital, Montreal, Canada. 6Laboratorio de Virologia Molecular, Regional Norte-Salto, Universidad de la Repblica, Salto, Uruguay. 7Department of Neuroscience, Baylor College of Medicine, Houston, Texas, USA. 8Complex Traits Group and Department of Microbiology and Immunology, McGill University, Montreal, Canada. 9The Research Institute of the McGill University Health Centre, Centre for the Study of Host Resistance, Department of Medicine and Department of Microbiology and Immunology, McGill University, Montreal, Canada. 10Department of Biochemistry, Microbiology and Immunology, and Emerging Pathogens Research Centre, University of Ottawa, Ottawa, Canada. 11These authors contributed equally to this work. Correspondence should be addressed to N.S. (nahum.sonenberg@mcgill.ca). Received 24 January; accepted 19 March; published online 29 April 2012; doi:10.1038/ni.2291
nature immunology VOLUME 13
NUMBER 6
JUNE 2012
543
Articles a
WT eIF4E(S209A)
b
4 6 8 10 0 2 4 6 8 10
WT
eIF4E(S209A)
c
WT eIF4E (S209A) 108 107 106 5 10 104 103 102 10 1
d
MOI G
eIF4E(S209A) MEFs EV 0 0.01 WT eIF4E 0 0.01
Time (h) 0 2 G N,P -actin eIF4E
eIF4E(S209A) MEFs + EV eIF4E(S209A) MEFs + WT eIF4E

108 107 106 105 4 10 3 10 2 10 10 1
VSV titer (PFU/ml)
-actin eIF4E
Figure 1 Replacement of the serine at the eIF4E-phosphorylation site at position p-eIF4E 209 with alanine impairs VSV replication. (a) Immunoblot analysis of VSV glycoprotein (G), nucleocapsid protein (N) and phosphoprotein (P), as well as -actin (loading control throughout) and eIF4E, in lysates of wild-type and eIF4E(S209A) MEFs mock infected (0) or infected for 210 h (above lanes) with VSV at a multiplicity of infection of 1 plaque-forming unit (PFU) per cell. (b) Cytopathic effect of VSV in wild-type and eIF4E(S209A) MEFs at 10 h after infection as in a. Original magnification, 10. (c) Plaque assay of VSV titers 10 h after infection as in a. (d) Immunoblot analysis of VSV proteins (as in a), -actin, total eIF4E and phosphorylated (p-) eIF4E in lysates of eIF4E(S209A) MEFs transfected with empty vector (EV) or vector encoding wild-type eIF4E, mock infected (0) or infected for 20 h with VSV at a multiplicity of infection (MOI) of 0.01 PFU/cell. (e) Plaque assay of VSV titers in eIF4E(S209A) MEFs transfected and infected as in d. Data are representative of three independent experiments (error bars, s.e.m.).
eIF4G, a scaffolding protein; and eIF4A, a DEAD-box RNA helicase10. The subunit eIF4G interacts with eIF3, which is bound to the small ribosomal subunit, thereby establishing the critical link between the mRNA and the ribosome. Among translation-initiation factors, eIF4E is the least abundant, and it is thought to be limiting for translation11. Thus, regulating eIF4E activity is critical for cellular function. The mitogen-activated protein kinaseinteracting kinases Mnk1 and Mnk2 phosphorylate Ser209 of eIF4E12. Although the function of eIF4E phosphorylation in various biological contexts remains unclear, it has been shown to control the translation of certain mRNAs that encode proteins associated with inflammation and cancer13. Mnk1 and Mnk2 are the sole kinases known to phosphorylate eIF4E in mice 14. Although Mnk2 is constitutively active, Mnk1 is regulated by signaling cascades of the mitogen-activated protein kinases p38 and Erk in response to mitogens, growth factors and hormones15,16. Phosphorylation of eIF4E is altered during viral infection. Dephosphorylation of eIF4E occurs during infection with influenza virus, adenovirus, encephalomyocarditis virus (EMCV), poliovirus or vesicular stomatitis virus (VSV)1720. In contrast, infection with herpesvirus or poxvirus stimulates Mnk1-dependent phosphorylation of eIF4E2124. Although inhibition of Mnk1 suppresses the replication of herpesvirus and poxvirus2124, direct involvement of eIF4E phosphorylation in infection by DNA viruses has not been established. Furthermore, it is unclear how dephosphorylation of eIF4E affects the replication of RNA viruses. To address those issues, we studied mouse embryonic fibroblasts (MEFs) derived from mice in which the serine at position 209 of eIF4E was replaced with alanine (eIF4E(S209A) mice), which prevented phosphorylation of eIF4E at this critical regulatory site. We found that loss of eIF4E phosphorylation in eIF4E(S209A) mice and cells resulted in an enhanced type I interferon immune response that protected against viral infection. We also found that the phosphorylation status of eIF4E controlled IFN- production by regulating the translation of Nfkbia mRNA, which encodes IB, the repressor of NF-B. Impaired infection of eIF4E(S209A) cells with select RNA or DNA viruses demonstrated direct involvement of eIF4E phosphorylation in viral pathogenesis and host defense. RESULTS Impaired VSV replication in cells lacking eIF4E phosphorylation To understand how the phosphorylation of eIF4E affects the replication of VSV, a negative-stranded RNA virus, we used spontaneously
544
immortalized MEFs derived from wild-type or eIF4E(S209A) mice13. VSV-encoded mRNAs are translated via a cap-dependent mechanism25. By immunoblot analysis, we detected strong signals for most viral proteins in wild-type MEFs at 10 h after infection but detected only small amounts of VSV-specific glycoprotein, nucleocapsid protein and phosphoprotein in eIF4E(S209A) MEFs (Fig. 1a). To exclude the possibility that the lower abundance of viral proteins in infected eIF4E(S209A) MEFs was a consequence of impaired global protein synthesis, we determined that new protein synthesis in uninfected eIF4E(S209A) MEFs was similar to that in wild-type MEFs (Supplementary Fig. 1a). In contrast to infected wild-type MEFs, infected eIF4E(S209A) cells did not show a detectable cytopathic effect (Fig. 1b). Plaque assays showed a ~1,000-fold higher viral yield in wild-type MEFs than in eIF4E(S209A) MEFs (Fig. 1c). The inhibition of VSV replication in eIF4E(S209A) MEFs was independent of the genetic background of the host, as we observed less abundance of viral proteins in eIF4E(S209A) MEFs from either C57BL/6 mice (Supplementary Fig. 1b) or BALB/c mice (Supplementary Fig. 1c). To confirm that the lower susceptibility of eIF4E(S209A) MEFs to VSV infection was a consequence of the S209A substitution rather than an unintended consequence of the gene-targeting manipulation, we reintroduced a gene encoding wild-type eIF4E into eIF4E(S209A) MEFs. Expression of wild-type eIF4E in eIF4E(S209A) MEFs enhanced VSV replication, as shown by more accumulation of viral protein (Fig. 1d) and more production of infectious virus (Fig. 1e). These data demonstrated that the impaired replication of VSV in eIF4E(S209A) MEFs was directly associated with the mutated Eif4e allele encoding the S209A substitution that precludes eIF4E phosphorylation. Substitution of the serine at position 209 with phosphomimetic glutamic or aspartic acid residues does not recapitulate the activity of wild-type eIF4E. Phosphomimetic mutants of eIF4E have less capbinding ability26, less ability to transport mRNA and substantially attenuated ability to transform cells27,28. Thus, these mutants do not mimic phosphorylated serine and therefore could not be used to show that the restricted replication of VSV in eIF4E(S209A) MEFs resulted from the inability of eIF4E to be phosphorylated rather than from nonspecific effects of the S209A substitution. Thus, we used MEFs deficient in the eIF4E kinases Mnk1 and Mnk2 to study this. Phosphorylation of eIF4E is not observed in cells that lack both kinases14. We obtained spontaneously immortalized MEFs derived from wild-type mice and mice doubly deficient in both Mnk1 and Mnk2 (Mnk1-Mnk2-deficient) and infected these MEFs with VSV. As noted for eIF4E(S209A) cells, accumulation of VSV protein was
VOLUME 13 NUMBER 6 JUNE 2012 nature immunology
npg
VSV titer (PFU/ml)
N,P
Articles
a
SV (MOI) 0 pE2 E1 C
WT
eIF4E(S209A)
1 1 01 00 0. 0. 0.
b
WT eIF4E(S209A) Time (h) 0 4 5 6 0 4 5 6
1 1 01 00 0 0. 0. 0.
of translational changes in host mRNA(s) whose protein product(s) is (are) essential for the regulation of virus infection. Production of more type I interferon in eIF4E(S209A) cells To determine if the diminished virus replication in eIF4E(S209A) cells could have been due to enhanced production of type I interferon, we treated wild-type and eIF4E(S209A) MEFs with poly(I:C), a potent inducer of type I interferon (Supplementary Fig. 3a). We grew naive wild-type MEFs in control medium or in the presence of conditioned medium from poly(I:C)-treated wild-type or eIF4E(S209A) cells and then infected the MEFs with VSV. Control medium from mock-treated eIF4E(S209A) cells resulted in a VSV yield ~1,300-fold higher than that resulting from conditioned medium from poly(I:C)-treated eIF4E(S209A) cells (Fig. 3a), whereas control medium from wild-type cells resulted in a VSV yield only twofold higher than that resulting from conditioned medium from poly(I:C)treated wild-type cells (Fig. 3a). Thus, conditioned medium from poly(I:C)-treated eIF4E(S209A) cells had potent antiviral activity. To investigate whether IFN- production was responsible for the antiviral activity detected, we used semiquantitative RT-PCR to measure Ifnb1 mRNA. Stimulation of eIF4E(S209A) MEFs for 6 h with poly(I:C) produced approximately 12-fold more Ifnb1 mRNA than did stimulation of wild-type cells (Fig. 3b). We also detected much more Ifih1 mRNA (which encodes the RNA helicase Mda5) and Irf7 mRNA (which encodes the transcription factor IRF7), whose transcription is induced by type I interferon and whose protein products are required for poly(I:C)-dependent induction of IFN-31,32, after treatment of eIF4E(S209A) MEFs with poly(I:C) (~16-fold for Irf7 mRNA and ~4-fold for Ifih1 mRNA; Fig. 3b), whereas secretion of IFN- was sixfold greater from eIF4E(S209A) MEFs than from wildtype cells (Fig. 3c). These data demonstrated that eIF4E(S209A) MEFs produced more IFN-. To show that the greater abundance of IFN- in the medium of infected eIF4E(S209A) MEFs antagonized viral replication, we cultured cells in the presence of neutralizing antibodies to type I interferon before and during VSV infection. The addition of antibody to IFN- (anti-IFN-) alone or in combination with anti-IFN- enhanced the accumulation of VSV protein (Fig. 4a) and viral titers (Fig. 4b) in eIF4E(S209A) MEFs, albeit not to the same degree observed in wild-type MEFs (21% of wild type). The amount of anti-IFN- used (25 g/ml) was sufficient to neutralize the secreted IFN-, as we found that a concentration of this antibody as low as 0.25 g/ml was saturating for the enhancement of VSV replication (Supplementary Fig. 3b). In addition, in EMCV-infected wild-type and eIF4E(S209A) MEFs, control rat immunoglobulin G or isotypematched monoclonal antibody to IFN- did not cause more replication of EMCV, whereas the addition of anti-IFN- enhanced EMCV viral
P1 3CD 3D 1AB, 2C 1D
-actin IB eIF4E p-eIF4E
1C, 3C
Figure 2 Restricted viral protein synthesis in eIF4E(S209A) MEFs infected with Sindbis virus or EMCV. (a) Autoradiogram (top) of new protein synthesis in wild-type and eIF4E(S209A) MEFs pulse-labeled with [35S]methionine and mock infected (0) or infected for 17 h with Sindbis virus (SV; MOI above lanes); left margin, viral proteins. Below, immunoblot analysis (IB) of total and phosphorylated eIF4E and -actin. (b) Autoradiogram of wild-type and eIF4E(S209A) MEFs pulse-labeled with [35S]methionine, deprived of serum for 72 h and mock-infected or infected for 4, 5 or 6 h (above lanes) with EMCV (MOI, 10 PFU/cell); left margin, viral proteins. Data are representative of three independent experiments.
considerably impaired in infected Mnk1-Mnk2-deficient MEFs (Supplementary Fig. 2a). Although we detected a pronounced cytopathic effect in infected wild-type cells, Mnk1-Mnk2-deficient MEFs remained unaffected (Supplementary Fig. 2b). In sum, these results suggested that precluding phosphorylation of eIF4E via the S209A substitution or by deletion of eIF4E kinases impeded the replication of VSV. Impaired replication of RNA viruses in eIF4E(S209A) cells To investigate whether the replication of other RNA viruses was similarly impaired in eIF4E(S209A) MEFs or if the replication defect was specific to VSV, we infected wild-type and eIF4E(S209A) MEFs with the alphavirus Sindbis virus or with the picornavirus EMCV. Sindbis virusencoded mRNA has a 5-cap structure and is translated in an eIF4E-dependent manner29, similar to that of VSV25. EMCV mRNA lacks a 5-cap structure and is translated in an eIF4Eindependent manner via an internal ribosome entry site30. Infection of eIF4E(S209A) MEFs with Sindbis virus or EMCV resulted in less production of viral protein than did infection of wild-type MEFs, as determined by metabolic labeling with [35S]methionine (Fig. 2a,b). Because translation of EMCV proteins is cap independent, these results suggested that the diminished replication of EMCV in eIF4E(S209A) MEFs was probably a consequence
npg
Figure 3 More production of type I interferon in WT WT eIF4E(S209A) MEFs after poly(I:C) stimulation. eIF4E(S209A) WT eIF4E(S209A) eIF4E(S209A) 8 40 (a) VSV titers in naive wild-type MEFs grown Poly(I:C) 10 7 35 (g/ml) 10 overnight in the presence of conditioned 6 30 10 medium from wild-type or eIF4E(S209A) Ifnb1 5 25 10 4 cells left untreated () or treated for 6 h (+) 20 10 Irf7 3 15 10 with poly(I:C) (0.5 g/ml); after that overnight 2 10 10 incubation, MEFs were infected for 12 h with Ifih1 10 5 VSV (MOI, 0.1 PFU/cell). (b) Semiquantitative 0 1 Actb 0 0.25 0.5 1 Poly(I:C) + + RT-PCR of Ifnb1, Ifih1 and Irf7 mRNA Poly(I:C) (g/ml) among total RNA isolated from wild-type and eIF4E(S209A) MEFs treated for 6 h with various concentrations (above lanes) of poly(I:C). Actb encodes -actin (loading control). (c) Enzyme-linked immunosorbent assay of IFN- in conditioned medium of wild-type and eIF4E(S209A) MEFs treated for 6 h with various concentrations (horizontal axis) of poly(I:C). Data are representative of two independent experiments (error bars, s.e.m.).
0 0. 25 0. 5 1
VSV titer (PFU/ml)
0 0. 25 0. 5 1
IFN- (pg/ml)
NUMBER 6
JUNE 2012
545
Articles
Figure 4 The addition of neutralizing 9 10 antibody to IFN- enhances viral replication 8 10 7 in eIF4E(S209A) MEFs. (a) Immunoblot 10 70 6 analysis of VSV proteins in lysates of wild-type eIF4E(S209A) WT 10 60 5 10 and eIF4E(S209A) MEFs mock treated or Anti-IFN- + + 50 4 10 Anti-IFN- + + + + treated with anti-IFN- (25 g/ml) alone or in 40 3 VSV (MOI 0.1) + + + + + 10 30 combination anti-IFN- (20 g/ml), followed by 2 10 20 G infection for 12 h with VSV (MOI, 0.1 PFU/cell). 10 10 (b) VSV titers in wild-type and eIF4E(S209A) 0 1 N,P Anti-IFN- IgG + + + MEFs infected for 12 h with VSV (MOI, 0.1 PFU/ Anti-IFN- Anti-IFN- + + + + cell) in the presence or absence of neutralizing Anti-IFN- + + anti-IFN- and anti-IFN-. (c) EMCV titers in WT eIF4E(S209A) WT eIF4E(S209A) wild-type and eIF4E(S209A) MEFs (2 x 105) deprived of serum (0.5% FBS in DMEM) for 72 h, then treated for 24 h with control antibody (rat immunoglobulin G (IgG; 2.5 g /ml)) or neutralizing anti-IFN- or anti-IFN- and then infected for 48 h with EMCV (100 PFU). Data are representative of two (a) or three (b,c) independent experiments (error bars, s.e.m.).
VSV titer 6 (10 PFU/ml) EMCV titer (PFU/ml)
titers by ~100-fold (Fig. 4c). Thus, eIF4E(S209A) MEFs produced and secreted more IFN- than did wild-type MEFs and this greater production of type I interferon in eIF4E(S209A) cells antagonized the replication of two different RNA viruses.
Lower susceptibility of eIF4E(S209A) mice to VSV infection To assess the susceptibility of eIF4E(S209A) mice to infection with VSV, we infected female wild-type and eIF4E(S209A) mice intranasally with VSV and monitored the mice for 12 d. In this model, VSV is neurovirulent and results in fatal encephalitis, which is preceded by hindlimb paralysis. Although only ~50% of the wild-type mice survived this infection, 87% of the eIF4E(S209A) mice survived this infection (P = 0.0243; Fig. 5a). Wild-type mice secreted less type I interferon into the serum than did eIF4E(S209A) mice at 48 h after infection (P = 0.0507; Fig. 5b). These data established that eIF4E(S209A) mice produced more type I interferon in response to viral infection than did wild-type mice and were more resistant to the neurovirulent effects of infection with VSV. Less translation of Nfkbia mRNA in eIF4E(S209A) cells To investigate the molecular mechanism by which the production of type I interferon was enhanced in eIF4E(S209A) MEFs, we used genome-wide polysome analysis to compare wild-type and
a
100 90 80 70 60 50 40 30 20 10 0
Survival (%)
b 250
200
IFN- (pg/ml)
150 100 50 0 WT eIF4E(S209A)
eIF4E(S209A) MEFs13. Notably, Nfkbia mRNA was one of the mRNAs whose translation was lower in eIF4E(S209A) MEFs than in wildtype cells and was the only mRNA that encodes a protein conspicuously involved in signaling pathways that regulate the expression of type I interferon. IB inhibits NF-B, which is an essential nuclear factor in the activation of transcription3. A lower abundance of IB in eIF4E(S209S) cells would thus be expected to result in more translocation of NF-B to the nucleus. The amount of IB in eIF4E(S209A) MEFs was approximately 50% of that in wild-type cells, as determined by immunoblot analysis (Fig. 6a). Ribosome loading of Nfkbia mRNA was much lower in eIF4E(S209A) MEFs, as indicated by the prominent shift of Nfkbia mRNA toward lighter polysomal fractions (fractions 35 in eIF4E(S209A) MEFs, compared with fractions 69 in wildtype MEFs; Fig. 6b). This indicated that initiation of the translation of Nfkbia mRNA was lower in eIF4E(S209A) MEFs than in wild-type MEFs. In contrast, we detected no apparent change in the polysome distribution of eIF4E-insensitive Actb mRNA (which encodes -actin; Fig. 6b). Cap-dependent translation of VSV mRNA that encodes the virus matrix protein and translation of luciferase reporter mRNA driven by an EMCV internal ribosome entry site were not diminished by the eIF4E(S209A) substitution, as determined by polysome profiling (Supplementary Fig. 4ac) and luciferase assay (Supplementary Fig. 4d). In addition, depletion of IB through the use of short hairpin RNA (shRNA) in wild-type MEFs diminished VSV replication (Fig. 6c,d), which provided evidence that lowering the amount of IB was sufficient to diminish the production of VSV protein and VSV replication. Together these data demonstrated that translation of Nfkbia mRNA was selectively impaired as a consequence of dephosphorylation of eIF4E, which resulted in less IB protein. More NF-kB activity in eIF4E(S209A) cells Less IB in eIF4E(S209A) MEFs would be expected to result in enhanced NF-B activity6,7. We used electrophoretic mobility-shift assay (EMSA) to monitor site-specific DNA-binding activity of NF-B in nuclear extracts of wild-type and eIF4E(S209A) MEFs and detected more NF-B activity in nuclear extracts of eIF4E(S209A) MEFs than in those of wild-type MEFs under basal conditions (Fig. 6e). We observed an increase in the binding of NF-B to DNA in nuclear protein extracts of wild-type and eIF4E(S209A) cells at 15 and 30 min after incubation with tumor necrosis factor, a potent inducer of NF-B. However, the increase was greater in extracts of eIF4E(S209A) MEFs than in those of wild-type MEFs. As a control, we determined that binding of the transcription factors SP1 and AP-1 to DNA was similar in extracts of wild-type and eIF4E(S209A) cells (Fig. 6e). The difference between wild-type and eIF4E(S209A)
npg
WT eIF4E(S209A)
Figure 5 Enhanced survival of eIF4E(S209A) mice after intranasal infection with VSV. (a) Survival of female wild-type mice (n = 14) and eIF4E(S209A) mice (n = 17) infected intranasally with VSV (1 104 PFU/mouse) and monitored for 12 d (mice were killed when they developed hindlimp paralysis); results are presented as a Kaplan-Meier plot. P = 0.0243 (Mantel-Cox test). (b) Concentration of IFN- in blood from wild-type mice (n = 8) and eIF4E(S209A) mice (n = 7) infected for 48 h with VSV (1 104 PFU/mouse). Each symbol represents an individual mouse; small horizontal lines indicate the mean. *P = 0.0507 (t-test). Data are representative of two independent experiments.
546
0 1 2 3 4 5 6 7 8 9 10 11 12
Time after infection (d)
Articles
Figure 6 Less translation of Nfkbia mRNA in eIF4E(S209A) MEFs leads to more activation of NF-B. (a) Immunoblot analysis of IB, 01 01 00 00 0. 0 0. VSV (MOI) 0 total and phosphorylated eIF4E, and -actin IB G shRNA: in wild-type and eIF4E(S209A) MEFs deprived p-eIF4E IB of serum overnight and then stimulated with N,P eIF4E serum as described13. (b) Polysome profiles p-eIF4E -actin -actin of wild-type and eIF4E(S209A) MEFs eIF4E -actin (top left); semiquantitative RT-PCR analysis of WT cytoplasmic Actb and Nfkbia mRNA among total 80S eIF4E(S209A) IB Ctrl shRNA: RNA isolated from wild-type and eIF4E(S209A) 60S cells (top right); and distribution of ribosomal Polysomes 40S RNA in individual fractions collected after Actb sucrose-gradient (1050%) centrifugation eIF4E of cytoplasmic extracts of wild-type and (S209A) WT 2 3 4 5 6 7 8 9 10 Nfkbia eIF4E(S209A) MEFs, with semiquantitative TNF (min) 0 15 30 0 15 30 RT-PCR analysis of the distribution of Nfkbia Sedimentation RelA and Actb mRNA (bottom). Results for polysome eIF4E(S209A) WT NF-B p50 profiles are presented as absorbance at 254 nm Fraction: 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 (A254); 40S, 60S and 80S indicate ribosomal RNA SP1 subunits. (c) Immunoblot analysis of IB and Nfkbia -actin in extracts of wild-type MEFs transfected AP-1 Actb with vector for control (Ctrl) shRNA or shRNA targeting IB. (d) Immunoblot analysis of VSV proteins, total and phosphorylated eIF4E, and -actin in lysates of wild-type MEFs transfected with vector for control or IB-specific shRNA and mock infected (0) or infected for 17 h with VSV (MOI, above lanes). (e) EMSA of NF-B subunits RelA and p50, as well as of SP1 and AP-1 (controls), in nuclear extracts of wild-type and eIF4E(S209A) MEFs treated for 0, 15 or 30 min (above lane) with tumor necrosis factor (TNF; 1 ng/ml). Data are representative of two independent experiments.
4E (S 20 9A )
eI F
0. 01 0. 00 1
A254
eI
F4
E( S2
09
IB
Ctrl
A)
40 60S S 80 S
npg
MEFs in the formation of the NF-BDNA complex was due to lack of eIF4E phosphorylation in eIF4E(S209A) cells, as the expression of wild-type eIF4E in these cells resulted in 44% less DNA-binding activity of NF-B (Supplementary Fig. 5a). In addition, through the use of an NF-B-specific reporter plasmid, we confirmed that eIF4E(S209A) MEFs had more NF-B transcriptional activity under basal conditions (Supplementary Fig. 5b). Consistent with those results, expression of the NF-B target genes Ccl11 (ref. 33) and Il1b34 was enhanced 4-fold and 139-fold, respectively, in untreated eIF4E(S209A) MEFs relative to their expression in wild-type cells (Supplementary Fig. 5c). However, the greater abundance of IL-1 did not contribute to the antiviral effect detected in eIF4E(S209A) MEFs (Supplementary Fig. 6a,b). Thus, eIF4E(S209A) MEFs had
greater NF-B activity than did wild-type MEFs, which resulted in enhanced induction of NF-B target genes. To show that the greater abundance of IFN- was due to enhanced transcription via NF-B, we transfected luciferase reporter plasmids containing the complete mouse Ifnb1 promoter or the positive regulatory domain II (PRDII), which is a distinct part of the Ifnb1 promoter that shares 80% nucleotide homology with the binding site for NF-B, into eIF4E(S209A) and wild-type MEFs. We observed more luciferase activity for each promoter in eIF4E(S209A) MEFs than in wild-type MEFs (Supplementary Fig. 7a,b), which suggested that the greater production of IFN- in eIF4E(S209A) MEFs was caused by transcriptional activation of Ifnb1. We detected no difference between wild-type and eIF4E(S209A) MEFs in the translation
EMCV WT Nfkbia WT + EMCV Actb
40 60S S 80 S
a
A254
60S 80S 40S
WT WT + VSV
VSV + Nfkbia Actb WT Fraction: 1 2 3 4 5 6 7 8 9 10 11 12 Nfkbia Actb
b
60S 80S
WT +
Time (h) 0 4 6 8 10
A254
Polysomes
40S
Polysomes
WT Fraction: 1 2 3 4 5 6 7 8 9 10 Nfkbia Actb
RelA NF-B p50 G N,P -actin eIF4E p-eIF4E
IB
40 60S S 80 S
Sedimentation
WT + VSV Fraction: 1 2 3 4 5 6 7 8 9 10 11 12 Nfkbia Actb
Sedimentation
WT + EMCV Fraction: 1 2 3 4 5 6 7 8 9 10 Nfkbia Actb
40 S 60 S 80 S
1 2 3 4 5 6 7 8 9 10 11 12
1 2 3 4 5 6 7 8 9 10 11 12
Time (h) 0 RelA p50
NF-B
40 60S S 80 S
Figure 7 Less translation of Nfkbia transcripts in wild-type MEFs infected with VSV or EMCV. (a) Polysome profiles (as in Fig. 6b) of wild-type MEFs mock infected or infected for 8 h with VSV (MOI, 1 PFU/cell; left), and semiquantitative RT-PCR analysis of cytoplasmic Actb and Nfkbia mRNA among total RNA isolated from mock-infected () or VSV-infected (+) wild-type MEFs (right, top) and of the distribution of Nfkbia and Actb mRNA in fractions (collected as in Fig. 6b) of cytoplasmic extracts of mock-infected (WT) or VSV-infected (WT + VSV) wild-type MEFs (right, below). (b) Polysome profiles (as in Fig. 6b) of wild-type MEFs mock infected or infected for 6 h with EMCV (MOI, 0.1 PFU/cell) in medium supplemented with 2% serum (left), and semiquantitative RT-PCR analysis of cytoplasmic Actb and Nfkbia mRNA among total RNA from mock- or EMCV-infected wild-type MEFs (right, top) and of the distribution of Nfkbia and Actb mRNA in fractions (collected as in Fig. 6b) of cytoplasmic extracts of mock- or EMCV-infected wild-type MEFs (right, below). (c) EMSA of NF-B in nuclear extracts of wild-type MEFs infected for 010 h (above lanes) with VSV (MOI, 1 PFU/cell; top), and immunoblot analysis of VSV proteins, -actin and total and phosphorylated eIF4E in cytoplasmic extracts of those cells (below). (d) EMSA of NF-B in nuclear extracts of lungs from wild-type mice mock infected (0) or infected for 16 h with VSV (2 10 6 PFU/mouse); each lane is extracts from the lung of one mouse. Data are representative of two independent experiments.
40 S 60 S 80 S
0. 01 0. 00 1
16 16
NUMBER 6
JUNE 2012
547
Articles a
VV (PFU) 300 3,000
c
WT eIF4E(S209A) 108 107 106 105 104 3 10 102 10 1
VV protein
VV titer (PFU/ml)
impaired in eIF4E(S209A) MEFs relative to their replication in wildtype cells. Hence, absence of the eIF4E-phosphorylation site Ser209 resulted in less replication of DNA and RNA viruses in MEFs, which demonstrated that this post-translational modification of eIF4E acted as a determinant of host susceptibility to a broad spectrum of viruses. DISCUSSION The production of type I interferon is critical in the innate host response to limit viral replication. A key cellular component that regulates the production of type I interferon is the transcriptional activator NF-B3,35. Here we have established that activation of NF-B was regulated by a mechanism that involved phosphorylation of the cap-binding protein eIF4E. Translation of the mRNA encoding IB, the short-lived inhibitor of NF-B, was selectively lower in MEFs from mice with the S209A substitution of eIF4E that prevents eIF4E phosphorylation. Less IB resulted in more activated NF-B in eIF4E(S209A) cells, which in turn enhanced the production of type I interferon. Not only was the replication of three different RNA viruses impaired in eIF4E(S209A) MEFs but also eIF4E(S209A) mice were more resistant to the neurological damage caused by VSV infection. Together our results have demonstrated a key role for the phosphorylation of eIF4E in viral pathogenesis and host defense against VSV through its effect on the translation of mRNA that encodes a critical suppressor of the innate antiviral response. The production of IFN- is dependent on several transcription factors, including NF-B, that assemble on the DNA at a region immediately upstream of the Ifnb1 promoter to form the IFN- enhanceosome3,7,36. As the enhanceosome components exist in suboptimal concentrations in a cell, many cells do not produce IFN-. Only 20% of an infected cell population reportedly expresses Ifnb1 mRNA37,38. Diminishing the abundance of phosphorylated eIF4E provides a means for rapidly increasing NF-B activity and thereby promoting assembly of the IFN- enhanceosome, especially in cells that would otherwise fail to activate Ifnb1 transcription. The existence of virus-encoded proteins, including the poliovirus protein 3C, which cleaves the RelA subunit of NF-B39 and the influenza virus protein NS1, which prevents the activation of NF-B mediated by double-stranded RNA40, further illustrates the importance of neutralizing NF-B-mediated production of type I interferon for efficient propagation of virus. Diminishing the abundance of phosphorylated eIF4E promoted the activation of NF-B and production of type I interferon by inhibiting the translation of Nfkbia mRNA. Notably, dephosphorylation of eIF4E occurs late in infection with several viruses19,20. Our data have suggested that dephosphorylation of eIF4E enabled cells to rapidly produce type I interferon via a translational-control mechanism that selectively diminished IB and stimulated activation of NF-B. Although this would be of limited benefit to the cell during the first cycle of viral replication, it could effectively establish an antiviral state in the surrounding cells. Thus, diminishing the phosphorylation of eIF4E might be a cellular response aimed at limiting the spread of viral infection and protecting the host. Although enhanced production of type I interferon was the main mechanism that rendered eIF4E(S209A) MEFs more resistant to infection with RNA viruses, the production and secretion of other antiviral cytokines might also be involved, as the addition of neutralizing antibody to IFN- to infected eIF4E(S209A) MEFs failed to fully restore VSV replication. Even unstimulated eIF4E(S209A) MEFs secreted more of the NF-B-dependent cytokines CCL11 (eotaxin) and IL-1 than did wild-type MEFs. CCL11 (eotaxin) is a chemokine
b d
eIF4E
Phase Hoechst GFP
eIF4E(S209A)
WT
Figure 8 Less replication of DNA viruses in eIF4E(S209A) MEFs. (a) Immunoblot analysis of proteins from vaccinia virus (VV) in lysates of wild-type and eIF4E(S209A) MEFs (4 105) infected for 64 h with vaccinia virus (300 or 3,000 PFU). (b) Isoelectric focusing of total proteins from a and immunoblot analysis of eIF4E; arrow indicates phosphorylated eIF4E. (c) Vaccinia virus titers in wild-type and eIF4E(S209A) MEFs (4 105) infected for 64 h with vaccinia virus (300 PFU). (d) Phase-contrast images (left) and fluorescence images of Hoechst staining (middle) and green fluorescent protein (GFP; right) of wild-type and eIF4E(S209A) MEFs infected with an HSV-1 strain expressing enhanced green fluorescent protein fused to the late viral protein Us11. Original magnification, 20. Data are representative of three independent experiments (error bars, s.e.m.).
of Ifnb1 mRNA (Supplementary Fig. 4c). Together these data indicated that phosphorylation of eIF4E modulated the type I interferon response by controlling the translation of Nfkbia mRNA and, subsequently, the activation of NF-B. Virus-induced eIF4E dephosphorylation and NF-kB activation To demonstrate the physiological importance of the dephosphorylation of eIF4E, we assessed the translation of Nfkbia mRNA in wild-type MEFs infected with viruses (VSV or EMCV) known to dephosphorylate eIF4E19,20. Ribosomal loading of Nfkbia mRNA was much lower in VSV-infected wild-type MEFs than in uninfected wild-type MEFs, as indicated by the shift of Nfkbia mRNA toward lighter polysomal fractions (Fig. 7a, fractions 1112). We obtained similar results with wild-type cells infected with EMCV (Fig. 7b, fractions 810). In addition, VSV infection resulted in the activation of NF-B in wild-type cells (Fig. 7c) and wild-type mice (Fig. 7d). These results suggested that the cellular antiviral response to RNA viruses that induce dephosphorylation of eIF4E involved less translation of Nfkbia mRNA and activation of NF-B. We next sought to determine whether infection with DNA viruses would result in a similar phenotype. We infected wild-type and eIF4E(S209A) MEFs with the following two DNA viruses known to stimulate phosphorylation of eIF4E: vaccinia virus (a poxvirus) and herpes simplex virus 1 (HSV-1; a member of the herpesviridae)21,22. The replication of vaccinia virus (Fig. 8ac) and HSV-1 (Fig. 8d) was
548
npg
eI F4 E W (S 20 T 9A eI ) F4 E( W S T 20 9A )
Articles
known to attract eosinophils, and IL-1 is a proinflammatory cytokine that mediates a wide range of immune and inflammatory responses. Despite the enhanced secretion of IL-1 from eIF4E(S209A) cells, neutralization of IL-1 in combination with neutralization of IFN- did not render eIF4E(S209A) MEFs more susceptible to VSV infection. That observation further demonstrated that IFN- was the main cytokine that limited viral replication in eIF4E(S209A) MEFs. Systemically, however, secretion of these cytokines could further augment the antiviral phenotype observed in eIF4E(S209A) mice, as recruitment of several cells of the immune response to the site of infection could clear the virus more effectively. Viruses use various strategies to attenuate the type I interferon response. Indeed, both herpesvirus and poxvirus produce proteins that stimulate the assembly of eIF4F and promote the phosphorylation of eIF4E, which successfully maintains a high concentration of phosphorylated eIF4E2123. Accordingly, eIF4E(S209A) MEFs were more refractory to infection with vaccinia virus or HSV-1 than were wildtype MEFs. These viruses also use many strategies to prevent the phosphorylation of the eukaryotic translation-initiation factor eIF2 and are more resistant to type I interferon than are RNA viruses41,42. Thus, phosphorylation of eIF4E represents another example of the strategies viruses use to control the type I interferon response by post-translational modification of critical host translation-initiation factors. Our data have suggested that dephosphorylation of eIF4E acts as an arm of the host response to VSV and EMCV infection. In nature, the interaction between host defense mechanisms and viral countermeasures is balanced and results in limited acute viral replication and persistence in immunocompetent hosts. The S209A substitution of eIF4E distorted that balance by diminishing the synthesis of IB, which resulted in more production of type I interferon. One possibility is that an antiviral mechanism does exist that diminishes the activity of eIF4E. Notably, natural mutations in the gene encoding eIF4E in some plants confer resistance to viral infection 43,44. Together our findings establish a previously unknown translation-control mechanism involved in the production of type I interferon, a fundamental aspect of innate host defense. Furthermore, they raise the possibility that interfering with the phosphorylation of eIF4E in virus-infected cells could be an additional means of augmenting the cellular innate immune response to viral infection. METHODS Methods and any associated references are available in the online version of the paper.
Note: Supplementary information is available in the online version of the paper. ACKNOWLEDgMENTS We thank A. Frey (NYU School of Medicine) for rat antibody immunoglobulin G; M. Olivier (McGill University) for tumor necrosis factor; J. Bell (University of Ottawa) for antibody to VSV proteins and VSV-GFP; J. Pelletier (McGill University) for the retroviral vectors MSCV-GFP and MSCV-eIF4E; M. Gale (University of Washington) for the PRDII luciferase reporter Ifnb1PRDII-Luc; D.A. Muruve (University of Calgary) for the luciferase reporters Ifnb1-pGL3 and Nfkb1; R. Fukunaga (Kyoto University) for wild-type and Mnk1-Mnk2-deficient mice; J. Berlanga (Universidad Autnoma de Madrid) for Sindbis virus; and C. Lister for assistance. Supported by the Canadian Institutes of Health Research (MOP-7214 to N.S.), the US National Institutes of Health (AI 073898 and GM056927 to I.M., and T32 AI007647 to M.K.) and the Irma Hirschl Trust (I.M.). AUTHOR CONTRIBUTIONS B.H., M.J. and D.W. designed, did and analyzed experiments and wrote the manuscript; Y.V.S. and A.B.R. provided intellectual input and wrote the manuscript; M.K., T.A., N.R., I.T. and L.F. designed and did experiments; P.S., R.J.O.D., A.S., L.R. and J.H.F. contributed to experiments; R.C. critically reviewed the manuscript; M.C.-M. designed experiments and critically reviewed the manuscript; M.O. and E.B. provided reagents; and I.M. and N.S. designed experiments and wrote the manuscript. COMPETINg FINANCIAL INTERESTS The authors declare no competing financial interests.
Published online at http://www.nature.com/doifinder/10.1038/ni.2291. reprints and permissions information is available online at http://www.nature.com/ reprints/index.html.
1. Katze, M.G., He, Y. & Gale, M. Jr. Viruses and interferon: a fight for supremacy. Nat. Rev. Immunol. 2, 675687 (2002). 2. Kawai, T. & Akira, S. Antiviral signaling through pattern recognition receptors. J. Biochem. 141, 137145 (2007). 3. Maniatis, T. et al. Structure and function of the interferon- enhanceosome. Cold Spring Harb. Symp. Quant. Biol. 63, 609620 (1998). 4. Silverman, N. & Maniatis, T. NF-B signaling pathways in mammalian and insect innate immunity. Genes Dev. 15, 23212342 (2001). 5. Hoffmann, A., Levchenko, A., Scott, M.L. & Baltimore, D. The IB-NF-B signaling module: temporal control and selective gene activation. Science 298, 12411245 (2002). 6. Chu, W.M. et al. JNK2 and IKK are required for activating the innate response to viral infection. Immunity 11, 721731 (1999). 7. Apostolou, E. & Thanos, D. Virus infection induces NF-B-dependent interchromosomal associations mediating monoallelic IFN- gene expression. Cell 134, 8596 (2008). 8. Honda, K., Takaoka, A. & Taniguchi, T. Type I interferon gene induction by the interferon regulatory factor family of transcription factors. Immunity 25, 349360 (2006). 9. Colina, R. et al. Translational control of the innate immune response through IRF-7. Nature 452, 323328 (2008). 10. Gingras, A.C., Raught, B. & Sonenberg, N. eIF4 initiation factors: effectors of mRNA recruitment to ribosomes and regulators of translation. Annu. Rev. Biochem. 68, 913963 (1999). 11. Duncan, R., Milburn, S.C. & Hershey, J.W. Heat shock effects on eIF-4F. Regulated phosphorylation and low abundance of HeLa cell initiation factor eIF-4F suggest a role in translational control. J. Biol. Chem. 262, 380388 (1987). 12. Joshi, B. et al. Phosphorylation of eukaryotic protein synthesis initiation factor 4E at Ser-209. J. Biol. Chem. 270, 1459714603 (1995). 13. Furic, L. et al. eIF4E phosphorylation promotes tumorigenesis and is associated with prostate cancer progression. Proc. Natl. Acad. Sci. USA 10, 1413414139 (2010). 14. Ueda, T., Watanabe-Fukunaga, R., Fukuyama, H., Nagata, S. & Fukunaga, R. Mnk2 and Mnk1 are essential for constitutive and inducible phosphorylation of eukaryotic initiation factor 4E but not for cell growth or development. Mol. Cell. Biol. 24, 65396549 (2004). 15. Morley, S.J. Intracellular signalling pathways regulating initiation factor eIF4E phosphorylation during the activation of cell growth. Biochem. Soc. Trans. 25, 503509 (1997). 16. Waskiewicz, A.J., Flynn, A., Proud, C.G. & Cooper, J.A. Mitogen-activated protein kinases activate the serine/threonine kinases Mnk1 and Mnk2. EMBO J. 16, 19091920 (1997). 17. Huang, J.T. & Schneider, R.J. Adenovirus inhibition of cellular protein synthesis involves inactivation of cap-binding protein. Cell 65, 271280 (1991). 18. Burgui, I., Yanguez, E., Sonenberg, N. & Nieto, A. Influenza virus mRNA translation revisited: is the eIF4E cap-binding factor required for viral mRNA translation? J. Virol. 81, 1242712438 (2007). 19. Kleijn, M., Vrins, C.L., Voorma, H.O. & Thomas, A.A. Phosphorylation state of the cap-binding protein eIF4E during viral infection. Virology 217, 486494 (1996). 20. Connor, J.H. & Lyles, D.S. Vesicular stomatitis virus infection alters the eIF4F translation initiation complex and causes dephosphorylation of the eIF4E binding protein 4EBP1. J. Virol. 76, 1017710187 (2002). 21. Walsh, D. & Mohr, I. Phosphorylation of eIF4E by Mnk-1 enhances HSV-1 translation and replication in quiescent cells. Genes Dev. 18, 660672 (2004). 22. Walsh, D. et al. Eukaryotic translation initiation factor 4F architectural alterations accompany translation initiation factor redistribution in poxvirus-infected cells. Mol. Cell. Biol. 28, 26482658 (2008). 23. Walsh, D., Perez, C., Notary, J. & Mohr, I. Regulation of the translation initiation factor eIF4F by multiple mechanisms in human cytomegalovirus-infected cells. J. Virol. 79, 80578064 (2005). 24. Arias, C., Walsh, D., Harbell, J., Wilson, A.C. & Mohr, I. Activation of host translational control pathways by a viral developmental switch. PLoS Pathog. 5, e1000334 (2009). 25. Both, G.W., Banerjee, A.K. & Shatkin, A.J. Methylation-dependent translation of viral messenger RNAs in vitro. Proc. Natl. Acad. Sci. USA 72, 11891193 (1975). 26. Scheper, G.C. et al. Phosphorylation of eukaryotic initiation factor 4E markedly reduces its affinity for capped mRNA. J. Biol. Chem. 277, 33033309 (2002). 27. Wendel, H.G. et al. Dissecting eIF4E action in tumorigenesis. Genes Dev. 21, 32323237 (2007).
npg
NUMBER 6
JUNE 2012
549
Articles
28. Topisirovic, I., Ruiz-Gutierrez, M. & Borden, K.L. Phosphorylation of the eukaryotic translation initiation factor eIF4E contributes to its transformation and mRNA transport activities. Cancer Res. 64, 86398642 (2004). 29. Banerjee, A.K. 5-terminal cap structure in eucaryotic messenger ribonucleic acids. Microbiol. Rev. 44, 175205 (1980). 30. Jang, S.K. et al. A segment of the 5 nontranslated region of encephalomyocarditis virus RNA directs internal entry of ribosomes during in vitro translation. J. Virol. 62, 26362643 (1988). 31. Gitlin, L. et al. Essential role of mda-5 in type I IFN responses to polyriboinosinic: polyribocytidylic acid and encephalomyocarditis picornavirus. Proc. Natl. Acad. Sci. USA 103, 84598464 (2006). 32. Honda, K. et al. IRF-7 is the master regulator of type-I interferon-dependent immune responses. Nature 434, 772777 (2005). 33. Huber, M.A. et al. The IKK-2/IB/NF-B pathway plays a key role in the regulation of CCR3 and eotaxin-1 in fibroblasts. A critical link to dermatitis in IB-deficient mice. J. Biol. Chem. 277, 12681275 (2002). 34. Warton, K., Foster, N.C., Gold, W.A. & Stanley, K.K. A novel gene family induced by acute inflammation in endothelial cells. Gene 342, 8595 (2004). 35. Visvanathan, K.V. & Goodbourn, S. Double-stranded RNA activates binding of NF-B to an inducible element in the human -interferon promoter. EMBO J. 8, 11291138 (1989). 36. Lenardo, M.J., Fan, C.M., Maniatis, T. & Baltimore, D. The involvement of NF-B in -interferon gene regulation reveals its role as widely inducible mediator of signal transduction. Cell 57, 287294 (1989). 37. Zawatzky, R. & De Maeyer, E. De Maeyer-Guignard, J. Identification of individual interferon-producing cells by in situ hybridization. Proc. Natl. Acad. Sci. USA 82, 11361140 (1985). 38. Hu, J. et al. Chromosome-specific and noisy IFNB1 transcription in individual virus-infected human primary dendritic cells. Nucleic Acids Res. 35, 52325241 (2007). 39. Neznanov, N. et al. Proteolytic cleavage of the p65-RelA subunit of NF-B during poliovirus infection. J. Biol. Chem. 280, 2415324158 (2005). 40. Wang, X. et al. Influenza A virus NS1 protein prevents activation of NF-B and induction of / interferon. J. Virol. 74, 1156611573 (2000). 41. Cheng, G., Feng, Z. & He, B. Herpes simplex virus 1 infection activates the endoplasmic reticulum resident kinase PERK and mediates eIF-2 dephosphorylation by the gamma(1)34.5 protein. J. Virol. 79, 13791388 (2005). 42. He, B., Gross, M. & Roizman, B. The 134.5 protein of herpes simplex virus 1 complexes with protein phosphatase 1 to dephosphorylate the subunit of the eukaryotic translation initiation factor 2 and preclude the shutoff of protein synthesis by double-stranded RNA-activated protein kinase. Proc. Natl. Acad. Sci. USA 94, 843848 (1997). 43. Piron, F. et al. An induced mutation in tomato eIF4E leads to immunity to two potyviruses. PLoS ONE 5, e11313 (2010). 44. German-Retana, S. et al. Mutational analysis of plant cap-binding protein eIF4E reveals key amino acids involved in biochemical functions and potyvirus infection. J. Virol. 82, 76017612 (2008).
npg
550
VOLUME 13
NUMBER 6
JUNE 2012
nature immunology
Chemicals and antibodies. Chemicals, poly(I:C) and poly(dI:dC) were from Sigma. FuGENE 6 and RNaseIN were from Roche. Tumor necrosis factor was a gift from M. Olivier (McGill University). [35S]methionine-[35S]cysteine and [-32P]dATP were from PerkinElmer. Antibodies were as follows: anti-eIF4E (610270; 87; BD Biosciences); anti--actin (A5316; AC-74; Sigma); anti-IB (4812; Cell Signaling); antibody to phosphorylated eIF4E (NB 100-2234; Novus Biologicals); control rat antibody immunoglobulin G (a gift from A. Frey); neutralizing antibody to mouse IFN- (22100-1; RMMA-1), IFN- (22400-1; RMMB-1) or IFN- (22500-1; RMMG-1; all from PBL Biomedical Laboratories); goat immunoglobulin G antibody to mouse (AF784) and neutralizing anti-IL-1 (AB-401-NA; both from R&D Systems); polyclonal antibody to vaccinia virus (8101; Virostat); and antibody to VSV proteins (a gift from J. Bell). Superscript III reverse transcriptase was from Invitrogen. The retroviral vectors MSCV-GFP (empty vector control) and MSCV-IF4E27 were provided by J. Pelletier. The shRNA lentiviral plasmids were from Sigma (pLKO.1-puro control shRNA and IB shRNA clone 939 (5-CCGGGAGTCAGAATTCACAGAGGATCTCGAGATCCTCTGTGAATT CTGACTCTTTTTG-3). The PRDII luciferase reporter Ifnb1PRDII-Luc was provided by M. Gale. The Ifnb1-pGL3 and Nfkb1-luciferase reporters were provided by D.A. Muruve. 2012 Nature America, Inc. All rights reserved. Cells, mice and viruses. Mice were kept in pathogen-free housing. Research involving animals was done according to the regulations of the Canadian Council of Animal Care and with the approval of the McGill University Animal Care Committee. Wild-type and eIF4E(S209A) mice and MEFs have been described13. Mice were backcrossed seven times onto a pure BALB/c background. Mice were born at normal Mendalian ratios and developed normally. Wild-type and Mnk1-Mnk2-deficient mice were from R. Fukunaga14. BHK21, Vero cells and BSC40 cells were from the American Type Culture Collection. The Indiana serotypes of VSV and EMCV have been described45,46. Sindbis virus was a gift from J. Berlanga. VSV-GFP was from J. Bell. Vaccinia virus (Western Reserve strain)22 and HSV-1 Us11-GFP have been described47. Viral infection and metabolic radiolabeling. Viral infection and metabolic radiolabeling of cells have been described46. For in vivo experiments, female BALB/c mice were anesthetized with 2 l rodent cocktail (5 ml ketamine, 2.5 ml xylazine, 1 ml acepromazine and 1.5 ml sterile saline) per 1 g body weight and were infected intranasally with VSV at a dose of 1 104 PFU per mouse. Mice were monitored for 12 d after infection and were killed when hindlimb paralysis developed. Blood was collected 48 h after infection by cardiac puncture and the serum concentration of IFN- was determined by B16-Blue assay (InvivoGen). Rescue experiments. The retroviral vector MSCV-IRES-GFP (empty vector control) or MSCV-eIF4E-IRES-GFP27 was transfected into phoenix-293-T packaging cells. Virus-containing medium was used for infection of eIF4E(S209A) MEFs. Cells were sorted by flow cytometry for expression of green fluorescent protein after 48 h. Isoelectric focusing. Vertical-slab isoelectric focusing was done as described21. Plaque assays. Plaque assays were done as described with confluent monolayers of BHK-21 or Vero cells46. Titers of vaccinia virus were determined as described22. Poly(I:C) treatment, virus-infection-protection assay and enzyme-linked immunosorbent assay. Wild-type and eIF4E(S209A) MEFs were seeded at a density of 0.7 105 cells per well. Cells were treated for 6 h with the appropriate concentration of poly(I:C) through the use of the FuGENE 6 transfection reagent according to the manufacturers protocol (Roche). Cells were lysed and total RNA was isolated, followed by reverse transcription and semiquantitative PCR. The intensity of bands was quantified with ImageJ software. Conditioned supernatants were collected and the production of mouse IFN- was detected by enzyme-linked immunosorbent assay according to the manufacturers protocol (PBL Biomedical Laboratories). For measurement of antiviral activity,
ONLINE METHODS
wild-type cells were overlaid for 24 h with conditioned cell medium and were infected for 1724 h with VSV at an MOI of 0.1 PFU/cell. Polysome profiles. Cells from five plates 15 cm in diameter at ~80% confluence were used for the preparation of polysomal fractions as described9. RNA was isolated from half of each fraction with TRIzol reagent (Invitrogen). RNA extraction and semiquantitative RT-PCR. RNA was extracted from 0.7 105 cells with an RNeasy Mini Kit according to the manufacturers protocol (Qiagen). Superscript III reverse transcriptase, oligo(dT) and 1 g of total RNA or RNA of polysomal fractions was used according to the manufacturers protocol (Invitrogen) for reverse transcription. Each cDNA template (1 l) was analyzed by semiquantitative PCR with specific primers and Taq DNA polymerase according to the suppliers protocol (Fermentas). The reaction products were in quantifiable range, as determined by serial dilution of mRNA isolated from the cytoplasmic fraction of wild-type MEFs. Primers for Ifnb, Irf7 or Actb have been described9. Primers for Ifih1, Nfkbia and VSV-specific M protein were as follows: Ifih1 sense (TGTCTGTTCTGCAGAGGACAGCTT) and antisense (ACAGAAAGAT GAGGTGGTCCAGCA); Nfkbia sense (GACGCAGACCTGCACACCCC) and antisense (TGGAGGGCTGTCCGGCCATT); and M protein sense (GTGGCAGCCGCTGTATCCCATT) and anti-sense (CGCAGTGAGCGT GATACTCGGG). Preparation of nuclear extracts and EMSA. Lungs of mice infected with VSV were isolated and nuclear proteins were extracted as described48. Cells grown to a confluence of ~70% in a dish 10 cm in diameter were treated for the appropriate time with tumor necrosis factor (1 ng/ml). Nuclear extracts were prepared and EMSA was done as described49. Nuclear extracts were incubated with [-32P]dATP-radiolabeled oligonucleotides containing a consensus binding site for NF-Bc-Rel homodimeric and heterodimeric complexes (5-AGTTGAGGGGACTTTCCCAGGC-3; Santa Cruz Biotechnology) or containing a consensus binding site for AP-1c-Jun homodimer and Jun-Fos heterodimeric complexes (5-CGCTTGATGACTCAGCCGG AA-3; Santa Cruz Biotechnology) or a consensus binding site for SP1 (5-ATT CGATCGGGGCGGGGCGAGC-3). DNA-protein complexes were resolved from free-labeled DNA by electrophoresis through native 4% (wt/vol) polyacrylamide gels and the dried gels were analyzed by autoradiography. Cytokine assay. Supernatants of wild-type or eIF4E(S209A) MEFs (1 105 cells) were analyzed by the MILLIPLEX MAP Mouse Cytokine/Chemokine according to the manufacturers protocol (Millipore) 8 h after being seeded. Luciferase reporter assays. Plasmid DNA (1 g) and the renilla luciferase plasmid pRL-TK (0.01 g) were transfected into 3 105 cells in a six-well plate through the use of Lipofectamine 2000 (Invitrogen). In vitrotranscribed EMCVIRESfirefly luciferase46 mRNA (0.5 g) was transfected together with capped-renilla luciferase mRNA (0.05 g) into 3 105 cells. Luciferase activity was measured 48 h after transfection of plasmids or 3 h after transfection of mRNA as described9. Metabolic labeling and TCA precipitation. Wild-type and eIF4E(S209A) MEFs were grown for 2 h in methionine- and cysteine-free medium supplemented with 10% dialyzed serum. The medium was replaced with methionine-free DMEM containing 35S protein-labeling mix (10 Ci/ml) and 1% FBS. After 30 min, cell monolayers were lysed and the radioactivity incorporated into material insoluble in 5% trichloroacetic acid was measured. Total protein content was determined with a Bio-Rad Protein Assay. Incorporation was normalized to protein content and is presented as counts per minute. Nitric oxide production. Macrophages isolated from the bone marrow of B10A.Bcgr (Bacillus-Calmette-Gurinresistant) mice (B10R cell line) were stimulated for 24 h with the appropriate treatment. The production of nitric oxide was assessed by measurement of the accumulation of nitrites in the cell culture medium with the colorimetric Griess Reagent System according to the manufacturers protocol (Promega).
npg
doi:10.1038/ni.2291
nature immunology
Statistical analysis. GraphPad Prism statistical software, version 4.0, was used for statistical analysis. P values of 0.01 to 0.05 were considered significant.
45. Stojdl, D.F. et al. The murine double-stranded RNA-dependent protein kinase PKR is required for resistance to vesicular stomatitis virus. J. Virol. 74, 95809585 (2000). 46. Svitkin, Y.V. et al. Eukaryotic translation initiation factor 4E availability controls the switch between cap-dependent and internal ribosomal entry site-mediated translation. Mol. Cell. Biol. 25, 1055610565 (2005).
47. Benboudjema, L., Mulvey, M., Gao, Y., Pimplikar, S.W. & Mohr, I. Association of the herpes simplex virus type 1 Us11 gene product with the cellular kinesin lightchain-related protein PAT1 results in the redistribution of both polypeptides. J. Virol. 77, 91929203 (2003). 48. Liu, S.F., Ye, X. & Malik, A.B. In vivo inhibition of nuclear factor-B activation prevents inducible nitric oxide synthase expression and systemic hypotension in a rat model of septic shock. J. Immunol. 159, 39763983 (1997). 49. Jaramillo, M., Naccache, P.H. & Olivier, M. Monosodium urate crystals synergize with IFN- to generate macrophage nitric oxide: involvement of extracellular signalregulated kinase 1/2 and NF-B. J. Immunol. 172, 57345742 (2004).
npg
nature immunology
doi:10.1038/ni.2291
Articles
Nonclassical MHC class Ibrestricted cytotoxic T cells monitor antigen processing in the endoplasmic reticulum
Niranjana A Nagarajan, Federico Gonzalez & Nilabh Shastri
The aminopeptidase ERAAP is essential for trimming peptides presented by major histocompatibility complex (MHC) class I molecules. Inhibition of ERAAP by cytomegalovirus results in evasion of the immune response by this virus, and polymorphisms in ERAAP are associated with autoimmune disorders. How normal ERAAP function is monitored is unknown. We found that inhibition of ERAAP rapidly induced presentation of the peptide FYAEATPML (FL9) by the MHC class Ib molecule Qa-1 b. Antigen-experienced T cells specific for the Qa-1bFL9 complex were frequent in naive mice. Wild-type mice immunized with ERAAP-deficient cells mounted a potent CD8+ T cell response specific for Qa-1bFL9. MHC class Ibrestricted cytolytic effector cells specifically eliminated ERAAP-deficient cells in vitro and in vivo. Thus, nonclassical Qa-1bpeptide complexes direct cytotoxic T cells to targets with defective antigen processing in the endoplasmic reticulum. Cells generate complexes of peptide and major histocompatibility complex (MHC) class I (pMHCI) to display their intracellular protein milieu for immunosurveillance1. Normal cells present complexes of self peptide and MHC class I, which are not recognized by selftolerant cytotoxic CD8+ T lymphocytes (CTLs), whereas infected or transformed cells present complexes of foreign peptide and MHC class I that activate the appropriate CTLs2. The nature of the peptide repertoire presented by MHC class I molecules and the nature of the rapidly responding antigen-specific CTLs are key determinants of immunosurveillance. The pMHCI complexes are generated by the antigen-processing pathway in a series of concerted steps3. Generally, antigen processing begins with the cleavage of protein precursors by the proteasome, followed by transport of peptides into the endoplasmic reticulum by the transporter associated with antigen processing (TAP). After transport, the aminopeptidase ERAAP (endoplasmic reticulum aminopeptidase associated with antigen processing) trims peptides with amino-terminal extensions to generate the final peptides presented by MHC class I4. The peptides generated by this pathway can be presented by classical MHC class Ia molecules as well as nonclassical MHC class Ib molecules5, although the role of ERAAP in generating peptides for presentation by MHC class Ib molecules is unclear. MHC class Ib molecules are closely related to the ubiquitous MHC class Ia molecules H-2K, H-2D and H-2L in the mouse (HLA-A, HLA-B and HLA-C in humans)5. An important difference is that MHC class Ib molecules are not polymorphic; for example, there are only four known alleles that encode the mouse MHC class Ib molecule Qa-1 (ref. 6) and there are only two alleles that encode its human analog HLA-E6. Both MHC class Ia and MHC class Ib molecules loaded with appropriate peptides in the endoplasmic reticulum move to the cell surface as potential ligands for circulating CD8+ T cells or natural killer (NK) cells1,3. Mechanisms to evade the immune response frequently target key components of the antigen-processing pathway to inhibit the formation and presentation of pMHCI7,8. Counter-measures have therefore evolved for the detection of the failure of various components of the antigen-processing pathway. For example, inhibition of peptide transport by TAP causes considerable loss of pMHCI from the cell surface9. The absence of pMHCI results in loss of recognition by CD8+ T cells but allows activation of NK cells released from pMHCImediated inhibition7,8. Similarly, immune responses can be inhibited or enhanced by changes in ERAAP activity. Expression of ERAP1, the human homolog of ERAAP, is inhibited in tumor cells10, and human cytomegalovirus inhibits ERAP1 function by a microRNAbased mechanism11. Additionally, polymorphisms in ERAP1 are associated with the autoimmune disorders ankylosing spondylitis and psoriasis12,13. It is not known if or how the immune system senses defects in ERAAP function. Loss of ERAAP function does not generally cause a large change in pMHCI expression on the cell surface1416. Nevertheless, loss of ERAAP substantially alters the repertoire of peptides presented by MHC class I1517. CTLs are elicited in wild-type mice immunized with ERAAP-deficient cells and vice versa17, which suggests that CTL-mediated recognition of changes in the pMHCI repertoire might indicate ERAAP dysfunction. However, the pMHCI ligands expressed by ERAAP-deficient cells that activate CTLs remain undefined. Here we used T cells as probes to analyze the immunogenic pMHCI ligands expressed uniquely by ERAAP-deficient cells. We found that cells lacking ERAAP elicited CD8+ T cells specific for both classical pMHCIa and nonclassical pMHCIb complexes. We identified the highly conserved peptide FYAEATPML (FL9) presented by the MHC class Ib molecule Qa-1b exclusively in ERAAP-deficient cells. Notably, we found that abundant antigen-experienced T cells specific for the
npg
Department of Molecular and Cell Biology, Division of Immunology and Pathogenesis, University of California, Berkeley, California, USA. Correspondence should be addressed to N.S. (nshastri@berkeley.edu). Received 16 December 2011; accepted 1 March 2012; published online 22 April 2012; doi:10.1038/ni.2282
NUMBER 6
JUNE 2012
579
Articles a
WT
10 10 10
5 4 3 2
APC
ERAAP-KO ERAAP-MHCIa-TKO
b
CD8+IFN-+ cells (%)
50
c
BEko8Z response (A595) 0.6
WT ERAAP-KO 0.6
ERAAP-MHCIa-TKO ERAAP-TAP-DKO
ERAAP-2m-DKO 0.8
1.8
41.8
36.7
CD8
25
10 0
0.3
0.3
0.4
IFN-
0 10 10 10
0 104 105 106
0 104 105 106
Figure 1 Wild-type T cells respond to pMHCIa and pMHCIb expressed by ERAAP-KO WT ERAAP-KO ERAAP-KO cells. (a) IFN- production by wild-type antiQa-1-KO ERAAPQa-1DKO ERAAP-KO CTLs incubated with wild-type (WT), ERAAP-KO or 0.8 Qa-2-null 0.6 ERAAP-MHCIa-TKO splenic APCs. Numbers adjacent to outlined areas indicate percent IFN-+ cells among all CD8+ cells. (b) Frequency of CD8+IFN-+ cells in cultures of anti0.4 0.3 ERAAP-KO CTLs incubated with wild-type or ERAAP-KO APCs (left two groups) or APCs that lacked both ERAAP and MHCIa, TAP or 2-microglobulin (2m; right three groups). Each symbol represents an individual mouse; small horizontal lines indicate the mean. (c) The lacZ 0 0 response of BEko8Z hybridoma cells to lipopolysaccharide-stimulated spleen cells used as APCs (key), 0 1 1,000 104 105 106 measured by spectrophotometry and presented as absorbance at 595 nm (A595). DKO, double Leucinethiol (M) Splenocytes (per well) deficiency (in ERAAP and either TAP or 2-microglobulin). (d) Response of BEko8Z cells (as in c) to lipopolysaccharide-stimulated Qa-1b-deficient (Qa-1-KO) or Qa-2-deficient (Qa-2-null) spleen cells (APCs) also treated with the aminopeptidase inhibitor leucinethiol. (e) Response of BEko8Z cells (as in c) to ERAAP-KO APCs or ERAAP-KO APCs also deficient in Qa-1b (ERAAPQa-1DKO). Data are from one experiment representative of three independent experiments (a,b) or are representative of three independent experiments (ce).
CD8+
ER W AA T PKO M H C la TA P 2m
104 105 106 Splenocytes (per well)
BEko8Z response (A595)
Qa-1bFL9 complex existed even in naive wild-type mice and proliferated to yield cytotoxic effector cells that rapidly eliminated cells lacking ERAAP in vitro as well as in vivo. Thus, nonclassical MHC class Ib molecules are sensors for defects in peptide processing in the endoplasmic reticulum. RESULTS ERAAP-deficient cells present immunogenic MHC class Ib ligands The immune system efficiently detects differences between self and non-self. Therefore, to identify immunologically important changes caused by ERAAP deficiency, we immunized wild-type (C57BL/6J H-2b) mice with MHC-matched spleen cells from ERAAP-deficient (ERAAP-KO) mice14. We generated wild-type CD8+ T cell lines directed against ERAAP-KO spleen cells by restimulating host spleen cells with ERAAP-KO antigen-presenting cells (APCs) in vitro (which resulted in anti-ERAAP-KO cells). Similar to published results17, CD8+ T cell lines generated from immunized mice responded strongly by producing more interferon- (IFN-) and tumor necrosis factor (TNF) when cultured with ERAAP-KO APCs than their background responses to self APCs (Fig. 1a,b and Supplementary Fig. 1). Similar fractions of the same CD8+ T cells responded to ERAAP-KO cells that also lacked any classical H-2Kb or H-2Db MHC class Ia molecules18 (ERAAP-MHCIa-TKO cells; 30% versus 25%, on average; Fig. 1a,b). MHC class I molecules were nevertheless required for the IFN- response, because APCs that lacked both ERAAP and 2-microglobulin, an essential structural subunit of all MHC class I molecules19,20, failed to stimulate wild-type anti-ERAAP-KO CD8+ T cells (Fig. 1b). Likewise, APCs deficient in both ERAAP and TAP9 also failed to stimulate CD8+ T cells (Fig. 1b), which demonstrated that their responses required peptide(s) transported into the endoplasmic reticulum. We therefore concluded that loss of ERAAP function resulted in the presentation of immunologically distinct peptides by MHC class Ia and class Ib molecules to wild-type CD8+ T cells. To further characterize those MHC class Ib ligands expressed by ERAAP-KO cells, we generated the monoclonal CD8+ T cell hybridoma BEko8Z, with inducible expressesion of -galactosidase (LacZ) in response to triggering of the T cell antigen receptor (TCR), by fusing
580
wild-type anti-ERAAP-KO CD8+ T cells with BWZ.36-CD8 cells (TCR BW5147 thymoma cell line with TCR-inducible expression of lacZ that expresses the coreceptor CD8)21. The BEko8Z hybridoma produced lacZ when cultured together with ERAAP-KO cells but not when cultured with wild-type splenic APCs (Fig. 1c, left). Consistent with our earlier observations (Fig. 1a), the ligand for BEko8Z T cells was expressed by ERAAP-KO APCs and did not require expression of classical MHC class Ia molecules (Fig. 1c, middle). BEko8Z cells did not respond to ERAAP-KO APCs that lacked 2-microglobulin or TAP (Fig. 1c). Thus, the BEko8Z hybridoma was specific for a TAP-transported peptide that was presented by a nonclassical MHC class Ib molecule exclusively on the surface of ERAAP-KO cells. Mice express a large number of nonclassical MHC class Ib molecules that serve as ligands for various innate as well as adaptive immune responses5. Among the three MHC class Ib molecules known to present peptides, we considered H-2M3 the least likely because these molecules present amino-formylated peptides that should be refractory to the aminopeptidase activity of ERAAP. We tested the other two molecules, Qa-1 and Qa-2, as potential ligands for BEko8Z T cells. We used splenocytes from Qa-1b-deficient mice22 and Qa-2-deficient (B6.K1) mice23 as APCs, after treating them with leucinethiol, a potent inhibitor of aminopeptidases, including ERAAP23 (Fig. 1d). The BEko8Z hybridoma responded to leucinethioltreated splenocytes from wild-type mice as well as those from Qa-2deficient B6.K1 mice but failed to recognize untreated splenocytes or those from Qa-1b-deficient mice (Fig. 1d). We further confirmed the requirement for Qa-1b by showing that BEko8Z cells did not respond to APCs from mice with genetic deficiency in both ERAAP and Qa-1b (Fig. 1e). We concluded that BEko8Z T cells recognized a unique peptide presented by the MHC class Ib molecule Qa-1b on ERAAP-KO APCs. Presentation of FL9 by Qa-1b to BEko8Z T cells Qa-1b is best known for presenting the Qdm peptide derived from the endoplasmic reticulumtranslocation signal sequence of classical MHC class Ia molecules24,25. The finding that the BEko8Z hybridoma recognized ERAAP-MHCIa-TKO APCs that lack H-2Db, which is the source
npg
Articles a
BEko8Z response (A595) 0.5
R3 R4 F1
R1 R4
Clone 23D9.15 Fam49b ORF F1
c
0.8 1.0 0.5
0.4
0.4 0.25 0
cDNA pool
MTNPAIQNDFSYYRRTLSRMRINNVPAEGENEVNNELANRMSLFYAEATPMLKTL MSLFYAEATPML R13 MSLFYAEATPM R14 R15 MSLFYAEATP
c R 3 R 4 R 1
Figure 2 The gene encoding Fam49b is the source of the antigen presented by Qa-1 b in the absence of ERAAP function. (a) Response of the BEko8Z hybridoma (as in Fig. 1c) to leucinethiol-treated, Qa-1b-expressing L cells transfected with cDNA pools. (b) Location of the gene encoding Fam49b in clone 23D9.15 (above), and of the PCR primers used for truncations (named for the reverse (R) PCR primer used), and locations of PCR primers used to establish the carboxyl terminus of the final peptide, on the amino acid sequence encoded by R13 (below). Underlining indicates the essential carboxy-terminal leucine residue. ORF, open reading frame. (c) Response of the BEko8Z hybridoma (as in Fig. 1c) to leucinethiol-treated, Qa-1b-expressing L cells transfected with minigenes encoding the peptides in b. Vec, vector. Data are from one experiment (a) or one experiment representative of two (c; mean and s.e.m. of duplicates).
of the Qdm peptide in C57BL/6 mice, suggested that BEko8Z T cells were specific for a different peptide. Additionally, neither a synthetic Qdm peptide (AMAPRTLLL) nor its analogs with amino-terminal extension were able to stimulate BEko8Z T cells (data not shown). Next we identified the naturally processed peptide presented by Qa-1b to BEko8Z cells by screening cDNA libraries prepared from ERAAP-KO cells26. To increase the efficiency of the screen, we first determined that CD11c+B220 dendritic cells from ERAAP-KO spleens were better APCs for stimulating BEko8Z than were unfractionated spleen cells or other splenocyte subsets (Supplementary Fig. 2). We constructed a cDNA expression library with mRNA from splenic dendritic cells isolated from ERAAP-KO mice treated with B16 cells expressing the cytokine Flt3L27. We fractionated the cDNA library into small pools and transfected those into Qa-1b-expressing L cells (H-2k mouse fibroblasts). Because the BEko8Z ligand was generated only in the absence of ERAAP, we treated the transfected cells briefly with leucinethiol to inhibit aminopeptidases and then assessed their ability to stimulate the BEko8Z hybridoma. One of the 2,880 pools screened stimulated the BEko8Z T cell hybridoma (Fig. 2a). After rescreening 24 individual cDNA clones from that cDNA pool, we identified three stimulatory clones and chose clone 23D9.15 for further analysis (Supplementary Fig. 3). BEko8Z T cells responded more strongly to APCs cotransfected with 23D9.15 cDNA and Qa-1b cDNA and treated with leucinethiol than to APCs treated with dithiothreitol alone (Supplementary Fig. 4), which suggested that the 23D9.15 plasmid encoded the antigenic peptide. All the stimulatory cDNAs encoded the hypothetical protein Fam49b of the National Center for Biotechnology Information database (accession code NM_144846). The motif of the peptides that that binds Qa-1b is unknown. Therefore, to identify the minimal antigenic peptide in Fam49b, we cloned a series of deletion constructs of 23D9.15 cDNA into mammalian expression vectors and assessed their ability to generate the BEko8Z ligand (Fig. 2b). The antigenic activity was produced in cells transfected with cDNA fragments R1 and R4 of Fam49b. In the next series of truncations of DNA fragment R4, we found that the carboxyterminal leucine encoded by fragment R13 was absolutely essential for T cellstimulatory activity, as truncated R14 DNA or R15 DNA, which lacked sequence encoding this leucine residue, was unable to stimulate the T cell hybridoma (Fig. 2c). The antigenic activity thus mapped to the carboxyl terminus of peptide encoded by DNA fragment R13. Natural processing and presentation of FL9 by Qa-1b To establish the identity of the antigenic peptide and define its precise boundary at the amino terminus, we assessed the ability panel
nature immunology VOLUME 13 NUMBER 6 JUNE 2012
of synthetic peptides of various lengths, which all terminated at the same essential leucine residue, to activate BEko8Z cells (Fig. 3a). The nine-residue peptide FYAEATPML (FL9) and the ten-residue peptide LFYAEATPML (LL10) were the most potent stimulators of BEko8Z T cells (Fig. 3a). In contrast, further addition to or deletion of one or two residues from the amino terminus of FL9 (11- and 12-residue peptides or 8- and 7-residue peptides, respectively) resulted in substantial loss of activity. Thus, the nonapeptide FL9 defined the minimum core sequence for the stimulation of BEko8Z T cells. We analyzed cell extracts to define the naturally processed peptide recognized by the BEko8Z hybridoma. We fractionated the extracts by reverse-phase HPLC and assessed their antigenic activity. We took advantage of the ability of BEko8Z cells to detect FL9 as well as its 10-, 11- and 12-residue analogs with amino-terminal extension (Fig. 3a). Each peptide eluted with a different retention time and could be readily distinguished from the others (Fig. 3b). We used cDNA encoding the R13 fragment of Fam49b (Fig. 2b) to transfect Qa-1b-expressing COS-7 monkey kidney cells (Fig. 3c). Because generation of the final antigenic peptide from the Fam49b cDNA was much lower in transfected cells when ERAAP function was not inhibited (Supplementary Fig. 4), we did all subsequent experiments in the presence of ERAAP inhibition. We extracted peptides from transfected cells after leucinethiol treatment to inhibit ERAAP and fractionated the peptides by HPLC. We assessed the ability of the fractions to stimulate BEko8Z cells in the presence of Qa-1b-expressing fibroblasts as APCs. We detected a single peak of antigenic activity that corresponded to the synthetic peptide FL9 in HPLC-fractionated extracts of cells transfected with R13 cDNA (Fig. 3c). We detected a peak with the same retention time in extracts of Qa-1b-expressing COS-7 cells transfected with a minigene encoding FL9 peptide alone (Fig. 3d). Furthermore, we detected the antigenic activity only in extracts of cells that expressed the restricting MHC class Ib molecule Qa-1b (Fig. 3e). The requirement for the restricting MHC molecule for detection of antigenic activity is similar to that observed with naturally processed peptides presented by classical MHC class Ia molecules28,29. We concluded that the nonapeptide FL9 corresponded to the naturally processed peptide presented by Qa-1b in the absence of ERAAP function. The Qa-1bFL9 complex is an immunodominant T cell ligand The Qa-1bQdm complex is recognized by the receptor CD94NKG2A expressed on NK cells and some activated T cells. To assess whether FL9 peptide presented by Qa-1b (Qa-1bFL9) was also recognized by a subset of NK cells, we stained wild-type spleen cells with Qa-1bFL9 tetramers obtained from the Tetramer Core Facility of
581
npg
R 15 R 14 R 13
Ve
Articles a
Fam49b R13
12:MSL--------11: SL--------10: L--------FL9: FYAEATPML 8: -------7: -------
b
BEko8Z response (A595) 1.1 0.6 0.2 0.1 0 20
FL FL9
10 11 12
Qa-1 + FL9 FL FL9 0.4 0.2 0 20 10
0.70 9 10 0.35 11 12 8 7
30 40 Fraction
50
0.2 0 20
10
Qa-1 + R13 0.4 FL FL9
30 40 Fraction
50
R13 only 0.4 0.2 0 20 FL9 10 L9
0 102 103 104 105 106 Peptide (pM)
30 40 Fraction
50
30 40 Fraction
50
Qa-1bFL9 tetramer (Fig. 4c,d). Likewise, a similar fraction of CD8+ T cells that produced IFN- in response to ERAAP-MHCIa-TKO APCs also bound the Qa-1bFL9 tetramer (Fig. 4c,d). Notably, a large fraction of T cells that produced IFN- in response to ERAAP-MHCIaTKO APCs did not bind the Qa-1bFL9 tetramer (Fig. 4c,d), which suggested that other pMHCIb complexes were also presented by these APCs. As expected, we detected only a background frequency of IFN--producing QFL T cells among T cells stimulated with APCs with genetic deficiency in both ERAAP and Qa-1b (Fig. 4c,d). We concluded that Qa-1bFL9 was an immunodominant ligand recognized by CD8+ T cells in wild-type mice immunized with ERAAP-KO cells, although we also detected other immunogenic complexes of peptide and MHC class Ia or MHC class Ib. Abundant antigen-experienced QFL T cells in naive mice We assessed the abundance of QFL T cells in the spleens of naive mice. We labeled spleen cells with Qa-1bFL9 tetramers conjugated to two different fluorophores to enhance specificity, followed by magnetic enrichment of cells positive for the Qa-1 bFL9 tetramer as described before for MHC class Ia and MHC class IIrestricted T cells3032 (Fig. 5a,b). On average, we detected 1,030 QFL T cells in a naive wild-type spleens (Fig. 5a,b), which suggested that QFL T cells were present at a frequency of ~1 in 1 104 splenic CD8+ T cells in naive wild-type mice. Enrichment for CD8+ T cells with the Qa-1bFL9 tetramer was specific, as the number of CD4+ cells positive for the Qa-1bFL9 tetramer enriched from wild-type mice was 2% of the corresponding number of CD8+ cells (Fig. 5a,b and Supplementary Fig. 5). In contrast, the spleens of TAP-deficient mice contained less than 5% of the cells specific for Qa-1bFL9 present in wild-type spleens (Fig. 5a,b). Most notably, we also found cells specific for Qa-1bFL9 in mice that lacked Qa-1b expression, but only at a frequency of about 30% of those in wild-type spleens (Fig. 5a,b). The frequency of cells specific for Qa-1bFL9 was higher than that of typical pMHCIa-specific precursors, which range from 1 in 1 104 to 1 in 1 105 total CD8+ T cells in combined spleens and lymph nodes of naive mice31,33, in contrast to the numbers we found here in spleens alone. We concluded that CD8+ QFL T cells were relatively abundant in normal wild-type mice. Furthermore, although the transport of antigenic peptides by TAP was crucial, the restricting Qa-1b molecule was not absolutely essential for the development of QFL T cells, which indicated that other MHC class I molecules were also able to support the development of QFL T cells.
Figure 3 FL9 is the naturally processed peptide presented by Qa-1 b molecules. (a) Response of the BEko8Z hybridoma (below; assessed as in Fig. 1c) to Qa-1b-expressing L cells loaded with synthetic peptides with progressive amino-terminal truncation or extension in the polypeptide encoded by R13 (above). (b) Response of the BEko8Z hybridoma (as in a) to peptides fractionated by reverse-phase HPLC (resulting in elution of synthetic FL9 and its analogs with amino-terminal extension) and loaded on Qa-1b-expressing L cells. (c,d) Elution of antigenic activity from peptide extracts obtained from Qa-1b-expressing COS-7 cells transfected with a minigene encoding R13 (Qa-1 + R13; c) or FL9 (Qa-1 + FL9; d) and fractionated as in b (assessed as in a). (e) Elution of antigenic activity from peptide extracts of COS-7 cells that do not express Qa-1b, transfected with the minigene encoding R13, analyzed as in c,d. Downward arrows indicate peak fractions. Data are from one experiment representative of three.
the US National Institutes of Health. In contrast to our detection of a distinct NKG2A+ population that bound the fluorescent Qa-1bQdm tetramer, we did not detect binding of the Qa-1bFL9 tetramer to either the NKG2A+ or NKG2A subset of NK cells (Fig. 4a,b). We concluded that Qa-1bFL9 was a ligand for CD8+ T cells rather than for NK cells. We then determined the frequency of T cells specific for the Qa-1bFL9 complex (called QFL T cells here) in wild-type mice that responded to ERAAP-KO cells. We obtained spleen cells from wildtype mice immunized with ERAAP-KO cells and restimulated the spleen cells in vitro with ERAAP-KO cells. We then measured the response of QFL T cells to APCs deficient in ERAAP alone or deficient in both ERAAP and MHC class Ia or both ERAAP and Qa-1b (Fig. 4c,d). A substantial fraction of wild-type CD8+ T cells that produced IFN- in response to ERAAP-KO APCs also bound the
npg
a
105 104 105 104
b
360 Tetramer (MFI)
c
APC:
105 104
d 45
17.8 22.6 0.5 Tet+ (%) ERAAP-KO ERAAP-MHCIaTKO ERAAPQa-1 DKO 30
Qa-1Qdm
Qa-1FL9
103 10 0
2
103 10 0
2
15
Qa-1FL9
180
103 10 0
2
0
0 102 103 104 105
0 102 103 104 105
0 102 103 104 105
dm
FL
IFN-
AA
Figure 4 The Qa-1bFL9 complex is an immunodominant T cell ligand. (a) Flow cytometry of spleen cells from naive mice labeled with the Qa-1 bQdm or Qa-1bFL9 tetramer together with markers for other spleen cell populations, showing tetramer labeling on gated B220 NK1.1+TCR cells. (b) Staining of NKG2A+NK1.1+TCR cells in a with the Qa-1bQdm tetramer (Qdm) or Qa-1bFL9 tetramer (FL9), presented as mean fluorescence intensity (MFI). (c) IFN- production by anti-ERAAP-KO CTLs incubated with T celldepleted splenic APC populations (above plots), then labeled with Qa-1bFL9 tetramers. Numbers above outlined areas indicate percent Qa-1 bFL9 tetramerpositive IFN-+ cells in the CD8+CD4 population. (d) Frequency of Qa-1bFL9 tetramerpositive (Tet+) cells among the CD4CD8+IFN-+ cells in c. Data are representative of two experiments (a,b; mean and s.e.m. of six mice) or three experiments (c) or are pooled from three independent experiments with three mice per group (d; mean and s.e.m.).
582
VOLUME 13
NUMBER 6
JUNE 2012
nature immunology
ER
ER
PM
NKG2A
NKG2A
ER
KO C AA Ia -T P KO Q a1 D KO H
AA P-
Articles
Figure 5 QFL T cells are antigen experienced in naive mice. (a) Flow cytometry of spleen cells from naive wild-type, TAP-deficient (TAP-KO) and Qa-1b-deficient (Qa-1KO) mice, stained with Qa-1bFL9 tetramers labeled with phycoerythrin (Tet-PE) or allophycocyanin (Tet-APC) and assessed before and after (Enriched) magnetic enrichment of tetramer-positive cells. Numbers in plots indicate average number of Qa-1bFL9 tetramer positive cells. (b) Qa-1bFL9 tetramerpositive cells per spleen, after enrichment (as in a). Each symbol represents an individual mouse; small horizontal lines indicate the mean. *P = 0.0043 and **P = 0.0007 (Mann-Whitney U-test). (c) Flow cytometry of enriched wild-type and Qa-1b-deficient Qa-1bFL9 tetramerpositive cells (as in a). Numbers adjacent to outlined areas indicate percent CD44hiTCR+ cells. (d) Frequency of CD44hi cells among Qa-1bFL9 tetramerpositive cells after enrichment (Tet+) or among total CD8+ cells before enrichment (right; as in a). Data are representative of six experiments (a,c) or are pooled from six independent experiments (b,d; mean and s.e.m.).
a
WT
Before
Enriched
b
Tet+ cells (per spleen) 104 103
* **
c
WT 57
1,030
102
10
5 4 3
10 TAP-KO 42
10
Q WT a1 K TA O PKO
Qa-1KO
CD44
10
102 0
13.2
3 4 0 10 10 10 10 2 5
10 10
5 4 3 2
d
390
010 10 10 10
2 3 4 5
TCR 70 Tet+
+ 70 Total CD8
10 0
CD44hi cells (%)
Tet-PE
Qa-1KO
10
35
35
Tet-APC
W T 1 KO
a-
We found that QFL T cells in naive wild-type mice expressed markers of antigen experience. Most cells positive for the Qa-1bFL9 tetramer expressed the memory markers CD44 and CD122 (the interleukin 2 receptor -chain; Fig. 5c and data not shown), compared with about 20% of total CD8+ T cells in the wild-type spleen that expressed CD44 and CD122 (Fig. 5c,d and data not shown). In contrast, most cells positive for the Qa-1bFL9 tetramer in Qa-1b-deficient mice did not have the antigen-experienced CD44hiCD122+ phenotype seen in wild-type mice (Fig. 5c,d and data not shown). The fraction of cells positive for the Qa-1bFL9 tetramer in Qa-1b-deficient mice that were also CD44hi was similar to the fraction of total CD8+ T cells in the same mice that were CD44hi at steady state (Fig. 5d). Thus, although the development of QFL T cells in Qa-1b-deficient mice did occur at a somewhat lower efficiency, acquisition of the antigen-experienced CD44hi phenotype was absolutely dependent on Qa-1b expression. QFL T cells proliferate in response to ERAAP-KO cells MHC class Ibrestricted T cells participate in primary immune responses in a variety of situations but are not thought to generate strong secondary responses3436. To determine how QFL T cells responded to ERAAP deficiency, we immunized wild-type mice with ERAAP-KO cells and
measured the frequency of QFL T cells ex vivo by tetramer staining. We analyzed mice either 10 d after immunization (during the effector phase of the response) or 812 weeks after immunization (during the memory phase). CD8+ QFL T cell populations expanded during the effector phase in wild-type mice immunized with ERAAP-KO cells (Fig. 6a,b). Notably, we also detected memory QFL T cells in wild-type mice 812 weeks after the initial immunization with ERAAP-KO cells (Fig. 6a,b). We restimulated spleen cells from the same mice with FL9 peptide immediately after isolation; intracellular cytokine staining for IFN- showed that both effector and memory QFL T cells were readily activated (Fig. 6c,d). The activation and population expansion of these cells were specific to challenge with ERAAP-KO cells, as we did not detect QFL T cells at similar frequencies in naive mice (Fig. 6) or in mice challenged with wild-type cells (data not shown). We obtained these results (Fig. 6) with mice that were not rechallenged before measurement of memory responses, and they showed that memory QFL T cell populations had expanded and were maintained at a high frequency after a single encounter with ERAAP-KO cells. We concluded that in contrast to other pMHCIb-restricted T cells3436, T cells specific for the ligand Qa-1bFL9 formed functional memory cells after encountering ERAAP-KO cells. Elimination of ERAAP-KO cells by MHC class Ibspecific T cells We hypothesized that CD8+ T cells specific for ERAAP-KO cells elicited in wild-type mice may carry out immunosurveillance of
Figure 6 Effector and memory QFL T cells are generated in response to ERAAP-KO cells. (a) Flow cytometry of spleen cells obtained from wildtype mice before (Naive) or 10 d (Effector) or 8 weeks (Memory) after immunization with ERAAP-KO cells, assessed (as in Fig. 5a) immediately after isolation (without magnetic enrichment). Numbers adjacent to outlined areas indicate percent Qa-1bFL9 tetramerpositive cells in the B220CD4CD8+ population. (b) Frequency of Qa-1bFL9 tetramerpositive cells assessed as in a (n = 3 mice per group). (c) Flow cytometry of spleen cells obtained from wild-type mice before (Naive) or 10 d (Effector) or 12 weeks (Memory) after immunization with ERAAP-KO cells then restimulated with FL9 peptide immediately after isolation. Numbers below outlined areas indicate percent CD8+IFN-+ cells among B220CD4 cells. (d) Frequency of IFN-+ cells assessed as in c (n = 3 mice per group). Data are from one representative of three experiments (mean and s.e.m. in b,d).
npg
a
10 10 10
5 4 3 2
Effector
Memory
Naive
b
Tet+ (%)
0.8
0.4
Tet-PE
10 0
2
0.75
0 10 10 10 10
3 4 5
0.3
0.03
Ef fe
c
10 10 10
5 4 3 2
CD8+IFN+ (%)
Effector
Memory
Naive
1.0
0.5
CD8
10 0
2 3
0.95
0 10 10 10 10
4 5
0.18
0.08
Ef
ct o em r or y N ai ve
IFN-
fe
ct o em r or y N ai ve
Tet-APC
a-
W T 1 KO
NUMBER 6
JUNE 2012
583
Articles a
Specific lysis (%) 30 15
b
Specific lysis (%)
100
Specific lysis (%)
RMA RMA-S
30 20 10 0 0
B16 + FL9 B16
50
0 10 100 Leucinethiol (M)
d
Naive
ER W AA T PK M H ER O C A Ia A -T P KO -
2.5 5.0 Effector/Target
90 Targets lost (%)

80 60 40 20 0 100 80 60 40 20 0 100 101 102 103
**
e
40 Targets lost (%)
60 30 0
20
at approximately the same ratio (1:1) as in the input cell mixture. In contrast, ERAAP-KO cells were specifically rejected without affecting the recovery of self cells by wild-type mice primed with ERAAP-KO cells 7 d earlier (Fig. 7d). As expected, ERAAP-MHCIa-TKO cells were also rejected with similar efficiency in primed mice. Notably, the elimination or loss of target cells required CD8+ T cells rather than CD4+ T cells, as mice depleted of CD8+ cells no longer rejected the ERAAP-KO target cells, unlike mice depleted of CD4+ T cells, which did reject the ERAAP-KO target cells (Fig. 7e). In contrast to another study suggesting a role for NK cells in the rejection of ERAAP-KO cells37, we did not detect a role for NK cells in our experiments with mice depleted of NK1.1+ cells (Fig. 7b). We concluded that ERAAP deficiency elicited a potent MHC class Ibrestricted CD8+ T cell response in wild-type mice that effectively eliminated ERAAP-KO cells in vitro as well as in vivo. DISCUSSION The pMHCI repertoire is the immune systems window into the state of the cellular proteome. Mechanisms for evading the immune response often target the antigen-processing pathway to inhibit the presentation of pMHCI. Immune-surveillance mechanisms, in turn, detect defects in pMHCI presentation. Here we have described an MHC class Ibrestricted CD8+ T cell mechanism that sensed normal antigen processing in the endoplasmic reticulum. Immune responses are influenced by alterations in ERAAP activity. A microRNA encoded by human cytomegalovirus specifically targets the predominant isoform of human ERAAP (hERAP1) 11. Loss of hERAP1 function inhibits the presentation of viral epitopes and probably contributes to the chronic infections established by human cytomegalovirus. Likewise, changes in ERAAP expression in tumors might lead to the activation of NK cells37. Finally, polymorphisms in hERAP1 have been found to be associated with two autoimmune disorders, ankylosing spondylitis and psoriasis12,13. Thus, impaired ERAAP function can cause various immune disorders. Analysis of the MHC class Ibound peptide pool in ERAAPdeficient cells has shown a distinct shift toward the presentation of peptides with amino-terminal extensions, as well as peptides unique to ERAAP-deficient cells, by MHC class Ia molecules1417,38. Here we have shown that the changes in the pMHCI repertoire in ERAAP-KO cells also extended to MHC class Ib molecules. Furthermore, the unique immunogenic pMHCIb ligands on ERAAPKO cells were immunodominant, inducing a large fraction of wildtype anti-ERAAP-KO CD8+ T cells. An MHC class Ibspecific response of this magnitude was not detected in an earlier study17, although a small fraction of wild-type anti-ERAAP-KO CTLs in that study were activated by MHC class Iadeficient APCs. These differences in the relative abundance of MHC class Ia and MHC class Ibrestricted CTLs may have arisen from differences in the culture conditions for in vitro restimulation. Also, in our study here, the Qa-1bFL9 tetramer allowed direct measurement of ligandspecific CTLs. We have identified Fam49b as the source of the antigenic peptide presented by Qa-1b in ERAAP-KO cells. Fam49b is highly conserved in vertebrates from zebrafish to humans and seems to have ubiquitous expression, as noted in the gene atlas of the European Bioinformatics Institute (http://www.ebi.ac.uk/gxa/gene/ENSMUSG00000022378), although its function is unknown. Notably, high expression of FAM49B has been detected in patients with relapsing multiple sclerosis and in nonsmall cell lung cancer tissues39,40. We suggest, therefore, that presentation of the peptide FL9 by Qa-1b could be a highly conserved mechanism for the detection of ERAAP deficiency.
Cells
Targets:
2 100 101 10 103
PH ER KO C A Ia A -T E R KO PAA PM H ER KO C A Ia A -T P KO -
Primed
ER AA
CFSE
Naive
Primed
Figure 7 Wild-type CTLs specifically eliminate ERAAP-KO cells expressing unique pMHCIb complexes. (a) Lysis of wild-type, ERAAP-KO or ERAAPMHCIa-TKO spleen cells by anti-ERAAP-KO CTLs (as in Fig. 1a) at an effector/target ratio of 2.5:1. (b) Lysis of leucinethiol-treated RMA or RMA-S cells by wild-type anti-ERAAP-KO CTLs at an effector/target ratio of 2.5:1. (c) Lysis of B16 melanoma cells, with (B16 + FL9) or without (B16) stable expression of FL9, by wild-type anti-ERAAP-KO CTLs. (d) Flow cytometry (left) of cells from naive wild-type mice (Naive) or wild-type mice primed with ERAAP-KO cells (Primed), then challenged (7 d after priming) with wild-type (self) spleen cells labeled with 0.25 M CFSE or ERAAP-KO spleen cells labeled with 2.5 M CFSE. Right, pooled results of mice challenged with wild-type (self) spleen cells labeled with 0.25 M CFSE or ERAAP-KO or ERAAP-MHCIa-TKO spleen cells labeled with 2.5 M CFSE. *P < 0.0001 **P < 0.0005 (Mann-Whitney U-test). (e) Flow cytometry as in d, without depletion (None) or with depletion of CD4+ cells (CD4) or CD8+ cells (CD8) 24 h before challenge. Data are pooled from three independent experiments with CTLs from a total of 12 mice (a; mean and s.e.m.), are representative of three experiments (b), are from one experiment representative of three independent experiments (c) or are pooled from three (d) or two (e) independent experiments (mean and s.e.m.).
npg
ERAAP deficiency. We assessed whether CD8+ T cells from wildtype mice immunized with ERAAP-KO cells were able to eliminate ERAAP-KO target cells. Indeed, wild-type anti-ERAAP-KO CTLs (generated as in Fig. 1a) were cytotoxic and efficiently killed both ERAAP-KO target cells and ERAAP-MHCIa-TKO target cells in vitro (Fig. 7a). Furthermore, wild-type anti-ERAAP-KO CTLs also killed TAP-sufficient RMA mouse tumor cells that were treated for 5 h with leucinethiol to inhibit ERAAP, but did not kill leucinethioltreated TAP-deficient RMA-S mouse tumor cells (Fig. 7b). Notably, B16 mouse melanoma cells transfected with a minigene encoding FL9 were also effectively killed by wild-type anti-ERAAP-KO CTLs, despite the absence of ERAAP inhibition, but untransfected B16 cells were not (Fig. 7c). Thus, expression of FL9 was both necessary and sufficient for the elimination of target cells. We concluded that wild-type CTLs, elicited by ERAAP-KO cells, were able to detect and eliminate ERAAP-KO or FL9-presenting cells. To assess whether ERAAP-KO cells could also be eliminated in vivo, we injected wild-type mice with a mixture of self and ERAAPKO cells labeled with two different concentrations of the intracellular dye CFSE. The input cells could therefore be distinguished from each other and from unlabeled host cells by flow cytometry (Fig. 7d, left). We recovered self and ERAAP-KO fluorescent cells from naive mice
584
N on
e C D 8 C D 4
0 Depletion:
Articles
We used the Qa-1bFL9 tetramer to characterize lymphocytes specific for this ligand. We were unable to detect NK cells specific for the Qa-1bFL9 tetramer, in contrast to the readily detectable subset of NKG2A+ NK cells specific for Qa-1bQdm. However, a large fraction of the CD8+ T cells elicited by ERAAP-KO cells was positive for the Qa-1bFL9. Using Qa-1bFL9 tetramers, we also enriched a sizeable number of CD8+ QFL T cells from naive wild-type mice, which suggested that these cells were present at a relatively high frequency. Notably, even in naive mice, QFL T cells had an antigen-experienced, CD44hiCD122hi phenotype, similar to that of other innate-like MHC class Ibspecific T cells, such as NK T cells and mucosa-associated invariant T cells41. Whether QFL T cells share other developmental and functional characteristics with these innate-like effector cells is not known. Unexpectedly, we found QFL T cells in Qa-1b-deficient mice, albeit at a lower frequency than in wild-type mice. We did not, however, detect QFL T cells in TAP- or 2-microglobulin-deficient mice, which suggested that QFL T cells require the presentation of peptides by an MHC class Ib molecule other than Qa-1b for development. Notably, acquisition of the antigen-experienced phenotype did require Qa-1b expression, because QFL T cells in naive Qa-1bdeficient mice did not constitutively express CD44 or CD122. MHC class Ibrestricted T cells generally have an antigen-experienced phenotype in naive mice, possibly because they are selected by APCs of hematopoietic origin42,43. Our observations have shown that QFL T cells expressed markers of antigen experience only after encountering Qa-1b-presented peptides. One implication of these findings is that QFL T cells may have encountered their ligand even in naive mice. Because FL9 was produced exclusively in ERAAP-KO cells, the Qa-1bFL9 complex or a cross-reactive ligand might have been generated during transient ERAAP deficiency in wild-type mice, caused, perhaps, by natural transformation events, endogenous viruses or commensal microbes. CD8+ T cells elicited by immunization with ERAAP-KO cells eliminated both MHC class Ia and MHC class Ibexpressing targets cells that lacked ERAAP. Wild-type anti-ERAAP-KO T cell lines also killed tumor cells in which ERAAP was inhibited and ERAAP-sufficient cells that expressed FL9, which suggested that expression of this pMHCIb complex was sufficient for the elimination of target cells. QFL T cell populations expanded robustly in response to a single challenge with ERAAP-KO cells and, unusually for a pMHCIb-specific response, also formed stable memory T cells that retained their effector function without rechallenge. QFL T cells might therefore be able to mediate long-term immunosurveillance of ERAAP deficiency. Qa-1bFL9 complexes were rapidly presented by APCs treated for 5 h with an ERAAP inhibitor. Qa-1bFL9 complexes were also a specific sign of ERAAP dysfunction and were not presented by TAP-deficient cells, unlike the Qa-1-associated peptides isolated from TAP-deficient cells described before44. Thus, the peptides presented by Qa-1b reflect the state of the antigen-processing pathway. In conclusion, ERAAP deficiency resulted in the presentation of a highly conserved pMHCIb complex, Qa-1bFL9, on the cell surface. Abundant antigen-experienced CD8+ T cells specific for Qa-1bFL9 existed in naive mice, proliferated and effectively eliminated ERAAPKO cells. This response seems to combine the best of both worlds: the exquisite specificity of adaptive immunity, and the rapid responsiveness of the innate immune system. METHODS Methods and any associated references are available in the online version of the paper.
Note: Supplementary information is available in the online version of the paper. AckNowledGmeNtS We thank the Tetramer Core Facility of the US National Institutes of Health for tetramer reagents; K. Sderstrom and E. Engleman (Stanford University) for Qa-1b-deficient mice generated by H. Cantor and colleagues (Harvard University); H. Cantor (Harvard University) for lentiviral vectors expressing Qa-1b; D. King for peptide synthesis; H. Nolla for assistance with cell sorting; K.C. Lind and A.H. Bakker for discussions and comments on the manuscript; and H. Dani for technical assistance. Supported by Irvington Institute Fellowship Program of the Cancer Research Institute (N.A.N.) and the US National Institutes of Health (N.S.). AUtHoR coNtRIBUtIoNS N.A.N. and N.S. designed the study and wrote the manuscript; and N.A.N. and F.G. did the experiments. comPetING FINANcIAl INteReStS The authors declare no competing financial interests.
1. Shastri, N., Schwab, S. & Serwold, T. Producing natures gene-chips: the generation of peptides for display by MHC class I molecules. Annu. Rev. Immunol. 20, 463493 (2002). 2. Williams, M.A. & Bevan, M.J. Effector and memory CTL differentiation. Annu. Rev. Immunol. 25, 171192 (2007). 3. Peaper, D.R. & Cresswell, P. Regulation of MHC class I assembly and peptide binding. Annu. Rev. Cell Dev. Biol. 24, 343368 (2008). 4. Serwold, T., Gonzalez, F., Kim, J., Jacob, R. & Shastri, N. ERAAP customizes peptides for MHC class I molecules in the endoplasmic reticulum. Nature 419, 480483 (2002). 5. Rodgers, J.R. & Cook, R.G. MHC class Ib molecules bridge innate and acquired immunity. Nat. Rev. Immunol. 5, 459471 (2005). 6. Strong, R.K. et al. HLA-E allelic variants. Correlating differential expression, peptide affinities, crystal structures, and thermal stabilities. J. Biol. Chem. 278, 50825090 (2003). 7. Hansen, T.H. & Bouvier, M. MHC class I antigen presentation: learning from viral evasion strategies. Nat. Rev. Immunol. 9, 503513 (2009). 8. Tortorella, D., Gewurz, B.E., Furman, M.H., Schust, D.J. & Ploegh, H.L. Viral subversion of the immune system. Annu. Rev. Immunol. 18, 861926 (2000). 9. Van Kaer, L., Ashton-Rickardt, P.G., Ploegh, H.L. & Tonegawa, S. TAP1 mutant mice are deficient in antigen presentation, surface class I molecules, and CD48+ T cells. Cell 71, 12051214 (1992). 10. Fruci, D. et al. Altered expression of endoplasmic reticulum aminopeptidases ERAP1 and ERAP2 in transformed non-lymphoid human tissues. J. Cell Physiol. 216, 742749 (2008). 11. Kim, S. et al. Human cytomegalovirus microRNA miR-US41 inhibits CD8+ T cell responses by targeting the aminopeptidase ERAP1. Nat. Immunol. 12, 984991 (2011). 12. Evans, D.M. et al. Interaction between ERAP1 and HLA-B27 in ankylosing spondylitis implicates peptide handling in the mechanism for HLA-B27 in disease susceptibility. Nat. Genet. 43, 761767 (2011). 13. Strange, A. et al. A genome-wide association study identifies new psoriasis susceptibility loci and an interaction between HLA-C and ERAP1. Nat. Genet. 42, 985990 (2010). 14. Hammer, G.E., Gonzalez, F., Champsaur, M., Cado, D. & Shastri, N. The aminopeptidase ERAAP shapes the peptide repertoire displayed by major histocompatibility complex class I molecules. Nat. Immunol. 7, 103112 (2006). 15. Yan, J. et al. In vivo role of ER-associated peptidase activity in tailoring peptides for presentation by MHC class Ia and class Ib molecules. J. Exp. Med. 203, 647659 (2006). 16. York, I.A., Brehm, M.A., Zendzian, S., Towne, C.F. & Rock, K.L. Endoplasmic reticulum aminopeptidase 1 (ERAP1) trims MHC class I-presented peptides in vivo and plays an important role in immunodominance. Proc. Natl. Acad. Sci. USA 103, 92029207 (2006). 17. Hammer, G.E., Gonzalez, F., James, E., Nolla, H. & Shastri, N. In the absence of aminopeptidase ERAAP, MHC class I molecules present many unstable and highly immunogenic peptides. Nat. Immunol. 8, 101108 (2007). 18. Vugmeyster, Y. et al. Major histocompatibility complex (MHC) class I KbDb/ deficient mice possess functional CD8+ T cells and natural killer cells. Proc. Natl. Acad. Sci. USA 95, 1249212497 (1998). 19. Koller, B.H., Marrack, P., Kappler, J.W. & Smithies, O. Normal development of mice deficient in 2M, MHC class I proteins, and CD8+ T cells. Science 248, 12271230 (1990). 20. Zijlstra, M. et al. 2-microglobulin deficient mice lack CD48+ cytolytic T cells. Nature 344, 742746 (1990). 21. Sanderson, S. & Shastri, N. LacZ inducible, antigen/MHC-specific T cell hybrids. Int. Immunol. 6, 369376 (1994).
npg
585
Articles
22. Hu, D. et al. Analysis of regulatory CD8+ T cells in Qa-1-deficient mice. Nat. Immunol. 5, 516523 (2004). 23. Serwold, T., Gaw, S. & Shastri, N. ER aminopeptidases generate a unique pool of peptides for MHC class I molecules. Nat. Immunol. 2, 644651 (2001). 24. Lu, L., Werneck, M.B. & Cantor, H. The immunoregulatory effects of Qa-1. Immunol. Rev. 212, 5159 (2006). 25. Aldrich, C.J. et al. Identification of a TAP-dependent leader peptide recognized by alloreactive T cells specific for a class 1b antigen. Cell 79, 649658 (1994). 26. Karttunen, J., Sanderson, S. & Shastri, N. Detection of rare antigen-presenting cells by the lacZ T-cell activation assay suggests an expression cloning strategy for T-cell antigens. Proc. Natl. Acad. Sci. USA 89, 60206024 (1992). 27. Mach, N. et al. Differences in dendritic cells stimulated in vivo by tumors engineered to secrete granulocyte-macrophage colony-stimulating factor or Flt3-ligand. Cancer Res. 60, 32393246 (2000). 28. Falk, K., Rtzschke, O. & Rammensee, H.G. Cellular peptide composition governed by major histocompatibility complex class I molecules. Nature 348, 248251 (1990). 29. Malarkannan, S., Goth, S., Buchholz, D.R. & Shastri, N. The role of MHC class I molecules in the generation of endogenous peptide/MHC complexes. J. Immunol. 154, 585598 (1995). 30. Moon, J.J. et al. Naive CD4+ T cell frequency varies for different epitopes and predicts repertoire diversity and response magnitude. Immunity 27, 203213 (2007). 31. Obar, J.J., Khanna, K.M. & Lefrancois, L. Endogenous naive CD8+ T cell precursor frequency regulates primary and memory responses to infection. Immunity 28, 859869 (2008). 32. Moon, J.J. et al. Tracking epitope-specific T cells. Nat. Protoc. 4, 565581 (2009). 33. Haluszczak, C. et al. The antigen-specific CD8+ T cell repertoire in unimmunized mice includes memory phenotype cells bearing markers of homeostatic expansion. J. Exp. Med. 206, 435448 (2009). 34. Kerksiek, K.M., Busch, D.H., Pilip, I.M., Allen, S.E. & Pamer, E.G. H2M3-restricted T cells in bacterial infection: rapid primary but diminished memory responses. J. Exp. Med. 190, 195204 (1999). 35. Cho, H., Choi, H.J., Xu, H., Felio, K. & Wang, C.R. Nonconventional CD8+ T cell responses to Listeria infection in mice lacking MHC class Ia and H2M3. J. Immunol. 186, 489498 (2011). 36. Bouwer, H.G., Barry, R.A. & Hinrichs, D.J. Lack of expansion of major histocompatibility complex class Ib-restricted effector cells following recovery from secondary infection with the intracellular pathogen Listeria monocytogenes. Infect. Immun. 69, 22862292 (2001). 37. Cifaldi, L. et al. Natural killer cells efficiently reject lymphoma silenced for the endoplasmic reticulum aminopeptidase associated with antigen processing. Cancer Res. 71, 15971606 (2011). 38. Blanchard, N. et al. Endoplasmic reticulum aminopeptidase associated with antigen processing defines the composition and structure of MHC class I peptide repertoire in normal and virus-infected cells. J. Immunol. 184, 30333042 (2010). 39. Petroziello, J. et al. Suppression subtractive hybridization and expression profiling identifies a unique set of genes overexpressed in non-small-cell lung cancer. Oncogene 23, 77347745 (2004). 40. Gilli, F. et al. Learning from nature: pregnancy changes the expression of inflammation-related genes in patients with multiple sclerosis. PLoS ONE 5, e8962 (2010). 41. Yamagata, T., Benoist, C. & Mathis, D. A shared gene-expression signature in innate-like lymphocytes. Immunol. Rev. 210, 5266 (2006). 42. Urdahl, K.B., Sun, J.C. & Bevan, M.J. Positive selection of MHC class Ib-restricted CD8+ T cells on hematopoietic cells. Nat. Immunol. 3, 772779 (2002). 43. Cho, H., Bediako, Y., Xu, H., Choi, H.J. & Wang, C.R. Positive selecting cell type determines the phenotype of MHC class Ib-restricted CD8+ T cells. Proc. Natl. Acad. Sci. USA 108, 1324113246 (2011). 44. Oliveira, C.C. et al. The nonpolymorphic MHC Qa-1b mediates CD8+ T cell surveillance of antigen-processing defects. J. Exp. Med. 207, 207221 (2010).
npg
586
VOLUME 13
NUMBER 6
JUNE 2012
nature immunology
Mice. ERAAP-KO mice (B6.129.Arts1tm1Ucb) have been described14. ERAAPMHCIa-TKO mice and ERAAP-KO mice also deficient in TAP, 2-microglobulin or Qa-1 were generated in our facility by the crossing of ERAAP-KO mice with mice deficient in both H-2Kb and H-2Db, TAP1, 2-microglobulin or Qa-1b, respectively. Wild-type C57BL/6J, Qa-2-deficient B6.K1 and 2-microglobulindeficient mice were from the Jackson Laboratory or were bred in our facility. H-2T23-deficient Qa-1b-deficient mice were generated in the laboratory of H. Cantor and were a gift from E. Engleman. Mice were housed and all procedures were done with the approval of and in accordance with the institutional animal care guidelines of the University of California Berkeley. Antibodies, cell lines and peptides. Antibodies for flow cytometry were from BD Biosciences (antibody to CD8 (anti-CD8; 53-6.7), anti-CD4 (RM4-5), antiB220 (RA36B2), anti-TCR (H57-597), anti-CD44 (IM7), anti-NK1.1 (PK136), anti-Qa-1b (6A8.6F10.1A6), anti-CD11c (HL3), anti-H-2Kb (AF6-88.5), anti-H-2Db (KH95), anti-I-Ab (25-9-17), anti-IFN- (XMG1.2) and anti-TNF (MP6-XP22)) and eBioscience (anti-NKG2A (20D5)). Antibodies used for depletion of cells in vivo were from the Cell Culture Facility of the University of California, San Francisco. All peptides were synthesized by D. King (University of California Berkeley). Qa-1b-expressing L cells and COS cells were made by stable transduction of L cells (mouse fibroblasts lacking thymidine kinase (Lmtk cells)) and COS-7 cells, respectively, with lentiviral vector expressing Qa-1b (H. Cantor). B16.FL9 cells were made by transfection of B16.BL6 cells with a plasmid containing the minigene encoding FL9 and sequence neomycin resistance. Stable cell lines were derived from single cells by selection with limiting dilution of the aminoglycoside G418. CTL- and T cellactivation assays. Wild-type CTL lines that recognize ERAAP-KO cells were generated by immunization of C57BL/6J mice once intraperitoneally with 2 107 ERAAP-KO spleen cells. Spleens were collected from immunized mice 10 d after immunization. Responder spleen cells (5 106 cells per well in 24-well plates) were restimulated in vitro with an equal number of irradiated ERAAP-KO spleen cells and 20 U/ml of recombinant human interleukin 2 (BD Biosciences). T cells were used for IFN- assays 6 d after restimulation and for in vitro cytotoxicity assays 5 d after restimulation. For measurement of intracellular IFN-, CTL lines were collected and stimulated for 5 h with lipopolysaccharide-stimulated spleen APC populations that were depleted of CD4+ cells and CD8+ cells. GolgiPlug (BD Biosciences) was added for the final 4 h of the incubation period. Cells were then stained first for surface markers or with tetramers, followed by other surface markers, then were fixed, permeabilized and stained for intracellular IFN-. For ex vivo analysis of IFN- production, spleen cells were collected from immunized mice and plated for 5 h in 96-well plates with 1 M peptide in the presence of GolgiPlug. Cells were then stained as described above. Hybridomas and expression cloning. Wild-type anti-ERAAP-KO CTL lines collected 4 d after in vitro restimulation were fused to the CD8+TCR cell line BWZ.36 that expresses a TCR-inducible gene encoding lacZ. Putative positive clones were screened with lipopolysaccharide-stimulated ERAAP-KO spleen cells as APCs, and positive wells were subcloned by limiting dilution to obtain clones derived from a single cell. Once a stable hybridoma BEko8Z was established, it was used for screening of cDNA libraries as described 21. ERAAP-KO mice were immunized subcutaneously with 3 106 Flt3L-secreting B16 melanoma cells as described27. Two weeks later, their spleens were collected and blasted overnight with 200 ng/ml lipopolysaccharide (Sigma Aldrich). Samples were enriched forCD11c+ cells through the use of CD11c microbeads followed by magnetic enrichment with LS columns (Miltenyi Biotec). The resultant cells were pelleted and frozen, then homogenized with a rotor-stator homogenizer. An Oligotex mRNA isolation kit (Qiagen) was used for isolation of mRNA. Then, cDNA was synthesized and cloned into pCMV-SPORT6 with the SuperScript Plasmid kit from Invitrogen, and transformed into bacteria. Plasmid DNA was isolated from bacterial cultures grown in 96-well plates and was used for transfection of Qa-1b-expressing L cells. Leucinethiol (15 M in 500 M dithiothreitol) was added 42 h later, then BEko8Z hybridoma cells were added 6 h after that, followed by overnight incubation. T cell activation was detected by colorimetric cleavage of CPRG
ONLINE METHODS
(chlorophenylred--d-galactopyranoside; Roche) by TCR-induced lacZ. When a positive well was detected, 1 l cDNA from that well was transformed in bacteria and plasmid DNA was isolated from multiple individual colonies. Each individual cDNA was transfected into Qa-1b-expressing L cells. Each positive cDNA was sequenced and identified. Progressive truncations were made in the coding region of the gene encoding Fam49b with nested PCR primers; the resultant fragments were cloned into pcDNA1 and their activity was assessed by transfection into Qa-1b-expressing L cells. Once the candidate region was narrowed to R13 and the final leucine residue was determined to be essential, peptides with amino-terminal extension were synthesized and tested. Minigenes were cloned in the pcDNA1 expression vector and were designed to encode the desired amino acid sequence with an ATG initiation codon. The minigene encoding FL9, for example, encodes the amino acid sequence Met-FL9 (MFYAEATPML). Enrichment for tetramer-positive cells. Qa-1bFL9 tetramers were synthesized by the Tetramer Core Facility of the US National Institutes of Health. Samples were enriched for tetramer-positive cells as described32. Spleens were collected from mice and homogenized with nylon filters. Red blood cells were lysed and the cells were resuspended in 200 l sorter buffer (0.1% sodium azide and 5% FCS in PBS). Phycoerythrin- or allophycocyanin-labeled Qa-1bFL9 tetramers were added at a final dilution of 1:200. Cells were incubated for 30 min at 23 C, then were washed and resuspended in 0.45 ml sorter buffer. Anti-PE microbeads (50 l; Miltenyi Biotec) were added to each sample, followed by incubation for 15 min at 48 C. Cells were washed, and phycoerythrin-labeled cells were isolated by passage over an LS magnetic column (Miltenyi Biotec). The entire positively selected fraction was collected and stained with anti-B220, anti-CD4, anti-CD8, anti-TCR and anti-CD44. Tetramer-positive cells were gated as B220CD4CD8+TCR+ phycoerythrin-tetramer-positive and allophycocyanin-tetramer-positive cells. A fixed number of CountBright beads (Invitrogen) was added to each sample to allow counting of cells, and samples were acquired on a BD LSR II. Reverse-phase HPLC. COS-7 or Qa-1b-expressing COS cells were transfected with cDNA encoding either Fam49b R13, or a minigene encoding the FL9 peptide. Leucinethiol (15 M in 500 M dithiothreitol) was added 42 h later. After 5 h of leucinethiol treatment, cells were collected and peptides were extracted with 10% acetic acid. Extracts were filtered through a 10-kilodalton exclusion filter, and all the material less than 10 kilodaltons in size was fractionated with a C18 reversed-phase HPLC column (Grace Vydac). The fractionated material was collected in 96-well plates and lyophilized. Qa-1b-expressing L cells (5 104) and BEko8Z cells (1 105) were added to each well, followed by incubation overnight at 37 C. The presence of activating peptide was detected the next day by colorimetric conversion of CPRG. Synthetic peptides (1 pmol each) were also assessed during the same experiment and detected in the same way. Killing assays. For in vitro cytotoxicity assays, cells used as targets were labeled with the dye PKH26 (Sigma Aldrich), counted and plated. Effector CTL lines were collected and added at the appropriate ratio of effector cell to target cell, followed by incubation for 1 h and 45 min at 37 C, after which the dye YoPRO-1 (Invitrogen) was added to a final concentration of 1 M for visualization of apoptotic cells. Cells were washed and visualized on a flow cytometer. Cells that were both PKH26+ and YoPRO-1+ were considered apoptotic target cells. The frequency of apoptotic target cells was calculated as follows: 100 (percent PKH26+YoPRO-1+ E:T = X percent PKH26+YoPRO-1+ E:T = 0)/(100 percent PKH26+YoPRO-1+ E:T = 0), where E:T = X indicates the ratio of effector cells to target cells at time X and E:T = 0 indicates that ratio at time 0. For in vivo cytotoxicity assays, wild-type C57BL/6 and ERAAP-KO spleen cells were collected and labeled with 0.25 M and 2.5 M of carboxyfluorescein-succinimidyl ester (CFSE, Invitrogen), respectively. Wild-type (CFSE lo) cells were mixed with ERAAP-KO or ERAAP-MHCIa-TKO (CFSEhi) cells (5 106 cells of each genotype) and their ratio was measured immediately before injection (input ratio). Target cells were injected intravenously into naive mice or mice that had been primed 7 days before by intraperitoneal injection of ERAAP-KO spleen cells. Spleens from mice injected with target cells were collected 24 h later and the ratio of CFSElo cells to CFSEhi cells
npg
doi:10.1038/ni.2282
nature immunology
was measured by flow cytometry. Targets lost in a sample were calculated as follows: 100 100 (fraction CFSE hi cells/input fraction CFSE hi cells), where fraction CFSEhi cells was calculated as percent CFSEhi cells / (percent CFSEhi cells plus percent CFSElo cells). Mice that received ERAAP-MHCIaTKO cells as targets were depleted of NK cells with 200 g anti-NK1.1 36 h before challenge. In experiments in which the role of CD4 + cells and CD8+
cells was assessed, samples were depleted of CD4 + cells and CD8+ cells with 200 g anti-CD8 (YTS169.4) or anti-CD4 (GK1.5) 24 h before challenge. Statistical analyses. GraphPad Prism was used for all statistical analyses. A nonparametric Mann-Whitney U-test was used for estimation of statistical significance.
npg
nature immunology
doi:10.1038/ni.2282
Articles
Clonal deletion and the fate of autoreactive thymocytes that survive negative selection
Leonid A Pobezinsky1, Georgi S Angelov2, Xuguang Tai1, Susanna Jeurling1, Franois Van Laethem1, Lionel Feigenbaum3, Jung-Hyun Park1 & Alfred Singer1
Clonal deletion of autoreactive thymocytes is important for self-tolerance, but the intrathymic signals that induce clonal deletion have not been clearly identified. We now report that clonal deletion during negative selection required CD28-mediated costimulation of autoreactive thymocytes at the CD4+CD8lo intermediate stage of differentiation. Autoreactive thymocytes were prevented from undergoing clonal deletion by either a lack of CD28 costimulation or transgenic overexpression of the antiapoptotic factors Bcl-2 or Mcl-1, with surviving thymocytes differentiating into anergic CD4 CD8 double-negative thymocytes positive for the T cell antigen receptor ab subtype (TCRab) that preferentially migrated to the intestine, where they re-expressed CD8a and were sequestered as CD8aa+ intraepithelial lymphocytes (IELs). Our study identifies costimulation by CD28 as the intrathymic signal required for clonal deletion and identifies CD8aa+ IELs as the developmental fate of autoreactive thymocytes that survive negative selection. Immunocompetent T cells must be reactive to foreign pathogens but tolerant of self ligands. Those critical features of T cell immunity are imposed by selection events in the thymus that determine the developmental fate of each individual T cell depending on the specificity of its T cell antigen receptor (TCR). Differentiation in the thymus proceeds in an ordered sequence characterized by expression of the coreceptors CD4 and CD8 in which the earliest cells are CD4CD8 (double-negative (DN)) thymocytes that differentiate into CD4+CD8+ (double-positive (DP)) thymocytes that then terminally differentiate into CD4+ or CD8+ (single-positive (SP)) T cells1,2. Thymocytes at the DP stage of differentiation are the first cells to express endogenous TCR complexes and are the cells subjected to TCR-specific thymic selection. DP thymocytes are intrinsically short-lived cells whose continued survival requires signaling via the TCR by self ligands in the thymic cortex. TCR signaling rescues DP thymocytes from death by neglect and induces either positive or negative selection1,2. Positive selection is induced by low-affinity ligands and results in the differentiation of signaled DP thymocytes into conventional CD4+ SP (SP4) T cells or CD8+ SP (SP8) T cells with helper function or cytotoxic function, respectively, whereas negative selection is induced by high-affinity ligands that prevent DP thymoctyes that have received TCR signaling from continuing their differentiation into conventional SP T cells3,4. Thus, the thymus imposes central tolerance by generating mature SP4 and SP8 T cells that express TCRs without substantial autoreactive potential. The most definitive way of preventing autoreactive TCRs from appearing on mature SP T cells is the clonal deletion of DP thymocytes that bear autoreactive TCRs during negative selection in the thymus5. However, strong TCR signaling of DP thymocytes does not necessarily result in thymocyte death. Indeed, a few DP thymocytes are strongly signaled by agonist ligands to differentiate into specialized SP4 T cell subpopulations with regulatory or natural killer functions6,7, with such specialized differentiation referred to as agonist selection8. In a similar vein, developing DP thymocytes do not undergo clonal deletion when they receive strong signaling by agonist ligands in the thymic cortex9. Consequently, strong TCR stimulation of DP thymocytes during negative selection seems to be insufficient by itself to induce clonal deletion. However, it is not known what, if any, additional in vivo signals are needed during negative selection to induce thymocytes to undergo clonal deletion. A potentially useful insight may have been provided by longstanding in vitro studies showing that costimulation by CD28 is needed to induce the death of thymocytes that have received strong TCR signaling1013. Indeed in vitro costimulation by CD28 blocks upregulation of the antiapoptotic protein Bcl-2 by the TCR14, and transgenic overexpression of Bcl-2 rescues costimulated thymocytes from death induced by TCR signaling in vitro12,14. Although it has been observed only in vitro, a requirement for CD28 costimulation in thymocyte death that results from TCR signaling is potentially consistent with observations that clonal deletion is mediated by thymic dendritic cells and medullary thymic epithelial cells but not by cortical thymic epithelial cells that differ in their expression of the CD28-costimulatory ligands CD80 and CD86 (refs. 9,1517). However, those in vitro results have been directly contradicted by multiple in vivo studies that investigated a role for CD28 costimulation in clonal deletion. Those studies have shown
npg
Immunology Branch, National Cancer Institute, National Institutes of Health, Bethesda, Maryland, USA. 2Ludwig Center for Cancer Research, University of Lausanne, Epalinges, Switzerland. 3Science Applications International Corporation-Frederick, National Cancer Institute-Frederick Cancer Research and Development Center, Frederick, Maryland, USA. Correspondence should be addressed to A.S. (singera@nih.gov). Received 11 January; accepted 21 March; published online 29 April 2012; doi:10.1038/ni.2292
1Experimental
NUMBER 6
JUNE 2012
569
Articles
that autoreactive thymocytes are prevented from differentiating into either SP4 or SP8 T cells and that central tolerance is achieved regardless of the presence or absence of in vivo CD28 costimulation1822. Nonetheless, we undertook this study to determine if in vivo costimulation by CD28 during negative selection is important for either clonal deletion or central tolerance. Unlike the previously published in vivo studies1822, we distinguished between negative selection and clonal deletion by considering the possibility that DP thymocytes that have received strong signaling and that bear autoreactive TCRs might survive negative selection even though they do not differentiate into conventional SP4 or SP8 T cells. In fact, we found that in vivo costimulation by CD28 during negative selection was required for autoreactive thymocytes to undergo clonal deletion but that neither costimulation by CD28 nor clonal deletion was critical for central tolerance. We also found that tolerance was induced during negative selection even in the absence of CD28-mediated clonal deletion because strong TCR signaling diverted DP thymocytes from differentiating into conventional SP T cells bearing autoreactive TCRs and directed their differentiation into TCR+ DN thymocytes, a process we refer to as developmental diversion. Although developmentally diverted thymocytes expressed autoreactive TCRs, they were coreceptor negative and functionally anergic, which diminished their autoreactive potential. Moreover, when developmentally diverted TCR+ DN thymocytes left the thymus, they preferentially migrated to the intestine, where they received signals to express homodimers of the coreceptor CD8 (CD8) and were sequestered in the gut as CD8+ intraepithelial lymphocytes (IELs). Thus, we have identified CD28 as being critical for inducing autoreactive thymocytes to undergo clonal deletion during in vivo negative selection and have additionally identified CD8+ IELs as the fate of autoreactive thymocytes that avoid clonal deletion during negative selection. RESULTS Clonal deletion versus rescue of autoreactive thymocytes We began our assessment of intrathymic costimulation during negative selection by comparing thymocyte populations from wild-type mice and costimulation-deficient mice that lack either the costimulatory receptor CD28 (Cd28/ mice) or its two costimulatory ligands CD80 (B7-1) and CD86 (B7-2; called B7-deficient mice here). In our analysis, we specifically looked for a thymocyte subpopulation that was present in costimulation-deficient (Cd28/ or B7-deficient) mice but was absent from wild-type mice. Costimulation-deficient mice on two different genetic backgrounds (C57BL/6 and BALB/c) had only modestly altered thymocyte numbers and a slightly greater frequency of DN cells relative to those of wild-type mice (Fig. 1a, top). However, costimulation-deficient mice showed substantial enrichment for a specific subset of DN thymocytes that were TCR+ (Fig. 1a, bottom). Unlike DN thymocytes from wild-type mice, which included few TCR+ cells that were mostly natural killer T cells, as determined by staining with tetramers of the antigen-presenting molecule CD1d, DN thymocytes from costimulation-deficient mice included a high frequency of TCR+ cells that were not natural killer T cells (Fig. 1b,c). If impaired clonal deletion in costimulation-deficient mice were the basis for greater abundance of TCR+ DN thymocytes, then impairing clonal deletion in costimulation-sufficient mice should also result in more TCR+ DN thymocytes. To test this prediction, we attempted to impair clonal deletion in wild-type mice through the use of transgenes encoding the antiapoptotic proteins Mcl-1 and Bcl-2 (Fig. 2a). Mcl-1-transgenic mice and Bcl-2-transgenic mice had substantially more TCR+ DN thymocytes than did wild-type mice, and these thymocytes were CD5hi (Fig. 2a,b), consistent with their having received strong signaling in vivo. Thus, thymocyte expression of transgenes encoding prosurvival molecules in wild-type mice had the same effect as costimulation deficiencythat is, more TCR+ DN thymocyteswhich suggested that the TCR+ DN thymocyte subset contained cells that would otherwise have been clonally deleted. To directly test the possibility that the TCR+ DN subset was enriched for thymocytes bearing autoreactive TCRs, we examined mice expressing endogenous superantigens. BALB/c mice express the proviral proteins Mtv-6, Mtv-8 and Mtv-9, which specifically engage TCRs containing -chain variable region 3 (V3), V5, V11
npg
a
10
4 3 2 1 0
C57BL/6 WT 27610
6
BALB/c B7-DKO 11010

6
b
TCR DN thymocytes (%)
Cd28
WT 17810
6
Cd28
B7-DKO 5210
6
TCR DN thymocytes (%)
/ 6
/ 6
80 60 40 20 0
WT / Cd28 B7-DKO
C57BL/6 5 4 3 2 1 0
10
*** ***
TCR DN thymocytes (106)
18510
12310
80 60 40 20 0
WT / Cd28 B7-DKO
BALB/c 5 4 3 2 1 0
*** ***
**
**
10
CD4
10
3 2 1 0 10 10 10 10 10
CD8
Events (% of max)
100 80 60 40 20 0 100
60 40 20 0
Events (% of max)
B7-DKO (60) Cd28/ (31) WT (11)
100 80
B7-DKO (54) Cd28/ (27) WT (17)
100 80 60 40 20
TCR DN thymocytes B7-DKO C57BL/6 WT
10
10
10
10
10
10
10
10
10
TCR
0 100
101
102
10
2.2
2.3
2.9
2.2
4.0
4.5
103
104
CD1d-PBS57
Figure 1 CD28 costimulationdeficient mice have more TCR+ DN thymocytes. (a) Staining of CD4 and CD8 (top) in thymocytes from wild-type (WT), Cd28/ and B7-deficient (B7-DKO) mice on the C57BL/6 background (left) or BALB/c background (right), and surface TCR expression on gated DN thymocytes (bottom). Numbers at top of plots (above) indicate total number of thymocytes; numbers adjacent to outlined areas (above) indicate percent DN cells; numbers in parentheses (below) indicate percent DN cells that are TCR+ (bracketed lines). (b) Quantification (frequency and number) of TCR+ DN thymocytes in wild-type, Cd28/ and B7-deficient mice. *P < 0.01, **P < 0.001 and ***P < 0.0001, compared with wild-type (Students two-tailed t-test). (c) Staining of TCR+ DN thymocytes from wild-type and B7-deficient C57BL/6 mice with a tetramer of CD1d and the -galactosylceramide analog PBS57 (CD1d-PBS57). Data are representative of five independent experiments (a,b; mean and s.e.m.) or two independent experiments (c).
570
VOLUME 13
NUMBER 6
JUNE 2012
nature immunology
Articles a
10
4
Events (% of max)
C57BL/6 WT 6 135 10 Mcl-1-TG 147 106 Bcl-2-TG 78 106
100 80 60 40 20 0
103 102
60 40 20 0
80 60 40 20 0
TCR+ DN thymocytes (%)
Bcl-2-TG (60) Mcl-1-TG (42) WT (15)
100 80
Bcl-2-TG (80) Mcl-1-TG (53) WT (26)
b
100
WT
Mcl-1-TG 8 6 4 2 0
Bcl-2-TG
** **
* *
CD4
101 10
0
2.9 CD8
3.0
5.5
100 101 102 103 104
100
101
102
103
104
100
101
102
103
104
TCR
CD5
Figure 2 Effect of transgenes encoding Bcl-2 and Mcl-1 on the appearance of TCR+ DN thymocytes. (a) Staining of CD4 and CD8 in thymocytes from wild-type mice, Bcl-2-transgenic (Bcl-2-TG) mice and Mcl-1-transgenic (Mcl-1-TG) mice (left), and surface expression of TCR (middle) and CD5 and (right) on gated DN thymocytes (numbers as in Fig. 1a). (b) Quantification of TCR+ DN thymocytes in wild-type and transgenic mice. *P < 0.05 and **P < 0.001, compared with wild-type (Students two-tailed t-test). Data represent four independent experiments (mean and s.e.m.).
and V12, so BALB/c thymocytes expressing those TCRs undergo clonal deletion23. Indicative of TCR-specific clonal deletion in wildtype BALB/c mice, all TCR+ thymocyte subsets (SP4, SP8 and DN) had considerably fewer thymocytes bearing the superantigen-reactive V3+, V5+ and V11+ TCRs than did preselection DP thymocytes, whereas they had compensatorily more thymocytes bearing unreactive V8+ TCRs (Fig. 3a and Supplementary Fig. 1a). In contrast to wild-type BALB/c thymocytes, costimulation-deficient BALB/c thymocyte populations bearing Mtv-8- and Mtv-9-reactive TCRs (V5+ and V11+) had a uniquely greater frequency and number of TCR+ DN thymocytes but not SP4 or SP8 thymocytes (Fig. 3a). Thus, thymocytes bearing Mtv-8- and Mtv-9-reactive V5+ and V11+ TCRs were not deleted in costimulation-deficient BALB/c mice but instead appeared as TCR+ DN thymocytes. We additionally analyzed Mtv-6 reactive TCRs but were initially confused by a discrepancy in thymocytes from Cd28/ mice and those from B7-deficient mice (Supplementary Fig. 1). SP4 and SP8 thymocytes of Cd28/ BALB/c mice were depleted of Mtv-6 reactive V3+ TCRs (Supplementary Fig. 1a), as noted for other superantigenreactive TCRs (Fig. 3a), but SP4 and SP8 thymocytes in B7-deficient BALB/c mice were not depleted of these TCRs (Supplementary Fig. 1a), as reported before24. Our attempts to understand the basis for this
discrepancy led us to discover that the gene encoding Mtv-6 was absent from B7-deficient BALB/c mice (Supplementary Fig. 1b). This turned out to be due to the fact that this gene is on the same chromosomal segment as the genes encoding CD80 and CD86, which in B7-deficient mice was derived from embryonic stem cells of 129 origin that lacked the gene encoding Mtv-6. Consequently, B7-deficient mice did not have the gene encoding Mtv-6 in their genome, which explained the presence of thymocytes bearing the V3+ TCR in all B7-deficient BALB/c thymocyte subsets, including SP4 and SP8 cells. Having resolved the discrepancy noted above, we then sought to determine if transgenes encoding Mcl-1 and Bcl-2 would prevent the deletion of superantigen-reactive thymocytes in costimulationsufficient wild-type mice. Both the Mcl-1 transgene and the Bcl-2 transgene did in fact prevent clonal deletion of thymocytes in wildtype mice, as the frequency and number of superantigen-reactive TCRs were greater among TCR+ DN thymocytes in Mcl-1-transgenic mice and Bcl-2-transgenic mice than in nontransgenic wild-type mice, with the Bcl-2 transgene having a greater effect than the Mcl-1 transgene (Fig. 3b). On the basis of these results, we concluded that TCR signaling by high-affinity intrathymic ligands was sufficient to prevent autoreactive thymocytes from becoming SP thymocytes
WT Cd28/ B7-DKO 0.25
npg
a
12 0.35
BALB/c 1.5
V11 cells (10 )
V5 cells (10 )
V5 cells (%)
8 6 4 2 0 DP SP4
*** ***
V8 cells (%)
0.25 0.20 0.15 0.10 0.05 0 SP8 TCR+ DN
**
0.15 0.10 0.05 0 SP8 TCR DN

+
40
V8 cells (10 )
V11 cells (%)
10
0.30
***
******
0.20
*** **
60
20
*** ***
SP8 TCR DN
DP
SP4
+ SP8 TCR DN
1.0
0.5
DP
SP4
SP8 TCR+ DN WT
SP8 TCR+ DN
Figure 3 Expression of Mtv-reactive Mcl-1-TG CB6 + Bcl-2-TG V+ TCRs in pre- and post-selection b TCR DN thymocytes thymocyte subsets. (a) Expression of 8 20 8 50 0.6 1.5 2.0 0.6 * Mtv-9-reactive V5+ and Mtv-8- and * ** *** + TCRs and 40 * Mtv-9-reactive V11 6 *** 15 1.5 6 *** ** 0.4 1.0 0.4 control V8+ TCRs on various thymocyte 30 4 10 1.0 4 subsets from wild-type, Cd28/ and * 20 ** * * B7-deficient mice on the BALB/c 0.2 0.5 0.2 2 5 0.5 2 background. Horizontal dashed lines * 10 * indicate the frequency of preselection 0 0 0 0 0 0 0 0 DP thymocytes in wild-type mice +, Mtv-9-reactive V 5+ and Mtv-8- and expressing TCRs of each V specificity. (b) Quantification (frequency and number) of specific Mtv-6-reactive V 3 Mtv-9-reactive V11+ TCRs and control V8+ TCRs expressed by TCR+ DN thymocytes from the progeny of BALB/c mice crossed with C57BL/6 mice (CB6) that were wild-type, Bcl-2-transgenic or Mcl-1-transgenic. *P < 0.01, **P < 0.001 and ***P < 0.0001, compared with wild-type (Students two-tailed t-test). Data represent at least three independent experiments with at least five mice per group (mean and s.e.m.).
V11 cells (10 ) V3 cells (10 ) V5 cells (10 ) V11 cells (%) V8 cells (10 )
6 6
V3 cells (%)
V5 cells (%)
V8 cells (%)
NUMBER 6
JUNE 2012
571
Articles
status of their Cd8b1 locus because it remains methylated until it is permanently 127 10 113 10 80 10 160 10 103 100 demethylated when thymocytes first express 80 2 10 Thymus 60 CD8 to become DP cells25. In thymocyte Unmethylated 1 40 10 20 0 1.4 3.6 8.6 1.5 populations from Bcl-2-transgenic mice 10 0 0 1 2 3 4 10 10 10 10 10 20 that were enriched for TCR+ DN thymo40 CD8 Methylated 60 / cytes, the Cd8b1 promoter was methylated Cd28 (64) B7-DKO (56) 80 / 100 B7-DKO Cd28 in TCR DN precursor thymocytes but was Zap70/ (1) MHC-KO (4) demethylated in DP thymocytes and in their 100 TCR TCR+ post-selection SP4 and SP8 progeny (Fig. 4a 80 60 and Supplementary Fig. 2). The Cd8b1 Figure 4 TCR+ DN thymocytes are the 40 promoter was also demethylated in TCR+ progeny of DP thymocytes. (a) Methylation 20 status of the Cd8b1 promoter in Bcl-2DN thymocytes (Fig. 4a and Supplementary 0 transgenic thymocytes, determined by 100 101 102 103 104 Fig. 2), which showed that TCR+ DN thymoTCR bisulfite conversion and sequencing of cytes were the progeny of coreceptor-positive genomic DNA from sorted thymocyte (DP) thymocytes. subpopulations from three individual mice and presented as the frequency of methylated or Next we determined whether the genunmethylated Cd8b1 promoter sequences (Supplementary Fig. 2). *P < 0.0001 (Students eration of TCR+ DN thymocytes, like two-tailed t-test). (b) Staining of CD4 and CD8 (above) in thymocytes from mice of various genotypes (above plots) on the C57BL/6 background, and surface TCR expression on gated that of post-selection SP4 and SP8 thymoDN thymocytes (below; numbers as in Fig. 1a). Data are from two independent experiments. cytes, required TCR-mediated thymic selection signals. Because TCR signaling in but was not sufficient to induce clonal deletion, which additionally DP thymocytes is strictly dependent on the tyrosine kinase Zap70 required intrathymic costimulation. In addition, we concluded that (ref. 26), we assessed the effect of Zap70 deficiency on the appearance transgenes encoding the prosurvival molecules Mcl-1 and Bcl-2 pre- of TCR+ DN thymocytes (Fig. 4b). B7-deficient mice also deficient vented clonal deletion and that thymocytes rescued from clonal dele- in Zap70 were devoid of TCR+ DN thymocytes (Fig. 4b), which tion appeared in the thymus as TCR+ DN thymocytes. showed that the generation of TCR+ DN thymocytes was strictly dependent on signals transduced by Zap70. We further determined Derivation and selection of TCRab+ DN thymocytes whether the generation of TCR+ DN thymocytes required intra+ thymocytes are originally derived from TCR DN All TCR thymic expression of major histocompatibility complexes (MHCs) precursor thymocytes that received signals from the pre-TCR to because TCR-mediated thymic selection is MHC specific. Cd28/ differentiate into DP thymocytes, and it is in DP thymocytes that mice that were additionally deficient in MHC (Cd28/B2m/H2endogenously encoded TCR surface complexes are first expressed. Ab1/) were essentially devoid of TCR+ DN thymocytes (Fig. 4b). Consequently, because TCR+ DN thymocytes bear endogenously These results showed that expression of both MHC and Zap70 was encoded TCR surface complexes, it is likely that they are the prog- required for the in vivo generation of TCR+ DN thymocytes. eny of DP thymocytes. To determine if TCR+ DN thymocytes are We concluded that TCR+ DN thymocytes were the progeny of indeed derived from DP thymocytes, we examined the methylation DP thymocytes and that the differentiation of DP thymocytes into
10
B7-DKO
6
B7-DKO / Zap70
6
Cd28/
Cd28 MHC-KO
6
Sequences (%)
P SP 4 SP 8 D N
Figure 5 Effect of thymic selection on 250 B7-DKO SP4 SP8 TCR+ DN the expression of Mtv-reactive V+ TCRs. Bcl-2-TG BALB/c 200 (a) Frequency of cells expressing Mtv-9-reactive / B7-DKO WT Cd28 150 V5+ and Mtv-8- and Mtv-9-reactive V11+ 150 150 150 TCRs and control V8+ TCRs in various TCR+ 100 100 100 100 thymocyte subsets from wild-type, Cd28/ 50 and B7-deficient BALB/c mice, presented 50 50 50 0 as relative to that among preselection DP + DP INT TCR DN 0 0 0 thymocytes, set as 100%. (b) Surface DP INT PostDP INT PostDP INT PostSP4 expression of CD28 on intermediate and selection selection selection TCR+ DN TCR+ DN thymocyte subsets from B7200 200 200 2 deficient and Bcl-2-transgenic mice, 150 150 150 Antipresented as mean fluorescence intensity TCR 100 100 100 1 relative to that of preselection DP thymocytes, 50 50 50 set as 100%. (c) TCR-induced calcium 0 0 0 0 mobilization in B7-deficient thymocyte 0 50 100 150 200 250 DP INT PostDP INT PostDP INT Postpopulations depleted of CD8+ cells; selection selection selection Time (s) 300 300 300 + downward arrow indicates crosslinkage TCR DN SP8 of biotinylated monoclonal anti-TCR 200 200 200 250 (5 g/ml) by avidin. (d) Proliferation of 200 purified TCR+ DN and SP8 B7-deficient 100 100 100 150 100 thymocytes (2 104 cells per well) 0 0 0 50 stimulated with plate-bound anti-TCR (5 g) DP INT PostDP INT PostDP INT Post0 Med + IL-2 selection selection selection plus anti-CD28 (10 g) in the presence (+ IL-2) or absence (Med) of exogenous IL-2 (200 U/ml), assessed as incorporation of [3H]thymidine. Data are from five independent experiments (a; mean s.e.m.), represent three independent experiments (b,c; mean s.e.m. in b) or are representative of two independent experiments (d; mean and s.e.m. of triplicate cultures).
Events (% of max)
CD4
npg
V5 cells (%)
V5 cells (%)
V5 cells (%)
CD28 (%)
c
++
V11 cells (%)
V11+ cells (%)
V11 cells (%)
V8 cells (%)
V8 cells (%)
V8 cells (%)
572
VOLUME 13
NUMBER 6
JUNE 2012
Proliferation (103 c.p.m.)
[Ca ]i
nature immunology
Articles
TCR+ DN required TCR-mediated, MHC-specific thymic selection signals. As thymic selection signals normally induce DP thymocytes to differentiate into coreceptor-positive (SP4 or SP8) thymocytes, we refer to the altered differentiation of DP cells into TCR+ coreceptornegative (DN) thymocytes as developmental diversion. Developmental stage at which deletion and diversion occur During positive selection, DP thymocytes that have received TCR signaling upregulate their surface expression of CD4 and downregulate their surface expression of CD8 to become phenotypically CD69+CD4hiCD8lo intermediate thymocytes, and it is in intermediate thymocytes that differentiation into either the CD4 + lineage or the CD8+ lineage (CD4-CD8 lineage choice) occurs1. Consequently, we wondered if clonal deletion and developmental diversion also occur in intermediate thymocytes. To assess this possibility, we examined superantigen-reactive V+ TCRs in CD69+CD4hiCD8lo intermediate thymocytes from wild-type and costimulation-deficient BALB/c mice. Superantigen-reactive TCRs were present on intermediate thymocytes in both wild-type and costimulation-deficient BALB/c mice at frequencies that were essentially equal to those of preselection DP thymocytes (Fig. 5a). In fact, superantigen-reactive TCRs were present at a much greater frequency among intermediate thymocytes that had received TCR signaling than among SP4 or SP8 post-selection thymocytes (Fig. 5a), which indicated that DP thymocytes that had received strong signaling differentiated into intermediate thymocytes before undergoing clonal deletion or developmental diversion (Fig. 5a). That is, in wild-type BALB/c mice, superantigen-reactive TCRs were present in intermediate thymocytes, but all post-selection populations lacked such TCRs (Fig. 5a), which indicated that autoreactive thymocytes did not survive beyond the intermediate stage of differentiation. In costimulation-deficient BALB/c mice, superantigen-reactive TCRs were present in both intermediate and TCR+ DN thymocytes, but SP4 and SP8 post-selection thymocytes lacked such TCRs (Fig. 5a), which indicated that autoreactive intermediate thymocytes were developmentally diverted into TCR+ DN cells. These results indicated that signals for both clonal deletion and developmental diversion were provided during negative selection at the intermediate stage of differentiation, but they do not formally exclude the possibility that some DP thymocytes might become DN cells directly. For intermediate thymocytes to receive costimulation-dependent deletional signals, intermediate thymocytes must be in contact with B7-expressing cells. To determine if intermediate thymocytes were in fact in contact with B7-expressing cells in the thymus, we made use of the fact that CD28-B7 interactions specifically downregulate CD28 surface expression27. In fact, surface expression of CD28 was upregulated during the differentiation of DP cells into intermediate and TCR+ DN thymocytes in B7-deficient mice, whereas it was downregulated on cells from B7-sufficent mice (Fig. 5b and Supplementary Fig. 3), which showed that both intermediate thymocytes and TCR+ DN thymocytes in B7-sufficient mice (which survived because of transgenic overexpression of Bcl-2) were in contact with B7-expressing cells in the thymus and received strong costimulation in vivo. We concluded that CD4+CD8lo intermediate thymocytes were located in an area of the thymus where they were able to contact B7 costimulatory ligands. Furthermore, the intermediate stage of differentiation was a point in thymocyte development during which clonal deletion and developmental diversion both occurred. Because the intermediate stage of differentiation is also the point in thymocyte development at which CD4-CD8 lineage choice occurs, we suggest that it was in intermediate thymocytes that different TCR signals were translated into different lineage fates, depending on the intensity and duration of TCR signaling as well as the presence or absence of costimulation (Supplementary Fig. 4). Characterization of developmentally diverted thymocytes Next we analyzed the phenotype of developmentally diverted TCR+ DN thymocytes. This showed that during the differentiation of DP cells into developmentally diverted TCR+ DN thymocytes, a variety of molecules were upregulated (TCR, PD-1, CD5, CD69, CD122, Bcl-2 and 47), whereas other molecules were either unchanged or
a npg
10 TCR DN thymocytes
+ 4
RAG-KO host B7-DKO donor 98 10 5 weeks 10

4 + TCR IEL 2
103 10
2
103
2
TCR+ LN 32
b
B7-DKO
TCR DN thymocytes 99
In vitro Anti-TCR + IL-15 9 4d 82
Med 0
101 100 104 100 101 102 103 104 CD4-CD8 96
101 100 104 103 5 weeks CD8 102 101 100
94 100 101 102 103 104 74
9 104 103 91 Bcl-2-TG CD5 102 101 10

0
0 0
104 98 103 4d CD8 102 101
CD5
TCR SP8 LN T cells
103 102 101 100
22 100 101 102 103 104 CD8
100 101 102 103 104 CD8
B2m/ B2m/ B2m/ WT Figure 6 Developmentally diverted TCR+ DN thymocytes migrate to 104 the intestine, where they become CD8+ IELs. (a) Homing and fate of 9 8 0 43 103 developmentally diverted TCR+ DN thymocytes, assessed in the small 2 10 intestine (IEL) and lymph nodes (LN) of recombination-activating gene 101 2deficient (RAG-KO) recipient mice (right) 5 weeks after adoptive transfer 32 75 5 82 100 of 0.8 106 sorted CD5+ DN thymocytes (top) or CD5+CD8+ T cells 2 3 4 100 101 10 10 10 (bottom) from the lymph nodes of B7-deficient donor mice (left). (b) Flow CD8 cytometry of developmentally diverted TCR+ DN thymocytes (right) generated + DN thymocytes (2 104 cells per well; left) for 4 d in medium alone (Med) or with plateby culture of sorted B7-deficient or Bcl-2 transgenic CD5 bound monoclonal anti-TCR and soluble IL-15 (100 ng/ml; Anti-TCR + IL-15). (c) Expression of CD8 versus CD8 on TCR+ cells from the small intestines of mice of various strains (above plots). Numbers adjacent to or in outlined areas indicate percent cells in each gate throughout. Data are representative of two independent experiments. CD8
100 101 102 103 104 CD4 CD8 NK1.1 GL3
78 100 100 101 102 103 104 CD8
B7-DKO
Bcl-2-TG
NUMBER 6
JUNE 2012
573
Articles a
V5 cells (%)
+
10 8 6 4 2 0 8 6 4 2 0 4 3 2 1 0 8 6 4 2 0
c
V5+ cells (%)
6 4 2 0 8 6 4 2 0 2 1 0 8 6 4 2 0
**
**
V11 cells (%)
V11+ cells (%)
**
V12 cells (%)
V 4 cells (%)
V4+ cells (%)
V12+ cells (%)
Figure 7 Developmentally diverted TCR+ DN thymocytes become CD8+ IELs. (a) Expression of Mtv-reactive V5+, V11+ and V12+ TCRs and control V4+ TCRs in various thymocyte subsets and IELs from B7deficient BALB/c mice. (b) Total TCR+ CD8+ or CD8+ IELs in wildtype and B7-deficient BALB/c mice. *P < 0.002 (Students two-tailed t-test). (c) Expression of Mtv-reactive V5+, V11+ and V12+ TCRs and control V4+ TCRs in various T cell subsets from wild-type BALB/c mice. *P < 0.05 and **P < 0.01 (Students two-tailed t-test). (d) Proliferation of sorted TCR+CD8+ IELs and SP4 lymph node T cells (LNT; 2.5 104 cells per well) stimulated with soluble monoclonal anti-CD3 (5 g) and irradiated syngenic antigen-presenting cells in the presence of medium alone or recombinant IL-2 (200 U/ml) or IL-15 (100 ng/ml), assessed as incorporation of [3H]thymidine. Data represent three independent experiments (mean and s.e.m.).
Thymocytes
IELs
b
40
WT (BALB/c) B7-DKO (BALB/c)
d
Stimulation (103 c.p.m.) 105
+ 6 TCR IELs (10 )
30 20 10 0
CD8 CD8
+ +
4 SP 8 C D 8 + C D 8
SP
SP4 SP8 DN CD8+ + TCR TCR+
+ LNT TCR IELs
Bcl-2-TG
104
+ TCR T cells
CD8+ IELs SP4 LNT
103 Med + IL-2 + IL-15
downregulated (CD25, CD44, CD124, CD127, CD132 and Bcl-xL; Supplementary Fig. 5a,b). The expression pattern of these surface molecules provided some potentially useful insights into the generation and fate of TCR+ DN thymocytes. First, the upregulation of TCR, the activation markers CD5 and CD69 suggested that TCR+ DN thymocytes had received TCR signaling, and their high expression of the attenuator of TCR signaling PD-1 indicated that those in vivo TCR signals had been strong and persistent, as would be expected of autoreactive TCRs (Supplementary Fig. 5a). Because surface PD-1 molecules dampen TCR signal transduction28, we then assessed the in vitro reactivity of TCR+ DN thymocytes. We determined that developmentally diverted TCR+ DN thymocytes were anergic cells on the basis of the following three
findings: their TCRs did not induce calcium mobilization (Fig. 5c and Supplementary Fig. 6a), they were unable to produce their own interleukin 2 (IL-2; Supplementary Fig. 6b) and they required the addition of exogenous IL-2 to proliferate in response to stimulation with antibody to TCR (anti-TCR) and anti-CD28 (Fig. 5d). We then considered whether PD-1 expression was required for developmental diversion. However, germline deletion of the gene encoding PD-1 did not affect the generation of developmentally diverted TCR+ DN thymocytes in B7-deficient mice (Supplementary Fig. 6c). In addition, little or no surface expression of CD127 (IL-7 receptor -chain) indicated that developmentally diverted TCR+ DN thymocytes were probably not dependent on IL-7 for their in vivo survival (Supplementary Fig. 5a), unlike mature SP thymocytes and T cells. Indeed, despite their almost complete lack of CD127 expression, developmentally diverted TCR+ DN thymocytes had high expression of Bcl-2 and were not apoptotic in vivo (as shown by their lack of staining with annexin V or active caspase-3; Supplementary Fig. 5b), possibly because they expressed CD122 (IL-2 receptor -chain) and CD132 (common -chain), which conferred the potential to respond in vivo to IL-2 and IL-15 (Supplementary Fig. 5a). Moreover, developmentally diverted TCR+ DN thymocytes expressed the integrin 47 (Supplementary Fig. 5a), which indicated that they could potentially leave the thymus and migrate to the intestine. Developmentally diverted thymocytes become CD8aa+ IELs To determine their in vivo developmental potential, we adoptively transferred TCR+ DN thymocytes from B7-deficient donor mice
TCR DN thymocytes 16 97
+
npg
a
100
Bcl-2-TG Runx3
80 60 40 20 0 0 10 10
1
+/YFP
600
Bcl-2-TG +/YFP mice Runx3
b
Bcl-2-TG +/YFP Runx3
10
4 3 2 1 0
Anti-TCR + IL-15 Med 312
Anti-TCR + IL-15
Med
Events (% of max)
Runx3-YFP (MFI)
Events (% of max)
DP TCR+ DN Thymocytes SP8 TCR CD8 IELs

+ +
400 75 50 25
4d
39
100 509 80 21 60 40 20 0 100 101 102 103 104 10 10 10
4 3 2 1 0 0 1 2
D P D N S C P8 D 8 +
Runx3-YFP
CD5
10 10
CD8
10
10
10
Bcl-2-TG YFP/YFP Runx3

+
10 10
94
4d
10 10
0
3 4
TC R
10 10 10 10 10
10 10 10 10 10
Thymus IELs CD8 Figure 8 Runx3 is required for the (TCR) CD1d-PBS differentiation of TCR+ DN thymocytes into GL3 + IELs. developmentally diverted CD8 (a) Expression of the Runx3-YFP reporter in thymocyte subsets and IELs from Bcl-2-transgenic mice (left) and mean fluorescent intensity (MFI) of Runx3-YFP (right). (b) Expression of Runx3-YFP (middle) and CD8 and CD8 (right) on developmentally diverted TCR+ DN thymocytes generated by culture of sorted Runx3-sufficient (Runx3+/YFP) or Runx3deficient (Runx3YFP/YFP) Bcl-2-transgenic CD5+ DN thymocytes (2.5 104 cells per well; left) for 4 d in medium alone or with plate-bound monoclonal anti-TCR and soluble IL-15 (100 ng/ml). Numbers adjacent to or in outlined areas indicate percent cells in that gate; numbers in plots (middle) indicate MFI of Runx3-YFP. (c) Total TCR+ DN thymocytes and TCR+CD8+ IELs in Runx3-sufficient and Runx3-deficient Bcl-2-transgenic mice. *P < 0.01 (Students two-tailed t-test). Data are representative of three independent experiments (a,c; mean and s.e.m.) or two experiments (b).
CD4
Runx3-YFP
c
15
CD8 Bcl-2-TG Runx3

+/YFP
Bcl-2-TG Runx3YFP/YFP
Cells (106)
10
TCR+ DN Thymocytes
TCR+CD8+ IELs
574
VOLUME 13
NUMBER 6
JUNE 2012
nature immunology
Articles
into host mice deficient in recombination-activating gene 2 (Fig. 6a). At 5 weeks after transfer, most developmentally diverted TCR+ DN thymocytes had homed to the small intestine and differentiated into TCR+CD8+ IELs, whereas SP8 T cells remained mainly TCR+CD8+ T cells regardless of their homing destination (Fig. 6a). Because IL-15 is present in the gut and because TCR+ DN thymocytes are potentially responsive to IL-15 because of their expression of CD122 and CD132, we wondered if IL-15 might contribute to the differentiation of TCR+ DN thymocytes into TCR+CD8+ cells. To assess this, we stimulated developmentally diverted TCR+ DN thymocytes in vitro with anti-TCR and IL-15 (Fig. 6b). Indeed, by day 4 of culture, most TCR+ DN thymocytes that received such signaling had differentiated into TCR+CD8+ cells (Fig. 6b). The results presented above showed that developmentally diverted TCR+ DN thymocytes had the potential to become TCR+CD8+ IELs, and they resembled results obtained with wild-type TCR+ DN thymocytes29, but we did not know if developmentally diverted TCR+ DN thymocytes actually differentiated into TCR+CD8+ IELs in vivo. To assess this, we took advantage of the finding that TCR+CD8+ IELs were nearly completely absent from C57BL/6 mice deficient in 2-microglobulin (B2m/ mice; Fig. 6c). We crossed B2m/ mice with B7-deficient or Bcl-2-transgenic mice, each of which had a substantial frequency of developmentally diverted thymocytes (Figs. 1 and 2). TCR+CD8 IELs were present in the B7-deficient or Bcl-2-transgenic B2m/ progeny of each cross (Fig. 6c), which demonstrated that developmentally diverted T cells did in fact differentiate in vivo into TCR+CD8+ IELs. To confirm that finding, we sought to determine if TCR+CD8+ IELs displayed the same TCR V repertoire as that displayed by developmentally diverted TCR+ DN thymocytes. Indeed, in B7-deficient BALB/c mice, TCRs expressing superantigen-reactive V molecules were over-represented at essentially identical frequencies in both TCR+CD8+ IELs and TCR+ DN thymocytes (Fig. 7a). We concluded that developmentally diverted TCR+ DN thymocytes differentiated in vivo into TCR+CD8+ IELs. Developmental diversion and the origin of CD8aa+ IELs Having shown that developmentally diverted TCR+ DN thymocytes differentiated in vivo into TCR+CD8+ IELs, we wondered if all TCR+CD8+ IELs might be derived from cells that had undergone developmental diversion during negative selection in the thymus. If so, mice in which thymic clonal deletion was absent would have more TCR+CD8+ IELs than would wild-type mice. In fact, costimulationdeficient mice had many more TCR+CD8+ IELs than wild-type mice had (about ten times more; Fig. 7b), which suggested that developmental diversion was the main origin of TCR+CD8+ IELs. We then wondered if the TCR+CD8+ IELs present in normal wild-type mice were also the progeny of thymocytes that had survived negative selection and had undergone developmental diversion. In fact, we found that even in wild-type mice, TCR+CD8+ IELs specifically overexpressed superantigen-reactive TCRs, but other T cell populations did not (Fig. 7c). That finding suggested that in normal wild-type mice, substantial numbers of autoreactive T cells avoided clonal deletion and underwent developmental diversion, a conclusion that was also consistent with our observation that the frequency of superantigen-reactive TCRs in wild-type mice was always greater in TCR+ DN thymocytes than in SP4 or SP8 thymocytes (Figs. 3a and 5a and Supplementary Fig. 1a). The over-representation of autoreactive TCR specificities in CD8+ IELs in normal and experimental mice, none of which were autoimmune, made us question the TCR responsiveness of CD8+ IELs.
In fact, CD8+ IELs resembled TCR+ DN thymocytes in being unresponsive to stimulation with anti-TCR and antigen-presenting cells in vitro without exogenous cytokines (Fig. 7d). We concluded that TCR+CD8+ IELs were the progeny of cells that survived negative selection and underwent developmental diversion in the thymus. Role of the transcription factor Runx3 Finally, we sought to gain molecular insight into the differentiation of developmentally diverted TCR+ DN thymocytes into CD8+ IELs. Because such differentiation requires reactivation of Cd8a expression, which is a known function of the transcription factor Runx3 (ref. 30), we considered that Runx3 might be required for this differentiation. To analyze this possibility, we used Bcl-2-transgenic mice also heterozygous for an endogenous Runx3 allele re-engineered to encode yellow fluorescent protein (YFP) instead of Runx3 (Runx3+/YFP mice)31. Developmentally diverted TCR+ DN thymocytes in these Bcl-2-transgenic Runx3+/YFP mice were YFPlo, whereas TCR+CD8+ IELs were YFPhi (Fig. 8a), which indicated that Runx3 expression was upregulated at some point during the differentiation of TCR+ DN thymocytes into TCR+CD8+ IELs. To determine if that point was related to the reactivation of Cd8a, we stimulated TCR+ DN thymocytes in vitro with anti-TCR and IL-15 (Fig. 8b). After that in vitro stimulation, TCR+ DN thymocytes expressed Runx3 and differentiated into CD8+ cells (Fig. 8b). To determine if their differentiation into CD8+ cells required Runx3, we used Bcl-2-transgenic Runx3YFP/YFP mice, which were Runx3 deficient because they had two Runx3 alleles re-engineered to encode YFP instead of Runx3 (Fig. 8b). In vitro stimulation of TCR+ DN thymocytes that were Runx3 deficient did not induce their differentiation into CD8+ cells, although it did induce YFP expression (Fig. 8b). Thus, Runx3 was required for the reactivation of Cd8a expression by developmentally diverted TCR+ DN thymocytes and for their differentiation into CD8+ cells in vitro. Applying those in vitro observations to the developmental fate of TCR+ DN thymocytes in vivo, we wondered if the differentiation of developmentally diverted TCR+ DN thymocytes into TCR+CD8+ IELs would be impaired in Runx3YFP/YFP mice because they were Runx3 deficient. Indeed, Bcl-2-transgenic Runx3YFP/YFP (Runx3-deficient) mice had significantly fewer TCR+CD8+ IELs than did Bcl-2transgenic Runx3+/YFP (Runx3-sufficient) mice, even though they had equal numbers of developmentally diverted TCR+ DN thymocytes (Fig. 8c). We concluded that TCRs and IL-15 stimulated developmentally diverted TCR+ DN thymocytes to express Runx3, which promoted the reactivation of Cd8a and the differentiation of developmentally diverted TCR+ DN thymocytes into TCR+CD8+ IELs. Thus, we were able to integrate lineage choice during positive selection and lineage fate during negative selection into a unified picture of thymic development in which the developmental fate of TCR-signaled DP thymocytes is determined by TCR and costimulatory signals (Supplementary Fig. 7). DISCUSSION Here we have identified costimulatory signals from CD28 as being necessary for thymocytes to undergo clonal deletion during negative selection in vivo and have demonstrated that neither costimulation by CD28 nor clonal deletion was required for self-tolerance. Regardless of the presence or absence of CD28-mediated clonal deletion, strong TCR signaling prevented DP thymocytes bearing autoreactive TCRs from differentiating into conventional SP4 or SP8 T cells, so the absence of autoreactive TCRs on SP T cells was indicative of in vivo
575
npg
Articles
negative selection but was not necessarily indicative of in vivo clonal deletion. Indeed, DP thymocytes that had received strong signaling and bore autoreactive TCRs that survived negative selection did not differentiate into conventional SP T cells but instead underwent developmental diversion and differentiated into TCR+ DN thymocytes. Developmentally diverted TCR+ DN thymocytes, after leaving the thymus, migrated mainly to the intestine and received signaling via IL-15 to express Runx3 and to further differentiate into Runx3+CD8+ IELs. Clonal deletion and developmental diversion occurred during negative selection at the intermediate CD4hiCD8lo thymocyte stage of differentiation, the same point in thymocyte development at which lineage choice occurs during positive selection1. Thus, our study has distinguished in vivo negative selection from in vivo clonal deletion, has identified costimulation by CD28 as the in vivo signal required for clonal deletion during negative selection, and has identified TCR+CD8+ IELs as the ultimate fate of autoreactive cells that survive negative selection in the thymus. During thymic selection, DP thymocytes that have received weak TCR signaling undergo positive selection into SP4 or SP8 T cells, whereas most DP thymocytes that have received strong TCR signaling undergo negative selection, which prevents them from differentiating into SP4 or SP8 T cells. Notably, as we have shown here, the absence of autoreactive TCR specificities among in vivo SP T cells was indicative of negative selection but was not necessarily indicative of clonal deletion. Consequently, our study has provided a new conceptual model of thymic selection in which positive selection is the differentiation of signaled DP thymocytes into SP T cells and negative selection is the prevention of their differentiation into SP T cells; developmental diversion refers to their differentiation into mature T cells that are neither SP4 nor SP8 but are DN; and clonal deletion refers to their death before maturation. We used specific V+ TCRs to monitor the developmental fate of DP thymocytes that had received strong signaling by endogenously encoded proviral antigens of the Mtv family. DP thymocytes are the first cells in the thymus to express endogenously encoded TCR complexes and are the cells that are subjected to thymic selection. As observed before23, DP thymocytes bearing superantigenreactive TCRs underwent negative selection, as they did not differentiate into SP4 or SP8 T cells. However, our study has documented that clonal deletion of superantigen-reactive thymocytes during negative selection additionally required costimulation by CD28. In the absence of CD28 costimulation, superantigen-reactive thymocytes survived negative selection and underwent developmental diversion to TCR+ DN thymocytes. Similarly, superantigen-reactive thymocytes that were prevented from undergoing clonal deletion by the expression of a transgene encoding Bcl-2 or Mcl-1 also survived negative selection and underwent developmental diversion into TCR+ DN thymocytes. Thus, costimulation by CD28 during negative selection was required for in vivo clonal deletion, which was prevented by transgenic expression of either Bcl-2 or Mcl-1. The developmental diversion of surviving autoreactive thymocytes to TCR+ DN thymocytes precluded autoreactivity by removing the contribution of the coreceptors CD4 and CD8 to TCR signaling, which lessened their in vivo autoreactive potential, as did their expression of negative costimulatory receptors such as PD-1 (ref. 28). In addition, after leaving the thymus, developmentally diverted TCR+ DN thymocytes migrated to the intestine, where they were sequestered away from the rest of the body and were stimulated by IL-15 to express Runx3, which reactivated Cd8a expression and promoted their further differentiation into CD8+ IELs that were hyporesponsive to TCR stimulation. Their CD8 phenotype may have further diminished their
576
autoreactive potential, as CD8 surface complexes have been suggested to sequester the signaling kinase Lck away from the TCR32. Indeed, costimulation-deficient mice were free of autoimmunity despite our finding that they had many CD8+ IELs bearing autoreactive TCR specificities and despite their lack of functionally suppressive Foxp3+ regulatory T cells33. Developmental diversion occurs in TCR-signaled CD4hiCD8lo intermediate thymocytes that transcribe Cd4 but lack transcription of Cd8, with the result that differentiation into TCR+ DN thymocytes during negative selection requires only the termination of Cd4 expression. In fact, in intermediate thymocytes that transcribe Cd4 but lack transcription of Cd8, such termination occurs during MHC class Ispecific positive selection into SP8 T cells, but in that case it is mediated by the transcription factor Runx3 (refs. 30,34), which additionally induces reactivation of Cd8 to result in coreceptor reversal35. In contrast, developmental diversion into TCR+ DN thymocytes did not involve or require Runx3, which suggested that strong TCR stimulation of intermediate thymocytes during negative selection terminates Cd4 transcription independently of Runx3 and without reactivation of Cd8. Notably, Runx3-independent termination of Cd4 transcription by strong TCR signaling in intermediate thymocytes that transcribe Cd4 but lack transcription of Cd8 would also explain why negatively selected thymocytes become DN cells instead of SP4 T cells. By identifying developmental diversion and clonal deletion as alternative outcomes of negative selection, our study has resolved many longstanding experimental contradictions. Published in vitro experiments have demonstrated that thymocyte death as a result of TCR signaling requires CD28 costimulation1014 and can be prevented by transgenic Bcl-2 expression12,14, whereas published in vivo experiments have arrived at opposite conclusions1822. Our study here has shown that those contradictory observations were due mainly to the presumption that autoreactive thymocytes had been clonally deleted in vivo if they failed to differentiate into either SP4 or SP8 T cells. In fact, here we have documented that the absence of autoreactive TCR specificities on SP T cells in vivo indicated that autoreactive thymocytes had undergone negative selection but did not indicate that they had undergone clonal deletion. Our study has also resolved the discrepant observations that genetic deletion of the proapoptotic protein Bim interferes with clonal deletion36, whereas transgenic overexpression of the antiapoptotic protein Bcl-2 does not20. In fact, many features of Bim-deficient mice3739 resemble those of the Bcl-2-transgenic and Mcl-1-transgenic mice we studied here; thus, the results of our study would explain the high frequency of TCR+ DN thymocytes in mice deficient in Bim or the proapoptotic protein Puma37,40,41 as resulting from their being autoreactive thymocytes that survived negative selection and underwent developmental diversion. Our study has also explained why SP4 and SP8 T cells bearing Mtv-6 reactive V3+ TCRs have been noted before in B7-deficient BALB/c mice but not Cd28/ BALB/c mice24. The discrepancy between these mice does not indicate an as-yet-unknown B7-specific receptor that signals clonal deletion in Cd28/ BALB/c mice24 but is instead due to the fact that Mtv-6 is not encoded by the genome of B7-deficient BALB/c mice. Although it has resolved many longstanding experimental contradictions, our study does contradict the experimental finding obtained with perinatal mice showing that blockade of costimulation by in vivo injection of anti-B7 rescues superantigen-reactive TCRs that are then expressed on SP4 T cells42. In our study here, genetic deletion of B7 ligands in B7-deficient mice also rescued superantigen-reactive TCRs; however, the rescued TCRs were not expressed on SP4 T cells but instead were expressed only on developmentally diverted TCR+
npg
Articles
DN thymocytes and CD8+ IELs. We think the explanation for this disparity is that injection of anti-B7 depletes the thymus of superantigenbearing B7+ cells and thus eliminates intrathymic expression of the negative selecting (superantigen) ligand. Does developmental diversion occur in normal wild-type (costimulation-sufficient) mice? Costimulation-sufficient TCRtransgenic mice have a substantial number of TCR+ DN thymocytes37,41,43 that can become CD8+ IELs29,32,4346. However, transgenically expressed TCRs differ from endogenous TCRs in that they are expressed by early DN thymocytes before the DP stage of differentiation, so it is difficult to know if TCR+ DN thymocytes in TCR-transgenic mice are post-selection cells that result from developmental diversion or are preselection DN thymocytes that were developmentally arrested because their transgenically expressed TCR engaged high-affinity ligands in the thymic cortex47,48. Notably, we found that in normal, nontransgenic mice, many or all TCR+CD8+ IELs were the progeny of thymocytes that survived negative selection and underwent developmental diversion in the thymus. Indeed, normal wild-type mice had a substantial number of TCR+CD8+ IELs (~5 106) in which TCRs with autoreactive specificities were overexpressed, which indicated that a substantial number of autoreactive T cells avoided clonal deletion and underwent developmental diversion in normal mice. In conclusion, by distinguishing in vivo negative selection from in vivo clonal deletion, we have identified costimulation by CD28 as being critical for clonal deletion and have identified developmental diversion as an alternative outcome of negative selection. Moreover, we have identified TCR+CD8+ IELs as the ultimate fate of autoreactive T cells that survived negative selection and underwent developmental diversion in the thymus. Also, by showing that the fate of thymocytes undergoing negative selection was determined at the same point in differentiation as that of thymocytes undergoing positive selection, our study has integrated negative selection and positive selection into a unified picture of thymic selection. METhODS Methods and any associated references are available in the online version of the paper.
Note: Supplementary information is available in the online version of the paper. AcknowLedGmenTS We thank N. Taylor, J. DiSanto and R. Hodes for critical reading of the manuscript; A. Sharpe (Harvard University) for B7-deficient mice, T. Honjo (Kyoto University) for PD-1-deficient mice; D. Littman (New York University) for Runx3-YFP reporter mice; and S. Sharrow, A. Adams and L. Granger for flow cytometry. Supported by the Intramural Research Program of the National Cancer Institute Center for Cancer Research of the US National Institutes of Health. AUTHoR conTRIBUTIonS L.A.P. designed the study, did experiments, analyzed data and contributed to the writing of the manuscript; G.S.A., X.T., S.J. and F.V.L. did experiments and analyzed data; J.-H.P. and L.F. generated transgenic mice, and A.S. designed the study, analyzed data and wrote the manuscript. comPeTInG FInAncIAL InTeReSTS The authors declare no competing financial interests.
1. Singer, A., Adoro, S. & Park, J.H. Lineage fate and intense debate: myths, models and mechanisms of CD4- versus CD8-lineage choice. Nat. Rev. Immunol. 8, 788801 (2008). 2. Starr, T.K., Jameson, S.C. & Hogquist, K.A. Positive and negative selection of T cells. Annu. Rev. Immunol. 21, 139176 (2003). 3. Gascoigne, N.R. & Palmer, E. Signaling in thymic selection. Curr. Opin. Immunol. 23, 207212 (2011). 4. McCaughtry, T.M. & Hogquist, K.A. Central tolerance: what have we learned from mice? Semin. Immunopathol. 30, 399409 (2008). 5. Kappler, J.W., Roehm, N. & Marrack, P. T cell tolerance by clonal elimination in the thymus. Cell 49, 273280 (1987). 6. Bendelac, A., Savage, P.B. & Teyton, L. The biology of NKT cells. Annu. Rev. Immunol. 25, 297336 (2007). 7. Jordan, M.S. et al. Thymic selection of CD4+CD25+ regulatory T cells induced by an agonist self-peptide. Nat. Immunol. 2, 301306 (2001). 8. Baldwin, T.A., Hogquist, K.A. & Jameson, S.C. The fourth way? Harnessing aggressive tendencies in the thymus. J. Immunol. 173, 65156520 (2004). 9. Laufer, T.M., DeKoning, J., Markowitz, J.S., Lo, D. & Glimcher, L.H. Unopposed positive selection and autoreactivity in mice expressing class II MHC only on thymic cortex. Nature 383, 8185 (1996). 10. Kishimoto, H. & Sprent, J. Negative selection in the thymus includes semimature T cells. J. Exp. Med. 185, 263271 (1997). 11. Kishimoto, H. & Sprent, J. Several different cell surface molecules control negative selection of medullary thymocytes. J. Exp. Med. 190, 6573 (1999). 12. Punt, J.A., Havran, W., Abe, R., Sarin, A. & Singer, A. T cell receptor (TCR)-induced death of immature CD4+CD8+ thymocytes by two distinct mechanisms differing in their requirement for CD28 costimulation: implications for negative selection in the thymus. J. Exp. Med. 186, 19111922 (1997). 13. Punt, J.A., Osborne, B.A., Takahama, Y., Sharrow, S.O. & Singer, A. Negative selection of CD4+CD8+ thymocytes by T cell receptor-induced apoptosis requires a costimulatory signal that can be provided by CD28. J. Exp. Med. 179, 709713 (1994). 14. McKean, D.J. et al. Maturation versus death of developing double-positive thymocytes reflects competing effects on Bcl-2 expression and can be regulated by the intensity of CD28 costimulation. J. Immunol. 166, 34683475 (2001). 15. Ramsdell, F., Lantz, T. & Fowlkes, B.J. A nondeletional mechanism of thymic self tolerance. Science 246, 10381041 (1989). 16. Roberts, J.L., Sharrow, S.O. & Singer, A. Clonal deletion and clonal anergy in the thymus induced by cellular elements with different radiation sensitivities. J. Exp. Med. 171, 935940 (1990). 17. Takahama, Y. Medullary interplay for central tolerance. Blood 118, 23802381 (2011). 18. Dautigny, N., Le Campion, A. & Lucas, B. Timing and casting for actors of thymic negative selection. J. Immunol. 162, 12941302 (1999). 19. Jones, L.A., Izon, D.J., Nieland, J.D., Linsley, P.S. & Kruisbeek, A.M. CD28B7 interactions are not required for intrathymic clonal deletion. Int. Immunol. 5, 503512 (1993). 20. Sentman, C.L., Shutter, J.R., Hockenbery, D., Kanagawa, O. & Korsmeyer, S.J. Bcl-2 inhibits multiple forms of apoptosis but not negative selection in thymocytes. Cell 67, 879888 (1991). 21. Tan, R., Teh, S.J., Ledbetter, J.A., Linsley, P.S. & Teh, H.S. B7 costimulates proliferation of CD48+ T lymphocytes but is not required for the deletion of immature CD4+8+ thymocytes. J. Immunol. 149, 32173224 (1992). 22. Walunas, T.L., Sperling, A.I., Khattri, R., Thompson, C.B. & Bluestone, J.A. CD28 expression is not essential for positive and negative selection of thymocytes or peripheral T cell tolerance. J. Immunol. 156, 10061013 (1996). 23. Simpson, E. T cell repertoire selection by mouse mammary tumour viruses. Eur. J. Immunogenet. 20, 137149 (1993). 24. Buhlmann, J.E., Elkin, S.K. & Sharpe, A.H. A role for the B71/B72:CD28/CTLA4 pathway during negative selection. J. Immunol. 170, 54215428 (2003). 25. Carbone, A.M., Marrack, P. & Kappler, J.W. Demethylated CD8 gene in CD4+ T cells suggests that CD4+ cells develop from CD8+ precursors. Science 242, 11741176 (1988). 26. Negishi, I. et al. Essential role for ZAP-70 in both positive and negative selection of thymocytes. Nature 376, 435438 (1995). 27. Yu, X., Fournier, S., Allison, J.P., Sharpe, A.H. & Hodes, R.J. The role of B7 costimulation in CD4/CD8 T cell homeostasis. J. Immunol. 164, 35433553 (2000). 28. Riley, J.L. PD-1 signaling in primary T cells. Immunol. Rev. 229, 114125 (2009). 29. Gangadharan, D. et al. Identification of pre- and postselection TCR+ intraepithelial lymphocyte precursors in the thymus. Immunity 25, 631641 (2006). 30. Sato, T. et al. Dual functions of Runx proteins for reactivating CD8 and silencing CD4 at the commitment process into CD8 thymocytes. Immunity 22, 317328 (2005). 31. Egawa, T. & Littman, D.R. ThPOK acts late in specification of the helper T cell lineage and suppresses Runx-mediated commitment to the cytotoxic T cell lineage. Nat. Immunol. 9, 11311139 (2008). 32. Cheroutre, H., Lambolez, F. & Mucida, D. The light and dark sides of intestinal intraepithelial lymphocytes. Nat. Rev. Immunol. 11, 445456 (2011). 33. Tai, X., Cowan, M., Feigenbaum, L. & Singer, A. CD28 costimulation of developing thymocytes induces Foxp3 expression and regulatory T cell differentiation independently of interleukin 2. Nat. Immunol. 6, 152162 (2005). 34. Taniuchi, I. et al. Differential requirements for Runx proteins in CD4 repression and epigenetic silencing during T lymphocyte development. Cell 111, 621633 (2002). 35. Brugnera, E. et al. Coreceptor reversal in the thymus: signaled CD4+8+ thymocytes initially terminate CD8 transcription even when differentiating into CD8+ T cells. Immunity 13, 5971 (2000). 36. Bouillet, P. et al. BH3-only Bcl-2 family member Bim is required for apoptosis of autoreactive thymocytes. Nature 415, 922926 (2002).
npg
NUMBER 6
JUNE 2012
577
Articles
37. Hu, Q., Sader, A., Parkman, J.C. & Baldwin, T.A. Bim-mediated apoptosis is not necessary for thymic negative selection to ubiquitous self-antigens. J. Immunol. 183, 77617767 (2009). 38. Kovalovsky, D., Pezzano, M., Ortiz, B.D. & SantAngelo, D.B. A novel TCR transgenic model reveals that negative selection involves an immediate, Bim-dependent pathway and a delayed, Bim-independent pathway. PLoS ONE 5, e8675 (2010). 39. Suen, A.Y. & Baldwin, T.A. Proapoptotic protein Bim is differentially required during thymic clonal deletion to ubiquitous versus tissue-restricted antigens. Proc. Natl. Acad. Sci. USA 109, 893898 (2012). 40. Erlacher, M. et al. Puma cooperates with Bim, the rate-limiting BH3-only protein in cell death during lymphocyte development, in apoptosis induction. J. Exp. Med. 203, 29392951 (2006). 41. McCaughtry, T.M., Baldwin, T.A., Wilken, M.S. & Hogquist, K.A. Clonal deletion of thymocytes can occur in the cortex with no involvement of the medulla. J. Exp. Med. 205, 25752584 (2008). 42. Gao, J.X. et al. Perinatal blockade of b71 and b72 inhibits clonal deletion of highly pathogenic autoreactive T cells. J. Exp. Med. 195, 959971 (2002). 43. Wang, R., Wang-Zhu, Y. & Grey, H. Interactions between double positive thymocytes and high affinity ligands presented by cortical epithelial cells generate double negative thymocytes with T cell regulatory activity. Proc. Natl. Acad. Sci. USA 99, 21812186 (2002). 44. Lambolez, F., Kronenberg, M. & Cheroutre, H. Thymic differentiation of TCR+ CD8+ IELs. Immunol. Rev. 215, 178188 (2007). 45. Leishman, A.J. et al. Precursors of functional MHC class I- or class II-restricted CD8+ T cells are positively selected in the thymus by agonist self-peptides. Immunity 16, 355364 (2002). 46. Levelt, C.N. et al. High- and low-affinity single-peptide/MHC ligands have distinct effects on the development of mucosal CD8 and CD8 T lymphocytes. Proc. Natl. Acad. Sci. USA 96, 56285633 (1999). 47. Baldwin, T.A., Sandau, M.M., Jameson, S.C. & Hogquist, K.A. The timing of TCR expression critically influences T cell development and selection. J. Exp. Med. 202, 111121 (2005). 48. Takahama, Y., Shores, E.W. & Singer, A. Negative selection of precursor thymocytes before their differentiation into CD4+CD8+ cells. Science 258, 653656 (1992).
npg
578
VOLUME 13
NUMBER 6
JUNE 2012
nature immunology
ONLINE METhODS
Animals. BALB/c, C57BL/6, Cd28/, B2m/, Ab/b2m/, Zap70/, Rag2/ and human Bcl-2transgenic mice20 were maintained in our own colony at the US National Institutes of Health. Cd80/Cd86/ (B7-deficient) mice were provided by A. Sharpe, Runx3-YFP reporter mice were provided by D. Littman31 and Pdcd1/ mice were provided by T. Honjo. The Mcl-1transgenic construct was made by ligation of cDNA encoding Mcl-1 into a vector based on the human CD2 enhancer-promoter and was injected into fertilized C57BL/6 oocytes for the generation of Mcl-1-transgenic mice. All mice were cared for in accordance with guidelines of the US National Institutes of Health. Flow cytometry. Monoclonal antibodies with the following specificities were used: CD4 (GK1.5 and RM4.5), CD5 (53-7.3), CD8 (53-6.7), CD8 (H35-17.2), TCR (GL3), TCR (H57-597), V3 (KJ25), V4 (KT4), V5.1-5.2 (MR9-4), V6 (RR4-7), V8 (F23.1), V10 (B21.5), V11 (RR3-15), V12 (MR11-1), NK1.1 (PK136), PD-1 (J43), CD132 (4G3), CD122 (TM-1), CD44 (IM7), CD69 (H1.2F3), Bcl-2 PE-set (3F11), IL-4R (mIL4R-M1), IL-7R (SB/199), CD25 (PC61) or CD28 (37.51; all from BD Pharmingen); Bcl-XL (54H6) or the cleaved form of caspase-3 (D175; both from Cell Signaling; IL-2 (JES-5H4; eBioscience); or CD8 (5H10; Invitrogen). The Annexin-V FLUOS staining kit was from Roche Applied Science. PBS57-loaded or unloaded CD1d tetramers were provided by the Tetramer Core Facility of the US National Institutes of Health. Cells were analyzed on a FACSVantage SEM or LSR II (BD). Doublets and dead cells were removed from data analysis. For intracellular staining, live cells were first stained for surface proteins, then fixed and permeabilized, and then stained for intracellular proteins. Data were analyzed with software designed by the Division of Computer Research and Technology of the US National Institutes of Health or with FlowJo software by TreeStar. Cell sorting and purification. CD8+ and lymph node T cells were isolated by antibody-mediated depletion of CD4+ and immunoglobulin-positive cells through the use of magnetic beads (BioMag Qiagen). For the purification of TCR+ DN thymocytes, thymocyte samples were depleted of CD4+ and CD8+ cells by antibody-mediated depletion through the use of magnetic beads, followed by electronic sorting for CD5+CD4CD8B220GL3NK1.1 cells or CD5+CD4CD8B220GL3CD1dPBS57 cells. IELs were isolated as described49. For the purification of TCR+CD8+ IELs, IELs were electronically sorted as CD45+CD8+CD8CD4GL3 cells. Calcium mobilization. Cells were loaded at 31 C with the calcium-sensitive dye Indo-1 (1.8 mM; Invitrogen) and then coated at 4 C with biotinylated antiTCR. Cells were warmed for 2 min before stimulation and then applied to the flow cytometer. Antibody crosslinking was induced with avidin (4 g/ml, Sigma) and data acquisition was recorded for 5 min. Adoptive transfer. Purified TCR+ DN thymocytes or CD8+CD5+ lymph node T cells (0.8 106) were injected into host mice deficient in
recombination-activating genes. Five weeks after transfer, cells from the thymus, peripheral lymphoid organs and small intestine were isolated and analyzed by flow cytometry. In vitro T cell proliferation and differentiation cultures. For in vitro differentiation, purified TCR+ DN thymocytes were cultured for 45 d in medium only or were stimulated for 45 d with immobilized monoclonal anti-TCR (5-10 g/ml) in the presence of IL-15 (100 ng/ml; R&D Systems), then were collected and analyzed by flow cytometry. For in vitro proliferation, sorted TCR+ DN and SP8 thymocytes from B7-deficient mice were stimulated for 72 h with immobilized monoclonal anti-TCR (5 g/ml) and monoclonal anti-CD28 (10 g/ml) in the presence or absence of recombinant IL-2 (200 U/ml). For in vitro proliferation of IELs, sorted TCR+CD8 and SP4 lymph node T cells were stimulated for 48 h with anti-CD3 (5 g/ml) in the presence of 1 105 antigen-presenting cells (syngenic splenocytes irradiated with 3,000 rads) in the presence or absence of recombinant IL-2 (200 U/ml) or IL-15 (100 ng/ml). Cultures were pulsed with [3H]thymidine (1 Ci) 8 h before collection. For analysis of intracellular IL-2 production, T cells that had been stimulated for 4 d were treated for 4 h with phorbol 12-myristate 13-acetate (50 ng) and ionomycin (500 nM) in the presence of protein-transport inhibitor (BD Biosciences), followed by intracellular staining for IL-2. Genomic PCR. PCR analysis of DNA from the mouse tail was used for the detection of genomes of integrated Mtv-6 and Mtv-9 proviruses with the following primers: Mtv-6 forward, 5-GCTGGCTATCATCACAAGAGCG-3, and reverse, 5-GGAGTTCAACCATTTCTGCTGC-3; Mtv-9 forward, 5-ACC GCAGTCAAAGAACAGGTGC-3, and reverse, 5-CAGGAAACCACTT GTCTCACATCC-3. DNA-methylation analysis. Methylation was analyzed as described50. DNA was isolated with the ZR genomic DNA II kit (Zymo Rearch). Bisulfite conversion of total DNA was achieved with the EZ-DNA methylation-Gold kit (Zymo Research). The following PCR primers were used: CD8 promoter forward, 5-TTGAAAAGTTAAGGTTTTGATGTT-3, and reverse, 5-AAA CACTATTCCCCTCAATACTCTATC-3. PCR products were purified and cloned with the TOPO-TA cloning kit. Plasmid DNA from MiniPrep (Qiagen) was sequenced. Statistical analysis. Statistical significance was determined by Students t-test with two-tailed distribution.
npg
49. Cruz, D. et al. An opposite pattern of selection of a single T cell antigen receptor in the thymus and among intraepithelial lymphocytes. J. Exp. Med. 188, 255265 (1998). 50. Liu, B., Tahk, S., Yee, K.M., Fan, G. & Shuai, K. The ligase PIAS1 restricts natural regulatory T cell differentiation by epigenetic repression. Science 330, 521525 (2010).
doi:10.1038\ni.2292
nature immunology
Articles
Constitutive MHC class I molecules negatively regulate TLR-triggered inflammatory responses via the FpsSHP-2 pathway
Sheng Xu1,4, Xingguang Liu1,4, Yan Bao1, Xuhui Zhu1, Chaofeng Han1, Peng Zhang1, Xuemin Zhang2, Weihua Li2 & Xuetao Cao1,3
The molecular mechanisms that fine-tune Toll-like receptor (TLR)-triggered innate inflammatory responses remain to be fully elucidated. Major histocompatibility complex (MHC) molecules can mediate reverse signaling and have nonclassical functions. Here we found that constitutively expressed membrane MHC class I molecules attenuated TLR-triggered innate inflammatory responses via reverse signaling, which protected mice from sepsis. The intracellular domain of MHC class I molecules was phosphorylated by the kinase Src after TLR activation, then the tyrosine kinase Fps was recruited via its Src homology 2 domain to phosphorylated MHC class I molecules. This led to enhanced Fps activity and recruitment of the phosphatase SHP-2, which interfered with TLR signaling mediated by the signaling molecule TRAF6. Thus, constitutive MHC class I molecules engage in crosstalk with TLR signaling via the FpsSHP-2 pathway and control TLR-triggered innate inflammatory responses. Toll-like receptors (TLRs) are key pattern-recognition receptors used by cells of the innate immune system to detect the conserved components of pathogens, and they have critical roles in host defense against invading microbial pathogens1,2. After the recognition of pathogen-associated molecule patterns, TLRs initiate innate immune responses by activating signaling pathways that depend on the adaptor MyD88 or the adaptor TRIF and consequently induce the production of proinflammatory cytokines and type I interferon 3. Many factors have been identified as being essential for full activation of TLR responses; however, inhibitors of TLR pathways need further investigation, as inappropriate activation or overactivation of TLR signaling may result in inflammatory disorders such as septic shock or autoimmune diseases4. The identification of the negative regulators and details of the mechanisms by which TLR signaling is fine tuned remain to be fully elucidated. Major histocompatibility complex (MHC) class I molecules are expressed on all nucleated cells from vertebrates. Classical MHC class I molecules (HLA-I in humans; H-2 I in mice) are heterodimers composed of 2-microglobulin (2m) and a membrane-bound heavy chain with three extracellular domains and a cytoplasmic domain of about 40 amino acids5. The component 2m is essential for the stable expression of MHC class I, and 2m-deficient mice lack most if not all MHC class I molecules6. The main function of MHC class I molecules seems to be the presentation on the cell surface of small-peptide fragments of endogenously produced antigens for recognition by CD8+ cytotoxic T lymphocytes6. MHC class I molecules also contribute to the development and selection of CD8+ T cells in the thymus7, as well as to the education and tolerance of natural killer (NK) cells8,9. Normal expression of MHC class I molecules by target cells inhibits the lysis of target cells by NK cells. However, chronic interaction between MHC class I molecules and cognate inhibitory receptors is essential for the licensing of NK cells and for the acquisition of their killing ability9,10. In these cases, MHC class I molecules act as ligands to stimulate relevant receptors to trigger downstream signaling. In addition to the classical function of antigen presentation, MHC class I molecules can also mediate reverse signaling after ligation and have nonclassical functions1113. The aggregation of MHC class I molecules on the cell surface with agonist antibodies, T cell antigen receptors (TCRs) or the coreceptor CD8 activates signaling pathways in T cells, B cells, tumor cells or endothelial cells and elicits various biological effects, such as cell apoptosis, activation or proliferation1315. Crosslinkage of MHC class I molecules on human NK cells induces intracellular tyrosine phosphorylation and inhibits NK cell cytotoxicity directed against tumor target cells16. Such data suggest that MHC class I molecules can be ligands and signaling receptors themselves to mediate reverse signaling via association with other receptors or directly through aggregation. The activation of tyrosine kinases, including Lck, Lyn, Syk, Zap70 and Tyk2, occurs immediately after engagement of MHC class I molecules 17,18. So far, almost all data about the nonclassical functions of MHC class I molecules have been obtained from studies of lymphoid cells; however, whether myeloid cells or antigen-presenting cells (APCs) display MHC class I molecules with nonclassical functions and reverse signaling has not been elucidated. Intracellular MHC class II molecules do promote
npg
1National Key Laboratory of Medical Immunology & Institute of Immunology, Second Military Medical University, Shanghai, China. 2National Center of Biomedical Analysis, Beijing, China. 3National Key Laboratory of Molecular Biology, Chinese Academy of Medical Sciences, Beijing, China. 4These authors contributed equally to this work. Correspondence should be addressed to X.C. (caoxt@immunol.org).
Received 28 October 2011; accepted 6 March 2012; published online 22 April 2012; corrected online 4 May 2012; doi:10.1038/ni.2283
NUMBER 6
JUNE 2012
551
Articles
Figure 1 MHC class Ideficient mice are more susceptible to TLR challenge. (a) Enzyme-linked immunosorbent assay (ELISA) of TNF, IL-6 and IFN- in serum from B2m+/+ and B2m/ mice (n = 5 per genotype) 2 h after intraperitoneal challenge with PBS, LPS, poly(I:C) or CpG-ODN. (b) Survival of B2m+/+ and B2m/ mice (n = 10 per genotype) after lethal challenge with LPS (10 mg per kg body weight). P < 0.01 (Wilcoxon test). (c) Hematoxylin-and-eosin staining of lungs from B2m+/+ and B2m/ mice 8 h after challenge with PBS or LPS. Original magnification, 100. (d) ELISA of TNF, IL-6 and IFN- in serum from wild-type mice (n = 6 per group) reconstituted for 8 weeks with B2m+/+ bone marrow cells (B2m+/+B2m+/+) or B2m/ bone marrow cells (B2m/B2m+/+), followed by challenge with PBS or LPS and analysis 2 h later. *P < 0.05 and **P < 0.01 (Students t-test). Data are from three independent experiments (a,d; mean s.e.m.) or are representative of three independent experiments with similar results (b,c).
a
TNF (ng/ml)
4 3 2 1 0
**
IFN- (ng/ml)
IL-6 (ng/ml)
B2m B2m/
+/+
20 16 12 8 4 0
** **
S
2.0 1.5 1.0 0.5 0
** ** *
** **
S L Po PS ly (I: C ) C pG
S ly (I: C ) C pG
S
PBS
b
Survival (%)
100 80 60 40 20 0
B2m+/+ / B2m
c
B2m+/+
B2m
0 8 16 24 32 40 48 56 Time after LPS challenge (h) IFN- (ng/ml) 12 IL-6 (ng/ml) 2.5
TLR-triggered innate immune responses by maintaining activation of the kinase Btk19. Whether and how MHC class I molecules intersect with innate TLR signaling pathways and regulate TLR-triggered innate inflammatory response remains unknown. Here we demonstrate that deficiency in MHC class I resulted in enhanced TLR-triggered production of proinflammatory cytokines and type I interferon both in vitro and in vivo. Consistent with that, engagement of MHC class I molecules inhibited TLR-triggered signal transduction and production of cytokines. Constitutive membrane MHC class I molecules physiologically interacted with the tyrosine kinase Fps (Fes). After the stimulation of TLRs with ligands, MHC class I molecules became phosphorylated and recruited more Fps, which resulted in more potent activation of the phosphatase SHP-2 and, finally, suppressed TLR-triggered inflammatory responses. Therefore, our results demonstrate a nonclassical function for MHC class I molecules as negative regulators of a TLR pathway, which adds new insight into the fine tuning of TLR-triggered innate inflammatory responses. RESULTS MHC class I deficiency exacerbates TLR-triggered responses To investigate the role of MHC class I molecules in TLR-triggered innate inflammatory responses, we challenged 2-microglobulindeficient (B2m/) mice with various TLR ligands, including lipopolysaccharide (LPS), the synthetic RNA duplex poly(I:C) or an oligodeoxynucleotide based on the dinucleotide motif CpG (CpG ODN). MHC class Ideficient mice produced significantly more tumor necrosis factor (TNF), interleukin 6 (IL-6) and interferon- (IFN-) than did their littermates (control mice; Fig. 1a). We detected no substantial difference between B2m/ and wild-type mice in the development of myeloid cells or the expression of TLRs (data not shown), which excluded the possibility that the phenomenon noted above was due to the abnormal development of myeloid cells or TLR expression. Furthermore, after lethal challenge with LPS, almost all MHC class Ideficient mice died within 24 h, whereas 40% of their littermates (control mice) were alive at that time and about 25% of these survived the challenge (Fig. 1b). Consistent with that, we observed more severe infiltration of polymorphonuclear cells and interstitial pneumonitis in the lungs of MHC class Ideficient mice 8 h after LPS challenge (Fig. 1c). Furthermore, we also confirmed the greater production of TNF, IL-6 and IFN- in mice deficient in the MHC class I heavy chain (H2-K1/H2-D1/; called KbDb/ here) after TLR challenge (Supplementary Fig. 1a). Given that MHC class Ideficient mice have considerably fewer CD8+ T cells in the thymus, spleen and lymph nodes, we further reconstituted lethally irradiated wild-type mice with wild-type, B2m/ or
552
d
TNF (ng/ml)
4 3 2 1 0 PBS LPS
**
**
2.0 1.5 1.0 0.5 0
**
B2m+/+B2m+/+ B2m/B2m+/+
8 4 0 PBS LPS
PBS LPS
KbDb/ bone marrow cells and then assessed their responses to challenge with the TLR ligands. When challenged with LPS, chimeric mice reconstituted with B2m/ or KbDb/ bone marrow produced more TNF, IL-6 and IFN- than did chimeric mice reconstituted with wild-type bone marrow (Fig. 1d and data not shown). Thus, these data demonstrated that TLR-triggered inflammatory responses were greater in MHC class Ideficient mice, which indicated a suppressive role for MHC class I molecules in the TLR response. To investigate the role of MHC class I molecules in host resistance to pathogen infection, we challenged B2m/ mice and KbDb/ mice with Gram-negative Escherichia coli or Gram-positive Listeria monocytogenes. After infection with E. coli, the production of TNF, IL-6 and IFN- in MHC class Ideficient mice was significantly greater than that of their littermates (control mice; Fig. 2a and Supplementary Fig. 1b). Accordingly, MHC class Ideficient mice had a greater load of E. coli bacteria in the blood (Fig. 2b and Supplementary Fig. 1b), consistent with published findings that proinflammatory cytokines promote the dissemination of E. coli20. After infection with L. monocytogenes, MHC class Ideficient mice also produced more proinflammatory cytokines and had a smaller bacterial load in the spleen and liver (Fig. 2c,d and Supplementary Fig. 1c). Furthermore, consistent with published reports21, NK cells from MHC class Ideficient mice showed less cytotoxicity directed against target cells and produced less IFN- after infection with L. monocytogenes (data not shown), which excluded the possibility that the smaller bacterial load in MHC class Ideficient mice was a result of enhanced killing by NK cells. These data indicated that MHC class I molecules may help the host restrict inflammatory responses after bacterial infection and protect the host from inflammatory injuries mediated by innate immune responses. More cytokine production in MHC class Ideficient APCs APCs, especially macrophages, are the main mediators of TLRtriggered innate inflammatory responses in vivo. Consistent with the in vivo data presented above, peritoneal macrophages from B2m/ mice produced more TNF, IL-6 and IFN- than did those from their littermates (control mice) in response to stimulation with LPS, poly(I:C) or CpG ODN (Fig. 3a and Supplementary Fig. 2a). We obtained similar results with KbDb/ macrophages (Supplementary Fig. 1d). We also found exacerbated cytokine production in
npg
L Po PS ly (I: C ) C pG
LPS
PB
PB
LP
Po
PB
Articles a
TNF (ng/ml)
5 4 3 2 1 0 2h
Liver LM (log10 CFU)
Blood E. coli (Log10 CFU/ml)
IFN- (ng/ml)
IL-6 (ng/ml)
IFN- (ng/ml)
TNF (ng/ml)
6 4 2 0 2h 4h
1.5 1.0 0.5 0 2h 4h
IL-6 (ng/ml)
2.0
**
**
6 4 2 0
Spleen LM (log10 CFU)
**
10
B2m+/+
B2m/
2.5
B2m+/+ B2m/
c
4 3 2 1 0 PBS LM
B2m+/+
B2m/
d
1.5 1.0 0.5 0
B2m+/+ 10 8 6 4 2 0
1d 3d
B2m/ 10 8 6 4 2 0
1d 3d
**
4 3 2 1 0 PBS
** **
4h
4h 8h
LM
PBS
LM
Figure 2 MHC class Ideficient mice are more susceptible to infection with E. coli but are more resistant to infection with L. monocytogenes (LM). (a,b) ELISA of TNF, IL-6 and IFN- in serum (a) and analysis of bacterial loads in blood (b) of B2m+/+ and B2m/ mice (n = 4 per genotype) 2, 4 or 8 h (horizontal axes) after intraperitoneal infection with 1 10 8 E. coli. CFU, colony-forming units. (c,d) ELISA of TNF, IL-6 and IFN- in serum (c) and analysis of bacterial loads in spleen and liver (d) of B2m+/+ and B2m/ mice (n = 4 per genotype) after treatment with PBS or infection with 1 10 4 L. monocytogenes; serum was obtained 4 h after infection. In b,d, each symbol represents an individual mouse; small horizontal lines indicate the mean. *P < 0.05 and **P < 0.01 (Students t-test). Data are from three independent experiments with similar results (mean s.e.m.).
MHC class Ideficient bone marrowderived dendritic cells (Supplementary Fig. 2b). We then investigated the effect of knockdown of MHC class I on cytokine expression in macrophages in which TLRs were triggered. Macrophages in which the gene encoding H-2Kb was silenced by H-2Kb-specific small interfering RNA (siRNA) produced significantly more proinflammatory cytokines and IFN- in response to stimulation with LPS, poly(I:C) or CpG ODN than did those transfected with control siRNA (Fig. 3b,c). To further demonstrate that the enhanced TLR-triggered inflammatory response in MHC class Ideficient mice in vivo was due to the MHC class I deficiency in macrophages, we adoptively transferred MHC class Ideficient macrophages into wild-type mice depleted of endogenous macrophages via pretreatment with clodronate liposomes. After LPS challenge, mice reconstituted with MHC class Ideficient (B2m/ or KbDb/) macrophages produced more proinflammatory cytokines and IFN- than did those reconstituted with wild-type macrophages (Fig. 3d and data not shown). Therefore, these data suggested that MHC class I deficiency enhanced TLRtriggered inflammatory responses in macrophages and that this may have resulted in the greater susceptibility of MHC class Ideficient mice to lethal challenge with LPS observed above.
Ligation of MHC class I molecules via monoclonal antibodies induces downstream signals in lymphocytes and elicits biological functions1315. After stimulation with LPS, poly(I:C) or CpG ODN, macrophages crosslinked with antibody to H-2Kb (anti-H-2Kb) secreted less proinflammatory cytokines and IFN- than did macrophages treated with control antibody (Fig. 3e). Crosslinkage of another MHC class I molecule, H2-Db, also suppressed the TLR-triggered inflammatory responses in macrophages (Supplementary Fig. 2c), whereas crosslinkage of both H-2Kb and H2-Db resulted in even less production of these cytokines (data not shown). These data suggested that crosslinkage of MHC class I molecules exerted an inhibitory effect on TLR-triggered inflammatory responses in macrophages. CD8+ T cells attenuate TLR responses via MHC class I To determine the physiological relevance of our observation that reverse signals from MHC class I molecules suppressed TLR-triggered immune responses, we looked for cells in vivo that provided ligands for MHC class I molecules on APCs. The known natural ligands for MHC class I in vivo are TCRs, the coreceptor CD8, the lectin-like receptor CD94-NKG2 and killer immunoglobulin-like receptors expressed mainly on NK cells and CD8+ T cells, which made these
npg
a
TNF (ng/ml)
5 4 3 2 1 0
B2m+/+
B2m/ IL-6 (ng/ml)
8 6 4 2 0
IFN- (ng/ml)
IL-6 (ng/ml)
2 1 0
3 2 1 0
** *
6 4 2 0
IFN- (ng/ml)
TNF (ng/ml)
**
**
**
5 4
Ctrl
MHCI
2 1 0
* **
PB S Po LP ly S (I: C ) C pG
S Po LPS ly (I: C ) C pG
PB S L Po PS ly (I: C ) C pG
S Po LP ly S (I: C ) C pG
d
5 4 3 2 1 0 TNF (ng/ml)
B2m+/+B2m+/+ IL-6 (ng/ml)
B2m/B2m+/+ IFN- (ng/ml) 8 6 4 2 0 4 6 Time (h)
e
TNF (ng/ml)
3 2 1 0
IgG
**
H-2Kb IL-6 (ng/ml)
siRNA: Ctrl MHCI -actin
MHCI
2.0 1.5 1.0 0.5 0
3 2 1 0
**
IFN- (ng/ml)
* *
**
**
**
**
3 2 1 0
Figure 3 MHC class I reverse signaling inhibits TLR-triggered production of proinflammatory cytokines and type I interferon in macrophages. (a) ELISA of TNF, IL-6 and IFN- in supernatants of B2m+/+ and B2m/ macrophages stimulated for 4 h with PBS, LPS, poly(I:C) or CpG ODN. (b) Immunoblot analysis of MHC class I (MHCI) in macrophages 48 h after transfection with control (Ctrl) siRNA or siRNA specific for MHC class I. -actin serves as a loading control throughout. (c) ELISA of TNF, IL-6 and IFN- in macrophages transected with siRNA as in b (key), then stimulated for 4 h with LPS, poly(I:C) or CpG. (d) ELISA of TNF, IL-6 and IFN- in serum from wild-type mice (n = 4 per group) depleted of endogenous macrophages and then given adoptive transfer of wild-type macrophages (B2m+/+B2m+/+) or MHC class Ideficient macrophages (B2m/B2m+/+), followed by LPS challenge and analysis 2 h later. (e) ELISA of TNF, IL-6 and IFN- in macrophages after ligation with control antibody (immunoglobulin G (IgG)) or monoclonal antibody to MHC class I (H-2Kb) and stimulation for 4 h with PBS, LPS, poly(I:C) or CpG ODN. *P < 0.05 and **P < 0.01 (Students t-test). Data are from three independent experiments (a,ce; mean s.e.m.) or are representative of three independent experiments with similar results (b).
4 6 Time (h)
4 6 Time (h)
PB S
PB
PB S
S L Po PS ly (I: C ) C pG
PB
PB
PB
PB
** *
NUMBER 6
JUNE 2012
553
Articles a
5 TNF (ng/ml) 4 3 2 1 0 PBS LPS CD8+ T cell depletion 15 IL-6 (ng/ml) 10 5 0
WT
IFN- (ng/ml)
1.0 0.5 0
1.5 1.0 0.5 0
3 2 1 0
**
0.5 0
+ / +/
TNF (ng/ml)
IL-6 (ng/ml)
IFN- (ng/ml)
TNF (ng/ml)
**
NS
NS
1.0
NS
1.5 1.0 0.5 0
TNF (ng/ml)
**
**
NK cell depletion 1.5
b
**
2.0
M + CD8+ T cells
c
NS
* *
PBS
LPS
PBS
LPS
Figure 4 CD8+ T cells suppress TLR-triggered production of inflammatory cytokines in macrophages ** 3 8 ** 2 in an MHC class Idependent manner. (a) ELISA of TNF, IL-6 and IFN- in the serum of wild-type * * NS mice (n = 4 per group) left undepleted (WT) or depleted of CD8 + T cells or NK cells, followed by 6 2 LPS challenge and analysis 2 h later. (b) ELISA of TNF, IL-6 and IFN- in supernatants of B2m+/+ or B2m/ macrophages cultured alone (M) or at a ratio of 1:1 with CD8+ T cells (M + CD8+ 4 1 T cells) and simulated for 4 h with LPS. (c) ELISA of TNF in supernatants of macrophages cultured 1 2 alone (M) or with NK cells (M + NK) and simulated for 4 h with LPS. (d) ELISA of TNF in wild-type macrophages cultured in the presence (+) or absence () of CD8 + T cells, antiIL-10 0 0 0 and antiTGF-, with (+) or without () a Transwell system (to separate CD8 + T cells in the upper +/+ or B2m/ chamber), then stimulated for 4 h with LPS (100 ng/ml). (e) TNF in serum from B2m mice (left) or B2m/ mice given transfer of wild-type CD8+ T cells (B2m/ + CD8+; n = 4 mice per group), assessed 2 h after LPS challenge. (f) ELISA of TNF in macrophages cultured alone (M) or with CD8+ T cells (M + CD8+) or OT-I cells (M + OT-I) in the presence (+ OVA) or absence of ovalbumin peptide (amino acids 257264). (g) ELISA of TNF in wild-type macrophages cultured alone (M) or with naive (M + naive) or memory (M + mem) CD8+ T cells (sorted by flow cytometry as CD8+CD44lo or CD8+CD44hi cells, respectively) and stimulated for 2 h with LPS (100 ng/ml). NS, not significant; *P < 0.05 and **P < 0.01 (Students t-test). Data are from three independent experiments with similar results (mean s.e.m.).
0 M CD8+ T cell Transwell AntiIL-10 AntiTGF-
+ +
+ + +
+ + +
+ + +
M M + N K
+/
B 2m
+/
B 2m
B 2m
B 2m
B 2m
B 2m
TNF (ng/ml)
TNF (ng/ml)
B 2m + B 2m B2 /+ / m / + C D 8+
M M+ + CD O + 8+ T- O I + TO I VA
TNF (ng/ml)
cells the most likely candidates. In vivo depletion of NK cells or CD8+ T cells resulted in significantly enhanced production of TNF, IL-6 and IFN- in mice after LPS challenge (Fig. 4a), consistent with a published report22. We further cultured macrophages together with NK cells or CD8+ T cells in vitro and found that culture with CD8+ T cells resulted in significantly impaired production of TNF, IL-6, IFN- by LPS-stimulated macrophages, whereas culture with NK cells did not (Fig. 4b,c). To demonstrate the underlying mechanisms by which CD8 + T cells inhibited macrophage inflammatory responses to TLR ligands, we used a Transwell system and found that the inhibitory effect of CD8+ T cells was attenuated when the cells were physically separated (Fig. 4d). In addition, blockade of the inhibitory cytokines IL-10 or TGF- did not relieve the suppressive effect of CD8+ T cells (Fig. 4d). Therefore, CD8+ T cellmediated suppression of the innate inflammatory responses of macrophages was dependent on cell-cell contact. To elucidate whether MHC class I molecules were involved in the process, we cultured CD8+ T cells together with MHC class Ideficient macrophages. Notably, the suppressive effect of CD8+ T cells was completely abrogated (Fig. 4b). Additionally, when we adoptively transferred wild-type CD8 + T cells into B2m/ mice whose macrophages lack expression of MHC class I, we observed no significant attenuation of inflammatory responses after LPS stimulation (Fig. 4e). These data suggested a nonredundant role for MHC class I expression on macrophages in the suppression of innate responses by CD8+ T cells. We further determined whether engagement of TCRs by peptide MHC class I complexes was required. OT-I cells (which transgenically express an ovalbumin-specific TCR) tempered cytokine production by macrophages just as nontransgenic CD8+ T cells did in the absence of cognate peptide (Fig. 4f). However, after the addition of ovalbumin peptide, the suppressive effect of OT-I cells was abrogated, whereas the inhibitory effect of nontransgenic CD8+ T cells was not influenced by the addition of ovalbumin peptide (Fig. 4f and data not shown). We also found that naive CD8+ T cells had a greater effect than memory T cells had in suppressing cytokine production by macrophages in which TLRs were triggered (Fig. 4g). Together these data suggested
554
that naive CD8+ T cells suppress innate inflammatory responses in macrophages in an MHC class Idependent way. Enhanced TLR signaling in MHC class Ideficient macrophages To investigate whether MHC class I deficiency intersected with the TLR signaling pathways in macrophages, we examined the activation kinetics of the mitogen-activated protein kinase (MAPK) and transcription factor NF-B pathways, which are both downstream of TLR signaling. Activation of the MAPKs Jnk, Erk and p38, the kinases IKK and IKK and the NF-B inhibitor IB was enhanced in LPSstimulated MHC class Ideficient macrophages (Fig. 5a and data not shown). We also observed more phosphorylation of the transcription factor IRF3 (Fig. 5a). We obtained similar results with MHC class Ideficient macrophages stimulated with poly(I:C) or CpG ODN (Supplementary Fig. 3a). These data suggested that deficiency in MHC class I enhances TLR signaling in macrophages.
npg
B2m+/+ B2m/ LPS (min) 0 15 30 45 60 0 15 30 45 60 p-Erk p-Jnk p-p38
IgG H-2Kb LPS (min) 0 15 30 45 60 0 15 30 45 60 p-Erk p-Jnk p-p38
p-IKK-IKK p-IB -actin B2m+/+ B2m/ LPS (min) 0 30 60 90 120 0 30 60 90 120 p-IRF3 -actin
p-IKK-IKK p-IB -actin IgG H-2Kb LPS (min) 0 30 60 90 120 0 30 60 90 120 p-IRF3 -actin
Figure 5 MHC class I reverse signaling impairs TLR pathways in macrophages. (a) Immunoblot analysis of phosphorylated (p-) signaling molecules in lysates of B2m+/+ and B2m/ macrophages stimulated for 0120 min (above lanes) with LPS. (b) Immunoblot analysis of phosphorylated signaling molecules in lysates of macrophages treated with control antibody (IgG) or crosslinked with monoclonal antibody to MHC class I (H-2Kb) and stimulated with LPS. Data are from one experiment representative of three independent experiments with similar results.
VOLUME 13
NUMBER 6
JUNE 2012
nature immunology
M M +n a + ive m em
Articles a
IP: MHCI IP: IgG IB: Fps IB: MHCI
PM + + + +
Cyt 15 0 5 15 + + + +
b
LPS (min) 15 0 5 15 IP: Fps IP: IgG IB: MHCI IB: Fps + + + +
c
LPS (0 min)
Fps
MHCI
Merge
LPS (min) 15 0 5 15
Mock Flag-H2K(WT) Flag-H2K(del) Flag-H2K(mut) Myc-Fps IP: Flag IB: Myc IB: Flag
+ + +
+ + +
+ + +
+ + +
e
Mock Myc-Fps Myc-FpsSH2 Flag-H2K IP: Myc IB: Flag IB: Myc KbDb+/+
b b/
+ + +
+ + +
+ + +
LPS (5 min)
f K D 3 Figure 6 Membrane MHC class I molecules interact directly with Fps. 8 * 5 * * * * * * * * (a) Immunoassay of lysates of peritoneal macrophages stimulated for 2 4 6 3 0, 5 or 15 min with LPS, followed by immunoprecipitation (IP +) of 4 1 2 plasma-membrane (PM) or cytoplasmic (Cyt) proteins with IgG or 2 1 antiMHC class I and immunoblot analysis (IB) with anti-Fps or 0 0 0 antiMHC class I. (b) Immunoassay of lysates of macrophages stimulated for 0, 5 or 15 min with LPS, followed by immunoprecipitation with IgG or anti-Fps and immunoblot analysis with antiMHC class I or anti-Fps. (c) Confocal microscopy of macrophages stimulated for 0 or 5 min with LPS and labeled with antiMHC class I and anti-Fps. Original magnification, 600. (d) Immunoassay of lysates of macrophages mock transfected (Mock +) or transfected to express Flag-tagged wild-type H-2K (Flag-H2K(WT)), H-2K lacking the intracellular domain (Flag-H2K(del)) or H-2K with substitution of the intracellular tyrosine site (Flag-H2K(mut)), plus Myc-tagged Fps (Myc-Fps), then stimulated with LPS, followed by immunoprecipitation with anti-Flag and immunoblot analysis with anti-Myc or anti-Flag. (e) Immunoassay of lysates of macrophages mock transfected or transfected to express Myc-tagged wild-type Fps (Myc-Fps) or mutant Fps with deletion of the SH2 domain (Myc-FpsSH2), plus Flag-tagged wild-type H-2K, then stimulated with LPS, followed by immunoprecipitation with anti-Myc and immunoblot analysis as in d. (f) ELISA of TNF, IL-6 and IFN- in supernatants of KbDb+/+ and KbDb/ macrophages mock transfected or transfected to express wild-type H-2K or the H-2K mutants in d. *P < 0.01 (Students t-test). Data are representative of three independent experiments with similar results (ae) or are from three independent experiments (f; mean s.e.m.).
TNF (ng/ml)
o 2K ck (W T H 2K ) ( H del) 2K (m ut )
IL-6 (ng/ml)
H ock 2K (W T H 2K ) ( H del 2K ) (m ut )
IFN- (ng/ml)
As ligation of MHC class I molecules attenuated TLR-triggered inflammatory responses, we further investigated the effect of MHC class I ligation on TLR pathways. Crosslinkage of MHC class I molecules on macrophages resulted in less activation of the MAPK, NF-B and IRF3 pathways in response to LPS stimulation (Fig. 5b), whereas crosslinkage alone did not activate any of these pathways (data not shown). We obtained similar results after stimulation with poly(I:C) or CpG ODN (Supplementary Fig. 3b). Collectively, these data indicated that ligation of MHC class I molecules may have inhibited the TLR-triggered production of proinflammatory cytokines and type I interferon by impairing activation of MyD88- and TRIF-dependent pathways in macrophages. Membrane MHC class I molecules bind Fps We further explored the underlying mechanisms of the suppression of TLR signaling by MHC class I molecules. First, we sought to determine whether MHC class I molecules interacted directly with TLRs and served as cofactors in inhibiting TLR signaling by forming a complex. Immunoprecipitation with antiMHC class I demonstrated no such interaction between TLRs and MHC class I molecules, and we obtained similar results by coimmunoprecipitation with anti-TLR as well (data not shown). Confocal microscopy also showed that the distribution of TLR3, TLR4 and TLR9 in endosomes in B2m/ macrophages was similar to that in wild-type macrophages after stimulation with the ligands for these TLRs (Supplementary Fig. 4ac). The downregulation of surface TLR4 in response to stimulation with LPS was also similar in wildtype and MHC class Ideficient macrophages (Supplementary Fig. 4d). These results demonstrated that deficiency in MHC class I did not affect the endosomal distribution of TLR4, TLR3 or TLR9 in macrophages after activation of the TLR, which suggested the existence of other molecules that interact with MHC class I and temper TLR signaling. Given that activation of many tyrosine kinases is involved in the control of TLR signaling, we immunoprecipitated proteins from lysates of LPS-stimulated macrophages with antiMHC class I and
then used reverse-phase nanospray liquid chromatographytandem mass spectrometry to identify possible MHC class Iassociated tyrosine kinases or other molecules that may be involved in the suppression of TLR pathways. Among the kinases shown to be associated with MHC class I, the non-receptor kinase Fps attracted our attention, as Fps-deficient mice have been shown to be more susceptible to LPSinduced sepsis23. An immunoprecipitation assay confirmed that Fps did indeed interact with MHC class I molecules at steady state, and this interaction increased after LPS stimulation (Fig. 6a,b). In addition, plasma-membrane MHC class I molecules associated with Fps, but intracellular MHC class I molecules did not (Fig. 6a). Confocal microscopy showed a dispersed pattern of Fps in the cytosol of unstimulated macrophages, although small amounts associated with MHC class I. After stimulation with TLR ligands, Fps was activated and aggregated as clusters along the cytoplasmic side of the membrane and showed substantially enhanced colocalization with membrane MHC class I molecules (Fig. 6c). MHC class I binds Fps via intracellular phosphorylated tyrosines We went further to determine which domains of MHC class I and Fps were required for their interaction and for the suppression of TLR-triggered responses. Distinct from MHC class II molecules, MHC class I molecules have a relatively long tail and a tyrosinephosphorylation site in the cytoplamic domain (Tyr320 in human and Tyr321 in mice), whereas Fps contains a Src-homology 2 (SH2) domain. Thus, we hypothesized that these two molecules may interact directly. We transfected mouse macrophages to express Myc-tagged Fps together with Flag-tagged wild-type H-2K, H-2K lacking the intracellular domain or H-2K with substitution of the intracellular tyrosine site. Coimmunoprecipitation showed that Fps interacted with wild-type MHC class I molecules but not with those with a mutant intracellular domain or tyrosine site (Fig. 6d). Mutant Fps with deletion of the SH2 domain did not interact with MHC class I either (Fig. 6e), which suggested a nonredundant role for the SH2 domain in the interaction of Fps and MHC class I. Gain-of-function experiments also showed that overexpression of wild-type MHC
555
npg
o 2K ck (W T H 2K ) (d el H 2K ) (m ut ) H
Articles a
B2m+/+ LPS (min) p-Fps Fps B2m/ IP: Fps LPS (min) p-Fps Fps
0 15 30 45 60 0 15 30 45 60
IFN- (ng/ml)
b
lgG
TNF (ng/ml)
IL-6 (ng/ml)
C trl Fp s
H-2Kb IP: Fps
c
siRNA: Fps -actin
Ctrl 3 2 1 0
Fps
**
0 15 30 45 60 0 15 30 45 60
**
4 3 2 1 0
**
**
**
2.5 2.0 1.5 1.0 0.5 0
**
** **
S Po LP ly S (I: C C ) pG
Figure 7 MHC class I molecules suppress TLR-triggered response by B2m/ B2m+/+ maintaining Fps activation. (a,b) Immunoassay of B2m+/+ and B2m/ * * macrophages (a) or wild-type macrophages crosslinked with IgG or 2.5 3 6 * *** * * ** 5 ** 2.0 ** * antiH-2Kb (b), stimulated for 060 min (above lanes) with LPS, followed * ** ** 2 4 1.5 ** ** by immunoprecipitation with anti-Fps and immunoblot analysis of 3 ** 1.0 1 2 phosphorylated Fps (p-Fps; detected by probing for phosphorylated tyrosine) 0.5 1 0 0 0 or total Fps. (c) Immunoblot analysis of Fps in lysates of macrophages siRNA: siRNA: siRNA: transfected for 48 h with control or Fps-specific siRNA. (d,e) ELISA LPS Poly CpG LPS Poly CpG LPS Poly CpG of TNF, IL-6 and IFN- in supernatants of wild-type C57BL/6 macrophages (d) (I:C) (I:C) (I:C) or B2m+/+ and B2m/ macrophages (e) transfected with control or Fps-specific siRNA and stimulated for 4 h with PBS, LPS, poly(I:C) or CpG ODN. *P < 0.05 and **P < 0.01 (Students t-test). Data are representative of three independent experiments with similar results (ac) or are from three independent experiments (d,e; mean s.e.m.).
C trl Fp s C trl Fp s C trl Fp s
IFN- (ng/ml)
TNF (ng/ml)
IL-6 (ng/ml)
class I significantly suppressed the production of TNF, IL-6 and IFN- in KbDb/ macrophages, whereas overexpression of mutant MHC class I did not have this effect (Fig. 6f). Therefore, the tyrosine site of MHC class I and the SH2 domain of Fps were required for their interaction and for the suppressive effect of MHC class I molecules on TLR-triggered responses. MHC class I suppresses TLR signaling via Fps activation As tyrosine-phosphorylation of the intracellular domain of MHC class I was required for binding to the SH2 domain of Fps, we further determined which signaling molecules contributed to this phosphorylation. The kinase Src is suggested to phosphorylate MHC class I molecules in Jurkat human T lymphocyte cells24. It has been shown that Src is activated in macrophages after stimulation with TLR ligands20 and is able to restrain TLR-triggered cytokine production. Suppression of Src activity with its inhibitor PP1 substantially suppressed the phosphorylation of MHC class I and resulted in less FpsMHC class I association in macrophages after stimulation with LPS (Supplementary Fig. 5a,b). Fps activation was also much lower when Src activity was inhibited (Supplementary Fig. 5c), which suggested a nonredundent role for Src in the interaction of MHC class I and Fps and also in the activation of Fps. Thus, Src may be involved in phosphorylation of the cytoplamic tail of MHC class I, allowing the association of MHC class I with Fps, which then transduces downstream signals. To investigate the exact role of Fps in the negative regulation of TLR signaling by MHC class I molecules, we examined the activation of Fps after stimulation with TLR ligands. Indeed, there was much less TLR ligandinduced phosphorylation of Fps in B2m/ or KbDb/ macrophages (Fig. 7a and data not shown). Crosslinkage of MHC class I molecules further enhanced their interaction with Fps and phosphorylation of Fps (Fig. 7b and Supplementary Fig. 6), which indicated that MHC class I molecules may have increased or maintained the TLR-triggered activation of Fps. Silencing of Fps resulted in more production of TNF, IL-6 and IFN- in macrophages stimulated with LPS, poly(I:C) or CpG ODN (Fig. 7c,d), which suggested a negative role for Fps in TLR responses. We also found that knockdown of Fps was less potent in suppressing the TLR-triggered production of cytokines in B2m/ macrophages than it was in wild-type macrophages (Fig. 7e). Accordingly, the inhibitory effect of crosslinkage of MHC class I molecules on TLR signaling was also substantially attenuated in macrophages in which the gene encoding Fps was silenced (data not shown). Together these data suggested
556
that membrane MHC class I molecules suppressed TLR-triggered inflammatory responses by associating with Fps and maintaining its activation. Fps suppresses TLR-triggered response by activating SHP-2 Fps activates the cell-adhesion molecule PECAM-1 (CD31) to mediate the suppressive effect of the receptor FcRI on mast-cell activation25, and published data have suggested that PECAM-1 can recruit SHP-2 to suppress TLR4 signaling26. Accordingly, we found that activation of SHP-2 in B2m/ or KbDb/ macrophages was considerably attenuated after stimulation with LPS (Fig. 8a and data not shown), which indicated that SHP-2 may have been involved in the suppression of TLR signaling by MHC class I molecules. Deficiency in SHP-2 enhanced the production of proinflammatory cytokines and IFN- in macrophages stimulated with LPS, poly(I:C) or CpG ODN (Fig. 8b). However, the inhibitory effect of crosslinkage of MHC class I on TLR-triggered cytokine production was considerably attenuated in SHP-2-deficient macrophages (Fig. 8c). These data indicated that activation of SHP-2 might have contributed to the suppression of TLR signaling in macrophages by MHC class I molecules. Next we investigated whether Fps contributed to the activation of SHP-2 enhanced by MHC class I molecules. After stimulation with TLR ligands, the phosphorylation of SHP-2 was considerably attenuated in macrophages in which the gene encoding Fps was silenced (Fig. 8d), which suggested that SHP-2 functions downstream of Fps. To more convincingly demonstrate the link between SHP-2 and Fps, we did coimmunoprecipitation assays and found an association between SHP-2 and Fps in resting macrophages, which was enhanced by stimulation with LPS (Fig. 8e). However, MHC class I molecules did not directly interact with SHP-2 (data not shown). SHP-2 suppresses TLR-triggered TRIF-dependent signal pathways by binding to and inhibiting activation of the kinase TBK1 (ref. 27). However, the molecules that interact with SHP-2 in the MyD88-dependent pathway have not been identified so far. As activation of both NF-B and MAPKs was enhanced in MHC class Ideficient macrophages, we investigated whether SHP-2 interacted with MyD88, the signaling molecule TRAF6, the kinase IRAK1 and other molecules involved in early events in TLR signaling. Among those, only TRAF6 interacted with SHP-2, a result confirmed by reverse coimmunoprecipitation (Fig. 8f). The carboxyl terminus of SHP-2 was responsible for its interaction with TRAF6 (Supplementary Fig. 7a). A luciferase reporter assay also showed that SHP-2 suppressed TRAF6-triggered activation of NF-B in a
npg
S Po LP ly S (I: C C ) pG
PB
PB
PB
Articles
Figure 8 Binding of SHP-2 to Fps and Shp2fl/fl Shp2/ 2.5 5 3 ** enhanced activation of SHP-2 are required +/+ / ** ** B2m B2m 2.0 ** 4 ** for negative regulation of TLR responses by ** ** 2 ** LPS (min) 0 15 30 45 60 0 15 30 45 60 1.5 3 ** MHC class I molecules. (a) Immunoblot 1.0 2 p-SHP-2 1 analysis of total and phosphorylated SHP-2 0.5 1 in lysates of B2m+/+ and B2m/ peritoneal SHP-2 0 0 0 macrophages stimulated for 060 min (above lanes) with LPS. (b) ELISA of TNF, IL-6 and IFN- in supernatants of macrophages 6 3 4 IgG H-2Kb with loxP-flanked alleles encoding SHP-2 5 (Ptpn11fl/fl; called Shp2fl/fl here) and 3 4 ** 2 ** ** ** SHP-2-deficient macrophages (Ptpn11/; * ** ** ** 3 2 called Shp2/ here) stimulated for 4 h ** 1 2 with PBS, LPS, poly(I:C) or CpG ODN. 1 1 (c) ELISA of TNF, IL-6, IFN- in supernatants 0 0 0 of macrophages of the genotypes in b, treated Shp2 fl/fl / fl/fl / fl/fl / Shp2 fl/fl / fl/fl / fl/fl / Shp2 fl/fl / fl/fl / fl/fl / b (for crosslinking with IgG or antiH-2K LPS Poly(I:C) CpG LPS Poly(I:C) CpG LPS Poly(I:C) CpG of MHC class I) and stimulated as in b. LPS (min) 15 0 5 15 LPS (min) 15 0 5 15 (d) Immunoblot analysis of phosphorylated Ctrl siRNA Fps siRNA IP: SHP-2 + + + IP: Fps + + + SHP-2 in wild-type macrophages transfected LPS (min) 0 15 30 45 60 0 15 30 45 60 IP: IgG + IP: IgG + with control or Fps-specific siRNA and then p-SHP-2 IB: SHP-2 IB: Fps stimulated for 060 min (above lanes) with IB: Fps IB: SHP-2 SHP-2 LPS. Total SHP-2 serves as loading control. (e) Immunoblot analysis of Fps and SHPShp2fl/fl Shp2/ 2 in lysates of macrophages stimulated for LPS (min) 0 15 30 60 0 15 30 60 (kDa) LPS (min) 15 0 15 30 60 LPS (min) 15 0 15 30 60 170 0, 5 or 15 min with LPS and subjected to IP: TRAF6 + + + + IP: SHP-2 + + + + 130 immunoprecipitation with IgG or antiSHP-2 IP: TRAF6 IP: IgG + IP: IgG + 95 (left) or anti-Fps (right). (f) Immunoblot IB: Ub IB: SHP-2 IB: TRAF6 72 analysis of TRAF6 and SHP-2 in lysates of IB: SHP-2 IB: TRAF6 LPS-stimulated macrophages subjected IP: TRAF6 to immunoprecipitation with IgG or IB: TRAF6 anti-TRAF6 (left) or antiSHP-2 (right). (g) Immunoblot analysis of lysates of LPS-stimulated macrophages (genotypes, as in b) subjected to immunoprecipitation with anti-TRAF6 and probed with anti-ubiquitin (Ub; above) or anti-TRAF6 (below). kDa, kilodaltons. *P < 0.05 and **P < 0.01 (Students t-test). Data are representative of three independent experiments with similar results (a,dg) or are from three independent experiments (b,c; mean s.e.m.).
S S ly (I: C ) C pG S L Po PS ly (I: C ) C pG
IFN- (ng/ml) TNF (ng/ml) IL-6 (ng/ml)
Po
dose-dependent manner (Supplementary Fig. 7b). As ubiquitination of TRAF6 is required in MyD88-dependent signaling, we further compared the ubiquitination of TRAF6 in wild-type and SHP-2-deficient macrophages. SHP-2 deficiency resulted in more LPS-induced polyubiquitination of TRAF6 (Fig. 8g), which indicated that SHP-2 suppressed the MyD88 pathway by interacting with TRAF6 and controlling its ubiquitination. Collectively, these data suggested that MHC class I molecules negatively regulated TLRtriggered inflammatory responses by enhancing activation of the FpsSHP-2 pathway (Supplementary Fig. 8). DISCUSSION Reverse signaling by members of the TNF family28 (including FasL, 4-1BBL, CD40L, LIGHT, membrane TNF and so on) and the B7 family29,30 (including CD80, CD86, B7-H1, B7-DC and so on) has been under extensive exploration. Reverse signaling mediated by MHC class I molecules on the cell surface has also been demonstrated1115, but there have been no reports on this signaling and its subsequent effects on myeloid cells. Here we have provided evidence that MHC class I molecules on macrophages stimulated with TLR ligands and/or after ligation with certain monoclonal antibodies may interact directly with Fps, maintaining its phosphorylation. Fps then recruits SHP-2, and SHP-2 participates at least partially in the suppression of TLR signaling by MHC class I molecules. Therefore, we have demonstrated a previously unknown nonclassical function for MHC class I molecules in which they are involved, via reverse signaling, in the negative regulation of TLR-triggered innate inflammatory responses. Adaptive T cells are able to temper innate immune responses; however, details of the mechanisms involved are unknown31. In the
clinical context, patients with type I bare lymphocyte syndrome (also known as TAP-deficiency syndrome) lack almost all surface expression of MHC class I molecules and CD8+ T cells32,33. These patients have high concentrations of C-reactive protein and a high erythrocyte-sedimentation rate during acute infection, which suggests they have enhanced innate inflammatory responses33. We indeed found that CD8+ T cells suppressed the macrophage response to TLRs both in vitro and in vivo via MHC class I molecules on macrophages. However, we did not obtain details about the exact ligands of MHC class I on CD8+ T cells, which may be CD8, TCR or one of the NK cellrelated inhibitory receptors. Furthermore, other ligands or receptors for MHC class I molecules expressed on monocytes and macrophages have been suggested, such as the inhibitory receptors ILT-2 and ILT-4 (ref. 34). During activation in vitro or in vivo, those MHC class I ligands, and even unknown ligands, may interact in cis or in trans with MHC class I molecules on the membrane, providing the reverse signal to MHC class I molecules, and, finally, may orchestrate the inhibitory effect of MHC class I on TLR signaling. Accumulating evidence suggests that a pool of MHC class I molecules present at the plasma membrane can dissociate from the light-chain 2m and open their structure, resulting in MHC class I molecules known as open conformers or misfolding conformers35. These are sparse in quiescent cells but increase considerably in abundance when the cells are metabolically activated. The open conformers are not stable and tend to associate with certain receptors or even with themselves, which is believed to be the basis for the reverse signaling of the biological effect of MHC class I molecules35. Notably, the cytoplasmic tail of MHC class I, which is somewhat longer than that of MHC class II, has one conserved tyrosine residue, which provides a
557
npg
IFN- (ng/ml)
TNF (ng/ml)
IL-6 (ng/ml)
Po
S ly (I: C ) C pG
PB
LP
PB
PB
LP
Articles
recognition site for the SH2 domain of kinases11,24. In addition, phosphorylation of the tyrosine residue in the cytoplasmic domain of MHC class I molecules is also related to the formation of open conformers and may influence or reflect interaction with certain molecules and the initiation of MHC class I reverse-signaling pathways24,35,36. Moreover, Fps can be activated after stimulation by certain cytokines, such as GM-CSF, IL-4, IL-6, erythropoietin and so on, and has been detected as being associated with phosphorylated receptors of these cytokines, including IL-4R, gp130 and the common -chain3739. Therefore, we predicted and have proven that phosphorylation of Tyr321 of intracellular MHC class I was necessary for its association with the SH2 domain of Fps and that this interaction may contribute to maintenance of Fps activation and the suppressive function of Fps on TLR-triggered responses40. Fps is a member of the non-receptor tyrosine kinase family, which has critical roles in regulating cytoskeletal rearrangements and the inside-out signaling involved in receptor-ligand, cell-matrix and cellcell interactions41,42. Fps is expressed in hematopoietic cells of the myeloid lineage, including macrophages, neutrophils and so on, and is essential for the survival and terminal differentiation of myeloidprogenitor cells. Mice deficient in Fps expression are hyper-responsive to endotoxin challenge23, although the underlying mechanisms have not been fully elucidated. Here we have shown that Fps associated with SHP-2 and that this association had a nonredundant role in the suppression of TLR responses by MHC class I molecules. SHP-2 is a non-receptor tyrosine phosphatase whose ubiquitous expression pattern is similar to that of MHC class I molecules 43. SHP-2 functions as a negative regulator in various signaling pathways through its phosphatase activity. Here, SHP-2 inhibited TLR signaling by binding to and controlling the activation of TRAF6 and TBK1 of the MyD88-dependent pathway and TRIF-dependent pathway27, respectively. In addition to SHP-2, Fps may also suppress TLR signaling via other mechanisms. It has been reported that Fpsdeficient macrophages have impaired internalization of TLR4 when stimulated with LPS44, which may provide another explanation for the Fps-mediated suppression. In conclusion, our results have demonstrated that membrane MHC class I molecules interacted with Fps and subsequently induced SHP-2 activation in response to stimulation with TLR ligands and, finally, suppressed innate inflammatory responses. Therefore, constitutively expressed MHC class I molecules are required for maintenance of the quiescent state and fine tuning of TLR-triggered innate inflammatory responses in macrophages. Our findings provide new insight into the negative regulation of TLR signaling and indicate a previously unidentified nonclassical function for MHC class I molecules in the regulation of innate inflammatory responses. METHODS Methods and any associated references are available in the online version of the paper.
Note: Supplementary information is available in the online version of the paper. ACknoWLedgmentS We thank G. Feng (University of California, San Diego) for mice with loxP-flanked alleles encoding SHP-2; H. Shen (University of Pennsylvania School of Medicine) for L. monocytogenes; N. van Rooijen (Free University of Amsterdam) for liposomes; J. Long, X. Zuo and P. Ma for technical assistance; and T. Chen and Y. Han for discussions. Supported by the National Key Basic Research Program of China (2012CB910202 and 2010CB911903), the National 125 Key Project (2012ZX10002-014 and 2012AA020901), the National Natural Science Foundation of China (81123006) and the Shanghai Committee of Science and Technology (10DZ1910300). AUtHoR ContRIBUtIonS X.C. and S.X. designed the experiments; S.X., X.L., Y.B., C.H., X.Zhu, P.Z., W.L. and X.Zha. did the experiments; X.C., S.X. and X.L. analyzed data and wrote the paper; and X.C. was responsible for research supervision, coordination and strategy. ComPetIng FInAnCIAL InteReStS The authors declare no competing financial interests.
1. Kawai, T. & Akira, S. The role of pattern-recognition receptors in innate immunity: update on Toll-like receptors. Nat. Immunol. 11, 373384 (2010). 2. Iwasaki, A. & Medzhitov, R. Regulation of adaptive immunity by the innate immune system. Science 327, 291295 (2010). 3. ONeill, L.A. & Bowie, A.G. The family of five: TIR-domain-containing adaptors in Toll-like receptor signalling. Nat. Rev. Immunol. 7, 353364 (2007). 4. Liew, F.Y., Xu, D., Brint, E.K. & ONeill, L.A. Negative regulation of toll-like receptormediated immune responses. Nat. Rev. Immunol. 5, 446458 (2005). 5. Antoniou, A.N., Powis, S.J. & Elliott, T. Assembly and export of MHC class I peptide ligands. Curr. Opin. Immunol. 15, 7581 (2003). 6. Raulet, D.H. MHC class I-deficient mice. Adv. Immunol. 55, 381421 (1994). 7. Nitta, T. et al. Thymoproteasome shapes immunocompetent repertoire of CD8+ T cells. Immunity 32, 2940 (2010). 8. Orr, M.T. & Lanier, L.L. Natural killer cell education and tolerance. Cell 142, 847856 (2010). 9. Jonsson, A.H. & Yokoyama, W.M. Natural killer cell tolerance licensing and other mechanisms. Adv. Immunol. 101, 2779 (2009). 10. Elliott, J.M., Wahle, J.A. & Yokoyama, W.M. MHC class I-deficient natural killer cells acquire a licensed phenotype after transfer into an MHC class I-sufficient environment. J. Exp. Med. 207, 20732079 (2010). 11. Tscherning, T. & Claesson, M.H. Signal transduction via MHC class-I molecules in T cells. Scand. J. Immunol. 39, 117121 (1994). 12. Skov, S. lntracellular signal transduction mediated by ligation of MHC class I molecules. Tissue Antigens 51, 215223 (1998). 13. Pedersen, A.E., Skov, S., Bregenholt, S., Ruhwald, M. & Claesson, M.H. Signal transduction by the major histocompatibility complex class I molecule. APMIS 107, 887895 (1999). 14. Yang, J. et al. Targeting 2-microglobulin for induction of tumor apoptosis in human hematological malignancies. Cancer Cell 10, 295307 (2006). 15. Zhang, X., Rozengurt, E. & Reed, E.F. HLA class I molecules partner with integrin 4 to stimulate endothelial cell proliferation and migration. Sci. Signal. 3, ra85 (2010). 16. Rubio, G. et al. Cross-linking of MHC class I molecules on human NK cells inhibits NK cell function, segregates MHC I from the NK cell synapse, and induces intracellular phosphotyrosines. J. Leukoc. Biol. 76, 116124 (2004). 17. Skov, S., Odum, N. & Claesson, M.H. MHC class I signaling in T cells leads to tyrosine kinase activity and PLC-1 phosphorylation. J. Immunol. 154, 11671176 (1995). 18. Pedersen, A.E., Bregenholt, S., Skov, S., Vrang, M.L. & Claesson, M.H. Protein tyrosine kinases p53/56lyn and p72syk in MHC class I-mediated signal transduction in B lymphoma cells. Exp. Cell Res. 240, 144150 (1998). 19. Liu, X. et al. Intracellular MHC class II molecules promote TLR-triggered innate immune responses by maintaining activation of the kinase Btk. Nat. Immunol. 12, 416424 (2011). 20. Han, C. et al. Integrin CD11b negatively regulates TLR-triggered inflammatory responses by activating Syk and promoting degradation of MyD88 and TRIF via Cbl-b. Nat. Immunol. 11, 734742 (2010). 21. Sun, J.C. & Lanier, L.L. Cutting edge: viral infection breaks NK cell tolerance to missing self. J. Immunol. 181, 74537457 (2008). 22. Perona-Wright, G. et al. Systemic but not local infections elicit immunosuppressive IL-10 production by natural killer cells. Cell Host Microbe 6, 503512 (2009). 23. Zirngibl, R.A., Senis, Y. & Greer, P.A. Enhanced endotoxin sensitivity in fps/fes-null mice with minimal defects in hematopoietic homeostasis. Mol. Cell. Biol. 22, 24722486 (2002). 24. Santos, S.G., Powis, S.J. & Arosa, F.A. Misfolding of major histocompatibility complex class I molecules in activated T cells allows cis-interactions with receptors and signaling molecules and is associated with tyrosine phosphorylation. J. Biol. Chem. 279, 5306253070 (2004). 25. Udell, C.M., Samayawardhena, L.A., Kawakami, Y., Kawakami, T. & Craig, A.W. Fer and Fps/Fes participate in a Lyn-dependent pathway from FcRI to plateletendothelial cell adhesion molecule 1 to limit mast cell activation. J. Biol. Chem. 281, 2094920957 (2006). 26. Rui, Y. et al. PECAM-1 ligation negatively regulates TLR4 signaling in macrophages. J. Immunol. 179, 73447351 (2007). 27. An, H. et al. SHP-2 phosphatase negatively regulates the TRIF adaptor proteindependent type I interferon and proinflammatory cytokine production. Immunity 25, 919928 (2006). 28. Sun, M. & Fink, P.J. A new class of reverse signaling costimulators belongs to the TNF family. J. Immunol. 179, 43074312 (2007). 29. Greenwald, R.J., Freeman, G.J. & Sharpe, A.H. The B7 family revisited. Annu. Rev. Immunol. 23, 515548 (2005).
npg
558
VOLUME 13
NUMBER 6
JUNE 2012
nature immunology
Articles
30. Chen, L. Co-inhibitory molecules of the B7CD28 family in the control of T-cell immunity. Nat. Rev. Immunol. 4, 336347 (2004). 31. Kim, K.D. et al. Adaptive immune cells temper initial innate responses. Nat. Med. 13, 12481252 (2007). 32. Zimmer, J. et al. Clinical and immunological aspects of HLA class I deficiency. QJM 98, 719727 (2005). 33. Gadola, S.D., Moins-Teisserenc, H.T., Trowsdale, J., Gross, W.L. & Cerundolo, V. TAP deficiency syndrome. Clin. Exp. Immunol. 121, 173178 (2000). 34. Colonna, M. et al. Human myelomonocytic cells express an inhibitory receptor for classical and nonclassical MHC class I molecules. J. Immunol. 160, 30963100 (1998). 35. Arosa, F.A., Santos, S.G. & Powis, S.J. Open conformers: the hidden face of MHC-I molecules. Trends Immunol. 28, 115123 (2007). 36. Little, A.M., Nossner, E. & Parham, P. Dissociation of 2-microglobulin from HLA class I heavy chains correlates with acquisition of epitopes in the cytoplasmic tail. J. Immunol. 154, 52055215 (1995). 37. Izuhara, K., Feldman, R.A., Greer, P. & Harada, N. Interaction of the c-fes proto-oncogene product with the interleukin-4 receptor. J. Biol. Chem. 269, 1862318629 (1994). 38. Matsuda, T. et al. Activation of Fes tyrosine kinase by gp130, an interleukin-6 family cytokine signal transducer, and their association. J. Biol. Chem. 270, 1103711039 (1995). 39. Brizzi, M.F. et al. Granulocyte-macrophage colony-stimulating factor stimulates JAK2 signaling pathway and rapidly activates p93fes, STAT1 p91, and STAT3 p92 in polymorphonuclear leukocytes. J. Biol. Chem. 271, 35623567 (1996). 40. Filippakopoulos, P. et al. Structural coupling of SH2-kinase domains links Fes and Abl substrate recognition and kinase activation. Cell 134, 793803 (2008). 41. Hellwig, S. & Smithgall, T.E. Structure and regulation of the c-Fes protein-tyrosine kinase. Front. Biosci. 17, 31463155 (2011). 42. Greer, P. Closing in on the biological functions of Fps/Fes and Fer. Nat. Rev. Mol. Cell Biol. 3, 278289 (2002). 43. Lorenz, U. SHP-1 and SHP-2 in T cells: two phosphatases functioning at many levels. Immunol. Rev. 228, 342359 (2009). 44. Parsons, S.A. & Greer, P.A. The Fps/Fes kinase regulates the inflammatory response to endotoxin through down-regulation of TLR4, NF-B activation, and TNF- secretion in macrophages. J. Leukoc. Biol. 80, 15221528 (2006).
npg
NUMBER 6
JUNE 2012
559
Mice and reagents. C57BL/6J mice were from Joint Ventures Sipper BK Experimental Animals. Mice deficient in 2m (B6.129P2-B2mtm1Unc/J; 002087), OT-I mice (Tg(TcraTcrb)1100Mjb/J; 003831) and Mx-Cre mice (B6.Cg-Tg(Mx1-cre)1Cgn/J; 003556) were from Jackson Laboratories. Mice deficient in H-2Kb and H-2Db were from Taconic Farms. Shp2fl mice (provided by G. Feng) were crossed with Mx-Cre mice. Shp2fl/fl mice and Mx-Cre Shp2fl/fl mice were injected intraperitoneally with 250 g poly(I:C) every other day for a total of five doses before use. All mice were bred in specific pathogenfree conditions. All animal experiments were in accordance with the National Institute of Health Guide for the Care and Use of Laboratory Animals, with the approval of the Scientific Investigation Board of the Second Military Medical University, Shanghai. LPS (from E. coli serotype 0111:B4), CpG ODN and poly(I:C) have been described19. E. coli serotype 0111:B4 was from the China Center for Type Culture Collection. L. monocytogenes was provided by H. Shen. Antibody to MHC class I (anti-H-2Kb; AF6-88.5) was from Biolegend. Anti-TLR4 (ab22048), anti-TLR3 (ab62566) and anti-TLR9 (ab52967) were from Abcam. Antibody to Erk phosphorylated at Thr202Tyr204 (E10), to Jnk phosphorylated at Thr183-Tyr185 (G9), to p38 phosphorylated at Thr180-Tyr182 (9211), to IRF3 phosphorylated at Ser396 (4D4G), to IB phosphorylated at Ser32-36 (5A5), to IKK-IKK phosphorylated at Ser176-Ser180 (16A6), to SHP-2 phosphorylated at Tyr580 (3703), to SHP-2 (D50F2) and to the Myc tag (2272) were from Cell Signaling Technology. Anti-Fps (sc-25415), anti-TRAF6 (sc-7221, sc-8409), anti-ubiquitin (P4D1), anti-IRAK1 (sc-5288) and anti--actin (sc-130656) were from Santa Cruz. Anti-Flag (M2) was from Sigma. Cell culture and RNA-mediated interference. Thioglycollate-elicited mouse peritoneal macrophages were prepared and cultured in endotoxin-free RPMI-1640 medium with 10% FCS (Invitrogen) as described19. The siRNA targeting MHC class I -chain or Fps was from Dharmacon. Mouse peritoneal macrophages were transfected with siRNA duplexes through the use of INTERFERin reagent (Polyplus) according to a standard protocol. Cytokine detection. TNF, IL-6 and IFN- in supernatants and serum were measured with ELISA kits (R&D Systems). RNA quantification. A LightCycler (Roche) and SYBR RT-PCR kit (Takara) were used for quantitative real-time RT-PCR analysis as described19. Data were normalized to -actin expression. Immunoprecipitation and immunoblot analysis. Cells were lysed with celllysis buffer (Cell Signaling Technology) supplemented with protease inhibitor cocktail (Calbiochem). Protein concentrations of the extracts were measured with a BCA assay (Pierce). The immunoprecipitation assays and immunoblot assays were done as described19,20. Nanospray liquid chromatographytandem mass spectrometry. Macrophages (3 108) were stimulated for 15 min with LPS with or without crosslinkage of MHC class I, then were lysed for preparation of immunoprecipitates with antiMHC class I. Proteins were eluted and digested. Digests were analyzed by nano-ultra-performance liquid chromatographyelectrospray ionization tandem mass spectrometry19. Data from liquid chromatography tandem mass spectrometry were processed through the use of ProteinLynx
ONLINE METHODS
Global Serverversion 2.4 (PLGS 2.4); the resulting peak lists were used for searching the NCBI protein database with the Mascot search engine. Cell isolation and in vivo depletion. CD8+ T cells and NK cells were sorted from C57BL/6 or OT-I mice with a MoFlo XDP (DakoCytomation). Naive CD8+ T cells were sorted as CD44loCD62Lhi and memory CD8+ T cells were sorted as CD44hiCD62Llo. Cell purity was >97%. For depletion of autologous CD8+ T cells or NK cells in vivo, mice were given monoclonal anti-CD8 (2.43; American Type Culture Collection) or anti-NK1.1 (PK136; American Type Culture Collection), respectively (each at a dose of 200 g per mouse), 3 d before infection. Depletion frequency was confirmed as being >90%. Confocal microscopy. Macrophages, plated on glass coverslips in six-well plates, were left unstimulated or stimulated with LPS and then labeled with antiMHC class I, anti-Fps, anti-TLR4, anti-TLR3, anti-TLR9 or anti-EEA1 (C45B10; Cell Signaling Technology). Cells were observed with a Leica TCS SP2 confocal laser microscope. Establishment of the endotoxin-shock model and bacterial infection. The endotoxin-shock mouse model was established by intraperitoneal injection of LPS (15 mg per kg body weight (mg/kg)) as described20. Serum TNF, IL-6 and IFN- were measured by ELISA in mice given intraperitoneal injection of LPS (5 mg/kg), poly(I:C) (20 mg/kg) or CpG ODN (20 mg/kg). For bacterial infection, E. coli serotype 0111:B4 and L. monocytogenes in mid-logarithmic growth were collected, counted on agar plates and then resuspended in PBS. Mice were given intraperitoneal injection of 1 107 E. coli or intravenous injection of 1 104 L. monocytogenes. Serum was collected for measurement of cytokines (or colony-forming units for E. coli) and spleens or livers were lysed for measurement of colony-forming units (L. monocytogenes) as described20,45. Bone marrow transplantation. Bone marrow cells (1 107) from B2m/ mice, KbDb/ mice or their littermates (control mice) were transplanted via tail-vein injection into lethally irradiated wild-type mice (10 Gy). After 8 weeks, CD8+ T cells in the spleen and lymph nodes were counted and MHC class I expression on macrophages was analyzed by flow cytometry. Macrophage reconstitution. Bone marrow cells from B2m/ mice, KbDb/ mice or their littermates (control mice) were cultured for 7 d in mouse macrophage colony-stimulating factor (50 ng/ml; PeproTech) for the preparation of bone marrowderived macrophages. Clodronate liposomes were injected intraperitoneally into wild-type recipient mice (50 mg in 200 l per mouse) for depletion of endogenous macrophages. Then 2 d later, B2m+/+ or B2m/ bone marrowderived macrophages (1 107) were transplanted into the recipient mice by tail-vein injection 6 h before challenge with LPS. Statistical analysis. The statistical significance of comparisons between two groups was determined with Students t-test. The statistical significance of survival curves was estimated according to the method of Kaplan and Meier, and the curves were compared with the generalized Wilcoxons test. P values of less than 0.05 were considered statistically significant.
45. Xu, S. et al. IL-17A-producing T cells promote CTL responses against Listeria monocytogenes infection by enhancing dendritic cell cross-presentation. J. Immunol. 185, 58795887 (2010).
npg
nature immunology
doi:10.1038/ni.2283
review
Genetic variation in Toll-like receptors and disease susceptibility

Mihai G Netea1, Cisca Wijmenga2 & Luke A J ONeill3
Toll-like receptors (TLRs) are key initiators of the innate immune response and promote adaptive immunity. Much has been learned about the role of TLRs in human immunity from studies linking TLR genetic variation with disease. First, monogenic disorders associated with complete deficiency in certain TLR pathways, such as MyD88-IRAK4 or TLR3-Unc93b-TRIF-TRAF3, have demonstrated the specific roles of these pathways in host defense against pyogenic bacteria and herpesviruses, respectively. Second, common polymorphisms in genes encoding several TLRs and associated genes have been associated with both infectious and autoimmune diseases. The study of genetic variation in TLRs in various populations combined with information on infection has demonstrated complex interaction between genetic variation in TLRs and environmental factors. This interaction explains the differences in the effect of TLR polymorphisms on susceptibility to infection and autoimmune disease in various populations. One of the most important concepts to revolutionize the understanding of host defense against pathogenic microorganisms that has emerged during the past 20 years is the recognition of patterns of microbial structures by dedicated germline-encoded receptors known as pattern-recognition receptors (PRRs). This was proposed by Charles Janeway in 1992 (ref. 1) and gained supporting experimental evidence from a seminal study by Lemaitre and colleagues, who showed that Drosophila fruit flies that lack the hematocyte receptor Toll, which indirectly recognizes pathogens through the cytokine-like protein Spaetzle, are highly susceptible to infection with fungi and Gram-positive bacteria2. That first report was followed shortly by the discovery of Tolllike receptors (TLRs) expressed on cells of the mammalian immune system. These receptors recognize evolutionarily conserved structures of microorganisms and activate an inflammatory response3. The homology between the intracellular domains of TLRs and that of the type 1 receptor for interleukin 1 (IL-1) has provided the key to understanding the function of TLRs in the activation of innate hostdefense mechanisms4. During infection, the host inflammatory reaction is initiated by the recognition by PRRs of conserved structures of the pathogenic microorganisms known as pathogen-associated molecular patterns (PAMPs). Five major classes of PRR have been described: the TLRs, the CLRs (C-type lectin receptors), the NLRs (nucleotide-binding domain, leucine-rich repeatcontaining receptors), the RLRs (RNA helicase RIG-Ilike receptors) and the ALRs (cytoplasmic DNA receptor
1Department of Internal Medicine and Nijmegen Institute for Infection, Inflammation and Immunity, Radboud University Nijmegen Medical Center, Nijmegen, The Netherlands. 2Department of Genetics, University Medical Center Groningen, University of Groningen, Groningen, The Netherlands. 3Trinity Biomedical Sciences Institute, School of Biochemistry and Immunology, Trinity College Dublin, The University of Dublin, Dublin, Ireland. Correspondence should be addressed to L.A.J.O. (laoneill@tcd.ie).
Published online 18 May 2012; doi:10.1038/ni.2284
AIM2like receptors)3,5. Among the families of PRRs, TLRs were the first to be described and have been studied most intensively. TLRs activate an acute inflammatory reaction after engaging with PAMPs from all the major classes of microorganisms, a reaction that represents the first line of innate host defense. Subsequently, stimulation via the TLR initiates and modulates the adaptive cellular and humoral immune responses6. More than 10 years of effort has identified the main signaling pathways activated by TLRs (Fig. 1). Signaling is initiated by adaptors that contain TollIL-1 receptor (TIR) domains. MyD88 is the universal adaptor, as it interacts with all the TLRs except TLR3, the receptor for double-stranded RNA. MyD88 also has a death domain, which recruits members of the IRAK (IL-1 receptorassociated kinase) family of serine-threonine kinases; this launches signaling pathways that culminate in the activation of transcription factors, most notably NF-B7. The structure of the multiprotein complex of the MyD88IRAK family has been solved, and this complex has been called the Myddosome8,9. Two Myddosomes have been characterized, defined by the presence of IRAK1 or IRAK2. In each, six MyD88 molecules assemble and interact with four IRAK4 molecules, which in turn interact with four IRAK1 molecules or four IRAK2 molecules. The interfaces between the components have all been solved structurally and thus detailed knowledge of the amino acids involved in the interactions is available. Notably, the Myddosome is also used by the receptors for IL-1, IL-18 and IL-33, which makes MyD88 especially important for inflammation and host defense. This is also true for the IRAK molecules, with IRAK4 being important for the activation of T cells by IL-1 (in cells of the TH17 subset of helper T cells) and most likely IL-18 (in cells of the TH1 subset of helper T cells)10. This broadens the role of this system into adaptive immunity. The other adaptors for the TLR system have more restricted uses. Mal (TIRAP) interacts with MyD88 and is required for signaling by the bacterial lipopeptide receptor TLR2 with a small amount of stimulation and is essential for signaling by the lipopolysaccharide
535
npg
NUMBER 6
JUNE 2012
review
TLR4, in which two amino acid changes (D299G and T399I) were reported to decrease the interaction of the receptor with lipopolysaccharide16 and to increase the susceptibility of patients to sepsis due to infection with Gram-negative bacteria 17. During the subsequent decade, a multitude of studies described genetic variation in practically all TLRs. The genetic variation in these receptors can be broadly described as resulting in complete functional deficiency that leads to primary immunodeficiency syndromes or having a variable effect on the function of the receptors; the latter has been studied in the context of diseases in case-control association studies.
TRIF
TLRs Triacyl lipopetides TLR1 TLR2 Diacyl lipopetides TLR6 Flagellin TLR5 LPS Mannans TLR4
Mal
IRAK4
IRAK1
NF-B
Proinflammatory cytokines
Type 1 interferons
NF-B
IRF3
TRIF
TRIF TRAM
Mal
TLR7,TLR8 ssRNA
TLR9 DNA
TLR3 dsRNA
LPS Mannans TLR4
Figure 1 TLRs, their main ligands and the intracellular pathways that lead to the stimulation of proinflammatory cytokines. ssRNA, single-stranded RNA; dsRNA, double-stranded RNA.
receptor TLR4 (ref. 11). Mal is also required for signaling by the receptor RAGE (receptor for advanced glycation end products), which has been linked to several inflammatory, degenerative and hyperproliferative diseases12. The adaptor TRAM is used only by TLR4 and interacts with the adaptor TRIF at endosomes, which leads to activation of the transcription factor IRF3. TRIF is also the sole adaptor used by TLR3. Finally, the fifth TIR domaincontaining adaptor, SARM, is inhibitory for TRIF-dependent signaling and thus limits signaling by TLR3 and TLR4 (ref. 13). Through these complex mechanisms, TLRs serve an important role in host defense against infection. In addition, as a result of their interactions with endogenous ligands, they are also involved in the pathophysiology of inflammatory and autoimmune diseases14. Genetic variation in TLRs Soon after the first description of TLRs, genetic variability in these molecules was proposed to result in differences in susceptibility to infectious and inflammatory diseases15. The first genetic variation to be described in TLRs (in the year 2000) were polymorphisms in
536
Primary immunodeficiency in the TLR pathways Complete defects in two main TLR-dependent pathways have been described so far. One defect leads to greater susceptibility to pyogenic bacteria (MyD88-IRAK4 deficiency) and the other results in greater susceptibility to herpesviruses (TLR3-Unc93bTRAF3 deficiency). Studies have identified patients with homozygous or compound heterozygous mutations in IRAK4 (refs. 1825) or MYD88 (refs. 26,27) that abolish protein production and result in a primary immunodeficiency syndrome characterized by greater susceptibility to pyogenic bacteria. Deficiency in MyD88-IRAK4 results in defective cell stimulation after engagement of TLRs and IL-1 receptors with subsequently impaired production of proinflammatory cytokines, followed by invasive and localized bacterial diseases. The invasive infections, such as meningitis, sepsis, arthritis or osteomyelitis, are caused by Streptococcus pneumoniae and Staphylococcus aureus and, less frequently, by Pseudomonas aeruginosa and Salmonella species. Localized bacterial diseases such as cellulitis, furunculosis and folliculitis are caused mainly by S. aureus, followed by P. aeruginosa and S. pneumoniae1828. So far, 15 different mutations have been identified in IRAK4, including insertions and deletions and missense, nonsense and splice-site mutations, and three different autosomal recessive mutations have been described in MYD88 deficiency in patients suffering from recurrent infection with pyogenic bacteria. Two of the mutations that lead to defective MyD88 (deletion of Glu52 or the substitution L93P) affect amino acids in key positions for the interaction between MyD88 and IRAK4 in the Myddosome8. A third mutation results in the substitution R98C, which is also in a key position for the so-called type 3 death domaindeath domain interactions found in the Myddosome structure29. Interestingly, the life-threatening infections in people with defective MyD88 first occur during in early infancy, as expected in a primary immunodeficiency, but they become less frequent and less severe during early adolescence, after which no life-threatening infections have been documented. Although the invasive infections in early childhood account for a cumulative mortality of 3040% (ref. 24), all adult patients have had a favorable clinical course without major infections and apparently with no prophylaxis needed28. This may indicate that the development of proficient adaptive immune responses (mediated by either T cells or B cells) later in life may compensate for the defects in the inflammatory reaction24, and if that phenomenon could be demonstrated, it would represent a major shift in the understanding of immune host defense. However, it should also be acknowledged that although MyD88-IRAK4 deficiency results in mortality of only 3040% in childhood, this prognosis is true only in advanced countries with well-developed intensive-care treatment facilities and in patients under protection with antibiotic treatment; the natural course of the disease would otherwise most likely be fully lethal. Although they are not related to primary immunodeficiency, striking examples of how somatic mutations in MYD88 can contribute to
MyD88 MyD88
TRAM
npg
review
human malignancies for both chronic lymphocytic leukemia (2.9%) and diffuse large B cell lymphoma (29%) have been reported in two studies30,31. Both identified the same amino acid substitution, L265P, in the TIR domain of MyD88. The L265P mutant spontaneously assembled and activated the Myddosome complex, leading to the constitutive production of many cytokines, including IL-6, that are integral to the disease pathogenesis. An IRAK4 inhibitor was selectively lethal for B cell lymphoma cells, presumably acting by inactivating the constitutively active Myddosome complex, which suggests that an inhibitor of IRAK4 could have utility in treatment of lymphoma. The recognition of viral pathogens by the innate immune system is mediated by receptors from the following two classes of intracellular PRRs: the RLR family, and several members of TLRs that recognize nucleic acids (TLR9, which recognizes unmethylated bacterial or fungal DNA; TLR7 and TLR8, which recognize single-stranded RNA; and TLR3, which recognizes double-stranded RNA) 32. Although no defects in receptors of the RLR family have been reported so far, patients with autosomal dominant33 or recessive34 missense mutations in TLR3, patients with autosomal dominant mutation of TRAF3 (ref. 35), patients autosomal recessive mutations in UNC93B1, which encodes a molecule in the TLR3 pathway36, and patients with autosomal recessive or dominant mutation of TRIF37 all present with a clinical syndrome characterized by recurrent herpesvirus encephalitis. This disease occurs mainly in early childhood, in children 3 months to 6 years of age, during primary infection with herpes simplex virus type 1 (refs. 28,38). Interestingly, this immunodeficiency syndrome leads to greater susceptibility only to herpesvirus encephalitis and to no diseases caused by other pathogens. The functional defect in patients with such deficiency in TLR3-UNC93B1-TRIF-TRAF3 is probably the result of less capacity to release type I interferons. Type I interferons have a crucial role in antiviral host defense39,40, and in vitro experiments have shown loss of TLR3-dependent induction of interferons in these patients33. Why patients with this immunodeficiency are susceptible only to herpesvirus remains a mystery. TLR polymorphisms: individual and population levels Although the complete deficiencies in the TLR pathways described above have a strong, and sometimes devastating, effect at the level of the individual people affected (with the notable exception of TLR5 deficiency, discussed below), they are generally rare events with a limited effect on the scale of an entire population. Consequently, those mutations with severely deleterious effects are under strong purifying selection and will never increase in frequency. In contrast, the genes encoding TLRs are extremely polymorphic and encode many variant amino acid sites. The underlying nucleotide variation within a species is compatible with purifying selection driven by pathogens41. Before genome-wide association studies, TLRs were considered excellent functional candidates for involvement in enhanced susceptibility to and severity of both infections15 and autoimmune and inflammatory diseases4244. Polymorphisms in all TLRs have been described, and a wealth of studies have reported their association with enhanced susceptibility to and severity of infections. Several excellent reviews have already described those association studies in detail 15,4547; therefore, we will focus mainly on discussing the functional and evolutionary implications of these polymorphisms. A key obstacle, however, to understanding the true role of TLR polymorphisms in enhanced susceptibility to infectious disease has been the lack of availability of large infection cohorts for genetic-associations studies and the lack of replication of associations among different studies and across populations. Several suggestive examples emphasize this point. For example, the TLR4 haplotype that results in glycine at position 299
(Gly299) and isoleucine at position 399 (Ile399) has been shown in two of four studies to be associated with a greater risk of sepsis48 and is suggested to result in a defective response to lipopolysaccharide16. However, various studies have failed to replicate those data at both genetic and functional levels49. The lack of replication might due to the small sample sizes, population stratification or the fact that the effect of this haplotype is restricted to certain subgroups of patients defined by criteria such as severity of disease or age of onset. Other polymorphisms associated with enhanced susceptibility to disease are those in the adaptor Mal (encoded by TIRAP), which is part of the TLR2 and TLR4 pathways. Single-nucleotide polymorphisms (SNPs) in TIRAP (resulting in the substitutions C558T and S180L) were initially shown to be associated with protection against tuberculosis50,51. However, meta-analysis of 6,584 patients with tuberculosis and 7,294 uninfected control subjects has not demonstrated any evidence of involvement of the S180L substitution in protection against susceptibility to tuberculosis52. Still, reports suggesting an association, particularly in South Indian and South African populations with tuberculosis meningitis, continue to appear53,54. A published review has nicely highlighted the problems faced in attempting to ascribe host genetics to enhanced susceptibility to tuberculosis, including differences in phenotype definition in both patients with tuberculosis and uninfected control subjects, consideration of latent versus active tuberculosis disease, population substructures and subsequent differences in linkage-disequilibrium patterns, and differences in the Mycobacterium tuberculosis strains causing the disease55. The S180L form of Mal has also been associated with malaria, pneumococcal disease and bacteraemia50, severe sepsis5658, Chagas cardiomyopathy59, systemic lupus erythematosus60, failure of the vaccine against Haemophilus influenza serotype b61 and Behets disease62. For unambiguous establishment of such associations, replication studies of larger populations are needed, given the relatively low odds ratios for most of them. The evidence of an association between the S180L variant of Mal and the failure of the vaccine seems to indicate a rather strong effect (odds ratio, 5.6; P = 1.2 107), possibly because of the tightly controlled nature of the study, which involved administering a vaccine to a defined population with careful followup. However, this study might also have been confounded by the typing of an indirect SNP and the low frequency of the risk allele. Clearly, many challenges remain in the analysis of genetic susceptibility to infectious diseases. The structure of Mal has been solved 63, and the Leu180 form has differences in structure compared with that of the Ser180 form. The amino acid at position 180 is very close to the aspartic acid at position 96 in Mal in the crystal structure, and substitution with asparagine at position 96 impairs the function of Mal29,64, which suggests that this region of Mal may be important for the functioning of Mal protein (Fig. 2). A final case that exemplifies the difficulties encountered in assessing the function of TLR polymorphisms is TLR5 deficiency. TLR5 recognizes flagellin, an important PAMP of flagellated bacteria65. A loss-of-function mutation in TLR5 (replacement of sequence encoding Arg392 with a stop codon) that results in a total defect in flagellin recognition has been described66. Interestingly, the mutated allele is present in approximately 10% of European populations and as much as 23% of other populations. People homozygous for this mutation are not characterized by severe primary immunodeficiency, although there are unconfirmed studies suggesting enhanced susceptibility to infections caused by Legionella pneumophila66 and to recurrent cystitis67. A protective effect of this mutation on systemic lupus erythematosus and Crohns disease has also been suggested68,69. The moderate to high frequency of the mutation resulting in a complete TLR5 defect
537
npg
review
Figure 2 The structure of the Myddosome and Mal provide information on the molecular basis of why variants in MyD88 and Mal are associated with disease. (a) The Myddosome has a stoichiometry of 14 and is composed of six MyD88, four IRAK4 and four IRAK2 molecules7. (b) Glu52 of MyD88 is at a key interface between MyD88 and IRAK4. Deletion of this amino acid in human leads to a greater risk of pyogenic infection and death in childhood and would disrupt the MyD88-IRAK4 complex. (c) The structure of Mal shows that Asp96 and Ser180 are in close proximity 63. The variant Leu180 is associated with many diseases, and Asp96 of Mal cannot associate with MyD88. Changes in these amino acids alter the structure of Mal in this region. Leucine at position 180 of Mal would cause steric occlusion of a cavity in Mal that could alter signaling, whereas asparagine at position 96 would alter the distribution of negative charge (red), which is probably key for the interaction with MyD88.
a
IRAK2
b
Next IRAK4
IRAK4
T41 R20
IIIa
F51 F36 D46 V16 E52
MyD88
Ia
E69
IIa
MyD88
Ib
T48 Y58 R62 I61
R12 N78
T76
V43 A44
L35
IIIb
D100 D99
E104 L103
IIb
in human populations, without a severe primary immunodeficiency phenotype, suggests a redundant role for TLR5 in host defense70. In conclusion, although many studies have suggested associations of TLR polymorphisms with disease processes, such conclusions should be interpreted with caution. Any of several factors may provide the basis for variation and discrepancies, such as the small sample sizes of many studies, population stratification (in particular in admixture populations) and restriction of the effect to certain subgroups of patients. In addition, all of those studies were of cohorts of limited sample size, and positive findings were often not replicated. Now it is possible to perform much more powerful genetic association studies and to control for the confounding factors that might have inflated some of the published results (such as population stratification). For example, genome-wide association studies have linked TLR7 and TLR8 to the genetic susceptibility to celiac disease (which involves an abnormal intestinal immune response to dietary gluten), one of the most common autoimmune disorders71. The advance of whole-genome and whole-exome sequencing is expected to identify many more such cases. Finally, one important but underestimated aspect of the influence of TLR polymorphisms on human diseases is heterogeneity in the frequency of polymorphisms and haplotypes among populations, often the result of past and present pressures exerted by local infections. Evolutionary pressure of infection A crucial aspect of the prevalence of polymorphisms in genes encoding TLRs and other genes related to immunity and their effect on susceptibility to infectious and autoimmune diseases is represented by the evolutionary processes that have influenced and shaped their spread in modern human populations. As discussed in detail below, several studies have suggested that certain polymorphisms in genes encoding TLRs in modern populations have resulted from positive selection through protection from infection. Subsequently, these studies may explain, at least in part, the differences among various populations in disease susceptibility. Classic examples of such differences are the lower susceptibility to malaria of populations in sub-Saharan Africa due to variants of the gene encoding hemoglobin 72 and the greater prevalence of autoimmune diseases in European populations than in African populations73. Systematic analyses of the evolutionary dynamics of TLRs in humans have reported the presence of selective forces that have shaped the evolution of this class of PRRs74,75. A first important observation was that all TLRs have undergone processes of purifying selection (that is, the elimination of gene variants with deleterious effects), albeit of different intensities. Interestingly, it seems that the intracellular TLRs (TLR3, TLR7, TLR8 and TLR9), which recognize mainly (viral) nucleic acids, have been under stronger purifying selection than have TLRs associated with the cell membrane. That has
538
c
S180 D96 N96 S180 D96 L180
led to speculation that either viral infections have exerted a stronger evolutionary pressure than have bacterial infections during the evolution of Homo sapiens or membrane TLRs are partially redundant74 and alternative recognition mechanisms are available to take over the activation of host defense when membrane TLRs are defective. Still, TLR4 and TLR1 are also under clear evolutionary pressure, and strong evidence of recent selection in the TLR1-TLR6-TLR10 cluster has become apparent, especially in European populations74,76, which indicates that membrane TLRs also have nonredundant roles in host defense. Those data are supported by a study identifying positive selection for several TLRs, as well as IRAK4, in both African and European populations77. Notably, members of the gene family encoding IL-1 cytokines (related to TLRs through the TIR domain) mostly show signs of balancing selection in the same populations77, and this is in line with a study linking the IL-1 family of cytokines to an abundance of pathogen species and a greater likelihood of developing a disease caused by parasites78. Studies suggesting an effect of selection pressures acting at the level of TLRs are also supported by a report investigating the evolution of TLRs in several primate species41. The authors concluded that distinct signatures of positive selection are present in most TLRs, more so in virus-sensing TLRs than in those that do not sense viruses. Moreover, the strongest evidence of positive selection is found in TLR1 and TLR4 (ref. 41), two of the genes identified as target of selection by the earlier studies74,76. A similar analysis has been made of the evolution of the TIR domaincontaining adaptors79. MyD88 and TRIF have been shown to have evolved under purifying selection, which indicates their role is essential and nonredundant for host survival. MyD88 is the least polymorphic adaptor, with the genes encoding Mal and TRIF having the greatest nucleotide diversity. Nonsynonymous polymorphisms in Mal are common, whereas they are present at low frequency in the other adaptors. Overall, however, there is evidence of more constraints on the adaptors than on the TLRs, presumably because there are fewer adaptors and therefore less redundancy. The stronger selecting constraint on MyD88 and TRIF than on plasma-membrane TLRs indicates that the integrity of these adaptors is very important for host defense. This is probably due to the nonredundant roles of all of the TLRs and receptors for IL-1 (with TLR4 using both, and TLR3 using TRIF). Adaptations of Mal in Europe, of TRIF in Europe and of TRAM in Asia, possibly as a result of infectious diseases, have been reported. The determinants of TLR selection are most probably infections that cause high morbidity and mortality in people before they reach reproductive age. It is beyond the scope of this review to discuss all
npg
review
Figure 3 The major routes of migration of modern humans after the movement out of Africa 100,000 years ago, and the worldwide geographical distribution of haplotypes of TLR4 and TIRAP (which encodes Mal). Average allele frequencies of the African, European, Asian and New World continents were adapted from refs. 56 and 80. Migration routes and dates are adapted from ref. 90. The TLR4 allele encoding Gly299 (299 allele) was selected in Africa because it provided protection against malaria but was negatively selected in Eurasia because it resulted in greater susceptibility to septic shock. Constraints imposed by the migration routes of Homo sapiens have influenced the absence of this allele in the Americas, despite a high prevalence of malaria. In contrast, the allele encoding Gly299 and Thr399 (399 allele) has most probably followed a variable distribution influenced mainly by genetic drift80. The distribution of the S180L polymorphism of Mal (180 allele) shows balanced evolution: the moderately greater cytokine production in people heterozygous for this allele protects them from susceptibility to several major diseases such as tuberculosis, malaria and pneumococcal pneumonia, whereas an overshoot in cytokine release in people homozygous for the allele encoding Leu180 is deleterious in septic shock. The distribution of the Mal S180L polymorphism is heavily skewed toward the Indo-European populations in which it most probably occurred.
Wild-type TIRAP TLR4 180 allele Wild-type 299 allele 399 allele 299 & 399 allele Europe Asia
15,00035,000
40,000
50,00060,000 100,000 >40,000 (50,00060,000?)
Africa
New World Wild-type 299 allele 399 allele TLR4 TIRAP 299 & 399 allele Wild-type 180 allele
the studies and data that have accumulated to support this hypothesis, but we will put forward a few powerful examples in support of this proposition. In addition, we should add that immunological research is now at a crossroads in the study of selection pressures on TLRs, as new databases such as the 1000 Genomes Project will prove to be an invaluable resource for such studies in the coming years, as this project will identify most genetic variants with a frequency of at least 1% in the 25 different and very diverse populations studied. TLR4 is one of the best-studied PRRs, and the distribution of TLR4 polymorphisms, as well as of the SNP that results in the S180L substitution of Mal, varies among different populations (Fig. 3). This is most probably because TLR4 with glycine at position 299 has protective effects against mortality due to Plasmodium falciparum cerebral malaria in Africa, where the TLR4 allele encoding Gly299 is highly prevalent (up to 15% in some populations)80. However, the deleterious effects of the product of this allele on the severity of Gram-negative sepsis prevented its fixation (that is, an increase in allele frequency to 100%) in Africa and may have been the selective force that led to nearly complete loss of this allele in Europe and Asia80. In addition, the SNP encoding the S180L substitution of Mal that results in an enhanced cytokine response is proposed to have been under balancing selection by protection against tuberculosis, malaria, pneumococcal infection50 and septic shock56 in heterozygous carriers. Similarly, it has been suggested that the positive selection of certain TLR1 alleles in Europeans is due to an attenuated inflammatory response, with potential beneficial effects during sepsis 74. Many other studies have reported important effects of polymorphisms of genes encoding receptors of the other PRR classes on susceptibility to infections, and comprehensive reviews have been published on this subject15,4547.
npg
Therefore, it is likely that selective forces exerted by pathogens have had an effect on the genetic makeup of genes encoding PRRs. However, it is also important to note that several studies published in the past few years have argued that nonadaptive mechanisms, such as genetic drift and geographical constraints, have also had an important role in the evolution of TLRs. Genetic drift and historical constraints In addition to selection, genetic drift is an important genetic process by which the frequencies of genetic variants both within and across populations change by chance alone. Processes such as population expansion and population bottlenecks can strongly influence the frequency of neutral variation, and it is very likely that such drift has influenced the frequency of at least some TLR polymorphisms (Fig. 3). For example, the frequency of the nonsense TLR5 polymorphism that results in replacement of the sequence encoding Arg392 with a stop codon ranges from 10% in Europeans to 23% in other populations70,74. The absence of major infectious complications in people who lack a functional TLR5 suggests that genetic drift, rather than adaptive processes, is the main mechanism behind the spread of this TLR5 variant, and this suggestion is supported by genetic studies74. Genetic drift is also the most likely cause of the variation of the TLR4 haplotype that encodes the D299G and T399I substitutions, which shows considerable variation in frequency among populations80. An obvious example of the role of geographical and historical constraints in influencing the genetic makeup of populations is the dependence of the genetics of one population on the genetic makeup of the population from which it originated. For example, the
539
review
indigenous populations of South America, who have been living in warm climates for a relatively short period of time (the past 18,000 years at most), originate from human populations who lived in the cold climates of North-East Asia for much longer before that (estimated at 35,000 years). The effects of human-migration routes on the TLR4 and TIRAP polymorphisms have been studied56,80. Nonadaptive constraints such as population history explain why the TLR4 allele encoding Gly299 has been lost in Trio Indians from the Amazon jungle, where malaria is highly prevalent, despite the protective effect of the product of this allele against malaria in sub-Saharan African. The protective phenotype has probably been lost during the 25,000-year immigration of Homo sapiens to the Americas across the cold climates of northeast Asia, where the TLR4 allele encoding Gly299 was deleterious because of greater mortality from septic shock and possibly respiratory syncytial virus infection (the travel of these changes is presented in Fig. 3). Thus, the TLR4 haplotype encoding Gly299 that may prove protective against malaria in the Amazon tropical forest is paradoxically missing because of selective pressures in the opposite direction undergone by the ancestors of the target population during their migration through cold climates. An additional substantial geographical constraint was encountered by European populations during the final glacial maximum 20,000 years ago, a period in which the advancement of the ice cap up to Central Europe forced the remaining scattered populations to retreat to isolated regions of the Iberian, Italian and Balkan peninsulas81. This period, characterized by geographical constraints due to the isolation of populations in southern refuges, may have contributed to the spread of the TLR4 haplotype encoding Gly299 and Ile399 in European populations from the Iberian refuge82. In conclusion, all of the data discussed above suggest that in addition to adaptation through natural selection, the role of genetic drift and geographical factors should also be considered in the genetic history of TLRs. Consequences of TLR genetic variability for modern societies There is a substantial body of evidence indicating that variation in genes encoding TLRs and molecules involved in the cellular pathways associated with these receptors has evolved under selection driven by pathogens and infections83. However, a multitude of studies that accumulated during the past decade also support the proposal of the involvement of TLRs in enhanced susceptibility to severe immunodeficiency disorders. There is not yet overwhelming evidence to link TLRs to widespread susceptibility to infectious diseases, inflammatory disorders and autoimmune processes, although studies in this area may also have been hampered by power issues. TLRs can bind endogenous ligands (such as heat-shock proteins and HMGB1) that function as danger-associated molecular patterns and can thus initiate sterile inflammation14. In this manner, the host-defense mechanisms that have beneficial effects during infection can exert the deleterious effects of exaggerated inflammatory or autoimmune reactions (Fig. 4). The observation that celiac disease is associated with a locus containing both TLR7 and TLR8 is notable in this context71. This is particularly important, as both TLR7 and TLR8 are the target of strong purifying selection, possibly induced by resistance to viral infections74, and a viral trigger has long been proposed for celiac disease and other autoimmune disorders (such as infection with rotavirus84). Infectious pressures tend to select genetic variants of PRRs that result in strong immune responses required for the elimination of pathogens, but such selection would also favor genetic variants that enhance the inflammatory response and hence may lead to greater susceptibility to inflammatory or autoimmune diseases. It is important
540
Infectious pressures Genetic drift and bottlenecks Migration routes
Adaptive evolution
Nonadaptive evolution
PRR polymorphisms
Resistance to infection
Autoimmunity
Figure 4 Adaptive evolution caused by infections, combined with nonadaptive evolution caused by genetic drift, population bottlenecks and migration routes, all contribute to PRR polymorphisms in various populations. These polymorphisms can give rise to resistance to infection via the sensing of PAMPs but also to a greater risk of autoimmunity, in which endogenous ligands acting via PRRs cause tissue damage and inflammation.
to realize that such responses may have beneficial effects early in life by protecting against childhood infections, whereas later in life, when autoimmune diseases become more prevalent, they may exert deleterious effects. Although this field of research is still in its infancy, important examples illustrate and support this proposition. A survey has shown that alleles and haplotypes related to innate immunity that are clearly compatible with positive selection, most probably due to a protective effect against infections, are in almost all cases associated with an greater risk of autoimmune disease85. That study is supported by data showing that pathogens lead to the selection of loci that confer both resistance to infection and, at the same time, susceptibility to autoimmune diseases78,83. For example, in celiac disease, recent positive selection has targeted several risk loci for celiac known to be associated with susceptibility to this disease83,86,87. One of these loci contains SH2B3, which encodes a protein with a structure reminiscent of that of cytokine-signalinginhibitory proteins (such as those of the SOCS family). The polymorphism of SH2B3 that results in the substitution R262W is associated with enhanced cytokine responses after stimulation of cells with bacterial peptidoglycans88. The SNP that encodes the R262W substitution is present at high prevalence in European populations but is almost absent in Asia and Africa. The sweep of the positive selection is estimated to have occurred around 1,2001,700 years ago88, coinciding with the collapse of the Roman Empire, during a period in which Europa was afflicted with the Justinian plague. Although the protein encoded by SH2B3 is not a member of the TLR family, its ability to modulate TLR signaling probably illustrates the general pattern of the effects exerted by genes encoding molecules of the immune system and most probably by TLRs as well. On the basis of the evidence presented above, it is rational to propose that the present high prevalence of certain alleles and haplotypes of genes encoding TLRs is the consequence not only of their protective effects against infections but also of nonadaptive genetic processes (for example, genetic drift and migration routes) of the past. In addition, extensive genetic variation is a driving force behind evolution and selection, as it allows modern humans to respond to a broad variety of infectious agents. In contrast, it might also predispose modern human populations to dysregulated immune responses that result in greater susceptibility to autoimmune and inflammatory diseases.
npg
review
Perspectives The two challenges that have emerged from the genetic analysis of the TLR system are the limited types of infections evident in primary immunodeficiencies and the low risk of infection in people bearing SNPs in a gene encoding one of the components of the TLR system, which in certain cases does not replicate across populations. Evolutionary analysis, however, strongly supports the proposal of a role for the TLR system in host defense, with redundancy in certain components possibly explaining the low risk of infection for particular single variants. However, such redundancy could be useful in the targeting of TLRs therapeutically for the treatment of autoimmune diseases or conditions, such as severe sepsis, or diseases of tissue injury, such as ischemia-reperfusion injury, conditions in which TLRs may have a less redundant role. For example, inhibiting TLR2 in heart ischemia has a clear benefit with no obvious signs of a greater risk of infection89. Similarly, inhibiting IRAK4 in B cell lymphomas that have mutant MyD88 could have considerable therapeutic benefits, possibly with a relatively limited risk of infection. Furthermore, if genetic variation can be used to predict severe sepsis or vaccine failure, testing for such variation could be useful for patient selection and personalized therapy or may help in the design of more efficient adjuvants. Further analysis of the functional consequences of genetic differences in the TLR system is needed to provide a better picture of what to target and for what indication.
ACkNOWLedGMeNts Supported by the Netherlands Organization for Scientific Research (M.G.N. and C.W.), Science Foundation Ireland (L.A.J.O.) and the European Research Council (L.A.J.O.). COMPetING FINANCIAL INteRests The authors declare no competing financial interests.
1. Janeway, C.A. Jr. The immune system evolved to discriminate infectious nonself from noninfectious self. Immunol. Today 13, 1116 (1992). 2. Lemaitre, B., Nicolas, E., Michaut, L., Reichhart, J.-M. & Hoffmann, J.A. The dorsoventral regulatory gene cassette Spaetzle/Toll/Cactus controls the potent antifungal response in Drosophila adults. Cell 86, 973983 (1996). 3. Takeuchi, O. & Akira, S. Pattern recognition receptors and inflammation. Cell 140, 805820 (2010). 4. Gay, N.J. & Keith, F.J. Drosophila Toll and IL-1 receptor. Nature 351, 355356 (1991). 5. Unterholzner, L. et al. IFI16 is an innate immune sensor for intracellular DNA. Nat. Immunol. 11, 9971004 (2010). 6. Iwasaki, A. & Medzhitov, R. Regulation of adaptive immunity by the innate immune system. Science 327, 291295 (2010). 7. ONeill, L.A. & Bowie, A.G. The family of five: TIR-domain-containing adaptors in Toll-like receptor signalling. Nat. Rev. Immunol. 7, 353364 (2007). 8. Lin, S.C., Lo, Y.C. & Wu, H. Helical assembly in the MyD88-IRAK4-IRAK2 complex in TLR/IL-1R signalling. Nature 465, 885890 (2010). 9. Gay, N.J., Gangloff, M. & ONeill, L.A. What the Myddosome structure tells us about the initiation of innate immunity. Trends Immunol. 32, 104109 (2011). 10. Staschke, K.A. et al. IRAK4 kinase activity is required for Th17 differentiation and Th17-mediated disease. J. Immunol. 183, 568577 (2009). 11. Kenny, E.F. et al. MyD88 adaptor-like is not essential for TLR2 signaling and inhibits signaling by TLR3. J. Immunol. 183, 36423651 (2009). 12. Sakaguchi, M. et al. TIRAP, an adaptor protein for TLR2/4, transduces a signal from RAGE phosphorylated upon ligand binding. PLoS ONE 6, e23132 (2011). 13. Carty, M. et al. The human adaptor SARM negatively regulates adaptor protein TRIFdependent Toll-like receptor signaling. Nat. Immunol. 7, 10741081 (2006). 14. Beutler, B. Microbe sensing, positive feedback loops, and the pathogenesis of inflammatory diseases. Immunol. Rev. 227, 248263 (2009). 15. Schrder, N.W. & Schumann, R.R. Single nucleotide polymorphisms of Toll-like receptors and susceptibility to infectious disease. Lancet Infect. Dis. 5, 156164 (2005). 16. Arbour, N.C. et al. TLR4 mutations are associated with endotoxin hyporesponsiveness in humans. Nat. Genet. 25, 187191 (2000). 17. Lorenz, E., Mira, J.P., Frees, K.L. & Schwartz, D.A. Relevance of mutations in the TLR4 receptor in patients with Gram-negative septic shock. Arch. Intern. Med. 162, 10281032 (2002). 18. Picard, C. et al. Pyogenic bacterial infections in humans with IRAK-4 deficiency. Science 299, 20762079 (2003). 19. Medvedev, A.E. et al. Distinct mutations in IRAK-4 confer hyporesponsiveness to lipopolysaccharide and interleukin-1 in a patient with recurrent bacterial infections. J. Exp. Med. 198, 521531 (2003). 20. Davidson, D.J. et al. IRAK-4 mutation (Q293X): rapid detection and characterization of defective post-transcriptional TLR/IL-1R responses in human myeloid and nonmyeloid cells. J. Immunol. 177, 82028211 (2006). 21. Cardenes, M. et al. Autosomal recessive interleukin-1 receptor-associated kinase 4 deficiency in fourth-degree relatives. J. Pediatr. 148, 549551 (2006). 22. Chapel, H., Puel, A., von Bernuth, H., Picard, C. & Casanova, J.L. Shigella sonnei meningitis due to interleukin-1 receptor-associated kinase-4 deficiency: first association with a primary immune deficiency. Clin. Infect. Dis. 40, 12271231 (2005). 23. Comeau, J.L. et al. Staphylococcal pericarditis, and liver and paratracheal abscesses as presentations in two new cases of interleukin-1 receptor associated kinase 4 deficiency. Pediatr. Infect. Dis. J. 27, 170174 (2008). 24. Ku, C.L. et al. Selective predisposition to bacterial infections in IRAK-4-deficient children: IRAK-4-dependent TLRs are otherwise redundant in protective immunity. J. Exp. Med. 204, 24072422 (2007). 25. Szab, J. et al. Recurrent infection with genetically identical pneumococcal isolates in a patient with interleukin-1 receptor-associated kinase-4 deficiency. J. Med. Microbiol. 56, 863865 (2007). 26. von Bernuth, H. et al. Pyogenic bacterial infections in humans with MyD88 deficiency. Science 321, 691696 (2008). 27. Picard, C. et al. Clinical features and outcome of patients with IRAK-4 and MyD88 deficiency. Medicine (Baltimore) 89, 403425 (2010). 28. Bousfiha, A. et al. Primary immunodeficiencies of protective immunity to primary infections. Clin. Immunol. 135, 204209 (2010). 29. George, J. et al. Two human MYD88 variants, S34Y and R98C, interfere with MyD88-IRAK4-myddosome assembly. J. Biol. Chem. 286, 13411353 (2011). 30. Puente, X.S. et al. Whole-genome sequencing identifies recurrent mutations in chronic lymphocytic leukaemia. Nature 475, 101105 (2011). 31. Ngo, V.N. et al. Oncogenically active MYD88 mutations in human lymphoma. Nature 470, 115119 (2011). 32. Akira, S. & Hemmi, H. Recognition of pathogen-associated molecular patterns by TLR family. Immunol. Lett. 85, 8595 (2003). 33. Zhang, S.Y. et al. TLR3 deficiency in patients with herpes simplex encephalitis. Science 317, 15221527 (2007). 34. Guo, Y. et al. Herpes simplex virus encephalitis in a patient with complete TLR3 deficiency: TLR3 is otherwise redundant in protective immunity. J. Exp. Med. 208, 20832098 (2011). 35. Prez de Diego, R. et al. Human TRAF3 adaptor molecule deficiency leads to impaired Toll-like receptor 3 response and susceptibility to herpes simplex encephalitis. Immunity 33, 400411 (2010). 36. Casrouge, A. et al. Herpes simplex virus encephalitis in human UNC-93B deficiency. Science 314, 308312 (2006). 37. Sancho-Shimizu, V. et al. Herpes simplex encephalitis in children with autosomal recessive and dominant TRIF deficiency. J. Clin. Invest. 121, 48894902 (2011). 38. De Tige, X., Rozenberg, F. & Heron, B. The spectrum of herpes simplex encephalitis in children. Eur. J. Paediatr. Neurol. 12, 7281 (2008). 39. Chapgier, A. et al. A partial form of recessive STAT1 deficiency in humans. J. Clin. Invest. 119, 15021514 (2009). 40. Dupuis, S. et al. Impaired response to interferon-alpha/beta and lethal viral disease in human STAT1 deficiency. Nat. Genet. 33, 388391 (2003). 41. Wlasiuk, G. & Nachman, M.W. Adaptation and constraint at Toll-like receptors in primates. Mol. Biol. Evol. 27, 21722186 (2010). 42. Radstake, T.R. et al. The Toll-like receptor 4 Asp299Gly functional variant is associated with decreased rheumatoid arthritis disease susceptibility but does not influence disease severity and/or outcome. Arthritis Rheum. 50, 9991001 (2004). 43. Tao, K. et al. Genetic variations of Toll-like receptor 9 predispose to systemic lupus erythematosus in Japanese population. Ann. Rheum. Dis. 66, 905909 (2007). 44. Kiechl, S. et al. Toll-like receptor 4 polymorphisms and atherogenesis. N. Engl. J. Med. 347, 185192 (2002). 45. Misch, E.A. & Hawn, T.R. Toll-like receptor polymorphisms and susceptibility to human disease. Clin. Sci. (Lond.) 114, 347360 (2008). 46. Brouwer, M.C. et al. Host genetic susceptibility to pneumococcal and meningococcal disease: a systematic review and meta-analysis. Lancet Infect. Dis. 9, 3144 (2009). 47. Texereau, J. et al. The importance of Toll-like receptor 2 polymorphisms in severe infections. Clin. Infect. Dis. 41 (suppl. 7), S408S415 (2005). 48. Casanova, J.L., Abel, L. & Quintana-Murci, L. Human TLRs and IL-1Rs in host defense: natural insights from evolutionary, epidemiological, and clinical genetics. Annu. Rev. Immunol. 29, 447491 (2011). 49. Ferwerda, B. et al. Functional consequences of toll-like receptor 4 polymorphisms. Mol. Med. 14, 346352 (2008). 50. Khor, C.C. et al. A Mal functional variant is associated with protection against invasive pneumococcal disease, bacteremia, malaria and tuberculosis. Nat. Genet. 39, 523528 (2007).
npg
NUMBER 6
JUNE 2012
541
review
51. Hawn, T.R. et al. A polymorphism in Toll-interleukin 1 receptor domain containing adaptor protein is associated with susceptibility to meningeal tuberculosis. J. Infect. Dis. 194, 11271134 (2006). 52. Miao, R., Li, J., Sun, Z., Xu, F. & Shen, H. Meta-analysis on the association of TIRAP S180L variant and tuberculosis susceptibility. Tuberculosis (Edinb.) 91, 268272 (2011). 53. Selvaraj, P., Harishankar, M., Singh, B., Jawahar, M.S. & Banurekha, V.V. Toll-like receptor and TIRAP gene polymorphisms in pulmonary tuberculosis patients of South India. Tuberculosis (Edinb.) 90, 306310 (2010). 54. Dissanayeke, S.R. et al. Polymorphic variation in TIRAP is not associated with susceptibility to childhood TB but may determine susceptibility to TBM in some ethnic groups. PLoS ONE 4, e6698 (2009). 55. Stein, C.M. Genetic epidemiology of tuberculosis susceptibility: impact of study design. PLoS Pathog. 7, e1001189 (2011). 56. Ferwerda, B. et al. Functional and genetic evidence that the Mal/TIRAP allele variant 180L has been selected by providing protection against septic shock. Proc. Natl. Acad. Sci. USA 106, 1027210277 (2009). 57. Hamann, L. et al. Low frequency of the TIRAP S180L polymorphism in Africa, and its potential role in malaria, sepsis, and leprosy. BMC Med. Genet. 10, 65 (2009). 58. Song, Z. et al. Genetic variants in the TIRAP gene are associated with increased risk of sepsis-associated acute lung injury. BMC Med. Genet. 11, 168 (2010). 59. Ramasawmy, R. et al. Heterozygosity for the S180L variant of MAL / TIRAP, a gene expressing an adaptor protein in the Toll-like receptor pathway, is associated with lower risk of developing chronic Chagas cardiomyopathy. J. Infect. Dis. 199, 18381845 (2009). 60. Castiblanco, J. et al. TIRAP (MAL) S180L polymorphism is a common protective factor against developing tuberculosis and systemic lupus erythematosus. Infect. Genet. Evol. 8, 541544 (2008). 61. Ladhani, S.N. et al. Association between single-nucleotide polymorphisms in Mal/TIRAP and interleukin-10 genes and susceptibility to invasive haemophilus influenzae serotype b infection in immunized children. Clin. Infect. Dis. 51, 761767 (2010). 62. Durrani, O. et al. TIRAP Ser180Leu polymorphism is associated with Behets disease. Rheumatology 50, 17601765 (2011). 63. Valkov, E. et al. Crystal structure of Toll-like receptor adaptor MAL/TIRAP reveals the molecular basis for signal transduction and disease protection. Proc. Natl. Acad. Sci. USA 108, 1487914884 (2011). 64. Nagpal, K. et al. Natural loss-of-function mutation of myeloid differentiation protein 88 disrupts its ability to form Myddosomes. J. Biol. Chem. 286, 1187511882 (2011). 65. Hayashi, F. et al. The innate immune response to bacterial flagellin is mediated by Toll-like receptor 5. Nature 410, 10991103 (2001). 66. Hawn, T.R. et al. A common dominant TLR5 stop codon polymorphism abolishes flagellin signaling and is associated with susceptibility to legionnaires disease. J. Exp. Med. 198, 15631572 (2003). 67. Hawn, T.R. et al. Toll-like receptor polymorphisms and susceptibility to urinary tract infections in adult women. PLoS ONE 4, e5990 (2009). 68. Hawn, T.R. et al. A stop codon polymorphism of Toll-like receptor 5 is associated with resistance to systemic lupus erythematosus. Proc. Natl. Acad. Sci. USA 102, 1059310597 (2005). 69. Gewirtz, A.T. et al. Dominant-negative TLR5 polymorphism reduces adaptive immune response to flagellin and negatively associates with Crohns disease. Am. J. Physiol. Gastrointest. Liver Physiol. 290, G1157G1163 (2006). 70. Wlasiuk, G., Khan, S., Switzer, W.M. & Nachman, M.W. A history of recurrent positive selection at the toll-like receptor 5 in primates. Mol. Biol. Evol. 26, 937949 (2009). 71. Dubois, P.C. et al. Multiple common variants for celiac disease influencing immune gene expression. Nat. Genet. 42, 295302 (2010). 72. Sirugo, G. et al. Genetic studies of African populations: an overview on disease susceptibility and response to vaccines and therapeutics. Hum. Genet. 123, 557598 (2008). 73. Kalla, A.A. & Tikly, M. Rheumatoid arthritis in the developing world. Best Pract. Res. Clin. Rheumatol. 17, 863875 (2003). 74. Barreiro, L.B. et al. Evolutionary dynamics of human Toll-like receptors and their different contributions to host defense. PLoS Genet. 5, e1000562 (2009). 75. Ferrer-Admetlla, A. et al. Balancing selection is the main force shaping the evolution of innate immunity genes. J. Immunol. 181, 13151322 (2008). 76. The Wellcome Trust Case-Control Consortium. Genome-wide association study of 14,000 cases of seven common diseases and 3,000 shared controls. Nature 447, 661678 (2007). 77. Casals, F. et al. Genetic adaptation of the antibacterial human innate immunity network. BMC Evol. Biol. 11, 202 (2011). 78. Fumagalli, M. et al. Parasites represent a major selective force for interleukin genes and shape the genetic predisposition to autoimmune conditions. J. Exp. Med. 206, 13951408 (2009). 79. Fornarino, S. et al. Evolution of the TIR domain-containing adaptors in humans: swinging between constraint and adaptation. Mol. Biol. Evol. 28, 30873097 (2011). 80. Ferwerda, B. et al. TLR4 polymorphisms, infectious diseases, and evolutionary pressure during migration of modern humans. Proc. Natl. Acad. Sci. USA 104, 1664516650 (2007). 81. DeGiorgio, M., Jakobsson, M. & Rosenberg, N.A. Out of Africa: modern human origins special feature: explaining worldwide patterns of human genetic variation using a coalescent-based serial founder model of migration outward from Africa. Proc. Natl. Acad. Sci. USA 106, 1605716062 (2009). 82. Plantinga, T.S. et al. The evolutionary history of TLR4 polymorphisms in Europe. J. Innate Immun 4, 168175 (2012). 83. Fumagalli, M. et al. Signatures of environmental genetic Adaptation pinpoint pathogens as the main selective pressure through human evolution. PLoS Genet. 7, e1002355 (2011). 84. Stene, L.C. et al. Rotavirus infection frequency and risk of celiac disease autoimmunity in early childhood: a longitudinal study. Am. J. Gastroenterol. 101, 23332340 (2006). 85. Di Rienzo, A. Population genetics models of common diseases. Curr. Opin. Genet. Dev. 16, 630636 (2006). 86. Barreiro, L.B. & Quintana-Murci, L. From evolutionary genetics to human immunology: how selection shapes host defence genes. Nat. Rev. Genet. 11, 1730 (2010). 87. Abadie, V., Sollid, L.M., Barreiro, L.B. & Jabri, B. Integration of genetic and immunological insights into a model of celiac disease pathogenesis. Annu. Rev. Immunol. 29, 493525 (2011). 88. Zhernakova, A. et al. Evolutionary and functional analysis of celiac risk loci reveals SH2B3 as a protective factor against bacterial infection. Am. J. Hum. Genet. 86, 970977 (2010). 89. Arslan, F. et al. Myocardial ischemia/reperfusion injury is mediated by leukocytic toll-like receptor-2 and reduced by systemic administration of a novel anti-toll-like receptor-2 antibody. Circulation 121, 8090 (2010). 90. Cavalli-Sforza, L.L. & Feldman, M.W. The application of molecular genetic approaches to the study of human evolution. Nat. Genet. 33 (suppl.), 266275 (2003).
npg
542
VOLUME 13
NUMBER 6
JUNE 2012
nature immunology
Articles
TRIM28 prevents autoinflammatory T cell development in vivo

Shunsuke Chikuma1, Naomasa Suita2, Il-Mi Okazaki3, Shiro Shibayama4 & Tasuku Honjo1
TRIM28 is a component of heterochromatin complexes whose function in the immune system is unknown. By studying mice with conditional T cellspecific deletion of TRIM28 (CKO mice), we found that TRIM28 was phosphorylated after stimulation via the T cell antigen receptor (TCR) and was involved in the global regulation of CD4 + T cells. The CKO mice had a spontaneous autoimmune phenotype that was due in part to early lymphopenia associated with a defect in the production of interleukin 2 (IL-2) as well as incomplete cell-cycle progression of their T cells. In addition, CKO T cells showed derepression of the cytokine TGF-b3, which resulted in an altered cytokine balance; this caused the accumulation of autoreactive cells of the T H17 subset of helper T cells and of Foxp3+ T cells. Notably, CKO Foxp3+ T cells were unable to prevent the autoimmune phenotype in vivo. Our results show critical roles for TRIM28 in both T cell activation and T cell tolerance. After antigen recognition, T cells produce interleukin 2 (IL-2) and subsequently proliferate. A series of protein phosphorylations transmits a signal from the T cell antigen receptor (TCR) to the nucleus, thereby inducing Il2 transcription. Although the pathways that lead to the initial Il2 transcription in T cells are well characterized 1, the mechanisms that control its long-term regulation are not. Such longterm gene regulation may include active chromatin modification by nuclear proteins. After clonal expansion, T cells differentiate into effector populations. In addition to the classical T helper type 1 (TH1) and TH2 cells, T cells that express the transcription factor Foxp3 (Foxp3+ T cells; known as regulatory T cells (Treg cells)) and IL-17-producing helper T cells (TH17 cells) have important roles in regulating tolerance versus immunity2. TH17 cells uniquely express the transcription factor RORt and are thought to be the main cells involved in many inflammatory diseases. In contrast, Foxp3 + T cells have many regulatory effects on other T cells. TH17 and Foxp3+ T cells, despite their different functions, share many aspects in common. Both arise in the thymus as a consequence of strong self interactions, which suggests that these T cells develop from a similar precursor35. They require the morphogen TGF- to develop from peripheral naive T cells6,7. TGF- can induce both Foxp3 and RORt in naive T cells, and further signaling by TGF- and IL-6 maintains the continuous expression of Foxp3 and RORt, respectively6,8. TH17 and Foxp3+ T cells are also similar in their phenotypic plasticity. Foxp3+ T cells lose Foxp3 expression under inflammatory conditions and redevelop into memory-like inflammatory T cells that produce interferon- (IFN-) and IL-17 (ref. 9). The transfer of Foxp3+ T cells into recipient mice deficient in the invariant signaling protein CD3 results in their reprogramming into follicular helper T cells, accompanied by the loss of Foxp3 expression10. After the loss of TGF-, fully committed TH17 can lose their ability to produce IL-17 (ref. 11). In general, it is likely that Foxp3+ T cells and TH17 cells develop by similar mechanisms that require TGF- and are relatively unstable, unlike classic TH1 and TH2 cells. Various factors, including IL-2 and TGF-, are known to influence the expression of Foxp3 and RORt. The tripartite-motif (TRIM) proteins, which bear RING, B-box and coiled-coil domains, constitute a family of ~60 molecules with diverse cell-type distributions, subcellular localizations and biological functions12. TRIM28 (also called KAP1 or TIF1-), TRIM24 (TIF-1) and TRIM33 (TIF1-) are nuclear TRIM proteins with pleckstrin-homology domains and bromo domains that interact with histones. TRIM28 has been identified as a gene-silencing factor for nuclear receptors that interacts with KRAB (Kruppel-associated box) domaincontaining zinc-finger transcriptional factors1315; these cofactors associate with heterochromatin proteins16 and recruit the histone methyltransferase SETDB1 to repress genes17. The function of TRIM28 in early development is relatively well understood. For example, it is critical for silencing of endogenous retrovirus in embryonic stem cells to maintain their integrity18,19. After its phosphorylation, TRIM28 mediates the repair of heterochromatic DNA2022. Furthermore, some TRIM proteins have been linked to antipathogen responses in cells of the innate immune system23,24. Although TRIM28 is a multifunctional protein, its intrinsic functions in T cells have not been characterized. T cells and B cells have abundant expression of TRIM28, and TRIM28 can interact with the cytidine deaminase AID in B cells25. We found that phosphorylation of TRIM28 was regulated by stimulation via the TCR. Genetic deletion of TRIM28 from T cells resulted in a profound loss of naive T cells associated with incomplete cell-cycle progression. Furthermore, we found that TRIM28 not only regulated T cell homeostasis
npg
1Department of Immunology and Genomic Medicine, Graduate School of Medicine, Kyoto University, Kyoto, Japan. 2Advanced Medicinal Research Laboratories, Tsukuba Research Institute, Ono Pharmaceutical, Tsukuba, Japan. 3Division of Immune Regulation, Institute for Genome Research, The University of Tokushima, Tsukuba, Japan. 4Exploratory Research Laboratories, Tsukuba Research Institute, Ono Pharmaceutical, Tsukuba, Japan. Correspondence should be addressed to T.H. (honjo@mfour.med.kyoto-u.ac.jp).
Received 3 October 2011; accepted 26 March 2012; published online 29 April 2012; doi:10.1038/ni.2293
596
VOLUME 13
NUMBER 6
JUNE 2012
nature immunology
Articles
a
Trim28 (relative)
15 10 5 0
+ U C S D 28
b Rest (h):
(kDa) 110 110 21 21
16
24 44 p-TRIM28(Ser473) Pan-TRIM28 p-TCR- Pan-TCR-

IM 28
Restim (h): (kDa) 110 110 110 21 21
2 p-TRIM28 (Ser473) p-Trim28 (Ser824) Pan-TRIM28 p-TCR- Pan-TCR-
(kDa) 110 50
si TR
si C
CD3 + CD28
UV
trl
TRIM28 Akt
e 1.5
IL-2 (ng/ml)
1.0 0.5 0
si siC TR t IM rl s 28 si iG SA 1 TB 1
Figure 1 TRIM28 is regulated by TCR signaling and regulates IL-2 production. (a) Quantitative real-time PCR analysis of Trim28 transcripts in purified wild-type mouse CD4+ T cells stimulated for 16 h with platebound anti-CD3 and anti-CD28 (CD3 + CD28); results are presented relative to those of cells left unstimulated (US). (b) Immunoblot analysis of TRIM28 phosphorylated at Ser473 (p-TRIM28(Ser473)), total TRIM28 (Pan-TRIM28), phosphorylated TCR -chain (p-TCR-) and total TCR -chain (Pan-TCR-) in extracts of purified wild-type mouse CD4+ T cells left untreated (Rest) for 044 h (above lanes). Left margin, molecular size in kilodations (kDa). (c) Immunoblot analysis of TRIM28 phosphorylated at Ser473 or Ser824, total TRIM28, and phosphorylated and total TCR -chain in extracts of purified wild-type mouse CD4+ T cells left untreated for 16 h, then restimulated (Restim; above lanes) for 04 h with anti-CD3 and anti-CD28 or for 2 h with ultraviolet irradiation (UV). (d) Immunoblot analysis of TRIM28 in Jurkat T cells treated with control siRNA (siCtrl) or siRNA specific for TRIM28 (siTRIM28); analysis of the irrelevant protein Akt serves as a loading control. (e) Enzyme-linked immunosorbent assay of IL-2 in Jurkat cells treated with control or siRNA (siCtrl) or siRNA specific for TRIM28 (siTRIM28) or for GFI1 (siGfi1) or SATB1 (siSATB1; negative controls). Data are representative of three (ac,e) or two (d) experiments.
but also interacted with pathways involved in the differentiation of Foxp3+ T cells and TH17 cells, which led us to propose previously unrecognized roles for TRIM28 in T cell homeostasis and tolerance. RESULTS Regulation of TRIM28 phosphorylation by TCR signaling TRIM28 is a phosphorylated protein present in T cells26. We first examined the overt expression and regulation of TRIM28 in primary mouse T cells. A database inquiry indicated that TRIM28 was expressed ubiquitously, with relatively high expression in lymphocytes (Supplementary Fig. 1). Accordingly, we found that TRIM28 was expressed constitutively in CD4+ mouse T cells at the level of mRNA and protein (Fig. 1ac) and that its transcription was further upregulated by TCR-mediated signals (Fig. 1a). TRIM28 is known to associate with heterochromatin proteins via a motif involving Ser473, whose phosphorylation by protein kinase C results in the dissociation of TRIM28 from heterochromatin proteins and the subsequent derepression of several genes27. Immunoblot analysis showed that TRIM28 was phosphorylated on Ser473 in freshly isolated CD4+ T cells (Fig. 1b). We wondered if this basal phosphorylation was due to TCR-mediated self recognition, which is thought to occur in vivo28. To investigate this possibility, we cultured sorted CD4+ T cells without TCR stimulation. Basal phosphorylation decreased after 816 h of culture, when phosphorylation of the TCR -chain had also decreased (Fig. 1b). We hypothesized that the phosphorylation of Ser473 was tightly controlled by external TCR signaling transduced from self and non-self antigens. In support of that idea, after T cells had been allowed to rest for 16 h, stimulation with plate-bound antibody to CD3 (anti-CD3) and anti-CD28 resulted in rapid rephosphorylation of TRIM28 on Ser473 (Fig. 1c). Another known TRIM28-phosphorylation site, Ser824, which is phosphorylated during DNA damage21,22, was not phosphorylated after stimulation via the TCR (Fig. 1c). Thus, TRIM28 was stimulated by TCR signaling at the level of transcription and by its phosphorylation on Ser473. TRIM28 regulates IL-2 production from Jurkat T cells We hypothesized that TRIM28 was involved in TCR-induced T cell responses; to investigate this, we knocked down TRIM28 in a Jurkat
human T lymphocyte cell line through the use of small interfering RNA (siRNA). Knockdown of endogenous TRIM28 to ~30% of its normal abundance resulted in much less production of IL-2 by Jurkat cells (Fig. 1d,e). We also observed this diminished IL-2 production in cells stimulated with the phorbol ester PMA and the calcium ionophore ionomycin (data not shown), which indicated that the defect was not at the TCR-proximal signals. In contrast to knockdown of TRIM28, knockdown of TRIM24 or TRIM33 did not affect IL-2 production (data not shown), which suggested a specific role for TRIM28 in regulating IL-2 production. These results suggested that TRIM28 was involved in the regulation of genes encoding molecules involved in T cell function. T cell activation in vitro requires TRIM28 Germline deletion of TRIM28 in mice results in embryonic death29. To examine the function of TRIM28 in T cells in vivo, we bred mice with loxP-flanked Trim28 alleles29 and mice with transgenic expression of Cre recombinase driven by the promoter of the gene encoding the kinase Lck30 to obtain mice with conditional T cellspecific deficiency in TRIM28 (called CKO mice here). As expected, we observed an almost complete loss of peripheral T cell compartments in CKO mice (Fig. 2a). This loss of TRIM28 did not result in an obvious defect in thymic development (Fig. 2b and Supplementary Fig. 2). Instead, we found fewer CD4+ or CD8+ T cells in the spleen, lymph nodes (Fig. 2b and Supplementary Fig. 2) and peripheral blood (data not shown). Sorted CD4+ T cells from CKO mice produced as much IL-2 as wild-type T cells did when cultured in plates coated with anti-CD3 and anti-CD28 for up to 18 h; this was followed by considerable decrease in IL-2 (Fig. 2c), in agreement with the data obtained with Jurkat T cells (Fig. 1d,e). CKO T cells also proliferated less than wild-type T cells did after stimulation with anti-CD3 and anti-CD28 (Fig. 2d,e). This was not due to more cell death, because we observed similar staining with the DNA-intercalating dye 7-AAD in wild-type and CKO T cells during culture (Supplementary Fig. 3). We wondered whether the defective IL-2 production caused less proliferation or whether the CKO cells had an intrinsic defect in cell-cycle progression. To investigate this, we stimulated T cells for 16 h via the TCR, then washed the cells and recultured them with exogenous IL-2. Exogenous IL-2 did not augment the proliferation of CKO T cells at any concentration used (Fig. 2f). Thus, the CKO cells were defective not only in IL-2 production but also in cell-cycle progression.
597
npg
Articles a
(kDa) 110 110 Naive MP Treg TRIM28 LSD1 WT
7.1 83.5
IL-2 (ng/ml)
WT CKO WT CKO WT CKO
1.5 1.0 0.5 0 0 20 40 60 CD3 + CD28 (h) 80 WT CKO
e
Cells
WT CKO
g
Aurkb Ccna2 Ccnb1 Ccnb1 Ccne1 Ccne1 Cdc2a Cdc45l Cdca5 Cdca8 Cdca8 Chek1 Chek1 E2f1 E2f1 Evi5 Exo1 Kif11 Kntc1 Nuf2 Sgol1
No stimulation WT CKO
CD3 + CD28 WT CKO
b
Thymus
CKO
5.0 83.1
CFSE
34.8
22.6
[3H]-thymidine uptake (104 c.p.m.)
3 [ H]-thymidine uptake (105 c.p.m.)
4.8
3.5
15 10 5
WT CKO
6 4 2 0 0 2 20
WT CKO
LN
24.1 19.4 12.6 8.2
CD4
Spleen
10.5
4.2
0 CD3 (g/ml) 0 CD28 (g/ml) 0
1 0
10 0
1 1
10 1
200 2,000
IL-2 (U/ml)
Fold change: 5.0 15 Ccne1
1.0
0.2
CD8
h
Ccna2
40
Figure 2 CKO mice show mild T lymphopenia, defective IL-2 production and defective cell-cycle 30 10 progression. (a) Immunoblot analysis of TRIM28 in CD4+ T cells from wild-type (WT) and CKO 20 mice, subsorted as naive cells (Naive; CD62LhiCD25), memory-phenotype cells (MP; CD62lo) and 5 10 Treg cells (Treg; CD62hiCD25+); below, blot reprobed for LSD1 (loading control). (b) Flow cytometry of thymus, lymph node (LN) and spleen cells from wild-type and CKO mice. Numbers in quadrants 0 0 WT CKO WT CKO WT CKO WT CKO indicate percent CD4+CD8 cells (top left), CD4+CD8+ cells (top right) or CD4CD8+ cells CD3 + CD28 (h): 0 14 0 14 (bottom right). (c) Enzyme-linked immunosorbent assay of IL-2 in supernatants of purified wild-type and CKO CD4+ T cells stimulated for 080 h (horizontal axis) with anti-CD3 and anti-CD28. (d,e) Thymidine incorporation (d) and dilution of the cytosolic dye CFSE (e) in purified naive wild-type and CKO T cells stimulated with various concentrations of anti-CD3 and anti-CD28 (horizontal axis, d) or for 3 d with anti-CD3, anti-CD28 and splenocytes (e). (f) Thymidine incorporation in purified naive wild-type and CKO T cells first primed with anti-CD3 and anti-CD28 and then cultured with various concentrations (horizontal axis) of IL-2. (g) Heat map of microarray data showing cell cyclerelated genes that changed differently in wild-type cells versus CKO cells after stimulation of the TCR with anti-CD3 and anti-CD28. (h) Quantitative RT-PCR analysis of Ccna2 (encoding cyclin A2) and Ccne1 (encoding cyclin E1) in wild-type and CKO T cells left unstimulated (0) or stimulated for 14 h with anti-CD3 and anti-CD28 (14); results are presented relative to those of unstimulated wild-type cells. Data are representative of two (a,e,f,h) or three (c,d) experiments (mean and s.d. of triplicates in d,h), at least ten experiments (b), or one experiment with four mice per group (g).
Upregulation of cell cyclerelated genes requires TRIM28 To obtain insight into the mechanisms underlying the phenotype described above, we used a microarray assay to compare transcripts from naive (CD62LhiCD25) and regulatory (CD62LhiCD25+) CD4+ T cells from CKO and wild-type mice. Many more genes were upregulated in the CKO samples, and these included many non-T cell genes (Supplementary Tables 1,2 and Supplementary Fig. 4), which supported the proposal that TRIM28 is a global repressor of gene regulation. In contrast, after activation with anti-CD3 and anti-CD28, naive CKO T cells failed to upregulate many genes encoding molecules involved in cell-cycle progression (Fig. 2g and Supplementary Table 2). Quantitative RT-PCR analysis showed a profound loss of cell cycle related genes in CKO T cells in which the TCR was stimulated (Fig. 2h), which suggested specific involvement of TRIM28 in the expression of genes encoding molecules involved in cell-cycle progression. Accumulation of autoreactive TH17 cells in CKO mice Despite the T lymphopenia in young mice (~6 weeks of age), as the mice matured (1622 weeks of age), we observed obvious organ enlargement accompanied by more cells in the lymph nodes and spleen. We found that this was due to the preferential population expansion of the CD62LloCD44hi memory-phenotype compartment (Fig. 3a,b). The sorted CD62LloCD44hi population from CKO mice had less production of the classic effector cytokines IFN- and IL-4 after in vitro stimulation than did wild-type cells (Fig. 3c). In contrast, the same cells secreted large amounts of the inflammatory cytokine IL-17. Memory-phenotype cells presumably develop from T cell experience of self or environmental antigens and increase in number with age31. We wondered if the memory-phenotype T cells accumulated because of expansion of the self-reactive population. To assess this
598
possibility, we analyzed sorted CD4+ cells by a syngeneic mixedlymphocyte reaction (Fig. 3d). These cells preferentially produced IL-17 in response to syngeneic splenocytes, which suggested that memory-phenotype T cells developed naturally in response to self antigen during TRIM28 deficiency in vivo. Spontaneous autoimmunity with TRIM28 deficiency We also found that serum from CKO mice contained more immunoglobulin than did that of control mice (data not shown). Serum from CKO mice reacted with many tissue-specific and ubiquitous antigens in tissue lysates from syngeneic mice (data not shown). These data suggested that the spontaneous activation of B cells by autoreactive T cells led to the autoantibodies in CKO mice. As might be expected, CKO mice died at a younger age than did their littermates (Fig. 3e). Histological examination of moribund mice showed the infiltration of multiple tissues by mononuclear cells (Supplementary Fig. 5), which suggested that the early death in these mice was caused by autoimmune-mediated tissue damage. TRIM28 deficiency accelerates experimental autoimmunity To examine how CKO mice responded to a defined antigen, we used a model of experimental autoimmune encephalomyelitis (EAE) in which a synthetic peptide that mimics myelin oligodendrocyte glycoprotein (MOG) induces autoreactive T cells. The EAE that developed in CKO mice was slightly delayed but significantly augmented relative to that of their littermates (Fig. 4a). The effector phase of EAE is characterized by an IL-17-mediated pathology. It was has been reported that adjuvant supplemented with heat-killed Mycobacterium tuberculosis strain H37Ra induces TGF- expression for the optimal initiation of TH17 differentiation during EAE32. We found that the
npg
Articles a
64.0 CD4+ gating
CD4+ naive cells (106)
+ 6 CD4 mem cells (10 )
WT 2.1 16.3
CKO 8.6
b
8 6 4 2 0
P = 0.016
c
8 6 4 2 0
P = 0.002
5 Mem/naive 4 3 2 1 0
P = 0.020
8 IFN- (ng/ml) IL-4 (ng/ml) 6 4 2 0 WT CKO
6 4 2 0 IL-17 (ng/ml) WT CKO 400 IFN- (pg/ml) 300 200 100 0 0
2.0 1.5 1.0 0.5 0 WT CKO
23.4 54.4 CD8 gating

+
10.5 13.2
20.5 32.5
54.6 14.1
W C T KO
W T C KO
W C T KO
CD8+ naive cells (106)
6 4 2 0
CD8+ mem cells (106)
P = 0.000079
1.5 1.0 0.5 0
NS
2 Mem/naive
P = 0.012
d
IL-17 (pg/ml)
400 300 200 100 0 0 3 6 Time in culture (d)
CD62L
WT CKO
28.6 CD44
3.9
41.4
12.1
W C T KO
W C T KO
W C T KO
Figure 3 Accumulation of autoreactive TH17 cells in TRIM28-deficient mice. (a) Flow cytometry of lymph node cells from 23-week-old wild-type and CKO mice. Numbers in or below quadrants indicate percent cells in each throughout. (b) Quantification of memory-phenotype (mem) cells (left) and ratio of memory cells to naive cells (Mem/naive; right) for CD4 + or CD8+ T cells among lymph node cells from 10- to 23-week-old wild-type mice (n = 8 per group) and CKO mice (n = 8 per group), assessed by flow cytometry as in a. NS, not significant (P = 0.09); P values, Students t-test. (c) Cytokine production by purified CD4+ cells sorted (beyond the sorting in b) for enrichment of memory-phenotype T cells (CD62LloCD44hi), then stimulated for 24 h with PMA and ionomycin. (d) Production of IL-17 and IFN- by wild-type and CKO cells in an autologous mixed-lymphocyte reaction. (e) Survival of CKO mice (n = 26) and their wild-type littermates (n = 14), presented as a Kaplan-Meier curve. P = 0.0004 (log-rank test). Data are representative of at least ten experiments (a), are pooled from four independent experiments (b; average s.d.) or are from two experiments (c,d).
e 1.0
3 6 Time in culture (d)
0.8 Survival 0.6 0.4 0.2 0 0 200 400 Time after birth (d) 600 WT CKO
use of a MOG emulsion without that strong adjuvant also induced severe EAE in CKO mice (Fig. 4b). In both experiments, CD4+ T cells from the immunized CKO mice produced more IL-17 and slightly less IFN- in response to the MOG peptide than did those from wildtype mice (Fig. 4c,d), which suggested that lack of TRIM28 resulted in the preferential new induction of pathogenic TH17 cells, which exacerbated EAE. Accumulation of Treg cells in TRIM28 deficiency IL-2 maintains the number of Foxp3+ T cells33 and constrains the development of TH17 cells34 in mice. We therefore wondered if the CKO phenotype simply reflected IL-2 deficiency. Unexpectedly, we found significantly more peripheral CD4+CD25+Foxp3+ T cells in CKO mice than in wild-type mice (Fig. 5a,b). We also found that the proportion of Foxp3+ T cells among CD4+ single-positive thymocytes was much greater in CKO mice, which suggested that both thymic induction and peripheral induction of Foxp3+ T cells were promoted in CKO mice. We wondered if the absence of TRIM28 would affect the
new induction of Foxp3+ T cells from naive CD4+ cells, which are typically induced by stimulation of the TCR in the presence of exogenous TGF-. Unexpectedly, we found that a substantial proportion of naive CKO T cells became Foxp3+ after stimulation of the TCR without the addition of TGF- (Fig. 5c and Supplementary Fig. 6a). Because CKO mice had fewer CD4+ cells, the greater abundance of Foxp3+ T cells may not have resulted from an intrinsic increase in the population. To clarify this point, we intercrossed mice with loxP-flanked Trim28 alleles and mice with transgenic expression of Cre recombinase linked to green fluorescent protein (GFP-Cre) driven by the Foxp3 promoter35 and generated mice with conditional deficiency of TRIM28 specifically in Foxp3+ T cells. In these mice, T cell development and the number of peripheral T cells were not affected (data not shown); however, we observed a much larger CD4+Foxp3+ compartment in the lymph nodes (Fig. 5d,e) and among peripheral blood T cells (Supplementary Fig. 6b). These observations indicated that TRIM28 negatively regulated the induction of Foxp3 as well as the number of Foxp3+ T cells in the periphery.
npg
Figure 4 TRIM28 deficiency exacerbates EAE, with enhanced induction of T H17 WT (n = 13) WT (n = 6) 4 5 8 NS * cells. (a,b) Clinical scores of wild-type 1,000 10 NS * 2.0 ** * * CKO (n = 10) CKO (n = 6) 4 and CKO mice immunized with MOG (a) or 800 8 3 6 1.5 with a MOG emulsion without M. tuberculosis 3 600 * 6 2 4 * (b). *P < 0.05, greater severity in CKO 1.0 2 400 4 mice than in wild-type mice (one-tailed 1 2 0.5 1 200 2 Students t-test). (c,d) Production of 0 0 0 0 0 0 IL-17 and IFN- by CD4 + T cells obtained 0 5 10 15 20 25 0 5 10 15 20 25 from wild-type and CKO mice immunized Time after immunization (d) Time after immunization (d) as in a (c) or b (d) and restimulated in vitro with MOG peptide on day 30, assessed by cytometric bead array. Numbers in parentheses (n) indicate number of mice. *P < 0.0040 and **P < 0.0010 (two-tailed Students t-test). Data are representative of two experiments (average and s.e.m. in a,b, and mean and s.d. in c,d).
Suboptimal EAE clinical score
EAE clinical score
IFN- (ng/ml)
IL-17 (ng/ml)
W T C (n KO = (n 7) = 6) W T C (n KO = (n 7) = 6)
NUMBER 6
JUNE 2012
W C T (n KO = (n 6) = 6) W T C (n KO = (n 6) = 6)
IFN- (ng/ml)
IL-17 (pg/ml)
599
Articles a
CD25
LN
WT 9.9
CKO 23.7 CD4 gate

+
CD25 cells + (% of CD4 LN cells)
Foxp3 cells + (% of CD4 LN cells)
30 20 10 0
Foxp3 cells (% of CD4SP thymocytes)
b
30 20 10 0 Flox 9.6
GFP-Cre
+
+
c
10
WT 0.0 6.4 0.2 0.0 6.0 93.4
CKO 0.0
US
CD4 2.1 7.9 7.7 16.1 CD4 gate 3.6 Foxp3 3.3 Thymus 8.0 CD4SP gate CD4 6.2
+
5 0.5
+ 94.0 CFSE gate
CD25
WT CKO WT
WT CKO
WT CKO Lymph node
Foxp3
***
**
CD3
0.0
11.9 81.7
0.7
98.6 CFSE 0.9
d
LN
15.7 CD4+ gate
e
GFP-Cre (%)
+
5.7
30 25 20 15 10
Foxp3
f
Change in body weight (relative)
1.9 1.7 1.5 1.3
No transfer (n = 10) Naive only (n = 11) Naive + WT Treg (n = 10) Naive + CKO Treg (n = 11)
CD25
5 * * 1.1 0 Figure 5 Aberrant induction of Foxp3+ T cells in TRIM28-deficient mice. * WT Flox * * 0.9 (a) Flow cytometry of lymph node cells and thymocytes from wild-type and CKO 0.7 mice. For Foxp3 staining, cells first stained for surface molecules were then fixed and made 0.5 permeable and then stained with anti-Foxp3. CD4SP, CD4+ single-positive. (b) Frequency of CD25+ 0 1 2 3 4 5 6 7 or Foxp3+ cells among lymph node CD4+ T cells (n = 10 mice per group) and of Foxp3+ cells among CD4+ Time after transfer (weeks) thymocytes (n = 9 mice per group), assessed by flow cytometry as in a. *P = 0.00008, **P = 0.000014 and + T cells loaded with ***P = 0.000002 (two-tailed Students t-test). (c) Flow cytometry of purified naive wild-type and CKO CD4 CFSE and then left unstimulated (US) or stimulated for 3 d with monoclonal anti-CD3 (CD3) and mitomycin Ctreated CD4 wild-type splenocytes, then stained for CD4 and intracellular Foxp3. (d) Flow cytometry of lymph node cells from wild-type mice and mice with conditional deficiency of TRIM28 specifically in Foxp3+ T cells (loxP-flanked Trim28 alleles and Foxp3-driven GFP-Cre; Flox). (e) Frequency of GFP+ cells among lymph node CD4+ cells from the mice in d. *P = 0.00038 (two-tailed Students t-test). (f,g) Weight loss (f) and colitis (g) in Rag2/ recipient mice given no cells (No transfer) or naive CD4+ T cells alone (Naive only) or together with wild-type Treg cells (Naive + WT Treg) or CKO Treg cells (Naive + CKO Treg) to induce experimental inflammatory bowel disease. *P < 0.05 (e), wild-type Treg cells versus CKO Treg cells (two-tailed t-test); *P = 0.00093 and **P = 0.00018 (f; two-tailed t-test). Each symbol (b,e) represents an individual mouse; small horizontal lines indicate the mean. Data are representative of more than five experiments (a,b), six experiments (c) or three experiments (d,e) or are from two independent experiments of four (f,g; error bars, s.e.m.).
g 12
10 8 6 4 2 0
* **
Histological score
o tr N ans ai fe ve r on + W Na ly T iv T e + C N reg KO a T ive
n: 6 8 9 9
A function for TRIM28 in Treg cells We assessed the function of the Foxp3+ T cells present in greater abundance in CKO mice. Sorted CD4+CD25hiCD62Lhi cells from CKO mice (~95% Foxp3+) and wild-type CD4+CD25hiCD62Lhi cells inhibited the in vitro proliferation of naive CD4+CD25CD62Lhi T cells in which the TCR was stimulated to a similar degree (data not shown). We assessed the in vivo activity of the CD4+CD25hiCD62Lhi cells with an established model of inflammatory bowel disease in which naive CD4+CD25CD62Lhi T cell populations depleted of Treg cells cause
Mouse: Cells:
WT chimera WT 65.5 Q1-1 7.3 Q2-1
CKO chimera CKO 15.6 Q1 7.9 Q2 28.6 Q1-1 WT
Mixed chimera CKO 6.6 Q2-1 11.9 Q1 6.7 Q2
CD62L
CD4+ gate
Q3-1 6.7 CD44 3.3 Q1-1
Q4-1 20.6 6.0 Q2-1
Q3 8.7 12.2 Q1
Q4 67.8 12.1 Q2
Q3-1 10.6 9.5 Q1-1
Q4-1 54.3 12.6 Q2-1
Q3 7.5 10.1 Q1
Q4 73.9 10.0 Q2
severe colitis in recipient mice deficient in recombination-activating gene 2 (Rag2/ mice); we found that 11 of 11 mice became moribund or were dead within 8 weeks (Fig. 5f,g). Mice that received naive cells together with wild-type Treg cells showed ameliorated disease (1 of 10 became moribund). In this setting, all mice (11 of 11) that received naive T cells together with CKO Treg cells developed severe colitis (Fig. 5g), lost weight (Fig. 5f) and became moribund. Thus, aberrantly expanded Foxp3+ T cell populations showed defects in their function as regulatory T cells in CKO mice. The defective function of Treg cells indicated that the functional defects of this population contributed to the spontaneous autoimmunity of TRIM28 deficiency. However, the mice with deletion of TRIM28 only in the Foxp3+ population did not develop any signs of autoimmunity, at least for 7 months after birth (data not shown), which suggested that the autoimmune phenotype required deletion of TRIM28 from the entire T cell compartment. Cell-extrinsic effects of TRIM28 deletion in T cells We further investigated whether the abundance of TH17 and Foxp3+ T cells was greater in CKO mice because of cell-autonomous mechanisms. For this, we generated chimeras by reconstituting lethally irradiated wild-type C57BL/6 recipient mice with wild-type or CKO bone marrow or a 1:1 mixture of both (mixed chimeras). In this
Figure 6 Cell-extrinsic promotion of TH17 cells and Foxp3+ T cells by TRIM28-deficient T cells. (a) Flow cytometry of splenocytes from -irradiated wild-type C57BL/6 mice reconstituted for 24 months with with T cell depleted bone marrow from CD45.1+ congenic wild-type mice (WT chimera) or CD45.2+ CKO mice (CKO chimera) or a mixture of bone marrow from mice of both genotypes (Mixed chimera). Q, quadrant. (b) Ratio of memory-phenotype cells to naive cells (left) and of CD45+Foxp3+ T cells among CD4+ T cells (right) from the mice in a. (c) IL-17 in supernatants of sorted CD4+ wild-type and CKO T cells stimulated for 2 d with plate-bound anti-CD3 and anti-CD28. Data are representative of two independent experiments (error bars (b,c), s.d.).
npg
Foxp3
CD4+ gate
Q3-1 CD25
Q4-1
Q3
Q4
Q3-1
Q4-1
Q3
Q4
b
Mem/naive
P = 0.008 6 5 4 3 2 1
+ CD25+Foxp3 T cells (%)
P = 0.022 25 20 15 10 5 0 WT CKO WT CKO WT CKO n: 4 4 Mix 4 P = 0.011
P = 0.0025
c 200
IL-17 (pg/ml)
P = 0.011 P = 0.011
150 100 50
0 Cells : WT CKO WT CKO Bone Mix marrow : WT CKO n: 4 5 4
0 Cells : WT CKO WT CKO Bone Mix marrow : WT CKO n: 4 5 4
600
VOLUME 13
NUMBER 6
JUNE 2012
nature immunology
re
Articles
Figure 7 Deregulation of TGF- expression in TRIM28 deficiency. (a) Heat map of the Tgfb3 signal from microarray data of naive wild-type and CKO T cells left unstimulated (Naive US) or stimulated with anti-CD3 and antiCD28 (Naive CD3 + CD28) or wild-type and CKO CD4+CD25hiCD62Lhi Treg cells (n = 4 mice per group). FC (fold change), difference in expression in wild-type versus CKO cells; FDR, false-discovery rate. (b) Quantitative RT-PCR analysis of Tgfb3, Tgfb1 and Tgfb2 in naive and memory wildtype and CKO T cells left unstimulated (Resting) or stimulated with anti-CD3 and anti-CD28 (CD3 + CD28); results are presented relative to those of resting, naive, wild-type cells. *P < 0.05 (two-tailed t-test). (c) Chromatin-immunoprecipitation assay of the abundance of TRIM28 (top), trimethylated H3K9 (H3K9me3) or acetylated H3K9 (H3K9ac) at the mouse Tgfb3 locus (positions and sequences of PCR primers, Supplementary Fig. 7); background values with no antibody were subtracted from the values of immunoprecipitation with specific antibodies, and the resultant values were normalized to the maximum value. Data are representative of one experiment with four mice per group (a) or at least two (b) or three (c) experiments (mean and s.d. of triplicate wells in b,c).
Tgfb3 expression (relative)
Naive US Naive CD3 + CD28 Treg
WT
CKO
Tgfb3 5.0 1.0 0.2 Expression (fold)
FC P FDR 8.4 0.00008 0.026 6.9 0.0006 0.027 9.4 0.0018 0.11
b
20 15 10 5 0
TRIM28 (normalized signal)
Exon1 100 50 0 Promoter Exon7
6 4 2 0
H3K9me3 (normalized signal)
A B CD E F GH I J K L MNOPQR S Exon1 WT CKO Exon7 100 Promoter 50 0
40 20 0
W C T KO W C T KO
* *
A B CD E F GH I J K L MNOPQR S Exon1 Exon7 WT CKO Promoter
H3K9ac (normalized signal)
100 50 0
setting, the recipients of CKO bone marrow showed a loss of naive T cells and an accumulation of memory-phenotype and Foxp3+ T cells, thus reproducing the CKO phenotype, but the recipients of wild-type bone marrow did not (Fig. 6a,b). Notably, in the mixed chimeras, we observed naive-to-memory transition of both wild-type and CKO T cells (Fig. 6a,b). We also observed a significantly larger Foxp3+ population among wild-type and CKO bone marrowderived T cells from the mixed chimeras (Fig. 6a,b). Separately sorted CD4+ T cells of each genotype in the mixed chimeras produced IL-17, which was not obvious in the mice reconstituted with wild-type bone marrow alone (Fig. 6c). Thus, a lack of TRIM28 in T cells can promote the differentiation of bystander wild-type T cells into TH17 cells and Foxp3+ T cells. Functional involvement of TGF-b in TRIM28 deficiency The cell-extrinsic phenotype of CKO T cells suggested derepression of soluble factor(s) in the absence of TRIM28. Among the genes upregulated in CKO T cells in the microarray, the Tgfb3 signal was much stronger in CKO T cells than in wild-type T cells after stimulation or in the resting state (Fig. 7a and Supplementary Table 2). TGF-3 uses the same receptor used by TGF-1 and transmits signals biologically similar to or even more active than those transmitted by
A B CD E F GH I J K L MNOPQR S
Naive Memory Naive Memory Resting CD3 + CD28
a
IL-17 (ng/ml)
9 8 7 6 5 4 3 2 1 0
An C An t t tiTG An i-TG rl A F- ti-T F- b An 1 G 1 ti- +T F- pa G 3 n- F- TG 3 F An C An ti- trl tiTG An TG Ab F ti- FAn -1 TG 1 ti- +T F- pa G 3 n- F- TG 3 F
b
120 100 IL-17 (pg/ml) 80 60 40 20 0 Anti-pan-TGF : + + + + Sorted cells : WT CKO WT CKO Bone marrow : WT CKO Mix
TGF-1 (ref. 36), which raised the possibility of its involvement in T cell biology. Quantitative RT-PCR analysis showed that T cells from CKO mice expressed Tgfb3 transcripts in the resting state (Fig. 7b). We also found that the abundance of TGF-1 and TGF-2 was greater in resting and stimulated CKO T cells, respectively, than in wild-type T cells. These data suggested that the TGF- family is a general target of TRIM28-mediated regulation. In chromatin-immunoprecipitation assays, treatment with antiTRIM28 resulted in enrichment for TRIM28 in multiple regions of conserved sequence in Tgfb3 locus from primary mouse T cells, with the strongest binding to the 3-proximal region of exon 1 (region N; Fig. 7c and Supplementary Fig. 7). TRIM28 is known to suppress gene expression by recruiting SETDB1 and histone deacetylases17. We found much less trimethylation of histone 3 Lys9 (H3K9; a repression mark) in region N in CKO T cells than in wild-type T cells (Fig. 7c). In contrast, we detected more acetylation of H3K9 in region N and the surrounding regions in CKO T cells than in wild-type T cells (Fig. 7c). These data suggested direct regulation of Tgfb3 expression by TRIM28. TGF- signaling has a role in TH17 cells6,7,11. Thus, we tested the functional relevance of the TGF- endogenously produced by CKO T cells. Blocking antibodies to members of the TGF- family inhibited the production of IL-17 by T cells from CKO mice (Fig. 8a) and sorted T cells derived from lethally irradiated wild-type recipient mice reconstituted with CKO bone marrow (Fig. 8b). Among the
Figure 8 Functional involvement of TGF- production in TRIM28 deficiency. (a) IL-17 in supernatants of sorted memory-phenotype wild-type and CKO CD4+ T cells stimulated for 2 d with plate-bound anti-CD3 and anti-CD28 in the presence of control antibody (Ctrl Ab) or various antibodies to TGF- (horizontal axis). (b) IL-17 in supernatants of sorted CD4+ T cells from bone marrow recipients (as in Fig. 6), stimulated for 2 d with platebound anti-CD3 and anti-CD28 in the presence of control antibody () or anti-pan-TGF- (+); short horizontal lines indicate the average. (c) Foxp3 expression in wild-type and CKO T cells stimulated in vitro (as in Fig. 5c), followed by the addition of control antibody or anti-pan-TGF- to the culture for 3 d, analyzed by flow cytometry. (d) Clinical scores of wild-type and CKO mice immunized with MOG, with rat immunoglobulin (control (Ctrl) Ig) or anti-pan-TGF- (Anti-TGF-; 60 g per mouse) in the peptide emulsion, for the induction of EAE. *P < 0.05, lower scores in CKO mice treated with anti-pan-TGF- than CKO mice treated with rat immunoglobulin (one-tailed Students t-test). Data are representative of three (a) or two (bd) experiments (average and s.e.m. in d).
npg
WT
CKO
c
Foxp3+ cells in CD4+ (%)
5 4 3 2 1 + CKO
d4
Clinical score
3 2 1 0
**
WT Ctrl lg (n = 9) WT Anti-TGF- (n = 7) CKO Ctrl lg (n = 7) CKO Anti-TGF- (n = 6)
0 Anti-TGF- : + WT
5 10 15 20 25 Time after immunization (d)
W T C KO W T C KO
NUMBER 6
JUNE 2012
601
Articles
antibodies, one able to block all three TGF- family members (antipan-TGF-) inhibited the secretion of IL-17 most efficiently (Fig. 8a), which suggested that all three TGF- members contributed to the continuous expression of IL-17 by CKO T cells. Anti-pan-TGF- also inhibited the aberrant induction of Foxp3+ T cells in vitro (Fig. 8c). These findings indicated that derepression of the family of TGF- cytokines in the absence of TRIM28 created a milieu in which TH17 cells and Foxp3+ T cells preferentially developed. Furthermore, we determined whether blockade of the TGF- signal inhibited the TH17 pathology. Including anti-pan-TGF- in the MOG peptide emulsion has been shown to block TGF- signals in the encephalitogenic T cells and to inhibit the development of EAE32. A suboptimal amount of anti-pan-TGF- significantly attenuated the EAE in CKO mice (Fig. 8d), which suggested that the accelerated EAE in CKO mice was dependent on the overproduced TGF-. Finally, we assessed whether supplementation with IL-2 rescued CKO mice from EAE. We treated mice with recombinant IL-2 in complex with monoclonal antibody to IL-2 before inducing EAE. This treatment ameliorated the severity of EAE in wild-type mice (Supplementary Fig. 8), as described37. The same treatment had little or no effect on CKO mice (Supplementary Fig. 8). These data suggested that CKO mice were defective not only in IL-2 production but also in the IL-2 response that promotes tolerance in vivo. DISCUSSION The most notable finding of our study here was that the genes affected by the loss of TRIM28 in T cells encoded molecules that precipitated an autoinflammatory phenotype in vivo. Although autoimmunity is polygenic, TRIM28 deficiency created a series of environments in which autoimmunity is likely to develop. First, the loss of TRIM28 caused an early T lymphopenia with fewer CD4+ T cells and CD8+ T cells (an abundance about 70% and 30%, respectively, that in wild-type mice). In such lymphopenic conditions, autoreactive T cell populations may preferentially expand as a consequence of altered homeostasis38. T lymphopenia was probably associated with the diminished cell-cycle progression of T cells in response to stimulation of the TCR and treatment with IL-2 in CKO mice. Microarray analysis of primary T cells in which the TCR was stimulated demonstrated an essential role for TRIM28 in upregulating many cell cyclerelated genes. CKO mice or a T cell line in which TRIM28 was knocked down had much less IL-2, a cytokine critical for T cell proliferation. Our findings indicated that TRIM28 is required for the optimal homeostatic proliferation that T cells need to maintain their numbers, which is triggered by TCRand cytokine-mediated signals. Next, deficiency in TRIM28 resulted in the derepression of TGF-, especially the isoform TGF-3. Overproduction of TGF- may be generally linked to autoimmunity. Autoimmune-prone MRL mice, in the absence39 or presence40 of an additional lymphoproliferation (lpr) mutation, naturally have high expression of TGF-1, and this may account for the TH17 type of pathology seen in these mice41. In humans, a high serum concentration of TGF-1 is found in patients with systemic lupus erythematosus42 and is associated with the severity of rheumatoid arthritis43. Transgenic mice that overexpress TGF- in T cells develop accelerated EAE6, whereas blockade of TGF- protects mice from EAE32. Such observations have been attributed to direct TGF- signaling to T cells, because the lack of TGF- signaling in mice with transgenic expression of a dominant negative TGF- receptor II results in the absence of TH17 cellmediated pathologies32. TGF- is also suggested to create a cytokine milieu for the development of TH17 cells and Foxp3+ T cells by suppressing TH1 and TH2 cells, which normally inhibit the expression of Foxp3 and IL-17. Subsequently, it has
602
been reported that cell-autonomous production of TGF- by activated T cells drives the development of TH17 cells44. TGF- is also important for the stability of TH17 cells11. Although those studies focused mainly on the immunologically well-characterized isoform TGF-1, our study has shown that TGF-3 may also have an important role in T cell immunity. A comprehensive study of all the TGF- isoforms associated with autoimmunity will be needed to elucidate their roles in self-tolerance and autoimmunity. Nonetheless, these observations are generally consistent with our conclusion that overexpression of TGF-3 due to the absence of TRIM28 contributed to the development and accumulation of TH17 cells. Thus, TRIM28 deficiency perturbed the balance of cytokines, which led to autoimmune-prone conditions. Overproduced TGF- may have also accounted for the accumulation of Foxp3+ T cells in CKO mice. However, despite their greater abundance, Foxp3+ T cells showed an obvious defect in their regulatory ability, especially in the development of inflammatory bowel disease. From these data, we speculate that a dysfunction in Foxp3+ T cells contributed, at least in part, to the spontaneous disease phenotype of CKO mice. Unexpectedly, we found that deletion of TRIM28 only in the Foxp3+ T cell population did not cause the systemic autoimmunity observed in CKO mice. A possible explanation for this discrepancy could be the difference in the in vivo environment in which Treg cells developed in CKO mice and in mice with deletion of TRIM28 only in the Foxp3+ T cell population. Treg cells in CKO may receive considerably unbalanced cytokine stimulation from TRIM28-deficient T cells, which would result in loss of their suppressive function. In contrast, Foxp3+ T cellspecific deletion of TRIM28 would not provide such environmental effects to Treg cells. Sorted CD4+CD25hiCD62Lhi cells from young mice in the assay of inflammatory bowel disease were ~95% Foxp3+, which excluded the possibility of massive contamination by nonregulatory Foxp3CD25+ cells. However, this did not rule out the possibility that Treg cells from the inflammatory environment had already changed their regulatory ability. It is also possible that early deletion of TRIM28 by the construct of Cre driven by the Lck promoter (in CKO mice) caused critical defects in some important epigenetic modifications of Treg cell signature genes, whereas deletion of TRIM28 at a later stage by the construct of Foxp3-driven GFP-Cre resulted in only mild defects in the Treg cell phenotype. In short, Treg cells derived from CKO mice were defective either cell autonomously or environmentally. It will be necessary to determine whether the autoimmune phenotype was due solely to TRIM28 deficiency in effector cells or was due to a secondary effect of cytokines from TRIM28-deficient effector T cells on the Treg cell phenotype. However, in preliminary studies, we found that transfer of CKO effector cells into Rag2/ mice did not cause inflammatory bowel disease as effectively as transfer of wild-type effector cells did, probably because of poor proliferation of CKO effector cells in vivo (data not shown). At this stage, we propose that the autoimmune phenotype of CKO mice was caused by a combination of T cell abnormality and Treg cell insufficiency. It is notable that Foxp3 works as both a transcriptional activator and a transcriptional repressor for key genes in Treg cells. Genes activated by Foxp3 include those encoding the immunomodulatory receptor CTLA-4 (CD152) and the cytokine receptor CD25, whereas genes repressed include those encoding IL-2 and IFN-. It is unlikely that TRIM28 works as a corepressor for Foxp3, as we did not find that those Foxp3 signature genes were affected much in TRIM28-deficient T cells. In addition, we did not observe direct molecular interaction between Foxp3 and TRIM28 in preliminary overexpression experiments (data not shown). However, it is possible that TRIM28 works with gene-regulatory factors that are downstream of Foxp3 (such as SATB1)45 to ensure the full function of Treg cells. The biochemical
npg
Articles
interactions of TRIM28 with known transcriptional factors in T cells require further investigation. In conclusion, our study has demonstrated crucial functions for TRIM28 in the regulation of T cell homeostasis and tolerance. IL-2, cell-cycle proteins and TGF- are targets of TRIM28 whose deregulation contributed to autoinflammatory T cell development. Our findings shed light on the function of the TRIM family of proteins in the acquired immune system and may help to elucidate the complex pathology that underlies the development and progression of autoimmune disease. METHODS Methods and any associated references are available in the online version of the paper. Accession codes. GEO: raw and normalized microarray data, GSE 32224.
Note: Supplementary information is available in the online version of the paper. ACkNOwledgMeNTS We thank X. Zhou, A. Shimizu, S. Bailey-Bucktrout, K. Ikuta, T. Eagar, K. Ogasawara and M. Aida for discussions; K. Yurimoto for assistance in mouse genotyping; P. Chambon and R. Losson (Institut de Gntique et de Biologie Molculaire et Cellulaire, France) for mice with loxP-flanked Trim28 alleles; J. Takeda (Osaka University) for mice with Cre expression driven by the Lck promoter; J. Bluestone (University of California-San Francisco) for Foxp3GFP-Cretransgenic mice; S. Nagata (Kyoto University) for CD45.1+ mice; K. Kabashima (Kyoto University) for Rag2/ mice; and J. Bluestone and N. Minato for critical reading of the manuscript. Supported by the Ministry of Education, Culture, Sports, Science and Technology of Japan (KAKENHI Grant-in-Aid for Scientific Research Young Scientist (B) 21790465 and 23790534 to S.C.) and the Program for the Promotion of Fundamental Studies in Health Sciences of the National Institute of Biomedical Innovation (T.H.). AUTHOR CONTRIBUTIONS S.C. and T.H. designed the research; S.C. did experiments; S.C., N.S. and S.S. did microarrays and analyzed the data; I.-M.O. provided materials essential to the research; and S.C. and T.H. wrote the paper. COMPeTINg FINANCIAl INTeReSTS The authors declare competing financial interests: details are available in the online version of the paper.
1. Cantrell, D. T cell antigen receptor signal transduction pathways. Annu. Rev. Immunol. 14, 259274 (1996). 2. Weaver, C.T., Harrington, L.E., Mangan, P.R., Gavrieli, M. & Murphy, K.M. Th17: an effector CD4 T cell lineage with regulatory T cell ties. Immunity 24, 677688 (2006). 3. Jordan, M.S. et al. Thymic selection of CD4+CD25+ regulatory T cells induced by an agonist self-peptide. Nat. Immunol. 2, 301306 (2001). 4. Apostolou, I., Sarukhan, A., Klein, L. & von Boehmer, H. Origin of regulatory T cells with known specificity for antigen. Nat. Immunol. 3, 756763 (2002). 5. Marks, B.R. et al. Thymic self-reactivity selects natural interleukin 17-producing T cells that can regulate peripheral inflammation. Nat. Immunol. 10, 11251132 (2009). 6. Bettelli, E. et al. Reciprocal developmental pathways for the generation of pathogenic effector TH17 and regulatory T cells. Nature 441, 235238 (2006). 7. Mangan, P.R. et al. Transforming growth factor- induces development of the TH17 lineage. Nature 441, 231234 (2006). 8. Veldhoen, M., Hocking, R.J., Atkins, C.J., Locksley, R.M. & Stockinger, B. TGF in the context of an inflammatory cytokine milieu supports de novo differentiation of IL-17-producing T cells. Immunity 24, 179189 (2006). 9. Zhou, X. et al. Instability of the transcription factor Foxp3 leads to the generation of pathogenic memory T cells in vivo. Nat. Immunol. 10, 10001007 (2009). 10. Tsuji, M. et al. Preferential generation of follicular B helper T cells from Foxp3+ T cells in gut Peyers patches. Science 323, 14881492 (2009). 11. Lee, Y.K. et al. Late developmental plasticity in the T helper 17 lineage. Immunity 30, 92107 (2009). 12. Reymond, A. et al. The tripartite motif family identifies cell compartments. EMBO J. 20, 21402151 (2001). 13. Le Douarin, B. et al. A possible involvement of TIF1 and TIF1 in the epigenetic control of transcription by nuclear receptors. EMBO J. 15, 67016715 (1996). 14. Friedman, J.R. et al. KAP-1, a novel corepressor for the highly conserved KRAB repression domain. Genes Dev. 10, 20672078 (1996). 15. Moosmann, P., Georgiev, O., Le Douarin, B., Bourquin, J.P. & Schaffner, W. Transcriptional repression by RING finger protein TIF1 that interacts with the KRAB repressor domain of KOX1. Nucleic Acids Res. 24, 48594867 (1996). 16. Nielsen, A.L. et al. Interaction with members of the heterochromatin protein 1 (HP1) family and histone deacetylation are differentially involved in transcriptional silencing by members of the TIF1 family. EMBO J. 18, 63856395 (1999). 17. Schultz, D.C., Ayyanathan, K., Negorev, D., Maul, G.G. & Rauscher, F.J. 3rd SETDB1: a novel KAP-1-associated histone H3, lysine 9-specific methyltransferase that contributes to HP1-mediated silencing of euchromatic genes by KRAB zincfinger proteins. Genes Dev. 16, 919932 (2002). 18. Wolf, D. & Goff, S.P. TRIM28 mediates primer binding site-targeted silencing of murine leukemia virus in embryonic cells. Cell 131, 4657 (2007). 19. Rowe, H.M. et al. KAP1 controls endogenous retroviruses in embryonic stem cells. Nature 463, 237240 (2010). 20. White, D.E. et al. KAP1, a novel substrate for PIKK family members, colocalizes with numerous damage response factors at DNA lesions. Cancer Res. 66, 1159411599 (2006). 21. Noon, A.T. et al. 53BP1-dependent robust localized KAP-1 phosphorylation is essential for heterochromatic DNA double-strand break repair. Nat. Cell Biol. 12, 177184 (2010). 22. Ziv, Y. et al. Chromatin relaxation in response to DNA double-strand breaks is modulated by a novel ATM- and KAP-1 dependent pathway. Nat. Cell Biol. 8, 870876 (2006). 23. Gack, M.U. et al. TRIM25 RING-finger E3 ubiquitin ligase is essential for RIG-I-mediated antiviral activity. Nature 446, 916920 (2007). 24. Tsuchida, T. et al. The Ubiquitin Ligase TRIM56 Regulates Innate Immune Responses to Intracellular Double-Stranded DNA. Immunity 33, 765776 (2010). 25. Okazaki, I.M. et al. Histone chaperone Spt6 is required for class switch recombination but not somatic hypermutation. Proc. Natl. Acad. Sci. USA 108, 79207925 (2011). 26. Navarro, M.N., Goebel, J., Feijoo-Carnero, C., Morrice, N. & Cantrell, D.A. Phosphoproteomic analysis reveals an intrinsic pathway for the regulation of histone deacetylase 7 that controls the function of cytotoxic T lymphocytes. Nat. Immunol. 12, 352361 (2011). 27. Chang, C.W. et al. Phosphorylation at Ser473 regulates heterochromatin protein 1 binding and corepressor function of TIF1/KAP1. BMC Mol. Biol. 9, 61 (2008). 28. Stefanov, I., Dorfman, J.R. & Germain, R.N. Self-recognition promotes the foreign antigen sensitivity of naive T lymphocytes. Nature 420, 429434 (2002). 29. Cammas, F. et al. Mice lacking the transcriptional corepressor TIF1 are defective in early postimplantation development. Development 127, 29552963 (2000). 30. Takahama, Y. et al. Functional competence of T cells in the absence of glycosylphosphatidylinositol-anchored proteins caused by T cell-specific disruption of the Pig-a gene. Eur. J. Immunol. 28, 21592166 (1998). 31. Linton, P.J., Haynes, L., Tsui, L., Zhang, X. & Swain, S. From naive to effector alterations with aging. Immunol. Rev. 160, 918 (1997). 32. Veldhoen, M., Hocking, R.J., Flavell, R.A. & Stockinger, B. Signals mediated by transforming growth factor- initiate autoimmune encephalomyelitis, but chronic inflammation is needed to sustain disease. Nat. Immunol. 7, 11511156 (2006). 33. Nelson, B.H. IL-2, regulatory T cells, and tolerance. J. Immunol. 172, 39833988 (2004). 34. Laurence, A. et al. Interleukin-2 signaling via STAT5 constrains T helper 17 cell generation. Immunity 26, 371381 (2007). 35. Zhou, X. et al. Selective miRNA disruption in Treg cells leads to uncontrolled autoimmunity. J. Exp. Med. 205, 19831991 (2008). 36. Graycar, J.L. et al. Human transforming growth factor-3: recombinant expression, purification, and biological activities in comparison with transforming growth factors-1 and -2. Mol. Endocrinol. 3, 19771986 (1989). 37. Webster, K.E. et al. In vivo expansion of Treg cells with IL-2-mAb complexes: induction of resistance to EAE and long-term acceptance of islet allografts without immunosuppression. J. Exp. Med. 206, 751760 (2009). 38. King, C., Ilic, A., Koelsch, K. & Sarvetnick, N. Homeostatic expansion of T cells during immune insufficiency generates autoimmunity. Cell 117, 265277 (2004). 39. Kench, J.A. et al. Aberrant wound healing and TGF- production in the autoimmuneprone MRL/+ mouse. Clin. Immunol. 92, 300310 (1999). 40. Lowrance, J.H., OSullivan, F.X., Caver, T.E., Waegell, W. & Gresham, H.D. Spontaneous elaboration of transforming growth factor suppresses host defense against bacterial infection in autoimmune MRL/lpr mice. J. Exp. Med. 180, 16931703 (1994). 41. Zhang, Z., Kyttaris, V.C. & Tsokos, G.C. The role of IL-23/IL-17 axis in lupus nephritis. J. Immunol. 183, 31603169 (2009). 42. Caver, T.E., OSullivan, F.X., Gold, L.I. & Gresham, H.D. Intracellular demonstration of active TGF1 in B cells and plasma cells of autoimmune mice. IgG-bound TGF1 suppresses neutrophil function and host defense against Staphylococcus aureus infection. J. Clin. Invest. 98, 24962506 (1996). 43. Mnoz-Valle, J.F. et al. The functional class evaluated in rheumatoid arthritis is associated with soluble TGF-1 serum levels but not with G915C (Arg25Pro) TGF-1 polymorphism. Rheumatol. Int. 32, 367372 (2010). 44. Gutcher, I. et al. Autocrine transforming growth factor-1 promotes in vivo Th17 cell differentiation. Immunity 34, 396408 (2011). 45. Beyer, M. et al. Repression of the genome organizer SATB1 in regulatory T cells is required for suppressive function and inhibition of effector differentiation. Nat. Immunol. 12, 898907 (1038).
npg
NUMBER 6
JUNE 2012
603
ONLINE METHODS
Antibodies. Fluorophore-conjugated antibodies used for flow cytometry were as follows: anti-CD4 (RM4-5) conjugated to allophycocyanin or fluorescein isothiocyanate; anti-CD8 (53-6.7) conjugated to fluorescein isothiocyanate; and anti-Foxp3 (FJK16-s) conjugated to allophycocyanin or phycoerythrin (all from eBioscience). Antibodies for immunoblot analysis were as follows: rabbit monoclonal anti-TRIM28 (C42-G12; Cell Signaling Technology Japan), rabbit polyclonal anti-TRIM28 (10483; Abcam), antibody to TRIM28 phosphorylated wt Ser473 (644602; BioLegend), antibody to phosphorylated tyrosine (4G10: Nihon Millipore), anti-TCR (6B10.2; Santa Cruz Biotechnology), anti-Akt (sc1618; Santa Cruz Biotech), anti-IL-2 (JES6-1; R&D Systems) and anti-LSD-1 (C69G12; Cell Signaling Technology). The hybridoma producing anti-pan-TGF- (1D11) was from American Type Culture Collection. Animals. Mice with loxP-flanked Trim28 alleles were originally generated as described29 and were provided by P. Chambon. Mice with Cre expression driven by the Lck promoter were provided by J. Takeda30 and Foxp3-GFPCretransgenic mice were provided by J. Bluestone9. CKO mice were produced from the mating of heterozygous mice with heterozygous mice or of heterozygous mice with homozygous mice, and only mice homozygous for loxP-flanked Trim28 and expressing the transgene encoding Cre are called CKO mice here. Littermates served as wild-type controls. CD45.1+ mice were provided by S. Nagata. All mice were kept under specific pathogenfree conditions at the Institute of Laboratory Animals, Graduate School of Medicine, Kyoto University and were used according to approved protocols. Cell isolation and skewing assays. Single-cell isolates from the axillary, submandibular, inguinal and mesenteric lymph nodes and the spleen were cultured on 10-cm dishes in RPMI medium supplemented with 10% FBS at 37 C for 20 min for depletion of adherent cells. The unbound fraction was subjected to magnetic selection by autoMACS (Miltenyi Biotech) with antimouse CD4 magnetic beads. The positive fraction was labeled and sorted into a CD4+CD25CD62Lhi population with a FACSAria II (BD Bioscience) to obtain naive T cells. Naive cells (2 105) and splenocyte samples that had been depleted of T cells and treated with mitomycin C (2 106 cells) were cultured together for 3 d in the presence of monoclonal anti-CD3 (1 g/ml; 2C11; eBioscience) with neutralizing anti-IFN- (10 g/ml; XMG1.2; ThermoFischer Scientific) and anti-IL-4 (10 g/ml; 1D11; ThermoFischer Scientific). EAE. Each mouse received an emulsion containing 200 g MOG peptide with or without 250 g heat-killed M. tuberculosis strain H37 RA in a mixture containing 100 l complete Freunds adjuvant, and 100 l PBS on day 0. Each mouse received an intraperitoneal injection of 200 ng of pertussis toxin on days 0 and 2. Clinical scores were assigned as follows: grade 1, limp tail; grade 2, hindlimb weakness; grade 3, partial hindlimb paralysis; grade 4, complete hindlimb paralysis; grade 5, difficulty moving forward; grade 6, death. Data are from experiments in which no mice died (all mice below grade 6). For restimulation assays, 2 106 CD4+ T cells from the lymph nodes and spleens of immunized mice on day 26 and 2 106 spleen cell samples depleted of T cells were cultured together in the presence of MOG peptide at a concentration of 30 g/ml in each well of a 48-well plate. Cytokines in culture supernatants were measured after 48 h of culture. Inflammatory bowel disease and Treg cellmediated rescue assay. Naive cells (5 105) with or without Treg cells (CD4+CD25hiCD62Lhi; 5 104) were injected intraperitoneally into Rag2/ mice (on the C57BL/6 background). Mice were weekly monitored for body weight. Mice that lost 20% of their initial body weight were considered moribund and were killed for histological analyses. For moribund mice, the body weight value at the time of death was included in the calculation of mean body weight at the later time points. Histological changes were assigned scores from 0 (normal) to 4 (very severe colitis) according to published criteria46. Cumulative scores from proximal, distal and mid-colon sections (012) were used. Mixedbone marrow transplantation and analysis. Bone marrow cells from CKO or CD45.1+ congenic Trim28+/+ mice were prepared from the femur and tibia, and samples of were magnetically depleted CD4+ cells and CD8+ cells by
autoMACS. Recipient C57BL/6 mice were -irradiated at 12 Gy (6 Gy twice, 3 h apart), and were injected intravenously with T celldepleted bone marrow (5 105 cells) from mice of each genotype or a 1:1 mixture of both (total 1 106 cells). Mice were analyzed 24 months after the transplantation. Quantitative RT-PCR. CD4+ cells sorted by autoMACS (purity >95%) were stimulated by plate-bound anti-CD3 (2C11) and anti-CD28 (37.51; eBioscience; 1 g/ml each) at a density of 1 106 cells per ml in each well of a 24-well plate, then were collected and stored at 80 C until the time of analysis. RNA was extracted with TRI reagent (Sigma), and a High Capacity cDNA RT-Kit (Applied Biosystems) was used for reverse transcription. Gene transcripts were analyzed with Taqman Gene Expression Assays (Applied Biosystems gene or probe identifier: Trim28, Mm00495594_m1; Tgfb1, Mm00441729_g1; Tgfb2, Mm00436955_m1; Tgfb3, Mm01307950_m1). Synthetic primers used for SYBR Green detection were as follows: cyclin A2, 5-TTGAACAGTTGGCAGCAC-3 and 5-AGGAGTCGCTCGGAGTC-3; cyclin E1, 5-CTTGAAGTTAAGGAGGCCAC-3 and 5-ACCACTGGTCA CAGCCTATC-3; GAPDH, 5-AAAATGGTGAAGGTCGGTGT-3 and 5-TGCCGTGAGTGGAGTCATAC-3. Microarray. RNA was extracted from TCR-stimulated and unstimulated CD4+ T cells and Treg cells with an RNeasy Mini Kit (Qiagen) and quality was assessed with a 2100 Bioanalyzer (Agilent Technologies). Both cDNA and cRNA were synthesized with a Low Input QuickAmp Labeling Kit according to the manufacturers protocol (Agilent Technologies). Samples were hybridized to a Whole Mouse Genome Microarray Kit, 4x44K (G4122F; Agilent), and dried slides were washed and then scanned (G2505B; Agilent). Microarray images were analyzed with Feature Extraction software (version 9.5; Agilent Technologies) and data were subsequently imported into Expressionist software (Genedata) for downstream analysis. Probes were normalized by quantile normalization among all microarray data. Chromatin-immunoprecipitation assay. Sorted CD4+ T cells (1 107) were fixed for 10 min in 1 ml of 1% formaldehyde, were quenched for 5 min by the addition of a 1/10 volume of 1.5 M glycine, were washed in 2% BSA in PBS and were resuspended in 350 l sonication buffer (50 mM Tris, pH 8.0, 1% SDS, 10 mM EDTA and Complete Protease Inhibitor Cocktail (Roche)). Cells were sonicated with a Bioruptor (Cosmo Bio) with a cycle of 30 s sonication, and then were allowed 1 min of rest until DNA fragments an average of 5001,000 base pairs in length were obtained. Chromatin (100 l) was diluted 10 times in radioimmunoprecipitation (RIPA) buffer (final concentrations: 50 mM Tris, pH 8.0, 150 mM NaCl, 1 mM EDTA, 1% Triton X-100, 0.1% SDS and 0.1% sodium deoxycholate), were precleared by protein G Sepharose and were subjected to one chromatin-immunoprecipitation reaction. The antibodies used were as follows: polyclonal anti-TRIM28 (10483; Abcam), antibody to trimethylated H3K9 (ab8898; Abcam) and antibody to acetylated H3K9 (39137; Active Motif). The product of each overnight immunoprecipitation reaction was captured for 4 h by 10 l protein G Sepharose, then washed sequentially in RIPA buffer (once), high-salt RIPA buffer (RIPA buffer with 500 mM NaCl; once), LiCl wash buffer (10 mM Tris, pH 8.0, 0.25 M LiCl, 1 mM EDTA, 0.5% NP-40, 0.5% sodium deoxycholate; once) and TE buffer (10 mM Tris pH 8.0, 1 mM EDTA; twice). Elution and decrosslinking were done at 65 C for overnight in 200 l CHIP elution buffer (10 mM Tris pH 8.0, 300 mM NaCl, 5 mM EDTA and 0.5% EDTA). Reactions were treated for 30 min at 37 C with RNAse (2 g/ml) and for 60 min at 55 C proteinase K (50 g/ml) and DNA was cleaned with a Qiaquick PCR purification kit (Qiagen). SYBR Green and a 7900HT Fast Real Time PCR system (Applied Biosystems) was used for semiquantitative PCR was done by with input DNA as standards (primers for amplification, Supplementary Fig. 7). Statistical analysis. The statistical significance of differences between two groups was calculated with a two-tailed Students t-test unless otherwise indicated.
46. Maloy, K.J. Induction and regulation of inflammatory bowel disease in immunodeficient mice by distinct CD4+ T-cell subsets. Methods Mol. Biol. 380, 327335 (2007).
npg
nature immunology
doi:10.1038\ni.2293
Articles
DOCK8 functions as an adaptor that links TLR-MyD88 signaling to B cell activation

Haifa H Jabara1,15, Douglas R McDonald1,15, Erin Janssen1, Michel J Massaad1, Narayanaswamy Ramesh1, Arturo Borzutzky1, Ingrid Rauter1, Halli Benson1, Lynda Schneider1, Sachin Baxi1, Mike Recher1, Luigi D Notarangelo1, Rima Wakim2, Ghassan Dbaibo2, Majed Dasouki3, Waleed Al-Herz4, Isil Barlan5, Safa Baris5, Necil Kutukculer6, Hans D Ochs7, Alessandro Plebani8, Maria Kanariou9, Gerard Lefranc10, Ismail Reisli11, Katherine A Fitzgerald12, Douglas Golenbock12, John Manis13, Sevgi Keles11,14, Reuben Ceja14, Talal A Chatila14 & Raif S Geha1
The adaptors DOCK8 and MyD88 have been linked to serological memory. Here we report that DOCK8-deficient patients had impaired antibody responses and considerably fewer CD27+ memory B cells. B cell proliferation and immunoglobulin production driven by Toll-like receptor 9 (TLR9) were considerably lower in DOCK8-deficient B cells, but those driven by the costimulatory molecule CD40 were not. In contrast, TLR9-driven expression of AICDA (which encodes the cytidine deaminase AID), the immunoglobulin receptor CD23 and the costimulatory molecule CD86 and activation of the transcription factor NF-kB, the kinase p38 and the GTPase Rac1 were intact. DOCK8 associated constitutively with MyD88 and the tyrosine kinase Pyk2 in normal B cells. After ligation of TLR9, DOCK8 became tyrosine-phosphorylated by Pyk2, bound the Src-family kinase Lyn and linked TLR9 to a Srckinase Syktranscription factor STAT3 cascade essential for TLR9-driven B cell proliferation and differentiation. Thus, DOCK8 functions as an adaptor in a TLR9-MyD88 signaling pathway in B cells. The maintenance of serological memory requires the maturation of naive B cells into memory B cells and long-lived plasma cells1. This involves interactions between ligand-receptor pairs that include CD40 ligand and the costimulatory molecule CD40; the B cellactivation factor BAFF or the proliferation-inducing ligand APRIL and the transmembrane activator TACI or the B cell maturation antigen BCMA; and Toll-like receptor (TLR) ligands and TLRs. In addition, it involves the transcription factor STAT3. Ligation of TLR9 on naive B cells promotes proliferation, immunoglobulin secretion and maturation into memory B cells24 and acts in synergy with CD40 and TACI in causing B cell activation5,6. The TLR9 ligand CpG oligodeoxynucleotide (ODN) is an adjuvant for antibody responses in mice7,8. Defective B cell responses to ligation of TLR9 have been observed in patients with common variable immune deficiency characterized by poor antibody production and less generation of memory B cells and plasma cells9. After binding its ligand, TLR9 associates with and signals via the adaptor MyD88 (ref. 10). Resting B cells express several TLRs, including TLR4 in mice, and TLR6, TLR7, TLR9 and TLR10 in mice
1Division
and humans. MyD88 is important for antibody responses to T cell dependent antigens administered with lipopolysaccharide (LPS) or CpG as the adjuvant1113, although exceptions have been noted14. MyD88 does not seem to be involved in the antibody response to T celldependent antigens administered with alum as adjuvant, or certain bacterial antigens1517, but is important for the antibody response to viruses12,18,19. DOCK8 is one of 11 members of the DOCK180 superfamily 20. DOCK (dedicator of cytokinesis) proteins have characteristic DOCKhomology region 1 (DHR-1) and DHR-2 domains. The DHR-1 domain is important for targeting DOCK proteins to membranes through its binding of phosphatidylinositol-(3,4,5)-triphosphate21. The DHR-2 domain binds to GTPases of the Rac-Rho family and can function as an exchange factor for these GTPases21. The biological functions of DOCK8 include the regulation of cell migration, morphology, adhesion and growth20. A mouse genetic screen for mutations that disrupt the persistence of the antibody response has identified loss-of-function mutations in Dock8 (ref. 22). DOCK8-mutant mice have T cell
npg
of Immunology, Childrens Hospital and Department of Pediatrics, Harvard Medical School, Boston, Massachusetts, USA. 2American University of Beirut, Beirut, Lebanon. 3Department of Pediatrics and Department of Internal Medicine, Division of Genetics, Endocrinology & Metabolism, University of Kansas Medical Center, Kansas City, Kansas, USA. 4Department of Pediatrics, Allergy and Clinical Immunology Unit, Al-Sabah Hospital, Kuwait City, Kuwait. 5Division of Pediatric Allergy and Immunology, Marmara University, Istanbul, Turkey. 6Department of Pediatric Immunology, Ege University, Izmir, Turkey. 7Department of Pediatrics, University of Washington, Seattle, Washington, USA. 8Pediatric Clinic and Angelo Nocivelli Institute of Molecular Medicine, University of Brescia, Brescia, Italy. 9Agia Sophia Childrens Hospital, Athens, Greece. 10Institute of Medical Genetics, Centre National de la Recherche Scientifique, Unit Propre de Recherch 1142, University of Montpellier, Montpellier, France. 11Division of Pediatric Allergy and Immunology, Meram Medical Faculty, Selcuk University, Konya, Turkey. 12Department of Medicine, University of Massachusetts, Worcester, Massachusetts, USA. 13Department of Transfusion Medicine, Childrens Hospital, Boston, Massachusetts, USA. 14Division of Allergy and Immunology and Department of Pediatrics, University of California Los Angeles, Los Angeles, California, USA. 15These authors contributed equally to this work. Correspondence should be addressed to R.G. (raif.geha@childrens.harvard.edu). Received 27 June 2011; accepted 11 April 2012; published online 13 May 2012; doi:10.1038/ni.2305
612
VOLUME 13
NUMBER 6
JUNE 2012
nature immunology
Articles
Table 1 IgG antibody titers in immunized DOCK8-deficient patients
Frequency of protective antibody titer TT DOCK8-deficient patients (n = 5) Normal children (n in legend) 40% >99% HepB 0% >99% HiB 20% 8397% PV 20% 50100%
Frequency of protective IgG antibody titers to TT (n = 3,032 normal children (controls)), vaccine against hepatitis B (HepB; n = 147 controls), TT-conjugated vaccine against Haemophilus influenzae type B (HiB; n = 3,486 controls) and conjugated polyvalent vaccine against pneumococcus (PV; n = 18,906 controls) in immunized DOCK8deficient patients (n = 5) and controls who received a full course of immunization with the vaccines, compared with published values in normal controls25 and the vaccine-prescribing information. Antibody titers were obtained before replacement with gamma globulin. Values for protective antibody titers were provided by the clinical laboratory at which the test was done. Data are representative of two experiments.
lymphopenia, and their B cells do not develop into marginal-zone B cells or persist in germinal centers and undergo affinity maturation. Reconstitution experiments have indicated that DOCK8 expression in B cells is important for the normal persistence of germinal centers22, which suggests that DOCK8 expression in B cells has a key role in serological memory. Mutations of DOCK8 have been shown to account for a combined immunodeficiency in humans characterized by greater susceptibility to viral skin infections, severe allergy, high serum concentrations of immunoglobulin E (IgE), eosinophilia, T cell lymphopenia and impaired antibody responses23,24. We analyzed the response of B cells from DOCK8-deficient patients to the TLR9 ligand CpG. We found that DOCK8 mediated a previously unknown MyD88 signaling pathway essential for TLR9-driven B cell proliferation and immunoglobulin production. RESULTS Antibody response and memory B cells in DOCK8 deficiency We studied ten patients 3.515 years of age with homozygous mutations in DOCK8 (Supplementary Table 1). None had detectable DOCK8 protein in lysates of peripheral blood mononuclear cells (PBMCs) or Epstein-Barr virustransformed B cells (EBV-B cells; data not shown). All had typical clinical characteristics of DOCK8 deficiency (Supplementary Table 2). Five patients, from whom serum was available before the initiation of immunoglobulin-replacement therapy, had defective IgG antibody responses to tetanus toxoid (TT), vaccine against hepatitis B, TT-conjugated vaccine against Haemophilus influenzae type B or conjugated polyvalent vaccine against pneumococcus (Table 1). The IgM antibody response to TT was significantly lower in these patients than in normal children (Supplementary Fig. 1). Two of these patients, 8 and 15 years of age,
mounted a brisk early antibody response 8 weeks after a booster dose of TT, which fell below the protective amount 15 and 12 months later, respectively (Fig. 1a). This response was in contrast to the >99% of normal children in whom protective antibody titers persisted 5 years after TT booster vaccination25,26. Flow cytometry of PBMCs showed that the frequency of CD3+ T cells was significantly lower in the patients than in age-matched healthy controls, as reported before23,24, whereas the frequency of CD19+ B cells was normal or greater (Fig. 1b). There was a considerable deficiency in the frequency of circulating CD19+CD27+ memory B cells in all patients examined, with CD19+CD27 naive B cells accounting for nearly all (>95%) of their B cells (Fig. 1c,d). The frequency of circulating IgD+CD27+ marginal zonelike B cells was lower in the patients than in the controls (Supplementary Fig. 2), consistent with the findings obtained with DOCK8-mutant mice22. These results indicated that DOCK8 was important for the generation of memory B cells and serological memory in humans. Impaired B cell activation by CpG in DOCK8 deficiency The TLR9 ligand CpG ODN 2006 (called CpG here) acts selectively on human B cells27 and has no detectable effects on non-B cells28. PBMCs from DOCK8-deficient patients were considerably deficient in their ability to proliferate and to secrete IgM and IgG in response to stimulation with CpG compared with PBMCs isolated from blood from age-matched normal subjects that was shipped together with the patients blood (Fig. 2a). In contrast, DOCK8-deficient PBMCs proliferated and secreted IgM normally after stimulation with antibody to CD40 (anti-CD40) plus interleukin 21 (IL-21) and secreted about half the amount of IgG secreted by normal PBMCs (Fig. 2b). PBMCs from the patients proliferated and secreted IgE in response to anti-CD40 plus IL-4 to an extent similar to that of normal PBMCs (Fig. 2c). The considerably impaired response of DOCK8-deficient PBMCs to CpG could not have been explained simply by the lack of memory B cells. Highly purified naive B cells from normal subjects (>95% CD27; Supplementary Fig. 3a) proliferated in response to CpG to an extent similar to that of total B cells isolated from the same subjects (Fig. 2d). Consistent with published reports24, the concentration of IgM and IgG secreted by CpG-stimulated naive B cells was much lower, at about one-half and about one-third, respectively, of that secreted by CpG-stimulated total B cells (Fig. 2d). In contrast, DOCK8-deficient purified B cells, which were almost all naive, completely failed to proliferate or to secrete IgM and IgG in response to CpG (Fig. 2d). Proliferation and secretion of IgM and IgG in response to anti-CD40
npg
a
IgG anti-TT (IU/ml)
CD19+ cells (%)
CD3+ cells (%)
4.0 3.0 2.0 1.0 0.5 0 0 3 6 9
Pt 4 Pt 7
60 30 0
40 20 0 CD27
104 103 102 101 100
C 23.9
Pt 1.46
CD19+CD27+ cells (%)
5.0
90
**
60
40 30 20 10 0 100 90 80 70 60 C C
**
12
15
Time (months)
Pt
Pt
100 101 102 103 104 CD19
Pt
Figure 1 Impaired antibody responses, inability to maintain serological memory and fewer memory B cells in DOCK8-deficient patients. (a) Serial antibody titers after reimmunization with TT in two DOCK8-deficient patients 8 (Pt 4) and 15 (Pt 7) years of age. Dotted line indicates lower limit of the protective antibody titer. (b) Frequency of CD3+ (T) cells and CD19+ (B) cells among PBMCs from DOCK8-deficient patients (Pt) and controls (C). (c) Expression of CD27 and CD19 by PBMCs from DOCK8-deficient patients and controls, analyzed by flow cytometry. Numbers in top right quadrants indicate percent CD27 + B cells. (d) Frequency of CD27+ memory B cells and CD27 naive B cells in the CD19+ B cell population of DOCK8-deficient patients and controls. Each symbol (b,d) represents an individual subject; small horizontal lines indicate the mean. *P < 0.05 and **P < 0.001 (Students t-test). Data are representative of two experiments (a) or five experiments (bd).
**
CD19+CD27 cells (%)
Pt
NUMBER 6
JUNE 2012
613
Articles a b
[ H]Td (10 c.p.m.) 12.5 IgM (g/ml) 10.0 7.5 5.0 2.5 C + Pt + 6 IgG (g/ml) 4 2 C + Pt + 15 10 5
3
[3H]Td (103 c.p.m.)
25 20 15 10 5 C +
**
IgM (g/ml)
30 20 10
**
IgG (g/ml)
25 20 15 10 5 C +
**
0 CpG
Pt
0 CpG
Pt
0 CpG
Pt
0 -CD40 + IL-21
3
0 -CD40 + IL-21
0 -CD40 + IL-21
Pt
Figure 2 Impaired CpG-driven 20 25 40 100 50 B cell proliferation and 20 40 30 15 immunoglobulin production 30 10 20 50 10 20 in response to CpG in DOCK810 5 10 deficient patients. (ac) Proliferation 0 0 0 0 0 and immunoglobulin secretion Total Naive Total Total Naive Total -CD40 + IL-4 + + -CD40 + IL-4 B cells: Total Naive Total + + by PBMCs obtained from DOCK8C Pt C Pt C Pt C Pt C Pt deficient patients and controls and left untreated () or stimulated (+) with CpG (a), anti-CD40 (-CD40) plus IL-21 (b) or anti-CD40 plus IL-4 (c). Not all 10 patients could be studied in each assay. Proliferation is assessed as the incorporation of [ 3H]thymidine ([3H]Td). Each symbol represents an individual subject; small horizontal lines indicate the mean. (d) Proliferation and secretion of IgM and IgG by highly purified total and naive B cells from controls (n = 3) and purified total B cells from DOCK8-deficient patients (n = 2), left untreated or stimulated with CpG. *P < 0.05 and **P < 0.001 (Students t-test). Data are representative of four experiments (ac) or two experiments (d; mean and s.e.m.).
[ H]Td (10 c.p.m.) [ H]Td (10 c.p.m.) IgM (g/ml)
d
3
plus IL-21 were similar in normal naive B cells and DOCK8-deficient purified B cells (Supplementary Fig. 3b). The diminished secretion of IgG by DOCK8-deficient PBMCs in response to anti-CD40 plus IL-21 was most probably secondary to the lack of memory B cells, because purified normal naive B cells showed similar less secretion of IgG in response to anti-CD40 plus IL-21 than that of total B cells (Supplementary Fig. 3b). The inability of B cells from DOCK8deficient patients to respond to CpG was not due to more apoptosis and cell death, as determined by staining with annexin V and propidium iodide (Supplementary Fig. 4). These results indicated that B cells from DOCK8-deficient patients had an intrinsic and selective defect in their ability to respond to CpG. CpG activates known pathways in DOCK8-deficient B cells The stimulation of B cells with CpG ODN induces the expression of AICDA, which encodes activation-induced cytosine deaminase (AID), critical for isotype switching, and upregulates expression of the low-affinity IgE receptor CD23 and the costimulatory molecule CD86 (refs. 4,29). Stimulation with CpG induced similar AICDA mRNA expression in PBMCs and B cells from DOCK8-deficient patients and controls (Fig. 3a) and similarly greater frequency of CD19 + B cells that expressed CD23 and CD86 (Fig. 3b,c). The mean fluorescence intensity of these antigens was similar for DOCK8-deficient B cells and control B cells (Fig. 3b,c). The transcription factor NF-B and the mitogen-activated protein kinase p38 are activated by stimulation with CpG ODN and have an important role in inducing the expression of AICDA, CD23 and CD86 (refs. 3032). Stimulation with CpG caused similar phosphorylation of the inhibitor IB and p38 in EBV-B cells and PBMCs from DOCK8-deficient patients and controls (Fig. 3d and data not shown). CpG-driven secretion of IL-6 by EBV-B cells was not significantly different in patients versus controls (Fig. 3e). Stimulation with CpG ODN causes plasmacytoid dendritic cells (pDCs) to secrete interferon- (IFN-). This induction requires MyD88-dependent activation of the transcription factor IRF7 (ref. 33). Stimulation of PBMCs with CpG type A (ODN 2216), which activates TLR9 in human pDCs, resulted in similar secretion of IFN- by cells from patients and controls (Fig. 3f ). The DHR-2 domain of DOCK proteins can activate the GTPase Rac and/or the GTP-binding protein Cdc42 (ref. 21), and ligation of TLR9 activates Rac1 in dendritic cells34. To examine Rac1 activation,
614
we used a fusion of glutathione S-transferase and the GTPase-binding domain of p21-activated kinase to precipitate GTP-bound Rac1 and Cdc42 from cell lysates, followed by immunoblot analysis with antiRac1 and anti-Cdc42. Stimulation with CpG resulted in similar Rac1 activation in EBV-B cells from normal and DOCK8-deficient subjects (Fig. 3g) but caused no detectable activation of Cdc42 (data not shown). Thus, CpG activated Rac1 in B cells, but DOCK8 was not essential for this activation. DOCK8-dependent activation of STAT3 by CpG in B cells STAT3 is important for B cell proliferation and differentiation in response to anti-CD40 plus IL-21 (ref. 35,36). The stimulation of normal PBMCs with CpG induced phosphorylation of STAT3 at Tyr705 (Fig. 4a). In contrast, CpG-driven phosphorylation of STAT3 was defective in DOCK8-deficient PBMCs, but that driven by IL-21 was not (Fig. 4a). Phosphorylation of STAT3 driven by CpG was defective in purified DOCK8-deficient B cells, but that induced by IFN- was not (Fig. 4b), which indicated that the defect was B cell autonomous. IL-6 and IL-21 caused similar phosphorylation of STAT3 in purified B cells from patients and controls (Fig. 4c). Densitometryscanning analysis demonstrated that CpG-driven phosphorylation of STAT3 in B cells from DOCK8-deficient patients was 18% 7% that of cells from controls, whereas IL-21-driven phosphorylation of STAT3 was similar in these cells (Fig. 4d). The defect in CpG-driven phosphorylation of STAT3 in DOCK8-deficient B cells was not secondary to the lack of memory B cells, because CpG caused similar phosphorylation of STAT3 in purified naive B cells and unfractionated B cells from normal controls (Fig. 4e). These results demonstrated that DOCK8 deficiency selectively impaired CpG-driven phosphorylation of STAT3 in B cells. We investigated the role of STAT3 in the CpG-driven proliferation and differentiation of B cells by examining PBMCs from six patients with autosomal dominant hyper-IgE syndrome (AD-HIES) resulting from different dominant-negative mutations of STAT3 (Supplementary Table 3). PBMCs from these patients had significantly impaired proliferation and IgG secretion after stimulation with CpG (Fig. 4f). There was no significant difference between PBMCs from patients with AD-HIES and those from controls in their ability to proliferate and secrete IgE in response to anti-CD40 plus IL-4 (Fig. 4g). As reported before35, PBMCs from patients
npg
IgG (g/ml)
IgE (ng/ml)
Articles
Figure 3 Normal CpG-driven upregulation of the expression of AICDA, CD23 and CD86, C Pt 2.5 6 PBMCs B cells activation of NFB, p38 and Rac1, and CpG (min) 0 20 30 45 0 20 30 45 2.0 p-IB secretion of IL-6 in B cells from DOCK84 1.5 deficient patients. (a) Expression of AICDA p-p38 1.0 mRNA in PBMCs and purified B cells obtained IKK 2 0.5 from patients and controls (n = 2 per group) and 200 250 left untreated or treated with CpG; results are 0 0 200 CpG + + + + CpG presented relative to those of GAPDH (encoding 150 100 C Pt Pt C glyceraldehyde phosphate dehydrogenase). 100 50 (b,c) Frequency of CD19+ B cells with surface 1,250 0 80 0 expression of CD23 or CD86 (left), and mean 20 30 45 20 30 45 1,000 Time (min) Time (min) 60 fluorescence intensity of those cells (right), 750 before () and after (+) CpG stimulation of 40 200 200 500 PBMCs from patients and controls (n = 3 per 20 group). (d) Immunoblot analysis of the 250 phosphorylation (p-) of IB and p38 in PBMCs 0 0 100 100 + + + + CpG CpG from patients and controls (n = 3 per group) before (0) or 20, 30 or 45 min after stimulation Pt C Pt C 300 75 with CpG (top); the kinase IKK serves as a 0 0 C Pt C Pt loading control. Below quantification of those 200 50 results, presented as the ratio of phosphorylated C Pt Ctrl IB or p38 to total IKK in patients relative to CpG (min) 0 10 15 0 10 15 GDP GTP 100 25 that in healthy controls. (e) Secretion of IL-6 Rac1 precipitation by CpG-stimulated EBV-B cells from patients, 0 0 Lysates presented relative to CpG-stimulated cells CpG + + + + CpG from control subjects (n = 3 per group; Pt C C Pt mean for controls, 55 pg/ml). (f) IFN- secretion by pDCs in PBMCs from patients and controls (n = 3 per group) in response to stimulation with CpG type A (ODN 2216; mean for controls 22.4 pg per 100 pDCs); the number of pDCs per culture was calculated from the frequency of CD123 +BDCA-4+ cells in each sample. (g) Activation of Rac1 in EBV-B cells, assessed by precipitation assay with a fusion of glutathione S-transferase and the GTPase-binding domain of p21-activated kinase, followed by immunoblot analysis of Rac1. Bottom, immunoblot analysis of Rac1 in lysate aliquots of volume equal to that in the precipitation assay (loading control). Data are representative of two (a) or three (bg) experiments (mean and s.e.m. in af).
AICDA (relative 102) AICDA (relative 102)
CD19+CD23+ (MFI)
CD19+CD23+ (%)
p-IB/IKK (% of control)
IL-6 (% of control)
CD19+CD86+ (MFI)
CD19+CD86+ (%)
with AD-HIES did not proliferate or secrete IgG in response to anti-CD40 plus IL-21 (Supplementary Fig. 5). Because IL-6 caused phosphorylation of STAT3 in DOCK8deficient B cells, we examined whether exogenous IL-6 corrected their defective response to CpG. The addition of recombinant IL-6 to DOCK8-deficient PBMCs did not correct their defective proliferation or IgG secretion in response to CpG (Fig. 4h and data not shown). These results indicated that in addition to the phosphorylation of STAT3, other signals important for CpG-driven proliferation and differentiation were impaired in DOCK8-deficient B cells. The expression of TLR9 protein was similar in B cells from patients and controls (Supplementary Fig. 6a). Correct subcellular localization of TLR9 is essential for its function37. Lack of DOCK8 might result in mislocalization of TLR9, which could interfere with the ability of TLR9 to activate STAT3. We stimulated EBV-B cells from normal subjects and DOCK8-deficient patients for 90 min with CpG, made them permeable and simultaneously stained for TLR9 and the endoplasmic reticulum, the early endosomal marker EEA1 or the late endosomal-lysosomal marker LAMP-1 (CD107a). Fluorescence microscopy showed that the subcellular localization of TLR9 was similar in normal and DOCK8-deficient EBV-B cells (Supplementary Fig. 6b). Thus, DOCK8 had no detectable role in TLR9 trafficking in B cells. Syk is downstream of DOCK8 in CpG-activated B cells The tyrosine kinase Syk activates STAT3 in B lymphoma cells 38,39. Stimulation of normal PBMCs with CpG resulted in phosphorylation of Syk on residue Tyr352, the target of kinases of the Src family (collectively called Src here; Fig. 5a). In contrast, it caused negligible phosphorylation of Syk in DOCK8-deficient PBMCs (Fig. 5a). The defect in phosphorylation of Syk was specific to CpG, because
crosslinking of the B cell antigen receptor (BCR) with anti-IgM caused similar phosphorylation of Syk in cells from patients and those from controls (Fig. 5b). Densitometry-scanning analysis demonstrated that CpG-driven phosphorylation of Syk in PBMCs from DOCK8deficient patients was 16% 8% that of controls, whereas BCRdriven phosphorylation of Syk was not significantly different in these cells (Fig. 5c). We used the pentapeptide Syk inhibitor SYKINH-61 to determine whether Syk activation is important for the CpG-driven phosphorylation of STAT3. SYKINH-61 inhibited CpG-driven phosphorylation of STAT3 in normal PBMCs (85% 5% inhibition) but had no effect on IL-21-driven phosphorylation of STAT3 (Fig. 5d) or CpG-driven phosphorylation of p38 (data not shown). Moreover, SYKINH-61 inhibited the proliferation and IgG production driven by CpG in normal PBMCs in a dose-dependent manner but had no effect on the proliferation and IgG production driven by CD40 and IL-21 (Fig. 5e,f). We verified the role of Syk in CpG-driven phosphorylation of STAT3 by RNA-mediated silencing. The mouse B cell line CH12 faithfully reproduced the response of primary human B cells to CpG stimulation by showing phosphorylation of Pyk2, Syk and STAT3 (Supplementary Fig. 7a). Treatment of CH12 cells with short hairpin RNA targeting Syk resulted in much lower expression of Syk protein and considerably inhibited the CpG-driven phosphorylation of STAT3, but the introduction of nonsilencing short hairpin RNA did not (Supplementary Fig. 7bd). These results placed Syk upstream of the activation of STAT3 by CpG in B cells. CpG may bind to sensors other than TLR9 (ref. 34). To determine if the DOCK8-dependent CpG-driven phosphorylation of Syk and STAT3 proceeded through the engagement of TLR9 and its adaptor MyD88, we compared the responses of purified splenic B cells
615
npg
IFN-/pDC (% of control)
p-p38/IKK (% of control)
Articles a
C p-STAT3 STAT3 CpG IL-21 Pt CpG IL-21 Time (h) p-STAT3 IB STAT3
b
0.5
CpG 3 4
IFN- 3
CpG 0.5 3 4
IFN- 3 p-STAT3 STAT3
Pt
p-STAT3/STAT3 (% of control)
Pt
c
IL-6 IL-21 IL-6 IL-21
150 100 50 0
e
B cells: CpG p-STAT3 Naive + Total +
f
[3H]Td (103 c.p.m.) 8 6 4 2
[ H]Td (10 c.p.m.)
**
IgG (g/ml)
6 4 2
**
g
15 10 5
3
NS IgE (ng/ml)
20 15 10 5
NS
h
IgG (g/ml) 8 6 4 2 2 1 0
C Pt CpG
C Pt IL-21
STAT3
CpG IL-6 CpG CpG 0 Figure 4 CpG-induced IL-6 + + + + + + + CpG + DOCK8-dependent C AD-HIES Pt C C AD-HIES C AD-HIES AD-HIES C phosphorylation of STAT3 in B cells is essential for B cell proliferation and IgG production. (ac) Immunoblot analysis of STAT3 phosphorylated at Tyr705 and total STAT3 in PBMCs (a) and purified B cells (b,c) obtained from DOCK8-deficient patients and controls and left unstimulated () or stimulated with CpG or IL-21 (a), CpG or IFN- (b; times, above lanes), or IL-6 or IL-21 (c). (d) Quantification of STAT3 phosphorylated at Tyr705 in B cells obtained from DOCK8-deficient patients and controls (n = 3 per group) and treated with CpG or IL-21, presented as the ratio of phosphorylated STAT3 to total STAT3 in patients relative to that in controls. (e) Immunoblot analysis of phosphorylated STAT3 in total B cells and naive B cells purified from one donor and left unstimulated or stimulated with CpG. (f,g) Proliferation and immunoglobulin secretion by PBMCs obtained from STAT3-deficient patients with AD-HIES and controls (n = 6 (f) or 5 (g) per group) and left unstimulated () or stimulated (+) with CpG (f) or anti-CD40 plus IL-4 (g). Each symbol represents an individual subject; small horizontal lines indicate the mean. (h) IgG secretion by PBMCs obtained from DOCK8-deficient patients (n = 3) and left unstimulated or stimulated with CpG or IL-6 (2 ng/ml) alone or in combination. NS, not significant; *P < 0.01 and **P < 0.001 (Students t-test). Data are representative of three (ae,h) or five (f,g) experiments (mean and s.e.m. in d,h).
0 CpG
0 -CD40 + IL-4
3
0 -CD40 + IL-4
(>95% B220+ cells) from Tlr9/ and Myd88/ mice and congenic wild-type mice (as controls). Stimulation with CpG ODN 1826, which binds to mouse TLR9, resulted in the phosphorylation of Syk and STAT3 in B cells from wild-type mice (Fig. 5g,h), as in normal human B cells. In contrast, stimulation with CpG caused minimal phosphorylation of Syk and STAT3 in B cells from Tlr9 / and Myd88/ mice. Furthermore, CpG ODN 1826 did not cause proliferation or IgG secretion by B cells from Tlr9/ and Myd88/ mice (Supplementary Fig. 8). Together these results indicated that CpG ODN 1826 engaged TLR9-MyD88 to cause DOCK8-dependent
phosphorylation of Syk and STAT3, which was essential for the proliferation and differentiation of B cells. CpG activates a Pyk2-Src-Syk-STAT3 cascade in B cells Stimulation of normal PBMCs with CpG resulted in the tyrosine phosphorylation of proteins with molecular weights that corresponded to those of Src (5456 kilodaltons (kDa)), the Src target Syk (72 kDa) and Pyk2 (110 kDa), a tyrosine kinase that functionally interacts with Src (Fig. 6a). In contrast, it caused minimal protein tyrosine phosphorylation in PBMCs from DOCK8-deficient patients. This defect was
npg
a
CpG (min) 0 p-Syk BLNK C 60 90 Pt 0 60 90
b
C -IgM p-Syk Syk + Pt +
c
p-Syk/BLNK (% of control)
200
NS
d
SYKINH CpG IL-21 CpG IL-21 p-STAT3 STAT3
e
Proliferation (% of control)
125 100 75 50 25 0
**
150
Proliferation (% of control)
100
***
100 50
C Pt C Pt -IgM CpG
SYKINH (M)
0 0 0.05 0.1 0.5 SYKINH (M) CpG
0 0.05 0.1 0.5 -CD40 + IL-21
Figure 5 CpG-driven phosphorylation / / WT Tlr9 Tlr9 WT ** of STAT3 in B cells is dependent on 250 125 CpG IL-21 CpG IL-21 ** CpG (min) 0 5 30 60 0 5 30 60 Syk and requires TLR9 and MyD88. 200 100 p-Syk p-STAT3 (a,b) Immunoblot analysis of Syk 150 75 Syk STAT3 phosphorylated at Tyr352, as well 100 50 / WT Myd88 WT Myd88/ as BLNK (loading control; a) and 50 25 CpG IL-21 CpG IL-21 CpG (min) 0 30 60 90 120 0 30 60 90 120 0 0 total Syk (b), in control and DOCK8p-Syk p-STAT3 SYKINH (M) 0 0.05 0.1 0.5 SYKINH (M) 0 0.05 0.1 0.5 deficient PBMCs before and after STAT3 Syk -CD40 + IL-21 CpG stimulation with CpG (a) or anti-IgM (b). (c) Quantification of Syk phosphorylated at Tyr352 in PBMCs obtained from DOCK8-deficient patients and controls (n = 3 per group) and stimulated with CpG or anti-IgM; bands for phosphorylated Syk and BLNK were scanned by densitometry and the p-Syk/BLNK ratio was calculated and is presented relative to the mean ratio in controls, set as 100%. (d) Immunoblot analysis of total and phosphorylated Syk (Tyr352) in normal PBMCs (n = 3 subjects) left unstimulated or stimulated with CpG or IL-21 (control), alone (left) or in the presence of the Syk inhibitor SYKINH-61 (SYKINH; right). (e,f) Proliferation (e) and IgG production (f) by normal PBMCs (n = 4 subjects) stimulated with CpG (left) or anti-CD40 plus IL-21 (right) in the presence or absence (0) of various concentrations of SYKINH-61. (g) Immunoblot analysis of total and phosphorylated Syk (Tyr352) in splenic B cells obtained from Tlr9/ and congenic C57BL/6 wild-type mice (top) or Myd88/ and congenic BALB/c wild-type mice (bottom) and treated for 060 min (above lanes) with CpG ODN 1826. (h) Immunoblot analysis of total and phosphorylated STAT3 (Tyr705) in splenic B cells (genotypes as in g) left untreated () or stimulated with CpG or IL-21. *P < 0.05, **P < 0.01 and ***P < 0.001 (Students t-test). Data are representative of three (ad,g,h) or four (e,f) experiments (mean and s.e.m in c,e,f).
IgG (% of control) IgG (% of control)
616
VOLUME 13
NUMBER 6
JUNE 2012
nature immunology
Articles a
CpG (min) 0 (kDa) 150 100 p-Tyr 75 50 BLNK C 60 90 0 Pt 60 90
b
CpG (min) p-Pyk2 p-Src BLNK 0 C 30 60 0 Pt 30 60
c
WT CpG (min) 0 p-Pyk2 p-Src Syk 30 60 90 120 0
Myd88/ 30 60 90 120
d
Inhibitor CpG p-Pyk2 Pyk2 + Tyr A9 + PP2 +
e
Inhibitor CpG p-Src Src + Tyr A9 PP2 + +
f
Inhibitor CpG p-Syk Syk + Tyr A9 PP2 + +
g
Inhibitor p-STAT3 STAT3 Tyr A9 PP2 CpG IL-21 CpG IL-21 CpG IL-21
h 200
* Figure 6 DOCK8 mediates CpG activation of a Pyk2-Src-Syk-STAT3 cascade essential for B cell proliferation ** and differentiation. (a,b) Immunoblot analysis of tyrosine-phosphorylated proteins (p-Tyr; a) and phosphorylation 100 of Pyk2 at Tyr402 and of Src at Tyr416 (b) in lysates of DOCK8-deficient and normal control PBMCs left unstimulated or stimulated for various times (above lanes) with CpG. BLNK serves as a loading control. The greater abundance of BLNK in DOCK8-deficient PBMCs reflects the greater frequency of CD19 + B cells in DOCK8-deficient patients than in controls (23% versus 7% (a) or 19% versus 8% (b), respectively). (c) Immunoblot analysis of Pyk2 and Src (as in b) in lysates of splenic B cells from Myd88/ and congenic BALB/c wild-type mice, 0 Tyr PP2 Tyr PP2 treated as in a,b. (dg) Immunoblot analysis of the phosphorylation of Pyk2 at Tyr402 (d), Src at Tyr416 (e), A9 A9 Syk at Tyr352 (f) and STAT3 at Tyr705 (g) in normal PBMCs left untreated or treated with CpG alone or with the -CD40 + IL-21 CpG Pyk2 inhibitor tyrphostin A9 (Tyr A9) or the Src kinase inhibitor PP2. (h) IgG secretion by normal PBMCs (n = 3 subjects) treated with CpG (left) or anti-CD40 plus IL-21 (right; control), alone or in combination with tyrphostin A9 or PP2. *P < 0.05 and **P < 0.01 (Students t-test). Data are representative of three independent experiments (ag) or are from four experiments (h; mean and s.e.m.).
IgG (% of control)
selective to stimulation with CpG, because crosslinkage of the BCR caused similar tyrosine phosphorylation in PBMCs from DOCK8deficient patients and those from controls (data not shown), consistent with the normal activation of B cells from DOCK8-mutant mice after crosslinkage of the BCR22. Pyk2 is autophosphorylated on Tyr402 (ref. 40). Activated Src is autophosphorylated on tyrosine residues that correspond to Tyr416 in Src40 and is recognized by antibody to Src phosphorylated at Tyr416. Immunoblot analysis of lysates of PBMCs with a phosphorylationspecific monoclonal antibody to Pyk2 phosphorylated at Tyr402 and to Src phosphorylated at Tyr416 demonstrated that stimulation with CpG caused phosphorylation of Pyk2 and Src in normal PBMCs (Fig. 6b). In contrast, stimulation with CpG caused minimal phosphorylation of Pyk2 and Src in DOCK8-deficient PBMCs (Fig. 6b), which demonstrated that CpG-driven tyrosine phosphorylation of Pyk2 and Src was DOCK8 dependent. Pyk2 and Src were phosphorylated after CpG stimulation of B cells from wild-type mice but not those from Myd88/ mice (Fig. 6c), which indicated that TLR9-driven phosphorylation of Pyk2 and Src, like that of Syk and STAT3, was dependent on MyD88. We used selective inhibitors of Pyk2 and Src to examine the roles of these kinases in the CpG-driven activation of B cells. The Pyk2 inhibitor tyrphostin A9 considerably inhibited the CpG-driven tyrosine phosphorylation of Pyk2 (84% 3%), Src (81% 6%), Syk (95% 4%) and STAT3 (97% 2%; Fig. 6dg). Tyrphostin A9 had no detectable effect on the IL-6-driven phosphorylation of STAT3 or CpG-driven phosphorylation of p38 (data not shown). Tyrphostin A9 considerably inhibited the secretion of IgG driven by CpG (90% 2%) but not that driven by CD40 and IL-21 (Fig. 6h). We obtained similar results with the structurally different Pyk2 inhibitor PF562271 (data not shown). We verified the role of Pyk2 in the CpG-driven phosphorylation of STAT3 in CH12 B cells through the knockdown of Pyk2 with short hairpin RNA (Supplementary Fig. 7b-d). The Src inhibitor PP2 had a modest effect on the inhibition of CpGdriven phosphorylation of Pyk2 (28% 13%; Fig. 6d) but had a much greater effect on the inhibition of CpG-driven phosphorylation of
Src (80% 8%), Syk (71% 9%) and STAT3 (63% 11%; Fig. 6eg). The modest inhibition of Pyk2 phosphorylation was consistent with the ability of Src to activate Pyk2 (ref. 40) and suggested feedback amplification of the activation of Pyk2 by Src. The less-than-complete inhibition of Syk and phosphorylation of STAT3 could have been due to the ~20% residual Src activity. Alternatively, a Src-independent link between Pyk2 and Syk-STAT3 might exist. PP2 inhibited the secretion of IgG driven by CpG (70% 12%) but had no effect on IgG secretion driven by CD40 and IL-21 (Fig. 6h). We obtained similar results with the structurally different Src kinase inhibitor SU6656 (data not shown). Together these results suggested that ligation of TLR9 activated a Pyk2-Src-Syk cascade that resulted in the phosphorylation of STAT3. A DOCK8-MyD88-Pyk2 complex links to Lyn Our results placed DOCK8 and Pyk2 downstream of MyD88. To determine whether DOCK8 associates with MyD88, we transfected 293T human embryonic kidney cells with vectors encoding Myc-tagged DOCK8 and either hemagglutinin-tagged MyD88 or hemagglutinin-tagged MRTF-A (myocardin-related transcription factor A; control). DOCK8 precipitated together with MyD88 but not with MRTF-A (Fig. 7a). To determine whether DOCK8 associated with MyD88 in B cells, we immunoprecipitated proteins from normal EBV-B cells with anti-DOCK8 and probed for MyD88. DOCK8 associated weakly with MyD88 in unstimulated cells but this association increased after stimulation with CpG (Fig. 7b). Pyk2 is reported to interact with MyD88 in macrophages41. This interaction has been mapped to the proline-rich carboxy-terminal region of Pyk2 and the death domain of MyD88. When we immunoprecipitated proteins from normal EBV-B cells with anti-MyD88 and probed the immunoprecipitates with monoclonal antibody to Pyk2, we found that Pyk2 associated with MyD88 (Fig. 7c). This association was not altered detectably by stimulation with CpG and was independent of DOCK8, as Pyk2 precipitated together with MyD88 from lysates of EBV-B cells from DOCK8-deficient patients (Fig. 7c). We readily detected Pyk2 among proteins immunoprecipitated from
617
npg
Articles a
DOCK8-Myc MyD88-HA MRTF-AHA DOCK8 MyD88 MRTF-A IP: -HA + + + + + + Lys + +
IP: -DOCK8 +
IgG -DOCK8 + + + +
CpG Tyr A9 p-Tyr Src DOCK8 Src DOCK8
Lys
IP:
DOCK8 0 30
IgG 30
CpG (min) MyD88 DOCK8
f g
IP: CpG p-Src Lyn
-Lyn +
IgG +
C IP: -MyD88 0 30 Pyk2 MyD88 30
DOCK8 def IgG -MyD88 0 30
IP:
-DOCK8 + + +
IgG +
CpG (min)
CpG Tyr A9 Lyn DOCK8
in DOCK8 (HNKSPDFYEEVKIKL; amino acids 621-635) that is a potential binding site for the Src-homology 2 domain of Src. Probing DOCK8 immunoprecipitates from EBV-B cells with monoclonal antibody to phosphorylated tyrosine and to Src showed that stimulation with CpG caused tyrosine phosphorylation of DOCK8 and its association with Src (Fig. 7e). Both events were abrogated by the Pyk2 inhibitor tyrphostin A9. The Src Lyn is expressed preferentially in B cells43. When we immunoprecipitated proteins with anti-Lyn and probed the immunoprecipitates with a monoclonal antibody to Src phosphorylated at Tyr416, we found that stimulation with CpG caused tyrosine phosphorylation of Lyn in normal B cells (Fig. 7f). Furthermore, stimulation with CpG resulted in robust association of Lyn with DOCK8, which was abrogated by the addition of tyrphostin A9 (Fig. 7g). These results indicated that ligation of TLR9 in B cells resulted in Pyk2-dependent phosphorylation of DOCK8 and recruitment of Src and/or Lyn. DISCUSSION We have presented evidence here that DOCK8 functioned as an adaptor that linked TLR9 via MyD88 to a Pyk2-Src-Syk-STAT3 signaling cascade, and we have demonstrated that this pathway was essential for TLR9-driven B cell proliferation and immunoglobulin production. DOCK8-deficient patients had impaired ability to sustain a protective antibody response, similar to DOCK8-mutant mice, and lacked circulating CD27+ memory B cells. DOCK8-deficient B cells did not proliferate or secrete IgM or IgG in response to CpG. This defect was cell autonomous, was specific to CpG and was not accounted for by the lack of memory B cells. Because of constraints in obtaining sufficient amounts of blood from children for B cell purification, we used PBMCs, which we stimulated with the B cellselective TLR9 ligand CpG, and/or EBV-B cells. Whenever possible, we also used purified B cells from at least two patients. Stimulation with CpG upregulated the expression of AICDA, CD23 and CD86 normally in DOCK8-deficient B cells and resulted in normal activation of NF-B and p38 and normal IRF7-dependent secretion of IFN- in DOCK8-deficient PBMCs, which indicated that these events occurred independently of DOCK8. Unexpectedly, given the reported guanine nucleotideexchange activity of DOCK proteins for small GTPases21, CpG activated Rac1 independently of DOCK8. The activation of Rac1 by CpG in pDCs uses a non-TLR9 sensor and DOCK2 (ref. 34). A similar pathway could be operative in B cells. A central finding of our study was that CpG caused phosphorylation of STAT3 in B cells in a DOCK8-dependent manner. Impaired phosphorylation of STAT3 in DOCK8-deficient B cells was specific to stimulation with CpG and was not secondary to the lack of memory B cells. A critical role for STAT3 in CpG-driven proliferation and differentiation of B cells was demonstrated by the observation that these responses were impaired in patients with dominant-negative mutations in STAT3. The inability of IL-6, which caused phosphorylation of STAT3 in DOCK8-deficient B cells, to correct their defective response to CpG indicated that other DOCK8-dependent signals were also required for CpG-driven B cell proliferation and IgG secretion. These may include signals delivered by Pyk2, Src and Syk, which trigger activation of phospholipase C-, phosphatidylinositol-3-OH kinase and the B cell linker BLNK4446. We found that Syk had a critical role in the CpG-driven, STAT3dependent proliferation and differentiation of B cells. Stimulation of B cells with CpG resulted in the phosphorylation of Syk, which was considerably impaired in DOCK8-deficient B cells; this placed Syk downstream of DOCK8. The Syk-selective inhibitor SYKINH-61 blocked CpG-driven phosphorylation of STAT3, which placed Syk
d
CpG (min) Pyk2 DOCK8 MyD88 0
Lyn/ DOCK8
C IP: -DOCK8 30 IgG 30
MyD88 def -DOCK8 0 30
0.8 0.6 0.4 0.2 +
Lys
0 CpG Tyr A9
+ +
Figure 7 DOCK8 associates with MyD88 and Pyk2 and undergoes Pyk2-mediated tyrosine phosphorylation and association with Src and/or Lyn after ligation of TLR9. (a) Immunoblot analysis of 293T cells transfected with various combinations (above lanes) of vectors encoding Myc-tagged DOCK8 (DOCK8-Myc) and hemagglutinin-tagged MyD88 (MyD88-HA) or MRTF-A (MRTF-AHA), assessed before (Lys; right) or after (IP: -HA; left) immunoprecipitation with anti-hemagglutinin. (b) Immunoblot analysis of EBV-B cells left untreated or stimulated for 30 min with CpG, followed by immunoprecipitation with anti-DOCK8 or IgG. (c,d) Immunoblot analysis of control and DOCK8-deficient (DOCK8 def; c) or MyD88-deficient (MyD88 def; d) EBV-B cells left untreated or treated for 30 min with CpG, followed by immunoprecipitation with anti-MyD88 (c), anti-DOCK8 (d) or IgG (c,d). (e) Immunoblot analysis of normal B cells left untreated or stimulated with CpG in the presence (+) or absence () of tyrphostin A9, followed by immunoprecipitation with anti-DOCK8 or IgG, probed with antibody to phosphorylated tyrosine, anti-Src and anti-DOCK8 (loading control); below (Lys), immunoblot analysis of Src and DOCK8 in lysates before immunoprecipitation (equal lysate volume, top and bottom blots). (f) Immunoblot analysis of normal PBMCs left untreated or stimulated with CpG, followed by immunoprecipitation with anti-Lyn or IgG, probed with antibody to Src phosphorylated at Tyr416 or anti-Lyn (loading control). (g) Immunoblot analysis (top) of normal PBMCs (n = 3 subjects) left untreated or stimulated with CpG in the presence or absence of tyrphostin A9, followed by immunoprecipitation with anti-DOCK8 or IgG, probed with anti-Lyn and anti-DOCK8 (loading control). Below, quantitative analysis of band intensity in the blot above, presented as the ratio of Lyn to DOCK8. Data are representative of three experiments (mean and s.e.m in g).
npg
normal EBV-B cells with anti-DOCK8, and this association increased after stimulation with CpG (Fig. 7d). The constitutive association of DOCK8 and Pyk2 was independent of MyD88 and was present in EBV-B cells from a MyD88-deficient patient that had no detectable expression of MyD88 protein (Fig. 7d). However, stimulation with CpG did not increase the association of DOCK8 and Pyk2 in the absence of MyD88. These results suggested that DOCK8 formed a complex with MyD88 and Pyk2. The Tyr-Glu-Glu-Val-Lys motif (amino acids 628632) in the DHR-1 domain of DOCK8 also exists in cofilin-1, in which it is a target of Pyk2 phosphorylation42. This motif is located in a sequence
618
Articles
upstream of STAT3, and inhibited CpG-driven B cell proliferation and IgG secretion. Stimulation of B cells with CpG resulted in DOCK8dependent phosphorylation of Pyk2 and Src. We demonstrated that DOCK8 linked TLR9-MyD88 to a Pyk2-Src-Syk-STAT3 cascade in B cells, which we found was essential for the TLR9-driven activation of B cells. The observation that CpG-driven phosphorylation of Pyk2, Src, Syk and STAT3, as well as B cell proliferation and differentiation, were dependent on TLR9 and MyD88 indicated that engagement of TLR9-MyD88 by CpG resulted in DOCK8-dependent activation of B cells. CpG type A DNA has been reported to induce the tyrosine phosphorylation of proteins, including Syk, in monocytes and macrophages independently of TLR9 and MyD88 (ref. 47). Differences between that study47 and our study here in the oligodeoxynucleotides and target cells used may account for the difference noted in the requirement for TLR9 and MyD88. We found DOCK8 existed in a complex with MyD88 and Pyk2. MyD88 was not essential for the DOCK8-Pyk2 association, and DOCK8 was not essential for the MyD88-Pyk2 association, although it was essential for the phosphorylation of Pyk2 after TLR9 ligation. After stimulation with CpG, DOCK8 became more strongly associated with MyD88 and Pyk2, underwent tyrosine phosphorylation and associated with Src and/or Lyn. We propose the following model of DOCK8-dependent TLR9 signaling. Ligation of TLR9 by CpG causes recruitment and stabilization of a preexisting MyD88-Pyk2-DOCK8 complex, which results in the autophosphorylation and activation of Pyk2. Pyk2 then phosphorylates DOCK8, causing it to recruit Src kinases, including Lyn, via their Src-homology 2 domain, which releases them from autoinhibition. Src then activates Syk, which drives STAT3 activation. Our proposed pathway is probably a simplification. Src and Syk can phosphorylate Pyk2 (ref. 40), and Src and Pyk2 can act in synergy to cause STAT3 phosphorylation48,49. Future experiments should determine whether the DOCK8-dependent TLR9-MyD88 signaling pathway we have identified is used by other receptors that signal via MyD88 in B cells and other cells. Preliminary data have indicated that this pathway is activated by TLR4 ligation in PBMCs (data not shown). DOCK8 deficiency results in impaired immunological synapses in B cells22 and may impair the ability of T cells and dendritic cells to drive antibody production by B cells. Mice with selective DOCK8 deficiency in B cells, as well as mice in which the interaction between MyD88 and DOCK8 is disrupted, will help define the contribution of DOCK8dependent MyD88 signaling in B cells to the impaired serological memory of DOCK8 deficiency. Given that TLR9 ligands are vaccine adjuvants and given the role of TLR9 in autoantibody responses to self DNA50, the TLR9 signaling pathway we have described may be important for the development of better vaccines and the understanding and treatment of autoantibody-mediated diseases. METHODS Methods and any associated references are available in the online version of the paper.
Note: Supplementary information is available in the online version of the paper. ACKNOWLEDGMENTS We thank F. Uckun (University of Southern California, Los Angeles) for SYKINH-61; R. Treisman (Cancer Research UK) for hemagglutinin-tagged MRTF-A; J.-L. Casanova (The Rockefeller University), A. Puel and C. Picard (Hopital Necker-Enfants Malades, France) for the MyD88-deficient EBV-B cell line; K. Eurich for technical assistance; A. Rambhatla and A. Chen for help in DNA sequencing; members of the Geha laboratory for discussions; J. Kagan and H. Oettgen for comments on the manuscript; and the patients and their families for donating blood. Supported by the US Public Health Service (P01AI076210, T32AI007512 and R01AI083503 to R.S.G.; R21AI087627 to T.A.C.; and K08AI076625 to D.R.M.), the Dubai Harvard Foundation for Medical Research (R.S.G. and L.D.N.), the Swiss National Science Foundation (PASMP3-127678/1 to M.R.), the Clinical Immunology Society (E.J.) and the Manton Foundation (E.J. and L.D.N.). AUTHOR CONTRIBUTIONS H.H.J., D.R.M., E.J., M.J.M., N.R., A.B., I.Ra., H.B. and M.R. did the experiments; L.S., S.B., R.W., G.D., M.D., W.A.-H., I.B., S.B., N.K., H.D.O., A.P., M.K., G.L. and I.Re. provided patient blood samples; K.A.F. and D.G. provided mice; J.M. and L.D.N. provided advice; S.K., R.C. and T.A.C. sequenced DNA; and H.H.J. and R.S.G. wrote the manuscript. COMPETING FINANCIAL INTERESTS The authors declare no competing financial interests.
1. Yoshida, T. et al. Memory B and memory plasma cells. Immunol. Rev. 237, 117139 (2010). 2. Huggins, J. et al. CpG DNA activation and plasma-cell differentiation of CD27 naive human B cells. Blood 109, 16111619 (2007). 3. Jiang, W. et al. TLR9 stimulation drives naive B cells to proliferate and to attain enhanced antigen presenting function. Eur. J. Immunol. 37, 22052213 (2007). 4. Giordani, L. et al. IFN-alpha amplifies human naive B cell TLR-9-mediated activation and Ig production. J. Leukoc. Biol. 86, 261271 (2009). 5. Carpenter, E.L., Mick, R., Ruter, J. & Vonderheide, R.H. Activation of human B cells by the agonist CD40 antibody CP-870,893 and augmentation with simultaneous toll-like receptor 9 stimulation. J. Transl. Med. 7, 93 (2009). 6. Katsenelson, N. et al. Synthetic CpG oligodeoxynucleotides augment BAFF- and APRIL-mediated immunoglobulin secretion. Eur. J. Immunol. 37, 17851795 (2007). 7. Weeratna, R.D., Makinen, S.R., McCluskie, M.J. & Davis, H.L. TLR agonists as vaccine adjuvants: comparison of CpG ODN and resiquimod (R-848). Vaccine 23, 52635270 (2005). 8. Zhang, X.Q. et al. Potent antigen-specific immune responses stimulated by codelivery of CpG ODN and antigens in degradable microparticles. J. Immunother. 30, 469478 (2007). 9. Cunningham-Rundles, C. & Bodian, C. Common variable immunodeficiency: clinical and immunological features of 248 patients. Clin. Immunol. 92, 3448 (1999). 10. Latz, E. et al. Ligand-induced conformational changes allosterically activate Toll-like receptor 9. Nat. Immunol. 8, 772779 (2007). 11. Pasare, C. & Medzhitov, R. Control of B-cell responses by Toll-like receptors. Nature 438, 364368 (2005). 12. Guay, H.M., Andreyeva, T.A., Garcea, R.L., Welsh, R.M. & Szomolanyi-Tsuda, E. MyD88 is required for the formation of long-term humoral immunity to virus infection. J. Immunol. 178, 51245131 (2007). 13. Hou, B. et al. Selective utilization of Toll-like receptor and MyD88 signaling in B cells for enhancement of the antiviral germinal center response. Immunity 34, 375384 (2011). 14. Meyer-Bahlburg, A., Khim, S. & Rawlings, D.J. B cell intrinsic TLR signals amplify but are not required for humoral immunity. J. Exp. Med. 204, 30953101 (2007). 15. Gavin, A.L. et al. Adjuvant-enhanced antibody responses in the absence of Toll-like receptor signaling. Science 314, 19361938 (2006). 16. Park, S.M. et al. MyD88 signaling is not essential for induction of antigen-specific B cell responses but is indispensable for protection against Streptococcus pneumoniae infection following oral vaccination with attenuated Salmonella expressing PspA antigen. J. Immunol. 181, 64476455 (2008). 17. Seibert, S.A., Mex, P., Kohler, A., Kaufmann, S.H. & Mittrucker, H.W. TLR2-, TLR4- and Myd88-independent acquired humoral and cellular immunity against Salmonella enterica serovar Typhimurium. Immunol. Lett. 127, 126134 (2010). 18. Sin, J.I. MyD88 signal is required for more efficient induction of Ag-specific adaptive immune responses and antitumor resistance in a human papillomavirus E7 DNA vaccine model. Vaccine 29, 41254131 (2011). 19. Browne, E.P. & Littman, D.R. Myd88 is required for an antibody response to retroviral infection. PLoS Pathog. 5, e1000298 (2009). 20. Meller, N., Merlot, S. & Guda, C. CZH proteins: a new family of Rho-GEFs. J. Cell Sci. 118, 49374946 (2005). 21. Cte, J.F. & Vuori, K. GEF what? Dock180 and related proteins help Rac to polarize cells in new ways. Trends Cell Biol. 17, 383393 (2007). 22. Randall, K.L. et al. Dock8 mutations cripple B cell immunological synapses, germinal centers and long-lived antibody production. Nat. Immunol. 10, 12831291 (2009). 23. Zhang, Q. et al. Combined immunodeficiency associated with DOCK8 mutations. N. Engl. J. Med. 361, 20462055 (2009). 24. Engelhardt, K.R. et al. Large deletions and point mutations involving the dedicator of cytokinesis 8 (DOCK8) in the autosomal-recessive form of hyper-IgE syndrome. J. Allergy Clin. Immunol. 124, 12891302 (2009).
npg
NUMBER 6
JUNE 2012
619
Articles
25. Bonilla, F.A. et al. Practice parameter for the diagnosis and management of primary immunodeficiency. Ann. Allergy Asthma Immunol. 94, S1S63 (2005). 26. Broder, K. et al. Preventing tetanus, diphtheria, and pertussis among adolescents: use of tetanus toxoid, reduced diphtheria toxoid and acellular pertussis vaccines recommendations of the Advisory Committee on Immunization Practices (ACIP). MMWR Recomm. Rep. 55, 134 (2006). 27. Hartmann, G. et al. Delineation of a CpG phosphorothioate oligodeoxynucleotide for activating primate immune responses in vitro and in vivo. J. Immunol. 164, 16171624 (2000). 28. Cunningham-Rundles, C. et al. TLR9 activation is defective in common variable immune deficiency. J. Immunol. 176, 19781987 (2006). 29. Hartmann, G. & Krieg, A.M. Mechanism and function of a newly identified CpG DNA motif in human primary B cells. J. Immunol. 164, 944953 (2000). 30. Dedeoglu, F., Horwitz, B., Chaudhuri, J., Alt, F.W. & Geha, R.S. Induction of activation-induced cytidine deaminase gene expression by IL-4 and CD40 ligation is dependent on STAT6 and NFB. Int. Immunol. 16, 395404 (2004). 31. Arrighi, J.F., Rebsamen, M., Rousset, F., Kindler, V. & Hauser, C. A critical role for p38 mitogen-activated protein kinase in the maturation of human blood-derived dendritic cells induced by lipopolysaccharide, TNF-alpha, and contact sensitizers. J. Immunol. 166, 38373845 (2001). 32. Lim, W. et al. Distinct role of p38 and c-Jun N-terminal kinases in IL-10-dependent and IL-10-independent regulation of the costimulatory molecule B7.2 in lipopolysaccharide-stimulated human monocytic cells. J. Immunol. 168, 17591769 (2002). 33. Kawai, T. et al. Interferon-alpha induction through Toll-like receptors involves a direct interaction of IRF7 with MyD88 and TRAF6. Nat. Immunol. 5, 10611068 (2004). 34. Gotoh, K. et al. Selective control of type I IFN induction by the Rac activator DOCK2 during TLR-mediated plasmacytoid dendritic cell activation. J. Exp. Med. 207, 721730 (2010). 35. Avery, D.T. et al. B cell-intrinsic signaling through IL-21 receptor and STAT3 is required for establishing long-lived antibody responses in humans. J. Exp. Med. 207, 155171 (2010). 36. Diehl, S.A. et al. STAT3-mediated up-regulation of BLIMP1 Is coordinated with BCL6 down-regulation to control human plasma cell differentiation. J. Immunol. 180, 48054815 (2008). 37. Barton, G.M. & Kagan, J.C. A cell biological view of Toll-like receptor function: regulation through compartmentalization. Nat. Rev. Immunol. 9, 535542 (2009). 38. Matsuda, T. & Hirano, T. Association of p72 tyrosine kinase with Stat factors and its activation by interleukin-3, interleukin-6, and granulocyte colony-stimulating factor. Blood 83, 34573461 (1994). 39. Uckun, F.M., Qazi, S., Ma, H., Tuel-Ahlgren, L. & Ozer, Z. STAT3 is a substrate of SYK tyrosine kinase in B-lineage leukemia/lymphoma cells exposed to oxidative stress. Proc. Natl. Acad. Sci. USA 107, 29022907 (2010). 40. Avraham, H., Park, S.Y., Schinkmann, K. & Avraham, S. RAFTK/Pyk2-mediated cellular signalling. Cell. Signal. 12, 123133 (2000). 41. Xi, C.X., Xiong, F., Zhou, Z., Mei, L. & Xiong, W.C. PYK2 interacts with MyD88 and regulates MyD88-mediated NF-kappaB activation in macrophages. J. Leukoc. Biol. 87, 415423 (2010). 42. Bonnette, P.C. et al. Phosphoproteomic characterization of PYK2 signaling pathways involved in osteogenesis. J. Proteomics 73, 13061320 (2010). 43. Gauld, S.B. & Cambier, J.C. Src-family kinases in B-cell development and signaling. Oncogene 23, 80018006 (2004). 44. Haynes, M.P. et al. Src kinase mediates phosphatidylinositol 3-kinase/Akt-dependent rapid endothelial nitric-oxide synthase activation by estrogen. J. Biol. Chem. 278, 21182123 (2003). 45. Boudot, C. et al. Involvement of the Src kinase Lyn in phospholipase C-2 phosphorylation and phosphatidylinositol 3-kinase activation in Epo signalling. Biochem. Biophys. Res. Commun. 300, 437442 (2003). 46. Mcsai, A., Ruland, J. & Tybulewicz, V.L. The SYK tyrosine kinase: a crucial player in diverse biological functions. Nat. Rev. Immunol. 10, 387402 (2010). 47. Sanjuan, M.A. et al. CpG-induced tyrosine phosphorylation occurs via a TLR9-independent mechanism and is required for cytokine secretion. J. Cell Biol. 172, 10571068 (2006). 48. Turkson, J. et al. Stat3 activation by Src induces specific gene regulation and is required for cell transformation. Mol. Cell Biol. 18, 25452552 (1998). 49. Shi, C.S. & Kehrl, J.H. Pyk2 amplifies epidermal growth factor and c-Src-induced Stat3 activation. J. Biol. Chem. 279, 1722417231 (2004). 50. Green, N.M. & Marshak-Rothstein, A. Toll-like receptor driven B cell activation in the induction of systemic autoimmunity. Semin. Immunol. 23, 106112 (2011).
npg
620
VOLUME 13
NUMBER 6
JUNE 2012
nature immunology
ONLINE METHODS
Patients. Patients with established DOCK8 mutations (n = 10) or STAT3 mutations (n = 6) were studied. Informed consent was provided by adult donors or by the childrens parents or guardian. Protocols used for studies of humans were approved by the Committee on Clinical Investigations at Childrens Hospital, Boston. Cell preparation. Human PBMCs were isolated from heparinized blood obtained from patients and normal donors with Ficoll-Hypaque (GE Healthcare). Naive and total B cells were purified from PBMCs by negative selection with a B cell Isolation Kit II according to the manufacturers recommendations (Miltenyi Biotec) and contained >96% CD19+ cells. Cells were suspended in RPMI-1640 medium containing 10% heat-inactivated FCS (Hyclone), 2 mM l-glutamine, 50 g/ml streptomycin and 100 U/ml penicillin (medium). EBV-transformed cell lines from patients and controls were established as described51. Cell cultures. PBMCs (1 106 or 1.5 106 cells per ml for analysis of proliferation or immunoglobulin production, respectively) or B cells (1.5 106 cells per ml) were cultured with medium, CpG (ODN 2006 or ODN 2216; 0.1 M; Invivogen) or anti-CD40 (5 g/ml; 626; Santa Cruz Biotechnology) in the presence of recombinant human IL-21 (50 ng/ml; Miltenyi Biotech) or recombinant IL-4 (5 ng/ml; R&D Systems). The following kinase inhibitors were used: Syk inhibitor SYKINH-61 (a gift from F. Uckun); Pyk2 inhibitor tyrphostin A9 (AG 17; EMD Biosciences) and Src inhibitor PP2 (EMD Biosciences). Proliferation was assayed 4 d later by assay of the incorporation of [3H]thymidine. The production of IgM, IgG and IgE was measured in culture supernatants at 14 d by enzyme-linked immunosorbent assay (ELISA) as described52. CpGmediated upregulation of the expression of CD23 and CD86 was analyzed by flow cytometry after 48 h. IL-6 and IFN- in culture supernatants were measured after 24 h with ELISA kits from Invitrogen. Apoptosis was evaluated on day 3 after stimulation. IgM antibody specific to TT in serum was assayed with an ELISA kit from IBL International. Flow cytometry. The following monoclonal antibodies to human antigens, with the appropriate isotype-matched control antibodies, were used for staining: anti-CD3 (HIT3a), anti-CD19 (HIB19), anti-CD27 (M-T271), anti-CD23 (M-L233), anti-CD86 (2331(FUN-1)), anti-IgD (IA6-2) and mouse IgG1, -chain (MOPC-21; all from BD Biosciences); mouse IgG2a, -chain (eBM2a; eBioscience); and anti-CD123 (AC145) and anti-BDCA-4 (AD5-17F6; both from Miltenyi). Annexin Vfluorescein isothiocyanate was from Biovision. Data collected with a FACSCalibur (BD Biosciences) were analyzed with CellQuest software (BDBiosciences) or FlowJo software (TreeStar). Quantitative PCR. RNA was extracted from cultured B cells on day 5 with TRIzol (Invitrogen) and was reverse transcribed with Supercript II RT according to the manufacturers instructions (Invitrogen). The appropriate primers and an ABI Prism 7300 sequence detector (Applied Biosystems) were used for quantitative PCR analysis of AICDA. Normalization of RNA expression to the abundance of GAPDH RNA was calculated by the relative standard curve method as outlined in the manufacturers technical bulletin (Applied Biosystems). Immunoblot analysis. PBMCs, B cells or EBV-B cell lines (1 106 cells per condition) were left unstimulated or stimulated with CpG (2.5 M), IL-21 (10 ng/ml), IFN- (1000 U/ml; Biosource), recombinant IL-6 (2 ng/ml; R&D systems) or antibody to human IgM (10 g/ml; 109-005-043, Jackson ImmunoResearch). Total cell lysates were then separated by SDS-PAGE and transferred to nitrocellulose membranes. The following antibodies were used for immunoblot analysis: anti-DOCK8 (HPA003218, Sigma-Aldrich); anti-STAT3 (79D7), anti-BLNK (3587), anti-Pyk2 (5E2), anti-Lyn (C13F9), antibody to phosphorylated p38 (D3F9), antibody to IB phosphorylated at Ser32 and Ser36 (5A5), antibody to STAT3 phosphorylated at Tyr705 (D3A7), antibody to Syk phosphorylated at Tyr352 (2701), antibody to Pyk2 phosphorylated at Tyr402 (3291) and antibody to Src phosphorylated at Tyr416 (2101; all from Cell Signaling); anti-IB (sc-847), anti-IKK (Fl-419) and anti-hemagglutinin (12CA5; all from Santa Cruz Biotechnology); anti-Syk (4D10.1), anti-Src
(CT, NL19), anti-Myc (9E10) and antibody to phosphorylated tyrosine (4G10; all from Millipore). Results for densitometry of scanned bands were evaluated with Image J 1.440 software (US National Institutes of Health). Transient transfection and coimmunoprecipitation. A wild-type human DOCK8 construct was generated with a Myc tag by PCR amplification of pCR4-TOPO human DOCK8 cDNA (Open Biosystems) by standard cloning techniques. Lipofectamine LTX (Invitrogen) was used for transfection of 293T cells with plasmids encoding Myc-tagged DOCK8 and hemagglutinin-tagged MyD88 or MRTF-A (R. Treisman). After 16 h, cells were lysed with buffer containing 0.75% NP-40. Immunoprecipitation was done with an agarose conjugate of monoclonal antibody to hemagglutinin (HA-7; Sigma-Aldrich), followed by separation by SDS-PAGE, transfer to nitrocellulose membranes and immunoblot analysis with antiMyc, anti-MyD88 and anti-hemagglutinin. Anti-MyD88 (H00004615-PW2 pair; Novus) was used for immunoprecipitation and immunoblot analysis of MyD88 in EBV-B cells. Rac1-precipitation assay. RAC1 activation was assessed by precipitation from lysates with a fusion protein of glutathione S-transferase and the GTPase-binding domain of p21-activated kinase (Pierce Laboratories) as described53. Immunoblot analysis of precipitated samples and aliquots of the cell extracts for loading controls was then done with anti-Rac1 (89856D; Pierce Laboratories). Subcellular localization of TLR9 in CpG-stimulated B cells. EBV-B cells were stimulated for 90 min with CpG, then spun onto coverslips coated with polyd-lysine (50 g/ml; Sigma). Localization of TLR9 to endoplasmic reticulum was confirmed with ER Tracker Red marker and Alexa Fluor 488conjugated anti-TLR9 (green; 26C593.2; Imgenex). Cells were loaded with ER Tracker Red according to the vendors protocol (Invitrogen), were fixed in 4% paraformaldehyde, permeabilized with 0.04% saponin in 1% BSA54 and incubated with anti-TLR9 (2254; Cell Signaling). Coverslips were mounted on slides with Prolong Gold Antifade Reagent with DAPI mounting medium (Invitrogen). Cells on coverslips were fixed in 4% paraformaldehyde, permeabilized with 0.25% Triton X-100, blocked with 10% FCS in PBS, incubated for 1 h at 25 C with monoclonal antibody to EEA1 (early endosome marker; 1G11; Abcam) and monoclonal antibody to LAMP1 (late endosome-lysosome marker; H4A3; Abcam), then were washed and stained with Alexa Fluor 555conjugated antimouse IgG (red; A21424; Invitrogen). Counterstaining with anti-TLR9 and slide mounting was done as described above. All images were acquired with a Nikon Eclipse Ti inverted microscope and were processed with AR3.0 software (NIS-Elements) and/or Adobe Photoshop. Short hairpin RNAmediated knockdown of Pyk2 and Syk in CH12 cells. The pGIPZ lentiviral vectors encoding green fluorescent protein (GFP) as well as short hairpin RNA specific for mouse Pyk2 or Syk and the nonsilencing (scrambled) plasmid were from Open Biosystems. CH12 cells were transfected by electroporation with 0.5 g/ml of each plasmid, and transduced cells were selected with puromycin. Green fluorescent proteinpositive cells were subcloned and their expression of Pyk2 and Syk was analyzed by immunoblot. Mice. TLR9- and Myd88-deficient and the corresponding C57BL/6 and BALB/c wild-type control mice were used at 812 weeks of age according to the guidelines of the Animal Care Committee of Childrens Hospital of Harvard Medical School. Mouse cell cultures. B cells were purified from splenic cell suspensions by negative selection with CD43-conjugated magnetic beads from Miltenyi. Purified B cells (>95% B220+) were suspended in complete medium containing 50 M -mercaptoethanol and were cultured at a density of 1 10 6 cells per ml in medium or CpG ODN 1826 (3 g/ml; Invivogen). Proliferation was assayed day 4 by measurement of the incorporation of [3H]thymidine. IgG was measured in culture supernatants at day 6 by ELISA. For immunoblot analysis, CpG ODN 1826 (2.5 M) and recombinant mouse IL-21 (10 ng/ml; R&D Systems) were used. Statistical analysis. Student t-test was used for statistical analysis.
npg
doi:10.1038/ni.2305
nature immunology
51. Jabara, H.H., Fu, S.M., Geha, R.S. & Vercelli, D. CD40 and IgE: Synergism between anti-CD40 mAb and IL-4 in the induction of IgE synthesis by highly purified human B cells. J. Exp. Med. 172, 18611864 (1990). 52. Jabara, H.H., Brodeur, S.R. & Geha, R.S. Glucocorticoids upregulate CD40 ligand expression and induce CD40L-dependent immunoglobulin isotype switching. J. Clin. Invest. 107, 371378 (2001).
53. Gallego, M.D. et al. WIP and WASP play complementary roles in T cell homing and chemotaxis to SDF-1. Int. Immunol. 18, 221232 (2006). 54. Shen, X.Z., Lukacher, A.E., Billet, S., Williams, I.R. & Bernstein, K.E. Expression of angiotensin-converting enzyme changes major histocompatibility complex class I peptide presentation by modifying C termini of peptide precursors. J. Biol. Chem. 283, 99579965 (2008).
npg
nature immunology
doi:10.1038/ni.2305
Articles
Tespa1 is involved in late thymocyte development through the regulation of TCR-mediated signaling
Di Wang1,2,7, Mingzhu Zheng1,2,7, Lei Lei1,2, Jian Ji3, Yunliang Yao1,2, Yuanjun Qiu1,2, Lie Ma1,2, Jun Lou1,2, Chuan Ouyang1,2, Xue Zhang2,4, Yuewei He5, Jun Chi5, Lie Wang1, Ying Kuang5, Jianli Wang1, Xuetao Cao1,6 & Linrong Lu1,2
Signaling via the T cell antigen receptor (TCR) during the CD4+CD8+ double-positive developmental stage determines thymocyte selection and lineage commitment. Here we describe a previously uncharacterized T cellexpressed protein, Tespa1, with critical functions during the positive selection of thymocytes. Tespa1/ mice had fewer mature thymic CD4+ and CD8+ T cells, which reflected impaired thymocyte development. Tespa1 associated with the TCR signaling components PLC-g1 and Grb2, and Tespa1 deficiency resulted in attenuated TCR signaling, as reflected by defective activation of the ErkAP-1 and Ca 2+-NFAT pathways. Our findings demonstrate that Tespa1 is a component of the TCR signalosome and is essential for T cell selection and maturation through the regulation of TCR signaling during T cell development. Thymocyte development occurs through progressive steps of cellular differentiation, lineage commitment and selection. In the cortex of the thymus, immature CD4CD8 double-negative (DN) thymocytes (subcategorized as stages DN1DN4) develop into CD4+CD8+ double-positive (DP) thymocytes, which then give rise to CD4+ or CD8+ single-positive (SP) T cells1. This process is controlled by expression of the T cell antigen receptor (TCR) and signaling via the TCR. Progression beyond the DN3 stage depends on successful rearrangement of the gene encoding TCR and signaling via the preTCR, whereas differentiation from DP cell to mature SP cell depends on the expression and positive selection of TCRs. Only a few DP thymocytes that express a functional TCR are able to recognize self peptide in the context of major histocompatibility complex (MHC), which results in successful positive selection24. The extracellular milieu and the nature of the peptide-MHC complex presented by the thymic stromal cells heavily influence the duration and strength of TCR signaling, which in turn dictate positive selection, negative selection and lineage commitment4. The most proximal event associated with TCR ligation is the activation of tyrosine kinases, including Zap70, Itk, Lck and Fyn. After ligation of the TCR, the Src-family kinase Lck phosphorylates tyrosine residues in immunoreceptor tyrosine-based activation motifs of the invariant signaling protein CD3 complex, which is followed by the recruitment and activation of Zap70. Activated Zap70 phosphorylates the adaptor Lat, which results in the formation of the multimolecular signalosome complex known as the Lat signalosome, which includes PLC-1, Grb2, Vav, Gads, SLP-76 and Themis5,6. Recruitment to the Lat signalosome
1Institute
activates these signaling molecules, leading to subsequent activation of signaling cascades, including the mitogen-activated protein kinase Erk and calcium mobilization, which are hallmarks of TCR signaling and are critical for the survival, differentiation and maturation of thymocytes710. Deletion or manipulation of intracellular kinases such as Lck, Zap70, Tec-family kinases and Erk results in defective thymocyte selection4. Although many of the molecular components that regulate TCR signaling and thymocyte selection are well characterized, the interactions between these molecules and their assembly in signaling complexes have remained elusive. Additional signaling components, such as Themis1113 and LKB1 (ref. 14), have now been identified. Here we describe a previously uncharacterized protein, Tespa1, which we found to be involved in the positive selection of thymocytes and TCR-induced activation of T cells. We found that Tespa1 encodes a protein of 458 amino acids of unknown function that is highly conserved among vertebrate species. Tespa1 expression was tightly regulated during T cell development, with highest expression at the DP stage. Tespa1 deficiency resulted in failure of positive selection and substantially fewer of CD4+ and CD8+ SP cells in the thymus. Our data suggest that Tespa1 is a critical component of the TCR signalosome with an important role in thymocyte development through the regulation of TCR signaling. RESULTS Identification and cloning of Tespa1 Programmed gene expression is critical for T lineage differentiation as well as cell activation after stimulation. Through the use of
npg
of Immunology, Zhejiang University School of Medicine, Hangzhou, China. 2Program in Molecular and Cellular Biology, Zhejiang University School of Medicine, Hangzhou, China. 3Institute of Feed Science, Zhejiang University School of Medicine, Hangzhou, China. 4Department of Pathology and Pathophysiology, Zhejiang University School of Medicine, Hangzhou, China. 5Shanghai Research Center for Model Organisms, Shanghai, China. 6Institute of Immunology and National Key Laboratory of Medical Immunology, Second Military Medical University, Shanghai, China. 7These authors contributed equally to this work. Correspondence should be addressed to L.Lu (lu_linrong@zju.edu.cn). Received 17 January; accepted 28 March; published online 6 May 2012; doi:10.1038/ni.2301
560
VOLUME 13
NUMBER 6
JUNE 2012
nature immunology
Articles a
Homo Bos Mus Platypus Xenopus Homo Bos Mus Platypus Xenopus Homo Bos Mus Platypus Xenopus Homo Bos Mus Platypus Xenopus Homo Bos Mus Platypus Xenopus Homo Bos Mus Platypus Xenopus Homo Bos Mus Platypus Xenopus Homo Bos Mus Platypus Xenopus Homo Bos Mus Platypus Xenopus 58 50 58 56 60 115 90 117 114 119 162 136 164 173 160 222 196 224 230 218 282 255 283 247 256 342 315 341 402 375 384 462 435 444
Br ai n H ea r Li t ve r Lu ng Ly m p M hn us o c d Pa le e nc Sp rea le s e Th n ym us
c
Br ai n H ea r Ki t dn e In y te s Li tine ve r Lu ng Ly m p M hn us o c d Pa le e nc Sp rea le s e Th n ym us
Tespa1 Actb
Tespa1
Actb
d
Tespa1 expression (relative) 4 3 2 1 0 Thymic DN Thymic Thymic Thymic Splenic Splenic DP CD4+ SP CD8+ SP CD4+ CD8+ DAPI Merge
e
Tespa1-GFP DAPI Merge
f
0 min
Tespa1
Tespa1
DAPI
Merge
10 min
521 492 458
Figure 1 Structure, expression and subcellular localization of Tespa1. (a) Amino acid sequences of the mouse (Mus) isoform of Tespa1 (5830405N20Rik) and its human (Homo), bovine (Bos), platypus and Xenopus orthologs, aligned with the Clustal W program; color of highlighting indicates identity: blue, 100%; green, 80%; gray, 60%. (b) RNA blot analysis of Tespa1 RNA in various mouse tissues; Actb (encoding -actin) serves as a loading control throughout. (c) Semiquantitative PCR analysis of Tespa1 mRNA in various mouse tissues. (d) Real-time RT-PCR analysis of Tespa1 mRNA expression in various cell subsets sorted by flow cytometry; results are presented relative to Actb expression. (e) Fluorescence microscopy of HEK293 cells transiently transfected to express a fusion of Tespa1 and green fluorescent protein (Tespa1-GFP; green) and labeled with the nuclear dye DAPI (blue). Original magnification, 40. (f) Confocal microscopy of Jurkat cells labeled with anti-Tespa1 (red) and DAPI (blue) before (0 min) and after (10 min) stimulation with antihuman CD3 and antihuman CD28. Original magnification, 63. Data are representative of three experiments (error bars (d), s.d.).
a data-mining tool (the cDNA Digital Gene Expression Displayer developed by the Cancer Genome Anatomy Project of the National Cancer Institute), we compared expressed sequence tags from pools of lymphoid organs (including thymus, lymph nodes and spleen) with those from other tissues in mice. This analysis identified many genes with preferential expression in cells of the T lineage. Many of these genes encode molecules linked before to the development and function of T cells, including Rag1, Rag2, CD3e and Ifng. Twelve transcripts whose functions are not well characterized have been given designations beginning with Tse (for T cellspecific expression; Tse1Tse12). One of these, Tse5, encodes a molecule shown to be involved in T cell development; this gene has been called Skint1 (ref. 15). Tse1 has been characterized as encoding Themis, which is involved in thymocyte selection1113. Here we describe Tse3 (Mouse Genome Informatics accession code 5830405N20Rik), a candidate gene that encodes a previously uncharacterized protein with restricted expression patterns in lymphoid organs. We have called it Tespa1 (for thymocyte-expressed, positive selectionassociated 1) because of its role in thymocyte selection. We cloned full-length cDNA encoding Tespa1 from both mouse thymocytes and a human Jurkat T cell line. Mouse Tespa1 cDNA contained an open reading frame encoding a protein of 458 amino acids; human Tespa1 protein shared ~70% homology with the mouse protein. Alignment of orthologous vertebrate sequences showed that Tespa1 has an evolutionarily conserved amino terminus, but we identified no potential catalytic domains (Fig. 1a). Blast analysis identified no paralog gene encoding Tespa1 in the mouse or human genome. Next we analyzed tissue distribution of Tespa1 mRNA by RNA blot and RT-PCR analysis. Tespa1 mRNA expression was highly
restricted to lymphoid tissues, especially the thymus, and, to a lesser extent, spleen and lymph nodes (Fig. 1b,c). According to the Eurexpress Transcriptome Atlas Database for Mouse Embryo (an in situ hybridization database)16, Tespa1 also had restricted expression in fetal thymus (Supplementary Fig. 1a). By RT-PCR analysis, mouse Tespa1 expression was also detectable in various lymphoid cells, including B cells and mast cells (Supplementary Fig. 1b). Realtime PCR analysis showed that Tespa1 expression was highest in DP thymocytes and decreased with maturation (Fig. 1d). Expression of Tespa1 was slightly higher in CD4+ SP cells than in CD8+ SP cells (Fig. 1d). In addition, Tespa1 expression was upregulated after TCR stimulation in splenic CD4+ and CD8+ cells (Supplementary Fig. 1c), which indicated that Tespa1 expression may be regulated by TCR signaling. An overexpressed fusion of Tespa1 and green fluorescent protein showed cytoplasmic localization in HEK293 human embryonic kidney cells (Fig. 1e), which correlated with the mainly cytoplasmic localization of Tespa1 in Jurkat cells (Fig. 1f). Together these experiments identified Tespa1 as a gene with high expression in thymocytes that encodes protein that might have a role in T cell development and function. Defective T cell development in Tespa1/ mice To clarify the function of Tespa1 in T cell development, we generated Tespa1-deficient (Tespa1/) mice by targeting the second and third exons of Tespa1 and replacing them with a neomycine-resistance cassette (Supplementary Fig. 1d). Tespa1 mRNA and Tespa1 protein were undetectable in thymocytes and splenocytes from Tespa1/ mice (Supplementary Fig. 1e,f). We found that total cell numbers were not altered substantially in the thymus of Tespa1/ mice, and the
561
npg
Articles
Thymic DN cells (106)
10 10 10
3 2 1 0
3.02 81.2
0.896 93.4
Thymic CD4+ SP cells (106)
10 5 0
+/+
20 15 10 5 0 Tespa1+/+ Tespa1/ Tespa1/ C M M C M
20 15 10 5 0 Tespa1+/+ Tespa1/
Thymic CD8+ SP 6 cells (10 )
Tespa1+/+
Tespa1/
Thymic DP cells (107)
a
CD8
15
NS
NS 25
25
8 6 4 2 0
10 10
2.80
0 1 2
12.1
3 4
1.15
4.15
10 10 10 10 10
CD4
Tespa1
Tespa1/
/ Tespa1+/+ Tespa1
Tespa1+/+
Tespa1/
c
M
Tespa1+/+
Tespa1+/+ (CD45.2+) Rag1/

82.3 8.45
Tespa1 (CD45.2+)
4.27 78.5
+/+
C M C CD4 CD8
10
4
1.05
16.2
/
CD4
CD8
d
CD8
Tespa1+/+
Tespza1/
Tespa1 (CD45.2+) Rag1/

10
4
Tespa1 (CD45.2+)
10
4
3 3 3 10 10 10 Figure 2 Defective T cell development in Tespa1 / 7.36 5.65 2 2 2 6.94 10 10 10 mice. (a) Surface staining of CD4 and CD8 (left) on 82.6 69.3 1 1 1 10 10 10 Tespa1 +/+ and Tespa1 / thymocytes. Numbers in or 15.2 6.64 0.697 4.09 0 0 adjacent to outlined areas (or in quadrants) indicate 10 0 100 0 10 0 1 2 3 4 1 2 3 4 1 2 3 4 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 percent cells in each throughout. Right, quantification CD4 CD45.1 CD4 of DN, DP, CD4 + or CD8 + SP thymocyte subpopulations. FTOC (day 8) Each symbol represents an individual mouse; small horizontal lines indicate the mean (and s.d.). NS, not significant; *P < 0.01. (b) Thymus sections from Tespa1+/+ and Tespa1/ mice, stained with hematoxylin and eosin; the cortex (darker) and medulla (lighter) can be distinguished by the intensity of staining. Original magnification, 4 (main images) or 10 (insets). (c) Staining of T cells with anti-CD4 (green) and anti-CD8 (red) in frozen sections of Tespa1 +/+ and Tespa1 / thymus. M, medulla; C, cortex. Original magnification, 10. (d) Flow cytometry of cells from fetal thymic organ culture (FTOC) at day 8, stained with anti-CD4 and anti-CD8. (e) Flow cytometry of thymus cells from bone marrow chimeras deficient in recombination-activating gene 1 (Rag1 /) 8 weeks after transfer () of bone marrow cells from wild-type C57BL/6 (CD45.1 +) mice and Tespa1 +/+ or Tespa1 / (CD45.2 +) mice, assessed for expression of CD4 and CD8 (right) after gating on CD45.1 or CD45.2 (left). Data are pooled from three independent experiments with thirteen mice per group (a) or are representative of three experiments (b,c), three separate experiments (d) or four separate experiments with three mice per group (e).
80.2
0.872
94.3
CD45.2
proportion of DN1DN4 cells was normal (Supplementary Fig. 2a), which indicated that -selection in these mice was unaffected. In contrast, we observed considerably fewer CD4+ or CD8+ SP cells in thymus from Tespa1/mice than in thymus from their Tespa1+/+ and Tespa1+/ littermates (controls; Fig. 2a and data not shown). The architecture of the thymus was also considerably altered in Tespa1/ mice, as characterized by much smaller areas of the medulla. We also observed fewer SP thymocytes in the medulla of Tespa1/ thymus stained with antibody to CD4 (anti-CD4) and anti-CD8 (Fig. 2b,c). We observed no difference in expression of the chemokine receptor CCR7 on either CD4+CD8int or CD4+ SP thymocytes in Tespa1/ thymus (Supplementary Fig. 2b), which indicated no alteration in the regulation of thymic cortex-to-medulla migration 17 of these cells. As assessed by fetal thymic organ culture18, there was less generation of the CD4+ SP population in the absence of Tespa1 (Fig. 2d), which suggested a requirement for Tespa1 in late stages of T cell development. We generated chimeric mice by transferring Tespa1+/+ or Tespa1/ (CD45.2+) bone marrow cells along with bone marrow cells from wildtype C57BL/6 (CD45.1+) mice, at a ratio of 4:1, into irradiated host mice deficient in recombination-activating gene 1. The thymus of reconstituted mice had a lower proportion of both CD4+ SP cells and CD8+ SP cells derived from Tespa1/ cells than those derived from Tespa1+/+ cells (Fig. 2e), which indicated an intrinsic defect in Tespa1/ thymocytes during development from the DP stage to the SP stage. Although Tespa1 was also expressed in cells of other lymphoid lineages, especially in B and mast cells, Tespa1 deficiency did not affect the number of T cells or B cells or the generation of bone marrowderived mast cells (Supplementary Fig. 2ce). Thus, Tespa1 was essential for the late development of thymocytes but not cells of other types.
562
Positive selection requires Tespa1 To determine the specific developmental stage at which thymocytes were blocked in Tespa1/ mice, we quantified cells at five distinct developmental stages defined by the expression of TCR and the activation marker CD69. We found that Tespa1/ mice and their wild-type littermates did not have significantly different numbers of immature TCR loCD69lo cells (population 1; mostly DN and DP cells) or TCRintCD69lo cells (population 2; preselection DP cells). However, Tespa1/ mice had significantly fewer cells in population 3 (defined by higher expression of TCR and CD69 than that in population 2; thymocytes undergoing selection). We also observed fewer TCRhiCD69hi cells (population 4; post-positive-selection thymocytes) and TCRhiCD69lo cells (population 5; mature SP cells ready for export to the periphery) in Tespa1/ mice (Fig. 3a). Because upregulation of CD69 expression is a critical marker of successful positive selection, we further evaluated the surface expression of CD69 on Tespa1+/+, Tespa1+/ and Tespa1/ thymocytes (Fig. 3bd). We found that CD69 expression in DP cells and CD4+CD8int cells was downregulated by Tespa1 deletion in a dose-dependent manner. Tespa1/ mice also had considerably fewer CD4 +CD69hi and CD8+CD69hi thymocytes than did their littermates. These data indicated that Tespa1/ thymocytes were blocked at the initial and early stage of positive selection. However, surface expression of the adhesion molecule CD5 was equivalent in DP and CD4+CD8int thymocytes from wild-type and Tespa1/ mice (Fig. 3e). Because developing thymocytes depend on TCR signaling for survival through direct induction of the interleukin 7 receptor (IL-7R) and the prosurvival factor Bcl-2, we analyzed the expression of those two proteins in cells from Tespa1/and wild-type mice. We found that surface expression of IL-7R was lower
npg
CD8
Articles a
Tespa1+/+
12
4
Events (% of max)
1 (14)
3 4 5
10 3 10 2 10 1 10 0 10 2 3 4 0 1 10 10 10 10 10
4
2 (66.3)
3 (5.36)
4 (5.72)
5 (5.37)
Events (% of max)
TCR-
CD4
15 10 5 0
+ hi 6 CD4 CD69 cells (10 )
+ hi 6 CD8 CD69 cells (10 )
c
6 4 2 0 +/+ /
d
1.0 0.8 0.6 0.4 0.2 0
DP 20
+ int CD4 CD8
CD4SP
41.6
4 3
Mat/imm
Population 1 NS 80 60 40 20 0 +/+ /
Population 2 Population 3 NS 1.5 * 1.0 0.5 0 +/+ / +/+ /
Population 4 0.4 0.3 0.2 0.1 0 +/+
Population 5 0.5 0.4 0.3 0.2 0.1 0
Cells (107)
**
**
100 80 60 40 20 0 0 1 2 3 4 10 10 10 10 10
DP +/+ /
CD4 CD8 +/+ /
Events (% of max)
10 3 10 2 / 3 4 10 Tespa1 1 10 5 12 0 10 0 1 2 3 4 10 10 10 10 10
(17.9)
(74.5)
(2.86)
(1.43)
(1.38)
100 +/+ 80 / 60 40 20 0 4 0 1 2 3 10 10 10 10 10
DP
Events (% of max)
CD4+CD8int +/+ /
100 80 60 40 20 0
0 1
DP
+/+ /
CD4+CD8int +/+ /
10 10 102 103 10
CD69
+ int
g
100 80 60 40 20 0
IL-7R DP CD4+CD8int +/+ /
CD69
CD8
+/+ /
CD5
0 1 10 10 10 10 10
Bcl-2
+/+
Tespa1+/+
Tespa1/
**
*
CD69 (MFI)
15 10 5 0 +/+ +/ /
CD8SP 102 CD24

10
1
Mat/imm
+/+
90 80 70 60 50 40
*
10 10
31.6
0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0
+/+ / CD4SP
**
+/+ / CD8SP
+/+
+/
100
38.2
29.0
2
10 10 10 103 104
TCR-
Figure 3 Impaired development of SP thymocytes in Tespa1/ mice. (a) Flow cytometry analysis of the surface expression of CD69 and TCR (far left) on Tespa1+/+ and Tespa1/ thymocytes (top two rows); numbers in outlined areas at left indicate subpopulations gated at right, and numbers in parentheses above plots at right indicate percent thymocytes in each subpopulation. Below, quantification of each subpopulation. (b) Surface staining of CD69 on gated CD4+CD8int and DP thymocytes from Tespa1+/+ (+/+) or Tespa1/ (/) mice. (c) Quantification of CD4+CD69hi and CD8+CD69hi cells. (d) Mean fluorescence intensity (MFI) of surface CD69 staining on gated DP or CD4 +CD8int cells. (eg) Surface staining of CD5 (e) and IL-7R (f) or intracellular staining of Bcl-2 (g) on or in gated DP or CD4+CD8int thymocytes from Tespa1+/+ and Tespa1/ mice. (h) Surface staining of TCR- and CD24 on gated CD4+ SP (CD4SP) or CD8+ SP (CD8SP) thymocytes. Numbers adjacent to outlined areas indicate percent cells in the CD24 loTCRhi gate. Right, ratio of mature (CD24neglo) cells to immature (CD24hi) cells (Mat/imm). Each symbol (a,c,d) represents an individual mouse; small horizontal lines indicate the mean (and s.d.). *P < 0.01 and **P < 0.001 (t-test). Data are representative of two independent experiments with eight (a) or five (c) mice per genotype or five mice per group (d), three independent experiments (b,h), or three independent experiments with three mice per group (eg).
in Tespa1/ CD4+CD8int thymocytes (Fig. 3f ), whereas expression of Bcl-2 was not substantially altered in CD4+CD8int thymocytes (Fig. 3g). CD4+ or CD8+ SP thymocytes from both wild-type and Tespa/ mice also had similar surface expression of IL-7R (Supplementary Fig. 2g), and in vitro stimulation of CD4+ or CD8+ SP cells with recombinant IL-7 similarly upregulated Bcl-2 expression in Tespa1+/+ and Tespa1/ thymocytes (Supplementary Fig. 2h), which indicated that IL-7R responses were unaltered in Tespa1/ mice and that the lower number of CD4+CD8int and SP thymocytes in Tespa1/ mice reflected a defect secondary to impaired TCR signaling. In addition, the maturation of CD4+ or CD8+ SP thymocytes was considerably impaired in Tespa1/ mice, as shown by the lower ratio of mature (CD24lo) TCRhiCD4+ or TCRhiCD8+ thymocytes to immature (CD24hi) TCRhiCD4+ or TCRhiCD8+ thymocytes, respectively (Fig. 3h). After the initiation of positive selection, DP thymocytes downregulate their expression of CD4 and CD8, passing through a CD4+CD8int transitional stage before committing to the CD4+ or CD8+ lineage1921. To determine whether Tespa1 deficiency leads to defects during this stage, we did a two-stage differentiation assay in vitro22. When we stimulated purified wild-type or Tespa1/ DP thymocytes with anti-TCR and anti-CD2, the downregulation of CD8 was attenuated in Tespa1/ DP thymocytes, and fewer CD4+CD8int cells were generated in the recovery culture of Tespa1/ DP thymocytes (Fig. 4a). Together these findings suggested a specific requirement for Tespa1 at the positive selection stage of thymocyte development.
Impaired positive selection in TCR-transgenic Tespa1/ mice As developmental defects associated with thymocyte selection can be masked by compensatory changes in the TCR repertoire, we introduced single transgenes encoding TCRs into Tespa1/ mice to limit such compensation. We then analyzed the effect of Tespa1 deficiency in mice expressing a transgene encoding the MHC class IIrestricted OT-II TCR (OT-II Tespa1/ mice) or the MHC class Irestricted H-Y TCR (H-Y Tespa1/ mice). The thymocyte-development defect in Tespa1-deficient mice was much greater on the OT-II background than in nontransgenic polyclonal mice, with very few cells proceeding to the CD4+ SP stage (Fig. 4b). Female H-Y Tespa1/ mice had significantly fewer total CD8+ SP thymocytes and H-Y TCRexpressing CD8+ SP thymocytes that stained with the T3.70 monoclonal antibody to the H-Y TCR (T3.70hi) than did female H-Y Tespa1+/+ mice (Fig. 4c and Supplementary Fig. 2i). We also introduced a transgene encoding another MHC class Irestricted TCR, OT-I, onto the Tespa1-deficient background. We found significantly fewer CD8+ SP thymocytes in OT-I Tespa1/ thymus than in OT-I Tespa1+/+ thymus (Fig. 4d). We also observed lower expression of both IL-7R and CD5 on DP and CD4+CD8int cells from female H-Y Tespa1/ mice than on H-Y Tespa1+/ DP and CD4+CD8int thymocytes (Fig. 4e,f). We also evaluated alterations in thymic gene-expression profiles in the absence of Tespa1 by microarray analysis. Among the genes with altered expression, we identified 88 genes encoding molecules involved in seven different physiological processes (including the cell cycle, signal transduction and transcription regulation), which suggested abnormal thymocyte development in the absence of Tespa1
563
npg
Articles a
0.458
Stimulatory (20 h) Medium

98.1 10.0 36.5
Recovery (40 h) Medium

26.6 20.9
b
CD8
Anti-TCR + anti-CD2
104 10
2
OT-ll Tespa1
+/
OT-ll Tespa1
1.05
103 0.721
51.8
HY Tespa1 104 34.8 26.1 3 10

102 10
1
+/
HY Tespa1
19.8
28.1
104 103 102 101 10

0
OT-l Tespa1
15.4 56.7
+/
OT-l Tespa1
8.26 66.7
82.5
CD8
CD8
Tespa1
+/+
101 100 0 2 3 4 10 101 10 10 10

31.7 3.07
0.353
1.08
44.4
9.14
37.4
15.1
100 0 1 2 3 4 10 10 10 10 10
35.6
3.49
43.8
8.31
6.63
20.4
10.8
12.4
10 10 10 10 10
Tespa1
102
900 600 300
CD8SP cells 7 (10 )
10
4.15
1,200
Total thymocytes 7 (10 )
10
CD4
0.296 98.6 3.01 52.8 4.06 32.7
CD4 20.2 11.1
CD4 10 8 6 4 2 0 DP +/ / 1.5 1.0 0.5 0 CD4 CD8

+ int
CD4
101 10
0
0.402
1 2
0.678
17.2
27.0
32.8
30.4
100 10 10 103 104
1,000 52.7 800 600 400 200 0 0 1 2 3 4 10 10 10 10 10
+/ /
CD8
Total thymocytes (107)
V2
CD4SP cells (107)
0 0 1 2 3 4 10 10 10 10 10
T3.70
Total thymocytes (107) CD8SP cells 7 (10 )
Events (% of max)
20
15 15 Figure 4 Defective positive selection in Tespa1/ * 10 10 +/+ or mice. (a) Staining of CD4 and CD8 on Tespa1 5 5 / DP thymocytes sorted by flow cytometry Tespa1 ** 0 0 and left unstimulated (medium alone; left) or stimulated overnight with anti-TCR and anti-CD2 and analyzed immediately (middle) or washed and incubated for an additional 24 h in medium without stimulation (right). (bd) Flow cytometry of thymocytes from Tespa1+/+ or Tespa1/ mice expressing a transgene encoding the MHC class IIrestricted OT-II TCR (b) or the MHC class I restricted H-Y TCR (c) or OT-I TCR (d). Top row, staining of CD4 and CD8 on total thymocytes (b,d) or on a gated H-Y-specific T3.70hi population (female () mice; c). Middle row (b,c), staining with antibody to the OT-II-specific variable region V5 (b) or the H-Y-specific antibody T3.70 (c); numbers above bracketed lines indicate percent V5+ cells (b) or T3.70+ cells (c). Bottom row, quantification of total or SP thymocytes. *P < 0.05 and **P < 0.01 (t-test). (e,f) Surface staining of IL-7R (e) or CD5 (f) on gated DP and CD4+CD8int thymocytes from female H-Y Tespa1+/+ and Tespa1/ mice. Data are representative of three independent experiments (ad; average and s.e.m. in bd) or four experiments (e,f).
+/ /
6 5 4 3 2 1 0
+/ /
20
+/ /
6 5 4 3 2 1 0
+/ /
H-Y
100 +/ 80 / 60 40 20 0 100 101 102 103 104
+/ /
IL-7R
Events (% of max)
100 +/ 80 / 60 40 20 0 100 101 102 103 104
DP
H-Y
+ int CD4 CD8
+/ /
CD5
(Supplementary Fig. 3a). We further did transcriptome sequencing analysis and compared the abundance of annotated or nonannotated transcripts in CD69+ thymocytes from wild-type and Tespa1/ mice. These data showed that several genes encoding molecules involved in positive selection were either induced or repressed in the absence of Tespa1 (Supplementary Fig. 3be and Supplementary Table 1). Because immature DP thymocytes are highly susceptible to apoptotic stimuli23, we investigated the effect of various apoptotic stimuli on Tespa1/ thymocytes. Exposure of thymocytes to anti-CD3 aggregates the TCR-CD3 complex and thereby results in the lysis of vulnerable DP thymocytes24. DP thymocytes from both wild-type and Tespa1/ mice had similar apoptotic sensitivity to anti-CD3 in vivo (Supplementary Fig. 4a). TCR-induced expression of Nur77, a critical regulator of TCR-induced cell death24,25, was unaltered in the absence of Tespa1 (Supplementary Fig. 4b). In addition, specific deletion of V8+ CD4+ SP and V8+ CD8+ SP thymocytes induced by the superantigen Staphylococcus enterotoxin B was similar in the presence or absence of Tespa1 expression (Supplementary Fig. 4c). We detected no substantial difference in the number of apoptotic cells in wild-type and Tespa1/ thymus in vivo by the TdT-mediated dUTP nick-end-labeling apoptosis assay (Supplementary Fig. 4f). In vitro, DP and SP thymocytes from wild-type and Tespa1/ mice underwent a similar degree of apoptosis in response to stimulation with anti-CD3 and anti-CD28 or treatment with TNF or the cell-surface receptor Fas (CD95; Supplementary Fig. 4d,e). Furthermore, we found no obvious cell-cycle abnormalities in Tespa1/ total and DP thymocytes by staining with the DNA-intercalating dye DAPI (Supplementary Fig. 4g). Together these data indicated that Tespa1 was required for the positive selection of thymocytes expressing either MHC class I or MHC class IIrestricted TCRs but was dispensable for the survival or apoptosis of thymocytes. Altered peripheral T cells in Tespa1/ mice Tespa1 deficiency led to fewer peripheral T cells in spleen (Fig. 5a). Tespa1-deficient OT-II and OT-I mice also had fewer mature peripheral CD4+ cells and CD8+ cells, respectively (Fig. 5b). We also
564
observed considerably fewer mature peripheral total CD8+ and T3.70hi CD8+ cells in female H-Y Tespa1/mice (Fig. 5c and Supplementary Fig. 2j). We found fewer peripheral regulatory (CD4+CD25+Foxp3+) T cells in Tespa1/ mice than in wild-type mice, although the decrease in Foxp3+ regulatory T cells was less extensive in Tespa1-deficient spleen than in Tespa1-deficient thymus (Supplementary Fig. 5a,b). Peripheral CD4+ or CD8+ T cell populations from Tespa1/ mice had a greater frequency of memory (CD44hiCD62Llo) cells than did those from wild-type mice (Fig. 5d). Consistent with that phenotype, splenic CD4+ or CD8+ cells in Tespa1/ mice had more uptake of the thymidine analog BrdU in vivo (Fig. 5e). Together these experiments showed that deficiency in Tespa1 not only led to fewer mature T cells in the peripheral but also resulted in a phenotype for Tespa1/ T cells reminiscent of that of proliferating and memory cells, a frequent observation in partially lymphopenic mice that reflects homeostatic expansion of T cell populations. Impaired TCR-driven thymocyte differentiation in Tespa1/ mice Because positive selection was impaired in Tespa1/ thymocytes, we considered the possibility that TCR signaling may be affected in the absence of Tespa1. Strong and/or persistent TCR signaling drives the maturation of CD4+ SP thymocytes, and attenuation or cessation of TCR signaling at the CD4+CD8int stage leads to the redirection of MHC class IIrestricted thymocytes into the CD8+ lineage2628. We crossed Tespa1/ mice with 2-microglobulin-deficient (B2m/) mice, which lack conventional CD8+ SP cells because of a deficiency in peptide presentation by MHC class I. We found CD3hi CD8+ SP thymocytes in B2m/Tespa1/ thymus and spleen but not in B2m/Tespa1+/+ thymus and spleen (Fig. 6a), which indicated redirection of MHC class IIrestricted cells to the CD8+ lineage. Moreover, CD69 expression was lower in CD4+CD8int cells from B2m/Tespa1/ mice (Fig. 6b). To further evaluate the contribution of Tespa1 deficiency to CD8+ thymocyte differentiation, we generated MHC class IIdeficient (H2-Ab1/) Tespa1/ mice, which lack mature CD4+ SP thymocytes. H2-Ab1/Tespa1/ mice had fewer CD8+ SP thymocytes and CD4+CD8int thymocytes than did H2-Ab1/Tespa1+/+ mice (Fig. 6c).
npg
Articles a
CD8
Tespa1+/+
0.062
Cells (%)
10
30 20 10 0
+/ + +/ + / /
Cells (10 )
104 13.2 102 101
Tespa1/
6.67 0.035
40
**
NS 20 15 10 5 0
6
+/+ /
b
OT-II Tespa1
+/
c
10 OT-II Tespa1
1.53
4
HY Tespa1
79.7
+/
HY Tespa1
29.5
0.183
0.438
*
+
103 102
2.49
0.085
0.069
*
CD8
D 8
+
Te sp a1 Te sp a1
ic
ic
CD4
Te sp a1 Te sp a1
100 0 10 101 102 103 104

56.2 30.5
82.5
10.8
101 100 0 10 101 102 103 104

9.48 10.7 32.2 37.9
4 C
10
Tespa1
75.8
+/+
CD4
2.46
Tespa1
50.0
CD4
6.49
CD62Lhi CD62lo CD44lo CD44hi
CD62L lo CD44
hi
CD62 hi CD44
lo
10
BrdU cells (%)
9.38
0.016
CD4SP (% of max)
d
103 10
2
Splenic CD4+ Splenic CD8+
le n
le n
70.7
26.7
86.9
11.5
Sp
Sp
OT-I Tespa1 104 27.9 0.070 102
+/
OT-I Tespa1
CD4
CD4+ cells (%)
CD8+ cells (%)
100 101 102 103 104
WT KO WT KO
WT KO WT KO
CD8SP (% of max)
10
9.83
11.9
7.60
35.9
CD4
BrdU cells (%)
CD44
Figure 5 Phenotype of peripheral T cells in mice. (a) Surface staining of CD4 and CD8 on Tespa1+/+ and Tespa1/ splenocytes (left), and frequency (middle) and total cell number (right) of splenocyte subpopulations. Each symbol (middle) represents an individual mouse; small horizontal BrdU-FITC lines indicate the mean. (b,c) Expression of CD4 and CD8 on total splenocytes from OT-II (b, top), OT-I (b, bottom) or H-Y (c; female) Tespa1+/+ and Tespa1/ mice, with gating on the T3.70hi population in c. (d) Expression of CD62L and CD44 on CD4+ splenocytes from Tespa1+/+ and Tespa1/ mice (top), and frequency of CD4+ cells in the CD62LhiCD44lo and CD62LloCD44hi subpopulations (bottom row; each symbol represents an individual mouse and small horizontal lines indicate the mean). (e) BrdU uptake by CD4+ or CD8+ cells among Tespa1+/+ and Tespa1/ splenocytes (left); numbers above bracketed lines indicate percent BrdU + cells. Right, frequency of BrdU+ CD4+ or CD8+ cells. *P < 0.05, **P < 0.01 and ***P < 0.001 (t-test). Data are representative of two independent experiments with seven mice per group (a) or three independent experiments (be; error bars (a,e), s.d.).
+
Tespa1/
In addition, H2-Ab1/Tespa1/ CD4+CD8int cells had lower expression of CD3 (Fig. 6d), which suggested abnormal TCR-driven differentiation of CD4+CD8int transitional cells into CD8+ cells in the absence of Tespa1. However, Tespa1 deficiency did not substantially alter the surface expression of CD5 on CD4+CD8int cells from B2m/ or H2Ab1/ mice (Fig. 6b,d). These observations suggested that attenuation or cessation of TCR signaling that might account for the impaired development of both CD4+ SP thymocytes and CD8+ SP thymocytes. Diminished TCR-mediated responses in Tespa1/ mice To investigate the function of Tespa1 in the TCR-mediated activation of T cells, we assessed the ability of Tespa1/ splenic T cells, including total splenocytes and sorted CD44loCD62Lhi splenic T cells, to respond to stimulation with anti-CD3 and anti-CD28. The upregulation of CD69 was lower in Tespa1-deficient CD4+ or CD8+ T cells than in wild-type T cells (Fig. 7a,b). Similarly, CD69 expression was much lower in OT-II Tespa1/ splenocytes and thymocytes (gated on Va2+V5+ CD4+ cells) in response to ovala MHCI-KO Tespa1+/+ bumin peptide (Supplementary Fig. 5ce). 0.302 88.9 In an assay of dilution of the cytosolic dye / CD4+ or CD8+ T cells had Thymus CFSE, Tespa1
2.90 7.89 0.05
a much lower proliferative capacity in response to stimulation with anti-CD3 alone or with anti-CD3 and anti-CD28 than did wild-type T cells (Fig. 7c). We obtained similar results for the incorporation of [3H]thymidine by CD4+ T cells that had been infected with a lentivirus expressing short hairpin RNA targeting Tespa1, followed by stimulation with anti-CD3 (Supplementary Fig. 5f). In addition, production of both IL-2 and IFN- by Tespa1/ splenic T cells was much lower after stimulation with anti-CD3 (Fig. 7d). Together these observations indicated that Tespa1 contributed to TCR-mediated responses. Interrupted TCR signaling in Tespa1/ thymocytes To determine whether Tespa1 functions downstream of the TCR, we analyzed intracellular signaling events in total thymocytes and sorted DP thymocytes from wild-type and Tespa1/ mice after TCR crosslinking with anti-CD3 plus anti-CD4 or with anti-CD3 plus the phorbol ester PMA. We observed no substantial difference in the
CD3 Tespa1/
0.896 95.2
npg
hi
Tespa1+/+
0.736
Tespa1/
11.9
b
MHCI-KO Events (% of max)
100 80 60 40 20 0
0 10 10 1
+ int CD4 CD8
0.681 5.37
3.26 0.033 2.96
80.1 0.401 31.7
52.1 0.462
+/+ /
100 80 60 40 20 0
+/+ /
Figure 6 Impaired MHC-restricted thymocyte differentiation in Tespa1/ mice. (a) Staining of CD4 and CD8 on total thymocytes or splenocytes (left) and on gated CD3hi thymocytes or splenocytes (right) from B2m/ (MHCI-KO) Tespa1+/+ or Tespa1/ mice. (b) Surface staining of CD69 and CD5 on CD4+CD8int thymocytes from mice with genotypes as in a. (c) Staining of CD4 and CD8 on total thymocytes or splenocytes from H2-Ab1/ (MHCII-KO) Tespa1+/+, Tespa1+/ or Tespa1/ mice. (d) Surface staining of CD3 and CD5 on CD4+CD8int thymocytes from mice with genotypes as in c. Data are representative of three independent experiments with five mice per group.
Spleen CD8
10 1.12 3 10 10 10 10
2 1 0
78.2
0 1 2
20.6
3 4
84.4
10.2
6.33
90.3
7.05
60.8
10 10 10 10 10
CD4
10
10
10
0 10 10
10
Br
3
10
MHClI-KO Thymus
Tespa1+/+
9.97 80.2 2.68
Tespa1+/
5.16 85.5 1.28
Tespa1/
3.30 80.8 0.682
MHClI-KO Events (% of max)

100 80 60 40 20 0
0 10 10 1
CD69 CD4+CD8int +/+ /

100 80 60 40 20 0
10
2
CD5
+/+ /
10 3 10
21.6 1.48
11.6 1.59
9.39 1.41
Spleen
10 10 10
2 1 0 0 1
CD8
10 10 10 10 10
10
10
10 101 102 10
CD4
CD3
CD5
NUMBER 6
JUNE 2012
d CU D 8
10
4 3
10
565
100 100 80 80 60 60 4.35 9.78 40 40 20 20 0 0 0 10 101 102 103 104 100 101 102 103 104
12 10 8 6 4 2 0
Br d CU D 4
+/+ /
101
80 60 40 20 0
CD8
**
**
100 80 60 40 20 0
***
***
10 10
0 68.5 0 1 2 3
3.56
89.3
1.29
CD62L
10 10 10 10 10
100 100 80 80 60 60 3.39 7.67 40 40 20 20 0 0 100 101 102 103 104 100 101 102 103 104
Tespa1+/+
Tespa1/
12 10 8 6 4 2 0
+/+ /
**
Articles a
Total splenocytes (% of max) 100 80 60 40 20 0 0 1 2 3 4 10 10 10 10 10 CD69 Med 100 80 60 40 20 0 100 101 102 103 104 CD69 CD4 +/+ 2.08 / 1.75
+
Med CD4 0.886 1.41

+
Anti-CD3 CD8 9.08 10.5

+
Events (% of max)
42.8 27.1
42.7 25.9
80 60 40 20
IL-2 (pg/ml)
+/+ /
+/+ /
CD4
+/+ /
CD8
c
+/+ / 100
CD4+ Anti-CD3 +/+ / Anti-CD3 + anti-CD28 +/+ /
250 200 150 100 50 0
+/+ /
0 100 101 102 103 104 CD8

+
05 0. 1 0. 2 0. 5 05 2 0. 0. 0. 5
0.
b
CD44 CD62L splenocytes (% of max)
hi
Anti-CD3 Events (% of max) CD8 +/+ 7.69 / 5.09

+
80 60 40 20
0 0 1 2 3 4 10 10 10 10 10 CFSE
IFN- (ng/ml)
CD4 +/+ 25.3 / 12.2
CD8 +/+ 18.2 / 9.55
100
Anti-CD3 +/+ /
Anti-CD3 + anti-CD28 +/+ /
7 6 5 4 3 2 1 0
0
+/+ /
lo
01
0.
01
Anti-CD3 (ug/ml)
0.
Anti-CD3 (ug/ml)
Figure 7 Defective activation of Tespa1-deficient T cells. (a,b) CD69 staining on gated CD4+ or CD8+ cells from Tespa1+/+ and Tespa1/ total splenocytes (a) or sorted CD44loCD62Lhi splenocytes (b) left unstimulated (Med) or activated by crosslinking for 5 h with anti-CD3 (Anti-CD3). Numbers above and below bracketed lines indicate percent CD69 + cells (colors match key). (c) Flow cytometry analysis of CFSE dilution in Tespa1+/+ and Tespa1/ CD4+ and CD8+ splenocytes left unstimulated (grey shaded curves) or stimulated for 72 h in vitro with anti-CD3 alone (left) or anti-CD3 and anti-CD28 (right). (d) Enzyme-linked immunosorbent assay of IL-2 and IFN- secreted by Tespa1+/+ and Tespa1/ spleen cells after stimulation for 3 d in vitro with anti-CD3. Data are representative of three independent experiments (mean s.d. in d).
phosphorylation of Lck, Zap70, PLC-1, SLP-76, Vav, the kinases p38, Jnk and Akt or the inhibitor IB (Fig. 8a and Supplementary Fig. 6be). However, both total thymocytes and sorted DP thymocytes from Tespa1/ mice had much less TCR-mediated phosphorylation of Erk than did those from wild-type mice (Fig. 8a and Supplementary Fig. 6d,e). In addition, analysis of total phosphorylated tyrosine showed a band of around 42 kilodaltons (presumably tyrosinephosphorylated Erk29,30) with consistently less phosphorylation in Tespa1/ cells (Supplementary Fig. 6a). Moreover, both DP thymocytes and CD4+CD8int thymocytes from Tespa1/ mice had much less TCR- or mitogen-driven calcium flux than did wild-type cells (Fig. 8b,c). In addition, in sorted DP thymocytes, Tespa1 deficiency resulted in much less TCR- or mitogen-mediated activity of luciferase reporters for the transcription factors AP-1 and NFAT but not for the transcription factor NF-B (Fig. 8d,e).
We also evaluated the expression of several target genes of TCR signaling downstream of the ErkAP-1 and Ca2+-NFAT pathways during positive selection. The induction of most of these genes, including Gimap4, Egr2, CD52, Bach2, Id3, Egr1 and Zbtb7b, was much lower in Tespa1/ DP thymocytes after stimulation with anti-CD3 and antiCD28 (Supplementary Fig. 7a). Microscopy of actin polymerization at the immunological synapse in thymocytes stimulated by beads coated with anti-CD3 and anti-CD28 showed there was no obvious difference between wild-type and Tespa1/ cells (Supplementary Fig. 7b). In sum, these data suggested that Tespa1 has an important role in TCR-mediated activation of the ErkAP-1 and Ca2+-NFAT signaling pathways. Tespa1 is a component of the TCR signaling machinery Because we observed no obvious abnormalities in the proximal events associated with TCR ligation, in contrast to the defects we found in
npg
Ca2+ ux (Fluo-4 uorescence)
a
Anti-CD3 + anti-CD4 (min) p-Erk1/2 Erk1/2 p-p38 p38 p-Jnk Jnk p-PLC- PLC- -actin Tespa1 0 5
+/+
b
Tespa1/ 20 0 5 10 20 10
Anti-CD3 + anti-CD28 DP 60 40 20 0 100 200 300 400 Time (s)

+ int CD4 CD8
PMA + ionomycin DP 80 60 40 20 0 100 200 300 Time (s) +/+ /
d
5 4 3 2 1 0
AP-1 (fold)
Anti-CD3 + anti-CD28 +/+ /

NFAT (fold)
0.55 0.97 0.93 1.47 0.62 0.73 0.65 0.95
60 40 20
+/+ /
0 100 200 300 400 Time (s)
100 200 300 Time (s)
Figure 8 Interruption of TCR-driven Erk activation and calcium signaling in Tespa1/ thymocytes. (a) Immunoblot analysis of total and phosphorylated (p-) Erk, p38, Jnk and PLC- in extracts of sorted Tespa1+/+ and Tespa1/ DP thymocytes left unstimulated (0) or stimulated for 5, 10 or 20 min with anti-CD3 and anti-CD4. Numbers below lanes (top) indicate densitometry of phosphorylated Erk, presented relative to total Erk expression in that same lane (below). -actin serves as a loading control. (b,c) Calcium flux in DP thymocytes (top row) or CD4+CD8int thymocytes (bottom row) from Tespa1+/+ and Tespa1/ mice after stimulation (downward arrows) with anti-CD3 and anti-CD28 (b) or PMA and ionomycin (c). (d,e) Luciferase activity in sorted DP thymocytes (d) or total thymocytes (e) from Tespa1+/+ and Tespa1/ mice, transfected by nucleofection with luciferase reporter constructs (vertical axes) and allowed to rest for 30 min, then stimulated for 6 h with anti-CD3 and anti-CD4 (d) or PMA and ionomycin (e); results were normalized to renilla luciferase activity and are presented relative to those of unstimulated cells. Data are representative of three independent experiments (mean and s.d. in d,e).
566
VOLUME 13
NUMBER 6
JUNE 2012
nature immunology
NF-B (fold)
AP-1 (fold)
NFAT (fold)
80 60 40 20
+ int CD4 CD8
+/+ /
e
14 12 10 8 6 4 2 0
PMA + ionomycin +/+ / 7 6 5 4 3 2 1 0 +/+ / 5 4 3 2 1 0 +/+ /
NF-B (fold)
+/+ /
3.0 2.5 2.0 1.5 1.0 0.5 0
+/+ /
5 4 3 2 1 0
+/+ /
Articles a
Time (min) IB: Themis 0.68 0.67 0.81 0.43 0.45 0.57 IP: PLC-1 IB: Grb2 0.78 0.88 1.22 0.49 0.69 0.80 IB: PLC-1 IB: Grb2 0.38 0.93 1.05 0.51 0.74 0.78 IP: Themis IB: PLC-1 0.48 0.98 0.99 0.31 0.82 0.54 IB: Themis 0 Tespa1+/+ 5 10 0 Tespa1/ 5 10
b
Time (min) IP: Tespa1 IB: PLC-1 IP: Tespa1 IB: Grb2 Lysate: IB: Tespa1 0
Jurkat Jurkat 5 10 IgG Flag-Tespa1 + Time (min) 0 IP: Flag IB: PLC-1 IP: Flag IB: Grb2 Lysate: IB: Flag + 5 + 10 + IgG
Figure 9 Involvement of Tespa1 in the Lat signalosome. (a) Immunoblot analysis (IB) of sorted Tespa1+/+ and Tespa1/ DP thymocytes left unstimulated (0) or stimulated for 5 or 10 min with anti-CD3 and anti-CD4, then lysed and immunoprecipitated (IP) with anti-PLC-1 (top) or anti-Themis (bottom), probed with anti-Themis, anti-Grb2 or anti-PLC-1. Numbers below lanes indicate densitometry values, presented relative to the density of the precipitated bait protein. (b) Immunoblot analysis of Jurkat cells stimulated as in a, then lysed and immunoprecipitated with anti-Tespa1 or mouse immunoglobulin G (IgG; control antibody) and probed with anti-PLC-1 or anti-Grb2 (left), or transfected with Flag-tagged Tespa1 and stimulated as in a, then lysed and immunoprecipitated with anti-Flag or mouse immunoglobulin G (control) and probed with anti-PLC-1 or anti-Grb2 (right). Below, immunoblot analysis of lysates (without immunoprecipitation) with anti-Flag. Data are representative of three independent experiments.
the distal events in Tespa1/ thymocytes, we speculated that Tespa1 deficiency might affect assembly of the Lat signalosome. The strength and duration of the TCR-induced association of PLC-1 with Themis and Grb2, but not of the interaction between Themis and the E3 ubiquitin ligase c-Cbl, were much lower in both total thymocytes and sorted DP thymocytes in the absence of Tespa1, as assessed by coimmunoprecipitation (Fig. 9a and Supplementary Fig. 8). There was also less association of Lat with Grb2, Themis and PLC-1 in Tespa1-deficient thymocytes (Supplementary Fig. 8). Together these observations suggested that Tespa1 deficiency impaired assembly of the Lat signalosome in thymocytes. Further coimmunoprecipitation analysis showed that Tespa1 associated with PLC-1 and Grb2 but not with Themis or Lat in Jurkat cells and that this association increased with prolonged stimulation of the TCR (Fig. 9b and data not shown). Similarly, overexpression of Flagtagged Tespa1 in Jurkat cells increased the interaction between Tespa1 and PLC-1 or Grb2 after TCR stimulation (Fig. 9b). Together these results suggested that engagement of the TCR led to the recruitment of Tespa1 to PLC-1 and Grb2, which in turn facilitated assembly of the Lat signalosome (proposed model, Supplementary Fig. 9). DISCUSSION Here we have identified a previously unknown protein, Tespa1, that was required for the development and maturation of SP thymocytes. The absence of Tespa1 did not affect the early stages of thymocyte development, as reflected by normal numbers of DN thymocytes. However, Tespa1 was essential for the development of DP thymocytes into mature SP cells. Tespa1 deficiency resulted in notably fewer
mature CD4+ or CD8+ SP thymocytes, which indicated a developmental block specifically at the positive selection stage in the development of thymocytes with an TCR (but not those with a TCR). The proposal of critical function for Tespa1 in positive selection was also supported by the findings of defects in the generation of CD4+CD8int thymocytes and much lower surface expression of CD69 on Tespa1/ DP and CD4+CD8int thymocytes. It is well known that the development and selection of thymocytes depends on TCR signaling and can be disrupted by defects in signaling components involved in the TCR signaling pathway. The positive selection defect observed in Tespa1/ mice suggested the potential involvement of Tespa1 in TCR signaling. It was notable that the defect in differentiation into SP cells was more profound for CD4+ T cells than for CD8+ T cells and was evident in T cells expressing either endogenous or transgenic TCRs. This could be explained by the present model of differentiation into the CD4+ or CD8+ lineage, in which the differentiation of CD4+ T cells relies on stronger and more-sustained activation of Erk and is thus more susceptible to defects in Erk activation7. Tespa1 was identified by its high expression in DP thymocytes. Structurally, Tespa1 resembles an adaptor because of the absence of a defined catalytic domain. Indeed, Tespa1 interacted with the TCR signaling components PLC-1 and Grb2. Tespa1 deficiency also led to less assembly of the TCR-driven Lat signalosome. For these reasons, we speculate that Tespa1 might serve as a scaffolding protein to promote the recruitment of PLC-1 (or other components) to the Lat signalosome. Such observations are consistent with the findings that Tespa1 deficiency did not affect the most proximal TCR signaling, including activation of Lck and Zap70, but led to changes in downstream signaling, including activation of the ErkAP-1 and Ca2+-NFAT pathways. Mouse Tespa1 protein is phosphorylated at Ser316 (S* in the sequence KNS*LDQIVWEVMDR) in spleen, as identified by mass-spectrum analysis on the Phosphomouse website. Published studies have reported that human Tespa1 protein extracted from Jurkat cells is phosphorylated at Ser454 (S* in the sequence QKNLMGRKVKS*LDLSITQQKW) and can be downregulated by stimulation of the TCR31. Such findings also support the proposal of a role for Tespa1 in T cell signaling, as the phosphorylation and dephosphorylation of proteins are closely related to the activities of signaling proteins and directly mediate protein-protein interactions. It is notable that Tespa1 and Themis share many features in common. First, both proteins have restricted high expression in thymocytes, and structurally they have no predicted conserved domains shared with other proteins. Second, both are involved in TCR signaling, and their influence on Erk activation and calcium flux is very similar after deletion. Third, both interact with the Lat signalosome component Grb2, although in our experience Tespa1 and Themis did not seem to interact directly with each other. Future studies should delineate the relative contributions and potential interactions of these two proteins through the crossing of Tespa1/ mice and Themis/ mice. Finally, further characterization of the basic biological functions of these two proteins (and others) may provide additional insight into the present models of TCR signaling and thymocyte selection. In sum, our studies have suggested that Tespa1 may be a previously unsuspected missing link in the TCR-proximal signaling machinery. METHODS Methods and any associated references are available in the online version of the paper.
Note: Supplementary information is available in the online version of the paper.
npg
567
Articles
ACKnOWLeDgMents We thank X. Li, C. Xu and B. Li for discussions; H. Cantor, Y. Ke and H. Hu for critical reading; and A. Alison for assistance with manuscript editing. Supported by the National Natural Science Foundation of China (30972724 and 31070782 to L.L., and 30901311 and 31170842 to D.W.), the Zhejiang Provincial Natural Science Foundation of China (R2090202 to L.L., and Y2090401 to D.W.) and the National Basic Research Program of China (973 Program; 2011CB944100 and 2012CB945004 to L.L.). AUtHOR COntRIBUtIOns D.W., M.Z., L.W., Y.K., J.W., X.C. and L. Lu. designed research; D.W., M.Z., L. Lei., J.J., Y.Y., Y.Q., L.M., J.L., C.O., X.Z., Y.H., J.C. and L. Lu. did research; D.W., M.Z., L.W. and L. Lu. analyzed data; and D.W., M.Z. and L. Lu. wrote the paper. COMPetIng FInAnCIAL InteRests The authors declare no competing financial interests.
12. Fu, G. et al. Themis controls thymocyte selection through regulation of T cell antigen receptormediated signaling. Nat. Immunol. 10, 848856 (2009). 13. Johnson, A.L. et al. Themis is a member of a new metazoan gene family and is required for the completion of thymocyte positive selection. Nat. Immunol. 10, 831839 (2009). 14. Cao, Y. et al. LKB1 regulates TCR-mediated PLC1 activation and thymocyte positive selection. EMBO J. 30, 20832093 (2011). 15. Boyden, L.M. et al. Skint1, the prototype of a newly identified immunoglobulin superfamily gene cluster, positively selects epidermal T cells. Nat. Genet. 40, 656662 (2008). 16. Diez-Roux, G. et al. A high-resolution anatomical atlas of the transcriptome in the mouse embryo. PLoS Biol. 9, e1000582 (2011). 17. Kurobe, H. et al. CCR7-dependent cortex-to-medulla migration of positively selected thymocytes is essential for establishing central tolerance. Immunity 24, 165177 (2006). 18. Anderson, G. & Jenkinson, E.J. Investigating central tolerance with reaggregate thymus organ cultures. Methods Mol. Biol. 380, 185196 (2007). 19. Brugnera, E. et al. Coreceptor reversal in the thymus: signaled CD4+8+ thymocytes initially terminate CD8 transcription even when differentiating into CD8+ T cells. Immunity 13, 5971 (2000). 20. Lucas, B. & Germain, R.N. Unexpectedly complex regulation of CD4/CD8 coreceptor expression supports a revised model for CD4+CD8+ thymocyte differentiation. Immunity 5, 461477 (1996). 21. Suzuki, H., Punt, J.A., Granger, L.G. & Singer, A. Asymmetric signaling requirements for thymocyte commitment to the CD4+ versus CD8+ T cell lineages: a new perspective on thymic commitment and selection. Immunity 2, 413425 (1995). 22. Cibotti, R., Punt, J.A., Dash, K.S., Sharrow, S.O. & Singer, A. Surface molecules that drive T cell development in vitro in the absence of thymic epithelium and in the absence of lineage-specific signals. Immunity 6, 245255 (1997). 23. Penninger, J.M. & Kroemer, G. Molecular and cellular mechanisms of T lymphocyte apoptosis. Adv. Immunol. 68, 51144 (1998). 24. Bouillet, P. et al. BH3-only Bcl-2 family member Bim is required for apoptosis of autoreactive thymocytes. Nature 415, 922926 (2002). 25. Calnan, B.J., Szychowski, S., Chan, F.K., Cado, D. & Winoto, A. A role for the orphan steroid receptor Nur77 in apoptosis accompanying antigen-induced negative selection. Immunity 3, 273282 (1995). 26. Yasutomo, K., Doyle, C., Miele, L., Fuchs, C. & Germain, R.N. The duration of antigen receptor signalling determines CD4+ versus CD8+ T-cell lineage fate. Nature 404, 506510 (2000). 27. Itano, A. et al. The cytoplasmic domain of CD4 promotes the development of CD4 lineage T cells. J. Exp. Med. 183, 731741 (1996). 28. Liu, X. & Bosselut, R. Duration of TCR signaling controls CD4CD8 lineage differentiation in vivo. Nat. Immunol. 5, 280288 (2004). 29. Amaral, M.C., Casillas, A.M. & Nel, A.E. Contrasting effects of two tumour promoters, phorbol myristate acetate and okadaic acid, on T-cell responses and activation of p42 MAP-kinase/ERK-2. Immunology 79, 2431 (1993). 30. Hermiston, M.L. et al. Differential impact of the CD45 juxtamembrane wedge on central and peripheral T cell receptor responses. Proc. Natl. Acad. Sci. USA 106, 546551 (2009). 31. Mayya, V. et al. Quantitative phosphoproteomic analysis of T cell receptor signaling reveals system-wide modulation of protein-protein interactions. Sci. Signal. 2, ra46 (2009).
1. Starr, T.K., Jameson, S.C. & Hogquist, K.A. Positive and negative selection of T cells. Annu. Rev. Immunol. 21, 139176 (2003). 2. Germain, R.N. T-cell development and the CD4CD8 lineage decision. Nat. Rev. Immunol. 2, 309322 (2002). 3. Bosselut, R. CD4/CD8-lineage differentiation in the thymus: from nuclear effectors to membrane signals. Nat. Rev. Immunol. 4, 529540 (2004). 4. Singer, A., Adoro, S. & Park, J.H. Lineage fate and intense debate: myths, models and mechanisms of CD4- versus CD8-lineage choice. Nat. Rev. Immunol. 8, 788801 (2008). 5. Werlen, G. & Palmer, E. The T-cell receptor signalosome: a dynamic structure with expanding complexity. Curr. Opin. Immunol. 14, 299305 (2002). 6. Brockmeyer, C. et al. T cell receptor (TCR)-induced tyrosine phosphorylation dynamics identifies THEMIS as a new TCR signalosome component. J. Biol. Chem. 286, 75357547 (2011). 7. Fischer, A.M., Katayama, C.D., Pages, G., Pouyssegur, J. & Hedrick, S.M. The role of erk1 and erk2 in multiple stages of T cell development. Immunity 23, 431443 (2005). 8. Daniels, M.A. et al. Thymic selection threshold defined by compartmentalization of Ras/MAPK signalling. Nature 444, 724729 (2006). 9. Kane, L.P. & Hedrick, S.M. A role for calcium influx in setting the threshold for CD4+CD8+ thymocyte negative selection. J. Immunol. 156, 45944601 (1996). 10. Freedman, B.D., Liu, Q.H., Somersan, S., Kotlikoff, M.I. & Punt, J.A. Receptor avidity and costimulation specify the intracellular Ca2+ signaling pattern in CD4+CD8+ thymocytes. J. Exp. Med. 190, 943952 (1999). 11. Lesourne, R. et al. Themis, a T cell-specific protein important for late thymocyte development. Nat. Immunol. 10, 840847 (2009).
npg
568
VOLUME 13
NUMBER 6
JUNE 2012
nature immunology
Mice. Tespa1/ mice were generated by homologous recombination-mediated gene targeting in embryonic stem cells of strain 129 at the Shanghai Research Center For Model Organisms as described32 (strategy, Supplementary Fig. 1d). Mice on a mixed 129 C57BL/6 background were backcrossed onto the C57BL/6 background for six generations. Mice with transgenic expression of the OT-I , OT-II or H-Y TCR and B2m/ mice were from the Jackson Laboratory. H2-Ab1/ mice and Rag1/ mice were from the Model Animal Research Center of Nanjing University. All mice were housed in the University Laboratory Animal Center. Animal experiment protocols were approved by the review committee from Zhejiang University School of Medicine and were in compliance with institutional guidelines. Antibodies. Fluorescein isothiocyanateconjugated antibody to mouse CD4 (L3T4), CD69 (H1.2F3), CD5 (Ly-1), CD3 (145-2C11) or BrdU (PRB-1); phycoerythrin-conjugated antibody to mouse CCR7 (4B12), the H-Y TCR (T3.70), V8 (MR5-2) or CD127 (A7R34); phycoerythrin-indodicarbocyanineconjugated anti-mouse Foxp3 (150D/E4); and allophycocyanin-conjugated antibody to mouse CD62L (MEL-14) and TCR (H57-597) were from BD Biosciences. Anti-Erk1-Erk2 (9102), antibody to Erk1-Erk2 phosphorylated at Thr202 and Tyr204 (9101), anti-Jnk1-Jnk2 (56G8), antibody to Jnk1-Jnk2 phosphorylated at Thr183 and Tyr185 (9251), anti-p38 (9212), antibody to p38 phosphorylated at Thr180 and Tyr182 (9211), anti-Akt (C67E7), antibody to Akt phosphorylated at Ser473 (D9E), antibody to Zap70 phosphorylated at Tyr319 and Syk phosphorylated at Tyr352 (2701), anti-IB (44D4) and antibody to IB phosphorylated at Ser32 and Ser36 (5A5) were from Cell Signaling Technology. Anti-Grb2 (Y237), anti-Cbl (YE323), anti-PLC-1 (EP1898-7Y), anti-SLP76 (EPR2549), antibody to PLC- phosphorylated at Tyr783 (EP1898Y), antibody to Lat phosphorylated at Tyr191 (E225) and antibody to Cbl phosphorylated at Tyr731 (EP973Y) were from Epitomics. Anti-Themis (06-1328) was from Millipore. The customized monoclonal antibody to mouse and human Tespa1 was from Abmart. Generation of bone marrow chimeras. Bone marrow cells collected from congenic C57BL/6 (CD45.1+) mice were mixed at a ratio of 1:4 with bone marrow cells from Tespa1+/+ or Tespa1/ (CD45.2+) mice and the mixture was injected into irradiated Rag1/ mice (650 rads). Mice were analyzed 8 weeks after bone marrow transfer. In vitro assay of DP-to-CD4+CD8int development. Purified DP thymocytes were resuspended in RPMI-1640 medium and incubated overnight with platebound anti-TCR (H57; BioLegend) and anti-CD2 (RM2-5; BD Bioscience) as described22. Cells were washed extensively and analyzed immediately by flow cytometry (stimulatory culture) or incubated further for 24 h in the same medium before analysis (recovery culture). Ca2+ flux. Thymocytes in suspensions were first labeled for 1 h at 37 C with 4 g/ml Fluo4 (Invitrogen), then were washed with ice-cold PBS and resuspended in PBS. Cells (3 106) were labeled on the surface for 30 min
ONLINE METHODS
on ice with phycoerythrin-conjugated anti-CD4 (H129.19; BD Biosciences) and phycoerythrin-indodicarbocyanine-conjugated anti-CD8 (53-6.7; BD Biosciences), then were incubated with biotinylated anti-CD3 (5 g/ml; 145-2C11; eBioscience) and anti-CD28 (5 g/ml; 37.51; BioLegend). Labeled cells were warmed for 20 min at room temperature and then crosslinked with streptavidin (25 g/ml) or were stimulated with PMA (phorbol 12-myristate 13-acetate) and ionomycin immediately before flow cytometry. Mean fluorescence ratios were plotted after analysis with FlowJo software (TreeStar). Real-time RT-PCR analysis. SYBR Premix Ex TaqTM II and a 7500 Fast Real-Time PCR system were used for real-time PCR and data were analyzed with 7500 software v2.0.1 (Applied Biosystems). Results were normalized to -actin expression. Splenocyte CD69-upregulation assay. Splenocytes from Tespa1+/+ or Tespa1/ mice were stimulated for 5 h at 37 C with plate-bound anti-CD3 (5 g/ml) and anti-CD28 (5 g/ml). Cells were stained and analyzed after gating on CD4+ or CD8+ cells. For OT-II Tespa1+/+ or Tespa1/ thymocytes, 2 105 to 3 105 cells were incubated for 5 h at 37 C with 1 106 antigenpresenting cells plus various concentrations of ovalbumin peptide in roundbottomed 96-well plates. Cells were stained and analyzed by flow cytometry as described33. Immunoprecipitation and immunoblot analysis. Cells were lysed in NP-40 Lysis Buffer containing 50 mM Tris (pH 7.4), 150 mM NaCl, 1% NP-40, PMSF and protease inhibitor cocktail (Sigma). Cell lysates were immunoprecipitated overnight at 4 C with the appropriate antibodies. Protein A/G Sepharose beads (Roche) were then added and the samples were incubated for additional 45 h. After being washed three times, samples were resolved by SDS-PAGE gels (10%) and blotted. For immunoblot analysis, cells were lysed in SDS sample buffer by the addition of 1/4 volume of 5 SDS sample buffer directly into cell suspensions. Samples were then boiled for 5 min and separated by (10%) SDS-PAGE. Immunoblot analyses were done as described34. Luciferase reporter assay. The activity of AP-1, NF-AT and NF-B luciferase reporter plasmids was assessed as described34. Statistical testing. Statistical differences were calculated with the mean difference hypothesis of Students two tailed t-test with assumption of different variances and a confidence level of 95%. GraphPad Prism software was used for calculations.
32. Zhang, W. et al. Essential role of LAT in T cell development. Immunity 10, 323332 (1999). 33. Yachi, P.P., Ampudia, J., Gascoigne, N.R. & Zal, T. Nonstimulatory peptides contribute to antigen-induced CD8T cell receptor interaction at the immunological synapse. Nat. Immunol. 6, 785792 (2005). 34. Wang, D. et al. Ras-related protein Rab10 facilitates TLR4 signaling by promoting replenishment of TLR4 onto the plasma membrane. Proc. Natl. Acad. Sci. USA 107, 1380613811 (2010).
npg
doi:10.1038/ni.2301
nature immunology
Articles
TGF-b and retinoic acid induce the microRNA miR-10a, which targets Bcl-6 and constrains the plasticity of helper T cells
Hayato Takahashi1, Tomohiko Kanno2, Shingo Nakayamada1, Kiyoshi Hirahara1, Giuseppe Scium1, Stefan A Muljo3, Stefan Kuchen4, Rafael Casellas4, Lai Wei1, Yuka Kanno1 & John J OShea1
Distinct CD4+ T cell subsets are critical for host defense and immunoregulation. Although these subsets can act as terminally differentiated lineages, they have been increasingly noted to demonstrated plasticity. MicroRNAs are factors that control T cell stability and plasticity. Here we report that naturally occurring regulatory T cells (Treg cells) had high expression of the microRNA miR-10a and that miR-10a was induced by retinoic acid and transforming growth factor-b (TGF-b) in inducible Treg cells. By simultaneously targeting the transcriptional repressor Bcl-6 and the corepressor Ncor2, miR-10a attenuated the phenotypic conversion of inducible Treg cells into follicular helper T cells. We also found that miR-10a limited differentiation into the T H17 subset of helper T cells and therefore represents a factor that can fine-tune the plasticity and fate of helper T cells. The appropriate differentiation of CD4+ T cell subsets is important for proper adaptive immune responses, and cytokines in the milieu are the key exogenous factors that drive specification1. These subsets include the TH1, TH2 and TH17 subsets of helper T cells, and regulatory T cells (Treg cells). Treg cells can be further divided into the following two populations: naturally occurring Treg cells (nTreg cells) that arise in the thymus, and inducible Treg cells (iTreg cells) that can arise in the periphery from naive CD4+ precursor cells. Follicular helper T cells (TFH cells) are another functional subset that is critical for providing help to B cells in germinal centers2. The stability versus plasticity of subsets is an area of intense investigation3,4. One example of plasticity is the generation of TH1 cells from TH17 cells5,6. Another example is the conversion of polarized TH2 cells into interferon- (IFN-) producers in the setting of viral infection7. The nTreg cells are generally viewed as a more stable subset than iTreg cells8. However, even for nTreg cells, there is evidence that they can alter their phenotype. For example, nTreg cells can convert into TFH cells in the environment in Peyers patches9 or become TH17 cells after stimulation with interleukin 6 (IL-6) in vitro10. There are still other examples of conversion from other subsets, such as the conversion of TH2 cells into TFH cells11 and, conversely, the conversion of TFH cells into cells of other effector subsets, including TH1, TH2 and TH17 cells12. However, the molecular mechanisms that underlie the phenotypic changes in Treg cells and other cells are incompletely understood. MicroRNAs (miRNAs) are small noncoding RNAs that regulate gene expression at the post-transcriptional level by directly binding
1Molecular
to the mRNA of target genes. These miRNAs are critically involved in a wide variety of the biological processes of cells, from embryonic stem-cell pluripotency to cancer tumorigenicity13, and evidence for the involvement of miRNA in cells of the immune response continues to mount1422. For example, Drosha and Dicer are two key components of the machinery responsible for the generation of miRNA, and loss of these factors is associated with defects in lymphocyte differentiation and autoimmunity2325. As for the issue of CD4+ T cell plasticity, deficiency in Dicer results in the instability of helper T cells, unstable expression of the transcription factor Foxp3 in Treg cells and skewing toward a TH1 phenotype26. Similarly, CD4+ T cells deficient in Drosha show accelerated differentiation into TH1 cells and diminished differentiation into iTreg cells23. Collectively, these data indicate that miRNA is important in preserving the phenotypic stability of helper T cells and raise the question of which specific miRNA(s) is (are) responsible. In this study, we investigated the function of the microRNA miR-10a, which has high expression in nTreg cells and is induced by transforming growth factor- (TGF-) and retinoic acid (RA). We found that miR-10a controlled expression of the transcriptional repressor Bcl-6 and the corepressor Ncor2 and thereby limited the conversion of iTreg cells into TFH cells. In addition, miR-10a inhibited TH17 differentiation. This effect was dependent on induction of the transcription factor T-bet by RA. Our data demonstrate that miR-10a is one factor that preserves the phenotype of Treg cells by targeting and constraining transcription-factor pathways that promote alternative fates.
npg
Immunology and Inflammation Branch, National Institute of Arthritis, Musculoskeletal and Skin Diseases, National Institutes of Health, Bethesda, Maryland, USA. 2Program in Genomics of Differentiation, National Institute of Child Health and Human Development, National Institutes of Health, Bethesda, Maryland, USA. 3Laboratory of Immunology, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, Maryland, USA. 4Genomics & Immunity, Laboratory of Molecular Immunogenetics, National Institute of Arthritis, Musculoskeletal and Skin Diseases, National Institutes of Health, Bethesda, Maryland, USA. Correspondence should be addressed to J.J.O. (osheajo@mail.nih.gov). Received 7 February; accepted 12 March; published online 29 April 2012; doi:10.1038/ni.2286
NUMBER 6
JUNE 2012
587
Articles a
miR-10a expression (relative) 200 150 100 50 0
+
b
miR-10a expression (relative) 40 30 20 10 0
c
miR-10a expression (relative)
d
12
e
miR-10a expression (relative) 25 150 20 15 10 5 0 TGF- Anti-TGF-
* *
miR-10a expression 2 (relative 10 )
****
**
**
**
**** **** ****
100
*** ***
1 2
H
**
17
***
50
re g
re g
Figure 1 High expression of miR-10a in nTreg cells and induction of miR-10a by RA and TGF-. (a) Quantitative RT-PCR analysis of miR-10a in various T cell subsets. (be) Expression of miR-10a in naive CD4+ T cells stimulated with anti-CD3 and anti-CD28 and cultured in medium alone (None), IL-2 (50 U/ml) or TGF- (20 ng/ml; b), with various concentrations of TGF- (c), with various concentrations of ATRA plus TGF- (20 ng/ml) or anti-TGF- (d), or with anti-TGF-, TGF- (20 ng/ml) or LE540 (1 M; e). (f) Quantitative RT-PCR analysis of RAR in naive CD4+ T cells stimulated with anti-CD3 and anti-CD28 and various concentrations of TGF-. (g) Abundance of miR-10a in isolated naive CD4+ T cells stimulated with anti-CD3 and-CD28 in the presence of TGF- alone or TGF- plus ATRA (1 M), the RAR agonist AM580 (1 M) or the RAR agonist A7980 (1 M). Results are presented relative to the expression of the small nucleolar RNA snoRNA202 (ae,g) or -actin (f). NS, not significant. *P < 0.05, **P < 0.01, ***P < 0.001 and ****P < 0.0001 (Students t-test). Data are representative of two (a,c,eg) or three (b,d) independent experiments with similar results (error bars, s.e.m. of triplicates).
f
RAR expression (relative) 6
g
**** **** ***
RESULTS Induction of miR-10a by RA and TGF-b in iTreg cells To identify miRNAs with preferential expression in various T cell subsets, we profiled genome-wide miRNA expression via massive parallel sequencing27. Among the abundant miRNAs with disparate expression in T cells, we identified miR-10a-5p (called miR-10a here) as the most selective miRNA in nTreg cells (Supplementary Fig. 1a). In contrast, we did not detect its paralogous sequence miR-10b-5p in any helper T cell subset (data not shown). We next confirmed the selective expression of miR-10a by quantitative RT-PCR and found that TH1, TH2, TH17 and naive CD4+ T cells have very low expression of miR-10a, consistent with the sequencing data (Fig. 1a and Supplementary Fig. 1b). In contrast, nTreg cells had abundant expression of miR-10a. We found that iTreg cells had more modest expression of miR-10a than did nTreg cells; nonetheless, its abundance was significantly greater in iTreg cells than in TH1, TH2, TH17 or naive CD4+ T cells (Fig. 1a). As the generation of iTreg cells in vitro requires TGF-28, we next sought to determine if this factor was relevant to miR-10a expression. TGF- induced miR-10a in a dose-dependent manner, whereas IL-2 had no effect (Fig. 1b,c). Because the upregulation of miR-10a by TGF- was modest, we considered other factors that might affect its expression. RA is another factor that promotes iTreg cell differentiation and induces miR-10a in non-T cells2932. Additionally, the sequence encoding miR-10a resides in the gene cluster that encodes the Hox transcription factors, which includes genes regulated by RA. Consistent with published results32, we noted that all-trans RA (ATRA) induced miR-10a expression in a dose-dependent manner. Notably, ATRA-induced expression of miR-10a was abolished by blockade of TGF- (Fig. 1d). We also found that the induction of miR-10a expression was blocked in the presence of LE540, a total inhibitor of the RA receptor (RAR) family (Fig. 1e). These aspects of miR-10a regulation have not been appreciated previously, to our knowledge, and support the idea that RA and TGF- are both required for maximal induction of miR-10a. To help explain the interaction of RA and TGF-, we considered the possibility that these factors might impinge on each others signaling
588
pathways. Indeed, we found that TGF- induced expression of the RAR family member RAR in a dose-dependent manner (Fig. 1f). We also found that an RAR agonist, in the presence of TGF-, induced miR-10a expression equivalent to that induced by ATRA. In contrast, an agonist of the RAR family member RAR did not (Fig. 1g). Collectively, these results demonstrated that the RA-induced expression of miR-10a was mediated via RAR in a TGF--dependent manner, with the latter inducing expression of the former. Direct targeting of Bcl-6 and Ncor2 by miR-10a To understand how miR-10a might function in helper T cells, we used in silico analysis to predict potential target mRNAs with the PicTar algorithm for the identification of microRNA targets. Among the predicted targets, we noted that Bcl-6 mRNA and Ncor2 mRNA had potential miR-10a target sequences in their 3 untranslated regions (3 UTRs; Supplementary Fig. 2a). The target genes were also predicted by another algorithm, TargetScan. Moreover, the 3-UTR target sequences of Bcl-6 mRNA and Ncor2 mRNA were highly conserved (Supplementary Fig. 2b). Bcl-6 is a key transcriptional repressor thought to be a master regulator for TFH cells, and Ncor2 is a corepressor that forms a complex with Bcl-6 and RAR to suppress the transcription of genes regulated by these factors3337. If miR-10a were involved in the fate of the helper T cell phenotype, these targets would provide good potential mechanisms for transmitting its effects. To investigate whether miR-10a regulates the expression of Bcl-6 and Ncor2, we generated reporter constructs that included the 3 UTR of Bcl-6 or Ncor2 mRNA. These constructs included the miR-10a target sequences linked to a luciferase gene. In addition, we generated constructs with deletion of the miR-10a target sequence (Fig. 2a). Overexpression of miR-10a resulted in significantly less luciferase activity derived from constructs that expressed the target sequences but had no effect on the expression of constructs that lacked these target sequences (Fig. 2a). These results demonstrated that miR-10a was able to directly target sequences in the 3 UTR of Bcl-6 mRNA and that of Ncor2 mRNA. We next sought to evaluate whether miR-10a influences the abundance of endogenous Bcl-6 and Ncor2 protein. First we used the mouse B cell lymphoma cell line CH12, which constitutively expresses Bcl-6.
npg
N on AT e R AM A 58 A7 0 98 0
0 1 5 10 20 TGF- (ng/ml)
tiTG F T TG GF T - F- - GF + L + An E5 40 tiTG F
250 200 150 100 50 0 NS
iT
C D 4
N ai ve
nT
None IL-2 TGF-
0 1 5 10 20 TGF- (ng/ml)
102 101 1 10 102 103 104 ATRA (nM)
An
****
Articles
1.5 0.5 Figure 2 Direct targeting of Bcl-6 and Ncor2 **** 991 * Luc-Bcl6 by miR-10a and downregulation of protein ** 0.4 CH12 Luc-Bcl6-del * expression. (a) Luciferase reporter vectors (left): 1.0 691 0.3 100 miR-10a target sequences (magenta) in Luc-Ncor2 miR-10a 80 Luc-Ncor2-del 0.2 the 3 UTR of Bcl-6 mRNA (above) and Ncor2 0.5 Control 60 mRNA (below) included in Luc-Bcl6 and 0.1 40 Luc-Ncor2, respectively; asterisks indicate 0 0 the complementary sequences of 5 bases at 20 + + + + Bcl6 Ncor2 Luc reporter position 29 or 18 of miR-10a lacking in the 0 Bcl6-del + Ncor2-del + 0 1 2 3 4 10 10 10 10 10 + Control + Control mutant vectors Luc-Bcl6-del and Luc-Ncor2-del, miRNA vector Bcl-6 + + miR-10a + + miR-10a respectively. Middle and right, luciferase activity in NIH3T3 mouse fibroblasts transfected with the luciferase (Luc) reporters at left and miRNA iTreg 150 iTreg CH12 vectors, presented relative to renilla luciferase 100 0.6 * **** miR-10a activity. (b,c) Bcl-6 expression in CH12 cells 80 100 transduced with retroviral vectors encoding the Control 0.4 60 Bcl-6 human nerve growth factor receptor (hNGFR) 40 50 Bcl-6 and miR-10a or control vector, assessed by 0.2 Actin 20 flow cytometry after gating of the hNGFR+ 0.66 0.28 Actin 0 0 population (b) or by immunoblot analysis after 0 0.49 0.23 100 101 102 103 104 isolation of transduced (hNGFR+) cells by Bcl-6 magnetic beads (c). Bottom (c), immunoblot analysis of lysates with anti-actin; below lanes, 400 ** iTreg iTreg quantification of the ratio of Bcl-6 to actin by NS ** ** 100 Control 100 300 densitometry. (d) Bcl-6 abundance in naive ** + T cells transduced with control vector 80 CD4 miR-10a-5pT 200 50 or vector expressing miR-10a and stimulated 60 with anti-CD3 and CD28 under iTreg cell 40 Ncor2 100 conditions (TGF- (5 ng/ml), IL-2 (50 U/ml) 0 20 Actin and anti-IFN-), assessed by flow cytometry 0 0 0 1 2 3 4 + + + Bcl6 0.27 0.09 after gating on transduced cells (left). 10 10 10 10 10 Flag reporter Bcl6-del + + Bcl-6 Right, Bcl-6 expression presented as MFI. 4 Control + + miRNA vector (e) Immunoblot analysis of Bcl-6 in CD4+ T cells NS miR-10a + + + 0.15 iTreg ** cultured and transduced with miR-10a as in d 3 + + + + Control Sponge vector miR-10a-5pT + and isolated by magnetic beads (quantification 0.10 2 as in c). (f) Quantitative RT-PCR analysis of Bcl-6 mRNA in the cells in e. (g) Abundance of Bcl-6 in NIH3T3 cells transduced with 1 0.05 Ncor2 various combinations of three vectors (vector encoding Flag-tagged Bcl-6 that includes the 0 3 UTR of Bcl-6 mRNA and the indicator marker hNGFR; vector encoding miR-10a and the Actin 0 indicator marker Orange2; and vector encoding the miR-10a sponge target sequence and the 0.07 0.11 indicator marker 4GFP) or the corresponding control vectors, assessed by flow cytometry with anti-Flag after gating on GFP+Orange2+hNGFR+ cells. (h) Immunoblot analysis of Ncor2 in activated CD4+ T cells transduced with miR-10a (quantification as in c). (i) Quantitative RT-PCR analysis of Bcl-6 mRNA in the cells in h. (j) Bcl-6 expression in naive CD4+ T cells transduced with a control vector or vector expressing the miR-10a sponge target sequence miR-10a-5pT and stimulated under iTreg cellinducing conditions, evaluated by flow cytometry in the CD4 +hNGFR+ population (left); right, MFI. (k) Immunoblot analysis (left) of Ncor2 expression in isolated transduced cells from j. Right, pooled data. *P < 0.05, **P < 0.01, ***P < 0.001 and ****P < 0.0001 (Students t-test). Data are representative of three (a,ek) or two (b,c,d) individual experiments with similar results (error bars, s.e.m.). Ncor2 luciferase activity (relative) Bcl-6 luciferase activity (relative)
Events (% of max)
C o m ntr iR ol -1 0a
Bcl-6 mRNA expression
Events (% of max)
tro l m iR -1 0
Bcl-6 (MFl)
C on
tro l m iR -1 0
C on
Events (% of max)
Flag (MFI)
Bcl-6 (MFI)
tro
on
iR
-1
0a
Ncor2 mRNA expression
tro l m -5 iRpT 10
Ncor2
on
npg
We found that overexpression of miR-10a resulted in less Bcl-6 in these cells, as determined by flow cytometry and immunoblot analysis (Fig. 2b,c). Next we overexpressed miR-10a in primary helper T cells and again found that this resulted in lower Bcl-6 expression (Fig. 2d,e). The decrease in Bcl-6 protein was associated with a compensatory increase in the abundance of Bcl-6 mRNA (Fig. 2f). Conversely, overexpression of Bcl-6 protein by retroviral transduction suppressed the accumulation of endogenous Bcl-6 mRNA (data not shown), consistent with a published report showing that Bcl-6 regulates its own gene expression38. To simplify the complex regulation of Bcl-6 expression, we expressed a tagged Bcl-6 construct under the control of a heterologous promoter and manipulated the abundance of miR-10a by overexpression or by sequestration of miR-10a with a sponge (decoy) target sequence39 (Supplementary Fig. 3). We found that overexpression of miR-10 downregulated Bcl-6 expression, whereas expression of the miR-10a sponge target sequence restored the expression of Bcl-6 (Fig. 2g). Deletion of the miR-10a target sequence from the 3 UTR of Bcl-6 mRNA eliminated this regulation (Fig. 2g). Overexpression of miR-10a
in helper T cells also resulted in less Ncor2 protein but had little effect on the abundance of Ncor2 mRNA (Fig. 2h,i). To verify the effect of endogenous miR-10a in regulating its target proteins, we expressed the miR-10a sponge sequence in iTreg cells. We found that the miR-10a sponge sequence resulted in significant upregulation of the expression of Bcl-6 and Ncor2 protein (Fig. 2j,k). Given that Ncor2 is a corepressor of RAR, it seemed plausible that a lower abundance of Ncor2 protein would augment the RA-induced expression of miR-10a by relieving the inhibitory effects of Ncor2. To examine this, we knocked down Ncor2 expression in cells through the use of short hairpin RNA and cultured the cells with RA (Supplementary Fig. 4a,b). We observed that inhibiting Ncor2 shifted the RA doseresponse curve to the left, which meant the cells were more sensitive to RA and had more effective induction of miR-10a (Supplementary Fig. 4c). Manipulating the abundance of another miR-10a target, either by knocking down or overexpressing Bcl-6, had no effect on the RAinduced expression of miR-10a (Supplementary Fig. 4c and data not shown). These results suggested that miR-10a amplified its own induction by RA through a positive-feedback loop and targeting of Ncor2.
589
iR Co -1 nt 0a ro -5 l pT
iR Co -1 n 0a tro -5 l pT
Articles a
Foxp3GFP mouse Tcra mouse
/
Naive CD4 Foxp3 T cells Transfer In vitro differentiation Transfer iTreg cells
b
10
5 4 3 2
Naive CD4+ T cells

105
iTreg cells 0
105
4 3 2
Foxp3 (GFP)
10
10
3 2
Foxp3 (GFP)
10
10
10 10
98.8
CD44
10 0
99.8
0 10
2
10 0 0 10
2
100
10
3
10 0 0 10
2
10
10 10
10
10
10
10
105
CD4
CD62L
CD4
c
Peyers patch
Naive CD4+ Tcra/
Tcra/
iTreg
Tcra/
d
Peyers patch
10 10
5 4
Naive CD4+
Tcra/
Tcra/
TFH cells (%)
iTreg
25 20 15 10 5 0
+
8.9
20.4
103
PD-1
102 0 0 102 10
3
Figure 3 Conversion of iTreg cells into TFH cells in iTreg Peyers patches. (a) Protocol: naive CD4+ T cells Naive CD4+ isolated from mice with sequence encoding Tcra/ Tcra/ Gated on CD4+V+ cells 5 GFP knocked into Foxp3 (Foxp3GFP) were either 10 transferred directly into Tcra/ mice (control; Peyers * 40 104 * top row) or cultured for 5 d under iTreg cell conditions, patch 6.22 0.35 3 10 with Foxp3-expressing cells then isolated by cell 30 2 sorting and transferred into Tcra/ mice (bottom row). 10 ** 0 20 (b) Flow cytometry of transferred naive CD4+ T cells 2 3 4 5 and iTreg cells. Numbers adjacent to outlined 0 10 10 10 10 10 areas indicate percent cells in each throughout. 105 (c) Immunofluorescence microcopy of Peyers 0 Spleen 104 patches from Tcra/ mice that received no cells Transferred Naive Naive 22.6 iTreg iTreg + 0.51 cell CD4+ CD4 (right; control), naive CD4+ T cells (left) or iTreg cells 103 (middle), assessed with anti-B220 (green), anti-CD4 2 10 Tissue Peyers Spleen 0 (red) and anti-AID (blue) for the visualization of patch TFH cells. Scale bars, 250 m. (d) Quantification of 0 102 103 104 105 TFH cells in from Peyers patches from the mice in c, CD4 assessed by expression of PD-1 and CXCR5 after gating on the CD4+TCR+CD44hi7-AAD population. (e) GFP (Foxp3) expression in CD4+TCR+7-AAD T cells from the spleen and Peyers patches, analyzed by flow cytometry (left). Right, frequency of Foxp3+ cells (pooled data from n = 4 mice per group). Each symbol (d, right) represents an individual mouse; small horizontal lines indicate the mean. *P < 0.01 and **P < 0.001 (Students t-test). Data are representative of two individual experiments with similar results (error bars, s.e.m.).
npg
Foxp3 (GFP)
Conversion of iTreg cells into TFH cells constrained by miR-10a As Bcl-6 is pivotal for the generation of TFH cells3335, we hypothesized that miR-10a might function to limit the generation of TFH cells by suppressing Bcl-6. Treg cells can convert into TFH cells in Peyers patches, a scenario in which miR-10a might be functionally relevant9. As iTreg cells are more suited for effective retroviral transduction, we used these cells to manipulate miR-10a expression and assessed the conversion of iTreg cells into TFH cells in Peyers patches. Specifically, we isolated naive CD4+ T cells from mice with sequence encoding green fluorescent protein (GFP) knocked into the Foxp3 locus. We differentiated these cells in vitro into iTreg cells and transferred the cells into mice deficient in the T cell antigen receptor -chain (Tcra/ mice). We then assessed the appearance of TFH cells in Peyers patches 6 weeks later (Fig. 3a,b). Immunofluorescence staining showed most of the transferred naive CD4+ T cells resided in interfollicular regions. However, some of the transferred cells were present in germinal centers (Fig. 3c, left), which suggested that they had differentiated into TFH cells. To confirm TFH differentiation, we assessed expression of the TFH cell markers PD-1 and CXCR5 (Fig. 3d, left). Consistent with the immunofluorescence staining, the transferred CD4+ T cells were present, which had acquired characteristics of TFH cells. Similarly, transferred iTreg cells generated in vitro localized to germinal centers (Fig. 3c, middle) and acquired
590
expression of PD-1 and CXCR5; in fact, iTreg cells were more efficient in their ability to convert into TFH cells than were naive CD4+ T cells (Fig. 3d). As a control, we assessed Foxp3 expression in transferred naive and iTreg cells. Foxp3+ cells were barely detectable in Peyers patches and spleens after the transfer of naive CD4+ T cells (Fig. 3e). Transferred iTreg cells more easily lost Foxp3 expression in Peyers patches than in the spleen (Fig. 3e), which suggested that the environment of Peyers patches promoted conversion into TFH cells. We next evaluated the effect of sequestrating miR-10a on the conversion of iTreg cells into TFH cells. For this, we generated iTreg cells (97.7% 0.4% Foxp3+; data not shown) that expressed the miR-10a sponge sequence or a control sequence. We adoptively transferred those cells into Tcra/ mice and assessed the acquisition of TFH cell markers. We found that cells transduced with the miR-10a sponge vector more readily acquired PD-1 and CXCR5 in Peyers patches than did control cells (Fig. 4a). This was associated with the presence of significantly more germinal-center B cells (Fig. 4b). We next determined the effect of overexpressing miR-10a in iTreg cells, reasoning that enforced expression should attenuate their conversion into TFH cells by limiting the amount of Bcl-6. We observed that overexpression of miR-10a significantly less conversion of transferred iTreg cells into TFH cells in Peyers patches (Fig. 4c).
Foxp3+ cells (%)
ai
ve
B220/CD4/AID
CXCR5
iT
re
10
105
Articles a
10 10 10
5 4 3 2
iTreg (control) Tcra

/
iTreg (miR-10a-5pT) Tcra

/
b
TFH cells (%)
iTreg (control) Tcra/

5
iTreg (miR10a-5pT) Tcra/ 38.3

GC B cells (%)
c
50 40 30 20 10 0
iTreg (control) Tcra/

105 104 103
iTreg (miR10a) Tcra/
25.3
35.2
30 20 0
C on m tr iR 10 ol a5p T
Fas
TFH cells (%)
40
10
18.8
104 103 102 0
50.9
28.5
60 40 20 0
PD-1
PD-1
10 0
10
10 0
d
Events (% of max)
e
100 Figure 4 Conversion of iTreg cells into TFH cells constrained by miR-10a. (a) Flow cytometry (left) to assess the 20 80 ** TFH conversion of iTreg cells into TFH cells as follows: naive CD4+ T cells were transduced with vector encoding the 15 Non-TFH 60 miR-10a sponge target sequence (miR-10a-5pT) and hNGFR or a control vector, then cultured for 5 d with Naive 10 40 ** TGF- (10 ng/ml), IL-2 (50 U/ml) and 1 nM ATRA, followed by isolation of transduced hNGFR+ cells and 5 20 transfer into Tcra/ mice; acquisition of surface markers indicative of TFH conversion (CD4+V+hNGFR+ 0 0 2 3 4 5 010 10 10 10 CD44hiPD-1hiCXCR5hi) was measured 5 weeks later. Right, frequency of TFH cells. (b) Appearance of Bcl-6 +IgDloFas+GL-7+ germinal-center B cells in the mice in a, assessed by flow cytometry (left). Right, frequency B220 of germinal-center (GC) B cells (data pooled from five mice). (c) TFH cells in Peyers patches of Tcra/ mice that received naive CD45.1+CD4+ T cells transduced with vector encoding miR-10a and hNGFR or control vector, then differentiated into iTreg cells and purified by flow cytometry; expression of PD-1 and CXCR5 was analyzed 6 weeks after transfer by flow cytometry after gating on CD4 +CD45.1+hNGFR+CD44hi7-AAD population (left). Right, pooled data. (d,e) Expression of Bcl-6 (d) and miR-10a (e) by TFH cells (CD4+V+PD-1hiCXCR5hi) and non-TFH cells (CD4+V+PD-1 CXCR5) isolated from Peyers patches by cell sorting, assessed by flow cytometry (d) and quantitative RT-PCR (e); results in e are presented relative to snoRNA202 expression. Each symbol (ac, right) represents an individual mouse; small horizontal lines indicate the mean. *P < 0.05 and **P < 0.0001 (Students t-test). Data are representative of three (a) or two (b) independent experiments with similar results (three to five mice per experiment) or three independent experiments with similar results (c,d) or are pooled from three independent experiments (e; error bars, s.e.m.).
We next measured miR-10a and Bcl-6 in T FH cells (CD4+V+ PD-1hiCXCR5hi) and non-TFH cells (CD4+V+PD-1CXCR5) from Peyers patches. As expected, Bcl-6 was preferentially expressed in TFH cells rather than non-TFH cells (Fig. 4d), whereas the opposite
was true for miR-10a (Fig. 4e). Another population of helper T cells, designated follicular regulatory T cells (TFR cells), has been identified4042. These cells express both Foxp3 and Bcl-6 and exert a regulatory function in germinal centers. To assess miR-10a
a
100 Foxp3 cells (%) 80 60 40 IL-17A 20 0
2 3 4 0 0.1 1 10 10 10 10 ATRA (nM)
iTreg 0
10 10 10
5 4 3 2
ATRA (nM)
+ Foxp3 cells (%)
c
+ IL-17A cells (%)
90 80 70 60 50 0
1 0.1 0.03 0.1 0.01
10 0.1
0.1
100 50 0
*** *
****
TGF- (ng/ml) 2 4 6 10 20
10 0
47.8
0 10
2
85.6
94.7
b
IL-17A cells (%)
Foxp3 TH17
10
10 10
0 1 10 ATRA (nM)
100 80 60 40 IL-17A 20 0 0 0.1 1 10 102 103 104 ATRA (nM) TH17 (anti-IL-2) TH17 iTreg
npg
IL-17A+ cells (%)
0
10 10 10
5 4 3 2
31.4
ATRA (nM) 1 1.9 0.3 46.9
10 14.2 2.1
40 30 20 10 0
0.1 1 10 102 103 ATRA (nM)
d
IL-17A+ cells (% increase) 60 40
10 0 0 10
2
0.8
10
3 4 5 10 10
6.4
25.9
*** *** **
Foxp3 TH17 (anti-IL-2) 0

105 10 10
4 3 2
0 1 10 ATRA (nM)
20
IL-17A+ cells (%)
56.1
0.4
ATRA (nM) 1 83.9 1.3
10 80 1.8
80 60 40 20 0
** *
NS
2 4 6 10 20 TGF- (ng/ml)
IL-17A
10 0
0.4
0 10
2
0.8
2.2
10
10 10
Foxp3
0 1 10 ATRA (nM)
Figure 5 RA has biphasic effects on TH17 differentiation. (a,b) Foxp3-expressing cells (a) and IL-17A-producing cells (b) among naive CD4+ T cells polarized into TH17 cells with IL-6 (10 ng/ml), TGF- (10 ng/ml) and anti-IFN- alone (TH17) or with anti-IL-2 (TH17 (anti-IL-2)) or polarized into iTreg cells with IL-2 (50 U/ml), TGF- (10 ng/ml) and anti-IFN- (iTreg), incubated with various concentrations of ATRA, analyzed by flow cytometry. Left, ATRA dose-response curves. Middle, numbers in quadrants indicate percent cells in each. Right, frequency of Foxp3 + cells (a) or IL-17A+ cells (b; pooled data). (c) Frequency of TH17 cells generated from naive CD4+ T cells with IL-6, anti-IFN- and anti-IL-2 plus various concentrations of ATRA (horizontal axis) and TGF- (key). (d) The effect of TGF- dose on the ATRA-dependent generation of TH17 (IL-17A+) cells, calculated as the following ratio: frequency of IL-17A+ cells cultured in the presence of ATRA (10 nM) minus that in the absence of ATRA divided by that in the absence of ATRA. *P < 0.05, **P < 0.01, ***P < 0.001 and ****P < 0.05 (Students t-test). Data are from three independent experiments (a,b, left; mean s.e.m. of triplicates) or five (iTreg and TH17 (anti-IL-2)) or seven (TH17) independent experiments (a,b, middle and right; error bars (right), s.e.m.), or two independent experiments with similar results (c,d; mean s.e.m of triplicates).
NUMBER 6
JUNE 2012
C T D no FH 4 + n- ce T lls FH ce lls
N ai ve
C on
CXCR5
GL-7
C o m iR ntro 10 l a5p T
CXCR5
tro l m iR 10 a
0 10
10
10
105
0 102 103 104 105
0 102 103 104 105
591
Articles
Figure 6 Onset of experimental autoimmune encephalomyelitis delayed (but not abrogated) by miR-10a. (a) Appearance of neurological disease (EAE score; left) in mice deficient in recombination-activating gene 2 that received naive CD4+ T cells isolated from 2D2 mice and transduced with vector expressing miR-10a and hNGFR (n = 24 recipient mice) or control vector (n = 29 recipient mice), then stimulated for 3 d with anti-CD3 and antiCD28, anti-IFN- and anti-IL-4 and no exogenous cytokines, followed by population expansion for additional 2 d with IL-2 (50 U/ml); recipients were immunized with myelin oligodendrocyte glycoprotein peptide (100 g) after transfer of transduced cells. Right, area-under-curve (AUC) values for graph at left. *P < 0.0001 (two-way analysis of variance) and **P < 0.001 (Students t-test). (b) Expression of IL-17A and IFN- in 2D2 T cells in the central nervous system, lymph nodes, and spleen at day 11 after immunization as in a, analyzed by flow cytometry with gating on CD4+V+hNGFR+7-AAD cells (left). Right, frequency of IL-17A+ cells or IFN-+ cells; each symbol represents an individual mouse (n = 5 per group) and small horizontal lines indicate the mean. *P < 0.05 and **P < 0.001 (Students t-test). Data are from five independent experiments (a; error bars, s.e.m.) or are representative of two independent experiments with similar results (b).
a
EAE score
3 Control 2 miR10a 10
**
*
AUC 0 2 4 6 8 10 12 Time after immunization (d) 2D2 (miR10a) IL-17A+ cells (%) 7.58 0.76 30 20 10 0 20 15 10 5 0 15 10 5 0 14 16
8 6 4 2
l tro C on
IFN- cells (%)
b
10 10
5 4 3 2
2D2 (control) 20.6 0.79
m
NS NS NS
80 40 0
CNS
10
10 0
18.3
0102 103 104 105
28 0.55 0.043
IL-17A+ cells (%)
IFN- cells (%)
10 10
5 4
11
3.18
20 10 0
Biphasic effects of RA on TH17 differentiation RA was originally reported to induce iTreg cells and suppress TH17 differentiation2931; however, subsequent work has suggested that RA is necessary for normal immune responses and can enhance inflammation43,44. Such findings led us to reexamine the effect of RA on TH17 differentiation and to assess the potential effects of miR-10a. Consistent with published reports2931, we found that ATRA at a concentration of 10 nM inhibited IL-17A production and enhanced Foxp3 expression (Fig. 5a,b). However, a lower, physiological concentration of RA (1 nM)45 enhanced the proportion of IL-17Aproducing T cells (Fig. 5b). IL-2 inhibits TH17 differentiation46, and we observed more enhancement of IL-17A production by RA when IL-2 was neutralized (Fig. 5b). As a control, cells cultured under iTreg conditions without IL-6 did not produce IL-17A, even with a low dose of RA (Fig. 5b). Notably, the magnitude of IL-17A production, assessed by mean fluorescence intensity (MFI), was greater at all doses of ATRA (Supplementary Fig. 6), despite the lower proportion of IL-17A-producing cells under TH17 conditions without antibody to IL-2 (anti-IL-2). The discrepancy between the greater MFI and lower frequency of IL-17A+ cells after ATRA treatment underscored the complex biological effects of ATRA on TH17 cells. Under these culture conditions, ATRA substantially upregulated miR-10a expression in a dose-dependent manner (Supplementary Fig. 7). TGF- is another factor that regulates IL-17A production, with lower doses of TGF- promoting more IL-17A production47. However, TGF- also enhanced the responsiveness of T cells to ATRA by upregulating RAR expression (Fig. 1f). We therefore assessed the combined effects of RA and TGF- on TH17 differentiation in vitro (Fig. 5c,d). As reported before47, TH17 differentiation was more efficient at lower concentrations of TGF-. The ability of RA to enhance IL-17A production was more efficient at high concentrations of TGF-, but regardless of the concentration of TGF-, the biphasic effects of RA were evident (Fig. 5c,d). These results indicated that TH17
592
differentiation was fine-tuned by ATRA and TGF-. The attenuated TH17 differentiation expected to occur with higher concentrations of TGF- was antagonized by RA. Constraint of TH17 differentiation by miR-10a Given the ability of RA to influence TH17 differentiation and induce miR-10a, we sought to analyze the effect of miR-10a expression on TH17 differentiation in vivo with a disease model dependent on these cells. We therefore transduced naive CD4+ T cells from 2D2 mice (which have transgenic expression of a T cell antigen receptor specific for myelin oligodendrocyte glycoprotein) with an expression vector for miR-10a and assessed whether miR-10 overexpression influenced the severity of experimental autoimmune encephalomyelitis. After immunization with a peptide of myelin oligodendrocyte glycoprotein, mice that received 2D2 T cells that overexpressed miR-10a had a significantly delayed onset of neurological disease (Fig. 6a). The production of IL-17A in was significantly lower in miR-10a-expressing 2D2 cells in the lymph nodes, spleen and central nervous system, whereas there was no significant effect of miR-10a on IFN- production (Fig. 6b). These results indicated that miR-10a was able to function as a factor that limited TH17 responses in vivo. It is notable that the disease was delayed but not abrogated. In contrast, disease eventually developed in mice that received miR-10a-expressing 2D2 cells, and their disease severity was similar to that of mice that received 2D2 cells transduced with a control vector. This suggested that miR-10a fine-tuned but did not abrogate IL-17 expression. As miR-10a overexpression limited TH17 differentiation in vivo, we sought to determine how miR-10a might influence TH17 differentiation in vitro. When we used a low dose of RA to generate TH17 cells, which thus expressed miR-10a, sequestration of this miRNA with the sponge vector significantly enhanced TH17 differentiation (Fig. 7a). Conversely, when miR-10a was overexpressed under the
npg
expression in this subset, we immunized mice expressing GFP from the Foxp3 locus with sheep red blood cells and isolated the following three populations from spleens: TFR cells (Foxp3+CD4+V+PD-1hi CXCR5hi), Treg cells (Foxp3+CD4+V+PD-1CXCR5) and Foxp3 TFH cells (Foxp3CD4+V+PD-1hiCXCR5hi; Supplementary Fig. 5a). We found that TFR cells had higher expression of miR-10a than did Treg cells (Supplementary Fig. 5b). Together these results indicated that one function of miR-10a was to constrain the conversion of iTreg cells into TFH cells.
10 0 010 10 10
5 4 2
14.5
10
3
6.13 2.18 0.28
10
10
IL-17A+ cells (%)
IFN- cells (%)
9.36
0.9
LN 103
30 20 10 0
IL-17A
10 0 010
2
8.62
10
3
14
tro
10
tro on
IFN-
on
iR
iR
10
10
10
3 Sp 10
iR 10
Articles a
TH17 (anti-IL-2) + ATRA 80 IL-17A+ cells (%) 60 40 20 0
105 104 103
Control 27.2
b
*
Control 66.1
TH17 (anti-IL-2) + ATRA 100 IL-17A+ cells (%) 80 60 40 20 0
c
*
TH17 (anti-IL-2) Control 21.4 IL-17A cells (%) 80 60 40 20 0 NS
d
0.1
iTreg + ATRA Control Foxp3 cells (%) 80 60 40 20 0 NS
105 104 103
miR10a-5pT 40.2
miR-10a 31.2
105 10
4
miR-10a 21.8
miR10a-5pT 105 0.1

104 103
tro
0a
tro
tro
0a
tro
5p
IL-17A
IL-17A
IL-17A
IL-17A
m iR -1
m iR -1
m iR 10
0 102 103 104 105
4 0 102 103 10 105
4 0 102 103 10 105
4 0 102 103 10 105
Foxp3
Foxp3
Foxp3
Foxp3 iTreg + ATRA
Control Figure 7 TH17 differentiation constrained by miR-10a in the presence Control 60 NS 80 NS 0.5 of RA. (a) Flow cytometry of naive CD4+ T cells transduced with 0.07 retroviral construct expressing the miR-10a sponge target sequence 60 40 (miR10a-5pT) or scrambled sequence (control) and cultured with 1.5 nM ATRA under TH17 conditions (IL-6 (10 ng/ml), TGF- (2 ng/ml) 69.9 40 47.7 and anti-IFN-) plus anti-IL-2. (b,c) Flow cytometry of naive CD4 + 20 miR-10a miR-10a T cells transduced with retroviral construct expressing miR-10a or 20 5 105 0.5 10 0.08 4 control vector under T H17 conditions (IL-6 (10 ng/ml), TGF- (2 ng/ml) 4 10 10 0 0 and anti-IFN-) plus anti-IL-2, with ATRA (b) or without ATRA (c). 103 103 2 (d) Flow cytometry of naive CD4 + T cells transduced as in a and 2 10 10 47.3 64.5 0 0 cultured with 3.0 nM ATRA under iTreg conditions (IL-2 (50 U/ml), 2 3 5 2 3 4 5 0 10 10 104 10 0 10 10 10 10 + TGF- (2 ng/ml) and anti-IFN-). (e,f) Flow cytometry of naive CD4 Foxp3 Foxp3 T cells transduced as in b,c and cultured under iTreg conditions (IL-2 (50 U/ml), TGF- (2 ng/ml (e) or 5 ng/ml (f)) and anti-IFN-) with ATRA (e) or without ATRA (f). Left, expression of IL-17A and Foxp3; right, frequency of IL-17A + cells (ac) or Foxp3+ cells (df). *P < 0.05 (paired Students t-test). Data are representative of (contour plots) or pooled from (graphs) five to eleven independent experiments. Foxp3 cells (%)
iTreg
0a
Foxp3 cells (%)
m iR 10 m
102 0
0.03
10 0
C on
C on
C on
C on
0.4
0.2
a-
66.5
IL-17A
tro
IL-17A
-1
tro
on
iR
on
same conditions, TH17 differentiation was inhibited (Fig. 7b). When miR-10a was overexpressed in TH17 cells cultured without RA, the efficiency of TH17 differentiation was not affected (Fig. 7c). These effects of miR-10a on TH17 differentiation were not due to changes in Foxp3 expression (Fig. 7df). These data indicated that miR-10a was able to influence TH17 differentiation, but only in circumstances in which RA was present. Bcl-6 and Ncor2 influence IL-17A expression via T-bet The finding that miR-10a inhibited TH17 differentiation was unexpected. We had anticipated that miR-10a expression would
npg
promote TH17 differentiation by relieving Bcl-6-mediated inhibition of IL-17A. However, the effect of miR-10a effects was evident only in the presence of RA. In published work showing that Bcl-6 inhibits IL-17A, a role for RA was not considered34. We therefore mimicked the effects of miR-10a by directly downregulating the expression of its targets, Bcl-6 and Ncor2, and found that in the presence of RA, knocking down these factors inhibited TH17 differentiation (Fig. 8a). Thus, inhibiting the expression of Bcl-6 and Ncor2 recapitulated the effect of overexpressing miR-10a. Although this result may seem counterintuitive, it must be borne in mind that TH17 differentiation occurred in the presence of RA in these experiments.
a
10 10
5 4 3 2
Control 22.3 17.5
iNcor2 15.9
iBcl6 IL-17A+ cells (%) 30 20 10
WT
**
Tbx21
/
10 10 10
5 4 3 2
26.8
26.5
26.6
IL-17A+ cells (%)
Control
iNcor2
iBcl6
40 30 20 10 0
Tbx21/ NS NS
WT
10
IL-17A
0.1
0 10
2
0.1
0.1
IL-17A
10
tr iN ol co r2 iB cl 6
on
Figure 8 Ncor2 and Bcl-6 regulate IL-17A in a T-bet-dependent 0 100 1,000 manner. (a,b) Expression of IL-17A and Foxp3 by naive CD4+ T cells T-bet /) 800 obtained from wild-type (WT) mice (a) or T-bet-deficient (Tbx21 ATRA Actin mice (b) and transduced with retroviral vector encoding short hairpin 600 0.5 1,000 nM RNA specific for Ncor2 (iNcor2) or Bcl-6 (iBcl6) or a control vector 0.4 400 10 nM and cultured under TH17 conditions with anti-IL-2 plus 1.5 nM ATRA, 0.3 + cells; 0.1 nM 200 assessed by flow cytometry (left). Right, frequency of IL-17A 0.2 0 nM each symbol represents a T cell culture from a mouse, and small horizontal 0 4 0.1 0 0.1 10 1,000 0 102 103 10 105 lines indicate the mean of quadruplicates. (c,d) T-bet expression in naive 0 ATRA (nM) + T cells cultured under T 17 conditions and anti-IL-2 without T-bet CD4 H ATRA (nM): 0 100 ATRA (0 nM) or with 100 nM ATRA (c) or various concentrations of ATRA (d), assessed by immunoblot analysis (c) or flow cytometry (d). Bottom (c), quantification of T-bet relative to actin (by densitometry). *P < 0.05 and **P < 0.01 (Students t-test). Data are representative of two (a,b) or three (c,d) experiments with similar results (error bars, s.e.m.).
T-bet expression (relative) T-bet (MFI)
ATRA (nM)
on
Foxp3
iN
NUMBER 6
JUNE 2012
iB cl
Foxp3
tro
593
10
10 10
10 0 0 10
2
0.2
10
3
0.1
0.2
10 10
iR
-1
co
0a
a-
102 0
10 0
5p
r2
103
0.02
0.2
0.5
68.7
Articles
It was less obvious why downregulation of Bcl-6 led to the inhibition of TH17 differentiation. In this context, a relevant Bcl-6 target is T-bet, an important negative regulator of TH17 differentiation34,35,48. We speculated that diminishing Bcl-6 expression would enhance T-bet function and thereby suppress TH17 differentiation. To test this possibility, we knocked down Bcl-6 in wild type and T-bet-deficient TH17 cells. We found that in the absence of T-bet, decreasing Bcl-6 had no effect on TH17 differentiation (Fig. 8b). Thus, the lower IL-17A production associated with lower expression of Bcl-6 was T-bet dependent. In addition to forming a complex with RAR, Ncor2 forms a repressor complex with Bcl-6 (ref. 37). We therefore wondered whether the effect of Ncor2 might also be dependent on T-bet. We found that the knocking down Ncor2 influenced IL-17A production only when T-bet was present (Fig. 8a,b). The unexpected involvement of T-bet in mediating the regulation of TH17 differentiation by the miR-10a targets Bcl-6 and Ncor2 prompted us to evaluate the effect of RA on T-bet expression. We found that RA was an effective inducer of T-bet under TH17 conditions (Fig. 8c,d). Collectively, these results indicated that RA induced both miR-10a and T-bet and that miR-10a exerted its negative effects on TH17 differentiation together with T-bet. DISCUSSION The importance of miRNA in T cell biology is rapidly becoming ever more apparent. Although there is clear evidence that miRNA in general is critically involved in T cell stability2325, the precise roles of individual miRNAs have been less clear. However, this gap in knowledge is rapidly being filled, and several studies have reported the roles of individual miRNAs in helper T cells1422. Here we found that miR-10a was an RA-induced factor with high expression in Treg cells and influenced their plasticity. In this study, we found that nTreg cells had high expression of miR-10a and that miR-10a was induced by factors that promote the Treg cell phenotype; that is, RA and TGF-. Functionally, miR-10a limited the plasticity of Treg cells, but, unexpectedly, it did not do so by influencing Foxp3 expression or other factors involved in the homeostasis of Treg cells. Instead, miR-10a limited the ability of Treg cells to acquire the features of TFH cells by targeting Bcl-6. It has been reported that iTreg cells generated through the use of RA are more likely to retain Foxp3 expression and are more resistant to conversion into other lineages than are iTreg cells generated by other methods31. Those findings fit well with our own data and are consistent with the idea that the expression of miR-10a by nTreg cells and RA-treated iTreg cells would enhance their stability. Published studies have identified another subset of cells with suppressive function that express both Foxp3 and Bcl-6 (TFR cells)4042. Notably, we found high expression of miR-10a in TFR cells. We speculate that miR-10a might have particularly critical functions in TFR cells in that the environment of the germinal center would be expected to be rich in factors that drive Bcl-6 expression. Thus, miR-10a might have important functions in constraining the actions of Bcl-6 in these cells and allowing the continued action of Foxp3. RA is a ubiquitous metabolite of vitamin A with broad effects on cellular differentiation. Vitamin A and RA are particularly abundant in the liver and the gut, and the importance of RA in mucosal immunity has been well documented. The effects of RA include the trafficking and gut tropism of conventional T cells and Treg cells31,49. Regulation of miR-10a can now be added to the effects of RA. However, the effects of RA are by no means simple. Although RA can induce the expression of Foxp3 and inhibit IL-17A production29,30,
594
subsequent work has shown that vitamin Adeficient mice have poor TH17 responses during T. gondii infection43. Our results showing the biphasic effects of RA on TH17 differentiation are consistent with those findings. Also initially perplexing to us in terms of IL-17 regulation and the control of Bcl-6 expression was published work indicating that Bcl-6 inhibits IL-17A production34. We found the contrary result; that is, the presence of Bcl-6 (and Ncor2) was associated with higher expression of IL-17A. However, in the context of the ability of RA to induce T-bet, a potent inhibitor of IL-17A, that result now makes more sense. Although in some circumstances low expression of Bcl-6 might amplify IL-17A production, in the circumstances in which T-bet is present, lower abundance of Bcl-6 would have the opposite effect. Low Bcl-6 expression would facilitate the unopposed action of T-bet to efficiently inhibit IL-17A production. We would argue that through its effects on Bcl-6, under selective circumstances, miR-10a can inhibit IL-17A production. Although the effect of miR-10a on TH17 differentiation in vitro was evident only in the presence of RA, we were able to discern an effect in vivo when miR-10a was overexpressed in adoptively transferred 2D2 T cells. However, our data also showed that miR-10a modulated but did not abrogate disease. Thus, like other miRNAs, miR-10a is best thought of as tuning responses. Our data have also made it clear that the effects of RA on TH17 differentiation were complex and its effects were influenced by another ubiquitous regulator of T cell differentiation; that is, TGF-. Thus, the action of RA was by no means simple, as it affected the abundance of T-bet, Foxp3 and Bcl-6, depending on the context. In this study, we sought to delineate the function of miR-10a, a factor with high expression in nTreg cells that is inducible by RA and TGF-. Notably, RA seems to directly regulate miR-10a, as published work indicates that RAR binds to the locus encoding miR-10a (ref. 50). By targeting Bcl-6, this miRNA seems to functionally attenuate the differentiation of TFH cells and TH17 cells. Thus, it is one factor that contributes to maintenance of the cell-specific phenotype of Treg cells by targeting factors that could lead to conversion into other fates. It is likely that the regulatory role of miR-10a would be most prominent in circumstances in which cells are poised for phenotypic transition, with the gut being one place where this occurs. Thus, regulation of miR-10a would be a reasonable mechanism for fine-tuning factors that influence fate decision versus plasticity in subsets of helper T cells. METHODS Methods and any associated references are available in the online version of the paper.
Note: Supplementary information is available in the online version of the paper. ACKNOWLedGMeNTS We thank Y. Belkaid, W. Chen and A. Villarino for reading this manuscript; K. Moro and T. Tamachi for advice on manipulating Peyers patches; L. Naldini (San Raffaele Institute) for the LV-SFFV lentiviral vector; and J. Simone and J. Lay for cell sorting. Supported by the Japan Society for the Promotion of Science (Research Fellowship for Japanese Biomedical and Behavioral Researchers at the US National Institutes of Health to H.T. and K.H.), the Instituto Pasteur-Fondazione Cenci-Bolognetti (G.S.) and the Intramural Research Program of the National Institute of Arthritis, Musculoskeletal and Skin Diseases. AUTHOR CONTRIBUTIONS H.T. designed, did, analyzed and interpreted all the experiments and wrote the manuscript; T.K., S.N., K.H., G.S., S.K. and L.W. helped with experiments; Y.K. contributed to data interpretation, helped with experiments and wrote the manuscript; S.A.M. and R.C. contributed to experimental design and provided
npg
VOLUME 13
NUMBER 6
JUNE 2012
nature immunology
Articles
suggestions; and J.J.O. contributed to experimental design, analyzed and interpreted all acquired data and wrote the manuscript. COMPeTING FINANCIAL INTeReSTS The authors declare no competing financial interests.
23. Chong, M.M., Rasmussen, J.P., Rudensky, A.Y. & Littman, D.R. The RNAseIII enzyme Drosha is critical in T cells for preventing lethal inflammatory disease. J. Exp. Med. 205, 20052017 (2008). 24. Liston, A., Lu, L.F., OCarroll, D., Tarakhovsky, A. & Rudensky, A.Y. Dicer-dependent microRNA pathway safeguards regulatory T cell function. J. Exp. Med. 205, 19932004 (2008). 25. Zhou, X. et al. Selective miRNA disruption in Treg cells leads to uncontrolled autoimmunity. J. Exp. Med. 205, 19831991 (2008). 26. Muljo, S.A. et al. Aberrant T cell differentiation in the absence of Dicer. J. Exp. Med. 202, 261269 (2005). 27. Kuchen, S. et al. Regulation of microRNA expression and abundance during lymphopoiesis. Immunity 32, 828839 (2010). 28. Chen, W. et al. Conversion of peripheral CD4+CD25 naive T cells to CD4+CD25+ regulatory T cells by TGF- induction of transcription factor Foxp3. J. Exp. Med. 198, 18751886 (2003). 29. Mucida, D. et al. Reciprocal TH17 and regulatory T cell differentiation mediated by retinoic acid. Science 317, 256260 (2007). 30. Coombes, J.L. et al. A functionally specialized population of mucosal CD103+ DCs induces Foxp3+ regulatory T cells via a TGF- and retinoic acid-dependent mechanism. J. Exp. Med. 204, 17571764 (2007). 31. Benson, M.J., Pino-Lagos, K., Rosemblatt, M. & Noelle, R.J. All-trans retinoic acid mediates enhanced Treg cell growth, differentiation, and gut homing in the face of high levels of co-stimulation. J. Exp. Med. 204, 17651774 (2007). 32. Meseguer, S., Mudduluru, G., Escamilla, J.M., Allgayer, H. & Barettino, D. Micro-RNAs-10a and -10b contribute to retinoic acid-induced differentiation of neuroblastoma cells and target the alternative splicing regulatory factor SFRS1 (SF2/ASF). J. Biol. Chem. 286, 41504164 (2010). 33. Johnston, R.J. et al. Bcl6 and Blimp-1 are reciprocal and antagonistic regulators of T follicular helper cell differentiation. Science 325, 10061010 (2009). 34. Nurieva, R.I. et al. Bcl6 mediates the development of T follicular helper cells. Science 325, 10011005 (2009). 35. Yu, D. et al. The transcriptional repressor Bcl-6 directs T follicular helper cell lineage commitment. Immunity 31, 457468 (2009). 36. Chen, J.D. & Evans, R.M. A transcriptional co-repressor that interacts with nuclear hormone receptors. Nature 377, 454457 (1995). 37. Dhordain, P. et al. Corepressor SMRT binds the BTB/POZ repressing domain of the LAZ3/BCL6 oncoprotein. Proc. Natl. Acad. Sci. USA 94, 1076210767 (1997). 38. Kikuchi, M. et al. Identification of negative regulatory regions within the first exon and intron of the BCL6 gene. Oncogene 19, 49414945 (2000). 39. Gentner, B. et al. Stable knockdown of microRNA in vivo by lentiviral vectors. Nat. Methods 6, 6366 (2009). 40. Linterman, M.A. et al. Foxp3+ follicular regulatory T cells control the germinal center response. Nat. Med. 17, 975982 (2011). 41. Chung, Y. et al. Follicular regulatory T cells expressing Foxp3 and Bcl-6 suppress germinal center reactions. Nat. Med. 17, 983988 (2011). 42. Wollenberg, I. et al. Regulation of the germinal center reaction by Foxp3+ follicular regulatory T cells. J. Immunol. 187, 45534560 (2011). 43. Hall, J.A. et al. Essential role for retinoic acid in the promotion of CD4+ T cell effector responses via retinoic acid receptor . Immunity 34, 435447 (2011). 44. DePaolo, R.W. et al. Co-adjuvant effects of retinoic acid and IL-15 induce inflammatory immunity to dietary antigens. Nature 471, 220224 (2011). 45. Kane, M.A., Folias, A.E., Wang, C. & Napoli, J.L. Quantitative profiling of endogenous retinoic acid in vivo and in vitro by tandem mass spectrometry. Anal. Chem. 80, 17021708 (2008). 46. Laurence, A. et al. Interleukin-2 signaling via STAT5 constrains T helper 17 cell generation. Immunity 26, 371381 (2007). 47. Manel, N., Unutmaz, D. & Littman, D.R. The differentiation of human TH-17 cells requires transforming growth factor- and induction of the nuclear receptor RORt. Nat. Immunol. 9, 641649 (2008). 48. Lazarevic, V. et al. T-bet represses TH17 differentiation by preventing Runx1-mediated activation of the gene encoding RORt. Nat. Immunol. 12, 96104 (2011). 49. Iwata, M. et al. Retinoic acid imprints gut-homing specificity on T cells. Immunity 21, 527538 (2004). 50. Mahony, S. et al. Ligand-dependent dynamics of retinoic acid receptor binding during early neurogenesis. Genome Biol. 12, R2 (2011).
1. Weaver, C.T., Harrington, L.E., Mangan, P.R., Gavrieli, M. & Murphy, K.M. Th17: an effector CD4 T cell lineage with regulatory T cell ties. Immunity 24, 677688 (2006). 2. Crotty, S. Follicular helper CD4 T cells (TFH). Annu. Rev. Immunol. 29, 621663 (2011). 3. Zhou, L., Chong, M.M. & Littman, D.R. Plasticity of CD4+ T cell lineage differentiation. Immunity 30, 646655 (2009). 4. OShea, J.J. & Paul, W.E. Mechanisms underlying lineage commitment and plasticity of helper CD4+ T cells. Science 327, 10981102 (2010). 5. Lee, Y.K. et al. Late developmental plasticity in the T helper 17 lineage. Immunity 30, 92107 (2009). 6. Bending, D. et al. Highly purified Th17 cells from BDC2.5NOD mice convert into Th1-like cells in NOD/SCID recipient mice. J. Clin. Invest. 119, 565572 (2009). 7. Hegazy, A.N. et al. Interferons direct Th2 cell reprogramming to generate a stable GATA-3+T-bet+ cell subset with combined Th2 and Th1 cell functions. Immunity 32, 116128 (2010). 8. Rubtsov, Y.P. et al. Stability of the regulatory T cell lineage in vivo. Science 329, 16671671 (2010). 9. Tsuji, M. et al. Preferential generation of follicular B helper T cells from Foxp3+ T cells in gut Peyers patches. Science 323, 14881492 (2009). 10. Xu, L., Kitani, A., Fuss, I. & Strober, W. Cutting edge: regulatory T cells induce CD4+CD25Foxp3 T cells or are self-induced to become Th17 cells in the absence of exogenous TGF-. J. Immunol. 178, 67256729 (2007). 11. Zaretsky, A.G. et al. T follicular helper cells differentiate from Th2 cells in response to helminth antigens. J. Exp. Med. 206, 991999 (2009). 12. Lu, K.T. et al. Functional and epigenetic studies reveal multistep differentiation and plasticity of in vitro-generated and in vivo-derived follicular T helper cells. Immunity 35, 622632 (2011). 13. Yu, F. et al. let-7 regulates self renewal and tumorigenicity of breast cancer cells. Cell 131, 11091123 (2007). 14. Rossi, R.L. et al. Distinct microRNA signatures in human lymphocyte subsets and enforcement of the naive state in CD4+ T cells by the microRNA miR-125b. Nat. Immunol. 12, 796803 (2011). 15. Steiner, D.F. et al. MicroRNA-29 regulates T-box transcription factors and interferon production in helper T cells. Immunity 35, 169181 (2011). 16. Ma, F. et al. The microRNA miR-29 controls innate and adaptive immune responses to intracellular bacterial infection by targeting interferon-. Nat. Immunol. 12, 861869 (2011). 17. Stittrich, A.B. et al. The microRNA miR-182 is induced by IL-2 and promotes clonal expansion of activated helper T lymphocytes. Nat. Immunol. 11, 10571062 (2010). 18. Lu, L.F. et al. Function of miR-146a in controlling Treg cell-mediated regulation of Th1 responses. Cell 142, 914929 (2010). 19. Du, C. et al. MicroRNA miR-326 regulates TH-17 differentiation and is associated with the pathogenesis of multiple sclerosis. Nat. Immunol. 10, 12521259 (2009). 20. Lu, L.F. et al. Foxp3-dependent microRNA155 confers competitive fitness to regulatory T cells by targeting SOCS1 protein. Immunity 30, 8091 (2009). 21. Rodriguez, A. et al. Requirement of bic/microRNA-155 for normal immune function. Science 316, 608611 (2007). 22. Thai, T.H. et al. Regulation of the germinal center response by microRNA-155. Science 316, 604608 (2007).
npg
NUMBER 6
JUNE 2012
595
Mice. Foxp3-IRES-GFP knock-in mice on the C57BL/6 background have been described51. B6.129S2-Tcratm1Mom/J (Tcra/) mice, B6.129S6-Tbx21tm1Glm/J (T-bet-deficient) mice and C57BL/6-transgenic (Tcra2D2, Tcrb2D2)1Kuch/J (2D2) mice52 were from Jackson Laboratory. C57BL/6J mice, B6.CD45.1 mice and B6.129S6-Rag2tm1Fwa N12 (recombination-activating gene 2deficient) mice were from Taconic. All animal studies were done according to guidelines for the use and care of live animals from the US National Institutes of Health and were approved by the Institutional Animal Care and Use Committee of the National Institute of Arthritis, Musculoskeletal and Skin Diseases. Antibodies. Anti-CD44fluorescein isothiocyanate (IM7), antiB220phycoerythrin (RA3-6B2), anti-CD11bphycoerythrin (M1/70), anti-CD11cphycoerythrin (HL3), anti-CD25phycoerythrin (PC61), antiCD45.1phycoerythrin (A20), anti-IL-17Aphycoerythrin (TC11-18H10), anti-CD62Lallophycocyanin (MEL-14), anti-TCRallophycocyanin (H57-597), anti-B220Alexa Fluor 488, anti-GL-7Alexa Fluor 647 (GL-7), anti-hNGFRbiotin (C40-1457) and streptavidinphycoerythrin-indotricarbocyanine were from BD Bioscience. Anti-hNGFRfluorescein isothiocyanate, anti-hNGFRphycoerythrin or anti-hNGFRallophycocyanin (ME20.4), antiCD4allophycocyanin-indotricarbocyanine or anti-CD4biotin (RM4-5) and anti-TCRPacific blue (H57-597) were from BioLegend. Anti-Foxp3 fluorescein isothiocyanate or anti-Foxp3allophycocyanin (FJK-16s), antiAID (mAID-2), anti-T-beteFlouor660 (4B10) and anti-Bcl-6phycoerythrin (GI191E) were from eBioscience. Rabbit anti-Bcl-6 (4242) and antiFlagAlexa Fluor 647 (3916S) were from Cell Signaling Technology. Rabbit anti-Ncor2 (06-891) and mouse anti-actin (C4) were from Millipore. Antirat immunoglobulin G (IgG)Alexa Fluor 647 (A21247)and streptavidinAlexa Fluor 594 (S1122) were from Invitrogen. Anti-rabbit IgGIRDye800 (611-132-122) and anti-mouse IgG-IRDye700 (610-130-121) were from Rockland. Cell isolation and sorting. Peripheral T cells were obtained from spleen and lymph nodes of 8- to 12-week-old mice. Naive CD4+ T cells (CD4+CD25CD62LhiCD44lo) were isolated with a FACSAria II (BD) or MoFlo (Beckman Coulter) after enrichment of CD4+ T cells by AutoMACS with a mouse CD4+ T Cell Isolation Kit (Miltenyi Biotec). In Figure 5, the CD4+C D25CD11bCD11cCD62LhiCD44lo population was isolated as naive CD4+ T cells. For the isolation of naive CD4+ T cells and nTreg cells from Foxp3EGFP mouse, CD4+GFPCD62LhiCD44lo cells and CD4+GFP+ cells were isolated, respectively. For the isolation of iTreg cells, GFP+ cells were differentiated from the naive T cells of Foxp3EGFP mouse and isolated. For analysis of lymphocytes from Peyers patches, Peyers patches were enucleated from the small bowels and incubated for 15 min at 37 C in Hanks balanced-salt solution containing 10% FBS, 5 mM EDTA, 15 mM HEPES (pH 7.2) and 1 mM dithiothreitol. Peyers patches were washed with vigorous vortexing in Hanks balanced-salt solution containing 5% FBS and 25 mM HEPES (pH 7.2) and cell debris floating in supernatants were removed until supernatants became clear. Then, Peyers patches were mechanically smashed for the generation of single-cell suspensions. For some experiments, TFH cells (CD4+V+PD-1hiCXCR5hi) and nonTFH cells (CD4+V+PD-1CXCR5) were isolated from Peyers patches by flow cytometry. For some experiments, TFR cells (CD4+V+Foxp3+PD-1hiCXCR5hi), Treg cells (CD4+V+Foxp3+PD-1CXCR5), TFH cells (CD4+V+Foxp3PD1hiCXCR5hi) were isolated by flow cytometry from the spleens of mice with sequence encoding GFP knocked into Foxp3, which were immunized with sheep red blood cells as described12. Cell culture. All cultures of T cells used RPMI-1640 medium (Invitrogen) supplemented with 10% FCS, 2 mM glutamine, 100 IU/ml penicillin, 0.1 mg/ml streptomycin, 10 mM HEPES (pH 7.2), 1 mM sodium pyruvate and nonessential amino acids, and 2 M -mercaptoethanol. Cells were activated with plate-bound anti-CD3 (10 g/ml; 145-2C11; eBioscience) and CD28 (10 g/ml; 37.51; eBioscience). The following culture conditions were used, with the cytokine concentrations noted in figure legends. T H17 conditions included TGF- (R&D Systems), IL-6 (R&D Systems), 10 g/ml anti-IFN- (XMG1.2; BD) and 10 g/ml anti-IL-2 (S4B6; BD). For Figure 8a,b, anti-IL-4 (11B.11; NCI Frederick) was also used. The iTreg conditions included TGF-, human IL-2 and 10 g/ml anti-IFN-. ATRA, A7980, LE540 and AM580 were
ONLINE METHODS
used as described53. NIH3T3 cells were maintained in DMEM (Invitrogen) supplemented with 10% calf serum, 2 mM glutamine, 100 IU/ml penicillin, 0.1 mg/ml streptomycin, 1 mM sodium pyruvate and nonessential amino acids. For HEK293T human embryonic kidney cells, DMEM supplemented with 10% FBS, 2 mM glutamine, 100 IU/ml penicillin, 0.1 mg/ml streptomycin, 1 mM sodium pyruvate and nonessential amino acids was used. Experimental autoimmune encephalomyelitis. Naive CD4 +V 11 + CD25CD62Lhi CD44lo cells isolated from 2D2 mice were stimulated for 16 h in plates coated with anti-CD3 and anti-CD28 and were transduced with recombinant retrorvirus expressing miR-10a and the selection marker hNGFR or with the corresponding control virus. The cells were further cultured for 2 d under neutral conditions (anti-IFN- and anti-IL-4; 10 g/ml of each) and another 2 d in RPMI medium supplemented with 50 U/ml human IL-2. Transduced hNGFR+ cells did not express IFN- (<0.5%) or IL-17A (undetectable; data not shown), and isolated hNGFR+ cells (1.0 106) were adoptively transferred into mice deficient in recombination-activating gene 2. After subcutaneous immunization with 100 g myelin oligodendrocyte glycoprotein, amino acids 3555 (Peptides International) in complete Freunds adjuvant (day 0), the clinical scores of the mice were evaluated daily according to published criteria54. Mononuclear cells were isolated from the central nervous system (spinal cord and brain), lymph nodes and spleen, and cytokine production was analyzed by intracellular staining with flow cytometry. Retroviral vector. The miR-10a expression vector was constructed as follows. Genomic DNA encompassing mouse miR-10a (miRBase accession code, MI0000685; mmu-mir-10a) was cloned by PCR with the primers 5-ACCCACAGTGACTTTTCTGCTCC-3 and 5-GGACACCTCAGGTA GATGAGATTT-3 and was inserted between exons 10 and 11 of the gene encoding CD19 in pLenti vector (Invitrogen), called pLenti-Orange2miR-10a here. The miR-10a sequence is expected to be transcribed as a part of mRNA encoding fluorescence protein Orange2 and spliced out as a mirtron (microRNA located in the intron of the mRNA; Supplementary Fig. 3a), generating mature miR-10a. A control vector, pLenti-Orange2Control, was generated that transcribes only a part of the sequence encoding CD19. Orange2 marker sequence was replaced with truncated hNGFR to generate a retroviral version of the miR-10a expression vector and control (pMY-hNGFR-miR-10a and pMY-hNGFR-Control; Supplementary Fig. 3d) with pMYs-puro (Cell Biolabs) as a backbone vector. The miR-10a sponge vector was constructed as follows: eight tandem repeats of the miR-10a target sequence 5-CACAAATTCGGTAAACAGGGTA-3 were cloned into the LVSFFV vector39, called LV-SFFV-4GFP-miR-10a-5pT here (Supplementary Fig. 3a). Similarly, eight repeats of a scrambled target sequence were used for the creation of a control vector, called LV-SFFV-4GFP-scT here. The selection marker 4GFP, a mutant version of GFP with shorter half-life than that of the original, was replaced with truncated hNGFR to generate the retroviral vectors pMY-hNGFR-miR10a-5pT and pMY-hNGFR-scT (Supplementary Fig. 3d). The BLOCK-iT Lentiviral Pol II miR RNAi Expression System (Invigrogen) was used for the expression of short hairpin RNA specific for Bcl-6 (Mmi505152; Invitrogen) or Ncor2 (Mmi520051; Invitrogen). The plasmid pcDNA6.2-GW/EmGFP-miR-neg Control was used for the construction of negative-control vectors. Gene segments including EmGFP and short hairpin RNA were transferred into the pMYs-puro vector, which generated pMY-EmGFP-iBcl6, pMY-EmGFP-iNcor2 and pMY-EmGFP-Control (negative control). EmGFP was replaced with truncated hNGFR to generate pMY-hNGFR-iBcl6, pMY-hNGFR-iNcor2 and pMY-hNGFR-Control. For overexpression of Bcl-6, first pMY-IRES-hNGFR vector was made by replacement of the GFP sequence of the pMYs-IRES-GFP vector (Cell Biolabs) with hNGFR. The cDNA encoding Bcl-6 was subsequently subcloned into pMYIRES-hNGFR to generate pMY-Bcl-6-IRES-hNGFR (RV-Bcl-6). For expression of Flag-tagged Bcl-6 protein from cDNA encoding Bcl-6 with the 3 UTR or with a 3 UTR without the miR-10a-seed sequence, the pMYs-based expression vectors Bcl6 and Bcl6-del (Fig. 2g) were generated, which also express hNGFR, with 3 UTR sequences derived from luciferase reporter construct. Retroviral transduction. Retroviral vectors were transfected into PlatE cells to generate recombinant retrovirus in culture supernatants. For retroviral
npg
nature immunology
doi:10.1038/ni.2286
transduction of CD4+ T cells, naive CD4+ T cells were stimulated for 16 h with plate-bound anti-CD3 and anti-CD28 together with soluble anti-IFN-. Culture medium was replaced with retroviral supernatant containing 4 g/ml polybrene, followed by centrifugation at 770g for 2 h. After 4 h of incubation at 37 C, viral supernatants were replaced with T cell culture medium containing the appropriate cytokines and antibodies. T cell differentiation was evaluated by flow cytometry at 3 d after transduction. For retroviral transduction of HEK293T cells and CH12 cells, the spin-inoculation conditions were modified to 770g or 400g, respectively, for 90 min. Adoptive transfer and detection of TFH cells in Peyers patches. Naive CD4+GFPCD62Lhi CD44lo cells were isolated from mice with sequence encoding GFP knocked into Foxp3 and were transferred into Tcra/ mice. Some naive T cells were stimulated for 6 d with 10 ng/ml TGF- and 50 U/ml recombinant human IL-2 and then Foxp3+ (GFP+) cells were isolated by flow cytometry and transferred into Tcra/ mice. For overexpression or sequestration of miR-10a, naive CD4+ T cells were transduced with the appropriate recombinant retrovirus and hNGFR+ cells were isolated by flow cytometry or magnetic beads at 67 d after transduction and then were transferred into Tcra/ mice. Spleens and Peyers patches were collected and analyzed by flow cytometry or immunohistocmistry at 56 weeks after transfer. 2012 Nature America, Inc. All rights reserved. Flow cytometry. Cytokines, transcriptional factors and surface markers were evaluated by flow cytometry with a FACSCanto, FACSCalibur or FACSVerse (BD). For cytokine detection, cells were stimulated for 2 h with PMA (phorbol 12-myristate 13-acetate) and ionomycin with the addition of GolgiPlug (BD). For the exclusion of dead cells, cells were stained with 7-AAD (7-aminoactinomycin D; BD) and were washed with PBS twice before fixation. Cells were fixed and permeabilized with the Foxp3 Staining Buffer Set (eBioscience) or BD Cytofix/Cytoperm (BD) and were stained with fluorescent antibodies. Events were collected and analyzed with FlowJo software (TreeStar). RT-PCR. Total RNA was isolated with a mirVana miRNA Isolation Kit (Applied Biosystems). Then cDNA was synthesized with TaqMan Reverse Transcription reagents (Applied Biosystems). An ABI 7500 Fast Real-Time PCR System with Taqman site-specific primers and probes (Applied Biosystems) was used for Qquantitative PCR. For reverse transcription and quantification of miRNA, a TaqMan Reverse Transcription Kit was used in combination with TaqMan miRNA assays for snoRNA202 and human mir-10a (miRBase accession code, MI0000266; hsa-mir-10a). Results were normalized to the abundance of -actin or snoRNA202. Immunoblot analysis. Cell lysates were fractionated on 412% Bis-Tris gel (Invitrogen), followed by transfer to a nitrocellulose membrane. Membranes were
blocked for 30 min in Odyssey blocking buffer (LI-COR Biosciences) and analyzed by immunoblot with the appropriate antibodies. Bands were visualized with Odyssey infrared imaging system (LI-COR Biosciences). Band intensities were analyzed by ImageJ software and normalized to the intensity of the actin band and results are presented as a ratio of the band of interest to that of actin. Luciferase assay. The 3 UTR of Bcl-6 or Ncor2 was cloned from genomic DNA with the following primers, followed by ligation into the pmirGLO vector (Promega); the resultant plasmids were called Luc-Bcl6 and Luc-Ncor2, respectively: forward, 5-ATGAAGCATGGAGTGTTCCTCGCCCTT-3, and reverse, 5-ATCTGCAGGCAGACACGGATCTGAGA-3, for Bcl-6; forward, 5-TGAGACACTCTCGGACAGCGAGTGA-3, and reverse, 5-AAATGACA GAATGCCGCCGTGCACA-3, for Ncor2. The QuikChange II Site-Directed Mutagenesis Kit (Agilent Technologies) was used for deletion of predicted miR-10a target sequences from each 3 UTR, and the resultant plasmids were called Luc-Bcl6-del and Luc-Ncor2-del. NIH3T3 cells were transfected with a luciferase reporter, renilla internal control reporter and either the miR-10a expression vector or a control vector, and firefly and renilla luciferase activity was measured with the Dual-Glo Luciferase Assay System (Promega). Immunohistochemical analysis. Peyers patches were embedded in optimum cutting temperature compound (Sakura Finetechnical). Frozen sections were cut with a cryostat at a thickness of 20 m and were stored at 80 C before fixation. Sections were fixed with 4% paraformaldehyde or acetone, were stained with antibodies and were analyzed with a Leica SP5 NLO confocal microscope. Transfection of NIH3T3 cells. NIH3T3 cells were transfected with three expression vectors encoding miR-10a, the miR-10a sponge and Flag-tagged Bcl-6 through the use of Lipofectamine 2000 (Invitrogen). Then, 2 d after transfection, cells were collected and analyzed by flow cytometry. Statistical analysis. The appropriate analyses were used and are presented in each figure legend. A P value of less than 0.05 was considered significant.
51. Bettelli, E. et al. Reciprocal developmental pathways for the generation of pathogenic effector TH17 and regulatory T cells. Nature 441, 235238 (2006). 52. Bettelli, E. et al. Myelin oligodendrocyte glycoprotein-specific T cell receptor transgenic mice develop spontaneous autoimmune optic neuritis. J. Exp. Med. 197, 10731081 (2003). 53. Elias, K.M. et al. Retinoic acid inhibits Th17 polarization and enhances FoxP3 expression through a Stat-3/Stat-5 independent signaling pathway. Blood 111, 10131020 (2008). 54. Ghoreschi, K. et al. Generation of pathogenic T(H)17 cells in the absence of TGF-beta signalling. Nature 467, 967971 (2010).
npg
doi:10.1038/ni.2286
nature immunology
Articles
Divergent transcriptional programming of classspecific B cell memory by T-bet and RORa

Nathaniel S Wang1, Louise J McHeyzer-Williams1, Shinji L Okitsu1, Thomas P Burris2, Steven L Reiner3 & Michael G McHeyzer-Williams1
Antibody class defines function in B cell immunity, but how class is propagated into B cell memory remains poorly understood. Here we demonstrate that memory B cell subsets unexpectedly diverged across antibody class through differences in the effects of major transcriptional regulators. Conditional genetic deletion of the gene encoding the transcription factor T-bet selectively blocked the formation and antigen-specific response of memory B cells expressing immunoglobulin G2a (IgG2a) in vivo. Cell-intrinsic expression of T-bet regulated expression of the transcription factor STAT1, steady-state cell survival and transcription of IgG2a-containing B cell antigen receptors (BCRs). In contrast, the transcription factor RORa and not T-bet was expressed in IgA+ memory B cells, with evidence that knockdown of RORa mRNA expression and chemical inhibition of transcriptional activity also resulted in lower survival and BCR expression of IgA + memory B cells. Thus, divergent transcriptional regulators dynamically maintain subset integrity to promote specialized immune function in class-specific memory B cells. Memory B cells are long-lived antigen-experienced B cells that typically express a high-affinity B cell antigen receptor (BCR), rapidly expand their populations and differentiate into plasma cells after antigen rechallenge1,2. Although memory B cells that express immunoglobulin M (IgM) have specialized functions3, many antigen-primed B cells switch to non-IgM isotypes under the antigen-specific regulation of follicular helper T cells4. Furthermore, non-IgM classes of membranebound antibody have different abilities to transduce signals through their BCR on the basis of the constant region expressed5,6. However, little is known about the molecular signals required for the survival, activation or differentiation of class-switched memory B cells. Several studies have delineated how cytokines affect transcriptional programs differently in naive B cells that culminate in class-switch recombination7,8. Interleukin 4 (IL-4) and interferon- (IFN-) reciprocally regulate class switching to IgG1 and IgG2a7, whereas the ability to express TGFRII, the type II receptor for the cytokine TGF- (transforming growth factor-), in naive B cells is required for switching to IgA9. However, it remains to be studied whether programmatic differences initiated at the time of class switching extend into memory B cell compartments to control longevity, cell fate and memory B cell function in an antibody classspecific manner. The differentiation of effector cells of the immune system relies heavily on transcription factors that belong to the following three families: T-box, GATA and ROR 10. Members of all three families share the ability to directly interact with chromatinremodeling machinery to transactivate or repress gene targets in a cell contextdependent manner. Transcriptional regulators from each family induce molecular programs known to direct cells of the
1Department
immune system into functional subsets1113. T-bet, a member of the T-box family, has a critical role in inducing IFN- production by T helper type 1 (TH1) cells, natural killer (NK) cells and CD8+ cells to regulate antiviral immunity 12. T-bet expression by naive B cells is sufficient to promote class switching to IgG2a and is required for IFN--induced production of IgG2a in vivo14,15. In contrast, members of the GATA family are expressed by TH2 cells, basophils and mast cells and are crucial to IL-4 production and immunity to helminthes13. Although members of the ROR family are less well characterized, RORt and ROR are present in cells of the TH17 subset of helper T cells16 and are known to be involved in mucosal immunity to extracellular bacteria11. Although factors that belong to these families have been well characterized in several subsets of cells of the immune response, their roles in memory B cell development and the regulation of memory B cell function remain unknown. Here we focus on IgG2a+ and IgA+ memory B cells and provide evidence for the divergent programming of memory B cell function by the major transcriptional regulators T-bet and ROR. Temporal deletion of T-bet in IgG2a+ memory B cells established a central and selective role for this regulator in the survival and antigen responsiveness of IgG2a+ memory B cells in vivo. Differences in the expression of cytokine receptors, integrins and ROR highlighted the specialized development and unique properties of IgA+ memory B cells. Notably, a role for both T-bet and ROR in persistent BCR transcription indicated both dynamic and ongoing class-specific requirements for each of these factors in IgG2a+ and IgA+ memory B cells, respectively. Thus, we propose that expression of any non-IgM
npg
of Immunology and Microbial Science, The Scripps Research Institute, La Jolla, California, USA. 2Department of Molecular Therapeutics, The Scripps Research Institute, Jupiter, Florida, USA. 3Department of Microbiology & Immunology and Department of Pediatrics, College of Physicians and Surgeons of Columbia University, New York, New York, USA. Correspondence should be addressed to M.G.M.-W. (mcheyzer@scripps.edu). Received 22 December 2011; accepted 27 March 2012; published online 6 May 2012; doi:10.1038/ni.2294
604
VOLUME 13
NUMBER 6
JUNE 2012
nature immunology
Articles
Figure 1 B cellintrinsic T-bet is required for IgG2a formation. (a) Total switched (IgMIgD) and CD138 B220hi polyclonal splenic B cells (Gr-1CD4CD8, CD19+) in C57BL/6 wild-type (WT) mice and Tbx21/ (KO) mice. *P < 0.01 (Mann-Whitney test). (b) Flow cytometry (left) of cells from Rag1/ mice given injection of a 1:1 mixture of splenocytes from CD45.1+ wild-type and CD45.2+ Tbx21/ donors and immediately immunized with NP-KLH in adjuvant; numbers in quadrants indicate percent CD45.2+CD45.1 (Tbx21/) cells (top left) or CD45.2CD45.1+ (wild-type) cells (bottom right). Right, total B cells (Gr-1 CD4CD8, CD19+ or CD138+) from wild-type and Tbx21/ donors before transfer. (c) Splenic NP-specific response of Rag1/ recipient mice to the adoptive transfer of cells and immunization with adjuvant only (Adj) or NP-KLH in adjuvant (NP-KLH + adj), as in b, assessed at day 21 (left); numbers in outlined areas indicate percent switched (IgM IgD) B cells (Gr-1CD4CD8, CD19+ or CD138+). Right, total NP-specific cells before transfer (0), or at day 21 after immunization with adjuvant only (Adj) or in the wild-type or Tbx21/ compartment after immunization with NP-KLH in adjuvant (right two bars). (d) Class-specific frequency (left) and total number (right) of NP-specific switched B220hi B cells after transfer and immunization as in b. Numbers in outlined areas (left) indicate percent cells in each congenic compartment. *P < 0.01 (Mann-Whitney test). (e) Class-specific frequency (left) and total number (right) of switched NP-specific plasma cells (Gr-1CD4CD8, IgMIgD, NP+CD138+) after transfer and immunization as in b. Numbers in outlined areas indicate percent cells in each throughout. *P < 0.05 (Mann-Whitney test). Data are representative of two experiments with three and two mice per experiment (a) or two experiments with three and two Rag1/ recipient mice per experiment (be; mean and s.e.m.).
a
Total switched + B220 B cells
Before transfer WT KO
105 104 103 102
b
* *
CD45.2 (KO)
105 104 103 10 0
2
Before transfer 5913 Total B cells 4013

0 103 104 105 105 108
107
IgG1 IgG2a IgG2b IgG3
106
WT KO
c
105 10
4
CD45.1 (WT)
Adj 0.4 0.1
Total NP-specific switched B cells
NP-KLH + adj 124
104 103 102
103 0 0 102 103 10

4
NP
IgD
10
0 Adj WT KO Day 21
d
10 10 10
5 4 3
10 10 10
4 3
Total NP-specific hi + B220 IgG1 cells KO
Total NP-specific hi + B220 IgG2a cells
KO 21
WT 226
KO 707
WT 404
10
105
10
104
IgG2a
IgG1
102 0
2 3 0 10 10
102 0 0 10
3
104
10
104
10
10
WT KO
10
WT KO
Total switch NP-speci c plasma cells
e
105 104 103
CD45.1 (WT) WT 2910 31 KO

10
5
WT
antibody signifies a transcriptionally regulated commitment event across subspecialized, class-specific memory B cells.
IgG2a
104 103 10
2
102 0 0 102 10
3
RESULTS IgG2a memory formation requires T-bet in B cells To look for class-specific transcriptional programs, we first focused on IgG2a+ memory B cells, given evidence that B cellintrinsic expression of T-bet is sufficient to induce germline IgG2a transcripts15. Furthermore, IFN- provided by CD8+ T cells transferred from OT-I mice (which have transgenic expression of an ovalbuminspecific T cell antigen receptor) skews B cells toward IgG2a class switching in vivo in ways dependent on non-T cell expression of T-bet14. To address the B cellintrinsic requirement for T-bet in IgG2a class switching more directly and in polyclonal mice, we used mice with germline deletion of the gene encoding T-bet (Tbx21/ mice)17. These mice have defects in multiple cell types12, including defects in serum IgG15. Consistent with those characteristics, Tbx21/ mice had significantly fewer IgG3+, IgG2a+ and IgG2b+ class-switched (IgMIgD) CD38hi B cells (also CD19+B220+CD138) than did wild-type mice with intact IgG1+ B cells (Fig. 1a and Supplementary Fig. 1). Thus, the development of these putative memory B cell subsets was compromised in the absence of T-bet with no direct or indirect requirement for T-bet in IgG1 memory development. To provide Tbx21/ B cells a source of wild-type polyclonal T cell help, we generated mixed peripheral chimeras deficient in recombination-activating gene 1 (Rag1/) given equal numbers of wild-type and Tbx21/ spleen cells (Fig. 1b). After immunization of the recipient mice with the hapten NP-KLH (nitrophenylacetyl keyhole limpet hemocyanin), both wild-type and Tbx21/ B cell compartments produced robust class-switched NP-specific B cell responses (Fig. 1c). Notably, the antigen-specific IgG2a+ B cell response was >95% lower (Fig. 1d) and extended across antigen-specific germinalcenter (GC; CD38lo), memory (CD38hi) B cell (Supplementary Fig. 2a) and plasma-cell (CD138hi; Fig. 1e) compartments in the B cell specific absence of T-bet. In contrast, slightly higher IgG1+ NP-specific
289
10
4
548
105
IgG1 IgG2a
IgG1
B cell responses were induced in both donor populations (Fig. 1d) with similar distribution into GC, memory B cell (Supplementary Fig. 2a) and plasma-cell (Fig. 1e) compartments. We found no defects in IgG3+ or IgG2b+ NP-specific B cell compartments in the absence of T-bet (Supplementary Fig. 2b). Furthermore, total class-switched B cells, which would contain the KLH-specific response and polyclonal reactivities in the peripheral chimeras, showed a similar selective defect in IgG2a, with numbers in IgG1, IgG3 and IgG2b B cell compartments equivalent to those of wild-type mice (Supplementary Fig. 3). Therefore, B cellintrinsic expression of T-bet was required selectively for class switching to IgG2a and/or the survival of IgG2a+ B cells after class switching in vivo. Survival and function of IgG2a+ memory B cells require T-bet To enable analysis of T-bet expression in IgG2a + B cells after class switching, we crossed mice with loxP-flanked Tbx21 alleles (Tbx21F/F mice)18 to C57BL/6 mice with tamoxifen-inducible sequence encoding Cre recombinase in the ubiquitously expressed Rosa26 locus (Rosa26CreERT2 mice). Treatment of the intact progeny (called CreERT2 Tbx21F/F mice here) with 4-hydroxy tamoxifen (4-OHT) induced the loss of T-bet protein in >50% of T-bet-expressing splenocytes (Fig. 2a). Treated mice had a similarly lower number of IgG2a+ B cells in with no change in the number of IgG1+ B cells. Notably, all remaining IgG2a+ B cells had expression of T-bet protein (Fig. 2b) equivalent to that of treated CreERT2 Tbx21+/+ mice (data not shown). Thus, ablation of Tbx21 in vivo led to substantial loss of IgG2a+ memory B cells over 4 d with no exogenous stimulation of the BCR. As temporal deletion with 4-OHT in vivo targets all cells, we sought to determine whether B cellspecific loss of T-bet had caused the
605
npg
Articles
Figure 2 T-bet is required for the survival and function of memory B Naive F/F WT F/F IgG2a+ 25 15 WT 100 100 cells in vivo. (a) T-bet expression in total splenocytes (left) from CreERT2 * F/F 20 80 80 Tbx21+/+ (WT) mice and CreERT2 Tbx21F/F (F/F) mice given daily injection 10 15 60 60 of 4-OHT for 3 d, followed by collection of spleens on day 4, assessed 10 40 40 5 after the use of a forward- and side-scatter lymphocyte gate. Middle and 20 5 20 + splenocytes (middle) and IgG2a+ and IgG1+ B cells right, total T-bet 0 0 0 0 + + 2 3 4 5 0 10 10 10 10 0 102 103 104 105 IgG1 IgG2a (right) in the treated mice at left. *P < 0.05 (Mann-Whitney test). (b) T-bet T-bet T-bet expression in B220hiCD38+IgG2a+ B cells (IgG2a+; Gr-1CD4CD8IgM After transfer IgDCD19+CD138) or naive B cells (Naive; Gr-1CD4CD8IgM+IgD+CD soluble NP-KLH Before transfer 104 +B220+) from the treated CreERT2 Tbx21F/F mice in a. (c) Splenic NP 5 10.4 5 251 105 10 10 19 4 4 104 gate (outlined) of IgMIgD B cells (Gr-1CD4CD8, CD19+ or CD138+) 10 10 103 from CreERT2 Tbx21F/F mice 14 d after immunization with NP-KLH, before 3 103 103 10 2 transfer into Rag1/ recipient mice (in d). (d) Expression of CD138 and 2 0 10 10 0 0 487 B220 (middle) on NP-specific cells (left) obtained from Rag1/ recipient 102 2 3 4 5 WT F/F 0 102 103 104 105 0 10 10 10 10 0 102 103 104 105 mice 7 d after transfer of a mixture of the spleen cells in c treated for IgD IgD B220 1 h in vitro with 4-OHT and mixed at a ratio of 1:1 with spleen cells from WT F/F CD45.1+ C57BL/6 mice (also 14 d after immunization with NP-KLH) for * WT F/F 103 104 105 11 105 197 * transfer, followed by boosting of recipients with soluble NP-KLH. Numbers 3 10 104 104 in outlined areas indicate percent NP-specific IgMIgD cells (left) or 102 102 CD138B220+ cells (middle). Right, total NP-specific B220hi B cells from 103 103 101 2 mice as in c (F/F) and from a separate group of Rag1/ recipient mice 101 102 10 0 0 T2 Tbx21+/+ (WT) mice 3716 256 0 that received spleen cells obtained from CreER 100 10 + + + + 0 102 103 104 105 0 102 103 104 105 IgG1 IgG2a IgG1 IgG2a (also 14 d after immunization with NP-KLH) and treated for 1 h in vitro IgG1 with 4-OHT, mixed 1:1 as described above for the F/F group. (e) Frequency (left) and total number (right) of NP-specific IgG2a + and IgG1+ B cells after transfer and boosting from the separate groups of recipients as in d, with gating on B220hi cells. *P < 0.01 (Mann-Whitney test). (f) Total NP-specific plasma cells (CD138+) after the transfer in d. *P < 0.05 (Mann-Whitney test). Data are representative of two experiments with two mice and one mouse per genotype per experiment (a,b) or two experiments with three and two Rag1/ recipient mice per group per experiment (cf; mean and s.e.m.).
Events (% of max) Total T-bet+ spleen cells (107) Total B220hi cells (104)
W T F/ F
Events (% max)
NP
NP
CD138
npg
selective IgG2a deficit. In the next strategy, we induced an NP-specific B cell response in CreERT2 Tbx21F/F and CreERT2 Tbx21+/+ donor mice that had antigen-specific B220hi IgG1+ and IgG2a+ compartments (Fig. 2c,d and Supplementary Fig. 4a). We treated splenocytes from donors of each genotype in vitro for 1 h with 4-OHT. This treatment resulted in excision of >90% of the Tbx21F/F alleles (Supplementary Fig. 4b) and provided a control, in wild-type cells, for the toxicity of this treatment. We transferred each set of treated cells into separate Rag1/ recipient mice together with T-bet-sufficient splenocytes from wild-type congenic mice immunized with NP-KLH. We then challenged the resultant mixed peripheral chimeras 24 h later with soluble antigen without adjuvant (which was unable to induce a primary NP-specific response; data not shown) to promote a secondary
NP-specific B cell response in the recipient mice. Both Tbx21F/F and Tbx21+/+ 4-OHT-treated sources of NP-specific B cells induced secondary responses to soluble antigen with equivalent numbers of NPbinding CD138B220+ B cells and plasma cells (Fig. 2e). However, the NP-specific IgG2a+ recall response required T-bet, with <5% of Tbx21F/F B cells responding after 4-OHT treatment relative to the number of responding wild-type B cells (Fig. 2f ). The NP-specific IgG1 response remained intact and emerged to an extent equivalent to that in the absence of T-bet (Fig. 2f ). We noted the same trends across the IgG2a+ and IgG1+ plasma-cell compartments (Fig. 2f ). We constructed mixed chimeras 6 months after the initial priming of donor mice and found the same selective loss of IgG2a+ NP-specific cells after transfer and immunization (Supplementary Fig. 4c).
Total NP-speci c B cells
105 10 10
4 3 2
0.06 0.01
51
105 104 103
525
105 104
106 105
161
111
Bcl-6 mRNA (relative)
107
250 200 150 100 50 0 N Pre 7 14 M GC 5 14 Recall
Blimp-1 mRNA (relative)
Adj
NP-KLH + adj
NP-specific B cells
c
10
3 2
200 150 100 50 0 N 7 14 5 14 Primary Recall
<PE-A>: NP
CD138
IgG2a
NP
10 0
104 103
428
0 102 103
5 104 10
10 0
0 10
10
104 105
Time (d): 0
IgD
7 7 14 5 14 Adj Primary Recall
103
104 105
B220
CD38
Figure 3 T-bet expression during antigen-specific IgG2a memory development. (a) NP-specific 200 200 B cell response at day 7 in lymph nodes of C57BL/6 mice immunized with adjuvant only or 150 150 NP-KLH plus adjuvant, gated on Gr-1CD4CD8 cells, CD138+ or CD19+ B cells, and 100 100 IgD cells (left). Right, time course of the abundance of total NP-specific B cells before IgM 50 50 transfer (0), after immunization with adjuvant only (Adj) or during the primary and memory (Recall) response in mice immunized as at left with NP-KLH plus adjuvant. (b) NP-specific 0 0 N 7 14 5 14 TFH N Pre 7 14 M 5 14 B cell subsets (left) and B220 expression (right) of CD38 + and IgG2a+ cells at day 7 of the GC Recall Primary Recall response in mice immunized with NP-KLH plus adjuvant. (c) Expression of Bcl-6 mRNA in naive (IgM+IgD+CD23+) cells (N) at day 0 and in NP-specific CD19+B220hiIgG2a+ pre-GC (CD38+) cells at day 7 (Pre), GC (CD38) cells at days 7 and 14, memory (CD38+) cells at day 14 (M) and memory recall (CD38+) cells (Recall) at days 5 and 14 of the response (left), and Blimp-1 mRNA in IgG2a+CD138+ NP-specific plasma cells from mice immunized with NP-KLH plus adjuvant (right). (d) T-bet expression by antigen-responsive TFH cells and populations as in c (left) and by IgG2a+ plasma cells (CD138+; right). Numbers along horizontal axes (a,c,d) indicate time in days; mRNA expression (c,d) is presented relative to that of T-bet in naive B cells, set as 1. Data are representative of at least four experiments with two groups of two to three mice at each time point (mean and s.e.m. in a,c,d).
T-bet mRNA (relative) T-bet mRNA (relative)
606
VOLUME 13
NUMBER 6
JUNE 2012
nature immunology
Total switched NP-specific plasma cells
Total switched NP-specific B220hi cells
IgG2a
Total NP-specific B220hi cells
Articles
Figure 4 Evidence for T-bet activity in IgG2a+ memory B cells. 2 10 (a) Expression of mRNA for T-bet target genes at recall day 5 in 1 10 IgG2a+ memory B cells among NP-specific B cells obtained from the draining lymph nodes of C57BL/6 mice immunized with adjuvant and 1 0 10 10 NP-KLH and purified by flow cytometry, then gated as Gr-1 CD4CD8, 0 1 10 10 1 B cellpositive (CD19+ or CD138+), switched (IgMIgD) and NP+IgG2a+ 10 1 hiCD38+CD138 cells (Mem mRNA; top), presented 10 memory B220 0 10 2 relative to expression in naive B cells obtained and purified as described 10 1 10 above and gated as Gr-1CD4CD8 cells, B220+CD19+CD138 2 10 B cells and IgM+IgD+CD23+ cells. Below (GC mRNA), mRNA expression IgG2a+ memory IgG1+ memory at day 14 in mature IgG2a+ GC (CD38) cells, presented relative to ** 100 10 100 * 10 KO 10 expression in memory cells at recall day 5. (b) Expression of mRNA for 80 8 WT 80 T-bet target genes at recall day 5 as in a in NP+IgG2a+ plasma cells 60 6 60 5 10 40 4 40 (PC mRNA) purified by flow cytometry and gated as Gr-1 CD4CD8 cells, 20 2 20 CD138+ or CD19+ B cells, switched (IgMIgD) cells and CD138+ cells, 0 10 0 0 0 presented relative to expression in IgG2a+ memory cells purified by flow 0 10 10 10 10 T-bet + and IgG1+B220hiCD38+ cytometry as in a. (c) T-bet expression in IgG2a B cells (Gr-1CD4CD8 IgMIgD CD19+CD138) from wild-type and Tbx21/ (control) mice (left). Right, Mean fluorescence intensity (MFI) of T-bet expression in naive, IgG2a + (G2a) and IgG1+ (G1) cells from the mice at left. *P < 0.05 and **P < 0.001 (Mann-Whitney test). (d) Expression of T-bet mRNA and IFN- mRNA by single cells sorted by flow cytometry as Gr-1CD4CD8, B cellpositive (CD19+ or CD138+), switched IgMIgD and B220hiCD138CD38+IgG2a+ (left), presented in arbitrary units (AU); each symbol represents an individual cell, and small horizontal lines indicate the mean. Right, frequency of T-bet + and IFN-+ cells among total cells at left. Data are representative of two experiments with three mice and two mice per experiment (ac; mean and s.e.m.) or two experiments with at least 12 cells from two mice each per experiment (d; mean and s.e.m.).
Mem B mRNA (relative) GC mRNA (relative) PC mRNA (relative)
R ad C 51 xc r Ifn 3 Il2 g rb Ks St r1 M at1 ap C k1 d R k6 un C x1 al Kd m2 m 3 C a Sg cl3 lp lg
Events (% of max)
T-bet (MFI 102)
Expression (AU)
ve G 1 G 2a
Cells (%)
et
Thus, IgG2a+ memory B cells selectively required B cellintrinsic expression of T-bet to respond effectively to antigen rechallenge in vivo. T-bet persists during the development of IgG2a+ memory B cells Next we assessed changes in Tbx21 expression over the course of a primary and memory response in mice immunized with NP-KLH19,20 (Fig. 3a). Differences in the expression of CD138, B220 and CD38 on class-switched (IgMIgD) antigen-binding (NP+) CD19+ B cells provided direct access to IgG2a+ B cells (Fig. 3b). We used expression of the transcription factors Bcl-6 (ref. 21) and Blimp-1 (ref. 22), together with that of antibody-isotype and phenotypic markers, to further distinguish the pre-GC stage (day 7; B220hiCD38hiBcl-6lo), GC stage (days 7 and 14, primary; B220hiCD38lo Bcl-6hi), memory stage (day 14, primary; days 5 and 14, secondary; B220hiCD38hi Bcl-6lo) and plasma-cell stage (CD138hiBlimp-1hi) of the antigenspecific development of IgG2a+ memory B cells (Fig. 3c). As anticipated, primary antigen-responsive IgG2a + B cells had abundant expression of T-bet before entry into GCs, which decreased substantially in the presence of Bcl-6 in GCs at days 7 and 14 after priming (Fig. 3d). The immunization used here, based on an agonist of Toll-like receptor 4, promoted negligible IgM+ memory B cells, with most NP-specific memory B cells expressing immunoglobulins of the IgG isotype (data not shown). After antigen rechallenge, NP-specific IgG2a+ memory B cells had higher expression of T-bet that remained high for at least 14 d after recall (Fig. 3d, right). In the presence of Blimp-1, IgG2a+ plasma cells from all stages of the primary and memory response had low but detectable expression of T-bet (Fig. 3d, left). Thus, T-bet was expressed early after the initiation of class switching to IgG2a and continued to be expressed at all stages of antigenspecific IgG2a+ memory B cell development and responses in vivo. T-bet activity in IgG2a+ memory B cells As evidence of T-bet activity in antigen-specific IgG2a+ CD38hi memory B cells, transcription of a series of genes known to be targets of T-bet23 was higher at day 5 of the memory response in C57BL/6 wildtype mice immunized with adjuvant and NP-KLH (Fig. 4a, top). The differences in the expression of these target genes indicated that T-bet enabled separate functions in IgG2a+ memory B cells and naive
B cells. Studies have shown that Bcl-6 can directly bind to T-bet and repress the expression of T-bet target genes in T cells24. To determine whether a similar event occurs in B cells, we assayed those same T-bet target genes in GC B cells with a greater abundance of Bcl-6 and found that their expression was lower in those cells than in memory B cells (Fig. 4a, bottom). Furthermore, Blimp-1 has been shown to directly antagonize T-bet expression25. Plasma cells with a greater abundance of Blimp-1 protein had lower expression of most of the T-bet target genes assessed than did memory B cells or GC B cells (Fig. 4b). Although T-bet activity was present in IgG2a + memory B cells, expression of Bcl-6 or Blimp-1 resulted in much less T-bet activity in IgG2a+ GC B cells or IgG2a+ plasma cells, respectively. T-bet protein has broad expression across most IgG2a+CD38hi memory B cells, with low but detectable expression in IgG1+CD38hi memory B cells (Fig. 4c). At the single-cell level, most IgG2a+CD38hi memory B cells had intermediate expression of T-bet mRNA, with a minor fraction having substantial expression per cell (Fig. 4d). As anticipated on the basis of the population analysis, many of the T-bet-expressing memory cells also had detectable expression of IFN- mRNA. Thus, it is likely that IgG2a+ memory B cells are a source of IFN- in vivo. The BCR-driven IgG2a response requires T-bet expression As IgM+ memory B cells were not actively excluded from the experiments described above, we purified IgG2a+ B cells from unimmunized CreERT2 Tbx21F/F mice and treated the cells with 4-OHT in vitro. We transferred those treated IgG2a+ memory B cells into Rag1/ recipients and reactivated them through their BCR with antibody to IgG2a (anti-IgG2a). This in vivo stimulus induced robust production of IgG2a+ plasma cells with negligible numbers of CD19+ non-plasma cells remaining 4 d after transfer (Fig. 5a). Nevertheless, after treatment with 4-OHT, there were 75% fewer IgG2a + plasma cells with no expansion of the residual CD19+ B cell compartment. Although cell recovery was low, we noted similar trends in a second series of experiments after sorting NP-specific memory B cells from immunized CreERT2 Tbx21F/F mice, treating the cells with 4-OHT in vitro, transferring the cells into Rag1/ recipient mice and rechallenging the recipient mice with antigen (Supplementary Fig. 5). Furthermore, we noted IgG2a-selective defects in total classswitched B cells with this experimental design (data not shown).
607
npg
Tbe IF t N -
IF N -
N ai
Tb
R ad C 51 xc Ifn r3 g Il2 r Ks b St r1 a M t1 ap C k1 dk R 6 un C x1 a Kdlm2 m C 3a c Sgl3 lp lg
Articles
Before transfer Veh 4-OHT Figure 5 B cellintrinsic T-bet is required for survival, expression of an 10 902 756 10 10 * IgG2a-containing BCR and function. (a) Frequency (middle) and total 10 10 number (left) of IgG2a+ plasma cells (Gr-1CD4CD8CD138+) from 10 10 Rag1/ recipient mice (n = 3) given injection of class-switched memory 10 10 0 IgG2a+ B cells purified by flow cytometry (left; Gr-1CD4CD8IgMIgD 0 10 +B220+CD138CD38+) from the spleens of unimmunized CreERT2 CD19 0 10 10 10 10 0 10 10 10 10 0 10 10 10 10 IgG1 CD138 Tbx21F/F donor mice and treated in vitro with vehicle (Veh) or 4-OHT; 15 150 * 200 100 1 d after cell transfer, recipient mice were injected with anti-IgG2a and WT * * * 80 F/F 150 spleens were collected 3 d after injection. *P < 0.05 (unpaired t-test). 10 100 60 100 (b) Expression of STAT1 mRNA (far left) and mRNA encoding the BCR 40 5 50 50 IgG2a constant region (BCR mRNA; far right), and total cells (middle 20 left) among live (propidium iodidenegative (PI )) IgG2a+ or IgG1+ 0 0 0 0 class-specific CD138B220hiCD38+ B cells from unimmunized CreERT2 + + + + IgG2a IgG1 IgG2a+ IgG1+ IgG2a IgG1 FSC Tbx21+/+ or CreERT2 Tbx21F/F mice (n 3 per genotype), purified by CD4CD8, B cellpositive (CD19+ or CD138+) ** flow cytometry as Gr-1 150 * 10 100 100 WT WT 100 ** F/F 80 F/F 80 and IgMIgD (switched) and cultured for 2 d in the presence of 4-OHT. 75 100 10 60 60 Middle right, forward scatter (FSC) of PI cells sorted after 2 d from the 50 40 40 50 10 wells at left. Expression of mRNA is presented relative to that of wild-type 171 25 20 20 cells, set as 100%. *P < 0.05 (unpaired t-test). (c) Total number of cells 0 0 0 0 10 (far left) and BCR mRNA expression (far right) among the IgG2a + memory + IgG2a IgG2a FSC FSC B cells in b reactivated with anti-IgG2a after 48 h in culture and sorted as cells (middle left) and Dom/Dom in b after 4 d. Middle, forward scatter of total PI Stat1 WT / WT Stat1Dom/Dom Tbx21 Before transfer Day 21 total PI blasting cells (middle), and of actual PI cells sorted for mRNA 217 102 0.50.1 10 15 50 ** * analysis (middle right). *P < 0.05 and **P < 0.001 (unpaired t-test). * * 40 10 10 (d) T-bet expression in splenic switched B220 hiCD38+IgG2a+ B cells from 30 10 C57BL/6 wild-type mice (WT), mice with no translocation of STAT1 to the 20 5 0 10 nucleus (Stat1Dom/Dom) and Tbx21/ mice on the C57BL/6 background. 0 0 0 10 10 10 10 (e) Frequency of IgG2a+ and IgG1+ B220hiCD38+ B cells before (left) + + + + IgG2a IgG1 IgG2a IgG1 T-bet and 21 d after (right) adoptive transfer into Rag1/ recipient mice (n 5). *P < 0.05 and **P < 0.01 (unpaired t-test). Data are representative of three experiments with two recipients each (a), two experiments with three mice per genotype in each (b,c) or three experiments with two mice, two mice and one mouse per genotype per experiment (d,e; mean and s.e.m.).
Total IgG2a+ plasma cells (spleen) BCR mRNA (%)
5 4 3 2
5 4 3 2
IgG2a
IgG2a
STAT1 mRNA (%)
T F/ F W T F/ F
W T F/ F W T F/ F
Events (% of max)
Total cells (103)
50 10 K 0 15 K 0 20 K 0 25 K 0K
Events (% of max)
Total blasting cells
Events (% of max)
Total cells (102)
BCR mRNA (%)
W T F/ F
T F/ F
50 10 K 0 15 K 0 20 K 0 25 K 0K
0 50 10 K 0 15 K 0 20 K 0 25 K 0K
B220hi cells (% CD38+)
5 4 3
Thus, B cellintrinsic T-bet was required for the differentiation of plasma cells driven by the IgG2a-containing BCR in vivo. Transcription factor STAT1, survival and IgG2a require T-bet To examine how T-bet exerted its effect on the fate of IgG2a+ B cells, we used IgG1+ and IgG2a+ memory B cells from wild-type or CreERT2 Tbx21F/F mice for in vitro studies. Deletion of Tbx21 with 4-OHT in vitro resulted in a significant loss of transcription of the gene encoding STAT1 (the signal transducer of IFN-) selectively in IgG2a+ B cells and not IgG1+ B cells, on a per-cell basis (Fig. 5b, left). In the absence of T-bet, there was also a small but significant effect on cell survival in short-term cultures in the presence of the transcription
a
10 10 10
5 4 3
B220 B cells 142 249
hi
b
Events (% of max)
IgG2a B220
100 80 60 40 20 0
0 10
2
hi
IgA B220
hi
Mem GC
0 0 10
2
21
10
3
292
5
10
10
10
10
10
Events (% of max)
100 80 60 40 20 0
CXCR3 (MFI 102)
IgG2a
3 IgA 102
IgA
Ki67
factor BAFF, with a loss of IgG2a+ B cells and more IgG1+ B cells (Fig. 5b, middle). However, over the same period, there was a loss of >75% in mature transcripts encoding the IgG2a constant region on a per-cell basis (Fig. 5b, right). Secondary culture on plates coated with anti-BCR induced a significant number of B cell blasts over 48 h in vitro, whereas IgG2a+ B cells remained small, as assessed by forward scatter, in the absence of T-bet (Fig. 5c). We selected small live cells for quantitative PCR analysis of mature transcripts encoding the IgG2a constant region and detected exaggerated losses in the absence of T-bet (Fig. 5c, right). Thus, in the absence of T-bet in vitro, downregulation of the IgG2a-containing BCR and loss of Stat1 transcription were more pronounced than was the overall loss of IgG2a+ memory B cells. Next we investigated the IgG2a+ B cell compartment in a mouse strain with a phosphorylation defect that prevents translocation of STAT1 to the nucleus (the Domino mutation of Stat1, induced by N-ethyl-N-nitrosourea)26. Although T-bet expression was 50% lower in those mice than in wild-type mice, there was an almost complete loss of IgG2a+ B cells (Fig. 5d) with a compensatory greater abundance of IgG1+ B cells in these mice (Fig. 5e, left). Both trends become
Figure 6 Separable programs for IgG2a+ and IgA+ memory B cells. (a) Frequency of CD38+ and CD38 IgG2a+ B cells from spleen (left) and IgA+ B cells from Peyers patches (right) of unimmunized C57BL/6 mice, gated as Gr-1CD4CD8, B cellpositive (CD19+ or CD138+), switched IgMIgD and class-specific CD138B220hi B cells. (b) Staining of the proliferation marker Ki67 on memory (Mem; CD38 +) and GC (CD38) cells (all gated as in a) from the spleen (left) and Peyers patches (right) in a. (c) Expression of mRNA for cytokine receptors (top) or signal-transduction molecules (bottom) by the IgG2a+ or IgA+ splenic B cells in a, presented relative to that of IgG2a+ B cells. (d) Surface staining of CXCR3 (left) and 47 (right) on IgG2a+ (G2a), IgG1+ (G1), IgA+ (A) or naive splenic B cells. *P < 0.001 (unpaired t-test). Data are representative of two experiments with two mice in per experiment (mean and s.e.m. in c,d).
npg
CD38
10 10
mRNA (relative)
101
0
IgG2a IgA
10 8 6 4 2 0
* *
101 102 103
IgG2a
Ifn g Ifn rI IL grI 22 I IL ra2 17 Tg rc Tg fbrI fb rII
ai
CXCR3 Events (% of max)

100 80 60 40 20 0
010
2
47 (MFI 102)
1 IgA 10
mRNA (relative)
IgG2a IgA
8 6 4 2 0
100
IgG2a
St a S t1 Sm tat4 ad Sm 2 Smad3 ad D 4 ax x
101
ve
ai
4 7
10
10
10
2a G 1 A
2a G 1 A
0 102 103 104 105
ve
608
VOLUME 13
NUMBER 6
JUNE 2012
nature immunology
B220hi cells (% CD38+)
IgG2a
T FF
T F/ F W T F/ F
V 4- eh O H T
Articles
Figure 7 ROR regulates survival and Veh SR1001 *** Veh SR1001 ** 200 150 25 *** 10 BCR expression in IgA+ memory B cells. 1.90.1 105 2.70.3 ** ** 20 8 (a) Expression of mRNA by IgG2a+, IgG1+ or 150 104 ** 100 15 6 IgA+ cells from unimmunized C57BL/6 mice, 100 103 10 4 0.90.2 0.80.2 50 gated as Gr-1CD4CD8 cells, B cellpositive 50 0 5 2 (CD19+ or CD138+) cells, switched IgMIgD 0 0 0 0 +B220hiCD138 B cells, 2 3 4 5 G2a G1 A G2a G1 A G2a G1 A 0 10 10 10 10 and class-specific CD38 IgA+ IgG1+ RORt ROR Foxp3 IgG1 presented relative to expression in naive B cells. (b) Expression of mRNA by total cells (left) or Veh SR1001 150 ROR Ctrl 150 150 150 400 sorted IgA+CD38+B220hi switched B cells (right) 105 * ** ** 101 ** ** 4 300 from pooled Peyers patches and mesenteric 10 100 100 100 100 3 lymph nodes from unimmunized C57BL/6 mice 200 10 ** 50 50 50 50 as in a, assessed 4 d after transfection with 2 100 10 0 control siRNA (Ctrl) or ROR-specific siRNA 0 0 0 0 0 IgA ITPR1 HIF- IgA IgG2a IgG1 IgA 103 104 105 siRNA: 0 (ROR) and presented relative to that of cells siRNA: BCR CD38 transfected with control siRNA, set as 100%. (c) Frequency (left) and total number (right) of class-specific CD38+B220hi splenocytes from unimmunized C57BL/6 mice as in a, after treatment for 6 d with vehicle or 5 M SR1001. (d) Expression of mRNA by CD38+IgA+ cells purified by flow cytometry as in a and treated for 24 h with vehicle or 5 M SR1001. *P < 0.05, **P < 0.01 and ***P < 0.001 (unpaired t-test). Data are representative of two experiments (mean and s.e.m.). with two and three mice per experiment (a), three mice per experiment (b) or one mouse and three mice per experiment (c,d).
GATA-3 mRNA (relative) T-bet mRNA (relative)
2a G 1 A
ROR mRNA (relative)
2a G 1 A
d
ROR mRNA (%)
IgA BCR mRNA (%)
BCR mRNA (%)
C R trl O R
IgA
exaggerated after adoptive transfer of spleen cells from mice with the Domino mutation of Stat1 into Rag1/ mice without immunization (Fig. 5e, right). Thus, both T-bet expression and STAT1 activation seemed necessary for continued expression of the IgG2a-containing BCR and/or survival of IgG2a+ memory B cells in vivo. Separable programs for IgA+ and IgG2a+ memory B cells IgA has a dominant role at mucosal surfaces, binding to a variety of cells of the innate immune response, enhancing phagocytosis and triggering the local release of cytokines and inflammatory mediators27,28. Expression of TGFRII by B cells is required for class switching to IgA9, with some evidence for the involvement of TH17 cells in the coordination of mucosal IgA responses29. Transcription factors of the Runx and Smad families have been shown to promote IgA-specific germline transcripts as precursors for class switching to IgA 8,30. To compare the transcriptional programs of IgG2a+ versus IgA+ memory B cells, we isolated isotype-switched (IgMIgD) CD38+ B cells (also CD19+CD138B220+) that were largely Ki67 (noncycling) as quiescent memory B cell compartments from spleen or Peyers patches (Fig. 6a,b). IgG2a+ memory B cells had higher expression of the interferon receptors IFN-RI and IFN-RII than did IgA+ memory B cells. In contrast, IgA+ memory B cells had higher expression of the interleukin receptors IL-17Rc and IL-22R2 than did IgG2a+ memory B cells, which suggested differences in the responsiveness of these cells to IL-17 and IL-22 (Fig. 6c). Furthermore, whereas both memory subsets had similar expression of TGFR, IgA+ memory B cells had higher expression of TGF- signaling intermediates of the Smad family (Smad2, Smad3 and Smad4) and the TGF--associated adaptor Daxx (Fig. 6c). IgG2a+ memory B cells had the highest expression of the chemokine receptor CXCR3, whereas most IgA + memory B cells had the highest expression of integrin 47 (Fig. 6d). These data suggested that IgG2a+ memory B cells and IgA+ memory B cells had separable growth factor requirements, signaling propensity and tissue-homing potential (Supplementary Fig. 6). Survival and BCR expression in IgA+ memory cells require RORa Among the main transcription factors downstream of TGF- and IL-17 signaling, ROR had higher expression in IgA+ memory cells than in other memory and naive B cells (Fig. 7a). GATA-3, Foxp3 and RORt were detectable only in small amounts in naive B cells and IgG1+, IgG2a+ and IgA+ memory B cells. Germline deletion of the gene encoding ROR in mice produces a complex phenotype with
neurological defects and early death31. Hence, to probe the function of ROR in IgA+ memory B cells, we silenced ROR expression in normal IgA+ B cells through the use of small interfering RNA (siRNA) targeting ROR. We transfected whole mesenteric lymph nodes and Peyers patch cells with ROR-specific siRNA by nucleofection, which resulted in ~50% less ROR mRNA after 4 d of culture in vitro than that in cells transfected with control siRNA (Fig. 7b, left). This degree of ROR knockdown resulted in ~50% less mature IgA mRNA in the total cell population (Fig. 7c). IgA+ memory B cells isolated from these cultures had >80% less mature IgA BCR transcripts on a per-cell basis than did those transfected with control siRNA (Fig. 7b). Thus, we found that IgA+ B cells expressed ROR, along with evidence that it controlled dynamic BCR expression. To interfere with the function of ROR proteins, we used SR1001, which inhibits the transcriptional activity of ROR and RORt32. The presence of this functional inhibitor resulted in selectively fewer IgA+ memory B cells that survived in culture for 6 d, without affecting IgG1+ B cells (Fig. 7c). To establish the influence of this inhibitor on IgA+ memory B cells, we treated purified IgA+ memory B cells with SR1001 over short-term culture. As this drug inhibits only protein function, there was no significant effect on the expression of ROR mRNA (Fig. 7d, right). However, there was significantly less transcription of the known ROR target genes encoding the inositol triphosphate receptor ITPR1 and the transcription factor HIF-1. Notably, there was a significantly lower abundance of mature transcripts encoding the IgA-containing BCR per cell (Fig. 7d). Thus, we found that ROR affected the function of IgA+ memory B cells, along with evidence that it was required for the survival of IgA+ B cells and control of transcription of mRNA encoding the IgAcontaining BCR. DISCUSSION Antigen-specific B cell memory is central to long-term immunoprotection and develops across many different antibody classes. Here we focused on unique properties of IgG2a+ and IgA+ memory B cells to demonstrate that divergent transcriptional regulators maintain class-specific integrity of memory B cell subsets. Both T-bet and ROR regulated the transcription of mature BCRs, which seemed necessary for the survival of IgG2a+ and IgA+ memory B cells, respectively. In addition, we used conditional genetic models with temporal deletion of T-bet to demonstrate defects in the ability of IgG2a+ memory B cells to respond to antigen. Both IgG2a+
609
npg
C R trl O R
mRNA (%)
Total B220hi memory
mRNA (relative)
IgA
Articles
and IgA+ memory B cells expressed molecules of different transcriptional programs that reflected separate potentials for cytokine secretion, trafficking and survival that permitted flexibility in long-term antigen-specific immunoprotection. We propose that the events that lead to class switching after antigen experience lead to the imprinting of molecular programs whose persistence is essential for maintaining the identity of memory B cell subsets. The dynamic regulation of BCR transcription was a common component of both the ROR and T-bet programs. We found consensus binding sites for both transcription factors in the immunoglobulin 3 regulatory region known to control class switching and antibody secretion33. Thus, separable T-bet- and RORdependent transcriptional programs in IgG2a+ and IgA+ memory B cells also controlled the central recognition properties of antigenspecific B cell memory. Cognate follicular helper T cells (TFH cells) regulate many facets of the development of antigen-specific memory B cells4. Bcl-6 programs the development of TFH cells and their movement to B cell areas for cognate control of B cell immunity34. Cytokine production occurs in those follicular regions after contact with antigen-primed B cells in ways that can initiate antibody class switching35. In GCs, contact with TFH cells36 regulates the clonal composition of GC B cells and export into B cell memory compartments37. In GC B cells, Bcl-6 can antagonize the function of T-bet, and this dynamic molecular interaction may underpin the fate of GC B cells and the development of IgG2a+ memory B cells. There is evidence of cytokine-producing GC TFH cells that engage GC B cells in an antibody classspecific manner35. In this scenario, antibody classspecific regulation by cognate TFH cells may streamline affinity maturation and efficiently promote the development of memory B cells. T-bet regulates critical functions in IgG2a+ memory B cells. In naive T cells, signaling through the antigen receptor acts in synergy with IFN-induced activation of STAT1 to promote initial T-bet expression 12. We found a similar regulatory axis dependent on IFN-, STAT1 and T-bet in IgG2a+ memory B cells. Ablating T-bet expression interfered with this program, which resulted in downregulation of BCR transcription and loss of IgG2a+ memory B cells. Although the binding of antigens by memory BCRs is not required for survival38, we propose that memory B cells rely on tonic signaling through the BCR for survival, as do naive B cells39. Similar to antigen-specific plasma cells20, memory B cells may also retain antigen-presenting ability for extended periods in vivo. In this manner, continued BCR expression may also indirectly permit ongoing local contact with antigen-specific TFH cells40. T-bet activity establishes a molecular framework that governs the functions of IgG2a+ B cells. Although T-bet binds to target promoters regardless of cell type, its ability to transactivate and remodel loci is dependent on cell context23. The regulation of genes by T-bet is also coupled to the dynamic expression of its molecular antagonists Bcl-6 and Blimp-1 (refs. 24,25). Notably, T-bet can assort asymmetrically in T cells at cell division41, and Bcl-6 has been shown to assort unevenly across GC B cells42. Our studies have shown that coexpression of Bcl-6 with T-bet in GC B cells resulted in repression of the transcription of T-bet targets. However, rather than abolishing a T-betdefined subset, Bcl-6 transiently and reversibly altered programming of IgG2a to allow productive GC activity and the development of IgG2a+ memory B cells. ROR, like T-bet, is able to recruit chromatin-remodeling machinery and directly transactivate loci11. However, unlike the interactions of T-bet, the interactions of ROR with Bcl-6 and Blimp-1 have not been characterized. Although some studies have indicated Bcl-6 represses RORt and TH17 differentiation43,44, its cell-intrinsic effects on ROR
610
remain unknown. Expression of ROR is required during early development11; hence, mice deficient in ROR have severe developmental defects and die early. Studies of ROR targets in cells of the immune system have provided connections to the activation of IL-17, IL-17F, IL-22 and the receptor IL-23R (ref. 45), and the repression of IL-6 and tumor necrosis factor46. It is unknown whether activation of these targets is direct or indirect, and we found no difference in the expression of these targets in IgA+ memory B cells (data not shown). However, ROR is involved in several other pathways, including calcium signaling, circadian rhythm and cellular metabolism11, and there is evidence of an effect of SR1001 on ITPR1 and HIF-1 that might indicate other pathways in which IgA+ memory B cells may be unique relative to other classes. Antigen-specific memory B cell responses may also require subspecialized regulatory programs organized by antibody class. After antigen rechallenge, memory B cells need cognate T cell help to expand their populations and differentiate into plasma cells47. IgM+ memory B cells have functions and an ability to respond that are different from those of their IgG+ counterparts3,48. Downstream antibody classes engage signaling pathways different from those engaged by IgM based on expression of the constant region of the BCR5,6. Other class-specific memory B cell properties control migration; for example, T-bet drives IgG2a+ memory B cells to inflammatory sites via CXCR3 expression, and the integrin 47 guides IgA+ memory cells to mucosal tissues27. Enhanced transcription of STAT1 and IFN-R may uniquely sensitize IgG2a+ memory B cells to IFN- signals. Similarly, ROR can enhance the sensitivity of cells to calcium49, and the calcium-dependent kinase CamKIV can enhance ROR expression50. Those types of changes in IgA+ memory B cells may lower BCR-activation thresholds in a class-specific manner. Therefore, many attributes of class-specific B cell memory that are introduced after development may be further reinforced after antigen recall under the cognate guidance of memory TFH cells40. Memory B cells are generally considered functionally equivalent, differing only by the antibody isotype they express. Here we have unexpectedly shown that B cell memory was organized in classspecific subsets, each with separate central transcriptional regulators. Specifically, the transcriptional regulators T-bet and ROR controlled divergent IgG2a+ memory B cell subsets and IgA+ memory B cell subsets, respectively, to control separate functions in these unique class-specific memory B cell compartments. T-bet is used by many cell types in response to inflammatory stimuli, with a focus on the clearance of intracellular pathogen12. We have now shown that IgG2a+ B cell memory also relied selectively on a T-bet-dependent program to establish and maintain subset integrity. Similarly, IgA + B cell memory is specialized to protect the mucosal surfaces27, and the selective use of transcriptional regulator ROR enhances this unique memory B cell function. Notably, these unique developmental programs can be exploited for directed immunotherapeutic applications, the future formation of class-skewing vaccines, and the treatment of cancer and autoimmunity. METHODS Methods and any associated references are available in the online version of the paper.
Note: Supplementary information is available in the online version of the paper. AckNOWLedGMeNTS We thank B. Beutler (The Scripps Research Institute) for mice with the Domino mutation of Stat1; and N. Cereb, S.Y. Yang and L. Boring for assistance with construction of the loxP-flanked Tbx21 allele. Supported by the US National
npg
VOLUME 13
NUMBER 6
JUNE 2012
nature immunology
Articles
Institutes of Health (TL1 RR025772-03 to N.S.W.; AI042370 and AI076458 to S.L.R; DK080201 and MH092769 to T.P.B.; and AI047231, AI040215 and AI071182 to M.G.M.-W.), the Swiss National Science Foundation (S.L.O.), Novartis Jubliaeumsstiftung (S.L.O.) and the Roche Research Foundation (S.L.O.). This is manuscript 21154 from The Scripps Research Institute. AUTHOR cONTRIBUTIONS N.S.W., L.J.M.-W. and M.G.M.-W. conceived of and designed the project; N.S.W. did all experiments and analyzed the data for all experiments; L.J.M.-W. identified T-bet in memory B cells and contributed to the preparation of the manuscript; S.L.O. contributed to experimental design; S.L.R. provided Tbx21F/F mice; T.P.B. provided SR1001; and N.S.W. and M.G.M.-W. wrote the manuscript. cOMPeTING FINANcIAL INTeReSTS The authors declare no competing financial interests.
1. Kurosaki, T., Aiba, Y., Kometani, K., Moriyama, S. & Takahashi, Y. Unique properties of memory B cells of different isotypes. Immunol. Rev. 237, 104116 (2010). 2. McHeyzer-Williams, M.G., Okitsu, S.L., Wang, N.S. & McHeyzer-Williams, L.J. Molecular programming of B cell memory. Nat. Rev. Immunol. 12, 2434 (2012). 3. Dogan, I. et al. Multiple layers of B cell memory with different effector functions. Nat. Immunol. 10, 12921299 (2009). 4. Fazilleau, N. et al. The function of follicular helper T cells is regulated by the strength of T cell antigen receptor binding. Nat. Immunol. 10, 375384 (2009). 5. Engels, N. et al. Recruitment of the cytoplasmic adaptor Grb2 to surface IgG and IgE provides antigen receptor-intrinsic costimulation to class-switched B cells. Nat. Immunol. 10, 10181025 (2009). 6. Martin, S.W. & Goodnow, C.C. Burst-enhancing role of the IgG membrane tail as a molecular determinant of memory. Nat. Immunol. 3, 182188 (2002). 7. Snapper, C.M. & Paul, W.E. Interferon-gamma and B cell stimulatory factor-1 reciprocally regulate Ig isotype production. Science 236, 944947 (1987). 8. Stavnezer, J., Guikema, J.E. & Schrader, C.E. Mechanism and regulation of class switch recombination. Annu. Rev. Immunol. 26, 261292 (2008). 9. Cazac, B.B. & Roes, J. TGF-beta receptor controls B cell responsiveness and induction of IgA in vivo. Immunity 13, 443451 (2000). 10. Miller, S.A. & Weinmann, A.S. Common themes emerge in the transcriptional control of T helper and developmental cell fate decisions regulated by the T-box, GATA and ROR families. Immunology 126, 306315 (2009). 11. Jetten, A.M. Retinoid-related orphan receptors (RORs): critical roles in development, immunity, circadian rhythm, and cellular metabolism. Nucl. Recept. Signal. 7, e003 (2009). 12. Lazarevic, V. & Glimcher, L.H. T-bet in disease. Nat. Immunol. 12, 597606 (2011). 13. Zhu, J., Yamane, H., Cote-Sierra, J., Guo, L. & Paul, W.E. GATA-3 promotes Th2 responses through three different mechanisms: induction of Th2 cytokine production, selective growth of Th2 cells and inhibition of Th1 cell-specific factors. Cell Res. 16, 310 (2006). 14. Mohr, E. et al. IFN- produced by CD8 T cells induces T-bet-dependent and independent class switching in B cells in responses to alum-precipitated protein vaccine. Proc. Natl. Acad. Sci. USA 107, 1729217297 (2010). 15. Peng, S.L., Szabo, S.J. & Glimcher, L.H. T-bet regulates IgG class switching and pathogenic autoantibody production. Proc. Natl. Acad. Sci. USA 99, 55455550 (2002). 16. Yang, X.O. et al. T helper 17 lineage differentiation is programmed by orphan nuclear receptors ROR and ROR. Immunity 28, 2939 (2008). 17. Szabo, S.J. et al. Distinct effects of T-bet in TH1 lineage commitment and IFN- production in CD4 and CD8 T cells. Science 295, 338342 (2002). 18. Intlekofer, A.M. et al. Anomalous type 17 response to viral infection by CD8+ T cells lacking T-bet and eomesodermin. Science 321, 408411 (2008). 19. McHeyzer-Williams, L.J., Cool, M. & McHeyzer-Williams, M.G. Antigen-specific B cell memory: expression and replenishment of a novel B220- memory b cell compartment. J. Exp. Med. 191, 11491166 (2000). 20. Pelletier, N. et al. Plasma cells negatively regulate the follicular helper T cell program. Nat. Immunol. 11, 11101118 (2010). 21. Dent, A.L., Shaffer, A.L., Yu, X., Allman, D. & Staudt, L.M. Control of inflammation, cytokine expression, and germinal center formation by BCL-6. Science 276, 589592 (1997). 22. Martins, G. & Calame, K. Regulation and functions of Blimp-1 in T and B lymphocytes. Annu. Rev. Immunol. 26, 133169 (2008). 23. Beima, K.M. et al. T-bet binding to newly identified target gene promoters is cell type-independent but results in variable context-dependent functional effects. J. Biol. Chem. 281, 1199212000 (2006). 24. Oestreich, K.J., Mohn, S.E. & Weinmann, A.S. Molecular mechanisms that control the expression and activity of Bcl-6 in TH1 cells to regulate flexibility with a TFH-like gene profile. Nat. Immunol. 13, 405411 (2012). 25. Cimmino, L. et al. Blimp-1 attenuates Th1 differentiation by repression of ifng, tbx21, and bcl6 gene expression. J. Immunol. 181, 23382347 (2008). 26. Crozat, K. et al. Analysis of the MCMV resistome by ENU mutagenesis. Mamm. Genome 17, 398406 (2006). 27. Cerutti, A. & Rescigno, M. The biology of intestinal immunoglobulin A responses. Immunity 28, 740750 (2008). 28. Fagarasan, S., Kawamoto, S., Kanagawa, O. & Suzuki, K. Adaptive immune regulation in the gut: T cell-dependent and T cell-independent IgA synthesis. Annu. Rev. Immunol. 28, 243273 (2010). 29. Jaffar, Z., Ferrini, M.E., Herritt, L.A. & Roberts, K. Cutting edge: lung mucosal Th17mediated responses induce polymeric Ig receptor expression by the airway epithelium and elevate secretory IgA levels. J. Immunol. 182, 45074511 (2009). 30. Lebman, D.A., Nomura, D.Y., Coffman, R.L. & Lee, F.D. Molecular characterization of germ-line immunoglobulin A transcripts produced during transforming growth factor type -induced isotype switching. Proc. Natl. Acad. Sci. USA 87, 39623966 (1990). 31. Hamilton, B.A. et al. Disruption of the nuclear hormone receptor ROR in staggerer mice. Nature 379, 736739 (1996). 32. Solt, L.A. et al. Suppression of TH17 differentiation and autoimmunity by a synthetic ROR ligand. Nature 472, 491494 (2011). 33. Cogn, M. et al. A class switch control region at the 3 end of the immunoglobulin heavy chain locus. Cell 77, 737747 (1994). 34. Crotty, S. Follicular helper CD4 T cells (TFH). Annu. Rev. Immunol. 29, 621663 (2011). 35. Reinhardt, R.L., Liang, H.E. & Locksley, R.M. Cytokine-secreting follicular T cells shape the antibody repertoire. Nat. Immunol. 10, 385393 (2009). 36. Allen, C.D., Okada, T., Tang, H.L. & Cyster, J.G. Imaging of germinal center selection events during affinity maturation. Science 315, 528531 (2007). 37. Victora, G.D. et al. Germinal center dynamics revealed by multiphoton microscopy with a photoactivatable fluorescent reporter. Cell 143, 592605 (2010). 38. Maruyama, M., Lam, K.P. & Rajewsky, K. Memory B-cell persistence is independent of persisting immunizing antigen. Nature 407, 636642 (2000). 39. Lam, K.P., Kuhn, R. & Rajewsky, K. In vivo ablation of surface immunoglobulin on mature B cells by inducible gene targeting results in rapid cell death. Cell 90, 10731083 (1997). 40. Fazilleau, N. et al. Lymphoid reservoirs of antigen-specific memory T helper cells. Nat. Immunol. 8, 753761 (2007). 41. Chang, J.T. et al. Asymmetric T lymphocyte division in the initiation of adaptive immune responses. Science 315, 16871691 (2007). 42. Barnett, B.E. et al. Asymmetric B cell division in the germinal center reaction. Science 335, 342344 (2012). 43. Nurieva, R.I. et al. Bcl6 mediates the development of T follicular helper cells. Science 325, 10011005 (2009). 44. Mondal, A., Sawant, D. & Dent, A.L. Transcriptional repressor BCL6 controls Th17 responses by controlling gene expression in both T cells and macrophages. J. Immunol. 184, 41234132 (2010). 45. Yang, X.O. et al. Molecular antagonism and plasticity of regulatory and inflammatory T cell programs. Immunity 29, 4456 (2008). 46. Dzhagalov, I., Giguere, V. & He, Y.W. Lymphocyte development and function in the absence of retinoic acid-related orphan receptor . J. Immunol. 173, 29522959 (2004). 47. Aiba, Y. et al. Preferential localization of IgG memory B cells adjacent to contracted germinal centers. Proc. Natl. Acad. Sci. USA 107, 1219212197 (2010). 48. Pape, K.A., Taylor, J.J., Maul, R.W., Gearhart, P.J. & Jenkins, M.K. Different B cell populations mediate early and late memory during an endogenous immune response. Science 331, 12031207 (2011). 49. Gold, D.A. et al. RORalpha coordinates reciprocal signaling in cerebellar development through sonic hedgehog and calcium-dependent pathways. Neuron 40, 11191131 (2003). 50. Kane, C.D. & Means, A.R. Activation of orphan receptor-mediated transcription by Ca(2+)/calmodulin-dependent protein kinase IV. EMBO J. 19, 691701 (2000).
npg
NUMBER 6
JUNE 2012
611
Animals. Tbx21/ mice (Jackson Laboratories), C57BL/6 (B6) mice, Rag1/ mice, B6.CD45.1 mice, Tbx21F/F mice, mice with the Domino mutation of Stat1 and CreERT2 Tbx21F/F mice were housed in pathogen-free conditions. All experiments were done in compliance with federal laws and institutional guidelines as approved by The Scripps Research Institutional Animal Care and Use Committee. CreERT2 Tbx21F/F mice were generated by crossing of Tbx21F/F mice with Rosa26-CreERT2 mice (Jackson Labratories). For genotyping of mutant mice, the following primers were used at a final concentration of 0.4 M in a standard FastStart Reaction Mix according to the manufacturers guidelines (Roche): for B6.129S6-Tbx21tm1Glm/J (Tbx21/) mice, oIMR1719 (common; 5-TGGGCATACAGGAGGCAGCAACAAAT A-3), oIMR1718 (wild-type; 5-GACTGAAGCCCCGACCCCCACTCCT AAG-3) and oIMR1717 (mutant; 5-GCGCGAAGGGGCCACCAAAGA ACGGAG-3); and for B6.129-Gt(ROSA)26Sor tm1(cre/ERT2)Tyj/J (Rosa26CreERT2) mice, oIMR8545 (common; 5-AAAGTCGCTCTGAGTTGTT AT-3), oIMR8546 (wild-type reverse; 5-GGAGCGGGAGAAATGGATA TG-3) and oIMR8547 (mutant reverse; 5-CCTGATCCTGGCAATT TC G-3). For the detection of the unexcised locus in Tbx21F/F mice and cells, primers A (sense; 5-TATGATTACACTGCAGCTGTCTTCAG-3) and B (antisense; 5-CAGGAATGGGAACATTCGCCTGTG-3) were used, and for detection of deletions in the locus, F (sense; 5-AGCCATC TCTCCAGCCTA CA-3) and C2 (antisense; 5-CTCTGCCTCCCATCTC TTAGGAGC-3) were used for amplification. Flow cytometry. Draining lymph nodes and spleens were removed from unimmunized or immunized mice and single-cell suspensions were prepared in PBS with 5% (vol/vol) FBS. Cells (4 108 per ml) were incubated for 15 min with anti-CD16-32 (Fc block; 2.4G2) followed for 45 min at 4 C with the following fluorophore- or biotin-labeled monoclonal antibodies: allophycocyanin-conjugated anti-CD138 (281-2), allophycocyaninindotricarbocyanineconjugated anti-CD19 (1D3), phycoerythrinconjugated anti-CD138 (281-2), phycoerythrinTexas redconjugated anti-B220 (RA3-6B2), fluorescein isothiocyanateconjugated anti-IgG1 (A85-1), anti-IgG3 (R40-8L), Horizon V500conjugated anti-CD8 (53-6.7), Horizon V500conjugated anti-CD4 (Rm4-5) and biotin-conjugated anti IgG2ab (5.7; all from BD Biosciences); phycoerythrin-conjugated anti-T-bet (eBio4B10), phycoerythrin-indotricarbocyanine-conjugated anti-CD45.1 (A20), phycoerythrin-indotricarbocyanine-conjugated anti-CXCR3 (CXCR3-173), fluorescein isothiocyanateconjugated anti-CD45.2 (104), biotin-conjugated anti-IgG2b (RMG2b-1), biotin-conjugated anti-CD45.2 (104) and biotinconjugated anti-IgA (RMA-1; all from Biolegend); Alexa Fluor 700 conjugated anti-CD38 (90) and phycoerythrin-indodicarbocyanineconjugated anti-Gr-1 (Ly6G/C; RB68C5; both from eBiosciences); and allophycocyanin-conjugated-NP, phycoerythrin-conjugated NP, peridinin chlorophyll proteincyanine 5.5conjugated anti-IgM (331.12) and Pacific blueconjugated IgD (11.26; all prepared in-house). Immunoglobulin-specific antibodies were added in a separate step for 45 min at 4 C with normal mouse serum (1:50 dilution) before other reagents were added. Streptavidin-conjugated Qdot 655 (Invitrogen) was used as a second-step visualization reagent. For intracellular staining, cells were fixed, permeabilized and stained according to the protocol in the Foxp3 Staining Buffer Set (eBioscience). Cells were washed and resuspended in FBS in PBS and were analyzed with a FACSAria III with FACSDiva software (BD Biosciences). Data were analyzed with FlowJo software (TreeStar). Quantitative PCR. Cells (5 103) were sorted directly into lysis buffer (Qiagen) and mRNA was purified with RNeasy Kit (Qiagen), then cDNA was prepared as described51 with the First-Strand Synthesis System for RT-PCR (Invitrogen) with random hexamers. Platinum SYBR Green Supermix UDG reaction mix (Invitrogen) and the StepOnePlus Realtime PCR system were used for SYBR Green quantitative PCR, followed by analysis with Step One software (Applied Biosystems). Primers (Supplementary Table 1) were used at a final concentration of 0.25 M. Gapdh (encoding glyceraldehyde phosphate dehydrogenase) was used as an endogenous control. For measurement of T-bet expression, results from naive B cells were assigned a value of 1 and relative expression was assigned
ONLINE METHODS
accordingly. For measurement of STAT1 and IgG2a mRNA after switching, results from cells treated with 4-OHT were assigned a value of 1 and relative expression was assigned accordingly. For analysis of treatment with RORspecific siRNA or SR1001, results from cells treated with control siRNA or vehicle were assigned a value of 1 and relative expression was assigned accordingly. For single-cell quantitative PCR analysis of mRNA, single cells were directly sorted into 2x Reaction buffer containing SSIII RT Platinum Taq (from the SuperScript III One-Step RT-PCR System with Platinum Taq DNA Polymerase Kit; Invitrogen) plus outside primers for genes encoding IFN-, T-bet and GAPDH, at a final concentration of 0.25 M. The 96-well plates were incubated in a thermocycler for 15 min at 50 C, followed by 2 min at 95 C, then for 22 cycles of 15 s at 95 C followed by 4 min at 60 C; 1 l of product was used for the standard quantitative PCR described above. Adoptive transfer. Draining lymph nodes and spleen were removed from unimmunized or immunized mice and single-cell suspensions were prepared in PBS. For the generation of chimeras with germline deletion of specific genes in peripheral cells, 3.5 107 cells from Tbx21/ mice or mice with the Domino mutation of Stat1 and from B6.CD45.1 mice were mixed at a ratio of 1:1 (7 107 cells total) and were injected intraperitoneally into Rag1/ hosts. Mice were then immunized subcutaneously at the base of the tail with 400 g NP-KLH in adjuvant based on monophosphoryl lipid A (MPL). For the generation of chimeras with conditional deletion of genes in the periphery, splenocytes from CreERT2 Tbx21F/F, CreERT2 Tbx21+/+ and B6.CD45.1 mice were treated with 4-hydroxytamoxifen (4-OHT) (Sigma). Unfractionated CreERT2 Tbx21F/F or CreERT2 Tbx21+/+ splenocytes (1 107) were treated for 1 h in 4-OHT-containing media (RPMI medium with 10% FBS, 2 mM l-glutamine and 1 M 4-OHT), then were filtered and mixed at a ratio of 1:1 with 1 107 cells from B6.CD45.1 mice (2 107 total) and transferred by intraperitoneal injection into Rag1/ host mice. Mice were immunized 1 d later with 400 g NP-KLH in PBS. For experiments with depletion of plasma cells and naive B cells, cells were purified by flow cytometry, and 1 106 cells from CreERT2 Tbx21F/F mice were treated for 1 h with 4-OHT or vehicle and transferred by intraperitoneal injection into Rag1/ host mice. For transfer of IgG2a+ B cells, cells were purified by flow cytometry, and 1 104 cells from CreERT2 Tbx21F/F mice were treated for 1 h with 4-OHT or vehicle and transferred by intraperitoneal injection into Rag1/ host mice. Rag1/ hosts were given intraperitoneal injection of 75 g anti-IgG2a (R19-15, BD Biosciences) 1 d after transfer of cells. Immunizations. Mice were immunized subcutaneously at the base of the tail with 400 g NP-KLH (4-hydroxy-3-nitrophenylacetyl (Biosearch) conjugated to keyhole limpet hemocyanin (Pierce)) in MPL-based adjuvant supplemented with trehalose dimycolate (1 mg per 1 mg MPL; Sigma). Antigen rechallenge was done 120 d or more after priming with 400 g NP-KLH in MPL-based adjuvant. Soluble boost was done by mixture of 400 g NP-KLH in PBS. Treatment of mice with 4-OHT. After 4-OHT was dissolved in 100% ethanol at a final concentration of 20 mg/ml, it was injected intraperitoneally into unimmunized CreERT2 mice or CreERT2 Tbx21F/F mice at a dose of 0.5 mg per mouse for the first injection and 0.25 mg/ml for the two subsequent injections. Cell culture. IgG2a+ or IgG1+ cells were purified with a FACSAria III (BD). For survival studies, 8 103 cells were placed in medium (DMEM with 10% FBS, 2mM l-gluatmine and 1 M 4-OHT) and treated for 48 h with 200 ng/ml BAFF (R&D systems), and total cells were counted with a FACSAria III. For gene-expression analysis, 2 104 cells were placed in medium (DMEM with 10% FBS, 2 mM l-gluatmine, 50 M -mercaptoethanol and 1 M 4-OHT) and treated for 48 h with 200 ng/ml BAFF (R&D systems) and live (PI) cells were sorted directly into lysis buffer and processed as described above. For BCR-stimulation studies, 2 104 cells were placed in medium (DMEM with 10% FBS, 2mM l-glutamine and 1 M 4-OHT) and treated for 48 h with 200 ng/mL BAFF (R&D systems), and then were transferred into plates coated with biotin-conjugated anti-IgG2ab (5.7; BD Biosciences) or biotin-conjugated anti-IgG1 (RMG1-1; Biolegend) and supplemented with an additional 200 ng/mL of BAFF, followed by culture overnight at 4 C. Live (PI) cells
npg
nature immunology
doi:10.1038/ni.2294
were sorted directly into lysis buffer 48 h after transfer to coated plates and processed as described above for gene-expression analysis. Analysis with siRNA. For siRNA experiments, Peyers patches and mesenteric lymph nodes were homogenized in medium (DMEM with 10% FBS, 2 mM l-glutamine and 50 M -mercaptoethanol) from unimmunized C57BL/6 mice. Cells (3 106) were transfected with an Amaxa Mouse B cell Nucleofector Kit according to the manufacturers protocol (Lonza). Each well was transfected by nucleofection with 300 pmol Silencer Select Pre-designed siRNA directed against ROR (Ambion) or 300 pmol control siRNA with scrambled sequence. Cells were collected whole or were sorted to purity 4 d after transfection by nucleofection, and mRNA was isolated as above.
SR1001 treatment. SR1001 was dissolved in 10% DMSO, 10% Tween-80 and 80% water. Whole splenocytes were homogenized and 2 107 cells were treated for 6 d with 5 M SR1001 in medium (DMEM with 10% FBS, 2 mM l-glutamine and 50 M -mercaptoethanol). For IgA memory cultures, Peyers patches and mesenteric lymph nodes were homogenized, and IgA+ memory cells were purified by flow cytometry and cultured for 24 h with medium and 5 M SR1001. Live cells were collected and mRNA was obtained and analyzed as described above. Statistics. Mean values, standard error of the mean, unpaired t-tests and Mann-Whitney tests were calculated with Prism software (GraphPad). A P value of less than 0.05 was considered statistically significant.
npg
doi:10.1038/ni.2294
nature immunology
e r r ata a n d c o r r i g e n d a
Erratum: Solving vaccine mysteries: a systems biology perspective

Lydie Trautmann & Rafick-Pierre Sekaly Nat. Immunol. 12, 729731 (2011); published online 19 July 2011; corrected after print 10 November 2011 In the version of this article initially published, the volume number for reference 2 was incorrect. The correct volume number is 12. The error has been corrected in the HTML and PDF versions of the article.
Erratum: Writing a first grant proposal

Julian Gomez-Cambronero, Lee-Ann H Allen, Martha K Cathcart, Louis B Justement, Elizabeth J Kovacs, Kenneth R McLeish & William M Nauseef Nat. Immunol. 13, 105108 (2011); published online 19 January 2012; corrected after print 3 February 2012 In the version of this article initially published, the first authors surname was incorrect and e-mail address was missing. The correct name is GomezCambronero; the e-mail address is julian.cambronero@wright.edu. The error has been corrected in the HTML and PDF versions of the article.
Erratum: SAMHD1 restricts the replication of human immunodeficiency virus type 1 by depleting the intracellular pool of deoxynucleoside triphosphates
Hichem Lahouassa, Waaqo Daddacha, Henning Hofmann, Diana Ayinde, Eric C Logue, Loc Dragin, Nicolin Bloch, Claire Maudet, Matthieu Bertrand, Thomas Gramberg, Gianfranco Pancino, Stphane Priet, Bruno Canard, Nadine Laguette, Monsef Benkirane, Catherine Transy, Nathaniel R Landau, Baek Kim & Florence Margottin-Goguet Nat. Immunol. 13, 223228 (2012); published online 12 February 2012; corrected after print 4 April 2012 In the version of this article initially published, the number for Baek Kims second affiliation is incorrect in the author list. The correct number is 10. The error has been corrected in the HTML and PDF versions of the article.
Erratum: Body-barrier surveillance by the epidermal TCRs

Grzegorz Chodaczek, Veena Papanna, M Anna Zal & Tomasz Zal Nat. Immunol. 13, 272282 (2012); published online 12 February 2012; corrected after print 4 April 2012 In the version of this article initially published, the designation for DETCs that lack a V5 TCR is incorrect. The correct designation is V5. Also, on page 273, right column, second full paragraph, the designation for reporter mice in the first sentence is incorrect. The correct designation is IL-2p8GFP. The errors have been corrected in the HTML and PDF versions of the article.
npg
Corrigendum: The redox-sensitive cation channel TRPM2 modulates phagocyte ROS production and inflammation
Anke Di, Xiao-Pei Gao, Feng Qian, Takeshi Kawamura, Jin Han, Claudie Hecquet, Richard D Ye, Stephen M Vogel & Asrar B. Malik Nat. Immunol. 13, 2934 (2012); published online 20 November 2011; corrected after print 3 February 2012 In the version of this article initially published, the description of the Trpm2/ mice in the first paragraph of the Online Methods was incomplete. That section should read as follows: Trpm2/ mice (obtained from B.A. Miller) were generated and originally provided by GlaxoSmithKline46. Another group has independently generated Trpm2/ mice12; those were not used here. The new reference (46) is as follows: Knowles, H. et al. Transient receptor potential melastatin 2 (TRPM2) ion channel is required for innate immunity against Listeria monocytogenes. Proc. Natl. Acad. Sci. USA 108, 1157811583 (2011). The error has been corrected in the HTML and PDF versions of the article.
621

Imunology

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Imunology

Uploaded by

Copyright:

Available Formats

editOrial

2012 Nature America, Inc. All rights reserved.

nature immunology volume 13 number 6 june 2012

Immune to addiction: the ethical dimensions of vaccines against substance abuse

2012 Nature America, Inc. All rights reserved.

nature immunology volume 13 number 6 june 2012

2012 Nature America, Inc. All rights reserved.

volume 13 number 6 june 2012 nature immunology

2012 Nature America, Inc. All rights reserved.

Box 2 The TA-CD vaccine

nature immunology volume 13 number 6 june 2012

2012 Nature America, Inc. All rights reserved.

volume 13 number 6 june 2012 nature immunology

nEWS AnD vIEWS

DOCKing innate to adaptive signaling for persistent antibody production

BCR CD19 Ig- Ig- Cytoplasm

TLR9 activation by DNA

MyD88 Pyk2 DOCK8

Class switch to IgE

Survival and Ig production

nature immunology volume 13 number 6 June 2012

nEWS AnD VIEWS

2012 Nature America, Inc. All rights reserved.

Dont mess with ERAAP!

volume 13 number 6 June 2012 nature immunology

nEWS AnD VIEWS

2012 Nature America, Inc. All rights reserved.

Dont mess with ERAAP!

volume 13 number 6 June 2012 nature immunology

nEWS AnD VIEWS

QFL-specific CD8+ T cell

No recognition Cell surface AA 190198 FAM49b

2012 Nature America, Inc. All rights reserved.

Qa-1 Endoplasmic recticulum

nature immunology volume 13 number 6 June 2012

nEWS AnD VIEWS

2012 Nature America, Inc. All rights reserved.

Death diverted, but to what?

volume 13 number 6 June 2012 nature immunology

nEWS AnD VIEWS

2012 Nature America, Inc. All rights reserved.

Death diverted, but to what?

volume 13 number 6 June 2012 nature immunology

nEWS AnD VIEWS a b

4SP int Clonal diversion

IELp (TCR+ DN)

IELp (TCR DN)

2012 Nature America, Inc. All rights reserved.

nature immunology volume 13 number 6 June 2012

nEWS AnD VIEWS

Tespa1: another gatekeeper for positive selection

2012 Nature America, Inc. All rights reserved.

volume 13 number 6 June 2012 nature immunology

nEWS AnD VIEWS

Tespa1: another gatekeeper for positive selection

2012 Nature America, Inc. All rights reserved.

volume 13 number 6 June 2012 nature immunology

nEWS AnD VIEWS

nature immunology volume 13 number 6 June 2012

nEWS AnD VIEWS

2012 Nature America, Inc. All rights reserved.

volume 13 number 6 June 2012 nature immunology

Inflammasome jigsaw puzzles