Oncogene (2000) 19, 783 790

2000 Macmillan Publishers Ltd All rights reserved 0950 9232/00 $15.00
www.nature.com/onc

H-, K- and N-Ras inhibit myeloid leukemia cell proliferation by a

p21WAF1-dependent mechanism
M Dolores Delgado1, J Pedro Vaque1, Imanol Arozarena1,2, Marco A Lopez-Ilasaca1,3,
Carlos Mart nez1,4, Piero Crespo1,2 and Javier Leon*,1
1
Grupo de Biologa Molecular del Cancer, Departamento de Biologa Molecular y Unidad Asociada al Centro de Investigaciones
Biologicas (CSIC), Universidad de Cantabria, 39011 Santander, Spain

Mutated ras genes are frequently found in human cancer.

However, it has been shown that oncogenic ras inhibits
growth of primary cells, through pathways involving p53
and the cell cycle inhibitors p16INK4a and p19ARF. We have
analysed the eect of the ectopic expression of the three
mammalian ras genes on the proliferation of K562
leukemia cells, which are decient for p53, p16INK4a,
p15INK4b and p19ARF genes. We have found that high
expression levels of both wild-type and oncogenic H-, Kand N-ras inhibit the clonogenic growth of K562 cells.
Induction of H-rasV12 expression in K562 transfectants
retards growth and this eect is accompanied with an
increase of p21WAF1 mRNA and protein levels. Furthermore, p21WAF1 promoter is activated potently by
oncogenic ras and less pronounced by wild-type ras.
This induction is p53-independent since a p21WAF1
promoter devoid of the p53 responsive elements is still
activated by Ras. Finally, inhibition of p21WAF1 expression by an antisense construct partially overcomes the
growth inhibitory action of oncogenic H-ras. Altogether,
these results indicate that the antiproliferative eect of
ras in myeloid leukemia cells is associated to the
induction of p21WAF1 expression and suggest the existence
of p19ARF and p16INK4a-independent pathways for rasmediated growth inhibition. Oncogene (2000) 19, 783
790.
Keywords: Ras; p21WAF1; myeloid leukemia cells
Introduction
The mammalian ras gene family is composed of three
members, H-, K-, and N-ras that encode closely related
21 kDa proteins. Mutated ras genes carrying point
mutations mainly in codons 12, 13 or 61 are commonly
found in human neoplasias and in experimental cancer
(reviewed in Barbacid, 1989; Leon and Pellicer, 1993).
ras genes harboring such mutations are capable of
transforming immortalized cell lines as NIH3T3
broblast in vitro. However, they induce growth
inhibition and premature senescence in murine and
human primary broblasts (Serrano et al., 1997; Lin et
*Correspondence: J Leon, Dpto. Biolog a Molecular, Facultad de
Medicina, 39011 Santander, Spain
Current addresses: 2Instituto de Investigaciones Biomedicas (CSIC),
Madrid, Spain; 3Department of Cell Biology and Center for Blood
Research, Harvard Medical School, Boston, MA, USA; 4Laboratorio
Immunogenetica, Universidad Federal do Parana, Curitiba, Brasil
Received 12 August 1999; revised 4 November 1999; accepted 18
November 1999

al., 1998). The antiproliferative eect of oncogenic Ras

in primary cells is achieved at least through two
pathways. By the rst one, H-ras (Serrano et al., 1997;
Lin et al., 1998) and the Ras eector, Raf (Zhu et al.,
1998) up-regulate the cyclin-dependent kinases (CDKs)
inhibitor p16INK4a thus blocking Rb phosphorylation
and G1 progression. By the other pathway, H-ras
(Serrano et al., 1997; Lin et al., 1998) and Raf (Lloyd
et al., 1997), up-regulate p53 and subsequently p21WAF1,
an inhibitor of CDKs and a transcriptional target of
p53 (El-Deiry et al., 1993). This second pathway
impinges in the induction of p19ARF (p14ARF in human
cells) (Palmero et al., 1998). p19ARF in turn has been
shown to inactivate Mdm2, resulting in high levels of
p53 (reviewed in Sharpless and DePinho, 1999).
Thus, in murine broblasts devoid of either p53 or
p16INK4a, ras oncogenes provoke cell transformation
instead of growth arrest or senescence. On the other
hand, in human primary broblasts the inactivation of
both the p16INK4a/Rb and p19ARF/p53 pathways is
required for Ras transformation (for reviews see Lloyd,
1998; Malumbres and Pellicer, 1998). So far, the
antiproliferative response induced by activated Ras
has been mainly studied in primary cells and its
relevance to human cancer and especially to hematological malignancies, remains unknown.
The involvement of ras in myeloid cell proliferation
is suggested by the high frequency of N-ras activation
in acute myeloid leukemia (AML) and myelodysplasic
syndromes (20 to 45% of cases). However, ras
mutations are practically absent in chronic myeloid
leukemia (CML). Moreover, there is no correlation
between the presence of activating ras mutations and
the malignant progression of AML (reviewed in
Beaupre and Kurzrock, 1999), casting doubts so as to
the role of Ras oncoprotein in the processes underlying
in myeloid cell malignant transformation.
In an eort to understand Ras function in CML we
have studied the eect of the three ras genes in a
biologically relevant model such as K562 cells. K562
cells do not carry mutated ras genes (Todd and Iland,
1991) and the three ras genes are expressed at similar
levels as we have previously shown (Delgado et al.,
1992). Here we show that H-, K- and N-ras oncogenes
induce growth arrest, activation of p21WAF1 promoter
and up-regulation of its mRNA and protein levels. The
same eect, albeit much less pronounced, was exerted
by wild-type ras. This is a striking nding in view of
the lack of expression of p53 (Law et al., 1993),
p15INK4b (Otsuki et al., 1995), p16INK4a (Ogawa et al.,
1994; Otsuki et al., 1995) and p19ARF (Vonlanthen et
al., 1998; and this study) in K562 cells, indicating that

Ras inhibits growth and up-regulation p21

784

WAF1

in myeloid cells
MD Delgado et al

ras-mediated growth inhibition in myeloid cells takes

place by a p19ARF and p16INK4a independent mechanism.
Even more so, our data could help explain the absence
of activating ras mutations observed in human CML.

Results
Ras genes inhibit clonogenic growth of K562 cells
In order to accurately compare the eect of the three
ras genes on K562 cell growth, we subcloned the three
wild-type ras genes and its corresponding oncogenic
mutants in the same vector (pCEFL). In these
constructs, high constitutive expression of ras is
achieved by the elongation factor 1a promoter. They
were transiently transfected into COS cells and the
level of Ras proteins was analysed by immunoblot with
specic antibodies. As shown in Figure 1, H-, K- and
N-Ras proteins were markedly overexpressed, and the
expression levels of the wild-type proteins and their
mutated counterparts were similar. To assess that all
these constructs yielded similar expression levels in
vivo, we tested the biological activity by a standard
transformation assay in NIH3T3 cells. While the wildtype genes produced only a small number of
transformed foci (15 25 foci/mg of transfected
DNA), codon 12 mutants potently transformed
NIH3T3 and the three genes showed similar transformation potencies (3800 4500 foci/mg DNA) (data not
shown).
We then assessed the eect of ras oncogenes on
clonogenic growth of myeloid leukemia cells. For this
purpose we used the K562-derived cell line K526/S,
capable of growing attached onto tissue culture plates.
Cells were electroporated with equivalent amount
(10 mg) of the ras-encoding expression vectors, as well
as empty vector. Interestingly, the number of K562/S
colonies was drastically reduced upon transfection with
either H-, K- or N-ras oncogenes (Figure 2), only 2
10% of the number of colonies obtained after
transfection with the empty vector. A similar eect
was observed upon transfection with the wild-types.
Although the number of colonies obtained was higher
than with the mutated versions, the reduction was still
remarkable (20 55%) when compared to the empty
vector (Figure 2a). Among the three ras genes, H-ras
was the most potent inhibitor of cell growth (inset in
Figure 2a). A representative view of the K562/S colony
densities appearing after transfection is shown in
Figure 2b. The aforementioned results could also be
due to ras genes inhibiting the adhesion of K562/S cells

Figure 1 Expression of Ras proteins in COS cells transfected

with wild-type and mutated H, K and N-ras cDNAs subcloned
into the pCEFL vector. Ras proteins were analysed by
immunoblot with specic anti H-, K- and N-Ras antibodies.
Control (lane `C') represents samples from COS cells transfected
with the pCEFL vector
Oncogene

Figure 2 Eect of wild-type and mutated ras on the clonogenic

growth of K562 cells. (a) Exponentially growing K562/S cells
were transfected with 10 mg of the pCEFL vector (white bar), the
pCEFL-derived wild-type ras (stripped bars) or mutated-ras
(black bars). Cells were selected with G418 and the colonies were
counted 12 days after transfection. The maximum score (K562/S
cells transfected with pCEFL vector) was set to 100%. Bars
indicate mean standard deviations of four independent experiments. The inset shows the comparison of the growth inhibition
achieved by the oncogenic and wild-type versions of the three ras
genes. Inhibition by H-rasV12 was set to 100%. (b) Photographs
of the colonies arisen from K562/S cells transfected with the
indicated pCEFL-derived constructs after 12 days of selection in
the presence of G418. (c) Soft agar colonies of K562 transfected
with pCEFL or pCEFL H-rasV12. Histogram shows data from a
representative experiment

rather than inhibiting clonogenic growth. To rule out

this possibility, we tested the eect of H-ras V12 on the
growth on soft agar of parental K562 cells as shown in
(Figure 2c). This resulted in a dramatic reduction of
agar-grown K562 colonies, indicating that ras inhibited
K562 growth rather than adhesion to surfaces.
Similar results were obtained using dierent ras
expression vectors such as pBabe-Puro-H-ras V12 (not
shown). We also transfected murine wild-type N-ras
and N-ras K61 genes in inducible expression vectors
under the control of the MMTV promoter, inducible
by dexamethasone (Quincoces et al., 1997). Upon
selection in G418 and dexamethasone, a pronounced
decrease in the number of colonies was found after
transfection with oncogenic N-ras (89%), and to a
lesser extent with normal N-ras (42%), as compared to
the empty vector.

Ras inhibits growth and up-regulates p21

MD Delgado et al

Ras oncogene retards growth of stable transfectants

In order to study the eect of H-ras on K562 growth
rates, we generated sublines in which H-rasV12 was
constitutively expressed. In view of the ras-mediated
growth inhibition, we aimed for relatively low H-Ras
oncoprotein expression levels so we avoided the use of
pCEFL and other commonly used expression vectors,
that carry potent promoters. Therefore, we used
plasmid pT24c, which harbors the entire human HrasV12 gene driven by its own, relatively weak, TATAless promoter (Malumbres and Pellicer, 1998). We coelectroporated pT24c and pSV2neo and selected
transfectants G418-resistant. Several clones were isolated and the presence of exogenous human H-ras gene
was veried by Southern analysis. We chose a clone
termed KT24, which contained exogenous human H-

Figure 3 Eect of stable transfection of constitutive H-rasV12 in

K562 cells. (a) Presence of H-rasV12 exogenous sequences in
K562 stably transfected with the plasmid pT24c. High-molecular
weight DNAs from the indicated cell lines were subjected to
Southern analysis and hybridized to a human H-ras probe. The
size of markers in kb is indicated at the right. Arrows indicate the
extra-bands present in KT24 cells. (b) Immunoblot of H-Ras
protein in K562 and KT24 cell lines. As control, puried H-Ras
protein (`p21ras') was also loaded. (c) Metabolic activity, as
determined by WST-1 dye reduction, of K562 (&) and KT24 (&)
cell lines growing in RPMI supplemented with 0.5% FCS.
Histograms show mean values of three independent experiments,
with triplicate determinations for each point. Bars indicate
standard deviations from the mean

WAF1

in myeloid cells

ras sequences for further studies (Figure 3a). In these

cells, Ras expression levels were barely 2 3-fold over
those found in parental cells, as assayed by immunoblot (Figure 3b). KT24 cells growth rate exhibited no
signicant dierences with parental cells under standard growth conditions. However, when cells were
cultured in low serum (RPMI-0.5% FCS), the growth
rate of KT24 was markedly lower than that of parental
cells (Figure 3c).
As a parallel approach, we stably transfected K562
with the inducible H-rasV12 vector pDxHras. We
selected the subline termed KDxHT9 which showed a
clear induction of H-rasV12 mRNA upon addition of
dexamethasone, as shown by Northern analysis (Figure
4a). A second subline KDxHT18 was selected, in which
the expression of the exogenous ras gene was lower
than that in KDxHT9. The growth rates of K562,
KDxHT9 and KDxHT18 were monitored in the
presence or absence of 1 mM dexamethasone. In the
presence of dexamethasone, KDxHT9 showed the
slowest growth rate, in consistence with the highest
expression of H-ras oncogene (Figure 4b). KDxHT18
grew faster than KDxHT9 but slower than parental
cells, again in agreement with their relative H-ras

785

Figure 4 Eect of inducible expresion of H-rasV12 in K562

cells. (a) Northern analysis showing the induction of H-rasV12
mRNA expression in KDxHT9 and KDxHT18 cell lines exposed
to dexamethasone for the indicated times. Endogenous H-ras
expression in K562 parental cells exposed to dexamethasone
(`Dex') is also shown. Arrowheads indicate the position of 28S
and 18S rRNAs. Lower panel shows the lter stained with
ethidium bromide to assess RNA loading. (b) Cell proliferation
rates for K562 (&), KDxHT18 (~) and KDxHT9 (*) cells in the
absence (black) or the presence (white) of 1 mM dexamethasone.
Cell numbers were determined daily for 4 days. Bars indicate
standard deviations from the mean of three separate experiments
Oncogene

Ras inhibits growth and up-regulation p21

786

WAF1

in myeloid cells
MD Delgado et al

expression levels, further conrming that the expression

of activated ras genes impairs K562 cells growth.
Ras oncogenes do not induce differentiation or apoptosis
in K562
K562 cells can be induced to dierentiate into
erythroid or myelomonocytic phenotypes, a process
associated to growth arrest (Delgado et al., 1992). Thus
it would be conceivable that Ras-mediated growth
arrest of K562 cells is a consequence of the induction
of a dierentiation program. We tested this possibility
by analysing changes in the presence of erythroid
markers (positive benzidine reaction and expression of
e-globin) or myeloid markers (reduction of nitroblue
tetrazolium and expression of vimentin) upon the
induction of ras expression. However, we did not
detect any signicant dierence in dierentiation
markers in the ras-expressing cells as compared to
the parental line (data not shown).
In view of previous reports describing that ras
oncogenes induce apoptosis in some models (reviewed
in Downward, 1998), we next explored whether Ras
could impair cell growth by inducing apoptosis.
However, we did not observe any increase in apoptosis
in dexamethasone-stimulated KDxHT9 cells as determined by DNA laddering (Figure 5a) and DAPI
staining of the nuclei (not shown). To further conrm
this nding, we investigated whether Bcl-2 could rescue
the inhibition of clonogenic growth brought about by
oncogenic Ras. As shown in Figure 5b, co-transfection
of H-ras V12 with an excess of bcl-2 did not modify
Ras-mediated growth inhibition. Altogether these
results suggest that the Ras-mediated growth inhibition
on K562 is not due to the induction of apoptosis.
Ras genes induce the expression of p21WAF1
To explore possible mechanisms for the Ras-mediated
growth inhibition of K562, we analysed whether Ras
could up-regulate inhibitors of cyclin-dependent kinases. The antiproliferative eects of ras oncogenes in
human broblasts require the presence of either p19ARF
or p16INK4a (see Introduction). However, K562 cells are
decient in p16INK4a as well as p15INK4b and p53, while
there are conicting reports as to the expression of
p19ARF (Della Valle et al., 1997; Vonlanthen et al.,
1998). To clear this point we performed Southern blot
analysis of K562 DNA and demonstrated that the
p19ARF gene is homozygously deleted in this cell line
(not shown). Therefore, an involvement of p15INK4b,
p16INK4a, and p19ARF in Ras-mediated inhibition of
K562 growth was ruled out. Thus, it was conceivable
that ras eects could be explained by the induction of
the CDK inhibitor p21WAF1. Northern hybridization
demonstrated that p21WAF1 mRNA expression is upregulated in KDxHT9 cells upon induction of ras
expression by dexamethasone (Figure 6a). As a control
we determined the up-regulation of p21WAF1 mediated
by wild-type p53, using a K562 subline transfected with
a termosensitive p53 mutant (A1 in Figure 6a). A clear
increment of p21WAF1 protein was also observed 12 to
16 h after dexamethasone addition (Figure 6b). The
up-regulation of p21WAF1 mediated by Ras however was
smaller than that mediated by wild-type p53 (Figure
6b, lane A1). We also found that the expression of

Oncogene

Figure 5 The growth inhibition by Ras is not due to apoptosis.

(a) Internucleosomal DNA fragmentation assay for K562 and
KDxHT9 cells. Cells were untreated (`7') or treated (`+') for
48 h with 1 mM dexamethasone as indicated. Positive control (lane
`C') represents a sample from K562 cells exposed for 48 h to
15 nM okadaic acid, an inducer of apoptosis in K562 (Lerga et
al., 1995). (b) Exponentially growing K562/S cells were cotransfected with 5 mg of the pCEFL or pCEFL-H-rasV12 plus
40 mg of the pLTR-bcl2 or pLTRpoly-vector as indicated. Cells
were selected with G418, and the colonies were counted 12 days
after transfection. Colonies were scored as described in Figure 2.
Histograms show data from a representative experiment

p27KIP1, the other CDK inhibitor expressed in K562,

did not change upon induction of ras expression
(results not shown).
To determine whether the increase in p21WAF1 levels
induced by Ras is due to a transcriptional upregulation we assayed the activity of p21WAF1 promoter.
The results in Figure 6c show that wild-type ras
activated p21WAF1 promoter (5 14-fold), while stronger
up-regulation (25 32-fold) was found with the three
oncogenic ras versions. Other less potent ras expression
vectors induced weaker p21WAF1 transactivation (data
not shown). As an additional control, p53 induced a
very strong activation of the p21WAF1 promoter (Figure
6c).
The p21WAF1 reporter plasmid used in the previous
experiments contains a 2.3 kb promoter fragment. To
more precisely dene the Ras-responsive region we
constructed and tested a series of promoter deletions.
The Figure 7 shows the results with two of them:
p21DSac-Luc, which only lacks the 5' p53-responsive
element (containing 2.2 kb of the promoter) and
p21DPst-Luc, which contains the proximal 0.21 kb of
the promoter (Figure 7a). H-RasV12 transactivated to
a similar extent the three reporters. The eect of H-,

Ras inhibits growth and up-regulates p21

MD Delgado et al

WAF1

in myeloid cells

787

Figure 6 Ectopic expression of ras up-regulates p21WAF1 in

K562 cells. (a) Northern analysis of p21WAF1 mRNA expression
in KDxHT9 cells exposed to dexamethasone for the indicated
times. The lane labeled `A1' represents a sample from Kp53A1
cells (K562 cells expressing a p53Val135 mutant which induce
p21WAF1 expression when incubated at 328C (Ceballos and Leon,
submitted)) used as positive control for p21WAF1 expression. The
lane labeled `K562' shows the low expression levels of p21WAF1
mRNA in the parental K562 cell line. Arrowheads indicate the
position of 28S and 18S rRNAs. (b) Immunoblot of p21WAF1
protein in KDxHT9 cells exposed to dexamethasone for the
indicated times. `A1' is the positive control for p21WAF1
expression as described above. (c) Analysis of p21WAF1 promoter
activity in K562 cells co-transfected with the pGL3-basic vector
(`pGL3') or pGL3-Waf1-Luc (`p21-Luc') and the pCEFL-derived
wild-type or mutated ras vectors as indicated, or the pLPC-hp53
(`p53'). The p21WAF1 promoter activity was determined as
described in Materials and methods. The luciferase activity value
for K562 co-transfected with p21-Luc and H-rasV12 was set at
100% and the other data were normalized to this value. Bars
indicate the standard deviations from the mean of seven
independent experiments for mutated ras and three independent
experiments for wild-type ras

K- and N-Ras oncoproteins were compared on the

shorter p21DPst-Luc reporter, and we found no
signicant dierences (Figure 7b). Thus, this 0.21 kb
region must contain the Ras-responsive element(s).
To investigate whether p21WAF1 up-regulation was
responsible for Ras-mediated growth inhibition, we
transfected K562/S cells with a p21WAF1 expression
vector. p21WAF1 caused inhibition of colony formation
(Figure 8a) although this inhibition was smaller than
that induced by H-rasV12. We also cotransfected
K562/S cells with H-rasV12 and an expression vector
carrying an antisense p21WAF1 (pPuro-asp21). It was
found that the suppression of p21WAF1 by the antisense

Figure 7 Region of p21WAF1 promoter responsible for Rasmediated up-regulation. (a) Scheme of the p21WAF1 reporter
plasmids used. The p53 responsive elements are indicated. (b)
Analysis of p21WAF1 promoter activity in K562 cells cotransfected with the above p21WAF1 reporter plasmids and the
pCEFL-derived mutated ras vectors as indicated. The p21WAF1
promoter activity was analysed as described in Figure 6c. Bars
indicate the standard deviations from the mean of four
independent experiments

vector was capable of partially overcoming the growth

inhibitory eects of H-rasV12 (Figure 8b). Altogether,
the results suggest that the antiproliferative eect of ras
in these cells is at least partially mediated by the
induction of p21WAF1 expression.
Discussion
In this work we describe that ras genes induce cell
growth inhibition of K562 leukemia cells by a novel
mechanism independent of p53, p19ARF, p15INK4b and
p16INK4a, that may be related to p21WAF1 induction.
We show that high expression levels of oncogenic
mutated versions of H-, K- and N-ras inhibited K562
clonogenic growth. Likewise, overexpression of wildtype ras genes also impaired clonogenic growth but
their eect was weaker as compared to their mutated
counterparts. This growth arrest does not seem to be
associated with the induction of dierentiation or the
triggering of apoptotic processes. Although growth
inhibition mediated by H-ras is not unprecedented, this
is to our knowledge the rst report showing a
comparison of the antiproliferative eects of the three
ras genes and oncogenes. Interestingly, there are
dierences in the growth inhibitory capacity of the
three ras on K562 cells, H-ras being the most potent
inhibitor. It is noteworthy that there is a dierential
expression of the three ras genes during dierentiation
of K562, in which H-ras but not K- or N-ras is downregulated (Delgado et al., 1992).
Oncogene

Ras inhibits growth and up-regulation p21

WAF1

in myeloid cells
MD Delgado et al

788

Figure 8 Antisense-p21WAF1 partially overcomes the growth

inhibitory eect of H-rasV12. Exponentially growing K562/S
cells were transfected with 10 mg of pCEFL, pCEFL-H-rasV12
(`H-rasV12') or pCEFL-hp21 (`p21WAF1') as indicated (a) or cells
were co-transfected with 2 mg of pCEFL or pCEFL-H-rasV12
plus 40 mg of pPuro-antisense-p21 (`as-p21') or the pBabe-Puro
vector (`pBP') (b). Cells were selected with G418, and the colonies
were counted 12 days after transfection. Colonies were scored as
described in Figure 2. Bars indicate standard deviations from the
mean of three separate experiments

It has been an enigmatic nding that ras mutations are

frequent in AML but they are absent in CML, even in the
blastic phase of the disease (Watzinger et al., 1994) even
though CML in blast crisis share many morphological
and biological properties with AML cells. The results
presented here, showing that activated ras have growth
inhibitory eects in K562, may help to explain the
absence of ras mutations in CML. Interestingly, in AML
the frequencies of ras mutations depend on the particular
member of the family involved. As such, mutations in Nras are the most prevalent, while those in K-ras are less
frequently found and mutations in H-ras are very rare
(Beaupre and Kurzrock, 1999). These frequencies are
consistent with our results in K562, where H-ras
oncogene is the most potent growth inhibitor, followed
by K- and N-ras.
An antiproliferative eect of ras oncogenes has been
previously reported for murine and human primary
broblasts and shown to be dependent on the presence
of p19ARF, p16INK4a and p53. In murine cells lacking one
of these genes, the antiproliferative or senescence
response to oncogenic ras is abolished and ras induces
cell transformation instead (reviewed in Malumbres
and Pellicer, 1998). In human broblasts, by contrast,
the disruption of both p19ARF/p53 and the p16INK4a/Rb
pathways is required to prevent ras-mediated growth
arrest (Serrano et al., 1997). Likewise, transfer of
Oncogene

mutated ras genes into bone marrow-derived cells did

not lead to cell proliferation but promotes myeloid cell
dierentiation or growth retardation (Pierce and
Aaronson, 1985; Hawley et al., 1995). In contrast to
primary broblasts and bone marrow cells, K562 are
devoid of p53, p15INK4b, p16INK4a and p19ARF. To our
knowledge this is the rst report of an antiproliferative
eect of oncogenic ras in cells lacking these four genes,
demonstrating that ras oncogenes can block cellular
growth by mechanisms other than those so far
described. H-ras oncogenes also impairs growth of
human monocytic U937 cells (Maher et al., 1996).
However U937, unlike K562, express p16INK4a (Juan et
al., 1998) which may explain the antiproliferative eect
of Ras in U937 through the p16INK4a/Rb pathway.
How does Ras inhibit K562 proliferation in spite of
the absence of p19ARF and p16INK4a activities? The most
likely explanation involves the CDKs inhibitor p21WAF1,
which expression is up-regulated in K562 upon the
induction of ras. In this line, our data indicate that the
p21WAF1 promoter is potently activated by ras. The idea
that p21WAF1 is responsible for Ras-mediated growth
inhibition is supported by the abrogation of Ras
growth-inhibitory eect by anti-sense p21WAF1.
It has been reported that ras induces p21WAF1
expression in primary broblasts (Serrano et al., 1997;
Lin et al., 1998) and Raf does so in primary Schwann
cells (Lloyd et al., 1997). However, in both cases p21WAF1
up-regulation was p53-dependent. This is consistent with
p21WAF1 being a target gene of p53 (El-Deiry et al., 1993),
and ras up-regulating p53 (Serrano et al., 1997; Lin et al.,
1998). In contrast, in K562 the elevation of p21WAF1
proceeds through a p53-independent pathway, because
(a) K562 cells are p53-decient and (b) a p21WAF1
promoter devoid of the p53 responsive elements is still
activated by Ras in K562 (this study). A number of
responsive elements have been described in p21WAF1
human promoter (Gartel and Tyner, 1999). We have
found that Ras up-regulation depends on a promoter
fragment of only 0.21 kb. Further work is aimed to
precisely dene the Ras-responsive sequences. It has also
been described that high-intensity Raf signals upregulates p21WAF1 in a p53-independent fashion in murine
primary broblasts (Sewing et al., 1997; Woods et al.,
1997), although its dependence from p15INK4b, p16INK4a or
p19ARF was not explored. It is reported that Rasmediated up-regulation of p21WAF1 is blocked by Rho
(Olson et al., 1998). However, we previously showed that
K562 express Rho and that Rho ectopic expression
induces apoptosis in these cells (Esteve et al., 1998).
Altogether our results strongly suggest that oncogenic Ras impair growth of CML-derived cells through
the elevation of p21WAF1. Other mechanisms could
operate subsequently to p21WAF1 mediated growth
arrest, as autophagic degeneration (Chi et al., 1999)
or cellular senescence (Serrano et al., 1997; Fang et al.,
1999) although in murine broblasts p21WAF1 is not
essential for Ras-mediated senescence (Pantoja and
Serrano, 1999). In conclusion, we propose a new
pathway to explain the growth arrest mediated by
oncogenic Ras, as schematized in Figure 9. In some cell
types Ras inhibits growth through mechanisms dependent on p19ARF (pathways A and B in Figure 9) or on
p16INK4a (pathway C). In other cells as CML-derived
cells, Ras would lead to p21WAF1 up-regulation in a way
independent from p16INK4a, p19ARF and p53 (pathway

Ras inhibits growth and up-regulates p21

MD Delgado et al

D). Moreover our results in K562 suggest that this

eect is, to some extent, p21WAF1-independent (E).

WAF1

in myeloid cells

and carries human bcl2 cDNA in the pLTR poly vector

(Koskinen et al., 1995). The expression vector for antisensep21 (pPuro-asp21) was provided by A Carnero (Institute of
Child Health, London, UK).

789

Materials and methods

Cell lines and cell culture
K562 cell line derives from a chronic myeloid leukemia and
was purchased from American Type Culture Collection.
K562/S is a K562-derived cell line able to grow attached to a
plastic surface. The line was selected after multiple passages
on tissue culture plates, removing the non-adherent cells.
Cells were grown in RPMI-8% fetal calf serum (FCS).
Clonogenic assays on soft agar of K562 were carried out
using Iscove's medium supplemented with 15% FCS and
0.3% agar (Difco). NIH3T3 cells were grown in DMEM10% newborn calf serum and COS cells in DMEM-10%
FCS.
Expression vectors
cDNAs corresponding to human H-, K- and N-ras protooncogenes and activated oncogenes were cloned into the
pCEFL vector (Teramoto et al., 1996). This is a derivative of
pcDNA3 where the expression is driven by the human
elongation factor 1a promoter. cDNAs encoding human
wild-type H-, K- and N-ras (provided by R Weinberg,
Massachusetts Institute of Technology, Boston, MA, USA),
human K-rasV12 (K-ras4B isoform) (provided by R Weinberg) and N-rasE12 (provided by J de Gunzburg, Institute
Curie, Paris, France) were isolated as 0.6 kb BamHI-NotI
fragments. The human cDNA encoding for mutant H-rasV12
(provided by M Serrano, Centro Nacional de Biotecnolog a,
Madrid, Spain) was isolated as a 0.6 kb EcoRI-BamHI
fragment. pcDNA3-H-rasV12 has been described (Crespo et
al., 1994a). pBabe-Puro H-rasV12 was provided by M
Serrano and carries human H-rasV12 cDNA in the pBabePuro vector (Morgenstern and Land, 1990). The entire
human H-rasV12 oncogene, including its promoter (plasmid
pT24-c) was provided by M Barbacid (Instituto Carlos III,
Madrid, Spain). pDxH-ras have been previously described
(Quincoces et al., 1997), pLTR-bcl2 was provided by P
Koskinen (Turku Center for Biotechnology, Turku, Turkey)

Transfections and clonogenicity assays

For clonogenicity assays K562 or K562/S cells (26106) in
exponential growth were resuspended in 0.8 ml of RPMI-8%
FCS containing the appropriate amount of the indicated
plasmids and electroporated at 260 V and 1 mFa with a BioRad electroporator. The plasmids were prepared by the
Qiagen method. Forty-eight hours after electroporation,
G418 (500 mg/ml) or puromycin (1 mg/ml) was added and
the colonies counted 12 days after transfection. For stable
transfection, 107 cells were electroporated with the indicated
plasmids, selected with G418 as above and cell clones were
isolated by limiting dilution. COS cells were transfected with
DEAE-dextran (Crespo et al., 1994a). Transformation assays
of ras genes in NIH3T3 cells were carried out as described
(Crespo et al., 1994b).
Assessment of cell growth, differentiation and apoptosis
Growth was determined by cell counting in hemocytometer
and by the determination of metabolic activity by the
reduction of the tetrazolium salt WST-1 (Roche Molecular
Biochemicals). Morphological dierentiation was monitored
by examining xed slide preparations stained with MayGrunwald Giemsa and assessed using established cytological
criteria. Erythroid dierentiation was also assessed by the
fraction of hemoglobin-producing cells scored by benzidine
staining and by expression of e-globin as described (Delgado
et al., 1992). Monocytic dierentiation was also assessed by
the fraction of cells positive for the nitroblue tetrazolium dye
reduction and the expression of vimentin (Delgado et al.,
1992). The presence of internucleosomal fragmentation was
assayed as described (Lerga et al., 1995).
Northern and Southern analysis
Total RNA was isolated from cells by the acid guanidine
thiocyanate method and Northern blots were prepared
according to standard procedures. The probes for human
H-ras, e-globin and vimentin were as described (Delgado et
al., 1992). The probe for human p21WAF1 was a 0.5 kb
BamHI-EcoRI from plasmid pWZL-Hygro-hp21 (provided
by M Serrano). For Southern, 10 mg of genomic DNA were
digested with restriction enzymes as indicated, electrophoresed, blotted onto nitrocellulose and hybridized to 32P-labeled
probes.
The probe for human p19ARF was a 190 bp fragment of the
exon 1b obtained by PCR (forward primer: 5'-TGGTGCGCAGGTTCTTGGTGA-3', reverse primer: 5'-GTCTTCTAGGAAGCGGCTGCT-3').
Immunoblots

Figure 9 Ras may induce growth arrest through dierent

pathways. A In primary cells, Ras and its downstream eectors
Raf and MEK/MAPK induce p53 through p19ARF. B p19ARF
activates p53 through the inhibition of Mdm2, which is required
for p53 degradation. C Ras or Raf can induce p16INK4a in
primary broblasts, that in turn induces growth arrest. D As
suggested by the present work on K562 cells, Ras can also induce
p21WAF1 through a pathway independent from p19ARF and p53.
This eect is, to some extent, p21WAF1 independent E

Cell lysates were prepared as described (Serrano et al., 1997).

30 50 mg of protein per lane were separated on SDS PAGE
gels and transferred to PVDF membranes by standard
procedures. p21WAF1 was detected by the WAF-1 Ab-1
monoclonal antibody (Oncogene Research, Calbiochem).
Anti-K and N-Ras antibodies were from Oncogene Research.
Anti-H-Ras (F235 monoclonal antibody) and anti-p27KIP1
were from Santa Cruz. Immunoblots were revealed by
chemiluminescency (ECL, Amersham).
Transactivation assays
K562 cells (46106) were transfected by electroporation at
260 V and 1 mFa with 430 nmol (corresponding to 1.4 to
Oncogene

Ras inhibits growth and up-regulation p21

790

WAF1

in myeloid cells
MD Delgado et al

2 mg) of reporter plasmids and 2 to 10 mg of pCEFL-derived

ras expression vectors of human p53 wild-type expression
vector (pLPC-hp53wt, provided by M Serrano) as indicated.
The reporter plasmids used were: pGL3-Waf1-Luc (provided
for M Oren, Weizmann Institute, Rehovot) which contains a
2.3 kb genomic p21WAF1 DNA fragment (El-Deiry et al.,
1993) cloned into the pGL3-basic vector (Promega) and two
deletions derived from this plasmid. The construct termed
p21DSac-Luc was prepared by digestion with SacI and
recircularization. The construct termed p21DPst-Luc, was
prepared by digestion with BglII and PstI, blunting with
Klenow enzyme and religation. Constructs were conrmed by
sequencing. Transfected cells were seeded in duplicated plates
and 36 h after transfection cells were lysed with 80 ml of
passive lysis buer (Dual Luciferase Reporter System,
Promega) and the rey and Renilla luciferase activities were
measured in a Turner luminometer. To normalize for
transfection eciencies, 1 mg of the pRL-TK plasmid

(Promega) was co-transfected in each case. The co-transfection with ras expression vectors did not modify signicantly
the Renilla luciferase activity values. Promoter activity was
dened as the ratio between the light units generated by the
rey and the Renilla luciferases.

Acknowledgments
We thank Rosa Blanco for expert technical assistance and
Mariano Barbacid, Amancio Carnero, Jean de Gunzburg,
Paivi Koskinen, Moshe Oren, Manuel Serrano and Robert
Weinberg for plasmids. We are also indebted to Manuel
Serrano for helpful discussions. This work has been
supported by grants PM98-0109 from Spanish Ministry
of Education and Culture and Biomed 96-3532 from
European Community (J Leo n) and from Fundacion
Marcelino Botin (P Crespo).

References
Barbacid M. (1989). Annu. Rev. Biochem., 56, 779 827.
Beaupre DM and Kurzrock R. (1999). J. Clin. Oncol., 17,
1071 1079.
Chi S, Kitanaka Ch, Noguchi K, Mochizuki T, Nagashima
Y, Shirouzu M, Fujita H, Yoshida M, Chen W, Asai H,
Himeno M, Yokoyama S and Kuchino Y. (1999).
Oncogene, 18, 2281 2290.
Crespo P, Xu N, Simonds WF and Gutkind JS. (1994a).
Nature, 369, 418 420.
Crespo P, Xu N, Daniotti JL, Troppmair J, Rapp UR and
Gutkind JS. (1994b). J. Biol. Chem., 269, 21103 21109.
Delgado MD, Quincoces AF, Gomez-Casares MT, Martinez
CA, Cuadrado MA, Richard C and Leon J. (1992). Cancer
Res., 52, 5979 5984.
Della Valle V, Duro D, Bernard O and Larsen CJ. (1997).
Oncogene, 15, 2475 2481.
Downward J. (1998). Curr. Opin. Genet. Dev., 8, 49 54.
El-Deiry WS, Tokino T, Velculescu VE, Levy DB, Parsons
R, Trent JM, Lin D, Mercer WE, Kinzler KW and
Vogelstein B. (1993). Cell, 75, 817 825.
Esteve P, Embade N, Perona R, Jimenez B, del Peso L, Leon
J, Arends M, Miki T and Lacal JC. (1998). Oncogene, 17,
1855 1869.
Fang L, Igarashi M, Leung J, Sugrue MM, Lee SW and
Aaronson SA. (1999). Oncogene, 18, 2789 2797.
Gartel AL and Tyner AL. (1999). Exp. Cell Res., 246, 280
289.
Hawley RG, Fong AZ, Ngan BY and Hawley TS. (1995).
Oncogene, 11, 1113 1123.
Juan G, Ardelt B, Li X, Mikulski SM, Shogen K, Ardelt W,
Mittelman A and Darzynkiewicz Z. (1998). Leukemia, 12,
1241 1248.
Koskinen PJ, Ayer DE and Eisenman RN. (1995). Cell
Growth Dier., 6, 623 629.
Law JC, Ritke MK, Yalowich JC, Leder GH and Ferrell RE.
(1993). Leuk. Res., 17, 1045 1050.
Leon J and Pellicer A. (1993). In: The ras Superfamily of
GTPases. Lacal JC and McCormick F. (eds). CRC Press:
Boca Raton, pp. 3 36.
Lerga A, Belandia B, Delgado MD, Cuadrado MA, Richard
C, Ortiz JM, Martin-Perez J and Leon J. (1995). Biochem.
Biophys. Res. Commun., 215, 889 895.
Lin AW, Barradas M, Stone JC, van Aelst L, Serrano M and
Lowe SW. (1998). Genes Dev., 12, 3008 3019.

Oncogene

Lloyd AC, Obermuller F, Staddon S, Barth CF, McMahon

M and Land H. (1997). Genes Dev., 11, 663 677.
Lloyd AC. (1998). Curr. Opin. Genet. Dev., 8, 43 48.
Maher J, Baker D, Dibb N and Roberts I. (1996). Leukemia,
10, 83 90.
Malumbres M and Pellicer A. (1998). Front. Biosci., 3, 887
912.
Morgenstern JP and Land H. (1990). Nucl. Acids Res., 18,
3587 3596.
Ogawa S, Hirano N, Sato N, Takahashi T, Hangaishi A,
Tanaka K, Kurokawa M, Tanaka T, Mitani Ki, Yazaki Y
and Hirai H. (1994). Blood, 84, 2431 2451.
Olson MF, Paterson HF and Marshall Ch J. (1998). Nature,
394, 295 299.
Otsuki T, Clark HM, Wellmann A, Jae ES and Raeld M.
(1995). Cancer Res., 55, 1436 1440.
Palmero I, Pantoja C and Serrano M. (1998). Nature, 395,
125 126.
Pantoja C and Serrano M. (1999). Oncogene, 18, 4974 4982.
Pierce JH and Aaronson SA. (1985). Mol. Cell. Biol., 5, 667
674.
Quincoces AF, Polanco I, Thomson T and Leon J. (1997).
FEBS Lett., 416, 317 323.
Serrano M, Lin AW, McCurrach ME, Beach D and Lowe
SW. (1997). Cell, 88, 593 602.
Sewing A, Wiseman B, Lloyd AC and Land H. (1997). Mol.
Cell. Biol., 17, 5588 5597.
Sharpless NE and DePinho RA. (1999). Curr. Opin. Genet.
Develop., 9, 22 30.
Teramoto H, Crespo P, Coso OA, Igishi T, Xu N and
Gutkind JS. (1996). J. Biol. Chem., 271, 25731 25744.
Todd AV and Iland HJ. (1991). Leuk. Lymph., 3, 293 300.
Vonlanthen S, Heighway J, Tschan MP, Borner MM,
Altermatt HJ, Kappeler A, Tobler A, Fey MF, Thatcher
N, Yarbrough WG and Betticher DC. (1998). Oncogene,
17, 2779 2785.
Watzinger F, Gaiger A, Karlic H, Becher R, Pillwein K and
Lion T. (1994). Cancer Res., 54, 3934 3938.
Woods D, Parry D, Cherwinski H, Bosch E, Lees E and
McMahon M. (1997). Mol. Cell. Biol., 17, 5598 5611.
Zhu J, Woods D, McMahon M and Bishop JM. (1998).
Genes Dev., 12, 2997 3007.