Veterinary Dermatology 2003, 14, 47 56

Detection of papillomavirus-DNA in mesenchymal tumour cells and

not in the hyperplastic epithelium of feline sarcoids
Blackwell Science, Ltd

JENS P. TEIFKE,* BEVERLY A. KIDNEY, CHRISTIANE V. LHR and

JULIE A. YAGER
*Bundesforschungsanstalt fr Viruskrankheiten der Tiere, Friedrich-Loeffler-Institut, Boddenblick 5a,
D-17493 Greifswald, Insel Riems, Germany
Department of Veterinary Pathology, Western College of Veterinary Medicine, University of Saskatchewan,
Saskatoon, Canada
Washington Animal Disease Diagnostic Laboratory, Washington State University, Pullman, USA
Department of Pathobiology, Ontario Veterinary College, University of Guelph, Ontario, Canada
(Received 12 August 2002; accepted 1 December 2002)

Abstract We examined 12 formalin-fixed paraffin-embedded feline skin tumours which had the histopathological features of fibropapillomas for the presence of papillomavirus (PV) DNA using touchdown polymerase chain
recation (PCR), DNA sequencing and nonradioactive in situ hybridization. Nine of the tumours contained a 102bp PCR product demonstrated using consensus PV primers that amplify a portion of the L1 gene. The nucleotide
sequences are closely related, but not identical to that of ovine PV type 2, rabbit oral PV and reindeer PV. The
deduced amino acid sequences had strong homologies with the major capsid protein L1 of deer PV, bovine papillomavirus (BPV) 1 and BPV 2, and European elk PV. Although PV antigens were not detected in any of the
tumours by immunohistochemistry, PV DNA was demonstrated in individual mesenchymal cells or cell nests of
4/12 tumours by in situ hybridization. A nonproductive infection of mesenchymal fibroblast-like tumour cells with
a papillomavirus would explain the lack of PV antigen expression and the absence of PV DNA in the hyperplastic
epithelium. Because these tumours and their pathogenesis are similar to equine sarcoids, we suggest that they
should be reclassified as feline sarcoids instead of fibropapillomas.
Keywords: cat, DNA sequencing, feline sarcoid, fibropapilloma, in situ hybridization, papillomavirus, PCR,
skin, tumour.

INTRODUCTION
In contrast to equine sarcoids, which represent the most
common skin tumour in this species, histopathologically
similar tumours of the feline skin are diagnosed
and reported only rarely in dermatopathology and
have been classified as feline sarcoids or fibropapillomas.
Statistically, their incidence ranks far behind feline
fibrosarcomas, mast cell tumours, basal cell tumours
and squamous cell carcinomas, and they seem to be
pathogenetically different from feline papillomas, which
are caused by host-specific, feline papillomaviruses.13
In contrast to feline papillomas, papillomaviral
antigens were not detected in feline sarcoids by immunohistochemistry.4 Based on their resemblance to
equine sarcoids, it was nonetheless assumed that papillomaviruses might play a role in the development of
these tumours.2 Furthermore, papillomavirus DNA was
recently shown to be present in feline sarcoids.5 In this
study we describe the macroscopic and histological
features of feline sarcoids and their molecular

Correspondence: PD Dr Jens Peter Teifke. Tel.: +49 38351 7225;

Fax: +49 38351 7226; E-mail: jens.teifke@rie.bfav.de
2003 European Society of Veterinary Dermatology

characteristics as determined by polymerase chain reaction (PCR), DNA sequencing and nonradioactive in situ
hybridization. We present evidence that these tumours
may be the outcome of a nonproductive infection of a
nonpermissive host with a novel, as yet uncharacterized papillomavirus and therefore we propose the classification as feline sarcoids instead of fibropapillomas
as this more accurately reflects their pathogenesis.

MATERIALS AND METHODS

Tumours
Formalin-fixed, paraffin-embedded samples of 12 feline
sarcoids were retrieved from the archives of Yager-Best
Histovet (cases 3, 510), Histovet (cases 2, 4), the
Department of Pathobiology, University of Guelph,
Canada (case 1), and the Washington Animal Disease
Diagnostic Laboratory (WADDL), Washington
State University, Pullman, USA (cases 11, 12; Table 1).
Controls for the immunohistochemistry, PCR,
DNA sequencing and in situ hybridization included tissues from normal feline skin, two bovine
papillomas, five equine sarcoids selected from the
files of the Department for Veterinary Pathology,
47

2Y
6Y
5Mo
12Y
F
F (S)
M
F (S)
DSH
DSH
DSH
DSH
32141
57101
008405
0110713
9
10
11
12

Justus-Liebig-University, Giessen, Germany,6,7 and five

feline dermal schwannomas (peripheral nerve sheath
tumours) from Yager-Best Histovet. Fifty feline vaccinesite-associated sarcomas from the files of the Western
College of Veterinary Medicine, University of Saskatchewan, Saskatoon, Canada were included as additional
negative controls.8

DSH = Domestic Short Hair; DMH = Domestic Medium Hair; M = male; F = female; C = castrated; S = spayed; Y = year(s); Mo = month; NA = data not available.

No recurrence 12 months
No recurrence 18 months later
NA
NA
Surgical excision
Surgical excision
Surgical excision
Surgical excision
0.5 cm diameter
NA
NA
0.5 0.5 0.3 cm

Farm cat
Farm cat
NA
NA

NA
No recurence 15 months later
NA
Surgical excision
Surgical excision
Surgical excision
NA
Town cat
Stray cat
1 cm diameter
3 3 cm
2 1 1.5 cm
M (C)
M (C)
M (C)
NA
DMH
DSH

1Y
10Y
2Y

M (C)

D00-151
D00-152
76556
30973
93996
5a
5b
6
7
8

DSH

8Mo

Base of right ear

(med. crus of helix)
Bridge of nose
Right fore footpads
Left forelimb-1st Digit,
medial P3 & digital pad
Upper Lip
Face
Right hock
Right temple

3 cm diameter

Stray cat

Surgery 2
Regression
Surgical excision
F (S)
0555-93
4

DSH

8Y

Right nostril

NA

Barn cat

Lost to follow up
Successful amputation 2nd cat involved,
not biopsied
Regrew twice successful therapy with
resection
No recurrence 12 months later
Excision
Tail amputation
M
M (C)
DSH
DSH

2Y
1Y

Right nose margin of nares

Tip of tail

5 cm diameter
3 cm diameter

Barn cat
Barn cat

Follow-up

Regrew twice then radiation successful

Housing
Size

1 cm diameter
Upper left lip

Site of lesion
Age

M (C)

4Y

Sex

DSH

2984-91
3528-91
2707-91
2979-93-1
1a
1b
2
3

Breed
Tumour ID
Cat/case

Table 1. Case numbers, animals, history and follow-up

Barn cat

Treatment

J. P. Teifke et al.

Radiation

48

Tissue processing for histopathology and electron

microscopy
The tumours were fixed in 10% buffered neutral formalin for at least 24 h. The formalin-fixed tissues were
routinely embedded in paraffin, cut at 4 m, mounted
on organosilane-coated slides and stained with haematoxylin and eosin (H&E) and Toluidine blue. For electron microscopy, fresh tumour tissue was fixed in 2.5%
glutaraldehyde in 0.16 cacodylic buffer (pH 7.2) and
embedded in Epon-Araldite. Ultrathin sections were
stained with lead citrate and uranyl acetate.
Immunohistochemistry
The immunohistochemistry for the detection of PV
antigen was performed as described.9 In brief, dewaxed
and rehydrated sections were incubated with a commercially available rabbit polyclonal antibody against
PV group specific antigens (Dako Diagnostika, Germany), at a dilution of 1:3000 in Tris-buffered-saline
(TBS, 0.1 Tris-base, 0.9% NaCl, pH 7.6). As linkerantibody for the avidinbiotin complex (ABC)
method, a biotinylated goat antirabbit IgG1 (Vector,
USA) was used. The following monoclonal antibodies
(mAbs) were applied to analyse expression of intermediary filaments and cellular antigens: antivimentin
(clone V9, Dako) was diluted 1:40 in TBS; antihuman
smooth muscle actin (clone 1A4, Dako) was used at
1:50 in TBS; antihuman desmin (clone D33, Dako)
was diluted at 1:50 in TBS; and antihuman neuronspecific enolase (NSE, clone BBS/NC/VI-H14, Dako)
was used at 1:100 in TBS. For the detection of S-100
protein, a rabbit anticow serum (Dako) was used at a
dilution of 1:300. To detect cytokeratins, a ready-touse-mixture containing mAbs against cytokeratins 18,
10, 13, 1416, 19 (Linaris, Germany) was applied.
Improvement of staining quality was achieved using
antigen retrieval by microwave irradiation (700 W,
2 5 min) in 10 mmol L1 citrate buffer, pH 6.0 carried
out according to the manufacturers recommendations with the exception of sections stained for cytokeratin, which required incubation with 0.1% pronase for
30 min (Linaris) and for NSE-staining, which was performed without any antigen demasking.
Polymerase chain reaction
A 244-bp DNA fragment located in the conserved,
transformation-associated E5 open reading frame of
bovine papillomavirus (BPV) 1 was amplified with
the primer set E5+ and E5.10 This primer pair also
amplifies a 248-bp long product of BPV 2. To determine, if infection with canine oral PV (COPV) was
present in the feline tumours, oligonucleotide primers

2003 European Society of Veterinary Dermatology, Veterinary Dermatology, 14, 47 56

Papillomavirus-DNA in feline sarcoids

Table 2. Oligonucleotide primers used in PCR to detect papillomaviral DNA in feline sarcoids
Primer

Nucleotide sequence (5 3)

E5+

CAAAGGCAAGACTTTCTGAAACAT

E5
L1+
L1
E6+
E6
IFNR2
IDNT2
FS-F
FS-R

AGACCTGTACAGGAGCACTCAA
CTTGTTTGGGGCTTAAGAGG
TGCAGTGTGTACCTGTCCTG
GGCACTGTTATCAATGGAGC
CACATAGTTCTTTGTCCGCC
TTYAAYMGICCITAYTGG
ISTICCICKWGTRTTRTC
CTGGCTATATAGAGCT
CAGTAGGGGATAATAC

Genome position
37593782 (BPV 1)
37603783 (BPV 2)
3981 4002 (BPV 1)
3985 4006 (BPV 2)
71317150 (COPV)
73727391 (COPV)
113132 (COPV)
443 462 (COPV)
6712 6534 (BPV 1)
6601 6618 (BPV 1)
NA
NA

I = Deoxyinosine, K = G or T, M = Aor C, R = A or G, S = C or G,
Y = C or T, W = A or T, NA = not applicable

corresponding to the L1 and E6 genes of COPV were

used.9 Based on the amino acid sequence homologies
between different animal papillomaviruses, nucleotide
sequences in the most conserved regions of the L1
genes of eight animal papillomaviruses were aligned to
design consensus primers. As described elsewhere, the
consensus primers IFNR2 and IDNT2 were predicted
to amplify a 102-bp fragment of the L1 gene of BPV 1,
BPV 2, ovine PV 1 and ovine PV 2, European elk PV,
deer PV, canine oral papillomavirus (COPV) and the
papillomavirus of Mastomys coucha (Table 2).8 For
PCR analysis, two 5-m paraffin sections were placed
in individual reaction tubes, the sections were dewaxed
with xylene, rinsed with 100% ethanol, and the DNA
was extracted using a commercially available tissue kit
(QIAMP DNA mini kit, Qiagen, Germany). PCR utilized 2 L of the extracted genomic DNA as template.
BPV- and COPV-specific PCR was carried out for 35
cycles using 0.625 U Taq DNA polymerase. (Promega
Corp., USA) and the PTC-200 PCR-thermocycler (MJ
Research, USA).7 In brief, 25 L of the reaction mixture contained 10 m TrisHCl (pH 9.0), 50 m KCl,
0.1% Triton X-100, 200 of dATP, dGTP, dCTP
and dTTP, 0.25 of both the forward and reverse
primer, 3.5 m MgCl2, and autoclaved water. After
initial heating (3 min at 95 C), each cycle consisted
of a denaturation step (1 min at 95 C) followed by
primer annealing (30 s at 63 C for BPV; 52.5 C for
COPV) and primer extension (1 min at 72 C). The
final extension step was set at 10 min at 72 C. Cloned
genomic DNA of BPV 1 (pBPV-1; kindly provided
by P.M. Howley) and the plasmid pCP62V8, which
contains the full-length COPV-DNA (kindly provided

Table 3. Consensus nucleotide sequence of

the papillomaviral L1 gene detected in feline
sarcoids

49

by H. Shirasawa), were used as positive controls.

Sections of formalin-fixed, paraffin-embedded, nonlesional
feline skin served as negative controls and a sample
without DNA was included in each run to check for
contamination. The consensus primer set was used
in a touchdown PCR as recently described.8 Basically, the reaction mixture contained 50 m KCl and
10 m TrisHCl (pH 9.0), 1.5 m MgCl2, 200 of
each dNTP, 1 of each primer, 1.25 U Taq DNA
polymerase (Promega), 2 L of the template DNA, and
autoclaved water to a final volume of 50 L. The
amplification conditions followed a protocol for touchdown PCR, starting with 94 C for 5 min, followed by
16 cycles of 94 C for 1 min, an annealing temperature
that decreased by 1 C each cycle from 50 to 35 C, and
the extension at 72 C for 1 min. This first amplification
round was followed by 24 cycles of 94 C for 1 min,
50 C for 2 min, and 72 C for 1 min with a final extension step set at 72 C for 6 min. The amplification products were visualized in ethidium bromide stained 2%
agarose gels.
Molecular cloning and sequencing of PCR products
PCR resulting in an amplicon of the expected size were
subjected to cloning and sequencing. At least 2 ng of
each amplicon was ligated into pGEM-T Easy vector
according to the manufacturers instructions (Promega).
Six clones per sample were selected for sequencing and
0.4 g of the purified plasmid DNA was subjected to
automated sequencing with the LI-COR system using
fluorescent-labelled SP6 and T7 primers according to
the recommended protocol (MWG-Biotech, Germany). The nucleotide sequences were analysed using
the tools at http://www.expasy.ch/
tools and compared with nucleotide and protein
sequences with the National Library of Medicine
BLAST sequence similarity search program at http://
www.ncbi.nlm.nih.gov.
Probe synthesis for in situ hybridization by PCR
Based on the sequence data obtained from the nine
feline sarcoids after touchdown PCR, an additional
internal set of primers, FS-F and FS-R was designed
for the generation of a probe for nonradioactive in situ
hybridization. The sequence of the oligonucleotides
and their location within the consensus sequence of the
102 bp amplicon are depicted in Tables 2 and 3. Starting with 100 ng of cloned DNA, the expected 78 bp
product was amplified from all nine feline tumours
after 30 cycles (initially 3 min at 95 C, 1min at 95 C,
30 s at 46 C, 30 s at 72 C and finally 5 min at 72 C).

1
50
TTTAATCGGC CGTACTGGCT ATATAGAGCT CAGGGTATGA ATAATGGTAT
AGCTTGGAAC AATACTTTAT TTGTAACAGT AGGGGATAAT ACTAGCGGCA
CC
Shaded boxes indicate the location of consensus primers IFNR2 (green) and IDNT2 (red) and
open boxes correspond to the location to the internal primers used for the generation of the ISH
probe.
2003 European Society of Veterinary Dermatology, Veterinary Dermatology, 14, 4756

50

J. P. Teifke et al.

Digoxigenin-11-dUTP (DIG DNA Labeling Mix, Roche

Diagnostics, Germany) was included in the reaction
mixture using the cloned DNA from case 1 for probe
labelling.11
Nonradioactive in situ hybridization
Nonradioactive in situ hybridization was performed
according to a standardized protocol,9 with the modification of the proteolytic digestion of tissue sections;
50 g mL1 proteinase K (Roche) in 50 m TrisHCl
(pH 7.4) for 15 min at 37 C were applied. A 1:50 dilution of the PCR generated probe was directly applied
to the sections. The tissue sections were briefly counterstained with Mayers haematoxylin. Feline normal skin
and vaccine-site-associated sarcomas were used as negative tissue controls. A double-stranded DNA-probe
specific for chelonid herpesviral DNA served as negative control for the probe.12 As positive control for the
nonradioactive in situ hybridization procedure, equine
sarcoid tissue was hybridized with a probe specific for
BPV 1.6

RESULTS

anisokaryosis, were oval to elongate with finely stippled chromatin and single, medium-sized nucleoli.
Mitotic figures were infrequent and averaged no more
than one per 400 field. Most tumours contained small
to moderate numbers of scattered round granulated
cells (Fig. 4). On Toluidine blue staining the granules

Figure 2. Photomicrograph of a typical feline sarcoid showing the

resemblance to an equine sarcoid. The exophytic nodule is composed
of fibroblastic cells. The irregular epidermal hyperplasia is a
characteristic finding. H&E, Bar, 200 m (cat 1).

Gross findings
The case identification, patient information, and the
follow-up data are summarized in Table 1. The tumours
were located on the head, especially on the upper
lips, the philtrum or nares, on the footpads, or at the
tip of the tail. Clinical evaluation of these tumours
revealed well circumscribed, 15 cm diameter, occasionally ulcerated dermal nodules (Fig. 1).
Histopathological findings
The lesions were exophytic, moderately well demarcated but nonencapsulated, dermal masses composed
of spindle-shaped to stellate cells, growing in poorly
defined fascicles and interlacing bundles (Fig. 2). The
tumour cells were separated by small amounts of collagenous, poorly vascularized stroma. The neoplastic
cells were of medium size and had indistinct cellular
borders (Fig. 3). The nuclei, which showed moderate

Figure 1. Typical clinical appearance of feline sarcoid: a small

nodule close to the nasal philtrum (cat 1).

Figure 3. Medium-sized fibroblast-like tumour cells are orientated

in fascicles and interwoven bundles and are separated by small
amounts of collagenous stroma. H&E, Bar, 50 m (cat 6).

Figure 4. Nuclei of the neoplastic cells are moderately anisokaryotic,

oval to spindle shaped with finely stippled chromatin. Scattered
throughout the tumour are mast cells. H&E, Bar, 20 m (cat 6).

2003 European Society of Veterinary Dermatology, Veterinary Dermatology, 14, 47 56

Papillomavirus-DNA in feline sarcoids

51

sarcoids (cases 1, 4, 9, 10). Cytokeratins, smooth

muscle actin and desmin were not detected in the
mesenchymal tumour cells (Fig. 6).

Figure 5. Higher magnification of Fig. 2. Note the close proximity

of the tumour cells to the long thin rete ridges of epidermis H&E,
Bar, 65 m (cat 1).

were metachromatic, indicating that these were mast

cells The overlying epithelium was hyperplastic forming long, branching rete ridges (Fig. 2). The tumour
cells grew to the epidermis, sometimes showing a charactistic perpendicular palisading (Fig. 5). The tumour
typically obliterated most adnexal structures of the
dermis but at the interface with normal tissue, the
tumour cells exhibited an infiltrative growth.
Immunohistochemistry
Group specific PV antigen was not detected in any of
the feline sarcoids using immunohistochemistry. The
cytoplasm of all spindle-shaped tumour cells stained
strongly for vimentin. In all feline sarcoids except
case 5 (a and b), many tumour cells had light to moderate cytoplasmic staining for S-100 protein. Strong
immunostaining for -NSE was present in four feline

Table 4. Results of the assays for the

detection of PV antigen and DNA in feline
sarcoids

Electron microscopy
Ultrastructurally, two cell populations were present.
The predominant cell population was mesenchymal
(Fig. 7). The cells were elongated in profile. The cytoplasm was filled with abundant rough endoplasmic
reticulum with moderately distended cisternae. No
other distinguishing cytoplasmic features were noted.
The second population comprised mast cells containing electron dense, cytoplasmic granules. The cells were
lying amidst bundles of collagen fibrils showing the
typical periodic banding pattern.
Molecular cloning and sequencing of PCR-amplified
DNA
In 9 of 12 feline sarcoids (including one recurrent mass),
the touchdown PCR yielded the expected 102 bp
product (Fig. 8). Viral DNA was not amplified from any
of the feline sarcoids using BPV 1, BPV 2 or COPVspecific primer sets in a PCR. The same primer sets
amplified the expected fragments from cloned BPV
and COPV-DNA or from DNA extracted from bovine
fibropapillomas and equine sarcoids (Table 4). All
PCR products were cloned and six clones each were
sequenced. The nucleotide sequences of all feline sarcoids showed homology with short stretches of the
ovine PV 2 (82% identity, 77/93 nucleotides), BPV 2
(82% identity, 51/62 nucleotides), rabbit oral PV (91%
identity, 32/35 nucleotide), reindeer PV (95% identity,
23/24 nucleotides) and BPV 1 (95% identity, 21/22
nucleotides). Nucleotide sequences of cases 1, 2, 4, 6, 8,
11 and 12 are depicted in Table 3. Minor differences

Case

PV-IHC

BPV-PCR

COPV-PCR

TD-PCR

ISH

1a
1b
2
3
4
5a
5b
6
7
8
9
10
11
12
Bovine papillomas
Equine sarcoids
Feline schwannomas
VSS
Normal skin

+
+

+
+
+
+
+

+
+
+
+
+

+
+

PV-IHC: detection of papillomavirus group-specific antigen by immunohistochemistry;

BPV-PCR: PCR specific for bovine papillomavirus 1 and 2 DNA; COPV-PCR: PCR specific
for canine oral papillomavirus DNA; TD-PCR: touchdown PCR to detect papillomaviral
consensus DNA; ISH: detection of papillomavirus DNA by in situ hybridization using the
internal papillomavius L1-gene DNA-specific probe. Tissues listed in the last four lines served
as controls; VSS: vaccine site-associated sarcomas.
2003 European Society of Veterinary Dermatology, Veterinary Dermatology, 14, 4756

52

J. P. Teifke et al.

Figure 6. Immunohistochemical staining pattern in feline sarcoids: (a) anti-pan-cytokeratin mAb-mix stains suprabasal cells of the feline
epidermis; (b) anti-vimentin Mab V9 stains the cytoplasm of nearly all fibroblast-like tumour cells; (c) no immunoreactivity in the tumour cells
with antihuman desmin (mAb D33) and with (d) antihuman smooth muscle actin (mAb 1A4); (e) faint to strong cytoplasmic immunostaining
with antihuman neuron-specific enolase (NSE, mAb BBS/ NC/ VI-H14); (f ) cytoplasmic immunostaining for S-100 protein (polyclonal rabbit
anti cow S-100). ABC-method, AEC substrate, Mayers haematoxylin counterstain. Bar, 100 m.

were present in the nucleotide sequences obtained from

case 3 (deletion of the CTA at positions 1921) and
case 10 (insertion of a CTA triplet at position 20). A
comparison of the deduced amino acid sequences
obtained from the feline sarcoids resulted in 85% identity and 92% similarity to the major capsid protein L1
of BPV 1 and BPV 2 (amino acids 311338), 84% identity and 91% similarity to deer PV (amino acids 312
337), and 81% identity and 84% similarity to European
elk PV (amino acids 313339). Furthermore, significant homology (70% identities) was present to the
L1 protein of human papillomavirus (HPV) 1A, 23, 22,
49, 63, the cottontail rabbit PV, BPV 4 and COPV.

Papillomavirus-specific PCR products were not obtained

from normal feline skin, feline schwannomas, and the
vaccine site-associated sarcomas.
Nonradioactive in situ hybridization
Using the cloned papillomaviral DNA of case 1 as a
template, a 78-bp internal probe was synthesized and
used in the in situ hybridization experiments. Specific
nucleus-associated hybridization signals were detected
in individual fibroblast-like cells (Fig. 9) or cell nests
within the mesenchymal tumour cells (Fig. 10) of 4 of
12 sarcoids. All of the tumours with neoplastic cells
that stained in the in situ hybridization were also

2003 European Society of Veterinary Dermatology, Veterinary Dermatology, 14, 47 56

Papillomavirus-DNA in feline sarcoids

Figure 7. Electron micrograph showing two tumour cells amidst

collagen fibrils. Note the distended cisternae of rough endoplasmic
reticulum, typical of fibroblastic cells. 4200 (cat 1).

Figure 8. Agarose gel electrophoresis. Touchdown PCR using the

IFNR2 and IDNT2 primer set yielding a 102-bp amplicon starting
with DNA extracted from formalin-fixed, paraffin-embedded tissue
of a bovine fibropapilloma (lane 1); cat 2 (lane 2); cat 3 (lane 3); cat
5a and 5b (lanes 4 and 5); cat 8 (lane 6); cat 10 (lane 7); cat 7 (lane
8); cat 9 (lane 9); cat 6 (lane 10); feline schwannoma (lane 11).

positive by PCR (Table 4). The hybridization signal was

either confined to the cell nucleus or covered the entire
cell. No signals were detected in the hyperplastic epithelium overlying feline sarcoids.

DISCUSSION
Skin tumours in young cats, bearing a very close histological resemblance to equine sarcoids, were first identified in the late 1980s in Canada.2 Many of the affected
animals were farm cats, possibly explaining why similar lesions were not recognized in diagnostic laboratories serving predominantly urban areas (EJ Walder,
personal communication). Similar cases have since
been recognized in the UK, USA, Europe, Australia
and New Zealand, also primarily in cats from rural
environments.5,13,14 The frequent location on the face
and footpads suggests association with trauma, such as
cat bites or scratches. Like their equine counterparts,
feline sarcoids do not metastasize, but may recur after
surgical excision, especially when the location of the
tumour (nares, philtrum) makes complete removal
difficult.

53

Figure 9. Feline sarcoid; cat 6. Scattered throughout the dermal

proliferation are singular cells, strongly stained after in situ
hybridization using the papillomavirus L1-specific probe. In situ
hybridization; Mayers haematoxylin counterstain, Bar, 50 m.

Figure 10. Feline sarcoid; cat 8. Nest of dermal fibroblast-like

tumour cells with nucleus- or cell-associated hybridization signals
for papillomaviral L1 DNA. In situ hybridization; Mayers
haematoxylin counterstain, Bar, 50 m.

Histopathologically, the feline sarcoid tumour bears

a very close resemblance to the equine sarcoid, which
has a very distinct, indeed unique microscopic appearance. The main distinguishing feature between the
equine and feline sarcoids is the lack of mast cells in the
equine tumour. This also distinguishes feline sarcoids
from the other feline lesions or tumours in the list of
differential diagnoses, namely granulation tissue, scars,
well-differentiated fibrosarcomas and schwannomas.
The architecture of granulation tissue and scars is a
further differentiating feature. It can be surprisingly
difficult to differentiate some schwannomas from those
feline sarcoids that contain a low population of mast
cells. Given the immunohistochemistry results, in which
there is some support for a nerve sheath origin for the
feline sarcoid cells, it might be postulated that there is
a continuum between these entities.
The positive immunostaining of mesenchymal tumour
cells at least in some of the feline sarcoids for both
S-100 and NSE was an unexpected finding. It has
been assumed, based on the histological and
ultrastructural features of the tumour cells, that they

2003 European Society of Veterinary Dermatology, Veterinary Dermatology, 14, 4756

54

J. P. Teifke et al.

were of fibroblastic origin. The NSE staining may be

spurious. This antibody is often poorly specific and
produces a high background staining. In a recent audit
of stains commonly used in immunohistochemistry
panels it was the only one that was deemed of questionable value.15 If the NSE staining of feline sarcoid cells
is genuine, then hypothetical reasons for the expression
of intermediary filaments not normally detected in
mature fibrocytes include: (i) the tumour originates
from incompletely differentiated stem or precursor
cells, or (ii) the expression of these antigens could be
driven by viral oncogenes (E5, E6 or E7) or growth factors (and therefore be an effect of the papillomaviral
infection). Thus the touchdown PCR as described in
this work seems to be the method of choice for differentiation between feline sarcoids and other differentials especially schwannomas and fibrosarcomas.
The sporadic occurrence of feline sarcoids may contradict, at first glance, an infectious pathogenesis. One
explanation is that farm cats may be less likely to be
presented to a veterinarian for removal of a cutaneous
mass than pet cats. Nonetheless, the striking histopathological similarity to equine sarcoids, the current understanding of equine sarcoids as the outcome
of a BPV infection in a nonpermissive host, and the
higher prevalence of feline sarcoids among cats in rural
areas lead us to the hypothesis that feline sarcoids may
also be the result of a nonproductive infection with
BPV.16 This hypothesis is further supported by the
recently reported association of feline fibropapillomas
with papillomavirus DNA.5
The published consensus primer set IFNR2 and
IDNT2 allows the detection of DNA from papillomaviruses related to supergroup C (ungulate fibropapillomaviruses: BPV 1, BPV 2, BPV 5, ovine PV, European
elk PV, deer PV), supergroup D (BPV 3, BPV 4,
BPV 6) and supergroup E (COPV, cottontail rabit PV,
HPV 1, HPV 41, HPV 63) in formalin-fixed, paraffinembedded tissues.17 The cloned amplicons from the
feline sarcoids in our study showed significant homology, but not complete identity with papillomaviral
DNA sequences available in GenBank. This confirms
the observation of others that a 176-bp fragment of the
E1 gene had only 75 and 64% homology with BPV 1
and BPV 2, respectively.5 The deduced amino acid
sequences of the feline sarcoids examined in our study
had a significant similarity to those of the major capsid
protein L1 of BPV 1 or -2, the deer PV, European elk
PV and of several HPVs. Based on a maximal identity
of 92% with any of the available papillomavirus
sequences at the amino acid level, we conclude that the
feline sarcoids contain DNA of an uncharacterized
papillomavirus. The negative PCR results in 3 of 12
tumours with the histological appearance of feline
sarcoids could be due to degradation of DNA during
prolonged formalin fixation that may result in
papillomaviral DNA copy numbers below the sensitivity threshold of the employed PCR method. Alternatively, it is possible that DNA from a papillomavirus
which is not fully host cell-adapted is lost during

proliferation of the host cells according to the hitand-run-hypothesis.18 Furthermore, it cannot be completely excluded that some of the tumours in this study
finally were not sarcoids but schwannomas or even
fibrosarcomas.
The in situ hybridization demonstrated papillomaviral DNA in nuclei of the mesenchymal tumour cells in
four of the PCR-positive feline sarcoids. The signal was
present in individual cells scattered throughout the
tumours or in small groups of tumour cells. Hybridization signals were not seen in the overlying hyperplastic
epithelium in any of the feline sarcoids or feline
schwannomas examined. In comparison with equine
sarcoids, which were hybridized with a BPV 1 E5 genespecific probe, the feline sarcoids exhibited a much
smaller number of stained cells. Multiple factors could
contribute to this phenomenon. The probe that was
used on the feline sarcoids is only 78 bp long compared
with the 244-bp, BPV-specific probe, which may result
in a lower overall signal intensity. It is also possible that
nonproductive replication of the papillomavirus results
in lower mRNA copy numbers of the late genes. While
the probe in the feline sarcoids hybridization experiments was directed at a late gene, the BPV-specific
probe hybridizes with an early gene, and because of this
could yield higher signal intensity due to hybridization
to transcripts of early genes. Alternatively, loss of viral
DNA during tumour progression, as mentioned previously, also has to be considered. Using highly specific
and sensitive PCR assays, BPV and COPV DNA was
not amplified from any of the feline sarcoids. These
observations together with findings that the deduced
amino acid sequences obtained from the feline sarcoids
have only a 92% similarity and 85% identity with those
of BPV 1 and BVP 2, and no significant homology to
COPV lead us to conclude that a role for bovine papillomaviruses in the pathogenesis of feline sarcoids is
highly unlikely and we rejected this hypothesis.
Using a cross-reactive BPV 1-induced antibody, papillomaviral late (structural) antigens were detected by
immunohistochemistry, neither in the mesenchymal
tumour cells nor in the overlying hyperplastic epithelium in any of the feline sarcoids. This finding was consistent with earlier observations of Gross & Affolter.4 It
was therefore concluded that productive papillomaviral replication with synthesis of complete virions carrying cross-reactive antigens does not occur in feline
sarcoids.
In contrast to bovine fibropapillomas and in accord
with equine sarcoids, the feline sarcoids did not express
structural PV antigen in tumour cells and did not contain PV DNA in the overlying hyperplastic epithelium.
This strongly argues for a pathogenesis similar to that
in equine sarcoids. Therefore, feline sarcoids could be
the result of a nonproductive infection in an alien,
nonpermissive host, by a PV, related to but not
identical with bovine, ovine, or rabbit papillomaviruses.
We believe that, wherever possible, the nomenclature
should not only reflect the histological appearance, but
also the pathogenesis of an entity and therefore

2003 European Society of Veterinary Dermatology, Veterinary Dermatology, 14, 47 56

Papillomavirus-DNA in feline sarcoids

we recommend that the tumours considered in this
study be classified as feline sarcoids, instead of feline
fibropapillomas.

ACKNOWLEDGEMENTS
We gratefully acknowledge the excellent technical support
of Gabriele Czerwinski, the financial support of Pettrust,
Ontario, Canada, and the clients of Yager-Best Histovet for their cooperation in providing information
and further material from patients with feline sarcoids.
REFERENCES
1. Goldschmidt, M.H., Shofer, F.S. Skin tumors of the dog
and cat. Oxford: Pergamon Press, 1992: 29 36.
2. Yager, J.A., Wilcock, B.P. Epithelial tumours. In: Color
Atlas and Text of Surgical Pathology of the Dog and Cat.
London: Mosby-Year Book Europe, 1994: 2923.
3. Sundberg, J.P., Van Ranst, M., Montali, R. et al. Feline
papillomas and papillomaviruses. Veterinary Pathology
2000; 37: 110.
4. Gross, T.L., Affolter, V.K. Advances in skin oncology. In:
Kwochka, K. W., Willemse, T., Von Tscharner, C., eds.
Advances in Veterinary Dermatology, Vol. 3. Oxford:
Butterworth-Heinemann, 1998: 383.
5. Schulman, F.Y., Krafft, A.E., Janczewski, T. Feline cutaneous fibropapillomas: clinicopathologic findings and
association with papillomavirus infection. Veterinary
Pathology 2001; 38: 291 6.
6. Teifke, J.P., Hardt, M., Weiss, E. Detection of bovine
papillomavirus DNA in formalin-fixed and paraffinembedded equine sarcoids by polymerase chain reaction
and non-radioactive in situ hybridisation. European
Journal of Veterinary Pathology 1994; 1: 5 10.
7. Teifke, J.P. Histologische, immunhistologische und molekularbiologische Untersuchungen zu tiologie und Proliferationsverhalten epithelialer Tumoren der Haut und kutanen
Schleimhute der Haustiere. Berlin: Lehmanns 1998.

55

8. Kidney, B.A., Haines, D.M., Ellis, J.A. et al. Evaluation

of formalin-fixed paraffin-embedded tissues from vaccine site-associated sarcomas of cats for papillomavirus
DNA and antigen. American Journal of Veterinary
Research 2001; 62: 8339.
9. Teifke, J.P., Lhr, C.V., Shirasawa, H. Detection of canine
oral papillomavirus-DNA in canine oral squamous cell
carcinomas and p53 overexpressing skin papillomas of
the dog using the polymerase chain reaction and nonradioactive in situ hybridisation. Veterinary Microbiology
1998; 60: 11930.
10. Teifke, J.P., Weiss, E. Nachweis boviner PapillomvirusDNA in Sarkoiden des Pferdes mittels der PolymeraseKettenreaktion (PCR). Berliner und Mnchener Tierrztliche
Wochenschrift 1991; 104: 1857.
11. Emanuel, J.R. Simple and efficient system for synthesis
of non-radioactive nucleic acid hybridisation probes
using PCR. Nucleic Acids Research 1991; 19: 2790.
12. Teifke, J.P., Lhr, C.V., Marschang, R.E. et al. Detection
of chelonid herpesvirus DNA by nonradioactive in situ
hybridisation in tissues from tortoises suffering from
stomatitisrhinitis complex in Europe and North America.
Veterinary Pathology 2000; 37: 37785.
13. Gumbrell, R.C., Rest, J.R., Bredelius, K. et al. Dermal
fibropapillomas in cats. Veterinary Record 1998; 142:
376.
14. Schulman, F.Y., Krafft, A.E., Janczewski, T. Feline
cutaneous fibropapillomas (sarcoid): clinicopathologic
findings and association with papillomavirus infection
[Abstract]. Veterinary Pathology 2000; 37: 532.
15. Maxwell, P., McCluggage, W.G. Audit and internal quality control in immunohistochemistry. Journal of Clinical
Pathology 2000; 53: 92932.
16. Pfister, H. Biology and biochemistry of papillomaviruses. Reviews of Physiology, Biochemistry and Pharmacology 1984; 99: 11181.
17. Lowy, D.R., Howley, P.M., Papillomaviruses. In: Fields
Virology. Philadelphia: Lippincott, Williams & Wilkins,
2001: 2232.
18. Smith, K.T., Coggins, L.W., Doherty, I. et al. BPV-4
induces amplification and activation of silent BPV-1 in
a sub-line of C127 cells. Oncogene 1993; 8: 1516.

Rsum Nous avons examin 12 blocs parrafins de tumeurs cutanes du chat ressemblant des fibropapillomes
afin de rechercher par PCR, squenage de lADN et hybridisation non radioactive in situ la presence de papillomavirus (PV). Neuf tumeurs prsentaient un produit de PCR 102 bp, mis en evidence par un primre PV qui
amplifie une portion du gne L1. Les squences nuclotidiques taient proches, mais non identiques au PV ovin
de type 2, au PV oral du lapin, et au PV du renne. Les squences dacides amins avaient des homologies marques
avec la protine de la capside L1 du cerf, du papillomavirus bovin de type 1 et 2. Alors que les techniques immunohistochimiques nont pas permis de mettre en vidence dantignes de PV, de lADN de PV a t retrouv dans
les cellules msenchymateuses ou les cordons cellulaires de 4/12 tumeurs par hybridisation in situ. Une infection
non productive des cellules tumorales msenchymateuses fibroblastiques par un papillomavirus pourrait expliquer labsence dexpression dantignes de PV et labsence dADN de PV dans lpithlium hyperplasique.
Comme ces tumeurs sont semblables aux sarcodes quins, nous suggrons quelles soient reclasses comme des
sarcodes flins plutt que comme des fibropapillomes.
Resumen Examinamos 12 tumores cutneos felinos, fijados en formalina e incluidos en parafina, con caractersticas histopatolgicas de fibropapiloma, para determinar la presencia de DNA de papilomavirus (PV) utilizando
una touchdown PCR, secuenciacin del DNA e hibridacin in situ no radioactiva. Nueve de los tumores contenan un producto de la PCR de 102 pb, obtenido con cebadores consensuales (consensos primers) de PV que
amplifican una porcin del gen L1. Las secuencias de nucletidos estn ntimamente relacionadas, pero no son
idnticas al PV ovino tipo 2, PV oral del conejo o al PV del carib. Las secuencias de aminocidos deducidas
2003 European Society of Veterinary Dermatology, Veterinary Dermatology, 14, 4756

56

J. P. Teifke et al.
tenan fuertes homologas con la protena de la capside mayor L1 del PV del ciervo, PV bovino (PVB) 1 y el PVB
2, y PV del ciervo europeo. Mientras que los antgenos del PV no fueron detectados immunoqumicamente en
ninguno de los tumores, ADN PV fue demostrado, mediante hibridacin in situ, en clulas mesenquimatosas individuales o en grupos de clulas en 4/12 tumores. Una infeccin no productiva con un papilomavirus de las clulas
fibroblsticas mesenquimatosas explicara la falta de expresin de antgenos PV y la ausencia de DNA PV en el
epitelio hiperplsico. Debido a que estos tumores y su patognesis son similares a los sarcoides equinos, sugerimos
que deberan clasificarse como sarcoides felinos en lugar de fibropapilomas.
Zusammenfassung Wir untersuchten 12 Formalin-fixierte, in Paraffin eingebettete Hauttumoren der Katze, die
histopathologische Charakteristika von Fibropapillomen aufwiesen, auf die Prsenz von Papillomavirus (PV)
DNS mittels PCR, DNS-Sequenzierung und nicht-radioaktiver in situ Hybridisierung. Neun der Tumoren
enthielten ein mittels consensus PV Starter nachgewiesenes 102 bp PCR Produkt, dass einen Teil des L1 Gens
amplifiziert. Die Nukleotidsequenzen sind mit denen von ovinem PV Typ 2, oralem Kaninchen PV, und Rentier
PV verwandt. Die deduzierten Aminosurensequenzen zeigten deutliche Homologie mit dem major capsid protein L1 vom PV des Rehs, vom bovinen Papillomavirus (BPV) 1 und BPV 2 und vom PV des europischen Elchs.
Whrend mittels Immunohistochemie in keinem der Tumoren PV Antigene entdeckt wurden, wurde PV DNS
in individuellen Mesenchymalzellen oder Zellnestern in 4 der 12 Tumoren mittels in situ Hybridisierung nachgewiesen. Eine nicht produktive Infektion von mesenchymalen Fibroblasten-hnlichen Tumoren mit Papillomavirus knnte das Fehlen der PV Antigen-Exprimierung und die Abwesenheit der PV DNS im hyperplastischen
Epithel erklren. Da diese Tumoren und ihre Pathogenese dem equinen Sarkoid hnlich sind, schlagen wir vor,
dass sie an Stelle von Fibropapillomen als feline Sarkoide reklassifiziert werden.

2003 European Society of Veterinary Dermatology, Veterinary Dermatology, 14, 47 56