Materials and Methods

Tissue specimens—Tissues were obtained from the collections of the Prairie Diagnostics Services, Western College of Veterinary Medicine, University of Saskatchewan, Saskatoon, SK, Canada; the Institut de Pathologie et de génétique, Loverval, Belgium; Rest associates, London; and, the Pathology Institute, Vetsuisse Faculty, University of Berne, Berne, Switzerland. Fifty-four formalin-fixed paraffin-embedded skin biopsy specimens were examined. Specimens originated from BISC (n = 21), invasive SCC (22), and miscellaneous skin conditions other than skin cancers (eg, allergic dermatitis; 11) that were used as negative controls. Specimens from white cats or from locations typical for AK such as ear tips and eyelids were excluded from this study. Thirty-micrometer-thick sections were cut from each tissue block, with a new disposable microtome blade for each block, before DNA extraction. Two canine warts, which had histopathologic characteristics of typical PV-induced inclusion bodies, and 1 bovine fibropapilloma served as positive controls.

DNA extraction—The protocol of Albini et al²⁹ was used for DNA extraction. Briefly, each section was deparaffinized twice with 1.2 mL of xylene at room temperature (approx 20°C) for 10 minutes, washed with 100% ethanol, and then dried at 37°C for 30 minutes. Desiccated samples were suspended in a tissue lysis buffer (50mM Tris-HCl [pH 8.5], 1mM ethilenediaminetetraacetic acid, and 2.8% sodium dodecylsulfate) and proteinase K (20 mg/mL) and incubated at 56°C on a rocking platform overnight. After lysis, samples were transferred to a column^a and centrifuged to reduce viscosity. The DNA was precipitated with absolute ethanol and extracted with a commercial DNA kit.^b

Primers—Two sets of primers were used for the PCR assay. Because the sequences encoding E1 are highly conserved, the E1 regions of FePV (GenBank accession No. AF480454) and CaPV (GenBank accession No. NC001619) were aligned to design a set of consensus primers (ie, PapF, 5[.minute]-ATGGCGGGMTARAAAAGGTA-3[.minute] and PapR, 5[.minute]-AACAGCTGYTTTTTARCYTTTTT-3[.minute]) for narrow-range PCR assay, which is able to generate amplification products of approximately 341 bp of FePV, CaPV, or closely related PVs.

The second set of primers (ie, CP4, CP5, and PPF1 primers), also derived from E1, was selected for broad-range PCR assay with the objective of amplifying as many PVs as possible. With this set of primers, up to 64 HuPVs are identifiable.³⁰ The expected size of the PCR product was approximately 450 bp.

PCR assay and agarose gel electrophoresis— Polymerase chain reaction conditions for PapF and PapR were performed. Volumes of 30 mL were used. Each reaction contained 1 μL of genomic DNA, 200μM of each deoxynucleoside triphosphate, 0.3μM of each of the sense and anti-sense primers, and 2.5 units of a DNA polymerase.^c After an initial denaturation step at 95°C for 4 minutes, PCR assay was performed for 30 cycles at 95°C for 1 minute, 50°C for 1 minute, and 74°C for 1 minute, with a final elongation step at 74°C for 5 minutes. Deoxyribonucleic acid extracted from canine warts served as positive control, whereas DNA- and RNA-free water was used as negative control.

The PCR mix with CP4, CP5, and PPF1 primers was identical to the mix for narrow-range PCR assay, except that 0.45μM of the CP4 and CP5 primers and 0.3μM of the PPF1 primer were used. The PCR assay consisted of a denaturation step at 95°C for 10 minutes, followed by 40 cycles at 95°C for 1 minute, 47°C for 1 minute, and 74°C for 1 minute, with a final elongation step at 75°C for 5 minutes. An extract from 1 bovine fibropapilloma served as an additional positive control.

Polymerase chain reaction products were segregated by agarose gel electrophoresis, and bands were viewed under UV light after ethidium bromide staining. Bands on the gel were excised, and DNA was extracted with a gel extraction kit.^d Amplified DNA was sequenced by use of fluorescent sequencing and terminator chemistry.^e

Sequence analysis—Samples were considered positive for PV DNA if they met the following requirements: they had a band of the expected size after gel electrophoresis and the sequenced DNA had homology with E1 of previously sequenced PVs. Homologous DNA sequences were searched for by use of the National Center for Biotechnology Information GenBank via a BLAST search.^f Sequence alignments and phylogenetic trees were made from the clustal algorithm obtained by use of a software program.^g

Results

Macro- and microscopic analysis—A careful macroscopic selection and microscopic confirmation of affected specimens was a major prerequisite prior to the virologic analysis. Specimens representing invasive SCC had been resected from sun-exposed or white areas of 21 domestic shorthaired cats and 1 Persian cat (12 males and 10 females). Twenty-two specimens from ear tips (n = 11), nose (3), eyelids (3), digits (3), and lips (2) met criteria required for invasive SCC. These criteria included the macroscopic presence of scaly-to-crusty and erosive-to-plaque-like or ulcerative lesions (Figure 1). The growth process was always endophytic. Histologically, cords or islets of infiltrative cells were detected in all specimens. Furthermore, anisocytosis; anisokaryosis; large, hyperchromatic nuclei; prominent nucleoli; increased mitotic index; and abnormal mitoses were encountered in all specimens with variable intensity and in variable proportion. In addition, keratin pearls and intercellular bridges were present in 16 and 12 specimens, respectively.

View Full Size

Macroscopic and microscopic lesions of SCCs in cats. A—Photograph of invasive SCC in a cat with ulceration of the eyelid. B—Photomicrograph of a section of the invasive SCC from panel A at low magnification. Notice invasive proliferation of atypical keratinocytes with pearl formation (white arrow) and the cornified layer (red arrow). The basal membrane is not discernible. The dermis (long arrow) is invaded by cords of atypical keratinocytes (short black arrow, pointing towards such an invasive site). H&E stain; bar = 200 μm. C—Photomicrograph of a section of the invasive SCC from panel A at high magnification. Notice islets of keratinocytes with features of malignancy, such as anisokaryosis, anisocytosis, multinucleated cells (short arrow), and abnormal mitosis (long arrow). H&E stain; bar = 50 μm. D—Photograph of BISC in a cat with circular, crusted, erosive, and hyperpigmented plaques (arrow). E—Photomicrograph of a section of the BISC from panel D at low magnification. The basal membrane (white arrow) is intact, and the dermis is not invaded. Irregular acanthosis (long black arrow) is obvious. Notice that hair follicles are affected (short arrow). H&E stain; bar = 200 μm. F—Photomicrograph of a section of the BISC from panel D at high magnification. Notice acanthosis, hyperpigmentation (brown cells), clumped keratohyalin granules (white arrow), loss of polarity, and anisokaryosis (branched arrow). H&E stain; bar = 50 μm.

Citation: American Journal of Veterinary Research 67, 12; 10.2460/ajvr.67.12.2036

• Download Figure
• Download figure as PowerPoint slide

A second group of 21 tumor specimens met the criteria for BISC. These specimens were obtained from the face (n = 16), neck (12), and limbs (3) or were scattered (2). Thirteen domestic shorthair cats, 3 domestic longhair cats, 2 Siamese, 1 Persian, 1 Himalayan, and 1 Cornish Rex were affected. Macroscopically, the lesions were squamous crustosus and grossly circular. Two lesions had a single center, but 19 were multicentric (Figure 1). Microscopically, the following criteria were met for BISC: moderate-to-severe parakeratotic hyperkeratosis, acanthosis with papillomatous hyperplasia (n = 1) or irregular hyperplasias (20), loss of polarity, and scattered dyskeratotic keratinocytes atypia in all layers of the epidermis and usually also in the infundibulum and reaching the isthmus. Furthermore, the following types of atypia were recorded: enlarged nuclei, anisokaryosis, monster cells (bizarre multinucleated giant cells), and abnormal mitotic figures. Hyperpigmentation was found in all but 3 specimens. Fifteen of the 21 specimens had clumped keratohyalin granules, which were considered as suggestive for PV infection. However, other signs such as koilocytosis and nuclear inclusion bodies were not detected.

With the exception of the ulcerated lesions, the dermis of all samples was considered normal and not heavily inflamed (Figure 1). Thus, a total of 22 samples representing invasive SCC and 21 samples representing BISC were available for virologic analysis by PCR assay and sequencing.

PCR assays—The narrow-range PCR assay amplified PV DNA extracted from canine warts but not DNA extracted from bovine fibropapilloma (Figure 2). In contrast, the broad-range PCR assay amplified PV DNA from canine warts and bovine fibropapilloma. It was concluded that both PCR assays were able to specifically amplify selected PV DNAs.

View Full Size

Establishment of broad-range and narrow-range PCR assays for detection of carnivore PVs. Polymerase chain reaction products were loaded on agarose gels and stained with ethidium bromide. A—Amplification of cloned DNA by either broad-range (450 bp) or narrow-range PCR assay (341 bp). Lane 1 = Water in place of DNA added to the reaction. Lane 2 = DNA from a commercially available phagemid.^h Lane 3 = DNA from oral CaPV cloned into the phagemid.^h Lane 4 = DNA from CPV3 (GenBank accession No. DQ295066) cloned into the phagemid.^h M1 = Molecular weight marker (100-bp ladder). B—The DNA was extracted from tissues before being amplified by either the narrow range or the broad-range PCR assays. M2 = 1-kilobase ladder. M1 = 100-bp ladder. Lane 1 = Negative control with no DNA added to the reaction. Lane 2 = Extract from canine wart tissue, which had typical PV-induced inclusion bodies on histologic examination. Lane 3 = Extract from a tumor specimen of a cat with SCC (GenBank accession No. DQ085784). Lane 4 = Extract from SSC sample No. 39.

Citation: American Journal of Veterinary Research 67, 12; 10.2460/ajvr.67.12.2036

• Download Figure
• Download figure as PowerPoint slide

The narrow-range PCR assay was applied to samples from invasive SCC and BISC specimens; 1 BISC sample (BISC sample No. 15; Appendix; Figure 3) had positive results for PV DNA, whereas the others had negative results. Next, the broad-range PCR assay was applied to the same samples. Interestingly, 5 of 21 BISC samples (BISC sample Nos. 2, 5, 6, 10, and 15) as well as 4 of 22 SCC samples (SCC sample Nos. 15, 24, 28, and 29) had positive results for PV DNA (Figure 2). One of the samples that had positive results for PV DNA on the broad-range PCR assay (BISC sample No. 15) also had positive results for PV DNA on the narrow-range PCR assay. These results suggested that the broad-range PCR assay was indeed able to uncover PVs that were different from the known FePV and CaPVs.

View Full Size

Phylogenetic relationships of the newly detected PV sequences with known PV E1sequences (ie, CaPV, FePV, HuPV, and BoPV are represented by GenBank accession Nos. D55633, AF377865, AY330623, and AJ620206, respectively). IS = BISC. SCC = Invasive SCC. Units indicate the number of substitution events (percentage of nucleotides).

Citation: American Journal of Veterinary Research 67, 12; 10.2460/ajvr.67.12.2036

• Download Figure
• Download figure as PowerPoint slide

Sequence analysis—The nucleotide sequence of the amplified DNA was determined and the resulting sequences were compared to assess whether novel PV-like sequences had been detected. Indeed, use of the basic local alignment search tool^e revealed relatedness to PV E1 sequences for all 9 samples that were positive for PV DNA. Relation to HuPV, FePV, CaPV, rat PV, and BoPV was evident. Clustal alignments^g revealed the various degrees of relationship of the newly determined sequences among each other as well as in comparison with known PVs. Overall, CaPVs and FePVs were most closely related to the newly detected viral sequences, with a relative amino acid identity of 56% to 71%. Among the HuPVs, types 4, 55, 63, 65, 71, and 74 were aligned most frequently but type 71 most often had the closest relationship to the new sequences with 58% to 61% amino acid identity. Among the BoPVs, type 5 was the closest relative, having 55% to 62% amino acid identity. A phylogenic tree drawn from the aligned sequences divided the new sequences into 4 clusters (Figure 3). In cluster 1, BISC sample No. 15 was situated most closely with CaPV and FePV. Cluster 2 was occupied by BISC sample No. 10 and was between FePV and HuPV type 71. Cluster 3 was represented by BISC sample No. 2 and found close to HuPV type 71. The remaining sequences (SCC sample Nos. 15, 24, 28, and 29 and BISC sample Nos. 5 and 6) represented a fourth cluster, which was clearly distinct from BoPV type 5 on the most distant side and HuPV type 71, FePV, and CaPV on the less distant side. These results suggested the presence of thus far unidentified PVs in tissues representing invasive SCC and BISC. Interestingly, sequences obtained from specimens of 3 cats with invasive SCC (SCC sample Nos. 15, 24, and 28) had identical sequences. Furthermore, it was observed that not a single sample from invasive SCC specimens had been associated with the more classic FePV and CaPV. However, because the classification of PVs is based on the sequence of L1, the exact taxonomic position of these novel PV-like sequences could not be assigned.

Discussion

The purpose of our study was to detect and partially characterize PV DNA in samples representing in situ and invasive types of SCC in cats to learn more about PV variants in cats and about possible associations of these viruses with individual forms of SCC in cats. Two types of PCR assays, a narrow range and a broad-range, were applied to extend the range of targeted PVs as far as possible.

Careful macroscopic selection and microscopic confirmation resulted in the identification of 22 samples representing invasive SCC and 21 samples representing BISC, which were available for virologic analysis by PCR assay. Papillomavirus DNA was detected in 4 of 22 samples representing invasive SCC and in 5 of 21 samples representing BISC, whereas all nontumor control samples had negative results for PV DNA. Only 1 (BISC sample No. 15) of the 9 viral DNAs had been revealed by the narrow-range PCR assay. However, the same narrow-range PCR assay amplified PV DNA extracted from canine warts, which was expected because the primers had been chosen for their homology with conserved sequences within E1 of FePV and CaPV. Yet, the restricted range of these primers was confirmed, as they proved unable to amplify DNA from the more distantly related bovine fibropapilloma virus. These results indicate that the remaining samples with positive results for PV DNA did not harbor conventional FePV or CaPV.

Eight samples, which had negative results for PV DNA on narrow-range PCR assay, had positive results for PV DNA on broad-range PCR assay. The broad-range PCR assay made use of a second set of primers that were also derived from E1 but known to uncover a large variety of HuPVs.³⁰ In our study, this second set of primers also amplified DNA from bovine fibropapilloma virus as well as viral DNA from canine warts. Sequencing of the amplification products obtained from the 8 samples revealed novel PV-related DNAs, although relations to HuPV, FePV, CaPV, rat PV, and BoPV were evident. A phylogenic tree drawn from the aligned sequences divided the new sequences into 4 clusters. Three of those clusters had close relationship to CaPV, FePV, and HuPV. Interestingly, the fourth cluster, represented by 6 amplification products, was clearly distinct from BoPV type 5 and from HuPV type 71, FePV, and CaPV. Judging from the limited sequence information available, it appeared as if this fourth cluster represented a novel group of FePVs that had not been detected previously and that may be associated with SCC in cats. Notably, all PVs detected in association with invasive SCC were found to belong to this novel cluster.

This represents, to our knowledge, the first evidence of thus far unknown PV-like sequences associated with SCC in cats. Interestingly, some of the novel sequences were found in association with invasive SCC. Notably, previous attempts to detect conventional PV antigens in such lesions had failed,²³ which led to the hypothesis that invasive carcinomas of cats are probably not virally induced, whereas instances of Bowen's disease in cats are probably PV-induced.⁷ Our findings clearly challenge the former opinion, although a causative correlation between the disease and the novel PV strains has not yet been shown. Results of another study²⁴ did reveal PV antigens in 44% of BISCs. Although the proportion of BISC samples with positive results for PV DNA in our study is lower (5 of 21 BISC samples), it should be kept in mind that the broad-range PCR assay may not be able to reveal all variants of FePVs. Furthermore, it is well-known that PCR detection of viral nucleic acids in formalin-fixed and paraffin-embedded tissues may be decreased in comparison to fresh tissue.²⁹ Finally, the absence of PV DNA in SCC samples can also be explained by the so-called hit-and-run model, which postulates an initial transformation of the infected cell and a subsequent loss of PV DNA.³¹

Full proof of the existence of the novel PVs that are predicted through the results of our study still needs to be provided. However, we suggest that a great diversity of FePVs may exist that is in need of detection and characterization. The future use of the technique applied here will help in identifying more affected cats with papilloma-associated diseases. Virologic studies can be initiated with the aim to better characterize these novel viruses. Cloning and sequencing of the entire genomes of these viruses will allow phylogenetic comparisons with HuPVs as well as discrimination between benign and high-risk variants. Such studies can eventually provide insights into the molecular pathways underlying the pathogenesis of these viruses in cats.