Human papillomavirus in vulvar and vaginal carcinoma cell lines.

A number of reports associate human papillomavirus (HPV) with cervical cancer and cancer cell lines derived from this tumour type. Considerably fewer reports have focused on the role of HPV in carcinomas from other sites of female anogenital squamous epithelia. In this study we have tested for the presence of HPV in eight low-passage vulvar carcinoma cell lines and one extensively passaged cell line, A431. One cell line from a primary vaginal carcinoma was included. The presence of the HPV was evaluated by the polymerase chain reaction (PCR), by Southern blot analysis and by two-dimensional gel electrophoresis. General primer-mediated PCR was applied by using primers from the L1 region, E1 region and HPV 16 E7 region. Southern blot hybridisation was performed under low-stringency conditions (Tm = -35 degrees C) using a whole genomic HPV 6/16/18 probe mixture and under high stringency conditions (Tm = -18 degrees C) with the whole genomic probes of HPV 16 and 33. HPV 16 E6-E7 mRNA was assessed by ribonuclease protection assay (RPA). HPV was found in only one vulvar carcinoma cell line, UM-SCV-6. The identified type, HPV 16, was integrated in the cell genome and could be amplified with all primers used. Also E6-E7 transcripts were found in these cells. Five original tumour biopsies were available from the HPV-negative cell lines for in situ hybridisation. All these were HPV negative with both the HPV 6/16/18 screening probe mixture under low stringency and the HPV 16 probe under high stringency. The results indicate that vulvar carcinoma cell lines contain HPV less frequently than cervical carcinoma cell lines and suggest that a significant proportion of vulvar carcinomas may evolve by an HPV-independent mechanism.

Numerous studies have shown that most cervical carcinomas and cancer-derived cell lines contain human papillomavirus (HPV) genomes, usually integrated in the host cell genome. While HPV has been implicated in the aetiology of this particular type of cancer, data on the role of HPV at other genital sites are more fragmentary. There is some epidemiological evidence that HPV infection is a risk factor for vulvar carcinoma (Sherman et al., 1991). HPV DNA has also been found in carcinoma samples of the vulva and vagina (Buscema et al., 1988;Ikenberg et al., 1990;Anderssen et al., 1991;Bloss et al., 1991;Toki et al., 1991). HPV is known to infect the squamous epithelium of the vagina and vulva, and in doing so HPV may induce classical exophytic condylomata, flat lesions or low-grade intraepithelial neoplasia. Follow-up studies have shown that such lesions may progress to higher grades (Planner and Hobbs, 1988) in the same manner as in the cervix (Syrjinen et al., 1988), although this occurs infrequently. Moreover, koilocytic atypia has been found within and in the vicinity of vulvar and vaginal intraepithelial neoplasia (VIN, VAIN) and in a subset of invasive neoplasia (Zaino et al., 1987), which indicates some biological similarity with the pathogenesis of cervical neoplasia. Many of the in vitro data evaluating the presence and physical state of HPV in squamous carcinoma of the female genital tract have been obtained using cancer cell lines derived from the uterine cervix. There are no previous studies on the presence of HPV DNA in cell lines derived from extracervical genital squamous carcinomas. The major reason for this has been the lack of suitable cell lines. We have recently established and characterised cell lines from human vulvar and vaginal carcinomas (Grenman et al., 1990;Worsham et al., 1991). These cell lines, as well as the longestablished vulvar carcinoma cell line A43 1, have now been studied for the presence of HPV.

Materials and methods
Cell lines The establishment and characterisation of cell lines UM-SCV-1, 2, 3, 4, 5 and 6 have been described previously (Grenman et al., 1990;Worsham et at., 1991). Vulvar carcinoma cell line UM-SCV-7 and the vaginal carcinoma cell line UM-SCVA-1 were more recently established in our laboratory (S Grenman et al., unpublished). The cell lines are summarised in Table I Southern blot hybridisation Southern blotting followed the standard procedure. Briefly, 10Lg of restriction enzyme-digested or undigested cellular DNA was loaded into individual lanes and run into 1.0% agarose gels and transferred by Southern blotting to nylon filters (Gene Screen, Dupont, Boston, MA, USA) for subsequent hybridisation. Restriction endonucleases PstI and BanI, known to cut the HPV 16 genome, were used to digest the genomic DNA. Hybridisation was performed with vector-HPV in Mm and vagn cardino cal is S Hienen et a  Gaithersburg, MD, USA). For screening, the samples were hybridised overnight under low-stringency conditions (T,, = -35'C), using a whole genomic HPV 6/16/18 probe mixture. After screening, the filters were rehybridised with HPV 16 and HPV 33 probes under high stringency (T, = -18C). The filters were exposed at -70'C for 1, 3 and 8 days. Before rehybridisation, the probe was removed from the filter by boiling with 0.1% SDS-1 mM EDTA for 2 m, followed by rapid cooling at 20°C. Removal of the hybridised probes was confirmed by exposing the filter for 24 h.
The absence of bacterial DNA in the cell lines was confirmed by hybridisation of the filters with 16S RNA gene probe. Bacterial ribosomal genes are highly conserved and universally presented among bacterial species, particularly the 16S rRNA (Fox et al., 1980). the 16S rRNA gene probe was prepared by PCR using Escherichia coli DNA as a template to amplify a 1.3 kb product. The probe was labelled with 3P using a multiprime method. The primers used were T-ITGAGCTCAGATTGAACGCTG and ATTGGATCCA- CGATTACTAGCG (Kauppinen et al., 1994). E. coli DNA in decreasing concentrations (from 500 ng to 2 pg per lane), digested with HindlII, was used as a positive control.

Integration analysis
The physical state of the HPV genome was studied with two-dimensional gel electrophoresis. A 5 fig sample of undigested genomic DNA was electrophoresed in a 0.4% agarose gel at 20 V for 17 h. The sample was incubated at 56-C before electrophoresis. The resulting lane was cut from the gel and recast in a 0.8% gel. This gel was electrophoresed at 90g to the orginal direction, using 70 V for 4 h. DNA was transferred to a filter and hybridised with 32P-labelled HPV 16 probe.

Polymerase chain reaction (PCR)
The presence of HPV DNA was studied using PCR with four primer sets targeting the El, LI and HPV 16 E7 regions. The primers are shown in Table II. The reaction took place on 300ng of genomic DNA in a 5001 l reaction volume. The PCR solution contained 5 1il of 10 x PCR buffer (50 mM potassium chloride, 10 mM Tris-HCI, pH 8.8, 1.5 mM magnesium chloride, 0.1% Triton X-100), 0.75 units of DynaZyme DNA polymerase (Finnzymes, Espoo, Finland), 200 gM of each deoxynucleotide triphosphate, 20 pmol of the primers and sterile water. The template DNA was first denatured for 4.5 min at 95°C and then exposed to 35 cycles of denaturation at 95°C for 30 s, annealing at 55°C for 50 s and extension at 72°C for 60 s in a thermal cycler (Cetus, Norwalk, CT, USA). The amplification was completed by a 4 min extension step at 72C. We performed an additional PCR with GP5 and GP6 primers in more relaxed conditions: the template DNA was first denatured for 4.5 min at 95°C and then exposed to 40 cycles of denaturation at 94°C for 60 s, annealing at 40C for 2 min and extension at 72°C for 90 s. HPV plasmid DNAs and DNA extracted from CaSki cells were used as positive controls. No DNA was added to the PCR solution of the negative controls. Both the undigested and digested PCR products were electrophoresed in 3% agarose gel (NuSieve, FMC BioProducts, Rockland, ME, USA). Bands were visualised by ethidium bromide staining. The PCR products with GP5-GP6 primers were transferred to a filter and hybridised using a mixture of whole genomic biotinylated HPV 6/16/18 probes under lowstringency conditions (Tm = -35C). The hybrids were visualised with streptavidin-alkaline phosphatase complex using nitroblue tetrazolium as chromogen and 5-bromo4chloro-3-indolyl phosphate as substrate (Syrjinen and Syrjdnen, 1986). P-Globin was amplified to ensure that the samples were appropriate for PCR. The sensitivity of the PCR was tested with MY09-11 primers and CaSki cells. We mixed 10000, 1000, 100, 50, 10, 5 and I CaSki cells with 300 ng of DNA extracted from normal human fibroblasts and performed the PCR in the more stringent conditions as described above. The PCR was still positive with one CaSki cell (data not shown). probe served as a negative control. The hybridisation was performed overnight with 32P-labelled probe (specific activity 2.5 x 105 c.p.m). After RNAse digestion and ethanol precipitation the samples were analysed in 5% polyacrylamide gel in 8 M urea. The gel was exposed overnight at -70°C.

In situ hkbridisation of the original tumour samples
Original tumour samples of cell lines UM-SCV-IB, -3, 4, -5 and UM-SCVA-1 were available for HPV in situ hybridisation. The hybridisation was performed with the HPV 6/16/18 mixture and HPV 16 probe under low and high stringency respectively (Syrj&nen and Syrjanen, 1986).

Results
Southern blot analysis and two-dimensional gel electrophoresis Only UM-SCV-6 hybridised clearly after 1 day's exposure to the HPV 6/16/18 probe mixture under low stringency. Five other vulvar cell lines (UM-SCV-1A, -1B, -4, -5 and A431) and the vaInal cell line (UM-SCVA-1) showed some bands after 8 days' exposure (Figures 1 and 2). The restriction pattern of UM-SCV-6 was similar to that found with DNA extracted from CaSki cells, indicating the presence of HPV type 16. This was confirmed by rehybridisation with HPV 16 probe under high stringency (Figure 3). The other vulvar cell lines and the vaginal cell line with positive bands under low-stringency conditions also showed some faint bands after hybridisation with HPV 16 probe under high stringency conditions after long exposure (Figures 3 and 4). However, the band sizes were less than expected for HPV and the restriction pattern did not fit any of the known HPV types. Also, the original tumour biopsies from these cell lines were HPV DNA negative by in situ hybridisation (Table I). None of the cell lines hybridised with the HPV 33 probe under highstringency conditions (data not shown). The copy number of the HPV genome in UM-SCV-6 cells was some 200-300 as judged from the comparison of hybridisation signals to the CaSki cells (500-600 HPV 16 copies). HPV 16 DNA in UM-SCV-6 cell line was integrated, since undigested samples showed high molecular weight signals (Figure 3). The integration was confirmed by two-dimensional gel electrophoresis. Intense hybnrdisation signals were detected with slowly migrating high molecular weight DNA, which are compatible with integrated DNA sequences. No circular sequences were detectable, indicating that UM-SCV-6 contained no episomal sequences ( Figure 5). The absence of bacterial sequences in DNAs extracted from the cell lines was confirmed by hybridising the filters with bacterial 16S rRNA gene probe. E. coli DNA on the positive control filter yielded detectable signals down to the concentration of 200 pg. All cell lines were negative. Polymerase chain reaction P-Globin was amplified in all cell lines. PCR with all three primer sets from LI and El ORF amplified only DNA from UM-SCV-6 cells. All other cell lines were negative. As the Southern blot showed positive hybridisation signals with HPV 16 under high stringency not only with UM-SCV-6, but also with UM-SCV-1A, -lB, -4, -5, A431 and UM-SCVA-1, these cell lines were reanalysed with PCR using HPV 16 E7 primers. Only UM-SCV-6 yielded positive amplification. In order to detect other HPVs which might be less homologous with the GP5 and GP6 primers, we performed an additional PCR under more relaxed conditions. UM-SCV-6 and CaSki were positive, but all others remained negative even after hybridisation of the PCR products with a probe mixture of HPV 6/16/18 under low stringency (data not shown).
HPV 16 E7-E6 mRNA analysis HPV 16 E7-E6 transcripts were found only in UM-SCV-6 and CaSki cells as determined by ribonuclease protection assay. All other cell lines were negative. Hybridisation with the sense probe was negative with both CaSki and UM-SCV-5 cells (Figure 6).

S Henen et al
Most cervical cancer biopsies and cell lines derived from cervical cancer contain integrated HPV and have shown transcripts from E6 and E7 ORFs of the persisting HPV DNA (Schwartz et al., 1985;Yee et al., 1985). The role of HPV in the carcinogenesis of other sites of female genital squamous epithelium has been less clearly defined. In this study we analysed the presence of HPV in one  HPV. However, the sizes of the bands, the negative PCR data with several pnmer sets even under permissive annealing conditionss and the negative in situ hybridisation result of the original tumours of UM-SCV-IB, -3, -4, -5 and UM-SCVA-l argue against the presence of HPV. The possibility of bacterial DNA was also excluded. It is possible that some of the cellular sequences might cross-hybridise with HPV probes and become visible in autoradiography, but only after long exposure of the filter.
According to the data derived from clinical studies, there are two different types of vulvar cancer, one associated with HPV and the other not (Anderssen et al., 1991;Bloss et al., 1991;Toki et al., 1991). The present study supports this view.
Chronic vulvar irritations including hyperplastic dystrophy and lichen sclerosis et atrophicus, which are not associated with sexually transmitted diseases, have been strongly linked to invasive vulvar carcinoma (Pincus and Stadecker, 1987;Zaino, 1987). Moreover, it has been shown that the association of HPV and VIN decreases with age (Park et al., 1991), whereas the incidence of vulvar carcinoma increases as a function of age, unlike cervical cancer, which plateaus between the fifth and eighth decades (Finnish Cancer Registry database). Based on these observations and the results of the present study, it appears that cervical and vulvar carcinomas are not aetiologically identical and that factors other than HPV have a more important role in vulvar carcinogenesis. We recently analysed these vulvar and vaginal cell lines for the state of p53 gene and found that UM-SCV-6 and UM-SCV-1 contain wild-type p53, whereas all other cell lines contain mutated p53 (Hietanen et al., in press). This indicates that p53 mutations in vulvar carcinoma cell lines are frequent and are detected in HPV-negative cell lines. Furthermore, it is possible that p53 gene mutations are more important in vulvar carcinogenesis than HPV infection.
In summary, the present results show that HPV is only infrequently required in the establishment of vulvar SCC cell lines and suggest that other factors may be more essential to the abnormal growth of vulvar carcinoma ceUs.
HPV in vwr and -as cmin n cd lins S Hietanen et al