Detection of Human Papillomavirus in Warty Carcinoma of the Uterine Cervix: Comparison of Immunohistochemistry,In Situ Hybridization and In Situ Polymerase Chain Reaction Methods

PATHOLOGY RESEARCH AND PRAGlCE © Gustav Fischer Verlag

Detection of Human Papillomaviru5 in Warty Carcinoma of the Uterine Cervix: Comparison of Immunohistochemistry, In Situ Hybridization and In Situ Polymerase Chain Reaction Methods Nam Hoon Cho, Hee Jae J00 1, Hee Jeong Ahn 2 , Woo Hee Jung and Kwang Gil Lee Department of Pathology, Yonsei University College of Medicine, Seoul, Korea, lAjou University College of Medicine, and 2Pochon University Pundang CHA Hospital

Summary Warty carcinoma of the uterine cervix is a very rare specific variant of invasive squamous cell carcinoma, usually described as a hybrid feature of condyloma with invasive squamous cell carcinoma. Besides having the koilocytes and maturation of squamous epithelium, there is unequivocal stromal invasion. The feathery surface seems to be characteristic of warty carcinoma, differentiating it from condylomata acuminata or verrucous carcinoma. Using in situ polymerase chain reaction (lS-PCR) and in situ hybridization (ISH) as well as an immunohistochemical technique (IHC), we determined the amplification and expression of several human papilloma virus (HPV) types (6, 11, 33, 16 and 18) in nine warty carcinomas of the uterine cervix. We found amplified HPV predominantly in the nuclei of the feathery surface in all cases only when IS-PCR was applied, while it was detected only in five (55.6%) or six cases (66.7%) by conventional IHC or ISH, respectively. We found multiple types in the same lesion in six cases (66.7%) by IS-PCR, in comparison with ISH which detected a lower incidence (22.2%). This finding of coinfection is more consistent with that seen in lowgrade squamous intraepithelial lesions (LSILs) of the uterine cervix than in high SIL. It is important to be aware of these distinct lesions in postmenopausal women with their characteristic feathery and thin surface, as well as a hybrid form of exophytic condyloma and stromal invasion, and frequent coinfection of multiple different HPVs as a type of LSILs. Pathol. Res. Pract. 194: 713-720 (1998)

Key words: Warty carcinoma - In situ polymerase chain reaction - Human papilloma virus

Introduction Warty carcinoma was originally described as a vulvar and anal lesion, recently being classified as a variant of invasive squamous cell carcinoma (SCC) [8]. However, the nomenclature of this entity is still divergent and consequently the concept and biologic behavior are still unresolved because of the lack of cumulative data. The exophytic lesion has features similar to condylomata acuminata. This tumor seems to be less aggressive than the classic SCC despite the presence of stromal invasion [9]. The role of human papilloma virus (HPV) appears to be important and should be elucidated in the pathogenesis of this lesion in order to determine its biologic behavior. The molecular tools for detecting HPV types have been continuously developed and each tool has its advantages and disadvantages. Most sensitive methods such as polymerase chain reaction (PCR) and blot analysis have the disadvantage of inaccessibility to cellular localization and limitations on prospective Address for correspondence: Nam Hoon Cho, M.D., Department of Pathology, Yongdong Severance Hospital, Yonsei University College of Medicine, 146-92. Dogok-dong, Kangnam-ku, Seoul, 135-270, Korea. Seoul Youngdaong P.O. Box 1217, Seoul, Korea. Tel.: 02-3497-3542, Fax: 02-3463-2103 0344-0338/98/0194-0713$5.00/0

714 . Nam Hoon Cho et al. study, whereas in situ hybridization (ISH) and immunohistochemical stain (IHS) are less sensitive and specific than PCR/blot analysis. Stage-of-the-art technology in the form of amplified in situ localization, namely in situ PCR (lS-PCR), has recently been made available which appears to overcome both disadvantages [15, 6, 4,24]. To our knowledge, this is the first report of in situ PCR as a means of studying the involvement of HPV types in warty carcinoma of the uterine cervix.

Material and Methods Selection ofCases

A total of nine cervical biopsy specimens were initially diagnosed as warty carcinoma showing features of a condylomatous surface coexisting with focal, minute subepithelial invasion. Radical hysterectomy was performed in three cases. Two patients had a previous history of a subtotal supracervical hysterectomy for unknown disease more than 10 years before. Both patients subsequently underwent a trachelectomy of radical excision including the cervix and upper vagina. Immunohistochemistry (IHC)

The presence of HPV antigens was examined using the immunoperoxidase method. Paraffin sections for immunohistochemical staining were mounted on slides coated with poly-Llysine and stained with polyclonal antibody against the late antigen (L1) of papilloma virus. Briefly, slides were deparaffinized and rehydrated. They were then immersed in 10 mM sodium citrate, pH 6.0, and heated in a microwave oven for 5 min (700 Watt) and then sequentially placed in running cold water and soaked in phosphate-buffered saline (pH 7.2) for 15 min. They were then heated in a microwave for another 5 min. Subsequently, circled markings were made by Dako pencil on all sections and blocked with 3% hydrogen peroxide for 10 min. Sections were sequentially incubated with normal goat serum for 30 min and then placed in a 1:80 dilution of L1 HPV primary antibody overnight at 4 0c. Sections were additionally incubated with linked streptavidin-biotin (LSAB) for 30 min at room temperature (Dako, Carpinteria, CA, U.S.A.). Slides were washed with PBS and incubated with performed ABC-complex for 30 min at room temperature. Diaminobenzidine (DAB) was used as a chromogen with hematoxylin counterstain. Condylomata acuminata associated with pregnancy were selected as a positive control of the Ll antigen. HPVTyping by in Situ Hybridization (ISH)

Formalin-fixed, paraffin-embedded sections which underwent punch biopsy were deparaffinized in xylene, rehydrated and digested with proteinase K solution (40 Ilg/mJ) at 37°C for 10 min, washed with PBS, dehydrated and air-dried. The probes for HPV 6+11,16+18 and 31+33+35 (DIGENE, USA) were labeled by bio-11-dUTP and incubated at 37°C overnight after a denaturation process of heating at 87 °C for 8 min. Slides were washed in PBS and Signasure wash buffer (pH 8.0, DIGENE, USA), then they were incubated with

streptavidine-alkaline phosphatase complex at 37°C for 30 min and developed with nitroblue tetrazolium (NBBT) and bromo-chloro-indoxyl phosphate (BCIP). Slides were washed with distilled water, dehydrated and counterstained with hematoxylin. The positive and negative control probes were applied in every staining series. Selection ofHPV Primers

The selected primers were obtained from the E6 region, as follows: HPV 6 (59.1 nmole): 5'-TgTCACAAACCgCTgTgTgA-3' HPV 11 (81.8 nmole): 5'-TgTCACAAgCCgTTgTgTgA-3' HPV 16 (67 nmole): 5'-gTTTgCAgCTCTgTgCATA-3' HPV 18 (86.1 nmole): 5'-gTgTTCAgTTCCgTgCACA-3' HPV 33 (81.3 nmole): 5'-gTCTCCAATgCTTggCACA-3' HPVPF (51.9 nmole): 5'-AAgggCgTAACCgAAATCggT-3' Each primer was diluted by TE Buffer and kept in stock solution in a refrigerator. HPV 6 and HPVPF were diluted by the ratio of 1:2, HPV 16 by 1:3, and the other primers by 1:4 TE buffer. In Situ Polymerase Chain Reaction (IS-PCR)

Formalin-fixed, paraffin-embedded sections which underwent punch biopsy were deparaffinized in xylene and graded alcohols, and a minimal circle was depicted around the most representative area while the slides were air-dried. The tissues were permeabilized with 20 Ilg/ml proteinase K in distilled water for 10 min at 37°C. After tapping off and subsequent rehydration, complete air drying was followed for 7 min at 82 0c. The slides were then immersed in polymerase chain reaction mixture containing lOx PCR buffer 2.5 Ill, dNTP mixture (each 2.5 mM) 4.0 Ill, each HPV primer about 20 pmole, HPVPF primer 26 pmole, digoxigenin-11-dUTP (125 nmole) 1.0 Ill, TaKaRa Taq polymerase (5 UnitlJ.l1) 0.8 Ill, and sterile water to a total of 25 III per slide. A thermal cycling oven (TaKaRa cycler, IPN) was employed using the following parameters: initial denaturation at 94°C for 3 min followed by 25 cycles of annealing at 55 °C for 2 min and denaturation at 94°C for I min. Evaporation of the PCR mixture was markedly reduced by covering the samples with sticker coverslips. After completion of amplification, the slides were washed in Tris buffer several times, followed by detection steps. Detection of in situ generated amplifiers was accomplished by the linking of anti-digoxigenin-alkaline phosphatase antibody (1: 100 dilution in antibody diluent) at 37°C for 30 min. After washing by Tris buffer several times, the chromogen solution containing 45 III nitro-blue-tetrazolium (NBT), 5-bromo-4-chloro-3-indoryl-phosphate (BCIP) was incubated at room temperature. While inspecting the chromogen staining procedure, the completion of the incubation steps was determined at the time of the best contrast. The slides were dipped in distilled water and mounted in glycerin without the counterstain step. Specificity controls for the amplification step included the negative control by omitting the Taq DNA polymerase or primers. A few cervical squamous cell carcinomas which had previously been documented in our laboratory by polymerase chain reaction were used as positive controls.

In situ PCR of HPV in Cervical Warty Carcinomas . 71S

Table I. Clinical presentation of warty carcinoma of the cervix Case No.

Age

Menopause

Size (cm)

Stage

Therapy

PHx

Folow-up (yrs)

1 2 3 4 5 6 7 8

49 53 51 60 40 56 77 53

47 51 No 50 No 40 52 39

0.8xO.3xO.3 lxlxO.5 1.5xlxO.7 3x2.5x1.2 4.5x3x1.8 3x3x? 3x2.5x? 6x5x2.2

TAl TA2 TA2 IBI IB2 ITA ITA ITA

9

48

38

4x3.5xO.7

TAl

Cone RAH Cone RAH RATwithLSO RTx RTx Trachelectomy with RTx & CTx Trachelectomy

NC NC NC NC NC NC NC SH (18 yrs ago) SH (10 yrs ago)

NED NED NED NED NED NED NED NED NED NED

(0.7) (2.3) (3) (4.3) (2.5) (2.7) (4) (2) (18.2) (10.2)

PHx: past history, RAH: radical abdominal hysterectomy, LSO: left salpingo-oophorectomy, RTx: radiotherapy, CTx: chemotherapy, NC: non-contributary, SH: subtotal hysterectomy, NED: no evidence of disease Each signal of stage is according to FlOO classification.

Results Clinical Profiles

The mean age of the nine patients with warty carcinoma of the uterine cervix was 54.1 years ranging from 40 to 77 years. All but one patient were postmenopausal and multipara. Treatment consisted of cervical conization in two cases and radical hysterectomy or trachelectomy in five and radiotherapy was given after biopsy in the other 2 cases. Interestingly, two patients had a previous history of subtotal hysterectomy for unknown disease approximately 18 and 10 years before, and then both were treated with radical trachelectomy. One of them had postoperative adjuvant therapy including radio- and chemotherapy. The tumor sizes were variable, ranging from less than 1 em to 6 cm. A summary of

the clinical data is shown in Table I. All cases have remained disease-free following resection of the tumor. Gross Findings

Three tumors (case 2, 4 and 5) from patients who underwent hysterectomy revealed exophytic warty and papillary features located in the transformation zone of the posterior half of the cervicallumbus (Fig. lA). All masses seemed to be sharply demarcated at their margins. Two cases which underwent radical trachelectomy (case 8 and 9) showed a large polypoid mass involving the whole circumference of the residual cervix and involvement of the vaginal fornix. On cut section, the mass was predominantly endophytic like an inverted papilloma (Fig. 2A). It had a characteristic gyriform appearance which extended into the deep stroma with bulbous pushing borders.

Fig. 1. The common gross finding of warty carcinoma of the cervix. A: A discrete papillary mass growing as a large sessile lesion is noted in the posterior lip (arrow) mimicking condyloma acumintum. B: A low power of warty carcinoma of the cervix. The well-defined polypod mass grows intraluminally and spreads into the superficial stroma (hematoxylin-eosin, xlO).

716 . Nam Hoon Cho et aI.

Fig. 3. The tongue-like projection directly radiates from the condylomatous area showing a fair maturation and infiltrates into the stroma (hematoxylin-eosin, orig. magnif. xlOO).

difference in the pattern of rete pegs, in which those of warty carcinoma showed a tentacular and infiltrative cord appearance without an anastomosing pattern, in contrast to the pattern of condyloma acuminata which shows frequent anastomosis between the rete pegs. Identification of Virus by Immunohistochemical Stain

Fig. 2. The gross finding of more advanced warty carcinoma of the cervix. A: The mass is large and involves the whole circumference of the lower cervix and some parts of the fornix and upper vagina. On cut section, the mass shows inward growth like an inverted papilloma with the surface being intact and slightly eroded. Notice the deep extension of tumor into the deep stroma. B: The mass is endophytic and revealed confluent irregular gyri (hematoxylin-eosin, xlO).

Late gene production of the viral capsid protein, LI, was identified in five cases (55.6%) with conventional immunohistochemical stain, but generally the staining intensity and distribution were very weak and limited. The L1 antigen was detected only in the nuclei of surface epithelium. Identification of Virus by ISH

Light Microscopical Findings

All biopsied tissues revealed a biphasic pattern composed of a condylomatous and papillary configuration at the surface and SCC at the bottom. The malignant foci seemed to be directly transformed from condyloma (Fig. 3). Four resected specimens revealed the typical gross morphology of this tumor. Common findings were polypoid growth into the cervical lumen and infiltrative dyskeratosis (Fig. lB). One of the high stage tumors was basically endophytic and irregularly gyriformed (Fig. 2B). The uppermost surface of the condylomatous fronds was somewhat different from that of condyloma acuminata in that the former is more feathery and thinner than the latter (Fig. 4, A and B). The uppermost surface of condyloma acuminata reveals stout and thickened hyperkeratotic layers covering the undulating acanthotic epithelium (Fig. 4C). In addition, there is another distinct

ISH results are summarized in Table 2. Tested typing included low-risk types 6 and 11, intermediate-risk types 31, 33 and 35, and high-risk types 16 and 18. The HPV signals were absolutely detected within the nuclei of the uppermost layer and occasionally within those of koilocytes in the intermediate layer, but not within the basal level (Fig. 5A). The staining was classified into a 3-tier system according to the staining intensity and distribution: 1, minimal degree; 2, moderate degree; 3, high degree of localiziation and staining. Six cases revealed positive signals at the minimal degree (66.7%), and three of them showed combined infection. Only two cases (22.2%) demonstrated an infection of highrisk HPV type. Identification of Viral Typing by In Situ peR

HPV typing was performed with IS-PCR in all cases. Tested typing included low-risk types 6 and 11, inter-

In situ PCR of HPV in Cervical Warty Carcinomas . 717

Fig. 4. Comparison of the surface features of condyloma acuminatum and warty carcinoma. A: Feathery surface in a case of warty carcinoma; B: A magnified view of the feathery surface of a warty carcinoma, C: The tightly-packed and thick hyperkeratotic surface in the case of condyloma acuminatum (hematoxylin-eosin, orig. magnif.: A: x40, B: x200, C: x40).

Fig. 5. Identification of HPV infection in warty carcinoma of the uterine cervix using in situ hybridization and IS-PCR. A: ISH signals were localized in the nuclei of the superficial and intermediate layers. B: IS-PCR for HPV is intensely expressed along the nuclei on the surface. C: The nuclei in the superficial and intermediate zone, suggestive of koilocytes, demonstrates a strong staining reaction (BCIPINBT, orig. magnif.: A: x200, B: x40, C: x200).

Table 3. Comparison between clinical parameters and HPV typing Low-risk type

intermediate-risk

high-risk type

combined infection

0.534 0.288 0.529

0.169 0.570 0.087

0.677 0.592 0.395

risk-type in 2 of 7 (28.6%) cases. Case 9 revealed intermediate-risk type by ISH despite absence of type 33 by IS-PCR. Comparison of Clinical Parameters and HPV Typing

age 0.147 size 0.497 stage 0.151

The values mean p value obtained from X2 test (spss 7.5).

mediate-risk type 33 and high-risk types 16 and 18. The HPV signals were detected predominantly within the nuclei of the uppermost layer and occasionally within those of koilocytes in the intermediate layer (Fig. 5B and 5C). No signals were identified in the basal level of epithelium. We interpreted only those nuclei marked different from the negative control as positive. Negative control slides where either Taq polymerase or primers were omitted from the PCR reaction mixture revealed no signals. Quantitative grading system was applied in the same way as that in ISH. HPV types 11, 16 and 33 were commonly detected in up to 7 of 9 (77.8%) cases, whereas HPV type 6 was detected in only one case and type 18 was not detected in a single case. Coinfection of two or more different HPV types were identified in seven cases (77.8%). As shown in Table 2, ISH as well as IHC tended to be far inferior to IS-PCR. Cases 5, 8 and 9 failed to demonstrate the L1 antigen, although both cases had proven coinfection by IS-PCR. Moreover, L1 staining intensity does not seem to correlate with multiple virus types of infection or with specific HPV types. The sensitivity of ISH was also quite less than that of IS-PCR. Low-risk type was detected in 4 of 7 (57.1 %) cases which were detected in the IS-PCR; intermediate-risk type in 3 of 6 (50%) cases; and high

The clinical parameters including age, size and stage were compared with HPV typing by in situ PCR, then the results were interpreted as the statistical non-parametric analysis. There was no significant correlation between any parameters and HPV typing (Table 3).

Discussion Malignant warty lesion of the uterine cervix is a newly described entity which has been recently subclassified as a specific variant of SCC [8} and very similar to the tumor described in the vulva [19}. Synonymous expressions include "scc following condyloma acuminatum" [7, 21] and "scc associated with BuschkeLoewenstein's giant condyloma" [3]. Other entities such as verrucous carcinoma or papillary SCC are usually differentiated from this lesion with ease by histologic findings alone. Verrucous carcinoma has characteristic bulbous rete pegs deep in the stroma and a complete absence of koilocytes. It also does not exhibit an aggressive clinical behavior [9]. Papillary SCC has papillary fronds lined by epithelial cells simulating transitional cells without squamous maturation or koilocytes. This entity may be deeply invasive and behaves in an aggressive manner [9, 17, 18k]. In contrast to those entities, warty carcinoma is a hybrid of condyloma acuminata and SCc. It is, however, too difficult to glean from the literature the cumulative data about HPV typing and biologic behavior, as well as morphology. Warty carcinoma is characterized by the presence of koilocytes in the superficial and intermediate level of

In situ PCR of HPV in Cervical Warty Carcinomas . 719

the epithelium and a definite invasion into the stroma [9, 19]. We observed that the surface of warty carcinoma is commonly feathery and thin. Verrucous carcinoma or condyloma acuminata usually show a stout and smooth keratotic surface in contrast to warty carcinoma. In addition, warty carcinoma is commonly frequent in the postmenopausal period in contrast to condyloma acurninata occurring during reproductive age. This observation may be useful in the differential diagnosis, even in small biopsy specimens. HPVs are classified into cutaneous and mucocutaneous group [22]. The cutaneous group includes benign warts caused by HPV types 1,2 and 4, and epidermodysplasia verruciformis warts that progress to cancer associated with HPV types 5 and 8, namely the high-risk group. In the mucocutaneous group, the viruses may be broadly classified into those with a low risk and those with a moderate-to-high risk for progressing to cancer. HPV 6 and 11 primarily cause benign exophytic genital warts, or condylomata acuminata. These are the viruses present in more than 90% of condylomata, with about two-thirds caused by HPV 6 and the remaining one-third by HPV 11 [22]. Although others have identified HPV 6 and 11 in only about 20% of cervical LSILs [16J, generally they are associated with LSILSs and are only rarely associated with high-grade squamous intraepithelial lesions (HSILs) or SCe. High-oncogenic risk viruses are predominantly HPV type 16, 18 and 31, which are commonly detected in women with HSILs and SCe. A number of viruses are classified into the intermediate-risk types, and include HPV types 33,35,39,51 and 51 [22, 16]. These moderate- or intermediate-risk viruses are still incompletely described because of a relative lack of probes and the relatively recent discovery of some members of this group. We selected five types of HPV (6, 11, 33, 16 and 18) which are known as the most prevalent types to infect the uterine cervix and which are associated with the entire range of cervical neoplasia (benign exophytic condyloma, intraepithelial neoplasia and SCC). We found that synchronous infections with HPV types 11, 33 and 16 were frequent in warty carcinoma of the cervix. HPV type 6 was detected only in one case and type 18 was not found in any of our cases. Exophytic condylomata not associated with HPV type 6 are usually associated with HPV type 11, but can also be associated with type 16 or 21, though HPV 6 is the most common type found in exophytic condylomata [16]. One case (case 9) where HPV type 6 was demonstrated was associated with intermediate-risk type (31, 33 and 35) by ISH and with type 11, but not with type 33 by ISPCR. This unusual result can be explained by the fact that intermediate-risk type 31 or 35 except 33 must be related. Exophytic lesions in adults are hardly ever associated with HPV type 18 or cutaneous types such as 1 or 2 [16]. We also demonstrated that no HPV type 18 was infected in the warty carcinoma.

The reported incidence of infection by two or more different HPV types in a given genital tract lesion varies from 2 to 71 %, with most series reporting an incidence of 10 to 28% [13, 20, 23]. HPV type 16 and 18 have been detected in about 20% of LSILs of the cervix [16J. Infection with more than one type of HPV is common in women with LSILs [16]. In one study using PCR, more than one type of HPV were detected in 22% of LSILs [12]. Unlike LSILs, multiple types of HPV are not commonly detected in women with HSILs [16, 12J. We identified mixed infection in 77.8% of warty carcinomas. Such a comparatively high incidence of multiple virus infection appears to be another characteristic of warty carcinoma. Specifically, infection with high-risk HPV type 16 combined with infection with low-risk HPV type 11 and/or intermediate-risk HPV type 33 except one case (case 9) showing the coinfection of lowand intermediate-risk type. The rate of detection of two or more HPV types in a single genital tract lesion is approximately lO-fold greater than that identified with in situ hybridization analysis [13]. When multiple viruses are found infecting the same lesion, types other than the dominant type are present in extremely-low copy numbers because of the very early stage or regressing stage, or because of competitive inhibition of replication of minor types by the dominant type [13]. According to one hypothesis, multiple infections in a warty carcinoma occur when a lesion containing low-risk virus is followed by the introduction of intermediate- or high-risk virus once the lesion is formed. The more virulent virus would then inhibit the replication of the less virulent virus, which is then found in low copy numbers [13,2]. In situ PCR method represents an attempt to wed the techniques of PCR and ISH through the amplification of a specific nucleic acid sequence inside a single cell which could detect as few as one viral copy number per cell using the nonisotopic PCR in situ assay [15, 6, 4, 24]. It has been well established that IS-peR is easier for the detection of low numbers of virus in a quiescent state than ISH, which requires at least 20 viral genomes per cell for detection and the virus in a state of active proliferation [14J. We can identify that the amplification of signals predominantly occurred in the superficial and intermediate layer, but not through basal and parabasal cells, and that there is a greater intensity and/or distribution in hybridization signals when ISPCR is being applied in contrast to ISH or me. Generally, the efficiency of IS-PCR is much lower in archival tissue sections than in suspended cells [l1J. Possible explanations for this finding include mechanical factors such as poor thermal conduction, uneven convection patterns in the PCR mixture under a coverslip, or possible adsorption of Taq DNA polymerase to glass [11]. Alternatively, tissue-related factors including inhibitors of the Taq polymerase and poor quality of target DNA may be important in the reduced amplification efficien-

720 . Nam Hoon Cho et al. cy observed in tissue sections [11]. Our results demonstrated variable intensity and proportion of HPV typing with a common finding of strong signals in the uppermost surface. Not all koilocytes contained HPY. This can be explained by differences in membrane permeability, nuclear protein and other causes for inaccessibility to the viral DNA [lOJ. Another possibility may be the loss of amplification during the washing step necessary for the abolishment of nonspecific signals. False positive signals could cloud interpretations. False positive signals may in part be the result of artifacts occurring due to the diffusion and extracellular generation of amplifiers [5]. Irrespective of these false artifacts, in situ peR appears to be effective and sensitive for the detection of low copy DNA in archival tissues. In conclusion, warty carcinomas of the uterine cervix are a distinct variant of invasive squamous cell carcinoma intimately related with HPV types of at least two or more types occurring most frequently in postmenopausal woman.

References 1. Bergenon CA, Ferenezy A, Shah KV, Nahashfar Z (1987) Multicentric human papillomavirus infections of the female genital tract: Correlation of viral types with abnormal mitotic figures, colposcopic presentation, and location. Obstet Gynecol69: 736-742 2. Branch AD, Benefeld Bl, Franck ER, Shaw JF, Varban ML, Willis KK, Rosen DL, Robertson HD (1989) Interference between coinoculated viroids. Virology 163: 538-546 3. Dawson DF, Duckworth lK, Bernhardt H, Yung KJM (1965) Giant condyloma and verrucous carcinoma of the genital area. Arch Pathol 79: 225-231 4. Gosden 1, Hanratty D (1993) PCR in situ: a rapid alternative to in situ hybridization for mapping short, low copy number sequences without isotopes. Biotechniques 15: 78-80 5. Komminoth P, Long AA, Ray R, Wolfe Hl (1992) In situ polymerase chain reaction detection of viral DNA, single copy genes and gene arrangements in cell suspensions and cytospins. Diagn Mol Pathol1: 85-97 6. Komminoth P, Long AA (1993) In situ polymerase chain reaction. An overview of methods, applications and limitations of new molecular technique. Virchow Arch [B] Cell Pathol64: 64-73 7. Kovi 1, Tillman L, Lee SM (1974) Malignant transformation of condyloma acuminatum. Am 1 Clin Pathol 61: 702-710 8. Kurman RJ, Norris Hl, Wilkinson E (1992) Tumors of the cervix. In: Kurman RJ, Norris HK, Wilkinson E (Eds) Atlas of Tumor Pathology. Tumors of the cervix, vagina and vulva, pp 29-36. AFIP, Washington, D.C: 9. Kurman RJ, Norris Hl, Wilkinson E (19929 Tumors of the cervix. In: Kurman RJ, Norris HK, Wilkinson E (Eds) Atlas of Tumor Pathology. Tumors of the cervix, vagina and vulva, pp 73-76. AFIP, Washington, D.C.

10. Liu AH, Creadon G, Wysocki LJ (1992) Sequencing heavy- and light chain variable region genes of single Bhybridoma cells by total enzymatic amplification. Proc Natl Acad Sci USA 89: 7610-7614 11. Long AA, Komminoth P, Lee E, Wolfe Hl (1993) Comparison of indirect and direct in-situ polymerase chain reaction in cell preparations and tissue sections. Histochemistry 99: 151-162 12. Lungu 0, Sun XW, Felix 1, Richart RM, Silverstein S, Wright TC (1992) Relationship of human papillomavirus type to grade of cervical epithelial neoplasia. lAMA 267: 2493-2496 13. Nuovo Gl, Darfler MM, Impraim CC, Bromley SE (1991) Occurrence of multiple types of human papillomavirus in genital tract lesions, analysis by in situ hybridization and the polymerase chain reaction. Am 1 Pathol138: 53-58 14. Nuovo GJ (1992) In situ hybridization. In: Nuove GJ (Ed) PCR in situ hybridization, pp 83-156. Raven Pr., New York 15. Nuovo Gl (1992) PCR in situ hybridization. In: Nuovo GJ (Ed) PCR in situ hybridization, pp 157-220. Raven Pr., New York 16. Park T, Fujiwara H, Wright TC (1995) Molecular biology of cervical cancer and its precursors. Cancer 76: 1902-1913 17. Qizilbash A (1974) Papillary squamous tumors of the uterine cervix. A clinical and pathological study of 21 cases. Am 1 Clin Pathol61: 508-520 18. Randall ME, Anderson WA, Mills SE, Kim lA (1986) Papillary squamous cell carcinoma of the uterine cervix: a clinico-pathological study of nine cases. Int 1 Gynecol Pathol5: 1-10 19. Rastkar G, Okagaki T, Twiggs LB, Clark BA(1982) Early invasive and in situ warty carcinoma of the vulva: Clinical, histologic, and electron microscopic study with particular reference to viral association. Am 1 Obstet Gynecol 143: 814-820 20. Reid R, Greenberg M, Jenson AB, Husain M, Willett J, Davoud Y, Temple G, Stanhope CR, Sherman AI, Phibbs GD (1987) Sexually transmitted papillomavirus infections. I: The anatomic distribution and pathologic grade of neoplastic lesions associated with different viral types. Am 1 Obstet Gynecol156: 212-222 21. Shafeek MA, Osman MI, Hussein MA (1979) Carcinoma of vulva arising in condylomata acuminata. Obstet Gynecol54: 120-123 22. Stoler MH (1997) The biology of human papillomaviruses and their role in cervical carcinogenesis. In: Bonfiglio TA, Erozan YA (Eds) Gynecologic Cytopathology, pp 51-71. Lippincott-Raven Pr., Philadelphia 23. Twiggs LB, Okagaki T, Clark B, Fukushima M, Ostrow RS, Faras Al (1988) A clinical, histopathologic, and molecular biologic investigation of vulvar intraepithelial neoplasia. Int 1 Gynecol Pathol 7: 48-55 24. Zehbe I, Hacker GW, Rylander E, Sallstrom 1, Wilander E (1992) Detection of single HPV copies in SiHa cells by in situ polymerase chain reaction (in situ PCR) combined with immunoperoxidase and immunogold-silver staining (IGSS) techniques. Anticancer Res 12: 2165-2168 Received: February 2, 1998 Accepted in revised form: luly 7, 1998