GYNECOLOFIC

ONCOLOGY

41, 123-128 (1991)

Identification of Human Papillomavirus Type 16 in Primary and Recurrent Cervical Canker following kdiation Therapy’ ROBERTW. HOLLOWAY, M.D.,**2 MICHAEL P. FARRELL, PH.D.,* CARLOSCASTELLANO, M.D.,? WILLARD A. BARNES, M.D.,* GEORGELEWANDOWSKI,M.D.,* BENNETT JENSON,M.D.,* CARLOSSANTOS,M.D.,? GRACIELA RAMIREZ, M.D.,t AND GREGORIODELGADO, M.D.* *Department

of Obstetrics and Gynecology, Division of Gynecologic Oncology, and #Department of Pathology, Georgetown Washington, D.C. 20007; and TDepartment of Gynecology, National Cancer Institute, Lima, Peru

University Hospital,

Received October 26, 1990

acid (RNA) in cervical cancer cells is unknown. To date, no reports have specifically described the presence of HPV DNA or its hybridization characteristics in recurrent cervical carcinoma following treatment of the primary cancer with radiation therapy. This report describes the use of 35S-labeledRNA probes and in situ hybridization to identify and type HPV in patients with primary cervical cancer and recurrent lesions following initial treatment with radiation therapy. Comparisons of primary and recurrent cancer are made with regard to HPV type, in situ hybridization signal patterns, and hybridization signal strength.

Formalin-fixed, paraffin-embedded tissue blocks from 13 women with cervical carcinoma that recurred following radiation therapy were evaluated for the presenceof human papillomavirus (HPV) by in situ hybridization using ribonucleic acid 35S-labeled probes for HPV types 6, 11, 16, and 18. Ten of thirteen patients also had pretreatment biopsies from their primary tumors available for analysis. HPV 16 was detected in both primary and recurrent lesions in 4 women. In 1 case, HPV was detected in the primary tumor and not in the recurrence. HPV 16 was also present in three recurrent cancers from which primary lesions were not available for probing. Radiation therapy did not alter the hybridization signal strength or pattern, suggesting that the HPV genome copy number was not significantly affected. The persistenceof HPV 16 in recurrent cervical carcinoma is consistent with the theory that HPV plays a role in maintaining the malignant state. 0 1991 Academic Press, Inc. INTRODUCTION

Human papillomavirus (HPV) deoxyribonucleic acid (DNA) has been detected in dysplastic, malignant, and metastatic genital tract lesions [l-3]. HPV types 16 and 18 are detected in the majority of high-grade cervical dysplasia and cervical carcinoma [4]. While ionizing radiation inflicts cell death and sublethal injury in mammalian cells through damage to nuclear DNA (including disruption of nucleotide bonds, chromosomal aberrations, and point mutations), the effect of ionizing radiation on the molecular hybridization of HPV DNA and ribonucleic ’ This project was supported by Biomedical Research Support Grant RR 05360 and P.H.C. Grant CA-50182. ’ To whom reprint requests should be addressed at: Division of Gynecologic Oncology, The Watson Clinic, 1600 Lakeland Hills Blvd., Lakeland, FL 33805.

MATERIAL

AND METHODS

Formalin-fixed, paraffin-embedded tissue blocks from 13 women with cervical carcinoma that recurred following radiation therapy were obtained from the National Cancer Institute, Lima, Peru (Table 1). Patients were treated with a combination of teletherapy (4-MeV linear accelerator) and radium brachytherapy. Point A doses ranged from 8000 to 8500 cGy. The median patient age was 44 years and the median time to disease recurrence following radiation therapy was 19 months. Eleven women had squamous cell carcinoma and two had adenosquamous carcinoma of the cervix. Ten of the thirteen patients with recurrent carcinoma also had tissue blocks from their primary, preirradiated cancers available for HPV probing. Ten recurrent lesions were located within the radiation field and three were distant. Pathologic diagnoses were confirmed by one coauthor (B.J.) prior to in situ hybridization. Hospital records were reviewed to confirm pertinent clinical information. Each tissue block was processed for hybridization in two separate experiments. 123 CO90-8258/91$1.50 Copyright 0 1991 by Academic Press, Inc. All rights of reproduction in any form reserved.

124

HOLLOWAY

TABLE 1 Characteristics of Patients with Recurrent Cervical Carcinoma Case no.

-4s (years)

Stage

1 2 3 4 5 6 7 8 9 10 11 12 13

34 37 37 52 44 58 65 58 57 53 39 42 40

IIb IIb IIb IIIb IIIb IIIb IIIb Ib IIb IIb IIIb IIb IIb

6 40 20

14 28

16 29

13 19 1.5 12 83

Pathology G G G G G G G G G G G G G

III adsq III squ/k III squ/k II squ/k III squ/nk I squ/k III adsq II squ/k I squ/nk II squ/nk III squ/nk II squ/nk II squ/k

Site of recurrence Pelvis Pelvis Lung Pelvis Pelvis Pelvis Pelvis Pelvis Pelvis Pelvis Lung Pelvis Mediastinum

Note. Abbreviations: adsq = adenosquamous, squ = squamous, k = keratinizing, nk = nonkeratinizing. ” Represents time from completion of radiation therapy to discovery of recurrent cancer.

Genomes of HPV types 6b, 11, 16, and 18 were isolated from the appropriate PBR 322 plasmid vector and recloned into the polylinker of the pGem transcription vectors (Promega Biotec). The orientation of the HPV relative to the Sp6 and T7 promoters was determined by restriction enzyme digestion and gel electrophoresis. The pGem derivatives were purified by cesium chloride centrifugation and linearized by cutting in the polylinker with the appropriate restriction enzyme. After phenol extraction, ?SUTP-labeled single-stranded antisense RNA probes were synthesized with Sp6 RNA polymerase (Promega Biotech). Template DNA was digested with RNAase-free DNAase I (Worthington, DPRF grade), phenol chloroform extracted, and the “S-labeled probe separated from the reaction mixture on a Nensorb column (New England Nuclear). “S-labeled RNA probe was vacuum dried and frozen at -70°C until used (within 48 hr). Specific activity of the probes ranged from 0.9 to 1.02 x IO’ dpm/pg RNA. Sections of buffered, formalin-tixed, paraffin-embedded cervical cancer biopsies were mounted on slides pretreated with 3aminopropyltriethoxysilane. ?S-labeled RNA probes were resuspended at 0.25 pg/ml in 50% formamide/lO% dextran sulfate buffer, placed directly on tissue sections (6 PI/cm*), covered with siliconized glass coverslips, and sealed with rubber cement. Slides were heated in a 90°C water bath for 6 min, transferred to a -20°C freezer for 5 min, then incubated 12-16 hr at 42°C in a humidified box. Slides were washed 4 hr at 55°C in several changes of solution consisting of 2 x SSC adjusted to 40 mM sodium phosphate buffer, pH 6.8; 1% dried nonfat milk; and 0.1% sodium dodecyl sulfate. Standard autoradiography using Kodak NB-2 nuclear tract emulsion was performed and sections were stored in the dark at 4°C for 14 days. Slides were developed in Kodak D-19 developer, fixed, and stained with dilute hematoxylin and eosin. FIG.

1.

TABLE 2

In Situ Hybridization Results in Patients with Post-radiation Therapy, Recurrent Cervical Cancer

Time to recurrence” (months) 15

ET AL.

Probe preparation and in situ hybridization.

Case no. 1 2 3 4 5 6 7 8 9

10 11 12 13

G G G G G G G G G G G G G

Pathology

Primary cancer

III adsq III squ/k III squ/k II squ/k III squ/nk I squ/k III adsq II squ/k I squ/nk II squ/nk III squ/nk III squ/nk II squ/k

HPV 16 HPV 16 HPV 16 HPV 16 HPV 16 n.a. n.a. n.a.

Recurrent cancer

Hybridization pattern

HPV 16

None Focal strong + None Focal strong + None Uniform Uniform None None Focal strong + Uniform Uniform Uniform

HPV HPV HPV HPV HPV HPV

16 16 16

16 16 16

Note. Abbreviations: squ = squamous, adsq = adenosquamous, k = keratinizing, nk = nonkeratinizing, n.a. = not available.

Human skin grafts infected with HPV 11 grown under renal capsules in athymic mice [5] were used as positive controls for the in situ hybridization experiments. Various techniques of in situ hybridization of paraffinembedded tissues have been described in detail [6-91. We utilized features from each of these protocols and included three minor modifications (Fig. 1). Hybridization slides were viewed under standard light microscopy by two authors independently (R.W.H., M.P.F.), with adjudication by a third author (A.B.J.) when there was a difference of opinion. Criteria for positive results included at least a threefold increase of intranuclear silver grains compared to background binding of silver grains to nonmalignant cells or blank slide (i.e., a signal/noise ratio at least 3: l), more than one cell per x400 view with hybridization signal, and no such detection of hybridization signal in adjacent sections with the other three HPV probes. RESULTS

Results of the in situ hybridization experiments (Table 2) indicate that HPV 16 was the only virus type detected and was present in 8 of 13 cases. HPV 16 was identified in both primary and recurrent lesions (all within the radiated field) from 4 patients and in the primary lesion alone in 1 patient. HPV 16 was also identified in 3 recurrent cancers from which paraffin blocks of the primary (preirradiated) lesions were not available for probing; 2 of these recurrent cancers were outside the radiation treatment ports (Table 1). In situ hybridization signal densities (i.e., the density of silver grains deposited over malignant cells relative to background binding of silver grains) in primary and re-

HPV IN CERVICAL

CANCER FOLLOWING

RADIATION

THERAPY

125

reaction mixture. Recovery of probe by this method was efficient and allowed easy concentration of radioactive probe without excess salt and buffer which might alter hybridization. Finally, the length of posthybridization washing protocols can range from 1 hr to 3 days, using various salt and formamide concentrations and temperatures to vary stringency [6-131. Our washing protocol used salt and temperature conditions of moderate stringency and lasted 4 hr. We added 0.1% sodium dodecyl sulfate as a detergent to improve washing and 1% nonfat dried milk to nonspecifically bind radioactive probe. This detergent buffer has been used on nylon membranes in Southern blot hybridization, resulting in extremely low background signal [14]. There was no microscopic evidence of tissue destruction following washes with this buffer and background signal was acceptable, allowing easy identification of positive slides. The detection rate of HPV in cervical cancer using in situ hybridization in this investigation was comparable to that in other recent reports. HPV has been identified using DNA-DNA in situ hybridization in as many as 80% of cervical cancers [7,11,13,15]. Using single-stranded RNA probes, HPV 16 was identified in 12 of 17 (70%) invasive cervical cancers in another report, with an estimated sensitivity of 20 to 50 genome copies per cell [ 161. Because the sensitivity of our in situ hybridization system was not determined, the apparent lack of HPV 18 in these cancers may actually be secondary to a genome copy number below the sensitivity of our system. Two different patterns of in situ hybridization signal were recognized in this study. Schnieder et al. [16] previously described a uniform distribution of signal in all CIN III lesions and in 15 of 17 invasive carcinomas; focal signal grain density was characteristic of CIN II lesions in areas of epithelial maturation; however, it was also present in two moderately differentiated carcinomas. We found that cancers with the focal pattern had silver grains distributed more densely over relatively fewer malignant cells, evenly positioned throughout the biopsy. Although numbers were small, there did not appear to be a relationship between the tumor’s histologic grade and hybridization pattern. The significance of these differences in distribution and intensity of hybridization signal is unDISCUSSION certain, but may reflect the tumor’s biologic potential. Our in situ hybridization protocol was simplified by the Future studies concerning the molecular biology of HPV use of three techniques. While poly-L-lysine-coated glass and its transcriptional activity in cervical cancer cells may slides are traditionally used for mounting tissue sections, help to explain the observed variation in hybridization we found that up to 50% of sections were lost or damaged characteristics. during posthybridization washing. In contrast, glass slides Tissue from Case 10 hybridized for HPV 16 only in the coated with 2% (v/v) 3-aminopropyltriethoxysilane in primary tumor. The majority of the malignant cells in the acetone held tissue reliably through prolonged washing primary lesion did not hybridize (Figs. 2e and 2f) and and did not increase background hybridization signal. Sec- probably contained an HPV genome copy number below ond, a Nensorb (New England Nuclear) column was used the level of sensitivity for our system. The absence of to separate radioactive probe from the Sp6 polymerase signal in the recurrent tumor could be explained by a loss current lesions were compared for each patient. Although differences in hybridization signal were evident in comparing cancers from different patients, no appreciable differences in signal were noted in comparing primary and recurrent cancers from the same patient (Fig. 2). Two hybridization signal patterns were observed and termed “uniform” and “focal.” The uniform signal distribution was observed in five of eight HPV positive tumors in which malignant cells were covered with a relatively light coating of silver grains, equally distributed over nuclei and cytoplasm, approximately 5- to lo-fold more dense than background grains (Figs. 2a and 2b). A more focal signal distribution with approximately 50% of malignant cells showing a higher silver grain density (greater than 20-fold above background) was observed in three tumors. Silver grains were concentrated most heavily over cellular nuclei in the focal pattern. In Cases 2 and 4, strong hybridization signal was deposited over approximately 50% of malignant cells, evenly dispersed throughout the section (Figs. 2c and 2d). Case 10 demonstrated a focal distribution of hybridization signal which was limited to one discreet area of the tissue section, unlike the former two cases (Figs. 2e and 2f). A single HPV 11 infected human skin graft grown in a nude mouse and known to contain replicating virus was used as a positive control for the in situ hybridization technique. When hybridized under the conditions of this experiment with each of the four HPV probes, a strong signal occurred with the HPV 11 probe and a weaker signal resulted from the HPV 6 probe; no signal above background resulted with probes for HPV 16 and 18. Silver grain density in the HPV 11 control tissue was focally very high, concentrated in koilocytes and adjacent parabasal keratinocytes (Fig. 3). HPV 16 positive cervical cancers with the uniform pattern had less signal per cell than the HPV 11 control, but three cancers with the focal pattern for HPV 16 contained many malignant cells probing with a high-intensity signal similar to that of the mouse control (Figs. 2c, 2d, and 3). Probes for HPV 6, 11, and 18 produced no hybridization signal above background in any of the cervical cancers.

126

HOLLOWAY

ET AL.

FIG. 2. Tissue sections of primary and recurrent cervical cancer biopsies. Black dots are silver grains exposed by 35S-labeledRNA HPV 16 probe. (a) Uniform hybridization signal pattern from a pretreatment biopsy of Case 7 showing silver grains distributed over nuclei and cytoplasm ( x 150); (b) similar uniform pattern in the recurrent, postirradiated biopsy from Case 7 ( x 150); (c) f ocally strong intranuclear signal for HPV 16 in a pretreated cancer (Case 4, x 160); (d) similar high-density signal pattern in the recurrent biopsy from Case 4 following radiation therapy ( x 150); (e) hybridization signal limited to a specific area of the biopsy in Case 10 ( x 30); (f) X 150 magnification of case 10 showing focal pattern.

HPV IN CERVICAL

CANCER FOLLOWING

RADIATION

THERAPY

127

transcription level. 35S-labeled probes typically require 50 to 500 viral genomes per cell to produce signal above background for positive results with in situ hybridization [13,16]. An abundance of viral DNA would be expected only in cells permissive for viral DNA replication. Squamous carcinoma of the cervix is capsid antigen negative by immunocytochemistry [12,17], indicating that late gene expression is absent and viral replication is not taking place. Early sequence HPV mRNA transcripts which code for regulatory proteins have been detected in human cervical cancer and cervical cancer cell lines [18-201. Therefore, the strong hybridization signal observed, especially in the tumors with focal distribution, probably represents accumulated mRNA transcripts. Focal distribution of signal may indicate heterogeneity with respect to subtle aspects of the differentiation states of malignant cells, and this did not change following radiation therapy. Furthermore, none of the primary tumors that probed negative converted to positive following radiation therapy. These results tend to question any hypothesis that ionizing radiation might induce a more aggressive malignancy by altering the HPV gene copy number or expression. These in situ hybridization experiments on a limited number of cervical cancer specimens are the first to demonstrate the persistence of HPV DNA in postirradiated recurrent cervical cancer. Using hybridization signal strength as a crude measure of total cellular HPV DNA and RNA, these experiments suggest that radiation therFIG. 3. Section of HPV 11 infected human foreskin epithelia grown apy probably has little effect on the total HPV DNA and in an athymic mouse, probed with ‘%labeled HPV 11 RNA p robe. RNA content in recurrent cervical cancer. The persistence Basal epithelial cells are at the top and maturing keratinocytes are near of HPV nucleic acid and the likelihood that it is tranthe bottom of the photograph. Strong in situ hybridization signial is scribed in recurrent cervical carcinoma support the curlocated over nuclei in koilocytes and suprabasal keratinocytes (X 100). rent theory that HPV may play an active role in maintaining the malignant state. of viral genome or a reduction in copy number accompanying a change from an episomal to an integrated state, REFERENCES although this apparently did not occur in the other four HPV positive tumors. Given the small area of hybridiDe Villiers, E. M., Schneider, A., Gross, G., and zur Hausen, H. zation in this patient’s original tumor, the lack of signal Analysis of benign and malignant urogenital tumors for human in this particular recurrent lesion is more likely a result papillomavirus infection by labelling cellular DNA, Med. Microbial. of biopsy “sampling error” within the tumor. Zrnmunol. 174, 281-286 (1986). This study demonstrates the persistence of HPV nucleic Gissmann, L., Wolnick, L., Ikenberg, H., Koldovsky, U., acid in cervical carcinoma that recurred following therSchnurch, H. G., and zur Hausen, H. Human papillomavirus types 6 and 11 DNA sequences in genital and laryngeal papillomas and apeutic doses of radiation therapy. The HPV genome is Sci. USA 80, 560-563 in some cervical cancers, Proc. Natl. Ad. a small radiation “target” in comparison to the cancer (1983). cell DNA; consequently, much higher radiation doses Lancaster, W. D., Castellano, C., Santos, C., Delgado, G., Kurwould be required to reliably destroy HPV DNA. Under man, R. J., and Jenson, A. B. Human papillomavirus deoxyribothe conditions of our in situ hybridization experiments nucleic acid in cervical carcinoma from primary and metastatic sites, (denaturation of HPV DNA in the tissue section and use Am. J. Obstet. Gynecol. 154, 115-119 (1986). of anti-sense RNA probes), both viral DNA and mRNA Lorincz, A. T., Temple, G. F., Kurman, R. J., Jenson, A. B., and transcripts can be detected. Biopsies from the primary Lancaster, W. D. Oncogenic association of specific human papiland recurrent cancer showed similar hybridization signal lomavirus types with cervical neoplasia, JNCI 4, 671-677 (1987). density, suggesting that ionizing radiation did not signifKreider, J. W., Howett, M. K., Lill, N. L., Bartlett, G. L., Zaino, icantly alter the total HPV genome copy number or HPV R. J., Sedlacek, T. V., and Mortel, R. In viva transformation of

128

6.

7.

8.

9.

10.

11.

12.

13.

HOLLOWAY

human skin with HPV type 11 from condylomata acuminata, J. Viral. 59(2), 369-376 (1986). Nagai, N., Nuovo, G., Friedman, D., and Crum, C. P. Detection of papillomavirus nucleic acids in genital precancers with the in situ hybridization technique, Znt. /. Gynecol. Pathol. 6,36f-379 (1987). Stoler, M. H., and Broker, T. R. In situ hybridization detection of human papillomavirus DNAs and messenger RNAs in genital condylomas and a cervical carcinoma, Hum. Pathol. 17, 1250-1258 (1986). Hasse, A., Brahic, M., Stowring, L., and Blum, H. Detection of viral nucleic acids by in situ hybridization, Methods. Virof. 7, 189226 (1984). Cox, K. H., DeLeon, D. V., Angerer, L. M., and Angerer, R. C. Detection of mRNAs in sea urchin embryos by in situ hybridization using asymmetric RNA probes, Dev. Biol. 101, 485-502 (1984). Crum, C. P., Nuavo, G., Friedman, D., and Silverstein, S. J. A comparison of biotin and isotope-labeled ribonucleic acid probes for in situ detection of HPV-16 ribonucleic acid in genital precancers, Lab. Invest. 58, 354-359 (1988). Caussey, D., Orr, W., Daya, A. D., Roth, P., Reeves, W., and Rawls, W. Evaluation of methods for detecting human papillomavirus deoxyribonucleic sequences in clinical specimens, J Clin. Microbial. 26(2), 236-243 (1988). Collins, J. E., Jenkins, D., and McCance, D. J. Detection of human papillomavirus DNA sequences by in situ DND-DNA hybridization in cervical intraepithelial neoplasia and invasive carcinoma: A retrospective study, J. Clin. Pathol. 41(3), 289-295 (1988). Ostrow, R. S., Manias, D. A., Clark, B. A., Okagaki, T., Twiggs, L. B., and Faras, A. J. Detection of human papillomavirus DNA

ET AL. in invasive carcinomas of the cervix by in situ hybridization, Cancer Rex 47, 649-653 (1987). 14. Reed, K. C., and Mann, D. A. Rapid transfer of DNA from agarose gels to nylon membranes, Nucleic Acids Res. 13, 7207-7211 (1985). 15. Tase, T., Okagaki, T., Clark, B. A., Manias, D. A., Ostrow, R. S., Twiggs, L. B., and Faras, A. J. Human papillomavirus types and localization in adenocarcinoma and adenosquamous carcinoma of the uterine cervix: A study by in situ DNA hybridization, Cancer Res. 48(4), 993-998 (1988).

16. Schneider, A., Oltersdorf, T., Schneider, V., and Gissmann, L. Distribution pattern of human papilloma virus 16 genome in cervical neoplasia by molecular in situ hybridization of tissue sections, ht. J. Cancer 39, 717-721 (1987). 17. Kurman, R. J., Jenson, A. B., and Lancaster, W. D. Detection of human papillomaviruses by immunocytochemistry, in Advances in (R. A. DeLellis, Ed.), Masson, New York, immunocytochernistry p. 201 (1984). 18. Schneider-Gadicke, A., and Schwarz, E. Different human cervical carcinoma cell lines show similar transcription patterns of human papillomavirus type 18 early genes, EMBOJ. 5, 2285-2292 (1986). 19. Tsunokawa, Y., Takebe, N., Nozawa, S., Kasamatsu, T., Gissmann, L., zur Hausen, H., Terada, M., and Sugimura, T. Presence of human papillomavirus type-16 and type-18 DNA sequences expression in cervical cancers and cell lines from Japanese patients, Inr. J. Cancer 37, 499-503 (1986). 20. Pater, M. M., and Pater, A. Expression of human papihomavirus types 16 and 18 DNA sequences in cervical cancer cell lines, J. Med. Virol. 26(2), 185-195 (1988).