Glucocorticoid-enhanced neoplastic transformation of human keratinocytes by human papillomavirus type 16 and an activated ras oncogene

VIROLOGY

173,767-77

1 (1989)

Glucocorticoid-Enhanced Neoplastic Transformation of Human Keratinocytes by Human Papillomavirus Type 16 and an Activated ras Oncogene MAITHIAS DURST,* DAN GALLAHAN,* GILBERT JAY,t AND JOHNG S. RHIM+’ *Institut fiir Virusforschung, Deutsches Krebsforschungszentrum. 6900 Heidelberg, Federal Republic of Germany; tLaboratory of Virology, American Red Cross, Rockville, Maryland 20855; and *Laboratory of Cellular and Molecular Biology, National Cancer Institute, Bethesda, Maryland 20892 Received July 13, 1989; accepted August 25, 1989 Human papillomaviruses (HPV) are known etiological agents of benign proliferation of the skin and mucosa (papillomas and warts). They have also been implicated in the development of cervical dysplasia and anogenital carcinoma. The close association of HPV type 16 DNA with the majority of cervical carcinomas suggests the involvement of the virus in this type of cancer. We have developed an in vitro multistep model for human epithelial cell carcinogenesis. Primary human epidermal keratinocytes acquired an indefinite lifespan in response to transfection with HPV 16 DNA but did not undergo malignant conversion. Addition of Kirsten murine sarcoma virus (Ki-MSV), which contains an activated K-ras oncogene, to these cells induced morphological alterations associated with the acquisition of neoplastic properties. The frequency of transformation by Ki-MSV was markedly enhanced by the inclusion of glucocorticoid. At optimal conditions, a 125-fold stimulation was observed. These findings demonstrate the malignant conversion of human primary epithelial cells in culture by the cooperation of HPV type 16 and an activated ras oncogene and support a multistep process for neoplastic conversion. The availability of a human epithelial cell transformation model should o 1989 Academic PESS, IK. facilitate studies of the interaction between HPV and human epithelial cells.

servative estimates suggest infection in the range of 1O-309/0 of the normal population (16, 17). With the use of more sensitive detection methods, such as the polymerase chain reaction (IS), this percentage is expected to increase even more dramatically. A considerable proportion of HPV infections are subclinical and remain in a latent state for unknown periods of time. Because of the widespread nature of HPV infections and the long incubation period between initial infection and development of cancer (19) it is quite clear that HPV infection by itself is not sufficient for neoplastic conversion. Various environmental factors have been implicated in the progression to malignancy. As is the case for other DNA tumor viruses, the development of HPV-associated cancer is expected to be a multistep process. The ability of the HPV genome to immortalize keratinocytes in vitro could represent an initial step towards malignant transformation. This immortalizing property appears to be specific for HPV types associated with cervical cancer (HPV 16, 18, 3 1, and 33), as it has not been observed for HPV types which induce benign genital lesions (HPV 6 and 11) (8, 10, 12). A functional dissection of the viral genome which would permit the identification of the gene(s) required for immortalizing human cells is currently in progress. In order to define additional events which are necessary for malignant conversion of normal epithelial cells, we used the human keratinocyte cell line (HPK-1A)

Cancers of epithelial origin are the most prevalent histological type in man. Despite this fact, little is known about the mechanism involved in the transformation event. To a large extent, this problem can be attributed to difficulties in establishing appropriate cell culture systems. It was possible only recently to propagate primary human epithelial cellsin vitro (I). Although infections with simian virus 40 (SV40) or transfection with SV40 DNA led to immortalization of these cells in culture, they failed to show malignant growth when injected into nude mice (2-5). Recent studies have shown that transfection of primary human foreskin epithelial cells with the DNA of human papillomavirus (HPV) can also lead to immortalization of these cells in culture (6- 12). To date some 58 different HPV genotypes have been identified (13, 14). Besides being the etiological agents of benign warts and papillomas, some HPV types (HPV 6, 11, 16, 18, 3 1, 33, and 35) are frequently detected in precancerous lesions of the genital tract, particularly those of the uterine cervix. Since the DNA of only this subset of HPV types (most notably HPV 16) also persists in a high percentage of cervical carcinomas (15) they have attracted much attention as potentially oncogenic human pathogens. Large scale population screenings have shown, however, that genital HPV infections irrespective of HPV types are extremely widespread. Con’ To whom reprint requests should be addressed. 767

0042-6822/89

$3.00

766

SHORT COMMUNICATIONS TABLE 1 BIOLOGICALPROPERTIESOF HUMAN EPIDERMALKERATINOCMESTRANSFECTEDBY HPV 16 DNA AND EXPOSEDTO Ki-MSV

Passage in culture (Passage number)

Cells Primary human keratinocytes + Ki-MSV + HPV 16 DNA(HPK-1A) + HPV 16 + Ki-MSV (Ki-MSV/HPK-1A) Note. Ki-MSV(BaEV) was produced a Saturation density was measured 37” with changes of growth medium b Nude mice were inoculated with by karyotypic analysis.

<3 <3 >50 >50

Saturation density” (X 105/cm’)

Agar colony formation (%)

Number of animals bearing tumorsb

1.9 4.2


015 515

in human nonproducer cells by superinfection with Baboon endogenous virus (13). as the maximum number of cells obtained after initial plating of 5 X 1 O3 cells/cm* and then incubating at every 3 days. 1O7 cells. Tumor cells were reestablished in tissue culture and were found to resemble the cells of origin

which was established from a primary foreskin cell culture in response to HPV 16 DNA transfection (6). This cell line is of clonal origin and harbors multiple HPV 16 genomes arranged in a head-to-tail fashion and integrated at a single chromosomal site. Although the cell line has been in culture for more than 2 years, the HPV 16 integration locus was shown to be stable and viral DNA has remained transcriptionally active. HPK-1 A cells are nontumorigenic in nude mice, irrespective of the number of passages of these cells in culture. The lack of tumorigenicity of the HPK-IA cells has led us to inquire whether its growth properties might be further altered by addition of a virus containing an activated ras oncogene whose cellular equivalent has been detected in many human malignancies (20). Neither control nor Ki-MSV-infected primary human epidermal cells could be propagated serially beyond two or three subcultures (Table 1). In contrast, infection of the HPK-1A line at passage 59 (Fig. 1A) with Ki-MSV (BaEV) (Baboon endogenous virus) (21) resulted in an alteration in cell morphology. As early as 7 to 10 days after infection, the cells began to pile up in focal areas, forming small projections and releasing round cells from the foci (Fig. 1 B). Approximately one focus per 1 x 1O* cells was observed using a 1 to 10 dilution of the stock virus supernatant. The absence of any detectable morphological alteration induced by the helper virus (BaEV) alone implied that Ki-MSV was responsible for induction of the transformed morphology. Analysis of the biological properties of the Ki-MSVtransformed cells revealed that their saturation density was approximately two times higher than that of parental HPK-1A cells. Moreover, transformants grew in soft agar with colony-forming efficiencies from 0.1 to 0.3% (Table 1). The presence and expression of the K-ras oncogene in the selected transformants were confirmed by Western immunoblot analysis using a pan-

reactive mouse antibody against ~21. Using this analysis, parental HPK-IA cells (Fig. 2A) were found to express a low level of endogenous p21 protein as indicated by a characteristic doublet on a SDS-polyactylamide gel which represented the phosphorylated and unphosphorylated forms of the gene product (4). In addition to the endogenous ~21, the transformed KiMSV/HPK-IA line (Fig. 2B) expressed high levels of the activated K-ras gene whose product was distinguishable by its slower migration on SDS-polyacrylamide gel. A similar pattern of expression of the activated Kras gene was also detected in the cell line 129 Nu 2409 (Fig. 2C), which had been adapted from a nude mouse bearing a tumor derived by subcutaneous injection of Ki-MSV/HPK-1 A cells. The expression of the activated K-ras gene in the selected transformants is further supported by the presence of the exogenous ras gene, as shown by Southern blot analysis (data not shown). HPV 16 DNA and RNA data indicated that the copy number and gene expression have remained unchanged (data not shown). When 129J nude mice were inoculated subcutaneously with 10’ Ki-MSV-transformed HPK-1A cells (passage 3), the animals developed progressively growing tumors within 2 to 3 weeks (Table 1). Cells around the tumor periphery appeared to be poorly differentiated and showed a high mitotic index (Fig. 1C). Cells toward the center had differentiation characteristics which included the lack of mitotic figures and a high cytoplasmic to nuclear ratio. These histologic findings are consistent with the tumors being squamous cell carcinomas. In contrast, subcutaneous inoculation of 10’ HPK-1 A cells into nude mice produced only regressing cystic nodules that contained epidermal cells (Table 1). Evidence of the human origin of all the cell lines was obtained by isoenzyme analysis and species-specific cell membrane immunofluorescence. The relatedness

769

SHORT COMMUNICATIONS

cells have similar marker chromosomes. They are aneuploid human male with chromosome counts in the triploid range, as detected by conventional staining. Moreover, there were no major changes in chromosome number. While two new marker chromosomes (M7 and M9) were found in the Ki-MSV-transformed and its tumor cell lines, these are relatively weakly represented. The representation of chromosome 11 in the Ki-MSV-transformed and its tumor cell lines appeared to be consistently greater than in the parental HPK-1A line. The HPK-IA cell line could also be neoplastically transformed using Polybrene at a concentration of 10 pg/ml, followed by a 4-min shock with 30% DMSO, by transfection with a plasmid carrying the activated, cellular H-ras gene from the EJ bladder carcinoma (22).

M,

FIG. 1. Comparison of the morphology of HPV 16transfected human keratinocytes (HPK-1 A) at the 59th passage (A) and a focus of Ki-MSV-infected HPK-IA cells 21 days after infection with Ki-MSV (BaEV) (B). Invasive squamous cell carcinoma induced in nude mice inoculated with HPK-1A cells infected with Ki-MSV (C).

of the Ki-MSV transformants to the parent HPK-1A cells was further established by chromosome analysis. Untransformed HPK-1 A cells, transformed, and tumor

A

B

C

1 2

1 2

1 2

FIG. 2. Expression of p21 protein in Ki-MSV transformants. Cell lysates from HPK-1 A (A), Ki-MSV/HPK-1 A passage 10 (B), and 129 Nu 2409 (C)were immunoprecipitated either with control mouse serum (lane 1) or with a pan-reactive mouse antibody against p21 (lane 2) and then subjected to Western immunoblot analysis using the same pan-reactive antibody against p21 and 1251-labeled protein A. The molecular weight (h/l,) markers used are shown on the left. Mefhods: Subconfluent cultures of HPK-lA, Ki-MSV/HPK-IA ~10. and 129 Nu 2409 cells were rinsed with phosphate-buffered saline and lysed with a buffer containing 50 mM Tris-HCI, pH 7.5, 150 mhll NaCI, 0.1% SDS, 1 .O%Triton X-l 00, and 1 .O% Na deoxycholate. The protein concentration of the lysates was determined and equivalent amounts of protein from each cell line were immunoprecipitated with either a control mouse serum or a pan-reactive mouse antibody agarnst p21 (Cetus Corp.). The resulting immunoprecipitates were fractionated on a 12.5% polyacrylamide-SDS gel. After electrophoresis. proteins on the gel were electroeluted onto a nitrocellulose membrane which was then reacted against the pan-reactive antibody against p2 1. The immunocomplex on the nitrocellulose membrane was visualized by a subsequent reaction with ‘Z51-labeled protein A.

770

SHORT COMMUNICATIONS TABLE 2

EFFECTOF HYDROCORTISONE ON TRANSFORMATIONINDUCEDBYKi-MSV OF HUMAN EPIDERMALHPK-1 A CELLS~ Concentration h3) 0 1 2.5 5.0 10.0

Virus titer (FFU?ml) 4.0 x 1.2x 2.5 x 5.0 x 3.0x

1o3 lo5 lo5 1o5 lo5

Enhancement (fold)

30 62.5 125 75

a Human epidermal HPK-1 A cells were inoculated with Ki-MSV and fed with Dulbecco’s modified minimum essential medium with 10% fetal bovine serum, with and without various concentrations of hydrocortisone. The infected cultures were incubated at 37” in 5% CO, and refed with the same medium, and foci were counted 14 days after infection. ’ Focus-forming units.

Transfection of 10 pg of pSV,ras DNA resulted in the rapid induction of transformed morphology. Approximately 10 to 20 foci were detected per 1 X 1O5 cells. As early as 2 weeks after transfection, the cells began to pile up in focal areas, forming small projections, and over times started to release round ceils. These transformed foci could be stained with Giemsa and were countable 3-4 weeks after transfections without selection. The transfected cells showed not only morphological alterations but induced carcinomas when transplanted into nude mice. It has previously been shown that hydrocortisone significantly enhanced transformation of human fibroblasts by Ki-MSV (23). The frequency of transformation of HPK-IA cells by Ki-MSV was also markedly enhanced by the addition of hydrocortisone (Table 2). The stimulation was dependent upon hormone concentration and a 125fold enhancement in frequency was observed with 5 pg/ml hydrocortisone. A hydrocortisone concentration as low as 1 pg/ml also had a significant effect. In the hydrocortisone-treated human epidermal cells, Ki-MSV produced larger, well-defined foci which could be counted 7 days after infection. In contrast, in untreated human epidermal cells, foci were small and could not be counted until 14 days after infection. In this report, we describe an in vitro model system for human epithelial cell carcinogenesis which provides further evidence in support of a specific role of certain HPV types in the development of genital cancer. Malignant conversion of human epithelial cells requires at least two, and possibly more, alterations in cell growth properties. In our model system, one such measurable event is the immortalization of the “natural” epithelial target cells in response to HPV 16 transfection in v&o. So far, this biological property appears

to be specific for HPV types associated with genital cancer and is not found for papillomaviruses which induce only benign lesions (8, 10, 12). In our system, the second step necessary for malignant conversion of HPV-immortalized human epithelial cells is the expression of an activated K-ras oncogene. This correlates well with the fact that amplification or activation of members of the ras oncogene family are observed in a high percentage of human tumors (20), including cervical cancer (24), and are thus likely to influence tumor progession in vivo. The importance of the combined effects of HPV 16 and activated K-ras oncogene in the induction of neoplastic human epithelial cells is further implied by the inability of Ki-MSV alone to induce continued proliferation of primary epithelial cells under our assay conditions. The cooperation of HPV 16 and activated ras oncogene to induce transformation has previously been suggested with rodent kidney cells (2%27), but not with human fibroblasts (28), in a hormone-dependant manner. The present study is unique in that it demonstrates the cooperativity in human epithelial cells, the cell type from which carcinomas are derived. Similar cooperativity results have recently been reported using human cervical cells (29). It is interesting to note the marked enhancing effect of hydrocortisone on Ki-MSV-induced transformation in the HPV16-immortalized cells. This effect may be due to an increased level of expression of HPV genes (30) which by itself has no phenotypic consequences, but in the presence of the activated ras protein augments transformation. Alternatively, hydrocortisone may in some way increase the potency of the activated ras oncogene resulting in increased transformation, i.e., producing larger foci or earlier appearance of foci. In this cell culture system, transformation seems to be a stepwise process. Infection with Ki-MSV is necessary but not usually sufficient for transformation. The effect of the virus thus resembles that of an “initiator,” as described for chemical carcinogenesis. Similarly, hydrocortisone appears to act here as a tumor promoter. Such an action would be in keeping with in vivo studies suggesting that certain hormones may act as tumorpromoting agents in their ability to enhance the incidence of cancer (3 7). ACKNOWLEDGMENTS We thank W. Peterson for performing the karyotypic analysis of the human epidermal cell lines and P. Arnstein for the tumorigenicity assay of the human epidermal lines in nude mice.

REFERENCES 7. FUSENIG,N. E., ln “Biology of the Integument,” (1. Bereiter-Hahn, A. Maltoltey, and K. Richards, Eds.), Vol. 2. p. 409. SpringerVerlag, Berlin/Heidelberg, 1986.

SHORT COMMUNICATIONS 2. STEINBERG,M. L., and DEFENDI, V., Proc. Nat/. Acad. Sci. USA 76, 801-805 (1979). 3. RHIM, J. S., eta/., Science 227, 1250-l 252 (1985). 4. RHIM, J. S., eta/., Science 232, 385-388 (1986). 5. CHANG, S. E., Biochem. Biophys. Acta 823, 161-194 (1986). 6. D~~RsT.M., eta/., Oncogene 1, 251-256 (1987). 7. PIRISIeta/., 1. Viral. 61, 1061-l 066 (1987). 8. WOODWORTH,C. D., eta/., CancerRes. 48,4620-4628 (1988). 9. KAUER,P., and MCDOUGALL, J. K../. wfo/. 62, 1917-l 924 (1988). 10. SCHLEGEL,R., eta/., EMBOJ. 7, 3181-3187 (1988). 11. MCCANCE, D.J..eta/., Proc. Nat/. Acad. Sci. USA85,7169-7173 (1988). 12. PECORARO,G., et al., Proc. Nat/. Acad. Sci. USA 86, 563-567 (1989). 13. COGGIN,J. R., and ZUR HAUSEN, H., Cancer Res. 39, 545 (1979). and Human 14. PFISTER.H., and FUCHS, P. G., /n “Papillomaviruses Diseases,” (K. Syrjsnen, L. Gissmann, and L. G. Koss. Ed.%), p. 1. Springer-Verlag. Berlin/Heidelberg, 1987. 15. PETO, R., and ZUR HAUSEN, H., Eds. /n “Banbury Report 21: Viral Etiology of Cervical Cancer.” Cold Spring Harbor Laboratory, New York, 1986. 16. DE VILLIERS,E.-M., eta/., Lancet 2, 703-706 (1987). 17. SCHNEIDER,A., eta/., lnt. J. Cancer40,

198-201

(1987).

771

(1987). 18. SAIKI, R. J., eta/., Science 239,487-491 (1986). 19. ZUR HAUSEN, H., Lancet 2,489-491 20. BARBACID,M., In “Important Advances in Oncology,” (V. DeVita, S. Hellman, and S. Rosenberg, Eds.) p. 3 Lippincott, Philadelphia, 1985. 21. RHIM, J. S., eta/., lnt. J. Cancer 15, 791-795 (1985). 22. SHIH, C., and WEINBERG,R. A., Ce//29, 161-169 (1982). 23. RHIM, J. S.. Proc. Sot. Expt. B/o/. Med. 174,217-223 (1983). 24. RIOU. G. F., et al., In “Papillomaviruses: Molecular and Clinical Aspects” p. 47. (P. M. Howley and T. R. Broker, Eds.), A. R. Liss, New York, 1985. 25. MATLASHEWSKI,G., eta/., EMBOJ. 6,1741-1746 (1987). 26. CROOK, T., et al., Proc. Nat/. Acad. Sci. USA 85, 8820-8824 (1988). 27. PATER, M. M., et a/., Nature (London) 335, 832-835 (1988). 28. MATLASHEWSKI,G., eta/., lnt. 1. Cancer42, 232-238 (1988). (1989). 29. DIPAOLO, 1. A., et al., Oncogene 4,395-399 30. CHAN, W.-K., KLOCK. G., BERNARD,H.-U.,/. L&o/. 63,3261-3269 (1989). 31. TROLL, W. /n “Fundamentals in Cancer Prevention” p. 41. (P. N. Magee, Ed.) Univ. Park Press/Tokyo Univ. of Tokyo Press. Baltimore, Maryland, 1976.