Progression of human papillomavirus type 18 ... - Europe PMC

3 downloads 272 Views 2MB Size Report
Oct 8, 1990 - system, FEP-1811 cells at passages 12 and 32 exhibited features typical of .... strains were designated S1, S2, and S3, respectively. Each of.
Proc. Nati. Acad. Sci. USA

Vol. 88, pp. 570-574, January 1991 Cell Biology

Progression of human papillomavirus type 18-immortalized human keratinocytes to a malignant phenotype (human ceil transformatlon/squamous cell carcinoma)

PETER J. HURLIN*, PRITINDER KAUR, PATRICIA P. SMITH, NURIA PEREZ-REYES, REBECCA A. BLANTON, AND JAMES K. MCDOUGALL Fred Hutchinson Cancer Research Center, 1124 Columbia Street, Seattle, WA 98104

Communicated by Earl P. Benditt, October 8, 1990

induced in rodent cells by expression of transfected HPVs or their transforming genes have varied but include cell immortalization (4, 6) as well as transformation to tumorigenicity (11). In contrast, when transfected into human epithelial cells, the transforming potential of HPV-16 or HPV-18 sequences appears limited to causing cell immortalization (1221) and associated alterations in differentiation (19-21). There have been no reports of tumorigenic transformation of human epithelial cells after their transfection with HPV sequences alone, and the only mechanism described for the conversion of HPV-immortalized human epithelial cells in culture to tumorigenicity has been by transfection of an HRAS oncogene (22, 23). The observation that HPVs associated with anogenital cancers are capable of increasing the life-span of human epithelial cells or immortalizing them in culture suggests that a similar biological consequence occurs with some frequency in cells infected in vivo. It has been hypothesized that some epithelial cell lesions caused by HPV infection in vivo progress to malignancy over a period of many years (24). This hypothesis is therefore consistent with a model in which progression to malignancy ofcells infected with certain HPVs is related to an increase in the number of cell doublings caused by such infection and the concomitant increase in the probability that secondary events, necessary for malignant transformation, will occur. Furthermore, if epithelial cells infected by the "oncogenic" HPVs are more likely to progress to malignancy in vivo, then epithelial cells immortalized by such HPVs in culture would be expected to give rise to tumorigenic cells when passaged extensively in vitro. Therefore, we propagated HPV-18 immortalized cell lines continuously in culture for more than 150 doublings and assessed the ability of such cells to produce tumors in athymic mice. We report here results from one of these lines, designated FEP-1811, which acquired the ability to form invasive squamous cell carcinomas in athymic mice during their propagation in culture. FEP-1811 cell strains isolated at different passage levels expressed distinct transformed properties, suggesting that progressive change requiring multiple events was involved in development of the malignant cells.

We have developed a model system for proABSTRACT gression ofhuman epithelial cells to malignancy, using a human papillomavirus type 18 (HPV-18)-immortalized human keratinocyte cell line. Cells of cell line FEP-1811 were nontumorigenic in athymic mice through at least 12 passages in culture, but after 32 passages were weakly tumorigenic, producing tumors that regressed. After 62 passages they produced invasive squamous cell carcinomas that grew progressively. The progression to malignancy was associated with an increase in the efficiency of forming colonies in soft agar and with altered differentiation properties. In an organotypic culture system, FEP-1811 cells at passages 12 and 32 exhibited features typical of premalignant intraepithelial neoplasia in vivo, and cells at passage 68 exhibited features consistent with squamous cell carcinomas. No change in copy number of the transfected HPV-18 genome or in the level of expression of the viral transforming gene products E6 and E7 was detected between tumorigenic and nontumorigenic cells. Cytogenetic analysis of cells at early, middle, and late passage levels and cells cultured from tumors revealed that several chromosomal abnormalities segregated with the tumorigenic cell populations.

Human papillomaviruses (HPVs) infect epithelial cells at a variety of anatomical sites. Infection by most of the -60 HPV types identified is associated with benign lesions. However, a subset of HPVs, including HPV types 16, 18, 31, and 33 (HPV-16, -18, -31, and -33), are frequently detected in tumors of the anogenital track and are suspected etiological agents for such tumors (1). Evidence showing that these same HPV types are also present in histologically normal cells (2) and evidence from epidemiological studies (1) suggest that events additional to HPV infection are necessary for progression to malignant disease. Human epithelial cells, the in vivo target of HPV infections, can be cultured readily in vitro, but no cell culture system has been developed that permits HPV replication. Therefore, models for a role that HPVs may have in the etiology of malignancies that contain them have relied primarily on the results of transfection studies. From such studies, it is clear that the HPVs frequently detected in anogenital carcinomas, such as HPV-16 and -18, exhibit transforming functions in a variety of rodent cell types. HPV types associated more with benign epithelial lesions, such as HPV-6 and -11, have a reduced or no transformation potential (3). Transfection studies of subgenomic fragments of HPV-16 and -18 have identified the E6 and E7 genes as the principal transforming genes (3-8). In corroboration of these results are studies showing that the E6 and E7 proteins interact with the cellular p53 (9) and retinoblastoma (10) tumor-suppressor gene products, respectively. The transformed phenotypes

MATERIALS AND METHODS Cells and Tissue Culture. Primary human keratinocytes were cultured from neonatal penile foreskins and grown in keratinocyte growth medium (KGM) (Clonetics, San Diego). Human epithelial cells immortalized after transfection with HPV-16 and -18 (12, 16, 18) were maintained in KGM. HeLa cells were cultured in Dulbecco's modified Eagle's medium (GIBCO) containing 10%o (vol/vol) bovine calf serum (Hy-

Clone). The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. ยง1734 solely to indicate this fact.

Abbreviations: HPV, human papillomavirus; HPV-16, -18, -31, and -33, HPV types 16, 18, 31, and 33 (etc.); SV40, simian virus 40.

570

Cell

Biology: Hurlin et al.

Organotypic Culture. Organotypic "raft" cultures were prepared and processed as described by Kopan et al. (25) with the exception that human foreskin fibroblasts were used in the collagen matrix instead of mouse 3T3 J2 cells. Cultures were fed daily and harvested after 10 days of growth at the air/liquid interface. For harvesting, cultures were formaldehyde-fixed, paraffin-embedded, sectioned, and hematoxylin/ eosin-stained. Organotypic culture experiments were repeated at least twice. Assay for Anchorage Independence. Exponentially growing cells were diluted to 1000 cells per ml in KGM containing 0.33% SeaPlaque agar (FMC), and 2 ml were plated on 60-mm dishes containing 4 ml of 2% solidified bottom agar in KGM. Cells were fed twice weekly with KGM. After 2 weeks, colonies with a diameter of >60 Aum were counted microscopically. Tumorigenicity. Exponentially growing cells (5 x 106) were injected subcutaneously into 4- to 6-week-old BALB/c athymic mice that had been irradiated with 300 rad from a 'Co source 24 hr previously. DNA and RNA Analysis. Genomic DNA was analyzed for the presence of HPV-18 sequences by the method of Southern (26). RNA blot-hybridization (Northern) analysis was carried out as described (26). Radioimnunoprecipitation of the E6 and E7 Gene Products. Cellular proteins were labeled with 500 ,uCi (1 ,Ci = 37 kBq) of [35S]methionine (Translabel, ICN) and 500 ,Ci of [35S]cysteine (Amersham) in KGM containing no methionine or cysteine. Cell lysates were immunoprecipitated by using rabbit polyclonal sera raised against bacterial fusion proteins containing the HPV-18 E6-encoded polypeptide (a gift from Denise A. Galloway and Steve Jenison) or E7-encoded polypeptide as previously described (27). Proteins were separated on SDS/17.5% polyacrylamide gels and detected by fluorography. Cytogenetic Analysis. Chromosome preparations were performed in situ (28). Chromosome modal number and chromosomal aberrations were determined from at least 40 G-banded metaphases. RESULTS The FEP-1811 cell line used in this study acquired an indefinite life-span in culture (i.e., were immortalized) after transfection with a vector plasmid containing the entire HPV-18 genome interrupted in the El open reading frame (16). The cell line was propagated continuously in culture after immortalization, and populations were frozen for storage at various passage levels. Cell strains were established from FEP-1811 cells originally frozen at passage levels 12, 32, and 59 (cells were passaged at a split ratio of 1:4; thus each passage represents approximately two population doublings). These strains were designated S1, S2, and S3, respectively. Each of these strains was characterized as described below within 10 passages from the passage level at which they were originally frozen. Morphology and Growth Properties. The S1, S2, and S3 FEP-1811 cell populations exhibited morphological features distinct from one another. A morphologically heterogeneous population of cells was observed with S1 cells (Fig. 1C) that included small cuboidal cells and some large multinucleated cells similar to those observed in senescing populations of primary keratinocytes. S1 cells exhibited a poor cloning efficiency on plastic and were not capable of producing colonies in soft agar (Table 1). S2 cells consisted of small refractile cells (Fig. 1E) that proliferated rapidly and grew in a disorganized manner. These cells had a high cloning efficiency on plastic but formed colonies in soft agar only at a 0.25% efficiency (Table 1), and the colonies formed in soft agar could not be reestablished in culture. S3 cells (Fig. IG)

Proc. Natl. Acad. Sci. USA 88 (1991)

*

571

C

F

H

FIG. 1. Morphology and differentiation profiles of FEP-1811 cells. Normal human keratinocytes (A and B), FEP-1811 S1 cells (C and D), S2 cells (E and F), and S3 cells (G and H) were grown in monolayer culture (A, C, E, and G) and on collagen rafts (B, D, F, and G). Cells on collagen rafts were grown at the air/liquid interface for 10 days and paraffin-embedded preparations were stained with hematoxylin and eosin. [x74 (A, C, E, and G) and x93 (B, D, F, and G).]

exhibited a morphology similar to that of S2 cells but grew to higher density. S3 cells formed colonies at nearly the same efficiency on plastic as passage S2 cells but formed colonies in soft agar at a 4- to 5-fold greater efficiency (Table 1), and these colonies could be reestablished in culture. Differentiation Potential and Tumorigenicity. Primary human foreskin keratinocytes and FEP-1811 S1, S2, and S3 cells were cultured for 10 days at the air/liquid interface on collagen rafts. Use of this system results in the stratification of primary epithelial cells and accompanying morphological changes associated with their differentiation in vivo (25). As a

Table 1. Growth properties and tumorigenicity of FEP-1811 cells FEP-1811 Cloning efficiency Tumors that grow Tumor cells in agar,* % incidences progressively Si 0.00 (0.00) 0/10 0.25 (0.16) No S2 3/5 S2T1 ND 5/5 No S3 1.03 (0.86) Yes 12/12 S3T1 ND Yes 6/6 NHK 0.00 (0.00) ND, not done. *Number in parentheses is from a duplicate experiment. tNumber of animals with tumors per number of animals injected with 5 x 106 cells.

572

Cell Biology: Hurfin et al.

expected, normal human keratinocytes derived from foreskin formed an epidermal sheet (Fig. 1B) with features comparable to those of normal squamous epithelium. In contrast, FEP1811 S1 cells exhibited abnormal differentiation properties with parabasal layer crowding and some small nucleated cells extending up to the cornified layer (Fig. 1D). The differentiation profile of S1 cells is similar to that observed with mild-to-moderate dysplasia in vivo. S1 cells were tested for the ability to form tumors in athymic mice by injecting 5 x 106 cells s.c. No tumors or evidence of any in vivo growth were apparent with these cells over a period of >9 mo (Table 1). S2 cells exhibited a similar stratification profile compared to S1 cells but with more crowding at the parabasal layer and more nucleated cells extending up to and within the cornified layer (Fig. 1F). S2 cells exhibited features similar to those observed in moderate to severe displasia in vivo. When tested for tumor-forming ability, S2 cells formed tumors in three of five mice injected (Table 1). The tumors formed required =6 wk to reach a diameter of 0.4 cm. Two of the three tumors regressed, leaving only residual necrotic tumor tissue by 8 weeks, and the other tumor was removed for histological examination. The biopsied tumor was diagnosed as a poorly differentiated squamous cell carcinoma. Cells were also cultured from this tumor (designated S2Ti), and their tumorigenic properties were retested. S2T1 cells produced tumors in each of five mice injected (Table 1). These tumors required only 3 weeks to reach a diameter of 0.4 cm, but like tumors produced by S2 cells, these tumors also regressed. S3 cells completely lacked the ability to differentiate to form a cornified layer on qollagen rafts (Fig. 1H). The presence of small nucleated cells throughout the epidermal sheet formed by S3 cells is consistent with the histological appearance of severe dysplasias in vivo and squamous cell carcinoma. When tested for tumorigenicity, S3 cells produced tumors in 12 of 12 mice that grew progressively (Table 1), eventually forming very large tumors that caused the mice to become moribund. In comparison with the S2 strains, these tumors reached a diameter of 0.4 cm after 4 weeks. The tumors stained positive for high-mass cytokeratins and were diagnosed as poorly differentiated squamous cell carcinomas (Fig. 2). Cells were cultured from tumors produced by S3 cells (designated S3T1). These cells were retested for tumorigenicity and produced histologically identical tumors that grew at a faster rate, requiring only 3 weeks to reach a diameter of 0.4 cm. These tumors grew progressively into large invasive tumors (Table 1). Karyotype analysis of cells derived from biopsied tumors formed by FEP-1811 S2 and S3 cells and Southern blot analysis of cells derived from tumors formed by S3 cells revealed that these cells were human and derived from the FEP-1811 cell line (see below). Presence and Expression of HPV-18 Sequences in FEP-1811 Cells. Genomic DNA isolated from S1, S2, and S3 cell populations and from cells cultured from tumors produced by S3 cells was analyzed by Southern blot hybridization for the presence of the transfected HPV-18 sequences. By using an HPV-18 genomic probe, a single fragment of 7.9 kilobases (kb) in EcoRI-digested DNA (EcoRI releases the entire 7.9-kb HPV-18 genome from the pBR322 vector used for transfection) was detected in each cell strain (data not shown). When hybridized to the same DNAs undigested, a single band migrating with high-mass DNA was observed, indicating that the hybridizing HPV sequences were integrated into the cellular genome (data not shown). The RNA expression patterns of HPV-18 sequences in FEP-1811 S1, S2, and S3 cells were analyzed by Northern analysis with an HPV-18 genomic probe. The pattern of HPV-18 gene expression observed with these cell strains was comparable in both transcript size and level of expression (data not shown).

Proc. Natl. Acad. Sci. USA 88 (1991)

FIG. 2. A poorly differentiated squamous cell carcinoma formed subcutaneously in athymic mice by FEP-1811 S3 cells. (A) The subcutaneous location of the tumor (arrow). (B) Tumor invasion (arrow) of mouse skeletal muscle. (C) Immunohistochemical staining of tumor cells (arrow) with an antibody (34BE12; Enzo Diagnostics) reactive against high-mass cytokeratins. [x 16 (A) and x32 (B and C).]

Expression of the E6 and E7 gene products in S1, S2, and S3 cells was determined by radioimmunoprecipitation analysis. Cells from each of these cell strains expressed comparable amounts of the HPV-18 E7 protein (Fig. 3A). HPV-18 E7 expression in HeLa cells was used as a positive control. Expression of HPV-18 E6 protein was more difficult to determine in S1, S2, and S3 cells, as well as in HeLa cells. However, a protein migrating at 16-17 kDa, the expected size for the HPV-18 E6 protein (29), was present in FEP-1811 cells but not in normal human keratinocytes (Fig. 3B). Karyotype Analysis of FEP-1811 Cells. We previously showed that the FEP-1811 cell populations at passage 36 (S2) and 59 (S3) contained a number of chromosome aberrations (30). We have extended that study here to include S1 cells and FEP-1811 cells derived from tumors formed by S2 and S3 cells. A representative karyotype of cells derived from a tumor produced by S3 cells is shown in Fig. 4. A consistent feature of FEP-1811 cells at different passage levels has been the maintenance of a nearly diploid karyotype that included aberrations in chromosomes 3 [der(3)], 6 [der(6)], 7 [der(7)], 13 [der(13)], and 22 [der(22)] and monosomy 13. In addition to these changes common to cells at each passage level

Biology: Hurlin et al.

Cell

A NHK Hela i

M P

i

p

Proc. Natl. Acad. Sci. USA 88 (1991)

mosome 21 and the Y chromosome and the 5q+ aberration. The 5q+ aberration appeared to involve an interstitial deletion in the q arm and addition of an undetermined chromosome fragment(s).

FEP 1811 S2 S3 P i i

DISCUSSION HPV-16 and -18 are frequently found in anogenital carcinomas, and it is well established that immortalized human epithelial cells can be generated after transfection of primary cells with these HPV types. In this report we show that when an HPV-18-immortalized cell line was continuously propagated in culture, tumorigenic populations arose. FEP-1811 cells characterized at an intermediate passage level (S2 cells) formed tumors that subsequently regressed, and cells characterized -30 passages later (S3 cells) formed tumors that grew progressively. The tumors produced by both the S2 and S3 populations were histologically similar and diagnosed as poorly differentiated squamous cell carcinomas. Cells characterized at an early passage level (S1 cells) were not tumor-

2-

1-4 18-

18-

do

--

14-

B

291

18-

573

o

14-

igenic. FIG. 3. Expression of HPV-18 E7 (A) and E6 (B) proteins in FEP-1811 cells. Cultures of primary human keratinocytes (NHK), HeLa cells, and FEP-1811 S1, S2, and S3 cells were radiolabeled with [35S]methionine and [35S]cysteine, and cell lysates were immunoprecipitated with rabbit polyclonal antiserum against HPV-18 E7 (lanes i in A) or E6 (lanes i in B), or preimmune serum (lanes p). Immunoprecipitates were analyzed on SDS/15% polyacrylamide gels by fluorography. Arrows indicate the location of the E7 and E6 proteins. The numbers under lane M are for molecular size standards (in kDa).

analyzed, the tumorigenic populations (S2 and S3) contained three additional changes. These included loss of a chromosome 21 and the Y chromosome, and an aberration in chromosome 5 (5q+). These changes appeared to segregate specifically with tumorigenic populations. This is because none of the chromosomal abnormalities found in the tumorigenic populations (-21, -Y, and 5q+) were observed in 50 metaphases prepared from S1 cells, but 50%6 of metaphases prepared from S2 cells and 100o of metaphases prepared from cells cultured from a tumor produced by S2 cells (S2T1) contained these three specific changes. Furthermore, 100o of metaphases prepared from S3 cells showed loss of chro-

Itat11I 4 2

1

6

7

5

3

B--

A

8

9

10

11

12

16

17

18

C

a

-

I

13

.t 14

15

E

D kt 19

20

-F

, -

21

A

-

22

x

y

G--

FIG. 4. Karyotype of cells derived from tumors produced by FEP-1811 S3 cells. This nearly diploid karyotype was identical to that of S1 and S2 cells with the exception of monosomy 21, monosomy Y, and an aberration in one copy of chromosome 5 (5q+).

Our data do not rule out the possibility that the transfected HPV-18 sequences may have contributed directly to the development of tumorigenicity. The presence of chromosomal aneuploidy and chromosome aberrations in each of the eight immortal cell lines generated in this laboratory after transfection of either HPV-16 or -18 sequences (30) suggest that such transformation to tumorigenicity might have been the result of chromosomal instability induced through the action of the principal transforming proteins of these viruses, E6 and E7. If such a mechanism were operating, it would be expected that one would see karyotypic abnormalities arise continuously as these cell lines were passaged in culture. For each of the cell lines generated in this laboratory, including the FEP-1811 cell line, the abnormal karyotype present at early passage levels persisted with few changes after extensive passaging (30). Furthermore, if the E6 and E7 proteins were involved in causing FEP-1811 cells to become tumorigenic, then transformation to tumorigenicity would be expected in the other HPV-immortalized cell lines. None of these other cell lines generated in this laboratory (14, 16, 17) and no HPV-immortalized human epithelial cell lines generated in other laboratories (12, 13, 15, 18-21) have proved to be tumorigenic. Therefore, our results are more consistent with a model in which the progression to tumorigenicity observed with the FEP-1811 cell line resulted from alterations in the cellular genome that occurred in a spontaneous manner independent of E6 and E7 expression. Other than HPVs, the only other agent capable of immortalizing human epithelial cells in culture is simian virus 40 (SV40) (31). Similar to studies with HPV-immortalized human epithelial cells (22, 23), SV40-immortalized cells have been transformed into tumorigenic cells following the introduction of a RAS oncogene (32). In a study by Brown and Gallimore (33), it was shown that SV40-immortalized cells progressed to a malignant phenotype in an apparently multistep manner without the transfection of additional oncogenes. It was not determined whether the malignant cells that arose did so as a result of a mutation in a RAS gene. Similar to this latter study, our results suggest that the malignant cells in the FEP-1811 cell line arose from their immortal, nontumorigenic parental cells in a multistep fashion. This is based primarily on the distinct phenotypic properties observed between FEP-1811 cells at different passage levels (Table 1) and is exemplified by their tumorigenic properties. The S1 population was not tumorigenic, but both the S2 and S3 populations formed tumors that exhibited the histological appearance of poorly differentiated squamous cell carcinoma, but only tumors produced by the S3 population grew progressively, an important criterion for malignancy. The S2

574

Cell Biology: Hurlin et al.

population was weakly tumorigenic, producing tumors that regressed. This was also true for tumors formed by cells (designated S2T1) that were cultured from a tumor formed by S2 cells (Table 1). Tumors formed by S2T1 cells required only 4 weeks to reach a diameter of 0.4 cm compared with =7 weeks for tumors formed by the S2 cells but subsequently regressed. This indicated that in vivo selection occurred for cells within the S2 population that formed tumors at a rate comparable to that of the S3 population but that lacked the malignant phenotype of later passage S3 cells. It is likely, therefore, that the cells that produced tumors that regressed represented an intermediate phenotype and that these cells evolved out oftheir immortal, nontumorigenic parental cell population by some genetic event(s). It is also likely that the cells exhibiting the intermediate phenotype then progressed to a fully malignant phenotype as a result of subsequent genetic event(s). One candidate for an event involved in causing FEP-1811 cells to produce tumors that subsequently regress is the chromosome 5 aberration (5q+) identified. This aberration appeared to involve an interstitial deletion in the q arm of chromosome 5 and addition of an undetermined chromosome fragment(s). This aberration, along with loss of chromosome 21 and the Y chromosome, segregated with tumorigenic populations. It is not known whether these chromosome changes played a role in the conversion of FEP-1811 cells to malignancy. Of interest, however, is the proximity of the 5q aberration observed in tumorigenic FEP-1811 cells to the adenomatous polyposis coli locus (34). The demonstrated potency of RAS oncogenes in converting HPV- and SV40-immortalized human epithelial cells to malignancy (22, 23, 32) suggested that one of these genes may have been activated at some stage of the progression of FEP-1811 cells to malignancy. However, our initial studies examining RAS genes for activating mutations have revealed that no codon 12 mutations in HRAS, KRAS, or NRAS exist in any of the FEP-1811 cell populations (data not shown). We have not yet tested for the presence of other activating RAS mutations. We also have found no changes in the level of expression of the cellular MYC protein in the FEP-1811 cell populations (data not shown). Increased levels of MYC protein have been observed in some cervical carcinoma samples (35). The progression to tumorigenicity of FEP-1811 cells provides a model system from which the genetic, cytogenetic, and phenotypic alterations that potentially mark the progression to malignancy of human epithelial cells in vivo may be dissected. However, it is likely that the events responsible for progression to tumorigenicity and the sequence of occurrence of these events in our in vitro model system represent only one of many routes to tumorigenicity. Therefore, it is important to develop additional model systems to analyze and compare phenotypes that arise and to determine how such phenotypes may relate to in vivo cellular alterations associated with increased risk of epithelial malignancies.

Proc. Natl. Acad. Sci. USA 88 (1991)

3. 4.

5. 6.

7.

8. 9. 10. 11.

12. 13. 14. 15.

16. 17. 18.

19. 20. 21. 22.

23. 24. 25.

26. 27. 28.

29. 30. 31. 32.

We thank Dr. Jerry Radich for assistance in analysis for mutations in RAS genes and Drs. Denise Galloway and Steven Jenison for critical reading of the manuscript. This research was supported by a Public Health Service grant from the National Cancer Institute (CA-42792).

33.

1. zur Hausen, H. (1983) Prog. Med. Virol. 32, 15-21. 2. de Villiers, E.-M., Wagner, D., Schneider, A., Wesch, H.,

35.

34.

Miklew, H., Wahrendorf, J., zur Hausen, H. & Papendick, U. (1987) Lancet iU, 703-706. Storey, A., Pim, D., Murray, A., Osborn, K., Banks, L. & Crawford, L. (1988) EMBO J. 7, 1815-1820. Bedell, M. A., Jones, K. H., Grossman, S. R. & Laimins, L. A. (1989) J. Virol. 63, 1247-1255. Kanda, T., Furuno, A. & Yoshiike, K. (1988) J. Virol. 62, 610-613. Kanda, T., Watanabe, S. & Yoshiike, K. (1988) Virology 165, 321-325. Matlashewski, G., Schneider, J., Banks, L., Jones, L., Murray, N. & Crawford, L. (1987) EMBO J. 6, 1741-1746. Yutsudo, M., Okamoto, Y. & Hakura, A. (1988) Virology 166, 594-597. Werness, B. A., Levine, A. J. & Howley, P. M. (1990) Science 248, 76-79. Dyson, N., Howley, P. M., Munger, K. & Harlow, E. (1989) Science 243, 934-937. Yasumoto, S., Burkhardt, A. L., Doniger, J. & DiPaolo, J. A. (1986) J. Virol. 57, 572-577. Durst, M., Dzarlieva-Petrusevska, R. T., Boukamp, P., Fusenig, N. E. & Gissmann, L. (1987) Oncogene 1, 251-256. Pirisi, L., Yasumoto, S., Feller, M., Doniger, J. & DiPaolo, J. A. (1987) J. Virol. 61, 1061-1066. Kaur, P. & McDougall, J. K. (1988) J. Virol. 62, 1917-1924. Woodworth, C. D., Doniger, J. & DiPaolo, J. A. (1989) J. Virol. 63, 159-164. Kaur, P. & McDougall, J. K. (1989) Virology 173, 302-310. Kaur, P., McDougall, J. K. & Cone, R. (1989) J. Gen. Virol. 70, 1261-1266. Munger, K., Phelps, W. C., Bubb, V., Howley, P. M. & Schlegel, R. (1989) J. Virol. 63, 4417-4421. McCance, D. J., Kopan, R., Fuchs, E. & Laimins, L. A. (1988) Proc. Natl. Acad. Sci. USA 85, 7169-7173. Hudson, J. B., Bedell, M. A., McCance, D. J. & Laimins, L. A. (1990) J. Virol. 64, 519-526. Pecoraro, G., Morgan, D. & Defendi, V. (1989) Proc. Natl. Acad. Sci. USA 86, 563-567. DiPaolo, J. A., Woodworth, C. D., Popescu, N. C., Notario, V. & Doniger, J. (1989) Oncogene 4, 395-399. Durst, M., Gallahan, D., Jay, G. & Rhim, J. S. (1989) Virology 173, 761-771. zur Hausen, H. (1986) Lancet Ui, 489-491. Kopan, R., Traska, G. & Fuchs, E. (1987) J. Cell Biol. 105, 427-440. Maniatis, T., Fritsch, E. F. & Sambrook, J. (1982) Molecular Cloning:A Laboratory Manual (Cold Spring Harbor Lab., Cold Spring Harbor, NY). Jenison, J. A., Yu, X.-P., Valentine, J. M., Koutsky, L. A., Christensen, A. E., Beckman, A. M. & Galloway, D. A. (1990) J. Infect. Dis. 162, 60-69. Martin, G. M., Smith, A. C., Ketterer, D. J., Ogburn, C. E. & Disteche, C. M. (1985) Isr. J. Med. Sci. 21, 296-304. Banks, L., Spence, P., Androphy, N., Hubbert, G., Matlashewski, G., Murray, A. & Crawford, L. (1987) J. Gen. Virol. 68, 1351-1359. Smith, P. P., Bryant, E. M., Kaur, P. & McDougall, J. K. (1989) Int. J. Cancer 44, 1124-1131. Steinberg, M. L. & Defendi, V. (1979) Proc. Natl. Acad. Sci. USA 76, 801-805. Rhim, J. S., Jay, G., Arnstein, P., Price, F. M., Sanford, K. K. & Aaronson, S. A. (1985) Science 227, 1250-1252. Brown, K. W. & Gallimore, P. H. (1987) Br. J. Cancer 56, 545-554. Bodmer, W. F., Bailey, C. J., Bodmer, J., Bussey, H. J. R., Ellis, A., Gorman, P., Lucibello, F. C., Murday, V. A., Rider, S. H., Scambler, P., Sheer, D., Soloman, E. & Spurr, N. K. (1987) Nature (London) 328, 614-616. Riou, G., Barrios, M., Sheng, Z., Duvillard, P. & Lhomme, C. (1988) Oncogene 3, 329-333.