Vittorio Defendi

Wild-type (wt) p53 acts as a tumor suppressor, while certain mutant type (mt) p53 may exhibit 'oncogenic' function. We have recently demonstrated that human papillomavirus type 18 (HPV-18) E6 can partially overcome the growth-suppressive effects of wt p53, but it remains unclear what role p53 plays in cervical carcinogenesis. In this report, we have examined nine HPV-immortalized human cervical epithelial cell lines and 13 HPV-positive and two HPV-negative primary cervical cancers for p53 mutations by polymerase chain reaction--single-strand conformation polymorphism (PCR-SSCP). None of them contained p53 mutations in exons 5-9 where most p53 mutations in human tumors have been found. The entire p53-coding region of the two HPV-negative cervical cancers was sequenced and no mutations were noted. In order to examine the effects of wt p53 and mt p53 on HPV-immortalized human cells, we transfected HPV-immortalized cell lines with wt p53 and a mt p53 (mtp53Val-135). The results indicate that HPV-immortalized cells cannot tolerate large amounts of exogenous wt p53, while mt p53Val-135 can enhance transformation of these cells. The results support the notion that inactivation of wt p53 by E6 may be important for HPV-associated transformation and also suggests that mt p53 can act as an oncogene in HPV-immoralized human cells.

The authors have established an in vitro model system which demonstrates the progression of the transformed phenotypes of human cervical epithelial cells transfected with human papillomavirus (HPV) type 16 and 18 DNAs. Both viral DNAs exhibit immortalizing potential; however, only HPV 18-immortalized cell lines progress to exhibit anchorage-independent growth and, in a limited number of cases, tumorigenesis. In this paper, the authors have examined the genetic basis for this in vitro progression step by step, including immortalization, anchorage-independent growth, and tumorigenicity of the HPV-transfected human cervical epithelial cells by cell fusion. The results suggest that (a) all three transformed phenotypes, i.e., immortalization, anchorage-independent growth, and tumorigenesis, in this in vitro cervical carcinogenesis model are a result of recessive changes in genes or processes involved; (b) inactivation of p53 and retinoblastoma protein is not sufficient for immortalization of human cervical epithelial cells; (c) HPV expression per se does not account for immortalization of human cervical epithelial cells; (d) immortalization of human cervical epithelial cells initiated by HPV can occur through different processes, although one of them is the most preferred; and (e) probably only one group of recessive genes appears to be involved in the loss of anchorage-dependent growth for HPV-immortalized human cervical cells.

Cervical carcinoma develops through a progressive spectrum of premalignant intraepithelial lesions (CIN I-III), the majority of which are associated with human papillomavirus (HPV) types 16 and 18. We established HPV16 and HPV18 immortalized human cervical epithelial cell lines and used them as a model to investigate the genesis and progression of cervical malignancy. The cell lines when cultured in vitro in a system mimicking their in vivo environment exhibit cytologic atypia and a variety of defects in morphologic differentiation at early passage compared to their normal counterparts. With increased passage, these alterations progress to more severe grades, histologically similar to CIN III; however only a limited number of the cell lines are tumorigenic, mimicking the epidemiologic evidence on the rate of conversion from premalignant to invasive carcinoma. The observed changes are not associated with alterations of viral DNA integration or expression and may reflect specific cellular events or changes in virus-host interactions associated with malignant progression.

The human papillomaviruses (HPVs) are associated with specific benign and malignant lesions of the skin and mucosal epithelia. Cloned viral DNAs from HPV types 6b, 16, and 18 associated with different pathological manifestations of genital neoplasia in vivo were introduced into primary human cervical epithelial cells by electroporation. Cells transfected with HPV16 or HPV18 DNA acquired indefinite lifespans, distinct morphological alterations, and anchorage-independent growth (HPV18), and contain integrated transcriptionally active viral genomes. HPV6b or plasmid electroporated cells senesced at low passage. The alterations in growth and differentiation of the cells appear to reflect the progressive oncogenic processes that result in cervical carcinoma in vivo.

Previous reports from this laboratory indicate that cultured simian virus 40 (SV40)-transformed human keratinocytes express keratin proteins characteristic of simple epithelia that are not found in their untransformed counterparts. In this study we show by in vitro translation and RNA transfer blot analysis that the altered keratin synthesis reflects changes in the abundance of specific keratin mRNAs. SV40-transformed keratinocytes have a reduced abundance of transcripts for 58-, 56-, 52-, 50-, 48-, and 46-kDa keratin species, compared with uninfected cultured keratinocytes, but express significant levels of transcripts for 52-, 45-, and 40-kDa keratins, typical of simple epithelia. The SV40-transformed cells also express mRNA for a 48-kDa keratin that is unique to SV40-transformed keratinocytes. Analysis of the keratin genome with keratin-specific cDNAs as probes indicates that the changes in keratin transcription are not correlated with gross rearrangements of the keratin genome. These results suggest that analysis of viral transformation of cultured keratinocytes affords a novel approach to study mechanisms regulating keratin gene expression.

Transformation of human epidermal keratinocytes by the oncogenic virus SV40 is a stage-specific process in which normal patterns of differentiation are progressively altered over time following infection. Within the context of this scheme, we examined the keratins produced by the infected cells. Immunofluorescence studies indicated that viral infection led to the formation of variant cells visibly lacking the normal keratin cytoskeleton after about 10–15 serial passages (60–90 cell generations) post infection. Analyses of variant cell formation in clonal populations grown on palladium islands revealed that the variants were derived within 2–3 cell divisions from cells containing an apparently normal keratin cytoskeleton, but that variant formation depended upon cell density. Immunoprecipitation of ³⁵S-methionine labelled keratins from the infected keratinocytes revealed a gradual loss of the normal 46, 50, 56 and 58Kd keratin species over a period of many months after infection. The loss of the normal keratins was accompanied by the appearance of at least two species in the 48–52Kd size range not present in uninfected cells and the enhancement of a third, 40Kd, protein quite early after infection. Analysis of the altered keratin patterns on two-dimensional acrylamide gels using either isoelectric focusing (IEF) or non-equilibrium pH gradient electrophoresis (NEPHG) along the first dimension showed that the infected cells came more transformed with serial passage including at least five isoelectric forms not seen in uninfected cells. Translation of poly A⁺ RNAs from the infected cells indicated that the altered keratin synthesis probably reflects changes in the translatable mRNA pool.

The synthesis and distribution of fibronectin were studied in human epidermal keratinocytes infected by SV40, a system in which the acquisition of transformed properties occurs in a sequential and progressive manner. Immunofluorescence studies showed that cultured uninfected keratinocytes do not exhibit fibronectin on the superficial cell surface, but that virus-infected cells come to display superficial fibronectin-containing cables in a density-dependent manner after a certain point in the transformation process. Organized arrays of fibronectin-containing fibrils associated with the cell-substrate attachment complex were seen in uninfected keratinocytes and in virus-infected cells at all stages of the transformation process. Studies of fibronectin synthesis using metabolic labeling of cell proteins with 35S methionine showed that viral infection caused a striking increase in overall fibronectin synthesis, although with a much higher proportion of newly synthesized fibronectin being secreted into the cell culture medium than in the case of the uninfected cells.