Abridged from the Dermatology Progress in Foundation ... · Abridged from the Dermatology Foundation Progress in Dermatology Editor: Alan N. Moshell. MD. Molecular Biology of Human

Abridged from the

Dermatology Foundation

Progress in Dermatology

Editor: Alan N. Moshell . MD.

Molecular Biology of Human Papillomavirus Infection and Oncogenesis

Elliot J. Androphy Department of Dermatology, New England Medical Center and Tufts University School of Medicine, Boston, Massachusetts. U.S.A.

P apillomaviruses infect stratifying epithelial cells and cause benign tumors or papillomas. In certain circum­stances, viral infection can lead to malignant transfor­mation, and one goal of this review is to discuss the mechanisms by which human papillomaviruses (HPVs)

cause cancer. All papillomaviruses are composed of two comple­mentary strands of DNA with about 8000 nucleotides on each of the two strands. Their genome is present as a covalently closed circle (plasmid or episome) within the nucleus of infected cells. Papillo­maviruses are defined by both the animal species they infect and their nucleic acid sequence. Whereas all papillomaviruses are highly related, the differences in their DNA sequences are used to distinguish viral genotypes by DNA-DNA hybridization. Using criteria to define how great the differences must be in a new isolate, nearly 70 HPV genotypes have been reported [1] . These DNA sequence differences result in amino acid changes among the viral encoded proteins. Divergence is also found in the viral regulatory region, including the cis-acting sites for assembly of viral and cell encoded factors that control viral gene expression and replication. These variations in coding and non coding sequences must be re­sponsible for the species specificity and pathogenicity of each viral genotype, and have been used to construct a phylogenetic tree of HPV evolution [2,3] .

Whereas there are multiple HPV genotypes, these can be catego­rized into three classes. The first set includes those predominantly found in cutaneous lesions, most commonly types 1,2,3, and 4. The second category of HPV is a large group isolated in epidermo­dysplasia verruciformis (EV) . In this disease, patients have lifelong warts, often from an early age. EV represents one of the earliest recognized examples of a human virus infection leading to malig­nancy. In EV, individuals infected with specific HPV types, most notably types 5 and 8, develop squamous cell carcinomas from their warts. However, whereas there are many lesions on an individual infected with HPV 5 or 8, only a minority proceed to malignancy.

Reprint requests to: Elliot Androphy, Department of Dermatology, Box 166, 750 Washington St., New England Medical Center, Boston, MA.

Because these cancers are usually located on sun-exposed skin, ultra­violet (UV) light is believed to be a co-factor in tumor induction.

The third class of HPV s primarily infects mucosal epithelia such as the oropharynx and anogenital region. As observed with EV, only a subset of these have a predilection for progressing to squamous cell carcinoma. HPV 6 and 11 are commonly found in genital warts and cervical papillomas, whereas HPV 16 and 18 a.re less frequent. The frequency is opposite in cervical, penile, vulvar and anal cancers. Most cervical carcinomas contain types 16, 18,31,35, etc., whereas HPV 6 and 11 are rare (estimated to be 1- 5%) [4-6]. The former set ofHPVs is called "high risk," whereas HPV 6 and 11 are termed "low risk." High-risk HPV types can infect cutaneous skin and have been identified in squamous cell carcinomas. Because specific sub­sets of HPVs are found in cutaneous squamous carcinomas in EV and in cervical cancer, this observation implies that the HPVs are causative agents for these tumors. Both the latent period (estimated to be 5 -20 years) between infection and development of cancer, and the fact that all warts do not become malignant, even with "high­risk" HPV's, indicate that infection with HPV is necessary but not sufficient for development of cancer. Under the influence of the virus life cycle, changes must occur within the cell during the pro­gression from papilloma to neoplasia. These may be consequences of viral infection (see below) or co-factors, such as ultraviolet light or smoking.

BIOLOGY OF PAPILLOMAVIRUS INFECTION

All papillomavirus genomes encode 9 - 10 open reading frames and with variable splicing have the potential to synthesize 12-15 gene products. The reading frames labeled "E," for early, represent those genes in bovine papilloma virus (BPV), which were thought to be involved in episomal replication in cultured cells. The late, or "L," genes encode the viral capsid proteins. The number after E or L refers to the size of the reading frame peptide, with 1 being the largest. These designations are not rigorous, because E4 is not essen­tial for viral DNA replication and probably serves a late function (see below). In this review, I briefly describe the events of the life cycle and how the papillomavirus gene products mediate these events. Particular emphasis is placed on the viral oncogenes E5, E6,

0022-202X/94/S07.00 Copyright © 1994 by The Society for Investigative Dermatology, Inc.

248

VOL. 103, NO. 2 AUGUST 1994

and E7 and the cellular proteins with which they interact. However, whereas these viral genes are oncogenic in cultured cells, their pri­mary function must be to assure viral replication and not direct conversion of the host cell to a malignant state. Malignant progres­sion should be viewed as the pathogenic consequences induced by these oncoproteins as they establish the viral replication life cycle.

Entry Papillomaviruses gain access to the host by direct implan­tation through physical breaks in the epithelial barrier. The specific receptor and the specific cell type that bind the virus upon entry h ave not been identified. It is assumed that virus initially infects the basal cell , although this has not been proved, and it is conceivable that pathogenic viral infection can initiate in the upper layers of the epithelium. Basal cells are assumed to be infected with one or two copies of the circular viral DNA per cell; however, a few such molecules are not detectable with current technology. The ex­trachromosomal viral DNA resident in basal cells is believed to replicate in concert with normal cell division. Presumably, the viral genome is transported, along with the daughter cells, to upper levels of the epithelium. It is in these highly differentiated strata that viral RNAs are expressed at substantial levels [7]. It is presumed that differentiation-specific events transduce a stimulatory signal for viral transcription and DNA replication. Thus papillomaviruses replicate to high numbers only in expendable, terminally differen­tiated cells that are destined to be sloughed off, and they are not lytic like other viruses. There is no evidence that papillomaviruses dis­seminate within the infected animal through viremia or another systemic route.

Exit After many copies of the circular vira l DNA are synthesized in the upper strata, these genomes are incorporated into a particle, or cap sid, which consists of the L1 and L2 proteins. The vira l capsid protects its DNA during traversal through the epithelium as the keratinocytes terminally differentiate, as well as after being shed into the environment. Importantly, papillomaviruses do not bud from the cell's plasma membrane and thus do not incorporate a membrane-derived lipid envelope, which would cause them to be sensitive to environmental stresses such as heat, soaps, or desicca­tion. Presumably the capsid binds receptors on the host cell. Having assembled these new viruses, papilloma viruses are carried along with stratifying cells to the stratum corneum, from where they are released as the epithelium undergoes its normal maturation pro­gram. Several groups have reported that high-level expression of HPV and BPV L1 and L2 genes results in their self assembly into normal-appearing capsids, which have been demonstrated to expose native epitopes [8 -12]. These virus- like particles may be of utility for serologic diagnosis of HPV infection and as vaccines.

Inunune Response All viruses probably evolve mechanisms to avoid host surveillance. For example, mutations in critical viral antigens develop at high frequency in influenza and human immu­nodeficiency virus (HIV) . Viruses such as herpes simplex and vari­cella-zoster induce a strong immune response but avoid immune surveillance through latency, i.e., these viruses can remain as DNA buried in the cell's chromosomes, only to reactivate, produce their proteins, and release viral particles at a later time. Other viruses such as measles or HIV directly kill lymphocytes, whereas others release proteins that inhibit host defenses. How papilloma viruses avoid host immune surveillance is not known. Papillomavirus-encoded proteins are present at exceedingly low levels in the lower levels of the epithelium and may therefore reduce exposure to the appropri­ate immune effector cells. In addition, the enveloped viruses (see Entry) synthesize proteins that are inserted into the cell's plasma membrane prior to budding and release, and it is these proteins that are strongly antigenic and recognized by the host immune response. Papillomaviruses do not assemble in this manner, and indeed viral particles form in the nucleus. Thus papillomaviruses have evolved a means of replication that probably reduces exposure to the immune system.

Coordination of the Viral Replication Cycle After entering the appropriate host cell, the virus must establish the proper order in

MOLECULAR BIOLOGY OF HUMAN PAPILLOMAVIRUS 249

which its own genes will be expressed. The full circuitry that con­trols papilloma virus gene expression is not known. Within the non-coding region of each virus are DNA sequences that are recog­nized by cellular transcription factors [13 -17]. The papillomavirus genotypes vary in the type, array, and position of these sites, which probably have important consequences for pathogenicity.

In addition to these binding sites for cellular factors, papillomavi­ruses encode a DNA binding protein in a gene called E2. E2 binds the inverted palindrome 5' ACCG NNNN CGGT 3' with very high affinity (~1O- 11 M) [18] , and multiple copies of this DNA palindrome are found in every papilloma virus genome [19]. Their placement often varies among different genotype groups. E2 can stimulate, and in certain instances repress, viral transcription [20-22] and is also required for viral DNA replication (see below). ~hen E2 binds its recognition site, it usually stimulates transcrip­tion from the nearby promoters in a classical "enhancer" mode.

The E2 proteins average about 400 amino acids with a mono­meric molecular weight of approximately 50 kilodaltons (KD) and have distinct functional domains [23 - 25]. The carboxy terminal 100 amino acids of all papillomavirus E2 proteins constitutes the DNA binding domain. T his small region is sufficient for sequence­specific D NA binding and dimerization, which E2 must do to bind DNA [26]. The atomic structure of the BPV E2 protein bound to its cognate D N A site has been solved and found to fo ld as an unusual B-barrel dimer with a DNA recognition helix crossing each mono­mer [27].

The amino terminal half of the E2 protein is necessary for stimu­lation of gene expression, so these amino acids must interface with the cell's transcription machinery. This region of E2 represents its transcription activation domain. When E2 binds a specific segment on the viral D NA through its DNA binding domain, its transcrip­tion activation domain recnlits the cell ular factors that lead to syn­thesis of the viralmRNA. One such factor may be Spl, although there is evidence that E2 activates transcription in concert with a variety of basal promoter factors [28,29].

The E2 DN A binding and transcription activation dom.ains share considerable homology among the papillomaviruses. Between these two regions is a variable stretch of residues that al'e not con­served, and are genera lly thought to represent a " flexibl e hinge." The E4 reading frame overlaps this E2 hinge region and the papillo­maviruses differ substantially in their E4 proteins. E4 is not essential for viral transcription, replication, or transformation in vitro and is believed to serve a role in maturation of the viral capsid alld escape of the viral particle fr0111 the dense intermediate filament network of the epithelial cell [30,31].

The precise determinants that regulate expression of the viral genome have yet to be resolved, but most likely involve a complex interplay between the cellular factors that recognize the viral ge­nome and the E2 protein. As presented previously, papillomaviruses do not express high levels of their RNAs and proteins until the late stages of epithelial differentiation. To restrict overexpression of the viral genes, BPV and HPV has evolved tnlncated E2 proteins to act as transcriptional repressors [32-35]. These repressors lack the amino terminal trallscription activation domain but retain a func­tional DNA binding domain [36] . Therefore they compete with full-length E2 proteins for occupancy of the E2 binding sites on the viral genome. In addition, the E2 protein normally is a dimer, and formation of a heterodimer between a full- length and truncated E2 protein represents another mechanism for the inactivation of the E2-induced transcription [37]. Both full-l ength and truncated E2 can repress vira l transcription by interfering with the binding of cellular transcription factors to their recognition sites in the regula­tory region. As such, E2 has dual roles in viral transcription: activa­tion and repression. Dominance of the latter function may reduce vira l expression (and replication) in basal and parabasal keratino­cytes.

Replication of the Viral Genome The viral DNA must be selectively replicated to produce a large number of infectious prog­eny in each cel l. This would normall y be very detrimental except

250 ANDROPHY

that PV replication occurs in terminally differentiated cells. This highlights the central paradox of papillomavirus replication: it begins in a non-replicating cell layer in which the multitude of enzymes necessary for DNA synthesis are thought not to be present. Because papillomaviruses do not encode a DNA polymerase or the associated factors necessary to duplicate DNA, they must induce the cell to provide the enzymes and substrates necessary for DNA repli­cation. Consequently, the virus must mobilize these cellular factors to reproduce the viral genome. It is hypothesized that the viral oncogenes described below trick the cell into providing the neces­sary materials and enzymes to synthesize DNA. It is the E2 and E1 proteins that recruit t~ese to the viral DNA. .. .

Papillomavirus utihzes two protems, E2 and El, to Identify theLr genomes among the mass of host DNA. Mutations in the viral El gene interfere with autonomous replication of the viral DNA (re­viewed in [38]). An important observation was that the E1 protein binds to E2 protein [39-41]. It is believed thatE2 and El each bring a set of cellular factors to the DNA, and these factors replicate the viral DNA. El weakly binds a specific DNA sequence in the viral regulatory region [42], and this activity of El is greatly enhanced when it is complexed with E2 [43]. In BPV, the E1 binding site is adjacent to E2 sites, and this segment of viral DNA is sufficient for autonomous replication in murine cells when E 1 and E2 proteins are expressed [44]. The E1 protein has helicase activity, which is neces­sary for separating the DNA strands prior to their replication [45]. Both BPV and HPV E1 and E2 are necessary for viral DNA replica­tion [46-48].

IN VITRO MODELS TO STUDY PAPILLOMA VIRUS BIOLOGY

To understand the events triggered by papillomavirus infection and their biologic implications, ill vitro models using cultured cells have been employed. Transformation assays measure the ability of a gene to alter a variety of growth parameters. In some assays, the rate of cell division is increased, the nutritional requirements of the cells are altered, and they become growth factor independent, or the cells become resistant to agents that induce terminal differentiation (such as calcium). Another model incorporates the use of cells that are contact inhibited and remain as a monolayer at confluence. Intro­duction of an oncogene results in the loss of contact inhibition, and the transformed cells pile up on top of another as they continue to proliferate. These piled-up cells, or foci, are tumorigenic in immu­no compromised (nude) mice, whereas the original cell line does not form tumors in animals. This experimental system is commonly used to test whether a single gene or set of genes can induce malig­nant transformation hI vitro. In another system, primary normal cells such as keratinocytes are used because they divide only for a limited number of generations, and ultimately senesce. Continuous prolif­eration in vitro of normal human cells is exceedingly rare. Introduc­ing specific papillomavirus genes causes cultured keratinocytes to become "immortal." These cells appear morphologically un­changed but have acquired the ability to grow continuously, yet are not tumorigenic.

It is possible to reproduce more closely the normal keratinocyte differentiation program in vitro by cultivating keratinocytes at the air-media interface. In this model normal human keratinocytes or cervical cells undergo orderly stratification and differentiation when grown on a raft of permeable collagen and fibroblast matrix floating on the nutritional media. Introduction of either the whole viral genome or its oncogenes into these keratinocytes results in disruption of their maturation program, and these PV -containing keratinocytes histologically resemble early squamous cell carci­noma [49-51]. The low-risk HPVs do not induce these changes when introduced into the raft model.

Until recently it has not been possible to cultivate human papillo­mavirus in vitro, even using the raft model. It has been assumed that whereas this system approaches formation of a normal-appearing epithelium, some unknown critical step was not reproduced. This has greatly impeded understanding viral pathogenesis because it is not possible to test the effects of elimination of a specific viral gene's

THE JOURNAL OF INVESTIGATIVE DERMATOLOGY

function in the viral replicative cycle or pathogenesis. The difficulty in producing papillomavirus in vitro probably reflects the fact that its reproduction is strictly coordinated with epithelial differentiation and that papillomavirus transcription is tightly regulated.

Using the raft culture system, two groups have recently reported production of viral particles in vitro. In both, keratinocytes from an HPV-infected clinical specimen were expanded in culture and placed onto the collagen matrix. One group, using cells from a cervical lesion that contained autonomously replicating HPV 31, found that treatment of cells with multiple cycles of tumor-pro­moting agent (TP A) induced expression of the viral proteins and assembly of viral particles [52J. These HPV 31 viruses appear to be infectious. In the second system, infected cervical tissue was placed on a raft, and it was discovered that when this matrix included murine fibroblasts, they would provide the correct milieu for proper differentiation of the epithelium and for the viral reproductive cycle [53]. These successes provide much excitement, as it may be possible to derive from this model a full understanding of the connections between virus replication and epithelial cell differentiation. Their major limitation at present is the inability to introduce a cloned viral DNA genome into a normal, uninfected cell and achieve replication of the viral genome and production of infectious viral particles. For example, to test the relevance of a specific DNA sequence that might control expression of the viral capsid genes one might begin by mutating the candidate locus and testing whether virus can be produced. It can be expected that these impediments will be re­solved and that these advances will be of major utility for the study of human papillomaviruses.

Animal-based models have also been successful for producing specific HPVs. In this system normal human epithelial cells from cervix or skin are infected with HPV and placed under the renal capsule of an immunodeficient (nude) mouse for extended period, usually several months [54,55]. In this protected environment the epithelial cells differentiate and form a papilloma in which papillo­mavirus particles can be found. This has been successful for isolates ofHPV 11; however, successful production of high-risk HPV parti­cles has not been reported. Another model utilized a cell line derived from an HPV 16-infected cervical lesion in which the viral DNA was maintained extrachromosomal. Incubation of these cells in a chamber on the back of a nude mouse has been reported to allow stratified differentiation and production of HPV 16 particles [56].

DO PAPILLOMAVIRUSES CAUSE CANCER?

The association of specific HPV types with the development of epithelial malignancies, such as in epidermodysplasia verruciformis and cervical cancer, strongly support the association of HPV and cancer. The primary goal of the viral reproductive cycle cannot be oncogenesis, as suggested by the observation that viral particles are not detected in cervical dysplasias and cancers. Because papilloma­virus replication is strictly limited to differentiated cells of the epi­thelium the early causation of a dedifferentiated state would not provide an appropriate foundation for viral replication.

MOLECULAR BIOLOGY OF PAPILLOMA VIRUS ONCOGENES

The transforming properties of papillomavirus were first studied using BPV. Large quantities of BPV could be harvested from cattle papillomas. Infection of murine cell lines (NIH 3T3 and C127) with BPV caused the cells to become transformed by multiple cri­teria: loss of contact inhibition, anchorage-independent growth, and tumorigenicity in mice. Because BPV DNA replicated autono­mously in these cells, the replication functions of viral proteins could also be studied. Ll and L2 were not expressed and hence viral particles do not assemble. However, whereas similarities exist be­tween the oncogenes of BPV and HPV, there are also differences, and the following discussion will concentrate on the three onco­genes of HPV: E6, E7, and E5. For a comprehensive survey of their molecular biology the reader is referred to [57,58].

As HPV 16 and 18 were associated with the development of anogenital malignancies, the initial investigations of HPV trans-

VOL. 103, NO. 2 AUGUST 1994

fOrIning potential began with the genes from these viruses. Intro­duction of both their E6 and E7 genes into primary human kerati­nocytes rendered these cells resistant to calcium- and serum-induced terminal differentiation [59,60]. E6 and E7 cooperate in immortali­zation of primary rodent embryo fibroblasts and primary human keratinocytes [61-70]. When the high-risk E6 and E7 genes were transferred into epithelial cells cultured in the raft model, the cells exhibited features of dysplastic alterations with disorganized differ­entiation and abnormal mitoses. These high-risk E6 and E7 genes also immortalize other cell types that papillomaviruses never infect, such as smooth muscle cells [71]. Interestingly, HPV 16 E6 alone induced immortalization of normal human mammary epithelial cells, providing a useful model for the study of early breast cancer and indicating that this viral gene has targeted a conserved and common cell growth pathway [72]. These are remarkable findulgs, because the generation of immortal cell lines from any normal human tissue is an extremely rare event, even with other tumor viruses, oncogenes, and chemical inducers. These immortal cells are not tumorigenic in animals; thus not all steps in neoplastic progres­sion are provided by these genes. Continued passage of the HPV 18 containing immortal keratinocytes resulted in outgrowth of fully malignant cells [73]. This was probably due to further genetic changes in the cells, which may be a result of the papillomavirus oncogenes mechanisms of action (see below).

Important insights into the mechanism of action of the high-risk HPV E6 and E7 genes were gained from comparison with other DNA tumor viruses. The transforming proteins of adenovirus, sim­ian virus (SV) 40, and polyomavirus were known to bind an over­lapping set of cellular proteins. The identification of these common targets as tumor suppressor gene products led to the discovery, by analogy, that the high-risk HPV transforming proteins E6 and E7 complex ill vitro with the p53 and retinoblastoma (Rb) proteins, respectively.

The precise mechanisms by which p53 and Rb are involved in cell proliferation are not fully understood (reviewe~ in [~4]). Their sig­nificance cannot be overstated, because mutatIOns In p53 are the most common genetic abnormalities identified in human cancers [75]. This obse.rvation suggested th~t p53 ~light regula~e .c~1l divi­sion, and that Its loss would result In continuous cell diVISIOn and malignancy. However, p53 'knockout' mice, in which both p53 alleles have been deleted, grew normally illl/tero but as adults devel­oped tumors at high frequency. Because the mice were viable, the absence of p53 does not automatically result in unchecked cellular proliferation (reviewed in [76]) . p53 is now believed to act as a 'guardian' that protects cellular DNA [76]. When the cell's DNA is damaged (for example, by UV or x-ray irradiation), p53 arrests cell division and may allow time for DNA damage to be repaired. In­deed, p53 levels are elevated m UV-irradiated skin, and p53 muta­tions consistent with UV damage are found in cutaneous squamous cell carcinomas [77 ,78].

p53 has been reported to stimulate expression of two important genes. The first, mdm2, was identified as a protein that complexes with p53 and inhibits its transcriptional activation function. Be­cause p53 induces expression of mdm2, it appears that the latter may be a component of a regulatory feedback loop. The second gene, WAF-I, was identified as a target for p53 activation and gives a very important clue to p53 function [79]. This 21-kDa protein was simultaneously and independently isolated as an inhibitor of cyclin­dependent kinase [80] . These findings link the function of p53 to control of cell division and provide insights into how p53 deregula­tion may lead to cancer.

Rb was first identified as a gene that was found mutated in kindreds with hereditary risk of cancers, including retinoblastoma of the eye in early childhood. The Rb protein is a very large, 105-kDa nuclear phosphoprotein, and is a member of a family of pro­teins that appear to directly regulate entry into cell cycle division.

THE E6 ONCOPROTEIN

Papillomavirus E6 proteins consist of about 150 amino acids be­lieved to coordinate a zinc atom through two sets of cysteine repeats

MOLECULAR BIOLOGY OF HUMAN PAPILLOMAV1RUS 251

(cysteine x-x cysteine zinc fingers, where x is any amino acid) [81,82]. The human papilloma viruses E6 proteins have moderate homology at the amino acid level, indicating that while they share functions, they may also differ. HPV 16 E6 has a half-life of 30 - 60 min and is present in transformed and cancer-derived cell lines at extremely low levels [83]. Several reports have demonstrated that high- and low-risk E6 genes can stimulate transcription equally, suggesting that this function may be relevant to viral reproduction rather than correlate with oncogenic potential [68 ,84,85].

The high-risk HPV 16 and 18 E6 proteins interact with p53, as do SV40 large T and adenovirus Elb [86] . HPV and other viruses presumably interfere with the ability of p53 to block cell division and DNA synthesis so that viral DNA can replicate to high levels. Whereas errors in the viral DNA may occur durmg its synthesis and may not be repaired, viruses probably compensate by producing large numbers of infectious particles and evolving efficient mecha­nisms of spreading to the next host. Indeed these mutations permit genetic drift that can circumvent immune defenses. SV40 and ade­novirus prevent p53-mediated cell cycle arrest by synthesizing large amounts oflarge T and Elb, which effectively hold p53 in inactive complexes. E6 protem eliminates p53 functions through a novel mechanism. It has been shown that formation of the E6-p53 com­plex ill J/itro induced p53 degradation through a ubiquitin-depen­dent mechanism [87] . High-risk E6 binds a 100-kDa protein, called E6-AP (for E6 Associated Protein), which appears be required for complex formation between E6 and p53 and is necessary for degra­dation of p53 [88] . The gene for E6-AP has recently been cloned and is a member of the ubiquitin pathway for protein degradation [89] . It has been reported that low-risk (HPV 6 and 11) E6 also binds p53 but with reduced efficiency, yet are not capable of inducing p53 degradation [85]. This is controversial, for another group found that low-risk E6 did not bind p53 i'l J/itro [86]. Usulg another assay they reported that low-risk E6 proteins have the ability to stunulate degradation of a target protein with which it is in complex [90]. It is likely that the different interpretations reflect inability to accurately measure quantitative differences using ill vitro assays rather than invoking an alternative mechanism of action for low-risk E6 that does not involve p53. The induction of p53 degradation ill vitro also appears to hold in vivo. Introduction of HPV 16 E6 into both pri­mary human keratinocytes and mammary epithelial cells .leads to a marked decrease in the half-life of p53 [62,68,91].

Despite these important findin gs, there are several lines of evi­dence that suggest that E6 possesses other functions. DPV E6 does not bind p53, yet fully transforms murine Cl27 cells [92]. BPV E6 and HPV 6 E6, which do not induce p53 degradation ill J/itro or ill vivo, can immortalize human mammary epithelial cells, although they are much less efficient than HPV 16 [93]. HPV 8, which is found in cutaneous squamous cell carcinoma in epidermodysplasia verruciformis, does not bind or degrade p53 ill vitro [94], but, similar to BPV E6, transforms mouse cells [95]. These putative non-p53 functions are yet to be determined.

THE E7 ONCOPROTEIN

The E7 oncoprotein is an acidic 98 amino acid phosphoprotein that has been localized to the nuclear matrix [96] . Two cysteine-x-x-cys­terne motifs in the carboxy terminus mediate zinc binding and di­merization [97]. Recognition of the amino acid similarities between the HPV E7 proteins and the DNA tUIll.or virus transforming pro­teins SV40 Large T and adenovirus Ela facilitated elucidation of the biochemical properties ofE 7 [98,99]. These similarity regions were shown to bind the tumor suppressor gene product Rb and the related p107 proteins [100-103] . Both Rb and pl07 regulate cell cycle division but act at different stages. Detailed analyses have found that the pl07 binding region of E7 overlaps with but can be distin­guished from the Rb binding domain [104,105].

It is thought that when E7 binds Rb or pl07, a transcription factor normally bound to Rb/p107 is released, because E7 binds the same pocket as the factor [106,107]. This factor, called E2F, is a se­quence-specific DNA binding protein, and the DNA motif it rec­ognizes is found in many genes essential for cell division {reviewed

252 ANDROPHY

in [108]). Recently, E2F and a related family of genes have been cloned [109,110]. There is evidence that the events induced by the Rb/pl07 association with E7 can eventuate in a complex yet coor­dinated cascade of positive and negative signals that allow a cell to replicate its DNA and divide. Other regions ofE7 bind additional growth-related cellular factors. HPV 16 E7 has bee~l identified in complexes with histone kinase [111], p33cdk2 and cycl1l1A [112], and casein kinase II [113 ,1 14], which phosphorylates high-risk E7 more efficiently than it does low risk. Because Rb preferentially binds to phosphorylated E7 protein, this phosphorylation may in part relate to the oncogenic potential of high-risk E7 [llS].

Both HPV 16 and 18 E7 induced focus formation in murine cell line transformation assays [116,1 17]. Consistent with its interaction with Rb/pl07, introduction ofE7 alone into primary human kera­tinocytes resulted in an increased rate of proliferation for an ex­tended period of time, although they eventually senesced [118]. Introduction of high-risk E7 can induce keratinocyte immortaliza­tion in the absence of E6, although this was a rare occurrence and presumably other changes occurred in the cells that bypass the pS3 pathway. U sing very high-efficiency retroviral-mediated infection of primary human keratinocytes, it has been shown that low-risk HPV 6 E7 can cooperate with high-risk E6 to induce immortaliza­tion of primary human keratinocytes, and in the alternative mixing experiment, HPV 6 E6 cooperated with HPV 16 E7 to induce immortalization, although in both instances, the efficiency was less than with high-risk E6 and E7 [119]. These data suggest that the difference between the high- and low-risk E7 genes may be the efficiency at which they perform similar functions. Low-risk E7 also complexes with Rb/pl07, and whereas there are quantitative differences in Rb/pl07 association affinities with high-risk E7 [120 -122], the specific interactions that are responsible for high­risk E7 inducing transformation of cultured cells have not been definitively proved [123,124] . As with E6, there remain unidenti­fied properties of E7 that are required for its transforming activities. For example , mutations in E7 that do not affect Rb/p107 associa­tion interfere with its ability to transform and immortalize the cells [124].

The cottontail rabbit papillomavirus (CRPV) model has recently been developed and permits evaluation of the biologic requirements for specific CRPV genes irl vivo. Remarkably, inoculation of CRPV DNA into rabbit skin resulted in papillomas in which CRPV DNA replicated autonomously and virus particles were synthesized [12S,126] . Surprisingly, rabbit skin injected with viral genomes carrying mutations in the CRPV E7 gene that abrogated Rb binding ability also formed papillomas, indicating that the interaction with Rb was not absolutely necessary for development of a wart or for replication of the virus [126,127]. These data suggest that Rb bind­ing by E7 may be a function important for malignant transforma­tion; however, it is possible that pl07 binding is more relevant to papilloma formation and viral reproduction.

THE ES ONCOPROTEIN

The ES protein represents another fascinating and ingenious means that papillomaviruses have evolved to prime the cell for viral repro­duction . An ES gene has been identified in bovine, deer, elk, and some human papillomaviruses. These animal viruses produce a con­comitant dermal fibroblast proliferation (fibropapillomas), along with an epithelial component, in their hosts. ES proteins have not been established in all the HPVs, in part because it is so small; the BPV ES gene product is only 44 amino acids and was initially overlooked as too tiny to do anything. The HPV ES gene is often deleted or its gene not expressed in human cervical carcinomas, whereas E6 and E7 are always synthesized. This suggests that ES plays a role at an early stage of viral carcinogenesis. BPV -1 and HPV 6c ES have been reported to transform established murine fibroblast cell lines, whereas HPV 16 ES did not [92,128,129]. Introduction of HPV 16 ES into established murine keratinocyte lines induced their tumorigenicity in nude mice, suggesting a potential role in human malignancy [130].

Much of the current understanding of the biochemistry ofES has

THE JOURNAL OF INVESTIGATIVE DERMATOLOGY

been derived from studies using BPV. The BPV ES protein has been isolated from transformed rodent cells as a 14-kDa disulfide-linked dimer [131]. A pair of cysteine residues separated by one amino acid is believed to mediate dimerization. Immunoelectron microscopy detection of the BPV ES protein has revealed that the majority is oriented within the luminal leaflet of the Golgi stacks, with signifi­cantly less in the plasma membrane [132] . The abundance of ES within the Golgi suggests the possibility that ES may interact with a growth factor receptor (see below) in transit to the plasma mem­brane. Recently, BPV -1 ES protein has been identified in both basal and upper layer keratinocytes in naturally occurring bovine fibro­papillomas [133].

The small size ofES facilitated detailed structure-function analy­ses. The amino terminal two-thirds of the protein is extremely hydrophobic and has been shown to be responsible for membrane localization. Within this hydrophobic transmembrane domain is a single polar amino acid, glutamine (amino acid 17 in BPV ES), and its presence (or that of another charged residue) is important for the transforming abi lity of BPV ES [134]. The carboxy terminal 16 amino acids are charged and are absolutely required for transforma­tion of murine cells [13S]. This segment has been reported to be oriented extracellularly, but is probably too small to form a growth factor receptor.

As with E6 and E7, progress in elucidating the mechanism through which ES acts has resulted from studies that identify cellu­lar factors that bind ES. A small 16-kDa protein (called p16) has been found to associate with BPV ES [136] . p16 has been identified as a component of the vacuolar H + ATPases [137 -139]. This pro­ton transporter is responsible for acidification of cellular compart­ments such as Iysosomes. The region of ES that binds p16 has been mapped to the transmembrane hydrophobic region and complex formation is believed to involve glutamine 17. These observations have led to a model in which the hydrophobic region mediates insertion ofES into hydrophobic cellular membranes and the region of glutamine 17 attracts and binds to the p16 component of the cellular proton pump. Presumably this interaction inhibits acidifi­cation of Iysosomes and may thus interfere with receptor down­regulation (see below).

Recognizing that ES was too small to possess enzymatic function yet was localized to the same cellular compartments as growth fac­tor receptors, it was discovered that introduction ofBPV ES caused increased levels of the epidermal growth factor (EGF) receptor in mouse cells, and as a consequence, the murine fibroblasts became exquisitely sensitive to transformation by EGF [140] . In the ES­transformed cells, these EGF receptors were maintained at the plasma membrane in the active phosphorylated form, suggesting that they were not being downmodulated, as they normally are subsequent to ligand binding. This is consistent with the hypothesis that ES interferes with receptor processing as a consequence of inhibiting p16-mediated acidification of Iysosomes. Consistent with this, stimulation of transformation by BPV ES was not specific for the EGF receptor, and similar synergistic effects were observed with other receptors. A specific target of BPV -1 ES may be the platelet-derived growth factor (PDGF) receptor, and a physical complex between ES and the PDGF receptor has been reported [141,142] .

The EGF and PDGF receptors are activated by distinct ligands and in turn transduce signa ls through separate pathways. How can BPV ES interact with different growth factor receptors? These ob­servations may be explained by ES possessing two different activities that can distinguish the receptors. It has been found that the cys­teines in the carboxy terminus of ES are necessary for transforma­tion through the PDGF receptor, and that the ES transmembrane domain and glutamine 17 arc also critical [143]. The hydrophobic portion of each ES monomer in the disulfide-bonded dimer binds a p16 molecule, which in turn complexes with a PDGF-receptor monomer. Thus, bridged by p16, ES may mediate PDGF receptor dimerization and activation. An alternate mechanism would be nec­essary for the EGF class of receptors, because they do not require dimerization for activation. It has been reported that the carboxy terminus ofBPV E5 physically binds a 125-kDa protein in the alpha

VOL. 103, NO.2 AUGUST 1994

adaptin family [144]. This association required the cysteines to be reduced, that is, not disulfide linked, which is opposite of the PDGF receptor model. It is speculated that via the adaptin complex, E5 may interfere directly with EGF receptor downregulation, as adap­tins are responsible for the processing of this class of receptors. These observations predict dual functions of E5, which depend on the particular growth factor receptor expressed in a papilloma virus­infected cell.

The role ofE5 in viral pathogenesis has also been examined using the CRPV inoculation model. Injection of DNA containing inacti­vating E5 mutations into the skin of domestic rabbits resulted in reduced efficiency of papilloma formation [145] . This indicates a role ofE5 in the viral reproductive cycle, although it does not appear to be essential. Perhaps the function of E5 is to establish the appro­priate aC.tiv.ated cellular. environment in .a cell that is. terminally differentlatmg, thus settll1g the stage for Viral DNA repitcatlon that is more directly induced by E7, with inhibitory checkpoints re­moved by E6 and perhaps E7.

FURTHER EVIDENCE THAT HPV CAUSES CANCER

There are several lines of evidence that support the role of HPV in causing cancer. As presented above, the E5, E6, and E7 genes inter­act with cellular proteins that control cell growth. High-risk E6 and E7 induce the immortalization of normal human keratinocytes (and other cell types) and morphologically transform rodent cells. Inser­tion ofHPV 16 E6 and E7 into the mouse genome (transgenic) leads to epithelial tumors in tissues that express the viral genes [146,147]' and similar results have been reported using CRPV [148]. Analyses of in vivo malignancies also support the role of the virus in oncogen­esis . HPV DNA is found in greater than 90% of cervical cancers, and the metastases also retain the HPV DNA. Interestingly, the viral genome is often integrated in cancer cells, although it is not known whether this precedes neoplastic conversion or is a conse­quence of the multiple genetic changes that occur in tumor cells. Nonetheless, the HPV E6 and E7 genes are always retained and expressed in the tum.or t~ssue, wl~ereas the ?ther viral genes are often disrupted or lost dunng lI1tegratlon. Cell hnes denved from cerVical carcinomas produce HPV E6 and E7 proteins, for instance HeLa cells, synthesizing HPV 18 E6 and E7 after decades in culture. Treatment of cervical carcinoma cells with antisense oligonucleo­tides or introduction of antisense DNA into these cells inhibited E6 and E7 expression and resulted in decreased cell proliferation and loss of tumorigenicity [149-152]. Cell hybrids of normal fibro­blasts crossed with HeLa cells led to decreased expression ofE6 and E7 and loss of He La tumorigenicity [153]. These observations imply that the continued expression of E6 and E7 in both fresh cervical cancers and cervical tumor cell lines is necessary for the malignant state.

The epidemiologic association between HPV and cancer now seems certain. Several large studies have indicated that infection with specific HPV types is associated with a relative risk estimated to range from 10 to 200 times for development of cervical neoplasia [154]. Several recent studies have 'provided important information and have been reviewed recently [155,156]. First, repeated testing detects HPV in cytologically normal cervical specimens. These may represent subclinical infection or presence of the virus in secretions without true infection. There is reason to believe that HPV infec­tion of the cervix may be transient, such that in the infected cell the virus does not replicate or the infected cells are desquamated and lost. Prospective studies have suggested that infection of the cervix with high-risk HPV progresses to cervical intraepithelialneoplasia (CIN) stage 2 or 3 over a relatively brief period of perhaps 2-3 years. The success of Pap smear screening in lowering the incidence of cervical cancer indicates that most invasive cervical cancers do not arise from latent, subclinical HPV infection but develop from ab­normal intraepithelial precursor lesions.

HOW DOES HPV CAUSE CANCER?

The critical obstacle that papillomaviruses must overcome is that viral DNA must be synthesized in cells that are not replicating. The

MOLECULAR BIOLOGY OF HUMAN PAPILLOMA VIRUS 253

papillomavirus oncogenes are thought to induce the cellular events necessary to synthesize DNA. E1 and E2 usurp these pathways for replication of the viral genom.e. The E6 and E7 oncogenes inactivate the p53 and Rb pathways and thus abrogate cellular control mecha­nisms that regulate DNA synthesis. This is relevant because p53 might recognize that DNA is being synthesized outside the cell cycle and shut down the metabolic pathways involved in DNA replication. HPV 16 E6 was shown to block growth arrest that occurs normally after UV radiation and allowed the cells to con­tinue to divide [157] .

Under ordinary conditions there are no detrimental effects of papillomavirus induction of DNA replication enzymes in the upper, terminally differentiated squamous cells. Malignancy may arise from the stimulation of cell division by the viral oncogenes in an inappropriate cell before it has been irreversibly converted to the non-replicating population. Perhaps in this case there is unregulated duplication of the host DNA, resulting in gross chromosomal ab­normalities. Alternatively, it is possible that neoplasia develops due to unregulated expression of the viral oncogenes. For example, it may be that loss of autonomous replication leads to integration and uncoupling of a normal mechanism that keeps E6 and E7 expressed at low levels. As a result of E6 inhibition of the normal p53 growth arrest function prior to DNA repair, il~ured cells may accumulate DNA damage and pathogenic mutations. Cells containing E7 have also been shown to develop an increased rate of mutations [158]. Cultured keratinocytes immortalized by HPV 16 or 18 also display chromosomal instability. The genetic alterations that occur in the dividing cells may often be detrimental to survival, but the accumu­lation of such events over years most likely permits the selection and establishment of cancerous cells that proliferate and invade in an unregulated manner.

One significance of the elucidation of the functions and events triggered by the viral oncoproteins E5 , E6, and E7, and the viral transcription/replication proteins E1 and E2, is that these provide insights into normal metabolic pathways operative in epithelial cells. Furthermore, these discoveries should permit the rationale design of anti-viral and anti-cancer agents that we may someday use to treat our patients with HPV infection.

REFERENCES

I. De Villiers E: Heterogeneity of the human papillomavirus group.] Virol 63:4898-4903,1989

2. Chan SY, Bernard HU, Ong CK, Chan SP, Hofmann B, Delius H: Phyloge­netic analysis of 48 papillomavirus types and 28 subtypes and variants: a showcase of the molecular evolution of DNA viruses .] Viro/66:57 14 - 5725, 1992

3. van Rast M, Kaplan]B, Burk RD: Phylogenetic classification of human papillo­maviruses: correlation with clinical manifestations. ] Ge" Viro l 73:2653-2660, 1992

4. Durst M, Gissmann L, Ikenberg H, zm Hausen H: A papillol11avirus DNA from a cervical carcinoma and its prevalence in cancer biopsy sanlplcs from differenr geographic regions. Proe Nat! Acad Sci USA 80:38 12-3815,1983

5. McCance 0], C larkson PK, Dyson JL, Walker PG, Singer A: Human papillo­mavirus types 6 and 16 in multi focal intraepithelial neoplasias of the female lower genital tract. Br] Ob51el GYIltJeco/92: 1093- 1100, 1985

6. zur Hausen H: Human papillom3viruscs in the pathogenesis of anogenital cancer. Virology 184:9- 13,1991

7. Stoler MH, Wolinsky SM, Wl1itbcck A, Broker TR, Chow LT: Differentia­tion-linked human papillomavirus types 6 and 1 1 transcription in genital condylomata revealed by in situ hybridization with message-specific RNA probes. Virology 172:331 - 340, 1989

8. Zhou], Sun XV, Stenzel], Frazer IH: Expression of vaccinia recombinant HPV 16 L1 and L2 ORF proteins in epithelial cells is suRicient for assembly ofHPV virion-like particles. Virology 185:251-257, 199 1

9. Kirnbauer R,Booy F, Chen g N, Lowy DR, Schiller ]T: Papillomavirus L1 major capsid protein self-assembles into virus-like particles that arc highly immuno­genic. Proe Natl A ead Sci USA 89:12180 - 12184, 1992

10. Kirnbauer R, T aub J , Greenstone H, Roden R, Lowy DR, SchiIIer JT: ERicient self-assembly of human papillomavirus type 16 LI and LI-L2 into virus-like particles. ] Viro/67:6929 - 6936, 1993

11. Rose RC, Bonnez W , Reichman RC , Garcca RL: Expression of human papillo­mavirus type 11 LI protein in insect cells: in vivo and in vitro assembly of viruslike particles.] viroI67:1936- 1944, 1993

12. Hagensee ME, Yaegashi N , Galloway DA: Self-assembly of human papilloma­virus type 1 capsids by expression ofLl protein alone or by cocxpression of the Ll and L2 capsid proteins.] ViroI67:315 -322, 1993

13. Sousa R, Dost:ttnj N , Yaniv M: Control of p.api llomavirus gene expression. Bioe/'i", Biopll)'s Acta 1032: 19 -37, 1990

254 ANDROPHY

14. Chong T. Apt D. Gloss B. Isa M. Bernard HU: The enhancer of human papillo­mavirus type 16: binding sites for the ubiquitous transcription factors oct-1. NFA. TEF-2. NFl. and AP-l participate in epithelial cell-specific transcrip­tion.] Virol 65:5933 - 5943. 1991

15. Mack DH. Laimins LA: A keratinocyte specific transcription factor. KRF-1. interacts with AP-1 to activate expression of human papillomavirus type 18 in squamous epithelial cells. Proc Nat! Acad Sci USA 88:9102-9106. 1991

16. Bartsch D. Boye B. Baust C. zur Hausen H. Scwarz E: Retinoic acid-mediated repression of human papillomavirus 18 transcription and different li gand regulation of the retinoic acid receptor B gene in non-tumorigenic and tumor­igenic HeLa hybrid cells. EMBOJ 11:2283-2291.1992

17. Ishiji T. Lace MJ . Parkkinen S. Anderson RD. Haugen TH. CripeTP. XiaoJH. Davidson I. Chambon p. Turek LP: Transcriptional enhancer factor (TEF-1) and its cell specific co-activator activate human papillomavirus-16 E6 and E7 oncogene transcription in keratinocytes and cervical carcinoma cells. EMBO J 11:2271-2281.1992

18. Monini p. Grossman SR. Pepinsky B. Androphy EJ. Laim.ins LA: Cooperative binding of the E2 protein of bovine papillomavirus to adjacent E2-responsive sequences.J viro/65:2124-2130. 1991

19. Androphy EJ. Lowy DR. Schiller JT: Bovine papillomavirusE2 trans-activating gene product binds to specific sites in papillomavirus DNA. Natllre 325:70-73.1987

20. Spalholz BA. Yang YC. Howley PM: Transactivation of a bovine papilloma virus transcriptional regulatory element by the E2 gene product. Cell 42: 183-191.1985

21 . Haugen TH. Cripe TP. Ginder GO. Karin M. Turck LP: Trans-activation of an upstream early gene promoter of bovine papilloma virus-1 by a product of the viral E2 gene. EMBO J 6:145-152. 1987

22. Hawley-Nelson P. Androphy EJ. Lowy DR. Schiller JT: The specific DNA recognition sequence of the bovine papilloma virus E2 protein is an E2-depen­dent enhancer. EMBO J 7:525 - 531.1988

23. McBride AA. Romanczuk H. Howley P: The papillomavirus E2 regulatory proteins. J Bioi Clm" 266:1 84 11 - 18414. 1991

24. Giri I. Yaniv M: Structural and mutational analysis of E2 trans-activating pro­teins of papillomaviruses reveals three distinct functional domains. EMBO J 7 :2823 - 2829. 1988

25. Haugen TH. Turek LP. Mercurio FM. CripeTP. Olson BJ. Anderson RD. Seidl D. Karin M. Schiller J: Sequence-specific and general transcriptional activa­tion by the bovine papillomavirus-l E2 trans-activator require an N-tenrunal amphipathic helix-containing E2 domain. EMBO J 7:4245 - 4253. 1988

26. Prakash SS. Grossman SR. Pepinsky RE. Laimins LA. Androphy EJ: Amino acids necessary for DNA contact and dimerization imply novel motifs in the papil­lomavirus E2 trans-activator. Gelles & Dev 6:105-116.1992

27. H~gde RS. Grossman SR. Laimins LA. Sigler PB: Crystal structure at 1.7 Aofthe bovine papillomavirus-1 E2 DNA-binding domain bound to its DNA target. Nature 359:505-512.1992

28. Li R. KnightJD.Jackson SP. Tjian R. Botchan MR: Direct interaction between Sp1 and the BPV enhancer E2 ptotein mediates synergistic activation of transcription. Cell 65:493-505. 1991

29. Ham J . Dostatni N. Amos F. Yaniv M: Several different upstream promoter elements can potentiate tra.nsactivation by the BPV-l E2 protein. EMBO J 10:2931 - 2940.1991

30. Doorbar J . Ely S. Sterling J. Mclean C. Crawford L: Specific interaction be­tween HPV-16 E l -E4 and cytokeratins results in coll apse of the epithelial cel l intennediate filament network. Natllre 352:824-827.1991

31. SterlingJC. Skepper IN. Stanley MA: Immunoelectron microscopicallocaliza­tion of human papillomavirus type 16 L1 and E4 proteins in cervical keratino­cytes cultured in vivo.J /"veJt Dermato/l00:154-158. 1993

32. Hubbert NL. Schiller JT. Lowy DR. Androphy EJ: Bovine papilloma virus­transfonned cells contain multiple E2 proteins. Proc Natl Acad Sd USA 85:5864-5868.1988

33. Lambert PP. Hubbert NL. Howley PM. Schiller JT: Genetic assignment of multiple E2 gene products in bovine papiliomavirus-traJlSfonned cel ls.] Virol 63:3151-3154.1989

34. ChoeJ. Vaillancourt p. Stenlund A. Botchan M: Bovine papillomavirus type 1 encodes two forms of a transcriptional repressor: strucrural and functional analysis of new viral cDNAs. J ViroI63:1743-1755. 1989

35. Doorbar J. Parton A. Hartley K. Banks L. Crook T. Stanley M. Crawford L: Detection of novel splicing patterns in a HPV16-colltaining keratinocyte cell line. Virology 178:254-262. 1990

36. Lambert PF. Spalholz BA. Howley PM: A transcriptional repressor encoded by BPV-I shares a common carboxy-terminal domain with the E2 transactivator. Cell 50:69 - 78. 1987

37 . Barsoum J. Prakash SS. Han p. Androphy EJ: Mechanism of action of the papillomavirus E2 repressor-repression in the absence of DNA binding. J Viro l 66:3941- 3945. 1992

38. Lambert PF: Papillomavirus DNA replication.] virol 65:3417 - 3420. 1991 39. Mohr IJ. C lark R. Sun S. Androphy EJ. MacPherson p. Botchan MR: Targeting

the El replication protein to the papillomavirus origin of replication by com­plex fonnation with the E2 transactivator. Scie" ce 250:1694- 1699. 1990

40 . Blitz IL. Laimins LA: The 68-kilodalton El protein of bovine papillomavirus is a DNA binding phosphoprotein which associates with the E2 transcriptional activator in vitro.] V iro/65 :649-656. 1991

41. Lusky M. Fontane E: Formation of the complex of bovine papilloma virus E1 and E2 proteins is modulated by E2 phosphorylation and depends upon sequences within the carboxyl terminus ofE1. Proc Natl Acad Sci USA 88:6363 -6367. 1991

THE JOURNAL OF INVESTIGATIVE DERMATOLOGY

42. Wilson VG. Ludes MJ: A bovine papillomavirusEl-related protein binds specif­ically to bovine papillomavirus DNA.J viro/65:5314-5322. 1991

43. Seo YS. Friedmann M. Lusky M. G ibbs E. Kim HY: Bovine papillomavirus (BPV) encoded E2 protein enhances binding ofEl protein to the BPV replica­tion origin. Proc Nat! AClId Sci USA 90:2865-2869.1993

44. Ustav M. Ustav E. Szymanski p. Stenlund A: Identification of the origin of replication of bovine papillomavirus and characterization of the viral origin recognition factor E1. EMBO J 10:4321-4329. 1991

45. Yang L. Mohr 1. Fouts E. Lim DA. Nohaile M. Botchan M: The El protein of bovine papillomavirus 1 is an ATP-dependent DNA helicase. Proc Natl Acad Sci USA 90:5086 - 5090. 1993

46. Ustav M. Stenlund A: Transient replication ofBPV-l requires two viral poly­peptides encoded by the E I and E2 open reading frames. EMBO J 10:449-457.1991

47. Chiang C-M. Ustav M. Stenlund A. Ho T. BrokerTR. Chow LT: Viral El and E2 proteins support replication of homologous and heterologous papilloma­viral origins. Proc Na tl Acad Sci USA 89:5799 - 5803. 1992

48 . del Vecchio AM. Romanczuk H. Howley PM. Baker CC: Transient replication of human papillomavirus DNAs. J Virol 66:5949 - 5958. 1992

49. McCance OJ. Kopan R. Fuchs E. Laimins LA: Human papillomavirus type 16 alters human epithelial cell differentiation in vitro. Prot Natl Acad Sci USA 85:7169-7173.1988

50. Blanton RA. Perez RN. Merrick DT. McDougall JK: Epithelial cells immortal­ized by human papilloma viruses have premalignant characteristics in orgauo­typic culture. Am J PatllO/138:673 - 685. 1991

51. Hudson JB. Bedell MA. McCance OJ. Laimins LA: Immortalization and altered differentiation of human keratinocytes in vitro by the E6 and E7 open reading frames of human papillomavirus type 18.J ViroI64:519-526. 1990

52. Meyers C. Frattini MG. HudsonJB. Lail11ins L: Biosynthesis of human papillo­mavirus from a continuous cel l line upon epithelial differentiation. Science 257:97 1- 973.1992

53. Dollard SC, Wilson JL. Demeter LM. Bonnez W. Reichman RC. Broker TR. Chow L T: Production of humaJl papilloma virus and modulation of the in­fectious program in epithelial raft cultures. Gwes & DeY 6:1131-1142.1992

54. Kreider JW. Howett MK. Leure-Dupree AB. Zaino RJ. Weber JA: Laboratory production in vivo of infectious human papillomavirus type 11. J Virol 61:590-593.1987

55. Stoler MH. Whitbeck A. Wolinsky SM. Broker TR. C how LT. Howett MK. Kreider JW: Infectious cycle of human papillomavirus type 11 in human foreskin xenografts in nude mice.J V iroI64:3310-3318. 1990

56. SterlingJ. Stanley M. Gatward G. Minson T: Production of hum all papilloma­virus type 16 virions in a keratinocyte cel lline. J ViroI64 :6305-6307. 1990

57. DiMaio 0: Transforming activity of bovine and human papillomaviruses in cultured cells. Ad" Cm"er /~es 56:133-159.1991

58. Mansur CP. Androphy EJ: Cellu lar transformation by papillomavirus oncopro­teills. BBA Reviews i" Callcer 1155:323-345. 1993

59. Schlegel R. Phelps WC. Zhang YL. Barbosa M: Quantitative keratinocyte assay detects two biological activities of human papillomavirus DNA and identifies viral types associated with cervical carcinoma. EMBOJ 7:3181 - 3187. 1988

60. Barbosa MS. Schlegel 11.: The E6 and E7 genes of HPV-18 arc sufficient for inducing two-stage in vitro transformation of human kcratinocytcs. O'lCogent> 4:1529-1532.1989

61. Pirisi L. Yasumoto S. Feller M. Doniger J. DiPaolo JA: Transformation of human fibroblasts and keratinocytes with human papillomavirus type 16 DNA.J ViroI61:106I-1066. 1987

62. Hubbert NL. Sed man SA. Schiller JT: Human papillomavirus type 16 increases the degradation rate of p53 in human keratinocytes. J Viro166:6237 - 6241. 1992

63. Kaur p. McDougall JK: HPV-18 immortalization of human keratinocytcs. Virology 173:302-310. 1989

64. Kaur p. McDougall JK. Cone R: Immortalization of primary human epithelial cells by cloned cervical carcinoma DNA containing human papillomavirus type 16 E6/E7 open reading frames.J Ge .. ViroI70:1261-1266. 1989

65. Munger K. Phelps WC. Bubb V. Howley PM. R. S: The E6 and E7 genes of the human papillomavirus type 16 together arc necessary and sufficient for trans­formation of primary human keratinocytes. J Virol 63:4417 - 4421. 1989

66. Pecoraro G. Morgan D. Defendi V: Differential effects of human papillomavirus type 6. 16. and 18 DNAs on immortalization and transformation of human cervical epithelial cells. Proc Natl Acad Sci USA 86:563-567. 1989

67. Woodworth CD. Waggoner S. Barnes W. Stoler MH. DiPaolo JA: Human cervical and foreskin epithelial cells immortalized by human papillomavirus DNAs exhibit dysplastic differentiation in vivo. Catlcer Res 50:3709-3715. 1990

68. Sed man SA. Barbosa MS. Vass WC. Hubbert NL. Haas JA. Lowy DR. Sch.iller JT: The full-length E6 protein of human papillomavirus type 16 has trans­forming and trans-activating activities and cooperates with E7 to inll110rtalize keratinocytes in culture.J ViroI65:4860 - 4866. 1991

69. Barbosa MS. Vass WC. Lowy DR. Schiller JT: In vitro biological activities of the E6 and E7 genes vary among human papillomaviruses of different oncogenic potential.J Viro/65:292 -298. 1991

70. Pecoraro G. Lee M. Morgan D. Defendi V: Evolution of in vitro transformation and tumorigenesis of HPV16 and HPV18 immortalized primary cervical epithelial cells. AmJ PatllO/138:1-8 . 1991

71. Perez-Reyes N. Halbert CL. Smith Pl'. Benditt EP. McDougallJK: Immortali­zation of primary human smooth muscle cells. Proc Natl Acad Sci USA 89:1224-1228.1992

72. Band V. Zajchowski D. Kulesa V. Sager R: Human papilloma virus DNAs

VOL. \03, NO.2 AUGUST 1994

immortalize normal human mammary epithelial cells and reduce their growth factor requirements. P,oc Natl Acad Sci USA 87:463-467, 1990

73. Hurlin PJ, Kaur P, Smith PP, Perez RN, Blanton RA, McDougallJK: Progres­sion of human papillomavirus type 18-immortalized human keratinocytes to a malignant phenotype. Proc Noll Acad Sci USA 88:570 -574, 1991

74. Levine AJ, Momand J : Tumor suppressor genes: the p53 and retinoblastoma sensitivity genes and gene products. Biochi", Biophys Acta 1032:119-136, 1990

75. Vogelstein B, Kinzler KW: p53 function and dysfunction . Cell 70:523-526, 1992

76. Lane DP: p53, guardian of the genome. Natu,e 358:15-16, 1992 77. Hall PA, Mckee PH, Menage H, Dover R, Lane DP: High levels of p53 protein

in UV-irradiated normal human skin. Otlcogetle 8:203-207,1993 78. Brash DE, RudolphJA, SimonJA, LinA, McKenna GJ, Baden HP, HalperinAJ,

Ponten J: A role for sunlight in skin cancer: UV-induced p53 mutations in squamous cell carcinoma. Proe Noll Acad Sci USA 88:10124-10128, 1991

79. El-Deiry WS, Tokino T, Velculescu YE, Levy DB, Parsons R, TrentJM, Lin D, Mercer E, Kinzler KW, Vogelstein B: W AFl, a potential mediator of p53 rumor suppression. Cell 75:817 - 825, 1993

80. HarperJW, Adami GR, Wei N, Keyomarsi K, Elledge SJ: The p21 Cdk-inter­acting protein Cipl is a potent inhibitorofGl cyclin-dependent kinases. Cell 75:805-816,1993

81. Barbosa MS, Lowy DR, Schiller JT: Papillomavirus polypeptides E6 and E7 are zinc-binding proteins. J virol 63: 1404 - 1407, 1989

82. Grossman SR, Laimins LA: E6 protein of human papillomavirus type 18 binds zinc. Otlcogetle 4:1089-1093,1989

83 . Androphy E], Hubbert NL, Schiller JT. Lowy DR: Identification of the HPV-16 E6 protein from transformed mouse cells and human cervical carcinoma cell lines. EMBO J 6:989 - 992, 1987

84. Desaintes C, Hallez S. VanAlphen p. Burny A: Transcriptional activation of several promoters by the E6 protein of human papillomavirus type 16.J virol 66:325-333, 1992

85. Crook T, Tidy J. Vousden K: Degradation of p53 can be targeted by HPV E6 sequences distinct from those required for p53 binding and trans-activation. Cell 67:547 - 556, 1991

86. Werness BA, Levine AJ, Howley PM: Association of human papillomavirus types 16 and 18 E6 proteins with p53. Science 248:76-79. 1990

87. Scheffner M, Werness BA. Huibregtse ]M, Levine A], Howley PM: The E6 oncoprotein encoded by human papillomavirus types 16 and 18 promotes the degradation of p53. Cell 63:1129-1136, 1990

88. Huibregtse JM. Scheffner M, Howley PM: A cellular protein mediates associa­tion of p53 with the E6 oncoprotein of human papillomavirus types 16 or 18. EMBO J 10:4129- 4135, 1991

89. Huibregtse ]M, Scheffner M, Howley PM: Cloning and expression of the cDNA for E6-AP. a protein that mediates the interaction of the human papilloma­virus E6 oncoprotein with p53. Mol Cell BioI 13:775 - 784, 1993

90. Scheffner M, Munger K, Huibregtse ]M, Howley PM: Targeted degradation of the retinoblastoma protein by human papillomavirus E7-E6 fusion proteins. EMBOJ 11:2425-2431,1992

91. Band V, DeCaprio ]A. Delmolino L, Kulesa V, Sager R: Loss of p53 protein in human papillomavirus type 16 E6 immortalized human mammary epithelial cells.J Vi,0165:6671-6676. 1991

92. Schiller ]T, Vass WC, Lowy 0: Identification ofa second transforming region in bovine papillomavirus DNA. P,oe Noll Aead Sci USA 81 :7880-7884. 1984

93. Band V, Dalal S, Delmolino L, Androphy EJ: Enhanced degradation of p53 protein in HPV-6 and BPV-l E6 immortalized human mammary epithelial cells. EMBO J 12:1847-1852. 1993

94. Steger G, Pfister H: In vitro expressed HI'V 8 E6 protein docs not bind p53. Arch Vi,0/125:355 -3(\0.1992

95. Iftner T, Bierfelder S. Csapo Z, pfister H: Involvementofhuman papillomavirus type 8 genes E6 and E7 in transformation and replication. J Viral 62:3655 -3661,1988

96. Greenfield I. Nickerson], Penman S, Stanley M: Human papillomavirus 16 E7 protein is associated with the nuclear matrix. Proc Nat! Acad Sci USA 88:11217-11221.1991

97. Mcintyre M, Frattini M. Grossman S. L1imins L: Human papilloma virus type 18 E7 protein requires intact cys-x-x-cys motifs for zinc binding, dimerizatioll. and transformation but not for Rb binding.J Vi,0167:3142-3150. 1993

98. Vousden KH,]at PS: Functional similarity between HPV16 E7, SV40 large T and adenovirus Ela proteins. O,leogetle 4:153-158.1989

99. Phelps W, Bagchi S, BarnesJ, Raychaudhuri P, Kraus V, MungerK. Howley P, Nevins J: Analysis of trans activation by human papillomavirus type 16E7 and adenovirus 12S EIA suggests a common mechanism.J Vi,0/65:6922-6930, 1991

100. DeCaprio ]A. Ludlow ]W. Figge J, Shew J-Y, Huang C-M. Lee W -H, Marsilio E, Paucha E, Livingston OM: SV40 large tumor antigen forms a specific complex with the product of the retinblastoma susceptibility gene. Cell 54:275-283, 1988

101. Dyson N, Howley PM, Munger K, Harlow E: The human papilloma virus-16 E7 oncoprotein is able to bind to the retinoblastoma gene product. Scifllce 243:934 - 937,1989

102. Munger K. WernessBA, Dyson N. Phelps WC, HarlowE, Howley PM: Com­plex formation of human papillomavirus E7 proteins with the retinoblastoma tumor suppressor gene product. EMBO J 8:4099-4105, 1989

103. Dyson N, Guida P, Munger K, Harlow E: Homologous sequences in adenovirus Ela and human papillomavirus E7 proteins mediate interaction with the same set of cellular proteins.J Vi,0166:6893-6902, 1992

MOLECULAR BIOLOGY OF HUMAN PAPILLOMAVIRUS 255

104. Huang PS,l'atrick DR, Edwards G, Goodhart PJ, Humber HE, Miles L, Garsky VM. Oliff A, Heimbrook DC: Protein domains governing interactions be­tween E2F, the retinoblastoma gene product, and human papillomavirus type 16 E7 protein. Mol Cell BioI 13:953-960, 1993

105. Wu EW. Clemens KE, Heck DV. Munger K: The human papillomavirus E7 oncoprotein and the cellular transcription factor E2F bind to separate sites on the retinoblastoma tumor suppressor protein.J ViroI67:2402-2407. 1993

106. Shirodkar S, Ewen M. DeCaprio JA, Morgan]. Livingston OM, Chittenden T: The transcription factor E2F interacts with the retinoblastoma producr and a pl07-cyclin A complex in a cell cycle-regulated manner. Cell 68:157 - 166, 1992

107. Chellappan S, Kraus VB, Kroger B. Munger K. Howley PM, Phelps WC. Nevins ]R: Adenovirus-EIA, simian virus-40 tumor antigen, and human pa­pillomavirus-E7 protein share the capacity to disrupt the interaction between transcription factor-E2F and the retinoblastoma gene product. Proc Nat! Acad Sci USA 89:4549-4553. 1992

108. Nevins JR: E2F: A link between the Rb tumor supressor protein and viral oncoproteins. Scieucc 258:424 - 429, 1992

109. Kaelin WG, Krek W . Sellers WR. DeCaprio ]A, Ajchenbaum F. Fuchs CS, Chittenden T. Li Y, Farnham Pl. Blanar MA, Livingston OM, Flemington EK: Expression cloning of a cDNA encoding a retinoblastoma-binding pro­tein with E2F-like properties. Cell 70:351 - 364, 1992

110. Girling R, Partridge ]F, Bandara LR. Burden N. Totty NF, Hsuan J]. La Than­gue NB: A new component of the transcription factor DRTF1 / E2F. Nature 362:83-87,1993

Ill. Davies R. Hicks R, Crook~, Morris] , Vousden K: Human papillomavirus type 16 E7 associates WIth a hIstone HI kinase and with p107 through sequences necessary for transfonnation .J Vi,0/67:252 1-2528, 1993

112. Tommasino M, Adamczewski JP, Carlotti F.Barth CF, Manetti R. Contorni M. Cavalieri F. Hunt T. Crawford L: HPV 16 E7 protein associates with the protein kinase p33-cdk2 and cyclin A. Oncogene 8:195-202,1993

113. Barbosa MS, Edmonds C , Fisher C, Schiller JT. Lowy DR. Vousden KH: The region of the HPV E7 oncoprotein homologous to adenovirus Ela and SV40 large T antigen contains separate domains for Rb binding and casein kinase II phosphorylation. EMBOJ 9: 153 - 160. 1990

114. FirzlaflJM. Galloway DA. Eisenman RN, Luscher B: The E7 protein of human papillomavirus type 16 is phosphorylated by casein kinase II. Nelli BioI 1:44-53, 1989

115. Imai Y. Matsushima Y. Takashi S, Terada M: Purification and characterization of human papillomavirus type 16E7 protein with preferential binding capac­ity to the underphosphorylated form of retinoblastoma gene product. J Vi,ol 65:4966-4972.1991

116. Vousden KH. DOlliger J , DiPaoloJA, Lowy DR: TheE7 open reading frame of human papillomavirus type 16 encodes a transforming gene. Ollcogetle Res 3:167-175,1988

117. Tanaka A, Noda T, Yajim. H, Hatanaka M,Ito Y: Identification of a transfonn­ing gene of human papillomavirus type 16.J Vi,0163: 1465 - 1469. 1989

118. Halbert CL, Demers GW, Galloway DA:TheE7 gene of human papillomavirus type 16 is sufficient for immortalization of human epithelial cells. J Viral 65:473-478. 1991

119. Halbert CL, Demers GW. Galloway DA: The E6 and E7 genes of human papiUomavirus type 6 have weak immortalizing activity in human epithelial cells.J ViroI66:2125-2134. 1992

120. Gage JR. Meyers C. Wettstein FO: The E7 proteins of the nononcogenic human papillomavirus type 6b (HPV-6b) and of the oncogenic HPV-16 differ in retinoblastoma protein binding and other properties. J Vi,ol 64:723 - 730. 1990

121. Heck DV, Yee CL, Howley PM, Munger K: Efficiency of binding the retino­blastoma protein correlates with the transforming capacity of the E7 oncopro­teins of the human papillomaviruses. P,oe Nat! Acad Sci 89:4442 - 4446, 1992

122. Munger K, Yee CL, Phelps WC, Pietenpol JA, Moses HL. Howley PM: Bio­chemical and biological differences between E7 oncoproteins of the high- and low-risk human papilloma virus types arc determined by amino-tcnnina.l se­quences. J ViroI65:3943 - 3948, 1991

123. Phelps W C. Munger K. Yee CL. Barnes JA. Howley PM: Structure-function analysis of the human papillomavirus type 16 E7 oncoprotein. J vi,ol 66:2418 -2427.1992

124. Jewers RJ. Hildebrandt p. Ludlow ]W. Kell B, McCance OJ: Regions of human papillomavirus type 16 E7 oncoprotein required for immortalization of human keratinocytes. J Viral 66: 1329 - 1335, 1992

125. Brandsma ]L, Yang ZH, Barthold SW. Johnson EA: Usc of a rapid. efficient inoculation method to induce papillomas by cottonta il rabbit papillomavirus DNA shows that the E7 gene is required. Proc Nat! Acad Sci USA 88:4816-4820. 1991

126. Meyers C, Harry J. Lin Y -L. Wettstein FO: Identification of three transfonning proteins encoded by cottontail rabbit papillomavirus. J Vi,0166: 1655 - 1664, 1992

127. Defeo-Jones D. Voucolo GA. Haskell KM. Hanobik MG. Kiefer OM. McAvoy EM. Ivey-Hoyle M. Brandsm. ]L, Oliff A. Jones RE: Papillomavirus E7 protein binding to the retinoblastoma protein is not required for viral induc­tion of warts. J Virol 67:716 - 725. 1993

128. Groff D. Lancaster WD: Genetic analysis of che 3' early region transformation and replication functions of bovine papillomavirus type 1. Vi,ology 150:221-230.1986

129. Chen SL, Mounts P: Transforming activity of E5a protein of human papilloma­virus type 6 in NIH 3T3 and C127 cells.J Vi,0164:3226 - 3233. 1990

256 ANDROPHY

130. Leptak C, Ramon y, Cajal S, Kulke R, Horowitz BH, Riese DJ, Dotto GP, DiMaio D: Tumorgenic transformation of murine keratinocytes by the E5 genes of bovine papillomavirus type 1 and human papilloma virus type 16. ] ViroI65:7078-7083, 1991

131. Schlegel R, Wade Glass M, Rabson MS. Yang YC: The E5 transforming gene of bovine papilloma virus encodes a small, hydrophobic polypeptide. Scie"ce 233:464 - 467, 1986

132. Burkhardt A, Willingham M, Gay C,Jeang KT. Schlegel R: The E5 oncopro­tein of bovine papillomavirus is oriented asymmetrically in Golgi and plasma membranes. Virology 170:334-339, 1989

133. Burnett S,Jareborg N, Dimaio D: Localization of bovine papillomavirus type-l E5 protein to transformed basal kcratinocytcs and pcrtnissivc differentiated cells in fibropapilloma tissue. Proc Na t! Acad Sci USA 89:5665 -5669, 1992

134. Horwitz BH. Weinstat DL, DiMaio D: Transfonning activity of a 16-amino­acid segment of the bovine papillomavirus E5 protein linked to random se­quences of hydrophobic amino acids.] ViroI63:4515-4519, 1989

135. Settleman J, Fazeli A, Malicki J, Horwitz BH, DiMaio D: Genetic evidence that acute morphologic transformation. induction of cellular DNA synthesis. and focus formation are mediated by a single activity of the bovine papillomavirus E5 protein. Mol Cell Bioi 9:5563 - 5572, 1989

136. Goldstein DJ, Schlegel R: The E5 oncoprotein of bovine papillomavirus binds to a 16 kd cellular protein. EMBO ] 9: 137 -145, 1990

137. Goldstein DJ, Finbow ME, Andresson T, Mclean P.Smith K,Bubb V, Schlegel R: Bovine papillomavirus E5 oncoprotein binds to the 16K component of vacuolar H+-ATPases. Nawr. 352:347 - 349, 1991

138. Goldstein DJ, Kulke R, Dimaio D, Schlegel R: A glutamine residue in the membrane-associating domain of the bovine papillomavirus type 1 E5 onco­protein mediates its binding to a transmembrane component of the vacuolar H+-ATPase.] ViroI66:405-413, 1992

139. Goldstein DJ, Andresson T, SparkowskiJJ, Schlegel R: The BPV-l E5 protein, the 16 kDa membrane pore-forming protein and PDGF receptor exist in a complex that is dependent on hydrophobic transmembrane interactions. EMBO ] 11 :4851 - 4859, 1992

140. Martin P, Vass WC, Schiller JT, Lowy DR, Velu TJ: The bovine papillomavirus E5 transforming protein can stimulate the transfonning activity ofEGF and CSF-1 receptors. Cell 59:21-32, 1989

141. Petti L, Nilson LA, DiMaio D: Activation of the platelet-derived growth factor receptor by the bovine papillomavirus E5 transforming protein. EMBO] 10:845 - 855, 1991

142. Petti L, DiMaio D: Stable association between the bovine papillomavirus E5 transforming protein and activated platelet-derived growth factor receptor in transformed mouse cells. Proc NaIl Acad Sci USA 89:6736-6740.1992

143. Cohen BD, Goldstein DJ, Rutledge L, Vass WC, Lowy DR, Schlegel R, Schiller JT: Transformation-specific interaction of the bovine papilloma virus E5 on­coprotein with the platelet-derived growth factor receptor transmembrane domain and epidermal growth factor receptor cytoplasmic domain.] Virol 67:5303-5311, 1993

144. Cohen BD, Lowy DR, Schiller JT: The conserved C-terminal domain of the BPV E5 oncoprotein can associate with an alpha-adaptin-Iike moleculc: a possible link between growth factor receptors and vi.ral transformation. Mol Cell Bioi 13:6462-6468, 1993

THE JOURNAL OF INVESTIGATIVE DERMATOLOGY

145. Brandsma JL, Z.-H. Y, DiMaio D, Barthold SW, Johnson E, Xiao W: The putative E5 open reading frame of cottontail rabbit papillomavirus is dispens­able for papilloma formation in domestic rabbits. ] Viral 66:6204-6207, 1992

146. Lambert PI', Pan H, Pitot HC. Liem A,Jackson M, Griep AB: Epidermal cancer associated with expression of human papillomavirus type 16 E6 and E7 onco­genes in the skin of transgenic mice. Proc Natl Acad Sci USA 90:5583 - 5587, 1993

147. Griep AB, Herber R,Jeon S, Lohse JK, Dubielzig RR, Lambert PI': Tumorigen­icity by human papillomavirus type 16 E6 and E7 in transgenic mice correlates with alterations in epithelial cell growth and differentiation.] Viro167: 1373-1384,1993

148. Peng X, Olson RO, Christian CB, Lang CM, Kreider JW: Papillomas and carcinomas in transgenic rabbits cartying EJ-ras DNA and Cottontail rabbit papillomavirus DNA.] Viral 67:1698-1701. 1993

149. von Knebel Doeberitz M, OltersdorfT, Schwarz E, Gissmann L: Correlation of modified human papilloma virus early gene expression with altered growth properties in C4-1 cervical carcinoma cells. Cm"er Res 48:3780- 3786. 1988

150. Storey A, Oates D, Banks L, Crawford L, Crook T: Anti-sense phosphorothioate oligonucleotides have both specific and non-specific effects on cells con­taining human papillomavirus type 16. Nllcleic Acids Res 19:4 109-4114, 1991

151. von Knebel Doeberitz M, Rittmuller C, zur Hausen H , Durst M: Inhibition of tumorgenicity of cervical cancer cells in nude mice by HPV E6-E7 anti-sense RNA. Jill] Callcer 51:831 - 834,1992

152. Steele C, Cowsert LM, Shillitoe EJ: Effects of human papillomavirus type 18-specific antisense oligonucleotides on the transformed phenotype of human carcinoma cell lines. Callcer Res 53:2330-2337,1993

153. Bosch 1', Schwarz E, Boukamp P, Fusenig NE, Bartsch D, zur Hausen H: Suppression in vivo of human papillomavirus type 18 E6-E7 gene expression in nontumorigenic HeLa x fibroblast hybrid cells.] Virol 64:4743-4754, 1990

154. Schiffman MH, Bauer HM, Hoover RN, Glass AG, Cadell DM, Rush BB, Scott DR, Sherman ME, Kurman RJ , Wacholder S, Stanton CK, Manos MM: Epidemiologic evidence showing that human papillomavinls infection causes most cervical intraepithelialneoplasia.] Natl Cm"er Illst 85:958-964,1993

155. Kiviat NB. Koutsky LA: Specific human papillomavirus types as the causal agents of most cervical intracpithclial neoplasia: implications for current views and treatment.] Natl Cm".r IlISt 85:934-935,1993

156. Schiffman MH: Recent progress in defining the epidemiology of human papil­lomavirus infection and cervica l neoplasia.] Nat! CIlI"er Illst 84:394-398, 1992

157. Kessis TD, Slebos RJ, Nelson WG, Kastan MD, Plunkett BS, Han SM, Lorincz AT, Hedrick L, C ho KR: Human papillomavirus 16 E6 expression disrupts the p53-mcdiated cellular response to DNA damage. Proc Nat! Acad Sci USA 90:3988-3992, 1993

158. Hashida T. Yasumoto S: Induction of chromosome abnonnalities in mouse and human epidermal keratinocytes by the human papillomavirus type 16 E7 oncogene.] Cell Virol72:1569-1577, 1991