A brief synopsis of the role of human papillomaviruses in cervical carcinogenesis

A brief synopsis of the role of human papillomaviruses in cervical carcinogenesis Mark H. Stoler, MD Charlottesville, Virginia The expansion of our knowledge in the realms of pathology, epidemiology, and molecular biology of human papillomaviruses (HPV) has defined them as the major and best understood class of true human tumor viruses. The interaction of the papillomavirus genome with its host cell produces the majority of cytologic abnormalities at which cervical cancer screening is directed. The epidemiologic pattern of HPV infection accounts for the established association of cervical neoplasia with sexual activity. The molecular interactions of the HPV genome with its host cell suggest a plausible mechanism for its carcinogenic action. This presentation will succinctly review current knowledge of HPV biology to facilitate an understanding of the ctinical significance of this virus. (Am J Obstet Gynecol !996;175:1091-8.)

Key words: H u m a n papillomavirus, papillomavirus, cervix, pathogenesis

An appreciation of h u m a n papillomavirus (HPV) biology is central to an understanding of cervical carcinogenesis and is currently influencing pathologic classifications, as well as clinical management of the lesions in this spectrum. The expansion of our knowledge in the realms of pathology, epidemiology, and molecular biology of these viruses has defined them as the major and best understood class of true h u m a n tumor viruses. It is the purpose of this presentation to briefly review our understanding of HPV biology, with special reference to the mechanisms by which the HPV induces abnormal cervical disease. Papillomaviruses in all vertebrate species induce primarily, albeit not exclusively, squamous epithelial neoplasms of the anogenital tract. In h u m a n beings, more than 70 molecular types have been cloned, some two dozen of which are trophic for the anogenital tract? The circular, 7900 base-pair, double-stranded DNA genome of the papillomaviruses is divided into early and late regions, encoding seven early and two late genetic open reading frames (ORFs), which, through gene splicing, code for all viral gene products, z' ~ In addition, there is a noncoding region, most often referred to as the upstream regulatory region (URR), which contains the sequences regulating the expression of the ORFs. In order of occurrence distal to the URR, E6 and E7 encode proteins that are capable of inducing cell proliferation and transformation. E6 and E7 are also involved in the regulation of viral

From the Division of Anatomic Pathology, University of Virginia Health Sciences Center Charlottesville. Supported in part by PHS grant no. CA 43629. Reprint requests: Mark 14, Stoleg, MD, University of Virginia Health Sciences"Center, SurgicalPathology-Box 214, Charlottesville, VA 22908. Copyright © 1996 by Mosby-YearBook, Inc. 0002-9378/96 $5.00+ 0 6/0/75663

gene expression/replication and the recruitment of the host machinery for these purposes. E1 is involved in genome maintenance and replication. E2 encodes the major transregulatory proteins that interact with the URR. E4 encodes several proteins, some of which bind to and disrupt the cytoplasmic keratin network, producing in appropriately differentiated cells, the cytoplasmic halo referred to as koilocflosis. E5 seems also to be involved in cell transformation because it interacts with cell membrane growth factor receptors. It also contains 3' regulatory and polyadenylation sequences for all the E region genes. ORFs L2 and L1 encode the minor and major viral capsid proteins, respectively (Fig. 1). The two dozen or so viruses making up the mucosotropic group of HPVs may be broadly classified into those with a low risk of lesion progression to cancer versus those with a moderate to high risk. 4Viruses classified as low risk are defined by the fact that they are almost never found in invasive cancers. In contrast, high-risk viruses are those that are most often found in invasive cancer. It should be noted that infection with a high-risk virus does not imply the inevitable development of cancer. It should also be pointed out that the molecular epidemiologic pattern of most of the moderate- or intermediate-risk viruses is incompletely developed because of a relative lack of probes and the recent description of members of this group. The four mucosotropic viruses, HPVs 6, 11, 16, and 18 form the prototypes for the high- versus low-risk groupings, and together account for approximately two thirds of the HPV-associated anogenital neoplasms. ~' ~ Types 6 and 1 l primarily cause benign exophytic genital warts or condylomata acuminata. These are the viruses present in greater than 90% of condylomas with about two thirds caused by HPV 6 and one third by HPV 11. They are also associated with low-grade squamous intraepithelial le1091

1092

Stoler

October 1996 AmJ Obstet Gynecol

@ circular viral episome

I E71

E1

I

I E5 J L2

I

o.i~oo

o~,~oo

SIMPLE LINEAR REPRESENTATION OF THE CIRCULAR HPV GENOME ORFS DISTRIBUTEDBASEDON 3 POSSIBLETRANSLATIONFRAMES

O

+

lost chromosom(

circular viral episome

me~=~=;~E2

E45

L2

L1

U R R E 6 E7

E1

E2'~,~===,mr---,~,-~=

Fig. 1. Organization of HPV genome and pattern of HPV integration. All papillomaviruses have similar genomic organization. In low-grade lesions the circular episomaI form is the dominant form. However, in most cancers, integration occurs. Integration of HPV viral episome into the host chromosome is a random event in terms of location in the host genome. Howevm; the viral genome is usually disrupted in the E2 0RF, which is believed to lead to loss of E2 proteins' transregulatory functions. sions (LSIL) and are only rarely associated with highgrade squamous intraepithelial lesions (HSIL) or invasive squamous cancers. Related viruses that p r o d u c e a similar s p e c t r u m in the cervix are HPVs 42, 43, 44. In contrast, HPV16 is the most likely virus to infect the uterine cervix and is closely associated with the entire range of intraepithelial and invasive squamous neoplasia, as well as the less c o m m o n l y observed cervical glandular neoplasia. T h e m o d e r a t e - to high-risk types most closely related to HPV 16 include types 31, 33, 35, 52, and 58. HPV 18 is the o t h e r cancer-associated prototype that is also most associated with n o n s q u a m o u s cervical neoplasms. T h e viruses most related to type 18 include types 45 and 59, as well as type 39. Morphologically similar lesions at different mucosal sites are usually caused by the same mucosal viruses. Thus laryngeal and conjunctival papillomas that are pathologically and biologically equivalent to a condyloma are most often caused by HPVs 11 and 6. In contrast; the b o w e n o i d dysplasias of the vulva, penis, anus, and oral cavity- are most often associated with HPV 16 (Table I). As will b e c o m e clear, all HPV types, even the high-risk viruses, must induce the pathologic equivalent of a wart, condyloma, or LSIL, because this is the disease that supports viral replication and virion production.

N u m e r o u s epidemiologic studies have linked cervical cancer to sexual behavior, suggesting a venereal cause. 7-n The finding that either female promiscuity or being a m o n o g a m o u s female p a r t n e r of a promiscuous p a r t n e r confers increased risk for cervical cancer supports the c o n c e p t of a sexually transmitted agent. T h e strongest epidemiologic risk factor is the n u m b e r of sexual partners. Young age at first intercourse and early age of parity, b o t h in part markers of promiscuit3; also seem to confer risk. Beginning sexual activit); however, during adolescence may also confer risk on the basis of an altered h o r m o n a l environment. In most epidemiologic studies, after sexual factors are controlled for, cigarette smoking still remains a risk factor. M o r p h o e p i d e m i o l o g i c studies demonstrate that the precursors of cervical cancer precede invasive cancer with LSIL having the highest prevalence in patients in their early twenties, HSIL in the late twenties and early thirties and invasive cancer in w o m e n aged 40 to 50 years. In m o r e recent cohorts there is a trend toward earlier age for each of these stages. T h e pathologic study of the uterine cervix is very m u c h the histopathologic and cytopathologic study of papillomavirus-associated neoplasia. In 1956 Koss and Durfee 1~ c o i n e d the t e r m hoilocytotic atypia to describe the cells

Volume 175, Number 4, Part 2 AmJ Obstet Gynecol

Stoler

1093

Table I. Association of HPV types with various clinical lesions (types in bold are most common or prototypes) Clinical lesion

HPV types

Common warts (verruca vulgaris) Plantar warts Flat warts Epidermodysptasia verruciformis Butcher's warts Condyloma acuminatum (genital warts) Laryngeal papillomatosis LSIL or equivalent HSIL [nvasive cervical cancers Bowenoid high grade dysplasia Invasivesquamous cancer of other anogenital sites

2,4 1 3, 10, 28, 41 5, 8, as well as 9, 12, 14, 15, 17, 19, 20-25, 36-38, 47 and 49 7 6, 11, 42, 43, 44 11, 6 Essentiallyall mucosotropic types (>24 types) Same as invasivecervical cancer 16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59, 66, 67 & 68 Mainly 16 Mainly 16

derived from fiat '°war~ lesions" of the cervix because of their histologic resemblance to skin warts. Koilocytotic atypia is often referred to as being pathognomonic for HPV infection or as HPV cyropathic effect because it is the cytologic abnormality most often described as being associated with the virus. In fact, cells with this change are those in which virion is most often detected and are highly correlated with productive HPV infection. Howevm, wtologic or histologic absence of KA does not in any way imply absence of HPV, or more specifically absence of pathogenic HPV gene expression. 1~KA, which u n d e r the Bethesda classification is considered a LSIL, is the most common definite abnormality in cytologically screened populations today, being present in 1% to 4% of Papanicolaou ("Pap") smears. In 1976 Meisels and Fortin 14and Purola and Savia1~ recognized the existence of fiat and inverted condytomas of the uterine cervix and reported that these lesions may not only be found in the spectrum of cervical intraepithelial neoplasia (CIN) but also were associated with the h u m a n papillomavirus. These landmark studies and numerous subsequent morphologic studies have contributed greatly to our understanding of the role of these viruses in carcinogenesis. Supporting morphologic data come from studies of the distribution of HPV capsid protein or HPV virions in iutraepithelial neoplasias. With a broadly reactive, group-specific antisera against the L1 protein, the papillomavirus capsid antigen, the expression of which is a highly differentiation-dependent p h e n o m e n o n , is present in 50% to 60% of condylomas or LSILs with a decreasing frequency as cytohistologic grade increases.~S Similarly, if one searches cell nuclei for virions, using transmission electron microscopy, there is an inverse correlation with cytologic lesion grade for the detection of virus. ~7 These epidemiologic and pathologic insights have been greatly bolstered by data derived from modern molecular biologic techniques. Molecular detection methods can be applied in a variety of ways. Most commonly a cellular sample is analyzed for the presence of HPV DNA by dot blot, Southern blot, or some analogous technique. More recently, sensitivity is improved by coupling these methods to an amplification technology such as the poly-

merase chain reaction. 4' is, 19Inherent in these methods is the destruction of the lesion; hence, the results are usually correlated with some other i n d e p e n d e n t morphologic observation. 2° These types of studies have been strongly complemented by many direct analyses that have used in situ hybridization to directly demonstrate the presence of HPV DNA or messenger RNA in defined groups of diseases. 21 As noted above, there are approximately 25 HPV types that infect the genital tract. All can be found in LSILs or in samples from cytologically "normal" women, and, although no single type predominates, HPV 16 is probably the single most common virus at the cervix. = Whereas the prevalence of HPV DNA in LSILs is in excess of 90%, the same can be said for HSIL and invasive squamous carcinoma of the cervix. However, the type spectrum in the high-grade lesions is much more restricted, with HPV types 16, 18, 31, and 45 accounting for almost 80% of the invasive cancers. Squamous cell carcinomas account for -80% of cervical cancers, with the remainder being made up of primarily endocervical adenocarcinomas and a small n u m b e r of the fortunately rare, small-cell neuroendocrine carcinomas. 23Studies with sensitive methods that analyze these nonsquamous cancers and their precursors also demonstrate a very high prevalence of HPV DNA. The virus most closely related to progressive cervical squamous neoplasia is HPV 16. 24Although it accounts for fewer cervical infections, HPV 18 is more consistently associated with adenocarcinomas and small-cell neuroendocrine cancers of the cervix and less frequently with invasive squamous cancer. 25-2sThe absolute prevalence of some of the more recently described types (e.g., HPVs 31, 33, 35, 39, 42, 43, 44, 45, 51, 52, 56, and others) may be underestimated because they have not been generally available for large-scale screening. Thus HPV genetic material is present in more than 90% of premalignant and malignant squamons lesions of the uterine cervix. Detractors of the HPV-cancer link point out that the finding of HPV DNA in a lesion is insufficient to explain pathogenicit3: Unlike any other candidate agent, it is clear that HPV DNA is not only present, but that in every disease linked to these viruses, HPV messenger RNA is

1094

$toler

October 1996

AmJ ObstetGynecol

expressed in these lesions.29 The presence of viral RNA and protein expression provides the essential pathogenic link. The patterns of viral mRNA expression vary with morphologic diagnosis in a tightly regulated and differentiation-dependent m a n n e r (vida infra). Thus in lowgrade lesions all viral genes are expressed as a manifestation of vegetative viral replication. In contrast, in HSIL and invasive cancer, there is a restricted pattern of viral gene expression involving primarily E6 and E7. Just as naturally occurring malignant lesions have been found to contain HPV DNA and express viral E6 and E7 mRNAs, so too have many long established cervical carcinoma cell lines, such as HeLa, SiHa and CaSki, been found to harbor integrated HPV 16 or 18 DNAs from which the transforming E6 and E7 regions are actively transcribed. These observations further strengthen the relationship between these HPVs and carcinogenesis. Whereas the presence of actively transcribed HPV DNA within lesions has established a strong molecular association, in vitro cell transformation experiments have pointed to an active role for these viruses in carcinogenesis?~32 HPV 16, 18, 31, and 33, but not HPV 6 or 11, are capable of transforming primary baby rat kidney epithelial cells in collaboration with an activated cellular oncogene, Ha-ras, thus mirroring general concepts of multistep carcinogenesis. HPV 16 DNA can immortalize cup tured primary foreskin keratinocytes or primary cervical cells in culture. Although not inhibitory to the formation of artificial epithelia by keratinocytes cultured on collagen rafts, HPV 16 can prevent cellular differentiation, thus inducing in these artificial epithelia morphologic transformation that mimics CIN. It is noteworthy that in this system the transformed phenotype is not apparent until the cells have been passaged many generations, again suggesting the need for additional genetic events and mimicking the long progression times seen in naturally occurring lesions. Because there seems to be a correlation between specific HPV types and the potential for clinical malignancy, there also may be an association between the physical state ofHPV DNAwithin the cell and the malignant potential of the associated epithelial proliferation. 33 In benign HPVinfected lesions, the viral DNAs exist as extrachromosomal plasmids, mostly as monomeric circular molecules. However, in many cancers, HPV DNAs are integrated into host chromosomes. Viral integration most frequently disrupts the E2 0RF, which encodes the transcription regulatory proteins. Loss of these regulatory proteins is believed to be the basis for potential dysregulation of the expression of the transforming E6 and E70RFs (Fig. 1). Concurrent with the revelations of HPV biology, there has been an explosion of information about the roles of cellular oncogenes in carcinogenesis. Several classes of oncogenes, including those encoding growth factors, growth factor receptors, GTP binding proteins, protein

kinases, and DNA binding proteins have been shown to be relevant to the control of cell growth. C-myc, as well as c-Ha-ras amplification can be documented in some cervical cancers and correlates with advanced clinical stage at the time of analysis.34' 3~As in other tumor systems, the inactivation of tumor suppressor genes, or antioncogenesas they are called now, may also be significant. Fusion of HPV 18 expressing HeLa cells with normal h u m a n fibroblasts or keratinocytes results in the repression of the malignant phenotype of the HeLa cell? 6 O n transplantation into nude mice, the loss of chromosome l l f r o m the hybrid cells results in the reversion to malignant phenotype, suggesting the presence of a tumor suppressor gene at this site? 7 The first of the antioncogenes to be characterized was the h u m a n retinoblastoma gene. 3a It is either completely absent or has significant deletions in tumors from patients with retinoblastoma, breast cancer, and in some other epithelial tumors. It has been reported that the transforming E7 protein of HPV 16 has structural and functional similarities to the E1A antigen of adenovirus and the large T antigen of SV40, all of which bind to the retinoblastoma protein and in doing so functionally inactivate the retinoblastoma protein? 9 Similar complexing and inactivation of the p53 suppressor gene by high-risk but not low-risk HPV E6s have been demonstrated. ~° Thus a molecular model for HPV-induced carcinogenesis is beginning to emerge, involving the interaction of HPV gene products with what is recognized to be a tightly controlled network of cellular oncogenes and antioncogenes, which regulate cell proliferation (Fig. 2). Model of HPV-induced carcinogenesis 41

It is clear that papillomaviruses must infect the "reserve, basal, or stem" cell population of the cervical transformation zone, cells that have the potential to differentiate along squamous, glandular, or neuroendocrine lines. In cells committed to squamous differentiation, there is an orderly program of maturation throughout the epithelial thickness at both the morphologic and molecular level. The only cells capable of cell division are the basal or parabasal cells. If HPV DNA is present in morphologically normal basal cells, papillomavirus gene expression is inhibited in such cells to essentially maintenance levels, because productive HPV gene expression is tightly regulated and permitted only in cells that have begun squamous differentiation, which concurrently have lost the ability to divideY-44The eventual outcome of this coordination between host and virus is the production of a histologically low-grade squamous intraepithelial lesion. Such lesions can regress or maintain themselves for extended periods. Continuing this scenario, morphologic progression, which is correlated with cellular hyperproliferation, occurs if the coordinate link between differentiation and viral early gene expression is lost. Possible mechanisms could include viral integration

Stoler

Volume 175, Number 4, Part 2 Am J Obstet Gynecol

~m p16" ..... " "--" 'C,~..~ CDKs

,,.,

/ ~

mdm2

f.~-f

RB-p ,

J,

I,

cyclins

1095

SV4OT

~" ~." cE1A

t,

"~7~ ''~'-'''E2F/Ti~. ~%

i

jp~cychns

i E6

G1

S

G2

M

[ ] positive effect ~.~ negative effect Fig. 2. Conceptual model of the interaction of HPV with cell cycle control checkpoints. Progression through the cell cycle is essential for cell division, a hallmark of neoplasia. Many proteins listed are members of protein families and also have more than one function, so this diagram could be much more complex. Cyclins (of which there are many family members) complex with their corresponding cyclindependent kinases (cdks) to influence progression through various cell cycle stages. For example, D-cyclins act during the G1 to S transition. Cyclin/cdk complexes in part influence the phosphorylation state of the retinoblastoma protein (RB). Phosphorylated retinoblastoma protein releases transcription factor E2F, which can have wide ranging effects on cellular transcription, including activation of many genes involved in DNA synthesis, including cyclins. Some cyclins, such as D 1, competitively bind to the RB pocket as do viral oncoproteins, such as HPV E7, simulating or complementing the effect of phosphorylation of the RB protein. Cellular p53 also acts as repressor or controller of cell division. It does this in part by stimulating expression of p21, an inhibitor ofcdk expression, pl 6 is also a cdk inhibitor. Thus viral oncoprotein (e.g., HPV 16 E6) mediated degradation of p53 removes some functional brakes on cell division.

or m u t a t i o n disrupting E2-mediated regulation such that the viral o n c o g e n e s E6 and E7 are inappropriately expressed in a p o p u l a t i o n of cells that retain the capacity to divide, thereby initiating cell proliferation and m o r p h o l o g i c progression. These changes could be promoted (albeit with relatively low frequency) by smoking, o t h e r viruses, or r a n d o m m u t a t i o n and are biologically manifest by the latency and relative rarity of high- versus low-grade lesions. Progression clearly occurs most frequently, albeit n o t exclusively, with high-risk viral types, and results most often in a m o d e r a t e or severe squamous dysplasia. T h e hyperproliferative basaloid cells of these lesions are t h e n at m u c h greater risk for the acquisition of additional genetic errors, (perhaps u n d e r the influence of the same external m u t a g e n s or host genetic predisposition), which further promotes the d e v e l o p m e n t of the fully malignant phenotype, most often an invasive squamous cell carcinoma. T h e different subtypes of squamous cancer are u n d o u b t e d l y related to the multistep and somewhat rand o m nature of the process, and the p r o p o r t i o n of different types just reflects the relative likelihood of different genetic pathways to a "successful" cancer. For glandular

neoplasms, similar considerations apply. Reserve cells, which are already c o m m i t t e d to glandular differentiation, are, by principles discussed above, n o t g o i n g to be productive of virus. T h e usual result is probably an abortive or latent infection. Much less frequently, dysregulation of viral early g e n e expression occurs in these nonpermissive cells. This leads to hyperproliferative lesions of glandular cells that we morphologically recognize as a d e n o c a r c i n o m a in situ. It seems that HPV type 18 is m o r e successful in i n d u c i n g transformation in glandular cells than is HPV 16. D e p e n d i n g on the genetic switches that may eventually a c c o m p a n y this class of cell proliferation, the o u t c o m e may be an invasive a d e n o c a r c i n o m a , again most often of the endocervical type, but less freq u e n d y of o t h e r than endocervical differentiation. Essentially identical a r g u m e n t s can be m a d e for the developm e n t of small-cell n e u r o e n d o c r i n e carcinomas. These minors are also highly HPV associated almost exclusively with type 18, and their low incidence probably reflects the relative a b u n d a n c e of a susceptible n e u r o e n d o c r i n e c o m m i t t e d precursor cell p o p u l a t i o n and the rarity of "successful" viral i n d u c t i o n of cell proliferation in such

1096

Stoler

cells. None of the above precludes alternative pathways of carcinogenesis unrelated to HPV. For example, in the post-diethylstilbesterol era, many clear cell cancers of the cervix in young women may have developed along an HPV-independent path. And certainly, morphologically and biologically similar clear cell tumors occur in the upper female genital tract without any evidence of a relationship to HPV. In fact, cellular antioncogene mutations have been shown to be capable of substituting for the genetic effects of the viral oncogenes E6 and E7, in the rare HPV-negative cervical cancer. But in the susceptible uterine cervix, the ubiquity of HPV infection makes it highly likely that the initiating steps in cervical carcinogenesis are the direct result of an HPV infection. Fortunately, progression is a relatively rare occurrence when compared to the high prevalence of the virus. Clearly, some of the above is speculative, but all of it is entirely consistent with the wealth of scientific data that continues to accumulate regarding HPV-induced neoplasia.

The clinical utility of HPV detection and typing techniques Given the above considerations, some comment regarding the utility of detecting and typing HPV nucleic acids seems warranted. These comments will focus on the setting of an abnormal Pap smear result demonstrating possible preinvasive disease, the most common clinical practice situation in which questions about HPV testing arise. ~rhether to use HPV tests is the subject of ongoing controversy, as well as technologic evolution.19' 4s-48When confronted with a cervical biopsy or Pap smear result, the basic question for the pathologist is whether the morphologic changes present warrant a diagnosis within the spectrum of cervical neoplasia. In our experience, many pathologists have difficulty in recognizing lesions in this continuum.49Particularly" problematic is the separation of normal or equivocal changes from diagnostic lesions or, in current terms, changes of atypical squamous cells of u n d e t e r m i n e d significance (ASCUS) from true squamous intraepithelial lesions. There is a tendency to overcall any cell with nonspecific halos around the nucleus a koilocyte. A common mind-set seems to be that, because HPV infections are ubiquitous, one should try to be as diagnostically sensitive as possible. The danger, of course, is the overdiagnosis of lesions that can lead to overtreatment. Another argument for a confirmatory HPV test might be that, by making a diagnosis of an HPV-associated lesion, one has diagnosed a sexually transmitted disease, and this should be confirmed by an i n d e p e n d e n t method for medicolegal reasons. In addition, HPV typing information might be clinically useful for prognosis and therapy. For instance, HPV 18 may be associated with rapidly progressive neoplasia and some aggressive forms of cervical cancer. 5° In contrast, HPV 6 and 11 are almost never associated with cancer. Such a distinction might argue for differential therapy based on the presence of "high-risk" versus "low-risk" viruses.

October 1996 AmJ ObstetGynecol

Do the currently available tests help us with these problems? The marketers of HPV tests certainly do not focus on the false-negative rates that can have profound effects on clinical utility. In the setting of unequivocal morphologic abnormality, a pathologic lesion within the spectrum under discussion must be assumed to be HPV related unless proven otherwise. To prove otherwise is not so simple to do, given the issues of test sensitivity and the great heterogeneity of viruses and their proven high prevalence in unequivocal squamous intraepithelial lesions. Thus the pathologist's problem is reduced to the application of morphologic criteria that will allow histologic diagnosis to correlate with clinical reality. For example, "koilocytosis" is only highly correlated with the finding of viral DNAwhen well-defined perinuclear halos (koilocytosis) are present together with definite nuclear enlargement, alteration of N / C ratio and hyperchromatism (i.e., atypia). If such strict criteria are used, then much of the need for in situ hybridization testing is obviated. Furthermore, the question of use of HPV testing as an aid in borderline or low-grade cases is confounded by the limitations of the current molecular technology. In randomly selected material, roughly 40% to 70% of borderline lesions that are truly HPV associated will be positive with these assays.< 49-01This low sensitivity is a function of several factors. It is now recognized that in low-grade lesions there is tremendous HPV-type heterogeneity, yet most kits are limited in terms of their probe repertoire. In addition, even when a viral t?qoe is homologous to an available probe, the viral DNA needs to be present in sufficient concentration, as well as in a chemical state that permits hybridization to occur. In highgrade lesions, fewer types are present, so DNA heterogeneity is less of a problem. However, here DNA copy number is often greatly reduced such that assay sensitivity is still a major issue. In the setting of equivocal morphol0gic changes (atypical squamous cells of u n d e t e r m i n e d significance), a different set of problems apply. Under the best of conditions, approximately 25% to 50% of equivocal lesions will be HPV positive; with lower sensitivity assays this figure will be even less. To argue that an equivocal lesion is clinically significant because it is HPV positive by in situ hybridization or some other technique is to deny the significance of the weak morphologic findings. Conversely, the presence of a negative HPV test result only provides false assurance that the lesion is truly HPV negative because of previously cited issues of DNA heterology, copy number, and latency. For example, consider a biopsy demonstrating nonspecific halos or parakeratosis that is HPV transcriptionally silent but is DNA positive by an extremely sensitive amplification technique. The morphologic abnormality might be due to the presence of virus, but it may just as likely be a nonspecific reaction to some other stimulus. DNA positivity alone does not prove pathogenetic involvement (see above). What about the prognostic significance of HPV type? In

Stoler 1097

Volume 175, Number 4, Part 2 AraJ Obstet Gynecol

general terms the o f t e n - q u o t e d high-risk versus low-risk d o g m a may make some sense, but u n d e r scrutiny it does n o t h o l d up very well. HPV 16 is the virus that most c o m m o n l y infects the cervix. Yet the high frequency of HPV infection relative to clinical, particularly high-grade disease, dilutes the significance of an infection by a highrisk type in a low-grade lesion. More significantly, the reservoir of high-risk viruses is u n d o u b t e d l y in patients without lesions, and the factors triggering cycling between a s y m p t o m a t i c / l a t e n t infection and a b n o r m a l disease are largely unknown. The overall frequency of lesion progression and regression is probably no different today than it was before we knew anything about HPVs. Although it seems clear that lesions associated with highrisk viruses are the ones that may m o r e frequently progress, the fact is that most lesions do n o t progress. T h e p a t h o g e n e t i c m e c h a n i s m s discussed earlier support the c o n c e p t that, although strongly linked to neoplastic development, HPV infection and expression are probably onIy two of several steps necessary for p r o d u c t i o n of the full malignant phenotype. F u r t h e r m o r e , occasional cases of high-grade lesions with low-risk viruses do occur. In the individual case, one has no way of knowing which lesions will progress. H e n c e , it would seem p r u d e n t to design t r e a t m e n t on the basis o f lesion grade and clinical extent without regard for HPV type. ~3' ~2. a3 This is especially true because there are no specific antiviral therapies for papitlomaviruses either in general or by type. T h e r a p y is largely ablative, and there is no evidence that ablative therapy effects a virologic cure. Rather, ablation eradicates clinical lesions, which then allows the patient to be placed u n d e r l o n g - t e r m surveillance. However, these issues are far f r o m settled and as evidence, there is currently a multimillion dollar 5-year National Cancer Instit u t e - s p o n s o r e d trial directed at resolving the appropriate m a n a g e m e n t of low-grade cervical diseases and the role of HPV testing in their m a n a g e m e n t . O u r knowledge regarding cervical carcinoma and its pathogenesis has grown markedly. Research with molecuiar m e t h o d s for the detection and analysis of HPVs has b e e n instrumental in improving o u r u n d e r s t a n d i n g to the p o i n t that today we can clearly refocus our Inorphologic criteria. These lesions should be classified and r e p o r t e d according to the degree of specific m o r p h o l o g i c abnormality present irrespective of the presence or absence of histologic or o t h e r evidence of HPV infection. The descriptive r e p o r t i n g systems for precancerous lesions of the cervix, dysplasia/carcinoma in situ or CIN r e m a i n appropriate. T h e presence Or absence o f m o r p h o l o g i c evidence of viral infection, that is, "koilocytosis," m a y b e m e n t i o n e d as part of the diagnosis but should n o t modify the clinical m a n a g e m e n t of the patient. O n the basis of currently available information, w o m e n with any of the lesions in this disease s p e c t r u m should receive appropriate clinical evaluation, treatment, and follow-up irrespective of evid e n c e for the presence or absence of HPV or a specific viral genotype. The c u r r e n t challenge for the diagnostic

pathologist is to use established m o r p h o l o g i c criteria that facilitate specific diagnosis. REFERENCES

1. de Villiers EM. Human pathogenic papillomavirus types: an update. Curr Top Microbiol Immunol 1994;186:1-12. 2. Broker TR. Structure and genetic expression of papillomaviruses. Obstet Gynecol Clin North Am 1987;14:329-48. 3. HoMey PM. Papillomavirinae and their replication, In: Field BN, Knipe DM, editors. Field's virology. New York: Raven Press, 1990;2:chap 58. 4. Lorincz AT, Reid R,Jenson AB, Greenberg MD, Lancaster W, Kurman RJ. Human papillomavirus infection of the cervix: relative risk associations of 15 common anogeuital types. Obstet Gynecol 1992;79:328-37. 5. Van Ranst MA, Tachezy R, Delius H, Burk RD. Taxonomy of the human papillomaviruses. Papillomavirus Report 1993;3: 61-5. 6. Wilbur DC, Reichman RC, Stoler MH. Detection of infection by human papillomavirus in genital condylomata: a comparison study using immunocytochemistry and in situ nucleic acid hybridization. AmJ Clin Patbol 1988;89:505-10. 7. Munoz N, Bosch FX, Shah KV, Meheus A. The epidemiology of human papillomavirus and cervical cancer. New York: Oxford University Press, 1992. IARC Scientific Publication No. 119. 8. Koutsky LA, Galloway DA, Holmes KK. Epidemiology of genital human papillomavirus infection. Epidemiol Rev 1988;10:122-63. 9. Schiffman MH. Recent progress in defining the epidemiology of human papillomavirus infection and cervical neoptasia. JNCI 1992;84:394-8. 10. Schiffmaal MH; Bauer HM, Hoover RN, et al. Epidemiologic evidence showing that human pap01omavirus infection causes most cervical intraepithelial neoplasia.jNCI 1993;85: 958-64. 11. Schiffman MH. Epidemiology of cervical human papillomavirus infections. Curt Top Microbiol Immunol 1994;186:5581. 12. Koss LG, Din-fee GR. Unusual patterns ofsquamous epithelium of the uterine cervix: cytologic and pathologic study of koilocytotic atypia. Ann NYAcad Sci 1956;63:1245-61. 13. Bonfiglio TA, Stoler MH. Human papillomavirus and cancer of the uterine cervix [editorial]. Hum Pathol 1988; 19:621-2. 14. Meisels A, Fortin R. Condylomatous lesions of the cervix and vagina: I--cytologic patterns. Acta Cytol 1976;20:505-9. 15. Purola E, Savia E. Cytology of gynecologic condyloma acuminatum. Acta Cytol 1977;21:26-31. 16. Kurman RJ, Sanz LE, Jenson AB, Perry S, Lancaster WD. Papillomavirus infection of the cervix I: correlation of histology with viral structural antigens and DNA sequences. Int Gynecol Pathol 1982;1:17-28. 17. Toki T, Qikawa N, Tase T, et al. hnnmnohistochemical and electron microscopic demonstration of human papillomavirus in dysplasia of the uterine cervix. Tohoku Exp Med 1986;149:163-7. 18. Lorincz A. The detection of genital human papillomavirns infection using polymerase chain reaction [letter]. JAMA 1991;265:2809-10. 19. Lorincz A. Detection of human papillomavirus DNAwithout amplification: prospects for clinical utility. Lyon, France: IARC Scientific Publications, 1992;119:135-45. 20. Stoler MH, Broker TR. In situ hybridization detection of human papillomavirus DNAs and messenger RNAs in genital condylomas and a cervical carcinoma. Hum Pathol 1986; 17:1250-8. 21. Stoler MH. In situ hybridization [review]. Clin Lab Med 1990;10:215-36. 22. Bergeron C, Barrasso R, Beaudenon S, Flamant P, Croissant O, Orth G. Human papillomaviruses associated with cervical intraepithelial neoplasia - great diversity" and distinct distribution in low-grade and high-grade lesions. Am J Surg Pathol 1992;16:641-9. 23. Kurman R J, Norris H J, Wilkinson EJ. Tumors of the cervix,

1098

24. 25.

26.

27. 28.

29.

30. 31.

32. 33.

34.

35.

36. 37.

38. 39.

Stoler

vagina and vulva. In: Rosai J, editor. Atlas of tumor pathology. 3rd series, fascicle 4. Washington, DC: Armed Forces Institute of Pathology, 1992. Campion MJ, McCance DJ, Cuzick J, Singer A. Progressive potential of mild cervical atypia: prospective cytological, colposcopic, and virological study. Lancet I986;2:237-40. Tase T, Sato S, 'Wada Y, Yajima A, Okagaki T. Prevalence of human papillomavirus type 18 DNA in adenocarcinoma and adenosquamous carcinoma of the uterine cervix occurring in Japan. Tohoku Exp Med 1988;156:47-53. Duggan MA, BenoitJL, McGregor SE, NationJG, Inoue M, Stuart GC. The human papillomavirus status of 114 endocervical adenocarcinoma cases by dot blot hybridization. Hum Pathol 1993;24:121-5. Farnsworth A, Laverty C, Stoler MH. Human papillomavirus messenger RNA expression in adenocarcinoma in situ of the uterine cervix. Int Gynecol Pathol 1989;8:321-30. Stoler MH, Mills SE, Gersell DJ, Walker AN. Small-cell neuroendocrine carcinoma of the cervix. A human papillomavirus type 18-associated cancer. Am Surg Pathol 1991;15:2832. Stoler MH, Rhodes CR, Whitbeck A, Wolinsky SM, Chow LT, Broker TR. Human papillomavirus type 16 and 18 gene expression in cervical neoplasias. Hum Pathol 1992;23:11728. Storey A, Banks L. Human papillomavirus type 16 E6 gene cooperates with EJ-ras to immortalize primary mouse cells. Oncogene 1993;8:919-24. Matlashewski G, Schneider J, Banks L, Jones N, Murray A, Cravcford L. Human papillomavirus type 16 DNA cooperates with activated ras in transforming primary cells. EMBO J 1987;6:1741-6. McCance DJ, Kopan R, Fuchs E, Laimins LA. Human papillomavirus type 16 alters human epithelial cell differentiation in vitro. Proc Natl Acad Sci USA 1988;85:7169-73. Cullen AP, Reid R, Campion M, L:orincz AT. Analysis of the physical state of different human papillomavirus DNAs in intraepithelial and invasive cervical neoplasm. Virology 1991;65:606-12. Rion GF, Bourhis J, Le MG. The c-myc proto-oncogene in invasive carcinomas of the uterine cervix: clinical relevance of overexpression in early stages of the cancer. Anticancer Res 1990;10:1225-31. Riou G, Le MG, Favre M, Jeannel D, Bourhis J, Orth G. Human papillomavirus-negative status and c-myc gene overexpression: independent prognostic indicators of disrant metastasis for early-stage invasive cervical cancers. JNCI 1992;84:1525-6. Stanbridge EJ, Der CJ, Doerson cJ, et al. Human cell hybrids: analysis of transformation and tumorigenicity. Science 1982;215:252-9. Bosch FX, Durst M, Schwarz E, Boukamp P, Fusenig NE, Zur HH. The early genes E6 and E7 of cancer associated human papilloma viruses as targets of tumor suppression? Behring Inst Mitt 1991;89:108-21. Lee W-H, Shew J-Y, Hong FD, et al. The retinoblastoma susceptibility gene encodes a nuclear phosphoprotein associated with DNA binding activity. Nature 1987;329:642-5. Phelps WC, Yee CL, Mfinger K, HoMey PM. The human

October 1996 AmJ Obstet Gynecol

40.

41.

42.

43.

44.

45. 46.

47. 48.

49.

50.

51.

52. 53.

papillomavirus type 16 E7 gene encodes transactivation and transformation functions similar to those of adenovirus E1A. Cell 1988;53:539-47. Scheffner M, Werness BA, HuibregtseJM, Levine AJ, HoMey PM. The E6 oncoprotein encoded by human papillomavirus types 16 and 18 promotes the degradation of p53. Cell 1990; 63:1129-36. Stoler MH. The biology of human papillomaviruses and their role in cervical carcinogenesis. In: Bonfiglio TA, Erozan YS, eds. Gynecologic cytopathology. Philadelphia: Lippincott-Raven, 1996:51-72. Stoler MH, Wolinsky SM, Whitbeck A, Broker TR, Chow LT. Differentiation-linked human papiliomavirus types 6 and 11 transcription in genital condylomam revealed by in situ hybridization with message-specific RNA probes. Virology 1989;172:331-40. Demeter LM, Stoler MH, Broker TR, Chow LT. Induction of proliferating cell nuclear antigen in differentiated keratinocytes of human papillomavirus-infected lesions. Hum Pathol 1994;25:343-8. Dollard SC, Wilson JL, Demeter LM, et al. Production of human papitlomavirus and modulation of the infectious program in epithelial raft cultures. Gene Develop 1992;6: 1131-42. Crum CE Genital papiltomaviruses and related neoplasms: causation, diagnosis and classification (Bethesda). Mod Pathol 1994;7:138-45. Schiffman MH, Schatzkin A. Test reliability, is critically important to molecular epidemiology: an example from studies of human papillomavirus infection and cervical neoplasia. Cancer Res 1994;54:1944s-1947s. Schiffman MH. Validation of hybridization assays: correlation of filter in situ, dot blot and PCR with Southern blot. Lyon, France: IARC Scientific Publications, 1992;119:169-79. Schiffman MH, Bauer HM, Lorincz AT, et al. Comparison of Southern blot hybridization and polymerase chain reaction methods for the detection of human papillomavirus DNA. Clin Microbiol 1991;29:573-7. Sherman ME, Schiffman MH, Lorincz AT, et al. Toward objective quality assurance in cervical cytopathology: correlation of cytopathologic diagnoses with detection of highrisk human papillomavirus types. AmJ Clin Pathol I994;102: 182-7. Kurman R J, Schiffman MH, Lancaster WD, et al. Analysis of individual human papillomavirus types in cervical neoplasia: a possible role for type 18 in rapid progression. AmJ Obstet Gynecol 1988;159:293-6. Masih AS, Stoler MH, Farrow GM, Johansson SL. Human papillomavirus in penile squamous cell lesions: a comparison of an isotopic RNA and two commercial nonisotopic DNA in situ hybridization methods. Arch Pathol Lab Med 1993;117:302-7. Dalence CR. Diagnostic and Therapeutic Technology Assessment(DATrA): human papillomavirus testing in the management of cervical neoplasia. JAMA 1993;270:2975-81. Kurman RJ, Henson DE, Herbst AL, Noller ILL, Schiffman MH. Interim guidelines for the management of abnormal cervical cytology. JAMA 1994;271:1866-9.