Human immunodeficiency virus (HIV)-positive women have a higher prevalence of human papillomavirus (HPV) infection in the cervix and anus, as well as squamous intraepithelial lesions (SILs) at these sites, than do HIV-negative women matched for age and HIV risk factors. Similarly, HIV-positive homosexual or bisexual men have a higher prevalence of anal HPV infection and anal SIL than do HIV-negative homosexual or bisexual men. In HIV-positive individuals, the prevalence of HPV infection, the proportion infected with multiple HPV types, and the prevalence of anogenital SILs increase with decreasing CD4 count. This situation may reflect loss of systemic immune response to HPV antigens or local HPV-HIV interactions at the tissue or cellular level. Despite the high levels of anogenital SILs, to date, there has not been a significant increase in reported cases of invasive anogenital cancer in HIV-positive individuals. However, several years may be required for SIL to progress to invasive cancer, and the advent of newer therapies for HIV that are expected to prolong survival may paradoxically increase the risk of progression to cancer in individuals with SILs if these lesions do not regress spontaneously and remain untreated.
The association between human papillomavirus (HPV) and invasive cervical cancer has been recognized for many years, initially through the recognition that cervical cancer had the characteristics of a sexually transmitted disease, i.e., association with number of sexual partners and the age at first intercourse. In the last 10 years, with the advent of molecular probes for HPV, a clear association between the presence of HPV DNA in cervical cancer cells has been shown. Furthermore, increasing understanding of the function of HPV proteins known to be expressed in cervical cancer tissues, such as E6 and E7, has established strong biologic plausibility for HPV as a key factor in the pathogenesis of this disease.
Most of current understanding of the epidemiology of HPV infection and anogenital neoplasia was derived from human immunodeficiency virus (HIV)-negative study populations. In recent years, it has become apparent that the prevalence of anogenital HPV infection is higher among HIV-positive men and women than in their HIV-negative counterparts. Likewise, potentially precancerous lesions of the anogenital region are more common among HIV-positive men and women. While this increased prevalence has not yet led to a substantial increase in the incidence of anogenital cancer in these populations, the incidence of cancer may well increase as HIV-positive patients live longer as a result of improvements in medical therapy for HIV infection. The consequences of HPV infection may be a paradigm for a possible shift in complications of HIV disease from acute opportunistic infections to malignancies associated with chronic viral infection. An understanding of HPV infection and anogenital neoplasia in HIV-positive men and women is therefore important, particularly since many, if not all, of the HPV-associated anogenital cancers that occur in these patients may be preventable.
The goals of this article are to summarize current understanding of the role of HPV in the pathogenesis of anogenital neoplasia, to describe current knowledge of the epidemiology of HPV infection and anogenital neoplasia in HIV-positive men and women, to describe HIV-HPV interactions that may play a role in anogenital disease pathogenesis in HIV-positive individuals, and to speculate on the effect of improved therapy for HIV infection on the natural history of anogenital neoplasia.
Role of HPV in Pathogenesis of Anogenital Neoplasia
There are many different anogenital HPV types, and these are generally divided into oncogenic and nononcogenic types by virtue of the frequency of their association with invasive cervical cancer. Of the oncogenic types, HPV type 16 is the most important by virtue of its being the most common in cervical cancer, followed by HPV types 18, 31, and 45 (1). Many HPV types are also found, although with less frequency, in some cases of cervical cancer; these include HPV types 33, 35, 39, 52, 56, 58, 59, and 68. Lastly, there is a large group of nononcogenic types rarely if ever found by itself in cervical cancer, and the ones that are most prominent in the genital region are types 6, 11, 42, 43, and 44.
The mechanisms of oncogenicity of HPV are increasingly, but not completely, understood. The HPV E6 protein binds and degrades its cellular p53 protein target through a ubiquitin-mediated pathway, whereas the HPV E7 protein inactivates the cellular RB protein, as reviewed previously (2). Among their many functions, p53 and RB shut down the cell cycle and negatively regulate cell growth. The p53 protein also has been shown to stimulate DNA repair enzymes after DNA damage, presumably to limit the amount of chromosomal damage that occurs in a cell (3,4). Consequently, through the effects of E6 and E7, HPV infection may lead to genomic instability in cells that continue to cycle despite the presence of chromosomal damage (3,5). Consistent with this hypothesis, chromosomal mutations as shown by studies of loss of heterozygosity may be found in cervical cancer (6-8). The stage at which these changes occur is not yet known, but one or more genetic changes may be required for progression from high-grade squamous intraepithelial lesion (HSIL) to cancer.
Cervical HPV Infection and Cervical Squamous Intraepithelial Lesions (SILs)
HPV on the cervix typically infects in the transformation zone, where the squamous epithelium of the exocervix meets the columnar epithelium of the endocervix. This squamocolumnar junction is a relatively thin, highly metabolically active area of epithelium. Most HPV infections occur here, and most HPV-related lesions, including invasive cancer, arise from this area. A spectrum of histopathologic abnormalities resulting from HPV infection has been described. At the more benign end of the disease spectrum are changes described as low-grade SILs (LSILs). The primary features of LSILs include limited proliferation of basal or parabasal cells with a high nuclear-to-cytoplasmic ratio and koilocytosis (characterized by cells in the more differentiated cell layers with an enlarged, irregular nucleus surrounded by a clear area or halo). At the other end of the disease spectrum are changes described as HSILs. The cardinal feature of HSILs is marked proliferation of immature basal cells, which may replace most or all of the normal epithelium, and mitoses in the more superficial cell layers. The clinical importance of this grading scheme is that almost all the invasive cancers arise from HSILs and few, if any, arise from LSILs. Consistent with this observation, almost all HSILs contain high- or medium-risk oncogenic HPV types, whereas LSILs may contain a wider range of types (9).
Epidemiologic studies of cervical HPV infection suggest that the age-related prevalence of HPV infection as determined by polymerase chain reaction (PCR) is highest among women in their late teens and early twenties (10). These data suggest that most women acquire HPV infection relatively early after initiation of sexual activity. The age-related prevalence of cervical HPV infection declines thereafter, suggesting that immune mechanisms may either be clearing the HPV infection or reducing it to levels that are undetectable with the use of current technology. In addition, some of this age-related decline may represent a cohort effect, given changes in sexual behaviors that have occurred in the last few decades.
The age-related prevalence of cervical LSILs parallels that of infection, but only a small proportion of women who acquire HPV infection develop clinically detectable LSILs (10). An even smaller proportion of these women develop HSILs or invasive cervical cancer. Currently, the rates in the United States are approximately 8 per 100 000, and many of these cancers are occurring in women who never have Pap smear screening (11). In developing countries, where there is no routine cervical cytology screening, the rates are much higher, and cervical cancer constitutes a major cause of mortality among young women.
Anal HPV Infection and Anal SILs
In the general population, anal cancer is more common among women than among men. In the pre-HIV era, anal cancer was reported in women approximately four times more often than in men, with an incidence of 13 per 1 000 000 per year in the United States (12). In the last 15 years, the incidence of anal cancer has increased over 35% in women (Hauser A: personal communication). In Denmark, rates of anal cancer are lower than those in the United States; however, between 1957 and 1987, the incidence of anal cancer among Danish women more than tripled to 7.4 per 1 000 000 (13). Rates of anal cancer also rose among Danish men during that time, increasing by 1.5-fold to 3.8 per 1 000 000.
The precise incidence of anal cancer among men with a history of receptive anal intercourse has been difficult to determine because cancer registries do not collect information on sexual orientation or behavior. Men with a history of receptive anal intercourse, however, may be at especially high risk of anal cancer. Daling et al. (14) estimated that the incidence of anal cancer among homosexual men was approximately 35 per 100 000, rendering the incidence of anal cancer in this group several times higher than current rates of cervical cancer in women in the United States (11) and similar to rates of cervical cancer prior to the introduction of routine cervical cytology screening.
Anal cancer shares many biologic properties with cervical cancer. These cancers are similar histopathologically, and both are strongly associated with HPV infection (15,16). The anal canal has a transformation zone, where the columnar epithelium of the rectum joins the squamous epithelium of the anus, which closely resembles that of the cervix. Like the cervical transformation zone, the anorectal junction is a common site of anal HPV infection and HPV-associated SIL. The histology of anal SIL is similar to that of cervical SIL, and anal HSIL is associated with the same HPV types as cervical HSIL (15). Although studies have never been done to determine if untreated anal HSIL is the precursor to invasive anal cancer, this is likely to be the case, based on knowledge of the natural history of cervical HSIL. From a clinical standpoint, this is important because the anorectal junction is about 2 cm inside the anal canal and thus would not be visible with routine perianal inspection.
Anogenital HPV Infection and SILs in HIV-Positive Men and Women
The data collected on the epidemiology and natural history of HPV infection and SIL described above were derived from HIV-negative individuals. There are several reasons to suspect that these data may be different in HIV-positive men and women. First, both HIV and HPV are sexually transmissible. Behaviors that increase the risk of acquiring HIV infection may also increase the risk of acquiring HPV infection and vice versa. Second, since the immune response to HPV may play an important role in control of SIL, it is possible that, as HIV-related immunosuppression increases, immunity specific to HPV declines. Consequently, HIV-positive individuals may be at higher risk of both acquiring HPV and developing SIL related to HPV once HPV is acquired.
Clear patterns are emerging from studies of large numbers of HIV-positive men and women. These patterns confirm the increased risk of HPV infection and SIL in these groups. One of these studies is a prospective cohort study known as the WIHS Study, the Women's Interagency HIV Study. This ongoing study enrolled 2015 HIV-positive women and 577 HIV-negative women who were matched for age and HIV risk factors at six different cities around the United States. HPV testing of a cervicovaginal lavage specimen was performed on the women at baseline by use of PCR. The procedure used was a modification of that employed by Ting et al. (17) and used consensus primers from the L1 region of the HPV genome, followed by specific typing for 39 different HPV types. Among these high-risk, HIV-negative women, the prevalence of HPV was approximately 26% (Palefsky JM: unpublished data), relatively high for their age when compared with the data of Schiffman (10), which were collected in a population of women with Kaiser Permanente medical insurance. However, the HIV-positive women had even higher rates of prevalent HPV infection, and the rates were highest among those with the lowest baseline CD4 count. Among women with CD4 levels less than 200/mm3 at baseline, approximately 70% had detectable HPV infection (Palefsky JM: unpublished data). Independent risk factors for cervical HPV infection at baseline included HIV status, current smoker, younger age, and ethnicity (with African-American women having the highest rates of HPV infection).
Similar data have been obtained from a cohort study of 346 HIV-positive and 262 HIV-negative homosexual or bisexual men in San Francisco. By use of a PCR HPV testing method similar to that used for cervicovaginal lavage specimens in the WIHS study, anal HPV infection was characterized in this study population at baseline. Approximately 60% of the HIV-negative men had anal HPV infection. Similar to our findings in the cervix, the proportion of men with anal HPV infection was even higher among the HIV-positive group and was highest among those with the lowest CD4 levels. Among the HIV-positive men with CD4 counts less than 500/mm3, HPV infection was nearly universal (Palefsky JM: unpublished data). These data were similar to those reported previously among homosexual men in Seattle (18).
Another interesting feature of HPV infection among both HIV-positive women and men was the multiplicity of HPV types. Of HIV-negative women in the WIHS study who were HPV-positive, 16% had more than one HPV type, compared with 42% of HIV-positive women (Palefsky JM: unpublished data). Among HIV-negative men, 23% had more than one type of HPV in the anal canal, compared with 73% of HIV-positive men (Palefsky JM: unpublished data).
With respect to cervical cytologic changes in the WIHS cohort at baseline, a relatively high proportion of HIV-negative women had abnormal cervical cytology (16%). Similar to the HPV data, the proportion of HIV-positive women with abnormal cervical cytology at baseline increased inversely with the baseline CD4 levels, and 53% of women with CD4 levels less than 200/mm3 had abnormal cytology (Fruchter R: personal communication). Finally, among HIV-negative men in the San Francisco cohort study, 21% had anal cytologic abnormalities at baseline, while 72% of HIV-positive men with CD4 counts less than 200/mm3 had abnormal anal cytology (Palefsky JM: unpublished data). Taken together, these findings indicate that HIV-positive women and men have a higher prevalence of anogenital HPV infection, a higher number of HPV types, and a higher rate of prevalent anogenital lesions—each of these clearly associated with lower CD4 levels.
Relatively few data are currently available on the natural history of HPV infection and SIL in these large study groups, and these studies are in progress. However, some prospective data on the projected incidence of anal HSILs are available from the San Francisco men's cohort study. Men who entered the study without HSIL were stratified according to HIV status and baseline CD4 level. Among HIV-positive men who had CD4 levels less than 200/mm3 at baseline or between 200/mm3 and 500/mm3, the 4-year projected incidence of anal HSIL was greater than 50% (Palefsky JM: unpublished data). Even among those who entered the study with a CD4 level greater than 500/mm3, the projected incidence of HSILs was high (approximately 30%). Finally, HIV-negative men in the study had a projected 4-year incidence of 17%, putting them potentially at the highest risk, since many of these men will have a completely normal life span. These data indicate that a high proportion of HIV-positive men, as well as a substantial proportion of HIV-negative men, will develop HSILs if they are followed for a sufficient period of time.
Invasive Anogenital Cancer in HIV-Positive Women and Men
The above data indicate a high prevalence and incidence of anogenital HPV infection and putative precancerous anogenital lesions. An important question that has emerged in the last few years is whether these findings will translate into higher rates of invasive anogenital cancer. Thus far, there has been no significant increase in the rate of invasive cervical cancer among HIV-positive women in the United States or in the developing world (19). Data on anal cancer are not as clear, since studies using different methods lead to different conclusions. Rabkin and Yellin (19) did not find a significant increase in anal cancer in single, never-married men in the San Francisco Bay Area in the early 1990s when compared with the pre-HIV years. In contrast, Melbye et al. (20) reported increased relative risk of anal cancer compared with the general population with increasing proximity to an AIDS diagnosis, implying a role for increasing immunosuppression. However, these data should be interpreted cautiously, since the results may be confounded by the length of time that an individual may have been infected with HPV. Overall, it seems likely that there has been some increase (but not very large) in anal cancer, and the exact magnitude is still unknown.
Based on knowledge of the natural history of cervical disease, it seems likely that progression of HSIL to invasive cancer may require several years. Until recently, most HIV-positive individuals would have died of other HIV-related complications before HSIL would have had the chance to progress, which may explain why the high prevalence of anogenital HSIL in HIV-positive women and men has not led thus far to a dramatically large increase in cervical or anal cancer.
Despite the absence of a clear increase in cervical cancer in HIV-positive women, it is important to recognize that, once it develops, cervical cancer may be very aggressive in HIV-positive women, and extreme vigilance on the part of the clinician is very important (21,22). Conversely, cervical cancer may be an indicator of HIV infection, particularly in those regions with the highest HIV prevalence, and women with cervical SIL or cervical cancer should be strongly considered for HIV testing.
In the case of cervical disease, the clinical implications of these findings are that aggressive screening, treatment, and follow-up for cervical SIL are necessary. In the case of anal disease, the data indicating the need for an anal-screening program modeled on that used for cervical disease are compelling. Studies have been performed recently that validate the use of anal colposcopy as a diagnostic tool (23) and anal cytology as a screening tool (24). Barriers to implementation of such a screening program include the paucity of studies showing that anal HSIL progresses to invasive cancer and that a screening program would have an impact on the rate of anal cancer. However, such studies are very unlikely to be performed, given the ethical issues of following an individual without treatment once HSIL is diagnosed. Other barriers include the absence of widespread expertise in the performance of anal diagnostic techniques and the morbidity associated with most treatments for anal lesions. Were such a screening program to be implemented in the future, individuals most likely to benefit would include HIV-positive and HIV-negative men with a history of receptive anal intercourse. Because of data showing that women with cervical HSIL or vulvar cancer have a high prevalence of anal SIL and anal cancer (25,26), HIV-positive and HIV-negative women with high-grade cervical or vulvar lesions should also be considered for screening. Data on anal HPV infection and anal SIL in HIV-positive women without cervical disease are still emerging, but several earlier studies (27,28) have shown that, similar to cervical HPV infection and cervical SIL, the prevalence of anal lesions is highest among HIV-positive women with the lowest CD4 levels.
Mechanisms of HIV-HPV Interaction
Several mechanisms of interaction between HIV and HPV may play a role in the higher prevalence and incidence of anogenital SIL in HIV-positive women and men. One such mechanism is the systemic immune response to HPV. It is known that women who have iatrogenic immunosuppression secondary to organ transplantation are at increased risk of developing cervical and vulvar cancers when compared with age-matched women with normal immunity (29). Recently, one study (30) showed that a smaller proportion of women with HPV type 16 infection and cervical SIL have cytotoxic T-cell responses to the HPV type 16 E6 and E7 proteins than women who were HPV type 16 positive without cervical disease. Other studies of T-cell proliferative responses (31,32) also confirm that HPV-positive women without disease have higher response rates than women with lesions. These data suggest that, once a woman acquires HPV infection, having a cell-mediated immunity (CMI) response is associated with protection against lesion development. Conversely, if HIV infection leads to global impairment of CMI responses, including those to HPV antigens, then loss of that CMI response to HPV antigens may be important in the pathogenesis of anogenital disease and may explain in part the higher levels of HPV infection and SIL among those HIV-positive individuals with the lowest CD4 levels.
Local interactions at the tissue and cellular levels between HIV and HPV may also play a role in potentiation of anogenital neoplasia in HIV-positive individuals (Fig. 1). Although the viruses are likely to be found in different cell types (33) (HPV in the epithelium and HIV in local Langerhans' cells and stromal cells), one possibility for interaction between these two viruses locally is through production of HIV-1 tat, which has been shown to be secreted by HIV-infected cells and may be able to up-regulate expression of the HPV type 16 E6 and E7 genes (34). Aberrant cytokine expression by HIV-infected cells may also play a role in potentiating HPV infection. Perhaps most controversial is the possibility of co-infection with HPV and HIV in the same epithelial cell and the possibility of mutual transactivation between the viruses.
Taken together, these disparate observations may be used to construct a model for the interplay between HPV and HIV in the pathogenesis of anogenital SIL and cancer (Fig. 2). In this model, individuals acquire HPV infection relatively early on after initiation of sexual activity and, in many cases, before HIV infection. HIV is acquired at one or more moments in time, and these individuals may continue to acquire different strains of HPV as well. As long as the systemic and local immune responses remain intact, HPV replication and gene expression are controlled and there are no anogenital lesions. If HIV infection is allowed to persist and the immune response deteriorates, higher levels of viral replication and expression of HPV-transforming genes may lead to development of a low-grade lesion. With time, which may vary considerably from one individual to another, the disease may progress to HSIL. As described earlier, the high prevalence of HSIL does not appear to be accompanied by increases in the rate of anogenital cancer because of HIV-related mortality before progression of HSIL to cancer occurs.
Conclusions: What Does the Future Hold?
The natural history of anogenital HPV infection and SIL in the era of highly active antiretroviral therapy (HAART) may represent a paradigm for HIV-related complications of the future. HAART has been shown to dramatically reduce systemic HIV viral load and to reduce the incidence of some opportunistic infections. Because HAART has not been in use for an extended period of time, it is not yet clear how much it will prolong the survival of individuals with HIV infection. Two scenarios are possible (Fig. 3). If individuals go on HAART after developing HSIL and if partial or complete restoration of immune response to HPV occurs, then one might expect HPV levels to decrease and HSIL to regress. If that is the case, then a decreased incidence of anogenital cancer from current levels would be expected. Conversely, if HAART permits individuals to live longer but does not have a significant impact on the immune response to HPV, then the disease in individuals with HSIL might now have the time to progress to invasive cancer, resulting in an increased incidence of anogenital cancer. It is not yet clear which of these two scenarios is correct. However, the most that HAART therapy could be expected to achieve would be to restore the immune system of an HIV-positive individual to one as functional as that of an HIV-negative individual. Since most HSILs do not regress spontaneously, even in HIV-negative individuals, the second scenario seems more likely at this time. Clearly, further natural history studies of men and women on HAART are needed, as are studies of the interactions between HIV and HPV in the pathogenesis of anogenital cancer.
Supported by Public Health Service (PHS) grants CA54053 and CA63933 from the National Cancer Institute, National Institutes of Health (NIH), Department of Health and Human Services, as well as by PHS grant AI-DE34989 from the National Institute of Allergy and Infectious Diseases and the National Institute of Dental Research, NIH.
We acknowledge the subjects and investigators of the University of California, San Francisco, Anal Cancer Study and the Women's Interagency HIV Study.