Squamous-cell carcinoma of the cervix (cervical cancer) has long been known to have all the epidemiologic characteristics of a sexually transmitted infectious disease. Annually, there are about 500,000 new cases of cervical cancer worldwide, and most of these are in nations with few resources. It is now established that cervical cancer is caused by infection of the genital tract with one of about a dozen sexually transmitted human papillomaviruses (HPVs). (1) Cervical cancer may well be the first major human neoplasm for which we have identified a single, necessary etiologic factor. Both epidemiologic and laboratory studies support this conclusion. In all parts of the world, infections with genital HPVs appear to account for nearly 100 percent of cervical cancers (2) (and unpublished data).
Of the approximately 30 HPVs that infect the genital tract, HPV-16 is responsible for about 50 percent of cervical cancers, and four types (HPV-16, 18, 31, and 45) together account for 80 percent of the cancers. Additional HPV types account for the remaining 20 percent. The association of HPVs with cervical cancer is very strong, with odds ratios of more than 50 in many case-control studies. (1) Cohort studies conclusively link the development of the precursor lesions of cervical cancer to infection with HPV. (1) A woman's risk of cervical cancer increases with the number of her sexual partners, but in many areas with a high incidence of cervical cancer, promiscuous sexual behavior of women's partners is an additional risk factor.
The compelling epidemiologic evidence of the role of HPV in cervical cancer is complemented by equally strong data on the role of oncoproteins E6 and E7 of high-risk HPVs in the molecular pathogenesis of cervical cancer. The E6 and E7 proteins, which in laboratory studies can immortalize human keratinocytes, are invariably expressed in HPV-associated cancers. Inhibiting the functions of E6 and E7 converts HPV-transformed cells back to the normal phenotype. In most cancers, the HPV genome is integrated into the cellular DNA. The integration of the virus disrupts the E2 region of the HPV genome and increases expression of the E6 and E7 oncoproteins. The oncoproteins exert their effect by forming complexes with the cellular tumor-suppressor proteins -- E6 with p53, and E7 with Rb. The E6 and E7 proteins of high-risk HPVs bind the tumor-suppressor proteins more efficiently than do the E6 and E7 proteins of low-risk HPVs. The inactivation of the tumor-suppressor proteins presumably distorts the cell cycle and results in genetic instability and additional cellular genetic alterations. Gain of chromosome 3q and loss of 3p may be important in the transition from an HPV-induced in situ cancer to an invasive cancer.
In summary, the evidence that HPV causes cervical cancer is as strong as that for any other human carcinogen, and there is optimism that HPV-based vaccines and cervical-cancer screening programs to detect high-risk HPVs in the genital tract will reduce the number of cervical cancers worldwide. However, the human immunodeficiency virus (HIV) pandemic now raging in Asia and Africa may pose a major obstacle to this goal, because HPV infections proceed at a faster pace in HIV-infected women (3) and are expected to lead to a higher incidence of cervical cancer in areas of high HIV prevalence. Cervical cancer was added to the list of AIDS-defining illnesses in 1993.
In this issue of the Journal, Sun et al. (4) compare the natural history of HPV infections and the persistence of this virus in a cohort of 220 HIV-infected and 231 HIV-negative women in the New York City area who had no Pap-smear abnormalities at recruitment and were monitored for HPVs for 12 months. Genital tract HPVs were detected by polymerase-chain-reaction (PCR) assays. After four examinations, the cumulative prevalence of HPV was 62 percent in the HIV-negative women, 74 percent in HIV-infected women with CD4 counts of greater than or equal to 500 cells per cubic millimeter, and 95 percent in HIV-infected women with <500 CD4 cells per cubic millimeter. Persistent infection, as defined by detection of the same HPV type in two or more examinations, was found in 24 percent of the HIV-infected women and 4 percent of the HIV-negative women. HPV persistence was associated with HIV infection as well as with the degree of immunosuppression. New HPV infections were detected equally frequently in HIV-infected and HIV-negative women. The authors conclude that an increased persistence of high-risk HPVs may explain the higher prevalence of these infections as well as the increased risk of cervical neoplasia in HIV-infected women.
HPV infections have also been implicated in cancers at anogenital sites other than the cervix (vulva, vagina, perineum, anus, and penis) and at nongenital sites (oral cavity, head and neck, and esophagus), but at these sites the role of HPV in the pathogenesis is less clear, and in some instances it is disputed. Among these cancers, the pathogenesis of squamous-cell carcinoma of the vulva, a relatively rare cancer, is well understood. Vulvar cancers are etiologically heterogeneous and are readily separated into those that are HPV-related and those that are not. (5) Only a minority of vulvar cancers are HPV-related, and they resemble cervical cancers in having basaloid morphologic features, an adjacent squamous intraepithelial lesion, an association with risk factors that are consistent with sexual transmission, a strong association with HPVs, and similar chromosomal changes. Women with the more common, keratinizing carcinomas of the vulva are older and do not have the risk factors that suggest sexual transmission. Their cancers are negative for HPVs.
Epidermoid anal cancer, a rare disease that is more common in women than men, may also have a dual cause. In this issue of the Journal, Frisch et al. (6) describe a large, population-based case-control study of anal cancers in Denmark and Sweden in which all 417 available patients with anal cancer diagnosed from 1991 to 1994 and two sets of controls (population controls and controls with rectal adenocarcinoma) were interviewed by telephone with regard to exposure histories and risk factors. In addition, paraffin sections of tumor tissues from 94 percent of the interviewed patients were examined for HPVs with a PCR-based assay. The investigators found that nearly all risk factors that were linked to anal cancer (e.g., high number of sexual partners, unmarried status, receptive anal intercourse, and history of sexually transmitted diseases) were consistent with sexual transmission. Daling et al. (7) reached a similar conclusion in an earlier investigation in the United States. Frisch et al. note that 93 percent of cancer tissues from women, as compared with only 69 percent of tissues from men, were HPV-positive, suggesting that at least in men some anal cancers may be unrelated to HPVs. The vast majority of both the men and the women with anal cancer reported no history of anal intercourse, thus leaving unanswered the question of how the anal canal becomes infected in patients with HPV-positive cancers.
Male homosexuality and HIV infection do not seem to contribute much to anal cancers in Denmark and Sweden, (6) but they have been implicated as independent risk factors in the recent increased incidence of anal cancer in the United States. (8) Homosexual men, regardless of their HIV status, have a high prevalence of anal HPV DNA. HIV infection in homosexual men is associated with the presence and progression of anal intraepithelial neoplasia, and it has been suggested that anal cytologic analysis may be of value in screening those at high risk for anal cancer. (9)
The focus in research so far has been on the role of genital tract HPVs in cancers of the lower genital tract. But recent studies show that cutaneous HPVs are widely seeded in normal skin, (10) which raises the intriguing possibility that HPV may also contribute to the development of nonmelanoma skin cancers.
Keerti V. Shah, M.D., Dr.P.H.
Johns Hopkins School of Public Health
Baltimore, MD 21205