Activation of Kaposi's Sarcoma-Associated Herpesvirus Lytic Gene Expression during Epithelial Differentiation

Since its initial description over a decade ago (14), Kaposi's sarcoma-associated herpesvirus (KSHV) has been recognized as a significant viral pathogen, particularly for immunocompromised hosts and persons infected with human immunodeficiency virus type 1 (HIV-1). KSHV is the etiologic agent of Kaposi's sarcoma (KS), having been demonstrated in biopsy specimens of all forms of KS despite distinct differences in the geographic origin, age, and gender of affected persons (29). KSHV is also associated with multicentric Castleman's disease (80) and primary effusion lymphoma (11). The precise modes of KSHV transmission are not clearly defined; however, epidemiological data suggest that both sexual and nonsexual routes are possible. The association between KSHV transmission and sexual activity has been largely defined in North American men who have sex with men (MSM), where risk factors include sex with a partner who has KS (66), a history of sexually transmitted infections (STIs) (55), and an increased number of sexual partners (55). Other risk factors, however, such as deep kissing with an HIV-positive partner (66) and orogenital contact (25), have led to consideration of an oral source of transmission. In a cohort of MSM from San Francisco, similar prevalences of KSHV infection from 1978 to 1996 were found, despite a reduction in HIV-1 seroprevalence and the institution of “safer” sexual practices (65). This raises the issue that sexual activity may be a marker for other types of intimate contact in this population. In keeping with this idea, PCR-based studies of the male genitourinary tract have described only infrequent, low-level shedding of KSHV DNA in genital secretions (22). Infectious virions have not been demonstrated in semen. Similarly, only infrequent and low-level shedding of KSHV DNA has been observed from male urethral and anorectal secretions (66). KSHV DNA has only rarely been detected in vaginal secretions (8, 46, 93), and heterosexual sex has not been clearly associated with KSHV transmission (76). Evidence for nonsexual transmission is supported by data derived from studies in Africa, Italy, Egypt, and Japan which document KSHV infection in groups with very low risk of STI, including children (1, 8, 42, 91, 92). In regions of Africa where KSHV is endemic, a cumulative increase in KSHV seroprevalence from birth through adolescence has been observed, with more than 40% of children over 14 years of age infected and many infants seroconverting during the first year of life (2, 30, 56). Evidence for transmission through breast milk has not been found (7), and congenital infection is thought to be rare (54). The fact that children and other groups without risk of STI are infected with KSHV implies salivary transmission, as is seen with other human herpesviruses (HHVs), such as cytomegalovirus, HHV-6, HHV-7, and Epstein-Barr virus (EBV). KSHV DNA has been detected frequently in human saliva (6, 44), and when found, it occurs at a titer 2 to 3 log units higher than those at other anatomic sites (66). Infectious KSHV has been documented in saliva (87). For these reasons, the oral cavity is unique in its ability to support lytic replication of KSHV. Published studies report that 11% to 68% of HIV-positive MSM infected with KSHV in the absence of KS shed viral DNA in saliva on at least one occasion (9, 22, 66). One study demonstrated that 39% of KSHV-seropositive MSM shed viral DNA from the oral cavity on more than a third of the days sampled, regardless of HIV-1 status (66). Current data indicate that KSHV does not replicate in salivary glands (21, 22). Epithelial cells lining the oral cavity are a probable source, given that nucleated oral epithelial cells from nonkeratinized mucosal surfaces have been shown to support viral infection in vivo, by the detection of KSHV-specific latent and lytic mRNA transcripts (66). Other viruses, such as EBV (31, 49, 74, 96) and human papillomavirus (HPV) (reviewed in reference 51), undergo lytic viral replication in differentiating epithelia. The organotypic raft culture model, originally developed to study keratinocyte differentiation (70), has been of great value in the study of viral replication in developing epithelium (57). Organotypic raft cultures accurately reproduce the process of epithelial differentiation in vitro, beginning with undifferentiated keratinocyte monolayers that become confluent and polarized, forming tight junctions prior to the initiation of keratinocyte differentiation. A proportion of the keratinocytes remain in the basal layer and maintain proliferative potential, while other cells leave the basal layer, lose mitotic capability, and differentiate into mature epithelial cells, forming the multilayered structure of the epithelium (39). Keratinocytes follow a programmed pattern of gene expression during differentiation, each step of which is characterized by the expression of keratin, and keratin-associated, proteins (59). This in vitro model has demonstrated that HPV productively infects cutaneous and mucosal epithelia by a process that is tightly linked to epithelial differentiation (40). In order to test the hypothesis that activation of KSHV lytic gene expression occurs during epithelial differentiation, we employed an organotypic raft culture model that utilized keratinocytes isolated from normal tonsils. We found that the process of epithelial differentiation activated KSHV lytic-gene expression and resulted in the production of infectious virus at the epithelial surface.