Background: Kaposi's sarcoma-associated herpesvirus (KSHV), also known as human herpesvirus 8, may be the infectious cause of KS. Its prevalence in the general population, on the basis of detection of the virus genome, is controversial. To investigate the seroprevalence, we measured antibodies to a recombinant capsid-related (lytic cycle) KSHV antigen and a latent antigen complex.
Methods: We selected potentially immunoreactive capsid-related proteins of KSHV by expressing them as recombinant proteins and testing them in western blot assays. We used a truncated recombinant protein encoded by KSHV open reading frame 65 (orf 65) to develop a diagnostic enzyme-linked immunosorbent assay (ELISA) and tested sera from HIV-infected individuals with KS, HIV-uninfected patients with "classic" KS, other HIV risk groups, and blood donors. We also compared the antibody response to this capsid-related protein to the response to latent antigen(s) in an immunofluorescence assay.
Findings: 77/92 (84%) sera from KS patients reacted with the KSHV orf 65 protein and 84/103 (81.5%) reacted with KSHV latent antigen(s). The dominant immunogenic region of orf 65 is within the carboxyterminal 80 aminoacids, a region with little sequence similarity to the related Epstein-Barr virus, suggesting that orf 65 is a KSHV specific antigen. Only three sera from patients with haemophilia (1/84) or from intravenous drug users (2/63) had KSHV specific antibodies in the orf 65 assay whereas none of these sera reacted with latent antigen. Antibodies to KSHV were also infrequently found in UK and US blood donors by either assay (UK, 3/174 with orf 65 and 4/150 with latent antigen; US, 6/117 with orf 65 and 0/117 with latent antigen). They were more common among HIV-infected gay men without KS (5/16 by orf 65 ELISA, 10/33 by IFA), HIV-uninfected STD clinic attenders (14/166 by IFA), and Ugandan HIV-uninfected controls (6/17 by orf 65 ELISA, 9/17 by IFA). Antibody reactivity to the orf 65 protein (ELISA) and to latent antigen(s) (IFA) was concordant in 89% of 462 sera tested but reactive blood donor sera were discordant in both assays. Four AIDS-KS sera were unreactive in both assays.
Interpretation: The distribution of antibodies to both a capsid-related recombinant protein and latent antigen(s) of KSHV strongly supports the view that infection with this virus is largely confined to individuals with, or at increased risk for, KS. However, infection with KSHV does occur, rarely, in the general UK and US population and is more common in Uganda. Antibodies to latent antigen(s) or to orf 65 encoded capsid protein will not detect all cases of KSHV infection, and a combination of several antigens will probably be required for accurate screening and confirmatory assays.