In this study, the LIPS technology based on luciferase-tagged antigens was employed to generate quantitative antibody profiles against selected antigens from HIV, influenza, and the seven known cancer-associated viruses. To our knowledge, this is the first study to examine infection status and antibody profiles in parallel against these nine different viruses. HIV-infected individuals were the focus of this study since HIV-positive patients are at higher risk for developing a variety of infection-related cancers. Our results with the three groups of HIV patients revealed that 68% of the HIV samples were also coinfected with at least four of the seven cancer-associated viruses (HBV, KSHV, EBV, and MCV). In the case of HBV, the detected antibodies could also reflect patients who have cleared infection. Nevertheless, the high rate of cancer-associated virus exposure was in sharp contrast to the control blood donors who were only infected with the two, EBV and MCV, ubiquitous cancer-associated viruses. Mechanistically the higher antibody levels against EBV and MCV in the HIV-infected patients compared to the healthy controls likely reflects the loss of immune control over these agents resulting in increased viremia and antibody production. On the other hand, these finding for EBV and MCV were not due to a generalized increase in antibody production in HIV-positive patients because antibody levels against the HA2 antigen of influenza were lower in the HIV patients.
Previous studies have shown that the pathogenesis of OLP involves EBV infection of epithelial cells in the oral cavity [25
]. Although it was expected that the anti-EBV antibody levels would be higher in the OLP group compared to the KS group, this was not the case. The antibodies against the EBV p23 and p18 antigens in the OLP group were not statistically different than the levels found in the KS group or NHL groups. It is possible that altered anti-EBV antibody responses in the OLP patients might exist in saliva, but this needs to be studied further. Despite these negative findings, the OLP patients did show two features of having impaired humoral immunity. First, many of the OLP patients showed blunted anti-p24 HIV antibody responses. A few KS patients also showed low levels of anti-p24 HIV antibodies. However, these low levels were not a general defect of humoral immunity because these same OLP and KS patients showed comparably high levels of antibodies against other HIV antigens (i.e., gp41 and RT) and many of these same patients showed robust humoral responses against the HBV core antigen. Additional dilution experiments involving the titration of anti-p24 antibodies in patient serum samples failed to demonstrate that the low titers of anti-p24 antibodies were due to p24 antigen-antibody immune complexes (Burbelo, unpublished). As previously reported, the low anti-p24 antibody responses in the patients are more likely due to T-cell exhaustion [26
] and may be a marker of AIDS progression [27
]. The OLP group also showed the poorest humoral responses against influenza. These results are consistent with other reports showing that HIV patients with low CD4 counts have depressed immunity against influenza [28
]. It is perhaps appropriate to point out that many of the OLP patients were often coinfected with several cancer-associated viruses including 78% with KSHV, 90% with HBV, and 8% with HCV. Although other studies have shown that KSHV infection is frequently found in cohorts of men who have sex with other men (~20–45%) [30
] and is highly correlated with HIV infection [30
], the high frequency (78%) of KSHV infection in patients with OLP was unexpected. Based on the high prevalence of KSHV infection in the OLP patients and the ability of KSHV to propagate in the oral cavity [34
], it is tempting to speculate that the KSHV infection might contribute to the immunosuppression seen in the oral cavity of some OLP patients.
The KS patients showed the highest frequency of infection with these different cancer-associated viruses. Although the KS patients showed the highest levels of anti-KSHV antibodies, this was not a consistent trend and there were selected OLP and NHL patients with equally high antibodies. The KS patients also had the highest antibody levels against MCV VP1 capsid protein compared to the control blood donors, OLP, or NHL patients. Since MCV and KSHV share a tropism for cells in the dermis [2
], the higher antibody levels seen against MCV-VP1 in the KS group could reflect an enhanced immune attack on the skin.
EBV has been implicated as contributing to HIV-associated NHL, especially immunoblastic and central nervous systems types. In contrast to several previous studies correlating antibodies with a particular infectious disease condition [12
], antibody profiles against EBV were not informative biomarkers of NHL. These findings are also consistent with a recent study using a protein array of 40 different EBV proteins, which failed to detect informative NHL-specific EBV humoral responses [35
]. This negative data regarding antibody profiling may reflect the pathogenesis of HIV-associated NHL, which predominantly involves a loss of T-cell immunity and genomic rearrangements [36
]. Nevertheless, the NHL subgroup did show the lowest anti-HBV core antibody levels among the HIV-infected groups, despite having a similar prevalence of HBV infection. It is possible that the lower anti-HBV core antibodies in the NHL patients is a surrogate of impaired immune function, but this requires further exploration.
It is important to point out that our study has several limitations. First, asymptomatic HIV-positive individuals were not studied to determine the prevalence of infection for these viruses. Second, the HIV individuals used for analysis may not necessarily be representative patients of OLP, KS, or NHL and it is highly likely that the prevalence of many of these viral coinfections may be markedly different in other cohorts of patients. The high rate of HBV infection seen in our cohort is of particular importance because HBV infection has a significant negative impact on HIV outcome [38
]. While the HIV patient serum samples studied here were from the 1980–1990s, more recent vaccination and other preventive measures have greatly changed the level of HBV infection in the United States [39
]. It is also important to point out that HIV-infected individuals show a low rate of response to vaccination and therefore remain at risk for HBV infection [41
]. Similarly, the incidence of KS, and to a lesser extent NHL, has decreased with HAART [42
], which may also influence the relative antibody levels seen in those patients. Another limitation is our selection of antigens. Along these lines, antibodies were only examined against two HPV proteins, not against the capsid, but it is likely we underestimated the prevalence of asymptomatic HPV infection in these subjects. Although the long-term followup of the HIV patients is lacking, it is possible that some of these HIV-positive subjects may have gone on to develop clinical problems related to some of these viral coinfections. For example, it is possible that some of the OLP patients were infected with KSHV ultimately when on to develop KS.
Our study shows the feasibility of generating highly quantitative antibody profiles against multiple viruses including HIV, influenza, and seven cancer-associated viruses. The findings of varying anti-viral antibody levels against influenza, KSHV, HBV, and MCV likely involve multiple causes including alterations of specific immune pathways in the HIV-infected individuals that are pathogen-specific, the levels of immunosuppression, and/or the time of viral infection in relation to infection with HIV. Since many of these viruses can cause not only cancer, but other illnesses, serological screening may be helpful for the clinical management of these patients. For example, HBV and HCV infections are also associated with liver disease in HIV-infected individuals, which is also a major cause of morbidity [44
] and thus serological screening may have utility for diagnosis and monitoring. The ability to individually profile these cancer-associated viruses, along with many other infectious agents, has the potential to provide high quality humoral response data that could be correlated with other medically relevant inventories and be used as a platform for developing a generalized disease surveillance technology.