Associate Professor Johns Hopkins University, United States
Introduction:: The persistence of virus in latent form is a key obstacle to a potential cure for HIV. The latent infection inevitably results in viral rebound after termination of combination anti-retroviral therapy (cART), under which a combination of anti-retroviral drugs is administered. Recent work showed that autologous immunoglobulin G (IgG) antibodies control viral outgrowth in the latent reservoir of HIV to variable degrees in different individuals. The neutralizing ability of these autologous neutralizing antibodies (AnAbs) is due to their recognition of the envelope env protein that buds off the membrane of mature HIV virion. As a result, the HIV env gene is under strong selective pressure; it contains five hypervariable regions that diversify of HIV membrane signature, allow the virus to evade the host adaptive immune response, and create a cycle of co-evolution between the env gene and the AnAbs. This suggests co-administration of anti-retroviral drugs and allogenic antibodies targeting AnAb-resistant variants as a potential approach to achieve long-term HIV remission.
Materials and Methods:: Using clinical data from the Multicenter AIDS Cohort Study (MACS) cohort, our lab previously developed and validated a mechanistic model that captures the dynamics of cell-cell and cell-virus interactions during HIV infection. As part of this modeling process, the MACS patients were divided into four subgroups based on the length of time from HIV seroconversion to AIDS diagnosis. Within each of these subgroups we generated a population of 1,000 virtual patients, each with different parameter values defining their HIV model dynamics, but each fitting within the distribution of the clinical training dataset. These virtual populations were subsequently validated against another clinical dataset, and used to simulate potential therapies. Here, we implement the effects of nAbs as a reduction in rate of virus infectivity and do not differentiate between autologous and allogenic nAbs. Within the context of our model, these changes are reflected in the rate of infectivity of CD4+ T cells by HIV, the rate of infectivity of macrophages by HIV, and the rate of infectivity of CD4+ T cells by infected macrophages.
Results, Conclusions, and Discussions:: Using these virtual-patient mechanistic models of HIV infection, here we study the population variability in the predicted outcome of therapeutic immunization with nAbs. This population model gives an estimate of the overall likelihood of success or failure, and in addition, insight into for which virtual patients is it most likely to work, and which parameters govern the responsiveness to therapy? In our simulations, the timing and longevity of nAb treatment are important determinants of therapeutic success, as measured by low viral load and high T-cell recovery (Figure 1). In addition, we can also vary the strength of nAb activity; at the population level, the fraction of individuals experiencing viral rebound increases with the attenuation of nAb activity. The overall trend appears to be a saturable response that is well described by the Hill equation. However, we found that effects of nAb activity are not permanent; in the case of diminishing nAb activity, viral rebound ultimately follows after the cessation of cART (ATI). In practice, this would mean autologous and allogenic nAbs combined would have to continue to confer specificity to the evolving viral reservoir in order for therapeutic immunization to have a successful outcome. Given the invasiveness of the procedures required to carry out therapeutic immunization, our mechanistic model allows one to simulate complex virtual clinical trials in silico prior to conducting it under clinical settings, with validated virtual patients capturing variability across HIV patient population.
