Worldwide, over 35 million people are currently infected with HIV, and 2.3 million individuals are newly infected each year. While combination antiretroviral therapy can suppress virus replication, improving life expectancy and quality, it cannot eradicate the virus. A latent reservoir of virus exists in long-lived lymphocytes and can re-initiate the infection (“rebound”) whenever treatment is stopped. Consequently, current therapy must be taken for life, and new research efforts are underway to find a permanent cure for HIV. Two general approaches are being taken to prevent HIV rebound and hence allow therapy to be completely stopped (“cure”). One approach, often called a “sterilizing cure”, is to purge all the remaining latent virus (so called Berlin patient, Mississippi baby). Another approach, often called a “functional cure”, is to instead equip the immune system with the ability to control virus that reactivates from latency (so called elite controllers or post-treatment controllers). This provides the proof-of-concept for immunologically-mediated control. The studies we will analyze in this program are part of a larger effort to use therapies that enhance the immune response (known as ”immunotherapy” and ”therapeutic vaccine”) to induce viral control either by clearing latent virus, boosting anti-viral immune responses, or both. The goal of this collaboration is to investigate the effect of each component of the potentially-curative intervention (vaccine and immunotherapy) and evaluate their synergy.
We use statistical inference methods (developed by SISTM team) applied to mathematical models describing the dynamics of virus and immune cells (developed by Harvard team). Such approaches are superior to regression-based methods that ignore the underlying biological mechanisms. A Bayesian framework which estimates parameters at a population level allows us to explicitly test for differences between treatment groups, and, can overcome parameter identifiability issues that arise in complex, high-dimensional models. Indeed, these model can both provide mechanistic understanding of the biological process when administrating a treat- ment and help predicting the outcome for a new study. Previous studies have demonstrated that mechanistic models can accelerate the development of new drugs by facilitating in silico simulation of clinical trials (e.g. for Hepatitis C antivirals). We aim at providing such a tool for the development of immunotherapies and vaccines for HIV cure.
