Ph.D Thesis Colloquium : Vemparala Bharadwaj

Two of a kind: Insights into the mechanisms driving post-treatment and natural HIV control from mathematical modeling. 

More than half-a-million people die each year due to HIV/AIDS, making it an important global public health problem. Antiretroviral therapy (ART), the current standard-of-care, halts disease progression but is not curative. Interrupting ART typically results in viral rebound. Efforts to develop interventions that elicit long-term viral remission in people living with HIV (PLWH) are underway. These efforts are motivated by the existence of two rare PLWH phenotypes: natural controllers (NCs) and post-treatment controllers (PTCs). NCs have the innate ability to control viral replication without ART, while PTCs develop this ability after months/years of ART. In my thesis, I developed mathematical models to study these phenotypes, drew quantitative insights by constraining them to pre-clinical data from experiments on macaques, and made recommendations for improving treatment outcomes.

My first study involves NCs. Evidence suggests that superior immune responses by CD8 T cells drive natural control. However, the mechanistic bases are unknown. To address this question, I developed an in vivo–ex vivo multiscale model that simultaneously predicts multiple virological and immunological biomarkers. The model described data from a pre-clinical study and estimated that the CD8 T cells of NCs had a 2-fold higher recruitment and/or maximal killing rate of infected cells than those of non-controllers. Previous models that could be constrained only by virological biomarker data fail to identify any such CD8 T cell signature. Finally, the model predicted the strength of CD8 T cell responses necessary for natural control, which sets targets for CD8 T cell–based therapies under development.

PTCs are the focus of my second study. Recent studies suggest that memory CD8 T cells may underlie PTCs, with their potency improving with early ART initiation. Memory cells are largely absent in chronic infections like HIV if left untreated. Thus, how memory cells formed early in infection are retained through ART and mount a response post-treatment leading to viral control is unclear. I addressed this question as follows. First, I developed a model of HIV infection incorporating memory CD8 T cell dynamics based on the following hypothesis drawn from recent experimental advances: the survival of memory CD8 T cells is progressively diminished with their exposure to antigen. The model, for the first time, described longitudinal data where the role of ART initiation time was explored. The model exhibited biostability: the progressive disease and viral control states are the two stable states, with the latter made more accessible by early ART. The model also predicted that an optimal ART initiation window exists for maximum post-treatment control, in agreement with independent experiments on macaques. Finally, I estimated the memory CD8 T cell pool size and its quality necessary to achieve post-ART control, which may guide translational efforts.