Stochastic analysis of protein-mediated and microRNA-mediated feedback circuits in HIV

After infecting a CD4 T cell, Human Immunodeficiency Virus (HIV) can either replicate and kill the cell or enter latency, a dormant state of the virus where viral gene-expression is turned OFF. This cell fate decision between viral replication and latency is controlled by the viral regulatory protein, Tat. This protein is known to activate its own production, creating a positive feedback circuit. Our previous work has shown that a stochastic model of this feedback circuit exhibits bimodal distributions of Tat levels, even though this circuit lacks deterministic bistability. The modes of the distribution correspond to infected cells with high Tat levels (corresponding to viral replication) or no Tat at all (corresponding to HIV latency). Experimental evidence points to an additional positive feedback loop mediated through a microRNA: a host microRNA targets Tat mRNA for degradation and Tat protein blocks synthesis of this microRNA. Here we investigate the interplay between Tat-mediated and microRNA-mediated positive feedback loops using deterministic and stochastic modeling. Our results show that these positive feedbacks together can exhibit deterministic bistability if the microRNA-mRNA interaction is sufficiently strong. Intriguingly, stochastic analysis reveals bimodal distributions for Tat even for parameter regimes where the coupled feedback system is not bistable. In summary, addition of the micro-mediated feedback loop can lead to bimodal Tat levels for wide a range of parameter values, and suggests a role for microRNAs in the viral cell fate decision in vivo.