Abstract Body

While the magnitude, timing, persistence, and quality of HIV released from the source CD4+ T cells undergoing latency disruption are unknown, both multi-day sustained HIV release and instantaneous burst stochastic models of HIV replication have been proposed. Which model best represents latency disruption may have important implications for detecting replication-competent virus.

We used HIV gag RNA RT-PCR to quantify HIV release from isolated primary CD4+ T cells undergoing latency disruption in limiting dilution cultures. Viral inhibition cultures with efavirenz revealed released virus without new rounds of infection and were compared to respective viral outgrowth cultures. Sustained release and instantaneous burst stochastic models were implemented using the Gillespie algorithm with parameters derived from applying the deterministic model to experimental data.

While many outgrowth wells accumulated high amounts of virus, there were also many with declining virus that was nevertheless replication-competent, confirmed by virus transfer from the original wells to new outgrowth wells. Culture wells with >77% probability of being seeded by a single virus-producing cell were analyzed for viral release. Half of cells releasing virus did so in the first 4 days of culture. The accumulated virus for a well varied from less than 120 to 30,000 HIV gag RNA copies. Although a low amplitude instantaneous burst dynamic pattern was observed, the vast majority of accumulated HIV RNA was attributed to a higher amplitude virus release pattern sustained for 2-6 days. The intrinsic decay half-life of virus in cultures in which infection was blocked was 3 days. Applying this result to the deterministic model, we estimated an average total sustained release per cell of 5500 HIV RNA copies.

Our results are consistent with a multi-day sustained virus release stochastic model that predicts, compared to an instantaneous burst model, higher virus extinction probabilities. Because extinction probabilities are highest at extremely low infected cell numbers, such as single infected cells at limiting dilution, the sustained release stochastic model predicts, and our results show, that replication-competent virus will be present in cultures that are negative for viral outgrowth. Our results further demonstrate that single virus-producing cells can survive and produce virus over several days without succumbing to viral cytopathic effects.