Models identify cellular proliferation as driver of HIV persistence

Antiretroviral therapy (ART) limits human immunodeficiency virus (HIV) replication and prevents further infection of CD4+ T cells, the cellular subset targeted by HIV. However, ceasing therapy results in viral rebound from a “reservoir” of HIV. It is currently unknown if the rebound is the result of viral replication in a small population of infected CD4+ T cells or is alternatively due to proliferation of the infected cells themselves. HIV might be functionally cured by decreasing the reservoir so that rebound after stopping ART is unlikely, but whether therapeutics should target viral replication or infected cell proliferation is unclear. Research Associate Dan Reeves from the Schiffer group in the Vaccine and Infectious Disease Division addressed this critical question in a recent publication in Nature Communications.

HIV integration site analyses provide an ideal assay to investigate the mechanism that sustains HIV reservoir. When HIV infects a T cell, the virus integrates into the cellular chromosomal DNA at a random location, making independent integrations at the same location in different cells highly unlikely. Based on this knowledge, cells that are newly infected by viral replication can be identified by their unique integration sites and sequences. Contrarily, infected cells that arise from the proliferation of an infected cell will share integration sites, making them distinguishable by their repeated, “clonal” integration sites. Published integration sites determined from individuals on ART for several years indicated substantial clonality¹ yet more than 50% of observed integration sites were unique. Taken directly, these data suggest proliferation occurs, but that viral replication makes for the driving force that is perpetuating the HIV reservoir.

In the current study, the authors incorporated other data² that real reservoir sizes were orders of magnitude larger than the observed sequence sample size. Thus, they developed and applied mathematical models inspired by ecology to extrapolate existing data to the whole reservoir. Contrary to raw data that suggested that most integrated HIV sequences occur only once within a patient, these models found that when the entire reservoir size is considered, most HIV integration sequences belong to small but nonetheless clonal families. Due to the nature of sampling, the HIV integration sequences that are accessible are limited compared to the total reservoir. Because of this, traditional empirical analyses allow sequences belonging to smaller clones to be obscured by the larger clones that predominate. However, Reeves’ modeling resolves reservoir dynamics on a larger scale, elucidating previously obscured clones. Reeves said that these findings indicate that “not only does cellular proliferation contribute to maintenance of the HIV reservoir, it more likely is the main mechanism of persistence, generating more than 99% of new sequences during ART.”

To expand on these findings, the authors next investigated the temporal dynamics of the HIV reservoir. They used a mechanistic modeling approach to indirectly simulate viral evolution, a surrogate marker of viral replication, as some previous studies noted evidence of viral replication persisting after ART initiation. However, the authors’ model found that the contribution of viral replication rapidly declines after ART initiation and becomes negligible within days. However, they found that longer-lived infected cells carry a “fossil record” of past viral replication that is detectable for months following the original viral replication, explaining previous discrepancies in the literature and demonstrating how this fossil record can be incorrectly interpreted as ongoing viral replication in patients on ART.

Through analysis of HIV DNA clonal distribution with mathematical modeling, this study demonstrated that infected cell proliferation, not viral replication, is the primary driver of the HIV reservoir, a distinction that has previously remained elusive. While the Schiffer group is clear that their work cannot completely rule out the hypothesis that viral replication persists after ART initiation, proliferation appears to be the far more dominant phenomenon. Furthermore, their dynamical models subsequently explain how replication artifacts within long-lived infected cells may have misinformed previous studies. These artifacts should be avoidable, and the authors suggest at least a six-month washout period for reservoir studies going forward, so that the fossil record no longer interferes with interpretation. Reeves said that the next steps for this project are to follow up what drives the HIV sequence distribution, or why there are “a few large clones, and many small ones,” a phenomenon they and others believe is dictated by selective proliferation of certain T cell subsets upon recognition of cognate antigen. Additionally, an ongoing American Foundation for AIDS Research-funded clinical trial collaboration between the Schiffer and Hladik groups is testing the efficacy of the lymphocyte antiproliferative agent MMF on the reduction of the HIV reservoir, as evidence suggests that blocking proliferation could meaningfully reduce the reservoir size. The findings of this study provide insight to an area that is critically important for HIV cure research and are currently informing design of future therapeutics both in the lab and the clinic.

¹ (Wagner et al. Science 2014, Maldarelli et al. Science 2014),

² (Ho et al. Cell 2013)

Reeves DB, Duke ER, Wagner TA, Palmer SE, Spivak AM, Schiffer JT. 2018. A majority of HIV persistence during antiretroviral therapy is due to infected cell proliferation. Nature Communications. Nov 16;9(1):4811. doi: 10.1038/s41467-018-06843-5.

The work was supported in part by the National Institutes of Health, the Washington Research Foundation and the Martin Delaney Collaboratory for HIV research.

Fred Hutch/UW Cancer Consortium members Josh Schiffer (FH/UW) and Thor Wagner (Seattle Children’s) contributed to this work.