Persistence by proliferation?

See allHide authors and affiliations

Science  11 Jul 2014:
Vol. 345, Issue 6193, pp. 143-144
DOI: 10.1126/science.1257426

The persistence of HIV-1 infected cells in individuals on antiretroviral therapy (ART) presents an obstacle for cure of infection. ART is the best available remedy for millions of infected people, but treatment must be life-long because HIV establishes latent infection that is unaffected by antiretrovirals and is invisible to immune surveillance. Because decades of treatment may be unsustainable, there is intense interest in reversing latency. If quiescent HIV in CD4+ T cells can be identified and activated without enhancing new infection, HIV-targeted immune response might be able to control or even clear infection. On page 179 in this issue and in this week's Science Express, Maldarelli et al. (1) and Wagner et al. (2), respectively, raise a new challenge for these efforts suggesting that proliferation of latently infected cells may be a key factor in sustaining this durable viral reservoir.

Latent HIV proviruses (viral genome integrated into the host cell DNA) are found most often in resting CD4+ T cells within the central memory arm of the immune system (35), although other T cell subpopulations have been implicated (6, 7). The latent pool shows minimal or absent decay (8), which is not fully understood. One possible explanation is that ongoing low-level HIV replication during ART replenishes the pool. However, viral genetic diversity does not increase over time in individuals during ART (9), at odds with this view.

The persistent pool of HIV-1.

Antiretroviral therapy can prevent the creation of new latently infected cells, but it does not affect cells in which latency was initially established. Intermittent bursts of viremia originate in part from this latent reservoir. Forcing these cells to exit the latent state without enhancing new infection could make the virus vulnerable to clearance by an HIV-targeted immune response. Blocking the proliferation of these latently infected T cells could deplete the pool if its stability is driven by such multiplication.


Other evidence suggests that virus emerges from the pool of latently infected cells periodically. Even patients whose viral load is well suppressed show intermittent bursts of viremia (“blips”), and in many patients viremia is detectable in specialized assays (10, 11). Given that the pool of latently infected cells must be primarily established before ART, it is difficult to understand why such periodic induction of the pool does not lead to it running dry.

Homeostatic proliferation of infected transitional memory T cells (6) has been proposed as a source that could maintain the pool, but this does not explain persistence in the dominant central memory reservoir. Latently infected stem cell–like memory T cells could proliferate (7), and it will be of great interest to compare integration patterns seen in these cells and in more differentiated cell populations.

Maldarelli et al. and Wagner et al. harvested DNA from the blood cells of HIV-infected individuals after a decade of successful ART, and analyzed the distribution of sites of proviral integration in the human genome. Typically, HIV favors integration in regions of the genome that are transcriptionally active (12), but a unique pattern was seen in rare proviruses from well-suppressed patients. Both groups found expanded proviral clones that were enriched for proviruses in or near a limited set of cellular genes, some of which encode products involved in controlling the cell division cycle or cancer progression. This supports the hypothesis that disruption of these genes by proviral insertion promotes growth or persistence of the host cell (13). Maldarelli et al. and Wagner et al. identify the host gene encoding the basic leucine zipper transcription factor 2 (BACH2) as a frequent site of HIV integration. BACH2 is a transcriptional regulator that controls CD4+ T cell senescence and cytokine homeostasis (14). Thus, the new findings suggest a link between the persistence of latently infected cells and proviral integration in genes related to cell proliferation and cancer.

Further experiments should strengthen these ideas. There is as yet no molecular evidence that such integrations of HIV-1 lead directly to the proliferation of latently infected cells, but it should be possible to engineer viral integration into specific sites of the host cell genome and demonstrate cell proliferation. In addition, there is as yet no proof that the proviruses encode for replication-competent HIV genomes. Maldarelli et al. did carry out the Herculean task of single-genome amplification and sequencing tiny amounts of HIV RNA recovered from the plasma of some patients studied. This verified a close similarity of circulating viral envelope sequences to those found in integrated proviral genomes in expanded clones. However, like prior studies (11), such sequencing is limited to a small portion of the HIV genome, and cannot eliminate the possibility of inactivating mutations in other parts of the proviral genome, making the virus incompetent to replicate. Given that the cells harboring quiescent HIV-1 are only a tiny minority of the total CD4+ T cell population examined by Maldarelli et al. or Wagner et al., and that years of ART have allowed for years of selection, alternative interpretations of the data are possible. For example, it is not yet ruled out that the expanded T cell clones detected could be expanding for other reasons (e.g., in response to stimulation by a specific antigen). There may be other reasons for preferential viral integration into the genes described as well. There may also be “survivor bias” in the detection of replication-incompetent genomes. Indeed, given the model, it is puzzling that no increase in the total number of HIV DNApositive cells was observed.

The findings of Maldarelli et al. and Wagner et al. raise additional issues. Lentiviral vectors are used extensively in therapeutic gene transfer, so monitoring for related events of proliferation-promoting integration with these vectors during gene therapy is important. Indeed, clonal expansion was observed in the case of a lentiviral-based gene correction of the blood disorder beta-thalassemia in which integration at the site of a proto-oncogene increased cell proliferation (15). In this case, the host gene encoding high-mobility group AT-hook 2 (HMGA2) produced a truncated mRNA due to vector insertion within the gene. HMGA2 is a transcription regulatory protein. The truncated mRNA removed a binding site for a microRNA that negatively controls HMGA2 expression. The result was increased accumulation of HMG2A mRNA and protein. HMGA2 was not an integration target in the cells studied by Maldarelli et al. or Wagner et al., raising questions about the differences between latent HIV infections and beta-thalassemia gene therapy.

Both studies also mention the concern that HIV integration could contribute to the development of cancers by insertional mutagenesis. However most HIV-related malignancies are not T cell cancers, and even most HIV-related lymphomas are of B cell origin. HIV cancers are not thought to harbor integrated HIV DNA, although this could be reinvestigated.

If blocking proliferation of latently infected cells proves to be necessary, it will complicate efforts to clear the latent reservoir. But clearance of this reservoir is crucial to achieve a cure of HIV infection.


View Abstract

Stay Connected to Science

Navigate This Article