One of the most widely discussed and publicized ideas for targeting the HIV reservoir is “kick and kill” or “shock and kill.” The aim is to kick dormant, latent HIV into revealing itself so the cells that contain the virus are visible to the immune system; the kill aspect involves trying to enhance the immune response so it’s able to recognize and destroy these HIV-infected cells. The ultimate goal is to deplete the latent HIV reservoir that persists in people on antiretroviral therapy (ART).
The RIVER trial, launched in the UK in December 2015, represents the first large, randomized controlled evaluation of the approach. Preliminary results were first reported at the AIDS 2018 conference in Amsterdam, and have now been described in detail in a paper in Lancet HIV. While the interventions were unable to reduce the HIV reservoir, the negative outcome still holds important lessons for the cure research field. New candidates for both the kick and the kill are in the research pipeline.
RIVER recruited a total of 60 participants with primary HIV infection—all men—and randomized them to receive either a standard ART regimen plus the integrase inhibitor raltegravir or standard ART, raltegravir, a prime-boost therapeutic HIV vaccine combination and a short course of the HDAC inhibitor vorinostat (a latency-reversing agent reported to have some activity in prior trials). The vaccines consisted of chimpanzee adenovirus and modified Vaccinia Ankara strain virus vectors carrying HIV antigens designed to induce T cell immune responses focused on parts of the virus that mutate the least (conserved regions).
The results showed that after 16-18 weeks there were no differences between the two groups in any measures of HIV persistence, including HIV DNA and the quantitative virus outgrowth assay (QVOA). A laboratory test assessing the ability of CD8 T cells to kill HIV-infected cells indicated that therapeutic vaccination helped maintain this capacity: killing activity declined from baseline in the ART only arm of the study but was maintained in the intervention arm. Vorinostat increased histone acetylation after dosing (the mechanism by which it can trigger the activity of latent HIV), however sensitive tests for low-level HIV RNA did not reveal any concomitant increases suggestive of HIV latency reversal.
In discussing their findings, the authors posit a number of possible reasons why this particularly kick and kill combination did not prove effective. Fewer doses of vorinostat were used than in some prior trials, and the evidence of latency reversal that has been described (increases in HIV RNA associated with dosing) was obtained in people with chronic rather than primary HIV infection. Another possibility is that even if vorinostat did cause some latently infected cells to reveal themselves to the immune system by expressing HIV antigens, the HIV-specific CD8 T cells induced and/or maintained by therapeutic vaccination may not have been capable of recognizing these antigens. Presentation of the antigens to CD8 T cells could also be inhibited by the HIV Nef protein, which is known to have the capacity to interfere in this process.
The researchers also acknowledge the argument that analytical treatment interruptions (ATIs) may be a better test of the effects of interventions than laboratory measures of the HIV reservoir, but at the time RIVER was designed ATIs were not in favor due to safety concerns, and they still believe that an ATI wouldn’t be justified in this case. They conclude that:
“RIVER helps to set the standard for how future trials might be done because of the new insights gained with regard to trial design, the inclusion of community representation, the need for better interventions, the necessity for clarity regarding the most relevant measures of the reservoir, and probably the use of an analytical treatment interruption approach.”
An example of a next generation kick and kill study was presented last week at CROI by Ole Schmeltz Søgaard. Known as ROADMAP, the trial compared romidepsin (an HDAC inhibitor that appears to have greater latency-reversing activity than vorinostat) to romidepsin plus the broadly neutralizing antibody (bNAb) 3BNC117 in people with chronic HIV infection on long-term ART. The bNAb was included in hopes of both promoting killing of HIV-infected cells via antibody-dependent cellular cytotoxicity (ADCC) and enhancing HIV-specific T cell immunity (as recently reported in a study of 3BNC117 combined with another bNAb, 10-1074).
A total of 20 participants were enrolled (17 men and three women); 11 were randomized to receive 3BNC117 and romidepsin, nine to romidepsin. There were two cycles of romidepsin administration: at weeks 0, 1, 2 and weeks 8, 9, and 10. The 3BNC117 group received a single infusion of the bNAb two days prior to these cycles. After 24 weeks, participants underwent an ATI in order to evaluate time to viral load rebound (defined as two consecutive measures equal to or greater than 200 copies/mL). Two participants opted out of the ATI, while a third was excluded after stopping ART while receiving romidepsin.
Søgaard reported that, disappointingly, there was no significant difference in time to viral load rebound between the two arms of the study: it took an average of 28 days in the romidepsin group compared to 17.5 days in the 3BNC117 plus romidepsin group. The slight difference in favor of the romidepsin group wasn’t statistically significant and Søgaard noted it was driven by one outlying participant who didn’t experience rebound until 12 weeks into the ATI.
HIV DNA levels measured at the midpoint between the romidepsin cycles and immediately prior to the ATI showed no significant changes. HIV-specific CD8 T cell responses remained stable over the same period and weren’t enhanced by 3BNC117 administration or negatively affected by romidepsin.
As was the case in the RIVER trial, Søgaard concluded that this particular combination was ineffective and that superior strategies are needed.
At CROI, Sharon Lewin—a leading scientist in HIV cure research—delivered an impressive, comprehensive update on the field, providing reasons to hope that better results will be obtainable in the future.
One key development is that the scientific understanding of HIV latency has improved significantly in recent years. The picture is now far more nuanced compared to when the HIV reservoir was first identified in the late 1990s.
The first relatively new concept highlighted by Lewin is that not all latent HIV remains stably latent (inactive) in people on effective ART. The HIV DNA that’s integrated into the genomes of infected cells can be transcribed into HIV RNA (which in turn can make HIV proteins), either intermittently or possibly also continuously in some cases.
A number of factors can affect the transcription of HIV DNA into HIV RNA, including the location of the infected cell. Lymph nodes, the gastrointestinal tract and the female genital tract are sites where HIV RNA is most frequently detected. Time of day also has an influence because proteins involved in the circadian cycle can activate HIV transcription, leading to fluctuations in HIV RNA levels.
Lewin described this phenomenon as “reservoir activity” and it’s important to stress that it does not represent ongoing HIV replication—any new infectious HIV viruses that are produced would be blocked from infecting new cells by the presence of ART. Gaining an understanding of the processes involved in reservoir activity should help researchers develop better latency-reversing agents.
Another aspect of HIV latency that is now coming into view is the importance of where exactly the virus lands when it integrates into the genome of a cell. As a loose analogy, if you think of the genome as a factory for producing all the proteins a cell needs to go about its business, HIV DNA tends to land in machinery that gets switched on regularly, which gives the virus opportunities to hijack that machinery to make more HIV RNA (and potentially more copies of infectious HIV).
But HIV DNA can also land in the genomic equivalent of a darkened factory storage room nobody goes into—in that case, the virus may be trapped and unable to reactivate. Between these two extremes there are likely a range of possibilities, from HIV DNA being integrated in a spot where the transcription of HIV RNA is likely, to HIV DNA being integrated into a genomic dead end from which it can never emerge. Again, these new findings have implications for latency-reversing agents because they demonstrate that there’s a spectrum of reversibility that will need to be addressed.
The potential importance of where latent HIV resides in a cell’s genome has been emphasized by recent studies of elite controllers (as presented at CROI by Chenyang Jiang), which suggest that their immune responses can clear the more active HIV reservoir, leaving behind only those cells containing HIV integrated in places from which it cannot reactivate—essentially, the intact HIV that remains in their bodies may be trapped, and unable to replicate or cause harm.
A further wrinkle in the latency story is the ability of CD4 T cells containing integrated HIV to proliferate, duplicating the viral genes they contain along with the cell. As Lewin noted, it’s now recognized that a large proportion—typically around half—of the HIV reservoir in people on ART has been created by what is referred to as clonal proliferation of CD4 T cells. The proportion of the HIV reservoir generated by clonal proliferation can be identified because the viruses are genetically identical and integrated into exactly the same spot in the genome of CD4 T cells that contain them.
The mechanisms driving the proliferation of latently infected CD4 T cells are still being elucidated but Lewin listed several possibilities:
- Homeostatic proliferation: this is a normal part of the life of CD4 T cells that facilitates the survival of the cell and maintenance of overall CD4 T cell numbers. Evidence that it plays a role in maintaining the HIV reservoir was first published back in 2009.
- Antigen-specific proliferation: the process by which CD4 T cells proliferate after recognizing a specific antigen (e.g. an antigen derived from HIV or other infectious agents such as CMV) and becoming activated to respond. Lewin cited several studies reporting new data on the importance of antigen-specific proliferation in HIV persistence (including two at CROI: Simonetti et al and Mendoza et al)
- HIV integration site driven proliferation: in this scenario, it’s been proposed that HIV can influence the proliferation of CD4 T cells by integrating into certain places in a cell’s genome. At CROI, John Coffin described findings that may support this possibility, but only for a limited number of integration sites.
There are published studies showing that latently infected CD4 T cell clones contribute to viral load rebound after ART interruption, and also that HIV reservoir activity in these cells can in some cases generate sufficient HIV RNA to be measurable by standard viral load tests (making it necessary for clinicians to be alert to the possibility of mistakenly attributing detectable HIV RNA in people on ART to treatment failure).
The last but critical point made by Lewin about the HIV reservoir is the distinction between intact and defective HIV. It’s been known for some time that the vast majority of HIV DNA present in people on ART represents defective viruses that are incapable of replicating (although in some cases capable of generating viral proteins), but researchers have lacked the tools to easily distinguish replication-competent HIV.
The recently developed intact provirus detection assay (IPDA) is an example of progress on this front. Lewin cited an encouraging new analysis by Gregory Laird and colleagues indicating that levels of intact HIV decline on ART, while the amount of defective HIV DNA remains relatively stable. Based on the case of the elite controller Loreen Willenberg, in whom no intact HIV can be detected, Lewin also suggested that perhaps a cure should be defined as the elimination of all intact virus (rather than the absence of any detectable HIV genetic material).
Moving on to the future of kick and kill, Lewin selected some examples of advances that offer reasons for optimism. Earlier this year, results of an evaluation of the candidate latency-reversing agent AZD5582 were published in the journal Nature. The compound belongs to a class of drugs called SMAC mimetics, which appear to have considerably less potential to cause toxicity compared to the drugs studied to date. In studies in both HIV-infected humanized mice and SIV-infected macaques, potent induction of viral RNA expression was observed in blood and multiple tissues after dosing.
Immunomodulatory compounds such as immune checkpoint inhibitors (e.g. anti-PD1 antibodies that are approved for cancer) and toll-like receptor (TLR) agonists may have the ability to both reverse HIV latency and enhance killing of infected cells. Concerns about serious immune-mediated side effects are currently limiting the use of the former, but might eventually be addressed by different formulations and dosing strategies. Two studies are testing TLR agonists in combination with dual bNAbs (one with the addition of therapeutic vaccination): TITAN and JAWS. Both trials involve ATIs. JAWS is taking place at the University of California San Francisco (UCSF) but was only due to start enrolling very recently (and is not yet registered in clinicaltrials.gov), so may well be delayed by the current COVID-19 emergency.
The field of gene therapy is contributing another kill strategy: Chimeric Antigen Receptor (CAR) T cells, which are genetically modified to efficiently target virus-infected cells for destruction. Blake Rust presented results of a small CAR T cell experiment in SHIV-infected macaques at CROI, which showed some potential to control viral replication—one animal experienced an 89 day delay in viral load rebound after ATI, and two out of the four total animals displayed post-ATI suppression of viral load. Other research groups are also pursuing the idea.
At the Pre-CROI Community HIV Cure Research Workshop—an event co-sponsored by TAG that was held via webinar this year—James Riley described the status of efforts to translate the CAR T cell approach for use in people with HIV. A pilot trial combining CAR T cells with CD4 T cells gene-modified to lack the CCR5 co-receptor is now underway the University of Pennsylvania.
The lack of success observed to date with kick and kill may seem disheartening, but it does not necessarily spell doom—the ever-improving understanding of the biology of HIV latency combined with the expanding array of therapeutic candidates suggests the stubborn persistence of the HIV reservoir can yet be overcome.
Comments