Rallying CROI

The 24th Conference on Retroviruses and Opportunistic Infections, which took place from February 13-16, is the preeminent US-based scientific meeting on HIV research.

By Richard Jefferys and Michael Dumiak

The annual Conference on Retroviruses and Opportunistic Infections (CROI) oscillates back and forth between the east and west coasts of the US from year to year. In 2017 it was Seattle that played host. Much of the science presented there represents subtle advances toward treating, preventing, and curing HIV infection, rather than the showstoppers that were a mainstay in some years. The HIV field seems poised to benefit from slow and steady progress, whether it be in the use of broadly neutralizing antibodies (bNAbs) to design vaccine immunogens or prevent infection through passive administration, or the use of the therapeutic vaccination in cure research.

The challenges and promise of bNAbs

Preventive HIV vaccine research occupied a relatively small plot on this year’s conference agenda. But what discussion there was on vaccines centered primarily around bNAbs. William Schief, professor in the Department of Immunology and Microbiology at The Scripps Research Institute (TSRI) and director of Vaccine Design for IAVI’s Neutralizing Antibody Center at TSRI, addressed efforts to create vaccine immunogens capable of eliciting these highly desirable antibodies, and did not sugarcoat the magnitude of the challenge. The task involves tracing the long and tortuous maturation pathway that B cells follow in order to produce bNAbs in some rare HIV-infected individuals, and then attempting to figure out if immunization regimens might be able to guide B cells down that pathway in the broad swath of the human population that needs an effective HIV vaccine. “This is something nobody has ever had to worry about for a vaccine before,” Schief emphasized.

He then reviewed the preliminary progress that has been made toward this goal with a class of bNAbs that target the CD4 binding site, so-called VRC-01 class antibodies (see page 17). An initial assessment of the likely frequency of the naïve B-cell precursors that can give rise to these antibodies suggest they are present at a reasonable frequency in the human population (Science 351, 1458, 2016). “Basically everybody in this room, in your resting lymph nodes, you have about 90 VRC01-class B-cell precursors,” Schief explained.

Schief and colleagues have designed an immunogen that contains 60 copies of an engineered CD4 binding site on a nanoparticle, dubbed eOD-GT8 60mer, capable of activating VRC01-class B-cell precursors, and shown that it appears to accomplish this task in mouse models (Science 353, 1557, 2016). A Phase I clinical trial of this vaccine immunogen in humans is scheduled to begin in the first quarter of 2018. It will be tested along with GlaxoSmithKline’s ASO1B adjuvant, which has been tested with other vaccine candidates, including those for malaria and herpes zoster.

Now researchers have turned their focus to investigating the best strategy for boosting the response induced by eOD-GT8 60mer in hopes of taking B cells further down the path toward producing VRC01-like bNAbs (Cell 166, 1459, 2016). The US Food and Drug Administration has indicated that when candidate boosters are ready for human trials they can be tested in individuals who have already received the eOD-GT8 60mer prime, rather than first being required to undergo assessment alone.

Schief also offered a glance at work assessing whether the same strategy can be employed to try and generate antibodies resembling the potent bNAb PGT121. While VRC01 interacts with the CD4 binding site of the HIV Envelope (Env), PGT121 is specific for a region known as the N332-supersite, which has been shown to be a common target in HIV-infected individuals who develop bNAbs (PLoS Pathog. 12, e1005369, 2016). In a pair of papers published last year, Schief and colleagues reported the successful design of stabilized Env trimers that were able to activate the inferred naïve precursors of PGT121-producing B cells (Immunity 45, 483, 2016). When given in an immunization series involving Env trimers with progressively fewer mutations compared to the wild-type or native Env trimer, antibodies capable of neutralizing tier-2 HIV isolates were generated (Cell 166, 1445, 2016). However, Schief stressed that these experiments involved mice genetically manipulated to only possess precursors of PGT121-producing B cells. The frequency of appropriate naïve B-cell precursors in humans is currently being investigated.

Michel Nussenzweig, the Zanvil A. Cohn and Ralph M. Steinman professor at Rockefeller University and Schief’s collaborator on the PGT121 studies, discussed clinical studies involving passive administration of bNAbs to either help control or prevent HIV infection. Nussenzweig’s group has conducted several trials involving passive administration of the bNAbs 3BCN117, which targets the V3 glycan supersite, and 10-1074, which is a CD4 binding-site targeting antibody. Both demonstrated antiretroviral activity after single infusions in HIV-infected individuals not on antiretroviral therapy (ART; Nature 522, 487, 2015; Nat. Med. 23, 185, 2017). 3BCN117 has also been studied in the context of treatment interruption: administration of four doses was associated with a significant delay in viral load rebound, averaging around 10 weeks (Nature 535, 556, 2016); a longer delay than that documented in similar studies with VRC01 (N. Engl. J. Med. 375, 2037, 2016), Nussenzweig noted. However, the rapid development of resistance is a problem with bNAbs administered singly, so trials are now exploring combinations.

Nussenzweig gave a preview of an unpublished macaque experiment conducted in a partnership with Malcolm Martin at the National Institute of Allergy and Infectious Diseases, in which a combination of 3BCN117 and 10-1074 was given for two weeks (three injections per week) starting three days after infection with the pathogenic hybrid SIV/HIV strain SHIVAD8. Interestingly, many of the animals have maintained low viral loads and preserved CD4+ T cells for many months after the administration of the bNAb combination, suggesting a prolonged beneficial impact on virus-specific immunity. Depletion of CD8+ T cells has been performed in some of the macaques, which caused an increase in viral load. This further supports the idea that the short-term bNAb intervention promoted immunological control of the SHIVAD8 challenge virus. The results were published shortly after the conference (Nature 543, 559, 2017). Nussenzweig pointed to the potential for bNAbs to modulate immunity in various ways, including via antibody-dependent cellular cytotoxity (ADCC) and the formation of immune-stimulating antibody-antigen complexes; these capacities underlie the mounting interest in studying bNAbs in the context of HIV cure research (see Best in Class).

With regard to prevention, Nussenzweig’s lab was involved in macaque studies, again in collaboration with Martin, which demonstrated the rationale for the ongoing trials of passive immunization with VRC01 for the prevention of HIV infection (the Antibody Mediatied Prevention studies, a joint effort of the HIV Prevention Trials Network and the HIV Vaccine Trials Network). In this case macaques received a single infusion of one of four bNAbs—3BCN117, 10-1074, VRC01, or a long-acting version of VRC01 (VRC01LS)—followed by weekly, low-dose intra-rectal challenges with SHIVAD8 until viremia was detected. The median time to infection in control animals was a little over three weeks, whereas significant delays in acquisition were observed in all bNAb recipients, ranging from eight to 14.5 weeks. Importantly for the human trials, protection was achieved at low antibody concentrations (Nature 533, 105, 2016).

Zimbabwe's example

CROI’s opening session put forth someone who knows quite a bit about progress in dealing with HIV. The Sudanese cardiologist and HIV researcher James Hakim, who gave the annual N’Galy-Mann Lecture, described his move to Zimbabwe in the early 90s and how he organized an increasingly robust HIV research network, establishing the University of Zimbabwe’s Clinical Research Center, and becoming head of the National Institute of Allergy and Infectious Diseases’ (NIAID) AIDS clinical trial unit in Harare. For a country still torn by political violence and wild economic gyrations, Zimbabwe has made dramatic progress. The country collects a three percent tax on personal and corporate income for a national AIDS Trust Fund (the AIDS Levy). This money, along with help from the US President’s Emergency Plan For AIDS Relief (PEPFAR) and the Global Fund to Fight AIDS, Tuberculosis and Malaria, helps provide antiretroviral therapy (ART) to over 900,000 people out of a population of 1.2 million living with HIV in the country. This corresponds to 86.8 percent of those in need on treatment, with 86.5 percent virally suppressed, almost reaching the United Nations goal of 90-90-90 (90 percent diagnosed, 90 percent on treatment, and 90 percent of those virally suppressed) set in 2014 as a means toward ending the AIDS epidemic. “Progress towards these targets is testament to the robust global coalition to fight AIDS and the resilience of the Zimbabwean people,” Hakim said. —MD

In pursuit of a cure

The pursuit of an HIV cure has gained momentum and a higher profile at CROI over the past decade, and the trend continued this year. One of the presentations that drew intense attention highlighted a possible role for therapeutic vaccination as a means to suppress HIV viral load in the absence of ongoing ART. The study was described by Beatriz Mothe, HIV Unit associate investigator at the Institut de Recerca de la Sida (IrsiCaixa) in Barcelona Spain, and involved HIV vaccine candidates designed by Tomas Hanke and colleagues at the University of Oxford that encode antigens designed to focus T-cell immune responses on conserved regions of the virus (primarily from Gag, Pol, Env, and Vif proteins).

In an initial trial, 24 individuals who began ART within three months of becoming HIV infected received immunizations with two vector-based candidates: a prime derived from a chimpanzee adenovirus (ChAdV63), and a booster from a modified Vaccinia Ankara strain (MVA). Both vaccines carried the conserved HIV antigen inserts, dubbed HIVconsv. As reported in a poster at last year’s CROI, the vaccine regimen successfully induced T-cell responses to the included antigens—close to half of all detectable HIV- specific T cells were targeting those regions of the virus (CROI 2016, Abstract 320). Fifteen of the participants (14 men and one woman) were then enrolled in a follow-up protocol. These participants received three infusions of the anti-cancer drug romidepsin, a histone deacetylase (HDAC) inhibitor that has been shown to induce replication of latent virus in the reservoir, in combination with additional boosters of the MVA HIVconsv vaccine given before and after the drug.

Eight weeks after the final MVA immunization, all participants interrupted ART until their viral load exceeded 2,000 copies/mL. So far, 13 have reached this stage, with eight of them rapidly experiencing viral load rebounds necessitating ART reintroduction. Five individuals, however, have been able to maintain low viral loads, with three below the limit of detection (20 copies/mL). One of these individuals has maintained a suppressed viral load for over six months. Mothe pointed out that while cases of post-treatment control of viral load have been reported in some studies of early ART, the frequency has been around 10 to 15 percent at most, compared to 38 percent of the cohort in this trial.

Additional studies are underway to try and better understand this outcome, but Mothe noted there is some evidence that the vaccine-induced T-cell responses to conserved HIV antigens, which were successfully boosted by the additional MVA immunizations, are contributing. The role of romidepsin will likely be challenging for researchers to tease out because there was no control group. The drug did not have a measureable effect on the size of the HIV reservoir overall, based on pre- and post-administration comparisons of HIV DNA levels. It also caused an array of side effects that are known to be associated with HDAC inhibitors, including one case of sepsis.

Transient low-level increases in HIV viral load during romidepsin infusions suggest it did, however, exert latency-reversing effects. But Mothe also highlighted that 60 percent of participants saw similar blips of viremia after receiving the MVA vaccine candidate, which is consistent with evidence that latent virus can reside in HIV-specific CD4+ T cells and therefore may be stimulated to replicate by HIV antigens (AIDS Res. Hum. Retroviruses 28, 835, 2012). Despite the small sample size and remaining unknowns, the trial represents something of a milestone—it is the first time an evaluation of the strategy known as “kick and kill,” which combines a latency-reversing agent with an immune-enhancing approach, has been associated with an increased frequency of viral load suppression after ART interruption.

Jintanat Ananworanich, associate director for Therapeutics Research at the US Military HIV Research Program (MHRP), emphasized the rarity of post-treatment control in a talk about a cohort of extremely early treated individuals (now numbering over 400 individuals) who have been recruited in Bangkok, Thailand. Ananworanich and colleagues have previously reported that study participants at Feibig stage I—estimated to be within 10-17 days of HIV acquisition, and diagnosed by having detectable HIV RNA but negative antibody and p24 antigen tests—have a significantly smaller viral reservoir than individuals diagnosed later (J. Virus Erad. 2, 43, 2016). The logical question that flowed from this observation was whether the initiation of ART at Feibig stage I might be associated with an increased possibility of achieving virological remission when ART is interrupted.

Eight individuals (seven men and one woman) were recruited in a study designed to address this question. A small group was chosen to minimize any negative consequences of treatment interruption with contingency plans to enroll an additional seven volunteers if at least one case of post-treatment control to a viral load of less than 50 copies/mL was documented in the initial group. The median time on ART prior to interruption was 2.8 years. Viral load was monitored every three to seven days and the criteria for restarting ART were two viral load measurements above 1,000 copies/mL. All participants rebounded after a median of 26 days (with a range of 13 to 48 days). There were no significant safety issues, and while the size of the HIV DNA reservoir transiently increased, it declined to levels similar to baseline when treatment was restarted. However, Ananworanich noted that four of six individuals who were initially non-reactive on HIV antibody tests, due to how quickly after infection they had begun ART, seroconverted and subsequently remained HIV positive by traditional antibody tests after the interruption.

Ananworanich also highlighted MHRP’s involvement in a clinical trial underway in adults employing therapeutic vaccines based on adenovirus serotype 26 (Ad26) and MVA vectors. The next step will be to combine vaccination with a toll-like receptor (TLR)7 agonist, as this approach has shown promise in the simian immunodeficiency virus (SIV)/macaque model (Nature 540, 284, 2016).

What a difference a day makes

In a symposium dedicated to cure research, Louis Picker, associate director of the Vaccine & Gene Therapy Institute at Oregon & Health Science University, provided an update on work involving his much-discussed cytomegalovirus (CMV)-based vaccine vector. As has been extensively documented, a version of the vector encoding SIV antigens reliably leads to robust control of pathogenic SIV challenges in half the macaques that receive it (Nature 473, 523, 2011). This control of SIV is associated with the induction of unusually broad CD8+ T-cell responses, some of which are class II-restricted—an antigen presentation pathway typically thought to only be utilized by CD4+ T cells (Science 340, 1237874, 2013). Many of the animals in these studies even appear to eventually clear SIV infection (Nature 502, 100, 2013), and Picker noted there are two possible explanations for this outcome: either the SIV reservoir is progressively eliminated by vaccine-induced SIV-specific immune responses, or these immune responses initially limit the formation of the virus reservoir to such an extent that the size eventually decays to zero over time.

In an attempt to distinguish between these possibilities, Picker and colleagues conducted an experiment in which the CMV-based SIV vaccine was tested as a therapeutic intervention in macaques started on ART at various times after SIV infection. A version of the vaccine encoding tuberculosis antigens was used as a control. The reasoning was that if vaccine-induced T-cell responses were capable of clearing the SIV reservoir, then immunization might lead to a lack of viral load rebound after ART withdrawal. But this was not what the researchers found. Receipt of the CMV-based SIV vaccine was not associated with prevention of SIV rebound, leading Picker to conclude that, in the prevention setting, the approach is working by limiting the formation of the viral reservoir rather than actively clearing it.

The study did produce some interesting observations, however. A feature of the experimental design called for ART to be initiated in the macaques when an inflammatory signature indicating monocyte activation, which was associated with initial establishment of the reservoir in prior studies, was observed. This made it possible for Picker to segregate macaques into different groups depending on how many days after SIV challenge ART was first administered: days 4/5, day 6, day 7, days 8/9, or day 12.

Picker showed that the six animals treated starting on days 4/5 had dramatically lower viral load peaks (which rose by a log in each subsequent group) and significantly smaller SIV reservoirs compared to those treated later. Furthermore, all the macaques in days 4/5 group did not display any rebound of SIV viral load when ART was interrupted after 600 days. Only one of the remaining 35 animals showed a similar lack of rebound, and that was one out of 13 macaques in the day 6 group.

To test whether CD8+ T cells were controlling SIV replication, a depletion experiment was conducted, but no return of SIV viral load was observed. Cells transferred from the non-rebounding macaques to naive animals were also unable to establish an infection. Necropsy studies revealed some evidence of SIV DNA, but no replication-competent virus could be detected. The sole exception was the macaque from the day 6 group, in which a rebound of SIV viral load occurred eight months after ART was stopped, just prior to necropsy. Picker drew a parallel between this animal and the famous human cases of HIV remission that have been described, most notably the Mississippi baby, in whom no sign of HIV could be detected for over two years after an ART interruption before viral load ultimately rebounded (N. Engl. J. Med. 372, 786, 2015).

Picker concluded that the SIV reservoir established within the first five days of infection is likely unstable, perhaps because the virus has not yet entered long-lived cells, and can wane during sustained suppression of viral replication by ART. But just one day appears to make a difference. Data from this study indicates that the reservoir had become more permanently established in the macaque that eventually rebounded. As has been suggested by the example of the Mississippi baby, latently infected cells can clearly persist in an inactive state for extended periods, making it challenging to establish whether a cure has been achieved and, Picker argued, supporting the need for strategies that attempt to expose and deplete the latent reservoir.

In addition to the Mississippi baby, the two other widely publicized cases of HIV remission in humans are known as the Boston patients. These were two HIV-infected individuals who underwent stem cell transplants required for the treatment of concomitant cancers while remaining on ART. Post-transplant, after their immune systems had been successfully reconstituted by the donor stem cells, HIV reservoirs could no longer be detected. Both ultimately underwent an ART interruption with careful monitoring, and no HIV could be detected for three and eight months, respectively, before a sharp and sudden rebound in viral load occurred accompanied by symptoms of acute retroviral syndrome, necessitating the reinstitution of treatment (Ann. Intern. Med. 161, 319, 2014). At CROI, a poster presentation from the research group of Nathan Cummins, assistant professor of medicine at the Mayo Clinic in Rochester, Minnesota, described an additional HIV remission case similar to the Boston patients.

This individual underwent a stem cell transplant for the treatment of acute lymphoblastic leukemia and afterward displayed a dramatic reduction in measures of the viral reservoir, which were extremely low or undetectable from day 56 post-transplant. ART was eventually interrupted, and no HIV could be detected for 288 days, at which point the virus reemerged. The rebound was slower and less dramatic than in the Boston patients and symptoms were not observed, but the resumption of ART was nonetheless required. This new case underscores Picker’s point that latently infected cells can persist for long periods in a dormant state before a reactivating event occurs. The presenter of the poster, Stacey Rizza, associate professor of medicine at the Mayo Clinic, noted that the individual had been involved in a car accident shortly before the viral load rebound, leading to speculation that while he was not seriously hurt, perhaps stress-related inflammation precipitated the activation of a latently infected cell or cells.

Worlds within worlds

HIV researchers over the last few years are turning more and more attention to microbial environments, particularly the single-celled flora in the intestines. After all, HIV is a disease of the gut. But scientists are now on to other biomes: the biota found in the vagina. Last summer in Durban researchers presented data suggesting that women carrying the uncommon Prevotella Bivia bacteria are 19 times more likely to have genital inflammation and 13 times more likely to contract HIV than those without this kind of vaginal bacteria. Variants in the vaginal biome also seem to play a role in the varying effectiveness of a vaginal gel used as pre-exposure prophylaxis (PrEP), but not when PrEP is administered orally.

The vaginal microbiome continues to be an environment attracting great interest. Scott McClelland, an epidemiologist at the University of Washington, cites bacterial vaginosis as a longtime indicator of increased HIV acquisition risk. But McClelland emphasized that while the composition of the vaginal microbiome may boost the chances of acquiring HIV, women with standard microbiota can still become infected. “Women do not have to have a particular microbial profile to acquire HIV,” he said, adding that it’s possible that treating bacterial vaginosis may alter these odds, but further research is needed.

Women with greater number of partners and more frequent sexual activity are at greater risk of bacterial vaginosis, and, incidentally, of HIV, said Sharon Hillier, director of reproductive infectious disease research at the Magee-Women’s Hospital of the University of Pittsburgh School of Medicine. It raises the possibility, Hillier said, that the relationship between a vaginosis-type biota and the reduced effectiveness of topical PrEP may have less to do with the bacteria itself than a coincidental lower adherence to using the gel or unmeasured differences in behavior.

Therapy and resistance

Another focus at CROI was the further refinement of antiretroviral (ARV) therapy with novel, more convenient drugs. The San Diego biotech and pharma company Gilead is pursuing an investigational integrase inhibitor called Bictegravir. The goal is to combine the integrase inhibitor with the company's antiretrovirals emtricitabine and tenofovir, which are already combined in the single pill known as Truvada. The only existing three-drug combination pill is GlaxoSmithKline’s (GSK) Triumeq, which is produced by GSK subsidiary Viiv. Gilead is also testing a novel compound which interferes with the assembly and disassembly of the HIV capsid, the part of the virus that holds its genetic blueprint. Though it’s still in beginning stages, Gilead reports this investigational ARV has the potential to be a low-dose, long-acting injectable treatment. New ARVs may become critical if the spread of drug-resistant virus widens. Last year for the first time more than 50 percent of people living with HIV had access to treatment, said Anne Derache, a post-doc at the new Africa Health Research Institute in Mtubatuba, in KwaZulu-Natal, South Africa. Derache examined results from the ANRS 12249 TASP trial, a randomized study evaluating the effect of early ARV therapy initiated regardless of the CD4+ T-cell count in HIV-infected individuals in Hlabisa, KwaZulu-Natal to assess drug-resistant virus. The results showed the prevalence of pre-treatment drug resistance was about nine percent for both recently and chronically infected volunteers. If this number were to reach 15 percent, Derache warned, that would compel a change in the composition of first-line treatment. “We need to perform surveillance studies of resistance before antiretroviral therapy initiation,” she said. —MD

Richard Jefferys is Coordinator, Michael Palm Basic Science, Vaccines & Prevention Project at the Treatment Action Group. Michael Dumiak reports on global science, technology, and public health and is based in Berlin.