Skip to main content

CROI: Progress on Prevention and Cure

The 21st edition of the Conference on Retroviruses and Opportunistic Infections thawed the chilly climes with encouraging news on several fronts

By Richard Jefferys

The headlines from last year’s Conference on Retroviruses and Opportunistic Infections (CROI) were dominated by the report of an infant from Mississippi who may have been cured of HIV infection as a result of receiving antiretroviral therapy (ART) soon after birth (see A Toddler Stole the Show, IAVI Report, Spring 2013). This year, when experts from all avenues of HIV research gathered in Boston March 3-6, there was an echo of this media attention when details on another possibly cured baby were presented, along with an update revealing the child in Mississippi remains off ART with no sign of a return of HIV replication.

While this news stoked enthusiasm over HIV cure research, the field also received sobering news about two individuals in Boston who had experienced a prolonged absence of detectable virus after receiving stem cell transplants for concomitant cancers; HIV remained undetectable for an extended period after ART interruption, but ultimately rebounded to high levels requiring treatment reinitiation.

In the prevention realm, three studies of long-acting antiretrovirals (ARVs) offered hope of more user-friendly pre-exposure prophylaxis (PrEP) options, and additional compelling evidence emerged that for HIV-infected individuals, suppression of viral load by ART massively reduces, or even obliterates, the risk of HIV transmission. Vaccines contributed only quietly to the busy symphony of science this year, primarily through the incremental advances reported in the accelerating quest to induce broadly neutralizing antibodies (bNAbs) against HIV.

Baby makes two?

Following on the apparent success of curing an HIV-infected baby in Mississippi with early initiation of ART, Deborah Persaud, associate professor of pediatrics at Johns Hopkins University School of Medicine, presented the case of a second baby who for now remains on ART, but has no detectable virus.

In this case the baby’s mother had relatively advanced HIV infection at the time of delivery, with a viral load of 138,811 copies and a CD4+ T-cell count of 70—she had been prescribed ART but had not adhered to the regimen. Although ART was administered during labor, the baby was ascertained to be HIV infected based on a positive HIV DNA test at four hours of life, with a viral load level of 217 copies/ml of blood at 36 hours, and another viral load measurement of 32 copies in a cerebrospinal fluid sample drawn at day six as part of a work up for sepsis.

Based on knowledge of the Mississippi baby, infectious disease specialist Audra Deveikis at Miller Children’s Hospital in Los Angeles County instituted a treatment regimen including the ARVs AZT, 3TC, and nevirapine at four hours after birth, which was supplemented with lopinavir two weeks later. Nevirapine was discontinued after 3.4 months, and the infant is now nine months of age and remains on AZT/3TC/lopinavir. Persaud, who was at the center of the media maelstrom generated by the first baby reported on last year at CROI, took pains to emphasize that ART has not yet been interrupted in this case.

However, the possibility of a cure is suggested by the results of virological analyses. The baby’s viral load became persistently undetectable (<20 copies/ml blood) after 11 days of follow up, and multiple subsequent tests for HIV DNA and replication-competent virus were all negative. Persaud noted that the viral outgrowth assay used to measure replication-competent virus did suggest the presence of low levels of integrated HIV proviruses that did not replicate after stimulation (described as “non-induced proviruses”) at one month of age, but not afterward. The technique of digital droplet polymerase chain reaction (PCR) for HIV DNA (see In Pursuit of a Cure, IAVI Report, Jan.-Feb. 2012) showed levels below 1.6 copies per million peripheral blood mononuclear cells (PBMCs) at the most recent time point. The infant now tests HIV antibody negative. Doctors are considering interrupting treatment if HIV remains undetectable at two years of age.

During the same presentation, Persaud provided an update on the child in Mississippi, who is now 41 months old and has been off ART for close to two years. HIV RNA continues to be undetectable by extremely sensitive viral load tests capable of measuring as little as one to two copies of virus per ml of blood, and no replication-competent virus can be found. However, HIV DNA has been consistently detectable at low levels that border on the detection limits of the assay, averaging around four copies per million PBMCs. To assess whether cells transferred from the mother might be the source of the HIV DNA, samples were evaluated for the presence of maternal alleles. One allele could be detected at an extremely low frequency (0.001%) at 24 months, but not at 40 months of age. Immunological analyses have documented no HIV-specific antibody or CD8+ T-cell responses. Persaud characterized the outcome not as a cure but rather as ongoing “remission” of HIV infection.

Based on the cases of these two babies, the International Maternal Pediatric Adolescent AIDS Clinical Trials Network (IMPAACT) is planning to launch a study of very early treatment in babies born to mothers who have not received adequate prevention against mother-to-child HIV transmission.

Diminishing reservoir with early treatment

In the vast majority of cases, where HIV is not cleared so dramatically, early treatment may still reduce the long-lived reservoirs of HIV that form quickly after infection, and typically reignite viral replication if treatment is interrupted. In a separate talk in the session on pediatrics, Persaud described results from a cohort study of 144 children with a median age of 14.3 years who were divided into three groups based on when viral load suppression was achieved—14 within one year after birth, 53 within 1-5 years, and 77 after five years. The amount of HIV proviral DNA detected in each group was starkly different: 4.2 copies, 19.4 copies, and 70.7 copies per million PBMCs, respectively. Similarly, earlier suppression of virus was associated with a significantly increased likelihood of indeterminate or negative results on HIV antibody tests—proportions were 86%, 19%, and 3%, respectively. Persaud emphasized that the relatively limited size of the reservoir in early treated children and adolescents makes them good candidates for inclusion in cure research studies because the number of latently HIV-infected cells that need to be cleared is likely very small.

In addition to the excitement generated by the reports of the now possibly two babies cured of HIV infection because of extremely early initiation of ART, a poster presentation about a possible adult equivalent also drew much attention. The case report by Hiroyu Hatano, assistant professor of medicine at the University of California-San Francisco School of Medicine, involved a man who was recruited into a PrEP demonstration project in San Francisco and acquired HIV infection during the 13-day interval between his last screening visit and the initiation of the ARV Truvada for PrEP.

A baseline sample taken on the day he started Truvada showed a viral load of 220 copies. As soon as this result became available seven days later, he was switched to a conventional ART regimen. Viral load at the time of the switch was 120 copies and was followed by a reading of “detected” but below 40 copies after 22 days. A test for cell-associated HIV RNA measured 4.7 copies per million CD4+ T cells at 22 days. All subsequent viral load tests have been negative, and no HIV DNA or replication-competent virus can be detected now. Hatano described the case as representing treatment of “hyperacute” HIV infection. A treatment interruption is planned after one year to assess if the virus may have been cleared.

Does lack of virus equal cure?

While all this news was encouraging, other research suggests the road to an HIV cure will likely still be a long one. Timothy Henrich, associate physician at Brigham and Women’s Hospital and assistant professor of medicine at Harvard Medical School, presented data that should give pause to anyone inclined to assume that the absence of detectable HIV equates to a cure.

Henrich previously reported on two HIV-infected individuals from Boston, who after receiving stem cell transplants to treat cancers while on continuous ART showed no measurable levels of virus over a period of several years (J. Infect. Dis. 207, 1694, 2013). Designated patients A and B, one had acquired HIV infection perinatally and required the transplant for recurrent Hodgkin’s lymphoma, while the other became HIV infected as a result of sexual transmission and had a diagnosis of myelodysplastic syndrome. Both were originally heterozygous for the CCR5Δ32 mutation, but received stem cells from wild-type individuals, unlike Timothy Brown, the lone adult considered cured of HIV, who received a transplant from a CCR5Δ32 homozygous donor.

Last spring, ART was interrupted in both patients A and B under a careful research protocol involving monitoring for viral load rebound every one to two weeks. Early signs were promising, as multiple tests continued to find no detectable HIV RNA or DNA. But after twelve weeks, patient A had a viral load of 900 copies, which rapidly increased to 127,000 copies just three to four days later. A drop in CD4+ T-cell count was also documented. Although ART was reinstituted, patient A only stayed on therapy for several days due to problems with adherence, resulting in a viral load that peaked at several million copies, along with accompanying symptoms of acute HIV infection including fever, malaise, and aseptic meningitis. ARV resistance was detected and a new treatment regimen was initiated that has since led to HIV suppression and an ongoing increase in CD4+ T-cell counts.

HIV remained undetectable in patient B for a longer period of eight months, but about a week after the last negative virology results, symptoms of acute HIV infection developed and viral load was measured at just under two million copies. ART was immediately restarted and HIV replication was successfully controlled. Sequencing experiments are ongoing, but Henrich noted that there is a clear relationship with the original infecting virus, ruling out any possibility of a new infection or superinfection. Samples taken five days after HIV rebounded in patient A also show an extremely homogenous viral population, suggesting that the source was perhaps just one latently infected cell.

Henrich suggested the rapid development of acute infection symptoms in both of these patients was due to the new, transplanted immune system not having encountered HIV antigens prior to the recrudescence of viral replication. Analyses of HIV-specific cellular immune responses by Marcus Altfeld at the Ragon Institute supported this suspicion—they were undetectable after the stem cell transplants and did not emerge until after viral load rebounded (after 108 days in patient A, and 13 days in patient B).

Henrich highlighted several lessons from these data. The estimated magnitude of the reduction in HIV reservoirs in these two individuals was at least three logs, but this only achieved a delay of viral load rebound, indicating the presence of long-lived sources of replication-competent virus in tissues that were not sampled in the study. And despite the potential risks, analytical ART interruptions followed by long-term clinical monitoring represent the only definitive means for establishing whether a cure of HIV infection may have been obtained.

Shortcomings of single agents for reversing latency

The central challenge in HIV cure research is eradicating the long-lived pool of cells containing latent, integrated HIV proviruses. A variety of strategies are being explored that might be able to awaken this slumbering reservoir, thereby facilitating the elimination of the infected cells. One strategy involved the use of so-called latency-reversing agents (LRAs), and among these are histone deacetylase (HDAC) inhibitors, a class of cancer drugs that target the cellular proteins involved in locking down integrated HIV and preventing viral gene expression in latently infected cells. HDAC inhibitors appear able to reverse latency in laboratory models, and human trials have suggested that the US Food and Drug Administration (FDA)-approved HDAC inhibitor, vorinostat, may be able to induce intracellular RNA production by latent HIV in ART-treated individuals (Nature 2012, doi: 10.1038/nature11286).

However, at CROI, Gregory Laird, a graduate student in the laboratory of Robert Siliciano at Johns Hopkins University, reported findings indicating single LRAs have essentially no effect on latently infected CD4+ T cells sampled from HIV-infected people on suppressive ART. The rarity of these cells make this type of study challenging—their frequency is around one in a million CD4+ T cells—but sufficiently large volumes for this study were collected from volunteers via apheresis.

Laird tested a panel of candidate LRAs including three HDAC inhibitors (vorinostat, romidepsin, and panobinostat), the acetaldehyde dehydrogenase inhibitor disulfiram, the bromodomain inhibitor JQ1, and a protein kinase C (PKC) agonist, Bryostatin 1.

Results were compared to those achieved by activating CD4+ T cells with phorbol 12-myristate 13-acetate (PMA)/ionomycin, which is known to maximally induce replication of latent HIV. Outgrowth of replication-competent HIV was measurable after PMA/ionomycin stimulation in CD4+ T cells from 11 of 13 donors, whereas, no virus was measurable in any of the samples exposed to LRAs. Similarly, an assay that quantified extracellular HIV messenger RNA (mRNA) in the cultures could detect none after exposure to LRAs, with the sole exception of a CD4+ T-cell sample from one individual in which viral mRNA became detectable in response to Bryostatin 1. The last assay Laird looked at was for intracellular HIV mRNA, a measure that has been used to assess activity in human trials of LRAs. Low levels were apparent at baseline in the majority of cases, and only increased significantly after treatment with Bryostatin 1 and PMA/ionomycin.

Although HIV mRNA increases were documented in the published clinical trial results of vorinostat, Laird pointed out that care needs to be taken with this particular assay because HIV mRNA can be generated as a result of the transcription of the human genes into which HIV has integrated. This phenomenon is distinct from the activation of the virus itself, as it does not lead to HIV replication and the production of virions, the desired goal of LRAs. HIV mRNAs generated by the activity of human genes are referred to as readthrough transcripts. Laird has developed an assay to specifically measure HIV readthrough transcripts, and in an analysis involving five donors, Laird showed that their levels are increased by vorinostat, suggesting that this mechanism likely accounts for the previously reported HIV mRNA increases in recipients of the drug. These analyses were published shortly after CROI ended (Nat. Med. 2014, doi: 10.1038/nm.3489).

Having called into question the value of single LRAs in HIV cure research, Laird concluded his talk on a more upbeat note by offering a glimpse at new data generated with combinations of LRAs. In several cases, significant synergy was observed, particularly when either romidepsin or panobinostat was combined with Bryostatin 1. Laird found levels of latency reversal with these duos were close to that achieved with PMA/ionomycin. While Bryostatin 1 may be too toxic for use in people infected with HIV, several analogs and other compounds from the PKC agonist class are now being studied.

Memory T-cell proliferation expands HIV reservoir

The persistence of memory CD4+ T cells is normally beneficial because these cells are specific for previously encountered pathogens and help prevent disease in the case of second exposure. The induction of CD4+ T-cell memory also makes an important contribution to the lasting effectiveness of many vaccines. But because HIV establishes latency in memory CD4+ T cells, their long-lived nature is actually a key obstacle to curing the infection. One of the mechanisms by which CD4+ T cells persist is proliferation; the cells divide, making new copies, which can in turn later repeat the process. Several studies at CROI identified proliferation as a mechanism by which latently HIV-infected CD4+ T cells can increase in number and persist.

Mary Kearney from Frank Maldarelli’s laboratory at the National Cancer Institute reported in a poster that 7%-55% of HIV DNA sequences identified in 14 study participants were identical, and that the numbers were seen to increase in chronically infected individuals initiating treatment at low CD4 counts (<100 cells). This finding suggests that when CD4+ T-cell numbers increase in response to HIV suppression, some of the proliferating CD4+ T cells are latently infected and so the integrated HIV is copied right along with the genome of the cell. In some cases, Kearney’s results indicate, more than half of the detectable HIV DNA reservoir can be derived from the proliferation of a single latently infected CD4+ T cell.

Frank Maldarelli presented a complementary late-breaker poster suggesting that the specific genes into which the virus integrates may also contribute to the persistence of latently infected CD4+ T cells. Thor Wagner, assistant professor of pediatric infectious diseases at the University of Washington, echoed this theme in his presentation. He has also observed preferential integration of HIV into cellular genes associated with proliferation and survival. Wagner also speculated that HIV integration is affecting the function of these genes in ways that promote CD4+ T-cell survival.

Alarming HIV Rates Among Black MSM  

Reports have repeatedly shown high rates of HIV infection among black men who have sex with men (MSM) in the US, and at this year’s Conference on Retroviruses and Opportunistic Infections, researchers began asking when something will actually be done to address the growing disparity between black and white MSM.

Eli Rosenberg, assistant professor in the department of epidemiology at Emory University, pointed out that the severe impact of HIV infection on black MSM in the US has not been linked to differences in behaviors, suggesting more research is needed to better understand the factors contributing to HIV risk in this population. To this end, Rosenberg and colleagues conducted a prospective study comparing HIV incidence in black and white MSM in Atlanta. Baseline testing documented an HIV prevalence of 44% and 13%, respectively.

An HIV-uninfected cohort of 260 black MSM and 302 white MSM were subsequently followed for an average of about 1.5 years. Rosenberg cited several significant differences between the two groups, including a relative lack of health insurance and higher rates of poverty among the black MSM. The results showed a stark difference in HIV incidence: 6.6 per 100 person-years in black MSM compared to 1.7 per 100 person-years in white MSM.

The differential was even more notable in the 18-24 year old age group: 12.1 per 100 person-years (representing 16 infections) versus one per 100 person-years (one infection). Rosenberg said that this translates to more than 1 in 10 young black MSM in Atlanta becoming HIV infected every year.

In a multivariate analysis, lack of health insurance and having a black partner were factors significantly associated with becoming HIV infected. Rosenberg said these findings add to the evidence that implicating social factors and HIV prevalence in sexual networks, not individual characteristics and behaviors, are the primary drivers of the distressing rates of HIV infection in the black MSM population. In the question and answer period following Rosenberg’s talk, Jeff Klausner from the University of California at Los Angeles called for greater urgency in moving from observational studies to implementing pre-exposure prophylaxis (PrEP) and other interventions that could avert HIV infections in black MSM. Klausner specifically lamented the unwillingness of Gilead—the pharmaceutical company that manufactures the ARVs Truvada and tenofovir, both of which have been shown to be effective PrEP drugs—to market PrEP to vulnerable populations and their providers, and drew applause when he concluded: “We need to think whether we can keep talking about observational studies showing this high incidence and then not doing anything about it.” –RJ

Improving PrEP with long-acting ARVs

Last year at CROI, results from an experiment in rhesus macaques offered an encouraging indication that a long-acting analog of the FDA-approved integrase inhibitor dolutegravir, named GSK744, could have a future as intermittently administered PrEP (see A Toddler Stole the Show, IAVI Report, Spring 2013). This year, three new studies bolstered this possibility, generating considerable excitement and a slew of mainstream media stories highlighting the preventive potential of the approach.

Chasity Andrews, a postdoctoral fellow at the Aaron Diamond AIDS Research Center in New York City, who presented the first macaque study last year, described new work that aimed to establish the minimum protective drug levels of GSK744 using the same model. Twelve macaques received a single 50mg/kg dose of GSK744 followed by weekly intra-rectal challenges with the hybrid simian/human immunodeficiency virus (SHIV)162p3, while four animals served as controls. Protection was sustained for at least five challenges, with between six and 17 challenges required to infect all the treated macaques.

Drug levels of GSK744 were then analyzed based on what is called the protein-adjusted 90% inhibitory concentration or PAIC90 (the concentration required to inhibit HIV replication by 90% in vitro, adjusted for the binding of the drug to proteins that occurs in vivo). This analysis demonstrated that levels three times the PAIC90 were completely protective, while those above the PAIC90 were 97% protective. Only one animal with drug levels above the PAIC90 became infected. Because human pharmacokinetic studies of injectable GSK744 have shown that levels are maintained at four times the PAIC90 for at least 16 weeks following an 800mg dose, Andrews was able to conclude that protective concentrations can be achieved by quarterly injections that are administered by two shots, one into each gluteal muscle. A Phase II safety and tolerability trial of this drug as PrEP in high-risk men who have sex with men (MSM) is slated to start this spring. Data from 480 volunteers to date suggest the drug is safe, with some volunteers reporting instances of dizziness and grade one rash, but nothing more serious.

Following Andrews’ talk, J. Gerardo Garcia-Lerma, a research microbiologist at the US Centers for Disease Control and Prevention (CDC), debuted findings from a study exploring the efficacy of GSK744 against vaginal transmission in rhesus macaques. An initial pharmacokinetic assessment of a single 50mg/kg dose in six female pigtailed macaques revealed that drug levels were maintained above the PAIC90 in vaginal tissue for at least a month, although the amount of GSK744 in these tissues was approximately four- to five-fold less than seen in plasma. This led to the selection of a monthly dosing regimen.

Garcia-Lerma explained that female pigtailed macaques were used because they have lunar menstrual cycles, allowing efficacy to be tested throughout all phases of the cycle. Six animals received three monthly GSK744 injections and six were given a placebo at the same time points. After the first dose, biweekly, low-dose, intra-vaginal challenges with SHIV162p3 were administered for a period of 12 weeks, followed by cessation of drug and continued monitoring for infection. All placebo recipients were infected after a median of two to four challenges, but no macaques in the GSK744-treated group acquired infection. Garcia-Lerma concluded that these results strongly support further efforts to establish the efficacy of GSK744 for PrEP in women.

In a complementary poster presentation by Chasity Andews’ group, protection against a single, high-dose intra-vaginal SHIV162p3 challenge was reported in macaques. This study employed a somewhat different model involving female rhesus macaques treated with the injectable hormonal contraceptive Depo-Provera to purposely thin the cervicovaginal epithelium and therefore mimic the most vulnerable phase of the menstrual cycle. Eight macaques were given 50mg/kg of GSK744 at the start of the study and four weeks later, and four animals served as controls. The high dose SHIV162p3 challenge was administered one week after the first dose of GSK744, with additional challenges at weeks five and seven in the GSK744 recipients. All control animals were infected after a single challenge, while six of the eight macaques in the GSK744 group remained uninfected throughout. In the two treated animals that became infected, viral load was first detected at three and seven weeks after the last challenge, respectively.

The development of long-acting ARVs that could represent potentially more user-friendly PrEP options is timely given that several trials have recently found that PrEP recipients appear to struggle with adherence to daily regimens. “Daily dosing is burdensome to the user,” said Ariane van der Straten, director of the Women’s Global Health Imperative at RTI International, an independent, nonprofit institute.

van der Straten described a new analysis of adherence measures in the four-arm VOICE trial, which compared vaginally administered tenofovir gel, oral tenofovir, and oral Truvada to placebo and was unable to show efficacy with any of the interventions. Multiple behavioral measures of adherence were used in the trial, including self-reports, interviews, and monthly returns of unused product, but none showed any correlation with biological measures of drug levels. Rates of non-adherence, defined as no product taken during the previous week prior to drug sampling, were high (>60%), but behavioral measures failed to reflect this reality.

Immune Activation and HIV Transmission  

Erin M. Kahle and colleagues from the Partners in Prevention herpes simplex virus (HSV)/HIV Transmission Study and the Couples Observational Cohort Study presented a poster at the Conference on Retroviruses and Opportunistic Infections describing a novel case-control analysis designed to assess whether levels of a variety of cytokines influenced HIV transmission risk.

Researchers compared 120 couples with genetically linked HIV transmission to 321 couples with no transmission events. Significant associations were reported for two cytokines linked with immune activation—interleukin (IL)-10 and interferon γ presenting (IP)-10—both in the partner acquiring HIV infection and the HIV-infected transmitting partner. The researchers suggest that further studies are needed to better understand the role of immune activation in determining both susceptibility to HIV and infectiousness. –RJ

Zeroing transmission

While it’s now well accepted that suppression of HIV by ART reduces the risk of sexual transmission of HIV, it is unclear how protective treatment is if condoms aren’t used. The PARTNER study, which has recruited over 1,000 discordant couples at 75 European clinical trial centers, was designed to address this question, and results of an interim analysis were presented at CROI by Alison Rodger, senior lecturer and honorary consultant in Infectious Diseases, Infection & Population Health at University College London.

Rodger described data from 767 couples that met criteria for inclusion in the analysis. The criteria included: reporting condomless sex in the partnership, a viral load below 200 copies in the HIV-infected partner, and reporting no PrEP or PEP (post-exposure prophylaxis) use in the HIV-uninfected partners. Of these, 282 were MSM, 245 were heterosexual couples in which the female partner was HIV infected, and 240 were heterosexual couples in which the male partner was HIV infected. Virus sequencing was used to confirm or rule out an epidemiological link for any infections that occurred in HIV-uninfected partners.

In the group of 767 couples, there were zero documented transmissions from HIV-infected partners on suppressive ART. However, Rodger stressed that the limited duration of follow up, an average of around a year per couple, still leaves some uncertainty. Based on the 95% confidence interval, the data are compatible with up to a 4% risk of transmission over ten years for any sex, or a 10% risk over ten years for anal sex. The PARTNER study is now being extended until 2017 for the MSM couples in order to try and narrow the confidence interval associated with the findings. Rodger noted that, in the absence of ART, around 50 to 100 HIV transmissions would have been expected in this group.

Stable trimer advances vaccine research

HIV vaccine researchers recently made a major advance in the pursuit of antibody-based vaccines—the creation of a stable envelope trimer. This work, done by the laboratory of John Moore, professor of microbiology and immunology at Weill Cornell Medical College (see Keystone in Rio: Breakthroughs, Predictions, and Surprises, IAVI Report, Winter 2013), is viewed as addressing one of the major roadblocks to developing vaccine candidates capable of inducing antibodies against this unstable protein structure, which is the target of the bNAbs that researchers are trying to induce through vaccination. Accurately recapitulating this critical component of HIV has been enormously challenging, with many attempts resulting in disappointment. Loosely speaking, stabilizing HIV’s envelop trimer has been the biochemical equivalent of baking a soufflé—researchers would think they had succeeded only to have it collapse into disarray the moment it was removed from the oven.

At CROI, Moore presented the first immunogenicity data generated with the new and better-behaved gp140 trimer protein, which is named BG505 SOSIP.664 and was derived from a clade A virus isolated from a Kenyan infant. Explaining the rationale for the experiment, Moore noted that previous gp140 immunization protocols have used versions of the molecule that unfold and do not retain a homogenous, consistent trimeric form (see diagram, below). BG505 SOSIP.664 is also sourced from a founder virus, and these viruses have been shown to typically generate strong autologous neutralizing antibody responses during primary infection—in around 25% of cases, these responses eventually evolve to become broadly neutralizing. The hope is that the new trimer will provide a platform for the design of immunogens capable of inducing bNAbs; however, the early data suggests there is still a long way to go.

Taming the Tricky Trimer  

Unlike previous best attempts which are unstable and display a diversity of shapes (PNAS 2014, doi: 10.1073/pnas.1319512111 and PNAS 2013, doi:10.1073/pnas.1314351110), BG505 SOSIP.644 (far right) shows a regular, homogenous conformation (PLoS Pathog. 2013, doi:10.1371/journal.ppat.1003618). Image courtesy of Andrew B. Ward of The Scripps Research Institute and John P. Moore.

In collaboration with Hansi Dean and colleagues at the International AIDS Vaccine Initiative, experiments were carried out in rabbits involving BG505 SOSIP.664 proteins generated in different cell lines, with or without digestion of surface glycans. Immunizations with a 30μg dose of the trimeric proteins were given at 0, 4, and 20 weeks, along with the ISCOMATRIX adjuvant (a particulate adjuvant comprising cholesterol, phospholipid, and saponin), with monomeric BG505 gp120 used as a comparator. Assays run by David Montefiore at Duke University demonstrated high titers of neutralizing antibodies in all immunized rabbits against the autologous BG505 virus, which is classified as tier 2 in the system that grades difficulty of neutralization (with tier 1 being the easiest to neutralize). By contrast, antibodies induced by monomeric BG505 gp120 showed significantly less activity. However, responses to heterologous tier 2 viruses with BG505 SOSIP.664 were far lower. Some neutralization was observed of a subtype C isolate, superior to that achieved with the monomeric gp120, but there was no effect against a subtype B virus. Cross-neutralization was only seen for subtype B viruses classified as tier 1B. For heterologous viruses, Moore said, “We haven’t gotten anywhere close to where we want to get.”

But it appears researchers are on the right track. A retrospective analysis by Montefiore suggests that the autologous neutralization titers seen with BG505 SOSIP.664 exceed those achieved by any prior gp140 immunogen, including a prior attempt at developing a stabilized trimer from Moore’s laboratory. This offered some encouragement that it represents a platform that can be built upon.

Efforts are now underway to map the epitopes being targeted by the neutralizing antibody response against BG505, and to identify means of improving the immunogenicity of BG505 SOSIP.664. Moore said one possibility for improvement is lessening the induction of antibodies against the V3 loop of the HIV envelope protein. Future plans also include developing multiple trimers based on viral sequences that appear to have invoked bNAb responses in HIV-infected individuals, and testing trimer cocktails derived from different HIV isolates from the same or different clades. Moore also cited the use of BG505 SOSIP.664 in structural studies as critical to better defining bNAb targets.

Ian Wilson, Hansen Professor of structural biology at The Scripps Research Institute, sounded the latter theme in a symposium session on visualizing HIV. Wilson covered the X-ray crystallography and cryo-electron microscopy (EM) work described in the last issue of IAVI Report, and published recently (Science 2013, doi: 10.1126/science.1245625, Science 2013, doi: 10.1126/science.1245627). Wilson stressed that SOSIP trimers are providing a new framework for studying bNAb interactions with the HIV envelope, for example, revealing how various glycan-dependent bNAbs can target the same site in different ways. These observations are cause for optimism because they suggest there are multiple opportunities for the immune system to successfully attack vulnerabilities in HIV’s armor.

Following Wilson, Walter Mothes, associate professor of microbial pathogenesis at Yale University School of Medicine, outlined his laboratory’s efforts to elucidate the dynamic behavior of the HIV envelope trimer, with a view to better understanding how bNAbs exert their inhibitory activity. The work is being conducted in collaboration with Scott Blanchard at Weill Cornell Medical College and Jason Gorman and Peter Kwong at the Vaccine Research Center at the National Institute of Allergy and Infectious Diseases. Mothes has employed a method called single molecule fluorescence resonance energy transfer (smFRET), which is capable of measuring microscopic movements over millisecond timescales. Molecules of interest are tagged with dyes and the distance between the tagged molecules can be measured based on computerized assessments of the transfer of energy between the dyes. For these studies, three HIV viruses were generated with one dye in the V1/V2 region and a second in the V4 and V5 loops of gp120. Two of the viruses were derived from neutralization-sensitive NL4-3 laboratory-adapted isolates and one represented a neutralization-resistant clinical isolate, JR-FL.

Analysis of the movement of the tags by smFRET showed that HIV envelope, unbound to any ligands, is dynamic and transitions between three distinct states based on the distance between the tagged variable loops. Mothes described these states as low, intermediate, and high, with low representing a “ground state.” Comparing the neutralization-sensitive and neutralization-resistant viruses revealed interesting differences: NL4-3 transitioned between states more frequently and spent more time in an intermediate form in which the envelope is more open, whereas JR-FL mostly remained in a ground state. “It opens up much less frequently,” Mothes explained. A similar reticence to remain in a potentially vulnerable conformational state was seen when JR-FL was studied interacting with ligands on target cells (CD4 and co-receptors).

Mothes next looked at how a panel of different bNAbs (including VRC01, PG16, and 2G12) affected the dynamics of the HIV envelope, finding that they all stabilized the ground state, and thus inhibited the transition to other conformational states required for the virus to successfully infect cells. A potent entry inhibitor drug in development by Bristol-Myers Squibb achieved the same effect. Mothes’ work offers a novel complement to techniques such as X-ray crystallography and electron cryomicroscopy (cryo-EM), which capture immobilized forms of the HIV envelope, by tracking the same molecules in flux.

Richard Jefferys is Coordinator, Michael Palm Basic Science, Vaccines & Prevention Project at the Treatment Action Group.