Skip to main content

A Toddler Stole the Show

The report of a child functionally cured of HIV stirred up the typically staid proceedings of the Conference on Retroviruses and Opportunistic Infections

By Richard Jefferys and Regina McEnery

The diminutive person who loomed largest at CROI was not in attendance.

An anonymous child from Mississippi, just 30 months old, drew headlines and became grist for multiple corridor confabs at the 20th Conference on Retroviruses and Opportunistic Infections (CROI) in Atlanta when it was reported—at a press conference before the science meeting opened, and then more fully during a session on HIV eradication—that the child had been functionally cured of HIV infection by early antiretroviral therapy (ART). The case was described by Deborah Persaud, a researcher and clinician-scientist at the Johns Hopkins Children’s Center who has decades of experience in pediatric HIV.

The story begins with an unnamed pregnant woman who received no prenatal care, and whose HIV infection was not diagnosed until she was in labor, leaving no time to administer AZT to reduce the risk of mother-to-child transmission. The rural hospital where the baby was delivered referred the infant to Hannah Gay, a clinician at the University of Mississippi Medical Center. Due to the high risk that HIV had been transmitted to the child, Gay decided to initiate a three-drug treatment regimen of AZT, 3TC and nevirapine, rather than the more standard dual combination of AZT and nevirapine that is typically given to children exposed to HIV.

That turned out to be a good call: a sample taken at 30 hours after birth tested positive for HIV DNA, while a second taken at 31 hours documented a viral load level of 19,812 copies. Therapy was then altered slightly after seven days, when the protease inhibitor lopinavir was substituted for nevirapine.

Over the course of the next several weeks, repeat viral-load testing revealed stepwise declines to 2,617 copies, 516 copies, 265 copies and then less than 48 copies, which is below the limit of detection of the assay they were using. About 18 months later, however, both mother and child were lost to care. When they returned, it transpired that therapy had been stopped. (Due to privacy concerns, Persaud declined to explain the exact circumstances.)

When Gay performed a new round of viral load measurements, expecting the typical rebound that has been described in many studies of pediatric ART interruption, she was surprised to find that HIV remained undetectable. Gay contacted Katherine Luzuriaga, professor in the division of Pediatric Immunology & Infectious Diseases at the University of Massachusetts Medical School, and an expert in the immunology of pediatric HIV infection. Luzuriaga in turn contacted Deborah Persaud, due to her experience in measuring latent HIV reservoirs in children. Persaud was the first to describe latent HIV in children (J. Clin. Invest. 105, 995, 2000).

Genetic studies confirmed the mother-child relationship and demonstrated that neither possessed known beneficial human leukocyte antigen alleles or protective CCR5 genotypes that could have made the child less susceptible or resistant to HIV.

Meanwhile, multiple laboratories searched for trace amounts of HIV in samples taken at 24 or 26 months of follow-up: Doug Richman at the University of California at San Diego employed digital droplet PCR to quantify HIV DNA in peripheral blood mononuclear cells. Tae-Wook Chun at the US National Institute of Allergy and Infectious Diseases (NIAID) looked for HIV RNA in plasma using a modified version of the Roche Amplicor test, which carries a limit of detection of two copies per ml of blood. Michael Piatak at the Frederick National Laboratory for Cancer Research—a branch of the National Cancer Institute—also assessed viral RNA in plasma using a single-copy assay, and Persaud herself looked for replication-competent HIV in resting CD4+ T cells using a co-culture approach.

Although the majority of these tests were negative, in two instances HIV proviral DNA was detected at levels slightly above the lower limit of the assays, suggesting that residual virus might have remained in the child’s blood or tissue. The single-copy assay for HIV RNA —so named for its detection limit of 0.3 copies per ml of plasma—also scored positive for one HIV RNA copy/ml at the 24-month time point. Importantly, Persaud was unable to detect any replication-competent HIV in a sample of 22 million resting CD4+ T cells.

To complement the virology, Luzuriaga undertook an array of immunological investigations. CD4+ T-cell counts and immune activation levels were comparable to age-matched HIV-uninfected children. No HIV-specific antibody responses were detectable by ELISA and Western Blot tests, and an analysis of CD4+ T-cell or CD8+ T-cell responses to Gag and Nef peptide pools also came up empty.

Persaud and her colleagues concluded that the case represents the “first well-documented case of ‘functional cure’ in an HIV-infected child.”

But, inevitably, after the data came questions and debate. David Margolis, director of the School of Medicine at the University of North Carolina and an HIV cure researcher, pointed out that the term functional cure was originally suggested to mean containment of HIV by virus-specific immune responses. Such responses have not yet been detected in the child.

“I would just like to point out that I don’t really like the term functional cure because I think it is very confusing for the public and for patients,” Margolis told Persaud during the question-and-answer period following her presentation.

Persaud noted that appropriate terminology for the pediatric population is not yet clear and, because follow-up off ART was relatively short—a little over 10 months—the researchers chose this term to try and convey the uncertainty regarding whether a complete cure had been achieved.

Another questioner asked whether the transfer of maternal cells containing HIV into the baby could explain the viral load readings. But Persaud had considered the possibility and calculated that an implausibly large transfer of around 250 ml of maternal blood across the placenta would be required to explain the amount of HIV RNA detected.

Additional questions posed at the CROI press conference included whether HIV might have been cleared even without ART, as was reported in a few rare published studies from the 1990s (N. Engl. J. Med. 332, 833, 1995; Lancet 347, 213, 1996). Persaud cited an analysis of viral genetic sequences from 1998, which concluded that sample contamination and mix-ups appeared to explain the majority of these findings, and that conclusive evidence for transient infection was lacking (Science 280, 1073, 1998).

A suggestion commonly heard in CROI hallway discussions was later articulated in a Wall Street Journal op-ed piece by Mark J. Seidner, a postdoctoral fellow in the division of infectious diseases at Massachusetts General Hospital and Harvard Medical School. Seidner, who works in Uganda, suggests ART did not cure HIV infection in the infant but acted as a form of post-exposure prophylaxis (PEP).

Yet, as Persaud explained at CROI, the detection of viral RNA and DNA within 48 hours after birth is widely accepted as evidence that infection occurred in utero. Further, studies involving viral load measurements in large numbers of infants receiving PEP have not documented any similar examples of progressive declines followed by an absence of evidence of infection (J. Pediatr. 160, 60, 2012).

To some extent, this represents a debate over exactly what constitutes established HIV infection. Persaud and colleagues hypothesize that early treatment prevented the formation of a latent HIV reservoir in the infant; for some scientists, this amounts to a kind of prophylaxis more than a cure.

In fact, researchers are probing this very question in HIV-infected adults. The largest study thus far assessing the impact of earlier treatment on the size of reservoirs is being conducted in Thailand by Jintanat Ananworanich (see Is It Ever Too Early?IAVI Report, Sep.-Oct. 2012). They have found that the earlier treatment is started, the smaller the reservoir size in blood and colon six or 12 months later. 

In terms of the broader implications of the case, Persaud described plans to conduct studies of prompt HIV treatment in infected infants through the International Maternal Pediatric Adolescent AIDS Clinical Trials group, in order to assess whether the results can be duplicated. In a separate poster presentation at the conference, Luzuriaga presented data on five HIV-infected youths (average age 16) who have been on ART since around two months of age and show no detectable levels of replication-competent virus. Luzuriaga suggested that such individuals are prime candidates for careful studies of treatment interruptions and interventions that seek to achieve a cure.

Amidst the debate and controversy, there was one very clear take-home message from the work of Persaud and colleagues: the pediatric population should be considered an essential part of the cure research agenda.

Searching for the sequel  

A late-breaker poster presentation from the newly formed European HIV Cure and Transplant Consortium (EHCTC) provided an update on attempts to repeat the cure of HIV infection obtained in Timothy Brown, widely known as the “Berlin patient” for undergoing two bone marrow transplants there to treat his leukemia. Results appear to be mixed.

Two HIV-infected individuals on antiretroviral therapy (ART) with different types of cancer were identified as candidates to receive cord blood-derived stem cells from donors homozygous for the CCR5Δ32 mutation. In one of them—a man from the Netherlands who was kept on ART throughout—the transplant was followed by an apparent rapid decay in HIV DNA levels. But he died two months later due to recurrence of the cancer and severe pneumonia.

In the second case, a man from Madrid, cord blood stem cells that were wild type for CCR5 had to be used due to concerns about the viability of cells from the CCR5Δ32 donor. In this case, a much slower decay of HIV RNA and DNA was seen, and the individual remains on ART but, happily, is in remission from the cancer. The EHCTC continues to work to identify additional candidates for CCR5-negative stem cell transplants. —RJ

No cure-all treatment

CROI 2013 also saw the public debut of data from the first multi-dose trial of a potential latency-reversing agent, the anticancer drug vorinostat (SAHA). Sharon Lewin, Professor and Director of Infectious Diseases at Alfred Hospital and Monash University in Melbourne, previewed the toxicity data at last year’s meeting (see A Slew of Science In Seattle, IAVI Report, Mar.-Apr. 2012). This time, Lewin described the effects of 14 days of vorinostat on HIV RNA expression in the latent reservoirs of 20 participants on long-term suppressive ART.

Echoing results from the single-dose study by Margolis published last year (Nature 487, 482, 2012), Lewin’s regimen induced a statistically significant 2.78-fold increase in cell-associated HIV RNA. However, Lewin found that there was no reduction of HIV reservoirs as measured by proviral DNA levels, consistent with studies from the laboratory of Robert Siliciano at Johns Hopkins University. Those studies found that induction of HIV RNA expression alone may be insufficient to cause the death of latently infected cells (Immunity 36, 491, 2012).

Siliciano himself delivered the opening plenary talk on some of the key challenges in cure research. In addition to the potential need for interventions, such as therapeutic vaccines to deliver the coup de grâce to latently-infected cells induced to express HIV RNA, Siliciano highlighted several other issues that continue to trouble researchers.

Among them is the question of whether laboratory cell line models of HIV latency accurately reflect what occurs in vivo. Siliciano noted that latency-reversing agents like vorinostat can show differential activity in different experimental models. To try and definitively assess the activity of candidate anti-latency approaches, Siliciano is now collaborating with Margolis to collect large numbers of latently infected resting CD4+ T cells from HIV-positive individuals via leukopheresis. Until now, the relative scarcity of these cells has precluded this type of analysis.

The accurate measurement of HIV reservoirs in HIV-infected individuals on ART is another difficult task. While results from various assays are commonly reported, exactly how they relate to each other has not been clear. Siliciano cited a recent comprehensive analysis that involved multiple laboratories testing samples from well-defined cohorts of people with suppressed viral loads. The study found no correlation between the gold standard viral outgrowth test for replication-competent HIV and most assays for viral DNA or RNA (PLoS. Pathog. 9, e1003174, 2013). Because viral outgrowth is time-consuming and cumbersome to measure, Siliciano’s laboratory is now working on a simplified version.

Siliciano also addressed a related question: What amount of reduction in the HIV reservoir might prevent viral load rebound after ART interruption, and for how long? His colleagues, led by Alison Hill from Harvard University, have developed a mathematical model that suggests at least a three log decline in the reservoir of latently infected cells would be required to prevent rebound for a period of several years, but late recrudescence of viremia might still be possible even with greater reductions.

Genetic protection

One proposed method of limiting HIV’s ability to replicate in the absence of ART is gene therapy. The idea is to genetically modify vulnerable cells in ways that prevent the virus from gaining entry or, short of that goal, replicating. The widely publicized case of Timothy Brown, the lone adult cured of HIV infection, has given impetus to this strategy: Brown received a transplantation of cells from a donor homozygous for the CCR5Δ32 mutation, abrogating expression of this key HIV co-receptor.

For Brown, however, stem cell transplantation was medically required to treat his cancer, and the cells he received came naturally modified. Genetically modifying all immune cells in otherwise healthy HIV-infected individuals, on the other hand, presents daunting challenges. On the plus side, Patrick Younan, a post-doctoral research fellow in the laboratory of Hans-Peter Kiem at Fred Hutchinson Cancer Research Center, described intriguing results from a small pig-tailed macaque study that hints that modification of all CD4+ T cells may not be necessary.

The experiment involved transplantation of stem cells modified with a gene encoding a peptide inhibitor of virus entry named C46—which works very much like the viral entry inhibitor drug Fuzeon—in tandem with a gene for green fluorescent protein (GFP) to allow identification of successfully altered cells. Two macaques received transplants with both C46 and GFP, while two controls were given cells containing only the GFP marker. All animals were subsequently challenged with the SIV/HIV hybrid SHIV 89.6P. Younan showed that, as is typical in studies using this virus, CD4+ T-cell counts plummeted during acute infection.

However, T-cell counts rebounded in the two recipients that received C46-modified cells. The viral loads in these animals dropped to levels significantly lower than controls—320-fold lower in one case and 1,477-fold lower in the other.

Analyses of the proportion of gene-modified CD4+ T cells to the total T-cell count prior to challenge demonstrated blood levels of 20% and 55%, respectively. During the acute phase of infection, levels rose to 92%, consistent with positive selection of virus-resistant CD4+ T cells. However, the percentage subsequently declined to pre-challenge levels in parallel with an increase in numbers of unmodified CD4+ T cells.

Younan and colleagues hypothesized that gene modification might be protecting SHIV-specific CD4+ T-cell responses and allowing them to mediate superior helper functions, leading to better immune control of viral replication. The result is improved protection of unmodified, susceptible cells. Supporting their suspicions, SHIV-specific CD4+ T cells were undetectable in controls, but responses to Env, Gag, Nef and Pol were present in the recipients of gene-modified cells. Further investigation revealed that approximately 85% of these SHIV-specific CD4+ T cells were descendants of modified cells. This implies that gene therapies may not need to protect the entire pool of susceptible cells in order to be effective. Additional experiments are now underway in hopes of confirming and extending the findings.

Novel evidence of the benefits of protecting virus-specific CD4+ T cells was presented by Adrienne Swanstrom, a graduate student at the University of Pennsylvania. Swanstrom outlined an unusual strategy to gain insight into the role of infection of CD4+ T cells in SIV (and by extension, HIV) pathogenesis.

Swanstrom and colleagues created an altered version of the highly pathogenic SIVmac239 that does not bind the CD4 molecule, while retaining the ability to interact with the co-receptor CCR5. This virus, designated iMac-ΔD385, was used to challenge two macaques. Peak viral loads were comparable to those typically observed with wild-type SIVmac239. But they were subsequently rapidly reduced and maintained below the limit of detection, between 30 and 50 copies/ml. No loss of CD4+ T cells occurred in blood, gut, or lymph nodes.

In assessing where infection was occurring in gut and lymph node samples, SIV p27 staining identified three populations: CD3+ T cells, CD68+ macrophages, and an as-yet unidentified CD3-, CD68- cell type. Localization of infection in lymph nodes showed a reverse pattern to that seen in pathogenic SIV, with few infected cells in the cortex and the bulk of infection occurring in the medulla.

Analysis of humoral immune responses revealed that the decline in viral load coincided with the generation of neutralizing antibodies. These antibodies, in vitro studies showed, had broad activity against iMac-ΔD385, SIVmac251 and the heterologous SIVsmE660. Such antibody responses are rarely seen in wild-type SIVmac239 infection.

The results appear consistent with the hypothesis of Swanstrom and colleagues that sparing CD4+ T cells from infection would improve adaptive immunity. Investigations in additional macaques are now planned in order to evaluate humoral and cellular immune responses in more detail, and study whether animals infected with iMac-ΔD385 can resist challenge with pathogenic SIV.

Scars of inflammation

The laboratory of Ashley Haase at the University of Minnesota has pioneered research into the association between HIV infection and fibrotic (scarring) damage to the lymph tissue (Semin. Immunol. 20, 181, 2008). At CROI, Joyce Sanchez, a research fellow in the laboratory, presented results from studies measuring lymph tissue fibrosis in HIV controllers.

Sanchez explained that fibrosis occurs in both lymph nodes and gut-associated lymphoid tissue (GALT) in HIV infection due to persistent immune activation and inflammation. The extent of the damage increases as disease progresses. Imaging analyses of the total area of lymph tissue samples occupied by collagen permits the quantification of such damage. Contrary to expectations, Sanchez found that HIV controllers displayed significantly increased levels of lymph tissue fibrosis compared to uninfected controls, regardless of whether viral load was undetectable or low (less than 2,000 copies/ml). There were also no statistically significant differences in fibrosis between the HIV controllers and HIV-infected groups with progressive disease either on or off ART.

Sanchez concluded that even the relatively low levels of viremia and immune activation seen in HIV controllers are sufficient to sustain fibrotic damage to lymphoid tissue. On a less pessimistic note, Sanchez highlighted recent results obtained from Timothy Brown, whose GALT displayed fibrosis that matched HIV-uninfected controls: the percentage of tissue occupied by collagen was 6.8%, compared to 7% in the uninfected group and 15.9% in HIV controllers. The implication, Sanchez noted, is that curing HIV infection can reverse the scars of the past.

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

...But HIV prevention got some play at the conference

Perhaps because the annual Keystone symposia on HIV Vaccines was held a month early this year—in February—CROI offered far fewer oral abstracts and posters related to vaccine science than in previous years. Even the plenaries, which set the tone for the conference and highlight big-picture science themes, ignored AIDS vaccines.

Still, CROI did offer up some jewels from the vaccine front, including fresh data from efforts to develop an AIDS vaccine based on a persistent replicating rhesus cytomegalovirus (rhCMV) viral vector containing several simian immunodeficiency virus (SIV) genes. Louis Picker, a professor of pathology at Oregon Health & Science University, who has been studying rhCMV vectors for nearly a decade, delivered fresh data in an expansive 30-minute talk that bolstered what his previous experiments suggest: some Indian rhesus macaques vaccinated with the rhCMV vectors, either alone or in combination with another viral vector vaccine, not only suppressed but cleared SIV following intravaginal challenge.

Picker conducted what he called an “acid test” for induced immunity—injecting naïve rhesus macaques with 60 million peripheral blood mononuclear cells taken from five SIV-infected animals who had been able to suppress the virus for at least 17 months post-challenge. When the cells from SIV-infected monkeys on fully suppressive highly active antiretroviral therapy or from monkeys able to control SIV without antiviral drugs were injected into SIV-uninfected animals, their transfer led to rapidly detectable infection. But in the case of the rhCMV vaccine-protected animals, no infection was detected.

“The implication is that there is no residual SIV in the rhCMV/SIV vector-vaccinated, long-term protected animals,” said Picker. “The SIV infection, which was demonstrable there early on, after challenge, is now gone, cleared, nada.”

These results build on previous work by Picker and colleagues that compared the immunogenicity of rhCMV vaccine candidates in a four-arm trial involving 61 macaques previously exposed to CMV. Twelve macaques were given the rhCMV/SIV viral vector-based vaccine, an equal number received an rhCMV/SIV vector-based candidate followed by a replication-defective adenovirus serotype 5 (Ad5) vector-based candidate encoding the full SIVmac239 genome, and nine received a DNA prime/Ad5 boost (encoding the full SIVmac239 genome) regimen. The 28 control animals received no vaccine. That study found that 12 of 24 macaques vaccinated with the replication-competent rhCMV viral vaccine candidate encoding the SIV proteins Env, Pol, Gag and Vpr/Vpx demonstrated early and complete control of viral replication for more than a year after repeat, homologous, low-dose SIVmac239 challenge (Nature 473, 523, 2011).

The study found no demonstrable differences in the number of challenges needed to infect the animals. And the remaining 12 macaques that received the CMV vaccine only and five of the 12 macaques that received the rhCMV/Ad5 vaccine combo were not able to control viral replication.

Exactly why some of the rhCMV-vaccinated macaques hit the biological equivalent of a grand slam, while others struck out, is not entirely clear. In previous studies, Picker and colleagues found that the persistent replicating vector induces effector memory T cells, which Picker suggests are better at protecting from virus in mucosal tissues than the central memory T cells induced by non-replicating vectors (see Nat. Med. 15, 293, 2009). Picker and colleagues also determined the mechanism that enabled CMV—an almost ubiquitous virus—to overcome pre-existing immune responses and superinfect monkeys already infected with the virus. They found that to superinfect its host, rhesus CMV needs genes that prevent major histocompatibility complex (MHC) class I of infected host cells from presenting CMV proteins to CD8+ T cells (Science 328, 102, 2010).

Picker thinks one of the reasons half of the rhCMV-vaccinated monkeys in his studies were unable to control SIV infection could be because they failed to generate enough effector memory T cells early enough. He said necropsies conducted in the protected animals showed responses “higher than what we see even in general vaccinees.” Picker’s lab is now working with the Nonhuman Primate Consortium funded by the US National Institute of Allergy and Infectious Diseases (NIAID) to look at the animals eight days after onset of infection. This is when you see responses in monkeys “going one way or another,” said Picker.

Picker said genetic polymorphisms in monkeys might also explain why roughly half the monkeys across five different trials have been able to suppress SIV while the other half have not. “But in some ways, it doesn’t really matter because in humans it might be 0% or 100%,” said Picker. “We’re not going to know that until we take [the CMV vector] into clinical trials.”

In any case, Picker said the CD8+ T-cell responses in the rhCMV-vaccinated animals appeared to violate all the rules. He noted that the responses are extraordinarily broad, targeting many different SIV epitopes found in all or most vaccinated animals—42 animals tested by Picker had a 100% response to the same epitope. The CD8+ T-cell responses also seem to rely on MHC class II molecules to recognize the virus—a trait more associated with CD4+ T cells—though Picker said the CD8 responses in his model are also capable of mediating SIV-infected cell recognition via class I molecules.

“Basically, with these vectors we can actually program T-cell responses that have effector memory differentiation that have entirely non-overlapping T-cell epitope recognition,” said Picker. These findings, he said, might provide proof of principle that the cellular immune system can mediate potent clinically relevant protection, including viral clearance, “by itself, without antibodies, against a highly pathogenic lentivirus.”

The Envelope, please

On the antibody side, University of Iowa researcher Hillel Haim presented findings from a recently published study that looked more closely at how the cleavage of the HIV Envelope gp160 protein affects its recognition by a wide range of antibodies, including broadly neutralizing antibodies (bNAbs) (J. Virol. 87, 1884, 2013).

Understanding this process will presumably enable researchers to design immunogens that more closely mimic the trimeric functional form of Env that appears on the viral surface, one that bNAbs must recognize to thwart infection.

Haim, who recently opened his own lab on the Iowa City campus, developed the premise for his research project while working as a post-doc in the lab of Joseph Sodroski, a professor of microbiology and immunology at the Dana-Farber Cancer Institute in Boston. Much of Sodroski’s work has focused on elucidating the structure and function of HIV Env in its various conformational states. Last year, he described an 11-Ångstrom structure of the unliganded HIV Env trimer using single-particle cryo-electron microscopy (Nat. Struct. Mol. Biol. 19, 893, 2012).

Because the Env spikes that adorn the surface of HIV are highly unstable and structurally dynamic, scientists, including Sodroski, have not been able to obtain a higher-resolution crystal structure of the unliganded Env trimer with methods more commonly used in structural biology. Knowledge of the atomic-scale structure of the Env spike could improve HIV drug design and remove much of the guesswork from selecting the best immunogens for vaccine candidates.

In the absence of a crystallized image of the unliganded trimer, Haim decided to examine the pattern of antibody binding to the Env trimer expressed on human osteosarcoma (HOS) cells, which allow clear differentiation between cleaved and uncleaved forms of Env. Haim and his colleagues used a cell-based ELISA system to measure binding of both cleavage-competent (wildtype) and uncleaved (mutant) Env against a panel of 16 bNAbs of varying breadth and potency. In a parallel test, they stabilized the HOS cells with the fixing compound glutaraldehyde to measure the effect of chemical fixation on the antigenicity of Env. “Fixation provides a snapshot of the different conformations that Env can assume at different times,” said Haim.

Haim said he and his colleagues found that both cleavage and chemical fixation were associated with an increase in binding by antibodies, such as VRC01, that target exposed epitopes. But they found cleavage and chemical fixation to be associated with a decrease in binding by antibodies like b13 and 17b that target the less visible (cryptic) epitopes overlapping the CD4-binding and co-receptor-binding sites, respectively. This would suggest, said Haim, that cleavage results in structural changes to Env that decrease flexibility and affect antibody recognition.

Based on his experimental results, Haim has developed a mathematical model to predict how well antibodies neutralize HIV. The model incorporates the efficiency of antibody binding, the level of Env reactivity (a measure of the sensitivity of the Env to the bound antibody) and the degree of structural change (perturbation) that occurs when an antibody binds to Env—which he calls the perturbation factor (PF).

The degree of perturbation caused by each antibody is strongly associated with how exposed the target epitope is on Env. Antibodies with a low PF, like VRC01, have target epitopes that are always visible on Env (i.e., more frequently sampled), while antibodies with a high PF, such as 4E10, have target epitopes that are masked and harder to find because they are less frequently sampled.

Haim’s model found that antibodies with a low PF neutralized both tier 1 and tier 2 viruses, compared to those with a high PF, which neutralized only tier 1 viruses. Haim said the model suggests that antibody recognition and the frequency at which the epitope targets of antibodies are sampled before binding both determine how potent an antibody will be.

In another talk, University of San Diego scientist Gabriel Wagner reported that the lack of broad and potent neutralizing responses to heterologous viruses might predispose HIV-infected individuals to superinfection. Wagner and colleagues measured the breadth and potency of neutralizing antibody responses before and after super-infection in a cohort of men who have sex with men (MSM) who were not receiving antiretroviral therapy. They then compared the responses to a matched cohort of individuals who had not been superinfected. The researchers used gene sequencing tools to estimate the frequency of superinfection in the MSM cohort.

They found that neutralizing antibody breadth and potency to a heterologous virus panel was significantly weaker three months after initial infection among those who would become superinfected. A year after infection, those who had become superinfected still had less neutralizing breadth and potency to the heterologous viruses.

Wagner said the study provides evidence that a broad and potent neutralizing antibody response to heterologous virus may be a correlate of protection against superinfection and “likely critical in the preventive vaccine effort,” though he also said increased viral genetic diversity through superinfection does not seem to accelerate the response.

Cynthia Derdeyn of the Emory Vaccine Center at Emory University also presented recently published findings that shed light on how the family of PGT128-like antibodies develop (see IAVI Report March 7 blog, Development of broadly neutralizing antibodies against HIV: It’s complicated). Previous studies suggest that the PGT128-like bNAbs isolated in HIV-infected individuals target a sugar residue connected with an amino acid at position 332, at the base of the so-called V3 loop of HIV’s Envelope protein.

But Derdeyn’s findings suggest the mere presence of a sugar residue at position 332 of HIV Env may not always suffice for the development of PGT128-like antibodies. Derdeyn said the same set of mutations can result in different outcomes, depending on when they appear. Not only are specific changes in Env likely to be important, she said, but neutralization breadth might also be programmed by sequential or simultaneous exposure to viral escape pathways, which result in a wave of antibody evolution.

PrEP update

Two years ago, there was something of a PrEP rally at the CROI meeting in Boston, with upbeat talk about the biomedical prevention strategy (the acronym derives from pre-exposure prophylaxis) dominating scientific discussion and hallway conversation (see A PrEP Rally, IAVI Report, Mar.-Apr. 2011).

This year, the news was a bit more sobering. Researchers wrestled with the latest spate of disappointing findings—this time from the Vaginal and Oral Interventions to Control the Epidemic (VOICE) trial of 5,029 high-risk women. The study was launched in 2009 from 15 sites in Uganda, South Africa, and Zimbabwe to evaluate the efficacy of daily topical and oral antiviral drugs in preventing HIV transmission among high-risk heterosexual women. The three products tested were tenofovir gel, oral tenofovir (TDF) and oral Truvada—a combination of tenofovir and emtricitabine.

Jeanne Marrazzo, a researcher from the University of Washington, which led the VOICE trial, said analysis of blood samples from 773 women—including 185 who acquired HIV—suggests adherence was low across all study groups. Less than a third of the blood samples from women assigned to either the Truvada or oral TDF arms, and less than a quarter of samples from women assigned to the gel arm had detectable levels of drug. As might be expected, none of the daily topical or oral regimens evaluated in the trial could be linked to a decline in HIV acquisition.

Less expected and more alarming was what the trial revealed about HIV incidence in participating cohorts: it came to 5.7% across all groups, and climbed to as much as 8.8% among unmarried women from South Africa who are under the age of 25, which is a remarkably high rate of infection. “The rate of acquisition did not differ product to product,” noted Marrazzo. “Adherence was low across the bar.”

Marrazzo said behavioral analyses will try to get at the root of the poor adherence. “Obviously, we were disappointed,” she said. Trial investigators also plan to examine drug levels in vaginal fluid to see if there is any relationship between the use of the products—particularly the tenofovir gel—and efficacy.

Yet Marrazzo’s findings were not exactly unanticipated. In the fall of 2011, one arm of the trial was stopped for futility after a data safety monitoring board (DSMB) determined that it would be impossible to show any difference between oral TDF and placebo in preventing HIV infection (see VAX Sep. 2011 Global News). Two months later, the DSMB reached the same conclusion on the topical administration of 1% tenofovir gel, a microbicide. The remaining arm of the trial tested daily administration of Truvada.

NIAID, the primary funder of the trial—which was earlier projected by the Microbicide Trials Network to cost US$100 million—noted that the DSMB found no safety concerns with oral TDF, which is currently used to treat HIV. The VOICE trial was sponsored by NIAID, the Microbicide Trials Network, Gilead Sciences (the manufacturer of tenofovir and Truvada), and CONRAD, a Virginia-based research institute developing contraceptive products and options to prevent HIV and other sexually transmitted infections.

The poor adherence mirrored findings from another trial, FEM-PrEP, which was also stopped early for futility (see Vaccine Briefs, IAVI Report, Mar.-Apr. 2011). The trial was conducted in Africa and led by Family Health International (FHI) 360 in North Carolina. Tim Mastro, who oversees research and programs in global health, population, and nutrition at FHI 360, remarked at CROI that the PrEP findings are “sobering” and reflect a “tremendous disconnect between perception and willingness to take product.”

Study investigators were unable to say why the women who enrolled in VOICE didn’t stick to their assigned regimens, but they hope a sub-analysis now underway will be able to shed light on some of these behavioral questions.

In a related PrEP talk, Chasity Andrews of the Aaron Diamond AIDS Research Center in New York reported results of a study of 16 Indian rhesus macaques that suggests an alternative to daily PrEP. The study evaluated the ability of GSK744, a long-lasting second generation drug, to protect monkeys from SHIV. GSK744 belongs to a family of drugs called strand transfer inhibitors, which block integrase, the protein HIV uses to insert its genetic material into chromosomes.

Clinical studies have suggested that a single intramuscular injection of GSK744 can be detected for up to 48 weeks, which makes it an attractive option for both HIV treatment and PrEP. Andrews said eight monkeys who were injected intramuscularly with a long-acting form of GSK744 twice over a four week period were protected against repeat intrarectal SHIV challenge. All eight control monkeys became infected after a median of two rectal challenges. Andrews said the results suggest that GSK744, when injected quarterly or monthly, might be an effective alternative to daily oral PrEP or topical PrEP.

Meanwhile, an animal study led by the US Centers for Disease Control and Prevention in Atlanta found, for the first time, that a tenofovir intravaginal ring protected all six female pigtailed rhesus macaques following repeat low dose vaginal challenge by SHIV 162p3. By comparison, 11 of 12 control animals became infected after a median of four exposures. The treated animals remained fully protected after 16 weekly challenges with the SHIV strain. The polyurethane reservoir ring containing 120 mg of TDF delivered, on average, 2.3 mg/day in vitro in macaques, resulting in high tissue concentrations of the medication.

These studies have yet to be replicated in human trials. But they may well foreshadow where PrEP research is heading in the near future: the development of products and strategies to solve the seemingly intractable problem of adherence. —Regina McEnery