Bethesda Marriott Hotel, Bethesda, MD
The AIDS Vaccine Research Subcommittee (AVRS) met in public session on May 19-20, 2009, in the Natcher Conference Center on the campus of the National Institutes of Health (NIH) in Bethesda, MD.
AVRS members present: Eric Hunter (chair), James Bradac (executive secretary), Jay Berzofsky (ex officio), Deborah Birx, Larry Corey (ex officio), Kevin Fisher, Nancy Haigwood, Paul Johnson, Jeffrey Lifson, Bonnie Mathieson (ex officio), Juliana McElrath, Nelson Michael (ex officio), Gary Nabel (ex officio), Douglas Nixon, Michel Nussenzweig, Louis Picker, Jerald Sadoff, George Shaw.
Other NIH personnel participating:
Dr. Hunter called the meeting to order and asked the committee members and observers to introduce themselves.
Julie McElrath identified the two major scientific questions emerging from the STEP trial, namely (1) What are the reasons for the lack of vaccine efficacy, and (2) What biological mechanisms explain the increased HIV-1 acquisition in the vaccine group? The HVTN is accepting proposals for access to clinical specimens from the STEP trial for ancillary studies. To date, 30 proposals have been submitted and 17 approved, with 5 more pending. In several cases experiments have already been conducted and data are under analysis.
Two independent factors have been identified that significantly correlate with increased rate of HIV-1 infection: circumcision (or rather lack thereof) and seropositivity to herpes simplex virus type 2 (HSV-2). Ad5 immunity is not correlated with risk of infection. Further analysis of vaccine-induced T cell responses failed to confirm the previously observed association of increased breadth of Gag-specific responses with set point viral load (sVL). Neither the total number of T cell responses nor the breadth of Nef and Pol response are associated with sVL.
In response to questions, McElrath agreed that we have some idea of what cellular response we want, but not how to induce it. The key may lie in the expression of cytokines (particularly perforin) by T cells as they proliferate in response to HIV exposure. Investigators are looking for correlates of perforin expression, which may have to do with vector and/or prime-boost regimen.
Morgane Rolland reviewed the data emerging from studies of viral genetics during early infection. By amplifying and sequencing virus from successive samples, investigators were able to trace the evolution of the viral population. In 75 percent of infected vaccinees, the original infection was due to a single variant, the same pattern found in infected placebo recipients, and there was no phylogenetic clustering that would indicate a “sieve effect” blocking infection by specific variants of HIV, or because of a “selection effect” that helps to drive virus mutation following infection. However, when data were analyzed based on potential CTL epitope sequences rather than full-length gene sequences, signs of CTL-mediated selection in infected vaccines was detected (expanded on by Peter Gilbert). That is, the virus that infected vaccinees was more likely to be different from the STEP vaccine than the virus infecting placebo recipients, indicating a possible sieve effect.
Peter Gilbert presented data on a “global sieve analysis” that measures the mismatch between vaccine and infection epitopes at the 8- and 11-peptide levels. Further “local sieve analysis” at the level of single amino acid sites produced a set of “signature sites” that represent mismatches between vaccine and placebo viral sequences. Results showed that vaccine does exert a sieve effect on Gag and especially Nef (but not Pol), but it does so primarily before seroconversion. The strongest signature was at site Gag 84, which is restricted by several A-list alleles. However, this selection effect did not lead to a vaccine effect on early post-infection markers of disease progression. Nevertheless, this may provide guidance in the development of improved T cell-based vaccines. Further analyses are currently underway to refine and extend these findings.
In response to questions, Gilbert added that these early vaccine effects are washed out in later analysis due to the magnitude of the natural response. Few data are available on continued sexual exposure and other behavioral factors. Epitope mapping is a relatively fragile tool, but ongoing analyses should be able to find more robust patterns. It could be useful to compare these results with finding from acute infection; the data are too new to have done all possible analyses.
Alan Bernstein reported on steps being taken to update Global HIV Vaccine Enterprise’s Scientific Strategic Plan (2005), as well as increases in the available resources and partners. The planning committee has identified two central questions – what is required to mount a protective immune response to HIV, and how can we use that knowledge to design and test new vaccine candidates? – and four broad scientific priorities:
Other themes include identifying roadblocks and opportunities, making organization and funding changes to integrate basic, preclinical and clinical research, and attracting and sustaining new investigators. Publication of the new Scientific Strategic Plan is scheduled for January 2010.
1. The problem of eliciting broadly neutralizing antibodies (bNAbs) against HIV
Gary Nabel noted that the fight against polio required researchers to develop a vaccine that could recognize three serotypes in order to convey protection. HIV, which possesses a potentially infinite number of serotypes, requires a different approach. A wide range of neutralizing antibodies are generated by circulating memory B cells during natural HIV-1 infection, and between 20 and 25 percent of infected sera display some cross-clade breadth of neutralization. Investigators have used structure-based design to create immunogens based on alternative forms of Env gp120 outer domain, cloaks and trimers. Vaccination with one such immunogen induced CD4 binding site (CD4bs)-specific Abs in the rabbit model, but those Abs weren’t neutralizing. Other immunogens are more useful for diagnostics, and many of them don’t seem to do much at all. Over the next few years, investigators will optimize these immunogens and validate them in nonhuman primates and humanized mice. Further understanding of B cell biology will facilitate progress in this field, which could lead to rationally designed immunogens that elicit broadly neutralizing Abs. Phase 1 clinical trials might be expected in early to mid 2011.
2. Induction of anti-HIV bNAbs
Dimitri Dmitrov noted that bNAbs to HIV-1 are highly divergent from germline Abs, and they are typically elicited later than isolate-specific Abs because it takes a long time to mature to a specific, highly mutated Ab. Indeed, it appears that HIV is so “smart” that it exploits gaps in the germline repertoire. However, it may be possible to speed up this process by using two or more “intermediate” immunogens to guide the immune system through this complex maturation pathway, producing not just a faster response but an exponential growth in Ab expression.
Michael Cancro suggested that bNAbs are not easily made because such clonotypes are rarely found in, or actively excluded from, primary or memory B cell repertoires. This raises the question of which processes dictate the size and composition of pre-immune B cell pools, and whether the factors that control these processes can be manipulated. The primary fate of mature B cells is determined by B cell receptor (BCR) signaling and by cytokines such as B lymphocyte stimulant (BLyS) and B cell activating factor (BAFF), which induce B cell proliferation and immunoglobulin secretion. Homeostasis and survival are closely integrated: take BLyS away and all transitional B cells die; provide enough BLyS and they all survive. Under ordinary circumstances, between 90 and 97 percent of transitional B cells are deleted before they can mature. However, mice treated with BLyS produce larger pools of pre-immune B cells and gp120-specific Abs. This suggests that a similar relaxation of the deletion process in humans might also provide a window of opportunity to shuttle useful clonotypes into the primary B cell pools.
In response to questions, Cancro explained that the mouse experiments are a work in progress – investigators have not moved beyond gp120 and thus don’t know how broad the response is. In addition, it’s one thing to use BLyS to generate B cells but quite another to use it as a new approach to immunization. One phenomenon they have noticed is a disproportionate increase in the number of marginal zone B cells, but they believe that the BLyS treatment increases the overall diversity of the pre-immune B cell population, since so many more immature cells are getting through.
Michel Nussenzweig explained that B cells are edited twice to eliminate self-reactivity, once as immature cells in bone marrow and again as circulating cells in the periphery. B cells that produce NAbs are not that uncommon, but they do develop slowly, and most bNAbs are singular events that cannot be reproduced by vaccination. By looking at populations of B cells from HIV-infected patients, it is possible to make the following observations:
In response to questions, Nussenzweig added that Patient 1, despite being a LTNP who has a high number of bNAbs and a high proportion of CD4bs-specific Abs, may represent an individual infected with a defective virus rather than one exhibiting a desirable immune response.
Abe Pinter reported on the isolation and characterization of a new class of NAbs that target quaternary neutralization epitopes (QNEs), whose activity is dependent on the juxtaposition of multiple domains on adjacent subunits of the Env trimer. The first HIV-1 QNE was defined by Ab2909, which has unprecedented neutralizing potency for SF162 Env, binding to intact virions rather than soluble Env proteins. Reactivity requires the presence of both V2 and V3 domains, but type-specificity maps to the V2 domain; Ab2909 is more or less sensitive and potent depending on mutations in these loops, but it is more than 4000 times more potent against the JR-FL mutant than standard anti-V3 mAbs. Similar QNEs are readily isolated in NHPs infected with SHIV, and investigators are accumulating data from human sera. Recent data suggest that some human QNEs are conserved and immunogenic; for example, one human serum (CAP-256) possesses very broad and potent neutralizing activity against a wide range of clade C Env proteins, including gp160 and the V1/V2 region of gp120. While some QNEs may be highly immunoreactive, however, they are not necessarily highly immunogenic; more information is needed about the structures to guide vaccine design, as well as an experimental system for testing viral constructs. One possible vaccine approach would be to prime with vector and boost with anti-QNE bNAbs.
Pascal Poignard described IAVI’s Protocol G, which screened over 2,000 patients who remained asymptomatic 3 years after infection to isolate bNAbs from “elite neutralizers” (about 1 percent of donors) whose sera demonstrate potent neutralization across clades. Successive assays led to the characterization of two new Abs (PG9 and PG16) that show potent neutralization across a 16-virus, cross-clade panel. Both Abs preferentially recognize a QNE on the Env trimer involving highly conserved residues in V1/V2 and V3 (reminiscent of Ab2909), and both are inhibited by soluble CD4, suggesting that they target a CD4-sensitive conformational epitope that might be a HIV-1 vaccine target. Further experiments are currently underway to complete epitope mapping, measure serum concentrations and neutralization in PBMCs, and begin Ab production for structural studies.
In response to questions, Poignard added that about 34 percent of Protocol G donors had fairly broad neutralization, but only 1 percent had potent neutralization; because PG9 and PG16 were isolated from a single patient, it is unknown how common they are. The rate of Ab recovery is very low, and there are many technical and methodological problems to be solved.
3. Optimal approaches for the induction of Abs
Robert Carter explained that Ab production follows different pathways in the short and long term. The marginal zone of the spleen contains specialized populations of B lymphocytes and macrophages that can be rapidly activated within hours of antigen presentation. However, defects in B cell receptors (notably the CD19 complex) interfere with the cell’s ability to capture antigen and to send and receive signals from T helper cells, macrophages, follicular dendritic cells, and antibody-producing B cells. Researchers have learned that chronic activation (as in HIV infection) also interferes with these functions. Antibody production by newly formed B cells is much slower, because it must wait for somatic changes in the bone marrow.
4. How infection impacts the B cell response (1) to HIV itself and (2) to vaccination
Iñaki Sanz described techniques for identifying and isolating B cell subsets that may have opposing functions in response to HIV infection. Memory B cells are long-lived lymphocytes that are very plastic and can be specific to a wide range of antigens; effector B cells are short-lived plasmocytes that produce large volumes of Abs when activated by memory cells. In HIV infection, however, BLyS and BAFF inhibition disrupt the maturation of naïve B cells but not the activation of B effector cells, suggesting that B effector cells com from extrafollicular reactions, not from the germinal center. Analysis of serum from HIV026, one of the 1 percent “elite neutralizers,” revealed a high proportion of plasma B effector cells. In a broader screen for cross-clade polyreactive anti-Env NAbs, however, only one cell in 1,000 matched HIV026 for Env. The next step is to isolate bNAbs involved and assay them for specificity. Researchers have not yet tested virus-like particles (VLPs) as a possible vehicle for antigen, in part because of size limitations on the molecule.
Barton Haynes reported that Ab response early HIV infection doesn’t show up until day 13, in the form of Abs to gp41, and the first NAbs don’t appear until day 77. The reason for this delay is unclear. Analysis of sera from three early patients showed a sudden peak in plasma (effector) B cells at day 7, many of them producing Abs for unknown (non-HIV) proteins, followed by a decline in the number of naïve B cells and the loss of 50 percent of germinal centers in the first 80 days of infection. If the earliest response to HIV infection is a massive but untargeted B cell activation, then it seems clear that B memory cells for a protective response must be present before infection. Further research will focus on B cell germlines, Ab evolution, and the cytokine repertoire of marginal zone B cells.
In response to questions, Haynes added that a similar, dysfunctional B cell response has been observed in SHIV-infected macaques; it would be desirable to have more detailed observations for days 1, 2 and 3 in NHPs. It would also be useful to study T-cell proliferation in response to gp41, which is by far the immunodominant response in early infection. If autologous response occurs at 6 weeks, how would it be possible to move it back to 2 or 3 weeks? Acute viral load may play a role in the disruption of B cell response, and anything that reduces that viral load might help to rescue that response. Treatment reduces acute viremia but also reduces Ab response.
George Shaw described the single genome analysis (SGA) of Env protein in three CHAVI subjects from day zero to one year after infection. SGA permits investigators to infer the founder or transmitted virus lineage and to trace viral escape through subsequent mutations in response to late-appearing NAbs. In all three cases, the antigenic profile of the transmitted virus matched the YU2 prototype, which replicates successfully in CD4+ T cells but not in macrophages. By following the patterns of mutation as the virus escapes neutralization, investigators discovered that the earliest NAb response in one subject targeted the conserved V1 and V3 domains of Env, a location that suggest a conformational epitope or QNE rather than a linear epitope. In another subject the NAb target was in the V1/V2 and C2 regions of Env, An important positive finding was that NAb titers in the range of 1:100 to 1:1000 are able to select for complete virus escape, indicating that in vivo, Abs are able to neutralize virus whether it is transmitted “cell-free” or cell-to-cell – which portends well for vaccine-elicited NAbs, if they can be elicited.
In response to questions, Shaw said that the site of transmission might influence its preference for CD4+ T cells rather than macrophage, but the transmitted virus appears to be depended on a CD4+, CCR5+ environment; he has no idea whether it would replicated in a dendritic cell. Early autologous Abs are largely monospecific and do not work cooperatively.
The meeting adjourned at 5:30 p.m. and reconvened the following morning at 8:30 a.m.
Stuart Shapiro presented an update on the Vaccine Discovery Branch, whose mission is to work with other branches, divisions and programs to address the basic biology of the virus, basic immunology, and the interactions of virus and host, as well as to develop new concepts that could lead to bNAbs, high levels of functionally optimal CTLs, and protective mucosal responses. New programs in FY2009 are HIT-IT and Basic HIV Vaccine Discovery; new programs for FY2010 will include the B Cell Biology Network, a systems biology approach to HIV immune response, and Mechanisms and Prevention of Sexual Transmission of HIV/SHIV. CHAVI is now producing 50 papers per year, and basic research should benefit from the diversion of funds away from large vaccine trials.
In response to questions, Shapiro explained that both the B Cell Network and the systems biology program would be a “virtual center” made up of cooperating centers with common interests. HIT-IT is well positioned to produce preliminary data that would allow investigators to then apply for vaccine research funds from the Gates Foundation. Subcommittee members asked that DAIDS provide a pie chart of money distributed through these and other programs.
James Bradac updated the subcommittee on new awards made in FY2009 through the Phased Innovation Award program, the HIV Vaccine Research and Design program, and the Integrated Preclinical/Clinical AIDS Vaccine Development program. In addition, five new DAIDS-sponsored vaccine candidates are currently in the pipeline for Phase 1 trials by 2011. Recent research breakthroughs include promising results from a recently published NHP study of a RhCMV-based vaccine that provided good protection against low-dose rectal challenge. Upcoming workshops include an IAVI/NIAID sponsored meeting on 21 May addressing vector-mediated Ab genes and a Bolger conference 15-16 Jun on cellular mucosal immunity.
Alan Fix reported that many studies were halted or altered in response to STEP, and no new trials were launched in FY2008. This makes it hard to maintain sites and networks that have little or no activity. However, several trials will begin enrollment in late FY2009 or early FY2010, including studies of mucosal, heterologous vectors, and several ancillary studies. Young and Early Clinical Investigator (YECI) Scholar Awards have already gone to five promising researchers, and the new Rapid HIV Point-of-Care Diagnostics award will stimulate the development of FDA-approved, low cost diagnostics for resource limited settings.
In their discussion of the B cell presentations and other topics, AVRS members were asked to address six general questions:
These questions informed but did not necessarily guide the discussion that followed. Participants noted that the Haynes findings with regard to early targeting of gp41 are another example of the virus’ cunning mechanisms – CTL responses against Env would eliminate the target for gp120-specific NAbs. However, breadth is the composite of many Abs, and the search for a single, perfect bNAb is futile. The same is true of perfect epitopes – we need to be thinking in terms of polyclonal, combinational and conformational epitopes.
Several participants expressed disappointment that the natural repertoire of B cells doesn’t include the necessary range of Abs and responses, and that it takes so long to develop a full repertoire. This means that research must also address how to make the best use of a partial or imperfect repertoire, at least in the short term, and thereby give the immune system time to augment it. Participants felt sure that we still have much to learn from LTNPs, who mount a defense consisting of multiple Abs that neutralize virus as well as isolates – a strong argument for the combination approach. Why do 25 percent of patients make the responses we want them all patients to make, and how do we get everyone to make responses as potent at the “elite neutralizers”? What is the optimum combination and intervals, and should immunization be passive or active?
If the solution to any of these questions involves a heterologous insert, it would not be a logistical problem to bring subjects back for a “booster” a year or even two years after their initial vaccination. The problem is that the field doesn’t have a candidate to bring forward, and the field is unenthusiastic about any candidate that doesn’t elicit NAbs. In addition, it will take 12 months just to develop the production technology for a stable trimer.
At present it takes the immune system about 12 months to mount a meaningful defense against HIV, during which time the vaccine is not protective and may actually increase the risk of infection. The combination of prime and boost seems to make a big difference, but we don’t know why. This may be a question for systems biology, but it’s certainly a scientific question worth of research: what are the determinants and mechanisms of the prime-boost phenomenon, particularly with regard to its timing? Such a question seemed too boring in the past but has a new importance, as do questions about somatic maturation in B cells.
NHP studies can usefully define the number and quality of neutralization epitopes, but these results may not be directly applicable to humans. Nor is it clear whether we should follow up on all epitopes or stick with the most promising, especially since some of them seem to lead in opposite directions. A compromise would be to do both – pursue epitopes broadly in order to narrow the field, and then combine the best candidates in a mosaic vaccine. Work is just beginning on such a mosaic and won’t be mature for another two years. The logical endpoint would be a comparative trial of (1) different inserts in the same vector and (2) the same vector with different inserts – an important trial that would also be impossibly large and expensive. We need a way to do it cheaper and faster, but for now we need to collect and understand more information about different vectors and inserts.
AVRS heard exciting presentations on systems biology at its last meeting, and there are good groups at work in this field, but it’s not clear that those groups will have access to the best samples available from NHP and human trials. It’s not clear that the field needs a coordinating committee or standardized protocol to collect additional samples, but at present there is no systematic comparison of the existing sample sets. It would be useful to have a firm recommendation that every efficacy trial collect sequential, weekly samples of serum (including Abs and virus) from all subjects. The Gates Foundation expressed willingness to convene a meeting on this topic and asked participants to suggest the names of experts who should be invited. In addition, GHVE will hold a meeting on structural biology in spring 2010 and would welcome suggestions on speakers and topics, including what samples should be collected. Dr. Hunter will try to pull together the subcommittee’s disparate opinions on these questions.
The meeting adjourned at 12 noon.
The next meeting of AVRS will be on September 15-16, 2009, in Bethesda MD.
back to top
Last Updated June 27, 2011