A data-driven approach to sequencing HIV vaccine immunogens

For several years now, researchers have been engineering HIV vaccine immunogens that can stimulate the immune system to do something it rarely accomplishes on its own — generating antibodies that can effectively neutralize most, if not all, HIV variants in circulation. These so-called broadly neutralizing antibodies (bnAbs) are considered the linchpin of an effective vaccine but inducing them is a monumental task.

For starters, they develop only occasionally in people living with HIV, and only after years of exposure to an ever-evolving virus. And, when these bnAbs do appear, they have many unusual properties — they are extensively mutated, for example — which make them difficult to induce through vaccination.

As a result, scientists have had to develop new ways to train, or, more accurately, coax the immune system to generate bnAbs. This requires a stepwise process. The first step is to engineer vaccine immunogens capable of priming the immune system. The goal of these priming immunogens is to engage specific subsets of naïve B cells that have the capacity to make bnAbs.

Results from early-stage discovery medicine trials show that several vaccine immunogens in development successfully initiate this process of bnAb development. Now, researchers must design subsequent immunogens to boost these responses and coax these B cells along the maturation pathway.

This work is being guided by detailed molecular, functional, and structural analyses of B-cell and antibody responses induced by the priming immunogens. These analyses were the subject of the workshop “B Cell Immune Repertoire Analysis for HIV Vaccines,” held by the U.S. National Institute of Allergy and Infectious Diseases’ (NIAID) Division of AIDS on August 13, 2024. At the virtual meeting, HIV vaccine experts reviewed clinical data for some of the most advanced immunogens. They also described the complex and labor-intensive methods used to characterize the immune responses these immunogens induce and brainstormed ways to make these processes more efficient and cost-effective, all while maximizing the amount of data collected. Fine-tuning these processes will become even more important as sequential vaccination regimens advance into iterative trials using platforms such as mRNA.

Progress in design and analysis

In recent years researchers have made substantial progress in designing vaccine immunogens that can prime naïve B-cell precursors of HIV-specific bnAbs and initiate the stepwise process of maturing these responses.

Several researchers at the workshop presented their progress in this area, including Leo Stamatatos, a professor at the Fred Hutchinson Cancer Research Center, who shared data from HVTN 301, an early-stage clinical trial of the vaccine immunogen known as the 426c.Mod.Core nanoparticle in combination with the adjuvant 3M-052/Alum. Analyses of data from this trial showed that the vaccine immunogen can activate and expand naïve B-cells that can potentially give rise to bnAbs that target the CD4-binding site on HIV’s outer Envelope protein that is used to dock onto and infect its target cells.

Bill Schief, vice president of antigen design and selection at Moderna who also holds positions at Scripps Research and IAVI, reported on a trio of trials — IAVI G001, G002, and G003 — testing another vaccine immunogen, known as the eOD-GT8 60-mer protein. This antigen also successfully primes naïve B cells that are specific to the CD4-binding site. All but one volunteer in the G001 trial generated the desired B-cell responses, and the frequency of desired responses was even higher among participants in the G002 and G003 trials, according to Schief, in which the vaccine immunogen was administered using Moderna’s mRNA platform.

Rogier Sanders, professor at the Amsterdam University Medical Center (UMC), and colleagues also had success in the IAVI C101 trial priming B-cell precursors using an immunogen known as BG505 SOSIP GT1.1, which is designed to mimic the trimeric structure of the native HIV Envelope protein. Three doses of this BG505 immunogen even seemed to induce B-cells that accumulated several of the mutations that would be necessary for them to make bnAbs.

Barton Haynes of Duke University, who leads a Consortium for HIV/AIDS Vaccine Development (CHAVD) program funded by the NIH, presented data from the HVTN 133 trial of yet another bnAb-focused vaccine immunogen, this one a peptide liposome immunogen that targets the membrane-proximal external region (MPER) of HIV Envelope, not the CD4-binding site.

The HVTN 133 trial was stopped prematurely due to a case of anaphylaxis in a study volunteer, however, Haynes and colleagues still evaluated thousands of antibodies generated by vaccine recipients. At the meeting, Haynes reported that the MPER peptide liposome induced heterologous antibody responses in serum, the most potent of which was able to neutralize 15% of harder-to-neutralize tier-2 global HIV strains.

This cluster of clinical data epitomizes the substantial progress in designing HIV vaccine candidates to induce bnAbs. Simultaneously, researchers are making major advances in the behind-the-scenes work of evaluating how these immunogens work in people by fully characterizing the types of immune responses they induce.

At the workshop, researchers described several of the highly technical, labor-intensive approaches computational scientists at the Vaccine Research Center (VRC) at NIAID, the HVTN, Scripps, Duke, Vanderbilt, and UMC Amsterdam are using to sort and quantify vaccine-specific B-cell responses, sequence B-cell receptor genes, and synthesize and screen monoclonal antibodies. Several groups are now applying next-generation sequencing methods and other newly developed technologies with alluring names such as TRAP and RATP-Ig (pronounced “rat pig”). The TRAP assay allows researchers to differentiate and analyze responses from plasma and memory B cells, which is useful when determining the durability of vaccine responses. RATP-Ig is a new high-throughput method for isolating immunoglobulin proteins from serum and generating monoclonal antibodies from individual B cells. 

One of the biggest challenges for all these analyses is selecting and developing high-quality probes and reagents that can capture antigen-specific B cells from either plasma or mucosal samples. The better the probe, the more effective the analysis. “Most of our problems are solved right there,” said Madhu Prabhakaran, a senior biologist at the VRC.

This is also one of the major bottlenecks in the process as the probes are vaccine and study specific. “I can’t over-emphasize how laborious the process is,” said Rachel Parks, a staff scientist at the HVTN.

Another challenge is increasing efficiency, which will become critical as samples are collected from more and more volunteers in future trials. “As we increase scale, you run into problems,” said Bryan Briney, an associate professor at Scripps Research. Some of the more labor-intensive methods can take weeks or months to generate results from a single sample, which isn’t optimal now and certainly won’t be as the field conducts more and more iterative trials. As computational scientists advance some of the next-generation approaches the goal is to make this process more efficient and cost effective. 

Another issue the field is contending with is understanding the genetic variation in human immunoglobulin genes on a global scale, and the effect of this variation on vaccine responses. The G001 trial provides a pertinent illustration of this.

Data published from G001 indicate the eOD-GT8 60-mer protein immunogen successfully primed naïve B-cell responses in all but one volunteer. Subsequent analyses explained this outlier. Genotypic studies showed that this individual lacked the specific alleles in their antibody genes that would allow them to generate specific CD4-binding site classes of bnAbs. “The vaccine needs to work in almost everyone you immunize,” said Schief, so understanding the genetic diversity in antibody genes on a global scale is a priority for the field.

Genetic variation in human antibody genes is quite extensive but poorly documented. The largest database of variation in antibody genes is the International Immunogenetics Information System (IMGT) out of the University of Montpellier, France. It was established in 1989, however, half of the alleles in this database have not been independently confirmed and much of the data is based on a limited number of studies involving individuals primarily of European descent.

To address these shortcomings, the Adaptive Immune Receptor Repertoire (AIRR) community of The Antibody Society is creating an additional reference set, the AIRR-C. Scientists are adding many new alleles of importance to HIV vaccine researchers to this database from genomic studies of populations worldwide. And they are finding that as more genomic data is collected, many more novel alleles are identified.

The goal is to use this information, along with the detailed analyses of B-cell responses, to interpret results from clinical trials of HIV vaccine candidates and inform the design of future immunogens in sequential immunization strategies. Everyone at the workshop agreed that determining the ideal sequence of immunogens to induce a broad and potent bnAb response against the virus remains one of the most critical questions facing HIV vaccine researchers today.