Research Briefs
By Andreas von Bubnoff
Dozens of New Broadly Neutralizing Antibodies Identified
In the last two years, many potent, HIV-specific broadly neutralizing antibodies (bNAbs) have been isolated from chronically HIV-infected individuals. Now, a study led by Michel Nussenzweig, a professor of molecular immunology and a Howard Hughes Medical Institute investigator at Rockefeller University, adds dozens of new CD4 binding site specific bNAbs (isolated from four chronically HIV-infected individuals who have high levels of bNAbs to HIV in their serum) to the collection (1).
The new bNAbs recognize the CD4 binding site on Env that is also the target of other known bNAbs including VRC01, which was isolated last year by researchers at the Vaccine Research Center (VRC) and is one of the most potent bNAbs identified so far. According to Nussenzweig, some of the best of the newly isolated bNAbs are broader and more potent than VRC01, while others are equivalent.
Nussenzweig and colleagues isolated single memory B cells from the blood of the HIV-infected individuals using protein baits that consist of parts of the HIV Env protein. They then used polymerase chain reaction (PCR) to amplify the variable regions of the light and heavy chains of the antibody genes that are expressed in these memory B cells to make monoclonal antibodies (mAbs).
The study is a follow up to a 2009 study in which Nussenzweig and colleagues used a similar approach to isolate HIV-specific mAbs from six HIV-infected individuals with high titers of broadly neutralizing sera against different HIV strains (2). In the 2009 study, recombining some of the HIV-specific mAbs they isolated could reconstitute the broad neutralizing activity of the serum in two of the individuals. Still, the neutralizing activity of any single antibody alone was less broad than the sera, suggesting that mixtures of antibody were responsible for the neutralizing activity in those two patients. They did not isolate single bNAbs in the 2009 study.
Nussenzweig says they were likely able to isolate single bNAbs in the most recent study because they took into account that HIV-specific bNAbs typically show a high degree of somatic hypermutation that makes them different from the germline genes they are derived from. Nussenzweig and colleagues used different primers that were designed to PCR amplify even highly mutated antibody genes from B cells by binding to less mutated parts of the antibody sequences.
When Nussenzweig and colleagues tested the same gp140 bait they used in their 2009 study together with the new primers, they could isolate VRC01 and VRC01-like bNAbs from the same individual from whom VRC01 was isolated last year, as well as many CD4 binding site specific bNAbs from two of the same individuals from whom they did not isolate bNAbs in their 2009 study. This suggests that the primers, and not the bait, account for the difference in the success in isolating bNAbs between their 2009 and their 2011 studies, Nussenzweig says. However, they did find that using the new primers in these two individuals with a bait designed to more specifically identify CD4 binding site specific antibodies resulted in isolation of an even larger fraction of CD4 binding site specific bNAbs.
Peter Kwong, chief of the structural biology section at the VRC, who was not involved in Nussenzweig’s study, says “it’s really exciting that they have many new broadly neutralizing HIV antibodies.” And because the antibodies were amplified from single cells, they represent true pairs of heavy and light chains of the antibodies. “I think that that is a big strength of this paper,” says Kwong.
For the first time, Nussenzweig and colleagues also showed that some of the newly identified CD4 binding site specific bNAbs are expressed in the HIV-infected individuals’ antibody-producing plasma cells, and that some of the corresponding secreted antibodies were in their serum.
“I think this study represents a landmark in the field,” says Davide Corti, director of the antibody discovery unit at the company Humabs BioMed, who was involved in isolating another CD4 binding site specific bNAb, known as HJ16, last year (see Research Briefs, IAVI Report, Jan.-Feb. 2010). Corti, who was not involved in the isolation of the latest bNAbs, says the study shows that at least in chronically infected people with high levels of bNAbs in their serum, which make up only a few percent of HIV-infected people, broad and potent antibodies are “not a gold nugget.”
Like other recently described HIV-specific bNAbs, the most potent new bNAbs Nussenzweig and colleagues identified show an unusually high degree of somatic hypermutation. But, to their surprise, they found that the variable portions of the heavy chain of the most potent new bNAbs and VRC01 have about 68% of their amino acid sequence in common and also share a common origin in that they are derived from two related germline genes, IgVH1-2 or IgVH1-46.
The most potent new bNAbs are also similar to VRC01 in that their binding to gp140 induces a similar conformation change in gp140 as does binding of CD4. In addition, the crystal structure of the Fab region of one of the new bNAbs, 3BNC60, was similar to VRC01, suggesting that it binds gp120 in a similar way.
This suggests that despite the high degree of somatic hypermutation of the CD4 binding site specific bNAbs Nussenzweig and colleagues identified, they are surprisingly similar in structure and sequence. “The path to achieve this particular type of antibodies seems to be fairly narrow,” says Johannes Scheid, a graduate student in Nussenzweig’s lab and first author of the study.
1. Science 2011, doi: 10.1126/science.1207227
2. Nature 458, 636, 2009
Researchers Gather Clues About How Broadly Neutralizing Antibodies Develop
There has been an increasing interest in trying to understand how HIV-specific broadly neutralizing antibodies (bNAbs) develop in HIV-infected individuals (see Vaccines to Antibodies: Grow Up!, IAVI Report, July-Aug. 2010). Now, a new study adds significantly to the understanding of this process. In the study, led by Peter Kwong and John Mascola of the Vaccine Research Center (VRC), researchers for the first time combined deep sequencing technology with bioinformatics to identify bNAb sequences from millions of variable heavy chain sequences in HIV-infected individuals (1). They also identified the likely precursors from which these antibodies develop through affinity maturation—a process through which an antibody’s variable region becomes different from its germline precursor as the result of somatic hypermutation—which increases the affinity of the antibody to its targets.
To isolate the new VRC01-like antibodies, the researchers used the same CD4 binding site specific bait that was used last year to isolate three bNAbs (VRC01, 02, and 03) from a chronically HIV-infected individual called donor 45. Using this bait, they isolated seven new VRC01-like antibodies from two additional chronically HIV-infected individuals—donor 74 from IAVI’s Protocol G cohort of chronically infected individuals, and donor 0219 from the Center for HIV/AIDS Vaccine Immunology’s 001 cohort.
All of these VRC01-like antibodies—which included the bNAb VRCPG04 isolated from donor 74 and VRC-CH31 from donor 0219—are similar to VRC01 in their breadth and potency, and in that they are CD4 binding site specific, highly affinity matured, and derived from the germline gene IgVH1-2. In addition, the crystal structure of gp120 bound to VRC03 and VRCPG04 showed that they bind gp120 similarly to the way VRC01 binds gp120. Despite this, the sequences of the variable regions of these VRC01-like antibodies were not more similar to each other than to completely unrelated antibodies. “From analysis of 
sequence identities, we were unable to determine if a particular sequence was a VRC01-like antibody or not,” says Kwong, who is chief of the structural biology section at the VRC.
In an attempt to identify VRC01-like antibodies just by deep sequencing analysis, the researchers also determined the sequences for hundreds of thousands of variable heavy chains expressed by B cells from donors 45 and 74 with 454 pyrosequencing, a next-generation sequencing technology.
Because just comparing sequences would likely miss VRC01-like antibodies that differ in sequence but not in structure, they selected sequences that were derived from the germline gene IgVH1-2, highly affinity matured, and related to known VRC01-like antibodies previously isolated from these donors. They also made the assumption that antibodies with different sequences but similar structures probably share a similar precursor or intermediate along their affinity maturation pathway. To find such intermediates, they did a phylogenetic analysis of the variable heavy chain sequences from donors 45 and 74, using them to construct an evolutionary tree.
To calibrate the trees, they included the variable heavy chain sequences that came from known VRC01-like antibodies from different donors, and found that these known VRC01-like antibodies shared the same precursor sequences with each other, and with certain unknown antibody sequences from the donors. This suggested that these unknown sequences encoded VRC01-like antibodies that could neutralize HIV. And, indeed, when they tested the neutralization capabilities of these sequences—by combining the encoded antibodies with light chain sequences from known VRC01-like bNAbs—they found that the resulting antibodies could neutralize HIV, some quite potently, whereas antibodies encoded by sequences that didn’t share common precursors with known VRC01-like antibodies could not neutralize HIV.
The researchers also used the sequence of CDRH3, one of the most variable parts of the heavy chain variable region of the antibodies, as a unique marker for antibodies that are likely derived from the same B cell in the germline to reconstruct how they accumulate somatic hypermutations on their way to the fully affinity-matured bNAb. This identified several maturation lineages of VRC01-like neutralizing antibodies within donor 74, suggesting that VRC01-like antibodies are not that rare, even within the same individual.
“The unique feature of this study is the deep sequencing analysis of the HIV Env-specific B-cell receptors from multiple HIV infected subjects, combined with their functional characterization,” says Leo Stamatatos, director of the viral vaccine program at Seattle BioMed, who was not involved in the study. “The amount of work it involves is out of this world.”
Kwong says the common precursors of VRC01-like antibodies identified by this phylogenetic analysis could be used to guide development of a vaccine that stimulates the immune system to generate these precursors. The B-cell sequences of vaccinees could then also be checked for these precursors to see if the vaccine-induced affinity maturation goes in the right direction, he adds.
“We now have come up with tools that will actually give us the entire maturation pathway for every single antibody that we choose to examine with this deep sequencing method,” Kwong says. “[The] next step [is] to make immunogens that will be effective in eliciting or producing these maturation series.”
1. Science 2011, doi: 10.1126/science.1207532