Researchers in the laboratory of Frederick Alt at the Howard Hughes Medical Institute and Program in Cellular and Molecular Medicine (PCMM) at Children's Hospital Boston, led by Leng-Siew Yeap and Joyce K. Hwang, have fundamentally changed our understanding of how the crucial mutagenic activity of activation-induced cytidine deaminase (AID) is targeted during antibody maturation.

In an article published online in Cell on November 19, 2015, they demonstrate that the DNA sequence encoding the antigen-binding variable (V) region of B-cell antibodies lies in a genomic location privileged for mutational diversification by AID.  Their study further reveals that no specialized mechanism beyond sufficient exposure to AID is required to generate antibodies whose neutralizing function depends on high levels of variable-region mutations and deletions.

The Alt lab has a long history of work on the genetic processes that constitute antibody gene diversification. V(D)J recombination is an early immune process during which variable (V)-region exons encoding the antigen-recognition sites of B lymphocyte antigen receptors are assembled from component gene segments.  When B cells are activated by a specific antigen, these BCRs can be secreted as specific antibodies.  In response to antigen, somatic hypermutation (SHM) hugely increases antibody variety/specificity by introducing point mutations into the antigen-binding V region (a million times more frequently than in the genome at large). Finally, class switch recombination (CSR) further modifies the genetic regions encoding a B-cell's IgH constant region to customize antibody effector functions—the ways in which the antibody attacks and eliminates a pathogen.

AID (activation-induced cytidine deaminase) is a relatively small (24 kDa) enzyme that is sometimes referred to as the master regulator of antibody diversification, because it initiates both SHM and CSR.  For SHM, AID introduces point mutations in variable-exon DNA, and for CSR AID introduces DNA double-strand breaks in switch (S) regions flanking the constant-region exons to allow them to recombine and replace the initially expressed constant region with another one.  An important unresolved and debated question with respect to antibody maturation has been whether AID uses the same downstream pathways for generating SHMs in V exons and DSBs for CSR.

Antibodies are composed of Ig heavy and light chains, each of which includes variable regions that contain complementarity-determining regions (CDRs), which are the most frequent sites of somatic hypermutation (SHM).  Each V exon has three CDRs, two of which are encoded by the germline V segment; the other is encoded by sequences assembled at V, D, and J junctions during V(D)J recombination. Because CDRs are the portion of variable exons that make direct antigen contact, SHMs in these regions are critical for generating antibody diversity.  CDRs are separated from each other in the V exon by framework regions that are much less variable, as they are responsible for general antibody structure.

Thus, the organism has at its disposal millions of different B cells, each with different antibodies on its surface. In germinal centers (GCs), which are specialized areas of immune activity in the spleen and in structures called Peyer's patches (PP) in the small intestine, mature B lymphocytes proliferate in response to foreign antigens and undergo both SHM and CSR to optimize the resulting immune response.

The Alt team, led by Dr. Yeap and M.D./Ph.D student Hwang, set about learning more about how AID is targeted.  A strong motivation behind this multi-year project was to transcend the limitations of previous approaches to study V-region exons.  One seemingly intractable problem was the difficulty of obtaining enough sequences to accurately follow the targeting of AID during antibody responses. 

Previous researchers in this field had attempted to elucidate AID targeting in mice by inserting a segment of DNA containing a V-region sequence isolated from one mouse into the genome of another (transgene), and saw that they were sometimes mutated by AID.  However, there were problems with the reproducibility of findings using this approach that resulted from integration of the transgenes at various random genomic sites and in multiple copies.  Thus, while some potentially interesting observations were made, most workers in the field considered them unconvincing. Nevertheless, the Alt lab felt that this technique, approached in a new and more controlled fashion, could prove quite rewarding.

Dr. Alt identifies the starting point for this line of inquiry: "How does substrate DNA sequence influence targeting and outcome of AID activity? How does AID targeting lead to mutations in V-region exons and breaks in S-region exons? Are those totally different processes, as many people would argue, or can you explain them as different outcomes of the same reaction?"

The Alt lab built upon the passenger alleles approach by generating B cells that have one productive (receptor-producing) V-gene allele on one parental chromosome with an identical "passenger" V gene allele that cannot be translated into protein or selected in an immune response. This approach eliminates bias by selection for antigen or against detrimental mutations that inactivate the B-cell receptor, as well as eliminating potential position or copy number effects of randomly inserted V exon transgenes.  Alt explains, "We designed this gene-targeted knock-in approach along with our immunization conditions to see how mechanistic factors -- as opposed to the cellular selection process that molds normal immune responses -- affect AID targeting."

For this work, the Alt lab also developed and employed an elegant animal system, called RAG-deficient blastocyst complementation.  This approach allowed them to very rapidly generate mice in which all B cells derived from embryonic stem (ES) cells that harbor both the productive test VDJ allele and the identical nonproductive "passenger" allele incapable of encoding proteins.  The second phase was to carry out short-term immunization of those offspring, i.e., enough to begin SHM but not sufficient to allow positive selection of B cells expressing a particular antibody. GC B cells were then sorted and subjected to allele-specific high-throughput sequencing, which yielded much more SHM data than prior researchers in this field had been able to obtain.

The team analyzed thousands of V-region sequences from mouse GC B cells and plotted the number of mutations per sequence. The findings were both striking and unexpected: not only did the productive and passenger alleles mutate in essentially the same locations (i.e., predominantly in CDRs) and to the same extent in a given mouse, but the same amount of SHM occurred even in bacterial sequences inserted into the passenger location.  The major difference between V-exon and bacterial sequences, which are totally divergent from V exons, was that the latter had AID hotspot mutated sequences that were totally unrelated to CDRs.  These findings revealed that the location of the V exon in the genome, rather than the V exon sequence per se, targeted it for high-level SHM.

Another notable finding was that both passenger allele sequences and bacterial sequences underwent substantial levels of deletions in regions of the test sequences that had high levels of AID SHM activity.  Moreover, the spots where the deletions occurred matched those where AID caused the most SHMs.  Such deletions were not routinely observed in prior studies that focused mainly on the productive V exon, which would be inactivated by deletions and, unlike the passenger, lead to death of the B cells in which it was inactivated.  

In studies of B cells activated for CSR, the team made a finding that was both very surprising to the field and complementary to their finding of AID-initiated DSBs and deletions in V exons of GC B cells: namely that V exons and S regions underwent SHMs at similar levels when normalized for numbers of AID target sites.  Thus, the ability of S regions to support frequent DSBs relies on their having many more AID target sites rather than any differential AID targeting or outcome at any given target. 

Dr. Alt summarizes these aspects of the work succinctly: "Many publications have said that AID mutational activity in SHM and deletional activity in CSR can be separated, that they're different. We've shown instead that they're mechanistically linked, intimately associated." 

A potentially important practical implication of the findings from these studies came from experiments in which the Alt lab separated B cells from test animals into those derived from the spleen and those from the intestinal Peyer's patches. Again, both sets of cells showed AID targeting for mutations and deletions in the same places across the productive and the non-selected passenger alleles (including bacterial sequences), though the Peyer's patch B cells consistently demonstrated many sequences with much more extensive mutation than the splenic B cells.  Alt's team suggests that these findings could result from the known chronic activation of Peyer's patch B cells, which might produce high levels of SHM and deletion of the V-exon repertoires due to prolonged AID exposure in the absence of cellular selection for antibody affinity maturation. 

Antibodies are among the fastest-growing group of drugs, and many current vaccine strategies seek to elicit generation of highly mutated neutralizing antibodies.  The implications of the findings in this publication have stimulated considerable interest among vaccinologists.  For example, a hugely diverse repertoire of antibodies that have already been highly mutated but have not yet been selected would provide a rich resource for the development of vaccines and therapeutic antibodies. One translational proposal made in the study is that the highly mutated PP germinal center B-cell antibody reservoir might be tapped to elicit antibodies with abundant SHM and/or deletions: for example, certain anti-HIV-1 broadly neutralizing antibodies.

The paradigm-changing data presented by the Alt lab provide a simple explanation for something many researchers expected to be more complicated, and practically speaking may help inform new approaches to elicit therapeutic antibodies.