Embracing diversity gives antibodies the power to bind

Abstract

The ability to respond to a diverse range of antigens is the hallmark of the adaptive immune system. In particular an expansive repertoire of antibodies is generated by a number of processes both during B-cell development and a functional B-cell response. One of the primary drivers of diversification is V(D)J recombination during development. Human antibody genes are formed from random selection of one of ~40 V genes, 23 D and 6 J gene segments to form the heavy chain, and ~70 V and 9 J gene segments that form the light chain of the prospective antibody. Diversity is further increased by the imprecise nature of joining the V, D and J genes together leading to insertions and deletions in that joining area. This site of V(D)J joining encodes the CDR3 region that is often in the heart of the antigen-binding region of the antibody. As any given heavy chain can pair with any light chain the number of possible combinations has been estimated at 5 × 1013.1 Further antibody diversification can occur during a productive germinal center reaction in which activated B cells undergo somatic hypermutation and class switch recombination.2 On top of this a recent study by Lanzavecchia and colleagues showed that the genes for some Plasmodium protein-binding antibodies contain large DNA fragments from the collagen-binding LAIR1 protein spliced directly into the CDR3 region.3 While an exciting new mode of antibody diversification, this specific insertion was only found in 2 out of 500 Plasmodium-exposed Kenyan individuals, thereby raising the possibility that this was a rare phenomenon as opposed to a common mechanism. A new study by the same group addresses this question through the broad screening of two large cohorts in Tanzania and Mali.4 These studies demonstrate that roughly 5–10% of Africans in malaria endemic regions had detectable levels of LAIR1-integrated antibodies, with <1% of Europeans showing similar levels. Strikingly in these new donors, they discovered other methods by which LAIR1 could insert itself into the antibody sequence to create antibodies capable of recognizing Plasmodium-infected cells. In two donors, LAIR1 was inserted into the switch μ region of the heavy chain resulting in B cells that were able to produce two antibody variants through alternative splicing, one expressing LAIR1 in the elbow region of the antibody and one expressing no LAIR1. The antibodies carrying the LAIR1 insert, as well as the original variable region, were bispecific. In contrast, the previously identified antibodies, with LAIR1 inside the CDR3, lost their original antigen specificity. The identification of a permissive site for inserting additional binding domains to create bispecific antibodies might be a useful tool in biotechnology. Strikingly, a third donor also had a LAIR1 insertion in the switch μ region, but this insertion was accompanied by deletions removing large chunks of the variable domain. The result of these changes was the creation of a Plasmodium-binding ‘antibody’ that in fact had no light chain at all, akin to the antibodies found in camelids5 (Figure 1). Interestingly, in all the antibodies identified, the LAIR1 insertions had undergone somatic hypermutation. By making synthetic constructs corresponding to the unmutated ancestor of one of the antibodies it was shown that these precursor antibodies had detectable, but very low affinity for Plasmodium-infected red blood cells (iRBCs). The somatic mutations not only increased affinity to iRBCs but also decreased binding to collagen. The ability of an antibody to bind collagen could of course result in autoimmunity, emphasizing the importance of the germinal center not only in enhancing affinity but decreasing autoreactivity. This supports the finding that autoantibodies can be ‘redeemed’ in the germinal center,6 and notably many broadly neutralizing antibodies for HIV are derived from germline precursors that are autoreactive.7 Clearly in the search for strong pathogen binding, the immune system is willing to tolerate some risk of autoimmunity. However, in this case, despite the increased breadth and opsonizing activity, individuals carrying LAIR1-containing antibodies were not protected against febrile malaria compared to other exposed individuals from malaria endemic areas. Perhaps the most exciting finding from the new study was that this mechanism of antibody diversification appears to be universal and not merely a quirk of malaria infection. This was hitherto unclear, as Plasmodium infection is known to induce genomic instability, which has been suggested to be the cause of the high prevalence of Burkitt's Lymphoma in areas where malaria is endemic.8 However, European donors also carried inserted sequences at a frequency of roughly 1 in every 1000 switched memory B cells with a fraction of those cells containing inserts with the potential to create bispecific antibodies. These inserts were not from LAIR 1 but almost exclusively from other genes that are highly expressed in B cells. This suggests that B cells in germinal centers are using templates from whatever RNA is at hand to make these modifications. Such insertions have probably been missed in previous studies of antibody repertoires as the alignment algorithms would not have recognized these sequences as consistent with productive V(D)J recombination. In the case of the LAIR1 antibodies these insertions were found because antibody sequences were identified from monoclonal populations, demonstrating the power of examining antigen specific B cells at the single-cell level. These interesting findings demonstrate that this once unknown form of antibody diversification is actually a relatively common method used to diversify the antibody repertoire. Much like the process of somatic hypermutation, random insertion will generate novel clones with diversified specificities that can then be selected into subsequent responses. These findings highlight the critical need for B-cell diversity and the germinal center machinery that enhances it, even at the risk of a loss of certain specificities or autoimmunity. The authors declare no conflict of interest. Different modes of LAIR1 insertion to create Plasmodium-specific antibodies. LAIR1 sequences can be inserted into antibodies either at the site of VDJ joining to create an additional binding loop (a), or into the switch μ region either creating a bi-specific antibody with an additional binding loop (b), or creating a truncated single-chain Camel-like antibody (c). Adapted from Pieper at el.4