Heavy-chain receptor editing unbound

With all of the certainty of death and taxes, questions regarding the mechanisms of immunological tolerance remain fixtures across the landscape of immunology. From Paul Ehlirch’s “horror autotoxicus” (reviewed in ref. 1) to the current synthesis of tolerance by apoptosis, anergy, and receptor editing (2), the mechanisms of immunological tolerance remain, despite substantial research effort, enigmatic.

In PNAS, Kumar et al. (3) and Sun et al. (4) refine our understanding of receptor editing, a principle component of immune tolerance, by demonstrating that at least one form of this tolerance pathway is not driven by reaction with self-antigens.

Receptor Editing and Autoimmunity

B lymphocytes respond to specific antigen ligands, via antigen receptors [B-cell receptors (BCR)] that are assembled by variable (V), diversity (D), and joining (J) [V(D)J] recombination, a process of directed genomic rearrangements that fuses V, D, and J gene segments into functional immunoglobulin genes (5, 6). The combinatorial and junctional diversity created by V(D)J rearrangement is thought to be capable of generating approximately 10¹⁴ distinct BCRs that are first expressed on the surface of immature B cells. This extraordinary potential for diversity has costs: V(D)J recombination often generates self-reactive BCRs that may instigate autoimmune disease (7, 8).

Immature B cells expressing autoreactive antigen receptors (BCRs) can be tolerized by at least three mechanisms: anergy, clonal deletion, and receptor editing (2). Whereas anergy and deletion inactivate or remove self-reactive clones, receptor editing alters BCR specificity through secondary Vκ→Jκ light (L)-chain rearrangements or, more rarely, by altering the variable region of heavy (H)-chains by the insertion of a V_H gene segment into an established V_HDJ_H rearrangement. Editing within the Igκ locus is straightforward: primary Vκ→Jκ rearrangements generally retain flanking Vκ and Jκ elements bearing recombination signals (RSs) optimally oriented for efficient recombination (9). In contrast, H-chain editing requires an atypical V(D)J recombination event mediated by a physiologic RS adjacent to the inserted germ-line V_H gene segment with a cryptic RS (cRS) located near the 3′ end of the rearranged V_H gene element (2, 9–11). V_H replacement reactions can rescue the development of B cells that would otherwise be eliminated by apoptosis and were first noted in mice with autoreactive transgenic BCRs (12, 13). Nonetheless, replacement of nonautoreactive (14–16) and nonproductive (17) VDJ knockins have been observed as well. Under antigen-dependent

Kumar et al. and Sun et al. refine our understanding of receptor editing, a principle component of immune tolerance, by demonstrating that at least one form of this tolerance pathway is not driven by reaction with self-antigens.

models of receptor editing, autoantigen initiates V_H → V_HDJ_H replacement, whereas diversification of the B-cell repertoire via V_H gene replacement is presumably antigen independent (13, 15–18).

The possibility of antigen-independent V_H → V_HDJ_H replacement is crucial, as receptor editing implies an editor. The difference between random secondary V_H → V_HDJ_H rearrangements that are permissive for B-cell survival vs. secondary rearrangements induced by autoantigen is not trivial: is H-chain receptor editing a chance event or is it induced by the signaling of self-reactive BCRs?

The question of antigen-dependent vs. -independent H-chain editing was previously addressed by Davila et al. (11), who demonstrated short-lived molecular intermediates of V_H → V_HDJ_H rearrangements in pro-B but not pre- and immature B cells. Moreover, Davila et al. observed normal levels of H-chain editing activity in pro-B cells of μMT mice homozygous for truncations of the immunoglobulin μ membrane exon and unable to assemble functional BCRs (19). The presence of V_H → V_HDJ_H rearrangements in pro-B cells from μMT mice is inconsistent with an editing process driven by self-antigen; instead, these results implied that V_H → V_HDJ_H replacement occurs spontaneously in pro-B cells. In this scenario, self-antigen does not drive receptor editing but rather selects for edited B cells that are not autoreactive.

The results of Davila et al. (11) were controversial, as Zhang et al. (18) had reported exactly reciprocal observations, finding evidence for H-chain editing only in pre-B cells; in subsequent work, Liu et al. reported V_H → V_HDJ_H replacement events driven by BCR signaling in human immature/transitional (new emigrant) B cells (20). Zhang et al. suggested that that H-chain editing induced by autoreactivity is a significant factor in diversifying the murine and human B-cell repertoires (18, 21–23).

Editing Without Autoreactivity

Kumar et al. (3) and Sun et al. (4) substantially advance this earlier work through the generation of novel experimental models and address the issue of antigen-dependent vs. antigen-independent H-chain editing; both papers conclude that, although V_H → V_HDJ_H replacement can increase BCR diversity, it does not serve as a significant mechanism for self-tolerance.

Sun et al. (4) build on an earlier study of H-chain editing in a knockin mouse line that expressed a nonfunctional form (D23^stop) of the V_HDJ_H rearrangement from D23 hybridoma cells that secrete DNA-binding antibody (14, 17). Remarkably, substantial numbers of B cells were present in mice homozygous for the D23^stop allele; B-cell development had been rescued by V_H → V_HDJ_H and forbidden V_H to J_H rearrangements (17). Sun et al. (4) generate analogous D23^prod mice that express a functional V_HDJ_H knockin allele and BCR that, when complemented by the D23 L-chain, are autoreactive. In mice heterozygous for the D23^prod allele, a small but consistent fraction (4%) of mature B cells expressed the WT locus. These cells carried a D23^prod knockin allele that had been inactivated by out-of-frame V_H → V_HDJ_H replacements.

When the D23 L-chain was crossed onto D23^prod mice, evidence for autoreactivity was observed, but there was no increase in V_H replacement. Indeed, even when the D23^prod allele was crossed onto the κ-macroself genetic background that renders all κ⁺ B cells autoreactive (24), the frequency of V_H replacement events at the D23^prod allele did not increase. These data appear to virtually eliminate any role for V_H replacement in mitigating autoimmunity.

Kumar et al. (3) generate a remarkably authentic model for studying V_H → V_HDJ_H replacements by using somatic cell nuclear transfer to generate mouse lines bearing the Igh and Igκ gene rearrangements of a single IgA/κ plasmacyte. The transnuclear mouse lines generated are exceptionally interesting for a variety of reasons, particularly the demonstration that the IgA constant region can support all stages of B-cell development except the B1-cell compartment. The transnuclear BCR, VHQ52^NT;Vκgr32^NT, did not appear to be autoreactive and showed no evidence of spontaneous signaling, suggesting that it did not recognize any self-ligand. Nonetheless, in heterozygous VHQ52^NT-HT mice carrying WT Igh alleles, approximately 20% of B cells expressed IgM rather than IgA because the VHQ52^NT rearrangement had been inactivated by an out-of-frame V_H → V_HDJ_H replacement. These nonproductive V_H insertions were present in pro-B cells and were highly diverse, suggesting substantial rates of secondary V_H replacement rather than the selective expansion of rare clones. These authors, like Davila et al. (11), conclude that V_H → V_HDJ_H replacements begin in early pro-B cells and occur even in the absence of pre-BCR signaling. These observations also minimize the possibility that H-chain editing is induced by the signaling autoreactive BCR.

These new papers (3, 4) emphasize the difficulties in distinguishing biological selection from induction. Viewed in this context, spontaneously edited V_HDJ_H rearrangements that spare B-cell development or promote survival may give the appearance of a directed process when in fact they are not. The issue of induction of receptor editing vs. selection for edited receptors remains an important debate, but for V_H replacement, at least, the editor is blind.