Shared VH1-46 gene usage by pemphigus vulgaris autoantibodies indicates common humoral immune responses among patients

Abstract

Pemphigus vulgaris (PV) is a potentially fatal blistering disease caused by autoantibodies (autoAbs) against desmoglein 3 (Dsg3). Here, we clone anti-Dsg3 antibodies (Abs) from four PV patients and identify pathogenic VH1-46 autoAbs from all four patients. Unexpectedly, VH1-46 autoAbs had relatively few replacement mutations. We reverted antibody somatic mutations to their germline sequences to determine the requirement of mutations for autoreactivity. Three of five VH1-46 germline-reverted Abs maintain Dsg3 binding, compared with zero of five non-VH1-46 germline-reverted Abs. Site-directed mutagenesis of VH1-46 Abs demonstrates that acidic amino-acid residues introduced by somatic mutation or heavy chain VDJ recombination are necessary and sufficient for Dsg3 binding. Our data suggest that VH1-46 autoantibody gene usage is commonly found in PV because VH1-46 Abs require few to no mutations to acquire Dsg3 autoreactivity, which may favour their early selection. Common VH gene usage indicates common humoral immune responses, even among unrelated patients.

Figure 1

Anti-Dsg3 mAbs reproduce the PV phenotype. (a)Indirect immunofluorescence staining of human skin with scFv mAbs demonstrates cell surface binding typical of PV. Negative control is an scFv mAb against an irrelevant Ag. Positive control is a previously characterized anti-Dsg3 scFv mAb, (D31)2/29. Scale bar, 20 μM. (b) VH1-46 mAbs, but not VH5a mAb, induce suprabasal skin blisters in human skin organ culture. Scale bar, 100 μM. Data are representative of 3 independent experiments.

Figure 2

VH1-46 anti-Dsg3 mAbs have relatively few CDR replacement mutations. Frequency of replacement (R, black) and silent (S, white) mutations in the CDRs and FWRs are shown for (a) VH1-46 anti-Dsg3 mAbs and (b) select members of other anti-Dsg3 clonal lineages. “>N” = N/0. VH1-46 Abs have relatively low R:S ratios in the CDRs compared to the FWRs, represented by the proportion of the bar that is black. In contrast, VH1-69 anti-Dsg mAbs ((D31)2/29 and (D3)1d/2c), have more frequent R mutations and elevated R:S ratios in the CDRs compared to the FWRs, with statistical evidence of positive antigen-driven selection (shown in Supplementary Table 3).

Figure 3

VH1-46 mAbs require few to no somatic mutations to bind Dsg3 (a) F706 germline-reverted (GL) and (b) F779 GL Abs do not bind Dsg3. (c) 3.2 GL Ab binds Dsg3 by ELISA with reduced affinity (specificity of binding confirmed by surface plasmon resonance, Table 3). (d) PVE4-8 GL and (e) 4.2 GL Abs retain binding to Dsg3 by ELISA. Black/gray hatched lines indicate GL1/2 Abs, respectively. (f-j) Germline-reverted non-VH1-46 mAbs do not demonstrate binding to Dsg3 by ELISA. GL LC/HC recombinant Abs were generated for those mAbs that were insoluble in their fully GL state. (k) Indirect immunofluorescence (IIF) staining of human skin confirms Dsg3 autoreactivity of PVE4-8 GL and 4.2 GL Abs in native skin tissue. IIF binding of 3.2 GL was not detectable, presumably due to low affinity. Scale bar, 20 μM. Error bars indicate s.e.m. Data are representative of 3-5 independent experiments.

Figure 4

Acidic amino acid residues in the CDRs confer Dsg3 binding.(a) Dsg3 autoreactivity of F779 somatically mutated (SM), F779 germline-reverted (GL) in which all somatic mutations have been removed, F779SM (D/E>N/Q) in which only two acidic amino acid residues in the heavy chain CDR2 and light chain CDR1 are mutated to their polar analogs (F779SM D/E>N/Q), and F779GL + D/E, in which only the two acidic somatic mutations are reintroduced into the GL Ab. Binding of F779GL1+ D/E (a) and F706GL2 + D/E (b) to native human skin tissue was confirmed by indirect IF. (b) Dsg3 autoreactivity of F706SM, F706GL2, and F706GL2 + D/E is similar to that seen for F779. (c) 4.2GL1/GL2 both retain binding to Dsg3, but removal of two acidic residues in the heavy chain CDR3 (4.2GL1/GL2 D>N) abolished binding by ELISA. (d) PVE4-8GL retains binding to Dsg3, which depends on 3 acidic residues in the heavy chain CDR3 (PVE4-8 GL HC D>N). Scale bar, 20 μM. Errors bars indicate s.e.m. Data are representative of 3 independent experiments.