Naive antibody gene-segment frequencies are heritable and unaltered by chronic lymphocyte ablation

Abstract

A diverse antibody repertoire is essential for an effective adaptive immune response to novel molecular surfaces. Although past studies have observed common patterns of V-segment use, as well as variation in V-segment use between individuals, the relative contributions to variance from genetics, disease, age, and environment have remained unclear. Using high-throughput sequence analysis of monozygotic twins, we show that variation in naive V(H) and D(H) segment use is strongly determined by an individual's germ-line genetic background. The inherited segment-use profiles are resilient to differential environmental exposure, disease processes, and chronic lymphocyte depletion therapy. Signatures of the inherited profiles were observed in class switched germ-line use of each individual. However, despite heritable segment use, the rearranged complementarity-determining region-H3 repertoires remained highly specific to the individual. As it has been previously demonstrated that certain V-segments exhibit biased representation in autoimmunity, lymphoma, and viral infection, we anticipate our findings may provide a unique mechanism for stratifying individual risk profiles in specific diseases.

Fig. 1.

Heritable variation observed in naive V_H-gene use profiles. (A) Hierarchical clustering of naive V_H-gene profiles from twins and biological replicates. Genetically distinct groups can be distinguished with high confidence by approximately unbiased multiscale bootstrapping (AU 100%). (B) V_H-gene use profile correlation between biological replicates (r = 0.994), monozygotic twins (r = 0.989, r = 0.972), and unrelated individuals (r = 0.870). (C) Significant heritable differences in V-gene profiles between twin pairs observed for germ lines IGHV1-2, IGHV1-3, IGHV1-8, IGHV4-b, IGHV5-a, and IGHV7-4–1 (*P < 4.7e-04).

Fig. 2.

Limits of heritable variation observed in naive V(D)J recombination repertoire. (A) Significant heritable differences in D-gene profiles between twin pairs observed for nine D-segments (*P < 5.3e-4). (B) Correlation coefficients for D-segment use within biological replicates, within twin pairs, and between twin pairs. Reported with 95% confidence intervals. (C) V-D observed recombination frequencies are highly correlated to expected frequencies (r = 0.95 ± 1.7). (D) D-J observed recombination frequencies are highly correlated to expected frequencies (0.97 ± 0.01).

Fig. 3.

Impact of chronic immunosuppressive therapy on twin B1. (A) SHM load in variable domains of IgM repertoire. In healthy twins A1, A2, and B2, 30–40% of unique clones contain >4 bp SHM. In treated twin B1, >97% of the IgM repertoire appears naive. Of the sequences from the TCRB repertoire of each twin, 10,000 were sequenced as a negative control for SHM detection. (B) Histograms of CD27⁺ cells gated on IgM⁺/CD20⁺. Healthy twins A1, A2, and B2 have 70–72% CD27⁻ cells, but treated twin B1 has 97.5% CD27⁻ cells.

Fig. 4.

Unique V(D)J clone overlap between IgM and class-switched IgG and IgA. Unique counts shown in upper diagonal, percent overlap [(A ∩ B)/min(A,B)] in the lower diagonal, and unique clones in sample along diagonal. In the upper diagonal, counts for twins A are in green, for twins B are in orange. Log₁₀ heatmap colors sequence counts. Percentage of overlap in the bottom diagonal is dark gray if over 1%, light gray if 0.1–1%, and white if less than 0.1%.

Fig. 5.

Ratio of V-gene use in activated to naive repertoire. (A) Ratio of heavy-chain V-gene use between class-switched IgA and IgG to naive IgM. Twins A1, A2, and B2 show significant correlation between naïve and class-switched V-gene use (Pearson's r: P = 3.6e-4, 3.3e-4, 2.0e-8, respectively). The MS Twin B1, indicated by red open circles, shows a different ratio in V-gene use between class-switched IgA and IgG to naive IgM. The null ratio of 1 is indicated with an orange line. Elevated V-genes are indicated in a separate scale to the right. Individual isotypes IgM, IgG, and IgA are shown in Fig. S5. (B) Same as above for V_κ and V_λ: (Pearson's r: P < 1e-10 for A1, A2, and B2).