Defective germinal center selection results in persistence of self-reactive B cells from the primary to the secondary repertoire in Primary Antiphospholipid Syndrome

Abstract

Primary antiphospholipid syndrome (PAPS) is a life-threatening clotting disorder mediated by pathogenic autoantibodies. Here we dissect the origin of self-reactive B cells in human PAPS using peripheral blood and bone marrow of patients with triple-positive PAPS via combined single-cell RNA sequencing, B cell receptors (BCR) repertoire profiling, CITEseq analysis and single cell immortalization. We find that antiphospholipid (aPL)-specific B cells are present in the naive compartment, polyreactive, and derived from the natural repertoire. Furthermore, B cells with aPL specificities are not eliminated in patients with PAPS, persist until the memory and long-lived plasma cell stages, likely after defective germinal center selection, while becoming less polyreactive. Lastly, compared with the non-PAPS cells, PAPS B cells exhibit distinct IFN and APRIL signature as well as dysregulated mTORC1 and MYC pathways. Our findings may thus elucidate the survival mechanisms of these autoreactive B cells and suggest potential therapeutic targets for the treatment of PAPS.

Fig. 1. PAPS patients have similar phenotypic B cell features compared to HD.

A Composite UMAP projection of B cell subsets, generated using 12 parameters analyzed in spectral flow cytometry of PAPS patient and HD samples (n = 30). Composite sample was derived from 5000 representative cells. B Frequencies of B cell subsets from PAPS patients (n = 15) and HD (n = 15). C Composite UMAP projection and frequencies of antigen-experienced B cells. D BCR isotype repartition in patients and HD samples. E Frequencies of aN and rN B cells from patients (n = 15) and HD (n = 15). F Composite UMAP projection and frequencies of DN B cells from patients (n = 15) and HD (n = 15). Significance was determined using a Two-tailed Mann–Whitney test. Data are presented as mean values ±SEM. aN activated naive B cells, DN double negative, HD healthy donor, PAPS primary antiphospholipid syndrome; rN resting naive B cells.

Fig. 2. Reduced diversity in BCR repertoire of PAPS patients.

A Schematic of peripheral B cell isolation from PAPS (n = 10) and HD (n = 7) blood followed by CITE seq including single-cell transcriptomic, antibody repertoire sequencing and ADT. VDJ sequencing was used to obtain paired heavy- and light-chain V(D)J with an average of 5,231 cells recovered per library. B–D IGHV (B), IGKV (C), and IGLV gene (D) usage frequencies in peripheral B cells from PAPS patients (red) and HD (blue). E, F IGHD families (E) and IGHJ gene (F) usage frequencies in peripheral B cells from PAPS patients (red) and HD (blue). G, H CDR3 length (G) and somatic hypermutations per VH segment (H) (nucleotide exchanges compared with the nearest germline gene segment). I, J Chao1 (I) and D50 diversity index (J) according to isotype in PAPS patients (red) and HD (blue). For IGHA diversity, n = 9 in patient group, as one patient did not have IgA⁺ B cells. Data are presented as mean values ±SEM. Significance was determined using a Two-tailed Mann–Whitney test. ADT antibody derived tag, HD healthy donor, PAPS primary antiphospholipid syndrome.

Fig. 3. PL-reactive B cells are part of the natural repertoire and persist after the naive B cell stage in triple-positive PAPS.

A Flowchart of single B cell immortalization and supernatant screening. B Frequency of PL-reactive in each B cell subsets. C IgG subclasses repartition in non-PL and PL-reactive clones. D Kappa/lambda ratio of non-PL and PL-reactive clones in each B cell subsets. aPL antiphospholipid antibody, CL cardiolipin, PL phospholipid.

Fig. 4. Naive pool of aPL B cells is largely polyreactive.

A HEp-2 reactivity of non-PL and PL-reactive clones, derived from PAPS patients, considering IgM⁺CD27^-, IgM⁺CD27⁺ and IgG clones. Pie charts represent the frequencies of reactive (solid) and nonreactive (open) clones, with the number of clones tested (n) indicated in the center (left). B ELISA reactivity of non-PL and PL-reactive clones, derived from PAPS patients, against nucleosome, thyroglobulin, myosin, insulin and tetanus toxin. C K-means clustering plot showing cluster analysis of clones, derived from PAPS patients, reactivity in ELISA (left) and representation of the dimensional repartition of clones according to the PL reactivity (top right) and the number of different antigens recognized among nucleosome, thyroglobulin, myosin, insulin and tetanus toxin (down right). D Pie charts represent the number of reactivity detected among aPL (solid) and non-aPL clones (open), derived from PAPS patients, with the number of clones tested (n) indicated in the center. Significance was determined using a χ2 test. aPL antiphospholipid antibody.

Fig. 5. Phospholipid-reactive B cells harbor specific BCR in PAPS patients.

A IGHV gene frequency of aPL B cell clones (red) compared to non-aPL clones (blue) for each patient. The box depicts the relative frequencies of IGHV3-11, 3-21, 3-74, 4-34, and 4-39 usage in aPL clones and non-aPL clones for each PAPS patient. B IGHD family frequency of aPL compared to non-aPL clones. Significance was determined using a χ² test. C IGHJ gene frequency of aPL compared to non-aPL clones for each patient. D Number of CDR3 positively charged amino acid and arginine residues in CDR3 in aPL (n = 184) compared to non-aPL clones (n = 192). E Number of additional nucleotides in aPL (n = 182) compared to non-aPL clones (n = 188). F Length of CDR3 in aPL (n = 184) compared to non-aPL (n = 192) clones. G Number of SHM per VH segment (nucleotide exchanges compared with the nearest germline gene segment) in naive B cells, unswitched and switched memory B cells in aPL (n = 173) compared to non-aPL (n = 184) clones. H Selected IGHV gene frequency in CD27⁺ and CD27^- population of aPL and non-aPL clones for each patient. I IGHV gene frequency of bone marrow plasma cells from two PAPS patients. Data are presented as mean values ±SEM. Significance was determined using a Two-tailed Mann-Whitney test, unless otherwise stated. aPL antiphospholipid antibody, HD healthy donor, PAPS primary antiphospholipid syndrome, R arginine, SHM somatic hypermutation.

Fig. 6. Peripheral B cells from PAPS patients exhibit mTORC1 and MYC transcriptomic signature.

A UMAP projection and clustering of 78,355 B cells analyzed by scRNA-seq. B Bar plot showing cell frequencies from PAPS patients (n = 10) and HD (n = 7) that contributed to cluster c1 to c14. C UMAP projection showing six annotated B cell subtypes. D Plot showing the top 10 GSEA signaling pathway significantly enriched in total B cells from PAPS patients as opposed to HD. E Plots depicting the enrichment of MYC and mTORC1 GSEA signaling pathways in naive and memory B cells from PAPS patients as opposed to HD. F Heatmap listing top genes of mTORC1 signaling pathway enriched in PAPS patients. G Projected enrichment score of indicated pathway for each cell onto UMAP from APS patients. The darker shading represents higher enrichment. H Dot plot for expression of B cell cytokine receptors among naive clusters in all samples. Colors represent minimum-maximum normalized mean expression, and sizes indicate the proportion of cells expressing respective genes. I TACI MFI on naive B cells from two representative APS and two HD (left); TACI and BAFF MFI ratio on naive B cells from APS patients as opposed to HD (right) using FCM. Data are presented as mean values ±SEM. Significance was determined using a Two-tailed Mann–Whitney test. FCM flow cytometry, FDR False Discovery Rate, GSEA Gene Set Enrichment Analysis, HD healthy donor, PAPS primary antiphospholipid syndrome.

Fig. 7. MYC-mTORC1 dysregulation is associated with follicular differentiation and impaired positive selection.

A Circos plot depicting the IGHV:IGLV and IGHV:IGKV pairings for a HD and two representative PAPS patients. Pairing between IGHV3-11, 3-21, 3-74, 4-39, or 4-34 (pink segment), and IGKV (blue segment) or IGLV (green segment) gene are shown by the connecting lines inside the circle with thickness corresponding to pairing frequencies. B Frequencies of IGHV3-21, 3-74, 4-39 clonotypes in IGHM and IGHG/A isotypes in PAPS patients (n = 10) and HD (n = 7). C Violin plot representing normalized ADT level for B cell signaling and activation markers in IGHV4-39 clonotypes from APS patients (dark red), HD (dark blue), other clonotypes from PAPS patients (light red) and HD (light blue). D Dot plot for expression of mTORC1 and MYC selected genes among naive and memory cells according to clonotypes in PAPS patients and HD. E Dot plot for expression of pre-GC and plasmablast selected genes in naive cells according to clonotypes in PAPS patients and HD. F Violin plot representing pre-GC signature, SELL and CXCR5 gene expression according to clonotypes in patients and HD. G Frequency of plasmablast (CD38⁺⁺CD27⁺⁺) after B cell stimulation with CD40L + IL-21 + IL-4 for 6 days. Two-tailed Mann–Whitney test was used to compare groups. H Representative lineage trees for two clones in APS patients. I Representative amino acid VH alignment for IGHV4-39 clone (clone 1 depicted in H). J Frequency of positively charged amino acid induced by SHM in VH genes (top) of switched IgG/A B cells from HD (n = 8191) and PAPS patients (n = 9406). K Frequency of arginine residues induced by SHM in VH genes CDR of switched IgG/A B cells from HD (n = 7229) and PAPS patients (n = 6101). Data are presented as mean values ±SEM. Significance was determined using a Two-tailed Mann–Whitney test. ADT antibody-derived tag, CDR complementary determining region, HD healthy donor, MFI Mean fluorescence intensity, PAPS primary antiphospholipid syndrome, SHM somatic hypermutation.