Single cell analysis of RA synovial B cells reveals a spectrum of ectopic lymphoid B cell activation and hypermutation characterized by NR4A nuclear receptor expression Supplemental material A A H A A H - - - - - CD3 SSC FSC SSC SSC SSC FSC-A FSC-W SSC-W DAPI CD45 CD19 B RA172_SYN RA195_SYN 105 5 SM USM CD45+ 10 SM USM CD45+ 4 105 10 DN N 105 4 DN N 10 37.5 49.3 4 BV605 3 4 49.1 10 10 10 CD27 3 7.2 10 1.1 3.3 CD27 CD27 3 2 3 10 10 57.6 CD3 2 31.6 10 10 0 CD3 PECy5 0 10.8 2 2 21.6 37.2 10 5.5 2 3 4 5 10 0 0 10 10 10 10 0 102 103 104 105 IgDIgD FITC 0 IgD 5 0 102 103 104 105 10 0 102 103 104 105 105 CD19CD19 PEDazzle594 CD19 - hi 4 - hi CD24 CD27 10 CD24 CD27 4 10 CD3 2.4 1.9 BV605 103 103 CD27 CD27 CD27 102 102 0 0 47.5 39.0 0 102 103 104 105 2 3 4 5 CD24 A647 0 10 10 10 10 CD24 CD24 RA221_SYN RA134_SYN 5 10 SM USM 105 SM USM CD45+ 104 DN N CD45+ 4 DN N 10 CD24 105 21.8 105 BV605 3 4.9 BV605 41.6 10 103 37.4 1.7 4 4 32.1 CD27 CD27 10 CD27 10 CD27 2 2 10 38.6 10 3 0 3 10CD3 10 0 12.1 34.3 CD3 48.7 CD3 PECy5 CD3 PECy5 2 3 4 5 0 10 10 10 10 0 102 103 104 105 2 20.8 2 2.7 10 IgD FITC 10 IgD FITC 0 0 IgD 5 5 2 3 4 5 10 0 102 103 104 105 10 0 10 10 10 10 CD19 PEDazzle594 CD19 PEDazzle594 CD19 - hi CD19 CD24 CD27 104 104 2.6 7.6 CD24-CD27hi BV605 BV605 103 103 CD27 CD27 CD27 CD27 2 102 10 0 0 60.8 2 3 4 5 0 102 103 104 105 0 10 10 10 10 CD24 A647 CD24CD24 A647 CD24 RA134_BLD SM USM CD45+ fig. S1. Flow cytometry schema for sorting and DN N analysis of B cells. (A) Flow cytometry gating strategy CD24 for sorting of B cells. (B) Flow cytometry analysis for CD27 B cell phenotyping for each sample. CD27 vs. IgD plot CD3 shows % of switched memory (SM), un-switched memory (USM), double negative (DN) and Naïve (N). - hi IgD CD27 vs. CD24 identify CD24 CD27 plasmablast. CD19 CD24-CD27hi CD27 CD24 A B C D Isotype BCR not recovered M BCR recovered D A G K L N/A # of cell E F Plasma (ii) IGHA Plasma(ii) ● IGHD Plasma (i) IGHG Plasma(i) ● IGHM memory NR4A+ ● LMNA+ Naive(iii) ● NR4A+ Naive(ii) ● Naïve (iii) Memory ● Naïve (ii) LMNA+ ● Naïve (i) −2.5 0.0 2.5 5.0 7.5 % point increase in SHM (vs Naive) fig. S2. Evidence of class-switched recombination and clonal expansion in RA synovium. (A) Bar graph showing percentage of cells with productive V-J spanning pair out of total barcode recovered per sample, RA134 PBL (1559 cells), RA134 SYN (694 cells), RA172 SYN (457 cells), RA195 SYN (1240 cells) and RA221 SYN (607 cells). (B) Bar graph displaying percentage recovery of paired heavy-light chain, single heavy or light chain or non-canonical pairing out of total, RA134 PBL (1348 cells), RA134 SYN (187 cells), RA172 SYN (140 cells), RA195 SYN (850 cells) and RA221 SYN (225 cells). (C) Bar graph showing number of cells where BCR was recovered or not recovered per cluster and per sample. (D) Pairing of heavy and light chains in RA195 SYN. Each point represents a cell. Clones that occurred 2 or more times are shown. The amino acid sequence of the CDR3 of the heavy and light chains are shown below and the left of the x and y axes, respectively. The color of the label represents the isotype. (E) Bar graphs depicting number of cells expressing IGHA, IGHD, IGHG and IGHM per B cell cluster. (F) Difference in SHM rate in cell populations vs Naive(i) using linear mixed effect models with capture:cluster random effect. A B Immediate early genes early BCR signaling B � NF signaling DZ CD40 signaling MYC fig. S3. (A) Heatmap display log2 fold change of immediate early genes, NF�B signaling genes, CD40 signaling genes and MYC genes (Ye et al, 2010; Ollila et al, 2003; Holmes et al, 2020). (B) NR4A+ cluster maintains expression of some DZ genes. Heatmap display log2 fold change of BCR signaling genes and genes associated with DZ GC B cells in each B cell cluster (Holmes et al, 2020). 100X 200X CD20 IgD PCNA CD21 CD20 CD138 CD3 CD68 CD20 fig. S4. Immunofluorescent staining of synovial tissue samples used for B cell single cell RNA sequencing. 5 µm synovial serial sections were stained with Hematoxylin and Eosin or antibodies specific for proliferating cell nuclear antigen (PCNA, white), IgD (green), and CD20 (red) to detect large CD20+IgD-PCNA+ germinal center B cells. Synovial sections were stained with a combination of antibodies specific for CD20 and CD21 to detect CD21+ follicular dendritic cells (red) in germinal centers populated by CD20+ B cells (green). We also stain paraffin sections with antibodies to detect CD138+ plasma cells (red), CD3+ T cells (green), CD68+ macrophages (white), and CD20+ B cells (blue) in the ELS in the inflamed synovia of RA patient. Representative 200x magnification pictures are shown. Scale bars represent 100 µm. A B non-zero expressing cell (% of each subset) Log2FC C D Log2FC non-zero expressing cell (% of each subset) fig. S5. Single cell RNA sequencing data of synovial B cells from AMP phase I (n= 14 RA, Zhang et al., 2019) were re-analyzed and B cell subsets parallel to current study were identified using a supervised classification technique SingleR. (A) Heatmap displays log2 fold change of cytokine and chemokine receptors in each B cell subset. (B) Bar graphs display percentage of cells of each subset that express CXCR5, CCR7, CXCR4 or CCR6. (C) Heatmap displays log2 fold change of cytokines and chemokines in each B cell subset. (D) Bar graph represent percentage of cells of each subset that express IL6, TNFSF9, CD70, TGFB1, CCL5 or LTB. A Cluster CD83 NR4A2 NR4A1 FOSB KLF6 PPP1R15A Dimension1 score FOS B CD69 NR4A1 DUSP1 DUSP2 CD83 ZFP36 6 NR4A2 BTG2 NENormalized Expression Cluster NR4A1 EIF1 6 Naive(i) FOSBCD83 DUSP1CD83 4 KLF6 NR4A2 NR4A2 4 Naive(iii) NR4A1 CDKN1ANR4A1 PPP1R15AFOSB NR4A3FOSB 2 Naive(ii) FOS KLF6 CREMKLF6 2 CD69 PPP1R15A NFKBIDPPP1R15A 0 NR4A+ DUSP2FOS GPR183FOS ZFP36CD69 CD69 0 PC1 Naive(i) DUSP2 NormalizedSRGN Expression Cluster BTG2 JUNDDUSP2 20 Naive(ii) EIF1 ZFP36 6 ZFP36 TXNIP CD79B DUSP1BTG2 NormalizedDNAJA1BTG2 Expression ClusterNormalizedNaive(i) Expression Cluster10 Naive(iii) EIF1 6YPEL5EIF1 6 NR4A+ CDKN1ADUSP1 4 DUSP1 Naive(i)Naive(iii) 6 Naive(i) 4ZNF331 Naive(iii)4 0 NR4A3CDKN1A 2 SLC2A3CDKN1A Naive(ii) Expression (logcounts) Naive(iii) CREMNR4A3 2KDM6BNR4A3 Naive(ii)2 Naive(ii)−10 NFKBIDCREM 0 CREM NR4A+ 4 NFKBID 0LY9 NFKBID NR4A+0 NR4A+−20 GPR183GPR183 IGHG1 SRGN GPR183 PC1 PC1 SRGN MCL1SRGN 2 JUNDJUND IDS JUND 2020 20 DNAJA1DNAJA1 DNAJA1 YPEL5YPEL5 MTRNR2L12YPEL5 1010 10 ZNF331ZNF331 LMNAZNF331 0 0 0 SLC2A3 TXNIPSLC2A3 0 SLC2A3KDM6B TCL1AKDM6B −−1010 −10 0 −10 10 20 −10 0 10 20 KDM6BLY9 RCSD1LY9 −20 PC1 LY9 IGHG1 −20 −20 IGHG1MCL1 IGHMIGHG1 IDS RP11MCL1−160E2.6 MCL1MTRNR2L12 TMSB10IDS IDS LMNA CD79BMTRNR2L12 MTRNR2L12TXNIP LMNA LMNATCL1A FCER2TXNIP TXNIPRCSD1 PLPP5TCL1A IGHM RCSD1 TCL1ARP11−160E2.6 IGHM RCSD1TMSB10 RP11−160E2.6 IGHMCD79B TMSB10 RP11−FCER2160E2.6 CD79B TMSB10PLPP5 FCER2 CD79B PLPP5 FCER2 DUSP1 PLPP5 JUN JUNB BTG1 CXCR4 C Dimension2 score ZFP36L2 LY9 JUN DUSP1 VPS37B 7.5 MT−ND4 DUSP1 MT−ND5 JUN NCF1 NENormalized Expression5.0 Cluster JUNB CD37DUSP1 6 BTG1 IGHMJUN Memory CXCR4 IGHDJUNB 4 2.5 Naive(iii) ZFP36L2 TCL1ABTG1 LY9 FAM129CCXCR4 2 NR4A+ VPS37B BACH2ZFP36L2 0 LMNA+ MT−ND4 LY9 0.0 FOXP1VPS37B PC2 MT−ND5 NormalizedGADD45BMT Expression−ND4 Cluster LMNA+ NCF1 RPS19 20 Memory MT−ND5 S100A6 HOPX CD37 6 RPL13ANCF1 NormalizedMemory Expression7.5 Cluster10 Naive(iii) IGHM VIMCD37 6 NR4A+ IGHD 4 CD99IGHM Naive(iii) Memory0 TCL1A S100A10 4 Naive(iii) 2 IGHD NR4A+ Expression (logcounts) 5.0 −10 FAM129C CRIP2TCL1A 2 NR4A+ BACH2 0 S100A6FAM129C LMNA+ −20 FOXP1 S100A4BACH2 0 LMNA+ PC2 2.5 GADD45B LGALS1FOXP1 PC2 RPS19 CRIP1GADD45B 20 RPL13A LMNARPS19 20 ITGB1RPL13A 10 0.0 VIM HLAVIM−C 10 CD99 CD99 0 0 S100A10 HLAS100A10−A IGHA1 −10 −10 −10 0 10 −10 0 10 CRIP2 ACTBCRIP2 S100A6 S100A6 −20 PC2 ACTG1S100A4 −20 S100A4 GAPDHLGALS1 LGALS1 SH3BGRL3CRIP1 CRIP1 HOPXLMNA LMNA PDE4DITGB1 ITGB1 HLA−C HLA−C HLA−A HLA−A IGHA1 IGHA1 ACTB ACTB ACTG1 ACTG1 GAPDH GAPDH SH3BGRL3 SH3BGRL3 HOPX HOPX PDE4D PDE4D D fig. S6. A gene expression continuum from naïve to NR4A+ B cells. (A) Principal component analysis projecting gene expression data of Naïve (i), Naïve (ii), Naïve (iii), NR4A+, LMNA+ and memory B cell subsets onto two dimensional space. (B) Heatmap display expression of top 40 loading genes for dimension 1 with increasing score from Naïve (i), Naïve (ii), Naïve (iii) and NR4A+ subset. Plots on the right show example of genes increase (NR4A1, DUSP1) or decrease (TXNIP, CD79B) as cells transition from naïve to NR4A+ state. (C) Heatmap display expression of top 40 loading genes for dimension 2 with increasing score from Naïve (iii), NR4A+, LMNA+ and memory subset. Plots on the right show example of genes decrease (JUN, DUSP1) or increase (S100A6, HOPX) as cells transition from NR4A+ to memory state.