IL-23 Promotes a Coordinated B Cell Germinal Center Program for Class-Switch Recombination to Igg2b in BXD2 Mice

IL-23 Promotes a Coordinated B Cell Germinal Center Program for Class-Switch Recombination to IgG2b in BXD2 Mice

This information is current as Huixian Hong, Min Gao, Qi Wu, PingAr Yang, Shanrun Liu, of October 2, 2021. Hao Li, Peter D. Burrows, Daniel Cua, Jake Y. Chen, Hui-Chen Hsu and John D. Mountz J Immunol published online 17 June 2020 http://www.jimmunol.org/content/early/2020/06/16/jimmun ol.2000280 Downloaded from

Supplementary http://www.jimmunol.org/content/suppl/2020/06/16/jimmunol.200028 Material 0.DCSupplemental http://www.jimmunol.org/

Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication by guest on October 2, 2021 *average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2020 by The American Association of Immunologists, Inc. All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. Published June 17, 2020, doi:10.4049/jimmunol.2000280 The Journal of Immunology

IL-23 Promotes a Coordinated B Cell Germinal Center Program for Class-Switch Recombination to IgG2b in BXD2 Mice

Huixian Hong,* Min Gao,† Qi Wu,* PingAr Yang,* Shanrun Liu,* Hao Li,‡ Peter D. Burrows,x Daniel Cua,{,1 Jake Y. Chen,† Hui-Chen Hsu,* and John D. Mountz*,‖

IL-23 promotes autoimmune disease, including Th17 CD4 T cell development and autoantibody production. In this study, we show that a deficiency of the p19 component of IL-23 in the autoimmune BXD2 (BXD2-p192/2) mouse leads to a shift of the follicular T helper cell program from follicular T helper (Tfh)–IL-17 to Tfh–IFN-g. Although the germinal center (GC) size and the number of GC B cells remained the same, BXD2-p192/2 mice exhibited a lower class-switch recombination (CSR) in the GC B cells, leading to lower serum levels of IgG2b. Single-cell transcriptomics analysis of GC B cells revealed that whereas Ifngr1, Il21r, and Il4r genes exhibited a synchronized expression pattern with Cxcr5 and plasma cell program genes, Il17ra exhibited a synchronized Downloaded from expression pattern with Cxcr4 and GC program genes. Downregulation of Ighg2b in BXD2-p192/2 GC B cells was associated with decreased expression of CSR-related novel base excision repair genes that were otherwise predominantly expressed by Il17ra+ GC B cells in BXD2 mice. Together, these results suggest that although IL-23 is dispensable for GC formation, it is essential to promote a population of Tfh–IL-17 cells. IL-23 acts indirectly on Il17ra+ GC B cells to facilitate CSR-related base excision repair genes during the dark zone phase of GC B cell development. The Journal of Immunology, 2020, 205: 000–000. http://www.jimmunol.org/ ollicular T helper (Tfh) cells have been reported to express transcription factor and the cytokine IL-21 are key initial events well-characterized transcription factors and different cy- that enable the establishment of the Tfh cell program and upreg- F tokines that define this Th cell subset as well as molecules ulation of adhesion molecules, leading to effective, long-term, that promote T–B cell interactions (1, 2). Upregulation of the Bcl6 and productive interactions between T and B cells in the germinal center (GC) (3–6). The spectrum of CD4+CXCR5+PD- 1+ICOS+CD40L+ Tfh cells has increased over the past several *Division of Clinical Immunology and Rheumatology, Department of Medicine, years to include multiple subpopulations (5, 7, 8). Although University of Alabama at Birmingham, Birmingham, AL; †Informatics Institute, ‡ classical Tfh cells produce IL-21, sublineages and pathogenic Tfh the University of Alabama at Birmingham, Birmingham, AL; Beth Israel Deaconess by guest on October 2, 2021 Medical Center, Harvard Medical School, Boston, MA; xDepartment of Microbiol- cells, including IFN-g–producing Tfh–IFN-g, IL-4–producing ogy, University of Alabama at Birmingham, Birmingham, AL; {Discovery Research, ‖ Tfh–IL-4, and IL-17–producing Tfh–IL-17, have been identified Merck Research Laboratory, Boston, MA; and Department of Medicine, Birmingham VA Medical center, Birmingham, AL in both humans and mice (9–11). In BXD2 autoimmune mice, 1Current address: Janssen Pharmaceuticals, Spring House, PA IL-17 was shown to contribute to spontaneous GC formation and autoantibody production by enhancing and prolonging the inter- ORCIDs: 0000-0002-5592-9349 (M.G.); 0000-0002-7201-4212 (S.L.); 0000-0002- 2171-8826 (H.L.); 0000-0002-5752-5647 (P.D.B.); 0000-0001-8829-7504 (J.Y.C.); action between Tfh cells and B cells through the desensitization of 0000-0002-0317-5725 (H.-C.H.). chemokine receptor responses through regulator of G-protein Received for publication March 16, 2020. Accepted for publication May 14, 2020. signaling (RGS) 13 and RGS16 (5, 12–15). Such a response + + This work was supported by the National Institutes of Health (R01-AI-071110, R01 promotes the migration arrest of CXCR4 - or CXCR5 -respond- AI134023, and P30-AR-048311 to J.D.M. and R01-AI-083705 to H.-C.H.), U.S. ing cells and thereby is critical to enable prolonged interactions of Department of Veterans Affairs grants (I01BX004049 and 1I01BX000600 to J.D.M.), and the Lupus Research Alliance (a distinguished innovator award to J.D.M. and an GC B cells with other cells for GC B cell affinity maturation. alliance target identification award to H.-C.H.). Similarly, Tfh–IFN-g cells can promote the development of au- The single-cell RNA sequencing data presented in this article have been submitted to toreactive B cells outside the GC through the T-bet transcription the Gene Expression Omnibus database (https://www.ncbi.nlm.nih.gov/geo/query/ program (16). Despite these previous insights, it is not clear if acc.cgi?acc=GSE145922) under accession number GSE145922. these different Tfh cytokines operate concomitantly or sequen- Address correspondence and reprint requests to Dr. John D. Mountz, Department of Medicine, Division of Clinical Immunology and Rheumatology, University of tially in different niches or stages of GC B cell development to Alabama at Birmingham, Shelby Interdisciplinary Biomed Research Building, facilitate pathogenic autoantibody formation. It is also unknown if Room SHEL 307, 1825 University Boulevard, Birmingham, AL 35294-2182. the functions of one lineage can be replaced by the others. E-mail address: [email protected] It is also perplexing that activation-induced cytidine deaminase The online version of this article contains supplemental material. (AID) expression and activity primarily occurs in the highly Abbreviations used in this article: Ad, adenovirus; AID, activation-induced cytidine proliferative CXCR4+ dark zone (DZ) GC B cells, leading to deaminase; B6, C57BL/6; contig, contiguous sequence of overlapping DNA fragments; CSR, class-switch recombination; DZ, dark zone; GC, germinal center; KEGG, Kyoto extensive somatic hypermutation (SHM) (17, 18). In contrast, Tfh Encyclopedia of Genes and Genomes; LZ, light zone; NEIL, nucleotide excision repair- cells interact primarily with CXCR5+ GC B cells in the light zone like; NHEJ, nonhomologous end joining; PALS, periarteriolar lymphatic sheath; PB, plasmablast; PNA, peanut agglutinin; POLD, DNA polymerase d; qRT-PCR, quantitative (LZ) for fine-tuning of selection and plasma B cell development RT-PCR; RF, rheumatoid factor; RGS, regulator of G-protein signaling; scRNA-seq, (19–21). We previously have shown that the spontaneous autore- single-cell RNA sequencing; SHM, somatic hypermutation; Tfh, follicular T helper; active GCs in BXD2 mice lead to the development of autoanti- UAB, University of Alabama at Birmingham; WT, wild type. bodies through the upregulation of AID, which promotes SHM Copyright Ó 2020 by The American Association of Immunologists, Inc. 0022-1767/20/$37.50 and class-switch recombination (CSR) (22, 23).

www.jimmunol.org/cgi/doi/10.4049/jimmunol.2000280 2 IL-23 PROMOTES A COORDINATED B CELL GC PROGRAM

In addition to AID, both SHM and CSR require Bcl6-directed 660 anti-GL7 (clone GL7), FITC anti–PD-1 (clone J43), and Alexa Fluor programs of both B cell survival (24) and DNA damage and re- 647 anti–IL-17A (clone eBio17B7). Peanut agglutinin (PNA) was conju- pair genes (25–27). For CSR, nonhomologous end joining (NHEJ) gated with biotin (Vector Laboratories, Burlingame, CA) and detected by Alexa Fluor 488–conjugated streptavidin (Invitrogen). Dead cells were first requires transcription initiating upstream of the IgH constant excluded from analysis with allophycocyanin–eFluor 780 Organic Via- region undergoing CSR, followed by AID-mediated dsDNA bility Dye (eBioscience). breaks and then double-stranded break repair, resulting in a syn- For cytokine-producing T cell analysis, cells were stimulated for 5 h with apse and joining of the two switch regions (28). Known classical PMA (50 ng/ml; Sigma-Aldrich) and ionomycin (750 ng/ml; Sigma- Aldrich) in the presence of GolgiPlug (BD Biosciences). Cells were NHEJ players include Ku70 (XRCC6), Ku80 (XRCC5), XRCC4, stained for surface markers and then fixed and permeabilized with Cytofix/ DNA ligase IV, and DNA-PKcs (29), whereas the nonclassical (or Cytoperm solution (BD Biosciences) before intracellular staining (5). alternative) NHEJ pathway uses XRCC1 and DNA ligase III Data were acquired using standard flow cytometry (LSR II; BD Biosci- for DNA end joining (30, 31). Other DNA damage/repair gene ences) and analyzed with FlowJo_v10 software. FACS sorting was performed families, including histone variants H2a and H3a, which encode using a FACSAria cell sorter (BD Biosciences). Both instruments were lo- cated in the University of Alabama at Birmingham (UAB) Comprehensive core histone proteins with substantial variations in their amino and Flow Cytometry Core in the Shelby Biomedical Research Building. C-terminal regions, DNA polymerase d (POLD) (32, 33), and nucleotide excision repair-like (NEIL) (34, 35), have been less Histology and confocal imaging analysis well studied in CSR. Although AID is considered the master reg- Frozen sections of mouse spleens were subjected to histology and confocal ulator of DNA breaks and NHEJ that enable CSR and we previously imaging. Immunofluorescent staining was performed as previously de- showed that a deficiency of IL-17RA in the lupus-prone BXD2 scribed (14). Images were taken using a Nikon A1 at the UAB High Resolution Imaging Facility. Imaging data analysis was carried out using mice affected the expression of AID in vivo and diminished the

the ImageJ software, version 1.47, developed by the U.S. National Insti- Downloaded from development of IgG autoantibodies (12), how IL-23 or IL-17 reg- tutes of Health (42). ulate AID-induced NHEJ to promote CSR in immunized or spontaneously autoreactive GC B cells is largely unknown. ELISA IL-23 is an IL-12–related cytokine composed of a p40 subunit The quantification of Ig and autoantibody levels in mouse serum was that it shares with IL-12 and a unique p19 subunit (36). In this performed using an ELISA method as previously described (14, 22). 2/2 study, we have generated BXD2-p19 mice to determine if Real-time quantitative RT-PCR IL-23 acts through modulating Tfh cell lineage development to http://www.jimmunol.org/ regulate spontaneous GC formation and class-switched autoanti- The expression of genes in sorted GC B cells and Tfh cells was determined using quantitative RT-PCR (qRT-PCR) following the previously described body production in BXD2 mice (5, 12, 14, 15, 22, 23, 37, 38). We method (12, 43). Spleen cells were purified using flow cytometry, gated to 2/2 found that although BXD2-p19 mice did not have an impaired either live CD4+PD-1+CXCR5+ Tfh cells or live CD19+GL7+Fas+ GC spontaneous GC response, there was a significantly lower IgG2b B cells. Primers used are listed in Supplemental Table I. response that could not be compensated for by other Tfh cell Single-cell library generation and sequencing subsets. By performing single-cell transcriptomics to elucidate + + + mechanisms, we showed that IL-23 promotes AID-mediated CSR Spleen cells were FACS sorted into live CD19 GL7 Fas GC B cells. 3 by upregulating histone variant genes, including H1fx, H2afv, Single cells were captured into Gel Bead-In EMulsions using a 10 chromium controller, from which single-cell 59-biased transcriptome li- by guest on October 2, 2021 H2afx, H2afz, and H3f3a as well as base excision repair genes braries and V(D)J libraries were prepared using a 59 Library and Gel Bead Pold4 and Neil1. Expression of Il17ra, but not Ifngr1, Il4ra,or Kit (PN-1000014) and Single Cell V(D)J Enrichment Kit for Mouse Il21r, was synchronized with the expression of Aicda and other B Cells (PN-1000016, 103 Genomics), according to the 103 Genomics GC program genes. Our results suggest a novel concept that manual. Final 59-biased transcriptomic libraries were sequenced using an Illumina NextSeq 500 with a targeting minimum of 50,000 read pairs/cell various subsets of Tfh cells act at different stages of GC B cell and 26 3 98–bp read lengths for the sequencing cycles. Final enriched development and that IL-23–generated Tfh–IL-17 cells are the V(D)J libraries were sequenced using an Illumina MiSeq with a minimum important GC Tfh cell program to ensure efficient AID-mediated of 5000 read pairs/cell and 150 3 150–bp read lengths. CSR in spontaneously autoreactive GCs in BXD2 mice. Processing of single-cell sequencing raw data Raw sequencing data were processed with the Cell Ranger pipeline software Materials and Methods (v.3.0.2; 103 Genomics). Raw base call files generated by NextSeq 500 or Mice MiSeq (Illumina) were converted to FASTQ files using cellranger mkfastq with default parameters. For single-cell 59 transcriptome data, cellranger C57BL/6 (B6) and BXD2 recombinant inbred mice were purchased from count was run with –transcriptome = refdata-cellranger-mm10-3.0.0 for 2/2 The Jackson Laboratory. B6–IL-23 p19 mice were kindly provided by each sample, in which reads were aligned against the mouse genome 2/2 Merck Research Laboratories (Palo Alto, CA) (39, 40). B6-Il17ra mice (mm10) using the spliced transcripts alignment to a reference aligner (44), 2/2 were obtained from Amgen (Thousand Oaks, CA). B6–IL-23 p19 and and the unique molecular identifiers were counted for each gene. The gene 2/2 B6-Il17ra mice were backcrossed with BXD2 mice using a marker- expression matrix in the outputs of cellranger count was used for down- associated speed congenic approach for seven generations to generate stream analysis. Single-cell V(D)J BCR data were processed by the 2/2 2/2 BXD2-p19 and B6-Il17ra mice (5). cellranger vdj pipeline to perform quality control, assembling, quantification, and annotation of paired V(D)J transcript sequences. Briefly, after Administration of adenovirus LacZ and adenovirus IL-23 trimming and filtering, a de novo assembly algorithm from cellranger vdj Adenovirus (Ad) IL-23 and AdLacZ (2 3 109 PFU per mouse), generous pipeline was applied to the reads of each cell barcode. A set of assembled gifts from Dr. J. Kolls (University of Pittsburgh) (41), were administered contiguous sequences of overlapping DNA fragments (contigs), which i.v. Spleen tissue and cells were analyzed 5 d later. represent the best estimation of transcript sequences, were generated via the Smith–Waterman algorithm (45) in cellranger vdj pipeline. The as- Flow cytometry analysis and sorting sembled contigs in each cell were aligned against all of the sequences from mouse reference genome assembly mm10 in this pipeline. The filtered Single-cell suspensions of spleen were harvested and subjected to standard contig annotation files were used for downstream analysis. cell-surface staining or intracellular staining following the manufacturer’s instructions. Anti-mouse Abs included BioLegend Pacific Blue anti-CD19 Classification of B cells using single-cell transcriptome data (clone 6D5), PE anti–IL-23R (clone 12B2B64), allophycocyanin anti–IFN- and single-cell BCR data g (clone XMG1.2), PE-Cy7 anti-IgM (clone RMM-1), BD Biosciences PE anti-CD95 (clone Jo2), PE-Cy7 anti-CXCR5 (clone 2G8), Brilliant Seurat (46) (v.3.0.0), implemented using the R package, was applied to Violet 510 anti-IgD (clone 11-26c.2a), Thermo Fisher Scientific eFluor exclude low-quality cells in different single-cell experiments. Cells that The Journal of Immunology 3 expressed fewer than 200 genes were filtered out. Genes that were not result was visualized by bar and dot chart using “barplot” and “dotplot” detected in at least three single cells were excluded. Based on these cri- functions, respectively. The “cnetplot” function was used to visualize the teria, we retained the following numbers of genes/cells in each sample: enrichment map and extract the potential biological complexities associ- 14,255 genes/2437 cells in BXD2-p192/2 no. 1, 13,366 genes/1084 cells ated between genes and diseases. in BXD2-p192/2 no. 2, 14,233 genes/2586 cells in BXD2-p192/2 no. 3, 14,479 genes/3158 cells in BXD2-wild type (WT) no. 1, and 14,195 genes/ SeqGeq single-cell RNA sequencing data analysis 1936 cells in BXD2-WT no. 2. The filtered contig annotation files gen- Transcriptome gating analysis was carried out using the SeqGeq platform erated from single-cell V(D)J BCR data were merged with filtered single- (v1.6.0; BD Life Sciences – Informatics, Ashland, OR). Following the use cell transcriptome databased on barcode information in R. Single cells with of gene expressed versus library size quality control, B cells were gated as overlapped barcodes between BCR data and transcriptome data were BCR gene-positive cells. SeqGeq analysis was used to gate on cell subsets identified as B cells. The transcriptome datasets for B cells were loaded based on the expression of gene-encoding Tfh cytokine receptors and then into the R package Seurat. The Seurat data matrix object for each dataset enabled analysis in quantitation of expression of significantly altered gene was created using the “CreateSeuratObject” function. The Seurat objects pairs including the histone variant genes, DNA excision repair genes, and for B cells were filtered using the function “subset.” The individual B cells Rgs13 and Aicda. The axis labels show that this indicates relative ex- Ms4a1 were selected for positive expression of the coding gene for pression of the indicated genes on a hyperbolic arcsine (Arcsinh) trans- (CD20) and CD19, key B cell markers. Only cells that expressed the VDJ formation scale (50, 51). The arcsinh values are calculated by applying the regions of BCR as well as Ms4a1 (the coding gene for Cd20)andCd19 arcsinh equation divided by the scale argument to the measured intensity were used for further analysis. The numbers of single cells used for the value data that are displayed in a linear-like fashion. The statistical anal- transcriptomics analysis for each mouse are 566, 299, 984, 1136, and 567 2 2 2 2 2 2 ysis was calculated as a Chi-Square significance of each percentage rela- p19 / p19 / p19 / for BXD2- no. 1, BXD2- no. 2, BXD2- no. 3, BXD2-WT tive to the percentage for the indicated gene pair for GC B cells expressing no. 1, and BXD2-WT no. 2, respectively. the Il17ra+ gene in WT BXD2 compared with the same gene pair in other cytokine receptor–positive B cells in both WT and BXD2-p192/2 mice. Identification of differentially expressed genes Statistical analysis Downloaded from The procedures for normalization, variable gene selection, scaling the data, and identifying differentially expressed genes were performed by following Results are shown as the mean 6 SD or mean 6 SEM. Pearson’s normality the Seurat documentation (v.3.0.0) (47, 48). In brief, the data were nor- test was used to determine the normal distribution of each dataset. A two- malized using the function “NormalizeData,” which uses a global-scaling tailed, unpaired Student t test was used when two normally distributed normalization method “LogNormalize” to normalize the gene expression groups of datasets were compared for statistical differences. The Mann– measurements for each cell to the total gene expression. The scale factor Whitney nonparametric test was used when the two datasets for compar- was 10,000. The highly variable genes were identified by the “FindVar- ison were not normally distributed. An ANOVA test was used when more http://www.jimmunol.org/ iableFeatures” function, and the method of selection was “vst.” Next, the than two groups were compared for statistical differences. Multiple com- data were scaled using the function “ScaleData” in Seurat. The guided parisons were performed using Tukey multiple comparisons test. Fisher clustering was performed based on the information of samples and exact test was used for contingency tests. Analysis of correlation between cells. The B cell cluster was identified as described above. The samples variables was performed using linear regression in GraphPad Prism soft- 2/2 2/2 2/2 BXD2-p19 no. 1, BXD2-p19 no. 2, and BXD2-p19 no. 3 were ware. The p values of ,0.05 were considered significant. identified as “ko” cluster, and the samples BXD2-WT no. 1 and BXD2-WT no. 2 were identified as “wt” cluster. Differential gene expression analyses Data availability were performed using the Seurat function “FindMarkers.” To identify differentially expressed genes, the Wilcoxon rank-sum test was performed The single-cell transcriptome data and single-cell V(D)J BCR sequencing with the default threshold of 0.25 for log fold change and a filter for the data used in this study have been deposited in the Gene Expression Om- 2 nibus database under the accession number GSE145922 (https://www.ncbi. minimum percentage of cells in a cluster of .25%. Differentially by guest on October 2, 2021 expressed genes were isolated by comparing significantly upregulated nlm.nih.gov/geo/query/acc.cgi?acc=GSE145922). Raw files used in the genes and downregulated genes defined as having an adjusted p value of figures that support the findings of this study are available from the cor- ,0.05. The top differentially expressed genes were visualized by heatmap responding authors upon reasonable request. using the ComplexHeatmap R package. Results Single-cell trajectory analysis IL-23 regulates the Tfh cell switch from Tfh–IFN-g to R package Monocle (2.6.4) (47, 48) was used to investigate the develop- Tfh–IL-17 but does not change Tfh cell frequency mental trajectories of single cells by a method similar to the procedure 2/2 described in Qiu et al. (47, 48). The Seurat objects from the last step We have generated p19-deficient BXD2 mice (BXD2-p19 )to were loaded into Monocle. To reduce the bias caused by technical varia- determine if the lack of IL-23 can alter Tfh cell lineage devel- tion, sequencing depth, and capture efficiency, data normalization was opment and if such an alternation can influence spontaneous GC performed using the functions “estimateSizeFactors” and “estimateDis- and autoantibody production. The percentage and the absolute persions.” The low-quality cells (i.e., doublets, triplets or libraries that + + + were accidentally formed from multiple cells) were filtered out. The number of classical Tfh cells (CD4 PD-1 CXCR5 ) were similar 2/2 minimum expression threshold was set at 0.1. Genes were filtered out in BXD2 and BXD2-p19 mice (Fig. 1A, 1B). Importantly, based on the average expression level and unusual variations across cells. however, the IL-23 deficiency significantly suppressed the ex- For single-cell trajectories in B cells, a set of ordered genes that define pression of IL-17 while enhancing the expression of IFN-g among B cell development were collected to order cells for supervised trajecto- 2 2 Tfh cells in the BXD2-p19 / mice (Fig. 1C, 1D). In addition, we ries. These genes included Ighm, Ighd, Ighg2b, Ighg2c, Ighg1, and other 2/2 key genes involved in GC development pathways. The expression profiles identified that Tfh cell phenotype between BXD2-p19 and 2 2 were reduced to two dimensions using the DDRTree algorithm in the BXD2-Il17ra / mice are not identical in that there was no de- function “reduceDimension.” Cells were ordered into a trajectory using the crease in the percentages of IL-17 or increase in the percentage of “orderCells” function. The function “plot_pseudotime_heatmap” was used IFN-g in IL-17RA–deficient BXD2 mouse Tfh cells compared with to visualize modules of genes that covaried across pseudotime. The “plot_genes_in_pseudotime” function was used for individual graphs to WT BXD2 mice (Fig. 1C, 1D). These results suggest that although depict the kinetic trend of the expression levels of genes with significant IL-23 promotes Tfh–IL-17 development, the effects of IL-23 and changes through pseudotime. IL-17 on autoreactive GC development may be different. We next determined if excessive production of IL-23 in vivo Kyoto Encyclopedia of Genes and Genomes pathway analysis can promote Tfh–IL-17 development. This was tested using an The clusterProfiler (49) package (v3.10.1) was applied to identify enriched IL-23–expressing Ad system. We previously showed that after pathways based on top differentially expressed genes (adjusted p value AdIL-23 administration to B6 mice, the level of serum IL-23 ,0.05). The ranked gene list was prepared following the the clusterPro- filter documentation. The number of selected genes associated with Kyoto increased over baseline by day 3, achieved its maximal level Encyclopedia of Genes and Genomes (KEGG) pathways was assessed via by day 5, and then returned to baseline by day 10 (52). Con- 2 2 the function “enrichKEGG.” The p value cutoff was 0.05. The enrichment sistent with the pro–IFN-g effects of BXD2-p19 / on Tfh cells 4 IL-23 PROMOTES A COORDINATED B CELL GC PROGRAM Downloaded from

FIGURE 1. IL-23 regulates switching of Tfh–IL-17 and Tfh–IFN-g.(A) Representative flow cytometry analysis of different subsets of CD4+ T cells based on their expression of PD-1 and CXCR5 in the indicated strains of mice. (B) The percentage of PD-1+CXCR52, PD-1+CXCR5+, and PD-12CXCR5+ http://www.jimmunol.org/ cells in the CD4 T cell subset (A) are shown as bar plots. (C) Representative intracellular flow cytometry analysis of splenic IL-17A and IFN-g–producing CD4+ ICOS+CXCR5+ Tfh cells in the indicated strains of mice are shown. (D) The mean (6SD) of the percentage of IL-17A (left) and IFN-g (middle)– producing Tfh cells in (C) are shown as bar plots. The ratio between IL-17A– and IFN-g–producing Tfh cells in each strain is shown (right). (E) qRT-PCR analysis of the expression levels of the transcription factor and cytokine genes of different CD4 subsets in CD4+ PD-1+CXCR5+ cells from B6 mice after 5 d of PBS, AdLacZ, or AdIL-23 administration. Statistical differences were determined by an ANOVA test and Tukey multiple comparisons test for significance between groups (n = 4–6 in each group, two independent experiments). *p , 0.05, **p , 0.01, ***p , 0.005.

2/2

described above, the overexpression of IL-23 in B6 mice pro- and RF IgM levels in serum of BXD2-Il17ra mice were de- by guest on October 2, 2021 moted the expression of Tfh–IL-17 cytokine genes (Il17a and creased, as we previously reported (Fig. 2C) (5, 12). Consistent Il17f) (Fig. 1E) but hindered the expression of the Tfh–IFN-g with the serum autoantibody levels, there was an increase in IgM- transcriptional factor (Tbet) and cytokine (Ifng) genes in Tfh cells producing plasmablasts (PBs) and a decrease in IgG-producing after 5 d of AdIL-23 administration (Fig. 1E). AdIL-23 also PBs in the BXD2-p192/2 mice, whereas both IgM- and IgG- suppressed the expression of canonical Tfh cells (Bcl6 and Il21) as producing PBs were diminished in BXD2-Il17ra2/2 mice com- well as Tfh–IL-4 transcription factor and cytokine (Gata3 and Il4) pared with BXD2 mice (Fig. 2D–F). The results thus far suggested genes (Fig. 1E). These results suggest a pro-Tfh–IL-17 effect of that IL-23 promoted CSR in BXD2 mice. IL-23 and that such effects can lead to compensatory alteration in IL-23 is not required for GC formation other lineages of Tfh cells in the absence of IL-23. Unlike BXD2-Il17ra2/2 mice, which had smaller GCs and a lower IL-23 is important for B cell CSR frequency of Fas+GL7+ GC B cells, as we previously reported We have shown that IL-23 positively drives its own signal through (5, 12), the size of spontaneous GCs and percentages of Fas+GL7+ the induction of the IL-23R (52). The levels of Il23r in GC B cells GC B cells were slightly but not significantly increased in BXD2 in both AdLacZ- and AdIL-23–administered B6 mice were very mice lacking IL-23–p19 (Figs. 2D, 3A, 3B). The results indicated low. In contrast, there was an almost 100-fold higher level of Il23r that IL-23 is dispensable for GC formation. Consistent with this, in Tfh cells compared with GC B cells (Fig. 2A). Further, AdIL-23 the administration of AdIL-23 to BXD2-p192/2 mice did not administration significantly induced the expression of Il23r in Tfh change the frequency and numbers of GC B cells (Fig. 3C, 3D). 2 2 cells of both B6 and BXD2-p19 / mice (Fig. 2A, 2B), suggesting There was detection of IgD+PNA+GL-7+ GC B cells in AdIL-23– that Tfh cells, but not GC B cells, are the major target of IL-23. administered BXD2-p192/2 mice, suggesting that founder GC Having established a positive effect of IL-23 in promoting Tfh– B cells were recruited to the GC (53). However, there was an IL-17 and diminishing other lineages of Tfh cells, we next de- increased percentage of PNA+GL-7+ IgM2IgD2 switched GC termined if the lack of IL-23 would influence the spontaneous GC B cells in BXD2-p192/2 mice administered AdIL-23 compared response and autoantibody formation in BXD2 mice. Comparison with the AdLacZ group of mice at day 5 (Fig. 3E). Taken together, 2 2 of the serum autoantibody levels in B6, BXD2, BXD2-p19 / , these results suggested that IL-23 is dispensable for GC formation, 2 2 and BXD2-Il17ra / mice revealed that there were lower levels of but AdIL-23 can induce B cell CSR. IgG anti-DNA and rheumatoid factor (RF) in BXD2-p192/2 mice Decreased expression of IgG2b in the spleens of compared with BXD2 mice, although the anti-DNA and RF IgM 2/2 autoantibody levels were significantly increased (Fig. 2C). Inter- BXD2-p19 mice estingly, these results cannot be fully explained by the lower To further delineate the effects of IL-23 and therefore Tfh–IL-17 in signaling through IL-17RA because the IgG and IgM anti-DNA regulating GC B cell CSR, single-cell RNA sequencing (scRNA-seq) The Journal of Immunology 5 Downloaded from http://www.jimmunol.org/ by guest on October 2, 2021

FIGURE 2. IL-23 is important for B cell CSR. (A) qRT-PCR analysis of IL-23r expression by FACS-sorted CD19+GL-7+Fas+ GC B cells and CD4+ PD- 1+CXCR5+ Tfh cells from the spleens of B6 mice 5 d after PBS, AdLacZ, or AdIL-23 administration. (B) Flow cytometry analysis of IL-23R expression on CD4+ PD-1+CXCR5+ Tfh cells from the spleens of BXD2-p192/2 mice 5 d after AdLacZ or AdIL-23 administration. Left, Representative FACS plot. Right, Bar graphs of mean (6SD) of percentages and absolute numbers of IL-23R+ Tfh cells. Two-tailed Student t test (n = 4–6 per group). (C) ELISA results of anti-DNA and RF Abs in serum from the indicated strains are presented as OD absorbance at 450 nm. Each symbol represents an individual mouse; small horizontal lines indicate the mean (n = 5–15, two independent measurements). Results are mean 6 SD. (D) Representative confocal images from the indicated strains stained with anti-IgM (red), anti-CD4 (white), anti-IgG (green), and PNA (blue). Top, Objective lens, original magnification 34 power view; Lower, Digitally magnified views of the indicated areas on the top. (E) Left, Representative confocal images from the indicated strains stained with anti-IgG (top) and anti-IgM (bottom) (objective lens, original magnification 34 power view). Right, Digitally magnified views of the indicated areas in the left. (F) ImageJ quantitation of the fluorescent intensity (FL) shown in (E) of IgG and IgM in PALS. The mean (6SD) FL intensity is shown. At least four representative areas from three to four spleens were quantitated. The statistical difference was determined by ANOVA test and Tukey multiple comparisons test for significance between groups for (A, C, and F). *p , 0.05, **p , 0.01, ***p , 0.005, ****p , 0.001 for (A–C and F). was carried out using sorted GC B cells obtained from two reduced levels of IgG2b in the serum BXD2-p192/2 mice com- WT BXD2 mice (1703 cells) and three BXD2-p192/2 mice (1849 pared with WT BXD2 mice were verified using a conventional cells). Consistent with the circulating autoantibody results, a ELISA analysis (Fig. 4B). There was also a significantly lower Seurat-based differentially expressed gene analysis revealed that intensity of IgG2b plasma B cells in the periarteriolar lymphatic Ighm and Cd38 were two highly overexpressed genes, whereas sheaths (PALS) in the spleen of BXD2-p192/2 mice, confirming Ighg2b was a highly underexpressed gene in BXD2-p192/2 mice that the production of IgG2b in BXD2-p192/2 mice was signifi- compared with WT BXD2 mice (Fig. 4A). Elevated IgM and cantly compromised (Fig. 4C, 4D). 6 IL-23 PROMOTES A COORDINATED B CELL GC PROGRAM

FIGURE 3. IL-23 is dispensable for GC formation. (A) Representative FACS plot of splenic CD19+GL-7+Fas+ GC B cells from the indicated strains. (B) The percentages and absolute numbers of GC B cells are shown as bar graphs (n = 4–6 per group, two independent experiments). (C) Representative FACS plot of CD19+GL-7+PNA+ GC B cells from BXD2-p192/2 mice after 5 d of PBS, AdLacZ, or AdIL-23 administration. (D) The percentage and absolute + + + number of GC B cells are shown as bar graphs. (E) Left, Representative FACS plot of IgM versus IgD among CD19 GL-7 PNA GC B cells of BXD2- Downloaded from p192/2 mice 5 d after receiving AdLacZ or AdIL-23. Right, The percentage of IgM+/IgD+ and IgM2/IgD2 cells among GC B cells are shown as a bar graph (n = 4–6 per group). Results are mean 6 SD. Statistical differences were determined by ANOVA test and Tukey multiple comparisons test for signiﬁcance between groups (B) or two-tailed Student t test (D and E). *p , 0.05, **p , 0.01, ***p , 0.005.

IL-23 promotes the induction of DNA excision repair genes in others (54–56). In WT and BXD2-p192/2 mice, Il17ra gene ex- + BXD2 Ighg2b GC B cells pression was synchronized with the downregulation of Ighm, Ighd, http://www.jimmunol.org/ Because lower Ighg2b and IgG2b is a signature of BXD2-p192/2 and Il4ra. This was also a stage during which there was initially a mice, we took advantage of scRNA-seq transcriptomics analysis slow but progressive upregulation of Ighg1, Ighg2b, and Ighg2c and determined the top 100 downregulated genes in Ighg2b+ GC (Fig. 6A, 6B) and with the expression of the DZ chemokine re- B cells of BXD2-p192/2 versus WT mice (Fig. 5A). Pathway ceptor gene Cxcr4 (Fig. 6A, 6C). Consistent with our previous analysis conﬁrmed that cell cycle genes (Ccna2 and Ccnb2), ﬁndings (12–15), Il17ra was also synchronized with GC program systemic lupus erythematosus autoantibody-targeted histone var- signature genes including Aicda, Bcl6,andRgs13 (Fig. 6A, 6C, iant genes (H1fx, H2afv, H2afx, H2afz, H3f3a, and H3f3b), and 6D). Ifngr1 and Il21r exhibited steady but low levels from the Ighm+ and Ighd+ early preswitch phase to the midphase of GC

DNA base excision repair genes (Pold4 and Neil1) are among the by guest on October 2, 2021 top downregulated pathways in Ighg2b+ GC B cells derived from B cell development. A late-phase increase of Ifngr1 and Il21r was BXD2-p192/2 mice (Fig. 5B, 5C). Violin plot analysis further correlated with a further elevation of the Ighg genes, the plasma cell demonstrated that histone variant genes (H1fx, H2afv, H2afx, program genes Prdm1 and Id2, and the LZ chemokine receptor gene H2afz, and H3f3a), DNA excision repair genes (Pold4 and Neil1), Cxcr5 (Fig. 6A, 6C, 6E). calmodulin-related and cell cycle genes (Calm1 and Calm2), and Although cytokine receptor genes (Il17ra, Ifngr1, and Il21r), Rgs13 were all signiﬁcantly downregulated in Ighg2b+ GC B cells chemokine receptor genes, GC signature genes, and plasma cell in BXD2-p192/2 mice versus WT BXD2 mice (Fig. 5D). signature genes showed a similar pseudotime expression kinetics in WT BXD2 and BXD2-p192/2 GC B cells (Fig. 6A, 6C), the Il17ra synchronized with the expression of GC program genes lower expression of Ighg2b in BXD2-p192/2 GC B cells became in spontaneous autoimmune GCs of BXD2 mice apparent at the mid–GC phase and continued through the late To further analyze how different Tfh cytokine responses can be plasma phase of development (Fig. 6A, arrow, Fig. 6B). Also, in 2 2 associated with the induction of Ighg2b in GC B cells, we have BXD2-p19 / GC B cells, Il4ra expression kinetics were shifted applied a pseudotime trajectory analysis of single-cell tran- to the late phase of GC B cell development and were synchronized scriptomes of GC B cells derived from WT and BXD2-p192/2 with the expression of plasma cell program genes (Fig. 6A, 6E). mice (Fig. 6A). The GC B cell developmental stages (pseudotime Il17ra synchronized with the expression of histone variant 0–3.5) were ordered as early (0–1), mid (1–3), and late phase (.3) genes and DNA excision repair genes at the GC program phase by pseudotime analysis based on the developmental stages, which were based on the downregulation of Ighm and Ighd and the up- The above results suggest that Il17ra is the only Tfh cytokine receptor regulation of Ighg BCR genes (Fig. 6A, 6B). This organization gene that synchronized with the GC B cells during the Aicda+ DZ enabled us to determine GC B cell gene expression from the earliest phase of GC B cell development. We therefore determined if the (Ighm+Ighd+) to the latest plasma cell phase (Ighg+Prdm1+)ofde- major downregulated pathway genes (histone variants, cell cycle velopment. The GC B cell location was ordered based on the ex- genes, and excision repair genes) in Ighg2b+ GC B cells from BXD2- 2 2 pression of the DZ chemokine receptor gene Cxcr4 and the LZ p19 / mice also exhibited the same expression kinetics as Il17ra chemokine receptor gene Cxcr5 (Fig. 6A, 6C). genes. Indeed, the Monocle-based pseudotime analysis showed that The pseudotime expression of Tfh cytokine receptor genes, GC all major histone variant genes including H1fx, H2afv, H2afx, H2afz, B cell program signature genes, and plasma B cell program signature and H3f3a, calmodulin-related and cell cycle genes including Calm1, genes was then analyzed. Il4ra expression was synchronized with Calm2, Cdca3,andCcna2 genes, and DNA excision repair genes the expression of Ighm and Ighd in the early phase of GC B cell Pold4 and Neil1 were expressed at the Aicda+/Cxcr4+ GC DZ phase development in BXD2 mice (Fig. 6A, 6C), a result consistent with of development in GC B cells derived from both WT and BXD2- transcriptomic analysis of human naive IgD+ B cells as reported by p192/2 mice (Fig. 7A). The Journal of Immunology 7 Downloaded from http://www.jimmunol.org/ by guest on October 2, 2021

FIGURE 4. Decreased expression of IgG2b in the spleens of BXD2-p192/2 mice. GC B cells isolated from the spleen of BXD2 and BXD2-p192/2 mice were subjected to droplet-based scRNA-seq analysis. (A) Seurat heatmap visualization of the 50 most downregulated versus 50 top upregulated genes in B cells of BXD2 versus BXD2-p192/2 mice. Representative genes are shown on the left of the graph (n = 2–3 per group). (B) ELISA analysis of Ig isotype levels in sera of B6, BXD2, and BXD2-p192/2 mice shown as bar graphs. Statistical differences were determined by ANOVA test and Tukey multiple comparisons test for significance between groups (n = 11–26 mice per group). ***p , 0.005, ****p , 0.001. (C) A representative confocal image from the indicated strains stained with anti-IgM (red), anti-IgG2b (green), and PNA (blue). Upper, Objective lens, original magnification 34 power view; Lower, Digitally magnified views of the indicated areas in the top panels. (D) ImageJ quantitation of the fluorescent intensity of IgG2b in PALS. The mean (6SD) fluorescent intensity is shown in the lower panel. Statistical differences were determined by ANOVA test and Tukey multiple comparisons test for significance between groups. At least four representative areas from three to four spleens were quantitated. ***p , 0.005, ****p , 0.001. 8 IL-23 PROMOTES A COORDINATED B CELL GC PROGRAM Downloaded from http://www.jimmunol.org/

FIGURE 5. IL-23 promotes expression of DNA excision repair genes in Ighg2b+ GC B cells in BXD2 mice. GC B cells isolated from the spleen of BXD2 and BXD2-p192/2 mice were subjected to a droplet-based scRNA-seq analysis. (A) Heatmap visualization of the 100 top signiﬁcantly upregulated genes in Ighg2b+ B cells from BXD2 versus BXD2-p192/2 mice. Representative genes are shown to the right of the graph. (B) KEGG pathway analysis of prominent nodes for top downregulated genes in Ighg2b+ GC B cells from BXD2-p192/2 mice. The size of the node is shown as the number of genes associated with each pathway. (C) KEGG pathway enrichment analysis showing prominent pathways for the top downregulated genes in Ighg2b+ GC B 2/2 + 2/2 cells of BXD2-p19 mice. (D) Violin plots showing the indicated gene’s expression in Ighg2b B cells from BXD2 versus BXD2-p19 mice. by guest on October 2, 2021 The p value is shown in each panel. The statistical analysis was carried out using the “stat_compare_means” function in ggpubr package (B–D)(n = 2–3 mice per group).

To determine if Il17ra+ GC B cells indeed had higher transcript especially applicable during the CXCR4+ DZ stage of GC B cell levels of these GC B cell genes that may be involved in the DZ development in which GC B cells undergo extensive proliferation, cell cycle and AID-mediated excision repair, a SeqGeq gating AID induction, and a series of AID-mediated effector functions analysis was used to identify GC B cells that expressed only needed for B cell affinity maturation (57, 58). We previously Il17ra, Ifngr1, Il21r,orIl4r (Fig. 7B, 7C, upper panels). There showed that the IL-17RA signal is important in mediating mi- was a higher percentage of Rgs13+Aicda+, H1fx+H2afx+, gration arrest to both CXCL12 and CXCL13 (5, 12–15). We also Calm1+Calm2+, and Pold4+Neil1+ cells among the Il17ra+ GC showed that IL-17 and IL-21 were expressed by distinct Tfh cell B cells compared with Ifngr1+, Il21r+, and Il4r+ GC B cells in both subpopulations, with IL-17 predominating in the CXCR52ICOS+ WT BXD2 and BXD2-p192/2 mice (Fig. 7B, 7C, lower panels). subpopulation, whereas IL-21 expression was confined to the Among these, the distribution of Rgs13+Aicda+, Calm1+Calm2+, CXCR5+ subpopulation (5). The present study further suggests and Pold4+Neil1+ cells in Il17ra+ GC B cells of BXD2-p192/2 mice that IL-23–mediated induction of Tfh–IL-17 cells plays an es- were significantly lower than that in Il17ra+ GCBcellsofWT sential role in shaping the autoantibody formation response BXD2 mice (Fig. 7B–D). These results also showed that other through its effects on promoting CSR leading to IgG2b expression cytokines produced by Tfh cells cannot compensate for the loss during the DZ Aicda+ GC program stage of B cell maturation of IL-17RA signals in BXD2-p192/2 mice to enable effective lupus-prone BXD2 mice. The lack of such signaling in BXD2- expression of these gene programs that are needed for DNA p192/2 mice is associated with diminished autoantibody CSR to repair associated with AID-mediated CSR. the IgG2b but not the IgG1 isotype. Surprisingly, the diminished autoantibody production in BXD2- Discussion p192/2 mice was only partially identical to the effects of IL-17RA Ab-forming B cell development in the GC is a complex process deficiency on the spontaneous GC responses. We previously involving B cell commitment, migration, migration arrest, cell showed that IL-17 was important to upregulate RGS13 to enable cycle entry, AID induction, AID-induced DNA strand breaks, and GC B cell migration arrest and close interaction with GC T cells excision repair, leading to differentiation of GC B cells into Ab- (5, 12–15). Thus, IL-17 acts as a “glue” to hold the GCs together secreting plasma cells (57). The classical CXCR5+ LZ–oriented as well as to promote efficient T–B cell interactions, leading to IL-21–producing Tfh cell model cannot fully explain the orches- increased AID expression and the resultant increased CSR and tration of Tfh cells in directing all of these processes. This is SHM (5, 12). In the absence of IL-17RA, efficient GC T–B cell The Journal of Immunology 9 Downloaded from http://www.jimmunol.org/ by guest on October 2, 2021

FIGURE 6. Expression of Il17ra during the Cxcr4+ DZ Aicda+ phase of GC development. GC B cells isolated from the spleen of BXD2 and BXD2- p192/2 mice were subjected to droplet-based scRNA-seq analysis. Transcriptome trajectory analysis was performed using the Monocle 2 platform. (A) Monocle heatmap of normalized response curves for the representative genes in each indicated cluster across pseudotime in B cells–derived BXD2 (left) or BXD2-p192/2 (right) mice. Pseudotime trajectory was ordered by the expression of Ighm, Ighd, Ighg1, Ighg2b (indicated with a red arrow), and Ighg2c genes. Pseudotime 0–1 represents a status of GC B cells expressing the highest levels of both Ighm and Ighd genes. Pseudotime 1–2 represents a status when Ighm and Ighd gene expression is downregulated. Pseudotime 2–3 represents a status when the expression of Ighg genes is upregulated. Pseudotime .3 represents a status of GC B cells expressing the highest levels of all Ighg genes. (B–E) Pseudotime kinetics (solid line) of the indicated gene signature from the root (Ighm+Ighd+) of the trajectory to the outcome (Ighg+). Cells are color coded into different segments of the trajectory tree based on the expression patterns of Ighm, Ighd, and Ighg genes (n = 2–3 mice per group). interactions do not occur during GC development, and the de- and numbers of GCs in the absence of IL-23. These factors, how- velopment of GC cells and Ig CSR is abrogated. In the current ever, are not affected by the absence of IL-17RA. study, we found that although the expression of Rgs13 in GC Analysis of the lineages of Tfh cells provided insights into the B cells was reduced and there was a significant reduction in the differential effects of IL-23 p19 deficiency versus IL-17RA defi- percentages and numbers of IL-17–expressing Tfh cells in BXD2- ciency on GC and autoantibody development. We have found that the p192/2 mice, the percentages and numbers of GC B cells were absence of IL-23 diminished the percentages of Tfh–IL-17 cells and comparable between WT and BXD2-p192/2 mice. Also, the ad- dramatically skewed Tfh cells into Tfh–IFN-g cells, whereas the ministration of AdIL-23 to B6 mice did not enhance the size of GC absence of IL-17RA did not promote Tfh–IFN-g cell polarization. B cells in both B6 and BXD2-p192/2 mice. These results, therefore, One possible explanation for the enhanced Tfh–IFN-g in BXD2- suggest that there are additional factors that can maintain the size p192/2 mice is that in the absence of the p19 component, the 10 IL-23 PROMOTES A COORDINATED B CELL GC PROGRAM Downloaded from http://www.jimmunol.org/ by guest on October 2, 2021

FIGURE 7. Il17ra synchronized with the expression of histone variants and DNA excision repair genes at the GC program phase. GC B cells isolated from the spleen of BXD2 and BXD2-p192/2 mice were subjected to droplet-based scRNA-seq analysis. Transcriptome trajectory analysis was performed using the Monocle 2 platform. (A) Monocle heatmap of normalized response curves for the representative genes in each indicated cluster across pseudotime in B cells–derived BXD2 (left) or BXD2-p192/2 (right) mice. Pseudotime was defined as described in the legend of Fig. 6A. (B and C) Top, SeqGeq gating of GC B cells that expressed only Il17ra, Ifngr1, Il21r,orIl4r in BXD2 (B) and BXD2-p192/2 mice (C). Bottom, The percentage of the gated Rgs13+Aicda+, H1fx+H2afx+, Calm1+Calm2+, and Pold4+Neil1+ cells among Il17ra+, Ifngr1+, Il21r+,orIl4r+ GC B cells in BXD2 (B) and BXD2-p192/2 mice (C) are shown. In all panels, the scales were plotted using the arcsinh-transformed method implemented in the SeqGeq analysis module. Differences of the distribution of each gene pair in Il17ra+ B cells from WT BXD2 compared with the same gene pair in all other groups were determined using a Chi- Square test (n = 2–3 mice per group). *p , 0.05, **p , 0.01, ***p , 0.005, ****p , 0.001 by Fisher exact test. (D) Bar graph showing the numbers of cells that are either double positive or else with the indicated gene pairs in Il17ra+ versus Ifngr1+ GC B cells from WT BXD2 and BXD2-p192/2mice. Differences between groups were determined using a Chi-Square test. *p , 0.05, **p , 0.01, ***p , 0.005, ****p , 0.001 by Fisher exact test. The Journal of Immunology 11 dynamics of IL-12 p40/p35 heterodimer or p40/p40 homodimer B cells of BXD2-p192/2 mice. Additional genes that were iden- may be altered. Constitutive higher levels of p40 expression has tified included Pold4, Neil1, Calm1, Calm2, and Xrcc1. Although been shown to be critical for the IL-12 p40/p35 production (59). the role of POLD4 in immunity has not been identified previously, IL-12 p40/p40 homodimer, however, has been shown as an an- a recent study identified that two patients with biallelic mutations tagonist of IL-12 p40/p35 function and blocks IFN-g production affecting the POLD1 and POLD2 subunits of polymerase d (60). As IFN-g was enhanced in Tfh cells in BXD2-p192/2 mice, exhibited developmental defects and combined immunodeficiency it is unlikely that there was an excessive IL-12 p40/p40 produc- (32). Further, NEIL1, a human H2TH family glycosylase, and tion. Although we do not rule out the possibility that deletion of Calm1 have been implicated in the repair of lesions in transcription p19 potentially enhances IL-12 p40/p35 production, resulting in a bubbles (68) and histone H2AX phosphorylation (69), respectively. shift toward the IFN-g pathway, the administration of excessive The present scRNA-seq data therefore suggest a novel group of IL-23 using the Ad system leads to the induction of Th17 cytokine complex factors are at play in AID-mediated CSR. Importantly, the genes, but it also downregulated the expression of Th1, Th2, and current study also suggests that the T cell factor that facilitates these Tfh cytokine and transcription factor genes. These results there- processes is IL-17. We propose that such IL-17–producing cells be fore suggest that IL-23 most likely influences the Tfh cell polar- considered as T–GC helper cells, an underappreciated GC T cell ization program to regulate the Tfh cell lineage development. subpopulation that facilitates the selection and maturation of cen- The excessive skewing of the Tfh cells into other non–Tfh–IL-17 troblasts and can be distinguished from the well-established Tfh lineages of cells observed in the BXD2-p192/2 mice has provided cells that operate in the LZ to regulate a GC response. the needed signals to initiate and maintain the size of the spon- Interestingly, the effect of IL-23 deficiency is unique to the de- taneous GCs. They also suggest that there are unique IL-17–me- velopment of IgG2b-producing B cells. Because IgG1 transcripts and diated B cell maturation responses in GC B cells that cannot be circulating Ab levels were not affected in BXD2-p192/2 mice, these Downloaded from completely compensated for by other Tfh cytokines. Perturbation results suggest that IgG1 production either does not require IL-23 of IL-23 in the BXD2-p192/2 mouse provided an opportunity to and Tfh–IL-17 or was rescued by the elevation of other Tfh cyto- interrogate the pro–IL-17– versus pro–IFN-g–mediated GC B cell kines. This is consistent with an earlier model in which there is an developmental responses and signatures in an autoimmune mouse. hierarchy of cytokines for inducing isotype switching such that The LZ is more widely studied for centrocytes and T cell inter- IFN-g is dominant over IL-4 for the survival of IgG2a versus IgG1 actions because these GC B cells closely interact with the ca- and IgE switched GC B cells (70). In contrast, CSR to IgG2b has http://www.jimmunol.org/ nonical CXCR5+ Tfh cells for commitment to the plasma cell fate been shown to be TGF-b dependent but IFN-g and IL-4 indepen- (21, 61). However, our scRNA-seq differentially expressed gene dent (70, 71). This is also consistent with previous observations that and pseudotime results show that upregulation of Aicda is coor- a high dose of IL-17 enhanced IgG2b but not IgG1 CSR (72) and dinated with the upregulation of Bcl6, Cxcr4, and Pax5, which suggests that IL-17R signaling may be uniquely synchronized with have been shown to occur in the DZ (62). Interestingly, pseudo- AID and genes essential for NHEJ, leading to IgG2b formation. time analysis suggests that the Il17r expression in GC B cells Our results suggest a unique effect of IL-23 in orchestrating is synchronized with the expression of Aicda and other tran- IgG2b CSR in the BXD2 mouse model of lupus. The results scriptome signatures of DZ centroblasts, including Pax5, Bcl6, therefore underscore the necessity for a coordinated upregulation of by guest on October 2, 2021 Rgs13, Tcf3 (which encodes E2A), and Cxcr4. In contrast, Ifngr1 not only AID but also the DNA manipulation machinery that occurs and Il21r are best synchronized with the transcriptome signature in an environment that contains IL-23. The findings of the current of LZ centrocytes/PBs, including Prdm1, Irf4, Xbp1, Id2, and study are also relevant to the prominent Phase II clinical trial of Cxcr5. Pax5 directly promotes AID induction, and E2A has been ustekinumab (Stelara) for systemic lupus erythematosus patients shown to induce B cell CSR (63, 64). In contrast, Blimp1 and Id2 with active disease (73). Ustekinumab is a human mAb that binds repress AID and Pax5 and thereby suppress CSR (65, 66). The to the p40 subunit shared by both IL-12 and IL-23 and will not synchronized expression pattern of Il17ra with Aicda and other GC lead to a skewed B cell response to Tfh–IFN-g cells as demon- program genes suggest that Tfh–IL-17 cells may operate in the DZ to strated in the p19-deficient condition described in this study. facilitate AID and its downstream effector functions. This is con- Based on the results generated in this work, we anticipate that sistent with our previous results showing that IL-17 was expressed ustekinumab will diminish the IL-23–mediated Tfh–IL-17 cells predominantly in CXCR52ICOS+ CD4 T cells, whereas IL-21 was and their AID-mediated effector functions in the DZ and will also expressedinCXCR5+ICOS+ CD4 T cells in BXD2 mice (5). diminish the effects of IFN-g in assisting canonical Tfh cells in the Coexpression analysis further shows that Il17r+ GC B cells LZ. However, ustekinumab treatment also has been shown to in- expressed higher levels of genes related to genome instability duce subacute cutaneous lupus (74). It will be important to de- and DNA excision repair compared with Ifngr1+, Il4ra+, and termine if such effects are relevant to the undesirable skewing Il21r+ GC B cells. Although we did not find differences in toward Th2 responses by IL-12/IL-23 blockade. pseudotime gene expression kinetics between WT and p192/2 GC B cells, the pseudotime gene trajectory analysis has enabled Acknowledgments identification of novel DNA excision repair genes that might be We thank Drs. Michael Crowley and David K. Crossman at the UAB Heflin related to DZ B cell excision repair, leading to IgG2b class switch. Genomics Core Laboratory for carrying out the RNA sequencing assay and In the classical NHEJ excision repair, most studies thus far fo- assembling the Cell Ranger pipeline. cused on the role of H2afx and the ligase complex of DNA ligase IV, XRCC4, XRCC5 (Ku80), and XRCC6 (Ku70) (28, 67). Al- Disclosures though the current study does not negate the previous findings, the The authors have no financial conflicts of interest. scRNA-seq gene expression and trajectory analysis reveal that there are numerous novel factors that work separately during the DZ or the LZ phase to enable the CSR response. In the histone References variate family members, we have identified that the expression of 1. Crotty, S. 2014. T follicular helper cell differentiation, function, and roles in disease. Immunity 41: 529–542. H1fx, H2afv, H2afx, H2afz, and H3f3a were synchronized with the 2. Crotty, S. 2011. Follicular helper CD4 T cells (TFH). Annu. Rev. Immunol. 29: expression of Aicda, and these genes were downregulated in GC 621–663. 12 IL-23 PROMOTES A COORDINATED B CELL GC PROGRAM

3. Weinstein, J. S., E. I. Herman, B. Lainez, P. Licona-Limoń, E. Esplugues, 27. Lee, G. S., V. L. Brandt, and D. B. Roth. 2004. B cell development leads off with R. Flavell, and J. Craft. 2016. TFH cells progressively differentiate to regulate a base hit: dU:dG mismatches in class switching and hypermutation. Mol. Cell the germinal center response. Nat. Immunol. 17: 1197–1205. 16: 505–508. 4. Weinstein, J. S., B. J. Laidlaw, Y. Lu, J. K. Wang, V. P. Schulz, N. Li, 28. Jolly, C. J., A. J. Cook, and J. P. Manis. 2008. Fixing DNA breaks during class E. I. Herman, S. M. Kaech, P. G. Gallagher, and J. Craft. 2018. STAT4 and T-bet switch recombination. J. Exp. Med. 205: 509–513. control follicular helper T cell development in viral infections. [Published er- 29. Saribasak, H., R. W. Maul, Z. Cao, R. L. McClure, W. Yang, D. R. McNeill, ratum appears in 2018 J. Exp. Med. 215: 999.] J. Exp. Med. 215: 337–355. D. M. Wilson, III, and P. J. Gearhart. 2011. XRCC1 suppresses somatic hyper- 5. Ding, Y., J. Li, Q. Wu, P. Yang, B. Luo, S. Xie, K. M. Druey, A. J. Zajac, mutation and promotes alternative nonhomologous end joining in Igh genes. J. Exp. H. C. Hsu, and J. D. Mountz. 2013. IL-17RA is essential for optimal localization Med. 208: 2209–2216. of follicular Th cells in the germinal center light zone to promote autoantibody- 30. Boboila, C., C. Yan, D. R. Wesemann, M. Jankovic, J. H. Wang, J. Manis, producing B cells. J. Immunol. 191: 1614–1624. A. Nussenzweig, M. Nussenzweig, and F. W. Alt. 2010. Alternative end-joining 6. Laidlaw, B. J., Y. Lu, R. A. Amezquita, J. S. Weinstein, J. A. Vander Heiden, catalyzes class switch recombination in the absence of both Ku70 and DNA N. T. Gupta, S. H. Kleinstein, S. M. Kaech, and J. Craft. 2017. Interleukin-10 ligase 4. J. Exp. Med. 207: 417–427. from CD4+ follicular regulatory T cells promotes the germinal center response. 31. Yan, C. T., C. Boboila, E. K. Souza, S. Franco, T. R. Hickernell, M. Murphy, Sci. Immunol. 2: eaan4767. S. Gumaste, M. Geyer, A. A. Zarrin, J. P. Manis, et al. 2007. IgH class switching 7. Crotty, S. 2019. T follicular helper cell biology: a decade of discovery and and translocations use a robust non-classical end-joining pathway. Nature 449: diseases. Immunity 50: 1132–1148. 478–482. 8. Ding, Y., J. Li, P. Yang, B. Luo, Q. Wu, A. J. Zajac, O. Wildner, H. C. Hsu, and 32. Conde, C. D., O. Y. Petronczki, S. Baris, K. L. Willmann, E. Girardi, E. Salzer, J. D. Mountz. 2014. Interleukin-21 promotes germinal center reaction by S. Weitzer, R. C. Ardy, A. Krolo, H. Ijspeert, et al. 2019. Polymerase d defi- skewing the follicular regulatory T cell to follicular helper T cell balance in ciency causes syndromic immunodeficiency with replicative stress. J. Clin. In- autoimmune BXD2 mice. Arthritis Rheumatol. 66: 2601–2612. vest. 129: 4194–4206. 9. Morita, R., N. Schmitt, S. E. Bentebibel, R. Ranganathan, L. Bourdery, 33. Costantino, L., S. K. Sotiriou, J. K. Rantala, S. Magin, E. Mladenov, T. Helleday, G. Zurawski, E. Foucat, M. Dullaers, S. Oh, N. Sabzghabaei, et al. 2011. Human J. E. Haber, G. Iliakis, O. P. Kallioniemi, and T. D. Halazonetis. 2014. Break- blood CXCR5(+)CD4(+) T cells are counterparts of T follicular cells and contain induced replication repair of damaged forks induces genomic duplications in specific subsets that differentially support antibody secretion. [Published erratum human cells. Science 343: 88–91. 34. Minko, I. G., V. L. Vartanian, N. N. Tozaki, E. Coskun, S. H. Coskun, P. Jaruga, appears in 2011 Immunity 34: 135.] Immunity 34: 108–121. Downloaded from 10. Koutsakos, M., T. H. O. Nguyen, and K. Kedzierska. 2019. With a little help J. Yeo, S. S. David, M. P. Stone, M. Egli, et al. 2020. Recognition of DNA from T follicular helper friends: humoral immunity to influenza vaccination. adducts by edited and unedited forms of DNA glycosylase NEIL1. DNA Repair J. Immunol. 202: 360–367. (Amst.) 85: 102741. 11. Mountz, J. D., H. C. Hsu, and A. Ballesteros-Tato. 2019. Dysregulation of T 35. Dou, H., S. Mitra, and T. K. Hazra. 2003. Repair of oxidized bases in DNA follicular helper cells in lupus. J. Immunol. 202: 1649–1658. bubble structures by human DNA glycosylases NEIL1 and NEIL2. J. Biol. 12. Hsu, H. C., P. Yang, J. Wang, Q. Wu, R. Myers, J. Chen, J. Yi, T. Guentert, Chem. 278: 49679–49684. A. Tousson, A. L. Stanus, et al. 2008. Interleukin 17-producing T helper cells 36. Oppmann, B., R. Lesley, B. Blom, J. C. Timans, Y. Xu, B. Hunte, F. Vega, N. Yu, and interleukin 17 orchestrate autoreactive germinal center development in au- J. Wang, K. Singh, et al. 2000. Novel p19 protein engages IL-12p40 to form a http://www.jimmunol.org/ toimmune BXD2 mice. Nat. Immunol. 9: 166–175. cytokine, IL-23, with biological activities similar as well as distinct from IL-12. 13. Mountz, J. D., J. H. Wang, S. Xie, and H. C. Hsu. 2011. Cytokine regulation of Immunity 13: 715–725. B-cell migratory behavior favors formation of germinal centers in autoimmune 37. Hsu, H. C., T. Zhou, H. Kim, S. Barnes, P. Yang, Q. Wu, J. Zhou, B. A. Freeman, disease. Discov. Med. 11: 76–85. M. Luo, and J. D. Mountz. 2006. Production of a novel class of polyreactive 14. Wang, J. H., J. S. New, S. Xie, P. Yang, Q. Wu, J. Li, B. Luo, Y. Ding, pathogenic autoantibodies in BXD2 mice causes glomerulonephritis and ar- Arthritis Rheum. K. M. Druey, H. C. Hsu, and J. D. Mountz. 2013. Extension of the germinal thritis. 54: 343–355. 38. Mountz, J. D., P. Yang, Q. Wu, J. Zhou, A. Tousson, A. Fitzgerald, J. Allen, center stage of B cell development promotes autoantibodies in BXD2 mice. X. Wang, S. Cartner, W. E. Grizzle, et al. 2005. Genetic segregation of spon- Arthritis Rheum. 65: 2703–2712. taneous erosive arthritis and generalized autoimmune disease in the BXD2 15. Xie, S., J. Li, J. H. Wang, Q. Wu, P. Yang, H. C. Hsu, L. E. Smythies, and recombinant inbred strain of mice. Scand. J. Immunol. 61: 128–138. J. D. Mountz. 2010. IL-17 activates the canonical NF-kappaB signaling pathway 39. Cua, D. J., J. Sherlock, Y. Chen, C. A. Murphy, B. Joyce, B. Seymour, L. Lucian, in autoimmune B cells of BXD2 mice to upregulate the expression of regulators W. To, S. Kwan, T. Churakova, et al. 2003. Interleukin-23 rather than

of G-protein signaling 16. J. Immunol. 184: 2289–2296. by guest on October 2, 2021 interleukin-12 is the critical cytokine for autoimmune inflammation of the brain. 16. Jain, S., A. Stock, F. Macian, and C. Putterman. 2018. A distinct T follicular Nature 421: 744–748. helper cell subset infiltrates the brain in murine neuropsychiatric lupus. Front. 40. Murphy, C. A., C. L. Langrish, Y. Chen, W. Blumenschein, T. McClanahan, Immunol. 9: 487. R. A. Kastelein, J. D. Sedgwick, and D. J. Cua. 2003. Divergent pro- and 17. Cantaert, T., J. N. Schickel, J. M. Bannock, Y. S. Ng, C. Massad, T. Oe, R. Wu, antiinflammatory roles for IL-23 and IL-12 in joint autoimmune inflammation. A. Lavoie, J. E. Walter, L. D. Notarangelo, et al. 2015. Activation-induced cy- J. Exp. Med. 198: 1951–1957. tidine deaminase expression in human B cell precursors is essential for central 41. Reay, J., S. H. Kim, E. Lockhart, J. Kolls, and P. D. Robbins. 2009. Adenoviral- B cell tolerance. Immunity 43: 884–895. mediated, intratumor gene transfer of interleukin 23 induces a therapeutic an- 18. Victora, G. D., D. Dominguez-Sola, A. B. Holmes, S. Deroubaix, R. Dalla- titumor response. Cancer Gene Ther. 16: 776–785. Favera, and M. C. Nussenzweig. 2012. Identification of human germinal center 42. Schneider, C. A., W. S. Rasband, and K. W. Eliceiri. 2012. NIH image to imageJ: light and dark zone cells and their relationship to human B-cell lymphomas. 25 years of image analysis. Nat. Methods 9: 671–675. [Published erratum appears in 2015 Blood 126: 1262.] Blood 120: 2240–2248. 43. Wang, J. H., J. Li, Q. Wu, P. Yang, R. D. Pawar, S. Xie, L. Timares, C. Raman, 19. Kraütler, N. J., D. Suan, D. Butt, K. Bourne, J. R. Hermes, T. D. Chan, D. D. Chaplin, L. Lu, et al. 2010. Marginal zone precursor B cells as cellular C. Sundling, W. Kaplan, P. Schofield, J. Jackson, et al. 2017. Differentiation of agents for type I IFN-promoted antigen transport in autoimmunity. J. Immunol. germinal center B cells into plasma cells is initiated by high-affinity antigen and 184: 442–451. completed by Tfh cells. J. Exp. Med. 214: 1259–1267. 44. Dobin, A., C. A. Davis, F. Schlesinger, J. Drenkow, C. Zaleski, S. Jha, P. Batut, 20. Zhang, Y., L. Tech, L. A. George, A. Acs, R. E. Durrett, H. Hess, L. S. m. Chaisson, and T. R. Gingeras. 2013. STAR: ultrafast universal RNA-seq K. Walker, D. M. Tarlinton, A. L. Fletcher, A. E. Hauser, and K. M. Toellner. aligner. Bioinformatics 29: 15–21. 2018. Plasma cell output from germinal centers is regulated by signals from Tfh 45. Smith, T. F., and M. S. Waterman. 1981. Identification of common molecular and stromal cells. J. Exp. Med. 215: 1227–1243. subsequences. J. Mol. Biol. 147: 195–197. 21. Ise, W., K. Fujii, K. Shiroguchi, A. Ito, K. Kometani, K. Takeda, E. Kawakami, 46. Satija, R., J. A. Farrell, D. Gennert, A. F. Schier, and A. Regev. 2015. Spatial K. Yamashita, K. Suzuki, T. Okada, and T. Kurosaki. 2018. T follicular helper reconstruction of single-cell gene expression data. Nat. Biotechnol. 33: 495–502. cell-germinal center B cell interaction strength regulates entry into plasma cell or 47. Trapnell, C., D. Cacchiarelli, J. Grimsby, P. Pokharel, S. Li, M. Morse, recycling germinal center cell fate. Immunity 48: 702–715.e4. N. J. Lennon, K. J. Livak, T. S. Mikkelsen, and J. L. Rinn. 2014. The dynamics 22. Hsu, H. C., P. Yang, Q. Wu, J. H. Wang, G. Job, T. Guentert, J. Li, and regulators of cell fate decisions are revealed by pseudotemporal ordering of C. R. Stockard, T. V. Le, D. D. Chaplin, et al. 2011. Inhibition of the catalytic single cells. Nat. Biotechnol. 32: 381–386. function of activation-induced cytidine deaminase promotes apoptosis of ger- 48. Qiu, X., Q. Mao, Y. Tang, L. Wang, R. Chawla, H. A. Pliner, and C. Trapnell. minal center B cells in BXD2 mice. Arthritis Rheum. 63: 2038–2048. 2017. Reversed graph embedding resolves complex single-cell trajectories. Nat. 23. Hsu, H. C., Y. Wu, P. Yang, Q. Wu, G. Job, J. Chen, J. Wang, M. A. Accavitti- Methods 14: 979–982. Loper, W. E. Grizzle, R. H. Carter, and J. D. Mountz. 2007. Overexpression of 49. Yu, G., L. G. Wang, Y. Han, and Q. Y. He. 2012. clusterProfiler: an R package for activation-induced cytidine deaminase in B cells is associated with production of comparing biological themes among gene clusters. OMICS 16: 284–287. highly pathogenic autoantibodies. J. Immunol. 178: 5357–5365. 50. Parks, D. R., M. Roederer, and W. A. Moore. 2006. A new “Logicle” display 24. De Silva, N. S., and U. Klein. 2015. Dynamics of B cells in germinal centres. method avoids deceptive effects of logarithmic scaling for low signals and Nat. Rev. Immunol. 15: 137–148. compensated data. Cytometry A 69: 541–551. 25. Muramatsu, M., K. Kinoshita, S. Fagarasan, S. Yamada, Y. Shinkai, and 51. Finak, G., J. M. Perez, A. Weng, and R. Gottardo. 2010. Optimizing transfor- T. Honjo. 2000. Class switch recombination and hypermutation require mations for automated, high throughput analysis of flow cytometry data. BMC activation-induced cytidine deaminase (AID), a potential RNA editing enzyme. Bioinformatics 11: 546. Cell 102: 553–563. 52. Li, H., H. C. Hsu, Q. Wu, P. Yang, J. Li, B. Luo, M. Oukka, C. H. Steele, III, 26. Rada, C., J. M. Di Noia, and M. S. Neuberger. 2004. Mismatch recognition and uracil D. J. Cua, W. E. Grizzle, and J. D. Mountz. 2014. IL-23 promotes TCR-mediated excision provide complementary paths to both Ig switching and the A/T-focused negative selection of thymocytes through the upregulation of IL-23 receptor and phase of somatic mutation. Mol. Cell 16: 163–171. RORgt. Nat. Commun. 5: 4259. The Journal of Immunology 13

53. Shinall, S. M., M. Gonzalez-Fernandez, R. J. Noelle, and T. J. Waldschmidt. 63. Quong, M. W., D. P. Harris, S. L. Swain, and C. Murre. 1999. E2A activity is 2000. Identification of murine germinal center B cell subsets defined by the induced during B-cell activation to promote immunoglobulin class switch re- expression of surface isotypes and differentiation antigens. J. Immunol. 164: combination. EMBO J. 18: 6307–6318. 5729–5738. 64. Gonda, H., M. Sugai, Y. Nambu, T. Katakai, Y. Agata, K. J. Mori, Y. Yokota, and 54. Jenks, S. A., K. S. Cashman, E. Zumaquero, U. M. Marigorta, A. V. Patel, A. Shimizu. 2003. The balance between Pax5 and Id2 activities is the key to AID X. Wang, D. Tomar, M. C. Woodruff, Z. Simon, R. Bugrovsky, et al. 2018. gene expression. J. Exp. Med. 198: 1427–1437. Distinct effector B cells induced by unregulated toll-like receptor 7 contribute to 65. Lin, K. I., C. Angelin-Duclos, T. C. Kuo, and K. Calame. 2002. Blimp-1- pathogenic responses in systemic lupus erythematosus. [Published erratum ap- dependent repression of Pax-5 is required for differentiation of B cells to im- pears in 2020 Immunity 52: 203.] Immunity 49: 725–739.e6. munoglobulin M-secreting plasma cells. Mol. Cell. Biol. 22: 4771–4780. 55. Becker, A. M., K. H. Dao, B. K. Han, R. Kornu, S. Lakhanpal, A. B. Mobley, 66. Shaffer, A. L., K. I. Lin, T. C. Kuo, X. Yu, E. M. Hurt, A. Rosenwald, Q. Z. Li, Y. Lian, T. Wu, A. M. Reimold, et al. 2013. SLE peripheral blood J. M. Giltnane, L. Yang, H. Zhao, K. Calame, and L. M. Staudt. 2002. Blimp-1 B cell, T cell and myeloid cell transcriptomes display unique profiles and each orchestrates plasma cell differentiation by extinguishing the mature B cell gene subset contributes to the interferon signature. PLoS One 8: e67003. expression program. Immunity 17: 51–62. 56. Harris, D. P., S. Goodrich, K. Mohrs, M. Mohrs, and F. E. Lund. 2005. Cutting 67. Kotnis, A., L. Du, C. Liu, S. W. Popov, and Q. Pan-Hammarstro¨m. 2009. Non- edge: the development of IL-4-producing B cells (B effector 2 cells) is controlled homologous end joining in class switch recombination: the beginning of the end. Philos. Trans. R. Soc. Lond. B Biol. Sci. 364: 653–665. by IL-4, IL-4 receptor alpha, and Th2 cells. J. Immunol. 175: 7103–7107. 68. Makasheva, K. A., A. V. Endutkin, and D. O. Zharkov. 2020. Requirements for 57. Stewart, I., D. Radtke, B. Phillips, S. J. McGowan, and O. Bannard. 2018. DNA bubble structure for efficient cleavage by helix-two-turn-helix DNA gly- Germinal center B cells replace their antigen receptors in dark zones and fail cosylases. Mutagenesis 35: 119–128. light zone entry when immunoglobulin gene mutations are damaging. Immunity 69. Smallwood, H. S., D. Lopez-Ferrer, P. E. Eberlein, D. J. Watson, and 49: 477–489.e7. T. C. Squier. 2009. Calmodulin mediates DNA repair pathways involving H2AX 58. Bannard, O., and J. G. Cyster. 2017. Germinal centers: programmed for affinity in response to low-dose radiation exposure of RAW 264.7 macrophages. Chem. maturation and antibody diversification. Curr. Opin. Immunol. 45: 21–30. Res. Toxicol. 22: 460–470. 59. Jalah, R., M. Rosati, B. Ganneru, G. R. Pilkington, A. Valentin, V. Kulkarni, 70. Deenick, E. K., J. Hasbold, and P. D. Hodgkin. 2005. Decision criteria for re- C. Bergamaschi, B. Chowdhury, G. M. Zhang, R. K. Beach, et al. 2013. The p40 solving isotype switching conflicts by B cells. Eur. J. Immunol. 35: 2949–2955. subunit of interleukin (IL)-12 promotes stabilization and export of the p35 71. Vidarsson, G., G. Dekkers, and T. Rispens. 2014. IgG subclasses and allotypes: subunit: implications for improved IL-12 cytokine production. J. Biol. Chem. from structure to effector functions. Front. Immunol. 5: 520. Downloaded from 288: 6763–6776. 72. Shibui, A., E. Shimura, A. Nambu, S. Yamaguchi, W. J. Leonard, K. Okumura, 60. Gillessen, S., D. Carvajal, P. Ling, F. J. Podlaski, D. L. Stremlo, P. C. Familletti, S. Sugano, K. Sudo, and S. Nakae. 2012. Th17 cell-derived IL-17 is dispensable U. Gubler, D. H. Presky, A. S. Stern, and M. K. Gately. 1995. Mouse interleukin-12 for B cell antibody production. Cytokine 59: 108–114. (IL-12) p40 homodimer: a potent IL-12 antagonist. Eur. J. Immunol. 25: 200–206. 73.vanVollenhoven,R.F.,B.H.Hahn,G.C.Tsokos,P.Lipsky,K.Fei, 61. Victora, G. D., T. A. Schwickert, D. R. Fooksman, A. O. Kamphorst, M. Meyer- R.M.Gordon,I.Gregan,K.H.Lo,M.Chevrier,andS.Rose.2020.Main- Hermann, M. L. Dustin, and M. C. Nussenzweig. 2010. Germinal center dy- tenance of efficacy and safety of Ustekinumab through one year in a phase II namics revealed by multiphoton microscopy with a photoactivatable fluorescent multicenter, prospective, randomized, double-blind, placebo-controlled

reporter. Cell 143: 592–605. crossover trial of patients with active systemic lupus erythematosus. Arthritis http://www.jimmunol.org/ 62. Gitlin, A. D., Z. Shulman, and M. C. Nussenzweig. 2014. Clonal selection in Rheumatol. 72: 761–768. the germinal centre by regulated proliferation and hypermutation. Nature 509: 74. Tierney, E., S. Kirthi, B. Ramsay, and K. Ahmad. 2019. Ustekinumab-induced 637–640. subacute cutaneous lupus. JAAD Case Rep. 5: 271–273. by guest on October 2, 2021 Supplementary Materials

IL-23 promotes a coordinated B-cell germinal center program for class-switch recombination to IgG2b in BXD2 mice

Huixian Hong*, Min Gaoϯ, Qi Wu*, PingAr Yang*, Shanrun Liu*, Hao Li‡, Peter D.

Burrows§, Daniel Cua¶, Jake Y. Chenϯ, Hui-Chen Hsu* and John D. Mountz*,‖

* Division of Clinical Immunology and Rheumatology, Department of Medicine,

University of Alabama at Birmingham

ϯ Informatics Institute, the University of Alabama at Birmingham

‡ Beth Israel Deaconess Medical Center, Harvard Medical School,

§ Department of Microbiology, University of Alabama at Birmingham

¶ Discovery Research, Merck Research Laboratory; Current institution: the Janssen

Pharmaceutical Companies of Johnson & Johnson

‖ Department of Medicine, Birmingham VA Medical center

*Address correspondence to: John D. Mountz, MD., PhD., Dept. of Medicine, Division of

Rheumatology, Shelby Interdisciplinary Biomed Research Bldg., Room SHEL 307, 1825

University Blvd, Birmingham, AL 35294-2182; Email: [email protected]; Phone: 205-934-

8909

Running Title: IL-23 promotes a coordinated B-cell germinal center program Supplemental Table 1 – Primers used in qRT-PCR analysis

Name Sequence Murine Gapdh F AGGTCGGTGTGAACGGATTTG R TGTAGACCATGTAGTTGAGGTCA Murine Il17a F TTTAACTCCCTTGGCGCAAAA R CTTTCCCTCCGCATTGACAC Murine Ifng F ATGAACGCTACACACTGCATC R CCATCCTTTTGCCAGTTCCTC Murine Il21 F GGACCCTTGTCTGTCTGGTAG R TGTGGAGCTGATAGAAGTTCAGG Murine Il17f F TGCTACTGTTGATGTTGGGAC R AATGCCCTGGTTTTGGTTGAA Murine Il21 F GGACCCTTGTCTGTCTGGTAG R TGTGGAGCTGATAGAAGTTCAGG Murine Il4 F GGTCTCAACCCCCAGCTAGT R GCCGATGATCTCTCTCAAGTGAT Murine Bcl6 F CCGGCACGCTAGTGATGTT R TGTCTTATGGGCTCTAAACTGCT Murine Tbet F AACCGCTTATATGTCCACCCA R CTTGTTGTTGGTGAGCTTTAGC Murine Gata3 F CTCGGCCATTCGTACATGGAA R GGATACCTCTGCACCGTAGC Murine Il23r F TTCAGATGGGCATGAATGTTTCT R CCAAATCCGAGCTGTTGTTCTAT