Mice were bred on the C57BL/6JRj background and maintained under specific pathogen-free conditions at the animal facility of Ulm University. The Pou2af1^fl/fl mice (C57BL/6JRj background) were crossed to CD4-Cre (32) or IL21-Cre mice (33). Animal experimentation was done in accordance with German legislation for animal protection (reference number: 35/9185.81-3/1325). Mice were maintained in a specific pathogen-free facility under standard housing conditions with continuous access to food and water. Mice used in the study were 8-14-weeks old and were maintained on 12 h light, 12 h dark light cycle (6:00-18:00). To test T-dependent antibody responses and Germinal Center reaction, mice were immunized with 10⁸ Sheep Red Blood Cells (SRBCs) (Cedarlane Laboratories, Canada) in PBS by i.p. injection. To analyze antibody production during a secondary immune response, mice were immunized on day 0 and again on day 14. Blood was collected on days 0, 14 and 21. To assess antigen-specific antibody levels, mice were immunized with 100 µg NP-KLH (Biosearch Technologies) emulsified in Imject Alum (Thermo Fisher Scientific) by i.p injection. Serum was analyzed for the presence of antigen-specific immunoglobulins after 28 days.

The strategy for the generation of a conditional Pou2af1 allele is described elsewhere (34).

To examine the expression of BOB.1/OBF.1 in CD4⁺ T cells in the respective mouse strains, T cells were enriched by a magnetic column. Appropriate kits were used to isolate CD4⁺ T cells (Milteny Biotec). TFH cells were sorted on a BD FACS Aria cell sorter (BD Biosciences) based on their expression of CD3, CD4, CXCR5 and PD-1. Due to limited number of sorted TFH cells, cells were pooled from two or three mice of respective genotypes. Isolated or sorted cells were lysed in RIPA buffer and immunoblotting was performed following standard procedures using the following antibodies: anti-BOB.1/OBF.1 (self-made) and anti-β-actin (Sigma-Aldrich, Missouri, USA). Horseradish-peroxidase conjugated rabbit anti-mouse antibody was used as a secondary antibody. Membranes were developed with a chemiluminescence detection system (Bio-Rad, California, USA).

The cell staining procedure was described earlier (34). Data were acquired on a Gallios cytometer (Beckman Coulter, California, USA) and analyzed with Kaluza Analysis software, version 2.1 (Beckman Coulter, California, USA). Used antibodies are summarized in Table S1 . Intracellular Foxp3 staining was performed using the Treg detection kit (Milteny Biotec).

Paraffin-embedded sections of spleens from immunized mice were stained with biotinylated Peanut Agglutinin (PNA) (Vector Laboratories, California, USA). This was followed by detection with horseradish-peroxidase conjugated streptavidin and AEC chromogen. Subsequently, slides were counter stained with hematoxylin.

Immunoglobulin ELISA was performed as previously described (34). For determination of NP-specific antibody levels in sera from 28 d NP-KLH immunized mice ELISA plates were coated overnight at 4°C with 10 µg/ml NP(9)- and NP(27)-BSA. The next day, plates were blocked with 5% BSA in PBS overnight at 4°C. Subsequently, plates were washed, and serum samples were serially diluted in duplicate and incubated overnight at 4°C. Plates were then washed and incubated with alkaline phosphatase-conjugated goat anti-mouse IgM, IgG1 or IgG3 (SouthernBiotech, USA) diluted 1:1000 for 1h at 37°C. 4-nitrophenil phosphate-disodium salt (Serva, Germany) was used as substrate. Plates were read in a plate reader at 450 nm. Values were reported as relative absorbance.

CD4⁺ T cells were isolated from LNs of CD4-Cre (n=4) and Pou2af1^fl/fl x CD4-Cre (n=9) mice after immunization with SRBCs by magnetic separation using CD4⁺ T cell isolation kit (Milteny Biotec). Purity of isolated CD4⁺ T cells was ≥ 96% for all samples and is shown in Figure S14 . Extraction and purification of RNA were performed using RNeasy Mini Kit (Qiagen). RNA was eluted in RNase-free water and stored at -80°C. RNA quantity was measured using the Infinite 200 Pro (Tecan).

Quality was checked using an Agilent Tape Station and all samples displayed RIN values above 9. After quantity measurement by Qubit Fluorometer (Invitrogen) 200ng of RNA were used to generate sequencing libraries using the Illumina RNA-Seq Kit V2. Sequencing libraries were subsequently quality checked on D1000 screentapes (Agilent) and 10 samples each were sequenced using a NextSeq550 (Illumina) and 2 NextSeq 500/550 High Output Kits v2.5 (75 Cycles) at the Genomics Core Facility (Ulm University).

Hiqh-quality reads were mapped to the mouse genome (GRChm38) using STAR (2.0.9). Followed by the removal of multimapping reads, converted to gene specific read counts for annotated genes using featureCounts (2.0.0).

Data analysis was performed in R (4.0.2) using the IDE Rstudio (1.3.1056). Differential expression analysis was performed using deseq2 (1.29.16), and the shrinkage estimator used was “apeglm” (35). For computation of distance matricies, euclidian distance was used. Generally, ggplot2 (3.3.2) was used for data visualization. The 99^th percentile of the log2foldchanges was calculated ( Figure S15 ) and applied as a cutoff alongside an FDR cutoff of 0.05. This cutoff corresponds to an absolute foldchange of 1.1327. Clustering in the heatmaps was done using hierarchical clustering within the pheatmap (1.0.12) package. Gene ontology and visualization of the results was performed using clusterProfiler (3.18.0) and enrichplot (1.11.0.991). UMAP dimensionallity reduction has been performed using the package uwot (0.1.8), with the default parameters (except: n_neighbors = n/2). Waterfall plots were created using R (4.1.1) and the ggplot2 (3.3.5) package.

Results are expressed as arithmetic means ± standard deviation. Statistical analysis was performed with a two-tailed unpaired Student’s t-test or Mann-Whitney-U test as indicated, with GraphPad Prism V8 software. * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001.

To address the role of BOB.1/OBF.1 in T lymphocytes, we generated floxed Pou2af1 alleles, as previously described (34). To assess the T lymphoid function of BOB.1/OBF.1, we used the CD4- (32) and IL21-Cre (33) mice. CD4-Cre was chosen to specifically delete BOB.1/OBF.1 in double positive and mature T cells. IL21-Cre mice were used to address the role of BOB.1/OBF.1 predominantly in TFH cells, key regulators of the GC response. IL21 is primarily produced by TFH cells, but a production from TH17 and natural killer T cells has also been reported (36, 37). Since IL21 is also required for TFH cell differentiation, it is already upregulated in developing pre-TFH cells (38). Thus, crossing Pou2af1^fl/fl mice to IL21-Cre mice results in BOB.1/OBF.1 deficiency already in pre-TFH cells and not specifically in fully developed TFH cells. The Pou2af1 mRNA contains two translation initiation codons leading to the expression of two different isoforms – p40 and p34 (39). The transient p40-isoform serves as a precursor of a third p35 isoform, which is why predominantly the p34 and p35 isoform are detectable by immunoblot analysis. Conditional deletion of both p34 and p35 isoforms of BOB.1/OBF.1 in CD4⁺ T and TFH cells was confirmed by immunoblot analysis ( Figures 1A, B ). BOB.1/OBF.1 is completely absent in CD4⁺ T cells isolated from Pou2af1^fl/fl x CD4-Cre mice and CXCR5⁺ PD-1⁺ TFH cells isolated from Pou2af1^fl/fl x IL21-Cre mice. For all analyses, we compared the conditional knockouts to their respective Cre- and floxed-controls, ruling out any Cre- or floxed-dependent effects. Both Cre- and floxed-genes are functionally WT. For comparison, we included WT C57BL/6JRj and conventional Pou2af1 knockout mice generated by Nielsen et al., in which BOB.1/OBF.1 is deleted in all cell types permanently/stably (14). Both conditional knockouts are generally healthy and fertile. CD4-Cre or IL21-Cre expression itself have no effect on thymic T cell development (40) ( Figure 1C ).

First, we investigated the effect of conditional BOB.1/OBF.1 ablation on the T cell compartment in the periphery. Beforehand, numbers of splenocytes of both Pou2af1^fl/fl crossed either to CD4- or IL21-Cre mice were comparable to the WT situation ( Figure 2A ). Similarly, the overall number of CD3⁺ T cells was unaffected in both conditional knockouts in spleen and LNs ( Figure 2B ; Figure S1A ). However, the number of CD4⁺ T cells was significantly reduced in Pou2af1^fl/fl x CD4-Cre mice in spleen and LN ( Figure 2C ; Figure S1B ). The reduction of splenic CD4⁺ T cells observed in Pou2af1^fl/fl x CD4-Cre was comparable to conventional BOB.1/OBF.1-deficient animals. On the other hand, CD8⁺ T cell numbers were significantly increased in Pou2af1^fl/fl x CD4-Cre mice both in spleen and LN ( Figure 2D ; Figure S1C ). Numbers of CD4⁺ and CD8⁺ T cells in Pou2af1^fl/fl x Il21-Cre mice were comparable to the WT situation ( Figures 2C, D ; Figures S1B, C ). In accordance with cell numbers, Pou2af1^fl/fl x CD4-Cre mice featured markedly reduced percentages of CD4⁺ T cells, while at the same time increased percentages of CD8⁺ T cells resulting in a significantly altered CD4/CD8 T cell ratio in comparison to controls both in spleen and LN ( Figure 2E ; Figure S1D ). The CD4/CD8 ratio in the spleen and LN of Pou2af1^fl/fl x Il21-Cre mice was unaffected ( Figure 2E ; Figure S1D ). B cell numbers in both conditional knockouts were comparable to the WT situation in spleen and LN ( Figure 2F ; Figure S1E ). These results suggest a role for BOB.1/OBF.1 in CD4⁺ T cell biology.

To examine the contribution of T cell-specific BOB.1/OBF.1 expression for the GC reaction, we immunized Pou2af1^fl/fl mice crossed to either CD4- or IL21-Cre mice with Sheep Red Blood Cells (SRBCs). Flow-cytometric analyses of spleen and LN revealed significantly reduced levels of GL7⁺ CD95⁺ GC B cells after both CD4⁺ T and TFH cell-specific BOB.1/OBF.1 deletion ( Figure 3A ). GC B cell numbers were reduced by 60% in Pou2af1^fl/fl x CD4-Cre and by 35% in Pou2af1^fl/fl x IL21-Cre animals. Histological analysis of spleens confirmed the distinct reduction in the number of PNA⁺ GCs and revealed a significant decrease in GC size for Pou2af1^fl/fl x CD4-Cre mice ( Figure 3B ). In case of Pou2af1^fl/fl x IL21-Cre mice, no difference in the overall number of PNA⁺ stained GCs compared to controls was observed, but still the GC size was significantly reduced in the TFH cell-specific conditional knockout, which probably accounts for the diminished GC B cell numbers assessed by flow cytometry ( Figure 3B ). These findings demonstrate that the absence of GCs observed in conventional BOB.1/OBF.1-defcient mice cannot exclusively be traced back to the B cell compartment and defective B cell maturation. Our results provide evidence that specific deficiency of BOB.1/OBF.1 in T cells, while the B cell compartment stays immunocompetent, also impairs GC formation.

We and others have recently shown the impact of BOB.1/OBF.1 for TFH cell development and function (22, 23). TFH cells provide help for B cells regulating their proliferation and immunoglobulin class switching and are therefore relevant key player during GC reaction. TFH cells are characterized by the expression of the chemokine receptor CXCR5 together with at least one of the co-receptors PD1, ICOS or BTLA4 (22, 41). After immunization with SRBCs, both Pou2af1^fl/fl mice crossed to CD4- and IL21-Cre mice revealed significantly reduced levels of CXCR5^hi PD1⁺, CXCR5^hi ICOS⁺ and CXCR5^hi BTLA4⁺ TFH cells in the spleen and LN in comparison to respective controls ( Figures 4A–D ; Figure S2 ). In detail, in Pou2af1^fl/fl x CD4-Cre mice the frequency of CXCR5^hi PD1⁺ cells was reduced by 55%, of CXCR5^hi ICOS⁺ cells by 50% and of CXCR5^hi BTLA4⁺ cells by 30% in the spleen ( Figures 4B–D ). For Pou2af1^fl/fl x IL21-Cre mice a reduction by 40% of both CXCR5^hi PD1⁺ and CXCR5^hi ICOS⁺ cells was observed, while CXCR5^hi BTLA4⁺ cells were decreased by 35% ( Figures 4B–D ). Besides TFH cell numbers, also their percentage of CD4⁺ T cells was markedly reduced in both conditional knockouts in spleen and LNs ( Figure 4 ; Figure S2 ). Together, these data reveal a requirement for BOB.1/OBF.1 for efficient TFH cell development during GC response. Reduced TFH cell frequencies in Pou2af1^fl/fl x IL21-Cre mice provide evidence for a TFH cell intrinsic defect in the absence of BOB.1/OBF.1 revealing that this reduction is not exclusively a consequence of defective CD4⁺ T cell differentiation towards TFH cells.

Within GCs, B cells expressing high affinity antibodies develop and differentiate into antibody secreting plasma cells (42). We observed a significant reduction in the number of activated CD69⁺ B as well as antibody producing plasma cells in both conditional knockouts post immunization ( Figures 5A, B ). Since an efficient GC reaction results in the production of class-switched secondary Igs, we assessed IgM and class-switched IgG Ig levels in serum of the conditional knockout strains. For Pou2af1^fl/fl x CD4-Cre mice total Ig levels were assessed during a primary and secondary immune response and measured 0, 14 and 21 days after immunization by ELISA. Despite our findings of impaired GC formation, we could not detect a striking reduction in IgM or class-switched Igs in these mice ( Figure S3A ). Pou2af1^fl/fl x IL21-Cre mice revealed a reduction of IgM, IgG2b and IgG3 compared to the floxed-controls, but not when compared to Cre-controls ( Figure S3B ). Notably, we found a significantly increased frequency of regulatory T cells (Tregs) in Pou2af1^fl/fl x CD4-Cre mice compared to controls in spleen and LNs ( Figure 5C ; Figure S3C ). Tregs are critical for the maintenance of immune homeostasis and regulate the initiation and input to the GC response (43). Since we observed increased levels of Tregs, we wondered about affinity maturation and subsequent production of antigen-specific Igs upon BOB.1/OBF.1 ablation in CD4⁺ T cells. Indeed, we could not detect significant differences in antigen-specific IgM levels, whereas levels of high-affinity anti-NP9 IgG1 and IgG3 were significantly reduced in Pou2af1^fl/fl x CD4-Cre mice 28 days after immunization with NP-KLH ( Figures 5D–F ). However, also levels of low affinity anti-NP27 antibodies of IgG1 and IgG3 isotype were significantly reduced in Pou2af1^fl/fl x CD4-Cre mice ( Figure 5G ). The ratio of NP9/NP27 antibody titers can be used as an index of affinity maturation. The ratio of high-affinity to low-affinity IgG1 was significantly reduced in the conditional knockout mice 28 days after immunization, even though this defect relativized on day 49 ( Figure 5H ). The ratio of NP9/NP27 IgG3 titers was not affected in Pou2af1^fl/fl x CD4-Cre mice ( Figure 5H ). In total, these data suggest a contribution of BOB.1/OBF.1 in T cells for the regulation of antigen-specific immune responses.

To identify DEGs in CD4⁺ T cells in the absence of BOB.1/OBF.1 we isolated CD4⁺ T cells from LNs of Pou2af1^fl/fl x CD4-Cre and CD4-Cre controls after immunization with SRBCs and performed RNA-seq analysis. 158 genes were significantly differentially expressed (DEGs) in Pou2af1^fl/fl x CD4-Cre compared to controls at FDR 0.05 including 61 upregulated and 97 downregulated genes ( Figure 6A ; Table S2 ) Hierarchical clustering analysis suggested that certain groups of the differentially expressed genes might have similar functions or might be involved in the same pathways ( Figure 6B ; heatmap FDR 0.01).

Significantly enriched GO terms at FDR 0.1 were identified related to molecular function. Top hits related to molecular function contained structural molecule activity (18), structural constituent of ribosome (16), translation regulator activity (10), rRNA binding (9), translation regulator activity/nucleic acid binding (8) and translation factor activity/RNA binding (8) ( Figure S4 ). When narrowing analysis to T cell specific GO-terms, several genes associated with T cell differentiation (12), T cell activation (20) and T cell proliferation (11) at FDR 0.05 could be identified as differentially expressed in Pou2af1^fl/fl x CD4-Cre mice ( Figures 7A–C ). We also checked for the expression of key TFs and cytokines relevant for TH cell differentiation. Tbx21 and Infg were downregulated (FDR < 0.1) in BOB.1/OBF.1-deficient CD4⁺ T cells confirming previous reports of reduced TH1 cytokines in conventional BOB.1/OBF.1-deficient mice and a direct contribution of BOB.1/OBF.1 to the Infg promoter (21) ( Figure 7D ). Consistently, three genes (Pde4d, Itk, Irf8) involved in IFN-γ production and seven genes (Cd74, Ccl22, Ccl5, Irf8, Ifng, H2-Q7, Dapk1) involved in the response to IFN-γ were differentially expressed at FDR < 0.05 ( Figures 7E, F ). Rc3h1 and Rc3h2 associated with negative regulation of TFH and TH17 cell differentiation were upregulated in BOB.1/OBF.1-deficient CD4⁺ T cells (FDR < 0.1) ( Figure S5A ). Interestingly, Rc3h1 is also described as negative regulator of GC formation. However, the TFH cell signature gene Bcl-6 was unaltered. In line with previous reports (24), Zbtb32, which encodes a transcriptional repressor of Prdm1/Blimp1, a known modulator of T cell memory, was downregulated (FDR < 0.1) ( Figure 7D ). Several genes associated with Treg cell differentiation or T cell function/differentiation during GC formation could also be identified as differentially expressed, but without statistical significance ( Figures S5B, C ).

Waterfall plot shows the top 30 DEGs in Pou2af1^fl/fl x CD4-Cre mice compared to CD4-Cre controls ( Figure 6C ). The strongest DEG was 2010007H06Rik, which function is unknown so far. The most interesting genes among these top 30 DEGs regarding T cell signaling and function are summarized in Table 1 . Interestingly in the context of BOB.1/OBF.1 being a transcriptional co-activator, four genes with DNA-binding TF activity (Sox4, Ubn2, Irf8, Irf4) were identified as differentially expressed in Pou2af1^fl/fl x CD4-Cre mice. Genes associated with chemokine signaling (Ccr5, Ccl5, Ccl22) necessary for chemoattraction of T lymphocytes were all downregulated in Pou2af1^fl/fl x CD4-Cre animals. Several genes associated with Treg, TFH and TH cell development (Sox4, Kmt2a, Kmt2d, Irf8, Irf4, Cd74) could be identified among the top 30 DEGs upon BOB.1/OBF.1 ablation specifically in CD4⁺ T cells. In silico analyses revealed consensus octamer motifs within the first 3,000 bp upstream of the transcriptional start site in eight genes that are among the top 30 DEGs in Pou2af1^fl/fl x CD4-Cre mice (Rab4a, Sox4, Mid1, Fasl, Crppa, Mpeg1, Cep170, Dapk1). Moreover, all other genes listed in Table 1 reveal multiple putative Oct/BOB.1/OBF.1 binding sites in the promoter region that are either perfect or differ in one or two positions from the consensus site. However, all these sites harbor an adenine at position five of the octamer motif that is essential for ternary complex formation with Oct1 and BOB.1/OBF.1 (8, 44). Three representative possible BOB.1/OBF.1 binding sites for each gene are shown in Figure S6 .