Bhlhe40 Promotes CD4⁺ T Helper 1 Cell and Suppresses T Follicular Helper Cell Differentiation during Viral Infection

In response to a host infection, naive CD4⁺ T cells differentiate into effector subsets to promote pathogen clearance. This occurs during viral infections, in which CD4⁺ T cells differentiate into T helper 1 (Th1) or T follicular helper (Tfh) cells (1–3) defined by their phenotype and function. The function of Th1 cells is to sustain the CD8⁺ T cell response to enhance viral clearance (4,5), whereas the function of Tfh cells is to signal B cells to generate germinal center reactions necessary for long-lived Ab responses during the late phase of infection (1, 6,7). Th1 cells are identified by canonical expression of surface markers Signaling Lymphocytic Activation Molecule (SLAM), CXCR 3 (CXCR3), CXCR 6 (CXCR6), and C-X3-C motif chemokine receptor 1 (CX3CR1). Through their production of the proinflammatory cytokine IFN-γ, Th1 cells both attenuate viral replication and augment the effector function of CD8⁺ T cells to promote viral clearance (8–10). In contrast, Tfh cells are identified by surface markers Programmed Death 1 (PD-1), CXCR 5 (CXCR5), and ICOS. Functionally, Tfh cells produce B cell-stimulating cytokines IL-4 and IL-21 (11–13).

The lineage fate of Th1 and Tfh cells is determined by their different master transcriptional regulators T-bet and BCL-6, respectively (14–18). The molecular switch determining the bifurcation of Th1 and Tfh cells depends on the interplay and threshold levels between T-bet and BCL-6. Evidence suggests that in addition to promoting lineage-defining transcription, another function of BCL-6 and T-bet is to transcriptionally repress each other, which further ensures a distinct terminal differentiation process (15, 19,20). Other studies characterize transcriptional regulation of BCL-6 (21–24) and T-bet (25,26) and how these additional factors assist in fate commitment of CD4⁺ T cells. These findings imply that fate commitment does not rely on only one master regulator but rather the threshold expression of each lineage-defining transcriptional regulators and their interactions with other transcription factors. Despite the well characterized roles of T-bet and BCL-6, there remains a gap in knowledge as to how other transcription factors may be involved in the Th1/Tfh lineage commitment.

The transcription factor Bhlhe40 is a basic helix-loop-helix protein known to bind to class B enhancer box elements as a homodimer (27, 28). Bhlhe40 has cell-intrinsic roles in regulating a wide variety of cellular processes in both nonhematopoietic and hematopoietic cell types such as macrophages, B cells, NK T cells, CD8⁺ T cells, and CD4⁺ T cells (18, 27, 29–41). In CD4⁺ T cells, Bhlhe40 has been presented as a “double-edged sword” because it has both protective and pathologic capabilities (42). However, the role of Bhlhe40 has yet to be dissected in terms of regulating Th1 versus Tfh bifurcation. Mechanistically, Bhlhe40 has been shown to directly bind to Bhlhe40, Ifng, and Il10 gene loci in Th1 cells in vitro (30). In Tfh cells, Bhlhe40 has been reported to restrain Tfh cell proliferation through direct repression of cell cycle–related genes but does not affect Tfh-related genes such as Bcl6, Prdm1, Irf4, Maf, Batf, Foxo1, or Rc3h1 (43). This demonstrates a dual role of Bhlhe40 to simultaneously activate and repress genes in CD4⁺ T cells. However, prior studies do not fully address the molecular mechanism of how Bhlhe40 facilitates CD4⁺ T cell fate commitment. During both acute and chronic viral infection, our laboratory identified differential expression levels of Bhlhe40 in CD4⁺ T cell subsets. Specifically, we identified high Bhlhe40 expression in Th1 cells and low levels in germinal center Tfh cells. We hypothesized that Bhlhe40 functions as an activator of Th1-related genes while simultaneously repressing Tfh-related genes. In this study, we further demonstrate the multifaceted role of Bhlhe40 in Th1 and Tfh differentiation at both the cellular and transcriptional levels.

Five to eight C57BL/6 mice, 6–8 wk old, were obtained through the National Cancer Institute (Frederick, MD). Bhlhe40GFP transgenic (Tg) Reporter mice were generated by the Gene Expression Nervous System Atlas Project (44). These mice were initially obtained from Dr. Brian T. Edelson (St. Louis, MO). TCR transgenic SMARTA mice were purchased from The Jackson Laboratory (stock no. 030450) and crossed with Bhlhe40^−/− initially obtained from Dr. William Drobyski (Medical College of Wisconsin) to generate SMARTA CD45.1⁺ Bhlhe40^−/− mice. The mice were bred and maintained in a closed breeding facility, and mouse handling conformed to the requirements of the Institutional Animal Care and Use Committee guidelines of the Medical College of Wisconsin.

The mice were infected with 2 × 10⁵ PFU of LCMV Armstrong i.p. to establish acute infection or LCMV clone 13 (Cl 13) via retro-orbital i.v. into mice at 2 × 10⁶ PFU/mouse to establish chronic infection in three to four independent experiments. The viruses were prepared by single passage from BHK21 cells.

All flow cytometry data were acquired on either an LSRII or Celesta flow cytometer (BD Biosciences, CA) analyzed by FlowJo (Treestar, OR). Lymphocytes were isolated from the spleen, and serum was harvested from blood. Cell staining with GP66:A-b PE tetramer (National Institutes of Health tetramer core facility) was performed in conjunction with Abs against cell surface Ags at room temperature for 1 h. All other staining panels, including panels in which GP33-41 tetramer (made in-house) was used, were performed at 4°C for 30–60 min. Transcription factor staining was performed using the True Nuclear transcription factor buffer set (BioLegend). For intracellular cytokine staining, the splenocytes were incubated with gp61-80 peptide (0.5 mg/ml) for 5 h at 37 C in the presence of brefeldin and monensin. The cells were stained with anti–IFN-γ Ab using Cytofix/Cytoperm (BD) according to the manufacturer’s recommendations.

Recipient mice were irradiated with 6.5 and 5.5 Gy separated by 4 h. Bone marrow from various donor mice (CD4^−/− mice were maintained in-house, Bhlhe40^−/− bone marrow was kindly provided by Dr. William Drobyski (MCW) and Dr. Jie Sun (UVA) were mixed at the indicated ratios, and a total of ∼3,000,000 cells were transferred i.v. into recipient mice, that were then maintained on oral sulfamethoxazole for 2 wk. Experimental mixed bone marrow (MBM) chimera mice were then infected with LCMV Armstrong or LCMV Cl 13 at 8 wk postreconstitution.

LCMV-specific (GP66-77 tetramer+) activated CD4 T cells were harvested from the spleens of five LCMV Armstrong–infected mice on day 10 postinfection and were FACS sorted using a BD-FACS-Melody Sorter (BD Biosciences) (8). Sorted cells were loaded onto the 10 × Chromium Controller with a target cell number of 5,000–10,000 per mouse. Single-cell RNA sequencing (scRNA-seq) libraries were prepared using the Chromium single-cell 59 version 2 and version 3 reagent kits (10× Genomics) according to the manufacturer’s protocol. Five libraries were then quantified using the Kapa Library quantification kit and then were loaded onto an Illumina NextSeq 500 sequencer with the NextSeq 500/550 high-output kit version 2 (150 cycles; FC-404–2002; Illumina) with the following conditions: 26 cycles for read 1, 98 cycles for read 2, and 8 cycles for the i7 index read. The raw sequencing data were downloaded using the Python Run Downloader (Illumina) and then demultiplexed and converted to gene–barcode matrices using the Cellranger (version 2.1) mkfastq and count functions, respectively (10× Genomics). A total of ∼18,000 cells were recovered from five mice. The downstream analysis was performed in R (version 3.6.0) using the package Seurat (version 2.3.3) (45) and Monocle 2 (46). The number of genes detected per cell, number of unique molecular modifiers, and percentage of mitochondrial genes were plotted, and outliers were removed (number of genes > 3,000, number of unique molecular modifiers over ∼20,000, or percentage of mitochondrial genes > 5%) to filter out doublets and dead cells. The cell cycle score was calculated for all cells and regressed out.

LCMV Armstrong (2 × 10⁶ PFU) was spun down in PBS overnight to a single pellet. ELIZA coating buffer (BioLegend) was used to resuspend the LCMV Armstrong pellet and aliquoted into a 96-well plate overnight. The wells were incubated in blocking buffer, then in serially diluted serum, and lastly secondary Ab for 1 h at room temperature. The wells were washed with PBS in between incubations. BD OptEIA TMB Substrate Reagent buffer was diluted 1:1 then added to each well. Stopping solution (BioLegend) was used to inhibit the reaction. The wells were detected using a Tecan plate reader.

Spleens from day 7 LCMV Armstrong-infected mice were fixed with periodate-lysine paraformaldehyde and then snap-frozen in OCT tissue-freezing solution and stored at −80 C. Tissues were cut into 7-mm sections, placed onto super frost glass slides, and stored at −80 C until staining. Prior to immunostaining, the tissues were rehydrated and blocked with 2% BSA, 5% goat serum, and 2% FCS. Anti-B220 PE, anti-CD4 BV421, and anti GL-7 AF647-conjugated Abs were used to stain spleen sections. The images were obtained using a Nikon TI2-E inverted microscope at 40× magnification, and IMARIS software (Bitplane) version 9 was used to prepare the images.

IMARIS (version 9.6.1) was used to quantify immunofluorescent images. The spots were determined using a threshold expression for each fluorescent channel. Colocalization of spots were measured based on the “shortest distance to spots” (μM). Measurements over 1 μM were filtered out. The statistics were calculated as within the program as “number of spots per timepoint” for each channel and “total number of spots” for each channel and plotted as a percentage.

For SMARTA phenotyping experiments, 10,000–15,000 congenic SMARTA cells were transferred into B6 recipients 1 day prior to LCMV Armstrong infection. The phenotypes of SMARTA cells were analyzed at day 10 after acute infection.

SMARTA CD4⁺ subsets (Ly5.1⁺ CD4⁺ CD44⁺ CXCR6⁺ and Ly5.1⁺ CD4⁺ CD44⁺ CXCR5⁺) were sorted by flow cytometry from spleens from recipient 6–8-wk-old C57BL/6 mice 10 d after LCMV Armstrong infection. Approximately 150,000–300,000 cells from each CD4⁺ subset were used for library construction using the CUT&Tag-seq protocol from Epicypher. Anti-Bhlhe40 (1:100, NB100-1800; Novus Biologicals), anti-H3K27Ac (1:50, Epicypher, 13-0045), anti-H3K27Me3 (1:25, Thermo Fisher, MA5-11198), anti-H3K4Me3 (1:50, Epicypher, 13-0041), and IgG (1:50, Epicypher, 13-0042) were used in this study. Thirty-seven cycles of paired-end sequencing were performed on an Illumina NextSeq 500, and 5–10 million reads were generated for each sample.

FastQC (version 0.11.8) was used to check the sequencing read quality. Each dataset was downsampled to equal read depths. For standardization between experiments, Escherichia coli DNA derived from transposase protein production was used to normalize sample read counts based on the recommendation of the CUT&Tag protocol (47). Reads were aligned to the Mus musculus mm10 genome and that of E. coli (strain K12) using Bowtie 2 (version 2.2.5) (48). The peaks were called using SEACR (version 1.1) (49), annotated with HOMER (version 4.9.1) (50), and visualized using IGV (version 2.8.2) and seqMiner (ver 9.4) (51).

SMARTA CD4⁺ T cell conditions described from the CUT&Tag experiments were used for ATAC-seq experiments. ATAC-seq was performed based on the protocol by Buenrostro et al. (52). A total of 50,000 Ly5.1⁺ CD4⁺ CD44⁺ CXCR6⁺ and Ly5.1⁺ CD4⁺ CD44⁺ CXCR5⁺ cells were used for library construction. Sequencing was performed on Illumina NextSeq550 with 37 cycles of paired-end sequences. Per ATAC-seq protocol, 50 million reads were generated for each sample.

The sequencing read quality was checked using FastQC (version 0.11.8). Paired-end sequencing of the libraries was performed on an Illumina NextSeq 500 sequencer. The raw sequencing data were first processed by nf-core/atacseq pipeline (version 1.2.2) (53) with default settings. Sequencing reads were aligned to GRCm38 mouse genome by BWA (54). MACS2 (55) was used for peak calling with a threshold of FDR > 0.05 and consensus peaks that were found in at least two replicates were kept for downstream analysis. Differential analysis was then performed with DEseq2 (version 1.36.0) (56). The preprocessed data were analyzed with DiffBind (version 3.6.5) (57) to identify the open chromatin regions uniquely accessible in the consensus peaks sets of each condition. The identified condition specific peak sets were then exported in bed file format for motif analysis and gene annotation using Homer (version 4.1.0) (50). Peak tracks were visualized by IGV viewer (58) and seqMiner (ver 9.4) (51).

SMARTA CD4 T cell conditions described from the CUT&Tag experiments were used for bulkRNA-seq experiments. RNA-seq libraries were prepared using the SMART-seq protocol (59) and sequenced on Illumina NextSeq 500.

The raw sequencing data were first processed by nf-core/rnaseq pipeline (version 3.8.1) with the default settings. The sequencing reads were aligned to GRCm38 mouse genome by Salmon. Differential analysis was then performed with DEseq2 (version 1.36.0) (56). Gene set enrichment analysis was performed with cluster Profiler (version 4.4.4) (60) and gene set database msigdbr (version 7.5.1) (61). ggplot2 (version 3.4.0) was used for plotting.

The data are expressed as means ± SEM. p ≤ 0.05 was considered statistically significant. Statistical tests were performed using GraphPad Prism version 9.

The scRNA-seq data have been deposited in the GEO database (accession no. GSE158896). All other R code and analyses are available from the corresponding author upon request.

To assess the expression of Bhlhe40 in CD4⁺ T cells during acute LCMV Armstrong infection, we analyzed our data from single-cell (sc)RNA-seq on GP_66-81 (GP₆₆)-specific CD4⁺ T cells (8). This revealed six major transcriptionally distinct clusters that arise 10 d postinfection as visualized by uniform manifold approximation and projection (Fig. 1A). Clusters were divided into Th1, Tfh, and T central memory precursor cells (Tcmps) based on their expression of lineage-specific markers and comparison with published Th1, Tfh, and Tcmp gene sets. Within the Th1 compartment, there were two transcriptionally distinct Th1 subsets, identified as Ly6c^hi Th1 and Lag3^hi Th1 (8). Within the Tfh compartment, there were three distinct Tfh subsets: pre-Tfh, Tfh1, and germinal center (GC) Tfh. Pre-Tfh cells are precursor Tfh cells based on their expression of memory markers Il7r, Ccr7, Lef1, and Tcf7. Tfh1 cells expressed Th1-associated markers Tbx21, Ifng, Gzmb, and Cxcr6. Both subsets expressed canonical GC Tfh markers such as Pdcd1, Cxcr5, Bcl6, and Icos but comparatively lower than the GC Tfh population (8, 10). Although Bhlhe40 expression was highly enriched in Th1 subsets, Bhlhe40 varied in the Tfh subsets, showing a gradient-like trend in pre-Tfh, Tfh1, and GC Tfh cells. (Fig. 1B, 1C). To further analyze the role of Bhlhe40 during the lineage commitment of Th1 and Tfh subsets, the R package Monocle 2 was employed to analyze the virus-specific CD4⁺ T cell subsets. This program uses unsupervised algorithms to predict differentiation trajectories (46). Aligning expression of Bhlhe40 within these subsets, lineage trajectory analysis showed high expression levels in terminally differentiated Th1 subsets. This, along with coexpression with other Th1-related genes, implied that the level of Bhlhe40 expression was positively correlated with the terminal Th1 differentiation program (Fig. 1C, 1D, top). However, as pre-Tfh cells progressed toward either Tfh1 or GC Tfh cells, they gradually downregulated expression of Bhlhe40, implying that Bhlhe40 expression was negatively correlated with Tfh cells but more so with the germinal center (Fig. 1D, bottom). Overall, this suggests that Bhlhe40 is coexpressed with Th1-related genes while negatively coexpressed with GC Tfh-related genes.

Based on our scRNA-seq findings, we next tested whether Bhlhe40 was required for Th1 and Tfh differentiation. To do this, we generated a series of MBM chimeras to restrict Bhlhe40 deletion in CD4⁺ T cells. A total of 70% of Cd4^−/− bone marrow from congenic CD45.1⁺ mice were mixed with 30% of either wild-type (WT) or Bhlhe40^−/− bone marrow cells from CD45.2⁺ congenic mice, and these bone marrow mixtures were grafted into lethally irradiated CD4^−/− recipient mice (Supplemental Fig. 1A). With this method, the only source of CD4⁺ T cells was the WT or Bhlhe40^−/− CD45.2⁺ congenic mice (Supplemental Fig. 1B), whereas other immune cells, such as CD8⁺ T and B cells, were largely (70%) WT. After reconstitution, the chimera mice were infected with LCMV Armstrong, and immune responses were assessed 10 d postinfection. We used CXCR6⁺ to define Th1 cells, CXCR5⁺ to determine GC Tfh cells, and CXCR6⁻ CXCR5⁻ to define double-negative (DN) cells, which have been reported to be memory-like CD4⁺ T cells during chronic infection (10). Although our data showed no significant differences in the frequency of GP₆₆-specific CD4⁺ T cells or total number of CD4⁺ T cells (Fig. 2A, 2E, 2F), there was a 3-fold increase in frequency and absolute number of Tfh cells and a significant decrease in Th1 cells in the spleen compared with the WT controls (Fig. 2B, 2G, 2H). During acute infection, the DN population expressed the highest level of Il7r among three CD4⁺ subsets, suggesting that the DN population exhibits a memory-like phenotype (62). In this article, we showed that there were no significant changes of DN cells between Bhlhe40-deficient CD4⁺ chimera mice and their WT control counterparts (Fig. 2M). This suggests that Bhlhe40 does not regulate the memory-like subset during acute infection. Interestingly, GP₆₆-specific GC Tfh cells (PD-1⁺ CXCR5⁺) were present at higher frequencies in Bhlhe40-deficient CD4⁺ mice while also showing an increase in frequency and total number of GC Tfh cells compared with the WT control group (Fig. 2C, 2I). Furthermore, the expression of Th1-associated markers, such as Ly6C and SLAM, showed a significant decrease, whereas the Tfh-associated marker ICOS did not change in the Bhlhe40-deficient CD4⁺ T cells compared with WT CD4⁺ T cells (Fig. 2J). In line with reduced Th1 cell formation, the effector molecules Gzmb and IFN-γ were also decreased in Bhlhe40-deficient CD4⁺ T cells, suggesting a critical role for Bhlhe40 in sustaining Th1 cell effector function (Fig. 2D, 2K, 2L). This finding is consistent with Bhlhe40-deficient Th1 cells having reduced capacity to produce IFN-γ (18). This was also supported by changes in lineage-specific markers such as a significant decrease in T-bet and increase in BCL-6 expression in GP₆₆-specific CD4⁺ cells (Fig. 2L). To highlight the importance of Bhlhe40 in regulating GP₆₆-specific CD4⁺ cells during viral infection, we also assessed Th1 and Tfh subsets during LCMV Cl13 infection. Similar to acute infection, we found that Bhlhe40 deletion in CD4⁺ T cells reduces Th1 and increases GC Tfh responses (Supplemental Fig. 2A–E). Lastly, no differences in GP₃₃-specific CD8⁺ T cells were observed between the two groups of chimeras during acute or chronic infection (Supplemental Figs. 1C–E and 2F–H). These findings collectively indicate the role of Bhlhe40 in virus-specific CD4⁺ T cells as a negative regulator of Tfh differentiation and a positive regulator of Th1 differentiation.

To confirm the cell-intrinsic requirement Bhlhe40 in facilitating CD4⁺ T cell differentiation during viral infection in vivo, we used an adoptive transfer system. Congenic CD45.1⁺ WT or CD45.1⁺Bhlhe40^−/− SMARTA cells, which express transgenic TCRs for the LCMV-specific epitope GP_66-81 (63), were independently transferred into C57BL/6 recipient mice and subsequently infected with LCMV Armstrong 1 day later (Fig. 3A). In accordance with our previous findings, deletion of Bhlhe40 in virus-specific CD4⁺ T cells significantly augmented Tfh cell formation accompanied with reduced Th1 CD4⁺ T cell development (Fig. 3C, 3F). We also observed a decrease in Th1-associated phenotypic markers such as Ly6C and SLAM and an increase in GC Tfh-associated markers such as PD-1 and CXCR5 expression (Fig. 3D, 3G, 3H). Deletion of Bhlhe40 in transgenic CD4⁺ T cells showed significant changes in IFN-γ and TNF-α production upon ex vivo GP₆₆ peptide stimulation (Fig. 3E, 3I). Assessment of the lineage-specific transcription factors T-bet and BCL-6 confirmed that Bhlhe40 deficiency decreases Th1 and increases Tfh differentiation (Fig. 3J). Taken together, our data confirm the cell-intrinsic requirement of Bhlhe40 in promoting Th1 and suppressing Tfh differentiation during viral infection.

With increased Tfh responses in Bhlhe40-deficient CD4⁺ chimera mice, we next tested whether Bhlhe40-deficient CD4⁺ T cells could regulate B cell–mediated humoral immunity because Tfh cells are essential in providing help to B cells (64). In the same chimera model as shown above, we observed a 2-fold increase in both frequency and number of CD19⁺ B220⁺ GL-7⁺ CD95⁺ GC B cells with minimal differences in plasma cells at day 10 postinfection (Fig. 4A–C; Supplemental Fig. 3A–C). This could be due to the increased Tfh cell formation and production of helper signals to B cells in Bhlhe40-deficient CD4⁺ chimera mice. We also analyzed GC B cells and plasma cells in Bhlhe40-deficient CD4⁺ chimera mice 21 d after LCMV Cl13 infection and observed a more robust increase in GC B cells and plasma cells in Bhlhe40-deficient CD4⁺ T cells (Supplemental Fig. 3D–I). By limiting the deletion of Bhlhe40 in CD4⁺ T cells, the humoral responses were indirectly affected likely due to increased Tfh responses. To confirm the increase in GC Tfh and GC B cell responses were virus-specific and not spontaneous or potentially autoreactive, we have also assessed unchallenged GC B cells and total IgG Ab titers in these MBM chimera mice to show that our phenotype was in fact significant in a viral infection model (Supplemental Fig. 4D, 4E). To assess whether Bhlhe40-deficient CD4⁺ chimera mice exhibited augmented humoral immunity, we assessed the Ab production in serum. On day 10 postinfection, Ab titers in the blood showed an overall increase in total IgG, LCMV-specific IgG1, IgG2a, and IgG2c in serum from Bhlhe40-deficient CD4⁺ chimera mice compared with WT controls (Fig. 4D). In sum, these observations suggest that Bhlhe40 not only intrinsically regulates CD4⁺ T cell differentiation but also indirectly affects GC B cell–mediated humoral immunity during viral infection.

To understand the expression of Bhlhe40 in CD4⁺ T cell subsets over time and to confirm the scRNA-seq data at the cellular level, we infected (bacterial artificial chromosome Tg Bhlhe40 reporter mice with LCMV Armstrong and analyzed Bhlhe40 expression in CD4⁺ T cells at days 7, 10, and 14 (Fig. 5A) (44). During acute infection, the expression of Bhlhe40^GFP among GP₆₆-specific CD4⁺ CD44⁺ CXCR6⁺ cells peaked at day 7 followed by a gradual decline with intermediate expression then lowest at day 14 (Fig. 5B). Moreover, Bhlhe40^GFP expression was minimal in CD4⁺ CD44⁺ CXCR5⁺ cells through the course of infection. In addition to tracing the kinetics of Bhlhe40, these findings confirmed our scRNA-seq analysis in that Bhlhe40 expression was highest in Th1 opposed to GC Tfh cells. We also assessed Bhlhe40 expression in CD11C⁺ T-bet⁺ B cells, suppressive and nonsuppressive neutrophils, and macrophages to show that Bhlhe40 was ubiquitous and may have various functions in different cell types (22, 42, 65) (Supplemental Fig. 4B, 4C).

To confirm our scRNA-seq findings that Bhlhe40-expressing CD4⁺ T cells are not in the germinal center, we analyzed the B cell follicles of Bhlhe40^GFP Tg mice 10 d after LCMV Armstrong infection. Using IMARIS to help quantify the overlap of Bhlhe40^GFP+ cells, we compared this percentage of overlap with the percentage of CD4⁺ GL-7⁺ and CD4⁺ GL-7⁻ T cells. CD4⁺ T cells that expressed GL7⁺ were less than CD4⁺ T cells that did not express GL7⁻. As expected, the coexpression of CD4⁺ and Bhlhe40⁺ was not found within the germinal center (Fig. 5C). These findings showed that CD4⁺ T cells that also expressed Bhlhe40 are not found in the germinal center (Fig. 5D). Altogether, our work postulates that Bhlhe40 functions as a regulator of the GC reactions and may need to be downregulated for CD4⁺ T cells to enter the germinal center.

To better understand how Bhlhe40 regulates Th1 and Tfh differentiation, we decided to first characterize the chromatin structure of virus-specific CD4⁺ T cells and dissect how it contributes to their distinctive gene expression. To this end, cleavage under targets and tagmentation (CUT&Tag) was performed on SMARTA CD45.1⁺ CD4⁺ CXCR6⁺ Th1 and CD45.1⁺ CD4⁺ CXCR5⁺ Tfh cells 10 d postinfection. Genome-wide analysis of H3K4Me3 and H3K27Me3 was shown at the promoter, exons, introns, and intergenic regions (Fig. 6A). The majority of H3K4Me3 peaks were located ≤1 kb of the promoter regions, whereas H3K27Me3 peaks were shown to be prominent at the distal intergenic regions and introns. This is consistent with recent findings in Th1 cells (35, 66). There were minimal changes in the proportionality of H3K4Me3 and H3K27Me3 between Th1 and Tfh subsets (Fig. 6A). To further identify the differences between H3K4Me3 and H3K27Me3 at promoter regions in Th1 and Tfh cells, the read density of each histone mark was compared between the CD4⁺ T cell subsets showing unique and overlapping promoter regions. Th1 cells had 14,724 unique H3K4Me3 and 1,800 unique H3K27Me3 promoter regions. In contrast, Tfh cells had 16,609 unique H3K4Me3 and 2,726 unique H3K27Me3 promoter regions. Both subsets shared 4,401 H3K4Me3 and 1,616 H3K27Me3 promoter regions (Fig. 6B). Overall, the data show that the transcriptional signatures of each subset are characterized by distinct regulation of cis-regulatory elements.

To correlate the distinct gene expression profiles of Th1 and Tfh subsets with H3K4Me3 and H3K27Me3 modifications, promoter regions of differentially expressed genes from bulk RNA-seq data (p < 0.05 and log₂ > 0.5) of Th1 and Tfh cells were calculated for both H3K4Me3 and H3K27Me3 signals. Th1-related genes such as Cxcr6, Gzmb, Prdm1, and Id2 displayed high H3K4Me3 and low H3K27Me3 signatures in Th1 cells while displaying low H3K4Me3 and high H3K27Me3 signatures in Tfh cells. Similarly, Tfh-related genes such as Bcl6, Cxcr5, Tcf7, and Sostdc1 displayed a permissive signature in Tfh cells while showing a repressive signature in Th1 cells (Fig. 6C, 6D). The switch from a permissive to repressive methylation state in Th1 and Tfh cells showed the need for one subset-specific gene profile to be repressed for the other to be activated. Although Tbx21 encodes for the master transcriptional regulator of Th1 differentiation, it was not epigenetically repressed in Tfh cells. On the contrary, Tbx21 showed H3K4Me3 signals in both Th1 and Tfh cells (Fig. 6D). This aligns with findings suggesting that T-bet is needed for Tfh differentiation during acute viral infection (67–69). Interestingly, Tcf7 showed bivalent chromatin in Th1 cells but still maintained its permissible signature in both Tfh versus Th1 cells. This could be explained by higher H3K27Me3 than the H3K4Me3 signal found at the promoter region of the Tcf7 locus in Th1 cells (Fig. 6D) but was still poised and could become easily activated when needed. Overall, these findings highlight that histone methylation changes at the promoter regions significantly contribute to the unique gene expression profiles of Th1 and Tfh CD4⁺ subsets.

To identify other cis-regulatory elements such as enhancers that contribute to the CD4⁺ subset-specific differentiation, transposase-accessible chromatin using sequencing (ATAC-seq) and CUT&Tag was used. We found that Th1 and Tfh subsets possessed distinct chromatin-accessible regions (Fig. 7A). In addition, genome-wide analysis of H3K27Ac signals showed minimal proportionality differences between subsets (Supplemental Fig. 5A). Correlation analysis supported that the enhancer regions were closely correlated between replicates and unique to each CD4⁺ subset (Fig. 7B). We also analyzed the shared and differentially accessible regions between CD4⁺ subsets and saw that most of the differential chromatin-accessible regions between Th1 and Tfh subsets were located at the enhancer regions (Supplemental Fig. 5C). This indicates that enhancers play an important role in CD4⁺ subset-specific differentiation.

To identify the role of Bhlhe40 in regulating Th1- and Tfh-related genes at enhancer regions, we checked Bhlhe40 occupancy at active and inactive enhancers measured by ATAC⁺ H3K27Ac⁺ and ATAC⁺ H3K27Ac⁻, respectively. Of the Bhlhe40-occupied regions, ∼70% were active enhancers that had genes nearby related to the Th1 signature (Fig. 7C, left, top). These genes included Gzmb, Selplg, Fasl, and Ccl5. Bhlhe40 occupancy was enriched at transcription factor motifs T-bet and Prdm1 through HOMER motif analysis. The remaining 30% of Bhlhe40-occupied regions were inactive enhancers related to Tfh signature. Genes within this list included Il21r, Il6ra, and Cxcr5. Bhlhe40 occupancy was enriched at transcription factor motifs related to Tfh differentiation such as Egr2, Bach1, and Ascl2 (Fig. 7C, left, bottom). For instance, Bhlhe40 selectively occupied active distal enhancers (+2.3, −12.58, −16.6, and −18.6 kb from TSS) of Tbx21 locus (Fig. 7D, left) and occupied inactive distal enhancers (+4.1 kb from TSS) of Bcl6 locus (Fig. 7D, right) in Th1 cells. This suggests Bhlhe40 can directly activate enhancers of Th1-related genes and inactivate enhancers of Tfh-related genes.

To further assess the role of Bhlhe40 in regulating Th1- and Tfh-related genes at promoter regions, we analyzed Bhlhe40 occupancy at active and inactive promoters characterized by H3K4Me3⁺ and H3K4Me3⁻, respectively. Approximately 60% of Bhlhe40-occupied sites were active promoters that had nearby genes such as Il2rb, Cxcr3, and Selplg. Bhlhe40 occupancy was enriched at Th1-related transcription factor motifs such as T-bet and Prdm1 (Fig. 7C, middle, top). Roughly 40% of Bhlhe40-occupied sites were inactive promoters close to Tfh signature genes such as JunB, Tox2, and Cxcr5. In addition, Bhlhe40 was enriched at Tfh-related transcription factor motifs such as BCL-6, Ascl2, and PU.1 (Fig. 7C, middle, bottom). To confirm that the inactive promoters occupied by Bhlhe40 are repressed, we assessed Bhlhe40 occupancy at H3K27Me3⁺ and H3K27Me3⁻ sites. Of the Bhlhe40-occupied sites, ∼30% of H3K27Me3⁺ sites were close to Tfh signature genes such as Tox2. Transcription factor motifs enriched by Bhlhe40 included Tcf7, Bcl6, and Lef1 (Fig. 7C, right, top). The remaining 70% of Bhlhe40-occupied H3K27Me3⁻ sites were close to Th1-related genes such as Ifng, Cxcr3, Selplg, and Gzmb (Fig. 7C, right, bottom). This showed that Bhlhe40 could also bind to promoter regions to directly activate or suppress Th1- or Tfh-related genes, respectively (Supplemental Fig. 7). In addition, we assessed genome-wide binding sites of Bhlhe40 and the potential pathways that Bhlhe40 could be regulating (Supplemental Fig. 6A–C). Taken together, the results indicate that Bhlhe40 directly activates Th1-related gene promoters and enhancers and directly represses Tfh-related gene promoters.

In this paper, we demonstrate that Bhlhe40 functions as a transcriptional activator of Th1 lineage-specific genes and as a transcriptional repressor of Tfh lineage-specific genes. Our findings further suggest Bhlhe40 regulates the bifurcation of CD4⁺ Th1 and Tfh during acute LCMV infection at both the cellular and epigenetic levels. There have been long-standing debates over the origin and bifurcation of Th1 and Tfh CD4⁺ T cells (44, 70–72). Our work further characterizes the role of Bhlhe40 in Th1 and Tfh CD4⁺ subsets during acute infection. As Th1 cells decline over the course of acute infection, CD4⁺ T cells such as pre-Tfh and Tfh1 cells begin to downregulate Bhlhe40 expression as they commit to becoming GC Tfh cells. Although our data strongly show the role of Bhlhe40 in regulating terminally differentiated Th1 and Tfh cells, it would be worthwhile to characterize its role in facilitating the heterogeneity of CD4⁺ T cells. In agreement with recent findings (43), downregulation of Bhlhe40 expression is necessary for CD4⁺ T cells to enter the germinal center, thus defining the role of Bhlhe40 as a negative regulator of the germinal center reaction. Our work also shows that Bhlhe40 regulates not only CD4⁺ T cells but also their ability to provide “help” signals to B cells and their Ab production. Changes in GC B cells and virus-specific Abs suggest that Bhlhe40 signaling in Tfh cells ultimately affects the magnitude of its help to B cells. We have also seen similar phenotypic patterns in Bhlhe40-deficient CD4⁺ T cells during acute and chronic infection, highlighting the importance of Bhlhe40 in facilitating Th1 and Tfh differentiation. Thus, these findings collectively provide evidence of Bhlhe40 as a transcriptional repressor of Tfh lineage-specific genes in CD4⁺ T cells during viral infection at the cellular level. Upon CD28 activation, Bhlhe40 has been reported to significantly upregulate its expression early as 4 h in an autoreactive CD4⁺ T cell model (38). In the context of viral Ag stimulation, Bhlhe40 needs to be upregulated early on to activate the Th1-related gene profile while suppressing the Tfh differentiation program. Upstream signals of Bhlhe40 are suggested to be cytokines IL-2, IL-12, or anti-IL4, given that in vitro stimulation of Th1 cells upregulates Bhlhe40 expression (18, 30). Our HOMER motif analysis also provides evidence that Bhlhe40 putatively binds close by gene Il2rb in Th1 cells, suggesting that after viral Ag stimulation, Bhlhe40 is activated upon IL-2 stimulation. In this paper, we did not assess upstream regulation of Bhlhe40; therefore, it would be a worthwhile future endeavor to determine the signal that induces in vivo Bhlhe40 expression.

Mechanistically, evidence suggested Bhlhe40 has the capacity to function as a transcriptional activator and repressor in different cell types (18, 27, 29–36, 38, 40, 41). Despite recent findings of Bhlhe40 in Th1 and Tfh system (30, 43), our contribution is using a naturally occurring system to dissect the molecular mechanism of Bhlhe40 in defining CD4⁺ lineage choice. Through this, we were able to define the multifaceted role of Bhlhe40 functioning as an activator of Th1-related genes and repressor of Tfh-related genes in CD4⁺ T cells through analysis of H3 methylation and acetylation markers. Combined ATAC-seq and CUT&Tag analysis showed that Bhlhe40 binds to both promoter and enhancer regions to activate Th1 and repress Tfh gene profiles. In line with other studies (18, 30), our data showed that Bhlhe40 binds to Ifng gene locus to directly activate expression through active enhancer regions and binds to the Il10 gene locus to directly repress by inactive enhancer regions. Our data contradict the findings of Rauschmeier et al. (43) in which they suggested that Bhlhe40 does not affect the known positive or negative regulators of Tfh cell development (Bcl6, Prdm1, Irf4, Maf, Batf, Foxo1, and Rc3h1). There are potential reasons for these discrepancies. One reason for the discrepancy could be the analysis of different cell types. These findings assess the gene expression profile in Ag85b-specific CD4 cells, whereas our findings are from Bhlhe40-DNA interactions in LCMV-specific CXCR6⁺ Th1 cells. Another reason could be the use of different techniques. Bulk RNA-seq captures average gene expression levels, which could mask the subset-specific gene expression in the Ag85b-specfic CD4 T cells. In contrast, CUT&Tag is relatively sensitive and can provide specific insights into the binding location of Bhlhe40. The discrepancy in epigenomic and transcriptional data may suggest that there are other regulatory mechanisms downstream of Bhlhe40 regulation that facilitate CD4 T cell differentiation. Through our analysis, Bhlhe40 not only directly binds to the master transcriptional regulators of Th1 and Tfh to activate and suppress, respectively, but also binds to genes that are related to their differentiation. Through HOMER motif analysis, we were also able to find Bhlhe40 putative binding sites with itself at E-box sites; AP-1 family transcription factors such as JunB, JunD, and c-Jun; KLF family; and ATF family. This suggests that Bhlhe40 can not only homodimerize to regulatory DNA elements through recognition of E-box sites (28), it is also able to dimerize with different families of transcription factors to regulate Th1- and Tfh-related genes. It has been reported that in some instances Bhlhe40 and its close binding partner Bhlhe41 form heterodimers to regulate B1a-progenitor B cells and macrophages (30, 33, 42). Although our scRNA-seq data show very low expression of Bhlhe41 in all CD4 subsets during acute and chronic infection, it is possible Bhlhe41 could play a role when Bhlhe40 is deficient in a system. Further studies are warranted to confirm cofactor binding with Bhlhe40 to regulate CD4⁺ T cell differentiation.

The authors have no financial conflicts of interests.

This research was completed in part with computational resources and technical support provided by the Research Computing Center at the Medical College of Wisconsin. We thank the nf-core community for developing the nf-core infrastructure and resources for Nextflow pipelines. A full list of nf-core community members is available at https://nf-co.re/community.