KMT2D coordinates antiviral CD4⁺ T cell responses through opposing effects on T follicular helper and cytotoxic gene expression

SUMMARY

T follicular helper (T_FH) cells are essential for protective antibody responses. Histone modifications direct T_FH development and function; however, the role of specific chromatin modifiers in this process is not well understood. Lysine methyltransferase 2D (KMT2D) is a histone methyltransferase that acts at H₃K₄ to promote gene expression. Herein, we examined the contribution of KMT2D to T cell responses during acute lymphocytic choriomeningitis virus infection. Mice lacking KMT2D in T cells generated sufficient antiviral CD8⁺ T cell responses to resolve infection. However, these mice formed fewer T_FH cells and had diminished germinal center and antibody responses. Mechanistically, KMT2D sustained T_FH responses in part by promoting Thpok and Il21 expression through H₃K₄Me₁ deposition at gene enhancers. Consistent with loss of THPOK, KMT2D-deficient CD4⁺ T cells acquired a cytotoxic CD4⁺ T (CD4^CTL) cell phenotype involving elevated expression of RUNX3, EOMES, and cytolytic molecules. Our findings show KMT2D balances T_FH development and humoral immunity over CD4^CTL cells.

In brief

Distinct subsets of T cells confer immune protection against infections. Histone modifiers integrate TCRs and other signaling pathways with transcription factors to specify each subset and its function. Herein, Cohen et al. reveal a central role for KMT2D in advancing T_FH formation and antibodies while restricting cytolytic Th subsets.

Graphical Abstract

INTRODUCTION

Durable and high-affinity antibody responses are important for efficient control of virus infection and can provide lasting protection after vaccination or primary infection.^1,2 These antibody responses depend upon CD4⁺ T follicular helper (T_FH) cells that express C-X-C motif chemokine receptor 5 (CXCR5) and migrate into the B cell follicles to coordinate germinal center (GC) reactions. Within the GC, T_FH cells deliver critical signals to cognate B cells that are required for the development of affinity-matured antibodies, long-lived plasma cells, and memory B cells.^3–7

Virus-specific CD4⁺ T cells can also differentiate into the T helper 1 (T_H1) lineage that promotes antiviral CD8⁺ T cell responses and supports cellular defenses at sites of infection. T_FH and T_H1 lineages are defined by their expression of the master transcriptional regulators B cell lymphoma 6 (BCL6) and T-box 21 (Tbet), respectively.⁸ Naive CD4⁺ T cells that receive signals through surface receptors such as ICOS and interleukin (IL)-6 receptor (IL-6R) upregulate BCL6 and differentiate into T_FH cells.^9–11 Other critical transcription factors, such as BATF, cMAF, TCF1/LEF1, TOX2, and THPOK, act upstream and/or in parallel to BCL6 to support T_FH differentiation and function.^12–16 In contrast, signals such as IL-2, IL-12, interferon (IFN)-γ, and type I IFN promote T_H1 differentiation by upregulating the expression of Tbet and BLIMP1.^17–20

CD4⁺ T cell differentiation into distinct T_H lineages is enabled by epigenetic changes, including histone modifications, that alter chromatin accessibility and direct transcription factor binding.²¹ In T_FH cells, activating histone marks such as histone 3 (H3) lysine 27 (K27) acetylation (H₃K₂₇Ac) and H₃K₄ mono- (Me1) and tri- (Me3) methylation are enriched at the enhancer and promoters of highly transcribed genes such as Il21 and Bcl6. In contrast, genes associated with other T_H lineages, such as Prdm1 and Runx3, are kept in an inactive or poised configuration through a combination of repressive H₃K₂₇Me3 and activating histone marks.^22–25 Collectively, these epigenetic marks may facilitate T_FH gene expression and predispose T_FH cells toward expression of alternative helper cell programs under certain inflammatory conditions.

Despite these initial chromatin studies, relatively little is known about how specific histone-modifying proteins program T_FH responses. KMT2A (MLL1) promotes initial T_FH differentiation by upregulating TCF1 expression.²⁶ The H₃K₂₇me3 methyltransferase EZH2 cooperates with TCF1 to promote BCL6 expression and separately represses inhibitors of T_FH differentiation and survival.^25,27,28 Our group previously found that UTX was critical for sustaining T_FH responses during chronic virus infection by promoting IL-6Ra expression through its H₃K₂₇ demethylase activity.²⁹

UTX acts as part of a larger epigenetic complex that includes histone lysine methyltransferase 2D (KMT2D), an H₃K₄ mono-methyltransferase.^30,31 KMT2D and UTX cooperatively promote cell-type-specific enhancer activation through complementary histone modifications and scaffolding activity.^30,32–34 Germline mutations in KMT2D and UTX cause Kabuki syndrome, a rare multisystem disorder that frequently presents with low antibody titers and poor vaccine responses.^35,36 KMT2D also has distinct functions compared to UTX in many developmental or disease settings.^30,37–40

A growing body of work implicates KMT2D in immunity. KMT2D regulates B cell gut homing, GC proliferation, and plasma cell differentiation.^41–43 Among developing T cells, KMT2D promotes regulatory T cell (Treg) differentiation and thymic egress.^44,45 During Listeria monocytogenes infection, KMT2D increases the survival of activated CD8⁺ T cells.⁴⁶ KMT2D-primed enhancers had varying effects on T_H1/T_H2 gene expression in CD4⁺ T cells that were cultured with polarizing cytokines.^47,48 However, whether KMT2D is needed for CD4⁺ T helper cell responses in vivo is not known.

Here, we explored the role of KMT2D in virus-specific CD4⁺ and CD8⁺ T cells during lymphocytic choriomeningitis virus (LCMV)-Armstrong infection. Mice with a T cell-specific deletion of KMT2D (KMT2D^TCD) efficiently controlled the infection due to a largely intact antiviral CD8⁺ T cell response. However, KMT2D^TCD mice showed defects in the formation of virus-specific T_FH cells and associated GC and antibody responses that coincided with an outgrowth of cytotoxic CD4⁺ T (CD4^CTL) cells. Correspondingly, KMT2D sustained the expression of genes involved in T_FH development and function and limited those involved in cytotoxicity. Our findings link CD4⁺ T cell expression of KMT2D to humoral immune defenses.

RESULTS

Antiviral CD8⁺ T cells do not require KMT2D expression to resolve acute LCMV infection

To assess whether KMT2D is important for antiviral T cell responses, we generated mice with a T cell-specific deficiency in KMT2D (KMT2D^TCD; B6.Lck-Cre+;Kmt2d^fl/fl) by intercrossing Kmt2d^fl/fl mice that harbor loxP sites flanking exons 50–51 with mice carrying an Lck-Cre transgene. Wild-type (WT; either B6. Lck-Cre+;Kmt2d^WT/WT or B6.Lck-Cre-;Kmt2d^fl/fl) and KMT2D^TCD mice were infected with LCMV-Armstrong, which causes a systemic virus infection that is normally resolved within ~8 days in C57BL/6 mice by CD8⁺ T cells.

Virus was undetectable by plaque assay in the sera of 11/13 mice from both groups, indicating that T cell expression of KMT2D was not required to control LCMV-Armstrong infection (Figure 1A). Both groups contained similar numbers of total activated CD4⁺ T cells, whereas KMT2D^TCD mice had a modest decrease in activated CD8⁺ T cells (Figure 1B). We used qPCR to verify efficient deletion of Kmt2d by measuring the relative amount of genomic Kmt2d exons 50–51, which tightly correlates with KMT2D protein levels.⁴⁹ Kmt2d deletion ranged from 50% to 99% for both activated CD4⁺ and CD8⁺ T cell subsets, and the majority of KMT2D samples showed >95% deletion (Figure 1C). Consistent with loss of KMT2D methyltransferase activity, activated CD4⁺ and CD8⁺ T cells from KMT2D^TCD mice had a highly significant reduction in H₃K₄Me1 compared to WT (Figure 1D).

Figure 1. CD4⁺ T cells in KMT2D^TCD mice have modestly reduced cytokine expression.

WT and KMT2D^TCD mice were infected with LCMV-Armstrong. At 8–9 days after infection, viral titers in the serum, splenic T cell numbers, and CD4⁺ T cell cytokine expression were quantified.

(A) Amounts of infectious virus in sera were determined by plaque assay. The limit of detection is 50 plaque-forming units (PFU)/mL serum.

(B) Number of activated CD4⁺CD44^hi and CD8⁺CD44^hi T cells per spleen.

(C) qPCR was used to quantify the amount of KMT2D exons 50–51 in genomic DNA from CD8⁺CD44^hi and CD4⁺CD44^hi T cells. To account for differences in starting material, KMT2D exon 50–51 Ct values were normalized against an unrelated genomic region.

(D) The flow histogram shows H₃K₄Me₁ expression among CD4⁺CD44^hi T cells. The graph shows the gMFI of H₃K₄Me₁ among CD8⁺CD44^hi and CD4⁺CD44^hi T cells.

(E) Flow plots show CD4⁺ T cells bound by I-A^b GP_66–77 tetramer (GP66+), and the graph shows numbers of GP66+ cells per spleen.

(F) Splenocytes were stimulated ex vivo with GP_61–80 peptide and analyzed for cytokine expression. Flow plots show CD4⁺ T cells expressing IFN-γ and IL-2. The graphs show the number of CD4⁺ cells expressing the indicated cytokine and the gMFI of cytokines among cytokine+ cells.

(G) Number of polyfunctional CD4⁺ T cells that co-express IFN-γ, TNF, and IL-2 after GP_61–80 peptide stimulation.

For (C), data are from 1 experiment with 4 mice per group. (D) shows 1 of 3 representative experiments with 4–6 mice per group per experiment. (B) and (E) are pooled from 5–6 independent experiments with 3–6 mice per group per experiment. (F) and (G) are pooled from 3 independent experiments with 3–6 mice per group per experiment. Error bars depict mean ± SEM. Significance was determined using an unpaired Mann-Whitney U test (A) or unpaired Student’s t test (B–G) (*p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001). gMFI, geometric mean fluorescence intensity.

See also Figures S1 and S2.

Both groups generated similar numbers of D^bGP_33–41 and D^bNP_396–404 tetramer+ CD8⁺ T cells in the spleen by days 8–9 post-infection, although there was a trend toward decreased DbNP396–404 in KMT2DTCD mice (Figure S1A). DbGP33–41+ CD8⁺ T cells from each group expressed similar amounts of granzyme B (GZMB), suggesting comparable cytotoxic potential (Figure S1B). However, CD8⁺ T cells in KMT2D^TCD mice had impaired expression of IFN-γ and tumor necrosis factor (TNF) after ex vivo stimulation with LCMV peptides GP₃₃, NP₃₉₆, and NP₂₀₅, although this varied based on the epitope and cytokine measured (Figure S1C). The reduction in cytokine expression was likely unrelated to effector cell differentiation because there were only minor differences in the percentage of short-lived and memory-precursor effector cells among only D^bNP_396–404+ CD8⁺ T cells (Figure S1D). Similar trends for CD8⁺ T cell numbers and cytokine expression were observed at day 15 post-infection (Figures S1E and S1F).

We considered that the impaired expression of cytokines by CD8⁺ T cells in KMT2D^TCD mice might be due to alterations in the CD4⁺ T cell compartment. Therefore, we directly measured the CD8⁺ T cell-intrinsic function of KMT2D by engrafting WT and KMT2D^TCD CD8⁺ P14 cells (which express a T cell receptor [TCR] transgene recognizing the GP_33–41 epitope) into WT hosts, which were subsequently infected with LCMV (Figure S2A). At day 7 post-infection, the phenotype of KMT2D^TCD P14 cells in the spleen and blood broadly recapitulated that of endogenous GP₃₃-specific CD8⁺ T cells from KMT2D^TCD mice, including relatively normal granzyme expression and short-lived effector cell (SLEC) and memory precursor effector cell (MPEC) distribution but diminished cytokine expression (Figures S2B–S2D). Taking these results together, we find that KMT2D is dispensable for CD8⁺ T cell expansion, cytotoxicity, and clearance of acute virus infection but does support the expression of inflammatory cytokines. Given the relatively minor impact of KMT2D on CD8⁺ T cells, we focused the remainder of our study on the primary antiviral CD4⁺ T cell response.

CD4⁺ T cell co-expression of pro-inflammatory cytokines is reduced in KMT2D^TCD mice

CD4⁺ T cells coordinate immune responses through the expression of cytokines and surface ligands that provide help to multiple cell types, including B cells and CD8⁺ T cells. At 8–9 dpi, both groups contained similar numbers of I-A^bGP_66–77-tetramer⁺ (GP66+) CD4⁺ T cells in the spleen that recognize the immunodominant major histocompatibility complex (MHC) class II LCMV epitope (Figure 1E). In an intracellular cytokine staining (ICCS) assay, virus-specific CD4⁺ T cells from KMT2D^TCD mice expressed normal amounts of IFN-γ but lower amounts of TNF and IL-2 on a per-cell basis following stimulation with the GP_61–80 peptide compared to the WT (Figure 1F). Polyfunctional CD4⁺ T cells expressing all three cytokines are associated with superior antiviral immune responses,^50,51 and we observed fewer polyfunctional CD4⁺ T cells in KMT2D^TCD mice compared to WT (Figure 1G). Thus, similar to CD8⁺ T cells (Figures S1 and S2), KMT2D is dispensable for virus-specific CD4⁺ T cell proliferative expansion but needed for optimal co-expression of multiple cytokines.

KMT2D^TCD mice have impaired T_FH cell responses

Given that KMT2D-deficient CD4⁺ T cells made less cytokine than WT, we next assessed whether KMT2D is important for establishing virus-specific T_FH and T_H1 cells. T_H subsets can be distinguished by their differential expression of surface markers.^52–58 T_FH cells (identified as CXCR5+CXCR6− or CXCR5⁺PSGL1^lo) and T_H1 cells (CXCR5−CXCR6+) were found in the spleens of WT mice after infection (Figures 2A and 2B). In comparison, KMT2D^TCD mice had ~2- to 3-fold fewer T_FH cells but normal numbers of T_H1 cells (Figures 2A and 2B). Alternative gating strategies identifying T_FH and T_H1 cells through expression of CXCR5, SLAM, and FR4 yielded similar results (Figures S3A and S3B). Interestingly, KMT2D^TCD mice had significantly fewer virus-specific PSGL1^hiLy6C^hi CD4⁺ T cells (Figure S3C), a terminally differentiated subset of T_H1 cells that produce the highest amount of IFN-γ.⁵⁹ However, considering that CD4⁺ T cell expression of IFN-γ was comparable between groups (Figure 1F), it may be that KMT2D regulates Ly6C expression rather than terminal T_H1 differentiation. Virus-specific T helper cell responses were not simply delayed in KMT2D^TCD mice because the overall numbers of GP66+ CD4 T cells were significantly decreased in KMT2D^TCD mice compared to the WT at 15 days post-infection (dpi) (Figure S3D), and this was largely due to a decrease in T_FH cells, although we cannot rule out defects in early memory formation (Figure S3E).

Figure 2. KMT2D promotes T_FH responses during acute viral infection.

The abundance of splenic LCMV-specific T helper cells and their expression of critical transcription factors and surface receptor ligands were analyzed by flow and scRNA-seq at 8–10 dpi.

(A) Flow plots show CD4+GP66+ cells expressing CXCR5 and CXCR6 with quadrant gates set to discriminate T_FH (CXCR5+CXCR6−) and T_H1 (CXCR5−CXCR6+) cells. Using this gating strategy, the graphs show the percentage of CD4+GP66+ T cells that are T_FH and T_H1 and the number of CD4+GP66+ T_FH and T_H1 cells per spleen.

(B) Flow plots show splenic CD4+GP66+ cells expressing PSGL1 and CXCR5 with gates set on T_FH (PSGL1^loCXCR5+). Using this gating strategy, the graphs show the percentage of CD4+GP66+ T cells that are T_FH and the number of GP66+ T_FH cells per spleen.

(C) The gMFI of ICOS among GP66+ T_FH (CXCR5+) and non-T_FH (CXCR5−) cells.

(D) The gMFI of BCL6 among GP66+ T_FH cells (CXCR5+).

(E) The percentage of GP66+ T_FH cells (CXCR5+) that express TCF1 in each group.

(F) GP₆₆-tetramer+ CD4⁺ T cells were sorted from the spleens of WT and KMT2D^TCD mice and subjected to scRNA-seq. Example flow plots show pre- and post-sort enrichment of total CD4+GP66+ T cells. The post-sort GP66+ gate was lowered slightly to account for excitation-induced photobleaching.

(G) The transcriptome of WT and KMT2D cells is represented by the UMAPs. Cell clusters were defined based on single-cell gene expression profiles and are colored according to their assigned cluster. Clusters 9–15 are colored gray and were excluded from downstream analyses.

(H) Heatmap of select T_FH and T_H1 genes per cluster. The heatmap intensity indicates the average expression of each indicated gene and cluster, relative to the average expression for that gene across all clusters.

(I) Violin plots of the of T_FH, T_CMP, and T_H1 module scores in each cluster calculated using signature gene lists for each T_H population obtained from Ciucci et al.⁶⁰

(J) Pie charts show the proportion of cells assigned to each major cluster. Clusters that show significant differences between groups are indicated by asterisks in the legend.

Data were pooled from 3 independent experiments with 3–6 mice per group for (A)–(D). Error bars are mean ± SEM for phenotypic data. scRNA-seq data are from two samples per group. Violin plots show mean and 25^th and 75^th percentile expression. Significance was determined by unpaired Student’s t test (*p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001). MFI, geometric mean fluorescence.

See also Figure S3.

The reduced number of T_FH cells in KMT2D^TCD mice at 8–9 dpi is most likely not explained by changes in regulatory subsets, T_FH proliferation, or apoptosis. The frequencies of FoxP3+ T follicular regulatory (T_FR) cells and Tregs were lower in KMT2D^TCD mice (Figures S3F and S3G), consistent with a role for KMT2D in FoxP3 expression.⁴⁴ There was a slightly increased frequency of KMT2D^TCD T_FH cells expressing Ki67 (Figure S3H), suggesting no defect in T_FH proliferation. T_FH cells (CXCR5+) in each group expressed similar amounts of the pro- and anti-apoptotic molecules BIM and BCL2, indicating that KMT2D^TCD T_FH cells do not have a greater propensity to undergo apoptosis at this time point. Interestingly, non-T_FH cells (CXCR5−) from KMT2D^TCD mice had increased BCL2 and an elevated BCL2:BIM ratio, suggesting that non-T_FH cells may survive better than WT controls (Figure S3I).

Given that KMT2D^TCD mice generated fewer T_FH cells, we quantified the expression of a few regulators of T_FH responses, including ICOS, BCL6, and TCF1. GP66+ T_FH and non-T_FH cells from each group expressed comparable amounts of ICOS and BCL6 (Figures 2C and 2D). In contrast, TCF1 was expressed by significantly fewer KMT2D^TCD T_FH cells. (Figure 2E). Thus, KMT2D appears to be dispensable for CD4⁺ T helper cell expansion and survival, but it may support T_FH responses by promoting the expression of critical transcription factors.

Single-cell transcriptomic analyses reveal that KMT2D associates with T_FH cell populations

To more broadly characterize how KMT2D impacts CD4⁺ T cell responses to virus infection, we performed single-cell RNA sequencing (scRNA-seq) on total GP66+ CD4⁺ T cells sorted from the spleens of WT and KMT2D^TCD mice at 10 dpi (Figure 2F). Sequenced single cells pooled prior to sorting were demultiplexed using the Cell Ranger “multi” algorithm. After filtering low-quality single cells by a percentage of mitochondrial genes and total read and feature counts, we obtained 18,514 unique single-cell profiles (7,148 from the WT and 11,366 from the KMT2D^TCD groups) (Figure S3J). Dimensionality reduction by principal-component analysis and sample integration were performed prior to uniform manifold approximation and projection (UMAP) of single-cell profiles onto 2D space. Cell clustering yielded 8 major clusters (clusters 1–8) (Figure 2G) and 7 minor clusters (9–15) consisting of only 30–100 cells per sample that were excluded from further analysis. Cluster 1 likely represented T_FH-like cells based on the expression of genes such as Cxcr5, Bcl6, and Il21 and had the highest T_FH gene signature score (Figures 2H and 2I). Cluster 2 cells were likely CD4⁺ T central memory precursor (T_CMP) cells based on the high expression of genes like Ccr7, Il7r, and Tcf7⁶⁰ and a high T_CMP signature score. Cluster 3 was similar to a previously described Lag3^hi Th1 subset⁶¹ based on intermediate expression of Th1-associated genes such as Ifng and Tbx21 and high expression of Lag3 and Havcr2 (Figure 2H). Clusters 4 and 5 had high T_H1 signature scores and the highest expression of critical T_H1 genes, such as Ifng, Tbx21, Id2, and Ly6c2, and were likely fully differentiated T_H1 cells (Figures 2H and 2I). Lastly, clusters 6, 7, and 8 were likely cycling CD4⁺ T cells, based on the enrichment of cell cycle genes and comparatively low expression of prototypical T_FH and T_H1 genes, such as Tbx21 and Bcl6 (Figures 2H, 2I, and S3K).

There were 5-fold fewer KMT2D^TCD cells in cluster 1 (T_FH-like) compared to the WT, and they trended toward a decrease in cluster 2 (T_CMP-like) (Figure 2J). In contrast, KMT2D^TCD cells trended toward an increase in clusters 4 and 5 (T_H1-like), although this did not reach statistical significance. These findings are broadly consistent with our phenotypic data demonstrating that KMT2D is needed for T_FH, but not T_H1, responses (Figures 2A and 2B). In addition, the single-cell data show that KMT2D^TCD cells may have defects in T_CMP formation and have an elevated T_H1-like profile, which was not previously revealed by antibody staining against a limited number of surface markers.

KMT2D promotes T_FH functionality and the ensuing GC response

T_FH development occurs in multiple stages. At the outset of infection, naive CD4⁺ T cells differentiate into immature T_FH cells, which migrate toward lymphoid follicles and engage cognate B cells. This interaction enhances the expression of BCL6 and enables T_FH maturation into GC-T_FH cells that are indispensable for GC responses through the local production of factors such as CD40L and IL-21.⁶² Consistent with our findings above, there were significantly fewer virus-specific GC-T_FH cells in KMT2D^TCD compared to WT mice (Figure 3A). Fewer KMT2D^TCD T_FH cells expressed CD40L, although the amounts of CD40L expressed per CD40L⁺ cell were comparable between groups (Figure 3B). In addition, pseudobulk scRNA-seq analyses showed significantly lower expression of Il21 in the T_FH-like cluster from KMT2D^TCD GP66+ cells compared to the WT T_FH cluster (Figure 3C; Table S1).

Figure 3. KMT2D^TCD mice have reduced GC and antibody responses.

WT and KMT2D^TCD mice were infected with LCMV-Armstrong. GC-T_FH cells, T_FH effector function, and GC-B cells were analyzed in the spleen by flow cytometry and/or scRNA-seq at day 8–9 post-infection. LCMV-specific IgG was quantified in sera at days 8 and 15 post-infection.

(A) Flow plots show GP66-tetramer+ CD4⁺ T cells expressing CXCR5 and PD1 (programmed cell death protein-1), with gates drawn on GC-T_FH cells (CXCR5^hiPD1^hi). Graphs show the percentage of GP66+ GC-T_FH cells among CD4⁺ T cells and their number per spleen at 8–9 dpi.

(B) Flow plots show endogenous GP66+ cells expressing CXCR5 and CD40L, with quadrant gates drawn to distinguish CXCR5− and CXCR5+ cells that produce CD40L. The graphs show the percentage of GP66+ T_FH cells (CXCR5+) that express CD40L and the gMFI of CD40L among CD40L+ GP66+ T_FH cells at 8–9 dpi.

(C) The UMAP plot shows the expression of Il21 among all tetramer+ CD4⁺ T cells. The dotted area outlines cluster 1 (T_FH cluster) The violin plot of Il21 for cluster 1 is shown on the right, which indicates the false discovery rate (FDR)-adjusted p value for Il21 from pseudobulk differential expression analysis comparing WT and KMT2D^TCD cluster 1.

(D) Flow plots show B220+ cells expressing FAS and GL7, with gates drawn on GC-B cells (FAS^hiGL7^hi). The graphs depict the percentage of B cells (B220+) that are GC-B cells and the number of GC-B cells per spleen at 8–9 dpi.

(E) Flow plots show GC-B cell expression of CXCR4 and CD86, with gates identifying light zone (LZ; CXCR4^loCD86^hi) and dark zone (DZ; CXCR4^hiCD86^lo) GC-B cells. Graphs show the percentage of GC-B cells with an LZ or DZ phenotype and the ratio of LZ/DZ GC-B per spleen.

(F) Virus-specific IgG was quantified by ELISA. The left graph shows the OD₄₉₀ for serial dilutions of sera at 15 dpi; hyper-immune serum from a rechallenged LCMV-Armstrong immune WT mouse and serum from an LCMV-naive WT mouse are used for reference. Based on the dilution series, the area under the curve (AUC) was calculated for the indicated mice at 8–9 and 14–15 dpi.

scRNA-seq data in (C) are from two mice per group. All other data were pooled from 1–3 independent experiments with 3–6 mice per group per experiment. Error bars are mean ± SEM. Significance was determined using an unpaired Student’s t test (*p < 0.05 and **p < 0.01).

The diminished T_FH responses in KMT2D^TCD mice corresponded with an ~2- to 3-fold reduction in GC B cells (B220 +FAS+GL7+) compared to WT mice, though the total numbers of B cells (B220+) were similar in both groups (Figures 3D and S4A). The GC-B cells from KMT2D^TCD mice had a relative bias toward a light zone phenotype (CXCR4^loCD86^hi) (Figure 3E), which can result from restricted CD40 or IL-21R signaling.^63,64 In contrast, both groups had similar numbers of activated (non-GC) and resting B cells (Figures S4B and S4C). By day 15, KMT2D^TCD mice continued to show fewer GC-T_FH and GC-B cells (Figures S4D–S4F), although this did not reach statistical significance for GC-T_FH due to an outlier. Lastly, LCMV-specific immunoglobulin G (IgG) levels were significantly lower in KMT2D^TCD mice at day 15 post-infection (Figure 3F), a time point when GC output should substantially contribute to the circulating pool of virus-specific antibodies. Collectively, these findings suggest that KMT2D promotes T_FH development, function, and the ensuing GC and antibody responses.

CD4⁺ T cell-intrinsic expression of KMT2D is needed for T_FH responses

CD4⁺ T cell differentiation is strongly influenced by the immunological microenvironment. We considered that our previous findings might be due to extrinsic factors, such as decreased CD8⁺ T cell cytokine expression (Figures S1F and S2C) and Tregs (Figures S3F and S3G), that collectively impair T_FH responses in KMT2D^TCD mice. Alternatively, CD4⁺ T cell-intrinsic expression of KMT2D may be needed for T_FH cell formation and effector function. To distinguish between these possibilities, we performed adoptive transfers of CD4⁺ SMARTA transgenic cells that express an LCMV-specific TCR recognizing the GP_66–77 epitope. Equal numbers of WT and KMT2D^TCD SMARTA cells (CD45.1⁺) were engrafted into separate B6 hosts (CD45.2⁺), which were subsequently infected with LCMV; donor cell phenotypes were analyzed 7–8 days later (Figure 4A).

Figure 4. CD4⁺ T cell-intrinsic expression of KMT2D is needed for T_FH responses.

WT or KMT2D-deficient SMARTA cells were introduced into congenic (CD45.2+) mice followed by LCMV-Armstrong infection 1 day later. At days 7–8 post-infection, SMARTA cells were identified in the spleens of host mice by flow cytometry, along with their expression of cytokines, T_FH and T_H1 differentiation, critical surface receptor ligands, and transcription factors.

(A) Illustration of the experimental setup.

(B) Flow plots of CD4⁺ cells showing CD4 and CD45.1 expression, with gates set on CD4⁺CD45.1+ donor SMARTA cells. The graphs show the number of donor cells per spleen and their percentage among total splenic CD4⁺ cells.

(C) Splenocytes were stimulated ex vivo with GP_61–80 peptide, and donor cells were analyzed for cytokine expression. The left graph shows the frequency of donor cells expressing IFN-γ, TNF, and IL-2. The middle graph shows the gMFI of IFN-γ, TNF, and IL-2 among IFN-γ+, TNF+, and IL-2+ donor cells, respectively. The right graph shows the percentage of donor cells that co-express IFN-γ, TNF, and IL-2.

(D) The flow plots show donor cell expression of CXCR5 and CXCR6 with quadrant gates set to discriminate T_FH (CXCR5+CXCR6−) and T_H1 (CXCR5−CXCR6+) subsets. Using this gating strategy, the graphs show the percentage of donor cells that are T_FH or T_H1.

(E) Graphs show the percentage of donor cells that are T_FH (CXCR5+PSGL1^lo).

(F) Graphs show the percentage of donor T_FH cells (CXCR5+) that express CD40L and the gMFI of CD40L among CD40L+ donor T_FH cells.

(G) Splenocytes were stimulated ex vivo with GP_61–80 peptide and analyzed for cytokine expression. Flow plots show donor T_FH (CXCR5+) expression of IL-21 and CD4, with gates set on IL-21+ donor T_FH cells. The graphs show the percentage of donor T_FH cells expressing IL-21 and the gMFI of IL-21 among IL-21+ donor T_FH cells.

(H) Graphs show the gMFI of TCF1, ICOS, and BCL6 among donor T_FH cells (CXCR5+).

Data in (A)–(F) and (H) were pooled from 2 independent experiments with 3–4 mice per group per experiment. Data in (G) are from 1 experiment with 3–4 mice per group. Error bars depict mean ± SEM. Significance was determined using an unpaired Student’s t test (*p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001).

See also Figure S4.

The KMT2D^TCD SMARTA cell phenotype broadly recapitulated our findings for endogenous antiviral CD4⁺ T cell responses in KMT2D^TCD mice, with some exceptions. The donor cells, identified by co-staining for CD4 and CD45.1, showed comparable engraftment in uninfected recipients (Figure S4G) and expanded similarly after infection (Figure 4B). H₃K₄Me₁ was uniformly reduced in the KMT2D^−/− SMARTA group (Figure S4H), indicating that there was no selective outgrowth of cells lacking sufficient Cre recombinase activity and retaining KMT2D expression during infection. These data indicate that KMT2D^TCD SMARTA donors effectively compete with WT host cells for survival and proliferation cues.

Similar to our earlier findings, we observed a modest decrease in the percentage of KMT2D^TCD donor cells expressing IFN-γ, TNF, or IL-2 after GP_61–80 stimulation, and lower amounts of IL-2 were made in KMT2D^TCD donor cells on a per-cell basis. Moreover, there were ~10-fold fewer polyfunctional KMT2D^TCD donor cells compared to the WT (Figure 4C). A significantly lower percentage of KMT2D^TCD donors were T_FH cells, whereas the percentages of T_H1 cells were equivalent between groups (Figures 4D and 4E). Fewer KMT2D^TCD donor T_FH cells produced CD40L and IL-21 (Figures 4F and 4G), and KMT2D^TCD donor T_FH expressed lower amounts of TCF1 (Figure 4H). In contrast to endogenous KMT2D^TCD responses, KMT2D^TCD SMARTA T_FH cells had significantly increased expression of ICOS and BCL6 (Figure 4H). Although it is unclear why these two factors were elevated in this context, it is notable that elevated ICOS and BCL6 failed to overcome the T_FH defect in the KMT2D^TCD SMARTA group. It may be that ICOS- and BCL6-mediated gene regulation improperly functions in the absence of KMT2D. Collectively, these findings suggest that CD4⁺ T cell-intrinsic expression of KMT2D is necessary for T_FH development and function during acute virus infection.

KMT2D limits the formation of CD4^CTL cells during infection

Pseudobulk differential gene expression analyses of the scRNA-seq data (Figure 2) revealed that KMT2D^TCD cells highly overexpressed genes associated with CD4^CTL cells in every major cluster. For example, the transcription factors Runx3, Eomes, and Zeb2 were elevated in KMT2D^TCD cells, as were the effector genes Gzma, Gzmb, Gzmk, Prf1, and Fasl (Figure 5A; Table S1). The relative expression of cytotoxic genes was highest in cluster 5 (the T_H1-like cluster), followed by clusters 3, 4, 6, 7, and 8, and was lowest in cluster 1 (the T_FH-like cluster) (Figure 5B).

Figure 5. KMT2D limits the formation of cytotoxic CD4⁺ T cells.

In (A)–(I), WT and KMT2D^TCD mice were infected with LCMV-Armstrong. At days 8–11 post-infection, flow cytometry, scRNA-seq, and CUT&Tag analyses were used to analyze cytolytic molecules and transcription factors in splenic CD4⁺ T cells. (H) compares gene expression in KMT2D^TCD T_FH-like cells and ThPOK^−/− T_FH cells. In (J), WT and KMT2D^TCD SMARTA cells were retrovirally transduced with THPOK overexpression vector (THPOK-OE) and analyzed for T_FH differentiation in host spleens 8 days after transfer.

(A) Heatmap shows pseudobulk differential expression analysis of cytotoxicity-associated genes that are overexpressed in KMT2D^TCD cells. The heatmap intensity indicates the log₂ fold change in expression of KMT2D^TCD over WT cells in the same cluster.

(B) Violin plots of cytotoxic genes across each major cluster; boxplots depict the 25^th and 75^th percentiles of expression.

(C) The percentage of GP66+ cells that expressed GZMB based on flow cytometry.

(D) The flow plot shows polyclonal PD1+ CD4⁺ T cells expressing CXCR5 and PERFORIN, with quadrant gates distinguishing PERFORIN+ T_FH (CXCR5+) and non-T_FH (CXCR5−) cells. Graphs show the number of polyclonal T_FH and non-T_FH cells that expressed PERFORIN per spleen and the percentage of CD4⁺ T_FH and non-T_FH cells that express PERFORIN at 8–9 dpi.

(E) The number of GP66+ T_FH and non-T_FH cells that expressed EOMES per spleen and the percentage of GP66+ T_FH and non-T_FH cells that expressed EOMES at 8–9 dpi.

(F) The expression of Thpok among all tetramer+ CD4⁺ T cells is depicted by UMAP.

(G) Flow plots show GP66+ CD4⁺ T cells expressing THPOK and EOMES, with gates drawn on THPOK+ cells at 8 dpi. The left graphs show the percentage of GP66+ cells that express THPOK and the gMFI of THPOK among THPOK+ GP66+ cells at 8–11 dpi. Right graphs show THPOK expression in CD4⁺ T cells from uninfected mice.

(H) The regression analysis compares Z-normalized log2 fold expression changes of all cluster 1 DEGs from KMT2D^TCD cluster 1 and Tbet-deficient T_FH cells. Each dot represents the Z score of a gene. Selected genes are annotated in red text on the scatterplot.

(I) H₃K₄Me₁ enrichment at several genomic loci was measured by CUT&Tag followed by qPCR in CD44^hiCD4+ T cells at 9 dpi.

(J) The top row shows the illustration of the experimental setup. The flow plots show THPOK-OE transduced WT and KMT2D^TCD donor cells that are GFP+ and GFP− and expressing CXCR5 and CXCR6. Graphs show the percentages of GFP+ and GFP−cells from each donor group that are T_FH cells (CXCR5+CXCR6−) in both empty control and THPOK-OE transduction conditions.

Data in (C)–(E), (I), and (J) were pooled from 1–2 independent experiments with 3–6 mice per group; data in (G) were pooled from 2–4 independent experiments performed on day 0 and days 8–11 post-infection, respectively. gMFI values in (G) at 8–11 dpi show 2 of 4 representative experiments. Significance was determined by paired (J) or unpaired (C–G and I) Student’s t test (*p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001). scRNA-seq data represented in (A), (B), (F), and (H) are from two samples per group, and DEGs were determined by DESeq2 (FDR < 0.05).

See also Figures S5 and S7.

EOMES can drive CD4^CTL formation and function during LCMV infection.⁶⁵ EOMES, GZMB, PERFORIN, and FASL were all overexpressed in either polyclonal or tetramer+ CD4⁺ T cells from KMT2D^TCD mice in terms of the percentage and number of cells expressing these factors, as well as the amount made per cell (Figures 5C–5E and S5A–S5C). Consistent with scRNA gene expression, CD4^CTL molecules such as EOMES and PERFORIN were overexpressed in both T_FH and non-T_FH cells from KMT2D^TCD mice (Figures 5A, 5D, 5E, S5B, and S5C). EOMES is dually associated with the development of CD4^CTL- and IL-10-producing T regulatory 1 (T_R1) cells. However, KMT2D^TCD cells did not have elevated expression of T_R1-associated genes, including Il10, Il10ra, Cd27, Cd39 (Entpd1), and Lag3, and IL-10 was undetectable in GP_61–80-stimulated cells (Figures S5D and S5E). The cytotoxic phenotype was not associated with the loss of CD4 or the gain of CD8 expression, indicating that CD4⁺ T cell identity was intact in KMT2D-deficient CD4⁺ T cells. Reciprocal findings were observed for CD8⁺ T cells (Figures S5F and S5G). The elevated frequencies of EOMES+ CD4⁺ T cells in KMT2D^TCD mice persisted to day 15 post-infection (Figure S5H), suggesting that their cytotoxic gene expression was not transient. Lastly, in an adoptive transfer setting, only KMT2D^TCD SMARTA donor cells acquired the cytotoxic phenotype (Figures S5I and S5J), indicating that CD4⁺ T cell-intrinsic expression of KMT2D limits CD4^CTL development.

KMT2D activity is dispensable for early formation of T_FH and CD4^CTL lineages

We asked whether the T_FH and CD4^CTL phenotype observed was due to baseline differences in KMT2D-deficient CD4⁺ T cells prior to infection. Naive CD4⁺ T cells from KMT2D^TCD mice had significantly lower amounts of H₃K₄Me₁ than WT controls (Figure S6A). Unexpectedly, these cells expressed ~60% lower amounts of TCF1 compared to WT (Figure S6B). Despite lower TCF1, we found similar proportions of T_FH cells in each group at steady state (Figures S6C and S6D), and few CD4⁺ T cells expressed cytotoxic molecules such as EOMES and GZMB (Figures S6E and S6F). Moreover, proportions of GC-B cells were equivalent between groups (Figure S6G), suggesting similar T_FH activity.

Based on these and other findings (Figures 2, 3, and 4), it was unclear whether KMT2D programmed initial CD4⁺ T cell differentiation or acted later during infection to support T_FH cells and limit CD4^CTL development. To study the early role of KMT2D, we co-transferred WT (CD45^.1/.2) and KMT2D^TCD (CD45^.1/.1) SMARTA cells into B6 hosts (CD45.2⁺) and analyzed donor cells in the spleen 3.5 dpi (Figure S6H). Donor cells from each group were equally abundant at this time, and the KMT2D^TCD donor cells had significantly lower levels of H₃K₄Me₁ compared to the WT (Figures S6I and S6J). In contrast to later stages of infection, we observed similar percentages of T_FH cells among WT and KMT2D^TCD donor cells and negligible differences in their expression of CD40L, BCL6, and ICOS (Figures S6K–S6M). Surprisingly, TCF1 was only slightly reduced in KMT2D^TCD donor T_FH cells at 3.5 dpi (Figure S6M), which indicates that KMT2D is dispensable for the initial upregulation of TCF1 and T_FH differentiation that occurs shortly after virus infection.^12,13 This finding is consistent with a recent short hairpin RNA (shRNA) screen of chromatin regulators of early T_FH differentiation that did not identify Kmt2d.²⁶ Although GZMB and PERFORIN were upregulated in WT compared to KMT2D^TCD donor cells at day 3.5 (Figure S6N), expression was transient in WT CD4⁺ T cells and lost by 7 dpi, whereas expression was sustained in KMT2D-deficient donor CD4⁺ T cells up to day 7 (Figures S5I and S5J). Lastly, EOMES was not expressed in donor cells from either group at day 3.5 (Figure S6N), suggesting that it is not driving cytotoxic markers at this time point. Collectively, these findings indicate that KMT2D functions at later stages of infection to sustain T_FH responses and limit the formation of a stable CD4^CTL lineage.

KMT2D associates with enhancer elements important for T_FH lineages

Transcriptional control of T_FH and CD4^CTL development occurs through opposing regulatory mechanisms. Indeed, our scRNA analyses revealed that cytotoxic gene expression was lowest in T_FH-like clusters (Figure 5B). KMT2D is critical for gene enhancer activation; therefore, we considered that KMT2D might be needed to promote the expression of transcription factors important for T_FH formation and factors that repress CD4^CTL formation.

THPOK (encoded by Zbtb7b, herein referred to as Thpok) impedes CD4^CTL formation by repressing the transcription of Runx3 and Eomes, which drives CD4^CTL formation during LCMV infection.^14,65 THPOK can also promote T_FH differentiation and function by increasing the expression of factors such as Bcl6 and Maf.¹⁴ In our scRNA-seq analysis, Thpok expression was substantially reduced in KMT2D^TCD cells in every cluster compared to the WT (Figure 5F), suggesting that KMT2D regulates Thpok irrespective of T helper cell type. Consistent with transcriptomic data, GP66-tetramer+ CD4⁺ T cells from WT mice universally expressed THPOK, whereas most KMT2D^TCD cells lacked THPOK (Figure 5G), and among those still expressing THPOK, KMT2D^TCD cells expressed lower amounts per cell than the WT. EOMES was expressed almost exclusively in KMT2D^TCD GP66+ cells lacking THPOK (Figure 5G), consistent with THPOK-mediated repression of cytotoxic genes.^14,60,66,67 The loss of THPOK in KMT2D^TCD cells was acquired during infection, as naive CD4⁺ T cells from each group showed minimal differences in THPOK expression (Figure 5G). Moreover, initial T_FH differentiation after infection was intact in KMT2D^TCD donor cells (Figures S6D and S6E), unlike the profound early defects observed in THPOK^−/− CD4⁺ T cells.¹⁴

To understand the role of THPOK in KMT2D-mediated T_FH responses, we compared differentially expressed genes (DEGs) identified in KMT2D^TCD T_FH-like cells (cluster 1) in this study by pseudobulk analyses against DEGs identified by bulk RNA-seq of THPOK-deficient (Ox40Cre⁺Zbtb7b^+/+) T_FH cells (CXCR5^int PD1^int) relative to WT GC-T_FH controls at day 7 post-LCMV-Armstrong infection (GEO: GSE130474).¹⁴ Among 1,161 DEGs in cluster 1, approximately half were also differentially expressed in THPOK-deficient T_FH cells (Figure S7A; Table S1). Relative changes in expression between KMT2D^TCD and THPOK-deficient T_FH cells were highly correlated (Pearson r = 0.54) (Figure 5H). In contrast, there was no correlation of DEGs when comparing KMT2D^TCD T_FH-like cells to Tbet^−/− or Tet2^−/− T_FH cells, despite evidence that these factors promote or inhibit T_FH development (Figures S7B and S7C).^68,69 These data indicate that KMT2D and THPOK elaborate overlapping transcriptional programs for T_FH development.

To assess the potential for KMT2D to regulate Thpok expression, we first reanalyzed data from KMT2D:chromatin immunoprecipitation (ChIP)-seq studies performed on naive CD4⁺ T cells (GEO: GSE69162, Placek et al.⁴⁴). We identified KMT2D peaks relative to input control at multiple known Thpok enhancers that are located upstream or within intronic regions of Zbtb7b (Figure S7D). Next, to determine whether KMT2D methylation could regulate Thpok expression during virus infection, we fluorescence-activated cell-sorted activated CD4⁺ T cells from the spleens of WT and KMT2D^TCD mice at day 9 post-infection and subjected the cells to CUT&Tag for H₃K₄Me₁, followed by qPCR at several enhancers. H₃K₄Me₁ was significantly depleted at three Thpok enhancer sites in KMT2D^TCD cells compared to WT (Figure 5I). The GTE (general T lymphoid element) and proximal enhancer were previously shown to be important for Thpok expression in peripheral CD4⁺ T cells, whereas the activity of site A has not been studied outside of the thymus. H₃K₄Me₁ was diminished at an Il21 enhancer (CNS-49) in KMT2D^TCD CD4⁺ cells. For reference, both groups showed similarly low levels of H₃K₄Me₁ at HoxB1, which is not expressed in CD4⁺ T cells, and at a genomic “desert” region devoid of genes (Figure 5I). Our findings indicate that KMT2D directly targets Thpok and Il21 through its histone methyltransferase activity, possibly explaining how it promotes their expression.

Finally, to directly assess whether KMT2D supports T_FH responses by promoting THPOK, we retrovirally overexpressed Thpok or an empty vector in WT and KMT2D^TCD SMARTA cells that were engrafted into infected B6 hosts; T_FH differentiation was analyzed 8 days later (Figure 5J). We compared the percentages of T_FH cells among GFP+ donor cells, which were successfully transduced and overexpressed THPOK, against GFP− donor cells within each group. As expected,¹⁴ we observed that WT GFP + and GFP⁻ donor cells consisted of similar proportions of T_FH cells (Figure 5J), indicating that THPOK overexpression does not affect T_FH differentiation, most likely because WT cells express sufficient amounts of THPOK. As anticipated, the empty vector also had no impact on T_FH differentiation in either group. In contrast, THPOK overexpression significantly enhanced T_FH differentiation in GFP+ KMT2D^TCD donor cells (Figure 5J), which reached WT levels of T_FH cell formation. By comparison, TCF1 overexpression failed to rescue T_FH formation in KMT2D^TCD cells despite fully rescuing TCF1 amounts (Figures S7E and S7F), indicating that KMT2D employs TCF1-independent mechanisms to regulate T_FH differentiation.

In sum, our findings implicate a role for KMT2D methyltransferase activity in sustaining Thpok expression in CD4⁺ T cells responding to virus infection, which promotes T_FH formation, while restricting their differentiation into the CD4^CTL lineage.

DISCUSSION

T_FH cells act within the GC to mediate the formation of high-affinity, neutralizing antibodies that confer protection against viruses and other pathogens. This process unfolds through the iterative selection of affinity-matured B cells by T_FH cells, often over the course of weeks or months.⁷⁰ T_FH cells must persist within the GC and maintain specialized functions required for B cell help. We found that the histone methyltransferase KMT2D sustains T_FH responses during primary virus infection. Mice lacking KMT2D in T cells formed fewer T_FH cells and had a diminished virus-specific antibody response. In WT CD4⁺ T cells, KMT2D acted after initial CD4⁺ T cell differentiation to promote the expression of Thpok and Il21 by activating critical enhancers via its H₃K₄ mono-methyltransferase activity. CD4⁺ T cells lacking KMT2D failed to remain as T_FH cells and instead acquired a cytolytic transcriptional program. Our data implicate KMT2D in maintaining T_FH gene expression and limiting CD4^CTL development over the course of virus infection.

Among T_FH cells, there was transcriptional overlap between genes advanced by KMT2D and those affected by THPOK. KMT2D regulates broad cell-type-specific transcriptional programs through enhancer activation, yet THPOK overexpression alone was sufficient to rescue the overall defect in T_FH differentiation in KMT2D-deficient CD4⁺ T cells. Nevertheless, it is possible that KMT2D mediates other essential aspects of T_FH physiology independent of THPOK. Nearly half of the genes differentially expressed in KMT2D^TCD T_FH-like cells were expressed normally in THPOK-deficient T_FH cells, including genes like Mef2d, a transcriptional regulator of T:B adhesion and Il21 expression.⁷¹ KMT2D can associate with a distal CXCR5 enhancer in CD8⁺ T cells,⁴⁶ suggesting a possible role for KMT2D in T_FH migration. We showed that KMT2D may directly regulate effector genes such as Il21 through its methyltransferase activity. Thus, available evidence suggests a broader role of KMT2D beyond THPOK in regulating physiologically relevant T_FH cell gene expression.

KMT2D and UTX, an H₃K₂₇ demethylase, can function in the same epigenetic complex and regulate gene expression through complementary histone-modifying and scaffolding activity.^29,30,72 Our prior study found that UTX promotes T_FH responses during persistent virus infection by promoting their expression of Il6ra, and Il6ra was also downregulated in KMT2D^TCD cells based on our scRNA-seq data. Despite these commonalities, UTX and KMT2D likely regulate T_FH responses through shared and distinct mechanisms. UTX had no obvious effect on T_FH formation during acute infection and did not impact the expression of KMT2D-regulated genes such as ThPok.²⁹ Although Il6ra expression was decreased in KMT2D^TCD T_FH-like cells, IL-6 signaling is redundant for T_FH differentiation during acute LCMV infection.⁷³ Future studies are needed to assess whether KMT2D-mediated IL-6Ra is important for T_FH responses during persistent infection.

We do not know the degree to which KMT2D’s histone methyltransferase or scaffolding activity is needed for gene regulation in T cells. KMT2D is needed for UTX stability and nuclear localization.^32,74 UTX stabilization may be a critical demethylase-independent function of KMT2D in T_FH responses. It is possible that KMT2D and UTX are redundant for the expression of certain T_FH genes, and changes to their expression may be evident only in CD4⁺ T cells lacking KMT2D and UTX. In addition, KMT2D and UTX may have additive or synergistic effects on the expression of other sets of genes.

Cytotoxic CD4⁺ T cells are increasingly identified in a variety of disease contexts, although their precise function has been difficult to parse out. Protective CD4^CTL responses, identified in various viral and bacterial infections and during malignancy, exert their effects by killing infected or transformed cells. Conversely, maladaptive CD4^CTL responses are reported in settings such as recurrent childhood tonsillitis, in which cytotoxic T_FH cells eliminate GC-B cells and limit durable antibody protection.⁷⁵ It is interesting to speculate whether the cytotoxic features acquired by KMT2D-deficient T_FH cells during LCMV infection contributed to antibody defects in KMT2D^TCD mice.

The pathways involved in CD4^CTL formation vary between tissue and disease settings. In the intestines, intraepithelial lymphocytes (IELs) constitute a major constitutive population of cytotoxic CD4⁺ T cells that promote tissue homeostasis and microbial defense.⁷⁶ CD4⁺ T cells differentiate into IELs through progressive downregulation of Thpok and upregulation of Runx3 and Eomes. These transcriptional changes are associated with signaling molecules such as IFN-γ, IL-27, transforming growth factor beta (TGF-β), PD1, and Aryl hydrocarbon receptor (AhR).^66,77–79 Whether aberrant signaling through these factors triggers THPOK downmodulation and CD4^CTL formation in KMT2D-deficient CD4⁺ T cells during infection is a matter of future investigation.

Kabuki syndrome frequently presents with low IgA and IgG titers and poor vaccine responses.^41–43,80 KMT2D deficiency can contribute to B cell-intrinsic defects in migration and terminal differentiation in GCs in mice and in humans with Kabuki syndrome. Our study implies that there may be elevated CD4^CTL and deficient T_FH responses in Kabuki syndrome, causing weakened humoral immunity. Our findings highlight the need for further characterization of CD4⁺ T cells in Kabuki syndrome, for which clinical data are currently limited to immune cell frequencies in the blood.

Limitations of the study

Our findings suggest that KMT2D expression within CD4⁺ T cells promotes T_FH development and antibody responses in part through H₃K₄Me₁ deposition at Thpok and Il21 enhancers. However, we were unable to directly test the contribution of KMT2D’s methyltransferase activity to gene expression due to a lack of readily available mouse models or specific chemical inhibitors targeting the KMT2D catalytic domain. We showed that KMT2D-mediated THPOK expression is sufficient for T_FH differentiation but did not test whether this extended to other aspects of T_FH and CD4⁺ T cell responses, such as CD40L and cytokine expression.

RESOURCE AVAILABILITY

Lead contact

Additional information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Jason K. Whitmire (jwhitmir@email.unc.edu).

Materials availability

KMT2D^fl/fl mice are available upon request with standard institutional material transfer agreements and evidence of institutional approval to receive mice.

Data and code availability

• All experimental data are available within this article and will be shared by the lead contact upon request. scRNA-seq data have been deposited to the Gene Expression Omnibus under accession GEO: GSE282033.

• No original code was developed for this study; however, the code utilized in the data processing and analyses is available upon request.

• Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

STAR★METHODS

EXPERIMENTAL MODEL AND STUDY PARTICIPANT DETAILS

Mice

All mouse colonies were maintained in-house at UNC by the Whitmire lab. C57BL/6J (B6), B6.Ly5a (CD45.1), and B6.LckCre mice were purchased from Jackson laboratories. SMARTA TCR-Tg and P14 TCR-transgenic mice were backcrossed with B6.Ly5a mice to yield SMARTA.Ly5a and P14.Ly5a mice. Kmt2d^fl/fl mice⁴⁹ were kindly provided by Dr. Kai Ge (NIDKK) and were crossed with B6.LckCre mice to generate KMT2D^TCD mice (LckCre^+/wt;Kmt2d^fl/fl). Both LckCre^posKmt2d^wt/wt and LckCre^negKmt2d^fl/fl mice were used as WT controls. SMARTA and P14 TCR-Tg mice were maintained on a B6.Ly5a genetic background and intercrossed with KMT2D^TCD mice to generate SMARTA.KMT2D^TCD and P14.KMT2D^TCD mice. The LckCre and TCR-Tg transgenes were maintained in a hemizygous state for all associated mouse lines (LckCre^+/o, SMARTA^+/o, P14^+/0). Both male and female mice were included in repeated experiments; there was no discernable sex difference in the effects we assessed. Age and sex-matched adult mice ranging from 8 to 14 weeks old were used for all studies. All experiments and procedures were approved by the University of North Carolina-Chapel Hill Institutional Animal Care and Use Committee.

Virus

Stocks of LCMV-Armstrong were generated by infecting BHK-21 cells with plaque-purified isolates. Virus-containing supernatants were aliquoted and frozen at −80°C. The plaque-forming titer was quantified by plaque assay on Vero cell monolayers. Viral stocks tested negative for mycoplasma.

Cell lines

Vero cells are a female African green monkey kidney epithelial cell line, which were used to quantify infectious LCMV by plaque assay. BHK-21 cells, a fibroblast cell line derived from hamster of unknown sex, were used to grow LCMV for viral stocks. Both Vero and BHK-21 cells were used at low passage number and were negative for mycoplasma. Plat-E cells are derivative of the 293T female human kidney cell line and were used to generate retrovirus that was transduced into T cells. Vero cells and BHK-21 cells were cultured in DMEM supplemented with 5% FBS (heat-inactivated), L-Glutamine, penicillin and streptomycin. Plat-E were cultured in 10% DMEM supplemented with HEPES, Pen/Strep, L-Glutamine, and Non-Essential Amino Acids.

METHOD DETAILS

Genotyping

LckCre and KMT2D^TCD mice were genotyped by PCR. DNA was isolated from PBMCs following overnight incubation with Proteinase K. Oligonucleotide primers were synthesized by Eurofins Genomics and used to PCR amplify DNA fragments, using the primers listed in Table and GoTaq PCR master mix. The PCR products were visualized by ethidium bromide staining after they were run on a 2% agarose gel. SMARTA and P14 TCR transgenic mice were genotyped by flow cytometry performed on PBMCs. CD4, Va2, and Vb8.3 identified SMARTA mice, and CD8, Va2, and Vb8.1 identified P14 mice. CD45 congenic background was determined by antibody labeling of CD45.1 and CD45.2.

Virus

Adult mice were intraperitoneally injected with 2 × 10⁵ PFU of LCMV Armstrong for analyses performed 7 or more days after infection, and 1 × 10⁶ pfu of LCMV Armstrong for analyses performed 3.5 days after infection. Viral titers in infected mice were quantified by plaque assay using Vero cell monolayers. Stocks of LCMV were generated by infecting BHK-21 cells with plaque purified isolates and tested negative for Mycoplasma.

PBMC isolation

Blood was collected by cheek bleed into a tube containing sodium citrate solution. Blood was underlaid with Ficoll, and then tubes were centrifuged at 2000 rpm for 20 min at room temperature with low acceleration/deceleration to separate PBMCs from other blood components.

Splenocyte isolation

Single cell splenocyte suspensions were prepared by mechanically disrupting spleens with scissors and a syringe plunger over a 70 μm strainer. RBC lysis was performed using ACK lysis buffer, followed by two washes with 1% RPMI. Live cells were counted with trypan blue on a hemocytometer. 2 × 10⁶ live cells were seeded per well of a 96 well plate. All isolation steps were carried out on ice or at 4°C.

Adoptive transfer experiments

All cell transfer studies were performed by tail vein injection into C57BL/6 recipients one day prior to infection of hosts with LCMV Armstrong, unless otherwise specified. Donor cell responses were analyzed in the spleen and blood of host mice by at the indicated time after infection. For single-transfer studies, donor cells were CD45.1 congenic. For co-transfer studies, one donor group was CD45.1 congenic, while the other was CD45.1 CD45.2 congenic.

P14 transfers

CD8⁺ P14 recovered from the spleens of WT and KMT2D^TCD P14 TCR transgenic mice. Co-transfers of 5 × 10³ total P14 cells, consisting of equal numbers of congenic WT and KMT2D^−/− P14 cells were performed.

SMARTA transfers

Splenocytes containing CD4⁺ SMARTA were recovered from WT and KMT2D^TCD SMARTA TCR transgenic mice. For analyses at 7–8 days post infection, single transfers of 1 × 10⁴ SMARTA cells into B6 recipients were performed. For analyses at 3.5 days post infection, co-transfers of 1 × 10⁶ total SMARTA cells, consisting of equal numbers of congenic WT and KMT2D^−/− SMARTA were performed.

Flow cytometry

Dead cells were labeled with fixable viability dye according to the manufacturer’s instructions (Cytek), and Fc receptors were blocked with anti-CD16/32 antibodies. Tetramer labeling was performed in cRPMI for 1 h at 4°C (MHC-I tetramers) or 37°C (MHC-II tetramer) and in some instances was performed simultaneously with CXCR5 labeling in a one-step stain. In other instances, CXCR5 labeling was performed as a two or three step stain for 1 h in FACS buffer. Next, the remaining surface markers were labeled with antibodies for 30 min at 4°C in FACS buffer. Transcription factors were labeled for 1 h with the FoxP3 Fixation/Permeabilization kit according to the manufacturer’s instructions. CD40L and Perforin staining was performed in Permeabilization/Wash Buffer (Biolegend) on splenocytes fixed with 4% PFA for 30 min. Intracellular cytokines were identified after 5-h stimulation with LCMV peptides GP_61–80, GP_33–41, NP_205–212 or NP_396–405 in the presence of Brefeldin A. Fixation and cytokine labeling was carried out the same as for CD40L. Fixation conditions used for identification of IL-21 were 2% PFA for 20 min. All labeling steps were carried out in 50ul of staining solution. Samples were washed 2–3 times following each staining step. Fixation and labeling were carried out at 4°C unless otherwise specified. Samples were acquired on a 5 laser LSR Fortessa, and data analysis was performed in FlowJo software (TreeStar).

Cell preparation for single cell RNA sequencing and Cut&Tag

For scRNA-seq, splenocyte single cell suspensions were prepared from WT and KMT2D^TCD mice at 10 dpi by mechanical disruption. CD4⁺ T cells were negatively enriched using biotinylated antibodies against CD8, B220, and TER-119 with MojoSort Streptavidin Nanobeads according to the manufacturer’s instructions. Enriched CD4⁺ T cells were stained with fixable viability dye, GP66-tetramer, CD4, and PD1 as described above. Samples were multiplexed by labeling each sample with a unique Biolegend TotalSeq-B Hashtag antibody-oligo conjugate for 30 min on ice according to the manufacturer’s instructions. Live virus-specific CD4⁺ T cells (CD4+GP66+ cells) were sorted using the Sony SH800Z cell sorter (100um nozzle). Post-sort cell viability was determined by Propidium Iodide/Acridine Orange staining on a Luna-FX7 cell counter and was between 85% and 96% per sample. For CUT&Tag, splenocyte single cell suspensions were prepared from WT and KMT2D^TCD mice at 9 dpi as described above. Total splenocytes were stained with fixable viability dye, and antibodies against CD4, CD8 and CD44. Live CD4⁺CD44^hiCD8-cells were sorted using the SH800Z and kept in complete RPMI on ice prior to CUT&Tag.

CUT&Tag

CUT&Tag was performed for H₃K₄Me₁ utilizing pAG-Tn5 according to manufacturer guidelines on sorted CD44^hiCD4+ T cells. DNA was extracted from 5% of the cell population to serve as input. Briefly, nuclei were extracted in HEPES (20 mM, pH 7.9), KCL (10 mM), Triton X-100 (0.1%), glycerol (20%), spermidine (0.5 mM), and protease inhibitor. Approximately 50,000 nuclei were bound to Concanavalin A magnetic beads for 10 min at room temperature. Nuclei/beads were incubated with H₃K₄Me₁ antibody in Digitonin Wash buffer (20 mM HEPES, pH 7.5), NaCl (150 mM), spermidine (0.5 mM), digitonin (0.01%), and protease inhibitor containing 2 mM EDTA overnight at 4°C with elevated cap on a rocker. Nuclei/beads were then incubated with secondary antibody for 30 min at room temperature on a rocker. Nuclei/beads were washed twice in Digitonin wash buffer and then incubated with pAG-Tn5 in high salt digitonin wash buffer and rocked for 1 h at room temperature. After washing and resuspension in high salt digitonin buffer, MgCl₂ (10 mM) was added for tagmentation for 1 h at 37°C in a thermocycler. After wash in TAPS buffer (10 mM, pH 8.5) containing EDTA (0.2 mM), DNA was extracted in 10 mM TAPS containing SDS (0.1%) for 1 h at 58°C in a thermocycler. The extraction was quenched with Triton X-100 (0.5%) and amplified for 14 cycles with NEBNext HiFi PCR mix, using dual index oligos that anneal to Tn5 adapter sequences.¹ Amplified samples were purified with KAPA pure SPRI beads, and eluted in Tris-HCl (10 mM, pH 8). One WT sample was discarded due to low input DNA concentration and lack of nucleosomal banding patterns by tape station analysis.

Quantitative PCR

qPCR was performed on a CFX96 thermocycler using the following cycling conditions: (1) initial denaturation at 95°C for 3 min, (2) 39 cycles of 95°C for 5 s and 60°C for 30 s. Each reaction was in 10 μl total volume, and contained 5 μl SYBR Green Master Mix, forward and reverse primer (500 nM each), and template DNA obtained from H₃K₄Me₁ Cut&Tag (1:20 dilution) (see above). Melt-curve analysis was performed immediately after amplification to confirm the specificity of the PCR products. Each group contained 3–4 biological replicates, and reactions were run in duplicate. Enrichment of H₃K₄Me₁ was calculated as ΔCt, with input DNA as the reference. qPCR primers for Thpok GTE, Il21 CNS-49, and Kmt2d Exon 50–51 were previously described.^49,86,87 qPCR primers were designed for Thpok Proximal Enhancer and ThPok site A based on known genomic coordinates.^88,89

Single-cell RNA-seq and bioinformatic analyses

Sorted virus-specific CD4⁺ T cells from each sample were pooled together in roughly equal numbers, and ~20k cells were converted into single-cell cDNA libraries using the 10x Genomics Chromium GEM-X Single Cell 3^′ V4 kit. The multiplexed library was loaded onto the NextSeq P4 flow cell, and paired-end sequencing was performed on an Illumina NextSeq 2000 at the UNC High Throughput Sequencing core facility. Reads were mapped to genome build GRcM39 and demultiplexed through the 10x Genomics Cell Ranger multi pipeline (v8.0.0).⁸³

Raw unique molecular identifier (UMI) counts for each demultiplexed sample were imported into R (v4.4.0) for analysis using Seurat (v5.1.0).⁹⁰ Quality control (QC) filtering of low-quality cells was performed using the miQC (v1.12.0)⁹¹ package based on the flexmix model with a posterior cutoff of 0.75. Additional manual thresholds were applied to exclude cells with RNA counts above the 99^th percentile and features detected above the 99^th or below the 1^st percentile. Post QC processing was performed using Seurat. Raw UMI counts were log normalized using the LogNormalize method with a scale factor of 10,000 and then scaled to variance with mitochondrial gene percentage regressed out.

Dimensionality reduction by uniform manifold approximation and projection (UMAP) was performed following PCA reduction using 41 principal components (PCs). The number of PCs was determined based on the point that PCs account for 90% of the cumulative variance and less than 5% of the standard deviation. Analysis of cell cycle genes indicated they were not major drivers of PC segregation on the UMAP, and thus were not removed by regression for downstream analysis. To facilitate comparison between groups, samples were integrated using Harmony (v1.2.1).⁹² The integrated data were clustered with the Louvain with multilevel refinement algorithm in Seurat at a resolution of 0.61. Clusters were visualized by UMAP.

Pseudobulk differential expression analyses were performed by summing counts across all cells within each cluster and sample. Differential expression between WT and KMT2D^TCD clusters was assessed using DESeq2 (v1.44.0).⁹³ Genes with a false discovery rate (FDR)-adjusted p value of less than 0.05 were considered significantly differentially expressed. Gene signatures for T_FH, T_H1 and Tcmp cells were obtained from Ciucci et al.,⁶⁰ and gene signature scores were calculated using Seurat AddModule function for each gene list.

To relate gene expression changes in our scRNA-seq data to transcriptional regulation of genes by ThPok, Tet2, and Tbet, we re-analyzed publicly available bulk RNA-sequencing data. For ThPok, RNA-sequencing was performed on virus-specific WT (Ox40Cre+ ThPok^wt/wt; GP66+ CD4⁺ CXCR5^hiPD1^hi) and ThPok-deficient (Ox40Cre+ ThPok^fl/fl; GP66+ CD4⁺ CXCR5^intPD1^int) T_FH cells that were sorted from the spleen of LCMV-Armstrong infected mice at 7 dpi (Vacchio et al.¹⁴). For Tet2, WT (CD45^.1/.1 CXCR5+) and Tet2^−/− (CD4Cre+ Tet2^fl/fl; CD4⁺ CD45^.1/.2 CXCR5+) donor SMARTA T_FH cells were sorted from the spleen of LCMV-Armstrong infected mice at 7 dpi (Baessler et al.⁶⁹). For Tbet, WT (B6; CD4⁺ Thy1.1+ PSGL1^lo Ly6C^lo CXCR5^hiPD1^hi) and Tbet^−/− (B6.Tbx21^−/−; CD4⁺ Thy1.1+ PSGL1^loLy6C^lo CXCR5^hiPD1^hi) donor SMARTA T_FH cells were sorted from the spleen 8 days after LCMV infection (Weinstein et al. ⁶⁸).

Differential expression analyses were performed using DESeq2 on raw count matrices obtained from GSE130474 and GSE183316 for the ThPok and Tet2 datasets respectively. For the Tbet dataset, sequencing files were downloaded from GSE105806. Adaptor trimming and removal of low-quality reads was performed using Trim Galore (v0.6.7). Reads were then aligned to the mm10 genome using STAR (v2.7.11b) to obtain raw gene counts followed by differential expression analysis using DESeq2. Genes with an FDR adjusted p value of less than 0.05 were considered significantly differentially expressed.

To compare genes regulated by KMT2D versus ThPOK, Tet2, and Tbet in T_FH cells, the log2 fold-change in expression of genes from pseudobulk (KMT2D^TCD cluster 1/WT cluster 1) and bulk (e.g., ThPok-deficient T_FH/WT T_FH) RNA-seq datasets were Z-normalized using the R scale function. The Z scores of cluster 1 DEGs from pseudobulk and bulk data was depicted on a scatterplot, and a Pearson correlation coefficient of Z scores was calculated.

Retrovirus production

MSCV-IRES-GFP plasmid vector was purchased from Addgene. MSCV-Thpok-t2a-GFP and MSCV-Tcf1-t2a-GFP vectors were kindly provided by Dr. Justin Milner (UNC-Chapel Hill). Plat-E cells were used to generate retrovirus following transfection with MSCV vector and pCL-Eco plasmids using Transit-LT1 reagent according to the manufacturer’s instructions. 24 h later, media were changed to cRPMI. Retroviral supernatant was collected 48 and 72 h after Plat-E cell transfection, and stored at −80°C.

Retroviral transductions

CD4⁺ SMARTA cells were enriched from the spleens of CD45 congenic SMARTA mice on a KMT2D^fl/fl and KMT2D^TCD genetic background using the MojoSort CD4 T cell isolation kit and then pooled together in equal numbers. Pooled SMARTA cells were resuspended in RPMI supplemented with 10% heat-inactivated FCS, hIL-2 (100u/ml), and in vitro activated with plate bound anti-CD3/CD28 antibodies (5 μg/ml each) on a 6 well plate for 24 h in a 37°C incubator. Next, retroviral supernatant containing polybrene (4 μg/ml final concentration) was added to each well, and pooled SMARTA cells were spin transduced for 90 min at 2000g at 32°C. After spin-transduction, pooled SMARTA were incubated for an additional 5–6 h in a 37°C incubator and then washed twice with 0% RPMI. Finally, 2 × 10⁴ transduced SMARTA cells (containing 1 × 10⁴ SMARTA from KMT2D^wt/wt and KMT2D^fl/fl backgrounds) were transferred by tail vein injection into C57BL/6 mice that were infected with LCMV Armstrong one day earlier. Donor SMARTA cells were distinguished from each other and host cells by expression of CD4, CD45.1 and CD45.2. Donor cells which were successfully transduced, were identified by GFP expression.

QUANTIFICATION AND STATISTICAL ANALYSIS

Unpaired or paired Student’s t tests (parametric test) were used to assess the significance of normally distributed data, as indicated in each legend. An unpaired Mann-Whitney U test was used for non-normally distributed data. Statistical analyses and graphing were performed using Prism software (GraphPad Software Inc.). DESeq2 was used to identify differentially expressed genes using scRNA-seq and bulk RNA-seq data, with significant gene cutoff defined at FDR < 0.05. p values considered significant are indicated in the figure legends as: *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001. All data are presented as mean ± SEM values. Each figure legend indicates the number of mice that were analyzed and the number of repeated experiments.

Supplementary Material

SUPPLEMENTAL INFORMATION

Supplemental information can be found online at https://doi.org/10.1016/j.celrep.2025.115775.

KEY RESOURCES TABLE.

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
anti-mouse-B220-Biotin	Biolegend	Cat #103203; RRID:AB_312988
anti-mouse-B220-Pacific Blue	Biolegend	Cat #103230; RRID:AB_492877
anti-mouse-B220-FITC	Biolegend	Cat #103206; RRID:AB_312991
anti-mouse-B220-Violet Fluor 500	Cytek	Cat #850452U025
anti-mouse-B220-PerCP	Biolegend	Cat #103233; RRID:AB_893355
anti-mouse-B220-Spark Plus UV395	Biolegend	Cat #103297; RRID:AB_3097645
anti-mouse-BCL2-PE/Cy7	Biolegend	Cat #633511; RRID:AB_2565246
anti-mouse-Bcl6-PE	BD Biosciences	Cat #561522; RRID:AB_10717126
anti-mouse-Bcl6-PE/Cy7	BD Biosciences	Cat #563582; RRID:AB_2738292
anti-mouse-BCL6-Real Blue 780	BD Biosciences	Cat #569143; RRID:AB_3662681
anti-mouse-Bim-unconjugated	Cell Signaling	Cat #2933S; RRID:AB_1030947
anti-mouse-CD3-Purified	Biolegend	Cat #100340; RRID:AB_11149115
anti-mouse-CD4-SparkUV387	Biolegend	Cat #100492; RRID:AB_2922449
anti-mouse-CD4-SparkPlusUV395	Biolegend	Cat #100495; RRID:AB_3097649
anti-mouse-CD4-BV421	Biolegend	Cat #100563; RRID:AB_2563052
anti-mouse-CD4-Violet Fluor 500	Cytek	Cat #850041U025
anti-mouse-CD4-FITC	Biolegend	Cat #100510; RRID:AB_312713
anti-mouse-CD4-PE	Biolegend	Cat #100512; RRID:AB_312715
anti-mouse-CD4-APC	Biolegend	Cat #100516; RRID:AB_312719
anti-mouse-CD4-Alexa Fluor 700	Biolegend	Cat #100536; RRID:AB_493701
anti-mouse-CD8a-Biotin	Biolegend	Cat #100703; RRID:AB_312742
anti-mouse-CD8a-SparkUV387	Biolegend	Cat #100797; RRID:AB_2922451
anti-mouse-CD8a-SparkPlus UV395	Biolegend	Cat #104011; RRID:AB_3097648
anti-mouse-CD8a-APC	Biolegend	Cat #100712; RRID:AB_312751
anti-mouse-CD8a-Alexa Fluor 700	Biolegend	Cat #100729; RRID:AB_493702
anti-mouse-CD8a-BV421	Biolegend	Cat #100753; RRID:AB_2562558
anti-mouse-CD8a-VioletFluor500	Cytek	Cat #850081U025
anti-mouse-CD8b-Biotin	Biolegend	Cat #126604; RRID:AB_961291
anti-mouse-CD28 -Purified	Biolegend	Cat #102116; RRID:AB_11147170
anti-mouse-CD40L-PE	Biolegend	Cat #106505; RRID:AB_313270
anti-mouse-CD40L-APC	Biolegend	Cat #106509; RRID:AB_2561560
anti-mouse-CD44-FITC	Biolegend	Cat #103005; RRID:AB_312956
anti-mouse-CD44-Alexa Fluor 700	Biolegend	Cat #103026; RRID:AB_493713
anti-mouse-CD44-BV711	Biolegend	Cat #103057; RRID:AB_2564214
anti-mouse-CD44-BV711	BD Biosciences	Cat #563971; RRID:AB_2738518
anti-mouse-CD45.1-eFluor450	Thermofisher	Cat #48-0453-82; RRID:AB_1272189
anti-mouse-CD45.1-eFluor506	Thermofisher	Cat #69-0453-82; RRID:AB_2637182
anti-mouse-CD45.1-FITC	Biolegend	Cat #110706; RRID:AB_313495
anti-mouse-CD45.1-PE	Biolegend	Cat #110708; RRID:AB_313497
anti-mouse-CD45.1-APC	Biolegend	Cat #110714; RRID:AB_313503
anti-mouse-CD45.2-eFluor450	Thermofisher	Cat #48045482; RRID:AB_11042125
anti-mouse-CD45.2-FITC	Biolegend	Cat #109806; RRID:AB_313443
anti-mouse-CD45.2-PE	Biolegend	Cat #109808; RRID:AB_313445
anti-mouse-CD45.2-APC	Biolegend	Cat #109814; RRID:AB_389211
anti-mouse-CD45.2-Alexa Fluor 700	Biolegend	Cat #109822; RRID:AB_493731
anti-mouse-CD62L-FITC	Biolegend	Cat #104405; RRID:AB_313092
anti-mouse-CD69-BV711	Biolegend	Cat #104537; RRID:AB_2566120
anti-mouse-CD86-PE/Cy7	Biolegend	Cat #Check; RRID:AB_439782
anti-mouse-CD138-PE/Cy	Biolegend	Cat #142513; RRID:AB_2562197
anti-mouse-CD150-APC	Biolegend	Cat #115909; RRID:AB_493461
anti-mouse-CD150-Alexa Fluor 488	Biolegend	Cat #115915; RRID:AB_528743
anti-mouse-CXCR3-PE	Biolegend	Cat #126505; RRID:AB_1027656
anti-mouse-CXCR4-Biotin	Thermofisher	Cat #13999180; RRID:AB_10596966
anti-mouse-CXCR5-Purified	BD Biosciences	Cat #551961; RRID:AB_394302
anti-mouse-CXCR5-Biotin	BD Biosciences	Cat #551960; RRID:AB_394301
anti-mouse-CXCR5-PE	Thermofisher	Cat #12718582; RRID:AB_11217882
anti-mouse-CXCR5-Biotin	Thermofisher	Cat #13-7185-82; RRID:AB_2572800
anti-mouse-CXCR6-PE/Cy7	Biolegend	Cat #151119; RRID:AB_2721670
anti-mouse-EOMES-PE	Thermofisher	Cat #12-4875-82; RRID:AB_1603275
anti-mouse-EOMES-APC	Thermofisher	Cat #17-4875-80; RRID:AB_2866427
anti-mouse-Fas-PE	Biolegend	Cat #152608; RRID:AB_2632902
anti-mouse-FasL-Biotin	Thermofisher	Cat #12-5911-82; RRID:AB_466794
anti-mouse-FoxP3-FITC	Thermofisher	Cat #11-5773-80; RRID:AB_465242
anti-mouse-FR4-Real Blue 780	BD Biosciences	Cat #755354; RRID:AB_3687749
anti-mouse-GL7 -APC	Biolegend	Cat #144618; RRID:AB_2800675
anti-mouse-GzmB-FITC	Biolegend	Cat #515403; RRID:AB_2114575
anti-mouse- H3K4Me1 -Purified	Cell Sign.	Cat #5326S; RRID:AB_10695148
anti-mouse-ICOS-PE/Cy7	Biolegend	Cat #117421; RRID:AB_2860636
anti-mouse-IFN-γ-FITC	Biolegend	Cat #505806; RRID:AB_315400
anti-mouse-IFN-γ-APC	Biolegend	Cat #505810; RRID:AB_315404
anti-mouse-IFN-γ-PE/Cy7	Biolegend	Cat #505826; RRID:AB_2295770
anti-mouse-IgD-FITC	Biolegend	Cat #405703; RRID:AB_315025
anti-mouse-IgD-Alexa Fluor 700	Biolegend	Cat #405729; RRID:AB_2563340
anti-mouse-IgG1-FITC	BD Biosciences	Cat #562026; RRID:AB_10926376
anti-mouse-IL-2-APC	Biolegend	Cat #503809; RRID:AB_315303
anti-mouse-IL7Ra-PE/Cy7	Biolegend	Cat #135013; RRID:AB_1937266
anti-mouse-IRF4-PE/Cy7	Thermofisher	Cat #25-9858-82; RRID:AB_2573558
anti-mouse-Ki67 Set-FITC	BD Biosciences	Cat #Cat# 556026; RRID:AB_396302
anti-mouse-Ki67-Alexa Fluor 700	Biolegend	Cat #16A8; RRID:AB_2564285
anti-mouse-KLRG1-APC	Biolegend	Cat #138412; RRID:AB_10641560
anti-mouse-KLRG1-PE	Biolegend	Cat #138408; RRID:AB_10574313
anti-mouse-Lag3-PE/Cy7	Biolegend	Cat #125225; RRID:AB_2715763
anti-mouse-LEF1-Alexa Fluor 488	Cell Signaling	Cat #8490S; RRID:AB_10949502
anti-mouse-Ly108-APC	Biolegend	Cat #134610; RRID:AB_2728155
anti-mouse-Ly6C-FITC	Biolegend	Cat #128005; RRID:AB_1186134
anti-mouse-Ly6C-PE	Biolegend	Cat #128008; RRID:AB_1186132
anti-mouse-Ly6C-APC	Biolegend	Cat #128016; RRID:AB_1732076
anti-mouse-PD1-BV421	Biolegend	Cat #109121; RRID:AB_2687080
anti-mouse-PD1-PE	Biolegend	Cat #109104; RRID:AB_313421
anti-mouse-PD1-APC	Biolegend	Cat #109111; RRID:AB_10613470
anti-mouse-PD1-Alexa Fluor 488	BD Biosciences	Cat #568577; RRID:AB_3684384
anti-mouse-Perforin-APC	Biolegend	Cat #154303; RRID:AB_2721462
anti-mouse-PSGL1-Alexa Fluor 647	BD Biosciences	Cat #562806; RRID:AB_2737807
anti-mouse-RORgt-Real Yellow 775	BD Biosciences	Cat #571372; RRID:AB_3686452
anti-mouse-Tbet-PE	Biolegend	Cat #644810; RRID:AB_2200542
anti-mouse-Tbet-Real Blue 780	BD Biosciences	Cat #569090; RRID:AB_3684770
anti-mouse-TCF1-Alexa Fluor 647	BD Biosciences	Cat #566693; RRID:AB_2869823
anti-mouse-TER-119-Biotin	Biolegend	Cat #116203; RRID:AB_313704
anti-mouse-ThPok-Real Blue 780	BD Biosciences	Cat #568697; RRID:AB_3684476
anti-mouse-ThPok-Real Yellow 586	BD Biosciences	Cat #568129; RRID:AB_3684061
anti-mouse-Thy1.2-FITC	Biolegend	Cat #105306; RRID:AB_313177
anti-mouse-Thy1.2-Alexa Fluor 700	Biolegend	Cat #105319; RRID:AB_493724
anti-mouse-Tim3-BV711	Biolegend	Cat #119727; RRID:AB_2716208
anti-mouse-TNF-BV421	Biolegend	Cat #506327; RRID:AB_10900823
anti-mouse-TNF-Alexa Fluor 488	Biolegend	Cat #506315; RRID:AB_493329
anti-mouse-TNF-PE	Biolegend	Cat #506306; RRID:AB_315427
anti-mouse-UTX-Purified	Cell Sign.	Cat #33510; RRID:AB_2721244
anti-mouse-Va2-PE	Biolegend	Cat #127808; RRID:AB_1134183
anti-mouse-Vb8.1-FITC	Biolegend	Cat #118406; RRID:AB_1227786
anti-mouse-Vb8.3-FITC	Biolegend	Cat #156306; RRID:AB_2800702
anti-human IgG Fab2(H + L)-Alexa Fluor 647	Jackson	Cat #709-606-149; RRID:AB_2340581
anti-rabbit-IgG	Thermofisher	Cat #NBP172763; RRID:AB_11024108
IL21R-Fc chimera-Purified	R&D	Cat #596-MR-100
Streptavidin BV421	Biolegend	Cat #4414S
Streptavidin PE	Biolegend	Cat #405237
Streptavidin APC	Biolegend	Cat #405207
Streptavidin Alexa Fluor 647	Biolegend	Cat #405237
Ghost Dye Red 780 Viability dye	Cytek	Cat #13-0865-T100
Bacterial and virus strains
LCMV-Armstrong	Whitmire Lab	N/A, generated in house
Chemicals, peptides, and recombinant proteins
Streptavidin BV421	Biolegend	Cat #4414S
Streptavidin PE	Biolegend	Cat #405237
Streptavidin APC	Biolegend	Cat # 405207
Streptavidin Alexa Fluor 647	Biolegend	Cat # 405237
I-AbGP66-77 Tetramer-BV421	NIH Tetramer Core	N/A
I-AbGP66-77 Tetramer-APC	NIH Tetramer Core	N/A
I-Ab-control Tetramer-BV421	NIH Tetramer Core	N/A
I-Ab-control Tetramer-APC	NIH Tetramer Core	N/A
DbGP33-41 Tetramer-APC	NIH Tetramer Core	N/A
DbGP33-41 Tetramer-BV421	NIH Tetramer Core	N/A
DbGP396-404 Tetramer-APC	NIH Tetramer Core	N/A
DbGP396-404 Tetramer-BV421	NIH Tetramer Core	N/A
2-mercaptoethanol	Sigma-Aldrich	Cat# M7522
ACK lysing Buffer	Lonza	Cat# 10-548E
Agar:Agar	EMD	Cat# 1.01614.0500
Bovine Serum Albumin	Sigma-Aldrich	Cat# A4503
Brefeldin A Solution (1000x)	Biolegend	Cat# 420601
Cell strainer (30um)	Sysmex	Cat# NC9682496
Cell strainer (70um)	Greiner	Cat# 542070
DH5-alpha \ (Subcloning efficiency)	LifeTechnologies	Cat# 18265017
DMEM	Lonza	Cat# 12-614F
DMEM	Cytiva	Cat# SH30285.FS
EcoRI-HF	New England Biolabs	Cat# R3101S
EMEM	Sigma-Aldrich	Cat# 56416C
Ethidium Bromide	Calbiochem	Cat# 18H235208
Fetal Bovine Serum	VWR	Cat# 97068-085
Ficoll-Paque PLUS	Cytiva	Cat# 17144002
Fixation Buffer	Biolegend	Cat# 420801
FoxP3 Transcription Factor Staining Kit	eBioscience	Cat# 00-5523-00
KAPA pure SPRI beads	Rocke	Cat# 7983271001
GoTaq Green Master Mix	Promega	Cat# M7123
HEPES 1M	Gibco	Cat# 15630-090
L-Glutamine	Lonza	Cat# 17-605L
Max Efficiency Stbl2	LifeTechnologies	Cat# 10268019
MEM Non-Essential Amino Acids	Gibco	Cat# 11140-050
NEB5-alpha competent E. Coli (High Efficiency)	New England Biolabs	Cat# C2987H
NEBNext HiFi 2X PCR mix	New England Biolabs	Cat# M0541L
Neutral Buffered Formalin (10%)	Thermofisher	Cat# 22-050-104
PBS (10x)	Gibco	Cat# 70011-044
PBS (1x)	VWR	Cat# SH30028.01
Pen/Strep	VWR	Cat# BW17-602E
Percoll	GE Healthcare	Cat# 17-0891-02
Permeabilization Wash Buffer	Biolegend	Cat# 421002
Poly-D-Lysine	Gibco	Cat# M7522
Polybrene	Sigma	Cat# 10-548E
Protease inhibitor	Sigma	Cat# 11836170001
Proteinase K	Thermofisher	Cat# check
Recombinant Human IL-2	Peprotech	Cat# A4503
RPMI 1640	Lonza	Cat# 420601
RPMI 1640	VWR	Cat# NC9682496
Sodium Pyruvate (100mM)	Gibco	Cat# 542070
TAE buffer (10x)	Corning	Cat# 18265017
Trypan Blue	Lonza	Cat# 12-614F
UltraPure Agarose	Invitrogen	Cat# SH30285.FS
XhoI	New England Biolabs	Cat# R3101S
Zymo DNA Clean and Concentrator	Zymo	Cat# 56416C
Zymopure II Plasmid Midi Prep	Zymo	Cat# 18H235208
Zymopure II Plasmid Mini Prep	Zymo	Cat# 97068-085
Critical commercial assays
ArC™ Compensation Bead Kit	Thermofisher	Cat# A10346
MojoSort™ mouse CD4 T cell Isolation Kit	Biolegend	Cat# 480033
MojoSort™ Streptavidin Nanobeads	Biolegend	Cat# 480016
UltraComp eBeads™	Thermofisher	Cat# 01-2222-42
UltraComp eBeads™ Plus	Thermofisher	Cat# 01-3333-42
anti-mouse TotalSeq™ Hashtag 1	Biolegend	Cat# 155831; RRID:AB_2814067
anti-mouse TotalSeq™ Hashtag 2	Biolegend	Cat# 155833; RRID:AB_2814068
anti-mouse TotalSeq™ Hashtag 3	Biolegend	Cat# 155835; RRID: AB_2814069
anti-mouse TotalSeq™ Hashtag 6	Biolegend	Cat# 155841; RRID:AB_2814072
pAG-Tn5	Epicypher	Cat# 15-1017
Concanavalin A magnetic beads	Cell Signaling	Cat# 93569S
Deposited data
scRNA-seq: CD4+ T cells; Spleen	This paper	GEO: GSE282033
Experimental models: Cell lines
Vero-E6	Michael Buchmeier	The Scripps Research Institute, La Jolla, CA
BHK-21	American Type Culture Collection	Cat# CCL-10
Plat-E Cells	CellBioLabs (kindly provided by Dr. Milner, UNC)	Cat# RV-101
Experimental models: Organisms/strains
Mouse: C57BL/6J	Jackson Laboratory (bred at UNC)	Cat# 000664
Mouse: B6.Ly5a	Jackson Laboratory (bred at UNC)	Cat# 002014
Mouse: LckCre	Jackson Laboratory (bred at UNC)	Cat# 003802
Mouse: Kmt2dfl/fl	Bred in house	Jang et al.49
Mouse: SMARTA TCR-Tg (B6.CD45.1)	Backcrossed in Whitmire Lab	Oxenius et al.81
Mouse: P14 TCR-Tg (B6.CD45.1)	Backcrossed in Whitmire Lab	Pircher et al.82
Oligonucleotides
Genotyping for Lck-Cre (FW)	TGCAACGAGTGATGAGGTTC	Eurofins
Genotyping for Lck-Cre (RV)	ACAGCATTGCTGTCACTTGG	Eurofins
Genotyping for KMT2Dfl/fl (FW1)	ATAAACGGTGAGCTTGTGCG	Eurofins
Genotyping for KMT2Dfl/fl (FW2)	AAGTTCCATAGCCATTGCTC	Eurofins
Genotyping for KMT2Dfl/fl (RV)	TTTCACCACTGTCGACCAAG	Eurofins
Thpok GTE (FW)	AGCCCTGGGTTCTTCTCTTC	Eurofins
Thpok GTE (RV)	CCTCCTTTCTTTCCCCTTTG	Eurofins
Thpok Proximal Enhancer (FW)	TGAGCCACATGTCACTAGAG	Eurofins
Thpok Proximal Enhancer (RV)	GAAGTTTAGAGGACCTGGAG	Eurofins
Thpok Site A (FW)	CCAGCCACCAAAAGTTCATC	Eurofins
Thpok Site A (RV)	TGTGAGGGTACAGGAAGTAC	Eurofins
Il21 CNS-49 (FW)	GGACCTCAACACTACCTTTTGT	Eurofins
Il21 CNS-49 (RV)	CCTTCCCATCACTGCTCATG	Eurofins
Hoxb1 (FW)	AGTCGGAAGGAGATGGAAGC	Eurofins
Hoxb1 (RV)	CTGTCTTAGGTGGGTTTCTC	Eurofins
Gene desert region (FW)	TTTCCTGCCTCTGCCTTTTA	Eurofins
Gene desert region (RV)	AGGTGGCACAGTGGGTAAAG	Eurofins
Kmt2d exon 50-51 (FW)	GGGTGGAGAGCTGTCAGAATTATT	Eurofins
Kmt2d exon 50-51 (RV)	CATGAGCGGTAACTCCATCAGA	Eurofins
Recombinant DNA
MSCV-IRES-eGFP	Addgene	Cat# 20672
MSCV-ThPok-t2a-eGFP	Kindly provided by Dr. Justin Milner	N/A
MSCV-Tcf1 (p33) - t2a-eGFP	Kindly provided by Dr. Justin Milner	N/A
Software and algorithms
Cell Ranger	10x Genomics	Zheng et al.83
FlowJo v10	Treestar	https://www.flowjo.com
GraphPad Prism v10	Dotmatics	https://www.graphpad.com
IGV v2.18.2	Robinson et al.84	http://www.igv.org/
Loupe Browser v8.0.0	10x Genomics	https://www.10xgenomics.com
R Studio		N/A
Seurat v5.1.0	Stuart et al.85	http://www.satijalab.org/seurat; RRID: SCR_016341

High lights.

• KMT2D promotes antiviral CD4⁺ and CD8⁺ T cell cytokine expression

• CD4⁺ T cell expression of KMT2D enhances T_FH formation and antibody responses

• KMT2D limits the formation of cytolytic CD4⁺ T cells