Histone deacetylase 3 represses cholesterol efflux during CD4+ T-cell activation

Abstract

After antigenic activation, quiescent naive CD4⁺ T cells alter their metabolism to proliferate. This metabolic shift increases production of nucleotides, amino acids, fatty acids, and sterols. Here, we show that histone deacetylase 3 (HDAC3) is critical for activation of murine peripheral CD4⁺ T cells. HDAC3-deficient CD4⁺ T cells failed to proliferate and blast after in vitro TCR/CD28 stimulation. Upon T-cell activation, genes involved in cholesterol biosynthesis are upregulated while genes that promote cholesterol efflux are repressed. HDAC3-deficient CD4⁺ T cells had reduced levels of cellular cholesterol both before and after activation. HDAC3-deficient cells upregulate cholesterol synthesis appropriately after activation, but fail to repress cholesterol efflux; notably, they overexpress cholesterol efflux transporters ABCA1 and ABCG1. Repression of these genes is the primary function for HDAC3 in peripheral CD4⁺ T cells, as addition of exogenous cholesterol restored proliferative capacity. Collectively, these findings demonstrate HDAC3 is essential during CD4⁺ T-cell activation to repress cholesterol efflux.

Keywords: HDAC3; T cell activation; cholesterol regulation; immunology; inflammation; mouse.

Figure 1.. CD8⁺ T cells have intrathymic deletion of histone deacetylase 3 (HDAC3) in dLck-HDAC3 cKO, but CD4⁺ T cells initiate deletion at the recent thymic emigrant (RTE) stage.

(a) Profile of primary splenic T-cell populations from wild-type (WT) and dLck-Cre HDAC3 cKO mice including total T cells (TCRβ⁺), total CD4⁺ (TCRβ⁺ CD4⁺), and total CD8⁺ (TCRβ⁺ CD8⁺), as well as memory (CD44^hi CD62L^lo) and naive (CD44^lo CD62L^hi) from each of the CD4⁺ and CD8⁺ populations. Bar graph depicts mean ± standard deviation (SD). Total cell number from three independent experiments (n = 5 mice/group). Statistical significance was determined for the indicated comparisons with Mann–Whitney tests between each WT and cKO population. (b) Expression of HDAC3 in thymocyte populations from WT, dLck-Cre HDAC3 cKO and CD4-Cre HDAC3 cKO mice. Thymic populations are gated as in Figure 1—figure supplement 1, and quantification of normalized HDAC3 MFI (median fluorescent intensity) ± SD from three independent experiments is shown on the right (n = 2–4 mice/group). Statistical significance was determined for the indicated comparisons with a one-way analysis of variance (ANOVA) with Tukey’s multiple comparisons test. (c) Profile of HDAC3 deletion in dLck-Cre HDAC3 cKO mice. Splenocytes were gated on key populations including total CD4⁺ (CD4⁺), naive CD4⁺ (CD4⁺ CD44^lo CD62L^hi), and memory CD4⁺ (CD4⁺CD44^hiCD62L^lo). Naive cells were further gated in mature naive T cells (MNTs; CD45RB^hiCD55^hi) or recent thymic emigrants (RTEs; CD45RB^loCD55^lo). HDAC3-positive and -negative population frequencies are shown below. Bar chart on right quantifies the mean frequency ± SD of HDAC3⁺ or HDAC3⁻ events within each population (n = 8 mice/group from three independent experiments).

Figure 1—figure supplement 1.. dLck-Cre histone deacetylase 3 (HDAC3) cKO mice have normal thymocyte population numbers.

(a) CD4/CD8 profile in thymocytes from wild-type (WT) and dLck-Cre HDAC3 cKO. Gates for double negative (DN; CD4⁻ CD8⁻), double positive (DP; CD4⁺ CD8⁺), CD4 single positive (CD4 SP; CD4⁺ CD8⁻), and CD8 single positive (CD8 SP; CD4⁻ CD8⁺) are shown. Quantification of total cell number ± standard deviation (SD) from three independent experiments is shown on the right (n = 2–5 mice/group).

Figure 2.. Histone deacetylase 3 (HDAC3)-deficient CD4⁺ T cells from dLck-Cre HDAC3 cKO mice have reduced differentiated T_h-cell populations.

Identification of helper T-cell populations in vivo. Splenocytes were harvested from wild-type (WT) and HDAC3 cKO mice, and labeled for flow cytometry. Cells were first gated on HDAC3⁺ or HDAC3⁻ events, then gating for T_h1 (T-bet⁺), T_h2 (GATA3⁺), T_h17 (RORγt⁺), T_reg (Foxp3⁺ CD25⁺), and T_fh (CXCR5⁺ PD-1⁺ Bcl-6^hi) is shown (left). Bar plots on right represent pooled data for the total cell number ± standard deviation (SD) from three independent experiments (n = 4–5 mice/group in total). Non-T_fh CXCR5⁻ PD-1⁻ cells (dark gray histograms) were used as a negative control for Bcl-6 expression to set the gate on the Bcl-6 histograms. Statistical significance was determined for the indicated comparisons using ordinary one-way analysis of variance (ANOVA) with Tukey’s multiple comparisons test (*p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001).

Figure 2—figure supplement 1.. Histone deacetylase 3 (HDAC3)-deficient T cells from dLck-Cre HDAC3 cKO mice have reduced differentiated T_h-cell populations.

Identification of helper T-cell populations in vivo. Lymphocytes were harvested from mesenteric lymph nodes of wild-type (WT) and dLck-Cre HDAC3 cKO mice, and labeled for flow cytometry. Cells were first gated on HDAC3⁺ or HDAC3⁻ events, then gating for T_h1 (T-bet⁺), T_h2 (GATA3⁺), T_h17 (RORγt⁺), T_reg (Foxp3⁺ CD25⁺), and T_fh (CXCR5⁺ PD-1⁺ Bcl-6^hi) is shown (left). Bar plots on right represent pooled data for the frequency of CD4⁺ T cells ± standard deviation (SD) from three independent experiments (n = 4–5 mice/group). Non-T_fh CXCR5⁻ PD-1⁻ cells (dark gray histograms) were used as a negative control for Bcl-6 expression to set the gate on the Bcl-6 histograms. Statistical significance was determined for the indicated comparisons using ordinary one-way analysis of variance (ANOVA) with Tukey’s multiple comparisons test (*p<0.05, **p < 0.01, ***p < 0.001).

Figure 2—figure supplement 2.. Histone deacetylase 3 (HDAC3)-deficient T cells from Peyer’s patches have reduced T_fh cell numbers.

Lymphocytes were harvested from the Peyer’s patches of wild-type (WT) and dLck-Cre HDAC3 cKO mice, and labeled for flow cytometry. Cells were first gated on HDAC3⁺ or HDAC3⁻ events, then gating for T_fh (CXCR5⁺ PD-1⁺ Bcl-6^hi). Non-T_fh CXCR5⁻ PD-1⁻ cells were used as a negative control for Bcl-6 expression to set the gate on the Bcl-6 histograms. Bar plots on right represent pooled data for the frequency of T_fh cells (n = 3–5 mice/group from two independent experiments). Statistical significance was determined for the indicated comparisons using ordinary one-way analysis of variance (ANOVA) with Tukey’s multiple comparisons test.

Figure 3.. Histone deacetylase 3 (HDAC3)-deficient CD4⁺ T cells are capable of differentiation, but produce fewer cells than wild-type (WT).

(a) In vitro differentiation assays were performed to examine differentiation into the T_h1, T_h2, T_h17, and T_reg lineages characterized by transcription factor expression. Splenocytes were harvested and magnetically enriched for naive (CD44⁻) CD4⁺ T cells by negative selection. Cells in all assays were stimulated with 2 µg/ml plate-bound αCD3 and 0.5 µg/ml αCD28 for 4 days. For T_h1 differentiation, 1 µg/ml αIL-4 antibody and 10 ng/ml of IL-12 were added to the media. For T_h2 differentiation, 1 µg/ml of each αIFNγ and αIL-12 antibody, as well as 10 ng/ml of IL-4 was added to the media. For T_h17 differentiation, media was supplemented with 10 µg/ml of αIFNγ and αIL-4 antibody as well as 10 ng/ml of rIL-23, 5 ng/ml TGF-β1, and 20 ng/ml IL-6. For T_reg differentiation, media was supplemented with 10 µg/ml αIFNγ and αIL-4 antibody as well as 2 ng/ml TGF-β1, and 2 ng/ml interleukin-2 (IL-2). Unstimulated control samples did not receive αCD3/αCD28 stimulation, but did receive 10 ng/ml IL-7 to maintain cell survival during culture. Bar plots on the right show % of cells from the total culture that are positive for the transcription factor ± standard deviation (SD), total cell number ± SD, and geometric mean of expression ± SD (n = 2–4 mice/group from two to three independent experiments for each). Two nonlittermate, but age and sex matched, WT B6 controls were used in these experiments. Negative controls (black histograms) represent unstimulated, but stained samples from the same mouse as the stimulated sample shown. Statistical significance was determined for the indicated comparisons using an unpaired t-test.

Figure 3—figure supplement 1.. Histone deacetylase 3 (HDAC3)-deficient CD4 T cells are capable of producing lineage-specific cytokines during in vitro differentiation.

(a) In vitro differentiation assays were performed to examine differentiation into the T_h1, T_h2, and T_h17 lineages characterized by cytokine expression. Splenocytes from wild-type (WT) and dLck-Cre HDAC3 cKO mice were harvested and magnetically enriched for naive (CD44⁻) CD4⁺ T cells by negative selection. Cells in all assays were stimulated with 2 µg/ml plate-bound αCD3 and 0.5 µg/ml αCD28 for 4 days. For T_h1 differentiation, 1 µg/ml αIL-4 antibody and 10 ng/ml of IL-12 were added to the media. For T_h2 differentiation, 1 µg/ml of each αIFNγ and αIL-12 antibody, as well as 10 ng/ml of IL-4 was added to the media. For T_h17 differentiation, media was supplemented with 10 µg/ml of αIFNγ and αIL-4 antibody as well as 10 ng/mL of rIL-23, 5 ng/ml TGF-β1, and 20 ng/ml IL-6. Unstimulated control samples did not receive αCD3/αCD28, stimulation, but did receive 10 ng/ml IL-7 to maintain cell survival during culture. Bar plots on the right show % of cells from the total culture that are positive for the cytokine ± standard deviation (SD). Negative controls (black histograms) represent unstimulated, but stained samples from the same mouse as the stimulated sample shown (n = 2–4 mice/group from two to three independent experiments for each). (b) In vitro differentiation assays were performed as in (a), and examined for coexpression of lineage-defining transcription factors and cytokines. Dot plots are representative of two to three independent experiments (n = 2–4 mice/group). Negative controls (black dots) represent unstimulated, but stained samples from the same mouse as the stimulated sample shown. Statistical significance was determined for all indicated comparisons using an unpaired t-test.

Figure 3—figure supplement 2.. Histone deacetylase 3 (HDAC3)-deficient CD4 T cells are capable of producing IFNγ ex vivo.

Splenocytes from wild-type (WT) and dLck-Cre HDAC3 cKO mice were harvested and magnetically enriched for total CD4+ T cells by negative selection. Cells were stimulated with 20 ng/ml PMA and 1 µg/ml of ionomycin for 6 hr. After stimulation, (a) total CD4 and (b) memory (CD44^hiCD62L^lo) cells were analyzed by flow cytometry for expression of T-bet and IFNγ. Bar graph on the right shows % of memory cells that are IFNγ+ T-bet^hi ± standard deviation (SD; n = 4–5 mice/group from two independent experiments).

Figure 4.. Histone deacetylase 3 (HDAC3)-deficient CD4⁺ T cells have reduced proliferation and diminished mechanistic target of rapamycin (mTOR) signaling after in vitro stimulation.

(a) Splenocytes were isolated from wild-type (WT) or dLck-Cre HDAC3 cKO mice, and magnetically enriched for CD4⁺ T cells by magnetic negative selection. Enriched CD4⁺ T cells were labeled with CFSE and stimulated with 5 µg of plate-bound αCD3 and 1 µg/ml of αCD28 for 3 days and examined by flow cytometry for proliferation. Cells from dLck-Cre HDAC3 cKO were gated on HDAC3⁻ cells to eliminate contaminating HDAC3⁺ events. Bar graphs on right show proliferation index ± standard deviation (SD) and % divided ± SD (n = 6 mice/group from three independent experiments). One nonlittermate, but age- and sex-matched WT B6 control was used in this experiment. (b) Enriched CD4⁺ T cells were stimulated with 5 µg plate-bound αCD3 and 1 µg/ml soluble αCD28 and analyzed for CD69 expression after 16 hr. Bar graph on right shows MFI ± SD (n = 6–7 mice/group from three independent experiments). (c) Enriched CD4⁺ T cells were analyzed for interleukin-2 (IL-2) production 48 hr after stimulation with 2 µg of plate-bound αCD3 and 0.5 µg/ml of αCD28. Bar graph on right shows fold induction ± SD (n = 4 mice/group from two independent experiments). (d) Enriched CD4⁺ T cells were stimulated with 5 µg plate-bound αCD3 and 1 µg/ml soluble αCD28 and analyzed for CD25 receptor expression after 3 days. Bar graph on right shows percent CD25 positive ± SD (n = 8 mice/group from four independent experiments). (e) Enriched CD4⁺ T cells were stimulated with 5 µg plate-bound αCD3 and 1 µg/ml soluble αCD28 and analyzed for viability after 24 hr. Bar graph on right shows percent viability dye positive ± SD (n = 7 mice/group from four independent experiments). (f) Enriched CD4⁺ T cells were stimulated with 5 µg plate-bound αCD3 and 1 µg/ml soluble αCD28 and analyzed for expression of P2RX7 after 20 hr. Bar graph on right shows MFI ± SD (n = 4 mice/group from two independent experiments). (g) Enriched CD4⁺ T cells were stimulated with 5 µg plate-bound αCD3 and 1 µg/ml soluble αCD28 and analyzed for Early Growth Response Protein 2 (Egr2) expression after 16 hr. Bar graphs on right shows percent Egr2⁺ ± SD and Egr2 MFI of Egr2⁺ population ± SD (n = 4–5 mice/group from three independent experiments). (h) Enriched CD4⁺ T cells were stimulated with 5 µg plate-bound αCD3 and 1 µg/ml soluble αCD28 and analyzed for phosphorylation of ribosomal protein S6 after 4 hr. Bar graph on right shows percent p-S6 positive ± SD (n = 5 mice/group from three independent experiments). (i) Enriched CD4⁺ T cells were stimulated overnight with 5 µg plate-bound αCD3 and 1 µg/ml soluble αCD28, then a Mito Stress Assay was conducted to measure oxygen consumption rate (OCR). Data are pooled, and show mean OCR ± standard error of the mean (SEM; n = 3 mice/group from two independent experiments). Statistical significance for indicated comparisons in all panels except for (f) was determined by an unpaired t-test. For (f), statistical significance was determined for the indicated comparisons using one-way analysis of variance (ANOVA) and with Tukey’s multiple comparisons test.

Figure 4—figure supplement 1.. dLck-Cre histone deacetylase 3 (HDAC3) cKO cells have reduced basal extracellular acidification rate after activation.

Enriched CD4⁺ T cells were stimulated overnight with 5 µg/ml plate-bound αCD3 and 1 µg/ml soluble αCD28, then a Mito Stress Assay was conducted to measure extracellular acidification rate (ECAR). Data are pooled, and show mean ECAR ± standard deviation (SD; n = 3 mice/group from two independent experiments).

Figure 5.. Histone deacetylase 3 (HDAC3)-deficient CD4⁺ T cells have defective blasting, reduced cholesterol levels, and increased cholesterol efflux transporter expression.

(a) Enriched CD4⁺ T cells were stimulated with 5 µg plate-bound αCD3 and 1 µg/ml soluble αCD28 and analyzed for blasting by size (forward scatter area, FSC-A) and granularity (side scatter area, SSC-A) after 3 days. Bar graphs on right represent mean ± standard deviation (SD; n = 6 mice/group from three independent experiments). One nonlittermate, but age- and sex-matched wild-type (WT) B6 control was used in this experiment. Statistical significance was determined for the indicated comparisons using an unpaired t-test. (b) Splenocytes were isolated from WT or dLck-Cre HDAC3 cKO mice, and labeled for cholesterol using Filipin III 20 hr after stimulation with 5 µg of plate-bound αCD3 and 1 µg/ml of αCD28. Flow cytometry was conducted. Bar plot (right) shows volume-adjusted MFI ± SD quantified across three independent experiments (n = 6 mice/group). MFI was adjusted to approximate cell volume by taking MFI divided by FSC-W³. Statistical significance was determined for the indicated comparisons using one-way analysis of variance (ANOVA) with Tukey’s multiple comparisons test. (c) Enriched CD4⁺ T cells were labeled with AF488-conjugated Cholera Toxin Subunit B 20 hr after stimulation with 5 µg of plate-bound αCD3 and 1 µg/ml of αCD28. Bar graph on right represents mean ± SD (n = 4 mice/group from two independent experiments). MFI was adjusted to approximate cell volume by taking MFI divided by FSC-W³. Statistical significance was determined for the indicated comparison using an unpaired t-test. (d, e) Enriched CD4⁺ T cells from Rag1-GFP WT and Rag1-GFP dLck-Cre HDAC3 cKO mice were sorted by fluorescence-activated cell sorting (FACS) for mature naive T cells (GFP⁻). After sorting, cells were stimulated with 5 µg of plate-bound αCD3 and 1 µg/ml of αCD28, and RT-qPCR was conducted to examine gene expression of genes involved in cholesterol homeostasis. Bar graphs represent mean ± SD (n = 3–4 mice/group from three independent experiments). Statistical significance was determined for the indicated comparisons using an unpaired t-test.

Figure 6.. Defective proliferation and blasting in histone deacetylase 3 (HDAC3)-deficient CD4⁺ T cells is rescued by addition of exogenous cholesterol in vitro.

(a) Enriched CD4⁺ T cells were labeled with CFSE and stimulated with 5 µg of plate-bound αCD3 and 1 µg/ml of αCD28 in the presence or absence of 5 µg cholesterol-methyl-β-cyclodextrin for 3 days and examined by flow cytometry for proliferation. Bar graphs represent mean proliferation index ± standard deviation (SD; n = 5 mice/group from three independent experiments). (b) Splenocytes were isolated from wild-type (WT) or dLck-Cre HDAC3 cKO mice, and magnetically enriched for CD4⁺ T cells by magnetic negative selection. Enriched CD4⁺ T cells were stimulated with 5 µg plate-bound αCD3 and 1 µg/ml soluble αCD28 in the presence or absence of 5 µg cholesterol-methyl-β- cyclodextrin, and analyzed for blasting by size (forward scatter area, FSC-A) and granularity (side scatter area, SSC-A) after 3 days. Bar graphs represent mean ± SD (n = 6–8 mice/group from three independent experiments). (c) Enriched CD4⁺ T cells were labeled with AF488-conjugated Cholera Toxin Subunit B 20 hr after stimulation with 5 µg of plate-bound αCD3 and 1 µg/ml of αCD28 in the presence or absence of 5 µg cholesterol-methyl-β-cyclodextrin. Bar graph on right represents mean ± SD (n = 4 mice/group from two independent experiments). MFI was adjusted to approximate cell volume by taking MFI divided by FSC-W³. (d) Enriched CD4⁺ T cells were labeled with Filipin III 20 hr after stimulation with 5 µg of plate-bound αCD3 and 1 µg/ml of αCD28 in the presence or absence of 5 µg cholesterol-methyl-β-cyclodextrin. MFI was adjusted to approximate cell volume by taking MFI divided by FSC-W³. Bar graph on right represents adjusted MFI ± SD (n = 6 mice/group from three independent experiments). (e) Enriched CD4⁺ T cells were stimulated with 5 µg plate-bound αCD3 and 1 µg/ml soluble αCD28 in the presence or absence of 5 µg cholesterol-methyl-β-cyclodextrin and analyzed for CD25 receptor expression after 3 days. Bar graphs represent mean percent CD25⁺ ± SD (n = 4 mice/group from two independent experiments). (f) Enriched CD4⁺ T cells were stimulated with 5 µg plate-bound αCD3 and 1 µg/ml soluble αCD28 in the presence or absence of 5 µg cholesterol-methyl-β-cyclodextrin and analyzed for Early Growth Response Protein 2 (Egr2) expression after 16 hr. Bar graphs represent percent Egr2⁺ ± SD and normalized Egr2 MFI ± SD among Egr2⁺ population (n = 8 mice/group from four independent experiments). (g) Enriched CD4⁺ T cells were stimulated with 5 µg plate-bound αCD3 and 1 µg/ml soluble αCD28 in the presence or absence of 5 µg cholesterol-methyl-β-cyclodextrin and analyzed phosphorylation of S6 after 4 hr. Bar graphs represent percent p-S6⁺ ± SD (n = 4 mice/group from two independent experiments). Statistical significance was determined for all indicated comparisons in this figure using a one-way analysis of variance (ANOVA) with Tukey’s multiple comparisons test.

Figure 7.. Histone deacetylase 3 (HDAC3) loss results in increased expression and hyperacetylation of ABCA1 and ABCG1.

(a) Gene expression (RNA-Seq) of Abca1, Abcg1, and Hmgcr in selecting (TCRβ^int CD69⁺) thymocytes from OT-II and OT-II CD2-iCre HDAC3 cKO mice. Bar graphs show mean RPKM (reads per kilobase million) ± standard deviation (SD). The exactTest (edgeR software) was used to compare mRNA levels (RPKM) from individual genes in RNA-Seq datasets. (b) Snapshot of H3K27Ac (left) and H3K9Ac (right) ChIP-seq tracks for the Abca1, Abcg1, and Hmgcr loci in selecting thymocytes from OT-II (wild-type, WT) and OT-II CD2-iCre HDAC3 cKO (CD2-iCre HDAC3-cKO) mice. (c) 16610D9 thymocytes were treated with competitive HDAC3 inhibitor RGFP966 for 24 or 48 hr, and expression of Abca1, Abcg1, and Hmgcr was measured by RT-qPCR at timepoints. Bar graphs show mean expression ± SD (n = 3 from three independent experiments). Statistical significance was determined for all indicated comparisons using an unpaired t-test. (d) WT splenocytes were harvested and magnetically enriched for CD4⁺ T cells. Cells were cultured with 5 µg of plate-bound αCD3 and 1 µg/ml of αCD28 or left unstimulated. Cells were also treated with competitive inhibitor of HDAC3 RGFP966. After 24 hr, expression of Abca1 and Abcg1 was examined by RT-qPCR. Bar graphs show mean expression ± SD (n = 4–5 mice/group from three independent experiments).

Figure 7—figure supplement 1.. Histone deacetylase 3 (HDAC3) binds to Abcg1 gene locus, upstream of the Abca1 locus in human CD4⁺ T cells.

Bedgraph from GSM393952 showing called peaks (gray bars) from HDAC3 ChIP-Seq in human CD4⁺ T cells. Value of peak height is listed below the bars.

Author response image 1.