NCoR1 controls immune tolerance in conventional dendritic cells by fine-tuning glycolysis and fatty acid oxidation

Abstract

Dendritic cells (DCs) undergo rapid metabolic reprogramming to generate signal-specific immune responses. The fine control of cellular metabolism underlying DC immune tolerance remains elusive. We have recently reported that NCoR1 ablation generates immune-tolerant DCs through enhanced IL-10, IL-27 and SOCS3 expression. In this study, we did comprehensive metabolic profiling of these tolerogenic DCs and identified that they meet their energy requirements through enhanced glycolysis and oxidative phosphorylation (OXPHOS), supported by fatty acid oxidation-driven oxygen consumption. In addition, the reduced pyruvate and glutamine oxidation with a broken TCA cycle maintains the tolerogenic state of the cells. Mechanistically, the AKT-mTOR-HIF-1α-axis mediated glycolysis and CPT1a-driven β-oxidation were enhanced in these tolerogenic DCs. To confirm these observations, we used synthetic metabolic inhibitors and found that the combined inhibition of HIF-1α and CPT1a using KC7F2 and etomoxir, respectively, compromised the overall transcriptional signature of immunological tolerance including the regulatory cytokines IL-10 and IL-27. Functionally, treatment of tolerogenic DCs with dual KC7F2 and etomoxir treatment perturbed the polarization of co-cultured naïve CD4⁺ T helper (Th) cells towards Th1 than Tregs, ex vivo and in vivo. Physiologically, the Mycobacterium tuberculosis (Mtb) infection model depicted significantly reduced bacterial burden in BMcDC1 ex vivo and in CD103⁺ lung DCs in Mtb infected NCoR1^DC-/-mice. The spleen of these infected animals also showed increased Th1-mediated responses in the inhibitor-treated group. These findings suggested strong involvement of NCoR1 in immune tolerance. Our validation in primary human monocyte-derived DCs (moDCs) showed diminished NCOR1 expression in dexamethasone-derived tolerogenic moDCs along with suppression of CD4⁺T cell proliferation and Th1 polarization. Furthermore, the combined KC7F2 and etomoxir treatment rescued the decreased T cell proliferative capacity and the Th1 phenotype. Overall, for the first time, we demonstrated here that NCoR1 mediated control of glycolysis and fatty acid oxidation fine-tunes immune tolerance versus inflammation balance in murine and human DCs.

Keywords: FAO; Glycolysis; HIF-1α; NCoR1; OXPHOS; Th1; Tregs.

Graphical abstract

Fig. 1

NCoR1 depleted tolerogenic cDC1 DCs depicted enhanced glycolysis and OXPHOS as compared to control cells A. Representative line graph and scatter plots showing the glyco-stress parameters (glycolysis and glycolytic capacity) in unstimulated and 6h CpG activated control and NCoR1 depleted tolerogenic DCs, as measured using Seahorse extracellular flux analyzer. (n = 6) B. Scatter plot depicting the extracellular lactate accumulation in the culture supernatants of unstimulated and 6h CpG activated control and NCoR1 depleted tolerogenic DCs at a dilution of 1:10. (n = 5) C. Heatmap showing the differentially expressed genes (DEGs) enriched for the glycolytic pathway. RNAseq of unstimulated and 6h CpG activated NCoR1 KD as compared to control DCs were used to determine the DEGs. (n = 2) D. Bar graphs depicting the top enriched metabolic pathways from the DEGs in NCoR1 KD CpG activated tolerogenic DCs as compared to controls, using Reactome Pathway Database. (n = 2) E. RT-qPCR analysis showing the relative transcript expression of glut transporter Slc2a1, and glycolytic genes Hk2, Pkm2 and Ldha in unstimulated and 6h CpG activated NCoR1 KD tolerogenic DCs. (n = 5–7) F. Bar graphs demostrating the relative intensities of intracellular metabolites of NCoR1 KD CpG activated DCs as compared to controls as quantified using GC-MS. (n = 3) G. IGV snapshots showing the ChIPseq binding of NCoR1 at the Pfkp and Pfkfb3 gene loci in unstimulated and 6h CpG stimulated control DCs. H. Representative line graph and scatter plots showing the mito-stress parameters (basal respiration, maximal respiration, coupled ATP and spare respiratory capacity) in unstimulated and 6h CpG activated control and NCoR1 depleted tolerogenic DCs as measured using an extracellular flux analyzer. (n = 8) I. Box and whisker plots showing the min-max values for extracellular nitrite in the culture supernatant collected from unstimulated and 6h CpG activated control and NCoR1 depleted tolerogenic DCs. (n = 8) J. Box and whisker plots depicting the min-max values for mitochondrial DNA copies (Nd-1/Hk-II) quantified by RT-qPCR in unstimulated and 6h CpG activated control and NCoR1 KD DCs. (n = 5) K. Bar graphs with dots depicting the MFI levels of mitochondrial SOX, (n = 6) and cellular ROS, (n = 5) in unstimulated and 6h CpG stimulated control and NCoR1 KD DCs as quantified from flow cytometry. L. Bar graphs with dots showing the MFI levels of TMRM measured for mitochondrial potential estimation in unstimulated and 6h CpG activated control and NCoR1 KD DCs. (n = 6) *p ≤ 0.05, **p ≤ 0.01, ***p ≤0.001 and ****p ≤0.0001. p-value has been calculated using two tailed paired student’s t-test. Data shown in figure is combined from 3 independent experiments [A] and from 4 independent experiments [H]. Error bars represent SEM, except I and J, horizontal lines depicting mean.

Fig. 2

FLT3L differentiated primary BMDCs from NCoR1^DC−/− mice showed enhanced glycolysis and OXPHOS as observed in mutu-cDC1 line A. Representative line graph and scatter plots showing the glyco-stress parameters (glycolysis and glycolytic capacity) in unstimulated and 18h CpG activated control CD11c-Cre and NCoR1 ablated FLT3 differentiated BMDCs on day 9, measured using Seahorse extracellular flux analyzer. (n = 11) B. Scatter plot depicting the levels of extracellular lactate accumulated in the culture supernatants of control CD11c-Cre and NCoR1^DC−/− BMDCs before and after 18h CpG activation, at a dilution of 1:5. (n = 6) C. Bar graphs showing the relative transcript expression of glucose transporter Slc2a1, and glycolytic genes, Hk2, Pkm2 and Ldha in unstimulated and 18h CpG activated CD11c-Cre and NCoR1^DC−/− BMDCs. (n = 4–5) D. Representative line graph and scatter plots showing the mito-stress parameters (basal respiration, maximal respiration, coupled ATP and spare respiratory capacity) in unstimulated and 18h CpG activated control CD11c-Cre and NCoR1^DC−/− BMDCs on day 9 as measured using Seahorse extracellular flux analyzer. (n = 6) E. Bar graphs with dots depicting the MFI levels of cellular ROS in unstimulated and 18h CpG activated CD11c-Cre control and NCoR1^DC−/− BMcDC1s on day 9 as measured in F4/80⁻CD11c⁺CD24⁺ gated cells using flow cytometry. (n = 12) F. Bar graphs with dots showing the MFI levels of mitochondrial potential upon TMRM staining in CD11c-Cre control and NCoR1^DC−/− BMcDC1s on day 9 before and after 18h CpG stimulation as measured in F4/80⁻CD11c⁺CD24⁺ gated cells using flow cytometry. (n = 7) *p ≤ 0.05, **p ≤ 0.01, ***p ≤0.001 and ****p ≤0.0001. p-value has been calculated using two tailed unpaired student’s t-test. Data shown in figure is combined from 4 independent experiments [A] and from 3 independent experiments [D]. Error bars represent SEM.

Fig. 3

AKT-mTOR–HIF–1α axis in regulating glycolysis and tolerogenic cytokine production in NCoR1 depleted DCs A. Representative western blots showing the phosphorylated and total protein levels of AKT, mTOR and HIF-1α in control and NCoR1 depleted DCs after 6h CpG stimulation. Ratios of phosphorylated to total protein levels were further normalized to the housekeeping control β-actin. Corresponding bar plots depicted the densitometry analysis of the western blots. (n = 3–4) B. Bar graph with dots showing the relative mRNA expression of Hif-1α in 6h CpG activated control and NCoR1 depleted tolerogenic DCs. (n = 7) C. Heatmap showing the differentially expressed genes of HIF-1α pathway observed in the RNAseq data of control and NCoR1 KD 6h CpG stimulated DCs. (n = 3) D. IGV snapshot depicting the NCoR1 binding at the distal genomic regions of Hif-1α gene loci in unstimulated and 6h CpG activated control DCs. E. Line graph and box and whisker plots with dots depicting the levels of ECAR along with glycolysis and glycolytic capacity respectively upon pre-injection with either rapamycin (20nM) or KC7F2 (10μM) in control and NCoR1 KD DCs after 6h CpG activation, as measured using Seahorse extracellular flux analyzer. (n = 6) F. Scatter plots (percent positive cells) and bar graphs (MFI) showing the cytokine levels of IL-10 and IL-27 upon treatment with rapamycin (20nM) and KC7F2 (10μM) in CpG activated control and NCoR1 depleted tolerogenic DCs. (n = 7–8) *p ≤ 0.05, **p ≤ 0.01, ***p ≤0.001 and ****p ≤0.0001. p-value has been calculated using two tailed paired student’s t-test. Error bars represent SEM.

Fig. 4

Mitochondrial fuel/flux (Pyruvate oxidation, Glutamine oxidation and FAO) characteristics regulates tolerogenicity in NCoR1 depleted DCs A. Representative Western blot showing the protein level expression of pyruvate dehydrogenase (PDH) in control and NCoR1 depleted tolerogenic DCs after 6h CpG activation. The protein level was normalized to the housekeeping control β-tubulin. Corresponding bar graph depicted the densitometric analysis of the Western blot bands. (n = 4) B. Plots demonstrating the percent dependencies of 6h CpG activated control and NCoR1 depleted DCs on pyruvate [(first injection- UK-5099-2μM and second a combination of etomoxir (4μM)+BPTES (3μM) together] and glutamine oxidation [first injection - BPTES (3μM) and second a combination of UK-5099 (2μM) + etomoxir (4μM) together], measured from an extracellular flux analyzer. (n = 6) C. Heatmap showing the differentially expressed genes of TCA cycle observed in the RNAseq of unstimulated and 6h CpG activated control and NCoR1 depleted tolerogenic DCs. (n = 2) D. Scatter plots (percent positive cells) and bar graphs (MFI) demonstrating the cytokine levels of IL-10 and IL-27 upon treatment with UK-5099 (5μM), BPTES (5μM) and DON (5μM) in control and NCoR1 depleted tolerogenic DCs after 6h CpG stimulation. (n = 6) E. Line graph and scatter plot showing the FAO index in control and NCoR1 KD DCs after 6h CPG activation upon treatment with etomoxir (50μM) and Oligomycin (2μM), as measured using Seahorse Extracellular flux analyzer. (n = 6) F. Bar graph depicting the relative mRNA expression of Cpt1a in 6h CpG activated control and NCoR1 depleted tolerogenic DCs quantified using RT-qPCR. (n = 3) G. Heatmap showing the differentially expressed genes of FAO pathway in control and NCoR1 depleted DCs before and after 6h CpG activation using RNAseq data. (n = 2) H. Scatter plots (percent positive cells) and bar graphs (MFI) showing the cytokine levels of IL-10 and IL-27 upon treatment with etomoxir (50μM) in 6h CpG activated control and NCoR1 depleted DCs. (n = 6–7) I. Bar plots showing the normalized counts for Acly from RNAseq data and relative mRNA expression of Acaca and Fasn using RT-qPCR in 6h CpG stimulated control and NCoR1 KD DCs. (n = 3 for Acly and n = 5 for Acaca and Fasn) J. Representative microscopic images showing the Oil red-O stained neutral lipids in the control and NCoR1 KD DCs after 6h CpG activation. Corresponding dot plot depicted the quantification of the oil-red-O stain using absorbance at 500nm. (n = 6) K. Bar plot showing the normalized counts for Ppar-γ in the RNAseq data of 6h CpG activated control and NCoR1 depleted tolerogenic DCs. (n = 3) L. IGV snapshot showing the direct binding of NCoR1 at Ppar-γ gene loci in cDC1 before and after 6h CpG activation. M. Schematic illustration showing the non-incorporation of pyruvate into the TCA cycle, accumulation of citrate, reduced expression of TCA cycle genes and decreased glutamine oxidation in 6h CpG activated NCoR1 KD DCs. *p ≤ 0.05, **p ≤ 0.01, ***p ≤0.001 and ****p ≤0.0001. p-value has been calculated using two tailed paired student’s t-test. Data shown in figure is combined from 3 independent experiments [B]. Error bars represent SEM. . (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

Fig. 5

Global transcriptomic changes and T cell polarization upon combined inhibition of HIF-1α and CPT1a A. Scatter plots (percent positive cells) and bar graphs (MFI) showing the cytokine levels of IL-10 and IL-27 upon combined treatment with etomoxir (50μM) + KC7F2 (10μM) in 6h CpG activated control and NCoR1 depleted tolerogenic DCs. (n = 5) B. Heatmap showing the differentially expressed genes of glycolytic and FAO pathway and their regulators in the RNAseq data before and after treatment with combined inhibitors KC7F2 and etomoxir in 6h CpG stimulated control and NCoR1 KD DCs. (n = 3) C. Heatmap with bar plots showing the enriched GO terms associated with genes differentially regulated in 6h CpG activated NCoR1 KD before and after combined inhibition with HIF-1α and CPT1a. The asterisk in the heatmap denotes the significantly up or down regulated genes (p-adjusted <0.05) from DESeq2 results in CpG activated NCoR1 KD + Inhibitor treated as compared to only NCoR1 KD conditions. (n = 3) D. Box plots depicting the distribution of gene expression (z-score) belonging to tolerogenic and inflammatory pathways with and without combined inhibition of HIF-1α and CPT1a in 6 h CpG activated control and NCoR1 KD DCs. (n = 3) E. Bar graphs showing the normalized counts (DESeq2) of Il10, Il27, Ido1, Ctla4, Cd274 and Socs3 in 6h CpG stimulated control and NCoR1 KD DCs before and after combined inhibition of HIF-1α and CPT1a. (n = 3) F. Schematic outline showing the control and NCoR1 KD DCsT cell co-culture experiment with the naive CD4⁺T cells isolated from OT-II transgenic mice ex vivo in the presence and absence of combined inhibitors KC7F2 and etomoxir. G. Bar graph and histogram plot showing MFI levels and shifts respectively of eF670 labelled CD4⁺ T cells upon co-culturing with 6h CpG activated control and NCoR1 KD DCs before and after treatment with combined inhibitors of HIF-1α and CPT1a. Changes in the cell proliferation levels were measured for CD3⁺CD4⁺CD44⁺ gated effector T cells by flow cytometry after 72 h of co-culture. (n = 6) H. Scatter plots (percent positive cells) and bar graphs (MFI) showing the IFN-γ⁺ and T-bet⁺ cells gated on CD3⁺CD4⁺CD44⁺ effector T cells upon co-culturing with control and NCoR1 KD DCs, with and without combined inhibition of HIF-1α and CPT1a. Percent positive cells and MFI levels were analyzed by flow cytometry after 96 h of co-culture. (n = 3) I. Scatter plot showing the percent CD25⁺ and FOXP3⁺ double positive cells (Tregs) gated on CD3⁺CD4⁺CD44⁺ effector T cells upon co-culturing with control and NCoR1 KD DCs, with and without combined inhibition of HIF-1α and CPT1a. (n = 3) J. Schematic showing the outline of an in vivo experiment for polarization of naive T cells upon adoptive transfer of NCoR1 KD DCs intraperitoneally in OT-II transgenic mice with and without combined inhibition of HIF-1α and CPT1a. Nearest draining lymph nodes (DLNs) were harvested after 96 h to profile the Th1 and Tregs. K. Scatter plots (percent positive cells) and bar graphs (MFI) showing the IFN-γ⁺ and T-bet⁺ cells, gated on CD3⁺CD4⁺CD44⁺ effector T cells in DLNs after 96 h, representing the Th1 subtype in OT-II mice upon adoptive transfer of NCoR1 KD DCs with and without combined inhibition of HIF-1α and CPT1a. (n = 7) L. Scatter plot showing the percent CD25⁺ and FOXP3⁺ double positive cells (Tregs), gated on CD3⁺CD4⁺CD44⁺ effector T cells in DLNs after 96 h, representing the Treg subtype in the OT-II mice upon adoptive transfer of NCoR1 KD DCs, with and without combined inhibition of HIF-1α and CPT1a. (n = 7). *p ≤ 0.05, **p ≤ 0.01, ***p ≤0.001 and ****p ≤0.0001. p-value has been calculated using two tailed paired [A, G, H and I] and unpaired [K and L] student’s t-test. Error bars represent SEM.

Fig. 6

Mtb disease burden upon dual inhibition of HIF-1α and CPT1a in NCoR1^DC−/− mice ex vivo and in vivo A. Schematic showing the mCherry-Mtb (H37Rv) infection in primary NCoR1^DC−/− BMcDC1s gated on F4/80⁻CD11c⁺CD24⁺ cells with and without combined treatment with KC7F2 and etomoxir. The cells were harvested after 2h and 18h post infection (hpi) and analyzed by flow cytometry. B. Representative FACS contour plot depicting gating of Mtb infected cells, along with scatter plot and bar graph showing the cumulative percent infection and change in MFI levels with and without combined treatment with KC7F2 and etomoxir at 2h and 18h post infection in NCoR1^DC−/− BMcDC1s. (n = 6) C. Schematic outline showing the GFP-Mtb (H37Rv) infection in vivo in NCoR1^DC−/− mice with (inhibitors were injected intraperitoneally at day 3, day 8, day 12 and day 16 post infection) and without combined inhibition with KC7F2 and etomoxir. Mice were sacrificed on day 21 to check bacterial burden in CD103⁺ lung DCs (cDC1 counterpart in lungs) and Th cell phenotype in spleen. D. Scatter plot and bar graph showing the cumulative percent infection and change in Mtb burden (MFI) in lung DCs (gated on CD11c⁺CD103⁺ cells) in NCoR1 DC^−/− mice at day 21 with and without combined treatment with KC7F2 and etomoxir. (n = 8) E. Representative FACS dot plots depicting the gating of IFN-γ⁺ and T-bet⁺ cells gated on CD3⁺CD4⁺CD44⁺ effector T cells in the Spleen of treated and untreated mice. Scatter plots and bar graphs showed the cumulative percent infection and change in MFI levels at day 21 with and without combined treatment with KC7F2 and etomoxir in NCoR1^DC−/− mice. (n = 5) *p ≤ 0.05, **p ≤ 0.01, ***p ≤0.001 and ****p ≤0.0001. p-value has been calculated using two tailed unpaired student’s t-test. Error bars represent SEM.

Fig. 7

NCOR1 expression and T cell polarization in human PBMCs derived moDCs A. Scatter plot showing the correlation between NCOR1 and tolerogenic genes IL-10 and SOCS3 from a publicly available RNAseq dataset. (n = 5) B. Violin plot showing the normalized counts of NCOR1 in human conventional DCs in immature, mature, and in different tolerogenic setup (Dex, VD3 and rapamycin treated groups), from publicly available RNAseq datasets. (n = 4) C. Schematic showing the experimental outline used for differentiation of mature and tolerogenic (Dex-treated) moDCs to check the mRNA (RT-qPCR) and protein level expression of NCOR1. D. Bar plots depicting the relative mRNA expression of NCOR1, IL-10, IL-12, HK-2 and LDHA in mature and tolerogenic moDCs after 24h pIC stimulation. (n = 4–5) E. Representative microscopic images showing the NCOR1 protein levels (red) in mature and tolerogenic moDCs after 24h pIC stimulation. Corresponding bar graph depicted the quantification of the NCOR1 intensities in the microscopic images. (n = 5) F. Schematic showing the CD4 T cell proliferation and T cell subtype profiling after 24h pIC stimulation in the co-culture of CD4⁺ T cells with mature and tolerogenic moDCs with and without KC7F2 and etomoxir treatment. G. Representative FACS contour plots showing the gating of P1 and P2 populations of eF670 stained CD3⁺CD4⁺ T cells upon co-culture with mature and tolerogenic moDCs with and without combined treatment with KC7F2 and etomoxir after 72h after 24h pIC stimulation. Corresponding bar graph depicted the cumulative MFI shifts calculated from independent biological replicates and histogram showed the representative MFI shifts observed. (n = 5) H. Representative FACS contour plot showing the gating of IFN-γ⁺ and T-bet⁺ populations gated on CD3⁺CD4⁺T cells upon co-culturing with moDCs for 96 h. Scatter plots and bar graphs depicted the cumulative percent positive and MFI shifts of IFN-γ⁺ and T-bet⁺ cells in mature and tolerogenic moDCs conditions with and without combined treatment with KC7F2 and etomoxir at 24h pIC stimulation.(n = 5) *p ≤ 0.05, **p ≤ 0.01, ***p ≤0.001 and ****p ≤0.0001. p-value has been calculated using two tailed unpaired student’s t-test. Error bars represent SEM. . (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)