The immunoproteasome regulates ILC2 responses by modulating mitochondrial capacity

Abstract

Type 2 innate lymphoid cells (ILC2s) contribute to type 2 immunity but have also been associated with multiple inflammatory diseases, including airway inflammation and asthma. We report that beyond its function of degrading poly-ubiquitinylated proteins, the immunoproteasome (i-20S) is required for the proper function of ILC2s by controlling their mitochondrial capacity. We found that 90% of the catalytic β subunits of proteasomes in human ILC2s (hILC2s) are the immuno- (β5i) rather than constitutive (β5c) isoform. Specific, noncovalent, reversible inhibition of i-20S β5i (LMP7) in hILC2s induced ROS production, which inhibited aconitase, leading to altered mitochondrial function and reduced levels of ATP. Reprogramming of metabolic status by an LMP7 inhibitor impaired ILC2 activation, without significant cytotoxicity or preventing their recovery. Hence, the selective inhibition of i-20S in ILC2 cells did not kill them but reversibly depleted their ATP, preventing their activation and cytokine secretion. In mice, proteasome inhibition similarly blocked mitochondrial function and ILC2 activation, preventing airway inflammation in response to IL33 and asthma in response to house dust mites. These findings reveal a previously unappreciated linkage between proteasome blockade, central carbon metabolism, and mitochondrial function and identify a strategy to regulate immune cell metabolism in inflammatory diseases.

Keywords: airway inflammation; innate lymphoid cells; metabolism; proteasome.

Fig. 1.

The i-20S regulates human ILC2 activation. (A) hILC2s from healthy donors (HDs, n = 3) were amplified for 10 d. PSMB8 (β5i) and PSMB5 (β5c) were quantified by Western blot. (B) hILC2s from HDs (n = 4) were incubated for 6 h alone or with IL-33 (10 ng/mL). Gene expression of PSMB9 (β1i), PSMB10 (β2i), and PSMB8 ((β5i) was quantified by PCR. (C) hILC2s from HDs (n = 3) were cultured with several concentrations of either β5i inhibitor (PKS3053), β5c (WZ1831), or negative control (IPX-10132) and each IC50 was determined. (D and E) hILC2s from HDs (n = 6 to 20) were cultured for 6 h alone (unst) or with IL-33 (10 ng/mL) either alone or with β5i inhibitor PKS3053 at 6, 1.25, 0.25, or 0.05 μM. Gene expression of IL13 and IL5 was quantified by PCR and normalized to IL-33. (F) hILC2s from HDs (n = 8) were cultured for 24 h alone (unst) or with IL-33 (10 ng/mL) either alone or with β5i inhibitor PKS3053 at 1.25 μM. IL13 production was quantified by ELISA. (G) hILC2 from HDs (n = 5) were stained with Cell Trace Violet and cultured for 3 d alone (unst), with IL-33 (10 ng/mL) either alone or with β5i inhibitor PKS3053 at 1.25 μM. (H) Principal component analysis of unstimulated hILC2s (unst), IL-33 activated hILC2s (IL-33), and IL-33 activated hILC2s treated with PKS3053 (IL-33+PKS3053) (n = 3). (I) Venn diagram of differentially expressed genes (DEGs) in IL-33 vs. Unst, IL-33+PKS3053 vs. Unst and IL-33+PKS3053 vs. IL-33. (J) K-means clustering (K = 6) of DEGs induced by the conditions shown in (I). (K) Dot plot of cytokines expressed by unstimulated hILC2s, IL-33-activated hILC2s, and IL-33 activated hILC2s treated with β5i inhibitor PKS3053. All results are represented as means ± SEM. Statistical significance was evaluated using a Kruskal–Wallis test with Dunn’s multiple comparisons posttest or Mann–Whitney test. *P < 0.05, **P < 0.01, ***P < 0.001.

Fig. 2.

The i-20S is a major regulator of human ILC2 metabolism. (A–C) Oxygen consumption rate (OCR) was measured in hILC2s from HDs (n = 3) cultured 24 h alone (black), with IL-33 (10 ng/mL, green) or with β5i inhibitor PKS3053 (1.25 μM, pink). Basal respiration (B) and spare respiratory capacity (C) in hILC2s from HDs (n = 3) cultured in same conditions as in (A) were quantified. (D) hILC2s from HDs (n = 14) were cultured for 3 h alone (unst), with IL-33 (10 ng/mL) either alone or with β5i inhibitor PKS3053 (1.25 μM). ATP was quantified by luminescence. (E and F) hILC2s from HDs (n = 14) were cultured for 6 h alone (unst), with IL-33 (10 ng/mL) either alone or with β5i inhibitor PKS3053 (1.25 μM). TMRM and mitotracker were quantified by flow cytometry and normalized to IL-33. (G) hILC2s from HDs (n = 4) were cultured alone or with β5i inhibitor PKS3053 (1.25 μM) for 24 h. Viability was measured by flow cytometry. (H) hILC2s from HDs (n = 3) were cultured for 24 h alone (Med) or with β5i inhibitor PKS3053 (1.25 μM). After a wash, hILC2s were activated or not with IL-33 (10 ng/mL). ATP and IL13 expression were quantified respectively by luminescence and PCR and both normalized to IL-33. All results are represented as means ± SEM. Statistical significance was evaluated using a Mann–Whitney test. *P < 0.05, ***P < 0.001. ns = not significant.

Fig. 3.

The i-20S controls aconitase activity but not its expression. (A) Schema showing simplified TCA cycle including the following metabolites: pyruvate (blue), α-ketoglutarate (brown), and glutamine (pink). (B and C) hILC2s from HDs (n = 6) were cultured for 6 h alone, with IL-33 alone (10 ng/mL) or with β5i inhibitor PKS3053 (1.25 μM) supplemented or not by pyruvate (10 mM). ATP production and gene expression of IL13 were quantified respectively by luminescence and PCR and both normalized to IL-33. (D and E) hILC2s from HDs (n = 6) were cultured for 3 or 6 h alone, with IL-33 (10 ng/mL) either alone or with β5i inhibitor PKS3053 (1.25 μM) supplemented or not by glutamine (10 mM). ATP production and gene expression of IL13 were quantified respectively by luminescence and PCR and both normalized to IL-33. (F and G) hILC2s from HDs (n = 6) were cultured for 3 or 6 h alone, with IL-33 (10 ng/mL) either alone or with β5i inhibitor PKS3053 (1.25 μM) supplemented or not by α-ketoglutarate (10 mM). ATP production and gene expression of IL13 were quantified respectively by luminescence and PCR and both normalized to IL-33. (H and I) hILC2s from HDs (n = 4) were cultured for 3 h alone or with β5i inhibitor PKS3053 (1.25 μM) following by the addition or not of IL-33 (10 ng/mL). Citrate was quantified by a colorimetric assay. (J and K) hILC2s from HDs (n = 3) were cultured for 24 h alone or with β5i inhibitor PKS3053 (1.25 μM). ACO2 protein level was quantified by western blot. (L and M) hILC2s from HDs (n = 6) were cultured for 3 h alone or with β5i inhibitor PKS3053 (1.25 μM) following by the addition or not of IL-33 (10 ng/ml). Aconitase activity was quantified by a fluorometric assay. All results are represented as means ± SEM. Statistical significance was evaluated using a Mann–Whitney test. *P < 0.05, **P < 0.01. ns = not significant.

Fig. 4.

The i-20S inhibition induces ROS production in immune cells. (A) Heatmap of genes enriched in the ROS pathway in unstimulated hILC2s (Unst), IL-33 activated hILC2s, and IL-33 activated hILC2s treated with β5i inhibitor PKS3053 (IL33+PKS3053). (B–D) hILC2s from HDs (n = 12) were cultured for 3 h alone or IL-33 (10 ng/mL) either alone or not with β5i inhibitor PKS3053 (1.25 μM). ROS production was measured by flow cytometry. (E and F) PBMCs from HDs (n = 3) were cultured for 3 h alone or with β5i inhibitor PKS3053 (1.25 μM). (G) Representative histograms of human dermal fibroblasts from HDs (n = 3) cultured for 3 or 5 h alone, with β5i inhibitor PKS3053 (1.25 μM), WZ1831 (1.25 μM), or with negative control IPX10132 (1.25 μM). ROS production was measured by flow cytometry. All results are represented as means ± SEM. Statistical significance was evaluated using a Mann–Whitney test. ***P < 0.001. ns = not significant.

Fig. 5.

Blockade of ROS induction prevents ILC2 inhibition by targeting i-20S. (A–H) hILC2s from HDs (n = 6) were cultured for 3 (A–C) or 6 h (D–H) alone or IL-33 (10 ng/mL) either alone or not with β5i inhibitor PKS3053 (1.25 μM) incubated or not with ROS inhibitor NAC (N-Acetyl-D-cysteine, 10 mM). Aconitase activity (A) was quantified by a fluorometric assay; citrate (B) and ATP (C) were measured by luminescence assays and normalized to IL-33 + PKS3053 (B) and IL-33 (C); mitotracker (D and E) and TMRM (D and F) were quantified by flow cytometry and normalized to IL-33; and expression of IL13 (G) and IL5 (H) was quantified by PCR and normalized to IL-33. All results are represented as means ± SEM. Statistical significance was evaluated using a Mann–Whitney test. *P < 0.05, **P < 0.01.

Fig. 6.

Treatment with β5 inhibitor reduces ILC2 activation during airway inflammation by affecting their metabolism. (A) Model of IL-33 induced airway inflammation with four injections of PBS or IL-33 (500 ng) either alone or with IPI-β5-1 inhibitor (30 μg/kg). (B) Percentage of lung mILC2s (CD45⁺Lin⁻CD127⁺CD90.2⁺ST2⁺) in CD45⁺ from the lung of control, IL-33, and IL-33+IPI-β5-1 mice measured by flow cytometry (n = 12). (C) Percentage of eosinophils (Eos, CD45⁺CD11c⁻SiglecF⁺), macrophages (Mac, CD45⁺Ly6G⁻SiglecF⁺CD11c⁺F4/80⁺), neutrophils (PMN, CD45⁺Ly6G⁺SiglecF⁻CD11b⁺), and T cells (CD45⁺CD3⁺) in BAL of IL-33 injected mice and IL-33 injected mice treated with IPI-β5-1 (n = 6 or 8). (D and E) IL13 and IL5 expression was quantified by PCR in total lung from control mice, IL-33 injected mice, and IL-33 injected mice treated with IPI-β5-1 (n = 9). (F) IL-13, IL-4, IL-5, IL-6, IL-10, and TNF were measured by flow cytometry in BAL from WT mice, WT mice injected with IL-33 and WT mice injected with IL-33 and treated with IPI-β5-1 (n = 6). (G–I) Intracellular staining of IL-13 (G), IL-5 (H), and IL-4 (I) was quantified in lung mILC2s by flow cytometry (n = 6). (J–L) ROS production was measured in lung mILC2s by flow cytometry (n = 6). (M–O) TMRM and mitotracker were measured in lung mILC2s by flow cytometry (n = 6). All results are represented as means ± SEM. Data are cumulative of two to four independent experiments and statistical significance was evaluated using a Mann–Whitney test. *P < 0.05, **P < 0.01, ***P < 0.001. ns = not significant.

Fig. 7.

Treatment with β5 inhibitor prevents asthma in HDM-treated mice. (A) Unsupervised analysis of the lung from control, HDM (10 μg/mL), or HDM (10 μg/mL)+IPI-β5-1 (300 μg/kg) using t-SNE and mapping the following populations: eosinophils (CD45⁺F4/80⁺SiglecF⁺CD11c⁻), ILC2s (CD45⁺Lin⁻CD127⁺CD90.2⁺ST2⁺), tissue-resident alveolar macrophages (TR-AM, CD45⁺F4/80⁺SiglecF⁺ CD11c⁺), monocyte-derived alveolar macrophages (Mo-AM, CD45⁺F4/80⁺SiglecF⁻CD11c⁻), CD4 (CD45⁺CD3⁺CD4⁺), CD8 (CD45⁺CD3⁺CD8⁺), cDC1 (CD45⁺CD11b^-CD11c⁺), and cDC2 CD45⁺CD11b⁺CD11c⁺). (B and C) Percentage of lung mILC2s and eosinophils in CD45⁺ from the lung of control, HDM, and HDM+IPI-β5-1 mice measured by flow cytometry (n = 14). (D and E) IL13 and IL5 expression was quantified by PCR in total lung from control mice, HDM-injected mice, and HDM-injected mice treated with IPI-β5-1 (n = 9). (F) IL-13, IL-4, IL-5, IL-6, IL-10, and TNF were measured by flow cytometry in BAL from WT mice, WT mice injected with HDM and WT mice injected with HDM and treated with IPI-β5-1 (n = 6). (G–I) Intracellular staining of IL-13 (G), IL-4 (H), and IL-5 (I) was quantified in lung mILC2s of WT mice, WT mice injected with HDM and WT mice injected with HDM and treated with IPI-β5-1 by flow cytometry (n = 6). (J) Representative images of H&E staining from lung sections at a ×10 magnification. (K and L) Epithelium thickness (K) and inflammation score (L) in WT mice, WT mice injected with HDM and WT mice injected with HDM and treated with IPI-β5-1 (n = 8). All results are represented as means ± SEM. Data are cumulative of three independent experiments and statistical significance was evaluated using a Mann–Whitney test. *P < 0.05, **P < 0.01, ***P < 0.001.