USP2 promotes experimental colitis and bacterial infections by inhibiting the proliferation of myeloid cells and remodeling the extracellular matrix network

Abstract

Inflammatory bowel disease (IBD) is closely associated with dysregulation of genetic factors and microbial environment. Here, we report a susceptible role of ubiquitin-specific protease 2 (USP2) in experimental colitis and bacterial infections. USP2 is upregulated in the inflamed mucosa of IBD patients and in the colon of mice treated with dextran sulfate sodium salt (DSS). Knockout or pharmacologic inhibition of USP2 promotes the proliferation of myeloid cells to activate IL-22 and IFNγ production of T cells. In addition, knockout of USP2 in myeloid cells inhibits the production of pro-inflammatory cytokines to relieve the dysregulation of extracellular matrix (ECM) network and promote the gut epithelial integrity after DSS treatment. Consistently, Lyz2-Cre;Usp2^fl/fl mice exhibit hyper-resistance to DSS-induced colitis and Citrobacter rodentium infections compared to Usp2^fl/fl mice. These findings highlight an indispensable role of USP2 in myeloid cells to modulate T cell activation and epithelial ECM network and repair, indicating USP2 as a potential target for therapeutic intervention of IBD and bacterial infections in the gastrointestinal system.

Keywords: Bacterial infection; Cell proliferation; Colitis; Myeloid cells; USP2.

Graphical abstract

Fig. 1

USP2 plays a pro-inflammatory role in DSS-induced colitis. (A) USP2 mRNA expression level in colonic mucosa tissues from active ulcerative colitis (UC) patients and healthy people, the data were collected from GEO database (GDS3119) (n = 13). (B) A scheme of DSS treatment and body weight change of Usp2^+/+ (n = 3) and Usp2^−/− (n = 3) mice with DSS-induced colitis. (C) The representative image and lengths of colons and pathology scores of Usp2^+/+ (n = 8) and Usp2^−/− (n = 6) mice with DSS-induced colitis. (D) HE-stained images of colon sections of Usp2^+/+ and Usp2^−/− mice with DSS-induced colitis. (E) A scheme of DSS treatment and body weight change of Usp2^fl/fl (n = 24) and Lyz2-Cre;Usp2^fl/fl (n = 24) mice with DSS-induced colitis. (F) The representative image and lengths of colons and pathology scores of Usp2^fl/fl (n = 24; 15) and Lyz2-Cre;Usp2^fl/fl (n = 24; 19) mice with DSS-induced colitis. (G) HE-stained images of colon sections of Usp2^fl/fl and Lyz2-Cre;Usp2^fl/fl mice with DSS-induced colitis. ∗P ​< ​0.05, ∗∗P ​< ​0.01, ∗∗∗P < 0.001 (Student's unpaired t-test). Scale bars represent 600 μm and 200 μm. Graphs show mean ± S.D. (A-C, E-F). Data are representative of two (B–D) independent experiments or a combination of four (E–G) independent experiments.

Fig. 2

USP2 alters gene expression profile in myeloid cells after DSS treatment. (A) Schematic illustration of acquisition of colonic RFP⁺ cells from Lyz2-Cre;Ai14^LSL-tdTomato or Lyz2-Cre;Ai14^LSL-tdTomatoUsp2^fl/fl mice, flow cytometry analysis of purities of RFP⁺ cells in LPMCs from Lyz2-Cre;Ai14^LSL-tdTomato and Lyz2-Cre;Ai14^LSL-tdTomatoUsp2^fl/fl mice before and after flow sorting, and components of RFP⁺ cells in LPMCs from Lyz2-Cre;Ai14^LSL-tdTomato and Lyz2-Cre;Ai14^LSL-tdTomatoUsp2^fl/fl mice after DSS treatment, and proportions and numbers of RFP⁺ cells in LPMCs from Lyz2-Cre;Ai14^LSL-tdTomato (n = 4) and Lyz2-Cre;Ai14^LSL-tdTomatoUsp2^fl/fl (n = 4) mice before and after DSS treatment. (B) Heatmap of differentially expressed genes (DEGs) that meet the conditions of P-Value＜0.05 and |log₂Fold change|≥1 in the transcriptome analysis of colonic RFP⁺ cells from two groups of Lyz2-Cre;Ai14^LSL-tdTomato (n = 11) and Lyz2-Cre;Ai14^LSL-tdTomatoUsp2^fl/fl mice (n = 9) with DSS-induced colitis. (C) KEGG pathway enriched by DEGs in the transcriptome analysis of colonic RFP⁺ cells from two groups of Lyz2-Cre;Ai14^LSL-tdTomato (n = 11) and Lyz2-Cre;Ai14^LSL-tdTomatoUsp2^fl/fl (n = 9) mice with DSS-induced colitis. (D) GSEA analysis of cytokine and chemokine signaling pathway, NF-κB signaling pathway, and cell cycle signaling pathway from the transcriptome data of colonic RFP⁺ cells from two groups of Lyz2-Cre;Ai14^LSL-tdTomato (n = 11) and Lyz2-Cre;Ai14^LSL-tdTomatoUsp2^fl/fl (n = 9) mice with DSS-induced colitis. (E) Heatmap of the indicated genes related cytokine and chemokine signaling pathway from the transcriptome data of colonic RFP⁺ cells from two groups of Lyz2-Cre;Ai14^LSL-tdTomato (n = 11) and Lyz2-Cre;Ai14^LSL-tdTomatoUsp2^fl/fl (n = 9) mice with DSS-induced colitis. (F) qRT-PCR results of the signature genes for cytokines, chemokines in colonic RFP⁺ cells from Lyz2-Cre;Ai14^LSL-tdTomato (n = 12) and Lyz2-Cre;Ai14^LSL-tdTomatoUsp2^fl/fl (n = 12) mice with DSS-induced colitis. (G) Flow cytometry analysis and mean fluorescence intensities (MFIs) of p-p65 in RFP⁺ cells of LPMCs from Lyz2-Cre;Ai14^LSL-tdTomato (n = 4) and Lyz2-Cre;Ai14^LSL-tdTomatoUsp2^fl/fl (n = 4) mice with DSS-induced colitis. (H) Heatmap of the indicated genes related cell cycle from the transcriptome data of colonic RFP⁺ cells from two groups of Lyz2-Cre;Ai14^LSL-tdTomato (n = 11) and Lyz2-Cre;Ai14^LSL-tdTomatoUsp2^fl/fl (n = 9) mice with DSS-induced colitis. (I) qRT-PCR results of the signature genes for cell cycle in colonic RFP⁺ cells from Lyz2-Cre;Ai14^LSL-tdTomato (n = 12) and Lyz2-Cre;Ai14^LSL-tdTomatoUsp2^fl/fl (n = 12) mice with DSS-induced colitis. ∗P ​< ​0.05, ∗∗P ​< ​0.01, ∗∗∗P < 0.001 (Student's unpaired t-test for A, G or Student's paired t-test for F, I). ns, not significant. Graphs show mean ± S.D. (A, F, I, G). Data are representative of three (G) or a combination of three (F, I) or four (A) independent experiments.

Fig. 3

USP2 inhibits the proliferation of macrophages and dendritic cells to promote type I immune responses in the lamina propria during DSS-induced colitis. (A) Flow cytometry analysis of myeloid cells in LPMCs from Usp2^fl/fl (n = 4) and Lyz2-Cre;Usp2^fl/fl (n = 4) mice with DSS-induced colitis. (B) The proportions and numbers of CD11b⁺CD11c⁻ and CD11c⁺ cells in LPMCs from Usp2^fl/fl (n = 4) and Lyz2-Cre;Usp2^fl/fl (n = 4) mice with DSS-induced colitis. (C) The proportions and numbers of CD11b⁺F4/80⁺ macrophages and CD11b⁺Gr-1⁺ neutrophils in LPMCs from Usp2^fl/fl (n = 4) and Lyz2-Cre;Usp2^fl/fl (n = 4) mice with DSS-induced colitis. (D) The proportions and numbers of CD11b⁻CD103⁺, CD11b⁺CD103⁺, CD11b⁺CD103^- DCs in LPMCs from Usp2^fl/fl (n = 4) and Lyz2-Cre;Usp2^fl/fl (n = 4) mice with DSS-induced colitis. (E) Flow cytometry analysis of proportions of Ki67 in CD11b⁺F4/80⁺ macrophages, CD11b⁺Gr-1⁺ neutrophils, CD11c⁺CD11b⁻CD103⁺, CD11c⁺CD11b⁺CD103⁺ and CD11c⁺CD11b⁺CD103^- DCs of LPMCs from Usp2^fl/fl (n = 4) and Lyz2-Cre;Usp2^fl/fl (n = 4) mice with DSS-induced colitis. (F) Flow cytometry analysis and the proportions and numbers of IL-22-and IL-17A-producing T lymphocytes in LPMCs from Usp2^fl/fl (n = 4) and Lyz2-Cre;Usp2^fl/fl (n = 4) mice with DSS-induced colitis. ∗P ​< ​0.05, ∗∗P ​< ​0.01, ∗∗∗P < 0.001 (Student's unpaired t-test). Graphs show mean ± S.D. (B–F). Data are representative of two (A–F).

Fig. 4

USP2 in myeloid cells inhibits immune responses to bacterial infections in the gut. (A) Body weight change of Usp2^fl/fl (n = 16) and Lyz2-Cre;Usp2^fl/fl (n = 14) mice that were injected with 1 × 10⁹ colony-forming units (c.f.u.) of Citrobacter rodentium by gavage. (B) Bacterial counts of Usp2^fl/fl (n = 9) and Lyz2-Cre;Usp2^fl/fl (n = 11) mice at day 6 after injection of Citrobacter rodentium (1 × 10⁹ c.f.u.) by gavage. (C) Flow cytometry analysis and proportions of CD11b⁺F4/80⁺ macrophages, CD11c⁺CD11b⁻CD103⁺, CD11c⁺CD11b⁺CD103⁺ and CD11c⁺CD11b⁺CD103^- DCs in LPMCs from Usp2^fl/fl (n = 12) and Lyz2-Cre;Usp2^fl/fl (n = 10) mice that were treated as in (A). (D) Flow cytometry analysis and mean fluorescence intensities (MFIs) of Ki67 in CD11c⁺, CD11c⁺CD11b⁻CD103⁺ and CD11c⁺CD11b⁺CD103^- DCs of LPMCs from Usp2^fl/fl (n = 5) and Lyz2-Cre;Usp2^fl/fl (n = 5) mice that were treated as in (A). (E) Flow cytometry analysis and proportions of IFNγ-producing T lymphocytes in LPMCs from Usp2^fl/fl (n = 12) and Lyz2-Cre;Usp2^fl/fl (n = 10) mice that were treated as in (A). (F) qRT-PCR results of Ifng in colon tissues from Usp2^fl/fl (n = 14) and Lyz2-Cre;Usp2^fl/fl (n = 10) mice that were treated as in (A). ∗P ​< ​0.05, ∗∗P ​< ​0.01, ∗∗∗P < 0.001 (Student's unpaired t-test). ns, not significant. Graphs show mean ± S.D. (A–F). Data are a combination of two (A-C, E-F) independent experiments or representative of two (D) independent experiments.

Fig. 5

Knockout of USP2 in myeloid cells alters the extracellular matrix in the inflamed colon. (A) GSEA analysis of fibrogenic genes from the transcriptome data of colonic RFP⁺ cells from two groups of Lyz2-Cre;Ai14^LSL-tdTomato (n = 11) and Lyz2-Cre;Ai14^LSL-tdTomatoUsp2^fl/fl (n = 9) mice with DSS-induced colitis. (B) qRT-PCR results of the signature fibrogenic genes in colonic RFP⁺ cells from Lyz2-Cre;Ai14^LSL-tdTomato (n = 12) and Lyz2-Cre;Ai14^LSL-tdTomatoUsp2^fl/fl (n = 12) mice with DSS-induced colitis. (C) qRT-PCR results of ECM-related genes in colon tissues from Usp2^fl/fl (n = 8) and Lyz2-Cre;Usp2^fl/fl (n = 7) mice with DSS-induced colitis. (D) Masson staining and anti-α-SMA immunohistochemistry staining, the percentage of collagen-positive areas and the integral optical density (IOD) per mm² analysis of α-SMA in colon sections of Usp2^fl/fl (n = 7) and Lyz2-Cre;Usp2^fl/fl (n = 7) mice with DSS-induced colitis. (E) Schematic illustration of culturing WT colon organoids with conditioned medium. (F) qRT-PCR results of ECM-related genes in WT colon organoids cultured with conditioned medium. (G) Anti-ZO-1 and anti-Occludin immunohistochemistry staining, the integral optical density (IOD) per mm² analysis of ZO-1 and Occludin in colon sections of Usp2^fl/fl (n = 5; 6) and Lyz2-Cre;Usp2^fl/fl (n = 7; 11) mice with DSS-induced colitis. (H) Schematic illustration and the concentration of FD-40 in peripheral blood of posterior orbital venous plexus from Usp2^fl/fl (n = 10) and Lyz2-Cre;Usp2^fl/fl (n = 10) mice with DSS-induced colitis. ∗P ​< ​0.05, ∗∗P ​< ​0.01, ∗∗∗P < 0.001 (Student's paired t-test for B or Student's unpaired t-test for C-D, F–H). Scale bars represent 600 μm, 200 μm and 100 μm. Graphs show mean ± S.D. (B-D, F–H). Data are a combination of three (B) or two (C-D, G-H) independent experiments or representative of two (F) independent experiments.

Fig. 6

Pharmacological inhibition of USP2 inhibits DSS-induced colitis. (A) A scheme of DSS and ML364 treatment, and body weight change of WT mice that were treated successively with DMSO (n = 15) or ML364 (10 mg/kg, n = 13) for 8 days with DSS-induced colitis. (B) The representative image and lengths of colons of WT mice treated as in (A). (C) HE-stained images of colon sections of WT mice treated as in (A). (D) Flow cytometry analysis and proportions of CD11b⁺F4/80⁺ macrophages, CD11c⁺, CD11c⁺CD11b⁻CD103⁺, CD11c⁺CD11b⁺CD103⁺ and CD11c⁺CD11b⁺CD103^- DCs in LPMCs from WT mice that were treated successively with DMSO (n = 6) or ML364 (10 mg/kg, n = 7) for 8 days with DSS-induced colitis. (E) Flow cytometry analysis and proportions of Ki67 in CD11b⁺F4/80⁺ macrophages, CD11b⁻CD103⁺, CD11b⁺CD103⁺, CD11b⁺CD103^- DCs of LPMCs from WT mice that were treated successively with DMSO (n = 6) or ML364 (10 mg/kg, n = 7) for 8 days with DSS-induced colitis. (F) Flow cytometry analysis and proportions of IL-22- and IFNγ-producing T lymphocytes in LPMCs from WT mice that were treated successively with DMSO (n = 6) or ML364 (10 mg/kg, n = 5) for 8 days with DSS-induced colitis. ∗P ​< ​0.05, ∗∗P ​< ​0.01, ∗∗∗P < 0.001 (Student's unpaired t-test). Scale bars represent 600 μm, 200 μm. Graphs show mean ± S.D. (A-B, D-F). Data are a combination of three (A–C) or two (D–F) independent experiments.

Fig. 7

Pharmacological inhibition of USP2 inhibits ECM dysregulation in DSS-induced colitis. (A) Masson staining and anti-α-SMA immunohistochemistry staining, the percentage of collagen-positive areas and the IOD per mm² analysis of α-SMA in colon sections of WT mice that were treated successively with DMSO (n = 6; 12) or ML364 (10 mg/kg, n = 8; 8) for 8 days with DSS-induced colitis. (B) qRT-PCR results of ECM-related genes in colon tissues from WT mice treated successively with DMSO (n = 10) or ML364 (10 mg/kg, n = 7) for 8 days with DSS-induced colitis. (C) Anti-ZO-1 and anti-Occludin immunohistochemistry staining, the integral optical density (IOD) per mm² analysis of ZO-1 and Occludin in colon sections of WT mice that were treated successively with DMSO (n = 13; 6) or ML364 (10 mg/kg, n = 8; 5) for 8 days with DSS-induced colitis. (D) Schematic illustration and the concentration of FD-40 in peripheral blood of posterior orbital venous plexus from WT mice that were treated successively with DMSO (n = 5) or ML364 (10 mg/kg, n = 6) for 8 days with DSS-induced colitis. (E) A model for USP2 to modulate colonic infections and inflammations. USP2 inhibits the proliferation of macrophages and DCs to suppress T cells activation, which leads to the inhibition of repair and the susceptibility of bacteria, and promotes the production of cytokines and chemokines to dysregulate ECM, destroying the integrity of intestinal barrier, aggravating DSS-induced colitis. Pharmacologically inhibition of USP2 alleviates DSS-induced colitis. ∗P ​< ​0.05, ∗∗P ​< ​0.01, ∗∗∗P < 0.001 (Student's unpaired t-test). Scale bars represent 600 μm, 200 μm and 100 μm. Graphs show mean ± S.D. (A–D). Data are a combination of two (A–C) independent experiments or representative of two (D) independent experiments.