Loss of Deubiquitinase USP1 Blocks Pancreatic β-Cell Apoptosis by Inhibiting DNA Damage Response

Abstract

Impaired pancreatic β-cell survival contributes to the reduced β-cell mass in diabetes, but underlying regulatory mechanisms and key players in this process remain incompletely understood. Here, we identified the deubiquitinase ubiquitin-specific protease 1 (USP1) as an important player in the regulation of β-cell apoptosis under diabetic conditions. Genetic silencing and pharmacological suppression of USP1 blocked β-cell death in several experimental models of diabetes in vitro and ex vivo without compromising insulin content and secretion and without impairing β-cell maturation/identity genes in human islets. Our further analyses showed that USP1 inhibition attenuated DNA damage response (DDR) signals, which were highly elevated in diabetic β-cells, suggesting a USP1-dependent regulation of DDR in stressed β-cells. Our findings highlight a novel function of USP1 in the control of β-cell survival, and its inhibition may have a potential therapeutic relevance for the suppression of β-cell death in diabetes.

Keywords: Biochemical Mechanism; Cell Biology; Endocrinology.

Graphical abstract

Figure 1

USP1 Knockdown Protects β-Cell from Apoptosis Under Diabetic Conditions (A–F) INS-1E cells were seeded at 300,000 cells/well and transfected with either control scrambled siRNA (siScr) or siRNA specific to USP1 (siUSP1) and treated with (A and B) 22.2 mM glucose (HG), (C and D) a mixture of 22.2 mM glucose and 0.5 mM palmitate (HG/Pal), or (E and F) pro-inflammatory cytokines (2 ng/mL recombinant human IL-1β, 1000 U/mL TNF-α, and 1000 U/mL IFN-γ; Cyto) for 2 days. Representative Western blots (A, C, and E) and quantitative densitometry analysis (B, D, and F) of cleaved caspase 3 (Cl Casp3) and cleaved PARP (Cl PARP) protein levels are shown. Data are pooled from at least three independent cell line experiments. Data show means ±SEM. *p < 0.05 siScr treated compared with siScr control conditions. **p < 0.05 siUSP1-treated compared with siScr-treated conditions. ^#p = 0.05 compared with HG (B) or Cyto (F). See also Figure S1 for USP1 quantification.

Figure 2

USP1 Inhibitor ML323 Blocks β-Cell Apoptosis Under Diabetic Conditions (A) Chemical structure of ML323. (B–G) About 500,000 INS-1E cells/well treated with or without USP1 inhibitor ML323 were exposed to 22.2 mM glucose (HG; B and C), a mixture of 22.2 mM glucose and 0.5 mM palmitate (HG/Pal; D and E), or pro-inflammatory cytokines (2 ng/mL recombinant human IL-1β, 1000 U/mL TNF-α, and 1000 U/mL IFN-γ; Cyto; F and G) for 2 days. Representative Western blots and quantitative densitometry analysis of cleaved caspase 3 (Cl Casp3) and cleaved PARP (Cl PARP) protein levels are shown. Data are pooled from at least three independent cell line experiments. Data show means ±SEM. *p < 0.05 treated compared with control conditions. **p < 0.05 inhibitor-treated compared with treated conditions. ^#p = 0.05 compared with HG/P (E) or untreated control (G) alone. See also Figure S2 for PCNA ubiquitination.

Figure 3

USP1 Inhibitor SJB2-043 Blocks β-Cell Apoptosis Under Diabetic Conditions (A) Chemical structure of SJB2-043. (B–G) About 500,000 INS-1E cells/well treated with or without USP1 inhibitor SJB2-043 were exposed to 22.2 mM glucose (HG; B and C), a mixture of 22.2 mM glucose and 0.5 mM palmitate (HG/Pal; D and E), or pro-inflammatory cytokines (2 ng/mL recombinant human IL-1β, 1000 U/mL TNF-α, and 1000 U/mL IFN-γ; Cyto; F and G) for 2 days. Representative Western blots and quantitative densitometry analysis of cleaved caspase 3 (Cl Casp3) and cleaved PARP (Cl PARP) protein levels are shown. Data are pooled from at least three independent cell line experiments except for cl PARP in C (n = 2). Data show means ±SEM. *p < 0.05 treated compared with control conditions. **p < 0.05 inhibitor-treated compared with treated conditions. ^#p = 0.05 compared with Cyto (G) alone.

Figure 4

Inhibition of USP1 Promotes β-Cell Survival in Isolated Human Islets (A–J) Isolated human islets (150–200 islets/dish) treated with or without USP1 inhibitors ML323 or SJB2-043 were exposed to a mixture of 22.2 mM glucose and 0.5 mM palmitate (HG/Pal) (A, B, E, F, I, and J) or to pro-inflammatory cytokines (2 ng/mL recombinant human IL-1β and 1000 U/mL IFN-γ; Cyto; C, D, G–J) for 2 days. Representative Western blots and quantitative densitometry analysis of cleaved caspase 3 (Cl Casp3) protein levels are shown. (G) All lanes are from the same gel but were run noncontiguously. (I and J) Human pancreatic islets were fixed, paraffin embedded, and stained for the TUNEL assay and insulin. Representative images (I) and quantitative percentage of TUNEL-positive β-cells (J) are shown. Scale bars represent 50 μm. Data are pooled from at least three human islet preparations. Data show means ±SEM. *p < 0.05 stimuli treated compared with control conditions. **p < 0.05 inhibitor-treated compared with treated conditions.

Figure 5

Impact of USP1 Inhibition on Insulin Content, Secretion, and Expression of Functional β-Cell Genes in Isolated Human Islets (A–D) Isolated human islets were treated with or without USP1 inhibitors ML323 or SJB2-043 for 2 days (30 islets/dish for A–C; 150–200 islets/dish for D). (A) Insulin content analyzed after GSIS and normalized to whole islet protein. (B) Insulin secretion during 1-hr incubation with 2.8 (basal) and 16.7 mM glucose (stimulated), normalized to insulin content. (C) The insulin stimulatory index denotes the ratio of secreted insulin during 1-hr incubation with 16.7 and 2.8 mM glucose. (D) RT-PCR for NEUROD1, INS, PDX1, NKX2.2, NKX6.1, and SlC2A2, normalized to cyclophilin. Tubulin normalization delivered similar results. Pooled data are from four independent experiments from four different human islet donors. Data show means ±SEM. *p < 0.05 compared with control conditions.

Figure 6

Inhibition of USP1-Suppressed DDR (A, B, and K) INS-1E cells seeded at 300,000 cells/well were transfected with either control siScr or siUSP1 and treated with a mixture of 22.2 mM glucose and 0.5 mM palmitate (HG/Pal) or pro-inflammatory cytokines (2 ng/mL recombinant human IL-1β, 1000 U/mL TNF-α, and 1000 U/mL IFN-γ; Cyto). Representative Western blot (A and K) and quantitative densitometry analysis (B) of γ-H2AX and p-p53 proteins are shown. (C–F) About 500,000 INS-1E cells/well treated with or without USP1 inhibitors ML323 (C and D) or SJB2-043 (E and F) were exposed to a mixture of 22.2 mM glucose and 0.5 mM palmitate (HG/Pal; C), 22.2 mM glucose (HG; E), or pro-inflammatory cytokines (2 ng/mL recombinant human IL-1β, 1000 U/mL TNF-α, and 1000 U/mL IFN-γ; Cyto; D and F) for 2 days. Data are pooled from at least three independent cell line experiments. Tubulin loading control in Figure 5F is re-used from the same experiment shown in Figure 3F. (G–J and L) Isolated human islets (150–200 islets/dish) treated with or without USP1 inhibitors ML323 or SJB2-043 were exposed to a mixture of 22.2 mM glucose and 0.5 mM palmitate (HG/Pal) for 2 days. (G–J) Human pancreatic islets were fixed, paraffin embedded, and stained for γ-H2AX (G) or p-p53 (I) and insulin. Representative images (G and I) and quantitative percentage of γ-H2AX or p-p53-positive β-cells (H and J) are shown. Scale bars represent 50 μm. Data are pooled from three human islet preparations except for J (n = 2). (L) Representative Western blot of p-p53 protein is shown. All lanes are from the same gel but were run noncontiguously. Data show means ±SEM. *p < 0.05 stimuli treated compared with control conditions. **p < 0.05 inhibitor-treated compared with treated conditions.

Figure 7

USP1 Inhibitors Reduced β-Cell Apoptosis and DDR in Human T2D Islets Isolated human islets from two patients with T2D treated with or without USP1 inhibitors ML323 or SJB2-043 for 1 day. (A–C) Human pancreatic islets were fixed, paraffin embedded, and double stained for the TUNEL assay (A), γ-H2AX (B), or p-p53 (C) and insulin. Representative confocal images and individual quantitative percentage of TUNEL, γ-H2AX, or p-p53 positive β-cells are shown from 2 to 4 technical replica/group from two islet donors with confirmed T2D. Scale bars represent 50 μm. Data are means of TUNEL-, γ-h2AX-, or pp53- and insulin-co-positive β-cells from pooled data of 2–4 different human islet sections spanning the whole islet pellet for each experimental group. About 150–200 islets were plated for each group. The mean number of β-cells scored was 6,050 for each condition for each donor.