Enhancer of Zeste Homolog 2 (EZH2) Mediates Glucolipotoxicity-Induced Apoptosis in β-Cells

Abstract

Selective inhibition of histone deacetylase 3 (HDAC3) prevents glucolipotoxicity-induced β-cell dysfunction and apoptosis by alleviation of proapoptotic endoplasmic reticulum (ER) stress-signaling, but the precise molecular mechanisms of alleviation are unexplored. By unbiased microarray analysis of the β-cell gene expression profile of insulin-producing cells exposed to glucolipotoxicity in the presence or absence of a selective HDAC3 inhibitor, we identified Enhancer of zeste homolog 2 (EZH2) as the sole target candidate. β-Cells were protected against glucolipotoxicity-induced ER stress and apoptosis by EZH2 attenuation. Small molecule inhibitors of EZH2 histone methyltransferase activity rescued human islets from glucolipotoxicity-induced apoptosis. Moreover, EZH2 knockdown cells were protected against glucolipotoxicity-induced downregulation of the protective non-canonical Nuclear factor of kappa light polypeptide gene enhancer in B-cells (NFκB) pathway. We conclude that EZH2 deficiency protects from glucolipotoxicity-induced ER stress, apoptosis and downregulation of the non-canonical NFκB pathway, but not from insulin secretory dysfunction. The mechanism likely involves transcriptional regulation via EZH2 functioning as a methyltransferase and/or as a methylation-dependent transcription factor.

Keywords: ER stress; GLT; NFκB; diabetes; histone deacetylases; histone methyltransferase; insulin secretion.

Figure 1

Enhancer of zeste homolog 2 (EZH2) putatively binds gene targets regulated by glucolipotoxicity (GLT) and BRD3308. (A) mRNA isolated from INS-1E cells exposed to 25 mmol/L glucose and 0.5 mmol/L palmitate (GLT) with or without 10 μmol/L histone deacetylase 3 (HDAC3) inhibitor BRD3308 was spotted in triplicate onto an Affymetrix gene array chip (n = 1 array). On the heatmap, yellow represent upregulated genes and purple represents downregulated genes relative to the control, with the scale shown in the upper left corner. (A) shows the expression pattern of the 52 genes identified as regulated both by GLT and by BRD3308. See gene names and regulation patterns in Table S1. (B) Genes from (A) were analyzed by the ENCODE ChIP-Seq Significance Tool, and only EZH2 was identified as a factor binding 22 of the genes (top panel). The bottom panel shows a description of the subset of genes from (A) that was identified to contain EZH2 binding sites. The Tool employs a hypergeometric test with multiple hypothesis correcting using Benjamini–Hochberg to calculate binding significance. List1: Input list of the 52 genes from (A).

Figure 2

Moderate Enhancer of zeste homolog 2 (EZH2) attenuation protects against glucolipotoxicity (GLT)-induced apoptosis. (A) mRNA expression of Ezh2 was measured by real-time quantitative PCR in INS-1E cells expressing different shRNAs targeting EZH2 (EZH2 KDs; sh1, -3 and -4). Data presented as means + SEM of n = 4, analyzed by one-way ANOVA with Sidak’s multiple comparisons test (KDs vs. ev). For (A–C): ev, INS-1E cells transfected with empty vector; sh1, INS-1E cells transfected with shRNA sequence #1; sh3, INS-1E cells transfected with shRNA sequence #3; sh4, INS-1E cells transfected with shRNA sequence #4. (B) EZH2 protein expression was detected by Western blot in EZH2 KDs sh1, -3 and -4. Representative blot of n = 10. Data presented as means + SEM of n = 10, analyzed by one-way ANOVA with Sidak’s multiple comparisons test (KDs vs. ev). (C) Fifty-thousand EZH2 KDs sh1, -3 and -4 were exposed to 25 mmol/L glucose and 0.5 mmol/L palmitate (GLT) for 24 h. Apoptosis was detected as cytoplasmic accumulation of mono- and oligonucleosomes. Data presented as means + SEM of n = 5–7, analyzed by one-way ANOVA with Sidak’s multiple comparisons test. (D) mRNA expression of Ezh2 was measured by real-time quantitative PCR in INS-1E cells expressing a gRNA targeting exon 3 of EZH2 (EZH2 HET). Data presented as means + SEM of n = 3, analyzed by paired Student’s t-test. For (D,E): ctrl, failed INS-1E CRISPR clone; EZH2 HET, CRISPR/Cas9 modified INS-1E with gRNA targeting EZH2 exon 3. (E) Fifty-thousand EZH2 HET cells were exposed to 25 mmol/L glucose and 0.5 mmol/L palmitate (GLT) for 24 h. Apoptosis was detected as above. Data presented as means + SEM of n = 4, analyzed by one-way ANOVA with Sidak’s multiple comparisons test. (F) Twenty-five human islets in duplicate were exposed to 25 mmol/L glucose and 0.5 mmol/L palmitate (GLT) in the presence or absence of 0.01 μmol/L (n = 2) and 0.1–1 μmol/L (n = 5) EZH2 inhibitor GSK126 or GSK343 for 72 h. Apoptosis measured as above. Data presented as means + SEM of n = 2–5, analyzed by one-way ANOVA with post hoc t-test. Islet donors: 1, 2, 3, 5, 7 (Table S2). ns: not significant; * p < 0.05; ** p < 0.01; *** p < 0.001; **** p < 0.0001.

Figure 3

Moderate Enhancer of zeste homolog 2 (EZH2) attenuation protects against glucolipotoxicity (GLT)-induced endoplasmic reticulum (ER) stress. INS-1E cells expressing shRNAs targeting EZH2 (sh3 and -4) were exposed to 25 mmol/L glucose and 0.5 mmol/L palmitate (GLT) or vehicle for 3–24 h (A–D) and mRNA expression of ER stress markers Activating transcription factor 4 (Atf4) (A), Spliced X-box-binding protein-1 (sXbp1) (B), Binding immunoglobulin protein (Bip) (C) and C/EBP homologous protein (Chop) (D) was measured by real-time quantitative PCR. Data presented as means + SEM of n = 5. (E) INS-1E cells expressing shRNAs targeting EZH2 (sh1, -3 and -4) were exposed to 25 mmol/L glucose and 0.5 mmol/L palmitate (GLT) for 24 h and protein expression of Chop was measured by Western blot. Representative blot of n = 5. Data presented as means + SEM of n = 5. ev: INS-1E cells transfected with empty vector; sh1: INS-1E cells transfected with shRNA sequence #1; sh3: INS-1E cells transfected with shRNA sequence #3; sh4: INS-1E cells transfected with shRNA sequence #4. (A–D) analyzed by one-way ANOVA with Sidak’s multiple comparisons test. (E) analyzed by paired Student’s t-test. * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001 ev vs. ev + GLT; ^¤ p < 0.05, ^¤¤ p < 0.01, ^¤¤¤ p < 0.001, ^¤¤¤¤ p < 0.0001 sh3 vs. sh3 + GLT (A–D left panels) or sh4 vs. sh4 + GLT (A–D right panels); ^# p < 0.05, ^## p < 0.01, ^#### p < 0.0001 ev + GLT vs. sh3 + GLT (A–D left panels) or ev + GLT vs. sh4 + GLT (A–D right panels).

Figure 4

No consistent effect of Enhancer of zeste homolog 2 (EZH2) attenuation on insulin secretion and content under glucolipotoxic (GLT) conditions. Fifty-thousand INS-1E cells expressing shRNAs targeting EZH2 (sh3, upper panel and -4, lower panel; (A)) or gRNA targeting exon 3 of EZH2 (EZH2 HET; (B)) were exposed to 25 mmol/L glucose and 0.5 mmol/L palmitate (GLT) for 24 h. Insulin was detected in supernatants by ELISA. Glucose-stimulated insulin secretion (C) and insulin content (D) from twenty human islets exposed to 25 mmol/L glucose and 0.5 mmol/L palmitate (GLT) or vehicle in the presence or absence of 0.1–1 μmol/L EZH2 inhibitor GSK126 or GSK343 for 72 h. Islet donors: 6–10 (Table S2). (A,B) Data presented as means + SEM of n = 4, analyzed by one-way ANOVA with Sidak’s multiple comparisons test. ev: INS-1E cells transfected with empty vector; sh3: INS-1E cells transfected with shRNA sequence #3; sh4: INS-1E cells transfected with shRNA sequence #4; ctrl: failed INS-1E CRISPR clone; EZH2 HET: CRISPR/Cas9 modified INS-1E with gRNA targeting EZH2 exon 3. (C,D) Data presented as means + SEM of n = 5. (C) Analyzed by one-way ANOVA with post hoc t-test. ns: not significant; * p < 0.05.

Figure 5

Enhancer of zeste homolog 2 (EZH2) knockdown prevents downregulation of the non-canonical Nuclear factor of kappa light polypeptide gene enhancer in B-cells (NFκB) pathway. (A) INS-1E cells expressing shRNAs targeting EZH2 (sh3 and -4) were exposed to 25 mmol/L glucose and 0.5 mmol/L palmitate (GLT) or vehicle (CTRL). Expression levels of phosphorylated p100 (P-p100) and nuclear factor of kappa light polypeptide gene enhancer in B-cells inhibitor, alpha (IκBα) were detected by Western blot. Data presented as means + SEM. Representative blot of n = 5–7, analyzed by paired Student’s t-test. ev: INS-1E cells transfected with empty vector; sh3: INS-1E cells transfected with shRNA sequence #3; sh4: INS-1E cells transfected with shRNA sequence #4. (B) One-hundred mouse islets were exposed to 25 mmol/L glucose and 0.5 mmol/L palmitate (GLT) or vehicle in the presence or absence of 1 μmol/L EZH2 inhibitor GSK126 for 48 h. B-cell lymphoma 2 (Bcl2) mRNA expression was detected by real-time quantitative PCR. Data presented as means + SEM of n = 3, analyzed by one-way ANOVA with Sidak’s multiple comparisons test. ns: not significant; * p < 0.05.

Figure 6

Proposed model of Enhancer of zeste homolog 2 (EZH2)/histone deacetylase 3 (HDAC3) interaction and EZH2 function in the β-cell. (A) Panel (i): HDAC3 targets EZH2 by removal of the activating acetyl (Ac) mark, leading to reduced EZH2 histone-lysine N-methyltransferase (HMT) activity on histone 3, lysine 27 (H3K27) and increased transcription of deleterious genes. Panel (ii): HDAC3 and EZH2 function as co-activators. HDAC3 catalyzes removal of the histone 3, lysine 4 (H3K4) acetyl-group, leading to EZH2 H3K4 methylation (Me3) and transcriptional activation of deleterious genes. Panel (iii): HDAC3 and EZH2 function as co-repressors. HDAC3 catalyzes the removal of the acetyl-group from H3K27, leading to EZH2 H3K27 methylation and transcriptional repression. (B) Proposed model of EZH2-mediated apoptosis in the β-cell. Elevated concentrations of glucose and palmitate (GLT) indicated by red arrows induce EZH2-mediated endoplasmic reticulum (ER) stress via increased Activating transcription factor 4 (Atf4) and Spliced X-box-binding protein-1 (sXbp1) mRNA and C/EBP homologous protein (Chop) protein expression, and EZH2 potentiates apoptotic signaling by downregulating protective non-canonical Nuclear factor of kappa light polypeptide gene enhancer in B-cells (NFκB) signaling.