Sox5 regulates beta-cell phenotype and is reduced in type 2 diabetes

Abstract

Type 2 diabetes (T2D) is characterized by insulin resistance and impaired insulin secretion, but the mechanisms underlying insulin secretion failure are not completely understood. Here, we show that a set of co-expressed genes, which is enriched for genes with islet-selective open chromatin, is associated with T2D. These genes are perturbed in T2D and have a similar expression pattern to that of dedifferentiated islets. We identify Sox5 as a regulator of the module. Sox5 knockdown induces gene expression changes similar to those observed in T2D and diabetic animals and has profound effects on insulin secretion, including reduced depolarization-evoked Ca²⁺-influx and β-cell exocytosis. SOX5 overexpression reverses the expression perturbations observed in a mouse model of T2D, increases the expression of key β-cell genes and improves glucose-stimulated insulin secretion in human islets from donors with T2D. We suggest that human islets in T2D display changes reminiscent of dedifferentiation and highlight SOX5 as a regulator of β-cell phenotype and function.

Figure 1. Co-expression network analysis and association between eigengene and type 2 diabetes traits.

(a) Symmetrically arranged heatmap of the topological overlap matrix for which the rows and columns are sorted by the hierarchical clustering tree used to define modules. The red square denotes the T2D-associated co-expression module. (b) Box plot showing the value of the eigengene for the 168 open chromatin genes in islets from non-diabetic (ND; n=45) and T2D donors (n=19). The box shows averages±s.e.m. and the error bars denote the tenth and ninetieth percentile, respectively. (c–e) The module eigengene of the 168 open chromatin genes displayed against HbA1c (c; n=52), glucose-stimulated insulin secretion (d; n=48) and K⁺-stimulated insulin secretion (e; n=26). Statistical comparisons using linear regression. (f) The connectivity k_in of each gene is displayed against the r value for the Pearson correlation between the gene expression trait and T2D status. Grey dots denote genes in the T2D-associated module and red dots denote genes with islet-selective open chromatin. Data are from human islets from 64 donors. (g) Cumulative density function (CDF) plots of log2-transformed gene expression fold-change in freshly isolated versus expanded islets in microarrays from GSE15543. The blue line denotes the fold-change of the 168 open chromatin genes in GSE15543 and the purple line denotes the fold-change of the remaining genes in the array. (h) CDF plot of log2-transformed expression fold-change of genes in the T2D signature in Pdx1⁺/Ins^low (immature) versus Pdx1^high/Ins^high (mature) human β-cells. The CDF plot of the 168 signature genes in T2D islets is also displayed.

Figure 2. Characterization of Sox5 expression and effects of Sox5 knockdown.

The experiments were performed in INS-1 832/13 cells if not stated otherwise. Sox5-kd cells were transfected with Sox5 siRNA 48 h before the experiment, and control cells were transfected with negative control siRNA. (a) Fold-change of insulin secretion at 16.7 mM glucose compared to control-treated cells in response to gene silencing. Data for each siRNA is compared with cells treated with a negative control siRNA (n=3–5 experiments per siRNA). P values are corrected for multiple comparisons. Tmem196 was excluded due to undetectable expression. (b) Expression of Pdx1 and Mafa mRNA following Sox5 knockdown (Sox5-kd) relative to control cells (n=8). (c) Secreted insulin normalized to cellular protein content during a 1-h stimulation with glucose as indicated (n=4). (d) Ratio of secreted proinsulin to insulin following a 1-h stimulation with 2.8 or 16.7 mM glucose (G) in Sox5-kd and control cells (n=4 samples in 1 experiment). (e) Fold-change of insulin secretion in Sox5-kd and control cells after a 1-h incubation at 2.8 mM glucose with 10 mM L-leucine (L-leu) or 10 mM α-ketoisocaproic acid (α-KIC) relative to 2.8 mM glucose (n=3 per condition). (f) Insulin secretion normalized to cellular protein content in Sox5-kd and control cells in response to glucose (G), 50 mM K⁺ and 200 μM tolbutamide (tolb) as indicated (n=3 per condition). (g) Representative electron micrographs of human islets from T2D donors with high or low SOX5 expression (5–9 cells analysed from each of 6 donors with T2D). Mitochondria (M), plasma membrane (PM), insulin granules (G), nucleus (N) and α- and β-cells are indicated. Granules were defined as docked (indicated by arrows) when located within 150 nm from the PM. Scale bar, 2 μm in the lower and 0.5 μm in the larger magnification (corresponding to the marked area). Data are mean±s.e.m. *P<0.05; **P<0.01; ***P<0.001 using Student's t-test.

Figure 3. Metabolic characterization of Sox5-kd cells.

The experiments were performed in INS-1 832/13. Sox5-kd cells were transfected with Sox5 siRNA 48 h before the experiment, and control cells were transfected with negative control siRNA. (a) Mitochondrial oxygen consumption rate (OCR) measured by Seahorse XF24. Data in each group are normalized to the values at 2.8 mM glucose (n=3). Glucose, pyruvate, oligomycin (which inhibits ATP synthase in order to assess the mitochondrial proton leak), the uncoupler FCCP (which maximizes respiration) and rotenone (which inhibits the electron transport chain) were added as indicated. The bar graph shows average mitochondrial OCR in control (Ctrl) and Sox5-kd cells (n=3). (b) Fold-change of glucose-6-phosphate (Glucose-6-P), fructose-6-phosphate (Fructose-6-P), hexose phosphates and fumarate at 2.8 mM glucose in Sox5-kd cells versus non-treated control cells (NT) measured by gas chromatography/mass spectrometry (n=5). (c) Levels of aspartate, glutamate, glycerol-3-phosphate (Glycerol-3-P), lactate and alanine at 2.8 and 16.7 mM glucose in Sox5-kd and control cells. Data are normalized to control cells at 2.8 mM glucose. (n=5) Data are mean±s.e.m. *P<0.05; **P<0.01; ***P<0.001 using Student's t-test.

Figure 4. Effects of Sox5-kd on exocytosis and Ca²⁺ currents.

The experiments were performed in INS-1 832/13 cells unless stated otherwise. Sox5-kd cells were transfected with Sox5 siRNA 48 h before the experiment. Control cells were transfected with negative control siRNA. (a) Increase in cell capacitance (ΔC), reflecting exocytosis, evoked by a train of ten 500 ms depolarizations from −70 to 0 mV applied to simulate glucose-induced electrical activity. Bar graph shows total capacitance increase (ΣΔC) in Sox5-kd and control (Ctrl) cells. Data from 21–25 cells per group. (b) Capacitance increase (ΣΔC) in response to the first two depolarizations of the train (‘early exocytosis'). Data are from the same cells as in a. (c) The Ca²⁺ current (I) in response to the first depolarization of the train. The bars denote the integrated Ca²⁺-current (charge) in Sox5-kd and control cells. Data from the same cells as in a. (d) Average Ca²⁺-sensitivity of the exocytotic process (defined as the capacitance increase divided by the integrated Ca²⁺ current in response to the first depolarization) in Sox5-kd and control cells. Data from the same cells as in a. (e) Average capacitance increase in response to infusion of a solution containing high Ca²⁺ concentration (free [Ca²⁺]_i∼1.5 μM) (n=10–11). (f) Representative recordings of Fluo-5F fluorescence in Sox5-kd and control cells. Data are presented as the ratio of Fluo-5F fluorescence (F_i) normalized to the basal values at 2.8 mM glucose in control cells (F₀). The bar graph shows the area under the curve (AUC) of the fluorescence signal during stimulation with 20 mM glucose (n=10–25). (g) Current–voltage relationship for the peak Ca²⁺ current in the absence or presence of 2 μM isradipine (isr) in Sox5-kd and control cells (n=8–15 cells). The isradipine-sensitive component was obtained by subtracting currents recorded in the presence of isradipine from currents observed in the absence of the blocker. (h) Total capacitance increase (ΣΔC) normalized for cell size in human β-cells treated with siRNA targeting SOX5 (SOX5-kd) and control (Ctrl) cells (n=6–9 cells). One-sided t-test was used for statistical analysis. Data are mean±s.e.m. *P<0.05; **P<0.01 using Student's t-test.

Figure 5. Expression of the T2D-associated module after SOX5 perturbation and in animal models.

The cumulative density function (CDF) plot of the 168 open chromatin genes in human islets from T2D versus non-diabetic donors is displayed as a reference (dashed line) in all panels. (a) Cumulative density function (CDF) plots of log2-transformed expression fold-change of genes in the T2D-associated module in INS-1 832/13 cells transfected with Sox5 siRNA (Sox5-kd; red) or a Sox5 plasmid (green) relative to control cells (n=3). (b) CDF plots of log2-transformed gene expression fold-change for the T2D-associated module in islets from 4-week-old db/db mice (n=3) versus db/+ littermates (n=5) and from phlorizin-treated db/db mice (n=3) versus untreated db/db mice (n=5). Phlorizin was administered at 400 mg kg⁻¹ daily for 10 days. (c–f) As in (b) but for 10-week-old db/db mice treated with phlorizin (400 mg kg⁻¹ daily) for 7 days (n=7 untreated db/db and n=3 phlorizin-treated db/db) (c), for 14-month-old C57BL/6 mice (n=6) versus 8-week-old C57BL/6 mice (n=5) (d), for 6 or 13-week-old ob/ob mice versus ob/+ littermates (n=5 in each group) (e), and for human islets incubated at 20 mM glucose for 72 h versus islets incubated at 5.6 mM glucose (n=6) (f).

Figure 6. Characterization of the regulation of Sox5.

The experiments were performed in INS-1 832/13 cells if not stated otherwise. Sox5-kd cells were transfected with Sox5 siRNA 48 h before the experiment, and control cells were transfected with negative control siRNA. (a) Sox5 mRNA levels relative to control (Ctrl) cells after knockdown of putative regulators of Sox5 expression (n=4). (b) Yy1 mRNA levels after 48 h incubation with different concentrations of palmitate (n=3). (c) Effect of palmitate on Sox5 expression in control cells and after knockdown of Yy1 (Yy1-kd) (n=3). (d) Sox5 mRNA levels after 48 h treatment with different concentrations of valproic acid (VPA) relative to untreated cells (n=3). (e) Insulin secretion at 16.7 mM glucose after incubation with or without VPA for 48 h in non-transfected cells (NT), cells transfected with a negative control siRNA (Ctrl) and Sox5-kd cells (n=4 per condition). Insulin secretion was plotted against log2-transformed expression levels of Sox5 for each of the 12 conditions indicated in the bar graph, with values from three independent experiments. (f) Cumulative density function (CDF) plot of log2-transformed expression fold-change of genes in the T2D-associated module in cells treated with 1 mM VPA relative to control cells. (g) Total capacitance increase (ΣΔC) in cells treated with 0.3 mM VPA for 48 h relative to untreated cells (n=12–14 cells). (h) Serum insulin concentration at 0, 30 and 120 min during an intraperitoneal glucose tolerance test in 12-week-old female NMRI mice treated with 250 mg kg⁻¹ VPA or vehicle (Ctrl) daily for 7 days (n=3 mice per condition). Insulin levels were increased by 37% at 30 min and by 64% at 120 min of the IPGTT (P=0.046 and 0.037, respectively. All mice were normoglycemic and there was no difference in glucose tolerance between the groups. Data are mean±s.e.m. *P<0.05; **P<0.01; ***P<0.001. Student's t-test was used in a,b,c,d,g,h, and Pearson correlation test was used in e.

Figure 7. Effects of SOX5 knockdown in a human β cell line and of SOX5 overexpression in human islets.

(a) Insulin secretion (as per cent of total insulin content) from EndoC-BH1 cells during a 1-h incubation with 1 or 20 mM glucose after lentiviral RNAi-mediated SOX5 knockdown (n=3). (b) Insulin secretion (as % of total insulin content) from batch-incubated human islets after SOX5 lentiviral overexpression relative to islets transduced with control lentivirus. Islets were incubated for 1 h with glucose (G) and for 15 min with K⁺ as indicated (n=3–9 replicates from eight donors per condition, of which three had T2D). (c) Insulin secretion (as % of total insulin content) during a 1-h incubation with 16.7 mM glucose in non-diabetic and T2D islets with or without SOX5 overexpression, respectively (n=4–9 replicates from five non-diabetic donors and three T2D donors). (d) mRNA expression of genes in human islets after SOX5 overexpression relative to control islets (n=10 donors). Log2-transformed gene expression data were used for statistical analysis (paired t-test). Data are mean±s.e.m. *P<0.05; **P<0.01 using Student's t-test.