Agonist-independent Gαz activity negatively regulates beta-cell compensation in a diet-induced obesity model of type 2 diabetes

Abstract

The inhibitory G protein alpha-subunit (Gα_z) is an important modulator of beta-cell function. Full-body Gα_z-null mice are protected from hyperglycemia and glucose intolerance after long-term high-fat diet (HFD) feeding. In this study, at a time point in the feeding regimen where WT mice are only mildly glucose intolerant, transcriptomics analyses reveal islets from HFD-fed Gα_z KO mice have a dramatically altered gene expression pattern as compared with WT HFD-fed mice, with entire gene pathways not only being more strongly upregulated or downregulated versus control-diet fed groups but actually reversed in direction. Genes involved in the "pancreatic secretion" pathway are the most strongly differentially regulated: a finding that correlates with enhanced islet insulin secretion and decreased glucagon secretion at the study end. The protection of Gα_z-null mice from HFD-induced diabetes is beta-cell autonomous, as beta cell-specific Gα_z-null mice phenocopy the full-body KOs. The glucose-stimulated and incretin-potentiated insulin secretion response of islets from HFD-fed beta cell-specific Gα_z-null mice is significantly improved as compared with islets from HFD-fed WT controls, which, along with no impact of Gα_z loss or HFD feeding on beta-cell proliferation or surrogates of beta-cell mass, supports a secretion-specific mechanism. Gα_z is coupled to the prostaglandin EP3 receptor in pancreatic beta cells. We confirm the EP3γ splice variant has both constitutive and agonist-sensitive activity to inhibit cAMP production and downstream beta-cell function, with both activities being dependent on the presence of beta-cell Gα_z.

Keywords: G protein; G protein–coupled receptor (GPCR); cAMP; cell signaling; diabetes; insulin resistance; insulin secretion.

Figure 1

Sixteen-week HFD feeding of C57BL/6N mice results in mild hyperglycemia and glucose intolerance ameliorated by loss of Gα_zindependent of effects on insulin sensitivity.A–B, baseline metabolic phenotypes of 11-week-old chow-fed WT or KO mice. A, oral glucose tolerance (1 g/kg) as represented by glucose excursion (left) or AUC from zero (right). Sixteen mice/group. B, IP insulin tolerance (0.75 U/kg) as represented by glucose excursion normalized to baseline (left) or AUC of the normalized data from zero (right). N=9 to 12 mice/group. C–F, metabolic phenotypes of 27-week-old WT and KO mice after 16 weeks of control diet (CD) or high-fat diet (HFD) feeding. N = 7 to 10 mice/group. C, body weights. D, fasting blood glucose levels of 4 to 6 h. E, oral glucose tolerance (1 g/kg) as represented by glucose excursion (left) or AUC from zero (right). F, IP insulin tolerance (0.75 U/kg) as represented by glucose excursion normalized to baseline (left) or AUC of the normalized data from zero (right). Data represent mean ± SD and were analyzed as described in Experimental procedures. ∗p < 0.05; ∗∗p < 0.01; ∗∗∗∗p < 0.0001. In panel E, ##p < 0.01 for KO HFD versus WT CD at the 15-min time point. AUC, area under the curve; Gα_z, G protein alpha-subunit; ns, not significant.

Figure 2

Sixteen-week high-fat diet (HFD) feeding of Gα_zKO mice results in wholesale alterations in gene expression profiles as compared with WT HFD-fed controls.A–C, the transcriptomics profile of isolated islets from WT and Gα_z KO islets after 16-week HFD or control diet (CD) feeding. N = 3 to 5 independent islet preparations/group. A, the heatmap indicating the top 100 most variable significantly differentially expressed genes across the four groups (Benjamini–Hochberg–adjusted p-value < 0.05). Z-score indicates SDs away from the mean log cpm in either a positive or negative direction. Each column represents a single mouse. B, four-set Venn diagram representing the total number of differentially expressed genes between groups. C, principal component analysis of microarray data. The numbers are indicative of mouse ID. D–E, validation of differentially upregulated (D) and downregulated (E) genes by qRT-PCR. Three independent islet preparations per group. Data represent mean ± SD and were compared within each gene by unpaired t-test. p-values > 0.05 are labeled. ∗p < 0.05; ∗∗p < 0.01. AUC, area under the curve; Gα_z, G protein alpha-subunit; qRT-PCR, quantitative real-time PCR.

Figure 3

Sixteen-week HFD feeding of Gα_zKO mice significantly alters islet secretion–related gene profiles: an effect that correlates with enhanced insulin secretion at study end.A–D, transcriptomics profile of isolated islets from WT and Gα_z KO islets after 16-week high-fat diet (HFD) or control diet (CD) feeding. N = 3 to 5 independent islet preparations/group. A, the heatmap indicating the relative expression of genes involved in the most significantly enriched biological Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways based on genes differentially expressed in islets from HFD-fed WT versus KO mice (q < 0.05, false-discovery rate). Genes in more than one significantly enriched KEGG pathway are listed only once and assigned to the most significantly affected pathway. B, pathway enrichment analysis of KEGG pathways significantly upregulated or downregulated islets from HFD-fed WT versus HFD-fed KO mice. Colors are matched to those of the pathways shown in panel A. C, pathway enrichment analysis performed as shown in panel B showing HFD feeding has an inverse effect on significantly affected pathways in islets from WT versus KO mice. Colors are matched to those of the pathways shown in panel A. D, significantly altered genes in the pancreatic secretion pathway in islets from HFD-fed KO versus WT mice. Data were analyzed as described in Experimental procedures. E–F, insulin secretion (E) and glucagon secretion (F), both as normalized to total protein, from islets isolated from WT and Gα_z KO mice after 25 to 29 weeks of HFD feeding (8). N = 5 to 7 independent islet preparations/group. Data represent mean ± SD and were compared by unpaired t-test. ∗p < 0.05; ∗∗p < 0.01. Gα_z, G protein alpha-subunit.

Figure 4

Beta cell–specific loss of Gα_zprotects C57BL/6J mice from developing hyperglycemia and glucose intolerance after 28 weeks of HFD feeding independent of effects on insulin sensitivity.A–B, baseline metabolic phenotypes of 11-week-old chow-fed WT and βKO mice. A, oral glucose tolerance (1 g/kg) as represented by glucose excursion (left) or AUC from zero (right). Twelve to 19 mice/group. B, IP insulin tolerance (0.75 U/kg) as represented by glucose excursion normalized to baseline (left) or AUC of the normalized data from zero (right). N = 9 to 18 mice/group. C–F, metabolic phenotypes of 39-week-old WT and βKO mice after 28 weeks of control diet (CD) or high-fat diet (HFD) feeding. Ten to 14 mice/group. C, body weights. D, 4 to 6 h fasting blood glucose levels. E, oral glucose tolerance (1 g/kg) as represented by glucose excursion (left) or AUC from zero (right). F, IP insulin tolerance (0.75 U/kg) as represented by glucose excursion normalized to baseline (left) or AUC of the normalized data from zero (right). Data represent mean ± SD and were analyzed as described in Experimental procedures. ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001; ∗∗∗∗p < 0.0001. In panel E, #p < 0.05 for KO HFD versus WT CD at the 15-min time point. ˆ in panels B and F indicate time points in which young and/or CD-fed mice of both genotypes had to be rescued by glucose injection during ITTs (n = 2 of 18 WT and 1 of 9 βKO in panel B and 5 of 12 WT and 5 of 10 βKO in panel F). AUC, area under the curve; Gα_z, G protein alpha-subunit; ITT, insulin tolerance tests; ns, not significant.

Figure 5

Beta-cell replication is not affected by HFD feeding or loss of Gα_zin insulin-sensitive B6J mice.A, quantification of Ki67-positive beta cells by immunofluorescence of pancreas slide sections from WT and βKO mice after CD or HFD feeding. Data represent the percentage total number of beta cells with Ki67-positive fluorescence. All islets from 2 pancreas sections of 3 mice of each group were used in this analysis (n = 3). B, raw values for %Ki67-positive beta cells from panel A as plotted on the same Y-axis as the original HFD study in full-body Gα_z KO mice in B6N background, with the mean values (8) indicated by labeled lines. C, the beta-cell fractional area, or the insulin-positive area as a percentage of the total pancreas area from slide sections. Three pancreases/group. D, insulin-positive slide area from the 28-week B6J HFD-feeding study described in this work and the previously published 26- to 30-week B6N HFD-feeding study (8), both as normalized to their own WT CD group. N = 3 to 4 pancreas/group. E, the insulin content of islets isolated from WT and βKO mice after CD or HFD feeding. N = 3 to 9 independent islet preparations/group. F, compiled area under the curve (AUC) analyses of insulin tolerance tests from WT B6J mice (Figs. 1 and 4) and WT B6N mice (Fig. 1 and (8)) at baseline (BL: 11 weeks of age) and after 26 to 30 weeks of CD or HFD feeding. Data represent mean ± SD. Data in panels A and C–E were analyzed by one-way ANOVA with the following preselected comparisons: WT CD versus WT HFD, WT CD versus KO CD, KO CD versus KO HFD, and WT HFD versus KO HFD. The Holm–Sidak test was used post hoc to correct for multiple comparisons. In panel F, data were analyzed by two-way ANOVA with the Holm–Sidak test post hoc to correct for multiple comparisons. ∗p < 0.05, ∗∗∗p < 0.01, and ∗∗∗∗p < 0.0001. Gα_z, G protein alpha-subunit; ns, not significant.

Figure 6

Loss of beta-cell Gα_zpreserves islet glucose and incretin responsiveness in HFD-fed mice and ablates the inhibitory effect of EP3γ on islet GSIS and cAMP production.A, insulin secreted from WT and βKO mice after 28 weeks of CD or HFD feeding. Islets were treated with the indicated glucose concentrations with or without the addition of exendin-4 (Ex4) or sulprostone. N = 3 to 9 independent islet preparations per group. B, the proportion of islet preparations in panel A with a higher insulin secretion response with Ex4 than without. C, relative islet mRNA expression of the Ptger3 gene or the EP3γ splice variant (Ptger3γ) as measured by qRT-PCR. Cycle times were normalized within each group to beta actin, and the fold change in expression versus WT CD calculated via the 2^ΔΔCt method. N = 4 to 6 independent islet preparations/group. D, insulin secreted in stimulatory glucose ± sulprostone in islets isolated from WT and βKO B6J mice (left) or WT and KO B6N mice (right) when islets are adenovirally overexpressing HA-EP3γ or GFP control. E, intracellular cAMP levels ([cAMP]_i) in islets from GFP- or HA-EP3γ-expressing WT and ΚΟ B6N mice in response to IBMX ± sulprostone. F, insulin secreted from the islets shown in panel E. In panels D–F, n = 3 independent islet preparations per group. Data represent mean ± SD and were analyzed as described in Experimental procedures. ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001. G, the model of the signaling pathways studied in this work and whether they are known to be downregulated (red minus signs) or upregulated (green plus signs) in the beta-cell dysfunction of T2D. The asterisk after EP3γ indicates its partial constitutive activity. B6J, C57BL/6J; B6N, C57BL/6N; CD, control diet; Gα_z, G protein alpha-subunit; GSIS, glucose-stimulated insulin secretion; HA-EP3γ, hemagglutinin-tagged EP3γ; HFD, high-fat diet; IBMX, 3-isobutyl-1-methylxanthine; ns, not significant; qRT-PCR, quantitative real-time PCR; T2D, type 2 diabetes.