Glycerol-3-phosphate phosphatase operates a glycerol shunt in pancreatic β-cells that controls insulin secretion and metabolic stress

Abstract

Objective: The recently identified glycerol-3-phosphate (Gro3P) phosphatase (G3PP) in mammalian cells, encoded by the PGP gene, was shown to regulate glucose, lipid and energy metabolism by hydrolyzing Gro3P and to control glucose-stimulated insulin secretion (GSIS) in β-cells, in vitro. However, whether G3PP regulates β-cell function and insulin secretion in vivo is not known.

Methods: We now examined the role of G3PP in the control of insulin secretion in vivo, β-cell function and glucotoxicity in inducible β-cell specific G3PP-KO (BKO) mice. Inducible BKO mice were generated by crossing floxed-G3PP mice with Mip-Cre-ERT (MCre) mice. All the in vivo studies were done using BKO and control mice fed normal diet and the ex vivo studies were done using pancreatic islets from these mice.

Results: BKO mice, compared to MCre controls, showed increased body weight, adiposity, fed insulinemia, enhanced in vivo GSIS, reduced plasma triglycerides and mild glucose intolerance. Isolated BKO mouse islets incubated at high (16.7 mM), but not at low or intermediate glucose (3 and 8 mM), showed elevated GSIS, Gro3P content as well as increased levels of metabolites and signaling coupling factors known to reflect β-cell activation for insulin secretion. BKO islets also showed reduced glycerol release and increased O₂ consumption and ATP production at high glucose only. BKO islets chronically exposed to elevated glucose levels showed increased apoptosis, reduced insulin content and decreased mRNA expression of β-cell differentiation markers, Pdx-1, MafA and Ins-2.

Conclusions: The results demonstrate that β-cells are endowed with a "glycerol shunt", operated by G3PP that regulates β-cell metabolism, signaling and insulin secretion in vivo, primarily at elevated glucose concentrations. We propose that the glycerol shunt plays a role in preventing insulin hypersecretion and excess body weight gain and contributes to β-cell mass preservation in the face of hyperglycemia.

Keywords: Glucodetoxification; Glucose-stimulated insulin secretion; Glycerol shunt; Glycerol-3-phosphate phosphatase; Obesity; Pancreatic beta cell; Type 2 diabetes.

Figure 1

Increased body weight gain, fat mass and fed insulinemia with reduction in fed plasma TG in male BKO mice. (A) Schematic representation of gene targeting strategy for the generation of the G3PP-lox conditional allele. Wild-type, targeted and G3PP-lox alleles are represented. The loxP recombination sites inserted at the 5′ side of exon 1 and 3′ side of exon 2 in PGP gene, the FRT-flanked Neo-cassette FRT sites, and the NDEL1/NDEL2 primers for PCR genotyping are indicated. The breeding of mice to produce BKO mice is described in Methods. (B) The presence of the WT, Mip-Cre transgene or the floxed PGP alleles was evaluated in DNA from ear punch tissue fragments using the primers described in Methods. A specific amplification of a 452 bp DNA fragment corresponds to the floxed- PGP allele, a 312 bp fragment corresponds to the WT PGP allele and the presence of a 267 bp DNA fragment corresponds to the Mip-Cre transgene, as verified by genomic PCR. (C) G3PP deletion was validated using protein extracts from islets, brain, liver and skeletal muscle (SM) from mice, 4 weeks after deletion and used for western blot (WT, n = 4; fl/fl, n = 3; MCre, n = 3; BKO, n = 3). For all tissues α-tubulin was used as a housekeeping protein. Mice were kept on chow diet for 12 weeks after G3PP deletion following TMX treatment. (D) Body weight (Mcre, n = 6; BKO, n = 10). (E) Percentage of body weight gain (MCre, n = 6; BKO, n = 10). (F) Lean and fat mass expressed as percentage of BW (MCre, n = 7; BKO, n = 7 (one BKO mouse out of 8, died during experiment)). (G) Food intake (MCre, n = 6; BKO, n = 10). (H) Plasma parameters in fed state (For glycemia: MCre, n = 9 and BKO = 9; for insulin, C-peptide, TG, glycerol and FFA, MCre, n = 18; BKO, n = 19). (I) Weights of visceral adipose tissue (VAT), subcutaneous adipose tissue (SAT), liver, brown adipose tissue (BAT) and brain (MCre, n = 6; BKO, n = 7). (J) TG content for liver, VAT and skeletal muscle (SM) (MCre, n = 6; BKO, n = 7). Data are mean ± SEM. ∗p < 0.05, ∗∗p < 0.01 vs MCre (Two-way ANOVA (Panel D and G) and Student's t test (Panel E, F, H, I and J)).

Figure 2

Increased in vivo glucose induced insulin secretion with mild glucose intolerance and normal insulin sensitivity in BKO mice. (A) Glycemia during IPGTT in male mice, 16 weeks after G3PP deletion (MCre, n = 8; BKO, n = 10). Inset depicts area under the curve (AUC) for glycemia after 15 min and 60 min. (B) Insulinemia during IPGTT in male mice (MCre, n = 8; BKO, n = 10). Inset depicts AUC for insulinemia after 15 min and 60 min. (C) Glycemia during ITT in mice, 19 weeks after G3PP deletion (MCre, n = 6; BKO, n = 10). Inset depicts area above the curve (AAC) after 120 min. Data are means ± SEM. ∗p < 0.05, ∗∗p < 0.01 vs MCre (Two-way ANOVA and Student's t test (for AUC and AAC)).

Figure 3

Assessment of ex vivo glucose induced insulin secretion and metabolic parameters in BKO versus control mouse islets after 12 weeks of G3PP deletion. (A) Insulin secretion at 3, 8 and 16 mM glucose. (B) Insulin secretion as in A, but in the presence of palmitate/oleate (0.125 mM each). (C) Insulin secretion at 3 mM glucose plus 35 mM KCl. (D) Total insulin content. For A, B, C and D: MCre, n = 7; BKO, n = 8. (E) Glycerol release (MCre, n = 5; BKO, n = 5). (F) Oxygen consumption. (G) ATP production. (H) H+ leak (MCre, n = 5; BKO, n = 6). ∗P < 0.05, ∗∗P < 0.01, and ∗∗∗P < 0.001 vs. MCre (Two-way ANOVA (Panel A and B) and Student's t test (Panel C–H)).

Figure 4

Targeted metabolomics analyses in BKO and control islets, 12 weeks after G3PP deletion. (A) Insulin secretion at 3, 8 and 16 mM glucose after 1 h incubation for the metabolomics experiments. At the end of incubation, metabolites were extracted and analyzed by LC-MS/MS. (B) Gro3P. (C) Dihydroxyacetone-phosphate (DHAP). (D) 2-Phosphoglycolate. (E) Lactate. (F) ATP. (G) ADP. (H) ATP/ADP. (I) Acetyl-CoA. (J) Malonyl-CoA. (K) Isocitrate/citrate. (L) Succinate. (M) Malate. (N) NADH/NAD. (O) NADPH/NADP. (P) GSH/GSSH. (Q) cAMP. (R) Glutamate. (S) Aspartate. (T) Alanine. Islets from 3 mice were pooled for one measurement and there were 5 such measurements for MCre and BKO, separately. A total of 15 mice were used in each group. Means ± SEM (n = 5 for each group). ∗P < 0.05 and ∗∗P < 0.01 vs. MCre (Two-way ANOVA).

Figure 5

Assessment of chronic glucotoxicity in BKO and control islets. Islets incubated at 11 and 30 mM glucose for 7 days (Islets from 2 mice were pooled for one measurement). After the 7-days incubation, islets were collected to measure: (A) insulin content (MCre, n = 10 from 20 mice; BKO, n = 10 from 20 mice), (B) apoptosis (MCre, n = 6 from 12 mice; BKO, n = 6 from 12 mice) and (C–G) gene expression (mRNA) by rt-PCR (MCre, n = 8 from 16 mice; BKO, n = 8 from 16 mice). (C) Ins-2, (D), Pdx-1 (E) Mafa, (F) Txnip and (G) Bip. Data are presented as % of 11 mM glucose from the data shown in Fig 5S. Means ± SEM. ∗P < 0.05, ∗∗P < 0.01, and ∗∗∗P < 0.001 vs. MCre (Student's t test).

Figure 6

Model depicting the effect of G3PP β-cell specific deletion on energy homeostasis and body weight via increasing insulin secretion under high glucose concentration. The abbreviations are: Ac-CoA, acetyl-CoA; AQP, aquaporin channel; ATP, adenosine triphosphate; DHAP, dihydroxyacetone phosphate; FA-CoA, fatty acyl-CoA; FFA, free fatty acid; G3PP, glycerol 3-phosphate phosphatase; G6P, glucose 6-phosphate; GL/FFA cycle, glycerolipid/free fatty acid cycle; Glyceraldehyde-3-P, glyceraldehyde-3-phosphate; Gro3P, glycerol 3-phosphate; Mal-CoA, malonyl-CoA; MCF, metabolic coupling factor; OCR, oxygen consumption rate; TCA cycle, tricarboxylic acid cycle; TG, triglyceride.

FigS1

FigS2

FigS3

FigS4

FigS5

FigS6