G6PC2 Modulates Fasting Blood Glucose In Male Mice in Response to Stress

Abstract

The glucose-6-phosphatase catalytic 2 (G6PC2) gene is expressed specifically in pancreatic islet beta cells. Genome-wide association studies have shown that single nucleotide polymorphisms in the G6PC2 gene are associated with variations in fasting blood glucose (FBG) but not fasting plasma insulin. Molecular analyses examining the functional effects of these single nucleotide polymorphisms demonstrate that elevated G6PC2 expression is associated with elevated FBG. Studies in mice complement these genome-wide association data and show that deletion of the G6pc2 gene lowers FBG without affecting fasting plasma insulin. This suggests that, together with glucokinase, G6PC2 forms a substrate cycle that determines the glucose sensitivity of insulin secretion. Because genome-wide association studies and mouse studies demonstrate that elevated G6PC2 expression raises FBG and because chronically elevated FBG is detrimental to human health, increasing the risk of type 2 diabetes, it is unclear why G6PC2 evolved. We show here that the synthetic glucocorticoid dexamethasone strongly induces human G6PC2 promoter activity and endogenous G6PC2 expression in isolated human islets. Acute treatment with dexamethasone selectively induces endogenous G6pc2 expression in 129SvEv but not C57BL/6J mouse pancreas and isolated islets. The difference is due to a single nucleotide polymorphism in the C57BL/6J G6pc2 promoter that abolishes glucocorticoid receptor binding. In 6-hour fasted, nonstressed 129SvEv mice, deletion of G6pc2 lowers FBG. In response to the stress of repeated physical restraint, which is associated with elevated plasma glucocorticoid levels, G6pc2 gene expression is induced and the difference in FBG between wild-type and knockout mice is enhanced. These data suggest that G6PC2 may have evolved to modulate FBG in response to stress.

The Glucose-6-phosphatase catalytic 2 (G6PC2) catalyzes the hydrolysis of glucose-6-phosphate to glucose and inorganic phosphate thereby opposing the action of the β-cell glucose sensor, glucokinase, which catalyzes the formation of G6P from glucose (1, 2). In isolated G6pc2 knockout (KO) mouse islets, glucose-6-phosphatase activity (3) and glucose cycling (4) are both abolished and perfused pancreas analyses demonstrate that deletion of G6pc2 causes a leftward shift in the dose-response curve for glucose-stimulated insulin secretion (GSIS) (3). Under fasting conditions, where insulin levels are the same in wild-type (WT) and G6pc2 KO mice, this shift results in a reduction in fasting blood glucose (FBG) in KO mice (3, 5). These observations are consistent with data from genome-wide association studies showing that common single nucleotide polymorphisms in the G6PC2 locus, that have modest effects on G6PC2 expression (6, 7), are nonetheless associated with variations in FBG but not fasting plasma insulin (FPI) (8).

Because chronically elevated FBG has been associated with both increased risk for cardiovascular-associated mortality (9) and the development of type 2 diabetes (T2D) (10), this implies that G6PC2 is actually detrimental to health in the modern Western environment, which is associated with nutrient excess and longevity. In contrast, during evolution, the ability of G6PC2 to regulate FBG must have conferred a specific advantage. We hypothesized that factors that induce G6PC2 expression would confer a transient beneficial change in FBG. We therefore screened for factors that regulate G6PC2 promoter activity. In this study, we show that glucocorticoids stimulate G6PC2 expression and we present in vivo data suggesting that G6PC2 may have evolved to modulate FBG in response to stress.

Research Design and Materials

Animal care

The Vanderbilt University Medical Center Animal Care and Use Committee approved all protocols used. Mice were maintained on a standard rodent chow diet. Food and water were provided ad libitum. Adult male mice (∼4–7 mo) were used in these studies.

Generation of G6pc2 KO mice

The generation of G6pc2 KO mice on a mixed 129SvEv^BRD x C57BL/6J genetic background has been described (5). Mice were backcrossed onto pure C57BL/6J or 129SvEv^BRD genetic backgrounds using a speed congenic breeding strategy (3).

Mouse and human islet isolation

Mouse islets were isolated as previously described (3). Human islets were obtained by A.C.P. through the National Institute of Diabetes and Digestive and Kidney Diseases-funded Integrated Islet Distribution Program (https://iidp.coh.org/).

Cell culture

βTC-3 and 832/13 cells were cultured in DMEM and RPMI 1640 media, respectively, supplemented with 10% fetal bovine serum. Mouse primary islet cells were dispersed into single cells by repeated pipetting in a Versene/Trypsin-EDTA solution (Life Technologies).

Fusion gene analyses

The construction of a wild-type human G6PC2-luciferase fusion gene has been previously described (11). The mouse G6pc2 promoter was isolated using PCR and 129SvEv or C57BL/6J genomic DNA as the template. Cells were transfected using lipofectamine as described (11) and incubated for 18–20 hours in the indicated concentration of dexamethasone (Dex) or corticosterone before harvesting.

Analysis of islet and pancreas gene expression

Pancreatic RNA was isolated using the ToTALLY RNA kit, whereas islet RNA was isolated using the RNAqueous kit (Ambion). Gene expression was quantitated using real-time PCR. Pancreatic gene expression was quantitated after RNA isolation by using the Turbo DNA-free DNAse Treatment kit (Ambion) to remove trace gDNA followed by cDNA generation using the iScript DNA Synthesis kit (Bio-Rad) and then PCR using the 2′-deoxyuridine 5′-triphosphate-containing FastStart SYBR Green Master Mix in conjunction with Uracil-DNA-glycosylase (Roche).

Islet gene expression was quantitated using the primer-probe TaqMan approach from Life Technologies as described (12). Fold induction of gene expression was calculated using the 2(−ΔΔC(T)) method (13).

In vivo mouse studies

Intraperitoneal glucose tolerance tests (IPGTTs) were performed as previously described (3). Blood was isolated by retroorbital bleeding for the measurement of FPI and corticosterone, which were assayed using RIA by the Vanderbilt Diabetes Center Hormone Assay Core. The repeated physical restraint experimental paradigm has been previously described (14). Briefly, mice were immobilized in a 50-mL conical tube at 9 am each day for 1 hour.

Statistical analyses

Gene expression data were analyzed using a Student's t test: 2 sample assuming equal variance. Mouse data comparing 2 groups were analyzed using a Student's t test: 2 sample assuming unequal variance, whereas mouse data comparing multiple groups were analyzed by one-way ANOVA. A post hoc analysis was performed using the Bonferroni correction for multiple comparisons. The level of significance was as indicated.

Results

Dex stimulates human G6PC2 expression

Although chronically elevated G6PC2 expression increases FBG and is detrimental over the long term to human health, we hypothesized that the ability of G6PC2 to transiently regulate FBG must nonetheless confer an evolutionarily conserved benefit. We therefore screened factors known to regulate FBG to determine whether they could also modulate G6PC2 promoter activity. Figure 1A shows that the synthetic glucocorticoid Dex stimulates a marked, concentration-dependent activation of the human G6PC2 promoter, as assessed by fusion gene transient transfection assays in the βTC-3 islet-derived cell line. The effect of Dex was enhanced by cotransfection with a plasmid encoding the glucocorticoid receptor (GR) (Figure 1B), consistent with previous studies that observed a limiting intracellular concentration of this receptor (15). A similar marked stimulation of human G6PC2-luciferase fusion gene expression by Dex was observed in primary mouse islet cells (Figure 1C). Furthermore, Dex stimulated endogenous G6PC2 and SLC37A4 gene expression in primary human islets (Figure 1D). Human G6PC2-luciferase fusion gene expression was also induced by the endogenous glucocorticoid corticosterone (Figure 1E).

Figure 1.

Dexamethasone stimulates human G6PC2 expression. A, Induction of G6PC2-luciferase fusion gene expression in βTC-3 cells by Dex. B, Cotransfection of an expression vector encoding the GR (GREV) enhances the induction of G6PC2-luciferase fusion gene expression in βTC-3 cells by Dex (250nM). C, Induction of G6PC2-luciferase fusion gene expression in dispersed primary mouse islet cells by Dex (250nM). These primary cells represent a mixed population of islet cell types. D, Induction of endogenous G6PC2 and SLC37A4 gene expression in human islets by Dex. Human islets were treated for 3 hours with 250nM Dex. E, Induction of G6PC2-luciferase fusion gene expression in 832/13 cells by corticosterone (250nM). F, Localization of the G6PC2 GRE through the analysis of the effect of Dex (250nM) on truncated G6PC2-luciferase fusion gene expression in βTC-3 cells. G, Cross-species sequence alignment of the G6PC2 GRE and comparison with the consensus (16). H, Point mutation of the G6PC2 GRE reduces basal G6PC2-luciferase fusion gene expression in βTC-3 cells and abolishes the effect of Dex (250nM). Results show the mean ± SEM of 3–4 experiments; *, P < .05 vs control.

To localize the G6PC2 glucocorticoid-response element (GRE) a series of 5′ truncated G6PC2-luciferase fusion genes were analyzed by transient transfection. Figure 1F shows that Dex-stimulated fusion gene expression was abolished when the region of the promoter between −171 and −131 was deleted. Visual inspection of this region identified the presence of a putative GRE (16) that is conserved across multiple species (Figure 1G). A point mutation of a nucleotide in this putative GRE known to be required for GR binding (Figure 1G) (16), in the context of an otherwise intact promoter, had a limited effect on basal fusion gene expression but abolished the Dex response (Figure 1H).

Dex stimulates 129SvEv but not C57BL/6J mouse G6pc2 promoter activity

We next sought to confirm that Dex also regulates mouse G6pc2 gene expression in vivo. Surprisingly, 3 hours after a single Dex injection we observed that endogenous pancreatic G6pc2 expression was induced in 129SvEv (Figure 2A) but not C57BL/6J (Figure 2B) mice, whereas Slc37a4 gene expression was induced in both (Figure 2, A and B). Consistent with these in vivo gene expression data, Dex only stimulated endogenous G6pc2 expression in isolated 129SvEv islets (Figure 2C) but not C57BL/6J islets (Figure 2D), whereas Slc37a4 gene expression was induced in both 129SvEv (Figure 2E) and C57BL/6J (Figure 2D) islets.

Figure 2.

Dex stimulates 129SvEv but not C57BL/6J mouse G6pc2 promoter activity. A and B, Selective induction of G6pc2 gene expression in 129SvEv (A) but not C57BL/6J (B) mice by short term Dex treatment. Pancreatic RNA was isolated from mice (129SvEv, n = 6; C57BL/6J, n = 6) 3 hours after injection with PBS or Dex phosphate (13 μg/g). Results show the mean ± SEM; *, P < .05 vs control. C–E, Induction of G6pc2 and Slc37a4 gene expression in 129SvEv (C and E) and C57BL/6J (D) mouse islets by Dex. RNA was isolated from islets incubated for 3 hours in the presence or absence of variable concentrations (C and E) or 250nM (D) Dex. Results show the mean ± SEM of 3–9 experiments; *, P < .05 vs control. F and G, Dex-stimulates 129SvEv (F) but not C57BL/6J (G) G6pc2-luciferase fusion gene expression in 832/13 cells. Results show the mean ± SEM of 3 experiments; *, P < .05 vs control. H, The stimulation of C57BL/6J G6pc2-luciferase fusion gene expression in 832/13 cells by Dex is modulated by the alternate alleles of rs32980497. Results show the mean ± SEM of 3 experiments; *, P < .05 vs control. Different vectors (F and G, pGL4) and (H, pGL3) were used in these experiments, explaining the differences in basal expression.

An explanation for this unexpected result became apparent from sequence analyses, which revealed that the alternate alleles of a mouse single nucleotide polymorphism, rs32980497, affect a key nucleotide in the G6pc2 GRE, that is required for GR binding (16). 129SvEv mice, but not C57BL/6J mice, harbor the allele that supports GR binding (Figure 1G). As a result, Dex markedly activates the 129SvEv G6pc2 promoter (Figure 2F) but not the C57BL/6J G6pc2 promoter (Figure 2G), as assessed by fusion gene transient transfection assays. Four other nucleotides differ between the proximal C57BL/6J and 129SvEv G6pc2 promoters (data not shown) but substitution of the alternate allele of rs32980497 in the C57BL/6J promoter is sufficient to restore responsiveness to Dex (Figure 2H). The reverse experiment, mutating the equivalent nucleotide in the human G6PC2 promoter, abolishes the Dex response (Figure 1H).

The induction of G6pc2 in response to physical restraint modulates FBG in 129SvEv mice

Previous studies have shown that glucose-6-phosphatase activity and glucose cycling are increased in islets isolated from Dex-injected mice (17, 18). These observations, which were published before the G6PC2 gene was identified (19), can now be explained by the induction of G6PC2 gene expression by glucocorticoids and are entirely consistent with our model of G6PC2 function. To extend these in vitro isolated islet studies, we next sought to study the physiological consequences of glucocorticoid-stimulated 129SvEv G6pc2 expression in vivo.

For these studies the G6pc2 KO allele was backcrossed onto the 129SvEv genetic background using a speed congenic approach (3). Before analyzing the effect of G6pc2 deletion on the response to glucocorticoids in 129SvEv mice, we first sought to compare the effect of G6pc2 deletion on FBG and glucose tolerance in control 129SvEv mice. As seen in mixed genetic background (5) and C57BL/6J mice (3), FBG was reduced in 129SvEv G6pc2 KO mice relative to WT mice, whereas fasting insulin was unchanged (Figure 3, A and B). As in C57BL/6J mice (3), IPGTTs showed no difference in glucose tolerance between WT and G6pc2 KO 129SvEv mice (Figure 3C). These data suggest that G6pc2 regulates FBG rather than glucose tolerance in control 129SvEv mice as previously observed in C57BL/6J mice (3).

Figure 3.

The induction of G6pc2 in response to physical restraint (PR) modulates FBG in 129SvEv mice. A and B, Blood glucose (A) and plasma insulin (B) levels in control 6-hour fasted 129SvEv male WT (n = 16) and G6pc2 KO (n = 14) mice. Results show the mean ± SEM; *, P < .05 vs WT. C, IPGTTs using 2.0-g/kg glucose performed on 6-hour fasted conscious male WT (n = 13) and G6pc2 KO (n = 11) mice. D, Induction of G6pc2 and Slc37a4 gene expression in 129SvEv mice by PR. Pancreatic RNA was isolated after a 6-hour fast from control (C) mice and mice that had been physically restrained (PR) (n = 10–11). G6pc2 and Slc37a4 expression were quantitated using real-time PCR; *, P < .05 vs control. E, Weight change in 129SvEv (WT, n = 21; KO, n = 22) mice after PR. Results show the mean ± SEM; *, P < .05 vs initial body weight. The difference in weight loss between WT and KO was significant (**, P < .05). F, Induction of plasma corticosterone by PR. Blood was isolated after a 3-hour fast from control (C) mice and mice that had been physically restrained (PR) (n = 7–10); *, P < .0001, one-way ANOVA. G, Induction of G6pc2 and Slc37a4 gene expression in C57BL/6J mice by PR. Pancreatic RNA was isolated after a 6-hour fast from control (C) mice and mice that had been physically restrained (PR) (n = 8). G6pc2 and Slc37a4 expression were quantitated using real-time PCR; *, P < .05 vs control. H and I, Diagrams proposing that the induction of G6pc2 expression by PR will increase the difference in FBG between fasted WT and KO mice. The diagrams indicate that the actual values of the x-axis (glucose) may be shifted to the right (H) or left (I), relative to those in control mice, depending on the effects of PR on other aspects of metabolism. In either case, FPI levels are predicted to not differ between WT and G6pc2 KO mice, as observed in control mice (3). J, Blood glucose levels in 6-hour fasted conscious control (C) (WT, n = 13; KO, n = 11) or 6-hour fasted conscious physically restrained (PR) (WT, n = 8; KO, n = 10) 129SvEv male mice. Results show the mean ± SEM; *, P < .0001, one-way ANOVA; NS, not significant. K, Plasma insulin levels in 6-hour fasted conscious control (C) (WT, n = 16; KO, n = 14) or 6-hour fasted conscious physically restrained (PR) (WT, n = 21; KO, n = 24) 129SvEv male mice. Results show the mean ± SEM. L, Glucose levels in 6-hour fasted conscious control (C) (WT, n = 12; KO, n = 15) or physically restrained (PR) (WT, n = 6; KO, n = 7) C57BL/6J male mice. Results show the mean ± SEM.

To study the physiological consequences of glucocorticoid-stimulated 129SvEv G6pc2 expression in vivo, we examined the effect of a 10-day, repeated physical restraint experimental paradigm (14) on FBG in 129SvEv G6pc2 WT and KO mice. Figure 3D shows that this treatment stimulated endogenous pancreatic G6pc2 and Slc37a4 gene expression in 129SvEv mice with little change in body weight (Figure 3E). This paradigm is associated with elevated endogenous glucocorticoid levels (Figure 3F) but also the activation of many other autonomic and neuroendocrine axes, beside the hypothalamic-pituitary-adrenal axis, resulting in the release of epinephrine and norepinephrine (20). The observed induction of G6pc2 expression could therefore be mediated by factors in addition to glucocorticoids. This may explain why G6pc2 expression is also induced to a lesser degree in C57BL/6J mice (Figure 3G), despite the GRE mutation (Figure 1G).

We have previously shown that, under fasting conditions, insulin levels are identical in WT and G6pc2 KO mice, but FBG is lower in G6pc2 KO mice due to a leftward shift in the dose-response curve for GSIS (3). We hypothesized that the induction of G6pc2 expression by physical restraint would enhance this existing difference in FBG between WT and KO mice (Figure 3, H and I). However, we reasoned that the physiological benefit conferred by G6pc2 induction would depend on the overall combined effects of physical restraint on whole-body glucose metabolism in vivo, because FBG is determined by multiple factors other than G6PC2. We predicted that if physical restraint results in an elevation in FBG in WT mice then the induction of G6pc2 expression would contribute to that elevation and it would be prevented or reduced in G6pc2 KO mice (Figure 3H). On the other hand we predicted that if physical restraint results in hypoglycemia in G6pc2 KO mice then the induction of G6pc2 expression in WT mice would limit or prevent hypoglycemia (Figure 3I). After a 6-hour fast, FBG was unchanged in physically restrained 129SvEv WT mice but decreased in physically restrained 129SvEv G6pc2 KO mice (Figure 3J), indicating that the difference in FBG between 129SvEv WT and KO mice was enhanced by physical restraint and that the induction of G6pc2 expression protects against low blood glucose, consistent with the model shown in Figure 3I. Fasting insulin levels were unchanged in 129SvEv mice after physical restraint (Figure 3K), again consistent with the model shown in Figure 3I. Physical restraint had no effect on FBG in C57BL/6J mice (Figure 3L) consistent with the small change in G6pc2 expression (Figure 3G). In these studies no difference in FBG was observed between control 129SvEv (Figure 3J) or C57BL/6J (Figure 3L) WT and KO mice, because the difference in FBG is small such that it is only consistently observed with much higher n values (Figure 3A) (3). In contrast, a clear difference in FBG was observed in physically restrained 129SvEv WT and KO mice (Figure 3J).

Discussion

We show here that the synthetic glucocorticoid Dex induces human G6PC2 promoter activity (Figure 1). Dex also induces 129SvEv but not C57BL/6J mouse G6pc2 promoter activity, the difference being due to a mutation in the C57BL/6J promoter GR binding site (Figure 2). In response to stress generated by physical restraint, G6pc2 expression is induced, enhancing the difference in FBG between 129SvEv WT and KO mice and protecting against low blood glucose (Figure 3). Thus, although in modern society the chronic influence of G6PC2 on FBG affects the risks of cardiovascular-associated mortality and T2D, our data suggest that G6PC2 may have initially evolved to transiently modulate FBG under conditions of glucocorticoid-related stress. Interestingly, G6PC2 expression is decreased in islets isolated from individuals with T2D (21), but it is unclear whether this contributes to islet dysfunction, given that the resulting enhanced glycolytic flux would promote the cytotoxic effects of glucose (22), or whether, because this would also lead to enhanced insulin secretion, this represents a compensatory change in unhealthy islets designed to maintain glucose homeostasis.

The widely accepted dogma with respect to the effect of glucocorticoids on glucose metabolism in both rodents and humans in vivo is that they inhibit glucose uptake by inducing insulin resistance, stimulate hepatic glucose production, and inhibit insulin secretion, thereby inducing glucose intolerance (23, 24). Although the resulting increase in blood glucose is considered beneficial during periods of transient stress (25), prolonged elevation of glucocorticoids, as occurs in Cushing's disease (26), can lead to diabetes. In contrast, our studies (Figures 3) and others (27) demonstrate the complexity of glucocorticoid physiology by suggesting that, in some experimental paradigms, glucocorticoid action leads to improved FBG and/or glucose tolerance. Thus, in 6-hour fasted mice, this induction of G6pc2 by glucocorticoids limits the decrease in FBG (Figure 3I) which increases circulating glucose levels that could provide energy under stressful circumstances relative to KO mice. Although in the physical restraint paradigm the induction of G6pc2 expression protects against low blood glucose (Figure 3I), we predict that in other stress-associated paradigms this induction may contribute to a transient increase in FBG (Figure 3H), which could be a survival mechanism that is valuable for keeping the brain supplied with glucose in times of stress (25).

In summary, our data suggest that G6PC2 may have initially evolved to transiently modulate the sensitivity of GSIS to glucose under conditions of glucocorticoid-related stress. G6PC2 is thought to be expressed exclusively in pancreatic islet β-cells (2, 8) but future studies on β-cell-specific KO mice will be required to prove that low G6pc2 expression in rare cell types is not regulating FBG and FPI through other mechanisms.