Brain GLUT4 Knockout Mice Have Impaired Glucose Tolerance, Decreased Insulin Sensitivity, and Impaired Hypoglycemic Counterregulation
Article Figures & Tables
Figures
- Figure 1
Brain-specific deletion of GLUT4. A: Whole-brain tissue, gastrocnemius muscle, heart muscle, and white adipose tissue (WAT) were harvested and homogenized for Western blotting of GLUT4. GLUT4 was markedly reduced in the brains of BG4KO mice (KO) compared with WT, Lox, and Cre mice. GLUT4 protein concentrations in muscle, heart, and WAT were similar between BG4KO and control mice. It should be noted that 50 times more protein was loaded for the brain compared with muscle, heart, and WAT in order to detect GLUT4 protein levels because of their relatively lower expression compared with that in muscle, heart, and WAT. B: Quantification of brain GLUT4 protein content. BG4KO mice (n = 11; black bar) had a >99% reduction in brain GLUT4 levels compared with WT (n = 6; white bar), Lox (n = 6; horizontal lines), and Cre (n = 6; slanted lines) mice. *P = 0.0009 (ANOVA) vs. WT, Lox, and Cre mice. Data are expressed as mean ± SEM.
- Figure 2
Normal blood glucose but impaired glucose tolerance in BG4KO mice. A: BG4KO mice (black bars) have fed and fasting blood glucose concentrations and fed and fasting plasma insulin concentrations similar to those of WT (white bar), Lox (horizontal lines), and Cre (slanted lines) mice. B: BG4KO mice had fed and fasting blood insulin concentrations similar to those of the control groups. C: Intraperitoneal GTTs (2 mg/kg) were performed in 12-week-old male mice fed normal chow. BG4KO mice (black triangles) had significantly higher excursions in blood glucose compared with WT (white circles), Lox (black circles), and Cre (white triangles) mice, where indicated. D: No difference in blood glucose was observed between BG4KO and the control mice during an insulin tolerance test. *P = 0.006 (repeated-measures ANOVA) vs. WT, Lox, and Cre mice. Data are expressed as the mean ± SEM (n = 6–17 mice per group).
- Figure 3
Reduced insulin sensitivity and hepatic insulin resistance in BG4KO mice. A: Blood glucose during a hyperinsulinemic-euglycemic clamp. No difference in blood glucose before or during the last hour of the clamp was observed between BG4KO (black triangles) and WT (white circles), Lox (black circles), and Cre (white triangles) mice. B: Despite matched blood glucose concentrations, BG4KO mice required a significantly lower glucose infusion rate (Ginf) to maintain euglycemia during the final hour of the clamp compared with controls, indicating insulin resistance (*P = 0.002, repeated-measures ANOVA, BG4KO vs. WT, Lox, and Cre). C: Whole-body Rd values at baseline and during the last 30 min of the hyperinsulinemic clamp were not significantly different between BG4KO (black bar) and WT (white bar), Lox (horizontal lines), and Cre (diagonal lines) mice. D: 2-[14C]DG–determined glucose uptake specifically in muscle was similar between BG4KO (KO; n = 7) and WT (n = 6), Lox (n = 6), and Cre (n = 5) mice during the hyperinsulinemic clamp. E: EGP was not different at baseline among the groups, but EGP during the clamp was significantly higher in BG4KO mice compared with the three groups of control mice. F: The percent suppression of EGP from baseline to during the hyperinsulinemic clamp was impaired in BG4KO mice compared with WT, Lox, and Cre mice (BG4KO 41 ± 7 vs. WT 61 ± 36, Lox 77 ± 10, and Cre 70 ± 17% suppression; *P = 0.05, ANOVA) during the last hour of the hyperinsulinemic clamp. All data are expressed as mean ± SEM (n = 6–11 mice per group).
- Figure 4
Reduced glucose uptake in the brains of BG4KO mice. Glucose uptake in the brain was measured using 2-[14C]DG autoradiography in mice that underwent a hyperinsulinemic-euglycemic clamp. A: The areas of interest were precisely identified by Nissl staining (left panels), identifying the desired areas (i.e., the VMH; red dashed oval outline) and arcuate nucleus (ARC; red dashed triangular outline) and measuring the density of the outlined areas of interest on the exposed autoradiographic films (right panels). B: Quantification of glucose uptake in the brain during euglycemia using the Sokoloff equation in BG4KO mice (n = 5; black bars) and Nestin-Cre control mice (n = 5; white bars). Glucose uptake in the brain was significantly reduced in BG4KO mice compared with control mice in the VMH and ARC (*P < 0.05, t test). Data are expressed as the mean ± SEM. NTS, nucleus tractus solitarus.
- Figure 5
Intracerebroventricular (ICV) infusion of IDV, an inhibitor of GLUT4 transport, reduced the counterregulatory response to hypoglycemia. IDV (n = 9; black circles) or aCSF (n = 6; white circles) was infused for 90 min before and for the duration of a 90-min hyperinsulinemic-hypoglycemic (∼45 mg/dL) clamp in Sprague-Dawley rats. A: Blood glucose was precisely matched between rats treated with ICV infusions of either IDV or aCSF. B: Despite matched blood glucose concentrations, IDV-treated rats (black circles) required a significantly higher glucose infusion rate (Ginf) than control rats (white circles) (*P < 0.0001, repeated-measures ANOVA). The higher glucose infusion rate was attributed to the attenuated counterregulatory response to hypoglycemia caused by the reduced GLUT4-mediated glucose sensing. Epinephrine (C), norepinephrine (D), and glucagon (E) responses to hypoglycemia were significantly reduced in IDV-treated rats vs. controls (*P < 0.05, t test, for all hypoglycemic time points for epinephrine, norepinephrine, and glucagon and area under the curve). Data are expressed as mean ± SEM.
- Figure 6
BG4KO mice have an impaired counterregulatory response to hypoglycemia. BG4KO and control mice were subjected to hyperinsulinemic-hypoglycemic (∼30 mg/dL) clamps. A: Blood glucose was not different before or during the clamp between BG4KO (n = 5; black triangles), WT (n = 5; white circles), Lox (n = 6; black circles), and Cre (n = 11; white triangles) mice. B: Glucose infusion rates (Ginf) during the hypoglycemic clamp were not different among the groups. C: Insulin concentrations in the basal state and during the hypoglycemic clamp were not different between the groups: WT (white bars), Lox (horizontal lines), Cre (slanted lines), and BG4KO (black bars). Epinephrine (D) and glucagon (E) responses to hypoglycemia were significantly reduced in BG4KO (black bars) mice compared with WT (white bars), Lox (horizontal lines), and Cre (slanted lines) mice. Norepinephrine (F) and corticosterone (G) concentrations during the hypoglycemic clamp were similar in all groups. H: Representative c-fos immunostaining of the hypothalamic PVN in WT and BG4KO (KO) mice after 2 h of hypoglycemia. I: Quantification of c-fos immunostaining. The number of c-fos–positive cells in the PVN was greatly reduced in BG4KO mice (black bar) compared with WT (white bar), Lox (horizontal lines), and Cre (slanted lines) mice. *P < 0.03 (ANOVA) vs. WT, Lox, and Cre. Data are expressed as mean ± SEM (n = 5–11 mice per group).
- Figure 7
Impaired glucose sensing in BG4KO mice. Whole-cell current-clamp traces of GI neurons in slices of the VMH from Nestin-Cre control (A) and BG4KO (B) mice. The upward deflections represent action potentials, the resting membrane potential is given to the left of each trace, and the downward deflections are the voltage response to a constant-current pulse. Compared with GI neurons from controls (CON; white bars), GI neurons from BG4KO mice (KO; black bars) had markedly impaired changes in membrane potential (C) and input resistance (D) when glucose concentrations were changed from 2.5 to 0.5 mmol/L. E: Current–voltage relationships for the effects of glucose or insulin on a GI neuron from the VMH of control mice. The reversal potential for the effect of glucose (intersection of 2.5 vs. 0.1 mmol/L glucose; upward dotted arrow) is close to the theoretical Cl− equilibrium potential (ECl ∼ −60 mV), suggesting that glucose activates a chloride channel. In contrast, the reversal potential for the effect of insulin (intersection of 0.1 mmol/L glucose vs. 0.1 mmol/L glucose + 5 nmol/L insulin; downward dashed arrow) is near the theoretical K+ equilibrium potential (EK ∼ −99 mV), suggesting that insulin activates a potassium channel. *P < 0.002 (t test) vs. control. Data are expressed as mean ± SEM (n = 14–16 mice per group).
In this Issue
