The Rho guanine dissociation inhibitor α inhibits skeletal muscle Rac1 activity and insulin action

Abstract

The molecular events governing skeletal muscle glucose uptake have pharmacological potential for managing insulin resistance in conditions such as obesity, diabetes, and cancer. With no current pharmacological treatments to target skeletal muscle insulin sensitivity, there is an unmet need to identify the molecular mechanisms that control insulin sensitivity in skeletal muscle. Here, the Rho guanine dissociation inhibitor α (RhoGDIα) is identified as a point of control in the regulation of insulin sensitivity. In skeletal muscle cells, RhoGDIα interacted with, and thereby inhibited, the Rho GTPase Rac1. In response to insulin, RhoGDIα was phosphorylated at S101 and Rac1 dissociated from RhoGDIα to facilitate skeletal muscle GLUT4 translocation. Accordingly, siRNA-mediated RhoGDIα depletion increased Rac1 activity and elevated GLUT4 translocation. Consistent with RhoGDIα's inhibitory effect, rAAV-mediated RhoGDIα overexpression in mouse muscle decreased insulin-stimulated glucose uptake and was detrimental to whole-body glucose tolerance. Aligning with RhoGDIα's negative role in insulin sensitivity, RhoGDIα protein content was elevated in skeletal muscle from insulin-resistant patients with type 2 diabetes. These data identify RhoGDIα as a clinically relevant controller of skeletal muscle insulin sensitivity and whole-body glucose homeostasis, mechanistically by modulating Rac1 activity.

Keywords: GLUT4 translocation; glucose uptake; insulin sensitivity; skeletal muscle; type 2 diabetes.

Fig. 1.

RhoGDIα inhibits Rac1 activity in GLUT4myc L6 myotubes. (A) Heat map showing the Rac1 interactome of proteins differentially bound to active GTP-bound Rac1 (Rac1-GTP/GFP) compared to inactive Rac1 (Rac1-GDP/GFP). (B) RhoGDIα protein content in GLUT4myc L6 myotubes transfected with RhoGDIα siRNA (si-RhoGDIα) or control siRNA (si-Ctrl). (C) Basal and insulin-stimulated Rac1 activity (Rac1-GTP loading) in GLUT4myc L6 myotubes transfected with si-RhoGDIα or si-Ctrl. Right Inset: Rac1 activity relative to total Rac1 protein content. The cells were incubated $\pm$10 nM or 100 nM insulin for 10 min. (D) Rac1 protein content and (E) phosphorylated (p)PAK1 T423 in GLUT4myc L6 myotubes transfected with si-RhoGDIα or si-Ctrl. (F) Representative blots showing (B–E), (SI Appendix, Fig. S2B) and control coomassie staining. (G) RhoGDIα protein content in GLUT4myc L6 myotubes with recombinant adeno-associated viral vector-mediated overexpression of wildtype RhoGDIα (rAAV6:RhoGDIα). Control cells were transfected with an empty vector (rAAV6:MCS). (H) Basal and insulin-stimulated Rac1 activity in GLUT4myc L6 myotubes transfected with rAAV6:RhoGDIα or as a control rAAV6:MCS. Right Inset: Rac1 activity relative to total Rac1 protein content. The cells were incubated $\pm$10 nM or 100 nM insulin for 10 min. (I) Rac1 protein content in GLUT4myc L6 myotubes transfected with rAAV6:RhoGDIα or as a control rAAV6:MCS. (J) Representative blots showing (G), (I), and control ponceau staining. As indicated by the SuperPlots (35), the siRNA experiment was assayed in triplicates and repeated twice and the AAV experiment was assayed in duplicates and repeated three times. Total protein content was evaluated with a Student’s t test. Rac1 activity and protein phosphorylation were evaluated with a two-way ANOVA. Main effects are indicated in the panels. Significant interactions in two-way ANOVAs and significant Student’s t tests: effect of si-RhoGDIα/rAAV6:RhoGDIα ### (P < 0.001); effect of insulin (0 nM vs. 10 nM/100 nM insulin) */**/*** (P < 0.05/0.01/0.001); and effect of insulin dose (10 nM vs. 100 nM insulin) §§§ (P < 0.001). Data are presented as mean $\pm$ SEM with individual data points shown and the average from each experimental round.

Fig. 2.

Insulin phosphorylates RhoGDIα S101 to dissociate from and activate Rac1 in skeletal muscle cells. (A) Violin plot showing immunopurification of RhoGDIα followed by immunoblotting for Rac1 (Left; RhoGDIα–Rac1 interaction) or pRhoGDIα S101 (Right) $\pm$100 nM insulin for 10 min in GLUT4myc L6 myoblasts. (B) Representative blots for (A). (C) RhoGDIα protein content in GLUT4myc L6 myotubes with recombinant adeno-associated viral vector-mediated overexpression of a RhoGDIα S101A mutant (rAAV6:RhoGDIα S101A). Control cells were transfected with an empty vector (rAAV6:MCS). (D) Basal- and insulin-stimulated Rac1 activity in GLUT4myc L6 myotubes transfected with rAAV6:RhoGDIα S101A or control rAAV6:MCS. Right Inset: Rac1 activity relative to total Rac1 protein content. The cells were incubated $\pm$10 nM or 100 nM insulin for 10 min. (E) Rac1 protein content in GLUT4myc L6 myotubes transfected with rAAV6:RhoGDIα S101A or control rAAV6:MCS. (F) Representative blots showing (C), (E), and control ponceau staining. (G) RhoGDIα protein content in GLUT4myc L6 myotubes transfected with rAAV6:RhoGDIα or rAAV6:RhoGDIα S101A after siRNA-mediated knockdown of endogenous RhoGDIα. Control cells were transfected with rAAV6:MCS and control siRNA. (H) Basal- and insulin-stimulated Rac1 activity in GLUT4myc L6 myotubes transfected with rAAV6:RhoGDIα, rAAV6:RhoGDIα S101A, or control rAAV6:MCS after siRNA-mediated knockdown of endogenous RhoGDIα or si-Ctrl. Right Inset: Rac1 activity relative to total Rac1 protein content. The cells were incubated $\pm$100 nM insulin for 10 min. (I) Rac1 protein content in GLUT4myc L6 myotubes transfected with rAAV6:RhoGDIα, rAAV6:RhoGDIα S101A, or control rAAV6:MCS. As indicated by the SuperPlots (35), the experiments were assayed in duplicates and repeated three times. Total protein content was evaluated with a Student’s t test or one-way ANOVA. Rac1 activity was evaluated with a two-way ANOVA. For (H), the effect of rAAV6:RhoGDIα or rAAV6:RhoGDIα S101A was tested separately. Main effects are indicated in the panels. Significant interactions in two-way ANOVAs and significant Student’s t tests and one-way ANOVAs: Effect of rAAV6:RhoGDIα/rAAV6:RhoGDIα S101A ### (P < 0.001). Data are presented as mean $\pm$ SEM with individual data points shown and the average from each experimental round.

Fig. 3.

RhoGDIα is a negative regulator of GLUT4 translocation to the plasma membrane. (A) Basal and insulin-stimulated GLUT4 translocation (cell surface GLUT4myc) in GLUT4myc L6 myotubes transfected with RhoGDIα siRNA (si-RhoGDIα) or control siRNA (si-Ctrl). The cells were incubated $\pm$10 nM or 100 nM insulin for 15 min. (B) Phosphorylated (p)Akt S474 and (C) Akt2 protein content in GLUT4myc L6 myotubes transfected with si-RhoGDIα or si-Ctrl. (D) Representative blots showing (B), (C), and control coomassie staining. (E) Basal and insulin-stimulated GLUT4 translocation in GLUT4myc L6 myotubes with recombinant adeno-associated viral vector-mediated overexpression of wildtype RhoGDIα (rAAV6:RhoGDIα) or control empty vector (rAAV6:MCS). The cells were incubated $\pm$10 nM or 100 nM insulin for 15 min. (F) pAkt S474 in GLUT4myc L6 myotubes transfected with si-RhoGDIα or si-Ctrl. (G) Representative blots showing (F) and control ponceau staining. (H) Basal and insulin-stimulated GLUT4 translocation in GLUT4myc L6 myotubes transfected with rAAV vector-mediated overexpression of a RhoGDIα S101A mutant (rAAV6:RhoGDIα S101A). Control cells were transfected with rAAV6:MCS. The cells were incubated $\pm$10 nM or 100 nM insulin for 15 min. (I) Basal and submaximal insulin-stimulated (10 nM) GLUT4 translocation in GLUT4myc L6 myotubes transfected with si-RhoGDIα or si-Ctrl ± 25 µM C2-ceramide during the last 2 h of serum deprivation and the acute insulin challenge. Data were evaluated with two two-way ANOVAs to test the factors “siRNA” (si-Ctrl vs. si-RhoGDIα) and “insulin concentration” (0 nM vs. 10 nM) in nontreated and C2-ceramide–treated cells, respectively. The effect of C2-ceramide treatment was assessed by two two-way ANOVAs to test the factors “siRNA” and “C2-ceramide” (nontreated cells vs. C2-ceramide–treated cells) at 0 nM and 10 nM insulin, respectively. As indicated by the SuperPlots (35), the experiments were assayed in duplicate (F) or triplicate (A–C, E, H, I) and repeated twice (B and C), three (E, F, and H), or four times (A and I). Total protein content was evaluated with a Student’s t test. Unless otherwise stated previously in the figure legend, GLUT4 translocation and protein phosphorylation were evaluated with a two-way ANOVA. Main effects are indicated in the panels. Significant interactions in two-way ANOVAs and significant Student’s t tests: Effect of si-RhoGDIα/rAAV6:RhoGDIα/rAAV6:RhoGDIα S101A #/##/### (P < 0.05/0.01/0.001). Effect of insulin (0 nM vs. 10 nM/100 nM insulin) */*** (P < 0.05/0.001). Effect of insulin dose (10 nM vs. 100 nM insulin) § (P < 0.05). Data are presented as mean $\pm$ SEM with individual data points shown and the average from each experimental round. A.U., arbitrary units.

Fig. 4.

RhoGDIα suppresses insulin-stimulated glucose uptake in mouse skeletal muscle in vivo. (A) Schematic of the experimental design. A recombinant adeno-associated viral vector encoding RhoGDIα (rAAV6:RhoGDIα) was administered intramuscularly in gastrocnemius (also targeting the soleus), tibialis anterior (also targeting the EDL), and triceps brachii muscles, while the contralateral muscles were injected with an empty viral vector (rAAV6:MCS). (B) Representative blot of Flag-tagged and endogenous RhoGDIα in gastrocnemius or nontransfected control quadriceps muscle 2 or 4 wk after administration. (C) Representative image of Flag (RhoGDIα) staining of cross-sections from rAAV6:RhoGDIα- or rAAV6:MCS-treated tibialis anterior muscle. (D) Insulin-stimulated (0.5 U kg⁻¹ body weight) glucose uptake index in rAAV6:RhoGDIα- or rAAV6:MCS-treated gastrocnemius (Gast), EDL, soleus (SOL) and triceps brachii (Tri) muscles from chow-fed young adult mice 4 wk after rAAV6-administration. Saline, n =8/7/8/8 (Gast/EDL/SOL/Tri); Insulin, n =7/7/7/7. Data were evaluated with a two-way RM ANOVA for each of the muscles. (E) Insulin-stimulated (0.4 U kg⁻¹ body weight) glucose uptake index in Gast, EDL, SOL, and Tri muscles from 8 wk 60E% HFD-fed young adult mice 4 wk after rAAV6-administration. Saline, n = 8/7/6/8 (Gast/EDL/SOL/Tri); Insulin, n = 8/7/7/8. Data were evaluated with a two-way RM ANOVA for each of the muscles. The effect of diet was evaluated with a Student's t test comparing rAAV6:MCS-treated muscle in chow- and 60E% HFD-fed mice. (F) Rac1 protein content in rAAV6:RhoGDIα- or rAAV6:MCS-treated gastrocnemius muscle. Data were evaluated with a paired t test. (G) Representative blots showing RhoGDIα, (G), and control coomassie staining. (H) Rac1 mRNA expression in rAAV6:RhoGDIα- or rAAV6:MCS-treated TA muscle 2 or 4 wk after rAAV6-administration in chow-fed, male mice, n = 4. Data were evaluated with a paired t test. Main effects are indicated in the panels. Significant interactions in two-way RM ANOVAs and significant t tests: Effect of rAAV6:RhoGDIα vs. rAAV6:MCS ##/### (P < 0.01/0.001); Effect of insulin */*** (P < 0.05/0.001); Effect of HFD 60E% (†)/†† (P < 0.1/0.001). Data are presented as mean $\pm$ SEM or when applicable mean $\pm$ SEM with individual data points shown. A.U., arbitrary units.

Fig. 5.

Canonical insulin signaling proteins and glucose-handling proteins are not regulated by RhoGDIα overexpression. (A) Phosphorylated (p)PAK1 T423, (B) total PAK1, (C) pAkt S474, (D) pAkt T309, (E) total Akt2, (F) pTBC1D4 T649, and (G) total TBC1D4 protein content in gastrocnemius from chow-fed mice with recombinant adeno-associated viral vector encoding RhoGDIα (rAAV6:RhoGDIα) administered intramuscularly in muscles of the right leg, while the contralateral muscles were injected with an empty viral vector (rAAV6:MCS). Saline, n = 8; Insulin, n = 7. Total protein content was evaluated with a paired t test. Protein phosphorylation was evaluated with a two-way repeated measures (RM) ANOVA. (H) Representative blots showing (A–G) and control coomassie staining. (I) Representative TEM images showing skeletal muscle mitochondria in TA from rAAV6:RhoGDIα- or rAAV6:MCS-treated muscle of young adult mice. n = 6. Main effects are indicated in the panels. Data are presented as mean $\pm$ SEM with individual data points shown. A.U., arbitrary units.

Fig. 6.

Chronic RhoGDIα knockdown negatively affects insulin-stimulated glucose uptake and intracellular insulin signaling in skeletal muscle. (A) Schematic of the experimental design. A recombinant adeno-associated viral vector encoding RhoGDIα shRNA (rAAV6:RhoGDIα shRNA) was administered intramuscularly in gastrocnemius (also targeting the soleus) and tibialis anterior (also targeting the EDL), while the contralateral muscles were injected with a control vector (rAAV6:LacZ shRNA). (B) RhoGDIα, (C) Rac1, and (D) PAK1 protein content in rAAV6:RhoGDIα shRNA- or rAAV6:LacZ shRNA-treated gastrocnemius muscle. Saline, n = 8; Insulin, n = 8. (E) Representative blots showing (B–D) and (G–O) and control ponceau staining. (F) Insulin-stimulated (0.5 U kg⁻¹ body weight) glucose uptake index in rAAV6:LacZ shRNA- or rAAV6:RhoGDIα shRNA-treated gastrocnemius (Gast), extensor digitorum longus (EDL), and soleus (SOL) muscles. Saline, n = 8/7/6 (Gast/EDL/SOL); Insulin, n = 8/8/8. Data were evaluated with a two-way repeated measures (RM) ANOVA for each of the muscles. (G) Akt2, (H) TBC1D4, (I) GSK-3β, and (J) phosphorylated (p)Akt S474, (K) pAkt T309, (L) pTBC1D4 T649, (M) pGSK-3β S9, (N) GLUT4 and (O) HKII protein content in gastrocnemius muscle. Saline, n = 8; Insulin, n = 8. Total protein content was evaluated with a paired t test. Protein phosphorylation was evaluated with a two-way RM ANOVA. Main effects are indicated in the panels. Significant interactions in two-way RM ANOVAs and significant paired t tests: effect of rAAV6:RhoGDIα shRNA vs. rAAV6:LacZ shRNA #/##/### (P < 0.05/0.01/0.001) and effect of insulin ** (P < 0.01). Data are presented as mean $\pm$ SEM with individual data points shown. A.U., arbitrary units.

Fig. 7.

Elevated muscle RhoGDIα content is associated with whole-body glucose intolerance in mice and humans. (A) RhoGDIα protein content in lean normal glucose tolerant (NGT; n = 9), obese NGT (n = 11), and obese T2D (n = 10) subjects of mixed sex. (B) Representative blots showing RhoGDIα and control coomassie staining. The data were evaluated with a one-way ANOVA. (C) Experimental overview and representative blots of tissue protein content of RhoGDIα in chow- or 60E% HFD-fed mice 8 wk after recombinant adeno-associated viral vector-mediated overexpression of RhoGDIα (rAAV6:RhoGDIα) specifically in striated muscle after a single intravenous administration in young, adult mice. As a control, an empty vector was administered (rAAV6:MCS). (D) Body composition (FM: fat mass; LBM: lean body mass; BW: body weight) in gram. Data were evaluated with a two-way ANOVA separately for FM, LBM, and BW. (E) Blood glucose levels during a glucose tolerance test (GTT). Chow, n = 7/8 (rAAV6:MCS/rAAV6:RhoGDIα); HFD, n = 7/8. Data were evaluated with two two-way RM ANOVAs to test the factors “rAAV6” and “time point” (0’ vs. 30’ vs. 60’ vs. 90’ vs. 120’) in chow- and HFD-fed mice, respectively. The effect of HFD was assessed with five two-way ANOVAs to test the factors rAAV6 and “diet” at each time point, respectively. (F) Incremental Area Under the Curve for blood glucose levels during the first 60 min of the GTT in panel (E). Data were evaluated with a two-way ANOVA. (G) Plasma insulin levels during a GTT. Chow, n = 8/8 (rAAV6:MCS/rAAV6:RhoGDIα); HFD, n = 7/8. Data were evaluated with two two-way RM ANOVAs to test the factors rAAV6 and time point (0’ vs. 20’) in chow- and HFD-fed mice, respectively. The effect of HFD was assessed with two two-way ANOVAs to test the factors rAAV6 and diet at both time points, respectively. (H) Blood glucose levels during an insulin tolerance test. Chow, n = 7/8 (rAAV6:MCS/rAAV6:RhoGDIα); HFD, n = 7/8. Data were evaluated with two two-way RM ANOVAs to test the factors rAAV6 and time point (0’ vs. 30’ vs. 60’ vs. 90’ vs. 120’) in chow- and HFD-fed mice, respectively. The effect of HFD was assessed with five two-way ANOVAs to test the factors rAAV6 and diet at each time point, respectively. Main effects are indicated in the panels. Significant interactions in two-way ANOVAs and significant one-way ANOVAs: Effect of rAAV6:RhoGDIα vs. rAAV6:MCS (#) (P < 0.1); Chow vs. HFD †/††† (P < 0.05/0.001); rAAV6:RhoGDIα vs. rAAV6:MCS in HFD-fed mice ¤¤ (P < 0.01); Chow vs. HFD in rAAV6:RhoGDIα ‡‡‡ (P < 0.001); Obese NGT vs. Obese T2D § (P < 0.05). Data are presented as mean $\pm$ SEM or when applicable mean $\pm$ SEM with individual data points shown.