One exercise session can elevate insulin-stimulated glucose uptake (ISGU) in skeletal muscle, but the mechanisms remain elusive. Circumstantial evidence suggests a role for Akt substrate of 160 kDa (AS160 or TBC1D4). We used genetic approaches to rigorously test this idea. The initial experiment evaluated the role of AS160 in postexercise increase in ISGU using muscles from male wild-type (WT) and AS160-knockout (KO) rats. The next experiment used AS160-KO rats with an adeno-associated virus (AAV) approach to determine if rescuing muscle AS160 deficiency could restore the ability of exercise to improve ISGU. The third experiment tested if eliminating the muscle GLUT4 deficit in AS160-KO rats via AAV-delivered GLUT4 would enable postexercise enhancement of ISGU. The final experiment used AS160-KO rats and AAV delivery of AS160 mutated to prevent phosphorylation of Ser588, Thr642, and Ser704 to evaluate their role in postexercise ISGU. We discovered the following: 1) AS160 expression was essential for postexercise increase in ISGU; 2) rescuing muscle AS160 expression of AS160-KO rats restored postexercise enhancement of ISGU; 3) restoring GLUT4 expression in AS160-KO muscle did not rescue the postexercise increase in ISGU; and 4) although AS160 phosphorylation on three key sites was not required for postexercise elevation in ISGU, it was essential for the full exercise effect.
Introduction
Insulin resistance is a primary defect preceding type 2 diabetes (1), and skeletal muscle is responsible for the largest portion of insulin-stimulated glucose uptake (ISGU) (2). One exercise session can markedly elevate muscle ISGU without amplification of the normal level of insulin signaling at steps ranging from insulin receptor binding to Akt activation (3–7), suggesting that exercise may alter post-Akt signaling.
Akt substrate of 160 kDa (AS160; also called TBC1D4) is a Rab GTPase–activating protein that is phosphorylated by Akt (8,9). Phosphorylated AS160 (pAS160) controls insulin-stimulated GLUT4 translocation and glucose uptake (GU) (8–11). Arias et al. (12) discovered that exercise leading to greater ISGU was accompanied by elevated pAS160. Many studies provide support for the idea that AS160 may mediate greater postexercise (PEX) ISGU (PEX-ISGU) (13–23).
Exposure to AICAR (an AMPK-activating compound) or electrically stimulated contractions can induce subsequently enhanced muscle ISGU (24–27). Results using AS160-knockout (KO) mice demonstrated AS160 is essential for AICAR or contraction effects on ISGU (28). However, no research has directly tested if AS160 expression and/or phosphorylation is essential for enhanced PEX-ISGU.
Accordingly, we performed a series of experiments to address these putative mechanisms. We initially used an AS160-KO rat model (29) to test if AS160 is essential for elevated PEX-ISGU. We then used AS160-KO rats with an adeno-associated virus (AAV) approach to determine if rescuing muscle AS160 deficiency could restore the ability of exercise to improve ISGU. We next used AAV-delivered GLUT4 to eliminate the GLUT4 deficit in AS160-KO muscle. Finally, we used AS160-KO rats with AAV delivery of AS160 that was mutated to prevent phosphorylation on key sites (Ser588, Thr642, and Ser704) to evaluate their role in PEX-ISGU. Ser588 and Thr642 are Akt phosphomotifs that control insulin-stimulated GLUT4 translocation and GU (9,11). Phosphorylation of these sites can be enhanced in insulin-stimulated muscle PEX (13,15,19,23). Ser704 (an AMPK phosphomotif) also responds to exercise (14,22,24,25,30). In addition, we evaluated other GU modulators (AMPK, glycogen, Akt, TBC1D1, and GLUT4).
Research Design and Methods
Materials
Materials for SDS-PAGE and immunoblotting were from Bio-Rad Laboratories (Hercules, CA) or Research Products International (Mount Prospect, IL). Table 1 provides information about antibodies, materials, and reagents. Other reagents were from Thermo Fisher Scientific (Hanover Park, IL) or Sigma-Aldrich (St. Louis, MO).
Antibody, material, or reagent . | Source (cat. no.) . |
---|---|
Anti-actin | Cell Signaling Technology (4968) |
Anti-Akt | Cell Signaling Technology (4691) |
Anti-AS160 | MilliporeSigma (ABS54) |
Anti–γ3-AMPK | Dr. David Thomson, Brigham Young University, Provo, UT |
Anti-GLUT4 | MilliporeSigma (CBL243) |
Anti-HA tag | Cell Signaling Technology (3724) |
Anti-phospho Akt Ser473, pAktSer473 | Cell Signaling Technology (4060) |
Anti-phospho Akt Thr308, pAktThr308 | Cell Signaling Technology (13038) |
Anti-phospho AMPKα Thr172, pAMPKThr172 | Cell Signaling Technology (2535) |
Anti-phospho AS160 Ser318, pAS160Ser318 | Cell Signaling Technology (8619) |
Anti-phospho AS160 Ser588, pAS160Ser588 | Cell Signaling Technology (8730) |
Anti-phospho AS160 Thr642, pAS160Thr642 | Cell Signaling Technology (4288) |
Anti-phospho AS160 Ser704, pAS160Ser704 | Dr. Jonas Thue Treebak, University of Copenhagen, Copenhagen, Denmark |
Anti-phospho TBC1D1 Ser237, pTBC1D1Ser237 | MilliporeSigma (07-2268) |
Anti-rabbit IgG, HRP-linked | Cell Signaling Technology (7074) |
Anti-TBC1D1 | Cell Signaling Technology (4629) |
[3H]-2-deoxyglucose (2DG) | PerkinElmer, Waltham, MA (NET328001MC) |
[14C]-mannitol | PerkinElmer, Waltham, MA (NEC314250UC) |
Human recombinant insulin | Eli Lilly, Indianapolis, IN (Humulin R U-100) |
Tissue protein extraction reagent (T-PER) | Thermo Fisher Scientific, Pittsburgh, PA (PI78510) |
Bicinchoninic Acid Protein Assay Kit | Thermo Fisher Scientific, Pittsburgh, PA (PI23223) |
MemCode Reversible Protein Stain Kit | Thermo Fisher Scientific, Pittsburgh, PA (PI24585) |
Protein G Magnetic Beads | Thermo Fisher Scientific, Pittsburgh, PA (10004D) |
DynaMag Magnet | Thermo Fisher Scientific, Pittsburgh, PA (12321D) |
Luminata Forte Western HRP Substrate | MilliporeSigma, Billerica, MA (WBLUF0500) |
Polyvinylidene fluoride (PVDF) | MilliporeSigma, Billerica, MA (IPVH00010) |
Bovine serum albumin | Research Products International, Mount Prospect, IL (A30075) |
Leupeptin | Research Products International, Mount Prospect, IL (L22035) |
Collagenase type 2 | Worthington Biochemical Corporation, Lakewood, NJ (LS004177) |
Bio-Safe II Scintillation Cocktail | Research Products International, Mount Prospect, IL (111195) |
Antibody, material, or reagent . | Source (cat. no.) . |
---|---|
Anti-actin | Cell Signaling Technology (4968) |
Anti-Akt | Cell Signaling Technology (4691) |
Anti-AS160 | MilliporeSigma (ABS54) |
Anti–γ3-AMPK | Dr. David Thomson, Brigham Young University, Provo, UT |
Anti-GLUT4 | MilliporeSigma (CBL243) |
Anti-HA tag | Cell Signaling Technology (3724) |
Anti-phospho Akt Ser473, pAktSer473 | Cell Signaling Technology (4060) |
Anti-phospho Akt Thr308, pAktThr308 | Cell Signaling Technology (13038) |
Anti-phospho AMPKα Thr172, pAMPKThr172 | Cell Signaling Technology (2535) |
Anti-phospho AS160 Ser318, pAS160Ser318 | Cell Signaling Technology (8619) |
Anti-phospho AS160 Ser588, pAS160Ser588 | Cell Signaling Technology (8730) |
Anti-phospho AS160 Thr642, pAS160Thr642 | Cell Signaling Technology (4288) |
Anti-phospho AS160 Ser704, pAS160Ser704 | Dr. Jonas Thue Treebak, University of Copenhagen, Copenhagen, Denmark |
Anti-phospho TBC1D1 Ser237, pTBC1D1Ser237 | MilliporeSigma (07-2268) |
Anti-rabbit IgG, HRP-linked | Cell Signaling Technology (7074) |
Anti-TBC1D1 | Cell Signaling Technology (4629) |
[3H]-2-deoxyglucose (2DG) | PerkinElmer, Waltham, MA (NET328001MC) |
[14C]-mannitol | PerkinElmer, Waltham, MA (NEC314250UC) |
Human recombinant insulin | Eli Lilly, Indianapolis, IN (Humulin R U-100) |
Tissue protein extraction reagent (T-PER) | Thermo Fisher Scientific, Pittsburgh, PA (PI78510) |
Bicinchoninic Acid Protein Assay Kit | Thermo Fisher Scientific, Pittsburgh, PA (PI23223) |
MemCode Reversible Protein Stain Kit | Thermo Fisher Scientific, Pittsburgh, PA (PI24585) |
Protein G Magnetic Beads | Thermo Fisher Scientific, Pittsburgh, PA (10004D) |
DynaMag Magnet | Thermo Fisher Scientific, Pittsburgh, PA (12321D) |
Luminata Forte Western HRP Substrate | MilliporeSigma, Billerica, MA (WBLUF0500) |
Polyvinylidene fluoride (PVDF) | MilliporeSigma, Billerica, MA (IPVH00010) |
Bovine serum albumin | Research Products International, Mount Prospect, IL (A30075) |
Leupeptin | Research Products International, Mount Prospect, IL (L22035) |
Collagenase type 2 | Worthington Biochemical Corporation, Lakewood, NJ (LS004177) |
Bio-Safe II Scintillation Cocktail | Research Products International, Mount Prospect, IL (111195) |
AAV Vector Production and Purification
Serotype AAV9 Y731F tyrosine mutant AAV9 was prepared as described (31,32). Expression of WT AS160 (AAV-WT-AS160), three phosphomutants (S588A, T642A, and S704A) on AS160 (AAV-3P-AS160), and GLUT4 (AAV-GLUT4) was codon optimized and expressed from the cytomegalovirus promoter. Both AS160 constructs included an N-terminal hemagglutinin (HA) tag. Alkaline phosphatase (AP) reporter was expressed from the RSV promoter. AAV vector was purified (two rounds of isopycnic cesium chloride ultracentrifugation). Vector titer was determined by quantitative SYBR Green PCR (forward primer 5′-TTACGGTAAACTGCCCACTTG; reverse primer 5′-CATAAGGTCATGTACTGGGCATAA).
Animal Treatment
The University of Michigan Committee on Use and Care of Animals approved animal care procedures. We previously described the creation of AS160-KO rats using CRISPR/Cas9 technology (29). Rats were genotyped (Transnetyx, Cordova, TN) as described previously (29). Male WT and homozygous AS160-KO rats were studied.
AAV Administration
Rats (5–7 weeks old) were anesthetized (2.5% isoflurane/100% oxygen), their forelimbs were shaved, and analgesic (5 mg/kg carprofen) was subcutaneously injected. A 5- to 7-mm skin incision was made, and the exposed epitrochlearis was rinsed (sterile PBS); each epitrochlearis was injected with AAV-AP (0.68 × 1013 vg/mL) in some WT rats and with AAV-WT-AS160 (0.6 × 1013 vg/mL), AAV-3P-AS160 (0.6 × 1013 vg/mL), or AAV-GLUT4 (0.06 × 1013 vg/mL) in AS160-KO rats, and the incision was sutured. Both muscles from some WT rats were injected with AAV-AP. Both muscles from some AS160-KO rats were injected with AAV-WT-AS160 or AAV-GLUT4. Controls for the AS160-KO with AAV-GLUT4 experiment were WT rats undergoing an identical surgical procedure as AS160-KO rats receiving AAV-GLUT4, except their muscles were injected with vehicle (sterile PBS; sham treatment). To compare effects of AAV-WT-AS160 versus AAV-3P-AS160 in AS160-KO rats, one muscle per rat was injected with AAV-WT-AS160. The contralateral muscle was injected with AAV-3P-AS160. Paired muscles from other AS160-KO rats were injected with AAV-WT-AS160 versus sham treatment or AAV-3P-AS160 versus sham treatment. Muscle lysates were immunoblotted to confirm AS160 expression in muscles injected with AS160 constructs, but not in paired sham-treated muscles (Supplementary Material). Using other muscles, transduction efficiency (TE) was assessed by immunoblotting of single myofibers isolated from AS160-KO rat muscles with or without AAV-WT-AS160 or with or without AAV-3P-AS160 (Supplementary Material). Terminal experiments occurred 3–4 weeks postinjection.
Rats were fed rodent chow (Laboratory Diet no. 5L0D; LabDiet, St. Louis, MO) until fasted (1700 h the day before experiment). The exercise protocol was swimming in a barrel filled with water (35°C) for four 30-min bouts with 5-min rest between bouts (14). Some rats were anesthetized (intraperitoneal injection of 50 mg/kg ketamine/5 mg/kg xylazine) immediately PEX (IPEX) along with sedentary rats, and muscles were dissected. At 3 h PEX (3hPEX), other rats (3hPEX and sedentary) were anesthetized, and muscles were dissected.
Muscle Incubation
Muscles underwent two-step incubation in gassed vials (95% O2/5% CO2 at 35°C) (22). Muscles from IPEX experiments were incubated without insulin (step 1 10 min; step 2 15 min). Muscles from 3hPEX experiments were incubated with or without 0.6 nM insulin (step 1 30 min; step 2 20 min).
Muscle Homogenization
Frozen muscles used for 2-deoxyglucose uptake and immunoblotting or to measure γ3-AMPK activity were processed as previously described (22). Frozen muscles used for glycogen analysis were homogenized in ice-cold water.
GU
2-Deoxyglucose uptake was determined as described (33). Δ exercise GU (value with exercise of individual muscles minus mean GU of respective sedentary group) was calculated for IPEX muscles. Δ insulin GU (GU of muscles with insulin minus GU of paired muscles without insulin) was calculated for 3hPEX muscles. For the final experiment, paired muscles were injected with AAV-WT-AS160 or AAV-3P-AS160 and incubated with the same insulin dose (± insulin). Mean GU without insulin for all muscles with the same AAV treatment was subtracted from individual values for the respective AAV treatment group to calculate Δ insulin GU.
Immunoblotting
Lysate total protein concentrations were measured (bicinchoninic acid assay). Equal amounts of sample protein were heated (3–5 min at 95°C) in SDS loading buffer (GLUT4-evaluated samples were not heated), separated via SDS-PAGE, and transferred to polyvinylidene fluoride membranes. MemCode protein stain confirmed equal loading (34). Membranes were incubated with appropriate primary and secondary antibodies. Enhanced chemiluminescence of protein bands was densitometrically quantified (ProteinSimple, San Jose, CA). Individual values were normalized to the mean value for all samples on the membrane. Δ insulin protein phosphorylation (protein phosphorylation of muscles with insulin minus protein phosphorylation of paired muscles without insulin) was calculated for 3hPEX rats. Δ insulin protein phosphorylation for the final experiment (AAV-WT-AS160 vs. AAV-3P-AS160) used mean protein phosphorylation values from muscles incubated without insulin.
γ3-AMPK Activity Assay
Glycogen
Muscle glycogen concentration was determined using the Glycogen Assay Kit (catalog no. MAK016; Sigma-Aldrich, St Louis, MO) following the manufacturer’s protocol.
Statistics
Two-way ANOVA was used for comparisons among the four groups (sedentary or IPEX in WT or AS160-KO group and sedentary or 3hPEX ± insulin in same genetic group or Δ insulin for sedentary or 3hPEX in two genetic groups). Tukey post hoc analysis identified significant variance (Sigma Plot 14.5; Systat, San Jose, CA). A two-tailed t test (unpaired or paired, as appropriate) was used to compare two groups. Data were expressed as means ± SD. P values ≤0.05 were considered statistically significant.
Data and Resource Availability
Data from the current study are available from the corresponding author upon reasonable request.
Results
Genotype Analysis
Genotype was determined by quantitative PCR analysis of DNA from tail samples. AS160 protein was undetectable in muscles from AS160-KO rats not injected with AAV-WT-AS160 or AAV-3P-AS160 (Figs. 1A, 2A, and 4A). In muscles from AS160-KO rats, immunoblotting using anti-AS160 or anti-HA (AS160 constructs included an N-terminal HA tag) confirmed AS160 expression in epitrochlearis injected with AAV-WT-AS160 or AAV-3P-AS160, but expression was absent in contralateral sham-treated muscles (Supplementary Material).
AAV TE
The AAV vectors used in the study included HA-tagged AAV-WT-AS160, HA-tagged AAV-3P-AS160, and AAV-GLUT4 (without a tag). Because these vectors were packaged in the same AAV capsids, the transgene was expressed from the same expression cassette, and the vector was delivered using an identical method to the same muscle, we expected them to yield similar TE. To unambiguously determine TE, we assessed AAV-WT-AS160–injected muscles via immunoblotting with anti–HA tag. Quantification revealed 96.1 ± 3.1% TE (Supplementary Materials). To confirm that the transduction quantified with this vector was applicable to other vectors, we determined TE in AAV-3P-AS160–injected muscles to be 95.8 ± 1.4% (Supplementary Materials).
Exercise Effects on WT and AS160-KO Rats
The first experiment assessed the role of AS160 in exercise effects on GU using WT and AS160-KO rats. Exercise induced greater insulin-independent GU in muscles from WT and AS160-KO rats versus respective sedentary controls (P < 0.001) (Fig. 1B). Both IPEX GU and Δ exercise GU in AS160-KO rats were ∼50% lower than those in WT rats (P < 0.001) (Fig. 1C). The magnitude of IPEX effects on glycogen (P < 0.001) (Fig. 1D), γ3-AMPK activity (P < 0.05) (Fig. 1E), and pAMPKThr172 (P < 0.05) (Fig. 1F) did not differ between genotypes. TBC1D1 abundance was not different for WT versus AS160-KO groups (Fig. 1G). No genotype or IPEX effects were detected for TBC1D1 abundance or pTBC1D1Ser237 (Fig. 1G and H).
For both WT sedentary and WT 3hPEX rats, GU was greater in insulin-stimulated versus paired muscles without insulin (P < 0.01) (Fig. 2B). GU in insulin-stimulated WT 3hPEX muscles exceeded that in WT sedentary (P < 0.01). Δ insulin GU of WT 3hPEX animals exceeded values from WT sedentary (P < 0.001) (Fig. 2C).
GU, with or without insulin, was greater for AS160-KO 3hPEX rats versus sedentary controls (P < 0.001) (Fig. 2B). However, the exercise effect on insulin-stimulated muscles was insulin independent (i.e., no difference in Δ insulin GU between AS160-KO 3hPEX and AS160-KO sedentary) (Fig. 2C). Thus, exercise did not enhance muscle ISGU in AS160-KO rats. In addition, Δ insulin GU in WT rats exceeded that in AS160-KO rats in sedentary (P < 0.005) and 3hPEX (P < 0.001) groups. The lack of PEX effect on ISGU in AS160-KO rats was not because of deficient exercise effect on muscle γ3-AMPK activity, which was similarly elevated in WT 3hPEX and AS160-KO 3hPEX groups versus sedentary (P < 0.01) (Fig. 2D), with no genotype differences. pAktThr308 (P < 0.001) (Fig. 2E) and pAktSer473 (P < 0.001) (Fig. 2F) were similarly increased by insulin in 3hPEX and sedentary groups. No genotype differences were observed for Δ insulin pAkt (data not shown).
pSer588 (P < 0.01) (Fig. 2G) and pThr642 (P < 0.01) (Fig. 2H) were increased in insulin-stimulated versus basal muscles from WT rats, with or without exercise. In insulin-stimulated muscles of WT rats, exercise led to greater pSer588 and pThr642 versus sedentary (P < 0.01). There was a nonsignificant trend for an exercise effect on pSer704, but no insulin effect in WT rats (Fig. 2I). There were no exercise differences for Δ insulin pAS160 on any phosphosite in WT rats (data not shown). There was no detectable pSer588, pThr642, or pSer704 in AS160-KO rats. TBC1D1 abundance and pTBC1D1Ser237 did not differ among the groups (Fig. 2J and K). GLUT4 abundance was markedly lower in AS160-KO versus WT rats (P < 0.001) (Fig. 2L).
Exercise Effects on AS160-KO Rats With Restored AS160 Abundance
We next tested if AAV delivery of WT AS160 to muscle of AS160 KO rats rescued insulin resistance and restored exercise-enhanced ISGU. Immunoblotting confirmed that AAV delivery of WT AS160 eliminated muscle AS160 deficiency in AS160-KO versus WT rats (Fig. 3A). For both AAV-AP in WT rats and AAV-WT-AS160 in AS160-KO rats, 3hPEX GU with insulin values exceeded those in sedentary counterparts (P < 0.05) (Fig. 3B). In sedentary rats, Δ insulin GU was not different for WT rats with AAV-AP versus AS160-KO rats with AAV-WT-AS160 (Fig. 3C), demonstrating AAV-WT-AS160 delivery eliminated muscle insulin resistance in AS160-KO rats. Δ insulin GU of muscles from 3hPEX animals exceeded Δ insulin from AAV-treated sedentary controls (P < 0.01). Furthermore, Δ insulin GU at 3hPEX did not differ between WT rats with AAV-AP versus AS160-KO rats with AAV-WT-AS160, revealing that restoring muscle AS160 expression in AS160-KO rats rescued exercise-enhanced ISGU.
With or without exercise, pAkt with insulin exceeded basal (P < 0.05 for Thr308 and P < 0.05 for Ser473) (Fig. 3D and E) in muscles from WT rats with AAV-AP and in AS160-KO rats with AAV-WT-AS160 (P < 0.001 for Thr308 and P < 0.001 for Ser473). No genotype differences were observed for Δ insulin pAkt (data not shown). Insulin also increased pAS160 (P < 0.01 for Ser588 and P < 0.05–0.01 for Thr642) (Fig. 3F and G) regardless of genotype. There were nonsignificant trends for 3hPEX to exceed sedentary for pSer588 (P = 0.07) and pThr642 (P = 0.08) of WT rats with AAV-AP and for pThr642 (P = 0.06) of AS160-KO rats with AAV-WT-AS160. Exercise enhanced pSer704 with or without insulin only in muscles from AS160-KO rats with AAV-WT-AS160 (Fig. 3H) (P < 0.05). No exercise or genotype differences in Δ insulin pAS160 were observed on any phosphosites (data not shown). AAV-WT-AS160 treatment of muscles from AS160-KO rats prevented a significant deficit versus muscles from WT rats treated with AAV-AP (Fig. 3I).
Exercise Effects on AS160-KO Rats With Restored GLUT4 Abundance
AS160-KO rats were treated with AAV-GLUT4 to test if the ability of AAV-WT-AS160 to rescue PEX-ISGU was an indirect effect of restoring GLUT4 abundance. AS160 was undetectable in muscles from AS160-KO rats treated with AAV-GLUT4 (Fig. 4A). GLUT4 abundance of AAV-GLUT4–treated muscles from AS160-KO rats did not differ versus WT rats (Fig. 4B).
GU was greater in insulin-stimulated muscles versus paired muscles without insulin for WT sedentary (P < 0.05) and WT 3hPEX rats (P < 0.001) (Fig. 4C). Both GU with insulin (P < 0.01) and Δ insulin GU (P < 0.01) (Fig. 4D) for WT 3hPEX exceeded those for WT sedentary. Δ insulin GU for WT 3hPEX rats also exceeded AS160-KO 3hPEX with AAV-GLUT4 (P < 0.001).
For sedentary AS160-KO rats receiving AAV-GLUT4, GU of muscles with insulin tended (P = 0.058) (Fig. 4C) to exceed GU of paired muscles without insulin. A significant increase for AS160-KO rats receiving AAV-GLUT4 versus paired controls with insulin was detected based on a paired t test (P < 0.01). Δ insulin GU was not significantly different for AS160-KO sedentary with AAV-GLUT4 versus WT sedentary (Fig. 4D), suggesting AAV-GLUT4 prevented insulin resistance in AS160-KO rats without AAV-GLUT4. Within 3hPEX in AS160-KO rats receiving AAV-GLUT4, GU was not different with insulin versus without insulin. PEX versus sedentary AS160-KO rats receiving AAV-GLUT4 had greater GU, both with (P < 0.05) and without (P < 0.01) insulin. However, exercise did not enhance Δ insulin GU for AS160-KO rats receiving AAV-GLUT4, providing evidence that the role of AS160 in PEX Δ insulin GU was not simply an indirect consequence of lower GLUT4 abundance.
The 3hPEX values for muscle pAMPKThr172 exceeded sedentary control values of muscles, both without (P < 0.05) (Fig. 4E) and with AAV-GLUT4 (P < 0.01). Neither total TBC1D1 nor pTBC1D1Ser237 was altered by exercise or insulin regardless of genotype (Fig. 4F and G).
pAktThr308 (P < 0.001) (Fig. 4H) and pAktSer473 (P < 0.001) (Fig. 4I) were increased with insulin versus without insulin, regardless of exercise or genotype. In insulin-stimulated muscles from AS160-KO rats receiving AAV-GLUT4, 3hPEX exceeded sedentary for pAktThr308 (P < 0.01) and pAktSer473 (P < 0.05). Δ insulin pAktThr308 for 3hPEX exceeded sedentary within the same genotype (data not shown) (P < 0.05 for WT sham rats and P < 0.01 for AS160-KO rats receiving AAV-GLUT4). Δ insulin pAktSer473 for 3hPEX exceeded sedentary for AAV-GLUT4-KO with AAV-GLUT4 (data not shown) (P < 0.01).
Insulin increased pAS160Ser588 (P < 0.01–0.001) (Fig. 4J) and Thr642 (P < 0.001) (Fig. 4K) for WT rats. In insulin-stimulated muscles of WT rats, exercise led to greater pAS160Ser588 (P < 0.05) and pAS160Ser704 (P < 0.05) (Fig. 4L). Exercise also led to greater Δ insulin for pAS160Ser704 (data not shown) (P < 0.05).
Exercise Effects on AS160-KO Rats Expressing WT AS160 or 3P AS160
The final experiment determined the consequences of preventing AS160 Ser588, Thr642, and Ser704 phosphorylation on PEX-ISGU. As expected, injecting muscles from AS160-KO rats with AAV-WT-AS160 or AAV-3P-AS160 resulted in AS160 expression (Fig. 5A). AS160 abundance was moderately greater for AAV-WT-AS160–injected muscles from 3hPEX versus sedentary (P < 0.05), but no genotype differences (WT vs. 3P) in AS160 abundance were observed for sedentary or 3hPEX rats.
Insulin increased GU by muscles of both sedentary and 3hPEX AS160-KO rats injected with AAV-WT-AS160 or AAV-3P-AS160 (P < 0.05–0.001) (Fig. 5B). As expected, exercise elevated ISGU in muscles injected with AAV-WT-AS160 (P < 0.001). Exercise also increased ISGU in muscles injected with AAV-3P-AS160, demonstrating pAS160 on these sites was not required for this outcome (P < 0.001). Δ insulin GU of AAV-WT-AS160–expressing muscles was greater for 3hPEX versus sedentary (P < 0.001) (Fig. 5C). Δ insulin GU for 3hPEX versus sedentary was also elevated for muscles expressing AAV-3P-AS160 (P < 0.05). However, 3hPEX values from AAV-WT-AS160 muscles exceeded those of AAV-3P-AS160 muscles, indicating the 3P mutation attenuated the absolute value for this exercise effect (P < 0.005).
Insulin led to greater pAkt on Thr308 (Fig. 5D) and Ser473 (Fig. 5E) in sedentary AAV-WT-AS160– and AAV-3P-AS160–injected muscles (P < 0.05) and in 3hPEX AAV-WT-AS160– and AAV-3P-AS160–injected muscles (P < 0.001). No differences were detected between muscles expressing AAV-WT-AS160 versus AAV-3P-AS160 for Δ insulin pAkt (data not shown).
Insulin led to greater pAS160 on Ser588 (Fig. 5F), Thr642 (Fig. 5G), and Ser318 (Fig. 5I) of sedentary (P < 0.001) and 3hPEX (P < 0.001 for all sites except Ser588 [P < 0.01]) rats expressing WT AS160. Insulin did not elevate pSer704 in muscles expressing WT AS160, with or without exercise (Fig. 5H). Exercise did not enhance phosphorylation on any of these AS160 phosphosites in insulin-stimulated muscles expressing WT AS160. pSer588 of 3hPEX rats exceeded that of sedentary rats in muscles expressing WT AS160 incubated without insulin (P < 0.01).
There was no detectable pSer588, pThr642, or pSer704 in muscles expressing 3P AS160 (Fig. 5F–H). For muscles expressing 3P AS160, insulin resulted in increased pSer318 (Fig. 5I) in 3hPEX rats (P < 0.01), but not in sedentary animals. Insulin-stimulated muscles expressing 3P AS160 from 3hPEX rats versus sedentary rats had greater pSer318 (P < 0.05).
Muscle GLUT4 abundance was unaffected by exercise in AS160-KO rats injected with AAV-WT-AS160 or AAV-3P-AS160 (Fig. 5J). GLUT4 levels were markedly greater for muscles expressing 3P AS160 versus muscles expressing WT AS160 (P < 0.001). Thus, the lower Δ insulin GU for 3P AS160–expressing muscles versus muscles expressing WT AS160 was not attributable to GLUT4 deficiency.
Discussion
The current study tested if AS160 expression and/or site-selective phosphorylation is essential for improved PEX-ISGU. The major new observations were: 1) AS160 was essential for increased PEX-ISGU by isolated rat muscle; 2) rescuing muscle AS160 expression of AS160-KO rats restored enhanced PEX-ISGU; 3) restoring muscle GLUT4 abundance of AS160-KO rats did not rescue enhanced PEX-ISGU; and 4) although AS160 phosphorylation on Ser588, Thr642, and Ser704 was not required for elevated PEX-ISGU, it was essential for the full exercise effect.
The lack of improved PEX-ISGU by muscle from AS160-KO rats was not attributable to attenuated exercise effects on muscle γ3-AMPK activity or glycogen, two exercise effects proposed to elevate insulin sensitivity (3,24). Epitrochlearis TBC1D1 abundance was unaltered in AS160-KO versus WT rats, consistent with earlier results (29). Insulin did not elevate muscle pTBC1D1Ser237 in WT or AS160-KO rats, confirming previous results for insulin-stimulated human or rat muscle (19,35). pTBC1D1Ser237 was also not increased IPEX in either genotype. This observation differs from results for human muscle after various durations and intensities of exercise (cycling for 30 sec of all-out sprinting and 2 min at a work rate corresponding to ∼110% peak oxygen consumption or 20 min at ∼80% peak oxygen consumption) (36). The explanation for the lack of increased pTBC1D1Ser237 IPEX in the current study is uncertain, but it is not because the rat epitrochlearis cannot increase pTBC1D1Ser237. Electrically stimulated contractions by epitrochlearis increased pTBC1D1Ser237 (35). Notably, multiple lines of evidence argue against TBC1D1 being crucial for increased PEX-ISGU. Insulin-induced GU was not lower in mouse muscle expressing TBC1D1 mutated to prevent phosphorylation on Ser237, Thr505, Thr596, and Ser667 (37). Insulin-induced GLUT4 translocation was not lower in TBC1D1-KO versus WT rats (38). Insulin-stimulated pTBC1D1 was not enhanced PEX in rat or human muscle (17,19). Available evidence has not linked TBC1D1 to greater PEX-ISGU.
Consistent with earlier research (29,39), muscle GLUT4 abundance was lower in AS160-KO versus WT rats, possibly because AS160 deletion induces lysosome-dependent GLUT4 degradation (40). Decreased GLUT4 abundance can reduce ISGU by muscle from sedentary rodents (41). Restoring muscle GLUT4 of AS160-KO rats via AAV-GLUT4 delivery revealed that the essential role of AS160 could not be entirely attributed to lower GLUT4 abundance at 3hPEX. However, insulin-independent GU was elevated 3hPEX in muscles of AS160-KO rats with restored GLUT4 abundance, suggesting that AS160 expression modulates the PEX reversal of insulin-independent GU. A highly speculative explanation is that elevated insulin-independent GU at 3hPEX in muscles from AS160-KO rats may suggest a previously unsuspected role for AS160 in enabling PEX GLUT4 endocytosis. PEX enhancement of ISGU often coincides with reversal of most of the PEX elevation in insulin-independent GU (42). It is uncertain if elevated insulin-independent GU at 3hPEX played a role in attenuated PEX-ISGU by AS160-deficient, but GLUT4-restored, muscle. Additional research will be required to test if, at some later PEX time point, the elevated insulin-independent GU in AS160-KO muscle with normalized GLUT4 might reverse, and if this putative outcome might be accompanied by subsequently altered Δ insulin GU.
We next tested if restoring muscle AS160 expression of AS160-KO rats could rescue ISGU. Correction of the muscle AS160 deficit eliminated muscle insulin resistance in sedentary AS160-KO rats and restored the exercise-induced improvement in ISGU. Normalization of AS160 expression and ISGU (sedentary and PEX) attenuated the muscle GLUT4 deficit in AS160-KO rats.
We hypothesized that greater pAS160 on specific regulatory sites may be important for improved PEX-ISGU (3). Therefore, the final experiment tested if mutating AS160 to prevent phosphorylation on three key phosphomotifs would attenuate the exercise effect on ISGU. Δ insulin GU represents the ability of insulin to increase GU above basal levels, and the relative magnitude of the exercise increase in Δ insulin GU was similar for muscles expressing WT AS160 versus 3P AS160. These results revealed that pAS160 on Ser588, Thr642, and Ser704 was not required for exercise to improve ISGU. However, preventing phosphorylation of these sites reduced the absolute value of Δ insulin GU PEX. Thus, while phosphorylation on these sites was not essential for a PEX improvement in Δ insulin GU, having phosphorylatable sites was necessary for the full exercise benefit.
The lower absolute value for PEX Δ insulin GU in 3P AS160–expressing muscles was not caused by GLUT4 deficiency. Indeed, muscle GLUT4 abundance in AS160-KO rats expressing 3P AS160 greatly exceeded GLUT4 in muscles expressing WT AS160. These results are reminiscent of the findings for TBC1D4Thr649Ala knock-in mutant mice (Thr649 is equivalent to Thr642) (11). Knock-in mice versus WT controls were characterized by insulin resistance and elevated muscle GLUT4 abundance. The mechanism whereby AS160 phosphomutations cause greater GLUT4 content is unknown. A speculative scenario is that eliminating AS160 phosphorylation sites might inhibit GLUT4 exocytosis, leading to slower GLUT4 degradation and ultimately to greater GLUT4 accumulation.
Preventing AS160 phosphorylation on one site can alter phosphorylation on a separate site (25). Expressing AS160 with a mutation preventing phosphorylation on Ser711 (equivalent to Ser704) in mouse muscle resulted in reduced insulin-stimulated phosphorylation of Thr649 (equivalent to Thr642). AS160 harbors multiple exercise- and insulin-responsive phosphomotifs. To assess the possibility that the 3P mutation might indirectly influence other phosphosites, we also evaluated Ser318 phosphorylation. Preventing phosphorylation on Ser588, Thr642, and Ser704 did not significantly alter pSer318. It remains possible that the 3P mutation may influence AS160 phosphorylation on other phosphosites and thus indirectly contribute to the attenuated absolute value for Δ insulin GU PEX.
Elevated ISGU has been repeatedly reported in multiple species with various exercise protocols (3). PEX effects on epitrochlearis ISGU were similar to treadmill exercise effects on ISGU by rat hindlimb muscles (5) or single-knee extensor exercise effects on human vastus lateralis muscles (19). In their elegant single-knee extensor exercise model, Pehmoller et al. (19) found enhanced ISGU and pAS160 in exercised versus nonexercised muscles, providing compelling evidence of an exercise effect that was localized to the exercising muscle. Similarly, our exercise protocol enhanced epitrochlearis ISGU and pAS160 without altering either outcome in the extensor digiti quinti proprius, a rat forelimb muscle with similar mass and fiber type composition as the epitrochlearis (22). The correspondence between results from our exercise model and other exercise models provides evidence that the elevated PEX-ISGU in the current study was not attributable to nonspecific stress effects of the exercise protocol.
Rat epitrochlearis fiber type composition (type I 6–9%, IIA 12–17%, IIB 47–53%, and IIX 27–29%) (29,43,44) is similar to the overall fiber type profile of rat musculature (45). However, tissue analysis of one muscle cannot reveal fiber type–selective effects. We previously reported PEX enhancement of ISGU and pAS160 in isolated type I, IIA, and IIB fibers from rat epitrochlearis muscles, but neither outcome was elevated PEX in type IIX fibers from the same muscles (16,46). It would be valuable for future research to use genetic methods to evaluate fiber type–selective effects of AS160 on PEX-ISGU.
In conclusion, the novel observations in this study revealed that AS160 was essential for increased PEX-ISGU. These findings after in vivo exercise are consistent with observations from previous studies that used AS160-KO mice in response to prior muscle stimulation with AICAR or electrically induced contractions (28). The essential role of AS160 was not simply an indirect effect of GLUT4 deficiency. The results with 3P AS160 conclusively established that enhanced phosphorylation of Ser588, Thr642, and/or Ser704 was not essential for improved PEX-ISGU. Moreover, absolutely no phosphorylation on these sites was required for a PEX benefit on ISGU, although the full exercise effect was not achieved by 3P AS160–expressing muscles. These results support AS160-dependent mechanisms, but also provide evidence for future pursuit of AS160-independent mechanisms that mediate greater insulin sensitivity. Finally, it is crucial for future research to carefully address this topic in females.
A.Z. and E.B.A. contributed equally to this work.
This article contains supplementary material online at https://doi.org/10.2337/figshare.16930801.
Article Information
Acknowledgments. The authors thank Dr. Arun Srivastava (University of Florida) for providing the AAV9 tyrosine mutant Rep-Cap packaging plasmid, Dr. Jonas Treebak (University of Copenhagen) for generously supplying the pAS160Ser704 antibody, Dr. David Thomson (Brigham Young University) for generously supplying the γ3-AMPK antibody, Yongping Yue (University of Missouri) for AAV production and purification, Dominic Thorley, Gengfu Dong, and Jiahui Zhao (University of Michigan) for valuable technical assistance, and Katrina Fox (University of Michigan Unit for Lab Animal Medicine) for animal husbandry services.
Funding. This research was supported by a grant from the National Institute of Diabetes and Digestive and Kidney Diseases, National Institutes of Health (R01-DK-071771).
Duality of Interest. No potential conflicts of interest relevant to this article were reported.
Author Contributions. A.Z. and E.B.A. contributed to the conception and design of the experiments, performed experiments, analyzed data, created figures, and assisted in writing and editing the manuscript. H.W. and S.E.K. assisted in the experiments, analyzed data, created figures, and reviewed/edited the manuscript. X.P. generated the AAV reagents and reviewed/edited the manuscript. D.D. supervised the generation of the AAV reagents and reviewed/edited the manuscript. G.D.C. conceived the experiments, supervised the project, and wrote the first draft of the manuscript. All authors reviewed the manuscript and approved of its submission. G.D.C. is the guarantor of this work and, as such, had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.