Enhanced skeletal muscle insulin sensitivity after acute resistance-type exercise is upregulated by rapamycin-sensitive mTOR complex 1 inhibition

Acute aerobic exercise (AE) increases skeletal muscle insulin sensitivity for several hours, caused by acute activation of AMP-activated protein kinase (AMPK). Acute resistance exercise (RE) also activates AMPK, possibly improving insulin-stimulated glucose uptake. However, RE-induced rapamycin-sensitive mechanistic target of rapamycin complex 1 (mTORC1) activation is higher and has a longer duration than after AE. In molecular studies, mTORC1 was shown to be upstream of insulin receptor substrate 1 (IRS-1) Ser phosphorylation residue, inducing insulin resistance. Therefore, we hypothesised that although RE increases insulin sensitivity through AMPK activation, prolonged mTORC1 activation after RE reduces RE-induced insulin sensitising effect. In this study, we used an electrical stimulation–induced RE model in rats, with rapamycin as an inhibitor of mTORC1 activation. Our results showed that RE increased insulin-stimulated glucose uptake following AMPK signal activation. However, mTORC1 activation and IRS-1 Ser632/635 and Ser612 phosphorylation were elevated 6 h after RE, with concomitant impairment of insulin-stimulated Akt signal activation. By contrast, rapamycin inhibited these prior exercise responses. Furthermore, increases in insulin-stimulated skeletal muscle glucose uptake 6 h after RE were higher in rats with rapamycin treatment than with placebo treatment. Our data suggest that mTORC1/IRS-1 signaling inhibition enhances skeletal muscle insulin-sensitising effect of RE.

Inhibition of rapamycin-sensitive mTORC1 pathway activation. Rapamycin completely inhibited P70S6K phosphorylation in both basal and exercised states (Fig. 6a,b). Moreover, phosphorylation of IRS-1 at Ser632/635 and Ser612 were lowered by rapamycin in the basal state, and the exercise effects on these phosphorylations were diminished (Fig. 6a,d,e). However, IRS-1 Ser1100 phosphorylation levels were not different between the placebo and rapamycin groups (Fig. 6a,c).

Interaction between mTORC1/IRS-1 Ser pathway activation and insulin sensitivity after acute
Re. Insulin-stimulated TBC1D1 Ser231 phosphorylation was not affected by either prior RE or rapamycin ( Fig. 7a,d). In the placebo group, insulin-stimulated Akt Thr308 and Ser473 phosphorylations were lowered by prior RE, and the impairment was totally reversed by rapamycin (Fig. 7a-c). Following these upstream responses, although phosphorylation of TBC1D4 Ser597 and Thr651 in response to insulin was not different between control and exercised legs in the placebo group, these phosphorylations were significantly higher in the exercised leg in the rapamycin group (Fig. 7a,e,f). In Fig. 4e,g, we showed an elevation of p-TBC1D4 Ser597 and Thr651 by RE without insulin. This may be the reason why insulin-stimulated p-TBC1D4 Ser597 and Thr651 levels were higher in the exercised leg in rapamycin-treated rats. Furthermore, the increase in skeletal muscle glucose uptake by exercise under insulin stimulating conditions was significantly improved by rapamycin (Fig. 8a,b). These results suggest that prior RE-induced mTORC1 activation and subsequent IRS-1 Ser phosphorylation lowered the improvement of insulin sensitivity.

Discussion
The current study presents a new molecular mechanism regulating the insulin-sensitising effect of acute RE on skeletal muscle. Here, we provided evidence that a single bout of acute RE increases insulin-stimulated skeletal muscle glucose uptake, and this insulin-sensitising effect is augmented by inhibition of rapamycin-sensitive mTORC1 activation and subsequent IRS-1 Ser632/635 and Ser612 phosphorylation.
Rapamycin is well-known as a highly selective inhibitor of mTORC1 45 . Therefore, rapamycin was widely used to identify the role of mTORC1. Bentzinger et al. 59 have generated skeletal muscle-specific the mTORC1 component raptor knockout mouse; however, the mice have exhibited muscular dystrophy 59 and developed insulin resistance 60 . These phenotype affects exercise quality and post-exercise insulin sensitivity, although rapamycin did not change the total workload of RE (data are not shown) and insulin-stimulated muscle glucose uptake in both previous and present studies 61 . Therefore, the conventional raptor knockout model mouse was not the best method to determine the role of mTORC1 on insulin sensitivity after acute RE. In 2019, tamoxifen-inducible raptor knockout mice were newly generated, and it minimised chronic adaptation by raptor deletion 62 . Thus, in the future study, we can explain the specific role of mTORC1 on insulin-sensitising effect following RE by using the inducible raptor knockout model mice.
In previous studies, an increase in skeletal muscle insulin sensitivity after AE was not associated with enhanced proximal insulin signaling in humans and rodents 3,8,9 . Furthermore, insulin-stimulated IRS-1 Tyr phosphorylation did not differ between the rested and the exercised legs, although the glucose uptake of the exercised leg during insulin clamp was enhanced 9 . By contrast, the increase in insulin-stimulated skeletal muscle glucose uptake after AE was diminished by skeletal muscle-specific AMPKα1/α2 knockout mice 15 . Therefore, they suggested that AMPK activation, but not IRS-1/Akt signaling in response to exercise, was important for muscle insulin sensitivity. However, in the case of RE, we found that prolonged mTORC1/IRS-1 signal activation by exercise may attenuate insulin-stimulated Akt/TBC1D4 signal activation, although AMPK was activated immediately. These results suggest that exercise mode, causing elevated and prolonged mTORC1 activation, interrupts prior AMPK activation-related increases in insulin sensitivity in skeletal muscle.
Theoretically, skeletal muscle glucose uptake should reflect Akt signal activation. However, insulin-stimulated glucose uptake was improved by RE, although Akt signal was impaired in this state (Fig. 5). Interestingly, previous studies suggested that TBC1D4 phosphorylation at Thr649 and Ser711 (equivalent to Thr651 and Ser713 on rat, respectively) were important sites for the insulin-sensitising effect of prior AMPK activation [15][16][17][18] . Moreover, TBC1D4 Ser711 (rat; Ser713) was more reflective of increasing insulin sensitivity by prior exercise than was Thr649 (rat: Thr651) 63,64 . In the present study, we only measured TBC1D4 phosphorylation at Thr651; we found that the phosphorylation in response to insulin was impaired, as Akt, in the exercised leg. Although we did not measure Ser713, we could speculate that this phosphorylation under insulin stimulation might be lowered by impairment of Akt signaling. Nevertheless, a prior AMPK activation by RE might still have further facilitated insulin-stimulated Ser713 phosphorylation and glucose uptake. Therefore, we need to confirm this in a future study.
As the other limitation, based on the previous finding that skeletal muscle AMPKα1/α2 deletion diminished insulin-sensitising effect of in-situ muscle contraction and running exercise 15 , we expected the role of AMPK on RE-induced insulin-sensitising effects. If we could inhibit mTORC1 activation on skeletal muscle-specific AMPK knockout animals, we could directly identify whether AMPK knockout diminishes the enhanced insulin-sensitising effect of RE by inhibiting mTORC1. Additionally, we have not used a female rat for this present study because the menstrual cycle affects insulin sensitivity 65 . However, it is also important to show whether current evidence can replicate in female rats. Thus, the current evidence will be extended by an additional study confirming the role of mTORC1 on RE-induced insulin-sensitising effect in both male and female AMPK knockout animals. Values are means ± standard error. *P < 0.05 versus placebo injection within CON or RE legs, ♯ P < 0.05 versus CON leg for each group, § P < 0.05 versus response to insulin (interaction of insulin × RE). RE, resistance exercise; CON, unstimulated control.

Scientific RepoRtS |
(2020) 10:8509 | https://doi.org/10.1038/s41598-020-65397-z www.nature.com/scientificreports www.nature.com/scientificreports/ Overall, we provided evidence that mTORC1 activation and subsequent IRS-1 Ser phosphorylation opposed the insulin-sensitising effect of acute RE on skeletal muscle. Although mTORC1 activation was thought to be the most important target for skeletal muscle hypertrophy by chronic resistance training 35-37 , our results newly Values are means ± standard error. *P < 0.05 versus placebo injection within CON or RE legs, ♯ P < 0.05 versus CON leg for each group, § P < 0.05 versus response to RE (interaction of RE × rapamycin). RE, resistance exercise; CON, unstimulated control.

Methods ethical approvals. The study protocols were approved by the Ethics Committee for Animal Experiments at
Ritsumeikan University (BKC2018-033). We do confirm that all experiments were performed in accordance with relevant guidelines and regulations.
Animals were maintained at 22 °C-24 °C with 12-h light-dark cycles. Food (CE-2; CLEA Japan, Tokyo, Japan), and water were available ad libitum. After at least 1-week acclimatisation period, the animals were subjected to each experiment.
Resistance-type exercise. Acute RE was mimicked as previously described 66 . Briefly, overnight-fasted rats were anaesthetised with isoflurane, and the right whole gastrocnemius muscle was subjected to maximal isometric contraction using percutaneous electrical stimulation (5 sets of 3-s stimulation × 10 contractions per set with 7-s intervals between contractions and 3-min rest between sets) with an electric stimulator and isolator (SS-104J; Nihon Kohden, Tokyo, Japan). The stimulation protocol called for 100 Hz, 4 ms and ~50 V. The left gastrocnemius muscle was saved as a non-exercise control. Muscle samples were obtained at either 0 or 6 h after RE. Six hours after RE, rats were assigned to the study identifying the effect of mTORC1 activation on insulin sensitivity, because the time point showed marked elevation of mTORC1 activity after the exercise 67,68 . Tissues were rapidly harvested and frozen in liquid nitrogen and stored at −80 °C until analysis. This RE method was established because stimulation induces 8-10% gastrocnemius muscle hypertrophy with 12-18 sessions in rats 67,68 . In vivo insulin stimulation. The exercised rats were anaesthetised with 2% isoflurane in air and were intraperitoneally injected with either insulin (2 U/kg body weight dissolved in saline; Novo Nordisk A/S, Bagsvaerd, Denmark) or saline 10 or 30 min before muscle sampling. This amount of insulin stimulation for 10 to 30 min was previously shown to increase skeletal muscle Akt pathway activation and decrease blood glucose levels in rats 69,70 . Inhibition of mTORC1 activity. Rapamycin was used for mTORC1 inhibition, as previously shown 57 .
Briefly, rapamycin (1.5 mg/kg, 0.25 mg/mL in saline containing 0.5% dimethyl sulphoxide) or placebo (saline containing 0.5% dimethyl sulphoxide) was intraperitoneally injected 1 h before RE. Following the method of insulin stimulation, these rats were treated with insulin (2 U/kg body weight) at 5.5 h post-exercise, and then muscle samples were taken 30 min after insulin injection.
In vivo 2-deoxy-d-glucose uptake. 2-Deoxy-d-glucose (2DG) uptake method that we used was originally established by Saito et al. 71 and widely used as in-vivo 2DG uptake measurement with some optimisations in previous studies, including ours 32,72,73 . Particularly, the anaesthetised rats were administered 2DG (166 nmol/g body weight) into a vein 20 min before muscle sampling. At the time of muscle sampling, gastrocnemius muscles were rapidly harvested, then frozen in liquid nitrogen. The frozen tissues were homogenised ultrasonically in 10 mmol/L Tris·HCl buffer (pH 8.1), heated at 95 °C for 15 min, and centrifuged at 17,800 g for 15 min at 4 °C. The transported 2DG into muscle accumulates as 2DG-6-phosphate (2DG6P); thus, the 2DG6P concentration in the supernatant was assessed with the enzyme cycling method (Nonradioactive 2DG Uptake Assay Kit; Cosmo Bio, Tokyo, Japan). In this method, we firstly oxidised glucose-6-phosphate (G6P) with a low concentration of glucose-6-phosphate dehydrogenase (G6PDH) and nicotinamide adenine dinucleotide (NAD + ) to eliminate G6P in the lysate. As a second step, following the elimination of endogenous nicotinamide adenine dinucleotide phosphate (NADPH) and produced NADH, NADPH was produced through the oxidation of 2DG6P with a high concentration of G6PDH. The produced NADPH was used for quantification of 2DG6P with a microplate spectrophotometer (Bio-Rad, Hercules, CA, USA).
Statistical analysis. Data are presented as means ± standard error. Two-way analysis of variance with repeated measures and paired/unpaired Students t-tests were used to assess statistical significance within and between interventions, where appropriate. Post hoc analysis was performed using t-tests with Benjamini-Hochberg false discovery rate correction, when appropriate. The main effects have been indicated by lines unless stated otherwise. Statistical significance was defined as P < 0.05.

Data availability
The datasets generated during and analysed during the current study are available from the corresponding author on reasonable request.