Genetically determined exercise capacity affects systemic glucose response to insulin in rats

Abstract

Aerobic exercise capacity is inversely related to morbidity and mortality as well as to insulin resistance. However, exercising in patients has led to conflicting results, presumably because aerobic exercise capacity consists of intrinsic (genetically determined) and extrinsic (environmentally determined) parts. The contribution of both parts to insulin sensitivity is also not clear. We investigated sedentary and exercised (aerobic interval training) high-capacity runners (HCR) and low-capacity runners (LCR) differing in their genetically determined aerobic exercise capacity to determine the contribution of both parts to insulin sensitivity. LCR and HCR differed in their untrained exercise capacity and body weight. Sedentary LCR displayed a diabetic phenotype with higher random glucose, lower glucose infusion rate during hyperinsulinemic euglycemic clamping than HCR. Echocardiography showed equal morphological and functional parameters and no change with exercise. Four week of exercise caused significant improvements in aerobic exercise capacity, which was more pronounced in LCR. However, with respect to glucose use, exercise affected HCR only. In these animals, exercise increased 2-deoxyglucose uptake in gastrocnemius (+58.5%, P = 0.1) and in epididymal fat (+106%; P < 0.05). Citrate synthase activity also increased in these tissues (gastrocnemius 69% epididymal fat 63%). In our model of HCR and LCR, genetic predisposition for low exercise capacity is associated with impaired insulin sensitivity and impedes exercise-induced improvements in insulin response. Our results suggest that genetic predisposition for low aerobic exercise capacity impairs insulin response, which may not be overcome by exercise.

INTRODUCTION

Aerobic exercise capacity has been described as an independent predictor of morbidity and mortality (1). Low endurance exercise capacity shows a close relationship to all-cause morbidity and mortality (2). Here, exercise capacity is the strongest predictor of long-term survival, stronger than many other risk factors such as body mass index, diabetes, insulin resistance, or hypertension independent if the individual is healthy or a patient with cardiovascular disease (3). Low exercise capacity is also associated with insulin resistance (4). Exercise training increases total exercise capacity and reduces fat mass and symptoms of diabetes/metabolic syndrome (5). Exercise training increases insulin sensitivity in normal subjects and in offspring of parents with diabetes through stimulation of muscle glucose uptake (6) Exercise training increases insulin-dependent glucose uptake and sensitizes muscle glucose uptake to subsequent glucose uptake (7). Exercise reduces liver fat via an increase in VLDL contributing to improved hepatic insulin sensitivity (8). It has been suggested that exercise improves signaling through insulin receptor IRS also in the heart (6). However, exercise has been described to benefit many, but not all individuals with type 2 diabetes and the reason why some do not respond favorably seems under investigated (9). Individuals with T2DM have impaired exercise capacity, and slower oxygen delivery to skeletal muscle has been suggested as a cause (10). However, it has been reported that the magnitude of the training is greater with increased relative intensity (9), and that the extent of the training response is largely based on heritable factors. Epigenetics as well as muscular myokines have been suggested to contribute to such differences (9). However, there is evidence that DNA sequence differences are the most important contributors to the response to regular exercise (11). Thus, the effect of exercise training on insulin sensitivity remains a matter of debate (12) and more large scale interventional studies appear to be required (11).

Exercise capacity is composed of a genetically determined (intrinsic) and an acquired (training) part (13). The respective contribution of both to exercise capacity (and thereby presumably to health status) is not clear and cannot be properly separated in humans. In contrast, the animal model of high-capacity runners (HCR) and low-capacity runners (LCR) has been developed using selective breeding based on intrinsic exercise capacity (without exercise training) (14). It is the currently the only animal model that was bred selective for aerobic exercise capacity. LCR and HCR display differences for numerous health features including reduced life span (15), increased blood pressure, and higher cardiovascular risk (16). The differences between strains can be accounted for by selection pressure and random genetic drift (17). The differences of HCR and LCR are thought to be due to a network of interacting quantitative trait loci where a high number of subtle changes but not a single gene accounts for the observed differences (18). LCR show a metabolic syndrome-like phenotype with an increase in body weight, serum triglycerides, free fatty acids, fasting glucose, and insulin as well as insulin resistance (16, 19). HCR rats, in contrast, are resistant to adverse effects of high-fat diet (HFD) (obesity, insulin resistance) and numerous associated diseases (16).

Both rat lines were started from a genetically diverse founder population (N:NIH) and inbreeding was kept below 1% per generation maintaining genetic variability comparable to the human population (20). Thus, this model offers the opportunity to assess the influence of intrinsic (i.e., genetically predetermined) as well as acquired (training induced) exercise capacity and their interaction on insulin sensitivity on a polygenic background. The model is therefore well suited to deliver clinically relevant information.

We aimed in this study to assess the hypothesis that LCR present with impaired insulin response and that insulin sensitivity may be improved with exercise training.

MATERIALS AND METHODS

Materials

Chemicals were obtained from Carl Roth (Karlsruhe, Germany), Sigma Aldrich (Deisenhofen, Germany) (DTNB, acetyl CoA, Triton-X-100, oxaloacetic acid, cumarinic acid, sodium orthovanadate), Merck-Millipore (Darmstadt, Germany) (Glut4 Antibody07-1404), Serva (Heidelberg, Germany) (chemiluminescence reagent 42582), Bio-Rad (München, Germany) (AG1-X8 Resin formate) and Perkin Elmer (3-3H glucose, 14 C-deoxy glucose) and B.Braun (Melsungen, Germany) (Glucose 20%, 40%).

Animals

All animal procedures were approved by the Animal Welfare Committee of the University of Jena, Germany and the local authorities (22–2684-04-02-055/10). Animals had free access to food and water. During exercise tests or training, no access to water or food was possible. For the hyperinsulinemic-euglycemic clamp, animals were fasted overnight. Animals were handled and housed in accordance with National Institutes of Health (NIH) guidelines. The creation of the high-/low-capacity runner (HCR/LCR) rat model has been described previously (14). Briefly, bidirectionally selected lines were generated from a founder population of 80 male and 88 female N:NIH stock rats based on intrinsic aerobic treadmill running capacity. Thirteen families for each line were set up for a within-family rotational breeding paradigm that keeps the inbreeding at < 1% per generation. At each generation, young adult rats (11 wk of age) were tested for their intrinsic ability to perform a speed-ramped treadmill-running test until exhaustion. This test was performed on three separate days with the rats given one full day of recovery in between. The greatest distance in meters achieved out of the three trials was considered the best estimate of an individual rat’s aerobic exercise capacity (14). The highest scored female and male from each of the 13 families were selected as breeders for the next generation of high-capacity runners (HCR). The same process was used with lowest scored females and males to generate low-capacity runners (LCR). Male HCR and LCR rats (generation 28, 36–41 wk of age at time of final investigation) were used for this investigation and housed in pairs in a temperature-controlled environment with a 12-h light/dark cycle.

Echocardiography

The animals were anesthetized with Fentanyl/Midazolam-hydrochloride/Medetomidine-hydrochloride (0.005/2/0.15 mg/kg). Chests were shaved and the rats were examined in supine position with a 716 RMV scanhead (11–24 MHz) from Fujifilm Sonosite (Visualsonics), The Netherlands. Two-dimensional short-axis views of the left ventricle at papillary muscle level were obtained. We determined left ventricular wall thickness (PWT) and cavity size in both systole (LVESD) and diastole (LVEDD) by the American Society for Echocardiography leading edge method and averaged values from five measurements for each examination. Ejection fraction (EF) and fractional shortening (FS) were determined according to Teichholz et al. (21). To assess diastolic function, pulsed Doppler signals of the left ventricular inflow were assessed in the apical four-chamber view, and the sample volume was placed at the mitral tip level. The peak velocities of E- and A-waves and the deceleration time of the E-wave were assessed. Pulse-wave tissue Doppler was obtained from the four-chamber view. Early diastolic wave (É) and the late diastolic wave (Á) were detected and their peak velocities were measured. Furthermore, E/e′ was calculated as an index of preload.

Exercise Testing

Before exercise training was performed, a maximal exercise capacity test was performed by running to exhaustion (total exercise capacity) on a treadmill at 25° incline according to Hoydal et al. (22). After 15 min, adjustment to the treadmill [HCR (0.22 m/s) and LCR (0.07 m/s)], speed was increased every 2 min by 1.8 m/min until the animals were unable to run further. The test was repeated twice every other day. To take into account for anaerobic exercise, the last four steps (7.2 m/min) were subtracted and the result was set as 100% (maximum) aerobic exercise capacity.

Aerobic Interval Training

LCR/HCR rats were assigned to groups after the exercise test such that equal distribution of untrained exercise capacity was achieved between sedentary and exercised groups. Groups were HCR sedentary (HCR-S), HCR exercised (HCR-Ex), LCR sedentary (LCR-S), and LCR exercised (LCR-Ex). Exercise was performed as relative aerobic interval training (AIT) for 1.5 h/day, 5 days/wk, 4 wk in total, on a treadmill at 25° incline (23). After 15 min of warm-up, animals were trained for 1.5 h in AIT with 8 min at 85% and 2 min at 55% of maximal aerobic exercise capacity. Treadmill velocity was increased by 0.02 m/s weekly if the animal was able to successfully complete at least three of the five training sessions during 1 wk (22). Running speed and time were recorded during each session. Running distance was calculated from treadmill speed and running time.

Hyperinsulinemic-Euglycemic Clamp

After an overnight fast, the hyperinsulinemic euglycemic clamp was conducted as described (24). Briefly, rats were anesthetized, and the jugular vein was catheterized. The venous catheter was used for infusions in all protocols. The experimental period started with a bolus of 3-H glucose (0.16 nmol = 58 KBq) followed by continuous infusion of ³H-glucose (0.12 nmol/h) to assess basal glucose metabolism. Insulin (Humulin R; Eli Lilly) was given to stimulate glucose metabolism and glucose was given to maintain stable blood glucose levels. Stability was defined by steady euglycemia for 15 min. 2-deoxy-d-[¹⁴C]-glucose was given and blood samples were collected from the tip of the tail at different time points and centrifuged and plasma was stored at −20°C until analyzed. After the final blood sample in euglycemic clamp, animals were euthanized and heart, lung, liver, pancreas, spleen, kidney, soleus, gastrocnemius, epididymal fat pads, and brain were harvested and frozen in liquid nitrogen. Glucose uptake was determined by the quantification of ¹⁴C-phospho-deoxyglucose, glucose turnover as the disappearance of ³H-glucose, and glycolysis as the rate of ³H₂O production. Blood glucose refers to capillary blood glucose measured using the glucose oxidase method.

Isolation and Separation of Cellular and Vesicle Membranes

Membrane fractions were isolated from frozen rat tissue according to a modified protocol of Hirshman (25). Briefly, frozen samples of M. gastrocnemius and heart muscle were powdered in liquid nitrogen. Tissues (200 mg) were homogenized at 4°C (Dounce homogenizer; 1,500 rpm) in 600 µL homogenization buffer (20 mM Tris, 1 mM EDTA, 255 mM sucrose, 0.2 M PMSF at pH 7.4). Homogenates were centrifuged (Eppendorf 5810 R) at 18,000 g for 20 min. Vesicle membranes were isolated from supernatant by centrifugation at 72,000 rpm for 65 min (Beckman Coulter, TLA 120.2) and resuspended in SDS lysis buffer (1% SDS, 10 mM EDTA, 50 mM Tris pH 7.4). Pellets were resuspended in 600 µL of HB, homogenized, and centrifuged at 14,000 g. Again, pellets were resuspended in 600 µL of HB and layered on 400-µL sucrose cushion (1.12 M). Plasma membranes were pelleted from supernatants at 65,000 rpm for 20 min and resuspended in SDS lysis buffer.

Determination of Citrate Synthase Activity

Powdered tissue was homogenized in buffer (100 mM KCl, 50 mM MOPS, 0.5 mM EGTA) on ice, sonicated on ice, and centrifuged at 12,000 g. Supernatant was diluted 1:20 using the same buffer and citrate synthase activity was measured as described previously (26).

Denaturing Gel Electrophoresis and Western Blotting

Proteins were isolated from frozen tissue samples. Western blots were performed on a semidry Western blot apparatus. After SDS-PAGE electrophoresis, proteins of related membrane fractions were blotted to the same PVDF membrane and incubated with primary and secondary antibodies. Bands were visualized by chemiluminescence and semiquantified using ImageJ software. Antibodies against Glut4, insulin receptor β, Akt, and CD36 were obtained from Millipore (No. 07–1404), Cell Signaling (No. 30255, No. 9272), and Santa Cruz (sc-9154) and secondary antibody from GE Healthcare (No. NA934).

Statistical Analysis

Data are presented as means ± SE and were analyzed using two-way analysis of variance. Post hoc comparisons among the groups were performed using the Bonferroni method. Student’s t test was used for data shown in Table 1 as indicated. Differences among groups were considered statistically significant if P < 0.05 and main effects for population and treatment are given in figures and tables.

Table 1.

Exercise-induced improvement in exercise capacity of HCR and LCR during 4 wk of exercise

	HCR		LCR
	First Week	Last Week	First Week	Last Week
Running speed, m/min	13.9 ± 0.5	15.4 ± 0.4*	4.2 ± 0.3	7.1 ± 0.3***
Running time, min	67.1 ± 6.5	84.9 ± 1.7*	88.8 ± 1.2	88.4 ± 1.6
Distance run, m	850 ± 62	1218 ± 36***	341 ± 24	583 ± 19***
Increase in speed, %	11.8 ± 4.1		75.8 ± 6.4***
Increase in distance run, %	47.8 ± 9.8		75.1 ± 7.0*

Values are means ± SE; n = 8–10/group. HCR, high-capacity runners; LCR, low-capacity runners. *P < 0.05; ***P < 0.001 to first week of exercise training as analyzed with paired t test within each line (running, speed, time, and distance) or to HCR as calculated using unpaired t test for increase in speed and distance run.

RESULTS

HCR presented with higher intrinsic exercise capacity than LCR as measured with the standardized exercise test used to characterize and select animals for selective breeding at 11 wk of age. HCR were able to run for a longer time (HCR vs. LCR: 66.7 ± 0.9 vs. 14.1 ± 0.7 min; P < 0.001) and at a higher speed (43.0 ± 0.5 vs. 16.6 ± 0.4 m/min; P < 0.001) resulting in a higher distance run (1.750 ± 0.039 vs. 0.186 ± 0.013 km; P < 0.001). Sedentary and exercised animals did not differ in their exercise capacity before treatment started.

Table 1 shows exercise-induced improvement in exercise capacity in HCR and LCR. Both HCR and LCR were able to increase running speed and the distance run. Thus, total exercise capacity (aerobic and anaerobic capacity) significantly increased with exercise training. Both the (relative) increase in speed and in distance run were higher in LCR than in HCR indicating a stronger effect in LCR.

Table 2 shows echocardiographic parameters of HCR and LCR in sedentary and exercise-trained animals. In sedentary LCR, higher posterior wall diameter may indicate hypertrophy and increased E/e′ indicated disturbed diastolic function. Both were no longer present in exercise-trained animals. There were no effects of exercise on systolic function in HCR or LCR. Supplemental Table S1 (see https://doi.org/10.6084/m9.figshare.14958012) shows echocardiographic data of exercise-trained animals before and after 4 wk of exercise training.

Table 2.

Echocardiographic parameters of sedentary and exercise-trained HCR and LCR

	Sedentary		Exercise
	HCR	LCR	HCR	LCR	Effect of Intrinsic Exercise Capacity	Effect of Acquired Exercise Capacity	Interaction
HR	228 ± 10	236 ± 7	235 ± 10	243 ± 7	ns	ns	ns
LVPWD, mm	1.56 ± 0.06	1.76 ± 0.05b	1.64 ± 0.04	1.55 ± 0.05a	ns	ns	0.008
LVEDD, mm	7.57 ± 0.18	7.67 ± 0.14	7.63 ± 0.24	8.02 ± 0.18	ns	ns	ns
EF, %	60.3 ± 2.2	63.9 ± 1.9	60.28 ± 2.2	61.7 ± 0.9	ns	ns	ns
FS, %	33.9 ± 1.7	36.4 ± 1.4	33.6 ± 1.5	34.7 ± 0.6	ns	ns	ns
E/e′	15.1 ± 0.4	17.8 ± 1.1b	15.0 ± 1.3	14.3 ± 0.6a	ns	0.041	ns
E/A	1.90 ± 0.05	1.82 ± 0.10	1.84 ± 0.12	2.08 ± 0.16	ns	ns	0.050

Values are means ± SE; n = 8–11/group. E/A ratio with E, early diastolic wave; A, late diastolic wave; EF, ejection fraction; FS, fractional shortening; HCR, high-capacity runners; HR, heart rate; LCR, low-capacity runners; LVEDD, left ventricular end diastolic diameter; LVPWD, left ventricular posterior wall thickness; ns, not significant. ^asignificantly different vs. sedentary animals of the same line; ^bsignificantly different vs. HCR with the same treatment; E/e′ was calculated as an index of preload.

Table 3 shows morphological data at the time of hyperinsulinemic-euglycemic clamp (after 4 wk of treatment). At all times, LCR presented with higher body weight than HCR. Sedentary animals of both groups increased body weight during this period. Exercise instead led to a reduction in body weight in both HCR and LCR (data not shown). The reduction of body weight during exercise was significantly greater in HCR. LCR further presented with greater body size as indicated by larger tibias and greater weights of most organs compared with HCR. Relative weight of both gastrocnemius and soleus muscle was not affected by exercise. The reduction in fat pad liver and kidney weight in both lines indicated that exercise training led to weight loss. The loss of fat in these organs may be the reason for body weight reduction with exercise.

Table 3.

Morphological data of HCR and LCR at the time of euthanasia 24 h after the last exercise session or in sedentary animals determined during hyperinsulinemic-euglycemic clamp

	Sedentary		Exercise		Effect of Intrinsic Exercise Capacity	Effect of Acquired Exercise Capacity	Interaction between Intrinsic + Extrinsic Capacity
	HCR	LCR	HCR	LCR
Animal age, mo	8.82 ± 0.20	10.03 ± 0.16 b	7.99 ± 0.24a	8.89 ± 0.23ab	<0.001	<0.001	ns
Body weight, g	379 ± 19	523 ± 16b	313 ± 9a	441 ± 10ab	<0.001	<0.001	ns
Body weight change with 20 days of exercise resp. sedentary time, g	18.3 ± 2.2	15.7 ± 3.9	−38.0 ± 7.0a	−7.4 ± 7.4ab	0.03	<0.001	0.012
Body weight change with 20 days of exercise resp. sedentary time, %	4.9 ± 0.6	3.2 ± 0.8	−10.3 ± 1.8a	−1.5 ± 1.6ab	0.018	<0.001	<0.001
Tibia length, mm	41.6 ± 0.5	44.0 ± 0.3b	40.8 ± 0.4	43.0 ± 0.3ab	<0.001	0.015	ns
Heart weight/TL, g/m	27.5 ± 1.1	28.8 ± 1.1	26.2 ± 1.0	28.3 ± 1.02	ns	ns	ns
Ventricle weight/TL, g/m	24.5 ± 1.0	25.9 ± 1.0	22.8 ± 0.6	24.6 ± 0.6	ns	ns	ns
Lung weight/TL, g/m	36.3 ± 1.4	33.6 ± 1.2	39.3 ± 3.0	38.4 ± 1.6	ns	0.046	ns
Liver weight/TL, g/m	216 ± 11	256 ± 11 b	187 ± 7a	214 ± 7 ab	<0.001	<0.001	ns
Pancreas weight/TL, g/m	16.6 ± 1.2	15.4 ± 0.8	16.1 ± 1.1	16.7 ± 1.2	ns	ns	ns
Spleen weight/TL, g/m	17.1 ± 1.1	16.2 ± 0.6	12.0 ± 0.9a	14.7 ± 0.5 b	nd	nd	0.027
Kidney weight/TL, g/m	64.2 ± 3.7	67.8 ± 2.5	55.5 ± 2.0a	58.6 ± 0.9a	ns	<0.001	ns
Epididymal fat pad weight/TL, g/m	110 ± 10	260 ± 24b	54.1 ± 5.5a	181 ± 19ab	<0.001	<0.001	ns
Gastrocnemius weight/TL, g/m	118 ± 6	127 ± 4	104 ± 4a	121 ± 2b	0.002	0.020	ns
Soleus weight/TL, g/m	8.49 ± 0.52	9.66 ± 0.33	8.55 ± 0.48	8.53 ± 0.31	ns	ns	ns
Brain weight/TL, g/m	42.7 ± 1.4	46.6 ± 0.5 b	43.6 ± 0.7	44.8 ± 1.1	0.015	ns	ns

Values are means ± SE; n = 8–10/group. HCR, high-capacity runners; LCR, low-capacity runners; nd, not determined; ns, nonsignificant; TL, tibia length. ^asignificantly different vs. sedentary animals of the same line; ^bsignificantly different vs. HCR with the same treatment.

Table 4 shows fasting and random glucose levels as indicators for whole body glucose metabolism and insulin sensitivity. Random and fasting glucose levels were always higher in LCR than HCR. During hyperinsulinemic-euglycemic clamp, HCR and LCR did not differ in their basal glucose turnover and this was not affected by exercise. Insulin-stimulated glycolysis rates were significantly higher in HCR than in LCR. Exercise had no effect on whole body glycolysis in HCR and LCR and the difference between both remained. Similarly, insulin-stimulated glucose turnover was higher in HCR than in LCR. However, the difference between HCR and LCR persisted (Table 4).

Table 4.

Blood glucose levels, glucose turnover, and glycolysis rates as indicators for whole body glucose metabolism and insulin sensitivity

	Sedentary		Exercise
	HCR	LCR	HCR	LCR	Effect of Intrinsic Exercise Capacity	Effect of Acquired Exercise Capacity	Interaction between Intrinsic + Extrinsic Capacity
Random blood glucose level, mmol/L	4.21 ± 0.14	4.69 ± 0.13b	3.98 ± 0.14	4.63 ± 0.13b	<0.001	ns	ns
Fasting blood glucose level, mmol/L	3.43 ± 0.14	3.84 ± 0.14b	3.80 ± 0.14	4.01 ± 0.13	0.03	ns	ns
Basal glucose turnover, mg/kg/min	2.99 ± 1.03	2.11 ± 0.97	1.95 ± 1.03	1.66 ± 0.86	ns	ns	ns
Insulin-stimulated glucose turnover, mg/kg/min	8.44 ± 0.78	4.79 ± 0.73b	6.64 ± 0.78	5.76 ± 0.65	0.004	ns	ns
Insulin-stimulated glycolysis, mg/kg/min	13.3 ± 1.6	7.0 ± 1.6b	11.0 ± 1.5	6.5 ± 1.4b	0.002	ns	ns

Values are means ± SE; n = 8–10/group. HCR, high-capacity runners; LCR, low-capacity runners; ns, not significant. ^bSignificantly different vs. HCR with the same treatment.

Similarly, we found significantly higher insulin-stimulated whole body glucose disposal rates (Fig. 1A) in HCR than in LCR independent of exercise training as determined from glucose infusion. Exercise did not affect glucose disposal rates in HCR and LCR. However, insulin-induced changes in hepatic glucose metabolism were significantly more pronounced in HCR than in LCR (Fig. 1B). Furthermore, in the liver, exercise led to an increase in the response to insulin in HCR, whereas LCR did not show any changes indicating a stronger insulin response in HCR.

Figure 1.

Insulin-stimulated whole body glucose disposal rates (A) and change in hepatic glucose production upon stimulation with insulin (B). Insulin-stimulated 2-deoxyglucose uptake of musculus gastrocnemius (C), epididymal fat tissue (D), brain (E), and heart (F) in HCR and LCR with and without exercise training in sedentary (white bars) and exercise-trained (black bars) animals. Values are means ± SE, n = 8–10 animals. *P < 0.05, **P < 0.01, for treatment (T), population (P), or interaction of both (I), as indicated. HCR, high-capacity runners; LCR, low-capacity runners; ns, nonsignificant.

Figure 1 also shows 2-deoxyglucose (DG) uptake of M. gastrocnemius, epididymal fat, brain, and heart. Deoxyglucose uptake of the other harvested organs is shown in Table 5. DG uptake was not different between sedentary HCR and LCR in the major insulin responsive organs. However, in HCR, exercise led to increased insulin-stimulated 2-DG uptake in gastrocnemius (Fig. 1C), fat (Fig. 1D), and brain (Fig. 1E). In contrast, such an increase was not seen in LCR where only minor changes could be observed. In the heart, exercise led to a reduction in glucose uptake in LCR only (Fig. 1F). These results indicate that exercise induced an increased response to insulin stimulation in HCR compared with LCR.

Table 5.

Deoxyglucose uptake of selected organs during hyperinsulinemic-euglycemic clamp in sedentary and exercise-trained HCR and LCR

Uptake, µmol/min/g	Sedentary		Exercise
	HCR	LCR	HCR	LCR	Group Effect	Treatment Effect
Lung	1.21 ± 0.15	1.40 ± 0.06	0.68 ± 0.12a	0.77 ± 0.08a	ns	<0.001
Liver	2.03 ± 0.31	1.88 ± 0.08	1.16 ± 0.05a	1.97 ± 0.17b	ns	0.024
Spleen	0.36 ± 0.17	0.90 ± 0.09b	0.38 ± 0.09	0.36 ± 0.08a	0.021	0.021
Pancreas	0.95 ± 0.14	0.95 ± 0.11	1.20 ± 0.14	0.84 ± 0.09	ns	ns
Kidney	22.13 ± 8.25	18.74 ± 3.26	13.46 ± 3.66	17.87 ± 3.11	ns	ns
Soleus	1.62 ± 0.21	1.71 ± 0.18	1.82 ± 0.20	1.72 ± 0.17	ns	ns

Values are means ± SE; n = 6–9/group. HCR, high-capacity runners; LCR, low-capacity runners. ^asignificantly different vs. sedentary animals of the same line; ^bsignificantly different vs. HCR with the same treatment.

Figure 2 shows the effect of exercise training on mitochondrial citrate synthase activity in heart (Fig. 2A), M. gastrocnemius (Fig. 2C), and epididymal fat (Fig. 2B). Citrate synthase activity in heart was comparable in HCR and LCR and did not change with exercise. In epididymal fat and in gastrocnemius muscle, there was again no difference between sedentary HCR and LCR. However, exercise increased citrate synthase activity both in skeletal muscle and fat in HCR only, indicating differences in the response to exercise between HCR and LCR.

Figure 2.

Citrate synthase activity in heart (A), epididymal fat (B), and gastrocnemius (C) in HCR and LCR with and without exercise training in sedentary (white bars) and exercise-trained (black bars) animals. Values are means ± SE, n = 8–10. *P < 0.05, **P < 0.01, ***P < 0.001 for treatment (T), population (P), or interaction of both (I), as indicated; A: no differences were observed. HCR, high-capacity runners; LCR, low-capacity runners; ns, nonsignificant.

In Figure 3, we further analyzed Glut4 expression and localization as potential cause for changes in substrate uptake in skeletal muscle and heart after insulin stimulation during hyperinsulinemic-euglycemic clamp. Glut4 amounts were lower after chronic aerobic interval training in skeletal muscle (Fig. 3A) but not in hearts of HCR and LCR (Fig. 3B). Glut4 translocation was not different between HCR and LCR in skeletal muscle (Fig. 3C) or heart (Fig. 3D), independent of exercise. Thus, Glut4 expression and localization in cardiac and skeletal muscle may not explain changes in glucose metabolism observed.

Figure 3.

Glut4 expression (A and B) and localization (C and D) (plasma membrane vs. vesicles) in skeletal muscle (SM) M. gastrocnemius (A and C) and heart muscle (B and D) in HCR and LCR with and without exercise training in sedentary (white bars) and exercise-trained (black bars) animals. Values are means ± SE, n = 8–10 animals. *P < 0.05, **P < 0.01, for treatment (T), population (P), or interaction of both (I), as indicated; B–D: no differences were observed. HCR, high-capacity runners; LCR, low-capacity runners; ns, nonsignificant.

DISCUSSION

We demonstrate in our model of HCR and LCR that genetic predisposition for low exercise capacity is associated with impaired insulin sensitivity and impedes exercise-induced improvements in insulin response. Our results suggest that genetic predisposition for low aerobic exercise capacity impairs insulin response, which may not be overcome by exercise.

Long-term exercise in patients with diabetes has been shown to improve insulin resistance and may lower HbA1c levels (27). Other reports describe improved whole body glucose disposal and insulin-stimulated skeletal muscle glucose transport in humans as well as in rats (28). Exercise improves whole body insulin sensitivity in insulin-resistant subjects. Furthermore, exercise increases aerobic exercise capacity in patients with heart failure (17). As a consequence, exercise training has been included in guidelines for the treatment of type 2 diabetes (29). However, exercise may also fail to affect glucose metabolism. For instance, HbA1c or HOMA-IR in patients with reduced insulin sensitivity or metabolic syndrome was unchanged with exercise in some studies (30, 31). The potential reasons for such discrepancies are a matter of debate (32, 33). In this investigation, exercise training improved HCRs insulin response but could not elicit the same response in LCR. These results indicate that low genetically determined aerobic exercise capacity may not only be associated with reduced insulin response and the metabolic syndrome, but at the same time may lead to a reduced metabolic response to exercise. Our results are supported by the finding of differences in exercise effects in different mouse strains (34). Genetic factors may thus potentially also be responsible for a lack of exercise effects on glucose metabolism found in several investigations. Aerobic exercise training has been also found ineffective in patients with diabetes (35). Furthermore, exercise training in genetically predisposed populations seems to be frequently ineffective as has been shown for African, Arabic, Chinese, or Polynesian subjects (36). Thus, genetic predetermination appears to be important not only for total exercise capacity but also for the potential therapeutic response of training.

Exercise training can be applied in different intensities and ineffective exercise has been suggested as a cause for lack of exercise effects (37). In humans, protocols related to the current capacity of the individual (relative training) are preferred to absolute protocols where each individual receives the same amount of training independent of capacities (23, 38). Both in patients as well as in the animal model used in this investigation, a relative training protocol seems indispensable to avoid inadequate exercise training (overtraining in LCR or no training in HCR). We observed adverse effects with exercise before (39). Therefore, we tested for exercise capacity according to Hoydal et al. (22) in both lines investigated in this study. Although we were not able to assess V̇O_2max, we used the approximation suggested (22). We applied an aerobic interval training which has been shown to be most efficient although the assessment of each animal allowed for individually adapted exercise training (23). Our results indicate proper adaptation for each animal for the following reasons: most animals were able to perform exercise sessions and could follow the increased intensity of exercise every week which was omitted in case an animal was not able to complete at least three of five training sessions during 1 wk. The increase was determined from previous studies conducted by Wang et al. (40) in animals of the same line. Furthermore, both HCR and LCR improved their distance run during exercise and their performance during each training session (see also Table 1). Exercise effects on cardiac morphology and function were only minor and no indication of functional impairment with exercise training was found. Finally, exercise training led to slight to moderate reduction in body weight in both HCR and LCR without any indication for adverse effects. Thus, it appears safe to conclude that our exercise protocol was effective and did not lead to “overtraining.”

The effect of exercise on body weight has been suggested to mediate improvements in insulin sensitivity. This has been a matter of debate (12). However, it has been shown that even moderate exercise without weight loss is able to improve insulin sensitivity in patients (41). In our animals, we observed some weight loss in both lines. Thus, it appears safe to assume that differences in weight loss may not to be responsible for the observed differences in response to exercise. Improvement of glucose metabolism in LCR after exercise training was relatively small in our investigation compared with other studies. In other studies, exercise training has been performed in both HCR and LCR and effects of exercise reported in LCR only (42, 43). However, in contrast to our investigation, in these studies all animals were subjected to the same absolute amount of exercise, thus running the same distance. It may be possible that the exercise protocol has been too mild for effects on HCR. In our study, we used individual adaptation to prevent potentially inadequate protocols. Every animal was tested for its individual running capacity and exercise training has been adjusted (22, 23). We used a protocol of high-intensity interval training which has been described as the most efficient method to improve exercise capacity (22, 44). Comparable protocols are used in human athletes to maximize improvements. In our investigation, both HCR and LCR lost weight, decreased epididymal fat pad and liver weight. Our results thus indicate that our exercise protocol has been high-intensity training for both HCR and LCR.

Low exercise capacity is a predictor for the development of the metabolic syndrome and diabetes (45). Individuals with low exercise capacity have an increased risk of total and cardiovascular mortality. Furthermore, low exercise capacity is related to increased morbidity and mortality (46). Exercise capacity is in part predetermined by genetic predisposition (intrinsic) and may be improved by exercise training [acquired (13)]. These two components of exercise capacity cannot be separated in humans due to the high variation in environmental conditions and their contribution to morbidity and mortality seems unclear. Our animal model of high and low exercise capacity provides the ability to assess the influence of genetic predisposition on insulin sensitivity (15, 16); the genetic component originates on selection pressure and random genetic drift during breeding. We show that intrinsic genetic predisposition for low aerobic exercise capacity leads to impaired glucose metabolism in these rats including higher fasting and random glucose levels. Using the hyperinsulinemic-euglycemic clamp for the first time in these animals, lower glucose disposal, lower insulin-stimulated glucose turnover, and a reduced hepatic insulin response were found in LCR. There is currently no other outbred model of genetic predisposition for exercise capacity. Our results indicate that intrinsic exercise capacity seems responsible for a major part of differences in morbidity.

Exercise has been suggested to increase aerobic metabolism and fatty acid oxidation (47). We did not measure fatty acid metabolism in HCR and LCR in this investigation. However, we observed a reduction in body weight after 4 wk of exercise (Table 3) and reduced epididymal fat pad and liver weights in exercise-trained animals which all indicate increased use of fatty acids. In sedentary animals as well as during exercise, HCR have been found to rely more on fatty acid oxidation and less on glycogen use compared with LCR (48). Skeletal muscle has been found to consist of more oxidative fibers in M. gastrocnemius in HCR compared with LCR or in similar fiber type contribution in M. soleus and EDL (49) indicating higher reliance on oxidative metabolism in HCR compared with LCR. These findings by others support our results of reduced glucose metabolism and reduced response to exercise in LCR.

Exercise has been suggested to promote mitochondrial biogenesis and function (50). However, there have been conflicting results, which have been suggested to be caused by differences in intensity and duration of exercise (39). In a model of insulin resistance induced by HFD feeding in mice, exercise led to increased mitochondrial biogenesis (51). In humans, exercise has been found to increase mitochondrial citrate synthase activity in skeletal muscle of healthy controls but not in patients with diabetes (52). Our results seem to be compatible with the human data of increased citrate synthase activity in skeletal muscle. However, to our knowledge, there have been no reports investigating the effect of exercise training on mitochondrial activity depending on genetic predisposition. Our finding of increased citrate synthase activity with exercise in adipose tissue and skeletal muscle in HCR but not LCR suggests that genetic predisposition has a strong impact on mitochondrial response to exercise. This is supported by the finding of higher expression of mitochondrial biogenesis-related genes in skeletal muscle of HCR compared with LCR (53). Insulin sensitivity, mitochondrial biogenesis, and function are closely interrelated. Thus, mitochondria may represent a possible mechanistic link between genetic predisposition, impaired insulin sensitivity, and a reduced response to exercise.

Limitations

The model of high and low intrinsic exercise capacity allows discriminating the influence of both on insulin sensitivity. However, in humans, differences in intrinsic exercise capacity are much less pronounced and our results seem thus not directly transferable to humans but may help to better understand why aerobic exercise training in patients with diabetes often is found not (or only minimally) to have the desired insulin-sensitizing effect. We performed exercise training in relation to current capacity. Although this avoids overtraining and insufficient stimuli, the animals ran different distances. Furthermore, we determined proteins using Western blot after performing the hyperinsulinemic-euglycemic clamp. Thus, the extended time of insulin stimulation and anesthesia may have influenced our results regarding protein expression and localization. In our investigation, we used only male animals at about 9 mo of age. We had to limit ourselves to one sex only, as females and males are highly different in their exercise capacities. Thus, we are not able to ensure that the results are valid for females as well as for young or old animals. However, we estimate the effects of genetic predisposition for high exercise capacity to be of comparable direction independent of age. Thus, the small differences in age of our animals may have contributed to differences but should not have led to effects higher than normal biological variation as all animals were adult but not aged.

We demonstrate in our HCR and LCR models that genetic predisposition for low exercise capacity resulting from selection pressure impairs insulin sensitivity and impedes exercise-induced improvements. This may explain findings of previously controversial results in patients indicating that genetic background may impair improvements in some patients or populations with diabetes.

SUPPLEMENTAL DATA

Supplemental Table S1: https://doi.org/10.6084/m9.figshare.14958012.

GRANTS

The LCR-HCR rat model system is funded by the Office of Research Infrastructure Programs Grant P40OD021331 (to L.G.K. and S.L.B.) from the National Institutes of Health (Bethesda, MD). This study was supported by grants of the Deutsche Forschungsgemeinschaft (DFG) to T.D. (DO 602/8-1).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

M.S. conceived and designed research; A.M., C.S., and A.S. performed experiments; A.M., C.S., and A.S. analyzed data; M.S. and A.M. interpreted results of experiments; M.S. prepared figures; M.S. drafted manuscript; M.S., S.L.B., L.G.K., and T.D. edited and revised manuscript; M.S., A.M., C.S., A.S., S.L.B., L.G.K., and T.D. approved final version of manuscript.