Cardiovascular Risk Factors Emerge After Artificial Selection for Low Aerobic Capacity

See allHide authors and affiliations

Science  21 Jan 2005:
Vol. 307, Issue 5708, pp. 418-420
DOI: 10.1126/science.1108177


In humans, the strong statistical association between fitness and survival suggests a link between impaired oxygen metabolism and disease. We hypothesized that artificial selection of rats based on low and high intrinsic exercise capacity would yield models that also contrast for disease risk. After 11 generations, rats with low aerobic capacity scored high on cardiovascular risk factors that constitute the metabolic syndrome. The decrease in aerobic capacity was associated with decreases in the amounts of transcription factors required for mitochondrial biogenesis and in the amounts of oxidative enzymes in skeletal muscle. Impairment of mitochondrial function may link reduced fitness to cardiovascular and metabolic disease.

Several investigations link aerobic metabolism to the pathogenesis of cardiovascular disease. Large-scale epidemiological studies of subjects with and without cardiovascular disease demonstrate that low aerobic exercise capacity is a stronger predictor of mortality than other established risk factors (14). In patients with type 2 diabetes, low aerobic capacity is associated with reduced expression of genes involved in oxidative phosphorylation (5). In insulin-resistant elders, there is a 40% reduction in mitochondrial oxidative and phosphorylation activity, largely attributable to impaired skeletal muscle glucose metabolism (6). These observations are consistent with impaired regulation of mitochondrial function as an important mechanism for low aerobic capacity and cardiovascular risk factors linked to the metabolic syndrome. These risk factors include weight gain, high blood pressure, reduced endothelial function, hyperinsulinemia, and increased triglyceride concentration in blood. The working hypothesis of the present study was that rats selected on the basis of low versus high intrinsic exercise performance would also differ in maximal oxygen uptake, mitochondrial oxidative pathways, and cardiovascular risk factors linked to the metabolic syndrome.

In previous work, we began large-scale artificial selection for low and high aerobic treadmill-running capacity with the genetically heterogeneous N:NIH stock of rats as the founder population (7). Eleven generations of selection produced low-capacity runners (LCRs) and high-capacity runners (HCRs) that differed in running capacity by 347% (Fig. 1A). The founder population had a capacity to run for 355 ± 144 m (23.1 min) until exhausted. On average, the treadmill-running capacity decreased 16 m per generation in LCRs and increased 41 m per generation in HCRs in response to selection. At generation 11, the LCRs averaged 191 ± 70 m (14.3 min), and the HCRs ran for 853 ± 315 m (41.6 min). For this study, we used young adult rats (ages 16 to 24 weeks) derived from generations 10 and 11 to test our hypothesis that risk factors for common diseases segregate with variation in intrinsic aerobic capacity (8).

Fig. 1.

Eleven generations of selective breeding in rats resulted in two divergent strains. Values are mean ± 1 SD. (A) Response to selection for aerobic treadmill-running capacity across 11 generations (n = 2912 rats). On average, the LCR rats decreased 16 m per generation and the HCR rats gained 42 m per generation in distance run to exhaustion. (B) Mean blood pressures from sedentary female rats recorded via telemetry. For both day and night values, the LCR rats (n = 10) recorded significantly higher pressures than did the HCR rats (n = 10). (C) Maximal acetylcholine-induced vasorelaxation was lower (left) and the EC50 for acetylcholine was 7.8-fold greater (right) in ring segments from the common carotid arteries of LCR (n = 6) relative to HCR (n = 6) female rats. 100% dilatation corresponds to maximal vessel dilatation, achieved by adding exogenous NO in the form of sodium nitroprusside (SNP, fig. S1). Asterisks indicate P < 0.01 for HCRs compared with LCRs. For more details see fig. S1.

High blood pressure is associated with increased risk for stroke and ischemic heart disease (9). We found that, relative to the HCRs, the LCR rats had higher mean blood pressures during the day (105 ± 13 mm Hg compared with 89 ± 8 mm Hg), at night (98 ± 3 mm Hg compared with 91 ± 7 mm Hg), and for the combined 24-hour period (102 ± 6 mm Hg compared with 90 ± 7 mm Hg) (Fig. 1B). Extrapolating from human data (9), this 13% higher 24-hour blood pressure suggests that the LCRs are twice as likely to develop cardiovascular disease as the HCRs.

Endothelial dysfunction is an independent predictor of long-term cardiovascular disease progression and cardiovascular event rates (10). To assess endothelial function in the two strains of rats, we assayed nitric oxide–mediated (acetylcholine) vascular relaxation in isolated ring segments of carotid arteries. In this assay, higher vessel relaxation is interpreted as better endothelial function. For maximal absolute relaxation, the HCR rats demonstrated a 48% increase compared with the LCR rats. Furthermore, the concentration of acetylcholine that provoked a half-maximal response [median effective concentration (EC50)] was 7.8-fold greater in LCR than HCR rats (Fig. 1C and fig. S1).

LCR rats were insulin-resistant compared with the HCR rats, as demonstrated by higher fasting insulin levels and impaired glucose tolerance (Table 1 and fig. S2). Insulin C-peptide levels were normal in LCR rats, indicating that insulin secretion was preserved. However, insulin clearance was reduced in the LCR rats, as indicated by lower steady-state C-peptide/insulin molar ratios. These data indicate that hyperinsulinemia results mainly from reduced insulin clearance. Consistent with the clinical scenario of the metabolic syndrome, the LCR rats also had more visceral adiposity, higher plasma triglycerides, and elevated plasma free fatty acids compared with the HCR rats (Table 1).

Table 1.

LCR and HCR rats differed significantly for carbohydrate and lipid metabolic measures. Measurements were taken from male LCR (n = 8) and HCR (n = 8) rats. Blood was drawn at 0900 hours with food and water ad libitum to measure random blood sugar. Other metabolic measures were made on blood drawn after 12 hours of food and water deprivation.

LCR HCR % Difference LCR vs. HCR P value
Random glucose (mg/dl) 86 ± 6 75 ± 12 15% 0.036
Fasting glucose (mg/dl) 110 ± 9 92 ± 5 20% 0.0007
Insulin (pM) 684 ± 195 296 ± 172 131% 0.002
C-peptide (pM) 1590 ± 338 1077 ± 565 48% 0.061
C-peptide/insulin 2.4 ± 0.4 3.8 ± 1.2 -58% 0.013
Visceral adiposity/body weight (%) 1.55 ± 0.39 0.95 ± 0.32 63% 0.005
Triglycerides (mg/dl) 67 ± 24 25 ± 4 168% 0.013
Free fatty acids (meq/l) 0.64 ± 0.22 0.33 ± 0.04 94% 0.031

Because individuals with cardiovascular disease often show diminished capacity for adaptation to exercise training (11), we measured 12 variables to assess the general exercise capacity and left ventricular function both in sedentary control (C) and in exercise-trained (T) LCR and HCR rats (Table 2). Each rat was trained for 6 weeks on a treadmill at an intensity relative to its own individual maximal oxygen consumption (VO2max) (12). Consistent with a low tolerance for exercise, the C-LCR rats had a 58% lower VO2max, a 17% lower economy of running (i.e., higher oxygen cost of running), 23% less left ventricular weight, and a trend (P = 0.07) toward shorter left ventricular cell length compared with the C-HCR rats. Isolated left ventricular cells from C-HCR rats had better systolic and diastolic function relative to the C-LCR rats (Table 2). In response to training, both T-LCR and T-HCR rats showed significant improvement in all 12 of the measures of capacity (Table 2), with a uniformly greater training response in the T-HCR relative to the T-LCR rats for each measure except cell width.

Table 2.

Exercise capacity and isolated left ventricular cell variables for LCR and HCR rats separated in groups of sedentary control (C) and exercise-trained (T). Before exercise, the C-LCR and C-HCR rats differed significantly (indicated by asterisks for P < 0.01) for all variables except left ventricular cell length and width. Six weeks of exercise training significantly improved each of these 11 variables in both T-LCR and T-HCR rats (indicated by ‡ for P < 0.01). In each case except cell width, T-HCR rats improved more than T-LCR rats with training († for P < 0.05). PS, percentage cell shortening. Values are means ± 1 SD from six LCR and six HCR female rats.

C-LCR C-HCR % Difference % Change with training
Whole animal variables
VO2max (ml kg-0.75 min-1) 43 ± 2 68 ± 3 -58%* 38%‡ 44%‡†
Economy of running (ml O2 kg-0.75 m-1) 4.9 ± 0.1 4.2 ± 0.2 17%* -7%‡ -17%‡†
Left ventricular weight (mg kg-0.75) 1561 ± 176 1917 ± 88 -23%* 22%‡ 27%‡†
Left ventricular cell variables
Cell length (μm) 118 ± 2 124 ± 2 -5% 6%‡ 14%‡†
Cell width (μm) 23 ± 3 19 ± 3 20% 2% 2%
Systolic cell function
Cell shortening (%) 14.0 ± 1.2 17.1 ± 1.1 -22%* 30%‡ 39%‡†
Relative time to peak shortening (ms PS-1) 2.7 ± 0.2 2.3 ± 0.2 17%* -24%‡ -32%‡†
Systolic [Ca2+] (μM) 1.61 ± 0.03 1.73 ± 0.04 -7%* -16%‡ -23%‡†
Amplitude of [Ca2+] transient (μM) 1.20 ± 0.03 1.38 ± 0.05 -15%* -19%‡ -24%‡†
Diastolic cell function
Time to 50% relengthening (ms) 39.9 ± 1.2 35.2 ± 1.3 13%* -14%‡ -16%‡†
Diastolic [Ca2+] (μM) 0.41 ± 0.02 0.35 ± 0.02 17%* -7%‡ -20%‡†
Time to 50% decay of [Ca2+] transient (ms) 55.3 ± 1.4 45.9 ± 1.3 20%* -11%‡ -13%‡†

Mitochondrial dysfunction is associated with a wide range of human diseases (5). In view of the lower aerobic capacity and reduced cardiovascular function of LCR rats, we hypothesized that they have compromised mitochondrial oxidative function relative to the HCR rats. To test this hypothesis, we measured the cellular content of proteins required for mitochondrial biogenesis and function (5, 13) in soleus muscle, which is composed largely of highly oxidative fibers. The amounts of peroxisome proliferative activated receptor γ (PPAR-γ), PPAR-γ coactivator 1 α (PGC-1α), ubiquinol-cytochrome c oxidoreductase core 2 subunit (UQCRC2), cytochrome c oxidase subunit I (COXI), uncoupling protein 2 (UCP2), and ATP synthase H+-transporting mitochondrial F1 complex (F1-ATP synthase) were markedly reduced in the LCR rats in comparison with the HCRs. The uniform decline in these proteins is consistent with the hypothesis that reduced aerobic metabolism plays a causal role in the development of the differences between the LCR and HCR rats (Fig. 2). PGC-1α, particularly because it interacts with PPAR-γ, seems to be centrally positioned for influencing both energy metabolism and the progression of complex diseases. PGC-1α is a transcriptional coactivator involved in energy transfer pathways and mitochondrial biogenesis and permits PPAR-γ to interact with many transcription factors (14). PPAR-γ, a regulator of adipocyte differentiation, has been implicated in the pathology of numerous diseases including obesity and diabetes. Thiazolidinediones are selective ligands of PPAR-γ and effective for the treatment of type 2 diabetes, suggesting a pivotal role for PPAR-γ in complex diseases (15).

Fig. 2.

Proteins known to be integral for mitochondrial function and hypothesized to be inversely associated with diseases were more abundant in soleus skeletal muscles from HCR male rats relative to LCR male rats (n = 6 of each strain). Asterisks indicate P < 0.01 for HCRs compared with LCRs. Bars show mean ± 1 SD.

Body weight can have a substantial influence on both aerobic running capacity and the emergence of disease (16). Eleven generations of selective breeding for running capacity produced a correlated change in body weight. By generation 11, male LCR rats weighed 92 g more (39%) than HCR males, and similarly the LCR female rats weighed 44 g more (24%) than HCR females (fig. S3). Multiple regression analysis using weight and generation as predictors of running capacity revealed that changes in body weight explained 7% of the variation in distance run in HCR females, 7% in LCR females, 20% in HCR males, and 14% in LCR males. Thus, factors other than body weight account for the majority of the variation in distance run across the 11 generations of selection in both strains.

Because risk factors for complex diseases often emerge with aging (17), we measured indices of metabolic risk in 5-week-old male pups (fig. S4). At this age, the HCR and LCR lines had essentially identical body and visceral fat weights, with a 25% greater VO2max in the HCR relative to the LCR. The LCR pups showed 12% higher plasma glucose (P < 0.001) and plasma triglyceride values (P < 0.04) compared with the HCR pups. Thus, in our contrasting strains, metabolic changes preceded the increase in body weight (fig. S4), a result that is consistent with a role for hyperinsulinemia in weight gain. Although mechanistic arguments have been put forward for either pattern in humans, clinical studies have not resolved whether obesity precedes or follows the development of insulin resistance in type 2 diabetes (18).

In summary, the present study demonstrated that selection for low versus high intrinsic aerobic exercise capacity simultaneously generated a differential load of metabolic and cardiovascular risk factors. Rats with low aerobic capacity expressed low amounts of key proteins required for mitochondrial function in skeletal muscle, suggesting a mechanistic association. Although a direct cause-effect relationship has not been proven, our observations support the notion that impaired regulation of oxidative pathways in mitochondria may be a common factor linking reduced total-body aerobic capacity to cardiovascular and metabolic disease. This is in concert with previous epidemiological and clinical studies (16, 19).

Supporting Online Material

Materials and Methods

Figs. S1 to S4

References and Notes

References and Notes

View Abstract

Stay Connected to Science

Navigate This Article