Peripheral artery occlusive disease (PAOD) is a major health problem with limited treatment options affecting 8-12 million Americans (1, 2). The pathophysiology of PAOD is characterized by an impaired perfusion to the lower extremities. Exercise training is one treatment option that improves quality of life and tissue perfusion characteristics, by inducing increased capillary density (3), increasing the surface area for exchange of oxygen and other substances between capillaries and muscle fibers, thereby limiting ischemic symptoms (4). Exercise training has been shown to increase capillary number within healthy skeletal muscle through the process of angiogenesis, mediated, at least in part, by vascular endothelial growth factor (VEGF) driven pathways (5–8). The mechanisms by which exercise training improves intermittent claudication in PAOD patients remains unclear (2, 9, 10). Clinical studies of PAOD indicate that patients limited by intermittent claudication who engage in any amount of weekly physical activity beyond light intensity, have a lower mortality rate than their sedentary counterparts who perform either no physical activity or only light-intensity activities (11), even after adjusting for other predictors of mortality, which include age, and body mass index (BMI). Thus, exercise training appears to induce adaptive remodeling within ischemic skeletal muscle, but the pathways and mechanisms responsible for this remodeling are still not well-characterized.

Exercise is a complex stimulus with multifactorial outcomes. These outcomes are determined by at least two genetic components (12–17). One component establishes the innate aerobic exercise capacity that each individual possesses irrespective of any subsequent training or activity (12–17), and the second genetic component establishes the pattern of responses to active exercise training (1, 2, 11, 18–20). It is estimated that up to 70% of the variation in exercise capacity are due to the genetic predisposition of an individual (12, 13). Many studies have focused on the active exercise training component (1, 2, 6, 7, 11, 18–21), but even within these studies, there is a heterogeneous response in subjects to training protocols (2, 6, 7, 11, 18–22). It is this finding that suggests the importance of investigating intrinsic exercise capacity and its impact on the response to ischemic stress.

Active exercise training in healthy individuals increases both perfusion and energy demand in the working muscle, although under strenuous exercise, the increase in demand is thought to exceed the increase in perfusion. This relative ischemic stress generates a remodeling stimulus for angiogenesis, but there must be limits because progressive loss of perfusion with advancing occlusive disease also induces a relative ischemia lower work intensity, but does not generate a comparable expansion of the vascular bed to offset it. The process of vascular remodeling is multifactorial, with vascular endothelial growth factor (VEGF) and its receptor family including VEGFR-1 [Fms-like tyrosine kinase 1 (FLT-1)], VEGFR-2 [Kinase insert domain-containing receptor (KDR) also known as fetal liver kinase (FLK-1)], and neuropilin-1 (NP-1) all participating (6, 7, 19). VEGFR-2 appears to respond following exercise training as well as with ischemic challenge independent of any active exercise (4, 19, 20). Other potential angiogenic factors that may influence intrinsic exercise capacity associated responses to ischemic challenge are angiopoietin 1 and 2 (Ang1 and Ang2) and nitric oxide (NO) (8, 18, 20, 23). Both factors are known to be involved in exercise training induced angiogenesis and ischemia induced angiogenesis, as well as the combination of both stressors. The angiopoietins are important cytokines that assist in vascular development and remodeling (24). Ang 1 promotes maturation and stabilization of vessels and is expressed widely throughout tissues, whereas Ang 2 competes with Ang 1 by displacing it from the activating Tie2 receptor. The Tie2 receptor is expressed at sites of vascular remodeling. Thus Ang1 dominance is associated with a stable vasculature while Ang2 dominance is associated with active angiogenesis (24). NO can be increased with VEGFR2 or KDR activation (24). The increase in NO production can be critical to VEGF signaling and the remodeling response of the skeletal muscle. The influence of intrinsic exercise capacity on the angiogenic response following an ischemic event has not been studied.

Low aerobic capacity or low cardiovascular fitness, is a strong predictor of early mortality within PAOD patients (2, 3, 9, 11). Koch and Britton (14–17), developed a rat model to investigate the role of intrinsic aerobic exercise capacity on responses and adaptations to chronic disease. Studies with this model show that the low endurance running capacity phenotype (LCR) rats are more prone to develop hepatic steatosis (25) and more sensitive to high fat diet-induced insulin resistance (26). These LCR rats displayed a higher incidence of both cardiovascular and metabolic syndrome risk factors than the high endurance running capacity (HCR) rats (27). Previous studies have found that the LCR phenotype has decreased capillary density when compared to the HCR counterparts (14–17, 27), which suggest that the LCR may have a decreased tolerance in hind limb ischemia models of PAOD. Studies of skeletal muscle responses to the stress of high fat feeding indicate that while the LCR's are less tolerant, the inducible responses are slower in onset but largely intact (26). However, while resting differences between the phenotypes are suggestive, it remains unknown whether the induced responses of peripheral ischemic challenge are altered by the intrinsic aerobic phenotype. Therefore, this study was designed to test the hypothesis that an increased intrinsic aerobic capacity will manifest more compensation in response to peripheral ischemia using a an established unilateral hind limb femoral artery occlusion (19, 20).

To provide a functional measurement for comparison between the phenotypes, high frequency electrical stimulation of the gastrocnemius muscle was performed prior and following ligation in two separate experiments. Prior to ligation, there were not significant differences between the phenotypes in the time to exhaustion with high frequency pacing [LCR time (s) = 2,291 ± 186 vs. HCR time (s) = 2,571 ± 218]. Following ligation, both strains decreased the time to muscle exhaustion with pacing, consistent with having induced an injury, but there was a significant difference between the strains (LCR = 437 ± 47 vs. HCR = 824 ± 56). The LCR gastrocnemius muscle time to exhaustion was significantly less than the HCR (Figure 1A), suggesting greater functional impairment in the LCR as a result of the occlusion.

Similar to the findings with pacing, there also were not differences in resting perfusion (LCR number of microspheres/g tissue = 13,424 ± 688 vs. HCR number of microspheres/g tissue = 12,563 ± 823). Not surprisingly, ligation of the femoral artery significantly decreased perfusion in both the LCR (6,383 ± 366) and HCR (7,821 ± 424) strains, but the decrease in perfusion was more severe in the LCR phenotype both immediately after ligation and 7 days after ligation (LCR = 6,880 ± 513 vs. HCR 9,522 ± 721) (Figure 1B). These data indicate that while both phenotypes decreased tissue perfusion in response to the hind limb ligature, the extent of perfusion deficit, and the resulting pattern of recovery over the next seven days was significantly different as a function of the phenotype.

mRNA expression for angiogenic growth factors was assessed with quantitative real time PCR (qRTPCR). VEGF expression was significantly lower in the LCR phenotype under baseline (pre-ischemic conditions (Figure 2A). Neither the LCR nor the HCR animals demonstrated significantly altered VEGF mRNA expression following 7D of ischemic challenge and baseline differences disappeared between the phenotypes at this time point. 14 days after ligation, the LCR significantly increased VEGF mRNA expression in the gastrocnemius muscle compared to both control (^*Figure 2A) and 7D ischemic levels (^$Figure 2A). Moreover, VEGF mRNA expression also was significantly greater in the LCRs than the HCRs 14 days after occlusion (^#Figure 2A).

In contrast to VEGF, neither VEGF receptor (Flt-1) mRNA nor VEGF receptor (KDR) mRNA expression was significantly different between the phenotypes under baseline, pre-ischemic conditions. Neither the LCR nor the HCR animals demonstrated significantly altered Flt-1 or KDR mRNA expression following 7D of ischemic challenge. Similar to VEGF, Flt-1 mRNA expression increased significantly in the LCR ischemic muscle at 14 days, compared to either control (^*Figure 2B) or 7D ischemic levels (^$Figure 2B) and was also significantly greater than expression in the HCRs at 14 days post ischemia (^#Figure 2B). However, these changes exhibited in Flt-1 were not in KDR mRNA expression (Figure 2C).

Angiopoietin 1 mRNA expression was significantly higher in the LCR phenotypes under baseline conditions (^#Figure 2D) but neither phenotype significantly altered the expression of this growth factor and the differences present at baseline were no longer present at 7D or 14D.

Angiopoietin 2 mRNA expression was significantly higher in the HCR phenotype under baseline conditions (^#Figure 2E), and did not change in this phenotype at either 7 or 14 days after occlusion. In contrast, the LCR phenotype showed significant increases in Ang 2 expression 14 days after occlusion (^*,$Figure 2E), abolishing any differences that were present between the phenotypes before occlusion.

eNOS mRNA expression was significantly lower in the LCR phenotype under baseline conditions (^#Figure 2F). Neither the LCR nor the HCR animals demonstrated significantly altered eNOS mRNA expression following 7D of ischemic challenge and baseline differences disappeared between the phenotypes at this time point. 14 days after ligation, the LCR significantly increased eNOS mRNA expression in the gastrocnemius muscle compared to both control (^*Figure 2F) and 7D ischemic levels (^$Figure 3). Moreover, eNOS mRNA expression also was significantly greater in the LCRs than the HCRs 14 days after occlusion (^#Figure 2F).

Histology data indicate that the LCR has reduced capillary density (CD) (LCR = 417.81 ± 23.6) compared to the HCR in non-ischemic gastrocnemius muscle (HCR = 614.6 ± 57.5) (^#p < 0.05, Table 2), consistent with previous reports (3, 28). The HCR phenotype increased capillary density at both 7 (918.54 ± 197.6) and 14 days (1,015.41 ± 174.9) after occlusion (^*Table 2). In contrast, capillary density in LCRs was significantly lower at all time points (LCR 7 Day = 564.76 ± 40.5, LCR 14 Day = 507.48 ± 54.2) (^#Table 2), and in contrast to HCRs, capillary density did not change in LCRs after occlusion. There were no differences in capillary contacts per fiber (NCAF) between phenotypes under control conditions (LCR = 4.547 ± 0.220 vs. HCR = 4.280 ± 0.217), but both phenotypes increased NCAF significantly in response to ischemia. However, the increase in the HCRs was earlier in onset, present at both 7 (5.170 ± 0.276) and 14 days after occlusion (5.210 ± 0.255) (^*Table 2), while the LCR was significantly elevated only at the 14 day endpoint (6.220 ± 0.481) (^*Table 2). At 14 days, the NCAF values had become significantly greater in the LCR compared to the HCR animals (^#Table 2).

Similar to the NCAF results, the capillary to fiber ratio on the individual fiber basis (Cap/Fib) were not different under baseline conditions (LCR = 1.662 ± 0.089 vs. HCR = 1.555 ± 0.073), but increased in both phenotypes in response to occlusion. Again, the increase in the HCR phenotype was evident at both the 7 (1.896 ± 0.149) and 14 day (1.852 ± 0.097) time points compared to control (^*Table 2), but a similar increase above control levels was not seen in the LCR animal until 14 days (2.248 ± 0.183) (^*Table 2), at which point the NCAF values were significantly higher in the LCR than the HCR animals (^#Table 2).

The percentage of slow twitch muscle fibers in the mixed gastrocnemius muscles from the LCR and HCR animals was compared (Figure 3). The LCR rats had a significantly lower percentage of slow twitch fibers in the control gastrocnemius muscle compared to the HCRs (^*p < 0.05, Figure 3).

The development of intermittent claudication is a major decrement to the quality of life in patients (1, 2, 31–33). A mounting body of evidence suggests that exercise training can increase perfusion in ischemic tissue, improving quality of life and avoiding the necessity for surgical revascularization and other expensive therapies (1–3, 9, 33, 34). The majority of research on aerobic exercise training in PAOD patients has examined its effects upon vascular remodeling, including the mechanisms underlying changes in capillary density within the ischemic muscle (6, 7, 21, 22). However, many patients with intermittent claudication are unable to exercise or participate in exercise protocols due to pain in the ischemic limb or arthritic problems in joints, and exercise studies do not always provide definitive remodeling within tissue. Developing a better understanding of the genetic influences associated with intrinsic aerobic capacity and endurance capacity on the vascular remodeling process in response to ischemic challenge may help explain the variability in PAOD patients responses to aerobic exercise, and may provide new treatment options for individuals unable to participate in aerobic exercising protocols. The goal of this study was to investigate whether rats with low intrinsic running capacity have different vasculogenic responses following peripheral arterial occlusion than rats with high intrinsic running capacity.

The results of this study demonstrate that rats selectively bred, but not trained, for high endurance running capacity (HCR) differ in their response to ischemic stress in a model of PAOD compared to the low endurance running capacity (LCR) counterparts. These results provide strong evidence indicating that low endurance running capacity confers increased risk for ischemic injury and is subject to delayed and less effective adaptive response to ischemic stress.

LCR and HCR's were inherently different in capillary density before ischemia, with HCR rats having significantly higher capillary density than LCR rats. Under baseline conditions, the HCR strain also showed a higher relative expression of mRNA for VEGF, eNOS, Ang2 and significantly less relative expression of Ang1 compared to the LCR strain. These results are consistent with those found by Lloyd et al. (19), who showed similarly increased amounts of angiogenic markers in response to active aerobic exercise training protocols in rats, except that in the present studies, these were intrinsic differences predicated only on intrinsic aerobic phenotype, and not dependent on any active training component. Similar to the actively exercised rats, the HCRs appear to have an underlying increased angiogenic potential under basal conditions compared to the LCR phenotype.

LCR muscle cells showed a significantly larger area compared to the HCR, a finding consistent with Howlett et al. (35) who found that mean cross-sectional area of the HCR fiber was 35% lower compared to the LCR. The combination of smaller fiber area, more capillaries, and greater oxidative capacity in HCR rats may provide the basis for explaining the inherent higher VO_2max observed in these animals without aerobic exercise training, and would be consistent with increased resistance to ischemic injury.

Increased capillary density (CD) (capillaries/mm²) has been shown to strongly correlate with increased skeletal muscle oxygen conductance in this model (25). We now extend those observations by demonstrating that the HCRs may have inherently better protection for maintenance of tissue perfusion under ischemic conditions. This was functionally evident with the prolonged time to exhaustion with high frequency electrical stimulation, and by higher post-ischemic perfusion levels. Although the initial drop in blood flow after occlusion is greater in the LCRs, consistent with provoking a more potent ischemic stress, the LCRs do not demonstrate a measurable recovery in perfusion at seven days. The HCRs initially lose less perfusion with acute arterial occlusion, presumably associated with better ischemic tolerance, and also recover perfusion in the post-ischemic tissue, consistent with anatomic remodeling, altered arteriolar/collateral tone favoring dilation, or both. This flow preservation, suggesting an improved tolerance for ischemic stress by the HCR, is further supported by the HCR's ability to maintain muscle fiber area (Table 2). Furthermore, despite less ischemic pressure and negligible changes in fiber area, HCR rats showed a significant increase in capillary density at seven days despite negligible changes in the mRNA expression of angiogenic markers suggesting the possibility that the elevated baseline expression levels of the typical angiogenic growth factors in the HCRs were sufficient to drive any required angiogenesis. These might be more analogous to patients who develop claudication, but do not progress to critical limb ischemia.

Conversely, the LCR animals did not generate a measurable change in perfusion following seven days of ischemic pressure. These animals responded to seven days of femoral artery ligation with a significant loss of muscle fiber area consistent with atrophic loss of muscle mass. Perhaps because there was loss of muscle mass, there was a reduced ischemic signal source such that the LCR rats did not alter capillary density (CD), Cap/Fib ratio, or NCAF significantly from baseline levels, and perfusion did not recover. The LCR rats did not show significant changes in mRNA expression for any of the angiogenic factors from the baseline to the 7D ischemic time point, suggesting that the LCR rats initially respond to seven days of ischemia was a loss in fiber area and no angiogenic response, despite a substantial loss in perfusion. These might be more analogous to patients who develop critical limb ischemia.

Following 14 days of femoral artery occlusion, the HCRs continued to maintain fiber area but did not significantly alter any angiogenic markers from seven day levels, suggesting that after 14 days of ischemia, angiogenesis was no longer an active pathway within this phenotype and vascular remodeling processes had been completed. This finding is consistent with Lloyd et al. who reported that ischemia-induced angiogenesis typically alters mRNA within 7-10 days and returns toward baseline levels by about 14 days (19). Of particular interest in the present study was that HCRs appeared to increase the percentage of slow twitch fibers with 14 days of ischemia. A fast to slow fiber switch is normally seen in response to aerobic exercise training and aging (19), but ischemia and unloading is typically associated with a slow to fast phenotypic shift. The dynamic recovery response over a 2 week period in the HCRs appears to involve a shift in fiber type and an increase in fiber numbers, associated with higher pre-existing angiogenic signals and fueled by a better flow recovery. In contrast, after an initial fiber loss at 7 days, muscle fiber area in the LCRs significantly increased again 14 days after femoral artery occlusion, suggesting that there was a change in fiber size, but not fiber number, and an entirely different remodeling response. The increase in muscle fiber size may have been sufficient to recover a relative ischemia signal, as significant increases in mRNA expression for VEGF, Flt-1, eNOS, and Ang2 were observed at 14 days after occlusion. Together, these findings suggest that an angiogenic drive remains inducible in the LCR phenotype, but that the dynamic between muscle mass, muscle number, demand ischemia and vascular remodeling still is incompletely understood. Recent reports in mice (36), and subsequently in humans (37), have suggested a potential role for BAG-3 (Bcl-2-Associated Athanagene-3) coding variants in dictating the response to ischemia in skeletal muscle. This angiogenic stimulus may be in response to signals associated with hypertrophy of the LCR cells, but not with the occlusive stimulus, or may be occurring at 14 days as a synergistic action of the loading stimulus in combination with ischemic drive within the tissue.

The present study suggests that LCR rats have an altered response to hind limb femoral artery occlusion compared to the HCR rats. This altered response is partially due to the phenotypic differences under baseline conditions, including significantly less capillary density, less angiogenic factor potential, larger fiber area, and a smaller percentage of slow twitch muscle fibers within the gastrocnemius muscle in the LCR animals. With onset of occlusion, the LCRs also appear to have an altered inducible response compared to the HCR. These alterations include fiber area initially decreasing, followed by a rebound to a larger area, a potent angiogenic upregulation at 14D ischemia and no significant changes in the percentage of slow twitch muscle fibers. These data suggest that low intrinsic aerobic capacity may have increased injury associated with ischemic disease and will generate an altered response that may require a combination ischemic challenge, and signaling from some other stress, such as increased mechanical loading of the muscle, to adequately stimulate vasculogenic responses.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

The animal study was reviewed and approved by Institutional Animal Care and Use Committee (IACUC) East Carolina University.

EG and RL participated in designing, conducting, data analysis, and writing. MA and MZ participated in data analysis and writing. SB and LGK provided the animals and participated in experimental design. LCK participated in data analysis and manuscript review. All authors contributed to the article and approved the submitted version.

This study was supported by NIH grant number P40OD021331, National Institutes of Health, Office of Research Infrastructure Programs.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.