The effects of long-term moderate exercise and Western-type diet on oxidative/nitrosative stress, serum lipids and cytokines in female Sprague Dawley rats

Cardiovascular disease (CVD) is a major cause of morbidity and mortality in Western societies and developing countries [1]. Modifiable risk factors, such as obesity and sedentarism are highly prevalent in patients with CVD, and both can be improved by safe and effective lifestyle interventions [2]. Such lifestyle factors are diet and exercise that affect the concentration of low-density lipoproteins (LDL), which are known metabolic driver of CVD, such as atherosclerosis [3‐5]. In addition, it is well established that elevated concentrations of LDL cholesterol (LDL-C) promote atherosclerosis and increase the risk of non-fatal and lethal CVD events. LDL-C is subjected to oxidation, resulting in the formation of oxidized LDL (oxLDL), which is thought to be more aggressive than non-oxidized LDL in inducing atherogenesis [6, 7].

Elevated oxidative stress characterized by an increased generation of radical oxygen species (ROS) is a common manifestation in patients with CVD, obesity, and diabetes mellitus. These highly reactive compounds can modify many biomarkers including LDL-C [8‐10]. The modification of LDL-C by oxidation promotes the receptor-mediated uptake of oxLDL by macrophages, thereby causing cholesterol accumulation in the vasculature. Previous studies in humans and animals indicate that Western diet and physical inactivity induce free radical production, which may enhance the susceptibility of LDL to oxidation [3, 4, 11]. One of the principal pathways involved in the adaptation to exercise is nitric oxide (NO) signalling. NO triggers vasodilation and, thus, can mitigate the high shear stress during exercise. Furthermore, NO is involved in several other physiological and pathological processes, such as cell inflammation and adhesion as well as angiogenesis [12‐15]. NO is synthetised by three isoforms of the enzyme NO synthase (NOS), including neuronal (nNOS or NOS-1), inducible (iNOS or NOS-2), and endothelial isoform (eNOS or NOS-3). All of them are constitutively but not exclusively expressed in the cardiovascular system [10]. In particular, eNOS requires dimerization to maintain its normal function but under certain conditions the reduction of molecular oxygen by eNOS is not coupled anymore with the oxidation of the substrate L-arginine (L-arg), thus resulting in the production of superoxide instead of NO [10]. The main eNOS uncoupling motifs are the cofactor of NO synthesis, tetrahydrobiopterin (BH4), reduced bioavailability of L-arg, or high levels of the endogenous NOS inhibitor, asymmetrical dimethyl-arginine (ADMA) [8].

Although regular exercise is proven to reduce CVD risk and promote health benefits [12, 16‐19], constant physical activity also increases the production of free radicals and oxidative stress [4, 9]. This raises the question as to whether exercise-induced oxidative stress is beneficial or detrimental. In this regard, obese and normal-weight human adults have reportedly comparable oxLDL concentrations [20], whereas other studies showed that body weight reduction after the bariatric surgery or regular moderate exercise decrease oxLDL [21, 22]. In contrast, a single intensive exercise session appears to increase the susceptibility of LDL and other lipoproteins to oxidation in healthy adults [3, 4]. Thus far, a single in vivo study explored the effects of HFD and regular physical activity on plasma lipids, oxidative/nitrosative stress and LDL oxidation. Specifically, Elmas et al. showed that rats consuming a HFD exhibit increased oxidative stress in aortic and myocardial tissue [2]. Regular exercise appears to modify the balance of antioxidants and oxidants as well as NO metabolism in these tissues. However, it remains elusive whether constant moderate exercise is able to prevent or attenuate the oxidative/nitrosative stress, inflammation, and serum lipids in lean and obese rats.

In the present study, we investigated the impact of HFD and regular moderate exercise on the oxidative/nitrosative stress, serum lipids and cytokines. For this purpose, blood samples from young and old Sprague Dawley female rats that were subjected to forced treadmill exercise sessions for 10 months in combination with a HFD or ND were analysed.

The present study shows that long-term moderate exercise reduces the body weight gain and NOx but increases oxLDL in normal-weight and obese rats. Administering HFD to sedentary animals resulted in a marked increase in body weight gain and triglycerides; however, it failed to systematically alter oxLDL, NOx or any other lipid profile parameter.

In the SD rats used in this project, long-term moderate running exercise did not change TC, HDL-C, and non-HDL-C. In HFD fed animals, TG was slightly reduced by exercise. Furthermore, exercise increased oxLDL in both dietary groups. These results are in contrast to most previous studies in humans and rodents where regular moderate endurance exercise reduced TG, TC and LDL-C, whereas HDL-C increased [2, 26, 37‐39]. However, the results of individual studies may vary substantially [4, 37, 40] and comparability between studies is limited by different experimental settings. For example, Elmas et al. analysed the effects of a 6-week exercise intervention in SD rats where animals were forced to swim for 1 h/d on 5 days/week [2]. Animals that received a high-fat diet showed a decrease in TC and TG when compared to sedentary controls. In our study, the administration of a HFD increased TG, but not TC. Furthermore, exercise did not show a modulatory effect on TG, TC, or non-HDL-C. When compared to the present results, the lipid concentrations reported by Elmas et al. are quite different. These contrasting results may be explained the use of male animals that were considerably younger at the time of sacrification. Furthermore, the present study used a HFD with a different lipid composition and a less vigorous exercise intervention. Analytical differences may also have contributed to the different results. In our laboratory, all methods are strictly controlled by internal and external quality controls. Furthermore, the lipid concentrations that we obtained in young and old control animals are in line with a previous study in SD rats [11].

The HFD used in this study is particularly rich in SFA and MUFA and thus induces an imbalance of lipoprotein metabolism with consequent alterations of plasma lipoproteins. According to the manufacturer, this diet aims to mimic the situation in humans who follow a typical Western-type dietary regimen and live in an obesogenic environment. Furthermore, it reliably induces obesity and metabolic syndrome in mice and rats. Differences in the dietary lipid composition may explain differences in plasma lipoproteins and oxidative–nitrosative stress markers observed in other HFD intervention studies in rodents [2, 23, 41, 42]. The current National Cholesterol Education Program (NCEP) and American Heart Association (AHA) dietary guidelines recommend limiting fat intake to 30–40% of the total dietary calories [43]. However, increasing fat intake inside the recommended range may already have adverse effects on the lipid profile with increasing TC and LDL-C concentrations [43]. In contrast, reducing fat intake to 20% or less can also be troublesome due to a reduction of LDL-C and HDL-C and a contemporary rise in TG. This combination is typically associated with the formation of small and dense LDL particles with a high atherogenic potential [43]. With the aim to promote a healthier LDL/HDL ratio, the original AHA Step I fat recommendation advices for a 1:1:1 proportion of SFA: MUFA: PUFA in the diet.

In old animals the lipid profile was substantially different from that of young animals with markedly higher TC and TG concentrations. Moreover, HDL-C, non-HDL-C and NEFA were higher in old than in young animals. Such age-related changes of the lipid profile are expected and have been described by others before [44]. With the HFD used in this study, the concentrations of TG and NEFA increased, whereas all other parameters of the lipid profile remained unchanged. Previous studies that treated rodents with HFD reported mixed results. While some studies found increasing concentrations of TC, TG and LDL-C [2, 25, 41, 44] others did not [11]. However, a direct comparison of these studies is limited due to differences in study design and composition of the diets. For example, in 4–8-week-old Wistar rats the administration of HFD for 4–8 weeks resulted in increased body weight and adipose tissue weight, TC, TG, LDL-C concentrations [23, 42]. However, both studies did not include baseline measurements, which impedes a longitudinal evaluation of age-related effects. Zelzer et al. treated adult female SD rats for 12 weeks with a HFD comparable to the one used in the present study [11]. This intervention did not result in different TC or TG concentrations. Also, HDL-C and non-HDL-C were comparable between controls and HFD-treated animals. The limited comparability of different animal models is not surprising and has already been shown before [44]. Similar to animal models, also human studies that compared the lipid profile of obese and non-obese individuals yielded heterogeneous results [20, 21, 45]. Most existing studies showed higher TC, TG, LDL-C, and small dense LDL-C (sdLDL-C) concentrations in obese individuals when compared to normal weight controls. However, for HDL-C, inconsistent results have been reported [21, 45]. A controlled dietary intervention study by Egert et al. demonstrated that the substitution of a high-fat diet rich in saturated fatty acids with either a high-fat or a low-fat diet rich in mono-unsaturated fatty acids ameliorated the lipid profile, reducing TC, LDL-C/HDL-C ratio, LDL-C size and its susceptibility to oxidation [46].

A main finding of our study is the exercise-induced increase in oxLDL that is accompanied by a reduction of NO. Elevated oxidative stress is a common condition in sedentary and obese individuals that increases the generation of ROS and leads to the modification of many biochemical targets [8‐10] including LDL-C. OxLDL is supposed to be more aggressive than non-oxidized LDL in driving inflammation, atherogenesis, and ultimately the risk of CVD events [5‐7]. Nevertheless, exercise-induced ROS production seems to protect cells against oxidation by maintaining the cellular oxidant-antioxidant homeostasis [47, 48]. Furthermore, regular physical activity improves blood pressure control through an increased production of NO and other vasoactive substances [10]. In the present study, however, long-term moderate running exercise was associated with an increase in oxLDL and a reduction in NOx. While the increase in oxLDL was comparable in both dietary groups, the reduction in NOx was more pronounced in HFD animals. These results are in agreement with previous studies demonstrating an increased susceptibility of LDL-C and other lipoproteins to oxidation after a single intensive exercise session [3, 4]. Furthermore, in another study by Zelzer et al., the administration of a similar HFD to rats increased several markers of oxidative stress, such as malondialdehyde, but not oxLDL [11]. While Zelzer et al. measured oxLDL by ELISA, the other studies analysed lipoprotein oxidation indirectly [3, 4], which limits the comparability of results. When interpreting the modification of oxidative stress biomarkers by physical activity, it should be considered that chronic exercise as well as high-intensity training can increase oxidative stress through several mechanisms including increased mitochondrial oxygen consumption and activation of oxidase enzymes such as nicotinamide adenine dinucleotide phosphate (NADPH) oxidase, a major source of ROS [9]. This might at least partially explain the exercise-induced increase in oxLDL that we observed in the present study. Additional support for this hypothesis comes from the reduction in NOx by regular exercise. Under normal circumstances, the potent vasodilator NO is released during the conversion of L-arginine to L-citrulline. This reaction is catalysed by NOS. However, under certain circumstances, NOS can also produce superoxide (O2*⁻), which reacts avidly with vascular NO^· to form peroxynitrite (ONOO⁻). This metabolite is capable of impairing NOS dimerization and function [8, 49], which is called NOS-uncoupling. Such an uncoupling of NOS can occur in the absence of either L-arg or BH4, and increased concentrations of ADMA, a competitive inhibitor of NOS. In the present study, HFD and regular exercise increased h-arg, another substrate of NOS for the production of NO. Although h-arg competes with L-arg for NOS-binding sites, it seems to be a less efficient substrate for NO synthesis. According to März et al., the role of h-arg in NO metabolism is still insufficiently understood [50]. Some studies support the hypothesis that h-arg might increase arginine bioavailability by inhibition of the enzyme arginase, which competes with NOS for the utilization of the key substrate L-arg [50, 51]. If this is correct, the substantial increase in h-arg is likely to exert a protective effect under HFD. This concept is supported by a previous study showing that h-arg supplementation ameliorates blood glucose in mice on HFD [52]. The reduction of ADMA and SDMA observed in the present study further supports the beneficial effect of h-arg in HFD animals. However, h-arg competes with L-arg in more than one way. They both utilize the same transport system for cell entry, and high extracellular h-arg concentrations will result in reduced L-arg uptake [53]. This might at least partially explain the exercise-induced reduction of NOx. The inverse relationship between oxLDL and NOx strongly supports the concept of NOS uncoupling in exercising animals. Considering that NOx levels were lower in non-exercising HFD animals than in non-exercising ND animals, HFD consumption seems to have an independent NO-reducing effect, which is amplified by regular exercise. The very low NOx concentration in exercising HFD animals further supports the concept of an additive NO-reducing effect of HFD and exercise.

In this project, we also investigated the influence of regular exercise on the adipokine, cytokine and chemokine profile in both dietary groups. Obesity is typically associated with an increased secretion of adipokines and a mild tissue inflammation [36]. As expected, our HFD animals showed markedly higher leptin and adiponectin serum concentrations than ND animals regardless of exercise, whereas IGF-1 and resistin were comparable in both dietary groups. Insulin was not considered as blood was collected non-fasting. Exercise was effective in reducing leptin in ND, but not HFD, animals suggesting that the HFD overwhelmed the effect of exercise. The mild increase in serum adiponectin in exeHFD animals might simply reflect the weight loss in these animals. Interestingly, the HFD used in this study had only minor pro-inflammatory effects inducing increased serum concentrations only for MCP-1 and IL-5. Other pro-inflammatory cytokines, such as TNF-α, IFN-γ or IL-6 were comparable between the two dietary groups.

Consistent with previous studies [26, 54], our exercise protocol significantly reduced the serum concentrations of TNF-α, IL-1β, IL-2, MCP-1, MCP-3 and RANTES in HFD animals. MCP-3 and RANTES were also decreased in the exeND group. Both chemokines regulate the migration and infiltration of monocytes and macrophage into solid tissues [33]. The present results are in line with recent findings from Rocha-Rodrigues et al. showing that regular physical activity reduces inflammation in response to HFD administration [26]. Borst et al. have shown that in the context of obesity, visceral fat derived resistin, TNF-α and several other interleukins contribute to insulin resistance [35], whereas weight loss or visceral fat removal decrease serum IL-6 and increase the insulin sensitizing hormone adiponectin [34, 35]. The exercise-induced increase of IP-10, IL-12, and MMP activity in ND animals might reflect a mild activation of cellular immunity and tissue remodelling in response to exercise, which is masked in the HFD animals.

The present study has several strengths and limitations that should be considered when interpreting the results. The rather long intervention period with sufficiently sized groups allows robust conclusions about the effects of the different diets and physical activity as important modifiable lifestyle factors that impact CVD risk. Furthermore, our exercise protocol was well controlled and imitated a realistic activity regimen that would normally be considered as healthy. Another strength of this study is the use of validated and strictly quality controlled methods for the measurement of serum lipids. The unlimited access of the animals to food allowed them to compensate the exercise-related increase in energy expenditure through higher food consumption. Due to the unlimited access to food, we cannot account for natural occurring differences in food intake. Furthermore, we lack metabolic studies and blood collections were performed in a non-fasting state. Although this may have an impact on the results of several metabolic biomarkers, we intentionally decided to perform non-fasting blood collections to avoid unwanted psychological stress in the animals. It should also be mentioned that our method for the measurement of TG does not distinguish between TG and free glycerol. However, the blood concentration of free glycerol is < 1 mg/dL, which accounts for approximately 10 mg/dl of triglycerides [55]. Considering that in the SD rats used in this study TG concentrations ranged between 100 and 400 mg/dl, the unintentional detection of free glycerol does not represent a relevant confounder of our results.

Another weakness of our study is the lack of mechanistic information on the effects of exercise on plasma lipids. Considering that body weight was strongly correlated with TG, it can be speculated that the higher fatty acid intake in HFD animals, was the main driver of differences in plasma TG. The HFD used in this study contains approximately 60% of SFA and MUFA, which are incorporated into TG. Exercise instead, induced only minor differences in body weight and plasma lipids. This is not surprising, as our exercise regimen was rather moderate. Previous studies have shown that exercise induces beneficial effects on plasma lipids and lipoprotein lipase activity only above a certain threshold of energy expenditure [56, 57]. Another putative mediator of plasma lipids could be adiponectin, which is known to increase energy expenditure through fatty acid oxidation in target organs, such as liver and skeletal muscle [58]. This theory is supported by the inverse correlations of TC and HDL-C with adiponectin.

In summary, long-term moderate exercise may alter the delicate equilibrium between oxidative and anti-oxidative compounds leading to an uncoupling of NOS with higher oxLDL and lower NOx concentrations.

However, these metabolic effects do not necessarily compromise the beneficial reduction of pro-inflammatory cytokines, such as TNF-α, IL-1β, IL-2, MCP-1, MCP-3 and RANTES. This conundrum adds to the controversial role of oxLDL in pathologic developments like atherosclerosis and CVD and should trigger additional research that helps to understand lipoprotein oxidation and NO production in response to exercise and different dietary habits.