Exercise and Cardiovascular Disease
A New Perspective
Abstract—The oxidation of low density lipoprotein (LDL) has been suggested as a key event in atherogenesis. Paradoxically, exercise, which imposes an oxidative stress, is an important deterrent of cardiovascular disease. In study 1 the oxidizability of LDL was enhanced in exercisers compared with sedentary controls. The lag time of isolated LDL subjected to copper-induced in vitro oxidation was significantly shortened in the exercisers compared with sedentary subjects. This increased sensitivity was not due to a decreased presence of vitamin E. Instead, these findings suggested that the LDL of exercisers may contain increased amounts of preformed lipid peroxides, which account for the increased oxidizability. In study 2, a group×sex ANOVA revealed that male exercisers had a significantly longer mean lag time than male sedentary subjects and that females had similar mean lag times regardless of exercise group. This remained the case when statistical adjustment was made for age, body mass index, blood lipid levels, LDL, and plasma α-tocopherol levels. Study 1 exercisers had been in training for a shorter time (<1 year) than study 2 exercisers (>2 years). These findings suggest that truly “chronic” exercise (aerobic intensity over several months) decreases the susceptibility of a male exerciser’s LDL to undergo oxidation. Conversely, regular aerobic stress during an overall shorter time span creates a more oxidative environment in the body, thus increasing the susceptibility of LDL to undergo oxidation. The oxidative stress of aerobic exercise does not appear to adversely affect the oxidizability of LDL in women.
- Received December 29, 1997.
- Accepted February 18, 1998.
The oxidation of LDL is known to be involved in atherosclerosis.1 Current medical science considers sustained physical activity and exercise to be deterrents to cardiovascular ailments.2 Yet exercise represents a severe form of oxidative stress. Strenuous physical exercise is associated with a dramatic increase in oxygen uptake by both the whole body and particularly the skeletal muscle.3 The production of ROS is believed to occur, causing a series of biochemical and physiological changes indicative of oxidative stress.4
In general, the body has adequate antioxidant reserves to cope with the production of ROS under physiological conditions.5 However, when ROS production is excessive, a pro-oxidative environment may be created in the body, and with chronic exercise, this may lead to an unbalancing of antioxidative measures. Understanding this paradox will increase our understanding of when and how oxidation may have proatherogenic and antiatherogenic effects and how nutrition may be tailored to the oxidative demands of the exercising population. Study 1 was conducted to examine the oxidizability of LDL isolated from exercisers versus sedentary subjects. Study 2 was conducted to evaluate confounders in study 1.
All subjects were recruited through advertisements and flyers posted and distributed at the Emory University Campus, Atlanta, Ga, and in the surrounding community. Criteria for participation in studies included the following: (1) nonsmokers, (2) non–vitamin/mineral supplement users, (3) no known heart disease, (4) no lipid-modifying medications, (5) no current pregnancy, and (6) aerobically active <1 hour each week (sedentary group) or aerobically active at least 6 hours each week (exercise group). Individuals who participated in the study were volunteers and were not given any kind of remuneration. The protocols for both studies were approved by the human subjects review committee of the Emory University School of Medicine, and informed consent was obtained from all participants.
The susceptibility of LDL, isolated from 8 chronic exercisers and 9 sedentary controls, to in vitro oxidation was determined with the use of a 2.5 μmol/L copper system. Biochemical markers, including plasma lipid levels, plasma and LDL vitamin E levels, plasma TBARS, and plasma and LDL electrophoretic mobility were measured.
The susceptibility of LDL, isolated from 30 chronic exercisers and 33 sedentary controls, to in vitro oxidation was determined with the use of a 5 μmol/L copper oxidation system. Biochemical markers, including plasma lipid levels, plasma and LDL vitamin E levels, LDL fatty acid composition (linoleic, linolenic, and arachidonic), MPO protein levels, plasma TBARS, and plasma and LDL electrophoretic mobility were measured.
Vo2 peak was determined on a Marquette treadmill using a continuous, progressive protocol.6 Intensity of exercise was assessed by continuous monitoring of heart rates (polar heart rate monitors) and self-reported ratings of perceived exertion7 every minute during the test. Criteria for test termination included subject volition, attainment of 95% of age-predicted maximal heart rate, or attainment of a respiratory exchange ratio ≥1.1.
Blood (15 mL) was drawn from subjects from their forearm vein after an overnight fast immediately before and after an acute bout of exercise on a treadmill. Sodium heparin Vacutainer tubes were used and immediately placed on ice. An aliquot of plasma was frozen at −80°C, and an additional aliquot was kept on ice until further assays were performed that day.
Lipid Analysis, Isolation, and Oxidation of LDL
Fasting plasma total cholesterol, triglycerides, HDL cholesterol, and LDL cholesterol measurements were determined by using the Cholestech L*D*X analyzer (Cholestech Corp). LDL was isolated from heparinized plasma with a Beckman TL-100 table-top ultracentrifuge.8 LDL samples were subjected to copper-mediated oxidation, and the formation of conjugated dienes was followed.8 TBARS were measured in the plasma.9
α-Tocopherol levels in plasma and LDL were analyzed by high-performance liquid chromatography, and the eluant was monitored with a UV detector at 292 nm.10
Fatty Acid Determinations
The polyunsaturated fatty acid esters linoleic, linolenic, and arachidonic in LDL were analyzed by high-performance liquid chromatography on a Rabbit high-performance liquid chromatography system (Rainin), and the eluant was monitored on a UV detector at 192 nm.11
Determination of Plasma MPO
An ELISA was used to measure plasma MPO levels (Bioxytech Oxis International, Inc).
Differences between subject groups were assessed by Student’s unpaired t test. ANCOVA was used in study 2 to describe the relationship between lag times and exercise group×sex, after controlling for the effect of age in each model and for other covariates one at a time. All analyses were performed using SPSS-PC version 7.5 (SPSS Inc). Statistical significance was set at P<0.05. Values are expressed as mean±SD.
Characteristics of the study groups are shown in Table 1⇓. Exercisers and sedentary subjects were significantly different with regard to BMI, Vo2 peak, total cholesterol, LDL cholesterol, and triglycerides. Specifically, sedentary subjects had a more atherogenic lipid profile and a lower aerobic capacity than the exercisers. There were no significant differences in regard to age or HDL cholesterol between the 2 groups.
The mean lag times for both groups are shown in Table 2⇓. LDL isolated from the exercise group was significantly more readily oxidized compared with LDL isolated from the sedentary group when subjected to oxidation by 2.5 μmol/L copper. The α-tocopherol and TBARS levels of exercisers and sedentary subjects are also shown in Table 2⇓. There were no significant differences between the 2 groups in regard to LDL or plasma α-tocopherol. There was no difference in plasma lipid peroxide levels between exercisers and sedentary subjects as measured by TBARS. It should be noted that TBARS determined on samples oxidized in vitro may not accurately represent the events in vivo. The relative mobility of LDL isolated from the 2 study groups was not different by agarose gel electrophoresis.
A confounding variable in this study is sex. The exercise group was predominantly male and the sedentary group predominantly female. This may suggest that females rather than sedentary subjects have a decreased susceptibility of their LDL to undergo oxidation. Another difference that might affect lag time directly or via sex differences is LDL fatty acid composition. Study 2 was designed to investigate the impact of sex and LDL fatty acid composition on lag time in chronic exercisers and sedentary controls.
Characteristics of the study groups are shown in Table 3⇓. Thirty chronic exercisers (15 female, 15 male) and 33 sedentary controls (21 female, 12 male) participated in this experiment. There were limited LDL samples for fatty acid analysis. Because of this situation, a subset of the study subjects was used in the model for estimating marginal mean lag times when adjusting for this variable.
Initially, Student’s t test was performed to compare the mean of subject characteristic variables between the 2 groups. Exercisers and sedentary subjects were significantly different in regard to self-reported aerobic activity, Vo2 peak, BMI, total cholesterol, LDL cholesterol, HDL cholesterol, and triglycerides. Specifically, sedentary subjects had a more atherogenic lipid profile and a lower aerobic capacity than the exercisers. There was no significant difference in age between the 2 groups.
The mean lag time for both groups is shown in Table 4⇓. Lag time values obtained from LDL isolated from the exercise group were, on average, longer than those obtained from LDL isolated from the sedentary group when subjected to oxidation by 5 μmol/L copper. However, the difference between the 2 groups was not statistically significant. The post hoc power for lag times by group was 0.65. These findings contradicted those of study 1, wherein the exercise group had significantly shorter lag times than the sedentary group. In an attempt to explain this difference, biomarkers (α-tocopherol, polyunsaturated fatty acid composition of LDL, plasma MPO, and plasma lipid peroxide levels) were compared between the 2 groups by Student’s t test. Biomarker results are presented in Table 4⇓.
α-Tocopherol measured in the plasma of exercisers versus sedentary subjects was not significantly different. α-Tocopherol measured in the LDL of exercisers versus sedentary subjects was not significantly different; however, the mean level was higher in the exercise group. There was a statistically significant difference in plasma MPO protein levels between the 2 groups. There was no significant difference between the exercise group (n=17) compared with the sedentary group (n=16) with regard to LDL composition of linoleic, linolenic, or arachidonic fatty acid; however, the mean levels were higher in the exercise group. There was no difference in plasma lipid peroxide levels between exercisers and sedentary subjects as measured by TBARS. Additionally, the relative mobility of LDL isolated from the 2 study groups was not different as determined by agarose gel electrophoresis.
Next, sex differences for subject characteristics, mean lag time values, and mean biomarker values were compared between the exercise group and the sedentary group. Simple analyses with Student’s t test are shown in Tables 5⇓ and 6⇓, respectively. Female exercisers and female sedentary subjects were significantly different in regard to self-reported aerobic activity, Vo2 peak, BMI, total cholesterol, LDL cholesterol, and HDL cholesterol. Female sedentary subjects had a more atherogenic lipid profile and a lower aerobic capacity than the female exercisers. There was no significant difference in age between the 2 groups. While there was no significant difference in triglycerides between the 2 groups, the exercisers had, on average, lower values than those of sedentary subjects.
Similarly, male exercisers and male sedentary subjects were significantly different in regard to self-reported aerobic activity, Vo2 peak, BMI, total cholesterol, and LDL cholesterol. Male sedentary subjects had a more atherogenic lipid profile and a lower aerobic capacity than the male exercisers. There was no significant difference in age between the 2 groups. There was no statistically significant difference in HDL or triglyceride levels between the two groups; however, exercisers had, on average, lower triglycerides and higher HDL values than sedentary subjects.
When subjected to oxidation by 5 μmol/L copper, lag time values obtained from LDL isolated from the female exercise subjects were, on average, shorter than those obtained from LDL isolated from the female sedentary subjects (Table 6⇑). This difference was not statistically significantly different. Conversely, when subjected to oxidation by 5 μmol/L copper, lag time values obtained from LDL isolated from the male exercise subjects were, on average, longer than those obtained from LDL isolated from the male sedentary subjects (Table 6⇑). This was a statistically significant difference. Sedentary males, on average, had shorter lag times than male exercisers, female exercisers, and sedentary females.
There was no statistically significant difference in plasma α-tocopherol; LDL α-tocopherol; LDL composition of linoleic, linolenic, or arachidonic fatty acid; plasma MPO protein; or plasma TBARS between female exercisers and female sedentary subjects (Table 6⇑). However, there was a significant increase in MPO levels in male exercisers. Nor was there a significant difference in plasma α-tocopherol; LDL α-tocopherol; LDL composition of linoleic, linolenic, or arachidonic fatty acid; or plasma TBARS between male exercisers and male sedentary subjects (Table 6⇑).
Group×sex ANCOVAs were performed using age as a covariate in each model along with 1 other biomarker. Covarying the BMI, plasma α-tocopherol, LDL α-tocopherol, plasma total cholesterol, LDL cholesterol, HDL cholesterol, and triglyceride levels revealed no significant differences in lag times for sex or group but a significant interaction effect. Covarying the MPO protein revealed a significant difference in lag times for group and a significant interaction effect. Post hoc power ranged from 0.518 to 0.663. Estimated marginal means of lag times for these models are shown in Table 7⇓. Tests of simple effects revealed statistically significant differences in lag times between the exercising males and sedentary males (P<0.05) in these models. There were no significant differences between sedentary females and exercising females.
These studies address a paradox. Oxidative stress has been implicated in the pathogenesis of atherosclerosis.1 Paradoxically, exercise, which has received widespread acclaim and recommendation as a deterrent to cardiovascular disease, increases oxygen consumption4 and promotes the depletion of plasma antioxidants.5 This study explored the hypothesis that LDL isolated from chronic exercisers would have increased susceptibility to oxidation (compared with sedentary subjects) because of the oxidative environment in which it resides.
Findings of Study 1
In study 1 the chronic exercisers and sedentary subjects were significantly different in regard to fitness and lipid profile parameters. The sedentary subjects had a lower aerobic capacity, higher body-fat composition, and a more atherogenic lipid profile than the exercisers. The lag time of isolated LDL subjected to in vitro copper oxidation was significantly shortened in the chronic exercisers compared with that of the sedentary subjects. This increased sensitivity was not due to the decreased presence of vitamin E, since the amounts of plasma and LDL vitamin E were not different between the 2 groups. Instead, these findings suggest that the LDL of exercisers contained increased amounts of preformed lipid peroxides, which would account for the increased susceptibility to oxidation. It is possible that plasma lipoproteins contain trace amounts of lipid peroxides.12 13 Lipid peroxides could be formed by endogenous lipid peroxidation reactions and then transferred to LDL.
This study is the first to report the enhanced oxidizability of LDL from exercisers and suggests that if it were to occur in vivo, the location of the modification, in the artery wall or in the plasma, would be important to ascertain. The enhanced rate of oxidation of LDL isolated from chronic exercisers might suggest an ongoing oxidative clearance of LDL in exercisers if the LDL were oxidized in the plasma. Oxidized LDL is cleared rapidly from the circulation by the liver compared with native LDL.14 15 In exercisers, the oxidation of LDL in the plasma might itself account for some of the lipid-lowering effects of exercise and might actually be beneficial. Exercise is known to increase the degranulation of neutrophils and subsequently increase the plasma levels of MPO.16 Wieland et al17 and others18 have reported the oxidation of LDL by neutrophils and MPO. Conversely, if this oxidation occurs in the artery wall with subsequent foam cell formation, supplementation with antioxidants might be warranted to help prevent this phenomenon.
Limitations of this study include the small sample size and the imbalance between the sexes in the 2 groups. The exercise group was predominantly male and the sedentary group predominantly female. This may suggest that females rather than sedentary subjects have a decreased susceptibility of their LDL to undergo oxidation. Also, the fatty acid composition of the LDL, which can also affect lag time, was not measured.
Findings of Study 2
In study 2 both sexes were represented in the 2 study groups, and the sample size was increased. The chronic exercisers and sedentary subjects were significantly different with regard to fitness and lipid profile parameters. The sedentary subjects had a lower aerobic capacity, a higher body-fat composition, and a more atherogenic lipid profile than the exercisers.
The findings of study 2 suggest that chronic exercisers are continually undergoing increased oxidative stress, as indicated by a higher level of MPO protein in their plasma compared with the sedentary subjects. However, exercisers tended to have slightly longer lag times than the sedentary subjects, although this difference was not statistically significant. This finding is in sharp contrast to the findings of study 1.
A review of the literature revealed no studies that tested sex differences with regard to the susceptibility of LDL to undergo in vitro oxidation. Group×sex analysis in this study found that LDL isolated from sedentary males was significantly more susceptible to oxidation than was LDL isolated from exercising males. Both exercising and sedentary females had LDL with a similar susceptibility to oxidation as measured by lag time. This held true when adjustments were made for age, BMI, blood lipids, α-tocopherol levels, and MPO protein levels.
The longer lag time seen in these models of male exercisers may be explained by the increasing evidence that chronic exercise training upregulates antioxidant enzymes in skeletal muscle.19 20 21 22 The similar lag times of women in these models, regardless of exercise status and concomitant oxidative stress, may be a reflection of the apparent cardiovascular protection they are afforded by virtue of their premenopausal status.23 Premenopausal women (estradiol-rich) have a less atherogenic lipid profile than do their postmenopausal (estradiol-poor) counterparts. Moreover, estradiol has been shown to scavenge free radicals and decrease the oxidation of LDL.24 Another potential role for estradiol to influence LDL oxidation is its affect on NO. Estradiol has been shown to induce the production of NO.25
Study 1 Compared With Study 2
Study 1 found that exercisers had a greatly increased susceptibility to copper-induced oxidation of LDL compared with sedentary subjects. Study 2 found that male exercisers had a significantly decreased susceptibility to copper-induced oxidation compared with male sedentary subjects, whereas females, regardless of exercise group, had similar susceptibility. This remained the case even when adjustments were made for fitness and blood lipid characteristics, as well as plasma and LDL α-tocopherol levels.
Both studies addressed the issue of chronic exercise on the susceptibility of LDL to oxidation. In study 1, the exercise group was predominantly male with a significantly shorter lag time compared with the sedentary group, which was predominantly female. In study 2, the male exercisers had a significantly longer lag time compared with that of the male sedentary subjects and a longer lag time, though not significantly different, than the females regardless of exercise group. This discrepancy might be explained by differences in the definition of chronic in the 2 studies.
Study 1 included exercise subjects who had been recruited solely from undergraduate, physical education classes at Emory University. These subjects reported greater >6 hours of aerobic activity each week, but the specific intensity of the exercise and the total length of time these subjects had been training is unknown. Study 2 used exercise subjects who had been recruited from the Atlanta Track Club and the Emory University track team. These subjects also reported >6 hours of aerobic activity each week, with many of them reporting 7 or more hours. Because of the competitive nature of clubs from which they were recruited, it may be assumed that they were exercising at a higher intensity and greater frequency than were study 1 exercisers. These subjects, for the most part, reported that they had been in training for several years.
This difference in group exercise status between the 2 studies may help to explain the difference in lag time findings. It is known that chronic exercise causes adaptations of the body’s antioxidant enzyme systems. This adaptation would decrease the susceptibility of an exerciser’s LDL to undergo oxidation via a controlled and a concomitant decrease in the production of free radicals and ROS. It is not know how long it takes for these positive adaptations to occur. The findings of these 2 studies suggest that truly chronic exercise (aerobic intensity, over several months or perhaps years) decreases the susceptibility of a male exerciser’s LDL to undergo oxidation (as seen in study 2). Conversely, regular aerobic stress for an overall shorter time span creates a more oxidative environment in the body, increasing the susceptibility of LDL to undergo oxidation (as seen in study 1). The length of time it would take a human body to adapt and overcome the oxidative stress of aerobic exercise as measured by the oxidizability of LDL is not known.
Exercise is generally regarded as an oxidative stress to the body. The oxidative environment that may be present in the plasma of exercisers would likely increase the susceptibility of their LDL to undergo oxidation, if one assumes that the body’s antioxidant systems are not performing at an optimal level. It is hypothesized that if LDL becomes oxidized in the artery wall, it becomes atherogenic. However, if it becomes oxidized in the plasma, it may be rapidly cleared by the liver, thereby lowering blood cholesterol levels.
The Figure⇓ illustrates the potential mechanism by which exercise-induced oxidative stress might contribute to the antiatherogenic effects of exercise. Given the generally accepted beneficial effects of aerobic exercise, the enhanced susceptibility to oxidation of LDL seen in study 1 exercisers (predominantly male) might suggest an ongoing oxidative clearance of LDL from the plasma. If so, oxidation of LDL in the plasma may itself account for some of the lipid-lowering effects of exercise and may actually be beneficial. The decreased susceptibility to oxidation of LDL seen in study 2 male exercisers might suggest an adaptation of their antioxidant enzyme systems, thereby decreasing the production of free radicals and ROS. It should be noted that study 1 male exercisers had higher mean blood lipid levels than did study 2 male exercisers. This supports the hypothesis that short-term, aerobic activity may cause oxidative modification of LDL in the plasma and clearance via the liver, thus lowering blood cholesterol. Long-term aerobic exercise may enhance the resistance of LDL to oxidation, thus stabilizing already lower blood lipid levels. Aerobic exercise does not appear to affect the oxidizability of LDL isolated from females regardless of their exercise status. Therefore, the many benefits of aerobic activity can be achieved without the potentially deleterious consequences of this oxidative stress.
The presence of oxidized lipids in the plasma is often associated with disease states, and there is no compelling reason to speculate that exercise-induced oxidation is different from other forms of oxidation. A corollary of this hypothesis would suggest that antioxidant supplements may raise plasma lipid levels in exercisers. Given the results of study 2, this seems unlikely. It may be prudent to assume that the oxidative stress induced by exercise is counteractive to its potential benefits, but the tissue exposed to recurrent oxidative stress generates defense strategies to minimize or negate the consequences of such stress.
Still, such systems may fail when compounded with other factors, such as chronic inflammation, poor diet, inadequate antioxidants, and genetic factors. Until the mechanism that regulates the delicate balance between the pro-oxidant and antioxidant factors during chronic exercise are clearly understood, it may be judicious to ensure an adequate intake of dietary antioxidants for the sporadic exerciser as well as the truly chronic exerciser.
Selected Abbreviations and Acronyms
|BMI||=||body mass index|
|ROS||=||reactive oxygen species|
|TBARS||=||thiobarbituric acid–reactive substances|
|Vo2 peak||=||peak oxygen consumption|
This work was supported by grant No. 94120115 from the American Heart Association, Georgia Affiliate (S.P.), and generous funds from the Department of Gynecology and Obstetrics of the Emory University School of Medicine (S.P.). Support from the Emory University Graduate Program in Nutrition and Health Sciences (R.S.-B.) is gratefully acknowledged. We thank Marquette Inc for their generous loan of treadmills and Larry Price, PhD, for his statistical analysis expertise. N.S. acknowledges the support of the American Heart Association Fellowship, Georgia Affiliate, 1995.
Haskell WL. The influence of exercise training on plasma lipids and lipoproteins in health and disease. Acta Med Scand.. 1986;711:25–37.
Borg G, Linderholm J. Perceived exertion and pulse rate during graded exercise in various age groups. Acta Med Scand Suppl. 1967;472:194–206.
Santanam N, Parthasarathy S. Paradoxical actions of antioxidants in the oxidation of low density lipoprotein by peroxidases. J Clin Invest. 1995;95:2594–2600.
Aveldano MI, VanRollins M, Horrocks LA. Separation and quantitation of free fatty acids and fatty acid methyl esters by reverse phase high pressure liquid chromatography. J Lipid Res.. 1983;24:83–93.
Sevanian A, Bittolo-Bon G, Cazzolato G, Hodis H, Hwang J, Zamburlini A, Maiorino M, Ursini F. LDL- is a lipid hydroperoxide enriched circulating lipoprotein. J Lipid Res. 1997;38:419–428.
Steinbrecher UP, Witztum JL, Parthasarathy S, Steinberg D. Decrease in reactive amino groups during oxidation or endothelial cell modification of LDL: correlation with changes in receptor mediated catabolism. Arteriosclerosis. 1987;7:135–139.
Negelkerke JF, Barto KP, van Berkel TJ. In vivo and in vitro uptake and degradation of acetylated low density lipoprotein by rat liver endothelial, Kupffer and parenchymal cells. J Biol Chem.. 1983;258:12221–12227.
Wieland E, Parthasarathy S, Steinberg D. Peroxidase-dependent metal-independent oxidation of low density lipoprotein in vitro: a model for in vivo oxidation? Proc Natl Acad Sci U S A.. 1993;90:5929–5933.
Savenkova ML, Mueller DM, Heinecke JW. Tyrosyl radical generated by myeloperoxidase is a physiological catalyst for the initiation of lipid peroxidation in low density lipoprotein. J Biol Chem.. 1994;269:20394–20400.
Higuchi M, Cartier LJ, Chen M, Holloszy JO. Superoxide dismutase and catalase in skeletal muscle: adaptative response to exercise. J Gerontol.. 1985;40:281–286.
Quintanilha AT. The effect of physical exercise and/or vitamin E on tissue oxidative metabolism. Biochem Soc Trans.. 1984;12:403–404.
Laughlin MH, Simpson T, Sexton WL, Brown OR, Smith JK, Korthuis RJ. Skeletal muscle oxidative capacity, antioxidant enzymes, and exercise training. J Appl Physiol.. 1990;68:2337–2343.
Takanashi K, Watanabe K, Yoshizawa I. The effect of 17-sulfoconjugated, catechol estrogens upon lipid peroxidation of rat liver microsomes. Biol Pharm Bull.. 1995;18:1120–1125.
Bobadilla RA, Henkel CC, Henkel EC, Escalante B, Hong E. Possible involvement of endothelium-derived hyperpolarizing factor in vascular responses of abdominal aorta from pregnant rats. Hypertension.. 1997;30:596–602.