Epicardial fat gene expression after aerobic exercise training in pigs with coronary atherosclerosis: relationship to visceral and subcutaneous fat

Abstract

Epicardial adipose tissue (EAT) is contiguous with coronary arteries and myocardium and potentially may play a role in coronary atherosclerosis (CAD). Exercise is known to improve cardiovascular disease risk factors. The purpose of this study was to investigate the effect of aerobic exercise training on the expression of 18 genes, measured by RT-PCR and selected for their role in chronic inflammation, oxidative stress, and adipocyte metabolism, in peri-coronary epicardial (cEAT), peri-myocardial epicardial (mEAT), visceral abdominal (VAT), and subcutaneous (SAT) adipose tissues from a castrate male pig model of familial hypercholesterolemia with CAD. We tested the hypothesis that aerobic exercise training for 16 wk would reduce the inflammatory profile of mRNAs in both components of EAT and VAT but would have little effect on SAT. Exercise increased mEAT and total heart weights. EAT and heart weights were directly correlated. Compared with sedentary pigs matched for body weight to exercised animals, aerobic exercise training reduced the inflammatory response in mEAT but not cEAT, had no effect on inflammatory genes but preferentially decreased expression of adiponectin and other adipocyte-specific genes in VAT, and had no effect in SAT except that IL-6 mRNA went down and VEGFa mRNA went up. We conclude that 1) EAT is not homogeneous in its inflammatory response to aerobic exercise training, 2) cEAT around CAD remains proinflammatory after chronic exercise, 3) cEAT and VAT share similar inflammatory expression profiles but different metabolic mRNA responses to exercise, and 4) gene expression in SAT cannot be extrapolated to VAT and heart adipose tissues in exercise intervention studies.

obesity is a cardiovascular disease (CVD) risk factor, and there is a strong association between obesity and atherosclerosis (15). Even though the link between obesity, atherosclerosis, and CVD morbidity and mortality has been widely discussed, the mechanism(s) by which these are linked is not sufficiently elucidated and remains controversial (25). Adipose tissue functions as an endocrine organ with the potential to modulate pathophysiology by producing and releasing pro- and anti-inflammatory cytokines that act in a paracrine and endocrine fashion. With the increase in obesity and obesity-related diseases, the need to clarify adipose tissue phenotype is essential.

Evidence is increasing that different adipose tissue depots have different functions. This is especially apparent in visceral and subcutaneous adipose tissue (SAT). Visceral adipose tissue (VAT) is considered to be more pathogenic than SAT, perhaps due to its proximity to metabolic organs (7). Many epidemiological studies have pointed to VAT as a risk factor for cardiovascular disease, but the mechanistic link remains partially unclear (4). Epicardial adipose tissue (EAT) shares an embryologic origin with VAT (6, 24, 44) and is of particular interest due to its close anatomic association with coronary arteries and myocardium. EAT thickness (via echocardiographic data) strongly correlates with abdominal visceral adipose tissue (via MRI) (25), and EAT mass has been suggested to reflect intra-abdominal VAT (26). In addition, EAT is a source of proinflammatory cytokines and has been shown to have more macrophage infiltration than SAT in patients with coronary artery disease (31). Therefore, like VAT's correlation with CVD, EAT may play a role in cardiac disease by its association with increased intramyocardial triglyceride content (44) or by displaying a similar adipokine profile as VAT (1, 35).

EAT can arbitrarily be divided into coronary peri-vascular adipose tissue (cEAT), which is directly around or on the coronary artery adventitia, and EAT over the myocardium (mEAT) since the two components may be functionally distinct (8) despite their anatomic contiguity. cEAT may interact with vascular cells in a paracrine fashion or through vasa vasora to affect vascular function (17, 25, 44, 47). Despite data supporting AT role as an endocrine organ, little is known about the phenotype of cEAT and if/how it differs from mEAT and, in turn, how these adipose tissue pads influence coronary and cardiac function, respectively. cEAT and mEAT may provide insight into a mechanistic link between adiposity and cardiac abnormalities.

Exercise is known to improve CVD risk factors (37). While data show that exercise reduces VAT mass independent of SAT mass (48, 49) and reduces EAT thickness (30), only a few studies investigated the molecular characteristics of VAT and SAT and no studies have investigated the effect of aerobic exercise on the molecular characteristics of cEAT or mEAT. A pig model, specifically one of familial hypercholesterolemia, was selected not only because the cardiovascular system of the pig closely approximates the human, but it also provides a good large animal model for humans (11) and hypercholesterolemia is a risk factor for human CVD (52). Human hypercholesterolemia is characterized by two- to threefold increases in low-density lipoprotein (LDL) cholesterol, causing rapidly progressing atherosclerosis, which leads to premature CVD (50).

The purpose of this study was to investigate the effect of 16 wk of aerobic exercise training on the expression of 18 genes selected for their role in chronic inflammation, oxidative stress, and adipocyte metabolism in four adipose tissue depots in a pig model of familial hypercholesterolemia (FH) with coronary artery disease (CAD): SAT, VAT (omental adipose tissue), cEAT, and mEAT. Given the importance of chronic inflammation in atherogenesis and the improvement of CVD risk factors with exercise, we hypothesized that aerobic exercise training will reduce the inflammatory profile of mRNAs in both components of EAT and VAT but will have little effect on SAT.

METHODS

Experimental animals.

Thirteen castrated adult male FH pigs (10–11 mo of age) were used in this study. The Rapacz Familial Hypercholesterolemic (FH) model was developed at the University of Wisconsin by selective breeding (19, 21, 22, 33, 38, 39, 40, 41). These swine are characterized by a single missense mutation in the LDL receptor (LDLR) that decreases LDLR affinity for LDL, resulting in elevated total cholesterol levels between 180 and 240 mg/dl (20). We confirm the hypercholesterolemia of these pigs on a normal diet [total cholesterol = 360 ± 15 mg/dl, high-density lipoprotein (HDL) cholesterol = 30 ± 1 mg/dl, and total cholesterol/HDL cholesterol = 12 ± 1]. Pigs were randomly assigned to a sedentary (SED, n = 8) or aerobic exercised (EX, n = 5) group. Both SED and EX were fed 800 g/day of the University of Wisconsin gestation diet (a cholesterol-free, 3% fat, corn-soybean-based diet) and were allowed ad libitum access to water. Pigs were housed in rooms with a 12:12-h light-dark cycle at 20–24°C. It is a standard procedure to use castrated male pigs for exercise studies. Sex hormones can have a significant influence on adipose tissue biology. Whether castration influenced adipose tissue metabolism in this experimental pig model is unclear since no studies were done comparing fat depots in castrated animals with sex hormone-replaced animals.

Pigs underwent experimental treatment for 16–20 wk where SED pigs were restricted to their pens (2 × 4 m) and EX pigs performed a moderate-intensity (∼70% of maximum heart rate) daily aerobic exercise-training program on treadmills one time per day, 5 days/wk (42, 53, 54). Briefly, the aerobic exercise training protocol entailed a 5-min warm up at 2–2.5 miles per hour (mph), 15 min at 4 mph, and 20 min at 3 mph, and a 5 min cool-down at 2–2.5 mph. The duration and intensity were increased each week so that by week 10 of training, the pigs performed a 5-min warm up at 2–2.5 mph, 15 min at 6.5–7 mph, and 60 min at 4.5–5 mph, and a 5 min cool-down at 2–2.5 mph. The training program's effectiveness was ascertained from measurements of endurance time (pre- and post-stress testing) and heart weight/body weight ratios. The Animal Care and Use Committee at the University of Missouri approved all experimental protocols.

Tissue collection.

Intramuscular ketamine/xylazine and intravenous pentothal was used to deeply anesthetize the pigs at the end of the 16- to 20-wk experimental treatment, and euthanasia was achieved by heart extraction. The heart was placed in iced KREBS buffer, and cEAT and mEAT were removed from the heart surface by the same person, thereby minimizing the possibility that weight differences between the two adipose tissues were due to dissection technique differences. cEAT was defined and identified as the thick, contiguous adipose tissue, forming a ridge, and surrounding the coronary vessel extending from ∼10 mm up to ∼20 mm on each side of the vessel wall (Fig. 1). mEAT was defined and identified as the thinner contiguous adipose tissue beginning ∼10 mm to 20 mm away from the coronary vessel extending down on the ventricular myocardium. SAT was removed from the hip and is defined as the adipose tissue located beneath the skin but superficial to the underlying muscle. VAT was collected from omental adipose tissue interconnecting with the visceral organs in the abdominal peritoneal cavity. The adipose tissue from each depot was immediately snap-frozen in a metal beaker containing isopentane that had been sitting in liquid nitrogen long enough that it had begun to freeze. The adipose tissue was then removed from the isopentane and placed in a plastic bag sitting on dry ice. The bags were closed and the fat was stored in a −80°C freezer until analysis. Coronary vessels were collected for immunohistochemistry. Histological assessment of atherosclerosis in coronary arteries include Sudan IV-fatty streaks (distribution of atherosclerosis) and 5-μm sections were stained according to Verhoeff's method for elastin for grading of the lesion. To quantify Sudan IV staining, the percentages represent the percent of the lumen that is covered by lesions (i.e., fatty streaks) that stain positively for Sudan IV. To quantify IMT, for each pig we identified the area of the vessel with the greatest thickness (maximum IMT) and also the one with the least thickness (minimum IMT). Then, in each area we perform three separate measurements from the lumen to the external elastic lamina. Finally, we averaged those three measurements and report the maximum and minimum IMT.

Fig. 1.

Dissection procedure for coronary epicardial adipose tissue (cEAT) and myocardial epicardial adipose tissue (mEAT). A and B: intact right side and left side of the heart, respectively, before dissection. The black line indicates the approximate edge between cEAT and mEAT. RCA, right coronary artery; LAD, left anterior descending coronary vessel; LCX, left circumflex coronary vessel; RV, right ventricle, LV, left ventricle. C: cEAT layer is peeled back to reveal the LAD and LCX coronary vessels. D: cEAT layer is peeled back to reveal more of the LAD and LCX coronary vessel.

mRNA isolation.

Approximately 0.5 g of frozen tissue was homogenized with 5 ml of a monophasic solution of phenol and guanidine isothiocyanate (TRIzol reagent, Invitrogen, Carlsbad, CA) using a Polytron homogenizer. It was assumed that the small 0.5-g sample of fat was representative of the total mass of each fat depot. In addition, we found no evidence of a difference in the yield of total RNA per gram of tissue for cEAT vs. mEAT. The total RNA per gram values in relative units are cEAT = 416 ± 23 and mEAT = 349 ± 30 as the mean ± SE for all 13 pigs. The extracts were spun at 12,000 g for 10 min at 2–8°C. A Pasteur pipette was used to remove the fat layer above the extract.

RT-PCR.

The mRNA assay involved real-time quantitative PCR (13, 14). Transcriptor First Strand cDNA synthesis Kit from Roche Diagnostics was used on equal quantities of RNA to prepare the complementary DNA (cDNA). The Roche Lightcycler 480 Real-time RT-PCR system and Roche's Universal Probe Library of short hydrolysis Locked Nucleic Acid (LNA) dual hybridization probes combined with the primers recommended by their web-based assay design center [http://www.universalprobelibrary.com] were used for mRNA quantification. Integrated DNA Technologies (Coralville, IA) synthesized the primers. Eighteen mRNAs were targeted for their roles 1) in inflammation: adiponectin, interleukin -1 receptor antagonist (IL-1Ra), interleukin-6 (IL-6), interleukin-8/CXCL8 (IL-8), toll-like receptor 4 (TLR4), plasminogen activator inhibitor-1 (PAI-1), prostaglandin D₂ synthase (PGDS), angiotensinogen, vascular endothelial growth factor a (VEGFa), and visfatin; 2) as pro-oxidants: heme oxygenase, cytochrome c oxidase subunit 6; and as anti-oxidants: mitochondrial Mn superoxide dismutase-2 (MnSOD), endothelium-derived nitric oxide synthase (eNOS) and glutathione peroxidase 3 (GPX3); and 3) in adipocyte metabolism: fatty acid binding protein 4 (FABP4), perilipin, and uncoupling protein-2 (UCP-2). Cyclophilin was used as the “housekeeping gene”(see below).

In each assay, 55 ng per tube of total RNA (determined by absorption at 260 nm in a spectrophotometer) was used, and the ratio of right to left primers was 1 for each assay. The data were obtained as crossing point values (Cp) obtained by the second derivative maximum procedure as described by Roche Applied Science technical notes LC10/2000 and 13/2001 [http://www.roche-applied-science.com/sis/rtpcr/htc/index.jsp]. The Cp values are comparable to crossing threshold (Ct) values as defined by ABI (http://www.rdml.org), and will be referred to as Ct values in this paper. The average value of six replicates was used for cyclophilin values in each tissue of each pig and of duplicate samples for the other mRNAs. Cyclophilin was used as the recovery standard to normalize the data, which corrected for the effects of exercise on the recovery of cyclophilin in each tissue. No correction was necessary for SAT or cEAT Ct since exercise did not affect cyclophilin Ct (−0.01 ± 0.30 and 0.04 ± 0.42, respectively). Exercise did affect cyclophilin Ct for mEAT; therefore those data were corrected by −0.76 ± 0.26.

Relative quantification of the data was calculated using the comparative Ct method, which removes the need for standard curves. Ratios calculated from ΔCt are based on a log₂ scale (2^−ΔCt). This method is the comparative C_t procedure expressed in the ABI PRISM 7700 Sequence Detection System user Bulletin 2 for quantitative RT-PCR. The calculation of ratios was done without an efficiency correction by assuming that the amount of target molecules doubles with every PCR cycle.

Statistics.

A two-tailed Student t-test was performed to determine if differences between exercise and sedentary groups were significant at a P value of <0.05. Statistical analysis of mRNA values was based on the ΔCt values before log₂ transformation to ratios. A between-group comparison using the means of the ΔCt values was performed to compare cEAT to the other depots.

RESULTS

Experimental animals.

Representative pictures from EX and SED hearts are shown in Fig. 2. mEAT mass was significantly higher in the EX pigs (Table 1). There was a nonsignificant trend (P = 0.063) for total EAT mass to be higher in EX pigs. There was a significant correlation of heart weight with total EAT in all 13 pigs (adjusted r² = 0.48, P = 0.005) (Fig. 3) and by inference of total EAT with myocardial mass since EAT accounted for only ∼3% of the total heart weight. Although body weight was not different between SED and EX, the effect of aerobic exercise was apparent. Heart weight, expressed in absolute terms (g), and the heart weight/body weight ratio (g/kg) were higher in EX pigs. EX pigs showed significantly more duration of running than SED in the post-stress test (27.70 ± 0.13 vs. 21.00 ± 0.71 min). In addition, SED pigs showed a significant decrease in running time between the pre- and post-stress test (24.21 ± 1.22 vs. 21.00 ± 0.71 min). All pigs showed evidence of atherosclerosis (Fig. 4). As shown in Table 2, there was no significant difference between EX and SED IMT. Sudan IV staining revealed 9.66% in EX + SED, 14.31% in EX, and 6.6% in SED (EX vs. SED, P = 0.10).

Fig. 2.

Representative pictures of EX and SED hearts. EX, heart from aerobic exercise-trained familial cholesterolemia (FH) pig; SED, heart from sedentary FH pig.

Table 1.

Pig characteristics

	SED (n = 8)	EX (n = 5)
EAT content
Total peri-coronary epicardial adipose tissue (cEAT), g	4.10 ± 0.36	4.99 ± 0.57
Total peri-myocardial epicardial adipose tissue (mEAT), g	0.75 ± 0.16	1.55 ± 0.30*
Total EAT, g	4.85 ± 0.40	6.54 ± 0.82
Other characteristics
Heart weight, g	169.1 ± 4.3	191.4 ± 8.2*
Body weight, kg	49.4 ± 41.6	49.8 ± 1.5
Heart weight/body weight ratio, g/kg	3.4 ± 0.3	3.8 ± 0.1*
Heart AT % (total EAT/heart weight),	2.9 ± 0.2	3.4 ± 0.2
Pre-stress test, min	25.21 ± 1.22	26.54 ± 1.53
Post-stress test, min	21.00 ± 0.71†	27.70 ± 0.13*

Data are presented as means ± SE. EAT, epicardial adipose tissue; AT, adipose tissue; SED, sedentary familial hypercholesterolemia (FH) pig; EX, exercise-trained FH pig.

^*P < 0.05 for SED vs. EX.

^†P < 0.05 for pre-stress test vs. post-stress test.

Fig. 3.

Correlation of heart weight with total EAT weight for all 13 pigs. r² = 0.53; adjusted r² = 0.48, P = 0.005.

Fig. 4.

Intima-media thickness (IMT): grading of lesion of structural atherosclerosis. LAD-ex, left anterior descending coronary artery from EX pig; RCA-ex, right coronary artery from EX pig; LAD-sed, left anterior descending coronary artery from SED pig; RCA-sed, right coronary artery from SED pig. The vessels are stained for VVG to identify elastin (black) and collagen (dark-red; purple) fibers, which were are use to determine the IMT. Scale bar, 100 μm.

Table 2.

Mean maximum and minimum intima-media thickness values

	LAD-IMT Maximum, μm	RCA-IMT Maximum, μm	LAD-IMT Minimum, μm	RCA-IMT Minimum, μm
EX	631.4 ± 108.3	575.8 ± 99.6	208.4 ± 31.8	265.8 ± 33.0
SED	721.6 ± 161.8	451.1 ± 43.8	230.6 ± 23.7	228.7 ± 24.0

Data are presented as means ± SE. IMT, intima-media thickness, LAD, left anterior descending coronary artery, RCA, right coronary artery.

Sedentary pigs showed almost no difference in gene expression between cEAT and mEAT.

We found no evidence of a difference in the yield of total RNA per gram of tissue for cEAT vs. mEAT. As shown in Fig. 5, with the exception of a 60% greater amount of cytochrome c oxidase mRNA in mEAT, there was no statistically significant difference between cEAT and mEAT in the expression of any of the other 17 mRNAs examined under sedentary conditions. mEAT and cEAT differed from VAT and SAT in several respects. There was far less heme oxygenase, angiotensinogen, adiponectin, perilipin, and FABP4 in both cEAT and mEAT than in VAT or SAT. Gene expression in VAT was more comparable to that in SAT than to that in cEAT or mEAT for all genes except PGDS whose expression was elevated in VAT as well as cEAT or mEAT. Of the putative inflammatory markers, none was elevated in cEAT compared with VAT or SAT except for PGDS.

Fig. 5.

Comparison of peri-coronary EAT (cEAT), peri-myocardial EAT (mEAT), and visceral omental adipose tissue (VAT) gene expression to that in subcutaneous adipose tissue (SAT) of sedentary pigs. The values are expressed as the ratio of mRNA expression in mEAT, cEAT, and VAT to that in SAT in 8 sedentary pigs. The ratios are the means ± SE and derived from the ΔCt corrected for the recovery of cyclophilin. With the exception of cytochrome c oxidase (P = 0.005), none of the mRNAs were expressed at significantly different values in cEAT vs. mEAT. FABP4, fatty acid binding protein-4; GPX3, glutathione peroxidase 3; IL-8, interleukin-8; IL-1Ra, interleukin 1 receptor antagonist; UCP-2, uncoupling protein-2; eNOS, endothelial nitric oxide synthase; VEGFa, vascular endothelial growth factor a; TLR4, toll-like receptor 4; PAI-1, plasminogen activator inhibitor 1; PGDS, prostaglandin D₂ synthase; IL-6, interleukin 6. Statistically significant differences between expression in cEAT, mEAT, or VAT relative to SAT are shown as follows: *P < 0.05, **P < 0.025, ***P < 0.005.

cEAT was distinctly different from mEAT in response to aerobic exercise.

The effect of aerobic exercise on the relative expression of mRNAs was compared in cEAT, mEAT, VAT, and SAT (Fig. 6). Peri-coronary epicardial adipose tissue (cEAT) is distinctly different in its response to aerobic exercise than peri-myocardial epicardial adipose tissue (mEAT). mRNAs are downregulated in response to aerobic exercise in mEAT but not cEAT. Of the 18 mRNAs shown in Fig. 5, it is evident from Fig. 6 that aerobic exercise reduced the expression of inflammatory genes IL1-Ra, IL-6, IL-8, PAI-1, and PGDS and redox genes eNOS and cytochrome c oxidase in mEAT but not cEAT, in which superoxide dismutase was the only molecule to decrease.

Fig. 6.

Comparison of peri-coronary EAT (cEAT), peri-myocardial EAT (mEAT), VAT, and SAT gene expression in exercised pigs to that in sedentary pigs. The mRNA values are the ratios ± SE of the mRNA expression for adipose tissue from five exercised (EX) pigs compared with that in 8 sedentary (SED). Ratios greater than 1.0 indicate that the level of expression was enhanced by exercise, while ratios less than 1.0 indicate that the level of expression was reduced by aerobic exercise. The horizontal line at 1.0 indicates a ratio (SED/EX) where the expression of mRNAs were equal for SED and EX (i.e., the means were identical). A bar below this horizontal line indicates that EX lowered the expression of the given mRNA. Statistically significant differences between expression in EX mRNA compared with SED mRNA are shown as follows: *P < 0.05, **P < 0.025, ***P < 0.005.

VAT exhibited different mRNA expression responses to aerobic exercise training.

As shown in Fig. 6, after exercise, VAT showed a striking absence of changes in inflammatory and redox gene expression and instead a downregulation of genes involved with adipocyte metabolism such as FABP4, perilipin, adiponectin, and UCP-2 that were not observed in cEAT and mEAT. Exercise had no effect on cyclophilin, eNOS, PAI-1, IL1-Ra, cytochrome c oxidase, or IL-8 in VAT.

Only a limited number of genes in SAT were altered by aerobic exercise training.

Of the 18 mRNAs that were measured in SAT shown in Fig. 5, only VEGFa gene expression was enhanced, and IL-6 decreased by exercise in SAT (Fig. 6). Clearly exercise affects gene expression to a far lesser extent in SAT than in VAT fat where the expression of 6 genes was reduced by exercise or mEAT where the expression of 10 genes was reduced by exercise (Fig. 6).

DISCUSSION

The purpose of this study was to investigate the effects of 16 wk of aerobic exercise training on the inflammatory, oxidative stress, and metabolic mRNA phenotypes of four adipose tissue depots in a pig model of FH with CAD. The major finding that arose from the study was that aerobic exercise training increased the mass of mEAT but not cEAT at the same time as it reduced the inflammatory response in mEAT but not cEAT. These very different responses to exercise strongly suggest that these two components of EAT are functionally distinct. Other important findings were 1) exercise did not downregulate inflammatory and redox genes in VAT except for PGDS and superoxide dismutase, respectively, but it did lower expression of adipocyte-specific adiponectin, FABP4, and perilipin mRNAs, which was not observed in cEAT and mEAT; and 2) SAT did not respond to aerobic exercise training except that VEGFa mRNA went up and IL-6 mRNA went down.

Anthropometric data.

It is important to note that this was not a weight-loss study since both groups of pigs were fed the same amount of food each day and there was no body weight difference between EX and SED pigs at the end of the study. This suggested that the effects on adipose tissue phenotype were due to aerobic exercise training per se. Interestingly, despite no change in body weight, EX pigs had twice as much mEAT as SED pigs, which may be related to their 13% greater heart weight. The reason for the expansion of epicardial fat over the myocardium in concert with myocardial hypertrophy after chronic exercise in pigs is unclear. Our data showing a direct correlation between EAT and myocardial weights is in keeping with human autopsy studies, in which both left and right ventricular weight correlated directly with dissected epicardial fat weight, a relationship that suggests reciprocal regulation between the two tissues (9). The percentage of epicardial adipose tissue (expressed as total heart adipose tissue weight/heart weight) was also not altered by aerobic exercise. This finding is contrary to the findings of Kim et al. (29) who investigated the effect of exercise on epicardial adipose tissue thickness in obese, middle-aged men. These subjects underwent 12 wk of aerobic exercise training (60–70% maximum heart rate, 60 min, 3 days/wk). Epicardial adipose tissue thickness was measured by transthoracic echocardiography. Kim et al. (29) found a significant reduction in epicardial adipose tissue thickness. Besides the difference in species, another possible reason for the discrepancy between our studies and those of Kim et al. is that the human subjects were obese, while the FH pigs were not. Another possible difference is that food intake was restricted in exercised pigs to that of sedentary pigs. Given that exercise training has been shown to reduce atherosclerosis, it is surprising that we did not see reductions in Sudan IV staining or reductions in IMT in the EX pigs; rather there was a trend for an increase in Sudan IV staining for EX pigs and no difference in IMT between EX and SED. One explanation for this counterintuitive result is that the elevated total cholesterol levels present in these FH pigs overpowers the effect of exercise in coronary arteries. Junyent et al. (28) related IMT to established and emergent risk factors in humans with FH and found that physical exercise was independently associated with IMT. Others have found similar results with exercise (5, 34). Mayet et al. (34) found no disparity in IMT between elite female athletes and sex-matched control subjects and attributed this finding to “intermittent exposure to markedly elevated systemic arterial pressures during exercise, resulting in intermittent elevations in carotid wall tensile stress.”

Coronary peri-vascular epicardial adipose tissue was distinctly different from contiguous peri-myocardial epicardial adipose tissue in response to aerobic exercise.

The novel finding in this study was that aerobic exercise training reduced a composite of 10 inflammatory and redox genes in mEAT but only one antioxidant gene, superoxide dismutase, and no proinflammatory genes in cEAT. This is contrary to our hypothesis that aerobic exercise training would reduce the inflammatory profile of mRNA in both locations, in particular because the data were generated from pigs with CAD. Intuitively, exercise might have been anticipated to reduce atherosclerosis risk by dampening inflammation in cEAT especially because of its closer proximity to the coronary arteries than mEAT. This was not the case, putatively because of associated “inside-to outside” inflammatory signaling (35, 44) emanating from atherogenesis in underling intima-medial coronary plaques, which were demonstrated histologically in our experimental animals. Our designation of cEAT as epicardial adipose tissue directly on or around coronary artery, and mEAT as epicardial adipose tissue on the myocardium but not near or touching an artery was arbitrary but in keeping with radiological definitions of the two entities (27, 30). Our findings support a recent report of a difference between cEAT compared with mEAT from adult human donor hearts without macroscopic evidence of CAD in that cEAT was more proinflammatory, releasing greater amounts of monocyte chemoattractant-1 in vitro (8). Despite the significance of the differences between the SED and EX pigs, we acknowledge that a larger sample size might have potentially yielded more significant differences in comparing cEAT as opposed to mEAT expression data. However, if we assume the differences are real without regard to standard errors or statistical tests, the data are as follows: PAI-1 was −59% in mEAT and −14% in cEAT, eNOS was, respectively, −56% and −2%, cyclophilin was −41% and −3%, PGDS was −45% and −19%, IL-1Ra was −64% and +23%, IL-6 was −65% and −6%, cytochrome c oxidase was −30% and +40%, IL-8 was −65% and +20%, UCP-2 was −50% and −22%, while superoxide dismutase was −48% and −35% and both were significant. It was only with regard to UCP2 and PGDS that there might have been an effect in cEAT that might have been significant if we had a larger number of exercise-trained pigs, but in any case, the effect on UCP-2 and PGDS was, at most, half of that seen in mEAT.

The two EAT components seem to differ metabolically. In the SED pigs, mEAT had more cytochrome c oxidase, indicating more mitochondrial oxidizing units per unit of adipose tissue than EX pigs. However, there was significantly more mEAT mass in EX pigs, so that the total amount of cytochrome c oxidase may not decrease in EX pigs. Likewise, even though lipolytic perilipin and FABP4 expression was the same in cEAT and mEAT before and after exercise, it is possible that the higher mEAT mass postexercise reflects more total mEAT lipolysis that could possibly deliver free fatty acids locally to the hypertrophied myocardium. Superoxide dismutase decreased with aerobic exercise training in both cEAT and mEAT. The enzyme catalyzes the conversion of superoxide to H₂O₂ and oxygen (46). Exercise training enhances antioxidant capacity via reactive oxygen species scavenging (46) and has been shown to increase in rat white visceral adipose tissue with exercise (45). The disparate results may relate to the source of tissue analysis, the experimental animal model, and the presence of underlying CAD. In SED pigs, apart from cytochrome c oxidase, mEAT and cEAT showed few differences but differed from VAT and SAT. Compared with SAT, there was far less heme oxygenase, angiotensinogen, adiponectin, perilipin, and FABP4 in both cEAT and mEAT. It is unclear what this reflects, but this suggests a different function for EAT compared with omental VAT. In addition, with the exception of PGDS, none of the putative inflammatory markers such as IL-8, IL-1Ra, angiotensinogen, TLR4, visfatin, PAI-1, or IL-6 were elevated in EAT.

Aerobic exercise training did not significantly alter cEAT mass but increased mEAT mass. Pigs lack functional UCP-1 (2), which, when coupled with a decrease in UCP-2 in EX pigs' mEAT, in part may account for the significant increase in EAT mass in EX pigs. On the other hand, perilipin, a regulator of lipolysis, was not downregulated in VAT as it was in EAT. It is possible that the expansion of the myocardium that occurred during exercise also expands adipose tissue at the periphery of the organ and this is less inflamed because it is further away from the toxic influence of inflamed endothelium in the sedentary atherosclerotic heart, which has a poorer oxygen supply. If this is the case then it is curious that the greatest effects of exercise are not seen in the fat closest to the putative inflammatory source. It may be that adipose tissue in the growth phase of expansion is more insulin responsive and exhibits less inflammation. Further studies are needed to substantiate whether cEAT and mEAT have distinct functions. Hypothetically, cEAT may alter vasomotion like peri-vascular fat around arteries in other extracardiac sites (15, 16).

It has been previously suggested that, due to its immediate anatomic apposition to the heart without fascial boundaries, EAT might secrete a number of inflammatory cytokines that interact with and modulate the coronary arteries and myocardium via paracrine and vasocrine pathways (25). We found that IL-1β, IL-6, IL-8, PAI-1, and adiponectin were expressed in both cEAT and mEAT in our sedentary pig model, suggesting that similar paracrine interactions might exist in the pig heart, but whether these inflammatory genes were altered by underlying CAD is unknown due to the absence of a control group of similar strain pigs with normal coronaries.

Omental and epicardial adipose tissue exhibited different responses to aerobic exercise training.

Like VAT, EAT is, by definition, a visceral adipose tissue possessing several common characteristics with mesenteric/omental VAT: EAT and VAT share an embryologic origin in the splanchnopleuric mesoderm affiliated with the gut (6, 24, 44), EAT and VAT increase in obesity (27) and decrease with weight loss (30), and both EAT and VAT are associated with cardiovascular risk factors (6, 10). Given the similarities between EAT and VAT, a similar exercise response could have been expected. However, in this pig model, exercise did not downregulate inflammatory and redox genes in VAT except for PGDS and MnSOD, respectively, but it did lower expression of adipocyte-specific adiponectin, FABP4, and perilipin mRNAs, which was not observed in cEAT and mEAT. Adiponectin is an anti-inflammatory adipokine that inhibits TNF-α production and stimulates production of other anti-inflammatory cytokines (57). Adiponectin blood levels are reduced in obesity (3) and reduced in hypoxia (56). Increased physical activity is associated with increased levels of adiponectin (43, 57). The interpretation of the decrease in adiponectin in VAT but not in SAT is limited by the absence of any direct quantitation of VAT mass and blood adiponectin levels in our experiments especially since exercise training in humans has been shown to reduce VAT in the absence of significant weight loss (48, 49).

Most subcutaneous adipose tissue mRNAs did not respond to aerobic exercise training.

With the exception of VEGFa mRNA (up) and IL-6 mRNA (down), SAT showed no response to aerobic exercise training in 11 of the mRNAs that were affected by aerobic exercise training in VAT and/or EAT. Unlike VAT and EAT, SAT is derived from a different mesodermal embryologic origin, is not associated with increased metabolic risk, and may be protective against metabolic disorders (51). Both VEGFa (18) and IL-6 (12) have proinflammatory characteristics. IL-6 is currently considered one of the primary mediators of chronic inflammation, and adipose tissue accounts for 15–35% of circulating levels (55). VEGF is proangiogenic (23) and plasma levels have been shown to increase with exercise training (32). One possible explanation for the opposite response seen in VEGFa and IL-6 is that the increase in VEGFa mRNA in EX pigs' SAT increased blood flow, which flow may result in better oxygenation and less inflammation, and thus less IL-6 mRNA.

In conclusion, this is the first study to investigate the response of mRNA levels to aerobic exercise training in four adipose tissue depots of pigs with FH and CAD. The major finding is that aerobic exercise training reduced the inflammatory response in mEAT but not cEAT. This suggests that epicardial adipose tissue is not homogeneous in its responses to aerobic exercise training. After exercise, VAT showed decreased preferential expression of adiponectin and other adipocyte-specific genes. None of the 19 genes in SAT responded to aerobic exercise training, except IL-6, which went down, and VEGFa mRNA, which went up. This indicates that responses of other inflammatory mRNAs in subcutaneous adipose tissue cannot be extrapolated to omental and heart adipose tissues in exercise intervention studies.

GRANTS

This research was supported by National Institutes of Health Grants HL-52490 and T32-AR-048523 and The Cardiometabolic Disease Research Foundation, Memphis, TN.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).