Use of Fat-Fed Rats to Study the Metabolic and Vascular Sequelae of Obesity and β-Adrenergic Antagonism

Abstract

Obesity-associated cardiovascular disease exerts profound human and monetary costs, creating a mounting need for cost-effective and relevant in vivo models of the complex metabolic and vascular interrelationships of obesity. Obesity is associated with endothelial dysfunction and inflammation. Free fatty acids (FFA), generated partly through β-adrenergic receptor-mediated lipolysis, may impair endothelium-dependent vasodilation (EDV) by proinflammatory mechanisms. β-Adrenergic antagonists protect against cardiovascular events by mechanisms not fully defined. We hypothesized that β antagonists may exert beneficial effects, in part, by inhibiting lipolysis and reducing FFA. Further, we sought to evaluate the fat-fed rat as an in vivo model of obesity-induced inflammation and EDV. Control and fat-fed rats were given vehicle or β antagonist for 28 d. Serum FFA were measured to determine the association to serum IL6, TNFα, and C-reactive protein and to femoral artery EDV. Compared with controls, fat-fed rats weighed more and had higher FFA, triglyceride, leptin, and insulin levels. Unexpectedly, in control and fat-fed rats, β antagonism increased FFA, yet inflammatory cytokines were reduced and EDV was preserved. Therefore, reduction of FFA is unlikely to be the mechanism by which β antagonists protect the endothelium. These results reflect the need for validation of ex vivo models of obesity-induced inflammation and endothelial dysfunction, concurrent with careful control of dietary fat composition and treatment duration.

The prevalence of overweight and obese adults in the United States has increased by almost 20% over the last 3 decades.³⁶ Similar upward trends have been observed in persons between 6 and 19 y of age.³⁷ The obesity epidemic extracts a monetary cost of more than $92 billion on medical care alone⁵⁶ and a profound human price in the form of increased disease³⁵ and higher death rates.⁵⁰

Obese adults have a higher risk of morbidity and mortality due to cardiovascular disease.⁵⁷ In health, endothelial cells that line the luminal surface of blood vessels release mediators that facilitate the appropriate regulation of multiple processes, including vascular permeability, inflammation and cell adhesion, coagulation, maintenance of intercellular matrix, lipid metabolism, and vascular reactivity.^25,42 Dysregulation of these processes favors inflammation, coagulation, and vasoconstriction. Not surprisingly, endothelial dysfunction as measured by impairment of endothelium-dependent vasodilation (EDV) is an early and reliable predictor of cardiovascular events in humans.^43,47

Obese persons have increased serum free fatty acids (FFA).¹ Obesity^14,53 and FFA¹³ are associated with increased circulating inflammatory markers, specifically IL6, TNFα, and C-reactive protein (CRP). In addition, both obesity³² and FFA^8,46 are associated with impaired EDV, and FFA exert direct adverse inflammatory effects on the endothelium.¹⁸

Partly in response to stimulation of β-adrenergic receptors, FFA are the principle moiety secreted from adipocytes. β-Adrenergic antagonist drugs reduce morbidity and mortality in patients with coronary artery disease^9,26 and affect both the myocardium^2,38 and vasculature.^5,51,55 The mechanisms by which β-adrenergic antagonists exert protection remain unclear, but reduced generation of FFA might play a role.

The first aim of the present study was to determine whether β antagonism lowers serum FFA in fat-fed rats and whether the magnitude and direction of change in FFA is correlated with circulating inflammatory markers and EDV. The β₁- and β₃-receptor subtypes predominantly mediate lipolysis in rodent adipose;^12,30 however, to minimize the potential for compensatory upregulation of unopposed receptors in this study, the β₁β₂ antagonist propranolol was combined with the β₃ antagonist SR59230A (Sigma-Aldrich, St Louis, MO) to exert antagonism at all 3 receptor subtypes. To test whether the high-fat dietary treatment was associated with a metabolic milieu consistent with obesity, serum triglycerides, leptin, glucose, and insulin concentrations were measured.

An inexpensive, valid, and physiologically relevant in vivo system would be valuable for studying the pandemic of obesity³¹ and related endothelial dysfunction.¹⁷ We are unaware of studies of whether long-term high-fat feeding affects in vivo EDV in the rat, although a study validating the use of high-resolution ultrasonography to measure in vivo flow-mediated vasodilation in normal rats was published recently.¹⁷ The second aim of our study was to investigate the fat-fed rat as a model of human diet-induced endothelial dysfunction. To retain the complex metabolic-vascular interplay that occurs in the intact organism, we used in vivo measures of EDV to assess the integrated physiologic response. In humans, the dilator response of peripheral vessels is associated with coronary EDV response.⁴⁸ We studied the rat femoral artery, with the aim of demonstrating changes in vasodilator responses in this easily isolated peripheral vascular bed.

Materials and Methods

Animals.

Adult male Sprague–Dawley rats (CD IGS Rat, Charles River Laboratories, Wilmington, MA) were housed in environmentally controlled rooms with 12:12-h light:dark cycles. The study protocol was approved by the University of Colorado Health Sciences Center Animal Care and Use Committee.

Diets.

Preliminary feeding studies using the treatment diet revealed an association between high-fat feeding and increased serum FFA (P < 0.001) and body weight (P < 0.001), despite an average of 27% less dietary consumption by weight in the fat-fed group (data not shown). Ad libitum dietary treatment (continuation of normal diet or introduction of high-fat diet) was initiated at 1 mo of age and continued for 16 wk. The pelleted high-fat diet (Harlan Teklad, Madison, WI) was composed of 41.7% kcal from fat (33.4% kcal from lard, 8.4% kcal from soybean oil), 38.6% kcal from carbohydrate, and 19.7% kcal from protein (metabolizable energy, 4.3 kcal/g). Diets of similar percentage fat composition have been associated with in vitro vascular dysfunction in diabetic²¹ and normal¹¹ rats. Further, insulin-resistant rats fed a diet high in soybean oil manifested endothelial dysfunction in isolated renal arteries.¹⁰ The diet did not contain vitamin C. Vitamin E concentrations were minimized to approximate the nutrient requirements for rats so benefits of exogenous antioxidants on EDV were minimized. The standard control diet (Harlan Teklad) had the following composition: 14% kcal from fat, 60% kcal from carbohydrate, and 26% kcal from protein (metabolizable energy, 3.8 kcal/g). Approximately 55% of the total fat was contributed by soybean oil, with the remaining fat contributed by corn, wheat, and fish (approximately 20%, 16%, and 7%, respectively).

Preexperimental and anesthetic protocol.

Rats were assigned to 1 of 6 groups: control (no vehicle pellet); control + vehicle; control + β antagonist; high-fat (no vehicle pellet); high-fat + vehicle; high-fat + β antagonist. Vehicle pellets were composed only of biologically inert matrix material (that is cholesterol, cellulose, lactose, phosphates, and stearates). Treatment pellets were generated by direct incorporation of active drugs into the same inert matrix. For each treatment, 8 rats were used for serum and tissue sampling, and 8 rats were used for measurement of EDV. Vehicle or combined β-antagonist treatments were delivered by slow-release pellets (Innovative Research of America, Sarasota, FL) implanted at week 12 of the 16-wk dietary intervention. The dosage of the β₁β₂ antagonist propranolol was 0.1 mg/kg daily, and the β₃ antagonist SR59230A was administered at a dose of 1 mg/kg daily (Sigma–Aldrich, St Louis, MO). Doses were based on acute inhibition of isoproterenol-stimulated lipolysis in preliminary studies (data not shown) and on previously reported work.⁴⁹ Anticipating increased serum inflammatory cytokine concentrations with vehicle pellet implantation alone, rats that received dietary treatment only (that is, no surgical intervention) were used as control animals in cytokine analyses.

Rats were fasted overnight, and body weights were measured prior to anesthesia. General anesthesia was induced with intraperitoneal ketamine (80 mg/kg) and xylazine (12 mg/kg), and supplemental doses were given as needed for anesthesia maintenance; 95% O₂–5% CO₂ was delivered by mask. To minimize the effects of handling and endogenous catecholamine release, a 30-min equilibration time was allowed prior to measurement of EDV.

Hemodynamic indices and endothelium-dependent vasodilation.

Instrumentation was adapted from approaches outlined previously.⁴⁰ The left carotid artery was catheterized with polyethylene-10 tubing attached to a pressure transducer and data analysis software (Biopac Systems, Goleta, CA) for recording of mean arterial pressure (MAP). The distal aorta was cannulated via retrograde advancement of polyethylene-50 tubing through the left femoral artery, with the catheter tip positioned at the aortic bifurcation. A microport connected to the distal end of this catheter was used for drug infusion. The catheter was of sufficient diameter to occlude the artery; thus, drugs necessarily flowed into the right femoral artery. Blood flow in the right femoral artery was measured by a perivascular 0.5-mm flow probe attached to the flow sensor (Transonic Systems, Ithaca, NY). This is a modification of the published approach,⁴⁰ which describes cannulation of a small branch of the femoral artery for infusion, at a location distal to the flow probe. Resistance was calculated as MAP divided by flow. Because the small arteries and arterioles are the principle contributors to peripheral resistance, changes in flow measured by using this protocol did not reflect changes in tone in the conduit femoral artery per se but rather those in the downstream resistance arteries.

We characterized EDV by measuring flow changes in response to the vasodilator acetylcholine (Sigma–Aldrich). Cumulative doses of acetylcholine (0.25, 0.75, 2.5 μg/kg) were injected through the left femoral microport at 10-min intervals. Response to acetylcholine, termed the flow–time integral (FTI), was defined as the area under the flow–time curve⁷ in milliliters from the time of injection to the time at which the flow reached a plateau, defined as the time at which flow remained unchanged for ≥ 90 s (Figure 1). To validate the model, preliminary work was conducted to establish an FTI acetylcholine dose–response, and the portion of the response attributable to nitric oxide was quantitated by measuring attenuation by N_ω-nitro-L-arginine methyl ester (L-NAME), an inhibitor of nitric oxide synthase (Sigma–Aldrich). Further, endothelium-independent vasodilation was characterized by measuring the FTI in response to the nitric oxide donor sodium nitroprusside (Sigma–Aldrich).

Figure 1.

Flow data in response to a single injection of acetylcholine (black arrow). Red line represents plateau in flow curve. The area under the flow curve that lies between the acetylcholine injection and the plateau is used to quantitate FTI in milliliters and is outlined in black.

After the FTI in response to 0.25, 0.75, 2.5 μg/kg acetylcholine were measured, the dose response was repeated 10 min after a 10-min infusion of 10 mg/kg L-NAME. The difference in FTI before and after L-NAME was determined and expressed as ΔFTI. After a second equilibration period, 5.2 mg sodium nitroprusside was administered. Because this treatment caused a profound decrease in MAP, the absolute change in resistance as well as the percentage change from preadministration values were calculated to quantitate the endothelium-independent vasodilatory response.

Implantation of subcutaneous pellets.

General anesthesia was induced with isofluorane. The interscapular space was prepared for aseptic surgery, and a 0.5-cm incision in the skin was made. The pellets were placed in the subcutaneous space approximately 3 cm caudal to the incision and 3 cm apart. Bleeding was minimal and controlled with local pressure.

Procurement and processing of serum samples.

Blood was aspirated by direct cardiocentesis through a medial sternotomy. Samples were placed in serum separator tubes, allowed to clot at room temperature for 2 h, and then centrifuged at 90,000 × g for 15 min. Serum aliquots were stored at −80 °C.

Serum triglyceride levels were measured by enzymatic glycerol-blanked methodology.⁵⁴ Serum leptin and insulin concentrations were measured at the University of Colorado General Clinical Research Center Laboratory by using rat leptin and insulin radioimmunoassay kits (Linco Research, St Charles, MO). Serum glucose was measured at the University of Colorado Clinical Laboratory by oxygen-depletion enzymatic methodology.

Serum FFA were measured by using an enzymatic calorimetric kit (NEFA-C, Wako Chemicals USA, Richmond, VA).⁶ Both IL6 and TNFα were quantitated by ELISA (BioSource International, Camarillo, CA). Serum CRP was determined by using an immunoturbidimetric assay (Turbitex CRP Ultra, Biocon, Rockville, MD).

Data analysis.

Analyses were conducted by using Prism 4.0 for Macintosh (Graphpad Software, San Diego, CA). Unpaired t tests and 1-way ANOVA were used to compare normal populations. When nonnormal data was not resolved with transformation, the Mann–Whitney or Kruskal–Wallis nonparametric tests were used. When the overall ANOVA P value was less than 0.05, the Tukey method of multiple comparisons was applied.

Data measured in response to progressively larger doses of acetylcholine (that is, FTI) were analyzed by using 2-way ANOVA for repeated measures. The Bonferroni posttest was used to compare individual means.

Pearson and Spearman correlations were applied to normal and nonnormal data, respectively. Data are expressed as mean ± SE, and statistical significance set at a P level of less than 0.05.

Results

Body weight and serum metabolic parameters.

Body weight and metabolic data are provided in Table 1. High-fat feeding was associated with increased body weight in vehicle animals. Rats given β-adrenergic antagonists had a further increase in body weight when compared with vehicle rats in both control and fat-fed groups.

Table 1.

Body weight and serum metabolic indices

	Control		Fat-fed
	Vehicle	β antagonist	Vehicle	β antagonist
Body weight (g)	413 ± 12	536 ± 21a	592 ± 26b	692 ± 21a
Triglycerides (mg/dL)	45.1 ± 6.4	66.7 ± 9.9	160 ± 21.2c	190 ± 30.6
Leptin (ng/mL)	1.24 ± .16	2.01 ± 0.45	13.22 ± 1.16b	13.8 ± 1.80
Glucose (mg/dL)	201 ± 23	243 ± 18	273 ± 14	301 ± 17
Insulin (ng/mL)	0.11 ± 0.01	0.19 ± 0.02a	0.31 ± 0.04c	0.32 ± 0.02

Values are shown as mean ± SE.

^aP < 0.01 when compared with corresponding vehicle group

^bP < 0.001 compared with control vehicle group

^cP < 0.01 compared with control vehicle group

High-fat–fed rats had higher serum triglycerides than did the controls (Table 1). Treatment with β-adrenergic antagonist did not alter triglyceride levels when compared with respective vehicle groups. The same trend was observed in serum leptin levels: high-fat feeding was associated with increased concentrations, but β-adrenergic antagonist treatment did not have an added effect. Serum glucose concentrations were similar among all treatment groups; however, serum insulin levels were increased in fat-fed animals. β-Adrenergic antagonist treatment was associated with increased serum insulin levels in control, but not fat-fed, rats.

High-fat feeding was associated with increased serum FFA in vehicle rats (Table 1). Counter to expectations, serum FFA were higher in both control and obese rats that received β-adrenergic antagonist treatment compared with vehicle rats (Figure 2). There was a weak positive correlation between body weight and serum FFA (Spearman r = 0.52; P = 0.0001).

Figure 2.

High-fat feeding resulted in higher serum FFA in vehicle rats (P < 0.05). β-Adrenergic antagonism was associated with increased serum FFA in both control (+, P < 0.01) and fat-fed (*, P < 0.05) rats.

Inflammation.

Serum TNFα, IL6, and CRP data are shown in Figure 3. High-fat feeding was not associated with increased serum cytokines; in fact, a trend toward decreased cytokine concentrations in fat-fed rats emerged, with the change in serum CRP being significant (P < 0.001). Vehicle pellet implantation alone often was associated with increased cytokine concentrations. Cytokines were reduced variably in rats that received β-adrenergic antagonist treatment. Positive correlation between serum FFA and cytokine concentrations was not present.

Figure 3.

(A) Serum TNFα. TNFα concentrations were reduced in control rats that received β-adrenergic antagonist treatment. +, P < 0.01. (B) Serum IL6. IL6 concentrations were reduced in control rats that received β-adrenergic antagonist treatment. +, P < 0.01. (C) Serum CRP. β-Adrenergic antagonist treatment reduced CRP in both control and fat-fed rats. #, P < 0.001; *, P < 0.05.

Hemodynamic indices and endothelium-dependent vasodilation.

Baseline MAP, flow, resistance, and heart rate were unchanged by high-fat feeding (data not shown). Similarly, neither the FTI nor ΔFTI were changed by high-fat feeding (Figure 4). The FTI and Δ FTI were not altered by β-adrenergic antagonist treatment in either control or fat-fed animals.

Figure 4.

Femoral artery FTI in response to saline and acetylcholine. (A) Control and (B) fat-fed rats with vehicle pellets are compared with those that received β-adrenergic antagonist pellets. The increase in FTI with β-adrenergic antagonism was not significant (control, P < 0.11; fat-fed, P < 0.17).

The absolute change in resistance and percentage change from baseline with sodium nitroprusside administration were not significantly different between groups.

Discussion

High-fat dietary treatment was associated with increased body weight and higher serum FFA concentrations. This effect was accompanied by a metabolic profile reflecting the obese state; namely, hypertriglyceridemia, hyperleptinemia, and hyperinsulinemia. Unexpectedly, fat-fed rats did not have higher serum inflammatory markers or attenuated EDV. Also counter to expectations, β-adrenergic antagonism was associated with increased serum FFA, but with a concomitant decrease in serum inflammatory markers, in both lean and obese rats. Furthermore in both dietary groups, rats that received β-adrenergic antagonist treatment had preserved EDV despite increased FFA.

Although much information exists regarding the effects of β-adrenergic modulation on lipolysis and serum triglycerides, very little is known about the effects of long-term β-adrenergic antagonism on serum FFA. In the present study, elevated FFA with β-adrenergic antagonism may have been due to compensatory activation of alternative lipolytic pathways. The adenylyl cyclase–protein kinase A (PKA) pathway recruited for β-mediated lipolysis is also stimulated and inhibited by β-independent factors.²⁸ In addition, a report of fasting lipolysis in mice lacking adipose β-adrenergic receptors²³ suggests that alternate pathways of sympathetic nervous system-mediated lipolysis exist. Moreover, there is evidence of β-mediated, PKA-independent lipolysis involving the ERK–MAPK pathway,^41,45 as well as a β- and PKA-independent pathway mediated by atrial natriuretic peptide in humans.⁴⁴ Alternative lipolytic systems, and the degree to which compensatory regulation of these pathways may occur, remain to be elucidated.

The principle cause of increased FFA seen with β-adrenergic antagonism in the present study may have been attenuation of brown fat thermogenesis, a process that requires fatty acid utilization. In brown adipose, stimulation of the β₃-adrenergic receptor enhances thermogenesis by increasing uncoupling proteins.⁵² Consistent with this, administration of the β₃-adrenergic agonist CL316243 produced a profound decrease in FFA in obese Zucker rats.²⁷ The change in FFA in response to this potent lipolytic agent occurred despite decreased body weight and white adipose mass (that is increased white adipose lipolysis) and was attributed to the profound increase in fatty acid oxidation in the brown adipose. Consistent with this idea, obese rats given CL316243 had brown fat hypertrophy and increased factors associated with fatty acid oxidation.²⁰ In the present study, exposure to β-adrenergic antagonists may have had the opposite effect, resulting in attenuated white adipose lipolysis with a relatively larger inhibition of fatty acid uptake and oxidation in the brown adipose, producing a net increase in serum FFA.

The lack of association of high-fat feeding with increased CRP, IL6, or TNFα in this study lies in contrast to increased inflammatory cytokine concentrations observed in obese humans^14,53 and genetically obese rats¹⁹ but is consistent with a study that failed to observe increased plasma TNFα concentrations in rats fed a 35% kcal fat diet for 10 wk.³ Differences in rodent models (that is, rats genetically prone to obesity versus normal rats that are fat-fed) render extrapolation difficult; however, it should be considered that rodent models of dietary obesity may not reflect the proinflammatory obese state in humans. Alternatively, different dietary fatty acid composition²⁹ or greater dietary fat content (that is 50% kcal fat)⁴ might have augmented the inflammatory response. Indeed, a 10% reduction in dietary fat to 30% kcal corrected the insulin insensitivity induced by a 40% kcal fat diet;¹⁶ it is reasonable to consider that dietary fat content may similarly affect other sequelae of high-fat feeding.

Unexpectedly, increased FFA, associated with either high-fat feeding or β antagonism, did not attenuate EDV. This observation may be related to our findings that suggest β antagonists have antiinflammatory action. This result is consistent with a report of reduced TNFα gene expression and protein production in infarcted rat myocardium after metoprolol treatment³⁹ and with the association of β antagonism and reduced CRP in patients with cardiovascular disease.²² Therefore, β-adrenergic antagonism may reduce the inflammatory response to elevated circulating FFA.

The absence of endothelial dilator dysfunction with high-fat feeding and increased adiposity observed in this study lies in contrast to published work describing the effect of short-term high-fat feeding on isolated arterial rings.³³ As with inflammation, differences in the dietary fatty acid composition or absolute fat content may account for the apparent discrepancies between studies of diet and endothelial function.²⁹ Short-term high-fat dietary treatment attenuates ex vivo EDV in rats in the absence of concomitant obesity,³³ lending support to the idea that diet-derived fats and their composition may determine EDV. Consistent with the idea that exogenous fatty acids impact EDV, measurement of dilator function postprandially may reveal greater dysfunction.

Lastly, the lack of aberrant EDV with high-fat feeding may reflect a lack of measurable response to this condition in this particular vascular bed. Vasodilatory mechanisms are highly specific to stimulus and to the location and size of the vessel.^15,34,58 Although the flow response used to quantitate EDV in this study perhaps was not sufficiently sensitive, the trend toward improved EDV with β-adrenergic antagonism observed in both dietary treatment groups (control, P < 0.11; fat-fed, P < 0.17) is consistent with known beneficial effects of these agents on rat vasculature.²⁴ Given the inherent variability, contributed by both subject and investigator, of prolonged in vivo flow determination studies, a larger sampling size will be required to determine significance. Because 2-way repeated-measures ANOVA did not indicate an interaction effect, power analysis was conducted by averaging over dose. The FTI data for each of the 4 treatment groups were analyzed by using 2-sample unequal-variance t tests, and within-group standard deviations were estimated from the observed data. If a biologically relevant true mean difference of 40 mL between treatments is assumed, then a sample size of 14 fat-fed rats and 27 control rats would be required to achieve 0.80 power.

In conclusion, we found that β-adrenergic antagonism unexpectedly increased serum FFA in both control and fat-fed rats, but EDV was unchanged, and inflammatory cytokine concentrations were reduced. If FFA contribute to endothelial dilator dysfunction in obesity, then β-adrenergic antagonist treatment may attenuate the proinflammatory effects of FFA, a protective mechanism unrelated to a direct reduction in serum FFA concentrations. The absence of inflammation and attenuation of EDV with 16 wk of a moderate fat diet indicates that further investigation of in vivo effects of targeted dietary treatment is warranted, so that functional changes can be assessed in light of systems interrelationships in the intact animal.

Efficient, feasible, and clinically relevant study of the obesity epidemic requires a reliable in vivo model that can be subjected to repeated measures and interventions reflective of the human condition. Studies of ex vivo EDV must be validated in the intact rat, and dietary composition must be controlled carefully and representative of dietary intake in affected humans. To this end, long-term studies are in progress to compare the effects of varied dietary fat composition on cardiovascular inflammation and apoptosis in rats. In addition, the noninvasive modality of ultrasonography is being used to perform sequential echocardiographic studies of cardiac systolic and diastolic function and to investigate the measurement of flow-mediated vasodilation of peripheral arteries.