Aerobic exercise-induced irisin secretion is associated with improved endothelial function and reduced atherosclerosis in ApoE-deficient mice

Abstract

Exercise training improves endothelial function and reduces vascular inflammation. However, whether aerobic exercise training-induced secretion of irisin, a myokine cleaved from fibronectin type III domain-containing protein 5 (FNDC5), is associated with endothelial nitric oxide synthase (eNOS)-mediated improvements in endothelial function and inflammation, thereby affecting atherosclerosis, remains unclear. In this study, apolipoprotein E-deficient (ApoE^−/−) mice, an atherosclerotic disease model, were assigned to sedentary (ApoE^−/− SED) or aerobic exercise training (voluntary wheel running for 16 weeks; ApoE^−/− AT) groups, with sedentary wild-type (C57BL/6J) mice (WT-SED) as healthy controls. Atherosclerotic lesion area, aortic interleukin-6 and tumor necrosis factor-α protein expression, and intercellular cell adhesion molecule-1 and vascular cell adhesion molecule-1 mRNA levels were significantly higher, whereas acetylcholine-induced vasorelaxation and aortic eNOS expression were significantly lower in ApoE^−/− SED mice than in WT-SED mice. Aerobic exercise training improved these parameters in ApoE^−/− mice. Soleus FNDC5 mRNA expression and plasma irisin levels were reduced in ApoE^−/− SED mice but increased in ApoE^−/− AT mice. Circulating irisin levels positively correlated with soleus FNDC5 mRNA and aortic eNOS expression. These findings suggest that aerobic exercise-induced irisin secretion is associated with improved endothelial function and reduced vascular inflammation, possibly involving eNOS, which suppresses atherosclerosis.

Introduction

Atherosclerosis is a chronic, progressive vascular disease that results in cardiovascular events and remains the leading cause of death worldwide¹. Endothelial cells regulate the vascular tone, circulating blood cell adhesion, inflammation, and smooth muscle proliferation^2,3. Functional impairment of the endothelium leads to atherogenesis by disrupting normal vascular homeostasis. Reduced endothelial function impairs vasorelaxation and creates a pro-inflammatory vascular phenotype characterized by enhanced monocyte adhesion and infiltration into the arterial wall. These infiltrated monocytes subsequently differentiate into macrophages, internalize lipogenesis, and form foam cells, thereby promoting the development of atheromatous plaques^2,3. Therefore, endothelial dysfunction may be a critical risk factor during the early stages of the progressive process of atherosclerosis^2,4. Endothelial dysfunction is primarily characterized by the decreased bioavailability of nitric oxide (NO) produced by endothelial NO synthase (eNOS)^2,4. An improvement in endothelial NO production suppresses the development and progression of atherosclerosis⁵. Thus, it is clinically important to identify strategic therapies that target improvements in endothelial function. Regular aerobic exercise is recommended to reduce the risk of atherosclerotic vascular disease^6,7. Several human and rodent studies involving obesity and advanced age have reported that regular aerobic exercise promotes endothelium-dependent vasorelaxation through NO production^8–11. Similarly, in apolipoprotein E-deficient (ApoE^−/−) mice, a model of atherosclerosis, habitual aerobic exercise improves endothelial function, inhibits inflammation, and increases eNOS protein expression in the aorta^12–14. Hemodynamic shear stress on endothelial cells accelerates NO synthesis as a mechanism for increasing endothelial NO production. Additionally, arterial eNOS activation is regulated by several peptide hormones, indicating crosstalk between blood vessels and other organs through multiple cytokines^15–18. However, the crosstalk mechanism underlying exercise-induced suppression of atherosclerosis is not fully understood.

Hundreds of myokines are secreted from skeletal muscle and act in autocrine, paracrine, or endocrine manners on the bone, fat, brain, liver, blood vessels, and other organs¹⁹. In 2012, irisin was identified as an exercise-induced myokine cleaved from fibronectin type III domain-containing protein 5 (FNDC5)²⁰. Our study showed that irisin secretion promoted by aerobic exercise training might be associated with a reduction in arterial stiffness and fat accumulation in humans and rats with obesity or advanced age^11,21. In addition, chronic administration of exogenous irisin improved the circulating lipid profile and endothelium-dependent vasorelaxation through NO production^22–24. A recent study showed that regular aerobic exercise increases skeletal muscle FNDC5 mRNA expression and circulating irisin levels in ApoE^−/− mice²⁵. Furthermore, in ApoE^−/− mice, administration of irisin has been shown to attenuate atherosclerosis, accompanied by increased endothelium-dependent vasorelaxation and reduced vascular inflammation^26,27. Skeletal muscle-specific peroxisome proliferator-activated receptor γ coactivator-1α overexpression, regulator of FNDC5 expression, suppresses atherosclerosis with increase in FNDC5 mRNA expression²⁸. However, it remains unclear whether aerobic exercise training-induced endogenous irisin secretion is related to improvements in endothelial function and a reduction in inflammation, potentially involving eNOS, in the context of atherosclerosis.

We hypothesized that the aerobic exercise training-induced increase in irisin secretion might be associated with the suppression of atherosclerosis. This study aimed to clarify whether aerobic exercise training-induced irisin secretion from skeletal muscles is associated with improvements in endothelial function and arterial inflammation, possibly through changes in eNOS levels, in ApoE^−/− mice.

Materials and methods

Animals

This study was approved by the Committee on Animal Care at Ritsumeikan University (BKC2016-019) and was conducted in accordance with our institutional guidelines and the ARRIVE guidelines (https://arriveguidelines.org/). Eight-week-old male ApoE^−/− mice (The Jackson Laboratory, Bar Harbor, ME, USA) were used as an atherosclerotic disease model and randomly divided into the following two groups: (1) sedentary control (ApoE^−/− SED, n = 13) group and (2) aerobic exercise training (ApoE^−/− AT, n = 13) group. ApoE^−/− mice were fed a high-fat diet (41% kcal from fat, Research Diets, New Brunswick, NJ, USA) throughout the experimental period. Age-matched wild-type mice (C57BL/6J, n = 13; Jackson Laboratory Japan, Kanagawa, Japan) were fed a standard chow diet (13% kcal from fat; Oriental Yeast, Tokyo, Japan) and served as a healthy sedentary control group (WT-SED). ApoE^−/− mice and WT-SED mice were individually housed under controlled environmental conditions (12:12 h light-dark cycle) with ad libitum access to water and their respective diet throughout the 16-week experimental period. All the mice were sacrificed after 24 weeks. Sex and estrous cycle–related estrogen fluctuations can affect endothelial function^29,30; therefore, potential variability was reduced by restricting experiments to male ApoE^−/− mice to clarify the irisin–endothelium association.

Aerobic exercise training protocol

Aerobic exercise training in the ApoE^−/− AT group consisted of voluntary angled-wheel running using a low-profile disc-shaped wheel with a solid plastic running surface (diameter 15.5 cm, Med Associates Inc., Fairfax, VT, USA) for 16 weeks, beginning at 8 weeks of age. The mice had no access to running wheels before 8 weeks of age, after which they had ad libitum access to wheels in their individual cages. The running distance was electronically monitored and recorded daily by a magnetic switch interfaced to a computer using Wheel Manager Data Acquisition Software (Med Associates Inc., Fairfax, VT, USA)³¹.

Experimental protocol

The body weights of 12-h fasted mice were measured, and fasting blood samples were collected from the orbital eye sinus under deep general anesthesia induced by inhalation of 2.0% isoflurane (Viatris Pharmaceuticals, Tokyo, Japan), followed by euthanasia by cervical dislocation performed while the mice were deeply anesthetized. After sacrifice, the soleus and quadriceps femoris muscles and epididymal fat were quickly excised, rinsed in ice-cold saline, weighed, frozen in liquid nitrogen, and stored at −80 °C for further analysis. The aorta was resected from the ascending region to the aortic bifurcation, and the perivascular adipose and connective tissues were carefully removed. The aorta was used for the assessment of arterial function, stained with Oil Red O, embedded in OCT compound (Sakura Finetek, Torrance, CA, USA), and quickly frozen for hematoxylin-eosin (HE) staining or stored at −80 °C for real-time PCR. Post-intervention experiments were conducted more than 48 h after the running wheels were removed to avoid any acute effects of the aerobic exercise.

Citrate synthase activity

Quadriceps femoris muscle tissues from each group were homogenized in 10 volumes of buffer containing 250 mmol/L sucrose, 1 mmol/L Tris·HCl (pH 7.4), and 130 mmol/L NaCl using a Teflon homogenizer. Citrate synthase (CS) activity, initiated by the addition of oxaloacetate, was measured spectrophotometrically using a microplate reader (xMark Microplate Spectrophotometer; Bio-Rad Laboratories, Hercules, CA, USA) as previously described³¹.

Lipid measurement

Plasma triglyceride levels (FUJIFILM Wako Pure Chemical Corporation, Osaka, Japan) and low-density lipoprotein (LDL) cholesterol concentrations were measured using a Cholesterol Assay Kit-HDL and LDL/VLDL (Abcam, Cambridge, UK), according to the manufacturer’s instructions.

Oil red O staining

The aorta, from the descending region to the aortic bifurcation, was fixed in 10% neutral-buffered formalin (4% formaldehyde; FUJIFILM Wako Pure Chemical Corporation, Osaka, Japan) for 24 h at room temperature. The samples were incubated with Oil Red O solution (Muto Pure Chemicals, Tokyo, Japan) for 20 min at 37 °C. After incubation, samples were washed with 60% isopropanol (FUJIFILM Wako Pure Chemical Corporation, Osaka, Japan). Images of each aorta were captured using a digital camera (EOS Kiss X10; Canon, Tokyo, Japan) equipped with a macrolens (EF-S 60 mm F2.8 Macro USM; Canon, Tokyo, Japan). The percentage of the lesion area relative to the total lumen area was quantified using ImageJ software (version 1.54 g; NIH, Bethesda, MD, USA)³².

HE staining

The ascending aorta, which was embedded in OCT compound (Sakura Finetek Torrance, CA), was sectioned into 8 μm slices and stained with hematoxylin (FUJIFILM Wako Pure Chemical Corporation, Osaka, Japan) for 3 min and eosin (FUJIFILM Wako Pure Chemical Corporation, Osaka, Japan) for 1 min at room temperature. Cross-sectional images of the aorta were visualized at ×10 magnification using a microscope (BZ-9000; Keyence, Osaka, Japan). The percentage of the lesion area relative to the lumen area was determined using ImageJ software (version 1.54 g; NIH, Bethesda, MD, USA).

Real-Time RT-PCR

Total RNA was extracted from the soleus muscle, thoracic aorta, and abdominal aorta using Isogen reagent (Nippon Gene, Toyama, Japan) and either the RNeasy Mini Kit for the soleus muscle (QIAGEN, Hilden, Germany) or the ReliaPrep RNA Miniprep Systems for aorta samples (Promega Corporation, Madison, WI, USA), and then reverse-transcribed using OmniScript Reverse Transcriptase (QIAGEN, Hilden, Germany) according to previous studies¹¹. FNDC5 mRNA expression in the soleus muscle and vascular cell adhesion molecule-1 (VCAM-1) and intercellular cell adhesion molecule-1 (ICAM-1) expression in the aorta were quantified by real-time PCR (Prism 7500 Fast Sequence Detection System 2.2, Applied Biosystems, Carlsbad, CA, USA) using TaqMan Gene Expression Assays (FNDC5: Mm01181543_m1; VCAM-1: Mm_01320970_m1; ICAM-1: Mm_00516023_m1; Applied Biosystems, Carlsbad, CA, USA), as previously described¹¹. FNDC5, VCAM-1, and ICAM-1 mRNA expression levels were normalized to β-actin mRNA (Mm02619580_g1, Applied Biosystems, Carlsbad, CA, USA). All reactions were performed in duplicates.

Sandwich enzyme immunoassay

Plasma irisin levels were determined using a sandwich enzyme-linked immunosorbent assay (ELISA) kit (Phoenix Pharmaceuticals, Burlingame, CA, USA). The optical density was measured at 450 nm using a microplate reader (xMark Microplate pectrophotometer, Bio-Rad Laboratories, Hercules, CA, USA). Irisin concentrations were calculated using a four-parameter logistic fit of the log-log plot of the standard curve, as previously described¹¹.

Assessment of endothelial function

Vascular function was assessed using a wire myograph, as previously described³¹. Aortas were harvested and cleaned of perivascular adipose tissue in ice-cold physiological saline solution (PSS; pH 7.4), cut into 2-mm segments, and the proximal thoracic aortic rings were mounted on wires for isometric force measurements. The rings were placed in 5 mL baths containing physiological Krebs–Henseleit solution containing 130 mmol/L NaCl, 4.7 mmol/L KCl, 1.16 mmol/L CaCl₂, 1.17 mmol/L MgSO₄, 1.18 mmol/L NaH₂PO₄, 5.5 mmol/L glucose, and 0.026 mmol/L EDTA. The solution was aerated with 95% O₂ to 5% CO₂ (pH 7.4) and maintained at 37 °C for 1 h to allow for equilibration, during which the bath solution was changed every 20 min. The rings were stretched to an optimal resting tension of 20 mN^33,34 and then pre-constricted with 30 nmol/L prostaglandin F₂α³⁵ (PGF₂α; U-46619, Cayman Chemical, Ann Arbor, MI, USA). Endothelium-dependent vasorelaxation in response to acetylcholine (ACh; 10^− 9 to 10^− 4 mol/L) and endothelium-independent vasorelaxation in response to sodium nitroprusside (SNP; 10^− 9 to 10^− 4 mol/L) were measured using isometric force transducers³⁶. Both ACh and the SNP were administered cumulatively to the same aortic ring without washout between concentrations. Vasorelaxation at each contraction was expressed as a percentage of maximum relaxation, where 100% represents complete loss of tension induced by PGF₂α.

Immunohistochemistry

The aortic arch, embedded in OCT compound (Sakura Finetek, Torrance, CA, USA), was sectioned into 8 µm-thick slices and incubated with primary antibodies against eNOS (1:100, BD Biosciences, Franklin Lakes, NJ, USA), tumor necrosis factor-α (TNF-α; 1:100, GeneTex, Irvine, CA, USA), and interleukin-6 (IL-6, 1:100, GeneTex, Irvine, CA, USA) for 12 h at 4°C. After three washes with phosphate-buffered saline (FUJIFILM Wako Pure Chemical Corporation, Osaka, Japan), the sections were incubated with Alexa Fluor 488-conjugated anti-mouse secondary antibody for eNOS (1:400, Invitrogen, Waltham, MA, USA) or Alexa Fluor 594-conjugated anti-rabbit secondary antibody for TNF-α and IL-6 (1:400, Invitrogen, Waltham, MA, USA). Finally, the sections were washed three times and counterstained with 4’,6-diamidino-2-phenylindole (DAPI; Invitrogen, Waltham, MA, USA)³⁷. Fluorescence images were acquired using a fluorescence microscope (BZ-9000, Keyence Corp., Osaka, Japan). The expression levels of IL-6 and TNF-α in the vessel walls were quantified by measuring the red fluorescence intensity with ImageJ software (version 1.54 g; NIH, Bethesda, MD, USA)³⁸. These data were obtained via blind counting by 3 persons.

Statistical analysis

Values are mean ± standard error (SE). One-way analysis of variance (ANOVA) was used to compare differences among the WT-SED, ApoE^−/− SED, and ApoE^−/− AT groups. When ANOVA indicated a significant effect, Fisher’s post-hoc test was performed to adjust for multiple comparisons. The expression levels of IL-6 and TNF-α in the vascular wall were compared across groups via Kruskal-Wallis tests with Dunn’s multiple comparisons test. Pearson’s correlation coefficients were used to assess the relationships between circulating irisin levels and muscle FNDC5 mRNA expression, aortic eNOS protein expression, LDL cholesterol and triglyceride levels, and atherosclerotic lesion areas determined by Oil Red O staining across the three groups. Statistical significance was defined as p < 0.05. All statistical analyses were performed using the StatView software (version 5.0; SAS Institute, Tokyo, Japan).

Results

In the ApoE^−/− SED group, body weight and epididymal fat mass were significantly greater, while CS activity and average food intake were significantly lower than in the WT-SED group (p < 0.05, Table 1). Body weight and epididymal fat mass were significantly lower, whereas soleus muscle mass, CS activity, and average food intake were significantly higher in the ApoE^−/− AT group than in the ApoE^−/− SED group (p < 0.05, Table 1).

Table 1.

Animal characteristics.

	WT-SED	ApoE−/− SED	ApoE−/− AT
Body weight (g)	30.4 ± 0.9	37.9 ± 0.7#	30.9 ± 1.7†
Epididymal fat mass (mg)	791.8 ± 96.2	1881.4 ± 84.1#	936.5 ± 151.8†
Soleus muscle mass (mg)	15.6 ± 0.5	17.0 ± 0.6	20.1 ± 0.8#†
QF muscle mass (mg)	314.7 ± 10.8	347.3 ± 15.3	317.3 ± 13.7
CS activity in QF muscle (µmol/min/g protein)	10.0 ± 0.8	3.7 ± 0.4#	7.4 ± 0.7#†
Food intake (g/day)	3.04 ± 0.07	2.80 ± 0.04#	3.07 ± 0.07†
Running distance (km/week)	-	-	62.6 ± 4.8

Values are mean ± standard error (SE). WT-SED: sedentary wild-type group, ApoE^−/− SED: sedentary apolipoprotein E-deficient group, ApoE^−/− AT: aerobic exercise-training apolipoprotein E-deficient group, QF: quadriceps femoris, CS: citrate synthase. #p < 0.05 vs. WT-SED, †p < 0.05 vs. ApoE^−/− SED.

Plasma total cholesterol, plasma HDL cholesterol, plasma LDL cholesterol, and triglyceride levels were significantly higher in the ApoE^−/− SED group than in the WT-SED group (p < 0.05, Table 2), whereas plasma total cholesterol and plasma LDL cholesterol levels were significantly lower in the ApoE^−/− AT group than in the ApoE^−/− SED group (p < 0.05, Table 2).

Table 2.

Lipid profile.

	WT-SED	ApoE−/− SED	ApoE−/− AT
TC (mg/dl)	134.9 ± 6.4	1043.0 ± 9.5#	968.0 ± 29.5#†
HDL-C (mg/dl)	38.3 ± 2.4	371.6 ± 41.4#	356.9 ± 26.9#
LDL-C (mg/dl)	45.6 ± 6.6	319.8 ± 26.8#	219.8 ± 38.4#†
TG (mg/dl)	10.7 ± 3.4	51.6 ± 16.3#	39.2 ± 12.4#†

Values are mean ± standard error (SE). WT-SED: sedentary wild-type group, ApoE^−/− SED: sedentary apolipoprotein E-deficient group, ApoE^−/− AT: aerobic exercise-training apolipoprotein E-deficient group, TC: total cholesterol, HDL-C: high-density lipoprotein cholesterol, LDL-C: low-density lipoprotein cholesterol, TG: triglyceride. #p < 0.05 vs. WT-SED, †p < 0.05 vs. ApoE^−/− SED.

The atherosclerotic lesion area, as determined by Oil Red O and HE staining, was significantly greater in the ApoE^−/− SED group than in the WT-SED group (p < 0.05, Fig. 1). In contrast, the lesion area was significantly smaller in the ApoE^−/− AT group than in the ApoE^−/− SED group (p < 0.05, Fig. 1).

Fig. 1.

Percentage of atherosclerotic lesion area determined by Oil Red O staining from the descending aorta to the aortic bifurcation (each group: n = 3) and hematoxylin-eosin (HE) staining in the ascending aorta (each group: n = 6) in sedentary wild-type (WT-SED), sedentary apolipoprotein E-deficient (ApoE^−/− SED), and 16-week exercised ApoE^−/− (ApoE^−/− AT) mice. Representative images (A, C) and quantified lesion area percentages (B, D) are shown. Values are mean ± SE. Yellow arrows indicate atherosclerotic lesions. Scale bar = 250 μm.

Soleus FNDC5 mRNA expression and plasma irisin levels were significantly lower in the ApoE^−/− SED group than in the WT-SED group (p < 0.05, Fig. 2) and significantly higher in the ApoE^−/− AT group than in the ApoE^−/− SED group (p < 0.05, Fig. 2).

Fig. 2.

Soleus fibronectin type III domain-containing protein 5 (FNDC5) mRNA expression (A) and plasma irisin levels (B) in sedentary wild-type (WT-SED), sedentary apolipoprotein E-deficient (ApoE^−/− SED), and 16-week exercised ApoE^−/− (ApoE^−/− AT) mice (each group: n = 13). Values are mean ± SE. A.U., arbitrary units.

PGF₂α-induced pre-constriction at 30 nmol/L did not differ significantly among the three groups (data not shown). ACh-induced vasorelaxation (endothelium-dependent vasodilation) was significantly lower in the ApoE^−/− SED group than in the WT-SED group (p < 0.05, Fig. 3) and significantly greater in the ApoE^−/− AT group than in the ApoE^−/− SED group (p < 0.05, Fig. 3). However, no significant differences in SNP-induced vasorelaxation (endothelium-independent vasodilation) were observed among the three groups (Fig. 3). Moreover, aortic eNOS expression levels were significantly lower in the ApoE^−/− SED group than in the WT-SED group (p < 0.05, Fig. 3) and significantly higher in the ApoE^−/− AT group than in the ApoE^−/− SED group (p < 0.05, Fig. 3).

Fig. 3.

Acetylcholine- (ACh, A) and sodium nitroprusside- (SNP, B) induced vasorelaxation (each group: n = 3), representative immunofluorescence images showing endothelial nitric oxide synthase (eNOS) localization in the arterial wall (C), and quantified arterial eNOS protein expression (D, each group: n = 5) in sedentary wild-type (WT-SED), sedentary apolipoprotein E-deficient (ApoE^−/− SED), and 16-week exercised ApoE^−/− (ApoE^−/− AT) mice. In panel C, eNOS immunoreactivity (green) is predominantly localized to the luminal-most cell layer of the arterial wall, corresponding to the endothelial layer. Nuclei were counterstained with DAPI (blue). White arrows indicate eNOS-positive staining for eNOS. Values are mean ± SE. †p < 0.05 vs. WT-SED; #p < 0.05 vs. ApoE^−/− SED. Scale bar = 25 μm.

Aortic IL-6 and TNF-α expression levels were significantly higher in the ApoE^−/− SED group than in the WT-SED group (p < 0.05, Fig. 4), and significantly lower in the ApoE^−/− AT group than in the ApoE^−/− SED group (p < 0.05, Fig. 4).

Fig. 4.

Aortic tumor necrosis factor-α (TNF-α) and interleukin-6 (IL-6) protein expression in sedentary wild-type (WT-SED), sedentary apolipoprotein E-deficient (ApoE^−/− SED), and 16-week exercised ApoE^−/− (ApoE^−/− AT) mice (each group: n = 4). Representative immunofluorescence images (A) and quantified TNF-α (B) and IL-6 (C) expressions are shown. Values are mean ± SE. White arrows indicate positive staining for TNF-α and IL-6. A.U., arbitrary units. Scale bar = 25 μm.

Aortic ICAM-1 and VCAM-1 mRNA expression levels were significantly higher in the ApoE^−/− SED group than in the WT-SED group (p < 0.05, Fig. 5) and significantly lower in the ApoE^−/− AT group than in the ApoE^−/− SED group (p < 0.05, Fig. 5).

Fig. 5.

Aortic intercellular adhesion molecule-1 (ICAM-1, A) and vascular cell adhesion molecule-1 (VCAM-1, B) mRNA expression in sedentary wild-type (WT-SED), sedentary apolipoprotein E-deficient (ApoE^−/− SED), and 16-week exercised ApoE^−/− (ApoE^−/− AT) mice (n = 7–10 per group). Values are mean ± SE. A.U., arbitrary units.

Circulating irisin levels were positively correlated with muscle FNDC5 mRNA expression (p < 0.05, r = 0.455, Fig. 6) and aortic eNOS protein expression (p < 0.05, r = 0.603, Fig. 6), and negatively correlated with plasma LDL cholesterol (p < 0.05, r = −0.611, Fig. 6), plasma triglyceride levels (p < 0.05, r = −0.743, Fig. 6), and atherosclerotic lesion area determined by Oil Red O staining (p < 0.05, r = −0.838, Fig. 6).

Fig. 6.

Pearson correlation analyses between plasma irisin levels and muscle FNDC5 mRNA expression (A, each group: n = 13), aortic eNOS protein expression (B, each group: n = 5), plasma LDL cholesterol levels (C, each group: n = 10), plasma triglycerides levels (D, each group: n = 10), and atherosclerotic lesion area quantified by Oil Red O staining (E, each group: n = 3) in sedentary wild-type (WT-SED), sedentary apolipoprotein E-deficient (ApoE^−/− SED), and 16-week exercised ApoE^−/− (ApoE^−/− AT) mice.

Discussion

In the present study, we investigated whether aerobic exercise training-induced increases in irisin levels are related to vascular alterations in atherosclerosis-prone ApoE^−/− mice. In high-fat diet-fed ApoE^−/− mice, aerobic exercise training was associated with a reduction in the atherosclerotic lesion area. Furthermore, aerobic exercise training was accompanied by decreased protein expression levels of the inflammatory cytokines IL-6 and TNF-α, as well as reduced mRNA expression levels of the adhesion molecules ICAM-1 and VCAM-1 in the aorta. Aerobic exercise also improved endothelial function in ApoE^−/− mice, coinciding with increased aortic eNOS protein expression. Taken together, these findings suggest that aerobic exercise training is associated with improved endothelial function and reduced vascular inflammation, potentially involving the upregulation of eNOS in atherosclerosis-prone ApoE^−/− mice. Interestingly, aerobic exercise training increased circulating irisin levels along with skeletal muscle FNDC5 mRNA expression. Moreover, circulating irisin levels positively correlated with both muscle FNDC5 mRNA and aortic eNOS protein expression. These findings suggest that increased irisin secretion associated with elevated FNDC5 expression in skeletal muscle may be related to improvements in endothelial function and reduced vascular inflammation, possibly involving the eNOS pathway, which may consequently contribute to the attenuation of atherosclerosis progression in ApoE^−/− mice.

NO is a well-established endothelium-derived relaxing factor that plays a critical role in regulating endothelial function³⁹. In this study, aerobic exercise training increased ACh-induced vasorelaxation, but not SNP-induced vasorelaxation. These results suggested that aerobic exercise training is associated with improved endothelium-dependent vasorelaxation, potentially reflecting enhanced ACh-stimulated endothelial NO production². AT also increases aortic eNOS protein expression. Notably, circulating irisin levels were elevated in exercised ApoE^−/− mice and were positively correlated with aortic eNOS protein expression. Previous studies have shown that chronic irisin administration improves endothelial function by increasing NO production in ApoE^−/− mice²⁷. Furthermore, in human umbilical vein endothelial cells (HUVEC), irisin treatment enhances NO production via eNOS activation^24,27. A cross-sectional human study also reported a positive association between circulating irisin levels and endothelial function⁴⁰. Therefore, aerobic exercise training-induced increase in irisin secretion may be related to improved endothelial function, possibly involving changes in eNOS.

The progression of atherosclerosis involves the expression of adhesion molecules, which are induced by inflammatory cytokines or oxidized LDL (oxLDL), leading to the recruitment of monocytes to the vascular wall^41,42. Macrophages that differentiate from monocytes release inflammatory cytokines that contribute to foam cell formation⁴¹. Therefore, the suppression of arterial inflammation is an important therapeutic target in atherosclerosis⁴³. In the present study, aerobic exercise training was accompanied by increased circulating irisin levels and decreased mRNA and protein expression of adhesion molecules and inflammatory cytokines in the aorta. Previous in vitro studies have shown that irisin can reduced inflammatory mediators such as IL-6, TNF-α, VCAM-1 and ICAM-1 in HUVEC^26,44, and chronic administration of irisin to ApoE^−/− mice has been reported to decrease arterial inflammatory mediators^26,27. Based on these findings and those of previous studies, exercise-induced increases in irisin secretion may be related to reduced vascular inflammation. Moreover, vascular inflammation is promoted by the inhibition of NO synthesis and is suppressed by NO donor administration; thus, NO exerts anti-inflammatory effects⁴⁵. In the present study, AT increased the aortic eNOS protein expression. Therefore, the beneficial effects observed in arterial inflammation may involve changes in aortic eNOS-derived NO production.

In the present study, aerobic exercise training was accompanied by lower LDL cholesterol levels and reduced atherosclerotic lesion formation in ApoE^−/− mice. Moreover, aerobic exercise training-induced increases in circulating irisin levels negatively correlated with LDL cholesterol levels. Previous studies have shown that exogenous irisin treatment can reduce LDL levels by enhancing biliary cholesterol excretion and fecal cholesterol output, thereby reducing atherosclerotic lesions²³. Conversely, FNDC5 knockout mice exhibit elevated LDL cholesterol levels and develop hyperlipidemia⁴⁶. Taken together, these findings suggest that aerobic exercise training-induced increase in irisin secretion may be related to reduced LDL cholesterol levels and may contribute to the attenuation of atherosclerosis.

Irisin is primarily secreted from skeletal muscle, and a muscle-specific knockout of peroxisome proliferator-activated receptor γ coactivator-1α, a regulatory factor of FNDC5 gene expression, leads to a 72% decrease in circulating irisin levels in rats²⁰. In contrast, FNDC5 overexpression in the hind limb muscle increased circulating irisin levels in rats⁴⁷. Therefore, FNDC5 expression in skeletal muscle is closely related to circulating irisin levels. In the present study, aerobic exercise training was accompanied by increased FNDC5 mRNA expression in skeletal muscle as well as higher circulating irisin levels. Similar results have been reported in ApoE^−/− mice and obese rats^11,25,48. Furthermore, aerobic exercise training-induced upregulation of FNDC5 mRNA expression was positively correlated with circulating irisin levels in the present study. These results suggest that aerobic exercise training may enhance endogenous irisin secretion via increased FNDC5 expression in the skeletal muscles of ApoE^−/− mice.

The present study showed that circulating irisin levels decreased with atherosclerotic development in ApoE^−/− mice. Consistent with this finding, previous studies have reported lower circulating irisin levels in patients with cardiovascular disease or atherosclerosis than in healthy subjects⁴⁹. Conversely, increased circulating irisin levels may be associated with the prevention of cardiovascular diseases, as irisin levels are inversely related to the progression and severity of coronary atherosclerosis^50,51. In the present study, the increase in circulating irisin levels induced by habitual exercise may be related to improved endothelial function, reduced vascular inflammation, and lower LDL cholesterol levels, which may collectively contribute to the attenuation of atherosclerosis. Circulating irisin levels were also negatively correlated with atherosclerotic lesion area, as determined by Oil Red O staining. A meta-analysis showed that aerobic exercise training increases circulating irisin levels⁵², consistent with the findings of the present study. Thus, maintaining irisin levels through aerobic exercise training may be beneficial in preventing cardiovascular events. High cardiorespiratory fitness levels promoted by habitual aerobic exercise are associated with increased circulating irisin levels and reduced cardiometabolic risk⁵³.

In conclusion, aerobic exercise training was accompanied by improved endothelial function, reduced vascular inflammation, and a smaller atherosclerotic lesion area in HFD-fed ApoE^−/− mice. Circulating irisin levels were increased by aerobic exercise training and were positively correlated with skeletal muscle FNDC5 mRNA expression and aortic eNOS expression, and negatively correlated with atherosclerotic lesion area. Taken together, these findings suggest that aerobic exercise training-induced irisin secretion from skeletal muscles may be related to improved endothelial function and reduced vascular inflammation, possibly involving changes in eNOS, which may consequently contribute to the attenuation of atherosclerosis.

Author contributions

S.F. and M.I. conceived and designed research; K.I., S.F., and M.I. analyzed the data; K.I., S.F., M.U., R.I., K.N., M.K., and M.I. performed experiments; K.I., S.F., M.U., R.I., K.N., and M.I. interpreted results of experiments; K.I., M.K., R.I., K.N., and M.I. prepared the figures; K.I. and M.I. drafted the manuscript; K.I., S.F., M.U., R.I., K.N., M.K., and M.I. edited and revised the manuscript; and K.I., S.F., M.U., R.I., K.N., M.K., and M.I. approved final version of the manuscript.

Funding

This work was supported by the Ministry of Education, Culture, Sports, Science, and Technology of Japan through Grants-in-Aid for Scientific Research (KAKENHI: 25H01105 to M. Iemitsu).

Data availability

The data supporting the findings of this study are available from the corresponding author upon reasonable request.

Declarations

Competing interests

The authors declare no competing interests.