Oxysterols are a broad group of bioactive lipids derived from enzymatic and non-enzymatic oxidation of the cholesterol molecule that takes place both intra and extracellularly (Brown and Jessup, 2009). Side-chain oxysterols such as, 24S-hydroxycholesterol, 25-hydroxycholesterol (25-OH C), and 27- hydroxycholesterol (27-OH C), are enzymatically produced and efficient mediate cholesterol synthesis and uptake by regulating the activation of SREBP and the expression of the HMG-CoA reductase and low-density lipoprotein (LDL) receptor, respectively (Radhakrishnan et al., 2008). These oxysterols also maintained cell cholesterol homeostasis by the induction of ABCA1 and ABCG1 gene transactivation, favoring the exportation of cholesterol excess to apo A-I and mature high-density lipoprotein (HDL) and the cholesterol flow through the reverse cholesterol transport (RCT; Venkateswaran et al., 2000). The RCT is an anti-atherogenic system that drives cholesterol from the arterial wall to the liver allowing its excretion into bile and feces.

7-ketocholesterol (7-KC) and 7β-hydroxycholesterol (7β-OH C) may be directly involved in the atherogenesis considering their role in inflammation, endoplasmic reticulum stress and apoptosis and its prevalence in atherogenic oxidized LDL (Gargiulo et al., 2016). Conversely, they can be exported from cells via ABCG-1 transporter (Terasaka et al., 2007). In apo E knockout mice (apo E KO) 80% of the total amount of arterial oxysterols are represented by the compounds derived from carbon-7 cholesterol oxidation (7-KC > 7β-OH C > 7α-OH C; Vaya et al., 2001). In the same animal model, Shentu et al. (2012) demonstrated that 7α-OH C and 7β-OH C were the most abundant forms in the arterial compartment.

Because of their higher polarity as compared to cholesterol the oxysterols diffuse much better through phospholipid membranes. The half-life for 7α-OH C and 7β-OH C exchange between lipid vesicles is 100 times higher than that observed for cholesterol (Kan et al., 1992). Babiker et al. (1997) and Lund et al. (1996) observed that macrophages eliminate large amounts of cholesterol in 27-OH C form, which is mediated by HDL, acting as an additional pathway for the RCT.

Physical exercise is a well-established non-pharmacological strategy in the management of lipid metabolism, reducing plasma triglyceride (TG) concentration and small dense LDL and increasing HDL cholesterol and RCT (Kodama et al., 2007; Rocco et al., 2011; Ghafouri et al., 2015; Pinto et al., 2015). Together with its beneficial effects in other risk factors for cardiovascular disease, exercise training has an important role in preventing, reducing and stabilizing atherosclerotic lesion.

The adenosine monophosphate-activated protein kinase (AMPK) is a classical target of exercise in many tissues including the arterial wall, acting as a sensor of energy levels (Kim et al., 2014). Besides inducing mitochondrial biogenesis, fatty acid, and glucose oxidation, recent studies suggest that AMPK plays an important role in atherogenesis (Li et al., 2010a,b; Chen et al., 2016). There are no data in literature on the role played by exercise training in the amount of cholesterol and oxysterols in plasma and arterial wall and how this relates in vivo with the expression of genes involved in oxysterols metabolism in dyslipidemic apo E KO mice and in vitro with AMPK activation in macrophages. Arterial macrophages are intimately related to the development of atherosclerotic lesion by mediating the uptake of modified LDL but also excess lipid efflux by the RCT. We then analyzed the role of a 6-week AET in the distribution of oxysterol subspecies and total cholesterol (TC) in plasma and aorta of apo E KO mice and the effect of AMPK activation in J774 macrophages in cellular content and efflux of oxysterols.

Body weight and plasma TC, TG, HDL-c, and glucose were similar between T and S groups, at baseline and after intervention (Table 1). Food consumption was higher in T as compared to S (Table 1).

Time to exhaustion determined in treadmill exercise test (expressed as seconds or delta time), was higher in T animals as compared to S. Furthermore, the trained mice improved exercise tolerance when compared to the baseline period (Figures 1A,B). No significant changes were found in the S group.

In plasma, the most prevalent oxysterols were 7-KC and 7β-OH C, representing 34 and 29%, respectively, of the total oxysterol amount. Plasma concentrations of 7α-OH C, 27-OH C, and 25-OH C were similar, representing 18, 12, and 7%, respectively, of total plasma oxysterols. Exercise training did not change plasma oxysterol concentration and their proportions in T as compared to S group (Table 2).

In the aortic arch, there was a higher amount of 7β-OH C in T mice compared to S mice (Table 2). 7-KC and total oxysterol level in the aortic arch were similar between T and S mice. 7-KC levels corresponded to 53% of total oxysterols, followed by 7β-OH C (21%). 27-OH C, and 25-OH C were not detected in the arterial wall. There was a significant reduction in cholesterol levels after 6-weeks of AET in T group as compared to S group (32%; Figure 2).

The expression of Prkaa1 (AMPK α1) and Prkaa2 (AMPK α2)—key proteins in cellular energy metabolism activated by exercise; Cyp27a1—an enzyme that produces 27-OH C; Cd36—a scavenger receptor and Cat (catalase)—an antioxidant enzyme, was higher in T mice as compared to S (Figure 3). On the other hand, AET reduced the expression of Abcg1—a transporter of cholesterol and oxysterol to HDL; Ch25h—an enzyme that produces 25-OH C; Cyp7b1—an enzyme that metabolizes 27-OH C, and Olr1 (LOX-1)—a scavenger receptor that mediates the uptake of oxidized LDL (Figure 3). Abca1, Nr1h3 (LXRα) and Nr1h2 (LXRβ) mRNA were not modified by exercise training in the aorta (Figure 3).

Considering the increase in Prkaa1 and Prkaa2 mRNA by AET, the effect of AMPK activation in macrophage oxysterol levels was evaluated. The J774 macrophages were cultured in the absence or presence of 0.25 mM AICAR (AMPK activator), for 24 h and subsequently incubated in the absence or presence of HDL for 6 h, in order to stimulate sterol efflux. The AICAR concentration used represented the highest concentration that does not induce decrease in cell viability, which was observed by lactate dehydrogenase release in the medium (data not show). AMPK activation in the absence of HDL did not alter the concentration of cellular sterols compared to control cells (Table 3). However, the treatment with AICAR and HDL₂ decreased the macrophage concentration of 7α-OH C but did not change the levels of 7β-OH C, 7-KC, and TC (Table 3). In addition, AICAR treatment increased Abca1 and Cd36 and reduced Prkaa1 and Cyp27a1 expressions. Abcg1, Nr1h,3 and Nr1h2 expressions were not modified by the AMPK activation (Figure 4).

AET has beneficial effects on lipid metabolism, including the RCT, which contributes to the prevention of atherosclerosis (Rocco et al., 2011; Kadoglou et al., 2013). Considering the role of oxysterols in modulating cell lipid homeostasis and inflammation in the arterial wall, we investigated in the apo E KO mice the role of a 6-week AET program in oxysterol and cholesterol accumulation in the aortic arch and plasma, and the expression of genes related to their metabolism in the aortic arch.

The amount of cholesterol and oxysterols is regulated in macrophages by interplay of mechanisms including: the uptake of oxidized LDL, enzymatic and non-enzymatic—mediated synthesis and conversion and cell exportation through passive diffusion and active mechanisms. In this investigation, mice fed a chow diet avoiding the influence of high fat/cholesterol diets as an external source of oxysterols that may be carried by lipoproteins.

Even in the absence of changes in plasma TC, HDL-c, and oxysterols in dyslipidemic mice, AET increased the aortic arch concentration of 7β-OH C. This sterol can be obtained by the macrophage uptake of oxidized LDL rather than intracellular cholesterol oxidation, which in accordance to the enhanced expression of Cd36 observed in trained mice. Despite that, exercise training reduced TC content in the aortic arch—that dictates plaque development and correlates with lesion area—and increased Cat mRNA. These findings agree to the role of AET in increasing RCT and reducing atherosclerotic lesion development as previously reported (Meilhac et al., 2001; Rocco et al., 2011). Acute exercise is associated with an increase in oxysterols. In contrast, chronic exercise has been shown to increase of antioxidant defenses such as, catalase and nitric oxidase synthase (Meilhac et al., 2001; Cardenia et al., 2017). Thus, catalase seems to counteract the increase in oxysterols obtained by oxLDL and cholesterol oxidation, which favors the reduction in atherosclerotic lesion.

Enhanced transcriptional rate of Cd36 after exercise training have been described in humans and animal models as a mechanism for lipid clearance associated with RCT (Butcher et al., 2008; Pinto et al., 2015; Ruffino et al., 2016). In the present investigation, AET distinctly influenced the expression of Cd36 and Olr-1 (LOX1), respectively, increased (75%) and diminished (65%).

To our knowledge, few studies evaluated the influence of exercise training in the formation of oxysterols, with divergent results. Cardenia et al. (2017) observed in rat that acute exercise until exhaustion increased hepatic concentrations of 7α-OH C, 7β-OH C, 7-KC, and 27-OH C, that was prevented by 45 days of rich broccoli extract enriched-diet. On the other hand, Musman et al. (2016) showed that physical training for 4 weeks reduced oxysterols in plasma and isolated mitochondria of ob/ob mice undergoing ischemia/reperfusion. This was due to the decrease of cholesterol content in the mitochondria and to the increase of the antioxidant capacity of the cardiac tissue. Exercise protocol (acute vs. chronic), exercise intensity, and tissue analysis may explain the difference across studies.

Enhanced and decreased expressions of, respectively, Cyp27a1 and Cyp7b1 mRNA, may contribute to the enhancement of 27-OH C, although no detectable levels of 27-OH C were observed in our CG/MS analyses. Low or absent of 27-OH C was also reported in apo E KO mice aorta and macrophages by, respectively, Vaya et al. (2001) and Maor et al. (2000). This may be attributed to the rapid flow of 27-OH C outside cells as an alternative pathway for the RCT, as previously described (Björkhem et al., 1994). Björkhem et al. (1994) observed that macrophages eliminate large amounts of cholesterol as 27-OH C by passive diffusion and Lange et al. (1995) demonstrated that the transfer of 25-OH C from the red blood cell to plasma acceptors occurs 2,000 times faster than cholesterol. Similarly, the spontaneous exchange rate between phospholipids vesicles was more efficient for 7-oxysterols, such as, 7β-OH C, compared to cholesterol (Kan et al., 1992).

In humans with hypoalphalipoproteinemia increase in plasma concentration of 27-OH C has been reported, which may overlap their reduced cell cholesterol efflux rate (Nunes et al., 2014). Independently of genetics and metabolic fate, the HDL cholesterol range in plasma is linked to a 40% variation in 27-OH C associated to HDL particle (Karuna et al., 2011). The ABCA-1 or ABCG-1 knockout bone marrow transplantation to wild type animals reduced the 27-OH C/cholesterol ratio in plasma (Karuna et al., 2011). Nonetheless, the exact role of these transporters in the efflux of 27-OH C to HDL remains unknown. In the present investigation, a reduction in Abcg1 mRNA levels in exercised mice was found, which was interpreted as related to the reduction in cholesterol arterial levels.

By activating AMPK, AICAR systemically modulates the energetic balance and reflects the major effects of physical exercise in many organs including the liver, skeletal muscle, heart, adipose tissue, and the arterial wall compartment as well (Ruderman et al., 2003; Kim et al., 2014). Macrophages are the major cells involved in the pathophysiology of atherosclerosis. They contain scavenger receptors that uptake modified LDL leading to excess cholesterol supplying and to the inflammatory process that accompanies the development of atherosclerosis. In addition, by the RCT macrophages dictate the final amount of cholesterol in the arterial intima-media layer. Considering these aspects and the increased levels of Prkaa1 and Prkaa2 after exercise training in mice, we utilized J774 macrophages in a well-controlled condition in order to test the specific effects of AMPK activation on cellular sterol levels.

Similarly to the aortic arch isolated from trained mice, macrophages treated with AMPK activator presented enhanced levels of Cd36 after incubating with HDL₂, as an extracellular lipid acceptor. On the contrary, a reduction in Cyp27a1 and 7α-OH C was observed, with no changes in other sterols including TC. This was related to an increase in Abca1 mRNA expression that is reported by few authors as a transporter for oxysterols as well as free cholesterol in fibroblasts and neurons (Tam et al., 2006; Matsuda et al., 2013).

Then, the in vitro data do not exactly reflect the biology of the arterial oxysterols under influence of AET and AMPK cannot be defined as a mediator of the exercise effects on oxysterol metabolism and RCT in the experimental conditions utilized by us. An enhanced expression of Abcg1, its protein level and cholesterol efflux elicited by the AMPK activation, was previously demonstrated by others (Li et al., 2010a) but dealing with AICAR concentrations that were found by others (Santidrián et al., 2010; Russe et al., 2014; Chang et al., 2016) and us as cytotoxic. Thus, the exact role of AMPK in mediating the effects of exercise training in the arterial wall lipid homeostasis should be further investigated. Our study was limited by the lack of detectable levels of Cyp7a1 gene expression in the aorta of the mice and Cyp7a1, Cyp7b1, Olr1, and Prkaa2 in cultured macrophages. In addition, the validation of gene expression by measuring protein levels was not performed due to the limitation of the aortic arch extent.

In conclusion, in dyslipidemic mice AET increases the non-enzymatic-driven oxysterol, 7β-hydroxycholesterol, formed by cholesterol autoxidation, which is related to the enhanced expression of Cd36. The rapid diffusion of oxysterols, as a complementary pathway for the RCT, may also be favored by the increase and reduction of Cyp27a1 and Cyp7b1 expressions, respectively, which in turns favors 27-OH C desorption from cells. Together with its direct role in improving RCT as previously reported (Rocco et al., 2011), AET diminishes cholesterol accumulation in the arterial wall preventing atherosclerosis.

GF performed mice exercise training, cell treatment, carried out all analysis with help of the other authors, and participated in the manuscript preparation and experimental design. PP participated in experimental design and helped in mice exercise training and RT-qPCR analysis. RI helped in cell treatment and analysis. VB helped in animal care and analysis. MS helped in oxysterols analysis. VN helped in chromatographic analysis and its data interpretations. CN had a substantial contribution to the initial concept of this paper. SC performed animal surgery, and participated in experimental design; MP was responsible for experimental design, coordination of research and preparation of the manuscript. EN helped in statistics. All authors read and approved the final manuscript.