Perivascular Adipose Tissue Inhibits Endothelial Function of Rat Aortas via Caveolin-1

Perivascular adipose tissue (PVAT)-derived factors have been proposed to play an important role in the pathogenesis of atherosclerosis. Caveolin-1 (Cav-1), occupying the calcium/calmodulin binding site of endothelial NO synthase (eNOS) and then inhibiting nitric oxide (NO) production, is also involved in the development of atherosclerosis. Thus, we investigated whether PVAT regulated vascular tone via Cav-1 and/or endothelial NO pathways. Isometric tension studies were carried out in isolated thoracic aortas from Wistar rats in the presence and absence of PVAT. Concentration-response curves of phenylephrine, acetylcholine, and sodium nitroprusside were illustrated to examine the vascular reactivity and endothelial function. The protein expressions of eNOS and Cav-1 were also examined in aortic homogenates. Our results demonstrated that PVAT significantly enhanced vasoconstriction and inhibited vasodilatation via endothelium-dependent mechanism. The aortic NO production was diminished after PVAT treatment, whereas protein expression and activity of eNOS were not significantly affected. In addition, Cav-1 protein expression was significantly increased in aortas with PVAT transfer. Furthermore, a caveolae depleter methyl-β-cyclodextrin abolished the effect of PVAT on the enhancement of vasoconstriction, and reversed the impairment of aortic NO production. In conclusion, unknown factor(s) released from PVAT may inhibit endothelial NO production and induce vasocontraction via an increase of Cav-1 protein expression.

Most of blood vessels are directly surrounded by adipose tissue, known as perivascular adipose tissue (PVAT) except for cerebrovessels.
Initially, PVAT is considered as structural support for the vasculature, and it is usually removed in isometric vascular tone studies. In 1991, Soltis and Cassis demonstrate that PVAT attenuates vascular response of norepinephrine (NE) in isolated rat aorta preparations as a result of reuptaking NE by adrenergic nerves in PVAT [6]. However, Löhn and his colleagues show that the vascular responses to angiotensin II (Ang II), serotonin and phenylephrine (PE) are reduced in intact aortas with PVAT, not subject to the uptake by adrenergic nerves [7], suggesting that the anti-contractile effect of PVAT is most likely mediated by a transferable PVAT-released material, adventitium-derived relaxing factor (ADRF), which relaxes different kinds of arteries [7]. In contrast, Gao et al. disclose that PVAT enhances the contractile response of mesenteric artery to perivascular nerve stimulation through the production of superoxide [8]. Moreover, PVAT diminishes the vasodilating effect of acetylcholine (ACh) in coronary arteries of lean dogs via phosphorylation of endothelial nitric oxide (NO) synthase (eNOS) and inhibiting eNOS activity and NO production [9,10]. In addition, the paracrine role of PVAT may involve in endothelial dysfunction leading to local stimulation of atherosclerotic plaque formation [11].
Caveolin-1 (Cav-1), a structure protein of caveolae, is present in most of the cells and involved in the development of atherosclerosis [12]. This 22-kDa protein-rich striation, Cav-1, inhibits endothelial function by occupying of the calcium/calmodulin (Ca 2+ /CaM) binding site of eNOS [13]. Caveolae are small flaskshaped invaginations at the plasma membrane enriched with cholesterol, glycosphingolipids and Cav, which interacts with a variety of signal-transducing molecules [13]. In addition, adipose tissue is a main site for cholesterol storage [14], and the crucial role of cholesterol in caveolae biogenesis has been evident for quite some time. However, whether PVAT could (i) regulate caveolae formation, (ii) increase the expression of Cav-1, or (iii) change vascular tension in rat aorta is unclear. Therefore, we tried to examine the role of PVAT in regulation of vascular tone via Cav-1 and/or endothelial NO by using wire myography and protein expression analysis.

Ethics statements
All animals used in this study received humane care in compliance with the animal care guidelines of the institute. This study was performed in accordance with the criteria of the National Academy of Sciences, and approved by the Institutional Animal Care and Use Committee of National Defense Medical Center (Taipei, Taiwan).

Animals
Adult male Wistar rats (250 to 300 g) were provided by BioLasco Taiwan Co., Ltd. (Taipei, Taiwan) and raised in the Laboratory Animal Center of National Defense Medical Center (Taipei, Taiwan). The animals were maintained on a 12-hour light/dark cycle and were given free access to water and standard rat chow.

Vascular reactivity in the thoracic aortas
Segments of thoracic aorta (2 mm) were mounted in bath chambers of the wire myograph (8 mL) (Multi Wire Myograph System-Model 620 M, Danish Myo Technology, Aarhus, Denmark) in Krebs' solution at 37uC, continuously bubbled with 95% O 2 and 5% CO 2, in order to measure the isometric tension (Fig. 1).
The aortas were allowed to equilibrate under zero tension for 15 min to be stretched to their optimal lumen diameter for active tension development. The basic tension was determined based on the internal circumference (IC)/wall tension ratio of the segments by setting the IC 0 to 90% of what the vessels would have if they were exposed to a passive tension equivalent to that produced by a transmural pressure of 100 mmHg (IC 100 ). The diameter (I 1 ) was determined according to the equation I 1 = IC 1 /p [15], using specific software for normalization (DMT Normalization Module, LabChart 7, ADInstruments, New Zealand). In addition, the isolated PVAT was incubated in another bath chamber with Krebs' solution same as the above condition for 30 min before the isometric tension studies. After equilibration for 1 h, the intactness of endothelium was confirmed by ACh (1 mM) in PE (1 mM)precontracted vessels. More than 90% vasorelaxation induced by ACh was considered as no injury of the endothelium. Vascular contraction was assessed by the contractile response to PE (1 nM to 10 mM), and vascular tension was expressed as a percentage of the steady-state tension (100%) induced by external 60 mM KCl. Endothelium-dependent and -independent relaxations in PE (20 nM,1 mM)-precontracted vessels were assessed by measuring the dilatory response to ACh (1 nM to 10 mM) and SNP (0.1 nM to 1 mM), respectively. The vascular dilatation was expressed as a percentage of maximum response to PE (20 nM,1 mM). In some experiments, we replaced 2 mL of the PVAT-incubated solution in a donor bath chamber to the aorta in an acceptor chamber (Fig. 1). In control experiments, transfer of Krebs' solution in a donor bath chamber without PVAT incubation did not affect the tension of aorta in an acceptor chamber ( Fig. 2C and 2D). L-NAME (1, 3, 10, or 100 mM, 30 min), tempol (100 mM, 15 min), losartan (10 mM, 30 min), indomethacin (10 mM, 30 min) and CD (100 mM or 1 mM, 30 min) were incubated in the acceptor bath solution to examine the changes of aortic tension. In addition, we examined effects of high temperature on the substances released from PVAT by heating (70uC) the transfer solution for 5 min that could damage the structure of peptide.

Preparation of aortic homogenates
After incubated with PVAT for 15 min in the presence or absence of inhibitors, aortas were frozen with liquid nitrogen immediately and stored at 280uC before assay. Aortas were homogenized in 16 RIPA lysis buffer with protease inhibitors and phosphatase inhibitor at 4uC for 1.5 h. The aortic homogenates were then centrifuged at 4uC for 30 min at 16,000 g. The homogenates were collected and transferred to new centrifuge tubes to determine the protein concentrations by BCA protein assay kit.

NO measurement
Samples were composed of 2 ml aortic homogenates and 4 ml 95% ice alcohol. The nitrite/nitrate concentrations in all samples were measured by using chemiluminescence through a NO analyzer (Sievers 280 Sievers Instruments, Inc., Boulder, CO, USA). It was noted that the aortic NO concentration depicted in the study was actually the total nitrite and nitrate concentrations in aortas. With this method, nitrate is reduced to NO via nitrite.

Statistical analysis
All data were analyzed with Graphpad Prism software version 5.0 (Graphpad Prism software, Inc., La Jolla, CA, USA). All values were expressed as the mean 6 SEM. The contractile response to PE was expressed as percentage of the maximal contraction induced by KCl (60 mM). Results of isometric tension studies were analyzed by two-way ANOVA analysis. Protein expressions were expressed as a ratio of the optical density for each protein against b-actin or eNOS. Both Protein expressions and nitrate/nitrite levels were analyzed with unpaired Student's t-test. Differences were considered statistically significant at P,0.05.

PVAT-induced vasocontraction in aortas
The PE-induced vasocontraction was in a concentrationdependent manner in aortic rings with or without PVAT. In the presence of PVAT, the concentration-dependent vasocontraction induced by PE was significantly enhanced (P,0.05, Fig. 2A and 2B). Krebs' solution transfer did not influence the vasocontraction induced by PE ( Fig. 2C and 2D).

Effects of PVAT on aortic endothelial function
After denuding the endothelium (-E), the enhanced PE-induced vasocontraction by PVAT was inhibited ( Fig. 2E and 2F). Similarly, in the presence of L-NAME (100 mM, an inhibitor of NOS via inhibiting NO production in the endothelium in our aorta preparations), the enhanced PE-induced vasocontraction by PVAT was also abolished ( Fig. 2G and 2H). Furthermore, we assessed the modulating effect of PVAT on vasodilatation induced by ACh and SNP. Our results showed that PVAT significantly attenuated the ACh-induced vasodilatation (P,0.05, Fig. 2I and 2J), and increased the EC 50 of ACh from 65.666.7 nM to 164.9638.7 nM (P,0.05). We also observed that PVAT had no significant effect on SNP-induced vasodilatation ( Fig. 2K and 2L). These results suggest that inhibition of NO in the endothelium plays an important role in enhanced vasocontraction induced by PVAT.

Inhibitory effect of PVAT on vasodilatation induced by NO
In order to evaluate whether PVAT could inhibit NO activity to enhance vasocontraction, aortic rings were incubated with different concentrations of L-NAME before carrying out the concentration-dependent ACh-induced vasodilatation. The inhibitory effect of PVAT on the ACh-induced vasodilatation was not affected by 1 mM L-NAME pretreatment (P,0.05, Fig. 3A and 3B). However, this inhibitory effect of PVAT on the ACh-induced vasodilatation was abolished by increasing the concentration of L-NAME up to 3, 10, and 100 mM (Fig. 3C -3F). The contributions of PVAT-releasing factors or endothelial regulators in PVAT-induced vasocontraction In contrast to other evidences, the vasocontraction effects of PVAT were not significantly changed after pretreatment of aortas with tempol (a superoxide anion scavenger, 100 mM) (P,0.05, Fig. 4A and 4B) or losartan (an Ang II type 1 receptor antagonist, 10 mM) (P,0.05, Fig. 4C and 4D). This suggests that the PVATinduced vasocontraction is not mediated by superoxide anion and Ang II in the current study. Since prostacyclin (PGI 2 ) and endothelin-1 (ET-1) are known to be major vasoactive substances in the endothelium of thoracic aorta, we treated aortas with indomethacin (an inhibitor of COX, 10 mM) (P,0.05, Fig. 4E and 4F) or damaged the peptide structure of ET-1 by heating the PVAT-incubated solution (70uC for 5 min) (P,0.05, Fig. 4G and 4H). However, the inhibitory effect of PVAT on the ACh-induced vasodilatation was not changed. Thus, PVAT-enhanced vasocontraction was not caused by PGI 2 or through ET-1 from the endothelium.

Effects of PVAT on eNOS expression and activity
Our results showed that the protein expressions of eNOS (Fig. 5A) and two phosphorylated residues (Thr 495 and Ser 1177 ) were not changed after aortas incubated with PVAT transfer (Fig. 5B and 5C). In addition, there was no significant difference in AMPK (a modulator of eNOS activity) protein expression in aortas between with and without PVAT transfer (Fig. 5D). Thus, we suggest that the inhibitory effect of PVAT on vascular dilatation is not through modulating the expression or activity of eNOS.

Role of Cav-1 in enhanced contraction caused by PVAT
It has been shown that Cav-1 can occupy the Ca 2+ /CaM binding site of eNOS to inhibit eNOS activity [13]. Interestingly, the protein expression of Cav-1 was significantly increased in the presence of PVAT (P,0.05, Fig. 6A). The enhanced contractile effect of PVAT on the PE-induced vasocontraction was not affected by 100 mM CD pretreatment (P,0.05, Fig. 6B and 6E).
However, this contractile effect of PVAT on the PE-induced vasocontraction was abolished by increasing the concentration of CD up to 1 mM (Fig. 6C and 6F). In addition, we found that the nitrate/nitrite levels were significantly decreased after PVAT treatment (P,0.05, Fig. 6D), which was reversed by the pretreatment of CD (1 mM) (P,0.05, Fig. 6D).

Discussion
This study presented an inhibitory mechanism of PVAT on aortic endothelial function and we found that Cav-1, a negative regulator of eNOS, played an important role in PVAT-induced abnormal vascular responsiveness, i.e. enhanced vasoconstriction and impaired endothelial vasodilatation.
Adipose tissue is known to release numerous adipokines, and plays a paracrine role toward surroundings [16]. In 2002, Lò`hn et al. demonstrate that the anti-contractile effect of PVAT is mediated by a transferable factor, called ADRF [7]. Since that, the regulation of arterial tone by various adipokines has been confirmed [4,17]. The adipokines with (i) both vasocontractile and vasorelaxing properties include reactive oxygen species [18], leptin [19], TNF-a [20], IL-6, and apelin [16]; (ii) vasocontractile adipokines include Ang II [21] and resistin; (iii) vasorelaxing adipokines include adiponectin, omentin, visfatin, and ADRF. In our study, we also observed the transferable characteristic of PVAT, which enhanced PE-induced contraction in rat thoracic aortas. This observation was similar to Gao and colleagues' finding that PVAT enhances the mesenteric artery contractile response to perivascular nerve stimulation [8]. They consider that the increased contraction is due to the production of superoxide induced by PVAT. However, our study showed that the enhanced PE-induced contraction by PVAT remained unchanged even after inhibition of superoxide anions and Ang II. The opposing results could be due to different methods of these studies used: (i) mesenteric arteries of Wistar-Kyoto rats, (ii) an intact connection between PVAT and the blood vessels, and (iii) electrical field stimulation to elicit vascular contractile response in isometric tension studies. On the contrary, our observation was opposite of the point of ADRF supported by Löhn [7], Gao [22], and Lee [23]. In our current study, we modified Löhn's experiment methods by using the PVAT transfer solution only. The donor chamber in Löhn's (they also showed a tracing about transfer of solution from isolated PVAT evoked vasodilatation; however, the PVAT-induced vasodilatation became not obviously compared with transfer of solution from intact aortic preparation) [7] or Gao's [22] experiments were composed of both PVAT and vessels. In order to eliminate the effects derived from vascular smooth muscle and endothelium from the donor, we used solution with PVAT alone. We found that PVAT-induced vasocontraction and anti-vasodilatation effects in rat thoracic aorta which were different form Lee's (vasorelaxation of PVAT on PE-induced contractions of rat thoracic aortic rings) [23] and Owen's (no effect of PVAT on potassium chloride-induced contractions of rat thoracic aortic rings) [24] findings. Nevertheless, if the unknown factors released from PVAT caused vasodilatation, the antivasodilatation effect should not be observed in Li's [25] and Tune's studies [9,10]. In addition, the PVAT transfer in our study was incubated for 30 minutes independently and then replaced to an acceptor chamber for another 15 minutes for isometric tension study. The unknown factor(s) released from PVAT in our study must have a long-lasting characteristic which was different from Lee's (superfused rat thoracic aorta with PVAT superfusion) [23] and Owen's (PVAT was incubated with rat thoracic aorta for 30 min) [24] groups (the released substances from PVAT in their studies could be predominately short-lasting). Moreover, Li and colleagues claimed that an intact connection between PVAT and the blood vessel was important for transfer of whatever substance involved in this messaging [25]. It has been shown on multiple occasions that the presence of intact PVAT reduces the contractility of arteries to agonists [6,26,27,28,29,30,31]. This has been shown on almost all arteries, ranging from the aorta [6,26,27], mesenteric arteries [25,28], skeletal arteries [29,30], and subcutaneous arteries [31]. The reduced sensitivity to NA was independent of membrane potential, indicating that surrounded PVAT posed a diffusion limitation [25]. However, this concept from Li's [25] group could not explain that there was no effect of PVAT on ACh-induced vasodilation of mice resistance artery [29]. Therefore, the effects of PVAT on vasoreactivity vary among different vascular beds. In addition, when PVAT conditioned fluid is transferred from one myograph bath to another [7,22,23,31,32,33], there is a vasodilatation effect. In our experiments, we adopted a slightly different approach in that we used a transfer solution which was derived from PVAT only -i.e. without the presence of the attached artery [7,23]. In summary, we speculate that the fundamental differences of PVAT function between other publications and this study were owing to (i) the different degradations of substances released from PVAT (shortlasting vs. long-lasting characteristics), (ii) the varied methodology (clean and intact vessels vs. solution transfer experiment), and (iii) the diverse subjects (different species or vessels).
Intriguingly, the PVAT-enhanced PE-induced contraction was abolished by denuding the endothelium in our study. Although the anatomic barrier existed in the endothelium, paracrine effects of adipokines in PVAT may cause endothelial dysfunction (decreased NO production), hypercoagulability, increased chemotaxis and adhesion of monocytes to the endothelium [34]. Therefore, we proposed that the enhanced PE-induced contraction by PVAT could be due to the inhibition of endothelial function. Indeed, the ACh-mediated vasodilatation was reduced after PVAT treatment, whereas SNP-mediated endothelium-independent vasodilatation remained unchanged in this study. This suggests that the inhibitory effect of PVAT on vasodilatation is dependent on the endothelium.
There are several substances from the endothelium involving the regulation of vascular tone, including NO, endotheliumderived hyperpolarizing factor (EDHF), PGI 2 , and ET-1 [35]. It is highly unlikely that EDHF contributes to the vasocontraction effect of PVAT because EDHF plays a major role in the resistance arteries, but not in conductive arteries, in response to vasoactive substances or physical stimuli [36,37]. Since the enhanced contractile effect of PVAT remained unchanged after suppressing the production of PGI 2 by indomethacin, PGI 2 was excluded in this study. Furthermore, after heating PVAT transfer solution to destroy peptides (e.g. ET-1 and Ang II), the enhanced contractile effect of PVAT was not changed, and hence, peptides seemed not to participate in the enhanced contractile effect of PVAT. Interestingly, L-NAME concentration-dependently inhibited the inhibitory effect of PVAT on ACh-induced vasodilatation, indicating that PVAT-induced endothelial dysfunction is through the inhibition of NO/eNOS pathway.
It has been shown that eNOS is regulated by a very complex network of protein kinases (PKA, PKB and PKC), protein phosphatases 1 and 2, cofactors (tetrahydrobiopterin, flavin mononucleotide and nicotinamide adenine dinucleotide phos-phate), protein-protein interactions, and subcellular localization [38]. They regulate the phosphorylate sites of eNOS, e.g., Thr 495 , Ser 1177 , Ser 633 , Ser 615 and Ser 114 residues [38]. In particular, the Ser 1177 residue is thought to be the most critical regulatory site for eNOS activation [38]. In addition, Payne et al. has proved that PVAT impairs coronary endothelial NO production via a PKC-bdependent, site-specific phosphorylation of eNOS at Thr 495 [10]. Nevertheless, in our study, there was no significant difference in eNOS, p-eNOS Thr495 and p-eNOS Ser1177 protein expression after PVAT pretreatment. It has been shown that AMPK activates eNOS on phosphorylate site of Ser 633 [39], however, we did not detect the increased protein expression of AMPK after PVAT treatment. As for the impact of Ser 615 and Ser 114 on eNOS activity is controversial so far [38], and hence, we did not measure expression of Ser 615 and Ser 114 in this study. According to our results, we suggest that the reduction of NO level is not due to an increase of protein expression or activity of eNOS.
Cav-1 is present in most terminally differentiated cells, in particular in endothelial cells, adipocytes, and type II pneumocytes [40]. Cav-1 can inhibit NO production by occupying the Ca 2+ / CaM binding site of eNOS [13]. However, the interaction between Ca 2+ /CaM binding and regulation of phosphorylation in eNOS is still unclear. In this study, we first discovered that PVAT could enhance Cav-1 protein expression. In addition, the administration of CD, a cholesterol depleter to disassemble caveolae, eliminated the enhanced contractile effect of PVAT and reversed the decreased aortic NO level induced by PVAT. Figure 7. A proposed mechanism of perivascular adipose tissue (PVAT) in the regulation of vascular tone through the endothelium. We hypothesized that some unknown factor(s) released from PVAT entered to endothelium and caused caveolin-1 (Cav-1) formation. The Cav-1 probably competes the binding site on endothelial nitric oxide synthase (eNOS) with calcium/calmodulin (Ca 2+ /CaM), which results in eNOS inactivation and inhibition of nitric oxide (NO) production. The NO inhibition leads to enhanced vasoconstriction in rat aortas in isometric tension studies. SMC, smooth muscle cell; EC, endothelial cell. doi:10.1371/journal.pone.0099947.g007 This suggests that PVAT increases Cav-1 protein expression and attenuates endothelial NO production leading to enhancement of contraction in aortas.
The inhibition of NO production is regarded as a characteristic of endothelial dysfunction. It has been shown that the impairment of endothelial function was a risk factor for the development of atherosclerosis [41]. Moreover, perturbations in the NO signaling have been implicated in numerous disease processes ranging from hypertension, shock and inflammation. In addition, PVAT accumulation plays a much more important role in coronary risk than the other body fat distributions in non-obese men [42]. This raises the concern between PVAT, aorta and cardiovascular events now and in the future.
In conclusion, our results provide direct evidence of PVAT elicited enhancement of vasoconstriction in rat aorta and it was endothelium-dependentvia increasing Cav-1 expression and decreasing NO production (Fig. 7). Therefore, our study reveals the connection of PVAT and endothelial dysfunction as an aspect in the future studies of cardiovascular diseases.