11β-Hydroxysteroid dehydrogenase type 1 contributes to the regulation of 7-oxysterol levels in the arterial wall through the inter-conversion of 7-ketocholesterol and 7β-hydroxycholesterol

The atherogenic 7-oxysterols, 7-ketocholesterol (7-KC) and 7β-hydroxycholesterol (7βOHC), can directly impair arterial function. Inter-conversion of 7-KC and 7βOHC has recently been shown as a novel role for the glucocorticoid-metabolizing enzyme 11β-hydroxysteroid dehydrogenase type 1 (11β-HSD1). Since this enzyme is expressed in vascular smooth muscle cells, we addressed the hypothesis that inter-conversion of 7-KC and 7βOHC by 11β-HSD1 may contribute to regulation of arterial function. Incubation (4–24 h) of aortic rings with either 7-KC (25 μM) or 7βOHC (20 μM) had no effect on endothelium-dependent (acetylcholine) or -independent (sodium nitroprusside) relaxation. In contrast, exposure to 7-KC (but not to 7βOHC) attenuated noradrenaline-induced contraction (Emax) after 4 h (0.78 ± 0.28 vs 0.40 ± 0.08 mN/mm; p < 0.05) and 24 h (2.28 ± 0.34 vs 1.56 ± 0.48 mN/mm; p < 0.05). Both 7-oxysterols were detected by GCMS in the aortic wall of chow-fed C57Bl6/J mice, with concentrations of 7-KC (1.41 ± 0.81 ng/mg) higher (p = 0.05) than 7βOHC (0.16 ± 0.06 ng/mg). In isolated mouse aortic rings 11β-HSD1 was shown to act as an oxo-reductase, inter-converting 7-KC and 7βOHC. This activity was lost in aorta from 11β-HSD1−/− mice, which had low oxysterol levels. Renal homogenates from 11β-HSD1−/− mice were used to confirm that the type 2 isozyme of 11β-HSD does not inter-convert 7-KC and 7βOHC. These results demonstrate that 7-KC has greater effects than 7βOHC on vascular function, and that 11β-HSD1 can inter-convert 7-KC and 7βOHC in the arterial wall, contributing to the regulation of 7-oxysterol levels and potentially influencing vascular function. This mechanism may be important in the cardioprotective effects of 11β-HSD1 inhibitors.


Introduction
Pro-atherogenic 7-oxysterols form a large component (40%) of oxidized LDL (oxLDL), of which 7-ketocholesterol (7-KC) contributes w30% [1]. 7-KC is toxic to cells in the vessel wall, and can impair arterial function ex vivo [2]. Indeed, 7-KC and its metabolite 7b-hydroxycholesterol (7bOHC) inhibited endothelium-dependent, acetylcholine-induced relaxation of rabbit aortic rings in vitro [2]. In human umbilical vein endothelial cells (HUVECs), 7bOHC and 7-KC reduced the histamine-activated release of nitric oxide [3]. This inhibition of endothelial function by 7-oxysterols appears to be independent of their cytotoxic effects [4], but its mechanism is unclear. Importantly, 7-KC and 7bOHC differ in their proatherogenic potential, with 7-KC implicated as the major proinflammatory and cytotoxic oxysterol [5]. However, any differences between the functional effects of 7-KC and 7bOHC in the vasculature have not been addressed.
The balance between 7-KC and 7bOHC in tissues may be actively modulated. Recently, a novel route of metabolism of 7-oxysterols has been described, involving the enzyme 11b-hydroxysteroid dehydrogenase (11b-HSD) type 1. The primary role of 11b-HSD1 is to catalyse the pre-receptor generation of glucocorticoids, allowing tissue-specific amplification of glucocorticoid receptor activation [6]. Inactivation of glucocorticoids is catalysed by the type 2 isozyme of 11b-HSD (11b-HSD2) [7]. It is becoming increasingly apparent that 7-oxysterols are alternative substrates for 11b-HSD1 [8,9], and inhibition of the enzyme can result in accumulation of 7-KC [10]. Since both isozymes of 11b-HSD are present in the arterial wall [11e14], where they are able to inter-convert glucocorticoids [15], it is conceivable that inter-conversion of 7-oxysterols by these enzymes has a role in modulating vascular function.
We used mice with targeted disruption of the 11b-HSD1 gene (Hsd11b1) to investigate the hypothesis that 11b-HSD1 metabolises 7-oxysterols in the arterial wall, thus influencing 7-KC-and 7bOHCmediated modulation of arterial function.

Chemicals and stock solutions
All solvents were HPLC grade (Fisher, Hemel Hempstead, UK) and were prepared containing an anti-oxidant (0.01% w/v butylated hydroxytoluene (BHT)) to prevent oxidative degradation of the lipids [3]. Steroids and oxysterols were from Steraloids (Newport, Rhode Island, USA), derivatization reagents from Fluka (Buchs, Switzerland), tissue culture reagents from Lonza (Reading, UK) and other chemicals from Sigma-Aldrich (Poole, Dorset, UK). Deuterium-labelled internal standards for GCMS were obtained from CDN Isotopes (Qmx Laboratories, Essex, UK). Stock solutions (30 mg/ml in ethanol with 250 mg/ml BHT) of 7-KC, 7bOHC and 7aOHC (an optical isomer of 7bOHC) were freshly prepared as required. All steroids were prepared in 100% ethanol. Working solutions for tissue culture were prepared in standard Dulbecco's modified Eagle's medium (DMEM). Working solutions (25 mM 7-KC; 20 mM 7bOHC) for myography were prepared by diluting the appropriate stock solution in DMEM without L-Arginine (Arg) or phenol red, but containing 1% charcoal-stripped foetal calf serum. These were the maximum concentrations of 7-oxysterols that could be achieved without sample precipitation. The final concentration of vehicle (ethanol with 250 mg/ml BHT) was <0.2%.
Aortae for functional investigation were removed from mice, placed in PBS (37 C), cleaned of peri-adventitial fat and used for myography. Aortae for oxysterol analysis were processed as described below.
For extended exposures [14], aortic rings were placed in a 24 well plate and immersed in 1 ml DMEM (without L-Arg) containing either 7-KC (25 mM), 7bOHC (20 mM), or vehicle (ethanol with 50 mg/ml BHT) and incubated overnight in a humidified incubator (37 C; 5% CO 2 ). These vessels were then mounted in a myograph and functional studies performed, as described above, in the continued presence of the appropriate 7-oxysterol or vehicle.

In kidney
Murine kidneys contain both isoforms of 11b-HSD. Homogenates of kidneys from Hsd11b1 À/À mice (which lack 11b-HSD1) were used as a source of murine 11b-HSD2, with kidneys from C57Bl6/J mice as controls. Kidneys were homogenized in phosphate buffer as detailed [10]. Homogenates (400 mg/ml) were incubated with 7-oxysterols (20 mM) and the appropriate cofactor (2 mM): NAD þ or NADP þ for dehydrogenase reactions; NADH or NADPH for reductase reactions. In all assays conversion of dexamethasone (Dex) and 11dehydrodexamethasone (11-DHDex; 40 mM) was used as a positive control for confirmation of 11b-HSD isozyme activity [28].
A capillary gas chromatograph (Trace GC, Thermo) was coupled to an ion-trap, Polaris Q (Thermo, Hemel Hempstead, UK) mass spectrometer (MS) and equipped with a BPX5 capillary column Limits of detection were assigned as 3:1 signal to noise ratio. Compounds were quantified by the ratio of area under peak of interest to area under peak of internal standard against a standard curve.

Quantitation of steroids by high pressure liquid chromatography
Radio-labelled glucocorticoids were separated by reverse phase HPLC (Symmetry C8 column maintained at 35 C; column length, 15 cm, internal diameter 4.6 mm, pore size 5 mm, Waters, Edinburgh, UK) and quantified by on-line liquid scintillation counting (2 ml/ min; GoldFlow, Meridian, Surrey, UK). Total run time was 35 min (elution times of epi-cortisol, 11-dehydrocorticosterone and corticosterone were typically 12 min, 21 min and 31 min, respectively, with mobile phase of water:acetonitrile:methanol (60:15:25) at 1 ml/min). Dex and 11-DHDex were separated using a mobile phase of water:acetonitrile:methanol (55:20:25) at 1 ml/min with typical retention times for epi-cortisol (10 min),11-DHDex (12 min) and Dex (16 min). UV detection of all steroids was achieved at 240 nm and epi-cortisol was used as an internal standard. Steroids were quantified by the ratio of area under peak of interest to area under peak of internal standard against a standard curve.

Statistical analysis
All data are mean AE standard error of the mean (SEM) where n indicates the number of different animals. Values were compared using unpaired Student's t-tests or 1-way ANOVA with Dunnett's multiple comparison post-tests, as appropriate. All analyses were performed using Graph Pad Prism v5.0 (GraphPad Software Inc. San Diego, USA). Statistical significance was assumed when p < 0.05.

7-Oxysterols are metabolized by 11b-HSD1 but not by 11b-HSD2
As expected [15], glucocorticoids were inter-converted by incubation with intact mouse aortic rings. The velocity of reduction of 11-dehydrocorticosterone to corticosterone (Fig. 3A) proceeded considerably (w10Â) faster than the dehydrogenation of corticosterone to 11-dehydrocorticosterone. Reduction of 11dehydrocorticosterone was attenuated in mice lacking 11b-HSD1, whereas deletion of this enzyme produced only a small (though significant) increase in the dehydrogenation of corticosterone (to 11-dehydrocorticosterone) (Fig. 3A). The oxysterols 7-KC and 7bOHC were also inter-converted by incubation with intact mouse aortic rings. In contrast to glucocorticoids, however, the velocities of reduction of 7-KC (to 7bOHC) and of dehydrogenation of 7bOHC (to 7-KC) were similar following incubation with mouse aortic rings (Fig. 3B). Genetic disruption of Hsd11b1 significantly reduced the velocity of conversion of both 7-KC and 7bOHC (Fig. 3B), with 96 AE 6% of added substrates being recovered. 7-KC was not interconverted with 7aOHC in aortic rings (data not shown).

Discussion
This study shows for the first time that 11b-HSD1, but not 11b-HSD2, catalyses the conversion of 7-oxysterols in the vascular wall. Previous work has shown that murine and human 11b-HSD1 Incubations had no effect on acetylcholine (ACh)-mediated relaxation (C, D) whereas 7-KC (E) (but not 7bOHC (F)), produced a trend towards increased sodium nitroprusside (SNP)-mediated relaxation (p ¼ 0.054). Relaxations were expressed on a scale where the response to 5-HT represented 100% and return to baseline was expressed as 0%. All points represent mean AE SEM, compared by 1-way ANOVA with Tukey's post hoc test, n ¼ 6e8. converts 7-KC to 7bOHC in the liver and in cultured adipocytes [8,9]. We provide evidence that murine 11b-HSD1 reduces 7-KC to 7bOHC in the vessel wall but, furthermore, that it also oxidizes 7bOHC to 7-KC. Use of Hsd11b1 À/À mice confirmed that 11b-HSD1 was the sole enzyme responsible for metabolism of 7-KC and 7bOHC in the aortic wall and that deletion of 11b-HSD1 alters vascular 7-oxysterol concentrations. Functional investigations showed differential effects of 7-KC and 7bOHC on vascular function, suggesting that this 11b-HSD1-mediated inter-conversion of 7oxysterols may influence 7-KC-mediated inhibition of arterial contraction.
7-KC and 7bOHC have both been shown previously to inhibit endothelium-dependent vasorelaxation [32], cause endothelial cell death, and induce production of radical oxygen species [17,33]. This is consistent with the ability of oxidized lipids to impair the endothelium-dependent relaxation of aortic segments from hyperlipidaemic mice [34]. The lack of impact of exposure to oxysterols on endothelium-dependent relaxation was surprising given the previous indications that both 7-KC and 7bOHC inhibit endothelial function [2,3,32] ex vivo. One possible explanation for lack of effect on vasorelaxation is the use of a low concentration of 7oxysterol (20e25 mM) compared with previous studies (180e 270 mM; [2,3,32]). The concentrations used for our investigations were the highest we could achieve without precipitation and are consistent with that used (25 mM) to show 7-oxysterol-mediated smooth muscle apoptosis in vitro [35]. Furthermore, a recent investigation using high concentrations of 7-KC (205 mM) found no effect of ex vivo incubation on ACh-mediated relaxation of mouse aorta [36].
Intriguingly at the concentrations used in this investigation, there was an inhibition of smooth muscle cell contraction by 7-KC that was not observed with 7bOHC. The mechanism involved is unclear but the effect was selective for noradrenaline, suggesting an impairment in the a 1 -adrenoceptor signalling pathway. Impaired contractility is consistent with 7-KC at this concentration having detrimental effects on vascular smooth muscle cells [35]. These results suggest, therefore, that the balance of 7-KC and 7bOHC may have functional and structural implications in the arterial wall.
The concentrations of 7-oxysterols in the vessels of C57Bl6/J mice are consistent with those reported previously in human plasma and vessels [1,24]. Since circulating 7-oxysterols can be sequestered by cells in the vessel wall [37], we assessed the potential of vascular 11b-HSD1 to inter-convert 7-oxysterols in this tissue. Plasma 7-oxysterol levels were not altered in Hsd11b1 À/À mice although total plasma cholesterol was substantially lower. Consistent with previous reports of reduced intra-vascular cholesterol accumulation with inhibition of 11b-HSD1 [38], we found lower levels of all 7-oxysterols in the aortae of Hsd11b1 À/À mice. It was, therefore, difficult to assess intra-vascular 7-KC:7bOHC ratios, since 7-KC levels in particular were near to the detection limit, but the data suggest that 7-KC levels are disproportionately reduced in Hsd11b1 À/À mice, consistent with the enzyme acting predominantly as an oxidase (converting 7bOHC to 7-KC) in vivo.
The ex vivo incubation of aortic rings described here has not previously been used to assess inter-conversion of 7-oxysterols. This approach confirmed that the stability of 7-oxysterols can be preserved during incubation, as both 7-KC and 7bOHC were successfully recovered from DMEM. It had been postulated that 7oxysterols may be taken up by the vessels during incubation but the percentage recovery of 7-oxysterols from reaction mixtures did not support this. Preparation of concentrated stock solutions of the 7oxyserols proved unexpectedly difficult, despite using published methodology [3], with oxysterols precipitating at high concentrations. Based on our own experiences and advice from other groups 7-oxysterol solutions were prepared in DMEM containing FCS containing an antioxidant (BHT; to prevent oxidative degradation of the lipids [3]). It is unlikely that BHT would have a detrimental effect on vascular function as it did not alter histamine-induced NO production in cultured HUVECs [32].
Ex vivo assays clearly demonstrated that incubation of 7oxysterols with mouse aortic rings results in the conversion of 7bOHC to 7-KC and 7-KC to 7bOHC, but not inter-conversion of 7aOHC and 7-KC. This is consistent with results generated in rats [9,10] and humans [39] but contrasts with the demonstration that 11b-HSD1 in hamsters can inter-convert 7aOHC and 7-KC [40]. The ability of 11b-HSD1 to inter-convert 7-oxysterols explains why carbenoxolone, a non-selective 11b-HSD inhibitor, attenuates 7-oxysterol metabolism in rat hepatic microsomes [10]. Interestingly, in contrast to the predominant reductase direction (11-dehydrocorticosterone to corticosterone) shown for metabolism of glucocorticoids, murine vascular 11b-HSD1 showed similar activity as both reductase (7bOHC to 7-KC) and dehydrogenase (7-KC to 7bOHC) for inter-conversion of oxysterols, consistent with previous reports in liver [9,39]. Under these assay conditions, the reaction velocity for inter-conversion of oxysterols was considerably (approximately 10-fold) higher than for reduction of 11dehydrocorticosterone. This contrasts with the demonstration of similar reaction velocities observed in other studies [9,40] and is likely to be a consequence of study design as substrate concentrations were higher (w800Â) for the oxysterols than for the glucocorticoids.
Residual metabolism of glucocorticoids in aortae from Hsd11b1 À/À mice is consistent with vascular 11b-HSD2 expression [14,20]. Virtually no residual inter-conversion of 7bOHC and 7-KC Table 1 Exposure to 7-oxysterols caused an agonist-selective inhibition of contraction, but had no effect on relaxation, of mouse aortic rings in vitro. was observed in aortae from mice lacking 11b-HSD1. Lack of 7oxysterol metabolism by 11b-HSD2 was confirmed using kidney homogenates (since the kidney is rich in 11b-HSD2 [15]; using kidneys from Hsd11b1 À/À mice ensured that there was no interference from this isozyme). This finding is consistent with the previous attribution of 7-oxysterol metabolism solely to the action of 11b-HSD1 in hamster [40], rat [9,10], guinea pig [9,41] and human [39]. There was, however, a notable loss of substrate in the reaction mixtures; suggesting incomplete recovery of substrate, non-enzymatic degradation, or formation of alternative products [42]. There was no loss of substrate in blank samples (containing buffer but no tissue homogenate), confirming chemical stability of 7-oxysterols during the incubation.

A) Short (4 h) incubation
Direct action on the cells of the arterial wall may not present the only mechanisms through which oxysterols can influence regulation of arterial function and structure. Previous work in our group [43] has indicated that the ability of oxysterols to act as substrates for 11b-HSD1 also makes them potential competitive inhibitors of glucocorticoid metabolism. This presents the possibility that endogenous 7-oxysterols contribute to regulation of 11b-HSD1dependent glucocorticoid generation. Glucocorticoids can interact directly with the arterial wall to enhance vasoconstriction [44], impair endothelium-dependent relaxation [45], inhibit angiogenesis [27] and reduce vascular lesion formation. There is increasing evidence that these interactions are regulated by the activity of 11b-HSD1 [27,38]. However, it is notable that no systematic Incubations had no effect on acetylcholine (ACh)-mediated (C, D) or sodium nitroprusside (SNP)-mediated (E, F) relaxation. All points represent mean AE SEM, compared by 1-way ANOVA with Tukey's post hoc test, n ¼ 6e8. difference in vascular function has been observed in vessels from Hsd11b1 À/À mice [20], so whether alterations in either 7-oxysterol or glucocorticoids influences physiological vascular function remains uncertain. Perhaps interactions of oxysterols with 11b-HSD1 are more important in pathology. In healthy individuals, the maximum concentrations of 7-oxysterols [46,47] are lower than those in patients with atherosclerosis who may have levels of 7oxysterols in the micromolar range [1]. It is plausible that inhibition of 11b-HSD1-mediated glucocorticoid generation in conditions of 7-oxysterol excess may have an indirect impact on arterial function and remodelling.
Metabolism of 7-oxysterols by 11b-HSD1 may also have implications for the development of new therapies. Selective 11b-HSD1 inhibition prevents atherosclerosis [38] and is being developed for treatment of cardiovascular risk factors [48], but the mechanisms responsible for this atheroprotective effect have not been demonstrated. It is conceivable that the beneficial effects of 11b-HSD1 inhibition are a consequence of prevention of 7-oxysterol interconversion as well as glucocorticoid metabolism.

Conclusions
11b-HSD1 influences 7-oxysterol concentrations within the arterial wall. By altering the balance of 7-ketocholesterol and 7bhydroxycholesterol, 11b-HSD1 may modulate their specific effects on vascular function, especially in disease states in which oxysterol levels are increased.