11βOH-Progesterone Affects Vascular Glucocorticoid Metabolism and Contractile Response
Abstract Vascular smooth muscle (VSM) contains a bidirectional isoform of 11β-hydroxysteroid dehydrogenase (11β-HSD), the enzyme that can metabolize endogenous glucocorticoids to their respective 11-dehydro derivatives. 11βOH-progesterone (11βOH-P), a compound that can be produced in vivo, is as potent or more potent than licorice derivatives in inhibiting renal and hepatic 11β-HSD. When studied in homogenates prepared from primary cultures of rat VSM, 11βOH-P and its derivative, 11-keto-progesterone (11-keto-P), proved to be potent, directionally specific inhibitors of vascular 11β-HSD. 11βOH-P selectively inhibited the forward dehydrogenase reaction (corticosterone→11-dehydrocorticosterone), whereas 11-keto-P selectively blocked the reverse oxidoreductase reaction. To test the physiological effects, vascular rings were prepared from rat aorta. Rings were incubated in culture media containing either a submaximal concentration of corticosterone (10 nmol/L), 11-dehydrocorticosterone (100 nmol/L), 11βOH-P (1 μmol/L), 11-keto-P (1 μmol/L), or a combination of glucocorticoid and inhibitor for 24 hours. After the 24-hour incubation, rings were briefly stimulated sequentially with phenylephrine (10 nmol/L to 1 μmol/L) and angiotensin II (1 μmol/L). The immediate contractile response in rings incubated with both corticosterone and 11βOH-P was greater than in rings previously incubated with either the corticosterone or 11βOH-P alone (eg, response to 100 nmol/L phenylephrine in milligrams of force, mean±SE: corticosterone, 728±56, n=9; 11βOH-P, 325±105, n=4; both, 1132±122, n=8; corticosterone versus both, P<.01). In contrast, the immediate contractile responses to phenylephrine and to angiotensin II were attenuated in rings exposed previously to both 11-dehydrocorticosterone and 11-keto-P. Thus, 11βOH-P and 11-keto-P (and possibly structurally similar compounds) alter the vascular effects of glucocorticoids and may play a role in glucocorticoid-induced hypertension.
In clinical practice, a prolonged exposure to glucocorticoids has long been associated with a rise in blood pressure. Although glucocorticoids may affect salt and water handling by the kidney, these steroids appear to play their greatest role by influencing vascular resistance.1 2 Kornel and coworkers3 4 have clearly shown that both sodium and calcium ion influx increase in vascular smooth muscle cells after a minimum 6-hour incubation with glucocorticoids and that the rate of flux reaches a maximum within 48 hours. Moreover, vascular smooth muscle cells exposed to glucocorticoids for a similar time period demonstrate a greater number of catecholamine and angiotensin II cell receptors.5 6 7 8 Both of these observations in vascular tissue correlate temporally with enhanced vascular tone.
The physiological effect of endogenous glucocorticoids (corticosterone in the rat and cortisol in humans) is modulated, at least in part, by the presence of the enzyme 11β-HSD in liver,9 kidney,10 11 and vascular smooth muscle cells.7 12 13 Vascular 11β-HSD is bidirectional; thus, regulation of enzyme direction and enzyme expression are both important in controlling local glucocorticoid metabolism. Licorice derivatives have been the most widely studied of the exogenous 11β-HSD inhibitors. This group of inhibitors appears to affect the forward dehydrogenase reaction (corticosterone→11-dehydrocorticosterone) to a far greater degree than the reverse oxidoreductase reaction (11-dehydrocorticosterone→corticosterone).14 15
The discovery of “directional” specificity in exogenous 11β-HSD inhibitors has raised the possibility that endogenously produced “licorice-like” factors might exist. Recently, Takeda et al16 demonstrated that pregnenolone can be converted to corticosterone, deoxycorticosterone, 18-OH-corticosterone, or aldosterone in vascular smooth muscle cells. Furthermore, these investigators used reverse-transcription polymerase chain reaction to show that steroidogenic enzymes including CYP 11B1 (cytochrome P450 steroid 11β hydroxylase) and CYP 11B2 (aldosterone synthase) are expressed in vascular smooth muscle.17 CYP 11B1 expression is increased in the mesenteric arteries of adrenalectomized rats, likely secondary to high ACTH levels. The conversion of pregnenolone to the endogenous glucocorticoid corticosterone means that 11β-hydroxylated steroid intermediates, including 11β-OH progesterone, could be produced in vascular tissue. In addition, 11β-OH-progesterone and 11-keto-P have been isolated in venous drainage from the adrenal gland in animal studies18 and can be seen in humans under certain clinical conditions, eg, 17-hydroxylase deficiency. 11β-OH progesterone and related compounds are of particular interest because these agents are potent inhibitors of 11β-HSD dehydrogenase activity.19 The present studies, conducted in rat vascular tissue, clearly show that 11β-OH progesterone and its 11-keto metabolite inhibit vascular 11β-HSD in a directionally specific way and that the inhibition can be correlated with altered contractile response to known vasoconstrictors.
Primary Vascular Smooth Muscle Cell Cultures
The aortae from eight adult Sprague-Dawley rats were dissected under sterile conditions, and associated fat and connective tissue were removed. The aortae were next incubated at 37°C for 15 minutes with digestion mixture that consisted of collagenase (1 mg/mL), elastase (0.125 mg/mL), DMEM/F12 (Gibco BRL) with HEPES (pH 7.4), penicillin (100 μg/mL), and streptomycin (100 μg/mL). At the completion of the incubation, the adventitia was carefully stripped, the vessel longitudinally opened, and the luminal surface gently scraped with forceps to remove endothelial cells. The remaining tissues were then minced into 1-mm pieces and incubated with digestion mixture for an additional 90 minutes at 37°C by use of a shaking water bath. The digestion was stopped with 8 mL of DMEM/F12, 10% fetal calf serum, and 25 mmol/L HEPES. After digestion, the remnant tissue became a suspension of smooth muscle cells. The cells were passed over a sieve and centrifuged at 1000 rpm for 5 minutes. The supernatant was discarded, and the cells were resuspended in DMEM/F12, 10% fetal calf serum, and 25 mmol/L HEPES. The cells were then plated at a concentration of 1×104 viable cells/cm onto T-25 plates and placed in an incubator. The flasks were gassed with 5% CO2 and maintained at 37°C. Cells studied up to 40 passages have been shown to retain the expression of α-actin, a factor unique to vascular smooth muscle cells.
Cell counts were obtained before each plating and before any experiment. The medium for the cells was changed twice per week. Studies were usually conducted on the cells after 5 to 7 days, when the cells had reached confluence and were in a quiescent stage. Cells were made quiescent by removing all growth factors, including fetal calf serum, from the incubation medium for 48 hours before study.
11β-HSD Enzyme Kinetics and Directionality
Experiments were conducted with either homogenates prepared from the cultures of primary rat vascular smooth muscle cells or endothelium-intact 3-mm aortic vascular rings. Homogenates of the cultured cells were made in isotonic HEPES buffer at pH 7.4 and 290 mOsm/kg in the presence of the protease inhibitors leupeptin and aprotinin (Sigma Chemical Co) (0.001 mg/mL for each). The protein concentration of the homogenate was ≈0.5 mg/mL. The cofactor NADP+ (200 μmol/L) was added to homogenates when the dehydrogenase reaction was being measured and 200 μmol/L NADPH was added for studies assessing the oxidoreductase reaction. All samples were incubated at 37°C. Corticosterone or 11-dehydrocorticosterone was used as the substrate in these studies in concentrations ranging from 10 nmol/L to 1 μmol/L for the kinetics studies, with 10 nmol/L 3H-labeled and the remainder unlabeled steroid. Incubation times were for 30 minutes unless otherwise indicated in the homogenate experiments. The rings were incubated for 24 hours as outlined for the contractile response studies.
The reaction was stopped with the addition of methanol (1 mL) for all the experiments. With intact tissue, steroids and their respective metabolites were extracted into the methanol over a 24-hour period. For the homogenate studies, samples were centrifuged at 3600 rpm for 10 minutes after the addition of the methanol. The steroids present in the supernatant were separated by high-pressure liquid chromatography with a DuPont Zorbax C8 column eluted at 44°C at a flow rate of 1 mL/min using 60% methanol for 10 minutes. We observed the various steroid compounds by monitoring radioactivity on-line with a Packard Radiomatic Flo-One/Beta Series A-500 counter connected to a Dell Optiplex 425 S/L computer running Windows 3.1 and A505 Flo-One for MacIntosh (version 2.0A). We identified steroids by comparing the retention times with those of known standards. Results of the homogenate experiments were normalized to the protein concentration of the sample (Bradford protein assay, Bio-Rad Laboratories).
In the enzyme kinetics experiments, each data point represented the mean of at least three separate observations from cell homogenates. The Michaelis constant and maximal velocity were calculated from a double reciprocal plot (Lineweaver-Burk plot) drawn using Cricket Graph version 1.3.2 with a line of best fit determined from the data. We calculated the Ki for 11βOH-P and 11-keto-P by plotting the concentration of the progesterone metabolite against the slopes determined from the Lineweaver-Burk plots.
Contractility in Vascular Ring Preparations
After inducing general anesthesia with pentobarbital 50 mg/kg IP, rats underwent a median sternotomy with rapid removal of the thoracic aorta. The adventitia was removed, but the endothelium was left intact. The aorta was cut into 3-mm rings and placed in 2 mL of DMEM/F12 incubated at 37°C under 95% O2-5% CO2 either in the presence or absence of test steroids for 24 hours before study. For the contraction measurements, aortic rings were suspended by tungsten wires with 1 g of tension and placed in a vessel bath containing Earle’s balanced salt solution (Sigma) at 37°C aerated with 95% O2-5% CO2 at pH 7.4. Vessels were equilibrated for 10 minutes and then tested with phenylephrine (0.01 μmol/L to 1 μmol/L) and later a maximal concentration of angiotensin II (1 μmol/L). The intensity of the contraction was assessed by use of a Narishige micromanipulator and model FT03 force transducer (Grass Instrument Co). Measurements were recorded by a model 79D EEG and Polygraph Data Recording System (Grass Instrument Co). Adhering to this protocol, we and others test vessel viability by demonstrating the ability of the vessel to vigorously contract when exposed to known vasoconstrictors and relax back to baseline after treatment with acetylcholine.20
Where appropriate, data were expressed as mean±SE and analyzed by use of a one-way ANOVA to determine differences among the groups. Comparisons of differences within groups were made by the Student-Newman-Keuls method, the Tukey-Kramer test, or Student’s t test; a value of P<.05 was considered significant.
Specific Directional Effects of Progesterone Metabolites
In the first series of studies, homogenates were prepared from cultures of vascular smooth muscle cells. The homogenates containing the appropriate cofactor were then incubated with either corticosterone or 11-dehydrocorticosterone (10 nmol/L) and 11βOH-P, 11-keto-P, 3α,5β-tetrahydroprogesterone, or progesterone at a concentration of 1 μmol/L for 60 minutes (Fig 1⇓). 11βOH-P was selective in its ability to inhibit the forward dehydrogenase reaction while having no influence on the reverse oxidoreductase reaction. In contrast, 11-keto-P suppressed only the oxidoreductase reaction. The 3α,5β-tetrahydroprogesterone appeared to be effective in blocking the enzyme in both directions, whereas the parent compound, progesterone, impeded only the oxidoreductase reaction.
The experiments were then repeated, this time in intact vascular aortic rings. In this case, only the effects of 11βOH-P and 11-keto-P were assessed. Vascular rings were incubated for 24 hours with either corticosterone or 11-dehydrocorticosterone (10 nmol/L); 11βOH-P or 11-keto-P (1 μmol/L) was also present in the experimental groups. Again, a directionally specific pattern of inhibition was observed (see Table⇓). 11βOH-P almost completely suppressed the dehydrogenase reaction while having no effect on the oxidoreductase reaction. 11-Keto-P clearly suppressed the oxidoreductase reaction, but it also was active in partially inhibiting the dehydrogenase reaction after the 24-hour incubation.
Effect of 11βOH-P and 11-Keto-P on Enzyme Kinetics
Kinetics experiments were conducted using homogenates of vascular smooth muscle cells. From Lineweaver-Burk plots, the pattern observed was most consistent with the progesterone metabolites serving as directionally specific competitive inhibitors of the enzyme in vascular tissue (Fig 2⇓). 11βOH-P inhibited the forward dehydrogenase reaction with a Ki of 0.51 μmol/L, and 11-keto-P blocked the reverse oxidoreductase reaction with a Ki of 0.68 μmol/L.
Correlation With the Vascular Contractile Response
To test whether the contractile response of vascular rings would be affected by manipulating the direction of the enzyme, aortic rings with endothelium left intact were incubated for 24 hours in the presence or absence of 11βOH-P (1 μmol/L) alone, a submaximal concentration of corticosterone (10 nmol/L), or the combination of both agents. At the end of the 24-hour incubation, the contractile responses to graded concentrations of phenylephrine and a maximal concentration of angiotensin II were determined. Aortic rings exposed to 11βOH-P alone or corticosterone at a submaximal concentration (10 nmol/L) responded no differently than controls. However, rings exposed to both the corticosterone and 11βOH-P demonstrated a statistically significant increase in the response to 10 nmol/L and 1 μmol/L phenylephrine (Fig 3⇓). Rings previously incubated with both corticosterone and 11βOH-P and then briefly stimulated with angiotensin II (1 μmol/L) also showed an increased contractile response compared with corticosterone alone (Fig 4⇓). Thus, when the local dehydrogenase reaction is inhibited, naturally occurring glucocorticoids appear to be able to enhance the contractile response of vascular tissue to phenylephrine and possibly to angiotensin II.
The next series of experiments was designed to examine whether preventing the local conversion of the inactive metabolite back to the parent glucocorticoid (oxidoreductase reaction) influences the contractile response. Aortic rings were incubated overnight with either 11-keto-P alone (1 μmol/L), 11-dehydrocorticosterone (0.1 μmol/L), or both. The contractile responses of rings incubated with both 11-dehydrocorticosterone and 11-keto-P and then stimulated with phenylephrine and angiotensin II were clearly diminished. Thus, inhibition of the reverse oxidoreductase reaction exerted a potential effect on blood pressure by attenuating the contractile response of these two vasoactive agents (Figs 4⇑ and 5⇓).
The discovery of a bidirectional isoform of 11β-HSD in vascular smooth muscle has raised the possibility that local steroid metabolism is important in mediating vascular contractility. The regulatory processes involved in local steroid metabolism are complex; 11β-HSD activity in vascular smooth muscle can be modulated by enzyme expression, cofactor concentration and redox state, and the presence of licorice-like endogenous inhibitors that are synthesized locally or enter the cell from the circulation. With the observation that vascular tissue has the potential to synthesize steroids,16 11β-OH steroid intermediates formed as precursors of corticosterone seem to be logical candidates for the proposed locally generated licorice-like substances. We now have demonstrated that 11βOH-P and its derivative, 11-keto-P, are clear examples of 11β-OH steroid intermediates that are capable of inhibiting this enzyme in a directionally specific manner. Moreover, the pattern of inhibition correlates well with the biological effects of corticosterone in altering the contractile responses to both catecholamines and angiotensin II.
In experiments conducted on cell homogenates, 11βOH-P and 11-keto-P were directionally specific in their ability to inhibit vascular 11β-HSD. These studies were performed with short incubation times. When more prolonged incubations were performed (24 hours) with endothelium-intact vascular rings, a similar pattern of directionally specific inhibition was observed. 11βOH-P almost completely blocked the dehydrogenase reaction with no effect on the oxidoreductase reaction. 11-Keto-P showed effects on the oxidoreductase reaction as expected and surprisingly also influenced the dehydrogenase reaction. The effects of 11-keto-P on the conversion of corticosterone to its metabolite may be explained by the fact that 11-keto-P, a competitive inhibitor, is likely to serve as a substrate for vascular 11β-HSD and as such could be partially converted back to 11βOH-P over time. The preferential conversion of 11-keto-P back to 11βOH-P is consistent with our previous observation that vascular 11β-HSD is bidirectional, favoring the oxidoreductase reaction in intact cells.13
Like the licorice derivative carbenoxolone, 11βOH-P together with corticosterone has the potential to augment the contractile response of vascular smooth muscle to known vasoactive agents. By preventing the conversion of corticosterone to its inactive metabolite, the glucocorticoid can exert its full enhancing effect. What has been more surprising is the observation that 11-keto-P, when combined with 11-dehydrocorticosterone, can attenuate the contractile response of aortic rings to phenylephrine and angiotensin II. Presumably, blocking the conversion of 11-dehydrocorticosterone back to corticosterone allows for the attenuated effect on vessel contraction. The full clinical relevance of this finding remains to be explored.
The current series of experiments raises the specter that other endogenously produced, licorice-like compounds may exist that affect glucocorticoid action in vascular tissue and possibly sodium handling in the kidney. Glucocorticoid-induced hypertension, then, may be due to a complex of competing processes involving both renal electrolyte reabsorption and vascular reactivity. Both vascular and renal mechanisms will need to be more thoroughly investigated to develop a complete understanding of this important area.
Selected Abbreviations and Acronyms
|K i||=||inhibitory constant|
These studies were supported in part by a Grant-in-Aid from the American Heart Association, Rhode Island affiliate (Dr Brem), and a grant from the NIH (No. HL-52972) to Dr Morris. The authors wish to thank Charlene McGloin for her expert secretarial assistance and Nicholas Hill, MD, for his help in the vascular ring studies.
- Received September 20, 1996.
- Revision received November 12, 1996.
- Accepted February 24, 1997.
Kornel L, Manisundaram B, Nelson WA. Glucocorticoids regulate Na+ transport in vascular smooth muscle through the glucocorticoid receptor-mediated mechanism. Am J Hypertens. 1993;6:736-744.
Ullian ME, Walsh LG. Corticosterone metabolism and effects on angiotensin II receptors in vascular smooth muscle. Circ Res. 1995;77:702-709.
Hatakeyama H, Miyamori I, Fujita T, Takeda R, Yamamoto H. Vascular aldosterone biosynthesis and a link to angiotensin II-induced hypertrophy of vascular smooth muscle cells. J Biol Chem. 1994;269:24316-24320.
Hill NS, Klinger JR, Warburton RR, Pietras L, Wrenn DS. Brain natriuretic peptide: possible role in the modulation of hypoxic pulmonary hypertension. Am J Physiol. 1994;266(Lung Cell Mol Physiol. 10):L308-L315.