Evidence of Coexisting Changes in 11β-Hydroxysteroid Dehydrogenase and 5β-Reductase Activity in Subjects With Untreated Essential Hypertension
Abstract We compared corticosteroid metabolite excretion rates and patterns in a group of 68 subjects with untreated essential hypertension and a matched group of 48 normotensive control subjects. The ratio of tetrahydrocortisol plus allotetrahydrocortisol to tetrahydrocortisone and the ratio of allotetrahydrocortisol to tetrahydrocortisol were significantly higher in the hypertensive group. This is qualitatively similar to the situation found in patients with the syndrome of apparent mineralocorticoid excess or subjects treated with licorice or carbenoxolone where hypertension is known to arise from deficiencies of 11β-hydroxysteroid dehydrogenase and 5β-reductase activities. The equivalent ratios for corticosterone metabolites were not different between groups, but total corticosterone metabolite excretion was higher in the hypertensive group. Plasma cortisol levels were lower in hypertensive than in control subjects, but corticosterone levels were higher. This evidence supports a previous suggestion that the activities of these two enzymes may be reduced in essential hypertension, but the contribution of these changes to hypertension is not known.
Cortisol exerts its glucocorticoid effects through specific glucocorticoid (type II) receptors. In vitro, it also binds with an affinity equal to that of aldosterone to isolated renal mineralocorticoid (type 1) receptors.1 In vivo, this is now known to be avoided by means of target-tissue 11β-hydroxysteroid dehydrogenase (11β-HSD) action, which converts cortisol to cortisone.2 Inherited2 3 4 or drug-induced5 6 deficiency of this enzyme leads to severe hypermineralocorticoid-like changes such as hypertension, suppressed renin (and aldosterone) levels, and hypokalemia. The deficiency is characterized by an abnormally high ratio of cortisol to cortisone metabolites in the urine. Interestingly, these patients, as well as carbenoxolone-treated rats,4 7 also show higher-than-normal levels of 5α-reduced metabolites, suggesting a simultaneous impairment of 5β-reductase activity.
Although patients with essential hypertension have no overt signs of excess mineralocorticoid activity, more subtle changes, such as a clear positive correlation of blood pressure with sodium levels and a negative correlation with potassium levels, have been interpreted as suggesting a corticosteroid influence.8 Recent studies in small groups of hypertensive patients have produced evidence of a slower-than-normal clearance of cortisol9 and an increase in vascular sensitivity to cortisol10 that may be due to altered target-organ 11β-HSD activity. In this article we describe a study of a large group of subjects with untreated essential hypertension by conventional urine analysis that corroborates the 11β-HSD findings but also provides evidence of altered 5β-reductase activity.
Subjects and Procedures
Hypertensive subjects were studied after screening to exclude known causes of secondary hypertension. They were either referred by their general practitioner or attended voluntarily. Subjects were defined as hypertensive when their blood pressure consistently exceeded 140/95 mm Hg for 4 weeks. All blood pressures were measured three times per week between 8 and 10 am and between 6 and 8 pm. On each occasion, six measurements were taken by the same observer in the same room. The subject remained seated. None had ever been treated with antihypertensive drugs, and all other forms of drug therapy had been discontinued at least 6 months before the study. Female subjects were either in the follicular phase or were postmenopausal. None were taking oral contraceptives. Informed consent was obtained from all subjects, and the study complied with local ethical guidelines.
A control group of subjects matched for age, sex, and body mass index was studied simultaneously. In addition to having normal blood pressure, none had a history of hypertension in first-degree relatives.
Urinary Corticosteroid Metabolite Analysis
A 24-hour urine specimen was obtained for measurement of excretion rates of tetrahydrocortisol (THF), alloTHF, tetrahydrocortisone (THE), tetrahydrocorticosterone (THB), alloTHB, and tetrahydro-11-dehydrocorticosterone (THA). The method of Shackleton11 was used with minor modifications. Steroid conjugates were extracted (Sep-Pak C18 cartridges, Waters Chromatography Division, Millipore Corp) and hydrolyzed with Helix pomatia juice (IBF Biotechnics). Steroid metabolites were then extracted, also on Sep-Pak cartridges, and methyloxime trimethylsilyl ether derivatives were synthesized.
Gas chromatography–mass spectrometry analysis was performed on an ITS40 mass spectrometer (Finnigan MAT) coupled to a Varian 3400 gas chromatograph. This was fitted with a fused-silica capillary column (30 m×0.25 mm internal diameter; J&W Scientific) coated with a nonpolar stationary phase (0.25 μm DB5). The temperature program was as follows: 2 minutes at 100°C, 20° per minute to 180°C, 3° per minute to 280°C, and 280° to 60 minutes. To improve separation of the derivatives of THB, alloTHB, and THA, samples were rerun on a BPX 70–coated column (SGE) with a modified temperature program (basal, 100°C; 4° per minute to 300°C; and 300°C to 50 minutes). The helium flow rate was 7 mL/min. Metabolite derivatives were identified from an internal steroid spectrum library and were measured by total ion current. Androstanediol was used as an internal standard.
Plasma samples were taken at approximately 8 am at least 1 hour after an indwelling catheter had been placed into a vein in the right forearm with the subject in the recumbent position. Plasma cortisol concentration was measured by direct radioimmunoassay. Plasma corticosterone concentration was measured by radioimmunoassay after partial purification by paper chromatography. Results are expressed as mean±SEM and were analyzed by Student’s t test.
Table 1⇓ compares basal data for the hypertensive and control groups. The groups were comparable in terms of age, sex, and body mass index. The hypertensive group had significantly higher systolic and diastolic pressures in both upright and recumbent states. Their blood pressure levels define them as having mild to moderate hypertension according to the World Health Organization classification. Their mean heart rate was also significantly higher than control subjects in the recumbent but not in the upright position.
Steroid excretion rate and pattern differed between the groups (Tables 2⇓ and 3⇓). The alloTHF excretion rate was significantly higher in the hypertensive group than in the control group; that of THE was significantly lower. THF excretion was higher in the hypertensive group but not significantly so. The sum of excretion rates of these cortisol metabolites was not different between groups (3850±242 versus 3908±289 μg/24 h, hypertensive versus control group). The ratio of 11β-hydroxy metabolites to 11-dehydro metabolites (THF+alloTHF/THE) was significantly higher in the hypertensive group. This was also the case for the ratio of 5α- to 5β-reduced metabolites (alloTHF/THF). The excretion rates of all the corticosterone metabolites—THB, alloTHB, and THA—were significantly higher in the hypertensive group, resulting in a higher “total corticosterone metabolite” excretion rate (578.8±41.7 versus 335.8±18.8 μg/24 h, hypertensive versus control group). However, the proportion of the three metabolites was not significantly different between groups. In the hypertensive group, 31 subjects had a THF+alloTHF/THE ratio higher than the mean+2 SD (1.53) of the control group. The alloTHF/THF ratio was higher than the mean+2 SD of the control group in only 3 hypertensive subjects.
Plasma cortisol concentration was higher in the control group (382±22 nmol · L−1) than in the hypertensive group (291±19 nmol · L−1, P<.01). The reverse was true for plasma corticosterone concentration (control, 1.73±0.2 nmol · L−1; hypertensive, 3.52±0.47 nmol · L−1, P<.01).
The basic mechanism of the syndrome of apparent mineralocorticoid excess (SAME) has now been satisfactorily explained in terms of reduced 11β-HSD activity.2 In children3 and in a single adult case,2 severe hypertension was accompanied by a marked hypokalemic alkalosis, reduced renin-angiotensin system activity, and very low plasma concentrations of aldosterone. Qualitatively similar changes result from excessive consumption of licorice or treatment with carbenoxolone.5 6 In both of these situations, there is also evidence of altered 5-reductase activity. In SAME patients, Ulick et al4 reported reduced levels of urinary 5-dihydrocorticosteroid metabolites and a relative increase in the proportion of 5α-metabolites. Shackleton et al12 found high alloTHF/THF ratios. Similarly, glycyrrhetinic acid, the 11β-HSD inhibitor from licorice, is a potent inhibitor of rat liver 5β-reductase, causing accumulation of 5α-reduced metabolites.7 3β-HSD activity was also inhibited. This evidence suggests that alterations in 11β-HSD and 5-reductase activities may be linked, but the nature of this link is not known.
Some preliminary evidence shows9 that patients with essential hypertension have reduced 11β-HSD activity compared with normotensive subjects. This observation is based on the slower rate of 3H2O excretion after dosage with [3H]cortisol. Such patients also had increased vascular sensitivity to glucocorticoids.10 However, the patients had no other signs of mineralocorticoid excess. Our urinary analyses may provide further evidence of lower 11β-HSD activity. In a large group of untreated hypertensive patients, the ratio of cortisol to cortisone metabolites, an index of 11β-HSD activity, was significantly higher than in a matched, normotensive control group. Moreover, this was associated with a higher proportion of the 5α-metabolite alloTHF. However, no such differences were seen in the equivalent ratios for corticosterone metabolites. The apparent abnormality of 11β-HSD in the hypertensive subjects was more marked; ie, more ratios exceeded the mean+2 SD of the control group than that for the 5-reductase. In a much earlier comprehensive comparison of urinary corticosteroid metabolites in small groups of subjects, Kornel et al13 found lower levels of THE in the hypertensive group, whereas THF and alloTHF were not different. These results are also indicative of lower 11β-HSD activity but not 5-reductase activity in hypertensive individuals. In contrast to the current study, THB was lower in the hypotensive group, whereas other corticosterone metabolites were unchanged. The reason for the differences between the two studies is not clear but may relate to the different technologies used.
Although our data in hypertensive subjects suggest reduced enzyme activity compared with healthy subjects, the magnitude of the “abnormality” is small compared with that seen in patients with SAME, and gross changes in electrolyte status and the activity of the renin-angiotensin-aldosterone axis are absent. Thus, the significance of our observations and those of Walker et al9 for the development of essential hypertension remain a matter of conjecture. As mentioned earlier, there is some evidence that blood pressure in this condition (but not in healthy subjects) is positively correlated with body sodium levels and negatively correlated with potassium levels, and this has been interpreted as evidence of a mild mineralocorticoid excess.8 Aldosterone and 11-deoxycorticosterone levels are reported to be normal, often with low plasma renin and mild impairment of 11β-HSD activity, and reduced Hβ-HSD might provide an explanation. Alternatively, Walker et al10 have postulated a direct effect at the vascular smooth muscle level, where a lower rate of 11β-HSD action might result in increased exposure of the vascular mineralocorticoid (type 1) receptors to cortisol. It is relevant that dexamethasone, a synthetic glucocorticoid that suppresses cortisol secretion but does not bind to type I receptors, is reported to reduce blood pressure in patients with essential hypertension but not in healthy subjects.14 Morris and colleagues15 have recently isolated but not yet fully characterized compounds from human urine that inhibit 11β-HSD and that they have called glycyrrhetinic acid–like factors (GALFs). It will be of interest to learn what physiological mechanisms control their production rates and whether changes in these rates correlate with 11β-HSD indexes such as steroid metabolite ratios and [3H]cortisol clearance in hypertensive subjects.
It is not clear to us why the pattern of metabolism of the 17-deoxycorticosteroid corticosterone did not change in parallel with that of the 17α-hydroxycorticosteroids. However, it is of interest that although the total excretion of measured cortisol metabolites was not different between groups (plasma cortisol was slightly lower in the hypertensive group), the quantity of measured corticosterone metabolites was higher in the hypertensive group. Plasma corticosterone concentration was also higher in the hypertensive group although still within the laboratory normal range. A possible explanation that the zona glomerulosa supplies a greater proportion of total corticosterone in the hypertensive subjects deserves prospective study. In the Milan hypertensive rat, plasma corticosterone concentration is also raised compared with control animals, although the abnormality is quantitatively much more marked than in the human subjects shown here. Hepatic 11β-HSD activity is reported to be reduced in this rat model of essential hypertension, but the renal enzyme activity is not different between strains.16 Clearly, this discrepancy requires further investigation.
Finally, it is now clear that the renal 11β-HSD on which the variation in mineralocorticoid activity of cortisol depends is a distinct isoenzyme from that of the liver and that their activities may vary independently.17 Although the renal enzyme must be involved in determining urinary metabolite ratios in SAME, its contribution to ratios in essential hypertension patients cannot yet be decided.
This study was supported in part by a research grant from the National Research Council of Italy (CNR) targeted projects “Prevention and control of disease factors,” n.91.00173.41. The manuscript was prepared by Anne McGregor.
Reprint requests to Dr R. Fraser, MRC Blood Pressure Unit, Western Infirmary, Glasgow GII 6NT, Scotland.
- Received June 27, 1994.
- Revision received July 22, 1994.
- Accepted September 23, 1994.
Funder JW, Pearce PT, Smith R, Smith AT. Mineralocorticoid action: target tissue specificity is enzyme, not receptor mediated. Science. 1987;242:583-585.
Stewart PM, Corrie JET, Shackleton CHL, Edwards CRW. Syndrome of apparent mineralocorticoid excess: a defect in the cortisol-cortisone shuttle. J Clin Invest. 1988;82:340-349.
Beretta-Piccoli C, Davies DL, Brown JJ, Ferriss JB, Fraser R, Lever AF, Morton JJ, Robertson JIS. The relation of arterial pressure with plasma and body electrolytes is similar in Conn’s syndrome and essential hypertension. Clin Sci. 1982;63:89s-92s.
Walker BR, Stewart PM, Edwards CRW. 11β-Hydroxysteroid dehydrogenase in essential hypertension. J Endocrinol. 1991;129:282s.
Walker BR, Stewart PM, Padfield PL, Edwards CRW. Increased vascular sensitivity to glucocorticoids in essential hypertension: 11β-hydroxysteroid dehydrogenase deficiency revisited. J Hypertens. 1991;9:1082-1083.
Whitworth JA, Gordon D, McLachlan-Troup N, Scoggins BA, Moulds RWF. Dexamethasone suppression in hypertension: effects of cortisol on blood pressure. Clin Exp Hypertens A. 1989;11:323-335.
Morris DJ, Semafuko WEB, Latif SA, Vogel B, Grimes CA, Sheff MF. Detection of glycyrrhetinic acid-like factors (GALFs) in human urine. Hypertension. 1992;20:356-360.
Stewart PM, Mason JI. Cortisol to cortisone: glucocorticoid to mineralocorticoid. In: Proceedings of the Aldosterone and Hypertension Conference; March 17-20, 1994; Lorne, Australia. Abstract.