Impaired Cortisol Binding to Glucocorticoid Receptors in Hypertensive Patients
Abstract We compared glucocorticoid receptor binding characteristics and glucocorticoid responsiveness of human mononuclear leukocytes (HML) from hypertensive patients and matched normotensive volunteers. We also considered associations of these variables with plasma renin activity, aldosterone, cortisol, corticotropin, and electrolyte concentrations. We calculated binding affinity (Kd; nmol/L) and capacity (Bmax; sites/cell) for dexamethasone and cortisol from homologous and heterologous competition curves for specific [3H]dexamethasone binding sites on HML isolated from the blood of normotensive volunteers and subjects with essential hypertension. Glucocorticoid responsiveness of HML was evaluated as IC50 values (nmol/L) for dexamethasone and cortisol for the inhibition of lysozyme release. We measured plasma hormones by radioimmunoassay. Kd values (mean±SE) for cortisol in HML of hypertensive patients were higher than in control subjects (24.6±2.4 versus 17.5±1.7 nmol/L, P<.04). Binding capacity (4978±391 versus 4131±321 sites/cell), Kd values for dexamethasone (6.7±0.5 versus 5.7±0.3 nmol/L), and IC50 values for dexamethasone (3.4±0.3 versus 3.1±0.2 nmol/L) and cortisol (12.2±1.6 versus 9.5±0.3 nmol/L) were not significantly different. Patients with renin values less than 0.13 ng angiotensin I/L per second were markedly less sensitive to cortisol than those with higher values. Both Kd (30.3±2.5 versus 19.2±2.4 nmol/L) and IC50 values (15.5±1.8 versus 8.9±1.2 nmol/L) for cortisol were significantly higher in patients with lower renin values (P<.03). Other variables, including plasma hormone and electrolyte values and binding characteristics for dexamethasone, were not different. These data suggest that cortisol binding to glucocorticoid receptor is slightly impaired in patients with essential hypertension. In vivo, this could lead to inappropriate binding of cortisol to mineralocorticoid receptors. Hence, decreased sensitivity to cortisol is associated with renin suppression. This hypothesis is supported by evidence of hypertension and low renin activity, which others have described in patients with primary glucocorticoid resistance due to mutations of the glucocorticoid receptor.
Glucocorticoid hormones play a key role in the control of blood pressure and are associated with hypertension in patients with Cushing’s syndrome.1 Increases in blood pressure following cortisol or corticotropin administration are often associated with sodium and fluid retention, hypokalemia, and suppression of plasma renin activity, which are indicative of excess MR activity.2 It is also clear, however, from studies with synthetic steroids with negligible MR properties, that glucocorticoid hormones can raise blood pressure acting via type 2 adrenocorticosteroid receptors.3 The mechanisms of this glucocorticoid-induced hypertension are poorly understood but are known to involve modifications of vascular responsiveness to vasoconstrictors and vasodilators; the distribution of electrolytes between intracellular and extracellular compartments is also affected.4
Despite the fact that glucocorticoid hormones have potent effects on blood pressure, evidence that they are involved in the etiology of essential hypertension is equivocal. Two studies have demonstrated increased urinary excretion of the principal metabolites of cortisol in patients with essential hypertension.5 6 Numerous other studies have failed to demonstrate that corticosteroid production is excessive and, when abnormal circulating levels are present, they do not seem to be high enough for the elevated blood pressure. However, more recent molecular studies of processes regulating hormone access to receptor as well as postreceptor events have indicated a number of sites that potentially could alter sensitivity to glucocorticoid hormones without altering systemic hormone concentrations.7 If essential hypertension is a multifactorial, polygenic disease, then glucocorticoid hormones, which regulate the expression of so many of the genes controlling blood pressure, are likely to be involved secondarily.
Normally, plasma cortisol concentrations are maintained by negative-feedback control of pituitary corticotropin secretion, which is mediated via adrenocorticosteroid receptors. Exposure to excess cortisol could arise, therefore, only if receptor function were altered in a general or tissue-specific manner. For example, a point mutation in the glucocorticoid receptor gene causes primary glucocorticoid resistance with high corticotropin and cortisol levels in the absence of clinical features of Cushing’s syndrome.8 This condition is associated with hypertension because, although type 2 adrenocorticosteroid function is impaired and feedback control of corticotropin secretion is lost, cortisol is able to exert MR-like effects via type 1 receptors.9 A case of hypersensitivity to cortisol has also been described, again caused by a mutation of the glucocorticoid receptor gene.10 This is associated with the symptoms of Cushing’s syndrome but not hypertension (hypertension is not invariably present in Cushing’s syndrome).
Such mutations of the glucocorticoid receptor gene are rare events and could not be expected to account for essential hypertension. More recently, a common polymorphism of the glucocorticoid receptor gene has been analyzed in a population study of the genetic determinants of hypertension.11 One of the allelic variants of the receptor gene was found to be more common in individuals with a familial disposition to develop hypertension. In the same group of individuals, plasma angiotensinogen (a glucocorticoid-dependent variable) and cortisol concentrations were slightly elevated. Although it has yet to be proven that this polymorphism alters receptor genotype, the concept that glucocorticoid function modulates blood pressure is established.
In the present study, we compare glucocorticoid binding properties and glucocorticoid responsiveness in white blood cells of patients with essential hypertension with those of normotensive control subjects. These values are considered in relation to plasma hormone values to determine whether receptor binding affects the pituitary-adrenal axis and to determine whether overall glucocorticoid activity secondarily affects MR activity.
Eleven normotensive healthy volunteers (6 women and 5 men) age 25 to 45 years and 17 essential hypertensive subjects (10 women and 7 men) were sampled. All women were evaluated during the follicular phase, whereas hypertensive subjects were studied after screening to exclude known causes of secondary hypertension. None of the hypertensive subjects had evidence of cardiac, renal, or hepatic failure and diabetes mellitus. Alcohol and nicotine abuse were excluded. Antihypertensive therapy was discontinued for at least 21 days before the study.
All subjects were admitted between 8:00 a.m. and 9:00 a.m., and blood samples were collected after the subjects had been lying down for at least 30 minutes. Blood pressure was measured by conventional methods twice on the same occasion in each individual, supine and after standing for at least 5 minutes. The mean of the two readings was used for analysis. Height and weight were measured to calculate the BMI from the formula BMI=weight(kg)/height2(m). Venous blood was then drawn for routine biochemical and hormonal measurement and HML separation. A 24-hour urine specimen was obtained before visit to the clinic.
Peripheral blood obtained from fasting individuals was collected into citrate, and platelet-rich plasma was removed by low-speed centrifugation (600g for 5 minutes). The remaining cells were diluted 1:3 with PBS (0.85% NaCl and 6.7 mmol/L phosphate, pH 7.2) and fractionated on a Ficoll-Hypaque gradient according to the method of Boyum.21 The layer of mononuclear cells was removed and washed three times with PBS at room temperature by centrifugation and resuspension to reduce platelet contamination. The final pellet was resuspended in RPMI-1640 medium containing 10% fetal calf serum, 1% penicillin, 1% streptomycin, and 1% glutamine. The cell viability was consistently higher than 95%, as evaluated by trypan blue dye exclusion. There were no changes in cell number or viability during any incubation procedure.
Whole-Cell GR Assay
GR binding characteristics were measured by homologous and heterologous competition for specific [3H]dexamethasone-binding sites under steady-state conditions.22 Briefly, ≈3×106 HML resuspended in RPMI-1640 (supplemented with 1% glutamine) were equilibrated with 2 nmol/L [3H]dexamethasone and a range of concentrations of unlabeled dexamethasone (0 to 150 nmol/L) and cortisol (0 to 600 nmol/L) by incubation for 3 hours at 24°C. Nonspecific binding was measured by incubating the cells with a 1000-fold excess of unlabeled dexamethasone. HML were harvested on Whatman GF/C filter paper and washed with ice-cold PBS containing polyethylenimine (1 g/L) using a Titertek cell harvester (Flow). Filters were treated with 1% Triton X-100, and the radioactivity from lysed cells was measured by a liquid scintillation spectrometer ( Beckman LS 1801).
Binding constants [Kd and binding capacity (Bmax)] were calculated from Scatchard plots that were fitted using the computer program Ligand.23 The data were compatible with a single-site model.
In vitro glucocorticoid responsiveness was measured by inhibition of lysozyme synthesis by dexamethasone and cortisol as previously described by Panarelli et al.24 Briefly, aliquots of cell suspensions (0.5 to 1×106 cells) were cultured in a 96-well tissue culture plate at 37°C in 5% CO2 for 72 hours in the presence of increasing concentrations of dexamethasone (0 to 150 nmol/L) and cortisol (0 to 600 nmol/L). After we centrifuged the plates, we removed a 100-μL aliquot of supernatant from each incubation well and stored it at −70°C. Lysozyme activity in the supernatant was measured photometrically (Beckaman DU20) by the lysis of Micrococcus lysodeikticus using human recombinant lysozyme as a standard. Each sample was assayed in triplicate. For each subject, the concentration of hormone that inhibits 50% of basal production of lysozyme (IC50) was calculated.
Biochemical and Hormonal Measurement
Routine serum and urine measurements were carried out by automated analyzer. A specific radioimmunoassay kit was used for the determination of plasma cortisol (INCSTAR, Sorin Biomedica), aldosterone (Sorin Biomedica), renin activity (Sorin Biomedica), whereas plasma corticotropin was determined by specific immunoradiometric assay (INCSTAR, Sorin Biomedica). Routine hematological profiles were carried out by the cytofluorimetric method.
The mean±SEM for each variable were calculated. Linear regression and correlation were calculated, and Student’s t test was used for the comparison of variables.
The clinical and laboratory results are summarized in Table 1⇓. Serum uric acid, cholesterol, and triglyceride were significantly higher in the hypertensive group compared with the normotensive group. Although serum creatinine was slightly higher and urinary Na+/K+ ratio was lower in the hypertensive group compared with the normotensive group, the differences were not statistically significant. BMI, serum electrolytes, hormonal levels, and hematological variables were similar in the two groups.
Binding characteristics for dexamethasone and cortisol are represented in Fig 1⇓, and the IC50 values for the inhibitory effects of dexamethasone and cortisol on lysozyme activity in HML are summarized in Fig 2⇓. Both dexamethasone and cortisol showed Kd, but not Bmax, values that positively correlated with the IC50 values for the effects of the corresponding hormone on lysozyme activity (R2=.8, P<.001 for dexamethasone; R2=.9, P<.001 for cortisol). Neither Kd nor Bmax for dexamethasone and cortisol correlated with either cortisol or corticotropin plasma concentrations.
Dissociation constant values for cortisol in HML from hypertensive subjects were significantly higher than control subjects (24.6±2.4 versus 17.5±1.7 nmol/L, P<.04). Binding capacity (4978±391 versus 4131±321 sites/cell), Kd values for dexamethasone (6.7±0.5 versus 5.7±0.3 nmol/L), and IC50 values for dexamethasone (3.4±0.3 versus 3.1±0.2 nmol/L) and cortisol (12.2±1.6 versus 9.5±0.3 nmol/L) were not statistically different.
The hypertensive group was divided in two subgroups according to plasma renin activity: (1) lower renin activity (n=8) defined as a basal plasma renin activity <0.13 ng angiotensin I/L per second and a plasma renin activity after 4 hours of walking of less than the upper limit of 1 SD of the normal basal value; and (2) normal high renin (n=9). Table 2⇓ compares biochemical and hormonal values in the two subgroups. The dissociation constant and IC50 for dexamethasone and cortisol in these two subgroups of hypertensive subjects are summarized in Figs 3⇓ and 4⇓. The lower-renin hypertensive subgroup was markedly less sensitive, in vitro, to cortisol compared with the subgroup with higher renin levels. Both Kd (30.3±2.5 versus 19.2±2.4 nmol/L) and IC50 (15.5±1.8 versus 8.9±1.2 nmol/L) for cortisol were significantly high in the lower-renin subgroup (P<.03). The binding capacity, Kd, and IC50 for dexamethasone were not significantly different between the two subgroups. Although plasma corticotropin and cortisol were slightly higher and the urinary Na+/K+ ratio was decreased in the lower-renin subgroup, the differences were not statistically significant (see Table 2⇓).
Endogenous GR are bound with similar affinities to MR and GR receptors, although access to MR in key tissues is restricted by enzymes that metabolize cortisol.12 13 Excess cortisol can cause hypertension by activation of GR and, after overcoming the enzyme barrier that protects MR, by promoting sodium retention and plasma volume expansion.2 Normally, however, feedback inhibition of pituitary corticotropin secretion by cortisol would reduce adrenal steroidogenesis to maintain homeostatic control of circulating concentrations of glucocorticoid hormones. It follows that cortisol might be expected to influence blood pressure if (1) the amount of added cortisol exceeds the compensatory capacity of the adrenal gland, (2) enzymes that metabolize cortisol are deficient, or (3) if GR binding properties are abnormal. In previous studies addressing the second possibility, a deficiency of 11β-hydroxysteroid dehydrogenase has been shown to cause the syndrome of apparent MR excess.14 15 The deficiency is characterized by high ratios of cortisol to cortisone metabolites. In the patients with essential hypertension, although overt signs of MR excess are not apparent, there have been several reports suggesting abnormalities of cortisol metabolism.5 6 16 The present study has addressed the third possibility by measuring GR binding properties and by assessing glucocorticoid responsiveness of patients with essential hypertension and of normotensive control subjects.
Although the synthetic glucocorticoid dexamethasone is used most often in these types of tests, in our earlier experience with a rat genetic model of hypertension we noted that differences between Milan hypertensive and normotensive strains in binding for endogenous ligands (corticosterone and aldosterone) were greater than that for dexamethasone.17 Accordingly, we considered both dexamethasone and cortisol in measurements of steroid binding properties and in steroid-induced inhibition of white blood cell lysozyme synthesis. When comparing the normotensive group with the entire hypertensive group, we noted that affinity for cortisol, but not dexamethasone, was lower in hypertensive subjects than in control subjects. The reason cortisol, but not dexamethasone, binding affinities were affected is not clear. Although in vivo cortisol is metabolized more readily than dexamethasone, this cannot account for the present in vitro observations.18 In previous unpublished studies, we incubated white blood cells from normal volunteers with cortisol for periods up to 72 hours (ie, the time course for lysozyme assay). When cells and medium were extracted and analyzed chromatographically at the end of this time, only the parent compound was recovered. It follows that if metabolism does not affect cortisol activity in cells from normal individuals, a metabolic deficiency cannot account for the discrepancy between affinities for dexamethasone and cortisol in hypertensive subjects. It is possible that there are also minor differences in dexamethasone binding but that like the Milan rats, the magnitude is much lower for higher-affinity ligands.
In the case of Milan rats, there are a number of differences in the coding region of the GR gene that may be associated with low-affinity binding. In humans, mutations that cause reduced steroid binding are relatively rare. The few glucocorticoid resistance cases that have been reported appear to have variable etiologies and side effects.8 19 20 A more common mutation affecting approximately 10% of a healthy population has been described, but this is associated with increased not decreased sensitivity.
For both dexamethasone and cortisol, the IC50 values for steroid-induced inhibition of lysozyme release correlated with Kd values for receptor binding studies. As might be expected, therefore, there were no significant differences in lysozyme release between control subjects and hypertensive patients. For cortisol, IC50 values (in line with Kd values) were 25% higher in the hypertensive than in the normotensive group, but this difference was not statistically significant.
Among patients with essential hypertension, a significant subgroup with lower plasma renin but normal aldosterone concentrations has been identified. Working on the premise that MR excess is characterized by suppression of renin activity, many groups have focused on novel steroid hormones with potent MR activity to explain these low renin values. As outlined above, however, low renin values might also be a consequence of cortisol acting on MR. In the present context, impaired GR binding leads to reduced feedback inhibition of corticotropin secretion and raised plasma cortisol concentrations which, acting on MR in transporting epithelia, could cause an antinatriuresis, renin suppression, and hypertension. When comparing all hypertensive patients with the control group, there were no differences in plasma renin activity. However, when patients with high- and low-renin groups were analyzed separately, it became apparent that hypertensive subjects in the lower-renin group had higher Kd values and showed reduced sensitivity to cortisol-induced lysozyme release compared with the high/normal–renin subgroup; the latter group had white blood cells that showed similar steroid binding properties and steroid-sensitive lysozyme release to those from normotensive control subjects.
One would predict, theoretically, that if GR impairment causes MR activation, then this should be mediated by raised plasma cortisol concentrations which, in turn, should be driven by increased corticotropin secretion. No such differences in plasma cortisol and corticotropin were observed. It could be argued, however, that a single-point measurement of cortisol and corticotropin does not reflect the complex circadian control of the hypothalamus-pituitary-adrenal axis. A test using low-dose cortisol (not dexamethasone) to suppress corticotropin would be required that should also take account of the influence of the 11β-hydroxysteroid dehydrogenase enzymes.
The association of GR binding and suppression of renin activity might also suggest that electrolyte metabolism could be affected. Hypokalemia was not seen in patients with lower renin values, but there was a tendency for the urinary sodium/potassium ratio to be decreased. This lack of a significant effect is not surprising because no attempt was made in the present investigations to control dietary intake of electrolytes. Others have also failed to show a link of urinary or plasma electrolytes and blood pressure in patients with essential hypertension.
An alternative view of the present results is that changes in plasma renin activity and receptor binding characteristics are a consequence rather than a contributory cause of hypertension. Although there are considerable data to indicate that expression of steroid hormone receptors and, therefore, binding capacity might be regulated, there is no evidence that receptor affinity and specificity are regulated other than via steroid-metabolizing enzymes. Although there is evidence in vivo of altered steroid metabolism in patients with essential hypertension,5 6 16 it does not seem to apply to our in vitro studies of receptor binding and glucocorticoid-dependent inhibition of lysozyme release. We are unaware of any mechanism whereby blood pressure or the consequences of hypertension could account for our findings.
In contrast to our present findings, Walker et al25 report increased peripheral glucocorticoid sensitivity in vivo associated with higher affinity of GR for dexamethasone in vitro in subjects predisposed to high blood pressure. Although these results seem contradictory, it should be remembered that glucocorticoid hormones, independent of MR-like actions, increase blood pressure.4 It follows that increased affinity of GR for the ligand would potentiate glucocorticoid-dependent processes and therefore increase blood pressure. Taking these data together, therefore, a plot of blood pressure against GR should be U-shaped; one limb of the U should be associated with normal renin values and reduced plasma cortisol concentrations, and the other limb should be associated with low renin values and raised plasma cortisol concentration.
In conclusion, our results show that in patients with essential hypertension, there is a subgroup that exhibits slightly impaired glucocorticoid binding. This effect is specific to cortisol and not dexamethasone and leads to reduced glucocorticoid responsiveness of white blood cells in vitro. This difference in steroid binding may secondarily increase MR activity and therefore could be a contributory cause of low renin values and higher blood pressure in these patients. The molecular mechanism underlying this difference in receptor binding remains to be elucidated.
Selected Abbreviations and Acronyms
|BMI||=||body mass index|
|HML||=||human mononuclear leukocyte(s)|
- Received February 10, 1996.
- Revision received March 17, 1996.
- Accepted June 2, 1997.
Kenyon CJ, Morton JJ. Experimental steroid-induced hypertension. In: Swales JD, ed. Textbook of Hypertension. Oxford, UK: Blackwell Scientific Publications; 1993:494-500.
Soro A, Ingram MC, Tonolo G, Glorioso N, Fraser R. Evidence of co-existing changes in 11β-hydroxysteroid dehydrogenase and 5β-reductase activity in subjects with untreated essential hypertension. Hypertension. 1995;25:67-70.
Chrousos GP, Vingerhoed A, Brandon D, Eil C, Pugeat M, De Vroede M, Loriaux DL, Lipsett MB. Primary cortisol resistance in man. A glucocorticoid receptor-mediated disease. J Clin Invest. 1982;69:1261-1269.
Lamberts SWJ, Molijn GJ, Huizenga NTM, de Jong FH, Koper JW. The glucocorticoid receptor in health and disease. J Endocrinol. 1994;143(suppl 1):O1.
Watt GCM, Harrap SB, Foy CJW, Holton DW, Edwards HW, Davidson HR, Connor JM, Lever AF, Fraser R. Abnormalities of glucocorticoid metabolism and the renin-angiotensin system: a four corner approach to the identification of the genetic determinants of blood pressure. J Hypertens. 1992;10:473-482.
Edwards CRW, Burt D, McIntyre MA, de Kloet ER, Stewart PM, Brett LP, Sutanto WS, Monder C. Localization of 11β-hydroxysteroid dehydrogenase-tissue specific protector of the mineralocorticoid receptor. Lancet. 1989;ii:986-989.
Funder JW, Pearce PT, Smith R, Smith AL. Mineralocorticoid action: target tissue specificity is enzyme not receptor mediated. Science. 1988;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. 1998;82:340-349.
Panarelli M, Holloway CD, Fraser R, Kenyon CJ. Evidence of an impaired glucocorticoid receptor in Milan hypertensive rats. J Endocrinol. 1994;140(suppl):P21.
Walker BR, Williams BC, Edwards CRW. Regulation of 11β-hydroxysteroid dehydrogenase activity by the hypothalamic-pituitary-adrenal axis in the rat. J Endocrinol. 1994;141:467-472.
Boyum A. Isolation of mononuclear cells and granulocytes from human blood. Isolation of mononuclear cells by one centrifugation and of granulocytes by combining centrifugation and sedimentation at 1G. Scand J Clin Lab Invest. 1968;21(suppl):77-89.
Walker BR, Noon JP, Panarelli M, Best R, Holton DW, Edwards HV, Edwards CRW, Watt GCM, Webb DJ. Increased glucocorticoid receptor affinity and sensitivity in subjects predisposed to high blood pressure. J Hypertens. 1996;14(suppl 1):P655.