- Email: [email protected]
RALES, EPHESUS and redox
Journal of Steroid Biochemistry & Molecular Biology 93 (2005) 121–125
RALES, EPHESUS and redox夽 John W. Funder∗ Prince Henry’s Institute of Medical Research, P.O. Box 5151, Clayton 3168, Vic., Australia
Abstract In RALES, low doses of the mineralocorticoid receptor (MR) antagonist spironolactone, added to standard of care for severe heart failure, improved survival by 30% and lowered hospitalization by 35%. Animal studies with the selective MR antagonist eplerenone have similarly shown MR blockade to prevent the cerebral, renal and coronary vascular inflammatory response to elevated aldosterone levels. There is now general acceptance that aldosterone concentrations inappropriate for salt status have major deleterious effects on the cardiovascular system. In many instances, however (e.g. Randomized Aldactone Evaluation Study (RALES), EPHESUS) aldosterone levels are normal and salt status unremarkable and yet MR blockade has unquestioned benefits. In these instances, there is increasing evidence that coronary and cardiac MR are activated by normal circulating cortisol levels, in the cellular context of generation of reactive oxygen species (ROS) and/or alteration in intracellular redox status. MR in VSMC and cardiomyocytes are normally predominantly occupied by cortisol in tonic inhibitory mode. Blockade of 11 hydroxysteroid dehydrogenase type II (11HSD2) or ROS generation both serve to activate cortisol–MR complexes, thus mimicking the effects of mineralocorticoid/salt imbalance on blood vessels and the heart. In RALES and EPHESUS, it is likely that the antagonists are blocking normal levels of cortisol, not aldosterone, from activating MR in the context of tissue damage and ROS generation. If this is the case, MR antagonists may be of wide therapeutic potential in cardiovascular disease and not confined to those characterized by aldosterone/salt excess. Finally, the pathophysiologic roles of always-occupied MR in ‘unprotected’ tissues such as cardiomyocytes or neurons in response to altered intracellular redox status remain to be explored. © 2004 Elsevier Ltd. All rights reserved. Keywords: Reactive oxygen species; Aldosterone levels; Intracellular redox status
In September 1999, the results of the Randomized Aldactone Evaluation Study (RALES) were published [1], with earlier on-line publication in the light of the anticipated interest in the outcome. In the study, two groups of patients with progressive, moderately severe (NYHA Stage III) heart failure were compared. Both groups received standard care therapy (diuretics, ACE inhibitors, etc.); in addition, one group received spironolactone (Aldactone® ) at an average dose of 26 mg/day and the other, placebo. Patients were followed for 3 years, but the originally planned 36 months recruitment period was closed just over half way through, given the difference in morbidity and mortality between the two groups. 夽 Proceedings of the 16th International Symposium of the Journal of Steroid Biochemistry and Molecular Biology, ‘Recent Advances in Steroid Biochemistry and Molecular Biology’, Seefeld, Tyrol, Austria, 5–8 June 2004. ∗ Tel.: +61 3 9594 3557; fax: +61 3 9594 7161. E-mail address: [email protected].
0960-0760/$ – see front matter © 2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.jsbmb.2004.12.010
In brief, the administration of the mineralocorticoid receptor (MR) antagonist spironolactone produced a 30% improvement in survival and a 35% lower incidence of hospitalization. Given the time-honoured use of spironolactone as an aldosterone antagonist, a question that naturally arises from RALES is whether aldosterone is bad for the heart, simply put. Perhaps too simply; without aldosterone we die, which is not good for the heart. If the question is recast – “Is inappropriate MR activation bad for the heart?” – then the answer is yes. If we then further refine the question “Can aldosterone be responsible for inappropriate coronary/cardiac MR stimulation?”—then the answer again is yes, with the qualification “when aldosterone is too high for salt status”. Under physiological conditions, aldosterone levels can be extremely high, without any deleterious cardiovascular effects, provided that they are high in response to sodium deficiency or loss. There is no question that in a variety of animal models the combination of aldosterone and salt can produce spec-
122
J.W. Funder / Journal of Steroid Biochemistry & Molecular Biology 93 (2005) 121–125
tacular pathologic consequences and that these can be very largely blocked by administration of MR antagonists such as spironolactone or the novel much more selective MR antagonist eplerenone. When stroke-prone spontaneously hypertensive rats (SHRSP) are given 0.9% NaCl solution as their drinking fluid from 9 weeks of age, they progressively die of stroke over the following 2 months. If at 9 weeks they are given saline to drink and eplerenone in their chow, survival increases dramatically, so that at 19 weeks only 1/8 treated SHRSP died [2]. Vehicle treated rats showed major cerebrovascular lesions and consequent rarefaction of brain parenchyma; eplerenone-treated rats showed a cerebral injury score a quarter that in control, despite no significant difference in the very high blood pressures (>200 mmHg) in this model. In the kidney, a similar protective effect was seen with renal vascular lesions <5% of control values in eplerenonetreated SHRSP and urinary protein excretion <10%. A second rat model of experimental hypertension also has proven instructive [3]. When uninephrectomized rats drinking 0.9% NaCl solution are infused with angiotensin II (AII) for 3 weeks their blood pressure rises progressively. This elevation is AII-driven, in that it is unaffected by adrenalectomy or eplerenone administration from the start of AII infusion. What is not angiotensin-driven, however, is the coronary vascular and perivascular tissue response. In AII/salt rats, after 3 weeks coronary vessels show massive thickening of their walls (Fig. 1A) and widespread perivascular and interstitial inflammatory cell infiltration. In AII/salt/adrenalectomized rats, with the same blood pressure, the vessel walls are normal and inflammatory cells conspicuously absent (Fig. 1B). When, however, the adrenalectomized rats are infused with aldosterone (Fig. 1C), comparable vascular and interstitial changes to those in AII/salt rats can be seen. There is clearly no question that in these animal models inappropriate aldosterone for salt status can produce very deleterious coronary and cardiac effects, net of any blood pressure elevation; in the clinical situation of Conn’s syndrome, for instance, it is reasonable to assume a similar role for aldosterone, out of normal feedback control, in patients on a normal (i.e. high salt) western diet. On the other hand, in RALES and subsequently in the similar clinical trial of eplerenone in post-myocardial infarct heart failure (EPHESUS; [4]), aldosterone levels were normal and salt status
unremarkable, raising the question of what the spironolactone or eplerenone was actually blocking to have such a major positive effect. The remainder of this presentation is thus devoted to providing supporting evidence for the contention that many, perhaps most, coronary and cardiac effects of inappropriate MR activation are in fact glucocorticoid-driven, in the context of tissue damage and the generation of reactive oxygen species (ROS). To sustain this contention it is first necessary to revisit some of the historic background findings on MR expression and selectivity in terms of ligand binding. When the human MR was first cloned in Ron Evans’ laboratory [5] and an hMR cDNA used to probe mRNA preparations from rat tissues, MR were found to be expressed not only in epithelial tissues such as kidney and gut, but also in non-epithelial tissues such as hippocampus and heart. Secondly, when the relevant recombinant receptors were expressed in COS cells and the affinity of various steroids determined by their ability to compete with [3 H]aldosterone for binding, it was found that cortisol and aldosterone showed equivalent high affinity for MR and that the affinity of corticosterone was approximately three-fold higher. These findings raised two questions – first, of the potential physiological roles of non-epithelial MR; and secondly, of how aldosterone could ever selectively activate epithelial MR, given that circulating plasma concentrations of the physiological glucocorticoids are three orders of magnitude higher than those of aldosterone. Even factoring in the much higher binding of the glucocorticoids in plasma, their free levels are still a 100-fold those of aldosterone. While the answer to the first of these questions is not yet clear, an aldosterone specificity-conferring mechanism in epithelial cells was proposed shortly afterwards [6,7]. In addition to MR, epithelial aldosterone target cells express at very high abundance the enzyme 11 hydroxysteroid dehydrogenase type II (11HSD2). Expression of 11HSD2 is not confined to classic aldosterone target tissues; it is also expressed, for example, in placenta, amygdala and – importantly for the present discussion – in vascular smooth muscle cells. What 11HSD2 does is to convert cortisol to cortisone, which has negligible affinity for MR, thus putatively excluding cortisol from MR and allowing aldosterone to access and activate them. Aldosterone escapes the attention of the enzyme because in solution the very
Fig. 1. (A) Coronary artery inflammation and perivascular inflammatory cell infiltration in hearts from rats on 0.9% NaCl solution to drink and infused with angiotensin II for 3 weeks. (B) Abrogation of inflammatory response by prior adrenalectomy. (C) Restoration of response by concurrent aldosterone infusion (redrawn with permission from [3]).
J.W. Funder / Journal of Steroid Biochemistry & Molecular Biology 93 (2005) 121–125
Fig. 2. Binding of tritiated aldosterone in epithelial and non-epithelial tissues; relative in vivo receptor occupancy by non-radioactive aldosterone (circles) and corticosterone (squares) coadministered with tracer. Adapted from [8].
reactive aldehyde group at carbon 18 cyclizes with the hydroxyl at carbon 11, forming an 11,18 hemiacetal and thus protecting the 11 hydroxyl from enzymatic attack. When 11HSD2 is not operating—as in the congenital syndrome of apparent mineralocorticoid excess or following licorice abuse or carbenoxolone (glycyrrhetinic acid hemisuccinate) administration, cortisol activates MR producing sodium retention, volume expansion and blood pressure elevation. That the enzyme operates to prevent glucocorticoid occupancy rather than glucocorticoid activation of epithelial MR in retrospect was to draw a very long bow, given the intracellular concentration differences between glucocorticoid and aldosterone. This was shown to be the case in a study on rats, in which the physiological glucocorticoid is corticosterone rather than cortisol. Adrenalectomized rats were injected with [3 H]aldosterone alone or plus half-log increasing doses of non-radioactive aldosterone or corticosterone [8]. Fifteen minutes after injection, the animals were killed and tissues harvested for determination of MR bound [3 H]aldosterone, with the inference that in this way in vivo MR occupancy by aldosterone or corticosterone in various tissues can be established. The results of these studies are shown in Fig. 2. In the heart (lower left) aldosterone appears to have ∼3fold higher affinity for MR in vivo. The heart expresses negligible levels of 11HSD2; corticosterone has ∼3-fold higher intrinsic affinity for MR, but on the other hand is 10-fold more highly bound in plasma; the finding that in vivo aldosterone at equal doses has a three-fold advantage is therefore consistent. For the hippocampus (Fig. 2, lower right), the aldosterone competition curve is shifted far to the right, evidence for the very high reflection coefficient for aldosterone at the blood–brain barrier; corticosterone, despite its higher plasma binding, gets through the blood–brain barrier very much more easily, and thus at equivalent concentrations is actually a more
123
likely occupant. In the classical epithelia tissues (Fig. 2, upper panels) where in contrast to cardiomyocytes and hippocampal neurons high levels of 11HSD2 are expressed, the operation of the enzyme shifts the corticosterone curve to the right by approximately an order of magnitude—i.e. it debulks intracellular glucocorticoid, rather than completely metabolizing it to its inactive 11-keto congener. What this means is that under physiologic conditions, intracellular active glucocorticoid concentrations are still ∼10-fold those of aldosterone and that most epithelial MR are occupied but not activated by cortisol—or, in the rat, corticosterone. When 11HSD2 is blocked, marginally more epithelial MR will be cortisol occupied, but this is an unsatisfactory explanation for a change in activity: maybe, when the enzyme is blocked, something else happens to activate glucocorticoid complexes. The prime candidate for this change-of-state would appear to be the generation of NADH from NAD, the forgotten cosubstrate for 11HSD2. For every molecule of cortisol converted to cortisone, a molecule of NAD is converted to NADH. Resting levels of NAD in the cell are far (∼600 times) higher than those of NADH, so that for minimal changes in substrate NAD levels (e.g. a fall from 600 to 500) an increase in NADH levels of two orders of magnitude could be achieved. It is not clear how the resultant redox change activates MR–glucocorticoid complexes, though redox-dependent transcriptional changes have been previously reported, involving, for example, activation of corepressors [9,10]. A corollary of this potential mechanism for MR is that the conformation and thus the cofactor binding profile of an MR–glucocorticoid complex differs sufficiently from that of an MR–aldosterone (or, for that matter, an MR–spironolactone) complex to allow cortisol this bivalent activity. That such redox-dependent bivalent activity may reflect the operation of additional or alternate mechanisms, suggested by its action in acute non-genomic MR effects, to be discussed later. One indirect but instructive series of studies by Young et al. [11] is shown in Fig. 3. Uninephrectomized rats maintained on 0.9% NaCl solution to drink and given a single injection of deoxycorticosterone (DOC) in oil show promptly increased levels of coronary vascular inflammatory markers (ED-1, COX-2 and osteopontin) 8 days later (Fig. 3). When the animals were given carbenoxolone (CBX) in their drinking fluid, rather than receiving DOC by injection, the vascular inflammatory response in terms of these three markers was identical to that seen with DOC. That this effect of carbenoxolone is mediated via MR is shown for each marker by the ability of eplerenone (EPL) to block the effect of carbenoxolone down to control levels. Our interpretation of these data is that when VSMC 11HSD2 is blocked, the normally inactive corticosterone–MR complexes in VSMC are activated and mimic the effect of administered exogenous mineralocorticoid. That this is of wider relevance to MR pathophysiology than animals receiving carbenoxolone or clinical situations of licorice overindulgence is suggested by the data shown
124
J.W. Funder / Journal of Steroid Biochemistry & Molecular Biology 93 (2005) 121–125
Fig. 3. Response of inflammatory markers to deoxycorticosterone (DOC), carbenoxolone (CBX) or carbenoxolone plus eplerenone (EPL) in coronary vessels from uninephrectomized rats drinking 0.9% NaCl solution for 8 days. Adapted from [11]; (*) P < 0.05.
in Fig. 4. In these studies, pigs were subjected to coronary angioplasty, receiving vehicle, aldosterone or eplerenone for 5 days before and 28 days after the procedure [12]. What eplerenone does under these circumstances is to prevent constrictive fibrosis, and thus preserve the coronary lumen. Control animals received no exogenous mineralocorticoid and were provided with regular pig chow and water to drink: what, then, is the eplerenone antagonizing in vessel wall MR? The answer would appear to be the action of normal levels of cortisol, occupying the majority of VSMC MR and transformed from inactive to active by the redox changes consequent upon generation of reactive oxygen species in the context of tissue damage. If redox changes following 11HSD2 blockade or ROS generation can transform glucocorticoid–MR complexes from inactive to active, then a possible physiological role for cardiomyocyte or neuronal MR in the absence of 11HSD2 may be as follows. In such cells the intracellular concentrations of glucocorticoid are such that MR will be essentially always occupied over the course of diurnal variation in glucocorticoid levels. This contrast with classical glucocorticoid receptors (GR), which have much (∼30-fold) lower affinity for cortisol/corticosterone than MR, and thus are more or less occupied over the range of physiological glucocorticoid levels. An always-occupied MR is difficult to reconcile as responsive to changing glucocorticoid levels; what may determine its response, therefore, is the redox state of the cells.
Both cardiomyocyte and neurons, for example, face particular metabolic challenges; glucocorticoid–MR complexes, activated under conditions of metabolic stress, may thus constitute one of the mechanisms involved in the cellular response to such stress. The extent to which such responses are homeostatic or ultimately pro-apoptotic remains to be explored. Very recently, Mihailidou et al. have provided direct evidence for the importance of intracellular redox state in MR activation by glucocorticoids [13]. When isolated rabbit cardiomyocytes are treated with 10 nM aldosterone, they show a rapid non-genomic response measured as a ∼10-fold increase in pump current within 15 min. Cortisol alone at 100 nM has no agonist effect when given alone and stoichometrically blocks the agonist effect of aldosterone 10 nM down to ∼10% of aldosterone alone levels. When intracellular redox state is altered by infusion of oxidized glutathione (GSSG) directly into the myocytes via the wide-tipped pipette, no change to baseline current is seen. When, however, GSSG infused cells are exposed to cortisol, it mimics aldosterone, producing an agonist effect on pump current. These preliminary studies thus constitute the first direct demonstration of the bivalent activity of physiological glucocorticoids in MR dependent upon the intracellular redox state. In summary, in a broad spectrum of cardiovascular disease aldosterone levels are within the normal range and sodium status unremarkable. What is common to an atherosclerotic vessel and a cardiomyocyte in heart failure is tissue damage and generation of reactive oxygen species. Use of second generation MR antagonists, such as eplerenone and subsequent generations of even more selective potent MR blockers, should thus not be limited to those conditions in which aldosterone is elevated. There will clearly be a major advance over Aldactone in Conn’s syndrome and perhaps in the broader context of patients with inappropriate aldosterone to renin ratios [14]; given their extraordinary vasoprotective effect in animal studies [2,3], even at doses that do not lower blood pressure, they may well have a much broader application across the spectrum of cardiovascular disease. References
Fig. 4. Luminal diameter in pig coronary vessels subject to experimental coronary angioplasty 4 weeks earlier. Adapted from [12]; (*) P < 0.05.
[1] B. Pitt, F. Zannad, W.J. Remme, R. Cody, A. Castaigne, A. Perez, J. Palensky, J. Wittes, The effect of spironolactone on morbidity and
J.W. Funder / Journal of Steroid Biochemistry & Molecular Biology 93 (2005) 121–125
[2] [3]
[4]
[5]
[6]
[7]
[8]
mortality in patients with severe heart failure, New Engl. J. Med. 341 (10) (1999) 709–717. R. Rocha, C. Stier, Pathophysiological effects of aldosterone in cardiovascular tissue, Trends Endocrinol. Metab. 12 (7) (2001) 308–314. R. Rocha, C. Martin-Berger, P. Yang, R. Scherrar, J. Delyani, E. McMahon, Selective aldosterone blockade prevents angiotensin II/salt-induced vascular inflammation in the rat heart, Endocrinology 143 (2002) 4828–4836. B. Pitt, W. Remme, F. Zannad, J. Neaton, F. Martinez, B. Roniker, R. Bittman, S. Hurley, J. Kleiman, M. Gatlin, Eplerenone, a selective aldosterone blocker, in patients with left ventricular dysfunction after myocardial infarction, New Engl. J. Med. 348 (14) (2003) 1309–1321. J.L. Arriza, C. Weinberger, G. Cerelli, T.M. Glaser, B.L. Handelin, D.E. Housman, R.M. Evans, Cloning of human mineralocorticoid receptor complementary DNA: structural and functional kinship with the glucocorticoid receptor, Science 237 (4812) (1987) 268–275. J.W. Funder, P. Pearce, R. Smith, A.I. Smith, Mineralocorticoid action: target-tissue specificity is enzyme, not receptor, mediated, Science 242 (1988) 583–585. C.R. Edwards, P.M. Stewart, D. Burt, L. Brett, M.A. McIntyre, W.S. Sutanto, E.R. de Kloet, C. Monder, Localisation of 11 betahydroxysteroid dehydrogenase—tissue specific protector of the mineralocorticoid receptor, Lancet 2 (8618) (1988) 986–989. J.W. Funder, K. Myles, Exclusion of corticosterone from epithelial mineralocorticoid receptors is insufficient for selectivity of aldos-
[9] [10]
[11]
[12]
[13]
[14]
125
terone action: in vivo binding studies, Endocrinology 137 (12) (1996) 5264–5268. Q. Zhang, D. Piston, R. Goodman, Regulation of corepressor function by nuclear NADH, Science 295 (5561) (2002) 1895–1897. C. Fjeld, W. Birdsong, R. Goodman, Differential binding of NAD+ and NADH allows the transcriptional corepressor carboxyl-terminal binding protein to serve as a metabolic sensor, Proc. Natl. Acad. Sci. U.S.A. 100 (16) (2003) 9202–9207. M. Young, L. Moussa, R. Dilley, J.W. Funder, Early inflammatory responses in experimental cardiac hypertrophy and fibrosis: effects of 11 beta-hydroxysteroid dehydrogenase inactivation, Endocrinology 144 (3) (2003) 1121–1125. M.R. Ward, P. Kanellakis, D. Ramsey, J.W. Funder, A. Bobik, Eplerenone suppresses constrictive remodeling and collagen accumulation after angioplasty in porcine coronary arteries, Circulation 104 (2001) 467–472. A.S. Mihailidou, M. Mardini, T. Fraser, D. Knights, J.W. Funder, Agonist/antagonist activity of cortisol in cardiomyocyte mineralocorticoid receptors is determined by redox state, in: Proceeding of 30th Annual Meeting of the International Aldosterone Conference, New Orleans, 2004, p. 18. O. Olivieri, A. Ciacciarelli, D. Signorelli, F. Pizzolo, P. Guarini, C. Pavan, A. Corgnati, S. Falcone, R. Corrocher, A. Micchi, C. Cressoni, G. Blengio, Aldosterone to renin ratio in a primary care setting: the Bussolengo study, J. Clin. Endocrinol. Metab. 89 (9) (2004) 4221–4226.
RALES, EPHESUS and redox夽 John W. Funder∗ Prince Henry’s Institute of Medical Research, P.O. Box 5151, Clayton 3168, Vic., Australia
Abstract In RALES, low doses of the mineralocorticoid receptor (MR) antagonist spironolactone, added to standard of care for severe heart failure, improved survival by 30% and lowered hospitalization by 35%. Animal studies with the selective MR antagonist eplerenone have similarly shown MR blockade to prevent the cerebral, renal and coronary vascular inflammatory response to elevated aldosterone levels. There is now general acceptance that aldosterone concentrations inappropriate for salt status have major deleterious effects on the cardiovascular system. In many instances, however (e.g. Randomized Aldactone Evaluation Study (RALES), EPHESUS) aldosterone levels are normal and salt status unremarkable and yet MR blockade has unquestioned benefits. In these instances, there is increasing evidence that coronary and cardiac MR are activated by normal circulating cortisol levels, in the cellular context of generation of reactive oxygen species (ROS) and/or alteration in intracellular redox status. MR in VSMC and cardiomyocytes are normally predominantly occupied by cortisol in tonic inhibitory mode. Blockade of 11 hydroxysteroid dehydrogenase type II (11HSD2) or ROS generation both serve to activate cortisol–MR complexes, thus mimicking the effects of mineralocorticoid/salt imbalance on blood vessels and the heart. In RALES and EPHESUS, it is likely that the antagonists are blocking normal levels of cortisol, not aldosterone, from activating MR in the context of tissue damage and ROS generation. If this is the case, MR antagonists may be of wide therapeutic potential in cardiovascular disease and not confined to those characterized by aldosterone/salt excess. Finally, the pathophysiologic roles of always-occupied MR in ‘unprotected’ tissues such as cardiomyocytes or neurons in response to altered intracellular redox status remain to be explored. © 2004 Elsevier Ltd. All rights reserved. Keywords: Reactive oxygen species; Aldosterone levels; Intracellular redox status
In September 1999, the results of the Randomized Aldactone Evaluation Study (RALES) were published [1], with earlier on-line publication in the light of the anticipated interest in the outcome. In the study, two groups of patients with progressive, moderately severe (NYHA Stage III) heart failure were compared. Both groups received standard care therapy (diuretics, ACE inhibitors, etc.); in addition, one group received spironolactone (Aldactone® ) at an average dose of 26 mg/day and the other, placebo. Patients were followed for 3 years, but the originally planned 36 months recruitment period was closed just over half way through, given the difference in morbidity and mortality between the two groups. 夽 Proceedings of the 16th International Symposium of the Journal of Steroid Biochemistry and Molecular Biology, ‘Recent Advances in Steroid Biochemistry and Molecular Biology’, Seefeld, Tyrol, Austria, 5–8 June 2004. ∗ Tel.: +61 3 9594 3557; fax: +61 3 9594 7161. E-mail address: [email protected].
0960-0760/$ – see front matter © 2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.jsbmb.2004.12.010
In brief, the administration of the mineralocorticoid receptor (MR) antagonist spironolactone produced a 30% improvement in survival and a 35% lower incidence of hospitalization. Given the time-honoured use of spironolactone as an aldosterone antagonist, a question that naturally arises from RALES is whether aldosterone is bad for the heart, simply put. Perhaps too simply; without aldosterone we die, which is not good for the heart. If the question is recast – “Is inappropriate MR activation bad for the heart?” – then the answer is yes. If we then further refine the question “Can aldosterone be responsible for inappropriate coronary/cardiac MR stimulation?”—then the answer again is yes, with the qualification “when aldosterone is too high for salt status”. Under physiological conditions, aldosterone levels can be extremely high, without any deleterious cardiovascular effects, provided that they are high in response to sodium deficiency or loss. There is no question that in a variety of animal models the combination of aldosterone and salt can produce spec-
122
J.W. Funder / Journal of Steroid Biochemistry & Molecular Biology 93 (2005) 121–125
tacular pathologic consequences and that these can be very largely blocked by administration of MR antagonists such as spironolactone or the novel much more selective MR antagonist eplerenone. When stroke-prone spontaneously hypertensive rats (SHRSP) are given 0.9% NaCl solution as their drinking fluid from 9 weeks of age, they progressively die of stroke over the following 2 months. If at 9 weeks they are given saline to drink and eplerenone in their chow, survival increases dramatically, so that at 19 weeks only 1/8 treated SHRSP died [2]. Vehicle treated rats showed major cerebrovascular lesions and consequent rarefaction of brain parenchyma; eplerenone-treated rats showed a cerebral injury score a quarter that in control, despite no significant difference in the very high blood pressures (>200 mmHg) in this model. In the kidney, a similar protective effect was seen with renal vascular lesions <5% of control values in eplerenonetreated SHRSP and urinary protein excretion <10%. A second rat model of experimental hypertension also has proven instructive [3]. When uninephrectomized rats drinking 0.9% NaCl solution are infused with angiotensin II (AII) for 3 weeks their blood pressure rises progressively. This elevation is AII-driven, in that it is unaffected by adrenalectomy or eplerenone administration from the start of AII infusion. What is not angiotensin-driven, however, is the coronary vascular and perivascular tissue response. In AII/salt rats, after 3 weeks coronary vessels show massive thickening of their walls (Fig. 1A) and widespread perivascular and interstitial inflammatory cell infiltration. In AII/salt/adrenalectomized rats, with the same blood pressure, the vessel walls are normal and inflammatory cells conspicuously absent (Fig. 1B). When, however, the adrenalectomized rats are infused with aldosterone (Fig. 1C), comparable vascular and interstitial changes to those in AII/salt rats can be seen. There is clearly no question that in these animal models inappropriate aldosterone for salt status can produce very deleterious coronary and cardiac effects, net of any blood pressure elevation; in the clinical situation of Conn’s syndrome, for instance, it is reasonable to assume a similar role for aldosterone, out of normal feedback control, in patients on a normal (i.e. high salt) western diet. On the other hand, in RALES and subsequently in the similar clinical trial of eplerenone in post-myocardial infarct heart failure (EPHESUS; [4]), aldosterone levels were normal and salt status
unremarkable, raising the question of what the spironolactone or eplerenone was actually blocking to have such a major positive effect. The remainder of this presentation is thus devoted to providing supporting evidence for the contention that many, perhaps most, coronary and cardiac effects of inappropriate MR activation are in fact glucocorticoid-driven, in the context of tissue damage and the generation of reactive oxygen species (ROS). To sustain this contention it is first necessary to revisit some of the historic background findings on MR expression and selectivity in terms of ligand binding. When the human MR was first cloned in Ron Evans’ laboratory [5] and an hMR cDNA used to probe mRNA preparations from rat tissues, MR were found to be expressed not only in epithelial tissues such as kidney and gut, but also in non-epithelial tissues such as hippocampus and heart. Secondly, when the relevant recombinant receptors were expressed in COS cells and the affinity of various steroids determined by their ability to compete with [3 H]aldosterone for binding, it was found that cortisol and aldosterone showed equivalent high affinity for MR and that the affinity of corticosterone was approximately three-fold higher. These findings raised two questions – first, of the potential physiological roles of non-epithelial MR; and secondly, of how aldosterone could ever selectively activate epithelial MR, given that circulating plasma concentrations of the physiological glucocorticoids are three orders of magnitude higher than those of aldosterone. Even factoring in the much higher binding of the glucocorticoids in plasma, their free levels are still a 100-fold those of aldosterone. While the answer to the first of these questions is not yet clear, an aldosterone specificity-conferring mechanism in epithelial cells was proposed shortly afterwards [6,7]. In addition to MR, epithelial aldosterone target cells express at very high abundance the enzyme 11 hydroxysteroid dehydrogenase type II (11HSD2). Expression of 11HSD2 is not confined to classic aldosterone target tissues; it is also expressed, for example, in placenta, amygdala and – importantly for the present discussion – in vascular smooth muscle cells. What 11HSD2 does is to convert cortisol to cortisone, which has negligible affinity for MR, thus putatively excluding cortisol from MR and allowing aldosterone to access and activate them. Aldosterone escapes the attention of the enzyme because in solution the very
Fig. 1. (A) Coronary artery inflammation and perivascular inflammatory cell infiltration in hearts from rats on 0.9% NaCl solution to drink and infused with angiotensin II for 3 weeks. (B) Abrogation of inflammatory response by prior adrenalectomy. (C) Restoration of response by concurrent aldosterone infusion (redrawn with permission from [3]).
J.W. Funder / Journal of Steroid Biochemistry & Molecular Biology 93 (2005) 121–125
Fig. 2. Binding of tritiated aldosterone in epithelial and non-epithelial tissues; relative in vivo receptor occupancy by non-radioactive aldosterone (circles) and corticosterone (squares) coadministered with tracer. Adapted from [8].
reactive aldehyde group at carbon 18 cyclizes with the hydroxyl at carbon 11, forming an 11,18 hemiacetal and thus protecting the 11 hydroxyl from enzymatic attack. When 11HSD2 is not operating—as in the congenital syndrome of apparent mineralocorticoid excess or following licorice abuse or carbenoxolone (glycyrrhetinic acid hemisuccinate) administration, cortisol activates MR producing sodium retention, volume expansion and blood pressure elevation. That the enzyme operates to prevent glucocorticoid occupancy rather than glucocorticoid activation of epithelial MR in retrospect was to draw a very long bow, given the intracellular concentration differences between glucocorticoid and aldosterone. This was shown to be the case in a study on rats, in which the physiological glucocorticoid is corticosterone rather than cortisol. Adrenalectomized rats were injected with [3 H]aldosterone alone or plus half-log increasing doses of non-radioactive aldosterone or corticosterone [8]. Fifteen minutes after injection, the animals were killed and tissues harvested for determination of MR bound [3 H]aldosterone, with the inference that in this way in vivo MR occupancy by aldosterone or corticosterone in various tissues can be established. The results of these studies are shown in Fig. 2. In the heart (lower left) aldosterone appears to have ∼3fold higher affinity for MR in vivo. The heart expresses negligible levels of 11HSD2; corticosterone has ∼3-fold higher intrinsic affinity for MR, but on the other hand is 10-fold more highly bound in plasma; the finding that in vivo aldosterone at equal doses has a three-fold advantage is therefore consistent. For the hippocampus (Fig. 2, lower right), the aldosterone competition curve is shifted far to the right, evidence for the very high reflection coefficient for aldosterone at the blood–brain barrier; corticosterone, despite its higher plasma binding, gets through the blood–brain barrier very much more easily, and thus at equivalent concentrations is actually a more
123
likely occupant. In the classical epithelia tissues (Fig. 2, upper panels) where in contrast to cardiomyocytes and hippocampal neurons high levels of 11HSD2 are expressed, the operation of the enzyme shifts the corticosterone curve to the right by approximately an order of magnitude—i.e. it debulks intracellular glucocorticoid, rather than completely metabolizing it to its inactive 11-keto congener. What this means is that under physiologic conditions, intracellular active glucocorticoid concentrations are still ∼10-fold those of aldosterone and that most epithelial MR are occupied but not activated by cortisol—or, in the rat, corticosterone. When 11HSD2 is blocked, marginally more epithelial MR will be cortisol occupied, but this is an unsatisfactory explanation for a change in activity: maybe, when the enzyme is blocked, something else happens to activate glucocorticoid complexes. The prime candidate for this change-of-state would appear to be the generation of NADH from NAD, the forgotten cosubstrate for 11HSD2. For every molecule of cortisol converted to cortisone, a molecule of NAD is converted to NADH. Resting levels of NAD in the cell are far (∼600 times) higher than those of NADH, so that for minimal changes in substrate NAD levels (e.g. a fall from 600 to 500) an increase in NADH levels of two orders of magnitude could be achieved. It is not clear how the resultant redox change activates MR–glucocorticoid complexes, though redox-dependent transcriptional changes have been previously reported, involving, for example, activation of corepressors [9,10]. A corollary of this potential mechanism for MR is that the conformation and thus the cofactor binding profile of an MR–glucocorticoid complex differs sufficiently from that of an MR–aldosterone (or, for that matter, an MR–spironolactone) complex to allow cortisol this bivalent activity. That such redox-dependent bivalent activity may reflect the operation of additional or alternate mechanisms, suggested by its action in acute non-genomic MR effects, to be discussed later. One indirect but instructive series of studies by Young et al. [11] is shown in Fig. 3. Uninephrectomized rats maintained on 0.9% NaCl solution to drink and given a single injection of deoxycorticosterone (DOC) in oil show promptly increased levels of coronary vascular inflammatory markers (ED-1, COX-2 and osteopontin) 8 days later (Fig. 3). When the animals were given carbenoxolone (CBX) in their drinking fluid, rather than receiving DOC by injection, the vascular inflammatory response in terms of these three markers was identical to that seen with DOC. That this effect of carbenoxolone is mediated via MR is shown for each marker by the ability of eplerenone (EPL) to block the effect of carbenoxolone down to control levels. Our interpretation of these data is that when VSMC 11HSD2 is blocked, the normally inactive corticosterone–MR complexes in VSMC are activated and mimic the effect of administered exogenous mineralocorticoid. That this is of wider relevance to MR pathophysiology than animals receiving carbenoxolone or clinical situations of licorice overindulgence is suggested by the data shown
124
J.W. Funder / Journal of Steroid Biochemistry & Molecular Biology 93 (2005) 121–125
Fig. 3. Response of inflammatory markers to deoxycorticosterone (DOC), carbenoxolone (CBX) or carbenoxolone plus eplerenone (EPL) in coronary vessels from uninephrectomized rats drinking 0.9% NaCl solution for 8 days. Adapted from [11]; (*) P < 0.05.
in Fig. 4. In these studies, pigs were subjected to coronary angioplasty, receiving vehicle, aldosterone or eplerenone for 5 days before and 28 days after the procedure [12]. What eplerenone does under these circumstances is to prevent constrictive fibrosis, and thus preserve the coronary lumen. Control animals received no exogenous mineralocorticoid and were provided with regular pig chow and water to drink: what, then, is the eplerenone antagonizing in vessel wall MR? The answer would appear to be the action of normal levels of cortisol, occupying the majority of VSMC MR and transformed from inactive to active by the redox changes consequent upon generation of reactive oxygen species in the context of tissue damage. If redox changes following 11HSD2 blockade or ROS generation can transform glucocorticoid–MR complexes from inactive to active, then a possible physiological role for cardiomyocyte or neuronal MR in the absence of 11HSD2 may be as follows. In such cells the intracellular concentrations of glucocorticoid are such that MR will be essentially always occupied over the course of diurnal variation in glucocorticoid levels. This contrast with classical glucocorticoid receptors (GR), which have much (∼30-fold) lower affinity for cortisol/corticosterone than MR, and thus are more or less occupied over the range of physiological glucocorticoid levels. An always-occupied MR is difficult to reconcile as responsive to changing glucocorticoid levels; what may determine its response, therefore, is the redox state of the cells.
Both cardiomyocyte and neurons, for example, face particular metabolic challenges; glucocorticoid–MR complexes, activated under conditions of metabolic stress, may thus constitute one of the mechanisms involved in the cellular response to such stress. The extent to which such responses are homeostatic or ultimately pro-apoptotic remains to be explored. Very recently, Mihailidou et al. have provided direct evidence for the importance of intracellular redox state in MR activation by glucocorticoids [13]. When isolated rabbit cardiomyocytes are treated with 10 nM aldosterone, they show a rapid non-genomic response measured as a ∼10-fold increase in pump current within 15 min. Cortisol alone at 100 nM has no agonist effect when given alone and stoichometrically blocks the agonist effect of aldosterone 10 nM down to ∼10% of aldosterone alone levels. When intracellular redox state is altered by infusion of oxidized glutathione (GSSG) directly into the myocytes via the wide-tipped pipette, no change to baseline current is seen. When, however, GSSG infused cells are exposed to cortisol, it mimics aldosterone, producing an agonist effect on pump current. These preliminary studies thus constitute the first direct demonstration of the bivalent activity of physiological glucocorticoids in MR dependent upon the intracellular redox state. In summary, in a broad spectrum of cardiovascular disease aldosterone levels are within the normal range and sodium status unremarkable. What is common to an atherosclerotic vessel and a cardiomyocyte in heart failure is tissue damage and generation of reactive oxygen species. Use of second generation MR antagonists, such as eplerenone and subsequent generations of even more selective potent MR blockers, should thus not be limited to those conditions in which aldosterone is elevated. There will clearly be a major advance over Aldactone in Conn’s syndrome and perhaps in the broader context of patients with inappropriate aldosterone to renin ratios [14]; given their extraordinary vasoprotective effect in animal studies [2,3], even at doses that do not lower blood pressure, they may well have a much broader application across the spectrum of cardiovascular disease. References
Fig. 4. Luminal diameter in pig coronary vessels subject to experimental coronary angioplasty 4 weeks earlier. Adapted from [12]; (*) P < 0.05.
[1] B. Pitt, F. Zannad, W.J. Remme, R. Cody, A. Castaigne, A. Perez, J. Palensky, J. Wittes, The effect of spironolactone on morbidity and
J.W. Funder / Journal of Steroid Biochemistry & Molecular Biology 93 (2005) 121–125
[2] [3]
[4]
[5]
[6]
[7]
[8]
mortality in patients with severe heart failure, New Engl. J. Med. 341 (10) (1999) 709–717. R. Rocha, C. Stier, Pathophysiological effects of aldosterone in cardiovascular tissue, Trends Endocrinol. Metab. 12 (7) (2001) 308–314. R. Rocha, C. Martin-Berger, P. Yang, R. Scherrar, J. Delyani, E. McMahon, Selective aldosterone blockade prevents angiotensin II/salt-induced vascular inflammation in the rat heart, Endocrinology 143 (2002) 4828–4836. B. Pitt, W. Remme, F. Zannad, J. Neaton, F. Martinez, B. Roniker, R. Bittman, S. Hurley, J. Kleiman, M. Gatlin, Eplerenone, a selective aldosterone blocker, in patients with left ventricular dysfunction after myocardial infarction, New Engl. J. Med. 348 (14) (2003) 1309–1321. J.L. Arriza, C. Weinberger, G. Cerelli, T.M. Glaser, B.L. Handelin, D.E. Housman, R.M. Evans, Cloning of human mineralocorticoid receptor complementary DNA: structural and functional kinship with the glucocorticoid receptor, Science 237 (4812) (1987) 268–275. J.W. Funder, P. Pearce, R. Smith, A.I. Smith, Mineralocorticoid action: target-tissue specificity is enzyme, not receptor, mediated, Science 242 (1988) 583–585. C.R. Edwards, P.M. Stewart, D. Burt, L. Brett, M.A. McIntyre, W.S. Sutanto, E.R. de Kloet, C. Monder, Localisation of 11 betahydroxysteroid dehydrogenase—tissue specific protector of the mineralocorticoid receptor, Lancet 2 (8618) (1988) 986–989. J.W. Funder, K. Myles, Exclusion of corticosterone from epithelial mineralocorticoid receptors is insufficient for selectivity of aldos-
[9] [10]
[11]
[12]
[13]
[14]
125
terone action: in vivo binding studies, Endocrinology 137 (12) (1996) 5264–5268. Q. Zhang, D. Piston, R. Goodman, Regulation of corepressor function by nuclear NADH, Science 295 (5561) (2002) 1895–1897. C. Fjeld, W. Birdsong, R. Goodman, Differential binding of NAD+ and NADH allows the transcriptional corepressor carboxyl-terminal binding protein to serve as a metabolic sensor, Proc. Natl. Acad. Sci. U.S.A. 100 (16) (2003) 9202–9207. M. Young, L. Moussa, R. Dilley, J.W. Funder, Early inflammatory responses in experimental cardiac hypertrophy and fibrosis: effects of 11 beta-hydroxysteroid dehydrogenase inactivation, Endocrinology 144 (3) (2003) 1121–1125. M.R. Ward, P. Kanellakis, D. Ramsey, J.W. Funder, A. Bobik, Eplerenone suppresses constrictive remodeling and collagen accumulation after angioplasty in porcine coronary arteries, Circulation 104 (2001) 467–472. A.S. Mihailidou, M. Mardini, T. Fraser, D. Knights, J.W. Funder, Agonist/antagonist activity of cortisol in cardiomyocyte mineralocorticoid receptors is determined by redox state, in: Proceeding of 30th Annual Meeting of the International Aldosterone Conference, New Orleans, 2004, p. 18. O. Olivieri, A. Ciacciarelli, D. Signorelli, F. Pizzolo, P. Guarini, C. Pavan, A. Corgnati, S. Falcone, R. Corrocher, A. Micchi, C. Cressoni, G. Blengio, Aldosterone to renin ratio in a primary care setting: the Bussolengo study, J. Clin. Endocrinol. Metab. 89 (9) (2004) 4221–4226.