Third Generation Mineralocorticoid Receptor Antagonists; Why We Need a Fourth

Abstract

The first mineralocorticoid receptor (MR) antagonist, spironolactone, was developed almost 60 years ago to treat primary aldosteronism and pathological edema. Its use waned in part due to its lack of selectivity. Subsequently knowledge of the scope of MR function was expanded along with clinical evidence of the therapeutic importance of MR antagonists to prevent the ravages of inappropriate MR activation. Forty-two years elapsed between the first and MR-selective second generation of MR antagonists. Fifteen years later, despite serious shortcomings of the existing antagonists, a third generation antagonist has yet to be marketed. Progress has been slowed by the lack of appreciation of the large variety of cell types that express the MR and its diverse cell-type-specific actions, as well as its uniquely complex interactions actions at the molecular level. New MR antagonists should preferentially target the inflammatory and fibrotic effects of MR and perhaps its excitatory effects on sympathetic nervous system, but not the renal tubular epithelium or neurons of the cortex and hippocampus. This review briefly describes efforts to develop a third generation MR antagonist and why fourth generation antagonists and selective agonists based on structural determinants of tissue and ligand-specific MR activation should be contemplated.

Introduction

Spironolactone, the first pharmacological antagonist of a steroid receptor approved for clinical use, was developed to treat primary aldosteronism, hypertension and the edema of chronic heart failure and cirrhosis. It was directed against the Mineralocorticoid Receptor (MR) 30 years before the MR was cloned¹ by basing its structure on that of progesterone, an endogenous antagonist of the MR. Spironolactone is an agonist of the progesterone receptor (PR), an antagonist of the androgen receptor (AR), and at diuretic doses reduces potassium excretion, thus can cause life-threatening hyperkalemia. Despite these serious side effects, it was 40 years before a second generation more selective MR antagonist was marketed. This was due in part to a lack of understanding of the basic biology of the MR, its ligands and their regulation, as well as the mistaken belief that inhibition of the renin-angiotensin-aldosterone system sufficed to silence the many effects of inappropriate MR activation. Before truly selective antagonists, those that are not only selective for the MR and no other receptor, but also selectively inhibit some, but not all, MR functions can be developed, we need to understand how the MR becomes activated, particularly under circumstances in which its ligands are not particularly elevated.

The basic biology of the MR

MR are expressed in a wide variety of epithelial, endothelial and mesenchymal tissues, including cardiomyocytes, vascular endothelial and smooth muscle cells, renal tubular epithelia, neurons, macrophages, and adipocytes. It mediates a commensurately large variety of cell-type specific effects, many of which are unrelated to the vectorial transport of water and ions for which it was named^{2, 3}. The MR, like the other steroid receptors, is a ligand-activated nuclear transcription factor that regulates the expression of genes coding for proteins that mediate its functions. In addition, as do other steroid receptors, MR associated with the membrane initiate rapid events through common cell signaling pathways. While these signaling pathways are understood, the mechanisms by which they are activation by steroid receptors are not clear, nor is the mechanism by which the MR is associated with the membrane. Inactive steroid receptors destined for transcriptional function are bound to chaperone proteins that hold the receptor in a conformation suitable for ligand binding and are shed upon activation by a ligand. Receptors like the MR and glucocorticoid receptor (GR) that in the inactive state reside primarily in the cytoplasm are then transported to the nucleus where they form dimers and recruit transcription cofactors to form a transcription complex with DNA hormone response elements (HRE) to initiate transcription of cell type-specific effector proteins. Chaperone proteins, transcription co-factors and available HRE (those not masked by epigenetic changes) are cell and context specific, thus may provide a target for future manipulation of specific MR mediated functions^4–7.

Structural similarities among the ligand binding domains(LBD) of the MR and glucocorticoid, progesterone, and androgen receptors (GR, PR and AR) and among their endogenous ligands have physiological implications and are important constraints to the development of selective synthetic agonists and antagonists^{3, 8}. The MR has similar affinity for its primary physiological agonists, aldosterone and the main endogenous glucocorticoids cortisol and corticosterone, as well as for 11-deoxycorticosterone (21-hydroxyprogesterone, deoxycorticosterone, DOC) and progesterone^{3, 9, 10}.

DOC was the first adrenal steroid identified and synthesized and its acetate, DOCA, has been used experimentally as a mineralocorticoid, in part due to its ease of synthesis and lower cost compared to aldosterone^{3, 11, 12}. However DOC is not equivalent to aldosterone; it also a glucocorticoid^{12, 13}. Despite ample evidence of its glucocorticoid properties, the belief that DOC is a mineralocorticoid equivalent to aldosterone has prevailed, apparently due to statements published over 50 years ago^{12, 13}. Nor is DOCA equivalent to DOC. DOCA is inactive and requires hydrolysis to DOC by esterases that are not ubiquitous, as once assumed, and are particularly limited in muscle¹². DOC is formed from progesterone by 21-hydroxylase in tissues expressing 21-hydroxylase including the adrenal gland, placenta and ovary^{12, 13}. It is converted in the adrenal zona fasciculata to corticosterone by 11β-hydroxylase (CYP11B1) and in the adrenal zona glomerulosa to aldosterone by aldosterone synthase (CYP11B2). In those species that express 17α-hydroxylase in the adrenal, deoxycortisol is the primary substrate for CYP11B1 and cortisol is the primary glucocorticoid. In mammals DOC normally circulates at low concentrations in comparison to those of aldosterone, with a significant portion, >96% bound by corticosteroid binding globulin (CBG) and albumin, unavailable for receptor binding¹⁴. Accumulation of DOC in patients with 17α-hydroxylase or 11β-hydroxylase deficiency produces a mineralocorticoid excess hypertension^{12, 15}. While plasma concentration may increase by 10-fold in pregnancy due to the increase in its precursor progesterone, inactivation of DOC by the aldo-ketoreductase AKR1C3 is thought to prevent excessive activation of MR in colon and renal tubular epithelial cells by DOC¹⁵. Progesterone is an MR antagonist that binds the MR with similar or slightly higher affinity than aldosterone and attains physiological relevant inhibitory concentrations during pregnancy and at its zenith during the estrus cycle¹⁶.

Glucocorticoid synthesis in mammals is 2–3 orders of magnitude that of aldosterone. Most circulating cortisol and corticosterone, 96% or more, is bound to CBG or plasma albumin¹⁴, notwithstanding the amount of free glucocorticoid still greatly exceeds that of aldosterone even at the nadir of the circadian rhythm of the HPA axis when cortisol concentrations descend to the 3–10 μg/dl range ^{3, 17, 18}. Even in severe Primary aldosteronism (PA) aldosterone concentrations are in the 10s of ng/dl ^19–21. CBG is a crucial regulator of glucocorticoid availability. Elastases produced by inflammatory cells and certain bacteria cleave CBG, releasing cortisol/corticosterone at the site of inflammation and infection^{22, 23} and globally in septic shock ²⁴, thus maximizing MR occupation and increasing GR activation which tends to counter excessive inflammatory and repair response of MR activation. CBG is not merely a plasma transport protein; it is produced in various tissues and is an active participant in local glucocorticoid activity^25–27.

In aldosterone target cells, including renal tubular epithelial and vascular endothelial and smooth muscle cells, cortisol & corticosterone are converted to inactive 11β-dehydro metabolites by the unidirectional 11β-hydroxysteroid dehydrogenase type 2 (11β-HSD2) which thus provides pre-receptor selectivity for the lower concentrations of aldosterone. Glucocorticoids are thought to be the physiological MR ligand in most cells that do not express 11-HSD2, including the cardiomyocyte, mature adipocyte, macrophage, and most neurons ^{10, 27–29}. A select few aldosterone target neurons of the nucleus tractus solitarius that mediate salt appetite co-express MR and 11β-HSD2³⁰. 11β-HSD1 is a bidirectional enzyme that usually acts as a reductase that activates the inactive metabolites 11-dehydrocorticosterone and cortisone and increases intra-cellular concentrations of active glucocorticoid. Selective HSD1 antagonists have been developed to decrease the effects of excessive glucocorticoid action, a major prejudicial factor in stress, obesity, metabolic syndrome, and dementia^{31, 32}. However 11-HSD1 acts as a dehydrogenase in the absence of hexose-6-phosphate dehydrogenase (H6PD) which is required to regenerate NADPH, the obligatory co-factor for reductase activity ^{29, 33}. This occurs in a few types of cells; those of particular relevance to cardiovascular disease are preadipocytes and pre-sympathetic neurons of the paraventricular nucleus^{33, 34}. In cells expressing both MR and GR that do not have prereceptor selectivity conferred by an 11β-HSD, for example many hippocampal neurons and immune cells, it is thought that many or most MR are occupied by cortisol or corticosterone during most of the circadian cycle for glucocorticoids, most or all MR and some GR are occupied during the zenith of the cycle, and that most or all GR are activated by stress levels of glucocorticoids³⁵. The coordinated actions mediated by the MR and GR are critical to normal affect, response to stress, learning, and memory. The ultradian rhythm of glucocorticoid secretion mediates further change in relative MR and GR activation in hippocampal neurons³⁶. Such short term pulses of cort may be most relevant to the rapid non-transcriptional effects mediated by the MR and GR^{37, 38}. The balance between MR and GR mediated effects are crucially important in tissue repair, as well as neuronal function.

MR activation initiates inflammation and repair pathways, while GR activation tends to dampen the responses. Local treatment of a wound with an 11β-HSD1 antagonist decreases local glucocorticoid activity at the site and accelerates wound healing, particularly in the stressed and aged patient ^{39, 40}, while topical application of glucocorticoids suppresses excessive scar formation. Treatment with an MR antagonist also suppresses excessive inflammation and fibrosis after injury^{41, 42}.

Evidence for other mechanisms conferring ligand specificity to the MR has accumulated, but only recently have been identified⁴. For example tesmin is a co-activator for the MR when it is bound to aldosterone and deoxycorticosterone, but not to cortisol or spironolactone⁵. Work in the area of the structural requirements of ligand-discriminant co-activators may solve a mystery that is at the crux of the development of truly selective MR antagonists to treat cardiovascular, renal and metabolic diseases. Even in primary aldosteronism due to an aldosterone producing adenoma the amount of circulating aldosterone does not approximate that of endogenous glucocorticoid agonists, yet clearly patients with primary aldosteronism have significantly more severe cardiovascular remodeling for the same level and duration of hypertension and benefit from removal of the offending tumor or treatment with an MR antagonist^{3, 20, 43}. The MR is reported to be susceptible to activation in the absence of ligand under conditions of high reactive oxygen stress ^{44, 45}, reviewed in reference 3, however there is evidence that the MR cardiomyocytes is quiescent when bound to glucocorticoids under normal conditions, but is activated by oxidative stress⁴⁶. This would explain the beneficial effects of MR antagonists in conditions of excessive MR activation in which aldosterone concentrations are either normal or reduced.

Increased binding of GR by glucocorticoid levels elevated by stress, inflammation or obesity may also play a significant role in increased or altered MR transcriptional activity. The affinity of the GR for cortisol and corticosterone is 10-fold less than that of the MR. Most cells that express MR also express GR. Therefore under non-stressed conditions glucocorticoid-targeted MRs such as those in cardiomyocytes or hippocampal neurons are generally fully occupied by glucocorticoids; both MR and GR are occupied at stress levels of glucocorticoids. In addition to forming homodimers, the activated MR forms heterodimers with ligand activated GR. MR:MR and MR:GR dimers have different transcriptional efficacy in vivo as well as in vitro^47–49, reviewed in reference 10 and these differences may cause further dysfunction in injured tissue.

The First Generation Antagonists

Deoxycorticosterone was isolated 75 years ago based upon its mineral retaining properties, however studies with the purified compound soon demonstrated that in addition to stimulating sodium and water retention in exchange for the excretion of potassium and protons, it caused severe hypertension and heart failure ^{11, 50, 51} preceded by increased vascular tone both due to direct action upon vessels and through an increase sympathetic drive ^51–55. In the early 1960’s, less than decade after the isolation of aldosterone⁵⁶, spironolactone (Aldactone) was developed and approved for the treatment of primary aldosteronism and its associated hypertension, hypokalemia and alkalosis⁵⁷, essential hypertension, and the edema of congestive heart failure and cirrhosis⁵⁸. It and canrenone, 7α-thiomethyl spironolactone, one of several active spironolactone metabolites approved for clinical use in Europe, constitute the first generation of MR antagonists for clinical use⁵⁹. Notwithstanding growing evidence that mineralocorticoids acted directly in many tissues, including vessels, heart and brain ^60–65 and that spironolactone antagonized these effects, the prevailing dogma became that the antihypertensive effect of spironolactone was due solely or primarily to its diuretic and saluretic action⁶⁶, a misconception that lasted several decades. The structure of spironolactone resembles that of progesterone, an endogenous antagonist of the MR. Spironolactone is a PR agonist and AR antagonist within therapeutic ranges for MR blockade. The use of spironolactone and canrenone at doses for potassium sparing diuretic effects, was limited by significant hyperkalemia, as well as progestational and anti-androgen effects causing significant menstrual cycle disruption, gynecomastia and impotence. While lack of receptor selectivity is a significant problem for most uses of spironolactone, the anti-androgenic effect is useful in women with hirsutism, particularly when associated with hypertension, for example in polycystic ovarian syndrome^67–69. Drospirenone, one of a class of 17α-pregnane-21,17-carbolactones with 15,16-βmethylene modifications developed by Schering AG, now Bayer Healthcare^{59, 70} is a potent synthetic PR agonist and MR and AR antagonist currently used in birth control and menopausal hormone replacement regimens in combination with an estrogen. It is significantly more potent as an MR antagonist than spironolactone and has been suggested as a treatment for hypertension in women ^{71, 72}. Thus lack of receptor selectivity is an advantage under select circumstances, however as with other oral contraceptives, the risk for thrombosis of estrogen+drospirenone preparations must be assessed for each patient^{73, 74}.

An effort was made by several laboratories during the 1980’s to develop more selective MR antagonists^{59, 75}. Roussel-UCLAF developed highly soluble potent 7-alkyl spironolactone MR antagonists which were used for research but were not marketed for clinical use ^{76, 77}. RU28318 was used to definitively demonstrate the critical importance of the MR in normal hippocampal neuronal function mediated by cortisol & corticosterone⁷⁸ and of MR in the central modulaton of blood pressure by mineralocorticoid excess and in salt sensitive rats^{77, 79}. Ciba-Geigy produced a class of more selective MR antagonists by incorporating epoxy groups into spironolactone derivatives^{80, 81}, however testing and marketing of one of these, eplerenone, was delayed for 2 decades (and several pharmaceutical company restructurings), reviewed in reference 82.

Meanwhile, due to their side effects clinical use of spironolactone and canrenone as antihypertensive agents waned in favor of angiotensin converting enzyme inhibitors and later, angiotensin type 1 receptor (AT1R) antagonists that when paired with diuretics were thought to suppress the pernicious effects of excessive renin-angiotensin-aldosterone system (RAAS) activity, including that of aldosterone, notwithstanding the knowledge that aldosterone production often escaped control of the RAAS after chronic RAAS suppression^82–84. The accumulation of evidence from animal studies demonstrating that inappropriate activation of MR in the heart, vessels and kidneys led to inflammation, hypertrophy and fibrosis that were not prevented by angiotensin converting inhibition and were independent of hypertension ^85–90, led to the Randomized Aldactone Evaluation Study (RALES). The RALES trial was stopped early when it became clear that addition of a low dose of Spironolactone to standard therapy of patients with severe congestive heart failure significantly reduced death from all causes, including sudden cardiac death, as well as improved quality of life ⁹¹.

The second generation MR antagonist

Early results of the RALES trial spurred the completion of clinical testing and marketing in 2002 of 9–11α-epoxymexrenone or eplerenone (Inspra). Eplerenone is selective for the MR, however its low potency and difficulty, thus expense, of synthesis in comparison to spironolactone had retarded its testing for market^{59, 92}. The Eplerenone Post-acute Myocardial Infarction Efficacy and Survival Study (EPHESUS) ⁹³ and subsequent randomized clinical trials confirmed that the addition of either spironolactone or eplerenone to standard treatment regimens reduced arrhythmias and delayed pathological remodeling of the vessels, heart and kidneys in various clinical conditions, many of which are not associated with increased circulating aldosterone ^94–96. The doses used were below those required to lower blood pressure or have a diuretic effect. As predicted from animal studies, some done more than 50 years earlier ^{60, 97}, spironolactone and eplerenone were shown to decrease cardiac arrhythmias in patients through direct effects on cardiomyocyte excitation, as well as by decreasing sympathetic nervous system activation and normalizing hypokalemia and hypomagnesaemia, including in those with coronary artery disease and chronic kidney disease without heart failure^{96, 98–103}.

Evidence that mineralocorticoids enhanced central sympathetic excitation leading to hypertension accrued since the late 1970s^{63, 65, 104}. The advent of specific antibodies and tracers allowed the demonstration of MR in hypothalamic presympathetic neurons³⁴. Spironolactone and eplerenone are lipophilic and cross the blood brain barrier^{105, 106}. Lowering blood pressure with a thiazide diuretic as standard first line treatment of hypertension often activates the SNS, which among other effects, increases blood glucose. Addition of a low dose of Spironolactone to the diuretic or substituting the diuretic with the MR antagonist resulted in a similar decrease in blood pressure without increasing sympthoexcitation and producing insulin resistance^{107, 108}. Judicious use of an MR antagonist while monitoring serum potassium is now advocated in diabetic nephropathy to slow the progress of renal failure without further perturbing glycemic control ⁹⁴. The potential for serious hyperkalemia must always be considered even in patients with no known renal impairment, particularly if the MR antagonist is added to an antihypertension regimen that includes a moderate to high dose of an angiotensin ll receptor blocker or converting enzyme inhibitor.

Comparison of the First and Second Generation MR antagonists

The clear advantage of eplerenone over spironolactone is its MR selectivity and lack of effect on the AR and PR. Both eplerenone over spironolactone increase the incidence of life-threatening hyperkalemia due to renal tubular effects that is mitigated by using lower doses. Some spironolactone metabolites are also potent MR antagonists that extend its effective half-life to about 16h and allow its once a day dosing with lower peak effects⁵⁹. Eplerenone is metabolized by the CYP3A4 and CYP3A5 enzymes to form inactive metabolites. While its half-life is only 3–4h, the natriuretic effects of eplerenone persist for 12h, perhaps because the eplerenone-bound MRs remain in an inert conformation that is less dynamic that of MR bound to an endogenous ligand that is released relatively rapidly after transcription, allowing the receptor to be translocated to the cytosol ^{59, 109–111}, thus eplerenone is dosed twice a day despite its very short half-life. A comprehensive survey of clinically relevant interactions between eplerenone and other drugs that are metabolized by CYP3A enzymes has yet to be published ⁹², however dosing of eplerenone is complicated by certain genetic polymorphisms of CYP3A and co-administration of eplerenone with other agents that alter CYP3A activity ^{59, 92, 112, 113}. A major advantage of spironolactone over eplerenone is cost and once a day dosing. Spironolactone was a significantly better antihypertensive than eplerenone in two clinical studies, perhaps because of its greater potency and longer half-life^{59, 114, 115}. Despite side effects, results of clinical trials continue to show therapeutic benefits of MR antagonists in a growing list of conditions^{2, 96, 116, 117}, clearly indicating the need for better MR antagonists^{59, 75, 118–120}. The Dialysis Outcomes Heart Failure Aldactone Study and others have been demonstrated their utility in stemming the progression of chronic kidney disease ^121–123. Another is the dramatic discovery that central serous chorioretinopathy, a rapidly progressing and heretofore untreatable cause of permanent destruction of the retina, is reversed by spironolactone and eplerenone^{124, 125}.

The need for better MR antagonists prompted John Funder to challenge the research community to develop third and fourth generations of MR antagonists, emphasizing that the new drugs should also have long patent lives to allow ample time for the recuperation of the cost of drug research and development at a reasonable price for the patient¹²⁶. Ideal third generation MR inhibitors were defined as being potent, selective and affordable, to be used in patients with primary aldosteronism, now recognized as comprising ~10% of treatment resistant hypertensive patients^{20, 127}. Fourth Generation MR Antagonists were defined as agents that target MR-mediated processes contributing to vascular dysfunction, but preserve the ability for MR to regulate renal excretion of potassium¹²⁶. I would propose that MR in neurons of the hippocampus and cerebral cortex should also be spared.

The third Generation MR Antagonists, some of which are now in clinical trials, are non-steroidal, potent, MR selective, and should be easier, thus less expensive, to synthesize than Eplerenone. None have been approved by the FDA for clinical use at this time, but not for lack of effort. At the time of this writing, 273 studies were listed for mineralocorticoid receptor antagonist on ClinicalTrials.gov. Some, in addition to good MR selectivity and favorable metabolic profiles, are reported to have proportionately greater antagonism of MR-mediated cardiovascular inflammation, dysfunction and remodeling than upon MR-mediated exchange of sodium for potassium and hydrogen ions in renal tubular epithelia, however the mechanisms for this functional selectivity are not discussed.

The Dihydropyridine-like MR antagonists

It was shown 30 years ago that dihydropyridine voltage-dependent calcium channel blockers decreased MR activation by blunting the release of intracellular calcium that activates aldosterone synthesis by adrenal zona glomerulosa cells^128–130. It was later demonstrated that dihydropyridines also bind and antagonize the MR directly^131–134. Spironolactone, eplerenone and progesterone facilitate the translocation of MR from the cytosol to the nucleus, but prevent recruitment of transcriptional co-activators. In contrast, due to their size and conformation, dihydropyridine-bound MR, including MR with the S810L mutation, are not translocated to the nucleus^{59, 131, 133, 134}. This property is particularly useful in the treatment of patients with the MR S810L mutation responsible for a very severe form of familial pseudohyperaldosteronism which is greatly exacerbated by pregnancy. The constituent activity of the MR S810L is significantly enhanced by progesterone, spironolactone and eplerenone which promote its translocation to the nucleus, but do not prevent its transcriptional activity¹³⁵. Dihydropyridine-type MR antagonists would be the first selective treatment for patients with the S810L mutation of the MR.

A class of cyanoester dihydropyridines MR antagonists that are potent, selective and orally available, as well as devoid of calcium channel-blocker activity, were developed at Pfizer Pharmaceutical research labs by introducing a cyano group at C3 and various R6 substitutions^{132, 136}. One of these compounds performed as well as eplerenone in decreasing hypertension in the Dahl salt sensitive rat¹³⁶. Bayer Pharm AG has also developed a class of cyano-dihydropyridines with similar properties and promising preclinical profiles¹³⁴. Among these are the BR-4628 (ethyl (4 R)-5-acetyl-2,6-dimethyl-4-(2-methyl-4-oxo-4 H -chromen-8-yl)-1,4-dihydropyridine-3-carboxylate) that performed very well in preventing adverse tissue remodeling in animal models of mineralocorticoid excess or salt-sensitivity despite mild progestational effects ^{134, 137}. This led to the development of the dihydronaphthyridine BAY 94-8862, or finerenone^138–140, the only third generation non-steroidal MR antagonist registered for a phase 3 clinical trial as of 9/2015, though recruitment has not yet begun (ClinicalTrials.org).

Several multicenter, randomized, double-blind, placebo-controlled ARTS (minerAlocorticoid Receptor antagonist Tolerability Studies) are in progress to determine the safety and efficacy of finerenone in heart failure, diabetes and nephropathy^{116, 139, 141, 142}. Only some studies have an active-comparator (spironolactone or eplerenone) arm. The phase 2a ARTS- HFrEF trial compared the effects of finerenone, spironolactone or placebo added to standard therapy of patients with heart failure with reduced ejection fraction and mild renal compromise^{143, 144}. The phase 2b ARTS-HF trial is to compare finerenone to eplerenone ¹⁴¹. Finerenone and spironolactone decreased albuminuria and brain natriuretic peptide, markers of renal and heart function, respectively, by similar amounts, but the increase in serum potassium, incidence of hyperkalemia and worsening of renal function were less in the finerenone group compared to spironolactone. The mechanisms for the decreased inhibition of MR mediated sodium/potassium homeostasis was not addressed. Spironolactone lowered the blood pressure to a significantly greater degree than finerenone in this study¹⁴³, so the lower effect on hyperkalemia might have been due to a lower effective dose of finerenone. However, spironolactone crosses the blood brain barrier and decreases peripheral sympathetic drive^{3, 34, 104, 105, 107}. This is particularly relevant to the potential for increased gluconeogenesis and insulin resistance due to sympathetic nervous system activation in patients that already have sympathoactivation^{3, 107, 108}. Whether finerenone crosses the blood brain barrier to access presympathetic neurons is not published. The authors of the phase 2a ARTS-HFrEF report point out that the hypertension in patients in this study was adequately controlled before the addition of the MR antagonists, however most were on a combination of anti-hypertensive drugs including β-blockers. A reassessment of the step-wise increase of drugs in multi-drug anti-hypertensive regimens is in order with what is now recognized as a clinically significant role of MR antagonists in the control of inappropriate activation of the sympathetic nervous system ^{10, 107, 108, 145}. This may be particularly relevant for patients with type 2 diabetes¹⁰⁸. The ARTS-Diabetic Nephropathy (ARTS-DN; NCT01874431) compared the effect of the addition of finerenone to standard treatment in diabetics with mild-to-moderate chronic kidney disease receiving either an angiotensin type 2 receptor or converting enzyme inhibitor on albuminuria and serum potassium^{142, 144}. A steroidal MR antagonist was not included in the study, however finerenone decreased albuminuria to a similar degree as that reported for spironolactone in the previous ARTS-HF trial, and decreased the decline of renal function compared to placebo^{142, 146}. It would be very interesting to compare the effect of finerenone and spironolactone sympathetic activity and glycemic control.

The Pyrazoline-like MR antagonists

Pyrazoline derivative MR antagonists have been developed and tested at the Pfizer labs¹⁴⁷. One of these, PF-3882845 (3S, 3aR)-2-(3-chloro-4-cyanophenyl)-3-cyclopentyl-3,3a,4,5-tetrahydro-2H-benzo[g] indazole-7-carboxylic acid), is orally available, potent, selective for the MR, and significantly decreased the renal pathology and loss of function in an aldosterone + salt excess model with less risk for hyperkalemia as compared to eplerenone ^{147, 148}. The first clinical phase I trial was terminated due to safety concerns. Results of another (NCT01314898) that evaluated urinary sodium/potassium ratio was presumably finished early in 2015 (https://clinicaltrials.gov), but results have not been published. Preclinical data for another promising pyrazoline derivative from Pfizer, (R)-14c, [(R)-6-(1-(4-cyano-3-methylphenyl)-5-cyclopentyl-4,5-dihydro-1H-pyrazol-3-yl)-2-me thoxynicotinic acid], was recently described¹⁴⁹.

The Sulfonamide-based non-steroidal MR antagonists

The development of aryl sulfonamide-based non-steroidal MR antagonists by Pfizer that are smaller, more polar, with better pharmacodynamics and safety features than the pyrazoline PF-3882845 has been described, but is still in the pre-clinical phase of evaluation¹⁵⁰. Another sulfonamide MR antagonists has been studied further. SM-368229 (N-4,4-dimethyl-2-thioxo-1,4-dihydro-2H-3,1-benzoxazin-6-yl-thiophene-2-sulfonamide) is one of a series of sulfonamide derived MR antagonists synthesized in by the Dainippon Sumitomo Pharma Research Laboratories¹⁵¹. SM-368229 is a selective MR partial agonist-antagonist about 10-fold more potent than spironolactone which was as effective as spironolactone in preventing hypertension, proteinuria and pathological cardiac and renal remodeling, as well as the increase in markers of oxidative stress, inflammation and fibrosis markers in the aldosterone + salt excess model. The authors credit the partial agonist activity of SM-368229 for the decreased potassium retention in animals receiving SM-368229 in comparison to spironolactone at the therapeutic doses^{152, 153}. SM-368229 has no PR activity, but is a weak AR antagonist¹⁵¹. This company has also developed an interesting compound, DSR-71167, (2-([(2,2-difluoroethyl)amino]methyl)-29-fluoro-N-(3-methoxy-4-sulfamoylphenyl)biphenyl-4-carboxamide hydrochloride) that is a selective MR antagonist and weak carbonic anhydrase inhibitor which lowered the blood pressure without increasing plasma potassium in the DOCA-salt and Dahl salt sensitive models¹⁵⁴. These two compounds are still in early development.

Other potential MR antagonists

Clinical trials for several other MR antagonists have been conducted for which details of the compound are not readily available. Two in early stages of clinical testing in patients with hypertension and diabetic nephropathy is MT-3995 (Mitsubishi) and SC-3150 (Daiichi-Sankyo)^{123, 140, 155}. Phase II clinical trials have been completed for LY2623091 (Eli Lilly) was compared to a placebo and spironolactone in patients with primary hypertension (NCT02194465) and to eplerenone, but not a placebo, in men and women with chronic renal disease (NCT01427972). Interaction of LY2623091 with drugs that commonly interfere with CYP450 metabolism was also tested (NCT02300259). While these studies have been closed (www.clinicaltrials.gov) data has yet to be published. Oxazolidinedione derivative MR antagonists are being developed and studied at the Merck Research Laboratories, but as yet published information is limited to interesting preclinical studies ^156–158.

Despite ever growing evidence of the value of MR antagonists to treat diseases that are increasingly prevalent in the Developed World, major short comings of spironolactone and eplerenone, and concerted effort in drug development in this area for at least 20 years, third generation potent, non-steroidal, MR-selective antagonists have been created, but have yet to be marketed¹²⁰. The reasons surely vary, but probably combine a small improvement over the existing clinically approved MR antagonists and the cost of bringing a drug to market. The new generation MR antagonists appear to be selective for the MR, have a favorable pharmacokinetic and pharmacodynamic profiles with minimal interference with other potential therapeutic agents, and target excessive MR-mediated oxidative stress, inflammation and pathological structural remodeling. However evidence that any have less effect on potassium excretion than steroidal MR antagonists is not particularly strong, nor have mechanisms for sparing the effect of MR in renal tubular transport of ions been described even for finerenone, the closest candidate to becoming clinically available.

Partial MR agonists and those combining a third generation MR antagonist and carbonic anhydrase inhibitor are being developed to separate sodium excretion from potassium retention MR inhibition ^151–154.

Perspective: Fourth Generation MR Antagonists?

In addition to sparing the MR-medated regulation of ions in renal epithelia, inhibition of MR function in neurons is a complex issue that should be addressed in the development of antagonists of specific functions of the multi-tasking MR. The ability of non-steroidal third generation candidates to cross the blood brain barrier has not been reported, but lack of suppression of MR in pre-sympathetic neurons of the hypothalamus that are protected by the blood brain barrier may explain why finerenone was less effective than spironolactone in lowering the blood pressure in clinical studies. However, with the increasing incidence of obesity and metabolic syndrome, avoiding the effects of accelerated sympathetic drive on glycemic control should be a goal^{107, 108}. Conversely, the greatest number and concentration of MR in the body are in the hippocampus where they are required for a normal response to stress and for learning and memory^{3, 10, 32, 38, 159–161}. MR-mediated events are also required for neurogenesis, differentiation and migration, now known to be important for the adult, as well as developing human brain^{162, 163}. Neurons of the hippocampus express both MR and GR. As described above, both are thought to be occupied and activated by cortisol or corticosterone differentially depending on the circadian, metabolic and stress related changes in cort levels. A balance between activated MR and GR mediated function is crucial to normal affect and behavior. Abnormal HPA axis function and a decrease in the MR:GR ratio is seen in the brains of persons with major depressive disorder^164–166. Spironolactone impaired neuronal function and cognition in healthy humans and rats ^167–169, as did MR gene deletion in mice ^{162, 170, 171}. Aldosterone treatment restored impaired neuronal electrical activity underlying learning deficits in diabetic rats ¹⁷² and addition of an MR agonist to standard SSRI and SNRI therapy was found to significantly ameliorate the response in patients with major depression ¹⁷³. In both of these instances the balance between MR relative to GR actions was ostensibly restored. In stark contrast, treatment with spironolactone or eplerenone has been shown in multiple clinical studies to significantly improve mood, cognition and quality of life of patients with primary aldosteronism or severe cardiovascular disease ^{141, 174, 175}. The difficulty in parsing the effects of an MR antagonist on cognition is in the separation of the effect of trophic actions of MR within neurons from the devastating effects of cardiovascular disease on the function and integrity of the cerebral vessels, the microvasculature in particular, that leads to neuronal dysfunction and death^176–178. Once the neuronal destruction of vascular dementia occurs, inhibitory effects on the MR within hippocampal and cortical neurons are probably moot. An agent that prevents MR mediated inflammation, loss of cerebrovascular autoregulation and pathological vascular remodeling would protect the neurons.

It is a daunting proposition to inhibit excessive MR-mediated activation of pre-sympathetic neurons of the hypothalamus while preserving MR function of the “learning & memory” neurons in the hippocampus and cortex, however there are differences in MR-mediated transcriptional function that may allow selective repression or stimulation of the transcription of specific genes. The MR antagonists developed so far act as passive antagonists that prevent binding of the activated receptor to transcription co-activators or do not allow the ligand-bound receptor to enter the nucleus ¹³⁴. This approach does not take into consideration the many functions of the MR. Analysis of the structural basis for the differential recruitment of co-transcription factors based upon whether the MR is bound to aldosterone or cortisol has progressed to the point that it may soon be feasible to develop active antagonists or partial agonists that produce a ligand-MR conformation that allows the recruitment of only certain co-factors, thus select the genes would be transcribed ^{4–7, 179, 180}. For example, MR is a transcription factor for NADPH oxidase, crucial for many cell processes, which in excess is responsible in part for oxidative damage caused by excessive MR activation^181–183. MR also promotes the transcription of Nfr2, another transcription factor that promotes genes in major antioxidant pathways. However indirect stimulation of glutathione production by MR is not enough to neutralize the MR-mediated reactive oxygen species generation under pathological conditions¹⁸⁴. Might it be possible to make ligands that would selectively promote recruitment of co-activators for Nfr2 transcription? Such studies should also elucidate how injury to the heart resulting in the accumulation of reactive oxygen species propels glucocorticoid-bound MR into pro-inflammatory and pro-fibrotic transcriptional activities⁴⁶. Selective MR agonists are being considered in the treatment of depression and cognitive decline^{172, 173} (and Ronald de Kloet, personal communications). An approach that targets the conformation of the receptor which promotes or prevents the binding of specific co-transcription factors heralds a new age of MR agonists, as well as antagonists.

Another target that may be fruitful for the development of fourth generation MR antagonists (or agonists) is the mechanism for the formation of MR and GR homodimers and heterodimers, as these appear to have different transcriptional potential^{47–49, 185, 186}.

Two more strategies designed to alleviate excessive MR activation may be in the offing that deserve mention, though they are peripheral to the topic of MR antagonists. Specific antagonists of CYP11B2, or aldosterone synthase, the enzyme responsible for the last step in aldosterone synthesis are being developed that would be useful in the treatment of primary aldosteronism ^187–189. However, as discussed above, aldosterone levels are not elevated or are even suppressed in many of the syndromes in which MR antagonists are effective, including in the Dahl Salt-Sensitive rat that was frequently used to test the Third Generation MR antagonists described above^{3, 10}. The second strategy involves the interaction between the MR and G protein-coupled protein receptor, GPER, formerly GPR30. GPER is activated by physiological levels of aldosterone, but only supra-physiological concentrations of estrogen^190–192. In addition to being a candidate for mediating those rapid non-genomic effects of aldosterone that are not inhibited by MR antagonists^{193, 194}, effects mediated by GPER activation are cell and context dependent and appear to act in opposition to some deleterious effects of MR activation^{191, 195–197}.

Conclusion

More than half a century after developing the first MR antagonist we have yet to develop an ideal antagonist that will mitigate the cardiovascular, renal and metabolic pathology associated with inappropriate MR activation without interfering with important homeostatic functions. This is due in great measure to the many roles of the MR and the complexity of its interaction with different physiological ligands under differing redox conditions, as well as its molecular and functional interaction with the GR with which it shares ligands. Thus it does not suffice to have a potent selective MR antagonist with excellent pharmacokinetic and pharmacodynamic properties. A better understanding of the molecular constraints of co-factor binding and how pathological concentrations of reactive oxygen species alter the structure of the ligand-activated MR and dictate its transcriptional activity should provide the basis for developing targeted MR antagonists and agonists.

Clinical Studies for New Mineralocorticoid Receptor Antagonists: 2011–2015

	trials	subjects	compared	outcomes	Notes
Phase 1
PF-03882845 (pyrazoline~) Pfizer	4 different safety, PK/PD trials	100 H +/−simvastatin	placebo	4 Short studies PK/PD potassium	5 completed
PF-03882845		12 DM	placebo	Serum K	safety concern terminated
PF-03882845	NCT01488877	6 DM+DN	placebo spironolactone	PK/PD Serum K albuminuria	>1 severe hyperK Terminated
LY2623091 Eli Lilly (structure not published)	NCT02300259	48 H		PK/PD (CYP3A –drug interactions)	nsrp
LY2623091	NCT02242981	6 H m	14C labeled-LY2623091	PK/PD	nsrp
LY2623091	NCT01237899	32 H, m&f	placebo	Serum K	nsrp
Phase 2
BAY 94-8862 Finerenone (dihydropyridine)	NCT01345656	CHF+CKD 2a) 60 2b)360	placebo spironolactone	PK/PD Serum K; CV & renal biomarkers	nsrp
	NCT01955694	CHF+ LVSD +/−DM2 +/− CKD 72	placebo or * eplerenone 25mg q48h*	N-t BNP Serum K	nsrp
	NCT01807221	worsening HF+CKD 1058	placebo or spironolactone placebo or eplerenone	CV & renal biomarkers serum K	nsrp
	NCT01874431	DM2+DN 823	(+ RAS inhibitor) placebo	Albuminuria, serum K	nsrp
	NCT01968668	P 2a; DM2+DN (Japan) 96	placebo	Albuminuria, serum K	nsrp
LY2623091 Eli Lilly	NCT02194465	306 hypertension	Placebo, spironolactone +/− tadalafil	Change in BP	nsrp
LY2623091	NCT01427972	48; CKD; m&f	placebo or eplerenone	Proteinuria; PK/PD Serum K	nsrp
MT-3995	NCT01889277	DM2+DN	placebo	Albuminuria, serum K	nsrp
Phase 3
BAY 94-8862 finerenone	NCT02540993	P3: DM2+DN 4800 planned	placebo	Albuminuria, serum K	Enrollment not started

^*dose of eplerenone low even for CKD

ClinicalTrials.gov + references; H-healthy volunteers; DM diabetes mellitus; CHF cardiac failure; LVSD-left ventricular systolic dysfunction; DN-diabetic nephopathy; CKD-chronic kidney failure; RAS inhibitor- angiotensin converting enzyme or receptor inhibitor; nsrp: no study results posted