1.4 Hormones and receptors: fundamental considerations

The original endocrine physiologists viewed hormones as responses to homoeostatic challenge, any signal a call to arms; the word is thus derived from the classical Greek ωρμαειν—‘to arouse’. In the twenty-first century a hormone is a molecule—small or large, protein or lipid—secreted in a regulated fashion from one organ and acting on another. The definition is firmly based on the anatomy of the seventeenth century, the histology of the nineteenth, and the physiology of the twentieth. It has been shaped by convention and clinical specialization: gut hormones are the marches between endocrinology and gastroenterology, and the adrenal medulla the territory of the cardiovascular physician. It has been refined by concepts of paracrine—where the secretion of one cell type in a tissue acts on another cell type in the same tissue—and autocrine, where a particular cell type both secretes and responds to a particular signal. Inherent in the concepts of paracrine and autocrine are that the signal is not secreted into blood or lymph, to be distributed more or less throughout the body, but is made locally to act locally. A very good example of a signalling system with both paracrine and autocrine activities is the neuronal synapse.

Inherent in the concept of the signal is that of a receptor: a signal without a receptor is the sound of one hand clapping. Inherent in the concept of a receptor are two functions: that of being able to discriminate between different signals, and to propagate the signal by activating cell membrane or intracellular signal transduction pathways. Discrimination by a receptor between different circulating potential signals is, in the first instance, a function of the likelihood of a particular signal being able to interact with the receptor, for a period of time sufficient to alter the confirmation of the receptor and thus to trigger propagation. This interaction is commonly referred to as binding, and thus the circulating hormone as a ligand (that which is bound). If the structures of ligand and receptors are such that the initial interaction is followed by formation of strong intermolecular bonds between the two, lessening the possibility of dissociation and the receptor returning to an unliganded state, the receptor is said to have high affinity for the ligand (and vice versa). If the binding is followed by propagation of the ‘appropriate’ signal the ligand is classified as an agonist, or active hormone; if a molecule occupies the binding site on the receptor but does not so alter its structure as to propagate a signal, it is classified as a hormone antagonist (and often, by extension, a receptor antagonist). In the past couple of decades, the concepts of ‘agonist’ and ‘antagonist’ have needed to be refined, as noted subsequently in this chapter.

In symbols, the reversible interaction between hormone and receptor can be simply written as follows;

where [H] is the concentration of hormone, [R] the concentration of empty or unliganded receptor, and [HR] the concentration of occupied receptor, i.e. hormone-receptor complexes. The forward (to the right by convention) or association reaction is equally a function of hormone and receptor concentrations; the association rate constant [K1] is a reflection of the likeliness of apposition/goodness of fit of hormone and receptor, reflecting their structures plus extrinsic factors such as temperature, ionic strength of the milieu, and unstirred layers. The actual rate of the forward reaction is thus mulitfactorial, a function of the rate constant, the concentration of hormone, and the concentration of receptor, or

The dissociation of hormone receptor complexes [HR] is driven by one thing, and one thing only, the dissociation rate constant [K-1], a measure of the inherent probability of the two entities falling apart, under particular conditions of temperature, ionic strength, etc. The actual rate of dissociation is thus the product of K-1, the dissociation rate constant, and the concentration of hormone-receptor complexes, or

At equilibrium, by definition, the rates of the forward and reverse reactions are equal, i.e. for every molecule of hormone that associates with a receptor molecule, a preformed hormone–receptor complex dissociates, or

By simple rearrangement, this can be rewritten

The quotient of the two rates constants (K-1/K1) is termed the dissociation constant or Kd; its reverse (K1/K-1) is the less commonly used Ka or association constant of the reaction. The key outcome of all this relatively simple mathematics is to put a value on Kd, as a measure of affinity, or overall probability of the hormone–receptor complex being in existence, as follows.

If we where to choose a concentration of hormone which would half saturate the receptors, then [R] would equal [HR]. Under such circumstances the two terms can be cancelled in (6) above, and

where Kd equals [H], the hormone concentration at which half maximal receptor occupancy is achieved, and which has the dimensions of concentration, that is, molar.

From equations (1)–(7) there are a number of things that flow. First, in a simple binding system the dissociation of hormone from receptor is not accelerated by addition of excess hormone. What this does, when, for instance, 1000-fold nonradioactive hormone is added to a system containing tracer hormone–receptor complexes, is to operationally prevent (i.e. dilute 1000-fold) reassociation of tracer to receptor. Under such conditions then, the disappearance of tracer–receptor complexes over time thus provides an accurate estimate of the dissociation rate. There are receptors that oligomerize: in such circumstances binding of ligand can increase or decrease the affinity of the other binding sites for hormone, termed positive and negative cooperativity, respectively. Dissociation of bound tracer, for instance, is accelerated in systems displaying negative cooperativity.

Secondly, dissociation constants can only be derived from equilibrium studies, that is, those in which the rates of forward and backward reactions are equal. The association rate constant and dissociation rate constant are often very different, and are constant for a given set of physical circumstances; the actual rates of association and dissociation are determined by not just these constants, but also by the concentration of reactants, as noted above. Where this concept of equilibrium comes into play is in situations where binding is covalent, or essentially irreversible; under such circumstances Scatchard analysis, for example, is inappropriate for determining Kd. A practical case in point is triamcinolone acetonide (TA), a powerful synthetic glucocorticoid in clinical use, which (in contrast with dexamethasone or the physiological glucocorticoids) requires approximately 24 h to come into equilibrium in glucocorticoid receptor binding systems in vitro at 4 C; exposure for shorter time points will consistently underestimate the affinity of TA for the glucocorticoid receptor. Third, different binding systems respond differently to changes in physical conditions. Cortisol, for example, binds transcortin with an order of magnitude higher affinity at 4° C than at 37° C, across a number of species, and with clear differences in binding at physiologically relevant temperatures. In contrast, cortisol binding to glucocorticoid receptors is not particularly temperature dependent, but if anything is of a higher affinity at physiological than at lower temperatures.

Finally there is the inherent bias of endocrinology, that of seeing high-affinity binding as good (‘binds well to the receptor …’), and lower affinity binding as less good (‘binds poorly …’). The underpinnings of this bias is twofold, one theoretical and the other practical. Practically, particularly in often unstable broken cell preparations, the absence of high-affinity binding equates to experimental failure, a powerful driver of emotive language. Even if no experiment ever failed, however, an endocrinologist’s bias is to regard high-affinity binding as good, for the following reason. The higher the affinity the lower the concentration of signal required to half-maximally occupy, and, other things being equal, activate the ‘cognate’ receptor. There are two consequences of this, one of which appears to be biologically sound, the other less so. The latter is a notion of economy; that it is better for an organ to make less rather than more signal, in that it poses less of a demand on precursors and metabolism. This is experientially not the case; every molecule of thyroglobulin, with a molecular weight in excess of 600 000 yields 4–16 molecules of thyroxine, at first sight an example of conspicuous biological extravagance. The other concept underlying the bias has more biological purchase, in that the higher the concentration required to activate cognate receptors, the more likely is the hormone to cross-react with other receptors, acting as an agonist or antagonist, and thus reducing the specificity of the signalling system. It is, of course, entirely possible that there have evolved circumstances in which such ‘cross-reactivity’ may reflect physiology, and that our bias is Ockham’s razor cutting too close to the bone: on the whole, however, such a degree of cautious reductionism appears justified.

In contrast with the previous discussion, if we take a broader biological view that low-affinity binding can be ‘good’—when it enables rolling of platelets or leucocytes on endothelium, giving them time to ‘sniff the wind’ in terms of damage or inflammation. It is also not only advantageous, but functionally required, within the nervous system, where low-affinity binding of signal to receptor is a necessity for the time constants of neurotransmission.

When the electrical impulse underlying nerve conduction is translated into a chemical signal at a synapse or neuroeffector junction, minute quanta of neurotransmitter are released. Because the space into which the neurotransmitter is released is even more minute, the concentration of neurotransmitter becomes very high, so that receptors are rapidly occupied and activated. To achieve this, the ‘on-rate’ of neurotransmitter-receptor binding must be very rapid; and the off-rate (in contrast with hormone-receptor interactions) must also be very rapid, to enable the receptor to return to ground zero. Signal is rapidly cleared by reuptake, diffusion, and metabolism, so that quantal release of signal is followed essentially stochastically by a single response.

To achieve this rapid onset rapid offset binding and activation by neurotransmitters, receptors have to be low affinity, to allow the time constants that characterize neurotransmission. The nervous system does it by mass, ‘brute-forcing’ occupancy of low-affinity receptors, with a restricted spatial distribution of the mass of signal to allow the very high concentrations required, and very efficient mechanisms of rapidly reducing signal concentration. Reflecting this difference, hormones have time constants of minutes, hours, and days compared with the nervous system’s milliseconds; the endocrine system sacrifices time to allow its signals to be distributed all over the body, to ‘arouse’ the diversity of cells that express receptors to which the particular signal can bind. Its signals are broadcast like radio, in contrast with the nervous system landline telephone network.

One striking anthropomorphic illustration of this difference may be worth a thousand words of theoretical justification. First, picture a hummingbird in the National Geographic, its wings still blurred despite shutter speeds of 1/500 or 1/1000 of a second. If acetylcholine had the same high affinity for its receptors at the neuromuscular junction as progesterone has for progesterone receptors, then a hummingbird could beat its wings twice a minute, aerodynamically challenging and clearly no evolutionary advantage. Even less of an evolutionary advantage accrues if progesterone receptors had the same affinity for progesterone as cholinergic receptors for acetylcholine. Unless the efficiency of steroidogenesis were vastly improved, the placenta would need to be considerably larger: to maintain plasma progesterone at the levels required, other things equal, it would need to be the size of a 0.4 m³ (14 cubic foot) refrigerator. Other evolutionary considerations would be 9 months of somnolence that such levels of progesterone would almost certainly produce, difficult to reconcile with the additional 25 000 calories per day required to maintain the requisite levels of progesterone biosynthesis required.

We have mercifully evolved otherwise, and evolution has exploited a range of interactions between signals and receptors in terms of growth, development, homoeostasis, and cognition. Sometimes we can second-guess nature, perhaps to our own disadvantage in terms of realizing our own physiology.

One example, within the author’s area of experience, is that of the mineralocorticoid receptor. Mineralocorticoid hormones were defined in 1961 by Jean Crabbé as promoting unidirectional transepithelial sodium transport (1), a definition that has stood the test of time. The principal mineralocorticoid hormone, aldosterone, is secreted from the zona glomerulosa of the adrenal cortex in response to elevated plasma potassium concentrations, or increased levels of angiotensin II. In response to sodium deficiency, volume depletion or potassium loading, aldosterone incontestably acts via mineralocorticoid receptor in kidney and colon, salivary gland and sweat gland to retain sodium, a la Jean Crabbé and thus acting as a classic homoeostatic hormone. And yet …

When human mineralocorticoid receptors were first cloned (2), the highest levels of mRNA were found in the hippocampus, not a classical site of aldosterone action, and recapitulating earlier binding studies on rat tissue extracts (3). Second, in both studies, mineralocorticoid receptors were shown to have equivalent affinity for the physiological glucocorticoids (cortisol, corticosterone) as for aldosterone, raising obvious questions of how aldosterone ever occupies epithelial mineralocorticoid receptors, given the orders of magnitude higher circulating concentrations of glucocorticoids.

The answer to this question appears to be the coexpression, in epithelial tissues, of the enzyme 11β-hydroxysteroid dehydrogenase (4, 5), which converts cortisol and corticosterone to their inactive 11-keto metabolites cortisone/11-dehydrocorticosterone. Aldosterone is not similarly metabolized, because its signature aldehyde group at C18 cyclizes with the hydroxyl at C11, forming a stable hemiacetal which is not susceptible to enzyme attack by 11β- hydroxysteroid dehydrogenase 2.

The enzyme is expressed at high abundance in aldosterone target cells (3–4 × 10⁶ molecules/cell), and its operation—by metabolizing glucocorticoids (6) and probably by other mechanisms (7, 8)—appears sufficient to confer aldosterone selectivity on the epithelial mineralocorticoid receptor. When it is congenitally deficient, as in the autosomal recessive syndrome of apparent mineralocorticoid excess (9), cortisol activates epithelial mineralocorticoid receptors, leading to uncontrolled sodium retention and severe hypertension.

The enzyme 11β-hydroxysteroid dehydrogenase 2 is not found in nonepithelial tissues in which mineralocorticoid receptors are expressed at high (hippocampus) or modest (heart) abundance, and which thus aldosterone has prima facie little chance of occupying. An inescapable corollary of the last sentence is that such mineralocorticoid receptors are physiologically high-affinity glucocorticoid receptors.

In the syndrome of glucocorticoid remediable aldosteronism (10), aldosterone is secreted primarily in response to adrenocorticotrophic hormone (ACTH), with aldosterone synthase activity expressed throughout the adrenal cortex. The underlying genetic defect is a chimeric gene in which the 5′ end of the gene for 11β-hydroxylase is fused with the 3′ end of the gene coding for aldosterone synthase. This can happen because the two parent genes lie next to one another, on chromosome 8, and because they are 94% identical in terms of nucleotide sequence. What the condition reflects is the product of an unequal crossing over at meiosis in an ancestral gamete, reflecting the relatively small misalignment required (gene proximity) and the possibility of realignment (sequence homology). In evolutionary terms, however, what the condition illustrates is the probability that the two genes (for 11β-hydroxylase and aldosterone synthase) share a relatively recent ancestor, and that their degree of identity and juxtaposition represent a relatively recent gene duplication event. Compare this with the gene coding for the mineralocorticoid receptor (chromosome 4) and the glucocorticoid receptor (chromosome 5).

Mineralocorticoid receptors and glucocorticoid receptors have one area of high (about 90%) sequence identity, the DNA-binding domain, and another of considerable homology, the ligand binding domain, with 57% identity: the majority of the two molecules, including major activation domains, have minimal (less than 15%) identity. It would thus appear that the mineralocorticoid receptor and glucocorticoid receptor are rather more evolutionary distant than are the enzymes 11β-hydroxylase and aldosterone synthase. Although classically mineralocorticoid and glucocorticoid receptors were thought to share a common immediate ‘corticoid’ receptor ancestor (11), more recently evidence has emerged for mineralocorticoid receptors being the first of the mineralocorticoid/glucocorticoid/androgen/progestin receptor subfamily to branch off (12).

In evolutionary terms aldosterone is thus a Johnny-come-lately, pressed into service as organisms became amphibious, to activate a pre-existing high-affinity glucocorticoid receptor (which we now term the mineralocorticoid receptor). Mineralocorticoid receptor selectivity in epithelial aldosterone target tissues is produced by coexpression of the enzyme 11β-hydroxsteroid dehydrogenase 2 at high abundance, and the integrity of a system for Na⁺ retention out of seawater obtained by the expression of aldosterone synthase being yoked to surrogates of Na⁺ deficiency (angiotensin II, K⁺) rather than primarily to the brain hormone ACTH. To call aldosterone the cognate ligand for mineralocorticoid receptor—and the ascription ‘mineralocorticoid receptor’ itself—is thus understandable in terms of our historical knowledge of aldosterone, but it fails to recognize the previous, and current, physiological roles for mineralocorticoid receptors net of aldosterone. The rainbow trout, for instance, does not synthesize aldosterone. In an attempt to clone rainbow trout androgen receptors, an rtMR sequence was identified, related to rtGR but with much higher identity with mammalian mineralocorticoid receptors (13). Its physiologic role(s), like the pathophysiologic roles of mammalian nonepithelial mineralocorticoid receptors, await exploration.

A final fundamental consideration might thus be as follows. There are currently 49 members of the extended steroid/thyroid/retinoid/orphan receptor superfamily of ligand activated transcription factors in the human genome, evidence for enormous evolutionary scope and flexibility. One might thus be pardoned for asking why a ‘specific’ mineralocorticoid receptor did not evolve, responsive to a ligand with levels inversely related to Na⁺ status, rather than the complicated system of highly reactive C18 aldehyde groups and epithelial 11β-hydroxsteroid dehydrogenase 2. This is in fact an impertinent question, bluntly put this way: what is the appropriate question to ask is where is the evolutionary gain in the system being how it is.

For aldosterone and mineralocorticoid receptors, the past decade has provided more questions than answers. Among the latter, for the hormone, is the acceptance that aldosterone can have both genomic and acute, nongenomic effects, and that most but probably not all such rapid effects are via the classic mineralocorticoid receptor. In addition, there is now general consensus that the syndrome of primary aldosteronism represents 10% of all ‘essential hypertension’, and that such patients show higher cardiovascular morbidity and mortality than age-, sex- and blood pressure-matched patients with essential hypertension. For the receptor, the RALES, EPHESUS, and 4E trials (14–16) have shown the beneficial effects of mineralocorticoid receptor blockade in heart failure and essential hypertension. The functions and roles of nonepithelial mineralocorticoid receptors, constitutively (90–99%) occupied by glucocorticoids, have hardly begun to be properly addressed. The mechanisms whereby the physiological glucocorticoids show bivalent activity when bound to mineralocorticoid receptors—normally antagonist, but agonist (in the sense of mimicking aldosterone) in the context of redox change (11β-hydroxysteriod dehydrogenase 2 blockade, reactive oxygen species generation (7, 8) similarly remain to be established.

In fact, the terms agonist and antagonist need to be seen for what they are—effector definitions, like that proposed for mineralocorticoids almost half a century ago by Jean Crabbé. For most hormone receptor systems, the last 20 years—and the past decade in particular—has seen the growing emergence of tissue selective agents, agonist in some organs, antagonist is others. While most microarray analyses have provided a formidable list of genes, expression of which is doubled or halved by a classical agonist, similar lists can be complied for classical antagonists. Some classical antagonists, e.g. spironolactone for epithelial mineralocorticoid receptors, demonstrate inverse agonist activity in experimental myocardial infarction (17). Aldosterone and cortisol aggravate the infarct area; spironolactone at low (EC50 3–5 nm) concentration reduces the infarct area, in the absence of any other steroid.

Spironolactone thus has its ‘antagonist’ effects in the context of cardiac damage not just by competing with agonist steroids for occupancy of mineralocorticoid receptors, but by inducing expression of protective genes and lowering that of proapoptotic genes. It does this at relatively low concentrations, evidence that not all mineralocorticoid receptors, or even a majority, need to be occupied for such an effect. Even before the advent of microarray, it was clear that the effects on enzyme induction, for example, in cultured cells could show distinct dose–response curves, evidence for maximal effects on some readouts at submaximal receptor occupancy. This has not been widely incorporated into consideration of the clinical roles of aldosterone and mineralocorticoid receptors. An example of the former is the demonstration that relatively mild elevations of aldosterone in primary aldosteronism, which would have minimal incremental effects on nonepithelial receptor occupancy, are accompanied by demonstrable cardiovascular damage (18), even in the absence of an elevated blood pressure (19), compared with age-, sex- and blood pressure- matched controls. An example of the latter would appear to be the otherwise curiously low dose (x = 26 mg/day) of spironolactone which, when added to standard care, produced a remarkable 30% improvement in survival, and 35% lower hospital admission rate, in the RALES trial (14).

Given the achievements of the human genome project, we are faced with a mass of information of daunting proportions. This brief chapter has attempted to raise questions, and thus help shape the mindsets of those who face the exciting but very challenging task of reconciling the enormity of information with the demands of clinical endocrinology, from individual patients through populations. For a chance of success, we need a degree of comfort with the underlying mathematics, the biology, and as best we can guess the historical record, the evolution.