The Adrenal Cortex and Medulla

Mineralocorticoid receptors bind cortisol as tightly as they bind aldosterone, and adult human plasma carries far more cortisol (a hundredfold molar excess is a commonly cited figure). Yet sodium transport in the kidney follows aldosterone. How?

The answer is an enzyme beside the receptor. On the way you will follow one gland’s steroids and catecholamines, then use their controls to predict disease. The first look is in the endocrine system.

Two glands in one capsule: three cortical zones and a medulla

Each adrenal gland, on the upper pole of a kidney, is two organs in one capsule. The outer cortex makes steroids from cholesterol. The inner medulla is a modified sympathetic ganglion whose chromaffin cells, which arise from the neural crest, release adrenaline (epinephrine) and noradrenaline (norepinephrine) into the blood.

Zonation of the adult human cortex is a matter of which enzymes each zone expresses: one assembly line, with stations switched on or off in three workshops. In every zone, the mitochondrial side-chain cleavage enzyme (CYP11A1) cuts cholesterol to pregnenolone. This is the slow step, and ACTH speeds it, within minutes by raising cholesterol delivery to the enzyme (through the StAR protein) and over days by inducing the pathway’s enzymes. Then the routes split.

• Zona glomerulosa (outer): makes aldosterone, the main mineralocorticoid (a steroid acting on sodium and potassium balance). Pregnenolone becomes progesterone, 21-hydroxylase makes 11-deoxycorticosterone (DOC), and aldosterone synthase (CYP11B2), expressed only here, converts DOC through corticosterone to aldosterone. No 17-alpha-hydroxylase, so no cortisol.
• Zona fasciculata (middle): makes cortisol, the main glucocorticoid (named for its effect on blood glucose). 17-alpha-hydroxylase (CYP17A1) gives 17-hydroxyprogesterone, 21-hydroxylase gives 11-deoxycortisol, and 11-beta-hydroxylase (CYP11B1) gives cortisol. No aldosterone synthase.
• Zona reticularis (inner): CYP17A1 acts strongly as a 17,20-lyase, giving DHEA, then DHEA sulfate and androstenedione, weak androgens.

So 21-hydroxylase lies on the routes to cortisol and aldosterone but not to the androgens. Species differ: rats and mice express little adrenal 17-alpha-hydroxylase and make corticosterone, with little adrenal DHEA, so the main glucocorticoid a rat or mouse makes is corticosterone, not cortisol.

Preganglionic sympathetic fibres release acetylcholine onto nicotinic receptors of chromaffin cells, which depolarize and release catecholamines within seconds (hypothalamus and autonomic control). In adult humans most of the output, often given as about four-fifths, is adrenaline. Blood drains from cortex to medulla, and work mainly in rats indicates that this glucocorticoid-rich blood supports PNMT, the enzyme that converts noradrenaline to adrenaline.

ACTH drives cortisol; angiotensin II and potassium drive aldosterone

ACTH binds the MC2 receptor (melanocortin 2 receptor), found mainly on fasciculata and reticularis cells but also in the glomerulosa, a G-protein-coupled receptor that signals through Gs and cyclic AMP (cAMP and PKA). The feedback loop and its tests belong to endocrine axes and testing and emotion and stress.

Aldosterone answers to angiotensin II, the end product of the renin cascade (blood pressure regulation), and to a rise in plasma potassium. ACTH stimulates it too, but less, and with continuous exposure that effect fades, which is why loss of pituitary ACTH leaves aldosterone largely intact.

The glomerulosa cell is a voltage sensor. Potassium channels dominate its resting conductance, so it rests near the potassium equilibrium potential, and a rise in plasma potassium depolarizes it (membrane transport). Angiotensin II, through its AT1 receptor, closes potassium channels with the same result. Depolarization opens voltage-gated calcium channels, and calcium drives aldosterone synthase. In mouse adrenal slices, Hu and colleagues (2012) reported spontaneous voltage oscillations in 73 of 85 glomerulosa cells, at a mean of 0.44 Hz at 22 degrees C: an electrical rhythm in mouse cells, not a human hormone pulse rate.

In adult human plasma about 90 percent of cortisol is protein-bound, mostly to cortisol-binding globulin (CBG). On the free hormone hypothesis, only the unbound fraction acts. Aldosterone binds CBG weakly, so more of it circulates free. By stable-isotope dilution, Esteban and colleagues (1991) measured a mean cortisol production of 5.7 mg per square metre of body surface per day in 12 normal volunteers, lower than earlier estimates by the authors’ account.

Cortisol shifts fuel, restrains inflammation and supports pressure

Dinneen and colleagues (1993) blocked cortisol synthesis in normal volunteers with metyrapone, an 11-beta-hydroxylase inhibitor, and replaced cortisol by infusion either at a constant basal rate or on a schedule mimicking the night-time rise. After a mixed meal, the variable schedule increased the liver’s glucose release, reduced glucose uptake in a way that appeared to reflect insulin resistance, and raised palmitate appearance, suggesting lipolysis. Insulin and glucagon changed too, so not every effect is direct. Simmons and colleagues (1984) found that an eight-hour hydrocortisone infusion increased whole-body proteolysis in seven normal subjects. Those amino acids feed gluconeogenesis (integration of metabolism). Cortisol thus raises glucose supply and lowers its use by muscle and fat, opposing insulin (insulin, glucagon and glucose).

On inflammation, cortisol acts through the glucocorticoid receptor, a ligand-activated transcription factor (nuclear receptors). In mice and cultured cells, Auphan and colleagues (1995) found that glucocorticoids inhibited activation of NF-kappa-B, which switches on many inflammatory genes (TLRs and NF-kappa-B), and that inducing its inhibitor, I-kappa-B-alpha, could be a major part of that; Scheinman and colleagues reported a similar mechanism the same year. Later work found that this induction varies between cell types, and that the receptor also represses inflammatory genes in other ways, for example by binding NF-kappa-B and AP-1 proteins directly.

On pressure, Whitworth and colleagues (1984) infused cortisol into seven subjects for five days at rates meant to resemble ACTH stimulation, and systolic pressure rose from 108 to 129 mmHg, by a mechanism the authors left unresolved. Part of cortisol’s support of pressure is thought to be permissive: it keeps vascular smooth muscle responsive to noradrenaline and angiotensin II. Very high cortisol, as in severe Cushing syndrome, can also overwhelm the kidney’s guard enzyme (below) and act on the mineralocorticoid receptor itself, retaining sodium and lowering potassium.

Aldosterone acts on the epithelium itself, and an enzyme guards its receptor

By the late 1950s aldosterone was known to make mammals retain sodium. Did it act on sodium-transporting cells directly?

Jean Crabbé (1961) used the urinary bladder of the toad Bufo marinus, an epithelium that moves sodium from its urine side to its blood (serosal) side. He cut each bladder into paired halves and mounted each as a sheet between two chambers of aerated frog Ringer solution at room temperature. He measured the short-circuit current, the current needed to hold the voltage across the sheet at zero, which equals net active ion transport when one ion carries it. One half received d-aldosterone at 2.8 x 10^-7 mol per L on the serosal side; its partner was the control. If the effect required another hormone or blood flow, the sheet should not respond. After a latent period of about one hour, the treated half’s current rose above its control’s, and in eight bladders bidirectional sodium tracer fluxes confirmed that the extra current was net sodium transport.

That ruled in a direct, delayed action on the epithelium and ruled out, in this preparation, a systemic intermediary; pairing removed differences between animals. It did not show which mammalian nephron segment responds (the distal nephron and collecting duct, from later work: tubular transport and diuretics), and the delay suggested, without proving, new protein synthesis, which later toad-bladder work, including from Isidore Edelman’s laboratory, supported.

Worked problem, part 1. These numbers are invented, modelled on Crabbé’s design; he did not report them. Two paired halves each expose 1.00 cm2. Once the response has developed, the aldosterone half carries 100 microamperes and its control 40 microamperes. Assume sodium (charge +1) carries all the net current.

sodium flux (mol per s) = current (C per s) / F, with F = 96,485 C per mol

1. Extra current = 100 - 40 = 60 microamperes = 60 x 10^-6 C per s.
2. Extra flux = (60 x 10^-6 C per s) / (96,485 C per mol) = 6.22 x 10^-10 mol per s. C per s divided by C per mol gives mol per s.
3. Per hour: 6.22 x 10^-10 x 3,600 s = 2.24 x 10^-6 mol, so 2.24 micromol per cm2 per h.
4. Check: the control alone carries 40 x 10^-6 / 96,485 x 3,600 = 1.49 micromol per cm2 per h and the treated half 3.73; the difference is 2.24, a 2.5-fold rise.

If another ion carried part of the current, step 2 would overstate sodium transport; hence the tracers.

Part 2: compare the concentration with plasma. 2.8 x 10^-7 mol per L = 280 nmol per L. Take an illustrative adult plasma aldosterone of 15 ng per dL (made up; real ranges differ by laboratory, posture and salt intake), which is 150 ng per L. With a molar mass of 360.4 g per mol, 150 x 10^-9 / 360.4 = 4.16 x 10^-10 mol per L, or 0.416 nmol per L. The ratio 280 / 0.416 is about 670. The classic shows that the epithelium can respond, not that it responds at physiological concentrations.

Now the opening puzzle. Funder, Pearce, Smith and Smith (1988) reported that mineralocorticoid receptors bind aldosterone and the physiological glucocorticoids with equal affinity, and that 11-beta-hydroxysteroid dehydrogenase (11-beta-HSD) in targets such as kidney, colon and parotid protects them by converting cortisol to inactive cortisone (corticosterone to 11-dehydrocorticosterone in rats). Aldosterone escapes this conversion. Edwards and colleagues reached a similar conclusion that year; the epithelial type 2 isoform was characterized in the 1990s. What the receptor then does is taught in water, sodium and potassium.

Reading adrenal disease from the loops

Primary aldosteronism tests the renin-angiotensin-aldosterone loop: an adenoma or bilateral hyperplasia makes aldosterone without regard to renin. Sodium reabsorption rises and extracellular volume expands only modestly, because higher pressure raises sodium excretion (pressure natriuresis) until output matches intake, often called mineralocorticoid escape; edema is not typical. Pressure rises; potassium falls in some patients, but many are normokalemic; bicarbonate may rise (acid-base balance); renin falls. So aldosterone is read against renin. In a recruited cross-sectional cohort, Brown and colleagues (2020) found renin-independent aldosterone production forming a continuum across blood-pressure categories, and a 2025 Endocrine Society guideline suggests screening everyone with hypertension.

The modern experiment. Murim Choi, Ute Scholl and colleagues in Richard Lifton’s laboratory (2011) exome-sequenced four human aldosterone-producing adenomas against the same patients’ blood, then sequenced KCNJ5 in 18 more tumor and blood pairs. Recurrent mutations in KCNJ5, a potassium channel gene, appeared in 8 of 22 adenomas and were somatic: in the tumor, not the blood. Expressed with KCNJ3 in cultured 293T cells, the mutant channels conducted sodium and depolarized the cells. A depolarized glomerulosa cell opens calcium channels and keeps making aldosterone even when renin and angiotensin II are suppressed. Keep apart association, somatic status and function. It does not show that all primary aldosteronism has this cause (most of these adenomas lacked it), that a cell line behaves like a glomerulosa cell, or why the cells proliferate.

The other disorders follow from the same loops (directions relative to a healthy adult; cases vary).

Disorder	Cortisol	ACTH	Aldosterone	Renin	Plasma K+	Adrenal androgens	Blood pressure
Primary insufficiency (Addison)	Down	Up	Down	Up	Up	Down	Down
Central insufficiency	Down	Low or normal	Near normal	Normal	Normal	Down	Can fall
Cushing, adrenal adenoma	Up	Down	Normal	Normal or down	Normal or down	Down	Up
Cushing, pituitary or ectopic ACTH	Up	Up or normal	Normal or down	Normal or down	Normal or down	Up or high-normal	Up
CAH, 21-hydroxylase, salt-wasting	Down	Up	Down	Up	Up	Up (androstenedione; DHEA and DHEAS can be low)	Down
CAH, 11-beta-hydroxylase	Down	Up	Down	Down	Normal or down	Up	Up in many

In primary insufficiency high ACTH and related peptides from its precursor (POMC) darken skin and mucosa; in central insufficiency angiotensin II and potassium still drive the glomerulosa, so neither hyperkalemia nor hyperpigmentation appears. Cushing disease is only the pituitary form of Cushing syndrome. In congenital adrenal hyperplasia (CAH) from 21-hydroxylase deficiency, 17-hydroxyprogesterone piles up and is diverted into androgens, virilizing affected children (sexual differentiation). In 11-beta-hydroxylase deficiency DOC acts as a mineralocorticoid; 38 of 70 patients with blood-pressure data in one international cohort were hypertensive (Khattab and colleagues, 2017).

up down unchanged ? unsettled all compared with a healthy adult

ACTHunchanged

reninunchanged

mineralocorticoid activity (aldosterone plus DOC)unchanged

plasma potassiumunchanged

blood pressure (through the mineralocorticoid route)unchanged

Block one enzyme

none blocked block 21-hydroxylase block 11-beta-hydroxylase

The loop

ACTH feedback switch on (off: ACTH ignores cortisol) cortisol supplied from outside

In this sketch: arrows compare each item with a healthy adult and carry no numbers; one enzyme is blocked at a time; a block lowers its product and everything after it and piles up only its own substrate; ACTH raises or lowers the boxes from pregnenolone to the androgens but not DOC, corticosterone or aldosterone; with the feedback switch off, ACTH ignores cortisol (a switch in the sketch, not a disease or lesion); supplied cortisol with feedback intact lowers ACTH; and blood pressure is shown through the mineralocorticoid route only (cortisol itself also supports pressure, which is left out). These rules are made up to sketch the lesson’s logic. They are not measurements, and cases vary.

A block lowers what comes after it and piles up what comes before it. If cortisol falls with feedback intact, ACTH rises and drives the cortisol and androgen branches harder, and mineralocorticoid activity (aldosterone plus DOC) sets the direction of renin, potassium and pressure.

A recent trial probed this loop from above with crinecerfont, a CRF1 (CRH receptor) antagonist. In 182 randomized adults with classic 21-hydroxylase deficiency, it allowed a 27.3 percent least-squares mean glucocorticoid dose reduction at 24 weeks with androstenedione controlled, against 10.3 percent with placebo (Auchus and colleagues, 2024), consistent with excess CRH-ACTH drive. Replacement was still required.

Design problem

Mineralocorticoid receptors extracted from rat kidney bind aldosterone and corticosterone with about equal affinity, and free corticosterone in rat plasma far exceeds free aldosterone. Yet aldosterone controls sodium transport in the kidney. Design experiments to find out how the kidney’s receptors are kept for aldosterone, and say what each result would mean.

One way to do it

1. Confirm equal binding in vitro, with labelled steroids and receptors from targets (kidney, colon, parotid) and from hippocampus, a non-epithelial comparison. Equal affinity means selectivity is not in the receptor.

2. Ask what binds in vivo. Give labelled corticosterone to adrenalectomized rats; measure what is bound to kidney receptors and what the steroid has become. Most label as 11-dehydrocorticosterone points to local inactivation.

3. Block the enzyme (11-beta-HSD) with carbenoxolone or glycyrrhetinic acid from liquorice. Predict that corticosterone now occupies kidney receptors and, in the intact animal, that sodium is retained and potassium excreted.

4. Controls: vehicle; the inhibitor alone without glucocorticoid, to exclude a direct action on the receptor; and a glucocorticoid the enzyme cannot inactivate.

Funder and colleagues (1988) concluded that specificity is enzyme-mediated, not receptor-mediated. Limits: rat and corticosterone; the type 2 isoform came later; and saturation of the enzyme by very high cortisol was shown only in later work.