Pathology Of Cortisol Metabolism

Much of what we have learnt of the physiology of cortisol metabolism has been inferred from observations of the consequences of pathology. In this section, we review these pathological clinical syndromes and experimental models.

Measurement of Cortisol Metabolism in Humans

The established technique for assessing different routes of cortisol metabolism in humans is to measure the principal metabolites of cortisol in urine, using gas chromatography-mass spectrometry (GCMS) (120,121). Examples of major enzyme defects and associated changes in ratios of metabolites are given below. However, these ratios must be interpreted with caution, since they reflect only relative excretion and not flux or turnover through different metabolic pathways. For example, the ratio of the principal metabolites of cortisol/cortisone, i.e. (5a- + 5^- tetrahydrocor-tisol)/(5^-tetrahydrocortisone), is elevated if 11^-HSD2 is congenitally deficient (66) or inhibited by liquorice (63). However, in theory this could also result from a primary increase in 11^-HSD1 activity. Moreover, dramatic changes in 11^-HSD activity may occur in a tissue-specific manner, but be balanced so that there is no net change in cortisol/ cortisone metabolite ratio. An example is that carbenoxolone inhibits both 11^-HSD1 and 11^-HSD2 and, despite marked cortisol excess in kidney and deficiency in liver, there is no net change in urinary cortisol/cortisone metabolite ratio (64). Nevertheless, all syndromes in which renal 11^-HSD2 is deficient, including carbenoxolone administration, appear to be associated with increased urinary free cortisol/cortisone ratio (122,123), since this reflects specifically intrarenal cortisol and cortisone concentrations without 'adjustment' by 11^-HSD1 elsewhere (Table 18.6) (Figure 18.8).

Primary changes in A-ring reductases are predicted to change the ratio of cortisol/cortisone metabolites too. It was suggested that A-ring reduc-tase activity could be inferred from an 'A-ring re-ductase quotient' (i.e. free urinary cortisol/tetrahyd-rocortisol) (124) but this is not valid if the urinary free cortisol concentration is also altered, e.g. increased because of impaired intrarenal cortisol in-activation by 11^-HSD2 (125). Confusion over the interpretation of urinary metabolite ratios has therefore been an important limitation to the inves-

Figure 18.8 Sensitivity of urinary cortisol metabolite ratios to inhibition of 11P-HSD. Data are from six healthy men treated for 2 days with carbenoxolone followed by 2 days with glycyr-rhetinic acid plus carbenoxolone, or 2 days with glycyrrhetinic acid followed by 2 days with glycyrrhetinic acid plus car-benoxolone. Urinary cortisol metabolites were measured by gas chromatography-mass spectrometry (GCMS) in 24 h urine samples. The ratio of A-ring reduced metabolites of cortisol/ cortisone (i.e. tetrahydrocortisols/tetrahydrocortisone; THFs/ THE) was elevated following glycyrrhetinic acid but not carbenoxolone. This ratio reflects the balance between 11P-HSD1 and 11P-HSD2 activities; glycyrrhetinic acid inhibits only 11 p-HSD2 while carbenoxolone inhibits 11P-HSD1. In contrast, the ratio of urinary cortisol/cortisone is elevated with all manipulations. This ratio reflects intrarenal cortisol/cortisone ratio which is determined by 11P-HSD2 and not 11P-HSD1. Adapted from Best and Walker (123)

tigation of cortisol metabolism in clinical research (126).

Additional methods to assess specific enzymes include dynamic assessments and tissue-specific sampling. 11a3H-Cortisol (Figure 18.3) is cleaved by 11P-dehydrogenase (probably exclusively 11P-HSD2 in humans) to unlabelled cortisone and 3H-H2O, and can therefore distinguish impaired dehydrogenase from enhanced reductase (50,63,64, 127,128). However, the use of radioisotopes in human research is increasingly difficult to justify on ethical grounds. An alternative approach using stable isotope labelling with deuterium has been reported (129) and shows promise. Alternatively, to measure 11P-HSD1, cortisone can be administered orally and the rate of cortisol appearance in peripheral plasma taken to indicate first pass hepatic metabolism by 11P-HSD1 (Figure 18.4) (64,93,130,131). However, this is potentially con founded by the rate of clearance of cortisol from peripheral plasma, and frequent sampling has not always been employed to allow estimates of rate of appearance. A simple estimate of peak plasma cortisol obtained may not be such a specific index of 11P-HSD1.

Tissue-specific sampling requires measurements in venous effluent from the site of interest or from fluid in equilibrium with interstitial fluid at that site. This has been achieved with difficulty for human liver (26), lung (56), and abdominal subcutaneous fat (57). However, the high variability of these measurements may obscure important differences. These techniques would be enhanced by including tracer cortisol or cortisone and examining rates of conversion rather than just cortisol/cortisone ratios.

With these limitations in mind, it has still been possible to define several syndromes on the basis of biochemical measurements, several of which have since been explained by molecular defects.

Congenital Syndromes in Humans

11 fi-HSD2 Deficiency: 'Apparent Mineralocorticoid Excess Syndrome'

This autosomal recessive disorder is rare, with fewer than 50 cases reported worldwide since 1974 (132,133). It usually presents in childhood with hypokalemia (polyuria, myopathy), severe hypertension, and complications including stroke and cardiac arrest. Investigation reveals low plasma renin activity, low plasma aldosterone (and low levels of other mineralocorticoids including ^deoxycorticosterone), and hypokalaemic alkalosis which improves with mineralocorticoid receptor antagonists (spironolactone) or post-receptor sodium channel antagonists (amiloride). In the absence of a measurable mineralocorticoid in serum of these patients the cause remained obscure for some years, hence the term 'apparent' mineralocorticoid excess. It was then recognized that, although plasma cortisol concentrations are normal, the condition improves following suppression of cortisol with dexamethasone, while additional administration of cortisol induces profound mineralocorticoid excess (134,135). Moreover, analysis of urinary metabolites of cortisol reveals an elevated ratio of the metabolites of cortisol versus those of cortisone, and the half-life of 11a3H-

Cortisol (Figure 18.3) is dramatically prolonged, consistent with 11^-dehydrogenase deficiency. Total Cortisol metabolite excretion is reduced (consistent with a compensatory fall in cortisol production rate) but urinary free cortisol concentrations are elevated (consistent with impaired intrarenal cortisol metabolism).

Following cloning of human 11^-HSD2 in 1994, a series of mutations in the 11^-HSD2 gene have been reported. Patients with the syndrome of apparent mineralocorticoid excess are more commonly homozygotes than compound heterozygotes for such mutations (62), suggesting that there is a low prevalence of heterozygous inactivating mutations in the population and most cases arise from consanguinity. Heterozygotes in these families may have more subtle abnormalities of cortisol metabolism and milder hypertension (135,136), but their blood pressure has been easily controlled by conventional means, indicating that there is considerable redundancy or excess capacity in 11^-HSD2 activity in human kidney. Intriguingly, polymorphisms of a microsatellite in the 11 ^-HSD2 gene on chromosome 16p have recently been associated with variations in 11^-HSD activity and with the individual blood pressure response to salt loading

An additional feature of this syndrome has emerged quite recently. 11^-HSD2 is present in placenta (10,11), where it is thought to protect the fetus from inappropriate exposure to maternal cortisol

(138). Excessive glucocorticoid exposure of the fetus results in growth retardation. In retrospective studies, birthweight has been found to be low in patients with 11^-HSD2 deficiency (133,139). In some (140), but not all (14,141), studies there is an inverse relationship between placental 11^-HSD activity and birthweight in otherwise healthy individuals. Extensive epidemiological studies have linked low birthweight with subsequent adverse risk factors for cardiovascular disease, including hypertension and insulin resistance (142). Prenatal treatment of rats with either dexamethasone, which is a poor 11^-HSD2 substrate and so directly accesses the fetus, or carbenoxolone which inhibits 11^-HSD, lowers birthweight and programs permanent hypertension and hyperglycemia in the adult offspring (143-145). However, while the hypertension and insulin resistance in low birthweight babies may be amplified by subsequent obesity (146), current data do not suggest that obesity itselfis programmed by these early determinants of metabolic and cardiovascular development.

11fi-HSD1 Deficiency: 'Apparent Cortisone Reductase Deficiency'

This syndrome has been reported in just four patients (92,93,147,148). All are female and presented with hirsutism and menstrual irregularity. Investigation reveals mild ACTH-dependent adrenal androgen excess. Urinary cortisol metabolites are strikingly abnormal, with markedly elevated cortisone metabolite excretion and elevated total cortisol metabolite excretion. Urinary free cortisol/corti-sone ratio is normal. The conversion of cortisone, administered orally, into cortisol in peripheral plasma was markedly impaired in one affected patient, and unlike healthy volunteers a peak of cortisone was also detected in peripheral plasma (Figure 18.4).

Although all of these features are consistent with 11^-HSD1 deficiency, no mutation in the 11^-HSD1 gene exons or exon/intron boundaries has yet been reported (93,149). It remains possible that an alternative abnormality of cortisol metabolism, such as 5^-reductase deficiency, might be responsible for shifting the excretion of cortisol metabolites in favour of cortisone; hence the term 'apparent' cortisone reductase deficiency.

Given the proposed role of 11^-HSD1 to maintain local cortisol concentrations in adipose tissue (4), one might expect these patients to be thin with lipoatrophy. In fact, this appears not to occur; indeed, at least one of these patients is obese.

Combined 11P-HSD1 and11fi-HSD2 Deficiency

A small series of patients with a so-called syndrome of apparent mineralocorticoid excess 'type 2' have been described (128). In these four patients, cortisol-dependent mineralocorticoid excess has been documented, and impaired 11^-dehydrogenase activity has been confirmed by measurement of 11a3H-cortisol half-life and urinary free cortisol/ cortisone ratio. However, overall cortisol/cortisone metabolite ratios are not deranged and the conversion of oral cortisone into cortisol in peripheral plasma is impaired. This suggests that there is a combination of 11£-HSD1 and 11£-HSD2 deficiency, as occurs after carbenoxolone administra-

CORTISOL METABOLISM Table 18.7 Features of transgenic deletion of 11 ^-HSDs

Lower blood glucose after overfeeding with obesity Lower blood glucose after stress

Impaired activation of gluconeogenesis (PEPCK) on fasting


Altered neonatal lung maturation

Severe hypertension Hypokalaemia

Renal structural abnormalities Increased risk of early postnatal death tion (see below). The molecular defect remains to be elucidated. A non-coding mutation in 11^-HSD2 has been preliminarily reported in these patients (150), but whether this accounts for the syndrome remains unclear since it does not explain their impaired 11^-HSD1 activity.

Interestingly, these patients all come from the Mediterranean island of Sardinia, where there is an association between hypertension and elevated ratios of cortisol/cortisone metabolites (151), and a high prevalence of the polymorphisms in the 11^-HSD2 gene which have been associated with saltsensitive hypertension (137).

A-Ring Reductase Deficiencies

The congenital syndrome of 5a-reductase type 2 deficiency results in impaired generation of dihyd-rotestosterone in key androgen target tissues (44). While testosterone levels are relatively low before puberty, this results in feminization of males including ambiguous genitalia. However, after puberty higher circulating testosterone levels overcome the defect and males develop full secondary sexual characteristics, so that males reared as girls may be re-assigned as men. The consequences of this enzyme defect on cortisol metabolism have, however, not been studied.

Congenital syndromes of 5a-reductase type 1 or 5^-reductase deficiency have not been described.

Interactions Between A-ring Reductases and 11^-HSDs

A paradox in syndromes of 11^-HSD deficiency is that there are also changes in the relative excretion of 5a-tetrahydrocortisol and 5^-tetrahydrocortisol. These remain poorly understood. Following liquorice or carbenoxolone administration, the relative excretion of 5^-tetrahydrocortisol falls (63,64), which could be explained by inhibition of 5^-reduc-tase by these compounds (152). However, the same relative fall in 5^-reduced cortisol metabolites occurs in congenital 11^-HSD2 deficiency. No obvious artefact of measurement (see above) can explain this. It suggests that there is a more complex interplay between the different pathways of cortisol metabolism than we currently understand.

Animal Models

Transgenic knockout mice have been generated for several of these key glucocorticoid metabolizing enzymes.

11P-HSD2 Knockout Mice

These animals faithfully reproduce the phenotype of the syndrome of apparent mineralocorticoid excess (AME) in humans, with hypertension and hypokalaemia which can be ameliorated by dex-amethasone or spironolactone treatment and exacerbated by physiological doses of corticos-terone (Table 18.7) (65). Interestingly, they also exhibit structural abnormalities in kidney from infancy (distal tubular hyperplasia and hypertrophy) which do not appear to reverse with spironolactone, suggesting a possible basis for the difficulties in treating established hypertension in AME merely with cortisol suppression or anti-mineralocor-ticoids.

11P-HSD1 Knockout Mice

The generation of these mice has confirmed the importance of 11^-HSD1 in modulating glucocorticoid action in several target tissues (69). They show a complete inability to convert 11-dehyd-rocorticosterone (the equivalent of cortisone in humans) to corticosterone (equivalent to cortisol), indicating that 11^-HSD1 is the only major 11^-reductase enzyme, at least in mice. This defect af fects multiple tissues, and the following evidence has been obtained that it results in attenuated glucocorticoid action in liver, brain and possibly peripheral fat.

In the liver, PEPCK, the rate-limiting enzyme in gluconeogenesis which is downregulated by insulin and upregulated by glucocorticoids, shows an impaired induction in response to fasting in 11^-HSD — / — animals (Figure 18.6c). This is consistent with the interpretation of pharmacological experiments (see above) that 11^-HSD1 maintains in-trahepatic glucocorticoid receptor activation.

When 11^-HSD1 — / — animals are fed a high-fat cafeteria diet, they gain less weight and have lower plasma glucose levels than wild-type controls. The latter is consistent with effects on the hepatic gluconeogenic response to insulin, but the former may be explained by additional central effects of 11^-HSD1 on regulation of energy expenditure and/or an influence of 11^-HSD1 on adipocyte metabolism. Appetite and energy expenditure have yet to be examined. Interestingly, however, body fat distribution in animals on normal chow is not obviously different from wild type.

11^-HSD1 — / — mice are also hyper-corticos-teronaemic, consistent with a defect in negative feedback regulation of the hypothalamic-pituitary-adrenal axis.

5a-Reductase Knockout Mice

Mice with transgenic deletion of both isozymes of 5a-reductase have been reported (86,153). The 5a-reductase type 2 knockout, surprisingly, has no abnormality of prostate or genital development but in rodents 5a-reductase type 1 is co-expressed and could compensate. The 5a-reductase type 1 knockout also has no major phenotypic abnormality, except an increased risk of in utero death attributed to diversion of androgens to oestrogens instead of 5a-dihydrotestosterone (86). However, glucocorticoid metabolism and action has not been reported in these animals.

Pharmacological Manipulation

Liquorice and its Derivatives

Extracts from the liquorice root have been used as confectionery and in therapy for dyspepsia for dec ades (154). Some people habitually consume excessive quantities of liquorice, although whether there is a liquorice withdrawal syndrome—suggesting that the material is truly 'addictive'—has not been established. The active constituents of liquorice include glycyrrhizic acid and its metabolite glycyrrhetinic acid (Figure 18.9). A related hemisuccinate, car-benoxolone, remains a licensed medication for dyspepsia in UK.

A long-recognized side effect of liquorice ingestion is sodium retention, leading to hypertension and even heart failure. This is accompanied by hy-pokalaemia and suppression of plasma renin activity and aldosterone and was at one time attributed to direct activation of mineralocorticoid receptors by liquorice. However, these side effects are dependent upon the presence of cortisol. Stewart and Edwards showed that liquorice derivatives, particularly glycyrrhetinic acid, inhibit 11^-HSDs and enhance binding of glucocorticoids to mineralocor-ticoid receptors (24,61,63,155). Liquorice administration therefore reproduces the syndrome of apparent mineralocorticoid excess, including elevated ratios of cortisol/cortisone and of their metabolites in urine and prolonged half-life of 11a3H-cortisol (Table 18.6).

Further studies showed that carbenoxolone also induces cortisol-dependent mineralocorticoid excess and inhibits 11^-HSD in vitro (64). However, there were several differences between the effects of carbenoxolone and liquorice/glycyrrhetinic acid in humans (Table 18.6). While carbenoxolone increases urinary free cortisol/cortisone and prolongs the half-life of 11a3H-cortisol, it does not alter ratios of A-ring reduced urinary metabolites of cortisol versus cortisone (Figure 18.8). This paradox is explained by the observation that carbenoxolone, but not liquorice, inhibits the conversion of cortisone to cortisol by 11$-HSD1 in liver (64,156). Thus, inhibition of 11^-HSD2 in kidney is balanced by inhibition of 11^-HSD1 in liver and there is no change in overall equilibrium between cortisol and cortisone despite marked changes in intrarenal and intrahepatic cortisol concentrations (125). This is an important observation to bear in mind when interpreting urinary cortisol metabolite results in other clinical syndromes (see above). As a result, car-benoxolone produces additional clinical effects, including enhancing insulin sensitivity (Figure 18.4) (67). The reasons for this discrepancy are not clear, since carbenoxolone and glycyrrhetinic acid have

Glycyrrhetinic acid Carbenoxolone

Figure 18.9 11^-HSD inhibitors. Glycyrrhetinic acid is the principal active constituent of liquorice. Carbenoxolone is a synthetic hemisuccinate derivative

Glycyrrhetinic acid Carbenoxolone

Figure 18.9 11^-HSD inhibitors. Glycyrrhetinic acid is the principal active constituent of liquorice. Carbenoxolone is a synthetic hemisuccinate derivative

Figure 18.10 A-ring reductase activity in human obesity. The ratio 5^-/5a-tetrahydrocortisol reflects relative activities of 5a-and 5^-reductases. This study of 68 healthy men (filled symbols) and postmenopausal women (open symbols; triangles indicate those receiving oestrogen replacement therapy) aged 47-53 years from a cross-sectional cohort study revealed higher excretion of 5a-tetrahydrocortisol in subjects with central obesity. Adapted from Andrew et al. (163)

Figure 18.10 A-ring reductase activity in human obesity. The ratio 5^-/5a-tetrahydrocortisol reflects relative activities of 5a-and 5^-reductases. This study of 68 healthy men (filled symbols) and postmenopausal women (open symbols; triangles indicate those receiving oestrogen replacement therapy) aged 47-53 years from a cross-sectional cohort study revealed higher excretion of 5a-tetrahydrocortisol in subjects with central obesity. Adapted from Andrew et al. (163)

similar affinities for both 11^-HSD isozymes in vitro. It probably relates to pharmacokinetic factors, perhaps because carbenoxolone is water soluble whereas glycyrrhetinic acid is lipid soluble and hydrophobic.

The effects of liquorice derivatives are not selective for 11^-HSD enzymes. These compounds also inhibit enzymes metabolizing prostaglandins (157). In addition, they inhibit 5^-reductase in vitro (see below) (152).

5a-Reductase Inhibitors

Arguably the most successful example of therapeutic manipulation of pre-receptor ligand metabolism to date is the use of finasteride to inhibit

5a-reductase type 2. This is used in the treatment of benign prostatic hyperplasia and in prostatic carcinoma. It has a weak effect to reduce androgen receptor activation in skin and may be useful in the treatment of hirsutism and male-pattern baldness. Finasteride is relatively, but not completely, specific for the human type 2 5a-reductase isozyme. It has also been shown to alter cortisol metabolism in one small study (158), although the impact on glucocorticoid receptor activation was not assessed.

Selective 5a-reductase type 1 inhibitors exist (159), but have not been employed therapeutically.

Delicious Diabetic Recipes

Delicious Diabetic Recipes

This brilliant guide will teach you how to cook all those delicious recipes for people who have diabetes.

Get My Free Ebook

Post a comment