| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Endocrinology Unit, School of Molecular and Clinical Medicine, University of Edinburgh, Western General Hospital, Edinburgh EH4 2XU, Scotland, United Kingdom
 |
ABSTRACT |
|---|
 TOP ABSTRACT I. PreambleII. Blood Glucocorticoid Levels in...III. Corticosteroid Receptors and...IV. Tissue Metabolism of GlucocorticoidsV. 11ß-HSD1 and Glucocorticoid...VI. 11ß-HSD1 as a Drug Target...VII. SummaryACKNOWLEDGEMENTSREFERENCES |
|---|
"I am resolved to grow fat and look young till forty, and then slip out of the world with the first wrinkle." — John Dryden, The Maiden Queen, 1668
|
I. Preamble |
|---|
|
TOPABSTRACTI. PreambleII. Blood Glucocorticoid Levels in...III. Corticosteroid Receptors and...IV. Tissue Metabolism of GlucocorticoidsV. 11ß-HSD1 and Glucocorticoid...VI. 11ß-HSD1 as a Drug Target...VII. SummaryACKNOWLEDGEMENTSREFERENCES |
|---|
Endocrinologists have long puzzled over the strong morphological and metabolic similarities between the rare Cushing’s syndrome of endogenous or exogenous glucocorticoid excess and the metabolic syndrome (Walker and Seckl, 2001). Excessive cortisol secretion in Cushing’s syndrome is a classic cause of secondary obesity. Effects of excess cortisol on adipose tissue are complex, with an increase in central (i.e., visceral, abdominal, facial, and nape of neck) fat deposition, while peripheral fat is reduced. This may result from opposing effects of glucocorticoids that, on the one hand, increase lipolysis and downregulate lipoprotein lipase — thereby liberating free fatty acids (FFAs) from peripheral fat — but, on the other hand, stimulate preadipocyte differentiation and enhance substrate flux in favour of gluconeogenesis and triglyceride synthesis in central fat (Samra et al., 1998; Andrews and Walker, 1999). Additionally important in glucocorticoid-induced obesity is central stimulation of appetite (Udden et al., 2003), mediated by complex interactions between hypothalamic responses to glucocorticoids, their effects upon target appetite-controlling neurotransmitters such as neuropeptide Y (NPY), and both central and peripheral effects upon adipose-derived endocrine mediators such as leptin and gut hormones (Solano and Jacobson, 1999). Other central effects of increased glucocorticoid action include depression (Cohen, 1980), which intriguingly is also linked specifically to idiopathic visceral obesity (Ahlberg et al., 2002).
Raised plasma cortisol levels cause phenotypic changes in Cushing’s syndrome. Against this background, the importance of glucocorticoids in obesity has been investigated in animal models and in humans. Many studies have focused on glucocorticoid secretion and circulating blood levels of steroids. This research has shown that cortisol levels are modestly, if at all, elevated in patients with the metabolic syndrome and even may be reduced in simple obesity (i.e., obesity without diabetes or other disease states complicating increased fat mass) (Walker and Seckl, 2001). This conundrum recently was readdressed and some progress made in disentangling the mechanisms and importance of altered tissue sensitivity to glucocorticoids. This, then, is a story of cellular corticosteroid receptors, and particularly of prereceptor metabolism by enzymes, notably by the hitherto-obscure 11beta-hydroxysteroid dehydrogenases (11ß-HSDs).
|
II. Blood Glucocorticoid Levels in Obesity |
|---|
|
TOPABSTRACTI. PreambleII. Blood Glucocorticoid Levels in...III. Corticosteroid Receptors and...IV. Tissue Metabolism of GlucocorticoidsV. 11ß-HSD1 and Glucocorticoid...VI. 11ß-HSD1 as a Drug Target...VII. SummaryACKNOWLEDGEMENTSREFERENCES |
|---|
B. THE HPA AXIS AND CORTISOL DYNAMICS IN OBESITY
A number of case-control studies have suggested that obesity is associated with increased urinary free cortisol (UFC) excretion (Strain et al., 1980; Marin et al., 1992; Pasquali et al., 1993a). This association is particularly prominent in subjects with abdominal obesity, at least in women (Pasquali et al., 1993b). A caveat is that UFC, though a useful marker of HPA axis activation, forms a very small fraction of total cortisol metabolite excretion. More convincingly, recent large studies confirm that total cortisol production rate is somewhat enhanced in obesity in men as well as women (Andrew et al., 1998; Fraser et al., 1999; Stewart et al., 1999). This is further supported by evidence that HPA axis responsiveness to stimuli such as acute stress, CRH/AVP (Pasquali et al., 1999), adrenergic manipulations (Pasquali et al., 2000), hypoglycaemia, or a standard meal (Marin et al., 1992; Hautanen and Adlercreutz, 1993) is enhanced in idiopathic obesity, particularly visceral obesity (Pasquali et al., 2000). Moreover, the cortisol response to particular types of food may differ with the locus of adiposity, as viscerally obese women show lower HPA axis responses to high-fat and protein but greater responses to high-carbohydrate meals, compared to peripherally obese and lean subjects (Vicennati et al., 2002). However, in obesity, overall plasma cortisol levels are not consistently elevated. Indeed, peak plasma cortisol levels in the morning are typically low in the obese (Ljung et al., 1996; Phillips et al., 1998; Rosmond et al., 1998; Walker et al., 1999), though South Asians may be an exception (Ward et al., 2003). While some data in humans suggest that plasma cortisol is higher during the evening nadir in obesity, flattening the diurnal variation (Ljung et al., 1996; Rosmond et al., 1998), evidence is lacking for marked elevations of basal plasma cortisol in obesity. The combination of increased secretion with low morning plasma levels suggests that the diurnal variation of cortisol secretion is disrupted and/or that peripheral metabolism of cortisol is enhanced.
As explanation for the increased glucocorticoid production and exaggerated cortisol responses in obesity, both increased forward drive to the HPA axis — perhaps reflecting some form of primary meal-related abnormality of cortisol release — moderate chronic stress (Rosmond et al., 1998), and/or reduced sensitivity to HPA axis feedback have been suggested (Ljung et al., 1996; Di Blasio et al., 2003; Jessop et al., 2003). Indeed, obesity has been associated with reduced sensitivity to glucocorticoid feedback (Ljung et al., 1996; Jessop et al., 2003), an effect thought to be mediated via altered sensitivity of glucocorticoid receptor (GR). However, increased feedback sensitivity may not be applicable to obese men (Pasquali et al., 2002), although this sex is typified by a more-android adipose distribution! Several studies suggest that individuals with common GR polymorphisms demonstrate HPA axis abnormalities in addition to a peripheral phenotype of increased adiposity, insulin resistance, and hypertension (Bjorntorp et al., 1999; Rosmond et al., 2000; Ljung et al., 2002; Di Blasio et al., 2003). Whether this GR genotype is a primary cause of obesity or whether obesity reflects HPA activation remains unclear but is the topic of ongoing investigations. An additional complexity stems from recent data showing that, in the nucleus, GR homodimers are part of a macromolecular complex with specific coactivator and corepressor proteins that, in combination, determine the actions of glucocorticoids-GR upon a target gene. Variation in density or function of coactivators and corepressors may underlie aspects of interindividual variation in glucocorticoid sensitivity (Li et al., 2003).
For the metabolic syndrome, case-control and cross-sectional studies have suggested that high blood pressure associates with somewhat elevated cortisol concentrations (in blood, saliva, or urine) (Watt et al., 1992; Filipovsky et al., 1996; Stolk et al., 1996; Rosmond et al., 1998; Walker et al., 1998; Fraser et al., 1999; Brunner et al., 2002). Similarly, insulin resistance and glucose intolerance are associated with higher circulating cortisol levels (Phillips et al., 1998; Reynolds et al., 2001). Higher morning plasma cortisol concentrations and increased responsiveness to corticotropin (ACTH) occur in adults having the additional cardiovascular risk factor of low birth weight (Phillips et al., 1998,2000; Levitt et al., 2000; Reynolds et al., 2001). However, all these associations between cortisol activity and hypertension/insulin resistance are independent of obesity. In such studies, obesity again is associated with lower (not higher) plasma cortisol, no change in (renal) 11ß-HSD2, and no difference in dermal sensitivity (vasoconstrictor responses) to glucocorticoids.
From these observational studies, it is not possible to dissect causality but they do not strongly favour a unifying hypothesis that ascribes the associations between obesity and other features of the metabolic syndrome to enhanced cortisol secretion. Rather, a primary or secondary increase in HPA axis activity may occur in subjects with hypertension and/or insulin resistance that do not directly predispose to obesity. Clearly, this syndrome cannot result from simple HPA activation or generalised alterations in tissue glucocorticoid sensitivity, yet the parallels with Cushing’s remain persuasive.
Thinking this through, it is plausible that the observed combination of increased total cortisol secretion with lower plasma cortisol levels during peak secretion in obesity reflects increased peripheral cortisol clearance. More than 20 years ago, obesity was associated with increased cortisol clearance rates (Strain et al., 1980; Lottenberg et al., 1998). Increased glucocorticoid clearance could itself reduce plasma levels and lead to compensatory HPA axis activation. Any such activation would lead to adrenal hyperresponsivity to stress and other stimuli.
|
III. Corticosteroid Receptors and Obesity |
|---|
|
TOPABSTRACTI. PreambleII. Blood Glucocorticoid Levels in...III. Corticosteroid Receptors and...IV. Tissue Metabolism of GlucocorticoidsV. 11ß-HSD1 and Glucocorticoid...VI. 11ß-HSD1 as a Drug Target...VII. SummaryACKNOWLEDGEMENTSREFERENCES |
|---|
Glucocorticoids act predominantly via nuclear receptors. There are two subtypes, GR and mineralocorticoid receptors (MR). These receptors are essential for life, since knockout (KO) of either by homologous recombination in mice is lethal at (for GR) or soon after (for MR) birth (Cole et al., 1995; Berger et al., 1998). GR are near ubiquitous, albeit their density varies considerably between cell types. MR are highly restricted to aldosterone target organs (i.e., distal nephron, distal colon, sweat and salivary glands) as well as some central nervous system (CNS) regions, notably, the hippocampus, and cardiovascular structures, including the heart. Metabolic effects associated with abdominal obesity are thought to originate in visceral adipose tissue and the liver, key sites determining hepatic glucose insulin and lipid homeostasis, as well as in skeletal muscle, the principal organ of glucose uptake. These tissues all highly express GR but have little or no MR. The importance of GR in obesity has been indicated by both its attenuation with adrenalectomy and GR antagonists in animal models and the effects on various mice bearing transgenic GR manipulations. However, the phenotypes are complex, since GR in the CNS also play a key role in HPA axis control such that a generalised transgenic reduction of GR causes obesity (Pepin et al., 1989). This is probably because HPA axis overactivity and marked hypercorticosteronaemia override the reduced density of GR in peripheral tissues. Such outcomes are predicted by the relatively low affinity of GR, which is little occupied by basal levels of glucocorticoids but fully occupied under conditions of glucocorticoid excess (e.g., severe stress) (de Kloet, 1991).
GR density in key metabolic cells, at least in skeletal muscle and adipose tissue, shows considerable interindividual variation. Importantly, in humans, GR mRNA levels in skeletal muscle correlate with insulin sensitivity and blood pressure (Reynolds et al., 2002; Whorwood et al., 2002). Moreover, dermal (vascular) sensitivity to GR agonists correlates with component disorders of the metabolic syndrome (Walker et al., 1996,1998; Andrews et al., 2002) and is associated with a common bclI restriction fragment length polymorphism (RFLP) of the GR gene (Panarelli et al., 1998), though any functional implications of this particular polymorphism are unclear. Thus, it has been hypothesised that genetic and environmental variations in GR density in specific metabolic tissues underlie the metabolic syndrome. Moreover, a mechanistic basis for such tissue-specific variation in GR expression has been provided through demonstration that the GR gene has a number of alternate untranslated exon1/promoter sequences in the 5' region that, at least in lymphocytes and the brain, can drive tissue-specific expression under some circumstances (McCormick et al., 2000). However, in obesity, GR mRNA is not consistently elevated in adipose tissue (Lindsay et al., 2003; Wake et al., 2003). This may reflect the fact that simple obesity and the metabolic syndrome can exist as discrete associated metabolic defects within a continuum of (visceral) obesity. Again, mechanisms pertaining to other features of the metabolic syndrome appear to differ from those that operate in obesity.
|
IV. Tissue Metabolism of Glucocorticoids |
|---|
|
TOPABSTRACTI. PreambleII. Blood Glucocorticoid Levels in...III. Corticosteroid Receptors and...IV. Tissue Metabolism of GlucocorticoidsV. 11ß-HSD1 and Glucocorticoid...VI. 11ß-HSD1 as a Drug Target...VII. SummaryACKNOWLEDGEMENTSREFERENCES |
|---|
- and 5ß-reductases), 6ß-hydroxylase, 20-reductase, and 11ß-HSDs. In rats and mice, which lack 17-hydroxylase in their adrenal cortex, the principal glucocorticoid is corticosterone, which is subject to analogous metabolism. This review will focus on the role of 11ß-HSDs.
B. HISTORY OF 11ß-HSD
50 years ago, Amelung and colleagues (1953) discovered an enzyme that catalysed the interconversion of 11-hydroxy glucocorticoids (cortisol, corticosterone) and 11-keto forms (cortisone, 11-dehydrocorticosterone): 11ß-hydroxysteroid dehydrogenase (Figure 1). While 11-hydroxyglucocorticoids are active at receptors, 11-ketosteroids have very low affinity for GR and MR and appear to all intents and purposes inert. This 11ß-HSD activity subsequently was described in a broad range of cells and tissues (Monder and White, 1993). In the 1980s, Monder and his coworkers in New York purified an 11ß-HSD activity from rat liver (Lakshmi and Monder, 1988). Homogenates, microsomal preparations, and purified enzyme from rat liver catalysed both 11ß-dehydrogenation of cortisol to inert cortisone and, typically to a lesser extent, the 11ß-reduction of cortisone to active cortisol. Thus, 11ß-HSD was thought to represent one of several arcane pathways for glucocorticoid clearance, so no specific function was ascribed to it. Subsequently, two 11ß-HSD isozymes have been characterised, isolated, and their cDNAs cloned, the products of distinct and only distantly related genes (reviewed in White et al., 1997; Stewart and Krozowski, 1999; Seckl and Walker, 2001).
|
D. 11ß-HSD TYPE 1
1. Biochemistry and Distribution
In contrast to the well-defined functions of 11ß-HSD2, although 11ß-HSD1 was the first isoform to be identified and characterised, it remained until recently the Cinderella of this story. 11ß-HSD1 has a much-lower affinity (i.e., low µM Km) for cortisol and corticosterone, is a nicotinamide adenine dinucleotide phosphate (NADP(H))-dependent enzyme, and is widely distributed. In cell homogenates or microsomal preparations, 11ß-HSD1 readily converts cortisol to cortisone. Thus, for several years, it was thought to be responsible for inactivation of cortisol in the kidney. However, more-recent data in intact cells either transfected with 11ß-HSD1 cDNA or in primary cells in culture show that in most, if not all, intact cells and organs, this enzyme is a predominant 11-ketoreductase, reactivating inert cortisone into cortisol (Low et al., 1994a; Hundertmark et al., 1995; Jamieson et al., 1995; Rajan et al., 1996). The preponderant reductive direction in vivo appears to be due more to the intracellular redox context of 11ß-HSD1 in the inner leaflet of the endoplasmic reticulum than to any structural feature, such as glycosylation (Agarwal et al., 1995). The mechanisms determining reaction directionality have not been elucidated fully. Nonetheless, 11ß-HSD1 co-precipitates with the NADPH-generating enzyme, hexose-6-phosphate dehydrogenase (Ozols, 1995), which may dictate the predominant 11-ketoreductase action of 11ß-HSD1 in vivo (Draper et al., 2003). Short-term, post-translational changes such as enzyme phosphorylation also may be important, particularly to explain the apparent instability of the 11-ketoreductase activity in homogenates despite plentiful NADPH, but remain to be clarified. Biochemical investigation of expressed 11ß-HSD1 suggests that the reductase reaction has cooperative rather than Michaelis-Menten kinetics, unlike the dehydrogenase, ensuring cortisol generation across a wide range of substrate concentrations (Maser et al., 2002), as observed ex vivo in perfused liver (Jamieson et al., 2000).
In a few studies (e.g., in Leydig cells), 11ß-dehydrogenase activity has been reported in apparently intact cell preparations (Gao et al., 1997; Ge and Hardy, 2000). Others have found predominantly 11-ketoreduction in similar preparations (Leckie et al., 1998). The discordance of reported reaction direction in intact cells, largely involving Leydig cells and preadipocytes (Bujalska et al., 2002), remains unresolved but is an important issue, given the current enthusiasm to decrease/inhibit 11-ketoreductase as a therapeutic target. It has been suggested that some 11ß-HSD1 may have been liberated from possible damaged cells in studies reporting 11ß-dehydrogenase activity (Seckl and Walker, 2001), since this direction is more stable in homogenates and broken-cell preparations. More plausibly, adipose tissue dissociation and culture lead to early induction of adipocyte cytokine synthesis and release (Ruan et al., 2003). Cytokines have potent effects upon 11ß-HSD1 and other adipocyte products, which may explain some discordances in the human adipose 11ß-HSD1 activities between freshly isolated tissue and cultured cells.
11ß-HSD1 is expressed in kidneys of rats but negligible expression occurs in adult human kidney (Stewart et al., 1994). However, 11ß-HSD1 is widely expressed in other tissues in humans (Ricketts et al., 1998) and rodents (Moisan et al., 1990b), including liver (Agarwal et al., 1989; Tannin et al., 1990), adipose tissue (Bujalska et al., 1997; Napolitano et al., 1998), lung (Hundertmark et al., 1993), skeletal muscle, cardiac and vascular smooth muscle (Walker et al., 1991; Brem et al., 1995; Christy et al., 2003), anterior pituitary gland, brain (including hippocampus) (Moisan et al., 1990a,b; Lakshmi et al., 1991; Sakai et al., 1992), and adrenal cortex (Shimojo et al., 1996). Notably, these locations have high GR rather than MR expression, with the exception of hippocampus and heart, where MR act as high-affinity sites for physiological glucocorticoids rather than aldosterone in vivo (Whorwood et al., 1991).
These observations in cells suggested a novel role for 11ß-HSD1 involving reactivation rather than inactivation of glucocorticoid. This, indeed, occurs in intact organs, at least in the liver. Isolated perfused cat (Bush et al., 1968) or rat (Jamieson et al., 2000) liver models suggest that 11ß-HSD1, which is the only isozyme expressed in liver, is a predominant 11ß-reductase with a high capacity for reactivating 11-ketosteroid substrate over a broad range of substrate concentrations. These findings can be extrapolated to human liver in vivo, since historical work suggests that, on oral administration, cortisone (the first pharmacological glucocorticoid used in man) is rapidly activated to cortisol. Indeed, recent studies confirm that very little oral cortisone reaches the systemic circulation (A. Jamieson et al., 1999) and that hepatic vein cortisol/cortisone ratios are very high (Walker et al., 1992). 11-ketoreductase activity has been shown in other human tissues in vivo, including subcutaneous adipose tissue (Katz et al., 1999).
2. Substrate Levels
For 11-ketoreductase to play a physiological role in regulating receptor exposure to active glucocorticoids (as opposed to a pharmacological role when cortisone is administered), there must be a substantial pool of substrate inert 11-ketosteroids in the circulation and tissues. In vivo, the main source of 11-ketosteroid is 11ß-HSD2 dehydrogenation of cortisol and corticosterone, which predominantly occurs in the kidney (Whitworth et al., 1989). In humans, cortisone circulates at levels approximating 50–100 nmol/l (Walker et al., 1992). While this level is lower than diurnal peak cortisol of 400–600 nmol/l, cortisone is largely unbound, while 
95% of cortisol is sequestered by binding to plasma proteins such as CBG (Dunn et al., 1981). Moreover cortisone, unlike cortisol, shows no pronounced diurnal rhythm and so is constantly available for conversion to active glucocorticoid (Walker et al., 1992). Estimates of "free" cortisol levels are rather imprecise but approximate 0.5–1 nmol/l at the diurnal nadir. In the rat, plasma concentrations of 11-dehydrocorticosterone are also 50 nmol/l, though in the mouse, levels are lower ( 3–5 nmol/l) (Kotelevtsev et al., 1997). Thus, for at least the quiescent part of the diurnal cycle, circulating cortisone levels equal or exceed free cortisol levels and similar ratios pertain in rodents.
E. FUNCTIONAL STUDIES OF 11ß-HSD1 IN LIVER
Initial findings suggesting that 11ß-HSD1 increases effective intracellular glucocorticoid action were obtained in liver. Here, glucocorticoids oppose the actions of insulin, for example, by upregulating expression of the key enzyme for gluconeogenesis, phosphoenolpyruvate carboxykinase (PEPCK). In male rats, estradiol potently downregulates 11ß-HSD1 expression (Low et al., 1993) and, only in the presence of glucocorticoids, downregulates PEPCK expression (P.M. Jamieson et al., 1999). Such indirect studies, as well as the use of relatively nonselective, liquorice-based inhibitors (Walker et al., 1995; Andrews et al., 2003), indicate that impaired activity of 11ß-HSD1 in liver is associated with features of reduced glucocorticoid action and increased insulin sensitivity in hepatocytes. This has been supported by recent studies with selective 11ß-HSD1 inhibitors (Alberts et al., 2002,2003) (see below).
To explore this further, 11ß-HSD1 KO mice have been generated (Kotelevtsev et al., 1997). These mice appear to develop normally and are viable, fertile, and apparently normotensive. This model shows that 11ß-HSD1 is the sole major 11ß-reductase, at least in mice, since adrenalectomised 11ß-HSD1 KO mice cannot convert administered 11-dehydrocorticosterone to active corticosterone. Plasma corticosterone levels are modestly elevated at the diurnal nadir, presumably due to somewhat deficient feedback upon the HPA axis (11ß-HSD1 is expressed in hippocampus, PVN, and anterior pituitary) (Harris et al., 2001). However, despite slightly elevated basal plasma corticosterone levels, 11ß-HSD1-/- mice have a phenotype compatible with impaired intracellular glucocorticoid regeneration and reduced antagonism of insulin action. They show impaired induction of the key glucocorticoid-inducible hepatic enzymes PEPCK and glucose-6-phosphatase on fasting and an attenuated hyperglycemic response to novel environment stress or chronic high-fat feeding (Kotelevtsev et al., 1997). Importantly, the mice have more-than-adequate, stress-induced HPA axis responses (Harris et al., 2001) and do not exhibit hypoglycaemia with prolonged fasting (Kotelevtsev et al., 1997). Further aspects of the 11ß-HSD1 null phenotype affect hepatic metabolic control and cardiovascular risk factors (Morton et al., 2001). Ad lib-fed 11ß-HSD1-/- mice have lower plasma triglyceride and elevated "cardioprotective" high-density lipoprotein (HDL) cholesterol levels (Morton et al., 2001). This appears to be due to increased hepatic ß oxidation of lipids, rather than to altered lipogenesis. Indeed, hepatic expression of peroxisome proliferator-activated receptor alpha (PPAR), a key glucocorticoid-sensitive transcription factor driving lipid metabolism, is elevated in fed 11ß-HSD1-/- mice. The major murine HDL carrier, apolipoprotein AI, is increased in the null mice, whereas serum apolipoprotein CIII, which increases plasma triglycerides by inhibiting hepatic lipolysis and interfering with transfer of triglycerides to the liver, is reduced. Additionally, preliminary data suggest favourable effects upon haemostasis, as in liver, glucocorticoid-inducible A-fibrinogen transcript levels are reduced (Morton et al., 2001).
F. 11ß-HSD1 IN ADIPOSE TISSUE
1. Isozymes and Directionality
Glucocorticoids play a key role both in the regulation of adipose tissue metabolism and in the differentiation of preadipocytes into adipocytes (Gaillard et al., 1991). 11ß-HSD expression in adipose tissue was noted in the 1980s and thought to be a dehydrogenase (Monder and White, 1993). Studies in adipose cells derived from the mammary gland confirmed the presence of 11ß-dehydrogenase activity in vitro (Quirk et al., 1991). More recently, 11ß-HSD in adipose tissue has been re-examined in light of the 11-ketoreductase predominance of this isozyme in intact hepatocytes and whole liver.
In terms of isozymes, 11ß-HSD1, but not 11ß-HSD2, mRNA is expressed in rat white (epididymal) adipose tissue (WAT) (Napolitano et al., 1998) and in human adipose tissue (Bujalska et al., 1997; Sun et al., 1997; Paulmyer-Lacroix et al., 2002). The enzyme appears to be expressed similarly in adipocytes and the stromal vascular fraction (enriched with preadipocytes and vascular cells) in the rat, though perhaps is more highly concentrated in adipose stromal cells in humans (Bujalska et al., 1997). 11ß-HSD also is expressed in murine fibroblast-adipocyte 3T3-F442A and 3T3-L1 cell lines, which closely reproduce preadipocyte differentiation in vitro when cultured in the presence of inducing agents, including dexamethasone. These respond to glucocorticoids in a similar manner to adipose tissue in vivo (MacDougald et al., 1994; MacDougald and Lane, 1995). 3T3 cells exclusively express the 11ß-HSD1 isozyme and, in intact cells, the reaction direction is solely 11-ketoreduction (Napolitano et al., 1998). 11ß-HSD1 mRNA and activity in 3T3 cells increases markedly with differentiation of cells from their preadipocyte to mature adipocyte state. 11ß-HSD1 appears to be a late-differentiation gene in 3T3 cells. 3T3 cell in vitro differentiation is promoted by insulin and glucocorticoids. However, the increase in 11ß-HSD1 with differentiation does not appear to be mediated directly by these hormones, since, at least in differentiated 3T3 cells, insulin and glucocorticoids reduce 11ß-HSD1 expression. By contrast, intact human adipose stromal cells cultured in the presence of insulin and cortisol have been reported to show bi-directional 11ß-HSD1 activity, although 11-ketoreductase activity predominates in cells derived from omental fat, with further induction by glucocorticoids in mature cells (Bujalska et al., 1997).
Some recent findings suggest that 11ß-HSD1 reaction direction may alter in adipose tissue with differentiation state (Bujalska et al., 2002), although this was not apparent in a clonal 3T3 cell model (Napolitano et al., 1998). Thus, primary cultures of human stromal and mature adipocytes taken from patients undergoing elective abdominal surgery showed predominant 11ß-dehydrogenase activity in freshly isolated stromal cells from the omental compartment, whereas intact mature adipocytes had predominantly 11-ketoreductase activity. Following culture for 2 weeks, 11-ketoreductase activity predominated, despite no significant changes in 11ß-HSD1 mRNA. The authors speculated that since glucocorticoids inhibit cell proliferation, dehydrogenase in immature preadipocytes facilitates proliferation. However, once differentiation is initiated, 11-ketoreductase amplifies glucocorticoid levels, promoting adipogenesis. While these findings might be explained by cell breakage leading to liberation of enzyme into a homogenate-like state in the early phases of cell culture and/or by cytokine-inductive effects of adipose tissue dissociation and early culture in vitro, an intriguing biochemical mechanism for shifts in 11ß-HSD1 directionality has been advanced. This advocates varying developmental expression of hexose-6-phosphate dehydrogenase, which co-precipitates with 11ß-HSD1 and acts to generate NADPH inside the lumen of the endoplasmic reticulum in adipose tissue (Stegeman and Klotz, 1979). Intriguingly, while polymorphisms in the 11ß-HSD1 gene (HSD11B1) correlate rather poorly (Draper et al., 2002; Gelernter-Yaniv et al., 2003), if at all (Caramelli et al., 2001), with adipose distribution in humans, changes in the hexose-6-phosphate dehydrogenase gene may be key. Thus, in three patients with rare cortisone reductase deficiency (Phillipov et al., 1996), null mutations in the HSD11B1 gene have not been found. Recent data suggest that such individuals have more-subtle mutations in both HSD11B1 and in hexose-6-phosphate dehydrogenase, resulting in both low 11ß-HSD1 expression and low endoplasmic reticulum NADPH generation, with consequent loss of 11ß-HSD1 reductase activity (Draper et al., 2003). It remains to be established whether or not this mechanism applies to more-common cases of 11ß-HSD1 dysregulation in adipose tissue with obesity and associated insulin-resistance states, such as the polycystic ovary syndrome, which phenotypically most closely resembles cortisone reductase deficiency.
2. Depot Differences
Fat depots may differ in 11ß-HSD1 activity, although no clear consensus exists. Discrepancies may reflect species differences or assay conditions employed. Thus, in humans, cultures of omental stromal cells have more 11ß-HSD1 than subcutaneous stromal cells (Bujalska et al., 1999). In contrast, in various nonobese strains of mice, 11ß-HSD1 mRNA expression and activity are higher in freshly isolated peripheral than visceral fat depots (Morton et al., 2003). This is consistent with the idea that increased glucocorticoid reactivation correlates with increased differentiation potential of this depot. These interspecies differences may reflect the conditions of assay; ex vivo cultures in high glucocorticoid concentrations employed in humans perhaps favour increased activity in visceral fat because this depot has more GR (Masuzaki et al., 2001) and/or has greater capacity for glucocorticoid-induced differentiation of immature to mature (i.e., higher 11ß-HSD1-expressing) adipocytes.
3. Molecular Regulatory Mechanisms
Several studies have examined regulation of 11ß-HSD1 expression in adipose tissue. Insulin appears to reduce 11ß-HSD1 gene expression in fat cells. Although glucocorticoids reduce adipose 11ß-HSD1 expression in some systems, in others, they stimulate its activity, with effects reflecting the assay conditions and perhaps species. Proinflammatory cytokines such as tumor necrosis factor alpha (TNF) and interleukin-1 (IL1) increase and anti-inflammatory cytokines reduce 11ß-HSD1 in adipose cells in vitro (Handoko et al., 2000; Friedberg et al., 2003). As detailed earlier, both glucocorticoids and insulin exert effects, albeit in vitro. However, regulation by other nuclear receptors remains less certain. Thus, while PPAR
ligands such as thiazolidinediones downregulate 11ß-HSD1 in 3T3 cells in vitro and in epididymal fat in mice in vivo (Berger et al., 2001), the same was not observed in lean or obese Zucker rats (Livingstone et al., 2000b). Liver X receptor (LXR) agonists also partially downregulate 11ß-HSD1 in vitro and in vivo (Stulnig et al., 2002). The mechanism may be indirect, which may explain the discrepancies observed in such studies. In terms of more-direct effects, CCAAT/enhancer binding protein alpha (C/EBP) binds to several sites on the 11ß-HSD1 promoter and, when transfected into human hepatoma hepG2 cells, increases 11ß-HSD1 promoter activity (Williams et al., 2000). In contrast, C/EBPß acts as a dominant-negative repressor of 11ß-HSD1 transcription when added to C/EBP, although alone, it is a weak inducer. Electrophoretic mobility shift assays (EMSA) using nuclear extracts from undifferentiated and differentiated 3T3-F442A cells suggest that C/EBP also may regulate 11ß-HSD1 expression in adipocytes (Williams et al., 1993). These data suggest the possibility that 11ß-HSD1 regulation in the adipocyte by insulin and glucocorticoids may be mediated indirectly through changes in C/EBP-related proteins. Since under many circumstances, 11ß-HSD1 regulation in liver and adipose tissue is discordant, as with other well-characterised C/EBP-regulated genes such as PEPCK, the fine details of control are likely to be tissue specific. 11ß-HSD1 mRNA regulation in 3T3-F442A cells parallels glycerol-3-phosphate dehydrogenase, a key enzyme in triglyceride synthesis and a well-characterised marker of adipocyte differentiation (Moustaid et al., 1990).
G. AP2–11ß-HSD1 ADIPOSE OVEREXPRESSING TRANSGENIC MICE: A MODEL OF THE METABOLIC SYNDROME
The preponderance of data suggest that 11-ketoreductase activity is increased in adipose tissue in obese rodent models and humans (see below). The key question is whether increased adipose 11ß-HSD1 is a cause or a consequence of obesity and any associated metabolic syndrome. To dissect this, mice overexpressing 11ß-HSD1 selectively in adipose tissue have been generated, exploiting the adipocyte fatty acid binding protein (aP2) promoter (Masuzaki et al., 2001). Two lines have been selectively developed to exhibit 2- to 3-fold overexpression of 11ß-HSD1 in adipose tissue, which represents the degree of enzyme increase found in obese humans and animals. These aP2–11ß-HSD1 mice are viable and appear developmentally normal. The transgene is expressed in all adipose depots examined and in brown adipose tissue (BAT) but is absent from other 11ß-HSD targets such as the brain, liver, skeletal muscle, and kidney. The effect of the transgene is to approximately double corticosterone levels within adipose tissue, while circulating glucocorticoid levels are unaltered, at least under basal conditions. As a consequence of local intra-adipose glucocorticoid excess, aP2–11ß-HSD1 mice are modestly obese. Strikingly, the obesity is predominantly intra-abdominal, with a > 3-fold increase in the mesenteric fat depot, whereas peripheral fat depots are significantly, but much less spectacularly, increased. The visceral adipose expansion in these animals (which overexpress 11ß-HSD1 similarly in all adipose beds) may relate to much-greater levels of GR in visceral adipose than peripheral depots in mice (Masuzaki et al., 2001). Indeed, the transgenics show increased expression of the glucocorticoid target gene lipoprotein lipase, particularly in mesenteric adipose tissue that could drive lipid accumulation in this adipose depot. It is interesting that GR is not downregulated in the transgenic animals. Glucocorticoid autoregulation of GR is highly tissue specific and is only seen acutely in many tissues.
In association with this localised change in glucocorticoid exposure and the consequent increased visceral fat mass, aP2–11ß-HSD1 mice develop all major features of the metabolic syndrome. The animals are markedly glucose intolerant and insulin resistant, features exacerbated by high-fat feeding. The animals do not show obvious fasting hyperglycemia but this becomes markedly manifest after a glucose load, suggesting the deficit is in peripheral glucose uptake rather than a predominant increase in hepatic glucose production. They also show dyslipidemia with elevated FFA and triglyceride levels (Masuzaki et al., 2001). Glucocorticoids are potent secretagogues of leptin from adipose tissue. Serum leptin levels are elevated in the transgenic animals disproportionately to their degree of obesity, suggesting leptin resistance, as observed in human obesity.
aP2–11ß-HSD1 transgenic mice show adipocyte hypertrophy but not hyperplasia (Masuzaki et al., 2001). This is perhaps not surprising, as aP2 is a late differentiation gene in adipocytes, so expression and effects in preadipocytes would not be expected. Within adipocytes themselves, aP2–11ß-HSD1 mice show changes concordant with decreased sensitivity to insulin and/or increased corticosterone levels. Thus, transgenic adipose tissue has decreased expression of the insulin-sensitising factor adiponectin (acrp30, adipoQ) (Yamauchi et al., 2002) and increased expression of TNF, an adipose cytokine that causes insulin resistance (Hotamisligil et al., 1993). Both are commensurate with the effects of glucocorticoid excess (Fasshauer et al., 2002; Viengchareun et al., 2002). Serum TNF levels also are elevated. In contrast, adipose mRNA-encoding resistin, which may lead to insulin resistance (Steppan et al., 2001), is reduced in transgenic adipose tissue, perhaps as a consequence of glucocorticoid excess (Viengchareun et al., 2002). Intriguingly, angiotensinogen mRNA, which normally is expressed at low levels in adipose tissue, is strikingly elevated in aP2–11ß-HSD1 transgenic mouse fat. This glucocorticoid-regulated transcript may underpin the marked hypertension seen in these animals. The hypertension is associated with renin-angiotensin-aldosterone system (RAAS) activation and is strikingly sensitive to low doses of an angiotensin II receptor antagonist (Masuzaki et al., 2003). Finally, the animals show intriguing hyperphagia, indicating the potential for a novel pathway, other than leptin deficiency (though leptin resistance is not excluded), of signalling between the adipocyte and CNS appetite centres. The aP2–11ß-HSD1 transgenic mouse faithfully models the major features of the metabolic syndrome.
The aP2–11ß-HSD1 transgenic mouse permits speculation about the subtle but important phenotypic differences between Cushing’s syndrome and the metabolic syndrome. If the metabolic syndrome is, indeed, generated by local glucocorticoid excess inside adipose tissue and/or liver but with normal circulating cortisol levels, then tissue without the adipose-hepatic axis will be unaffected. We would not necessarily expect Cushingoid skin thinning, vascular fragility, delayed wound healing, myopathy, or CNS effects, which may well reflect the actions of systemic glucocorticoid excess upon skin, blood vessels, skeletal muscle, and brain without the compass of localadipose and hepatic glucocorticoid excess seen in the 11ß-HSD1 transgenic models.
H. LIVER GLUCOCORTICOID EXCESS:APOE-11ß-HSD1 TRANSGENIC MICE
Blood from visceral adipose tissue drains via the hepatic portal vein to the liver. Unsurprisingly, therefore, aP2–11ß-HSD1 transgenic mice have elevated levels of corticosterone and FFA in the portal plasma, implying that the liver is exposed to excess glucocorticoids. To address this and to examine whether excess 11ß-HSD1 in the liver per se can cause the metabolic syndrome, transgenic mice overexpressing 11ß-HSD1 in the liver have been generated under the ApoE promoter. These mice are viable and appear normal. As adults, the mice show modest insulin resistance and hypertriglyceridemia, along with substantially increased fat deposits in the liver. However, ApoE-11ß-HSD1 transgenic mice have normal weight with normal distribution and mass of fat depots and show normal glucose tolerance. Intriguingly, the animals are hypertensive, with activation of the RAAS, in this case due, apparently, to overexpression of angiotensinogen in the liver but not adipose tissue. Overexpression of 11ß-HSD1 in liver produces an attenuated metabolic syndrome phenotype without visceral obesity. This might be of pathogenic relevance in cases such as patients with the insulin resistance of myotonic dystrophy (Johansson et al., 2001), in which liver 11ß-HSD1 activity is raised.
I. DEFICIENCY OF 11ß-HSD1 IN ADIPOSE TISSUE: LESSONS FROM THE KO MOUSE
11ß-HSD1 expression in adipose tissue produces a striking metabolic syndrome phenotype. But what about enzyme deficiency (Figure 2)? Inhibitors have provided some conflicting data that may relate to difficulties in access of carbenoxolone to adipose tissue in vivo, at least in the doses administered to rodent models (Livingstone and Walker, 2003) and perhaps humans (Andrews et al., 2003). Adiposity in 11ß-HSD1 KO animals has been examined on both intrinsically obesity-resistant and obesity-prone genetic backgrounds. Intra-adipose corticosterone levels are lowered substantially in the 11ß-HSD1-/- animal in the face of modestly elevated plasma corticosterone concentrations (Morton et al., 2003). On the obesity-prone C57Bl/6J genetic background, high-fat diet-fed 11ß-HSD1 KO mice gain significantly less weight than controls, despite relative hyperphagia. This appears due to an enhanced metabolic rate, as judged by increased core body temperature (Morton et al., 2002). With high-fat diet on either genetic background, 11ß-HSD1 KO mice preferentially gain adipose tissue in the "metabolically safer" peripheral depots rather than in the "metabolically disadvantageous" visceral sites. While the explanation of fat redistribution in 11ß-HSD1-/- animals is uncertain, these mice show higher expression of the thiazolidinedione target adipogenic transcription factor PPAR receptor in all adipose tissue beds. Furthermore, 11ß-HSD1-/- mice show a greater increase of adipose PPAR receptors with high-fat feeding than wild-type (WT) mice. PPAR ligands cause fat redistribution to the periphery, which may underpin the favourable morphology seen (Kelly et al., 1999; Sewter et al., 2002). Additionally, 11ß-HSD1 null animals show greater induction of uncoupling protein-2 (UCP-2) in mesenteric adipose tissue than WT mice (Morton et al., 2002), which may allow local calorie wastage rather than storage as fat (Digby et al., 2000). UCP-2 is downregulated by glucocorticoids (Udden et al., 2001) and upregulated by PPAR activation (Kelly et al., 1998), so its induction in the 11ß-HSD1 null mouse adipose tissue is not unexpected.
|
mRNAs are reduced, whereas adiponectin is increased, again compatible with an adipose-mediated, insulin-sensitised phenotype. Thus, overall, the mouse shows improved glucose tolerance, increased insulin sensitivity, and reduced intratissue glucocorticoid levels in the face of modest hypercorticosteronaemia (Harris et al., 2001). These beneficial effects of deletion in 11ß-HSD1 in adipose tissue are accompanied by changes in hepatic gene expression consistent with increased ß oxidation of lipids in the liver (Morton et al., 2001). Although the relative importance of liver and adipose tissues in contributing to this beneficial phenotype remains to be determined, it is clear that loss of 11ß-HSD1 has few, if any, deleterious metabolic effects, at least as assessed in this mouse model. Indeed, the lifespan of this model is apparently normal and the animals resist the normal glucocorticoid-associated decline of memory and other cognitive processes with ageing (Yau et al., 2001). Thus, loss of this enzyme appears to be beneficial.
1. Why Might 11ß-HSD1 Exist?
However, the phenotype of the 11ß-HSD1-/- mouse and the various transgenic and pharmacological manipulations detailed here raise the evolutionary question of the overall biological purpose of such an apparently virtueless enzyme. It cannot exist merely as a convenient target for therapy. We can only speculate. Given that our vertebrate biological antecedents had to suffer considerably more starvation than plenty, perhaps the boost of glucocorticoid from local tissue reactivation, particularly during the diurnal nadir, provided an extra lift to metabolic processes underpinning calorie generation. This might well have assisted survival during periods of starvation. Even during relative nutritional plenty, the sustenance of some key metabolic organs, while others, including the HPA system, were quiescent (i.e., during sleep), may have had an advantage. However, in modern westernised existence with an excess of calorifically rich foods available at every street corner, perhaps such activity becomes redundant, particularly as obesity and its metabolic consequences become the problem. It is an intriguing notion that an enzyme putatively required to help survive starvation becomes a metabolic "appendix" in the presence of nutritional plenty and a target for therapy of the consequences of calorific excess.
|
V. 11ß-HSD1 and Glucocorticoid Metabolism in Obesity |
|---|
|
TOPABSTRACTI. PreambleII. Blood Glucocorticoid Levels in...III. Corticosteroid Receptors and...IV. Tissue Metabolism of GlucocorticoidsV. 11ß-HSD1 and Glucocorticoid...VI. 11ß-HSD1 as a Drug Target...VII. SummaryACKNOWLEDGEMENTSREFERENCES |
|---|
- and -ß) are activated in both animal (Livingstone et al., 2000a) and human obesity (Andrew et al., 1998,2002). The underlying cause of increased enzyme activity is unknown but these enzymes may be regulated by insulin, lipids, and substrate availability. However, here, we concentrate upon another pathway, the 11ß-HSDs.
A. 11ß-HSD1 IN OBESITY
Measurement of 11ß-HSD1 in vivo in humans is not straightforward (Walker, 2000). The original assessment relied upon measuring ratio of the most-abundant urinary metabolites of cortisol and cortisone (i.e., 5- and 5ß-tetrahydrocortisols/5ß-tetrahydrocortisone). However, this ratio also is influenced by activities of other enzymes, including 11ß-HSD2, 5-reductase, 5ß-reductase, and 3-HSDs. For example, carbenoxolone, which inhibits both renal 11ß-HSD2 and hepatic 11ß-HSD1, does not alter the ratio of cortisol/cortisone metabolites (Stewart et al., 1990). In obesity, the urinary cortisol/cortisone metabolite ratio has been reported to be increased (Andrew et al., 1998; Rask et al., 2002; Tiosano et al., 2003), unchanged (Fraser et al., 1999; Reynolds et al., 2001), or decreased (Stewart et al., 1999; Rask et al., 2001). This inconsistency may depend in part upon variations — for example, with sex — in 5- and 5ß-reductase activity. Recent findings suggest that increased fat content of the liver, which constitutes a risk factor for insulin resistance over and above the effect of obesity, is associated with increased urinary excretion of 5ß-reduced cortisol metabolites. Liver fat content was shown to be a more-powerful determinant of cortisol/cortisone metabolite ratios than the degree of obesity (Westerbacka et al., 2003). However, this important confounder has not been accounted for in most studies in which liver fat content was not measured. Another source of inconsistency of cortisol/cortisone metabolite ratio in obesity is the possible influence of tissue-specific disruption of 11ß-HSD1 activity. It is this last crucial point that turns out to be the key issue in obesity.
Studies in leptin-resistant obese Zucker rats revealed that obesity was associated with decreased 11ß-HSD1 expression and activity in liver but increased 11ß-HSD1 in omental adipose tissue (Livingstone et al., 2000a). This led to dissection of potential tissue-specific differences in obese humans. In humans, conversion of cortisone after oral administration to cortisol in peripheral plasma, reflecting first-pass metabolism by hepatic 11ß-HSD1, is impaired in obesity (Stewart et al., 1999; Rask et al., 2001,2002). In contrast, in subcutaneous abdominal adipose tissue, 11ß-HSD1 activity is increased both in vivo and in vitro (Rask et al., 2001,2002; Lindsay et al., 2003; Wake et al., 2003). Further studies have confirmed that the increased 11ß-HSD1 activity in biopsies is accompanied by increased 11ß-HSD1 mRNA (Paulmyer-Lacroix et al., 2002; Lindsay et al., 2003; Wake et al., 2003; Westerbacka et al., 2003). By analogy with the transgenic mouse having 3-fold overexpression of 11ß-HSD1 in adipose tissue (Masuzaki et al., 2001,2003), one would expect a Cushingoid phenotype to result from the similar magnitude of increase in adipose 11ß-HSD1 documented in human obesity. Interestingly, while increased subcutaneous adipose 11ß-HSD1 is associated with insulin resistance in obesity, it is not linked specifically with visceral adiposity or hypertension. In these respects, mice and men may differ but it remains to be determined whether 11ß-HSD1 mRNA or activity is increased in intact omental adipose tissue in obese subjects (Tomlinson et al., 2002).
The mechanisms underlying increased adipose 11ß-HSD1 in obesity are uncertain. 11ß-HSD1 transcription is highly regulated, including by many factors that are altered in obesity (e.g., cytokines, sex steroids, growth hormone (GH), insulin, PPAR and - agonists). In obese Zucker rats and ob/ob mice, downregulation of hepatic 11ß-HSD1 is reversible with other manipulations that induce weight loss (Livingstone et al., 2000b; Liu et al., 2003). Attempts to link the 11ß-HSD1 genotype with obesity have not been successful (Caramelli et al., 2001; Draper et al., 2002), although arguably the intermediate phenotypes (i.e., anthropometric measurements or urinary cortisol/cortisone metabolite ratios) employed are too insensitive to allow clear inferences about the influence of known polymorphisms in the 11ß-HSD1 gene.
B. 11ß-HSD1 AND GH
In addition to dysregulation in patients with idiopathic obesity, the central adiposity of GH deficiency and hypothalamic obesity may be associated with changes in cortisol metabolism and 11ß-HSD1 activation. 11ß-HSD1 is inhibited by GH and insulin-like growth factor-1 (IGF-1) in vitro (Napolitano et al., 1998; Tomlinson et al., 2001). In humans with hypopituitarism and acromegaly, this effect may be important (Moore et al., 1999; Trainer et al., 2001; Tiosano et al., 2003). Indeed, GH therapy reduces cortisol:cortisone metabolite ratios in patients with hypopituitarism (Weaver et al., 1994) and in idiopathic obesity (Tomlinson et al., 2003). As in studies of idiopathic obesity, the difficulty is in the interpretation of urinary cortisol:cortisone metabolite ratios, which do not reflect any tissue-specific changes. In vivo animal studies have shown that GH regulates liver 11ß-HSD1, with the direction of change dependent upon sex-specific patterns of pulsatile GH release (Low et al., 1994b). These findings may not explain changes in adiposity but may give a mechanism for improvements in insulin sensitivity in liver alone when GH is administered at the constant levels typical of female patterns that downregulate hepatic 11ß-HSD1. Whether such effects pertain to humans is not fully known.
C. 11ß-HSD IN THE BRAIN IN OBESITY
While not the main object of this review (for details of 11ß-HSDs in the CNS, see Seckl, 1997; Yau and Seckl, 2001), specific CNS changes in 11ß-HSDs in obese rodents have been reported that add to the complexity of tissue-specific dysregulation. In particular, obese Zucker rats have reduced 11ß-HSD1 mRNA in the hippocampus (Mattsson et al., 2003). The enzyme functions as a reductase in hippocampal neurons (Rajan et al., 1996). The reduced expression of 11ß-HSD1 may, in part, underlie the reduced sensitivity to glucocorticoid feedback inhibition of the HPA axis and consequent hypercorticosteronaemia of Zucker rats. Moreover, the feedback defect appears linked more to MR than GR, both in obese rodents (Mattsson et al., 2003) and humans (Jessop et al., 2003). Therefore, studies with GR-selective agonists like dexamethasone may be misleading (Weyer et al., 1997). Whether such effects occur in obese humans is unknown (and is probably unknowable) but 11ß-HSD1 is expressed in the human hippocampus (Sandeep et al., 2002) and may be targeted by 11ß-HSD inhibitors (Jellinck et al., 1993).
|
VI. 11ß-HSD1 as a Drug Target for Therapy in the Metabolic Syndrome |
|---|
|
TOPABSTRACTI. PreambleII. Blood Glucocorticoid Levels in...III. Corticosteroid Receptors and...IV. Tissue Metabolism of GlucocorticoidsV. 11ß-HSD1 and Glucocorticoid...VI. 11ß-HSD1 as a Drug Target...VII. SummaryACKNOWLEDGEMENTSREFERENCES |
|---|
First, studies showing increased 11ß-HSD1 activity in adipose tissue in humans mainly have investigated abdominal subcutaneous fat. While this is not necessarily as distinct from visceral adipose as other subcutaneous depots — and subcutaneous fat is perhaps the major source of adipose hormones and cytokines — this depot is not the same as visceral fat and probably contributes little, if anything, to the portal venous influx to the liver.
Second, mice are not humans in key aspects of their biochemistry. Thus, murine HDL cholesterol represents the majority of total cholesterol, while in humans, the preponderance is reversed, with most cholesterol transported in low-density lipoprotein (LDL) particles. Mice have corticosterone as their sole glucocorticoid, whereas in humans, cortisol predominates. Corticosterone has a higher affinity for GR and MR. Mice do not spontaneously get atheroma without a specific mutation (e.g., ApoE-/-), whereas humans do.
Third, while human subcutaneous adipose 11ß-HSD1 correlates with obesity and increased insulin resistance, it does not correlate closely with the distribution of body fat, notably in the visceral compartment (Westerbacka et al., 2003).
Fourth, it remains unclear whether 11ß-HSD1 activity is also increased in omental adipose tissue in obese humans and drives a Cushingoid phenotype of central obesity (Bujalska et al., 1997). One study has reported that 11ß-HSD1 in omental adipose obtained during surgery is not correlated with obesity. Indeed, when adipocytes were cultured, there was an inverse correlation between 11ß-HSD1 and obesity (Tomlinson et al., 2002), although the effects of anaesthesia and/or stress on the subsequent pattern of gene expression and 11ß-HSD1 activity and direction are difficult to anticipate. Attempts to demonstrate a close association between increased 11ß-HSD1 and intra-adipose cortisol levels or glucocorticoid-dependent gene transcription have been inconclusive (Lindsay et al., 2003; Wake et al., 2003). It has been proposed that the level of 11ß-HSD1 mRNA expression and total protein may not be the most-important determinant of net cortisol generation from cortisone, since the balance between reductase and dehydrogenase activities may be crucially dependent upon cofactor generation by hexose-6-phosphate dehydrogenase (Draper et al., 2003). While this hypothesis has not been tested definitively, it is notable that subjects with combined 11ß-HSD1 and hexose-6-phosphate dehydrogenase mutations putatively conferring impaired conversion of cortisone to cortisol are not universally lean (Phillipou et al., 1996; A. Jamieson et al., 1999; Draper et al., 2003).
Fifth, understanding the molecular mechanisms regulating 11ß-HSD1 is key. While insulin and other hormones exert indirect effects in rodents, the mechanisms in humans are poorly understood. However, recent data suggest that patients with glucose intolerance have paradoxically normal cortisol secretion, which is perhaps inappropriately high, given enhanced central and peripheral tissue sensitivity to glucocorticoids (Andrews et al., 2002). In these subjects, "normal" 11ß-HSD1 activity in adipose tissue with reduced hepatic activity suggests that tissue-specific changes in 11ß-HSD1 in hyperglycemia differ from those in primary obesity (Andrews et al., 2002). Thus, increased adipose 11ß-HSD1 appears to relate specifically to obesity, rather than to hyperglycaemia or insulin resistance (Westerbacka et al., 2003) per se.
Despite these caveats, 11ß-HSD1 inhibition is a tempting target for treatment of the metabolic syndrome and its complications. In this line, at least one selective inhibitor has been formulated and doubtless others will be developed.
A. INHIBITOR STUDIES: OLD AND NEW AGENTS
Carbenoxolone, the hemisuccinate of glycyrrhetinic acid, which is the main active derivative of liquorice, has been a prototype, though nonselective, 11ß-HSD1 inhibitor (Stewart et al., 1990). Carbenoxolone enhances insulin sensitivity, as measured by a euglycaemic hyperinsulinaemic clamp (Walker et al., 1995), an effect attributed to enhanced hepatic insulin sensitivity. Stable isotope glucose tracers have been used to demonstrate that, in lean patients with type 2 diabetes, carbenoxolone inhibits hepatic glucose production (Andrews et al., 2003). However, in contrast with enhanced adipose insulin sensitivity and glucose disposal in 11ß-HSD1 KO mice, carbenoxolone has no effect on glucose disposal, perhaps because carbenoxolone poorly inhibits adipose 11ß-HSD1 (Livingstone and Walker, 2003). Moreover, liver 11ß-HSD1 is downregulated in obesity, so that the incremental effect of inhibition may be smaller in the obese (Stewart et al., 1999; Rask et al., 2001,2002). Whether inhibition in liver alone, and resulting enhancement of hepatic insulin sensitivity, is sufficient to confer clinically useful metabolic benefits is uncertain.
Recently, a new, potent, nonsteroidal class of selective inhibitors of 11ß-HSD1 has been described, the arylsulfonamidothiazoles (Barf et al., 2002). These agents inhibit 11ß-HSD1 and enhance insulin action in liver, decreasing PEPCK and glucose-6-phosphatase expression and lowering blood glucose concentrations in KKAY diabetic and ob/ob obese mice (Alberts et al., 2002,2003). However, euglycaemic hyperinsulinaemic clamps in mice suggested that the selective 11ß-HSD1 inhibitor did not increase peripheral glucose uptake (Alberts et al., 2003). What remains to be established for 11ß-HSD1 inhibitors is whether their effects extend beyond the liver and include benefits in adipose tissue, where the most-beneficial target is likely to be.
|
VII. Summary |
|---|
|
TOPABSTRACTI. PreambleII. Blood Glucocorticoid Levels in...III. Corticosteroid Receptors and...IV. Tissue Metabolism of GlucocorticoidsV. 11ß-HSD1 and Glucocorticoid...VI. 11ß-HSD1 as a Drug Target...VII. SummaryACKNOWLEDGEMENTSREFERENCES |
|---|
|
ACKNOWLEDGEMENTS |
|---|
|
TOPABSTRACTI. PreambleII. Blood Glucocorticoid Levels in...III. Corticosteroid Receptors and...IV. Tissue Metabolism of GlucocorticoidsV. 11ß-HSD1 and Glucocorticoid...VI. 11ß-HSD1 as a Drug Target...VII. SummaryACKNOWLEDGEMENTSREFERENCES |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTI. PreambleII. Blood Glucocorticoid Levels in...III. Corticosteroid Receptors and...IV. Tissue Metabolism of GlucocorticoidsV. 11ß-HSD1 and Glucocorticoid...VI. 11ß-HSD1 as a Drug Target...VII. SummaryACKNOWLEDGEMENTSREFERENCES |
|---|
ligands inhibit adipocyte 11ß-hydroxysteroid dehydrogenase type 1 expression and activity. J Biol Chem
276
: 12629
–12635
