| HOME | HELP | CONTACT US | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
STARLING REVIEW |  |
Endocrinology, Division of Medical Sciences, Institute of Biomedical Research, Medical School, University of Birmingham, Edgbaston, Birmingham B15 2TT, UK
(Requests for offprints should be addressed to P M Stewart; p.m.stewart{at}bham.ac.uk)
 |
Abstract |
|---|
 Top Abstract Background11ß-Hydroxysteroid...11ß-Hydroxysteroid...11ß-HSD1, human...Polycystic ovary syndromeConclusionsReferences |
|---|
Conversely, 11ß-HSD1 has been linked to human obesity and insulin resistance, but also to other diseases in which glucocorticoids have historically been implicated (osteoporosis, glaucoma). Here, the activation of cortisol from cortisone facilitates glucocorticoid hormone action at an autocrine level. The molecular basis for the putative human 11ß-HSD1 ‘knockout’ – ‘cortisone reductase deficiency’ - has recently been described, an observation that also answers a long standing conundrum relating to the set-point of 11ß-HSD1 activity. In each case, these clinical studies have been underpinned by studies in vitro and the manipulation of enzyme expression in vivo using recombinant mouse models.
|
Background |
|---|
|
TopAbstractBackground11ß-Hydroxysteroid...11ß-Hydroxysteroid...11ß-HSD1, human...Polycystic ovary syndromeConclusionsReferences |
|---|
However, studies by Cope and Black showed that the biological activity of any glucocorticoid relates, in part, to the presence of a hydroxyl group at position C11 of the steroid structure, and the inactivation of this group to a C-11 oxo group inactivates the steroid (Cope & Black 1958). Cortisol (Kendall’s compound F), and the principal glucocorticoid in rodents, corticosterone (compound B), are active steroids whereas cortisone and 11-dehydrocorticosterone (compound A), possessing a C11 keto group, are inactive. Therefore, unbeknown to Hench, Kendall and Reichstein, the bioactivity of cortisone was dependent upon conversion to cortisol in the liver. This was supported by studies that showed orally administered cortisone was, in part, excreted as cortisol (Sprague et al. 1950, Axelrod 1953, Burton et al. 1953).
Subsequently, cortisol was characterized as the active ligand and shortly thereafter the first description of tissues converting cortisol to cortisone was published. Amelung and colleagues (1953) discovered that the enzymatic inter-conversion of cortisol in man, and corticosterone in the rodent was performed by the 11ß-HSD enzyme, shown diagrammatically in Fig. 1
. 11ß-HSD was designated the number EC1.1.1.146 by the Nomenclature Committee of the International Union of Biochemistry. 11ß-Dehydrogenase activity was subsequently observed in other tissues such as placenta (Osinski 1960, Bernal et al. 1980) and kidney (Jenkins 1966). This was surprising, given that Bush et al.(1968) described a liver enzyme in vivo which was capable of reductase activity, converting cortisone to cortisol. Isotopic studies (Hellman et al. 1971) and clinical studies measuring cortisol/cortisone (F/E) levels in patients with renal disease (Srivastava et al. 1973, Whitworth et al. 1989) confirmed that the kidney was an important site for cortisol to cortisone conversion. Selective venous catheterisation studies indicated significantly lower circulating F/E ratios in renal venous blood compared with peripheral values. In contrast, circulating F/E ratios were much higher in hepatic venous blood, confirming that the liver predominantly converts E to F (Walker et al. 1993). These data suggested the existence of either a single bi-directional enzyme or two distinct isoforms, one catalysing oxidation the other responsible for reduction (Monder & Lakshmi 1989).
|
.
|
Although both isozymes are members of the short-chain dehydrogenase superfamily, they share only 21% homology, which is largely at the sites of conserved regions observed in short chain dehydrogenase/reductase (SDR) members. The isozymes are derived from separate gene products (Tannin et al. 1991, Albiston et al. 1994).
|
11ß-Hydroxysteroid dehydrogenase type 2 (11ß-HSD2) |
|---|
|
TopAbstractBackground11ß-Hydroxysteroid...11ß-Hydroxysteroid...11ß-HSD1, human...Polycystic ovary syndromeConclusionsReferences |
|---|
11ß-HSD2 localisation
The tissue distribution of 11ß-HSD2 has been ascertained by RNA blot hybridisation in human tissues, and the enzyme is expressed in high amounts in adult mineralo-corticoid target tissues (Edwards et al. 1988), such as kidney (Agarwal et al. 1994, Albiston et al. 1994, Stewart et al. 1994) and colon (Whorwood et al. 1994). Immunohistochemistry (Krozowski et al. 1995) and in situ hybridisation studies (Whorwood et al. 1995) localised 11ß-HSD2 in the kidney to the renal collecting ducts in the cortex, most of the medulla, the connecting tubule and the distal convoluted tubule. Expression is also seen in surface epithelial cells in the colon (Whorwood et al. 1994), in salivary epithelial cells (Roland & Funder 1996) and in human skin (Kenouch et al. 1994). The expression of both 11ß-HSD isozymes, 11ß-HSD1 and 11ß-HSD2, in the kidney has been reported and shown to be species-specific, as high levels of 11ß-HSD1 mRNA have been found in the rat kidney, in proximal tubules (Agarwal et al. 1989), while in the human and sheep kidney, 11ß-HSD1 mRNA is barely detectable (Tannin et al. 1991, Yang et al. 1992).
A number of studies demonstrated NAD+-dependent dehydrogenase enzyme activity in human placenta (Brown et al. 1993, Lakshmi et al. 1993) with oxo-reductase predominating in the decidua (Lopez & Craft 1981, Baggia et al. 1990). Studies have shown that 11ß-HSD2 is expressed in the placental trophoblast (Brown et al. 1993, Krozowski et al. 1995, Stewart et al. 1995, Sun et al. 1997). The capacity for 11ß-HSD2 to inactivate glucocorticoids in the placenta is very powerful, and this is supported by the placenta being the most abundant source of the enzyme, per mg of wet weight tissue (Shams et al. 1998). 11ß-HSD2 expression increases to term in the human, baboon and rat placenta (Burton & Waddell 1994, Shams et al. 1998, Pepe et al. 1999). In contrast, in the mouse placenta expression is reduced from gestation day 13 (term=19.5 days) (Brown et al. 1996, Condon et al. 1997).
11ß-HSD2 is also documented in some malignant tissues; as expression is not seen in their non-malignant counterparts this has been implicated as part of the oncogenic process (see later).
The functional role of 11ß-HSD2 in mineralocorticoid receptor (MR)-expressing epithelia is to protect the MR from cortisol. The ligand for the MR in vivo is aldosterone but in vitro studies investigating MR binding (Krozowski & Funder 1983, Arriza et al. 1987) indicated that cortisol, corticosterone, and aldosterone all have equal binding affinities for the MR, with an affinity of around 0.5–1 nM (Arriza et al. 1988). The realisation of the importance of 11ß-HSD2 in this regard came through our understanding of an inherited form of hypertension – ‘apparent mineralocorticoid excess’.
Apparent mineralocorticoid excess (AME)
Defective peripheral conversion of cortisol to cortisone (F-E) in patients with AME, reflecting impaired activity of 11ß-HSD2, was first suggested by Ulick and co-workers in the late 1970s (Ulick et al. 1979).
AME is a monogenic cause of hypertension, with an autosomal recessive inheritance pattern. Approximately 100 cases have been reported worldwide, including several cases of affected siblings (Ulick et al. 1979, Shackleton et al. 1985, Monder et al. 1986, Shackleton & Stewart 1990). AME is characterised by low renin, low aldosterone hypertension, and hypokalaemia. Patients with AME present as children or young adults with severe hypertension and hypokalaemia, suppressed plasma aldosterone concentrations, low plasma renin activity, and extended half-life of plasma cortisol (Ulick et al. 1979, White et al. 1997). The profound hypokalaemia can cause rhabdomyolysis and nephrogenic diabetes insipidus. Other renal abnormalities including renal cysts and nephrocalcinosis have been observed (Milford 1995, Moudgil 2000). Other clinical features include intra-uterine growth restriction (IUGR), short stature, thirst, polyuria and post-natal failure to thrive. The hypertension can be treated with the blockade of mineralocorticoid receptors with the potassium sparing diuretic, Spironolactone, or by a low salt diet, thus it was initially thought that mineralocorticoid excess was causing the disorder. However, all known mineralocorticoids were found to be normal or reduced.
In the 1970s Ulick et al.(1979) suggested that the features of AME could be attributed to defective peripheral conversion of cortisol (F) to cortisone (E) due to impaired activity of 11ß-HSD. Urinary steroid metabolites from AME patients supported this hypothesis as AME was associated with abnormalities in the peripheral metabolism of cortisol, specifically an increase in the urinary excretion of the A-ring reduced cortisol metabolites (tetrahydrocortisol (THF) and allo-THF) versus those of cortisone (tetrahydrocortisone (THE)), suggesting defective 11ß-HSD2 activity (Mune et al. 1995). The excretion of 5ß-cortisol metabolites exceeds that of 5ß-cortisol metabolites, resulting in a high urinary allo-THF/THF ratio and suggesting an additional defect in 5ß-reductase activity (Shackleton et al. 1985, Monder et al. 1986). Typical THF+allo-THF:THE ratios range from 8 to greater than 70 in AME (reference range 0.7–1.1). However, the in vivo assessment of 11ß-HSD2 activity is complicated. The THF+alloTHF:THE measurement provides an index of overall 11ß-HSD activity with the principal contributors being 11ß-HSD1 in the liver and 11ß-HSD2 in the kidney. Therefore, unless the THF+alloTHF:THE is grossly abnormal it is unclear whether 11ß-HSD1 or 11ß-HSD2 is the main contributor to the ratio. The urinary free cortisol/urinary free cortisone ratio (UFF:UFE) may more accurately reflect renal 11ß-HSD2 activity. Normal individuals excrete ~twofold more UFE than UFF (reference range 0.5–0.8), reflecting 11ß-HSD2 activity in the kidney. In AME, UFE excretion is virtually undetectable (Palermo et al. 1996). However, others dispute the suggestion that the UFF:UFE ratio is more sensitive than the THF+alloTHF:THE ratio for detecting decreased 11ß-HSD2 activity (Ferrari et al. 2001).
Lack of 11ß-HSD2 in AME leads to a failure of the ‘protective’ mechanism preventing illicit MR activation and results in cortisol acting as a potent mineralocorticoid (Stewart et al. 1988). However, AME patients do not present with Cushingoid features. In the face of reduced metabolism, the presence of a normal negative feedback system causes a decrease in cortisol secretion rates, thus normal circulating cortisol levels are maintained (Stewart et al. 1988).
Therapeutic treatment of AME
The main aim of treating patients with AME should be to correct life-threatening hypokalaemia and to control blood pressure. Therapeutically, patients have been treated using a wide range of drugs, including potassium supplementation and the MR antagonist Spironolactone. Spironolactone has been of variable benefit, due to high doses being required to block the MR from cortisol activation. Triamterene and Amiloride, potassium sparing diuretics which act by blocking the tubular sodium channels, have also had some success in the treatment of AME. Dexa-methasone has had variable success, and works by suppression of cortisol secretion (dexamethasone itself is not metabolised to any great degree by 11ß-HSD2). Often, additional conventional anti-hypertension treatment is also required, such as angiotensin converting enzyme (ACE) inhibition in order to completely control blood pressure. Interestingly, Captopril (an ACE inhibitor) has been shown to be capable of enhancing renal 11ß-HSD2 activity and it is has been theorised that it could be useful in treating AME patients with partial enzyme activity (Riddle & McDaniel 1994). Finally, a ‘cure’ for AME has been reported in a patient following kidney transplantation presumably because of normal 11ß-HSD2 activity in the transplanted kidney (Palermo et al. 1998).
Molecular basis for AME
At the time of its description, it was thought that mutations in the gene encoding 11ß-HSD1 might be responsible for AME. However, no mutations were identified in the four unrelated AME patients whose gene was first sequenced (White et al. 1994). With the discovery of the 11ß-HSD2 isozyme and the elucidation of its gene structure, the focus for determining the causative mutation for AME moved to 11ß-HSD2. The HSD11B2 gene is 6.2 kb in length containing 5 exons and is located on chromosome 16q22 (Agarwal et al. 1995b). The HSD11B2 gene from AME patients was screened for mutations by sequence analysis, and mutations were initially found to be present in exons 3, 4 and 5 in all the patients studied (Mune et al. 1995). Subsequently, mutations have been identified throughout the gene, although exons 3–4 remain a hotspot for mutations. Mutations on both alleles were found in all cases and thus AME is an autosomal recessive disorder. At present, over 33 different mutations have been defined within HSD11B2 in approximately 60 kindreds; a comprehensive list of the mutations responsible for causing AME is shown in Table 2.
|
Recently, nine patients with AME from four families from Oman have been reported. Five mutations, four novel and the previously reported L114Ä6nt (Odermatt et al. 2001) were identified and, for the first time, mutations were identified within exon 1 of HSD11B2 (Quinkler et al. 2004). With the addition of this study, mutations have been identified in all exons and two introns of the HSD11B2 gene.
AME is most commonly found in consanguineous families (Mune et al. 1995, Wilson et al. 1995a,b, Stewart et al. 1996, White et al. 1997). Homozygosity in AME is thought to result from endogamy or a founder effect in Native American families with the R208C and E356Ä1nt mutations, and the L250S, L251P mutation (Mune et al. 1995, Wilson et al. 1995b). The fact that six kindreds are of Native American origin has prompted speculation as to a possible selective advantage of heterozygotes. Such individuals may have an increased ability to conserve salt under conditions of extreme sodium deprivation (White et al. 1997). The Oman population comprises 2.2 million people with a majority (73%) of Omanis with tribal customs consisting of a strong preference for inter-family marriages. In 1995, Wilson et al. (1995a) described a family from Oman with two boys having AME syndrome, the parents and three other siblings being heterozygous for a R208C mutation. The more recent Oman cases derive from four different tribes; however the parents of one kindred and the grandparents of another kindred were second degree relatives. This suggests that the homozygosity of those mutations may be explained by consanguinity in the tribal society. Therefore, the diagnosis of AME should be suspected in patients from tribal societies who present with the features of low birth weight, failure to thrive, polyuria and polydipsia, and hypokalaemia and hypertension (Quinkler et al. 2004). These data suggest that AME may be a common cause of hypertension in certain ethnic groups, such as the Oman population.
AME has generated interest in HSD11B2 as a possible candidate gene for analysis in populations with essential hypertension, more specifically those with low renin, low aldosterone hypertension. Abnormalities in 11ß-HSD2 activity have been shown in patients with essential hypertension either by increases in the plasma [11-3H]cortisol half life or the THF+alloTHF:THE ratio (Walker et al. 1993, Soro et al. 1995). Genetic association studies using polymorphic markers within HSD11B2 have been carried out to determine linkage between 11ß-HSD2 and hypertension. An Italian population was genotyped for a CA repeat polymorphism (11 alleles) in the first intron of HSD11B2, and an association between salt sensitivity (observed in AME hypertension) and shorter CA repeat length was observed, suggesting decreased HSD11B2 expression with shorter CA repeat length. However, functional analysis has failed to support these data (Agarwal et al. 2000). A further study by our group has shown weak associations between the short alleles of the microsatellite within the HSD11B2 gene, and type 1 diabetes mellitus and nephropathy - salt sensitivity and hypertension being phenotypes characteristic of diabetic nephropathy (Lavery et al. 2002). However, linkage and/or association studies to date have mainly been negative, probably because of underpowered studies; a direct genetic link remains to be defined (Brand et al. 1998, Smolenicka et al. 1998).
11ß-HSD2 knockout (KO) mouse
To further assess the role of 11ß-HSD2 in vivo, mice have been generated with targeted disruption of the 11ß-HSD2 gene. The homozygous mutation was generated by removing the genomic fragment encompassing exons 2–5 through specific recombination in mouse embryonic stem cells (Kotelevtsev et al. 1999). The mice appear normal at birth, but around 50% show motor weakness and die within 48 hours of birth. Female and male survivors were fertile and exhibited severe hypertension, hypokalaemia and polyuria, thus exhibiting a typical human AME syndrome phenotype. The epithelium of the distal tubule of the nephron displayed hypertrophy and hyperplasia. Interestingly, the heterozygote mice appeared to be like the wild-type littermates for all parameters studied, and to date no phenotype has been described. The 11ß-HSD2 KO mice will be a useful model to evaluate the role of cortisol in hypertension.
Malignancy and tumorigenesis
It has been well documented that glucocorticoids have an antiproliferative action by inducing cell cycle arrest at the G1 phase (Sanchez et al. 1993, Rogatsky et al. 1997). 11ß-HSD2 expression has now been described in a number of tumours including adrenal adenomas and carcinomas (Coulter et al. 1999), lung adenocarcinomas (Suzuki et al. 2000) and adrenocorticotrophin (ACTH)-producing small cell lung cancer (Parks et al. 1998). Recent studies of bone and the pituitary have shown that both 11ß-HSD isozymes may be expressed at these sites. Specifically, our group has described a shift from predominant 11ß-HSD1 expression in normal tissue to 11ß-HSD2 in tumours (Bland et al. 1999, Cooper et al. 2000, Eyre et al. 2001, Korbonits et al. 2001). In each case the presence of 11ß-HSD2 in neoplastic cells was associated with GR rather than MR expression, suggesting an alternative function for the isozyme that is distinct from its classical role in MR-rich tissues, such as the colon or kidney. In view of the fundamental transregulatory role of the GR and its potential impact on cell proliferation and differentiation, we have hypothesised that abnormal expression of 11ß-HSD2 in GR-rich tissues will confer growth advantage and may be an important component of tumour initiation and/or progression.
|
11ß-Hydroxysteroid dehydrogenase type 1 (11ß-HSD1) |
|---|
|
TopAbstractBackground11ß-Hydroxysteroid...11ß-Hydroxysteroid...11ß-HSD1, human...Polycystic ovary syndromeConclusionsReferences |
|---|
11ß-HSD1 is a low affinity NADP(H)-dependent bi-directional enzyme, capable of carrying out both 11-oxo-reductase and dehydrogenase reactions, interconvert-ing inactive cortisone and active cortisol (Fig. 2). The enzyme is microsomal (Ozols 1995) and NADP dependent; in a cell-free system it behaves mainly as a dehydrogenase and no reductase activity is detected in the purified preparation, as described by Ozols (1995). Homogeneous enzyme gave rectilinear Eadie plots and Km constants of 1.83 ± 0.06 µM for corticosterone and 17.3 ± 2.24 µM for cortisol. First order rate constants were an order of magnitude higher for corticosterone than cortisol, but maximal velocities were similar (Lakshmi & Monder 1988).
|
Set-point of 11ß-HSD1 activity
In original purification studies, liver 11ß-HSD1 was shown to be bi-directional although, in contrast with its dehydrogenase activity, the reductase activity was unstable in vitro (Lakshmi & Monder 1988). More recently, a series of studies suggests that the enzyme prefers the reductase direction unless cells are disrupted. This applies in primary cultures of cells from liver (Jamieson et al. 1995), adipose tissue (Bujalska et al. 1997), lung (Hundertmark et al. 1995), hippocampus (Rajan et al. 1996), and vascular smooth muscle (Brem et al. 1995). In a few studies, for example in Leydig cells, 11ß-dehydrogenase activity has been reported in apparently intact cell preparations (Phillips et al. 1989), but others have found predominant 11ß-reduction (Leckie et al. 1998) and argued that some 11ß-HSD1 must be liberated from damaged cells to detect 11ß-dehydrogenase activity. This striking change in directionality between intact cells and homogenates has never been satisfactorily explained, but is likely to reflect the specific intracellular localization of 11ß-HSD1 within the lumen of the endoplasmic reticulum (ER), where neighbouring enzymes may be powerful generators of the reduced co-substrate NADPH. Hexose-6-phosphate dehydrogenase (H6 PDH) is the first enzyme in the microsomal version of the cytosolic based pentose phosphate pathway, and is related to glucose-6-phosphate dehydrogenase (G6 PDH). It generates NADPH through the oxidation of various hexose-6-phosphates including glucose-6-phosphate and appears to be responsible for NADPH provision for 11ß-HSD1 thus regulating its set-point of activity (see later).
The human gene encoding 11ß-HSD1 (HSD11B1)
The human HSD11B1 gene was first cloned and isolated by Tannin et al. in 1991. A rat 11ß-HSD cDNA was used to probe a human cDNA library (testis) and the corresponding human 11ß-HSD cDNA was used to determine chromosomal localisation. HSD11B1 was localised to chromosome 1 by human/hamster cell hybrid panels, and clones containing the HSD11B1 gene were isolated from a chromosome 1 genomic library. The human cDNA predicted a protein of 292 amino acids and was 77% identical at the amino acid level to the rat enzyme (Tannin et al. 1991). The human HSD11B1 gene is localised to chromosome 1 (1q32–41) and consists of 6 exons (Fig. 3). When first cloned, the gene was thought to be approximately 9 kb in size; however, subsequent studies revealed a much larger than previously recognised intron 4 of approximately 25 kb, expanding the size of the 11ß-HSD type 1 gene to approximately 30 kb.
|
Structural analysis of 11ß-HSD1
Little is known about the 11ß-HSD1 protein as its crystal structure has yet to be determined. The structure has been modelled on other SDR family members where the crystal structure has been identified.
The amino acid sequence of 11ß-HSD1 was determined by Edman degradation of 11ß-HSD1 purified from rabbit liver microsomes. The protein was determined to have 291 amino acid residues. This sequence was later shown to have 80% homology with human liver 11ß-HSD1. Residues 17–49 were shown to be highly conserved, whereas the C-terminus shows most variability across species (Ozols 1995). The topology of 11ß-HSD1 was elucidated by utilising 11ß-HSD1 constructs with attached FLAG epitopes at the N- and C-terminal regions. 11ß-HSD1 is intrinsic to the ER membrane, the N-terminus region of 11ß-HSD1 traverses the ER membrane and the bulk of 11ß-HSD1 is orientated in the lumen, whereas the N-terminus is cytoplasmic (Odermatt et al. 1999).
Co-factor binding sites for pyridine nucleotides, NADP/NAD were localised to residues 41–47 within the 11ß-HSD1 sequence (GXXXGXG). A putative active site region identified by Tyr-X-X-X-Lys sequence was identified at residue 182 of 11ß-HSD1 (Ozols 1995). This sequence has been implicated as part of the enzyme active site in some members of the short-chain dehydrogenase family (Persson et al. 1991). Mutagenesis studies of Tyr-179 and Lys-183 have shown these residues to be essential for enzymatic activity of 11ß-HSD1 (White et al. 1994). Mutation of the serine residues within the YSASK region of rat 11ß-HSD1 has indicated that these residues play an important role in determining the rate of the catalytic reaction (Obeyesekere et al. 1998).
Within the primary structure of 11ß-HSD1, the protein was found to be associated with oligosaccharides, Asn-X-Ser glycosylation motifs, at three specific sites. The Asn-X-Ser sites are at positions 123–125, 162–164, and 207–209 of the protein. The importance of N-linked glycosylation to 11ß-HSD1 activity has been controversially discussed in the literature. Ozols reported that de-glycosylation of rabbit 11ß-HSD1 resulted in no alteration in oxo-reducing ability (Ozols 1995). However, Agarwal et al.(1990) showed that glycosylation of rat 11ß-HSD1 was essential for dehydrogenase activity (<50%), but did not alter reductase activity. Also muta-genesis of the three potential N-glycosylation sites resulted in either reduced or complete abolition of enzyme activity (Agarwal et al. 1995a). This study was one of the first to show that HSDs could be regulated by post-translational modification. However, the same group also reported that the activity of de-glycosylated 11ß-HSD1 purified from human liver showed no difference to native enzyme activity in both dehydrogenase and reductase directions (Odermatt et al. 1999, Blum et al. 2000a). In support of this finding, recent data conclusively show fully active non-glycosylated 11ß-HSD1 enzyme activity generated in E. coli, with kinetic properties for both dehydrogenase and reductase activities similar to those reported in mammalian systems (Walker et al. 2001). These data suggest glycosylation is not required for correct protein folding or enzyme activity (Blum et al. 2000b). Glycosylation of 11ß-HSD1 may still play a role in preventing protein aggregation, in addition to stabilising the overall structure within the ER (Ozols 1995).
Studies of purified protein have helped to elucidate some sites of post-translational modification, and the enzyme’s orientation within the ER. 11ß-HSD1 has been difficult to purify as the enzyme has remained largely insoluble due to its glycosylation state and N-terminal region. Expression of 11ß-HSD1 constructs in E. coli, containing an N-terminal 6 xHis tag, has yielded a purified (157-fold) protein exhibiting full enzyme activity, with Kms equivalent to those seen in mammalian systems (Walker et al. 2001).
11ß-HSD1 knockout mouse
An important advance in understanding the role of 11ß-HSD1 in glucocorticoid action was the creation of the 11 ß-HSD1 knockout mouse (–/–) (Kotelevtsev et al. 1997). The HSD11B1 homozygous mutant was generated by removing the genomic fragment encompassing exons 3 to 4 through specific recombination in mouse embryonic stem cells.
The –/– mouse is incapable of converting 11-dehydrocorticosterone to active corticosterone - this con-firmed the fact that the only enzyme capable of 11ß-reduction in vivo, at least in mice, is 11ß-HSD1, refuting suggestions of a third isozyme. The mice showed no visible abnormalities, and retained a fully active 11ß-HSD2 isozyme. The mice were fertile, producing normal litters. The adrenal glands of male 11ß-HSD1 –/– mice are enlarged, suggesting adrenocortical hyperplasia. This could be explained by stimulation of the hypothalamic–pituitary–adrenal (HPA) axis to restore corticosterone levels (Holmes et al. 2001). The overcompensation of the HPA axis, leading to greater corticosterone levels than wild-type mice, may be explained by the absence of 11ß-HSD1 in the CNS, removing a level of regulation (Holmes et al. 2001).
The mouse studies highlighted the importance of hepatic 11ß-HSD1 in activating gluconeogenic enzymes and regulating hepatic glucose output. A high fat diet led to 11ß-HSD1 –/– mice that had significantly lower fasting plasma glucose levels than weight-matched litter mates. Glucocorticoids stimulate gluconeogenesis, and phosphoenolpyruvate carboxykinase (PEPCK) and glucose-6-phosphatase (G6 Pase) catalyse key steps of gluconeogenesis. Upon starvation these enzymes are induced by glucocorticoids; however, the 11ß-HSD1 –/– mice failed to show normal induction of G6 Pase and PEPCK. Fed mice accumulated liver glycogen, reflecting lack of glucocorticoid-induced glycogenolysis. Thus 11ß-HSD1 –/– mice are able to resist the hyper-glycaemia seen in obese wild-type mice (Kotelevtsev et al. 1997).
|
11ß-HSD1, human obesity, insulin resistance and metabolic syndrome |
|---|
|
TopAbstractBackground11ß-Hydroxysteroid...11ß-Hydroxysteroid...11ß-HSD1, human...Polycystic ovary syndromeConclusionsReferences |
|---|
Studies have demonstrated that primary cultures of adipose stromal cells (ASCs) isolated from patients undergoing elective abdominal surgery, express 11ß-HSD1 but not 11ß-HSD2. Activity studies show conversion of inactive cortisone to active cortisol through the expression of 11ß-HSD1 which is significantly higher in omental fat than in subcutaneous fat (Bujalska et al. 1997). Cortisol and insulin treatment were shown to increase differentiation of preadipocytes to adipocytes (Bujalska et al. 1999). Exposure to cortisol also increased expression and activity of 11ß-HSD1, providing a fast-forward feedback system for the local generation of active glucocorticoid within omental adipose tissue (Bujalska et al. 1997). In vivo, such a mechanism would ensure a constant exposure of active glucocorticoid specifically to omental adipose tissue, suggesting that central obesity may reflect ‘Cushing’s disease of the omentum’ (Bujalska et al. 1997). However, in vivo studies are less clear-cut with increased (Paulmyer-Lacroix et al. 2002, Lindsay et al. 2003), unchanged or even reduced 11ß-HSD1 activity reported in obesity. Our own data are consistent with a reduction in hepatic 11ß-HSD1 reductase activity with increasing body mass index (BMI) (Stewart et al. 1999, Rask et al. 2001). Expression in adipose tissue in obesity is more contentious with increased 11ß-HSD1 expression and/or activity again reported in subcutaneous sites. Our data in omental fat suggests no change or even a slight reduction similar to that seen in the liver (Tomlinson et al. 2002). Conversely, enzyme expression increased in obese subjects on a strict calorie-restricted diet (Tomlinson et al. 2004).
Obesity is thought to result from an expansion of adipose tissue mass by hypertrophy of existing adipocytes by lipid accumulation, differentiation of preadipocytes (adipose stromal cells (ASC)) through to adipocytes, or proliferation of preadipocytes (Yau et al. 1995, Tomlinson & Stewart 2001). Cortisol has been shown to cause differentiation of preadipocytes to adipocytes (Hauner et al. 1987), and the expression of 11ß-HSD1 in vitro is sufficient to allow differentiation of omental preadipocytes by cortisone alone (Bujalska et al. 1999). This is supported by inhibition of 11ß-HSD1 with glycerrhetinic acid which prevented cortisone-induced adipocyte differentiation. However, glucocorticoids generally inhibit cellular proliferation, and 11ß-HSDs can regulate this process (Rabbitt et al. 2002). In addition, our group has recently demonstrated inhibition of human omental preadipocyte proliferation with cortisol, modulated by 11ß-HSD1. Thus, in whole adipose tissue, 11ß-HSD1 activity would lead to increased adipocyte differentiation, but simultaneous inhibition of preadipocyte proliferation, and thus the overall impact upon adipose tissue mass needs to be evaluated (Tomlinson & Stewart 2001). Our latest data, supported by data shown in the 11ß-HSD1 KO mouse (Morton et al. 2004), suggest that reduced expression of 11ß-HSD1 in liver and fat might represent an important protective mechanism offsetting the deleterious metabolic consequences of increasing BMI (i.e. reduced hepatic glucose output, reduced adipocyte differentiation). Of interest is that a BMI-related reduction in 11ß-HSD1 activity is not observed in type 2 diabetes mellitus (Valsamakis et al. 2004). Further prospective studies are required to evaluate the consequences of this upon the development of diabetes in obese subjects.
An adipose-specific 11ß-HSD1 transgene has been used to further in vivo studies of the role of 11ß-HSD1 in visceral obesity. 11ß-HSD1 cDNA was linked to the adipocyte fatty-acid-binding protein (aP2) promoter (Masuzaki et al. 2001), which caused a sevenfold amplification of 11ß-HSD1, leading to a viscerally obese phenotype (transgene mice 16% heavier than wild-type mice after 15 weeks). Significantly, this caused a 15–30% elevation in local adipose tissue corticosterone levels but circulating levels remained normal. Adipocyte number remained the same between transgenic and wild-type mice, although exposure of adipocytes to increased glucocorticoid levels resulted in lipid accumulation and an increased adipocyte size in transgenic mice. The transgenic mice were also markedly insulin resistant and glucose intolerant (Masuzaki et al. 2001).
Understandably, these in vitro, clinical and recombinant mouse studies have been the impetus for several pharmaceutical companies to develop selective 11ß-HSD1 inhibitors (selective in that they do not inhibit the related 11ß-HSD2). Many pharmaceutical companies have already published encouraging results using such inhibitors, which are discussed below, and many more are currently investigating newly developed compounds.
Biovitrum-Amgen has shown that arylsulphonamidothiazole compounds inhibit 11ß-HSD1 both in vivo and in vitro (Alberts et al. 2002, Barf et al. 2002) and have shown encouraging results in animal studies. The diethylamide derivative was shown to inhibit human 11ß-HSD1 with an IC50 of 52 nM, and an N-methylpiperazinamide form (BVT.2733) was shown to be specific for the mouse enzyme (IC50 of 96 nM). Both compounds showed >200-fold selectivity over human and murine 11ß-HSD2 (Barf et al. 2002). In the hyperglycaemic mouse strain KKA(y), this compound lowered hepatic PEPCK and glucose-6-phosphatase mRNA, blood glucose and serum insulin concentrations, supporting data that 11ß-HSD1 is a key regulator of gluconeogenesis. This raises the possibility that inhibition of 11ß-HSD1 might be used therapeutically to treat patients with insulin resistance.
Merck has shown that insulin sensitising thiazolidinediones peroxisome proliferator-activated receptor (PPAR)-
agonists may mediate their action, in part, through inhibition of adipose 11ß-HSD1, an effect demonstrated in cultured adipocytes (Berger et al. 2001). It was shown that thiazolidinedione and nonthiazolidinedione agonists of PPAR markedly inhibit expression of 11ß-HSD1 in 3T3-L1 adipocytes. This decrease in expression correlated with a significant decrease in the cellular conversion of cortisone to active cortisol. The half-maximal inhibitory effect of the thiazolidinedione, rosiglitazone, occurred at a concentration that supported a PPAR-mediated mechanism of action. It was also demonstrated that the inhibitory action of PPAR agonists on 11ß-HSD1 mRNA expression appears to take place at the level of transcription. In addition, treatment of diabetic db/db mice with rosiglitazone inhibited expression of 11ß-HSD1 in adipose tissue. This decrease in enzyme expression correlated with a significant decline in plasma corticosterone levels (Berger et al. 2001). Furthermore, the lipid lowering agent fenofibrate (a PPAR- agonist) inhibits 11ß-HSD1 in hepatocytes (Hermanowski-Vosatka et al. 2000).
More recently, Merck has identified a new 11ß-HSD1 specific inhibitor – adamantly triazole, and have shown that chronic oral administration in a murine model system lowered triglycerides, insulin, fasting glucose and reduced body weight (Thieringer et al. 2004). This suggests pharmacological inhibition of 11ß-HSD1 improves several key features of the metabolic syndrome.
Glaucoma
11ß-HSD isozyme expression has recently been described in the human and rodent eye (Stokes et al. 2000, Rauz et al. 2001, Suzuki et al. 2001), notably 11ß-HSD1 in ciliary epithelial cells, raising the possibility for a role in aqueous humour production and the regulation of intraocular pressure. Within the aqueous humour, cortisol concentrations exceed cortisone by approximately 14-fold (Rauz et al. 2001). This is highly suggestive of predominantly 11ß-HSD1 reductase activity. Treatment with carbenoxolone in a randomised, double-blind placebo-controlled study of glaucoma sufferers significantly reduced intra-ocular pressure (Rauz et al. 2001) and topical 11ß-HSD1 inhibitors may represent a future novel therapeutic strategy in glaucoma.
Glucocorticoid-induced osteoporosis
Glucocorticoid-induced osteoporosis (GIOP) was first described as a recognised feature of Cushing’s syndrome (Cushing 1932). Since then the increasing use of pharmacological doses of corticosteroids has brought attention to GIOP (Canalis 1996). Studies in vivo clearly show that administration of pharmacological doses of glucocorticoids stimulate bone resorption and inhibit bone formation, therefore decreasing bone mass and increasing the risk of fractures, and causing profound effects on bone cell replication, differentiation and function (Canalis & Delany 2002).
Several studies have demonstrated GR expression in osteoclasts and osteoblastic cells (Chen et al. 1977); GR
and MR are present in osteoclasts and osteoblasts (Beavan et al. 2001) and it is thought that many of the effects of glucocorticoids on bone are GR mediated. Studies by our group have demonstrated the expression of 11ß-HSD isozymes in bone. Predominant 11ß-HSD1 expression was demonstrated by enzyme activity studies, RT-PCR and immunohistochemistry, in osteoblast primary cultures (Bland et al. 1999), cell lines and normal adult bone (osteoblasts and osteoclasts) (Cooper et al. 2000). It is proposed that the expression of 11ß-HSD1 in osteoblasts facilitates the local synthesis of active glucocorticoids with consequent effects on osteoblastic proliferation and differentiation. In contrast to liver and adipose tissue, where reductase activity predominates, both reductase and dehydrogenase activities of 11ß-HSD1 are evident in bone chips and primary cultures of human osteoblasts (Cooper et al. 2000).
Tissue damage by pro-inflammatory cytokines is attenuated at both systemic and cellular levels by anti-inflammatory factors such as glucocorticoids. Within the human osteosarcoma cell-line (MG-63) pro-inflammatory cytokines, interleukin (IL)-1ß and tumour necrosis factor (TNF)- potently inhibited 11ß-HSD2 activity, and reciprocally raised 11ß-HSD1 activity and expression. In addition, a rise in 11ß-HSD1 reductase activity was observed in primary cultures of osteoblasts treated with TNF- (Cooper et al. 2001). These data suggest that local tissue metabolism of glucocorticoids is likely to be important in determining the sensitivity of both osteoblasts and osteoclasts to glucocorticoids in normal bone homeostasis. In addition, variation in 11ß-HSD isozyme expression and activity may explain individual differences in susceptibility to glucocorticoid-induced osteoporosis.
Importantly, 11ß-HSD1 reductase activity in osteo-blast cultures has been shown to correlate positively with age. Glucocorticoid treatment caused a time- and dose-dependent increase in 11ß-HSD1 activity with similar increases in 11ß-HSD1 mRNA expression (Cooper et al. 2002). These studies indicate that activation of glucocorticoids at an autocrine level within bone is likely to play an important role in the age-related decrease in bone formation and increased risk of glucocorticoid-induced osteoporosis (Cooper et al. 2002).
|
Polycystic ovary syndrome |
|---|
|
TopAbstractBackground11ß-Hydroxysteroid...11ß-Hydroxysteroid...11ß-HSD1, human...Polycystic ovary syndromeConclusionsReferences |
|---|
Over 10 years ago, our group postulated that defective 11ß-HSD1 activity may be an important mechanism in patients with PCOS (Stewart et al. 1990). The resulting increase in cortisol metabolic clearance rate would lead to stimulation of the HPA axis in an attempt to normalise serum cortisol levels, and might account for the ACTH-mediated hyperandrogenism in such patients. Alterations of peripheral metabolism of adrenal steroids, specifically increased 5-reductase and 11ß-HSD isozyme activities, have been identified in women with PCOS (Stewart et al. 1990, Rodin et al. 1994, Chin et al. 2000, Moghettii et al. 2002, Fassnacht et al. 2003, Tomlinson et al. 2003). This is perhaps best exemplified through patients with the opposite of AME – ‘cortisone reductase deficiency’.
Cortisone reductase deficiency
The existence of a deficiency in the peripheral enzymatic conversion of E to F (the opposite of AME) was first described in 1984 by Taylor et al., and has since been described in eleven individuals, all but one being female (Taylor et al. 1984, Savage 1991, Nikkila et al. 1993, Phillipov et al. 1996, Jamieson et al. 1999, Nordenstrom et al. 1999). The clinical characteristics of these patients are summarised in Table 3. This rare condition named ‘cortisone reductase deficiency’ (CRD) is a form of PCOS in which the underlying defect appears to be deficient activity of the 11ß-HSD1 enzyme. Patients with CRD share many clinical characteristics associated with PCOS and have invariably presented in adolescence or early adulthood with features of hyperandrogenism (acne, hirsutism, oligo-amenorrhoea, infertility). Obesity has been a feature of some cases. Serum androgens (dehydroandrosterone sulphate (DHEAS), androstene-dione and testosterone) have been elevated in each case, but are readily suppressed following dexamethasone administration.
|
-HSD1 activity). The defect in E to F conversion results in an increased metabolic clearance rate for cortisol; through the negative feedback mechanism the HPA axis is activated and ACTH secretion is increased to maintain normal circulating cortisol concentrations, but at the expense of ACTH-mediated androgen excess. This leads to a polycystic ovarian syndrome-like phenotype in females (Biason-Lauber et al. 2000), as summarised in Fig. 4. The phenotypic consequences of CRD in males are likely to be milder, with a report of premature adrenarche/puberty in a boy aged 7 years (Biason-Lauber et al. 2000). Dexamethasone, by suppressing endogenous ACTH drive has been used therapeutically to control the hyperandrogenism.
|
Recently, we have reported an extensive genetic investigation of HSD11B1 in three CRD kindreds, each with a single affected case. The CRD cases, one of which is male, all exhibited characteristically low urinary THF+allo-THF:THE ratios of between 0.03 and 0.04 (reference range 0.7–1.1), and have been described previously in the literature (Jamieson et al. 1999, Laing et al. 2002, Malunowicz et al. 2003). In all three CRD cases, two polymorphisms in complete linkage disequilibrium within intron 3 of HSD11B1 were identified; an A insertion (83557), and 40 base pairs downstream a T to G substitution (83597) (Fig. 3A
). The allele frequency for the 83557A/83597T-G haplotype was 14% in control populations.
Although the 11ß-HSD1 cDNA sequence was normal in one case, adipose tissue mRNA levels were 28-fold lower when compared with an unaffected sister, and the corresponding oxo-reductase activity was absent (0% conversion of E to F versus 14% in the unaffected sibling). In luciferase reporter assays, there was a 2.5-fold reduction in transcriptional activity in HSD11B1 constructs containing the intron 3 83557A/83597T-G mutation compared with wild-type, suggesting that this region of the gene acts as an intronic enhancer of HSD11B1 expression. These data are in keeping with published precedents suggesting a silencer/enhancer role for intronic sequences in many genes including HSD11B2 (Agarwal 2001). The impact of heterozygosity for the 83557A/83597T-G mutation upon cortisol metabolism is unknown, but the CRD phenotype cannot be explained by heterozygosity or homozygosity at this single locus because this was present in 25% and 3% of normals respectively.
As discussed previously, 11ß-HSD1 oxo-reductase activity requires NADPH. The purified enzyme behaves as an NADP-dependent dehydrogenase and the switch to oxo-reductase activity upon tissue homogenization (Agarwal et al. 1990) suggests a close association between 11ß-HSD1 and a NADPH generation system. The G6 PDH enzyme of the pentose phosphate pathway has been considered to be the major source of intracellular NADPH (Kletzien et al. 1994), but this is a cytosolic enzyme and the membrane-binding domain of 11ß-HSD1 directs the active site of the enzyme towards the ER lumen, away from the cytosol (Ozols 1995). Here H6 PDH, an enzyme of previously uncertain significance but related to G6 PDH, can catalyze the first two steps of the pentose phosphate pathway and generate NADPH (Kimura et al. 1979, Stegeman & Klotz 1979). H6 PDH is present in most tissues but is highly expressed in liver and adipose tissue, sites of 11ß-HSD1 oxo-reductase activity.
Sequencing of the H6PD gene, localized to chromosome 1p36.2 (see Mason et al. 1999), revealed mutations within exon 5 in all three CRD cases. Case 1 was heterozygous for 620 ins29 bp621, an insert of 29 bp between residues 620 and 621 that results in the inclusion of 3 new amino acids and a stop codon, truncating the protein by 171 amino acids. Cases 2 and 3 were homozygous for R453Q, a non-conservative missense amino acid change. The position of the mutations within H6 PDH are shown in Fig. 3B. When H6PD mutant cDNAs were synthesized and expressed in hepatic WRL68 cells, the 620 ins29 bp621 mutant was devoid of H6 PDH activity and the R453Q mutant demonstrated residual activity that was consistently less than 50% of wild-type. The impact of the R453Q mutation in the context of a normal HSD11B1 genotype is unknown. The allele frequency of R453Q in both Scottish and Indo-Asian controls is 21–22%, so that ~4% of the normal population is homozygous for this mutation. The combination of homozygosity for the H6PD R453Q and heterozygosity for the HSD11B1 intron 3 83557A/ 83597T-G mutations was not observed in our control subjects, and its presence in 2/3 CRD patients is unlikely to be due to chance (P=0.0008, Fisher’s exact test). Based on our normative allelic frequency data we can predict a prevalence rate for CRD of approximately 0.1% for the mutations HSD11B1 +/– and R453Q H6PD –/–. Alternatively, interaction with an unidentified third locus, modifying penetrance, might be implicated.
Thus, a combination of mutations in the HSD11B1 and H6PD genes interact to cause CRD manifesting in a reduction in 11ß-HSD1 expression and impaired provision of NADPH to an enzyme that is critically dependent on reduced co-factor for oxo-reductase activity. This is summarised in diagrammatic form in Fig. 5. A digenic triallelic mode of inheritance is proposed, in which three distinct alleles, from two (or more) loci (HSD11B1 and H6PD), are necessary for trait manifestation (Fig. 3A and B).
|
|
Conclusions |
|---|
|
TopAbstractBackground11ß-Hydroxysteroid...11ß-Hydroxysteroid...11ß-HSD1, human...Polycystic ovary syndromeConclusionsReferences |
|---|
11ß-HSD2, being a unidirectional dehydrogenase, controls renal cortisol concentrations and, via its pre-receptor role, protects the MR from illicit activation by cortisol. Mutations in the gene encoding 11ß-HSD2 lead to the monogenic form of hypertension – AME. It has been suggested that mild mutations in HSD11B2 may account for some cases of idiopathic hypertension.
More recently, attention has focused on the 11ß-HSD1 isoform and its role in human disease states, in particular features of the metabolic syndrome such as obesity and insulin resistance, and also glucocorticoid-related diseases such as osteoporosis and glaucoma. Thus far, studies have predominantly concentrated upon the liver and adipose tissue and the role of 11ß-HSD1 in regulating hepatic gluconeogenesis and fat mass. Obviously, this makes selective inhibitors of 11ß-HSD1 an attractive proposition for the pharmaceutical industry, who are currently rigorously searching for, and developing, 11ß-HSD1 inhibitors with the aim of finding a therapeutic target for the metabolic syndrome. In addition to the clear clinical impact of these inhibitors, they will also be crucial for increasing our understanding of the role of 11ß-HSD1 in vivo. Some caution should be considered here, as these inhibitors are likely to be systemic, and not tissue specific and it will therefore be of importance to tease out the effect of inhibition within individual tissues that show a wide range of 11ß-HSD1 expression. In addition, as many inhibitors are derived from high throughput assays in intact cell systems, it is exciting to speculate that they might be inhibitors of H6 PDH rather than acting directly upon the 11ß-HSD1 enzyme.
The ‘human 11ß-HSD1 knockout’, termed cortisone reductase deficiency has enabled greater understanding of the set-point of activity of 11ß-HSD1. Mutations were found in both HSD11B1 and the gene encoding hexose-6-phosphate dehydrogenase. It is now thought that H6 PDH is responsible for conferring reductase activity upon 11ß-HSD1 by the production of cofactor (NADPH). Further studies are clearly needed to assess the broader impact of regulation of H6 PDH expression and activity as a determinant of 11ß-HSD1 reductase activity and hence autocrine cortisol generation. The next generation of experiments may be to place the last twenty years of 11ß-HSD1 research in the context of H6 PDH activity and expression.
|
References |
|---|
|
TopAbstractBackground11ß-Hydroxysteroid...11ß-Hydroxysteroid...11ß-HSD1, human...Polycystic ovary syndromeConclusionsReferences |
|---|
Agarwal AK, Monder C, Eckstein B & White PC 1989 Cloning and expression of rat cDNA encoding corticosteroid 11ß-dehydrogenase. Journal of Biological Chemistry 264 18939–18943.
Agarwal AK, Tusie-Luna MT, Monder C & White PC 1990 Expression of 11ß-hydroxysteroid dehydrogenase using recombinant Vaccinia virus. Molecular Endocrinology 4 1827–1832.[Abstract]
Agarwal AK, Mune T, Monder C & White PC 1994 NAD+-dependent isoform of 11ß-hydroxysteroid dehydrogenase. Cloning and characterization of cDNA from sheep kidney. Journal of Biological Chemistry 269 25959–25962.
Agarwal AK, Mune T, Monder C & White PC 1995a Mutations in putative glycosylation sites of rat 11ß-hydroxysteroid dehydrogenase affect enzymatic activity. Biochimica et Biophysica Acta 1248 70–74.[CrossRef][Medline]
Agarwal AK, Rogerson FM, Mune T & White PC 1995b Analysis of the human gene encoding the kidney isozyme of 11 beta-hydroxysteroid dehydrogenase. Journal of Steroid Biochemistry and Molecular Biology 55 473–479.[CrossRef][ISI][Medline]
Agarwal AK, Giacchetti G, Lavery G, Nikkila H, Palermo M, Ricketts M, McTernan C, Bianchi G, Manunta P, Strazzullo P, Mantero F, White PC & Stewart PM 2000 CA-repeat polymorphism in intron 1 of HSD11B2: effects on gene expression and salt sensitivity. Hypertension 36 187–194.
Alberts P, Engblom L, Edling N, Forsgren M, Klingstrom G, Larsson C, Ronquist-Nii Y, Ohman B & Abrahmsen L 2002 Selective inhibition of 11 beta-hydroxysteroid dehydrogenase type 1 decreases blood glucose concentrations in hyperglycaemic mice. Diabetologia 45 1528–1532.[CrossRef][ISI][Medline]
Albiston AL, Obeyesekere VR, Smith RE & Krozowski ZS 1994 Cloning and tissue distribution of the human 11ß-hydroxysteroid dehydrogenase type 2 enzyme. Molecular and Cellular Endocrinology 105 R11–R17.[CrossRef][ISI][Medline]
