To evaluate the potential roles that both receptors and enzymes play in corticosteroid regulation of intestinal function, we have determined glucocorticoid receptor (GR), mineralocorticoid receptor (MR), and 11β-hydroxysteroid dehydrogenase (11β-HSD) expression in intestinal epithelial cells. GR and MR mRNA and receptor binding were ubiquitously expressed in epithelial cells, with receptor levels higher in ileum and colon than jejunum and duodenum. RNase protection analysis showed that 11β-HSD1 was not expressed in intestinal epithelial cells, and enzyme activity studies detected no 11-reductase activity. 11β-HSD2 mRNA and protein were demonstrated in ileal and colonic epithelia; both MR and GR binding increased when enzyme activity was inhibited with carbenoxolone. Duodenal and jejunal epithelial cells showed very little 11β-HSD2 mRNA and undetectable 11β-HSD2 protein; despite minor (<7%) dehydrogenase activity in these cells, enzyme activity did not alter binding of corticosterone to either MR or GR. These findings demonstrate the ubiquitous but differential expression of MR and GR in intestinal epithelia and that 11β-HSD2 modulates corticosteroid binding to both MR and GR in ileum and proximal and distal colon but not in duodenum or jejunum.
- glucocorticoid receptor
- mineralocorticoid receptor
glucocorticoid receptors (GR) and mineralocorticoid receptors (MR) bind endogenous glucocorticoids with high affinity. Intracellular availability of corticosterone (B) and cortisol (F) for both GR and MR can be modulated by 11β-hydroxysteroid dehydrogenase 1 (11β-HSD1) and 11β-hydroxysteroid dehydrogenase 2 (11β-HSD2). In mineralocorticoid target cells, aldosterone specificity is conferred on the nonselective MR by 11β-HSD2, which converts B and F to their inactive 11-keto metabolites, 11-dehydrocorticosterone (11-DHB) and cortisone (7, 9). In addition, 11β-HSD2 inhibits B binding to GR (20,23). In tissues that express MR but not 11β-HSD2, MR bind endogenous glucocorticoids and mediate glucocorticoid effects (5). When 11β-HSD2 activity is deficient (25, 29), mineralocorticoid activity is elevated, presumably a result of endogenous glucocorticoids binding and activating MR in mineralocorticoid target tissues.
Although 11β-HSD2 acts as a dehydrogenase only for endogenous glucocorticoids, 11β-HSD1 catalyzes the reversible conversion of B to 11-DHB. Inhibition of 11β-HSD1 in GH3 cells has been shown to potentiate the activity of B mediated via GR (26, 32), although in vivo 11β-HSD1 is thought to act as a reductase and has been suggested to potentiate glucocorticoid action by increasing the local tissue concentration of endogenous glucocorticoids (13, 16, 28). In addition to 11β-HSD1 and 11β-HSD2, several other 11β-HSD isoforms have been described but not cloned. They can be distinguished from 11β-HSD1 and 11β-HSD2 based on cofactor preference, affinity for substrate, and whether they act as an oxidase, a reductase, or both (10-12). The role these enzymes play in modulating corticosteroid access to GR and MR is yet to be defined, although in rat Leydig cells a novel 11β-HSD isoform appears to modulate the GR response (10).
In the present study we have evaluated GR, MR, and 11β-HSD expression and activity in duodenal, jejunal, ileal, and colonic epithelial cells to define the role that these receptors may play in intestinal function and that 11β-HSD isoforms may play in enhancing or limiting glucocorticoid effects on rat small intestine.
MATERIALS AND METHODS
Animals, intestinal dissection, and epithelial cell isolation.
Adult male Sprague-Dawley rats weighing 180–220 g were used in all experiments. Animals were killed by decapitation, and intestinal sections were rapidly removed as follows: duodenum, 5 cm distal to the pylorus; jejunum, 20 cm distal from the duodenum; ileum, 20 cm proximal to the cecum; proximal colon, half the colon distal to the cecum; and distal colon, half the colon proximal to the rectum. Cells were prepared from intestinal sections by a nonenzymatic technique (31). Briefly, intestinal sections were washed with ice-cold PBS and then incubated at 22°C (room temperature) in PBS containing 3 mM EDTA and 0.5 mM dithiothreitol for 90 min. Cells were removed by vigorous shaking, pelleted by centrifugation (40g for 5 min at 4°C), and then resuspended in DMEM containing 25 mM HEPES (DMEM-HEPES). Previous studies have demonstrated that this method results in the isolation of viable intact cells essentially free of stroma (30).
RNA isolation and recombinant clones.
Total RNA was prepared from intestinal cells and tissues by the guanidinium isothiocyanate method as previously described (4), and RNA concentration was determined by absorbance at 260 nm. All RNA probes were synthesized with the Promega transcription system (Promega) to a specific activity of ∼0.3 × 109 cpm/μg. The cDNA templates used to generate32P-labeled riboprobes were as follows: for GR, a 294-bp XbaI-BamH I rat GR cDNA fragment corresponding to the 3′ untranslated region; for MR, a 188-bpEcoR I-Stu I rat MR cDNA fragment corresponding to nucleotides 3130–3318; for 11β-HSD1, a 466-bpEcoR I-EcoR V fragment corresponding to nucleotides 1–466; and for 11β-HSD2, a 628-bp fragment corresponding to nucleotides 924–1552.
Solution hybridization/RNase protection assay.
A solution hybridization/RNase protection assay was used to quantify mRNA levels of MR, GR, and 11β-HSD enzymes. Total RNA (10–20 μg) was speed-vac dried and reconstituted in 25 μl of hybridization buffer (80% formamide, 40 mM PIPES, pH 6.7, 0.4 mM NaCl, and 1 mM EDTA) containing 5 μl of32P-labeled antisense RNA probe. Samples were hybridized overnight at 45°C, followed by digestion for 45 min at 37°C with 300 μl RNase buffer (300 mM NaCl, 10 mM Tris, pH 7.5, and 5 mM EDTA) containing RNase T1 (400 units) for MR, GR, and 11β-HSD2 mRNA measurements and both RNase A (40 μg/ml) and RNase T1 (400 units) for 11β-HSD1. Protected hybrids were then purified by proteinase K (10 mg/ml) digestion (37°C for 15 min) in the presence of 1% SDS, followed by isopropanol precipitation. Protected hybrids were reconstituted in diethyl pyrocarbonate-treated H2O and separated on 5% nondenaturing polyacrylamide gels.32P-labeled hybrids were visualized and quantified on a Fujix Bio-Imaging Analyzer (BAS1000 with Mac BAS; Fuji). Because GR and MR riboprobes were uniformly labeled with [32P]UTP but differed in the number of bases protected (294 for GR and 188 for MR), specific activity in molar terms (cpm/mol of riboprobe) differed between GR and MR. To allow assessment of relative expression of MR and GR mRNA, protected hybrids were corrected for specific activity (cpm/mol).
Cytosol steroid receptor binding assay.
Cells resuspended in TMD buffer (10 mM Tris, 100 mM NaMoO4, and 1 mM dithiothreitol, pH 7.4) were homogenized by hand (glass-Teflon, 4°C), and the homogenate was centrifuged (105,000 gfor 40 min at 4°C) to yield cytosol. Cytosols (250 μl) were incubated (22°C for 90 min) with 150 μl of TMD buffer containing 25–30 nM of [3H]B with or without 6 μM RU-38486 or 6 μM aldosterone. Bound and free steroid were separated on dextran-coated charcoal as previously described (21). A sample of the cytosol was taken for protein determination by the Bradford method (2). Binding displaced by 6 μM RU-38486 was taken as specific GR binding, and binding displaced by aldosterone in the presence of RU-38486 was taken as specific MR binding.
Steroids and thin layer chromatography.
Radioactive steroids were purchased from Amersham (Buckinghamshire, UK) and nonradioactive steroids (B and 11-DHB) from Sigma (St. Louis, MO), or were a gift from Roussel-Uclaf, Romainville, France (RU-38486). Carbenoxolone (CBX) was from Sigma. Ethyl acetate-extracted samples spiked with nonradioactive B and 11-DHB were separated by TLC on silica gel 60 F254 plates (Merck, Darmstadt, Germany) with 92% chloroform and 8% ethanol as the mobile phase. Silica gel plates containing separated radioactive steroids were exposed to a BAS-TR2040S imaging screen (Fuji) for up to 5 days.3H-labeled steroids were then visualized and quantified by phosphorimage analysis. [3H]B and [3H]11-DHB were identified by the comigration of UV-visualized nonradioactive B and 11-DHB.
Cells were added to pregassed (5% CO2- 95% O2) tubes containing [3H]B (25 nM) or [3H]11-DHB (20 nM) in DMEM-HEPES. After 90-min incubation at 22°C, medium was removed, and steroids were extracted with ethyl acetate and separated by TLC. Chopped rat liver (∼0.5-cm cube) was used as a positive control for 11-reductase activity.
Whole cell steroid receptor binding assay.
Intestinal cells were added to pregassed (5% CO2-95% O2) glass tubes containing 30–35 nM [3H]B with or without 6 μM nonradioactive RU-38486, aldosterone, or CBX. Tubes were then covered in Parafilm and incubated at 22°C for 90 min. After incubation, a sample of medium was extracted with ethyl acetate, and the cells were washed three times with 3 ml of ice-cold DMEM, followed by centrifugation (200g for 5 min at 4°C). Washed cells were resuspended in TMD buffer and homogenized, and the homogenate was centrifuged (105,000 g for 60 min at 4°C) to yield cytosol. Protein-bound steroid was then separated from free steroid with dextran-coated charcoal (21), and a sample of cytosol was taken for protein determination (2). Binding displaced by 6 μM RU-38486 was taken as specific GR binding, and binding displaced by aldosterone in the presence of RU-38486 was taken as specific MR binding.
Western blot analysis.
Total tissue homogenates were prepared from 1 g of frozen rat tissue or cells by homogenization in 4 vol of homogenizing buffer (0.25 M sucrose, 10 mM sodium phosphate, and 1 mM phenylmethylsulfonyl fluoride, pH 7.4). Protein concentration was determined by the Bradford method (2). Proteins (50–100 μg) were separated by 5–15% SDS-PAGE gradient gel electrophoresis under reducing conditions and then were transferred to nitrocellulose filters (Schleicher & Schuell) for 2 h at 4°C. After nonspecific sites were blocked with 5% skim milk powder in PBS (pH 7.4) containing 0.1% Tween 20 (PBS-T), the nitrocellulose blot was incubated overnight at 4°C with immunopurified rabbit anti-rat 11β-HSD2 (RAH23) polyclonal antibody (24) at a concentration of 1 μg/ml in the presence of 0.5% skim milk powder in PBS-T. The filter was washed with PBS-T and then incubated at room temperature for 60 min with a 1:5,000 dilution of goat anti-rabbit IgG antibody conjugated to horseradish peroxidase. The blots were washed in PBS-T for 60 min before the bands were visualized with a chemiluminescence kit (DuPont NEN, Boston, MA).
Data were compared by one-way ANOVA followed by Fisher’s protected least significant differences test. Differences ofP < 0.05 were considered significant. All data are expressed as means ± SE.
Expression of MR and GR mRNA in intestinal cells.
MR and GR mRNA were measured in epithelial cells isolated from the duodenum, jejunum, ileum, proximal colon, and distal colon. Kidney RNA was used as a positive control for both GR and MR mRNA, and yeast total RNA was used as a negative control. Figure1 A shows a typical phosphorimage after simultaneous solution hydridization/RNase protection analysis of GR and MR mRNA. GR mRNA expression was ubiquitous with no significant differences in levels of expression between different segments of intestine (Fig.1 B). Similarly, MR mRNA was expressed in all intestinal segments but with significantly (P < 0.05) greater levels in cells from ileum and colon than jejunum and duodenum; levels in colon epithelial cells in turn were significantly higher than those in ileal cells (Fig. 1 C). In Fig. 1,B andC, the relative expression of GR and MR was assessed by correcting values for riboprobe-specific activity (cpm/mol). As illustrated in Fig. 1 D, MR mRNA expression was higher than GR mRNA in epithelial cells isolated from all segments of the intestine.
GR and MR binding in cytosol.
To confirm the presence of GR and MR protein in intestinal epithelial cells, we performed cytosol binding assays on epithelial cell extracts. As shown in Fig. 2, specific binding of [3H]B to GR and MR could be measured in epithelial cells from the different intestinal segments. [3H]B binding to GR was significantly greater in ileum and proximal colon than in the other intestinal segments, reflecting the trend in GR mRNA levels. MR binding correlates directly with the MR mRNA concentrations, with MR binding significantly greater in ileal and colonic epithelial cells than in cells from jejunum and duodenum. MR binding in colonic cells was similarly significantly greater than in ileal cells. In contrast to the MR-to-GR mRNA ratio, the MR-to-GR binding ratio was <1 for all intestinal segments.
Expression of 11β-HSD in intestinal cells.
To determine whether 11β-HSD was expressed in intestinal epithelial cells we measured both 11β-HSD1 and 11β-HSD2 mRNA in cells isolated from the various intestinal segments. Figure3 A shows a phosphorimage after solution hydridization/RNase protection analysis of rat 11β-HSD1 mRNA. A faint band running at a higher position than 11β-HSD1 was seen in all samples, including the negative control (yeast total RNA) and probably represents a small quantity of undigested single-stranded riboprobe. There was no evidence for 11β-HSD1 mRNA expression in any of the gut epithelial cells. In contrast, liver and the medulla/cortex of the kidney showed high levels of expression, indicating that 11β-HSD1 mRNA could be detected in tissue known to express this isoform. As previously shown (15), multiple mRNA species are detected in kidney; the two major protected bands in kidney cortex/medulla correspond to the predicted full-length 11β-HSD1 and the smaller band to a truncated nonfunctional form of the enzyme. In renal papilla the major mRNA species is clearly smaller than full-length 11β-HSD1.
In contrast to 11β-HSD1 mRNA expression, 11β-HSD2 mRNA was expressed in cells derived from all regions of the intestine (Fig.3 B), with expression being 10 times higher in colon than ileum, and in the ileum threefold higher than in duodenum and jejunum (Fig. 3 C). To confirm 11β-HSD2 mRNA translation, whole cell homogenates were probed for 11β-HSD2 protein by Western blot with antibodies to rat 11β-HSD2. As illustrated in Fig. 4, 11β-HSD2-transfected Chinese hamster ovary cells transfected with the polyoma large T antigen and rat kidney demonstrated a single immunoreactive species of ∼40 kDa, corresponding to the predicted size of rat 11β-HSD2. In intestinal cell extracts, proximal colon, distal colon, and ileum were shown to contain an identical 40-kDa 11β-HSD2 immunoreactive species. A second immunoreactive species of ∼38 kDa was coexpressed in ileum and exclusively expressed in duodenum and jejunum but not present in colon.
To test whether the 38-kDa band detected by the 11β-HSD2 antibody was an active 11β-HSD enzyme or whether other 11β-HSD isoforms are present, we incubated cells with either [3H]B or [3H]11-DHB and determined whether these substrates were metabolized. Figure5 A is a phosphorimage of TLC-separated3H-labeled steroids extracted from the media after cells were incubated for 90 min at room temperature with [3H]B. There was very little conversion (<7%) of [3H]B to [3H]11-DHB in duodenal and jejunal cells. In contrast, ileal, proximal colonic, and distal colonic cells converted [3H]B to [3H]11-DHB with values of 16 ± 1%, 63 ± 5% and 59 ± 5%, respectively. The apparently equivalent levels of the immunoreactive 38-kDa protein in the different small intestinal segments (Fig. 4) and the difference in 11β-HSD activity (Fig. 5) suggest that the 38-kDa protein is not an active 11β-dehydrogenase. In addition, when cells were incubated with [3H]11-DHB there was no conversion to [3H]B, demonstrating that the 38-kDa immunoreactive protein was not an active 11-reductase (data not shown).
GR and MR binding in whole cells.
Because 11β-HSD activity is absent from cytosol extracts (20), we performed binding assays on whole cells in the presence or absence of CBX to determine whether 11β-HSD regulates B access to GR and/or MR. In accordance with the low levels of 11β-HSD activity in duodenum and jejunum, [3H]B binding to MR and GR was unaltered in the presence of CBX (data not shown). [3H]B binding to GR doubled in ileum, proximal colon, and distal colon when 11β-HSD was inhibited. [3H]B binding to MR in the presence of CBX increased approximately threefold in ileum and distal colon and approximately fourfold in proximal colon (Fig. 6). In media taken at the end of incubation, 44% of 30 nM of [3H]B was converted to [3H]11-DHB in ileum and 85% of 35 nM [3H]B in proximal and distal colon.
11β-HSD isoforms are known to regulate the response to endogenous glucocorticoids by modulating the intracellular concentrations of B and F. 11β-HSD2 acts as only a dehydrogenase for B and F; given its colocalization with MR in sodium-transporting epithelia and the increase in sodium retention when enzyme activity is compromised (27,29), 11β-HSD2 is thought to confer aldosterone specificity on MR (7,9, 17). 11β-HSD1, on the other hand, can act as both a dehydrogenase and reductase and has been suggested as possibly modulating glucocorticoid responses by either increasing or decreasing local tissue concentrations of endogenous glucocorticoids (13, 14, 16, 28). In the present study, we examined whether MR, GR, and 11β-HSD isoforms were present in intestinal epithelial cells and whether 11β-HSD isoforms modulated endogenous glucocorticoid access to these receptors.
In duodenum and jejunum there is good evidence that glucocorticoids regulate several aspects of electrolyte transport, whereas mineralocorticoids have minor or no effects (19). In agreement with these functional studies is the relatively high expression of GR and low expression of MR in duodenal and jejunal epithelial cells. The cellular response to endogenous corticosteroids is dependent on many factors, two of which are receptor concentration (1) and the presence of 11β-HSD (6, 8). The low level of MR in jejunal and duodenal cells suggests that these receptors may not be capable of inducing a major response. Furthermore, the minimal 11β-HSD2 activity plus the observation that B access to MR or GR was not altered by CBX suggests that, in vivo, B rather than aldosterone binds MR. Thus the MR in duodenum and jejunum may resemble hippocampal MR that bind B in vivo (22) and thus mediate physiological glucocorticoid actions (5).
Significant expression of MR, GR, and 11β-HSD2 is present in epithelial cells from ileum, proximal colon, and distal colon. When 11β-HSD activity is inhibited, an increase in B binding to both MR and GR is observed. These data suggest that, in vivo, MR in these intestinal segments bind aldosterone and that B binding to GR is also modulated by enzyme activity. This is supported by functional studies in which both mineralocorticoids and glucocorticoids have been shown to regulate electrolyte transport in both ileum and colon (19). When 11β-HSD activity was inhibited in ileum and colon, B binding to MR increased three- to fourfold, whereas, in contrast, binding to GR only doubled, despite a threefold increase in [3H]B in media (∼10 nM to 30 nM) for ileal cells and a sevenfold increase in [3H]B in media (∼5 nM to 35 nM) for proximal and distal colonic cells. Given that MR in the intestine and elsewhere have substantially higher affinity for B than do GR, the consistently enhanced occupancy of MR compared with GR when 11β-HSD2 is inhibited cannot be explained simply on the basis of increased [3H]B concentrations. A possible interpretation of these data is that 11β-HSD is in much closer association with MR than with GR. Whether this is because of some epithelial cells expressing GR only or to an intimate intracellular association of 11β-HSD2 with MR but not GR, so that the local microconcentration of [3H]B is lower for MR than for GR, requires further investigation.
In all intestinal segments expression of MR mRNA is greater than that of GR mRNA; however, GR binding was greater than MR binding. The discrepancy between GR and MR mRNA and binding levels suggests that RNA stability, protein stability, and/or translational efficiency may differ in intestinal cells. Recently, a splice variant of MR mRNA was described in rat colon that resulted in truncation of the steroid-binding domain (33). Although receptor binding was not performed on this truncated MR, aldosterone and B failed to increase transactivation of a luciferase reporter system, suggesting that neither steroid binds to the truncated MR. The riboprobe used in the present study would measure both of these MR mRNA species, so that an explanation for the discrepancy in the MR-to-GR mRNA and binding ratios may reflect expression of truncated MR in intestinal epithelial cells. Two GR isoforms have also been described, GR-α and a nonhormone binding form, GR-β. The GR binding assay only measures GR-α, whereas the GR riboprobe used in the RNase protection assay is complementary to a common region of both GR isoforms and thus would detect both GR mRNA species. Previous studies have shown that GR-β is expressed predominantly in epithelial cells (18), suggesting that this isoform may be present in intestinal epithelial cells. Although the presence of the GR-β would add to the discrepancy between receptor mRNA and binding levels, it may explain the discrepancy between the relatively uniform expression of GR mRNA levels and the variation between tissues in GR binding.
In addition to 11β-HSD1 and 11β-HSD2, other 11β-HSD isoforms have been described but not cloned. They can be distinguished from 11β-HSD1 and 11β-HSD2 based on cofactor preference, affinity for substrate, and whether they act as an oxidase, a reductase, or both (3,10-12). Western blot analysis with an antibody (RAH23) directed against the last 16 amino acids in the nonconserved COOH terminus of the cloned rat 11β-HSD2 protein (24) detected the 40-kDa 11β-HSD2 protein in ileum, colon, and kidney. In addition, this antibody detected a protein of ∼38 kDa in small intestinal epithelia but not colonic epithelia or kidney. The insignificant amount of 11β-dehydrogenase activity and the absence of 11-reductase activity in duodenal and jejunal epithelia argue against the presence of a novel 11β-HSD in these cells. If the immunoreactive 38-kDa protein is a novel 11β-HSD then it is either nonfunctional or inactive under our experimental conditions. Previous studies with the 11β-HSD2 antibody RAH23 in Western blot analysis of various rat tissues have shown multiple bands at 30 kDa, which correlated with levels of the 40-kDa 11β-HSD2 protein and therefore probably represent NH2-terminal degradation of 11β-HSD2 (24). In the present study the 38-kDa protein was present in the absence of detectable 11β-HSD2 and was consistently seen, so that it is unlikely to be a degradation product of the enzyme.
In summary, we have demonstrated the ubiquitous but differential expression of GR and MR in epithelial cells along the intestinal tract. In duodenum and jejunum epithelia there is very little 11β-dehydrogenase and no 11-reductase activity, suggesting that 11β-HSD isoforms do not modulate corticosteroid responsiveness in these cells. In addition, MR expression is low, supporting functional studies in which mineralocorticoids fail to regulate water and electrolyte transport in these intestinal segments (19). In contrast, ileum and colon epithelial cells express high levels of 11β-dehydrogenase activity that limits intracellular B availability to both GR and MR and would confer aldosterone specificity on MR in vivo.
We thank Rebecca Ridings for technical assistance and Professor John Funder for constructive criticism.
Address for reprint requests and other correspondence: K. E. Sheppard, Baker Medical Research Institute, PO Box 6492, Melbourne, Victoria 8008, Australia (E-mail:).
This work was supported by a block grant to the Baker Medical Research Institute from the National Health and Medical Research Council of Australia.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.
- Copyright © 1999 the American Physiological Society