The 11β-hydroxysteroid dehydrogenase (11βHSD) type 1 (11βHSD1) enzyme is an NADP+-dependent oxidoreductase, usually reductase, of major glucocorticoids. The NAD+-dependent type 2 (11βHSD2) enzyme is an oxidase that inactivates cortisol and corticosterone, conferring extrinsic specificity of the mineralocorticoid receptor for aldosterone. We reported that addition of a reducing agent to renal homogenates results in the monomerization of 11βHSD2 dimers and a significant increase in NAD+-dependent corticosterone conversion. Estrogenic effects on expression, dimerization, and activity of the kidney 11βHSD1 and -2 enzymes are described herein. Renal 11βHSD1 mRNA and protein expressions were decreased to very low levels by estradiol (E2) treatment of both intact and castrated male rats; testosterone had no effect. NADP+-dependent enzymatic activity of renal homogenates from E2-treated rats measured under nonreducing conditions was less than that of homogenates from intact animals. Addition of 10 mM DTT to aliquots from these same homogenates abrogated the difference in NADP+-dependent activity between E2-treated and control rats. In contrast, 11βHSD2 mRNA and protein expressions were significantly increased by E2 treatment. There was a marked increase in the number of juxtamedullary proximal tubules stained by the antibody against 11βHSD2 after the administration of E2. Notwithstanding, neither the total corticosterone and 11-dehydrocorticosterone excreted in the urine nor their ratio differed between E2- and vehicle-treated rats. NAD+-dependent enzymatic activity in the absence or presence of a reducing agent demonstrated that the increase in 11βHSD2 protein was not associated with an increase in in vitro activity unless the dimers were reduced to monomers.
the glucocorticoids corticosterone and cortisol circulate at 100 to 1,000 times the level of aldosterone in plasma and have similar affinities for the mineralocorticoid receptor (MR) to those of aldosterone (3, 14). Aldosterone binds the MR and modulates vectorial transport of sodium and water across epithelia in mineralocorticoid target tissues, such as the kidney and colon, where extrinsic specificity of the MR for aldosterone is provided by the enzyme 11β-hydroxysteroid dehydrogenase (11βHSD) type 2 (7, 8). This enzyme converts corticosterone and cortisol to their inactive metabolites 11-dehydrocorticosterone and cortisone, respectively, before they can interact with the MR (2, 7, 8, 31). The 11β-18-hemiacetal of aldosterone prevents it from being a substrate for the 11βHSD2.
Two 11βHSD isozymes have been cloned and characterized. 11βHSD1 is NADP+ dependent, has a high Km of 1–3 μM for corticosterone and cortisol, and is bidirectional, although in vivo and in intact cells it is primarily a reductase. It does not colocalize with the MR in the kidney (19, 30, 32). 11βHSD2 is NAD+ dependent, acts only as a dehydrogenase, has a Km for corticosterone and cortisol low enough to be relevant to circulating levels of free glucocorticoids (4–14 nM), and is colocalized with the MR in aldosterone target tissues (15, 31). 11βHSD2 has been cloned from kidney cDNA libraries of the human (2), sheep (1), rabbit (23), mouse (6), cow (29), and rat (42). In addition to classical aldosterone target cells, data from in situ hybridization studies indicate that 11βHSD2 mRNA is expressed in several other tissues, including the female rat reproductive system and the central nervous system (26–28).
The direction of 11βHSD1 activity, reduction or oxidation, in vitro is dependent on substrate concentration, cofactor, and type of preparation. Before the mid-1990s, when the existence of a second 11βHSD was confirmed (2), there was confusion about the bidirectionality of the 11βHSD enzyme in different tissues under different experimental conditions. Significant 11βHSD2 activity in the rat kidney was difficult to demonstrate until the recent report that the addition of the reducing agent dithiothreitol (DTT) to the incubation mixture of renal homogenates in the presence of NAD+ resulted in significant conversion of tritiated corticosterone to tritiated 11-dehydrocorticosterone (11). Western blots under nonreducing conditions indicated that most of the 11βHSD2 was present as an inactive dimer (11). With this in mind, we restudied the effect of estrogens on the mRNA, protein, and activity of 11βHSD1 and 11βHSD2 in vitro and in vivo by performing ribonuclease protection assays and Western blots and measuring oxidation of [3H]corticosterone by kidney homogenates with and without DTT, as well as by measuring the 24-h urinary excretion of adrenal cortical steroids. In addition, changes in expression and distribution of the 11βHSD2 were assessed by immunohistochemistry.
MATERIALS AND METHODS
Materials. Nicotinamide adenine dinucleotide (NAD+), nicotinamide adenine dinucleotide phosphate (NADP+), their reduced derivatives, DTT, mercaptoethanol, and unlabeled steroids were obtained from Sigma Chemical (St. Louis, MO). [1,2-3H]corticosterone (sp act 25 Ci/mmol) was tritiated by American Radiolabeled (St. Louis, MO). Channeled TLC plates (silica gel 60 Å) and reagent-grade solvents were purchased from Whatman (Clifton, NJ) and Fisher Scientific (Medford, MA), respectively. Ham's F-12 medium and Hanks' basic salt solution (BSS) without calcium and magnesium were purchased from the University of Missouri Cytology Core Facility.
Animals. All animal experiments were performed under protocols approved by the University of Missouri-Columbia Animal Care and Use Committee and the Harry S Truman Memorial Veterans Committee on Animal Studies. Sprague-Dawley rats from Harlan Industries (Indianapolis, IN) were housed under conventional husbandry conditions in American Association for Accreditation of Laboratory Animal Care-accredited facilities. Orchiectomies were done with animals under isoflurane anesthesia delivered in oxygen by a veterinary anesthesia machine with use of standard aseptic surgical procedures through a midline skin incision in the scrotum. The tunicae vaginalis were obliterated to prevent intestinal herniation and entrapment. There were 10 male Sprague-Dawley rats in each group for the urinary steroid excretion studies. Intact and castrated male rats received subcutaneous injections of control oil, 5 mg of estradiol valerate, or 5 mg of testosterone propionate. Injections were given 2 days after surgery, before the animals were placed in metabolism cages without urine collection funnels. After 4 days of acclimatization to the metabolism cages, the rats received a second injection of estradiol valerate, testosterone, or vehicle, and urine was collected for 48 h. On removal from the metabolism cages, the rats received a third injection and were placed in their home cages. The rats were anesthetized with isoflurane 48 h later, and blood was taken by cardiac puncture before tissue harvest. Two rats from each group were gravity-perfused under anesthesia with heparinized saline, followed by Streck's Tissue Fixative (STF; Streck laboratories, La Vista, NE), and their kidneys were used for immunocytochemistry. One kidney from each of the other rats was cut in half, and the pieces were frozen separately in liquid nitrogen and stored at -80°C for RNA and protein extraction at a later date. The other kidney was homogenized and used for enzymatic activity measurements. The mRNA and protein expression and enzyme activity studies were repeated with 6, 4, or 3 rats per group; statistical comparisons were made within each group.
Ribonuclease protection assay. Total RNA was prepared using UltraSpec RNA isolation solution from Biotecx (Houston, TX). Biotinylated RNA probes were made as follows. PCR products of 11βHSD1 (305 bp), 11βHSD2 (163 bp), and GAPDH (282 bp) were individually cloned into pCR3 plasmid, and the positive clones with the correct orientation were identified by restriction digestion and sequencing. The plasmids were linearized and transcribed with T7 RNA polymerase by use of a BrightStar BIOTINscript in vitro transcription kit from Ambion (Austin, TX) to make the cRNA probes. The RNA probes were gel purified by denaturing polyacrylamide gel. Twenty micrograms of total RNA from three individual rats of each group were hybridized with 11βHSD1 or 11βHSD2 probes, as well as with a GAPDH probe as an internal control for the amount of RNA. The ribonuclease protection assay (RPA) was performed according to the instruction manuals from Ambion (RPA II and the BrightStar BioDetect kits). The films were scanned and quantified with a Kodak Image Station 440 CF by use of Kodak ds 1D image analysis software. Enzyme mRNA densitometry values were normalized for sample loading differences by dividing them by the densitometry values of the GAPDH signals run concomitantly and were expressed as arbitrary units (au). Each lane represents RNA from a different individual rat.
Western blots for 11βHSD2 and 11βHSD1. Rat kidney microsomes were isolated by differential centrifugation, as described before (21). Microsomes were solubilized using Laemmli buffer with or without 10 mM β-mercaptoethanol (16) and run in a 12% PAGE, transferred by semidry blot to a polyvinylidene membrane, dried, blocked with 5% nonfat milk, and incubated with the 11βHSD2 antibody raised against the recombinant rat 11βHSD2 protein in sheep (11) or with a rabbit anti-11βHSD1 antibody [kindly provided by Dr. Mathew Hardy from Rockefeller University (17)]. The blots were then incubated with a peroxidase-labeled second antibody and developed using West Pico Chemiluminescence substrate from Pierce Chemical (Rockville, IL). The films were scanned and quantified with a Kodak Image Station 440 CF using Kodak ds 1D image analysis software. For the quantification of the 11βHSD2, a group of four control and estradiol-treated rat renal microsomal protein isolates were electrophoresed, blotted as above, developed using SuperSignal West Dura Extended Duration Substrate (Pierce Chemical), and quantified directly using the Kodak Image Station.
Enzyme activity assays. Kidney homogenates were prepared in ice-cold Tris buffer 0.1 M, pH 7.6, 0.25 M sucrose, and 5 mM magnesium chloride solution with short bursts of a polytron homogenizer. Microsomes were separated by differential centrifugation, as previously described (21), and then incubated at 37°C for 30 min with 0.5 mM NAD+ or NADP+, 0 or 10 mM DTT, and 10 nM [3H]corticosterone (NAD+) or 500 nM [3H]corticosterone (NADP+).
Measurements of urinary corticosterone and 11-dehydrocorticosterone. A 1-ml aliquot of urine mixed with [3H]corticosterone (∼3,000 dpm) was extracted with 25% dichloromethane in hexane, washed with water, evaporated under air, and reconstituted in ELISA buffer. An aliquot was used for measurement of recovery. ELISAs for corticosterone and 11-dehydrocorticosterone were performed using antibodies raised in sheep against corticosterone- and 11-dehydrocorticosterone-3-carboxymethoxylamine-chicken serum albumin conjugates and methods developed by us as previously described (9, 20). The anti-corticosterone antibody has a cross-reactivity of 0.83% with 11-dehydrocorticosterone, 2.7% with cortisol, 12.5% with deoxycorticosterone, 2.5% with progesterone, and 0.083% with 18-OH-deoxycorticosterone. The antibody against 11-dehydrocorticosterone has a cross-reactivity of 0.81% with corticosterone, <0.09% with cortisol, 0.75% with deoxycorticosterone, 0.11% with cortisone, and 0.1% with progesterone. Two consecutive 24-h urine collections were pooled after an acclimatization period to achieve an average 24-h excretion over 2 days.
Immunohistochemistry. The kidneys from perfused animals were fixed overnight at room temperature in STF, embedded in paraffin, and cut in 6-μm slices. Immunocytochemistry was performed within 2 wk of sectioning the tissue by use of an antibody we raised in sheep against a recombinant 11βHSD2 enzyme (11) at a dilution of 1:2,000, followed by an anti-sheep affinity-purified biotin-labeled antibody detected using a streptavidin-peroxidase system (Zymed Laboratories, S. San Francisco, CA) and diaminobenzidene (Sigma), and counterstained with Gil hematoxylin.
Statistical analysis. Differences in the measured variables between control and treated samples were evaluated by ANOVA, followed by a Bonferroni test where appropriate, with use of a STATISTICA 6.0 (StatSoft) package. Results are expressed as means ± SE.
Levels of male rat kidney 11βHSD1 mRNA, measured by RPA (Fig. 1, top), were unaffected by castration (0.324 ± 0.035 vs. 0.387 ± 0.039), decreased by estradiol valerate (E2) treatment in both intact (0.324 ± 0.035 vs. 0.069 ± 0.014; P = 0.000025) and castrated rats (0.387 ± 0.039 vs. 0.043 ± 0.008; P = 0.000003), and increased in testosterone-treated castrated rats (0.387 ± 0.039 vs. 0.484 ± 0.024; P < 0.05).
Western blot (Fig. 1, bottom) showed that 11βHSD1 protein expression in castrated male rats was about two-thirds that of intact animals, 117.8 ± 9.4 and 79.8 ± 1.2 au, respectively, a change that did not attain significance. 11βHSD1 protein in the kidneys was decreased by E2 treatment, in intact rats from 117.8 ± 9.4 to 24.97 ± 14.0 au (P = 0.0008) and in castrated rats from 79.8 ± 12.0 to 23.5 ± 10.0 au (P = 0.032). 11βHSD1 protein in castrated and testosterone-treated castrated rats was 79.8 ± 12.0 and 116.3 ± 28.0 au, respectively; the difference was not statistically significant. NADP+-dependent enzymatic activity in renal homogenates from intact E2-treated rats was ∼45% less than in those from intact, untreated animals when enzymatic activity was measured in incubations without DTT or in the presence of 1 mM DTT (Fig. 2). The addition of 10 mM DTT to the incubations produced a small, but significant, increase in the activity of NADP+-dependent 11βHSD activity in kidney homogenates from control animals, 27.17 ± 1.51%, compared with incubations of control homogenates without DTT, 21.83 ± 0.52%. The activity in kidney homogenates from E2-treated rats measured in the presence of 10 mM DTT, 25.11 ± 0.67%, was greater than that in aliquots of the same homogenates incubated with 0 or 1 mM DTT, but not significantly different from that in any of the control homogenates (Fig. 2).
Expression of 11βHSD2 mRNA (Fig. 3) was increased by castration, from 0.66 ± 0.05 in intact males to 1.00 ± 0.01 (P = 0.03). Testosterone replacement in the castrated males did not alter expression significantly (1.08 ± 0.06). E2 treatment increased 11βHSD2 mRNA in both intact, from 0.66 ± 0.05 to 1.47 ± 0.1 (P = 0.000004), and castrated, from 1.00 ± 0.01 to 1.64 ± 0.05 (P = 0.00003), male rats. Figure 4 depicts eletrophoresis gels of kidney proteins with and without 10 mM of the reducing agent β-mercaptoethanol. The first, in which no reducing agent was present in the buffer and 30% less protein from the E2-treated animals was loaded on the gel than for the controls, demonstrates the effect of E2 on 11βHSD2 protein expression (n = 4). In the second, β-mercaptoethanol was used as the reducing agent, and graded amounts of protein from the E2-treated animals were loaded to demonstrate the E2-induced increase in 11βHSD2 protein expression (n = 2). The dimer-to-monomer ratio was 1.3 for control rats, which was not significantly different from the ratio for E2 treatment, 1.6. In a similar experiment in which microsomal protein was extracted with the reducing agent β-mercaptoethanol, n = 4, the optical density quantified directly (±SE) was 7.0 ± 1.3 au for control and 35.2 ± 5.2 au for E2-treated rats; P = 0.0018.
11βHSD2 activity was very low for both treatment groups when the kidney microsomes were incubated in buffer alone, and it increased progressively when increasing concentrations of DTT were added to the incubation media (Fig. 5A). Although there was almost no NAD+-dependent activity when 20 μg of kidney microsomes from control and estrogen-treated rats were incubated in buffer without DTT, NAD+-dependent activity in all samples was increased in the presence of 1 and 10 mM DTT. To better appreciate the difference in activity in the different treatment groups, kidney microsomal protein was decreased to 5 μginthe incubation (Fig. 5B). Under these conditions, NAD+-dependent activity of kidneys from E2-treated rats was increased by 71% compared with control rats, 26.7 ± 2.6 compared with 15.6 ± 1.1% conversion of [3H]corticosterone. The 24-h excretion of corticosterone and 11-dehydrocorticosterone, or the ratio of corticosterone to 11-dehydrocorticosterone in the urine, a measurement of 11βHSD activity in vivo, was not significantly altered by castration or treatment of male rats with estradiol (Fig. 6).
The pattern of expression of the 11βHSD2 enzyme in the kidneys of control rats as assessed by immunohistochemistry was similar to that described by others (34) and consistent between animals of a given treatment group. Figure 7 is a composite of photomicrographs of two representative kidneys, one from a control rat and the other from an E2-treated rat, at ×100 and ×200 magnifications, respectively. 11βHSD2 immunoreactivity is seen primarily in the distal convoluted tubules and collecting tubules, with expression disappearing as the tubules penetrate the papilla. E2 treatment markedly increased the number of both tubules and cells stained within a cross section of tubule, as well as the intensity of staining. Kidneys of the E2-treated, but not of control, animals have 11βHSD2 immunoreactive cells in the straight segments of the proximal tubules in the medullary rays and outer strip of the medulla. There is considerable variability in the intensity of staining, even in adjacent cells in the proximal straight tubule, with one-third to more than three-fourths of the cells staining very strongly, and the rest staining weakly or not at all. The staining has a finely granular to reticulated appearance and is equally distributed throughout the cytoplasm. No 11βHSD2 immunoreactivity is seen in glomeruli or proximal convoluted tubules within the cortical labyrinth in any kidneys.
There has been confusion about the distribution, activity, and role of the 11βHSD enzymes in the kidney (11) that was only partially resolved by the discovery that there were two distinct members of this enzyme family. Although the colocalization of the 11βHSD2 with the MR limits access of glucocorticoids to this receptor, both 11βHSD enzymes modulate glucocorticoid access to the glucocorticoid receptor (GR) (35, 41), including GR of the hypothalamo-pituitary-adrenal feedback system (33).
In the present study, as in others (18), 11βHSD1 mRNA and protein are decreased to almost undetectable levels in the kidneys of animals treated with estradiol. However, in vitro NADP+-dependent activity decreased by <50% in incubations under nonreducing conditions, but not when the incubations were performed with 10 mM DTT in the media. This is further evidence that NADP+-dependent 11βHSD activity in the kidney may be mediated by an enzyme different from the 11βHSD1 or -2 isozyme, as previously suggested by us and others (12, 18). Measurements of the 11βHSD1 in microsomes are usually done by determining the enzymatic activity of the dehydrogenase reaction, because the reductase activity is very unstable (19). Even in liver microsomes, the dehydrogenase reaction is easily measured, whereas the reductase has been difficult to measure. 11βHSD1 acts exclusively as a reductase in intact hepatocytes in primary culture (13). Incubation of the kidney microsomes in the presence of 500 nM corticosterone and NADP+ makes it unlikely that the activity seen is due to a contribution of the 11βHSD2, since the activity of this enzyme is very low in the absence of NAD+. In addition, a high level of substrate produces substrate inhibition of the 11βHSD2 enzyme (21).
Estradiol treatment of both intact and castrated male rats resulted in a large increase in the mRNA and protein expressions of the 11βHSD2 enzyme. The demonstration by immunocytochemistry of a marked increase in 11βHSD2 expression induced by estradiol in proximal tubules of the corticomedullary junction corroborates the results of Western blot analysis. Despite this increase, when enzymatic activity was measured in the absence of a reducing agent, no significant activity in microsomes from either control or treated animals was seen, as previously reported (11) (Fig. 5A). Even in the presence of 10 mM DTT, which we have shown to optimize 11βHSD2 activity in kidney homogenates (11), estradiol treatment results in only a 73% increase in NAD+-dependent conversion of cortisol to 11-dehydrocorticosterone (Fig. 5B). It has been proposed that activity of the 11βHSD2 may be regulated by the formation of inactive dimers that act as a latent form of the enzyme (4).
Castration of male rats and estradiol treatment produced no change in the urinary excretion pattern or amounts of corticosterone and 11-dehydrocorticosterone. The increase in 11βHSD2 enzyme expression produced by estradiol treatment is not adequately represented by an increase in in vitro and in vivo enzymatic activity. Activity of 11βHSD1 in microsomes was similar in estradiol-treated and control rats when optimal reducing amounts of DTT were added. Estradiol may promote the formation of dimers of both the 11βHSD1 and -2 enzymes in an unknown manner. If regulated, dimerization or monomerization of the 11βHSDs would serve as a mechanism for rapid modulation of enzymatic activity, and thus MR- and GR-mediated events, at the cellular level.
Several inactivating mutations in the gene for 11βHSD2 have been described in the syndrome of apparent mineralocorticoid excess (AME). Lack of 11βHSD2 activity allows the more abundant endogenous glucocorticoids access to the MR in mineralocorticoid target tissues, increasing both the ratio of cortisol to cortisone metabolites in plasma and urine and the half-life of cortisol and producing clinical signs of mineralocorticoid excess (22, 40). Similar, though less severe, differences in cortisol metabolite excretion have been found in a subset of essential hypertensive subjects compared with control subjects (38). In a third of these patients, the half-life of cortisol exceeded 2 SD of the mean, although no 11βHSD2 gene mutation was found. This milder form of AME could be due to posttranslational changes in the protein-altering enzyme activity, perhaps by increased inactivation through dimer formation. Results of studies using cells transfected with the cDNA encoding the defective 11βHSD2 of patients with AME have been puzzling. There was small, but measurable 11βHSD2 activity in the intact cells, but no activity in homogenates (39). The mutations in these patients might not only produce enzymes that are less effective but may also promote the formation of inactive dimers that are more apparent when cells are homogenized.
Levels of 11βHSD1 and 11βHSD2 are independently regulated in the placenta and fetal tissues throughout gestation (32). The interconversion of cortisol and corticosterone differs in the mother, fetus, and placenta at midterm in baboons, as well as in other mammals (37). Estrogens, rather than adrenocorticoid substrate or product, were found to regulate this difference (24, 36). 11βHSD2 in the placenta maintains normal fetal blood glucocorticoid levels in the face of elevated maternal glucocorticoids early in the pregnancy, and relatively high 11βHSD2 levels in the fetal brain protect the developing brain throughout, when 11βHSD2 in the placenta decreases significantly near term, allowing maternal glucocorticoids, crucial for fetal lung maturation, to enter the fetal circulation (5, 25).
The data presented herein support the possibility that, in addition to the genomic regulation of 11βHSD enzyme expression by estrogens, increasing levels of estrogenic steroids exert a posttranscriptional regulation of 11βHSD2. Estrogens increase near term. 11βHSD2 dimerization in some organs, but not in others, in response to periparturient elevations of estrogens might be a mechanism to ensure that glucocorticoids pass to the maturing fetal lungs without loss of 11βHSD2 protection of MR where still needed, for example, in the kidney and brain.
In conclusion, both message and protein for the NADP+-dependent 11βHSD1 were reduced to very low levels by estradiol in both intact and castrated male rats. However, enzymatic activity did not change, particularly when the reducing agent DTT was added to the incubation media. These data support the suggestion that another NADP+-dependent enzyme with 11βHSD activity exists. There have been other reports of 11βHSD activity that does not appear to be due to either 11βHSD1 or -2, but as yet no other enzyme has been isolated or gene cloned (10, 12, 18). Despite a large increase in message and protein measured by both Western blot and immunocytochemistry for the 11βHSD2 enzyme, in vivo activity, as measured by the ratio of urinary corticosterone to 11-dehydrocorticosterone, did not change in estradiol-treated rats. We have previously reported that reducing agents separate inactive 11βHSD dimers into active monomers. In this study, NADP+-dependent enzyme activity in renal microsomes from estradiol-treated rats was lower than that of controls under nonreducing conditions but was restored to the level of control kidneys by the addition of DTT. NAD+-dependent activity in microsomes from control or estradiol-treated kidneys was minimal, and no difference between the treated and untreated groups could be appreciated unless DTT was added to the incubation media. We suggest that, in addition to transcriptional regulation, 11βHSD activity may be regulated in vivo by dimer formation or other posttranslational modification. The significance of the large increase in 11βHSD2 expression in proximal tubules of the cortical medullary junction induced by estradiol without a significant change in vivo or in vitro activity is yet unknown.
These studies were supported by medical research funds from the Department of Veteran Affairs and National Heart, Lung, and Blood Institute Grants HL-27255 and HL-27737.
Submitted 16 September 2002
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. Section 1734 solely to indicate this fact.
- Copyright © 2003 by American Physiological Society