Pendrin is Upregulated by Corticosterone and Participates in its Pressor Response

Abstract

NaCl restriction upregulates pendrin, in part, through increased circulating aldosterone and the intercalated cell (IC) mineralocorticoid receptor (MR). Since 11 β HSD2 enhances aldosterone binding to this receptor in other cells, we asked if pendrin abundance is reduced in NaCl-restricted 11 β HSD2 KO rats. However, pendrin abundance was greater in 11 β HSD2 KOs than in controls, possibly from enhanced glucocorticoid MR activation. The MR antagonist, spironolactone, reduced pendrin abundance in mice that do not produce aldosterone (aldosterone synthase KO). IC MR gene ablation also reduced pendrin protein abundance in corticosterone-treated, adrenalectomized mice. Therefore, the MR regulates pendrin independently of aldosterone. As such, we asked if glucocorticoids, the other MR ligands, change pendrin abundance and/or subcellular distribution in adrenalectomized wild type mice. We observed that corticosterone upregulated pendrin in a dose-dependent fashion through both increased total protein abundance and subcellular redistribution. At higher doses, corticosterone increased pendrin abundance from greater pendrin-positive cell number within the late distal convoluted tubule 2 (DCT2) rather than increased pendrin abundance per cell. Finally, we asked if pendrin contributes to the hypertension seen in rodent models of Cushing Syndrome. While corticosterone increased blood pressure in wild type mice, it had no effect in pendrin KOs. In conclusion, glucocorticoids upregulate pendrin by increasing pendrin total protein abundance through an MR-dependent pathway and through subcellular redistribution. Glucocorticoids increase pendrin abundance by increasing the number of pendrin-positive cells within the DCT2. In so doing, pendrin contributes to the hypertension seen in rodent models of Cushing Syndrome.

Graphical Abstract

INTRODUCTION:

Mineralocorticoid as well as glucocorticoid hormones bind to the mineralocorticoid receptor (MR) in connecting tubule cells and principal cells of the connecting tubule (CNT) and the cortical collecting duct (CCD) (1–6). Upon aldosterone binding, the receptor is activated, which upregulates transporters and channels, such as the epithelial sodium channel, thereby increasing Na⁺ absorption and K⁺ secretion (2, 3, 7). Because circulating levels of corticosterone/cortisol are generally much higher than of aldosterone, while receptor affinity is similar for both hormones, aldosterone receptor binding is thought to depend on 11 β Hydroxysteroid Dehydrogenase 2 (11 β HSD2) to reduce effective corticosterone/cortisol concentration via its oxidation (5, 8). As such, 11 β HSD2 confers ligand specificity.

While it is well-established that the IC MR directly regulates pendrin (9), the low 11 β HSD2 expression observed in ICs (9) should mitigate aldosterone binding to the MR. There is also evidence that aldosterone upregulates pendrin through indirect effects, such as changes in acid-base balance rather than through direct MR ligand binding (10). Therefore, pendrin’s response to IC MR activation might occur, at least in part, through the receptor’s interaction with another ligand. The low or absent 11 β HSD2 expression found in ICs might leave glucocorticoid binding to the IC MR unopposed, thereby making cortisol/corticosterone an important MR ligand. On the other hand, the 11 β HSD activity detected in ICs (11) may be sufficient to oxidize glucocorticoids and thereby promote aldosterone receptor binding.

The purpose of this study was therefore to determine 1) if 11 β HSD2 gene ablation reduces pendrin’s response to dietary NaCl restriction, 2) if MR antagonists reduce pendrin abundance in the known absence of aldosterone, 3) if corticosterone upregulates pendrin and 4) if pendrin gene ablation blunts the pressor response to corticosterone administration.

METHODS:

Animals:

We studied the IC-specific, mineralocorticoid receptor KO mice we reported previously (9), which were developed by breeding floxed MR mice (12) with mice expressing Cre recombinase driven by the B1-H⁺-ATPase promoter (13), encoded by ATP6V1B1. These IC MR KOs were compared to their floxed MR littermates. We also studied aldosterone synthase null mice (14) and compared them to their wild type littermates. We studied global pendrin KO mice and their wild type littermates on a C57Bl/6 background that were described previously (15). Finally, 11β HSD2 null rats and their littermate controls on a Fischer (F344coCrl) background were studied (16). Genotype was determined from tail biopsies by PCR with specific probes designed for each gene (Transnetyx, Cordova, TN). Mice and rats were euthanized under 1–4% isofluorane/100% oxygen anesthesia.

The Institutional Animal Care and Use Committee at Emory University approved all treatment protocols.

Animal Conditioning:

Treatment 1: Effect of NaCl intake in 11 β HSD2 null and control rats.

For 10 days, 11 β HSD2 KO and control littermate rats (16) drank water ad libitum and ate a standard, pelleted NaCl-restricted rodent diet containing 0.05% NaCl (TD.90228, Envigo, Madison, WI) or an identical diet with NaCl added to give 0.7% NaCl (TD.96329, Envigo).

Treatment 2: Effect of spironolactone in angiotensin II-treated mice that do not produce aldosterone.

Aldosterone Synthase null mice were given 14 days of angiotensin II at 490 ng/kg bw/min by minipump and a NaCl-restricted diet (TD #90228) prepared as a gel (0.6% agar, 74.6% water, and 24.8% mouse chow) that was supplemented with NaCl to give a diet that is 0.5% Na⁺, 0.74% K⁺ and 0.77% Cl^-. This diet provided mice 1 meq/day NaCl and 0.95 meq/day K⁺. Over the treatment period, the gelled diet was supplemented with either vehicle (100% ethanol) or spironolactone to give mice 0 or 500 μg spironolactone/day.

Treatment 3: Effect of intercalated cell mineralocorticoid receptor gene ablation in adrenalectomized, corticosterone-treated mice.

For 10 days, adrenalectomized floxed MR and IC MR KO mice were fed a balanced, NaCl-restricted diet (Zeigler Brothers, #538813) prepared as a gel (0.6% agar, 74.6% water, and 24.8% mouse chow), which was supplemented with NaCl to give a diet that was 0.62% Na⁺, 0.78% K⁺ and 0.95% Cl^-. This provided mice 1.3 meq/day NaCl and 0.91 meq/day K⁺. Over the treatment period, the gelled diet was supplemented with corticosterone to give 200 μg/day.

Treatment 4: Effect of corticosterone in adrenalectomized and in adrenal-intact mice.

For 10 days, adrenalectomized or adrenal-intact mice were fed a balanced, NaCl-restricted diet (Zeigler Brothers, #538813) prepared as a gel (0.6% agar, 74.6% water, and 24.8% mouse chow), which was supplemented with NaCl and either corticosterone or corticosterone’s vehicle (100% ethanol) to give a diet that was 0.62% Na⁺, 0.78% K⁺ and 0.95% Cl^-, thereby providing mice with 1.3 meq/day NaCl, 0.91 meq/day K⁺ and with 0, 200 or 800 μg/day corticosterone. These corticosterone doses reflect basal, physiological and stress doses of this glucocorticoid hormone (17).

Treatment 5: Effect of corticosterone on blood pressure in adrenal-intact wild type and pendrin KO mice.

Adrenal-intact wild type and global pendrin KO mice, reported previously (15), ate a standard rodent diet (LabDiet 5001, 0.4% Na⁺, 1.1% K⁺) and drank water ad libitum (basal condition). For 14 days mice were then fed a balanced, NaCl-restricted diet (Zeigler Brothers, #538813) prepared as a gel (0.6% agar, 74.6% water, and 24.8% mouse chow), supplemented with NaCl and corticosterone to give a diet that was 0.62% Na⁺, 0.78% K⁺ and 0.95% Cl^- (1.3 meq/day NaCl and 0.91 meq/day K⁺) and provided 800 μg/day of corticosterone.

In all experiments, mice were given the gelled diets between ~2 and 4 pm. They were sacrificed between 10 am and 6 pm. When sacrificing mice and rats, groups were alternated.

Measurement of serum and urine chemistries, osmolality, arterial blood gases and serum aldosterone.

Blood was collected for serum chemistries and arterial blood gases through the abdominal aorta under isofluorane anesthesia. Serum electrolytes and arterial blood gases were measured with an iSTAT Alinity V (Abbot Point of Care, Princeton, NJ). Serum aldosterone was measured by radioimmunoassay (Michigan State Veterinary Diagnostic Lab/IDEXX).

Antibodies:

For immunohistochemistry and immunoblot studies, we used a rabbit, anti-rat pendrin antibody (18) at 1:400,000 or 0.93 ng/ml IgG for immunohistochemistry with polywax sections and 1:10,000 or 37 ng/ml IgG for immunoblots and immunohistochemistry with paraffin sections. For immunofluorescence experiments, we used a goat anti-Aqp2 at 1:200 (Novus NBP1–70378), a Chicken anti-Calbindin D at 1:1000 (Novus NBP2–50028) and a Rabbit anti-Pendrin antibody at a 1:200 dilution (Bicell 20501).

Immunohistochemistry and Quantitative Analysis of Immunohistochemistry:

For standard immunohistochemistry, kidneys were fixed in situ with 3% paraformaldehyde, lysine, meta-periodate (PLP) and then embedded in paraffin or polyester wax [polyethylene glycol 400 distearate (Polysciences, Warrington, PA) and 10% 1-hexadecanol] as described previously (19). Immunoreactivity was detected using immunoperoxidase procedures. For polywax embedded tissue, blocking was done with 3% H₂O₂ for 45 minutes, followed by protein blocking using Dako Protein Blocker Serum-Free (Cat#X0909). Sections were incubated in the primary antibody diluted in Dako Antibody Diluent with Background Reducing Components (Cat#S3022) overnight at 4°C. Sections were rinsed with 1X PBS and incubated with MACH 2 Rabbit HRP-Polymer (Biocare Med Cat#RHRP520G) for 30 minutes, washed with PBS and incubated with diaminobenzidine (DAB substrate kit, Vector). Sections were washed in with distilled water, counter stained with hematoxylin, dehydrated with graded ethanols and xylene, mounted and observed by light microscopy.

For paraffin-embedded tissue, 2-micron-thick sections were cut and mounted on Superfrost Plus glass slides, as described previously (20, 21). Immunolocalization used standard immunoperoxidase procedures. Sections were dewaxed, rehydrated, and rinsed in distilled water. Endogenous peroxidase activity was blocked by incubating the sections in 3% H₂O₂ in distilled water for 45 min. For pendrin labeling, sections were blocked for 1 hour in M.O.M. blocker (MKB-2213, Vector Laboratories, Burlingame, CA) along with an unconjugated Fab fragment goat anti-mouse IgG (115–007-003, Jackson ImmunoResearch, West Grove, PA) and 15 min with 2.5% Normal Horse Serum (MP-2400, Vector Laboratories, Burlingame, CA). Sections were then incubated at 4°C overnight with primary antibody diluted in 1 x PBS, washed in PBS and incubated for 45 min in polymer-linked peroxidase-conjugated anti-mouse IgG diluted to 1:10 or 30 min in anti-rabbit IgG (Vector ImmPRESS, Vector Laboratories, Burlingame, CA), washed again with PBS, and exposed to diaminobenzidine (Vector DAB substrate kit) for 5 min. Sections were then washed in distilled water, dehydrated in graded ethanols and xylene, mounted, and observed by light microscopy.

Transporter subcellular distribution was quantified as described previously in bright field light micrographs (21, 22). High-resolution digital micrographs were taken of defined tubule segments using a Leica DM4500 microscope and a Zeiss Axiocam 705 5 digital camera (14.4-megapixel images, with 10X or 63X objectives) and Zeiss Zenlite software. Pixel intensity across a line drawn from the tubule lumen through the center of an individual cell was quantified with NIH ImageJ, version 1.34s software. Background pixel intensity was calculated as the mean pixel intensity outside the cell and was subtracted from the pixel intensity at each point. Total cellular expression was quantified by integrating total cellular net pixel intensity per cell. Total cell pixel intensity was determined as the pixel intensity of the region demarcated with the ImageJ freehand tool, which traced the cell membrane. For each cell, nuclear pixel intensity was quantified by integrating pixel intensity within the nuclear region that was outlined with the ImageJ freehand tool. Cross sectional cytoplasmic area was determined using previously determined calibrations that involved calculating the area within the plasma membrane minus the area within the nucleus of that cell.

When cell height or cytoplasmic area differed between the 2 groups being compared, we quantified pendrin label/cell in the most apical 1 μm, rather than the pendrin redistribution ratio. Immunolabel expression intensity in the most apical 1 μm was determined from cell height and quantitative analysis of cellular immunolabel intensity as described previously (21). Apical membrane length was determined by tracing the apical membrane using the freehand tool in ImageJ. The length of the traced membrane was converted from pixels to μm using previously determined calibrations. Length-adjusted apical expression was then calculated as the product of apical expression and apical membrane length.

To quantify apical label/cell or the redistribution ratio, a minimum of 15 cells of a given cell type were analyzed per kidney. When calculating total pendrin label/cell, we analyzed at least 20 cells of a given cell type from each kidney. The individual performing the microscopy and quantifying the results was blinded as to the treatment group of each animal. For each animal, data from all cells of a given subtype were averaged to achieve a single value for that cell type. This value was used in the statistical analysis. All images shown were taken from polywax sections, whereas labeling was quantified in sections from polywax or paraffin sections.

Immunoblots:

Immunoblots were performed using methods reported previously (23, 24). Whole kidney lysates were isolated by harvesting mouse kidneys and placing them in an ice cooled buffer (0.3 M sucrose, 25 mM imidazole, pH 7.2, containing 1x Roche Complete Protease Inhibitor Cocktail). Tissue was immediately homogenized using an Omni THQ Tissue Homogenizer (Omni International) and then centrifuged at 1000 x g for 15 min at 4^oC. To prepare whole cell lysates, kidneys were homogenized in Gentle Lysis Buffer (10 mM Tris·HCl, 10 mM NaCl, 2 mM EDTA, 0.5% NP-40, 1% glycerol, freshly added 0.18 μg/ml Na₃VO₄, 10 μg/ml PMSF, 5 μg/ml aprotinin, and 1 μg/ml leupeptin). To enable equal protein loading in each lane, protein content in the soluble fraction of homogenates was measured using a RC-PC protein assay kit (DC Protein Assay Kit, Bio-Rad, Hercules, CA) and then dissolved in Laemmli buffer.

Aliquots containing equal amounts of protein from these lysates were separated by SDS-PAGE on 8.5% acrylamide gels and then electroblotted to PVDF membranes (Immobilon, Millipore, Bedford, MA). Blots were blocked with Odyssey Blocking Buffer (LI-COR Biosciences) following the manufacturer’s instructions and then incubated with primary antibody overnight at 4°C, followed by incubation for 2 hours at room temperature with Alexa Fluor 680-linked anti-rabbit IgG (Invitrogen). Pendrin protein was detected by immunoblot using a rabbit anti-rat pendrin antibody (18). To correct for possible differences between lanes in lysate protein loading, membranes were Coomassie stained as reported previously (25). Signals were visualized with an Odyssey Infrared Imaging System (LI-COR Biosciences). Immunoblot and Coomassie band densities were quantified using software program Image J (NIH, available at http://rsb.info.nih.gov/). Pendrin band density was normalized to the density of the Coomassie gel band with the same mobility. To confirm protein loading, actin immunoreactivity was quantified in the same blots, using a rabbit, anti-actin antibody (Sigma Aldrich, #A2066)).

Volume Density Analysis:

The volume density of pendrin positive cells in mouse cortex was determined using previously described standard point-counting techniques (26). High resolution digital micrographs of pendrin immunolabel were taken at 20x magnification and overlayed with a standardized grid through NIH ImageJ software. Volume density was determined by counting the number of grid intersections in cortical sections at which pendrin-positive cells were located and dividing that number by the total number of grid intersections within that section. At least 15 micrographs were analyzed per kidney. The volume-density values for each image within a single kidney were averaged to yield a single data point per animal. The average for each animal was used in the statistical analysis.

Counting Pendrin-positive cells in the CNT and DCT2:

Kidneys were sectioned at a 5μm thickness, deparaffinized in xylene, rehydrated through a series of ethanol baths, boiled for 20 minutes in Simple Green antigen unmasking buffer (1.85% v/v Simple Green, 10mM sodium citrate, pH 6) and then cooled at room temperature for 20 minutes. Sections were blocked in PBS containing 1% bovine serum albumin, 0.2% powdered milk, 4% v/v fish gelatin, and 0.3% Triton X-100 for 30 minutes at room temperature before incubation with primary antibodies inside a humidified chamber at room temperature for one hour. The primary antibodies used included Aqp2 (Novus NBP1–70378 at a dilution of 1:200), Calbindin-D (Novus NBP2–50028 at a dilution of 1:1000) and Pendrin (Bicell 20501 at a dilution of 1:200). Sections were then incubated for 1 hour at room temperature with fluorescently conjugated secondary antibodies at a 1:250 dilution to visualize the primary antibodies and Hoechst to stain nuclei. Coverslips were mounted using ProLong Gold Antifade Mountant and sealed with clear nail polish. Slides were imaged on a Nikon A1R confocal microscope using NIS elements software.

Images of kidney sections were analyzed by two blinded, independent observers. At least seven tubule cross-sections of each nephron segment (CNT and DCT2) were analyzed in each kidney. Tubule cross-sections with less than five cells or ambiguous staining were excluded. Individual cells were identified by the presence of a nucleus, and/or a discrete membrane marker. The curving of the apical membrane towards the basolateral membrane at the lateral sides of a cell was also used to determine the number of individual cells among adjacent cells expressing the same membrane marker. Each cell was assigned as positive or negative for AQP2, Calbindin-D and Pendrin staining. Connecting tubule (CNT) segments were identified as positive for AQP2, Pendrin, and Calbindin, whereas distal convoluted tubule 2 (DCT2) segments were identified as positive for Pendrin and Calbindin, but negative for AQP2. The percentage of Pendrin-positive cells relative to total cells in DCT2 segments were calculated for each kidney. The percentage of Pendrin-positve cells in CNT were similarly calculated.

Quantitative PCR

Kidneys were homogenized in TRIzol RNA isolation reagent (Invitrogen). After TRIzol extractions, RNA was precipitated with isopropanol, washed twice with 70% ethanol, resuspended in diethyl pyrocarbonate (DEPC)-treated water and then quantitated by measuring the optical density at 260 nm, 280 nm, and 320 nm. Because the concentration was high and variable from sample to sample, RNA was diluted to a concentration range of 250–300 ng/μL and then re-quantified. Reverse transcription was performed using Superscript III (Invitrogen, Cat. No. 18080051) following the manufacturer’s instructions. Reactions (20 μL) contained 1 μg RNA, 4 μL 5× buffer, 10 mM DTT, 2.5 μM Random hexamers (Invitrogen, Cat. No. 100026484), 1.25 mM MgCl2, 250 nM dNTP, and 100 U Superscript III. After a final denaturation step, reactions were diluted three times by adding 40 μL of water. The Applied Biosystems QuantStudio 3 Real-Time PCR system was used to carry out the quantitative PCR.

Slc26a4 (pendrin) primers employed were the following:

Forward: GCTGGCCTCATCTCAGCTG

Reverse: GCAAGGGTTCCAGAAGCCT

GAPDH primers employed were the following:

Forward: AGGTCGGTGTGAACGGATTTG

Reverse: GGGGTCGTTGATGGCAACA

Quantitative PCR reaction components included 12.5 μL of SYBR Green PCR Master Mix (Applied Biosystems, Foster City, CA), 0.5 μL of each primer (1 μM), 8.5 μL of water, and 3 μL of cDNA. Pendrin transcript (mRNA) was normalized to GAPDH. The PCR cycling conditions used were an initial denaturation step at 95°C for 2 min, followed by denaturation at 95°C for 30 s, annealing for 30 s at 60°C, and extension at 72°C for 30 s. These short steps were repeated or cycled 40 times. The melting curve analysis confirmed the presence of only one fragment for the primer set.

Blood Pressure Measured by Tail Cuff

Systolic blood pressure was measured in conscious mice by tail cuff using a B-2000 (Visitech Systems), as we reported previously (22). To condition mice for tail cuff blood pressure readings, animals were placed in a platform for 15 min on 2 consecutive days. Over the next 3–4 consecutive days, mice were placed on the platform and at least four readings were taken. All conditioning and all blood pressure readings were performed at the same location under quiet, low-light conditions. Measurements and conditioning were performed by the same operator at the same time of day.

Statistics.

Data are presented as the mean ± SE. Each “n” used in the statistical analysis represents data from separate animals. To test statistical significance between two groups, a paired or unpaired Student’s t-test was used. Multiple groups were compared by ANOVA with a Holm-Sidak or a Bonferoni post test. The criterion for statistical significance was P < 0.05.

RESULTS:

11 β HSD2 gene ablation increases pendrin abundance.

Since aldosterone upregulates pendrin (27) and since 11 β HSD2 enhances aldosterone binding to the mineralocorticoid receptor (MR) in principal and connecting tubule (CNT) cells, we asked if 11 β HSD2 gene ablation reduces pendrin abundance under conditions in which circulating aldosterone concentration is high, such as with dietary NaCl restriction (28) (Treatment #1). Consistent with previous observations in mouse (28, 29), we observed that in both male and female control rats, pendrin abundance was on average higher following a NaCl-restricted than NaCl-replete diet (Treatment #1), although differences did not reach statistical significance (Figure 1). However, in both sexes pendrin total protein abundance was higher, rather than lower, in 11 β HSD2 KO rats relative to controls following either a NaCl-replete diet or a NaCl-restricted diet. Therefore, 11 β HSD2 gene ablation increases pendrin abundance.

Figure 1: 11 β HSD2 gene ablation increases pendrin abundance.

Male and female control (wild type) and 11 β HSD2 KO rat littermates were given a NaCl-restricted or -replete diet for 10 days (Treatment #1). Pendrin abundance was quantified in kidney lysates from each group by immunoblot. Panel A show a representative immunoblot. Pendrin band density in lysates from both males (closed circles) and females (open circles) are shown in Panels B-D. Lysates from males and females were run on the same blot. For each value, band density was expressed as a percent of the male controls given the NaCl-replete diet. *P< 0.05 ANOVA with a Holm-Sidak post-test.

We next explored the effect of 11 β HSD2 gene ablation on apical pendrin abundance of cells within the cortical labyrinth, which includes the DCT2 and the CNT. Figure 2 shows that following a NaCl-replete diet, pendrin abundance per cell in the most apical 1 μm was higher in cells from male KO than control rats. Therefore, eliminating 11 β HSD2 increases both total and apical pendrin abundance.

Figure 2: Apical pendrin abundance is higher in kidneys from male NaCl-replete 11 β HSD2 null than wild type littermates.

Panel A shows a typical micrograph of pendrin immunolabel in the cortical labyrinth and medullary rays of male NaCl-replete control (wild type) and a 11 β HSD2 null rats (Treatment #1). Panel B shows pendrin immunolabel per cell in the most apical 1 μm, cytoplasmic area, and pendrin label per cell within the cortical labyrinth. *P<0.05, unpaired t-test. Bar represents 10 μm.

Further experiments asked why pendrin abundance is greater in NaCl-replete 11 β HSD2 KO relative to control rats. Since pendrin protein abundance is highly regulated by changes in acid-base and electrolyte balance (30), we examined serum electrolytes and acid-base balance in these groups. Table 1 shows that in both sexes, serum K⁺ is lower in NaCl-replete KOs relative to controls, which is consistent with previous observations (16). Arterial pH and HCO₃^- concentration were also higher in the NaCl-replete KOs than in NaCl-replete controls of both sexes. However, while arterial pH and HCO₃^- concentration were higher, while serum K⁺ was lower, in the NaCl-restricted female KOs relative to their female wild type controls, in males, however, we saw no difference in arterial blood gases or serum K⁺ between NaCl-restricted KOs and controls. We conclude that while the alkalosis observed in the NaCl-replete 11 β HSD2 KO rats likely contributes to the increment in total and apical pendrin abundance, changes in arterial pH cannot explain the increase in pendrin abundance seen in the male NaCl-restricted 11 β HSD2 KOs.

Table 1:

Serum aldosterone, electrolytes and arterial blood gases in rats and mice

	Arterial blood gases			Electrolytes
	pH	pCO2	cHCO3-	Na+	K+		Aldo
		mm Hg	mM	mEq	mEq	n	pM	n
Female rats (Treatment 1)
NaCl-replete
Control	7.470 ± 0.013*	41.8 ± 1.5*	30.3 ± 0.5*	138 ± 1	3.4 ± 0.1*	7	NM
11 β HSD2 KO	7.539 ± 0.018*	55.3 ± 2.1*	47.2 ± 2.0*	140 ± 1	2.5 ± 0.2*	5	NM
NaCl-restricted
Control	7.425 ± 0.016*	46.3 ± 2.2	30.3 ± 0.9*	138 ± 1	3.5 ± 0.1*	5	NM
11 β HSD2 KO	7.500 ± 0.015*	46.0 ± 1.5	35.5 ± 1.2*	140 ± 1	2.7 ± 0.1*	7	NM
Male rats (Treatment 1)
NaCl-replete
Control	7.429 ± 0.027	50.6 ± 3.4	33.3 ± 0.3*	137 ± 1	3.4 ± 0.2*	4	NM
11 β HSD2 KO	7.502 ± 0.023	54.7 ± 1.9	42.7 ± 1.0*	139 ± 1	2.7 ± 0.1*	6	NM
NaCl-restricted
Control	7.448 ± 0.012	41.4 ± 1.8	28.5 ± 0.7	136 ± 1	3.7 ± 0.1	12	NM
11 β HSD2 KO	7.436 ± 0.018	44.5 ± 2.8	29.6 ± 1.3	137 ± 1	3.7 ± 0.3	9	NM
Angiotensin II-treated male aldosterone synthase KO mice (Treatment 2)
Vehicle	7.361 ± 0.042	46.4 ± 2.2	26.3 ± 1.5	149 ± 2	4.0 ± 0.2	3	NM
Spironolactone	7.333 ± 0.019	47.2 ± 1.5	25.1 ± 0.8	147 ± 1	4.2 ± 0.1	5	NM
Adrenalectomized wild type male mice (Treatment 4)
Vehicle	7.404 ± 0.0125#	35.8 ± 1.6#	22.3 ± 0.8	141 ± 1	4.3 ± 0.1	12	64 ± 8	4
Corticosterone 200 μg/d	7.449 ± 0.015	31.2 ± 1.4	21.5 ± 0.5	141 ± 1	4.5 ± 0.1	8	252 ± 56	3
Corticosterone 800 μg/d	7.465 ± 0.020#	29.3 ± 1.1#	21.0 ± 0.7	142 ± 1	4.3 ± 0.1	6	207 ± 42	5
Adrenalectomized male mice given corticosterone, 200 μg/d (Treatment 3)
Floxed MR	7.438 ± 0.022	32.2 ± 2.1	21.5 ± 0.8	144 ± 1*	4.4 ± 0.1*	12	NM
IC MR KO	7.421 ± 0.028	31.6 ± 2.3	20.4 ± 1.2	141 ± 1*	5.4 ± 0.2*	8	NM
Adrenal-intact male mice given vehicle or corticosterone, 800 μg/day for 10 days (Treatment 4)
Vehicle	7.375 ± 0.019	33.4 ± 2.4	19.3 ± 0.6	143 ± 1	4.1 ± 0.2	6	NM
Corticosterone 800 μg/d	7.442 ± 0.013*	28.4 ± 0.7	19.6 ± 0.8	144 ± 1	4.3 ± 0.1	5	NM
Adrenal-intact male mice given corticosterone, 800 μg/day for 2 weeks (Treatment 5)
Wild type	7.442 ± 0.022	30.9 ± 1.2*	21.1 ± 0.8*	142 ± 1	4.1 ± 0.1	5	NM
Pendrin KO	7.459 ± 0.016	35.1 ± 1.1*	25.2 ± 1.1*	142 ± 1	3.6 ± 0.1*	6	NM

cHCO₃^-, calculated HCO₃^- concentration; KO, knockout; MR, mineralocorticoid receptor; IC MR KO, intercalated cell mineralocorticoid receptor KO; aldo, aldosterone; NM, not measured

^*P < 0.05, t test

^#P < 0.05, ANOVA

MR antagonists reduce pendrin abundance independently of aldosterone.

Since changes in arterial blood gases cannot fully explain the increment in pendrin abundance seen in 11 β HSD2 null rats, we explored other mechanisms that might mediate 11 β HSD2’s effect on pendrin. It is well-established that eliminating 11 β HSD2 enhances MR activation by glucocorticoid hormones in principal cells (31). We therefore hypothesized that glucocorticoid hormones activate the IC MR, which upregulates pendrin. If so, MR antagonists should reduce pendrin abundance in the absence of aldosterone. To test this hypothesis, we examined the effect of MR inhibitors on pendrin abundance in male mice that do not produce aldosterone, i.e. aldosterone synthase null mice. These KO mice were given angiotensin II by minipump to activate the MR (32–34) and either spironolactone or vehicle (Treatment #2). As shown (Figure 3A & B), pendrin abundance was lower in aldosterone synthase null mice given spironolactone than in those given vehicle. Since arterial pH and HCO₃^- concentration, as well as electrolytes, were similar in both groups (Table 1), the fall in pendrin abundance seen with spironolactone administration cannot be attributed to changes in acid-base balance or electrolyte concentration. We conclude that MR antagonists reduce pendrin abundance through an aldosterone-independent mechanism.

Figure 3: Mineralocorticoid receptor antagonists and IC MR gene ablation reduce pendrin abundance independently of aldosterone.

Male aldosterone synthase null mice were given angiotensin II and a NaCl-replete diet with either spironolactone or its vehicle in their gelled diet (Treatment #2). Panel A shows a typical pendrin immunoblot, while Panel B shows pendrin band density of kidney lysates from each of these groups. In other experiments, adrenalectomized male floxed MR and IC MR KO mice were given 200 μg/day corticosterone for 10 days (Treatment #3). Panel C shows a typical pendrin immunoblot, while Panel D shows pendrin band density of lysates from each of these groups. V, vehicle; S, spironolactone; MR, mineralocorticoid receptor; KO, knockout; Flx, floxed. *P < 0.05, unpaired t-test.

Further experiments explored whether IC MR gene ablation changes pendrin abundance under conditions in which circulating aldosterone concentration is very low. To answer this question, we gave adrenalectomized, male floxed MR and IC MR null mice (9) 10 days of corticosterone (200 μg/day) (Treatment #3). Figures 3C & D show that pendrin total protein abundance was ~40% lower in kidneys from the IC MR KO than the floxed MR mice. This difference cannot be attributed to changes in acid-base balance since arterial pH and calculated HCO₃^- concentration were similar in the control and mutant mice (Table 1), although serum K⁺ was higher in the KO than in the floxed controls. Moreover, these differences cannot be attributed to aldosterone activation of the MR since circulating aldosterone concentration under these conditions is very low (Table 1). We conclude that IC MR gene ablation reduces pendrin abundance under conditions in which circulating aldosterone concentration is very low or absent.

Corticosterone upregulates pendrin

Since MR antagonists and IC MR gene ablation can reduce pendrin abundance independently of aldosterone, we hypothesized that pendrin is upregulated by the MR’s other known ligand, corticosterone. To test this hypothesis, adrenalectomized male and female wild type mice were given vehicle or corticosterone at a physiological (200 μg/day) or a stress dose (800 μg/day) for 10 days (Treatment #4) (35, 36). Arterial pH was slightly higher in adrenalectomized mice given corticosterone than in those given vehicle (Table 1) due to the lower pCO₂ observed in the former (37). Figures 4 A-D show that in both adrenalectomized males and females, corticosterone increased pendrin total protein abundance in a dose-dependent fashion. However, this increase in pendrin protein abundance was not accompanied by increased Slc26a4 (pendrin) mRNA (Figure 4E).

Figure 4: Corticosterone increases pendrin abundance.

Panels A & B show typical pendrin immunoblots of kidney lysates from adrenalectomized male and female mice given 10 days of vehicle or corticosterone at physiological (200 μg/day) or stress doses (800 μg/day, Treatment #4). Pendrin band density of lysates from vehicle- and corticosterone-treated males (closed circles, Panel C) and in females (Panel D, open circles) is shown. Panel E shows pendrin (Slc26a4 )/GAPDH transcript (mRNA) in kidneys from adrenalectomized male mice treated with corticosterone or vehicle. *P<0.05 by ANOVA with a Bonferoni posttest (males) and a Holm-Sidak posttest (females).

Further experiments explored the effect of corticosterone on pendrin subcellular distribution. Figure 5A shows pendrin immunolabel in the cortical labyrinth and the medullary rays of adrenalectomized, male mice that received vehicle or corticosterone at 200 or 800 μg/day. Within the cortical labyrinth, pendrin’s relative abundance in the most apical region of ICs (the redistribution ratio) was higher in corticosterone- than in vehicle-treated mice (Figure 5A & B). We conclude that corticosterone upregulates pendrin by increasing its total protein abundance as well as through subcellular redistribution, which increases pendrin’s relative abundance in the region of the apical plasma membrane.

Figure 5: Corticosterone increases pendrin’s relative abundance in the most apical region of the cell.

Adrenalectomized male mice were given vehicle or corticosterone at 200 or 800 μg/day (Treatment #4). Panel A shows pendrin label in the cortical labyrinth and medullary rays of mice from each group. The bar represents 10 μm. Panel B shows the pendrin redistribution ratio, Panel C shows label per cell and Panel D shows cytoplasmic cross-sectional area of pendrin-positive ICs within the cortical labyrinth of mice from each group. *P < 0.05 ANOVA, Holm-Sidak post test.

Because pendrin abundance was higher in kidney lysates from corticosterone-treated than vehicle-treated mice, we asked if this steroid hormone also increases pendrin label per cell. As shown, corticosterone did not increase pendrin label/cell at either dose (Figure 5C). Moreover, pendrin label per cell was lower in mice that received 800 than in those receiving 200 μg/day, which is in stark contrast to the increase in pendrin abundance we observed by immunoblot at this higher corticosterone dose (Figure 4). We therefore hypothesized that corticosterone treatment increases pendrin total protein abundance, at least at higher doses, by increasing the number of pendrin positive cells. To test this hypothesis, we measured the percent of cortical area occupied by pendrin-positive cells in mice that received 200 or 800 μg/day corticosterone or vehicle for 10 days (Treatment #4, Figures 6A & B). This ratio reflects both the cross-sectional area of pendrin positive cells as well as pendrin-positive cell number. However, since cytoplasmic cross-sectional area was similar in adrenalectomized mice that received vehicle or corticosterone at either 200 or 800 μg/day (Figure 5D), changes in this ratio should primarily reflect changes in cell number. While there was a trend towards an increase in the ratio of pendrin positive cell area to total cortical cell area (Figure 6A & B), differences between these groups were not statistically different (P = 0.088). To explore this issue further, we quantified the number of pendrin-positive cells in both the distal portion of the distal convoluted tubule, DCT2, as well as the CNT (Figure 6C & D), within the cortical labyrinth of kidneys from adrenalectomized male mice that received 800 μg/day of corticosterone or vehicle (Treatment #4). DCT2 segments were identified as those expressing Calbindin, but not AQP2, while the CNT labeled for both Calbindin and AQP2 (38, 39). We saw no difference in pendrin positive cell number in the CNT. However, pendrin-positive cell number was higher in the DCT2 of corticosterone- than vehicle-treated mice (Figures 6 C & D). There was also a trend towards fewer cells that are both AQP2- and pendrin-positive in the CNT of corticosterone relative to vehicle-treated mice (7.8 ± 1.9%, n=3, vehicle, versus 5.4 ± 2.0%, n=3, corticosterone, P = 0.44), although differences did not reach statistical significance. These cells might be progenitor cells or cells undergoing cell-type conversion.

Figure 6: Effect of corticosterone on the number of pendrin-positive cells.

Adrenalectomized male mice were given vehicle or corticosterone at 200 or 800 μg/day (Treatment #4). Panel A shows low magnification images of sections from each group labeled for pendrin. The bar represents 50 μm. The ratio of pendrin-positive cell area to total cortical area of mice from each group is shown in Panel B. Panel C shows cortical sections from adrenalectomized mice that received vehicle or 800 μg/day corticosterone that were stained for pendrin (red), calbindin (green), AQP2 (white) and nuclei (blue). Panel C, left-most column, shows a typical DCT2. The absence of AQP2 label indicates it is a DCT2 rather than a CNT. Panel C, second column from the left, shows merged images of pendrin, calbindin, AQP2 and nuclei of a typical CNT. The right 2 columns show an image of the same CNT tubule with AQP2 and pendrin label shown individually in grayscale. Arrows show cells positive for both pendrin and AQP2, whereas arrowheads show cells positive for pendrin, but negative for AQP2. Bars represent 10 μm. Panel D shows the percentage of pendrin-positive cells within the DCT2 and CNT. *P<0.05 unpaired t test.

We conclude that corticosterone increases apical pendrin abundance through subcellular redistribution. Corticosterone also increases pendrin total protein abundance, but at higher doses does so primarily by increasing pendrin-positive cell number. The increase in pendrin-positive cell number that occurs in response to corticosterone is found in the DCT2 rather than the CNT.

Pendrin gene ablation eliminates the increase in blood pressure seen in response to stress doses of corticosterone.

Figures 7A & B show that in kidney lysates from adrenal-intact, male mice, pendrin band density was on average higher in those given 10 days of 800 μg/day corticosterone than in those given vehicle (Treatment #4), consistent with observations in adrenalectomized mice. However, differences in these adrenal-intact mice did not reach statistical significance. Since glucocorticoid administration increases blood pressure both in rodents and in people, we examined blood pressure in adrenal-intact, male pendrin null and wild type mice under basal conditions and then after 2 weeks of corticosterone (800 μg/day, Treatment #5). Under basal conditions, blood pressure was 125 ± 2 mm Hg in wild type mice but 115 ±2 in pendrin null mice (Figure 7C). These values are consistent with our previous observation that blood pressure measured by telemetry is ~ 8 mm Hg lower in pendrin KO than wild type mice under basal conditions (40). However, while corticosterone increased blood pressure by ~10 mm Hg in wild type mice, it either had no effect or reduced blood pressure in pendrin KO mice (115 ± 2 under basal conditions and 110 ±1 mm Hg following corticosterone treatment, P = 0.0565, Figure 7C). We conclude that pendrin gene ablation eliminates the increment in blood pressure seen in response to glucocorticoid administration.

Figure 7: Pendrin gene ablation eliminates the increase in blood pressure seen with corticosterone administration.

Panel A shows a typical immunoblot of adrenal-intact male mice given 10 days of vehicle or 800 μg/day corticosterone (Treatment #4). Panel B shows pendrin band density in kidney lysates from each group. Panel C shows blood pressure measured by tail cuff in adrenal-intact male wild type and pendrin null mice given a standard rodent diet and then 14 days of a NaCl-replete gelled diet supplemented with corticosterone (800 μg/day, Treatment #5). While blood pressure rose with corticosterone administration in wild type mice (P = 0.004), it was the same or lower (P = 0.0565) in the pendrin null mice following glucocorticoid administration. *P<005 paired t test. ^#P=0.0565.

DISCUSSION

We observed previously that IC MR gene ablation blunts the increase in pendrin abundance seen in treatment models associated with increased circulating aldosterone, such as dietary NaCl restriction or aldosterone administration (9, 27, 28). In principal cells, aldosterone activates the MR when cortisol/corticosterone is oxidized and thereby inactivated by 11 β HSD2 (8). However, while intercalated cell transporters such as pendrin are clearly upregulated by aldosterone and the IC MR, pendrin abundance is higher, not lower, in kidneys from 11 β HSD2 KO than in control rats following either a NaCl-restricted or a NaCl-replete diet. As such, enhanced glucocorticoid MR activation provides another potential explanation for the increased pendrin abundance seen in the 11 β HSD2 KO rats. Therefore, this study focused on pendrin’s regulation by glucocorticoid hormones.

We measured circulating aldosterone concentrations of 0.06 to 0.2 nM in adrenalectomized male mice, which are all well below the aldosterone-MR Kd of 0.5 nM reported in other cell types (31, 41, 42). The aldosterone concentrations we measured likely overestimate true circulating aldosterone concentration in corticosterone-treated adrenalectomized mice due to weak corticosterone interference in aldosterone RIA assays (43). This interference may have contributed to the slightly higher aldosterone concentration we measured in corticosterone- than in vehicle-treated adrenalectomized mice, although these differences in aldosterone concentrations did not reach statistical significance. As such, the lower pendrin abundance observed in adrenalectomized, corticosterone-treated IC MR KO mice relative to floxed controls cannot be attributed to disruption of aldosterone-MR binding. Moreover, because MR antagonists reduce pendrin abundance in mice that do not produce aldosterone, the IC MR can regulate pendrin independently of aldosterone. However, since the IC MR has a unique structure (44), we cannot exclude the possibility that the receptor’s apparent affinity for aldosterone or corticosterone is much different in ICs than in other cell types. If so, significant aldosterone binding might occur at very low mineralocorticoid hormone concentrations. In that case, aldosterone might readily bind to the IC MR without the need for corticosterone/cortisol oxidation by 11 β HSD2. The mechanism by which aldosterone upregulates pendrin requires further investigation.

While aldosterone and corticosterone are both MR ligands, their effects on pendrin expression differ. Aldosterone increases both pendrin total protein abundance as well as pendrin label per cell (27). At least at higher doses, however, corticosterone does not increase pendrin abundance per cell but rather increases pendrin total protein abundance in kidney by increasing pendrin positive cell number within the DCT2. Corticosterone may increase pendrin-positive cell proliferation. Alternatively, it may induce the interconversion of AQP2-positive to pendrin-positive cells or the differentiation of progenitor cells with the potential to generate mature cells in the late DCT2, the CNT and the CCD (45, 46). These progenitor cells are thought to be a cell subset that express both AQP2 as well as intercalated cell markers. While signaling molecules, such as Notch, are known to mediate this interconversion (39), an effect of steroid hormones on the abundance of these cell subsets, to our knowledge, has not been reported previously. The signaling cascade through which corticosterone facilitates this increment in pendrin-positive ICs remains the topic of future studies.

Whether the MR modulates the abundance of cell populations in the kidney or instead modulates its abundance per cell, remains to be determined. Since both the MR and the glucocorticoid receptor (GR) are expressed within ICs (47), corticosterone administration might regulate pendrin, in part, through its interaction with the IC GR. The role of the IC GR, however, also remains a topic of future investigation.

Human Cushing Syndrome is associated with excess cortisol. This cortisol excess produces hypertension, truncal obesity, moon facies, hirsutism, glucose intolerance and osteoporosis. It most commonly occurs from excess pituitary production of ACTH but also occurs from increased ectopic synthesis of cortisol, which can arise from adrenal tumors (48). The hypertension seen in Cushing Syndrome comes about from both increased renal NaCl absorption as well as increased renal vascular resistance (48). The increment in renal NaCl absorption and the resulting intravascular volume expansion seen in rodent models of Cushing Syndrome occurs, in part, through upregulation of both the epithelial sodium channel and the thiazide-sensitive NaCl cotransporter (49, 50). The present study demonstrates that pendrin is also critically important to the hypertension generated in rodent models of this disorder.

Cortisol and corticosterone are stress hormones regulated by clock genes making them important in blood pressure’s circadian rhythm (49). In Cushing’s syndrome, normal nocturnal blood pressure dipping is disrupted, which is associated with poorer cardiovascular outcomes. The present study did not explore the role of pendrin in glucocorticoid hormone-induced circadian changes in blood pressure. However, because mice were given corticosterone in their food, corticosterone consumption and the increased circulating concentration that follows its consumption should have occurred more at night when they are active (36, 51). This expected nighttime increase in corticosterone should roughly correspond to the natural circadian rhythm of this glucocorticoid hormone in adrenal-intact mice (36).

While glucocorticoids likely upregulate pendrin through a direct effect on ICs, they might instead upregulate pendrin by inducing respiratory alkalosis (52). As such, we cannot completely exclude a role of indirect mechanisms in pendrin’s response to these steroid hormones. However, the higher pCO₂ we saw in the vehicle- than the corticosterone-treated mice may reflect differences in anesthesia sensitivity.

It is well-established that pendrin is upregulated in models of metabolic alkalosis. Therefore, the alkalosis seen in NaCl-replete 11 β HSD2 KO rats likely contributes to the increase in pendrin abundance we observed (30). However, the greater pendrin abundance observed in male NaCl-restricted 11 β HSD2 KO relative to NaCl-restricted controls cannot be explained by changes in acid-base balance.

Following either a NaCl-replete or a NaCl-restricted diet, previous studies showed that serum K⁺ is lower in both 11 β HSD2 KO rats and in 11 β HSD2 heterozygous mice relative to their wild type controls (16, 53–55). This fall in serum K⁺ was attributed to greater ENaC subunit abundance and activity seen in these mutant rodents (53–55), which should increase the driving force for renal K⁺ secretion. The lower serum K⁺ of NaCl-restricted KOs relative to controls seen in females, but not in males, raises the possibility that ENaC’s response to 11 β HSD2 gene ablation is sex-dependent. This question remains the topic of future investigation.

We have shown previously that ENaC abundance and activity are lower in the IC MR null than in floxed MR mice. This fall in ENaC abundance and function occurs partly because ablation of the MR within pendrin-positive ICs reduces principal cell ENaC expression and function through paracellular signaling, independently of MR expression within principal cells (23, 56, 57). Thus, the higher serum K⁺ seen in adrenalectomized, corticosterone-treated IC MR KOs relative to floxed MR controls likely reflects lower ENaC activity in the former relative to the latter. However, the effect of pendrin gene ablation on ENaC in corticosterone-treated mice remains the topic of future studies.

We conclude that corticosterone increases pendrin abundance in the most apical region of the cell through subcellular redistribution. These steroid hormones also increase pendrin total protein abundance through a mineralocorticoid receptor-dependent pathway. At least at stress doses, corticosterone increases pendrin total protein abundance by increasing pendrin-positive cell number. In so doing, pendrin contributes to the pathogenesis of the hypertension seen in rodent models of Cushing Syndrome.