Mapping the Transcriptome Underpinning Acute Corticosteroid Action within the Cortical Collecting Duct

Abstract

Key Points

• We report the transcriptomes associated with acute corticosteroid regulation of ENaC activity in polarized mCCD_cl1 collecting duct cells.

• Nine genes were regulated by aldosterone (ALDO), 0 with corticosterone alone, and 151 with corticosterone when 11βHSD2 activity was inhibited.

• We validated three novel ALDO-induced genes, Rasd1, Sult1d1, and Gm43305, in primary cells isolated from a novel principal cell reporter mouse.

Background

Corticosteroids regulate distal nephron and collecting duct (CD) Na⁺ reabsorption, contributing to fluid-volume and blood pressure homeostasis. The transcriptional landscape underpinning the acute stimulation of the epithelial sodium channel (ENaC) by physiological concentrations of corticosteroids remains unclear.

Methods

Transcriptomic profiles underlying corticosteroid-stimulated ENaC activity in polarized mCCD_cl1 cells were generated by coupling electrophysiological measurements of amiloride-sensitive currents with RNAseq. Generation of a principal cell-specific reporter mouse line, mT/mG-Aqp2Cre, enabled isolation of primary CD principal cells by FACS, and ENaC activity was measured in cultured primary cells after acute application of corticosteroids. Expression of target genes was assessed by qRT-PCR in cultured cells or freshly isolated cells after the acute elevation of steroid hormones in mT/mG-Aqp2Cre mice.

Results

Physiological relevance of the mCCD_cl1 model was confirmed with aldosterone (ALDO)-specific stimulation of SGK1 and ENaC activity. Corticosterone (CORT) only modulated these responses at supraphysiological concentrations or when 11βHSD2 was inhibited. When 11βHSD2 protection was intact, CORT caused no significant change in transcripts. We identified a small number of ALDO-induced transcripts associated with stimulated ENaC activity in mCCD_cl1 cells and a much larger number with CORT in the absence of 11βHSD2 activity. Principal cells isolated from mT/mG-Aqp2Cre mice were validated and assessment of identified ALDO-induced genes revealed that Sgk1, Zbtbt16, Sult1d1, Rasd1, and Gm43305 are acutely upregulated by corticosteroids both in vitro and in vivo.

Conclusions

This study reports the transcriptome of mCCD_cl1 cells and identifies a small number of ALDO-induced genes associated with acute stimulation of ENaC, including three previously undescribed genes.

Introduction

Aldosterone (ALDO) and cortisol both influence blood pressure: ALDO as the final effector in the renin-angiotensin-aldosterone system, promoting Na⁺ reabsorption in the kidney; cortisol, in addition to being thought of as a “stress” hormone, being linked to the circadian rhythm of blood pressure. Within the kidney, ALDO acts specifically in the cells of the “aldosterone-sensitive distal nephron” (ASDN), where the mineralocorticoid receptor (MR) and the enzyme 11-β-hydroxysteroid dehydrogenase type 2 (11βHSD2) are both expressed.¹ 11βHSD2 converts circulating cortisol to the inactive metabolite cortisone, conferring ALDO specificity to the ASDN that has previously been demarked as the late distal convoluted tubule (DCT2), connecting tubule (CNT), and collecting duct (CD).¹ There is now increasing evidence that 11βHSD2 is absent in the DCT^2,3; thus, the boundaries of the ASDN may need updating to include only the CNT and cortical CD (CCD). Genetic mutations or pharmacological inhibition of 11βHSD2 cause hypertension in both humans⁴ and rodents,⁵ highlighting the importance of 11βHSD2 to blood pressure control. In addition, recent evidence suggests that in states of glucocorticoid excess (e.g., chronic stress, obesity), cortisol may cause hypertension because of aberrant activation of renal sodium transport. Causal mechanisms are not fully understood but activation of the epithelial sodium channel (ENaC) via MR and the glucocorticoid receptor (GR) have been implicated.⁶

The classic actions of corticosteroids in the kidney involve hormones binding cytosolic steroid receptors within the epithelial cells lining the ASDN, modulating transcriptional processes, and subsequent stimulation of Na⁺ reabsorption.⁷ This involves the transepithelial movement of Na⁺ back to the circulation, first crossing the apical membrane via the epithelial Na⁺ channel (ENaC) and subsequent extrusion across the basolateral membrane via the Na⁺/K⁺ ATPase. ENaC is rate-limiting in this process and is subject to regulation by a number of different hormones and bioactive factors.⁸ Elucidation of the transcriptional targets of corticosteroid action in the kidney has involved many studies over the past two decades, which have made use of a variety of model systems, e.g., whole kidney homogenate, microdissected tubules,⁹ primary cells isolated by FACS,^10,11 and a number of cell lines, ^12–14 using ever-improving technology, e.g., SAGE analysis, microarrays, and more recently next generation sequencing. Key transcripts have been identified that encode target proteins including the serum and glucocorticoid-induced kinase 1 (SGK1) and the glucocorticoid-induced leucine zipper protein (GILZ), both of which have been shown to regulate ENaC activity.^15–17 However, there are discrepancies regarding the relative roles of these steroid-induced genes, particularly considering the phenotypic difference in mice after the nephron-specific deletion of MR¹⁸ compared with deletion of SGK1¹⁹ or GILZ.²⁰ Furthermore, the high concentrations of corticosteroids used to identify relevant genes in various model systems leave the physiological relevance of these transcripts to steroid-induced ENaC activity within the distal nephron unclear.

The objective of this study was therefore to apply an unbiased, reductionist approach to generate transcriptomic profiles associated with the acute Na⁺ retaining effects of corticosteroids in a physiologically relevant murine cell model of the CCD, mCCD_cl1 cells. These display many of the reported features of Na⁺ absorbing principal cells of the CD: predominant amiloride-sensitive Na⁺ conductance via ENaC, a K⁺ conductance mediated via the renal outer medullary K⁺ channel (ROMK),²¹ functional 11βHSD2 activity,²² as well as regulation of these transport pathways by physiological concentrations of hormones including aldosterone and arginine vasopressin.^22,23 We report the full transcriptome of the mCCD_cl1 cells and identify nine ALDO-induced genes. Lack of transcriptional effects of corticosterone (CORT) confirms 11βHSD2 activity, and following pharmacological inhibition of this enzyme, CORT modulated 151 genes. Identified ALDO-induced genes were validated in mCCD_cl1 cells and subsequently in primary CD cells isolated from a newly generated mT/mG-Aqp2Cre reporter mouse. Identified ALDO-induced targets were subsequently measured in both cultured primary principal cells treated with steroids or from principal cells isolated directly from reporter mice after acute injection of corticosteroids. Our findings confirm both Sgk1 and Zbtb16 as acute ALDO-induced genes, and we further report Rasd1, Sult1d1 and the unannotated Gm43305 that encodes a lncRNA.

Methods

Culture of mCCD_cl1 Cells

mCCD_cl1 cells (Prof. Bernard Rossier, University of Lausanne, Switzerland) were maintained in routine culture^22,24 at 37°C and 5% CO₂, in phenol-red free DMEM/F12 media supplemented with FBS (2%), triiodothyronine (1 nmol/L), sodium selenite (5 ng/ml), insulin (5 µg/ml), transferrin (5 µg/ml), L-glutamine (200 mmol/L), penicillin (100 U/ml), streptomycin (100 µg/ml), dexamethasone (DEX, 50 nmol/L), and epidermal growth factor (EGF, 10 ng / ml). Cells were used between passages 29 and 35. For experiments, cells were seeded onto 0.4 µm polyester filter membranes (Corning CoStar Snapwells) and grown for 9–11 days with the medium exchanged every 2 days. 48 hours before experimental protocols, the medium was replaced with DMEM/F12 medium supplemented with charcoal-stripped FBS (2%), penicillin (100 U/ml), and streptomycin (100 µg/ml). For the final 24 hours, the medium was replaced with DMEM/F12 with penicillin (100 U/ml) and streptomycin (100 µg/ml).

Quantification of Transepithelial Ion Transport

Transepithelial voltage (V_te) and resistance (R_te) were measured using an epithelial volt-ohm-meter with chopstick “STX” electrodes (World Precision Instruments, Hertfordshire, United Kingdom). The equivalent short circuit current (I_eq) was subsequently calculated by Ohm's Law. V_te is shown relative to an earth electrode in the basolateral bath;²⁵ a negative I_eq therefore reflects either apical to basolateral movement of cations, basolateral to apical movement of anions, or some combination of the two. Amiloride (10 µM, 10 minutes) was applied to the apical bath to determine the ENaC-mediated proportion of I_eq.

RNA Sequencing

Total RNA was extracted from cells using the RNeasy kit (Qiagen) after electrophysiological measurements. RNA was subject to quality control (RNA ScreenTape), and RIN values were ≥9.0. TruSeq-stranded mRNA-seq libraries were generated from each total RNA sample (24 libraries in total). Libraries were sequenced using the Illumina HiSeq 4000 Platform with 150 base pair paired-end reads. Reads were trimmed for quality (Cutadapt version 1.121) at the 3′ end using a threshold of Q30 and for adapter sequences of the TruSeq stranded mRNA kit (AGATCGGAAGAGC) using Cutadapt version 1.121.²⁶ Reads after trimming were required to have a minimum length of 50. The raw RNAseq data have been uploaded to the Sequence Read Archive at NCBI, with the project ID: PRJNA820455. The reference used for mapping was the Mus musculus genome from Ensembl, assembly GRCm38, annotation version 84. Reads were aligned to the reference genome using STAR2 version 2.5.2b.²⁷ The raw counts' table was filtered to remove rows consisting of predominantly near-zero counts (values had to be >0.05 across five samples), filtering on counts per million to avoid artifacts caused by library depth. Initial exploratory analysis indicated an outlier in the dataset, which was removed, and subsequent filtering and normalization were performed again for downstream analysis. After filtering, 17,962 genes remained. Differential analysis was carried out with EdgeR²⁸ (version 3.16.5) and compared with all possible combinations from the four experimental conditions.

Generation of a Principal Cell-Specific Reporter Mouse

Aqp2Cre mice²⁹ (The Jackson Laboratory) were crossed with mT/mG (tdTomato-GFP) mice³⁰ (Prof. Neil Henderson, The University of Edinburgh, United Kingdom). Both mouse strains were on a C57BL/6 background. To ensure kidney-specific Cre expression, female AQP2-Cre mice were bred with male mT/mG mice; female offspring, which were heterozygous for both AQP2-Cre and mT/mG, were subsequently bred with male mice homozygous for mT/mG. Genotyping for AQP2-Cre was carried out using a forward primer 5′-CTCTGCAGGAACTGGTGCTGG-3′ and reverse primer 5′-GCGAACATCTTCAGGTTCTGCGG-3′. Genotyping for mT/mG was carried out using the forward primer 5′-CTCTGCTGCCTCCTGGCTTCT-3′ and reverse primers: wildtype 5-CGAGGCGGATCACAAGCAATA-3′ and mutant 5′-TCAATGGGCGGGGGTCGTT-3′. Experiments were performed on male mice aged 10–30 weeks, under the authority of a UK Home Office Project License and after approval by the University's Animal Welfare and Ethical Review Board.

Isolation and Culture of Primary Principal Cells

Mice (wild-type, mT/mG^+/+- Aqp2Cre^−/− and mT/mG^+/+- Aqp2Cre^+/−) were terminated by rising CO₂ and PBS injected into the left ventricle to remove blood. Kidneys were excised, decapsulated, and stored in ice-cold PBS. Both kidneys were manually chopped into small pieces and then homogenized in gentleMACS C Tubes (Miltenyi Biotec, Surrey, United Kingdom).³¹ Digestion buffer contained the following: RPMI supplemented with Collagenase V (0.425 mg/ml), Collagenase D (0.625 mg/ml), Dispase II (1 mg/ml), DNase I (30 µg/ml), penicillin (100 U/ml), and streptomycin (100 µg/ml). Cellular suspensions were digested for 30 min at 37°C before a second dissociation in the gentleMACS Dissociator. An equal volume of neutralization buffer (PBS supplemented with FBS 2% vol/vol and EDTA 1 µM) was added, and cell suspensions were passed sequentially through 100 µm, 70 µm, and 40 µm cell strainers. Cells were pelleted by centrifugation, and red cell lysis was carried out using red blood cell lysis buffer (Sigma, Dorset, United Kingdom). Cells were pelleted and resuspended in 1 ml neutralization buffer.

Cells were analyzed and sorted using either a BD FACS Aria II or BD FACS Aria Fusion cell sorter. 405 nm, 488 nm, and 561 nm lasers were used for the excitation of DAPI, GFP, and tdTom, respectively. Wild-type C57BL/6 and mT/mG^+/+-Aqp2Cre^−/− kidney samples were processed first to define gates before processing mT/mG^+/+-Aqp2Cre^+/− samples. DAPI, at a final concentration of 0.1 μg/ml, was added to cells immediately before sorting. Cells were gated for a stable recording, singlets (plotting forward scatter area versus height), cells (forward scatter area versus side scatter area), and live cells (DAPI versus forward scatter area), followed by exclusive gates for both tdTom and GFP (Supplementary Figure 3). From the GFP gate, a further gate was added to remove autofluorescence events detected within that channel and a final gate to remove a mixed population of both tdTom and GFP, enabling collection of a “GFP+” population. Cell sorting was performed with a 100 µm nozzle, and because of the starting cell numbers and the relatively small percentage of GFP positive events, a yield sort followed by a purity sort was used to optimize sort/time efficiency. For initial validation studies, 100,000 tdTom cells and all possible GFP cells were collected. Once the GFP+ population was validated, only GFP+ events were collected. Flow cytometry data were analyzed using FCS Express 7 (De Novo Software, Pasadena, CA).

For downstream experiments, GFP+ sorted cells were then spun down, supernatant removed, and RLTplus buffer added for RNA extraction or instead directly plated onto gelatin-coated 12 well plates. Cells were maintained under identical conditions as those used for culture of mCCD_cl1 cells with the addition of gelatin coating of either the initial 12-well plate or the permeable membrane used for growing polarized monolayers.

Acute Steroid Treatment of mT/mG-Aqp2Cre Mice

Male mT/mG^+/+-Aqp2Cre^+/− mice, aged 10–30 weeks, were administered carbenoxolone (CBX) at 2.5 mg/kg BW per day by mouth (in drinking water) to inhibit endogenous 11βHSD2,³² control animals were given ad lib access to drinking water for 8 days. Animals were weighed every morning (between 0800 and 1000 hours) 5 days before CBX/control treatment and subsequently throughout. On day 9 at 08:30, mice were administered with a single dose of steroid: ALDO (10 µg/kg BW),³³ CORT (0.5 mg/kg BW), or solvent vehicle (5% EtOH) via ip injection. After 3 hours, mice were sacrificed by rising concentration of CO₂ and CD cells (GFP+) were then isolated by FACS.

Immunofluorescence Imaging

Kidneys were decapsulated, bisected (left kidney: longitudinal section and right kidney: transverse section), and immersed in MeOH-free 4% PFA for 2 hours at 4°C. Half kidneys were then washed twice with PBS, transferred to an 18% sucrose solution at 4°C overnight, and then embedded in OCT, frozen, and sectioned at 10 µm onto glass slides. Sections were permeabilized with 0.2% Triton X-100 for 10 minutes, blocked with 10% donkey serum for 1 hour, and subsequently washed with TBS-T (0.05% Tween 20). Sections were mounted in ProLong Diamond Antifade Mountant (Life Technologies, Paisley, United Kingdom) and imaged at ×40 using a Zeiss AxioScan.Z1.

All images were processed using Fiji,³⁴ and the same process was applied to images from both mT/mG^+/+- Aqp2Cre^−/− and mT/mG^+/+- Aqp2Cre^+/− kidneys. Channels were split, and a background subtraction (50 pixels) was performed on the 555-nm channel. For low magnification images, the 488-nm channel image was duplicated and a Gaussian blur (σ=10) applied. To separate the Cre-GFP signal from the tubule-generated autofluorescence, the blurred image was subtracted from the raw image and a threshold applied such that on the Cre-GFP remained (approximately 0.04%). This was used to generate a mask, which was changed to 16-bit and a Gaussian blur (σ=7) was applied. This was used as the green channel and the autofluorescence as the gray channel. For high magnification images, the 488-nm channel image was duplicated and a Gaussian blur (σ=3) was applied. The blurred image was again subtracted from the raw image and a threshold applied such that only the Cre-GFP remained (approximately 0.01%). This was used to generate a mask, which was changed to 16-bit, and a Gaussian blur (σ=2) was applied and used as the green channel.

qRT-PCR

Total RNA was extracted from cells, either grown as monolayers in culture on filter membranes or directly following FACS of CD cells, using a QIAGEN RNeasy Plus Micro Kit. The integrity of the RNA preparations was verified using the Agilent RNA 6000 Pico Kit and Agilent 2100 Bioanalyser, and samples with RIN values <7.5 were excluded. cDNA was transcribed from total RNA using Applied Biosystems High Capacity cDNA Reverse Transcription Kit (Life Technologies). For experiments using lysates from cells grown in culture, 500 ng RNA was used and cDNA was diluted 1:20 to correlate to the middle of the seven-point calibration curve generated from serial dilutions. For experiments using isolated primary CD cells from mT/mG-AQP2Cre mice, 1 ng RNA was used, as determined by the lowest yield across the samples. cDNA was diluted 1:10 to correlate with the middle of the calibration curve. qRT-PCR was performed using a Roche Light-Cycler 480 II using a probe-based assay (Roche Universal Probe Library, Sigma, Dorset, United Kingdom). Primers were designed using the ProbeFinder software within the Roche Assay Design Center. Samples were run in triplicate, and only Cq values with a standard deviation >0.3 were excluded. A selection of reference genes (Actb1, HPRT, Tbp, and 18S) was tested and included if expression remained unaltered across all samples (Supplementary Figures 2, 4, and 5). Negative controls included only reverse transcriptase negative, RNA negative, and H₂O.

Western Analysis

Cells were washed with ice-cold PBS (×3), lysed in lysis buffer,²⁵ and vortexed. Protein concentration was quantified by Bradford assay (BioRad, Hertfordshire, United Kingdom). Samples were prepared by adding a fixed mass of protein lysate to the sample buffer, reducing and denaturing by heating at 95°C for 5 minutes in the presence of 2-mercaptoethanol. Samples were subsequently fractionated on 10% SDS polyacrylamide gels, transferred to PVDF membranes, blocked, and probed with primary antibodies of interest and respective horseradish peroxidase-linked secondary antibodies. Primary antibodies against Thr^346/356/366-phosphorylated and total forms of the protein encoded by n-myc downstream regulated 1 (NDRG1), as well as total serum and glucocorticoid-regulated kinase 1 (SGK1), were purchased from the Dundee Protein Phosphorylation Unit, University of Dundee (Dundee, United Kingdom). The antibody against β-actin was from Sigma (Dorset, United Kingdom). Immunoreactive proteins were visualized by enhanced chemiluminescence and quantified by densitometric measurements, as described previously.³⁵

Data Analysis

Data are expressed as mean±95% confidence interval (CI). For western blotting and qRT-PCR, because of the uneven distribution of data expressed as fold-change, all data were log-transformed. All datasets were subject to normality testing (Shapiro-Wilks) followed by either parametric testing (unpaired t test or one-way or two-way ANOVA) or nonparametric testing (Mann-Whitney or Kruskal-Wallis test), where appropriate. Post hoc analysis was also carried out, where appropriate, and the details of specific tests used are included in the figure legends.

Results

Modulation of ENaC and SGK1 Activity by Corticosteroids in mCCD_cl1 Cells

Polarized mCCD_cl1 cells generated an average transepithelial voltage (V_te) of 22.8±11.3 mV and resistance R_te of 1.6±0.6 kΩ·cm², giving an average equivalent short-circuit current (I_eq) of 13.7±3.4 µA·cm⁻² (n=144). I_eq reflects ENaC-mediated Na⁺ transport as approximately 95% is inhibited by amiloride (10 µM).²⁴ CORT stimulated the amiloride-sensitive current (I_ami-3h) at concentrations ≥100 nM, consistent with the endogenous activity of the “protective” 11βHSD2 (Figure 1Ai). Inhibiting 11βHSD2 with CBX (10 µM, 30 minutes) revealed a concentration-dependent stimulation of ENaC-mediated Na⁺ transport by CORT at concentrations ≥1 nM (Figure 1Ai). Baseline current was not altered in cells treated with CBX alone. ALDO also stimulated I_ami in a concentration-dependent manner; this was independent of CBX pretreatment, consistent with ALDO not being a substrate for this enzyme (Figure 1Bi). CORT-induced Na⁺ transport correlated with increased activity and expression of the protein serum and glucocorticoid-induced kinase 1 (SGK1), Figure 1Aii and iii. SGK1 abundance under basal conditions is very low;³⁶ however, there is clear activity of this kinase as per the basal levels of phosphorylation of specific residues in a downstream target NDRG1. Both SGK1 activity and expression were increased with ALDO (Figure 1Bii and iii).

Figure 1.

Effects of acute corticosteroids on ENaC-mediated transport and SGK1 activity and expression in mCCD_cl1 cells. Amiloride-sensitive currents measured from polarized mCCD_cl1 cells after 3 hours (I_ami-3h) exposure to increasing concentrations of corticosterone (Ai) or aldosterone (Bi) in the absence (−CBX, solid line) or presence (+CBX, dashed line) of CBX (10 µM, 30 minutes preincubation) to inhibit 11βHSD2 activity. H₂O was used as vehicle control for CBX, and amiloride (10 µM, 10 minutes) was added after corticosteroid treatment. The activity (Aii) and expression (Aiii) of SGK1 was determined by measuring the abundance of P-NDRG1 or T-SGK1, respectively, in cell lysates after corticosteroid treatment and electrophysiological measurements. These are expressed relative to T-NDRG1 and β-actin, respectively. Upper panels show representative blots, and lower panels show densitometric data expressed as a fold change from the control (0) lane (log₁₀). Data are shown as mean ±95% confidence interval (95% CI) (n=7). Statistical significance was determined by a two-way ANOVA for the electrophysiological data, with a Dunnett's post hoc test, and a one-way ANOVA for Western blot data, because of separate blots for − and + CBX, with a Dunnett's post hoc test. CORT, corticosterone; CBX, carbenoxolone.

Mapping the Transcriptomes Associated with Acute Corticosteroid Stimulation of ENaC-Mediated Na⁺ Transport

Polarized mCCD_cl1 cells were treated for 3 hours with solvent vehicle, ALDO (3 nM) or CORT (100 nM), the latter in the absence or presence of CBX (10 µM). Consistent with the concentration response assays ALDO, or CORT in the presence of CBX, stimulated I_ami compared with control, whereas CORT in the absence of CBX did not alter I_ami (Figure 2A). cDNA libraries were generated across each of the four groups (n=6) that underwent 150 bp paired-end sequencing. Over 94% of trimmed reads were mapped to the genome, and of those, 97.0%–98.4% were mapped as pairs.

Figure 2.

Differentially expressed transcripts in corticosteroid-treated mCCD_cl1 cells. (Ai) I_eq measured from cells treated with either solvent vehicle (Veh), 3 nM Aldo, or 100 nM Cort for 3 hours; arrow indicates addition of corticosteroid. An additional CORT group was pretreated CBX (10 µM, 30 minutes), all other groups received vehicle control for this period. Amiloride (10 µM) was added to the apical bath for a final 10 minutes, and Aii shows the amiloride-sensitive current (I_ami-3h). Data are shown as mean ±95% confidence interval (CI) (left panel) and as individual points and mean±95% CI (right panel). Statistical significance was determined by one-way ANOVA with a Tukey's post hoc test used to compare groups to vehicle control, **P<0.01 and ***P<0.001. (B) Differentially expressed transcripts are plotted as log₂ fold change versus −log₁₀ false discovery rate (FDR). The horizontal dashed line represents the specified FDR threshold (0.05) and the vertical dashed lines indicate the specified fold change threshold (2) in both positive and negative directions. Each treatment, (Bi) ALDO, (Bii) CORT or (Biii) CBX+CORT, is compared with vehicle-treated control. Points passing only the FDR threshold are shown in green, passing only the fold change threshold are shown in blue and those passing both thresholds are shown in red. Individual transcripts passing both thresholds in ALDO-treated cells are labeled.

Differential gene expression analysis compared all possible contrasts of experimental conditions using thresholds of a minimum log₂ fold change of one and a false discovery rate <0.05 (Table 1). Volcano plots were generated and show the differential expression of genes in cells treated with vehicle compared with ALDO, CORT, or CBX+CORT (Figure 2B). Nine genes were identified in the ALDO group compared with the control group, and 151 genes were differentially expressed in the CBX+CORT group compared with the control group (Figure 2B). Table 2 lists all transcripts regulated by ALDO, and Table 3 lists the top 15 upregulated and all downregulated annotated genes in the CBX+CORT group. No transcripts were differentially expressed in the CORT group, and this finding correlates with a lack of stimulated I_ami. All nine transcripts that were differentially expressed in the ALDO versus control group were also differentially expressed in the CBX+CORT group versus control. The complete list of differentially regulated genes, as well as a counts table across the four experimental groups, can be found in the Supplementary Excel File. Five of the genes identified (Sgk1, Sult1d1, Gm43305, Rasd1, and Zbtb16) were subsequently validated by qRT-PCR in polarized mCCD_cl1 cells (Supplementary Figure S1). Cells were treated in a similar manner: CORT (100 nM) or vehicle for 3 hours, after preincubation with CBX (10 µM) or vehicle for 30 minutes, or treated with ALDO (3 nM) or vehicle for 3 hours. Electrophysiological measurements to monitor ENaC activity were made before RNA extraction (data not shown).

Table 1.

Differential analysis of gene expression after acute corticosteroid treatment of mCCD_cl1 cells

Experimental Conditions Compared	Upregulated Genes	Downregulated Genes
Vehicle versus ALDO	8	1
Vehicle versus CORT	0	0
Vehicle versus CBX + CORT	123	78
ALDO versus CBX + CORT	64	7
CORT versus CBX + CORT	82	22

Number of genes up or downregulated in each contrast made of experimental condition according to the thresholds on minimum log₂ fold change (1) and maximum false discovery rate (0.05).

ALDO, aldosterone; CORT, corticosterone; CBX, carbenoxolone.

Table 2.

Differentially expressed genes in mCCD_cl1 cells after aldosterone treatment

Gene Symbol	Gene Name	Log2FC	FDR
Upregulated genes (Veh versus Aldo)
Zbtb16	Zinc finger and BTB domain-containing 16	3.2	6.0×10−12
Sgk1	Serum and glucocorticoid-regulated kinase 1	2.9	0.0001
Tslp	Thymic stromal lymphopoietin	2.1	0.0477
Rasd1	Ras-related dexamethasone-induced 1	2.1	9.4×10−10
Gm16178	NA	1.6	0.0007
Sult1d1	Sulfotransferase family 1D, member 1	1.1	3.2×10−14
Gm43305	NA	1.1	0.0002
Defb1	Defensin β 1	1.0	0.0005
Downregulated transcripts (Veh versus Aldo)
Gm9694	NA	−3.0	0.0444

The threshold for false discovery rate (FDR) was 0.05, and the Log₂ fold change (Log₂FC) threshold was 1. NA, not available.

Table 3.

Differentially expressed genes in mCCD_cl1 cells after corticosterone treatment in the absence of 11βHSD2 activity

Gene Symbol	Gene Name	Log2FC	FDR
Top 15 upregulated annotated genes (Veh versus CBX+cort)
Zbtb16	Zinc finger and BTB domain-containing 16	5.7	7.4×10−18
Sgk1	Serum and glucocorticoid-regulated kinase 1	5.6	3.1×10−8
Rasd1	Ras-related dexamethasone-induced 1	4.2	7.0×10−19
Hif3a	Hypoxia inducible factor 3, α subunit	2.7	1.1×10−14
Sult1a1	Sulfotransferase family 1A, phenol-preferring, member 1	2.7	0.0005
Sult1d1	Sulfotransferase family 1D, member 1	2.6	3.0×10−10
Tsc22d3	TSC22 domain family, member 3	2.5	1.2×10−8
Myom2	Myomesin 2	2.5	0.0031
Htr6	5-Hydroxytryptamine receptor 6	2.4	0.0198
Slco4c1	Solute carrier organic anion transporter family, member 4C1	2.3	0.0012
Tekt4	Tektin 4	2.2	4.1×10−5
Per1	Period circadian clock 1	2.1	5.3×10−6
Abcb5	ATP-binding cassette, subfamily B (MDR/TAP), member 5	2.1	2.4×10−11
Adora2b	Adenosine A2b receptor	2.1	6.4×10−9
Arg2	Arginase type II	2.1	1.3×10−12
All downregulated annotated genes (Veh versus CBX+cort)
Hao2	Hydroxyacid oxidase 2	−2.2	6.8×10−6
Lipc	Lipase, hepatic	−1.8	0.0210
Il1f6	Interleukin 1 family, member 6	−1.7	0.0012
Mboat4	Membrane bound O-acyltransferase domain containing 4	−1.5	0.0026
Pabpn1l	Poly(A) binding protein nuclear 1-like	−1.4	8.5×10−6
Sprr2g	Small proline-rich protein 2G	−1.3	0.0348
Prl2c5	Prolactin family 2, subfamily c, member 5	−1.3	0.0305
Gimap1	GTPase, IMAP family member 1	−1.2	0.0364
Id4	Inhibitor of DNA binding 4	−1.2	0.0276

The threshold for false discovery rate (FDR) was 0.05, and the log₂ fold change (Log₂FC) threshold was 1. For clarity, only annotated genes are shown, and a full list can be found in the supplemental excel file. cort, corticosterone; CBX, carbenoxolone.

Generation and Validation of a Principal Cell-Specific Reporter Mouse

The mT/mG mouse line,³⁰ which ubiquitously express tdTomato (tdTom) in cell membranes (mT) or following Cre excision express enhanced GFP (mG), was crossed with the Aqp2Cre line.²⁹ Fixed longitudinal sections of kidneys from adult offspring were imaged for tdTom and GFP labeling. mT/mG positive and Aqp2Cre null mice, e.g., mT/mG^+/+-Aqp2Cre^−/−, demonstrated membrane labeling of tdTom in both cortex and medullary regions (Figure 3Ai). Distinct differences in labeling can be seen in the cortex where brush border membranes of the proximal tubules exhibit less evenly distributed membrane-associated markers than distal tubules and CDs (Figure 3Aii). This is consistent with the expression pattern previously reported in kidney tissue.²⁹ Negligible eGFP labeling was detected (Figure 3Aiii). Kidney sections from adult mice homozygous for mT/mG and hemizygous for Aqp2Cre, e.g., mT/mG^+/+-Aqp2Cre^+/− mice (B), displayed both tdTom and Cre-induced eGFP labeling in the cortex and medulla (Figure 3Bi–iii). The eGFP labeling of tubules is low, possibly indicating low recombination efficiency of the Aqp2Cre line.

Figure 3.

tdTom and GFP expression in kidney sections from mT/mG-Aqp2Cre mice. Longitudinal sections of fixed kidneys from adult mT/mG^+/+-Aqp2Cre^−/− mice (A) and mT/mG^+/+-Aqp2Cre^+/− mice (B) showing tdTom and GFP labeling. (i) Tiled images taken with ×40 objective across the section with 488-nm (gray, autofluorescence to show tissue morphology; green, GFP signal) and 555-nm (red, tdTomato signal) excitation light. (ii) Cortex and medullary regions are shown at higher magnification, regions of interest denoted in (i) by cyan and yellow boxes, respectively. (iii) Individual 555-nm and 488-nm channels for (ii) are shown. Scale bars: (i) 500 µm and (ii–iii) 50 µm.

Expression of nephron segment-specific markers was determined in tdTom and GFP+ populations (Table 4). While the tdTom population was enriched for markers consistent with the proximal tubule (NHE3) and the loop of Henle (NKCC2), the GFP+ population was enriched for markers of the CD. In particular, principal cell markers including ROMK, α-ENaC, and 11βHSD2, but also β-intercalated cell markers, V-ATPase β1 and pendrin, but not the α-intercalated cell marker, AE1 (Table 4). NCC expression was detected in both populations.

Table 4.

Validation of isolated primary collecting duct cells by qRT-PCR

Tubule-Specific Target	Symbol	Nephron Segment	tdTom Population		GFP Population		Cq Range of Standards	Population Enriched?
			Cq	Expression of GOI Relative to Reference Genes (log10)	Cq	Expression of GOI Relative to Reference Genes (log10)
NHE3	Slc9a3	Proximal tubule	30.6±1.2	0.56±0.47	34.4±0.5	−0.91±0.34***	29.3–34.7	tdTom
NKCC2	Slc12a1	Loop of Henle	30.3±1.0	−0.03±0.18	32.3±0.8	−3.00±3.81**	26.9–32.4	tdTom
NCC	Slc12a3	Distal convoluted tubule	30.4±1.3	−0.47±0.10	29.0±0.6	−0.33±0.13	25.1–30.7	—
11βHSD2	Hsd11b2	CD—principal cell	BLD	BLD	25.8±0.7	0.09±0.11	23.6–29.1	GFP
α-ENaC	Scnn1a	CD—principal cell	BLD	BLD	31.2±0.8	−0.10±0.18	29.1–32.8	GFP
ROMK	Kcnj1	CD—principal cell	BLD	BLD	33.7±0.6	−0.02±0.16	32.4–35.1	GFP
Pendrin	Slc26a4	CD—β-intercalated cell	BLD	BLD	30.5±0.8	−0.02±0.14	27.4–33.1	GFP
AE1	Slc4a1	CD—α-intercalated cell	35.4±0.9	−0.48±0.38	34.8±0.7	−0.33±0.29	29.7–35.7	—
V-ATPase β1	Atp6v1b1	CD—α-and β-intercalated cell	BLD	BLD	32.2±0.6	−0.06±0.14	28.4–34.2	GFP
UT-A1	Slc14a2	Medullary CD	BLD	BLD	31.9±0.8	0.20±0.22	29.1–35.2	GFP

Reference gene	Symbol		tdTom Population		GFP Population		Cq Range of Standards	Population Enriched?
			Cq	Expression (log10)	Cq	Expression (log10)
β-actin	Actb1	Reference gene	28.7±1.1	−1.13±0.43	27.6±0.7	−0.68±0.27	26.1–30.9	—
18S	Rn18s	Reference gene	15.7±1.4	−1.12±0.46	15.6±0.7	−1.13±0.36	13.1–18.2	—

Expression of nephron segment-specific genes in either tdTom-labeled (n=5) or GFP-labeled (n=5) populations of cells isolated by FACS. For each gene of interest (GOI), both the Cq (cycle quantification value) and transcript expression relative to the average expression of reference genes (log10) are shown. The range of Cq values detected across the seven-point standards is also shown. The population where the GOI was found to be enriched—determined either by statistical significance of the relative expression (log10) or where one population showed expression within the seven point standard and the other was below the limit of detection (BLD)—is highlighted in the final column. A hyphen denotes no difference in gene expression between tdTom and GFP populations. Statistical significance was determined by unpaired t test. CD, collecting duct; ROMK, renal outer medullary K⁺ channel. **P<0.01, ***P<0.001.

To determine functional properties, one isolated population was grown in culture over several weeks. Cells were subsequently seeded onto permeable inserts; after 9–11 days, baseline V_te and R_te were −20.3±3.1 mV and 5.1±0.1 kΩ·cm², respectively, giving rise to an I_eq of −3.9±0.6 µA·cm⁻² (values are mean±95% CI, n=25). It was noted that over numerous passages V_te reduced, R_te increased and thus I_eq decreased (data not shown). Cells were therefore used between passages 4 and 8 for all experiments. Similar to the mCCD_cl1 cells, application of amiloride (10 µM, 10 minutes) inhibited basal I_eq to negligible values, indicating a predominant amiloride-sensitive current in the primary principal cells. A concentration-response assay to CORT±CBX revealed that CORT stimulated I_ami-3h at concentrations >100 nM and preincubating primary principal cells with CBX (10 µM, 30 min) unveiled a concentration-dependent stimulation of I_ami-3h (Figure 4Ai). Stimulation of ENaC-mediated Na⁺ transport correlated with increased activity of SGK1 activity, as determined by phosphorylation of NDRG1-Thr^346/356/366 (Figure 4Aii).

Figure 4.

Effects of acute addition of CORT on ENaC-mediated transport and SGK1 activity and expression in cultured primary principal cells. (i) Amiloride-sensitive currents were determined in primary principal cells after 3h (I_ami-3h) exposure to increasing concentrations of CORT in the absence (−CBX) or presence (+CBX) of CBX (10 µM, 30 minutes preincubation) to inhibit 11βHSD2 activity. H₂O was used as vehicle control for CBX, and amiloride (10 µM, 10 minutes) was added after corticosteroid treatment. (ii) The activity of SGK1 was determined by measuring the abundance of P-NDRG1 in cell lysates after corticosteroid treatment and electrophysiological measurements, expressed relative to T-NDRG1. Upper panels show representative blots, and lower panels show densitometric data expressed as a fold change from the control (0) lane (log₁₀). Data are shown as mean±95% confidence interval (CI) (n=7). Statistical significance was determined by a two-way ANOVA for the electrophysiological data and a one-way ANOVA for Western blot data, because of separate blots for − and + CBX, with a Dunnet's post hoc test, ***P<0.001.

Corticosteroid Regulation of ENaC Activity and Identified ALDO-Induced Genes in Cultured Primary Principal Cells

Primary principal cells were treated (3 hours) with CORT (10 nM) in the absence/presence of CBX (10 µM, 30 minutes preincubation), ALDO (3 nM), or DEX (100 nM). Consistent with the concentration-response assays, CORT only stimulated I_ami-3h when cells were preincubated with CBX: 2.4±0.3 fold versus without CBX: 1.2±0.2 fold (Figure 5A, n=8). Both ALDO and DEX stimulated I_ami-3h by 2.0±0.3 fold (Figure 5B, n=8) and 3.6±fold (Figure 5C, n=8), respectively. Of the eight targets tested: Sgk1, Sult1d1, Gm43305, Rasd1, Zbtb16, Defb1, Gm16178, and Gm9694, expression of three were upregulated by CBX+CORT (10 nM)—Sgk1, Gm43305, and Zbtb16. No targets were altered either by CORT or CBX alone (Figure 6). In cells treated with ALDO, expression of three out of eight target transcripts tested was increased: Sgk1, Rasd1, and Zbtb16 (Figure 7A). Finally, in cells treated with DEX, all but one target tested (Gm9694) was upregulated (Figure 7B).

Figure 5.

Corticosteroids stimulate ENaC-mediated Na⁺ transport in cultured primary principal cells. I_eq was measured across polarized monolayers of primary principal cells after treatment with corticosteroids. (A) Cells were preincubated with H₂O (−CBX, left) or 10 µM CBX (+CBX, right) for 30 minutes. Cort (10 nM, dashed line) or solvent vehicle (solid line) was subsequently added for 3 hours; arrow denotes addition. (B) Aldo (3 nM, dashed line) and (C) DEX (100 nM, dashed line), or respective solvent vehicle (solid line), were added to cells for 3 hours. In all experiments, amiloride (10 µM) was subsequently applied for 10 minutes. Data shown in traces (Ai, Bi, Ci) are mean I_eq±95% confidence interval (CI) and in bar graphs (Aii, Bii, Cii) as individual points and mean I_ami-_3h±95% CI (n=8). Statistical significance in (A) was determined by two-way ANOVA and Tukey's post hoc test and in (B) and (C) by unpaired t test, ***P<0.001.

Figure 6.

Expression of identified corticosteroid-induced transcripts in primary principal cells treated with CORT. Cells were treated with CBX or solvent vehicle for 30 minutes before addition of Cort (10 nM) or solvent vehicle for a further 3 hours. Amiloride (10 µM) was added for a final 10 minutes. Transcript expression of GOI is relative to the average expression of reference genes (log₁₀): Actb1 and Hprt. Data are shown as individual points and mean ±95% confidence interval (CI), and GOI is indicated in bold italics above each graph. Statistical significance was determined by one-way ANOVA with Tukey's post hoc test or Kruskal-Wallis test with Dunn's post hoc test, where appropriate: *P<0.05, **P<0.01, and ***P<0.001.

Figure 7.

Expression of identified corticosteroid-induced transcripts in primary principal cells treated with ALDO or DEX. Polarized primary principal cells were treated with either (A) Aldo (3 nM) or (B) Dex (100 nM) for 3 hours. Amiloride (10 µM) was added for a final 10 minutes. Transcript expression of GOI is relative to the average expression of reference genes (log₁₀): Rn18S and Tbp (Aldo) or Rn18s, Tbp, and Hprt (Dex). Data are shown as individual points and mean ±95% confidence interval (CI) and GOI is indicated in bold italics above each graph. Statistical significance was determined by unpaired t test or Mann-Whitney test, where appropriate: **P<0.01 and ***P<0.001.

Corticosteroid Regulation of Identified ALDO-Induced Genes in Principal Cells Isolated from mT/mG-Aqp2Cre Mice

We measured the expression of six ALDO-induced targets, Sgk1, Sult1d1, Gm43305, Rasd1, Zbtb16, and Defb1, in isolated primary principal cells after acute treatment with steroid hormones or respective controls. Of these, four were upregulated in mice treated with ALDO: Sult1d1, Gm43305, Rasd1, and Zbtb16 (Figure 8A). These four transcripts were also upregulated in mice treated with CBX+CORT. Notably, Sult1d1, Gm43305, and Zbtb16 were also upregulated in the CORT group but also in the CBX group.

Figure 8.

Steroid-induced transcript expression in isolated primary principal cells after acute injection of corticosteroids. Expression of selected target genes were quantified in cells isolated from mT/mG-Aqp2Cre mice 3 hours after ip injection of solvent vehicle (Ctl), Aldo, Cort, after 8 days ad libitum access to H₂O. Two further groups were concomitantly treated with CBX (2.5 mg/kg BW per day by mouth) for 8 days with subsequent ip injection of solvent vehicle (CBX) or CORT (CBX+Cort) for 3 hours. Target genes were selected from ALDO-induced genes identified in the RNA sequencing dataset. Transcript expression of GOI is relative to the average expression of reference genes (log₁₀): Rn18S and Tbp. Data are shown as individual points and mean ±95% confidence interval (CI), and GOI is indicated in bold italics above each graph. Statistical significance was determined by one-way ANOVA with Tukey's post hoc test or Kruskal-Wallis test with Dunn's post hoc test, where appropriate: *P<0.05, **P<0.01, and ***P<0.001.

Discussion

We have mapped the transcriptomes underlying corticosteroid-regulated ENaC activity, identifying a small number of ALDO-regulated genes and larger number of CORT-regulated genes, the latter only when 11βHSD2 was inhibited. We used mouse mCCD_cl1 cells,²² a well-described model of ASDN to couple transcriptomic with electrophysiological analysis. ENaC activity was stimulated by low nanomolar concentrations of ALDO^22,37 and CORT at concentrations > 100 nM or when endogenous 11βHSD2 activity was inhibited by CBX. The “gatekeeping” activity of this enzyme extends to complete absence of transcriptional activity by CORT. Isolation of primary principal cells from a novel reporter mouse enabled assessment of steroid-induced ion transport and target gene expression, the latter both in vitro and in vivo.

Deep sequencing of polarized mCCD_cl1 cells revealed expression of genes associated with principal cells of CCD, including Aqp2 (AQP2), Hsd11b2 (11β-HSD2), Kcnj1 (ROMK), Kcnj10 (Kir4.1), Nr3c1 (GR), Nr3c2 (MR), Scnn1a (α-ENaC), Scnn1b (β-ENaC), and Scnn1g (γ-ENaC). With low expression of Slc12a3 (NCC) and absence of Pvalb (parvalbumin), Slc12a1 (NKCC2), and Slc9a3 (NHE3), these cells likely represent epithelia from DCT2 onwards. Although there is low expression of Atp6v1b1 (V-ATPase β1), there is no detectable expression of either Slc4a1 (AE1) or Slc26a4 (pendrin); thus, it seems mCCD_cl1 cells reflect a principal cell population. This expression profile aligns with principal cell populations identified by scSeq of mouse kidneys.^38,39 Transcriptomic profiling of the related cell line, mpkCCD_cl4,⁴⁰ revealed a similar pattern of transcripts, with the exception of Kcnj1, Kcnj10, and Nr3c2.⁴¹ More recently, RNAseq analysis of a subclone of these cells, mpkCCD_cl1, while also having a similar pattern of transcripts, revealed low expression of Nr3c2, but Kcnj1, Kcnj10, Kcnj16, and Scnn1b were absent.⁴² Functionally, the mpkCCD_cl4 cell line does not exhibit ALDO-sensitivity, requiring micromolar concentrations to exert a stimulatory effect on ENaC,^40,43,44 consistent with the lack of, or very low abundance of, the MR.

We confirm both Sgk1 and Zbtb16 as early ALDO-induced genes^16,33,45,46 and further identify Rasd1, Sult1d1, and an unannotated transcript Gm43305. SGK1 is well described as a steroid-induced protein that prevents ubiquitin-mediated removal of ENaC in the apical membrane through phosphorylation of the ubiquitin-ligase Nedd4-2.⁴⁷ Zbtb16, the promyelocytic leukemia zinc finger protein (PLZF), was identified as an early ALDO-induced gene in M1 cells expressing rat MR.⁴⁶ Overexpression of PLZF reduced basal I_SC but did not alter DEX-induced I_SC, suggesting that rather than mediating the stimulatory response of ALDO, PLZF may negatively regulate ENaC in the CD.⁴⁶ Although Rasd1 has not been directly linked to corticosteroid effects in the CD, it has previously been identified as a downregulated transcript in SAGE analysis of the outer medullary collecting duct in mice after 3 days of K⁺ depletion.⁴⁸ Rasd1 has been identified as an intercalated cell-enriched transcript in mice.⁴⁹ Because of the potential plasticity of CD cells, or the de-differentiation of isolated cells, it will be prudent to determine in which cell-type Rasd1 is expressed in vivo and whether its induction by ALDO relates to ENaC stimulation. It is of note that a related gene from the Ras superfamily: Kras (also known as KRas2a) was previously identified as a steroid-induced transcript in amphibian A6 cells⁵⁰ and when coexpressed with ENaC in oocytes, stimulated amiloride-sensitive currents.⁵¹ Although our data show expression of Kras, we did not detect significant changes in expression with any steroid treatment tested. Sult1d1 was also upregulated by ALDO, and two other family members, Sult1a1 and Sult1b1, were upregulated in cells treated with CBX+CORT. Sulfonation has been associated with inactivation of molecules to facilitate excretion,⁵² and it may also regulate intracellular bioavailability.⁵³ SULT1D1 has been associated with catecholamine sulfation in mouse kidneys, thought to enhance excretion.⁵⁴ However, with no human orthologue of Sult1d1, it may represent a species-dependent regulation of catecholamines. Finally, Gm43305 is an unannotated transcript encoding a long noncoding RNA located upstream of the olfactory receptor Olfr49 on chromosome 14. This ALDO-induced lncRNA relates to the stimulation of ENaC remains unknown, but it is of interest whether lncRNAs may be identified as yet another layer of regulation of ion transport processes in the distal nephron.

We compared acute ALDO-induced transcriptional events from this study with two recent studies that mapped the transcriptomes of isolated primary cells of the ASDN after chronic ALDO treatment.^10,11 In these two studies, mice kept on a low Na⁺ diet for 5 days or infused with exogenous ALDO by osmotic minipump for 6 days had plasma (ALDO) of approximately 1.5 nM, which is comparable to the 3 nM used in our own study. Of the ALDO-induced genes identified in the present study, both Sult1d1 and/or Sgk1 were identified as ALDO-regulated.^10,11 From the full transcriptome of the DCT/CNT/iCCD under control conditions,¹⁰ expressions of seven of nine of our ALDO-induced transcripts were detected, with the absence of two identified unannotated genes Gm43305 and Gm9694. The differences in ALDO-induced genes between our study and these others may reflect the acute time frame in which we measured transcriptional changes and may align with the concept of early versus late effects of ALDO in the ASDN.⁵⁵

ENaC-mediated Na⁺ transport in the ASDN is not normally responsive to CORT because of 11βHSD2 activity.^56,57 We confirmed ENaC cannot be stimulated by CORT at physiological concentrations in mCCD_cl1 cells²² and subsequently demonstrated that SGK1 expression/activity is also protected. Our transcriptomic data reveal that 11βHSD2, in fact, fully abolishes the transcriptional effects of CORT, and because no 11βHSD1 expression was detected, CORT is fully inactivated and cannot be reactivated in principal cells. Recent work suggests that the ASDN may, in fact, not include the late DCT/early CNT, where basal ENaC activity is much greater than in the CNT/CCD.⁵⁸ Dietary maneuvers that raise plasma (ALDO) produce different effects on ENaC depending on its location. While ENaC activity is stimulated in both the DCT/CNT and CNT/CCD in mice maintained on a high K⁺ diet,⁵⁹ ENaC activity is only stimulated in the CNT/CCD in mice maintained on a low Na⁺ diet.⁶⁰ The MR seems critical in both regions for steroid-induced ENaC activity as deletion abolishes these responses.^59,61 It is interesting to speculate whether the expression of the cellular machinery underpinning ALDO sensitivity differs in these locations giving rise to different basal, as well as ALDO-induced, ENaC activity. There are mixed reports regarding immunolocalization of 11βHSD2 in the DCT,^2,3,62 and recent scSeq of murine kidneys reports low/negligible expression.^38,39,63 We determined the transcriptomic effects of CORT in the absence of 11βHSD2: ENaC activity was robustly stimulated and differential gene expression analysis revealed modulation of a much larger number of genes compared with ALDO, including previously described steroid-induced targets: GILZ, α-ENaC, and Per1.^12,13 All genes identified after ALDO treatment were identified in cells treated with CBX+CORT, indicating a shared pathway of these hormones downstream of binding endogenous receptors. The relative roles that the MR, GR, or, indeed, both within the distal nephron, particularly where 11βHSD2 is absent versus present, remain incompletely understood and warrant further investigation.

Generating mT/mG-Aqp2Cre mice to isolate primary CD cells is a similar strategy to previous studies that used a TRPV5-eGFP reporter line to isolate primary CNT/CD cells¹⁰ or CD-specific cell isolation from wild-type mice using either DBA lectin⁶⁴ or an L1-CAM antibody.¹¹ In our study, the isolated GFP+ cells were, as expected, strongly enriched for genes characteristic of principal cells (encoding 11βHSD2, ROMK, α-ENaC). These cells also expressed genes encoding pendrin and AE1, associated with intercalated cells, albeit at a low abundance. This finding is consistent with groups who isolated primary CD cells¹¹ or DCT2/CNT/initial CCD cells.¹⁰ We do not think our results reflect nonspecificity of Cre-recombinase as Aqp2Cre is not expressed in intercalated cells.²⁹ Nor do we consider this to be contamination during isolation, e.g., because of autofluorescence: our gating strategy was stringent and we did not detect enrichment of markers associated with proximal tubule cells, which exhibit strong autofluorescence. It is possible that either the isolated primary principal cells de-differentiate rapidly or perhaps more likely, our results support the evidence that CD epithelia exhibit plasticity.^38,65,66 Indeed, scSeq analysis of murine kidney³⁸ reported a “transitional cell” type where high expression of these IC transcripts was detected at a similar level as Aqp2. The finding that the GFP+ cells contain transcripts associated with PCs but also ICs is consistent with this. Importantly, functional assessment of the isolated GFP+ cells revealed 11βHSD2 activity as well as predominant amiloride-sensitive currents that could be stimulated by ALDO; therefore, the isolated GFP+ cells phenotypically behave as Na⁺ absorbing principal cells.

Primary principal cells grown on permeable supports in culture developed V_t and R_t in a manner analogous to mCCD_cl1 cells. R_t, of note, was much larger in these cells at approximately 5 kΩ·cm², with a V_t of approximately −20 mV giving rise to an I_eq of approximately −3.5 µA·cm⁻². These currents are smaller than those recorded from mCCD_cl1,^22,24 mpkCCD_cl4,^35,40 as well as primary CD cells isolated by DBA lectin,⁶⁴ but were almost completely abolished by amiloride, indicating a Na⁺ absorbing phenotype via ENaC. Primary principal cells exhibited 11βHSD2 activity and ENaC-mediated currents were significantly stimulated by ALDO, CBX+CORT, as well as DEX. This correlates with currents measured in mCCD_cl1 cells, confirming these cells represent a relevant model of principal cells with the advantage that they do not de-differentiate over a small number of passages. Analysis of identified ALDO-induced genes in primary cells grown in culture revealed ALDO and CBX+CORT both upregulated Sgk1, Rasd1, and Zbtb16, but ALDO also upregulated Sult1d1, whereas CORT also upregulated Gm43305. Interestingly, DEX upregulated seven of eight ALDO-induced genes tested. Studies have shown that while DEX is considered a synthetic glucocorticoid, it can bind the MR as well as the GR.⁶⁷ In the transcriptomic data from the mCCD_cl1 cells, MR is expressed at 4× greater levels than GR, smaller than the 7× difference reported in the mpkCCD_cl1 cells,⁴² but similar to the 3x greater levels reported in the CD principal cell population identified in scSeq analysis of murine kidney.³⁸

Building on our in vitro experiments, principal cell-specific reporter mice were acutely administered ALDO or CORT±CBX.⁶⁸ Four of the identified ALDO-induced genes were upregulated in mice treated with either ALDO or CBX+CORT: Sult1d1, Gm443305, Rasd1, and Zbtb16. We noted that CBX treatment alone resulted in an upregulation in the expression of the ALDO-induced genes Zbtb16, Gm43305, and Sult1d1. We did not, however, detect changes in Sgk1 across any of the groups. Sgk1 has previously been shown to be upregulated in microdissected CNT/CCD after 1-hour ALDO treatment³³ and it is possible these rapid activation events were completed by our 3 hours collection point. However, from the data, it is clear that four of our identified corticosteroid-induced genes remained upregulated after 3 hours.

In summary, we report the transcriptional landscape associated with acute corticosteroid-induced ENaC activity in principal cells of the CD. In addition to the previously described ALDO-induced targets Sgk1 and Zbtb16, we identify three additional acutely upregulated targets Rasd1, Sult1d1, and the unannotated Gm43305. The potential role that these genes play in mediating ALDO-induced ENaC activity in principal cells of the CD is of interest and remains to be determined.

Disclosures

M.A. Bailey reports the following: Consultancy: River 2 Renal; and Advisory or Leadership Role: Kidney Research UK (Research Grant & Fellowships Committee); American Journal of Physiology (Associate Editor). S.M. Wilson reports the following: Ownership Interest: AstraZeneca PLC, Diageo, Legal & General Group, Reckitt Benckiser Group Plc, and Unilever Plc. All remaining authors have nothing to disclose.

Funding

This was work supported by Kidney Research UK: postdoctoral fellowship PDF_008_20151127 and innovation grant IN_001_20170302; the British Heart Foundation: research excellence award RE/13/3/30183; the Scottish Funding Council: St Andrews Research Funding Scheme; and the Society for Endocrinology: early career grant.

Author Contributions

M.A. Bailey and M.K. Mansley performed the conception and design of research; H.M. Costello, S. Loughlin, M.K. Mansley, and A.J. Roe performed experiments; C. Buckley, H.M. Costello, S.R. Loughlin, M.K. Mansley, and A.J. Roe analyzed the data; M.K. Mansley contributed in preparing figures and drafted the manuscript; M.A. Bailey and M.K. Mansley edited and revised the manuscript; and M.A. Bailey, C. Buckley, H.M. Costello, S.R. Loughlin, M.K. Mansley, A.J. Roe, and S.M. Wilson approved the final version of the manuscript.

Data Sharing Statement

The raw RNAseq data have been uploaded to the Sequence Read Archive at NCBI, with project ID PRJNA820455.

Supplemental Material

This article contains the following supplemental material online at http://links.lww.com/KN9/A232 and http://links.lww.com/KN9/A233.

Supplementary Figure 1. Validation of corticosteroid-induced transcripts in mCCD_cl1 cells.

Supplementary Figure 2. Expression of reference genes used for validation of identified corticosteroid-induced transcripts.

Supplementary Figure 3. Gating strategy for FACS of primary cells into tdTom and GFP labeled populations.

Supplementary Figure 4. Expression of reference genes used for measurement of identified corticosteroid-induced transcripts in primary principal cells.

Supplementary Figure 5. Expression of reference genes used for measurement of steroid-induced targets in isolated primary principal cells in mT/mG-Aqp2Cre mice following acute injection of corticosteroids.