Activation of the Hypoxia-Inducible Factor Pathway Inhibits Epithelial Sodium Channel–Mediated Sodium Transport in Collecting Duct Principal Cells

Significance Statement

The hypoxia-inducible factor (HIF) pathway is a key mediator of cellular adaptation to low oxygen tension. The aldosterone-sensitive distal nephron is the site of active and highly ATP-consuming sodium reabsorption, according to the requirement of sodium balance. The authors found that activation of the HIF pathway in cultured principal cells led to decreased amiloride-sensitive current (reflecting decreased epithelial sodium channel [ENaC] activity) and decreased expression of ENaC subunits, whereas HIF silencing led to increased amiloride-sensitive current and expression of ENaC subunits. Hypoxic control mice displayed decreased γENaC, whereas HIF1α knockout mice displayed increased γENaC. These findings suggest that the HIF pathway controls ENaC expression and activity, and may represent a negative feedback mechanism to prevent hypoxia and/or reactive oxygen species–induced cell damage under sustained stimulation of sodium transport.

Abstract

Background

Active sodium reabsorption is the major factor influencing renal oxygen consumption and production of reactive oxygen species (ROS). Increased sodium reabsorption uses more oxygen, which may worsen medullary hypoxia and produce more ROS via enhanced mitochondrial ATP synthesis. Both mechanisms may activate the hypoxia-inducible factor (HIF) pathway. Because the collecting duct is exposed to low oxygen pressure and variations of active sodium transport, we assessed whether the HIF pathway controls epithelial sodium channel (ENaC)–dependent sodium transport.

Methods

We investigated HIF’s effect on ENaC expression in mpkCCD_cl4 cells (a model of collecting duct principal cells) using real-time PCR and western blot and ENaC activity by measuring amiloride-sensitive current. We also assessed the effect of hypoxia and sodium intake on abundance of kidney sodium transporters in wild-type and inducible kidney tubule–specific Hif1α knockout mice.

Results

In cultured cells, activation of the HIF pathway by dimethyloxalylglycine or hypoxia inhibited sodium transport and decreased expression of βENaC and γENaC, as well as of Na,K-ATPase. HIF1α silencing increased βENaC and γENaC expression and stimulated sodium transport. A constitutively active mutant of HIF1α produced the opposite effect. Aldosterone and inhibition of the mitochondrial respiratory chain slowly activated the HIF pathway, suggesting that ROS may also activate HIF. Decreased γENaC abundance induced by hypoxia in normal mice was abolished in Hif1α knockout mice. Similarly, Hif1α knockout led to increased γENaC abundance under high sodium intake.

Conclusions

This study reveals that γENaC expression and activity are physiologically controlled by the HIF pathway, which may represent a negative feedback mechanism to preserve oxygenation and/or prevent excessive ROS generation under increased sodium transport.

The high blood supply to the kidney is required for elimination of metabolic waste products and xenobiotics in the urine. Despite this high oxygen supply, the kidneys are highly sensitive to hypoxia which plays a major role in the pathogenesis of both AKI and CKD.^1–4 This sensitivity of kidney tissue to hypoxic injury relies on the architecture of renal vasculature forming intrarenal shunts that decrease tissue oxygen availability. These anatomic factors combined with the high metabolic activity of renal tubule epithelial cells generate a steep cortico-medullary oxygen gradient resulting in relative hypoxia of the straight portions of the kidney tubule.⁵

The hypoxia-inducible factor (HIF) pathway is a key mediator of cellular adaptation to low oxygen tension.⁶ HIFα proteins are constitutively expressed but rapidly degraded by the ubiquitin-proteasome system after their hydroxylation by Prolyl Hydroxylase Domain (PHD) proteins in the presence of oxygen. Under hypoxic conditions, PHD activity is inhibited, leading to HIFα protein stabilization, dimerization with HIFβ, and transcriptional activation of target genes involved in glycolysis, pH regulation, angiogenesis, or erythropoiesis. Thus, the HIF signaling pathway promotes cell survival via the reprogramming of cellular metabolism and mitochondrial activity. Interestingly, several studies reported that HIF pathway activation improves renal function in different models of experimental kidney injury such as ischemia/reperfusion.^7,8 To this aim, PHD inhibitors have been developed and some of them are currently being tested in humans.⁹

The kidney constitutively produces low amounts of reactive oxygen species (ROS) via both NADPH oxidases, mostly NOX4, expressed by kidney tubule epithelial cells¹⁰ and the leak pathway of the mitochondrial respiratory chain.¹¹ Mitochondrial ROS production physiologically increases under conditions of stimulated ATP synthesis¹² which prevails after stimulation of active sodium reabsorption by kidney tubule cells.¹³ The H₂O₂ generated by kidney tubule cells may promote HIF pathway activation via PHD inhibition.¹⁴

One major renal function is the body fluid volume homeostasis via tubular reabsorption of filtered sodium. Luminal sodium flows into tubular cells along its electrochemical gradient via a segment-specific apical transporter. The epithelial Na⁺ channel (ENaC) is expressed along the aldosterone-sensitive distal nephron (ASDN) which includes the late distal convoluted tubule, connecting tubule, and collecting duct (CD).¹⁵ The ASDN is the site of the final adjustment of sodium reabsorption according to the requirement of sodium balance. In principal cells which are responsible for sodium reabsorption in this segment, sodium crosses the apical membrane through ENaC and is then extruded to the interstitial compartment via the basolateral Na⁺,K⁺-ATPase which hydrolyzes ATP.¹⁶ This active sodium reabsorption generates high levels of ATP consumption which is supplied by ATP produced by mitochondrial respiration and glycolysis. Sodium transport by principal cells is tightly controlled by hormones such as aldosterone¹⁷ and nonhormonal factors including intracellular sodium¹⁸ and extracellular osmolality.^19,20

We hypothesized that activation of the HIF pathway may contribute to the physiologic regulation of sodium reabsorption by inhibiting the ENaC-dependent sodium transport, thereby reducing ATP and oxygen consumption and ROS generation by principal cells. Several lines of experimental evidence may support this hypothesis: HIF activation increased natriuresis and reduced salt-sensitive hypertension in Dahl salt-sensitive rats²¹, whereas in the subtotal nephrectomy rat model of CKD, HIF pathway activation by a PHD inhibitor increased urinary sodium excretion.²² An important question is whether the renal HIF pathway is active only under pathologic conditions or is also involved in the regulation of physiologic processes such as sodium handling by CD principal cells. Using cultured CD principal cells and renal tubular cell–specific knock-out mice for HIF1α, we showed that the HIF pathway acts as a negative feedback that inhibits ENaC-mediated sodium reabsorption.

Methods

Cell Culture

Mouse cortical collecting duct (CCD) cells, mpkCCD_cl4 and mCCD_cl1 cells, were cultured as described previously.²³

For shRNA-mediated knockdown, the following shRNA pLKO.1 plasmids were employed: pLKO.1 scrambled shRNA (Addgene 1864) and pLKO.1-Hif1α shRNA (Sigma-Aldrich TRCN0000232222). Lentivirus was produced and purified according to standard protocols using the packaging vector psPAX2 and the envelope vector pMD2G.²⁴ To generate stable cell lines expressing shRNA, mpkCCD_cl4 cells were transduced with concentrated lentiviral particles. At 3 days after transduction, polyclonal shRNA-expressing cell lines were selected using 2 μg/ml puromycin (Sigma-Aldrich).

For expression of human HA-tagged mutants of HIF1α, HA-HIF1α (P402A/P564A), the following plasmid was used: HA-HIF1alpha P402A/P564A-pBabe-puro (Addgene, 19005).²⁵ Retrovirus was produced according to the protocol associated with the BABE plasmids available on the Addgene website (https://media.addgene.org/data/70/60/165f20c4-af64-11e0-90fe-003048dd6500.pdf). To generate stable cell lines expressing HA-HIF1α (P402A/P564A), mpkCCD_cl4 cells were transduced with concentrated lentiviral particles. At 3 days after transduction, polyclonal HA-HIF1α (P402A/P564A)–expressing cell lines were selected using 2 μg/ml puromycin (Sigma-Aldrich).

For electrophysiological measurements, cells were seeded on polycarbonate filters (Transwell, Corning Costar, Cambridge, MA, USA). Cells were then maintained in serum- and hormone-deprived medium for 12 hours before performing experiments²⁷.

Aldosterone, 17-hydroxyprogesterone, and amiloride hydrochloride were purchased from Sigma-Aldrich. Dimethyloxalylglycine (DMOG) was purchased from Enzo Life Science. All experiments were performed during at least 3 separate days, using different passages.

Transgenic Mice

To generate inducible and renal tubule–specific knock-out mice for Hif1α (Hif1α^PAX8/LC1) and their control littermates (Hif1α^lox/lox), we crossed successively Hif1α^lox/lox mice²⁶ with Pax8-rtTA transgenic mice (Pax8-rtTA^tg/0), which express the reverse tetracycline transactivator (rtTA) under the control of the Pax-8 promoter expressed by kidney tubule cells, and TRE-Cre transgenic mice (LC1-TRE-Cre^tg/0),²⁷ which express the Cre recombinase and luciferase (LC1) under the control of the tetracycline response element (TRE) which binds the rtTA in the presence of tetracycline. After breeding, we obtained (Hif1α^lox/lox; Pax8-rtTA^tg/0; LC1-TRE-Cre^tg/0) genotypes. HIF1α deletion was induced in renal tubular cells by doxycycline hydrochloride treatment (Alfa Aesar) (2 mg/ml and 2% sucrose in drinking water) for 15 days in 3–4-week-old mice.

Genotyping of those mice was performed by PCR analysis of DNA isolated from ear biopsy specimens using the following primers: HIF1α lox: sense 5′-GGAGCTATCTCTCTAGACC-3′ and antisense 5′-GCAGTTAAGAGCACTAGTTG -3′; Pax8-rtTA: ST1 5′-CCATGTCTAGACTGGACAAGA-3′ and ST2 5′-CTCCAGGCCACATATGATTAG-3′; LC1: Cre3 5′-TCGCTGCATTACCGGTCGATGC-3′ and Cre4 5′-CCATGAGTGAACGAACCTGGTCG-3′.

Animal Experiments

All animal experiments were approved by the Institutional Ethical Committee of Animal Care in Geneva and Cantonal authorities. Male transgenic mice aged 6–7 weeks were housed in metabolic cages for the last 3 days (included 2 days of adaptation) in order to measure food and water intake (Table 1). Daily urine samples were measured and collected in order to determine urinary creatinine and Na⁺ excretions using ion-specific electrodes (UniCel DxC800 Synchron Clinical System, Beckman Coulter). Mice aged 7–8 weeks were euthanized for kidney removal under anesthesia by intraperitoneal injection of ketamine and xylazine (100 mg/kg and 5 mg/kg, respectively) (GRAEUB, Swissmedic, Bayer Healthcare).

Table 1.

Blood parameters in C57Bl/6 mice exposed to normoxia (21% O₂) or hypoxia (8% O₂) for 48 hours (n=8)

Parameter	Normoxia	Hypoxia	P Value
pH	7.23±0.05	7.30±0.02	NS
pCO2 (mm Hg)	53.1±3.2	36.1±1.95	0.0004
HCO3− (mmol/L)	22.4±1.4	17.8±0.9	0.013
K+ (mmol/L)	7.13±0.14	6.51±0.11	0.005

All values are means±SEM.

For hypoxia experiments, male transgenic or wild-type mice aged 6–7 weeks were exposed either to normoxic (21% O₂) or to normobaric hypoxic (8% O₂) conditions for 48 hours in a hypoxic in vivo cabinet (BioSpherix, Parish, NY, USA). Then, blood samples were collected from the submandibular vein to measure several parameters (Table 2) using the EPOC blood analysis system (Siemens, Munchen, Germany) and the animals were euthanized by cervical dislocation for kidney removal. Hif1α^PAX8/LC1 mice, their littermates, and wild-type mice were subjected to low- (0.1%) and high- (3%) salt diet, for 7 days before euthanasia.

Table 2.

Baseline physiologic parameters of transgenic mice

Strains	Body Weight (g)	Food Intake (g)	Water Intake (ml)	Urine Volume (ml)
Hif1αlox/lox	19.87±0.88	2.91±0.17	6.71±0.34	2.34±0.30
Hif1αPAX8/LC1	19.79 ± 0.55	2.63 ± 0.21	5.99 ± 0.34	1.84 ± 0.31

Values are means±SEM from 6–8 mice in each group. Knock-out mice were compared with littermate controls.

For measurement of GFR, mice were anesthetized with isoflurane and a miniaturized imager device (Mannheim Pharma and Diagnostics, Mannheim, Germany) was mounted onto the animal’s back. The skin background signal was recorded for 5 minutes before intravenous injection of 150 mg/kg FITC-sinistrin (Mannheim Pharma and Diagnostics, Germany) and recording of transcutaneous fluorescence for 1 hour in conscious animals. mGFR (µl/min) was calculated from the decrease in fluorescence intensity over time (i.e., plasma half-life of FITC-sinistrin) using a two-compartment model, the mouse body weight, and an empirical conversion factor using the MPD Lab software (Mannheim Pharma and Diagnostics, Germany), as previously described.²⁸

Microdissection of Kidney Tubules

Briefly, the left kidneys of pentobarbital-anesthetized mice were infused via the abdominal aorta with an incubation solution (Hank's solution supplemented with 1 mM pyruvate, 0.1% BSA, 0.5 mM MgCl₂, 1 mM glutamine, and 20 mM Hepes, pH 7.4) containing collagenase (Worthington, 337 UI/mg, 0.18% wt/vol). Kidneys were cut into small pieces which were incubated for 20–25 min at 32°C in an oxygenated incubation solution containing 0.1% collagenase. Renal tubules, i.e., PCT, cTAL, and CCD, were microdissected under stereomicroscopic observation in an incubation solution supplemented with antiproteases (protease inhibitor cocktail tablets, Roche) at 4°C, as previously described.²⁹

RNA Extraction and Real-Time PCR

Total RNA was extracted from cultured cells or kidney tissues using the Nucleospin RNA II kit (Macherey-Nagel) according to the manufacturer’s instructions. RNA (500 ng) was used to synthesize cDNA using the qScript cDNA Supermix (Quanta Biosciences). Real-time PCR analysis was performed as previously described.³⁰ For isolated tubules, 40 tubules were pooled and RNA was isolated using the RNeasy Plus Micro kit (Qiagen); RNA was eluted in 14 ml of RNase-free water before cDNA synthesis (see above). Primers used are described in Table 3. Mouse acidic ribosomal phosphoprotein p0 was used as internal standard.

Table 3.

Primer sequences used for quantitative real-time PCR

Target Gene	Sense (5′–3′)	Antisense (5′–3′)
p0	AATCTCCAGAGGCACCATTG	GTTCAGCATGTTCAGCAGTG
αENaC	CAGACTTGGAGCTTTGACAAGGA	ACTTCTCTGTGCCTTGTTTATATGTGTT
βENaC	CAGACTGGGCCTATTGCTATCTAAA	ACATGCTGAGGCAGGTCTCTCT
γENaC	CCGAGATCGAGACAGCAATGT	CGCTCAGCTTGAAGGATTCTG
Na+,K+-ATPase α	TCCCTTCAACTCCACCAACAA	TTTGGGCTCAGATGCATTTG
Na+,K+-ATPase β	TCGGAGAAGAAGGAGTTTTTGG	GCAGCCATAAAATATCACGTAGAACA
HIF1α	CACCGATTCGCCATGGA	TTCGACGTTCAGAACTCATCTTTT
PDK1	TTCAAGAATGCCATGAGAGC	CCGGTCACTCATCTTCACAG
PGK1	AACAACCAAAGGATCAAGGC	CAGCAGCAACTGGCTCTAAG
VEGF	GGTGGACATCTTCCAGGAGT	TGATCTGCATGGTGATGTG
Glut1	AGCCCTGCTACAGTGTATCCT	CCGACCCTCTTCTTTCATCT
Hmox1	CCACTCCCTGTGTTTCCTTT	GCTGCTGGTTTCAAAGTT
EPO	AACCCATCGTGACATTTTCT	CACCCTGCTGCTTTTACTCT

Western Blotting

Cultured cells or kidney tissues were lysed as previously described.¹⁰ Then, 10 μg of protein was loaded on polyacrylamide gels. Protein concentrations were quantified using a BCA protein assay kit (Pierce), subjected to SDS-PAGE, and blotted onto polyvinylidene difluoride membranes (Immobilon-P; Millipore) using standard methods. The antibodies used are indicated in Table 4. Protein signals were quantified with ImageJ software. Results were expressed as the ratio of the densitometry of the band of interest to the loading control.

Table 4.

Primary antibodies used for western blotting

Name	Species	Dilution	Supplier
NHE3	Rabbit	1/500	Stressmarq
NKCC2	Rabbit	1/1000	Wagner et al. 200831
NCC	Rabbit	1/1000	Sorensen et al. 201332
NCC-TP58	Rabbit	1/1000	Sorensen et al. 201332
αENaC	Rabbit	1/1000	Wagner et al. 200830
βENaC	Rabbit	1/1000	Stressmarq
γENaC	Rabbit	1/1000	Stressmarq
β-actin	Mouse	1/20,000	Sigma
HIF1α	Mouse	1/1000	Santa Cruz
HA-tag	Mouse	1/1000	Covance
SGK1	Rabbit	1/1000	Sigma
Claudin-8	Rabbit	1/1000	Thermofisher
AQP2	Rabbit	1/1000	Hasler et al. 200333
Na+,K+- ATPase	Rabbit	1/1000	Carranza, 199634

Short-Circuit Current Recordings in Ussing Chambers

Confluent mpkCCD_cl4 cells grown on Snapwell filters (Snapwell, Corning Costar, Cambridge, MA, USA, 1.12 cm²) for 7 days were transferred to an Ussing chamber and the short-circuit current was measured under voltage clamp (0 mV) using dual silver-silver chloride electrodes connected to a VCC MC6 Multichannel Voltage/Current Clamp via silver/AgCl electrodes and 3 M KCl agar bridges (Physiologic Instruments, San Diego, CA, USA). Ussing chambers were filled with high-sodium buffer (120 mM NaCl, 10 mM NaHCO₃, 5 mM KCl, 1.2 mM CaCl₂, 1 mM MgCl₂, and 10 mM Hepes, pH 7.4) or low-sodium buffer (240 mM mannitol, 10 mM NaHCO₃, 5 mM KCl, 1.2 mM CaCl₂, 1 mM MgCl₂, and 10 mM Hepes, pH 7.4). Cells were equilibrated during 1 hour and current was recorded using Quick Data Acquisition DI100 USB board (Physiologic Instruments). Cell monolayers were voltage-clamped at 0 mV and continuously monitored for changes in short-circuit current. Buffers were maintained at 37°C and bubbled constantly with a mixture of 95% O₂ and 5% CO₂ throughout the experiment. The transport of Na⁺ across the apical and the basolateral membranes mediated by ENaC and Na⁺,K⁺-ATPase, respectively, was assessed by adding amiloride (100 μM) to the apical side and ouabain (1 mM) to the basolateral side. The Na⁺,K⁺-ATPase was activated by addition of the Na⁺ ionophore amphotericin B (50 µg/ml) to the apical side. Positive current corresponds to a flow of positive charges from the apical to the basolateral side of the monolayer. The amiloride-sensitive current reflecting ENaC activity was calculated as the total current minus the amiloride-resistant current. The Na⁺,K⁺-ATPase current was defined as the difference between the amphotericin B–induced current and the ouabain-resistant current.

Immunofluorescence

After dehydration and paraffin-embedding, kidneys sections of 5-µm thickness were used for analysis. Antigen retrieval was done with 10 mM citrate buffer, pH 6. Blocking of nonspecific binding was done with 2% BSA for 20 min. The sections were then incubated overnight at 4°C with primary γENaC antibody (diluted 1:50), followed by a 1-hour incubation with Cyanin3-conjugated goat anti-mouse (cat. no. M30010; Invitrogen) diluted 1:200 in 10% PBS-NGS at room temperature. Samples were mounted on microscope slides using Vectashield mounting medium (Maravai Life Science, San Diego, CA, USA). Fluorescence images were acquired using a Zeiss Axio Imager M2 (Carl Zeiss, Oberkochen, Germany). Negative controls were performed in the absence of primary antibody (not shown).

Statistical Analyses

Results are given as the mean±SEM from n independent experiments. A Shapiro–Wilk test was used to test the distribution of the population from which sample data were extracted. For parametric data, statistical differences were assessed using a two-tailed unpaired t test when two groups were compared or one-way or two-way ANOVA with multipair-wise comparison from Tukey when more than two groups were compared. For nonparametric data, statistical differences were assessed using the Mann–Whitney test when two groups were compared or by the Kruskal–Wallis test with a multipair-wise comparison from Dunn when more than two groups were compared. A P value <0.05 was considered significant.

Results

Activation of the HIF Pathway Inhibits Transepithelial Sodium Transport and Downregulates ENaC Subunits Expression in Mouse CD Principal Cells

RT-PCR experiments performed in isolated microdissected kidney tubule segments showed that HIF1α mRNA levels were much higher in CCDs than in proximal tubules and thick ascending limbs (Supplemental Figure 1, Supplemental Table 1). These results suggest that CD is potentially the major physiologic target of the HIF pathway along the kidney tubule.

We first exposed mpkCCD_cl4 to the PHD inhibitor DMOG which stabilizes HIFα proteins and thereby activates the HIF pathway. Activation of the HIF pathway was assessed by increased HIF1α protein abundance and by the upregulation of four classic HIF target gene mRNA levels (Figure 1, A, B, and F). We observed a decrease in amiloride-sensitive transepithelial current correlated with a decrease in protein and mRNA levels of all three ENaC subunits (Figure 1, C, D, and E). Both Na⁺,K⁺-ATPase α- and β-subunit mRNA levels, and Na⁺,K⁺-ATPase α-subunit protein level, decreased in response to 1 mM DMOG within a time-course similar to that of ENaC subunits (Figure 1, C–E). It should be noted that time-courses of increased Hif1α protein abundance and upregulation of HIF target gene mRNA were similar (Figure 1, A, B, and F). The downregulation of ENaC and Na⁺,K⁺-ATPase subunits was delayed and sustained in comparison with the rapid and transient upregulation of classic HIF target genes (Figure 1, C, E, and F).

Figure 1.

Activation of the HIF pathway by DMOG, a PHD inhibitor, decreases sodium transport via the downregulation of ENaC and Na⁺,K⁺-ATPase in cultured mouse CCD principal cells. mpkCCD_cl4 cells were exposed to 1 mM DMOG for 2, 6, and 24 hours. (A) Stabilization of HIF1α protein by DMOG assessed by western blot. β-actin was used as a loading control. A representative experiment from three is shown. (B and C) Western blots were quantified using ImageJ software. Results are expressed as a percentage of control values (dots) and are means±SEM (lines) from three experiments in each group. (D) Monitoring of the transepithelial current in mpkCCD_cl4 cells exposed to DMOG compared with control cells along the kinetic. Results are means±SEM from three independent experiments. (E) Decrease in ENaC subunits mRNA levels in response to DMOG assessed by real-time PCR and those of Na⁺,K⁺-ATPase subunits. (F) HIF pathway transcriptional activity was confirmed by real-time PCR analysis of HIF target genes PDK1, PGK1, VEGF, and GLUT1 mRNA levels. Results are expressed as fold of controls (dots) and means±SEM (lines) from three independent experiments are displayed. Unpaired Student’s t test was used in (C) and (D). Kruskal–Wallis test was used to compare data from (B). One-way ANOVA was used to compare data from (E) and (F). *P<0.05; **P<0.01; ***P<0.001; ****P<0.0001, versus controls.

After 24 hours of incubation with 1 mM DMOG, monolayers of mpkCCD_cl4 cells were mounted in Ussing chambers under voltage-clamp conditions. DMOG decreased amiloride-sensitive short-circuit current, confirming the downregulation of ENaC (Figure 2, A and B). As expected from the size of the sodium concentration gradient established between the apical chamber and cell cytosol, the ENaC-mediated current was larger in the presence of 130 mM than 10 mM apical sodium. Permeabilization of the apical membrane by amphotericin-B resulted in a large positive short-circuit current (Figure 2, A–C). The ouabain-sensitive portion of this current measures the Na⁺,K⁺-ATPase activity. Results show that DMOG inhibited Na⁺,K⁺-ATPase-mediated current in a proportional manner under either rate limiting (15 mM) or saturating (130 mM) sodium concentration (Figure 2, A–D), indicating an effect on Na⁺,K⁺-ATPase V_max which points to decreased number of active pump units.

Figure 2.

Activation of the HIF pathway by DMOG decreases sodium transport via the downregulation of ENaC and the Na⁺,K⁺-ATPase activity in cultured mouse CCD principal cells. mpkCCD_cl4 cells were grown to confluence on filters and exposed or not to 1 mM DMOG for 24 hours. (A) Representative short-circuit current recording in Ussing chambers under a zero-voltage clamp showing the effect of amiloride (100 μM, apical side), amphotericin B (50 μg/ml, apical side), and ouabain (1 mM, basolateral side) application on the short-circuit current in the presence of 10 mM (orange lines) or 130 mM (blue lines) apical NaCl. Amiloride-, amphotericin B-, and ouabain-sensitive currents were determined from the changes in short-circuit current after drug application in the presence of 10 mM (orange dots) or 130 mM (blue dots) apical NaCl (B–D). Results are means±SEM from four independent experiments. Statistical analysis was performed by Mann–Whitney U test; *P<0.05; **P<0.01; ***P<0.001.

To confirm this effect, we analyzed the effect of hypoxia (sealed bag) on HIF pathway activation and sodium transport in cultured mpkCCD_cl4 cells, a model of CD principal cells. Increased HIF1α protein levels (at 3, 6, and 12 hours) associated with an upregulation of PDK1 and PGK1 expression, two typical HIF target genes, confirmed the expected activation of the HIF signaling pathway under hypoxic conditions (Figure 3, A, B, and F). The amiloride-sensitive current, reflecting the transepithelial sodium transport, decreased in cells exposed to hypoxia for 6–12 hours (Figure 3D). Both mRNA and protein abundance of ENaC subunits were decreased, suggesting that hypoxia downregulates their mRNA transcription or stability (Figure 3, A, C, and E). The protein abundance of the Na⁺,K⁺-ATPase α-subunit and the mRNA levels of its β-subunit remained unchanged (Figure 3, A, C, and E), whereas the mRNA levels of its α-subunit decreased within a time-course similar to that of αENaC (Figure 3E).

Figure 3.

Hypoxia activates the HIF signaling pathway and decreases sodium transport via ENaC subunits downregulation in cultured mouse CCD principal cells. (A) Western blot analysis showing that increased HIF1α protein abundance induced by hypoxia correlates with decreased α-, β-, and γENaC abundance. The Na⁺,K⁺-ATPase α-subunit abundance was not significantly changed. β-actin was used as a loading control. A representative experiment from three is shown. (B and C) Western blots were quantified using ImageJ software. Results are expressed as a percentage of control values (dots) and are means±SEM (lines) from three experiments in each group. (D) Monitoring of the transepithelial current in mpkCCD_cl4 cells exposed or not to hypoxia (1% O₂) for 3, 6, and 12 hours. Results shown are means±SEM from four independent experiments. (E) Real-time PCR analysis under the same conditions revealed that hypoxia induced a downregulation of mRNA levels of ENaC subunits and Na⁺,K⁺-ATPase α-subunit. (F) HIF pathway activation was confirmed by increased mRNA levels of the HIF target genes PDK1 and PGK1, assessed by real-time PCR. Results are expressed as fold of control (dots) and are means±SEM (lines) from three independent experiments in (C) and (D). Unpaired Student’s t test was used to compare data from (C) and (D). One-way ANOVA was used to compare data from (B), (E), and (F). *P<0.05; **P<0.01; ***P<0.001; ****P<0.0001, versus control.

Altogether, these experiments strongly suggest that activation of the HIF pathway inhibits the transepithelial sodium transport via downregulation of ENaC and Na⁺,K⁺-ATPase in mpkCCD_cl4 cells.

Silencing of HIF1α Upregulates ENaC in Mouse CCD Principal Cells

To confirm the effect of HIF1α in the regulation of sodium transport in mpkCCD_cl4 cells, these cells were transduced with lentiviruses expressing shRNAs directed against HIF1α or a scrambled shRNA used as control. Real-time PCR experiments revealed that shHIF1α specifically decreased the HIF1α mRNA levels by >80% (Figure 4A). Exposure to hypoxia for 6 hours induced the stabilization of HIF1α in cells transduced with scrambled shRNA (pLKO), although this effect was blunted in cells transduced with shHIF1α (Figure 4B). Specific silencing of HIF1α stimulated the ENaC-mediated transepithelial current as compared with control cells (pLKO) (Figure 4C). γENaC was upregulated at both mRNA and protein levels, whereas α- and βENaC were increased at the mRNA level (Figure 4, D and F). Therefore, knockdown of HIF1α protein induced activation of transepithelial sodium transport and upregulation of γENaC in mpkCCD_cl4 cells.

Figure 4.

Knockdown of HIF1α by shRNA increases sodium transport and β- and γENaC subunit expression levels in cultured CD principal cells. mpkCCD_cl4 cells were transduced with lentiviruses expressing shRNAs targeting HIF1α (shHIF1α) or scramble shRNA (plko). (A and B) Knockdown of HIF1α was assessed by real-time PCR (A) or western blot (B) of the different cell lines exposed or not to hypoxia (1% O₂). (C) Increased transepithelial current in HIF1α-depleted cells (shHIF1α) compared with control cells (pLKO). Amiloride-insensitive current was unchanged. Results are means±SEM from three independent experiments. (D) Western blot analysis showing that HIF1α silencing is associated with increased β- and γENaC abundance. β-actin was used as a loading control. A representative experiment from three is shown. (E) Western blots were quantified using ImageJ software. Results are expressed as a percentage of control values (dots) and are means±SEM (lines) from three experiments in each group. (F) Increase in ENaC subunits mRNA levels in response to HIF1α silencing assessed by real-time PCR. Bar graphs represent fold of controls (dots) and are means±SEM (lines) from three independent experiments. Unpaired Student’s t test was used to compare plko and shHIF1α. **P<0.01; ***P<0.001; ****P<0.0001, versus plko.

Overexpression of a Constitutively Active Mutant of HIF1α Downregulates ENaC in Mouse CCD Principal Cells

To further demonstrate the role of HIF1α in the control of ENaC subunit expression, mpkCCD_cl4 cells were transduced with retroviruses expressing a human HA-tagged degradation-resistant mutant of HIF1α, HA-HIF1α^{(P402A/P564A)}. These mutations prevent the hydroxylation of HIF1α by PHDs, inducing its resistance to VHL-mediated degradation.³⁵ The stable expression of this human mutant HA-tagged HIF1α was confirmed by western blot with anti-HA antibodies that do not detect the mouse endogenous HIF1α (Supplemental Figure 2A). The increase in mRNA levels of HIF target genes PDK1 and VEGF indicated that the human HIF1α mutant was transcriptionally active (Supplemental Figure 2D). Cells stably expressing this degradation-resistant mutant of HIF1α exhibited a decrease in ENaC-mediated transepithelial current (Supplemental Figure 2B) correlated with a downregulation of ENaC β- and γ-subunit expression for both mRNA and protein levels (Supplemental Figure 2, A and C). Decreased αENaC abundance was observed at the protein level only and may reflect enhanced degradation of nonassembled αENaC proteins. Therefore, HIF1α stabilization resulted in the inhibition of sodium transport and downregulation of ENaC subunits in mpkCCD_cl4 cells.

Stimulation of Sodium Transport Activates the HIF Pathway in Mouse CCD Principal Cells

The renal tubule cell metabolism is a major determinant of renal oxygen consumption.³⁶ Stimulation of sodium reabsorption increases ATP consumption and thereby oxygen consumption through mitochondrial ATP synthesis. This increased oxygen consumption may generate cellular hypoxia and/or increased generation of ROS which may activate the HIF pathway.³⁷

We assessed the effect of aldosterone on mpkCCD_cl4 cells. Aldosterone (1 μM) for 24–48 hours stimulated sodium transport as shown by enhanced transepithelial current associated with increased aldosterone targets expression levels, such as SGK1, αENaC, and Na⁺,K⁺-ATPase α- and β-subunits (Figure 5, A–D). Aldosterone progressively increased HIF1α protein abundance that correlated with upregulation of HIF target genes mRNA levels, such as PDK1, PGK1, VEGF, and GLUT1 (Figure 5, C, D, and E). To determine whether induction of the HIF pathway by aldosterone relied on the mineralocorticoid receptor (MR), mpkCCD_cl4 cells were pretreated with 10 μM 17-α-OH progesterone (17OHP) before exposure to aldosterone.²⁹ As expected, 17OHP abolished both sodium transport stimulation and HIF activation by aldosterone (Supplemental Figure 3). To confirm that HIF activation resulted from the increase in sodium transport, mpkCCD_cl4 cells stimulated or not with aldosterone were exposed for 24 hours to a low apical sodium concentration (15 mM) that decreased ENaC-mediated sodium transport under both control and aldosterone-treated conditions as compared with the regular apical sodium concentration (150 mM) (Figure 5F). Aldosterone increased αENaC mRNA levels to a similar extent in the presence of either low (15 mM) or regular (150 mM) apical sodium concentration (Figure 5G), whereas upregulation of the HIF target gene mRNA levels, such as PDK1 and PGK1, was abolished when sodium transport was decreased by low apical sodium, as compared with regular apical sodium (Figure 5H). Therefore, aldosterone activates the HIF signaling pathway via the stimulation of sodium transport in mpkCCD_cl4 cells. Similar activation of the HIF signaling pathway assessed by the upregulation of HIF target genes in response to aldosterone (1 μM) was observed in mCCD_cl1 cells, a second mouse CD principal cell model (Supplemental Figure 4).

Figure 5.

Stimulation of sodium transport by aldosterone activates the HIF signaling pathway in cultured CD principal cells. mpkCCD_cl4 cells were treated with aldosterone (1 μM) for 24 and 48 hours to stimulate active sodium transport. (A) Transepithelial current increased in cells treated with aldosterone (Aldo), as compared with control cells. Results are means±SEM from three independent experiments. (B) Upregulation of the aldosterone target genes, αENaC, and α (NaKα) and β subunits (NaKβ) of the Na,K-ATPase, assessed by real-time PCR. (C and D) HIF pathway activation in cells treated with aldosterone. Increased HIF1α protein abundance assessed by western blot. β-actin was used as a loading control. A representative experiment from three is shown (C). (D) Western blots were quantified using ImageJ software. Results are expressed as a percentage of control values (dots) and are means±SEM (lines) from three experiments in each group. (E) Aldosterone treatment increased mRNA levels of the HIF target genes PDK1, PGK1, VGEF, and GLUT1 assessed by real-time PCR. Results are expressed as fold of control (dots) and are means±SEM (lines) from three independent experiments. (F) Exposure to low–apical sodium medium, i.e., 15 mM (Na15), instead of high–apical sodium medium, i.e., 150 mM (Na150), for 6 hours decreased transepithelial current in mpkCCD_cl4 cells treated or not with aldosterone. Results are means±SEM from three independent experiments. (G) Effect of aldosterone on αENaC mRNA levels assessed by real-time PCR in cells exposed to low (Na15) or high (Na150) apical sodium is unchanged. (H) HIF target genes PDK1 and PGK1 mRNA levels decreased in cells exposed to low–apical sodium medium (Na15), as compared with high apical sodium (Na150), under both control and aldosterone-treated conditions. Results are expressed as fold of control (dots) and are means±SEM (lines) from three independent experiments. Unpaired Student’s t test was used to compare transepithelial currents from control and aldosterone-treated cells in (A). One-way ANOVA was used to compare data from (B), (D), (E), (F), (G), and (H). *P<0.05; **P<0.01; ***P<0.001; ****P<0.0001, versus controls for (A), (B), (D), and (E).

Effect of ATP Depletion on HIF Pathway Activation in Cultured CD Principal Cells

We then assessed the effect of inhibition of the mitochondrial respiratory chain by either rotenone, which blocks complex I, or antimycin A, which blocks complex III. Both compounds abolish mitochondrial ATP synthesis and induce a large increase in ROS production.³⁸ Seahorse experiments showed that both compounds efficiently inhibited the respiratory chain (Supplemental Figure 5A) and thereby induced partial ATP depletion in mpkCCD_cl4 cells. Rotenone or antimycin A, which block mitochondrial respiration (Supplemental Figure 5A), only weakly inhibited ENaC-dependent transepithelial sodium transport, suggesting that increased glycolysis compensated the defective mitochondrial ATP synthesis or that glycolysis is the major pathway of ATP synthesis in these cells. Inhibition of mitochondrial respiration for 24 hours slightly decreased γENaC protein abundance (Supplemental Figure 5, B and C) and induced HIF1α stabilization that may rely on increased ROS generation secondary to mitochondrial uncoupling.³⁷

To further test the role of ATP depletion of HIF pathway activation in CD principal cells, we first assessed the effect of 2-deoxy-d-glucose on HIF activation and γENaC protein level. Results show that blocking glycolysis did not alter HIF1α or γENaC protein levels in mpkCCD_cl4 cells (Supplemental Figure 5, E and F). Because ATP depletion activates AMPK, we assessed the effect of AICAR, a classic AMPK activator, on ENaC activity and expression. Incubation of mpkCCD_cl4 cells with 1 mM AICAR for 6 and 24 hours increased ACC phosphorylation, confirming AMPK activation, but altered neither amiloride-sensitive transepithelial current nor γENaC protein abundance (Supplemental Figure 6). This result is at variance with a previous report showing inhibition of ENaC by AMPK.³⁹ This discrepancy may rely on the use of different cell subclones and/or on different experimental conditions. Altogether, our results indicate that ATP depletion led to neither HIF pathway activation nor γENaC downregulation in CD principal cells.

Kidney Tubule–Specific Hif1α Knockout Does Not Alter Baseline Sodium Transporter Abundance in Mouse Kidney

To investigate the role of the HIF signaling pathway on renal sodium handling and to determine the specific role of HIF1α, we generated inducible and kidney tubule–specific knockout mice for Hif1a (Hif-1a^PAX8/LC1). To delete HIF1α, we treated 3-week-old Hif1a^PAX8/LC1 mice and their control littermates with doxycycline (Supplemental Figure 7A). HIF1α mRNA levels were reduced by about 90% in renal cortex from Hif1α^PAX8/LC1 knockout mice (Supplemental Figure 7B). Western blotting experiments revealed that HIF1α protein was not detected in renal cortex of Hif1α^PAX8/LC1 mice exposed to hypoxia (8% O₂) and confirmed that HIF1α is mostly expressed in kidney tubule epithelial cells (Supplemental Figure 7C).

Baseline diuresis, urinary sodium excretion, and BP were similar in wild-type and Hif1α^PAX8/LC1 mice, suggesting that sodium balance is not altered by the absence of tubular HIF (Figure 6, A and B). Measurement of the GFR did not reveal any difference between wild-type and Hif1α^PAX8/LC1 mice (Figure 6C), indicating that the filtered load of sodium is identical in each mouse strain. We compared the protein abundance of the major kidney tubule sodium transporters in wild-type and Hif1α^PAX8/LC1 mice. Results show that tubular HIF1α knockout did not alter NHE3, NKCC2, NCC, and ENaC subunits expression levels (Figure 6, D and E). Similarly, claudin-8 and AQP2 abundance, taken as surrogates of paracellular permeability and transcellular water transport, were not altered under these baseline conditions in Hif1α knockout mice (Figure 6, D and E). These results indicate that the HIF pathway does not control kidney tubule sodium and water transport under basal conditions.

Figure 6.

Analysis of renal sodium transporter abundance in inducible and kidney tubule–specific HIF1α knockout mice under baseline conditions. Hif1α^PAX8/LC1 knockout mice and their control littermates (Hif1α^lox/lox) were housed in metabolic cages for 6 days with liberal access to food and water (n=5–10 per group). (A) The 24-hour urinary volume (left panel) and urinary sodium excretion (right panel). (B) Systolic (left panel) and diastolic (right panel) BP. (C) GFR. (D) Western blot analysis of renal sodium transporter abundance in renal cortex from Hif1α^PAX8/LC1 knockout mice and their respective control mice. β-actin was used as a loading control. (E) Western blots were quantified using ImageJ software. Results are expressed as a percentage of control values (dots) and are means±SEM (lines) from five animals in each group. Unpaired t test was used to compare control and HIF1α knockout mice in (A), (B), (C), and (E).

HIF Pathway Activation by Hypoxia Regulates Renal Sodium Transporter Abundance

We next assessed whether the HIF pathway may control kidney tubule sodium transport under pathologic hypoxic conditions. Wild-type mice (control) were exposed to either normoxic (21% O₂) or hypoxic (8% O₂) conditions for 48 hours and we analyzed the consequences on renal tubule sodium transporters expression. Activation of the HIF pathway by hypoxia in the mouse renal cortex was confirmed by increased HIF1α protein abundance, indicating that hypoxia-induced HIF protein stabilization (Supplemental Figure 7C). Results showed that hypoxia altered the pattern of segment-specific apical sodium transporter expression (Figure 7, A–C). The protein abundance of NHE3 and total and phosphorylated NCC was increased. In contrast, the abundance of NKCC2 and both α- and γ-subunits of ENaC was decreased. Claudin-8 abundance was decreased, as expected from its coupling to γENaC abundance,²³ and AQP2 abundance was not significantly altered (Figure 7, A and B).

Figure 7.

Hypoxic conditions modulate kidney tubule sodium transporter abundance in wild-type and HIF1α renal tubule–specific knockout mice. Hif1a^PAX8/LC1 knockout mice and their control wild-type littermates (Hif1a^lox/lox) were exposed or not to hypoxia (8% O₂) for 48 hours (n=4 per group). Sodium transporter proteins were detected by western blot from renal cortex lysates (A–C) and results were quantified using ImageJ software (B–D). β-actin was used as a loading control. Results are expressed as a percentage of control values (dots) and are means+SEM (lines) from four animals in each group. Unpaired Student’s t test was used to compare control and hypoxic mice. *P<0.05; **P<0.01; ***P<0.001.

To assess the role of HIF1α in the alteration of renal sodium transporter abundance in response to hypoxia, Hif1α ^PAX8/LC1 knockout mice and their wild-type littermates were exposed to normoxia (21% O₂) or hypoxia (8% O₂) for 48 hours. In kidney cortices of tubule-specific HIF1α knockout mice, the increase in NHE3 abundance in response to hypoxia was more pronounced and, remarkably, γENaC protein was upregulated (Figure 7, C and D). The alterations in γENaC abundance observed in response to hypoxia were confirmed by immunofluorescence in both wild-type and Hif1α knockout mice (Supplemental Figure 8).

The increased abundance of total and phosphorylated NCC in response to hypoxia was not significantly changed in HIF1α tubular cell–specific knockout (Figure 7, C and D). This increase in abundance and phosphorylation of NCC in response to systemic hypoxia may result from hyperventilation which generates a respiratory alkalosis and associated hypokalemia (see Table 1).⁴⁰ Both claudin-8 and AQP2 abundance increased in parallel with γENaC in hypoxic Hif1α knockout mice (Figure 7, C and D). In the physiologically hypoxic inner medulla, exposure to 8% O₂ for 48 hours did not alter ENaC subunits or claudin-8 abundance, whereas γENaC and claudin-8 protein abundance tended to be increased in hypoxic Hif1α knockout mice (Supplemental Figure 9).

Results confirm that HIF1α controls γENaC abundance in vivo and that its expression modulates the abundance of a specific subset of sodium and water transporters in response to systemic hypoxia.

High-Sodium Diet Activates HIF Pathway

Our results show that increased sodium transport upregulates HIF1α in CD (Figure 5). We therefore challenged Hif1α ^PAX8/LC1 knockout mice and their littermates with high- (3%) and low- (0.1%) salt diet. In wild-type mice given a low-salt diet for 1 week, Hif1α protein abundance was not increased in kidney cortex (Supplemental Figure 10), suggesting that the HIF pathway was not activated. In this context, Hif1α knockout did not alter the abundance of sodium and water transporters (Figure 8, A and B).

Figure 8.

HIF1a knockout upregulates gENaC subunit under high salt diet. Analysis of renal sodium transporter abundance in cortex of inducible and kidney tubule–specific HIF1α knockout mice under low- and high-sodium diet. Hif1a^PAX8/LC1 knockout mice and their control wild-type littermates (Hif1a^lox/lox) were fed with low-sodium diet (0.1%; LSD, [A] and [B]) or high-sodium diet (3%; HSD, [C] and [D]). Sodium transporter proteins were detected by western blot from kidney cortex lysates (A–C) and results were quantified using ImageJ software (B–D). E-cadherin was used as a loading control for LSD. β-actin was used as a loading control for HSD. Results are expressed as a percentage of control values (dots) and are means+SEM (lines) from three (LSD) or four (HSD) animals in each group. Unpaired Student’s t test was used to compare control and HIF1α knockout mice. *P<0.05; **P<0.01; ***P<0.001.

Wild-type mice given a high-salt diet for 1 week displayed increased Hif1α protein abundance (Supplemental Figure 10), indicating HIF pathway activation. In Hif1α knockout mice, this high-salt diet led to upregulation of γENaC (Figure 8, C and D). The absence of detectable cleaved γENaC is explained by the high-salt diet which inhibits aldosterone secretion and prevents γENaC cleavage. In contrast with Hif1α knockout mice exposed to hypoxia, claudin-8 and AQP2 abundance were not significantly increased (Figure 8, C and D). The absence of clear claudin-8 upregulation might be explained by the downregulation of aldosterone secretion by the high-salt diet²⁹ and inhibition of the aldosterone-dependent claudin-8 expression.⁴¹ Inhibition of vasopressin secretion by increased osmotic load (3% NaCl) may explain the absence of AQP2 upregulation.

Discussion

Our results demonstrate that HIF signaling is part of the physiologic network that controls the ENaC-dependent sodium reabsorption process in CD principal cells. In CD principal cells, we showed that activation of HIF inhibits transepithelial sodium transport. This effect is mediated by inhibition of both apical sodium entry via ENaC and basolateral sodium exit via Na⁺,K⁺-ATPase. This coordinated control may preserve intracellular sodium concentration and decrease ATP and oxygen consumption.

In cultured CD principal cells activation of HIF induced the downregulation of both β- and γ-subunits of ENaC. In vivo experiments showed that HIF activation significantly decreases γENaC abundance but not βENaC. The reasons for this discrepancy are unclear but may rely on the complexity of ENaC regulation which relies on both cell autonomous and exogenous factors,⁴² the latter being absent under cell culture conditions. It should be mentioned that CD principal cells and their closely related connecting tubule cells represent a tiny fraction of kidney epithelial cells. Therefore, changes in Na⁺,K⁺-ATPase abundance along the collecting system cannot be detected in whole kidney cortex extracts.

The effect of HIF on ENaC and Na⁺,K⁺-ATPase subunits expression is most likely indirect. Indeed, HIF is a transcriptional activator, and although HIF pathway activation may lead to inhibitory effects, these are to our knowledge indirect effects via induction of transcriptional repressors or miRNA.⁴³ To determine whether ENaC subunits share the typical footprints of HIF target genes, we performed in silico analysis of the mouse genomic DNA sequences 2000 base pairs upstream of the start codon of ENaC subunits (Supplemental Figure 11). We found only one putative binding site located relatively far away from the start codon in both α- and βENaC (scnn1a and b). In contrast, several potential HIF binding sites were identified along the putative promoter sequence of γENaC (scnn1g), but they are far more distant from the start codon than binding sites observed in bona fide HIF target genes such as Pdk1, Pgk1, and Slc2a1 (GLUT1). These observations suggest that ENaC subunits are not typical HIF-induced genes. Therefore, additional studies are required to determine the actual HIF target(s) leading to downregulation of sodium transport.

Our results, which clearly indicate that the HIF pathway modulates the expression of ENaC subunits in CD principal cells, are at variance with a previous study by Husted et al.⁴⁴ showing that oxygen modulates ENaC activity independently of both HIF and AMPK pathways. It should be mentioned that, firstly, we used much more drastic hypoxic conditions (close to anoxia versus 8% O₂) and, secondly, we observed a clear effect on ENaC subunit mRNA and protein levels, although Husted et al. mostly observed a functional effect. However, these two oxygen-dependent mechanisms of ENaC regulation are not mutually exclusive. On the other hand, our results strongly suggest that ATP depletion does not significantly contribute to the downregulation of ENaC subunits by hypoxia. Indeed, blockade of the mitochondrial respiratory chain, which almost immediately blocks mitochondrial ATP synthesis, did not alter γENaC abundance and ENaC-dependent sodium transport before 24 hours. Similarly, ATP depletion by inhibition of glycolysis by 2-deoxy-d-glucose neither activated HIF nor decreased γENaC expression. The delayed activation of HIF and downregulation of ENaC expression by aldosterone or mitochondrial respiratory chain inhibitors are rather compatible with an effect mediated via ROS accumulation.¹⁴

HIF activation remains undetectable in kidney cortex of mice under a standard or a low-sodium diet. This observation suggests that under these conditions, oxygen supply is sufficient to fuel mitochondrial ATP synthesis by kidney tubule cells. Although stimulation of ROS generation by angiotensin II and aldosterone^45,46 may locally activate the HIF pathway, this effect is not large enough to be detected at the level of the kidney cortex. However, under conditions of high sodium intake, energy consumption increases²⁹ and oxygen demand may exceed oxygen delivery and result in hypoxia.¹⁴ Alternatively, stimulation of mitochondrial ATP synthesis by increased active sodium transport in distal nephron segments may generate ROS¹² which activate the HIF pathway and downregulate transepithelial sodium transport. Activation of the HIF pathway may induce both metabolic adaptation and specific downregulation of apical sodium transporters in order to decrease ATP and oxygen consumption and ROS generation, to prevent tubular lesions.⁴⁷

This might be especially true in CD which expresses higher levels of HIF1α than proximal tubules and thick ascending limb.

In conclusion, our study shows that in CD principal cells, the HIF pathway specifically downregulates the abundance of γENaC leading to decreased ENaC activity and transepithelial sodium transport. This pathway may constitute a physiologic negative feedback on active sodium reabsorption to adjust metabolic demand and oxygen supply.

Disclosures

I.J. Frew reports Research Funding from Merck Healthcare. J. Loffing reports Ownership Interest in Novartis, Roche, and VIFOR; Honoraria from VIFOR; Patents and Inventions: receiving royalties for licensed antibodies (Abcam, Chemicon/Millipore); Scientific Advisor or Membership: editorial board American Journal of Physiology, Histochemistry and Cell Biology, and Pflügers Archiv; and Other Interests/Relationships with the American Society of Nephrology, the American Society of Physiology, Anatomische Gesellschaft, the Swiss Society for Anatomy, Histology, and Embryology, and the Swiss Society of Nephrology. All remaining authors have nothing to disclose.

Funding

This work was supported by the National Center of Competence in Research Kidney Control of Homeostasis and Swiss National Science Foundation (Schweizerischer Nationalfonds) grants 31003A_175471/1 to E. Feraille, 310030_143929/1 to J. Loffing, 31003A_144198/1 to E. Hummler, and 310030_184813 to R.H. Wenger. B.R. Kwak and S. Morel received a grant from the Fondation Ernst et Lucie Schmidheiny.

Supplemental Material

This article contains the following supplemental material online at http://jasn.asnjournals.org/lookup/suppl/doi:10.1681/ASN.2021010046/-/DCSupplemental.

Supplemental Figure 1. Relative abundance of HIF1α mRNA levels in microdissected mouse proximal tubules (PT), thick ascending limbs of Henle (TAL), and cortical collecting ducts (CCDs).

Supplemental Figure 2. Expression of a constitutively stable mutant of HIF1α, HA-HIF1α^{(P402A/P564A)}, decreases sodium transport via decreased ENaC subunit expression levels in cultured collecting duct principal cells.

Supplemental Figure 3. Activation of the HIF pathway by aldosterone is MR-dependent in mpkCCD₄ cells.

Supplemental Figure 4. Activation of the HIF pathway by aldosterone in mCCD_cl1 cells.

Supplemental Figure 5. Effect of inhibition of the mitochondrial respiratory chain and of glycolysis on HIF1a and γENaC abundance in mpkCCD_cl4 cells.

Supplemental Figure 6. AMPK activation does not alter the HIF pathway in mpkCCD_cl4 cells.

Supplemental Figure 7. Characterization of inducible and tubular-specific HIF1α KO mice.

Supplemental Figure 8. Immunofluorescence staining of γENaC in cortices from wild-type and tubule-specific HIF1α knockout mice under baseline and hypoxic conditions.

Supplemental Figure 9. Analysis of renal sodium transporter abundance in inner medulla from wild-type and tubule-specific HIF1α knockout mice under baseline and hypoxic conditions.

Supplemental Figure 10. Dietary conditions activate the HIF signaling pathway in mouse renal cortex.

Supplemental Figure 11. Diagrams of the putative promoter region (2 kb) of the mouse Scnn1 genes, Pgk1, Pdk1, and Slc2a1, three genes highly induced by hypoxia.

Supplemental Table 1. Average Ct in microdissected tubules.