Loss of sodium chloride co-transporter impairs the outgrowth of the renal distal convoluted tubule during renal development

Abstract

Background

Loss-of-function mutations in the sodium chloride (NaCl) co-transporter (NCC) of the renal distal convoluted tubule (DCT) cause Gitelman syndrome with hypokalemic alkalosis, hypomagnesemia and hypocalciuria. Since Gitelman patients are usually diagnosed around adolescence, we tested the idea that a progressive regression of the DCT explains the late clinical onset of the syndrome.

Methods

NCC wild-type and knockout (ko) mice were studied at Days 1, 4 and 10 and 6 weeks after birth using blood plasma analysis and morphological and biochemical methods.

Results

Plasma aldosterone levels and renal renin messenger RNA expression were elevated in NCC ko mice during the first days of life. In contrast, plasma ion levels did not differ between genotypes at age 10 days, but a significant hypomagnesemia was observed in NCC ko mice at 6 weeks. Immunofluorescent detection of parvalbumin (an early DCT marker) revealed that the fractional cortical volume of the early DCT is similar for mice of both genotypes at Day 4, but is significantly lower at Day 10 and is almost zero at 6 weeks in NCC ko mice. The DCT atrophy correlates with a marked reduction in the abundance of the DCT-specific Mg²⁺ channel TRPM6 (transient receptor potential cation channel subfamily M member 6) and an increased proteolytic activation of the epithelial Na⁺ channel (ENaC).

Conclusion

After an initial outgrowth, DCT development lags behind in NCC ko mice. The impaired DCT development associates at Day 1 and Day 10 with elevated renal renin and plasma aldosterone levels and activation of ENaC, respectively, suggesting that Gitelman syndrome might be present much earlier in life than is usually expected. Despite an early downregulation of TRPM6, hypomagnesemia is a rather late symptom.

DCT development, Gitelman syndrome, NaCl co-transporter, renal development, transient receptor potential melastatin type 6

INTRODUCTION

The renal distal convoluted tubule (DCT) plays a crucial role in the maintenance of extracellular fluid volume, regulation of arterial blood pressure and electrolyte homeostasis. Sodium reabsorption in the DCT constitutes 5–10% of the entire sodium absorption along the nephron and is primarily mediated by the thiazide-sensitive sodium chloride (NaCl) co-transporter (NCC) [1], which is uniquely expressed in the DCT. NCC is the target of thiazide and thiazide-like diuretic drugs that are commonly used for long-term treatment of arterial hypertension [2]. The importance of the DCT and NCC to human physiology is also confirmed by Gitelman syndrome, a genetic disease caused by loss-of-function mutations in NCC leading to an autosomal recessive renal tubulopathy characterized by hypokalemic alkalosis, hypomagnesemia, hypocalciuria, mild renal salt wasting and normal to low arterial blood pressure [3, 4].

Patients suffering from Gitelman syndrome are considered to be rarely symptomatic (muscle weakness and spasms, fatigue due to potassium and magnesium depletion) during childhood and the disease is generally diagnosed during adolescence or early adulthood [5]. The late clinical onset of Gitelman syndrome is in contrast to Bartter syndrome, which represents a group of inherited renal tubulopathies related to mutations in various genes [i.e. the Na-K-2Cl co-transporter NKCC2, the K⁺ channel ROMK, the Cl⁻ channel ClC-Kb or its subunit Barttin BSND and the calcium-sensing receptor (CSR)] that finally lead to impaired NaCl reabsorption in the thick ascending limb of Henle’s loop [6]. Patients with Bartter syndrome suffer from severe electrolyte disturbances, with tetany, polyuria and volume depletion that usually become symptomatic in the prenatal period [7]. The reasons for the rather late onset of the clinical manifestation in most Gitelman patients remains elusive.

Schultheis et al. [8] generated a NCC null mutant mouse model. These mice show a renal phenotype that recapitulates many of the findings in human Gitelman patients, including hypomagnesemia, alkalosis and hypocalciuria [8, 9]. Moreover, distinctive alterations in the epithelial structure of the DCT, showing marked atrophy of the DCT due to an almost complete absence of the early DCT portion (DCT1), were demonstrated in NCC knock-out (KO) mice [9]. The loss of a significant portion of the DCT epithelium was accompanied by a marked reduction of the abundance of the DCT-localized apical Mg²⁺ channel TRPM6 (transient receptor potential cation channel subfamily M member 6), which was suggested to explain the renal Mg²⁺ wasting seen in NCC-deficient mice and patients with Gitelman syndrome [10]. Moreover, adult NCC-deficient mice show a marked structural hypertrophy of the connecting tubule (CNT), which goes along with an upregulation of the epithelial Na⁺ channel (ENaC) in the kidney [9, 11], which may not only compensate for the loss of NCC activity in the DCT, but may also contribute to the susceptibility of NCC-deficient mice [12] and Gitelman patients for hypokalemia [3, 4].

The marked epithelial remodelling of the distal tubule described above may constitute a morphological basis for the pathophysiology of Gitelman syndrome. Although the underlying mechanism for the marked structural atrophy of the DCT is unclear, the loss of the DCT epithelium likely reflects the altered ion transport activity due to the loss of NCC. Consistently also, the inhibition of NCC with thiazide diuretics causes DCT cell apoptosis [13]. Interestingly, a loss of carbonic anhydrase II (CADII) in CADII knockout (ko) mice is also accompanied by a depletion of the cells usually expressing this particular gene (i.e. the intercalated cells in the renal collecting system) [14]. The loss of intercalated cells in CADII-deficient mice starts at about Day 11 after birth, with a bisection of intercalated cells by 6 weeks of age and an almost complete loss of intercalated cells by the age of 3–5 months [15]. Thus, in CADII ko mice, the depletion of intercalated cells occurs quite late. The time course of the occurrence of the DCT atrophy in NCC-deficient mice is not defined. This study addresses the question whether a rather late onset of DCT atrophy may explain the late clinical presentation of Gitelman syndrome. To reach this end, we evaluated DCT development and the expression of key distal tubule ion transporters and channels in NCC wild-type (wt) and ko mice at Days 1, 4 and 10 and Week 6 after birth.

MATERIALS AND METHODS

Animals

The generation of NCC-deficient mice has been described previously [8]. All mice used in this study had a C57/Bl6 background. Animals were bred in a standard, nonspecific pathogen-free animal facility. All experiments involving living animals were conducted in accordance with Swiss laws and approved by the veterinary administration of the Canton of Zurich, Switzerland.

Tissue processing

Mice at the age of 1, 4 or 10 days were anesthetized with isoflurane and killed by decapitation. The kidneys were then immediately removed for subsequent analysis. For immunohistochemistry, the kidneys were immersion fixed with paraformaldehyde (PFA) 3% for 12–16 h, incubated in 0.1 M phosphate buffer (pH 7.4, 300 mOsm) for 4 h and embedded in paraffin. For RNA extraction and western blot analysis, the kidneys were frozen in liquid nitrogen and stored at −80°C until future use. Mice at the age of 6 weeks were anesthetized with isoflurane and the kidneys were fixed by retrograde abdominal aortic perfusion of 3% PFA followed by rinsing with 0.1 M phosphate buffer (pH 7.4, 300 mOsm) and paraffin embedding. For RNA extraction and western blot analysis, perfusion was performed with phosphate-buffered saline (PBS) via the left ventricle of the heart and the kidneys were removed, frozen in liquid nitrogen and stored at −80°C.

Immunohistochemistry

Paraffin tissue sections of 4 µm were deparaffinized using Histo-clear and rehydrated through serial ethanol washes (Pathisto AS-2 automatic slide stainer). For epitope retrieval, slides were heated in 10 mM citrate buffer (pH 6.0) in a microwave (98°C for 10 min). After blocking for 10 min with 10% normal goat serum in PBS with 1% bovine serum albumin (BSA), the sections were incubated overnight in a humidified chamber at 4°C with the primary antibodies and diluted in PBS with 1% BSA. The primary antibodies and dilutions used for immunohistochemistry are shown in Supplementary data, Table S1. Secondary antibodies [Cy3-conjugated goat-anti-rabbit immunoglobulin G (IgG) (catalog code 111-165-144), dilution 1:1000 and fluorescein isothiocyanate–conjugated goat-anti-mouse IgG (catalog code 115-095-068), dilution 1:100, both from Jackson Immuno Research Laboratories, West Grove, PA, USA] were applied to the slides for 2 h at room temperature and diluted in PBS with 1% BSA. After final washing with PBS, sections were mounted using glycergel and coverslips. Images were acquired with a Leica DM6000 B fluorescence microscope on a Leica DFC350 FX fluorescence monochrome digital camera (Leica Microsystems, Wetzlar, Germany).

Proliferating cell nuclear antigen (PCNA) labelling index

PCNA-positive cell nuclei were detected in paraffin sections using immunohistochemistry as described above. DCT1 segments were identified by parvalbumin (PV)-positive immunostaining. The number of PCNA-positive DCT1 cell nuclei within a kidney cross section was counted and expressed as a percentage of all 4′,6-diamidino-2-phenylindole (DAPI)-stained cell nuclei in the DCT1 segments.

Terminal deoxynucleotidyl transferase (TdT)-mediated dUTP nick-end labeling (TUNEL) assay

Apoptotic cell nuclei were detected in paraffin sections with the TdT-mediated dUTP-biotin nick end labeling (TUNEL) method using In Situ Cell Death Detection Kit, Fluorescein (Roche, Basel, Switzerland), according to the manufacturer’s instructions.

Morphometric measurements

The fractional tubular volume of DCT segments was assessed using a planimetric point-counting method according to a previously described protocol [16]. In brief, PFA-fixed kidneys were cut into halves perpendicular to their longitudinal axis. After embedding in paraffin, the kidneys were cut into 4-µm thick sections and stained with polyclonal antibodies against NCC (NCC wt samples) and PV (NCC wt and NCC ko samples) as described above. Overviews were taken from each section at the microscope using the 10× objective. DCTs were identified according to their specific antibody-staining pattern. Morphometric analysis was then performed with a computerized image analysis system (Stereo Investigator, MBF Bioscience, Williston, ND, USA). A transparent grid with a distance of 50 μm between lines was electronically superimposed on the micrographs. Cortical areas extending between the renal capsule and the outer medullary boundary, identified by arcuate arteries, were evaluated. The intersections of grid lines falling on the total cortex area and those intersections falling on DCT segments were counted. The proportion of DCTs in the cortex was calculated as the percentage of DCT counts versus the total number of cortex counts and expressed as the fractional cortical volume of the DCT.

RNA isolation and quantitative polymerase chain reaction (qPCR)

Kidneys were harvested as described above. Total RNA from kidney tissue was isolated using the ReliaPrep RNA Tissue Miniprep System for the kidneys of 1, 4 and 10-day-old mice and the SV Total RNA Isolation System for the kidneys of 6-week-old mice (both from Promega, Madison, WI, USA), according to the manufacturer’s protocol. A total of 500 ng of total RNA was used for the generation of complementary DNA (cDNA) using the GoScript Reverse Transcription System (Promega) as per the manufacturer's protocol. This cDNA was further diluted to 1:5 and expression levels of PV, transient receptor potential melastatin type 6 (TRPM6) cation channel and renin were quantified by real-time qPCR using a Light Cycler II 480 (Roche). Relative gene expression was determined after normalization to ribosomal RNA gene expression. The primers used for reverse transcription qPCR (RT-qPCR) are listed in Supplementary data, Table S2.

Plasma ion measurement

Blood collection. In 6-week-old mice, blood was sampled by puncture of the inferior vena cava of isoflurane-anesthetized animals. In 10-day-old pups, trunk blood was collected after anesthesia with isoflurane and decapitation. Whole blood was centrifuged for 10 min at 6000 rpm and plasma was removed and frozen at −80°C until further use.

Plasma Na⁺, K⁺ and Ca²⁺ measurement. In 6-week-old mice, blood ions (Na⁺, K⁺, Ca²⁺) were measured with the ABL80Flex Blood Gas Analyzer (Radiometer, Copenhagen, Denmark). In 10-day-old pups, plasma electrolytes (Na⁺, K⁺, Ca²⁺) were measured by flame photometry (EFOX 5053, Eppendorf, Hamburg, Germany).

Plasma Mg²⁺ measurement. Plasma Mg²⁺ measurements were performed by the Zurich Integrative Rodent Physiology facility (University of Zurich) using photometric methods (UniCel DxC 800, Beckman Coulter, Brea, CA, USA).

Aldosterone measurement

Blood of 10-day-old pups was collected as described above. Plasma aldosterone was measured by an enzyme-linked immunosorbent assay (ELISA) (Aldosterone ELISA Kit, item no. 501090, Cayman Chemical, Ann Arbor, MI, USA) according to the manufacturer’s instructions.

Western blot analysis

Sample preparation. Kidneys were homogenized in ice-cold lysis buffer [mannitol 200 mM, HEPES (4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid) 80 mM, potassium hydroxide 41 mM] containing protease inhibitor (Complete Ultra, Roche) and phosphatase inhibitor (PhosSTOP, Roche). Tissue homogenization was performed using MagNA Lyser Green Beads (Roche) and a Precellys 24 tissue homogenizer (Bertin Instruments, Montigny-le-Bretonneux, France) at 2 × 20 s (5000 rpm). The homogenized samples were centrifuged for 10 min (4600 rpm) and the protein-containing supernatant was removed and stored at −80°C.

Sodium dodecyl sulfate–polyacrylamide gel electrophoresis. Equal amounts of protein (25–50 µg) from kidneys of wt and ko mice were denatured in Laemmli buffer (sodium dodecyl sulfate, β-mercaptoethanol, bromophenol blue, glycerol and 0.5 M Tris-HCl buffer, pH 6.8) and loaded on 8–12% polyacrylamide gels.

Transfer and antibody incubation. After electrophoretic separation, proteins were transferred to a nitrocellulose membrane. Equal protein loading and electrophoretic transfer were confirmed using REVERT Total Protein Stain (LI-COR Biosciences, Lincoln, NE, USA) and visualization of protein bands with an Odyssey infrared imaging system (LI-COR Biosciences) prior to antibody incubation. The membrane was blocked for 10 min with Odyssey blocking solution (LI-COR Biosciences) and then incubated at 4°C for 16 h with the primary antibodies (Supplementary data, Table S1) diluted in Odyssey blocking buffer and PBS (1:5). After repeated washes with PBS, blots were incubated at room temperature for 2 h with secondary antibodies [goat-anti-rabbit IRDye 800 (product no. 926-32211) and goat-anti-mouse IRDye 680 (product no. 926-32220), 1:20 000; both from LI-COR Biosciences] in Casein blocking solution and water (1:10). Quantification of immunoreactive bands was carried out with the Odyssey infrared imaging system and software (LI-COR Biosciences) and normalized for the total protein signal obtained for each sample after the electrophoretic transfer (see above).

Table 1

Plasma electrolytes in NCC wt and NCC ko mice at age 10 days and 6 weeks

	10 days		6 weeks
	NCC wt	NCC ko	NCC wt	NCC ko
Plasma Na+ (mmol/L)	141.0 ± 15.78	140.1 ± 21.36	147.3 ± 2.06	147.6 ± 1.81
Plasma K+ (mmol/L)	7.85 ± 1.65	8.54 ± 1.32	3.68 ± 0.30	3.78 ± 0.25
Plasma Ca2+ (mmol/L)	2.53 ± 0.25	2.59 ± 0.51	1.13 ± 0.16	1.15 ± 0.11
Plasma Mg2+ (mmol/L)	0.96 ± 0.15	0.97 ± 0.09	1.07 ± 0.03	0.75 ± 0.06*

Plasma Na⁺ (mmol/L), K⁺ (mmol/L), Ca²⁺ (mmol/L) and Mg²⁺ (mmol/L) in NCC wt versus NCC ko mice at the age of 10 days (n = 11) and 6 weeks (n = 5 for Na⁺, K⁺ and Ca²⁺ and n = 3 for Mg²⁺). Values are mean ± SD. Plasma Mg²⁺ (mmol/L) in NCC wt versus NCC ko mice at the age of 6 weeks reveals significant hypomagnesemia in NCC ko mice (*P = 0.00007).

Generation of affinity purified rabbit-anti-mouse TRPM6 antibody

The new antibody against TRPM6 was obtained by immunizing rabbits (Pineda Antikörper-Service, Berlin, Germany) with keyhole limpet-hemocyanin-coupled synthetic peptides corresponding to the amino acid sequence within mouse TRPM6 (NH2-CERDKNRSSLEDHTRL-COOH). After affinity purification against the immunizing peptide (Pineda Antikörper Service), the antibody was characterized by using standard western blotting with WT mouse kidney samples (as described above). Preincubation of the antiserum (concentration 1:2000 in PBS) with the peptide used for immunization (concentration 20 µg/mL in PBS) for 2 h at room temperature with subsequent centrifugation (13 000 rpm for 10 min) completely inhibited the specific binding of the antibody to the nitrocellulose western blot membrane (Supplementary data, Figure S1).

Statistical analysis

Unpaired two-tailed t-test, one-way analysis of variance (ANOVA) and two-way ANOVA were used to compare the groups (GraphPad Prism, version 8.0.2, GraphPad Software, San Diego, CA, USA). Data are given as mean ± standard deviation (SD) (Table 1) and mean ± standard error of the mean (SEM) (in Figures 1–6). P-values <0.05 were considered statistically significant.

RESULTS

Overt hypomagnesemia in NCC ko mice at age 6 weeks

At the age of 10 days, the plasma concentrations of Na⁺, K⁺, Ca²⁺ and Mg²⁺ were similar for NCC wt and NCC ko mice (Table 1). For both groups of mice, the measured plasma K⁺ levels were rather high, likely reflecting some artificial hemolysis due to the difficult blood sampling in these young animals. At the age of 6 weeks, the plasma concentrations of Na⁺, K⁺ and Ca²⁺ were all in the normal range and similar between NCC wt and NCC ko mice. However, plasma Mg²⁺ levels were significantly lower in NCC ko versus NCC wt mice. The different plasma Ca²⁺ concentrations of 10-day- and 6-week-old mice are likely related to technical reasons (i.e. different methods used for blood collection and plasma Ca²⁺ measurements in 10-day-old mice versus 6-week-old mice).

Increased renal renin expression and hyperaldosteronism in NCC ko mice at age 10 days

At the age of 1 day, the messenger RNA (mRNA) expression levels for renin were significantly higher in the kidneys of NCC ko mice than in NCC wt mice (Figure 1A). Renin mRNA expression also tended to be higher in NCC ko versus NCC wt mice at the age of 4 days, 10 days and 6 weeks. However, the differences did not reach statistical significance (Figure 1A). Nevertheless, consistent with continuous activation of the renin–angiotensin–aldosterone system (RAAS), plasma aldosterone levels were found to be increased in NCC ko mice at the age of 10 days compared with the corresponding levels in NCC wt mice (Figure 1B). Due to the limited amounts of blood that could be collected from very young pups, plasma aldosterone levels could not be measured at the ages of 1 and 4 days.

FIGURE 1

Renin gene expression and plasma aldosterone. (A) Gene expression of renin in kidneys of NCC wt versus NCC ko mice at the ages of 1, 4 and 10 days and 6 weeks analyzed by RT-qPCR. Significant difference in NCC wt versus NCC ko at age 1 day (*P = 0.0002). Values are mean ± SEM. n = 3. (B) Plasma aldosterone (pg/mL) in NCC wt versus NCC ko mice at the age of 10 days after birth analyzed by ELISA with significant difference (*P = 0.0097). Values are mean ± SEM. n = 4.

DCT outgrowth in NCC wt mice occurs mainly during the first 10 postnatal days

Immunofluorescent staining for the DCT-specific NCC showed an increase of NCC expression in the renal cortex of NCC wt mice during the first 10 days of life (Figure 2A). Consistently, morphometric analysis revealed that the fractional cortical tubular volume of the DCTs becomes progressively higher until Day 10 after birth (Figure 2B). At the age of 10 days, the fractional cortical tubular volume for the DCT reaches ∼7%, which is similar to the value seen in adult NCC wt mice. In parallel with DCT outgrowth, NCC mRNA expression (Figure 2C) and NCC protein expression (Figure 2D) increased up to Day 10 and remained at the same level. Although not statistically significant, the abundance of NCC protein tended to increase further after Day 10, suggesting a posttranscriptional mechanism for the additional increase in NCC abundance from Day 10 to Week 6.

FIGURE 2

Fractional cortical volume of DCT and NCC mRNA and protein abundance. (A) NCC protein abundance in kidneys of NCC wt mice at different ages (1, 4 and 10 days and 6 weeks after birth). (B) Fractional cortical volume (%) of whole DCT in NCC wt mice. DCT were identified due to NCC immunostaining. Significant difference in NCC wt (Day 1) (*P = 0.002) versus NCC wt (6 weeks). Values are mean ± SEM. n = 3. (C) Gene expression of NCC in kidneys of NCC wt mice at different ages (1, 4 and 10 days and 6 weeks after birth) analyzed by RT-qPCR. Significant difference in NCC wt (Day 1) (*P = 0.0002) versus NCC wt (6 weeks) and NCC wt (Day 4) (*P = 0.002) versus NCC wt (6 weeks). Values are mean ± SEM. n = 3. (D) Protein expression of NCC in total kidney homogenates of NCC wt mice at different ages (1, 4 and 10 days and 6 weeks after birth) probed on western blot. Significant difference in NCC wt (Day 1) (*P = 0.0006) versus NCC wt (6 weeks) and NCC wt (Day 4) (*P = 0.0145) versus NCC wt (6 weeks). Values are mean ± SEM. n = 3.

Impaired postnatal development of DCT1 in NCC ko mice

For obvious reasons, we could not use the detection of NCC to identify DCTs in NCC ko mice. Therefore we relied, as in previous studies, on detection of the calcium- and magnesium-binding protein PV, which is a marker for early DCT (DCT1) [9]. Immunostaining of kidneys from 10-day-old WT mice confirmed the reliability of this approach for the developing mouse kidney. As in adult kidneys, PV expression in kidneys of 10-day-old pups started precisely at the transition from the NKCC2-positive thick ascending limb to the NKCC2-negative early DCT (DCT1) (Supplementary data, Figure S2A), which was also characterized by a very low abundance of calbindin (CB). PV abundances ceased abruptly at the transition from the weak CB-positive early DCT (DCT1) [17] to the strong CB-positive late DCT (DCT2). DCT1 and DCT2 are both NCC-positive (Supplementary data, Figure S2B). Interestingly, the morphometric quantification of the fractional cortical volume of PV-positive tubules (i.e. DCT1) versus NCC-positive tubules (i.e. DCT1 and DCT2) revealed that the ratio of DCT1 to DCT2 varies with age. As observed previously in rat kidney [18], the developing DCT (Days 1 and 4 after birth) is mainly composed by DCT2 cells, while at later stages of development (Day 10 and Week 6 after birth) DCT1 cells appear to dominate (Figure 3). The reason for this developmental switch from a DCT2 predominance to a DCT1 predominance is unclear.

FIGURE 3

DCT1:DCT2 ratio at different ages. Percentage of fractional cortical volume of DCT1 (based on PV immunostaining) and DCT2 with respect to entire DCT fractional cortical volume (based on NCC immunostaining) in NCC wt mice at different ages (1, 4 and 10 days and 6 weeks). Values are mean ± SEM. n = 3 (1 and 4 days), n = 4 (10 days) and n = 6 (6 weeks).

When we then compared kidneys from NCC wt and NCC ko mice, we found that the abundance of PV-positive tubules (DCT1s) in the renal cortex increases progressively with age in NCC WT mice. In contrast, PV-positive tubules are barely visible in kidneys of NCC ko mice at all analyzed time points (Figure 4A and B). Morphometric analysis confirmed this observation. In NCC wt mice, the fractional cortical volume of PV-positive tubules increases with age, suggesting a constant outgrowth of this segment. In contrast, after an initial minor outgrowth (until Day 4), the development of DCT1 appears to cease in the kidneys of NCC ko mice (Figure 4C). Consistent with impaired outgrowth of DCT1, PV gene expression tended to be lower in the kidneys of NCC ko mice than in NCC wt mice (Figure 4D), although differences reached statistical significance only at 6 weeks of age.

FIGURE 4

Fractional cortical volume of DCT1 and PV mRNA abundance. (A) PV protein expression in kidneys of NCC wt mice at different ages (1, 4 and 10 days and 6 weeks after birth). (B) PV protein expression in kidneys of NCC ko mice at different ages. (C) Fractional cortical volume of DCT1 (%) in NCC wt versus NCC ko mice at different stages of development (1, 4 and 10 days and 6 weeks after birth). DCT1 were identified due to PV immunostaining. Significant differences in NCC ko compared with age-corresponding NCC wt at age 1 day (*P = 0.005), 10 days (*P = 0.005) and 6 weeks (*P = 0.00004). Values are mean ± SEM. n = 3. (D) Gene expression of PV in NCC wt and NCC ko mice. Significant difference in NCC ko (6 weeks) (*P = 0.00004) versus NCC wt (6 weeks). Values are mean ± SEM. n = 3.

Reduced DCT cell proliferation in NCC ko mice at age 4 and 10 days

To test whether reduced cell proliferation may explain the impaired outgrowth of DCT1 in NCC ko mice, we used immunodetection of the PCNA, which is a reliable cell proliferation marker. As shown in Figure 5, the number of PCNA-positive cell nuclei is drastically lower in PV-positive DCT1 cells of NCC ko mice than in PV-positive DCT1 cells of NCC wt mice for both 4 and 10 days of age.

FIGURE 5

DCT1 cell proliferation in NCC wt and NCC ko mice at age 4 and 10 days evaluated by PCNA immunodetection. (A) Double immunostaining for PV to identify DCT1 segments (D1) and PCNA to identify proliferating cells in paraffin kidney sections of 4-day-old NCC wt and NCC ko mice. (B) Double immunostaining for PV and PCNA in 10-day-old NCC wt and NCC ko mice. (C) PCNA-positive DCT1 cell nuclei within kidney cross sections of 4- and 10-day-old NCC wt versus NCC ko mice expressed as a percentage of all DAPI-stained cell nuclei in the DCT1 segments. Significant differences in NCC ko (4 days) versus NCC wt (4 days) (*P = 0.0002) and NCC ko (10 days) versus NCC wt (10 days) (*P = 0.0001). Values are mean ± SEM. n = 3 (4 and 10 days).

Sporadic DCT cell apoptosis in NCC ko mice at age 4 and 10 days

Since the morphometric analysis suggested that the fractional volume of DCT1 decreases from Day 4 to Day 10 in NCC ko mice, we analyzed whether the frequency of apoptotic cell death in the DCTs of NCC ko mice might be increased at this age. For the detection of apoptosis, we used the TUNEL assay and combined it with immunostaining for PV to identify early DCT (DCT1). Apoptotic cell nuclei were detected very rarely in developing renal tubules and the interstitium of the nephrogenic zone in both groups of mice at the ages of 4 and 10 days (Supplementary data, Figure S3). Only one apoptotic DCT cell was detected in the analyzed NCC wt kidney samples (one paraffin section each from three wt mice at the age of 4 days and three wt mice at the age of 10 days) and five apoptotic DCT cells were found in NCC ko kidney samples (one paraffin section each from three ko mice at the age of 4 days and three ko mice at the age of 10 days).

Reduced protein expression of the DCT-specific Mg²⁺ channel TRPM6 in NCC ko mice

Western blot analysis revealed a significant reduction in the protein expression of the DCT-specific Mg²⁺ channel TRPM6 in NCC ko versus wt mice at 1 day, 4 days, 10 days and 6 weeks (Figure 6A and B; Supplementary data, Figures S4 and S5). In contrast, the abundance of the Ca²⁺ reabsorbing transient receptor potential vanilloid 5 (TRPV5) Ca²⁺ channel, which is expressed in DCT2 and CNT, was similar in NCC wt versus ko mice (Figure 6A and B).

FIGURE 6

TRPM6, TRPV5, α-ENaC, β-ENaC and γ-ENaC protein abundance. (A) Protein expression of TRPM6, TRPV5, α-ENaC, β-ENaC and γ-ENaC in total kidney homogenates of NCC wt and NCC ko at different ages (1, 4 and 10 days and 6 weeks after birth). (B) Difference in protein expression (%) of NCC ko at different ages (1, 4 and 10 days and 6 weeks) compared with average of NCC wt of age-corresponding group. Day 1: significant difference in NCC ko (1 day) versus NCC wt (1 day) for TRPM6 (*P = 0.022). Day 4: significant difference in NCC ko (4 days) versus NCC wt (4 days) for TRPM6 (*P = 0.0002), α-ENaC cleaved (*P = 0.012), β-ENaC (*P = 0.013) and γ-ENaC (*P = 0.0075). Day 10: Significant difference in NCC ko (10d) versus NCC wt (10d) for TRPM6 (*P = 0.0014), α-ENaC cleaved (*P = 0.023), β-ENaC (*P = 0.016), γ-ENaC (*P = 0.012) and γ-ENaC cleaved (*P = 0.008). 6 weeks: Significant difference for NCC ko (6w) versus NCC wt (6w) for TRPM6 (*P = 0.002) and γ-ENaC cleaved (*P = 0.000035). Statistical analysis was performed combining the densitometry results obtained from two independent western blots (A) and Supplementary data, Figures S4 and S5). Values are mean ± SEM. n = 3 (1d, 4d), n = 9 (10d, 6w).

Compensatory upregulation of ENaC in response to loss of NCC

Protein expression analysis for α-, β- and γ-ENaC revealed a significant upregulation of the cleaved forms of α- and/or γ-ENaC at 4 days, 10 days and 6 weeks, while β-ENaC protein abundance was downregulated at Days 4 and 10 (Figure 6A and B; Supplementary data, Figures S4 and S5).

DISCUSSION

The renal DCT has a remarkable structural plasticity. The DCT responds to an increased tubular workload with a pronounced epithelial hypertrophy, while a reduced transport activity in the DCT is associated with epithelial hypotrophy [19]. Previous studies on adult mice have shown that genetic deletion of the thiazide-sensitive NCC is associated with marked epithelial atrophy of the early DCT (DCT1). In the present study, we addressed the question whether NCC deficiency impairs outgrowth of the DCT during renal development or whether DCT atrophy occurs at later stages of life due to a regression of the already developed DCT.

In general, no grave phenotypic features such as growth retardation, severe electrolyte disturbances or metabolic abnormalities were observed in NCC-deficient mice, neither during the developmental period nor at adulthood. However, our study shows that NCC ko mice lack significant outgrowth of the DCT during early perinatal development. The disturbed DCT growth during renal development in NCC ko mice results in marked DCT atrophy at adulthood. We showed previously that the DCT atrophy concerns mainly the early DCT (DCT1) rather than the late DCT (DCT2) [9]. We did not address differential effects on DCT1 and DCT2 morphology in the present study. However, our observation that the DCT1-specific marker protein PV is much lower in kidneys of NCC ko mice than the DCT1 and DCT2 marker protein TRPM6 indicates that DCT1 is preferentially affected during renal development. The structural atrophy of the DCT is presumably accompanied by changes in DCT tubular function and may explain some of the clinical features of Gitelman syndrome. Signs of volume depletion with increased renal renin expression and hyperaldosteronism, common symptoms in human Gitelman patients, and previously also described in adult NCC ko mice [8, 12], were already present in mice at the age of 1 and 10 days, respectively. The very high upregulation of renal renin expression at Day 1 is surprising, but might reflect the more difficult adaptation of newborn NCC ko mice to postnatal life. Perhaps, directly after birth, the NCC ko mice lose a lot of sodium until activation of the RAAS system leads to a compensatory upregulation of ENaC and possibly other Na⁺-retaining mechanisms. Consistent with this idea, our immunoblot data did not show clear evidence for a significant up-regulation of ENaC in NCC ko mice at Day 1 after birth. However, at Day 4, Day 10 and Week 6, proteolytic cleavage of ENaC subunits was observed, which is indicative of ENaC activation [20]. Overall, our data strongly suggest that renal tubular dysfunction leading to renal sodium wasting and volume contraction manifests in NCC ko mice at a very early age.

Hypomagnesemia is another key characteristic of Gitelman syndrome [3, 4]. Interestingly, low plasma magnesium levels were only found in adult NCC ko mice, not in NCC ko 10-day-old pups, even though expression of the DCT1-specific Mg²⁺ channel TRPM6 was significantly decreased in NCC ko mice at both the age of 10 days and 6 weeks. Most probably, NCC ko pups are born with a normal body magnesium content. We hypothesize that magnesium wasting in NCC ko only starts postnatally and a prolonged time period is necessary to establish a relevant magnesium depletion. Additionally, the transition from breast-feeding nourishment to lab chow may also contribute to the development of hypomagnesemia. It would be interesting to know the situation in newborn human Gitelman patients. Given that most human patients do not become symptomatic during early childhood [5], one can speculate that plasma magnesium is also not, or at least not drastically, reduced at a neonate age, as demonstrated in this NCC ko mouse model.

Hypokalemia is another cardinal feature in human Gitelman patients [3, 5]. Interestingly, and consistent with previous studies on adult mice [8], the NCC ko mice did not show any hypokalemia at any of the analyzed ages. This difference between human Gitelman patients and NCC ko mice may reflect real species differences but could be also related to the exceedingly higher K⁺ intake of mice compared with humans. On the standard mouse chow with 0.8% K⁺, adult mice ingest ∼30–40 mg of potassium [21]. The average daily K⁺ intake of adult humans is in the range of 2.9 g for men and 2.3 g for women [22]. For a 25-g mouse and a 70-kg human, this equals a daily K⁺ intake of 1.2–1.6 g potassium/kg body weight in the mouse and 0.04–0.03 g/kg body weight in the human. Thus, normalized for body weight, mice have an ∼40 times higher daily K⁺ intake than humans. The idea that the high K⁺ intake on standard diet may protect NCC ko mice from hypokalemia is supported by the observation that the mice become rapidly hypokalemic when K⁺ intake is restricted [12].

Gitelman syndrome patients and NCC ko mice are hypocalciuric [3, 4, 8]. Likewise, thiazide diuretics reduce urinary Ca²⁺ excretion [23]. Although NCC inhibition with thiazides increases calcium uptake by DCT cells [24, 25], micropuncture and biochemical studies on NCC ko mice and thiazide-treated mice indicated that the reduced renal calcium excretion in these mice is due to an enhanced passive paracellular calcium reabsorption in the proximal tubule rather than due to an increased transcellular calcium transport in the distal tubule [9, 23].

The loss of NCC leads to major structural remodelling of the DCT, which includes a reduction of the individual cell height and the tubular length of the DCT [19]. Our present study suggests that the drastic shortening of the DCT in adult NCC-deficient mice is most probably the consequence of incomplete outgrowth of DCT1 during early perinatal kidney development, indicated by the significantly lower rate of PCNA-detected DCT1 cell proliferation in NCC ko mice compared with NCC wt mice for both age 4 and 10 days. Apparently a few PV-positive tubuli are formed in both NCC wt and NCC ko mice at the age of 4 and 10 days. However, this minor outgrowth of the DCT ceases in the kidneys of NCC ko mice. Even though apoptosis was more frequently observed in NCC ko mice, apoptotic events occurred only rarely in both genotypes and its significance remains unclear. Finally, early DCT is almost completely absent from the kidneys of 6-week-old NCC ko mice. This marked structural atrophy diminishes the available epithelial cell surface area along the DCT and may hence explain the reduced TRPM6 abundance and renal Mg²⁺ wasting in NCC ko mice reported in a previous study [8] and in this study. Likewise, pharmacological inhibition of NCC by prolonged treatment with thiazides (e.g. hydrochlorothiazide) or thiazide-like diuretics (e.g. metolazone) often causes hypomagnesemia in humans [26, 27] and mice [19, 23], which is also thought to be related to a downregulation of TRPM6 at the mRNA and protein levels. We do not know about the effect of the genetic loss or pharmacological inhibition of NCC on the structural integrity of the DCT in humans. However, the observation of an almost complete loss of the early DCT in NCC-deficient mice [9] and of massive apoptotic death in the early DCT of thiazide-treated rats [13] suggests that the loss of epithelial surface area in the DCT may explain the hypomagnesemia observed in Gitelman patients and humans treated with thiazide diuretics.

In summary, the genetic loss of NCC leads to remarkable structural and functional changes in mouse kidneys. Our data indicate that NCC deficiency mainly impairs outgrowth of the developing DCT rather than causes a regression of already formed DCTs. Consistently an activation of the RAAS with a compensatory upregulation of ENaC and a downregulation of TRPM6 is detectable quite early in life. Together with the observation in humans that the clinical manifestation and onset of Gitelman syndrome varies considerably [28–30], one can speculate that signs of Gitelman syndrome are already present at a very young age but are not diagnosed, as laboratory testing for plasma magnesium and potassium as well as urinary calcium and magnesium excretion are not routinely performed in children.

ACKNOWLEDGEMENTS

The authors thank Dominique Loffing-Cueni, Dario Mattle and Eszter Banki for the initial characterization of the TRPM6 antibody. The TRPV5 anti-body was a kind gift of Olivier Bonny (University of Lausanne). The NCC ko mice were a kind gift of Patrick J. Schultheis (Northern Kentucky University) and Gary E. Shull (University of Cincinnati). The technical assistance of Michèle Heidemeyer is kindly acknowledged.

FUNDING

C.S. is a member of the University of Zurich MD/PhD programme. J.L. is supported by research funds from the Swiss National Centre for Competence in Research ‘Kidney.CH’ and by a project grant from the Swiss National Science Foundation (310030_173276/1).

CONFLICT OF INTEREST STATEMENT

The results presented in this article have not been published previously in whole or part, except in abstract format. J.L. reports grants from the Swiss National Science Foundation and grants from the National Centre of Competence in Research ‘Kidney.CH’ during the conduct of the study. Moreover, his group receives royalties for outsourced antibodies from Millipore and Abcam outside the submitted work.