Reduced Expression of Magnesium Transport Proteins in the Distal Convoluted Tubule of Clcnkb‐Deficient Mice May Explain Urinary Magnesium Wasting in Classical Bartter Syndrome

ABSTRACT

Aim

The kidney thick ascending limb (TAL) plays a key role in the transport of sodium, chloride, potassium, calcium, and magnesium. Bartter syndrome is a hypokalemic, salt‐losing tubulopathy caused by impaired TAL function. Pathogenic variants in SLC12A1 or KCNJ1 cause antenatal Bartter syndrome, and variants in CLCNKB result in classical Bartter syndrome. Although all variants impair TAL electrolyte transport, their effects on mineral handling differ. In contrast to antenatal forms, classical Bartter syndrome is frequently associated with hypomagnesemia, a feature also found in Gitelman syndrome resulting from pathogenic variants in SLC12A3, expressed in the distal convoluted tubule (DCT). The mechanisms underlying these distinct clinical phenotypes are not understood.

Methods

Clcnkb‐ and Slc12a3‐deficient mice as well as human kidneys were investigated. The abundance and localization of electrolyte and mineral transporters and stereological parameters were assessed by immunohistochemistry. Gene and protein expression was determined in Clcnkb‐deficient mice.

Results

In human kidney, both ClC‐K channels and their essential subunit Barttin were found in TAL, DCT, and collecting system. Clcnkb‐deficient mice showed complete ablation of Clcnkb and reduced DCT and connecting tubule volumes. Furthermore, DCT magnesium channels and select basolateral magnesium transporters were markedly reduced in Clcnkb‐deficient mice, similar to Slc12a3‐deficient mice. In contrast, the overall abundance of calcium transport proteins was largely preserved.

Conclusion

ClC‐K channels are expressed in the TAL and DCT in mouse and human kidneys. Clcnkb ablation reduces TRPM magnesium channel expression and DCT tubule volume, providing a likely explanation for the frequently observed hypomagnesemia in classical Bartter syndrome.

Practitioner Points

• Patients with classical Bartter syndrome caused by pathogenic CLCNKB variants often develop hypomagnesemia, a feature not seen in the antenatal forms caused by SLC12A1 or KCNJ1 mutations.

• In both human and mouse kidney, ClC‐K channels and their subunit Barttin are present, not only in the TAL but also throughout the distal nephron, enabling a direct effect of pathogenic variants in CLCNKB to impair electrolyte handling beyond the TAL.

• Clcnkb‐deficient mice display reduced magnesium transporter expression in the distal convoluted tubule, providing a mechanistic explanation for the frequent hypomagnesemia observed in classical Bartter syndrome.

1. Introduction

Bartter syndrome is caused by pathogenic variants in genes that are necessary to maintain electrolyte transport across the thick ascending limb (TAL) [1, 2, 3, 4, 5, 6]. The syndrome is characterized by excessive sodium chloride wasting into urine, which causes hypotension, hypokalemia, and metabolic alkalosis [7, 8, 9]. In Bartter syndrome, divalent cation handling is also affected. However, the extent to which calcium and magnesium homeostasis is perturbed in Bartter syndrome patients varies depending on the underlying genetic defect [10]. The reason for these differences has not been investigated in detail.

The TAL is accountable for reabsorbing around 25% of calcium and up to 60% of magnesium filtered by the glomerulus [11, 12, 13]. Paracellular reabsorption of calcium and magnesium in TAL relies on the transepithelial voltage gradient and hence intact TAL transport function [14, 15]. The transport of sodium chloride to the basolateral side, coupled with the recycling of potassium on the apical side, forms the transepithelial voltage gradient, which is critical for calcium and magnesium reabsorption across the segment.

The furosemide‐sensitive cotransporter NKCC2, encoded by the SLC12A1 gene, cotransports sodium, potassium, and two chloride ions into the TAL cell [4, 16, 17]. The renal outer medullary potassium channel (ROMK), encoded by the KCNJ1 gene, is also situated in the apical membrane. Here, it facilitates the recycling of potassium back into the lumen to provide sufficient potassium to fuel NKCC2‐dependent transport [5, 16, 18, 19]. In contrast, the ClC‐Kb channel encoded by the CLCNKB gene and its essential Barttin subunit encoded by the BSND gene enable the exit of chloride ions across the basolateral membrane [6, 20, 21]. Pathogenic variants in these genes lead to Bartter syndrome as they strongly impair electrolyte transport across the TAL. Furthermore, as this collapses the lumen‐positive voltage, divalent cation reabsorption via the paracellular shunt is also impaired [1, 2, 3, 4, 5, 6].

The antenatal forms of Bartter's syndrome result from pathogenic variants in SLC12A1 and KCNJ1. These antenatal forms cause severe hypercalciuria and nephrocalcinosis. In contrast, alterations in magnesium balance, such as hypomagnesemia and hypermagnesuria, are not a common feature [4, 5, 9, 22]. The classical form of Bartter syndrome resulting from pathogenic variants in the CLCNKB gene is more variable and frequently less severe than the antenatal forms with respect to electrolyte balance and impairment of TAL transport. In line with these findings is the observation that hypercalciuria is less common in classical Bartter syndrome and typically less pronounced [2, 3, 10, 23]. However, hypomagnesemia is far more common in classical Bartter syndrome. This difference cannot easily be explained by impaired TAL function, because hypomagnesemia is generally not seen in the antenatal form [2, 3, 10, 23]. Classical Bartter syndrome has been studied in mice with targeted deletion of Clcnkb encoding the ClC‐K2 protein (equivalent to the CLCNKB gene encoding the ClC‐Kb protein in humans). Clcnkb‐deficient mice display a Bartter‐like phenotype with marked renal sodium chloride and potassium losses, hypotension, and hypokalemic metabolic alkalosis [21, 24]. In line with the classic Bartter phenotype, Clcnkb‐deficient mice display urinary magnesium wasting [24, 25]. Depending on the Clcnkb‐deficient mouse line, mice are either normo‐ or hypercalciuric [24, 25]. This is consistent with the phenotypical spectrum associated with the classical Bartter phenotype.

ClC‐Kb is expressed throughout the TAL, but also in the distal convoluted tubule (DCT), connecting tubule (CNT), and collecting duct [21, 24, 25]. When challenged, Clcnkb‐deficient mice also show a reduced response to thiazide diuretics, in line with reduced expression of the sodium chloride cotransporter NCC, encoded by the SLC12A3 gene, in the DCT [21, 24, 25] and display abnormalities in the morphology of the DCT, which could impair renal magnesium handling [24]. Gitelman syndrome is caused by pathogenic variants in the SLC12A3 gene encoding the thiazide‐sensitive sodium chloride cotransporter in the DCT [26]. A hallmark feature of Gitelman syndrome is hypomagnesemia [26]. Mice with targeted deletion of the Slc12a3 cotransporter display hypomagnesemia and atrophy of the early DCT [27]. This study, therefore, aimed to investigate the molecular mechanisms underlying disturbed mineral balance in classical Bartter syndrome by detailed evaluation of Clcnkb‐deficient and Slc12a3‐deficient mouse lines with particular focus on the magnesium and calcium transport machinery in the distal nephron.

2. Results

2.1. Human ClC‐Kb and Its Mouse Homolog ClC‐K2 Are Expressed Along the Distal Nephron

To understand the distribution of ClC‐Kb in human kidney, we localized the protein along the length of the nephron. Antibodies targeting ClC‐Kb also react with the highly homologous ClC‐Ka channel, which is expressed in parts of the nephron, especially the medullary TAL and thin ascending limb [24]. The antibodies used to localize ClC‐K were first validated by localizing the mouse ClC‐K homolog in WT mice and Clcnkb‐deficient mice. Clcnkb mRNA expression was completely abolished in the Clcnkb‐deficient mice (Figure 1A). In WT mouse kidney, the antibody recognized epitopes in the medullary and cortical TAL, distal nephron and collecting system (Figure 1B–F), as evidenced previously [24]. In Clcnkb‐deficient animals, staining was largely reduced in the outer stripe of the outer medulla (OSOM) and cortex (CTX). Co‐staining with the furosemide‐sensitive cotransporter NKCC2, a marker of the TAL, and the thiazide‐sensitive cotransporter NCC, a marker of the DCT, revealed colocalization with ClC‐K along the length of the TAL and staining in the DCT, which was absent in the Clcnkb‐deficient mice (Figure 1G–J), in line with ClC‐K2 expression at these sites. ClC‐K expression was maintained in the medullary thin limbs of Clcnkb‐deficient mice, likely due to expression of ClC‐K1 (homolog of human ClC‐Ka) (Figure S1A,B). Co‐staining with AQP2 revealed expression in the collecting system, staining that remained in the Clcnkb‐deficient mice (Figure 1K,L), indicating that this signal is not ClC‐K2.

FIGURE 1.

Localization of ClC‐Kb in mouse kidney. (A) Complete loss of Clcnkb gene expression in Clcnkb‐deficient mice confirmed by qPCR. ∗p < 0.05 compared with WT. (B–F) Immunostaining for ClC‐K in WT mouse kidney revealed expression in thin limb, medullary and cortical TAL, DCT, and collecting system. In Clcnkb‐deficient mice, staining in the cortex and outer medulla was reduced. (G–L) Co‐immunostaining revealed ClC‐K colocalization with NCC in the DCT (G, H), NKCC2 in the TAL (I, J), and AQP2 in the collecting system (K, L). In Clcnkb‐deficient mice, ClC‐K immunoreactivity was absent in the distal nephron, but not in collecting ducts. The scale bar represents 50 μm (B) or 200 μm (C–L). *p < 0.05 is considered statistically signigficant.

In human kidney, a similar pattern of ClC‐K distribution was evident, with strong immunoreactivity in the TAL and distal nephron, as well as thin limbs in the inner medulla (IM) (Figure 2A–D). Co‐staining with NKCC2 revealed colocalization with ClC‐K along the whole length of the TAL (Figure 2F–H). Furthermore, ClC‐K colocalized with the DCT marker NCC (Figure 2E), and weakly with AQP2 in the collecting system, predominantly in the ISOM and IM, while mainly staining AQP2‐negative cells in the CTX and OSOM (Figure 2I–L) in line with previous findings showing expression in intercalated cells [28, 29].

FIGURE 2.

Localization of ClC‐Kb in human kidney. (A–D) In human kidney, ClC‐K showed strong expression in TAL and distal nephron, and weaker expression in some cells in the collecting duct (arrows). (E–L) Co‐immunostaining of human kidney revealed ClC‐K colocalization with NKCC2 in the TAL (F–H), NCC in the DCT (E), and AQP2 in the collecting system (I–K), and AQP2‐negative tubules in the inner medulla (IM) (L). The scale bar represents 200 μm.

ClC‐K function is dependent on the interaction with the essential Barttin subunit encoded by the BSND gene, required for ClC‐K localization to the plasma membrane and its subsequent activity [6]. In mouse WT kidney, Barttin was expressed along the ascending loop of Henle, the distal nephron, and in intercalated cells in the collecting duct (Figure S2A–E). Barttin expression appeared strongly reduced in Clcnkb‐deficient animals in the TAL in the OSOM and CTX, DCT and collecting duct (Figure S2B,C,G–L), while expression was mildly reduced in the inner stripe of the outer medulla (ISOM) TAL and preserved in AQP2‐negative tubules, likely thin limbs, in the inner medulla (Figures S2D,E and S3A,B). Bsnd mRNA expression was not reduced in the Clcnkb‐deficient mice (Figure S2F), suggesting the reduction of Barttin expression seen is an effect of Barttin protein stability due to the absence of ClC‐K2. In the human kidney, expression of Barttin was found along the length of the TAL (Figure S4A–C), which was confirmed by colocalization with NKCC2 in cortex and medulla (Figure S4F–H). Furthermore, colocalization with NCC (Figure S4E) and AQP2 (Figure S4I–L) revealed expression of Barttin in the DCT and in AQP2‐negative intercalated cells in the connecting tubule and collecting duct in the cortex and outer medulla. Weak expression was detected in the inner medullary thin limbs (Figure S4L).

2.2. Clcnkb‐Deficient Mice Display Marked Reduction of DCT Marker Proteins and Loss of Magnesium Channel Expression

The Clcnkb‐deficient mice investigated here have an increased urinary magnesium excretion but show no changes in urinary calcium excretion [24]. To investigate whether alterations in distal tubular transport could occur, expression of transporters in the DCT was analyzed by qPCR. Clcnkb‐deficient mice showed a marked reduction in the tubular expression of Slc12a3 and an even stronger reduction in the expression of parvalbumin, a marker of the early DCT encoded by the Pvalb gene (Figure 3A,B). Furthermore, expression of Trpm6 and Trpm7 was markedly reduced in Clcnkb‐deficient mice in comparison to WT (Figure 3C,D), reflecting disrupted magnesium transport pathways in the DCT. Epidermal growth factor (EGF) is a marker that has been associated both with magnesium wasting and distal tubular mass [30, 31, 32, 33, 34]. Expression of Egf was markedly reduced in Clcnkb‐deficient mice, while epidermal growth factor receptor (Egfr) expression did not change significantly (Figure 3E,F). Expression of Slc41a3 isoforms was measured because recent evidence has shown that Slc41a3‐deficient mice develop hypomagnesemia, supporting a physiological role of Slc41a3 in magnesium homeostasis [35]. In particular, a DCT‐specific isoform (Slc41a3‐Isoform 2) may contribute to basolateral magnesium extrusion. Unlike the more widely expressed Isoform 1, Isoform 2 is enriched in the DCT through the use of an alternative promoter. Moreover, isoform 2 has been shown to facilitate magnesium transport in vitro [36]. The expression of the Slc41a3 isoform 2 was significantly decreased in the Clcnkb‐deficient mice, while expression of isoform 1 was significantly increased (Figure 3G,H). Expression of the Fxyd2, the γ‐subunit of the Na⁺/K⁺‐ATPase, expressed at high levels in the TAL and DCT, was significantly decreased in the Clcnkb‐deficient mice (Figure 3I), while expression of Cnnm2, encoding the magnesium‐transport regulatory protein showed no significant difference between genotypes (Figure 3J).

FIGURE 3.

Reduction of DCT markers in Clcnkb‐deficient mice. (A, B) Gene expression quantified by qPCR of Slc12a3 and the DCT1 marker Pvalb was markedly reduced in Clcnkb‐deficient mice compared to WT controls. (C–F) Expression of magnesium transporters Trpm6 (C) and Trpm7 (D) was significantly decreased in Clcnkb‐deficient mice. Egf expression was also reduced (E), while Egfr expression remained unchanged (F). Expression of the Slc41a3 isoform 1 was increased, while expression of isoform 2 was decreased in Clcnkb‐deficient mice compared to WT controls, similarly to Fxyd2 (G–I). Expression of Cnnm2 was unchanged (J). (K–N) Immunohistochemistry for NCC (K–L) and parvalbumin (PV) (M, N) showed a strong reduction in Clcnkb‐deficient mice, consistent with DCT atrophy. The arrow points out a dilated tubule in the Clcnkb‐deficient mice. The scale bar represents 50 μm (K, M) or 200 μm (L, N). (O–Q) Kidney lysates were subjected to immunoblotting for NCC and pro‐EGF and signals were densitometrically quantified. Molecular masses (in kDa) are indicated on the left. Clcnkb‐deficient mice showed a decrease in the total renal expression of NCC and pro‐EGF. *p < 0.05 is considered statistically significant.

In agreement with the above qPCR data, protein expression of NCC, encoded by Slc12a3, and parvalbumin showed marked reductions in Clcnkb‐deficient mice by immunohistochemistry (Figure 3K–N). Furthermore, some of the remaining NCC‐positive tubules were dilated in the Clcnkb‐deficient mice. Expression of DCT proteins was further investigated by immunoblotting. Consistent with the immunohistochemical findings, total renal NCC expression as well as the expression of Pro‐EGF, the uncleaved EGF precursor, was reduced in the Clcnkb‐deficient mice (Figure 3O–Q). To evaluate expression of TRPM6 in kidney, a mouse monoclonal antibody was raised and its specificity was verified by two independent lines of experiments. First, immunohistochemical staining for TRPM6 on paraffin‐embedded mouse kidney tissue showed that staining was absent after preincubation with the immunizing peptide (Figure S5A,B). Second, the staining was absent in kidney‐specific Trpm6‐deficient mice in comparison to WT (Figure 4A). Fluorescence localization of TRPM6 was conducted with AQP2, Calbindin‐D28k, and NCC (Figure 4B–D). No apparent co‐expression of TRPM6 and AQP2 was observed, while Calbindin‐D28k was partly co‐expressed with TRPM6, and both localized to other distinct regions. NCC and TRPM6 were completely colocalized in the apical membrane of the DCT, confirming localization of TRPM6 in the DCT. In Clcnkb‐deficient mice, TRPM6 expression per cell was markedly reduced compared to WT mice (Figure 4E,F). A TRPM6 antibody clone was validated for Western blotting in Trpm6‐deficient mice, showing a specific band around 240 kDa only in WT mice (Figure 4G). TRPM6 protein expression was significantly reduced in the Clcnkb‐deficient mice (Figure 4H,I).

FIGURE 4.

Decreased expression of magnesium channel TRPM6 in Clcnkb‐deficient mice. (A) Validation of a monoclonal anti‐TRPM6 antibody demonstrated loss of staining in kidney‐specific Trpm6 knockout mice. (B–D) Fluorescence colocalization of TRPM6 was performed with AQP2, Calbindin‐D28k (CaBP‐28K), and NCC. No colocalization of TRPM6 and AQP2 was observed, whereas NCC and TRPM6 were fully colocalized at the apical membrane of the DCT. Calbindin‐D28k partly colocalized with TRPM6, with both proteins showing additional expression. (E, F) Immunostaining for TRPM6 in WT and Clcnkb‐deficient kidneys confirmed a marked reduction of TRPM6 protein in the DCT of knockout mice. The scale bar represents 200 μm. (G) Validation of monoclonal anti‐TRPM6 antibody demonstrated absence of a band around 240 kDa in Trpm6‐deficient mice. (H, I) Membrane fractions were subjected to immunoblotting for TRPM6, and signals were quantified by densitometry. Molecular masses (in kDa) are indicated on the left. Clcnkb‐deficient mice showed a decrease in the total renal expression of TRPM6. ∗p < 0.05 compared with WT.

2.3. Expression of Distal Nephron Calcium Transporters Is Relatively Preserved in Clcnkb‐Deficient Mice

Clcnkb‐deficient mice showed a change in expression of genes related to transcellular calcium transport, which are mainly expressed in the late DCT (DCT2) and CNT. Here, the mRNA expression of the rate‐limiting apical calcium channel Trpv5 as well as Trpv6 (which is expressed at low levels in kidney but may contribute to calcium transport) was increased (Figure 5A,B). The expression of the Calb1 gene, encoding the intracellular buffer Calbindin‐D28k was not changed. The sodium/calcium exchanger (NCX1, encoded by the Slc8a1 gene) and PMCA4 (Atp2b4) are the main basolateral transporters for calcium extrusion localized to the distal nephron [37, 38]. Slc8a1 expression was not changed, while Atp2b4 expression was significantly increased (Figure 5D,E). Expression of Klotho was significantly decreased, while expression of Egr1, a marker of FGF23 signaling, was significantly increased (Figure 5F,G). Expression of TRPV5 and Calbindin‐D28k was evaluated by immunohistochemistry. Expression of TRPV5 per tubular surface area seemed slightly decreased in the knockout mice, while Calbindin‐D28k expression was more markedly reduced (Figure 5H–K), potentially due to tubular dilation, while overall levels appeared maintained between WT and Clcnkb‐deficient mice. Expression of calcium transport proteins was further investigated by immunoblotting. Expression of both basolateral transporters, NCX1 and PMCA4, was not changed in the Clcnkb‐deficient mice. Although Clcnkb‐deficient mice showed a trend toward a decrease in Calbindin‐D28k expression, this was not significant (Figure 5L–O).

FIGURE 5.

Expression of distal nephron calcium transporters in WT and Clcnkb‐deficient mice. (A–E) Gene expression analysis of the apical calcium channel Trpv5 (A), Trpv6 (B), intracellular calcium buffer Calb1 (C), and basolateral calcium extrusion transporters Slc8a1 (NCX1, D) and Atp2b4 (PMCA4, E). Slc8a1 showed no significant changes, while Trpv5, Trpv6, and Atp2b4 were significantly upregulated, and Calb1 showed a trend toward reduction in Clcnkb‐deficient mice. (F, G) Expression of the Klotho gene (Kl) was significantly decreased in the knockout mice, while Egr1 was significantly upregulated. (H–K) Immunofluorescence staining of TRPV5 and Calbindin‐D28k (CaBP‐28K) in distal nephron tubules showed that expression of TRPV5 relative to tubular surface area was slightly decreased, whereas Calbindin‐D28k expression was more markedly reduced. The arrows point out dilated tubules in the Clcnkb‐deficient mice. The scale bar represents 50 μm (H, J) or 200 μm (I, K). (L–O) Kidney lysates were subjected to immunoblotting for NCX1, PMCA4 and CaBP‐28K, and signals were densitometrically quantified. Molecular masses (in kDa) are indicated on the left. No significant differences between wildtype and Clcnkb‐deficient mice were detected. *p < 0.05 is considered statistically significant.

2.4. Clcnkb‐Deficient Mice Display Dilated Tubules and a Decreased Cortical Volume of the Distal Convolution

Morphometric point counting‐based stereological analysis was used to determine the relative volume of the DCT1, DCT2, and CNT segments of the distal convolution. NCC was used to identify the DCT; TRPV5 was used to identify the DCT2 and CNT, where an overlap between NCC and TRPV5 staining was used to identify the DCT2 segment. This demonstrated substantial remodeling of distal nephron segments in Clcnkb‐deficient kidneys compared with WT controls. All analyzed tubular segments, DCT1, DCT2, and CNT, exhibited markedly reduced cortical representation in animals lacking Clcnkb, indicating an overall loss of distal tubular mass within the cortical reference space (Figure 6A–C). DCT1 and DCT2 segments displayed a similar degree of reduction, with cortical volume fractions decreased by approximately 40% in Clcnkb‐deficient kidneys compared with WT controls. CNT segments exhibited a more pronounced reduction of approximately 60%, although this was not significantly different from the reduction in the DCT.

FIGURE 6.

Tubular dilation and decreased volume of the distal convolution in Clcnkb‐deficient mice. (A–C) Morphometric point counting‐based stereological analysis was used to analyze the relative volume of the different segments of the distal convolution. NCC was used to identify the DCT, and TRPV5 was used to label the DCT2/CNT, where an overlap between NCC and TRPV5 staining was used to define the DCT2. Representative images are shown in A and B. NCC is labeled in green, TRPV5 in magenta. Arrows indicate DCT1 (green), DCT2 (white, due to colocalization) and CNT (magenta). Scale bar represents 50 μm. Quantification is shown in C, demonstrating a lower cortical volume fraction in Clcnkb‐deficient mice compared to WT mice in all analyzed segments. (D) Kidneys were stained with hematoxylin and eosin for qualitative assessment of renal morphology, showing dilated tubules in all Clcnkb‐deficient mice. Representative images are shown. From left to right, scale bars represent 500, 200, and 100 μm. (E) Segment‐specific analysis of tubular dilation demonstrated the presence of dilated/cystic tubules within all analyzed distal nephron segments in Clcnkb KO kidneys, indicating that cystic remodeling in Clcnkb KO kidneys occurs in the whole distal convolution. *p < 0.05 is considered statistically significant.

Kidneys were stained with hematoxylin and eosin for qualitative assessment of renal morphology, showing dilated tubules in all Clcnkb‐deficient mice, which were absent in WT mice (Figure 6D). Clcnkb‐deficient mice showed a varying degree of dilation of the renal pelvis and an increase in interstitial mass compared to WT animals. Segment‐specific analysis of tubular dilation demonstrated the presence of dilated/cystic tubular profiles within all analyzed distal nephron segments in Clcnkb knockout kidneys (Figure 6E). Numerically, DCT2 segments exhibited the highest number of enlarged/cystic profiles, although differences between the segments were not statistically significant. In contrast to the results in the Clcnkb‐deficient mice, only a single enlarged tubular profile was observed across all WT animals, likely reflecting a longitudinally sectioned tubule rather than true cystic dilation. These findings indicate that cystic remodeling in Clcnkb knockout kidneys occurs in the whole distal convolution.

2.5. No Change in Expression of Transporters Involved in Calcium and Magnesium Transport in the Proximal Tubule

To investigate whether alterations in proximal tubular transport could occur, we determined the expression of the Slc9a3 gene encoding the sodium/proton exchanger NHE3, which is the main transporter involved in creating the driving force for paracellular calcium and magnesium reabsorption [39]. No significant difference was found between WT and Clcnkb‐deficient mice (Figure 7A). In general, magnesium is reabsorbed in the proximal tubule via the paracellular shunt, but to a much lower extent than calcium [40], due to the lower permeability of the proximal tubular Claudin 2 (CLDN2) for magnesium over calcium [41]. We determined the expression of the Cldn2 gene, as well as the vitamin D‐metabolizing enzymes Cyp27b1 and Cyp24a1, also localized to the proximal tubule. Here, no significant differences were observed between WT and Clcnkb‐deficient mice (Figure 7B–D).

FIGURE 7.

Expression of transporters involved in proximal tubule calcium and magnesium handling in WT and Clcnkb‐deficient mice. (A–D) Gene expression quantified by qPCR of the sodium/hydrogen exchanger Slc9a3 (A), the tight junction protein Cldn2 (B), Cyp24a1 (C), and Cyp27b1 (D) in kidneys from Clcnkb‐deficient mice normalized to WT. No significant differences were observed between genotypes, indicating unaltered proximal tubular transporter expression.

2.6. Alterations in Calcium and Magnesium Transporting CLDNs in the TAL of Clcnkb‐Deficient Mice

The expression of the Slc12a1 gene encoding the furosemide‐sensitive transporter NKCC2, expressed in the TAL, was not altered between the groups (Figure 8A). Investigation of gene expression of claudins in the TAL revealed no change in Cldn10b, which primarily mediates paracellular monovalent cations permeability in the TAL (Figure 8B). The mRNA expression of the divalent cation‐permeable Cldn16 showed a non‐significant trend for an increase (Figure 8C), while no changes were observed in the expression of its binding partner Cldn19 (Figure 8D). Protein expression of NKCC2, CLDN16, and CLDN10 by immunohistochemistry did not reveal any marked differences (Figure 8E–J).

FIGURE 8.

Expression of calcium‐ and magnesium‐transporting claudins in the TAL is preserved in Clcnkb‐deficient mice. (A) Slc12a1 expression encoding the furosemide‐sensitive transporter (SLC12A1/NKCC2) quantified by qPCR was unchanged between WT and knockout mice. (B–D) Analysis of expression of claudin genes in the TAL showed no difference in Cldn10b (B), Cldn16 (C), or Cldn19 (D). (E–J) Immunohistochemistry of NKCC2 (E–H), CLDN10 (I) and CLDN16 (J) in TAL segments revealed no major changes in staining patterns in Clcnkb‐deficient mice. The scale bar represents 50 μm (E, I) or 200 μm (F–H, J).

2.7. Expression of the Distal Nephron Magnesium Transporter TRPM6 Is Markedly Reduced in Slc12a3‐Deficient Mice

Hypomagnesemia is a defining feature of Gitelman syndrome due to pathogenic variants in the SLC12A3 gene encoding the thiazide‐sensitive cotransporter. Hypomagnesemic Slc12a3‐deficient mice exhibit atrophy of the early DCT segment and magnesium wasting [27, 42]. Slc12a3‐deficient mice showed markedly reduced parvalbumin expression (Figure 9A,B). However, Slc12a3‐deficient mice did not show a dilation of TRPV5‐expressing tubules, and no difference in protein expression of TRPV5 or Calbindin‐D28k was observed by immunohistochemistry (Figure 9C–F). Since the phenotype with respect to magnesium loss is similar between Clcnkb‐ and Slc12a3‐deficient mice, we investigated the expression of TRPM6 in these animals as well. Here, we find alterations in TRPM6 expression as observed in Clcnkb‐deficient mice (Figure 9G,H).

FIGURE 9.

Reduced TRPM6 expression in Slc12a3‐deficient mice. (A, B) Immunohistochemistry for parvalbumin (PV) revealed a marked reduction in Slc12a3‐deficient mice, consistent with early DCT atrophy. (C–F) No differences were observed in protein expression of TRPV5 or Calbindin‐D28k (CaBP‐28K) between Slc12a3‐deficient and WT mice. (G, H) TRPM6 expression was markedly reduced in Slc12a3‐deficient mice, similar to the reduction seen in Clcnkb‐deficient animals, supporting the link between DCT atrophy and impaired magnesium transport. The scale bar represents 50 μm (A, C, E, G) or 200 μm (B, D, F, H).

3. Discussion

In this study, we aimed to establish the role of ClC‐Kb in tubular mineral handling, in particular transport of magnesium and calcium. This is largely due to alterations in the distal convolution, a key site for both calcium and magnesium handling, where these two transport pathways exhibit partially distinct segmental distributions. Transcellular magnesium reabsorption occurs in the DCT1 and DCT2 through the apical magnesium channel TRPM6 [43, 44, 45]. In contrast, transcellular calcium reabsorption is largely mediated by the apical calcium channel TRPV5 expressed in the late distal nephron, particularly in DCT2 and CNT [46, 47, 48]. We show that loss of Clcnkb results in reduced expression of genes important for magnesium transport in the distal convolution, while calcium transport pathways are largely preserved.

This is supported by the following observations: (i) ClC‐K expression in TAL and DCT cells was validated using kidney tissue of Clcnkb‐deficient mice. Similar CLC‐K immunoreactivity was found in the TAL and DCT of the human kidney. (ii) Expression of the essential ClC‐Kb subunit Barttin (BSND) paralleled ClC‐Kb distribution in both mouse and human kidney. The similar distribution of both Barttin and ClC‐K in humans and mice reflects a similar organization in the human kidney, corroborating the translational relevance of the Clcnkb‐deficient mouse model. Barttin expression was reduced in Clcnkb‐deficient mice. (iii) DCT expression of Trpm6, Trpm7, and isoform 2 of Slc41a3 transcripts and TRPM6 protein levels were markedly reduced in Clcnkb‐deficient mice, consistent with impaired magnesium reabsorption observed in these animals [24]. We found no decrease in the magnesium‐transport regulating protein Cnnm2, which is also expressed in the TAL [49]. (iv) Clcnkb‐deficient mice displayed a reduced expression of DCT markers, such as NCC (SLC12A3) and parvalbumin on mRNA and protein level. Furthermore, Egf expression, which has been associated with distal tubular mass and magnesium homeostasis, was strongly reduced in Clcnkb‐deficient mice. (v) In addition, stereological analysis revealed decreased relative volumes of main segments in the distal convolution, including DCT1, DCT2, and CNT in Clcnkb‐deficient mice. (vi) In contrast, expression of distal nephron calcium transporters, mainly expressed in the DCT2 and CNT (Trpv5, Trpv6, Calb1/CaBP‐28K, Slc8a1/NCX1, Atp2b4/PMCA4), was increased or maintained at RNA and protein level, with only a reduction in Klotho gene expression. (vii) No significant changes were observed in proximal tubular transporters or claudins involved in proximal tubular and TAL Ca²⁺ and Mg²⁺ handling. (viii) Finally, Slc12a3‐deficient mice, a model of Gitelman syndrome with known hypomagnesemia from Mg²⁺ wasting, also displayed reduced TRPM6 expression and preserved TRPV5 expression. Together, these results indicate that ClC‐K2 is important for maintaining TRPM6 expression and distal magnesium transport pathways, whereas distal calcium transport is not dependent on the presence of ClC‐K2. This mirrors the different clinical presentations in Bartter syndrome, with antenatal Bartter syndrome, caused by pathogenic variants in SLC12A1 or KCNJ1, primarily leading to hypercalciuria, and classical Bartter syndrome, due to CLCNKB variants, being frequently associated with hypomagnesemia but with less severely affected urinary calcium handling [2, 3, 10, 23]. Moreover, the overlap with Gitelman syndrome, where pathogenic variants in SLC12A3 similarly cause impaired TRPM6‐dependent magnesium reabsorption, highlights the potential shared pathophysiological basis of magnesium wasting in these disorders and the challenging nature of distinguishing Gitelman syndrome from classic Bartter syndrome in clinical practice [50].

Gitelman patients and mice with Slc12a3 deficiency display hypomagnesemia and hypocalciuria, the latter resulting primarily from increased proximal tubular reabsorption [26, 27, 42]. In these mice, ablation of Slc12a3 caused atrophy of the early DCT1 segment, accompanied by hypertrophy of the CNT. In our study of Slc12a3‐deficient mice, the calcium transport machinery of the late DCT and CNT was preserved. This is in line with previous findings showing unchanged expression of TRPV5 and NCX in Slc12a3‐deficient mice [27, 51]. In contrast, TRPM6 was found to be reduced in Slc12a3‐deficient mice [51], consistent with our observations in both Slc12a3‐ and Clcnkb‐deficient mice. Slc12a3‐deficient mice display DCT1 atrophy [27, 51], which is related to impaired cell proliferation in DCT1 during development due to loss of NCC‐dependent transport [51]. Absence of basolaterally expressed ClC‐K2 would also inhibit NCC‐dependent sodium chloride transport. Moreover, reduced chloride efflux and consequent intracellular chloride accumulation would inhibit with‐no‐lysine (WNK) kinases, which normally stimulate NCC through the SPAK/OSR1 pathway [52]. Similar to the Slc12a3‐deficient mice, Clcnkb‐deficient mice developed a reduction in DCT1 cortical volume fraction consistent with DCT1 atrophy. However, in contrast to the Slc12a3‐deficient mice, where DCT2 and CNT volumes were not reduced [27], Clcnkb‐deficient mice displayed a general reduction in the cortical volume fraction of the entire distal convolution. In addition, some tubules in the distal convolution of Clcnkb‐deficient mice were dilated, surrounded by flattened cells. This flattening of cells could contribute to the observed reduction in cortical volume of the distal convolution.

Similarly to Slc12a3‐deficient mice, NCC inhibition by thiazides leads to hypomagnesemia. Furthermore, chronic thiazide treatment has been reported to induce structural remodeling of the DCT and apoptosis of DCT cells in some studies [53, 54]. However, in a study by Nijenhuis et al., NCC expression was significantly increased and no evidence of DCT apoptosis was detected after thiazide treatment in mice, while thiazide markedly reduced both Trpm6 mRNA expression and TRPM6 protein abundance, indicating that the reduction in TRPM6 expression was not secondary to loss of DCT viability [55]. This may be in line with our findings, since the expression of TRPM6 per cell as evaluated by immunohistochemistry is reduced. This reduction may be further amplified by volume loss associated with atrophy of the DCT, resulting in an overall decrease in TRPM6 abundance.

Tubular dilation is seen in all transgenic models of Bartter syndrome following ablation of Clcnkb, Slc12a1, and Kcnj1 [24, 25, 56, 57, 58], together with marked alterations in kidney volume and flow‐induced hydronephrosis. The severity of these alterations will also lead to additional effects on tubular mineral and electrolyte transport. Following deletion of Slc12a1 and in homozygous Slc12a1 ^I299F mice with late‐onset manifestation of antenatal Bartter syndrome, bilateral hydronephrosis developed [56, 57], with focal dilation of tubules, which at later stages will further disturb electrolyte and mineral handling by the kidney. In the antenatal mouse model with Kcnj1 deletion, similar findings of hydronephrosis and dilation of tubules in the distal tubule and collecting ducts were shown. In the Clcnkb‐deficient animals we observed a varying degree of hydronephrosis and dilation of the distal convolution, although we did not observe dilation of the collecting duct. As dilation is seen in all models of Bartter syndrome, these changes are likely secondary to primary dysfunction of the TAL.

In Kcnj1‐deficient animals, morphometric analysis of the cortex revealed a significantly reduced volume of cells from the proximal tubule, while the volume of distal tubules was significantly increased [57]. Further detailed analysis of Kcnj1‐deficient mice showed an increased DCT volume, while the CNT remained unaltered between genotypes [59]. Furthermore, NCC expression was increased in Kcnj1‐deficient mice, and they showed an exaggerated response to thiazide [59, 60], well in line with an increase in DCT function. This is in stark contrast to the Clcnkb‐deficient mice having a reduced expression of NCC and parvalbumin, showing a blunted response to thiazides [24] and a reduced DCT volume, indicating impaired DCT structure and function. In line with the location of the TRPM6‐dependent magnesium‐transporting machinery in the DCT, ablation of Clcnkb leads to its reduced abundance, which we observe at both the gene and protein levels. This would further augment renal magnesium wasting from the DCT in addition to the TAL, in comparison to that of the antenatal Bartter forms. As most Bartter cases are diagnosed clinically and genetically without histology material, no general comparison in subtypes of Bartter patients has been made with respect to segmental renal cortical volumes. In a kidney biopsy from a patient carrying a pathogenic variant in CLCNKB, there was a decrease in the expression of both parvalbumin and NCC in line with tubular remodeling in comparison to two control patients [61].

We found a marked decrease in Egf mRNA and expression of Pro‐EGF in Clcnkb‐deficient mice. EGF has been shown to stimulate magnesium transport via TRPM6, and pharmacological inhibition of the EGF receptor or pathogenic variants in the pro‐EGF gene leads to hypomagnesemia [30, 31, 32, 33]. Furthermore, EGF is expressed in the TAL and DCT [62, 63] and has been used as a marker of distal tubular mass. In fact, urinary EGF correlates with kidney volume following donor nephrectomy [34]. Moreover, in the general population, lower urinary EGF is associated with lower serum magnesium levels [34]. These findings are in line with the magnesium‐wasting in Clcnkb‐deficient mice. It would be of interest to further investigate the urinary excretion of EGF in patients with different forms of Bartter syndrome.

In the Clcnkb‐deficient mice investigated here, urinary calcium excretion does not differ from wildtype [24]. In classical Bartter syndrome patients with pathogenic variants in CLCNKB, hypercalciuria is less frequently observed and also less severe compared to antenatal Bartter syndrome. The same holds true for the development of nephrocalcinosis [10, 23]. The reason for this discrepancy has not been clearly established. We find that expression of the DCT/CNT calcium transport machinery is not visibly impaired in the kidney of Clcnkb‐deficient mice. While this has not been examined in mouse models of antenatal Bartter syndrome, there do not appear to be any changes in the CNT volume in Kcnj1‐deficient mice, albeit DCT volume is increased (here DCT1 and DCT2 were not discriminated) [59]. Therefore, it is likely that calcium transport is maintained in both classical and antenatal forms in these segments. The more severe calcium wasting in antenatal Bartter could result from differential effects on the TAL. In fact, while antenatal Bartter is likely to fully impair TAL transport, other efflux systems may allow for chloride ions to exit from the TAL and could present some explanation as to why hypercalciuria is not as severe and penetrant in patients with classical Bartter syndrome [10, 23].

ClC‐K1 has traditionally been regarded as segment‐specific, being expressed in the thin ascending limb. Our investigations have found it expressed in the medullary TAL by immunohistochemistry [24], although not in the cortical portion where paracellular calcium and magnesium transport occurs. However, patch‐clamp analysis did detect ClC‐K1‐like channel currents at low frequency in the cortical TAL, in agreement with a previous study describing ClC‐K1 channel activity in the segment [64], and transcriptome analysis of individually dissected adult mouse renal tubules [65], suggesting that Clcnka might be expressed in the cortical TAL at low levels. Others have shown that, while both loss of Clcnka and Clcnkb affect TAL transport, only loss of Clcnkb but not Clcnka reduced NCC expression, in agreement with our previous findings showing an absence of ClC‐Ka activity in the DCT [24]. Interestingly, in this mouse strain, wasting of both calcium and magnesium was observed following embryonic Clcnkb deletion [25], though this was not seen following postnatal deletion, in line with the variability observed in classical Bartter patients. We find that Clcnkb‐deficient mice show no significant alteration in the expression of TAL claudins; however, transtubular calcium transport in the TAL is likely markedly reduced since the transepithelial voltage gradient is expected to be inhibited or fully abolished following the ablation of Clcnkb.

A limitation of the present study is that we focused on a limited set of transport proteins considered most relevant to the primary aims of the study, but additional changes in gene and protein expression are likely present and could influence the urinary excretion of calcium or magnesium. Furthermore, transporter abundance does not necessarily reflect functional activity. Several key transport proteins, including NKCC2, NCC, and TRPV5, are subject to important post‐translational regulation through phosphorylation and altered membrane trafficking, which were not assessed in the current study. However, Clcnkb‐deficient mice were previously shown to display no difference in phosphorylated, active NKCC2, while phosphorylated NCC was strongly decreased, in line with the decrease in total NCC [24]. Consequently, the observed changes in expression should be interpreted within this context, and future studies addressing post‐translational regulation may provide additional insights into segment‐specific adaptations in Clcnkb‐deficient mice.

4. Methods

4.1. Animal Experiments

Animals were generated as described in detail previously [24]. In brief, exons 5–10 of the Clcnkb gene were flanked by loxP sites using a targeting vector. The resulting animals were mated with a Cre‐expressing strain to generate constitutive Clcnkb‐deficient mice. Clcnkb‐deficient mice were subsequently backcrossed with C57BL/6 mice. All mice were housed in the animal facility (Jena, Germany) and maintained under a 12‐h light/dark cycle, with ad libitum access to water and food. For RNA isolation and immunoblotting (5 WT and 7 Clcnkb ^−/−), mice were killed by cervical dislocation. The kidney tissue was dissected, rinsed in ice‐cold PBS, and then immediately flash frozen in liquid nitrogen. For immunohistochemistry experiments, mice (5 WT and 7 Clcnkb ^−/−) were deeply anesthetized and perfused transcardially with 4% paraformaldehyde in PBS at pH 7.4. All animal experiments were approved under license 02‐009/13 by the Thüringer Landesamt für Verbraucherschutz in Germany. The fixed tissues were dehydrated for paraffin wax embedding. Fixed kidneys from Slc12a3‐deficient mice have been previously described [42, 66]. Conditional Trpm6 mutant (floxed) mice were generated by IVF using oocytes from B6NCrl;B6N‐A<tm1Brd>Trpm6<tm1a(KOMP)Wtsi>/CipheOrl females and sperm from FlpE transgenic mice that constitutively express FLP‐recombinase to remove the lacZ‐neo cassette of the KOMP allele. In the resulting conditional Trpm6 mutant mice, exon 7 of the Trpm6 gene is flanked by loxP sites. Trpm6 floxed animals were then crossed with B6.129P2(Cg)Pax8tm1.1(cre)Mbu/J (JAX #028196) to generate animals with kidney‐specific Trpm6 deletion [67].

4.2. Processing of Tissue From Human Kidney

Kidney tissue was collected from patients undergoing nephrectomy for renal carcinoma. Samples were taken from tumor‐free regions of the resected kidneys. Following dissection, the tissue was immersion‐fixed in 10% formalin and then transferred to phosphate‐buffered saline (PBS) prior to dehydration and paraffin embedding. All tissue samples (n = 5) were collected with written informed consent from the participants. The study and its procedures were approved by the Biomedical Research Ethics Committee of Southern Denmark (license S‐20140159).

Monoclonal antibodies targeting TRPM6 were generated in mice in accordance with Danish legislation under animal experimental permit #2014‐15‐0201‐00043. The antibodies were raised against an intracellular epitope in the extreme C‐terminal domain of the mouse TRPM6 protein (CNLGFGQTIEPTEELPERDKNRSSLEDHTRL) using previously described procedures [68]. In brief, 30 μg of the peptide was conjugated to inactivated diphtheria toxoid and emulsified with GERBU adjuvant. NMRI mice were immunized with two doses over 14 days. Three days before spleen cell fusion, mice received an intravenous booster injection of the peptide conjugated with adrenaline. Hybridoma production was performed using the SP2 myeloma cell line as a fusion partner, as detailed [69]. Screening was done by ELISA, and positive clones were isolated through limiting dilution. Clone #16 was selected based on its ability to detect TRPM6 by immunohistochemistry and its absence of staining in Trpm6‐deficient mice (Figure 4) and was used as an undiluted culture supernatant in immunohistochemistry applications, while Clone #10 was selected for Western blotting, showing a specific band of the expected size that was absent in Trpm6‐deficient mice (Figure 4). All antibodies used in the manuscript are listed in Table 1.

TABLE 1.

Antibodies.

Antibody against	Host/Clone	Source	Dilution
AQP2	Rabbit	Novus NB110‐74682	1:500
BSND	Rabbit	Sigma HPA053836	1:1000 or 1:500
Calbindin‐D28k	Mouse, clone #2	Developed in‐house	1:1000
CLDN10	Rabbit	Invitrogen 38‐8400	1:100
CLDN16	Mouse, clone #18	Developed In‐house [70], biotinylated [71, 72]	1:50
ClC‐K	Rabbit	Alomone ACL‐004	1:1000 or 1:100
EGF	Goat	R&D Systems AF2028	1:1000
NCX1	Rabbit	Alomone ANX‐011	1:500
Parvalbumin	Rabbit	Sigma HPA048536	1:100
PMCA4	Mouse	Abcam AB2783	1:1000
SLC12A1 (NKCC2)	Mouse, clone #1	Developed in‐house [70]	1:10
SLC12A3 (NCC)	Mouse, clone #3	Developed in‐house [66, 73]	1:100 (IHC) or 1:500 (WB)
TRPM6	Mouse, clone #10 (WB) & 16 (IHC)	Developed in‐house, described here	Undiluted
TRPV5	Mouse, clone #10	Developed in‐house	Undiluted

4.3. Immunohistochemical Staining of Paraffin‐Embedded Tissue

Paraffin‐embedded kidney sections were processed for immunohistochemistry as previously described [68]. Briefly, tissue sections were rehydrated, and antigen retrieval was carried out by boiling in a Tris‐EGTA buffer (10 mM Tris, 0.5 mM EGTA, pH 9.0). To block endogenous peroxidase activity and free aldehyde groups, sections were treated with a solution of 0.6% hydrogen peroxide and 50 mM ammonium chloride in PBS. Primary antibodies were applied in PBS containing 0.1% Triton X‐100 and incubated overnight at 4°C. After washing, sections were incubated with horseradish peroxidase (HRP)‐conjugated secondary antibodies (DakoCytomation, Denmark). Detection of HRP activity was performed using the DAB+ Substrate Chromogen System (K3467, DakoCytomation). Finally, sections were counterstained with hematoxylin and visualized using an Olympus BX51 light microscope. For qualitative assessment of renal morphology, sections were stained with hematoxylin and eosin using standard histological procedures.

4.4. Morphometric Quantification of DCT1, DCT2, and CNT Cortical Volume Fractions

The relative cortical volume fractions of DCT1, DCT2, and CNT segments were estimated using point‐counting stereology on immunolabeled kidney sections obtained from WT (n = 5) and Clcnkb‐deficient (n = 5) mice. Images were acquired at 40× magnification and systematically sampled across the entire region of the renal cortex. Between 20 and 35 cortical images were analyzed per animal. A square lattice grid containing 980 test points (area per point = 5000 pixel²) was superimposed onto each image using FIJI/ImageJ software. Grid points overlying tubular profiles positive for NCC alone, TRPV5 alone, or both NCC and TRPV5 were counted manually. NCC‐positive/TRPV5‐negative tubular profiles were classified as DCT1, NCC‐positive/TRPV5‐positive profiles as DCT2, and TRPV5‐positive/NCC‐negative profiles as CNT. The percentage of cortical reference volume occupied by each tubular segment was estimated according to stereological principles as Vv = (Pp/Pt)*100, where Vv is volume fraction, Pp represents the number of grid points overlying the tubular cells of interest and Pt represents the total number of sampled grid points. For each animal, counts from all analyzed images were pooled and normalized to the total number of sampled points to obtain final segmental cortical volume fractions expressed as percentages.

4.5. Quantification of Enlarged/Cystic DCT1, DCT2, and CNT Tubules

Tubular lumen morphology was analyzed on the same cortical images used for stereological quantification of tubular volume fractions outlined above. For each animal, the whole cortical area of the section was analyzed. Tubular profiles were classified based on immunolabeling patterns as DCT1 (NCC‐positive/TRPV5‐negative), DCT2 (NCC‐positive/TRPV5‐positive), or CNT (TRPV5‐positive/NCC‐negative). Luminal cross‐sectional area was measured in FIJI/ImageJ, and tubular lumens with an area greater than 30 000 pixels² were classified as enlarged/cystic tubules. Counts from all analyzed cortical images were summed for each animal before genotype‐based comparisons.

4.6. Immunoblotting

Kidneys were homogenized using a TissueLyser II (Qiagen) in ice‐cold dissection buffer (0.3 M sucrose, 25 mM imidazole, and 1 mM EDTA, pH 7.2) containing phosphatase and protease inhibitors (Halt) followed by a low‐velocity spin (1000× g, 5 min, 4°C). For analysis of TRPM6, membrane fractions were prepared by a second centrifugation of 17 000× g, 30 min, 4°C. Standard procedures were utilized for sample preparation and SDS‐PAGE using 4%–15% gradient polyacrylamide gels (Criterion TGX Stain‐Free Protein Gels, BioRad). Equal quantities of total protein were loaded per lane as determined by the fluorescent Stain‐Free signal. The maximal deviations in total protein concentration between samples on individual blots were ±10%. Immunoblots were developed using SuperSignal West Femto chemiluminescent substrate (Thermo Scientific) or Western Lightning Pro Chemiluminescent Substrates (Revvity). Signal intensity in specific bands was quantified using ImageJ software.

4.7. Immunolabeling of Tissue for Fluorescent Microscopy

Immunostaining was carried out as previously described when using primary antibodies from two different species [68]. For double labeling with two antibodies raised in the same species and targeting distinct epitopes, a tyramide signal amplification (TSA) approach was used as detailed previously [74]. In brief, the first primary antibody was detected using HRP‐conjugated secondary antibodies, followed by signal development with a Cy3‐conjugated TSA substrate (TSA Cyanine 3, Perkin Elmer, Waltham, MA, USA). To enable subsequent staining, the sections were reboiled in Tris‐EGTA (TEG) buffer to strip bound antibodies while preserving the TSA signal. A second round of immunolabeling was then performed using the second primary antibody, followed by detection with Alexa Fluor 488‐labeled secondary antibodies.

4.8. RNA Extraction and Semi‐Quantitative Real‐Time PCR (qPCR)

Following isolation, kidneys were snap‐frozen in liquid nitrogen and stored at −80°C until RNA extraction. Total RNA was isolated using TRIzol reagent (Invitrogen, Carlsbad, CA, USA) as previously described [38, 74]. Samples were treated with DNase (Thermo Fisher Scientific Baltics UAB, Vilnius, Lithuania) according to the manufacturer's instructions. RNA was then reverse transcribed into complementary DNA (cDNA) using the iScript cDNA Synthesis Kit (Bio‐Rad, Copenhagen, Denmark). Quantitative PCR was carried out using the iTaq Universal SYBR Green Supermix (Bio‐Rad, Foster City, CA, USA) and gene‐specific primers as listed below. The primer sequences used are shown in Table 2. The data are presented relative to the expression of the geometric mean of 3 housekeeping genes, TATA‐binding protein (Tbp), Glyceraldehyde‐3‐phosphate dehydrogenase (Gapdh), and 18S ribosomal RNA (18s).

TABLE 2.

Primer sequences.

Gene	Forward primer	Reverse primer
18S	GCAATTATTCCCCATGAACG	GGCCTCACTAAACCATCCAA
Atp2b4	ATCTGCAGGGTTCCCAGATA	CTTAATGGACCTGCGAAAGC
Bsnd	CTATCCTAAGATCACCTTTGTG	TCATTTGGATGTGAGTGAAG
Calb1	GTTATATGATCAGGAATGGCAAC	GCAAAAGAATAAGAGCAAGG
Cldn10b	AGGAGTTCCCCTCCATGCT	ACCGCAGCGATCATTAGTC
Cldn16	GCCATATTCTCCACTGGGTT	AGTCATCAGCGTTCACCATC
Cldn19	CAGGTGCAATGCAAACTCTACG	ACTTCATGCCCACGACACTG
Cldn2	GGCTGTTAGGCACATCCAT	TGGCACCAACATAGGAACTC
Clcnkb	AAAGAAATGGAACTGTTGGC	TCTCCTGTTCATTGTTGAAG
Cnnm2	ACTACTTCGTCCTCATTCTTC	ACTTATTTTCCCCTGGAGAG
Cyp24a1	CCCAAAGGAACAGTCTTAAC	GGTCTAAACTTGTCAGCATC
Cyp27b1	AGTGTTGAGATTGTACCCTG	CGTATCTTGGGGAATTACATAG
Egf	CAAGTTCGTGACATTGTTTC	GAATTTCTAGTTCTCGTGGG
Egfr	CTGTCGCAAAGTTTGTAATG	GTACTCTTGGGTGTGAAAAC
Egr1	CAGAGTCCTTTTCTGACATC	GAGAAGCGGCCAGTATAG
Fxyd2	GTAAGAAACATAGGCAGGTC	AAGTTAAGGATCTGAGGCAG
GAPDH	TGATGGCATGGACTGTGG	CAGCAATGCATCCTGCAC
Kl	GGTTGCCCACAACCTACTTT	TGGGAGCTTAAGGCGATAGA
Pvalb	CGCTGAGGACATCAAGAAGG	CCGGGTTCTTTTTCTTCAGG
Slc12a1	GGCTTGATCTTTGCTTTTGC	CCATCATTGAATCGCTCTCC
Slc12a3	AACAGAATTAAGTCCCTTCG	CAGGTAAGGATGATCAGA
Slc41a3‐isoform 1	ACGTCCCTGATCATTGGCTT	CCAGCGTCATTTCCAGGTTT
Slc41a3‐isoform 2	GTCGCCATACCTATCCTGCT	CCAGCGTCATTTCCAGGTTT
Slc8a1	CTCCCTTGTGCTTGAGGAAC	CAGTGGCTGCTTGTCATCAT
Slc9a3	TGCCTTGGTGGTACTTCTGG	TCGCTCCTCTTCACCTTCAG
TBP	CAGCCTTCCACCTTATGCTC	TTGCTGCTGCTGTCTTTGTT
Trpm6	AAAGCCATGCGAGTTATCAGC	CTTCACAATGAAAACCTGCCC
Trpm7	TTTGGTGTTCCCAGAAAAGC	ACCAAGTTCCAGGACCACAG
Trpv5	CTGGAGCTTGTGGTTTCCTC	TCCACTTCAGGCTCACCAG
Trpv6	AGTACTACACATACTCATCCTG	AACATCACAATGTTGCCTTC

4.9. Statistics

Values are presented as means ± standard error of the mean. Normality of the datasets was tested using the Shapiro–Wilk test. Comparisons between two groups were performed using an unpaired Student's t‐test or, if samples were not normally distributed, a Mann–Whitney test, where p < 0.05 was considered statistically significant.

4.10. Writing and Editing

ChatGPT version 5.0 was employed for editing and improving the quality of the manuscript text. The authors subsequently checked and take full responsibility for the final manuscript text.

5. Conclusion

Loss of Clcnkb disrupts distal tubular structure and causes downregulation of TRPM6 and other magnesium transporters. These alterations provide a mechanistic basis for impaired tubular magnesium reabsorption following Clcnkb deficiency. In contrast, calcium transport pathways in the DCT and CNT remain largely preserved. These findings highlight a molecular basis for magnesium wasting in classical Bartter syndrome.

Author Contributions

Karsten Skjødt: investigation, methodology, validation, writing – review and editing, formal analysis. Lars Lund: methodology, writing – review and editing, resources. Christian A. Hübner: methodology, writing – review and editing, investigation. R. Todd Alexander: investigation, writing – review and editing, validation. Kirsten Madsen: resources, writing – review and editing, methodology. Milos Bogdanovic: methodology, validation, visualization, writing – review and editing, investigation, formal analysis, data curation. J. Christopher Hennings: investigation, methodology, writing – review and editing. Ida Appel: methodology, validation, visualization, writing – review and editing, investigation, formal analysis. Dominique Eladari: methodology, writing – review and editing, resources. Henrik Dimke: conceptualization, investigation, funding acquisition, writing – original draft, writing – review and editing, validation, methodology, visualization, supervision, project administration, formal analysis. Régine Chambrey: methodology, writing – review and editing, resources. Marleen L. A. Kortenoeven: methodology, investigation, funding acquisition, writing – review and editing, visualization, validation, formal analysis, data curation, project administration. Ernst‐Martin Fuchtbauer: investigation, methodology, validation, writing – review and editing, formal analysis.

Funding

This work was supported by Danmarks Frie Forskningsfond (3101‐00194B, 8045‐00011B, 1133‐00044B), Novo Nordisk Fonden (0087966), Carlsbergfondet (CF24‐2108).

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Figure S1: Localization of ClC‐K in the inner medulla. Co‐immunostaining revealed expression of ClC‐K in AQP2‐negative tubules in the mouse inner medulla, which was not changed in Clcnkb‐deficient mice. The scale bar represents 200 μm.

Figure S2: Localization of BSND in mouse kidney. (A–E) Barttin expression in WT mouse kidney was observed along the ascending loop of Henle and distal nephron, while expression was reduced in Clcnkb‐deficient mice in cortex and outer medulla, but preserved in the inner medulla. (F) qPCR showed that Bsnd mRNA expression was not changed in the Clcnkb‐deficient mice. (G–L) Co‐staining showed colocalization of Barttin with NCC (G, H) and NKCC2 (I, J). Additionally, Barttin expression was seen in AQP2‐negative cells in the cortical and outer medullary collecting duct (K, L). The scale bar represents 50 μm (A) or 200 μm (B–E, G–L).

Figure S3: Localization of BSND in the inner medulla. Co‐immunostaining revealed expression of BSND in AQP2‐negative tubules in the mouse inner medulla, which was not changed in Clcnkb‐deficient mice. The scale bar represents 200 μm.

Figure S4: Localization of BSND in human kidney. (A–D) In human kidney, Barttin was found throughout the TAL, DCT and CD. (E–L) Barttin colocalized with NKCC2 in cortex and medulla (F–H) and with NCC in the cortex (E). Barttin showed expression in AQP2‐negative cells in the collecting ducts (I–K). A weak expression was detected in the inner medullary thin limbs (L). The scale bar represents 200 μm.

Figure S5: Peptide preabsorption. Immunohistochemical staining for TRPM6 on paraffin‐embedded mouse kidney tissue shows that staining is absent after preincubation with the immunizing peptide.

Data Availability Statement

The data that supports the findings of this study are available in the Supporting Information of this article.