Elevated aldosterone and blood pressure in a mouse model of familial hyperaldosteronism with ClC-2 mutation

Gain-of-function mutations in the chloride channel ClC-2 were recently described as a cause of familial hyperaldosteronism type II (FH-II). Here, we report the generation of a mouse model carrying a missense mutation homologous to the most common FH-II-associated CLCN2 mutation. In these Clcn2R180Q/+ mice, adrenal morphology is normal, but Cyp11b2 expression and plasma aldosterone levels are elevated. Male Clcn2R180Q/+ mice have increased aldosterone:renin ratios as well as elevated blood pressure levels. The counterpart knockout model (Clcn2−/−), in contrast, requires elevated renin levels to maintain normal aldosterone levels. Adrenal slices of Clcn2R180Q/+ mice show increased calcium oscillatory activity. Together, our work provides a knockin mouse model with a mild form of primary aldosteronism, likely due to increased chloride efflux and depolarization. We demonstrate a role of ClC-2 in normal aldosterone production beyond the observed pathophysiology.

T he adrenal steroid hormone aldosterone is an important regulator of blood pressure and electrolyte homeostasis. In the kidney, its binding to the mineralocorticoid receptor causes increased Na + reabsorption via the apical epithelial sodium channel ENaC and the basolateral Na + /K + -ATPase 1 , raising blood pressure. This effect is accompanied by increased K + secretion via the apical ROMK channel 2 , lowering serum potassium levels.
Excessive aldosterone production despite suppressed plasma renin activity and normal or low extracellular potassium levels (relative autonomy) is characteristic of primary aldosteronism (PA), the most common cause of secondary hypertension 3,4 . It accounts for about 6% of hypertensive patients in primary care and about 10% in specialized hypertension centers 5,6 . Compared with essential hypertension, PA is associated with an increased risk of cardiovascular disease 7,8 . Approximately one third of PA cases feature benign, typically unilateral tumors of the adrenal gland (aldosterone-producing adenomas), caused by somatic mutations in ion channels and pumps [9][10][11][12] . About two thirds have bilateral adrenal hyperplasia (idiopathic hyperaldosteronism), which may be due to an increased number and size of so-called aldosterone-producing cell clusters that similarly carry somatic mutations [13][14][15] .
Mendelian forms of PA are rare and often associated with early onset 16 . Known forms are inherited in an autosomal−dominant fashion and include familial hyperaldosteronism type I (FH-I) with mutations in CYP11B2 (aldosterone synthase) 17,18 , FH-III due to mutations in the potassium channel KCNJ5 9,19 and FH-IV caused by mutations in the T-type voltage-gated calcium channel CACNA1H 20,21 . Mutations in an L-type calcium channel (CAC-NA1D) that is also expressed in the central nervous system are found in a syndrome of PA, seizures, and neurologic abnormalities 10 .
Recently, mutations in the chloride channel ClC-2 encoded by the CLCN2 gene that is expressed in the zona glomerulosa were identified as cause of FH-II 22,23 . This syndrome had initially been described in an extended Australian kindred in 1992 24 . By exome sequencing of this kindred, we identified a heterozygous CLCN2 mutation (p.Arg172Gln, located at the cytoplasmic end of transmembrane helix C) that was present in eight subjects 22 ; incomplete penetrance was observed. In three additional unrelated kindreds with PA, the identical p.Arg172Gln was identified. Further families showed mutations located in the N-terminus (p. Met22Lys, p.Tyr26Asn), C-terminus (p.Ser865A), or the cytoplasmic interface of transmembrane helix K (p.Lys362del) 22 . In an independent study, Fernandes-Rosa and colleagues identified a single case with a de novo N-terminal p.Gly24Asp mutation 23 . By electrophysiology in transfected mammalian cells or Xenopus oocytes, all mutations were shown to cause gain of channel function; mutant channels also led to increased aldosterone production when expressed in an adrenocortical cancer cell line 22,23 . The underlying pathophysiology of PA due to CLCN2 mutations was inferred to be increased chloride conductance of glomerulosa cells (which have high intracellular chloride concentrations), depolarization, activation of voltage-gated calcium channels, calcium influx, increased expression of aldosterone synthase and PA 22,23 . Most recently, a somatic CLCN2 mutation was found in a single aldosterone-producing adenoma 25 .
Mutations in adrenal ion channels appear to differ in their ability to cause increased proliferation. KCNJ5 mutations that are found in aldosterone-producing adenomas are typically associated with massive bilateral adrenal hyperplasia when present in the germline, requiring bilateral adrenalectomy 9,19,26 . In contrast, germline mutations in the CACNA1H gene are associated with only microscopic glomerulosa hyperplasia 20 . Some subjects with CLCN2 mutations showed bulky adrenal glands or a small adrenal nodule upon computed tomography, but in other cases, adrenal imaging was unremarkable 22 , raising the question whether these mutations affect glomerulosa size in vivo and are responsible for proliferation in aldosterone-producing tumors.
Besides the identification of CLCN2 mutations in PA, very little has been known about any role of anion channels in zona glomerulosa (patho)physiology 27,28 . To investigate signaling pathways involved in PA and hypertension due to ClC-2 mutations, we here generate and characterize a mouse model carrying a heterozygous Clcn2 mutation at the position homologous to the most common mutation in humans (p.Arg172Gln). We study the effect of increased anion permeability on intracellular calcium, a known determinant of glomerulosa aldosterone production 27 .
Here we show that Clcn2 R180Q/+ mice have elevated Cyp11b2 expression and plasma aldosterone levels. Aldosterone:renin ratios are elevated in male mice, and blood pressure is slightly elevated, consistent with mild PA. The cellular correlate of these findings is an increased calcium oscillatory activity in adrenal glomerulosa cells. These results, combined with the finding of elevated renin levels in Clcn2 −/− mice, point to an important role of ClC-2 in adrenal physiology and disease.

Results
Generation of Clcn2 R180Q/+ mice. By sequence alignment, we identified Arg180 in the mouse ClC-2 protein (NP_034030) as the residue homologous to human Arg172 that is commonly mutated in FH-II 22 . Mouse Arg180 is located in a highly conserved protein region and is encoded by CGG as in humans (Fig. 1a). We introduced a point mutation homologous to human Arg172Gln 22 (chr16:20,712,754C > T (m38); p.Arg180Gln (NP_034030)) on exon 5 of the mouse Clcn2 gene by using CRISPR/Cas9-based genome editing (Fig. 1b, see Methods). Because FH-II is an autosomal dominant disease with heterozygous mutations in humans, mice with the heterozygous p.Arg180Gln mutation (Clcn2 R180Q/+ , Fig. 1c, d) were used for further characterization.
Both male and female Clcn2 R180Q/+ mice were viable and fertile, and no obvious changes in morphology or behavior were seen. Compared to C57BL/6N mice, no differences in litter size, mortality rate until the age of 3 weeks or sex distribution among litters were seen ( Supplementary Fig. 1a-c). When Clcn2 R180Q/+ mice were mated with WT mice, about half of surviving offspring were Clcn2 R180Q/+ ( Supplementary Fig. 1d).
Clcn2 R180Q/+ mice show no zona glomerulosa hyperplasia. Sanger sequencing of adrenal cDNA confirmed expression of the mutant allele in the adrenal glands of Clcn2 R180Q/+ mice (Fig. 1d). Total adrenal weight was not significantly different in Clcn2 R180Q/+ mice compared to controls (Fig. 2a). Similarly, by quantitative realtime PCR, overall expression levels of Clcn2 in adrenal gland did not differ between Clcn2 R180Q/+ mice and controls (Fig. 2b), and there was no sex dimorphism in Clcn2 expression (Fig. 2c). H&E staining showed normal adrenal zonation and morphology in Clcn2 R180Q/+ mice (Fig. 2d, e). Qualitative assessment of Cyp11b2 expression and zona glomerulosa morphology using in situ hybridization revealed no evidence of nodular glomerulosa hyperplasia or formation of aldosterone-producing adenomas in Clcn2 R180Q/+ mice (Fig. 3a, b).
In humans with PA and hypertension, renin levels are typically suppressed, contributing to an elevated aldosterone:renin ratio (ARR) 4 . PRC was unchanged between WT and Clcn2 R180Q/+ mice, even in male and female subgroups (Fig. 4f, g, Table 1, see Discussion). Whereas many mouse strains have two renin genes, C57BL/6 only has Ren1 31 . We thus examined relative renal expression levels of the Ren1 gene in the kidney by quantitative real-time PCR. No significant change in Ren1 expression was seen in the entire Clcn2 R180Q/+ group (1.16 ± 0.14-fold of WT, mean ± SEM, P = 0.9812; one-way ANOVA, Sidak´s MCP) (Fig. 4h), but subgroup analysis showed suppression of Ren1 expression in males (Fig. 4j).
Clcn2 R180Q/+ mice show decreased urinary sodium and chloride. The majority of patients with PA do not have hypokalemia 4 , and many patients with FH-II are similarly normokalemic 22 . We determined plasma and spot urinary electrolytes (potassium, sodium, chloride) in Clcn2 R180Q/+ mice and WT controls. Whereas no changes in plasma electrolytes or urinary potassium levels were observed (Fig. 6a- 2  20  22  21  11  12  13  14  15  16  17  18  19  3  4  5  6  7  8  9 Fig. 6f) were lower in Clcn2 R180Q/+ mice (Fig. 6b), suggesting increased retention, lower salt intake or a more dilute urine. Interestingly, in subgroup analysis, this effect was only significant in female, but not male mice.
Clcn2 R180Q/+ mice have slightly elevated blood pressure. To assess whether Clcn2 R180Q/+ mice develop hypertension, we measured blood pressure of unanaesthetized 3-month-old male mice by telemetry, using catheters implanted in the carotid artery (see Methods). Systolic and diastolic blood pressures as well as mean arterial pressure (MAP) of Clcn2 R180Q/+ mice were slightly but significantly increased when compared to WT  Clcn2 R180Q/+ mice show an altered high-salt response. Hypertension in states of high aldosterone is salt-dependent 34 , and impaired suppression of aldosterone upon salt loading is diagnostic of PA in humans 4 . We therefore fed Clcn2 R180Q/+ mice and WT controls a high-salt diet (HSD) (1.71% sodium) for 4 weeks, added 1% NaCl to the drinking water, measured plasma aldosterone as well as PRC and harvested adrenal glands and kidneys for RNA extraction. Plasma aldosterone dropped after HSD (Fig. 4a) in both genotypes, whereas PRC showed no significant change (Fig. 4f). We next measured adrenal Cyp11b2 expression and renal Ren1 expression. This revealed incomplete suppression of relative Cyp11b2 expression in Clcn2 R180Q/+ mice (2.85 ± 0.34) compared to WT mice (1.34 ± 0.23, mean ± SEM, P < 0.0001; Mann-Whitney test) (Fig. 4d).
Aged Clcn2 R180Q/+ mice show no evidence of end-organ damage. To study whether elevated aldosterone levels in Clcn2 R180Q/+ mice are associated with end-organ damage, we harvested kidneys and hearts from 11-month-old mice. H&E staining ( Supplementary Fig. 4), as well as staining for collagen (fibrosis, Supplementary Fig. 5) or macrophages (inflammation, Supplementary Figs. 6, 7) did not reveal qualitative differences between Clcn2 R180Q/+ mice and controls. Aged mice also did not show evidence of adrenal hyperplasia ( Supplementary Fig. 4).
Altered glomerulosa calcium signaling in Clcn2 R180Q/+ mice. Physiological stimuli of glomerulosa aldosterone production such as hyperkalemia and AT-II lead to an increase of the cytosolic     37 ) resulted in visible activity in some cells, either as single, isolated spikes or as temporally clustered bursts of activity (burst plots in Fig. 8a, b; Supplementary Movies 1 and 2). Raising the concentration of AT-II above physiological levels to 1 nM further increased the extent of bursting activity (Fig. 8a, b).
To examine whether calcium signaling in Clcn2 R180Q/+ mice is altered, we assessed the oscillatory activity in individual cells. To account for the potential lack of independence between cells from the same animal, linear mixed modeling was performed as described in Methods. The number of spikes in the absence of AT-II was insufficient for statistical analysis. Mean activity was significantly higher in  (Fig. 8c). These results support upregulated calcium signaling   Clcn2 R180Q/+ increases [AT-II]-dependent calcium bursting.
As activity in cells appears to be clustered, we defined bursts of calcium activity based on interspike interval analysis (see Methods). Again, analysis was performed using a linear mixed model to account for potential non-independence of cells from the same animal (see Methods). The number of calcium bursts per second and cell was significantly increased in Clcn2 R180Q/+ mice compared to WT at 1 nM AT-II for both, 3 and 5 mM of K + (Supplementary Fig. 8a). A trend towards an increased number of bursts at 20 pM of AT-II was not significant (Supplementary Fig. 8a). Burst duration was not significantly altered ( Supplementary Fig. 8b), and within bursts, the spiking frequency did not significantly differ between Clcn2 R180Q/+ mice and controls (Supplementary Fig. 8c). These results suggest that the augmented ClC-2 activity in Clcn2 R180Q/+ mice mainly increases the likelihood of initiating bursting activity in the presence of AT-II, likely through depolarization compared to wildtype controls.
Clcn2 −/− mice show increased renin levels. To study whether ClC-2 plays a role in physiological aldosterone production beyond FH-II, we investigated Clcn2 −/− mice 29 . Knockout mice had unaltered plasma aldosterone levels and Cyp11b2 expression (Fig. 4a, c). However, this occurred at the expense of elevated PRC and renal Ren1 expression (Fig. 4f, h), suggesting that activation of the renin-angiotensin system is required to maintain normal aldosterone levels in Clcn2 −/− mice, and that ClC-2 is involved in normal zona glomerulosa function.

Discussion
The Clcn2 R180Q/+ mouse model characterized here represents a genetic mouse model of FH-II with a rather mild phenotype. Previously characterized genetic mouse models with elevated aldosterone levels include a line derived from a mutagenesis screen with mutations in several candidate genes 38   in blood pressure and incomplete suppression of aldosterone synthase expression upon high-salt challenge (1.71% sodium diet, corresponding to the very high end of the spectrum of human salt consumption 46,47 ). A trend towards elevated aldosterone levels in Clcn2 R180Q/+ mice after HSD was not significant, possibly due to small sample sizes. Unlike in most humans with FH-II, PRC was not suppressed in Clcn2 R180Q/+ mice, likely due to only slightly elevated blood pressure. In line with this interpretation, male Clcn2 R180Q/+ mice (with more severe phenotype) showed significantly lower Ren1 expression levels than WT mice. Similar observations (mildly elevated aldosterone, normal plasma renin concentration (PRC)) have been made in mice with tissue-specific deletion of TASK channels 42 . In such a setting, extended periods of high-salt treatment might be necessary to produce end-organ damage 48,49 . Unlike aldosterone synthase-transgenic mice with more pronounced hyperaldosteronism 44   not show upregulation of renal aldosterone target gene expression by real-time PCR. Interestingly, ARR was only elevated in male, but not in female mice. We excluded differences in Clcn2 expression as cause of this observed sex dimorphism. Changes during the mouse estrus cycle may cause increased variability in parameters of the renin-angiotensin-aldosterone system in female rodents 50 ; in humans, aldosterone levels rise during the luteal phase 51 . In addition, sex dimorphism in adrenocortical tissue renewal was recently described, with female mice showing a 3-fold higher turnover than males 52 , and sex dimorphism has previously been reported in a mouse model of PA 41 .
The rather mild phenotype observed in mice is consistent with incomplete penetrance in humans 22 : In FH-II families, index cases typically had severe hyperaldosteronism and hypertension, which may be due to selection bias. However, in a large Australian kindred, three of eight mutation carriers did not have hypertension. Among these three subjects, ARRs were normal in one individual, borderline in a second and elevated in the third. These considerations suggest that factors other than the CLCN2 mutation contribute to disease expression. The mild phenotype of Clcn2 R180Q/+ mice could allow for the identification of such elusive genetic, epigenetic and/or environmental modifiers: Investigation of congenic mouse strains as well as effects of the microbiome or food on aldosterone levels and blood pressure would be instrumental 53 . A weakness of the current model is the absence of cardiac or renal damage. Homozygous mice (Clcn2 R180Q/R180Q ) may show a more severe phenotype and could enable studies of target organ damage.
The absence of macroscopic or microscopic adrenal hyperplasia in Clcn2 R180Q/+ mice suggests that CLCN2 mutations in humans and in mice may have a predominant effect on aldosterone synthesis rather than cellular proliferation; microscopic images of FH-II patients were unavailable 22 . A CLCN2 mutation has only very recently been described in an aldosteroneproducing adenoma; however, this was small tumor 25 , similar to those with CACNA1D and ATPase mutations.
The Clcn2 R180Q/+ mouse model also allowed us to further investigate the impact of the mutation on the cellular level. Calcium imaging in acute slice preparations revealed only very low activity of the zona glomerulosa in unstimulated adrenal glands from either WT or Clcn2 R180Q/+ mice, similar to results in WT mice by Penton and colleagues 35 . Stimulation with AT-II led to an increase in calcium signaling and the appearance of calcium bursts, further defining the mechanism by which aldosterone synthesis is regulated under physiological conditions; Penton and colleagues even observed continuous activity without intermittent silent periods upon stimulation with extremely high AT-II and K + concentrations. Clcn2 R180Q/+ mice showed an increase of the mean activity per cell compared to controls in the presence of AT-II, demonstrating a higher sensitivity for stimulation, accounting for the increased aldosterone production in mice and patients. A significant increase in the number of bursts could only be observed at 1 nM of AT-II while the trend was also visible at 20 pM but not significant, likely due to the smaller overall number of bursts and small effect size. In contrast, frequencies during bursting and burst lengths remained unchanged. These results support a main role of ClC-2 in cellular depolarization 22 and initiation of calcium signals in the form of bursts.
Future studies may compare the phenotype, adrenal morphology and calcium signaling of the Clcn2 R180Q/+ model with other models of familial hyperaldosteronism. Mouse models with gain-of-function mutations in calcium channels have so far not been described, but would be particularly interesting to assess. The physiological role of ClC-2 in zona glomerulosa function, suggested by elevated renin expression in Clcn2 knockout mice, also warrants further investigation.
Clcn2 R180Q/+ mice can be obtained exclusively for non-commercial research purposes (material transfer agreement). Recipients must treat, keep and use mice according to applicable regulations (e.g., animal welfare and protocols) and pay for packing, shipping, and handling. Recipients must not distribute the mice, their progeny or other materials derived from these mice without written consent of the provider and must cite this publication.
Animals and sample collection. The Clcn2 −/− strain was kindly provided by Dr. Thomas Jentsch, the generation of this mouse model was described by Bösl et al. 29 . All strains were bred and housed at the Zentrale Einrichtung für Tierforschung und wissenschaftliche Tierschutzaufgaben (Heinrich Heine University Düsseldorf), the Forschungseinrichtungen für Experimentelle Medizin (Charité-Universitätsmedizin Berlin) and/or the Forschungszentrum Jülich (only Clcn2 R180Q/+ mice studied at age 11 months). They were maintained under specific pathogenfree conditions in a 12-h light/dark cycle with ad libitum access to food and water. All animal experiments were approved by the local authorities (Landesamt für Natur, Umwelt und Verbraucherschutz Nordrhein-Westfalen and Landesamt für Gesundheit und Soziales Berlin) and performed under consideration of all relevant ethical regulations.
To harvest organs for RNA extraction, mice were anesthetized using ketamine 100 mg kg −1 + xylazine 10 mg kg −1 i.p., followed by cervical dislocation. After opening of the thoracic cavity, blood was collected by cardiac puncture and transferred into Vacutainer EDTA tubes (Becton Dickinson). Blood samples were centrifuged (10 min, 2000g, 4°C), and the supernatant (plasma) was stored at -20°C until further analysis. Adrenal glands were harvested and weighed, and kidneys and brains were collected. All organs were stored in tubes containing RNAlater RNA stabilization reagent (Qiagen; 4°C overnight, then -20°C) until RNA extraction. For histology, animals were anaesthetized using ketamine 100 mg kg −1 + xylazine 10 mg kg −1 i.p., and organs were fixed by cardiac perfusion. A 23G blood collection set (Becton Dickinson) was inserted into the left ventricular chamber. After opening of the right atrium, the heart was perfused with 10 mL PBS containing 10 U/mL heparin, followed by 10 mL 4% paraformaldehyde in PBS (pH 7.4; Santa Cruz). Adrenal glands were harvested and weighed. In addition, kidneys, brains, hearts, and testis were removed. All organs were fixed in 10% formalin solution (neutral buffered; Sigma Aldrich) for 18-24 h. 11-month-old mice were anesthetized with isoflurane (1 ml/l open drop). Following cervical dislocation, hearts, kidneys, and adrenal glands were harvested and kept in PBS + 0.03% NaN 3 for 72 h after fixation until further processing. Samples were dehydrated in ethanol (70%, 80%, 96%, and 100%) and xylene, followed by embedding into paraffin (Merck Millipore) and storage at 4°C (for in situ hybridization) or room temperature until further investigations. Spot urine was collected over three days in the morning followed by 1 week of recovery and blood collection via cardiac puncture (26G needle, 1 ml syringe) into Microtainer (Lithium-Heparin) tubes (Becton Dickinson). One group of mice per genotype received a HSD (E15431-34 containing 1.71% sodium, Ssniff Spezialdiäten) and 1% NaCl via drinking water for 4 weeks. Mice were euthanized by cervical dislocation under anesthesia, and blood and organs were collected as above.
For Sirius red staining, sections of 11-month-old mice were deparaffinized and rehydrated as above, incubated in hematoxylin for 8 min and washed in tap water for 5 min. Sections were incubated in 0.25 g/L Direct Red 80 (Sigma Aldrich) in 1.3% picric acid (Sigma Aldrich) for 60 min, washed in 0.5% acetic acid for 1 min, dehydrated (70%, 95% and 100% ethanol and xylene) and mounted using VectaMount.
Aldosterone ELISA. Plasma aldosterone levels were measured by enzyme-linked immunosorbent assay (ELISA, RE52301; IBL International GmbH) according to the manufacturer's instructions.
Determination of PRC. PRC was determined by measuring the capability of the sample to generate Angiotensin I (AT-I) from excess renin substrate 55 . AT-I concentrations were determined using a radioimmunoassay (RIA) with the following modifications: high binding RIA tubes (Greiner Bio-One) were coated with 300 µl of 1:60,000 diluted anti-AT-I antibodies for 20 h at 4°C. After washing with TRIS buffer, tubes were incubated with 400 µl blocking buffer containing 0.5% BSA for 2 h at room temperature to prevent unspecific binding. Tubes were washed again and stored at -20°C. For the RIA, 3 × 50 µl of the pre-treated samples and 200 µl 125 I-labeled AT-I with an activity of 5000 cpm (Amersham Biosciences) were added to the coated tubes und incubated for 20 h at room temperature. After three times washing with 1 ml washing buffer, the AT-I concentration was measured using a γ-counter (Berthold LD 2111). The concentration was determined according to the standard curve.
Determination of electrolytes. Plasma and urine electrolytes (K + , Na + , Cl − ) were measured at the Animal Phenotyping Facility at Max-Delbrück-Centrum (MDC) Berlin using an AU480 Clinical Chemistry Analyzer (Beckman Coulter).
Blood pressure measurements. Blood pressure was measured via telemetry. A telemetric catheter (DSI PhysioTel PA-C10; weight 1.4 g; Data Science International) was implanted in 3-month-old male mice under general anesthesia using ketamine 100 mg kg −1 + xylazine 10 mg kg −1 i.p. Via a median neck incision, the right common carotid artery was located and dissected. The carotid artery was ligated near the bifurcation, and two loops were placed caudal to the ligature.
Caudally the blood flow of the common carotid artery was stopped using a miniature clamp, an angled cannula tip was inserted, and the catheter was placed into the vessel. After releasing the mini clamp, the catheter was pushed forward by 1.5 cm towards the aortic bifurcation and fixed by closing the loops. A subcutaneous cavity was prepared for the transmitter. The wound was closed using 3-0 Prolene. Rimadyl 5 mg/kg s.c. was given intraoperatively and on the first, second, and third postoperative day. Mice were fed a standard diet (V1124-300 containing 0.24% sodium, ssniff Spezialdiäten). After 1 week of recovery, blood pressure and heart rate were recorded for twelve days (every 5 min, each 10 s). Locomotor activity was recorded based on signal strength changes that occurred with changes in the orientation of the animal relative to the receiver or changes in the distance of the animal from the receiver. Quality controlled sensor data were analyzed by mixed effect modeling using the lme4 56 R package, with animal and time of day as random effects, day-night status and genotype as fixed effects. Significance of genotype, was done through likelihood ratio test comparison with a model omitting the tested predictor as implemented in the lmtest 57 R package. Nights were defined as periods when light was off (6 p.m. to 6 a.m. or 7 p.m. to 7 a.m.). Blood pressure data were plotted as loess regressions using the ggplot2 58 R package, separately by genotype, showing 95% confidence interval of loess parameters in the figure.
Calcium imaging of acute adrenal slices. To prepare the adrenal slices, mice were anesthetized with isoflurane (1 ml l −1 open drop), followed by cervical dislocation. Both adrenal glands were rapidly removed and transferred to ice-cold bicarbonatebuffered saline (BBS: 125 mM NaCl, 2 mM KCl, 26 mM NaHCO 3 , 0.1 mM CaCl 2 , 5 mM MgCl 2 , 10 mM glucose, continuously gassed with 5% CO 2 /95% O 2 ). Surrounding fat was removed and adrenal glands embedded in 4% low-melting temperature agarose in BBS. Agarose blocks were mounted on a vibratome (7000 smz-2, Campden Instruments) and cut into 100-150 µm slices at 4°C. Slices were kept at 35°C for 30 min in BBS, followed by storage for up to 6 h in BBS supplemented with 2 mM CaCl 2 at room temperature. Adrenal slices were placed in a cell culture insert (Merck Millipore) in one well of a 24 well plate for staining. The outer reservoir was filled with 800 µl of BBS (supplemented with 2 mM Ca 2+ ). Calbryte 520 AM (AAT Bioquest) dissolved in DMSO + 2% Pluronic F-127 was diluted to a concentration of 9.2 µM in 200 µl BBS (supplemented with 2 mM Ca 2+ ) and filled into the inner reservoir on top of the adrenal slice. Loading occurred on top of a heating block set to 37°C for 1-1.5 h while continuously gassing the outer reservoir with 5% CO 2 /95% O 2 . After dye loading, slices were transferred to the recording chamber. The chamber was continuously perfused with recirculating solution from a reservoir gassed with 5% CO 2 /95% O 2 and heated in line via a heating coil, resulting in a temperature of the perfusion solution of 33 ± 1°C. During camera exposure, fluorescence was excited using a blue LED (pE-300 ultra, CoolLED), passing through a 474/27 nm bandpass filter (AHF Analysentechnik) and recorded using a 40 × /0.8 NA objective (Olympus), a 432/ 523/702 triple-band filter (AHF Analysentechnik) and an OptiMOS camera (Qimaging). Images were taken every 100 ms with an exposure setting of 10-15 ms. Regions of interest corresponding to individual cells were chosen manually from recorded images using Fiji 59,60 . Criteria for selection were: (1) appearance in glomerular structures located near the capsule and (2) visible activity during the perfusion with 1 nM angiotensin II (AT-II). In total, slices were prepared and recorded from eight wildtype (4 male, 4 female; 9 slices at 3 mM [K + ], 12 slices at 5 mM [K + ]) and eight Clcn2 R180Q/+ mice (4 male, 4 female; 11 slices at 3 mM [K + ] and 10 slices at 5 mM [K + ]).
Calcium signals were extracted and analyzed using a custom Python script (Supplementary Software 1). Zona glomerulosa calcium signals in acute adrenal slices were identified based on sudden increases in intracellular calcium concentration, followed by a slow decay 35 . For calcium spike detection, background noise and the signal decay caused by dye bleaching were removed by subtracting the minimum within a rolling window of 50 frames (5 s). Afterwards, the change in signal from one time point to the next was calculated to identify spikes generated by sudden rises of the intracellular calcium concentration. This method allowed for successful identification of isolated spikes as well as spikes within bursts, even if the baseline of the calcium concentration did not reach basal levels. To separate true activity from random background fluctuations, a manually set threshold for spike registration (above the background level as determined in parts of the recording without activity) was applied. Remaining artifacts from motion of the slice itself and perfusion were removed after manual inspection of all detected spikes. The code and a sample dataset can be found in Supplementary Code 1. Calcium spikes in zona glomerulosa cells appeared to become more temporally clustered with increasing AT-II (Fig. 8a, b). To define these clusters of activity (bursts) and separate them from isolated spikes and one another, we first determined the intervals between successive calcium spikes. These interspike intervals are shorter during bursts (one spike quickly follows the next) while they get longer outside of bursts. Plotting the distribution of the interspike intervals reveals a relatively large peak towards shorter values (few seconds) corresponding to the typical distances between spikes during bursting ( Supplementary Fig. 9). Increasing [K + ] and AT-II leads to more bursting (Fig. 8) resulting in more recorded interspike intervals during bursts and thus a more defined distribution at shorter values ( Supplementary Fig. 9). These distributions were similar in Clcn2 R180Q/+ and controls ( Supplementary Fig. 9a, b), with a trend towards more bursting at increasing [K + ] and AT-II levels represented by a more defined distribution of values below 5 s ( Supplementary Fig. 9a, b). Based on these interspike interval distributions, we defined bursts as clusters of activity (with at least three spikes) containing less than 4 s of quiescence.
Single cell calcium signaling data were analyzed by mixed effect modeling using the lme4 56 R package, with individual animals as random and genotypes as fixed effects with random intercepts. Inclusion of individual slices or sex into the model did not significantly improve the model as determined by likelihood ratio comparisons. Significance testing was performed through likelihood ratio test comparison with a model omitting the fixed effect using a double sided χ2 test.
Statistics. Raw data were analyzed using Microsoft Office Excel 2016 and GraphPad Prism 7.00 or 8.2.0 as well as R for blood pressure analysis and calcium imaging (see above for details). Statistical tests were performed two-sided and are specified in the figure legends. To account for pseudoreplication, linear mixed modeling was used in the analysis of blood pressure and imaging data 61 .
Reporting summary. Further information on research design is available in the Nature Research Reporting Summary linked to this article.

Data availability
Datasets generated during the current study are included in the article and supplementary information files, or otherwise available from the corresponding author upon request.

Code availability
Custom code is available as Supplementary Software 1.