Dissociation of sodium-chloride cotransporter expression and blood pressure during chronic high dietary potassium supplementation

Dietary potassium (K+) supplementation is associated with a lowering effect in blood pressure (BP), but not all studies agree. Here, we examined the effects of short- and long-term K+ supplementation on BP in mice, whether differences depend on the accompanying anion or the sodium (Na+) intake and molecular alterations in the kidney that may underlie BP changes. Relative to the control diet, BP was higher in mice fed a high NaCl (1.57% Na+) diet for 7 weeks or fed a K+-free diet for 2 weeks. BP was highest on a K+-free/high NaCl diet. Commensurate with increased abundance and phosphorylation of the thiazide sensitive sodium-chloride-cotransporter (NCC) on the K+-free/high NaCl diet, BP returned to normal with thiazides. Three weeks of a high K+ diet (5% K+) increased BP (predominantly during the night) independently of dietary Na+ or anion intake. Conversely, 4 days of KCl feeding reduced BP. Both feeding periods resulted in lower NCC levels but in increased levels of cleaved (active) α and γ subunits of the epithelial Na+ channel ENaC. The elevated BP after chronic K+ feeding was reduced by amiloride but not thiazide. Our results suggest that dietary K+ has an optimal threshold where it may be most effective for cardiovascular health.


Contents
Administration. Male C57Bl6/J mice (Janvier, France) of 10-12 weeks of age were housed under standard conditions with a 12/12 hour dark/light cycle (18:00 lights off) and continual free access to rodent chow and water. In an initial cohort (cohort 1), the animals were implanted with radiotelemetric devices (detailed below) to record their BP a week before any dietary manipulations commenced, during which they were individually caged. To limit potential telemetry probe malfunctions during mouse handling and prevent changes in BP due to volume depletion this group followed the same dietary interventions but their blood and urine was not sampled. Diets were prepared from powdered commercial rodent diet (Teklad Diet: TD.08251.Envigo. USA), being nominally K + , Na + and Clfree, with ionic compounds (Sigma-Aldrich) added back to generate respective modified diets. 12 week old mice were fed a control diet of KCl (1.05% K + ) and NaCl (0.3% Na + ) (NK/NS) or a high NaCl (1.57% Na + ) (HS) diet for 7 weeks. Subsequently, mice were stratified to receive either a high KCl (5.25% K + ) or a high potassium citrate (KCit, 5.25% K + ) diet (HK) in combination with the control or high NaCl intake for 3 weeks. Following high K + feeding animals were fed a zero K + diet (0K) with either control or high NaCl level for 2 weeks. Animals maintained on HS/0K diet were treated with 37.5 mg/kg body weight hydrochlorothiazide (HCTZ, Sigma-Aldrich) administered as two i.p. doses 24 hours apart, with BP being recorded following the 2 nd dose. In a subsequent experiment, animals fed a chronic NS/HK (high KCl) diet were treated with HCTZ (as above) and after a 3 day washout period received a single i.p. dose of amiloride (5 mg/kg body weight, Sigma-Aldrich) with BP being recorded for the first 12 hours following injection. HCTZ and amiloride were prepared from solid by dilution in DMSO and then further diluted in sterile physiological saline solution before injection at the stated doses. A scheme of the diets and study outline is shown in Figure 1.
Telemetric Blood Pressure Measurements. BP recordings and associated surgery was performed using PA-C10 and HDX-11 radiotelemetry devices (Data Sciences International (DSI), USA) as described previously (1, 2). In vivo BP was recorded using Ponemah software (ver. 6.4, DSI) (two 1 minute recordings every 5 minutes). BP data was collected in a session of up to 3 days and the average BP for each hour was calculated. Where recordings were collected over multiple days, values from matched times of day were averaged and BP values are presented over a 24 hour period from 18:00 (ZT hour 0), the time at which lights in the animal facility automatically switch off. A four factor fixed-effects single-component model was then applied to plot the rhythmicity of the BP using cosinor curve fitting (1, 3). Curves were generated from the following equation: Y = mesor+amplitude*cos(period*(X-acrophase)).
No default constraints were defined for curve fitting. Initial values were: mesor = 1 (rule = *YMID), amplitude = 0.5 (rule = *YMAX-YMIN), period = 0.268 (rule = initial value, to be fit), acrophase = 1 *(Value of X at YMAX). The effect of diets on BP was analysed by comparison of fits analysis for the curves. The derived Midline Estimated Statistic Of Rhythm (MESOR) value is presented as the mean BP across a 24 hour period.
Tail cuff plethysmography. In a subsequent cohort of animals (cohort 2) conscious BP was recorded from the tail using an occlusion cuff and volume-pressure recording (VPR) sensor (Coda equipment, Kent Scientific). Animals were progressively acclimatised to the restraint and cuffing system over 3 days as previously described (4,5). VPR traces were obtained from individual animals sequentially not alongside each other, as in our experience animals from multiple channels don't look as relaxed and generally data collected in lower throughput has less variation between cycles. Recording settings; maximal cuffing pressure of 250mmHg, occlusion cuff deflation over 15 seconds for 1 cycle, 5 seconds inter cycle interval.
An initial 5 acclimatisation cycles were programmed, followed by up to 15 further repeating cycles where accepted values (tail volume > 15µL from calm animal) could be used for analysis. Individual cycles were interrogated by the researcher to discard potential movement artefacts. The mean from a minimum of 5 accepted cycles was calculated and is presented as the BP value of an individual for SBP or DBP. All recordings were performed during the dark phase, approximately 19:00 -22:00, in a blackout room under a single red light. BP was recorded from NK and HK fed animals in the same session. Animals being recorded from were moved to the recording room from a housing room during the light phase and had at least 6 hours acclimatisation to the area. In using this system to assess effects of HCTZ on BP, animals were briefly anesthetised by vaporised isoflurane and HCTZ injected (for 37.5mg/kg body weight) at least 4 hours before recordings were initiated.
Physiological phenotyping. Dietary effects on renal water and solute handling were determined by individually housing animals in metabolic cages (Tecniplast, Italy). Mice had ad libitum access to water while fed the different diets as detailed in the 'study design' section. Animals were acclimatised to the cage for 2 days, after which urine was collected over a 24-hour period, starting at 09:00, and for the same period the volume of water drank and food eaten was determined. Urine was centrifuged at 1000 g for 10 min before storage at -20°C until required.
Plasma and urine analysis. Animals were anaesthetised under 5% isoflurane for 60 s, Li-heparin capillaries inserted into the retro orbital plexus, and blood collected directly into Li-heparin coated tubes. Blood was immediately centrifuged at 5000 g for 2 minutes and the upper plasma layer collected and flash frozen.
Samples were stored at -20°C until required. Plasma and 24-hour urine samples were analysed for K + , Na + , Aldosterone concentrations were determined using an enzyme immunoassay kit (EIA-5298; range: 20-1,000 pg/ml; DRG International, Springfield, NJ) as per the manufacturer's protocol. Plasma renin activity was determined with an in-house kinetic assay (Erasmus Medical Center, Rotterdam, the Netherlands) as previously described (6) with detection limit = 0.17 ng Ang I/mL per hour. Plasma copeptin (7) was determined using a mouse specific ELISA (CEA365, Cloud-Clone Corporation).
Immunoblotting. Animals were euthanized by cervical dislocation, kidneys removed and protein homogenates prepared as previously described (1). Standard procedures were utilized for SDS-PAGE using 4-15% gradient polyacrylamide gels (Criterion TGX Precast Protein Gels, BioRad). Equal quantities of total protein were loaded per lane as determined by Coomassie blue staining. The maximal deviations in total protein concentration between samples on individual blots were ± 10%. Primary antibodies used for immunoblotting are listed in Supplemental Table 3. All antibodies have been extensively characterized in previous studies. Secondary antibodies (1:5000, Dako) were incubated for 1 hour at room temperature before immunoblots were visualised using the Enhanced Chemiluminescence system (GE Healthcare) or SuperSignal West Femto Chemiluminescent Substrate (Thermo Scientific) by an ImageQuant LAS 4000 imager (GE Healthcare). Signal intensity in specific bands were quantified using Image Studio Lite (Qiagen) densitometry analysis.
Real-time quantitative PCR (RT-qPCR). Animals were maintained on a similar control diet to the main study (0.25% Na + , 0.67% K + -Rodent Maintenance diet 1, Special Diet Services) under standard facility conditions with 12 h dark/light cycle. Whole kidney was snap frozen at ZT time 6 hours (13:00 hours -light phase) or 12 hours later at ZT 18 hours (01:00 -dark phase). RT-qPCR was performed as described (8) for Scnn1a (encoding αENaC) using the Roche Universal Probe Library. Scnn1a mRNA level was normalised to the mean expression of 3 control genes; 18S, cyclophilin A and TBP and data standardised to show the relative fold change in mRNA level at 01:00 compared to 13:00.
Immunohistochemistry. Half a kidney was immersion fixed in 4% paraformaldehyde (PFA) for 24 h at 4°C.
Tissue preparation, sectioning and immunolabeling was as described using primary antibodies targeting NCC, pNCC, γENaC or proliferating cell nuclear antigen (PCNA) and HRP conjugated secondary antibodies (Dako). Sections were visualised using a slide scanner system (Slide Scanner Olympus VS120 using a × 40 objective). For cell proliferation, cells were classed as PCNA positive (PNCA+) if there was clearly visible labelling of the nuclei. The number of PNCA+ cells were counted in 5 non-overlapping fields of view (each 0.31mm 2 ) across a section and the data reported as PCNA+/1.55mm 2 for each animal.

Liquid Chromatography Mass Spectrometry (LC-MS/MS) and
Bioinformatics. 150 μg of mouse kidney cortex homogenates were solubilized, digested to peptides, and quantified as described (9). Peptide samples were labelled individually using the TMTpro™ Label Reagent (Thermo Scientific), fractionated using high pH fractionation and analyzed by LC-MS/MS using an Easy nLC-1200 coupled to a Tribrid Fusion mass spectrometer (Thermo Scientific) for protein identification and quantification as described (10). Raw data were searched using SEQUEST and MASCOT against the reviewed uniprot mouse protein database (dated 13-01-2022) and quantified using Proteome Discoverer, version 2.4 (Thermo Scientific). The parameters for Proteome Discoverer were: precursor mass tolerance, 10 ppm; fragment mass tolerance, 0.02 Da; maximum miss cleavage, 2; static modification: cysteine carbamidomethylation, TMTpro modification on peptide N-terminal and lysine; variable modification: N-terminal acetylation, methionine oxidation. Percolator was used to calculate false discovery rate (FDR). Only rank 1 and high confidence (with a target false discovery rate (FDR) q-value below 0.01) peptides were included in the results. Protein quantification was based on normalized relative TMT reporter ion intensities. Only unique peptides were used for quantification. Proteins with a quantification P value lower than 0.05 were considered significantly changed and subjected to further downstream analysis. The mass spectrometry data have been deposited to the ProteomeXchange Consortium via the PRIDE partner repository with the dataset identifier PXD035354 (Username: reviewer_pxd035354@ebi.ac.uk, Password: dT1rUiz5). For gene ontology (GO) analysis, data were analyzed using Cytoscape with ClueGO Plugin version 2.5.9, using all without IEA as evidence. Related terms that share similar associated genes were fused to reduce redundancy. P < 0.05 was considered significant. Ingenuity pathway analysis (IPA, 2022 release) was performed using proteins with significantly changed abundances and the Ingenuity Knowledge Database (gene only) was used as the reference set for analysis.
Data processing and statistical analysis. Data was curated in Microsoft Excel. A 4 factor cosine curve was fitted to the expected regularly repeating pattern of BP, as previously described (1). Datasets were further analysed using Graph Pad Prism v9.1. For comparison of two groups, data meeting the statistical assumptions of normality were assessed using an unpaired Students t-test with level of significance set as 0.05. Comparisons of more than two groups were performed using one-or two-way (regular or repeated measurement) ANOVAs followed by a Dunnett or Tukey multiple comparison test (see individual figure   legends). An exception was the analysis of urine and plasma data where individual feeding groups, from a dataset of three or more, were compared to the control group using an unpaired Students t-test. In this instance the level of significance was set as 0.033 to correct for the false discovery rate (FDR) (11). Data is plotted as mean ± standard error (SEM) alongside individual values from independent animals, unless otherwise stated.
Study approval. The use of experimental animals is in agreement with a license issued by the Animal Experiments Inspectorate, Ministry of Food, Agriculture, and Fisheries, Danish Veterinary and Food Administration.
Representative panels (samples run on the same gel but non-contiguous) of NCC and pNCC protein expression in whole kidney by western blotting are shown above quantification graphs. Neither NCC or pNCC protein was significantly altered by long term (3 weeks) high NaCl (HS) feeding compared to animals fed control NaCl (NS) diets. Data is shown as mean ± SEM with individual data points representing different animals. Statistical testing by students 2-way t-test was performed.

Supplemental Figure 6. BP increases after 1 week and then is further elevated after 3 weeks of high KCl
feeding. Diet key: NS = normal NaCl (0.3% Na + ), NK = normal KCl (1.05% K + ), +KCl= high KCl (5.25% K + ), +KCit = high K citrate (5.25% K + ). Telemetric SBP recording over 24 hours for NS/+KCl fed animals was increased during dark period after 1 week and 3 weeks feeding. A. 24 hours recording, B. quantification of full recording and stratified by dark and light periods. Dark period was 12 hours from ZT hour 0 (18:00) as shown by bar above x axis. SBP recording over 24 hours for NS/+KCit fed animals was increased during dark period after 1 week and 3 weeks feeding; C. 24 hours recording, D. quantification of full recording and stratified by dark and light periods). Data is shown as mean ± SEM with individual data points representing different animals. Individuals shown in B and D make up the mean value represented in A and C. *P<0.05, **P<0.01. Data between groups is analysed by T-test corrected for false discovery rate. Figure 7. Sirius red labelling of kidney sections shows no evidence for hypokalemic nephropathy. Diet key: NS = normal NaCl (0.3% Na + ), NK = normal KCl (1.05% K + ), HK= high KCl (5.25% K + ), 0K = zero K + diet for 2 weeks. Short and chronic HK feeding was for 4 days and 3 weeks respectively. Representative images are shown from comparable sectional depths of paraffin embedded whole kidney tissue. Images from ≥5 animals/dietary condition were collected and compared. Mean ± SEM. Data compared using 1-way ANOVA with Tukeys multiple comparisons test.

Supplemental
Supplemental Figure 8. Immunohistochemical localization of NCC and γENaC after dietary K + manipulation. Diet key: NS = normal NaCl (0.3% Na + ), NK = normal KCl (1.05% K + ), HK= high KCl (5.25% K + ), 0K = zero K + diet. Representative images are shown from comparable sectional depths of paraffin embedded whole kidney tissue. Images from ≥5 animals/dietary condition were collected and compared. After 2 weeks of NS/0K diet, qualitative staining intensity of NCC and pNCC in the kidney is greatly increased relative to NS/NK intake, whereas staining intensity is decreased after feeding of a NS/+KCl diet (NS/HK) short-term (4 days) or chronically (3 weeks). After feeding a NS/OK or NS/NK diet γENaC labelling is predominantly intracellular, whereas after short-term NS/+KCl feeding it is both intracellular and located in the apical plasma membrane domain. On a chronic NS/HK diet γENaC is predominantly observed in the apical membrane domain. Scale bar represents 500µm.
Supplemental Figure 10. Correlations of NCC and ENaC with urine aldosterone concentration. Supplemental Figure 11. Gene ontology (GO) analysis by ClueGO for proteins signficantly changed by short-term or chronic KCl feeding. Short-term feeding is 4 days, whereas chronic is 3 weeks. GO terms are based on a p<0.02 cut-off. A) GO terms overrepresented in the pool of proteins increased in abundance after short-term KCl feeding. B) GO terms signficantly overrepresented in the pool of proteins decreased in abundance after short-term KCl feeding. C) GO terms signficantly overrepresented in the pool of proteins increased in abundance after chronic KCl feeding. D) GO terms signficantly overrepresented in the pool of proteins decreased in abundance after chronic KCl feeding. The number of genes in each GO term group is shown in each GO term subgroup, with the percentage of genes associated with a particular biological function shown in the individual pie charts.