High salt intake activates the hypothalamic–pituitary–adrenal axis, amplifies the stress response, and alters tissue glucocorticoid exposure in mice

Abstract Aims High salt intake is common and contributes to poor cardiovascular health. Urinary sodium excretion correlates directly with glucocorticoid excretion in humans and experimental animals. We hypothesized that high salt intake activates the hypothalamic–pituitary–adrenal axis activation and leads to sustained glucocorticoid excess. Methods and results In male C57BL/6 mice, high salt intake for 2–8 weeks caused an increase in diurnal peak levels of plasma corticosterone. After 2 weeks, high salt increased Crh and Pomc mRNA abundance in the hypothalamus and anterior pituitary, consistent with basal hypothalamic–pituitary–adrenal axis activation. Additionally, high salt intake amplified glucocorticoid response to restraint stress, indicative of enhanced axis sensitivity. The binding capacity of Corticosteroid-Binding Globulin was reduced and its encoding mRNA downregulated in the liver. In the hippocampus and anterior pituitary, Fkbp5 mRNA levels were increased, indicating increased glucocorticoid exposure. The mRNA expression of the glucocorticoid-regenerating enzyme, 11β-hydroxysteroid dehydrogenase Type 1, was increased in these brain areas and in the liver. Sustained high salt intake activated a water conservation response by the kidney, increasing plasma levels of the vasopressin surrogate, copeptin. Increased mRNA abundance of Tonebp and Avpr1b in the anterior pituitary suggested that vasopressin signalling contributes to hypothalamic–pituitary–adrenal axis activation by high salt diet. Conclusion Chronic high salt intake amplifies basal and stress-induced glucocorticoid levels and resets glucocorticoid biology centrally, peripherally and within cells.

assessed for each extraction by calculating the corticosterone concentration of a known sample following extraction. The average extraction efficiency throughout the experiments carried out was 82 ± 7%.
Copeptin & aldosterone measurement. Mice were euthanised by decapitation (between 0715 and 0800 local time), trunk blood was collected on ice, and separated by centrifugation and stored at -80⁰C. After a single thaw of all samples, copeptin (CEA365Mu; Cloud-Clone Corp., USA; lower limit of detection 9pg/ml) and aldosterone (ADI-900-173; Enzo Life Sciences, UK; sensitivity of detection 4.7pg/ml) and were measured by commercial ELISA. Antibody-based methods of aldosterone can over-estimate aldosterone due to crossreactivity with corticosterone and other steroids. The assay used has the following cross reactivity with other steroids: 11-Deoxycorticosterone (0.3%), Corticosterone (0.19%), Progesterone (0.20%), and <0.001% for Cortisol, DHT, Estradiol, Testosterone. The expression of renin in kidney homogenate was also measured to confirm suppression of the renin-angiotensin-aldosterone system with high salt intake. Corticosterone Binding Globulin binding capacity. Terminal plasma samples were diluted (1:100) and stripped of endogenous steroids using dextran-coated charcoal. To calculate CBG specific binding, non-specific binding had to be subtracted from total binding. To measure total binding, samples were incubated with radioactively labelled corticosterone ([1,2,6,7-3H]corticosterone). To measure non-specific binding, samples were first incubated with high concentrations of unlabelled corticosterone in addition to radioactively labelled corticosterone. This meant that it was likely that the [1,2,6,7-3H]corticosterone would bind to unsaturated non-specific binding proteins. Free [1,2,6,7-3H]corticosterone was removed by further incubation with dextran-coated charcoal. The remaining radioactively labelled corticosterone bound to CBG was quantified by scintillation spectrophotometry. The maximal binding capacity (Bmax), measured in disintegrations per minute, was estimated using non-linear regression. RNA isolation and quantitative PCR. Tissues were snap frozen in dry ice and stored at -80⁰C. RNA was extracted from mouse tissue, using the RNeasy Mini Kit (Qiagen, Hilden, Germany), by adding 300 -600 μL RLT buffer (from the RNeasy Mini Kit) and a 5 mm stainless steel bead (Qiagen) to a 2 mL tube containing specific mouse tissue for homogenization, carried out using TissueLyser II (Qiagen) at 30 Hz for 1 min. The volume of RLT buffer was dependent on tissue weight as per manufacturer's instructions. Samples were centrifuged for 3 min at 8000 x g, and supernatant mixed with 1 volume of 70% ethanol (50% ethanol for liver tissue) and transferred to a spin column. Various spin steps and a DNAase step (RNase-free DNase Set, Qiagen) were then carried out according to manufacturer's protocol before RNA was eluted in 30-50 μL RNase free water (Thermo Fisher). RNA concentrations, in ng/mL, were measured using a nanodrop 1000 spectrophotometer (Thermo Fisher). RNA integrity was assessed using automated electrophoresis (Bioanalyzer 2100, Agilent, CA, USA). The software assigns an RNA integrity number (RIN) to a total RNA sample, which is an objective metric of RNA quality. RIN ranges from 1 (completely degraded RNA) to 10 (highly intact RNA). RNA samples of a RIN of 7 or above were taken forward for reverse transcription.
cDNA was synthesised from 500ng of RNA, using the High Capacity cDNA Reverse Transcription (RT) kit (Applied Biosystems, CA, USA), where 10μL of 2x RT buffer and 1 μL of 20x enzyme mix were added to the diluted RNA and made up to 20 μL with nuclease free water. This was then reverse transcribed on a Veriti Thermal Cycler (Applied Biosystems). Negative controls were included, where the reverse transcriptase enzyme or RNA was replaced with an equal volume of water.
Primers for RT quantitative real-time polymerase chain reaction (RT-qPCR) were designed using the Universal Probe Library (UPL) Assay Design Centre. All designed UPL primers were synthesised by Integrated DNA technologies (IDT, IA, USA). Primers were designed where the amplicon spanned an intron, excluding the first or last intron. The resulting cDNA from the RT reaction was amplified and quantified using qPCR. Equal volumes of cDNA from samples were pooled together to construct an 8 point standard curve. The top standard of pooled samples were diluted 1:5 with nuclease-free water before serial dilutions (1:2) were carried out for standards 2-7. The final standard was used as a blank (nuclease-free water). The remaining cDNA was diluted 1:40 to fit in the middle of the standard curve. Sample/standard (2μL) was added to 8 μL of the qPCR master mix in triplicate in a 384 well plate. Primers and dual hybridisation probes used in RT-qPCR reactions are illustrated in Supplemental For all qPCR assays, the Roche LightCycler 480 (Roche Diagnostics Ltd., Basel, Switzerland) was used to perform thermocycling and fluorescence detection, using the specific cycling conditions. Threshold cycle/crossing point (Cq) values for each well were quantified, using the automated LightCycler 480 software. Each triplicate was analysed and excluded if the Cq SD was > 0.5. For each gene of interest and reference gene, the standard curve was analysed and had to meet criteria of standard curve efficiency being within the range of 1.7-2.1 and error < 0.05. Gene expression was analysed as relative expression as each gene was normalised against the mean expression of a minimum of 2 reference genes. A panel of reference genes were tested for each tissue (including Actb, Gapdh, Hprt, Rn18s and Tbp) and appropriate reference genes were selected for each tissue on the basis that they did not differ between treatment groups. The reference genes were: Actb and Tbp for adrenal, hippocampus and anterior pituitary; Actb and Hprt for liver; Actb and Rn18s for kidney cortex/medulla and aorta; Actb and Gapdh for heart; Gapdh and Tbp for hypothalamus. Data were then log transformed to give symmetry and normality tests were conducted. RT-qPCR data were handled in accordance to the MIQE guidelines. Q values adjusted p values controlling for the false discovery rate (FDR), were also reported to adjust for multiple comparisons in gene expression data for tissues.     glucocorticoid receptor (Nr3c1) and C) 11b hydroxysteroid dehydrogenase type 1 (Hsd11b1) from mice fed either 0.3% Na diet (Control, open circles) or 3% Na diet (High Salt; grey circles) for 14 days. Individual values are shown with group mean±SD; statistical comparisons were made using Student's unpaired t test with two-tailed p values stated. Figure S6. mRNA abundance in kidney cortex and medulla of A) FK506 binding protein 5 (Fkbp5); B) glucocorticoid receptor (Nr3c1); C) 11b hydroxysteroid dehydrogenase type 1 (Hsd11b1) and D) 11b hydroxysteroid dehydrogenase type 2 (Hsd11b2) from mice fed either 0.3% Na diet (Control, open circles) or 3% Na diet (High Salt; grey circles) for 14 days.
Individual values are shown with group mean±SD; statistical comparisons were made using Student's unpaired t test with two-tailed p values stated. Figure S7. mRNA abundance in heart and aorta of A) FK506 binding protein 5 (Fkbp5); B) glucocorticoid receptor (Nr3c1) and C) 11b hydroxysteroid dehydrogenase type 1 (Hsd11b1) from mice fed either 0.3% Na diet (Control, open circles) or 3% Na diet (High Salt; grey circles) for 14 days. Individual values are shown with group mean±SD; statistical comparisons were made using Student's unpaired t test with two-tailed p values stated.