Caffeine intake antagonizes salt sensitive hypertension through improvement of renal sodium handling

High salt intake is a major risk factor for hypertension. Although acute caffeine intake produces moderate diuresis and natriuresis, caffeine increases the blood pressure (BP) through activating sympathetic activity. However, the long-term effects of caffeine on urinary sodium excretion and blood pressure are rarely investigated. Here, we investigated whether chronic caffeine administration antagonizes salt sensitive hypertension by promoting urinary sodium excretion. Dahl salt-sensitive (Dahl-S) rats were fed with high salt diet with or without 0.1% caffeine in drinking water for 15 days. The BP, heart rate and locomotor activity of rats was analyzed and urinary sodium excretion was determined. The renal epithelial Na+ channel (ENaC) expression and function were measured by in vivo and in vitro experiments. Chronic consumption of caffeine attenuates hypertension induced by high salt without affecting sympathetic nerve activity in Dahl-S rats. The renal α-ENaC expression and ENaC activity of rats decreased after chronic caffeine administration. Caffeine increased phosphorylation of AMPK and decrease α-ENaC expression in cortical collecting duct cells. Inhibiting AMPK abolished the effect of caffeine on α-ENaC. Chronic caffeine intake prevented the development of salt-sensitive hypertension through promoting urinary sodium excretion, which was associated with activation of renal AMPK and inhibition of renal tubular ENaC.

Scientific RepoRts | 6:25746 | DOI: 10.1038/srep25746 (ENaC), and sodium hydrogen (Na/H) exchanger are major regulators of renal tubular sodium reabsorption and of blood pressure 18,19 . Among them, ENaC mediates apical entry of sodium in the aldosterone-sensitive distal tubule and collecting duct and accounts for the rate-limiting step for sodium reabsorption in these nephron segments 20 . ENaC activity is stimulated by kinases such as SGK1 and PKA and inhibited by PKC, ERK1/2 and AMPK 21 . Caffeine can activate AMPK in different type of cells, including the vascular endothelial cells, skeletal muscle cells and hepatocytes [22][23][24] . However, it is unknown whether caffeine regulates ENaC function in distal nephron through AMPK or other kinases.
In the present study, we hypothesized that long-term caffeine consumption inhibits ENaC function in distal nephron, which in turn increases urinary sodium excretion and prevents hypertension induced by high salt intake. To test this hypothesis, the Dahl's salt sensitive rat, a salt-sensitive hypertension model 25,26 , was examined in this study. We provide in vivo and in vitro evidence that AMPK activation by caffeine promoted urinary sodium excretion and lowered blood pressure through inhibition of the ENaC function in salt-sensitive rats. The present results also suggest that the acute and chronic actions of caffeine are mediated via different sites of the kidney tubules.

Results
Effect of chronic caffeine intake on blood pressure. Despite lower water and food intake in the first three days, chronic caffeine intake had no effect on water and food intake in Dahl-S rats thereafter (Fig. 1a,b), but significantly decreased their body weight compared with controls from the 12th day of treatment (Fig. 1c). Similarly, although caffeine intake temporally increased blood pressure on the second day, long-term caffeine intake attenuated high salt-induced increase in systolic, but not diastolic blood pressure in a time dependent manner (Fig. 1d,e). Also, 24-hr ambulatory systolic blood pressure on the 15th day was lower in caffeine-treated rats. This anti-hypertensive effect of chronic caffeine was more obvious at night when the rats were more active (Fig. 1f). 24 hr ambulatory diastolic blood pressure did not differ between the two groups ( Fig. 1g).
Effect of chronic caffeine intake on sympathetic nerve activity and vascular function. On the second day, caffeine initially increased heart rate and locomotor activity, after that, there were no significant difference in heart rate and locomotor activity between control and caffeine-treated Dahl-S rats (Fig. 2a,b). Accordingly, the plasma catecholamine concentration was not different between the two groups (Fig. 2c). In addition to sympathetic nerve activity, electrical field stimulation (EFS)-induced mesenteric artery constriction was not affected by caffeine (Fig. 2d). Similarly, both of endothelium-dependent and independent relaxations of mesenteric arteries were almost equal between the two groups ( Fig. 2e,f). These results indicate that neither sympathetic nerve activity nor vascular function is responsible for the anti-hypertensive effect of chronic caffeine intake.
Chronic caffeine intake increases urinary sodium excretion. To investigate whether chronic caffeine intake has a diuretic or natriuretic effect, we measured the 24-hr urinary volume and urinary sodium concentration of caffeine-treated rats and control rats. Interestingly, despite not altering 24-hr urinary volume (Fig. 3a), chronic caffeine intake increased the urinary sodium concentration and sodium excretion (UNaV) (Fig. 3b,c) without affecting plasma sodium concentration (Fig. 3d). Therefore, increase of urinary sodium excretion would mainly contribute to the anti-hypertensive effect of caffeine.
Chronic caffeine intake inhibits renal αENaC. We further examined which renal sodium transporter involved in the effects of caffeine. Long-term administration of caffeine decreased the α -ENaC protein expression in cortical collecting duct of Dahl-S rats, whereas amounts of β -and γ -subunit proteins of ENaC were unchanged. The protein level of the thiazide-sensitive NaCl cotransporter (NCC) in the distal tubule was also unchanged by chronic caffeine intake ( Fig. 4a and Supplementary Fig. 1). To establish the importance of ENaC in the effect of caffeine on urinary sodium excretion, Dahl-S rats were given an intraperitoneal injection of amiloride (3 mg/kg), a specific ENaC inhibitor. The ENaC-dependent sodium reabsorption was evaluated by the change in urinary sodium excretion after amiloride intervention. Caffeine-treated Dahl-S rats had a weaker ENaC-dependent sodium reabsorption than controls (Fig. 4b). In contrast, hydrochlorothiazide (12.5 mg/kg) inhibition of renal tubular transport of sodium by NCC was similar in control and caffeine-treated rats (Fig. 4c). Thus, the natriuretic effect of chronic caffeine intake was dependent on renal α -ENaC activity rather than NCC.
Caffeine treatment attenuates the ENaC function by activating AMPK. To investigate the mechanism underlying caffeine-induced increase in urinary sodium excretion mediated byα -ENaC inhibition, the expressions of kinase proteins were measured in cortical collecting duct cells (M1-CCD). Caffeine exposure for 24 hr decreased α -ENaC protein expression without affecting protein levels of SGK1, ERK1/2 and PKCα . An exception was α AMPK and its phosphorylated form, which was increased in response to caffeine ( Fig. 5a and Supplementary Fig. 2a). Thus we determined whether the effect of caffeine on α -ENaC subunit expression was dependent on AMPK. Caffeine exerted dose-dependent effects to increase α AMPK and phosph-AMPK protein levels in association with reduced α -ENaC. Furthermore, compound C, an AMPK inhibitor, abolished the effects of caffeine on α AMPK, phosph-AMPK and α -ENaC protein levels ( Fig. 5b and Supplementary Fig. 2b). In addition, electrophysiological assessment of ENaC activity in M1-CCD cells demonstrated that caffeine exposure for 24 hr decreased the open probability of ENaC (Fig. 5c). These data indicate that caffeine treatment reduces protein levels of the α -ENaC by a mechanism involving AMPK and decreases the open-probability of ENaC in the renal collecting duct that leads to reduced sodium reabsorption and increased sodium excretion.

Discussion
The present study demonstrates that chronic caffeine administration promotes urinary sodium excretion and reduces blood pressure in salt sensitive hypertensive rats. Furthermore, this effect of caffeine is independent of the sympathetic nerve activation and vascular function. The natriuretic action of caffeine is mediated by inhibition of (a-c) The water, food intake and body weight analyses after high salt (control) or high salt plus caffeine (caffeine) intervention (n = 8). (d,e) Systolic blood pressure and diastolic blood pressure analyses in control or caffeine groups (n = 9). (f,g) 24 hr systolic and diastolic blood pressures in conscious Dahl-S rats on the 15 th day of caffeine intervention (n = 9). All data are presented as means ± SEM. *P < 0.05 compared with control. renal α ENaC function without affecting the thiazide-sensitive NaCl cotransporter. In both in vitro and in vivo studies, caffeine up-regulated α AMPK level which suppressed αENaC expression and activity.
Caffeine is a major ingredient in a number of the most widely consumed non-alcoholic beverages 27 . Although the diuretic and natriuretic effects of caffeine have been recognized for long term, the mechanism responsible for caffeine effect was not fully elucidated. Thomsen et al. reported that acute caffeine intake increases lithium clearance that reflects sodium reabsorption in proximal tubule 28 . Rieg et al. showed that the acute diuresis and natriuresis produced by caffeine were related to blockade of adenosine A 1 receptors 29 . However, the chronic effects of caffeine on renal sodium handling are rarely investigated. The diuretic effect of caffeine is short-lived and is easy to develop tolerance 30 . In addition, no dehydration was observed in habitual coffee drinkers 31 . These findings suggest that other mechanism might be involved when chronic caffeine use. In the present study, we found that (e,f) Endothelium-dependent vasodilation induced by acetylcholine (Ach) was not significantly different in control and caffeine group (n = 5). Endotheliumindependent vasodilation induced by nitroglycerin (NTG) was also similar in control and caffeine group (n = 5). Results are expressed as a percentage of the maximum contraction elicited by 10 −5 M phenylephrine (PE). All data are presented as means ± SEM. *P < 0.05 compared with control. chronic caffeine intake dominantly increased 24-hr urinary sodium excretion but not urinary volume in Dahl-S rats compared with control rats. This natriuretic effect but not diuretic effect suggested a primary action on renal sodium handling in chronic caffeine intake.
In our study, we showed that the urinary sodium excretion increased after caffeine intervention without change in plasma sodium concentration. Similarly, Burge et al. reported that long-term diuretics (amiloride and hydrochlorothiazide) administration at therapeutic dose have no influence on plasma sodium concentration 32 . Since caffeine displays a much weaker natriuretic effect than diuretics 9 , chronic caffeine administration would not affect plasma osmolality.
It is well documented that thiazide-sensitive NaCl cotransporter (NCC), amiloride-sensitive epithelial sodium channel (ENaC), and sodium hydrogen (Na/H) exchanger are major regulators of renal tubular sodium reabsorption thus affect blood pressure regulation 18,19 . These epithelial sodium transporters can be regulated by several kinases such as protein kinase C, with-no-lysine kinase (WNK) 4 and 1, ERK 1/2, and serum-and glucocorticoid-inducible protein kinase 1 (SGK1), as well as AMPK 16,21 . Disturbances of signaling via these kinase pathways can result in human sodium retention and hypertension [33][34][35][36] . Previous studies in oocytes and epithelial tissues showed that AMPK inhibits sodium transport through increasing Nedd4-2-dependent ENaC retrieval from the membrane 37,38 . Some studies reported that caffeine can activate AMPK in different type of cells [22][23][24] . In the in vivo study, chronic caffeine intake decreased the protein levels and function of ENaC, but not NCC, which also affects renal electrolyte transport and blood pressure in the distal convoluted tubule 39 . Thus, caffeine appears to increase urinary sodium excretion by inhibiting renal ENaC activity secondary to the AMPK pathway. However, whether caffeine inhibits the open probability of ENaC by regulating its expression or ability still needs further investigation.
We also confirmed that acute caffeine intake induced an initial increase in blood pressure in association with a transient increase in locomotor activity and heart rate in Dahl-S rats, which is likely associated with sympathetic nerve activation. By contrast, these effects failed to be observed in long-term caffeine intake. Importantly, long-term administration of caffeine lowered blood pressure, which was not associated with sympathetic nerve activation and cardiovascular changes in Dahl-S rats. It suggested that chronic hypotensive effects of caffeine could be caused by its natriuretic effects in Dahl-S rats. Some studies have reported that systolic blood pressure, rather than diastolic blood pressure, are affected by urinary sodium excretion 40,41 . This may partially explain why caffeine failed to prevent high salt induced increase in diastolic blood pressure in our study. In addition, a recent study also showed that ambulatory systolic blood pressure was inversely correlated with urinary caffeine and its metabolites in adults from a general population, which indicated a potential protective effect of caffeine on blood pressure 42 . Thus, chronic caffeine intake might be an effective lifestyle intervention to promote salt excretion in people who normally consume a high salt diet. However, its application in human still warrants further determination in future.
In conclusion, we have demonstrated that chronic caffeine intake prevents salt-sensitive hypertension in Dahl-S rats. The beneficial effect of caffeine is associated with activation of renal AMPK that inhibits ENaC activity, which subsequently increases urinary sodium excretion and maintains blood pressure during high salt diet. However, chronic effect of caffeine on animal models needs to be further validated in human. These findings provide insight into the physiological role of caffeine in a long-term regulation of blood pressure through affecting renal sodium handling.

Materials and Methods
Animal Treatment. Six-week-old male Dahl salt-sensitive rats (Dahl-S) were obtained from Vital River Company, Beijing, China and housed under controlled temperature (21-23 °C) with a 12/12 hr light-dark cycle with free access to food and water. Animals were anesthetized by inhaling of 2% isoflurane (v/v), and surgically implanted with BP telemetric transmitters (TA11PA-C40, Data Sciences International, Minnesota, USA). After recovering from the surgery for 10 days, animals were randomly assigned to two groups: The control group was given a high salt chow containing 8% (w/w) NaCl and normal drinking water. The caffeine group was fed the 8% NaCl chow and 0.1% caffeine in their drinking water. The dietary intervention lasted 15 days. All of the experimental procedures were performed in accordance with protocols approved by Institutional Animal Care and Use Committee of the Institute of Third Military Medical University, and all experiments were performed in accordance with the National Institutes of Health guidelines for the use of experimental animals.
Blood pressure and locomotor activity measurement. 24-hour ambulatory blood pressures and locomotor activity were measured by radiotelemetry in conscious, unrestrained rats as previously described 43 . We collected data for 10 seconds every 30 minutes and used the mean values of 24 hours for the analysis.
Urinary Samples Analysis. On the 16th day, rats were transferred to individual metabolic cages (Tecniplast, Italy). Both groups maintained the same dietary intervention as mentioned before and had free access to food and water. The 24-hr water consumption and urinary excretion of water and sodium were measured. Urinary sodium concentrations were measured using a flame photometer (Spectrum, Shanghai China). Urinary sodium excretion (UNaV) was calculated by this formula: UNaV = [urine sodium concentration] × [24-hour urine volume].

Vascular Reactivity and Electrical Field Stimulation.
Vascular reactivity and electrical field stimulation were performed as previously described 44 . After rats were anesthetized with pentobarbital sodium (100 mg/kg body weight ip), the mesenteric vascular bed was removed and placed in a cold (4 °C) Krebs solution containing (mM): 118 NaCl, 25 NaHCO 3 , 11 D-glucose, 4.7 KCl, 1.2 KH 2 PO 4 , 1.17 MgSO 4 , and 2.5 CaCl 2 . The second branches of mesenteric arteries were dissected out and the connective tissue was removed. The arterial segments (2-2.5 mm in length) were mounted in a myograph. Vascular rings were bathed in Krebs solution aerated with 95% O 2 and 5% CO 2 at 37 °C (pH 7.4) and were stretched to the optimum baseline tension (2.5 mN). The rings were equilibrated for 60 min before the start of an experiment. High K + (60 mM)-containing Krebs solution was added to test contractility. Isometric contractions were recorded using a computerized data acquisition system (PowerLab/8SP; AD Instruments Pty Ltd., Castle Hill, Australia).
Electrical Field Stimulation (EFS) was achieved using a stimulator (Grass SD9) connected to two platinum electrodes placed on each side of the ring parallel to its longitudinal axis. The frequency-response curves (2-32 Hz, 30 V, 30 s trains and 1 ms duration) were obtained. There was an interval of 1 min between each stimulus to allow for recovery of basal tone. To evaluate the neural origin of the EFS-induced contractile response, the nerve impulse propagation blocker, tetrodotoxin (TTX, 0.1 mM), was added to the bath 30 min before the second frequency-response curve was determined.
Preparation of renal cortical collecting duct. Rats were anesthetized with sodium pentobarbital (100 mg/kg body weight ip), the circulatory system was perfused via the left ventricle with Eagle's minimal essential medium (MEM) containing collagenase (1 mg/ml), soybean trypsin inhibitor (2 μ g/ml), aprotinin (2 μ g/ml). After perfusion, kidneys were removed, cut into coronal slices, and incubated for 10-15 min at 37 °C in the same solution used for the perfusion. After collagenase washout, slices were kept in ice-cold MEM containing soybean trypsin inhibitor (2 μ g/ml) throughout the microdissection procedure. The renal cortical collecting duct (CCD) segments were separated manually using fine forceps and identified with characteristic branching indicative of CCDs. The pools of CCDs containing 10-20 microdissected tubules with the total tubular length of ~10 mm/pool were transferred in 5 μ l of Dulbecco's modified Eagle's medium (DMEM)/ Ham's F-12 (1:1) medium into 1.5 ml of Eppendorf tubes and refrigerated in − 70 °C for protein extraction 45 .
The amiloride and hydrochlorothiazide test. Another group of rats were used to evaluate the effect of caffeine on ENaC and NCC function at distal nephron. The procedure of fifteen-day's dietary intervention was the same as mentioned at Animal Treatment section. After that rats were transferred to metabolic cages. A 24-hr urine samples were collected in metabolic cages. Then the rats were administrated of amiloride (3 mg/kg, ip), an ENaC antagonist, or hydrochlorothiazide (12.5 mg/kg, ip), a NCC antagonist. After the amiloride or hydrochlorothiazide administration, another 24-hr urine samples were collected. Urinary sodium concentrations were measured using a flame photometer (Spectrum, Shanghai China).
Electrophysiology. For single channel recordings of ENaC, M1-CCD cells were studied under voltage-clamp conditions using standard methods as previously described. A patch pipette with the resistance of 6-8 MΩ was fabricated from a borosilicate glass capillary (1.5 mm od, 0.86 mm id, Sutter Instrument, Germany) on a Sutter Puller (P97, Sutter Instrument, Germany). The bath solution was (in mM): 110 NaCl, 4.5 KCl, 1 MgCl 2 , 1 CaCl 2 , 5 Hepes, 5 Na-Hepes (pH 7.2) wih a pipette solution of: 110 NaCl, 4.5 KCl, 0.1 EGTA, 5 Hepes, 5 Na-Hepes (pH 7.2). Single-channel currents were recorded using an EPC-10 patch-clamp amplifier (HEKA Instrument, Germany) and PATCHMASTER 8.0 software (HEKA Instruments, Germany). The data were acquired by application of 0 mV pipette potential and were sampled at 5 KHz and low-pass filtered at 1 KHz. During post hoc analysis, data were further filtered at 50 Hz and single-channel events were listed and analyzed by clamp fit 10.2 software (Molecular Devices, Sunnyvale, CA, USA). The total number of functional channels in a patch was determined by observing the number of peaks detected on the current amplitude histograms during at least 10-min recording period. NPo, the product of the number of channels and the open probability, or the open probability (Po) of ENaC before and after application of test agents was calculated using Clampfit 10.2 (Molecular Devices, Sunnyvale, CA, USA). In single-channel records, the control ENaC activity was recorded for 3-4 min after forming the cell-attached mode and ENaC activity had stabilized. We usually recorded at least 30 min of any experimental manipulation.
Statistical Analysis. All data are expressed as mean ± SEM. Statistical significant differences between two groups were evaluated by means of Mann-Whitney U tests. P values < 0.05 were considered to be statistically significant.