The acute pressure natriuresis response is suppressed by selective ETA receptor blockade

Abstract Hypertension is a major risk factor for cardiovascular disease. In a significant minority of people, it develops when salt intake is increased (salt-sensitivity). It is not clear whether this represents impaired vascular function or disruption to the relationship between blood pressure (BP) and renal salt-handling (pressure natriuresis, PN). Endothelin-1 (ET-1) regulates BP via ETA and ETB receptor subtypes. Blockade of ETA receptors reduces BP but promotes sodium retention by an unknown mechanism. ETB blockade increases both BP and sodium retention. We hypothesized that ETA blockade promotes sodium and water retention by suppressing PN. We also investigated whether suppression of PN might reflect off-target ETB blockade. Acute PN was induced by sequential arterial ligation in male Sprague Dawley rats. Intravenous atrasentan (ETA antagonist, 5 mg/kg) halved the normal increase in medullary perfusion and reduced sodium and water excretion by >60%. This was not due to off-target ETB blockade because intravenous A-192621 (ETB antagonist, 10 mg/kg) increased natriuresis by 50% without modifying medullary perfusion. In a separate experiment in salt-loaded rats monitored by radiotelemetry, oral atrasentan reduced systolic and diastolic BP by ∼10 mmHg, but additional oral A-192621 reversed these effects. Endogenous ETA stimulation has natriuretic effects mediated by renal vascular dilation while endogenous ETB stimulation in the kidney has antinatriuretic effects via renal tubular mechanisms. Pharmacological manipulation of vascular function with ET antagonists modifies the BP set-point, but even highly selective ETA antagonists attenuate PN, which may be associated with salt and water retention.


Introduction
Hypertension is the single biggest contributor to global disease burden and remains a major risk factor for cardiovascular disease despite current multimodal therapeutic strategies [1,2]. Blood pressure (BP) rises with salt intake and this effect is exaggerated in a significant minority of people [3]. The mechanisms of such salt-sensitivity remain contentious. Since salt intake commonly far exceeds recommended upper limits [4], there is an unmet need to further understand and target the disruption to normal sodium handling that leads to salt-sensitivity.
The computational model of Guyton and others places the kidneys as the central regulator of long-term BP, achieved through relative homeostasis of extracellular fluid. Thus, the expansion of intravascular volume, after sodium and water ingestion for example, causes renal arterial pressure to rise, which in turn induces sodium excretion [5]. The pressure natriuresis (PN) relationship can be measured empirically and is attenuated in settings of experimental hypertension. The Guyton model has influenced hypertension research for more than 50 years but is challenged by an alternative hypothesis that salt-sensitivity is a manifestation of impaired vasodilatation rather than inadequate sodium and water excretion [6]. This is based on observations that dilatation of the systemic vasculature normally accommodates salt-induced

Renal function studies
The original protocol for PN [21] was adapted, as described previously [22]. It was applied to adult rats under non-recovery anaesthesia induced by intraperitoneal thiopental; (50 mg/kg, 50 mg/ml; Archimedes Pharma, Reading, U.K.) and maintained by intravenous thiopental of the same concentration. Rats received an intravenous (jugular) infusion of physiological saline (pH 7.4; 1 ml/hour/100 gbw) containing 2% (weight:volume) bovine serum albumin to maintain euvolaemia. 0.25% fluorescein isothiocyanate (FITC)-inulin (Sigma-Aldrich Company Ltd, Gillingham, U.K.) was added to the infusate to measure glomerular filtration rate (GFR) by inulin clearance. An arterial line (carotid) was used for blood sampling and real-time BP measurement via a calibrated transducer and multi-channel data acquisition system (Powerlab; ADInstruments, Oxford, U.K.).
Following laparotomy, ligatures were pre-placed around the coeliac and mesenteric arteries, and the aorta distal to the left kidney. Left renal artery blood flow (ml/min) was measured by a calibrated Doppler ultrasound probe (PR-probe; Transonic, Ithaca, U.S.A.) positioned around the renal artery stripped of periarterial fat and nerves [21]. Laser Doppler spectroscopy measured regional changes in renal blood flow [22,23]. A laser Doppler patch probe (MSP100XP; ADInstruments) was glued to the dorsal surface of the left kidney to measure cortical flux, and medullary flux was measured by a needle probe (MNP110XP; ADInstruments) inserted through the capsule to a depth of ∼5 mm and orientated toward the hilus. Urine was collected from the left ureter.
After 30 min for equilibration, urine collections were made over three sequential periods, each lasting 30 min. After the first (baseline) collection, BP was acutely increased by ligation of the cranial mesenteric and coeliac arteries (ligation 1), and urine was collected. The distal aorta was then ligated (ligation 2), after which the final urine collection was made. The entire procedure was performed under homeostatic temperature control at 37 • C.
Urinary sodium concentrations were measured by a sodium-selective electrode (9180 Electrolyte Analyser; Roche Diagnostics Ltd., Burgess Hill, U.K.). Mean renal blood flow values were calculated, along with urine flow rate (UV), urinary sodium excretion rate (UNaV) and fractional excretion of sodium (FENa). GFR was calculated from the

Continuous measurement of BP in conscious rats
Radiotelemetry devices (TA11-CA P40; Data Sciences International, Hertogenbosch, Netherlands) were pre-calibrated and implanted into the abdominal aorta of male Sprague Dawley rats under anaesthesia with inhalational isoflurane (Isoflo; Zoetis Animal Health Ltd., Sandwich, U.K.) using aseptic technique [22]. Recovery from surgery was aided by warm airflow and analgesia with 0.5mg/kg s.c. buprenorphine (Buprecare; Animal Care UK, York, U.K.) every 12 h for 3 days.
The start of every day was defined as the start of the dark period at local time 7 p.m. Baseline systolic BP (SBP), diastolic BP (DBP) and heart rate data prior to salt supplementation have been published [22] and are not presented here. Radiotelemetry recording was then extended in the same rats for three consecutive weeks to limit the number of animals used, and we present these data. During the first of these weeks, rats received additional dietary salt after the 6 p.m. but before the 7 p.m. data collection, in order to determine the BP response to salt-loading. The salt was in the form of an ice-cube (∼10 ml)-sized gelatin (Dr. Oetker UK Ltd, Leyland, U.K.) block containing soluble meat extract (80 mg sodium chloride/ml; 410 mg beef extract/ml; Bovril; Unilever UK Ltd, Leatherhead, U.K.) so that, for every rat, daily sodium intake was 140 mg of sodium/day based on 15 g/day daily intake of pelleted food (Benevenga et al., 1995). This equated to a 1% sodium diet, four times the concentration of sodium in the unsupplemented diet. One gelatin block was placed in every rat's cage on top of a cardboard tube used for environmental enrichment. Voluntary, rapid consumption of the block was visualized in every case.
To determine whether the BP response to salt-loading was modified by ET receptor antagonists, the gelatin block was supplemented [25] with atrasentan (5 mg/kg) during the second week, and then, during the third week, atrasentan and A-192621 (10 mg/kg). Data presented here were acquired during the middle five days of each week (1 kHz over a 1-min period in every hour). Data from the first and seventh days were excluded to remove confounding stress effects associated with weekly staff changes and routine husbandry.
At the end of the experiment, rats underwent euthanasia by carbon dioxide asphyxiation followed by cervical dislocation.
Numbers of rats varied between groups according to experimental losses, and control groups contained larger numbers to ensure they remained contemporary over the duration of the study (∼18 months). PN studies were designed to obtain a power >80% if group sizes were six rats, UNaV in rats receiving atrasentan was 50 + − 25% [26] of the expected value in controls [22], and medullary blood flow was 20 + − 5% greater in controls over atrasentan-treated rats [27]. Additional rats were included in every group to account for an expected dropout rate of 25% due to experimental mortality.
Rats receiving vehicle formed the control group for each set of statistical comparisons. Non-normal data (Anderson-Darling test) were compared following log or square-root transformation to generate a Gaussian-distributed data set amenable to parametric analysis, or, where this was not possible, non-parametric testing.
For acute PN studies, baseline mean BP (MBP) was the independent variable used as a surrogate marker of renal perfusion pressure. After confirming that baseline MBP and increments for all groups were similar to control, data were compared with control with two-way repeated measures analysis of variance (ANOVA). The results are presented as the effect on dependent variables from ligation (all rats), ET receptor antagonist and the interaction between the two of them. Where there was an effect from an ET receptor antagonist, Dunnett's post hoc tests were applied to determine whether this effect was observed during individual clearance periods. As additional analysis, dependent variables were plotted against MBP and compared using regression analysis. Linear regression, including polynomial and reciprocal terms, was used first to model curves, and if an adequate fit could not be obtained, non-linear regression was attempted. Regression lines were compared by analysis of covariance (ANCOVA) with Tukey post hoc tests (linear) and extra sum of squares F-tests (nonlinear), to determine whether one regression line fitted all data sets.
For radiotelemetry, BP and heart rate were compared with the values obtained during the previous recording period, using one-way repeated measures ANOVA, with rat identification as a random factor, and Sidak's post hoc tests. within an autoregulatory range (horizontal dotted lines) previously described for Sprague Dawley rats during experimental PN [22]. Weights are shown. There was no difference in weight between groups (P=0.122). (C) Mean BP during individual clearance periods. Arterial ligation increased mean BP (P<0.001) above the mean BP of the previous clearance period (ligation 1, P<0.001; ligation 2, P=0.003). Mean BP and ramps in mean BP were similar between groups during every clearance period (interaction P=0.347). (D) GFR during individual clearance periods. Arterial ligation increased GFR (P=0.004) after ligation 1 (P=0.007) but not ligation 2 (P=0.977). GFR did not differ between groups during every clearance period (interaction P=0.127). Most values remained within an autoregulatory range (horizontal dotted lines) previously described for Sprague Dawley rats during experimental PN [22]. Every rat contributed three data points into every panel (baseline and after ligations 1 and 2). In panels (A and B), they are combined. In panels (C and D), there is one data point per rat per time-point. Bars show mean; vertical dashed lines divide clearance periods; *P<0.05 compared with vehicle or previous clearance period. All comparisons were made with two-way repeated measures analysis of variance (ANOVA) with Dunnett's post hoc tests.
For all ANOVAs, data were transformed where necessary to generate normality and equality of variance in residuals. For all tests, statistical significance was set at P<0.05.

Effects of ET receptor antagonists on renal haemodynamics
MBP (interaction P=0.347) and GFR (interaction P=0.127) were similar across all groups during individual clearance periods ( Figure 1C,D). Ramps in MBP increased medullary flux (P=0.042) but did not affect cortical flux (P=0.379) or renal artery flow (P=0.701).
Atrasentan suppressed medullary flux during PN (P=0.001, Figure 4A). It increased by only ∼20% from baseline, reaching only half of vehicle levels after ligation 2 (P=0.002; Figure 4D). The regression curve of proportionate increases in medullary flux and MBP that was obtained with vehicle was lost with atrasentan ( Figure 5A). Renal artery flow was decreased (P=0.014; Figure 4B) but was no different to vehicle during individual clearance periods ( Figure  4E). Cortical flux was unaffected (P=0.378; Figures 4C,F and 5C).
A-192621, however, did not suppress the two-fold increase in medullary flux observed with vehicle (P=0.819; Figure 4A) although the curvilinear relationship of medullary flux with MBP was lost ( Figure 5D). Renal artery flow was decreased (P=0.003; Figure 4B) by around one third at baseline (P=0.012) and after ligation 1 by ∼4 ml/min, P=0.029; Figure 4E), but there was no influence on cortical flux, which remained close to baseline levels ( Figure  4C,F). With combined ET receptor blockade, suppression of the increase in medullary flux was, again, similar to atrasentan (P=0.002; Supplementary Figure S4A,D). However, unlike all the other groups, cortical flux increased overall to ∼25% greater than vehicle (P=0.001; Supplementary Figure S4C). This was most marked after the second ligation (P=0.003; Supplementary Figure S4F).

Effects of salt and ET receptor antagonists on BP
Salt increased SBP and DBP by ∼3-5 mmHg, and heart rate decreased by ∼15 beats/min ( Figure 6A-C).

Discussion
The major finding of our study was that administration of the ET A receptor antagonist, atrasentan, severely suppressed PN but nevertheless reversed the salt-induced increase in systolic and diastolic BP. In clinical trials, selective ET A receptor antagonists also lower BP but consistently cause sodium and water retention [19]. Our study provides  mechanistic insight, indicating that fluid retention may reflect attenuated PN due to vascular ET A blockade rather than inappropriate sodium reabsorption from off-target blockade of ET B receptors in the distal renal tubule.

. Regression analysis of renal perfusion responses following induction of experimental pressure natriuresis (PN) in Sprague Dawley rats after endothelin A (ET A ) and ET B , receptor antagonism (A) Medullary flux. A curve could not be fitted to the dataset from ET
The acute PN relationship can be determined experimentally by sequentially increasing BP (and thus renal perfusion pressure), which substantially increases blood flow through the renal medulla and induces a rapid, marked natriuresis [21]. Intact glomerular autoregulation limits the rise in GFR and filtered sodium load [27], so natriuresis mostly reflects decreased renal tubular sodium transport in response to augmented intrarenal hydrostatic pressure and pressure-mediated release of autocrine and paracrine agents. Therefore, suppression of PN can occur if there is a decrease in medullary blood flow or the signaling that it activates, or through a direct effect on sodium transporter activity itself. Here, we demonstrated that suppression of PN by atrasentan was linked with suppression of the normal increase in medullary blood flow.
We considered that this might reflect a systemic vasodilatatory response to atrasentan that reduced the hemodynamic load on the kidneys since, overall, MBP and renal artery flow were decreased in the atrasentan group. However, during individual clearance periods, we did not find that atrasentan reduced renal artery flow, and the small number of GFR data points that were above and below the autoregulatory range were evenly distributed across the vehicle and ET receptor antagonist-treated groups. This suggests that the reduction in circulatory capacity following ligation of major arteries overwhelmed any vasodilatatory response to atrasentan. In addition, atrasentan suppressed diuretic and natriuretic responses at BPs, renal artery flow rates and GFRs similar to those in vehicle-treated rats, and, when it was combined with A-192621 to block both ET A and ET B receptors, medullary blood flow and natriuresis were suppressed without any reduction in renal artery flow.
We also considered that vasodilatation with atrasentan might promote sodium and water retention through increased renal sympathetic activity, activation of the renin-angiotensin-aldosterone system (RAAS), or release of vasopressin [28]. Although we did not adopt alpha blockade because of potential confounding systemic effects on BP and RAAS that occur even when administered intrarenally [29,30], renal sympathetic activity should have been minimal following stripping of the renal artery. Hormonal clamping was not employed because circulatory levels of hormones above physiological ranges [21] could have additional confounding effects, and because it is difficult to determine which hormones should be included and which should not [28]. We have previously shown that our protocol for inducing PN does not modify urinary aldosterone excretion [22], a marker of RAAS activity, and angiotensin II-mediated sodium and water retention would be expected to decrease rather than increase following ET A receptor blockade [10]. Plasma vasopressin levels were not measured. Vasopressin is released during ET A blockade following overstimulation of unblocked endothelial ET B receptors [28]. However, similar suppression of medullary blood flow and natriuresis were observed during combined ET receptor blockade. Therefore, we do not believe that the small overall reduction in renal artery flow, increased renal sympathetic activity, RAAS activation or vasopressin release explain the magnitude of the suppression of PN observed with atrasentan. Instead, we conclude that atrasentan decreased sensitivity to the BP signal, and decreased natriuresis by interfering with ET A -receptor mediated regulation of intra-renal blood flow, and, subsequently, decreased tubular sodium transport.
By contrast, and somewhat surprisingly, blockade of ET B receptors with A-192621, enhanced the natriuretic response, despite a slight decrease in renal artery flow, and no influence on the normal increase in renal medullary blood flow that initiates natriuresis. This not only suggested that A-192621 was augmenting the inhibitory influence that increased medullary blood flow exerts on tubular sodium transport but also that the suppression of PN by atrasentan was not due to reduced renal artery flow or off-target ET B receptor blockade. We conclude that ET B receptors inhibit natriuresis when BP is acutely increased.
Pro-natriuretic ET A receptors and anti-natriuretic ET B receptors are at odds with their relative putative roles [10] and clinical applications of ET receptor blockade. Potent, long-lasting ET A receptor-mediated vasoconstriction and inhibition of natriuresis [31] is generally regarded as deleterious in the context of BP regulation [32], while renal tubular ET B receptors promote natriuresis at lower BPs than those reached during experimental PN, possibly through a combination of ENaC inhibition and renal medullary vasodilatation that is enhanced by high salt intake [12,33]. However, two previous studies have also demonstrated anti-natriuretic responses to ET A receptor blockade and pro-natriuretic effects during ET B receptor blockade, although mechanisms were not identified. In those studies, rats were sodium-loaded by i.v. [26] or intra-renal [24] routes, but renal blood flow was not measured. Our study proposes physiological mechanisms to explain these discrepancies: namely that if BP or sodium intake is sufficiently increased, stimulation of vascular ET A receptors promotes sodium excretion by increasing medullary perfusion and initiating PN. ET B receptors, however, inhibit the natriuretic response by promoting tubular sodium reabsorption, either by interfering with pro-natriuretic signaling initiated by increased medullary perfusion or by directly promoting sodium transporter activity. There is already evidence that ET B receptors within the proximal tubule increase sodium transport through sodium-potassium ATPase, the sodium-hydrogen anti-porter (NHE3), and the sodium-glucose transporter 2 (SGLT2) [34][35][36]. However, ET A receptors that mediate constriction of the medullary vasa recta by pericytes [37] would reduce rather than increase medullary perfusion. Therefore, we speculate that ET A receptors upstream of the vasa recta increase medullary perfusion and are only activated when BP increases acutely. ET-1 is known to constrict cortical and juxtamedullary afferent and efferent arterioles in vitro via both ET receptor subtypes [38,39], and intravital microscopy during infusion with ET-1 and the selective ET A receptor antagonist, BQ123, suggests a greater vasoconstrictive response in the afferent arterioles supplying cortical glomeruli [40]. This could promote diversion of flow through juxtamedullary glomeruli and into medullary vasa recta during ET A receptor activation. Support for such a mechanism comes from the sodium and water retention observed in knockout mice that lack vascular smooth muscle ET A receptor signalling [20]. A pro-natriuretic role for the small number of ET A receptors present within the collecting duct has also been proposed [41]. However, our data suggest that loss of ET A receptor-mediated PN can overwhelm the loss of anti-natriuretic effects mediated by renal ET A receptors [31,41,42] and might contribute to the sodium and water retention observed clinically with ET A receptor antagonists [43]. This would explain why this adverse effect is observed even with highly selective ET A antagonists and is unlikely to be eliminated by further increasing ET A receptor selectivity.
Using radiotelemetry, the gold-standard approach to measuring BP in rodents [44], we found that a high salt intake over several days increased BP, confirming a previous study that identified salt-sensitivity in Sprague Dawley rats [45]. We confirmed the BP lowering effect of atrasentan, consistent with vasodilatation, previously described in water and salt-loaded rats [31]. We also confirmed the increase in BP from additional A-192621 that is consistent with the putative vasodilatory and pro-natriuretic roles of ET B receptors. An important outcome of our study is that these effects on BP, recorded over days, were in opposition to those predicted by the Guyton hypothesis and the PN data, recorded over hours. This is directly relevant to the current debate surrounding the relative contributions that PN and vasodilatation make to BP regulation [6]. It suggests that the Guyton hypothesis does not predict the effect of increased dietary salt on BP in the face of salt-sensitivity, and is consistent with recently published data demonstrating an enhanced rather than attenuated PN response in salt-sensitive C57BL/6J mice [46]. Instead, the picture that emerges from our study is that the BP set-point is established by a systemic vascular response that can be manipulated by dietary salt and ET receptor blockade.
The Guyton hypothesis may still have clinical relevance. Pure vasodilator agents, such as dihydropyridine calcium channel antagonists, elicit an increase in heart rate in response to a reduction in BP [47]. In our radiotelemetry study, when atrasentan reduced BP there was no change in heart rate, which would be consistent with an increase in cardiac preload from impaired PN, in opposition to the new BP set-point. Thus, the Guyton hypothesis and PN may provide insight into the systemic hemodynamic load that target organs experience rather than BP, per se. Specifically within the context of therapeutic ET A receptor blockade, enhanced vasodilatation may be achieved at the cost of disrupted PN and the inability to efficiently excrete acute sodium and fluid loads. This would explain the apparent paradox of reduced BP but also sodium and water retention observed clinically with ET A receptor antagonists [19].

Limitations
A limitation of our work is that the experimental studies were performed exclusively in healthy, male rats. Although ET A receptor blockade also reduces sodium excretion in salt-challenged female rats [48], the contribution that ET-1-mediated changes in medullary blood flow make to PN may be less in females than males [49]. In addition, the effect of ET receptor blockade on BP and PN may be different in disease states and following chronic salt and ET antagonist administration. Therefore, the effect of atrasentan on medullary blood flow during experimental PN in female rats, in both sexes in disease states, and after chronic salt-loading and ET receptor blockade should be the subject of further study.
Our conclusions are based on the assumption that in vitro receptor affinity and selectivity are maintained in vivo. We mitigated for loss of receptor selectivity in the PN and radiotelemetry studies by including combined ET receptor blockade groups.
The potential implications for cardiovascular risk from our data remain to be resolved. The recent SONAR study [50] achieved reductions in cardiovascular risk associated with decreasing proteinuria, suggesting that benefits from blocking ET A receptor-mediated glomerular inflammation and fibrosis in diabetic CKD [51,52] outweigh any deleterious effects on sodium excretion that may occur in the clinical setting. Nevertheless, our study supports the use of ambulatory BP measurement and measurement of sodium intake and excretion to allow detailed analysis of daily fluctuations in BP and sodium balance when determining the impact of pharmacological interventions on cardiovascular risk.
In summary, regulation of BP by ET-1 is complex. It includes opposing and non-putative roles of both ET receptor subtypes at the levels of the renal and non-renal vasculature, and sodium transport. The influence that ET-1 has on BP is dynamic, according to salt intake and the level of BP. We propose that while blockade of ET A receptors may reduce BP, highly selective ET A receptor antagonists may also interfere with renal sodium handling, which might lead to adverse effects of salt and water retention that would offset any beneficial effects on cardiovascular risk.

Clinical perspectives
• ET A receptor antagonists have considerable clinical potential in resistant hypertension and chronic kidney disease because they reduce BP and proteinuria but patients also develop sodium and water retention by an unknown mechanism.
• Although ET A blockade decreased BP in salt-fed rats, and ET B blockade reversed these effects, novel and opposing effects of ET A (pro-natriuretic) and ET B (anti-natriuretic) receptors were identified during experimental PN.
• Suppression of PN may explain the sodium and water retention observed clinically with ET A receptor antagonists, despite reductions in BP. Increasing ET A receptor selectivity of antagonists may further decrease BP but would not eliminate sodium and water retention.

Data Availability
The data that support the findings of this study are available from the corresponding author upon reasonable request.