Pressure natriuresis and the renal control of arterial blood pressure

Abstract

The regulation of extracellular fluid volume by renal sodium excretion lies at the centre of blood pressure homeostasis. Renal perfusion pressure can directly regulate sodium reabsorption in the proximal tubule. This acute pressure natriuresis response is a uniquely powerful means of stabilizing long-term blood pressure around a set point. By logical extension, deviation from the set point can only be sustained if the pressure natriuresis mechanism is impaired, suggesting that hypertension is caused or sustained by a defect in the relationship between renal perfusion pressure and sodium excretion. Here we describe the role of pressure natriuresis in blood pressure control and outline the cascade of biophysical and paracrine events in the renal medulla that integrate the vascular and tubular response to altered perfusion pressure. Pressure natriuresis is impaired in hypertension and mechanistic insight into dysfunction comes from genetic analysis of blood pressure disorders. Transplantation studies in rats show that blood pressure is determined by the genotype of the kidney and Mendelian hypertension indicates that the distal nephron influences the overall natriuretic efficiency. These approaches and the outcomes of genome-wide-association studies broaden our view of blood pressure control, suggesting that renal sympathetic nerve activity and local inflammation can impair pressure natriuresis to cause hypertension. Understanding how these systems interact is necessary to tackle the global burden of hypertension.

Jessica Ivy is a British Heart Foundation PhD student at the University of Edinburgh. Her PhD combines in vivo approaches for cardiovascular and renal physiology with molecular techniques to explore the mechanisms of salt-sensitive hypertension. Matt Bailey is a renal physiologist with a PhD from the University of London. He had postdoctoral training at UCL, CNRS in Saclay, France and Yale University. He is a Reader at the University of Edinburgh and, in addition to research, directs the British Heart Foundation 4 year PhD programme at the Centre for Cardiovascular Science.

Hypertension is the leading risk factor for the global burden of disease (Lim et al. 2012). At the turn of the millennium, 1 billion people were estimated to have hypertension; by 2025 it is predicted that >1.5 billion people will be affected (Kearney et al. 2005), an increase outstripping the predicted population expansion. The overall prevalence shows no gender bias and hypertension is evident in all societies, affecting all races/ethnicities. Hypertension is the major modifiable risk factor for cardiovascular and renal disease and there is no cut off for risk, which doubles with every 20 mmHg increase in systolic blood pressure (SBP; Lewington et al. 2002). The prevalence of both diseases is increasing. One hundred years ago, cardiovascular disease was a minor cause of death, but now causes one-third of deaths worldwide (Lim et al. 2012). Similarly, chronic kidney disease (CKD) now affects 10% of the world's population (Eckardt et al. 2013). Clearly, hypertension presents a global health challenge and exerts a large societal and economic burden.

High blood pressure is considered the point at which the benefit of treatment outweighs that of inaction (Evans & Rose, 1971) and is currently defined as an average SBP of >140 mmHg and/or a diastolic blood pressure (DBP) of >90 mmHg. This quantitative threshold has reduced over the years and even those in the pre-hypertensive range (SBP 120–139 mmHg) have an increased risk profile (Toprak et al. 2009). The benefits of treating high blood pressure are strongly evidenced by clinical trial but at a population level, control rates are poor (Chobanian, 2009). Part of the challenge is identifying affected individuals since they often do not feel unwell and routine blood pressure (BP) screening in the clinic is an unreliable indicator of the steady state. Ambulatory BP monitoring can reveal ‘masked’ hypertension (Peacock et al. 2014) but is not routinely employed. Even when hypertension is confirmed, compliance to antihypertensive medication is poor and the gap in our knowledge means that the vast majority of hypertension is ‘essential’, of undefined causality. Research can bridge this knowledge gap but there is no silver bullet: physiological control of BP resides in complex, multi-system interactions, as indicated by the modes of actions of antihypertensive approaches developed since the 1950s (Table 1). Notably, these treatments all influence renal sodium handling, either directly via inhibition of tubular transport proteins, or indirectly via modulation of cardiovascular function and endocrine/neuroendocrine pathways. Thus, impaired renal function is likely to be an important driver for hypertension. Indeed, the relationship between renal perfusion and sodium excretion – the acute pressure natriuresis response (PN) – has long been considered vital for BP homeostasis. PN is abnormal in most, if not all, models of hypertension. In this article, we briefly review the role of PN in the long-term control of BP and discuss factors that regulate the renal response to altered perfusion pressure. We also examine the mechanistic insights into blood pressure gleaned from genetic analysis of experimental and human hypertension. These studies focus attention on vascular–tubular crosstalk in the renal medulla and suggest that tubulointerstitial inflammation in the medullary is an important early event, impairing PN directly or via increased renal sympathetic nerve activity (RSNA).

Table 1.

Time line of the development of major antihypertensive therapies

Year(s)	Treatment	Mechanism of action
1900	Thiocyanates	Reduction of cardiac output/adrenal suppression
1904	Dietary sodium restriction*	Reduction of extracellular volume
1923–1953	Sympathectomy	Sympatholytic
1931	Reserpine	Catecholamine depletion (sympatholytic)
1947–1950	Ganglion blockers	Sympatholytic
1949	Hydralazine	Direct vasodilatation (including renal arterial)
1958	Thiazide diuretics*	Major natuiresis and diuresis
1957	Aldosterone antagonist	RAAS suppression
1964	β-Receptor antagonists	Sympatholytic (peripheral/central)
1970	α2-Receptor agonists	Sympatholytic (central)
1975	α1-Receptor antagonists	Sympatholytic (peripheral)
1977	ACE inhibitors*	RAAS suppression
1977	Calcium channel blockers*	Direct vasodilatation. Increased renal perfusion, GFR, natuiresis and diuresis
1993	Angiotensin II receptor blockers*	RAAS suppression.
2000	Renin inhibitors*	RAAS suppression.
2005	Baroreceptor stimulator device	Sympatholytic.
2009	Catheter-based renal denervation	Sympatholytic
2013	ROX arteriovenous coupler	Reduction of arterial volume and peripheral vascular resistance

Most of these strategies, including the current primary treatment options for hypertension (^*), will either directly or indirectly affect renal function and improve the pressure natriuresis response. ACE, angiotensin converting enzyme; RAAS, renin–angiotensin–aldosterone system.

Pressure natriuresis and the long-term control of blood pressure

BP is the product of cardiac output and total peripheral resistance, with cardiac output itself the product of heart rate and stroke volume. From these simple equations it is easy to understand why cardiac or vascular theories dominated hypertension research for many years. This viewpoint was fundamentally challenged by a mathematical model published by Guyton, Coleman and Granger (Guyton et al. 1972), which generated the hypothesis that long-term, steady-state BP was influenced primarily by effective intravascular volume. Effective intravascular volume is influenced by vascular tone and also by extracellular fluid volume (ECFV), the latter determined by sodium balance. If BP rises, so too does renal arterial pressure (RAP) and the kidney responds by increasing sodium excretion and reducing ECFV; the converse applies if blood pressure falls. This renal response to changes in RAP is called the acute pressure natriuresis response (PN) and can be assessed experimentally as shown in Fig. 1A and B. In theoretical analysis, PN can offset completely any change in BP, regardless of origin. Set-point BP is therefore the point at which ECFV and PN are in equilibrium. It follows that a change in BP can only be sustained if the acute PN response is impaired, making hypertension a disease of the kidney. It is evident that the acute PN relationship is shifted to the right in experimental hypertension and the gradient of the slope can also be blunted (Mullins et al. 2006). If RAP is experimentally servo-controlled to prevent the kidney from sensing the increase in BP, the PN response is lost and hypertension is sustained/exaggerated (Hall et al. 1988). During the physiological adaptation to altered sodium intake, the acute PN curve is modulated by hormonal and neuronal inputs: the renin–angiotensin system (Wadei & Textor, 2012) and RSNA are both major regulators (Kubota et al. 1993). In response to high sodium intake, for example, angiotensin II is suppressed, the PN curve is shifted to the left and the gradient becomes steeper. This amplification of the PN curve permits a greater sodium excretion for any given BP, preventing expansion of ECFV. Both renal (down-regulation of epithelial sodium transporters) and extra-renal (vasodilatation and reduced aldosterone secretion) components are involved in the adaptation and if these hormonal and neuronal systems do not respond appropriately to dietary sodium intake, the ability of PN to regulate BP is impaired.

Figure 1. The relationship between blood pressure and sodium excretion.

A, the acute pressure natriuresis response can be induced experimentally by imposing serial pressure ramps upon the kidney using arterial constriction. B, the relationship between blood pressure, taken as a surrogate of renal perfusion pressure, and sodium excretion is flattened in experimental hypertension. C, a rise in blood flow through the vasa recta stimulates local production of paracrine agents such as nitric oxide (NO) and ATP, which can inhibit tubular sodium reabsorption at multiple sites. The rise in renal interstitial hydrostatic pressure (RHIP) reduces sodium reabsorption in the proximal tubule.

The mechanisms of the acute PN response are not fully understood. Natriuresis does not require increased glomerular filtration rate (Haas et al. 1986; Roman, 1988), rather proximal tubule sodium reabsorption is inhibited directly by a cascade of paracrine and biophysical events originating in the renal medulla (Fig. 1C). Medullary blood flow through the vasa recta network is poorly autoregulated and thus rises with increased RAP (Roman et al. 1988). Renal interstitial hydrostatic pressure (RIHP) also rises (Roman & Kaldunski, 1988; O'Connor & Cowley, 2010) and the increased pressure is transduced throughout the kidney due to renal encapsulation (Garcia-Estan & Roman, 1989). Sodium reabsorption in the proximal tubule is thus inhibited due to a change in the Starling forces across the epithelium (Schafer, 1990) and also by internalizing sodium transporters, including Na⁺–H⁺ exchanger isoform 3 (NHE3; McDonough, 2010). How the latter occurs is not defined but paracrine signalling may contribute: ATP, for example, is released by proximal tubule cells in response to stretch and inhibits NHE3 by activation of purinergic P2Y₁ receptors (Bailey, 2004).

As discussed below, other nephron sites may contribute to the overall natriuresis following increased RAP and the close anatomical relationship between the vasa recta and thick limb of Henle is perfectly arranged for paracrine signalling. Several agents, including ATP (Burnstock et al. 2014), 20-hydroxyeicosatetraenoic acid (20-HETE), reactive oxygen species (O'Connor & Cowley, 2010) and endothelin-1 (Garvin et al. 2011) are important modulators of the PN response. The powerful and pervasive effect of these factors is exemplified by nitric oxide, released as vasa recta blood flow increases (O'Connor & Cowley, 2010). NO augments the PN response by blunting the myogenic component of autoregulation (Dautzenberg et al. 2011) and inhibiting Na⁺–K⁺–Cl^– cotransporter (NKCC2)-mediated sodium transport in the loop of Henle.

The paracrine milieu of the kidney exerts a major influence on the acute PN response. Defining the regulation of these signalling systems and how they interact will shed light on the mechanisms of PN. It is unlikely that a purely physiological approach will unlock the secrets of this complex signalling environment. A complementary strategy is to probe the genetics of hypertension to identify important pathways for BP control.

The genetics of hypertension: BP follows the kidney

BP is a quantitative physiological trait with a normal distribution in the general population and ∼50% of that variability is genetically determined (Harrap et al. 2000). Inferentially, the ability of the kidney to efficiently regulate BP must also be genetically or epigenetically inherited and kidney transplantation studies in rat models of polygenic hypertension support this hypothesis. These studies (reviewed by Rettig & Grisk, 2005) repeatedly showed that hypertension could be normalized by grafting a control kidney into a bilaterally nephrectomized hypertensive rat. The converse was also true, leading to the concept that BP follows the kidney. In humans, essential (Curtis et al. 1983) and genetic hypertension (Botero-Velez et al. 1994; Khattab et al. 2014) can be resolved by kidney transplantation with a graft from a normotensive donor. An unambiguous mechanistic interpretation of these is challenging. The benefit for BP control may relate to the replacement of a kidney ravaged by exposure to hypertension with a ‘normal’ one, rather than alleviation from a genetically predetermined defect in the PN response. Against this argument, a normotensive rat will still develop hypertension even if the donor rat is young and pre-hypertensive (Kopf et al. 1993) or had long-term antihypertensive therapy to normalize BP prior to donation (Rettig et al. 1990). The hypertension generated in normotensive rats receiving a hypertensive kidney is associated with impaired natriuresis and a positive sodium balance but this occurs after the rise in BP, not before (Grisk et al. 2004). Clearly there is a complexity here that has not yet been resolved.

What is the heritable component that causes BP to follow the kidney? Brenner hypothesized that nephron endowment at birth was critical, and those with fewer nephrons have impaired PN and a genetic predisposition to hypertension (Brenner et al. 1988). Other studies suggest that epigenetic factors may affect the abundance of renal sodium transporters or the regulation of the PN response by angiotensin II (Dagan et al. 2010) or RSNA (Dagan et al. 2008) The effect on BP of receiving a hypertensive kidney is markedly attenuated if the donor rat was sympathectomized as a neonate (Grisk et al. 2002). Cross-transplantation studies in gene-targeted mice also suggest that the renal sensitivity to factors modulating PN is a key element to the development of hypertension. Angiotensin II infusion does not induce hypertension in angiotensin II receptor (AT1R)-null mice and reciprocal transplantation studies showed that the lack of BP response was mostly due to absence of the receptor in the kidney (Crowley et al. 2006). Recent studies have shown that deletion of AT1R in the proximal tubule alone is sufficient to protect against angiotensin II-induced hypertension. In the absence of this receptor, angiotensin II does not increase NHE3 expression and the PN response remains intact (Gurley et al. 2011).

The distal nephron influences overall natriuresis: evidence from monogenic hypertension

The acute PN response is mediated by inhibition of sodium transport in the proximal tubule but the overall natriuresis can be blunted if distal nephron sodium reabsorption is inappropriately activated (Ashek et al. 2012; Nguyen et al. 2013). Thus, distal nephron processes contribute importantly to long-term BP regulation, a concept underscored by the 15 Mendelian blood pressure disorders (8 causing high blood pressure, 7 causing low), which affect renal sodium handling in the loop of Henle, distal convoluted tubule and collecting duct. Detailed analysis of these disorders confirms a major renal contribution to abnormal BP and also highlights inputs from other systems. This complexity is illustrated by our own work on 11β-hydroxysteroid dehydrogenase type 2, an enzyme which deactivates glucocorticoids and confers aldosterone specificity on the mineralocorticoid receptor (Hunter et al. 2014).-null mutations in the encoding gene cause ‘apparent mineralocorticoid excess’, conventionally defined as a renal hypertensive disorder caused by unregulated sodium reabsorption in distal nephron. Indeed, isolated case studies report remission of the disease following kidney transplantation (Khattab et al. 2014) and the Hsd11b2-null mouse has an increased tubular sodium reabsorption due to enhanced bioactivity of the epithelial sodium channel (ENaC; Bailey et al. 2008). The acute PN response has not been measured in Hsd11b2-null mice but a chronic natriuresis develops as the disease progresses (Evans et al. 2012b). This increased sodium excretion contracts ECFV but does not normalize BP and our data indicate that activation of the sympathetic nervous system may contribute both to the origins and maintenance of hypertension (Bailey et al. 2008). The enzyme is also expressed in areas of the brain (Chapman et al. 2013) and genetic deletion here does not affect kidney function but induces both a salt-appetite and a salt-sensitive hypertension (Evans et al. 2012a). A similarly complex story emerges from other Mendelian BP genes, particularly ENaC. Deletion of the α-subunit in receptor cells of the tongue alters the taste perception for sodium chloride (Chandrashekar et al. 2010); deletion in the colon impairs reabsorption in the intestinal tract (Malsure et al. 2014). ENaC is also expressed in the vascular endothelium and activation, particularly in a high dietary salt setting, may promote endothelial dysfunction and vascular stiffness (Warnock et al. 2014).

Studies exploring genotype–phenotype relationships in rare conditions are extremely valuable and reshape our concepts of normal and abnormal BP control. Mechanistically, these disorders suggest that abnormal PN – or renal sodium handling – is at the centre of hypertension and highlight the importance of sodium homeostasis at the whole body level. Nevertheless, the physiological drive for sodium balance is an ancient one and it is informative that several of the affected proteins are expressed at sites which influence (i) our appetite for and perception of salt, (ii) our central processing of ingested sodium, and (iii) our ability to transport sodium across the gastrointestinal and renal epithelia.

Genome-wide association studies (GWASs) and new pathways for BP control

The large-scale GWASs examine an array of single nucleotide polymorphisms (SNPs) with coverage across the genome and use linkage disequilibrium to associate genotypic variants to observable phenotypes. This approach assumes that recent traits, such as hypertension, are more often than not associated with a given haplotype and a SNP is used as a surrogate for a functional trait. Nevertheless, ascribing function to a given genetic variant is very difficult and requires extensive physiological investigation to translate a genetic parts list to potential therapeutic utility.

The outcomes for BP from GWASs are ostensibly disappointing. The Wellcome Trust Case Control Consortium, for example, did not identify a significant association between BP and any of the assessed SNPs (Wellcome Trust Case Control Consortium, 2007). More recent studies have used much larger discovery groups to identify statistically significant associations. Unfortunately there is little consensus and the 29 loci thus far identified account for <1% of BP variability (Padmanabhan et al. 2012). These association studies are nevertheless intriguing since ascribing function to the identified variants may identify novel factors influencing renal sodium handling and BP control. For example, a SNP in UMOD, which encodes uromodulin, is associated with a reduced risk of hypertension (Padmanabhan et al. 2010) and recent data show that umod-null mice have a lower BP and an increased renal capacity for excretion salt (Graham et al. 2014). Uromodulin is the most abundant protein in the urine and may contribute to the renal inflammatory response to infection (Rampoldi et al. 2011).

It is increasingly recognized that hypertension has an inflammatory component and accumulation of macrophages and T-cells in the kidney, vasculature and brainstem is reported in experimental models (Rodriguez-Iturbe et al. 2014). GWASs may help unravel some of the contributing pathways and have identified CD247, a gene critical to T-cell activation, as a candidate for BP control (Ehret et al. 2009). Genetic disruption of CD247 attenuates hypertension and confers renoprotection in Dahl Salt-Sensitive rats (Rudemiller et al. 2014). Other studies point to a key role for the T-cell in the renal response to hypertension. Immune-deficient mice are protected from the pressor effects of chronic angiotensin II infusion (Crowley et al. 2010). The adoptive transfer of T-cells into these animals restores the full profile of angiotensin II hypertension (Guzik et al. 2007; Marvar et al. 2010). Figure 2 provides an overview of the current working hypothesis for the role of T-cells in hypertension. Infiltration of T-cells into the kidney correlates with the severity of hypertension, causes salt-sensitive hypertension and impairs the PN response (Franco et al. 2013). T-cells promote oxidative stress, release of inflammatory cytokines and local production of angiotensin II, all of which may suppress PN directly or, as discussed below, by increasing RSNA.

Figure 2. Simplified diagram of the proposed cycle of inflammation produced by the adaptive immune response in hypertension.

Stimuli such as angiotensin II (AngII), aldosterone or mechanical stress caused by high blood pressure cause the production of reactive oxygen species (ROS) and damage to renal and vascular tissues leading to the shedding of neoantigens. ROS also promotes the release of chemokines and adhesion molecules from damaged tissues. Neoantigens are then presented by antigen presenting cells (APCs), to T-cells within the thymus, which egress towards the sites of chemokine and adhesion molecule signalling where they proliferate and accumulate. The kidney injury, vasoconstriction and increase in sympathetic nervous system (SNS) outflow caused by the inflammatory milieu perpetuates the hypertension and creates a cycle of injury and increasing blood pressure. NOS, nitric oxide synthase.

It is likely that variants in the immune system confer some degree of hypertensive risk, part of which may be maladaptation of the PN response to high salt intake. In 1929, Cannon proposed that homeostasis was supported through storage by inundation, noting that sodium could be held, without water, in the skin to guard ECFV against ‘intolerable excess’ (Cannon, 1929). Recent studies provide molecular mechanisms for this storage: hypertonic storage of sodium chloride in the skin attracts macrophages and, via a tonicity-responsive enhancer binding protein–vascular endothelial growth factor-C (TONEBP–VEGF-C)-dependent pathway, promotes hyperplasia of the local lymphatic system to guard against hypertension (Machnik et al. 2009). If this local response is blocked, BP is elevated (Wiig et al. 2013). In this setting, the immune system plays a beneficial role for BP homeostasis, perhaps by controlling the amount of sodium that the kidney ‘sees’. It is also the case, however, that increased local concentration of sodium chloride activates pro-inflammatory T-helper cells (Kleinewietfeld et al. 2013) via a complex pathway involving TONEBP and serum and glucocorticoid-regulated kinase 1 (SGK1). T-helper cells can also launch an autoimmune response against neoantigens, proteins no longer recognized as ‘self’. Heat shock protein 70, expressed in the kidney, is one such antigen and induces a strong proliferation of blood lymphocytes in essential hypertensives (Pons et al. 2013).

Thus the contribution of the immune system to the normal/abnormal control of BP by the kidney is clearly complex but intrarenal inflammation will suppress the PN response (Fig. 3). This may provide novel therapeutic opportunities: mycophenolate mofetil, an immunosuppressant used to prevent transplant rejection, improves PN and lowers BP in hypertensive rats (Franco et al. 2013).

Figure 3. Efferent and afferent pathways of the renal sympathetic nervous.

Efferent preganglionic neurons originate from medullary centres such as rostral ventrolateral medulla (RVLM) and the nucleus of the solitary tract (NTS), where afferent inputs from end-organs (including the kidney) and baroreceptors are integrated. Post-ganglionic efferents synapsing at coeliac, supramesenteric and inferior mesenteric ganglia, innervate the renal tubules, juxtaglomerular apparatus (JG cells) and afferent and efferent arterioles. An increase in efferent activity evokes renin release, antinatuiresis and renovascular resistance. Renin release will stimulate the renin–angiotensin–aldosterone cascade, thus increasing renal tubular reabsorption of sodium but angiotensin II aided by circulating aldosterone can also have central nervous system effects (Blaustein et al. 2012; Biancardi et al. 2014). Inputs to the afferent sensory neurons of the kidney are relayed through the dorsal root ganglia (DRG) to higher brain centres such as the paraventricular nucleus (PVN), where they are integrated and the appropriate efferent output generated. For simplicity the renal efferent and afferent neurons are drawn separately, but note that the renal nerve is a compound nerve. The asterisk indicates regions of the sympathetic nervous system that have been manipulated for hypertension treatment (for a comprehensive review of renal sympathetic anatomy, see DiBona & Kopp, 1997; Johns, 2014).

Sympathetic over-activity, renal function and hypertension

That sympathetic over-activity contributes to human hypertension is well established (Fisher & Paton, 2012). It is increasingly evident that this may reflect an impairment of renal function since increased RSNA will alter renal haemodynamics and stimulate sodium transport in several nephron segments, including the distal convoluted tubule (Terker et al. 2014). Overall, increased RSNA shifts the PN curve to the right and impairs the ability of the kidney to buffer against a hypertensive insult. The origins of increased RNSA are many and varied but local inflammation in the brainstem, vasculature, or fat are important driving forces (Zubcevic et al. 2011).

The kidney itself may also be the driver for increased global sympathetic tone. It has long been known that sympathetic outflow is only reduced in kidney transplant patients if the diseased kidney is removed (McHugh et al. 1980). Conversely, β-blockers are only effective in transplant patients who retain their original kidneys (Huysmans et al. 1988). Since β-blockers can cross the blood–brain barrier the effect on BP may be centrally mediated (Gourine et al. 2008). Nevertheless, these findings suggest that the diseased kidney is the cause of the sympathetic overdrive. Finally experiments in rats with phenol-induced renal damage or renal insufficiency, show that dorsal rhizotomy to specifically transect afferent fibres abrogates hypertension (Campese & Kogosov, 1995; Ye et al. 2002). How the kidney becomes the source of increased afferent activity is not clear but a chain of events in which transient renal ischaemia (Faber & Brody, 1985) activates inflammation in the renal medulla/pelvis seems likely. The kidney is richly innervated by sympathetic neurons, which project to the tubules, vasculature and juxtaglomerular granular cells. Sensory afferent fibres are also present, particularly in the renal pelvis. These sense stretch and can, by diminution of sympathetic outflow to the contralateral kidney, induce natriuresis. Thus both afferent sensory neurons and efferent ‘effector’ neurons, exert integrated control over ECFV and hence BP (see Johns et al. 2011 for review). This renocentric view, illustrated in Fig. 3, does not include the multiple layers through which the brain and the kidney can ‘talk’ to each other. Activation of the systemic renin–angiotensin system, for example, will influence central cardiovascular control mechanisms. This may be particularly important in the hypertensive state when angiotensin II can extravasate into the brainstem (Biancardi et al. 2014). This is intrinsically pro-inflammatory; central interactions with aldosterone will also increase sympathetic outflow (Xue et al. 2011).

The interruption or resetting of RSNA is an attractive goal for hypertensive therapies. Pharmacological agents go some way to achieve this, with β-blockers, α-adrenoceptor antagonists and angiotensin receptor blockers (ARBs) all able to reduce sympatho-excitation. These agents interrupt peripheral sympathetic outflow to some degree but more complete or regional manipulations may be required for long-term resetting of the sympathetic nervous system. This concept is the spur for the development of bilateral renal denervation therapy (RDX) for hypertension. Ablation of the renal efferent and afferent neurons within the adventitia of the renal artery interrupts both the efferent signalling responsible for antinatriuresis and renin release and afferent output, which contributes to the setting of global sympathetic tone. In preclinical models RDX can evoke a large and sustained reduction in BP irrespective of the underlying cause of hypertension (Winternitz et al. 1980; Katholi et al. 1983; Dagan et al. 2008). Thus, RDX is very effective in obesity-induced or programmed hypertension, moderately effective in the spontaneously hypertensive rat (SHR) and the Goldblatt model but has little effect in the Dahl salt-sensitive rat. In such models there is a strong relationship between the reduction in BP and the reduction in renal catecholamine content, which suggests that increased RSNA is a common, but not universal, component of hypertension.

RDX has recently been used in humans and is attractive since the procedure is minimally invasive, involving the percutaneous insertion of a device into the renal artery, usually from the femoral artery, and can ostensibly provide a cure for elevated BP. RDX also offers a ‘plan B’ for patients with resistant hypertension and this was taken forward in the proof of concept Symplicity-HTN-1 and -2 open label clinical trials. The initial results were promising, with a ∼30 mmHg reduction in office BP being reported at 9–12 months (Krum et al. 2009; Schlaich et al. 2009; Esler et al. 2010), persisting to 36 months in the follow-up for Symplicity HTN-1 (Krum et al. 2014), suggesting that that functional re-innervation of the kidney was minimal. Nevertheless, the number of participants was small and patients continued with their antihypertensive medication throughout the trial. A Hawthorne effect on BP of improved adherence to antihypertensive medication cannot be excluded (Fadl Elmula et al. 2013).

Symplicity HTN-3 was the first large randomized, double-blinded, placebo-controlled trial using radiofrequency nerve ablation and incorporated ambulatory blood pressure (ABP) as a secondary end-point. The trial failed to reach its primary end-point (a 10 mmHg reduction in office BP) and was halted early in 2014 (Bhatt et al. 2014). This highlights the importance of properly controlled trials for soft variables like BP (Howard et al. 2013) and the main question now is the continued validity of renal denervation as a therapeutic concept. Data in animal studies suggest that RDX may be beneficial in some patients but not others. It is also observed that complete denervation in humans in remains a challenge (Hart et al. 2013). For now, however, the damage may have been done, with other clinical studies being halted early in the light of Symplicity-HTN3.

Nevertheless, device-based interventions remain an attractive strategy for BP control, particularly in resistant hypertension. Table 1 includes two alternatives to RDX, a device that introduced an arteriovenous anastomosis and another that electrically stimulates the baroreflex. Both devices show sustained reductions in BP in small clinical trials and both are currently being investigated in larger, multicentre studies. The hypotensive effect of these devices may not directly reflect an improved PN response. The stent developed by ROX Medical shunts a controlled volume of arterial blood into the venous system, whereas baroreflex pacing increases vagal tone and causes a global suppression of sympathetic drive (Donazzan et al. 2014). It is likely that the sustained improvement in BP is multifactorial, as recently reviewed for the ROX coupler (Burchell et al. 2014). As described above, any manoeuvre that suppresses RSNA would improve the ability of the kidney to excrete sodium and BP control.

Conclusions and future directions

This review has focused on the role of the pressure natriuresis mechanism in the long-term control of BP, a concept that emerged from the Guyton–Coleman–Granger model of long-term cardiovascular control. This model, now incorporated into a larger-scale simulation of human physiology (Hester et al. 2011), leads to the conclusion that hypertension must reflect underlying renal dysfunction. This hypothesis has exerted a profound and lasting influence on the research landscape. It is clear that the PN response is attenuated in hypertension but is this relationship causal? This question is difficult to answer experimentally but a recent computational approach, based on the novel assumption that sodium excretion is independent of BP, predicts that angiotensin II-dependent hypertension can evolve without an impaired PN mechanism, reflecting instead increased sympathetic drive to the peripheral vasculature (Averina et al. 2012). The underlying assumption of this model has been challenged (Judge & Dorrington, 2013), yet the importance of in silico approaches, which stimulate debate and advance understanding of complex physiological systems, is undeniable.

It is clear that BP homeostasis is intimately associated with sodium homeostasis and the distribution of sodium between fluid compartments and within tissues. The kidneys provide the principal route for sodium excretion, even if storage within the skin/tissues determines how much sodium they see. Renal sodium handling therefore underpins the long-term stability of BP and the concept that hypertension is caused or sustained by an impaired acute PN response retains contemporary relevance: BP homeostasis is sacrificed to maintain sodium balance and this may reflect an early stage of hypertension that is not readily detectable in the clinic. Indeed, the daytime ‘office’ BP may be normal but if BP doesn't dip in the night, this may indicate an impaired PN response (Bankir et al. 2008) strongly associated with cardiovascular risk (Sega et al. 2005). Tackling this by adjusting the timing of antihypertensive mediation is an effective means of BP control (Hermida et al. 2014) and future physiological studies must consider the operation of the acute PN response within a 24 h time frame to provide a more sophisticated view of BP control.

Glossary

BP

blood pressure

CKD

chronic kidney disease

ECFV

extracellular fluid volume

ENaC

epithelial sodium channel

DBP

diastolic blood pressure

GWAS

genome-wide association study

NHE3

Na⁺–H⁺ exchanger isoform 3

PN

acute pressure natriuresis response

RAP

renal artery pressure

RDX

renal denervation

RHIP

renal interstitial hydrostatic pressure

RSNA

renal sympathetic nerve activity

SBP

systolic blood pressure

SNP

single nucleotide polymorphism

Additional information

Competing interests

None declared.

Funding

J.R.I. is supported by a British Heart Foundation 4 year PhD studentship. M.A.B. thanks the British Heart Foundation Centre of Research Excellence, Kidney Research UK, NC3Rs and the Diabetes and Wellness Foundation for research funding.