Physiological Mechanisms of Dietary Salt Sensing in the Brain, Kidney, and Gastrointestinal Tract

Abstract

Excess dietary salt (NaCl) intake is strongly correlated with cardiovascular disease and is a major contributing factor to the pathogenesis of hypertension. Salt-sensitive hypertension is a multi-system disorder that involves renal dysfunction, vascular abnormalities, and neurogenically-mediated increases in peripheral resistance. Despite a major research focus on organ systems and these effector mechanisms causing salt-induced increases in arterial blood pressure, relatively less research has been directed at elucidating how NaCl is sensed by various tissues to elicit these downstream effects. The purpose of this review is to discuss how the brain, kidney and gastrointestinal tract sense NaCl including key cell types, the role of NaCl versus osmolality, and the underlying molecular and electrochemical mechanisms.

Graphical Abstract

INTRODUCTION

Excess dietary salt intake is strongly correlated with cardiovascular disease and is regarded as a major contributing factor to the pathogenesis of hypertension ^1,2. Meta-analyses and systemic reviews of randomized clinical trials report a strong positive relation between Na⁺ intake and systolic arterial blood pressure (ABP) ^1,2. Importantly, NaCl or Na⁺ restriction leads to a decline in systolic ABP ^1,2. For example, a meta-analysis of randomized clinical trials reported a 2,300 mg reduction in 24-hr urinary Na⁺ excretion is associated with a decline in systolic ABP of 5.8 mmHg (after adjustment for age, ethnic group, and ABP status)³. These reductions in ABP due to NaCl-restriction are generally larger in older adults, African Americans, and hypertensive adults ^4–6.

The hypertension field has focused on organ systems and effector mechanisms causing salt-induced increases in ABP or how such systems respond to a high salt diet (vascular, renal, neurogenic), whereas relatively less research has been directed at how specialized cells may directly NaCl concentrations to elicit activation of these downstream effectors. The purpose of this review is to discuss the underlying cellular and biochemical mechanisms of NaCl-sensing within the brain, gastrointestinal tract and kidney. The important distinction is to discuss how cells detect changes in NaCl versus how cells respond to chronic high NaCl feeding. The roles of skin and immune cells in salt-sensing and homeostasis will be discussed elsewhere as part of this series.

CENTRAL NERVOUS SYSTEM

A high NaCl diet raises plasma or cerebrospinal fluid (CSF) NaCl concentrations by 2–6mM in salt-sensitive human cohorts and many experimental models of salt-sensitive hypertension including the Dahl-salt-sensitive rat, Spontaneously Hypertensive rat, Grollman renal wrap rat, and deoxycorticosterone-salt rat ^7,8. These small changes in extracellular NaCl concentrations are physiological meaningful as systemic or intracerebroventricular infusion of hypertonic NaCl to raise Na+ concentrations by 5–10mM increases sympathetic nerve activity (SNA) and ABP ^7,8. Importantly, chronic intracerebroventricular infusion of hypertonic NaCl to increase CSF Na⁺ concentrations by 5–10mM in rats produces a sympathetically-mediated hypertension ^9,10.

The brain houses specialized cells that detect changes in osmolality and NaCl concentrations of ~1% ^11,12. These specialized cells are located in circumventricular organs known as the organum vasculosum of the lamina terminalis (OVLT) and the subfornical organ (SFO) ^11,12. OVLT and SFO are midline structures juxtaposed to the third ventricle, densely vascularized, and lack a complete blood brain barrier thereby integrating electrolyte, endocrine and inflammatory signals from the blood and cerebrospinal fluid ^11,12. Neurons in the OVLT and SFO are intrinsically-sensitive to discrete changes in osmolality or extracellular NaCl ^11–15 (Figures 1A, B). Importantly, human brain imaging studies using positron-emission tomography or functional magnetic resonance report acute hypertonic saline infusion to increase plasma osmolality by ~1–2% activates the lamina terminalis (Figure 1C) ^16,17.

Figure 1. OVLT Neurons Detect Changes in Extracellular NaCl Concentrations.

A, In vitro electrophysiological recording in current clamp showing concentration-dependent excitation of rat OVLT neuron to hypertonic NaCl ¹³. B, In vivo single-unit recording of rat OVLT neuron in which intracerebroventricular infusion of 0.5M NaCl (2.5uL over 5 min) increased neuronal discharge and ABP ²⁶. C, fMRI images in the horizontal (a) and sagittal (b) planes of the human brain during IV infusion of hypertonic NaCl (0.5M NaCl at 11.4mL/kg/hr over 25 min) ¹⁶ (Copyright 2003 National Academy of Sciences U.S.A.).

Landmark studies reported that electrolytic ablation of the anteroventral third ventricular region which encompasses the OVLT and SFO prevented or attenuated every model of salt-sensitive hypertension associated with elevations in plasma or CSF NaCl concentrations ^18–20. To the extent that such experiments have been performed, selective ablation of the OVLT or SFO also attenuates deoxycorticosterone or angiotensin II-salt hypertension ^21–23. Consistent with this notion, acute optogenetic or chronic chemogenetic stimulation of OVLT neurons increase SNA and ABP ²⁴. OVLT or SFO neurons mediate NaCl-induced sympathoexcitatory and/or pressor responses ^13,25. Although limited data exists regarding the activity of SFO or OVLT in salt-sensitive models (or humans), in vivo single-unit recordings reveal the neuronal discharge of NaCl-responsive OVLT neurons was elevated in Dahl-Salt-Sensitive rats fed high salt. Interestingly, NaCl-responsive neurons in the Dahl-Salt-Sensitive animals showed exaggerated responses to NaCl. The downstream circuits that mediate changes in SNA and ABP during hypernatremia are discussed elsewhere ⁷ but involve the hypothalamic paraventricular nucleus and rostroventrolateral medulla.

NaCl versus osmolality?

The secretion of VP and thirst are stimulated by a central osmoreceptor-dependent mechanism, as such responses are evoked by infusion of any impermeable solute that osmotically withdraws water from cells (e.g. cellular dehydration) ¹¹. However, the regulation of SNA and ABP by OVLT and SFO neurons is notably different. First, OVLT neurons respond to both hypertonic NaCl and mannitol in vitro, but hypertonic NaCl evokes a greater increase in neuronal discharge than equi-osmotic mannitol ²⁶ (Figure 2A). Second, intracarotid or intracerebroventricular infusion of hypertonic NaCl evokes a greater increase in OVLT discharge than equi-osmotic sorbitol ²⁶ (Figure 2B, C). Third, intracerebroventricular infusion of hypertonic NaCl increases SNA and ABP whereas equi-osmotic infusion of mannitol or sucrose does not ²⁶ (Figure 2B, C). Limited data are available comparing the effects of Na+ versus Cl− ions on cellular excitability or regulation of SNA.

Figure 2. Hypertonic NaCl Evokes a Greater Response in OVLT Neurons than Hyperosmolality.

A, In vitro whole-cell recording of OVLT neuron demonstrating hypertonic NaCl evokes a greater response versus equi-osmotic mannitol ²⁶. B, In vivo single-unit recording of OVLT neuron to illustrate intracarotid infusion of hypertonic NaCl (50uL over 15 s) evokes a concentration-dependent increase in neuronal discharge and ABP. Intracarotid infusion of equi-osmotic sorbitol increased discharge and ABP but the magnitude was smaller ²⁶. C, Intracerebroventricular infusion of 0.5M NaCl (2.5uL per 5 min) increased OVLT neurons discharge and ABP in rats. Equi-osmotic infusion of sorbitol produced a smaller increase in OVLT discharge without any changes in ABP (²⁶).

Cellular Mechanisms of NaCl-Sensing

Putative osmosensory mechanisms within these cell populations have been extensively reviewed elsewhere ¹¹. Since SNA and ABP are regulated differently by NaCl versus osmolality, the following section will focus on the putative signaling mechanisms of NaCl-sensing. We readily acknowledge a large number of factors may modulate NaCl-sensing (e.g. neurohumoral signals such as angiotensin II or aldosterone).

Amiloride-Sensitive Channel.

Non-voltage dependent amiloride Na⁺ channels play a pivotal role in Na⁺ homeostasis and are expressed in the kidney, taste buds and also the brain. Intracerebroventricular administration of the amiloride analog benzamil largely attenuates sympathoexcitation, VP secretion, and pressor responses to central infusion of hypertonic NaCl ^27,28. Chronic intracerebroventricular administration of benzamil attenuates multiple experimental models of salt-sensitive hypertension including the Dahl-Salt-Sensitive, deoxycorticosterone-salt, aldosterone-salt, angiotensin II-salt, and stroke-prone-salt but not renovascular hypertension ^27,29–32. Since benzamil or amiloride block a number of non-voltage dependent sodium channels including the epithelial sodium channel, acid-sensing ion channel (ASIC), sodium-hydrogen exchanger, and sodium-calcium pump, any one of these ion channels could represent the putative NaCl sensing element in OVLT/SFO neurons.

NaX and ASIC1.

NaX (Scn7a) channel has a threshold of approximately 150mM extracellular Na⁺ and is primarily expressed within glial or ependymal cells within the OVLT and SFO ³³. Although thirst and VP secretion are normal in NaX-knockout mice ^34,35, pressor responses to intracerebroventricular infusion of hypertonic NaCl or deoxycorticosterone-salt were attenuated ³⁶. Na⁺ influx through the ependymal or glial cells causes a proton driven activation OVLT neurons through ASIC1a ³⁶. Surprisingly, none of the NaCl-sensitive OVLT neurons in this study were osmosensitive ³⁶ yet the majority should be ^14,26. Additionally, the magnitude of the NaCl stimulus exceeded the physiological range reported in salt-sensitive models of hypertension. It is noteworthy that there are species differences in resting osmolality and plasma/CSF NaCl concentrations ¹¹ across humans (~285–288mOsm/L), rats (~290–295mOsm/L) and mice (~310–315mOsm/L). Therefore, channels that exhibit activation of thresholds >150mM Na⁺ may sense such changes in a mouse with comparatively higher Na+ concentrations in the plasma and CSF.

Stretch-Inactivated Cation Channels.

Stretch-inactivated cation channels are partially inactivated at a normal cell volume (or membrane stretch) but exhibit cell-volume dependent gating ³⁷. Although these channels mediate osmosensory processes in the magnocellular VP neurons and the OVLT ^11,37, these may still play a role in NaCl-sensing. Increases in external Na⁺ concentrations increase the osmosensory gain as a result of modifying stretch-inactivated cation channel open probability ³⁸. Whether stretch-inactivated cation channels play a role in OVLT or SFO neurons in the context of salt-sensitive hypertension has not been explored.

KIDNEY

Renal mechanisms of NaCl sensing and handling play central roles in in the maintenance of circulatory volume, body fluid and electrolyte homeostasis, and ABP. Normally, the kidneys filter a large amount of NaCl, with the vast majority being reabsorbed in the proximal tubule and the remainder being titrated in the more distal segments of the nephron to maintain NaCl balance in the body. The capacity of the kidney to respond to increases (or decreases) in NaCl intake is quite large, ensuring that NaCl homeostasis is maintained. As a component of this regulation, the kidney senses NaCl.

The macula densa (MD) is the chief NaCl sensor in the kidney. MD cells are localized in a strategically central position in each nephron at both the vascular pole entrance of the kidney filter (glomerulus) and at the early distal convoluted tubule. MD cells provide key physiological control of basic kidney functions including glomerular filtration rate, renal blood flow and renin release via sensing alterations in NaCl concentrations in the distal tubule fluid and by transducing this information into intracellular signaling and the synthesis and release of paracrine acting mediators that target various other neighboring cell types in the juxtaglomerular apparatus (JGA) ^39,40. Importantly, MD cells sense NaCl concentrations in the tubular fluid rather than transport NaCl across the MD epithelium, unlike other renal tubular epithelial cell types. This is achieved by a paucity of Na⁺-pump (Na⁺/K⁺-ATPase) expression and activity at the basolateral cell membrane ⁴¹ and a unique set of cell membrane ion transporters. These include the two main apical sodium entry pathways Na⁺:2Cl⁻:K⁺ cotransporter (NKCC2) and Na⁺/H⁺ exchanger (NHE2), and additional Na⁺(H⁺)/K⁺-ATPase and K⁺ channel (ROMK) ^39,41–43. The most prominent basolateral ion transporters in MD cells are chloride and cation channels, Na⁺/H⁺ exchanger (NHE4), and Cl⁻/HCO₃⁻ exchanger ^39,42–44. Collectively, the activity of these ion transporter pathways permit MD cells to closely track variations in NaCl concentrations of distal tubular fluid, with the highest sensitivity in the 10–60 mM NaCl range ³⁹. Transduction of the luminal NaCl signal by MD cells involve changes in intracellular Na⁺ and Cl⁻ concentrations, pH, basolateral membrane potential that trigger MD cell calcium signaling, Nos1-mediated NO release, MAP kinases p38 and ERK1/2, COX2-mediated PGE₂ release (for renin control), or ATP/adenosine release (for hemodynamic control, TGF). These traditional MD sensing mechanisms and functions have been reviewed previously in detail ^39,40,43,45 and are depicted in Figure 3 (left panel).

Figure 3. Traditional and novel salt sensing mechanisms and functions of MD cells.

ERK1/2: extracellular signal regulated kinase 1 and 2; Cox2: cyclooxygenase 2; Nos1: nitric oxide synthase type 1; PGE2: prostaglandin E2; NO: nitric oxide; SGLT1: sodium-glucose cotransporter 1; Pappa2: pregnancy-associated plasma protein a2.

Advanced research techniques may help improve understanding of the dynamic changes in Na⁺ concentrations in the local renal tissue microenvironment and their role in Na⁺ sensing. Examples include ²³Na magnetic resonance imaging ⁴⁶ and intravital multiphoton microcopy (MPM) ^47–49 of tissue Na⁺, which can be measured similarly to the measurement of proximal tubule bicarbonate reabsorption ⁵⁰. Due to the physiological significance of MD cells being key salt sensors in the kidney, future non-invasive measurements of the physiological variations in the MD tubular fluid Na⁺ concentration with high spatial and temporal resolution at the single nephron level will open new possibilities for future research.

State-of-the-art transcriptomic analysis, genetic animal models and imaging techniques has identified novel mechanisms of MD cell salt sensing. The expression of the sodium-glucose cotransporter 1 (SGLT1) at the MD luminal membrane couples salt and metabolic sensing by these cells in conditions of hyperglycemia, and via Nos1-dependent NO production and altering hemodynamics (tubuloglomerular feedback response) this mechanism contributes to the development of glomerular hyperfiltration in diabetes ^51–53. Further emphasizing the important role of MD Nos1, the activity of the Nos1b splice variant in MD cells plays an important role in the development of salt-sensitive hypertension in a gender-specific manner ^54–57. In addition, alterations in dietary salt intake leads to changes in MD cell ultrastructure, including the length and density of their basal cell processes network (maculapodia) ⁴⁷, and the robust level of MD cell synthesis of classic and novel intracellular and secreted proteins such as Pappa2 ⁴⁹. These new structural and functional MD mechanisms may play further important roles in salt sensing and cellular communications in the JGA, especially in chronic adaptations to salt intake. Interestingly, it was shown recently that Pappa2, a metalloproteinase that is highly and specifically expressed in MD cells and surrounding tubule segments can protect against salt-induced hypertension ^58–60. Finally, additional mechanisms of MD cell salt sensing are emerging that are expected to further increase our understanding of this important renal cell type.

In addition to MD cells as NaCl sensors, sensory nerves originating in the T8-L2 dorsal root ganglion innervate the pelvic wall, renal vasculature, and to a lesser extent, the tubules in the renal cortex ^61,62. Activation of renal sensory fibers alters SNA, renal function, and ABP ^61,62. These sensory fibers respond to both chemosensitive and mechanosensitive stimuli including renal ischemia, chemokines, increased pelvic pressure but also urinary NaCl concentrations ^61,63,64. These effects appear NaCl-dependent as responses to responses to equi-osmotic mannitol or urea are attenuated or absent ⁶³. The molecular identity of NaCl sensing in renal afferent nerves is unknown.

GASTROINTESTINAL-HEPATIC SENSING OF SALT INTAKE

More than 50 years ago it was shown that sodium absorption from the gastrointestinal tract had rapid physiological effects that occurred more quickly than could be explained by changes in systemic Na⁺ concentration or osmolarity. Initial studies focused on the impact on fluid intake or VP secretion and suggested that this “pre-systemic” sensing ingested Na⁺ could occur at the level of the intestinal-hepatic portal vein region. Interestingly, hepatic denervation eliminated the post-prandial natriuresis and decrease in renal SNA of dogs fed a high NaCl meal ⁶⁵. The role of hepatic portal sensing of Na⁺ in cardiovascular regulation, particularly in a chronic perspective, is unclear.

Hepatic Portal Osmosensing.

Early studies on the potential of feed-forward signals arising from the hepatic portal region demonstrated that infusion of hypertonic saline solutions into the portal vein produced a rapid antidiuretic response ⁶⁶ and increased hepatic afferent nerve activity ^67–69. The physiological significance of these observations was demonstrated by Baertschi and colleagues showing that intragastric infusions of hypertonic NaCl in conscious rats increased plasma VP levels without a change elicited in plasma osmolality ^70–73. Similarly, studies showed hepatic effects on renal excretion ^65,66.

Osmosensing Versus NaCl Sensing.

Electrophysiological recordings mostly suggested that afferent nerves were activated by hyperosmotic solutions ^67–69 with a possible selectivity of Na^{+ 67}. Adachi et al ⁶⁸ recorded from bundles of hepatic afferent nerves in rat and found some increased in response to hypertonic saline (~30%), some to water (~20%), and some to neither solution (~50%). The nerve bundles that responded generally to hypertonic solutions responded most vigorously to hypertonic NaCl solutions. Morita and Abe ⁷⁴ show recordings of hepatic afferents responding to portal vein injections of hypertonic NaCl and NaHCO₃, but not LiCl or mannitol, suggesting the presence of Na⁺ sensors. Furthermore, bumetanide blocked the response to hypertonic saline suggesting that the Na⁺,K⁺,2Cl⁻ transporter may be involved in the Na⁺ sensing mechanism. Modern techniques that allow recording from many single afferent neurons instead of bundles of nerve fibers find that large numbers of vagal afferent neurons responded hypertonic saline infusion into the intestines ⁷⁵. In contrast, few neurons in thoracic dorsal root ganglia responded to these osmotic stimuli. Zimmerman et al ⁷⁶ recorded Ca⁺⁺ signals in SFO neurons while infusing solutions into the stomach of mice and reported that these neurons respond to osmotic load, though from their data it appears that the response is more robust when the osmotic load is Na⁺. Consistent with these observations that the afferent neurons respond to changes in osmolality rather than specifically Na⁺, intragastric infusion of hypertonic solutions, not specifically Na⁺, stimulated VP secretion ⁷². On the other hand, hepatic portal administration of hypertonic Na⁺ but not other hypertonic solutions reduced Na⁺ intake in Na⁺-deprived rats ⁷⁷.

Overall, the issue of whether hepatic afferents respond to Na⁺ versus hyperosmolality seems to not yet be settled. It would be important to test both hypertonic saline and other hypertonic solutions on a large population of afferent neurons, but this would also require an understanding of where the relevant afferent neurons are (vagal versus DRG) and where the sensory endings are (e.g., in the intestinal lining or in the hepatic-portal area). The possibility that an interaction between osmotic stimuli and Na⁺ exists, consistent with observations that hyperosmotic stimuli containing Na⁺ exert a more powerful effect than other hyperosmotic stimuli, should also be considered.

Vagal Versus Spinal Pathway.

There is also uncertainty regarding the pathways by which this osmotic/Na⁺ information is conveyed to the central nervous system. Both afferent recording studies as well as functional lesion studies provide conflicting data as to whether the relevant afferent pathways are vagal or spinal (or both). Some of the studies recording from hepatic afferent nerves do not distinguish between vagal versus spinal afferents, though the more recent studies monitoring Ca⁺⁺ signals in vagal afferent neurons (or brain neurons following vagal transection) show clearly that some vagal afferents respond ^75,76, whereas few DRG neurons appear to ⁷⁵. Hepatic vagal transection eliminated the effect of hepatic portal Na⁺ on VP secretion in dogs ⁷⁸ or Na⁺ appetite in Na⁺-deprived rats ⁷⁷, and attenuated c-Fos expression in some brain regions ⁷⁹. In regard to a spinal afferent pathway, splanchnic nerve transection has either attenuated ⁷² or did not alter ⁷⁹ VP secretion to intragastric hypertonic saline.

Role of Hepatic Portal Na⁺ Sensing in Cardiovascular Regulation.

Despite the evidence for this pre-systemic signal arising from the hepatic portal area playing a role in rapid responses involved in fluid homeostasis, a role in longer term Na⁺ balance and cardiovascular regulation is uncertain. Morita et al. ⁸⁰ reported that denervation of the hepatic portal area in rats had no effect on Na⁺ balance or ABP while rats were maintained on a standard 0.45% NaCl diet. However, denervated rats fed a high NaCl diet showed a greater increase in Na⁺ balance and an increased ABP. This increase in Na⁺ balance in response to a high salt diet is similar to what has been reported for rats treated with carbon tetrachloride to induce liver cirrhosis, which is accompanied by hepatic denervation ⁸¹. In contrast, Carlson et al ⁸² reported that hepatic denervation increased mean ABP by ~10 mm Hg, independent of the NaCl content of the diet. Taken together, these observations, coupled with the conflicting data regarding precise location of the relevant afferent nerves and whether they are vagal or spinal, raises questions as to the role of pre-systemic Na⁺ sensing in chronic fluid and cardiovascular regulation.

Interestingly, Tsuchiya et al ⁸³ report that rats fed a high Na⁺ diet for 4 weeks showed a marked decrease in hepatic afferent nerve response to hypertonic saline infusion into the hepatic portal vein. Similar results were observed with a high K⁺ diet (8% K⁺, compared to 0.8%) consistent with decreased hepatic expression of NKCC1. These observations raise the possibility that high dietary NaCl intake can induce chronic changes in Na+ sensing relevant to the long-term regulation of cardiovascular function.

Two additional points should be discussed. First, whereas the preceding discussion has focused on hepatic portal sensors, gut microbiota or immune cells associated with the gastrointestinal tract may contribute to salt-sensitive hypertension; this important area of research has been recently reviewed ⁸⁴. Second, the preceding discussion of hepatic Na⁺/hyperosmotic sensing should not be confused with the water drinking-induced pressor response (osmopressor response) that has be described in autonomically-impaired people and experimental animals ⁸⁵.

PERSPECTIVE / SUMMARY

Multiple systems contribute to salt-sensitive hypertension including impaired renal salt handling, vascular dysfunction, and sympathoexcitation. Although the brain, kidney, and gastrointestinal tract contain specialized cells that detect or monitor NaCl concentrations, the cellular and electrochemical mechanisms are still not fully understood. Whether the primary signal is NaCl versus Na⁺ versus Cl⁻ and whether these are sensed by different mechanisms or cells remains unknown. This is particularly clinically relevant as dietary NaHCO₃ and NaCl increases ABP in salt-sensitive humans ⁸⁶, but the magnitude of the hypertension is attenuated after NaHCO₃ despite a clear relationship between ABP and serum Na⁺ concentration. Lastly, salt-sensitive hypertension is influenced by biological variables including sex, age, and ethnicity ^4–6, yet there is extremely limited data to test whether such factors impact the ability of these specialize cells to detect NaCl concentrations and predispose these populations to salt-sensitive hypertension.

Abbreviations:

ABP

Arterial blood pressure

ASIC

Acid-sensing ion channel

CSF

Cerebrospinal fluid

MPM

Intravital multiphoton microcopy

JGA

Juxtaglomerular apparatus

MD

Macula Densa

OVLT

Organum vasculusom of the lamina terminalis

NaCl

Salt

Na⁺

Sodium

SFO

Subfornical organ

SNA

Sympathetic nerve activity

VP

Vasopressin