Dietary salt, vascular dysfunction, and cognitive impairment

Abstract

Excessive salt consumption is a major health problem worldwide leading to serious cardiovascular events including hypertension, heart disease, and stroke. Additionally, high-salt diet has been increasingly associated with cognitive impairment in animal models and late-life dementia in humans. High-salt consumption is harmful for the cerebral vasculature, disrupts blood supply to the brain, and could contribute to Alzheimer’s disease pathology. Although animal models have advanced our understanding of the cellular and molecular mechanisms, additional studies are needed to further elucidate the effects of salt on brain function. Furthermore, the association between excessive salt intake and cognitive impairment will have to be more thoroughly investigated in humans. Since the harmful effects of salt on the brain are independent by its effect on blood pressure, in this review, I will specifically discuss the evidence, available in experimental models and humans, on the effects of salt on vascular and cognitive function in the absence of changes in blood pressure. Given the strong effects of salt on the function of immune cells, I will also discuss the evidence linking salt consumption to gut immunity dysregulation with particular attention to the ability of salt to disrupt T helper 17 (Th17) cell homeostasis. Lastly, I will briefly discuss the data implicating IL-17A, the major cytokine produced by Th17 cells, in vascular dysfunction and cognitive impairment.

This article is part of the Spotlight Issue on Brain, Heart, and Vessels Crosstalk.

1. Introduction

Salt is essential for life, but its worldwide consumption exceeds the minimal requirements.¹ Although controversy exists regarding the ideal level of salt intake,² high-salt consumption has been consistently associated with increased risk of serious cardiovascular events³ and is responsible for more than half of diet-related deaths and two-thirds of diet-related disability-adjusted life-years.^3–5 Dietary salt has a major role in the pathogenesis of hypertension,⁶ a leading cause of death and disability worldwide,⁷ and lowering blood pressure (BP) by reducing salt intake⁸ is one of the most cost-effective ways to reduce cardiovascular risk.^9,10 Thus, major global strategies have been implemented to reduce salt food content, to educate the public, and to help consumers select food products with lower sodium content.¹¹ However, since most of these efforts have not been fully successful, the use of salt substitute is emerging as an alternative strategy to reduce the incidence of serious cerebrovascular events among persons at risk.^12,13

Despite the deleterious effects of salt on BP, accumulating evidence indicates that the risk of serious cardiovascular events associated with excessive salt consumption is independent by its effect on BP.¹⁴ Therefore, maintaining low dietary salt intake may be beneficial even in subjects without hypertension.^15,16 Indeed, excessive salt intake is harmful for the brain and increases the risk for stroke^2,3,17–21 and cerebral small vessel disease (CSVD),^22,23 a common cause of vascular cognitive impairment in the elderly. Increasing evidence also indicates that high-salt consumption impairs cognitive function²⁴ independently of known risk factors, including hypertension and APOE genotype.²⁵ However, the mechanisms by which excessive salt intake may alter brain function are not well understood.

This review will examine the evidence linking excessive salt consumption to vascular dysfunction and cognitive impairment independently by its effects on BP. In addition, it will discuss the novel evidence demonstrating that salt can alter gut immunity homeostasis and in turn promote cerebrovascular and cognitive dysfunction.

2. High-salt diet and endothelial-dependent vasodilation in animal models

Numerous animal studies have demonstrated that high-salt diet (HSD) mediates endothelial dysfunction independently of the increase in BP.¹⁶ In 1987, Luscher et al.²⁶ provided one of the first evidence of the deleterious vascular effects of HSD. In this study, HSD increased the vasoconstrictive response to serotonin or norepinephrine of aortic rings isolated from normotensive Dahl salt-resistant (Dahl-R) rats fed a HSD, an effect lost in the absence of the endothelium indicating that HSD altered the ability of endothelial cells (ECs) to limit serotonin- or norepinephrine-induced constriction.²⁶ Similarly, HSD enhanced vasoconstriction to endothelin-1 (ET-1) and the α-adrenergic agonist phenylephrine in aortic rings of normotensive Sprague–Dawley (SD) rats.²⁷ Furthermore, HSD impaired the release of endothelial nitric oxide (eNOS) and the nitric oxide (NO)-dependent vasodilation of aortic rings in response to cholinergic receptor activation by methacholine.²⁸ These effects were rescued by the free radical scavenger TEMPOL suggesting that the NO deficit was likely dependent on increased vascular oxidative stress.^28,29 More recent data indicated that the NO deficit associated with HSD may depend on the activation of the nucleotide binding domain and leucine-rich repeat protein-3 (NLRP3) inflammasome.³⁰ Genetic deletion or pharmacological inhibition of NLRP3 inflammasome restored endothelium-dependent vasodilation.³⁰ In addition, exposure of aortic ECs to high-salt medium increased reactive oxygen species (ROS) and NLRP3 expression, effects abrogated by ROS scavengers (Figure 1).³⁰

The vascular effects of HSD are not limited to the aorta. Boegehold^31,32 showed that the arteriolar dilation in response to acetylcholine (ACh) was reduced in spinotrapezius muscle arterioles from normotensive Dahl-R rats and C57Bl/6J mice fed HSD. Inhibition of eNOS suppressed dilation of arterioles in Dahl-R rats fed a control diet but not in those fed a HSD further indicating that HSD impairs the ability of endothelial NO to regulate resting vascular tone.³¹ Flow-dependent dilation, a response in part mediated by endothelial NO, was also attenuated by HSD.³³ In the same vascular bed, HSD attenuated the arteriolar myogenic responsiveness,³⁴ an effect later associated with reduced NO availability.³⁵ As in the aorta, ROS scavengers rescued the arteriolar response to ACh in HSD rats, further supporting a role of ROS in NO breakdown.³⁶ However, pharmacological inhibition of NADPH oxidase or xanthine oxidase abolished the oxidative stress associated with HSD,³⁷ but did not restore the response to ACh indicating that other sources of ROS could be involved.³⁷ Accordingly, eNOS uncoupling has emerged as a potential source of ROS in the vascular wall of HSD mice.³² In arterioles of HSD mice, the eNOS inhibitor L-NMMA reduced basal and ACh-induced superoxide levels, pointing at eNOS uncoupling as a potential source of ROS in HSD.³² Indeed, administration of the eNOS substrate L-arginine, increased levels of tetrahydrobiopterin (BH₄), an essential cofactor of eNOS, suppressed ROS production and improved the vasodilator response ACh of HSD arterioles (Figure 1).³⁸ Altogether, these data indicate that HSD could increase microvascular ROS through eNOS uncoupling.³² In support of this, HSD did not alter the vascular smooth muscle responsiveness to NO arguing against any significant ROS-mediated inactivation of NO within the vascular smooth muscle cells and suggesting that ECs are the more likely site of ROS generation in HSD.^32,39 Additional mechanisms contributing to the increased ROS levels and the associated loss of microvascular NO in HSD may include reduced activity of antioxidant enzymes including Cu/Zn superoxide dismutase (SOD) (Figure 1).³⁹

The attenuation of NO-mediated arteriolar vasodilation has been later demonstrated in several other vascular beds (Figure 1). In cremaster arterioles, exposure to HSD time-dependently attenuated the vascular response to Ach,^40–42 which was restored by returning animals to the control diet.⁴² Incubation of mesenteric arteries with methacholine or ACh increased NO release and induced vasodilation in vessels from ND but not HSD rats.^43,44 ROS scavenging or NADPH oxidase inhibition restored NO levels and improved the NO-dependent vasodilator responses.^43,44 Similar findings were also reported in mesenteric vessels from HSD mice.⁴⁵ Interestingly, treatment with the transient receptor potential vanilloid type 1 (TRPV1) agonist capsaicin reduced ROS, restored NO levels, and partially rescued the vascular response to ACh.⁴⁵ In Wistar rats, HSD decreased renal endothelium-dependent vasodilatation,⁴⁶ an effect possibly linked to diminished eNOS expression.⁴⁶ In another study, HSD increased ET-1 and ROS levels in the kidney.⁴⁷ Both effects were attenuated by endothelial ET-1 deletion or blocking of ET_AR.⁴⁷ Since inhibition of NADPH oxidase also suppressed ROS levels, a possibility is that ET-1 could increase oxidative stress by activating NADPH oxidase.⁴⁷

Accumulating evidence indicates that HSD also alters cerebrovascular function.^48–53 Liu et al.⁵⁰ found that a short-term HSD impaired ACh response in rat pial arterioles. Response to NO donors was not altered further indicating that HSD selectively alters the ability of brain ECs to release NO without altering the ability of vascular smooth muscle cells to relax in response to NO.⁵⁰ Elevated dietary salt intake completely abolished the ACh-induced dilation of isolated middle cerebral arteries and significantly reduced that of basilar arteries,⁴⁸ and these effects were lost in endothelium-denuded vessels or when L-NMMA, an eNOS inhibitor, was added to the perfusion chamber.⁴⁸ The vasodilator response to bradykinin, which is dependent on COX-1-derived metabolites,⁵⁴ was not altered by HSD,⁴⁸ further indicating a selective impairment in endothelial NO availability. Consistent with this, NO production, assessed by diaminofluorescein-2 (DAF-2) fluorescence, was decreased in cerebral arteries of HSD rats.⁴⁸ The impairment of flow- and ACh-induced dilatation of middle cerebral artery induced by HSD was associated with reduced levels of the antioxidant enzyme, glutathione peroxidase 4.⁵⁵ However, differently from the periphery, ROS scavenging did not restore the response to ACh indicating that factors other than oxidative stress could be involved.^48,53

More recently, to investigate the cerebrovascular effects of HSD in C57Bl/6J mice, we measured resting cerebral blood flow (CBF) quantitatively using MRI with arterial spin labelling.⁵³ We found that HSD attenuated resting CBF both in the cortex and in the hippocampus.⁵³ Since resting CBF is highly dependent on endothelial NO,^56,57 we then tested the effects of HSD on the local CBF increase produced by ACh, a response also dependent on endothelial NO.⁵⁸ We found that the CBF response to ACh was attenuated in HSD mice.⁵³ In line with this, NO basal levels, assessed by DAF fluorescence, were reduced in cerebral vessels isolated from HSD mice.⁵³ Additionally, HSD suppressed the ability of cerebral vessels to produce NO in response to Ach. On the contrary, the increases in CBF evoked by mechanical stimulation of the whiskers (functional hyperaemia) were not reduced in HSD indicating that HSD selectively impairs the ability of ECs to produce NO, without altering the ability of neurons or astrocytes to regulate CBF. Administration of the eNOS substrate L-arginine in the drinking water normalized endothelial-dependent vasodilation and microvascular NO levels providing further evidence that suppression of eNOS-derived NO mediates the cerebrovascular effects of HSD.⁵³ Of note, the cerebrovascular alterations induced by HSD could be rescued after returning dietary salt intake to normal levels.⁵³ eNOS catalytic activity is dynamically regulated by phosphorylation.⁵⁹ Thr⁴⁹⁵ phosphorylation reduces eNOS catalytic activity and NO production while Ser¹¹⁷⁷ increases them. We found that HSD had no effect on Ser¹¹⁷⁷ but increased eNOS phosphorylation at Thr⁴⁹⁵ in cerebral microvessels.⁵³ The increase in pThr⁴⁹⁵ was dependent on the activation of Rho-kinase (ROCK) since administration of the ROCK inhibitor Y27632 ameliorated the endothelial dysfunction in HSD mice (Figures 1 and 2).⁵³ Collectively, these observations further indicate that HSD profoundly alters the endothelial regulation of cerebral perfusion by mediating eNOS inhibition and reduced endothelial NO production.

3. HSD and endothelial-dependent vasodilation in humans

Independent of its effects on BP, increased sodium intake exerts adverse effects on vascular function also in humans. An acute increase in sodium intake impaired endothelial function in young adults in the absence of BP changes.⁶⁰ In this study, subjects received either salt supplement consisting in a total amount of 4.6 mg/day of sodium or placebo for 5 days. Forearm blood flow (FBF) was measured by strain-gauge venous occlusion plethysmography, and the response to intra-arterial infusions of ACh and SNP was assessed.⁶⁰ Consistent with the observations in animal models, vasodilation induced by ACh was attenuated after salt loading whereas that to the NO donor was not altered.⁶⁰ In a similar study, the ACh-induced vasodilatation was reduced following HSD (5 days, 9.2 mg/day Na) in men but not in women.⁶¹ The eNOS inhibitor L-NMMA did not blunt the vasodilator response to ACh indicating that HSD attenuates the NO component of ACh-mediated vasodilation.⁶¹ Furthermore, the reduction of the brachial artery flow-mediated dilation (FMD) associated with HSD (7 days, 7–8 mg/day Na) was greater in men than women further indicating that men might be more sensitive to the deleterious effects of HSD on the vasculature.⁶² The same HSD regimen attenuated the cutaneous vasodilatation in response to local heating in healthy, normotensive adults who manifested salt resistance.⁶³ The vascular response to heating includes an initial vasodilation largely due to an axon reflex with a small NO contribution,⁶⁴ followed by a secondary increase in blood flow which is predominately mediated by endothelial NO.⁶⁴ Remarkably, in the absence of BP changes, only the NO-dependent response was impaired in HSD subjects and the local infusion of ascorbic acid rescued the dysfunction, suggesting a role for oxidative stress in mediating the effect.⁶³ In a subsequent study, the same group demonstrated that the endothelial-dependent dilation induced by shear stress was also attenuated by HSD whereas the endothelium-independent dilation induced by sublingual nitroglycerine administration was not altered.⁶⁵ Although controversial,⁶⁶ a high-salt meal seems sufficient to suppress the FMD, in the postprandial phase, in healthy normotensive men and women.⁶⁷ Of interest, the addition of potassium counteracted the deleterious effects of sodium.⁶⁸ HSD (5 days, 9.2 mg/day Na) reduced plasma levels of nitrite and nitrate, an index of NO production.⁶⁹ However, the vascular response to mental stress, dependent on NO production,⁷⁰ was not altered by HSD.⁶⁹ Since dietary salt reduction improved endothelial function in overweight and obese normotensive subjects,⁷¹ a possibility is that salt could contribute to the endothelial dysfunction associated with other cardiovascular risk factors.⁷¹ In a cross-sectional study involving pregnant women, high dietary salt intake, assessed by 24-h sodium excretion, was associated with a reduction in circulating NO levels and attenuation of flow- and ACh-mediated dilation of the brachial artery.⁷² Consistent with previous findings, HSD did not alter the endothelium-independent dilation induced by SNP.⁷² The increase in forearm skin blood flow induced by post-occlusive reactive hyperaemia and ACh was attenuated in HSD subjects.^73–75 Since the post-occlusive reactive hyperaemia is not dependent on NO,⁷⁶ these data suggest that HSD may also impair NO-independent pathways regulating blood flow. The administration of antioxidants prevented the vascular effects of HSD, pointing at ROS generation as a potential mediator of the effect.⁷⁴ Lastly, changes in plasma sodium concentration per se could affect the stiffness and deformability of ECs leading to altered endothelial function and vascular tone.^77,78 Altogether, the evidence presented above strongly supports the concept that, in humans as in rodents, HSD has deleterious effects on vascular function beyond those attributable to the increase in BP.

4. HSD and cognitive impairment in animal models

There is now considerable evidence indicating that excessive salt intake causes cognitive impairment in rodents.⁷⁹ HSD altered anxiety-like behaviour and short-term memory (radial arm water maze test) in aged F344/brown Norway (F344/BN) rats.⁸⁰ The increased oxidative stress in the brain of HSD aged F344/BN rats suggests that ROS could play a role.⁸⁰ SD rats fed a HSD showed no anxiety-related behaviour but developed cognitive impairment assessed by the Morris water maze (MWM) test,⁸¹ an effect associated with dendritic spines loss.⁸¹ Since, in both studies, HSD increased systolic arterial pressure, its potential effect cannot be excluded. In a recent study, cognitive function in young and aged SD rats fed a HSD was assessed by the novel object recognition (NOR) test and the T-maze.⁸² HSD had no effects on short- and long-term memory in adult rats.⁸² However, HSD impaired long-term memory in aged rats indicating that aging and dietary salt may have an additive effect on cognitive function.⁸² In C57BL/6J mice, HSD impaired retention but not the acquisition of spatial memory.⁸³ The performance of HSD mice in the MWM test was not significantly different from that of control mice during the training phase.⁸³ However, during the probe test, HSD mice spent less time in the target quadrant than mice fed a control diet indicating a deficit in the retention of spatial memory.⁸³ Levels of superoxide, assessed by DHE staining, were higher whereas the activity of several antioxidant enzymes was reduced in the hippocampus of HSD mice.⁸³ However, whether hippocampal oxidative stress contribute to cognitive impairment was not tested. Nevertheless, this is one of the first evidence that HSD may impact cognition independently from its effect on BP.⁸³ In the object-place recognition task, the time spent exploring the object that had been moved to a new location was less in HSD mice indicating impairment in short-term memory.⁸⁴ Long-term memory, assessed by contextual fear conditioning, was also impaired in HSD mice⁸⁴ and, in contrast to the object-place recognition task, the deficit was already present after 4 weeks of HSD⁸⁴ indicating that long-term memory may be more susceptible to HSD than short-term memory. The effect was associated with a suppression of synaptic plasticity induced by long-term potentiation (LTP) in the CA1 region of the hippocampus⁸⁴ and with a down-regulation of proteins mediating memory and learning.⁸⁴ In line with previous evidence, ROS generation was increased whereas antioxidant capacity was reduced in the hippocampus of HSD mice further suggesting a role for oxidative stress.⁸⁴ Evidence from our lab further attests to the ability of HSD to induce profound alterations in cognitive function that involves multiple domains.^53,85 Using the NOR test, we demonstrated that HSD mice spent equal time exploring novel and familiar objects, indicating a failure to identify the novel one.⁵³ Of note, return to normal diet was associated with normalization of the performance at the NOR test, indicating that these deficits are reversible within the time frame of our experiments.⁵³ To further investigate the effect of HSD on cognitive function, we used the Barnes maze, a hippocampus-dependent task requiring spatial memory to learn and retain the location of an escape hole.⁸⁶ We found that HSD mice were able to learn the location of the escape hole. However, when the escape hole was moved to the opposite quadrant, both primary latency and distance travelled were significantly longer in HSD mice.^53,85 Lastly, we examined whether HSD affects nesting building behaviour which is considered to be equivalent to the activities of daily living typically altered in patients with cognitive impairment.⁸⁷ The ability of the mice to build a nest, assessed by the Deacon rating scale,⁸⁷ and the amount of nesting material used were reduced in HSD mice, attesting to disruption of this behaviour. In a follow-up study, we found that HSD induced phosphorylation of tau epitopes that promote neuronal dysfunction including Ser²⁰²/Thr²⁰⁵ and Thr²³¹ (Figure 2).⁸⁵ HSD not only promoted hyperphosphorylation of tau, but also its aggregation as demonstrated by increased levels of insoluble tau in HSD brains.⁸⁵ Of note, HSD did not impair cognitive function in tau-null mice and in mice injected with anti-tau antibodies (HJ8.8), demonstrating the contribution of tau to the cognitive effects of HSD.⁸⁵ Cyclin-dependent kinase 5 (CDK5) is a major kinase regulating tau phosphorylation, and its activity is tightly regulated by its binding partner p35.⁸⁸ However, in neurodegenerative conditions, activation of calpains, Ca²⁺-dependent proteases, leads to cleavage of p35 into p25 leading to overactivation of CDK5.⁸⁹ Since reduced endothelial NO may lead to tau phosphorylation by activating CDK5,⁹⁰ and because calpain activity is regulated by their nitrosylation,⁹¹ we tested the effects of HSD on calpains and Cdk5 activity. Cdk5 activity was increased in HSD mice, and administration of the CDK5 peptide inhibitor TFP5 attenuated phosphorylation of tau and prevented cognitive dysfunction.⁸⁵ By using the biotin switch assay, we also found that HSD suppressed calpain nitrosylation and this was associated with increased calpain activity and p25/p35 ratio indicative of increased p35 cleavage (Figure 2).⁸⁵ Importantly, nitrosylation was markedly suppressed in eNOS-null mice, but not in nNOS-null mice, indicating the key role of eNOS-derived NO in regulating nitrosylation and activity of calpain.⁸⁵ Overall, these findings indicate that a deficit in endothelial NO is crucial in mediating the cognitive dysfunction associated with HSD by setting in a motion a cascade of events that results in de-nitrosylation of calpain, activation of CDK5, and phosphorylation of tau.⁸⁵

Gilman et al.⁹² also found an impairment of nesting behaviour in mice drinking water containing 4% of NaCl for 7 days. Increased salt intake did not significantly affect anxiety-related behaviours in the elevated plus maze, open field, and marble burying test but promoted social interaction indicating that HSD might lower behavioural inhibitions thereby reducing engagement in relevant behaviours that promote survival.⁹² Similarly, HSD did not induce anxiety but impaired spatial memory and learning in both the MWM and the shuttle box test.⁹³ As shown before,⁸⁵ these effects were associated with increased tau phosphorylation and impairment in neuronal autophagy.⁹³ However, the association between autophagy, tau phosphorylation, and cognitive function in HSD mice was not investigated in sufficient mechanistic details.

The cognitive effects of HSD are also present in female mice and have been associated with increased tau phosphorylation,⁸⁵ blood–brain disruption,⁹⁴ and evidence of apoptosis and neuroinflammation.⁹⁴ Furthermore, HSD may have deleterious consequences during pregnancy. Adult offspring from dams exposed to HSD spent significantly less time in the novel arm in the Y-maze test, showed a reduced preference for displaced object in the novel object-place recognition test, and failed to remember the location of the escape hole in the Barnes maze test.⁹⁵ LTP was suppressed, and levels of synaptic proteins were attenuated in adult mice born from dams exposed to HSD attesting to the potential long-lasting effects of HSD on synaptic plasticity.⁹⁵

5. HSD and cognitive impairment in humans

The relationship between salt intake and cognitive function in humans is not well established as in animal models²⁴ and needs to be better elucidated. The evidence linking high-salt intake and poor cognition is hindered by numerous limitations including sample size and suboptimal assessment of sodium intake.²⁴ Since most of these studies are cross-sectional, it is also difficult to establish a cause-and-effect relationship. In addition, the effect of salt has been mainly investigated in older adults. However, considering that the pathological processes associated with cognitive impairment and dementia may start many years before the development of symptoms, assessing the effect of salt intake in mid-life may provide better insights into the association between excessive salt intake and cognitive impairment. Finally, in contrast to animal models, it is extremely difficult to exclude the contribution of other dietary components that might also influence cognition.

Despite the limitations, a positive association between high sodium intake and cognitive decline has been found in most of the studies which yielded significant results.²⁴ In a relatively large cross-sectional study where participants were asked a single self-reported question, preferring a salty diet was negatively associated with a higher prevalence of cognitive impairment.⁹⁶ Salerno-Kennedy and Cashman⁹⁷ found that sodium intake was significantly higher in subjects with altered cognitive function assessed by the Mini Mental State Examination (MMSE) test. However, out of the 44 subjects enrolled in this study, only 4 had a MMSE score indicative of cognitive impairment (MMSE <24).⁹⁷ In addition, sodium intake was estimated by a semiquantitative food frequency questionnaire and urinary sodium amount was not measured.⁹⁷ A negative correlation between MMSE score and sodium was also demonstrated by Rondanelli et al.⁹⁸ in the elderly, whereas, in another study, the association between sodium intake and cognitive function was dependent on the level of physical activity.⁹⁹ Specifically, this study indicated that the combination of low levels of physical activity and high levels of sodium intake is particularly detrimental to cognitive health.⁹⁹ Impairment of cognitive function, particularly immediate recall memory, was associated with excessive salt and insufficient potassium intake in a cohort of heart failure patients.¹⁰⁰ In this study, levels of salt intake were estimated to be around two times higher than what was recommended by the World Health Organization.¹⁰⁰ An increase in the 24-h urinary sodium to potassium ratio, indicative of an imbalance in the intake of sodium and potassium, was also associated with mild cognitive impairment (MCI).¹⁰¹ A negative correlation between salt intake and MMSE score was also found by Afsar¹⁰² in patients with essential hypertension. In a more recent study, the correlation was still present after adjusting for hypertension.²⁵ Here, sodium intake was assessed by collecting 24-h urine samples over 7 consecutive days to estimate salt intake at baseline.²⁵ Changes in salt intake of participants during the follow-up study were determined using the spot urine method and Tanaka equation.²⁵ This study, conducted in over 2000 healthy participants, used a combination of MMSE, Montreal Cognitive Assessment, and Mattis Dementia Rating Scale scores to assess cognitive function and found that the risk of dementia as well as the progression of cognitive impairment increased progressively with salt intake.²⁵ Since salt intake may cause changes in plasma sodium,¹⁰³ few studies have also investigated the link between serum sodium concentration and cognitive function.^104–107 Both low (126–140 mmol/L) and high serum sodium (>146 mmol/L) were associated with cognitive decline indicating that the association of serum sodium level with cognitive function may have a U- or J-shape.^104,106 However, another study found no association between serum sodium and risk of dementia.¹⁰⁵

The association between salt intake and cognitive impairment was not found in a large prospective study conducted in post-menopausal women over a 9-year follow-up period.¹⁰⁸ When stratified by hypertension status, hypertensive women with higher sodium intake had a higher risk of developing cognitive impairment, although the effect was not significant.¹⁰⁸ In line with this, a cross-sectional study, where all participants underwent a research clinic visit to assess cognitive function and obtain nutritional information, found that lower sodium intake was associated with increased odds of clinically significant impairment on the MMSE test.¹⁰⁹ In another study, women who reported ‘sometimes’ adding salt to their food after cooking displayed better cognitive function than those who never added salt.¹¹⁰ On the contrary, ‘usually’ adding salt during cooking was associated with poorer cognitive function in men.¹¹⁰ However, salt intake was not properly measured, and cognitive function was assessed by a brief telephone-based method.¹¹⁰ Other studies have reported no association between higher dietary sodium intake and cognitive decline.^111,112 Similarly, sodium intake did not correlate to the presence of cerebrovascular alterations including white matter hyperintensities (WMHs) or cortical atrophy.¹¹¹ However, these results were recently challenged by a study demonstrating that the risk for progression of CSVD was significantly higher in the participants with higher salt intake.²² Another recent study further suggested that HSD may increase dementia risk in older adults independently of known risk factors, including hypertension and apolipoprotein E genotype.²⁵ The effect was present even after adjusting for other confounders including smoking, blood lipids, fasting plasma glucose, and baseline WMH volume.²⁵ Lastly, in a Korean cohort of patients with mild to moderate Alzheimer's Disease (AD), reducing salt intake was found to slow the progression of the disease over the 3-year follow-up.¹¹³ These data indicate that salt may not only increase the risk of dementia but could also have detrimental effects on the progression of the disease.

6. HSD, gut immunity, and IL-17

The effects of salt on immune cells have been known since almost three decades.^114,115 However, over the last 10 years, several studies have contributed to elucidate the underlying mechanisms.^116–129 For example, it has been demonstrated that increased extracellular sodium boosts the activity of pro-inflammatory¹¹⁸ and inhibits that of anti-inflammatory macrophages.¹¹⁹ Furthermore, HSD has shown to profoundly affect the adaptive immune system as demonstrated by the enhancement of pathogenic Th17 cell differentiation in the gut of HSD mice.^116,117 Th17 cells are highly abundant in barrier tissues including the intestine and the skin and produce IL-17 which is critical in regulating mucosal immunity.¹³⁰ Dysregulation of Th17 cells has been associated with several inflammatory conditions including autoimmune diseases and cancer.^{116,117,130–135} Two studies published in 2013, reported that, in mice, HSD was associated with a marked increase in the frequency of Th17 cells in the gut lamina propria, whereas no significant changes were observed in the mesenteric lymph nodes or spleen.^116,117 Accordingly, the ability of Th17 cells to produce IL-17A was also increased by increasing concentrations of sodium in vitro.^116,117 p38/MAPK, the osmosensitive transcription factor NFAT,¹¹⁶ and the serum glucocorticoid kinase 1 (SGK1),¹¹⁷ which controls sodium transport and salt homeostasis in non-immune cells, were identified as critical for inducing the Th17 cell phenotype (Figure 3). Importantly, the severity of inflammatory diseases associated with Th17 response, including colitis and experimental autoimmune encephalomyelitis, is also exacerbated by HSD.^{116,117,131–135} A positive correlation between sodium intake and increased clinical and radiological disease activity has been also demonstrated in multiple sclerosis patients.¹³⁶ However, this association needs to be further investigated.¹³⁷ Sodium has a direct effect on T cells, and salt-induced Th17 response does not seem to be mediated by dendritic cells (DCs).¹³⁸ Nevertheless, there is evidence that elevated sodium is associated with DC activation via SGK1, NLRP3 inflammasome, NADPH oxidase, and superoxide production.^139–141 More recent data further demonstrated that high-salt activates DCs via the p38/MAPK and STAT1 signalling pathways exacerbating pathology in a murine model of systemic lupus erythematosus, a chronic inflammatory autoimmune disease.¹⁴² Although salt seems to exert a direct effect on T cells and DCs, there is also evidence that changes in gut microbiota composition associated with HSD may influence the Th17 response¹²² and may play a role in DC activation.¹⁴³ HSD suppressed intestinal Lactobacillus murinus levels, leading to reduced availability of indole-3-lactic acid (ILA), a tryptophan metabolite, which, in vitro, suppressed Th17 cell response (Figure 3).¹²² In addition, levels of Prevotellaceae, which has been associated with Th17 response,¹⁴⁴ were increased in the caecum of HSD mice.¹⁴³

The effects of salt on immune cells are not limited to Th17 cells and DCs. HSD also alters the differentiation and function of other T helper cell subtypes, including T_regs, T_h2, and follicular helper T cells.^{120,121,124–127,129,145} Remarkably, the changes induced by elevated salt in T_regs resemble those found in T_regs of patients with autoimmune diseases,¹⁴⁶ further suggesting that the dysregulation of the immune system induced by HSD could exacerbate the severity of these chronic inflammatory conditions. Like Th17 cells, the effects of salt on T_regs are dependent on SGK1.^{120,121,125,127} In addition, β-catenin signalling has emerged as a key player in the interferon-γ (IFN-γ)/IL-10 imbalance induced by high salt in T_regs.¹²⁴ More recent data also demonstrated that elevated extracellular sodium may alter T_regs function by inhibiting oxidative phosphorylation and disrupting mitochondrial respiration.¹²⁹

Consistent with these observations, we also found that HSD altered the balance between Th17 and T_regs in the lamina propria of the distal small intestine and increased IL-17A circulating levels.⁵³ Remarkably, we also demonstrated that the gut Th17 response contributes to the cerebrovascular and cognitive dysfunction associated with HSD.⁵³ Indeed, the harmful effects of salt on the brain were absent after genetic deletion of IL-17A or administration of anti-IL-17A blocking antibodies, strongly indicating that IL-17A plays a key role in mediating the effects of HSD on the cerebral vasculature.⁵³ Furthermore, administration of recombinant IL-17A reproduced the effects of HSD in mice fed a normal diet.⁵³  In vitro studies using mouse brain ECs confirmed the deleterious effects of IL-17A on endothelial function.⁵³ IL-17A induced eNOS phosphorylation on Thr⁴⁹⁵ leading to a reduction in baseline NO production and suppression of the increase in NO produced by ACh.⁵³ Because these effects were prevented by the administration of the ROCK inhibitor Y27632, the data indicate that IL-17A may activate ROCK, one of the major kinases mediating inhibitory eNOS phosphorylation (Figure 2). Altogether, these data demonstrated that HSD induces a Th17 response in the gut that increases circulating IL-17A, which, in turn, acts on cerebral ECs to suppress endothelial NO production, leading to reduction in cerebral perfusion and cognitive dysfunction.⁵³

Our findings are in line with accumulating evidence demonstrating that IL-17A alters vascular function.¹⁴⁷ For example, IL-17A-deficient mice are protected from the endothelial dysfunction and the oxidative stress induced by ANGII in isolated aortic rings.¹⁴⁸ Furthermore, the ACh-induced relaxation is reduced in the aortas of IL-17A-treated mice.¹⁴⁹ In a recent study, we demonstrated that IL-17A derived from gut and dural T cells activates IL-17RA on cerebral ECs and border-associated macrophages and induces neurovascular uncoupling and cognitive impairment in DOCA-salt hypertension.¹⁵⁰ Inhibition or genetic deficiency of IL-17A also attenuated atherosclerosis in ApoE^−/− mice indicating that IL-17A could play a role in atherosclerosis.^151,152 IL-17A has powerful effects both on ECs^153–157 and vascular smooth muscle cells,¹⁵⁸ and the vascular effects of IL-17A are potentiated by other cytokines, including TNF-α.¹⁵⁶ Since, as mentioned above, IL-17A plays a key pathogenic role in autoimmune disease, a possibility is that it might also contribute to the cardiovascular risk associated with these conditions.¹⁵⁹ In support of this hypothesis, K14-IL-17Aind/+ mice, a model of severe psoriasis-like skin inflammation, presented increased IL-17A plasma levels, vascular oxidative stress, reduced vascular NO availability, and endothelial dysfunction.¹⁶⁰ Furthermore, because autoimmune diseases are also associated with increased risk of cognitive impairment,^161–164 another possibility is that the vascular effects of IL-17A could promote cognitive decline.⁵³ Indeed, there is also increasing evidence linking IL-17 to cognitive impairment.¹⁶⁵ Numbers of circulating Th17 cells and/or circulating IL17 are increased in patients with MCI¹⁶⁶ and AD¹⁶⁷ and are linked to AD progression.^168,169 More recently, levels of IL-17A were detected in post-mortem brain tissue from AD and vascular dementia subjects.¹⁷⁰ Evidence from animal studies also supports a role of IL-17A in cognitive impairment. IL-17A circulating levels are increased in the brain of 3xTg-AD mice,¹⁷¹ an AD mouse model, before evidence of tau pathology in the brain,¹⁷² and accumulation of meningeal ɣδT cells, a major source of IL-17A, promotes cognitive decline in 3xTg-AD mice.¹⁷³ Since IL-17A neutralization did not affect Aβ or phospho-tau levels, IL-17A could promote cognitive impairment by directly altering synaptic function.¹⁷³ This conclusion is supported by a recent study showing that IL-17A impaired synaptic plasticity and induced cognitive impairment in EAE mice.¹⁷⁴ Additionally, neutrophils surrounding Aβ deposits could release IL-17A, further suggesting a potential role in AD pathophysiology.¹⁷⁵ In line with this, neutralization of IL-17A rescued neuroinflammation and memory impairments induced by injection of Aβ into cerebral ventricles.¹⁷⁶ IL-17A also promoted Aβ accumulation in a model of post-operative cognitive impairment in mice¹⁷⁷ whereas Th17 cells have been found in the brain of APP/PS1 mice.¹⁷⁸ Altogether, there is considerable evidence linking IL-17A to cognitive impairment. However, additional studies will be needed to further elucidate the molecular and cellular mechanisms by which IL-17A alters cerebrovascular and cognitive function.

7. Conclusion

The evidence presented above demonstrates that HSD has profound effects on vascular function and may contribute to cognitive impairment both in animal models and in humans. Remarkably, these effects are independent of the BP elevation, indicating that the deleterious effects of HSD are not secondary to the effect of HSD on systemic BP. Although considerable progress has been made in the understanding of the molecular mechanisms underlying the vascular and cognitive effects of HSD, several questions remain to be addressed.

Oxidative stress has emerged as the main culprit for the deleterious effects of HSD on vascular function. However, we have demonstrated that suppression of eNOS activity is an additional mechanism by which HSD could compromise tissue perfusion and vascular reactivity. Whether other mechanisms are involved remains to be defined. A recent study by Bailey et al.¹⁷⁹ shed light on the transcriptional changes induced by HSD in the brain of WKY rats revealing profound changes in genes linked to neurovascular function even in the absence of BP increase. Although this study is missing cellular resolution, these transcriptome data further attest to the powerful effects of HSD on neurovascular function. The development of single-cell RNA sequencing approaches could open the way to a better understanding of the effects of HSD on the different cellular components of the neurovascular unit. In addition, because of the differences in the transcriptome of arterial, capillary, and venous brain EC¹⁸⁰ and because of EC heterogeneity across multiple tissues,¹⁸¹ these studies could further advance our understanding on the effects of HSD on the vasculature and potentially reveal tissue specific mechanisms contributing to vascular dysfunction.

Another major question that needs to be addressed is the translational relevance of the preclinical findings. Although most people consume too much sodium, HSD commonly used in animal studies constitutes an approximation of the highest levels of human salt consumption which are only found in certain regions of Central and East Asia.¹ Furthermore, modelling human HSD in rodents is complicated by major species differences in salt absorption, turnover, utilization, and metabolism. The uncertainty in estimating minimal salt requirements in mice, the relatively short exposure in animal models compared to life-long exposure in humans, and the well-known underestimation of human salt consumption makes the interpretation of animal findings even more challenging.¹⁸² For all these reasons, preclinical data need to be interpreted with caution and additional clinical validation will be required to establish their relevance to human diseases.

This is particularly true for the link between excessive salt consumption and cognitive impairment in humans. Additional studies will be required to demonstrate a more direct association. The inclusion of novel cerebrospinal fluid, neuroimaging, or blood biomarkers of cognitive impairment should be considered. The evaluation of cognitive function in large epidemiological studies investigating the effects of salt on cardiovascular health could also be considered and would vastly increase the amount of data available to draw more confident conclusions regarding the effects of salt on cognitive function. Finally, though the impact of salt on cognition may be relatively limited compared to other risk factors, reducing its intake may be a cost-effective strategy to help preserve not only vascular but also cognitive function.

Similarly, the link between salt intake and immune dysregulation, which has been consistently found in mice, will have to be further investigated in humans. Because dysregulation of peripheral immune system may contribute to cognitive decline,¹⁸³ a better understanding of the link between salt and immune function could help in further elucidating the effects of salt on cognition. As for the role of IL-17A in mediating cognitive impairment, additional studies will be needed to elucidate its cellular and molecular brain targets and to further support its significance in humans consuming excessive amounts of salt.

Acknowledgements

Figures were created with the BioRender scientific illustration software.

Funding

This work was supported by the following grants: RO1NS130045 – 01 (National Institute of Neurological Disorders and Stroke [NINDS]) and CAF211776-01 (Cure Alzheimer’s Fund).

Data availability

There are no new data associated with this article.

References

Author notes