Hypertension, neurodegeneration, and cognitive decline

Abstract

Elevated blood pressure is a well-established risk factor for age-related cognitive decline. Long linked to cognitive impairment on vascular bases, increasing evidence suggests a potential association of hypertension with the neurodegenerative pathology underlying Alzheimer’s disease. Hypertension is well known to disrupt the structural and functional integrity of the cerebral vasculature. However, the mechanisms by which these alterations lead to brain damage, enhance Alzheimer pathology, and promote cognitive impairment remain to be established. Furthermore, critical questions concerning whether lowering blood pressure by antihypertensive medications prevents cognitive impairment have not been answered. Recent developments in neurovascular biology, brain imaging, and epidemiology, as well as new clinical trials, have provided insights into these critical issues. In particular, clinical and basic findings on the link between neurovascular dysfunction and the pathobiology of neurodegeneration have shed new light on the overlap between vascular and Alzheimer pathology. In this review, we will examine the progress made in the relationship between hypertension and cognitive impairment and, after a critical evaluation of the evidence, attempt to identify remaining knowledge gaps and future research directions that may advance our understanding of one of the leading health challenges of our time.

Introduction

The impacts of elevated blood pressure (BP) on brain health significantly contribute to cognitive decline over the life course^1–3. Although associations between chronically elevated BP and cognitive impairment were noted by the end of the 19^th century, systematic investigations into the relationship did not begin in earnest until the 1940s-1960s. The introduction of the Kety and Schmidt method to measure cerebral blood flow (CBF) led to the discovery that cerebrovascular resistance is increased in hypertensives⁴ and correlates with hypertensive retinopathy⁵. Shortly thereafter, pioneering observations of diminished psychomotor speed in hypertensive air traffic controllers by Spieth⁶ and intellectual decline in patients with hypertension by Wilkie and Eisdorfer⁷ provided the initial evidence associating chronically elevated BP with worsened cognition.

More recently, hypertension has emerged as a pathogenic factor both in cognitive impairment on vascular bases and in Alzheimer’s disease (AD)⁸. This hypertension-dementia relationship has led to hypertension being targeted as a treatable condition that could delay the onset of cognitive deterioration. Longitudinal studies indicate preservation of cognition following adequate BP control⁹, particularly in intensive lowering regimens¹⁰, although such outcomes are not universally observed¹¹. Results from investigations into the differential efficacy of antihypertensive agents in combating cognitive decline are similarly conflicting^12–14. Consequently, emphasis has been placed on the identification of early biomarkers of hypertension-induced brain damage and novel therapeutic targets. To this end, recent studies have focussed on identifying the white matter tracts and brain regions affected in hypertension, the mechanisms by which hypertension impairs cognition, and the nature of the association of hypertension with AD pathology.

The present review will explore the consequences of hypertensive sequelae on cerebrovascular health, cognition, and dementia risk; the associations between hypertension and AD pathology; and potential therapeutic targets and strategies to mitigate the impacts of hypertension on cognitive decline. Finally, we will attempt to highlight outstanding questions that remain unaddressed.

Hypertension and cognitive decline: epidemiological evidence

BP trends over the life course, as observed in the Framingham Heart Study¹⁵, suggest an age-related pattern of systolic (SBP) and diastolic (DBP) BP changes in hypertension. Characterized initially by increasing DBP and SBP in early adulthood, DBP falls in late-life concurrent with continual elevation of SBP. Accordingly, mean arterial pressure progressively increases until reaching an asymptote around the seventh decade, whereas pule pressure, beginning in middle-age, steadily rises throughout the remaining lifespan¹⁵.

The hypertension-cognition relationship changes over the life course (for in-depth reviews, see references^16,17). Thus, elevated BP in the 4^th to 5^th decade of life is associated with late-life cognitive deterioration^1–3, while a U-shaped relationship is noted in the elderly where both high and low BP portend cognitive impairment¹⁸. Of interest, elevated SBP through mid-life followed by a rapid drop in late-life is associated with severe cognitive decline¹⁹ and brain atrophy²⁰, suggesting that late-life reductions in BP may be driven by pathological brain changes impacting autonomic control, as has been advocated for AD²¹. Whether preventing such late-life BP decline would be beneficial remains to be established.

BP variability (BPV), defined as variations in BP stratified by very short term (beat-to-beat), short term (24-hours), mid-term (day-to-day), or long term (visit-to-visit)²² variability, has emerged as an additional contributor to cognitive impairment, particularly in the elderly. In recent meta-analyses combining 53 (9 very short-, 12 short-, 9 mid-, 23 long-term) and 19 (4 short-, 4 mid-, 11 long-term) studies, worse cognitive performance was associated with short, medium, and long term BPV^23,24. In contrast, low BPV over the very-short term elevates the risk of poor cognitive outcomes, perhaps due to impairment of the ability to adjust perfusion commensurate with demand (e.g., postural changes)²³.

In summary, hypertension, particularly in mid-life, is associated with worse long-term cognitive outcomes, and increased short term, mid-term, and long term BPV are emergent risk factors for cognitive decline and dementia, especially in the elderly.

Hypertension and the pathobiology of cerebrovascular impairment

In this section, we will first review basic concepts on the blood supply of the brain, then examine the consequences of hypertension on the structure and function of cerebral vessels.

The cerebral blood supply and the neurovasculome

The carotid arteries enter the skull and merge to form the circle of Willis, a collateral flow loop from which the anterior, middle, and posterior cerebral arteries originate²⁵. Branches of these vessels travel along the surface of the brain within the subarachnoid space, forming an anastomotic network (pial arteries)²⁵, before diving into the substance of the brain (penetrating arteries and arterioles) while ensheathed within an extension of the subarachnoid space delimited by the vascular basement membrane and glia limitans (perivascular space; PVS). As these penetrating arterioles advance into the brain parenchyma, the PVS disappears, and the vascular basal lamina and glia limitans merge²⁵. At this level, arterioles are surrounded by a single layer of smooth muscle cells which, as the vessels get smaller, becomes discontinuous until replaced by pericytes at the capillary level²⁵.

Interactions between cerebral vessels and neural elements are crucial determinants of CBF. Structural and functional associations between vascular cells, glia, and neurons, dubbed the neurovascular unit (NVU), regulate blood-brain barrier (BBB) maintenance and the adjustment of cerebral perfusion commensurate with local metabolic demands²⁵. Owing to the segmental specificity of these neurovascular interactions, the concept of the NVU has been recently expanded to include large extra- and intra-cerebral vessels, as well as meningeal vessels and lymphatics (the neurovasculome)²⁶.

Hypertension alters cerebrovascular structure

Hypertension is often preceded by arterial stiffening²⁷ which, as a result of diminished vascular compliance, drives an elevation in pulse pressure²⁷ that exposes the downstream circulation to increased tensile strain²⁸, triggering adaptive remodeling responses throughout the cerebrovascular tree in an attempt to preserve vessel integrity (Figure 1)^16,29. Remodeling is defined as inward or outward depending on whether luminal diameters increase or decrease, and eutrophic or hypertrophic based on the nature of changes to the vascular wall. Eutrophic remodeling is characterized by an altered luminal diameter sans modification of wall thickness²⁹, whereas increased wall thickness consequent to cellular hyperplasia/hypertrophy and deposition of extracellular matrix signifies hypertrophic remodeling²⁹. Although indeterminate, stiffening and remodeling involve the convergence of hemodynamic stress, endothelial dysfunction, immune infiltration, and inflammation^30,31, as well as the actions of overlapping mediators, such as cytokines, angiotensin-II (AngII), endothelin, and oxidative stress³².

Figure 1: Structural alterations and segmental pathology induced by hypertension.

Top: Vascular remodeling proceeds in either an outward or inward manner in response to mechanical, cellular, inflammatory, and oxidative factors. Bottom: The predominant vascular structural alterations associated with hypertension are indicated according to the affected segment of the neurovasculature. Key pathological outcomes are also depicted. Details are provided in the text. Abbreviations: WML, white matter lesion; ICA – internal carotid artery; MCA – middle cerebral artery.

Large extracranial and intracranial cerebral arteries

Hypertension contributes to formation of atherosclerotic plaques in both extracranial and intracranial cerebral arteries^16,33, increasing the risk of ischemic stroke^34,35. Extracranial lesions are characterized by elevated accumulation of lipids in vertebral and carotid arteries³⁶, increasing the risk for artery-to-artery embolism³⁶, while intracranial lesions affecting the circle of Willis and its branches display a fibrous expression more likely to result in local vascular occlusion³⁴. How hypertension promotes atherosclerosis remains incompletely understood, but AngII is thought to contribute by generating reactive oxygen species (ROS) and upregulating cytokines, chemokines, and growth factors³⁷. The resultant cellular and molecular milieu triggers an inflammatory response in the vascular wall that precipitates endothelial dysfunction, neointima formation, and atherosclerotic lesions³⁷.

In addition to atherosclerosis, concurrent vascular stiffening and stenosis resulting from sclerosis of the vascular wall (arteriosclerosis) negatively affects the cerebral blood supply and downstream microvasculature. These changes are implicated in CBF reduction³⁸; diminution of cerebrovascular reserves³⁹, which can elevate stroke risk⁴⁰; and increased hydrodynamic impacts on the microvasculature⁴¹. Modification of the extracellular matrix is a major contributor to arterial stiffening, and is characterized, in part, by collagen and fibronectin accumulation, metalloprotease-mediated elastin fragmentation, and pro-fibrotic cascades related to transforming growth factor-β³⁰.

Pial vessels and penetrating arterioles

Pial and penetrating arteries and arterioles are uniquely sensitive to chronic elevations in BP¹⁶. In humans, hypertension-induced microatheroma in pial arteries and small perforating arteries (300 to 800 μm external diameter^42–44) can occlude vessels and instigate lacunar infarction or microinfarction⁴⁴. Moving further into the brain, small penetrating arteries and arterioles, from 300 μm to 20 μm^42–44, arising from either the first segment of the middle cerebral artery or terminal branches of the pial arteries, converge on deep white matter territories¹⁶. As these vessels display vulnerability to shifts in upstream pressure, scant collateralization, and limited anastomosis¹⁶, they are strongly implicated in formation of the white matter damages attendant to vascular dysfunction^45,46.

Three vascular lesions predominate within these smaller vessels: lipohyalinosis, fibrinoid necrosis, and arteriolosclerosis. Lipohyalinosis involves asymmetric deposition of an amorphous, glass-like material composed of collagen and degenerated smooth muscle tissue into the vascular wall, while fibrinoid necrosis, observed in more advanced lesions, is signified by infiltration of fibrin and its degradation by-products^42–44. More common is arteriolosclerosis⁴³, characterized by myocyte degradation and loss of elastin alongside concentric accumulation of fibro-hyaline materials and collagens in the vascular wall, leading to stenosis and reduced elasticity (hyaline arteriolosclerosis)^42–44, or the extensive intimal fibromuscular proliferation, luminal narrowing, and necrotic degeneration often associated with malignant hypertension (hyperplastic arteriolosclerosis)^42–44. Capillaries and pre-capillary arterioles are also affected, marked by loss of endothelial cells and pericytes (string vessels), increased thickness of the basal lamina and accompanying tortuosity, fibrin deposition, and overall rarefaction⁴⁷.

Collectively, these changes form the basis of hypertension-associated small vessel disease (SVD), a disorder affecting arterioles, venules, and capillaries of the subcortical and periventricular white matter in a segment specific manner^42–44; atheromatous lesions are typically observed in larger arteries/arterioles situated upstream of distal perforating vessels supplying the deep parenchymal tissues, which are in turn affected by arteriolosclerosis, lipohyalinosis, and fibrinoid necrosis (Figure 1)^42–44. The sequelae of SVD, detectable via magnetic resonance imaging (MRI), include WML, small ischemic foci (lacunes and microinfarcts) and microhemorrhages⁴⁸, venous collagenosis⁴⁹, and enlarged PVS⁵⁰. SVD is also influenced by short term and long term BPV, evidenced by brain imaging and autopsy data (arteriolosclerosis, microinfarcts, WMLs, enlarged perivascular spaces, atherosclerosis of the circle of Willis,)^51,52.

Hypertension impairs cerebrovascular function

The brain is dependent on a continuous and highly regulated supply of oxygen- and glucose-rich blood. Accordingly, neurovascular mechanisms assure that perfusion is commensurate with the brain’s regionally and temporally diverse energetic and waste clearance requirements²⁵. In this section, the mechanisms regulating CBF are reviewed, focusing on factors disrupted in hypertension which may lead to cognitive impairment. A more comprehensive description of CBF regulation is found in recent reviews^25,26,53,54.

Autoregulation

The cerebral vasculature maintains CBF within approximately ±20 mmHg of baseline⁵³, dampening the BP changes attendant to daily life. Traditionally, autoregulatory responses have been studied using stepwise changes in BP coupled with measurement of CBF; the stabilization of perfusion during these manipulations is known as static autoregulation⁵⁵. In the 1990s, the introduction of transcranial Doppler flowmetry enabled assessment of flow velocity in response to rapid changes in BP⁵⁶, leading to the characterization of dynamic autoregulation⁵⁷, which refers to the latency between BP changes and corresponding adjustments to vascular resistance⁵³. These studies revealed that while rapid fluctuations in BP escape autoregulation and result in flow changes, they are not as large as if the autoregulatory process was absent.

On a cellular level, autoregulation depends on the ability of smooth muscle cells to constrict or dilate in accordance with intravascular pressure (myogenic response). Multiple mechanisms have been implicated, including (a) ion channels and mechanosensors, which adjust intracellular calcium levels commensurate with fluctuating transmural pressures⁵⁸, (b) alteration of contractile apparatus calcium sensitivity⁵⁹, and (c) depolarization-mediated amplification of intracellular calcium signaling⁶⁰.

Animal studies have provided evidence that hypertension shifts the static autoregulatory curve to the right, rendering the brain more susceptible to reductions in BP¹⁶. In contrast, most human studies suggest that both static and dynamic autoregulation are preserved among hypertensives⁵³, except for malignant hypertension (SBP: 180–260 mmHg), where failure of the autoregulation mechanism causes flow to passively follow BP⁶¹.

Overall, autoregulation appears to be a resilient mechanism that is largely maintained in human hypertension, but the available data have limitations that need to be considered. Since in most studies flow was assessed in a single vessel by transcranial Doppler, regional changes in dynamic autoregulation could have been missed. Furthermore, the inherent risk associated with lowering BP has been an obstacle to assessing the lower limit of static autoregulation in patients with hypertension.

Neurovascular coupling

Neurovascular coupling, which pertains to the distribution of blood to active brain regions commensurate with their metabolic demands²⁵, requires participation of all constituents of the NVU at each level of the cerebrovasculaure, and is driven by diffusible mediators (nitric oxide (NO), prostanoids, adenosine, ions, etc.) and segment-specific intrinsic vascular mechanisms²⁵. In brief, activation of neurons deep within the brain parenchyma initiates a series of vascular alterations which begin at the level of the capillary endothelium and are transmitted upstream in a retrograde fashion through intramural signaling, resulting in propagation of smooth muscle cell relaxation and attendant vasodilation^62–64. This retrograde propagation ultimately reaches larger pial vessels, which must relax to facilitate adequate blood flow to the activated region and prevent flow steal from neighboring vascular territories²⁵.

At present, a paucity of human studies have examined neurovascular coupling in hypertension. In untreated hypertensives, increased baseline BP is associated with diminished regional alterations to CBF during cognitive tasks⁶⁵. Similar impairments are reported for blood flow responses to visual stimuli in the retina⁶⁶ or posterior cerebral artery⁶⁷. These observations, albeit limited, support the presence of impaired functional hyperemia in hypertensives and, in turn, increased susceptibility to vascular insufficiency.

Endothelial cells: vasoregulation and the BBB

Cerebrovascular endothelial cells regulate vasomotor tone by releasing vasoactive messengers in response to select chemical and mechanical stimuli⁵⁴. Perhaps the most well-studied endothelium-dependent vasoactive mechanism is NO-mediated vasodilation in response to cholinergic agonists activating endothelial nitric oxide synthase (eNOS)⁶⁸. Signaling in a paracrine manner, NO reaches the vascular smooth muscle cells where it initiates cGMP-mediated relaxation of the smooth muscle⁶⁹. As mentioned above, the cerebral capillary endothelium contributes to functional hyperemia. Release of K⁺ by neuronal depolarization activates endothelial inward rectified K⁺ channels, leading to a hyperpolarization cascade transmitted upstream through inter-endothelial junctions which, upon reaching smooth muscle cells, culminates in their relaxation⁶².

Endothelial cells are also the site of the BBB⁷⁰. The bidirectional exchange of molecules between blood and brain is regulated by (a) molecular transporters on their surface, which gate substrate entry, (b) tight junctions between adjacent endothelial cells, which largely prevents paracellular solute flux, and (c) paucity of vesicular transport, which limits transcytosis⁷⁰.

In humans, endothelial dysfunction in peripheral arteries precedes BP elevation and correlates with the severity of hypertension post-development⁷¹. While direct in vivo evaluation of cerebral endothelial function in humans is not feasible, inhibition of NO synthesis does not attenuate arteriolar blood flow in the retina of hypertensive patients, suggesting NO deficiency concomitant with endothelial dysfunction⁶⁶, and endothelium-dependent vasodilatory responses to acetylcholine are diminished in arteriolar samples collected post-mortem from SVD/AD mixed pathology patients⁷².

These endothelial aberrations may affect cognition: impaired endothelium-dependent vasodilation in peripheral arteries is associated with brain microhemorrhages⁷³ and WMLs⁷⁴, both of which are implicated in cognitive decline^75,76; and some hypertensive patients present with BBB impairment, which can contribute to brain damages and SVD (e.g., WMLs)⁷⁷.

Perivascular spaces, waste clearance and neuroimmune regulation

PVS and the vessels within have emerged as major routes for the elimination of potentially-deleterious by-products of brain activity, such as amyloid-β (Aβ) and tau⁷⁸. Several clearance systems have been proposed⁷⁸. A transvascular pathway facilitates the passage of select molecules through the vascular wall by way of abluminal transporters⁷⁰. Since the brain parenchyma lacks conventional lymphatic vessels, other pathways are thought to use the PVS as a conduit to carry solutes out of the brain. In the perivascular pathway, solutes are thought to exit the parenchyma through retrograde transport along the vascular basal lamina or PVS, eventually mixing with the CSF in the subarachnoid space and draining into the cervical lymph nodes⁷⁹. Running opposed to this route is the glymphatic pathway. Aided by aquaporin-4 (AQP4) water channels in astrocytic endfeet, convective and diffusive forces are hypothesized to drive CSF from the periarterial space into the parenchyma where it blends with the interstitial fluid and picks up waste products. This waste rich fluid is thought to exit into the perivenous space, travel to the subarachnoid compartment, and drain into the meningeal lymphatics⁸⁰.

Cerebral vessels and PVS are also involved in immune regulation, and the constituent vessels of the cerebrovasculature are surrounded by an assortment of immune cells. Yolk sack derived myeloid cells, collectively termed border associated macrophages, populate the meninges, PVS, and choroid plexus⁸¹, and have emerged as key correlates of the neurovascular and cognitive dysfunction associated with hypertension^82,83.

In humans, hypertension is linked to enlarged PVS^84,85. Although incompletely understood, a proposed mechanism invokes hypertension-induced stiffening of large arteries exposing the microvasculature to enhanced pulsatile pressures⁸⁶, which increases the mechanical stress under which the PVS is placed and drives their enlargement⁸⁵. In support, a recent neuroimaging study found that the glymphatic removal of an intravenously administered tracer was reduced in SVD-positive participants, most of whom were hypertensive⁸⁷. This hypertension-clearance dynamic is discussed further in Impaired clearance of amyloid and tau peptides.

Hypertension and neurovascular dysfunction: mechanistic studies

Angiotensin II

Arguably the most well-established models of hypertension involve use of pressor doses of AngII that elicit a sharp increase in BP, or sub-pressor doses (slow-pressor model) which drive a gradual elevation in BP over several days. The slow-pressor model has gained in popularity for its potential to capture the progressive increase in BP of primary hypertension⁸⁸. Neurovascular dysfunction, i.e., impaired functional hyperemia and endothelial vasodilation, is provoked by both acute and chronic AngII treatment^82,89,90, and is driven by the interaction of AngII with one of its receptors, AngII receptor type 1 (AT1R), on vascular and perivascular cells (Figure 2)^82,89, as well as brain structures involved in autonomic regulation, such as the subfornical organ and periventricular hypothalamus⁹⁰.

Figure 2: Potential mechanisms of neurovascular dysfunction in AngII hypertension.

Circulating AngII interacts with AT1R on endothelial cells to initiate BBB disruption (tight junction remodeling and suppression of MFSD2A, leading to increased transcytosis) and enable its entry into the PVS. Next, circulating and brain-derived AngII engage with AT1R on PVM, leading to NOX2 activation, vascular oxidative (superoxide) and nitrosative (peroxynitrite) stress, reduced NO, further BBB disruption, and neurovascular dysfunction. Abbreviations: NO, nitric oxide; SMC, smooth muscle cell; EC, endothelial cell; PVS, perivascular space; PVM, perivascular macrophage; ROS, reactive oxygen species; BBB, blood-brain-barrier; AngII, angiotensin II; NOX2, NADPH oxidase 2; AT1R, angiotensin II receptor type I; TJ, tight junction.

Intriguingly, the effects of AngII on neurovascular function may be independent of BP elevation. Direct neocortical application of AngII, which does not increase BP, elicits neurovascular uncoupling⁸⁹; non-pressor doses of AngII induce neurovascular dysfunction⁹¹; and topical application of the AngII receptor blocker (ARB) losartan to the neocortex rescues functional hyperemia without lowering BP⁸⁹. Additional evidence for BP independent effects is provided by the BP high mouse (BPH)⁸² and SHR⁹², genetic models of lifelong hypertension, as well as the deoxycorticosterone acetate+salt (DOCA-salt) model of salt-sensitive hypertension⁹³, in which neurovascular dysfunction can be rescued despite increased BP^82,91,93. Furthermore, elevations in BP provoked by an α-adrenergic agonist fail to induce neurovascular dyfunction⁸², suggesting that not all pressor agents will affect these regulatory mechanisms. While these experimental observations highlight the BP-independence of neurovascular dysfunction, the pathogenic impacts of elevated BP in human hypertension cannot be underestimated. After all, antihypertensive intervention dramatically lowers stroke risk and⁹⁴, in some studies, improves cognitive function^10,95; however, as the data on the efficacy of BP lowering against cognitive decline is conflicting (see Therapeutic Interventions), BP-independent effects warrant further consideration.

Cytokines

In recent years, interleukin-17 (IL-17), a proinflammatory molecule, has received increased attention for its contributions to neurovascular dysfunction. In the AngII⁹⁶ and DOCA-salt⁹³ models, IL-17 is required for the development of neurovascular dysfunction and cognitive impairment, whereas neutralization of IL-17⁹⁶, inhibition of its receptor⁹⁶, or elimination of its cellular sources⁹³ prevents such outcomes. Intriguingly, IL-17 knock-out mice treated with DOCA-salt develop increased BP without neurovascular dysfunction⁹³, attesting to the BP-independent effects of hypertension-associated mediators.

Oxidative stress

Oxidative stress is a principal driver of the cerebrovascular sequelae of hypertension¹⁶. While several enzymatic sources have been identified⁹⁷, a NOX2 containing NADPH oxidase appears to be the prime source of the ROS involved in neurovascular dysfunction, particularly in AngII-induced hypertension^82,89, but also in other models, such as salt-sensitive⁹³ and in BPH mice⁸². Accordingly, both genetic deletion and pharmacological inhibition of NOX2 prevents neurovascular dysfunction and/or cognitive impairment in models of hypertension^82,89,93. However, how ROS precipitate these alterations requires elucidation. In AngII hypertension, peroxynitrite, a diffusible reaction product of NOX-2-derived superoxide with NO⁹⁸, is the ultimate mediator of the neurovascular dysfunction⁹⁹. Alternative mechanisms may include NO scavenging, and redox modifications of proteins, lipids and DNA, among others⁹⁸.

Innate immunity: Perivascular macrophages and microglia

Perivascular macrophages (PVM) are found closely apposed to the outer vascular wall of intracerebral arterioles and venules within the PVS⁸¹. PVM express both AT1R and IL-17RA, are richly endowed with the ROS producing enzyme NOX2^100,101, and have emerged as key players in the neurovascular and cognitive dysfunction observed in animal models of hypertension^82,93. In the slow-pressor model, AngII disrupts the BBB and subsequently enters the PVS where it interacts with AT1R expressed on PVMs to elicit vascular impairments, as evidenced by the restoration of neurovascular function which follows PVM depletion or deletion of At1r or Nox2 from PVM⁸². The disruption of the BBB is mediated by remodeling of tight-junctions and an increase in vesicular transport consequent to suppression of Mfs2da¹⁰², a negative regulator of transcytosis¹⁰³. Although activation of endothelial AT1R initiates the BBB dysfunction, PVMs are required for its full expression (Figure 2)¹⁰². In addition, depletion of PVMs abrogates the neurovascular dysfunction and cognitive impairments attendant to life-long hypertension in BPH mice⁸², as well as to salt-sensitive hypertension in the DOCA-salt model⁹³, supporting the inter-model generalizability of their contributions.

Less is known on the role of microglia. In the slow-pressor model, microglia acquire a proinflammatory phenotype at sites of BBB leakage, and their depletion (along with PVM) by the colony-stimulating factor 1 receptor inhibitor PLX5622 partially curtails cognitive impairments without rescuing BBB integrity¹⁰⁴. However, how microglia promote cognitive impairment independently of BBB disruption and whether PVMs are also involved remains to be determined.

Adaptive immunity: T-cells

T cells, namely of the T helper 17 (T_h17) and γδ17 subtypes, and IL-17, their principal effector cytokine¹⁰⁵, have emerged as crucial players in the induction of hypertension¹⁰⁶ and its sequelae, both centrally⁹³ and peripherally¹⁰⁶. Concerning central immunity, a recent study using the DOCA-salt paradigm revealed T cell-dependent mechanisms of neurovascular dysfunction (Figure 3): in the circulation, IL-17 derived from gut localized and circulating T_h17 and γδ17 cells acts on cerebral endothelial IL-17RA to reduce NO production and disrupt endothelial vasoactivity; in the brain, IL-17 secreted from γδ17 cells situated within the dura acts on IL-17RA in PVMs to induce vascular oxidative stress and concomitant neurovascular uncoupling. While endothelial IL-17RA deletion partially restores cognition, antagonizing the central sources (T-cells) or targets of IL-17 (PVMs) rescues cognition in full⁹³, revealing a previously unappreciated involvement of meningeal immunity in the cognitive effects of salt-sensitive hypertension.

Figure 3: Putative contributions of meningeal immunity to neurovascular dysfunction in salt-sensitive hypertension.

DOCA-salt treatment leads to production of IL-17 from T cells in the small intestine and dural immune compartment. Gut-derived circulating IL-17 interacts with IL-17RA on endothelial cells to disrupt endothelial function. IL17 produced in the meninges enters the subarachnoid space, travels to the PVS, and engages with IL-17RA on PVMs to induce vascular oxidative stress (via NOX2) and neurovascular dysfunction. Abbreviations: NO, nitric oxide; SMC, smooth muscle cell; EC, endothelial cell; PVS, perivascular space; PVM, perivascular macrophage; ROS, reactive oxygen species; IL-17, interleukin 17; Nox2, NADPH oxidase 2; IL-17RA, interleukin 17 receptor A; T_h17, T helper 17 cell; IL17γδT – IL-17-producing gamma delta T cell.

How does hypertension promote cognitive decline?

The most profound impacts of hypertension on cognition appear to center on executive function², motor speed, and attention¹⁰⁷, domains classically associated with vascular cognitive impairment¹⁶. Typically, memory is involved more so in AD-related cognitive decline than in vascular cognitive impairment, although this is more of a trend than a rule (see ref.¹⁰⁸).

In this section, the potential mechanisms underlying hypertension-induced cognitive impairment are discussed, with a particular focus on two hallmarks of vascular cognitive impairment: ischemic and hemorrhagic brain lesions and white matter disease¹⁶.

Ischemia, hemorrhage, white matter lesions and atrophy

Hypertension is major risk factor for both hemorrhagic and ischemic stroke which, in turn, is associated with a 3–6 fold increased risk of cognitive impairment¹⁶. Lacunar infarcts, microinfarcts, and microbleeds similarly portend cognitive deterioration^109–111.

WML are among the most common lesions associated with hypertension¹¹². SVD, often a consequence of hypertension¹¹³, is a major contributor to WML burden¹¹⁴, and associations between SVD, WML, and cognitive decline have been reported in the epidemiological literature. In the Radboud University Nijmegen Diffusion Tensor and Magnetic Resonance Cohort (RUN DMC), baseline SVD severity and progression were independently correlated with dementia risk over 14 years¹¹⁵; SVD progression over time was accelerated by baseline vascular risk factors, a phenomenon primarily driven by hypertension¹¹⁶; and reduced processing speed was associated with WML load⁷⁵. Supporting these observations, WML risk loci overlapped with BP traits in a meta-analysis of genome-wide association studies, with subsequent Mendelian randomization suggesting a causal association of WML volume with genetically predicted SBP¹¹⁷.

Mechanistically, hypoxia-ischemia following occlusion or rarefaction of the microvessels supplying the subcortical white matter likely contributes to white matter disease consequent to SVD^16,118. Indeed, reduced cerebrovascular reactivity to CO₂ was observed to precede the progression of normal-appearing white matter to WML in middle aged subjects with moderate-to-severe WML burden at baseline¹¹⁹; arteriolosclerosis and a hypoxic milieu have been found in WMLs post-mortem¹²⁰; and increased severity of arteriolosclerosis in watershed regions located between arterial territories has been shown to associate with greater WML and neurofibrillary tangles at autopsy¹²¹. However, reverse causation has been suggested¹²² wherein reductions in blood supply could be consequent to BBB dysfunction-induced damage of the microvasculature via extravasation of neurotoxic molecules, edema, and microvascular compression¹²³, or by WMLs lowering local CBF requirements secondary to deafferentation, tissue atrophy, and reduced energy demands¹²⁴.

Disrupted functional connectivity

Hypertension-induced loss of white matter integrity and consequent derangement of functional connectivity may affect cognition. Continued progression of WMLs¹²⁵, microstructural alterations¹²⁶, and network functional deficits^127,128 correlate with cognitive decline in hypertensive patients^126–128. In recent years, location-based image analyses have revealed that strategically placed WMLs are particularly relevant to cognitive decline¹¹⁰. In a study employing white matter tractography and resting state fMRI¹²⁶, hypertension was associated with early alterations in the superior longitudinal fasciculus, the forceps minor, and the anterior thalamic radiation. Hypertensive patients performed worse in the domains ascribable to these affected tracts in subsequent cognitive testing (executive functions, processing speed, and memory)¹²⁶. These observations are corroborated by an analysis combining imaging in middle aged adults from the UK Biobank cohort with genetic causal inference approaches, which found elevated SBP to associate with alterations to the external capsule, anterior corona radiata, and anterior thalamic radiation, the latter two of which are linked to cognitive decline¹²⁹.

Potential contributions of hypertension to AD

Once considered to be purely a neurodegenerative disease, an increasing body of evidence indicates the involvement of vascular brain lesions in AD, such that most cases diagnosed clinically present mixed vascular and AD pathologies^8,130. Relative to age-matched controls, AD brains display elevated intracranial atherosclerosis and numerous microvascular alterations¹³¹. Furthermore, BBB permeability is increased¹³² and hemodynamic responses to neuronal activity are suppressed¹³³ in the prodromal period, implicating vascular factors early in disease progression.

Epidemiological associations of hypertension with AD biomarkers

Amyloid positron emission tomography (PET) data concerning the link between amyloid pathology and hypertension are conflicting. In the ARIC cohort, heightened vascular risk factor burden in mid-life was associated with elevated PET Aβ in old age; however, hypertension alone did not significantly correlate with amyloid burden¹³⁴. In the British 1946 birth cohort, increased BP in middle age associated with late-life WML burden and brain atrophy, but not PET Aβ¹³⁵. While these findings suggest that hypertension fails to exacerbate Aβ burden, this does not preclude the possibility that amyloid pathology in some of the hypertensive patients was driven, in part, by vascular contributions. Of note, the impact of hypertension on Aβ may be modulated by genetic risk factors. In carriers of the ε4 allele of the apolipoprotein E gene, mid-life vascular burden correlated with increased brain amyloid in ARIC participants at 20-year follow-up¹³⁴, while in the Rotterdam study, hypertension was associated with Aβ burden 7 years after initial assessment¹³⁶.

As for tau, increased pulsatility secondary to aortic stiffness correlated with tau burden in the rhinal and entorhinal cortices of dementia-free subjects from the Framingham Heart Study¹³⁷, and a clear age-related pattern of BP was delineated in the Chinese Alzheimer’s Biomarker and LifestylE (CABLE) cohort, where higher mid-life SBP, late-life lower DBP, and increased pulse pressure in both mid- and late-life was associated with tau-related biomarkers and cognitive dysfunction, but not amyloid burden¹³⁸.

Intriguingly, inconsistencies in BP might be important for both amyloid and tau pathologies. An association of BPV with phosphorylated tau (p-tau) and decreased Aβ was observed in the CSF of older participants from the Alzheimer’s Disease Neuroimaging Initiative¹³⁹.

Overall, these observations present a plausible association of hypertension with AD pathology, particularly at the level of tau, although it remains unclear whether this relation is causative or incidental, or if the effects of hypertension on cognition involve or simply co-occur with AD pathology.

Hypertension and AD pathology: mechanistic considerations

Amyloidogenesis and tau hyperphosphorylation

Hypertension promotes tau hyperphosphorylation and Aβ accumulation in animal models (Figure 4). In transgenic models of brain amyloid build up, AngII-induced hypertension increases Aβ deposition¹⁴⁰ and p-tau immunoreactivity¹⁴¹. Similar findings are reported in other models, such as SHR¹⁴². While the underlying mechanisms have yet to be elucidated, AngII-induced elevation of β-¹⁴⁰ and γ-secretase¹⁴³ activity, which shifts APP processing toward β-amyloidogenesis, could be involved. Experimentally, genetic ablation of AT1R curtails γ-secretase activity and subsequent Aβ generation¹⁴³, and AT1R blockers abrogate amyloid pathology: telmisartan reduces Aβ formation in stroke-resistant SHR¹⁴⁴ and valsartan diminishes Aβ burden in Tg2576 mice¹⁴⁵. Regarding tauopathies, AngII¹⁴⁶ and IL-17^93,147 disrupt eNOS-mediated NO production and reduce NO bioavailability. Since suppression of endothelial NO has been linked to tau phosphorylation via activation of cyclin-dependent kinase 5¹⁴¹, AngII and IL-17 could promote p-tau through this mechanism, although supporting evidence is missing.

Figure 4: Potential mechanisms associating hypertension and AD.

The oxidative and inflammatory sequela of hypertension could promote AD pathology by increasing Aβ and p-tau (A), and disrupting perivascular and glymphatic clearance (B). Hypertension may enhance Aβ accumulation through increased processing of APP by secretase enzymes. Tau phosphorylation may be elevated under hypertensive conditions consequent to reductions in endothelial NO bioavailability and attendant activation of cyclin-dependent kinase 5. Abbreviations: SMC, smooth muscle cell; EC, endothelial cell; PVS, perivascular space; AngII, angiotensin II; IL-17, interleukin 17; AQP4, aquaporin 4; Aβ, amyloid-β; APP, amyloid precursor protein; sAPPβ, soluble peptide APPβ; BM, basement membrane; p-tau, phosphorylated tau.

Impaired clearance of amyloid and tau peptides

Aβ clearance is facilitated by the glymphatic and perivascular pathways^148,149, as well as utilization of transporters situated within the BBB , namely the low-density lipoprotein receptor-related protein 1 (LRP1) for egress and the receptor for advanced glycation end products (RAGE) for ingress⁷⁰; thus, disruption of these systems would be expected to worsen amyloid accumulation.

Hypertension-induced alterations in perivascular pumping, a fluid transport mechanism wherein vascular wall kinetics drive CSF flow in the PVS¹⁵⁰, could aggravate proteinopathy by disrupting Aβ and tau clearance. In a study using particle tracking velocimetry concurrent with measurements of arterial diameter, acute elevations in BP induced by intravenous infusion of AngII in mice altered arterial wall motion during the cardiac cycle, leading to reduced CSF flux in the PVS consequent to increased backflow¹⁵⁰. This diminished CSF flow within the PVS could precipitate a similar decrease in glymphatic transport, a phenomenon recently demonstrated in the SHR model using dynamic contrast-enhanced MRI-aided quantification of the glymphatic transport of a contrast agent¹⁵¹. These perturbations are likely to diminish Aβ and tau removal. Indeed, TGN-020-mediated inhibition of AQP4 in astrocytic end-feet¹⁴⁹, which are crucial for glymphatic flow⁸⁰, suppressed tau protein clearance in mice¹⁴⁹. Considering recent observations that changes in vasomotion caused by neurovascular coupling or optogenetically-induced vascular constriction-dilation facilitate CSF flow^152,148, the neurovascular uncoupling observed in experimental and human hypertension^67,82,89,93 could compromise metabolite clearance. Lastly, RAGE, which is activated in hypertensive models¹⁵³, modulates the entry of Aβ into the brain and influences β- and γ-secretase activity¹⁵⁴, suggesting that hypertension could promote Aβ accumulation by enhancing Aβ influx¹⁵³ and/or β-amyloidogenesis through RAGE induction¹⁵⁴.

Evidence for disruption of these pathways in human hypertension is limited. Elevated peripheral expression of extracellular RAGE binding protein was associated with increased risk of cognitive decline in the Rotterdam Study¹⁵⁵, suggesting a role for RAGE in AD pathogenesis; however, this cross-sectional association was attenuated in a longitudinal setting¹⁵⁵, indicating potential reverse causation. Human studies of SVD, in which the majority of participants were hypertensive, showed evidence of reduced glymphatic clearance¹⁵⁶, but a specific role of elevated BP has not been established.

Collectively, these clinical and basic observations suggest that hypertension could contribute to AD pathology through induction of amyloidogenesis and, based mainly on animal studies, diminution of amyloid/tau clearance (Figure 4), although the degree to which hypertension affects these pathways in humans requires elucidation.

Summary: how does hypertension promote cognitive deterioration?

Hypertension likely contributes to cognitive decline through a confluence of pathogenic mechanisms (Figure 5). First, vascular damages, inflammation, neurovascular dysfunction, and BBB disruption threaten the health of the cortical tissue and subcortical white matter. The resulting WMLs, microinfarcts, microbleeds, and disruption of metabolite clearance may contribute to network dysfunction, neuronal loss, and brain atrophy. Lastly, by fostering Aβ and p-tau accumulation, hypertension could conceivably contribute to the overlap between vascular and AD pathology to accelerate the development of cognitive impairment.

Figure 5: Hypertensive sequalae which underlie cognitive impairment.

Hypertension promotes structural alterations to the cerebrovasculature concurrent with NVU functional deficits. The resultant microbleeds, microinfarcts, and local hypoxia-ischemia drive neuronal loss and degradation of white matter tracts (especially in thalamo-cortico circuits), leading to brain atrophy and network disruption. Additionally, hypertension may provoke β-amyloidogenesis (upregulated secretase activity) and disrupt clearance of toxic metabolites, giving rise to Alzheimer’s-associated proteinopathies. Collectively, these events likely contribute to cognitive impairment. Abbreviations: WML, white matter lesion; CBF, cerebral blood flow; NVU, neurovascular unit.

Therapeutic interventions

A recent post-mortem investigation of samples taken from 4 separate birth cohorts (1905–1914, 19-15–1919, 1920–1924, and 1925–1930) found that although neurodegenerative pathologies did not differ by birth year, dementia incidence decreased over time concurrently with atherosclerosis and arteriosclerosis¹⁵⁷. The dramatic reduction in vascular pathology, possibly due to better control of vascular risk factors, including hypertension, could have contributed to the decrease in dementia incidence. Considering epidemiological associations of hypertension with dementia, elevated BP presents a particularly inviting modifiable target in the effort to diminish global dementia burden.

A role for antihypertensive therapy?

As previously discussed (see Hypertension and cognitive decline: epidemiological evidence), the relationship between hypertension and cognitive decline is most pronounced when considered in mid-life^1–3. Thus, while randomized clinical trials would best evaluate the efficacy of antihypertensive therapy in dementia prophylaxis, the temporal characteristics of the hypertension-dementia relationship necessitate reliance on long-term observational studies.

Several large cohort longitudinal analyses indicate an association of antihypertensive treatment with lesser cognitive decline^2,158,159. Consistent with the relationship of hypertension chronicity to dementia risk¹⁶⁰, therapy duration appears important in prophylaxis: in the Rotterdam cohort, each additional year of antihypertensive treatment, prior to age 75, reduced dementia risk by 8%¹⁵⁹.

Clinical trial data are comparatively conflicting, likely due to issues surrounding the age of treatment onset and duration of medication use. Indeed, the most recent update of the Cochrane Review found that pharmacological treatment of hypertension in patients without prior cerebrovascular disease provides only low certainty evidence for the prevention of dementia onset and cognitive deterioration¹¹, with investigators citing insufficient study duration as a probable confound. In contrast, a recent meta-analysis of 12 major clinical trials, including the recently completed SBP Intervention Trial – Memory and Cognition in Decreased Hypertension (SPRINT-MIND) study, reported a significant association of BP control with decreased risk of dementia or cognitive impairment¹⁶¹.

Intriguingly, intensive BP-lowering regimens may offer increased benefit relative to standard therapy. Recent findings from secondary analyses of the SPRINT-MIND data set indicate diminished risk of cognitive decline^10,95, white matter lesion development¹⁶², and probable dementia⁹ in patients undergoing intensive blood-pressure-lowering (SBP < 120 mmHg). Intensive regimens, but not standard, were also found to ameliorate cognitive decline in patients with high BPV¹⁶³. Overall, these findings are corroborated by a recent meta-analysis where late-mid-to-late-life BP correction dramatically mitigated dementia risk, with intensive lowering showing the most benefit¹⁶⁴. Taking into account the demonstrated safety of BP lowering interventions in the elderly¹⁶⁵, aggressive BP control warrants further consideration as a means to delay dementia onset.

Comparative assessments of differing antihypertensive agents have yielded inconsistent results. While some meta-analyses show no heterogeneity of antihypertensive class on risk of incident dementia¹², an emerging body of evidence seemingly supports improved outcomes for drugs that preserve AngII synthesis (such as ARBs) over those that inhibit it (ACE inhibitors)¹³. The BBB permeability of drugs may also be consequential, with recent reports suggesting BBB-permeable ARBs¹⁴ and β-blockers¹⁶⁶ are superior to their impermeable counterparts. Lastly, the recent development of a selective aldosterone synthase inhibitor¹⁶⁷ and powerful angiotensinogen silencing antisense oligos¹⁶⁸ could allow for blood pressure control in treatment resistant hypertensive patients.

Conclusions

The data presented herein show that hypertension remains among the most insidious factors influencing cognition over the lifespan, largely due to impacts on cerebrovascular structure and function which threaten the health of the brain¹⁶. Despite advances in the field, numerous questions remain regarding the temporality of hypertension-induced cognitive decline, the underlying mechanisms, and potential interventions.

• Because the best clinical evidence for preserving cognition in hypertension currently supports BP control, it is important to consider whether certain drug classes are more efficacious than others. While many studies fail to show heterogeneity among the antihypertensive classes¹², AngII-preserving drugs¹³ and BBB-crossing antihypertensives¹⁴ and β-blockers¹⁶⁶, as discussed in the previous section, warrant further exploration.

• Another important question is whether intensive BP control provides greater cognitive benefit than standard treatment. Results from SPRINT-MIND secondary analyses suggest this to be the case¹⁰, particularly in those with increased BPV¹⁶³. While this more aggressive approach does not compromise cerebral perfusion¹⁶⁵, the risk/reward ratio of intensive treatment requires further validation (will adverse effects emerge as treatment length increases, or in the elderly?), and it remains to be determined whether aggressive control in mid-life, when the hypertension-cognitive decline relationship is strongest^1–3, will still provide greater benefit.

• Does combating cognitive deterioration by controlling BP also reduce AD pathology? Since the neurohumoral dysfunction underlying hypertension may promote Aβ and p-tau accumulation, it would be of interest to assess if antihypertensive medications reduce AD pathology independently of effects on BP, as suggested by animal studies¹⁴⁵. If so, antihypertensives could be combined with Aβ immunotherapy, which has recently been approved by the FDA.

• Considering preclinical evidence that the cerebrovascular sequelae of hypertension are not entirely ascribable to elevated BP^82,89,93,102, further elucidation of the mechanistic underpinnings of neurovascular and cognitive impairment in hypertension are needed. Are immune cells a major driving force? In addition to the growing body of experimental evidence connecting PVMs to hypertensive pathology^82,93,102, the recently unveiled contributions of meningeal T-cells to neurovascular dysfunction in a model of salt-sensitive hypertension⁹³ links meningeal immunity to hypertension-induced cognitive impairment.

• What about pericytes, which may be involved in BBB regulation and cerebral perfusion¹⁶⁹? Reduced capillary pericyte coverage is observed alongside BBB disruption in slow-pressor AngII hypertension¹⁷⁰, and mouse models of pericyte deficiency exhibit impairment of the BBB¹⁷¹ and CBF regulation¹⁷². However, the precise contributions pericyte alteration to the neurovascular and cognitive dysfunction in hypertension remain to be determined and would be an area of interest for future studies.

• The events linking vascular dysfunction with disruption of neuronal function and cognitive impairment have yet to be defined. How does derangement of specific vascular cells by hypertension contribute to neuronal dysfunction? scRNA-seq-based studies could help unveil the hypertension-induced molecular changes which occur within these cells, potentially leading to new diagnostic insights. Is a mismatch between energy demand and blood supply consequent to hemodynamic insufficiency enough to drive cognitive decline? Or are events such as loss of endothelial trophic support and diminished clearance of metabolites of equal (or greater) importance? Answering these questions would provide new leads and, potentially, new therapeutic targets.

• How do aging and sex modulate these relationships? At present, most animal studies have been performed in young males⁸⁸. Given that age and sex modulate the development and expression of hypertension in the AngII, SHR, Dahl salt-sensitive, and DOCA-salt models¹⁷³, future research into the influence of these factors on neurovascular and cognitive dysfunction in hypertension would be illuminating and relevant to the human disease, which affects mainly aged individuals of both sexes.

These are a few of the outstanding questions that remain to be addressed. Rapid advances in neurovascular biology, an increased appreciation for the role of the neurovasculome in health and disease, and powerful new methodological approaches will expand our knowledge of the impacts of vascular sequela on cognitive function in hypertension. In closing, synchrony of biomarker and clinical-pathological studies with basic science mechanistic investigations will be essential for the advancement of new diagnostic tools and treatment interventions in the effort to preserve cognitive health over the lifespan.

Non-standard Abbreviations and Acronyms

Aβ

Amyloid-β

AD

Alzheimer’s disease

AngII

Angiotensin II

APOE4

Apolipoprotein E ε4

AQP4

Aquaporin 4

ARB

Angiotensin receptor blocker

AT1R

AngII type 1 receptor

BBB

Blood-brain barrier

BP

Blood pressure

BPV

BP variability

CBF

Cerebral blood flow

CSF

Cerebrospinal fluid

DBP

Diastolic blood pressure

IL-17

Interleukin 17

IL-17RA

IL-17 receptor A

MAP

Mean arterial pressure

MRI

Magnetic resonance imaging

PET

Positron emission tomography

PP

Pulse pressure

p-tau

Phosphorylated tau

PVM

Perivascular macrophage

PVS

Perivascular space

RAGE

Receptor for advanced glycation end products

SBP

Systolic blood pressure

SHR

Spontaneously hypertensive rat

SVD

Small vessel disease

WML

White matter lesion