Vascular Contributions to Cognitive Impairment and Dementia: The Emerging Role of 20-HETE

Abstract

The accumulation of extracellular amyloid-β and intracellular hyperphosphorylated tau proteins in the brain are the hallmarks of Alzheimer’s disease. Much of the research into the pathogenesis of Alzheimer’s disease has focused on the amyloid or tau hypothesis. These hypotheses propose that amyloid-β or tau aggregation is the inciting event in Alzheimer’s disease that leads to downstream neurodegeneration, inflammation, brain atrophy, and cognitive impairment. Multiple drugs have been developed and are effective in preventing the accumulation and/or clearing of amyloid-β or tau proteins. However, clinical trials examining these therapeutic agents have failed to show efficacy in preventing or slowing the progression of the disease. Thus, there is a need for fresh perspectives and the evaluation of alternative therapeutic targets in this field. Epidemiology studies have revealed significant overlap between cardiovascular and cerebrovascular risk factors such as hypertension, diabetes, atherosclerosis and stroke to the development of cognitive impairment. This strong correlation has given birth to a renewed focus on vascular contributions to Alzheimer’s disease and related dementias. However, few genes and mechanisms have been identified. 20-HETE is a potent vasoconstrictor that plays a complex role in hypertension, autoregulation of cerebral blood flow and blood-brain barrier integrity. Multiple human genome wide association studies have linked mutations in the CYP4A genes that produce 20-HETE to hypertension and stroke. Most recently, genetic variants in the enzymes that produce 20-HETE have also been linked to Alzheimer’s disease in human population studies. This review examines the emerging role of 20-HETE in Alzheimer’s disease and related dementias.

Alzheimer’s Disease

Dementia is a global healthcare crisis. The term dementia is not a specific disease but refers to a spectrum of conditions that affect memory, language, judgement and behavior. Although minor loss of cognitive function is normal in aging, dementia is a clinical diagnosis that is characterized by cognitive impairment that is severe enough to interfere with activities of daily living.

Alzheimer’s disease (AD) and vascular dementia are the most common forms of dementia.¹ One in ten people above the age of 65 years old and 32% of the population above the age of 85 years old are affected by AD.^2,3 The cost for treating dementia in 2020 was $305 billion in the US and this cost is expected to triple to $1.2 trillion by 2050 with the aging population.^2,3 These costs were calculated from Medicare claims data, self-reported out of pocket spending, net nursing home spending and formal and informal home care.² Worldwide, around 50 million people have dementia, with about 10 million new cases reported each year. The worldwide cost in 2015 was estimated to be $818 billion per year.⁴

In 1901 the German psychiatrist Alois Alzheimer identified the first case of what would later come to be known as Alzheimer’s disease in a 50-year-old woman.⁵ Soon thereafter, the medical community would characterize this condition by its histologic features of “plaques” and “tangles” in the brain. Extracellular “plaques” are composed of amyloid-β (Aβ) peptides most predominantly Aβ42.⁶ Neurofibrillary “tangles” are hyperphosphorylated tau protein aggregates.⁶ These plaques and tangles became the hallmark of AD. It is of no surprise that academic research programs and pharmaceutical companies began focusing on developing therapies that would eliminate or decrease the production of plaques and tangles.

One hundred twenty-two years later, there is still no cure or effective treatments for AD. Current therapies such as acetylcholinesterase inhibitors temporarily alleviate symptoms but are not effective in delaying the progression of this insidious disease.⁷ Over 200 Investigational New Drug (IND) applications and clinical trials have been initiated and failed in the last decade alone.⁸ Some have proposed that the failure in AD therapies has been due to the focus on either the amyloid or tau hypothesis of AD.^6,8–10 Many of the proposed therapies, although effective at decreasing the generation of plaques and tangles, were not effective in slowing or preventing the progression of AD in human trials.

Most recently a new amyloid-β directed monoclonal antibody (Aducanumab/Aduhelm) was conditionally approved by the FDA in June of 2021 despite opposition by an independent scientific advisory committee.¹¹ Experts in the field questioned the drug’s efficacy in ameliorating the cognitive decline in patients and the marginal symptomatic benefits in the face of side effects such as brain bleeding and swelling.¹² The FDA has mandated a Phase IV trial to be conducted by the drugs’ sponsor Biogen in order to determine the future of the drug. The approval has been surrounded by controversy and is now the subject of a U.S. House of Representatives committee investigation.¹³ The lackluster results from Phase III trials¹² of Aducanumab continue to underscore the need for a broader holistic view regarding the pathogenesis of Alzheimer’s disease and consideration of therapeutic options beyond the amyloid hypothesis.

Vascular Contributions to Cognitive Impairment and Dementia

Traditional perspectives on AD and stroke have categorized these two diseases as distinct and unrelated disorders. However, decades of research have revealed a significant overlap between AD and cardiovascular/cerebrovascular disease. This overlap has given way to a burgeoning vascular hypothesis of AD.^14,15 Vascular Contributions to Cognitive Impairment and Dementia (VCID) is now the term that encompasses all cardiovascular and cerebrovascular risk factors that predispose an individual to the development of AD and related dementias.¹⁶ The National Institute of Health (NIH) has also recently designated VCID as a critical and emerging area of research.¹⁶

Hypertension and aging are primary risk factors for the development of cerebrovascular disease (CVD), stroke, vascular cognitive impairment (VCI), and are increasingly implicated in AD.^14,15 Seventy million patients are hypertensive, and 40% of patients with dementia have elevated blood pressure.^2,3,17 The Medicare costs for the treatment of hypertension and stroke are $46 and $130 billion per year, respectively.^17,18 Hypertension has synergistic deleterious effects on the cerebral circulation that accelerates the loss of cognitive function with aging, but the mechanisms remain unknown. Furthermore, studies investigating the management of hypertension as a preventative therapy for dementia are lacking. Recent advances in neuroscience, neurovascular biology, genomics, biomarkers, microscopy, and clinical imaging have provided new insights into VCID.¹⁴

Studies utilizing magnetic resonance imaging (MRI), positron emission tomography (PET), two-photon microscopy (TPM), phosphorescence lifetime microscopy (PMI), optical coherence tomography (OCT) have all contributed to the field.¹⁹ In particular MRI has been used as a non-invasive diagnostic modality for dementia.²⁰ MRI studies have also revealed that blood-brain barrier (BBB) breakdown at the level of the cerebral microcirculation can occur prior to the accumulation of plaques and tangles in the brain of AD patients.²¹ Additionally, MRI studies have also revealed that an increase in capillary blood flow can actually precede changes in upstream larger vessels which suggests that the cerebral microcirculation plays an integral role in functional hyperemia and that impairment may be implicated in cognitive dysfunction.^22–24

Cerebral Microcirculation and the Neurovascular Unit

The cerebral microcirculation (Fig. 1) refers to the network of small vessels less than 100 μm in diameter, which function as the primary distribution system for the exchange of oxygen and nutrients in the brain.^19,25 Current in vivo imaging modalities are not capable of characterizing resolutions (< 10 μm) necessary to fully visualize the entire cerebral microcirculation. However, ex vivo confocal microscopic techniques and whole-brain clearing protocols can be used to visualize even the smallest capillaries (< 10 μm) in diameter (Fig. 2).

Fig. 1.

The cerebral microcirculation. Pial arteries ensheathed with smooth muscle cells become penetrating arteries that dive into the parenchyma of the brain surrounded by a perivascular area known as the Virchow-Robin space filled with cerebral spinal fluid. Penetrating arteries begin to loose smooth muscle cell coverage as they branch into arterioles and capillaries and are replaced by pericytes. Astrocytic end feet help form the glial limitans and synapse with both neurons and cells of the neurovascular unit.

Fig. 2.

A) Thick coronal brain slices (500 μm) are rendered translucent using a novel commercially available clearing technique (Visikol ®). B) Z-stacked confocal image of cleared thick sections allows for visualization of the cerebrovasculature in the neocortex after in vivo perfusion of tomato-lectin (green). C) Colored density rendering of the 3D confocal image allows for ideal mapping of the extensive cerebrovascular network.

We used a new commercially available clearing technique (Visikol Inc, Hampton, NJ) to image the cerebral microcirculation in 500 μm brain slices (Fig. 2). Imaging of the tissue using confocal microscopy in 0.5 μm Z-stacked steps and reconstructed in 3D reveal the extensive capillary network of the neocortex (Fig. 2) and hippocampus (Movie 1).

Movie 1.

3D rendered Z-stacked confocal image of cleared rat brain tissue perfused with tomato-lectin (green). The extensive capillary network of the hippocampus is visualized in high resolution.

NIHMS1761266-video-Movie_1.pptx ^{(12MB, pptx)}

Larger vessels such as the MCA will branch into pial vessels that travel along the surface of the brain. These pial vessels transition into penetrating arteries that then become arterioles and capillaries (Fig. 1). Studies have shown that there is a capillary within every 10–20 μm of a neuron and that neurodegeneration and localized ischemia can occur when this close spatial relationship is disrupted.^26,27

Tissue blood flow is determined by the pressure gradient across resistance that is provided by this microcirculatory network. Resistance is a component of vascular diameter and viscosity of the microcirculatory bed.²⁸ Metabolic demands are variable and regionally determined. These changes which occur secondary to differences in metabolic demand are referred to as neurovascular coupling. Neurovascular coupling is largely determined by various cell types and mediators within the neurovascular unit (NVU).^29,30 The cellular components of the NVU are astrocytes, microglia, pericytes, neurons, smooth muscle cells, and endothelial cells (Fig 1). The NVU is a relatively recent term and has garnered much attention in the field of neuroscience. Through the intricate coordination of the NVU, the microcirculation plays an active role in cerebral blood flow, maintenance of the BBB oxygen and nutrient supply as well as cerebral homeostasis. The BBB is composed of tight junctions between endothelial cells, the basement membrane, and astrocytic foot processes. Glucose and amino acids can pass through the BBB via transport proteins, and lipid-soluble/nonpolar molecules may cross via diffusion.

The NVU is therefore a critical coordination hub that links neural activity to cerebral blood flow. The concept of the NVU was formalized at the 2001 Stroke Progress Review Group meeting of the National Institute of Neurological Disorders and Stroke.²² Since then, there has been tremendous interest and advances in the understanding of the NVU.

The structure and function of the NVU varies from large vessels branching of off the circle of Willis to the capillaries in the hippocampus. Larger cerebral vessels such as the anterior, middle and posterior cerebral arteries begin to arise from the circle of Willis and run along the surface of the brain inside of the subarachnoid space and supply the pial arteries on the surface of the brain. Pial arteries are ensheathed in multiple layers of vascular smooth muscle cells (VSMCs) and are distinguishable from the endothelium by a pronounced elastic lamina.³¹ Penetrating arterioles branch off of the pial arteries and dive into the parenchyma of the brain through an extension of the subarachnoid space known as the Virchow-Robin Space (VRS) composed of the vascular basement membrane and glial limitans (Fig. 1).²² The layers of VSMCs that surround the pial artery begin to thin at the level of the penetrating arteriole (Fig. 1).³¹ Eventually only a single or discontinuous layer of VSCMs surrounds the most distal portion of the penetrating arterioles, the elastic lamina becomes less pronounced, perivascular nerves are infrequent and ensheathing astrocytic end-feet are abundant.³¹ Penetrating arterioles then branch into intraparenchymal arterioles with only a single layer of VSCMs and give rise to capillaries where the VSMC layer is replaced by pericytes situated in the endothelial cell basement membrane (Fig. 1). Astrocytic end-feet are evident as well as abundant neural processes adjacent to the capillary basement membrane at this level.

Cerebral Pericytes

Of particular interest to VCID has been the role of pericytes inside of the NVU. Pericytes are contractile mural cells that, along with smooth muscles cells, have the potential to regulate cerebral blood flow.^32–34 Pericytes also play a multifunctional role in maintaining BBB integrity,^35,36 supporting new capillary growth³⁷ and possess stem cell properties.³⁸

First and second-order pericytes express alpha-smooth muscle actin (aSMA). However, whether or not higher-order pericytes express aSMA and can constrict and dilate in response to metabolic demand remains controversial.³⁹ Much of this controversy was based on histological findings that indicated higher-order pericytes did not exhibit immunostaining for aSMA. However, recent studies by Alarcon-Martinez et al. pointed out that standard histological fixation methods could lead to depolymerization of aSMA and explain the lack of staining during traditional post-fixation protocols.⁴⁰ Alarcon-Martinez then went on to report that actin depolymerization could be avoided using snap freeze fixation methods in methanol, and identified the expression of aSMA in higher-order pericytes.⁴⁰

Previous in vitro studies of brain slices, by Atwell et al. reported that higher order pericytes constrict rather than dilate under ischemic conditions.⁴¹ However, later studies by the same investigators revealed that pericytes can dilate in response to prostaglandin E2, but require nitric oxide to first inhibit synthesis of the vasoconstrictor, 20-HETE.³⁹ In this landmark study, Atwell’s group also indicated that capillaries dilate before arterioles when blood flow demands are stimulated by sensory inputs and that dilation of pericytes at the origin of capillary beds contributes 84% of the total increase in cerebral blood flow.³⁹ Additionally, Atwell and his co-workers reported that ischemic conditions can induce pericyte constriction in vivo and may lead to pericyte death in a state of constricted rigor.³⁹ They speculated that this mediates the no-reflow phenomena following traumatic brain injury and transient ischemic stroke. Pericyte death and stalled capillary flow (functional rarefaction) are also associated with BBB impairment, cerebrovascular pathology, neurodegeneration, cognitive impairment, and AD and related dementias.^{21,23,24,42,43}

Cytokines and Neuroinflammation

Alzheimer’s disease and related dementias are associated with neuroinflammation. The presence of immune cells and antigens surrounding plaques and tangles has been extensively reported.^44–46 These reports were some of the first evidence that began to question the supposition that the brain was an immune delimited organ. Later reports of elevated levels of complement and cytokines in the cerebrospinal fluid of AD patients further supported the hypothesis that neuroinflammation contributes to the development of AD.

However, it is important to note that neuroinflammation is still generally considered to be a downstream consequence of Aβ aggregation. The view that inflammation is a consequence and not a cause of AD is supported by the assumption that systemic diseases rarely are initiated by inflammation but rather require an insult to trigger an inflammatory response.⁴⁷ So the question remains, what is the initial insult? Is it simply the accumulation of Aβ or tau? If so, why haven’t current pharmacologic interventions which reduce Aβ and tau in the brain failed to slow the progression of AD? Is Aβ and tau aggregation in AD simply a final downstream biomarker of an earlier and more complex disease etiology? Perhaps the study of injury or impairment to the cerebrovasculature can provide insights into the initiation of neuroinflammation?

The BBB maintains a barrier between inflammatory cells and cytokines in the circulation and the immune-privileged CNS.⁴⁸ Impairment of cerebral autoregulation in TBI and stroke, distention of the cerebral microcirculation and breakdown of the BBB may enhance the extravasation of immune cells and initiation of cytokine signaling cascades. Cytokines are classified as either pro-inflammatory (IL-1, IL-6, IL-17, and TNF-α), anti-inflammatory (IL-3, IL-4, IL-10) or hematopoietic (IL-3, IL-5, and G-CSF). Pro-inflammatory cytokines have been linked to neurodegeneration and dementia.^49–54

Inflammatory cytokines, in particular IL-1, IL-6, IL-17, and TNF-α, have garnered significant attention in the field of behavior and neurodegenerative diseases. IL-1, IL-6, and TNF-α have been extensively studied in sickness behavior syndrome induced by acute bacterial and viral infections. Psychological changes associated with infections known as social withdrawal behavior have been associated with elevated IL-1, IL-6, IL-17, and TNF-α levels.⁵⁵ These cytokines and their receptors were reported in the brain, decades before the recent discovery of the meningeal lymphatic system.⁵⁶ More recently, Faraco et al. reported that mice fed a high salt diet exhibit cognitive impairment via upregulation of gut-initiated TH-17 leading to increased levels of IL-17 and consequent endothelial dysfunction and reduction of resting CBF.⁵¹ A diet rich in salt has long been associated with increases in blood pressure and cardiovascular pathology.^57,58 However, dietary salt has also been linked to morbidity and mortality independent of blood pressure.⁵⁹

Cerebral inflammation could also be initiated independent of BBB leakage. In a startling discovery that overturned decades of research, Louveau and co-workers provided the first evidence of a lymphatic system in the CNS. The existence of lymphatic vasculature in the CNS sheds new light on the field of neuroimmunology and the role of inflammation in neurodegenerative diseases. One such notable example was recently reported when De Mesquita et al. examined the effects of aging on the meningeal lymphatic system.⁶⁰ The investigators reported that CSF clearance of macromolecules such as Aβ is essential in young, healthy brains. Pharmacologic ablation of the meningeal lymphatic system resulted in cognitive impairment in the absence of cerebrovascular system dysfunction in elderly animals.⁶⁰ The meningeal lymphatic system may be the key to a better understanding of mechanisms that link neuroinflammation to Aβ and tau accumulation and cognitive dysfunction.

Vascular Clearance of Amyloid-β

Zlokovic and his coworkers have produced an impressive body of work that suggests BBB impairment and/or breakdown may occur in AD before the presence of Aβ deposition, neurodegeneration, and cognitive decline.^{23–25,42,43,61,62} Clinical studies by Zlokovic’s group using in vivo advanced dynamic contrast-enhanced MRI reported that BBB breakdown occurs prior to cognitive dysfunction and changes in amyloid and tau accumulation in AD patients.²¹ One proposed mechanism involves reduced clearance of Aβ and tau aggregates via loss or impairment of BBB embedded transport proteins.

Under physiologic conditions in healthy individuals, ~85% of Aβ in the brain is transported from the brain to the blood via BBB transport proteins embedded in the capillary endothelium. The remaining ~15% of Aβ is cleared via the CSF and the recently described cerebral lymphatic system.⁶³ Clearance of Aβ occurs mainly at the capillary level.⁶³ Therefore, BBB leakage and/or endothelial dysfunction that alters the expression of transport proteins along cerebral capillaries dramatically impedes Aβ clearance and likely plays a key role in the accumulation of Aβ in AD and related dementias.

The known proteins that clear Aβ are, apolipoprotein E (APOE) and apolipoprotein J (APOJ), also known as clusterin (CLU).⁶⁴ Autosomal dominant AD is caused by mutations in amyloid precursor protein and loss of function mutations in presenilin 1 & 2 that lead to impairment in the γ-secretase complex and an increase in relative ratios between Aβ42/Aβ40. The long Aβ42 peptide is found abundantly in extracellular plaques but the shorter Aβ40 peptide is less prone to aggregation. However, autosomal dominant AD accounts for <1% of AD cases. The majority of AD cases are sporadic late-onset AD that is most often linked to increased expression of APOE4.^10,42 The genetic basis of AD remains to be determined; however sequence variants in APOE, particularly in APOE4 reduce Aβ clearance leading to BBB breakdown, vascular pathology, loss of neurons and cognitive dysfunction. AD patients also exhibit reduced APOE expression in cerebral microvessels.^42,65 These findings underscore the critical role of the cerebral vasculature and changes in the expression of transport proteins in the capillary endothelium in the pathogenesis of AD. The key remaining questions to consider are whether the expression of amyloid and tau transport proteins are altered in patients with cardiovascular/cerebrovascular risk factors such as hypertension, diabetes and atherosclerosis all of which have been associated with endothelial dysfunction? Previous studies have associated cerebrovascular damage with neuroinflammation and cognitive dysfunction, but few genetic mechanisms and pathways have been identified. Our recent studies suggest that one such mechanism may involve 20-HETE.

20-HETE

20-hydroxyeicosatetraenoic acid (20-HETE) is an eicosanoid that is formed by the omega hydroxylation of arachidonic acid via enzymes of the cytochrome P450 (CYP) 4A and F families.⁶⁶ The enzymes that produce 20-HETE differ among species. In humans, CYP4A11 and CYP4F2 are the predominant isoforms that catalyze the formation 20-HETE. In rats, the analogous isoforms are CYP4A1, CYP4A2, CYP4A3 & CYP4A8, with CYP4A1 as the most catalytically active.⁶⁷ In mice, CYP4A12 is the only enzyme capable of producing 20-HETE.⁶⁸ The other isoforms, CYP4A10 and 4A14 are involved in the omega-oxidation of short chain fatty acids and leukotrienes.

20-HETE has a myriad of biological activities, including; vasoconstriction, endothelial dysfunction, restenosis, and angiogenesis.⁶⁹ The diverse role of 20-HETE has been extensively reviewed elsewhere.⁷⁰ The focus of this review is to highlight the role of 20-HETE in hypertension, aging, and cerebral blood flow regulation. In the peripheral vasculature, 20-HETE is a potent vasoconstrictor that promotes endothelial dysfunction and is prohypertensive.⁶⁹ In the kidney, 20-HETE inhibits sodium reabsorption in the proximal tubule and thick ascending loop of Henle and is therefore antihypertensive via its natriuretic effects.⁷¹

GPR75

The vasoconstrictive effects of 20-HETE are mediated by activation of PKC, MAPK, tyrosine kinase, and Rho kinase signaling cascades which ultimately induce Ca²⁺ intracellular influx via depolarization of VSCMs secondary to inhibition of large conduction Ca-sensitive potassium channels.^71–74 The signal transduction pathways activated by 20-HETE implied the existence of one or more G protein-coupled receptors. However, identification of a 20-HETE receptor remained elusive until Garcia et al. recently identified an orphan G_q protein-coupled receptor (GPR75) as an effector for the vascular actions of 20-HETE.⁷⁵ The elegant study utilizing click chemistry was a breakthrough in the field of eicosanoid research.⁷⁶ Garcia’s group reported that GPR75 was expressed in human umbilical vein endothelial cells and that 20-HETE activation of GPR75 receptor overexpressed in these cells resulted in increased formation of reactive oxygen species (ROS) activity and upregulation of the expression of angiotensin converting enzyme (ACE). Increased formation of vascular 20-HETE production seen in transgenic mouse models and following administration of dihydrotestosterone is associated with the development of angiotensin II and 20-HETE-dependent hypertension.⁷⁷ GPR75 has also been identified as an endogenous receptor for RANTES/CCL5.⁷⁸ Dedoni et al reported that activation of GPR75 via RANTES/CCL5 attenuated the neurotoxic effects of Aβ deposition.⁷⁸ These results suggest that the neuroprotective effects of CCL5 in the brain might be associated with inhibition of 20-HETE activation of GPR75 in neurons and the vasculature that have recently been reported to express this receptor.⁷⁹

20-HETE in Stroke and VCI

Numerous genome-wide association studies (GWAS), have identified genetic polymorphisms in CYP enzymes that produce 20-HETE are associated with hypertension and stroke.^80–86 These include rs2108622 (V433M) and rs1558139 in CYP4F2 and rs1126742 (F434S), rs9333025, and rs389011 in CYP4A11 in Japanese, Australian, Swedes, Chinese, South India, Japanese and American subjects. Gao et al. confirmed that the rs12912592 polymorphism in CYP11 is a risk factor for stroke in a Chinese Han population, they also identified a variant in rs28681535 that has a protective role.⁸⁷ Among these single nucleotide polymorphisms (SNPs), rs2108622 (V433M, CYP4F2) and rs1126742 (F434S, CYP4A11) was confirmed to reduce 20-HETE production in vitro. In animal studies, Dahl salt-sensitive (SS) rats have a deficiency in the expression of CYP4A enzymes and the formation of 20-HETE. They displayed cerebral vascular dysfunction and enhanced blood-brain barrier (BBB) leakage, which may increase the susceptibility to the development of stroke.⁸⁸ Thus, available evidence seems to suggest that reduced 20-HETE levels may predispose individuals to the development of stroke, however, the mechanisms involved remain to be determined.

On the other hand, numerous studies indicate that cerebral ischemia increases the activity of phospholipase A2, which hydrolyzes membrane phospholipids to liberate arachidonic acid (AA).^89,90 An increase in AA promotes the subsequent production of HETEs and epoxyeicosatrienoic acids (EETs).⁹¹ 20-HETE levels increase following subarachnoid hemorrhage (SAH)^90,92 and ischemic stroke.^93,94 Plasma levels of 20-HETE, EETs and DiHETEs are markedly elevated in acute ischemic stoke patients and may serve as a biomarker for the severity of the injury.⁹⁵ Blockade of 20-HETE attenuates the initial fall in CBF and delayed cerebral vasospasm following SAH and infarct size following transient cerebral ischemia and traumatic brain injury.^96,97

Recent studies have indicated that pericytes constrict and contribute to the no-reflow phenomena following transient ischemia and traumatic brain injury (TBI).^25,42,98 Loss of pericytes has also been associated with capillary rarefaction, reduced CBF and BBB leakage in various animal models and AD patients.^19,22,62 Elevated levels of 20-HETE following cerebral ischemia and traumatic brain injury have been postulated to contribute to pericyte constriction and the no-reflow phenomena. Indeed Attwood and coworkers have found that pericytes constrict following ischemia and administration of a 20-HETE inhibitor reduces capillary stalling and infarct size following ischemia reperfusion injury.³⁹ In this regard, Poloyac et al reported that administration of HET0016, an inhibitor of the synthesis of 20-HETE, attenuated the delayed fall in CBF following transient cerebral ischemia and greatly reduced infarct size.⁹⁹ However, other studies have found that the reduction in infarct size following ischemia reperfusion injury was not due to recovery of CBF in the ischemic penumbra, but rather due to a reduction in the generation of ROS and a direct neuroprotective effect of blockade of 20-HETE synthesis in ischemic neurons.⁹³

20-HETE promotes endothelial dysfunction, proliferation of vascular smooth muscle cells and inward vascular remodeling.^70,100 Given the strong evidence that 20-HETE plays an important role in cerebral vasospasm and ischemic stroke, it seems logical to assume that it may contribute to vascular dementia by promoting cerebral hypoxia.

20-HETE in Alzheimer’s Disease

In a preliminary report examining elderly patients from the Atherosclerosis Risk in Communities (ARIC) cohort, we found that these inactivating variants of CYP4A11 and CYP4AF2 are associated with AD.⁸⁹ More specifically, in behavioral tests we reported that these inactivating variants were significantly linked to multiple aspects of cognitive impairment including delayed word recall, impaired digit symbol substitution, mistakes in word fluency tests and cognitive impairment.⁸⁹ MRI imaging revealed that these variants were also associated with the loss of hippocampal, cortical and signature AD region volumes, as well as an increase in white matter hyperintensities indicative of AD.⁸⁹ Most recently Ni et al. using peripheral blood samples from 744 patients from the Alzheimer’s Disease Neuroimaging Initiative (ADNI) cohort reported that a decrease in expression of CYP4A11 mRNA in peripheral blood was a significant diagnostic biomarker for AD.¹⁰⁴ These results suggest that there is an association between the CYP4A variants, 20-HETE and AD.

There is also some recent evidence suggesting that Aβ and 20-HETE may interact and contribute to the fall in CBF at the capillary level in AD patients. A recent study indicated that Aβ42 constricted pericytes and reduced capillary diameter in human brain slices obtained from tissue removed during neurosurgical procedures. The constriction of the pericytes was dependent on endothelin and increased ROS.¹⁰⁵ Previous studies have indicated that the vasoconstrictor effect of endothelin, angiotensin II and other Gq dependent vasoconstrictors in cerebral vascular smooth muscle cells is partially dependent on 20-HETE and can be attenuated by inhibitors of the synthesis of 20-HETE.^106,107

Other studies indicate that the expression of CYP4A is elevated in the liver and reduced in the kidney of Tg2572 transgenic mouse model of Alzheimer’s disease that is associated with elevated production of Aβ42.¹⁰⁸ Since 20-HETE constricts pericytes³² it is tempting to speculate that upregulation of the expression of CYP4A and elevated production of 20-HETE may contribute to the effects of Aβ on pericytes and reductions in capillary perfusion seen in AD patients. Alternatively, Aβ mediated reductions in the expression of CYP4A may impair pericyte function and in time lead to capillary damage and rarefaction. However, the mechanisms involved remain to be determined.

20-HETE and the Regulation of Cerebral Blood Flow

One clue to unraveling this link between 20-HETE and AD may be related to its essential role in the regulation of the myogenic response of cerebral arteries and autoregulation of cerebral blood flow. The myogenic response, also known as the Bayliss effect, is an inherent property of blood vessels to constrict in response to increases in transmural pressure and to dilate in response to decreases in pressure.¹⁰⁹ The myogenic response plays an essential role in autoregulation of CBF to maintain a constant blood flow to the brain despite transient changes in perfusion pressure.¹¹⁰

Inhibitors of the formation of 20-HETE have been reported to block the myogenic response of renal and cerebral arteries in vitro, and autoregulation of blood flow in vivo.^{72,73,88,111–113} Our lab has reported that MCAs dissected from 20-HETE deficient Dahl salt-sensitive (SS) rats exhibit impaired myogenic response and autoregulation of cerebral blood flow. Moreover, the transfer of the wild-type Cyp4A1,A2,A3 & A8 genes on chromosome 5 of the Brown Norway (BN) rat, or by the transgenic knock-in of wild type Cyp4A1 on to the SS background rescued myogenic response and autoregulation of cerebral blood flow. Additionally, the SS rat exhibited leakage of the BBB in response to acute elevations in pressure, and these effects were not evident in control consomic and transgenic strains.⁸⁸

Epidemiologic studies have shown that elderly individuals are more susceptible to cerebrovascular disease, stroke, and cognitive impairment in the face of superimposed hypertension than young hypertensive individuals.¹¹⁴ However, there has been a lack of studies examining specific mechanisms that may explain why younger individuals are protected from the cerebrovascular damage of hypertension.¹¹⁵ In an attempt to answer this question, Toth et al. proposed that age-related loss of the myogenic response in cerebral arteries and consequently impaired autoregulation of CBF could exacerbate hypertension-induced microvascular damage leading to BBB leakage, neuroinflammation and cognitive impairment in elderly subjects.¹¹⁶ Toth and coworkers induced hypertension in young and old mice via angiotensin II. They reported that myogenic tone and autoregulation of CBF in MCAs were markedly impaired in the aged hypertensive mice versus the young hypertensive group.¹¹⁶ They found that the enhanced myogenic tone in young hypertensive mice was dependent on upregulation of vascular 20-HETE production and that this compensatory response was not seen in elderly hypertensive mice.¹¹⁶ Additionally, Toth’s group reported that the loss of 20-HETE mediated myogenic tone and autoregulation was associated with BBB leakage, neuroinflammation, and cognitive decline with aging.¹¹⁶ These findings were the first to implicate age-related loss of 20-HETE production with cerebrovascular pathology and susceptibility to cognitive impairment. The mechanism remains to be determined. Since vascular 20-HETE production is strongly induced by testosterone, perhaps the loss of function is related to the decline in testosterone levels in elderly males.

20-HETE, GPR75, and the Neurovascular Unit

An important step in characterizing 20-HETE’s function in the cerebrovascular system was to identify the cell types and regions of the brain that express 20-HETE producing enzymes (CYP4A) and the 20-HETE receptor (GPR75). Our lab recently identified the cell types in the NVU and found that CYP4A and GPR75 were expressed in VSMCs and endothelial cells of pial and penetrating arterioles (Fig. 3). We also were the first to report that both CYP4A and GPR75 are expressed in pericytes and neurons.⁷⁹ CYP4A, but not GPR75, was expressed in astrocytes.⁷⁹ Both CYP4A and GPR75 were also abundantly expressed in neurons of the entorhinal cortex, hypothalamus, and a subpopulation of neurons in the neocortex and hippocampus.⁷⁹ We also found that the expression of CYP4A in first and second-order pericytes was associated with a decrease in capillary diameter and stalling of red blood cell flux.⁷⁹

Fig. 3.

Expression of CYP4A in pial and penetrating arterioles. (A–E) Confocal images of a pial and penetrating arteriole in the neocortex. A penetrating arteriole is distinguished from veins by the Virchow-Robin space (VRS). CYP4A (red) is abundantly expressed in pial and penetrating arterioles labeled by in vivo injection of FITC-Dex (green). CYP4A is also co-expressed in astrocytes (GFAP/purple) surrounding penetrating arterioles. White arrows indicate absence of CYP4A expression on the endothelial cell layer of capillaries. Scale bar = 100 μm. (F–J) Single plane confocal image of a penetrating arteriole cut in cross section. Endothelial cells (EC) stained with tomato-lectin (Lectin/purple) that line the lumen co-express CYP4A (red) and yield a magenta color. Vascular smooth muscle cells (VSMC) immunostained with aSMA (green) co-express CYP4A (red) and are seen in yellow. (J) Red staining in the adventitia (blue arrows) depicts astrocytic end feed which express CYP4A (red). Representative images were repeated in multiple vessels in 3–5 non-adjacent sections obtained from at least 3 animals. Reprinted with permission from Frontiers in Neurology.

20-HETE and Pericyte Contractility

Recent studies by our lab found that pericyte constriction could be impeded by administration of an inhibitor of 20-HETE synthesis and restored with a 20-HETE agonist in vitro.¹¹⁷ Similar findings were seen in the MCA and penetrating arterioles (PA). Administration of the inhibitor increased the diameter of both MCAs and penetrating arterioles (PA) and impaired the myogenic response to elevations in pressure. These responses were restored by the 20-HETE agonist. Additionally, in vivo studies by our group revealed that administration of a 20-HETE antagonist increased CBF and impaired autoregulation of CBF in both the superficial and deep cortex of rodents as measured by laser Doppler flowmetry and two-photon microscopy (Fig. 4).¹¹⁷ These findings highlight the emerging role of 20-HETE in mural cell contractility and the regulation of capillary perfusion. According to Toth et al., 20-HETE levels decrease with age.¹¹⁶ Could age-related loss of 20-HETE and contractibility in pericytes explain the increased susceptibility of cerebrovascular disease in elderly individuals? Could the resulting loss of autoregulation of CBF with age lead to transmission of high pressure and microbleeds in the setting of hypertension?

Fig. 4.

Left) Comparison of relative cerebral blood flow, estimated by red blood cell velocities of cortical microvessels in response to HET0016 a 20-HETE antagonist using two-photon laser scanning microscopy in Sprague Dawley rats. Right) A representative image of cortical cerebral microvessels and the sight of red blood cell velocities measurements (red arrows). Mean values ± SEM are presented. * indicates P < 0.05 from the corresponding values in HET0016-treated versus untreated rats. Reprinted with permission from the Journal of Prostaglandins.

EETs in Ischemic Stroke and VCI

EETs are abundantly produced in astrocytes, endothelial cells and neurons in the brain.¹¹⁸ Several studies suggest that EETs are protective in ischemic brain injury. The first suggestion came from studies showing that transient ischemic preconditioning reduced infarct size following ischemia reperfusion in the rat.¹¹⁹ The protective effect was associated with upregulation of the expression of CYP2C11 epoxygenase,¹²⁰ suggesting that increased formation of EETs may serve as an endogenous protective mechanism in stroke.¹¹⁹ In other studies elevating the levels of EETs by deleting soluble epoxide hydrolase (sEH) that degrades EETs reduced infarct size and neurological deficit following transient middle cerebral artery occlusion compared with wild type control mice.¹²¹ This effect was associated with higher CBF in the knock out mice. The mechanisms of protection by EETs in ischemic stroke have been attributed to cerebral vasodilation, promotion of angiogenesis, suppression of platelet aggregation, oxidative stress, and post-ischemic inflammation.¹²²

Both 20-HETEs and EETs have been implicated in VCI given their important role in regulation of the cerebral microcirculation. Autopsy analysis of the brains of patients with VCI revealed the ratio of DiHETEs to EETs was elevated, suggesting upregulation of sEH.¹²³ Immunohistochemistry indicated upregulation of sEH in microvascular endothelium. In another study, type 2 diabetes was associated with loss of spatial memory and upregulation of cerebrovascular sEH expression. This finding suggests that upregulation of sEH, decreased levels of EETs, endothelial dysfunction and subsequent loss of neurovascular coupling may contribute to diabetic induced VCI. To date however, there is little information available regarding the role of EETs or if sEH inhibitors have a beneficial effects in Alzheimer’s disease.

Translational Impact

The 20-HETE pathway is an attractive target for pharmaceutical intervention. 20-HETE analogs, inducers, and antagonists have all been identified.¹⁰⁶ Many of these compounds are lipid-soluble, cross the BBB, and are effective at nanomolar concentrations. A decrease in CYP4A11 mRNA levels in the blood have recently been proposed as an early diagnostic biomarker for AD.¹⁰⁴ Precision medicine is now capable of identifying individuals with CYP4A/F polymorphisms. CYP genotyping is routine in most pharmacogenomic studies to predict drug metabolism and responses. Clinicians may someday utilize this type of information to more aggressively manage hypertension and prevent the possible onset of AD and related dementias in patients with inactivating mutations in CYP4A11 and CYP4F2 genotypes.

Conclusion

Vascular contributions to cognitive impairment and dementia is an emerging and critical area of research. Although links to cerebrovascular disease and dementia continue to be reported, very few genes and mechanisms have been identified. 20-HETE is a potent vasoconstrictor that has recently been shown to be produced by cells of the neurovascular unit and is a critical mediator in the myogenic response and autoregulation of cerebral blood flow. Genetic variants in 20-HETE production have been previously linked to hypertension and stroke and there are now preliminary reports that they may also be associated with Alzheimer’s disease and related dementias. The 20-HETE pathway could play an important role in understanding vascular contributions to dementia and is an attractive target for potential pharmaceutical intervention.