The Alzheimer’s Disease Brain,Its Microvasculature,and NADPH Oxidase

INTRODUCTION

Signs of altered vascular structure evident at the very earliest preclinical stages of sporadic Alzheimer’s disease (AD) have prompted long-standing suspicions that the deterioration of the brain’s microcirculation is a key event in the pathogenesis of the disease [1–3]. In support of this impression, a recent large-scale multifactorial analysis of brain images in AD concluded that vascular dysregulation was the earliest/strongest pathological factor associated with the development of AD preceding, in order, amyloid deposition, reduced glucose utilization, impaired neuronal activity and grey matter atrophy [4].

MICROVASCULAR PATHOLOGY

Cortical capillaries in post-mortem specimens from patients with AD have irregular pouches and excrescences on their vessels (a “lumpy-bumpy” appearance), general thickening of the basement membrane, a reduced number and length of tight junctions, a loss of the fine perivascular neuronal plexus and the appearance of pits or lacunae in many of the vessel walls [3]. Microvascular density is unchanged or reduced but the number of string, kinked or coiled vessels is increased. [1, 5–8]. These vascular changes stand in contrast to the smooth cylindrical shapes of the cortical capillaries revealed by electron microscopy in postmortem specimens derived from elderly non-demented subjects [3].

Pathological changes in the microvasculature engage all the major cellular elements of the capillary wall, the endothelium, the pericytes and the astrocytes. Endothelial cells and pericytes present with pinocytotic vesicles and cytoplasmic inclusions and a loss of pericytes is discernible very early in the course of the disease [8]. Endothelial cells and pericytes as well as perivascular astrocytic processes contain enlarged mitochondria with disrupted mitochondrial cristae reduced in number and density. The endothelial expression of the glucose transporter protein GLUT-1 is also reduced. Prominent perivascular microglial proliferation adds to the pathological picture [1, 9].

The vascular changes in the brain in AD are not uniformly distributed and are most obvious in the basal forebrain and the hippocampus. Abnormally contoured blood vessels can be found throughout the cortex but, again, are most abundant in the basal forebrain and the hippocampus [2]. The highest density of microvascular pathology is found in the hippocampal CA1/subiculum, a region highly susceptible to ischemia and severely involved in the initial stages of AD [6, 10]. Functional impairments in this region disrupt the connections between the hippocampus, the basal forebrain and the association cortices that subserve memory [11, 12]

BLOOD-BRAIN BARRIER

This disruption appears to start early even in the normal brain. Advanced imaging and data processing techniques have been applied to living human subjects with and without cognitive impairment to quantify the blood-brain barrier (BBB) regional permeability (K_trans) constant [13]. Measurements in 12 different regions of the brain, in the hippocampus and in different cortical, subcortical, and white matter regions have revealed an age-dependent increase in the K_trans values in the hippocampus, its CA1 region and the dentate gyrus, starting as early as the third decade of life. The increase in K_trans in the hippocampus was significantly greater in subjects with minimal cognitive impairment than in age-matched normal controls. No significant changes in BBB permeability during aging were found in cortical, subcortical, or white matter regions, except for the caudate nucleus, in either the normal subjects or in those with minimal cognitive impairment. BBB breakdown during aging appeared to be a predominantly hippocampal phenomenon [13].

In keeping with the increased BBB permeability and vascular leakage observed in individuals with minimal cognitive impairment, a significantly greater cerebrospinal fluid (CSF)/plasma albumin ratio was found in these individuals than in age-matched normal controls. Higher CSF levels of soluble platelet-derived growth factor receptor β (sPDGFRβ) in subjects with minimal cognitive impairment also reflected the increased permeability of the BBB. sPDGFRβ CSF levels were found to correlate with K_trans values in the CA1 and dentate gyrus subdivisions of the hippocampus [13]. Levels of sPDGFRβ increased in parallel with rising scores on the clinical dementia rating scale and this was true regardless of CSF levels of amyloid-β (Aβ) or phosphorylated tau suggesting that BBB breakdown was an early biomarker of impaired cognition independent of Aβ and phosphorylated tau [14].

sPDGFRβ is an established biomarker of pericyte injury [15]. These cells maintain the integrity of the BBB but, in normal subjects, starting at about age 20, CSF sPDGFRβ levels begin to rise continuously, suggesting that pericyte injury is a normal feature of aging [16, 17]. Corresponding to the data presented earlier by Iturria-Medina et al. (2016) [4], and by Nation et al. (2019) [14], the increase in CSF sPDGFRβ begins earlier than the decline of Aβ₄₂ in the CSF, implying that pericyte damage, and with it, vascular damage, occurs before Aβ₄₂ deposition [16]. CSF Aβ₄₂ begins to decline at about age 40 as it is deposited in the brain and starts to form plaques. Pericyte damage may thus contribute to the pathogenesis of AD very early on in the development of the disease [16]. Pericyte degeneration is recognized by many, but not all, investigators as a feature of AD [8, 18–20].

HIPPOCAMPUS

What explains the vulnerability of the hippocampus in AD and specifically its CA1 region? When neurovascular coupling in the CA1 region in the mouse is compared to that in the visual cortex, two major differences stand out [21]. First, resting blood flow in the hippocampus is lower than in the cortex because vascular density is lower in the hippocampus and because red blood cell velocity and flux is lower in individual capillaries. This leads to a lower resting blood oxygen saturation which, in turn, limits the capacity of the hippocampus to generate ATP. Second, hippocampal capillaries dilate less frequently and to a lesser extent than cortical capillaries, possibly because hippocampal capillaries have fewer pericytes than cortical tissue and because these pericytes have a longer and less contractile phenotype [22, 23]. Increased neuronal activity in the hippocampus causes fewer and smaller dilations of local blood vessels and a smaller increase in overall blood volume. The hippocampus appears to be less able to respond to fluctuations in energy demand than the cortex [21].

MICROVASCULAR NADPH OXIDASE

In addition, hippocampal microvessels have distinguishing biochemical features which render them vulnerable to deterioration [24]. Protein levels of NADPH oxidase (NOX), the major source of superoxide (O₂^–) in the vascular endothelium, are significantly higher in cortical and hippocampal microvessels than in cerebellar microvessels where the pathological changes of AD appear less extensive [24, 25]. Corresponding to this high NOX activity, levels of O₂^– are also significantly greater in cortical and hippocampal microvessels. These biochemical features are coupled with significantly lower levels of manganese superoxide dismutase in hippocampal and cortical microvessels than in cerebellar microvessels.

Other biochemical alterations serve to further augment the oxidative environment of hippocampal and cortical microvessels. Thus, although there are no differences in endothelial nitric oxide synthase (eNOS) expression between the microvessels of the three brain regions, the bioavailability of tetrahydrobiopterin, an essential cofactor for eNOS activity, is significantly reduced in hippocampal and cortical microvessels compared to the cerebellum while levels of oxidized tetrahydrobiopterin, i.e., 7,8 dihydrobiopterin, are increased [24]. Endothelial nitric oxide (NO) is an essential factor in the regulation of vascular tone and regional blood flow, leukocyte–endothelial interactions, platelet adhesion and aggregation, and vascular smooth muscle cell proliferation. Deficient vascular NO results in endothelial dysfunction, vasoconstriction, and inflammation [26, 27]. However, in the absence of adequate amounts of tetrahydrobiopterin, eNOS becomes uncoupled and produces O₂^– in addition to NO, and O₂^–, in turn, reduces the bioavailability of NO by oxidizing NO to the free radical peroxynitrite (ONOO^–) effectively preventing its physiologic functions [28–30] (Fig. 1). The oxidation of NO to ONOO^– further amplifies the level of oxidative stress in the mouse hippocampal and cortical microvasculature.

OPEN IN VIEWER

Fig. 1

At optimal concentrations of tetrahydrobiopterin (BH₄), endothelial nitric oxide synthase (eNOS) converts L-arginine to L-citrulline and nitric oxide (NO). This reaction becomes uncoupled when BH₄ is oxidized to dihydrobiopterin (BH₂) by superoxide (O₂^–) generated by activated NADPH oxidase (NOX). The uncoupled reaction now produces (O₂^–) as well as L-citrulline and NO. NO and O₂^– react to produce the peroxynitrite radical (ONOO^–). The ratio of BH₄/BH₂ regulates the ratio of NO/O₂^– generated by eNOS.

NADPH OXIDASE AND AGING

In mouse and man, brain NOX expression and O₂^– production rise significantly with age [31]. For example, in normal elderly wild-type (WT) mice, examination of midbrain tissue homogenates reveals a significant increase in the expression of NOX compared to young WT controls. Levels of angiotensin II, a potent vasoconstrictor, and a known activator of NOX, are also increased in the brains of aging WT mice. Along with the increased expression of NOX, there is a two-fold increase in NOX dependent O₂^– production in aging WT brains associated with increased levels of brain tissue lipid peroxides. The increased production of O₂^– is much less evident when young and old NOXKO mice are compared. A significant reduction in cerebral capillary and neuronal density in aging WT mouse brains accompanies the increase in O₂^– production. In contrast, capillary density and neuronal density are well preserved in aging NOXKO mouse brains. Apocynin, a putative NOX inhibitor, significantly reduces the O₂^– level but neither rotenone, a mitochondrial complex I enzyme inhibitor, nor oxypurinol, a xanthine oxidase inhibitor, can achieve this effect [31].

In elderly rats, the increase in brain NOX activity leads to the increased nuclear translocation of NF-_Kβ, a transcription factor that induces the expression of pro-inflammatory genes, and to a corresponding increase in the levels of the proinflammatory cytokines IL-1β, IL-6, and TNF-α. All these phenomena are significantly inhibited by long-term dietary antioxidant treatment with agents such as N-acetyl-cysteine, a glutathione precursor [32].

In common with the findings in rodents, postmortem human midbrain tissues that include the hippocampus and ventral tegmental area excised from young, middle-aged, and elderly adults without diagnosed neurodegenerative diseases, also show significant increases in NOX expression in elderly adults together with significant increases in O₂^– production and angiotensin II levels. The increase in O₂^– production may explain the increased DNA damage found in the midbrain regions of elderly adults and the significant reductions in capillary and neuronal density [31].

An earlier study that examined the changes in NOX expression in the superior and middle temporal gyri and cerebellum of patients as they progressed from preclinical AD to mild cognitive impairment and the late-stage of the disease found that compared to control subjects, significant elevations in NOX activity were limited to the temporal gyri of patients with mild cognitive impairment and that increased NOX activity was not found in the pre-clinical or late-stages of AD. No significant changes were found in the cerebellum at any stage of the disease [33]. However, a follow-up study of postmortem samples taken from the frontal and temporal cortices confirmed the significant elevation of NOX expression in subjects with mild cognitive impairment and demonstrated its persistence as the disease advanced to its late stages. Most important, an inverse correlation was observed between NOX activity and cognitive ability [34].

NADPH OXIDASE ACTIVATION

Why do NOX levels rise with age in the hippocampal and frontal regions of the brain? There is no clear answer, but vascular endothelial NOX can be activated by numerous stimuli. The levels of angiotensin II, for example, a potent NOX stimulant, increase with normal aging in the frontal and temporal cortices especially after age 65 [31, 36]. Hyperglycemia and hyperinsulinism, two essential features of type 2 diabetes, known to predispose to AD, also raise endothelial NOX levels [37–39]. Breakdown of the BBB with age allows thrombin, another NOX activator, to enter the brain and further damage the endothelial barrier [40–42].

The profuse development of abnormally contoured and deformed capillaries in the frontal and hippocampal regions with age also predisposes to the increased expression and activity of NOX. Shear stress is detected by mechanosensors on endothelial cells [43]. In structurally normal vessels, laminar unidirectional flow and high shear stress (pulsatile shear stress) promote the survival of endothelial cells by increasing the production of metabolites that maintain vascular stability such as NO. Disturbed or turbulent flow, on the other hand, with little or no mean flow rate, exerts oscillatory shear on the endothelium that increases in proportion to the degree of vessel deformity. When severe, it can damage and even denude the capillary endothelial lining. The delivery of oxygen, glucose and other essential nutrients is compromised [2]. Oscillatory shear, in contrast to laminar flow, impairs endothelial homeostasis, and induces high levels of oxidative stress by up-regulating NOX which, in turn, triggers an inflammatory response by activating NFκB and the release of cytokines and adhesion molecules as well as the migration of leukocytes across the vascular endothelial barrier [38]. Under these conditions, as noted earlier, NO is converted to peroxynitrite, and its protective actions are lost [44–46].

APOCYNIN

Many molecules have been screened to identify specific NOX inhibitors with low off-target effects, but none have so far progressed to clinical use. Of these, Apocynin (acetovanillone), a naturally occurring acetophenone found in the roots of certain plants, is perhaps the most studied NOX inhibitor in experimental models of disease [47]. Apocynin is a prodrug and requires myeloperoxidase-mediated dimerization to inhibit NOX. NOX is a multienzyme complex composed of regulatory subunits that consist of membrane bound gp91^ph ∘x and p22^ph ∘x and cytosolic p47^ph ∘x, p67^ph ∘x, and p40^ph ∘x. On activation, the cytosolic regulatory component p47^ph ∘x becomes heavily phosphorylated, and the entire cytosolic complex migrates with the small GTPase Rac to the membrane, where all components assemble to form the active oxidase. The membrane-integrated protein gp91^ph ∘x is the catalytic core of the enzyme responsible for the electron transfer from NADPH to molecular oxygen for O₂^– production [33]. Diapocynin inhibits the activation of NOX by preventing the translocation of p47phox to the plasma membrane. However endothelial cells and vascular smooth muscle cells do not have myeloperoxidase and therefore cannot convert apocynin into its active dimer. In the absence of myeloperoxidase, apocynin continues to display its off-target effects and acts as a strong scavenger of H₂O_2, and as a weak scavenger of O₂^– [47–51].

Despite these limitations, apocynin has been used in many experimental paradigms to reduce the tissue damaging consequences of NADPH oxidase activation. In gerbils, pretreatment with apocynin significantly attenuates neuronal degeneration, oxidative DNA damage, lipid peroxidation and the activation of microglia and astrocytes in the CA1 region of the hippocampus that follow 5 minutes of bilateral common carotid artery occlusion [52]. In rat brain, increased expression and activity of NADPH oxidase together with a significant increase in the levels of reactive oxygen species and a decrease in antioxidative enzyme activity can be measured a week after a 20-minute occlusion of both common carotid arteries. Levels of glutathione are significantly decreased at that time. A sharp increase in hippocampal microglial activation is observed together with a significant rise in the levels of inflammatory markers. All these post ischemic changes are far less evident in animals pretreated with apocynin and during the 7-day ischemic period [53].

In another experimental paradigm, apocynin was shown limit the dysfunction of liver cells and pancreatic beta cells that follows exposure to non-esterified fatty acids (NEFA) such as oleate and palmitate and the activation of NOX. Treatment with apocynin restored normal levels of NOX activity and reactive oxygen species production and reversed the deleterious effects of NEFA on insulin metabolism [54, 55]. In rats, the acute application of isoproterenol was used to create a model of myocardial infarction. Apocynin reduced the production of O₂^– and the signs of oxidative stress attributed to the activation of NOX and ameliorated the histological signs of myocardial necrosis [56, 57]. In a model of polymicrobial sepsis in rats induced by cecal ligation and puncture, a combination of niacin and apocynin but neither one alone prolonged the life of the animals. Combined therapy, but again, neither one alone, significantly decreased the activity of NOX, increased the levels of NADPH and glutathione, decreased the levels of malondialdehyde and reduced NFκβ, and inflammatory cytokine expression. As well, combined therapy attenuated the histological signs of lung injury [58]. Niacin enhances the antioxidative effects of apocynin by generating NADPH and by inhibiting the expression of NOX [58, 59]. Indeed, NADPH, the antioxidant cofactor, acting alone, can be tissue protective. In mice, rats and rhesus monkeys with cerebral ischemia reperfusion injuries, NADPH sharply raised brain levels of glutathione and ATP, decreased the levels of reactive oxygen species and significantly reduced the infarct volume. The therapeutic effects of NADPH were evident even if it was started many hours after reperfusion [60]. The combination of NADPH and apocynin limited ischemic brain damage even better [61].

SODIUM OXYBATE

The tissue protective effects of sodium oxybate (SO) are comparable to apocynin and have been demonstrated in brain, gut, heart, kidney, liver, lung, pancreatic β cells and skeletal muscle [62]. These widespread cellular protective effects may be attributed to SO’s capacity to provide energy to the cell and to attenuate oxidative stress by generating NADPH and inhibiting the expression of NOX. SO’s catabolism in both the cytoplasm and mitochondria of the brain generates succinic semialdehyde which, in turn, is rapidly converted to the metabolic intermediate succinate with the concomitant reduction of NAD to NADH [63]. NADH is transformed by mitochondrial nicotinamide nucleotide transhydrogenase to NADPH [64]. Succinate is metabolized to form ATP, but it is also processed through the malate-pyruvate shuttle to stimulate the formation of additional NADPH [65, 66]. SO also generates NADPH by shifting intermediary metabolism in the direction of the pentose phosphate pathway [67, 68]. SO is thus both a source of energy to the cell and a powerful antioxidant. SO further promotes an antioxidative and anti-inflammatory environment in the cell by acting epigenetically on histone deacetylase 3 to suppress the expression of NOX [69–71]. SO is a known inhibitor of histone deacetylase 3, a class 1 histone deacetylase, the most widely expressed histone deacetylase in the brain [72, 73].

SO’s tissue protective effects were first demonstrated by Laborit who showed that SO could prolong the survival of mice subjected to the toxic effects of irradiation or high-pressure oxygen [67]. He attributed this protection to the activation of the pentose phosphate pathway and the increased formation of NADPH. The activation of the pentose phosphate pathway by SO was later confirmed by Taberner and his colleagues [68]. More recently, SO has been shown to increase the level of NADPH and the NADPH/NADP ratio in pancreatic islets exposed to palmitate. Palmitate has a lipotoxic effect on islet cells. It activates NADPH oxidase and the release of reactive oxygen species. These actions inhibit glucose stimulated insulin secretion and reduce NADPH levels and the NADPH/NADP ratio. Pre-treatment with SO blocks this inhibition and maintains normal levels of NADPH and a normal NADPH/NADP ratio [74]. Similarly, treatment with SO prevents the toxic effects of isoproterenol on the heart. Isoproterenol activates NADPH oxidase and produces diverse signs of oxidative stress, metabolic impairment, and histological damage. The demand for oxygen exceeds the capacity of the coronary circulation to provide it and globules of fat (triglyceride) accumulate in myocardial fibers when fatty acids, the major myocardial energy substrate, cannot be metabolized normally. Mitochondrial swelling and myocardial necrosis are evident. The application of SO before and after isoproterenol completely prevents these histological signs of metabolic and tissue stress [75, 76].

SO has repeatedly been shown to protect the brain from ischemic reperfusion injuries. A brief period of middle cerebral artery occlusion leads to an increase in NADPH-oxidase activity in the arteries serving the ischemic penumbra of the infarcted region. NADPH-oxidase activity is even temporarily increased within the arteries serving non-ischemic parts of the brain [77]. Pre- and post-treatment with SO significantly reduces the size of the infarct [78]. But even delayed treatment with SO after a brief period of ischemia protects the brain [79, 80]. SO’s tissue protective powers are not limited to its antioxidative effects. SO also sustains the supply of energy to the cell. In the brain, pre-treatment with SO maintains normal levels of ATP even under hypoxic condition [81]. This attribute may be particularly germane given the hypoxic threat to the vulnerable CA1 region of the AD brain.

SO may specifically be able to target and ameliorate the oxidative stress which develops in AD in the vascular endothelium of the microcirculation in the hippocampus. Monocarboxylate transporters carry SO into human capillary endothelial cells [82]. Unlike glucose, its uptake is not compromised by the reduced expression of glucose transporters, and it can provide energy to the hippocampal capillary endothelium in place of glucose. Indeed, SO reduces brain glucose utilization without reducing cerebral blood flow or oxygen consumption [83]. Inhibition of histone deacetylase 3 by SO also reduces the epigenetic expression of NOX and reduces the production of free radicals and the associated inflammatory response [72, 85]. By raising NADPH levels in the capillary endothelium, SO should quench the high levels of O₂^– which are found in these cells in the elderly and in AD [31]. As well, SO should maintain the formation of NO. NADPH raises the bioavailability of tetrahydrobiopterin by reducing dihydrobiopterin and, in this way, prevents the uncoupling of eNOS and the oxidation of NO to the ONOO^– radical [86]. As noted earlier, NO contributes to vascular homeostasis by inhibiting vascular smooth muscle contraction, platelet aggregation and leukocyte adherence [26, 27]. Leukocyte adherence and microvascular stasis are important contributors to ischemia reperfusion injury. In liver, for example, reperfusion injury significantly increases serum levels of hepatic enzymes and serum levels of endothelin-1. Endothelin-1, in addition to its vasoconstrictor properties, strongly promotes leukocyte adhesion. The apoptotic index rises, tissue malondialdehyde levels markedly increase and superoxide dismutase activity falls. SO reverses all these effects [87].

CONCLUSION

The nocturnal application of SO may have another singular advantage. SO reduces the level of Aβ in the brain by increasing the expression of neprilysin, a protease [88]. The increase in slow wave activity produced by SO is also associated with the increased recruitment of aquaporin-4 to perivascular sites [89]. Perivascular localization of aquaporin-4 facilitates the clearance of interstitial solutes, including Aβ, through a network of perivascular pathways in the brain termed the glymphatic system. Perivascular aquaporin-4 localization is reduced in individuals with AD and reduced localization is strongly associated with increased neurofibrillary and Aβ pathology [90].

SO has been used safely night after night for decades by many patients with narcolepsy [91]. It has also been used safely and effectively in elderly patients to reduce daytime fatigue and sleepiness in clinical trials of patients with Parkinson’s disease [92]. Long-acting formulations of SO are now being developed which may make it possible to maintain a prolonged and steady anti-oxidative and anti-inflammatory environment in the brain with high ratios of NADPH/NADP throughout the night. Applied to patients with early cognitive impairment, this metabolic environment may delay the progress of AD.