Targeting NADPH Oxidase and Phospholipases A₂ in Alzheimer’s Disease

Abstract

Alzheimer’s disease (AD) is marked by an increase in the production of extracellular beta amyloid plaques and intracellular neurofibrillary tangles associated with a decline in brain function. Increases in oxidative stress are regarded as an early sign of AD pathophysiology, although the source of reactive oxygen species (ROS) and the mechanism(s) whereby beta amyloid peptides (Aβ) impact oxidative stress have not been adequately investigated. Recent studies provide strong evidence for the involvement of NADPH oxidase and its downstream oxidative signaling pathways in the toxic effects elicited by Aβ. ROS produced by NADPH oxidase activate multiple signaling pathways leading to neuronal excitotoxicity and glial cell-mediated inflammation. This review describes recent studies demonstrating the neurotoxic effects of Aβ in conjunction with ROS produced by NADPH oxidase and the downstream pathways leading to activation of cytosolic phospholipase A₂ (PLA₂) and secretory PLA₂. In addition, this review also describes recent studies using botanical antioxidants to protect against oxidative damage associated with AD. Investigating the metabolic and signaling pathways involving Aβ NADPH oxidase and PLA₂ can help understand the mechanisms underlying the neurodegenerative effects of oxidative stress in AD. This information should provide new therapeutic approaches for prevention of this debilitating disease.

Increase in Oxidative Stress Is an Early Sign of AD Pathogenesis

Alzheimer’s disease (AD) is the most prominent age-related neurodegenerative disease characterized by accumulation of beta amyloid plaques and intracellular neurofibrillary tangles in the brain. Impairment of cognitive function and memory in this disease is attributed to degeneration of synapses in specific brain regions [1]. Although manifestations of AD have been shown to involve multiple factors, including inherent genetic mutations and environmental conditions, increases in oxidative stress associated with amyloid beta peptide (Aβ) accumulation are considered important underlying factors in the development of this disease [2–5]. Aβ peptides are comprised of 39–43 amino acids and are generated by proteolytic cleavage of amyloid precursor protein (APP) by beta and gamma secretases. Considerable evidence indicates that the cytotoxic effects of Aβ are due to its aggregation to the oligomeric form and not the monomeric or fibrillar forms [6–8]. Oligomeric Aβ can diffuse into cells and alter cellular pathways leading to disruption of synaptic plasticity and inhibition of long-term potentiation [9]. These abnormalities occur before the appearance of Aβ plaques [10].

A number of studies have attempted to elucidate the mechanism(s) whereby Aβ enhances reactive oxygen species (ROS) production and stimulates oxidative signaling pathways in the AD brain. Studies using redox proteomics have identified oxidatively modified proteins as early markers of AD pathogenesis [11]. Earlier studies by Butterfield [12] implicated the possible role of methionine-35 residue of Aβ in conferring neurotoxicity. Transgenic PDAPP mice with an M631L mutation in the APP molecule (corresponding to M35 in Aβ) showed markedly decreased oxidative indices and immunoreactive plaques. Interestingly, this mutation did not prevent memory loss in this mouse model of AD [13]. In recent years, studies have provided compelling evidence for the role of oxidative stress in sustained inflammatory responses in the AD brain [14–20]. Novel strategies have been proposed to prevent neuronal damage due to oxidative stress [21].

Phospholipases A₂ in AD—a Link to Oxidative and Pro-inflammatory Pathways

Phospholipases A₂ (PLA₂) are ubiquitous enzymes known for their role in the maintenance of cell membrane integrity and the production of lipid mediators that regulate cell functions [22, 23]. More than 20 species of PLA₂ are present in mammalian cells and are categorized as Ca²⁺-dependent cytosolic (cPLA₂), Ca²⁺-independent (iPLA₂), and secretory (sPLA₂) subtypes. In the central nervous system (CNS), PLA₂ activity has been implicated to play an important role in neurodegenerative diseases [24]. Recent studies also linked PLA₂ activity to ROS production and lipid peroxidation. A major objective of this review is to describe recent studies linking cPLA₂ and sPLA₂ to oxidative pathways in AD.

Cytosolic Phospholipase A2

In earlier studies, an increase in functional cPLA₂ protein or mRNA expression in AD brain has been demonstrated by immunohistochemical analysis [25, 26], reverse transcriptase polymerase chain reaction [27] or detection of lipid hydrolysates [28]. cPLA₂ activity is regulated by a number of protein kinases, which phosphorylate the protein on specific serine residues. Extracellular signal-regulated kinase (ERK)-dependent cPLA₂ phosphorylation on Ser505 is critical for cPLA₂ activity [29]. However, studies to determine cPLA₂ activity in human AD brain have provided for contrasting results. For example, increases in levels of activated cPLA₂ were observed in the hippocampus of human AD brain and in the hAPP mouse model of AD [30]. The increase in cPLA₂ activity in AD brain was marked by the increase in 4-hydroxynonenal, an index of lipid peroxidation [31, 32]. On the contrary, other studies indicated a reduction in cPLA₂ activity in blood and brain samples from AD patients [33]. During the early stages of AD, cognitive impairment associated with aberrant cholinergic and glutamatergic activities has been linked to a decrease in both cPLA₂ and iPLA₂ activities [34]. A study utilizing magnetic resonance imaging also demonstrated reduced phospholipid turnover in the pre-frontal cortex of AD patients [35]. Furthermore, decreased cPLA₂ activity was found in cerebrospinal fluid from AD, vascular, and mixed AD–vascular demented patients, as compared to healthy controls [33]. Although the reason for the differences in results is not yet clear, it is recognized that this enzyme is regulated by complex factors depending on the cell types and intracellular signaling pathways. Since cPLA₂ and its hydrolytic products are critically important in maintenance of phospholipids in cell membrane, more studies are needed to understand cell specific factors regulating cPLA₂ under physiological and pathological conditions.

Secretory Phospholipase A₂

Mammalian sPLA₂s are generally small molecular weight proteins (~14–19 kDa) with six to eight disulfide bridges. There are more than ten isoforms of sPLA₂s in mammals including groups IA-B, IIA-F, III, V, IX, X, XIA-B, XII, XIII, and XIV [22, 36]. Recent studies have focused on sPLA₂-IIA because this sPLA₂ is a pro-inflammatory protein and is upregulated in coronary artery diseases, atherosclerosis, sepsis, arthritis, and infection [37]. In rodents, most studies on sPLA₂-IIA are limited to rats because many mouse strains lack functional sPLA₂-IIA [38]. Transgenic mice overexpressing human sPLA₂-IIA showed increases in hepatic cholesterol and in collagen deposits in macrophages, suggesting a pro-inflammatory role for this protein [39, 40]. Our studies indicated upregulation of sPLA₂-IIA in reactive astrocytes in response to injury due to focal cerebral ischemia in rats [41]. In a subsequent study, we also showed an increase in sPLA₂-IIA mRNA expression in AD brain as compared to age-matched controls [42]. These results are in agreement with the recent report indicating an increase in sPLA₂ activity in cerebrospinal fluid of AD patients as compared to healthy control subjects [43].

Substantial cross-talk seems to exist between the path-ways that regulate the expression/activation of cPLA₂- and sPLA2-dependent inflammatory responses [44]. It appears that this cross-talk is linked to oxidative stress mediated by NADPH oxidase [45].

NADPH Oxidase is an Important Source of Oxidative Stress in the CNS and in AD

Recent studies have demonstrated the role of NADPH oxidase as an important nonmitochondrial source of ROS in many cell and tissue types, including the CNS [46–48]. Although NADPH oxidase activity is necessary for cell signaling under physiological conditions, its hyperactivation has been shown to contribute to the pathogenesis of a variety of diseases, including neurodegenerative diseases including vascular and neurodegenerative diseases and stroke [49, 50].

There are at least seven known isoforms of NADPH oxidases, NOX1-5, and Duox1 and 2, each with a unique combination of subunits [51]. NOX2 is well studied in phagocytic cells, macrophages, and endothelial cells and is comprised of the subunits p47phox, p67phox, p40phox, and Rac 1 in the cytosol and gp91phox and p22phox in the membrane fraction (plasma membranes or other subcellular membranes). Activation of NOX2 is dependent on phosphorylation of the cytosolic subunits, e.g., phosphorylation of p47phox by protein kinase C in human monocytes [52] and by Src-mediated tyrosine kinase in lung endothelial cells [53]. Subsequently, the cytosolic subunits form a complex and translocated to the membrane-associated gp91phox subunit. The gp91phox subunit has six trans-membrane domains and four heme-binding histidines in transmembrane domains III and V (see Bokoch et al. [54] for a recent review). FAD-binding and NADPH-binding domains are present in the carboxy-terminus of gp91phox [55]. NOX1, 2, and 4 are expressed in neurons, astrocytes, and microglia in the CNS [48]. Although the subcellular sites of NOX isoform expression in cells in the central nervous system have not been clearly determined, recent studies suggest that multiple NOX isoforms can be expressed within the same cell [56].

Recent studies have demonstrated the ability for Aβ to upregulate different NADPH oxidase subunits in microglial cells and increase enzyme activity in AD patients as compared to age-matched controls [57–59]. These studies suggest a relationship between ROS produced from NADPH oxidase and oxidative stress in AD.

Involvement of NADPH Oxidase and cPLA₂ in N-Methyl-D-Aspartic Acid-Induced Neuronal Excitotoxicity

Recent studies with cultured neurons have demonstrated the involvement of NADPH oxidase in excitotoxicity induced by ionotropic glutamate receptors, including the N-methyl-D-aspartic acid (NMDA) subtype [50, 60–62]. NMDA-induced activation of NADPH oxidase can trigger signaling pathways leading to activation of ERK1/2 [63], the protein kinase required for activation of cPLA₂ (Fig. 1). Studies in our laboratory with primary cortical neurons demonstrated that oligomeric Aβ can induce ROS production through NADPH oxidase. In turn, ROS from NADPH oxidase can stimulate ERK1/2 phosphorylation and activation of cPLA₂ and result in a release of arachidonic acid (AA) [61]. This study further demonstrated that the excitotoxic effect of Aβ is inhibited by NMDA receptor antagonists, including memantine, a drug used to treat AD patients. Stimulation of this signaling pathway should have important physiological consequences since (1) cPLA₂ activation modulates membrane phospholipid homeostasis that is known to be altered in a number of neurodegenerative diseases [64, 65], (2) cPLA₂-dependent production of AA leads to an increase in the synthesis of bioactive eicosanoids [66, 67], (3) AA itself can act as a retrograde messenger in neurons thereby modulating learning and memory processes [68], and (4) lysophospholipids, e.g., lysophosphatidylcholine, produced by cPLA₂ may alter the membrane microenvironment and thereby further alter membrane protein functions [36, 69].

Fig. 1.

The role of Aβ in the coupling of NADPH oxidase and NMDA receptor signaling pathways in neurons

The involvement of cPLA₂ in neuronal excitotoxicity, dysfunction, and death in vivo and in vitro has been suggested by several studies. However, despite some indirect evidence, the mechanism whereby cPLA₂ alters mitochondrial membrane and triggers mitochondrial apoptotic pathways remains to be investigated. A study by Kriem et al. demonstrated the role of cPLA₂ in mediating neuronal apoptosis induced by oligomeric Aβ [70]. Using selective inhibitors, cPLA₂ and iPLA₂ were shown to play a role in Aβ-mediated loss of mitochondrial membrane potential and increase in ROS in astrocytes [71]. Neurons from cPLA₂ knockout (KO) mice show less NMDA-induced injury as compared with wild-type controls [72]. Transgenic mice overexpressing human APP in neurons exhibited a reduced deficit in learning, memory, and behavioral dysfunctions and diminished Aβ-induced neurotoxicity when crossed with cPLA2 KO mice [30]. Similar reduction of Aβ-induced neurotoxicity was observed using cPLA2 inhibitors [70]. Furthermore, squalestatin was shown to be neuroprotective against Aβ-induced neurotoxicity by inhibiting cPLA2 activation [73].

Although there is convincing evidence suggesting a role for glial cell NADPH oxidase in Aβ-induced neurotoxicity [74], relatively few studies have investigated the role of NADPH oxidase in neurons. Antisense oligonucleotide knockdown of p22phox expression inhibited Aβ-induced neuronal apoptosis [75]. The NADPH oxidase inhibitor apocynin was also effective in diminishing the cytotoxic effects of Aβ [76]. Furthermore, diminished oxidative stress, neurovascular dysfunction, and behavioral deficits were observed in the Tg2576 mouse model of AD when these mice were crossed with gp91phox KO mice [77].

NADPH Oxidase and cPLA₂ and sPLA₂ in Glial Cells

Relatively high levels of NADPH oxidase are found in astrocytes and microglial cells, as compared with neurons, and ROS produced from glial cells has been shown to cause neuronal damage [74, 78]. ROS produced by NADPH oxidase modulates the cytokine-induced activation of NF-κB, a transcription factor that regulates the expression of pro-inflammatory genes, including cyclooxygenase-2 (COX-2), sPLA₂-IIA, and inducible nitric oxide synthase (iNOS; Fig. 2). Our studies showed that IL-1β-induced sPLA₂-IIA mRNA and protein expression in rat astrocytes could be inhibited by polyphenol antioxidants and apocynin, a known NADPH oxidase inhibitor [79]. These results are consistent with other studies indicating that Aβ and cytokines potentiate ROS production via NADPH oxidase activation in glial cells [80, 81]. Arachidonic acid is another activator of NADPH oxidase that causes superoxide release from microglia and induces their proliferation [82]. Using neuron–glial cell coculture, per-oxynitrite was shown to be produced by NO release from iNOS and ROS from NADPH oxidase in glial cells, and peroxynitrite is the potent cytotoxic factor that kills neurons.

Fig. 2.

Effects of Aβ on cytokine-induced inflammatory responses in glial cells: the role of NADPH oxidase

Although studies with phagocytes indicate that cPLA₂ targets NADPH oxidase [83, 84], the role of cPLA₂ in the activation of NADPH oxidase in neurons and glial cells has not been well characterized. Studies with neutrophils demonstrate the formation of a complex between the C2 domain of cPLA₂ and the p47 phox-PX domain of NADPH oxidase [84]. In BV-2 microglial cells, which lacks sPLA₂-IIA, Aβ could induce activation of NADPH oxidase and this was linked to a rapid increase in cPLA₂ phosphorylation and a delayed increase in expression which was attributed to the NF-κB pathway. Aβ also enhanced production of PGE₂ through conversion of AA by COX-2. In this study, Aβ also led to induction of iNOS and release of NO, but this was thought to be mediated through activation of the PGE₂ receptor and the down-stream PKA-CREB pathway. Interestingly, inhibition of cPLA₂ by antisense led to a decrease in NADPH oxidase activity and release of superoxides, PGE₂ formation, iNOS expression, and NO production [85]. Taken together, this study provided results demonstrating a close relationship between NADPH oxidase and cPLA₂ in Aβ-mediated inflammatory responses in microglial cells.

In wild-type mice, Aβ-induced neurotoxicity is enhanced by microglia that produce superoxide from NADPH oxidase, a response that is inhibited in gp91phox KO mice [78]. Aβ also has been shown to increase ROS production in astrocytes through the calcium-dependent activation of NADPH oxidase, and oxidative stress results in glutathione depletion and neuronal death [86]. Peroxynitrite appears to be an important factor mediating neurotoxicity in vitro [87] since generation of pathophysiological concentrations of superoxide and NO are necessary but not sufficient to induce neurodegeneration [88]. It is noted that these studies do not exclude the possibility that neurotoxicity is mediated by other factors, such as pro-inflammatory cytokines and protein modification through cysteine S-nitrosylation [89–92].

Targeting NADPH Oxidase Activity for Neurodegenerative Diseases

Recognition of the important role of NADPH oxidase in generating ROS that regulate receptor signaling pathways has highlighted an urgent need to develop novel approaches and pharmacological agents to regulate this enzyme system [93, 94]. In this regard, it is important to identify compounds that directly modulate the activities of NADPH oxidase, either through binding to specific protein subunits or through scavenging the ROS produced. The recent study by Jaquet et al. provided a list of small molecules useful as NOX inhibitors [95]. Although compounds such as diphenyleneiodonium, apocynin, atorvastatin, and 4-(2-amino-ethyl)benzenesulfonyl fluoride have been used to inhibit NADPH oxidase, these compounds are rather nonspecific. On the other hand, gp91ds-tat, a peptide inhibitor, can specifically block NOX2 [48]. Since different cell types can produce different amount of ROS, it is also important to develop cell-special inhibitors for NADPH oxidase, in particular, for targeting microglial cells [96, 97]. Of particular interest is dextromethorphan, a noncompetitive NMDA receptor antagonist and antitussive agent that has been shown to inhibit several NADPH oxidase-mediated responses, including degeneration of dopaminergic neurons in a 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine Parkinson’s model and endotoxic shock in mice [98, 99]. Another approach is to inhibit NADPH oxidase-mediated responses including inflammation and oxidative stress by nutritional means using natural products from plants [100]. Indeed, the use of plant antioxidants has gained considerable popularity in recent years [95].

Prevention of Neurodegenerative Diseases by Natural Antioxidants

Fruits and vegetables are known to contain phenolic compounds that exhibit antioxidant properties and have been shown to exert health benefits and reduce risk of major diseases, including cardiovascular and neurodegenerative diseases (see review in [101]). The underlying mechanisms for neuroprotection, however, remain incompletely understood but likely involve the ability of polyphenols and plant extracts to scavenge ROS and counteract Aβ formation/aggregation directly or indirectly [102, 103]. Some phytochemicals may exert neurohormetic effects and protect neurons from injury by upregulating cell survival pathways [104]. There is also evidence that some polyphenols, such as curcumin, may inhibit Aβ fibril formation and destabilize preformed fibril Aβ [102, 105–107]. Other studies have demonstrated that some natural products provide neuroprotection by targeting multiple cellular signaling pathways [108, 109]. Recent in vitro and in vivo studies have indicated that purified botanical compounds or plant extracts effectively prevent Aβ-induced neurotoxicity (Table 1) as well as provide neuroprotection in animal models of AD (Table 2). Since our recent review has described the use of phenolic compounds such as resveratrol, curcumin, and apocynin as neuroprotective agents [101], these compounds are not included in Tables 1 and 2.

Table 1.

Polyphenols and botanical extracts tested against amyloid-beta toxicity

Polyphenol/plant name	Cell type	Effects	References
Apigenin	Cortical neurons	+	[122]
Baicalein	PC12 cells	+	[123]
			[124]
Biflavonoids	SH-SY5Y cells	+/−	[125]
Catechin	PC12 cells	+	[126]
			[127]
	Cortical neurons	+	[128]
Cocoa	PC12 cells	+	[127]
Cyanidin 3-O-glucoside	SH-SY5Y cells	+	[129]
Daidzein	Hippocampal neurons	+/−	[130]
EGCG	PC12 cells	+	[131]
	Hippocampal neurons	+	[132]
			[133]
Epicatechin	PC12 cells	+	[127]
			[124]
	Cortical neurons	+	[128]
	Hippocampal neurons	−	[133]
Gallic acid	Hippocampal neurons	+	[133]
Genistein	SH-SY5Y cells	+	[134]
	Hippocampal neurons	+/−	[130]
	Hippocampal neurons	+/−	[135]
	Cortical neurons	+	[136]
		−	[122]
Ginkgo biloba (EGb 761)	Hippocampal neurons	+	[137]
	N2a neuroblastoma	+	[138]
	SH-SY5Y cells	+	[139]
	PC12 cells	+/−	[140]
Grape seed	PC12 cells	+	[141]
			[142]
Hibifolin	Cortical neurons	+	[143]
	PC12 cells	+	[124]
Hypericum perforatum	Hippocampal neurons	+	[144]
Icaritin	Cortical neurons	+	[145]
Kaempferol	PC12 cells	+	[146]
			[124]
	Cortical neurons	+	[122]
Morin	HT22 neuroblastoma	+	[106]
Mulberry (leaf extract)	Hippocampal neurons	+	[147]
Myricetin	Cortical neurons	+	[148]
Naringenin	PC12 cells	+	[149]
		−	[124]
	Cortical neurons	−	[122]
Puerarin	PC12 cells	+	[150]
Pycnogenol	PC12 cells	+	[151]
Quercetin	Cortical neurons	+/−	[152]
		−	[122]
	HT22 neuroblastoma	+	[106]
Rhizoma acori graminei	PC12 cells	+	[153]
Scutellarin	PC12 cells	+	[124]
Smilacis chinae	Cortical neurons	+	[128]
Tea extracts	Hippocampal neurons	+	[133]

Table 2.

Polyphenols and botanical extracts tested in AD animal models

Polyphenol/plant name	Animal model	Effects	References
Bacopa monniera	PSAPP mice	+	[154]
Baicalein	Aβ infusion (i.c.v.) mice	+	[155]
Blueberry	Tg2576 mice	+	[156]
Cabernet Savignon	Tg2576 mice	+	[157]
EGCG	Tg2576 mice	+	[158]
			[159]
	Aβ infusion (i.c.v.) mice	+	[160]
	PS2 mice	+	[160]
	LPS (i.c.v.) mice	+	[160]
Epicatechin	Aβ infusion (CA1) rats	+	[161]
Ferulic acid	Aβ infusion (i.c.v.) mice	+	[162]
			[163]
			[164]
	Tg2576 mice	−	[165]
Fustin	Aβ infusion (i.c.v.) mice	+	[166]
Garlic	Tg2576 mice	+	[167]
			[168]
	TgCRND8 mice	+	[169]
Ginkgo biloba	Tg2576 mice	+	[170]
			[171]
	TgAPP/PS1 mice	+	[172]
			[173]
Ginseng	Tg2576 mice	+	[174]
Ginsenoside	Rb1or M1Aβ infusion (i.c.v.) mice	+	[175]
Glycyrrhiza uralensis	Aβ infusion (i.c.v.) mice	+	[176]
Grape seed	Tg2576 mice	+	[177]
Green tea catechins	Aβ infusion (i.c.v.) rats	+	[178]
Luteolin	Aβ infusion (i.c.v.) mice	+	[179]
Nobiletin	APP–SL 7–5 mice	+	[180]
	Aβ infusion (i.c.v.) rats	+	[181]
Oroxylin A	Aβ infusion (i.c.v.) mice	+	[118]
Pomegranate	Tg2576 mice	+	[182]
Rosmarinic acid	Tg2576 mice	+	[165]
Silibinin	Aβ infusion (i.c.v.) mice	+	[183]
Soy isoflavone	Aβ infusion (i.c.v.) rats	+	[184]

Dietary flavonoids are potent inhibitors of NADPH oxidase. Figure 3 shows the chemical structures of some flavonoids as described in Table 1. A study of 45 compounds indicated that flavanols inhibit NADPH oxidase through an apocynin-like mechanism [110]. Other studies demonstrate the ability of dietary polyphenols to inhibit NADPH oxidase, suggesting that these polyphenols may serve as novel therapeutic agents in neuroinflammation [111]. Some polyphenols, like those from grape seed extract, also have been shown to regulate NADPH oxidase subunit expression [112].

Fig. 3.

Chemical structures of some flavonoids described in Table 1

Dreiseitel et al. used a kinetic photometric model to compare the potency of a number of anthocyanidins for inhibition of sPLA₂ [113]. Anthocyanidins have been shown to ameliorate cognitive deficits in AD patients and improve immunocompetency of these patients [114]. The polyphenol genistein was shown to be a potent inhibitor of sPLA₂ in inflammatory exudates and in snake venom-induced mouse paw edema [115]. It appears that anti-inflammatory activities of many plant flavonoids are associated with inhibition of PLA₂ [116]. These studies provide strong rationale to search for novel compounds as specific inhibitors of PLA2s [117–121].

Summary

In summary, AD progression is marked by an increase in oxidative stress and chronic inflammation that is attributed in part to the toxic effects of Aβ. Studies in recent years have identified NADPH oxidase as an important source of ROS that contributes to Aβ-induced neuronal damage and glial cell activation. ROS produced by NADPH oxidase activate MAPK and subsequently cPLA₂, a key enzyme that releases AA from phospholipids for the synthesis of eicosanoids (Fig. 1). cPLA₂ activation and AA release have been associated with neuronal excitotoxicity, impairment of mitochondrial dysfunction, and neuronal apoptosis. In addition, ROS produced by NADPH oxidase can activate NF-κB to promote pro-inflammatory gene transcription, thereby enhancing the synthesis of sPLA₂, iNOS, and COX-2, enzymes that play a role in neurodegenerative diseases (Fig. 2). There is compelling evidence for the beneficial effects of antioxidants from botanical sources as inhibitors of NADPH oxidase and PLA₂, and these polyphenolic compounds may become useful therapeutic agents to alleviate oxidative stress and chronic inflammation in neurodegenerative diseases, including AD.