ReviewNeuroscience

G Protein–Coupled Receptors, Cholinergic Dysfunction, and Aβ Toxicity in Alzheimer’s Disease

See allHide authors and affiliations

Science Signaling  20 Oct 2009:
Vol. 2, Issue 93, pp. re8
DOI: 10.1126/scisignal.293re8

Abstract

The β-amyloid (Aβ) peptide is associated with the pathogenesis of Alzheimer’s disease (AD). Evidence gathered over the last two decades suggests that the gradual accumulation of soluble and insoluble Aβ peptide species triggers a cascade of events that leads to the clinical manifestation of AD. Aβ accumulation has also been associated with the cholinergic dysfunction observed in AD, which is characterized by diminished acetylcholine release and impaired coupling of the muscarinic acetylcholine receptors (mAChRs) to heterotrimeric GTP-binding proteins (G proteins). Although the mechanism of Aβ-mediated toxicity is not clearly understood, evidence shows that Aβ accumulation has an effect on the oligomerization of the angiotensin II (AngII) AT2 (angiotensin type 2) receptor and sequestration of the Gαq/11 family of G proteins. Sequestration of Gαq/11 results in dysfunctional coupling and signaling between M1 mAChR and Gαq/11 and accompanies neurodegeneration, tau phosphorylation, and neuronal loss in an AD transgenic mouse model. Collectively, these results provide a putative link among Aβ toxicity, AT2 receptor oligomerization, and disruption of the signaling pathway through M1 mAChR and Gαq/11 and potentially contribute to our understanding of the cholinergic deficit observed in AD.

The Amyloid Hypothesis

Alzheimer’s disease (AD) is the most common neurodegenerative disorder afflicting the elderly. AD is clinically characterized by progressive neuronal loss and inflammation, memory impairment, cognitive deficits, and behavioral changes (1). Neuropathologically, AD-afflicted brains are characterized by two proteinaceous aggregates: amyloid plaques, which are mainly composed of the β-amyloid protein (Aβ), and neurofibrillary tangles (NFTs), which are made up of hyperphosphorylated aggregates of the tau protein (2). Two major hypotheses have driven pharmaceutical research in the search for a treatment for AD: the amyloid hypothesis (3) and the cholinergic hypothesis (4). Although considerable progress has been made toward understanding the pathophysiology of AD, many questions remain unanswered, including the potential link between amyloid pathology and the cholinergic deficit observed in AD patients and the relationship between Aβ generation and two other aspects of AD, neuronal death and NFTs. Evidence suggests that heterotrimeric GTP-binding protein (G protein)–coupled receptors (GPCRs) may provide a common underlying mechanism, and although these observations require further confirmation, they broaden our insight into an unanticipated spectrum of therapeutic targets for AD research and, perhaps, neurodegeneration in general.

Aβ is derived from proteolysis of the β-amyloid precursor protein (APP), a type I integral membrane protein, after sequential cleavage by the β- and γ-secretases. The γ-secretase is a tetrameric complex that cleaves APP within its transmembrane domain and so liberates the intact Aβ peptide, which ranges in length from 39 to 43 residues (5). The majority of Aβ produced is 40 amino acids in length (Aβ40), whereas a small proportion (~10%) is the 42-residue variant (Aβ42). Compared with Aβ40, Aβ42 is more hydrophobic, aggregates faster, and is more toxic, and it is the major Aβ species found in amyloid plaques (2, 5). According to the amyloid hypothesis, the chronic excessive accumulation and deposition of Aβ in amyloid plaques, Aβ42 in particular, initiates a pathogenic cascade, which leads to the development of AD (3); however, observations in AD mouse models and in the normal aging human brain reveal that amyloid plaques are also associated with minor neuronal alterations and do not necessarily correlate with the degree of cognitive impairment, which suggests that Aβ accumulation may not be strictly correlated with Aβ toxicity (68). A revision of the original hypothesis proposes that soluble Aβ oligomers, but not insoluble fibrillar Aβ deposits in amyloid plaques, are neurotoxic and responsible for the synaptic dysfunction in the brains of AD patients and in AD mouse models (3). Accordingly, the revised amyloid hypothesis provides a more appropriate explanation for familial forms of AD. In this regard, more than 30 mutations in APP (the substrate) and 180 mutations in the two presenilin genes (the protease) have been identified (www.molgen.ua.ac.be/ADMutations). These missense mutations account for less than 0.5% of all AD cases (9), whereas sporadic AD is more common and is caused by a combination of environmental and genetic factors.

The most basic and complex forms of Aβ—soluble monomers and insoluble fibrillar Aβ, respectively—display only a minor degree of toxicity (10). As a result, the concept of “Aβ-derived diffusible ligands” (ADDLs) or “soluble Aβ toxic oligomers” (11) has advanced during the last decade. Several different Aβ oligomeric species have been identified (1214), including ADDL-like oligomeric aggregates isolated from postmortem AD patient brains (15), which act as synaptotoxic ligands, inhibiting long-term potentiation (LTP) (11, 16, 17). Currently, there is no consensus as to which toxic Aβ aggregate is most relevant in vivo, although it is plausible that various toxic aggregates may coexist together. Aβ40 appears to be much less toxic, and even protective, in comparison with Aβ42 (18). Thus, the relative amount of Aβ42 might influence the equilibrium between toxic and inert assemblies of Aβ and may provide an explanation for the lack of correlation between the absolute amount of Aβ and Aβ toxicity (1921).

Tau

Tau is a highly soluble, phosphorylated protein that stabilizes and promotes the polymerization of microtubules in neurons, primarily through its microtubule-binding domain. Phosphorylation of the microtubule-binding domain of tau is believed to be crucial in regulating microtubule stabilization (22). Splicing generates two sets of tau isoforms, each containing either three (3R) or four (4R) microtubule-binding repeats with a differential affinity for microtubules. The 4R:3R tau ratio is approximately equal in the normal brain; however, disturbances, usually increases in the 4R tau, occur in most neurodegenerative tauopathies, including AD (23, 24). Tau abnormalities, whether they are due to mutations in the tau gene that result in amino acid changes or an altered 4R:3R tau ratio, cause deposition in the brain of highly phosphorylated tau in an aberrant conformation (24). As a result, hyperphosphorylation of tau leads to the detachment of tau from microtubules and the consequent destabilization of these vital structures (25). Unbound tau can then aggregate into insoluble paired helical filaments (PHFs) that coalesce into NFTs (23). Normally, tau is confined to axons, but in neurodegenerative tauopathies, it becomes missorted to cell bodies and dendrites (26). Thus, the three most visible AD-related changes in tau are hyperphosphorylation, aggregation, and mislocalization.

The Cholinergic Hypothesis

The degeneration of cholinergic neurons is considered one of the seminal features of AD. Cholinergic neurons, which are thought to be involved in cognition and attention (27), provide the primary source of cholinergic innervations to the cerebral cortex, hippocampus, and amygdala (28), major areas of the brain that are affected in AD. Cholinergic dysfunction is characterized by a reduction in acetylcholine (Ach) synthesis due to reduced choline acetyltransferase (ChAT) and choline uptake, cholinergic neuronal and axonal abnormalities, and degeneration of cholinergic neurons (29). In this context, cholinergic deficits correlate well with amyloid plaque formation, NFTs, and the clinical severity of dementia (30). Consequently, the cholinergic hypothesis suggests that a dysfunction in ACh-containing neurons contributes, to a large extent, to the cognitive and behavioral deficits observed in AD patients (4) and has prompted the development of treatment strategies aimed at restoring cholinergic function through the use of acetylcholinesterase inhibitors and cholinergic muscarinic receptor agonists.

As a key neurotransmitter involved in learning and memory (31), ACh binds to two distinct receptor subtypes in the brain: nicotinic (nAChR) and muscarinic (mAChR). Whereas nAChRs are ligand-gated ion channels, mAChRs are metabotropic receptors. Five distinct mAChR subtypes (M1 to M5), each encoded by a different gene, have been identified in the CNS (32, 33). The M1 mAChR is the most abundant subtype in the cerebral cortex and hippocampus, the two main brain regions that develop amyloid plaques and NFTs (32). Loss of basal forebrain cholinergic neurons, along with reduced ChAT, ACh release, and M2 mAChR binding, cause a presynaptic cholinergic deficit in AD (3437). In contrast, the abundance of the predominantly postsynaptic M1 mAChR remains relatively unchanged (35, 38), although some reports that indicate normal amounts of radioligand binding to the M1 mAChR conflict with other reports of reduced M1 mAChR immunoreactivity (39, 40). One study suggests that M1 mAChR abundance is elevated in the AD brain (41).

M1 mAChRs selectively couple to the Gαq/11 family of G proteins. Stimulation of Gαq/11-coupled receptors can activate phospholipase C-β (PLCβ) and can lead to inositol phospholipid breakdown, protein kinase C (PKC) activation, and intracellular calcium mobilization (42). M1 mAChR–mediated activation of signaling cascades involving PKC and Src potentiates N-methyl-d-aspartate (NMDA) glutamate receptor function in hippocampal CA1 pyramidal cells (43). Potentiation of NMDA receptor currents regulates long-lasting forms of synaptic plasticity. NMDA receptor activity is also essential in the molecular processes that underlie learning and memory, which suggests a potential mechanism through which M1 mAChRs might exert cognition-enhancing effects (43). In accordance with these observations, M1 mAChR–deficient mice display a cortical memory dysfunction, such as impaired cross-talk between the hippocampus and the cortex, specific to short-term memory and memory consolidation (44).

Muscarinic Receptor Signaling in Alzheimer’s Disease

Several studies indicate that the interaction of GTP with muscarinic receptors is impaired in the AD brain (4547), which suggests that the coupling between receptors and G proteins may be dysfunctional. However, not all studies have found such a difference in the AD brain (48). Although these discrepancies might be partially explained by methodological approaches, researchers have sought to more directly measure muscarinic receptor–activated phosphoinositide (PI) hydrolysis in the postmortem human brain. Unfortunately, this strategy has not resolved the controversy: Analysis of phosphatidylinositol-4,5-bisphosphate (PIP2) hydrolysis suggests that there is not a significant impairment in PI signaling in the AD brain (49), whereas other studies show that both G protein and muscarinic receptor–stimulated activation of PI hydrolysis is impaired in the AD prefrontal cortex (4951). Moreover, a small study utilizing three brain samples found no difference in PI hydrolysis between control and AD samples but did find a correlation between the postmortem interval and muscarinic receptor–stimulated PI hydrolysis (52). In contrast, a study conducted by Greenwood et al. (50) utilized 16 control samples and did not find a correlation between the postmortem interval and muscarinic receptor–stimulated PI hydrolysis.

As previously suggested (52), the agonal state (the condition of an individual immediately preceding death, such as prolonged hypoxia or oxidative stress) of patients should be considered and may compromise biochemical analysis of brain samples. Although agonal influence on PI signaling has not been measured in AD brain samples, studies conducted on PI hydrolysis in samples with a short postmortem interval, as well as with a considerably lower brain tissue pH or acidosis associated with a prolonged agonal state, indicate that pH and the postmortem interval do not adversely affect measurements of PI signaling (53, 54).

A deficit in AD in muscarinic receptor activation is associated with dysfunctional activation of Gαq/11 (38, 50, 51, 55). Gαq/11 abundance appears to remain unaltered in the AD brain (5557). Moreover, Gαq/11-stimulated signaling cascades are critical in neuronal communication, synaptic plasticity, and neuronal survival (58, 59). Given that constitutive and muscarinic agonist–dependent Gαq/11 signaling appears to be impaired in AD patient brain samples (discussed above) (38, 50, 51), these findings suggest that, although the abundance of M1 mAChRs and Gαq/11 is apparently not altered, the AD brain displays a disturbance in effective coupling between M1 mAChRs and Gαq/11. Nevertheless, studies have also suggested that M1 mAChR abundance is reduced in the human AD brain (40) and in an AD mouse model (60).

Muscarinic Agonists

Several muscarinic agonists with varying degrees of selectivity for the M1 mAChRs have been evaluated in different animal models for their efficacy in the treatment of AD (61). In particular, the development of a transgenic mouse model of AD (3xTg-AD), harboring the human PS1M146V, APPSwe, and tauP301L transgenes, which is characterized by the age-associated development of Aβ plaques, NFTs, cholinergic dysfunction, and cognitive impairments (6264), has permitted the evaluation of potential therapeutic compounds. Moreover, treatment of 3xTg-AD mice with the M1 mAChR–selective agonist AF267B resulted in a decrease in amyloidogenic processing of APP, a reduction in both the amyloid and tau pathologies, and a reversal of cognitive deficits (65). These findings established the efficacy of M1-selective agonists for the treatment of AD in mice. AF267B (commercial name NGX267) is now in early Phase II clinical trials (61).

A recent report has also opened the door for the development of highly selective allosteric modulators of the muscarinic receptors for the treatment of AD. Characterization of an allosteric modulator specific for the M1 mAChR indicates that TBPB [1-(1′-2-methylbenzyl)-1,4′-bipiperidin-4-yl)-H-benzo[d]imidazol-2(3H)-one] can potentiate M1-induced regulation of NMDA receptor currents in hippocampal pyramidal cells, increase nonamyloidogenic processing of APP, and decrease Aβ production in vitro (66). This and other studies suggest that selective activation of the M1 mAChR with an allosteric modulator may provide a highly selective alternate approach for the treatment of AD.

The Janus kinase 2 (JAK2)−signal transducer and activator of transcription factor 3 (STAT3) signaling pathway has been suggested to enhance the responsiveness of the M1 mAChR to agonist stimulation under normal conditions (67). In contrast, Aβ accumulation or neurotoxic insults, such as oxidative stress (68), deficient protein degradation (69), and age-dependent decreased trophic-factor release (70), may lead to decreased JAK2−STAT3 activation and disturb cholinergic neurotransmission potentially through desensitization of M1 mAChRs. In addition, phosphorylated STAT3 is reduced in the hippocampus of sporadic AD patients and in an AD mouse model (67), which suggests that STAT3 inactivation may correlate with the pathogenesis of AD. Nevertheless, the breakdown of M1 mAChR–mediated intracortical signal transduction is hypothesized to underlie the limited efficacy of cholinergic replacement therapies.

Angiotensin Receptors in Alzheimer’s Disease

The classic function of the renin-angiotensin system (RAS) is to regulate systemic blood pressure and maintenance of water and electrolyte homeostasis (71). These functions appear to be under the control of angiotensin II (AngII) acting through AT1 receptors. Angiotensins have been implicated in other functions, such as learning and memory, cerebral protection, renal blood flow, and natriuresis, as well as pathological states, such as stress and depression (71). Although a definitive link between the RAS and AD remains elusive, several studies suggest that AngII, acting through AT2 receptors, mediates various central nervous system functions, including behavior (72), neuronal apoptosis (73), and memory (74). The AT2 receptor has also been reported to disrupt G protein activation, in particular, activation of Gαi/o and Gαq/11, by inhibiting AT1-stimulated activation of these G proteins (75). To further define the nature of the disruption of Gαq/11 activation in the pathophysiology of AD, AbdAlla et al. (76, 77) demonstrate that the prevalence of AT2 receptor oligomers in the human AD brain is accompanied by sequestration of Gαq/11 and reduced constitutive and muscarinic agonist–dependent Gαq/11 activation.

The AT2 receptor has been previously reported to heterodimerize with the AT1 receptor and to perturb AT1-mediated signaling by preventing the coupling of the AT1 receptor with Gαi/o and Gαq/11 (75). Murine hippocampal injection of Aβ initiates AT2 receptor homodimerization and oligomerization (76). Aβ has also been reported to interact directly with and to trigger homodimerization of APP (78, 79). Aβ binds to monomers and trimers of the 75-kD neutrophin receptor (p75NTR) and induces receptor activation and apoptosis (80). Evidence suggests that the cellular prion protein (PrP) is a specific binding partner for Aβ42 oligomers and mediates the inhibitory effect of Aβ42 on synaptic plasticity (81). Accordingly, Aβ has been shown to induce neuronal death in cell culture and in the brain in vivo (13, 17, 82), to impair mAChR signaling, and to suppress inositol phospholipid hydrolysis and calcium release in rodent neuronal cultures (83). Given that the mechanism of the Aβ-mediated dimerization of the AT2 receptor has not been elucidated, it remains unclear how the interaction of Aβ with other proteins might contribute to the overall effects observed in the work by AbdAlla et al. (76, 77, 83).

Oxidative Stress in Alzheimer’s Disease

Oxidative stress is closely associated with the neuropathology of AD. Oxidative stress in AD patients occurs as a result of genetic factors (such as the apolipoprotein E ε4 allele); germline mutations [such as in the APP, Presenilin-1 (PS1), and Presenilin-2 (PS2) genes]; environmental factors; life-style–related factors (such as smoking); and certain health conditions, such as diabetes mellitus, brain injury, hypertensive vascular disease, and hypercholesterolemia (84). Oxidative stress is found in tissues and fluids from patients with AD (living and postmortem brains) and cognitive diseases such as mild cognitive impairment (MCI) and Down syndrome.

Reactive oxygen species (ROS) are formed under normal conditions, and although they are chemically unstable and highly reactive, ROS concentrations are generally kept relatively low by efficient antioxidant systems (85). However, in some situations, ROS generation can exceed the ability of the body to neutralize them, which results in oxidative stress (86). The brain is particularly vulnerable to oxidative stress because of its high energy and oxygen consumption rate, abundance of peroxidizable fatty acids, and relative paucity of antioxidant enzymes relative to other organs (87). The brain’s high content of transition metals, which can catalyze the formation of ROS and low antioxidant content in comparison with other organs, are additional potential risk factors. Indeed, several studies have found markers of oxidative stress in the AD brain, such as lipid peroxidation, protein and DNA oxidation, and glycoxidation (87).

The ability of toxic Aβ peptide species to induce protein oxidation and to inhibit the activity of oxidation-sensitive enzymes is consistent with the hypothesis that Aβ can induce severe oxidative damage. In this regard, treatment with an antioxidant or immunization with an Aβ peptide decreases AT2 receptor dimerization and reduces the Aβ plaque load, respectively, in an AD mouse model (76). Exposure to stressful conditions—such as water deprivation, noise, and flashing lights—triggers an increase in corticosterone concentrations and induces AT2 receptor oligomerization in AD transgenic mice. Similar to findings in AD patients, Gαq/11 associates with AT2 receptor oligomers only in stressed animals (76), which suggests that AT2 receptor oligomers sequester Gαq/11, which, in turn, may impair coupling between and signaling through M1 mAchR and Gαq/11. Indeed, constitutive and M1 mAChR–stimulated Gαq/11 activation are reduced in stressed AD transgenic mice (76).

To assess whether AT2 receptor sequestration of Gαq/11 is directly involved in reduced Gαq/11 activation, AbdAlla et al. (76) decreased AT2 receptor amounts in stressed AD transgenic mice. The consequent reduction of AT2 receptor oligomer formation was accompanied by a restoration of M1 mAChR signaling and amelioration of the Gαq/11 signaling deficit. Moreover, decreased AT2 receptors also reduced tau phosphorylation and Aβ generation. These results correlate with previous findings that demonstrate a reduction in or exacerbation of Aβ and tau pathologies in AD transgenic mice treated with M1 mAChR agonists or antagonists, respectively (65).

Tissue transglutaminase (tTG) mediates the cross-linking of transmembrane proteins by catalyzing the formation of stable intermolecular amide bonds (88). The abundance and activity of tTG is elevated in the AD brain and in cerebrospinal fluid from AD patients (89, 90). Moreover, hippocampal injection of tTG results in AT2 receptor oligomerization, sequestration of Gαq/11, and reduced Gαq/11 activation (76). tTG activity is higher in animals subjected to stressful conditions compared with nonstressed animals, and hippocampal injection of Aβ results in a dose-dependent increase in tTG activity (76), suggesting that AT2 receptor dimerization and oligomerization can be facilitated by oxidative stress and tTG. In this regard, transfection of cells with NOX-3, a ROS-generating NADPH (nicotinamide adenine dinucleotide phosphate, reduced form) oxidase, induces oxidative cross-linking of tyrosine residues in the cytoplasmic loops of transmembrane helices III and IV and dimerization of AT2 receptor monomers. Further oligomerization of the AT2 receptor dimers requires the action of tTG-2. Similar to the studies in the AD transgenic mice and AD patients, tTG-2–induced AT2 receptor oligomerization reduces Gαq/11 activation. In contrast, coexpression of an AT2 receptor mutant devoid of the cytoplasmic tail prevents AT2 receptor cross-linking and oligomerization in vitro and in vivo and delays the development of neurodegenerative symptoms in stressed AD mice. Thus, silencing of AT2 receptors by using RNA interference and expression of an AT2 receptor mutant disrupts AT2 receptor oligomerization, restores activation and signaling through M1 mAChR and Gαq/11, and alleviates the AD pathology in a mouse AD model (76, 77).

The Role of Other GPCRs in AD

An increasing number of GPCRs have been implicated in the pathophysiology of AD, and more specifically, in modulation of the amyloidogenic and nonamyloidogenic processing of APP. For example, the β2-adrenergic receptor (91) and the orphan GPCR GPR3 (92) participate in the mediation of Aβ generation by affecting the activity of the γ-secretase complex. Other studies have implicated the serotonin receptors 5-HT2a, 5-HT2c, and 5-HT4 in addition to the metabotropic glutamate receptor 1α (mGluR) in the stimulation of the nonamyloidogenic proteolysis of APP (9396) and in the inhibition of Aβ generation (97). Furthermore, 5-HT6 receptor antagonists can improve cognitive deficits and elevate cholinergic neurotransmission (98).

In contrast to the AT2 receptor, which indirectly potentiates the toxicity of Aβ by sequestering Gαq/11 and preventing M1 mAChR signaling, the GPCR glucagon-like peptide 1 (GLP-1) receptor can protect murine hippocampal neurons against Aβ-mediated toxicity following treatment with GLP-1 (99), although the mechanism of action remains to be elucidated. Another study suggests that treatment of neuronal cultures with an antagonist of the GPCR amylin receptor can effectively block the deleterious effects of Aβ-induced neurotoxicity (100). Thus, these studies implicate various GPCRs at multiple levels of APP processing and Aβ generation and suggest potential avenues for the development of efficacious therapeutic targets for AD.

Summary

The evidence suggests that oxidative stress is an element of the pathogenesis of AD; however, it remains unclear whether oxidative stress in AD causes damage or is a response to the damage. The studies by AbdAlla et al. (76, 77) suggest that oxidative stress induced by Aβ accumulation leads to an increase in ROS generation and dimerization of the AT2 receptors. The consequent further elevation of Aβ and tTG amounts initiates oligomerization of AT2 receptor dimers and sequestration of Gαq/11 by the AT2 receptor oligomers. The subsequent decrease in Gαq/11-coupling to the M1 mAChR and the reduction in M1 mAChR activation accompanies hippocampal neurodegeneration, tau phosphorylation, and neuronal loss, contributing to the development and exacerbation of the neuropathology of AD (Fig. 1). Taken together, these studies suggest a putative link among oxidative stress, Aβ toxicity, AT2 receptor oligomerization, and disruption of the M1 mAChR and Gαq/11 signaling pathway, and provide a potential explanation for the cholinergic deficit, neurodegeneration, and tau phosphorylation observed in AD.

Fig. 1

A proposed model of the pathogenesis of Alzheimer’s disease. Oxidative stress and Aβ accumulation trigger increased ROS generation by NOX-3, resulting in dimerization of the AT2 receptors. A further increase in Aβ accumulation coupled with an increase in transglutaminase abundance causes cross-linking and oligomerization of AT2 receptor dimers. The AT2 receptor oligomers sequester Gαq/11, which prevents coupling of Gαq/11 to the M1 mAChR. Gαq/11 signaling is involved in synaptic plasticity and memory consolidation, and it counteracts AD pathology. Sequestration of Gαq/11 results in tau phosphorylation and neuronal degeneration. The ultimate outcome is the clinical manifestation of AD.

Further work is needed to determine the role of this proposed mechanism in sporadic AD. Nevertheless, these studies challenge the current thinking in the AD field, which usually confines itself to one aspect of the molecular pathogenesis of AD at any given moment in time and compels the field to integrate different views and alternative hypotheses in a consistent molecular framework. Patients who fail to respond effectively to currently available cholinergic treatments may benefit from combining their treatment with an antiamyloid therapy to decrease AT2 dimerization and to alleviate the effect of reduced M1 mAChR–induced Gαq/11 signaling.

Acknowledgments

101.This project was supported by a Methusalem grant from the Flemish government, the Fund for Scientific Research (Flanders, Belgium), and the Federal Office for Scientific Affairs (IUAP P6/43/), Belgium.

References and Notes

View Abstract

Stay Connected to Science Signaling

Navigate This Article