Abstract
Alzheimer’s disease (AD) is the most common cause of dementia and is characterized by the progressive loss of memory and cognition in the aging population. However, the etiology of and therapies for AD remain far from understood. Astrocytes, the most abundant neuroglia in the brain, have recently aroused substantial concern due to their involvement in synaptotoxicity, amyloidosis, neuroinflammation, and oxidative stress. In this review, we summarize the candidate molecules of astrocytes, especially receptors and transporters, that may be involved in AD pathogenesis. These molecules include excitatory amino acid transporters (EAATs), metabotropic glutamate receptor 5 (mGluR5), the adenosine 2A receptor (A2AR), the α7-nicotinic acetylcholine receptor (α7-nAChR), the calcium-sensing receptor (CaSR), S100β, and cannabinoid receptors. We describe the characteristics of these molecules and the neurological and pharmacological underpinnings of these molecules in AD. Among these molecules, EAATs, A2AR, and mGluR5 are strongly related to glutamate-mediated synaptotoxicity and are involved in glutamate transmission or the clearance of extrasynaptic glutamate in the AD brain. The α7-nAChR, CaSR, and mGluR5 are receptors of Aβ and can induce a plethora of toxic effects, such as the production of excess Aβ, synaptotoxicity, and NO production triggered by changes in intracellular calcium signaling. Antagonists or positive allosteric modulators of these receptors can repair cognitive ability and modify neurobiological changes. Moreover, blocking S100β or activating cannabinoid receptors reduces neuroinflammation, oxidative stress, and reactive astrogliosis. Thus, targeting these molecules might provide alternative approaches for treating AD.
INTRODUCTION
Alzheimer’s disease (AD) causes the deterioration of memory and other cognitive abilities of the elderly and is characterized by the extracellular accumulation of amyloid-β peptide deposits (mainly Aβ42) and the intracellular aggregation of hyperphosphorylated tau protein in neurofilament tangles, as well as reactive astrocytosis and microglia-mediated neuroinflammation [1, 2]. As the most prevalent cell type (5- to 10-fold more numerous than neurons) in the central nervous system (CNS) [3], the astrocyte is believed to be a potential target for treating AD. Through their close structural connections with synapses, astrocytes are fundamental for metabolic and homeostatic regulation, synapse formation and elimination, as well as ionic equilibrium, and are also indispensable for the transport of essential neurotransmitters such as glutamate in the CNS [4, 5].
Glutamate, the major excitatory neurotransmitter in the mammalian CNS, is extremely important for the regulation of learning, memory, and cognition [6, 7]. However, long-lasting exposure to high levels of glutamate causes excitotoxicity [8, 9], which is regarded as an important mechanism in various brain pathologies, including AD [10, 11]. Astrocytes are the primary protectors against glutamate-mediated excitotoxicity and play a key role in glutamate homoeostasis [11, 12].
Furthermore, astrocytes synthesize and secrete substantial quantities of Aβ and contribute to the overall amyloid burden in the AD brain [13]. Astrocytes have increased levels of the three essential components for Aβ production: amyloid-β protein precursor (AβPP), β-secretase (BACE1), and γ-secretase [14, 15]. Recently, increasing evidence has suggested that potential receptors on astrocytes can bind to Aβ and lead to the neurotoxicity observed in AD [16–18].
Astrocytes undergo a series of molecular, morphological, and functional changes, which are referred to as reactive astrogliosis, in response to CNS damage or the progression of the disease. Recently, the reactive proliferation of astrocytes in AD has been observed [14, 19]. In the early stages of neurodegenerative processes in AD animal models, astrocytic atrophy is characterized by a markedly reduced arborization of astrocytic processes, the loss of tripartite synapses, and a resultant disturbance of synaptic connectivity [20]; at later stages, hypertrophic and hyperplastic astrogliosis-induced neuroinflammation is observed in response to the proximity of astrocytes to extracellular accumulations of Aβ [20–22]. Reactive astrocytes release inflammatory factors, mainly TNFα, granulocyte-macrophage colony-stimulating factor, and S100β, which cause further activation of astrocytes and microglia, thus perpetuating inflammatory signaling cycles [23, 24]. Oxidative damage is also an early event in this brain pathology, and increased levels of oxidative markers have been detected in the brains of subjects with mild cognitive impairment, the early phase of AD [25]. Aβ exposure increases reactive oxygen species production in cultures of astrocytes [26, 27]. In addition to the participation in neuroinflammation and oxidative stress, the application of Aβ causes abnormal intracellular Ca2 + transients and mitochondrial depolarization in astrocytes long before an impairment is visible in neurons [28].
In general, reduced glutamate clearance, excessive Aβ production and secretion, increased inflammation and oxidative stress coupled with abnormal Ca2 + homeostasis in astrocytes have been proposed to underlie the synaptic injury and cognitive decline in AD. Although the exact regulatory mechanism of astrocytes in AD remains elusive, several candidate molecules have been proposed as promising targets in these processes. These molecules include excitatory amino acid transporters (EAATs), metabotropic glutamate receptor 5 (mGluR5), the Adenosine A2A receptor (A2AR), the α7-nicotinic acetylcholine receptor (α7-nAChR), the Calcium-sensing receptor (CaSR), S100β, and cannabinoid receptors. In this review, we summarize the key molecules, especially these receptors and transporters of astrocytes, and their contribution to the pathogenesis of AD. The rescue or antagonism of these molecules might provide an alternative approach to improving the cognitive deficits and neuropathological symptoms of AD.
RECEPTORS AND TRANSPORTERS INVOLVED IN GLUTAMATE-MEDIATED SYNAPTOTOXICITY IN AD
Excitatory amino acid transporters
The accumulation of residual glutamate in the synaptic cleft may induce a series of toxic effects, including oscillations of intracellular calcium influx, the activation of apoptosis-related genes, and the induction of neuronal death [8]. This finding suggests that the regulation of glutamate levels in the synaptic cleft is significant for preventing excitotoxic neuronal injury.
Glutamate-mediated synaptotoxicity and neuronal death represent causal phenomena in the neurodegeneration of AD. The elevated levels of Aβ impair glutamate uptake by astrocytes [29–31] and impair long-term potentiation and facilitate long-term depression by inhibiting glutamate uptake [32]. The main approach to removing synaptic glutamate is uptake by the high-affinity glutamate transporters, aka, the EAATs in astrocytes. Overall, 80–90% of the released extracellular glutamate is cleared by EAAT1 (the rodent analogue is referred to as glutamate aspartate transporter, GLAST) and EAAT2 (the rodent analogue is referred to as glutamate transporter 1, GLT-1) on astrocytes [33]. Lower levels of EAAT1 and EAAT2 have been found in the hippocampus and gyrus frontalis medialis of sporadic AD patients during the early clinical stage, which was particularly evident in the vicinity of amyloid plaques [34]. A reduction in EAAT2 levels was also observed in the temporal cortex [19]. EAAT2 has been shown to form dynamic signaling platforms through interactions with the water channel aquaporin-4 to regulate the neuroprotective effects against Aβ toxicity [35]. Decreased EAAT1 and EAAT2 levels have also been found in HA1800 cells treated with Aβ oligomers (AβO) and in cultured rat astrocytes exposed to Aβ40 [27, 36]. Similarly, the loss of GLT-1 and GLAST is also a common phenomenon in AD animal models and has been linked to the deposition of Aβ plaques and gliosis [27, 37–39]. When bred with partial GLT-1 loss GLT-1 (+/–) mice, which reduces GLT-1 to levels that are comparable to the losses in AD patients, the cognitive deficits of AβPPswe/PS1ΔE9 mice occurred earlier, at 6 months of age, and the ratio of insoluble Aβ42/Aβ40 was also increased [40]. In contrast, when bred with GLT-1 transgenic mice or treated with a novel GLT-1 translational activator, LDN/OSU-0212320, APPSw,Ind AD mice showed improved learning and memory functions, restored synaptic integrity, and reduced Aβ deposits. The observed beneficial effects lasted one month after the cessation of compound treatment [41].
Partial GLT-1 loss can also cause insulin/Akt signaling abnormalities in an AD mouse model [42]. GLT-1 heterozygosity in AβPPswe/PS1ΔE9 mice induced the sustained activation of Akt and disturbed the CNS insulin signaling cascade by causing a decrease in insulin receptor β (Irβ) activity and an increase in insulin receptor substrate 1 (IRS-1) activity, as well as a reduction in insulin degrading enzyme activity [40]. Therefore, the neuropathology of AD might be modified by restoring EAATs on astrocytes due to the roles of these transporters in glutamate uptake and insulin/Akt signaling.
Metabotropic glutamate receptor 5
mGluR5, a Gq-protein coupled receptor, is an important glutamate receptor in the CNS that is involved in neuronal survival, synaptic plasticity, and neural circuit formation [43, 44]. Recently, it has been proposed that the dysregulation of mGluR5 at the synaptic level might be the neurological basis for AD [45]. mGluR5 is not only expressed on neurons but is also widely expressed by astrocytes throughout the CNS gray and white matter [46]. Recently, a strong enrichment of mGluR5 in reactive astrocytes surrounding Aβ plaques was detected in brain sections from APPswe/PS1dE9 mice [47]. In vitro studies have also illustrated that AβO exposure increases the expression of mGluR5 and its downstream InsP3 receptor type 1 in hippocampal astrocytes [48]. Another study has demonstrated the diffusional trapping and aberrant clustering of mGluR5 on the plasma membrane of cultured rat embryonic astrocytes treated with AβO [18]. The clustering of mGluR5 has been found to intensify calcium influx and cause synapse deterioration, thus impeding the diffusion of membrane-attached AβO at synapses [47]. It has been reported that AβO increases the astrocytic release of ATP after treatment with the mGluR5 agonist DHPG; ATP then slows the diffusion of mGluR5 in astrocytes and neurons cocultured with astrocytes, but not in pure neuronal cultures [18]. Conversely, the mGluR5 antagonist MPEP prevented the ATP-induced slowing of mGluR5 diffusion [18]. Thus, astrocytic mGluR5 clusters and the subsequent synaptotoxicity might contribute to the pathogenesis of AD, while antagonism of the astrocytic mGluR5 might be an alternative approach to impeding the synaptotoxicity induced by AβO accumulation.
Adenosine A2A receptor
The adenosine receptors are a class of purinergic G-protein-coupled receptors that have adenosine as an endogenous ligand. A2AR belongs to one of the four types of adenosine receptors (A1, A2A, A2B, and A3) [49], and this receptor is widely expressed by various cell types in the brain, including neurons, microglia, and astrocytes [50, 51]. The A2AR has been suggested to be a main targeting molecule in AD [52–54] and controls the fundamental mechanisms governing neurodegeneration [55, 56]. A2AR upregulation in aged animal models and in the frontal cortex of AD patients has also been reported [57–59]. Moreover, both epidemiological and animal studies have demonstrated that the chronic consumption of caffeine, an adenosine receptor antagonist, is associated with a decreased incidence of dementia [60–64].
The role of the A2AR in the neurological disease is attributed, at least in part, to glial A2ARs and their involvement in neuroinflammatory and neuro-modulatory processes [65]. AD patients have increased levels of A2AR in astrocytes of the hippocampus compared to an aging cohort without dementia [59]. Similar to humans with AD, the level of astrocytic A2AR in the hippocampus of aging mice expressing human amyloid precursor protein (hAPP) was also increased, while the conditional knockout of astrocytic A2AR promoted memory loss in aging hAPP mice [59]. The knockout of astrocytic A2AR enhanced long-term contextual memory and increased the level of Arc/Arg3.1, an immediate-early gene required for long-term memory, in young and aging mice [59].
Given that the A2AR is enriched in glutamatergic synapses, this receptor may control synaptic plasticity and N-methyl-D-aspartate receptor (NMDAR) activity, a main determinant of neurodegeneration [66]. It has also been reported that the A2AR controls glutamate uptake by astrocytes both in sporadic AD patients and in cultured astrocytes and gliosomes [34, 68], which protects neurons from Aβ-induced neurotoxicity [34]. An Aβ42-induced decline in GLAST and GLT-1 expression was retrieved in primary cultured astrocytes dissected from the neocortex of global A2AR knockout mice [67]. Similar results were also obtained in an astrocyte preparation (gliosomes) from the cerebral cortex of rats that were intracerebroventricularly (icv) injected with Aβ42 [68]. These experiments indicated that antagonism of the astrocytic A2AR might be an efficient approach to modifying the glutamatergic dysfunction and damage in AD.
The glutamate-mediated synaptotoxicity engendered by astrocytic EAATs, A2AR, and mGluR5
The malfunction of the glutamatergic system is one of the main characteristics of the AD brain. At tripartite synapses, glutamate receptors and transporters and their downstream signaling pathways, which include CAMK, p38-MAPK, and JNK, finally lead to the neurodegeneration and cell death in AD [69]. Pathological observations have illustrated that the attenuation of glutamatergic transmission is an early indicator of neurodegeneration [69], and the decline of EAATs on astrocytes might be attributed to reduced glutamate metabolism. In the later stages of the disease, the diffusional trapping and aberrant clustering of mGluR5 in the reactive astrocytic membrane could be responsible for a further aggregation of Aβ and Aβ-mediated synaptotoxicity and neuronal death, which is further aggravated by the failure of glutamate reuptake caused by a decline in EAATs [70] and an increase in the A2AR level. Moreover, cholinergic dysfunction, particularly the alteration of α7-nAchRs on astrocytes, was found to be initiated in the early stages of the disease [71], prior to glutamatergic dysfunction, and is closely related to memory loss. Aβ engages astrocytic α7-nAChRs to induce glutamate release from astrocytes, with the extracellular concentration of glutamate approaching tens of micromolar [31].
ASTROCYTIC Aβ RECEPTORS
α7-nicotinic acetylcholine receptor (α7-nAchR)
The dysregulation of the cortical cholinergic system is known as one of the major characteristics of AD. Nicotinic acetylcholine receptors (nAChRs), the crucial neurotransmitters of the cholinergic system, are ligand-gated ion channels that form a central, cation-permeable channel consisting of five subunits, and the opening of the channel is gated to bind to acetylcholine [72]. One of the most widely distributed forms of nAChRs in the CNS is the homomeric α7-nAChR [73, 74], which has high permeability to Ca2 + [75, 76]. The α7-nAChR appears to participate in the development and differentiation of the CNS and, in particular, plays a crucial role in neuroprotection in AD [73, 77]. Among the nine subunits of the nAchRs, the α7-nAChR is the major subunit expressed by astrocytes of the human brain [74].
Recently, the astrocytic α7-nAChR has been reported to be involved in the pathogenesis of AD [72, 78]. α7-nAChRs on astrocytes were shown to be elevated in the hippocampus, entorhinal cortex, and temporal cortex of sporadic AD patients and subjects carrying the Swedish APP 670/671 mutation (APPswe) compared to aged-matched controls [74]. Furthermore, the increase in the astrocytic α7-nAChR level was positively correlated with the extent of neuropathological variations, such as the number of neuritic plaques, which illustrated the involvement of the astrocytic α7-nAChR in the Aβ cascade and deposition [74].
Interestingly, Aβ has been reported to bind to the α7-nAChR with high affinity [79]. The impact of Aβ on the α7-nAChR depends on the concentration and the state of aggregation of Aβ, which may thus display either neuroprotective or neurodegenerative effects in different stages of the disease [72]. It has been reported that at physiologically equivalent picomolar concentrations, Aβ42 peptides could activate α7-nAChRs and enhance spontaneous intracellular Ca2 + signals, thus leading to the release of glutamate from neurons, which further enhances long-term potentiation and memory [80]. While at low micromolar concentrations, both Aβ42 and Aβ25 - 35 inhibit α7-nAChR channels, trigger intercellular Ca2 + influx, and enhance the release of glutamate in cultures of purified rat and human astrocytes [31, 80]. Secreted glutamate triggers toxic effects, including an increase in NO production, p-tau oligomers, and caspase-3 activity, which then destroy the synaptic spines, the first step that leads to neuronal death [31, 80]. It has also been argued that the release of astrocytic glutamate induced by the interaction of Aβ and the astrocytic α7-nAChR activates the extrasynaptic NMDA receptors on neurons [81]. Increased gliotransmitters, such as ATP and D-serine, have also been detected [16]. Furthermore, Aβ exacerbates tau phosphorylation via α7-nAChR activation in mouse astrocytes [72]. Correspondingly, increases in the levels of α7-nAChR mRNA and protein were observed in primary cultures of astrocytes isolated from neonatal rat brains exposed to 0.1-100 nM Aβ42 [82], which might be a defensive or compensatory effect. However, 10-fold higher concentrations of Aβ42 did not alter the levels of the α7-nAChR in astrocytes, which exhibited pronounced neurotoxicity [82].
Because of the complicated roles of α7-nAChRs, which are neuroprotective or neurodegenerative, both α7-nAChR agonists and antagonists have been reported to protect against AD. For instance, pharmacological trials have suggested that α7-nAChR agonists enhance cognition and are neuroprotective [83], although it is not clear whether these effects result from activation or desensitization because α7-nAChRs rapidly desensitize following activation [84]. Similarly, α7-nAChR knockout mice crossed with Tg2576 mice transgenic for mutant hAPP showed severe learning and memory deficits and had increased levels of soluble oligomeric Aβ [85]. Furthermore, it has been demonstrated that α7-nAChR activation by nicotine reduces Aβ-induced cell apoptosis, while pretreatment with a selective α7-nAChR antagonist (methyllycaconitine) prevents the neuroprotective effect of nicotine [86]. However, α-bungarotoxin, a highly selective α7-nAChR antagonist, inhibits Aβ-induced glutamate release from astrocytes [31].
Despite the rapid desensitization of α7-nAChRs, the relations to the forms of Aβ aggregation (monomers, oligomers, and fibrils) as well as the Aβ-dose-dependent effects reduce the feasibility of exploring simple therapeutic strategies, and the promising alternative therapies are proposed. These therapies involve the application of positive allosteric modulators of α7-nAChRs that function only in the presence of endogenous agonists and do not affect the desensitization of the receptor [87, 88].
Calcium-sensing receptor
CaSR is a ubiquitously distributed class C G-protein-coupled receptor (GPCR) with a unique sequence that differs from that of other GPCR families, and the CaSR can bind to various G proteins, including Gqα, Giα, and G11α [89]. CaSRs are predominantly formed as homodimers (CaSR/CaSR) or heterodimers (CaSR/mGluR) on the plasma membrane, although these receptors also function as monomers [90]. The dimers are polymerized at the endoplasmic reticulum and glycosylated in the Golgi prior to being transported to the cell surface [91]. The main function of the extracellular CaSR is to regulate the homeostasis of free calcium, and research has also illustrated the roles of the CaSR in axon and dendrite development, cell proliferation and differentiation, and insulin secretion [92].
A growing body of evidence has indicated that the CaSR expressed by astrocytes plays vital roles in inflammatory and neurodegenerative diseases such as AD [93–95]. Exogenous Aβ42 oligomers bind to astrocytic CaSRs at the plasmalemma, thereby activating a set of intracellular signaling pathways that further exacerbate the intracellular accumulation and oversecretion of endogenous Aβ42 oligomers by hindering their proteolysis. Concomitantly, Aβ42 oligomers accumulating in the extracellular milieu can diffuse to reach more neurons and astrocytes and bind to the CaSR receptors on these cells, which leads to the production and release of additional Aβ42 oligomers [96]. Furthermore, the interplay of Aβ42 oligomers and CaSRs can also elicit the excess production and secretion of NO and the increased expression of nitric oxide synthase-2 (NOS2) and vascular endothelial growth factor-A (VEGF-A) by astrocytes as well as the increased expression and activity of GTP cyclohydrolase 1 [3]. Highly selective allosteric CaSR antagonists (calcilytics), such as NPS 2143 and NPS 89626, efficiently suppress these neurotoxic effects in cultured cortical human astrocytes [96, 97]. Similar results have also been obtained in human adult astrocytes that were isolated from a normal temporal cerebral cortex and exposed to fibrillar or soluble Aβ25 - 35 peptides, the Aβ42 proxy that evokes surplus endogenous Aβ42 production/accumulation [3]. The administration of NPS 2143 suppressed endogenous Aβ42 secretion, lowered the Aβ42/Aβ40 ratio, decreased the total CaSR protein complement, transiently increased the proteasomal chymotrypsin activity, and blocked excess NO production [3]. The progressive extracellular accrual and diffusion of Aβ42 oligomers in cultured human astrocytes were also blocked upstream by NPS 2143 [98, 99].
A recent research work also indicated that Aβ/CaSR signaling shifts hAPP from being acted on by α-secretase, which produces neurotrophic/neuroprotective soluble sAβPPα, to the cleavage of hAPP by β-secretase to create AD-driving Aβ42 oligomer peptides, while both NPS 2143 and NPS 89626 rescue the extracellular shedding of sAβPPα and suppress the neurotoxic effects of Aβ/CaSR signaling [96, 97]. It has also been shown that NPS 2143 drives the plasma membrane translocation of hAPP and increases the membrane translocation and specific enzymatic activity of ADAM10 (the principle α-secretase of hAPP, which facilitates the cleavage of hAPP into sAβPP rather than Aβ42) in cultured human astrocytes, thereby restoring sAβPPα extracellular shedding and fully suppressing Aβ42 oligomer oversecretion, but the expression of hAPP remains unaffected [17]. Moreover, Aβ25 - 35 exposure significantly increases both the intracellular exosomal levels of phosphorylated tau protein and the activity of the upstream tau kinase GSK3β in cultured cortical adult human astrocytes, while NPS 2143 hinders these effects [100]. The roles of the CaSR in both Aβ production and tau phosphorylation indicate that antagonism of astrocytic CaSRs might be a promising therapeutic approach for the treatment of AD.
Astrocytic Aβ receptors closely related to calcium signaling
Although astrocytes cannot generate propagating action potentials like neurons, astrocytes undergo a set of reactive functional changes involving intracellular Ca2 + transients and intercellular Ca2 + waves via Aβ/α7-nAChR signaling and Aβ/CaSR signaling [98]. Ca2 + signaling is responsible for the local release of gliotransmitters such as glutamate, D-serine, GABA, and taurine, which modulate astrocyte-astrocyte and astrocyte-neuron signaling. Moreover, in vitro studies have shown that the application of Aβ to mixed cultures of hippocampal neurons and astrocytes causes abnormal intracellular Ca2 + transients and mitochondrial depolarization in astrocytes long before an impairment is visible in neurons [28]. Specifically, the interaction of Aβ oligomers and astrocytic α7-nAChRs results in Ca2 + signaling changes and the release of gliotransmitters, including glutamate. The release of astrocytic glutamate induced by the interaction of Aβ and the α7-nAChR activates the extrasynaptic NMDA receptors on neurons [101]. The NMDAR inhibitor nitromemantine, an improved memantine derivative, has been shown to block the synapse-destroying effects of pathological Aβ/α7-nAChR signaling and thus preserve cognition.
Aβ/CaSR signaling results in an overproduction/oversecretion of newly synthesized Aβ42 oligomers, NO, and VEGF-A, particularly by normal adult human astrocytes [3, 102]. The latter effect could significantly increase the pool of Aβ- and NO-producing nerve cells, thus favoring the progressive spread of a self-sustaining and self-reinforcing ‘infectious’ mechanism of neural and vascular (i.e., the blood-brain barrier) cell damage [103]. Nevertheless, the broad extracellular accrual and spreading of neurotoxic Aβ42 oligomers could be blocked well upstream by administering an allosteric CaSR antagonist (calcilytic) [3, 102]. The interaction of Aβ with other receptors and/or nonreceptor-mediated Aβ mechanisms presumably increases the secretion of other toxic factors, such as proinflammatory cytokines, chemokines, and reactive oxygen species.
CALCIUM-BINDING PROTEIN β
S100β belongs to the S100 family of proteins, a family of calcium binding proteins with low-molecular weights that are often involved in Ca2 +-regulated processes, such as intracellular calcium homeostasis, protein phosphorylation, enzyme activities, cell proliferation and differentiation, the dynamics of cytoskeletal components, and the protection of cells from damage due to oxidative stress [104, 105]. S100β is abundantly expressed by astrocytes in the central and peripheral nervous system as well as by some populations of neurons, but to a substantially lesser extent [106]. S100β exerts trophic actions and stimulates neurite outgrowth as a mitogen or a neurotrophic factor when released at nanomolar (putatively physiological) concentrations, while at high micromolar concentrations, S100β is thought to behave as a danger-associated molecular pattern molecule and participate in neuroinflammatory processes that lead to cell injury and apoptosis [107].
An accelerated age-associated increase in the level of S100β has been found in the cerebral cortex of prematurely aged mice, along with an increase in activated astrocytes [108]. S100β overexpression precedes by decades the neurological changes in Down’s syndrome (a natural model for the study of Alzheimer’s disease due to the universal distribution of neuritic plaques) [109], precedes by months the appearance of Aβ deposits in the APPV717F transgenic mouse (an animal model of familial AD) [110] and precedes by months the activated astrocytes in SMAP6 mice (an animal model of aging).
Previous research has suggested that the overexpression of S100β plays a pathogenetic role in AD, particularly in the formation of Aβ plaques, in diffuse or neuritic forms [111, 112]. Increased S100β expression has been found in the reactive astrocytes of the AD brain [112–114]. Animal studies have also illustrated that the overexpression of human S100β aggravates cerebral amyloidosis and gliosis in the Tg2576 AD mouse model, which is rescued by the pharmacological blockade of S100β biosynthesis with arundic acid [115]. Moreover, mice that overexpressed S100β showed learning and memory impairment in all behavioral tests compared to nontransgenic controls [116]. S100β induces the expression of AβPP at both the gene and protein levels in a time- and dose-dependent manner in primary cultured rat cortical neurons, which serves as another way to increase Aβ accumulation [117].
The proposed pathogenic mechanism of S100β involves IL-1 overexpression in microglia. The diffuse immunogenic amyloid plaques at the early stage of AD could recruit the microglia via elevated levels of IL-1. IL-1 would then stimulate the activation of astrocytes and the astrocytic overexpression of S100β, which aggravates neuronal injury in the later stage by promoting dystrophic neurite formation in diffuse amyloid deposits [111]. It has also been proposed that the upregulation of S100β is a key component in the regenerative feedback loop of Aβ generation through an interaction with IL-1β [118]. S100β also causes the overexpression of inducible nitric oxide synthase (iNOS) and the subsequent release of nitric oxide [119] and activates nuclear factor-κB (NF-κB), a key transcription factor mediator of inflammatory responses [120]. The dysregulation of the astrocytic S100β protein along with its receptor, receptor for advanced glycation end products (RAGE), and the implications of this dysregulation regarding pathological events have also been studied in AD, especially the Aβ-induced signal transduction in glial cells [121]. The critical role of RAGE in Aβ clearance has also been emphasized [122]. Additionally, Esposito and his colleagues proposed a link between astrocytic-derived S100β and tau hyperphosphorylation in human neural stem cells, suggesting that in addition to Aβ plaques, S100β can contribute to NFT formation in AD [123]. Specifically, S100β, through the interaction with RAGE, causes Wnt pathway disruption and tau protein hyperphosphorylation. Overall, this evidence suggests a role of astrocytic S100β in the pathogenesis of AD, including amyloidosis, tau hyperphosphorylation and neuroinflammation, thereby providing a promising potential molecular target. Blocking S100β biosynthesis in reactive astrocytes might be a promising therapeutic strategy to delay AD progression.
CANNABINOID RECEPTOR 2 (CB2)
The endocannabinoid signaling system consists of the endocannabinoids N-anandamide and 2-arachidonoyl-glycerol (2-AG), the cannabinoid receptors CB1 and CB2 and the enzymes that synthesize and degrade endocannabinoids. Various authors have proposed cannabinoids as potential treatments in AD; some authors have identified the neuroprotective role of cannabinoids through the upregulation of Notch-1 signaling or the suppression of ERK1/2, NF-κB phosphorylation, and cyclooxygenase (COX-2) expression in a CB1-dependent way in neurons, whereas others authors have proposed that cannabinoids exert anti-inflammatory effects by altering functions of microglia and astrocytes [124–126].
Of the two receptors, CB2 has been shown to be abundantly and selectively expressed in neuritic-plaque-associated astrocytes in hippocampus and entorhinal cortex slices dissected from the brains of AD patients [127]. The activation of CB2 by the agonist WIN 55,212-2 has been shown to prevent the decreased viability of primary cultured astrocytes and exert anti-oxidative and anti-inflammatory effects despite the existence of Aβ42 [128]. It has been demonstrated that WIN 55,212-2 increases the expression of the anti-oxidant Cu/Zn superoxide dismutase and inhibits increases in TNF-α and IL-1β levels as well as the upregulation of the p-65, COX-2, and iNOS proteins in cultured astrocytes [128]. Similarly, a decline in GFAP and S100β levels, NF-κB pathway activation, and iNOS and IL-1β levels has also been observed in Aβ-stimulated cultured newborn rat astrocytes treated with cannabidiol, a nonpsychotomimetic cannabinoid derived from cannabis [129, 130]. Furthermore, the administration of WIN 55,212-2, 2-AG, or methanandamide (a synthesized chiral analog of anandamide) has been shown to prevent the hemichannel activity and inflammatory profile induced by Aβ in astrocytes [131]. The activation of CB2 by its agonist MDA7 has been found to attenuate the activation of astrocytes, normalize CB2 expression, promote Aβ clearance, reverse the synaptic plasticity impairment and restore learning and memory impairments in mice bilaterally injected with Aβ40 fibrils [132]. These experiments support the possibility of activating the astrocytic cannabinoid receptors as an efficient way to improve the cognitive ability of and reduce inflammation in AD patients.
CONCLUSIONS
In this review, we summarized the potential therapeutic targets, especially the receptors and transporters of astrocytes and their regulatory mechanisms, during the onset and development of AD (as shown in Fig. 1). The contribution of astrocytes in the neuropathology of AD occurs, at least in part, via the excitotoxicity that results from the accumulation of extracellular glutamate. The processes of astrocytes, through the interaction with synapses, play a crucial role in glutamatergic synaptic transmission. Both isotropic and metabotropic glutamate receptors located at the plasma membrane of astrocytes, for instance, NMDARs and mGluR5, can drive glutamate-mediated signal transduction. More importantly, astrocytic glutamate transporters, especially GLT-1 and GLAST, are responsible for the uptake of excess glutamate in the synaptic cleft, thereby terminating the neurotoxic effect [33]. mGluR5 directly contributes to intracellular calcium homeostasis and ATP release [18], while the A2AR regulates synaptic activity through NMDARs and is also involved in glutamate uptake [34, 68].
Overview of the receptors and transporters of astrocytes potentially involved in the pathogenesis of Alzheimer’s disease. Schematic representation of the molecular mechanisms of EAATs (GLT-1 and GLAST), the A2AR, the α7-nAChR, cannabinoid receptors (especially CB2), the CaSR, mGluR5, and S100β in AD pathogenesis. Aβ, amyloid-beta; APP, amyloid-β protein precursor; CaSR, calcium-sensing receptor; NOS2, nitric oxide synthase-2; VEGF, vascular endothelial growth factor; A2AR, adenosine receptor 2A; Arc, activity-regulated cytoskeleton-associated protein; GLT-1, glutamate transporter-1; GLAST, glutamate aspartate transporter; Glu, glutamate; Gln, glutamine; GS, glutamine synthetase; NMDAR, N-methyl-D-aspartate receptor; Akt, protein kinase B; IRs-1, insulin receptor substrate 1; IRβ, insulin receptor β; IDE, insulin degrading enzyme; mGluR5, metabotropic glutamate receptor 5; ATP, adenosine triphosphate; S100β, calcium-binding protein β; RAGE, receptor for advanced glycation end products; p, phosphorylation; α7-nAchR, α7-nicotinic acetylcholine receptor; CB2, cannabinoid receptor 2; TNFα, tumor necrosis factor α; IL-1β, Interleukin-1β; GFAP, glial fibrillary acidic protein; NF-κB, nuclear factor-κB; Cox-2, cyclooxygenase-2; SOD, superoxide dismutase; iNOS, inducible nitric oxide synthase.
The deposition of Aβ plaques in the AD brain depends on the splicing rate of AβPP, the synthesis and secretion of Aβ42 in neurons and astrocytes, and the extracellular aggregation of Aβ42 as well as the phagocytosis and clearance of extracellular Aβ. Astrocytes, the main class of neuroglia, are involved in the production, secretion, and phagocytosis of Aβ. For instance, S100β not only increases the expression of AβPP but also influences the clearance of Aβ by interacting with the S100β receptor RAGE [117, 121]. Some of the potential molecules in astrocytes, such as the CaSR, mGluR5, and the α7-nAChR, are Aβ receptors. It has been suggested that the antagonism of mGluR5, the α7-nAchR, and the CaSR in astrocytes could ameliorate the pathological symptoms of AD by increasing Aβ clearance, lowering the intracellular accumulation and oversecretion of Aβ, inhibiting the elevation of intracellular calcium, decreasing the production of nitric oxide and glutamate release and mitigating tau phosphorylation [18, 98]. For instance, the interaction of the CaSR and AβO produces more β-secretase-cleaved Aβ42 oligomers, while inhibiting the CaSR via its antagonist NPS 89626 rescues the extracellular shedding of sAβPPα [93, 96]. The mGluR5 clusters in the reactive astrocytes that surround Aβ plaques cause intracellular calcium elevation and synaptotoxicity [47]. Positive allosteric modulators of α7-nAChRs have been suggested as a promising approach to preventing the neurotoxicity induced by Aβ [133].
Reactive astrocytes also play an inflammatory role in AD etiology. The activation of cannabinoid receptors, especially CB2, or blocking S100β prevents the neuroinflammatory and oxidative effects and improves cognitive and memory performance despite the existence of Aβ42 [128, 132]. The function of these molecules in glutamatergic synaptic transmission, Aβ production and clearance, and the inflammatory and oxidative stress responses implies an indispensable role of astrocytes in the occurrence of AD, and modulating these molecules might provide alternative ways to modify the cognitive deficits and neuropathology of AD.
It is worth mentioning that some other molecules synthesized by astrocytes, such as BACE1 and cytokines, also play important roles in the etiology of AD; regarding Aβ production and neuroinflammation, the detailed involvement of these molecules has been previously reviewed [14, 134]. Furthermore, the distinction between human and rodent astrocytes can be taken into consideration in exploring the mechanism of AD. Despite the fact that both human and rodent astrocytes are involved in the clearance, accumulation, proteolysis, and release of Aβ, neurotoxicity induced by glutamate, neuroinflammation, and oxidative stress, the number, structure, morphology, and diversity of human astrocytes differ greatly from those of rodent astrocytes [135]. For example, the whole human adult brain contains a mean glia/neuron ratio of one-to-one [136], while this ratio in the rat cerebral cortex is 0.4 [137]. The length of the primary process of human astrocytes is much longer than that of rodent astrocytes. Accordingly, the data gathered in rodent astrocytes should be carefully and properly transposed to in vivo studies.
Footnotes
ACKNOWLEDGMENTS
This work was supported by National Natural Science Foundation of China (grant numbers 81801088, 81471415, 81873740), the Research Foundation of Xi’an Medical University (grant numbers 2017DOC01 and 2017GJFY01), Key Project of Shaanxi Province-the field of Social Development (grant number 2018ZDXM-SF-040), Scientific Research Fund of Shaanxi Provincial Education Department (grant number 18JS104), Projects of International Cooperation and Exchanges Natural Science Foundation of Shaanxi Province of China (grant number 2018KW-038). We acknowledge the helpful support of the Xi’an Medical University’s key disciplines of molecular immunology.
