MAPK, β-amyloid and synaptic dysfunction: the role of RAGE

Abstract

Genetic and biological studies provide strong support for the hypothesis that accumulation of β amyloid peptide (Aβ) contributes to the etiology of Alzheimer’s disease (AD). Growing evidence indicates that oligomeric soluble Aβ plays an important role in the development of synaptic dysfunction and the impairment of cognitive function in AD. The receptor for advanced glycation end products (RAGE), a multiligand receptor in the immunoglobulin superfamily, acts as a cell surface binding site for Aβ and mediates alternations in the phosphorylation state of mitogen-activated protein kinase (MAPKs). Recent results have shown that MAPKs are involved in neurodegenerative processes. In particular, changes in the phosphorylation state of various MAPKs by Aβ lead to synaptic dysfunction and cognitive decline, as well as development of inflammatory responses in AD. The present review summarizes the evidence justifying a novel therapeutic approach focused on inhibition of RAGE signaling in order to arrest or halt the development of neuronal dysfunction in AD.

Alzheimer’s disease & β-amyloid synaptic dysfunction

Alzheimer’s disease (AD) is typically characterized by the presence of senile plaques composed of the β-amyloid (Aβ) peptide, and intracellular neurofibrillary tangles. Excessive accumulation of Aβ is a key feature of AD pathology [1]. The Aβ peptide is derived from cleavage of the amyloid-precursor protein (APP) by α- and γ-secretases [2]. Aβ peptides of various lengths (Aβ₄₂, Aβ₄₀ and Aβ₃₆) are produced after γ-secretase cleavage, with the 42-amino acid peptide (Aβ₄₂) being among the most neurotoxic amyloidogenic fragments [3,4]. Aβ is found in the form of soluble peptides composed of monomers, oligomers and as large aggregated fibrils that lead to precipitates in the form of senile plaques [5]. Aβ is normally produced in the brain [6–8] and low levels are detectable in the cerebrospinal fluid of healthy subjects [9].

Although senile plaques and tangle formation are generally thought to play a key role in the development of cognitive deficits, recent studies in AD animal models have highlighted a dichotomy between behavioral deficits and neuropathological findings. In fact, memory impairment and synaptic loss occur prior to extensive accumulation of Aβ or the occurrence of senile plaques in both mouse models and AD patients [10–17].

Memory deficits for recently experienced events and declarative episodic memory (the means for conscious retrieval of previously experienced information), are often the initial cognitive manifestations of AD [18–21]. It has been hypothesized that the cognitive deficits during the early phase of AD are a consequence of synaptic dysfunction induced by the deposition of soluble Aβ [22]. Supporting evidence for this hypothesis in recent reports shows that low concentrations of oligomeric soluble Aβ that allow neuronal cells to remain viable, were able to induce an impairment of synaptic function in the hippocampus and entorhinal cortex [23,24]. These two areas of the brain play a crucial role in learning and memory [25–27], and are affected during early stages of AD [28].

Long-term potentiation (LTP), a type of synaptic plasticity that sustains learning and memory within critical areas such as the hippocampus and cortical parahippocampal regions [29,30], is now known to be easily affected by Aβ. Specifically, both the native human oligomers and synthetic Aβ₄₂ have been shown to acutely inhibit LTP at concentrations as low as 200 nM [31–36].

Interestingly, extracts from the cerebral cortex of AD patients containing Aβ oligomers were able to inhibit LTP in the hippocampus leading to reduction of spine density and disruption of learning behavior [37]. A few reports demonstrated that administration of Aβ at concentrations higher than those inducing pure inhibitory effects on LTP [24], depressed synaptic transmission and impaired a second form of long-term synaptic plasticity, long-term depression (LTD), by increasing the synaptic removal of glutamate receptors [38–41]. When combined, these results suggest that deposition of oligomeric Aβ plays an important role in the development of synaptic dysfunction and disruption of learning/memory in AD and emphasize the need to identify receptors and signal transduction pathways responsible for neuronal impairment at different phases of AD development.

Receptor for advanced glycation end products & β-amyloid

β-amyloid is a pleiotropic peptide, capable of binding to receptors at several different membrane locations, including the receptor for advanced glycation end products (RAGE) [42]. RAGE is a multiligand receptor from the immunoglobulin superfamily and has been extensively studied for it roles in migration and differentiation of neuronal cells during development, perturbation of neuronal cells by Aβ, and for its role in the inflammatory response [43–46]. There are many forms of RAGE, ranging from the full-length receptor, which includes an intracytoplasmatic domain, to the extracellular binding domains found in the truncated forms. The full-length form of RAGE consists of the extracellular domain and a single transmembrane region followed by a short intracellular domain. The intracellular domain of RAGE is necessary for RAGE-dependent signaling, including the activation of the mitogen-activated protein kinase (MAPK) signaling pathway [23,47–49]. Activation of these pathways by full-length RAGE involves members of the Rho-family of GTPases in addition to RAS [44,45,50]. A recent report suggested that a d-domain on the cytoplasmic tail of RAGE could provide a means for direct docking of MAPK to the cytoplasmic domain of full-length RAGE [51]. The current view is that the cytosolic domain of RAGE serves as a scaffolding for the initiation of signal transduction (Figure 1).

Figure 1. Receptor for advanced glycation end products (RAGE) is a member of the immunoglobulin superfamily of cell surface proteins interacting with several ligands, including β-amyloid petide (Aβ 1–42), advanced glycation end products (AGEs), amphoterin and members of the S100 family.

RAGE binding is able to activate multiple signaling pathways, such as activation of the transcription factor NF-κB; through a redox-dependent activation (O₂) of Ras-ERK1/2 pathway, Cdc42/Rac pathway, p38 MAP (p38) and SAPK/JNK (JNK) kinase pathways.

Experimental evidence indicates that RAGE is most probably the receptor that mediates and focuses Aβ effects on neuronal and non-neuronal target cells (neurons, glia and microglia) [23,43,45,49,52,53]. During the progression of AD, RAGE expression changes in both neuronal and non-neuronal brain cells [42]. Furthermore, introduction of a wild-type RAGE transgene targeted to neurons in the AD-type transgenic mouse model expressing mutant human APP (mAPP) [54] accelerated Aβ-mediated neuronal perturbation [23]; these results indicate that RAGE is an important cofactor for Aβ-dependent neuronal dysfunction. As RAGE is expressed in both neuronal and non-neuronal cells, it is important to identify the cell-specific activity of RAGE, under Aβ load. The involvement of RAGE in neurons during Aβ-induced synaptic dysfunction, in the entorhinal cortex and hippocampus is of interest [23,24]. RAGE activation in microglia also induces the release of proinflammatory cytokines, such as IL-1β and TNF-α [49,52,55].

Truncated forms of RAGE, termed soluble RAGE (sRAGE), contain the extracellular binding domain of RAGE and represent a nonfunctional form of RAGE that actively competes for RAGE ligands, thereby reducing the likelihood of deleterious activation of full-length RAGE. sRAGE also has potent anti-inflammatory properties, through its actions as a decoy for RAGE ligands such as Aβ. Notably, sRAGE not only binds to Aβ in its soluble state but also prevents Aβ aggregation, thus reducing the formation of senile plaques [56,57]. These studies suggest that RAGE expressed in neuronal and non-neuronal cells plays a role in the neurodegenerative processes of AD.

Mitogen-activated protein kinases

The MAPK cascade is prototypical for a family of signaling cascades that share a motif of three linked kinases regulating each other by sequential phosphorylation. For the purposes of this review, we will use the term MAPKs to refer to the entire superfamily of signaling cascades (comprising Erk, Janus kinase [JNK] and p38 MAPK). MAPKs are expressed in neuronal cells in the mature CNS, during dynamic states in response to various external stimuli such as growth factors, glutamate and hormones, stressors (e.g., hypoxia and oxidative stress) and pathogens such as Aβ.

Erk1/2 (p44/p42)

The p42 and p44 MAPKs (Erk1 and Erk2) are molecules that play a critical role in the intracellular signaling cascade and whose phosphorylation is induced as part of the cellular response to various different stimuli; serial phosphorylation leads to the GTP bound form of the small G protein Ras, the protein kinase Raf and phosphorylation of MEK. Many different substrates can be phosphorylated by Erk1/2 in the cytoplasm as well as at the level of the nucleus [58]. Once activated in the cytoplasm, Erk1 and Erk2 translocate into the nucleus and interact with nuclear substrates to induce specific patterns of gene expression [59,60]. Erk1 and Erk2 are expressed throughout the developing brain [61,62]; animal models with deletion of Erk1 gene are characterized by subtle phenotypic changes, whereas Erk2 gene deletion results in early embryonic lethality [63]. Erk1 and Erk 2 are also expressed in nondividing, mature neuronal cells. In mature neuronal cells, Erk1/Erk2 react to external stimuli such as growth factors and the binding of glutamate to their receptors, contribute to information processing in dendrites, modulate structural changes in dendritic spines and interact with scaffolding and structural proteins at the synapse [64].

Recent studies on human mental retardation suggested that Erk1/Erk2 plays a critical role in learning [65]. In addition, results have shown that the Erks are involved in several forms of synaptic plasticity in the hippocampus and other brain areas critically involved in learning and memory processes [58]. These include the NMDA receptor-independent forms of LTP in hippocampal area CA1 [66], several forms of LTP in the dentate gyrus [67] and LTP in the amygdala, all of which are associated with fear-dependent learning [68,69]. Importantly, Erk1/Erk2 is required for LTP induction and maintenance for ‘rewiring’ the visual cortical circuitry following monocular deprivation [70]. Another role for the two Erk isoforms in synaptic plasticity has been proposed, specifically that Erk2 is active in the CA1 area of the hippocampus during NMDA-dependent LTP induction [71]; additional results demonstrate that LTP maintenance depends on late activation of Erk2 [71,72]. Mazzucchelli and colleagues reported increased LTP associated with improved long-term memory in the striatum of Erk1 knockout mice, thus proposing a role for Erk1 in synaptic plasticity [62]. However, a compensatory increase in Erk2 activity was described in several brain areas of Erk1 knockout mice [61,62,73], and an analogous compensatory effect was reported in Erk2 conditional knockout mice presenting increased Erk1 signaling [74]. Thus, it is possible that both Erk isoforms are involved in long-term synaptic plasticity. Beyond the consolidated role in synaptic plasticity, recent evidence has also suggested participation of the ERK signaling cascade in modulation of synaptic transmission; in fact, Ras, acting through the Erk1/Erk2 MAPK signaling pathway, enhances glutamate-mediated synaptic transmission through AMPA receptors [75].

During the last 10 years, the signal transduction pathways involved in Aβ-dependent neurotoxicity have become a major focus in AD research. There is a general consensus that Aβ induces Erk1/Erk2 abnormalities as shown in the brain of AD patients [76] and animal models [23,77]. For example, fibrillar Aβ₄₂ is capable of causing Erk1/Erk2 activation. Moreover, sustained activation of Erk1/Erk2 by Aβ is implicated in the mechanism leading to abnormal phosphorylation of Tau, the progression of neuronal degeneration [78,79] and cell death [80,81].

Synaptic dysfunction from RAGE signaling has been demonstrated in a recent study showing acceleration of Aβ-mediated hippocampal perturbation along with an increase in the regional phosphorylation of Erk1/Erk2 in a double transgenic mouse for mAPP and wild-type RAGE [23]. This genetic manipulation led to synaptic impairment and cognitive abnormalities, suggesting that persistent activation of Erk1/Erk2 through RAGE/Aβ signaling contributes to an impairment of neuronal functionality.

In addition to Aβ, persistent activation of Erk1/Erk2 has been shown to play a crucial role in neuronal impairment following oxidative stress induced by glutamate neurotoxicity [82]. On the other hand, even brief treatment with cell-derived soluble human Aβ influences Erk1/Erk2 phosphorylation and the phosphorylation of Erk cytoplasmatic targets such as CaMKII and Akt/PKB [83], thereby decreasing Erk1/Erk2 phosphorylation [84]. It has been demonstrated that a cytosolic domain of the RAGE receptor is responsible for decreased Erk1/Erk2 phosphorylation. Thus, both hyper- or hypo-phosphorylation of Erks by Aβ may contribute to synaptic dysfunction in AD, possibly at different stages of the neurodegenerative process. Indeed, studies conducted in human brain and mouse models of AD suggest stage-dependent Erk activity levels with increased and reduced amounts of Erk phosphorylation, respectively [74,77].

P38 MAPK

There are four isoforms of p38 MAPKs in mammals; α, β, γ and δ. These can be separated into two groups, p38α and p38β, which are inhibited by the pharmacological agents SB203580 and SB202190, at micromolar and nanomolar concentrations [85–87]. Among p38 isoforms, the α variant is expressed in the widest range of cells. Together with the JNK family, the p38 MAPKs are also known as stress-activated protein kinases. The classical pathway of activation for the p38 MAPKs occurs via dual phosphorylation of their Thr–Gly–Tyr motif by MKK3 and MKK6 [88,89], found in the activation loop. Many p38 MAPK targets have been described, both in the cytoplasm and in the nucleus;, including protein kinases (MAPK-activated protein kinases, MAPK-interacting kinase, mitogen- and stress-activated kinase), which in turn phosphorylate transcription factors (p53, ATF-2, NFAT and STAT1), cytoskeletal proteins (e.g., the microtubule-associated protein Tau), and other proteins with enzymatic activity, such as the glycogen synthase and cytosolic phospholipase A2 [90].

As reported for Erk, p38 MAPK is also widely expressed in the adult brain. Several studies examined the role of p38 MAPKs in regulating synaptic plasticity. For example, the p38 MAPK signaling cascade participates in the induction of hippocampal LTD, a form of long-term synaptic plasticity implicated in learning and memory. More specifically, p38 MAPK is activated in response to synaptic stimulation resulting in a form of LTD that depends on the metabotropic glutamate receptor. Inhibition of the p38 MAPK pathway by pharmacological compounds specifically blocks this form of LTD without affecting LTP [91]. Other authors have shown that NMDA-dependent hippocampal LTD also requires p38 MAPK activation [75], clearly demonstrating that p38 MAPK contributes to synaptic plasticity and is involved in learning and memory processes.

Recent results obtained in murine-type models of AD using exogenously applied Aβ contributed to our understanding of the role of p38 MAPK in neuronal dysfunction. Phosphorylation of p38 MAPK selectively increases in double transgenic mice overexpressing both mAPP and full-length RAGE; the level of phospho-p38 MAPK was reduced in double transgenic mice overexpressing mAPP and a dominant-negative form of RAGE targeted to neurons [23]. Work by Origlia and colleagues has shown that the suppression of LTP by Aβ could be prevented in entorhinal cortex treated with the p38 MAPK inhibitor SB203580 [24]. Furthermore, a recent report shows that Aβ at low micromolar concentrations was able to drive synaptic removal of AMPA receptors leading to a loss of dendritic spines and a depression of synaptic transmission [40]. Using pharmacological inhibition, these authors suggest a role for p38 MAPK in the trafficking of glutamate receptors, mainly at the level of the synaptic cleft. Thus, p38 MAPK activation by RAGE/Aβ signaling may trigger a downstream signaling cascade that contributes to LTP inhibition, to depressed synaptic transmission and to loss of synaptic spines that characterizes the early phases of AD.

Activation of the p38 MAPK cascade has been implicated in neural responses to various stressors and pathogens, including high levels of Aβ [92,93]; as aforementioned, p38 MAPK and JNK are also known as stress-activated protein kinases. Elevated levels of Aβ in the brain of an AD mouse model result in microglial activation through p38 MAPK phosphorylation, leading to the release of proinflammatory cytokines [94–96].

There also exists somewhat controversial results regarding a clear and direct activation of p38 MAPK by Aβ in neurons. One report suggests that activation of p38 MAPK facilitates Aβ cytotoxicity in primary cultures of cortical neurons [97], while other authors found no effect [98]. p38 MAPK is required for LTP inhibition owing to the presence of proinflammatory agents such as IL-1β and lipopolysaccharide (LPS) [99,100]. Anwyl and colleagues reported that blocking p38 MAPK phosphorylation is sufficient to prevent Aβ-dependent inhibition of LTP in hippocampal slices [33]. Furthermore, a novel p38 α MAPK inhibitor was capable of suppressing brain proinflammatory cytokine upregulation and attenuating synaptic dysfunction and behavioral deficits in an AD mouse model [101–103]. Thus, Aβ is able to induce strong activation of p38 MAPK in neurons following the release of proinflammatory cytokines by microglia. This may represent an important trigger for the development of neuronal impairment during the progression of AD.

Thus, it is possible that binding of Aβ to RAGE in neurons and microglia causes oxidative stress and inflammation leading to cellular perturbation [52]. p38 MAPK activation is in fact necessary for RAGE-dependent NF-κB activation, induction of target gene expression and secretion of proinflammatory cytokines from monocytes [47]. It is also known that NF-κB, the oxidant-sensitive transcription factor, contributes to the proinflammatory response by regulating gene expression. For example, accumulation of advanced glycation end products (the category of RAGE ligands linked to aging, vascular pathology, inflammation and hyperglycemia) generates oxidative stress following the activation of NF-κB [104]. Therefore, it follows that the RAGE signaling axis involved in p38 MAPK activation in neuronal and non-neuronal cells contributes to the development of inflammatory responses and neuronal perturbation, and is induced by an increased Aβ load during the course of AD.

JNK

The stress-activated, protein kinases C-JNK is involved in cellular regulatory functions such as gene expression, cell proliferation and programed cell death. For example, activation of the JNK pathway is critical for naturally occurring cell death (apoptosis or necrosis during development as well as for cell death associated with neurodegenerative diseases). JNK can be distinguished from other MAPKs by the sequence of the tripeptide dual phosphorylation motif that is required for activation (Thr183-Pro184-Tyr185). There are three known JNK genes (JNK1, JNK2 and JNK3). JNK1 and JNK2 are ubiquitously expressed while JNK3 is restricted to the brain, heart and testes [89,105–107]. Differential splicing results in multiple isoforms for all of the JNK genes [108,109]. Each JNK is expressed in both a short (46 kDa) and a long form (54 kDa) [105], and each isoform of JNK has different binding and phosphorylation specificities [108,109].

Targeted gene disruption for each of the JNKs helped to define the differential functions for JNK1, JNK2 and JNK3 in various cell types. JNK1 mutant mice show a progressive learning impairment and motor defect, along with loss of anterior and posterior commissure, and dendritic alteration in the hippocampus [110]. In JNK2-deficient mice, LTP was impaired both in the CA1 and CA3 regions [111]. The involvement of JNKs in synaptic plasticity is supported by strong expression of different JNK isoforms in adult hippocampi and cortical areas involved in the learning and memory processes [112,113]. JNK is able to phosphorylate several cytoskeletal proteins that undergo ultrastructural changes during LTP maintenance in the hippocampus [114]. Specifically, increased phosphorylation of JNK was associated with IL-1β-dependent inhibition of LTP in the hippocampus [115] and disruption of metabotropic glutamate receptor-dependent LTD in JNK1 knockout mice [116]. In addition, Xu and colleagues showed robust JNK activation during exploration of a manipulated new environment, indicating that the JNK cascade has a role in dynamic brain responses to environmental stimuli [117].

Initial research on the role of the JNK cascade in neurodegenerative diseases was focused on cell death. Several in vitro and in vivo studies reported alteration to the JNK signaling cascade, thus it was considered to be potentially involved in neuronal death associated with neurodegenerative diseases such as Parkinson’s disease and AD [118,119]. The JNK/c-Jun cascade is active in neurons of AD brains, suggesting its involvement in these abnormal processes, ranging from tau-phosphorylation to neuronal death. Immunocytochemical investigations of JNK and p38 MAPK in AD brains have demonstrated the presence of active JNK at a stage characterized by the presence of neurofibrillary degeneration [120], primarily concentrated in the entorhinal cortex [121]. Recent studies have been conducted on the effects of Aβ on JNK phosphorylation, particularly as it leads to synaptic dysfunction. JNK pharmacological inhibitors prevent LTP impairment by either synthetic or native human Aβ₄₂ in the hippocampus [33]. By contrast, a recent paper by Origlia and colleagues demonstrated that pharmacological inhibition of p38 MAPK, but not JNK, was able to rescue RAGE-dependent LTP inhibition in entorhinal cortex slices treated with a low concentration of oligomeric Aβ₄₂ [24]. These differences are not entirely irreconcilable, as the discrepancy may be a result of different concentrations of Aβ₄₂ used, as well as differences in the sensitivity of JNK to different pharmacological compounds and the different brain areas investigated in the two studies. Compelling results indicate that high levels of Aβ are able to induce phosphorylation of JNK [97,122–127]. However, these studies were performed using chronic treatments with relatively high Aβ concentrations in the μm–mM range. Furthermore, acute treatment with low concentrations of oligomeric Aβ₄₂, in turn, does not produce changes in the phosphorylation of JNK in hippocampal slices [128]. Nonetheless, engagement of RAGE in neuronal and non-neuronal cells by different ligands has been shown to result in JNK activation [129–131]. Supporting this, Aβ inhibition of LTP in perforant path granule cell synapses was accompanied by the phosphorylation of JNK and increased expression of RAGE [132]. These findings suggest that a high Aβ load may induce RAGE engagement, activating the JNK cascade thereby contributing to synaptic dysfunction in AD.

Expert commentary

The aim of pharmacological therapy in AD patients is to arrest or at least slow down the development of dementia without interfering with basal functional mechanisms in neural cells. To this end, drugs treating multiple pathological mechanisms and clinical symptoms in AD, such as cholinesterase inhibitors and/or cholinomimetics agonists (donepezil, rivastigmine and galantamine) [133], and drugs that selectively inhibit glutamate exotoxicity, such as memantine [134,135], an NMDA receptor antagonist, have been proposed as therapeutically appropriate. However, a lack of knowledge of the mechanisms of these drugs, their adverse effects (especially among cholinesterase inhibitors), and emerging evidence that their efficacy is short-lived and limited by the progression of AD cast some doubt on their efficacy and value. Increasing evidence suggests that regenerative agents may be the therapeutic potential of neurogenesis to promote the regeneration of lost neurons and neural circuitry to restore cognitive function in AD [136–138].

Currently, efforts are ongoing to develop effective strategies to decrease the Aβ load [139] — research into drugs that regulate the production and/or the accumulation of Aβ are ongoing. Genetic and biochemical studies imply that Aβ is a key molecule in the pathogenesis of AD, suggesting that interference with Aβ production could be a promising therapeutic strategy. Selective inhibition of APP cleavage is therefore critical and several compounds acting as inhibitors of γ-secretase activity have been proposed [140]. However, the findings that suggest that secretases participate in several different signaling events and, more importantly, that in the normal brain, low levels of Aβ may be an important modulator of synaptic plasticity and memory acting through the nicotinic receptors [141], bring forth severe potential limitations in this approach, and dampen earlier enthusiasm regarding the expected efficacy of such drugs. Additionally, passive immunotherapeutic strategies targeted to Aβ in the brain [142] may be negatively influenced by our lack of understanding of the normal physiological role of low Aβ levels in the brain [6,141]. On the other hand, the passive immunotherapeutic strategies for AD treatment might have a benefit by normalizing the excess levels of Aβ in the brain but yet leaving enough Aβ for the normal physiological function. Given that APP knockout mice are surviving well, APP-like proteins might substitute the effect of APP or Aβ on the neuronal function. Aβ monomers were shown to support the survival of developing neurons under conditions of trophic deprivation and to protect mature neurons against excitotoxic death [143]. In addition, given the antioxidant activity of Aβ, excessive removal of Aβ might enhance, rather than reduce, neuronal oxidative stress [144]. Therefore, particular caution is required when designing effective and safe approaches to AD therapy.

It becomes obvious that additional and alternative strategies to interfere with Aβ signaling are required. In this review, evidence has been presented suggesting that increasing Aβ load results in RAGE-related signaling with consequent changes in the phosphorylation of MAPKs. A clarity of understanding is beginning to emerge as our knowledge of the relationships between Aβ, RAGE signaling and MAPKs are elucidated: it appears that modulation of different MAPKs, such as Erk, p38 MAPK and JNK isoforms may contribute to the development of synaptic dysfunction and impairment of cognitive functions in AD. It seems fairly certain that p38 MAPK and JNK are involved in the development of synaptic dysfunction characterizing the progress of neuronal perturbation in AD [24,33,145]. Moreover, the activation of these two MAPKs by Aβ facilitates the development of cytotoxicity in primary cortical neurons [97,146].

Consistent with these findings, it has been reported that RAGE is not only expressed in neurons but also in glia and microglia where it may contribute to the development of inflammatory responses and the progression of neurodegeneration prior to accumulation of soluble Aβ or the formation of fibrillar aggregates or senile plaques; during all phases of AD, varying levels of soluble Aβ and differing states of aggregation may contribute to changes in the temporal and spatial pattern of MAPK activation in the brain, particularly in areas such as the hippocampus and parahippocampus that are critically involved in synaptic plasticity and learning behavior. Therefore, it follows that RAGE interaction with increasing levels of Aβ induces progressive involvement of different MAPKs in neuronal and non-neuronal cells along with an induction of an inflammatory response and increasing synaptic dysfunction. Thus, combining pharmacological inhibition of RAGE with prevention of cellular perturbation by acting on different MAPKs is an attractive strategy for the prevention and/or arrest of cognitive decline from AD.

Five-year view

Findings from the aforementioned studies help us to understand the mechanisms through which MAPKs mediate cascades, particularly RAGE-dependent signal transduction (activation of MAPKs), contributing to neuronal inflammation and neurodegenerative sequelae observed in AD. Activation of p38 MAPK is found in the brain during the early stages of AD [121,146–150], both clinically and in mouse models [147,151,152]. Inhibition of p38 MAPK activation blocks Aβ-mediated cytokine production and neuronal death [103,141]. Blockade of RAGE significantly attenuates Aβ-mediated, sustained neuronal and microglial stress and improves cognitive function in AD mice [23,52]. These studies provide a substantial argument for targeting RAGE, in addition to MAPKs, as a primary therapeutic strategy in AD. Fortunately, small molecule antagonists of the receptor have already been developed and have demonstrated a protective effect in animal models [153]. The RAGE inhibitor has an excellent safety profile, and has been well-tolerated for over 10 weeks in patients with AD in oral treatments according to results of a Phase II clinical study [154]. RAGE inhibitors, in addition to those for MAPKs, have the potential for significant therapeutic advances in the near future.

Key issues.

• Alzheimer’s disease (AD), a progressive neurodegenerative disease typically affecting older patients, is characterized by a cognitive decline leading to dementia. Hallmark features of AD include the presence of neurofibrillary tangles, primarily composed of the hyperphosphorylated protein tau, and by deposits of β-amyloid (Aβ) in the form of senile plaques.

• Aβ is an end product resulting from the sequential cleavage of amyloid precursor protein by β-and γ-secretase. Aβ peptides of various lengths (e.g., Aβ₄₂ Aβ₄₀ and Aβ₃₆) are produced after γ-secretase cleavage; the 42 amino-acid peptide Aβ₄₂ is considered one of the most neurotoxic of the amyloid fragments. Aβ₄₂ progressively accumulates in the brain of AD patients and murine models of AD and participates in plaque formation.

• Receptor for advanced glycation end products (RAGE) is a multiligand receptor in the immunoglobulin superfamily. RAGE is expressed in neuronal and non-neuronal cells and mediates the effects of Aβ.

• Mitogen-activated protein kinase (MAPK) refers to the entire superfamily of signaling cascades (comprising ERKs, the JNKs and the p38 MAPK). MAPKs are expressed in neuronal cells in the mature CNS, and contribute to regulation of the dynamic states of neural cells when responding to various external stimuli such as glutamate, growth factors and pathogens such as Aβ.

• Long-term potentiation and long-term depression represent two forms of synaptic plasticity, the maintenance of which is dependent on the synthesis of new proteins. Both long-term potentiation and long-term depression are related to learning and memory processes in critical brain areas such as the hippocampus and entorhinal cortex.

• Synaptic dysfunction indicates an alteration of a single or of multiple forms of synaptic responses, such as synaptic transmission that regulates intercellular communication, synaptic plasticity (short- and long-term synaptic plasticity), thereby regulating the dynamics of neuronal circuitry. Aβ, particularly Aβ₄₂, is capable of inducing synaptic dysfunction at different concentrations by binding RAGE and activating MAPK cascades.

• Drugs treating multiple pathological mechanisms and clinical symptoms of AD such as cholinesterase inhibitors or cholinomimetics agonists, and drugs that selectively inhibit glutamate exotoxicity have been proposed as therapeutically competent to arrest or slow cognitive decline in AD.

• We propose an alternative therapeutic strategy that combines pharmacological inhibition of Aβ/RAGE signaling with prevention of cellular perturbation (through affecting the action of certain MAPKs in order to prevent and/or arrest the development of synaptic dysfunction and cognitive decline in AD).