Nanotechnology Building Blocks for Intervention with Alzheimer’s Disease Pathology: Implications in Disease Modifying Strategies

Alzheimer’s disease (AD) is the most common neurodegenerative disorder manifested by progressive loss of memory followed by irreversible dementia [1-3]. AD neurodegeneration is characterized by the loss of neurons preceded by cell membrane and cytoskeleton damage due to a complex of molecular pathways that finally lead to formation of amyloid-β (Aβ) plaques and neurofibrillary tangles in the brain tissue [1-3].


Introduction
Alzheimer's disease (AD) is the most common neurodegenerative disorder manifested by progressive loss of memory followed by irreversible dementia [1][2][3]. AD neurodegeneration is characterized by the loss of neurons preceded by cell membrane and cytoskeleton damage due to a complex of molecular pathways that finally lead to formation of amyloid-β (Aβ) plaques and neurofibrillary tangles in the brain tissue [1][2][3].
Current treatments of AD can only partially alleviate symptoms, but there exists no cure to stop or slow neurodegeneration. The first major class of drugs used to treat AD was cholinesterase inhibitors which inhibit the proper functioning of the enzyme that breaks down acetylcholine. These drugs maintain a higher level of acetylcholine in the synaptic cleft thereby promote signaling and improve neuronal function; however, they do not prevent the underlying neurodegeneration [1][2][3]. On the other hand, all current diseasemodifying strategies targeting different molecular pathways involved in AD have failed in phase 3 clinical trials [4].
As reported in our recent detailed studies [2,3], major molecular pathways involved in AD can potentially be targeted through nanotechnology methods [5] and molecular building blocks (MBBs) [6] as it is shown in Figure 1. In summary, the current and envisioned applications of nanotechnology in neurology consist of neuroprotection, neuroregeneration and drug delivery beyond the blood brain barrier (BBB).
Among the above-mentioned applications, nanotechnology neuroprotective approaches can be of great implication in effective disease-modifying strategies for AD, especially when the targeted molecular pathways are involved in more initial phases of AD pathogenesis. This report is focused on such neuroprotective approaches through nanotechnology molecular building blocks (MBBs). To this end, we first elaborate on molecular pathways of Alzheimer's disease targeted by each group of nanotechnology MBBs.

Protection against Glutamate Excitotoxicity
A major pathological event in AD is synaptic failure [9], characterized by excessive synaptic release of glutamate and/or impairment of glutamate clearance mechanisms in the synaptic cleft. Glutamate together with D-serine are the primary excitatory neurotransmitters in the brain that activate voltage-dependent calcium channels through the N-methyl-D-aspartate (NMDA) receptors. Excessive synaptic glutamate, in turn, leads to a disruptive influx of calcium ions causing an overabundance of intracellular calcium, triggering several damaging pathways in neurons, and ultimately neuronal death ( Figure 2) [1,5,6]. Therefore, a disease modifying approach against AD synaptic pathology could be preventing the excessive glutamate signaling by blocking the NMDA-receptor channels [5].
From the structural point of view, NMDA receptors are coassemblies of different receptor subunits from different subunit families: NR1 and NR2 as shown in Figure 3. Among these subunits, NR1 has the essential role in NMDA receptor function [8]. Co-assembly of different members of the above mentioned subunit families leads to composition of different NMDA receptor subtypes with discrepant biophysical and pharmacological properties [9]. A variety of binding sites for ligands exist in the structure of NMDA. These ligand/binding site interactions are almost subunit selective and result in regulation of NMDA receptor function [9] as shown in Figure 3.
The topologies of NMDA receptor subunits are almost the same. Each subunit consists of an extracellular N-terminal domain, four membrane (M 1-4) domains and an intracellular carboxyl terminal [8,9]. The extracellular domain (N-terminal) forms the glycosylated large part of each subunit, responsible for binding with ligands. Among agonist ligands, glycine binds to NR1 and glutamate binds to NR2 binding sites. The first and second intra-membranous domains are interconnected through an intramembrane loop. The M2 domain forms the receptor channel pore. The intracellular domain (carboxyl terminal) contributes to regulation of the receptor through second messenger system [8,9].
As discussed before, if excessive NMDA signaling is not controlled, a high influx of calcium leads to the impairment of mitochondrial function, cellular energy failure and finally neurodegeneration [5]. Previously developed drugs which act as NMDA antagonists were discontinued due to blocking the normal NMDA signaling, a process that can also promote neurodegeneration [5]. Selective blockade of the excessive NMDA signaling is possible through optimization of NMDA receptor partial antagonism via nanotechnology molecular building blocks (MBBs) and their derivatives [9,10].
A derivative of nanotechnology building blocks diamondoids, known as Memantine, shows effective partial NMDA receptor antagonism [6]. Memantine is a derivative of adamantine (1-amino-3,5-dimethyladamantane) the smallest diamondoid molecule [10] ( Figure 4). Diamondoids are cage-like saturated hydrocarbons and one of the promising nanotechnology MBBs for biological applications. Memantine acts as a low-affinity and noncompetitive receptor antagonist which unbinds following strong synaptic activation due to its voltage-dependency and fast unblocking kinetics. Therefore, Memantine prevents excessive glutamate from setting off too high a calcium influx, while still allowing normal NMDA receptor to function at normal concentrations of glutamate [5]. Memantine has become of great interest to researchers owing to such self-regulated mechanism that controls the amount of NMDA activation [11].

Protection against Oxidative Stress
There is compelling evidence on initial role of mitochondrial derived oxidative stress in AD pathogenesis [12]. Mitochondria are crucial organelles to the eukaryotic cells for both energy generation and regulation of programmed cellular death (apoptosis). During the cellular energy metabolism in the mitochondria, semi reduced oxygen species (ROS), including super oxide and hydrogen peroxide, are normally produced [13]. According to the AD mitochondrial cascade hypothesis [14], during aging, which is the main risk factor for sporadic AD, the continual damage to mitochondrial DNA (mtDNA) leads to the malfunction of proteins of mitochondrial electron transfer chain (ETC) [12][13][14]. The resultant decrease in ETC efficiency per se causes overproduction of ROS [13].
Oxidative stress happens when the production of ROS exceeds their neutralization. The ROS, unless neutralized, can contribute to some cytopathological reactions, including oxidation of mitochondrial DNA (mtDNA) and to a lesser degree nuclear DNA, as well as protein oxidation and lipid peroxidation. These pathological reactions, in turn, lead to cell death and degeneration [12][13][14]. It is noteworthy that, in addition to mitochondrial dysfunction, the interaction between toxic Aβ aggregates and cell membrane is another major source of oxidative stress in AD pathology.
Fullerene is one of the main nanotechnology building blocks with strong antioxidant and free radical scavenger properties, owing to its specific chemical structure that allows it to be linked (i.e. functionalized) to several active chemical groups in a 3-dimensional orientation [15][16][17]. Fullerene molecule is composed of a three-dimensional array of carbon atoms, evenly spaced in a pattern similar to rhombuses on a soccer ball [15][16][17].
In addition, hydrated C 60 fullerene (C 60 HyFn) is shown to protect neurons against Aβ induced neurodegeneration in vivo, presumably owing to both its anti-oxidant and anti-amyloid properties [19]. The molecular mechanism of C 60 HyFn anti-amyloid properties is not clear yet [19]. Anti-amyloid properties were also attributed to sodium fullerenolate Na4[C 60 (OH)~30] (NaFL), a water soluble polyhydroxylated fullerene derivative [20]. NaFL destroys mature amyloid fibrils and at the same time prevents from Aβ aggregation and fibril formation. Thus, increased concentration of Aβ oligomers due to Aβ fibril break-down might be a source of potential toxicity in vivo. Of course, in vitro studies reported NaFL toxicity is exceptionally low. However, this is to be confirmed in pre-clinical in vivo trials [20].

Protection against Amyloid-β induced Synaptic and Neuronal Damage
In early stages of AD, symptoms are more due to Amyloid-β (Aβ) damage to synapses rather than its damage to neurons themselves. For example, the early memory impairment is more related to Aβ synaptotoxicity than its neurotoxicity [21]. Therefore, the synaptic degeneration and its presentations in early clinical symptoms of AD precede neuronal degeneration and death [22,23].
Growing evidence suggests the culprit molecules in synaptic A key functional alteration is excessive glutamate release from presynaptic compartment and/or impaired glutamate clearance by excitatory amino acid transporters (EAAT) of astrocytes. Excessive glutamate over activates NMDA receptors, causing excitotoxicity through pathological influx of Ca2+. Excessive intracellular Ca2+ subsequently triggers various pathological events such as membrane lipid peroxidation, formation of reactive oxygen species, mitochondrial fragmentation and ultimately cell death [5].  Figure 3: Schematic representation of NMDA receptor and its potential binding sites for ligands [10]: As it is shown most of NMDA receptor subtypes have quaternary structure, composing of two NR1, and two NR2 subunits (here only one pair of subunits is schemed). The large extracellular domain consists of two regions: NTD and ABD. The ABD region contains the site for binding with ligands. NR1 allows binding with glycine (or D-serine) and glutamate occupies the NR2 binding site. Other agonist and antagonist ligands could bind at the same sites (as pointed with large arrows). The thick orange arrows show the site for binding with allosteric modulators (like endogenous zinc) and the thin orange arrows indicates the putative sites for both positive and negative allosteric modulators. The ion channel domain also provides binding sites for pore blockers such as endogenous Mg++, Ketamine and Memantine, all acting as uncompetitive antagonists.  HCl). NAMENDA is the commercial medicine name for the neutralized Memantine [6] degeneration are the soluble oligomeric species of Aβ (AβO), and the process is independent of the levels of amyloid precursor protein (APP) expression, Aβ monomers and amyloid depositions [22,23]. The exact downstream mechanisms of AβO on synaptic degeneration are not clear yet. However, postsynaptic terminals are shown to be more affected than pre-synaptic terminals in AβO induced toxicity ( Figure  2) [5]. Interestingly, AβO induced synaptic disruptions are more functional in the beginning, and are followed by structural disruptions later in the disease process. The functional synaptic impairment mainly include a reversible disruption in long term potentiation [24,25].
Some signaling pathways have been suggested to account for the functional toxic effects of AβO. The backbone of these pathways is a putative synaptic terminal receptor with high affinity to AβO, which is not yet fully characterized [5,25]. However, there is strong supporting evidence regarding association of AβO with some synaptic terminal receptors, including Fyn, Arc, glutamate receptors and PSD 95 protein.
It is hypothesized that specific interactions of AβO with PSD-95 protein lead to a displacement of Fyn, which is tethered by PSD-95 protein to the signaling chain [26]. Displacement of Fyn receptor will then inhibit the c-AMP response element-binding (CREB) protein signaling ending in synaptic dysfunction and eventually neuronal death [26].
Not only the extracellular AβO, but also the intra-membranous and intracellular AβO can have toxic effects on synaptic terminals. We have discussed the intra-membranous oligomerization of Aβ and its subsequent permeabilization effect on the plasma membrane as a neurotoxic mechanism elsewhere [2,3]. In addition to the same intra-membranous oligomerization mechanism, intracellular Aβ oligomerization is proposed as a mechanism for synaptic degeneration following intra terminal accumulation of Aβ [27].
Aβ-synaptotoxicity may also occur through localized apoptotic changes as evidenced by caspase activation in neurites [28]. Caspase is the name for a family of cysteine proteases that are responsible for regulating apoptosis in mammalian cells. However, it is not completely proven whether the activation of caspases in AD causes classic apoptotic effects or if it merely performs some apoptosis-associated molecular dysfunctions in the neurons [28]. Some caspases, especially caspase-6, have been recently suggested as the first molecular responses of the insulted neurons and as triggers for further disruptions in the structure of neuronal cytoskeleton and subsequently in the neuronal trafficking [28,29].
Blocking toxicity of AβO has been widely studied as a disease modifying strategy for AD. To date, dendrimers are one of most validated nanotechnology designed structures with anti-amyloid properties [34]. Dendrimers are highly branched macromolecular structures with globular shape ( Figure 5) and a densely packed surface [30][31][32][33][34]. Beside their potent anti-amyloid effects, however, most dendrimeric structures exert cytotoxic effects on neurons. Therefore, designing dendrimers with minimal toxicity has been the focus of several recent studies. Recently, minimally toxic dendrimers like poly-L lysine dendrimers D3 and D5 were found with strong anti-amyloid properties as demonstrated in human neuroblastoma SH-SY5Y cells. In addition, D3 lysine dendrimers, attenuated Aβ induced increases in K + channel current [35]. Interestingly, following intraventricular injection of fluorescence-labelled D3 dendrimers to rat brain, dendrimers were found distributed to both parietal cortex and hippocampus, major regions involved in AD, without any sign of cell death throughout the brain [35].
Interestingly, in a recent study, cationic phosphorus-containing dendrimers (CPD) (generations 3 and 4) were proved capable of interfering with both Aβ and MAP-Tau aggregation processes [37]. In this study, the fragment of amyloid peptide (Aβ 1−28 ), which is the core region responsible for aggregation, was used as a model for full-length Aβ monomer [37]. The cytotoxicity associated with formed fibrils and intermediate products of Aβ 1−28 aggregation decreased in the presence of CPDs via regulation of the fibrillization process [37].
Generally, dendrimers regulate Aβ fibrillization in a concentration dependent manner. In low concentrations, dendrimers facilitate Aβ fibrillization, whereas in higher concentrations, they block Aβ fibrillization [33,34], either via their binding with peptide monomers or through blocking the end of protofibrils and fibrils [34]. The therapeutic importance of this dualistic behavior of dendrimers is rooted in the recently described protective, rather than toxic role of Aβ fibrillization. Based on this paradigm, Aβ fibrillization is a natural mechanism to neutralize the toxic AβOs, which can be considered as an alternative way to compensate for the impaired Aβ clearance mechanisms in AD brain. In this regard, lower concentration of dendrimers could contribute to disease modification of AD through neutralizing toxic AβOs via enhancement of Aβ fibrillization [33,34]. An exception may be gallic acid-triethylene glycol dendrimer decorated with 27 terminal morpholine groups ([G3]-Mor) that facilitate amyloid fibirilization in high concentrations [38]. Interestingly, in line with the abovementioned notion, the results of G3-Mor and Aβ interaction showed concomitant increase in Aβ fibril formation and decrease in Aβ-induced toxicity [38].

Conclusions
Thus far, all disease modifying strategies designed to slow or halt the Alzheimer's disease molecular pathogensis have failed in phase-3 clinical trials [4]. Future perspective on finding a cure for AD is not bright, without any considerable shift in the paradigms explaining the underlying mechanisms of AD, establishing and validating relevant animal models used in preclinical studies [39] and methodologies used in drug design and discovery. While toxic side effects are still the major concern about different nanoparticles, biochemical features of these particles show potentials of a methodological breakthrough in targeted drug delivery to specific molecules involved in AD pathogenesis.
Recently, disease modifying potentials have been suggested with Memantine [40,41], a derivative of adamantane the smallest diamondoid molecule [42,43]. Diamondoids are key molecular building blocks for nanotechnology with several biological implications [44], including NMDA receptor antagonism as shown in NAMENDA. Clinically, NAMENDA is beneficial for treatment of moderate to severe AD.
As the currently dominant paradigm of AD pathogenesis emphasizes on Aβ induced toxicity, several nanotechnology building blocks have shown neuroprotective effects against different mechanisms through which Aβ damages neurons. For instance, Aβ damage to synapses and neuronal cell membrane could be restrained via anti-amyloid (e.g. dendrimers) and anti-oxidant (e.g. fullerenes) nanotechnology building blocks respectively.
It is noteworthy that beside efforts towards effective disease modifying agents for AD, nanotechnology research is also focusing on drug delivery systems that can allow targeted delivery of both current symptomatic and envisioned disease-modifying agents to the brain [2,45], thus providing higher efficacy with minimal toxicity.
In conclusion, nanotechnology can provide potent agents to modify the course of certain pathological mechanisms involved in AD. However, translation of these strategies to actual modification of Alzheimer's disease course in human subjects lies largely not only on biocompatibility of these agents, but also on the accuracy of our current paradigms about molecular mechanisms underlying AD [46].