Currently, the pharmacotherapy for AD and PD consists of drugs approved by the Food and Drug Administration (FDA) that regulate neurotransmitter levels. Unfortunately, they only provide valuable but modest symptomatic benefits, being unable to modify the course of these diseases [
9,
148]. These treatments are also accompanied by limitations. For instance, AChE inhibitors offer relatively short-lasting positive effects in AD patients [
149] and display cholinomimetic actions on the gastrointestinal tract that result in diarrhea, nausea, and vomiting [
150]. The efficacy of PD medicines also decreases over time, and the chronic treatment often culminates in motor complications (e.g.: L-DOPA-induced dyskinesia) [
151].
4.3.1. Oxidative Stress as a Target in Alzheimer’s and Parkinson’s Diseases
Oxidative stress is one of the major contributors to the pathogenic cascade that leads to neurodegeneration in AD and PD [
207,
208]. Evidences of reactive species (RS)-mediated injuries, with increased levels of oxidative markers and damaged cell components, were observed in AD and PD brains [
209]. A decline in the pool of endogenous antioxidants and a decrease in the activity of antioxidant enzymes were also reported [
25,
210].
The brain is particularly prone to oxidative stress-induced damage. Although the brain constitutes only
2 % of the total body weight, it is responsible for more than 20 % of the body's oxygen consumption, with a significant amount of oxygen being converted into reactive oxygen species (ROS) [
208,
211,
212]. Despite this massive oxygen consumption, the brain presents a lower content of endogenous antioxidants (e.g.: glutathione, catalase) in comparison to other tissues, thus being more sensitive to cellular redox dyshomeostasis [
211,
213]. In addition, redox-active metals (e.g.: iron, copper) accumulate in specific brain regions and catalyze the formation of ROS [
120,
208]. Finally, the high levels in polyunsaturated fatty acids in the brain increase the susceptibility to lipid peroxidation and subsequent formation of toxic compounds [
208,
211,
214].
The increased oxidative stress in NDs is strictly connected to other pathological events, namely mitochondrial dysfunction, dopamine oxidation, neuroinflammation, and accumulation of protein aggregates (e.g. Aβ, α-syn) (
Figure 10) [
215].
Although RS are generated in several cellular compartments,
mitochondria is one of the main sources of the overproduction of ROS [
121]. The formation of ROS in mitochondria occurs primarily at the ETC present in the inner mitochondrial membrane (
Figure 11) [
122,
208]. The mitochondrial ETC consists of a series of membrane-bound complexes (complexes I, II, III and IV) [
117], which generate a proton gradient across the inner mitochondrial membrane through electron transfers, leading to the production of ATP by ATP synthase (complex V) [
208]. Metabolic intermediates formed during the Krebs cycle are used for oxidative phosphorylation [
118]. During the ETC, a small proportion of electrons occasionally leak and directly reduce O
2 to
[
118,
119], which in turn is converted into other ROS such as H
2O
2 and HO
• [
121]. The formation of
occurs mainly in complexes I and III [
119,
216]. Enzymes from the Krebs cycle (e.g.: α-ketoglutarate dehydrogenase, pyruvate dehydrogenase, aconitase) may also generate ROS [
216,
217].
Mitochondrial ETC is one of the primary targets of the harmful effects inflicted by high levels of ROS [
117,
219]. The oxidative damage at this level leads to the inhibition of ATP synthesis and the increased production of ROS in a vicious and detrimental cycle, contributing to cell dysfunction and cell death [
122,
219,
220]. Mitochondria contain other components susceptible to oxidative damage, namely several iron-sulfur centers, proteins and unsaturated fatty acids in the inner membrane, and mitochondrial DNA (mtDNA), all of which are important for proper mitochondrial function [
122]. Considering that mtDNA encodes some of the subunits of the complexes that constitute the ETC, the oxidative damage of mtDNA leads to the defective production of these proteins and subsequent mitochondrial dysfunction [
219].
Since neurons have limited glycolytic capacity, they are particularly dependent on mitochondrial oxidative phosphorylation to meet their high energy requirements [
220,
221,
222]. In addition to ATP synthesis, mitochondria are involved in other crucial cellular functions such as the synthesis of amino acids and steroids, β-oxidation of fatty acids, Ca
2+ homeostasis, and regulation of apoptotic cell death [
223]. Therefore, improper mitochondrial function compromises neuronal survival and contributes to neurodegeneration [
217,
223].
In PD,
DA oxidation is associated with a selective vulnerability of dopaminergic neurons to oxidative stress [
218]. Despite the essential role of DA in neurotransmission, DA contains a catechol group that may participate in the generation of ROS and metal chelation [
224]. Dopamine is normally stored in monoaminergic vesicles under a low pH environment that prevents its oxidation [
225]. However, DA may undergo enzymatic and non-enzymatic decomposition in the cytosol, which is accompanied by the formation of ROS (
Figure 12) [
226].
In the presence of O
2, DA generates
and electron-deficient DA semiquinones and DA quinones (
Figure 12) [
226,
228,
229]. The reaction rate of DA semiquinone formation is slow, but it is accelerated by redox-active transition metals [
213]. The spontaneous cyclization of DA quinone yields leucoaminochrome whose further autoxidation forms aminochrome and
[
228]. Aminochrome participates in redox-cycling reactions that results in the formation of
and in the depletion of cellular nicotinamide adenine dinucleotide (NADH) and nicotinamide adenine dinucleotide phosphate (NADPH) [
230]. Dopamine quinone and aminochrome also form adducts with cellular nucleophiles modifying their function [
231,
232]. These include DNA, biothiols (e.g.: glutathione), α-synuclein and proteins involved in ATP synthesis (complexes I, III and V of the ETC), proteasomal degradation (parkin), microtubule stabilization (α- and β-tubulin) and axonal transport (actin) [
227]. Therefore, the formation of these adducts will contribute to mitochondrial dysfunction, impairment of the axonal transport, inhibition of the proteasomal system, disruption of cytoskeleton architecture, and formation of α-synuclein aggregates in PD [
227]. Aminochrome also polymerizes into neuromelanin, a brain pigment that contributes to neurodegeneration by triggering neuroinflammatory processes [
208].
The oxidative deamination of DA by MAOs uses O
2 and generates H
2O
2 and ammonia as by-products (
Figure 12) [
233]. Due to the increased expression with age in neuronal tissue [
102,
234], MAO-B becomes the predominant isoform involved in DA metabolism [
208]. Monoamine oxidase B is mainly found in glial cells [
105,
235], but the H
2O
2 produced during DA deamination can permeate cell membranes and induce toxic effects in the neighboring neurons [
208]. In fact, compared with astrocytes, neurons are more vulnerable to H
2O
2 due to the lower content in antioxidants involved in its detoxification (e.g.: GPx and glutathione) [
235]. The H
2O
2 generated from MAO-B activity in astrocytes is also associated with increased amyloid plaque deposition [
111].
Neuroinflammation represents a set of inflammatory processes occurring in the central nervous system that involve the action of glial cells in CNS (microglia, oligodendrocytes, astrocytes), non-glial resident myeloid cells (macrophages and dendritic cells) and peripheral leukocytes [
236,
237]. Neuroinflammation plays an important role in the progression of NDs [
226]. For instance, in AD, microglia are activated by the presence of Aβ and co-localize with the plaques [
238]. However, instead of efficiently removing the Aβ deposits, microglia release pro-inflammatory mediators that lead to neuronal damage [
239]. In PD, extracellular αSYN aggregates can also interact with and activate surrounding glial cells to trigger a deleterious pro-inflammatory response [
240]. In NDs, the expression of NADPH oxidases (NOXs) in activated microglia and reactive astrocytes is increased, resulting in the excessive formation of
[
226,
241]. The activation of RS-producing enzymes in glial cells is associated with neurotoxic effects, which arise not only from the direct oxidative damage in neurons, but also from the intracellular redox signaling that exacerbates the pro-inflammatory response [
241,
242].
4.3.1.1. Targeting Oxidative Stress with Mitochondria-Targeted Antioxidants
Considering the involvement of oxidative stress in the pathophysiology of NDs, the rationale for using exogenous antioxidants to prevent delay, or remove the oxidative damage is evident [
212,
243]. In fact, several exogenous antioxidants showed promising results in animal and cellular models [
211,
212]. However, the results obtained in clinical trials were inconclusive, negative, or showed little benefit in NDs [
244]. Numerous factors contribute to the discrepancy between pre-clinical and clinical results. In addition to aspects associated with the design of clinical trials (e.g.: posology, duration of treatment, age, and disease stage of the patients), most known dietary antioxidants display poor bioavailability and are unable to cross the BBB, affecting their delivery into the brain [
211,
212,
244,
245].
A common strategy used to overcome these pharmacokinetic limitations is the introduction of minor structural modifications on the antioxidant scaffold. The resulting derivatives may improve the targeting and drug-like properties while preserving or enhancing the antioxidant profile of the parent compounds [
246,
247].
Aside from the pharmacokinetic constraints, the lack of clinical efficacy of antioxidants may also result from the uniform distribution of antioxidants across all tissues and organs following administration, with only a small fraction being taken up by mitochondria [
244,
248], the main source and the target of ROS. Therefore, the development of antioxidants that selectively accumulate within mitochondria and tackle oxidative damage is of particular interest [
249]. Compounds lacking mitochondriotropism but with relevant biological activities towards mitochondrial targets usually need to be directed to mitochondria [
250]. In this sense, several approaches were developed to deliver antioxidants and other bioactive molecules to mitochondria, but one of the most widely used is their conjugation with lipophilic cations such as triphenylphosphonium (TPP
+) [
249,
251].
Lipophilic TPP
+ cations can diffuse across phospholipid bilayers because their positive charge is surrounded and dispersed over a large hydrophobic surface area, which decreases the activation energy for membrane permeation [
252,
253,
254]. In response to the plasma and mitochondrial membrane potentials (
, respectively), these compounds accumulate within the mitochondrial matrix against the concentration gradient [
252] (
Figure 13A). Then, TPP
+ conjugates are taken up from the intracellular space to the mitochondrial matrix in response to the
(-140 to -160 mV), leading to 100 to 500-fold accumulation within the mitochondrial matrix [
218,
254].
The increased accumulation of lipophilic TPP
+ conjugates enhances the compounds’ potency and decreases the external dose required, limiting extramitochondrial metabolism that results in inactivation, excretion, or toxicity [
122,
256]. However, the extensive accumulation of these compounds within the mitochondrial matrix can disrupt membrane integrity and thereby compromise cellular respiration and ATP production (281, 294).
Following oral or intravenous administration, lipophilic TPP
+ conjugates are rapidly taken up by the organs most affected by mitochondrial dysfunction (e.g.: liver, heart, brain) [
252,
257]. Therefore, targeting antioxidants in mitochondria stands out as a promising strategy in the discovery of new therapies for oxidative stress-related disorders.
Over the last decade, TPP
+ cations have been conjugated with dietary antioxidants such as hydroxybenzoic [
258] and hydroxycinnamic acids [
259]. These compounds displayed remarkable antioxidant properties and were able to protect neuroblastoma cells against the oxidative damage induced by 6-hydroxydopamine or H
2O
2 [
260]. Moreover, in studies performed in skin fibroblasts from male sporadic PD patients (sPD), the caffeic acid-based TPP
+ conjugate AntiOXCIN4 restored mitochondrial membrane potential and mitochondrial fission, decreased autophagic flux, and enhanced cellular responses to stress by improving the cellular redox state and decreasing ROS levels [
261]. To circumvent the drawbacks associated with the use of TPP
+ cation, its replacement with nitrogen-based cationic carriers (e.g. isoquinolinium, imidazolium, picolínium) was recently performed (
Figure 13B) [
262]. This chemical modification resulted in decreased cytotoxicity while maintaining the compounds’ antioxidant properties and their ability to accumulate within mitochondria [
262].
4.3.2. Adenosine Receptors as a Target in Alzheimer’s and Parkinson’s Diseases
Adenosine is a purine nucleoside that may act as a neurotransmitter as neuromodulator in the CNS [
263]. It is involved in several physiological and pathophysiological processes in the brain, including motor function, sleep/wake cycle, learning and memory, pain, and astrocytic activity [
264]. To perform its physiological roles, adenosine binds to four distinct G-coupled protein adenosine receptors (ARs), designated as A
1, A
2A, A
2B and A
3. Adenosine receptors (ARs) represent a group of glycoproteins containing seven transmembrane domains and are coupled to different G proteins [
265] (
Figure 14). While adenosine A
1 and A
3 receptors are coupled to inhibitory G proteins, A
2A and A
2B ARs are coupled to stimulatory G proteins. The A
2A and A
2B ARs preferably interact with members of the G
s family of G proteins, stimulating adenylyl cyclase to produce cyclic AMP (cAMP) and leading to the activation of a series of downstream signaling pathways. In contrast, A
1 and A
3 ARs inhibit the adenylyl cyclase activity by interaction with G
i proteins (
Figure 14) [
266].
ARs are widely distributed in the human body and participate in a broad range of physiological and pathophysiological processes [
196]. While A
1Rs and A
2Rs can be predominantly found in specific parts of the CNS, A
2BRs and A
3Rs are mainly located in peripheral tissues [
267] .
In the CNS, A
1Rs are widely distributed in neocortical and limbic systems and are linked to cognitive functions [
196,
268,
269]. A
2ARs are highly expressed in striatal areas [
196,
268] and participate in the regulation of motor behavior and the management of dopamine-mediated responses [
197]. A
2ARs co-localize with dopamine D
2 receptors (D
2Rs) on GABAergic striatopallidal output neurons, where they form heteromer complexes [
270]. These receptors within the heteromeric complex exert opposite effects on motor behavior, in which A
2A AR agonism induces antagonistic effects on D
2Rs. For instance, stimulation of dopamine D
2Rs enhances motor activity, while A
2A ARs decrease this effect by decreasing the affinity and response of D
2Rs to their ligands [
269,
271].
Excessive A
2A AR function has been linked to neuronal damage [
272], and increased A
2A AR expression is a characteristic feature of PD progression [
273]. The cellular mechanisms responsible for A
2A AR-mediated neurodegeneration remain elusive. However, evidence suggests that the activation of A
2A ARs leads to increased glutamate release, increased Ca
2+ entry, and enhanced long-term potentiation, all of which may culminate in excitotoxic damage [
271]. The localization of A
2A ARs at the basal ganglia, coupled with their pathophysiological role in PD, makes these receptors attractive drug targets to treat this disease [
271]. A
2A AR antagonism decreases motor impairment by enhancing dopamine D
2R-mediated signaling. Moreover, A
2A AR antagonism modulates cholinergic, glutamatergic, and GABAergic functions in the CNS [
269]. Blockade of A
2A AR signaling with selective A
2A AR receptor antagonists was shown to be beneficial, not only by enhancing the therapeutic effects of L-DOPA, but also by reducing dyskinesia from long-term L-DOPA treatment [
270].
Recent studies have also been disclosing a close association between A
2A ARs and cognitive impairment in AD. For instance, abnormally high levels of A
2A AR were detected in the hippocampus and in the cortex of AD patients [
262,
274] and in APP/PS1 transgenic AD mice [
275]. Remarkably, activation of A
2A ARs with agonists and optogenetic agents led to severe impairments in spatial discrimination in wild-type mice [
276]. The involvement of A
2A ARs in in hippocampal-dependent spatial reference memory was also shown in A
2A AR knock-out studies in an Aβ
1-42-based mice model of AD [
277]. The memory deficits in APP/PS1 mice were reverted by the blockade of A
2A ARs with a selective antagonist or by downregulation driven with shRNA interference [
275]. Finally, recently it was shown that the improvement of spatial memory deficits by A
2A AR antagonists in APP/PS1 mice results from the promotion of synaptic plasticity of adult-born granule cells [
278]. Thus, the blockade of A
2A AR activation with selective antagonists can be of great therapeutic benefit to AD patients.
4.3.2.1. A2A Adenosine Receptor Antagonists
The knowledge acquired over the last decades concerning the involvement of adenosine on motor functions, mainly through modulation of A
2A AR, makes A
2A AR antagonists promising non-dopaminergic agents for the treatment of PD motor symptoms. Over the last decades, the development of potent and selective ligands for ARs has been a dynamic area. Excellent reviews were recently published on this topic [
279,
280]. A small number of selective A
2A AR antagonists reached advanced clinical trials for the treatment of motor symptoms in PD, namely the xanthine derivative istradefylline (KW-6002) and the non-xanthine derivatives Tozadenant (SYN115), Preladenant, and KW-6356 (
Figure 15) [
268,
273].
Istradefylline was approved for the adjunctive treatment of PD in Japan in 2013 and by the FDA in 2019, being the first non-dopaminergic drug approved by FDA for PD in the last two decades [
281]. Preladenant and Tozadenant underwent clinical evaluation for the treatment of PD (Preladenant: NCT00406029, NCT01227265; Tozadenant: NCT02453386, NCT03051607) [
282]. Unfortunately, the clinical evaluation for both drug candidates were discontinued due to the lack of efficacy (Preladenant) or safety (Tozadenant) in phase 3 clinical trials [
283].
KW-6356 is a new, selective, nonxanthine A
2A receptor antagonist/inverse agonist. Compared to istradefylline, KW-6356 exhibits approximately 100-times higher affinity for the human A
2A receptor and a prolonged drug residence time [
284]. In a phase 2b clinical study in patients with PD, KW-6356 was safe and effective in the adjunctive treatment with L-DOPA (NCT03703570) [
285]. Moreover, in a phase 2a clinical trial, KW-6356 monotherapy was well tolerated and more effective than placebo in patients with early, untreated PD (NCT02939391) [
273].