Mitochondrially-Targeted Therapeutic Strategies for Alzheimer’s Disease

Abstract

Alzheimer’s disease (AD) is an irreversible, progressive neurodegenerative disease and the most common cause of dementia among older adults. There are no effective treatments available for the disease, and it is associated with great societal concern because of the substantial costs of providing care to its sufferers, whose numbers will increase as populations age. While multiple causes have been proposed to be significant contributors to the onset of sporadic AD, increased age is a unifying risk factor. In addition to amyloid-β (Aβ) and tau protein playing a key role in the initiation and progression of AD, impaired mitochondrial bioenergetics and dynamics are likely major etiological factors in AD pathogenesis and have many potential origins, including Aβ and tau. Mitochondrial dysfunction is evident in the central nervous system (CNS) and systemically early in the disease process. Addressing these multiple mitochondrial deficiencies is a major challenge of mitochondrial systems biology. We review evidence for mitochondrial impairments ranging from mitochondrial DNA (mtDNA) mutations to epigenetic modification of mtDNA, altered gene expression, impaired mitobiogenesis, oxidative stress, altered protein turnover and changed organelle dynamics (fission and fusion). We also discuss therapeutic approaches, including repurposed drugs, epigenetic modifiers, and lifestyle changes that target each level of deficiency which could potentially alter the course of this progressive, heterogeneous Disease while being cognizant that successful future therapeutics may require a combinatorial approach.

1. INTRODUCTION

Alzheimer's disease is the most common neurodegenerative disease in patients over 60 years and there are currently more than 50 million individuals affected globally at an annual cost in excess of US$ 1 trillion. Without breakthrough therapy, more than 152 million people will be affected by 2050 (https://www.alz.co.uk/research/WorldAlzheimerReport2018.pdf). While our understanding of the genetics of the disease has increased considerably, there are still no effective prevention or therapeutic strategies available to stop or slow disease progression. Current treatment only provide temporary symptomatic relief when administered in a timely manner [1]. While cholinesterase inhibitors are prescribed for all stages of AD dementia, N-Methyl-D-aspartate (NMDA) receptor antagonist memantine or a combination of cholinesterase inhibitors and glutamate inhibitors is FDA approved for moderate to severe AD. The nutraceutical Huperzine A improves cognition and daily function but is not used in the USA as it is not regulated by the US Food and Drug Administration (FDA) [2-4]. Early-onset familial AD (fAD) comprise <10% of AD cases where ~ 60% have at least one affected first-degree relative and 13% of them are inherited in an autosomal dominant manner [5, 6]. Mutations in the amyloid precursor protein (APP; chromosome 21), presenilin-1 (PSEN1; chromosome 14), and presenilin-2 (PSEN2; chromosome 1) cause fAD [7]. These mutations drive amyloidosis [8-11], resulting in an early disease onset (20-30 years of age) [12, 13], although most of the cases are diagnosed between 45 and 60 years. Most AD cases are late-onset sporadic (sAD) (>95% prevalence). While age is considered to be a principal risk factor for sAD [14], apolipoprotein E (APOE), the strongest genetic risk factor for sAD [15] and meta-analysis of genome-wide association studies (GWAS) have identified novel genes sets that are associated with significant sAD risk including: bridging integrator 1 (BIN1), clusterin (CLU), complement component (3b/4b) receptor 1 (CR1), disabled homolog 1 (DAB1), PI-binding clathrin assembly protein (PICALM), Sortilin- like receptor 1 (SORL1), triggering receptor expressed on myeloid cells 2 (TREM2), the membrane-spanning 4-domains, subfamily A (MS4A), ATP-binding cassette transporter A1 and A7 (C1 and 7), methylenetetrahydrofolate dehydrogenase 1 (MTHFD1) and CD33 [16, 17]. These genes are involved in innate immunity, inflammation, lipid metabolism, production and clearance of amyloid-β, and endosomal vesicle recycling [18-20].

Female gender is an additional sAD risk factor that is partly explainable by aging and lengthened lifespan. Other factors that may increase the risk for sAD include chronic inflammation, lack of physical activity, depression, social isolation, diabetes, obesity hypertension and smoking [21, 22]. Cumulative evidence reveals that mitochondrial dysfunction is a key player in AD pathogenesis. While the human brain accounts for ~2% of the body weight, but it consumes ~20% of glucose-derived energy [23]. This very high energy demand makes it particularly vulnerable to mitochondrial dysfunction [24]. Glucose hypometabolism is consistently evident in the hippocampus and cortical regions and precedes the clinical diagnosis of AD by decades and quite accurately predicts cognitive decline in normal aging [25] as well as progression from mild cognitive impairment (MCI) to AD [26, 27]. Indeed, mitochondrial dysfunction and glucose hypometabolism consistently precede Aβ plaque and neurofibrillary tangles in AD pathology [28]. Further, the ROS arising from oxidative phosphorylation (OXPHOS) adversely affects cellular structures such as membranes, proteins, DNA and lipids when produced in excess. mtDNA are particularly vulnerable to mitochondrial ROS (mtROS) because they are not only close to the respiratory chain, but unlike nDNA, they do not have protective histones. Damaged mtDNA further alters the respiratory chain and trigger a vicious cycle by enhancing the generation of free radicals that results in neuronal dysfunction and aging phenotype [29]. This oxidative stress is further compounded by inflammatory processes associated with alterations in cellular metabolism [24, 30, 31]. Type 2 diabetes mellitus (T2DM) confers an additional risk factor for AD [32-36], suggesting that deranged metabolism contributes to AD development [11, 37, 38]. Several recent studies demonstrate epigenetic modifications in aging and neurodegeneration. These include histone modifications, DNA methylation, and microRNA expression that alter gene expression and genome architecture [39-42]. There is growing evidence that the mitochondrial DNA (mtDNA) like nuclear DNA, can be controlled by epigenetic mechanisms [43-46]. While several diseases including AD are associated with differential mtDNA methylation, it has not yet been elucidated whether the disease results from the methylation or the methylation results from the disease [47]. Epigenetic changes to nucleoid proteins may influence the regulation of mtDNA gene expression as these lack histones. There is an urgent need to develop effective disease-modifying drugs for AD and a paradigm that alleviates the bioenergetic deficit in vulnerable neurons of affected brain regions may achieve better outcomes in AD patients. This can be achieved by restoring optimal mitochondrial function in AD.

1.1. The Mitochondrion

The human mitochondrion holds two to ten copies of 16.5 kilo base (kb) double stranded, closed circular mitochondrial DNA (mtDNA) [48-50]. mtDNA codes only 13 polypeptides, two rRNAs (12S and 16S) and 22 tRNAs that are essential for the oxidative phosphorylation system [51]. The cell’s nuclear genome encodes the rest of the mitochondrial proteins, metabolic enzymes, DNA and RNA polymerases, ribosomal proteins, and mtDNA regulatory factors, such as mitochondrial transcription factor A. Nuclear mitochondrial proteins are synthesized in the cytoplasm and then transported into mitochondria. mtDNA are primarily maternally inherited with many thousands of copies per cell. An equilibrium between mitochondrial fission and fusion regulates the number and morphology of mitochondria [52, 53]. This is crucial for metabolism, energy production, calcium (Ca²⁺) signaling, reactive oxygen species (ROS) production, apoptosis, and senescence [54-58]. The trafficking of mitochondrial components between different mitochondria is facilitated by fusion. mtDNA is hyper-mutable compared with nuclear DNA (nDNA) because they replicate more frequently and also because they do not have protective histones and are close to the respiratory chain. As mtDNA mutations arise in the maternal lineage, a heteroplasmic state is established as a mixture of the wild-type and mutant mtDNAs. As cells divide, the mutant mtDNAs are randomly passed onto the daughter’s cells and the percentage of mutant mtDNAs in different cell lineages drift toward either pure mutant or normal (known as homoplasmy) [59, 60]. The dynamic process by which mtDNA mutations accumulate (clonal expansion) results in a decline in cellular energy output and a total cellular mitochondrial dysfunction. The ratio of mutant to normal mtDNA, therefore, determines the severity of the disease. Severely damaged mitochondria are unable to fuse and this results in the fragmentation of mitochondria [61].

1.2. Mitochondria in AD

AD mitochondria are fundamentally altered in many ways. Energy metabolism is impaired in AD [62-65] and PET studies consistently demonstrate glucose hypometabolism in the brain [66] that precede the onset of the histopathological hallmarks and symptoms [67, 68]. The activities of the mitochondrial complex IV cytochrome c oxidase (COX), pyruvate dehydrogenase complex (PDHC), mitochondrial isocitrate dehydrogenase, α-ketoglutarate dehydrogenase (αKGDH), and ATP synthase complex are decreased [67, 69, 70] and this compromises the mtΔΨ and ATP production [71]. Aged mitochondria can induce neuronal death through altered mitochondrial dynamics and biogenesis/mitophagy processes; impaired Ca²⁺ homeostasis; mutations on mtDNA; aberrant activation of apoptotic pathways; oxidative stress; and altered bioenergetics [72]. The mitochondrial dynamics are disrupted in AD [73-75], resulting in altered sizes and shapes (including both enlarged, very small, and elongated mitochondria [76-79], with fewer cristae [77, 80, 81] and reduced expression of mitochondrially encoded electron transport chain (ETC) enzymes [65, 82, 83]. These alterations in mitochondrial dynamics are also evident in the peripheral blood of AD patients and may be potential blood based biomarkers for the disease [84]. In AD, misfolded proteins impair mitochondrial activity through direct interaction with mitochondrial structures and impede trafficking and dynamics. This results in impaired bioenergetics and quality control pathways, and trigger mitochondria-dependent apoptosis [85]. Aβ peptides can localize on the endoplasmic reticulum (ER) mitochondrial associated membranes (MAMs) and induce the release of cytosolic Ca2+ that leads to mitochondrial Ca²⁺ overload, stimulation of mitochondrial respiration and increased ROS generation [86-89]. Aβ peptides have been shown to interact with components of the mitochondrial matrix [90] and in excess, Aβ alters mitochondrial dynamics by differentially modulating fission/fusion proteins [91, 92]. Finally, the effects of aggregating Aβ can be mitigated by efficient mitochondrial proteostasis [75, 93]. Since mitochondria are structurally and functionally altered in AD [94-96], compounds that can induce and/or restore their bioenergetic capacity present an attractive strategy AD therapy. We review nascent developments of mitochondrially targeted approaches, including repurposed drugs, epigenetic modifiers, and lifestyle changes that target each level of deficiency, which show promise for AD treatment although it is likely that successful future therapeutics may require a combinatorial approach.

2. POSSIBLE MITOCHONDRIALLY TARGETED INTERVENTIONS IN AD

2.1. Editing mtDNA Mutations

The frequency of mtDNA deletions and mutations increases with age in human somatic tissues [97], with the substitution rate of mtDNA being an order of magnitude higher than that of the nuclear DNA (nDNA) mutations [98, 99]. While these alterations do not result in a significant change in the absolute copy number, they reduce respiratory activity and are considered to be important drivers of aging [100-102] and the pathophysiology of AD [103-105]. The common mtDNA 5 kb deletion is increased > 15 times in AD brains [106, 107] and somatic mtDNA control region (CR) mutations are increased 73% in AD brains [108]. With the rapid development of gene editing, mtDNA modifications with fewer side effects can provide a new pathway to developing target-oriented molecular networks, interactions, and mitochondrial biology.

A gene editing tool that does not involve CRISPR and is capable of precise editing of mtDNA in vitro has been recently described [109]. It is based on an interbacterial toxin A called DddA, which deaminates cytidines within dsDNA and has been engineered into non-toxic and inactive halves that are activated when coupled on target DNA by adjacently bound programmable DNA-binding proteins. Fusion of the split-DddA halves, transcription activator- like effector array proteins, and a uracil glycosylase inhibitor create RNA-free DddA-derived cytosine base editors (DdCBEs) that catalyze C•G-to-T•A conversions in human mtDNA with high target specificity and product purity.

Mitochondrial-targeted transcription activator-like effector nucleases (mitoTALENs) are able to rectify mtDNA mutations in an mtDNA disease patient cell culture [110, 111] and also correct induced mtDNA mutation in mouse models [112]. Mitochondrially targeted zinc- finger nucleases (mtZFNs) [113] can also be used to specific remove mtDNA mutations without any interaction with the cell’s nuclear DNA. mtZFNs can eliminate pathogenic mtDNA mutation in mouse models [114]. Adult cells accrue age related mtDNA mutations and are heteroplasmic and the extent (threshold) of the mutant mtDNA mutation load determines when onset of clinical symptoms [115]. Shifting the heteroplasmic equilibrium can ameliorate the onset of clinical symptoms [116]. While CRISPR and mitoTALENs can rectify mtDNA mutations, mtZFNs selectively eliminate mitochondria that harbor mtDNA mutations, ultimately repopulating cells with healthy mitochondria [117]. This is ideal for clinical applications as it eliminates the risk normal genes being erroneously being altered by gene-editing techniques [112-114, 118-121].

2.2. Intercellular Mitochondrial Transplantation

The transplantation of functional mitochondria directly into defective cells is a novel approach for combating mitochondrial dysfunction. This results in enhanced bioenergetics, reduced ROS production and restoration of mitochondrial function. It was recently discovered that astrocytes may release extracellular mitochondrial particles that are endocytosed into injured neurons via an actin-dependent mechanism and restore neuronal viability and recovery after stroke [122-124]. In vitro cytoplasmic hybrid (cybrid) cell system was the original mitochondrial delivery method in an AD model. Cybrids result from the fusion of mt-DNA depleted (rho0) cells with mitochondria from AD patient platelets [125-127]. Peptide- mediated allogeneic mitochondrial delivery (PMD) is a novel technique that facilitates the incorporation of mitochondria into Parkinson’s disease (PD) rat models. Direct microinjection of Pep-1-modified allogeneic mitochondria into medial forebrain bundle (MFB) significantly improves mitochondrial uptake by neurons compared to either the injection of naïve mitochondria or xenogeneic PMD. As a result, respiration is enhanced while ROS is quenched and neuronal viability improves and the locomotor activity of PD rats is restored [128]. The successful uptake of mitochondria by target tissues will likely depend upon the amount, quality of mitochondria and route of organelle delivery.

2.3. Cellular Therapy

Cell-based therapies in the treatment of AD are being explored in many in vivo and in vitro models as a promising alternative to reverse neurodegeneration. This can be achieved by the direct replacement of injured neurons or by paracrine induction of neuronal repair [129]. Neurons and glial cells can be generated from embryonic stem cells (ESCs), neural stem cells (NSCs), neural progenitor cells (NPCs), mesenchymal stem cells (MSCs), induced pluripotent stem cells (iPSCs), induced neuronal cells (iN), and induced neuronal progenitor cells (iNPCs). MSCs, which are non-hematopoietic stem cells capable of differentiating into a multitude of cell lineages [130] are the most commonly used cells in tissue engineering and regenerative medicine because they can promote host tissue repair through several different mechanisms, including donor cell engraftment, release of cell signaling factors, and the transfer of healthy organelles to the host. Transplantation of these cells into AD animal models induces mitochondrial biogenesis and reverses cognitive defects and extends lifespan [131-133]. This nascent therapeutic technique has great promise for NDDs, including AD [134-136].

2.4. Targeting Mitochondrial Biogenesis

Impaired mitochondrial bioenergetics is a key player in the pathogenesis of AD. As a consequence, therapeutic strategies that enhance mitochondrial mass and activity may be beneficial [137-140]. Cells increase their mitochondrial mass by a process called mitochondrial biogenesis. This process involves an intricate coordination of nuclear gene expression, protein import and mtDNA transcription [141-145], including inter alia, mitochondrial transcription factor A (TFAM), peroxisome proliferator-activated receptor gamma coactivator 1-alpha (PGC- 1α) and other proteins, including nuclear respiratory factors (NRF -1 and NRF-2), uncoupling proteins (UCP2), thyroid hormone, glucocorticoid, estrogen, and estrogen-related receptors (ERR) a and b [146-148]. PGC-1α additionally regulates mitochondrial genome copy number, mitochondrial dynamics and OXPHOS [149]. Mitochondria are fluid and constantly fuse (fusion) and divide (fission) [150]. Fusion involves three GTPases; mitofusin (Mfn) 1 and Mfn2, and optic atrophy protein 1 (OPA1), while fission is mediated by GTPase dynamin-related protein 1 (Drp1) [151]. Upregulating TFAM increases mtDNA. Mitochondrial bioenergetics can be improved by extraneous manipulation of mitochondrial genome, including by the use of synthetic TFAM that can target mitochondria. Treating 3xTg-AD mice with human TFAM (hTFAM) enhances the expression of transthyretin which blocks Aβ aggregation, and reduces oxidative stress and enhances cognitive function [152, 153]. Recombinant-human TFAM (rhTFAM) improves cellular bioenergetics in mtDNA disease cultured cells [154, 155] as well as in aged lab mice where memory is improved [156, 157] and they are able to run two times longer on their rotating rods than their age matched controls [157]. The pirinixic acid derivate MH84, an in vitro dual g-secretase/proliferator activated receptor gamma (PPARγ) modulator, improves mitochondrial dysfunction in cellular and Thy-1- APP -SL mice (that harbor the Swedish and London mutation of human APP) [158]. Dumont et al., however, have shown that overexpression of PGC-1a In the Tg19959 mouse model of AD induced mitochondrial abnormalities and neuronal death [159, 160].

2.5. Targeting the Proteasome

The ubiquitin proteasome system (UPS) and mitochondrial systems are tightly interdependent (Fig. 1). Mitochondrial dysfunction and impairment of the ubiquitin proteasome system are both hallmarks of aging and implicated in the etiopathogenesis of AD. It is not clear which one triggers the vicious cycle of dysfunction [161-164]. Defective proteostasis impairs mitochondrial function. In order to restore mitochondrial function, cells initiate the mitochondrial unfolded protein response (mtUPR) [165-167], which induces mitochondrial biogenesis [167]. If this fails to maintain mitochondrial ΔΨm, mitophagy ensues to eliminate the dysfunctional mitochondria [168] and maintain a healthy population of mitochondria [150]. A viable strategy to treat NDDs is to activate the proteasome using small molecules to get rid of protein aggregates [169-171]. Betulinic acid is a triterpene natural product that selectively enhances the chymotryptic-like site of proteasome activity. Synthetic modifications have produced active analogs, suggesting a complex structure activity relationship (SAR) [172, 173]. Pyrazolone is another small molecule proteasome activator [170] that is neuroprotective in animal models of amyotrophic lateral sclerosis [174]. Another potent activator of the proteasome is the p38MAPK inhibitor PD169316, which enhances Proteolysis Targeting Chimeric (PROTAC)-mediated and ubiquitin-dependent protein gradation. This downregulates both overexpressed and endogenous α-synuclein in a bimolecular fluorescence complementation (BiFC) assay [175, 176] without any adverse effects on overall protein turnover while increasing the viability of cells overexpressing toxic α-synuclein assemblies [173]. Synthetic peptides based upon the HbYX motif are the most common class of proteasome gate openers. While their activity depends on the activator protein they are modeled after, they all increase the turnover of oxidized proteins [177-181]. Selective phosphodiesterase-4 inhibitors like rolipram, activate proteasome function, reduce aggregated tau levels, and improve cognitive performance and ameliorate the early stages of neurodegeneration in mouse models of tauopathy by increasing cAMP levels [171]. The FDA-approved drug for adult chronic myeloid leukemia, Nilotinib, clears misfolded and damaged proteins by autophagy via its enhancement of parkin levels [182]. It is in phase 2 trials for AD (NCT02947893). It is important to recognize that it is important to restore autophagic flux and not simply enhance autophagy that might merely result in accumulation of autophagosomes and undigested autolysosomes, but enhance all stages from autophagosome biogenesis, lysosome fusion and degradation of loaded autolysosomes [183, 184]. Finally, since activating Nuclear factor erythroid-derived 2 -like 2 (Nrf2) increases proteasome activity, the antioxidant 3H-1,2-dithiole-3-thione (D3T), which upregulates both 20S and 19S proteasome subunits, are promising therapeutic targets [185, 186]. As a proof of concept, genetic activation of the proteasome ameliorates the aging process and elongates lifespan in different models, including C. elegans, human fibroblasts and yeast cells [187, 188].

Fig. (1).

The ubiquitin proteasome system (UPS) and mitochondrial systems are tightly interdependent. Proteasome activation is a promising strategy to treat or prevent AD as it helps prevent the accumulation of toxic protein aggregates. (A higher resolution / colour version of this figure is available in the electronic copy of the article).

2.6. Targeting Mitophagy

Defects in mitophagy closely track with AD pathogenesis. Mitophagy is the quality control process by which damaged and inefficient mitochondria are eliminated and are regulated by mitochondria fission- and fusion- promoting proteins [189-191]. Mitophagy is crucial in neurons [192-195] which are post-mitotic, have high energy demands and strictly aerobic and are therefore particularly sensitive to mitochondrial dysfunction [190, 196-200]. They need an efficent system to eliminate damaged mitochondria and decrease ROS -induced death [201, 202]. Damaged mitochondria are eliminated through focal mitophagy that results in reduced ROS and neuroprotection [201, 202]. Urolithin A, which is produced by human gut microbiota, induces mitophagy which degrades dysfunctional mitochondria that would otherwise accumulate with age. This extends Caenorhabditis elegans lifespan and improves rodent muscle function [203]. Urolithin A can cross the blood-brain barrier (BBB) and is neuroprotective against NDDs. The aliphatic polyamine, Spermidine, induces autophagy in a SIRT1- independent manner [204] and induces the formation of mitophagosomes and decreases the aggregation of dysfunctional mitochondrial through the PINK1/Parkin pathway [205] and restores mitochondrial activity in aged cardiomyocytes [206] as well as protect against age-induced memory impairment [207]. Metformin and resveratrol, which regulate PINK1/parkin and sirtuin activating compounds (STACS) or nicotinamide adenine dinucleotide (NAD) precursors such nicotinamide riboside (NR) or nicotinamide mononucleotide (NMN), induce mitophagy [208]. PMI (P62- mediated mitophagy inducer), is a novel potent inducer of mitophagy. Its actions are independent of the PINK1/parkin pathway and do not affect the mitochondrial network or induce mitochondrial membrane potential [209]. Inactive glyceraldehyde-3-phosphate dehydrogenase (iGAPDH) induces mitophagy. iGAPDH is a molecular sensor that detects and tags damaged mitochondria as GAPDH is inactivated by mitochondrial ROS. Mitochondria-associated iGAPDH promotes the elimination of damaged mitochondria via a lysosomal-like structure, a hybrid organelle of late endosome and lysosome [162, 210]. Exogenously expressed, catalytically inactive iGAPDH eliminate damaged mitochondria [162, 210]. Neurons may be protected by modulating GAPDH to eliminate damage mitochondria or mitochondria that are producing excessive amounts of ROS.

2.7. Targeting Mitochondrial ROS

An ideal strategy for preventing and treating AD is to eliminate detrimental neuronal ROS without affecting ROS signaling cascades. Anti-oxidants targeting mitochondria potently sequester reactive oxygen intermediates (ROIs) and confer greater protection against oxidative damage than untargeted cellular antioxidants. Compounds such as CP2, a tricyclic pyrone that can cross the BBB and accumulate in the mitochondria can selectively target mitochondria-derived ROS have the potential of enhancing neuronal viability [211]. In wild-type mice, it elicits a mito-hormetic effect by mildly inhibiting complex I of the mitochondrial ETC and increasing respiratory capacity and coupling. In APP, PSEN1 and APP/PSEN1 mouse models of AD, CP2 prevents cognitive impairment and reduces Ab plaques and phosphorylated tau. Other mitochondria-targeted antioxidants such as (10-(6′-plastoquinonyl) decyltriphenyl-phosphonium) (SkQ1), MitoQ, MitoTEMPOL and MitoVitE are neuroprotective and more efficient compared to untargeted antioxidants such as 6- hydroxy-2,5,7,8-tetramethylchroman-2-carboxylic acid (Trolox) [212]. Similar antioxidants include: 4,5-dihydroxybenzene-1,3-disulfonate (Tiron), also accumulate within the mitochondria by permeabilizing the mitochondrial membrane [213] and astaxanthin is a mitochondrion-permeable antioxidant, that crosses the blood-brain barrier and effectively prevents and treats macular degeneration [214-216]. Szeto-Schiller (SS) tetrapeptides can deliver and localize anti-oxidants into the inner mitochondrial membrane with an approximate 1000- 5000- fold accumulation [217-220]. Novel XJB peptides, which consist of an electron and ROS scavenger (4-NH2-TEMPO) conjugated to the Leo-D-Phe-Pro-Val-Orn fragment of gramicidin S can specifically target the XJB peptides to mitochondria. Of these, XJB-5-131, improves mitochondrial function and enhances neuronal viability in Huntington’s disease animal models [221, 222]. Nano carriers such as the biodegradable poly-lactide-co-gylcolide (PLGA) like PLGA CoQ10 nanoparticles are an alternative vehicle to target anti-oxidants to mitochondria [221, 223].

N-acetyl-5-methoxytryptamine (Melatonin) is potently neuroprotective against oxidative stress in the brain compared to vitamins C and E at physiological concentrations [224-229]. Melatonin and its metabolites are broad-spectrum antioxidants [230-232] with therapeutic potential in aging and AD [233-236]. Melatonin also down regulates caspase-3 levels [237], that are elevated in AD brains [238] and are linked directly to neuronal apoptosis [239]. Caspase-3 activation increases β-secretase activity and Aβ production [240]. Melatonin also increases anti-apoptotic Bcl-2 expression in AD transgenic mice and ischemic brains [241], and has anti-fibrillogenic effects [242]. It also alleviates behavioral deficits associated with apoptosis and cholinergic system dysfunction [243, 244], inhibits amyloid pathology [245] and increases survival in AD transgenic mouse models [246]. Melatonin reverses the pro-fibrillogenic activity of ApoE4 and neutralizes its neurotoxic combination with Aβ [242].

2.8. Targeting Cardiolipin

Targeting cardiolipin SS-31 (elamipretide) is a mitochondrially specific small peptide that binds to cardiolipin on the inner mitochondrial membrane. Cardiolipin helps organize the components of the ETC into super complexes for more efficient OXPHOS [247]. SS-31 also prevents mitochondrial permeability transition pore formation that leads to mitochondrial swelling and apoptosis when mitochondria are stressed [248]. In preclinical models of AD, SS-31 protects against anesthesia-induced cognitive impairment and promotes mitochondrial and synaptic health [248, 249]. SS-31, which is in clinical trials for mitochondrial myopathy, Leber’s hereditary optic neuropathy, Barth syndrome and Huntington’s diseases are being repurposed for potential use in AD (stealthbt.com/clinical).

2.9. Targeting Mitochondrial Cholesterol

Mitochondrial function can be severely disrupted in cholesterol accumulation and this may contribute to the progression of AD. Cholesterol accumulation in the mitochondria reduces the fluidity of membranes [250], ATP generation [251-254] and mitochondrial glutathione (GSH) import [255-259]. Altered membrane lipids are directly linked to brain mitochondrial dysfunction [260]. There are increased levels of lysosomal cholesterol transporter Niemann-Pick type C protein 1 (NPC1) in the hippocampus and frontal cortex of patients with AD and AD-Tg mice [261]. Steroidogenic acute regulatory protein (STARD1), which modulate mitochondrial cholesterol trafficking is elevated in the pyramidal hippocampal neurons of AD patients [262, 263]. 2-Hydroxypropyl-β-cyclodextrin (HPCD) effectively lowers cholesterol through multiple mechanisms and is FDA approved. HPβCD, which is in phase 2b/3 trials for Niemann-Pick type C (NPC), reduces cholesterol accumulation defect in animal models [264] and can be administered intranasally when amalgamated with polymeric microspheres made of chitosan or sodium alginate [265]. Cytochrome P450 46A1 (CYP46A1) modulates brain cholesterol turnover by regulating its elimination. In the APP23 AD mouse model, inhibiting CYP461A1 results in Ab accumulating with extensive neuronal death. Decreasing CYP46A1 gene expression increases cholesterol concentration in normal mouse hippocampal neurons resulting in hippocampal atrophy and cognitive deficits [266]. There are ongoing preclinical tests aimed at restoring AD brain cholesterol metabolism by targeting CYP46A1 [218]. The clinical trial is anticipated to start in 2021 http://www.brainvectis.com.

2.10. Targeting Mitochondrial Membrane Potential

Ursodeoxycholic acid (UDCA) is safe and has a limited side effect profile and has been indicated for primary biliary sclerosis for over 30 years [267]. Drp1 inhibitors have been explored as a therapeutic avenue in AD. In vitro and in vivo studies in AD models of UDCA and related compound tauroursodeoxycholic acid (TUDCA) [268-271] have revealed a putative protective effect [270-273]. UDCA restores mitochondrial membrane potential in sAD and PSEN1 mutant fibroblast through its effects on Drp1. It does this without impacting both sAD and PSEN1 mutant fibroblasts via its actions on Drp1 while having no significant effect on mitochondrial morphology. Drp1 inhibitors are now explored as a therapeutic avenue in AD [274].

2.11. Epigenetic Modifiers

Epigenetic mechanisms may mediate the risk for AD [275], and molecular modulation of epigenetic mechanisms may protect against age related cognitive decline [276]. A few epigenetic compounds are in clinical trials for AD. These include Suberoylanilide hydroxamic acid (SAHA, Vorinostat) pan-histone deacetylase (HDAC) inhibitor that can induce autophagy in cardiomyocytes and mitigate ischemia/reperfusion injury if administered during reperfusion. If administered before or after ischemia, SAHA induces and autophagy mitochondrial biogenesis that mitigates mitochondrial dysfunction and ameliorates oxidative stress [277]. Vorinostat is FDA-approved for cutaneous T-cell lymphoma and is currently in phase I trial for patients with AD (NCT03056495). ORY-2001, a safe, well tolerated selective dual LSD1-MAO-B inhibitor that is capable of crossing the BBB and regulating histone methylation, has been shown to significantly improve cognition in transgenic AD models [278].

3. NON-PHARMACOLOGIC LIFESTYLE AND NUTRITIONAL INTERVENTIONS

3.1. Calorie Restriction

Calorie restriction (CR) aims to decrease caloric intake while maintaining all the essential nutrients so that there is no malnutrition. It enhances life span and prevents age-related diseases, including neurological deficits, brain atrophy, and cognitive decline [279]. CR induces mitochondrial biogenesis [280] in a NO•- mediated manner resulting in enhanced mitophagy and the production of new, more efficient mitochondria that have reduced membrane potential, produce less ROS, consume increased levels of oxygen and exhibit an improved ATP/ROS ratio - leading to decreased energy expenditure [281]. It is viable non-pharmacologic strategy to improve healthy brain aging [282].

CR neutralizes the harmful effects of ROS and oxidative damage [283-286] and slows down age-associated transcriptional changes [287]. CR induces expressions of sirtuins, such as SIRT1, SIRT3, SIRT5, and SIRT7 [288]. CR also inhibits the PI3K/AKT pathway, induces mitophagy and maintains mitochondria homeostasis [289].

3.2. Endurance Exercise

Endurance exercise (EE) may help retard the neurodegenerative process in AD. In AD mouse models, EE increases mtDNA repair capacity in the hippocampus and activates mitochondrial uncoupling proteins (UCP), which regulates mitochondrial proliferation and control the production of mitochondrial-derived ROS and activates autophagy. Ultimately, EE abrogates the deleterious effects of free radicals, the production of total cholesterol, and insulin resistance while enhancing vascularization and angiogenesis, and improving glucose metabolism and neurotrophic functions that result in neurogenesis and synaptogenesis. These improve memory and cognitive function [290-292].

Endurance exercise (EE) increases mitochondrial biogenesis in most brain regions [293] and may help retard the neurodegenerative process in AD. The induction of continuous oxidative stress induces mitohormesis- a series of counteractive mechanisms that enhances mitochondrial health and mitigates ROS-induced neurotoxicity [281, 294]. This is especially crucial in the hippocampus, which is particularly sensitive to oxidative stress [295]. Simultaneously, EE improves mtDNA repair capacity in the mouse hippocampus and activates mitochondrial uncoupling proteins (UCP) that regulate mitochondrial proliferation [296] and production of mitochondrial-derived ROS [297, 298]. At the same time EE alters mitochondrial proteostasis and mitochondrial unfolded protein response (UPRmt) markers and stimulates the OXPHOS component from mtDNA in (neuropeptide Y) NPY-producing neurons in the lateral hypothalamus of mice [299]. Since EE has been shown to be helpful in retarding the progress of Parkinson’s disease (PD) [300], it may be a promising therapeutic option for AD [301].

3.3. Ketogenic Diet

The ketogenic diet (KD) is a high-fat and low-carbohydrate diet aimed at reducing carbohydrate to ≤10% of consumed energy in order to shift to the utilization of ketone bodies (KBs) from fatty acids (FAs) for energy [302]. In a 24 h period, the adult brain utilizes between 100-120 g of glucose which is ~ 20% of its basal metabolism [303]. The KD aims to supply sufficient protein for growth and development but not enough carbohydrates for the metabolic requirements [304, 305]. KD is biochemical fasting [306], which promotes the utilization of KBs instead of glucose, as the main fuel source by organs, including the CNS [307]. Ketogenic diet benefits AD patients [308]. While the neuroprotective mechanism hasn’t been fully elucidated, it enhances neuronal mitochondrial biogenesis and function [309, 310]. Ketone bodies (KBs) prevent the entry of Ab into mitochondria, thus preserving respiratory chain function, and improving cognition [311] by abrogating the bioenergetic deficit in AD brains [312]. Alternatively, by improving mitochondrial function, Ab production is decreased while the soluble APPa production is enhanced and this promotes neurite growth by binding to p75 receptor of BDNF [313]. In animal studies and clinical trials, KD induced stabilization of synaptic functions due to enhanced mitochondrial biogenesis [314], reduced ROS production, and enhanced cellular bioenergetics [315]. In rat cultured hippocampal neurons, KB protects against Ab toxicity [316, 317]. In AD and aging mouse models, motor function and cognition are improved by KD, while human studies show improved cognitive outcomes (global cognition, memory and executive functions) regardless of the severity of cognitive impairments previously detected [318]. 2-deoxy-D- glucose (2-DG) stimulated ketogenesis improves mitochondrial bioenergetics and delays progression of AD by reducing both amyloid precursor protein and Aβ oligomers [319]. A new experimental drug J147 has been shown to be effective against AD and aging in mouse models of accelerated aging [320, 321]. It is currently under Phase 1 clinical trial. A J147 derivative called CAD-31 enhances the use of free fatty acids for energy production by shifting of the metabolic profile of fatty acids toward the production of ketone bodies [321]. It targets mitochondrial ATP synthase. However, the best KD treatment outcomes are expected in the pre-symptomatic stages of AD.

CONCLUSION

AD is a complex, mostly sporadic age-dependent disease that is becoming increasingly prevalent, partly because the global population and average lifespan continue to increase. With only symptomatic treatments currently available, it presents a major threat to human health and is of great socioeconomic concern. Mitochondrial impairments ranging from mtDNA mutations, epigenetic modification of mtDNA, to oxidative stress, altered gene expression, impaired mitobiogenesis, altered protein turnover and changed organelle dynamics (fission and fusion) may drive the neurodegenerative process in AD. We discuss potential therapeutic approaches, including repurposed drugs, epigenetic modifiers, and lifestyle changes that target each level of deficiency and lead to the development of effective therapy (Fig. 2).

Fig. (2).

Schematic illustration of mitochondrial impairments observed in AD and the potential therapeutic approaches for each level of deficiency. The red arrow depicts the possible causes of mitochondrial impairment, while potential therapeutic approaches are depicted in green. (A higher resolution / colour version of this figure is available in the electronic copy of the article).

LIST OF ABBREVIATIONS

αKGDH

α-Ketoglutarate Dehydrogenase

Aβ

Amyloid Beta

ABCA

ATP-Binding Cassette Subfamily A

AD

Alzheimer’s Disease

APP

Amyloid Precursor Protein

BBB

Blood-Brain Barrier

BIN1

Bridging Integrator 1

CLU

Clusterin

CNS

Central Nervous System

COX

Cytochrome C Oxidase

CR

Calorie Restriction

CR1

Complement Component (3b/4b) Receptor 1

CYP46A1

Cytochrome P450 46A1

Drp1

Dynamin-related Protein 1

EE

Endurance Exercise

ER

Endoplasmic Reticulum

ERR

Estrogen-related Receptors

ESC

Embryonic Stem Cells

fAD

Early-onset Familial Alzheimer’s Disease

GWAS

Genome-wide Association Studies

HDAC

Histone Deacetylase

hTFAM

Human TFAM

iN

Induced Neuronal Cells

iPSCs

Induced Pluripotent Stem Cells

KD

Ketogenic Diet

MAMs

Mitochondrial Associated Membranes

Mfn

Mitofusin

MSCs

Mesenchymal Stem Cells

mtDNA

Mitochondrial DNA

MTHFD1

Methylenetetrahydrofolate Dehydrogenase 1

mtROS

Mitochondrial Reactive Oxygen Species

mtUPR

Mitochondrial Unfolded Protein Response

mtZFNs

Mitochondrially Targeted Zinc-finger Nucleases

NDD

Neurodegenerative Diseases

NMDA

N-Methyl-D-aspartate

NPCs

Neural Progenitor Cells

NPC1

Niemann-Pick Type C Protein 1

NRF

Nuclear Respiratory Factors

NSCs

Neural Stem Cells

OPA1

Optic Atrophy Protein 1

OXPHOS

Oxidative Phosphorylation

PD

Parkinson’s Disease

PDHC

Pyruvate Dehydrogenase Complex

PGC-1α

Peroxisome Proliferator-activated Receptor Gamma Coactivator 1-Alpha

PICALM

PI-Binding Clathrin Assembly Protein

PSEN1

Presenilin 1

PSEN2

Presenilin 2

rhTFAM

Recombinant-Human TFAM

ROI

Reactive Oxygen Intermediates

ROS

Reactive Oxygen Species

sAD

Late-onset Sporadic Alzheimer’s Disease

SAHA

Suberoylanilide Hydroxamic Acid

SORL1

Sortilin-Related Receptor-1

STARD1

Steroidogenic Acute Regulatory Protein 1

T2DM

Type 2 Diabetes Mellitus

TFAM

Mitochondrial Transcription Factor A

TREM2

Triggering Receptor Expressed on Myeloid Cells 2

TUDCA

Tauroursodeoxycholic Acid

UCP

Uncoupling Proteins

UDCA

Ursodeoxycholic Acid

UPS

Ubiquitin Proteasome System

AUTHOR'S CONTRIBUTION

All authors listed have made a substantial, direct and intellectual contribution to the work, and approved it for publication.

CONSENT FOR PUBLICATION

Not applicable.

FUNDING

This work was supported by the European Regional Development Fund- Project ENOCH 750 (No. CZ.02.1.01/0.0/0.0/16_019/0000868).

CONFLICT OF INTEREST

The authors declare no conflict of interest, financial or otherwise.