key: cord-0028078-p7rl2wlg
authors: Trushina, Eugenia; Trushin, Sergey; Hasan, Md Fayad
title: Mitochondrial complex I as a therapeutic target for Alzheimer's disease
date: 2021-11-09
journal: Acta Pharm Sin B
DOI: 10.1016/j.apsb.2021.11.003
sha: e3fd988aa3d77f12e504cac663b0d3528845fb01
doc_id: 28078
cord_uid: p7rl2wlg

Alzheimer's disease (AD), the most prominent form of dementia in the elderly, has no cure. Strategies focused on the reduction of amyloid beta or hyperphosphorylated Tau protein have largely failed in clinical trials. Novel therapeutic targets and strategies are urgently needed. Emerging data suggest that in response to environmental stress, mitochondria initiate an integrated stress response (ISR) shown to be beneficial for healthy aging and neuroprotection. Here, we review data that implicate mitochondrial electron transport complexes involved in oxidative phosphorylation as a hub for small molecule-targeted therapeutics that could induce beneficial mitochondrial ISR. Specifically, partial inhibition of mitochondrial complex I has been exploited as a novel strategy for multiple human conditions, including AD, with several small molecules being tested in clinical trials. We discuss current understanding of the molecular mechanisms involved in this counterintuitive approach. Since this strategy has also been shown to enhance health and life span, the development of safe and efficacious complex I inhibitors could promote healthy aging, delaying the onset of age-related neurodegenerative diseases.

Alzheimer's disease (AD) is a common neurodegenerative disorder in the elderly without a cure 1 . Common hallmarks of AD include extracellular plaques formed by amyloid beta (Ab) peptides, neurofibrillary tangles comprised of hyper-phosphorylated Tau protein (pTau), reactive microgliosis, dystrophic neuritis, and loss of neurons and synapses 2 . However, failure of recent clinical trials focused on the prevention of Ab and pTau production or their clearance questions the amyloid cascade hypothesis and the role of Ab and pTau in the underlying disease mechanism. The outcomes of the most recent studies conducted using advanced biochemistry and multi omics systems biology approaches in well characterized cohorts of AD patients, patient-derived cells and tissues suggest a multifactorial nature of AD where several mechanisms on a whole organismal level become individually or synergistically affected in a disease-stageand sex-specific manner 3e5 . These pathways include reduced glucose uptake and utilization, insulin resistance, altered autophagy and proteostasis, increased inflammation and oxidative stress, and mitochondrial dysfunction 6e8 . Conclusive evidence demonstrating that brain hypometabolism precedes clinical presentation and the development of Ab aggregates provided a justification for interventions focused on metabolism and mitochondrial function as potential disease-modifying strategies that could block the disease progression 9 . However, most treatments aimed at boosting mitochondrial function or reducing the pathology associated with increased production of reactive oxygen species (ROS) have failed clinical trials 10, 11 . Unexpectedly, partial reduction of the activity of the complexes involved in the oxidative phosphorylation (OXPHOS) and electron transport chain (ETC) machinery using genetic or pharmacological down modulation approaches has been shown to provide significant health benefits, improving mitochondrial function and cellular energetics in multiple model systems in vitro and in vivo. In particular, partial inhibition of mitochondrial complex I using small molecules has emerged as a therapeutic strategy for multiple human conditions, including cancer and neurodegenerative diseases. This counterintuitive strategy has been shown to increase longevity and health span, which ultimately could delay the onset of neurodegenerative diseases of aging, such as AD 9 . Indeed, the induction of mild energetic stress via partial complex I inhibition with subsequent mitochondria-mediated stress response may increase resilience to the greater stress associated with aging and ensure that mechanisms found instrumental for protection against AD, including inflammation, synaptic function, proteostasis, mitochondrial dynamics and function, and oxidative stress, remain in control 12, 13 . Below, we discuss the current understanding of the neuroprotective mechanisms behind complex I inhibition with respect to mitochondrial signaling via integrated stress response (ISR) and progress in the development of safe and efficacious partial complex I inhibitors.

Most of the energy required for cellular functions is produced by mitochondria (Fig. 1) . These organelles are abundant, occupying up to 25% of the cytoplasmic volume. The mitochondrion is the only cellular organelle other than the nucleus that has its own DNA (mtDNA) and transcriptional and translational machinery (Fig. 1A) . These features, together with the unusual dynamics of mitochondrial division and fusion (reminiscent of bacteria), have led to the theory of an ancient endosymbiosis of a nucleated cell and an aerobic prokaryote 14 . Such cooperation provides the host with significantly increased energy supply, making mitochondria a "power plant" of the cell, while mitochondria enjoy the protection and resources of the host, including the outsourcing of the most of protein synthesis essential for their function. Successful symbiotic integration required the development of a robust communication system. The extensive arsenal of signaling molecules utilized by mitochondria for intracellular communication is discussed below.

Cell populations in the brain are diverse, and each cell type has distinctive energy demands and metabolic profiles. The conventional view is that neurons depend on energy produced by mitochondria via OXPHOS 15 . In OXPHOS, a series of metabolic reactions leads to the oxidation of glucose or its metabolites, such as pyruvate and lactate, to produce energy in the form of adenosine triphosphate (ATP) (Fig. 1B) 16 . OXPHOS is the most efficient metabolic pathway, producing approximately 36 molecules of ATP per one molecule of glucose compared to 2 ATP molecules produced during glycolysis, a cytoplasmic process that also uses glucose but does not require mitochondria 17 . Mitochondrial morphology is essential to maintaining OXPHOS. The organelle has two membrane compartments (Fig. 1A) . The outer membrane delimits the organelle and allows the passage of small molecules and ions to maintain mitochondrial homeostasis. An inner membrane consists of multiple folds called crista and defines the mitochondrial matrix as a closed compartment. OXPHOS machinery is located at the inner mitochondrial membrane, while the tricarboxylic acid (TCA) cycle that produces essential components to power OXPHOS takes place in the matrix.

During OXPHOS, substrates such as nicotinamide adenine dinucleotide (NADH) or flavin adenine dinucleotide (FADH 2 ) produced in the TCA cycle donate electrons that are transferred through the mitochondrial ETC via a series of redox reactions coupled to a final phosphorylation of adenosine diphosphate (ADP) to produce ATP, CO 2 and water. This process requires oxygen, and oxygen consumption could be measured using oxygen electrodes or a Seahorse Extracellular Flux Analyzer to inform on the OXPHOS activity 18 . The ETC (Fig. 1B) includes four protein complexes: NADH dehydrogenase (complex I), succinate dehydrogenase (complex II), cytochrome c oxidoreductase (complex III), and cytochrome c oxidase (complex IV) 19 . Electrons moving through the ETC promote the translocation of protons (H þ ) from the matrix to the intermembrane space, establishing an electrochemical gradient or proton-motive force (PMF). The energy generated from the PMF is used to phosphorylate ADP to ATP via ATP synthase (complex V) (Fig. 1B) . The PMF is regulated by the electrical potential difference established between the cytoplasm and the matrix, known as mitochondrial membrane potential (Djm), and the proton gradient (DpH) across the inner mitochondrial membrane 20 . Under normal physiological conditions, PMF is dominated by Djm, which accounts for over 70% of its potential 21, 22 . However, maintaining the PMF at high potential can lead to a breakdown of the membrane with subsequent formation of ROS, including hydrogen peroxide (H 2 O 2 ) and superoxide (O 2 e ) 23 .

Hence, the PMF buildup during OXPHOS is counterbalanced by ATP synthesis, during which protons re-enter the matrix diminishing the PMF. Under steady state conditions, the rate of electron transport equilibrates with proton translocation resulting in sufficient energy production and minimal generation of ROS 19 . During OXPHOS, H 2 O 2 and O 2 e are produced as byproducts in mitochondria, primarily by complexes I and III, and are sequestered by the antioxidant enzymes, including superoxide dismutases, glutathione peroxidase, glutaredoxins, thioredoxins and catalases 7 . Under disease conditions, this balance may be altered, leading to excessive ROS production and cellular damage. The balance amongst mitochondrial function, ROS production and the antioxidant defense is essential for normal function, since neuronal cells comprise most (80%e 90%) of the energy demand of the brain and primarily depend on OXPHOS 24 . Thus, levels of ROS, energy metabolites, metabolites of the TCA cycle, and Djm are tightly linked to mitochondrial function, so changes in concentrations of all these metabolites could be used for intracellular signaling. In essence mitochondria serve as signaling organelles. The NADH and FADH 2 are converted to NAD þ (complex I) and FAD (complex II), respectively, with H 2 O formed (complex IV) as a biproduct. This electron transfer causes protons to flow from mitochondrial matrix to intermembrane space, creating an outward gradient of positively charged protons. The inner mitochondrial membrane bound F 0 subunit of complex V (ATP synthase) uses this electrochemical gradient to rotate causing conformational changes to F 1 subunits that convert ADP to ATP. Changes in the concentrations of all these metabolites could be used for intracellular communication. ADP indicates adenosine diphosphate; ATP, adenosine triphosphate; FADH 2 , flavin adenine dinucleotide; NAD þ , nicotinamide adenine dinucleotide; NADH, nicotinamide adenine dinucleotide; OXPHOS, oxidative phosphorylation; TCA, tricarboxylic acid.

Mitochondria have long been recognized as central to ATP production, calcium buffering, and initiation of the apoptosis. In recent years, it became apparent that mitochondrial involvement in regulating cellular fate is more complex 25, 26 . Mitochondria communicate with the rest of the cell by releasing metabolites, mtDNA, and ROS, by changing their size and motility and by interacting with other subcellular organelles (Fig. 2) . For example, fluctuations in the TCA cycle metabolites (citrate, a-ketoglutarate, succinate and fumarate) induce epigenetic modifications, including nuclear DNA methylation, histone acetylation, and protein hydroxylation and acetylation 27, 28 . Changes in the levels of nicotinamide adenine dinucleotide (NAD þ ), a direct product of mitochondrial complex I function, affect the activity of sirtuins, the essential regulators of multiple cellular functions linked to improved mitochondrial function, increased health span and longevity 29e32 . Alterations in ROS levels control hypoxic responses, immunity, and stem cell function 33 . Recent findings demonstrate that mitochondrial ROS is an essential component of signaling that mediates antioxidant (redox) balance and stabilizes hypoxiainduced factor 1 a (HIF-1a) 34 , an important mediator of life-span extension linked to mitochondrial ISR 35 . Recent studies also identified mitochondrial ROS signaling as a mitohormetic process where an increase in sublethal levels of ROS could predispose cells to a better response to increased oxidative stress in the future 36 . Furthermore, mitochondria regulate immune response by releasing mtDNA and through the peptides (e.g., humanin and mitochondrial open-reading-frame of the twelve S rRNA-C, MOTS-c) encoded by mtDNA 37, 38 . Changes in mitochondrial dynamics (fission, fusion, axonal trafficking, biogenesis and mitophagy) are also important determinants of mitochondrial function and quality control 39, 40 . Multiple mechanisms are in place to respond to abnormal mitochondrial dynamics to ensure organelle preservation, including mitochondrial unfolded protein response (mtUPR, a process to maintain monoconidial proteostasis) 41 , enhanced biogenesis (a mechanism to produce new mitochondria), and mitophagy (a process that removes damaged organelles) 42 . Changes in Djm play a key role in mitochondrial homeostasis signaling for selective elimination of damaged organelles through mitophagy by recruiting PTEN-induced kinase 1 (PINK1) and Parkin proteins to the mitochondrial membrane 43 . It is also a driving force for the translocation of ions and proteins essential for mitochondrial function. Mitochondria interact with other organelles, including the endoplasmic reticulum (ER), to modulate lipid homeostasis, immune response, and cell death 44, 45 . Finally, changes in ATP levels associated with either increased energy utilization or reduced mitochondrial capacity led to an increase in the cellular adenosine monophosphate (AMP)/ATP ratio, which activates AMP-activated protein kinase (AMPK), a master regulator of cellular energy homeostasis 46e49 . Active AMPK initiates a robust signaling cascade to restore energy balance. This dynamic process involves changes in lipid and glucose metabolism, mitochondrial dynamics and biogenesis, autophagy, and protein synthesis. Directly relevant to aging and neurodegenerative diseases is the AMPK-dependent reduction of inflammation and increase in levels of sirtuins, signaling molecules that regulate vast networks of metabolic and non-metabolic enzymes essential for healthy aging 50e52 . Numerous pathways affected by AMPK have been shown to be neuroprotective brining attention to AMPK as a drug target for neurodegenerative diseases 46 . However, the development of direct AMPK activators has been proven difficult given the delicate balance required for cellular energy homeostasis 46 . Nevertheless, it is now broadly accepted that indirect AMPK activation via exercise or caloric restriction is associated with increased health span and slowing down the progression of age-related neurodegenerative diseases 53 . The availability of such a robust signaling arsenal allows mitochondria to successfully adapt to environmental changes, ensuring sustained energy production and cell survival.

The most common mitochondrial stressors that induce mtUPR and ISR include fluctuations in energy sources, mtDNA mutations, changes in Djm, Ca 2þ and other ions, increased ROS production, and inhibition of the OXPHOS complexes 26, 42, 54 . The mechanisms of the mtUPR and the ISR studied in Caenorhabditis elegans and mammalian cells converged on changes in gene expression of chaperones, proteases, detoxification enzymes, and the engagement of mediators of metabolic and epigenetic reprograming, including activation of AMPK (Fig. 2) . In C. elegans, initiation of this signaling cascade extended life span 55 with epigenetic modification transmitted over four generations through histone H3K4 methylation 56 . Data in mammalian cells suggest that activation of ISR depends both on the nature of the mitochondrial stressor and the metabolic state of the cell 57 . Multiple studies conducted to date indicate that mild mitochondrial stress associated with the inhibition of OXPHOS complexes could induce an adaptive stress response that promotes health and longevity and delays the development of neurodegenerative diseases. Early evidence from studies in model organisms have demonstrated that mutations that decrease the activity of the mitochondrial respiratory chain resulted in a 20%e300% increase in the mean adult life span in C. elegans 58e63 . Similar effects on longevity were achieved with RNA interference (RNAi) reduction in expression of the ETC components. The complete ablation of major ETC subunits resulted in severe phenotype and shorter lifespan indicating that only mild decrease in ETC activity was beneficial 64 . In flies, the RNAi of five genes encoding components of mitochondrial respiratory complexes I, III, IV, and V resulted in increased life span 65 . This phenomenon was not associated with altered assembly of respiratory complexes or reduced ATP production. Targeted RNAi of two complex I genes in adult tissues or in neurons alone was sufficient to extend Drosophila melanogaster life span 65 .

In mice, decreased expression of proteins involved in the ETC, especially the matrix arm subunits of complex I, increased longevity by 30% and was associated with improved complex I assembly, higher complex I-linked state 3 respiration and decreased ROS production 66, 67 . Partial inhibition of complex IV and cytochrome c oxidase activity not only increased longevity in mice but also protected from neurodegeneration 68 . The severe deficiency in complex IV or mild deficiency in complex III expression in neurons resulted in a reduction of ROS and Ab plaques in the APP/PS1 mouse model of AD 69 . Similarly, the depletion of mtDNA also led to a decrease in plaque accumulation in the same AD mouse model 70 . Inhibition of complex V has been linked to mitohormetic signaling, which increased neuronal survival in response to toxic agents in vitro and in vivo where mechanistic pathways converged on AMPK and nuclear factor kappa-light-chain-enhancer of activated B cells (NF-kB) 71 . Furthermore, the uncoupling of OXPHOS using molecules that depolarize mitochondria causing an ATP decrease and activation of AMPK has been considered as a therapeutic strategy for aging, obesity, neurodegeneration, and cancer 72e77 . Despite showing an improvement of mitochondrial function and oxidative metabolism via adaptive stress response, the OXPHOS uncouplers had multiple off-target effects and toxicity, which limited their clinical use.

Longitudinal RNA sequencing analysis identified mitochondrial complex I as a hub in a module of genes whose expression was negatively correlated with lifespan in Nothobranchius furzeri, an African turquoise killifish 78 . Partial pharmacological inhibition of complex I with picomolar concentrations of the small molecule rotenone reversed aging-related regulation of gene expression rejuvenating the transcriptome and increasing life span in N. furzeri by 15% 78 . This is particularly interesting since high concentrations of rotenone are devastating for the organism due to a high ROS production 79 . In humans, data generated in a cohort of 2200 ultranonagenarians (and an equal number of controls) have shown that mutations in subunits of complex I that resulted in partial loss of its activity had a beneficial effect on longevity, while the simultaneous presence of mutations in complexes I and III or in complexes I and V appeared to be detrimental 80 . Thus, the beneficial adaptive stress response could be induced by multiple mitochondria-targeted stressors, including inhibition of OXPHOS complexes, mtDNA depletion, and mitochondrial uncoupling, among others. However, clinical translation in most cases is impeded by the lack of selectivity, specificity, and deleterious side effects.

Molecular mechanisms linked to life-extending interventions associated with mild inhibition of OXPHOS across species included adaptive response to energetic stress via activation of AMPK 81 . Additional important outcomes involved protection against oxidative stress that was attributed to the decreased rate of OXPHOS leading to overall lower production of ROS, AMPKinduced activation of the nuclear factor E2-related factor 2 (NRF2) signaling pathway, and mitohormetic response where sublethal ROS production associated with the ETC inhibition increased antioxidant defense via retrograde ROS signaling 36,82e84 . Furthermore, AMPK activation enhanced autophagy mediating the removal of damaged organelles and misfolded proteins to improve cellular proteostasis 85 , while increasing the production of "young" mitochondria via biogenesis. Importantly, these mechanisms overlap with the outcomes of nonpharmacological interventions, such as exercise and caloric restriction, known to reduce oxidative damage and inflammation and improve health, life span, and cognitive function 86e88 .

While inhibition of mitochondrial ETC to achieve healthy aging and prevent neurodegeneration appears counterintuitive, broad application of metformin (1,1-dimethylbiguanide), an inexpensive U.S. Food and Drug Administration (FDA)-approved drug to treat type II diabetes mellitus (T2DM), supports the feasibility of such an approach in humans. Metformin is a natural product derived from the medicinal herb 'goat's-rue', Galega officinalis. It has a robust safety record having been used in herbal medicine since medieval times 89 . Metformin became the most prescribed antidiabetic drug after the results of a prospective study conducted in overweight patients with T2DM with a median follow up of over 10 years where blood glucose control with metformin reduced the incidence of diabetes-related endpoints and all-cause mortality 90 . Metformin exerts its glucose-lowering effect by inhibiting hepatic gluconeogenesis and opposing the action of glucagon 89, 91, 92 . Among other multiple targets, metformin could inhibit mitochondrial complex I to result in defective cyclic AMP and protein kinase A signaling in response to glucagon and the stimulation of AMPK 93 . Metformin can cross the blood-brain barrier (BBB) and have specific effects on the central nervous system. Biological, clinical, and epidemiological data suggest that T2DM increases risk of mild cognitive impairment (MCI), vascular dementia and AD. Clinical trials have found that application of antidiabetic drugs including metformin protected against cognitive decline in patients with MCI and AD, improving executive functioning, learning, memory, and attention 94e96 (Table 1) . These antidiabetic drugs positively affected mitochondrial and synaptic function, reduced neuroinflammation, and improved brain metabolism 97 . Interestingly, a recent systematic review reported that metformin reduced mortality and diseases of aging (cardiovascular disease and cancer) in patients who did not have diabetes, demonstrating that the effect of metformin on health span is independent of its antidiabetic properties 98 . Thus, metformin appears to mimic mechanisms involved in caloric restriction and exercise shown to slow the aging process, improve memory, and reduce oxidative stress 99e104 . However, a few reports based on data generated in experimental animal models and collected in studies in diabetic patients suggest that metformin could increase amyloid accumulation and risk of developing AD 105e109 . These effects have been linked to overactivation of AMPK and vitamin B 12 deficiency potentiated by metformin, which contribute to cognitive impairment 110, 111 . Furthermore, it remains uncertain to what extent 113 . Thus, while increasing evidence supports strong therapeutic potential for metformin as a neuroprotective therapy for neurodegenerative diseases of aging, additional safety and feasibility studies and mechanistic studies aimed at evaluating the contribution of complex I inhibition in different tissues to the drug efficacy are needed to identify potential risk factors, windows of therapeutic opportunity, and regimens 114, 115 .

Recent studies have identified other ETC inhibitors with a wide range of biological properties, including antioxidant, anticancer, anti-inflammatory, and cardio-and neuroprotective effects 116e120 . Resveratrol, a promising therapeutic compound that activates sirtuins 121 , has been shown to reduce the activity of mitochondrial complexes I, III, and V 122e124 . Similar to metformin, resveratrol stimulates key signaling pathways, including antioxidant defenses, reduction of inflammation via inhibiting NF-kB signaling, and AMPK activation, leading to improved mitochondrial function and biogenesis through sirtuin 1/AMPK/peroxisome proliferatoractivated receptor gamma coactivator 1 alpha (PGC1a) pathway and vitagenes, which prevent the deleterious effects triggered by oxidative stress 125, 126 . Results of studies conducted in in vitro and in vivo models of AD provided evidence that resveratrol normalizes cholinergic neurotransmission and brain-derived neurotrophic factor (BDNF) expression, reduces oxidative stress, promotes Ab peptide clearance and anti-amyloidogenic cleavage of APP, and reduces neuronal apoptosis 127 . Application of resveratrol was also beneficial in models of metabolic disorders, Huntington's disease (HD), and Parkinson's disease (PD), amyotrophic lateral sclerosis, stroke, and alcohol-induced neurodegenerative disorders 128 . However, the use of resveratrol in humans has been challenging, with limited bioavailability, pronounced adverse side effects, and inconsistent results were reported in healthy and unhealthy participants of clinical trials 129 . While data generated to date strongly support the importance of resveratrol for human health, the design of better analogs with greater potency, solubility, and bioavailability are needed 122 . Taken together, these data demonstrate that mild inhibition of the OXPHOS complexes engages a multifaceted mitochondria-mediated signaling cascade that improves multiple mechanisms of AD pathogenesis, including inflammation, mitochondrial dysfunction, abnormal energy and lipid homeostasis, the ER and oxidative stress making this therapeutic approach appealing 7, 9, 40, 130 .

Partial inhibition of complex I with small molecules emerged as a promising strategy to induce beneficial ISR. Table 1 lists complex I inhibitors that are in clinical trials for various human conditions, including T2DM, cancers, metabolic disorder, obesity, inflammatory and infectious diseases. Only metformin, resveratrol, berberine, and epigallocatechin-3-gallate were trialed in a limited number of studies for neurodegenerative diseases, including AD, HD, MCI, multiple sclerosis, and Down syndrome. Metformin improved cognitive function in patients with amnestic MCI, while resveratrol, berberine and epigallocatechin-3-gallate did not show statistically significant improvements in cognitive performance in patients with AD, HD, or MCI. While all four complex I inhibitors penetrate the BBB, the therapeutic effect of resveratrol, berberine and epigallocatechin-3-gallate was limited, probably due to a poor stability, short half-life, and a very low bioavailability (<1%) in contrast to metformin, which is stable and has better bioavailability. Therefore, modifications of current complex I inhibitors or the development of new small molecules with improved druglike properties and bioavailability are needed to increase therapeutic efficacy for neurodegenerative diseases.

Complex I is the largest (970 kDa) multisubunit complex of the ETC with 14 central subunits involved in the oxidation of NADH to NAD þ at the flavin mononucleotide domain (FMN), transfer of the electrons along eight canonical iron-sulfur clusters to ubiquinone and its reduction, and proton pumping (Fig. 3) 131 . There are an additional 31 accessory subunits that are not directly associated with energy production 131 . Structures of bacterial and mammalian complex I have been determined by X-ray crystallography and cryogenic electron microscopy (cryo-EM) at high resolution, providing new insights into its assembly, proton-pumping machinery, the enzyme's catalytic mechanism, and dysfunctions associated with disease-causing mutations 131e133 . Complex I contributes significantly to the formation of ROS 134 . Interestingly, there are more than 60 complex I inhibitors that have a differential effect on the enzyme kinetics or ROS production, where molecules including rotenone, piericidin A, and rolliniastatin 1 and 2 increase ROS, while inhibitors such as stigmatellin, mucidin, capsaicin, and coenzyme Q2 prevent ROS formation 134 . Similarly, some mutations in complex I could preserve the conversion of NADH to NAD þ and, therefore, complex I activity while completely blocking pathological ROS production 135 . These data suggest that it is possible to develop safe and efficacious complex I inhibitors that are selective to the target and do not induce mitochondrial dysfunction associated with increased ROS production. These observations help to address concerns associated with the development of complex I inhibitors for chronic use in the elderly population to treat/delay the development of AD. For example, it is well established that mitochondrial complex I inhibitors such as 1-methyl 4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) or 3-nitropropionic acid could be used to mimic PD and HD, respectively. Indeed, the metabolic product of MPTP, MPP þ , binds complex I at two sites and induces significant ROS formation. However, the pathophysiology of PD involves other mechanisms that affect mitochondrial function, including altered mitophagy and biogenesis, where the involvement of complex I inhibition has been called into question 136, 137 . Moreover, a recent study demonstrated that a-synuclein aggregation was less common in complex I deficient neurons in the substantia nigra, implying that partial complex I inhibition in PD may be a beneficial compensatory mechanism for reducing increased ROS production 138 . Similarly, the involvement of complex I inhibition in HD pathogenesis may not be a primary mechanism underlying the disease pathogenesis 139, 140 . Our data generated in primary mouse neurons from HD mice 118 and in a bacterial artificial chromosome (BAC)-mediated transgenic mouse model of HD (unpublished observations) demonstrate that partial reduction of complex I activity improves multiple mechanisms affected by the expression of mutant huntingtin protein. Thus, a mounting body of evidence supports the feasibility of targeting complex I as a therapeutic strategy for neurodegenerative diseases. However, to develop safe and efficacious complex I inhibitors, it is imperative to determine the binding site, extent of inhibition of complex I activity, selectivity, and levels of ROS production. We recently identified a small molecule tricyclic pyrone compound (CP2) that penetrates the BBB and accumulates in mitochondria where it mildly inhibits the activity of complex I 117, 141, 142 . CP2 is bioavailable, has low toxicity in vitro and in vivo, and has good drug-like properties and safety profile, demonstrating the lack of off-target activity against human receptors and ion channels 141e144 . The genome-wide associations study using 196 human lymphoblastoid cell lines from healthy individuals with diverse age, sex and racial background demonstrated the safety of CP2 application at therapeutic doses 144 . The effect of CP2 on the activity of each of the respiratory complexes examined using enzymatic assays and mitochondria isolated from the mouse or postmortem human brain confirmed selective and specific inhibition of complex I 117, 142 . The bioenergetics studies conducted in mouse primary cortical neurons using a Seahorse Extracellular Flux Analyzer (Agilent Technologies, Inc.) demonstrated that CP2 improved cellular bioenergetics augmenting spare respiratory capacity, an indicator of mitochondrial ability to produce energy under conditions of increased workload or stress, which is essential for long-term survival and function 145 . Similarly, CP2 increased mitochondrial respiratory control ratio and reduced proton leak, suggesting better coupling efficiency of the neuronal ETC, greater bioenergetic reserve, and enhanced ability to withstand stress. In vivo efficacy of chronic CP2 administration was examined in independent cohorts of male and female mice that express mutant human amyloid precursor protein (APP), mutant human presenilin 1 protein (PS1), mutant APP and PS1 (APP/PS1) or mutant APP, PS1 and human Tau protein (3xTgAD) starting in utero for 14 months, at pre-or symptomatic stages of the disease 141e143 . In all studies, chronic CP2 treatment did not induce toxicity or affect development. In all treatment paradigms, animals were allowed to have CP2 in drinking water ad lib. Remarkably, in all treatment groups, CP2 improved energy homeostasis in the brain and periphery (glucose uptake and utilization, glucose tolerance, and insulin resistance), synaptic activity, long-term potentiation, dendritic spine maturation, cognitive function and proteostasis (reduced Ab and pTau levels), and reduced oxidative stress and inflammation in the brain and periphery, ultimately blocking the ongoing neurodegeneration ( Fig. 4) 142, 143 . We observed increased levels of ATP consistent with improvement of brain energy homeostasis and reduced levels of ceramides, indicative of the release of the ER stress prominent in patients with AD 142 . Therapeutic efficacy was monitored using translational in vivo biomarker fluorodeoxyglucose-positron emission tomography (FDG-PET), phosphorus-31 magnetic resonance imaging ( 31 P MRI), and blood-based metabolomics. Interestingly, this treatment augmented mitochondrial dynamics and function, including restoration of axonal trafficking in neurons from CP2-treated PS1 and APP/PS1 mice 117 . While CP2 was demonstrated to be selective and specific complex I inhibitor that lacks the off-target activities 142, 144 , it was shown to interfere with (Tyr144, His95, His91) . The FMN and ubiquinone are the main sites of reactive oxygen species (ROS) production. The membrane arm comprises three antiporter type subunits with discontinuous helices (ND5, marine; ND4, cyan; ND2, pink) corresponding to three potential proton translocation sites (black arrows). In the proximal part of the membrane arm (PP module) the p-bulge helix of ND6 (orange) and the discontinuous helix of ND1 (red) are highlighted. Residues constituting a fourth putative proton pathway (dashed arrow) are found in subunits ND2 and ND4L. In the center of the membrane arm a series of protonable residues (basic, blue; acidic, red) extends from subunit ND5 to subunit ND1 and terminates below the ubiquinone reduction site with a loop comprising a cluster of highly conserved acidic residues. Conformational changes linked to the redox chemistry of ubiquinone are proposed to induce an electric pulse that ultimately triggers proton translocation events in the membrane arm. Reprinted from Ref. 131 with the permission from the Elsevier.

the formation of Ab aggregates 146e148 , which could also contribute to its beneficial properties.

Further translational support for this therapeutic strategy was provided by the cross-validation of transcriptomic data generated in CP2-treated AD mice with the human brain transcriptome data available through the co-expression meta-analysis in the Accelerating Medicines Partnership Program for Alzheimer's disease database (ampadportal.org). Beneficial changes in gene expression associated with CP2 treatment in APP/PS1 mice overlap with signatures established in patients with AD, female patients in particular, supporting high translational potential of this approach 142 . Major translational targets included the immune system response and multiple pathways involved in synaptic function and neurotransmission, which underlie early pathology in patients with AD 149 . Since CP2 improved axonogenesis and dendritic spine morphology and function, it is feasible that this treatment could also induce neuronal regeneration.

Molecular mechanisms of neuroprotection converged on the AMPK activation and the downstream signaling that resulted in increased resistance to oxidative stress, augmented mitochondrial bioenergetics, improved glucose uptake and utilization, increased production of sirtuins 1 and 3, reduction of glycogen synthase kinase 3 beta (GSK3b) activity, significant reduction in levels of pTau and Ab, and increased autophagy and levels of BDNF and synaptic proteins in vivo 117, 142, 143 . With CP2-inhibited complex I activity, the overall energy levels in the brain measured using 31 P MRI were not decreased, which could be attributed to enhanced mitochondrial biogenesis and bioenergetics and improved brain energy homeostasis 142 . The translational relevance of this approach is emphasized by the fact that the intervention was started after the onset of Ab neuropathology 150 , cognitive symptoms 151 , bioenergetic dysfunction 152 , and progressive neurodegeneration 153 . These data provide further support for brain energy rescue as a novel concept for treatment of neurodegenerative diseases of aging 9, 142, 143 . Furthermore, similar to metformin and resveratrol, CP2 also enhanced health and life span in chronologically aged wild-type mice and mice fed with a high-fat diet (our unpublished observations), implying that the activation of mitochondria-induced ISR using complex I as a small molecule druggable target could delay the onset or block the progression of age-related neurodegenerative diseases.

We summarized here evidence for a novel therapeutic approach to exploit the incredible ability of mitochondria to engage multifaceted neuroprotective stress response triggered by partial complex I inhibition. This approach promises relief for multiple human conditions, including, but not limited to mitochondrial diseases, HD, PD, and amyotrophic lateral sclerosis, and to promote healthy aging to delay the onset of neurogenerative diseases, AD in particular, where age is the greatest risk factor. There is a mounting body of evidence generated in model organisms and humans in support of the safety of chronic application of complex I inhibitors. However, a better understanding of the molecular mechanisms is required to establish safety in translation to humans, including the development of biomarkers that inform on mitochondrial function and the capacity to induce the beneficial stress response. Further therapeutic developments should produce selective and specific complex I inhibitors capable of penetrating the BBB with excellent safety profile. UH3NS 113776, all to Eugenia Trushina, USA]. Its contents are solely the responsibility of the authors and do not necessarily represent the official view of the NIH. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. 

Eugenia Trushina is an inventor on the patent US20180044295A1 ("Compounds for modulating mitochondrial function"). The authors declare no conflict of interest. Figure 4 Partial inhibition of mitochondria complex I with small molecule compound CP2 activates multiple AMP-activated protein kinase-dependent mechanisms leading to neuroprotection in mouse models of Alzheimer's disease.

Alzheimer's disease facts and figures

Epidemiology of Alzheimer disease

Dissecting complex and multifactorial nature of Alzheimer's disease pathogenesis: a clinical, genomic, and systems biology perspective

Sex-specific genetic predictors of Alzheimer's disease biomarkers

Metabolic network failures in Alzheimer's disease: a biochemical road map

Mitochondria and mitochondrial cascades in Alzheimer's disease

Oxidative stress, synaptic dysfunction, and Alzheimer's disease

Brain glucose hypometabolism and oxidative stress in preclinical Alzheimer's disease

Brain energy rescue: an emerging therapeutic concept for neurodegenerative disorders of ageing

Why have antioxidants failed in clinical trials?

Novel targets for mitochondrial medicine

Centenarians as models of resistance and resilience to Alzheimer's disease and related dementias

Protective genes and pathways in Alzheimer's disease: moving towards precision interventions

Ancient invasions: from endosymbionts to organelles

Brain energy metabolism: focus on astrocyte-neuron metabolic cooperation

Regulation of substrate utilization by the mitochondrial pyruvate carrier

Quantifying intracellular rates of glycolytic and oxidative ATP production and consumption using extracellular flux measurements

Neuromethods techniques to investigate mitochondrial function in neurons

Regulation of oxidative phosphorylation, the mitochondrial membrane potential, and their role in human disease

Assessing mitochondrial dysfunction in cells

Estimation of membrane potential and pH difference across the cristae membrane of rat liver mitochondria

The influence of respiration and ATP hydrolysis on the proton-electrochemical gradient across the inner membrane of ratliver mitochondria as determined by ion distribution

The causes and functions of mitochondrial proton leak

Fueling thought: management of glycolysis and oxidative phosphorylation in neuronal metabolism

Evolution of mitochondria as signaling organelles

Mitocellular communication: shaping health and disease

Metabolic signaling to the nucleus in cancer

Mitochondrial complex III is essential for suppressive function of regulatory T cells

NAD þ in aging, metabolism, and neurodegeneration

Modulating NAD þ metabolism, from bench to bedside

SIRT3 deacetylates and activates OPA1 to regulate mitochondrial dynamics during stress

Mitochondrial SIRT3 mediates adaptive responses of neurons to exercise and metabolic and excitatory challenges

Mitochondrial ROS regulation of proliferating cells

Mitochondrial reactive oxygen species trigger hypoxia-induced transcription

ROS function in redox signaling and oxidative stress

Unraveling the truth about antioxidants: mitohormesis explains ROS-induced health benefits

VDAC oligomers form mitochondrial pores to release mtDNA fragments and promote lupus-like disease

Mitochondrial DNA stress primes the antiviral innate immune response

Determinants and functions of mitochondrial behavior

Mitochondrial dynamics and transport in Alzheimer's disease

Mitochondrial unfolded protein response: an emerging pathway in human diseases

Mitochondrial dynamics in adaptive and maladaptive cellular stress responses

Mitochondrial membrane potential

ER-to-mitochondria miscommunication and metabolic diseases

MitochondriaeER tethering in neurodegenerative diseases

AMP-activated protein kinase: the current landscape for drug development

AMP-activated/SNF1 protein kinases: conserved guardians of cellular energy

AMPK: a nutrient and energy sensor that maintains energy homeostasis

AMPK: guardian of metabolism and mitochondrial homeostasis

AMPK activation: role in the signaling pathways of neuroinflammation and neurodegeneration

AMPK and SIRT1: a long-standing partnership?

Mitochondrial sirtuins and molecular mechanisms of aging

Mitochondrial dysfunction in Alzheimer's disease and progress in mitochondria-targeted therapeutics

Mitochondrial import efficiency of ATFS-1 regulates mitochondrial UPR activation

Two conserved histone demethylases regulate mitochondrial stress-induced longevity

N 6 -Methyldeoxyadenine is a transgenerational epigenetic signal for mitochondrial stress adaptation

Distinct mitochondrial defects trigger the integrated stress response depending on the metabolic state of the cell

A systematic RNAi screen identifies a critical role for mitochondria in C. elegans longevity

Rates of behavior and aging specified by mitochondrial function during development

Mitochondrial electron transport is a key determinant of life span in Caenorhabditis elegans

A systematic RNAi screen for longevity genes in C. elegans

Altered quinone biosynthesis in the long-lived clk-1 mutants of Caenorhabditis elegans

Mitochondrial oxidative phosphorylation is defective in the long-lived mutant clk-1

Relationship between mitochondrial electron transport chain dysfunction, development, and life extension in Caenorhabditis elegans

Extension of Drosophila life span by RNAi of the mitochondrial respiratory chain

Low abundance of the matrix arm of complex I in mitochondria predicts longevity in mice

Early mitochondrial dysfunction in long-lived Mclk1 þ/e mice

Increased longevity and refractoriness to Ca 2þ -dependent neurodegeneration in Surf1 knockout mice

Cytochrome c oxidase deficiency in neurons decreases both oxidative stress and amyloid formation in a mouse model of Alzheimer's disease

Mitochondrial DNA damage in a mouse model of Alzheimer's disease decreases amyloid beta plaque formation

A review of the inhibition of the mitochondrial ATP synthase by IF1 in vivo: reprogramming energy metabolism and inducing mitohormesis

AMP-activated protein kinase regulates the cell surface proteome and integrin membrane traffic

Expression of the uncoupling protein 1 from the aP2 gene promoter stimulates mitochondrial biogenesis in unilocular adipocytes in vivo

Targeting mitochondrial cardiolipin and the cytochrome c/cardiolipin complex to promote electron transport and optimize mitochondrial ATP synthesis

Dinitrophenol-induced mitochondrial uncoupling in vivo triggers respiratory adaptation in HepG2 cells

FR58P1a; a new uncoupler of OXPHOS that inhibits migration in triple-negative breast cancer cells via Sir-t1/AMPK/b1-integrin pathway

Effects of caffeine on metabolism and mitochondria biogenesis in rhabdomyosarcoma cells compared with 2,4-dinitrophenol

Longitudinal RNA-Seq analysis of vertebrate aging identifies mitochondrial complex I as a small-moleculesensitive modifier of lifespan

Mechanistic investigations of the mitochondrial complex I inhibitor rotenone in the context of pharmacological and safety evaluation

The cooccurrence of mtDNA mutations on different oxidative phosphorylation subunits, not detected by haplogroup analysis, affects human longevity and is population specific

AMP-activated protein kinase (AMPK) controls the aging process via an integrated signaling network

Decreased reactive oxygen species production in cells with mitochondrial haplogroups associated with longevity

Mitochondria in health, aging and diseases: the epigenetic perspective

Dietary restriction and AMPK increase lifespan via mitochondrial network and peroxisome remodeling

AMPK and mTOR regulate autophagy through direct phosphorylation of Ulk1

Caloric restriction can improve learning and memory in C57/BL mice probably via regulation of the AMPK signaling pathway

Effects of 2 years of caloric restriction on oxidative status assessed by urinary F2-isoprostanes: the CALERIE 2 randomized clinical trial

Autophagy and aging: maintaining the proteome through exercise and caloric restriction

The mechanisms of action of metformin

Effect of intensive blood-glucose control with metformin on complications in overweight patients with type 2 diabetes (UKPDS 34)

Metformin selectively targets redox control of complex I energy transduction

Metformin activates AMPK through the lysosomal pathway

Metformin improves mitochondrial respiratory activity through activation of AMPK

Comparative efficacy and acceptability of antidiabetic agents for Alzheimer's disease and mild cognitive impairment: a systematic review and network meta-analysis

Effects of the insulin sensitizer metformin in Alzheimer disease: pilot data from a randomized placebo-controlled crossover study

Metformin in amnestic mild cognitive impairment: results of a pilot randomized placebo controlled clinical trial

The effect of diabetes medication on cognitive function: evidence from the PATH through life study

Metformin reduces all-cause mortality and diseases of ageing independent of its effect on diabetes control: a systematic review and meta-analysis

Anti-inflammatory effects of metformin irrespective of diabetes status

Caloric restriction and exercise "mimetics'': ready for prime time?

Effect of metformin on life span and on the development of spontaneous mammary tumors in HER-2/neu transgenic mice

Caloric restriction and aging: studies in mice and monkeys

Caloric restriction improves memory in elderly humans

Metformin use associated with reduced risk of dementia in patients with diabetes: a systematic review and meta-analysis

Antidiabetic drug metformin (GlucophageR) increases biogenesis of Alzheimer's amyloid peptides via up-regulating BACE1 transcription

Metformin, other antidiabetic drugs, and risk of Alzheimer's disease: a population-based casecontrol study

Increased risk of cognitive impairment in patients with diabetes is associated with metformin

The CAMKK2-AMPK kinase pathway mediates the synaptotoxic effects of Ab oligomers through Tau phosphorylation

Roles of AMP-activated protein kinase in Alzheimer's disease

Association between metformin and vitamin B 12 deficiency in patients with type 2 diabetes: a systematic review and meta-analysis

Cognitive impairment and vitamin B12: a review

GLUCOPHAGEâ (metformin hydrochloride extended-release tablets)

Metformin: a potential candidate for targeting aging mechanisms

Brain insulin resistance in type 2 diabetes and Alzheimer disease: concepts and conundrums

The therapeutic potential of metformin in neurodegenerative diseases

Therapeutic potential of resveratrol in Alzheimer's disease

Modulation of mitochondrial complex I activity averts cognitive decline in multiple animal models of familial Alzheimer's disease

Tricyclic pyrone compounds prevent aggregation and reverse cellular phenotypes caused by expression of mutant huntingtin protein in striatal neurons

Cardioprotective actions of grape polyphenols

Resveratrol: a double-edged sword in health benefits

Small molecule SIRT1 activators for the treatment of aging and age-related diseases

Resveratrol directly binds to mitochondrial complex I and increases oxidative stress in brain mitochondria of aged mice

Effects of resveratrol on the rat brain respiratory chain

Inhibition of mitochondrial proton F0F1-ATPase/ATP synthase by polyphenolic phytochemicals

Resveratrol: a focus on several neurodegenerative diseases

Neuroprotective action of resveratrol

Neuroprotective effects of resveratrol in Alzheimer disease pathology

Resveratrol as a therapeutic agent for neurodegenerative diseases

Significance of resveratrol in clinical management of chronic diseases

Molecular and cellular mechanisms underlying the pathogenesis of Alzheimer's disease

Structure and function of mitochondrial complex I

Atomic structure of the entire mammalian mitochondrial complex I

High-resolution cryo-EM structures of respiratory complex I: mechanism, assembly, and disease

Differential effects of mitochondrial complex I inhibitors on production of reactive oxygen species

Structural basis for a complex I mutation that blocks pathological ROS production

Mitochondria and Parkinson's disease: clinical, molecular, and translational aspects

Mitochondrial complex I inhibition is not required for dopaminergic neuron death induced by rotenone, MPP þ , or paraquat

Neuronal complex I deficiency occurs throughout the Parkinson's disease brain, but is not associated with neurodegeneration or mitochondrial DNA damage

Mitochondrial dysfunction in Huntington's disease; interplay between HSF1, p53 and PGC-1a transcription factors

Mutant huntingtin impairs axonal trafficking in mammalian neurons in vivo and in vitro

Modulation of mitochondrial complex I activity averts cognitive decline in multiple animal models of familial Alzheimer's disease

Partial inhibition of mitochondrial complex I ameliorates Alzheimer's disease pathology and cognition in APP/PS1 female mice

Partial inhibition of mitochondrial complex I reduces Tau pathology and improves energy homeostasis and synaptic function in 3xTg-AD mice

A genome-wide association study in human lymphoblastoid cells supports safety of mitochondrial complex I inhibitor

Bioenergetic analysis of isolated cerebrocortical nerve terminals on a microgram scale: spare respiratory capacity and stochastic mitochondrial failure

Inhibition of Alzheimer's amyloid toxicity with a tricyclic pyrone molecule in vitro and in vivo

Novel tricyclic pyrone compounds prevent intracellular APP C99-induced cell death

A novel tricyclic pyrone compound ameliorates cell death associated with intracellular amyloid-beta oligomeric complexes

Large-scale proteomic analysis of Alzheimer's disease brain and cerebrospinal fluid reveals early changes in energy metabolism associated with microglia and astrocyte activation

Quantitative histological analysis of amyloid deposition in Alzheimer's double transgenic mouse brain

Behavioral changes in transgenic mice expressing both amyloid precursor protein and presenilin-1 mutations: lack of association with amyloid deposits

Defects in mitochondrial dynamics and metabolomic signatures of evolving energetic stress in mouse models of familial Alzheimer's disease

The mechanistic link between selective vulnerability of the locus coeruleus and neurodegeneration in Alzheimer's disease

We thank Drs. Thomas Chung and Graham Johnson for constructive suggestions, and Ms. Shelly Gochnauer for help with the manuscript. This research was supported by grants from the National Institutes of Health [grant numbers RF1AG55549, R01NS107265, RO1AG062135, AG59093, AG072899, UG3/