*3.2. Neurodegenerative Diseases*

The neurodegenerative diseases include AD, PD, ALS and Multiple Sclerosis (MS), injury to the central nervous system (CNS) through chronic low-grade hypoxia, the rarer Huntington's disease (HD), Wilson's disease and Freidreich's Ataxia. In all these diseases, impaired ATP generation causes a failure of cellular homeostasis, with a number of consequences, including the ionic imbalance, altered Ca2+-dependent transmission of information in the CNS and ultimately, necrotic or apoptotic cell death, depending on ATP depletion.

The first study to implicate ATP synthase in AD etiology found decreased level of the entire complex in the hippocampal tissue of AD patients through BN-PAGE analysis [111]. Whether the impaired complex stability in AD patients during detergent extraction also reflects changes in the catalytic activity of ATP synthase is still debated in the literature. An early study found no significant decrease in the ATP synthase catalytic activity in isolated mitochondria from AD patient hippocampal tissue, motor cortex and platelets [112]. However, there is evidence of post-translational modifications occurring in AD with consequences on the ATP synthase activity. In the hippocampus of AD patients, the α subunit was shown to be excessively nitrated in comparison to age-matched control brains [113], a posttranslational modification that has been shown to inhibit the ATP synthase catalysis [92]. The α subunit can be subjected to glycosylation with O-linked β-N-acetylglucosamine (O-GlcNAcylated) on the Thr432 residue. This modification was reduced in brains of AD patients, Tg AD mice and in Aβ-treated mammalian cell cultures–resulting in reduced ATP levels [114]. Molecular modelling and co-IP experiments with deletion mutants of the α and β subunits showed that Aβ directly blocks the O-GlcNAcylation of this Thr432 residue.

Interestingly, the O-GlcNAcylation of Thr432 that had been inhibited by Aβ was rescued by treatment with the O-GlcNAcase inhibitor. These findings are particularly noteworthy as the authors demonstrated a chemical mechanism for the interaction of the Aβ peptide with mitochondrial ATP synthase, which could provide a potential therapeutic target for AD [114]. Another post-translational modification was shown in the presence of a lipid peroxidation product, 4-hydroxy-2-nonenal (4-HNE) [115]. This product was shown to modify the α subunit of ATP synthase. It caused a 35%-decrease of ATP hydrolysis in the hippocampal tissue of early stage AD individuals with mild cognitive impairment [116] and a 30%-decreased ATP synthase activity in the entorhinal cortex [117], suggesting that oxidative stress precedes the presence of Aβ in the affected tissue.

Multiple studies have pointed to a decrease in the levels of ATP synthase subunits in AD models. A decreased expression was observed in several of the nuclear ATP synthase genes in the posterior cingulate cortex, hippocampal field CA1, middle temporal gyrus, entorhinal cortex and posterior cingulate neurons [118]. A study using induced pluripotent stem cell (iPSC) -derived hippocampal neuronal cells, with familial associated presenilin 1 (PS1) mutation M146, observed a decreased level of the ATP synthase complex while PS1 expression was kept at physiological levels [119]. Neuroblastoma cells expressing the *ApoE4* allele of the *ApoE* gene, the major genetic risk factor for sporadic AD, showed a reduction in the levels of all detected ATP synthase subunits [120]. A reduced expression of the catalytic β subunit mRNA levels by over 50% was found in the mid-temporal cortex of AD patient brains [121]. In another study linking Aβ peptides with ATP synthase in AD, rats receiving a bilateral intrahippocampal injection of Aβ showed a significant decrease in the levels of β subunit of ATP synthase [122]. Gene expression analysis of the entorhinal cortex of AD patient brains showed reduced expression of γ, δ, c, and β subunit genes [123]. The subunits β, d, e, and F6 were down-regulated in the early-onset AD, as revealed by the iTRAQ quantitative mass spectrometric technique [124], whereas another proteomics analysis of hippocampal subcellular fractions from a murine AD model showed a decreased level of the peripheral stalk subunit d [125]. Interestingly, the d subunit gene was firstly thought to be genetic risk factor for AD in a genome wide association study [126].

On the contrary, increased expression was found for the ATP synthase α subunit gene in a Transgenic Swedish APP mouse (Tg2576) model for AD, with increased levels of amyloid plaque formation in the brain [127]. Moreover, a transgenic mice line (J20 Tg) producing a mutant form of APP, corresponding to the Swedish and Indiana familial forms of AD, showed a 12.2-fold increase in the α subunit level in a whole brain homogenate [128]. One might speculate that the increased subunit levels are due to an adaptive response only occurring in the AD animal models. Yet, in 2004 a study in AD patients by Manczak et al. showed increased mRNA levels for the mitochondrial *ATP6* and *ATP8* genes in brains, while increased levels of the δ subunit in the frontal cortex by immunofluorescence analysis [129].

ATP synthase is further controlled by possible interactions with other proteins modulating its activity and influencing neurodegeneration. Selective loss of the peripheral stalk subunit of ATP synthase, OSCP, was found in the brains of AD individuals and in an AD mouse model [130]. OSCP loss and complex interactions with Aβ leads to reduced ATP production, elevated oxidative stress and activated permeability transition [130]. The authors suggested that the restoration of OSCP ameliorates Aβ-mediated mouse and human neuronal mitochondrial impairments, including the effects on ATP synthesis and the resultant synaptic injury [130]. This finding demonstrating the OSCP involvement in AD is of particular interest, given that this subunit is the molecular-binding site for the CypD, the matrix prolyl-cis-trans-isomerase, which has been shown to modulate ATP synthase catalytic activity [91] and the ATP synthase transition to the permeability transition pore [131]. A previous study in an AD mouse model showed that neuronal and synaptic stress due to the interaction of CypD with mitochondrial Aβ are attenuated in CypD-deficient cortical mitochondria. CypD deficiency protected neurons from Aβ- and oxidative stress-induced cell death, in a mechanism involving the permeability transition pore [132], which was also confirmed in other AD transgenic mouse models [133]. Moreover, CypD levels, which

increased in aging mice, have been shown to decrease ATP synthase activity and to promote mitochondrial dysfunction [134]. Compared with non-synaptic mitochondria, the synaptic mitochondria showed a greater degree of age-dependent accumulation of Aβ and deficits in mitochondrial function, as shown by increased mitochondrial permeability transition and decline in respiration [135]. In an AD animal model (5×FAD mice) the genetic depletion of CypD mitigates OSCP loss via ubiquitin-dependent OSCP degradation and attenuates OSCP/Aβ interaction preserving the ATP synthase function, mitochondrial bioenergetics and improved mouse cognition [136]. The authors' interpretation is that CypD is a critical mediator that promotes OSCP deficits in AD-related conditions, providing a promising therapeutic strategy to correct mitochondrial dysfunction for AD therapy.

The most frequent form of neurodegenerative disorder affecting movement, PD, is caused by death of dopaminergic neurons in the mesencephalic region called substantia nigra pars compacta. In astrocytes derived from *PINK1*-knockout mice, proliferation defects were associated with a decrease in mitochondrial mass, membrane potential and ATP production as well as an increase in cellular ROS. Treatment of wild-type astrocytes with the ATP synthase inhibitor oligomycin was sufficient to mimic the proliferation phenotype observed in *PINK1*-deficient murine cells [137]. Protein aggregation and mitochondrial dysfunction are two central pathogenic processes in both familial and sporadic PD. However, the way in which these two processes converge to cause neurodegeneration was only recently proposed. Protein aggregation causes α-synuclein to switch from its physiological role to a pathological toxic gain of function form. Under physiological conditions, monomeric α-synuclein improves ATP synthase efficiency [138]. On the other hand, aggregation of α-synuclein monomers generates beta sheet-rich oligomers localized in the mitochondria in close proximity to several mitochondrial proteins including ATP synthase. Oligomers induce selective oxidation of the ATP synthase β subunit and mitochondrial lipid peroxidation. These oxidation events are proposed to increase the probability of permeability transition pore opening, triggering mitochondrial swelling and, ultimately, cell death [139]. Moreover, the protein DJ-1, linked to early onset PD, if defective, binds the ATP synthase β subunit. The interaction with the wild-type form of DJ-1 decreased the mitochondrial uncoupling and enhanced ATP production, while mutations in *PARK7* gene encoding DJ-1 (or *PARK7*-knockout) increased mitochondrial uncoupling and depolarized neuronal mitochondria [140]. The Authors suggested that this observation may depend on the presence of a leak at the level of the c-ring of ATP synthase in the membrane, which may be closed by pharmacological treatment [140].

ALS is an appalling neurodegenerative disease characterized by the loss of spinal motor neurons, which is rapidly progressive and lethal [141]. The most common genetic form of the disease is caused by GGGGCC repeat expansion in the *C9ORF72* gene. It was found that poly(GR) preferentially binds to the ATP synthase α subunit and promotes its degradation via the ubiquitin-proteasome pathway. Moreover, inducing the expression of *ATP5F1A* gene in poly(GR)-expressing neurons or reducing the poly(GR) level in adult mice after disease onset, rescued poly(GR)-induced neurotoxicity [142].

In a small subset of patients, the disease is caused by mutations in superoxide dismutase 1 [143]. Transgenic mice producing the mutant protein display mitochondrial alterations, including swelling, respiratory inhibition and an elevated generation of ROS [144,145]. Another form of ALS is caused by mutant forms of fused in sarcoma or translocated in liposarcoma (FUS), which is a multifunctional DNA/RNA-binding protein associated with neurodegeneration. In both cellular and animal models, the expression of wild-type or an ALS-associated mutant (p.Pro525Leu) FUS disrupts the formation of the mitochondrial ATP synthase supercomplexes and suppresses the activity of ATP synthase, resulting in mitochondrial cristae loss followed by mitochondrial fragmentation. Expression of FUS increases levels of the β subunit which is not properly assembled, and importantly, the downregulation of this subunit by RNA interference partially rescues neurodegenerative phenotypes [146]. In other studies, fibroblasts from patients affected by frontotemporal dementia and ALS presented mitochondrial ultrastructural alterations and fragmentation of

the mitochondrial network together with respiratory chain deficiency. A missense mutation was identified (c.176C>T; p.Ser59Leu) in the *CHCHD10* gene which encodes a mitochondrial coiled-coil helix protein, whose function is unknown. Blue native-PAGE analysis of patient muscles revealed altered ATP synthase assembly which might contribute to the described abnormal organization of cristae morphology in these patients [147].
