Interrogating Parkinson’s Disease Associated Redox Targets: Potential Application of CRISPR Editing

Abstract

Loss of dopaminergic neurons in the substantia nigra is one of the pathogenic hallmarks of Parkinson’s disease, yet the underlying molecular mechanisms remain enigmatic. While aberrant redox metabolism strongly associated with iron dysregulation and accumulation of dysfunctional mitochondria is considered as one of the major contributors to neurodegeneration and death of dopaminergic cells, the specific anomalies in the molecular machinery and pathways leading to the PD development and progression have not been identified. The high efficiency and relative simplicity of a new genome editing tool, CRISPR/Cas9, make its applications attractive for deciphering molecular changes driving PD-related impairments of redox metabolism and lipid peroxidation in relation to mishandling of iron, aggregation and oligomerization of alpha-synuclein and mitochondrial injury as well as in mechanisms of mitophagy and programs of regulated cell death (apoptosis and ferroptosis). These insights into the mechanisms of PD pathology may be used for the identification of new targets for therapeutic interventions and innovative approaches to genome editing, including CRISPR/Cas9.

Graphical Abstract

“The art of medicine consists in amusing the patient while nature cures the disease.”

Voltaire

1. Introduction

Parkinson’s disease (PD) – the condition first described by James Parkinson at the beginning of the XIX century – still remains one of the most debilitating diseases worldwide. PD is a progressive neurodegenerative disorder involving the loss of dopaminergic (DA) neurons in the substantia nigra (SN) of the midbrain [1]. PD is characterized by severe motor deficits such as bradykinesia, postural instability and tremor. Most patients with PD exhibit impaired cognition, especially deficits in attention, executive functioning, visual-spatial processing and varying degree of dementia [2, 3]. Not only does PD affects the patient’s quality of life, but as populations age, patients with PD are expected to impose an increasing social and economic burden [4].

Morphological hallmarks of PD include loss of 50-70% of DA neurons, and the presence of Lewy Bodies (LBs) enriched in abnormal α-synuclein (α-syn) filaments [5, 6]. Whether these LBs contribute to disease pathogenesis or serve as markers of underlying cellular dysfunction or adaptation remains controversial.

Most cases of PD are sporadic and approximately 10% are familial in etiology [7]. Familial PD is a genetically inherited form of PD accompanied by a variety of disturbances suggesting that mutations in one or more genes can lead to the synthesis of abnormal proteins associated with increased aggregation and/or changes in enzymatic activity. On the contrary, sporadic PD neither has apparent genetic linkage nor clear idea on the contribution of genetic risk variants and environmental risk factors to this form of PD [8]. Among many possible genetic variants potentially contributing to the sporadic PD, in this review we will focus on those associated with aberrant redox mechanisms and leading to triggering of cell death programs, including apoptosis and ferroptosis. The emphasis will be on the enzymatic pathways of selective phospholipid peroxidation as mechanisms of apoptotic and ferroptotic cell death programs in the context of the SN iron dishomeostasis that lead to the disturbed redox balance and H₂O₂ (ROS) accumulation feeding the enzymatic lipid peroxidation. We will also discuss the role of lipid peroxidation as a source of lipid mediators (oxidized PUFA) that affect inflammation and abnormalities associated with the PD progression [9–13].

Currently, symptomatic drug therapy is the main strategy for PD. No mechanism based therapeutic approaches have been developed that may prevent, control or reduce the pathological manifestations of PD [14, 15]. Moreover, the symptomatic therapeutic modalities used are associated with a variety of side effects. In the field of cell replacement therapies, a few studies showed the potential of generation of DA neurons from human embryonic stem cells (hESCs) followed by implantation of these cells in animal models of PD [16–19]. The initial results showed the increase in dopamine levels in the brain of experimental animals [16, 17]. However, this approach has a lot of unsolved problems, including risk of brain tumors, phenotype instability of hESC-derived DA neurons, immunologic responses, not to mention ethical concerns and the necessity to examine the safety and efficacy of the treatment for PD patients.

Based on the possible pathogenic role of α-syn and the demonstrated capacity of microglia to degrade abnormal intracellular α-syn filaments and prevent damage of DA neurons [7], vaccines against α-syn might be a good therapeutic strategy. However, no studies developing this approach have been published so far.

There is an obvious urgency in mechanistic studies of a new type aimed at integrative understanding of genetic and environmental factors triggering multiple aberrant metabolic pathways and erroneous interactions of various macromolecules during overall PD pathogenesis. In this regard, promising opportunities may be associated with CRISPR/Cas9 system – a new tool developed during the last five years – which enables quick and precise genome editing in almost any living creature [20–22]. Several recent publications indicate that CRISPR/Cas9 has a potential to broaden and accelerate the fundamental research aimed at understanding of pathogenic mechanisms of neurodegenerative diseases and lead to new therapies, especially for PD [23–25].

This new technology can be applied for genomewide screening experiments to simultaneously assess large numbers of genes and identify previously unknown PD-associated genetic variants that enable subsequent generation of adequate cellular and animal models effectively applicable for drug discovery. There are several excellent recent reviews considering different applications of CRISPR/Cas9 technology in relation to PD [26–29]. This review is focused on the aspects of direct reprogramming of host neuronal and immune cells in the context of targeting the redox mechanisms of lipid signaling and neuro-inflammation associated with genetic programs of regulated cell death (Table 1).

Table 1. Genes implicated in Parkinson’s disease impairment of redox metabolism and lipid peroxidation.

Here we reviewed most promising candidate genes and processes that highlight PD pathology: mitophagy, iron homeostasis, ferroptotic and apoptotic cell death programs.

Gene	Protein	Function
ACSL4	Acyl-CoA Synthetase Long-Chain Family Member 4	Preferably converts arachidonic acid into fatty acyl-CoA esters, and thereby plays a key role in ferroptosis
ALOXs	Lipoxygenases	Involved in peroxidation of polyunsaturated fatty acids in ferroptosis
APP	Amyloid precursor protein	Regulator of synapse formation, neural plasticity and iron export
FBXO7	F-box only protein 7	Mediates the ubiquitination and subsequent proteasomal degradation of target proteins; regulation of mitophagy
GCLC/GCLM	Glutamate-cysteine ligase	Biosynthesis of glutathione
GPX4	Glutathione peroxidase 4	Reduces membrane phospholipid hydroperoxides to suppress ferroptosis
GSS	Glutathione synthetase	Biosynthesis of glutathione
IREB2	Iron-responsive element-binding protein 2	Regulator of TFRC, FPN, DMT1 translational activity
MAPT	tau	Microtubule-associated protein; regulates APP trafficking in neurons
NME4 (NDPK-D)	Nucleoside diphosphate kinase D	Synthesis of nucleoside triphosphates other than ATP; regulator of mitophagy
PARK2	Parkin	Ubiquitin ligase, regulator of mitophagy
PARK7	Nucleic acid deglycase DJ-1	Redox-sensitive chaperone and a sensor for oxidative stress. Required for mitochondrial morphology and function
PINK1	PTEN-induced putative kinase 1	Kinase, phosphorylates Parkin and ubiquitin to regulate mitophagy
SLC40A1	Solute carrier family 40 member 1 (Ferroportin)	Transmembrane protein that transports iron from the inside of a cell; promotes ferroptosis
SNCA	α-synuclein	Regulator of dopamine biosynthesis, release and transport; suppressor of apoptosis

2.1. Overview

In the early 2000s, Clustered Regularly Interspaced Short Palindromic Repeats (CRISPR) has been found in the genomes of different prokaryotes, providing them with immunity against viral infections [30–33]. Since then, three types and twelve subtypes of CRISPR/Cas systems have been discovered [34]. The best characterized is the type II CRISPR system [35–38] that consists of the endonuclease Cas9, bacterial CRISPR RNA (crRNA) and trans-activating crRNA (tracrRNA) (Fig. 1). Later these naturally occurring bacterial RNAs were fused together to generate the single guide RNA molecule (gRNA) [37]. The target activity of CRISPR/Cas9 systems is provided by protospacer adjacent motif (PAM) – a sequence within an invading DNA that helps bacteria to recognize pathogenic genetic material from its own [39, 40]. Only if the spacer sequence is fully homologous to PAM the CRISPR/Cas9 system will target viral or plasmid genetic material by introducing double-stranded (ds)DNA breaks in the invading DNA [41]. These discoveries have led to the realization that the CRISPR/Cas9 system can be used as a new tool for genome editing in a variety of organisms. For this, the DNA sequence of interest is selected and a dsDNA break is introduced in a selected region by simply adding designed gRNA, Cas9 protein and your repair template (a plasmid) if necessary (Fig. 1). The technology offers new tools for a broad spectrum of genetic manipulations, including introduction of point mutations [42, 43], deletion and insertion of short and large DNA fragments into an organism’s genome [44, 45], epigenome editing [46] and CRISPR-based activation and inactivation of different genes [47, 48]. A number of genome editing technologies preceded CRISPR/Cas9 system: zinc-finger nucleases (ZFNs) [49, 50] and transcription activator-like effector nucleases (TALENs) [50–52]. However, CRISPR/Cas9 genome editing technology possesses a number of advantages over previous approaches. First, CRISPR/Cas9 system is more precise as it utilizes specific gRNAs fully complementary through Watson-Crick base pairing to the PAM and targeted DNA sequence. It makes CRISPR/Cas9 technology more specific and easier to design. Second, CRISPR/Cas9 technology saves a lot of research time. Modern gRNA designing tools allow a researcher to create a perfect guide RNA for a genomic locus of interest within an hour. The simple principle of CRISPR/Cas9 genome editing and an easy to perform procedure allows generation of genetically/epigenetically modified cells within a few weeks. Third, an important point about CRISPR is its versatility. By changing gRNA sequence one can target almost any sequence of any living creature from algae to humans.

Fig. 1. CRISPR/Cas9 genome editing system.

The CRISPR/Cas9 two-component system is composed of Cas9 endonuclease and a single guide RNA (gRNA) molecule. Endonuclease Cas9 generates site-specific double-strand breaks in the genomic locus of interest by direct pairing of gRNA molecule with the target DNA. The targeting specificity relies on the presence of PAM sequence (protospacer adjacent motif) in the DNA sequence of interest. Upon Cas9 action double-stranded breaks are introduced into the DNA, which in mammalian cells are usually repaired either by non-homologous end joining (NHEJ), an error-prone mechanism that generates random insertion and deletion mutations at the targeted locus (used for knock-outs), or by homology-directed repair (HDR), for which an exogenous repair template is provided.

2.2. Application of CRISPR/Cas9 genome editing to Parkinson’s disease

Difficulties in deciphering the PD pathogenic mechanisms are due, to a large extent, to the multiplicity of mutations with Mendelian autosomal dominant (AD), SNCA, LRRK2, VPS35, HTRA2, EIF4G1, GBA, or autosomal recessive (AR) inheritance – PARK2 (encoding for Parkin protein), PINK1, PARK7 (known as DJ-1), ATP13A2, PLA2G6, FBXO7 [53–61]. With the familial forms of PD accounting only for 10% cases [14], and the rest of them considered sporadic [58], it is difficult to sort through many genetic risk variants [62, 63] to determine if they could be causally linked to PD pathogenesis. Application of CRISPR/Cas9 and CRISPR for screening and identification of new therapeutic redox targets may be quite valuable.

3. The role of mitochondria in Parkinson’s disease

3.1. Parkin/Pink1-dependent mitophagy

Mitochondria are unique power-houses of cells with multiple additional signaling and metabolic functions [64]. As the major electron transporting redox organelle, mitochondria can also generate reactive oxygen species (ROS) essential for the dynamic regulation of their own organization via fission, fusion and mitophagy [64–66]. Discoordination of mitochondrial redox machinery may lead to their damage which, in turn, triggers self-elimination through selective autophagy, known as mitophagy. Mitophagy has been observed in degenerating SN neurons of patients with PD or the related Lewy body dementia (LBD) [67]. Inefficient mitophagy or excessive or prolonged activation of mitophagy may contribute to neurodegeneration [11, 68]. In response to different physiological stimuli, various signaling cascades can be activated to trigger specific types of mitophagy [69]. Impairment of this essential cellular process perturbs mitochondrial function and leads to accumulation of damaged mitochondria within the cells and furthermore promotes cell and tissue damage [68]. Mitochondrial damage and excessive ROS production have also been considered as causative pathogenic factors in animals with PD-associated mutations [8, 70].

PTEN-induced putative kinase 1(PINK1), a cytosolic E3 ubiquitin ligase (Parkin), and F-box only protein 7 (Fbxo7) are essential for proper mitochondrial functioning (Fig. 2). Their disruption can affect mitochondrial structure, function and sensitize cells to pro-oxidant factors (Table 1) [71–73].

Fig. 2. Mitophagy in neurons.

Left: following stress, PINK1 is stabilized on the mitochondrial outer membrane (MOM), promoting Parkin recruitment to the mitochondria. Parkin ubiquitinates several MOM proteins. Adaptor proteins (p62, OPTIN) recognize phosphorylated mitochondrial proteins and initiate autophagosome formation through binding with LC3, thus promoting mitophagy. During apoptosis mitochondrial inner membrane (MIM) phospholipid cardiolipin (CL) interacts with cytochrome c (cyt c) and forms a peroxidase complex that catalyzes CL oxidation. The cyt c/CL complex is capable of promoting covalent hetero-oligomerization of α-syn with cyt c into high molecular weight aggregates that occur in Lewy neuritis of PD patients. Right: another protein NDPK-D (NME4) plays a key role in mitophagy initiation. In healthy mitochondria, NME4 interacts with MIM and OPA1. In this topology, NME4 has NDP kinase (phosphotransfer) activity. Depolarization of mitochondrion initiates redistribution of CL and its externalization to the mitochondrial surface; hexameric NME4 forms a complex with CL and acts like a rotary machine transferring CL from MIM to MOM through the intermembrane space (IMS) interacting simultaneously with MIM and MOM. Binding to CL suppresses NDP kinase activity, but facilitates the intermembrane transfer of the phospholipid.

Importantly, PINK1 and Parkin act in a common pathway to control turnover of mitochondria via ubiquitin-dependent mitophagy [71, 74]. PINKI-Parkin pathway regulates mitochondrial dynamics, biogenesis, transport and recruitment of autophagic machinery, to ensure proper elimination of defective organelles [68, 74, 75]. Normally, PINK1 is imported into mitochondria to reach the mitochondrial inner membrane (MIM) where it is cleaved by several proteases and degraded by the ubiquitin-proteasome system [74]. The import of PINK1 depends on mitochondrial membrane potential. Following the loss of mitochondrial membrane potential, active PINK1 starts to accumulate on the mitochondrial outer membrane (MOM), promoting Parkin translocation to the damaged mitochondria [76, 77]. Upon phosphorylation, Parkin is activated to initiate ubiquitination of numerous MOM proteins, induce recruitment of ubiquitin-binding mitophagy receptors to promote capture of damaged mitochondria by the autophagosome [68, 78, 79].

In flies, mutations in PINK1 and Parkin homologues cause mitochondrial dysfunction associated with PD, such as loss of DA neurons, mitochondrial enlargement and disintegration, muscle degeneration and shortened lifespan [71, 80]. However, PINK1 and Parkin knockout mice failed to reproduce PD related conditions observed in human patients [81]. Moreover, PINK1- and Parkin-independent pathways of mitophagy occur in neuronal cultures and in vivo [82–85]. Recently, significant phenotypic differences between several PINK1 knockout animal models and human PD patients have been documented [86–88]. Yang et al. observed marked neuronal loss in CRISPR-modified PINK1 knockout rhesus monkey that was not reproduced in studies on PINK1 deficient mice and pigs - the difference possibly associated with the primate-specific expression and function of PINK1 [89]. It was assumed that PINK1 functions may be diverse and the severity and complexity of PINK1 depleted phenotypes can depend on the type and the number of mutations in the single gene. Furthermore, the mosaicism of CRISPR/Cas9-mediated mutations can cause different degrees of PINK1 loss [89].

CRISPR/Cas9 technology can be also used in searches of new genes, transcription factors or other regulators of Parkin/PINK1 expression in cells. For example, Potting et al. applied genome-wide CRISPR/Cas9 knockout screening and discovered 53 positive and negative regulators of PARKIN abundance in the context of transcriptional repression and mitophagy [90].

Another protein, DJ-1 (Table 1), has been suggested to play a key role in regulating oxidative stress, ROS formation, mitochondrial function and autophagy [91, 92]. Mutations in the DJ-1 gene (PARK7) have been linked to autosomal-recessive early on-set PD [53]. DJ-1 is mainly localized to the cytoplasm, however, under normal conditions only a small portion of DJ-1 is present in mitochondria. Upon decrease of mitochondrial membrane potential, DJ-1 translocates into mitochondria promoting clearance of damaged organelle by mitophagy [93]. Hao et al. reported that DJ-1 knockout leads to mitochondrial dysfunction in age-dependent manner in both Drosophila and mouse. DJ-1 knockout flies manifest similar phenotypes as PINK1 and Parkin mutants: male sterility, shortened lifespan, and reduced climbing ability [94]. Moreover, DJ-1 up-regulation can rescue PINK1 deficiency in mutant flies and protect neurons against oxidative stress-induced cell death in a rat model of PD [94, 95]. While DJ-1 and PINK1/Parkin may participate in two parallel pathways whose function critically impacts mitochondrial activity the precise relationships between these three genes remain enigmatic [94]. DJ-1 has been also shown to reduce α-syn aggregation and toxicity [96]. It has been demonstrated that DJ-1 can directly interact with monomeric α-syn in vitro and in cells and reduce its dimerization. In contrast, DJ-1 mutants failed to antagonize α-syn dimerization and abrogate the protective effects of wild-type DJ-1 [97].

Fbxo7 plays a role in modulating of mitochondrial homeostasis [98]. The PD-associated mutations in this gene may result in mitochondrial dysfunction through impairment of mitochondrial respiration or accumulation of defective mitochondria via impaired mitophagy [73]. In response to mitochondrial depolarization Fbxo7 can promote mitophagy through direct interaction with Parkin and Pink1 [98]. Under stress-induced conditions cytosolic Fbxo7 relocates to the mitochondria. Moreover, Fbxo7 participates in Parkin translocation to the mitochondria via direct physical interaction. Importantly, the loss of Fbxo7 expression results in a significant inhibition of Parkin recruitment to depolarized mitochondria and subsequent autophagic clearance of the dysfunctional organelle [98, 99].

Interestingly, overexpression of Fbxo7 in Drosophila Parkin mutants significantly rescued the Parkin depleted phenotypes, including loss of DA neurons, muscle degeneration, and mitochondrial disruption, but failed to rescue the effects of PINK1 silencing in mammalian cells [98]. These observations are compatible with recent reports that PINK1 plays additional roles in neuronal health through activation of cytosolic signaling pathways [100].

Several attempts to use CRISPR/Cas9 technology to model and treat PD-related conditions associated with impairment of Parkin/PINK1-dependent mitophagy have been made. However, knockout of Parkin and PINK1 in animal models of PD did not reproduce PD related behavior and pathological alterations observed in PD patients [81, 86–89]. These results suggest that PD pathology is much more complex and depends on the proper execution of many other mechanisms. Mutations and improper functioning of Parkin/PINK1 can launch mechanisms that compensate for the loss of their function and stabilize mitochondrial dynamics. Moreover, execution of mitophagy can occur independently from Parkin/PINK1 pathway through other mechanisms. One of many possible applications of CRISPR can be the unearthing such mechanisms and deciphering their role in PD. First, meta-analysis of previous research of such mechanisms and pathways should be performed with subsequent testing in vitro by generation of CRISPR gRNA libraries and their use in screening with CRISPR/Cas9. Interestingly, Burchell et al. showed that Fbxo7 overexpression rescued Parkin depletion - but not PINK1 silencing [98]. This suggests that other protein(s) can be involved in Parkin/PINK1-driven mitophagy and maintenance of mitochondrial functions or that PINK1 has independent functions regulating neuronal health [100–102]. Hence, revealing the respective genes and their protein products may be another challenging employment of CRISPR/Cas9 screening system.

3.2. α-synuclein

Missense mutations and duplications or triplications of the SNCA gene, which encodes for α-syn (Table 1), have been linked to PD with autosomal dominant inheritance and to sporadic forms of PD as well [57, 60, 61]. Elevated expression levels of endogenous α-syn induce mitochondrial dysfunction, characterized by impaired respiration and mitochondrial membrane depolarization [12]. Aggregation of endogenous α-syn monomers generates beta sheet-rich oligomers that localize to the mitochondria in close proximity to several mitochondrial proteins including ATP synthase. Oligomeric complexes are able to impair complex I dependent respiration, oxidize ATP synthase beta subunit, induce production of ROS that subsequently stimulates peroxidation of lipids in the MIM and neuronal death [12, 103, 104]. These pro-oxidant events increase the probability of permeability transition pore (PTP) opening, leading to equilibration of ionic gradients, swelling of the matrix, cristae unfolding, rupture of the MOM ultimately triggering neuronal death [12]. Inhibition of oligomer induced oxidation prevents the pathological induction of PTP [105]. Furthermore, suppression of lipid peroxidation by bis-allylic deuterated polyunsaturated fatty acids (D-PUFAs) protects glio-neuronal cell cultures from death [104]. α-Syn can also exert its pro-oxidant effects via interactions with a mitochondrial intermembrane space hemoprotein, cytochrome c (cyt c) (Fig. 2) [79]. This interaction is realized with the participation of a mitochondria-specific phospholipid, cardiolipin (CL). The latter forms a complex with cyt c that displays a peroxidase activity and oxidizes CL. Notably, α-syn can bind with cyt c/CL to yield a triple complex retaining a high pro-oxidant potential capable of damaging mitochondria. The presence of CL in the cyt c-CL complex is required to mediate cyt c interaction with α-syn. This unexpected pro-oxidant catalytic activity of α-syn represents a source of persistent oxidative stress in affected DA neurons that leads to chronic neurodegeneration. Interestingly, α-syn-cyt c aggregates were detected in SN DA neurons of rotenone-treated rats and LBs of PD patients [106].

In spite of these strong circumstantial pieces of evidence for a pro-oxidant role of role α-syn in mitochondrial and, possibly, extra-mitochondrial compartments, the intricacies of its redox partnerships still remain enigmatic. CRISPR/Cas9 genome editing technology can be utilized for direct modification of a SNCA gene transcriptional activity, thus, leading to stable and long-term effect on gene expression. α-Syn transcriptional activity is regulated through wide range of genetic and epigenetic mechanisms, particularly DNA methylation patterns. It was shown that increased α-syn expression levels can occur due to demethylation of CpG dinucleotides at SNCA intron 1 [107]. Interestingly, brains of PD patients show hypomethylation at SNCA intron 1 and therefore, high levels of α-syn. Kantor et al. developed a system for target editing of SNCA methylation profiles within intron 1. This group established a lentiviral vector, harboring gRNA, an engineered nuclease null form Cas9 (dCas9) fused with the catalytic domain of DNA-methyltransferase 3A (DNMT3A) to target hypomethylated CpG islands in the SNCA intron 1 region. The system was successfully applied to human induced pluripotent stem cell (hiPSC)-derived DA neurons from PD patients with the SNCA gene triplication. The study showed downregulation of SNCA mRNA levels, 25% decreased protein abundance in hiPSC-derived DA neurons compared to controls without lentiviral construct. Moreover, CRISPR/dCas9 epigenome editing influenced PD-related phenotype. Results demonstrated increased cellular viability and mitochondrial functioning by alleviating the susceptibility to oxidative stress [24]. While mutations in SNCA gene are likely causative to PD progression, the exact contribution of α-syn to PD development is not fully understood.

3.3. Cardiolipin-driven mitophagy

Dependently on the context of mitochondrial injury, mitophagy can also occur independently of Parkin/Pink1 [108–110]. One of the pathways triggering elimination of dysfunctional mitochondria is realized via signaling by a mitochondria-specific phospholipid, CL. Normally CL is synthesized and localized to the MIM, particularly its matrix leaflet. However, in depolarized mitochondria, CL is redistributed to the MOM [84]. This process includes CL transfers through the MIM, intermembrane space (IMS) and culminates in CL externalization on the mitochondrial surface where it is recognized by the microtubule associated protein 1 light chain 3 (LC3). The latter targets the injured mitochondria to autophagosomes and lysosomal degradation [111]. The transfer of CL through the IMS occurs with the assistance of the hexameric nucleoside diphosphate kinase D (NME4, NDPK-D, or NM23-H4) (Table 1), encoded by the NME4 gene (Fig. 2). RNAi knockdown of endogenous NDPK-D decreased CCCP-induced CL externalization and mitochondrial degradation [112]. Externalization of CL to the mitochondrial surface and enhanced mitophagy is triggered by a number of mitochondrial protonophoric uncouplers (such as CCCP) as well as an inhibitor of mitochondrial complex I, rotenone which is also known as a chemical PD inducer in experimental animals [83, 113].

Apart from CL binding and further externalization to MOM NDPK-D interacts with OPA1, a MIM-located dynamin-related GTPase, responsible for MIM fusion and mitochondrial quality control [114]. OPA1/NDPK-D interaction helps to fuel OPA1 with GTP to maintain OPA1 functions. Interestingly, OPA1 is a negative regulator of NME4/NDPK-D-supported CL transfer. Loss of NDP kinase activity induced by mitophagic stimuli weakens NME4/NDPK-D interaction with OPA1, thus releasing NME4/NDPK-D to crosslink MIM and MOM, facilitating CL transfer, resulting in the mitochondrial fragmentation observed during mitophagy [111, 114].

CL has the capacity to buffer synucleinopathy in SNCA mutated neurons (containing A53T and E46K mutations) by withdrawing α-syn monomers out of potentially pathological oligomers and fibrils [11]. α-Syn has been reported to bind to both mitochondrial membranes and mitochondria associated endoplasmic reticulum (ER) membranes [115, 116]. CL translocates to the MOM where it can facilitate α-syn refolding from aggregated β-sheet forms back to monomers comprising α-helices, effectively buffering synucleinopathy. The observed effects of CL on refolding of α-syn fibrils may be unique to CL-containing mitochondrial and mitochondria associated membranes. The increased abundance and prolonged CL exposure on the MOM in SNCA-mutants needed to refold mutant α-syn alter membrane dynamics and may initiate the depolarization of mitochondrial membranes, leading to increased LC3 recruitment to the MOM. Mutated α-syn had a significantly reduced ability to competitively inhibit LC3 binding to CL, as far as they compete for the same binding site on CL, thus increasing mitochondrial stress and triggering excessive mitophagy. Furthermore, co-culture of SNCA-mutant neurons with their isogenic controls results in transmission of mitochondrial pathology in non-mutant control neurons [11]. Thus, CL has two effects on α-syn: facilitating the formation of cyt c/CL-α-syn triple complexes with peroxidase activity capable of catalyzing prooxidant reactions, including those of covalent oligomerization of α-syn with cyt c and CL itself. These oligomers may become components of LBs relevant to the PD pathogenesis. Independently of this, CL by itself can suppress the formation of physical aggregates of α-syn thus counteracting the development of PD inducing reactions. One can imagine that CRISPR/Cas9 manipulated forms of α-syn as well as of cyt c may be instrumental in getting better insights in these conflicting roles of CL in PD development.

Intracellular quality control mechanisms are increasingly recognized as key factors in aging and NME4/NDPK-D-dependent mitophagy can also be an active player in this process [111]. Up to date no mechanistic studies on CL, NDPK-D in PD pathology have been conducted. The role of NDPK-D in CL redistribution and mitophageal signaling has been documented in animal models and cells treated with mitochondrial respiratory inhibitors (rotenone) or uncouplers (CCCP). However these results have not been confirmed on model animals that reproduce PD phenotype. The exact role of CL in modulation of α-syn pathology in relation to PD has not been established. It is also possible that abnormal α-syn itself may affect NME4/NDPK-D activity and facilitate CL translocation from MIM to MOM, thus promoting mitophagy.

CRISPR/Cas9 is a good tool to test whether NME4/NDPK-D activity is the only mechanism that promotes CL transgression to the MOM. Furthermore, creation of transgenic animal and cell models carrying various combinations of mutations in NDPK-D and α-syn may solve the puzzle of involvement of these genes and their protein products in PD pathogenesis.

3.4. ROS

Mitochondria synthesize ATP but also leak electrons to generate ROS, that are both cellular signaling molecules and damaging oxidants [117]. The generation and extensive accumulation of ROS in the mitochondria leads to the formation of mitochondrial PTP and mitochondrial membrane hyperpolarization. The damaged mitochondria then release the pro-apoptotic protein cyt c into the cytoplasm, where it activates nuclear fragmentation and caspase-3-dependent apoptosis [118]. In neurodegenerative diseases, ROS can cause damage to macromolecules such as proteins, lipids, polysaccharides, or nucleic acids in neurons. Abundance of fatty acids prone to peroxidation; high intracellular concentrations of transition metals (especially iron), capable of catalyzing the formation of reactive hydroxyl radicals; low levels of antioxidants; reduced capability to regenerate are the intrinsic properties of neurons, that make them highly vulnerable to the harmful effects of ROS. Furthermore, oxidative damage to proteins, lipids and DNA has been observed in post-mortem brain samples from PD patients [119].

3.5. Lipid mediators

Chronic pro-inflammatory immune activity is being increasingly associated with neurodegenerative disorders such as PD [14]. Oxidation of free polyunsaturated fatty acids (PUFAs) produces lipid mediators that exert diversified effects on the tissue homeostasis in health and disease [120]. Free PUFAs are normally esterified into cellular (phospho)lipids, but can be hydrolyzed and released by Ca-dependent phospholipases A2 (PLA2) and act as substrates for oxidation reactions catalyzed by several enzymes, especially lipoxygenases (LOXs), cyclooxygenases (COXs), cytochrome P450 isoforms and peroxidases [121]. A plethora of lipid mediators, including thromboxanes, pro-inflammatory prostaglandins and leukotrienes, regulators with both anti-inflammatory and proresolution activities, such as the lipoxins, resolvins and protectins are formed [122].

Recently a new pathway for selective generation of lipid mediators from a mitochondrial-specific phospholipid, CL, has been reported. This novel pathway localizes to mitochondria, and begins with the direct cyt c-catalyzed oxidation of PUFA-CL with subsequent hydrolysis of oxidation products, oxidized CL (CLox), by two types of Ca²⁺-independent phospholipase A2 (iPLA2) specific towards oxidatively modified phospholipids. The presence of sufficient amounts of oxidizing equivalents (H₂O₂ or lipid hydroperoxides) is essential for the peroxidase activity of cyt c/CL complexes and CL oxidation [120]. The brain has an unprecedented variety of CL species that can be utilized for the biosynthesis of diversified lipid mediators via cyt c-catalyzed oxidation processes [123]. Tyurina et al. demonstrated that this Ca²⁺-independent pathway, that triggers the oxidation of polyunsaturated CLs and accumulation of their hydrolysis products (oxidized linoleic, arachidonic acids and mono-lyso CLs), is activated in vivo after acute tissue injury [120].

Mitochondria are an active player in PD pathology. Long term exposure of mitochondria (especially those located in DA neurons of SN) to pro-oxidant conditions would cause CL externalization leading to its binding with cyt c to a peroxidase complex. Disrupted electron transport will inevitably generate superoxide anions that will dismutate directly or via Mn-SOD catalyzed reaction to yield H₂O₂. The latter can feed the peroxidase reaction of cyt c/CL complex to directly oxidize mitochondrial phospholipids, particularly CL. Among the oxidation products, there will be signals of apoptotic death (Fig. 3) as well as other oxygen-containing derivatives that can be hydrolyzed by Ca²⁺-independent PLA2γ (iPLA2γ) present in mitochondria (Fig. 3). These reactions will yield a plethora of pro- and anti-inflammatory lipid mediators – eicosanoids and docosanoids [120]. It is believed – but not experimentally proven – that these reactions may contribute to PD associated pathogenic processes.

Fig. 3. Schema showing a mitochondrial pathway for generation of lipid mediators.

Enzymatic oxidation of cardiolipin (CL) by cytochrome c (cyt c) in mitochondria during apoptosis and its hydrolysis by iPLA_2γ yields mono-lyso CL as well as a diversified series of pro-inflammatory and pro-resolving lipids mediators.

CRISPR/Cas9 protocol can be employed to test how Parkin, PINK1, Fbxo7, α-syn, CL, PLA2 and iPLA2 over-expression/downregulation/knockout affect the process of mitophagy, integrity of DA neurons and contribute to PD pathology.

4. Iron accumulation

Iron is essential for life as a part of many enzymes of oxidative metabolism. Yet free ionic iron or “loosely-bound” small molecular complexes of iron may be potentially toxic because of their very high redox activity, particularly with regards to their involvement in the poorly controlled production of ROS in oxygen-rich environments. Understanding mechanisms of iron metabolism in the brain is of great importance as iron may be involved in the pathogenesis of several neurodegenerative disorders.

Strictly controlled levels and activity of iron and ROS are crucial for normal cell and organismal function, although the aberrant accumulation of iron and excessive ROS production is recognized in a number of chronic degenerative conditions [124, 125]. In eukaryotic organisms, excessive mitochondrial iron-dependent production of ROS may result in cell death and damage to mitochondrial (mt)DNA by interfering with essential metabolic pathways [126].

Iron accumulation occurs in the brain tissues of aging animals, including humans, in areas primary responsible for the execution of motor and cognitive functions, particularly SN, globus pallidus and dentate nucleus [127]. The mechanisms underlying extensive iron accumulation in these structures likely involve changed balance between all major elements of the complex system fulfilling import, export, storage, redistribution and precise chaperon-mediated delivery of iron to the particular protein clients in neurons [128]. However, the exact knowledge about the role and contribution of individual components of this sophisticated machinery to the overall PD related iron dishomeostasis is incomplete [128]. Iron is believed to play a key role in PD pathogenesis. Excessive iron accumulation in SN of PD patients promotes death of DA neurons [9, 129, 130]. High sensitivity of these cells to various damaging agents can be related to the presence of dopamine, the major neurotransmitter released by DA neurons. Importantly, interaction between iron and dopamine can promote the production of various toxic metabolic intermediates and end-products (e.g. dopamine-o-quinone, amino-chrome, 5,6-indolequinone) that can be involved in redox-cycling reactions and cause mitochondrial dysfunction and α-syn oligomerization in DA neurons [131–137]. Notably, in vitro studies implicated elevated iron levels in triggering α-syn aggregation to toxic fibrils that are present in LBs [137, 138].

Tau protein associates with amyloid precursor protein (APP) (Table 1) and regulates its trafficking in neurons where APP stabilizes surface presentation of transmembrane iron exporter, ferroportin (FPN) (Table 1), and facilitates iron efflux from neurons (Fig. 4) [139–141]. Tau deficiency prevents APP from being transported to the neuronal surface that further leads to increase in intracellular APP levels in tau-deficient neurons, thus preventing it from interaction with FPN and impairs iron trafficking in DA neurons [141]. Iron accumulation within the neurons, especially DA neurons, involved in PD pathology, leads to neuronal death observed in in vivo and in vitro PD models and PD patients [140]. High amounts of insoluble tau levels with simultaneous reduction of soluble tau proteins is observed in familial and sporadic forms of PD even in individuals without dementia [140–142]. In a recent study, Lei et al. reported that tau-knockout mice develop iron accumulation and SN neuronal loss, with concomitant cognitive deficits and Parkinsonism. Importantly, changes in tau levels and iron accumulation were specific to the SN and were absent from any other brain regions [141]. Interestingly, accumulation of pathological neurofibrillary tangles of the microtubule-associated protein tau has been shown to contribute to PD pathology [143]. Moreover, tau protein encoded by MAPT gene (Table 1) was shown to be a risk factor for late-onset sporadic PD, possibly due to its involvement in regulation of axonal transport, synaptic function, DNA stabilization and protection from heat damage and oxidative stress. In 2017 Hallmann et al. used CRISPR/Cas9 genome editing in stem cell model of fronto-temporal dementia (FTD) to repair FTD-associated mutation in the MAPT gene. For the FTD stem cell model researchers utilized hiPSC-derived neural progenitor cells (NPCs) from patients carrying N279K MAPT mutation. It has been demonstrated that FTD NPCs-differentiated astrocytes with mutation in the MAPT gene expressed increased levels of 4R-TAU isoforms and showed changes in the transcriptome profiles as well as higher vulnerability to oxidative stress compared to the control group [144].

Fig. 4. Iron homeostasis in the brain.

Iron uptake in neurons occur via transferrin receptor 1 (TfR1), it is followed by the internalization of Tf/TfR1 complex inside the cell and iron dissociation from Tf. Divalent metal transporter 1 (DMT1) facilitates iron transfer into the cytosol, where cytosolic iron chaperones of PCBP family, PCBP1 and PCBP2, transfers iron to the appropriate enzymes and proteins. Excess iron can be stored inside the cavity of extracellular protein, ferritin, or exported by transmembrane iron carrier, ferroportin (FPN). The amyloid precursor protein (APP) that is transported to the cell membrane by microtubule-associated protein tau is essential for FPN stabilization at the neuronal surface; therefore facilitating iron efflux from neuronal cells. Pathological accumulation of iron due to impairment in iron homeostasis can provoke iron-dependent oxidations of DNA, proteins and lipids that threaten cell health and integrity. Treatment with iron chelators might reduce excessive iron accumulation and restore iron balance in cells.

Ceruloplasmin (Cp) participates in iron transport by oxidizing Fe²⁺ to Fe³⁺ resented by FPN. Dysfunction of Cp is widely observed in PD-affected SN and suggests it may contribute to the pathogenic pro-oxidant iron accumulation [145]. Cp is also essential for transferrin (Tf) dependent delivery of iron into neuronal cells through that specific binding to the Tf receptors (TfR1) (Table 1) on the cell surface, which triggers endocytosis of Tf-TfR1 complex [146]. Endosomal iron can then be transported across the membrane to the cytosol by endosomal divalent metal transporter (DMT1) (Fig. 4) [147]. Tf is required for both iron uptake and efflux; when the cell requires iron for the proper functioning Tf binds to TfR1, promoting iron import into the cell, alternatively Tf removes iron from the FPN/Cp complex when the cell is iron replete and redistributes excess iron throughout the body [148]. Depletion of Tf in serum and SN has been described in PD as a contributor to iron accumulation in the brain. Furthermore, Tf is mis-compartmentalized in the mitochondria of SN neurons in PD that could further contribute to aberrant regulation of Tf in the brain [148].

In the cytosol, iron may be transported by binding to metallochaperones that deliver iron to target apoproteins [127]. One of these chaperones, poly(rC)-binding protein 1 (PCBP1), transfers iron to ferritin, an intracellular protein that detoxifies excess Fe²⁺ by oxidizing and storing the resultant Fe³⁺ inside its cavity [127, 149]. PCBP1-depleted cells show reduction in ferritin iron loading and simultaneous increase in cytosolic iron levels [150]. Another protein from the PCBP family, PCBP2, acquires ferrous iron from DMT 1 and transfers it to the appropriate enzymes and proteins in iron-PCBP1/PCBP2 complex. Importantly, PCBP2 facilitates iron export by binding to FPN [151]. PCBP1 and PCBP2 may function as iron chaperones for enzymes containing mono- and dinuclear iron centers, especially the family of non-heme iron-containing dioxygenases such as LOXs – enzymes involved in regulation of ferroptotic cell death program (see below). Depletion of one or both PCBPs corresponds with impaired iron delivery to the target cytosolic proteins [152].

Iron homeostasis in all mammalian cells is regulated by cytosolic iron response proteins (IRP½). In neurons IRP2 is the main sensor of labile intracellular iron, it binds to transcripts of ferritin, TfR1 and other target genes to control the expression of iron metabolizing proteins at the post-transcriptional level [9, 153]. In normal DA neurons and other cells under iron-depleted conditions iron-regulatory elements (IREs) (Table 1) located on the 3’-UTR of DMT1 and TfR1 mRNAs stabilize these mRNAs by IRP binding followed by enhanced translation of proteins. Alternatively, when intracellular iron levels are elevated, IRPs bind to IREs in the 5’-UTR of DMT 1, ferritin, APP and α-syn mRNAs, thus leading to repression of translation from these mRNAs [9]. Jiang et al., showed that impairment of IREs regulation through IRPs binding promoted DMT1 upregulation in the SN of 6-hydroxydopamine (6-OHDA)-induced PD rats. Overexpression of DMT1 caused increased iron influx, resulting in damaged mitochondrial function and increased ROS generation [154].

Iron chelators might reduce harmful iron-dependent oxidations of DNA, proteins and lipids [155]. For instance, clioquinol specifically reduces iron levels in the SN and prevents the onset of both motor dysfunction and cognitive decline in tau-deficient mice [141]. Unfortunately, treatment with iron chelators such as clioquinol, deferiprone or deferasirox could also remove iron from tissues unaffected by the disease or promote other side effects. Among the side-effects of iron-chelating compounds are nephrotoxicity, reduction of striatal dopamine, and neutropenia or agranulocytosis [148].

CRISPR/Cas9-based drug screening technologies may contribute to the development of new small molecule regulators that will target pathological iron accumulation in PD more selectively and effectively. Importantly, CRISPR/ Cas9 drug screening strategy can be applied for personalized treatment of PD patients. Based on unique array of genetic variants and specific disturbances of every PD patient, appropriate treatment strategies and medicines can be designed targeting the specific disease drivers in a given patient.

Overall, dysregulated iron homeostasis in neurons and glia of the SN is an established feature of PD. Iron homeostasis is regulated through multiple pathways and cellular mechanisms. APP and tau protein regulate FPN functioning and stabilize FPN in the cellular membrane to support iron export from the cell. Moreover, many other proteins, Tf, TfR1, DMT1, PCBPs and ferritin, are important for proper iron metabolism inside the cell. Interestingly, translational levels of these iron homeostatic proteins are regulated by intracellular iron levels and specific regulatory elements IRPs and IREs. The impairments in these mechanisms can lead to dysregulation of DMT 1, FPN, Tf, ferritin or APP translational activity and iron accumulation. CRISPR/Cas9 epigenome editing technology can be used for validation of the effects of up- and down-regulation of transcriptional and translational activity of iron homeostatic genes and their protein products in DA neurons. However, dysregulation of iron levels can also occur due to mutations in the reported genes, which can influence normal protein functioning and underlie the abnormalities in iron homeostasis. Disturbances in some of these proteins are essential and sufficient for development of PD phenotype, while impairments in others are just the consequences. However, which ones are really involved in PD is not clear. CRISPR/Cas9 multiple screening approaches can assist in separating the causative iron homeostatic genes associated with the PD pathology from those that are not directly related to the PD pathogenesis.

5. Programmed Cell Death

5.1. Apoptosis

Apoptotic death of DA neurons has been implicated in PD [156]. In the context of this review, we would like to focus on the mechanisms of intrinsic apoptosis because its early stages take place in mitochondria, strongly depend on changes in redox environment and include CL oxidation as a required stage of the execution of the entire program [157]. As has been described above, mitochondrial injury triggers trans-migration of unoxidized CL from the matrix side of the MIM to the mitochondrial surface thus signaling mitophagy and elimination of dysfunctional mitochondria. These pro-survival pathways may be effective in maintaining mitochondrial health within the cells. However, if the injurious factors continue to progress and the cell cannot cope with the intensity of the impact via mitophagy, the next initiated program is apoptotic elimination of the “high risk” cell continuously producing massive ROS. This stage is characterized by the formation of cyt c/CL complexes with a strong peroxidase activity towards bound CLs [10]. Within these complexes, CL induces substantial conformational changes and mobility of the protein such that hexa-coordinated iron loses one of the coordination bonds (Fe-S/Met80) and converts into a penta-coordinated state capable of binding small molecule oxidants, such as H₂O₂. The latter may be provided by the disrupted electron transport chain whereby cyt c interacting with CL obtains a highly negative redox potential precluding the acceptance of electrons from the respiratory complex III [158]. As a result, the flow of electron may be re-directed towards other acceptors, among them molecular oxygen to yield superoxide anion-radical. Spontaneous or MnSOD-catalyzed dismutation of the latter produces H₂O₂ that feeds the catalytic peroxidase cycle of cyt c/CL complexes and perpetuated CL peroxidation.

Several small molecule regulators such as mitochondria-targeted GS-nitroxides and imidazole-substituted fatty acids have been employed to effectively suppress the peroxidase CL oxidation and apoptosis [159]. However, their effectiveness in the context of PD has not been assessed. Because PD pathogenesis has been strongly associated with apoptosis and accumulation of specific CL peroxidation products (Fig. 5), it seems reasonable to suggest that optimized combination of small molecules inhibitors may be effective in anti-PD therapy. However, such a development may require detailed genetic studies unequivocally identifying all the participating protein-based mechanisms. These include above mentioned transmembrane transporters of CL to the mitochondrial surface, electron transporters generating and providing H₂O₂, kinases participating in phosphorylation of cyt c thus determining its affinity towards interactions with CL, protein mechanisms of CL biosynthesis and remodeling and numerous mitochondrial proteins capable of binding CL hence “out-compete” cyt c. Obviously, this extensive and time-consuming research may benefit from the utilization of the effective CRISPR/Cas9 technology.

Fig. 5. Cardiolipin oxidation products in substantia nigra.

Figure shows accumulation of specific cardiolipin oxidation (CLox) products in substantia nigra (SN) using rat rotenone model of PD [164], Full MS spectra of CL (left panel) and CLox (right panel) from SN of rats exposed to rotenone (10 days after treatment). Insert: Significantly higher levels of CLox molecular species were detected in SN of rotenone treated rats and associated with key pathological features of PD such as bradykinesia, postural instability/gait disturbances and rigidity.

5.2. Ferroptosis – Overview

Ferroptosis is a form of iron-dependent, oxidative cell death with specific role of lipid hydroperoxides accumulating in conditions of GPX4 insufficiency [160, 161]. Ferroptosis can be inhibited by lipophilic radical scavengers (eg, ferrostatin-1, different homologues of vitamin E, and Trolox), by inhibitors of LOX (eg, baicalein) and by iron chelators (deferoxamine) [161, 162]. Ferroptosis has been implicated in the pathological cell death associated with degenerative diseases, including PD [163]. The sensitivity to ferroptosis is strongly dependent on the access to oxidizable PUFA and their esterification in specific phospholipids, iron homeostasis, as well as on the sufficiency of the thiols.

Glutathione peroxidase 4 (GPX4) (Table 1) catalyzes the reduction of lipid peroxides to lipid alcohols. Inactivation of GPX4 through depletion of reduced glutathione (GSH) with erastin, or with a direct GPX4 inhibitor RSL3, ultimately induces intensive lipid peroxidation that causes cell death [165]. Hydroperoxy derivatives of PUFA-PEs also cause ferroptotic death when added to cells with inactivated GPX4 [166]. GSH, a cofactor of GPX4, is an essential intracellular antioxidant synthesized from glutamate, cysteine and glycine by cytosolic glutamate-cysteine ligase (GCL) and glutathione synthetase (GSS) (Fig. 6a) (Table 1). Due to the overwhelming importance of several amino acids for biological activity of GPX4, amino acid metabolism is tightly linked to the regulation of ferroptosis [167].

Fig. 6. Ferroptotic cell death in neurons.

a. Ferroptosis is triggered by accumulation of lipid hydroperoxides (PL-OOH). Glutathione peroxidase 4 (GPX4) catalyzes the reduction of lipid peroxides to lipid alcohols (PL-OH). Inactivation of GPX4 through depletion of reduced glutathione (GSH) with erastin, or with the direct GPX4 inhibitor RSL3, ultimately induces intensive lipid peroxidation that causes ferroptotic cell death. GSH, a cofactor of GPX4, is an essential intracellular antioxidant synthesized from glutamate (that is exchanged for cystine by system x_c-), cysteine and glycine by cytosolic glutamate-cysteine ligase (GCL) and glutathione synthetase (GSS). Lipoxygenases (LOXs), nonheme, iron-containing proteins, normally utilizes free PUFAs as their substrates.

b. LOX substrate specificity is switched from free PUFAs to PUFA-containing polyunsaturated phosphatidylethanolamines (PEs). Phosphatidylethanolamine-binding protein 1 (PEBP1) complexes with 15-LOXs and changes their substrate competence to produce hydroperoxy-PE (PE-OOH). Inadequate reduction of PE-OOH due to dysfunction of GPX4 leads to ferroptosis.

Cysteine availability limits the biosynthesis of glutathione, furthermore, some cells can bypass the requirement for cystine import via the cystine/glutamate antiporter system x_c- by converting methionine to cysteine through transsulfuration pathway [165]. Glutamate and glutamine are also important for ferroptotic cell death regulation [168]. Glutamate is exchanged for cystine by system x_c-, so high extracellular concentrations of glutamate inhibit system x_c- and induce ferroptosis [160]. Thus, accumulation of extracellular glutamate could serve as a natural trigger for inducing ferroptosis in physiological contexts. Free PUFAs are substrates for synthesis of lipid signaling mediators, yet they must be esterified into membrane phospholipids and undergo oxidation to become ferroptotic signals [166]. Previous studies suggest that phosphatidyl-ethanolamines (PEs), containing arachidonic acid (AA) or its elongation product, an isomer of docosatetraenoic acid, adrenic acid (AdA), are key phospholipids that undergo oxidation and provoke ferroptotic cell death [166, 169].

Acyl-CoA synthetase long-chain family member 4 (ACSL4) (Table 1) is involved in the biosynthesis and remodeling of PUFA-PEs in cellular membranes and, furthermore, promotes cell sensitivity to ferroptosis. Loss of function mutations in ACSL4 deplete the substrates for lipid peroxidation and increase resistance to ferroptosis [169, 170]. Alternatively, cells that are supplemented with arachidonic acid or other PUFAs are sensitized to ferroptosis [170].

LOXs (Table 1), non-heme, iron-containing proteins, utilize free PUFAs as their substrates (Fig. 6a) [171]. Interestingly, LOX substrate specificity can be switched from free PUFAs to PUFA-containing PEs (Fig. 6b). The mechanism that alters the substrate specificity of LOX remains to be enigmatic; however for 15-lipoxygenases (15-LOXs) the requirement in Phosphatidylethanolamine-Binding Protein 1 (PEBP1) has been demonstrated.PEBP1 forms complexes with 15-LO1 and 15LO2, two isoforms of 15LOX and changes their substrate competence to produce hydroperoxy-PE [172]. Notably, increased levels of lipid peroxidation products have been consistently observed in the SN of PD patients and in the brains of PD animal models [173, 174].

Several studies reported potential links between ferroptosis and the pathogenesis of neurodegenerative diseases. Inhibitors of ferroptosis, ferrostatins and liprox-statins, were found to be effective protectors against neurodegeneration in models of PD, Huntington’s, and Alzheimer’s disease models [163, 175, 176]. In line with the key role of GPX4 in ferroptosis, insufficiency of the enzyme – due to chemical inactivation, deletion or loss of function mutations - cause ferroptotic cell death characterized by extensive lipid peroxidation. Therefore, CRISPR/Cas9 genome/epigenome editing aimed at the restoration of the GPX4 activity might be beneficial. Another application of the CRISPR/Cas9 genome editing technology may be promising in studies of 15-LOX/PEBP1 interactions in determining selectivity and specificity in PUFA-PE oxidation and switching the substrate specificity from free to esterified PUFA. Oxidation of PUFA-PE by 15-LOXs is considered as the key mechanism that generates lipid signals of ferroptotic cell death. There are, however, studies that implicate other isoforms of LOXs (3-5-, 12-) in triggering ferroptotic death [177–179]. CRISPR/Cas9 silencing or activation of respective LOX isoforms may give more quantitative description of their role and significance in lipid peroxidation, execution of ferroptosis and involvement in PD pathology.

Empirical research has established many potential metabolic abnormalities that may represent the specific key mechanisms of PD pathogenesis. However, the diversity of these findings and the lack in understanding the connections between them slow down the progress in the development of specific treatments. Several papers in a recent special issue of Science (December 14, 2018) “Illuminating the Brain” provided new insights into an integrative functional genomic/transcriptomics analysis, a powerful approach to gain comprehensive information on the entire the genome as well as epigenomic and transcriptomic features of the healthy and diseased (mostly psychiatric conditions) of human brain across various regions and cell types [180–182]. The success of this approach in several neuropsychiatric conditions (e.g. schizophrenia, autism spectrum disorder (ASD), bipolar disorder) assures that that it will also be applied to PD patients. Simultaneously, new advancements based on CRISPR/Cas9 genome/epigenome editing protocols are being rapidly developed. As in sporadic PD, individual genomic variations are highly specific, patient-derived iPSCs may provide important personalized insights into the genetic basis of PD heritability, facilitate creation of patient-specific experimental cell models, drug discovery and develop new strategies for effective treatment [27, 183–186].

Among many potentially important targets for CRISPR/Cas9 based research we focused in this brief review on aberrant redox mechanisms and pathways related to redox metabolism, mitochondria and regulated cell death programs that have been identified as strongly associated with the PD pathogenesis, yet without definitive conclusions leading to design of new therapies. With the appreciation of the significant role of the post-translational modifications of proteins and lipid protein interactions, we are optimistic that the power and relative simplicity of the CRISPR/Cas9 genome editing and screening protocols will generate new information on metabolic impairments essential for PD pathogenesis. This enthusiasm is supported by the fact that employment of genome-editing technology has already generated new animal models in which the introduction of custom-made modifications into the genome replicated key features of PD [26].

Highlights:

• CRISPR/Cas9 genome editing in Parkinson disease

• Redox lipidomics of signaling in cell death programs

• Pathogenic impairments of iron regulation

• New therapies of neurodegeneration

• Cardiolipin signals of mitochondrial injury