Genome Integrity and Neurological Disease

Neurological complications directly impact the lives of hundreds of millions of people worldwide. While the precise molecular mechanisms that underlie neuronal cell loss remain under debate, evidence indicates that the accumulation of genomic DNA damage and consequent cellular responses can promote apoptosis and neurodegenerative disease. This idea is supported by the fact that individuals who harbor pathogenic mutations in DNA damage response genes experience profound neuropathological manifestations. The review article here provides a general overview of the nervous system, the threats to DNA stability, and the mechanisms that protect genomic integrity while highlighting the connections of DNA repair defects to neurological disease. The information presented should serve as a prelude to the Special Issue “Genome Stability and Neurological Disease”, where experts discuss the role of DNA repair in preserving central nervous system function in greater depth.


Nervous Systems and the Brain
Nervous systems within the animal kingdom vary tremendously in structure and complexity [1,2]. Simple nervous systems can range from the non-centralized neural network in jellyfish to sea sponges who lack a true nervous system altogether. Unlike most invertebrate species, the nervous systems of vertebrates are complex, centralized, and specialized. While there is diversity among vertebrates, they share a fundamental nervous system structure, i.e., a central nervous system (CNS) that consists of a brain and spinal cord and a peripheral nervous system (PNS) composed of peripheral sensory and motor nerves. As a whole, the nervous system works coordinately to receive and transmit complex information throughout the body, directing all aspects of thought, behavior, and physical activity. The brain serves as the master receiver, organizer, and distributor of sensory and visual information from the body, controlling our thoughts, memory and speech, physical movements, and the function of most organs [3,4]. Neurons are the basic functional units of the brain, producing electrical pulses to transmit information across great distances within the interconnected circuitry of the body. Glia cells, i.e., astrocytes, oligodendrocytes, and microglia, primarily operate in the mature CNS to support neurons [5], with data indicating unique and specialized roles for neuroglia in preserving the proper operation of the nervous system. Astrocytes function to maintain brain homeostasis and neuronal metabolism by producing antioxidants, recycling neurotransmitters, and supporting the blood-brain barrier. Oligodendrocytes are responsible for generating the insulating myelin sheath around axons, enabling easy and efficient transfer of electrical signals between nerve cells. Microglia are specialized macrophages and operate to remove damaged neurons or infections to safeguard brain health. Disruption of any one of these components can lead to biochemical and cellular abnormalities that give rise to CNS dysfunction and disease.

Neurodegenerative Diseases
The National Institute of Neurological Disorders and Stroke within the United States of America lists over 400 distinct neurological diseases on its website, and it is almost certainly not exhaustive (https://www.ninds.nih.gov/disorders/all-disorders, accessed on 2 March 2022). These include neurogenetic diseases, developmental disorders, degenerative diseases, and trauma or infection-induced diseases. Neurodegenerative diseases, in particular, involve the progressive loss or death of nerve cells and directly affect the lives of hundreds of millions of people worldwide, encompassing diseases such as Alzheimer disease (AD), Parkinson disease (PD), Huntington disease (HD), and amyotrophic lateral sclerosis (ALS). The symptomatic degeneration and accompanying brain atrophy present in these disorders result in impaired memory or thought processes, confusion and anxiety, mood swings, poor coordination, and/or loss of movement control. While the causes of neurodegenerative diseases are very heterogeneous in nature and are still being deciphered, the most common risk factor is age, a phenomenon known to involve DNA damage and genomic instability, epigenetic alterations, loss of proteostasis, mitochondrial dysfunction, and altered intercellular communication, to name a few of the hallmarks [6]. In combination with the numerous basic research and clinical investigations conducted to date, current evidence, therefore, implicates the above molecular mechanisms, as well as oxidative stress, protein aggregation, cytoskeletal abnormalities, synaptic dysfunction, and cell death pathways as underlying drivers of neurodegenerative disease [7]. This Special Issue is dedicated to more exhaustively enumerating the link between DNA damage, genomic stress, neuronal cell loss, and neurological disease.

Exogenous and Endogenous Threats to the Brain: Oxidative DNA Damage
The compositional integrity of organisms is constantly challenged by a barrage of insults from both external (exogenous) and internal (endogenous) sources, which often possess the ability to generate lipid, protein, RNA, and/or DNA damage [8]. Some of the most common external sources include sunlight, ionizing radiation (IR), and chemicals or pollutants found in the environment. In addition, many agents used as clinical therapeutics have either intended or incidental harmful effects on macromolecule integrity [9]. Given that damage to the first three cellular components (i.e., lipids, proteins, and RNA) can be resolved by degradation and molecular turnover, alterations to the integrity of the genetic blueprint (DNA) are thought to be the most threatening to organism function and health. Beyond the array of external chemical and physical agent exposures, exogenous stress can also include physical trauma (e.g., accidents), lifestyle choices (e.g., drug or alcohol use), or unexpected changes in one's personal life or mental health (e.g., due to the loss of a loved one, work problems, etc.) [10]. Collectively, the impact of external stress can be significant, leading to deleterious changes in the composition of macromolecular structures and the effectiveness of biochemical processes.
In addition to the numerous external threats, several endogenous agents can cause macromolecular damage [11], including to DNA. Perhaps the best known intracellular damaging agents are reactive oxygen species (ROS), formed as natural by-products during mitochondrial respiration in the process of generating energy (ATP). ROS can create a range of base and sugar modifications upon reacting with DNA, some of which promote strand breakage. Through reactions with phospholipids, ROS can generate epoxide or aldehyde species, which can, in turn, react with DNA to produce severe genomic lesions, such as interstrand crosslinks that covalently connect the two strands of the helix [12]. Apart from ROS-induced DNA damage, S-adenosylmethionine, a common co-substrate involved in methyl group transfers, including during epigenetic landscaping, has been found to react with DNA, producing alkylated base products such as N3-methyladenine and O6-methylguanine [13]. Lastly, DNA, as a chemical, exhibits intrinsic instability, mainly in the form of deamination of cytosine to uracil or the loss of purine or pyrimidine bases, creating an abasic site that lacks instructional coding information. Collectively, it has been estimated that a typical human genome will experience over 100,000 alterations per day under normal physiological conditions [8].
ROS are not only continuously being produced and leaked by mitochondria but are also commonly induced by exogenous stressors, such as those described above. Indeed, IR is well known for generating ROS through the radiolysis of water [14], a fact that seeded the free-radical theory of aging proposed by Denham Harman in 1952 [15]. He argued that aging and progressive degenerative disease occur due to the gradual accumulation of harmful oxidative damage during the lifetime of an organism. Although the free radical theory has encountered setbacks, the preponderance of evidence supports the notion that oxidative macromolecular damage accumulates with age [16]. Since the discovery of the intrinsic instability of DNA and the role of genomic stress in disease, many researchers have pursued the idea that the DNA template is the primary pathological target of ROS.
It is important to emphasize that, unlike the PNS, the CNS is protected from most exogenous threats by the vertebral column that encircles the spinal cord and the skull that houses the brain [17]. In addition, the CNS is shielded from many circulating external factors, such as toxins, by the blood-brain barrier [18]. Thus, unlike the PNS, which appears to be more susceptible to the deleterious effects of external damaging agents, the CNS is mainly challenged by endogenous stressors. The brain is one of the most active and energetically demanding organs in the human body, consuming roughly 49 mL of oxygen per minute, with nearly all oxygen being utilized for the oxidation of carbohydrates (almost exclusively glucose) to produce high-energy ATP [19]. As oxidative phosphorylation via the electron transport chain in mitochondria is the primary source of ATP generation, the risk of ROS production is high in neuronal cells [20]. The primary front-line defense against ROS is comprised of a battery of scavenging molecules and enzymes, collectively known as antioxidants [21]. Thus, how much oxidative damage is generated depends on the balance between ROS production and neutralization by antioxidants. Since the brain has been reported to maintain a comparatively low level of antioxidant systems, there is an increased need for efficient oxidative damage coping mechanisms, including DNA repair processes that resolve ROS-induced DNA damage and preserve genome integrity [22].

DNA Repair Mechanisms
Since evidence implicates oxidative stress and associated macromolecular damage as a critical element in the etiology of neurodegenerative disease, we focus the remainder of the review on DNA damage and DNA repair. Oxidative DNA damage includes a range of alterations, such as base modifications (e.g., 8-oxoguanine or cyclopurines), protein-DNA adducts, DNA-DNA crosslinks, abasic sites, strand breaks, and clustered lesions [23,24]. Depending on the chemical, structural, and positional nature, DNA damage can promote mutagenesis, genomic instability, or stress responses triggered upon replication or transcription arrest [25] (Figure 1). Considering the non-dividing nature of terminally differentiated neurons, it is likely the loss of genome integrity or the persistent stalling of RNA polymerase II at blocking damage that activates a cell death response, giving rise to neurodegeneration. Additionally, unresolved DNA damage may promote cell cycle activation and reentry via a DNA damage response mechanism, leading to apoptotic cell death as well [26].
As a means of understanding the fundamental contributions of the different DNA repair pathways (and sub-pathways), we provide next an overview of the major corrective mechanisms in humans: (1) direct reversal, (2) mismatch repair (MMR), (3) base excision repair (BER), (4) single-strand break repair (SSBR), (5) ribonucleotide excision repair (RER), (6) nucleotide excision repair (NER), and (7) double-strand break repair (DSBR). In general, excluding direct reversal (see more below), DNA repair involves some variation of the following biochemical steps: (a) damage recognition; (b) damage processing, excision, or resolution; (c) where necessary, replacement of the removed genetic material; and (d) strand break closure. We describe the nuclear DNA repair mechanisms below, yet it is important to recognize that separate, albeit often related, corrective pathways exist within the mitochondrial compartment to deal with mitochondrial DNA (mtDNA) damage [27]. Table 1  As a means of understanding the fundamental contributions of the different DNA pair pathways (and sub-pathways), we provide next an overview of the major correc mechanisms in humans: (1) direct reversal, (2) mismatch repair (MMR), (3) base exci repair (BER), (4) single-strand break repair (SSBR), (5) ribonucleotide excision repair (R (6) nucleotide excision repair (NER), and (7) double-strand break repair (DSBR). In gen excluding direct reversal (see more below), DNA repair involves some variation of the lowing biochemical steps: (a) damage recognition; (b) damage processing, excision, or r lution; (c) where necessary, replacement of the removed genetic material; and (d) str break closure. We describe the nuclear DNA repair mechanisms below, yet it is importa recognize that separate, albeit often related, corrective pathways exist within the mitoc drial compartment to deal with mitochondrial DNA (mtDNA) damage [27]. Table 1 i cluded as a reference of the DNA repair pathways, the central protein components and m biochemical activities, and the connections to inherited (neurodegenerative) disorders.

Direct Reversal
In direct reversal, the biochemical event entails the precise removal of subtle base modifications from DNA, immediately restoring the genome to its original, undamaged state ( Figure 2). For instance, it was discovered early during DNA repair investigations that organisms possess an evolutionarily conserved mechanism for resolving O 6 -methylguanine adducts. The human protein, termed O 6 -methylguanine DNA methyltransferase (MGMT), like its bacterial functional counterparts (i.e., Escherichia coli Ada and Ogt), utilizes a conserved active site cysteine residue to directly remove alkylation damage from DNA [28][29][30]. This repair step results in the permanent inactivation of the protein and its consequent degradation, commonly referred to as a "suicide" mission [31].

Mismatch Repair
To cope with mispaired nucleotides or small insertions/deletions that arise as DNA replication errors, organisms have evolved the MMR system [38] (Figure 3). MMR, which is likely intimately connected to the replication machinery to ensure faithful removal of the newly synthesized material, entails a scanning and recognition step, followed by the excision of the portion of the daughter strand that contains the incorrectly inserted nucleotide or bulge/loop structure. A complex of MutS homologs carries out damage detection, i.e., either MSH2/MSH6 (MUTSɑ, for base-base mismatches and small loops) or MSH2/MSH3 (MUTSβ, for small and large loops) [39,40]. Subsequent excision of the misincorporated nucleotides is mediated by exonuclease 1 (EXO1) or the MUTLɑ complex (MLH1 and endonuclease PMS2), in coordination with the MUTS recognition complex, replication protein A (RPA), replication factor C (RFC), proliferating cellular nuclear antigen (PCNA), and DNA polymerase δ (POLδ) [41][42][43][44]. Following damage excision, the remaining steps include re-replication of the now missing section of DNA (likely by the replicative polymerase POLδ) and closure of the remaining nick (likely by ligase 1 (LIG1) in cooperation with PCNA). In addition, recent investigations have uncovered nine human proteins (ALKBH1-8 and FTO), sharing homology with the bacterial DNA repair enzyme ALKB as part of the family of Fe(II)-and α-ketoglutarate-dependent dioxygenases [32,33]. Of the nine paralogs, ALKBH2 and ALKBH3 are bonafide nuclear DNA repair enzymes, possessing the ability to directly remove 1-methyladenine, 1-ethyladenine, 3-methylcytosine, and specific etheno adducts from damaged polynucleotides, sometimes with single-stranded specificity, in an α-ketoglutarate-dependent reaction [34][35][36]. ALKBH1 is also a nucleic acid demethylase with the ability to target both RNA and DNA substrates but localizes to the mitochondria [37]. Most of the other ALKB homologs demethylate RNA or proteins, functioning in gene expression regulation or epigenetic programming, or have biochemical activities that have yet to be defined.

Mismatch Repair
To cope with mispaired nucleotides or small insertions/deletions that arise as DNA replication errors, organisms have evolved the MMR system [38] (Figure 3). MMR, which is likely intimately connected to the replication machinery to ensure faithful removal of the newly synthesized material, entails a scanning and recognition step, followed by the excision of the portion of the daughter strand that contains the incorrectly inserted nucleotide or bulge/loop structure. A complex of MutS homologs carries out damage detection, i.e., either MSH2/MSH6 (MUTSα, for base-base mismatches and small loops) or MSH2/MSH3 (MUTSβ, for small and large loops) [39,40]. Subsequent excision of the misincorporated nucleotides is mediated by exonuclease 1 (EXO1) or the MUTLα complex (MLH1 and endonuclease PMS2), in coordination with the MUTS recognition complex, replication protein A (RPA), replication factor C (RFC), proliferating cellular nuclear antigen (PCNA), and DNA polymerase δ (POLδ) [41][42][43][44]. Following damage excision, the remaining steps include re-replication of the now missing section of DNA (likely by the replicative polymerase POLδ) and closure of the remaining nick (likely by ligase 1 (LIG1) in cooperation with PCNA).

Ribonucleotide Excision Repair
Driven by the observation that as many as 1 million or more ribonucleotides can be incorporated during a single round of chromosome replication, recent attention has been put into deciphering the molecular steps of RER [45,46] (Figure 3). In short, RER looks a lot like long-patch BER (see next), except that the initial incision event is catalyzed by RNaseH2, a heterotrimeric complex that also degrades RNA:DNA hybrids [47]. After strand cleavage 5′ of the faulty ribonucleotide, the replicative polymerases (POLδ/POLε) carry out strand displacement synthesis with RFC and PCNA, followed by removal of the 5′-flap containing the damage via flap endonuclease 1 (FEN1) and nick ligation by LIG1 [48]. There appears to be an alternative topoisomerase 1 (TOPO1)-mediated processing mechanism, which we will not discuss here.

Base Excision Repair
BER is the primary pathway for coping with spontaneous decay products (e.g., apurinic/apyrimidinic (AP) sites for base loss or uracil from cytosine deamination) and oxidative base or sugar lesions [49] (Figure 4). Classic BER begins with the removal of a substrate base by a DNA glycosylase, resulting in the generation of an AP site intermediate.

Ribonucleotide Excision Repair
Driven by the observation that as many as 1 million or more ribonucleotides can be incorporated during a single round of chromosome replication, recent attention has been put into deciphering the molecular steps of RER [45,46] (Figure 3). In short, RER looks a lot like long-patch BER (see next), except that the initial incision event is catalyzed by RNaseH2, a heterotrimeric complex that also degrades RNA:DNA hybrids [47]. After strand cleavage 5 of the faulty ribonucleotide, the replicative polymerases (POLδ/POLε) carry out strand displacement synthesis with RFC and PCNA, followed by removal of the 5 -flap containing the damage via flap endonuclease 1 (FEN1) and nick ligation by LIG1 [48]. There appears to be an alternative topoisomerase 1 (TOPO1)-mediated processing mechanism, which we will not discuss here.

Base Excision Repair
BER is the primary pathway for coping with spontaneous decay products (e.g., apurinic/apyrimidinic (AP) sites for base loss or uracil from cytosine deamination) and oxidative base or sugar lesions [49] (Figure 4). Classic BER begins with the removal of a substrate base by a DNA glycosylase, resulting in the generation of an AP site intermediate. In humans, there are 11 known DNA glycosylases: uracil DNA glycosylase (UNG), methylpurine DNA glycosylase (MPG), nth endonuclease III-like (NTHL1), 8-oxoguanine DNA glycosylase (OGG1), mutY homology (MUTYH), thymine DNA glycosylase (TDG), single-strand-selective monofunctional uracil DNA glycosylase (SMUG1), methyl CpG binding domain protein 4 (MBD4), and the nei endonuclease VIII-like proteins (NEIL1, NEIL2, and NEIL3) [50]. The AP site intermediate is then primarily incised by APE1, creating an SSB with a priming 3 -hydroxyl group and a damaged 5 -abasic fragment [51]. The missing nucleotide is then replaced, and the 5 -end processed to create a regular 5 -phosphate group to permit the ligation and completion of the repair response. The primary gap-filling polymerase in humans is POLβ, which also possesses the ability to remove 5 -abasic groups (i.e., the deoxyribose phosphate) left behind by APE1 incision [52]. X-ray cross-complementing protein 1 (XRCC1) in complex with DNA LIG3α plays a role in sealing the nick during classic BER [53,54]. Besides the single-nucleotide (a.k.a., shortpatch) BER event described above, humans possess a long-patch BER process that entails strand-displacement synthesis, typically by a replicative polymerase (POLδ/POLε) in collaboration with PCNA, 5 -flap resolution by FEN1, and nick sealing by LIG1 [55]. NEIL2, and NEIL3) [50]. The AP site intermediate is then primarily incised by APE1, creating an SSB with a priming 3′-hydroxyl group and a damaged 5′-abasic fragment [51]. The missing nucleotide is then replaced, and the 5′-end processed to create a regular 5′phosphate group to permit the ligation and completion of the repair response. The primary gap-filling polymerase in humans is POLβ, which also possesses the ability to remove 5′-abasic groups (i.e., the deoxyribose phosphate) left behind by APE1 incision [52]. X-ray cross-complementing protein 1 (XRCC1) in complex with DNA LIG3ɑ plays a role in sealing the nick during classic BER [53,54]. Besides the single-nucleotide (a.k.a., shortpatch) BER event described above, humans possess a long-patch BER process that entails strand-displacement synthesis, typically by a replicative polymerase (POLδ/POLε) in collaboration with PCNA, 5′-flap resolution by FEN1, and nick sealing by LIG1 [55].  . Classic base excision repair copes with various abnormal or modified bases, while singlestrand break repair is a related, specialized pathway that resolves non-conventional 3 or 5 termini in DNA. Full pathway descriptions and protein details can be found in the text and Table 1. The large red line indicates the substrate base, whereas the smaller red box designates the 5 sugar fragment prior to removal by POLβ (Short-patch) or during strand-displacement synthesis (Long-patch). The black slashes in SSBR represent various non-conventional ends processed by specific enzymes as discussed in the text.

Single-Strand Break Repair
Related to BER, yet more appropriately considered an independent pathway, SSBR has been designed to deal with strand breaks that harbor a non-conventional terminus created as an intermediate during DNA enzymatic processing or through direct chemical damage to DNA [56] (Figure 4). The most prominent of the damaged termini include: 3 -phosphate, a product of ROS attack of the sugar moiety in DNA and certain enzymatic events; 3 -phosphoglycolate and 5 -aldehyde, common to IR exposure and also products of ROS attack; 3 -phospho-α,β-unsaturated aldehyde, an intermediate of a dual functional DNA glycosylase with AP lyase activity; 3 -TOPO1 adduct, a trapped enzyme intermediate due to the presence of interfering DNA damage or exposure to a therapeutic agent like camptothecin; 5 -hydroxyl group, a product of IR exposure and certain alkylators; 5adenylate (AMP), an intermediate of failed DNA ligation; and the 5 -deoxyribose phosphate that arises following APE1 incision during BER. These different lesion types must be resolved to create normal 3 -hydroxyl and 5 -phosphate ends to permit repair synthesis and/or nick ligation [57].
SSB lesions are often initially recognized by poly(ADP)ribose polymerase 1 (PARP1), a strand break sensor that operates to coordinate the repair response through DNA bindingmediated post-translational modifications [58]. Then, depending on the type of termini present at the strand break ends, different enzymes are needed to facilitate the SSBR event. APE1, the major AP endonuclease in BER, also plays a significant role in SSBR, specifically processing 3 -phosphoglycolates and 3 -α,β-unsaturated aldehydes and to a lesser extent 3 -phosphate groups [59,60]. Aprataxin (APTX) copes with 3 -phosphoglycolates but maintains a more critical role in resolving 5 -adenylate intermediates [61]. Polynucleotide kinase 3 -phosphatase (PNKP), as the name suggests, removes 3 -phosphate blocking groups and phosphorylates 5 -hydroxyl termini [62,63]. Tyrosyl-DNA phosphodiesterase 1 (TDP1) has evolved to remove 3 -TOPO1 adducts, which are likely processed to small peptides before being converted to 3phosphate and eventually 3 -hydroxyl ends [64]. FEN1, the predominant 5 -flap endonuclease, as well as other structure-specific nucleases, such as the complex consisting of excision repair cross-complementing 1 and xeroderma pigmentosum complementation group F (i.e., ERCC1/XPF), also appear to play a general role in termini clean-up [65][66][67]. FEN1, for instance, assists in the removal of 5 -aldehyde and 5 -adenylate groups after strand displacement repair synthesis events [68,69]. POLβ excises 5 -deoxyribose abasic fragments that arise during classic BER and appears to also contribute to 5 -adenylate removal via its associated lyase activity. After termini resolution and gap-filling, SSBR, like BER, is typically completed by either the XRCC1/LIG3α complex or LIG1 [70].

Nucleotide Excision Repair
NER is divided into two major sub-pathways, global-genome (GG-NER) and transcriptioncoupled (TC-NER), molecular processes that cope with large, helix-distorting adducts, most notably UV-light induced (6-4)photoproducts and cyclopyrimidine dimers [71,72] ( Figure 5). The distinction between the two pathways is centered around their damage recognition mechanism. In the case of GG-NER, which, as the name implies, handles substrates throughout the genome, damage identification is mediated by rad23 homolog B (RAD23B) and the XPC protein, which recognize disrupted base pairs within the duplex; damage verification is performed in concert with RPA, XPA, and XPD, a helicase within the transcription complex TFIIH [73,74]. For TC-NER, the substrate lesion, when positioned on the transcribed strand, is revealed by the stalling of RNA polymerase during transcription elongation [75]. This event calls into action numerous factors, including the TFIIH complex, XPA, RPA, and the TC-NER proteins, Cockayne syndrome A (CSA, a.k.a., ERCC8), and CSB (a.k.a., ERCC6) [76,77]. Following the recognition and verification of the damage, the two NER sub-pathways involve identical proteins that carry out excision of the damagecontaining strand (ERCC1/XPF, 5 incision; XPG, 3 incision), replacement of the missing DNA segment (POLδ/POLε), and ligation of the remaining nick (LIG1 in replicating or XRCC1-LIG3α complex in non-replicating cells) [78].

Double-Strand Break Repair
Cells have adopted two primary corrective responses to cope with DSBs: homologous recombination (HR) and non-homologous end-joining (NHEJ) (Figure 6). DSBs exist in two primary forms, one-ended and two-ended. The former is an intermediate of replicative stress, arising after fork collapse at sites of DNA damage (e.g., SSBs), complex sequences (e.g., microsatellites), or active RNA synthesis and R-loop formation. The latter is created upon the breakage of both strands of DNA, either via chemical attack or enzymatic processing. Beyond just the cell cycle phase, several factors dictate pathway choice, i.e., HR or NHEJ [79,80]. We focus here on only the key molecular features of the two processes.
Int. J. Mol. Sci. 2022, 23, 4142 19 of 30 Figure 6. Double-strand breaks can be resolved via homologous recombination or non-homologous end-joining. Full pathway descriptions and protein details can be found in the text and Table 1.
Homologous recombination is crudely drawn, with more extensive details available in [81].
HR, which also operates to introduce genetic diversity into the gene pool during meiosis, faithfully resolves both one-ended and two-ended DSBs using a homologous sequence (typically the sister chromatid) as a template, thereby restricting the process to primarily the S/G2 phases of the cell cycle [81,82]. In brief, HR entails the following coordinated steps: (i) damage recognition, signaling, and some initial processing by the MRN Figure 6. Double-strand breaks can be resolved via homologous recombination or non-homologous end-joining. Full pathway descriptions and protein details can be found in the text and Table 1. Homologous recombination is crudely drawn, with more extensive details available in [81]. HR, which also operates to introduce genetic diversity into the gene pool during meiosis, faithfully resolves both one-ended and two-ended DSBs using a homologous sequence (typically the sister chromatid) as a template, thereby restricting the process to primarily the S/G2 phases of the cell cycle [81,82]. In brief, HR entails the following coordinated steps: (i) damage recognition, signaling, and some initial processing by the MRN complex (composed of meiotic recombination 11 homolog 1 (MRE11), RAD50, and Nijmegen breakage syndrome 1 (NBS1)); ataxia telangiectasia mutated (ATM); and RB binding protein 8 (RBBP8; a.k.a., CtIP and SAE2), (ii) long-range resection of the DSB end to create the necessary RPA-coated 3 -single-stranded tail for strand invasion (by exonuclease 1 (EXO1) and DNA2, likely in coordination with the Bloom helicase (BLM)), (iii) strand exchange into the undamaged sister chromatid and D-loop formation via RAD51 filament formation and the activities of BRCA2, the RAD51 paralogs, RAD54, among possible others, (iv) transfer of genetic information from the sister chromatid to the damaged chromosome through orchestrated DNA movements (e.g., branch migration) and DNA synthesis most likely by POLδ, and (v) resolution of the intertwined duplexes, the so-called Holliday junction (by BLM and TOPO3α or a nuclease complex, such as MUS81/EME1 (essential meiotic structure-specific endonuclease 1)). There are alternative forms of HR and different mechanisms of proceeding through DSB resolution that we will not cover here (see references above and [83]). Still, in the end, the process typically results in faithful removal of DSBs from the genome.
NHEJ, which likely evolved to execute programmed processes, such as class switch recombination that facilitates antibody diversification, operates independently of the cell cycle yet is most prominent in G1 [84]. The classic pathway entails the primary steps [85]: initial recognition of the DSB by the Ku complex, consisting of KU70 and KU80, and DNA-PK (catalytic subunit); alignment of the two ends of the two-ended DSB; and end-processing by MRN, Artemis, other clean-up enzymes (see SSBR), or DNA polymerases (e.g., POLλ and POLµ). These biochemical steps facilitate annealing and ligation by the XRCC4-like factor (XLF), a paralogue of XRCC4 and XLF (PAXX), XRCC4, and LIG4. NHEJ tends to be error-prone as small segments of DNA are often added or removed during the end-joining step. Recent work has identified two sub-pathways of NHEJ; classic and alternative (a.k.a., microhomology-mediated end-joining), with the latter involving a distinct set of proteins (e.g., PARP1, XRCC1, FEN1, MRE11, NBS1, LIG3, and POLθ), more commonly inducing extensive genomic alterations (typically deletions), and often being upregulated in cancer as a compensatory system [86].

DNA Repair Defects in Neurodegenerative Disease
Armed with the knowledge of the previous section, we next review the evidence for the involvement of DNA damage and DNA repair mechanisms in neurodegenerative disease. We focus initially on inherited disorders that experience neurodegeneration and subsequently on associations of DNA repair defects with broader neurodegenerative disorders.

Inherited Disorders Involving Defects in DNA Repair and Neurological Abnormalities
The suspected participation of DNA damage in neurodegeneration was validated upon the discovery that inherited disorders involving neurological abnormalities, such as xeroderma pigmentosum (XP), stem from defects in the ability to efficiently respond to and clear genomic stress [87]. XP, characterized by extreme sun sensitivity and a high occurrence of ultraviolet radiation-induced skin cancers, was the first human genetic disorder to be shown to harbor a genuine DNA repair defect [88,89]. XP is comprised of complementation groups, XPA-XPG, which identify factors of the NER sub-pathways that have specific roles in removing obstructive, helix-distorting DNA adducts, such as ultraviolet irradiation photoproducts. Besides being cancer-prone, approximately a quarter of XP-affected individuals exhibit progressive neurological degeneration. Since mutations in genes that affect only GG-NER (e.g., XPC) result in strictly cancer predisposition, whereas mutations in TC-NER-associated genes (i.e., CSA, CSB, XPA, XPB, XPD, XPF, and XPG [90]) give rise to neurological defects, it is believed that persistent transcription blocking lesions are particularly lethal to neuronal cells. The primary role for NER in the brain is presumed to be the removal of certain forms of oxidative DNA damage, such as 8-oxoguanine, apurinic/apyrimidinic (AP or abasic) sites, cyclopurines, DNA-DNA crosslinks, and protein-DNA adducts, particularly since in many instances alternative resolution mechanisms are absent, namely replication-associated HR [91,92].
Neurological problems occur not only due to deficits in TC-NER but also due to defects in other DNA repair processes [93,94]. Ataxia telangiectasia (AT) is a rare, monogenic recessive disorder characterized by radiosensitivity, increased cancer risk (particularly lymphomas), reduced immunological function, and progressive neurodegeneration that gives rise to ataxia, chorea, myoclonus, and neuropathy [95]. The defective gene in AT, i.e., AT mutated (ATM), encodes a serine/threonine-protein kinase that functions as a central signaling enzyme in the response to DSBs [96,97]. Its role in the faithful resolution of recombinogenic DNA DSBs explains the hypersensitivity of AT cells to IR, a physical agent that creates DSBs, and the pronounced genomic instability. However, given that homology-direct repair is not fully operational in non-dividing cells due to the lack of a paired sister chromatid, the precise role of ATM in protecting against progressive neuronal degeneration is less clear [98]. Supportive of a role for DNA DSBs in the neurodegenerative features of AT, inherited mutations in the nuclease meiotic recombination 11 homolog (MRE11), a protein that collaborates with ATM in the recognition and response to DNA DSBs, results in AT-like disorder (ATLD), a disease that exhibits similar neurological defects as AT [99]. While in these two diseases, some of the pathologies might arise from problems during development when cells are rapidly dividing, the absence of microcephaly in AT or ATLD individuals points towards brain atrophy mainly originating later in life [100]. Recent studies have indicated that DNA DSBs may arise in non-replicating, mature neurons through the action of ROS or the cleavage activity of TOPOIIβ as part of normal gene expression regulation [101]. However, the precise resolution mechanism(s) and how it might engage ATM remains unclear. Evidence also implicates ATM in mitigating cellular oxidative homeostasis, possibly through cytoplasmic or mitochondrial functions, and a nuclear DNA repair response to less complex oxidative DNA damage, such as SSBs, particularly in the context of transcription [98,102].
Besides ATM and MRE11, defects in other proteins that operate to clear DSBs or replicative stress have been tied to distinct disorders that involve neurological disease, cancer predisposition, and a range of other clinical phenotypes: Nijmegen breakage syndrome 1 (NBS1), ataxia telangiectasia and Rad3-related (ATR), DNA ligase 4 (LIG4), Cernunnos (a.k.a., XLF or NHEJ1), and components of the Fanconi anemia pathway (Table 1). However, unlike AT, these disorders mostly exhibit microcephaly, likely reflective of failed cell duplication during embryogenesis and consequent poor brain development [100].
Defects in components of SSBR (e.g., APTX, TDP1, PNKP, and XRCC1) have also been associated with genetic disorders that present with neurological disease (namely ataxia) [56]; yet these syndromes exhibit no overt cancer predisposition or other shared clinical manifestations with the diseases mentioned above ( Table 1). The restrictive nature of the clinical presentation in SSBR diseases to neurological deficits presumably stems from the fact that elevated endogenous strand breaks in non-dividing neurons result in transcription arrest and the activation of cell death pathways. Conversely, such SSB intermediates are resolved faithfully by replication-directed HR in dividing cells. Additionally, recessive mutations in Senataxin (SETX), an RNA-DNA helicase, have been linked to apraxia with oculomotor ataxia type 2 (AOA2), whereas dominant mutations in the gene have been detected in a juvenile-onset form of ALS [103]. Amassing evidence indeed suggests that defects in R-loop resolution, i.e., the ability to disentangle RNA-DNA hybrids, which are frequent intermediates of transcription, result in loss of RNA transcriptome homeostasis and genome integrity, culminating in neuronal cell death [104,105]. Lastly, the accumulation of nuclear DNA strand breaks can result in the hyperactivation of PARP1, leading to NAD overconsumption and mitochondrial dysfunction that contribute to the neurodegenerative process [106].

Associations of DNA Repair Defects with Classic Neurological Diseases
Besides the direct link between mutations in established DNA repair genes and the inherited disorders mentioned above (Table 1), DNA damage accumulation or defects in DNA damage processing have been implicated in nearly all neurological diseases, including sporadic and familial cases of AD, ALS, and PD [107][108][109]. In these latter diseases, oxidative DNA damage, which often arises in concert with mitochondrial dysfunction, is typically higher in disease tissue than in comparative control tissue. Evidence indicates that defects in BER, the primary system for handling oxidative DNA damage, can promote neurological degeneration in AD models, implying at minimum a protective role for DNA repair in disease etiology [110]. In addition, AD is associated with the accumulation of DNA DSBs [111], particularly in vulnerable neuronal and glial cell populations, likely due to reduced expression of DSB response proteins (e.g., MRE11, RAD50, and BRCA1) [112]. Likewise, in ALS, mutations in the superoxide dismutase 1 (SOD1) gene lead not only to defective ROS scavenging and oxidative homeostasis but to the impaired control of important DNA damage response gene expression (e.g., of SPY1) [113]. Moreover, mutations in TARDBP, which account for~4% of familial ALS cases, result in the pathogenic mislocalization of the TDP43 protein from the nucleus to the cytoplasm, causing defects in cytoplasmic-nuclear trafficking and reduced levels of proteins in the nucleus, such as the NHEJ factor XRCC4/LIG4 [114]. FUS, another protein found to be defective in sporadic and familial forms of ALS, has been demonstrated to have more direct roles in DNA repair, recruiting XRCC1/LIG3 to SSBs and promoting strand break resolution, as well as modulating DSBR efficiency, possibly through interactions with the chromatin modifier HDAC1 [115][116][117]. Mutations in FUS lead to cytoplasmic aggregation and loss of normal protein function in the nucleus, culminating in genomic instability and neuronal cell death. Regarding PD, studies indicate that α-synuclein, a protein found to misfold and aggregate into the characteristic clumps called Lewy Bodies, activates the ATM kinase [118]. Elevated levels of poly(ADP)ribose, a symptom of persistent DNA damage, have also been reported to accelerate the fibrillization of α-synuclein [119]. Finally, similar to what has been observed in AD, compromised DNA repair, namely BER or NER, may serve as a risk factor for PD-like pathology [120,121].
Other proteins linked to inherited neurological disorders have also been described to regulate or participate in the response to DNA damage [122][123][124][125]: (i) the Huntingtin protein (defective in HD) is recruited to DNA damage sites by ATM and serves as a repair complex scaffold; (ii) ATNX3, defective in spinocerebellar ataxia 3 (SCA3), accumulates at sites of DNA damage, participates in the ATM response, and interacts with the SSBR protein PNKP and TC-NER complexes; and (iii) survival motor neuron (SMN1/SMN2) proteins, in which defects give rise to spinal motor atrophy (SMA) disease, associate with the HR protein RAD51 and influence the levels of the NHEJ factor DNA-PKcs and the RNA/DNA helicase SETX.
Given the lack of DNA replication in terminally differentiated cells, pathways such as MMR, which copes with replication errors, and classic HR, which resolves collapsed replication forks, are not thought to constitute major neuroprotective mechanisms. Seemingly consistent with this notion, defects in MMR give rise to microsatellite instability (MSI) and, almost exclusively, cancer susceptibility, with genetic links to a range of hereditary colorectal cancers. That said, it is worth noting that inherited defects have been identified in MMR components, the FANCD2/FANCI-associated nuclease 1 (FAN1), and LIG1 in individuals with HD, suggesting a role for an undefined DNA surveillance mechanism (presumably not classic MMR) in suppressing pathological CAG repeat expansion that defines HD [126]. Defects in factors that participate in DSB resolution and/or HR not only give rise to chromosome rearrangements, an outcome common to cancer and other pathological conditions associated with genomic instability, but also to immunological dysfunction and premature aging features, in addition to neurological disease.

DNA Damage, DNA Repair, and Neuronal Cell Function
The previous sections highlight the fact that defects in DNA repair can lead to genomic stress that is responsible for the neuronal cell deficiencies in neurodevelopmental or neurodegenerative diseases. However, emerging evidence is now pointing to a more nuanced role of DNA damage and repair in mitigating neuronal cell identity and function. Indeed, early work by Suberbielle et al. discovered that the natural behavior of young adult mice, i.e., exploration of a new environment, and increased neuronal activity causes the formation of DNA DSBs in neurons. This DNA strand breakage occurs at higher levels in the dentate gyrus, an area of the brain that plays a crucial role in learning, memory, and spatial coding [127]. With the advent of genome-wide DNA damage techniques, scientists now appreciate that DNA damage is in some cases generated in a region-specific manner with a physiological purpose. For instance, Madabhushi et al. reported that stimulation of neuronal activity triggers the formation of DNA DSBs in the promoters of a subset of early-response genes that are crucial for experience-driven changes to synapses, learning, and memory (i.e., Fos, Npas4, and Egr1), enabling the resolution of topological constraints and gene expression [128]. Additionally, Gadd45γ, a member of the growth arrest and DNA damage (Gadd45) protein family, operates after an initial wave of DNA DSB formation and DNA methylation in the prelimbic prefrontal cortex, directing a DNA repair-mediated DNA demethylation process that promotes temporal immediate early gene expression necessary for fear memory consolidation [129]. Studies have suggested functions for other DNA repair proteins, such as PARP-1 and FEN-1, in learning and memory formation as well [130,131]. Although the functional relevance remains unclear, more recent genome-wide analyses have revealed that post-mitotic neurons accumulate exceptionally high levels of DNA SSBs, as assessed by recurrent repair synthesis events, within specific genomic hotspots, namely enhancers at or near CpG dinucleotides and sites of DNA demethylation [132]. Thus, to maintain the integrity of neurons and their genome, DNA repair seems to preferentially protect essential genes within specific repair hotspots [133]. The above findings emphasize that DNA damage and repair may be targeted to specific genomic locations, possibly in unique ways in different cell types, brain regions, and biological contexts. Neuronal cell survival and functionality, therefore, may depend not only on the resolution of global DNA damage but on precise and efficient DNA repair processing in critical genomic domains. As such, DNA repair defects and associated DNA damage accumulation might not only lead to neurodegeneration directly but neuronal cell dysfunction more broadly, perhaps causing more widespread consequences on the CNS network.

Summary and Future Perspectives
DNA damage, and, as such, defective DNA repair, plays a causal role in numerous pathological states, including neurodegenerative phenotypes. The accumulation of DNA damage in neuronal cells either drives the loss of genome integrity or interferes with gene regulatory processes or transcriptional events, thereby promoting activation of cell death responses and consequent cell loss. As is likely evident from the presentation above, depending on the proliferative nature of the cell, a preference for certain DNA repair pathways exists, where for example, dividing cells are more reliant on systems such as MMR, RER, and HR, whereas non-dividing cells favor NHEJ, SSBR, or NER (Figure 7). Not surprisingly, given that neurons are non-dividing in nature, defects in DNA repair pathways such as SSBR and NER, which resolve obstructive DNA lesions largely independently of replicative status, have been linked to neurodegenerative outcomes. Conversely, pathways such as MMR, which predominantly cope with DNA replication errors, are not directly associated with neurological complications when impaired. Although there have been associations with BER defects and the development of neurological disease, BER gene mutations, which result in defective processing of oxidative DNA damage, are mostly restricted to cancer predisposition disorders without clear neurodegenerative features [134]. Nevertheless, as pointed out above, evidence that indicates a role for both BER and NER in regulating neurological disease risk and severity is mounting. It will be interesting going forward to determine specifically whether central BER defects play causal roles in neurodegenerative disease, particularly given the prominent role of the pathway in the repair of oxidative DNA damage. For each pathway, only a representative protein is shown, typically the main recognition factor. * refers to general BER and not the long-patch sub-pathway, which engages several replication proteins and thus is relatively inactive in non-dividing cells.
As noted earlier, ribonucleotide incorporation into genomic DNA during replication is frequent. Of course, in non-dividing cells, faulty incorporation during chromosome duplication is not relevant. However, in situations of DNA repair synthesis, unwanted insertion of ribonucleotides could presumably still occur. Though it is presently unclear how this might affect the integrity of neuronal cell function, pathogenic mutations in RNAseH2, the main enzyme that initiates RER, give rise to Aicardi-Goutières syndrome (AGS), a monogenic type I interferonopathy characterized by neurodevelopmental defects and neuroinflammation [135]. Notably, recent evidence indicates that DNA damage-dependent signaling rather than type I interferon signaling underlies neurodegeneration in this class of disease [136]. While a prominent role for RER in the context of neurodevelopment is obvious, i.e., where replication is active, studies are needed to more precisely flush out the contribution of the repair system in terminally differentiated cells, such as neurons.
It is now well established that DNA DSBs can be generated in the neuronal genome through multiple mechanisms. While NHEJ is implicated in the resolution of such damage [128], the contributions of ATM and other HR-related factors are less clear [137]. Recent reports of an RNA-directed HR process involving RAD52 could provide non-dividing cells with the ability to carry out faithful repair of DSBs [138][139][140][141], as opposed to the errorprone mechanism of NHEJ that can give rise to chromosome instability. Delineating the molecular choreography that occurs at DSBs within the neuronal genome, both in terms of damage recognition and ultimately (ideally faithful) resolution, is a key topic of investigation looking into the future.
To date, studies have largely focused on the effects of DNA repair defects on neuronal cell function and viability. But as noted earlier, it is important to recognize that the nervous system is composed of an integrated network of neuronal and glial cell types. Determining the consequences of persistent DNA damage and the comparative role of the different DNA repair mechanisms in the distinct cell populations represents a critical step in defining the factors that influence health of the nervous system as a whole. Moreover, compared to the CNS, virtually nothing is known about DNA repair processes in the PNS or enteric nervous system, the other two major nervous systems in mammals. A broader understanding of DNA damage and repair in all nervous systems and within the cellular Figure 7. The proliferative character of cells allows them to preferentially use different DNA repair pathways. Dividing cells are more prone to use replication-associated pathways, e.g., mismatch repair (MMR), ribonucleotide excision repair (RER), or homologous recombination (HR), whereas non-dividing cells are more likely to apply non-homologous end-joining (NHEJ), single strand break repair (SSBR), or transcription-coupled nucleotide excision repair (TC-NER). Direct reversal (DR), global-genome NER (GG-NER), and base excision repair (BER) can be used in almost all cell types, regardless of proliferation status. The gray area indicates the level of pathway utilization in dividing (top) or non-dividing (bottom) cell types. For each pathway, only a representative protein is shown, typically the main recognition factor. * refers to general BER and not the long-patch sub-pathway, which engages several replication proteins and thus is relatively inactive in non-dividing cells.
As noted earlier, ribonucleotide incorporation into genomic DNA during replication is frequent. Of course, in non-dividing cells, faulty incorporation during chromosome duplication is not relevant. However, in situations of DNA repair synthesis, unwanted insertion of ribonucleotides could presumably still occur. Though it is presently unclear how this might affect the integrity of neuronal cell function, pathogenic mutations in RNAseH2, the main enzyme that initiates RER, give rise to Aicardi-Goutières syndrome (AGS), a monogenic type I interferonopathy characterized by neurodevelopmental defects and neuroinflammation [135]. Notably, recent evidence indicates that DNA damage-dependent signaling rather than type I interferon signaling underlies neurodegeneration in this class of disease [136]. While a prominent role for RER in the context of neurodevelopment is obvious, i.e., where replication is active, studies are needed to more precisely flush out the contribution of the repair system in terminally differentiated cells, such as neurons.
It is now well established that DNA DSBs can be generated in the neuronal genome through multiple mechanisms. While NHEJ is implicated in the resolution of such damage [128], the contributions of ATM and other HR-related factors are less clear [137]. Recent reports of an RNA-directed HR process involving RAD52 could provide non-dividing cells with the ability to carry out faithful repair of DSBs [138][139][140][141], as opposed to the error-prone mechanism of NHEJ that can give rise to chromosome instability. Delineating the molecular choreography that occurs at DSBs within the neuronal genome, both in terms of damage recognition and ultimately (ideally faithful) resolution, is a key topic of investigation looking into the future.
To date, studies have largely focused on the effects of DNA repair defects on neuronal cell function and viability. But as noted earlier, it is important to recognize that the nervous system is composed of an integrated network of neuronal and glial cell types. Determining the consequences of persistent DNA damage and the comparative role of the different DNA repair mechanisms in the distinct cell populations represents a critical step in defining the factors that influence health of the nervous system as a whole. Moreover, compared to the CNS, virtually nothing is known about DNA repair processes in the PNS or enteric nervous system, the other two major nervous systems in mammals. A broader understanding of DNA damage and repair in all nervous systems and within the cellular network of a particular nervous system will not only shed new light on the etiology of neurodegenerative diseases but can pave the way for new therapeutic strategies. Given the decades' worth of research that has been poured into understanding the role of DNA repair in neurons within the CNS, it would appear that neuroscientists interested in genome stability and the DDR in the nervous system will have plenty of opportunities to keep themselves busy for the next several decades.