**2.3 Mismatch repair (MMR)**

Mismatch repair (MMR) corrects mismatches and small insertions or deletions during DNA replication, thus eliminating potentially pre-mutagenic bases. Repair involves recognition of the mismatch by MutSα (MSH2 and MSH6 proteins), or by MutSβ (MSH2 and MSH3 proteins) in the case of small insertions/deletions (1-10 nucleotides). MutLα (a heterodimer of MLH1 and PMS2 proteins) is then recruited and serves to coordinate the process that involves, among others, the PCNA protein for strand discrimination and exonuclease 1, DNA Pol δ and a DNA ligase, for DNA repair (Kunkel & Erie, 2005).

#### **2.4 Homologous recombination (HR) and non homologous end joining (NHEJ)**

Non homologous end joining (NHEJ) is the major pathway for the repair of DSBs because it can function throughout the cell cycle and does not require a homologous chromosome. Rather, NHEJ involves rejoining of what remains of the two DNA ends, tolerating nucleotide loss or addition at the rejoining site. When a DSB occurs during G0, G1, and early

Variants and Polymorphisms of DNA Repair Genes and Neurodegenerative Diseases 571

and genotype frequencies between patients and controls (Coppedè et al., 2007a). Subsequently, 91 sporadic Turkish AD patients and 93 matched controls have been genotyped for the presence of *OGG1* Ser326Cys, *APEX1* Asp148Glu (rs1130409), *XRCC1* Arg280His (rs25489) and *XRCC1* Arg399Gln (rs25487) polymorphisms, but none of them was associated with increased AD risk (Parildar-Karpuzoğlu et al. 2008). Also a small case-control study performed in Poland with 41 AD patients and 51 controls failed to find significant differences in *OGG1* Ser326Cys allele frequencies between groups (Dorszewska et al., 2009). A borderline association with AD risk (*P* =0.06) was observed for the *XRCC1* Arg194Trp (rs1799782) polymorphism in a group of 98 Turkish AD patients and 95 healthy subjects (Doğru-Abbasoğlu et al., 2007), but a recent study failed to replicate this association in a larger case-control group of over 200 Chinese AD patients (Quian et al., 2010). Overall, five common functional polymorphisms of BER genes have been investigated as possible AD risk factors, but none of them resulted significantly associated with increased AD risk (Table 2). Also polymorphisms of NER genes have been evaluated as candidate AD risk factors. Particularly, two common polymorphisms of the *XPD (ERCC2)* gene (namely, rs238406 and rs13181), and a silent mutation in exon 11 (T>C at codon 824) of the *XPF (XRCC4)* gene have been investigated in 97 Turkish AD patients and in 101 matched controls, but none of them resulted to be associated with AD risk (Doğru-Abbasoğlu et al., 2006). Poly-ADP-ribose polymerase-1 (PARP-1) is a zinc-finger DNA binding protein that is activated by DNA SSBs or DSBs. The primary function of PARP-1 is in DNA repair processes through the detection of DNA damage and the prevention of chromatide exchanges. PARP-1 poly-ADP-ribosylates several proteins involved in DNA repair including histones, thus inducing local relaxation of the chromatin structure and facilitating the access of repair proteins to damaged DNA. There is evidence for widespread DNA SSBs and DSBs in AD brains, as well as increased PARP-1 activity (Love et al., 1999). Two independent groups evaluated *PARP-1* gene polymorphisms as putative AD risk factors. Infante and coworkers screened 263 Spanish AD patients and 293 matched controls for the presence of two *PARP-1* promoter polymorphisms (–410 and –1672). If evaluated independently, nor *PARP-1* –410 neither *PARP-1* –1672 resulted associated with increased AD risk (Table 2). However, *PARP-1* – 410 and *PARP-1* –1672 polymorphisms resulted in linkage disequilibrium and some haplotypes were associated with increased AD risk. Particularly, haplotypes 2-1 and 1-2 were significantly overrepresented in AD individuals and associated with an increased risk for the disease with an adjusted OR of 1.42 and 5.38, respectively (Infante et al., 2007). More recently two *PARP-1* exonic polymorphisms, 414C>T (rs1805404) and 2456T>C (rs1136410), have been evaluated in 120 Chinese AD patients and 111 matched controls (Liu et al., 2010). Again, none of the polymorphisms resulted independently associated with increased AD risk (Table 2). However, authors found that the distributions of haplotype 3-TT and haplotype 4-CC were significantly associated with an increased risk of AD, whereas the haplotype 1-TC showed a protective effect, with OR of 12.2 and 0.52, respectively (Liu et al., 2010). Overall, both studies support the hypothesis that *PARP1* 

**3.1 Searching for BER gene variants in DNA extracted from post-mortem AD brain**  Mao and colleagues extracted nuclear DNA from post-mortem brain specimens of 14 late stage AD patients and 10 neurologically healthy controls. They identified and characterized novel *OGG1* mutations (a single base deletion C796del, and two base substitutions leading

haplotypes might affect AD risk.

S phase, the Ku heterodimer (Ku70/Ku80) recognizes DSB ends, aligns them, protects them from excessive degradation, and ultimately prepares them for ligation. The Ku heterodimer is capable of interacting with the nuclease (Artemis-DNA-PKcs) complex, the polymerases (μ and λ), and the ligase (XLF-XRCC4-DNA ligase IV) complex. If complementary ends are not present at the break, the Artemis-DNA-PKcs complex resects some of the overhangs to create single-strand overhangs with short stretches of micro-homology. When necessary, polymerization of missing nucleotides is performed by DNA polymerases. Then, the XLF-XRCC4-LigaseIV complex seals the DSB. When homologous recombination (HR) is used for repair, it is promoted by the recombinase RAD51, the human homolog of the *E. coli* RecA protein, which binds to 3′-tailed single strands at the end of DSBs in a helical fashion and promotes pairing with homologous DNA sequences (usually the sister chromatid) as a prelude to strand invasion and repair of the DSBs. Strand invasion is the invasion of the 3′ end of the single-stranded DNA overhang into the region of complementarity in the intact sister chromatid. The process is directed by RAD51 which forms a nucleoprotein filament that directs homology search, strand pairing, and invasion of the homologous chromosome. Rad51 is assisted in this process by several RAD family members (RAD51B, RAD51C, RAD51D, XRCC2, XRCC3, RAD54, and RAD52). During strand invasion, RAD51 creates a four-stranded Holliday junction intermediate. Then, the invading strand is extended by DNA polymerase η and the Holliday junction is resolved by a RAD51C and XRCC3 directed mechanism. Several nucleases and helicases, such as the RecQ family members, also participate in resolving Holliday junctions. Since eukaryotic genomes contain dispersed repeated DNA, repair of DSBs by HR can occur not only through an interaction with the sister chromatid or the homolog chromosome, but also with repeats on non-homolog chromosomes. Numerous factors affect the decision to repair a DSB via these pathways, and accumulating evidence suggests these major repair pathways both cooperate and compete with each other at double-strand break sites to facilitate efficient repair and promote genomic integrity (Kass & Jasin, 2010).

#### **3. Polymorphisms of DNA repair genes and Alzheimer's disease**

Alzheimer's disease is a complex multi-factorial neurodegenerative disorder and represents the most common form of dementia in the elderly. In 2006, the worldwide prevalence of AD was 26.6 million. It has been estimated that following the global aging of the world's population this number will quadruple by 2050, suggesting that 1 in 85 persons worldwide will be living with the disease, which is clinically characterized by a progressive neurodegeneration in selected brain regions, including the temporal and parietal lobes and restricted regions within the frontal cortex and the cingulate gyrus, resulting in gross atrophy of the affected regions and leading to memory loss accompanied by changes of behaviour and personality severe enough to affect work, lifelong hobbies or social life (Brookmeyer et al., 2007). Increasing evidence reports oxidative DNA damage in affected brain regions of AD patients, paralleled by a decrease in DNA repair activities, particularly concerning the BER pathway (Lovell et al., 2000; Weissman et al., 2007). This has driven current research to focus on common polymorphisms of BER genes as candidate AD risk factors. Studied genes are those encoding for 8-oxoguanine DNA glycosylase (OGG1), APEX1 and XRCC1. Particularly, we screened 178 Italian late onset AD patients and 146 matched controls for the presence of the *OGG1* Ser326Cys gene polymorphism (rs1052133), observing no difference in allele

S phase, the Ku heterodimer (Ku70/Ku80) recognizes DSB ends, aligns them, protects them from excessive degradation, and ultimately prepares them for ligation. The Ku heterodimer is capable of interacting with the nuclease (Artemis-DNA-PKcs) complex, the polymerases (μ and λ), and the ligase (XLF-XRCC4-DNA ligase IV) complex. If complementary ends are not present at the break, the Artemis-DNA-PKcs complex resects some of the overhangs to create single-strand overhangs with short stretches of micro-homology. When necessary, polymerization of missing nucleotides is performed by DNA polymerases. Then, the XLF-XRCC4-LigaseIV complex seals the DSB. When homologous recombination (HR) is used for repair, it is promoted by the recombinase RAD51, the human homolog of the *E. coli* RecA protein, which binds to 3′-tailed single strands at the end of DSBs in a helical fashion and promotes pairing with homologous DNA sequences (usually the sister chromatid) as a prelude to strand invasion and repair of the DSBs. Strand invasion is the invasion of the 3′ end of the single-stranded DNA overhang into the region of complementarity in the intact sister chromatid. The process is directed by RAD51 which forms a nucleoprotein filament that directs homology search, strand pairing, and invasion of the homologous chromosome. Rad51 is assisted in this process by several RAD family members (RAD51B, RAD51C, RAD51D, XRCC2, XRCC3, RAD54, and RAD52). During strand invasion, RAD51 creates a four-stranded Holliday junction intermediate. Then, the invading strand is extended by DNA polymerase η and the Holliday junction is resolved by a RAD51C and XRCC3 directed mechanism. Several nucleases and helicases, such as the RecQ family members, also participate in resolving Holliday junctions. Since eukaryotic genomes contain dispersed repeated DNA, repair of DSBs by HR can occur not only through an interaction with the sister chromatid or the homolog chromosome, but also with repeats on non-homolog chromosomes. Numerous factors affect the decision to repair a DSB via these pathways, and accumulating evidence suggests these major repair pathways both cooperate and compete with each other at double-strand break sites to facilitate efficient repair and promote

genomic integrity (Kass & Jasin, 2010).

**3. Polymorphisms of DNA repair genes and Alzheimer's disease** 

Alzheimer's disease is a complex multi-factorial neurodegenerative disorder and represents the most common form of dementia in the elderly. In 2006, the worldwide prevalence of AD was 26.6 million. It has been estimated that following the global aging of the world's population this number will quadruple by 2050, suggesting that 1 in 85 persons worldwide will be living with the disease, which is clinically characterized by a progressive neurodegeneration in selected brain regions, including the temporal and parietal lobes and restricted regions within the frontal cortex and the cingulate gyrus, resulting in gross atrophy of the affected regions and leading to memory loss accompanied by changes of behaviour and personality severe enough to affect work, lifelong hobbies or social life (Brookmeyer et al., 2007). Increasing evidence reports oxidative DNA damage in affected brain regions of AD patients, paralleled by a decrease in DNA repair activities, particularly concerning the BER pathway (Lovell et al., 2000; Weissman et al., 2007). This has driven current research to focus on common polymorphisms of BER genes as candidate AD risk factors. Studied genes are those encoding for 8-oxoguanine DNA glycosylase (OGG1), APEX1 and XRCC1. Particularly, we screened 178 Italian late onset AD patients and 146 matched controls for the presence of the *OGG1* Ser326Cys gene polymorphism (rs1052133), observing no difference in allele and genotype frequencies between patients and controls (Coppedè et al., 2007a). Subsequently, 91 sporadic Turkish AD patients and 93 matched controls have been genotyped for the presence of *OGG1* Ser326Cys, *APEX1* Asp148Glu (rs1130409), *XRCC1* Arg280His (rs25489) and *XRCC1* Arg399Gln (rs25487) polymorphisms, but none of them was associated with increased AD risk (Parildar-Karpuzoğlu et al. 2008). Also a small case-control study performed in Poland with 41 AD patients and 51 controls failed to find significant differences in *OGG1* Ser326Cys allele frequencies between groups (Dorszewska et al., 2009). A borderline association with AD risk (*P* =0.06) was observed for the *XRCC1* Arg194Trp (rs1799782) polymorphism in a group of 98 Turkish AD patients and 95 healthy subjects (Doğru-Abbasoğlu et al., 2007), but a recent study failed to replicate this association in a larger case-control group of over 200 Chinese AD patients (Quian et al., 2010). Overall, five common functional polymorphisms of BER genes have been investigated as possible AD risk factors, but none of them resulted significantly associated with increased AD risk (Table 2). Also polymorphisms of NER genes have been evaluated as candidate AD risk factors. Particularly, two common polymorphisms of the *XPD (ERCC2)* gene (namely, rs238406 and rs13181), and a silent mutation in exon 11 (T>C at codon 824) of the *XPF (XRCC4)* gene have been investigated in 97 Turkish AD patients and in 101 matched controls, but none of them resulted to be associated with AD risk (Doğru-Abbasoğlu et al., 2006). Poly-ADP-ribose polymerase-1 (PARP-1) is a zinc-finger DNA binding protein that is activated by DNA SSBs or DSBs. The primary function of PARP-1 is in DNA repair processes through the detection of DNA damage and the prevention of chromatide exchanges. PARP-1 poly-ADP-ribosylates several proteins involved in DNA repair including histones, thus inducing local relaxation of the chromatin structure and facilitating the access of repair proteins to damaged DNA. There is evidence for widespread DNA SSBs and DSBs in AD brains, as well as increased PARP-1 activity (Love et al., 1999). Two independent groups evaluated *PARP-1* gene polymorphisms as putative AD risk factors. Infante and coworkers screened 263 Spanish AD patients and 293 matched controls for the presence of two *PARP-1* promoter polymorphisms (–410 and –1672). If evaluated independently, nor *PARP-1* –410 neither *PARP-1* –1672 resulted associated with increased AD risk (Table 2). However, *PARP-1* – 410 and *PARP-1* –1672 polymorphisms resulted in linkage disequilibrium and some haplotypes were associated with increased AD risk. Particularly, haplotypes 2-1 and 1-2 were significantly overrepresented in AD individuals and associated with an increased risk for the disease with an adjusted OR of 1.42 and 5.38, respectively (Infante et al., 2007). More recently two *PARP-1* exonic polymorphisms, 414C>T (rs1805404) and 2456T>C (rs1136410), have been evaluated in 120 Chinese AD patients and 111 matched controls (Liu et al., 2010). Again, none of the polymorphisms resulted independently associated with increased AD risk (Table 2). However, authors found that the distributions of haplotype 3-TT and haplotype 4-CC were significantly associated with an increased risk of AD, whereas the haplotype 1-TC showed a protective effect, with OR of 12.2 and 0.52, respectively (Liu et al., 2010). Overall, both studies support the hypothesis that *PARP1*  haplotypes might affect AD risk.

#### **3.1 Searching for BER gene variants in DNA extracted from post-mortem AD brain**

Mao and colleagues extracted nuclear DNA from post-mortem brain specimens of 14 late stage AD patients and 10 neurologically healthy controls. They identified and characterized novel *OGG1* mutations (a single base deletion C796del, and two base substitutions leading

Variants and Polymorphisms of DNA Repair Genes and Neurodegenerative Diseases 573

association study for risk assessment but a genetic screening performed on brain DNA specimens searching for novel *OGG1* variants. The authors created the mutant proteins by site-directed mutagenesis observing that the C796del mutant OGG1 lacks glycosylase activity, whereas both Ala53Thr and Ala288Val substitutions result in 40–50% reduced activity (Mao et al., 2007). Therefore, we cannot exclude that the activity of the OGG1 protein might be partially impaired by rare gene variants in some AD subjects. However, given the limited sample-size of the studied case-control group, further studies are required

**4. Polymorphisms of DNA repair genes and amyotrophic lateral sclerosis** 

Amyotrophic lateral sclerosis (ALS), also known as motor neuron disease (MND), is one of the major neurodegenerative diseases alongside AD and PD. It is a progressive disorder characterized by the degeneration of motor neurons of the motor cortex, brainstem and spinal cord. The incidence of the disease is similar worldwide and ranges from 1 to 3 cases per 100,000 individuals per year, with the exception of some high-risk areas around the Pacific Rim. Several studies report increased oxidative DNA damage and a compromised DNA repair activity, particularly BER activity, in spinal cords and other tissues of ALS patients (Bogdanov et al., 2000; Ferrante et al., 1997; Kikuchi et al., 2002; Kisby et al., 1997). Missense mutations in the gene encoding *APEX1* were found in DNA obtained from 8 of 11 ALS patients, including the common *APEX1* Asp148Glu polymorphism (Hayward et al., 1999), that was subsequently associated with increased ALS risk in a Scottish cohort of 117 ALS patients and 58 controls, and in an Irish group of 105 ALS individuals and 82 controls (Greenway et al., 2004). The analysis of 88 English ALS patients and 88 matched controls still revealed an increased frequency of the variant allele in the ALS cohort, even if not statistically significant (Tomkins et al. 2000). We have recently performed the largest casecontrol study aimed at clarifying the role of *APEX1* Asp148Glu in sporadic ALS pathogenesis. No difference in *APEX1* Asp148Glu allele and genotype frequencies was found between 134 ALS patients and 129 controls of Italian origin, nor was the polymorphism associated with disease age or site of onset, or duration of the disease, suggesting that it might not play a major role in ALS pathogenesis in the Italian population (Coppedè et al., 2010a). The ALSGene database (www.alsgene.org) is a public database containing all the ALS genetic association studies, genome-wide association studies and updated meta-analyses of the literature. A meta-analysis of the four studies described above revealed a significant increased frequency of the variant 148Glu allele in ALS cases with respect to controls, suggesting a protective role for the wild type 148Asp variant with an OR = 0.78 (95%CI=0.62-0.97) (www.alsgene.org). Our analysis of the *OGG1* Ser326Cys polymorphism in 136 ALS patients and 129 matched controls of Italian origin revealed a significant association of the variant allele with increased ALS risk (Coppedè et al., 2007b) (Table 3). At best of our knowledge this study is the first in the literature addressing this issue, still pending replication in other populations. More recently, we screened over 400 individuals, including 206 ALS patients and 203 matched controls of Italian origin for the presence of *XRCC1* Arg194Trp, Arg280His and Arg399Gln polymorphisms, observing a significant increased frequency of the 399Gln variant allele and a borderline significant decreased frequency of the 194Trp allele in ALS patients with respect to controls (Coppedè et al., 2010b). Interestingly, others have evaluated the same *XRCC1* polymorphisms and two

to confirm this hypothesis.


Table 2. DNA repair gene polymorphisms and risk of Alzheimer's disease

to Ala53Thr or Ala288Val amino acidic changes, respectively) in 4 of 14 AD subjects. Particularly, two AD patients carried the C796 deletion, one patient had the Ala53Thr substitution, and another patient carried the Ala288Val substitution. No mutations were found in any of 10 studied age-matched controls (Mao et al., 2007). This study is not an

al., 2007a OGG1 Ser326Cys 178/146 0.19/0.18 1.04 (0.70-1.55)

et al., 2009 OGG1 Ser326Cys 41/51 0.29/0.21 1.60 (0.81-3.14)

OGG1 Ser326Cys 91/93 0.29/0.23 1.32 (0.83-2.11)

Asp148Glu 91/93 0.33/0.31 1.08 (0.70-1.68)

Arg280His 91/93 0.06/0.10 0.53 (0.24-1.14)

Arg399Gln 91/93 034/0.33 1.05 (0.68-1.63)

Arg194Trp 98/95 0.11/0.06 2.06 (0.97-4.37)

Arg194Trp 212/203 0.31/0.31 1.04 (0.70-1.52)

(XPD exon 6) 97/101 0.40/0.42 0.94 (0.63-1.41)

(XPD exon 23) 97/101 0.41/0.36 1.24 (0.83-1.86)

(XPF exon 11) 97/101 0.37/0.35 1.09 (0.72-1.64)

(-410) 263/293 0.35/0.33 1.08 (0.84-1.38)

(-1672) 263/293 0.17/0.18 0.94 (0.69-1.27)

(rs1805404) 120/111 0.53/0.59 0.76 (0.53-1.11)

(rs1136410) 120/111 0.52/0.59 0.75 (0.52-1.09)

to Ala53Thr or Ala288Val amino acidic changes, respectively) in 4 of 14 AD subjects. Particularly, two AD patients carried the C796 deletion, one patient had the Ala53Thr substitution, and another patient carried the Ala288Val substitution. No mutations were found in any of 10 studied age-matched controls (Mao et al., 2007). This study is not an

**Variant allele frequency AD/Controls** 

**Odds Ratio (95% CI)** 

**Number of subjects AD/Controls** 

**Reference Polymorphism** 

APEX1

XRCC1

XRCC1

XRCC1

XRCC1

ERCC2 rs238406

ERCC2 rs13181

ERCC4

PARP-1

PARP-1

PARP-1

PARP-1

Table 2. DNA repair gene polymorphisms and risk of Alzheimer's disease

Coppedè et

Parildar-Karpuzoğlu et al. 2008

Dorszewska

Parildar-Karpuzoğlu et al. 2008

Parildar-Karpuzoğlu et al. 2008

Parildar-Karpuzoğlu et al. 2008

Doğru-Abbasoğlu et al., 2007

Quian et al., 2010

Doğru-Abbasoğlu et al., 2006

Doğru-Abbasoğlu et al., 2006

Doğru-Abbasoğlu et al., 2006

Infante et al., 2007a

Infante et al., 2007a

> Liu et al., 2010

> Liu et al., 2010

association study for risk assessment but a genetic screening performed on brain DNA specimens searching for novel *OGG1* variants. The authors created the mutant proteins by site-directed mutagenesis observing that the C796del mutant OGG1 lacks glycosylase activity, whereas both Ala53Thr and Ala288Val substitutions result in 40–50% reduced activity (Mao et al., 2007). Therefore, we cannot exclude that the activity of the OGG1 protein might be partially impaired by rare gene variants in some AD subjects. However, given the limited sample-size of the studied case-control group, further studies are required to confirm this hypothesis.
