Genetic Polymorphisms in Oxidative Stress and Inflammatory Pathways as Potential Biomarkers in Alzheimer’s Disease and Dementia

Oxidative stress and neuroinflammation are important processes involved in Alzheimer’s disease (AD) and mild cognitive impairment (MCI). Numerous risk factors, including genetic background, can affect the complex interplay between those mechanisms in the aging brain and can also affect typical AD hallmarks: amyloid plaques and neurofibrillary tangles. Our aim was to evaluate the association of polymorphisms in oxidative stress- and inflammation-related genes with cerebrospinal fluid (CSF) biomarker levels and cognitive test results. The study included 54 AD patients, 14 MCI patients with pathological CSF biomarker levels, 20 MCI patients with normal CSF biomarker levels and 62 controls. Carriers of two polymorphic IL1B rs16944 alleles had higher CSF Aβ1–42 levels (p = 0.025), while carriers of at least one polymorphic NFE2L2 rs35652124 allele had lower CSF Aβ1–42 levels (p = 0.040). Association with IL1B rs16944 remained significant in the AD group (p = 0.029). Additionally, MIR146A rs2910164 was associated with Aβ42/40 ratio (p = 0.043) in AD. Significant associations with cognitive test scores were observed for CAT rs1001179 (p = 0.022), GSTP1 rs1138272 (p = 0.005), KEAP1 rs1048290 and rs9676881 (both p = 0.019), as well as NFE2L2 rs35652124 (p = 0.030). In the AD group, IL1B rs1071676 (p = 0.004), KEAP1 rs1048290 and rs9676881 (both p = 0.035) remained associated with cognitive scores. Polymorphisms in antioxidative and inflammation genes might be associated with CSF biomarkers and cognitive test scores and could serve as additional biomarkers contributing to early diagnosis of dementia.


Introduction
Age, as a complex process, is an important risk factor for neurodegeneration. Alzheimer's disease (AD) is the most prevalent neurodegenerative disease, highly related to aging. Roughly 10-30% of people above 65 suffer from AD [1,2]. Earlier onset of clinical symptoms is uncommon and usually linked to familial genetic heritability [3]. Two major pathophysiological hallmarks of AD are deposition of amyloid β (Aβ) in neuritic plaques and accumulation of neurofibrillary tangles from tau protein. AD is often preceded by mild cognitive impairment (MCI), which is considered to be the pre-dementia stage of the disease. Severe memory and learning decline can be observed during disease progression from MCI to AD [4].
Oxidative stress is a biological process, driven by the imbalance between production of reactive oxygen species (ROS)-superoxide radical anion (O 2 − ), hydrogen peroxide (H 2 O 2 ), hydroxyl radical (HO − ), nitric oxide (NO) and peroxynitrite (ONOO − )-and antioxidative defense mechanisms [5]. In the AD brain, oxidative stress can be manifested through

Subjects
Our study included patients with cognitive impairment as they were coming for clinical evaluation and lumbar puncture appointment at the Department of Neurology, University Medical Centre Ljubljana, Slovenia, between June 2019 and July 2021. Inclusion criteria were age above 55 and diagnosis of AD or MCI. We excluded patients with physical diseases significantly affecting cognitive performance and dementia due to diseases other than AD. Patients and their caregivers underwent a structured interview to obtain demographic and clinical data. Additional information was obtained from medical records.
The study protocol was approved by the National Medical Ethics Committee of Republic of Slovenia (0120-523/2017-4) and all the subjects provided written informed consent in accordance with the Declaration of Helsinki.
Population-based group of elderly patients without diagnosed cognitive impairment was selected as a control group. All of them participated in previous pharmacogenetic studies and provided written consent to participate in further genetic studies. Information on all diagnoses and treatments was available for all participants. Inclusion criteria were age above 65. Exclusion criteria were cancer, neurodegenerative or other neurological disease, inflammatory diseases, joint dementia or anxiety.

Assesment
Dementia was diagnosed using a standardized clinical assessment and patients' history of cognitive decline. Cognitive screening was performed with the Mini-Mental State Examination (MMSE) [51]. A comprehensive diagnostic work-up, including structural brain imaging, blood laboratory tests, neuropsychological assessment and cerebrospinal fluid (CSF) dementia biomarker testing, was performed.
Patients were diagnosed as having dementia using DSM V criteria [52] after a consensus meeting with clinicians and neuropsychologists, taking into account all available information.
According to CSF biomarker levels, dementia criteria and Winblad & Peterson MCI diagnostic criteria [53], patients were stratified in three groups: AD, MCI (AD) and MCI (NOT AD). Locally validated biomarker cut-off levels were used for Aβ 42 (>570 pg/mL), Aβ 42 / 40 (>0.07), p-tau 181 (<60 pg/mL) and total tau (<400 pg/mL). Patients with elevated total and p-tau 181 and reduced Aβ 42 and Aβ 42 / 40 levels and with impaired daily activities were defined as AD group. Patients with MCI and AD CSF biomarker profile and normal daily functioning were included in MCI (AD) group. Patients with normal biomarker levels that had MCI and preserved daily functioning were defined as MCI (NOT AD) group.

Genotyping
Genomic DNA was isolated using the E.Z.N.A. ® SQ Blood DNA Kit II (Omega Bio-tek, Inc., Norcross, GA, USA) from peripheral blood samples according to the manufacturer's protocol. Genotyping was performed for 20 single-nucleotide polymorphisms (SNPs) in 13 genes; 5 of them are involved in oxidative stress mechanisms (SOD2, CAT, GPX1, KEAP1, NFE2L2), while the other 8 are important in the inflammation process (IL1B, MRNA146A, IL6, TNF, CARD8, NLRP3, GSTP1, NOS1). Polymorphism selection was performed according to published literature, as potentially functional SNPs were selected, with a minor allele frequency of at least 0.05. All of the studied SNPs were genotyped with competitive allele-specific PCR (KASP assays, LGC Biosearch Technologies, Hoddesdon, UK), according to manufacturer's instructions.
Additionally, APOE rs7412 and rs429358 were genotyped for the assessment of APOE4 status using real-time PCR-based Taqman assay (Applied Biosystems, Foster City, CA, USA). Combination of APOE rs429358 (p.Cys112Arg) and rs7412 (p.Arg158Cys) defines three polymorphic alleles, APOE2, APOE3 and APOE4, and was used for adjustment in statistical analysis. While APOE3 is the most common among different populations, APOE4 is considered to significantly increase AD risk [54,55]. Ten percent of samples were genotyped in duplicate as quality control and all the results were concordant.

Statistical Analysis
Continuous variables were described with median and interquartile range (25-75%), while categorical variables were described with frequencies. Interquartile range was determined using weighted averages if more than two samples were included in the group and using Tukey's hinges if two samples were included in the group. Fisher's exact test, Mann-Whitney and Kruskal-Wallis tests were used to compare patients' characteristics between groups. The agreement of genotype frequencies with Hardy-Weinberg equilibrium (HWE) was examined by chi-squared test. Both dominant and additive genetic models were used in the analysis. Logistic regression was used to evaluate the association of selected SNPs with binary categorical variables and to calculate the odds ratios (ORs) and their 95% confidence intervals (CIs). Fisher's exact test was used if there were no subjects within one of the groups and for dependent categorical variables with more than two categories. Mann-Whitney or Kruskal-Wallis tests with post hoc Bonferroni corrections for pairwise comparisons were used to evaluate the association of SNPs with MMSE and biomarker levels. Bonferroni correction was used to account for multiple comparisons to decrease the chance of false-positive results. The significance threshold was set to 0.0025 and p-values below 0.0025 were considered statistically significant, while p-values between 0.0025 and 0.050 were considered nominally significant. IBM SPSS Statistics version 27.0 (IBM Corporation, Armonk, NY, USA) was used for all analyses. All tests were two-sided and the level of significance was set at 0.05. GraphPad Prism version 9 (GraphPad Software, LLC., San Diego, CA, USA) was used for preparation of figures.
For the comparison of genotype frequencies between cases and controls, this study had 80% power to detect ORs of approximately 2.6 or more for polymorphisms with minor allele frequency between 0.20 and 0.40, while we could detect ORs of 3.6 or more for

Patients' Characteristics
Our study included 150 subjects: 88 patients with memory deficits, of which 54 were AD patients, 14 MCI patients with pathological CSF biomarker levels (MCI (AD)), 20 MCI patients with normal CSF biomarker levels (MCI (NOT AD)) and 62 populationbased controls without diagnosed cognitive impairment. Median age of all dementia patients at enrolment was 77 (72.25-80) years and differed significantly from controls (69.2 (66.48-73.18), p < 0.001). There were more female subjects in both groups (dementia patients: 56.8%, controls: 66.1%, p = 0.309). Clinical characteristics of all patients with dementia and of individual groups (AD, MCI (AD) and MCI (NOT AD)) are summarized in Table 1. AD patients were significantly older compared to patients with MCI (AD) and MCI (NOT AD) (p = 0.010). APOE4 carriers were most frequent in the AD group, but differences were not significant (p = 0.138). Significant differences in all CSF biomarker levels (Aβ, Aβ 42/40 , total tau and p-tau 181 ) were observed between groups (all p < 0.001). AD patients also achieved significantly lower results in the cognitive test (p < 0.001).   Genotype frequencies of all 20 investigated SNPs in SOD2, CAT, GPX1, IL1B, MIR146A, IL6, TNF, CARD8, NLRP3, GSTP1, NOS1, KEAP1 and NFE2L2 genes in the whole cohort are presented in Table S1. Genotype frequencies of NOS1 rs2293054, KEAP1 rs9676881 and NFE2L2 rs6706649 were not in agreement with HWE in the control group. However, due to the fact that the control population was not randomly selected, we included those genes in further analysis.

Association of Investigated SNPs with Dementia and AD Susceptibility
A comparison of genotype frequencies between all dementia patients and controls is presented in Table 2. Subjects who were heterozygous for TNF rs1800629 polymorphism were less likely to have dementia (p = 0.017, OR = 0.368 (0.163-0.834)). The protective association with dementia susceptibility remained nominally significant, even after adjustment for age and APOE carrier status (p = 0.011, OR = 0.274 (0.101-0.741)). In the dominant model, the association was also nominally significant, both in univariable and multivariable analysis (both p = 0.049, Table 2). On the other hand, all carriers of two NOS1 rs2293054 G alleles had dementia (p = 0.019).   Genotype distribution of investigated polymorphisms in different dementia pathologies is presented in Table S2. At least one polymorphic NFE2L2 rs35652124 C allele was more common in AD compared to both MCI groups (p = 0.021). No significant or nominally significant differences were observed for other investigated SNPs.
Additionally, we compared genotype frequencies of only AD patients with controls (Table S3). Carriers of one polymorphic IL1B rs16944 C allele tended to have lower risk for AD (p = 0.049, OR = 0.315 (0.099-0.997)). Association remained nominally significant, even after the adjustment for age and APOE carrier status (p = 0.039, OR = 0.206 (0.046-0.924)). Carriers of two polymorphic NOS1 rs2293054 G alleles were present among AD patients only (p = 0.012). Carriers of at least one polymorphic TNF rs1800629 allele were less likely to have AD, when adjusting for age and APOE carrier status (p = 0.025, OR = 0.246 (0.072-0.840)).

Association of Investigated SNPs with CSF Biomarker Levels and MMSE
Among investigated SNPs, three were associated with different CSF biomarkers among all patients with dementia ( Table 3). Carriers of two polymorphic IL1B rs16944 C alleles had higher Aβ levels (p = 0.020) (Figure 1a). On the other hand, decreased Aβ levels were observed in carriers of one polymorphic NFE2L2 rs35652124 C allele (p = 0.031). The association did not reach statistical significance or nominal significance in the dominant model (p = 0.053). For NFE2L2 rs6721961, decreased total tau was observed in carriers of at least one polymorphic T allele (p = 0.020).   The association of IL1B rs16944 C allele with Aβ remained nominally significant when only patients with AD were included in the analysis (Figure 1b, Table S4); carriers of two polymorphic C alleles had higher Aβ levels (p = 0.038). Additionally, carriers of at least one polymorphic C MIR146A rs2910164 allele had higher Aβ 42/40 ratio (p = 0.043).
Nominally significant associations with the MMSE cognitive test were observed in the dementia group for five of the selected SNPs (Table 4). Higher test scores were observed in carriers of two polymorphic alleles for CAT rs1001179 (p = 0.022), KEAP1 rs1048290 (p = 0.019, Figure 2a) and KEAP1 rs9676881 (p = 0.019). Furthermore, carriers of at least one polymorphic GSTP1 rs1138272 T allele had higher test scores (p = 0.005). Conversely, polymorphic NFE2L2 rs35652124 C allele was associated with lower test scores in both the additive (p = 0.030) and dominant model (p = 0.024). In AD patients, lower cognitive test scores were observed in patients with the heterozygous IL1B rs1071676 genotype (p = 0.004), while carriers of two polymorphic KEAP1 rs1048290 alleles (p = 0.035, Figure 2b) and KEAP1 rs9676881 (p = 0.035) had higher MMSE scores (Table 4). the dementia group for five of the selected SNPs (Table 4). Higher test scores were observed in carriers of two polymorphic alleles for CAT rs1001179 (p = 0.022), KEAP1 rs1048290 (p = 0.019, Figure 2a) and KEAP1 rs9676881 (p = 0.019). Furthermore, carriers of at least one polymorphic GSTP1 rs1138272 T allele had higher test scores (p = 0.005). Conversely, polymorphic NFE2L2 rs35652124 C allele was associated with lower test scores in both the additive (p = 0.030) and dominant model (p = 0.024). In AD patients, lower cognitive test scores were observed in patients with the heterozygous IL1B rs1071676 genotype (p = 0.004), while carriers of two polymorphic KEAP1 rs1048290 alleles (p = 0.035, Figure  2b) and KEAP1 rs9676881 (p = 0.035) had higher MMSE scores (Table 4).      served in carriers of two polymorphic alleles for CAT rs1001179 (p = 0.022), KEAP1 rs1048290 (p = 0.019, Figure 2a) and KEAP1 rs9676881 (p = 0.019). Furthermore, carriers of at least one polymorphic GSTP1 rs1138272 T allele had higher test scores (p = 0.005). Conversely, polymorphic NFE2L2 rs35652124 C allele was associated with lower test scores in both the additive (p = 0.030) and dominant model (p = 0.024). In AD patients, lower cognitive test scores were observed in patients with the heterozygous IL1B rs1071676 genotype (p = 0.004), while carriers of two polymorphic KEAP1 rs1048290 alleles (p = 0.035, Figure  2b) and KEAP1 rs9676881 (p = 0.035) had higher MMSE scores (Table 4).

Discussion
Our study evaluated the effect of selected SNPs from oxidative stress and inflammation pathways on the risk for AD and assessed their associations with CSF biomarkers and cognitive test results. Among inflammatory genes, polymorphisms in IL1B and TNF showed a protective function in the development of dementia or AD, while MIR146A was associated with CSF biomarkers. Observed associations of polymorphisms in CAT, GSTP1, NFE2L2 and KEAP1 with dementia support the importance of antioxidative mechanisms in AD.
Several studies evaluating the effect of genetic polymorphisms in the oxidative stress-and inflammation-related pathways on AD risk are summarized in multiple metaanalyses [57][58][59]. Although the association of genetic variability with most prominent AD CSF biomarkers (Aβ 42 , total-tau, P-tau 181 ) is readily studied, oxidative stress-and inflammation-related genes are usually overlooked. Only a couple of studies assessed the association of polymorphisms in cytokine-encoding genes with AD biomarker levels [60][61][62]. To see a broader picture of the influence of oxidative stress and inflammation pathways on AD genetic background, we selected a comprehensive list of SNPs and assessed their correlations with disease pathology.
Among antioxidative genes, CAT, GSTP1, NOS1, NFE2L2, and KEAP1 were associated with dementia or AD. CAT and GSTP1 genetic variability was thoroughly studied in oxidative-stress-related diseases, especially different types of cancer, while in neurodegeneration, NFE2L2 and KEAP1 have recently gained the interest of researchers. On the other hand, no significant association in NOS1 on AD susceptibility is published to date.
We observed an association of CAT rs1001179 with MMSE in all patients with dementia. To date, no CAT polymorphisms are associated with AD or MCI [63]. Catalase is a ubiquitous antioxidative enzyme, catalyzing the reductions of hydrogen peroxide to water. Aβ can cause an accumulation of hydrogen peroxide, since Aβ-driven reductions in catalase activity have been observed [64,65], thus, promoting oxidative stress in AD [66]. The CAT rs1001179 polymorphism leads to alterations in the transcription factor binding site at the promoter region and affects catalase blood levels [67]. Although previous work found no significant effect of CAT rs1001179 on AD risk [63], our association between MMSE and rs1001179 implies catalase might contribute to dementia.
Observed association of GSTP1 rs1138272 with higher MMSE scores is, to the best of our knowledge, a novel finding. GSTP1 rs1138272 is a missense polymorphism that can alter catalytic activity of glutathione S-transferase (GST) in oxidative stress [68,69]. GST gene family is important in glutathione-induced detoxification mechanisms, preventing oxidative damage of biomolecules [70,71]. Increased levels of GST pi (π), encoded by GSTP1, were found in an AD mouse model [72]. Moreover, increased activity of both catalase and GSTs has been found in CSF [73][74][75] and blood samples [76] of patients with different dementia types. Previously, the GSTP1 rs1695 polymorphism was associated with increased AD risk in different studies and one meta-analysis [58,77,78]. Association of GSTP1 rs1695 with MMSE was also reported [79], but these results were not confirmed in our study.
The presence of the NOS1 rs2293054 G allele only in patients with dementia or AD, in our study, was a surprising outcome. NOS1 encodes the neuronal isoform of the nitric oxide synthase that is the main source of brain nitric oxide (NO) [80]. Although NOS1 rs2293054 has not yet been studied in AD, association with clinical phenotypes of ischemic stroke [81] and Parkinson's disease treatment [82] was observed. In accordance with our results, higher AD risk was found in another functional NOS1 promoter polymorphism in two independent cohorts [83,84]. Although the investigated antioxidative enzymes were not associated with CSF biomarker levels in our study, they could serve as additional biomarkers of dementia.
Multiple effects of NFE2L2 and KEAP1 on dementia and AD were found in our study. An association with Aβ 42 levels and MMSE scores was found for NFE2L2 rs35652124 in the combined group with dementia. NFE2L2 rs6721961 was also associated with tau levels in AD and MCI. Additionally, we found association with MMSE in KEAP1 rs1048290 and rs9676881, both in the whole cohort as well as in AD patients only. Nuclear factor E2related factor 2 (NRF2, encoded by NFE2L2 gene) and kelch-like ECH-associating protein 1 (KEAP1) are two of the main regulators of redox balance, crucial for human stress response. They are involved in response to oxidative stress, but also important in metabolism of xenobiotics and inflammatory response (reviewed in [13]). Consistent with our results, different studies investigating NFE2L2 and KEAP1 polymorphisms found no association with AD risk; however, one NFE2L2 haplotype was linked to faster disease progression [85]. In the same study, no association with CSF biomarkers and MMSE in AD patients was observed, which is in contrast with our findings. NRF2 is aberrantly expressed in different brain cells, with an observed age-associated decrease [86] and reduced expression in AD patients [87]. NRF2 deficit was associated with AD pathology: NRF2 knockout in mice leads to an increase in Aβ and tau levels [88,89] and worsens cognitive decline [90]. The NRF2-KEAP1 signaling axis is thoroughly studied in neurodegeneration, with the focus on potential therapeutic application [13,91,92]. Our observed associations for NRF2 and KEAP1 support the potential role of the NRF2-KEAP1 axis in neurodegeneration and could partially explain the missing connection between oxidative stress and neurodegeneration.
In the present study, the effect of IL1B, MIR146A and TNF on dementia or AD susceptibility confirmed previously published data [27,[60][61][62]. We observed a protective effect of the TNF rs1800629 A allele in a combined group of all patients with AD or MCI, while the results did not reach significance in the AD group only. TNF-α is one of the most studied inflammatory cytokines in AD as TNF-α plays a central role in the cytokine cascade during microglial activation. Subsequently, TNF-α can induce neuronal death by activating TNF receptor 1 if the NF-κB pathway is inhibited [93]. TNF rs1800629 polymorphism is part of the promoter region and altered expression of TNF-α and soluble TNF receptors has been linked to TNF rs1800629 and other promoter polymorphisms [94][95][96][97][98]. Previously, the protective role of this polymorphism was observed in a Finnish population [61] but was not confirmed in another independent cohort [60]. Consistent with our results, no differences among AD and MCI were observed [99]. TNF-α levels in blood and CSF were associated with MCI and AD [100], and elevated TNF-α appears to correlate with disease progression [101]. However, the results regarding the association of TNF rs1800629 with biomarker levels differ among studies: TNF rs1800629 was previously associated with AD CSF Aβ 42 and P-tau 231 in some studies [61,62], while no association was observed in other studies [60,102], similarly to our results, suggesting further studies are needed to evaluate the association of this polymorphism with AD biomarkers.
In our study, IL1B polymorphisms were associated with AD risk, pathology and cognitive decline. IL1B rs16944 genotype frequencies tended to differ between AD patients and controls. The association between IL1B rs16944 and higher Aβ 42 levels was observed in both the all dementia and AD group. Additionally, the heterozygous IL1B rs1071676 genotype was associated with lower MMSE scores in AD. IL-1β is another cytokine that has been extensively studied in AD. IL-1β induces synaptic loss by simultaneously activating multiple pathways that require both pre-and post-synaptic activity [103]. Different IL1B polymorphisms can affect IL-1β expression [104][105][106]. IL1B rs16944 is located in the promoter region and can affect IL-1β expression, while rs1071676 in 3 untranslated region of IL1B gene could modify miRNA binding. Selected SNPs did not reach significant associations with AD patients in any previous studies, including meta-analyses [59,[107][108][109]. However, increased CSF p-tau and t-tau levels were found in AD carriers of the IL1B rs1143623 polymorphic G allele [62], which was not replicated in our study. With that in mind, our results add to the importance of IL1B genetic variability in AD or dementia.
MiR-146a is another important factor for inflammation, interacting with IL-1β and TNF-α [110]. An inhibitory function of miR-146a on IL-1β secretion through IRAK1 was found [111,112]. In our study, an SNP in the coding region of primary hsa-miR-146a, MIR146A rs2910164, was associated with higher Aβ 42/40 ratio in carriers of at least one polymorphic allele. The rare C allele was previously linked to increased AD risk and it was proposed that it could affect binding to miRNA target genes [27]. Consistent with our results, MIR146A rs2910164 was not associated with MMSE scores [113]. On the other hand, a different MIR146A polymorphism, rs57095329, was more important for miR-146a expression and AD susceptibility in the same study. To the best of our knowledge, none of the studies investigated the association of this polymorphism with CSF biomarkers. However, the expression of miR-146a was altered in AD patients [114][115][116]. Taken together, these findings support the important interplay between cytokines (TNF-α, IL-1β) and regulatory elements (miRNA-146a) in neuroinflammation.
Our study has some limitations. The sample size was relatively small and some clinical parameters, especially cognitive test scores, were not available for all patients. The smaller sample size is partly due to the fact that the patients were included during their lumbar punction appointment to assess CSF biomarkers. Although it is difficult to detect the contributions of many factors in a study with a smaller sample size, a similar effect can occur due to phenotypic heterogeneity in larger studies as well. We also accounted for multiple comparisons in the statistical analysis. Another limitation was the observed deviation from HWE for some SNPs in the control group that can be partially explained with the non-random selection of patients. A larger and more homogenous control group would help in overcoming that issue. On the other hand, our study had several strengths. All the patients were recruited from the same department and evaluated according to the same protocol. We comprehensively assessed the simultaneous influence of several clinical and genetic parameters on AD risk and pathology. We were the first to assess the genetic variability in oxidative stress and inflammatory pathways among Slovenian patients. Furthermore, only a few studies focused on the association of selected genes with CSF biomarkers. Usually, a single gene effect on AD pathology is of interest, so our pathway-based approach helps to elucidate the broader picture of disease mechanisms.
We were among the first to highlight the potential of the KEAP1-NRF2 axis in AD and dementia, opening up a new window with therapeutic potential. Due to the versatile role of the KEAP1-NRF2 axis in numerous diseases with oxidative stress and inflammation as underlying pathological features, the protective effects of NRF2 have been a research focus lately [117]. In the field of autoimmune diseases, NRF2-activator dymethyl fumarate is already approved for use in psoriasis and MS. However, activation of NRF2 could be one of the promising candidates in a multi-target therapy approach in AD. To date, several NRF2 activators for AD treatment have advanced to preclinical studies in mouse AD models and clinical trials [118]. Genetic variability in this pathway could also contribute to a better understanding of the mechanisms underlying the KEAP1-NRF2 axis in the search for suitable therapy approaches.

Conclusions
Our data suggest that genetic variability in oxidative stress-and inflammation-related genes might affect susceptibility for AD and MCI, and may be associated with CSF biomarkers levels and cognitive test scores. Observed associations support the important role of oxidative stress and neuroinflammation pathways in AD pathogenesis and could contribute to a better understanding of dementia and help identify additional biomarkers contributing to early diagnosis of cognitive decline.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/antiox12020316/s1, Table S1: Genotype frequencies of selected polymorphisms; Table S2: Comparison of genotype frequencies among patients with different types of dementia; Table S3: Comparison of genotype frequencies among patients with AD and controls; Table S4: Association of investigated polymorphisms with cerebrospinal fluid biomarkers among patients with AD.  Informed Consent Statement: Informed consent was obtained from all subjects involved in the study.
Data Availability Statement: All the data are presented within the article and in Supplementary Material. Any additional information is available on request from the corresponding author.