Introduction

Alzheimer’s disease (AD) is the most common cause of dementia in the elderly, affecting nearly 35 million people worldwide [1]. This devastating disease is characterized by progressive neurodegeneration with multifactorial etiology. AD dysfunctions include calcium dysregulation, proteolysis failure, altered cell signaling, mitochondrial dysfunction, and oxidative stress, which lead to synaptic dysfunction, nerve cell death, and neurodegeneration [2, 3].

Oxidative stress is accepted to play a key role in the pathology of AD. This process results from an imbalance between elevated free radical production and the decrease in either free radical scavenging or the mechanisms used to repair oxidized macromolecules. The overproduction of free radicals is capable of damaging proteins, lipids, and DNA, leading to cellular instability events, such as chromosomal damage and cell death [4,5,6,7]. These events can be easily measured by the micronucleus assay, considered a reliable biomarker of genetic instability in numerous applications [8,9,10,11].

Genes involved in cellular mechanisms of oxidative damage repair are strong candidates for AD [12]. These genes include SIRT1 (Silent Information Regulator Type 1), located on chromosome 10q21.3, and SOD2 (Superoxide dismutase 2), on chromosome 6q25.3 [13]. Since the SIRT1 protein can increase life span through the regulation of cellular metabolism, it is a possible protective factor for AD [14]. The gene SOD2 encodes a protein involved in the repair of oxidative damage and is located in a region that shows evidence of a relationship with AD in genome-wide association studies (GWAS) [15]. Despite the large number of genes supposedly related to AD, to date, the ε4 allele of the APOE gene on 19q13.2 is considered the major risk factor for the disease in several populations [16]. The study of genetic variants associated with the risk for AD is useful to understand the mechanisms of the disease, aiding in complementary diagnosis.

Additionally, lipid peroxidation is one of the most important manifestations of oxidative stress and results in the production of toxic aldehydes, such as malondialdehyde (MDA), a mutagenic compound measured by the chemical determination of thiobarbituric acid reactive substances—TBARs [17,18,19,20]. The search for biomarkers of genomic instability is fundamental to improving the implementation of diagnosis and treatment. In this context, the main objectives of this study were to validate the association of single-nucleotide polymorphisms (SNPs) in the SIRT1 and SOD2 genes with late-onset Alzheimer’s disease and to estimate the level of cellular and genomic damage and lipid peroxidation in AD patients.

Materials and methods

Subjects

All patients fulfilled the clinical criteria for probable AD according to the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV) and the National Institute of Neurological and Communicative Diseases and Stroke/Alzheimer’s Disease and Related Disorders Association (NINCDS/ADRDA). The complete evaluation of dementia were performed using imaging exams (computed tomography or magnetic resonance imaging); standard laboratory tests (complete blood count, serum electrolytes, serum glucose, blood urea nitrogen, vitamin B12, folate, thyroid function, and syphilis serology); the Clinical Dementia Rating (CDR) scale, which categorize functional impairment of dementia in values as 0 = normal; 0.5 = borderline; 1 = mild; 2 = moderate; and 3 = severe; and the Mini-Mental State Examination (MMSE) [21,22,23,24,25].

For controls samples, the geriatrician meticulously selected only participants with no family history of Alzheimer’s and the non-dementia elders using the Mini-Mental State Examination (MMSE) (mean in controls samples = 21 ±5). MMSE evaluates that cognitive status and several factors as sociocultural variables, age, and education levels could affect individuals scores [26, 27].

For genotyping, a total of 332 non-consanguineous subjects, comprising 109 late-onset AD patients (mean age: 82.3 ± 7.6) and 223 no demented healthy controls (mean age: 81.2 ± 9.9) from a single geographical location in Vitória, ES, Southeast of Brazil, were enrolled in the study. Peripheral blood (10 ml) was collected in 5% ethylenediaminetetraacetic acid (EDTA) tubes (Vacuette, Greiner Labortechnik, Germany).

In this study, we used the cytokinesis-block micronucleus cytome (CBMNcyt) assay in peripheral blood lymphocytes collected with heparin and the buccal micronucleus cytome (BMNcyt) assay in exfoliated buccal cells. For these two assays, cells were collected from 20 subjects. Both the AD and the control groups consisted of 7 females and 3 males (mean age: 81.0 ± 7.3 and 82.0 ± 8.7, respectively, p = 0.909), with patients presenting the disease for a mean duration of 3.8 ± 3.0 years (range: 1–10 years). All patients were on cholinergic therapy at the time of enrollment in the study. During a face-to-face interview, volunteers or their caregivers were asked to answer an adapted questionnaire from the International Commission for Protection against Environmental Mutagens and Carcinogens [28] to minimize confounding factors; all subjects in these protocols were purposely selected to be non-smokers, non-alcohol drinkers with no recent X-ray exposure. AD patients and controls were matched for sex, age, and ethnic background.

For the TBARs study, blood samples were collected from 50 subjects by venipuncture with EDTA. Both the AD and the control groups consisted of 20 females and 5 males (mean age: 80.0 ± 5.5 and 80.5 ± 10.7, respectively, p = 0.884), with patients presenting the disease for a mean duration of 4.3 ± 3.0 years (range: 1–10 years).

This study was approved by the Research Ethics Committee of Emescam (School of Health Sciences of Santa Casa of Vitória), and written informed consent was obtained from each subject or from his/her surrogate prior to his/her inclusion into the study. All experiments were conducted in accordance with the Declaration of Helsinki.

Genotyping

For genotyping, the samples were stored at 4 °C prior to analyses. The concentration and purity of genomic DNA were measured using a NanoDrop 2000 Spectrophotometer (Thermo Fisher Scientific, DE, USA). SNPs in the APOE (rs429358: T > C, RefSeq NG_007084.2 and rs7412: C > T, RefSeq NG_007084.2), SIRT1 (rs2273773: C > T, RefSeq NG_050664.1) and SOD2 (rs4880: A > G, RefSeq NM_000636.3) genes were genotyped using real-time quantitative polymerase chain reaction (qPCR).

For each SNP, three standard genotypes identified through Sanger sequencing were used in all reactions of qPCR. The sequencing analysis was conducted using BioEdit software v.7.2.5 for Windows (Ibis Biosciences, Carlsbad, CA, USA) in ABI PRISM 3130XL Genetic Analyzer/HITACHI (Applied Biosystems, Carlsbad, California, USA).

Genomic DNA (30 ng/μl) was employed for qPCR on a 7500 Fast Real-Time PCR System according to the manufacturer’s instructions (TaqMan SNP Genotyping Assay – Applied Biosystems, Carlsbad, CA, USA). Genotypes were analyzed using SDS v.2.0.5 software.

CBMNcyt assay

The lymphocytes were cultured according to the methodology developed by Fenech and Morley (1985) [29]. For each subject, 2000 binucleated cells with well-preserved cytoplasm (1000 cells from each of two replicate cultures) were scored for the presence of micronuclei (MN), nucleoplasmic bridges, and buds, following the criteria described by Bolognesi and Fenech (2013) [30]. Frequencies of cells carrying these abnormalities were expressed per 1000 binucleated cells scored. The cell cycle alteration was expressed by the variations of the Nuclear Division Index (NDI), which was calculated by counting 1000 cells per subject according to Eastmond and Tucker (1989). The analyses were performed using an optical microscope (× 1000).

BMNcyt assay

Subjects were asked to rinse their mouths with water, and a premoistened cytobrush was used to sample cells from both sides of the interior cheek. The swab was immersed in 5 ml cold saline (0.9% w/v, aqueous NaCl) in a conical tube. The samples were centrifuged at 1500 rpm for 10 min, and the resulting pellets were washed three times more with 2 ml of saline under the same centrifugation conditions. The cell suspension was dropped onto a slide, air-dried, and fixed in 80% methanol. Slides were stained with pure Leishman for 2 min followed by 15 min in 10% Leishman aqueous solution, rinsed in distilled water, and air-dried. We used the criteria of scoring described by Thomas et al. (2009). The BMNcyt assay measures biomarkers of DNA damage (MN and buds), cell death (condensed chromatin, pyknosis, karyorrhexis and karyolysis), and cytokinetic defects (binucleated cells). Two thousand cells per subject (1000 from each duplicate slide) were scored blindly by the same person, using optical microscopy (× 1000).

TBAR assay

In this work, we adopted a modified method from Buege and Aust (1978) [31]. Blood samples were kept in an ice bath until the time of centrifugation, which was performed at 1500 rpm for 10 min. Plasma samples were stored at − 80 °C until analysis. To a volume of 1 ml of plasma, 2 ml of aqueous acid solution (15% trichloroacetic acid, 0.375% TBA, 0.25 N HCl, and 2.5 mM butylated hydroxytoluene—BHT), diluted in ethanol, was added. BHT, an antioxidant, was added to prevent MDA formation, which could result in falsely elevated TBA reactivity, during the assay. Samples were heated for 30 min at 95 °C, cooled at room temperature, and centrifuged at 3000 rpm for 10 min. Absorbance values were read in a spectrophotometer at 532 and 572 nm. Absorbance at 572 nm was subtracted from absorbance at 532 nm. MDA values were estimated with the extinction coefficient of MDA-TBA complex at 532 nm = 1.56 × 105 cm-1 M-1 [32].

Statistical analysis

To test the association between AD and the SNPs, Pearson’s chi-square and Fisher’s exact tests and logistic regression were performed. Additionally, the odds ratio (OR), confidence interval (CI, 95%), and Hardy-Weinberg equilibrium (H-WE) were calculated. Statistical analysis was performed using SPSS software v23.0 for Windows (IBM Corporation, Armonk, New York, USA). For the CBMNcyt, BMNcyt, and TBARs assays, the normality of the data was checked using the D’Agostino-Pearson test. Comparisons of mean values between groups were examined by parametric (unpaired Student’s t test) or non-parametric (Mann-Whitney U) tests, when applicable. Data were expressed as the mean ± standard deviation, and p < 0.05, for a two-tailed test, was considered statistically significant.

Results

Genotyping

The characteristics of each sample group, such as ethnic background, sex, age, APOE status, and level of education, are shown in Table 1. The education of almost 90% of the literate of the total sample had less than 4 years of study. No significant differences between the samples were observed in relation to ethnic background, sex, or age. However, a low education level appeared to affect the predisposition for the disease.

Table 1 Characteristics of the sample
Full size table

The genotypic frequencies of the SNPs and results of logistic regression analysis are shown in Tables 2 and 3, respectively. As expected, the APOE ε4ε4 genotype frequency was significantly increased in AD patients. We also found a significant value for the APOE ε3ε3 genotype. A positive association with AD was detected for a SIRT1 variant but not for SOD2. All of the polymorphisms were in H-WE equilibrium in both the controls and AD patients.

Table 2 Genotype frequencies of polymorphisms in the study
Full size table
Table 3 Logistic regression analysis of polymorphisms in the study
Full size table

Genomic instability and oxidative stress

The results of the CBMNcyt assay in blood lymphocytes and the BMNcyt assay in epithelial buccal cells are summarized in Table 4. We detected higher frequencies of MN, nucleoplasmic bridges, and buds in lymphocytes from AD patients, compared with controls. Moreover, the evaluation of epithelial buccal cells also revealed elevated frequencies of MN, buds, binucleated cells, condensed chromatin, karyorrhexis, and pyknosis in AD patients, in relation to controls.

Table 4 The CBMNcyt assay in peripheral blood lymphocytes and the BMNcyt assay from AD patients and healthy controls.
Full size table

We performed a TBAR assay to estimate the level of MDA, a biomarker of lipid peroxidation. The mean level of MDA in plasma was significantly higher in the AD patients than in the controls (Fig. 1).

Fig. 1
figure 1

Comparison of MDA levels in plasma of AD patients (n = 25) and healthy controls (n = 25)

Full size image

Discussion

The current study allowed the identification of positive associations of SIRT1 (rs2273773) and APOE (rs429358 and rs7412) polymorphisms with AD. Significant results were also found in AD patients for chromosomal damage and a higher level of plasma malondialdehyde. Our results support the hypothesis that oxidative stress may play an important role in the etiology of AD.

Several lines of evidence suggest that oxidative damage is involved in the aging process and, particularly, in neurodegenerative diseases such as AD. The brain metabolism of AD individuals appears to be more affected by reactive oxygen species damage than that of healthy older adults [17, 33]. This may partially explain the increase in oxidative damage to nucleic acids and a mitochondrial dysfunction in the brain affected by AD [34,35,36].

Our results show that individuals carrying the CT genotype of the SIRT1 rs2273773 variant have a greater probability of developing AD than those who carry the TT genotype. Data provide evidence for implication of the TT genotype as a protective factor for AD. We did not find a positive result for the CC genotype, probably because of the sample size. SIRT1 is one of the sirtuins involved in the control of apoptosis, cell survival, the modulation of reactive oxygen species (ROS) levels, and mitochondrial biogenesis and function [37]. It constitutes a molecular link between aging and human degenerative disorders [38].

The SIRT1 rs2273773 polymorphism has been analyzed in several studies for a role in neuroprotection against the formation of ROS in AD and other neurotoxic conditions [39, 40]. As SIRT1 can regulate the aging and metabolic processes involved in the pathogenesis of AD, it may represent a potential therapeutic target. Our results support the hypothesis that the SIRT1 rs2273773 polymorphism is associated with AD and reinforces the supposition that oxidative stress can be involved in its pathogenesis.

Some genetic variants increase the risk of developing AD, particularly the ε4 allele of the Apolipoprotein E (ApoE) gene, whose product is involved in the transport of cholesterol [41, 42]. In our work, an AD patient who carries two copies of this allele is 8 times more likely to develop the disease than those who develops the disease but carries only one or no copies of the ApoE-ε4 allele (data not shown). Carrying the APOE-ε4 allele is an established risk factor for sporadic AD [43, 44]. Our results showed that APOE-ɛ4ɛ4 AD patients present an increased risk of developing AD, as reported worldwide, including in populations from North America, Europe, and Asia [44,45,46]. Regarding the APOE-ɛ3ɛ3 genotype, our findings indicated that this SNP acts as a protective factor, which is in accordance with the previous reports in an Italian population [47]. The ɛ3 allele may act similarly to an antioxidant agent [48], protecting against AD [49]. However, there is no consensus regarding the effective role of the APOE-ɛ3 allele in the pathology of AD in different populations [49, 50].

SODs are crucial to the endogenous signaling system that protects cells and tissues from oxidative stress [51]. Although functional studies have shown a role for SOD2 in the disease, in the present study, the SOD2 polymorphism (rs4880) did not exhibit an association with AD. Early studies employing the CBMNcyt assay to assess chromosomal damage in cells of AD patients reported a higher frequency of MN in peripheral lymphocytes of these patients than in those of controls [52, 53]. In our experiments, we additionally reported the frequencies of nucleoplasmic bridges and buds in the lymphocytes of AD patients. The inclusion of these markers of chromosomal damage allows the assessment of complementary events of genomic instability. The BMNcyt assay in MN-exfoliated buccal cells is a potential biomarker of genome instability. This assay has been used to measure increased risk for accelerated aging, cancer, and neurodegenerative diseases [54] as well as frailty syndrome and mild cognitive impairment [55, 56]. In our study, we adopted the cytome approach to score not only MN but also other biomarkers of nuclear abnormalities (buds), cytokinetic defects (binucleated cells), and cell death (condensed chromatin, karyorrhectic, karyolitic and pyknotic cells).

Using the buccal micronucleus cytome assay, Thomas and Fenech (2007) [57] detected a slight increase in MN frequency in their AD cohort, but this increase failed to reach significance. They also reported significantly lower frequencies of condensed chromatin cells and karyorrhectic cells in AD patients than in controls. Their results are in the opposite direction of the results obtained in our study; we found significantly increased frequencies of these markers in AD patients compared with controls. Additionally, we detected significantly enhanced frequencies of buds, binucleated cells, and pyknosis in AD patients. Such different results may be partially explained by the features of the cohorts: our AD cohort comprised patients who had been diagnosed with the disease for a mean period of 3.8 ± 3.0 years and were undergoing treatment with cholinesterase inhibitors at the time of enrollment in the study. In contrast, the study by Thomas and Fenech (2007) [57] comprised patients newly diagnosed with AD prior to the commencement of any treatment. We suggest that DNA damage increases with the progression of the disease, due to the elevated oxidative stress, a key factor in the degenerative neuronal death and progression of Alzheimer’s disease [7, 58]. Nevertheless, these changes indicate that the buccal cells of AD patients show significant alterations in cellular kinetics and may be useful as predictive biomarkers in identifying individuals with elevated risk for developing or having AD.

We used the TBAR assay to estimate the level of MDA, the most abundant aldehyde resulting from lipid peroxidation [20]. Measurement of MDA by TBAR assay is the most widely used method to estimate the overall lipid peroxidation level [59]. Similar to the previous studies, we detected a significantly higher plasma level of MDA in AD patients than in controls, supporting the hypothesis that oxidative stress may be related to the etiology of AD [60, 61]. Since MDA is a toxic and mutagenic product and can damage genomic material and the membranes of cells [18, 19], we suggest that the higher frequencies of DNA damage observed in the CBMNcyt and BMNcyt in AD patients may be a consequence of elevated oxidative damage.

Therefore, our results suggest that AD patients show increased oxidative stress and genomic instability as compared with controls and the CT polymorphism of SIRT1 is associated with increase the risk of the AD.

Conclusion

In conclusion, our results show that SIRT1 and APOE variants are associated with AD, a condition with increased oxidative stress and genomic instability characterized by elevated lipoperoxidation and genomic and cell damage. Part of the discrepancy observed in our results and previous publications could be explained due to the presence of confounders factors in the samples not evaluated in this study such diet, antioxidant intake and inflammation. Although further research is required to achieve an overall understanding of the disease, we believe that our findings are important for the characterization of genes and other factors related to AD that could be useful in the future to improve strategies of treatment and diagnosis of the disease.