Lower Prevalence of Alzheimer’s Disease among Tibetans: Association with Religious and Genetic Factors

Subjects

From May 2014 to September 2014, a population-based cross-sectional survey with a multistage cluster sampling design was carried out. Subjects were recruited from the Qinghai province of China. First, four representative cities were selected in Qinghai. Then, six counties within the selected cities were chosen randomly. Lastly, 65 rural villages within the selected counties were sampled randomly.

Then, to be selected, participants had to be unrelated Tibetans aged >60 years who had been living for at least 10 years in Qinghai and whose ancestors had been living in Qinghai. Participants were excluded in the presence of: 1) dementia caused by poisoning or systemic diseases; 2) pseudo-dementia caused by depression; 3) cerebrovascular or nervous system diseases (including vascular dementia); 4) severe acute disease or with malignant tumor; or 5) recent history of suffering from cold, chronic inflammation, acute disease, metabolic disease, or infectious disease.

This study was approved by the Ethics Committee of our hospital. Written informed consent was obtained from each participant or from the legal guardians in case of cognitive impairment. The study was done in accord with the ethical standards of the Committee on Human Experimentation, and in accord with the Helsinki Declaration of 1975.

A total of 4,060 subjects participated in the survey. Among them, eight were diagnosed with vascular dementia. Due to missing data, 3,974 participants were finally included in the analyses.

Interviews

The investigation group consisted of experts and interviewers. The evaluation panel included three neurologists and three neuropsychologists with special expertise in cognitive impairments. There were three groups of interviewers, each including one senior neurology graduate student, one junior neurologist, and one interpreter. All interviewers and experts received the same training on neuropsychological diagnosis and assessment. The inter-rater reliability for videotapes of cognitive tests and diagnostic procedures was required to exceed 0.90 [11].

Each interviewing group conducted individual semi-structured interviews with participants and their close caregivers at the residence. Interviews usually lasted 1.5 h. Detailed data on socio-demographic characteristics, medical history, lifestyle, current medications,and family history were collected. One of the inter-viewers tested the participant with a series of neuropsychological tests. First, the cognitive function was assessed with the Mini-Mental State Examination [12, 13]. Second, functional autonomy was evaluated with the Activity of Daily Living Scale [14]. Third, the differentiation between degenerative dementia and vascular dementia was assessed by the Hachinski Ischemic Scale [15]. The other interviewer took a detailed history of all cognitive impairments including time and mode of onset, possible triggers, affected domains, the course of the condition, impact on daily activities, changes in mood or behavior, results of computed tomographic scans or magnetic resonance imaging, and treatment and its effects. Last, standardized general and neurological examinations were performed.

All interviewers involved in diagnosing AD were blinded to lifestyle and genetic variables.

Diagnostic criteria of AD

All information was recorded on the questionnaire of basic information, reviewed by the panel and interviewers, and diagnoses were made each workday. If consensus was not reached, the expert would return to the residence to reevaluate the participant for a final diagnosis the following day.

Diagnostic criteria for dementia were based on the fourth edition of the DSM-IV [16]. Diagnostic criteria for AD was based on the criteria issued by the National Institute of Neurological and Communicative Disorders and Stroke-Alzheimer’s Disease and Related Disorders Association [17]. Diagnostic criteria for vascular dementia was based on the criteria of the National Institute of Neurological Disorders and Stroke-Association International pour la Recherche et l’Enseignement en Neurosciences [18]. Other types of dementia (including mixed dementia, frontotemporal dementia, dementia with Lewy bodies, Parkinson’s disease with dementia, etc.) were not analyzed in the present survey.

Dietary habits, smoking, and alcohol

Consumption of beef, mutton, dairy products, bean products, chicken, fish, and vegetables were defined as eating more than 1.25 kilograms per week. Eggs intake habits were defined as eating more than three eggs per week [19–21]. Smoking was defined as smoking ≥100 cigarettes a year. Alcohol consumption was defined as drinking more than 1 time a week for at least 1 year [22]. Kowtow was defined as performing this religious action more than three times per day. Reciting sutra, turning prayer beads, and turning a prayer wheel were defined as doing so for more than 3 hours per day.

Laboratory measurements

Venous blood samples (5 ml) were obtained after an overnight fast, stored in EDTA-K₂ tubes, and stored at –20°C. A DNA extraction kit (Promega, Madison, WI, USA) was used to extract DNA. Finally, matrix-assisted laser desorption/ionization-time of flight (MALDI-TOF) was used to test the different genotypes (Complete PLEX Gold Genotyping Reagent Ser 384; Sequenom, Inc., San Diego, CA, USA). Single-nucleotide polymorphisms (SNPs) were selected by reviewing previous studies [1, 23]. Thirteen SNPs were selected (Supplementary Table 1).

PCR amplification primers and single-base extension primers of SNP loci were designed using the Genotyping Tools of the Sequenom Company and the MassARRAY Assay Design software (Supplementary Table 1). PCR amplification was carried out using multiple PCR technology in 384-well plates. The reaction volume was 5μl. PCR reaction conditions were: 94°C for 4 min; 45 cycles of 94°C for 20 s, 56°C for 30 s, 72°C for 1 min; and final extension at 72°C for 3 min. PCR products were treated with shrimp alkaline phosphatase (SAP) to remove the free dNTPs. The single base extension reaction was carried out after SAP treatment. Conditions were: 94°C for 30 s; followed by 43 cycles at 94°C for 5 s, 52°C for 5 s, and 80°C for 5 s; and final extension at 72°C for 3 min. PCR products were purified using resin. Dried resin was added with the extension product. The plate was sealed and rotated at low speed for 30 min. The resin was centrifuged and removed.

A Mass ARRAY Nanodispenser RS1000 spotting instrument was used to spot the purified extension products on a 384-well SpectroCHIP (Sequenom, Inc., San Diego, CA, USA) chips. The chips were analyzed by MALDI-TOF. The results were analyzed using the TYPER 4.0 software (Sequenom, Inc., San Diego, CA, USA).

Statistical analysis

Validity and reliability of the questionnaire of basic information were tested using the Cronbach’s alpha. Participants were divided into the AD and non-AD groups. To identify associations, subjects with cognitive impairment were matched 1:10 with subjects without impairment. For genotypes, subjects with cognitive impairment were matched 1:1.5 with subjects without impairment. Estimates of the prevalence of AD were calculated for the overall population and for subgroups stratified by age and gender. Age-standardized and gender-standardized prevalence was calculated based on the population distribution of China in 2010 [24]. The 95% confidence intervals (CIs) were calculated for prevalence, adjusted for age and gender. Analyses were performed using SAS 9.4 (SAS Institute Inc., Cary, NC, USA). The analysis of gene polymorphisms data was performed using SHEsis [25]. Two-sided p-values <0.05 were considered significant.

Characteristics of the participants

A total of 3,974 participants completed this survey. The rate of participation was 88.37% . Tables 1–3 present the characteristics of the participants. Compared to non-AD participants, patients with AD were older (p <  0.001), fewer were married (p = 0.026), and more of them lived with their children (p = 0.015) (Table 1). Dietary habits are presented in Table 2. Compared to non-AD participants, less patients with AD performed religious activities (all p <  0.05) (Table 2). Compared to non-AD participants, more patients with AD suffered from chronic obstructive pulmonary disease (p <  0.001) and had a history of head trauma (p <  0.001). Compared to subjects with AD, subjects without AD were eating less dairy products (p = 0.026), more bean products (p <  0.001), more chicken (p <  0.001), more fish (p = 0.002), and more eggs (p = 0.002).

Prevalence of AD

The prevalence of AD among residents aged 60 years and older was 1.33% (95% CI, 0.98–1.69). AD was more prevalent in females (prevalence ratio = 1.59; 95% CI, 1.10–2.09) compared to males (prevalence ratio = 0.92; 95% CI, 0.44–1.39). Meanwhile, the age group ≥85 years old (prevalence ratio = 12.05; 95% CI, 7.10–17.00) had the highest prevalence of AD (Table 4).

Association with non-hereditary factors

Variables with p-values <0.05 in univariate analyses (Supplementary Table 2) were entered in a multivariate logistic regression analysis (Table 5). Age of 75–79, 80–84, and ≥85, eating beef, and head trauma were positively associated with AD. Kowtow, turning prayer beads, and eating chicken were negatively associated with AD.

Gene polymorphisms

Among all patients, 39 specimens in the AD group and 56 specimens in the no AD group were available for SNP analyses. Age (p = 0.306) and gender (p = 0.524) were comparable between the two groups.

Linkage disequilibrium (D’ >  0.5, r² <  0.8) was found in the loci for CLU, TP53INP1, and TFAM, indicating that haplotypes could be analyzed for these three genes. In the AD and no AD groups, the frequency of all other SNPs loci conformed to the Hardy-Weinberg equilibrium (p >  0.05), insuring the reliability of their application to evaluate larger groups.

When not considering the impact of the APOE ɛ4 allele, allele A of rs2279590 and allele C of rs2306604 were more frequent in the AD group, while allele C of rs11136000 had a lower frequency in the AD group (Table 6). The genotypes AA+GA of rs2279590 were positively associated with AD, while the genotypes GC of rs9331888 and GG+GC of rs9331888 were negatively associated with AD (Supplementary Table 3).

There was no difference in the distribution of ApoE alleles between the two groups (p = 0.876) (Supplementary Table 4). Within the ApoE ɛ4 carriers, there was no difference in any locus between the two groups (Supplementary Table 5; all p >  0.05). Within the ApoE ɛ4 non-carriers, allele A of rs2279590 and allele C of rs2306604 were more frequent in the AD group, while allele C of rs11136000 was less frequent in the AD group (Supplementary Table 6).

In the analysis of haplotypes, the CLU haplotype AG of rs2279590 and rs9331888, and the CLU haplotype TG of rs11136000 and rs9331888 were more frequent in the AD group. The CLU haplotype GG of rs2279590 and rs9331888, and the CLU haplotype CG of rs11136000 and rs9331888 had a lower frequency in the AD group.

Within the ApoE ɛ4 carriers, there were no differences in haplotypes between the two groups. Within the ApoE ɛ4 non-carriers, the CLU haplotype AG of rs2279590 and rs9331888, and the CLU haplotype TG of rs11136000 and rs9331888 showed a higher frequency in the AD group. The CLU haplotype GG of rs2279590 and rs9331888, the CLU haplotype CG of rs11136000 and rs9331888, and the TFAM haplotype GT of rs1937 and rs2306604 showed a lower frequency in the AD group (Supplementary Table 7).

Multivariate analyses for the association with non-hereditary and genetic factors

History of head trauma was not included in the model because of the too small number of subjects. The model showed that AA+GA of rs2279590 was positively associated with AD, while GG+GC of rs9331888 and kowtow were negatively associated with AD (Table 7).