Anatoliy I. Yashin

Hemoglobin A1c (HbA1c), an important diagnostic biomarker for type 2 diabetes (T2D), is also associated with aging, cognitive performance, and mortality. To identify epistatic interactions, we assessed 133 known gene variants associated with HbA1c among 3778 non-diabetic subjects of European ancestry in the Long Life Family Study (LLFS). We applied Bayesian imputation based association mapping (BIMBAM) to identify significant pairwise epistatic interactions among 133 genetic variants. To take into account confounding effects, we adjusted age, sex, field centers, body mass index (BMI), and genetic principal components (PCs). This analysis yielded seven pairs with log10(BF) > 10; of those, six pairs were confirmed using a full-term mixed regression model. Specifically, these included significant interactions of HK1-rs17476364 with variants in GCK (rs2971670, rs4607517) or G6PC2 (rs560887), as well as between HK1-rs16926246 and the same variants (p values for each term ≤ 7.14 × 10⁻³). All epistatic interactions between HK1 and GCK, and between HK1 and G6PC2 were replicated at p < 0.05 in at least one of two large independent studies (namely, Framingham Offspring Study; Health and Retirement Study). The present study revealed that HK1 and GCK interact to contribute to regulating levels of HbA1c and are likely to be involved in molecular mechanisms underlying healthy aging processes.

Background: APOE4 is the strongest genetic risk factor for late‐onset Alzheimer's disease (AD), with a higher effect in females than in males. Identifying factors that may reduce the detrimental impact of APOE4 on AD is an important research objective. Accumulating evidence points to a key role of the innate and adaptive immune system in AD, including natural killer (NK) and cytotoxic CD8+ T cells infiltrating the Central Nervous System (CNS) in AD, contributing to neuroinflammation and neuronal death. Here, we investigated how polymorphism of the HLA‐C (Human Leucocyte Antigen‐C) gene, a major regulator of NK and CD8+ T cells, may influence AD risk in the presence of the APOE4 allele in males and females. Method: We assessed the association of the HLA‐C genetic variation in study participants of the UK Biobanks (UKB), who were double carriers of the APOE4 allele, with AD reported in their mothers and fathers who carried at least one APOE4 allele (1562 and 798 in the case and 7572 and 8336 in the control groups, respectively). The analysis used a logistic regression model in the REGENIE and was repeated using the PLINK2 software package. The PLINK2 clumping procedure was used to reduce the number of statistical tests. Result: The SNPs rs2074488 and rs2074491 in the HLA‐C region, genotyped in UKB participants, were significantly associated with lower chances of AD in mothers who carried at least one copy of the APOE4. Three SNPs in the HLA‐C region (+/‐2KB): rs13203722, rs13218306, and rs12207404 showed deleterious associations with AD in mothers. Both analyses (using PLINK2 and REGENIE) supported these results. The protective SNPs and the deleterious ones were not in linkage disequilibrium. A similar analysis did not produce significant associations of HLA‐C genetic variants with AD in fathers. Conclusion: The detected significant associations of HLA ‐ C genetic variants in offspring with AD odds in mothers carrying at least one APOE4 allele suggest that HLA ‐ C may modulate (attenuate or aggravate) the impact of APOE4, the strongest genetic risk factor, on AD occurrence. This finding further supports the role of infections and other factors initiating immune response in AD.

Background: Infections and neuroinflammation are increasingly recognized as key contributors to the pathogenesis of Alzheimer’s disease (AD). Emerging evidence suggests that the immune‐regulating genes GDF15 and MIR146A are promising targets for developing treatments aimed at reducing the risks of age‐related diseases associated with chronic inflammation, including AD. The shared involvement of these genes in inflammation and immune modulation suggests potential overlap or interaction between respective regulatory networks. Here we hypothesize that interactions between polymorphisms in the GDF15 and MIR146A genes may confer protection against AD. Method: We analyzed the Framingham Heart Study (FHS) data using logistic regression model with an interaction term in REGENIE software package. AD outcome was defined as a binary variable: group 1 (cases) included participants who survived past age 85 without AD diagnosis ( N =1,421); group 0 (controls) included participants diagnosed with AD at any age ( N =257). To account for linkage disequilibrium (LD) and optimize multiple testing correction, we used the PLINK2 clumping procedure, which results in a higher threshold for Bonferroni correction (BCT). Result: The interaction between SNPs rs1059369 ( GDF15 ) and rs2910163 (upstream MIR146A ) was associated with significantly reduced odds of AD (β = 0.57, p ‐value = 1.16E‐02, BCT = 2.5E‐02). Similar interaction effect was seen for rs4808795 ( GDF15 ) and rs2910163 ( MIR146A ) (β = 0.57, p = 1.15E‐02, BCT = 2.5E‐02). The rs1059369 (missense variant) is in strong LD with the intergenic rs4808795 (D’=1, R ² =0.99), suggesting that the latter may represent the effects of rs1059369 rather than having its own functional significance. SNP rs2910163 (upstream MIR146A ) and rs2910164 (MIR146A) are also in strong LD: (D’=1, R ² =0.77). The associations of SNPs rs1059369 and rs2910164 have also been linked to immune and aging‐related traits in separate analyses and can be found in respective publications. Conclusion: The results of this study support the importance of interactions that microRNAs and GDF15 may play in AD risk. They also suggest that detected associations of interactions of SNP rs2910163 related to MIR146A with rs1059369 and rs4808795 SNPs related to the GDF15 gene may be considered when developing interventions aiming to protect against AD.

Background: Accumulating evidence points to a major role of compromised immunity in Alzheimer’s Disease (AD). The MIR146a and COMT genes are both known for their involvement in immune responses, as well as in AD‐related traits. COMT also regulates the metabolism of catecholamines, which is linked to inflammation and trained immunity. We hypothesized that the interplay between MIR146a and COMT genes may influence AD risk. Method: To test this hypothesis, the logistic regression model with the interaction term was applied to the analysis of the Cardiovascular Health Study (CHS) data, males and females combined. A binary AD trait was created: bn_AD_2 =1 (case) included participants who survived age 85 and did not have AD before or after this age ( N =1,421); bn_AD_2 =0 (control) included participants who acquired AD at any age ( N =257). Framingham Heart Study (FHS) data were used for replication. The case and control groups there included 903 and 354 participants, respectively. Result: The analysis of CHS data found that the interaction between SNPs rs2910163 in MIR146a and rs4818 in COMT genes is significantly associated with AD occurrence (β=‐0.48, p =1.62E‐02, Bonferroni=1.67E‐02). This finding was confirmed in FHS data, for the same SNPs (β=‐0.41, p =3.63E‐02). Notably, rs2910163 is in strong LD with rs2910164, and both rs2910164 and rs4818 have been associated with multiple health‐related traits in other studies. Conclusion: Our study findings emphasize the importance of factors regulating immune responses in AD development and support the role of GxG interactions in AD risk.

Background: Results of recent analyses indicate that axon demyelination may play an important role in AD pathology. The MBP gene encodes a myelin basic protein involved in axon myelination in the nervous system including the central nervous system. Polymorphisms in this gene, as well as variations in expression, have been associated with multiple sclerosis (MS). Evidence of the involvement of this gene in Alzheimer’s disease (AD) is emerging. The EIF2B5 gene codes for a subunit of eukaryotic translation initiation factor 2B (eIF2B), a master regulator of translation. Recent studies connected EIF2B5 and its polymorphisms with various health disorders including MS and AD, as well as hypomyelination. The latter was also associated with reduced MBP. Considering this, and that both MBP and EIF2B5 have been associated with AD, we hypothesized that the interplay between these two genes might also be associated with AD.Methods: To test this hypothesis, we analyzed data on genotyped individuals (males and females combined) from the Health and Retirement Study using a logistic regression model with the interaction term. Education, smoking status, sex, and the first five principal components were included as observed covariates. The binary AD trait was defined as “0” if AD onset was observed at any age, and as “1” if a person survived to age 85 and did not get AD during their life. Linkage disequilibrium (LD) test and SNP‐clumping procedures were used to reduce the number of tests in SNPxSNP interaction analysis and make Bonferroni correction less conservative.Results: We found a statistically significant positive association of the interaction between intronic rs12959006 in MBP gene and missense variant rs843358 in the EIF2B5 gene with AD. The rs12959006 was associated with MS and rs843358 in survival from cancer in the literature. Several other SNPs that are in high LD with the SNPs found in this study have also been separately associated with health traits in the literature.Conclusion: The results of this study support the hypothesis that the interaction between MBP and EIF2B5 variants is significantly associated with AD. Our finding implies that regulation of translation and axon demyelination jointly contribute to AD.

Semantic fluency impairment has been attributed to a wide range of neurocognitive and psychiatric conditions, especially in the older population. Moderate heritability estimates on semantic fluency were obtained from both twin and family-based studies suggesting genetic contributions to the observed variation across individuals. Currently, effort in identifying the genetic variants underlying the heritability estimates for this complex trait remains scarce. Using the semantic fluency scale and genome-wide SNP genotype data from the Long Life Family Study (LLFS), we performed a genome-wide association study (GWAS) and epistasis network analysis on semantic fluency in 2289 individuals aged over 60 years from the American LLFS cohorts and replicated the findings in 1129 individuals aged over 50 years from the Danish LLFS cohort. In the GWAS, two SNPs with genome-wide significance (rs3749683, p = 2.52 × 10⁻⁸; rs880179, p = 4.83 × 10⁻⁸) mapped to the CMYAS gene on chromosome 5 were detected. The epistasis network analysis identified five modules as significant (4.16 × 10⁻⁵ < p < 7.35 × 10⁻³), of which two were replicated (p < 3.10 × 10⁻³). These two modules revealed significant enrichment of tissue-specific gene expression in brain tissues and high enrichment of GWAS catalog traits, e.g., obesity-related traits, blood pressure, chronotype, sleep duration, and brain structure, that have been reported to associate with verbal performance in epidemiological studies. Our results suggest high tissue specificity of genetic regulation of gene expression in brain tissues with epistatic SNP networks functioning jointly in modifying individual verbal ability and cognitive performance.

Background: Several mechanisms have been hypothesized for the pathogenesis of Alzheimer’s disease and related dementias (ADRD); however, we are still far from effective prevention and treatment of the disease, thus more basic research is needed to understand the etiological neuropathogenesis. We hypothesized that a part of increased ADRD‐risk could be due to a complex interplay of genes and daily‐life psychosocial stress.Method: We used the Health and Retirement Study (HRS) surveys and Medicare‐linked data for the discovery analysis and the Long‐Life Family Study (LLFS) dataset for replication analysis, creating a synthetic stress measure using the algorithm based on stress proxy‐indicators, as described in Singh et al., 2015. The two datasets were compared by evaluating samples’ stress levels and socioeconomic status. We performed the HRS gene‐by‐stress discovery GWAS analyses on age‐at‐onset of ADRD using linear regression with SNP, STRESS, SEX, and ancestry principal components as other terms in model. We fitted the same regression model in the LLFS replication analysis, adjusting it additionally with field centers and Huber–White familial clustering.Result: The HRS GWASs resulted in associations at p‐value of order 10⁻⁸ in both Blacks and Whites. Although most of the top 25 significant SNPs were mapped to non‐coding regions, we identified four characterized genes (BCR, MCC, KCNQ5, and DCP2) in Whites and three genes (EMCN, DPYD, and INSYN2A) in Blacks. Published literature from human and experimental studies suggested that five out of the eight genes were known for their association with neurodevelopmental disorders including Autism, intellectual disability, but also for cancers including colorectal carcinoma, gastric carcinoma, and leukemia. Although LLFS samples presented higher socioeconomic status and lower stress levels compared to HRS samples, the associations of one‐third tested SNPs from HRS Whites replicated in LLFS samples.Conclusion: Our findings suggest that daily‐life stress may contribute to the risk of ADRD via complex interactions with genes. The associations with non‐coding genetic variants may indicate their influence through gene‐regulatory roles. In addition, the associations observed with cancer genes may possibly point towards the involvement of a common mechanistic feature, e.g., angiogenesis, which needs to be investigated further.

Background: Accumulating evidence suggests that infections can play a major role in Alzheimer’s disease (AD), however, mechanism is unclear as multiple pathways may be involved. One possibility is that infections might contribute to AD directly by promoting neuronal death and neurodegeneration.Method: We investigated the relationship between history of infectious diseases (a potential AD risk factor) and brain hippocampal volume, HV (a biomarker of neurodegeneration and AD neuropathology) in a subsample of the UK Biobank participants with respective information. Infectious disease diagnoses were based on ICD‐10 codes. The left/right HV was measured by the MRI. Analysis of variance (ANOVA) and Welch Two Sample t‐test were used to examine the statistical significance.Result: The HV was significantly lower in older women (age 65+) with prior history of infections, compared to the same age women without such history (p‐value = 0.014 and 0.029, for the left and right HV, respectively). The effect size increased with age, and this increase was faster for the right vs. left HV. The difference in HV between male subjects with and without history of infections was not statistically significant.Conclusion: In this study, prior infections were significantly associated with the lower left/right HV in women aged 65+, which supports the role of infectious diseases in neurodegeneration and preclinical AD pathology in females. The effect size increased with age, in line with the increase in brain vulnerability to stressors due to aging. The faster increase for the right vs. left HV might indicate that females’ verbal memory declines faster with age compared to visual‐spatial memory, following the exposure to infectious disease, which deserves further investigation. The observed sex differences may reflect a higher vulnerability of the female brain to infection‐related factors, which in turn might contribute to the higher risk of AD in women compared to men.

We performed a genome-wide association study (GWAS) of human extreme longevity (EL), defined as surviving past the 99th survival percentile, by aggregating data from four centenarian studies. The combined data included 2304 EL cases and 5879 controls. The analysis identified a locus in CDKN2B-AS1 (rs6475609, p = 7.13 × 10⁻⁸) that almost reached genome-wide significance and four additional loci that were suggestively significant. Among these, a novel rare variant (rs145265196) on chromosome 11 had much higher longevity allele frequencies in cases of Ashkenazi Jewish and Southern Italian ancestry compared to cases of other European ancestries. We also correlated EL-associated SNPs with serum proteins to link our findings to potential biological mechanisms that may be related to EL and are under genetic regulation. The findings from the proteomic analyses suggested that longevity-promoting alleles of significant genetic variants either provided EL cases with more youthful molecular profiles compared to controls or provided some form of protection from other illnesses, such as Alzheimer’s disease, and disease progressions.