CLEC3B p.S106G Mutant in a Caucasian Population of Successful Neurological Aging

Abstract

A number of efforts are underway to better understand the role of genetic variation in successful aging and longevity. However, to date, only two genes have been consistently associated with longevity in humans: APOE and FOXO3, with the APOE ɛ2 allele also protective against dementia. Recently, using an exome-wide SNP array approach, a missense variant CLEC3B c.316G>A (rs13963 p.S106G) was reported to associate with longevity in two independent cohorts of Japanese and Chinese participants. Interestingly, CLEC3B p.S106G is more frequent in Caucasian populations. Herein, we examined the frequency of CLEC3B p.S106G in a Caucasian series of 1,483 neurologically healthy individuals with a specific subset >80 years of age. Although our findings do not support an association between CLEC3B p.S106G and aging without neurological disease (p = .89), we confirmed the association between the APOE ε2 allele and better survival without neurological disease (p = .001). Further assessment of healthy aged cohorts that retain intact neurological function will be critical to understand the etiology of neurodegenerative disease and the role of age at risk.

Aging is a biological process that is accompanied by the general decline of an organism. Many theories have been proposed to explain the biology of aging, such as the accumulation of somatic mutations (1), the overactivity or hyperfunction of developmental processes resulting in cellular stress and shortening of telomeres (2,3), and the accumulation of protein aggregates due to high levels of oxidative stress (4,5). All of the proposed biological mechanisms are feasible but are not mutually exclusive, as a broad number of genes and pathways are probably involved in the aging process (6).

Longevity in humans can be defined as the survival beyond the population’s mean life expectancy (eg, 75–80 years old) and is not to be confused with healthy aging (health span), which is a combination of old age and freedom of specific disorders or desirable performance on functional tests (7). The heritability of age at death ranges from 15% to 30%, and is the greatest at the oldest ages (centenarians and supercentenarians) (8–10). Despite ongoing research efforts, to date only two genes have shown consistent association across multiple genetics of aging studies: Apolipoprotein E (APOE) and Forkhead Box O3 (FOXO3) (8,11,12). FOXO3 is a member of the forkhead family of transcription factors and probably triggers apoptosis in the absence of survival factors (13). APOE encodes for apolipoprotein E and is important for lipid and lipoprotein metabolism. APOE has three distinct common isoforms (ɛ2, ɛ3, and ɛ4) that are defined by two nonsynonymous variants (rs429358 and rs7412). The ApoE isoform ApoE4 increases the risk of developing cardiovascular disease and Alzheimer’s disease, and ApoE2 is protective against Alzheimer’s disease and is associated with healthy aging (>80 years old) (14).

A recent study by Tanisawa and colleagues identified an association between variation in the C-Type Lectin Domain Family 3 Member B (CLEC3B) gene and longevity in two independent cohorts of diverging genetic and cultural background by the use of an exome array (15). Specifically, the missense variant CLEC3B c.316G>A (p.S106G rs13963) is associated with extreme longevity (p = 1.87 × 10⁻⁸, odds ratio = 1.50) in two independent nonagenarian/centenarian cohorts of Japanese and Chinese populations, where the “G” allele is more frequent in nonagenarians/centenarians than in younger controls (15). CLEC3B, also known as TNA, encodes for tetranectin, which is found in plasma and extracellular matrix. Tetranectin is proposed to be a potential inhibitor of tumor invasion and metastasis and has been reportedly used as a prognostic biomarker for several different types of infirmities, including multiple cancer types and cardiovascular diseases. Furthermore, a mouse model with deleted tetranectin displayed motor deficits resembling Parkinson’s disease symptoms, including limb rigidity, bradykinesia, and Lewy body-like inclusions within the substantia nigra pars compacta (16).

To determine whether the CLEC3B p.S106G variant is associated with longevity and neurologically healthy aging in individuals of European descent, a total of 1,483 individuals from two independent neurologically healthy cohorts were genotyped for this variant and combined into one overall neurologically normal series, and the association between CLEC3B p.S106G and aging without neurological disease was assessed. In addition, the APOE ε4 and ε2 alleles were also examined for association with survival without neurological disease.

Methods

Population Demographics

Two independent cohorts of neurologically healthy Caucasians were included in this study, resulting in a total of 1,483 unrelated neurologically healthy individuals. The first cohort is defined as clinical healthy aging (CHA) and consists of 511 unrelated neurologically healthy Caucasian individuals who are at least 80 years old (median = 88 years; maximum = 101 years), have one or more siblings who are at least 85 years old, have no family history of dementia and have intact neuronal health (no history of Parkinson’s disease, stroke, uncontrolled epilepsy, multiple sclerosis, head injury with residual brain damage, brain tumor, and brain surgery) (17). Additional information about the methods for sample recruitment, inclusion and exclusion criteria, and participant evaluation for the CHA cohort can be found in the Supplementary Material. The other neurologically healthy cohort consisted of 972 unrelated Caucasian samples collected at the Mayo Clinic Jacksonville (MCJ). The controls were recruited from 1998 to 2015 through Mayo Clinic as a part of the Udall Parkinson’s Disease Research Center of Excellence. The ages of the MCJ controls upon sample collection ranged from 18 to 92 years old, with a median of 67. All MCJ controls are free of neurodegeneration and family history of mental/cognitive impairments upon sample collection.

To maximize power to detect an association between the genetic variants examined and age without neurological disease and also because the distribution of age overlapped between the CHA and MCJ control cohorts, rather than compare the frequency of genetic variants between these two cohorts, we instead combined the two neurologically normal cohorts into an overall series of 1,483 neurologically normal controls (Table 1). This combined series was utilized in all analyses. As the probability of living longer without neurological disease increases as age without neurological disease increases, age at blood draw is a measure of survival without neurological disease and therefore was the primary focus of the study.

Genotyping

The CHA and the MCJ cohorts were genotyped for the CLEC3B p.S106G (rs13963), APOE p.C130R, and p.R176C (rs429358, rs7412) using ABI TaqMan allelic discrimination assays on an Applied Biosystems QuantStudio 7 flex Real-Time PCR System (Thermo Fisher Scientific, Waltham, MA) using standard protocols. The genotype data were analyzed using QuantStudio Real-time PCR software (Thermo Fisher Scientific, Waltham, MA). The two APOE variants (rs429358, rs7412) determine the APOE genotypes (ε2/ε2, ε2/ε3, ε2/ε4, ε3/ε3, ε3/ε4, or ε4/ε4) that are used to infer the three common APOE isoforms (E2, E3, and E4). CLEC3B and APOE genotypes are displayed in Table 1. Observed variant frequencies did not differ noticeably from those expected under Hardy–Weinberg equilibrium (Supplementary Table 2).

Statistical Analysis

To evaluate whether the minor allele of the CLEC3B rs13963 variant is more (or less) common as age at blood draw increases, we used proportional odds logistic regression models, where CLEC3B rs13963 (the dependent variable) was examined as an ordered categorical variable under an additive model (ie, 0, 1, or 2 copies of the minor allele), and age (the independent variable) was analyzed as a continuous variable. The primary analysis involved all controls, and the model was adjusted for sex, presence of APOE ε4, and presence of APOE ε2. Odds ratios and 95% confidence intervals were estimated and are interpreted as the multiplicative increase in the odds of a greater number of minor alleles of CLEC3B rs13963 corresponding to each 10-year increase in age at blood draw. Similarly, to examine whether the APOE ε4 or ε2 alleles systematically increase or decrease in frequency as age at blood draw increases, binary logistic regression models were used, where APOE ε4 and ε2 (the dependent variables) were both considered under dominant models (ie, presence of absence of ε4; presence of absence of ε2) owing to the small number of ε4/ε4 and ε2/ε2 participants, age (the independent variable) was again examined as a continuous variable, and models were adjusted for sex. Odds ratios and 95% confidence intervals were estimated and are interpreted as the multiplicative increase in the odds of presence of the APOE ε4 allele (for analysis of APOE ε4) or presence of the APOE ε2 allele (for analysis of APOE ε2) corresponding to each 10-year increase in age at blood draw.

In secondary analysis, we examined the association between CLEC3B rs13963 and age at blood draw in the separate subgroups of males, females, carriers of APOE ε4, noncarriers of APOE ε4, carriers of APOE ε2, and noncarriers of APOE ε2; proportional odds logistic regression models were adjusted for presence of APOE ε4 and ε2 in analysis of males and females and for sex in analysis of carriers and noncarriers of APOE ε4 and ε2. Similarly, we also assessed the associations of APOE ε4 and ε2 with age at blood draw separately in males and females. All statistical tests were two sided, and p-values less than .05 were considered as statistically significant. Statistical analyses were performed using R Statistical Software (version 3.2.3; R Foundation for Statistical Computing, Vienna, Austria). The R vglm function was used for proportional odds logistic regression analysis, whereas the R glm function was used for binary logistic regression analysis.

Results

The association between CLEC3B rs13963 and age is given in Table 2, where the proportion of carriers of the minor allele of rs13963 is shown for six different age categories (≤50, 51–60, 61–70, 71–80, 81–90, and >90 years). These age categories were utilized for purposes of presentation only; age was analyzed as a continuous variable in all statistical tests of association. In the overall sample of 1,483 controls, there is no evidence of association between occurrence of the minor allele of CLEC3B rs13963 and age (p = .89); the proportion of carriers of the minor allele is highest in the age ≤ 50 years (59.4%) and age > 90 years (62.8%) groups. This lack of association is consistent in the specific subgroups of males, females, carriers of APOE ε4, noncarriers of APOE ε4, carriers of APOE ε2, and noncarriers of APOE ε2 (all p ≥ .26). The lack of difference in age according to CLEC3B rs13963 genotype is further illustrated in Figure 1. As the “G” allele of CLEC3B rs13963, that is significantly more common in nonagenarians/centenarians compared with younger controls in the study by Tanisawa and colleagues (15), is the major allele of this variant in our study, it is important to clarify that examining the association between age and number of minor (“A”) alleles of CLEC3B rs13963, as we have done, is statistically equivalent in terms of p-value to assessing the association between age and number of major (“G”) alleles.

We did observe a significant association between occurrence of the APOE ε2 allele and increased age (p = .001, Table 3); the ε2 allele increased in frequency from 14.3% in the age ≤ 50 years group to 28.5% in the age > 90 group (Figure 2). The higher frequency of the ε2 that was observed in the older participants was consistent for both males and females, though this finding is slightly more pronounced in the latter group (Table 3). On the other hand, there is no evidence of an association between APOE ε4 and survival without neurological disease in the overall cohort or when examining males and females separately (all p ≥ .78, Table 3). Of note, when excluding the 49 participants with an APOE ε2/ε4 genotype, the lack of association between age and presence of ε4 remains consistent (p = .49), as does the significant association between ε2 and older age (p = .009).

We performed two additional sensitivity analyses examining associations with age for CLEC3B rs13963, APOE ε4, and APOE ε2 in the overall series of 1,483 controls. First, rather than considering the given variant as the dependent variable and age as the independent variable as we had previously done, we assessed these associations using linear regression analysis where age was the dependent variable and the given variant was the independent variable (Supplementary Table 3). Consistent with our original results, no significant association with age is observed for CLEC3B rs13963 (p = .79) or APOE ε4 (p = .92), whereas APOE ε2 was significantly associated with older age (p = .001). Second, we utilized a case–control design where participants of age > 90 (N = 145) were classified as cases, participants of age ≤ 70 (N = 602) were classified as controls, and the remaining participants between ages 70 and 90 were excluded (Supplementary Table 4). Again, consistent with our previous findings, in logistic regression analysis, there is not a significant association with case–control status for CLEC3B rs13963 (p = .97) or APOE ε4 (p = .80), whereas APOE ε2 is associated with an increased likelihood of being a case with age > 90 years (p = .002).

Discussion

The evidence of a genetic component to longevity and healthy aging comes from the observation of families with multiple long-lived members across generations. Longevity is a complex and polygenic trait with an estimated heritability that ranges from 15% to 30% (9,10). Attempts have been made to identify genetic variants associated with successful aging to provide new insights into the physiological processes that determine the increased risk for age-related diseases such as neurodegenerative, cardiovascular, cancers, and arthritis. Analysis of families, founder populations, and isolates has been the main focus in genetic studies of longevity and healthy aging; however, several larger population association studies have been performed without successful replication (7).

The present study attempted to replicate a genetic study of longevity that found the association of extreme longevity with a missense mutation (CLEC3B p.S106G) in two independent cohorts of nonagenarians/centenarians of Japanese and Chinese descent (15). The alternative allele frequency of the CLEC3B p.S106G variant “A” is much higher in Asian populations (86% in East Asians, 65% in South Asians) in comparison to Caucasians (40% in Europeans). Though it is important to highlight that the longevity-related outcome examined in our study (survival without neurological disease) differs from that utilized in the aforementioned study by Tanisawa and colleagues (extreme longevity) (15). The association between the CLEC3B p.S106G variant and longevity in a cohort of neurologically healthy aged Caucasians could not be replicated in this study; this finding is also supported by the lack of association for this locus in a European longevity GWAS (11). Specifically, there are no indications that the presence of the alternative “A” allele significantly increases the risk of death or occurrence of neurological disease, as the minor allele of CLEC3B p.S106G was observed at a relatively consistent frequency across the different age groups that were examined. In a secondary analysis, the data were separated by sex, which is a known influencer of the aging process, and the APOE isoforms ɛ2 and ɛ4, which have been associated with protection and increased risk of Alzheimer’s disease, respectively. No link was found between the tested CLEC3B genotype and neurologically healthy aging in the different subgroups of sex and APOE isoforms. These findings suggest that the role of CLEC3B p.S106G in longevity may be specific to Asian populations, although a recent longevity GWAS in the Han Chinese population also failed to nominate CLEC3B (18). Interestingly, tetranectin levels have been suggested to be a biomarker for neuronal and vascular diseases such as Parkinson’s disease (16,19–21), epilepsy (22), multiple sclerosis (23), and coronary artery disease (24).

The different ApoE isoforms regulate the aggregation of amyloid-β in synapses at distinct rates, either increasing or decreasing the risk of neuronal death. Individuals that carry the ɛ4 ApoE isoform are at an increased risk of Alzheimer’s disease compared with those carrying the more common ɛ3 allele, whereas the ɛ2 allele in general has a decreased neurodegeneration risk (25). Even though no association between the ɛ4 isoform and younger age was found in the studied Caucasian cohort (p = .92), the ɛ2 allele was significantly associated with older age at study (p = .001), which is consistent with previous studies (8,11,14). The reason for no statistical difference in ɛ4 carriers between younger and older ages is unclear; however, it may suggest that there are other complex interactions, either genetic or environmental, that are protective to the older and neurologically healthy ɛ4 carriers. It may also be the case that having selected a series of neurologically healthy participants without a family history of disease we have specifically selected against ɛ4 carriers in the younger participants and the association is confounded.

Several limitations of our study should be acknowledged. First, our study contained far fewer individuals of age 95 or older than did the original study by Tanisawa and colleagues (15), and this could have contributed to the lack of association between CLEC3B p.S106G and survival without neurological disease that we observed. Second, though the participants included in our study were free of neurological disease, lack of presence of other comorbidities was not an inclusion criteria; though our findings do not indicate that CLEC3B p.S106G is associated with longevity occurring without neurological disease, we cannot rule out the possibility that this variant may be associated with general healthy aging in a Caucasian population. Finally, without available genome-wide population control markers, population stratification could potentially have affected our results.

Longevity and healthy aging genetic associations, other than APOE and FOXO3, have been so far limited to individual studies and could not be replicated across independent cohorts (7). The lack of consistency is probably explained by the multifactorial nature of longevity and healthy aging, and adding to that, the genetic heterogeneity, small sample sizes and the combination of data from populations of different lifestyles and genetic backgrounds. The absence of clear genetic associations and study replication failures could also be attributed to the “dilemma” that older individuals do carry the same risk alleles as the general population; however, they may also carry a series of rare protective alleles that are misrepresented in population studies. Furthermore, because most of the association peaks contain variants and haplotypes located in noncoding/intergenic regions, functional and regulatory studies of such polymorphisms and their likely effect in candidate genes or pathways could provide valuable insights that are lacking in the current genomic association studies. For example, an intronic variant that may disrupt or strengthen protein binding motifs that may significantly affect gene expression levels and efficiency of pathways that lead to longer and healthier life spans. Future studies with larger data sets of better-characterized cohorts and next-generation sequencing approaches will be needed to resolve the genetic architecture of longevity and healthy aging.

Funding

O.A.R. is supported by the National Institutes of Health (NIH; R01 NS78086; U54 NS100693), the U.S. Department of Defense (W81XWH-17-1-0249), and the Little Family Foundation. This study was supported in part by the Mayo Clinic Center for Individualized Medicine and The David Eisenberg Professor at Mayo Clinic.

Acknowledgments

We thank all those who have contributed to our research, particularly the patients and families who donated blood samples for this work. The Mayo Clinic is an American Parkinson Disease Association (APDA) Information and Referral Center, an APDA Center for Advanced Research, Lewy Body Dementia Association (LBDA) Research Center of Excellence, is supported by the Mangurian Foundation for Lewy body research and was a Morris K. Udall Parkinson’s Disease Research Center of Excellence (NINDS P50 #NS072187).

Conflict of Interest

None reported.

References

Author notes