In this study, we demonstrated that LTL in American Indians has a strong genetic component, with heritability estimates ranging from 51% to 62%. A genome-wide linkage scan identified significant evidence of linkage for LTL on chromosomes 13q12.11, 18q22.2 and 3p14.1 in the Oklahoma population. In addition, we observed suggestive evidence for linkage on chromosome 1q42.2 and 20q13 from Oklahoma families, and one suggestive linkage on chromosome 7q33 in the Arizona population. No significant or suggestive evidence of linkage was obtained in the Dakota pedigrees. There is no overlap between observed linkage peaks of different centers, suggesting potential genetic heterogeneity among American Indians from different geographic regions.
The strongest evidence of linkage for LTL in our genome-wide scan was localized to chromosome 13q12 in the Oklahoma population. This region has not been previously reported to harbor loci affecting interindividual variation in LTL in any ethnic groups, thus may represent a novel genetic locus influencing LTL. The one-LOD unit support interval (8.8 Mb) of this linkage signal contains over 50 annotated genes. Among these, two genes could represent promising candidate genes for LTL in American Indians. One is the well-known aging gene Klotho (KL)
, which is located ~ 10 Mb downstream from the peak 13q LOD score. This gene encodes a type-I membrane protein and functions as an ageing-suppressor gene [20
]. Overexpression of this gene extended life span in mice, and klotho
-deficient mouse (klotho
−/−) manifested a wide range of aging-related phenotypes, such as short life span, atherosclerosis, and osteoporosis [21
]. In human population studies, genetic variants in the KL
gene have been associated with longevity [24
] and several age-related disorders, such as cardiovascular disease and its associated risk factors [24
] and cognitive function [26
], all of which are consistent with its association with life span [21
]. Another possible candidate gene located in this 13q region is poly (ADP-ribose) polymerase family, member 4 (PARP4
]. The PARP
enzymes recognize DNA strand damages, and DNA binding by PARP
controls telomere length and chromosomal stability by triggering its own release from DNA ends. Telomeres are the terminal DNA structure of chromosomes and are, therefore, potential targets of PARP
. Mice lacking PARP
displayed telomere shortening and chromosomal instability, lending further support for an important role of PARP
in telomere maintenance [28
Apart from KL
, the 13q linkage peak also includes known candidate genes for inflammation, e.g., arachidonate 5-lipoxygenase-activating protein (ALOX5AP
), and cancer, e.g., breast cancer 2 early onset (BRCA2
), all of which may be involved in the aging process. In a recent GWAS meta-analysis, the gene encoding zinc finger protein 676 (ZNF676
) was related to the regulation of human telomere homeostasis [15
]. Interestingly, several genes encoding zinc finger proteins, such as zinc finger, DHHC-type containing 20 (ZDHHC20)
, zinc finger, MYM-type 2 (ZMYM2)
, and zinc finger, MYM-type 5 (ZMYM5)
, are also located within the 13q linkage region identified in our study. The possible role of these zinc finger proteins in telomere maintenance warrants further research.
We also identified evidence for linkage on chromosome 18q22.2 in the gene region of docking protein 6 (DOK6
), a member of the DOK family of intracellular adaptors that play an important role in RET (rearranged during transfection) signaling cascade [30
]. Activated RET signaling causes phosphorylation of key docking tyrosines that bind to several adaptor proteins, resulting in the activation of downstream signal transduction pathways [31
], thereby controls key cellular processes, such as cell proliferation, differentiation, and survival [32
]. Aberrant RET signaling has been associated with papillary thyroid carcinoma, multiple endocrine neoplasia types 2 syndromes, and Hirschsprung's disease [33
]. Genetic defects in genes encoding docking proteins have the potential to cause abnormal interaction with the RET signaling, which in turn may result in aging neurons and contribute to aging-related disorders such as Parkinson's disease [34
] and Alzheimer's disease [35
In a previous GWAS for LTL measured by Southern blot analysis, two SNPs (rs2162440 and rs7235755) on chromosome 18q12.2 were significantly associated with telomere length in the gene region of VPS34/PIK3C3
in Caucasians [36
]. Our linkage signal on chromosome 18q22.2 is ~28Mb downstream of this VPS gene region. Given the relatively large map distance between these two regions, it is uncertain whether these two loci belong to a same genetic locus influencing telomere variation in different populations. It is also possible that the VPS
locus in Caucasians and the DOK6
locus in American Indians may represent a long-distance cis
-regulatory element influencing telomere variation.
Another linkage peak identified in our genome-wide scan was located on chromosome 3p14.1 in the gene region of ADAMTS9
metallopeptidase with thrombospondin type 1 motif, 9). As a member of the ADAMTS
has been implicated in proteoglycan cleavage, organ shape control during development, and angiogenesis inhibition. Genetic polymorphisms in the ASAMTS9
gene have been associated with body fat distribution [37
] and Alzheimer's disease [40
]. Previous studies reported association and replication of genetic variants in the telomerase RNA component (TERC)
gene, located on 3q26, with telomere length variation [12
], but the genomic region we identified in American Indians localizes on the short arm of chromosome 3, and thus may represent a novel locus for LTL.
Except for the above-mentioned results, we also observed several loci with marginal evidence of linkage. Although these signals do not meet the genome-wide significance threshold, these genomic regions may still provide valuable information that is worthy of further investigation.
The major strength of this study includes the large, multi-generational pedigrees with well-characterized phenotypes including demographic, clinical and environmental information. The lack of overlap for linkage regions from different study centers further highlights the potential differences in genetic architecture between American Indians from diverse geographic regions. However, the genetic background of American Indians is likely to be more homogeneous than other population-based studies from urban areas.
In summary, we identified strong evidence for novel genetic loci affecting variation in leukocyte telomere length on chromosomes 13q12, 18q22.2 and 3p14.1 in American Indians who suffer from high rates of diabetes and cardiovascular disease. Several other loci with suggestive linkage were also localized. Our findings are independent of adjustments for multiple covariates, including age at enrollment, sex, center, BMI and total triglyceride, suggesting that these factors may not contribute to the observed linkage signals for telomere length. Our linkage results, coupled with plausible biological functions of the potential candidate genes related to aging, such as Klotho, PARP4, DOK6, and ASAMTS9, make these genomic regions good candidates for further investigation of causal variants influencing LTL in this minority population. Future research to fine map these candidate regions and to determine causal variants, including rare variants and structural variants, will provide valuable information on telomere biology and aging-related disorders.