Frailty and mortality are not influenced by mitochondrial DNA haplotypes in the very old☆

Inherited genetic variation of mitochondrial DNA (mtDNA) could account for the missing heritability of human longevity and healthy aging. Here, we show no robust association between common genetic variants of mtDNA and frailty (an “unhealthy aging” phenotype) or mortality in 700, more than 85-year-old, participants of the Newcastle 85+ study. Conflicting data from different populations underscore our conclusion that there is currently no compelling link between inherited mtDNA variants and aging.


Introduction
Human longevity shows heritability of w25%, but large-scale, nuclear, genome-wide association studies have not yet clearly established all the responsible genes (Beekman et al., 2013;Kirkwood et al., 2011). In addition to nuclear genes, there may also be contributions from the maternally inherited, extranuclear, mitochondrial genome (mitochondrial DNA [mtDNA]). mtDNA codes for 13 respiratory chain proteins that are essential for the production of adenosine triphosphate, which is required for all active intracellular processes. There is an emerging evidence that mitochondrial dysfunction plays a key role in cellular aging, and the accumulation of somatic mutations of mtDNA is associated with an aging phenotype in mice and humans (Schon et al., 2012). It is, therefore, of great interest that different polymorphic variants of mtDNA appear to be enriched in cohorts of healthy older individuals, raising the possibility that our mitochondrial genome may influence how we age and how long we live. However, although several studies support this hypothesis (Niemi et al., 2003(Niemi et al., , 2005Yang et al., 2012), there are conflicting findings and a lack of consistency (Courtenay et al., 2012;De Benedictis et al., 1999;Feng et al., 2011;Finnila et al., 2000;Ivanova et al., 1998;Ross et al., 2001;Tanaka et al., 1998Tanaka et al., , 2000. Moreover, there have been no studies of mtDNA and healthy aging phenotypes.
In an attempt to resolve these issues, we carried out a comprehensive study of mtDNA haplogroups in the Newcastle 85þ study. Ten mtDNA haplogroup markers effectively tag the most common subgroups of mtDNA found in 95% of the local population. The Newcastle 85þ study provides a unique opportunity to study the effects of these haplogroups both on survival up to and beyond age 85 and on frailty, an "unhealthy aging" phenotype, in a representative population-based cohort of the very old.

Study cohort
The Newcastle 85þ study has been reported and includes a sociodemographically representative 1921 birth cohort recruited at age w85 through general practice patient lists (n ¼ 845) (Collerton et al., 2009). An assessment of frailty was performed at baseline using 2 robust and validated measures: the Rockwood frailty index (RFI) (Rockwood and Mitnitski, 2007) and the Fried frailty status (FFS) (Fried et al., 2009), as described (Collerton et al., 2012). RFI was available for 811 (96.0%) of the cohort and FFS for 552 (65.3%). The cohort has been followed for mortality from baseline assessment (June 2006 to Sept 2007) until April 30, 2012.

Molecular genetic analysis
mtDNA haplogroups were determined using a stepwise algorithm (Torroni et al., 1996) by primer extension of multiplex polymerase chain reaction products with the detection of the allelespecific extension products by matrix-associated laser desorption/ ionization time of flight (Sequenom MassARRAY, San Diego, CA, USA). DNA was available for 752 participants; 52 cases were excluded from the analysis because of either heteroplasmic status (n ¼ 5) or inability to detect haplogroup/low-quality DNA (n ¼ 47), leaving 700 with valid mitochondrial haplogroup data.

Statistical analysis
The frequency of mtDNA haplogroups in the incident Newcastle 85þ cohort (n ¼ 700) was compared with 3 ethnically matched population control data sets representing different ages using chi-squared analysis with pairwise comparisons of samples for each haplogroup: (1) a local birth cohort (n ¼ 344, neonatal cord blood samples), North Cumbria Community Genetics Project (Elliott et al., 2008); (2) a national mid-age cohort, the 1958 Medical Research Council cohort (n ¼ 2889, 52% male), which has previously been shown to be representative of control subjects in our region (Chinnery et al., 2007); and (3) a local cohort of healthy older subjects (n ¼ 93, 35% male; mean age 69, standard deviation ¼ 8.5).
Mitochondrial haplogroup data were available for 85.8% (696/ 811) of those with RFI available and 91.4% (477/552) of those with FFS available. These were the samples used in the principal analyses. Linear regression was used to determine the relationship between mtDNA haplogroups and RFI, and ordinal logistic regression was used to determine the relationship between mtDNA haplogroups and FFS, both before and after controlling for sex, years of education, and smoking status. A count of chronic diseases was used as an additional control for the FFS. The relationship between mtDNA haplogroups and survival was determined by Cox proportional hazards analysis, both before and after controlling for sex, total cholesterol, body mass index, hypertension, diabetes, ethnicity, and smoking status. The median follow-up period was 58 months during which 336 deaths occurred.

Results
The overall frequency distribution of mtDNA haplogroups in the incident Newcastle 85þ cohort was compared with the birth cohort (p ¼ 344), mid-age cohort (p ¼ 2889), and old-age cohort (p ¼ 93) ( Table 1). The only significant difference in distributions was for the local older cohort that differed from each of the other cohorts in the "other" haplogroup category only.
There was no significant association between mtDNA haplogroups and frailty (RFI or FFS) before and after controlling for the potential confounding variables (Tables 2 and 3). Although we observed an association between haplogroup X and increased mortality (p ¼ 0.025), this was not apparent after controlling for total cholesterol, body mass index, hypertension, diabetes, ethnicity, and smoking status (Table 4). Likewise, the trend toward reduced mortality associated with haplogroup K (p ¼ 0.041), which remained after controlling for other variables (p ¼ 0.023), did not withstand correction for the multiple haplogroups under investigation.

Discussion
We found no robust evidence of an association between mtDNA haplogroups and either frailty or survival beyond age 85 or any   Our findings do not support a role for mtDNA in promoting healthy aging or longevity. The absence of an age-associated stratification of mtDNA haplogroups is in agreement with previous findings in large European cohorts (Benn et al., 2008;Dato et al., 2004). On the other hand, our findings contrast with the results of several smaller European (De Benedictis et al., 1999;Dominguez-Garrido et al., 2009;Ivanova et al., 1998;Niemi et al., 2003;Ross et al., 2001) and Far-Eastern studies (Feng et al., 2011;Zhang et al., 2003).
Although it is conceivable that these geographic differences reflect different environmental constraints, or ethnic differences in the nuclear genetic background, it is perhaps more likely that the relatively small size of these study groups led to false-positive associations because mtDNA haplogroup studies are particularly sensitive to population stratification. In keeping with this, none of the previously reported positive findings have been replicated directly, and it is not clear why a particular haplogroup would be associated with longevity in one context and not the other. Likewise, we were unable to replicate previous findings of a genderspecific association with aging. If present, such an association would be difficult to explain mechanistically. Although we cannot exclude the possibility that a larger study cohort would reveal an association between mtDNA and longevity and/or healthy aging phenotypes, our findings of a lack of an association with 2 sensitive and reliable measures of frailty and no direct evidence of an effect on survival suggest that any contribution from mtDNA would be modest at best. Our results, therefore, turn the spotlight away from mtDNA back to the nuclear genome, in the search for genes predisposing to longevity and healthy aging.

Disclosure statement
The authors report no conflicts of interest. The data contained in the manuscript being submitted have not been previously published, have not been submitted elsewhere, and will not be submitted elsewhere while under consideration at Neurobiology of Aging. All authors have reviewed the contents of the manuscript being submitted, approved of its contents, and validated the accuracy of the data.
Ethical approval for the study is in place.  Bold text indicates p < 0.05. a Seven binary variables were created of "in haplogroup H" versus "not in haplogroup H" type, and 7 models were run entering each binary variable separately. Model 1 is unadjusted and model 2 is adjusted for sex, ethnicity, total cholesterol, body mass index, hypertension, diabetes, and smoking.
Newcastle Biomedical Research Centre based at Newcastle upon Tyne Hospitals NHS Foundation Trust and Newcastle University. The views expressed are those of the author(s) and not necessarily those of the NHS, the NIHR, or the Department of Health. The funders had no role in the study design; in the collection, analysis, and interpretation of data; in the writing of the paper; and in the decision to submit the paper for publication. Thanks are especially because of the 85-year olds of Newcastle and North Tyneside for the generous donation of their time and personal information to make the study possible. In addition, the authors thank the research nurses, biomarker technicians, data manager, project secretary, and NHS North of Tyne (Newcastle Primary Care Trust) and local general practices.