FOXO3A is one of the very few genes where single nucleotide polymorphism variants are fairly reliably shown to correlate with statistical variations in human aging across populations. Generally genetic correlations with aging do not replicate between study groups, which implies that the effects of genetic variations on aging are individually very small and overall highly varied and complicated. As these researchers show, even in the case of FOXO3A variants it is hard to establish associations reliably:
In this study, we investigated the association of variation in the second consistently confirmed longevity-associated gene, FOXO3A, with aging-related phenotypes in the oldest-old. The Discovery sample was 1200 participants randomly drawn from 1651 eligible members of the Danish 1905 Birth Cohort Study. For replication of the findings observed in the Discovery sample, we used data on 1279 individuals from two additional population-based surveys of oldest-old Danes: the Danish 1910 and 1915 Birth Cohort Studies. We explored four phenotypes known to predict their survival, that is, cognitive function, self-reported health, hand grip strength, and activity of daily living (ADL). Moreover, we included data on five self-reported diseases: diabetes, cancer, cardiovascular disease, osteoporosis, and bone fracture, as these are either investigated in genetic association studies and/or are supported by foxo animal models.
We found FOXO3A variation to show nominal significant association with two of the investigated phenotypes; ADL and bone fracture. This does, however, not exclude relevance of FOXO3A variation for the remaining seven phenotypes or for the physiological processes behind these phenotypes; it is possible that more statistical power is needed to detect such associations, especially if these are of small effect size. Furthermore, as the foxo3a protein has a wide array of downstream targets, which themselves affect a wide range of cellular and physiological processes, it may simply be difficult to pinpoint the candidate phenotypes, which FOXO3A potentially affects. The association between FOXO3A variation and bone fracture was not accompanied by a concurrent association with osteoporosis. However, these two phenotypes cannot be expected to be completely interchangeable, as osteoporosis is often underdiagnosed and undertreated.
We could not formally replicate the associations of FOXO3A variation with ADL and bone fracture in another sample of Danish oldest-olds. There are a number of possible reasons for this. First of all, it may indicate that these were merely chance findings or that the Replication sample could be underpowered with respect to small effect sizes. However, another explanation could be that the individuals of the Replication sample were slightly older (94.7-100.9 years) than the individuals of the Discovery sample (92.2-93.8 years). This could potentially be of importance if the associations are not constant with age.