by
|
by |
|
| This article will appear in a forthcoming issue of Trends in Genetics. | |
Abstract
Heritability of Lifespan
The Iceland study used a new statistical method, the minimum founder test (MFT), that was developed specifically to suit this kind of population database. The MFT is straightforward, if computationally intensive, and works as follows. First, define the trait of interest, in this case longevity. Second, identify the set, S, of all individuals with the trait. Third, interrogate the database to identify the minimum ancestor set, F(S,Y), of people who were born not earlier than a given time point Y. An ancestor set is a set of individuals within the database such that each person in S has at least one ancestor (or him/herself) in the set. As Y goes backwards in time, the function F(S,Y) gives values that can only decrease or remain the same. If the trait has a heritable component, F(S,Y) will decrease faster than if the trait is independent of ancestry. Evidence of heritability and its statistical significance can then be judged by comparing F(S,Y) with results obtained for 500 randomly selected control sets, each control set being matched to S in terms of sample size and the distribution of dates of birth.
The sample set S considered in the Iceland study comprised all babies born between 1870 and 1900 who lived to at least the 95% percentile of the sex-specific lifespan distributions. This totalled 1531 individuals. The minimum founder set was traced back to the year 1500, by which time it had approached an asymptote of about 600. In line with expectation if longevity is inherited, the founder set for the long-lived babies was smaller than the average founder set for randomly selected control sets. The difference, which averaged about 5-10% (depending on the value of Y), was consistently more than two standard deviations from the control mean.
Genes versus Environment
So far, so good. There are several difficulties, however, in going further with such data. These highlight the particular challenges in assessing the genetic contribution to human longevity. Although it is well known from animal models that genes influence lifespan [12-16], human studies are confounded by the necessary reliance on observational, rather than experimental, methods and by the challenge of teasing out the genetic contribution from other forms of inheritance. In humans, 'inheritance' has a broader definition than mere genetic transmission, being 'anything acquired or possessed by descent or succession'.
Nongenetic inheritance can affect human traits as much as genetic inheritance. A wide range of transmissible phenotypic characteristics can be important, including lifestyle, habits, wealth, employment, education, land tenure and social status. Regrettably, the data are almost invariably inadequate to control for these variables. In some circumstances, indirect inference using appropriate surrogate variables, such as the longevity of kin, can be used to throw light on the probable balance between genetic and environmental factors affecting human life-history variables [17,18], but the problem is particularly acute when addressing the familial inheritance of lifespan [19].
Gudmundsson et al. [11] argue that Iceland has exceptional socioeconomic uniformity, having been comprehensively poor for most of its history yet enjoying great equality under present-day healthcare arrangements. Similar claims can be made for other populations, such as the villages of the Valserine Valley in the French Jura studied by Cournil et al. [10]. Unfortunately, there can be no certainty that uniform general conditions guarantee the absence of fine-grained heterogeneity in environmental factors, such as land quality, that might contribute to nongenetic inheritance of lifespan. It is, therefore, hard to dispute the overall conclusion by Gudmundsson et al. [11] that, 'we have clearly shown that there is a strong familial component to longevity.' But there remains a frustrating degree of uncertainty about their second conclusion that, 'it is likely that this familial component is a genetic one.'
In resolving the problem of shared environment, twin studies are the most reliable indicator of a genetic contribution to human longevity, although even these are not entirely without difficulty. The upbringing of monozygotic (MZ) twins is often different from that of dizygotic (DZ) twins and could account for greater behavioural and developmental similarities between MZ twins than between DZ twins, which in turn might contribute to lifespan. Nevertheless, we can regard the estimates of lifespan heritability coefficients from twin studies as the upper bound for a genetic contribution to human longevity, and it would be surprising if environmental factors made more than a modest contribution to these. It is striking, therefore, that the estimates for heritability coefficients, although clearly significant, are uniformly low (Table 1) only rarely exceeding 30%. It is unfortunate that the minimum founder test applied to the Icelandic data does not furnish a quantitative estimate of heritability to compare with these figures. However, the impression gained from the modest difference in the sizes of founder sets between the test and control samples is that the Iceland data do not contradict the inference that heritability of human longevity is only modest.
Which genes?
Given that there is evidence for heritability of human lifespan, at least some of which is likely to be genetic, an important question is the nature of the genetic contribution. What kinds of genes contribute to longevity and how many of them are there? The first question is addressed by the evolutionary theories of ageing and by the tests that have been made of these [16]. It seems likely that several kinds of genes might be involved in determining longevity, particularly those that regulate the levels of somatic maintenance and repair. Thus, there are good a priori grounds for supposing that potentially many genes affect human longevity.
Gudmundsson et al. [11], however, concluded that, 'the genetic component to longevity may depend on one or a few genes.' This inference was based on an analysis of how an individual's chance (risk ratio) of becoming long-lived was affected by having long-lived kin. It was found that the risk ratios were consistently greater than unity, ranging from 1.3 for first cousins to 1.8 for full sibs. The authors remarked that this rate of decrease of the risk ratio with meiotic distance is more consistent with one or a few genes contributing to longevity, but qualified by the observation that the data are consistent with an additive multilocus model (multiple genes working independently). In the case of the invertebrate model systems Caenorhabditis elegans and Drosophila melanogaster, a number of single gene mutations have been identified that have major effects on longevity. However, these organisms are known to exhibit considerable life history plasticity, and in the case of C. elegans several of the mutations map onto the genetically controlled alternative developmental pathway that results in the long-lived stress-resistant dauer larva. It will surprise many if single genes with major effects on human longevity, other than those involved in pathological conditions such as progeria, are found to exist.
Longevity and Gender
Several studies have examined the extent to which longevity might be transmitted differentially between parents and offspring of either sex. Although the interpretation of these studies is subject to considerable uncertainty about whether the factor being transmitted is genetic or environmental, it is interesting to review the consistency, or rather inconsistency, of the patterns that have been reported. Cournil et al. [10] found the heritable component of longevity to be much stronger for daughters than sons. Some have found similar effects [8,9], whereas others have reported weaker inheritance by daughters, chiefly because of a weak father-daughter correlation [2,3,7].
Gudmundsson et al. [11] also examined this question with respect to whether long-lived individuals confer a late-life survival advantage on their progeny. And they examined how this transmission was influenced by sex. Over the age range 70-90, mortality rates for male offspring of long-lived fathers and for male and female offspring of long-lived mothers were reduced by 25%. A smaller reduction of 14% was observed for the female offspring of long-lived fathers.
Although there are grounds to suppose that there could be sex-specific transmission of genetic factors that influence longevity (e.g. X or Y chromosomes, mitochondrial DNA), a study of maternal and paternal transmission of longevity to male and female offspring among British aristocrats during the period from 700 to 1875 (reference 20) suggests that there might be a complex interplay of environmental and genetic effects. When the period for which the aristocrat records are available was divided into two phases, pre-1700 and post-1700, significant differences were found. The significance of the year 1700 as a phase-boundary is that, among this relatively privileged group, 1700 marks the point at which the demographic transition commenced. In the period between 700 and 1700, the probability of longevity of men but not women was dependent on longevity of the parents, especially the father. In the later period (1701-1875), longevity of both men and women was dependent on mothers, and the longevity of women but not men was also dependent on fathers. The most likely explanation for the contrast between these periods is that it was the environment, not the genetics, that underwent important changes. However, we should also consider the possibility that as patterns of survivorship have changed, there might have been some alteration in the selection forces that act on the genetic determinants of the human life history.
Conclusion
Human longevity is perhaps as complex as any trait can be because length of life is influenced not only by 'longevity assurance' mechanisms, such as antioxidant defenses and DNA repair, but also by susceptibility to a wide range of diseases, such as heart disease, dementia, cancer and stroke [21]. Longevity is also influenced to a considerable extent by environment and probably by gene-environment interactions, as well as by the intrinsic actions of chance [22]. Nevertheless, in view of the great changes that are occurring in the age structure of populations around the world, it is increasingly important to understand the nature of the genetic contribution to human longevity, which in turn could help us to understand better the complex nature of ageing. The opportunity to address these questions within comprehensive population studies, such as that in Iceland, represents a welcome advance.
Artificial Chromosomes: Ideal Vectors? - a review of progress. From Trends in Biotechnology, 2000, 18:5:218-223. Full text available from BioMedNet.
Gene Therapy: The First Decade - summarizes the approaches taken, results achieved, lessons learned, and important recent developments. From Trends in Biotechnology, 2000, 18:3:119-128. Full text available from BioMedNet.
Centromeres: The Missing Link in the Development of Human Artificial Chromosomes - a review of recent work. From Current Opinion in Genetics & Development, 1998, 8:219-225. Full text available from BioMedNet.
Artificial Chromosomes Coming to Life and Principles for Human Gene Therapy Studies - two recent articles from Science magazine. Registration required for full text.
Delivering the Goods, Gene Therapy's Trials and Tribulations, Gene Therapy Soldiers On, and Genomes and Gene Therapy - four recent articles on gene therapy. From The Scientist.
Introduction to Gene Therapy - includes molecular biology background information, genetic transfer technology, applications, and additional resources. From a course taught at Vanderbilt University.
Institute for Human Gene Therapy - provides news, information, and resources on gene therapy. From the University of Pennsylvania.
Gene Therapy - a journal Web site that allows free access to tables of contents and abstracts, but requires registration for the full text of articles.
Related HMS Beagle articles: