[
1987]
To my knowledge, a theory of "developmentally programmed aging" has never been explicitly stated, although the notion that aging has some relationship to development has certainly been proposed many times. In the preceding chapter (36), Dr. Hayflick has made a brief description of the central idea of developmental programming within aging. In order to discuss relevant evidence in this chapter, I would like to propose the following, somewhat more specific and operational definition: The theory of developmentally programmed aging posits that aging involves events controlled in ways recognizably similar to those that operate during development. This definition is perhaps a little less extreme than it might have been, since it uses the phrase "aging involves events" rather than the phrase "aging is caused by events." However, I think it captures most of the usual connotations of "developmentally programmed aging," and it at least has the virtue of testability. Of course, to test the theory, as defined, requires evidence of several sorts. In particular, it requires (a) that we understand how some aging events are controlled, (b) that we understand how some developmental events are controlled, and (c) that we know how to recognize whether there is or is not similarity between the two. A central message of what follows is that we are really only at the beginning of being able to test this theory, although some lines of approach do appear
[
1987]
The concept of developmentally programmed senescence has been outlined by Leonard Hayflick (this volume), and examples from development have been used as exemplars of "developmentally programmed senescence" (Richard Russell, this volume). Unlike development, senescence has probably evolved in the absence of direct selection for increased longevity, perhaps as a direct result of the absence of such selection. (For an excellent review see Charlesworth.) A popular evolutionary model that has received experimental support suggests that senescence may result from pleiotropic effects of selection for adaptive life history characteristics. In the literature on aging, less rigorous arguements have been used to suggest that in human evolution, a delay in the age of senescence has been indirectly selected for by means of so-called longevity assurance or longevity-determinant genes. However, all explanations for the evolution of senescence are theoretical, and, with few exceptions, remain largely untested. Like Dr. Hayflick and Russell, I will assume that by developmental programming we mean genetic specification. I will use a general definition so as not to preclude examples that fail to meet one or more of the rigid criteria defined by Russell (this volume). This less rigid definition of programmed aging is necessary, because unlike development, where genetics has been successfully applied for 50 years, examples of genetic specification of senescent processes are quite few. In the literature on aging, it is still not widely accepted that mutants can alter fundamental patterns of senescent events in well-defined ways. One purpose of this presentation is to outline a few examples. In senescence, large batteries of new genes are not differentially regulated; this is quite unlike development, where many genes are differentially regulated. The molecular etiology of senescence is unknown in almost every instance and, as such, makes the study of aging a fascinating area for inquiry. If senescence is unlike development in lacking differential gene regulation, what are the approaches that are likely to yield useful results in the analysis of senescence and the aging process? The developmental genetic paradigm is a useful, indeed essential, theoretical construct for approaching the aging process in an experimental context. The lack of a suitable model organism in which classical and molecular genetics can be productively combined with other experimental techniques has impeded our understanding of senescence. Despite a general lack of evidence for genetic specification, there are instances where genetic specification is clearly evident; the analysis of mutational events that alter normal senescence in well-defined ways demonstrates this point. These instances also provide experimental models for dissecting the aging