Research on aging requires the ability to measure aging, and therein lies a challenge: it is impossible to measure every molecular, cellular, and physiological change that develops over time, but it is difficult to prioritize phenotypes for measurement because it is unclear which biological changes should be considered aspects of aging and, further, which species and environments exhibit “real aging.” Here, I propose a strategy to address this challenge: rather than classify phenotypes as “real aging” or not, conceptualize aging as the set of all age-dependent phenotypes and appreciate that this set and its underlying mechanisms may vary by population. Use automated phenotyping technologies to measure as many age-dependent phenotypes as possible within individuals over time, prioritizing organism-level (i.e., physiological) phenotypes in order to enrich for health relevance. Use those high-dimensional phenotypic data to construct dynamic networks that allow aging to be studied with unprecedented sophistication and rigor.
Introduction
“Aging” is a convenient label for a diverse set of biological changes that develop in a high proportion of individuals within a population over an average lifespan. The concept that aging is worth studying is essentially a hypothesis that these conditions share causal mechanisms and that working to identify and combat those shared mechanisms is a viable strategy to improve the quality and duration of life. However, research in this area is undermined by preconceived notions that aging is a universal, intrinsic process that is distinct from disease and transcends species and environments. This leads to an excess of effort spent attempting to define and measure “true aging.” There is no such thing. Instead, I propose de-emphasizing the classification of phenotypes as aging (or not) and, instead, measuring the multi-dimensional set of all possible age-dependent phenotypes, which can change between populations and environments. By measuring these phenotypes over time, aging can be modeled as a dynamic network. This eliminates the need for arbitrary cutoffs, such as dictating a distinction between normal aging and disease and trying to only measure the former. Many phenotypes progressively change with time, and we cannot measure everything; because the ultimate goal is to understand and ameliorate declines in health and well-being, I propose taking the “top-down” approach of initially focusing on organism-level phenotypes that change with age. The field now has the hardware to measure such phenotypes and the computational infrastructure to analyze them: tools such as metabolic and behavioral monitoring chambers and video monitoring paired with machine vision are making automated, high-dimensional longitudinal phenotyping of model organisms a reality. A network approach has several benefits: (1) it quantifies aging in high-dimensional space, allowing the nuanced assessment and comparison of interventions; (2) network structure may elucidate phenotypic clusters, suggesting shared mechanisms; and (3) comparison of the networks of different populations, e.g., mice and humans, may improve preclinical models of aging by identifying the preclinical phenotypes that are the most predictive of human phenotypes. Although a high-dimensional network is more complex than single endpoints such as lifespan or a handful of health-span parameters, it is also a more accurate and useful picture of reality.
Aging Is Population Dependent
What we call “aging” is a set of phenotypes (generally considered deleterious) that develop in a high proportion of individuals within a population over an average lifespan (a population, in this case, meaning any set of individuals under study, which in practice is a species or subspecies in a particular environment). This set depends on context, namely the specific species or subspecies and environment: our description of the aging phenotypes of a human is not identical to our description of the aging phenotypes of a mouse. Further, because aging is considered the “normal” trajectory for individuals, the set of phenotypes we term “aging” depends entirely on what we subjectively consider a “normal” version of each species and a “normal” environment. In short, the phenotypes we term aging are based on a reference point. It follows then that there is no formal requirement that the mechanisms of aging be shared between populations—as the phenotypes of aging vary by population, the mechanisms driving those phenotypes might vary by population as well. Therefore, the hypothesis that the age-dependent decline experienced by another species is causally similar to our own is just that—a hypothesis. The temptation to presuppose this hypothesis, i.e., to assume that the biological mechanisms of aging must be universal and thus span species and environment, is an erroneous extension of the observation that, with rare exception, all biological entities deteriorate with time. It is true that all versions of aging are ultimately driven by the passage of time itself, but that is hardly a level of causality we can or want to modulate, and there is no requirement that the aging of different populations share any layer of causality beyond this. We are interested in identifying the modifiable, biological mechanisms that affect the rate and nature of physiological decline, and there are no a prioriconclusions we can make about the similarity of such mechanisms between species and environments. In short, all things break down; this does not mean all things break down for the same reasons.
However, although not a formal requirement, some aging mechanisms may indeed be shared between populations. On one hand, aging may in fact be largely universal, with most aging phenotypes across many species and environments being driven by the same set of common mechanisms (Figure 1, left). In this case, all roads lead to Rome, and we can study aging in whatever population we choose without losing human relevance. Alternatively, there may be some common mechanisms that cross species and environments, but they may only affect a subset of aging phenotypes and/or affect phenotypes only partially. This model is perhaps the most consistent with current data. There are a number of cellular and molecular phenotypes that arise with age in multiple species, e.g., protein aggregate formation, DNA damage accumulation, and reduced ATP generation by mitochondria, suggesting there are common aging processes at work, although it is less clear whether delaying or reversing those cellular and molecular phenotypes leads to therapeutic benefit. In terms of interventions, inhibition of insulin/insulin-like growth factor 1 (IGF-1) signaling (IIS) extends the lifespan of worms, flies, mice, and perhaps humans as well as improves (or slows the decline of) a number functionally relevant phenotypes in all of those species. These data, along with similar results for other nutrient-sensing pathways such as TOR, are evidence that at least some mechanisms of functional decline are conserved. Nevertheless, inhibition of IIS is far from a panacea—not only do animals with reduced IIS signaling still decline over time, a number of age-dependent phenotypes are unaffected, or even worsened, compared to age-matched control animals. While mutations in some IIS pathway components, notably FOXO3, have been associated with human longevity, the maximum lifespan extension achieved to date by IIS inhibition drops when moving from invertebratesto mammals, and the same mutations that dramatically extend lifespan in mice do not cause similar lifespan extension in humans (e.g., growth hormone receptor deletion), suggesting important rewiring. Lastly, there is the possibility that the mechanisms of aging are so specific to species and the environment that virtually no important mechanisms are shared between multiple populations, and model organisms will, at best, be a poor abstraction of human aging. The data from nutrient-sensing pathway inhibition suggest that reality is not as unfortunate as this; however, we still know of no mechanism that equally affects age-dependent phenotypes in all species and environments, so it is premature to discount this possibility entirely. Further, it is not difficult to think of species whose version of aging is likely “population specific,” that is, unlikely to share any causal similarity with the aging of humans or standard model organisms, e.g., semelparous species such as Pacific salmon. Some might say that salmon do not experience “true aging,” but as discussed above, this is merely an argument from an arbitrary reference point. The more correct statement, and the one that should guide our thinking, is that the aging of salmon is probably not like human aging and thus has little relevance for our purposes.
Figure 1. Hypothetical Models of How the Mechanisms of Aging Might Differ between Populations and Environments
Circles represent sets of causal mechanisms driving age-dependent decline for specific populations and environments. Overlap between circles represents shared causal mechanisms.
(Left) Aging is largely universal. A set of common mechanisms drive a high percentage of aging phenotypes across many populations. Targeting these common mechanisms ameliorates most aspects of aging in both model organisms and human clinical trials.
(Middle) Aging is partly conserved, partly population specific. A common set of mechanisms exists but drives only a subset of aging phenotypes or affects phenotypes only partially. Targeting these common mechanisms is beneficial, but only partially; most of the age-dependent decline within a population is driven by mechanisms unique to that population.
(Right) Aging is extremely population-specific. Almost all drivers of aging phenotypes change as species or environment changes, with virtually no mechanisms that are common across many species and environments. One should not assume that studying aging in any non-human context will have clinical relevance; models must be critically evaluated. Many human-relevant aging mechanisms are impossible to study in model organisms, and aging mechanisms even vary between human populations living in different environments.
.......
The rest at the source: https://www.scienced...405471219300390
