Today's open access paper mounts an interesting argument, based on the use of a large data set for phenotypic aging in mice. They looked at transcriptomic and proteomic data for a sizable number of genes in a variety of different tissues, then grouping these into phenotypes by related function, or relation to specific age-related declines. Differences in expression by age in these phenotypic groups of genes were observed directly in mice and in human data sets.
The researchers then looked the effects on phenotypes of a few very well studied interventions widely thought to slow aging in mice: growth hormone signaling inhibition, mTOR inhibition, and intermittent fasting. The authors argue, based on their data, that these interventions are essentially compensatory rather than age-slowing, in that they appear to be changing phenotypes (mostly for the better) in a similar way in youth and old age, but they are not slowing the age-related change in those phenotypes. At least insofar as those phenotypes are assessed by the selected transcriptomic and proteomic data.
This is a very interesting view, given the present consensus that, yes, these interventions genuinely slow aging, setting aside some arguments as to whether mTOR is extending life in animal models only because it reduces cancer risk. It is a good illustration of the state of the present debate over strategies for intervention in aging, shaped by the lack of a strong consensus on how to define aging in a way that is useful for the assessment of therapies in animal models or human trials. One can always look at obvious external signs of dysfunction, such as grip strength, but it will never be completely clear, given only those biomarkers, as to whether a therapy helps because it is compensating, or because it legitimately does in fact address mechanisms of aging.
It is a reasonable supposition that better therapies will be better because they reverse underlying mechanisms of aging, and therefore will produce lasting benefits to patients in many aspects of health. As a strategy, this is the right way forward, but the expectation of better outcomes for aging-targeting therapies is by no means a given for any specific therapy and specific age-related condition. If we can point to interventions such as mTOR inhibition that appear to slow the age-related decline of a great many of those aspects of health, and show that they are in fact only broadly compensatory instead, it muddies the waters considerably when it comes to steering the research and development communities towards better approaches to therapy.
Deep phenotyping and lifetime trajectories reveal limited effects of longevity regulators on the aging process in C57BL/6J mice
A large body of work, carried out over the past decades in a range of model organisms including yeast, worms, flies and mice, has identified hundreds of genetic variants as well as numerous dietary factors, pharmacological treatments, and other environmental variables that can increase the length of life in animals. Current concepts regarding the biology of aging4 are in large part based on results from these lifespan studies. Much fewer data, however, are available to address the question of whether these factors, besides extending lifespan, in fact also slow aging, particularly in the context of mammalian models.
It is important to distinguish lifespan vs. aging because it is well known that lifespan can be restricted by specific sets of pathologies associated with old age, rather than being directly limited by a general decline in physiological systems. In various rodent species, for instance, the natural end of life is frequently due to the development of lethal neoplastic disorders: Cancers have been shown to account for ca. 70-90% of natural age-related deaths in a range of mouse strains. Accordingly, there is a strong need to study aging more directly, rather than to rely on lifespan as the sole proxy measure for aging.
Deep phenotyping represents a powerful approach to capture a wide range of aging-associated phenotypic changes, since it takes into account alterations at molecular, cellular, physiological, and pathological levels of analysis, thereby providing a very fine-grained view of the consequences of aging as they develop across tissues and organs. The approach is therefore ideally suited to assess genetic variants, pathways, dietary or pharmacological factors previously linked to lifespan extension and, potentially, delayed aging. Deep phenotyping examines hundreds of parameters, many of which are expected to differ between young and old animals (hereafter called age-sensitive phenotypes; ASPs); these can be collectively used to address if and how a given intervention interacts with the biological processes underlying the signs and symptoms of aging.
We here refer to the mechanisms of aging as the sets of processes that underlie age-dependent phenotypic change. Accordingly, an intervention that targets the mechanisms underlying aging should slow the transformation of a phenotypically young to a phenotypically aged organism. In other words, the intervention should attenuate the age-dependent change in ASPs (the delta in phenotype between young and old). For instance, a specific intervention or genotype could ameliorate the age-dependent loss of neurons by promoting processes concerned with maintaining the integrity of neurons over time.
An intervention could mimic a targeting of age-dependent change by affecting ASPs directly (i.e., independently of age-dependent change in these phenotypes). For instance, a specific genetic variant may increase the number of neurons by promoting neurogenesis during brain development, without affecting the rate of subsequent age-dependent neuron loss. This variant would regulate neurodevelopmental processes but would not affect the mechanisms underlying age-dependent change. Although this would also result in increased neuronal numbers in old age, it cannot be taken as evidence of a slowed progression of aging because the rate of age-dependent change remains unaltered
Here, we employ large-scale phenotyping to analyze hundreds of markers in aging male C57BL/6J mice. For each phenotype, we establish lifetime profiles to determine when age-dependent change is first detectable relative to the young adult baseline. To cover key genetic longevity interventions and study their effects on aging in mice, we here chose genetic models targeting the mTOR pathway as well as growth hormone signaling. In parallel to our studies in mice, we applied multi-dimensional phenotyping combined with stratification based on genetic expression variants in GHRHR and MTOR in a human population across a wide age range, spanning from 30 to 95 years. The analyses in humans complement our work in animal models and allowed us to address, in parallel to the work in mice, whether or not a potential genetic modification of human ASPs occurs in an age-independent fashion or not.
We examine these key lifespan regulators (putative anti-aging interventions; PAAIs) for a possible countering of aging. Importantly, unlike most previous studies, we include in our study design young treated groups of animals, subjected to PAAIs prior to the onset of detectable age-dependent phenotypic change. Many PAAI effects influence phenotypes long before the onset of detectable age-dependent change, but, importantly, do not alter the rate of phenotypic change. Contrary to a general expectation that 'anti-aging' treatments should produce a broad change in aging rate across many phenotypes, our study shows that the PAAIs we examined - that are concerned with some of the very core mechanisms proposed to be involved in aging - often did not seem to work through targeting age-dependent change.
In conclusion, the PAAIs examined (i.e. mTOR loss of function, Ghrhr loss of function, intermittent fasting-based version of dietary restriction) often influenced age-sensitive traits in a direct way and not by slowing age-dependent change. Previous studies often failed to include young animals subjected to PAAI to account for age-independent PAAI effects. However, any study not accounting for such age-independent intervention effects will be prone to overestimate the extent to which an intervention delays the effects of aging on the phenotypes studied. This can result in a considerable bias of our view on how modifiable aging-related changes are.