Biological Restraints on Indefinite Survival

Abstract

Multiple observations that organismal life span can be extended by nutritional, genetic, or pharmacological intervention has raised the prospect of transforming medicine with the goal of slowing, stopping, or even reversing age-associated disease and maintaining or restoring health and resilience in the increasing numbers of elderly across the world. The potential for such an enterprise is supported in theory by plant and animal models of negligible senescence, most notably the small, freshwater organism Hydra spp. The existence of some very long-lived species, including bowhead whale, Greenland shark, and giant tortoises, suggests that increased healthy life spans in humans, significantly higher than the current known maximum life span of about 120 years, may be possible. Here we discuss the biological restraints on human life extension based on the evolutionary basis of aging and our current genetic and molecular insights into the processes responsible for age-related loss of function and increased disease risk.

While life is eternal and has existed on our planet for nearly 4 billion years, most individual life forms have to deal with limits. Throughout human history, mortality, one of those limits, has always been difficult to accept. This goes back to the more than 5000-year-old Epic of Gilgamesh in which the hero sets out to conquer death but fails. Likewise, according to the Bible, there was no aging and death when Adam and Eve still lived in paradise. Only after they had committed the original sin, they brought God's curse of death upon themselves and every other living thing as well. But also in real life, humans have always tried to cheat death. China's powerful first Emperor, Qin Shi Huang, refused to accept his mortality, always seeking the fabled elixir of life. He died nevertheless and left us his immortalized terracotta army at his tomb close to his capital city of Xianyang (close to current Xi'an).

Interestingly, things have not changed a great deal since those days. Quests for immortality are still common, most of them easily recognizable as quackery, but scientists contemplating technology's power to conquer aging and death play an increasingly important role as part of new, well-funded private research organizations (Eisenstein 2022). Most notably in this respect is the concept of “longevity escape velocity,” the idea that life span will be manufactured by medical technology faster than the rising risk of death due to aging (de Grey 2004). Rather than seeking immortality through elixirs of life, rejuvenation is now often considered possible through regeneration (e.g., by turning somatic cells back into partially undifferentiated stem cells) (de Magalhães and Ocampo 2022).

Meanwhile, demographic studies suggest that our prospects of breaking through the limits of human life span have not really gotten much better since Qin Shi Huang's crude attempts (Colchero et al. 2021). What has dramatically improved, however, is health span. Owing to the explosion of new technology in the nineteenth century the two great killers of the past (i.e., famine and infectious diseases) have been largely tamed, at least in technologically developed countries. And over the last decades, progress in medical research has greatly increased our capacity to optimize the health of older people, resulting in significantly improved survival of the oldest old. However, more recently, we have begun to see diminishing returns of these efforts, with improvements in survival with age rapidly declining after age 100. Indeed, while the number of centenarians has exploded over the last 30 years, their mortality rate of ∼50% per year has not changed (Modig et al. 2017). After steadily rising since reliable data have become available in the 1950s, the maximum reported age at death has not increased since the 1990s (Dong et al. 2016). Subsequent extreme value statistical analyses from large data sets of elderly individuals indicates that improvement in the human life span, defined by us as the oldest known person, has ceased (Gbari et al. 2017; Einmahl et al. 2019). Indeed, the world's oldest known person ever is still Jeanne Calment, who died in 1997 at the age of 122 years, 164 days (Allard et al. 1998). No other person, before or since, is known to have reached even the age of 120 years.

Conclusions that the limits of human life span has been reached are somewhat clouded by two issues. First, life expectancy at birth steadily increased from a mere 50 years in the most developed countries in 1900 to over 80 years in most developed countries today. This long-term pattern led some authors to conclude that life span itself is fluid and perhaps even malleable (Oeppen and Vaupel 2002). However, it should be noted that, most recently, that is, since 2010, the rise in life expectancy has decelerated or even reversed slightly in many developed nations (Cardona and Bishai 2018). This pattern was evident even before the appearance of COVID-19, which accentuated that reversal.

Second, experimental laboratory populations of worms, flies, and mice have had their life spans increased, sometimes dramatically. In the nematode, Caenorhabditis elegans, for instance, genetic mutations and environmental manipulation have increased median and maximum longevity nearly 10-fold (Shmookler Reis et al. 2009). Survival of individual laboratory mice of standard genetic backgrounds eating standard mouse diets seldom exceeds 3 years, but a single mutation in the growth hormone receptor produced a mouse that lived almost 5 years (Pilcher 2003). Moreover, in worms and flies, the typical age-related increase in mortality has been found to slow, stop, or even decrease at later ages (Curtsinger et al. 1992). Various attempts have been made to extrapolate these findings of a mortality “plateau” at the oldest ages to human population studies. Most notably, from studying data on all inhabitants of Italy aged 105 and older, it has been concluded that human mortality rate reached a plateau after the age of 105, suggesting there may be no theoretical limit to human longevity (Barbi et al. 2018). However, even if this were true (it has been disputed [Gavrilov and Gavrilova 2019]), the probability of survival at this age is so low that the influence of such a mortality plateau on maximum human life span would be miniscule (Beltrán-Sánchez et al. 2018).

Of what, if any, relevance are studies of short-lived, genetically modified laboratory species to the future of human longevity? Are there other species, species with longer life spans or that fail to show signs of aging without genetic manipulation, that might be more relevant? Here we will review the biological restraints on indefinite survival. First, we will discuss the evolutionary logic underlying species-specific limits to life span with a focus on possible exceptions in terms of organisms that show negligible senescence. “Senescence” we should emphasize as we use the term is synonymous with “aging,” the widespread decline in physiological function that is associated with increasing age. It is not shorthand for cellular senescence, a different concept entirely. We will then discuss the possible primary mechanisms of aging that prevent extended species life span. Finally, we will consider the biomedical advances needed to break through the basic biology that underlies current limits to human life span.

EVOLUTIONARY BASIS OF LIMITED LONGEVITY

Mathematical modeling of evolutionary processes has revealed that aging exists because of the decline in the force of natural selection after the age of first reproduction due to unavoidable environmentally imposed death (Rose 1991). Indeed such modeling led W.D. Hamilton to conclude that senescence “cannot be avoided by any conceivable organism” (Hamilton 1966) and that conclusion was widely accepted in the evolutionary biology community, at least for organisms where there was a clear distinction between somatic and germ lines. However, that strong statement was eventually challenged and alternative models parameterized differently were developed that allowed for organisms that do not undergo senescence, or even do the reverse, display negative senescence, that is, improve survival and reproduction with increasing age (Baudisch 2005). At some point, however, mathematical models need to be validated against the real world, and, in fact, organisms that escape aging were devilishly difficult to discover. Even the exquisitely well-studied bacterium Escherichia coli, which had always been assumed not to age was discovered to do so (Rang et al. 2011). In fact, the only reasonably well-documented species not observed to undergo signs of senescence appear to be cnidarians in the genus Hydra (Martı´nez 1998; Schaible et al. 2015).

Recently, however, a number of other species, generally those that live exceptionally long lives, have been claimed to undergo “negligible senescence,” a useful term, but one that disguises an important distinction. That distinction is between species that age very slowly such as Greenland sharks, which may live several centuries (Nielsen et al. 2016), and those that do not age at all. In practical terms, it is a distinction between species that grow old but do so slowly, and those that show no signs of senescence at all. Such species would remain eternally youthful, in other words, the very grail of human aspirations. To the extent that non-aging species exist, particularly if they are widespread, it would give some credence to the possibility that we will break the limits of human longevity and perhaps create eternal youth for ourselves.

EVIDENCE FOR THE ABSENCE OF SENESCENCE

The astronomer Carl Sagan often said that extraordinary claims (in science) require extraordinary evidence. Eternal youth in any species would be an extraordinary claim, so what sort of extraordinary evidence would validate such a claim? We feel that three legs are required to support that particular stool. First, there must be evidence from survival patterns. Senescence is often measured as the rate of increase in mortality with age. The most common method is to evaluate the Gompertz exponent, which is the slope of a line relating age to the logarithm of age-specific mortality (Finch 1990). The shallower the slope, the slower the aging. A slope of zero indicates the absence of mortality senescence. So no increase in mortality rate with age is one important bit of supportive evidence. It is necessary but not sufficient. The other aspect of demographic senescence is reproductive rate. A non-aging organism should also display no decrease in reproductive rate with increasing age. Finally, there is physical or physiological evidence. Non-aging organisms should show no signs of deteriorating physical capabilities with increasing age.

As noted above, the poster children for lack of senescence are laboratory populations of species of the cnidarian genus, Hydra, small, freshwater animals consisting of a base, stalk, and mouth surrounded by tentacles. In the laboratory, they reproduce primarily by budding. We emphasize laboratory populations to make the point that in the real world of nature, there is no such thing as immortality. Environmental dangers are omnipresent, whether they are predators, pathogens, poisons, famine, or climatic catastrophe. The laboratory environment minimizes or eliminates such hazards and thus may reveal the potential for immortality. It is this potential that Hydra provides the strongest evidence to support. Martı´nez found no increase in mortality, or decrease in budding rates in several groups of Hydra vulgaris over 4 years (Martı´nez 1998). That study was later followed by Schaible and colleagues (including Martı´nez), who evaluated mortality rates of hundreds of Hydra of three strains and two species in two different laboratories for up to 8 years. They also observed no increase in mortality or decrease in budding rate over that period (Schaible et al. 2015). Despite these results, a legitimate question is whether 8 years is sufficient time to detect senescence in Hydra, if it exists. For an animal that develops in weeks and reproduces several times per month, it would seem like this should be sufficient time, but it is difficult to be certain. It is too bad that no physical or physiological function assays were done over this time as they would confirm whether Hydra really achieve something approaching eternal youth. Assays such as spontaneous contraction rate, prey capture ability, and regeneration rate, which have been used in other Hydra studies (Yoshida et al. 2006), could have made a most compelling case for eternal youth.

Other species with at least some claim to lacking senescence include the so-called immortal jellyfish, Turritopsis nutricula, another cnidarian with the unique life history feature that, under stress, it can reverse developmental stages from a fully independent, sexually mature jellyfish-like adult back to its juvenile Hydra-like state, then develop once again back to its adult stage (Piraino et al. 1996). From this unique life history feature, it has developed the reputation (and nickname) of potential immortality, although no reports exist that it can make these transformations an unlimited (or even large) number of times.

Other claims for the absence of senescence are based virtually entirely on long life, sometimes combined with information that the animals are still reproductively active at later ages. Greenland sharks have developed such a reputation despite the fact that no mortality trajectories with age are available. Recently though, two papers have reported no age-related increase in mortality in multiple species of turtles. Da Silva and colleagues take direct aim at Hamilton's claim that, due to evolutionary dynamics, senescence is inescapable, by analyzing mortality data from zoo populations of 52 turtle and tortoise species (da Silva et al. 2022). They find that approximately three-quarters of those species exhibit mortality patterns statistically consistent with an absence of senescence in a Gompertzian sense and that a few additional species display mortality senescence but at lower rates than humans. How compellingly does this challenge Hamilton's claim? To state the obvious, evolutionary hypotheses should be critically evaluated in the environments in which evolution occurred (i.e., in nature). Da Silva and colleagues do point out that evidence from separate studies carried out in nature on three of the species they analyzed, all provided evidence for senescence. The most thorough of these field studies, a 24-year field study of more than 1000 painted turtles (Chrysemys picta), found evidence for both mortality senescence and reproductive senescence. Importantly, in that study, although egg production increased slightly with age, egg quality decreased, so that fewer eggs actually hatched in older females. So merely counting egg production in older individuals is not sufficient. Overall Darwinian fitness indeed declined with age (Warner et al. 2016).

A related question is how to interpret Gompertzian slopes. Does a shallower slope inevitably suggest slower aging or could there be contextual factors that need be considered? For an example of how context matters, consider that American women in the year 1900 showed a slower rise in mortality with age than they did in the year 2000 (Fig. 1; Bell and Miller 2005), yet this hardly suggests that women aged more slowly in 1900. Life expectancy was 30 years longer in 2000 and women were healthier at every age. The difference in slope simply reflects a bigger change in early life relative to later life mortality over the twentieth century. Considering both Gompertz slope and another survival metric, in this case life expectancy (although it could have been a number of others), helps interpret the plot. To help interpret the Gompertz slopes in da Silva et al., we note that life expectancies of many of the turtle species seem low in zoos. In a number of species, reported life expectancies are shorter than the age of first reproduction, which would be unusual in iteroparous species and none of the species show greater-than-human life expectancies. So whether the Gompertz slopes indicate lack of senescence or a high rate of nonsenescent deaths throughout life is not clear.

Figure 1.

Age-specific mortality in American women in 1900 and 2000 (Bell and Miller 2005). Note the lower Gompertz slope in 1900. Clearly this does not represent slower aging. Life expectancy among American women in 1900 was 49 years and in 2000 was over 79 years. Women were healthier in 2000 at all ages. The change in slope reflects greater progress in reducing mortality at younger compared with older ages.

A second study by Reinke and colleagues reports mortality patterns generated from individual mark-capture data from long-term field studies of 77 species of turtles, tortoises as well as other ectothermic vertebrates, including crocodiles, snakes, lizards, frogs, and salamanders (Reinke et al. 2022). They also report apparent lack of mortality senescence in a number of species, again, particularly among the turtles. While provocative, particularly so as these are studies from nature, there is still no evidence on reproductive senescence or its absence or on changes in physical function with age. One limitation of both of these papers is the small number of animals alive at later ages, where senescence would be the most evident if it exists. Determining age-specific mortality rate with confidence requires a significant number of individuals of the age in question. One thing is clear, however, irrespective of these published survival patterns. Even in the longest-lived giant tortoise species, individuals do not maintain eternal youth. They eventually become aged. Harriet, a Galápagos tortoise reputedly 170 years old at her death in 2006 succumbed to a heart attack, and Jonathan, supposedly the oldest currently living terrestrial vertebrate at an estimated age somewhere between 160 and 190 years has been blind from cataracts, has lost his olfactory sense, and consequently has had to be fed by hand since 2015 (Austad 2022). If the fountain of youth is to be found, it may not be among the oldest turtles.

To be sure, both studies are of considerable interest, but both also fall short of demonstrating a lack of senescence in any of the species in question.

PRIMARY MECHANISMS SETTING LIMITS TO LIFE

In spite of challenges posed by seemingly immortal or very slowly aging species, the decline in the force of natural selection after the age of first reproduction remains the best explanation of why we age. This would lead to the accumulation in the germline of mutations not subject to purifying selection because they are acting only late in life (Medawar 1952). However, this tells us little about the nature of the mechanisms that ultimately determine life span of a species. If aging is caused by late-acting genetic variants accumulating by genetic drift, why are there so many similarities in aging phenotypes among individuals of the same species, and even between species? It was George Williams who first came up with the idea that such late-acting gene variants have important functions at an early age, driving their selection through fitness effects (Williams 1957). His argument was that genes, or in fact genetic variants, can have beneficial effects at a young age, which is why they have been selected during evolution, but adverse effects at later ages. To illustrate this “pleiotropic gene theory of aging,” Williams himself provided an example in the form of a hypothetical gene variant positively selected during evolution because it had a favorable effect on bone calcification in the developmental period. However, as a side effect it produced depositions of calcium in the arterial walls. Because the adverse effect of calcium deposition in the arteries only shows up late in life, evolution would not select against this hypothetical gene variant.

At the time it was not realized that many if not most genes have multiple functions. But by now many examples of pleiotropic genes or gene regulatory pathways have been identified. For example, pathways that cause inflammation, a hallmark of aging that has been found to be associated with multiple diseases, from arthritis, cardiovascular disease and cancer to diabetes mellitus, chronic kidney disease, non-alcoholic fatty liver disease, and autoimmune and neurodegenerative disorders, must have been selected because aggressive immune response protects the young against infection (Finch 2010).

Other pleiotropic pathways are those that are part of the somatotropic axis. Growth and reproduction are obviously under positive selection because they enhance fitness. However, they have multiple late-acting adverse effects, as was dramatically illustrated by the discovery, first in the nematode worm, that weak mutations that dampen such activities increase life span. Inhibiting insulin/IGF-1 signaling in worms, flies, and mice extends life span, as mentioned earlier, sometimes to a very large extent, at least in the worm (Kenyon 2005). Dampening the somatotropic axis intertwines with the paradigm for all interventions in aging: dietary restriction (DR) (Brown-Borg 2015).

As expected, there is a price tag on this increased longevity and it has to be paid in the wild under natural conditions (Jenkins et al. 2004). While trade-offs in the control of longevity remain incompletely understood (Maklakov and Chapman 2019), it is highly unlikely that interventions in normal patterns of growth and reproduction will provide a rational strategy to extend limits to life span.

With the exceptions mentioned, the observed life span gains by dampening growth and reproduction in organisms more complex than worms are modest and appear to diminish with complexity (Vijg and Campisi 2008). For DR, it also remains difficult to distinguish between a true life span effect or the amelioration of an artificially unhealthy diet in laboratory animals (Wolf 2021). This is especially true for the various inbred mouse and rat strains, as suggested by the possible absence of a DR effect in wild mice (Harper et al. 2006).

While interventions in the many pleiotropic pathways associated with early fitness have the potential to increase health span with modest gains in mean life span, they are likely not a good strategy to extend species-specific life span limits. For that purpose we need to closely examine the pathways that allow a species to maintain its somatic tissues for long periods of time. According to Kirkwood's disposable soma theory (Kirkwood 1977), life span is the product of a balance between investments in growth and reproduction on the one hand and somatic maintenance on the other. If this is true, we may be able to intervene in the sources of the wear and tear that require somatic maintenance or possibly manipulate the somatic maintenance pathways themselves. However, the main problem here is that there are not just a few cellular defense systems, but a great many.

In his imperfectness model, Gladyshev argues that aging is a consequence of numerous forms of damage collectively termed the deleteriome (Gladyshev 2016). While in an organism such as Hydra such damage could be diluted out by the processes of continuous cell division and selection, explaining its potential immortality (above), in organisms with fully differentiated nonrenewable cells and structures, this same damage will accumulate and cause aging. As the processes that generate these deleteriomes are under genetic regulation, different species will have different capabilities to deal with them, which may explain their different life spans. It is tempting to speculate that evolution used loci involved in somatic maintenance to change life span over time, such as the delayed aging and longer life spans that developed in insular opossums as compared to those on the mainland due to greatly reduced exposure to predation (Austad 1993). This increase in life span from 31 to 45 months in a short time span (i.e., over 4000–5000 years of separation) is difficult to explain from systematically eliminating the late-life adverse effects from genes critical for maintaining early-life fitness.

It has been argued that somatic DNA damage would be central to this deleteriome and affects most, if not all, aspects of the aging phenotype, making it a potentially unifying cause of aging (Schumacher et al. 2021). Advances in sequencing technology have now provided ample evidence that one of the consequences of DNA damage, somatic mutations, accumulate in most if not all human and animal tissues with age (Zhang and Vijg 2018), at a rate that was found to be tissue-specific (Fig. 2). Moreover, in primary fibroblasts from different rodent species, an inverse correlation has been found between mutation burden induced by the same dose of a mutagen and maximum life span (Zhang et al. 2021). This finding has now been extended to a much broader range of species (Cagan et al. 2022).

Figure 2.

The number of single-nucleotide variants (SNVs) per cell as a function of age in human B lymphocytes, liver hepatocytes, and lung bronchial cells. (Figure created from data in Zhang et al. 2019, Brazhnik et al. 2020, and Huang et al. 2022.)

Also, the capability of DNA repair to remove DNA damage correlates with life span (Tian et al. 2019), something that has been suggested since the 1970s (Hart and Setlow 1974). Because the DNA of the genome is the central repository of all functional information, with somatic mutations both inevitable and irreversible, the possibility cannot be ruled out that interventions at that level, if at all possible, would be sufficient to modify species-specific life span. However, it remains unclear whether somatic mutations or DNA damage in general are a key causal factor in age-related functional decline and disease apart from cancer. Moreover, genome maintenance and its regulation are extremely complex, making it highly unlikely that interventions at that level will emerge soon.

CONCLUSIONS AND FUTURE PROSPECTS

Based on demographic data, there is now strong evidence that human life span has a practical natural limit. As death rates reach the annual level of 30%–50%, as is the case at extreme old age, it becomes pointless to speculate on an unlimited life span without significant amelioration of the fundamental aging processes. A natural limit to a species’ life span is in keeping with Hamilton's conclusion that senescence cannot be avoided due to the decline in force of natural selection after the age of first reproduction. Exceptions to this rule, most notably Hydra, are explained from a continuous state of cell renewal similar to embryos. For all other examples of extremely long-lived species, the evidence for a lack of senescence is inconclusive. Breaking through the glass ceiling of evolution's iron law that has given each species a formidable repertoire of late-acting adverse genetic variants, some of which have beneficial effects critical to early life, seems a bridge too far even for the twenty-first century's powerful technology.

The immediate conclusion that can be drawn from this is that the focus of geroscience should be on improving health span, which, in contrast to radical life extension, is testable. Indeed, this has already been very successful in adding life to years, and while there is evidence for diminishing returns (Dong et al. 2016), new geroscience approaches may well reinitiate further improvements (Kaeberlein et al. 2015). For instance, it is by no means clear that we have yet discovered the optimal lifestyle interventions to preserve and prolong health and certain drugs currently in early phase clinical trials may be capable of delaying the onset of groups of diseases simultaneously (DeVito et al. 2022).

To substantially delay the aging process that caps maximum life span extension in our species, it will be necessary to mimic evolution, but then in real time, to eliminate the multiple causes of aging while retaining early fitness and species’ identity. This would require approaches to increase cellular defense, which in view of the utter complexity of the deleteriome seems beyond our current technological capacity, although what the future holds is, of course, unknown.