| Home | M.R. Rose | Reply 1 | J. Campisi | D. Promislow | Reply 2 | F.E. Yates | Reply 3 |
|
American Aging Association Web Forum 1 Responses to George Martin's 2009 Aging Cell
Paper on Posting date
2010/09/26.
This forum can be cited as Martin, GM, Yates, FE, Promislow, D and MR Rose
(2010) “Responses to George Martin's 2009 Aging Cell Paper on Epigenetic
Gambling and Aging” American Aging Association Web Forum 1 (eds. GM Martin and RJ Martin),
AmericanAging.org/Forum/1/. This forum is open for comment. Please send comments by visiting AGE Forum Comments.
Your message will be sent to editors of this forum, Dr. George M. Martin, Dept. of Pathol.,
Univ. of Washington, Seattle, WA 98195 and Dr. Rolf J.
Martin, MMT Corporation, Sherman, CT 06784. Contents of this
Forum:
The following paper by Dr. George M. Martin is
the subject of this forum. We thank the publisher Wiley and Aging
Cell for making this paper available online so that readers of this forum can have single-click access to it. George
M. Martin. 2009. Epigenetic gambling and epigenetic drift as an antagonistic
pleiotropic mechanism of aging. Aging
Cell 8:761-764. PDF
Commentary
on George Martin's “Short Take” on Epigenetic Gambling M.R. Rose I welcome the
stimulus supplied by this “Short Take” by my long-standing colleague, George
Martin; I hope that it leads to important research. As such research
proceeds, it is important to be clear as to exactly what is meant by bet
hedging as a life-history strategy.
Does it refer to (i) entirely stochastic switching between phenotypic
states with an increase in phenotypic variance over all environments, or (ii) does it refer to producing different
phenotypes via an adaptive plasticity that ends up reducing the average
phenotypic variance over this range of environments? Or (iii) both, depending on specific
circumstances? Or (iv) neither? Explicit definition of the hypothesized
bet-hedging phenotype is required to predict what its evolutionary
consequences will be in formal mathematical models (e.g. Frank and Slatkin,
1990). Among the
salient issues for theoretical delimitation are the hypothesized
relationships among average value of the varying phenotype, variance of the
phenotype, correlation between phenotypic states over environments,
correlations over time between environments, correlations among the
phenotypes of different genotypes – possibly mediated by interactions between
individuals with different phenotypes, and so on. All these things
matter, and need to be defined. The
theoretical evolutionary literature on these points is both long-standing
(e.g. Gillespie, 1974) and continuing to be developed (e.g. Wilbur and
Rudolf, 2006). As with many
quantitative and theoretical issues in gerontology, formal evolutionary
theory is neglected at peril. Second, the
evidential basis for making inferences about what evolution has done, or
might do in general, also requires attention to salient experimental
details. It is an important question
whether or not the N2 strain of C. elegans mentioned in Martin’s article
(link to be supplied) is well-adapted to the environmental conditions under
which its phenotypic variance in longevity is estimated. Evolution by natural selection responds to
environmental particulars (e.g. Leroi et al., 1994a), and this stricture
applies with even greater strength to the evolution of phenotypic plasticity
(e.g. Leroi et al., 1994b), when bet hedging is the evolutionary phenomenon
of interest. An arbitrary genotype
assayed in an arbitrary environment will not necessarily reveal what
evolution tends to produce when it has acted on an array of genotypes
maintained within a particular spectrum of varying environmental conditions
for enough generations for natural selection to respond to the specific
experimental paradigm. On the other
hand, the burgeoning techniques of experimental evolution (vid. Garland and
Rose, 2009) can demonstrate the evolution of bet hedging under conditions
that are well-defined with respect to both the evolving biological
populations and the specific choice of environmental regime (e.g. Beaumont et
al., 2009). Such work reveals both
that bet hedging can be favored under particular regimes of environmental
variation, as well as providing an opportunity to have experimental evolution
show us what the molecular genetics of such bet-hedging adaptations can
involve. Thus I suggest
that more scientific progress will be made if experimental evolutionists set
up evolving lineages of organisms that undergo biological aging under
conditions that are supposed to favor bet hedging, in order to see if it then
evolves as supposed. And if it does so
evolve, its mechanistic basis can then be uncovered using whole-genome
methods, such as re-sequencing of replicate lines in which bet hedging has
evolved compared to replicate lines which have not evolved bet hedging, among
other –omic technologies. This may be
contrasted with intuitive or narrative theories followed up with the search
for cellular mechanisms which might underlie the realization of a
hypothesized verbal scenario. The
latter scientific strategy would seem to be the opposite of strong inference. As with many
other important questions concerning evolution, we can now evaluate Martin’s
idea for the evolution of bet hedging using experimental evolution under
defined conditions and, furthermore, then proceed to identify the mechanistic
foundations of such products of evolution (cf. Beaumont et al., 2009). Ideally, the predictions to be tested in
the case of Martin’s hypothesis will be formally defined, perhaps by
simulation, prior to experimentation, so that there will be the least
possible ambiguity about the hypothesis being evaluated. Evolutionists
are always happy when other biologists are interested in basic evolutionary
hypotheses, and we particularly welcome opportunities to (i) refine such
hypotheses formally, (ii) determine their empirically salient consequences or
corollaries in explicit mathematical or simulation models, and then (iii)
definitively test them experimentally. Michael Rose
Department of Ecology
and University of
California Literature cited Beaumont HJE,
Gallie J, Kost JC, Ferguson GC, Rainey PB (2009) Experimental evolution of
bet hedging. Nature 462, 90-94. Frank SA, Slatkin M (1990) Evolution in a variable environment. Amer. Nat. 136,
244-260. Garland T, Rose
MR, eds (2009) Experimental evolution: Concepts, methods, and applications of
selection experiments. Berkeley, Calif: University of California Press. Gillespie JH
(1974) Natural selection for within-generation variance in offspring number.
Genetics 76, 601-606. Leroi AM,
Chippindale AK, Rose MR (1994a)
Long-term laboratory evolution of a genetic trade-off in Drosophila
melanogaster. I. The role of genotype x environment
interaction. Evolution 48, 1244-1257. Leroi AM, Kim
SB, Rose MR (1994b) The evolution of
phenotypic life-history trade-offs: an
experimental study using Drosophila melanogaster. Amer. Nat. 144,661-676. Wilbur HM,
Rudolf VHW (2006) Life-history evolution in uncertain environments: Bet
hedging in time. Amer. Nat. 168, 398-411.
George
Martin's reply to M.R. Rose
The first point to make in this response to his commentary is that I have concluded that, from the point of view of heuristics, Michael’s use of the time honored nomenclature of “bet hedging” should trump my invention of the term “epigenetic gambling”. If there are indeed genetic loci modulating the degrees of expression variegation in response to environments in which species evolve, the experimentalist should be completely agnostic about what classes of loci are involved. There is a currently a surge of new interest, new methods and new data concerning epigenetic modes of transcriptional regulation. Indeed, a highly relevant model along these lines appeared shortly after the publication of my brief 2009 publication in Aging Cell (Feinberg and Irizarry 2010). The gene actions responsible for bet hedging, however, could be based upon transcription, translation, post-translational modifications, or combinations of these mechanisms. Michael’s first point is of course of seminal importance if one is to design quantitative theoretical models and tests of those models. I was silent on those essential details because I had little basis for a focus upon one of several alternatives. I do imagine, however, a model that allows for a large dynamic range of gene expression rather than a system of phenotypic switching among two or a few states. I left open the important question of the degree to which such bet hedging is determined during development among isogenic organisms and isogenic families of cells and the degree to which it continues to be dynamic over the life course. The empirical data strongly support an increase in the variance of numerous phenotypes as functions of age, but this would be consistent with either model. Regarding Michael’s second point, I would imagine that the N2 wild type strain of C. elegans has by now has been well adapted to the more usual mode of husbandry, which requires living bacteria as food. Perhaps others might comment upon the degree to which this strain has adapted to a very different environment – suspension cultures with yeast extract and otherwise axenic and defined media. I could not agree more with Michael’s recommendation that “..more scientific progress will be made if experimental evolutionists set up evolving lineages of organism…” As for me, I would like to go back to my first love in science – somatic cell genetics – as a means of discovering mutant loci that can either decrease or expand cell to cell variations in gene expression. At age 83, however, all bets are off on whether or not I will succeed! Literature cited Feinberg, AP,
Irizarry, RA (2010). "Evolution in health and medicine Sackler colloquium:
Stochastic epigenetic variation as a driving force of development,
evolutionary adaptation, and disease." Proc Natl Acad Sci U S A 107 Suppl 1: 1757-1764. Rose, MR (1991). Evolutionary biology of aging. New York, Oxford University Press.
Comment
sent by Dr. J. Campisi Judith Campisi, Ph.D.
Commentary
on George Martin’s Epigenetic Gambling paper D. Promislow Dr. Martin’s thought-provoking paper focuses on the
adaptive significance of phenotypic variation early in life, and the way in
which early-age variation in traits such as reproductive output could
translate to variation late in life for survival. Evolutionary biologists
have long thought about the potential benefits of producing variable
offspring (i.e., ‘bet hedging’), as well as the related idea of how the
ability to respond to external perturbations can be beneficial (i.e., ‘adaptive
phenotypic plasticity’). It is surprising that these classic evolutionary
concepts have rarely been discussed in the context of aging, given the
extreme variability of age at death seen even under highly controlled
conditions. Thus, Martin’s article is both thought provoking and timely. But where Martin focuses on the selective forces
shaping variation, and its relevance to aging, here I point to the flip side
of this coin—namely, robustness. A plastic trait is one that responds to a
perturbation by changing its phenotype. In contrast, we usually think of a
robust trait as one that does not change in the face of a perturbation. This
is akin to Waddington’s notion of canalization (Waddington 1952). In a recent study of gene expression in yeast, Proulx et
al. (2007) examined the degree to which genes varied in response to
environmental and mutational perturbations. Their analysis suggests that
genes under relatively strong selection have evolved the ability to alter
expression in response to environment perturbations, but not to change in the
face of novel mutations. Proulx and colleagues suggest that trade-offs
between between environmentally-induced variation (plasticity) and
genetically-induced variation (or lack thereof) could lead to aging. Imagine that selection favors yeast cells that can
respond to environmental variability. The range of possible environments
(e.g., temperature, ethanol concentration, pH, etc.) in which a single yeast
cell grows might be unpredictable within the lifetime of that cell, but the
overall range might be quite predictable over a few tens of generations. In
this context, selection will have the opportunity to favor plastic responses.
However, changes due to a mutation at a specific gene are so rare that it
might make more sense to select for a robust (or canalized) response, whereby
expression patterns at other genes do not change in the face of a novel
mutation. A recently published study of robustness in Drosophila
(Frankel et al. 2010) shows that both environmental and genetic
robustness can be determined, at least in part, by the same molecular
machinery. As a result, conflicts may arise. Genes under strong selection for
plasticity early in life (Martin’s epigenetic gamble) may lack the robustness
required late in life as somatic mutations accumulate. In light of the work of Martin and Proulx et al.,
aging research might benefit from closer scrutiny of age-related changes in
plasticity and robustness, and the degree to which conflict between these two
basic phenomena might play a central role in the evolution and maintenance of
variation in ages at death. Daniel Promislow Department of Genetics University of Georgia Athens, GA 30602-7223 Literature cited Frankel N, Davis GK, Vargas D, Wang S, Payre F, Stern DL (2010) Phenotypic robustness
conferred by apparently redundant transcriptional enhancers. Nature. 466,
490-493. Proulx SR, Nuzhdin S,
Promislow DE (2007). Direct selection on genetic robustness revealed
in the yeast transcriptome. PLoS ONE. 2, e911. Waddington CH (1952) Canalization of the development
of quantitative characters. In Quantitative Inheritance. (ECR Reeve
, CH Waddington, eds). London: Her
Majesty’s Stationery Office, pp. 43-46.
Dr.
Martin’s response to commentary by Daniel Promislow Just as he did while I was developing my short essay on epigenetic gambling, Dan Promislow continues to educate this one time surgical pathologist about evolutionary biology. As noted in my response to Michael Rose’s commentary, I am increasingly fond of the much older term -“bet hedging”, as it leaves open the possibilities of a wide range of molecular mechanisms. Just as nature has evolved varying degrees of the fidelity with which DNA-dependent DNA polymerases copy genes, it is conceivable that she has also permitted variable degrees of fidelity at nodes of gene action that are much more distal – for example, the fidelities with which gene products can be activated or deactivated by kinases and phosphatases. It is not a bad hypothesis, however, to assume that variable degrees of transcriptional regulation are among the mechanisms that play roles in the evolution of various types of trade-offs. Such trade-offs have been a major focus of the Promislow lab for many years, especially as regards the biology of aging. That branch of biogerontology had its origins with the famous 1957 paper by George C. Williams (Pleiotropy, Natural Selection, and the Evolution of Senescence. Evolution 11;4: 398-411). I thank Dan for also referring to Waddington’s famous canalization concept, where the relevant gene actions are in development. I would love to know Dan’s secret for how well he keeps up on both the very old and the super-recent highly relevant literature. The Frankel et al. paper must have been hours old when he discovered it, just in time for his commentary! The extension of these ideas about the evolution of robustness and plasticity from the Proux, Nuzhdinof and Promislow paper on yeast to the Frankel et al. paper on fruit flies is most welcome. It is notable that both papers focus upon transcriptional regulation. The Frankel paper mentions a potential role for micro RNAs, an attractive idea. Their paper reminded me of the 2010 Feinberg and Irizarry paper, an abstract of which I included in my response to Michael Rose’s commentary. Both papers suggest particularly important roles for distal enhancers.
Commentary
on George Martin’s Short Take
The recent death of George Williams provoked me to check on where we stand now on the question: At what organizational level does natural selection, acting like a biased filter – a sort of biological Maxwell’s Demon, favor for their passage into the future some elements (to be defined) of living organisms, or of some influence they can convey? Below I highlight my understanding of the views current in 1966 when his (since classic) book: Adaptation and Natural Selection appeared. At about the same time a persuasive case for genetic determinism (e.g., “one gene, one protein”) came from Jacques Monod in a book entitled (English translation edition) Chance and Necessity (1971). The”chance” Monod referred to was vicissitudes of the external environments of organisms.
Traditionally, the strength of an (ineffable) property called “fitness”, of the entities in play in evolution, is the proximate substrate for a selection process. Fitness is a phenotypic attribute, though partly dependent on the underlying genotype of an organism. Fitness is identifiable retrospectively by the “average” contribution made by elements possessing it, to the gene pool of the next generation. From a game theoretic viewpoint, the “winning” of the evolutionary game that results from natural selection is simply that the game goes on for a group, allowing some modifications of average features of players in the surviving populations.
As we all know, Williams strongly believed that the chosen element on which natural selection acted was the individual gene, manifested through its associated individual phenotype. That phenotype had been affected by both genes and (some to be defined) environment(s). His basic position became dominant in evolutionary biology. He added the notion that some genes, when expressed, can have multiple effects that, in the phenotype, may be in competition - a situation called “antagonistic pleiotropy” (Williams, 1957).
Opposing the Williams position even now is that of “inclusive fitness” insisting that some genetic alleles in an individual can promote the “fitness” - a preferred ability to survive and reproduce- not just of the individual, but also of other individuals that possess that allele, as in “kin selection” (W.D. Hamilton – who wrote a series of profound, mathematical papers in the 1960s. See Harman, 2010). But, there are even more arguments against William’s position, coming from considerations of dynamics. (I write as a dynamacist, not a geneticist.) The natural world is full of synergies, mostly positive, but some negative (e.g., antagonistic pleiotropy). Peter Corning claims, and I agree, that “synergy ranks up there with such heavyweight concepts as gravity, entropy, and information as one of the keys to understanding how the world works and how we got here…Moreover, synergy has been a creative dynamo and a prolific source of innovation in evolution…” ( p 4 in Nature’s Magic: Synergy in Evolution and
the Fate of Humankind, Cambridge Uni. Press, 2003. This is a profound, richly documented, technical work.)
Today, how far we have already come from the perspective of genetic determinism. We now find influential aleatory events – chances- appearing both outside and inside, that affect organisms in many places : in DNA replication that is not always faithful in its transcription to RNAs; in the operations of multiple transcription regulators, including the dependence of some gene expressions on metabolic substrates. Recently recognized multiple classes of RNAs - including micro and inhibitory – that influence gene expression , introduce sources of phenotypic variability not previously foreseen. Translation of RNAs to proteins also opens up to non-genetic influences. In summary, the injection of chance at multiple points in development, and in daily operations of cells, can greatly increase the variety of organisms in the populations to which they belong. That enhanced variety encourages the evolution of new forms and functions.
Now George Martin’s essay , rich in ideas, proposes a novel expansion of the set of processes that can increase the range of variability in a population. He suggests that “random changes in gene expression (cellular epigenetic gambling or bet hedging) evolved as an adaptive mechanism to ensure survival of members of a group in the face of unpredictable environmental challenges. Once activated, it could lead to progressive epigenetic variegation (epigenetic drift) amongst all members of the group.” He provides details to put flesh on this skeleton. I welcome George’s ideas and his indication that they may help us understand distributions of lifespans of individuals in a population of a conspecific tribe, as well as the wide range of median or maximum lifespans across species (in Class Mammalia for example). With that focus he rightly considers manifestations of antagonistic pleiotropy as determinants of lifespans. HIs interest here is in aging and senescence and I think he got it right. But, rather than try to reinforce the arguments George offered and defended so well in his essay, I want to add something extra.
Having introduced synergies as agents with great creative potential, I would enlarge the scope of applicability of George’s concept of epigenetic gambling as a source of variabilities in populations – so important for evolution – noting that it is just as likely (or more likely) to lead to positive synergies as to the dreary negativism of antagonistic pleiotropy. The universe (and biosphere) show many levels of synergies in which those at one level serve as the building blocks for the next level. As they do, new dynamics arise at each level (in physics these are represented as the “equations of motion”). The idea that there is a vicious competition among “selfish genes” seems to me to be not only impoverished, but wrong. It is now obvious that genes do not usually operate as individualists, but rather in huge cooperative networks, synergistically. That synergism buffers against chance, yet still leaves room for rare and subtle changes.
Instead, the functional benefits of positive synergies – survival advantages- created by “bets” (George’s apt metaphor) invent novelties of various kinds, structural, functional and behavioral. It is these cooperative effects that have shaped progressive evolution of complexity In nature. Bad genes are the exception in multilevel evolution that I think clearly selects for positive synergisms. So, I say to George, your new idea has an importance and power extending way beyond how they explain why we grow old and die!
Martin seeks a plausible mechanism for the epigenetic gambling. Noise can always be found at the kT level of thermal energies for molecules. Classically, the total energy of a system corresponds to the sum of all the molecular energies of translation, rotation, vibration, plus electronic and nuclear energies, but new discoveries of quantum coherences and micro-motors in cell motions indicate the presence of constraints and possibilities previously unsuspected. Some of these are well discussed in an unusual (and controversial) book by Mae-Wan Ho (1993, 1998).
In 1992, on request, I published an extensive article on fractal applications in biology, looking for an understanding of local and global stabilities (and instabilities) that permit organisms to show regularities of forms and functions, and yet adapt – a condition of phenotypic plasticity, a variable capability that at the group level must affect the distributions of phenotypes in next generations. I described some rough tests for the presence of fractal time, fractal spaces, and chaotic dynamics in living systems based on the advances in nonlinear mechanics by 1992. Much has happened since then, and I recommend the new paper by Aon, Cortassa and Lloyd (2010) for a substantive update. I mention these articles as possible dynamic bases for the epigenetic gambling Martin proposes. As is now well known, modern nonlinear mechanics allows for deterministic operations sometimes to appear as “noise”.
The progression from genes, to kins to groups as substrates for natural selection has created many disputes, still ongoing, though I believe that a multi-level view will prevail. The new book by Harman focuses on the life of George Price, and the problem of altruism (Harmon, 2010). Price took a theoretical, mathematical approach to such questions, and eventually attracted warm support from Hamilton and John Maynard Smith. Harmon now explains the issues for evolutionary biologists, and a review of Harmon’s book by H. Allen Orr (2010) introduces the history competently. I see a common theme running through all the levels at which natural selection might operate, as I have indicated above. The theme is synergies. Hermann Haken has been a leader in technical investigations of synergies (1977). More recently, Peter Corning takes the lead ( 2003, 2005). These synergies depend on variations, so the proposal by George Martin greatly enriches the field, both practically and theoretically.
Professor of Medicine Emeritus, UCLA School of Medicine 1950 Sawtelle Blvd. #330 Los Angeles, Ca. 90025-7014 Literature cited Aon MA, Cortassa S, Lloyd L (2010) Chaos in biochemistry and physiology. Encyclopedia of Molecular Cell Biology and Molecular Medicine. Hoboken NJ: Wiley & Sons (in press). Corning P (2003) Nature’s Magic: Synergy in Evolution and the Fate of Humankind. Cambridge, UK: Cambridge University Press Corning PA (2005) Holistic Darwinism: Synergy, Cybernetics, and the Bioeconomics of Evoluton. Chicago: University of Chicago Press. Haken H (1977) Synergetics – An Introduction. Berlin: Springer-Verlag. Harmon O (2010) The Price of Altruism: George Price and his Search for the Origins of Kindness. New York: W W Norton. Ho M-W (1993, 2nd ed 1998) The Rainbow and the Worm: The Physics of Organisms. Singapore: World Scientific. Monod J (1971) Chance and Necessity. New York: A A Knopf. Orr HA (2010) Is Goodness in Your Genes? (a review of Harmon’s book cited above.) The New York Review October 14, 2010, pp 32-34. Williams G (1957) Pleiotropy, natural – selection, and the evolution of senescence. Evolution 11, 398-411. Yates FE (1992) Fractal Applications in Biology: Scaling time in biochemical networks. In: Methods in Enzymology 210 636 – 675. New York: Academic Press.
George Martin's response to F. Eugene Yates
Gene Yates was the inspiration behind the idea that epigenetic drifts can eventually result in sufficiently extreme excursions of biochemical oscillations such that they will have escaped from the physiological feedback controls that normally function to bring them back into functioning windows of homeostasis. One has to be a good detective to track down his original papers, as they may appear in book chapters not indexed by PubMed (a situation that one hopes will be addressed some day soon by the National Library of Medicine). Readers who wish to learn more about the intellectual background of his ideas on this subject should read the section on “From homeostasis to homeodynamics” in John Urquhart’s delightful biographical sketch (Urquhart, 2009). That essay also briefly outlines his interest in the biology of aging. For many decades, Gene was one of the very few integrative physiologists who showed up at meetings on the biology of aging (see Martin, 2002). Sad to say, this is still the case.
In his exceptionally thoughtful essay, Gene demonstrates his versatility by considering some deeper questions in genetics and evolution, questions that continue to be debated (see, e.g., Wilson & Wilson, 2007). In an early draft of my paper on Epigenetic Gambling, I invoked the notion of a sort of quasi kin selection, the kin in my case being cohorts of neighboring parenchymal liver cells that (except for variable degrees of ploidy) are expected to be genetically identical, yet with patterns of gene expression that may vary substantially. By that formulation, one can envisage patterns of gene expression that will survive certain environmental hepatotoxins and, thus, permit liver regeneration and survival of the host, despite the sad loss of some fallen brethren to the ravages of the toxic onslaught. (For other scenarios of environmental challenges, losers and winners would switch roles.) Comparable stories could be proposed for a variety of other tissues within multicellular organisms.
The key idea and very controversial notion in my Short Take was that the degrees of variegation of gene expression can be genetically modulated. Our tiny lab has now begun some pilot studies (with very little money) in attempts to identify, via mutagenesis, unusual degrees of gambling at the business of gene expression. We will try to isolate what we are calling “High Stakes Gamblers” and “Low Stakes Gamblers”. As an incentive to our semi-starving lab workers, I will offer each of them a one hundred dollar bill and a round trip to Las Vegas to try their luck at gambling – but only if our quest for such putative mutants is successful!
George Martin Literature cited Martin GM (2002) Help wanted: physiologists for
research on aging. Sci Aging Knowledge Environ. Mar 6;2002(9):vp2.
Urquhart J (2009) Living history: F. Eugene Yates. Adv. Physiol. Edu. 33: 234-242; doi:10.1152/advan.90165.2008
Wilson DS, Wilson EO (2007) Rethinking the
theoretical foundation of sociobiology. Q. Rev. Biol. Dec;82(4):327-48. |
| Top | M.R. Rose | Reply 1 | J. Campisi | D. Promislow | Reply 2 | F.E. Yates | Reply 3 |