« May 2009 | Main | July 2009 »

June 25, 2009

The bitter fountain of youth

"When stress predicts a shrinking gene pool, trading early reproduction for longevity can increase fitness, even with lower fecundity." That's the title of a paper that Will Ratcliff, Mike Travisano, Peter Hawthorne and I just published in PloS-One. This was a spin-off from Ratcliff's work on the timing of reproduction in bacteria, but our main conclusions should apply broadly to plants and animals, with important implications for human health. Our entire paper is available on-line, but here is some additional background and explanation.

Earlier, I blogged about our research at UC Davis showing that tomatoes grown using organic methods have higher concentrations of a specific chemical (Mitchell, et al. 2007). Plants make this chemical to defend themselves against insects, which may be why there was more of it in tomatoes not protected by artificial pesticides. Surprisingly, this chemical actually seems to benefit human health. At the time, I thought this might just be coincidence, and wrote that "some of the natural insecticides plants make... are likely to be harmful to humans, rather than beneficial."

Now, I'm not so sure. It turns out that many toxins, including natural insecticides, can have health benefits in low doses, a phenomenon known as hormesis (Mattson & Cheng. 2006). Other forms of mild stress, such as dietary restriction (calorie restriction, intermittent fasting) or high temperature, have also been shown to increase longevity.

How can stress be beneficial? Some stresses trigger various protection mechanisms, such as antioxidants or heat-shock proteins, which may increase lifespan, even relative to individuals not exposed to stress. But why aren't these protective mechanisms turned on all the time, rather than only under stress? Don't individuals with longer lifespans leave more descendants than those with shorter lifespans? Not necessarily.

What if some mechanisms that increase lifespan also delay sexual maturity or decrease the rate of reproduction? For example, what if the blood pressure that maximizes lifespan is lower than that which maximizes reproduction? Then a gene for lower blood pressure would not necessarily increase in frequency over generations. A trade-off between early reproduction and longevity (and later reproduction) was central to the "antagonistic pleiotropy" hypothesis of Williams (1957). Our paper builds on this widely accepted hypothesis.

Given trade-offs between early and late reproduction, when will natural selection favor genes that potentially increase longevity but delay reproduction? Sometimes, resources not used for reproduction can be invested in growth, increasing reproduction in future years. Also, more experienced individuals may care for their offspring better. But what if delaying reproduction doesn't increase either the number of offspring or their survival?

We showed that delaying reproduction can still increase Darwinian fitness, that is, proportional representation in the gene pool, provided that overall population size is decreasing. Hamilton (1966) pointed out that an offspring added to a smaller population represents a larger fraction of the total gene pool. Therefore, if total population is increasing, offspring produced earlier have a larger effect on fitness. But if population size is decreasing, then offspring produced later have a larger effect on fitness. This means that delaying reproduction can sometimes increase fitness, even if delay does not increase the number of offspring.

Most populations will alternate between increasing and decreasing in numbers. If the population is stable or increasing, delaying reproduction can only decrease fitness. This is especially true if there is a high risk of death from causes unrelated to reproduction. But if the size of the gene pool is likely to decrease, delaying reproduction can increase fitness. This is especially true if risks directly or indirectly associated with reproduction are large relative to other risks.

Our mathematical models show that the best strategy is to delay reproduction only when an individual's chance of surviving to reproduce later is high, and only when an individual has reliable information predicting a decrease in overall population size. This is where stress comes in.

Past population declines were often caused by shortages of food, which can affect both the amount and types of food eaten. For example, natural insecticides in plants often have an unpleasant taste. Over most of our evolutionary history, therefore, these plants may have been eaten only when preferred foods, like meat or fruit, were not available. Consumption of these "famine foods" would therefore have been a reasonably good predictor of population decline, so they may trigger physiological changes (lower testosterone, etc.) that increase longevity while tending to delay reproduction.

A remarkable result, seen in both nematode worms and fruit flies, is that food odors can reverse the beneficial effects of dietary restriction on longevity (Libert, et al. 2007). If an individual smells food, others may be eating that food, so population size may be increasing. In that case, delaying reproduction would be a losing strategy, even if reproducing now increases the chance of an early death.

What about humans? Our models assumed that individuals reproduce only once, then die, like salmon or soybeans. However, we expect that some of our results will apply to species, like humans, with more complex life histories. One result for humans that is consistent with our hypothesis is that artificially sweetened soft drinks are just as likely to cause metabolic syndrome (related to diabetes) as sugared soft drinks are (Lutsey, et al. 2008). Like food odors, sweet foods may have been correlated, over much of our evolutionary history, with abundance, and therefore with impending increases in population size. If we want to live longer, maybe we should instead eat foods whose chemical composition or flavor remind our bodies of past famines. The health benefits we get from eating vegetables like kale may be due, in part, to the chemicals that give them their slightly bitter taste.

High levels of toxins, including natural ones, are still presumably harmful. But low doses of plant toxins, perhaps especially those found in traditional famine foods, may often improve health. This assumes that our hypothesis is correct, so you might want to wait for the results of experiments we are planning before making major changes in your diet.

We are also assuming that most people would consider some decrease in potential reproduction to be acceptable. For the many humans that already choose to limit their own reproduction, this need not result in any decrease in actual family size. For example, if people don't expect to marry until after college, the risks of early fertility may outweigh the benefits, even apart from health effects of hormone levels etc. in the teenage years on health later in life. Delaying puberty might, however, result in larger adults, with possible negative implications for automobile fuel economy and other resource issues.

Another popular hypothesis has been that individuals benefit from delaying reproduction in a bad year and waiting until conditions are better. This may increase the number of offspring produced, but we show that it does not increase proportional representation if the entire population also reproduces more in the good year.

"How is putting our entire kingdom to sleep for 100 years better for my family than losing one daughter, however much we love her?" asked the queen. "In 100 years, our other children would have had countless grandchildren. Meanwhile, those in neighboring kingdoms will multiply. By the time the impenetrable thorn forest you put around our kingdom dies and we awake, our enemies will vastly outnumber us."

"Not necessarily", replied the fairy scientist, "My computer models predict 100 years of wars, famines, and plagues. It's true that your population won't grow, but those of your enemies will shrink. This would have been a winning strategy, even if there were another way to save your daughter's life."

Ratcliff, Travisano, Hawthorne, and Denison. Can you spot the model?


Hamilton WD. 1966. The moulding of senescence by natural selection. Journal of Theoretical Biology. 12 : 12-45

Libert S, Zwiener J, Chu X, VanVoorhies W, Roman G, Pletcher SD. 2007. Regulation of Drosophila life span by olfaction and food-derived odors. Science. 315 : 1133-7

Lutsey PL, Steffen LM, Stevens J. 2008. Dietary intake and the development of the metabolic syndrome: The atherosclerosis risk in communities study. Circulation. 117 : 754-61

Mattson MP, Cheng A. 2006. Neurohormetic phytochemicals: Low-dose toxins that induce adaptive neuronal stress responses. Trends in Neurosciences. 29 : 632-9

Mitchell AE, Hong YJ, Koh E, Barrett DM, Bryant DC, et al. 2007. Ten-year comparison of the influence of organic and conventional crop management practices on the content of flavonoids in tomatoes. Journal of Agricultural and Food Chemistry. 55 : 6154-9

Williams GC. 1957. Pleiotropy, natural selection, and the evolution of senescence. Evolution. 11 : 398-411

June 2, 2009


I will be visiting in-laws in Hawaii and then at the International Congress on Nitrogen Fixation, so may not post again until late June.

Maybe by then I will be able to write something about our recently accepted paper on the evolution of aging.

Or maybe I'll have good news about an NSF grant. One grad student has a little money left in his own small grant, but the other two are also doing interesting and important work and need to pay for supplies, flow cytometer use, etc.