Evolution Of Senescence

Why do we grow old? Should not natural selection favor an ageless phenotype in which there is no decline of vigor or reproductive output with age? How could it be an evolutionary advantage for individuals to lose their vigor and reproductive capabilities as they age? In this section, we will see how the fitness measures derived above allow us to address these important questions.

Let us start with a population of ageless individuals who show no senescence over their entire lifetime. Being ageless is not the same as being immortal. Individuals who do not age still can die through accidents, predation, disease, and so on. They are ageless in the sense that the chances of dying in an interval of time do not depend upon their age. Let d be the probability of an individual dying in a unit of time. We regard d as being independent of age and a constant throughout the entire lifetime, reflecting the ageless phenotype of the individual. The individual is also regarded as being ageless with respect to reproduction by letting mb, the probability of having mated times the expected number of offspring in a time unit, also be independent of age and a constant throughout the entire lifetime. Given these ageless parameters, the probability of an individual living to age x is

Then, the net reproductive rate of an ageless individual is mb

x =0 x =0 d using the well-known formula for the sum of a geometric series [sn = a + ag + ag2 +----

+ agn-1 = a(1 - gn)/(1 - g) ^ a/(1 - g) as n ^-<x>, where a and g are constants with -1 < g < 1]. We can also apply the sum of a geometric series to Euler's equation to obtain the Malthusian parameter for this ageless population as mb

1 = £ e-rxlxmb = mbJ2 [(1 - d)e-T = 1 _ (1-d)g_r ^ r = ln(1 - d) - ln(1 - mb) (15.14)

When d and mb are small numbers, Taylor's series approximation of r from equation 15.14

Now suppose a mutation occurs in this ageless population such that the bearers of this mutation senesce and die at age n - 1. The net reproductive rate of the mutant individuals is n-1

For large n and any d < 1 (that is, some death occurs from causes unrelated to age), the term (1 - d)n goes to 0, and hence the term in brackets in equation (15.15) goes to 1. Thus, if n is large enough (depending on d), then R0 « R0 and the mutation is selectively neutral as measured by the net reproductive rate. Similarly, one can show that the mutant's Malthusian parameter obeys the implicit approximation r' « mb{ 1 - [(1 - d)e r']n} - d & mb - d for large n

Once again, as long as senescence is delayed to an old age, the mutant phenotype is neutral. As we saw in Chapter 5, neutral mutations will inevitably become fixed in a population over long periods of time. This means that if mutations can occur that kill their bearers at a sufficiently advanced age, such mutations are effectively neutral and some will go to fixation, thereby destroying the agelessness of the initial population.

In Chapter 13 we discussed some of the selective pressures on Huntington's chorea, one of several neurodegenerative diseases in humans associated with trinucleotide repeats which have a late age of onset. Langbehn et al. (2004) found that the empirical relationship between age of onset and CAG repeat number is well described by the equation

where Age is the age of the individual, CAG is the number of CAG repeats (see Chapter 13 and Figure 13.11), and S (Age, CAG) is the probability of having no neurological symptoms to the given age with the given repeat number. If we make the assumption that all reproduction stops with the onset of the neurological symptoms, the net reproductive rate for a bearer of Huntington's chorea is max age

Using the life history data in Table 15.1 in equation 15.18, we can calculate the net reproductive rate of bearers of a newly formed Huntington's allele (one that just crossed the threshold repeat number and reached a value of 36 repeats, as discussed in Chapter 13) to be 0.9941 versus the normal net reproductive rate from Table 15.1 of 0.9968. Theneurolog-ical symptoms are relatively mild when they first occur and then get progressively worse, eventually resulting in death. The assumption that all reproduction stops with the onset of symptoms is therefore overly conservative, so the actual difference in net reproductive rates associated with a newly formed Huntington's is even less than that indicated above. Hence, the lethal neurodegeneration of Huntington's disease is essentially neutral with respect to natural selection when an allele reaches the 36 repeat threshold. Of course, as discussed in Chapter 13, there are other targets of selection on Huntington's disease, including the family and the repeats themselves. Focusing just upon the selection on the repeats themselves and ignoring the family-level selection, we saw in Chapter 13 that selection at the genomic level favors an increase in repeat number, which in turn is associated with an earlier age of onset (equation 15.18 and Figure 13.11). As the age of onset is lowered, there is now stronger individual-level selection against Huntington's disease. For example, the net reproductive rate for bearers of a Huntington's allele with a CAG repeat number of 56 using equation 15.18 and the life history data in Table 15.1 is 0.4407 versus the normal net reproductive rate of 0.9968, resulting in substantial selection at the individual level against the Huntington allele. These calculations show how important the age of onset is in determining the fitness impact of an allele that affects life history parameters. Even lethal genetic diseases are effectively neutral when the age of onset is old enough. As a result, individual selection alone cannot prevent the evolution of senescence via genetic drift leading to the fixation of nearly neutral alleles with deleterious effects of late age of onset.

We now turn our attention to another class of mutations with life history effects. Suppose, as before, a mutation occurs that kills its bearers at age n - 1. However, we now assume that this same mutation increases earlier reproduction from mb to mb' such that mb' > mb. For example, suppose this mutation is associated with transferring the energy used in maintaining viability after age n - 1 to reproduction at earlier ages. This mutation is therefore associated with a pattern of antagonistic pleiotropy (Chapter 11) because it is associated with traits that have opposite effects on fitness. The net reproductive rate of the individuals with this antagonistic pleiotropic mutant is

As before, the term in brackets in equation 15.19 goes to 1 as n increases, so if the age of onset of the deleterious effects of this mutant is old enough, then its net reproductive rate is approximately mb'/d, which is greater than the net reproductive rate of the nonmutants of mb/d. Similarly, one can show that the Malthusian parameter for this pleiotropic mutant is, for n large, approximately mb' - d > mb - d. Once again, by either fitness criterion, bearers of this pleiotropic mutant are actually favored by natural selection as long as the deleterious effects have a late age of onset. In this case, our initial ageless population will evolve senescence due to the positive action of natural selection; that is, it is adaptive to senescence.

Many mutations have been found that are associated with beneficial effects early in life and deleterious effects later in life. For example, in Chapter 11 we discussed several gene loci associated with resistance to falciparum malaria. Since most of the mortality associated with this parasite occurs in childhood, the beneficial effects of these malarial resistance genes are primarily expressed at an early age. However, the deleterious effects of these same genes (often associated with the chronic effects of anemia) are often not clinically significant until later in life. Many life history trade-offs have been documented in the fruit fly Drosophila melanogaster and in particular between larval survival and adult size. Bochdanovits and de Jong (2004) examined the trade-off between larval survival and adult size by an analysis of global gene expression. This quantitative genomic approach revealed 34 genes whose expression explained 86.3% of the genetic trade-off between larval survival and adult size. Fourteen of these genes had known functions that suggest that the trade-off is at the level of cellular metabolism and is due to shifts between energy metabolism and protein biosynthesis regulated by the RAS signaling pathway. These and other studies indicate that mutations associated with patterns of antagonistic pleiotropy are common.

Both effectively neutral, late-age-of-onset mutations and antagonistic pleiotropic mutations will lead to the evolution of senescence. An empirical demonstration of the evolution of senescence is provided by an experiment on the flour beetle Tribolium castaneum (Sokal 1970). Sokal allowed adults of two strains (wildtype and black bodied) to live three days after emergence from the pupal stage and then killed them. As a control, he allowed adults of the same two strains to live as long as they were able under identical laboratory conditions. He ran these lines for 40 generations and then let adults from all lines, both experimentals and controls, to live as long as they could under the specified laboratory conditions. The results are shown in Table 15.3.

Table 15.3 reveals that the experimental populations had a shorter life span than the controls for both strains and both sexes. Thus, in just 40 generations, these stocks evolved

Table 15.3. Adult Life Spans of Flour Beetles from Two Strains Raised for 40 Generations with All Adults Either Killed at Three Days of Age (Experimental) or Allowed to Live as Long as Possible under Laboratory Conditions (Control)

Average Adult Life Span in Days

Strain Sex Experimental Control

Wildtype Female 5.0 7.5

Wildtype Male 12.5 14.5

Black bodied Female 5.6 5.9

Black bodied Male 5.0 8.8

increased senescence in response to an environmental regimen in which no adult lived past three days of age, thereby rendering any deleterious fitness effects past that age irrelevant as a contributor to fitness. These effects could be due to a combination of mutants of either type discussed above, although the speed of the response indicates selection on at least some genetic variants with antagonistic pleiotropy. Note also that these results are yet another example of genetic assimilation (Chapter 14). In this case, an environmentally imposed early death (by the experimenter) becomes genetically assimilated as an innate decrease in life span. These experimental results indicate that we grow old because the evolutionary forces of genetic drift and natural selection actually favor senescence over agelessness. Once again, if we take the gametic perspective rather than the individual perspective, there is no mystery as to why selection favors those phenotypes that transmit more gametes even at the price of the senescence of the individuals that are the temporary bearers of those gametes.

Was this article helpful?

0 0

Post a comment