I. Introduction
A. Definition of Senescence
1. Senescence in humans usually connotes a rapid deterioration of physical
and mental capacity
that begins at an advanced age.
2. Generally, this deterioration leads to increased proability of mortality per unit time.
3. As will become evident shortly, similar rapid increases in probability
of mortality, associated
with physiological deterioration, occur in many, if not most, organisms.
4. In every-day life, we recognize senescence in terms of the outward
manifestation of the
aging process.
5. However, from an evolutionary perspective, we wish to define senescence
in terms of its
effects on fitness.
6. Consequently, evolutionary biologists use the following definition
of senescence: Senescence
(or aging) is a persistent decline in age-specific fitness components of an
organism due to
internal physiological deterioration. (From M. R. Rose)
a. By fitness component, evolutionary biologists mean either probability
of survival per
unit time or fecundity (number of offspring produced per unit time)
b. Consequently, evolutionary biologists consider senescence to occur
if, as organisms
age, either the probability of mortality increases or the number of offspring
produced
per unit time decreases.
B. Recoginzing senescence
1. Senescence manifested as an increased probability of mortality over
time can be recognized by
the form of the survivorship curve for a population.
2. A survivorship curve, l(x), plots the probability that
an indivdual is still alive as a function of
age, x:
a. Absence of senescence means that the probability of mortality does
not change as
the organism ages.
b. Another way of saying this is that for any given time interval (e.g.
a year in humans),
the probability that an individual will die does not depend on its age.
c. If we designate that probability of mortality as for all ages
, then the
probability an individual is alive after one time interval is .
d. The probability that an individual survives two time intervals is
simply
P[survives interval 1 and survives interval 2] = P[survives interval 1] x
P[survives interval 2]
= p2.
e. More generally, the probability that it survives x time intervals
is just px
f. But this just means that the survivorship curve is given by
l(x) = px
g. But if we recognize that p can be rewritten as a power of
e, i.e. p=e-a, then we
have
.
l(x) = (e-a)x = e-ax
h. In other words, with no senescence, the survivorship curve of a
population is
exponentially decreasing, as in part A of the following figure (red line).
i. Alternatively, if one plots the logarithm of l(x) as a function
of age, one obtains
a straight line,as depicted in part B of the figure (red line):
ln[l(x)] = ln[px] = x ln[p] (note that ln(p)
is a constant < 0)
4. Survivorship curve for a population that senesces
a. By definition, a senescent population is one in which the probability
of survival
per unit time decreases as an individual ages.
b. This may be represented by
p(x) = pax
(note that since p < 1, p(x) decreases as x increases).
c. Then,
l(x) = p(1) p(2) p(3) … p(x)
= pa p2ap3a … pxa
= pax(x-1)/2
d. Taking logarithms of both sides of this equation yields
ln[l(x)] = x(x-1) ln[pa/2]
e. This is a concave-downward function, as portrayed in part B of the
previous figure (blue line).
f. On a linear scale, this survivorship takes the form of one of the
red curves in part B
of the figure.
5. Conclusion: A survivorship curve the shape of the blue curves
in part A of the previous figure
is indicative of senescence.
C. Example: Senescence in Drosophila melanogaster. ( Miquel et al. 1976.
Effects of temperature on the
life span, vitality and fine structure of Drosophila melanogaster. Mech.
Ageing Dev. 5: 347-70)
1. Laboratory populations set up and proportions of original cohort
of adults surviving
ascertained approximately every five days.
2. Different populations maintained under different temperatures (18,
21, 27 and 30°C).
3. Results: As is evident from the figure, the shape of the survivorship
curve is strongly
concave downward, indicative of senescence.
All figures,
unless otherwise indicated, are from M. R. Rose. 1991. Evolutionary
Biology of Aging. Oxford Univ. Press, New York.
4. The steep decline in proportion surviving after a particular age,
which differs for the
different temperature treatments, signifies the "normal" lifespan.
5. Note that life span is greatly decreases as temperature increases.
6. Nevertheless, for a given temperature, there is clearly rapid senescence
once the normal
life span has been achieved.
II. Genetic Variation for Lifespan
A. Strain Differences
1. "Common garden" experiments reveal that different strains of Drosophila
and other
organisms differ in average lifespan.
2. Example: Ganetzky and Flanagan. 1978. On the relationship
between senescence and
age-related changes in two wild-type strains of Drosophila melanogaster.
Exp.
Gerontol. 13: 189-196.
a. Used two inbred lab strains of D. melanogaster (Canton-S and Oregon-R).
b. Estimated adult survivorship curves under standardized laboratory
conditions
c. Senescence begins at substantially eariler age in the Canton-S strain
B. Selection for altered lifespan
1. A response to artificial selection on lifespan indicates
a. presence of genetic variation for longevity
b. capability of life-span to evolve.
2. Example: selection for prolonged lifespan in Drosophila
subobscura (Wattiaux, J. M.
1968. Cumulative parental effects in Drosophila subobscura.
Evolution 22: 406-421.)
a. Established laboratory populations and imoposed one of two treatments
on
each population.
b. In both treatments, males and females were allowed to mate soon after
emergence.
c. Control treatment: Eggs were collected from females approximately
one week after
they emerged as adults, and these eggs were used to found the next generation.
d. Experimental treatment: Eggs were collected from females that
had survived for six
to eight weeks, and these eggs were used to found the next generation.
e. Because there was substantial adult mortality before six weeks, the
experimental
treatment effectively imposed selection for greater female longevity.
f. After a number of generations of selection, the survivorship
curve was estimated
for the experimental and control populations.
g. Results: for both males and females, average longevity was
greater in the
experimental populations.
h. Similar results have been obtained with D. pseudoobscura, D. melanogaster,
and
the flour beetle Tribolium castaneum.
i. Conclusion: Evidence from experiments on four species indicates
that natural
populations are capable of evolving greater longevity if the appropriate selection
is applied. In otherwords, these populations show the possibility of
evolutionarily
postponing senescence.
III. The Issue
A. Results of these genetic analyses of senescence raise a fundamental
evolutionary question: if species
harbor genetic variants for which senescence is postponed, why has natural
selection not caused
these variants to become fixed? i.e., why have organisms not evolved
longer life spans.
1. A priori, the evolution of prolonged senescence might be expected:
individuals that live longer
would be expected to produce more offspring in their lifetime, and thus pass
on more genes
conferring longer life span to the next generation than individuals that do
not live as long.
2. Clearly, however, this has not occurred, at least not in the four
species discussed above, and
probably in many other species, including humans.
3. And just as clearly, the answer to this question lies in some evolutionary
process or processes
that act to prevent the evolution of greater longevity.
4. In the sections below, we will examine what those processes are thought to be.
5. In particular, we will consider two evolutionary theories of aging:
a. The mutation accumulation theory, and
b. The antagonistic-pleiotropy theory.
6. Before doing so, however, we must examine a basic evolutionary principle
that underlies both
of these theories: the intensity of selection acting on a trait expressed
late in life is expected
to be lower than the intensity of selection acting on a trait expressed early
in life.
IV. Age and the Intensity of Selection A. Restatement of the principle
1. Consider an gene the effects of which are expressed only at a given age .
2. Then, if this allele increases the probability of survival over a
fixed time interval by a given
amount, the intensity of selection favoring that allele will decrease as the
age at which
its effects are expressed increases.
B. Illustration of principle.
1. First, consider a population fixed for allele A at locus A that has the following characteristics:
a. The probability of an adult surviving to age is given by the following table:
Age
1 2
3 4
5 6
l(x) .7
.49 .34 .24
.17 0
p(x) .7
.7 .7
.7 .7
0
(Note: p(x) = l(x) / l(x-1) )
b. The expected number of offspring produced at a given age is the same
at all ages;
designate this number by k.
2. The average number of offspring produced by an indvidal in its lifetime is then
mAA = .7k + .49k + .34k + .24k + .17k = 1.94k
3. This is by definition the fitness of the average individual.
4. Now, assume there arises a mutation a that has the following property:
it increases the
probability of survival p(x) of aa and Aa individuals
by 10 percent in at one age.
5. Let us calculate the selection coefficient favoring this allele for different ages of expression.
6. First, assume that the allele's effects are expressed at age 1.
The allele thus causes some
advantageous character to occur at an early age.
a. This effect increases p(1) from 0.7 to 0.77.
b. The remaining values are unchanged, but the change in causes
to
change for all ages, giving the following values of the survivorship parameters
for
the new genotypes:
Age
1 2
3 4
5 6
l(x) .77
.54 .38 .26
.18 0
p(x) .77 .7
.7 .7
.7 0
c. Calculating, as before (see 2. above) the average number of offspring
produced by
individuals of these genotypes gives
mAa = maa = .77k + .54k +
.38k + .26k + .18k = 2.13k
d. The relative fitnesses of the three genotypes, relativized to the
fitness of the Aa
genotype, are then
WAA = mAA / mAa = 1.94k / 2.13k = 0.909
WAa = mAa / mAa = 2.13k / 2.13k = 1
Waa = maa / mAa = 2.13k / 2.13k = 1
e. The selection coefficient is thus WAa - WAA
= 1 - 0.909 = 0.091
7. Next, let us assume that the effects of allele a on fitness
are expressed only very late
in life, at age 5.
a. This means that p(5) is increased by 10 percent, from
0.7 to 0.77, while the other p(x)
are not affected.
b. This change in turn means that l(5) changes from 0.17 to 0.18,
while the other l(x)
are not affected.
c. The values of p(x) and l(x) for the new genotypes are
then
Age
1 2
3 4
5 6
l(x) .7
.49 .34 .24
.18 0
p(x) .7
.7 .7
.7 .77
0
d. Calculating, as before (see 2. above) the average number of offspring
produced by
individuals of these genotypes gives
mAA = .7k + .49k + .34k + .24k + .18k = 1.95k
e. The relative fitnesses of the three genotypes, relativized to the
fitness of the Aa
genotype, are then
WAA = mAA / mAa = 1.94k / 1.95k = 0.991
WAa = mAa / mAa = 1.95k / 1.95k = 1
Waa = maa / mAa
= 1.95k / 1.95k = 1
f. The selection coefficient is thus WAa - WAA
= 1 - 0.991 = 0.009
g. This selection coefficient is much smaller than the value of .091
when the a allele acts early
in the life span, i.e. selection favoring the mutant allele is weaker when
the effects
of that allele occur late in life, even if the effect of the allele on age-specific
survival
is the same magnitude.
8. Intuitive explanation of this result:
a. when a mutation affects survival late in life, most individuals are
dead before the
effects of the allele are expressed, and therefore a beneficial allele
can only increase
the fitness of a small fraction of the individuals that carry it.
b. By contrast, when a mutation affects survival early in life, most
individuals carrying
the mutant allele are alive when the allele is expressed and thus benefit
from having
that allele.
V. Evolutionary Theories of Senescence
A. Two general theoretical explanations for the evolution of senescence:
1. Antagonistic pleiotropy theory
2. Mutation acumulation theory
B. Antagonistic pleiotropy theory (Due to Medawar and Williams)
1. Explanation
a. Considers deleterious mutations that have effects in both young and
old individuals.
b. Three types of such mutations:
i. Those that are deleterious when young, advantageous when old.
ii. Those that are deleterious when young and deleterious when old
iii. Those that are advantageious when young but deleterious when old.
c. Of course, timing of expression of deleterious effects is continuous,
ranging from
very young through intermediate ages, to very old, but for conceptual purposes
we divide this continuum into "young" and "old".
d. Because of the relationship between age of expression and intensity
of seleciton
described in previous section, mutations with deleterious effects expressed
early
but advantageous effects expressed late (category i.) will be selected against
because the selection against the allele due to its early deleterious effects
will
be much stronger than, and overwhelm, selection for the allele due to its
late
advantageious effects.
e. They will therefore not become fixed and will persist at very low
frequencies
reflecting a mutation-selection balance.
f. Similarly, mutations with deleterious effects both early and
late (category ii) will
also tend to be eliminated and held at low frequencies by selection, which
opposes
these mutations at all life stages.
g. By contrast, the relationship between age of express and intensity
of selection
implies that mutaitons with advantageous effects early in life, but deleterious
effects
late in life (e.g. category iii above) will tend to be favored by selection
because
the intensity of positive selection early in life is greater than the intensity
of
negative selection late in life.
h. These mutations will accumulate, and their deleterious effects late
in life will cause
late-life fitness (prob. of survival per unit time or age-specific fecundity)
to decrease,
i.e. their accumulation will lead to senescence.
2. Experimental evidence for antagonistic pleiotropy--genetic correlations
a. One corollary of the antagonistic pleiotropy theory is that at least
some category iii
alleles will evolve to high frequency but will not become fixed.
b. Variation at such loci should give rise to negative genetic correlations
between
longevity and early-life fitness: genotypes with, say, high early fecundity
should
exhibit short longevity, while genotypes with low early fecundity should have
greater longevity.
c. Quite a few studies have attempted to detect such negative genetic
correlations, with
some successes. In particular, the body of literature on Drosophila
in general
supports the existence of these types of negative genetic correlations.
d. Example: Rose and Charlesworth. 1981. Genetics 97: 173-186.
Additive genetic correlation between early fecundity (days 1-5 of adult life)
and longevity = -1.43.
(NOTE: The correlation is greater than 1 in absolute value than because
it is an
estimate of the true correlation. While the true correlation can not
be
greater than 1 in absolute value, estimates of the true correlation may
be because of errors of measurement. The value -1.43 indicates a strong
negative genetic correlation approaching a true correlation of -1.)
3. Experimental evidence for antagonisitc pleiotropy--selection experiments
a. Because of the expected negative genetic correlation between longevity
and early
fitness described above, selection for increased longevity should cause a
correlated
decline in early fitness.
b. Experiment: Rose. 1984. Evolution 38: 1004-1010.
i. Performed selection experbed above.
ii. One set of population selected for late reproduction, and hence for
increased longevity.
iii. Another set of populations was a control with no selection imposed.
iv. At end of selection experiment, measured age-specific fecundity
of
females in the two treatments.
v. Results--found that lines selected for increased longevity showed
both
increased longevity and reduction in early fecundity (days 1-10 of
adult life:
4. Conclusion: Both approaches indicate that at least some portion
of senescence in Drosophila
is explained by antagonistic pleiotropy.
C. Mutation acumulation theory. (Due to Medawar)
1. Explanation
a. Assumes that there are at least three types of deleterious mutations:
i. Those for which deleterious effects are restricted to young individuals
ii. Those for which deleterious effects are restricted to old individuals
iii. Those for which deleterious effects are expressed in both young
and old
old individuals.
b. Of course, timing of expression of deleterious effects is continuous,
ranging from
very young through intermediate ages, to very old, but for conceptual purposes
we divide this continuum into "young" and "old".
c. Because of the relationship between age of expression and intensity
of seleciton
described in previous section, deleterious mutations expressed early (e.g.
categories
i. and iii. in a. above will be strongly selected against.
d. They will therefore not become fixed and will persist at very low
frequencies
reflecting a mutation-selection balance.
e. By contrast, the relationship between age of express and intensity
of selection
implies that deleterious mutations expressed late in life (e.g. category ii
above)
will be only weakly selected against.
f. Such mutations are reasonably likely to be fixed by genetic drift
and, if not
fixed, to occur at fairly high frequencies.
g. These mutations, because they are detrimental, cause a decline in
probability of
survival in old individuals, i.e. senescence.
2. Evidence for mutation accumulation: increased genetic variation in
age-specific fitness with
age.
a. One expectation of the mutation accumulation theory is that genetic
variation for
age-specific fecundity or survival should be virtually absent at young ages.
b. This expectation arises because the selection intensity against mutations
with
deleterious effects on fecundity or survival is expected to be high, and thus
such mutations are kept at very low frequencies, producing very little genetic
variation.
c. By contrast, mutations that cause deleterious effects late
in life are not strongly
selected against and may achieve relatively high frequences, producing substantial
genetic variation for late fecundity and/or survival.
d. Thus, one pattern predicted by the mutation accumulation theory is
that genetic
variation for age-specific fitness should increase with age.
e. Experimental test: Rose and Charlesworth 1981.
i. Measured additive genetic variance for daily egg production, starting
at age of peak fecundity.
ii. Results:
No detectable change in geneic variance as females age:
3. Experimental evidence for mutation accumulation: selection experiments.
a. Logic and protocol of the experiment:
b. Select for late reproduction and extended longevity; this should
have the
effect of greatly decreasing the frequencies of alleles causing senescence.
c. Then reverse selection by selecting for early fecundity and survival.
d. Two possible outcomes of reversed selection:
i. Results in decrease in late fecundity and/or a decrease in longevity.
ii. Results in no decrease in late fecundity or longevity.
e. Interpretation of outcomes:
i. Outcome i. indicates alleles with deleterious effects on late
fitness have
advantageous effects on early fitness and vice versa--indicates antagonistic
pleiotropy explains senescence
ii. Outcome ii. indicates that alleles with deleterious effects on late
fitness
do not have advantageous effects on early fitness--indicates mutation
accumulation contributes to senescence.
f. Example: Mueller, L. D. 1987. Proc. Nat. Acad. Sci. 84: 1974-1977.
i. When applied selection for early fecundity in D. melanogaster,
did
not get loss of late fecundity initially (first 30-40 generations), indicating
that for this fitness
component in the population examined, there was no antagonistic pleiotropy.
ii. However, after 120 generations, late fecundity had declined substantially,
suggesting that senescence
in this laboratory population waisdue largely to mutation accumulation.
D. Conclusions
1. Analysis of genetic variation for senescence in Drosophila
indicates that both antagonistic
pleiotropy and mutation accumulation contribute to senescence.
2. Little information is available for other oragnisms.
IV. Implications of Evolutionary Theories of Senescence
A. The major cause of senescence is the decreasing strength of selection
as organisms age, which allows
mutations causing reduced late fitness to increase in frequency or even become
fixed. (Note that this
is true regardless of whether antagonistic pleiotropy or mutation accumulation
applies.)
B. Implies that there will be no "general cause" of senescence that
acts across species, or even across
populations.
1. In any species or population the exact causes of senescence will
depend on what mutations
having deleterious effects on late fitness increase in frequency, and what
characters those
mutations affect.
2. Also implies that senescence in a particular organism may have multiple
causes, representing
different characters affected by different deleterious mutations contributing
to senescence.
C. Suggests that studying the "cause" of senescence in non-human animals
will not be informative
directly about the causes of senescence in humans, and thus will not provide
direct information on
the types of medical intervention that will be needed to eliminate senescence
in humans.
D. Nevertheless, exploration of the causes of senescence in other organisms
may indirectly benefit
research in human gerontology by leading to the development of a robust research
protocol that
can be applied to the study of human senescence.
E. Hope for a "magic bullet" cure of senescence, however, are slim,
because evolutionary considerations
suggest that once a cure is developed for one senescence factor and lifespan
is extended somewhat,
other senescence factors will kick in to limit lifespan.
