1.2. INCREASED RESISTANCE TO STRESS, ASSURES LONGEVITY
"Gerontogenes" (genes that affect the rate of aging) (40) can be denned operationally to refer to genes that can be altered such that a longer than normal maximum lifespan is the result (41). The last two decades of research in aging have demonstrated overwhelmingly that gerontogenes exist and modulate the rate of aging. The first direct demonstration that genes play a role in the aging process was carried out in the nematode Caenorhabditis elegans (42). Despite original prejudices, that the aging process is "ineluctable" (43) or that genes controlling longevity cannot be selected for (44), these results and others have shown that the process of aging, just as other biological processes, is specified by the gene. This is not to say that aging is programmed. Statements by noted developmental biologists that aging must be programmed (45) to prevent competition with offspring are untenable for the nematode C. elegans,
which has billions of descendants by the time its hypothetical "death program" kicks in to kill it (46).
1.2.1. THE EXISTENCE OF GERONTOGENES
A. Gerontogenes affect mortality rate. Johnson, working with recombinant-inbred (RI) strains of C. elegans, demonstrated that genes significantly affect length of life (42). Both life expectancy (mean lifespan at birth) and maximum lifespans were lengthened by up to 65% and twofold, respectively, in some of the longest-lived strains, over the longevity of their parental strains - two common wild-types, N2 and Bergerac (48). These lifeextensions were modulated by overall changes in age-specific mortality rates. Whenever one assesses age-specific mortality in wild-type strains, mortality rates are observed to increase exponentially with age, especially when population size is relatively small, <200 worms or so (48, 49). In the RI strains, this exponential rate of increase in mortality rate is reduced from doubling every 4 days to doubling about every 10 days. Initial mortality rate at age of maturity remained essentially unchanged (48).
Note that limitations of small populations lead to the failure to detect a late-life leveling off of age-specific mortality that can only be observed in large populations ofnematodes (50-52). This issue will be addressed again below in the discussion of major gerontogene mutants.
B. Gerontogenes do not affect development. Subsequent studies on these RI strains showed that longevity is determined independently of other life-history traits such as length of development, length of the reproductive period or total reproductive effort (48). In these initial studies we could only detect phenotypic changes resulting from the simultaneous effect of the many genes affecting longevity and other life-history traits. With the advent of the Human Genome Project and its worm equivalent, it became possible to assess the effect of genes throughout the nematode genome. Subsequent studies on the RI strains (46, 53) took advantage of these resources to localize gerontogenes and to determine whether genes affecting length of life also affected other life history traits; they did not.
Moreover, unlike development, where new genes are turned on and off at each developmental stage, there are much more subtle changes in gene activity during aging (54). It seems clear that the simplistic concept that aging is an extension of development (45) is unwarranted; a larger discussion of these issues can be found in several books and reviews (55-57).
C. Gerontogenes may modulate overall health status. Long-lived RI strains move more even late in life (48). Indeed, the rate of loss of movement in these long-lived strains serves as an effective predictor of when any particular strain will die (48). Similar effects are observed in gerontogene mutants such as age-1, where mutant worms maintain vigorous movement well past the time when most of their wild-type controls have already died. Total lifetime movement of these gerontogene mutants is greatly increased over that observed in the wild type (58). The cause of nematode death is unclear but a typical worm loses the ability to move and then to respond to touch about a day before overt physical degeneration is observed. Dead worms are rapidly invaded by bacteria through the mouth, anus and vulva and decay of the entire worm can occur in only a matter of a few hours. At least one long-lived variant (age-1) is much more resistant to this bacterial invasion (59).
1.2.2. MAJOR GERONTOGENE MUTANTS
A number of gerontogene mutants that show lengthened life have been identified over the last several years. These genes lengthen life from just barely detectable amounts under certain conditions (rad-8, clk-2, clk-3, gro-1, spe-26) (60) to almost two-fold (age-1 and daf-2). Those showing the largest changes are also involved in dauer formation. Dauers are an alternative larval development stage leading to an extended diapause; dauers are also a dispersive stage that fills an important role in the maintainance of wild populations of C. elegans (61). In the lab, dauers are formed when food is exhausted or when nematode density is high.
There is a major cost associated with the extended longevity of all the gerontoene mutants so far detected. All of them affect some aspect of normal development (for example, making dauers under abnormal conditions). As will be seen below, all gerontogene mutant studies show elevated resistance to environmental stresses and we suggest that this is the major mechanism leading to life extension and extended health (Fig. 1.1).
Fig. 1.1. Possible interactions between Age mutants (including tkr-1} in regulating the stress respom daf-2, which has homology to the insulin receptor, has been proposed to activate age-1, a putati PIS kinase, which in turn negatively regulates daf-16, a putative forkhead transcription factor (3' TKR-1 also acts in concert with daf-16; this may occur by directly activating daf-16, downregulating the daf-2/age-1 pathway, or by being a downstream effector of daf-16 (Murakami & Johnson, preparation). Other signaling components may also be involved in regulating daf-16, or the transcriptional response to stress in general.
A. Molecular action of gerontogenes age-1. The first gerontogene mutant identified was age-1, it is the only mutant yet studied that was identified by its longevity phenotype alone (62, 63). The age-l(hx546) reference allele has a life expectancy 70% longer than wild type and a maximum lifespan that is 105% longer (49, 64). age-1 mutations have little effect on fertility, length of reproduction or rate of development (41, 63-66) but are dauer constitutive (Daf-c) at the semi-lethal temperature of 27°C (67, 68). age-1 (hx546) is resistant to H+ (69), paraquat (70), and UV (71), is thermotolerant (72, 73), and has reduced frequencies of deletions in mitochondrial DNA (74). Three other alleles of age-1 were independently isolated on the basis of longevity alone (63) and are also stress resistant (Str) but show subtle variations among themselves.
Three more alleles of age-1 have been isolated by selecting for increased thermotolerance (G. Lithgow, personal communication). All new age-1 alleles are also Daf-c at 27°C.
daf-23. Mutations in another gene, previously called daf-23, also lead to extended longevity and increased resistance to stress (75). daf-23 mutants are daf at 25°C and age-1 mutations fail to complement daf-23 mutations (63, 67, 64 Morris et al. (68) have cloned the daf-23 locus and demonstrated that it shows structural homology with mammalian phosphatidylinositol-3-OH kinase (PI3I and have also suggested that daf-23 and age-1 are the same gene, based on this failure to complement.
Morris et al. (68) failed to demonstrate any mutation in PI3K for the age-1 (hx546) reference allele. Murakami, Kliminskaya and Johnson have complete sequencing of the three alleles (zIO, zll, z25) identified (63) and found no sequence changes in either exon or intron regions of PI3K. On the basis of these results, the formal possibility exists that the age-1 gene is not the same gene as daf-23 and does not encode PI3K. (Non-allelic failure to complement has been observed both in the C elegans spe-6 gene (76) and in other species; for example, in the mouse HoxB cluster (77), "non-allelic non-complementation suggests that these two genes function together ..."). However, we will assume for now that PI3K is the protein encoded by the age-1 gene.
daf-2. Mutants in daf-2 are Daf-c at 25°C and result in a more than two-fold extension of life expectancy in the adult phase (45). daf-2 interacts with daf-12 to cause an almost four-fold increase in life expectancy (75). daf-2 bears structural homology to the human insulin receptor (78).
Other Age mutants. There are other examples of genes in C. elegans in which mutations result in longer adult life expectancy and maximum lifespan. In all of these cases the Age phenotype results from hypomorphic or nullomorphic mutations:
(40) Mutations in rad-8 extend overall longevity, although most of the extension is from slowed development (79).
(41) Two of the six mutant alleles of spe-26, a gene specifying proper segregation of cellular components affecting sperm activation, result in life extensions of about 80% for the hermaphrodite and the mated male (71, 80), although the details are contentious (81).
(42) Wong et al. (82) reported that two of four alleles of clk-1, both of which have altered cell cycle and developmental timing, also have increased life expectancy. (A close examination of their data shows that much of the extension results from slowed development.) The clk-1 gene encodes a 187-amino acid protein showing some homology with yeast COQ7, a gene required for synthesis of coen-zyme Q (83); they suggest that clk-1 mutant alleles have reduced metabolic rates that are responsible for the extended longevity. Lakowski & Hekimi (84) have extended these studies to include clk-2, clk-3, and gro-1, all of which have modest (typically 20-30%) extensions of lifespan, but only for some alleles. There are several additional gerontogene mutants that complement all known Age genes (63, 85), suggesting that there may be many unidentified gerontogenes in C. elegans.
tkr-1. The genes just described are hypo- or nullomorphic (reduced levels of function). Another class of gerontogene has been identified; this class, represented by tkr-1, is longevous when the gene is overexpressed. Upregulation of tkr-1 leads to a life extension averaging about 65% (86). This gene encodes a tyrosine kinase that responds to stress and may modulate the general cellular response to environmental stress (86).
daf-16. The Age phenotype oi age-1, daf-2, spe-26, clk-1 and tkr-1 is suppressed by mutant alleles of daf-16 (71, 75, 78, 87), a gene necessary for dauer formation. These results show that the Age phenotype of these mutants is dependent on a functional daf-16 protein and suggest a possible involvement of the dauer formation pathway in life extension (78, 87). The daf-16 gene product appears to be a transcription factor homologous to the Forkhead family of transcription factors (88, 89).
B. Role of stress resistance in specifying gerontogene action. All of the gerontogene mutants in C. elegans that have been tested are more resistant to environmental stresses such as reactive oxygen species (ROS) (69, 70), high temperature (72, 73) and UV radiation (71). Increased thermotolerance is seen in all of the life-extension mutants tested; however, daf-4 and daf-7 mutants (Daf-c) are thermotolerant but do not show Age phenotypes (73). Of the three stressors, UV is the most consistent predictor of life extension (71). The increased UV resistance of age-l/(daf-23), daf-2, spe-26 and clk-1 is suppressed by daf-16 (70) as are all other stress-resistance traits tested. Finally, tkr-1 also seems to mediate increased resistance to numerous stresses as well. These results are summarized in Fig. 2.
C. Effect of gerontogenes on health and survival. The questions about the relationship between development and aging and the extended health span of these mutants have been most effectively evaluated for age-1. Like the RI strains, age-1 has normal development and overall reproduction as well as length of reproduction (64-66). Thus, despite earlier suggestions to the contrary (64, 65), there appears to be little reason to suggest that there is any trade-off between reproduction and survival in the age-1 mutants.
1.2.3. GERONTOGENES IN THE EVOLUTION OF SENESCENCE
Evolutionary theory suggests that gerontogenes have not evolved to have an adaptive function but rather as a by-product of other adaptive responses or due to the lack of selection on late-life events (90, 91). All of the data in C. elegans are consistent with this view. All of the major gerontogene mutants have alterations in other aspects of development (especially dauer formation) that reduce evolutionary fitness. Even age-1 (the only gene selected solely for its life extension characteristics) has an altered ability to form dauers at high temperature, resulting in reduced competitive ability with the wild type (G. Lithgow, personal communication). In contrast, little evidence for trade-offs was detected in the RI strains.
Fig. 1.2 Summary of mean lifespan and UV resistance of Age mutants. Observations are presented as Age mutant/wild type ratios, as data are compiled from independent studies. Typical data are shown. Life expectancies were taken from original publications as follows: age-1(41), elk-1(82), daf-2 (75), spe-26 (71), and tkr-1 (86). All the UV-resistance data are from Ref. 86 or Murakami & Johnson, in preparation. The alleles used here are age-l(hx546), clk-l(e2519), daf-2 (el 370), spe-26 (he 138), and tkr-l(zls3000).
1.2.4. FUTURE OF AGING RESEARCH
The gerontogene mutants in C. elegans have elicited great interest. But the answer to the question of whether similar genes exist in mammals remains unresolved. The relationship between the increased response to stress observed in the gerontogene mutants and that seen in dietary-restricted (DR) rodents is consistent with an interpretation that the increased longevity due to DR may result from an increased resistance to environmental stresses. Although human longevity has only a small genetic component (92), there is reason to be hopeful that major gerontogenes exist in mammals and humans. One of the best bits of data relate to the extended longevity of Mm. Calment and her near relatives, which is consistent with the effects of a major recessive gerontogene that by chance were present in Mm. Calment and her near relatives.