The invertebrate model organisms are a fertile hunting ground for new discoveries about aging. They are valuable because of their short lifespans, and because powerful methods have been developed to conduct genomic screens and to test specific molecular hypotheses. However, because the worm Caenorhabditis elegans and the fruit fly Drosophila melanogaster are generally believed to be post-mitotic as adults, they have been assumed to be poor models for aging in the dividing tissues of mammals. Not so. Stem cells in Drosophila are teaching us much about those of mammals, including their role in cancer (Nystul & Spradling, 2006; Fuller & Spradling, 2007; Gonzalez, 2007; Rebollo et al., 2007). Aging of stem cells and their somatic niches in the gonad of adult Drosophila have recently been investigated as a cause of loss of fecundity with age (Wallenfang et al., 2006). Somatic stem cells have also recently been discovered in the mid-gut of the adult fly (Micchelli & Perrimon, 2006; Ohlstein & Spradling, 2006, 2007), opening the way to the use of Drosophila to understand aging of these cells.
The last 18 months have seen some important theoretical work on the evolution of asymmetrical cell division, a key step towards the evolution of aging itself. For aging to evolve, some entity, be it a single cell or a multicellular organism, must experience an age-related decline in the force of natural selection. An organism that reproduces by perfectly symmetrical division will not be subject to this decline, because death of any individual at any time will result in loss of 100% of its reproduction; it becomes an evolutionary dead end. In contrast, if division is asymmetrical, into a parent and an offspring, then death of the parent after production of progeny has started does not cause total loss, and hence extrinsic mortality can cause the force of natural selection to decline with age. It is thus understandable that aging occurs in single-celled organisms that divide asymmetrically, such as budding yeast Saccharomyces cerevisiae (Kaeberlein et al., 2007) and the bacterium Caulobacter (Ackermann et al., 2003), despite the absence of a distinction between soma and the germ line in these creatures. Aging has been observed even in single-celled organisms with ostensibly symmetrical cell division, for instance Escherichia coli (Stewart et al., 2005; Fredriksson & Nystrom, 2006) and fission yeast Schizosaccharomyces pombe (Barker & Walmsley, 1999), possibly associated with asymmetrical distribution of damaged proteins, as has been demonstrated in budding yeast (Aguilaniu et al., 2003) and, recently, in mammalian embryonic stem cells (Hernebring et al., 2006). This type of asymmetrical division of damage is rejuvenating for the progeny, but at the expense of the parent, which retains the unrepaired damage. Stimulated by these findings, a recent study (Ackermann et al., 2007) investigated whether there might be some evolutionary advantage to asymmetrical cell division of damage in single-celled organisms. The study assumed that damage accumulates with time, impairs survival or fecundity, can be repaired but at a cost, and can be distributed symmetrically or asymmetrically between offsprings. Asymmetry created variation in the level of damage between progeny, and hence resulted in a survival or fertility advantage for individuals that went on to produce less damaged progeny, resulting in a higher growth rate for the lineage. Interestingly, once asymmetrical division evolved in the model, it became advantageous to reduce the level of repair, because damage was now eliminated by the asymmetrical division. The cellular economics of the costs of asymmetrical division vs. those of repair may therefore have been fundamental in setting the scene for the evolution of aging, and continues to be of key importance in maintaining the integrity of the germ line.
Dietary restriction (DR) extends lifespan of diverse organisms including yeast, worms and flies as well as many species of mammals. The precise mechanisms involved have not been worked out for any of these organisms. It is therefore not clear if this is a case of evolutionary conservation or, instead, of evolutionary convergence; time will tell. Recent results from C. elegans and Drosophila imply that sensory perception, specifically chemosensation, may be an important first step in the detection of nutrient status during DR. A genome-wide RNA expression profile of the response to DR in Drosophila had previously shown up-regulation of expression of odorant binding genes. A new study (Libert et al., 2007) has shown that simply exposing dietarily restricted flies to the odour of yeast was sufficient to reduce their lifespan, although not as much as did feeding them on yeast paste. Interestingly, there was no parallel reduction of fecundity, as is generally seen with DR itself, implying at least partially separate control of the two responses. Mutation of a gene encoding an odorant-binding protein (Or83b) also greatly increased lifespan and attenuated the response to DR, and disruption of G protein function in the neurons that expressed Or83b had similar effects. Interestingly, ablation of amphid, sensory neurons has previously been shown to extend lifespan of C. elegans, and, recently, alteration of G protein signalling in two different types of amphid neurons has been reported to have similar effects (Lans & Jansen, 2007). It is not yet known if either of these effects are involved in mediating the response to DR in C. elegans. These findings imply that, in these two invertebrates, neural encoding of environmental information is an important early step in determining lifespan and, in the case of Drosophila, the response to DR, and it will be important to identify the downstream mechanisms at work. A possible role for chemosensation in the DR response in rodents would also bear investigation.
Signalling pathways influencing lifespan continue to be elucidated, although we need to learn much more about the types of damage and pathology that are ameliorated by altered signalling activity. Recent findings from C. elegans have also illuminated signalling pathways that are important for the response of lifespan to DR. SKN-1 is a transcription factor orthologous to mammalian NFE2-related factors, and is involved in control of response to oxidative stress. Worms lacking SKN-1 function failed to increase their lifespan in response to DR; they showed normal extension in response to reduced insulin signalling. The absence of SKN-1 in two amphid neurons was critical for blocking the response to DR, and laser ablation of these two neurons also prevented the response. The sensory role of the amphids may not be relevant in this context, because mutants lacking the sensory cilia in these neurons showed a normal DR response (Bishop & Guarente, 2007). In another study, a systematic screen of all 15 of the worm forkhead transcription factors revealed that just one of them, pha-4, orthologous to the mammalian Foxa family, is essential for the response. Again, the effect is specific to extension of lifespan by DR, because insulin pathway mutants produced normal extension of lifespan in pha-4 mutant worms. Interestingly, smk-1, previously shown to be an essential coregulator of daf-16, the key forkhead transcription factor in extension of lifespan by insulin pathway mutants, is also required for extension of lifespan by DR. It is therefore possible that smk-1 is also a cofactor for pha-4 in the response to DR (Panowski et al., 2007). Understanding exactly how these different pathways act together and in different tissues of the worm to coordinate the response to DR will be important future work.
Silent information regulator 2 (Sir2) has been suggested to be essential for extension of lifespan by DR in yeast, C. elegans and Drosophila, implying strong evolutionary conservation of the role of this pathway. However, recent findings have suggested that DR can extend lifespan in yeast independent of all sirtuins (Tsuchiya et al., 2006). In addition, in C. elegans extension of lifespan by overexpression of Sir-2.1 is dependent upon daf-16, while the response to DR is not. Two recent studies of C. elegans have developed a new, and highly effective, method for extension of lifespan by DR: food deprivation. Both studies showed that deletion of Sir2.1 does not affect the increase in lifespan in response to food deprivation (Kaeberlein et al., 2006; Lee et al., 2006). Further work is needed to determine the nature of any role of sirtuins in the response of lifespan to the various experimental paradigms for DR in these invertebrates.
Exciting work has come from the development of powerful invertebrate models to understand human aging-related diseases. Aging is the key risk factor for the major chronic and killer diseases of industrialised societies: cardiovascular disease, cancer and neurodegeneration. It has long been known that DR not only extends lifespan, but also delays or attenuates multiple forms of aging-related pathology in rodents. The normal aging process itself is thus acting as a major risk factor for aging-related disease, and when aging is slowed down by DR, so is the onset of multiple aging-related diseases. This observation points to the prospect of a broad spectrum, preventative medicine for the diseases of aging, a truly inspiring possibility. There is therefore a growing interest in discovering whether single gene mutations that extend lifespan can also ameliorate aging-related pathology. This topic is still in its infancy, but some interesting findings have already emerged. Reduction of insulin and target of rapamycin signalling has been shown to attenuate normal decline in heart function with age in Drosophila (Wessells et al., 2004). Recently, two studies in C. elegans have produced parallel findings for genetic models of cancer and neurodegeneration. Mutations in the gld-1 gene cause overproliferation of germ line cells of C. elegans, and these cells eventually escape from the gonad and kill the animal. The lifespan of long-lived worms with two different mutations of the worm insulin receptor daf-2 were not shortened by mutation of gld-1, and this was a consequence both of increased apoptosis of the cells and decreased cell division of the germ line cells (Pinkston et al., 2006). In a second study, a worm model of Alzheimer's disease was used, in which Aβ1–42, an aggregate-prone peptide implicated in the disease in humans, was expressed in muscle. RNA-interference with the insulin receptor daf-2 extends lifespan of wild-type worms, and it also delayed the paralysis induced by expression of Aβ1–42, again suggesting that slowing the aging process slows disease progression. The neuritic plaques formed by Aβ1–42 are characteristic of Alzheimer's disease, but there is uncertainty as to exactly how neurotoxicity is caused. The results of the study with C. elegans strongly implicated oligomers of Aβ1–42 as the toxic species. The aging process reduced the ability of cells to detoxify small protein aggregates, and the study suggested a mechanism for amelioration of this problem by altered activity of the insulin signalling pathway (Cohen et al., 2006).
Testing the generality of these conclusions will be greatly helped by the recent development of a number of new model systems in which to study health and disease during aging, particularly in Drosophila. About 77% of the genes implicated in specific human diseases have one or more Drosophila homologues (Reiter et al., 2001), and see http://superfly.ucsd.edu/homophila/. Fly models of human diseases have been created by inserting a human gene or mutant versions thereof into the fly, or by mutating the corresponding fly orthologue to correspond to a human disease state. The ability to rapidly do enhancer/suppressor screens and to test drugs in the invertebrates will make them a rich source of candidate interventions into disease over the next several years. Fly models of several neurodegenerative disorders including Alzheimer's disease (Chee et al., 2006; Crowther et al., 2006; Kinghorn et al., 2006) and Parkinson's disease (PD) (Whitworth et al., 2006) have been produced, and recapitulate many of the features of the human disease (for a review see Marsh et al., 2006). Sometimes fly models of human neurodegenerative disease have produced unexpected phenotypes that have turned out to be highly informative. For instance, parkin is an E3 ubiquitin protein ligase, and loss of function mutations in the gene encoding it in humans causes early onset PD. Drosophila has a parkin orthologue. Mutants in this gene cause loss of dopaminergic neurons in the fly brain, which is characteristic of PD in humans. However, the mutants in the fly also cause muscle loss and male sterility, neither of which are present in human PD. These findings could be used to argue that the fly parkin mutants are a poor model for PD. However, mitochondrial defects are a conserved feature of parkin mutants, and the male sterility of parkin mutants is attributable to a mitochondrial defect in sperm that could also be important in muscle loss. These seemingly disparate phenotypes may therefore have commonalities at the biochemical level that are also important in PD (Park et al., 2006; Whitworth et al., 2006; Yang et al., 2006).
The vast majority of studies of aging with invertebrates rely, often almost exclusively, on genetic manipulation. This has been, and will continue to be, a very powerful approach. However, increasingly it is becoming necessary to understand exactly what genetic manipulations are achieving, at the biochemical, cellular, tissue and systemic level. Because the model invertebrates are so small, they are not ideal subjects for biochemical and physiological work, but future studies will increasingly need to include this kind of approach if they are to realize the full potential of what these little creatures have to tell us about aging.