Early middle age is a difficult time, not least for male fruitflies when sperm production falls. The unexpected reason for this decline seems to be that, as tissues age, maintaining functional stem cells becomes difficult.
Stem cells arose early in evolution and underlie the body plans of most multicellular animals. So we tend to picture stem cells and their niches — the microenvironments in which they operate — as finely tuned systems with near perfect, almost magical, capabilities. A report by Yamashita and colleagues (Cheng et al., page 599 of this issue)1 reminds us, however, that stem cells are no exception to François Jacob's description2 of evolution as “tinkering”.
The subject of the Yamashita group's studies is the germline stem cells (GSCs) in the testis of adult male fruitflies (Drosophila melanogaster). The authors discovered that by day 20 of adulthood — early middle age for a fruitfly — a significant fraction of GSCs had become arrested; that is, they had ceased the cell-cycle activity and asymmetric cell division that lead to each GSC producing one self-renewing cell and eventually one differentiated sperm cell. This slows sperm production without changing stem-cell number.
Normally, GSCs align one of their two centrosomes — key organelles controlling the direction of division — with the niche, which in this case is the periphery of a cluster of support cells called the hub (Fig. 1a). This arrangement ensures that, on division, one daughter remains a stem cell while the other one differentiates. Cheng et al.1 found that centrosomes in the arrested stem cells had become misaligned. Remarkably, it seems that centrosome misalignment is a by-product of the dedifferentiation of older germ cells that enter the niche to replace GSCs that have turned over; dedifferentiation randomizes centrosome orientation. These observations provide a much clearer picture of the ageing Drosophila testis, and hint at processes that may limit cell production in many other tissues: stem-cell homeostatic mechanisms are not perfect, and cannot maintain full activity even into middle age.
A case in point is the signature ability of stem cells to divide asymmetrically — to self-renew. Centrosome alignment seems to be a highly adaptive mechanism that acts to ensure self-renewal by limiting stem-cell loss or over-proliferation3. Keeping one centrosome, the maternal centrosome4, permanently localized within the cell cytoplasm adjacent to the boundary between the GSCs and the niche-generating hub ensures that divisions will be oriented perpendicular to the niche (Fig. 1a). Because the niche microenvironment that maintains the stem-cell state exists only in a one-cell-wide strip around the hub, such an orientation guarantees that just one GSC and one differentiating germ cell will be produced by each division (Fig. 1b). This elegant system, or a variant of it, seems to be used in other stem cells, such as Drosophila neuroblasts5,6. It may have evolved from a fundamental molecular asymmetry in all dividing cells — the non-equivalence of newly duplicated centrosomes — and hence may be an ancient stem-cell mechanism.
We must now ponder the fact that, despite the operation of this system, many GSCs are lost with age anyway and have to be replaced. Moreover, replacement cells seem to have difficulty realigning with the niche, and during realignment they cannot function as GSCs (Fig. 1c). However, it is still too early to blame these problems on centrosome misalignment as such. Age-related changes in niche adhesion or signal reception could underlie stem-cell loss, and dedifferentiation itself might be the slow step in replacement and simply be a prerequisite to alignment. Nonetheless, the system as a whole comes under stress by early middle age, and problems only get worse in older flies, in which niche signals begin to fail as well7.
A major contribution of this work1 is the finding that new GSCs are regularly produced by dedifferentiation of downstream germ cells (Fig. 1d). The GSC daughter that leaves the niche divides four times, but incompletely, to generate 2-, 4-, 8- and 16-cell interconnected germline 'cysts'. It was already known from lineage marking that male GSCs turn over and are replaced at a moderate rate8, but the replacement of stem cells by the dedifferentiation of germ cells up to the 8-cell stage had been observed only under experimental conditions9,10.
The data of Cheng et al. suggest that, in an old male, all the stem cells that started out in the niche are likely to have undergone replacement by a dedifferentiated cell at least once. Consequently, we now know that GSC activity in male Drosophila resides not just in the GSCs themselves, but also in the pool of 2-, 4- and 8-cell germline cysts. This is very similar to a long-standing but controversial model of germ-cell production in the mammalian testes. According to this model, sperm production can be sustained by both individual stem cells and a pool of 'potential stem cells' that correspond to the 2-, 4-, 8- and 16-cell germline cysts, which in mice are called A-type spermatogonia11,12. The discovery that dedifferentiation and GSC replacement occur regularly in the normal Drosophila testis opens these fascinating processes up to detailed mechanistic study in an accessible system.
How general are these discoveries likely to be? We already know that many other stem cells depend on niches, undergo turnover and replacement, and decline with age13. But do they do so in the context of a growing pool of arrested stem cells? Similar cells have not been described during ageing in the Drosophila female germ line14. In the ovary, separate small niches, each supporting just two or three GSCs, are isolated within each of 16 ovarioles. This cellular organization probably reduces the opportunities for contact between niches and downstream germ cells. In addition, replacement seems to rely on changing the direction of GSC division, to keep both daughters in the niche, rather than on dedifferentiation15.
But there are many other systems that are compatible with the possibility of such changes. One attractive idea, particularly in large, long-lived animals such as mammals, is that a decline in stem-cell activity with age is actively programmed in order to contain the escalating probability of tumour initiation13. It is hard to see how this can be a significant factor in Drosophila, however, which is too small and short-lived to suffer significantly from terminal cancer. Short of finding some currently unknown genetic or environmental factors that put Yamashita and colleagues' flies1 under stress, we are left with the possibility that stem-cell maintenance is simply less perfect than we might have imagined. Why entrust something as essential as sperm production to a makeshift system that breaks down regularly? “Why indeed?” we hear Jacob asking us. “It works.”