An appreciation for mutations associated with DNA repeats and the impact on human health (Cooper and Krawczak 1993) predates the excitement over the massive expansion associated with many neurodegenerative diseases. The mutations associated with many diseases caused by small changes in repeat length can be easily explained by replication slippage or, in the cases of polyalanine diseases, recombination. Both are classic, long-known mutation mechanisms (Drake 1970). Not unexpected are additional repeat destabilizing effects of mutations in genes involved in DNA replication, repair, and recombination. Though much is known, a remarkable and exciting question remains unanswered: how does a repeat expand to lengths of 1000 to 11 000 copies from an initial length of 100 copies or less during a single intergener-ational transmission? The field may be only slightly closer to understanding this question now, compared with 15 years ago. Early reviews discussed many of the same models for repeat instability presented here, before there was sup portive experimental data. While expandable human repeats are unstable in mice, their behavior does not accurately recapitulate the patterns of intergen-erational transmission seen in humans. Human gametogenesis and early embryogenesis, where expansive instability may occur, are simply not tractable experimental systems. In the absence of a model experimental system that recapitulates intergenerational massive expansion, progress in understanding mechanisms for massive expansion may be slow.
While mysteries remain, much has been learned. One important take-home lesson is that the standard model systems, bacteriophage, bacteria, yeast, mice, and human cells, exhibit different and variable responses to long repeats. Human cells can maintain thousands of d(CGG) ■ d(CCG) and d(CTG) ■ d(CAG) repeats in quite stable fashion, while showing greater instability with d(ATTCT) ■ d(AGAAT) and d(GAA) ■ d(TTC) repeats. Bacterial cells, on the other hand, have great difficulty maintaining several hundred repeats. Different experimental systems exhibit different patterns of repeat instability, and conclusions learned from one system may not always apply to another. One must also keep in mind that during intergenerational transmission in humans a unique "window of opportunity" must exist for expansion from about 100 to thousands of repeats, and that once that window has closed, the repeats become complacent, so to speak. Either the window is missing in other systems, or it cannot be pried open. Therefore, we must continue to utilize model systems, keeping in mind the limitations and implications of each, with the ultimate goal of understanding processes that explain expansion.
A second take-home lesson is that all disease-expanding repeats are unique with their own personalities and peculiarities in terms of alternative DNA conformations (Table 1). Moreover, there is not a simple feature that correlates with expansion. In addition, different repeats can behave differently in a model system. For example, d(CTG) ■ d(CAG) repeats associated with DM1 or d(CGG)n ■ d(CCG)n repeats associated with fragile X syndrome can form slipped mispaired structures (Pearson and Sinden 1996) that may block replication, and they undergo rapid deletion in E. coli (Kang et al. 1995a; Bowater et al. 1996; Ohshima et al. 1996a; Hashem et al. 2002). Conversely, SCA10 d(ATTCT) ■ d(AGAAT) repeats do not form a structure that can block replication, but rather may support replication in human cells, and they are quite stable in E. coli at lengths at which d(CTG) or d(CGG) repeats are very unstable. The point to be made here is that a single pathway for massive expansion may not exist, although it cannot be presently excluded. Likewise, multiple pathways exist for the small changes in repeat length observed in somatic cells throughout life. Alternative DNA conformations associated with certain repeats are probably very important for repeat instability in some pathways; however, they may be less important for other repeats.
In summary, repeat instability remains a major problem for human health and no simple mechanism or biochemical pathway may direct massive expansion for all repeats. Moreover, given the interdependence of replication, repair, and recombination, under the global regulation and coordination of checkpoint control, many players and pathways will be expected to have an influence on repeat instability. Maybe in the last 15 years we have learned enough to know where to begin to address in new ways this complex biological phenomenon.
An additional goal is to learn how to manipulate repeat length in a therapeutic fashion to delay or prevent disease-causing expansion, or to reverse the expansion process, preventing or alleviating the genetic source of the problem. Initial investigations related to this question have recently been described (Gorbunova et al. 2003; Yang et al. 2003; Pineiro et al. 2003; Gomes-Pereira and Monckton 2004a; Hashem et al. 2004a).
Acknowledgements We thank Albino Bacolla, John J. Bissler, Sharon F. Edwards, and Michael Leffak for critically reading the manuscript. We thank Alan Maness for assistance in preparation of the figures. Work in the authors' laboratories is supported by NIH grants ES05508 to R.R.S. and CA 74175 to V.N.P.
Was this article helpful?