Introduction
The aggregation mechanisms of the polyglutamine-containing sequences relevant to Huntington disease and other expanded-CAG-repeat diseases are of continued great interest. In this issue of Nature Chemical Biology, Colby et al. describe an elegant analysis of the nucleation mechanism of aggregation for one such polyglutamine-rich fragment, the exon I fragment of huntingtin (httex1), as it takes place within a cell model1. The results confirm and extend previous in vitro studies of polyglutamine aggregation nucleation2, 3 and support earlier speculations that nucleation kinetics might influence ages of onset in these diseases2.
The universal observation of polyglutamine aggregates in brain tissue from persons with expanded-CAG-repeat disease and from animal models, plus a repeat-length dependence of aggregation that recapitulates trends in disease ages of onset, have driven a sustained interest in disease mechanisms involving polyglutamine aggregation4. Although recent cell studies5 suggest that large, easily detected inclusions may be protective rather than toxic, interest in aggregation-based disease mechanisms has not waned, as there are a variety of other polyglutamine aggregates of different sizes and morphologies that remain under suspicion as agents of cell dystrophy and death6, 7.
In vitro experiments show that simple polyglutamine peptides follow a classical nucleated growth pathway8 in which a rare nucleation event precedes an aggressive aggregate growth (elongation) phase2, 3. These in vitro experiments have been typical 'ensemble' studies containing a relatively large number of protein molecules (Fig. 1). Concerned that the number of molecules of huntingtin (htt) in a cell is likely to be too low to yield uniform aggregation behavior in each cell, Colby et al. applied stochastic models to analyze aggregate formation in a cell model of httex1 aggregation (Fig. 1). Thus, rather than study the total mass of aggregated and non-aggregated httex1, they followed the numbers of cells with and without aggregates over time and fit the data to equations representing various aggregation mechanisms.
Figure 1: Nucleation of polyglutamine aggregation.
(a) Test tube experiments involve ensemble measurements in which there are sufficient molecules that even rare events occur with a predictable and reproducible frequency. (b) Cells contain so few molecules that rare events may occur only in some of the cells, leading to stochastic appearance of cells with aggregates. (c,d) Two mechanisms of nucleation consistent with cell experiments. In one, (c), the initiating aggregation rate is limited by both rare formation of a molecule in an abnormal conformation (one concentration term) and its collision with a molecule in the normal conformation (a second concentration term). In the second, (d), the initiating aggregation rate is limited by the collision of two molecules, both which must be in the rare, abnormal conformation (two concentration terms). Once a small aggregate is formed, it grows through cycles of elongation reactions with the normal conformation of polyglutamine (the third line in each of the mechanisms). Molecules of normal conformations are indicated by a wavy line; molecules in the abnormal conformation are designated as horseshoes.
Ingrid McNamara
Full size image (49 KB)The results are equally consistent with two possible nucleation mechanisms in which kinetic control of aggregation involves a sparsely populated, abnormally folded state of the disease protein (Fig. 1). In contrast to the oligomeric assemblies normally considered in the classical nucleation theory8, both of these mechanisms posit that the critical structure in nucleation is a misfolded monomer. In one mechanism (Fig. 1c), identical to that previously proposed by Chen et al.2 on the basis of ensemble measurements, the nucleus is a misfolded state of the monomer, and spontaneous aggregation depends on the productive reaction of this nucleus with a native-state monomer. The alternative mechanism requires a productive interaction between two molecules in the misfolded state (Fig. 1d).
The studies of Colby et al. provide an important bridge between the earlier studies of simple polyglutamine peptides in the test tube and the behavior of more complex polyglutamine proteins in the cellular environment. The efficiency of nucleation in the cell calculated by Colby et al. appears to be roughly 1,000 times lower than that observed in vitro in ensemble studies3. There are several possible rationales for this discrepancy, including a scenario in which many nascent aggregation reactions are efficiently neutralized by proteasome surveillance and thus never mature into macroscopic aggregates. Colby et al. also show that their experimentally determined kinetic parameters project to a reasonable approximation of the kinetics of decline in Huntington's disease for a htt concentration of 100 nM (the intracellular concentration of htt is, however, unknown). This ability to project time-dependent neuronal loss and hence Huntington disease age of onset purely on the basis of nucleation kinetics is an intriguing outcome of these experiments. Given the many added layers of complexity in cells as compared to in vitro experiments, the degree of agreement between the results of Colby et al. and the earlier ensemble measurements2, 3, with respect to both mechanism and projected ages of onset2, is fairly astounding and gives us confidence in continuing to think of polyglutamine aggregation as a well-behaved and ultimately simple nucleation-dependent process.
At the same time, there is much to be learned before we have a full understanding of the aggregation mechanisms of disease-related polyglutamine proteins. For example, in contrast to simple polyglutamine peptides7, htt fragments in vitro form oligomeric and protofibrillar aggregates in addition to amyloid-like fibrils9. Also, tissues from individuals with Huntington disease contain mixtures of both elongation-competent and elongation-incompetent polyglutamine aggregates6. Htt aggregation in the cell might thus occur by parallel pathways leading, with differing kinetics, to multiple aggregates of differing morphologies, functionalities and toxic activities7.
In addition, studies of polyglutamine aggregates in disease and animal models often reveal several, and in some cases many, independent microaggregates in single cells6, 10, 11. Yet the simple nucleated growth models for htt aggregation discussed here suggest that, because nucleation is a very rare event, and because elongation is much more efficient than nucleation, each cell should never contain more than one aggregate (Fig. 1). How can we account for this discrepancy? Do multiple aggregates arise via coagulation kinetic mechanisms that do not depend on rare nucleation events, or, alternatively, from the breakdown of larger aggregates within the cell? Or are there aspects of the cellular environment (such as high viscosity, compartmentalization and perhaps others) that effectively create multiple, virtual reaction chambers each capable of sustaining an essentially independent nucleation and aggregation reaction? Further studies using approaches similar to that of Colby et al. may help to address some of these important issues.
Although the studies of Colby et al. do not address the issue of aggregate pathogenicity, and although they leave a number of important questions about the aggregation process to be answered by future experiments, the results confirm the small nucleus size and slow initiation of polyglutamine aggregation, and they also support the idea that treatments that either reduce the intracellular concentration of the expanded polyglutamine protein or target nascent aggregation nuclei may prove to be viable therapeutic approaches for this family of devastating diseases.
