The major human neurodegenerative diseases, including Alzheimer’s, amyotrophic lateral sclerosis, Parkinson’s, and Huntington’s diseases, are associated with accumulation and aggregation of misfolded proteins. In most cases, the majority of aberrantly aggregated proteins are found in the cell cytoplasm. However, in disorders caused by the expansion of a trinucleotide repeat, including Huntington’s disease and spinocerebellar ataxia, the corresponding aggregates of proteins containing the encoded polyglutamine expansions are predominantly nuclear. Whether differences in intracellular location matter for the toxicity generated by such proteins has not been determined. On page 173 of this issue, Woerner et al. (1) report that the location does indeed matter, with toxicity arising from the cytoplasmic accumulation of a pair of artificial proteins designed to mimic the properties of amyloid aggregates. Surprisingly, forcing the same artificial proteins into the nucleus substantially reduces their toxicity.
Woerner et al. established a cell culture system in which artificial β-sheet proteins, previously shown to form fibrillar amyloid aggregates (2), can be targeted to accumulate in the cytoplasm or nucleus by inclusion of a nuclear export sequence (NES) or nuclear localization sequence (NLS), respectively (see the figure). With this approach, they demonstrate that only the cytoplasmically targeted proteins, but not the nuclear counterparts, enhance cell death. The authors propose that the reduced toxicity of the nuclear proteins and their aggregates, despite accumulating in amounts comparable to those in the cytoplasm, may be the result of the chaperone-like activity of a highly abundant nucleolar protein called nucleophosmin-1 (NPM1), which they show interacts with nuclear but not cytoplasmic aggregates.
These discoveries add to other emerging evidence that compartment-enriched chaperones—which form complexes with misfolded proteins in the cytosol (3) or nucleus (4)—may play central roles in ameliorating damage, possibly by generating compartment-specific conformers of aggregated proteins with different propensities for cellular toxicity. Woerner et al. report that nuclear β-sheet–containing proteins produce aggregates that have reduced solubility and weaker affinity for an amyloid-specific dye compared to their cytosolic counterparts, underscoring possible differences in aggregate conformation between the two subcellular compartments. Whether any of the nuclearly enriched chaperones (4) contribute to these possible conformational changes or whether they shield the surfaces of the nuclear amyloid-like protein aggregates (thereby making them more innocuous) has not been established.
So why is misfolded protein accumulation and aggregation in the cytoplasm toxic? Woerner et al. used a proteomic approach to implicate the THOC2 subunit of a messenger RNA (mRNA) export complex that facilitates mRNA delivery to the cytoplasm. In primary neurons, THOC2 is mislocalized to the cytoplasm in cells with cytoplasmic β-sheet aggregates, although its interaction with those is unlikely to be direct as the aggregates are distinct from cytoplasmic redistributed THOC2. Components of the nuclear pore complex and nuclear import receptors are misaccumulated in the cytoplasm, strongly implicating diminished nuclear import and export in the affected cells. Not yet determined is whether nuclear proteins or nuclear pore components, and if so which ones, are trapped by the cytoplasmic amyloid aggregates, thus preventing their proper nuclear localization and function.
Using similar assays, Woerner et al. show that expression of disease-linked fragments of polyglutamine-containing huntingtin protein or amyotrophic lateral sclerosis–causing mutants in the transactivation element (TAR) DNA binding protein–43 (known as TDP-43) also inhibit mRNA export when expressed in cultured cells, suggesting that errors in nucleocytoplasmic transport may be common to multiple neurological conditions. That said, expression of mutant huntingtin in primary cultures of cortical neurons led preferentially to nuclear aggregation, which did not impair nuclear mRNA export. This is consistent with evidence that intranuclear inclusions of polyglutamine-containing huntingtin fragments are not toxic per se (5). In the widely used R6/2 Huntington’s disease mouse model in which aggregates of a mutant huntingtin fragment accumulate intranuclearly in most neurons, Woerner et al. report impaired RNA export in the small proportion of neurons that accumulate aggregated huntingtin in the cytoplasm. These findings, and the consensus from analyses of human samples and most mouse models, raise the question of whether the much rarer cytoplasmic aggregates are primary contributors to toxicity in Huntington’s disease, rather than the more abundant intranuclear ones.
The finding by Woerner et al. that cytoplasmic aggregates diminish nuclear import and/or export adds to the growing recognition that diminished nucleocytoplasmic transport may be a common component of multiple human neurodegenerative diseases, including Huntington’s (6), Alzheimer’s, amyotrophic lateral sclerosis, frontotemporal dementia, and Parkinson’s, where components of the import and/or export machinery are mislocalized and found to interact with disease-associated mutant proteins. Coupled with nuclear “leakiness” that dramatically accelerates during aging (7), altered cytoplasmic localization offers one explanation for the age-dependence of these neurodegenerative disorders.
How nuclear import-export is inhibited in the various diseases is still unclear. Recently, expression of a hexanucleotide expansion within the C9orf72 gene, which is the most frequently inherited cause of both amyotrophic lateral sclerosis and frontotemporal dementia, has been reported to disrupt nuclear import and/or export (8–10), but how this defect arises is not firmly established. One study identified a direct interaction between the hexanucleotide repeat–containing RNAs and Ran GTPase-activating protein (Ran-GAP), a factor required for nuclear import (10). Other studies implicated import inhibition by repeat associated non-AUG (referred to as RAN)–dependent translation-produced polydipeptides encoded by expansion-containing RNAs (8, 9).
To this controversy, Woerner et al. demonstrate that nuclear and/or cytoplasmic transport defects can be attributed to a proteotoxicity caused by cytoplasmic accumulation of β-sheet proteins and their aggregates. Additionally, the recent finding that RAN translation is not restricted to diseases with noncoding region repeat expansions, but also occurs across repeats located in an open reading frame such as in Huntington’s disease (11), provides a new perspective on potential mechanisms underlying toxicity in this disorder. A critical next step will be to determine whether RAN-encoded peptides can directly provoke nucleocytoplasmic transport defects previously reported in Huntington’s disease (6), and whether there is compartment-selective toxicity, as now demonstrated for the β-sheet proteins.
To the unresolved, key question of how cytoplasmic accumulation of aberrant proteins and/or their aggregation provokes diminished nuclear import and/or export, it must be noted that the focus in Alzheimer’s disease (12), Huntington’s disease (5, 13), and most recently in inherited amyotrophic lateral sclerosis (14), has reversed. An initial focus was on the large aggregates seen with conventional pathology. Most investigators have refocused on oligomeric assemblies of the misfolded protein as the most important contributors to neuronal dysfunction that leads to the characteristic disease symptoms (15). Seen from this prospective, location definitely matters, but the β-sheet protein aggregates (and other aggregates in the various disorders) may actually be protective, with toxicity arising from oligomeric species that are hard to detect.
子夜星河 