24 Sep 2002

Convergent biochemical and genetic evidence indicate that the main component of Alzheimer's plaques, the amyloid β peptide (Aβ), plays an initiating role in a complex cascade that culminates in dementia and ultimately death (1,2). Aβ is produced by proteolytic cleavage of the amyloid precursor protein (AβPP). Extensive studies of the mutations at or just outside Aβ's β- or γ-secretase cleavage sites have shown that these mutations increase the total production Aβ or selectively increase levels of Aβ ending at residue 42 (3). Less well studied are the 5 point mutations within the Aβ sequence that are associated with hereditary diseases similar or identical to Alzheimer's disease (AD). Clustered around the central hydrophobic core of Aβ, these mutations include: the A21G Flemish mutation, E22K Italian mutation, E22G Arctic mutation, E22Q Dutch mutation and the D23N Iowa mutation.

Mutation	Name	Clinical name/description	Effect on Aβ production vs. Wt	Effect on Aβ aggregation	Aβ-mediated toxicity
A21G	Flemish	More pronounced cerebral amyloid angiopathy (CAA) than AD, larger-than-usual parenchymal plaques, and AD-like neurofibrillary changes.	~2-fold higher Aβ40 and 42.	Slightly less prone to form fibrils and protofibrils (PF).	Both fibrillar and soluble preparation of G21 peptides are toxic to cultured neurons, but are not toxic to smooth muscle cells.
E22K	Italian	Very similar to the "Dutch" disease.	Unknown.	Increased propensity to form fibrils. Ability to form PF not yet examined.	Toxic to smooth muscle cells. Effect on neurons unknown.
E22G	Arctic	Indistinguishable from idiopathic AD.	Total Aβ unchanged, but a slight decrease in the Aβ42/40 ratio.	Increased propensity to form fibrils and PF.	Toxic to neuroblastoma cells. Effect on primary neuronal or cerebrovascular cells unknown.
E22Q	Dutch	Hereditary cerebral hemorrhage with amyloidosis, Dutch type (HCHWA-D). Fulminant CAA with more parenchymal amyloid deposits than in age-matched controls. Little or no neurofibrillary changes.	Unchanged.	Increased propensity to form fibrils and PF.	Toxic to smooth muscle and cerebral endothelial cells. Toxic to neuroblastoma cells. Effect on primary neuronal or cerebrovascular cells unknown.
D23N	Iowa	Severe CAA, cortical microinfarcts and white matter loss reminiscent of HCHWA-D; plus widespread presence of NFTs.	Unchanged.	Increased propensity to form fibrils. Ability to form PF not yet examined.	Toxic to smooth muscle cells. Effect on neurons unknown.

What can these mutations and their effects tell us about AD? The phenotypes observed with these mutations cover a broad spectrum, ranging from almost pure cerebral amyloid angiopathy (CAA) to typical AD pathology including plaques and neurofibrillary tangles (NFTs). The Dutch mutation causes a disease referred to as hereditary cerebral hemorrhage with amyloidosis, Dutch type (HCHWA-D). This disease is marked by fulminant CAA, and although parenchymal amyloid deposits are more abundant than in age-matched controls, neurofibrillary changes are seldom observed (4,5). In vitro, Aβ peptides containing the E22Q substitution form fibrils and protofibrils faster than wild-type Aβ (6-9), and are more toxic to cultured cerebrovascular cells (10). This mutation does not alter Aβ production (11), suggesting that the changes in Aβ aggregation and consequent cytotoxicity are causative. The clinical and pathological presentations of the Italian mutation are identical to those observed in the HCHWA-D (12), and the effect of this mutation on Aβ aggregation and toxicity is also similar(13).

The recently described Iowa mutation is associated with clinical and pathological features, which resemble both CAA and AD. Pathologically, those carrying the mutation demonstrate severe CAA, cortical microinfarcts, and white matter loss reminiscent of HCHWA-D; however, the widespread presence of NFTs mirrors that seen in AD (14). Assessment of aggregation propensity and cytotoxicity demonstrated that Aβ peptides containing the D23N (Iowa) substitution behaved in a manner similar to peptides containing the E22Q mutation. In addition, cells transfected with ABPP D23N did not show increased amyloidogenic processing of AβPP (15).

Like the Iowa mutation, the Flemish mutation (A21G) is associated with a disease phenotype that overlaps both HCHWA-D and AD. In the Flemish disease, CAA was more pronounced than in AD but neurofibrillary changes were similar. In vitro, Aβ peptides containing the A21G substitution behave very differently from either E22Q or D23N peptides, i.e. the Flemish peptides were less prone to aggregation than wild type peptides (8,9) and were not toxic to human smooth muscle cells (16), but were toxic to neuronal cultures (17). Moreover, in cells transfected with AβPP, A21G total levels of secreted Aβ were approximately two-fold higher than in cells transfected with wild-type AβPP (18), indicating that increased amyloidogenic processing contributed to the observed pathology.

In contrast to the other intra-Aβ mutations, the Arctic mutation (E22G) is associated with a disease phenotype indistinguishable from AD (19). Interestingly, in vitro experiments revealed that Aβ peptides bearing the E22G substitution are more prone to form protofibrils than wild-type Aβ (20) and are toxic to human neuroblastoma cell lines (21). Moreover, in transfected cells the E22G mutation does not alter levels of secreted total Aβ, but causes a small significant decrease in the Aβ42/A 40 ratio, indicating that E22G does not increase amyloidogenic processing of Aβ and must mediate pathogenesis by another mechanism, possibly by forming toxic assemblies.

It is intriguing that mutations at residues 21 and 22 of Aβ lead to such different phenotypes. Although a substantial body of data already exists on the effects of these mutations on AβPP processing, Aβ aggregation, and toxicity, many gaps remain. For instance, it still remains to be determined if the Dutch, Italian and Arctic mutations really do not effect Aβ production. Perhaps, like all the other known Aβ mutations, they do indeed increase Aβ levels, which elude detection by conventional ELISA methods due to their increased propensity for aggregation. Similarly, although a wealth of data exists pertaining to the relative ability of the various peptides to form aggregates and cause toxicity, many of these studies occurred before the discovery of protofibrils (9, 22) or Aβ derived diffusible ligands (ADDLs) (23). Moreover, no single study has characterized the relative toxicity for each peptide on a single cell type. Obviously, it will be tremendously important to assess the toxicity of different well-defined assemblies (monomer, oligomers, ADDLs, protofibrils and fibrils) on primary neuronal and cerebrovascular cells. So often in the past, detailed study of rare inherited forms of human disease have yielded important insights into the more common idiopathic disease. For this reason, it is imperative that the disease-causing intra-Aβ mutations continue to receive careful attention and that their secrets not be lost in a tangle of incomplete or conflicting data.-Dominic Walsh, Brigham and Women's Hospital, Boston, Massachusetts.

See also:

Tagliavini, F., Rossi, G, Padovani, A, Magoni, M, Andora, G, Sgarzi, M, Bizzi, A, Savioardo, M, Carella, F, Morbin, M, Giaccone, G, and Bugiani, O. (1999) Alzheimer's Reports 2, S28.

Aβ Mutations—What Do They Tell Us?

Quick Links

Tools

Comments

Make a Comment

References

Further Reading

Papers

Annotate