This is a very exciting development that strongly supports the amyloid hypothesis of AD causation. It appears to be symmetrical with the discovery of α-synuclein duplication (and triplication) in otherwise phenotypically normal individuals in the causation of PD. It suggests that increased expression of wild-type APP—whether from enhanced gene dosage, as in these families, or perhaps from alterations in regulatory elements of the APP gene in other families yet to be discovered—can directly cause AD.

View all comments by Dennis Selkoe

This is a very nice piece of work. First, it highlights the important issue of gene duplication in neurodegenerative disease. Second, the paper is very important for answering the question "Is duplication of APP alone (as in Down syndrome) sufficient to cause AD?" The authors have narrowed the AD region down to just four genes. This almost answers this question, but the genome may still have suprises up its sleeve, so it would be great for other labs to carry out similar analyses and see what the minimal duplicated region is to cause AD.

View all comments by Elizabeth M. Fisher

This is a very interesting paper that is totally consistent with the Aβ hypothesis. The observation that duplication of APP causes early-onset AD and CAA is consistent with the observations in Down syndrome (DS) and confirms that AD pathology in DS is due to APP overexpression. The presence of CAA is also interesting and supports data from the transgenic mice, DS, and other FAD mutations that overproduction of Aβ40 leads to CAA, while increased Aβ42 is associated with parenchymal plaques. This also fits with the observations in PD, where duplication and triplication of α-synuclein have been associated with early-onset PD.

I am surprised by the frequency of the duplications in their early-onset FAD samples (8 percent of FAD cases) and by the fact that each of the duplications is different but not associated with any other phenotype despite the presence of other genes in the duplicated region.

Lastly, these studies really beg the question: Does variation in APP expression contribute to risk for late-onset AD? We posed this possibility in the first mutation papers back in...  Read more

This is a very interesting paper that is totally consistent with the Aβ hypothesis. The observation that duplication of APP causes early-onset AD and CAA is consistent with the observations in Down syndrome (DS) and confirms that AD pathology in DS is due to APP overexpression. The presence of CAA is also interesting and supports data from the transgenic mice, DS, and other FAD mutations that overproduction of Aβ40 leads to CAA, while increased Aβ42 is associated with parenchymal plaques. This also fits with the observations in PD, where duplication and triplication of α-synuclein have been associated with early-onset PD.

I am surprised by the frequency of the duplications in their early-onset FAD samples (8 percent of FAD cases) and by the fact that each of the duplications is different but not associated with any other phenotype despite the presence of other genes in the duplicated region.

Lastly, these studies really beg the question: Does variation in APP expression contribute to risk for late-onset AD? We posed this possibility in the first mutation papers back in 1991/1992 but I don't think the question has been adequately addressed yet. This new data makes it all the more important to tackle this question.

View all comments by Alison Goate

This brief paper describes a rare genetic abnormality in five different families with autosomal dominant, early-onset Alzheimer disease with cerebral amyloid angiopathy. Using three different techniques, this research group was able to establish that a very small portion of chromosome 21 is duplicated in these families. The chromosomal region includes the locus for APP, as well as other genes. Because the duplicated region is slightly different among the five families but in all families the APP locus is included, the conclusion that overproduction of APP is responsible for the disease is compelling. The disease phenotype is similar in all five families. It consists of progressive AD with abundant dense-core and diffuse amyloid deposits as well as neurofibrillary tangles, giving support to the hypothesis that overproduction of the amyloid protein initiates a cascade of events that leads to both plaques and tangles.

In addition, the patients studied by Rovelet-Lecrux et al. have severe cerebral amyloid angiopathy (CAA), and this vascular amyloid deposition is primarily...  Read more

This brief paper describes a rare genetic abnormality in five different families with autosomal dominant, early-onset Alzheimer disease with cerebral amyloid angiopathy. Using three different techniques, this research group was able to establish that a very small portion of chromosome 21 is duplicated in these families. The chromosomal region includes the locus for APP, as well as other genes. Because the duplicated region is slightly different among the five families but in all families the APP locus is included, the conclusion that overproduction of APP is responsible for the disease is compelling. The disease phenotype is similar in all five families. It consists of progressive AD with abundant dense-core and diffuse amyloid deposits as well as neurofibrillary tangles, giving support to the hypothesis that overproduction of the amyloid protein initiates a cascade of events that leads to both plaques and tangles.

In addition, the patients studied by Rovelet-Lecrux et al. have severe cerebral amyloid angiopathy (CAA), and this vascular amyloid deposition is primarily composed of Aβ40. Mutations in Aβ have been discovered in several autosomal dominant cerebral amyloid angiopathies, leading to the hypothesis that amino acid substitutions within the Aβ sequence generate a peptide that has a propensity to aggregate in the microvasculature. However, CAA is also found in sporadic AD, as well as Down syndrome, and is now reported in these families with duplication of the APP locus. Thus, overproduction alone of wild-type Aβ can precipitate development of CAA.

In the patients reported here, the parenchymal amyloid deposits are primarily composed of Aβ42, surrounded by Aβ40. A tendency of Aβ40 to aggregate in the vasculature versus primarily parenchymal deposition of Aβ42 suggests that these peptides are either produced or processed differently within different cell types. Distinct pathological structures characterize the brains of AD patients with mutations of the presenilins (PS1 and 2), where overproduction of Aβ42 is a consistent finding. The comparison of the nature of lesions within each of these genetic variations of AD may help us understand molecular events that are upstream of amyloid deposition.

View all comments by Barbara Tate

This study by Rovelet-Lecrux et al. is interesting in several aspects: It points out that gene duplication may be a more common cause of early-onset familial Alzheimer disease than previously suspected, and it highlights the coexistence of two forms of amyloid pathology, parenchymal and vascular.

The authors did not comment on whether immunohistochemical stains for synuclein or ubiquitin were performed. It will be interesting to study the interactions between the pathological burden of amyloid (both parenchymal and vascular) and Lewy body pathology. Specifically, it will be interesting to determine whether the coexistence of Lewy body pathology with Alzheimer-type pathology in early-onset AD is limited to people harboring presenilin 1 mutations, or whether it is a more general feature seen in some patients independent of the gene mutation (or duplication).

View all comments by Joy Snider

Obviously, the excitement of the paper derives from the fact that simple duplication of the APP gene can cause AD. Scientists had always assumed that it is the extra copy of APP that is responsible for the Alzheimer disease pathology that develops in virtually all individuals over the age of 40 with Down syndrome, but there was not definitive proof that this was the case. While this paper does not prove absolutely that it's the extra dose of APP in the duplicated segments that causes AD—the authors point out that the duplicated segments they studied in the families contained between five and 12 known genes—it certainly makes a good case for APP being the cause. If so, we can conclude that a simple overdose of the wild-type gene has the same clinical consequences as does one dose of the mutated FAD APP genes.

There are precedents for this; for example, either overexpression of the wild-type EGF receptor or else expression of mutant EGF receptors can cause oncogenesis, while either overexpression of wild type α-synuclein or expression of specific mutants of α-synuclein can...  Read more

Obviously, the excitement of the paper derives from the fact that simple duplication of the APP gene can cause AD. Scientists had always assumed that it is the extra copy of APP that is responsible for the Alzheimer disease pathology that develops in virtually all individuals over the age of 40 with Down syndrome, but there was not definitive proof that this was the case. While this paper does not prove absolutely that it's the extra dose of APP in the duplicated segments that causes AD—the authors point out that the duplicated segments they studied in the families contained between five and 12 known genes—it certainly makes a good case for APP being the cause. If so, we can conclude that a simple overdose of the wild-type gene has the same clinical consequences as does one dose of the mutated FAD APP genes.

There are precedents for this; for example, either overexpression of the wild-type EGF receptor or else expression of mutant EGF receptors can cause oncogenesis, while either overexpression of wild type α-synuclein or expression of specific mutants of α-synuclein can cause Parkinson disease (Singleton et al., 2003).

The data in Rovelet-Lecrux et al. are consistent with the previous finding that overexpression of wild-type APP in neurons in vitro causes it to enter the cell cycle and then die via a signal transduction pathway that is very similar to, if not identical with, the pathway by which FAD APP causes neurons to die in vitro (McPhie et al., 2003). Dissecting the elements of this signal transduction pathway may explain why overexpression of APP, but usually not expression of FAD APP, causes cerebral amyloid angiopathy in addition to Alzheimer disease.

View all comments by Rachael Neve

This is a very interesting paper. As the daughter of an Alzheimer's patient it is very important to know that progress is being made. Well done.

View all comments by Deborah Uetz