Alzheimer disease rages in the brain long before plaques form, and even before the earliest measurable cognitive symptoms. Most in the field agree that early interventions are the best hope of nipping memory loss and cognitive decline in the bud. Which raises the question—when exactly, and where, does AD start?

To find some answers, Floyd Bloom of Neurome Inc., and the Scripps Research Institute, both in La Jolla, California, has been scrutinizing young PDAPP mice, looking for the earliest signs of pathological change (see ARF related news story and ARF news story). Using morphological, behavioral, and biochemical measures, Bloom and colleagues have pinpointed defects in the hippocampus as early as 4-5 months of age in these mice, long before plaques appear at 18 months. Their goal is to define early changes in the mouse model as a testing ground for therapeutics, and to understand the early memory loss that occurs in people with AD.

The progression of AD varies among different mouse models, leading Bloom and colleagues to extend their analysis to another transgenic model, the Tg2576 mouse. In a paper in this week’s PNAS Early Edition, first author J. Steven Jacobsen at Wyeth Research in Princeton, New Jersey, and colleagues report that the Tg2576 mice undergo similar, but not identical changes at 4-6 months of age compared to the PDAPP mice. Like in the PDAPP mice, the first change seen was a decrease in spine density in the outer molecular layer of the dentate gyrus in the hippocampus by 4 months of age. But unlike earlier findings in the PDAPP mice, the researchers did not find a decrease in hippocampal volume in the Tg2576 line. The published work follows a presentation at last year’s Society for Neuroscience meeting (see ARF related news story).

The loss of dendritic spines observed at 4 months of age coincided with a decrease in basal synaptic transmission and long-term potentiation in hippocampal slices, as well as behavioral changes in the hippocampal-based learning test of contextual fear conditioning. The cause of these synaptic problems remains to be found. The researchers did detect a measurable rise in the fraction of soluble Aβ42 after 6 months, correlating with the emergence of spatial memory defects. It took a detailed analysis of soluble Aβ species by Karen Hsiao Ashe and colleagues to identify a specific dodecameric form of Aβ as a candidate for causing the memory problems that show up at 6 months in these mice (see ARF related news story), and the same type of analysis may be required to shed light on the even earlier events.

The decrease in spine density in the dentate gyrus in two different models suggests this may be a common early effect of aberrant APP processing, consistent with other studies that this region is highly sensitive to aging. If the early changes seen in the mouse model mimic the early memory defects in humans (as the authors deem likely), then the animals and techniques described will be valuable for testing early treatments.—Pat McCaffrey


  1. The first sentence of Pat McCaffrey's news summary is both enlightening and puzzling: "Alzheimer disease rages in the brain long before plaques form...." It was not that long ago that the century-old adage "amyloid in plaques is the major problem in AD" was modified to "amyloid in neurons." We have come a long way, and it is satisfying to see, with less than a fortnight apart, two major papers pointing to early amyloid peptide-related defects, that is, a new molecular structure referred to as Aβ*56 (Lesné et al., 2006) and new cell-functional consequences in vivo (Jacobsen et al., 2006).

    I disagree with McCaffrey's second sentence that "Most in the field agree that early interventions are the best hope of nipping memory loss and cognitive decline in the bud." I am convinced most in the field actually know that this is the only way forward, instead of trying to treat the late symptoms that actually signal an already irreversibly established pathology. Evidently, current clinical practice of treatment must be continued and efforts to improve on them even stepped up as fast and as far as possible. But the grand aim of research must be identifying the most "early defects" as the tell-tale signs of upcoming, and hopefully reversible, pathology.

    We were the first to approach those early signs in our APP mouse models, even when the field was still concentrating on amyloid plaques (Moechars et al., 1996, 1999). The second publication is better known and cited than its predecessor from 1996, but their message was and is the same: Behavioral and cognitive functional defects occur in the absence of any amyloid plaques, on which note we converge on McCaffrey's question: "…when exactly, and where, does AD start?"

    Bloom and colleagues analyzed two APP transgenic models and come to the conclusion that LTP is impaired at age 4-6 months concomitant with decreased spine density in the outer molecular layer of the dentate gyrus. This parallel is an important new piece of evidence; all other defects in LTP and in cognition, followed by increased Aβ peptides and amyloid plaque load, were demonstrated and are well known in these and other APP models (for review see Van Dooren et al., 2006).

    McCaffrey then rightfully states: "The cause of these synaptic problems remains to be found." This is her cue to switch to the other publication I referred to above: The detailed analysis of soluble Aβ species by Dr. Karen Ashe and colleagues, identifying dodecameric Aβ as cause of memory problems in APP mice (Lesné et al., 2006). The Aβ*56 species is a stable molecular complex of Aβ peptides in brain, which appears together with memory defects in APP transgenic mice and causes memory defects when injected in brain of young rats. Thereby, a strong piece of evidence is provided for Aβ*56 to take up center stage in efforts to nail down the real culprit in AD.

    One cannot escape thinking that this might not be "the" but "a" cause of synaptic problems, since several other types of Aβ complexes, isolated or synthesized, prove pathologically active. The precise structural relationship of Aβ dimers, Aβ oligomers, globular oligomers, ADDLs, Aβ*56, etc., remains to be established, but these species point without a doubt to a series of interconvertible molecular forms of Aβ. If they also equilibrate with each other in vivo, it may remain impossible to define which of these forms is the more or most potent in wrecking synaptic functions in the hippocampus. In this respect, I maintain my comparison of amyloid peptides to the ancient dodecaeder, on display in the Gallo-Roman museum in Tongeren, Belgium (see ARF comment). I referred to the dodecaeder as an object with a definite shape, structure, even beauty, but of which we don't know the function or purpose. I humbly confess I never intended to predict the actual dodecamer form of Aβ*56 as the cause of AD!

    The increase with aging of amyloid peptide oligomers of sorts, including dimers, trimers, and dodecamers, correlating with decreased spine density in the dentate gyrus and with the most early defects in cognition in the mouse models, has to be shown to mimic the early memory defects in humans. If so, the mice and tools described will be valuable for testing early treatments, in full agreement with McCaffrey once again!


    . A specific amyloid-beta protein assembly in the brain impairs memory. Nature. 2006 Mar 16;440(7082):352-7. PubMed.

    . Early-onset behavioral and synaptic deficits in a mouse model of Alzheimer's disease. Proc Natl Acad Sci U S A. 2006 Mar 28;103(13):5161-6. Epub 2006 Mar 20 PubMed.

    . A specific amyloid-beta protein assembly in the brain impairs memory. Nature. 2006 Mar 16;440(7082):352-7. PubMed.

    . Expression in brain of amyloid precursor protein mutated in the alpha-secretase site causes disturbed behavior, neuronal degeneration and premature death in transgenic mice. EMBO J. 1996 Mar 15;15(6):1265-74. PubMed.

    . Early phenotypic changes in transgenic mice that overexpress different mutants of amyloid precursor protein in brain. J Biol Chem. 1999 Mar 5;274(10):6483-92. PubMed.

    . Transgenic mouse models for APP processing and Alzheimer's disease: early and late defects. Subcell Biochem. 2005;38:45-63. PubMed.

  2. I think we in this field have to be careful with overinterpreting phenomena in our APP mouse models. Transgenic mouse models expressing AD-associated mutant forms of the amyloid-β precursor protein (APP), or both mutant APP and mutant presenilin-1 (PS1), develop robust amyloid pathology with abundant neurotic plaques that recapitulate many of the features of the Aβ deposits found in humans with AD. As they age, they also show other AD-like features including decreased synaptic density, reactive astro- and microgliosis, and the presence of plaque-associated inflammatory proteins. However, these transgenic models show little evidence of overt neuronal loss and do not, without additional genetic manipulation, develop NFT pathology.

    The APP and APP/PS1 mice also develop cognitive deficits. In most studies, these deficits are observed coincident with the earliest biochemical signs of Aβ accumulation, consistent with early aggregation events, yet the cognitive deficits show limited progression as the mice age and are not tightly linked to the degree of amyloid pathology. Such deficits also appear highly reversible as Aβ immunotherapies rapidly reverse the cognitive deficits even when they have little effects on overall Aβ load. As a result, it seems likely that the APP mice recapitulate only part of the cognitive decline that is seen in AD patients.

    The absence of a more complete recapitulation of AD-type pathology in APP mouse models has been used to argue against a primary role of Aβ accumulation in the development of AD. Although such models do demonstrate that Aβ accumulation and amyloid deposition alone are not sufficient to cause overt neuronal loss in mice, it is inappropriate to conclude, based on such data, that Aβ does not drive these changes in humans. The APP and PS mutations included in the transgenes used to generate these mice do cause AD; therefore, the lack of a complete pathological phenotype in these models simply demonstrates that current transgenic mice are not capable of producing all the features of the human disease.

    The inability to develop a more complete animal model of AD has been a significant problem because it imposes limits on the ability to dissect the pathogenic cascade, which hinders therapeutic studies aimed at downstream targets. It is, therefore, important to understand why APP mice fail to develop the complete spectrum of AD pathology. Further insight into the phenotype of the mice, such as is found in the current manuscript, is of interest, but still does not address the fundamental difference between our APP mouse models and humans with AD, namely the profound neuronal loss.

    On another note, more and more evidence is emerging that amyloid deposition precedes the onset of clinical symptoms perhaps by more than a decade. If this is true, then we have to be careful about statements regarding cause and effect.

Make a Comment

To make a comment you must login or register.


News Citations

  1. Get Out the Rulers—Stunted Dendrites May Flag Early Alzheimer's
  2. Mutant AβPP Retards Growth in Hippocampus before Plaques Form
  3. Translational Biomarkers in Alzheimer Disease Research, Part 5
  4. Aβ Star is Born? Memory Loss in APP Mice Blamed on Oligomer

Further Reading


  1. . A specific amyloid-beta protein assembly in the brain impairs memory. Nature. 2006 Mar 16;440(7082):352-7. PubMed.

Primary Papers

  1. . Early-onset behavioral and synaptic deficits in a mouse model of Alzheimer's disease. Proc Natl Acad Sci U S A. 2006 Mar 28;103(13):5161-6. Epub 2006 Mar 20 PubMed.