In this week's early online edition of PNAS, collaborative research directed by Floyd E. Bloom of Neurome Inc. and the Scripps Research Institute, both in La Jolla, California, shows that dramatic changes occur in the brains of mice expressing mutant human forms of AβPP long before any Aβ has been deposited. This study is the latest in an ongoing trend pointing to damage in young adult AD mouse models prior to amyloid plaque formation, fed to date primarily by electrophysiological and behavioral tests. It is also the first report of the use of quantitative three-dimensional MRI in an AD mouse model.

First author Jeffrey Redwine et al. examined mice expressing a mutant human AβPP originally found in cases of familial AD. Using high-resolution magnetic resonance microscopy, Redwine et al. measured the volume of different regions of mouse brains at various stages of growth. At 40 days of age, the authors found no difference between wildtype mice and their transgenic littermates, indicating that embryonic and early postnatal development are unaffected by mutant APP overexpression. Already by 100 days (i.e., young adulthood), however, the transgenic mice had a 12 percent reduction in hippocampal volume. The authors found this reduction to persist unchanged until 21 months. Intriguingly, it appears to result from poorer hippocampal growth because the hippocampi in wildtype mice actually grew by about 18 percent between 40 days and 21 months. There was no growth difference between wildtype and transgenic cerebellums, but the corpus callosum was about 25 percent shorter in transgenic mice of all ages.

The stunted hippocampal growth was most evident in the molecular layer of the dentate gyrus, the projection area of the perforant pathway that is severely affected early on in human AD. The researchers could not distinguish whether the loss in the dentate gyrus results from loss of projection neurons or resident dendrites, or both.

One surprise of this study was that the hippocampus apparently continues to grow in normal adult mice. Another was that volumetric losses in the transgenic animals preceded any detectable deposition of Aβ. This leads Redwine et al. to speculate that the mutant AβPP may cause early pathologic changes that slow down normal growth of the hippocampus. Among possible culprits, the authors mention increased soluble Aβ, citing its reported ability to bind to neuronal receptors and to agrin, a protein known to function in dendritic growth and synapse formation. Finally, the authors write that their findings reinforce the importance of developing strategies to detect AD early; one of the promising candidates for this goal is measuring the volume of the hippocampus and other selected brain regions by MRI. This study also suggests that prospective imaging studies of FAD family members even in their twenties may yield valuable insight into the natural history of this disease.—Tom Fagan and Gabrielle Strobel


  1. The quantitative analyses of volumes of different brain structures in an APP mutant mouse line by Redwine et al. provide valuable new information concerning the possible relationships between altered APP processing, amyloid deposition, and cellular damage in Alzheimer's disease. The data reveal reductions in the size of the hippocampal dentate gyrus and the corpus callosum that occur prior to evidence of amyloid deposits. Indeed, both of these brain structures are smaller in the APP mutant mice as early as 40 days of age with further divergence from age-matched non-transgenic control mice at 100 days of age. Previous studies of this line of APP mutant mice have suggested no cell loss in the hippocampus, although numbers of dentate granule cells have not been quantified in a rigorous manner. However, there are data suggesting that numbers of synapses are decreased in APP mutant mice. It will be important to establish that similar decreases in the sizes of the dentate gyrus and corpus callosum occur in other lines of APP mutant transgenic mice, and to determine whether the increased production of amyloid beta-peptide is the factor causing the reductions in these brain structures.

    The differences in the sizes of the dentate gyrus and corpus callosum at such an early age suggests a possible abnormality in brain development. The nature of such an abnormality is unknown, but could involve impaired neurogenesis, neurite outgrowth or synaptogenesis. Indeed, recent studies have shown that neurogenesis is impaired in the dentate gyrus of a different line of APP-mutant transgenic mice, and that amyloid β-peptide at subtoxic concentrations can impair neurogenesis (Haughey et al., 2002).

    The reduction in the size of the corpus callosum in the APP-mutant mice examined by Redwine et al. could be the result of decreased numbers of axons, decreased axon size, or reduced numbers of myelinating oligodendrocytes. It was recently reported that oligodendrocytes are vulnerable to amyloid toxicity, and that the vulnerability of oligodendrocytes is in the corpus callosum of presenilin-1 mutant knockin mice (Pak et al. 2003. Presenilin-1 mutation sensitizes oligodendrocytes to glutamate and amyloid toxicities and exacerbates white matter damage and memory impairment in mice. Neuromolecular Med., in press.)


    . Disruption of neurogenesis by amyloid beta-peptide, and perturbed neural progenitor cell homeostasis, in models of Alzheimer's disease. J Neurochem. 2002 Dec;83(6):1509-24. PubMed.

  2. This study by Redwine and colleagues is very well done and meticulous. It convincingly demonstrates reduced volume of the dentate gyrus by 100 days of age in the PDAPP transgenic mouse model of Alzheimer’s disease. Aβ-containing plaques have been shown to accumulate in this model by 240-300 days. Thus, this paper is crucial in demonstrating a frank, fairly gross structural change well in advance of the appearance of one of the major traditional markers of AD pathology.

    It is important to note that two independent methods were used and that they both arrived at similar conclusions. Meticulous magnetic resonance microscopy (MRM) was used (T2 images, 11.7 T equipment) to determine the volumes of the total brain, hippocampus and cerebellum. Genu–splenium length of the corpus callosum was also measured. Ages studied were 40 days, 100 days, and 12 and 21 months. MRM data were obtained from perfusion-fixed mouse brains while still in the skull. Brain was then removed, sectioned at 50µm, and stained with thionin and cresyl violet. These stained sections were used to obtain unbiased stereological estimates of the regional volumes in 11 of the 49 mice used for MRM. These Nissl-stained sections also allowed a more detailed analysis of the subregions of the hippocampus.

    The MRM data show early volume reduction in the hippocampus only, and not in the cerebellum or the total brain. The length of the corpus callosum was reduced in transgenic mice at all time points measured. The stereological analysis of Nissl sections further showed that the volume reduction in the hippocampus was attributable to the dentate gyrus (defined as granule cell layer plus molecular layer, not hilus). The MRM data and the Nissl data were highly correlated (r = +0.75, r2 = 0.57, p = 0.007).

    What do these data mean? For over 100 years the plaques and neurofibrillary tangles described by Alois Alzheimer in his patient, Auguste D, have dominated thinking about the pathophysiology of AD. Yet, for decades there has been evidence of a need to escape the shadow of these markers. The once-pivotal study of Blessed et al. 1968 showed no relationship between plaque density and cognitive impairment in Alzheimer’s disease once their data were appropriately analyzed. Other studies of animal models have described behavioral and morphological losses in the absence of plaques, even in the mouse model studied by Redwine and collaborators (e.g., Dodart et al., 2000). The Redwine article uses advanced methods and provides more detail to tear away yet one more segment in the fabric of plaques and tangles.

    Such findings emphasize that from the clinical to the neuropathological levels we have too often been using discovery methods that are not appropriately revealing. It is clear that the molecular, cell biological and neuropathological (e.g., Braak and Braak, 1997; Morsch et al., 1999) stigmata of AD have been ravaging the brain for years or decades prior to the clinical detection of the disease. Does the person in whom these stigmata are present have AD even though it is not detected with current clinical methods? They certainly do. And is it possible that a person whose brain is devoid of plaques and NFT may also already be traveling down the AD road? This may also be certainly so, and the Redwine et al. article underscores this possibility (but does not constitute proof, for, after all, this transgenic mouse is not a full replicate of the human disease).

    This is not to say that plaques and NFT must be relegated to the realm of the inconsequential. It is, rather, to assert that plaques and NFT must be regarded as latecomers in the pathological cascade, the light under the lamppost, whose predecessors (undetected until recently) may well play more critical roles in the pathophysiology of AD. Thus, as Redwine and coworkers suggest, binding of Aβ peptides to agrin may be critical to the dendritic failure in AD first described by Buell and Coleman, 1979. Or (and?) the translocation of a C-terminal fragment of APP to the nucleus (see ARF related news story) may lead to critical changes in transcription. Similarly, it is now known that altered posttranslational processing of tau is found in the absence of (prior to?) the formation of NFTs (many papers by Peter Davies and collaborators), and that these changes have consequences for microtubule integrity (Drewes et al., 1997) and the expression of synaptic molecules (Callahan et al., 2002). The study of the plaques and NFT under the lamppost has not been in vain and may have been for a time the best we could do. With the molecular, cell biological, imaging and other tools currently at our disposal it is now time to emphasize seeing the previously unseen. This must then surely lead us to a much more complete understanding of the mechanisms of AD and to earlier detection and effective therapeutic intervention.


    . The association between quantitative measures of dementia and of senile change in the cerebral grey matter of elderly subjects. Br J Psychiatry. 1968 Jul;114(512):797-811. PubMed.

    . Neuroanatomical abnormalities in behaviorally characterized APP(V717F) transgenic mice. Neurobiol Dis. 2000 Apr;7(2):71-85. PubMed.

    . Frequency of stages of Alzheimer-related lesions in different age categories. Neurobiol Aging. 1997 Jul-Aug;18(4):351-7. PubMed.

    . Diagnostic criteria for neuropathologic assessment of Alzheimer's disease. Neurobiol Aging. 1997 Jul-Aug;18(4 Suppl):S85-8. PubMed.

    . Neurons may live for decades with neurofibrillary tangles. J Neuropathol Exp Neurol. 1999 Feb;58(2):188-97. PubMed.

    . Dendritic growth in the aged human brain and failure of growth in senile dementia. Science. 1979 Nov 16;206(4420):854-6. PubMed.

    . MARK, a novel family of protein kinases that phosphorylate microtubule-associated proteins and trigger microtubule disruption. Cell. 1997 Apr 18;89(2):297-308. PubMed.

    . Progressive reduction of synaptophysin message in single neurons in Alzheimer disease. J Neuropathol Exp Neurol. 2002 May;61(5):384-95. PubMed.

  3. When introduced as a clinical tool in the 1980s, MRI transformed the field of clinical neuroscience. Visualizing the brain with exquisite anatomical resolution was one thing, but to accomplish this feat noninvasively, so that human subjects could be imaged repeatedly over time, was too good to be true. Finally, we had a technique allowing us to study the natural course of disease in living patients. Up to that point, we’d relied mainly on dead human tissue, or if we wanted to study disease in a living system, we turned to animal models and used invasive mapping techniques. In a perfect illustration of science’s dialectical course, the current study—which is as powerful as it is elegant—shows that MRI’s utility can be retrofitted to the study of dead mice.

    Why so much excitement over this seemingly retrogressive trend? One reason is that after a few decades of perfecting imaging protocols designed to highlight different aspects of brain anatomy and function, we now need to verify what it is we are actually imaging. Verification is best accomplished in animal models of disease against independent measures that typically require in vitro analysis. The current study nicely illustrates the importance of performing animal imaging to gain insight into sources of signal change.

    Volumetric MRI, an imaging approach that has gained in popularity in the last few years, uses MRI maps to estimate and compare the volume of different brain regions, either between groups or over time. Volumetric MRI has become particularly common in assessing patients with neurodegenerative disorders, where there isn’t a clear "pathological" signal as there is in stroke or multiple sclerosis, and these studies typically find diminished volume in targeted regions of the brain.

    What is the underlying mechanism that leads to diminished brain volume? Either implied or explicitly stated, many volumetric studies make the assumption that decreases in brain volume suggest areas of cell loss. This interpretation, however, became problematic when a number of studies found diminished volume in states or brain regions where we know there is no cell loss. Age-related volume loss observed in the neocortex is a good example, because age-related cell loss does not occur in these regions. The current study, as well as other studies coupling neuroimaging and postmortem analysis, suggests that shrinkage occurs independent of cell loss. The emerging view is that diminished brain volume most likely reflects shrinkage of extracellular tissue constituents, and that volumetric measures are more sensitive to white-matter over gray-matter changes. The precise mechanism that governs this change remains unknown, but with the primary finding made by Redwine and colleagues, these investigators are perfectly poised to pursue this question. If, in fact, volumetric changes are related to changes in synaptic function, as the authors postulate, then volumetric MRI may be capturing a snapshot of the chronic functional state of the brain, and may be closer in kind to fMRI.

    Another general reason for the excitement over mouse imaging is that as we perfect our models of disease—and it appears that mice transgenically expressing FAD genes are at least good models of Aβ toxicity—we could use imaging to gain insight into mechanisms of disease. By following mice longitudinally over their natural life span of around two years, we can establish selectively vulnerable brain regions, and more precisely map pathological spread over time. In this regard, the current study is somewhat disappointing. The transgenic mice were sacrificed for the study because of the long time it takes to image them; this in many ways negates the main advantage of in vivo imaging, namely longitudinal follow-up. I would imagine that this or other groups are working to acquire reliable volumetric data in shorter time periods. If so, then this MRI approach will not only address questions of where and when, but would be ideally suited to test pharmacological intervention.

    As indicated by their title, the authors seem to place a great emphasis on the selective changes that they report on in the dentate gyrus. Measuring the volume of the hippocampal subregions requires visualization of the precise boundaries between subregions, and this can only be accomplished with histological staining. Thus, the main finding of this study cannot be made using volumetric MRI. In any case, a minor concern regarding this dentate gyrus observation should be voiced. In their analysis, the authors chose not to include the hilus of the dentate gyrus in their measurements. Thus, as shown in figure 5, the measurement of the dentate gyrus includes fewer cellular layers compared with measurements from the CA1, CA3, and the subiculum. Since volumetric measurements are more sensitive to changes in white matter over gray matter, a concern is that the slightly greater difference they observed in the dentate gyrus over CA1 might simply reflect this selection bias.

  4. The study by Redwine et al. provides important new evidence for FAD mutant AβPP and/or elevated Aβ causing preplaque structural changes in a well-established FAD transgenic mouse strain. The excellent comments on this paper reinforce the point that the study of how mutant APP/Aβ may relate to synaptic dysfunction and selective regional volume loss in brain certainly is an important area of study. A major question is how AβPP or Aβ is responsible for these changes. Increasing evidence favors a role for soluble Aβ oligomers in this pathological process (multiple studies; i.e., Walsh et al., 2002). Increasing evidence also indicates that preplaque Aβ accumulation occurs within nerve cells in FAD transgenic mice and AD brain (also, multiple studies, most recently Shie et al., 2003; see ARF online discussion). The biological interactions of these different pools of brain Aβ, especially at synapses, appears to be an area that may be especially important in understanding why structural abnormalities occur prior to plaque formation. The remarkable early histological studies that provided so much insight into AD seem, indeed, to become superceded as plaques appear less essential to brain dysfunction. Newer pathological studies are continuing to provide clues that demonstrate, for example, abnormal Aβ accumulation in processes and at synapses prior to plaques in FAD-transgenic mice. These subcellular increases may be playing a role in the early preplaque APP/Aβ-related changes that are increasingly being reported—such as now in this magnetic resonance microscopy study—but they have not yet been mentioned here.


    . Naturally secreted oligomers of amyloid beta protein potently inhibit hippocampal long-term potentiation in vivo. Nature. 2002 Apr 4;416(6880):535-9. PubMed.

    . Early intraneuronal Abeta deposition in the hippocampus of APP transgenic mice. Neuroreport. 2003 Jan 20;14(1):123-9. PubMed.

  5. The is an interesting and somewhat unexpected finding. At 100 days of age PDAPP mice show a small but significant decrease in hippocampal volume (12.3%). This is before there is measurable biochemical or immunohistochemical Abeta deposition. The main quesiton raised by this study is whether the change is due to APP, soluble Abeta, or even non-specific transgene effects. Further studies will be needed to sort out these possibilities.

    View all comments by Todd E. Golde

Make a Comment

To make a comment you must login or register.


No Available References

Further Reading

No Available Further Reading

Primary Papers

  1. . Dentate gyrus volume is reduced before onset of plaque formation in PDAPP mice: a magnetic resonance microscopy and stereologic analysis. Proc Natl Acad Sci U S A. 2003 Feb 4;100(3):1381-6. PubMed.