08 Nov 1998

Are both Aβ and the intracellular hyperphosphorylation of tau necessary for the induction of morphological alterations in dendrites in AD? What effect might abnormally shaped dendrites have on synaptic transmission? Using triple labeling for neurofilament (SMI-32), hyperphosphorylated tau (Alz-50) and β-amyloid (R1280), Robert Knowles, Brad Hyman and colleagues addressed these hypotheses via confocal microscopy (abstract 107.8). In nondemented brains, they found that the morphology (curvature and curvilinear length) of neurofilament-positive, Alz-50-negative dendrites which pass through plaques were unaffected by the presence of Aβ deposits. In the AD brain, these same neurofilament-positive, Alz-50-negative dendrites exhibited a threefold increase in curvature when within Aβ deposits. When dendrites were both Alz-50-positive and within Aβ deposits, their curvature increased fourfold and their curvilinear length was increased by 31 percent. These measurements allow for an analysis of the functional consequences of such alterations using "Genesis" cable property computer simulations. The net result: a 30 percent delay in synaptic transmission. Considering the integration and timing necessary within large-scale neural networks, such delays within even a subset of dendrites could have serious impact on function.

What accounts for these changes in dendritic morphology? It appears that Aβ alone is not enough; dendrites in the non-demented brains were unaffected by plaques. The likely culprit is a loss of normal function in hyperphosphorylated tau. However, even Alz-50 negative dendrites exhibit abnormal morphology if they are within AD plaques (but not non-AD plaques). Are there other components of AD plaques which are absent from the plaques in non-demented brains (e.g., extracellular matrix components, complement proteins, etc.) which could account for the morphological changes? Another possibility is that Alz-50 is not sensitive enough to detect all tau hyperphosphorylation. Earlier markers such as AT8 and MC1 might reveal abnormal tau within plaque associated dendrites. Paul Coleman asked how to reconcile these data with previous reports that dendritic length in AD is decreased compared to controls. Knowles responded that these data were within-subject measures rather than comparisons between AD and control, while acknowledging that differential shrinkage between diseased brains and controls is also possible.

One also has to wonder whether changes in synaptic transmission in the dendrites within AD plaques would really affect processing. As dendritic spike timing became delayed, couldn't the aged brain compensate for these changes? Brad Hyman responds that these timing changes would be essentially random within a large network. While an individual neuron might compensate for increasing delays within one branch of a circuit, there would be no way for an entire network to compensate for small delays from multiple sites. So even if there are "healthy appearing" neurons within the AD brain, their dysfunction could be contributing to AD long before they become neurofibrillary tangles.—Brian Cummings