The cytoskeleton in Alzheimer disease

Summary

The strongest physical correlate with the severity of dementia in Alzheimer’s disease and its most rational cause are the loss of neocortical and hippocampal synapses. Evidence, showing that β-amyloid causes that loss is weak despite the popularity of that hypothesis. Other changes can better explain that damaging phenomenon. Axonal terminals are dependent on axoplasmic flow, and that function requires intact microtubules and the motor proteins kinesin, dynein and dynamin.

It has been known since the earliest electron microscopic studies of AD that neuronal microtubules are lessened in number. Tubules are normally in equilibrium with unpolymerized tubulin, and the stability of the formed elements is dependent on normal binding of tau to the tubule. But, as is well known, tau is abnormally hyperphosphorylated in AD leading to tangle formation and to dissolution of the tubules. Tangles are insufficient in number to account for the cortical loss of neurons and synapses, but hyperphosphorylated tau in the unpolymerized pre-tangle state undoubtedly plays a role. Abnormalities in the motor proteins are now being investigated (some have already been found) and these too would contribute to the loss of synapses in AD by way diminished axoplasmic flow.

Most investigators of Alzheimer disease (AD) seem confident that the cause of the clinical symptoms is the deposition of amyloid. Since the original isolation and sequencing of this specific peptide by Glenner and Wong (1984), about 3,000 papers have been published concerning it. Such a torrent of information is hard to resist, but there are a number of flaws in the argument that this amyloid is the direct cause of the dementia. Among them is the fact that amyloid inoculated into the brain of the monkey does not induce the Alzheimer-like changes (Podlisny et aI., 1993). A cognitively normal elderly human may have a high concentration of cortical amyloid. There is only poor correlation between the amount of amyloid in the human brain and the severity of the disease (Arriagada et aI., 1992), except in an area such as the entorhinal region (Cummings et aI., 1995) where plaques are clearly a secondary feature. Since the amyloid burden does not correlate with the severity of the disease, one must ask whether pharmacologic reduction of that burden will be clinically useful.

A number of other pathogenic theories have arisen including abnormal calcium influx in neurons (Mattson et a1., 1992), excitotoxic neuronal damage (Cowburn et a1., 1988), abnormal oxidative stress (Bea1 et a1., 1994), physiologic (corticosteroida1) stress (Gould et a1., 1991), and inflammation (McGeer et a1., 1994). There is indeed evidence for each of these possibilities, but none is very strong, at least in terms of primary causation.

A very strong correlate with the clinical severity is synaptic loss in neocortex (Scheff et a1., 1990; Terry et a1., 1991) and hippocampus (Sze et a1., 1997). The loss of synapses causes disconnection (Hof et a1., 1994) among the many functional parts of the brain, and so this is a biologically rational cause of dementia. It is important to realize that a neuron in the central nervous system can lose many of its axonal terminals but survive as an apparently intact soma. This fact explains why cell counts correlate so poorly with function in some degenerative diseases, and how there is greater synaptic loss than neuronal loss in the Alzheimer cortex (Scheff et a1., 1990; Terry et a1., 1991). It also might provide an optimistic view that the synaptic terminals might regenerate given the appropriate stimuli and circumstances. The question then is what causes the loss of synapses, and this writer would have it that synapses die and neurites become dystrophic largely because of abnormal and deficient axoplasmic flow (Suzuki et a1., 1967). Techniques to measure this function in the human are not yet available.

Axop1asmic flow is that essential neural mechanism by which substrates and organelles are moved from the cell body out through both axons and dendrites to their endings, and by which other products are returned from terminals to the cell body. Essential to this process are intact microtubu1es as is proven by the fact that dissolution of microtubu1es (as by colchicine) arrests the process. The protein kinesin provides the motive force for rapid transport along the tubules in the anterograde direction (Vale et a1., 1985), while dynein is the motor for retrograde movement (Paschal et a1., 1987). Both motor proteins gain energy by hydrolyzing ATP, and both must act along intact micro tubules (Fig. 1).

Electron microscopic studies of cortical biopsies performed more than 30 years ago first demonstrated a paucity of microtubu1es in the neurons of AD (Terry et al., 1964). This is true of cells both with and without tangles. Microtubules are in equilibrium with their component monomers alpha and beta tubu1in, so that total tubulin might be normal in a cell although polymerized functional tubules are diminished. In such a case, axop1asmic flow would be deficient.

The effects of synaptic loss, itself due to deficient axop1asmic flow, are multiple. First, there will be disconnection between the several coordinated parts of the brain and thus dementia (Hof et al., 1994). Second, the degenerative synapses will elicit microglial activation, and their cytokines cause further damage (McGeer et al., 1994). Third, the neuron will receive less trophic factor from the target areas since the terminals are not available to pick it up and the retrograde axop1asmic flow is inadequate to transport the factors back to the cell body. Therefore, the soma will ultimately die by way of apoptosis (Cotman et a1., 1994) secondary to its synaptic loss (Fig. 2).