Aluminum modifies the properties of Alzheimer's disease PHF tau proteins in vivo and in vitro.

Hyperphosphorylated adult human CNS tau (PHF tau or A68) forms paired helical filaments (PHFs) in neurofibrillary tangles (NFTs), neuropil threads, and dystrophic neurites associated with senile plaques (SPs) during the progression of Alzheimer's disease (AD). While amyloid fibrils in SPs are composed of beta-amyloid (A beta), NFTs and SPs contain similar associated components such as ubiquitin, alpha 1-antichymotrypsin (ACT), apolipoprotein E (ApoE), heparan sulfate proteoglycans (HSPGs), and aluminum salts. Thus, SPs and NFTs may result from specific interactions between PHF tau, A beta, and these other components. In fact, intracerebral injections of PHF tau induce co-deposits of A beta, ACT, and ubiquitin (Shin et al., 1993). To examine this issue further, we probed interactions between PHF tau, aluminum salts, and other plaque and tangle components. We investigated in vivo interactions of PHF tau and aluminum chloride (AlCl3) with other plaque and tangle components by injecting PHF tau with and without AlCl3 into the rodent brain. PHF tau co-injected with AlCl3 formed aggregates that persisted much longer in the rat brain, and induced longer-lived co-deposits of A beta, ubiquitin, ACT, and ApoE than PHF tau alone. Injections of PHF tau with AlCl3 also induced neurons near the injection site to acquire PHF tau-like properties as monitored with antibodies (AT8, T3P, PHF1) that recognize defined PHF tau epitopes containing phosphoserine residues (Ser202, Ser396, Ser404). Injections of AlCl3 alone as well as injections of normal adult and fetal CNS tau, several different synthetic peptides, neurofilament proteins, ACT, HSPGs, or ApoE with and without AlCl3 failed to induce co-deposits of A beta or alter the immunoreactivity of tau in rodent neurons. To determine if aluminum salts interact directly and specifically with PHF tau in situ, we pretreated sections of AD hippocampus with 10 mM AlCl3 and then probed these sections by immunohistochemistry with antibodies to PHF tau as well as to a number of other plaque and tangle components. Preincubation of these sections with AlCl3 diminished PHF tau immunoreactivity in NFTs and SPs using the PHF tau-specific antibodies AT8, T3P, and PHF1, while the immunoreactivity of other plaque and tangle proteins (A beta, ubiquitin, ACT, HSPGs, ApoE) was not abolished. We also examined the effects of AlCl3 on PHF tau and normal adult human CNS tau in vitro. AlCl3 had no effect on normal adult human CNS tau, while increasing concentrations of AlCl3 (from 0.1 to 1.0 mM) induced PHF tau to aggregate at the top of the stacking gel, and at high concentrations (0.3 and 1.0 mM) of AlCl3, PHF tau completely failed to enter the gel. These studies suggest that aluminum binds to PHF tau, induces these proteins to aggregate, and retards their proteolysis. Further, since intracerebral injections of PHF tau with and without AlCl3 in rats appear uniquely capable of inducing co-deposits of a number of proteins found in authentic AD SPs and NFTs (including A beta, ubiquitin, ACT, and ApoE), we speculate that the contributions of PHF tau to plaque and tangle formation in AD may be modulated by aluminum.