08 Dec 1999

A newcomer has joined the ranks of suspects implicated in early Alzheimer's disease. In tomorrow's Nature, Li-Huei Tsai of Harvard Medical School and her colleagues report that the protein p25 deregulates the kinase cdk5, enabling it to hyperphosphorylate tau. This leads to neurofibrillary tangles and, perhaps, triggers apoptosis, Tsai argues. "We believe that the production and accumulation of p25 in the Alzheimer's brain very likely plays a role in the pathogenesis of the disease," says Tsai.

P25 is a proteolytic fragment of cdk5's normal regulator, p35, which controls cdk5 while the kinase plays its well-established role in the formation of the cerebral cortex during embryogenesis and its emerging role in neurotransmitter release in the adult. Yet p25 is a stable protein that occurs freely in the cytosol, whereas p35 is anchored to the neuron's membrane and gets degraded in the cell's proteasome, the authors show. These biochemical differences allow p25 to activate cdk5 constitutively, sending it on a "phosphorylation spree," in the words of Eckhard Mandelkow of the Max-Planck Institute in Hamburg. In an accompanying News and Views article, he calls Tsai's work a major advance in the search for the kinase that hyperphosphorylates tau.

Tsai and colleagues first show that p25 is accumulated 20- to 40-fold in the brains of AD patients, occurring in the same cortical neurons that also harbor NFTs. Tsai suggests that p25 contributes to the disease's early stages, in part because it appeared in more neurons than contained NFT, its amount increased with advancing disease, and it no longer existed in a brain sample from a terminal-stage patient whose brain had already shrunk to about half the normal weight. Transfection experiments in cultured COS-7 and primary rat cortical neurons showed that p25/cdk5 phosphorylates tau and induces the disintegration of microtubules. Transfected neurons then withdrew their processes.

The other contribution of this study lies in the observation that 90 percent of the neurons expressing p25/cdk5 were dead by apoptosis within three days, supporting the notion that hyperphosphorylation of tau can, in fact, induce neuronal death. Whether p25/cdk5 actually causes cytoskeletal damage and apoptosis in people with AD remains to be proven; for example, by testing whether inhibiting cdk5 or p25 can slow or reverse neurodegeneration in Alzheimer's animal models. Calling the study "extremely interesting," Bruce Yankner points out its thoroughness. "There are a lot of experiments in it, and every conclusion is very well substantiated," he adds. For her part, Tsai is now trying to finger the enzyme that cleaves p35, and to learn what conditions induce it.—Gabrielle Strobel