The cascade hypothesis also may explain why therapeutic strategies that target amyloid have been unsuccessful, Jagust says: By the time people have lots of plaques, they are in late stages of the disease. So many neurons have been damaged by amyloid and tau (and perhaps other unknown processes) that breaking up plaques at this stage does not slow the disease. Amyloid therapies given at an earlier stage may work, but those studies have not been done yet.
But though the sequence of events described above is a working hypothesis for many scientists, it is not yet a consensus view. Some believe that early tau changes in a brain area called the entorhinal cortex (which were shown to precede measurable amyloid changes in a 1997 study) may be the initiating event. It goes to show that no one knows what the first lesion is, says Dr. Gary Small, who directs the UCLA Center on Aging.
Whatever the case, it's becoming clear that both abnormal protein structures start building up in the brain many years before major symptoms crop up. Small, who directs the UCLA Center on Aging, has developed a PET scanning method that labels both amyloid and tau problems in patients' brains. In a study presented in December at a professional meeting, his team scanned 42 volunteers with normal cognition, then followed them for two years.
Subjects whose initial scans showed more plaques and tangles suffered more cognitive decline over two years. In addition, rescanning their brains revealed increased brain pathology.
"It correlates extremely well with disease progression," Small says of his diagnostic method.
Some promising results with tau therapies in animal models of Alzheimer's disease were reported at the Honolulu conference. Researchers are still years away from knowing whether the strategy will work in patients, however. Meanwhile, some two dozen other approaches are being tested in clinical trials, including more amyloid-targeted therapies (some of which aim to block formation of the protein) and drugs that combat inflammation in the brain.