Three years ago, however, Genentech made a strategic decision to re-enter the neuroscience research field. At an investor and analyst meeting in March 2008, CEO Art Levinson said that finding new drugs for diseases like Alzheimer's was a new, long-term priority for the company. Tessier-Lavigne is a neuroscientist recruited to join Genentech to help kick start this new research effort. The new understanding about APP came about as an accident, he says, during the course of research into the normal process of nerve cell degeneration that occurs as embryos develop in the womb. When the brain and spinal cord are first being formed, excess numbers of nerve cells are generated. Ultimately many of these nerve cells are killed off, or "pruned," says Tessier-Lavigne, so that only the right number exist to form into the brain, spinal column and other parts of the nervous system. Working on various cell culture experiments in the lab, Genentech researchers discovered something called death receptor 6 that is responsible for this normal embryonic self-destruction mechanism of nerve cells. But when Tessier-Lavigne and his team went looking for the substance that activated death receptor 6 to prune back embryonic nerve cells, they were shocked to learn that it was APP. "We almost fell out of our chairs," he says. The reason for their surprise is that APP has long been linked to Alzheimer's. The protein resides in nerve cells and is normally broken down by various enzymes into harmless fragments that are then disposed of by the body. But in some people these APP fragments aren't cleared from the body but instead form sticky clumps known as beta amyloid. Eventually beta amyloid forms into plaques that attach to and kill nerve cells in the brain, which is widely believed to be the cause of Alzheimer's.