Genentech Scientists Report Antibody Delivery Into Brain
From BioWorld Today (May 26, 2011)
A Long Shot for Alzheimer's Hits its Target
In back-to-back papers, scientists from Genentech Inc. reported two potential advances in the fight against Alzheimer's disease: They have managed to make an antibody that inhibits beta-secretase and to deliver that antibody into the brain to lower amyloid beta.
Ryan Watts told BioWorld Today that of the two advances, "by far the greater challenge was getting the antibody into the brain." But now that it works, at least preclinically, the method ?? a bispecific antibody whose second binding site targets an iron receptor in the blood-brain barrier's ?? should have broad applicability to brain disorders.
The therapeutic end of the antibody targets "one of the most popular Alzheimer's drug targets: beta-secretase," said Watts, who is associate director of neuroscience at Genentech and a co-author on both papers. The enzyme is the first step in processing amyloid precursor protein to amyloid-beta, which is a likely underlying culprit in the dementia that plagues Alzheimer's victims. "Many companies have gone after BACE1 with small molecules," he added.
But such inhibitors have been hard to get into the brain, and so despite the fact that BACE1 has been in industry's sights for more than a decade by now, Phase I trial results are only now emerging. Gamma secretase inhibitors, which interfere with the next step in amyloid processing, are further along, but have been unsuccessful to date. So far, the conclusion from unsuccessful trials such as Eli Lilly's semagacestat has been that what's needed are better inhibitors, rather than novel targets. (See BioWorld Today, Aug. 18, 2010, and Sept. 2, 2010.)
Watts and his team worked with an antibody that they identified by screening a synthetic antibody library. In cell culture, the antibody was highly specific to beta-secretase 1, and the researchers found that their antibody bound noncompetitively. Asked whether that was an advantage or a disadvantage compared to competitive binding, Watts said that it is "unique" among BACE inhibitors, adding that the noncompetitive binding is probably what accounts for the antibody's specificity.
The antibody also reduced the concentrations of amyloid beta in mice and monkeys. (It was ineffective in one specific mouse model, probably because the APP is processed intracellularly, before the antibody can get to it at the cell surface.)
But at first, there was a rub. The antibody, Watts said, was "very effective at reducing peripheral levels of BACE activity . . . however, the brain reductions were very modest."
So Genentech antibody engineer Mark Dennis and his team, somewhat paradoxically, decreased the antibody's binding strength for the receptor that was meant to be its ticket into the brain: the transferrin receptor.
The transferrin receptor's regular job is to transport iron into the brain. It has been eyed by more than one group as a way through the notoriously hard-to-breach blood-brain barrier, which keeps most molecules ?? and certainly ones that are the size of antibodies ?? out of the brain.
But Watts likened the transferrin receptor to a ski lift: Too-tightly bound proteins "remain on the ski lift," rather than getting off and entering the brain.
Dennis said that even if the antibody hops off the ski lift at the top of the slope ?? or in this case, the inside of the blood-brain barrier ?? that in itself is not sufficient for it to do any good. Antibodies could hitch a ride back out before spreading out, or simply "get stuck in the blood vessels," he told BioWorld Today.
The Genentech team developed a bispecific antibody that ultimately mastered the fine art of letting go. After injection, compared to high-affinity antibodies the lower-affinity version showed higher uptake into the brain, "broad distribution" within the brain after uptake and was localized near neurons.
Though the team is "eagerly pursuing" the Alzheimer's antibody, and contended that the transferring antibody will prove useful for other diseases as well, Watts cautioned that "this is a proof-of-concept study in a mouse model. We're thrilled with the data and the concept, but a lot of work remains to be done to turn this into therapies."
Still, he added, when the first data came in showing that the antibody did not just get into the brain, but spread out and found its target once it was there, "it was probably the most exciting day I've had . . . in this study."
Posted: May 2011