"I had been a little skeptical about the choice of this antibody for the API trial," said Todd Golde, University of Florida, Gainesville, who has now seen the safety profile and dosing possibilities. "After seeing these data, I think it was a good choice." Crenezumab's selection was also based on Genentech's willingness to make the compound available for its study.

Scientists started developing passive antibodies after an active immunotherapy caused encephalitis in some Alzheimer's disease (AD) patients (see ARF related news story). Passive immunotherapies had their own side effects, however, most notably non-life-threatening microhemorrhages and vasogenic edema (also called amyloid-related imaging abnormalities, or ARIA), especially in people who carried the ApoE4 allele (see ARF related news story). These antibodies, including bapineuzumab, gantenerumab (see ARF related news story), and solanezumab, are in clinical trials, and Phase 3 results for bapineuzumab and solanezumab are scheduled to come out this year. All are built on the same immunoglobulin (IgG1) backbone, which the authors say activates microglia a bit too strongly, possibly causing ARIA.

Crenezumab differs from these, in that it possesses an IgG4 backbone—which has a milder effect on microglia because it binds more weakly to cell surface immunoglobulin receptors. Looking in mixed cell cultures from rats, first author Oskar Adolfsson of AC Immune used confocal microscopy to show that the antibody activated microglia just enough to engulf Aβ oligomers. But crenezumab did not initiate a major inflammatory response. P38MAP kinase, which starts a neurotoxic cascade when exposed to Aβ42 was activated, but no more than in cells treated with Aβ42 alone. Treatment with the same antibody on an IgG1 backbone activated this pathway much more. TNF-α, which is a downstream proinflammatory molecule in this pathway, also fell in the presence of crenezumab relative to the IgG1 antibody—more evidence that inflammation declined. Crenezumab still cleared plaques. In living hAPP(V717I)/PS1 transgenic mice injected with the antibody, multiple two-photon laser-scanning microscope images revealed that plaque volume declined over three weeks. The lack of a substantial inflammatory response is crucial for the success of this antibody, Watts said, because he and his colleagues believe that an inflammatory response causes the blood-brain barrier to break down and leads to vasogenic edema.

The lack of vasogenic edema lifts the limit on dosing imposed on other passive antibodies and allowed the researchers to administer more drug in Phase 1 trials, said Watts. A higher dose could mean the antibody will be more effective. Though efficacy data will come later, the researchers saw that in Phase 1 trials, plasma Aβ levels rose in a dose-dependent fashion, suggesting that the drug reached and broke up brain plaques for disposal in the blood. Further evidence of vasogenic edema was absent, even with the highest doses. "When you go into a prevention trial like that of the API, people are cognitively normal," he said. "You want to have as little risk as possible."

Without the prominent microglial response, plaque removal may be slower with crenezumab than with other antibodies, said Roy Weller, University of Southampton, U.K., and that could be a good thing. "It may be better than trying to eliminate amyloid all at once," he told Alzforum. The drug still has to prove its worth in humans, however. "The most positive results come from mouse studies," said Weller.

What does crenezumab bind to? In vitro ELISA tests showed that crenezumab bound to Aβ fibrils and oligomers, with a slightly lower affinity for monomers. The drug also bound to plaques in slices of brain from both hAPP(V717I) transgenic mice and humans with AD. Epitope mapping showed that the antibody binds to amino acids 12-23 of Aβ42, which contains the hydrophobic region responsible for aggregation. Crenezumab both prevented Aβ assembly and broke up aggregates in vitro as revealed by thioflavin T staining. The drug also rescued rat neurons from Aβ42 cytotoxicity, in that these cells did not show the same drop in mitochondrial metabolism seen in control cells exposed to Aβ42.

"I think it's an encouraging antibody and a very clever design—it certainly merits clinical evaluation, said Dave Morgan, University of South Florida, Tampa. Morgan is encouraged that the researchers do not see vasogenic edema and can deliver higher, potentially more effective doses than possible with other antibodies.

Donna Wilcock, University of Kentucky College of Medicine, Lexington, found the study intriguing but thought that more results are needed. For instance, data about the activation of both microglia and p38MAP kinase from the mouse brains, not just cultured cells, would have better solidified the mechanism of crenezumab's action in living animals. "More in-vivo studies have to be done to demonstrate that the drug is not going to initiate the adverse events we've seen with other antibodies," she told Alzforum.

Though antibody doses used were high in the Phase 1 trial and the researchers are relatively confident in crenezumab's safety, vasogenic edema may not have surfaced because of the trial’s small sample size. Therefore, Watts and colleagues continue to monitor blood-brain barrier integrity in current Phase 2 trials. According to Watts, the crenezumab studies will carry on regardless of the outcome of Phase 3 results for bapineuzumab and solanezumab. "We've uniquely engineered crenezumab around its safety profile," Watts said. (For a conversation with Watts about the API trial, see ARF related news story.)—Gwyneth Dickey Zakaib.

Reference:
Adolfsson O, Pihlgren M, Toni N, Varisco Y, Buccarello AL, Antoniello K, Lohmann S, Piorkowska K, Gafner V, Atwal JK, Maloney J, Chen M, Gogineni A, Weimer RM, Mortensen DL, Friesenhahn M, Ho C, Paul R, Pfeifer A, Muhs A, Watts RJ. An effector-reduced anti-β-amyloid (Aβ) antibody with unique Aβ binding properties promotes neuroprotection and glial engulfment of Aβ. J Neurosci 2012 July 11; 32(28):9677-9689. Abstract