After a disappointing interim analysis, Roche and its collaborator AC Immune are halting two phase 3 trials of the antiamyloid antibody crenezumab.
Dr. Richard J. Caselli
CREAD 1 and CREAD 2 enrolled patients with prodromal-to-mild sporadic Alzheimer’s disease. The preplanned interim safety and efficacy analysis determined that neither study was likely to meet the primary endpoint of change from baseline on the Clinical Dementia Rating-sum of boxes score.
There were no unexpected safety signals associated with the drug, despite a quadrupling of the phase 3 dose from that used in phase 2. The company in its press release said that it will continue to conduct the Autosomal Dominant Alzheimer’s Disease (ADAD) trial as part of the Alzheimer’s Prevention Initiative (API). ADAD is a large South American trial of crenezumab in Colombian families with familial Alzheimer’s caused by mutations in the presenilin-1 gene (PSEN1).
Roche did not release any data but said the trial results will be discussed at an upcoming scientific meeting.
“While the results with crenezumab are disappointing, they meaningfully contribute to our understanding of Alzheimer’s disease,” Sandra Horning, MD, Roche’s chief medical officer and executive vice president for global development, said in an interview. “We gratefully acknowledge the participants in the CREAD trials and the efforts of everyone involved in this important program.”
The decision was not a surprise to researchers who have followed the antibody’s development. It advanced into phase 3 with lackluster phase 2 cognitive, imaging, and biomarker data. Its selection as the therapeutic agent for the ADAD trial was a key driver in its continued development, securing Roche $100 million in federal funds to help launch ADAD, the first-ever Alzheimer’s primary prevention study.
Despite its failure in sporadic Alzheimer’s, there is still some hope that crenezumab might benefit people with the PSEN1 mutation, said Richard Caselli, MD, professor of neurology at the Mayo Clinic Arizona in Scottsdale and associate director and clinical core director of the Arizona Alzheimer’s Disease Center.
“The Colombian trial is aimed at dominantly-inherited AD due to a PSEN1 mutation, so it is different enough to imagine it still might make a difference in patients in whom amyloid metabolism is actually defective due to functionally altered amyloid precursor protein or gamma secretase,” he said in an interview. “Possibly some might argue that many of the patients in the crenezumab trial likely had additional pathologies so that even if the AD component responded, the overall clinical picture might not reflect it due to the other components. That would be interesting if proven and could even argue against equating young-onset with late-onset AD, at least for clinical purposes, as is currently envisioned.”
Michael Wolfe, PhD, had a different take on the matter.
“Although amyloid-beta [Abeta] production is not necessarily altered in sporadic AD, there is essentially the same pathology, presentation, and progression with familial and sporadic AD, suggesting a common molecular mechanism,” said Dr. Wolfe, who is the Mathias P. Mertes Professor of Medicinal Chemistry at the University of Kansas, Lawrence. “It’s hard to say Abeta is the pathogenic species in familial but not sporadic AD.
Dr. Michael S. Wolfe
“To me, the failures of the antiamyloid approaches are because the drugs are given too late, are targeting the wrong form of Abeta, or are targeting an enzyme [for example, beta secretase1] that has other important functions. Most likely it’s a combination of these reasons. One could argue that even if some form of Abeta is the pathogenic entity, it is not a practical target because intervention may need to be initiated many years before the onset of symptoms.”
Despite the long string of failed antiamyloid antibodies, it’s not yet time to give up on the approach, said James Kupiec, MD, chief medical officer at ProMIS Neurosciences of Toronto.
“I understand where the pessimism [around antiamyloid antibodies] is coming from, and I also understand the enthusiasm from these companies to pursue them,” said Dr. Kupiec, who formerly headed Pfizer’s neuroscience research unit. “Targeting plaque is clearly not going to do the job. But in my opinion, the deeper pathophysiologic questions have not been adequately addressed. I’m not willing to throw in the towel. The correct molecular species [of amyloid] has not been appropriately or adequately tested in studies with monoclonal antibodies.”
The antibodies that have been failing for 5 years now were designed in the early 2000s, Dr. Kupiec pointed out, when knowledge of the various amyloid species was still immature. Newer candidates can target specific conformations of the protein – monomers and oligomers – before they aggregate into insoluble sheets. “Solanezumab was the first of these, paving the way for this new generation of antibodies,” Dr. Kupiec said.
Because they target soluble Abeta, not amyloid plaques, these domain-specific antibodies are less likely to elicit ARIA (amyloid-related imaging abnormalities), the inflammatory reaction that’s been associated with plaque dissolution in other antibody trials. ARIA has been a dose-limiting step for antiamyloid antibodies – one that conformationally targeted antibodies could avoid, Dr. Kupiec said.
“There may be some limited success with the these, and there may be enough of a treatment effect to secure approval,” he said. “The question is: Can we generate a higher effect size with an antibody that is more selective to the toxic forms of Abeta?”
PMN310 is ProMIS’ attempt to thread this needle. In preclinical studies, the antibody did not bind to amyloid monomers, plaques, or vascular Abeta aggregates. The company expects to take this antibody into phase 1 trials later this year.
“If we have a molecule that doesn’t bind to monomers or to plaques, but only to the toxic oligomer, then that is an something well worth testing in the clinic,” he said.
Dr. Caselli and Dr. Wolfe have no financial disclosures.