Just after the first A4 trial volunteers were screened in San Diego on February 28, study leader Reisa Sperling, Brigham and Women's Hospital, Boston, was already laying out her vision for A5, A6, and even A7 trials at a Keystone symposium in Colorado held March 2-7. Called Alzheimer's Disease—From Fundamental Insights to Light at the End of the Translational Tunnel, the meeting ran in parallel with the first-ever Keystone symposium devoted to PD, called Parkinson's Disease: Genetics, Mechanisms, and Therapeutics. A panel on combination therapy generated the most animated discussion at this joint meeting. At issue was whether those future trials will combine two different treatments, as Sperling hopes. The conversation exposed how difficult it might be to test combination therapies in the current climate of negative AD trials. While some academics believe combination trials should begin right away, colleagues in industry are less enthusiastic, seeing all manner of obstacles.
Combination therapy is hardly a new concept in AD. Last year, at the urging of Rusty Katz, who was then at the U.S. Food and Drug Administration, the Washington-based advocacy group ACT-AD and the Critical Path Institute, Tucson, Arizona, hosted a meeting on the topic. The upshot was that the FDA encourages combination trials for AD, and many in the field believe the climate is ripe and the tools are in hand to try (see Feb 2013 news story). At Keystone, Dennis Selkoe, also from Brigham and Women's, convened a panel with Sperling, Ryan Watts from Genentech, San Francisco, and Charles Albright and Michael Ahlijanian of Bristol-Myers Squibb, Wallingford, Connecticut, to gauge support for the idea and discuss how it could be realized. Participants decided to forgo the customary Keystone sojourn to the ski slope, instead packing this mid-afternoon session and peppering it with comments.
Researchers from industry, both on the panel and in the audience, broadly agreed that two or more pharmaceutical companies would willingly collaborate to run such a trial. There are precedents for conducting combination trials, for example in cancer and HIV. "Collaboration is the least of our problems," noted Albright. The issue is not if this can be done, but whether it should be, he said. Albright claimed that because no single agent has yet proven successful in AD, putting two drugs together would amount to a Hail Mary pass: If it failed, it could cripple investment in AD drug development. The risk tolerance of investors in AD has reached an all-time low even for monotherapy, said Albright. "That we might do something even more risky, considering the drug-dosing and safety hurdles that need to be crossed, seems unrealistic," he said. "We need to focus on succeeding with a single agent so that we can spur more investment in this arena, not less."
Watts echoed some of those concerns, conceding that thinking about dose-ranging alone was enough to give him a headache. He reminded the audience that other combination trials, in cancer for example, tended to be conducted on drugs that were first developed individually. "If you see no efficacy unless you combine, then that is a bigger challenge," he said. "Without evidence of clinical efficacy, the risk is extremely high to test two unproven new molecular entities."
Ahlijanian agreed that dosing could be problematic, and cautioned that the safety and tolerability issues may be complex. He said that AD patients can react differently than normal individuals. "Combining drugs that have unknown effects in a patient population may be fraught with danger," he said.
Others worried about regulatory approval and reimbursement decisions that might follow on from combination trials, specifically when trying to determine which of the agents was responsible for any positive effects.
For her part, Sperling considered many of these concerns surmountable. She said the major rationale for a combination trial is that AD represents a combination of pathologies. "We know it is a complex disease, so the idea that targeting a single mechanism will give us a home run is unrealistic," she said. She noted that tackling different forms of Aβ may be as important as targeting Aβ and another entity, such as tau. Sperling argued that combination trials should start now because the length of AD trials means that waiting for each therapy to show efficacy means another decade, and that is too long. She suggested 2 x 2 designs, where drugs can be tested as monotherapies and in combination in the same trial.
Selkoe put it in practical terms. "If we chose solanezumab, which arguably has shown the best effect so far, and decided to go with the safest and most effective dose, would it be irrational to add a BACE inhibitor or an anti-tau antibody?" he asked. Could we envision doing this in mice and in humans in a year?
Watts was even more direct, asking "Merck, are you guys ready to combine your BACE inhibitor with Lilly's antibody?" Following a brief pause, Merck’s Mark Forman—acknowledging that such a decision would be made higher up the pay scale—replied that this seems feasible in animals, but insisted that in the clinic it makes more sense to first see how BACE inhibitors work alone. "We would argue that we have a level of Aβ reduction that is unprecedented and should test [our inhibitor] as a monotherapy," he said. Lilly scientists agreed that combination testing in animals is conceivable in principle.
There remain many issues to iron out, even for preclinical modelling. As in previous discussions of AD combination trials, proponents recommended adaptive trial designs such as those that have been used in cancer trials, while others asked which outcome measures could be used. "To adapt a trial based on cognitive change would take too long," said Sperling. She suggested biomarkers, but Albright noted that none of those have robustly shifted in trials in response to drug. "We have diagnostic and pharmacodynamic markers, but no surrogate markers. Do we have an outcome that we can adapt on?" he asked.
Dennis Dickson, Mayo Clinic, Jacksonville, Florida, advised that the heterogeneity of the disease be taken into account lest trials be underpowered. Selkoe agreed, noting that a substantial portion of participants in the solanezumab trial turned out to be amyloid negative or not have AD, and those who did may have multiple pathologies. Selkoe suggested correlating soluble oligomers in the CSF with disease and phenotyping participants in other ways, including tau imaging. Sperling agreed that endophenotyping would be valuable and said more imaging and biomarker analysis should be part of any new trial. For example, some patients in the A4 trial will undergo brain imaging for tau (see below). Watts said one reason why Genentech opted for the API trial was because its cohort is more homogenous. "We can come out with negative trials time and again, when really the drug has worked for some of those volunteers," he said
Others wondered what drug combination to use and when to begin. Some suggested using a BACE inhibitor and solanezumab. Albright questioned what that would achieve, since both therapies essentially target the same molecule, soluble Aβ. Sperling said she'd combine a BACE inhibitor with an antibody that targets an aggregated form of Aβ, or tau. Watts found that idea compelling, and suggested targeting inflammation in combination with Aβ therapies. He hinted that bi-functional molecules that could be developed as a single drug might circumvent some of the problems associated with combination therapy.
LEARNing from A4
Researchers hope to learn much from upcoming secondary prevention trials. Sperling announced that A4 has now recruited a site in Melbourne, Australia, bringing the total number of centers to 60. Some of the A4 patients will undergo tau PET imaging, a first in any AD trial. The Alzheimer's Association recently announced an $8 million grant for a so-called LEARN arm of A4. LEARN, which stands for Longitudinal Evaluation of Amyloid Risk and Neurodegeneration, is to track cognitive and clinical change in volunteers who test negative in PET scans for brain Aβ accumulation at a screening visit for A4. The grant also funds tau imaging in 100 participants who do have brain amyloid (50 in the treatment arm and 50 in the placebo group), and a further 50 in the amyloid-negative control group. Sperling told Alzforum that which tau ligand will be used has yet to be determined. "We have most experience with T807, but we want to consider all agents," she said. Sperling and colleagues describe the A4 trial in a Focus piece in the March 19 Science Translational Medicine.
A4 will run for three years, at which point Sperling hopes to have A5 in place. For that, she would like to test a BACE inhibitor. For A6, Sperling has previously proposed Combination Therapy in Early AD, aka COMBAT, a trial that could test both a secretase inhibitor and an antibody. She told Alzforum that a combination of an anti-amyloid and anti-tau therapy would be preferable. Subjects in the LEARN trial would be ideally suited to transition into one of these subsequent trials, depending on whether they showed signs of developing Aβ pathology and/or more neurofibrillary tangles. “We have always wondered how we can identify those who are just about to show signs of tau spreading from the medial temporal lobe. Now A4 can perhaps give us that information,” Sperling said.—Tom Fagan.
No Available Further Reading