25 July 2012. At the Alzheimer’s Association International Conference held 14-19 July 2012 in Vancouver, Canada, Gabrielle Strobel met up with Luca Santarelli, Senior Vice President and Head of Neuroscience at Hoffmann-La Roche. Santarelli is perhaps best known for his work on depression—while a young scientist at Columbia University, he discovered that antidepressants such as Prozac work by inducing neurogenesis (Santarelli et al., 2003). But Santarelli moved into drug discovery research at Roche/Genentech in 2005 and now oversees the Alzheimer’s therapy development program, among others, at the parent company in Basel, Switzerland. Below are excerpts from the conversation, updated after news on the first bapineuzumab trial broke on 23 July.
Q: There is concern in the Alzheimer’s disease research community about pharma getting out of CNS, particularly AD. What is Roche’s position?
A: We are very much into this area. We have a large CNS program. For example, we have gantenerumab in Phase 3. It is the first monoclonal antibody targeting prodromal AD. The study’s name is SCarlet RoAD. It is in the middle of recruiting 770 patients and will complete recruitment next year. [Editor’s note: See on ClinicalTrials.gov.] The strategy is to go early in the course of disease where we hope to achieve most benefit.
Q: What else is different about it?
A: A significant percentage of patients involved in AD studies, even at the mild to moderate stage, do not display amyloid pathology. This diagnostic inaccuracy almost certainly adds noise to these trials. In SCarlet RoAD we are identifying prodromal AD patients by deploying a molecular diagnostic strategy. In this case, we use CSF Aβ along with cognitive testing to identify the patients with underlying AD pathology. For the current trial, we use a commercially available research-grade test, but we are developing a companion diagnostic assay test for future use.
Q: What makes you confident that gantenerumab will work?
A: This compound has published results showing that it reduces brain amyloid three times faster than other published amyloid-lowering agents. [Editor’s note: See Ostrowitzki et al., 2011.] The antibody binds to aggregated Aβ in fibrils and plaques. It is fully human, and the trial uses a subcutaneous formulation. If positive, the study could be pivotal.
Q: Is this Phase 2/3 going to be enough for a New Drug Application?
A: Confirmation in a separate trial is often required, but that depends on the effect size of your pivotal trial, and on the perceived medical need and other regulatory considerations at the time the data become available. We are now assessing what additional trial will be appropriate.
Q: How are the talks with the FDA going?
A: In our interaction with the FDA, we have experienced a collaborative spirit, where the agency is open to new ideas and willing to give us helpful advice on how to make progress on these novel trial designs. For example, the SCarlet RoAD trial does not use ADAS-cog. We use CDR-SOB; it is a single primary that is a global measure with some functional components.
Q: To read the trade and general media, you’d think the entire world turns on the results of the bapineuzumab and solanezumab Phase 3 trials that will read out in the next few months. Indeed, the first of four bapineuzumab trials on 23 July reported negative topline results in ApoE4 carriers (see ARF related news story). There’s considerable worry for the future of anti-amyloid drug development if these Phase 3 trials show no clinical significance. What do you think can be reasonably expected from them?
A: It is hard to predict the outcome of the additional Phase 3 trials. We will have to be patient and wait for the results. Hopefully, the outcome of the remaining trials will help provide support to the amyloid hypothesis. I would add, however, that in AD it is important to use doses that have previously demonstrated amyloid removal in patients; patients should be included early in their illness, at the prodromal stage or even earlier; and biomarkers indicative of Aβ amyloid pathology should be used to confirm the diagnosis before randomization.
Q: Venture out and make a prediction, will you?
A: I see three scenarios. 1) Bapineuzumab meets the endpoint in the ApoE4 non-carriers and the trials form basis for approval. That seems optimistic. 2) We get a mixed result with a signal but no basis for approval. That seems more likely. 3) The program is totally negative. The first two are good news for the field. The last would be a setback. But there are other approaches, like going earlier, that might yield good results and that are getting underway. The field will have to wait longer to gain confidence. Incidentally, the intense attention focused on bapineuzumab reflects also some of the challenges of charting a new path.
Q: Will Roche’s approach and investment into gantenerumab change if bapineuzumab and/or solanezumab were to be negative?
A: We will continue. As far as gantenerumab, we have targeted a different population, and there is a fundamental issue when targeting this disease late. [Editor’s note: Roche’s internal communication rules prevent Santarelli from commenting on Genentech’s crenezumab, which is being tested in patients with mild to moderate AD; see the ABBY Trial. See also ARF related news story.]
Q: The gantenerumab trial is viewed as innovative because it enrolled a prodromal population based on memory impairment short of dementia and a CSF Aβ cutoff. How much confidence do you have in the CSF biomarker assays that are currently available? There is discussion about their center-to-center and lot-to-lot variability, and none of them are FDA validated.
A: I am very confident in them. Just think about the mild to moderate trials. These studies are using clinical criteria to include patients, which means an estimated 20 to 30 percent of patients randomized in these trials do not have amyloid pathology. That was true—some 20 percent amyloid-negative people in the AD group—even in ADNI. Even with their limitations, the current assays do increase the homogeneity of your group considerably. You have 90 percent of people who truly have prodromal AD because you filter the clinical diagnosis through a biomarker. The ADNI and work from multiple longitudinal studies worldwide have really shown that. Some people think that accuracy of diagnosis may be even more important to show treatment benefit than early stage.
Q: There is great emphasis in the field on standardizing biomarkers to where they are robust and stable for use in multicenter trials and clinical care. The FDA issued a Guidance on Companion Diagnostics last year, but the concept is not broadly known among the wider research community. Roche being in part a diagnostics company, how do you approach biomarker development for your AD portfolio?
A: For us it is important to identify a way to diagnose AD early that is reproducible, high quality, and can become a gold standard for hospitals everywhere in world. That is how we approach this. Being both a pharma and a diagnostic company, we have a unique opportunity of developing a diagnostic and using it to evaluate the efficacy of our compound. Our strategy for AD is now focused on CSF, but we are also looking very actively in plasma.
Q: Have you found anything in plasma yet?
A: Nothing definitive yet. We are working with a number of companies.
Q: Besides gantenerumab and crenezumab, what else does Roche have in its pipeline for neurodegeneration?
A: We have an oral BACE1 inhibitor in Phase 1. There, we are in a pack with others. We also have a monoamine oxidase B (MAO-B) inhibitor in a Phase 2b dose-ranging study. This is a compound that targets the inflammatory aspect of age-related neurodegeneration. We have internal data showing that this mechanism is beneficial for AD. The inhibitor’s properties are linked to reducing the release of reactive oxygen species. This compound was discovered by Roche, then taken over by Evotec, which conducted trials for a different indication, and more recently, we took it back to evaluate it in AD.
Q: How about preclinical?
A: In the preclinical stage, we are active on several different fronts on AD. For example, we have two monoclonal antibodies that target tau, one internally developed at Roche, and one discovered at AC Immune. Additionally, we are developing a series of molecules that are highly brain penetrant. We use a technique we call brain shuttle. [Editor’s note: See ARF related news story.] The tau antibody is one molecule we are developing that way. Tau is thought to be released from cells and from there to spread to other cells, so you will need a technology that allows you to get a high amount of therapeutic antibody in, more than the 0.5 percent or so that we have seen so far. Brain shuttle has demonstrated the potential to greatly increase brain antibody concentration.
Q: What about combination trials? Researchers increasingly say it may be too much to ask of any single treatment to dramatically improve symptomatic AD. The FDA issued a guidance already in 2010 calling for trials of two unapproved drugs simultaneously in serious diseases including cancer and AD. You could combine an amyloid antibody and a BACE1 inhibitor, an antibody and your MAO-B inhibitor, or an amyloid and a tau antibody. Pipedream? In the works? Is Roche interested?
A: We were just discussing whether to do this for gantenerumab and our BACE inhibitor, in fact. There you’d have synergy between clearance and production. It should be possible. The problem with this example is that there is a time delay, so we may have to gather more data on the BACE inhibitor before we take that tack. We have that approach in oncology. For AD, it’s not out of the question at all.