More candidate therapies have landed on the scrap heap of Alzheimer’s clinical trials, freeing up sites, investigators, and participants to try something else. At the 8th Clinical Trials on Alzheimer’s Disease conference, held November 5-7 in Barcelona, Spain, Rachelle Doody of Baylor College of Medicine in Houston closed the book on Roche’s monoamine oxidase B MAOB inhibitor sembragiline. Doody presented a straight-up intent-to-treat analysis of all 542 patients with probable Alzheimer’s who had entered a Phase 2 study called MAyflOwer RoAD. MAO B inhibition was being tried as a neuroprotective approach targeted to reactive astrocytes. Sembragiline is a safer, more selective follow-on compound to a previous MAO inhibitor, lazabemide, which inhibited both MAO A and B and was toxic to the liver, Doody said.
This trial was fairly standard, comparing two doses of drug to placebo for one year, plus 12-week follow-up. Twenty percent of the patients dropped out. The drug was safe but showed no difference between groups on its primary endpoint, the ADAS-cog, Doody showed. Because the observed drug concentrations in this study predict almost complete occupancy of the MAO-B enzyme in vivo at both doses, the hypothesis was tested. “Sembragiline at these doses does not slow cognitive decline in AD,” Doody said.
On one secondary endpoint, the Behavioral Pathologies in Alzheimer’s (BEHAVE-AD) scale, the drug did show a difference. The placebo group developed behavioral disturbances driven by a large minority of patients, Doody said. Another secondary, the ADCS-ADL scale of daily function, showed a trend but fell short of statistical significance. A post hoc analysis indicated that patients whose behavioral symptoms were more severe may have benefitted from sembragiline; however the study was not designed to assess behavioral symptoms and this signal would need to be reproduced, Doody said. Roche has discontinued development of sembragiline.
The axe also fell on a histamine H3 receptor antagonist by the French pharmaceutical company Laboratoires Servier. Called S 38093, this drug underwent two large Phase 2 trials, one as single treatment and one in addition to donepezil. They were conducted in 40 different countries, though not the United States. At CTAD, Servier’s Annette Merdes gave a postmortem on this program. Together, more than 1,300 patients with mild to moderate Alzheimer’s disease participated in these studies. The drug was reasonably safe, though treated patients did have more falls than people on placebo. However, neither of three doses—2, 5, or 20 mg/day—taken for a year, outdid the placebo on any measure of cognition, function, clinical global impression, or caregiver burden. This program, too, saw a trend for a benefit on neuropsychiatric symptoms in one trial, but that did not hold up in the companion study. “We saw no clinically relevant finding. The program has been stopped,” Merdes told the audience.
In a warning perhaps to the flurry of early stage trials that are ramping up, Merdes noted that further analysis of the placebo response in this trial indicated that the patients with the mildest disease, who had been diagnosed most recently, were most prone to demonstrating a strong placebo effect.
The end came for yet another treatment approach, the NGF gene therapy Cere-100. This attempt at neuroprotection had, in various incarnations, been undergoing clinical evaluation ever since Mark Tuszynski of the University of California, San Diego, started the line of research in 1999 (Dec 1999 news). Earlier this year, the company developing Cere-100 announced the end of the program, and at CTAD, UCSD’s Mike Rafii presented the data of the final study. This was a multicenter ADCS trial controlled by sham surgery, an invasive way of blinding a study that had generated controversy in the past (Kim et al., 2005).
Of 88 screened patients, 49 were enrolled and assigned 1:1 to Cere-110 or control surgery. Both the procedure and the AAV2 vector used to ferry in the NGF gene were safe, Rafii reported. Of the six deaths over the course of this two-year trial, four occurred in the placebo group and the other two were not related to study procedures. The problem, as so often, was efficacy: Two years after the surgery, the treated patients were no better off than placebo recipients on the ADAScog and ADCS-ADL; they were even trending worse than the placebo group on the MMSE and CDR-sb. FDG-PET and MRI volumetry were also not different between the groups, suggesting that the NGF expression, to the extent it occurred, was unable to boost the brain’s metabolism or slow neurodegeneration.
On a procedural note, the trial did show that it is feasible to conduct multicenter trials of controlled steretotactic surgery and that patients will enroll knowing they might end up with sham surgery. To maintain the blind, the study employed an unblinded surgery team and a separate, blinded neurology team, Rafii said. This implies that if the field revives gene therapy and new candidate therapies roll around, the necessary trials to evaluate them can be done. Applauding the bravery of patients in this study, Mary Sano of Mount Sinai School of Medicine, New York, agreed, saying “This seems safe and doable now.”
For CERE-110 itself, the epilogue in due time will come in the form of neuropathologic examination of how many cholinergic neurons remained in these patients, and whether NGF was expressed and adequately distributed in the brain.
Getting exhausted with all the flameouts? Almost there. Two more negative answers, then this Phase 2 story ends on a more positive note. Phase 2 is, after all, a better place for ineffective drugs to die than Phase 3. At CTAD, the “Alzheimer’s is Type 3 Diabetes” hypothesis posted two negative answers on therapeutic validation. For one, an investigator-initiated trial showed that the insulin analog detemir is ineffective in treating Alzheimer’s disease, at least in a four-month course tested in this trial. Suzanne Craft from Wake Forest University Medical School in Winston-Salem, North Carolina, told the audience that she compared the long-acting insulin analog to regular insulin and placebo, both delivered intranasally. In 11 AD patients per group, 20 international units of detemir twice daily had no effect on a memory Z score and MRI volumetry, but the same amount of regular insulin did show a signal on both measures. Meanwhile, the Alzheimer’s Therapeutic Research Institute at the University of Southern California is conducting the NIA-funded Phase 2/3 trial of intranasal regular insulin called SNIFF. This multicenter study had to stop temporarily over a problem with the insulin nebulizer, but is now back on track, Craft said.
Metformin, a potent and widely used insulin sensitizing drug, fell on its sword. Steven Arnold of the University of Pennsylvania, Philadelphia, reported that his study enrolled 20 people with mild to moderate AD, with about the same late-60s age distribution as Craft’s study. All completed an eight-week course of up to 4,000 mg/day of metformin. This was primarily a biomarker study, Arnold said. Metformin failed to budge readouts on resting cerebral perfusion as measured by ALS MRI, nor CSF Aβ, tau, or phospho-tau. Cognitive outcomes were similar, with neither the Montreal Cognitive Assessment (MOCA), the functional rating scale (FRS), or some other paper and pencil tests showing any change, Arnold said.
After five no-gos, here is a new contender. The serotonin receptor antagonist ITI-007 is slowly wending its way toward Phase 2. This drug is being developed primarily for schizophrenia, as a reportedly safer alternative to current antipsychotic drugs such as clozapine and benzodiazepines, both of which are prescribed off-label to treat disturbed behavior and sleeplessness in dementia, as well. ITI-007 is what used to be called a “dirty drug” in that it hits serotonergic, dopaminergic, and glutamatergic receptors, though its developers claim that its higher potency for 5-HT2A than the other receptors give it a wide dosing range that can be exploited to treat different conditions. In schizophrenia, doses of up to 60 mg per day have completed one Phase 3 study while another is ongoing, and two Phase 3 trials in bipolar disorder are about to start (see company press release; clinicaltrials.gov).
For an indication in Alzheimer’s, the company Intra-Cellular, which licensed ITI-007 from Bristol Myers Squibb, thus far has only conducted one small safety and pharmacokinetics study of up to 20 mg/day, which Robert Davis of Intra-Cellular previously reported at CTAD (see Dec 2014 conference news). In Barcelona, Davis merely added data from a small trial conducted in France in 19 people with insomnia, saying that doses as low as 1 to 10 mg/day helped people stay asleep, with longer slow-wave sleep in the first half of the night and longer stage 2 sleep in the second half. Poor sleep is increasingly being implicated both in increasing risk and worsening behavioral symptoms of Alzheimer’s disease. A Phase 2 trial of ITI-007 to soothe neuropsychiatric symptoms of Alzheimer’s is being planned to start in summer 2016, Davis said in Barcelona. “This drug is very safe. We think it will be useful to test behavioral disturbances in dementia, such as depression, anxiety, and aggression.” For Phase 1 results, see Part 5 of this series.—Gabrielle Strobel
- NGF Gene Therapy Trial Begins
- New Treatments for Alzheimer’s Behavioral Symptoms on Horizon
- Truly New to Déjà Vu: Phase 1 Means ‘Check!’ for Some, ‘Out!’ for Others
- Kim SY, Frank S, Holloway R, Zimmerman C, Wilson R, Kieburtz K. Science and ethics of sham surgery: a survey of Parkinson disease clinical researchers. Arch Neurol. 2005 Sep;62(9):1357-60. PubMed.
No Available Further Reading