Of the 199 windows on the wavy walls of the Lou Ruvo Center for Brain Health in Las Vegas, no two are on the same plane. A specialized plaster coating on the walls ensures that the acoustics ring true amid this architectural free-for-all. The setting uncannily echoed the vibe at Opportunities and Challenges in Combination Therapy for Alzheimer’s Disease, an advisory panel meeting that took place at the Frank Gehry-designed building on May 28. Twenty-one clinicians, computational modelers, regulators, and leaders from foundations and pharma gathered to chart a path for using repurposed drugs in combination trials for Alzheimer’s. In 13 presentations and a two-hour discussion, they debated how drugs should be selected and whether combo trials were even ready for prime time. In the end, the panel concluded that combination therapy, in one form or another, may be needed to truly make a mark on AD. Leaders from the Alzheimer’s Association and the Alzheimer’s Drug Discovery Foundation (ADDF) agreed to request proposals for combination therapies on repurposed drugs, with details to come.

Combinations Among the Chaos.

The Cleveland Clinic at the Lou Ruvo Center for Brain Health, designed by architect Frank Gehry, made a dramatic setting for the combination therapy advisory panel. The interior is pictured above. [Image credit: Cygnusloop99, Wikimedia Commons.]

A push for combination trials has emerged at conferences and formed the basis of several small meetings in recent years (e.g., Feb 2013 conference news). The advisory panel in Vegas had the specific goal of discussing combination trials using repurposed drugs, in particular drugs aimed at targets other than amyloid. The meeting was co-hosted by Howard Fillit of the ADDF in New York and Jeffrey Cummings of the Cleveland Clinic in Las Vegas, who respectively fund and run AD trials on drugs already approved for other conditions. Because these drugs have passed safety hurdles, they could be used right away if they also work for AD, Fillit told the group. The focus on non-amyloid therapies is purely a function of efficiency, he said. Because pharmaceutical companies already pour millions into amyloid-based therapies, it makes sense for research foundations, clinics, and academic and philanthropic institutions to focus their efforts elsewhere, he said. However, others at the meeting considered it unnecessary to exclude amyloid.

In his opening talk, Fillit made a case for targeting biological processes of aging. Unlike most organs in the body, whose cells are continually renewed, most neurons and their networks in the brain must endure for a lifetime, Fillit said. This makes brain circuitry particularly vulnerable to malfunctions in metabolism and repair that may underlie or exacerbate AD, he said. Drugs that counteract mitochondrial dysfunction, epigenetic damage, faltering protein homeostasis, and inflammation therefore could slow disease progression, claimed Fillit. He added that treating co-morbidities associated with AD risk, such as vascular disease and elevated cholesterol, could also help.

Targets Galore. The process of aging takes its toll on many pathways. Howard Fillit proposed targeting multiple ones in combination therapy for AD. [Image courtesy of Howard Fillit, ADDF].

Diane Stephenson of the Coalition Against Major Diseases (CAMD) at the Critical Path Institute in Tucson said laying out these myriad aging pathways was a step forward from past discussions on combination therapy. “His rationale makes a lot of sense,” Stephenson told the group during her talk, which outlined the successes and future logistical challenges combination trialists confront. 

Uncharted Territory
Who will run these trials? Cummings noted that expired patents mean pharmaceutical companies may not recuperate the cost of running a combination trial for repurposed drugs. In the case of new investigational drugs, most companies are reluctant to run combination trials before monotherapy trials have read out. Multiple companies could join forces to contribute different parts to the combo, but that has its own commercial and intellectual-property hurdles, Cummings said. He concluded that while involving pharma in early stage combination therapy trials is desirable, other research institutes and foundations may be more willing to get the ball rolling.

If a combination of repurposed drugs passes muster in Phase 2, several scenarios are possible. If each drug in the combination is currently available for other indications, clinicians who know the positive results could prescribe the combo off-label to AD patients. While this would make the new cocktail immediately available to patients, Cummings told Alzforum that the downsides would be significant. They include liability for the prescribing physician, possibly lack of reimbursement for patients, low visibility within the clinical AD community, and uncertainty about the long-term safety of the combination. Alternatively, a pharmaceutical company could run a Phase 3 trial and seek regulatory approval for the combination.

Richard Mohs of Eli Lilly and Company told the panel that companies could be enticed to take on such an endeavor if a new formulation, dose, or combination pill would give them new intellectual property, and they were convinced of the biological rationale behind the combination. Lilly has two combo contenders of its own in development—the anti-Aβ antibody solanezumab is in Phase 3 and the BACE inhibitor ADZ3293 being co-developed with AstraZeneca is in Phase 2/3. Lilly scientists recently presented preclinical data suggesting that a combination of an anti-Aβ antibody and a BACE inhibitor wiped out amyloid more efficiently in mice than either alone (see Jul 2014 conference news). However, Mohs told Alzforum that given the concerns about the safety of BACE inhibitors, he would like to see each drug developed individually first.

If trialists go for FDA approval of a combination, what would the regulatory road look like? According to Rusty Katz, the former director of the Food and Drug Administration’s division of neurology products, the only way to know is to ask the FDA. There is precedent for combination therapy in other areas, such as oncology, Katz said, but for neurology, combo trials are terra incognita. “This is just as new to the FDA as it is to researchers,” he told Alzforum.

Katz delivered a rousing talk at last year’s Clinical Trials in Alzheimer’s Disease (CTAD) conference in Philadelphia, urging the research community to go for “big effects” and try combination trials (see Dec 2014 conference news). The talk motivated Cummings and Fillit to organize the current panel. The agency drafted guidelines for initiating combination trials in 2013, but Katz said several aspects of the guidelines are open to interpretation. For example, they state that in vitro or animal studies could be used to demonstrate that a combination of two drugs is superior to either on its own. After that, a clinical trial could theoretically test the combination against placebo. However, Katz said he questions whether the FDA would want to rely too heavily on non-clinical data to make the assessment that the combination is warranted. He also raised the possibility of eliminating the placebo group altogether in early trials—testing only drug A, drug B, and drug A+B—to indicate that the combination is superior. Some researchers, including Stephen Salloway of Brown University in Providence, Rhode Island, were skeptical of this approach, saying that placebo groups should always be included. While the FDA’s guidelines were designed specifically for new investigational drugs, not repurposed ones, combination therapies of repurposed drugs would still need to demonstrate safety, proper dosage, and efficacy. Whether some steps could be skipped in the trial pipeline will depend on the drugs and the populations in which they were tested previously, Katz said. The best way to obtain approval for combinations is to involve the FDA as early as possible, he emphasized.

Finding Targets in the Biological Haystack
How can researchers identify targets and drugs for AD combination trials? Barry Greenberg of the University Health Network in Toronto suggested quantitative systems pharmacology (QSP) could help. This uses computational approaches to generate hypotheses about how biological pathways might respond to drugs and potential on- and off-target effects that could emerge in clinical trials. The problem with current drug-discovery methods, Greenberg said, is the focus on a single target without consideration for the entire networks of genes and proteins that link to it. “If you modulate a target, you are not just modulating that target. You are modulating a homeostatic system, and that is one of the major reasons drugs fail in clinical trials,” he said. Not only can a drug engage more than one target, but even when a drug affects only a single protein, for example, the consequences of modulating the networks connected to that protein may be widespread or unexpected. Furthermore, these networks differ among tissues, genetic backgrounds, and disease states. The complexity balloons in combination therapy, where more than one target will be hit at once. Greenberg cited examples of therapy failures that could have been avoided with proper QSP. They include Vioxx, which targeted cyclooxygenase-2, an enzyme involved in multiple biological pathways in different tissues, and γ-secretase inhibitors, which targeted a protein complex with essential pleiotropic functions.

Greenberg championed deep analysis of failed trials to facilitate improvements in trial design. He proposed that combination trials include non-pharmacological therapies. Citing the recent FINGER trial in Finland, in which behavioral interventions appeared to work as well as drugs, Greenberg said dietary and exercise intervention should be given the same weight as drugs in combination therapies (see Jul 2014 conference news). 

Gerard Schellenberg of the University of Pennsylvania in Philadelphia proposed that large genetic-association studies and whole-exome sequencing approaches could also yield potential targets for combination therapy.

The Short List.

Some genes that were expressed more (green) or less (yellow) in AD brains overlapped with the druggable genome (purple). [Image courtesy of Anil Jegga.]

Anil Jegga of the Cincinnati Children’s Hospital and University of Cincinnati in Ohio suggested a multi-pronged strategy for identifying compounds for combination therapies. Pooling data from studies that measured hippocampal gene expression in AD patients versus controls, Jegga generated a list of differentially expressed genes and determined which of them appeared within the “druggable genome”—a collection of genes already targeted by drugs or with druggable potential, such as kinases. He filtered the list using knowledge of pathways thought to be involved in AD, such as immune responses, and came up with a pool of candidates to target. He proposed using this pool to identify drugs that could be tested as AD therapies. Jegga had previously used a similar approach to design a drug combination for cystic fibrosis, which has worked in preclinical studies so far. 

Designed for Success
Once trialists identify a drug combination, how can they test it? Klaus Romero and colleagues at CAMD have designed a clinical trial simulator that received qualification from the FDA and European Medicines Agency in 2013 (see Alzforum Webinar and Jul 2013 news). The simulator is not designed for combination trials; creating such a model, which would need to account for the interaction between drugs, would require six to eight months, Romero estimated.

Diane Wuest of GNS Healthcare in Boston, a startup comprising computer scientists, systems biologists, and statisticians, stressed a big-data approach. Her company takes multiple sources and types of information, ranging from patient data to genetic associations to biomarker analyses, and combines them to predict which drugs to test, which biomarkers to use, or to make models that guide researchers in designing clinical trials or tweaking adaptive trials on the fly. 

Michael Donohue of the University of California, San Diego, stressed the importance of creating long-term disease-progression models that span the entire course of the disease. Using ADNI data, which at the time included longitudinal data spanning only nine years, Donohue generated computational models that extrapolated the trajectory of markers and outcomes, such as amyloid PET, CSF biomarkers, hippocampal volume, and cognitive performance across several decades. Donohue hopes these models will facilitate selection of biomarkers and outcomes suitable for clinical trials of people at different stages of disease. He proposed using a multivariate approach—incorporating several biomarkers and cognitive tests for a given disease stage and patient type—for inclusion and exclusion criteria, rather than selecting patients based on single measures. Donohue also proposed creating a patient registry that contains biomarker and cognitive data and could be accessed by clinical trial sponsors.

Combination Front Runners
Two talks at this meeting highlighted early success with repurposing. Joao Siffert of Avanir Pharmaceuticals described how a combination of two drugs originally approved to treat uncontrollable laughing and crying could calm agitation in AD patients. AVP-923, a combination of the cough suppressant dextromethorphan and quinidine (a drug that prevents liver breakdown of dex) is approved, so the researchers conducted Phase 2 studies at 44 centers to measure the combination’s effects on agitation in Alzheimer’s (see Oct 2014 conference news). Phase 3 trials are beginning, Siffert said.

Daniel Cohen of Pharnext in Issy-les-Moulineaux, France, described an ongoing combination trial for AD in Bordeaux. PXT00864, a combination of drugs approved for spasticity, improved scores on the ADAS-Cog11 in people with mild AD, while the scores of those on placebo declined (see Dec 2014 conference news). 

Where to?
The goal of the meeting was to generate ideas for combination therapies that could commence immediately. However, when late afternoon rolled around, most of the discussion had not advanced beyond big-picture planning. The panel decided that the selection of combinations would require another meeting, perhaps one that convened experts in each area of biology ripe for targeting, such as neuroinflammation, mitochondrial function, and epigenetics. Furthermore, while some researchers wanted to charge forward to combination trials and try something, others expressed caution and called for more research into signaling pathways and drug interactions. Romero suggested that trials could move forward so long as real-time results are continuously fed into modeling efforts to improve future trials.

As the day’s proceedings inched to their end, Maria Carrillo of the Alzheimer’s Association offered an idea: Since the group was not ready to nominate combinations for testing that day, ADDF and the Alzheimer’s Association would put out a joint request for proposals for combination therapies on repurposed drugs to the research community. While the specifics of this funding did not materialize at the meeting, the panel concluded with a sense of resolution that combination trials were at least on the horizon.—Jessica Shugart


Make a Comment

Comments on this content

  1. The proposal of combination trials, necessarily with amyloid-modulating interventions, is a substantial challenge, as discussed in an earlier post. Basically the different synthesis rates of Aβ1-40 and Aβ1-42 (Potter et al., 2013) and different aggregation rates of these two amyloid peptides (Garai et al., 2013) results in a complex relationship between monomers, oligomers, and aggregated forms of both peptides. Add to that the differential effect of the different Aβ forms on electrophysiological or functional readouts of neuronal circuits (for instance through glutamate or α7 nAchR) where many non-linear processes play a role, and you have a very complex model that goes beyond simple linear (or additive) hypotheses.

    A possible novel approach that has been mentioned by Barry Greenberg and is already used in other disease indications, such as oncology, cardiovascular, and inflammation, is mechanism-based quantitative systems pharmacology (Geerts et al., 2013). This computer-modeling approach is based on integration of a substantial amount of biological, biochemical, and clinical domains and constrained by existing clinical data in AD patients. Computer-modeling is part of the standard toolbox of other successful industries, such as automotive, micro-electronics, and aeronautics.

    We have a paper in review in which we simulated the complex pharmacodynamic interactions between acetylcholine esterase inhibitors, such as donepezil and galantamine, with memantine, antipsychotics, and smoking on cognitive impairment in schizophrenia. It turns out that these interactions can go from synergism over additive to negative interference, depending upon the nature of the antipsychotic. This might explain the lack of clear clinical benefit, especially when the effect sizes are relatively small.

    Computer modeling of these complex pharmacodynamic interactions might provide interesting insights that could help design clinical trials for combination therapies in this patient population. 


    . Increased in vivo amyloid-β42 production, exchange, and loss in presenilin mutation carriers. Sci Transl Med. 2013 Jun 12;5(189):189ra77. PubMed.

    . Quantitative analysis of the time course of Aβ oligomerization and subsequent growth steps using tetramethylrhodamine-labeled Aβ. Proc Natl Acad Sci U S A. 2013 Feb 26;110(9):3321-6. PubMed.

    . Quantitative systems pharmacology as an extension of PK/PD modeling in CNS research and development. J Pharmacokinet Pharmacodyn. 2013 Jun;40(3):257-65. Epub 2013 Jan 22 PubMed.

Make a Comment

To make a comment you must login or register.


News Citations

  1. Combination Trials: Companies Can Play Well in One Sandbox
  2. Anti-Amyloid Therapies Combine Forces to Knock Out Plaques
  3. Rusty Unleashed: Forget Disease Modification, Go for Big Effect
  4. Healthy Lives, Healthy Minds: Is it Really True?
  5. AD Trial Simulation Tool Receives Regulators’ Blessings
  6. A New Drug to Calm Agitation, Uncontrollable Laughing and Crying, in Alzheimer’s?
  7. New Ideas for Alzheimer’s Treatment: What’s on Offer in 2015?

Therapeutics Citations

  1. Solanezumab
  2. AZD3293
  3. AVP-923

Webinar Citations

  1. Computational Modeling—Will it Rescue AD Clinical Trials?

External Citations

  1. guidelines 

Further Reading


  1. . A review: treatment of Alzheimer's disease discovered in repurposed agents. Dement Geriatr Cogn Disord. 2013;35(1-2):1-22. PubMed.