When the Dominantly Inherited Alzheimer’s Network Trials Unit (DIAN-TU) convened its nine-company pharma consortium in Washington, D.C., earlier this month, it was clear the meeting was more than a regular checkup on DIAN-TU’s many ongoing projects. Even while the observational cohort study, and the first secondary prevention trial that grew out of it, are humming along and generating data, this latest gathering was squarely about preparing for the next phase. Where to go from here was the question of the day. If the DIAN-TU scientists led by Randy Bateman of Washington University, St. Louis, can persuade their consortium, the second trial aimed at preventing clinical symptoms in carriers of autosomal-dominant Alzheimer’s disease (ADAD) mutations will become significantly more aggressive. Now that it has been established that therapy trials in this far-flung, middle-aged population can indeed be done, the goal for the next trial is to shoot for a much larger biomarker and therapeutic effect. This DIAN-TU NexGen trial is to start in 2016.

Nine biopharma companies currently participate in the DIAN-Trials Unit Pharma Consortium. They are funding planning meetings and lending their scientists’ expertise in hopes of seeing their own investigational drugs tested in largely asymptomatic mutation carriers as a bridge to an indication in all AD. Elan, Mithridion, Novartis, and Pfizer are former founding members.

Part of the motivation to reach for a larger effect is personal. “The DIAN families participate enthusiastically in the cohort study and the first trial, but they are asking: ‘Why are you being so conservative? Why can’t we do more?’” Bateman told the pharma researchers. Part of the motivation is scientific. Recent data have shown that carriers of autosomal-dominant Alzheimer’s mutations overproduce Aβ and may need stronger intervention to hit the therapeutic target. The certainty that they will develop this disease makes the families, as well as the regulators, more willing to accept the risks inherent in taking investigational drugs. Finally, part of the motivation is a cautiously growing confidence around safety. The first trial in this population was shaped by concerns around giving such medications to asymptomatic people; however, experience since then has shown that the DIAN population tolerates at least the current regimen well. 

What form would more aggressive treatment take? The DIAN-TU scientists want to achieve three things for the NexGen Trial: They want to push the dose of the investigational drugs to maximal effectiveness, they want to combine investigational drugs, and they want to make the trials flexible by adapting more frequently. Each of these goals departs from the conventional way of doing things in Alzheimer’s therapy development. To achieve them, the DIAN-TU leaders have started a conversation with their academic and industry colleagues. Below is a summary of the issues at play.

But first, a quick update on where current DIAN projects stand. After all, they are the foundation for the NexGen trials. The observational cohort study is fully enrolled, at 416 participants from families with a known pathogenic APP or presenilin mutation. It has entered its longitudinal phase with a second, five-year grant cycle, for an 11-year funded study. The study is adding brain imaging with the tau PET tracer T807/AV1451. FDG-PET, a measure of neuronal glucose metabolism, is being phased out to reduce both cost and burden, because it is less specific to AD than imaging the disease’s core molecular pathologies. The observational cohort study is beginning to analyze longitudinal data on its original set of clinical and biomarkers. Its ongoing ancillary studies include exome chip genotyping and generating fibroblast-derived iPSC lines for differentiation into cortical neurons.

The DIAN Expanded Registry has become an active outreach tool. It has brought in nearly 220 people with a known ADAD mutation in their family. Of those, 175 have been referred to trial sites. Of the 813 registry members, 130 are physicians, and all tend to be highly interested in research participation, Bateman said. When someone with a family history first registers, they can obtain exploratory genetic testing to establish whether that history is due to an ADAD mutation; this has led to the discovery of 24 new families, Bateman told the consortium scientists. More generally, the advent of the trial has stimulated a groundswell of interest in genetic counseling and predictive testing among asymptomatic relatives, Bateman noted. “Fifteen percent chose this option when the observational cohort started in 2008, but with the trial there was a surge of requests, and by now we are up to 35 percent,” he said.

Current and future sites of the DIAN observational study and DIAN-TU trials. [Image courtesy of DIAN-TU.]

DIAN is adding sites both for the observational study and clinical trials. New centers are coming online in Argentina, Korea, and Japan, which is planning a national four-site DIAN project. Beyond these, some 30 clinical sites worldwide collectively know about 3,000 more potential participants (see table). The DIAN team is supportive of a separate effort to fund a European extension of the global DIAN through the Innovative Medicines Initiative 2 program. Enrolling a total of 1,000 DIAN and DIAN-TU participants worldwide within the next two years is realistic, Bateman said.

The connection between DIAN’s observational and trial components is key, Bateman said at the meeting. Nine of the current treatment trial’s 26 active sites were original DIAN observational sites. Being an observational site is the most efficient way to become a therapeutic site, Bateman said. Much of the infrastructure is in place, as are motivated and informed participants who pass the trial’s screening procedures quickly and stay with the trial. Recent therapeutic trials in early stage Alzheimer’s disease have had screen failure rates of 80 percent; for the DIAN-TU trial, that number is below 10 percent, Bateman said. To tighten the link between the observation study and therapy trials, the DIAN observational study will add the CogState test battery, and it will conduct its clinical and cognitive assessments according to International Committee on Harmonization (ICH) guidelines for good clinical practice. When administered in this way, the data can carry over from the observation study into any treatment trial a DIAN participant may decide to join. This helps the observational study transition interested participants into future therapy trials, and provides run-in data to better evaluate if the tested drug made a dent in the participant’s cognitive trajectory.

The current DIAN-TU treatment trial evaluates the ability of solanezumab or gantenerumab to change a biomarker readout. This trial is two-thirds enrolled, and a first look at the data will come in 2016.

In the interim, the challenge at hand is to plan ahead for larger registration trials based on a cognitive endpoint. Ideally, such trials should be designed to evaluate more drugs more quickly. This is where the NexGen goals of ramping up the dose, combining drugs, and adapting more frequently come in. First, about dose. Because mutation carriers overproduce Aβ42 from an early age, higher doses of Aβ-reducing agents may be necessary to make a difference on both pathology and symptoms. Tantalizingly, Phase 1, 2, and 3 results with aducanumab, crenezumab, and solanezumab, respectively, have not only shown robust target engagement, but also hinted that a cognitive signal could be had with higher doses. Upping the dose always raises safety questions but here, too, the ground has shifted. The first trial was strongly guided by the Hippocratic obligation of doing no harm to outwardly healthy middle-aged people; however, since 2012, exposure in 100 participants randomized to treatment thus far has not generated safety concerns. Hence the stance has evolved to, “We should do more in people we know are destined to die young of this disease.”

At the meeting, academic and industry researchers discussed how to go about dosing. One option is to design a trial with individualized titration. There, each patient takes successively higher doses until they show the first side effect, at which point the trial backs off and continues at the highest dose the participant tolerates well. Another is to start a whole treatment group off on a high dose and to dial it down if needed. Typically, drug-development programs try to find what is called the maximum tolerated dose (MTD) for a drug early in the clinic, in Phase 1. In D.C., the DIAN-TU group considered using healthy volunteers to find the MTD for candidate drugs, so as not to use the limited DIAN-TU population for both safety and efficacy evaluation of the same medication.

A conversation about dosing cannot get very far without discussing specific drugs. The pharma consortium discusses drug classes, but not a company’s individual proprietary drugs. For example, with a BACE inhibitor, the consensus was that no one yet knows which dose might score a cognitive effect. On the other hand, several such compounds have extensive pharmacokinetic and pharmacodynamics biomarker data that allow a trial design team to choose a maximum dose based on a desired percentage of CSF Aβ reduction.

From the perspective of the patient, “maximum tolerated dose” is an important concept in the private risk-benefit calculation. One pharma scientist put it this way: “If I was 45 and knew I’d start losing my memory at 50 but taking this drug for the next five years could stave it off, I’d be interested. But if I also knew I was going to wake up nauseous every morning during those five years and my memory loss would start at 51 instead, I’d be less sure.” Both a person’s tolerance for discomfort and the regulators’ tolerance for risk may rise the closer a mutation carrier gets to his or her family’s average age at onset, the scientists agreed.

Combination therapy is the second goal on which DIAN-TU researchers have set their sights for their NexGen trial. The idea here is to test two or more investigational drugs together in the same patient as a way of reaching for a bigger, and perhaps faster, effect than each therapy achieves on its own. Combination therapy has become the norm in other illnesses, such as cardiovascular disease, diabetes, and AIDS, but most drugs were first developed and approved on their own. In Alzheimer’s, researchers broadly agree that the current approach of mono-therapy parallel group trials puts the field on too slow a track to clinch large effects within the 2025 timeline of the National Plan to Address Alzheimer’s Disease.

Repeated calls to conduct combination trials in Alzheimer’s disease before mono-therapies are approved are beginning to gain traction throughout the field. Groups ranging from ACT-AD to the New York Academy of Sciences have hosted discussion of the topic (see Feb 2013 conference seriesMar 2014 conference news). Stakeholder groups have weighed in (Stephenson et al., 2014Perry et al. 2015). The FDA has issued guidance, and Rusty Katz, formerly of the FDA, has appealed to the field to tackle combination therapy as a way of obtaining truly large treatment effects (see Dec 2014 conference news). On April 12, 2015, Maria Carrillo of the Alzheimer’s Association hosted a meeting to prepare a consensus paper among academic and industry leaders on how to get combination trials on the road.

In the DIAN-TU population, combination therapy could initially involve an anti-Aβ antibody plus a BACE inhibitor in asymptomatic carriers who are years away from expected onset, or an anti-amyloid plus an anti-tau therapy in carriers who are close to onset or mildly symptomatic. That said, the DIAN-TU Pharma consortium scientists emphasized the importance of articulating both a scientific rationale and a larger framework for combination therapy that would support future types of combination as new mono-therapies enter clinical trials (see Alzforum Therapeutics database). Some preclinical data showing additive effects of combining a BACE inhibitor and an anti-Aβ antibody have started to appear (see Apr 2013 conference news on Jacobsen et al., 2014Jul 2014 conference news on DeMattos et al., 2014). Even so, the DIAN-TU researchers discussed the need for more preclinical studies. They agreed that dosing for combinations should be settled in Phase 1 and proof of concept in Phase 2 before attempting a Phase 3 trial. One relatively straightforward option for the 2016 NexGen trial might be to test two drugs as both mono-therapy and combination in a form of 2x2 factorial design.

At the DIAN-TU planning meeting, scientists signaled openness to testing drugs in combination, but an inherent disincentive also hung in the air. The Food and Drug Administration has never required a minimum effect size, and indeed all approved AD drugs have small effects. This, then, is enough for a company to bring a product to market. From a business perspective, shooting for a large effect might be seen as adding more risk than benefit. While the DIAN-TU scientists see the scientific rationale and perhaps a moral incentive, they have to persuade their upper management to go along.

Will DIAN-TU test combinations of drugs from different companies? In theory, the pharma consortium as a group is positioned to do that. The companies already work together to support the same initiative. Roche and Lilly each have given an active investigational drug into the same trial against a shared placebo group. But joining what companies call their “asset” with that of a competing company in the same patient at the same time—or even in the same mouse at the same time? That prospect did seem a bit like asking them to share toothbrushes. Even if pharma scientists were willing, it was not their decision to make that day. “It would take 1½  years to get the contract for that in place,” said one researcher. “Outside of this room, we are competitors,” said another.

Thus far, sharing trial-design and other types of expertise has fit comfortably under the precompetitive mantle of this consortium. Sharing drugs is harder, but the sense in the room was that it is not out of the question. Of the current pharma consortium members, only Lilly/AstraZeneca and Eisai/Biogen may have enough “assets” far enough in the clinic to launch a combination trial by 2016; other combinations would require cooperation between two competing companies. Even so, the group agreed to get to work on this issue, perhaps by engaging the Alzheimer’s Association to help facilitate preclinical studies of drugs from different companies conducted by an outside contract research organization.

The third proposed pillar for the DIAN-TU NexGen trial is to make it “frequently adaptive.” This means that the trial would look sooner at predefined interim outcomes—for example, a target engagement biomarker that is known to respond to the chosen drug within weeks or a few months. This enables a faster decision on whether to continue to a cognitive outcome or eliminate the drug. Even for the current, ongoing trial, the original two-year biomarker readout has been moved up. Instead, there will now be three interim assessments: one when everyone has been enrolled for one year, one when half have been enrolled for two years, and a final one when everyone has been in the trial for two years. The NexGen trials could follow a similar scheme of interim looks, but start them at six months.

Frequently adaptive also means shrinking the time between ending evaluation of one drug and starting another, or adding an additional drug arm. This becomes possible because swapping a drug or adding an arm will not require a wholly separate trial, but merely represent adaptations within an existing trial infrastructure and standing placebo pool. The European Prevention of Alzheimer’s Project (EPAD) is pursuing a similar approach in its ongoing planning phase.

In this first NexGen discussion in D.C., the industry scientists were open to the proposal of a frequently adaptive NexGen trial. That said, they made clear that the amount of upfront work to plan such an undertaking is formidable. What’s the threshold for biomarker decisions? Will the trial adapt based on an individual person’s or a group’s responses? Can it use a person’s longitudinal cognitive assessments to estimate their cognitive benefit, if any? How to model all this statistically? How to avoid Type 1 error? Questions abounded. Overall, however, the tenor was that adaptive trial platforms are well suited for a small patient population such as in DIAN.

Indeed, adaptive trial designs may be up and coming more broadly in late-onset AD therapy development. One such trial is already underway (see Nov 2012 conference newsAlzforum Therapeutics database). To some, this trial signifies an attempt to reinstate Phase 2 proof-of-concept studies at a time when some drug developers have chosen to largely skip Phase 2 and leap from Phase 1b directly to Phase 2/3 pivotal trials.—Gabrielle Strobel

Comments

  1. The proposal of combination trials, necessarily with amyloid-modulating interventions (as they are the only disease-modifying drugs close to the clinic), is a substantial challenge, both in terms of selecting drugs and more specifically doses.

    Based on the human SILK data, Aβ1-40 is synthesized faster than Aβ1-42 and cleared more slowly (Potter et al., 2012). Together with biochemical observations on the aggregation rates of these two amyloid peptides (Garai and Frieden, 2013), this results in a complex relationship between monomers, oligomers, and aggregated forms of both peptides. Assuming that oligomers might be formed from breakdown of plaques, it is clear that many non-linear processes play a role. When combining BACE-inhibitors with antibodies against either monomers (as with the Lilly drug) or against aggregated forms (as with the Roche drug), the results are likely to be quite different.

    Using a mechanism-based Quantitative Systems Pharmacology and humanized computer model (Geerts et al., 2014) based on integration of a substantial amount of biological, biochemical and clinical domain expertise, we recently showed that the choice of the epitope for the antibody matters, both in terms of biomarkers (CSF levels of Aβ) and functional outcome (effect on ADAS-Cog). The model can take into account different baseline amyloid levels and disease states and is constrained by existing clinical data on amyloid modulation in AD patients.

    For instance, a high level of BACE inhibition likely reduces levels of the more neuroprotective Aβ1-40 monomer more than antibody-mediated clearance of the monomers at clinically relevant target exposure, because of the relatively high synthesis rate of Aβ1-40. Interestingly, the computer model provides a possible hypothesis for the observed changes in free unbound CSF Aβ1-40 and Aβ1-42 in the solanezumab trial.

    On the other hand, clearing aggregated forms with an antibody against aggregated β-amyloid might get rid of the oligomers formed from breakdown of the plaques and leave Aβ1-40 monomer levels largely determined by the effect of BACE-inhibitors. This suggests that baseline amyloid load is another factor that drives variability.

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

    References:

    . 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.

Make a Comment

To make a comment you must login or register.

References

News Citations

  1. Combination Drug Trials: Time to Open a New Front in AD?
  2. Combination Trial Debate Energizes Keystone Symposium
  3. First Stab at Combination Therapy Yields Additive Effect
  4. Anti-Amyloid Therapies Combine Forces to Knock Out Plaques
  5. CTAD: Adaptive Antibody Trial to Try Bayesian Statistics

Therapeutics Citations

  1. BAN2401

Paper Citations

  1. . Charting a path toward combination therapy for Alzheimer's disease. Expert Rev Neurother. 2015 Jan;15(1):107-13. Epub 2014 Dec 26 PubMed.
  2. . Building a roadmap for developing combination therapies for Alzheimer's disease. Expert Rev Neurother. 2015 Mar;15(3):327-33. PubMed.
  3. . Combined treatment with a BACE inhibitor and anti-Aβ antibody gantenerumab enhances amyloid reduction in APPLondon mice. J Neurosci. 2014 Aug 27;34(35):11621-30. PubMed.

Other Citations

  1. Alzforum Therapeutics database

External Citations

  1. Innovative Medicines Initiative 2
  2. International Committee on Harmonization (ICH)
  3. DIAN-TU treatment trial
  4. National Plan to Address Alzheimer’s Disease
  5. DeMattos et al., 2014

Further Reading

Papers

  1. . Dominantly Inherited Alzheimer Network: facilitating research and clinical trials. Alzheimers Res Ther. 2013 Oct 17;5(5):48. PubMed.
  2. . Preclinical trials in autosomal dominant AD: Implementation of the DIAN-TU trial. Rev Neurol (Paris). 2013 Oct;169(10):737-43. PubMed.
  3. . Developing an international network for Alzheimer research: The Dominantly Inherited Alzheimer Network. Clin Investig (Lond). 2012 Oct 1;2(10):975-984. PubMed.