Despite the litany of failed Alzheimer’s drugs, academic and industry researchers at the 13th International Conference on Alzheimer’s and Parkinson’s Diseases, held March 29 to April 2 in Vienna, appeared confident that they are moving in the right direction. Asked why the field continues to go after the amyloid hypothesis even after billion-dollar losses, Samantha Budd Haeberlein of Biogen in Cambridge, Massachusetts, expressed the consensus view when she said, “The science for Aβ is very compelling.” Researchers stressed that they have learned from the past and believe the new trials stand a greater chance of success than previous ones, in part thanks to the recent emphasis on biomarkers.
At the same time, scientists conceded that therapeutic strategies must broaden beyond just targeting amyloid itself. “Instead of a linear amyloid cascade, it’s more of a swirling eddy,” quipped Brad Hyman of Massachusetts General Hospital, Boston. His point was that many other processes, such as inflammation and tau pathology, interact with amyloid to influence neurodegeneration. For an update on non-amyloid treatments, see May 2017 conference news. Roche’s Paulo Fontoura said the previous view—that amyloid is necessary and sufficient for the disease—has evolved into an acknowledgement that curbing amyloid may offer but a partial solution, and that combination therapy may be necessary for a truly effective treatment. In addition, the field still needs to identify the right form of Aβ to target. That goal is close, he believes. “We are on the verge of developing the first successful treatments,” Fontoura predicted. So what, in this big scheme of things, was the amyloid-related news at AD/PD?
Researchers in Vienna presented no breakthroughs. Instead, they offered updates on three drugs since Alzforum’s last coverage from the CTAD conference (see Dec 2016 conference news; Dec 2016 conference news). AstraZeneca reported Phase 1 data from the new anti-Aβ antibody MEDI1814; Eisai discussed Phase 2 pharmacodynamic data for a BACE inhibitor going on to Phase 3; and the biotech firm Alzheon in Lexington, Massachusetts, promoted its rationale for resurrecting the failed drug tramiprosate. Previously, Alzheon had noted a treatment benefit for the subset of participants who carried two copies of the APOE4 allele, and in Vienna, Martin Tolar said data from an open-label extension of a Phase 3 trial conducted in 2006 suggested a cognitive advantage over the original placebo group a year later. A version of the drug will enter Phase 3 this year.
Antibody Therapy: Sweeping Out Aβ
The strategy that has advanced the furthest through the development pipeline is that of clearing Aβ with antibodies. The 12 anti-Aβ antibodies currently in clinical trials target different forms of the peptide, employ distinct mechanisms of action, and produce varying side effects. Amyloid-related imaging abnormalities (ARIA) are of particular concern. These light spots on MRI scans reflect fluid buildup from leaky blood vessels (ARIA-E) or micro-hemorrhages (ARIA-H). Some researchers believe ARIA results from inflammation triggered by antibodies that bind amyloid in blood vessels.
In Vienna, Fabrizio Piazza of the University of Milano-Bicocca, Italy, noted that some people with cerebral amyloid angiopathy (CAA) spontaneously develop ARIA-like events characterized by edema in blood vessels. This phenomenon is known as CAA-related inflammation (CAA-ri), and Piazza has correlated its occurrence with the presence of endogenous autoantibodies against Aβ (see Piazza et al., 2013). These autoantibodies mobilize Aβ, removing it from plaques and leading to more deposition and damage in blood vessels. The condition models therapeutic-induced ARIA, Piazza said. “Aβ removal leads to a temporary overload at the vessels, particularly for people in advanced disease stages who have many plaques,” he explained.
One therapy option might be to pretreat participants at advanced stages of Alzheimer’s pathology with ponezumab, to lower vascular amyloid deposits before busting up parenchymal plaques, he suggested. Pfizer discontinued ponezumab for AD but is now developing it for CAA. Piazza leads the Inflammatory Cerebral Amyloid Angiopathy and Alzheimer’s Disease Biomarkers International Network. iCAβ is evaluating potential diagnostic and prognostic biomarkers of CAA and ARIA. Such biomarkers might help clinicians manage and treat ARIA, he believes.
In the meantime, some antibodies are being engineered to restrain microglial activation and thus lessen ARIA. One of these is MEDI1814, which recognizes monomeric Aβ42 and is being developed in collaboration with Eli Lilly. Built on an IgG1 backbone, MEDI1814 carries mutations that weaken its binding to microglial Fc receptors, potentially lowering the incidence of ARIA (see Nov 2015 news). In Vienna, Thor Ostenfeld of AstraZeneca Neuroscience Innovative Medicines & Early Development, Cambridge, U.K. reported results from the recently completed Phase 1 study. It enrolled 77 people between 55 and 85 years who were clinically diagnosed with mild to moderate AD and had low CSF Aβ42. In the first study phase, 45 participants received a single dose of drug or placebo in a 3:1 ratio. The tested doses were 25, 100, 300, 900, or 1,800 mg by IV, or 100 mg subcutaneously. In the second phase, the other 32 participants received three doses each over the course of two months. Dose levels for this group were 300, 900, or 1,800 mg by IV or 200 mg subcutaneously.
Participants were followed for four months. They had no serious adverse events, Ostenfeld reported. The most common complaints—dizziness, headache, gastrointestinal problems—seemed unrelated to dose. The researchers saw no irritation at the site of injection or reactions to the infusion. They found no evidence of ARIA-E or ARIA-H on MRI scans, supporting the idea that this antibody lowers the risk for this side effect.
The pharmacokinetic data resembled that for other antibodies. The antibody’s half-life in the blood was 17 to 20 days. There, it stabilized Aβ42, with plasma levels of the peptide rising after dosing. A small amount of MEDI1814 entered the brain, with CSF levels hovering around 0.1 to 0.6 percent of those in plasma, as is commonly seen with other therapeutic antibodies. About one-third of participants still had detectable CSF antibody a month after dosing stopped, Ostenfeld noted. MEDI1814 bound to Aβ42 in the central nervous system, as seen by the disappearance of free CSF peptide. A single IV dose of 300 mg cut the levels of unbound CSF Aβ42 in half, while multiple doses of 900 mg took it to near zero. For the subcutaneous formulation, about 33 percent of the dose became bioavailable, Ostenfeld said.
Crucially, the researchers saw a treatment effect on total CSF Aβ42. At the highest doses of drug, this biomarker tripled or quadrupled. Researchers have long debated how they would expect CSF Aβ42 to change if a drug worked. In Vienna, Ostenfeld said that because CSF Aβ42 levels fall in Alzheimer’s disease as the peptide becomes bound up in plaques, this rise may indicate mobilization of the peptide. A similar effect was seen in preclinical studies in rats and monkeys, Ostenfeld said. Aβ40 levels remained unchanged, demonstrating the antibody’s selectivity.
Other scientists stress the importance of determining a full dose-response curve for new therapeutics in order to pick the most effective dose. This was not always done for previous anti-amyloid therapeutics, including solanezumab, which fell short due to a puny effect size in Phase 3 trials of mild AD. Privately, researchers agree that more extensive dose ranging might have prompted a higher solanezumab dose in those trials (see Nov 2016 news; Dec 2016 news). Because AstraZeneca researchers found nearly complete engagement of CSF Aβ42 at high doses of MEDI1814, they believe they have the necessary dosage data, Ostenfeld said in Vienna. He told Alzforum that his team will take the lessons from solanezumab into account.
BACE Inhibitors: Squelching Production
BACE inhibitors represent a different anti-amyloid approach, seeking to turn off the spigot rather than pump up outflow. Several are in trials, with researchers encouraged by early signs that they appear safe enough for long-term use (see Oct 2016 conference news). In Vienna, Bruce Albala of Eisai Inc, Woodcliff Lake, New Jersey, presented safety and pharmacodynamic data from an ongoing Phase 2a trial of elenbecestat (formerly E2609). The study started in 2014 and enrolled 71 people who had prodromal or mild AD verified by a positive amyloid scan. Over the course of 18 months, they received either 5, 15, or 50 mg/day elenbecestat or placebo. A 12-week interim analysis determined that all doses were well-tolerated at that time, Albala said.
To model the pharmacodynamics, the researchers combined data from this trial with that from previous, smaller studies on healthy volunteers. They found that the drug dampened Aβ production to a similar degree in controls and patients (see Jul 2012 news). Specifically, with 17 ng/ml of elenbecestat in cerebrospinal fluid, CSF Aβ1-x levels fell by half. This concentration can be achieved by administering 20 mg/day of the drug, Albala said. At the 50 mg/day dose, modeling predicts that CSF Aβ1-x will fall by 70 percent, he added.
Eisai had initially planned to run a Phase 2b safety and efficacy study on another 630 participants. That trial would have used a Bayesian adaptive design, in which doses and other parameters could be adjusted as it went along based on interim data. The trial included other innovative features, such as using the new cognitive composite ADCOMS as an outcome measure. However, based on the Phase 2a data, Eisai decided instead to accelerate elenbecestat directly into two large pivotal Phase 3 trials with a more conventional design. These two-year studies, called MISSION AD 1 and 2, will enroll 1,330 people with early AD. The trials will test the 50 mg dose, with the primary endpoint being the Clinical Dementia Rating Sum of Boxes. Andrew Satlin at Eisai told Alzforum that this design was chosen after consultation with health authorities.
An audience member questioned whether BACE inhibitor trials are worth doing, given the recent failure of Merck’s verubecestat to budge endpoints in the Phase 3 EPOCH trial in mild to moderate AD. Albala said that Eisai is testing its inhibitor at an earlier stage of the disease, similar to the ongoing APECS trial of verubecestat in prodromal AD (see Feb 2017 news). The Eisai trials run longer, confirm amyloid positivity with PET or CSF, and use a different outcome measure than the EPOCH trial, Albala added.
Let’s Try This One Again: Preventing Aggregation
Tramiprosate has resurfaced to once again take a stab at amyloid. An analog of the amino acid taurine, tramiprosate was originally developed by Neurochem in Montreal. It failed to improve cognition in AD patients in an 18-month Phase 3 trial and was rebranded as a nutraceutical and directly sold to consumers (see Aug 2007 news; Nov 2007 conference news).
In 2013, Alzheon licensed the drug following a re-analysis of the original data that suggested a slowing of decline in the APOE4 homozygote subgroup (see Abushakra et al., 2016). APOE4 homozygotes make up about 2 percent of the general population and 10 percent of people with AD worldwide (see Ward et al., 2012).
In Vienna, Tolar presented data from an open-label extension of tramiprosate. The Phase 3 trial had comprised 1,053 participants taking either 100 mg or 150 mg twice/day, or placebo. Seven hundred and thirty-six participants continued past the blinded phase on the higher dose for another year. In this extension cohort, 104 people carried two APOE4 alleles. Among this group, those who had been on drug for the full 2.5 years maintained a cognitive benefit of 3.9 points on the ADAS-Cog compared to those who were switched from placebo, Tolar claimed. People with one copy of APOE4 had about half that benefit. ADAS-Cog scores stabilized for people who switched from placebo to active drug but never caught up to the active group; this suggests a disease-modifying effect, Tolar said. Among homozygotes, those with mild AD benefited more than those with moderate AD, again implying that early intervention may be more effective.
How might tramiprosate help AD patients? Previous studies claimed the drug blocked aggregation of Aβ42, but provided no mechanism. Alzheon researchers examined the drug’s interaction with Aβ in vitro using mass spectrometry and nuclear magnetic resonance spectroscopy, and in the April 24 CNS Drugs, reported that tramiprosate binds the Lys16, Lys28, and Asp23 amino acid side chains of Aβ42. Multiple molecules of the drug become tethered to each Aβ42 monomer, enveloping the peptides and preventing them from forming oligomers, the authors said. The same mechanism inhibited the elongation of existing oligomers. In vitro, a 1,000-fold excess of tramiprosate over Aβ42 reportedly prevented oligomer formation (see Kocis et al., 2017).
Tolar believes the evidence supports a new study. The researchers developed a tablet prodrug they call ALZ-801. This prodrug is converted to tramiprosate in vivo, but is more easily absorbed by the body and lasts longer in blood, the researchers claim (see Nov 2015 news).
In Phase 1 studies of about 160 healthy elderly participants, ALZ-801 appeared to cause less nausea than tramiprosate, and this side effect cleared up after one week on the drug, Tolar said in Vienna. The other most common complaints were headaches, dizziness, and falls, but these did not relate to dose.
The new formulation cut the person-to-person variability in drug levels in half, Tolar added. Because 150 mg twice daily of tramiprosate achieved a therapeutic effect in the Phase 3 trial, the researchers looked for the ALZ-801 dose that would produce an equivalent CSF exposure. This turned out to be 265 mg ALZ-801 twice daily. The researchers calculated that this dosage should result in a brain concentration of about 550 nM ALZ-801, or about 1,300 to 3,700 times the concentration of soluble CSF Aβ42 seen in AD patients, Tolar noted. This suggests the drug reaches the necessary concentrations to interfere with oligomer formation and have a therapeutic effect, he predicted. Alzheon plans to start a Phase 3 trial of 265 mg twice-daily ALZ-801 in APOE4 homozygotes this year.—Madolyn Bowman Rogers
- Non-Amyloid Treatments: Inflammation, Epigenetics, Regeneration
- Much ‘Adu’ About a Little: Phase 1 Data Feeds the Buzz at CTAD
- Emerging Alzheimer’s Therapies Test the Waters at CTAD
- Truly New to Déjà Vu: Phase 1 Means ‘Check!’ for Some, ‘Out!’ for Others
- Lilliputian Effect Size Fells Phase 3 Trial of Solanezumab, Leaving Its Future Uncertain
- At 2nd Kloster Seeon Meeting, Renewed Optimism for Targeting BACE1
- Wave of New BACE Inhibitors Heading to Phase 2
- Merck Pulls Plug on Phase 2/3 BACE Inhibitor Trial
- FDA Deems U.S. Alzhemed Trial Results Inconclusive
- Philadelphia: European Trial of Alzhemed Ends, Marketing Morphs to Supplement
- Piazza F, Greenberg SM, Savoiardo M, Gardinetti M, Chiapparini L, Raicher I, Nitrini R, Sakaguchi H, Brioschi M, Billo G, Colombo A, Lanzani F, Piscosquito G, Carriero MR, Giaccone G, Tagliavini F, Ferrarese C, Difrancesco JC. Anti-amyloid β autoantibodies in cerebral amyloid angiopathy-related inflammation: Implications for Amyloid-Modifying Therapies. Ann Neurol. 2013 Feb 11; PubMed.
- Abushakra S, Porsteinsson A, Vellas B, Cummings J, Gauthier S, Hey JA, Power A, Hendrix S, Wang P, Shen L, Sampalis J, Tolar M. Clinical Benefits of Tramiprosate in Alzheimer's Disease Are Associated with Higher Number of APOE4 Alleles: The "APOE4 Gene-Dose Effect". J Prev Alzheimers Dis. 2016;3(4):219-228. PubMed.
- Ward A, Crean S, Mercaldi CJ, Collins JM, Boyd D, Cook MN, Arrighi HM. Prevalence of apolipoprotein E4 genotype and homozygotes (APOE e4/4) among patients diagnosed with Alzheimer's disease: a systematic review and meta-analysis. Neuroepidemiology. 2012;38(1):1-17. PubMed.
- Kocis P, Tolar M, Yu J, Sinko W, Ray S, Blennow K, Fillit H, Hey JA. Elucidating the Aβ42 Anti-Aggregation Mechanism of Action of Tramiprosate in Alzheimer's Disease: Integrating Molecular Analytical Methods, Pharmacokinetic and Clinical Data. CNS Drugs. 2017 Jun;31(6):495-509. PubMed.