Background

This is a tablet drug affecting amyloid precursor protein (APP) and its processing products. It inhibits APP cleavage by the enzyme BACE, the rate-limiting step in Aβ generation. The rationale of BACE inhibition is that it represents an upstream interference with the amyloid cascade. BACE inhibition is sometimes envisioned as long-term maintenance therapy to limit Aβ production after an initial round of immunotherapy to remove existing amyloid deposits.

In 2012, Janssen licensed this thiazine-based small-molecule compound from Shionogi, which had discovered and conducted preclinical research on it,  (Koriyama et al., 2021).

Findings

In 2013, Janssen ran a series of Phase 1 trials of JNJ-54861911. A first single-ascending-dose study in 56 healthy volunteers was followed by a second in 70 healthy elderly volunteers. Both assessed safety parameters as well as pharmacological measures relating to drug exposure and concentrations of Aβ fragments in CSF and plasma. Both were conducted in Belgium, and a similar study in 24 healthy volunteers was conducted in Japan. Additional studies in 46 healthy volunteers evaluated the effect on the concentration, metabolism, and excretion of JNJ-54861911 and of various other drugs commonly used by the elderly, including coffee, anxiolytic, or diabetes drugs. One trial in 64 people assessed whether JNJ-54861911 affects heart function. The data were published in peer-reviewed journals (Timmers et al., 2016; Timmers et al., 2017; Timmers et al., 2018).

In December 2013, the first trial in prodromal Alzheimer's disease began, enrolling 45 people in Belgium, the Netherlands, Spain, and Sweden. Study participants were cognitively impaired as measured by the CANTAB Elect test battery, and had evidence of amyloid deposition either as per a pathological Aβ/tau CSF assay result or per flutemetamol amyloid PET scan. Participants took either 10 or 50 mg of JNJ-54861911 or placebo once daily for four weeks, and were assessed on a range of biomarker outcomes related to drug exposure, metabolism, and target engagement, i.e., the concentration of various APP and Aβ fragments in CSF and plasma. 

In November 2014, Phase 2 began with a multinational trial in Europe. It compared a six-month, once-daily course of 10 or 50 mg of JNJ-54861911 to placebo in 114 people with a CDR rating of 0 to 0.5, plus evidence of amyloid pathology supplied either by CSF or PET—i.e., people who had asymptomatic to predementia Alzheimer's disease. This study assessed safety, exposure, target engagement, and a downstream effect in the form of CSF tau concentration, but not cognition/efficacy.

In March 2015, Janssen listed another Phase 1 trial in Japan to compare a one-month course of 10 or 50 mg to placebo in 18 people who were clinically normal as measured by a CDR of zero but had brain amyloid deposition as evidenced by low CSF Aβ42 levels. Called "asymptomatic at risk of AD," this population represents an earlier stage of AD pathophysiology than predementia or prodromal AD, as people with measurable impairment were excluded. This trial measured markers of drug exposure and target engagement. Data from this trial, and the prodromal AD trial above, were formally published (Timmers et al., 2018).

In March 2015 at the AD/PD conference in Nice, France, Janssen reported results of a single-ascending- and a multiple-ascending-dose study in healthy elderly participants. According to this presentation, the inhibitor was safe and well-tolerated in 94 people studied up to that point. It reportedly entered the blood and CSF with favorable pharmacokinetics and pharmacodynamics, and dose-dependently reduced Aβ1-37, Aβ1-38, Aβ1-40, and Aβ1-42. Reduction of the BACE cleavage product sAPPβ tracked reduction of Aβ, whereas levels of sAPPα rose (Apr 2015 news). A 5 mg dose was reported to reduce CSF Aβ concentration by half; 25 mg by 80 percent, 50 mg by 90 percent. 

In July 2015, a long-term safety and tolerability study began enrolling 100 patients from previous Phase 1 and 2 trials who were willing to continue their randomized, blinded treatment for another year. Participants were then invited to an additional year of open-label active treatment for all.

In October 2015, a Phase 2/3 study called EARLY began enrolling asymptomatic people at risk of developing Alzheimer's dementia. Their risk was determined by a CDR score of zero combined with CSF or PET evidence of brain amyloid accumulation. The trial enrolled people aged 60 to 85. To limit screen failures, 60- to 64-year-old candidates had to have either a family history of dementia, previously known ApoE4 genotype, or previously known biomarker evidence of amyloid deposition. The primary endpoint was slowing of cognitive decline, as measured by change on the Alzheimer's Disease Cooperative Study Preclinical Alzheimer Cognitive Composite (ADCS-PACC) between baseline and 54 months of treatment with either 10 or 25 mg of drug, or placebo. The 10 mg dose was later lowered to 5 mg. Secondary outcomes included 10 different functional, clinical, neuropsychological, exposure, and biomarker measures. This trial was to enroll 1,650 participants at 121 locations in Europe, Australia, Japan, North America, and Mexico, and run until 2023. For details on this trial, and a Phase 1 data summary, see Aug 2016 conference news.

In January 2016, Janssen added a Phase 1 study in 32 healthy adults in Germany to evaluate drug interactions between JNJ-54861911, the antidiabetic metformin, and rosuvastatin.

On May 17, 2018, Janssen announced the discontinuation of this program, citing findings of elevated liver enzymes among some of 600 people exposed to atabecestat thus far (May 2018 news). At the 2018 CTAD conference, Janssen reported that EARLY participants who took 25 mg atabecestat scored four points worse than the placebo group on the RBANS at three months, and one point worse on the PACC at six and 12 months. Treatment groups reported more depression, anxiety, and sleep problems than controls (Nov 2018 conference news). A preliminary analysis of cognitive and safety results was later published (Henley et al., 2019). Full trial data revealed that the atabecestat-induced cognitive decline reversed within six months when participants stopped taking drug, similar to what was reported for other BACE inhibitors (Sperling et al., 2021; Dec 2020 news).

In 2020, Janssen published results of the Phase 2 trial and its long-term safety and tolerability extension. Its 114 participants were randomized to 10 or 50 mg atabecestat daily, or placebo, for six months. Midway through the trial, the dose was halved to 5 or 25 mg, due to a high dropout rate in the 50 mg group and liver enzyme elevations in both groups. Ninety participants continued into a one-year double-blind extension at doses of 10 or 25 mg per day or placebo, and 77 continued into the open-label phase where all received 5 or 25 mg atabecestat. Consistent with previous results, atabecestat elevated liver enzymes. In the extension study, the treated groups showed a trend toward decreasing RBANS scores, consistent with worsening cognition reported in the Phase 3 trial and with other BACE inhibitors. Open-label dosing stopped with the rest of the atabecestat program in May 2018 (Novak et al., 2020). 

Investigation of the mechanism of liver toxicity points to an immune response to atabecestat or a metabolite is responsible. In one trial volunteer who discontinued the drug due to elevated liver enzymes, a liver biopsy revealed inflammation with infiltration of immune T and B cells, and hepatocyte death. Treatment with prednisolone returned enzyme levels to normal (De Jonghe et al., 2020). Drug-reactive T cells were subsequently detected in trial participants with liver damage; these cells arose from  atabecestat/metabolite binding to antigen-presenting cells (Thomson et al., 2020).

See all trials at clinicaltrials.gov.

Clinical Trial Timeline