Off-target effects proposed as reason for phase I study death


Off-target drug effects may be the reason for the phase I accident in France that left one person dead and five others hospitalized (see 1 dead in phase I trial, NeuroSens, January 27, 2016). However, other causes are also being explored by the 12-person Temporary Specialist Scientific Committee (TSSC) set up by France’s Agence national de securité du medicament et des produits du santé (ANSM).

Serious toxicities occurred with the experimental drug BIA 10‐2474, “as if something gave way or swung suddenly at a specific dose or concentration threshold which is typical of an on-off effect,” the TSSC noted in the minutes of a meeting held February 15.

BIA 10-2427 is a fatty acid amide hydrolase (FAAH) inhibitor that was undergoing testing in a first-in-man phase I trial in Rennes, France. A total of 64 healthy volunteers aged 18-55 years participated in a single ascending dose stage to test eight doses (0.25 to 100 mg) versus placebo. This phase of the study ran from July 9 to October 9, 2015, with no significant toxicities reported. For the multiple ascending dose study that began on October 6, 48 healthy volunteers were scheduled to receive either placebo or 2.5 mg, 5 mg, 10 mg, 20 mg, 50 mg or 100 mg for 10 days. The next dose level was initiated only if no toxicities were observed with the lower doses. The 2.5-mg dose was estimated to be 1/40th of the highest dose with no observable adverse effect level based on pharmacokinetic testing.

The six volunteers on active drug in cohort 5 (50 mg) received the first dose on January 6, 2016 (two others received placebo). On the evening of January 10 (day 5), one volunteer who had received five consecutive 50-mg doses was hospitalized after presenting with signs and symptoms resembling a stroke. Biotrial, the clinical research organization conducting the study, did not consider the acute symptoms to be related to the study drug and the remaining five volunteers received their sixth dose the next morning. The trial was suspended on January 11. The remaining five subjects were hospitalized between January 13 and 15, 2016.

In five of six volunteers, there was rapid onset of neurological symptoms of varying clinical and radiological severity. One subject died. MRIs in the six volunteers showed consistent evidence of microcerebral tissue damage, described as highly variable in severity, with unusual topography, and primarily affecting the hippocampus and pons. Cerebral hemorrhage and brain necrosis were reported in some study participants. There were no seizures or peripheral neurological symptoms. No biological, metabolic or immunological abnormalities were observed.

No neurological toxicities were observed in subjects in the prior dosing cohorts. In the single-dose study there were 18 adverse events, such as orthostatic hypotension, reflex tachycardia, PR or QT interval prolongation, mild dizziness and headache. Two adverse events of possible significance (diplopia, blurred vision) were considered to be not relevant by trial investigators. Subjects exposed to BIA 10-2427 in the earlier stages of the study have since been investigated and no clinical or MRI anomalies have been detected.

In addition, no neurological toxicity was reported in a phase II trial of PF-04457845, an FAAH inhibitor that was in development as a treatment for pain due to osteoarthritis (Huggins et al. Pain 2012;153:1837-1846).

Among its findings, the TSSC noted that several volunteers were included in the study despite having risk factors for treatment-related adverse effects. One volunteer had a history of severe head injury with loss of consciousness; and another had a prolonged PR interval (>240 msec) on ECG and blood pressure >140/90 mmHg on four pre-dose readings. Other problems with the trial included too rapid escalation from 20 to 50 mg/day, and a lack of neuropsychological assessment.

The committee found that toxicology studies in four species were conducted properly and no neurological toxicities were found despite the use of high doses administered for prolonged periods. However, there were reports of cerebral damage, especially in the hippocampus with gliosis and inflammatory cell infiltration, in rodents receiving high doses (150-500 mg/kg over four weeks). Such damage was considered by the committee to be common in rodents and did not constitute a safety signal.

The TSSC noted that pharmacokinetic studies demonstrated that complete inhibition of FAAH could be achieved at doses from 1.25 to 5.0 mg, and questioned the need to escalate to 80-100 times higher than this if FAAH inhibition was the putative mechanism of action of the drug. Moreover, studies of the single-dose cohorts had demonstrated non-linear pharmacokinetics, with the area under the curve (AUC) increasing more rapidly than dosing. The committee speculated that this was due to a slowing of elimination at doses >40 mg/day. BIA 10-2427 has at least four active metabolites, of which two reach measurable plasma concentrations. Pharmacokinetic variability has been shown to be higher for these metabolites, but little is known about their clinical effects or toxicity. The committee noted that there is a potential for intracerebral metabolism, which could have resulted in CNS effects if one or more of the metabolites was highly reactive or toxic at low concentrations.

The TSSC considered several hypotheses for explaining the observed toxicity. Administration error was considered to be unlikely. A pharmacodynamic interaction was also rejected since the five patients affected were unlikely to have consumed the same foods (e.g. chocolate) or medications, blood tests ruled out the use of narcotics, cannabis and alcohol, and a common genetic factor that might alter FAAH inhibitor response was statistically implausible.

The more likely mechanism, according to the committee, was an off-target effect that has not been identified. BIA 10-2427 is an irreversible FAAH inhibitor with relatively poor specificity for FAAH, so there may have been interactions with other CNS hydrolases. This proposed effect might be augmented due to the non-linear pharmacokinetics at doses >40 mg.

Alternative mechanisms were also suggested. FAAH is a hydrolase that degrades anandamide, which preferentially interacts with the CB1 receptor. However, anandamide’s effects are not limited to the endocannabinoid system. It affects ion channels by activating TRPV1 (transient receptor potential vanilloid 1); is an agonist of PPAR (peroxisome proliferator-activated receptor), which is involved in inflammation; interacts with NMDA (N-methyl D-aspartate) glutamate receptors; and can activate transcription pathways (e.g. MAPK) involved in apoptosis. Stimulation of endocannabinoid receptors by anandamide is not known to result in the toxicities that occurred. If FAAH is inhibited, anandamide is degraded via the cyclooxygenase pathway, resulting in leukotrienes and prostanoids. The committee noted that these have known vasomotor effects that might be compatible with some of the lesions observed in the affected subjects.

The TSSC investigation is ongoing and the committee plans to produce comprehensive good-practice recommendations on first-in-man trials.

For a free download of the TSSC minutes, 07 March 2016:



Recommend to a Colleague Go back to home page