A Phase 1 drug trial in France has now left one volunteer dead and five others hospitalized– four with neurological damage– at the University of Rennes Central Hospital (CHU Rennes).
A Phase 1 drug trial is the first time an experimental drug is given to humans after a series of exhaustive laboratory and animal studies. Phase 1 trials are conducted in healthy volunteers–those without the medical condition in question–and 84 other patients had safely received the drug since the trial began in July. Biotrial, a clinical research organization with a 150-bed research unit in Rennes, has been the trial site.
The drug in question, code named BIA 10-2474, is under development by Portugal-based pharmaceutical company Bial-Portela & Ca., S.A., for pain relief, a therapeutic area where non-addictive drugs are in great need. Unlike oxycodone working on opioid receptors or naproxen inhibiting cyclooxygenase enzymes, BIA 10-2474 is an inhibitor of an enzyme called fatty acid amide hydrolase, or FAAH.
What's so important about FAAH? Our bodies make several fatty acid amides that include anandamide, a natural stimulator of the cannabinoid receptors upon which chemicals in the marijuana plant act. Anandamide is referred to as an endocannabinoid. The rationale is that a drug blocking FAAH will allow naturally-occurring anandamide to accumulate and act on cannabinoid receptors in a manner that won't produce the psychoactive effects of cannabis.
Other pharmaceutical companies have been developing their own FAAH inhibitors, such as Janssen/Johnson & Johnson's JNJ-42165279 for social anxiety disorder, Merck's MK-4409, and Pfizer's PF-04457845 against osteoarthritis pain, insomnia, Tourette syndrome, and cannabis withdrawal, and Vernalis' V158866.
The diversity of indications reflects the spectrum of actions endocannabinoids have in animal research models. But none of these drugs have been associated with any type of brain injury in human research volunteers.
Few details, best-informed speculation
How and why this tragedy occurred has been a focus of considerable speculation in the scientific and medical communities. But crucial details on the drug involved, such as in vitro and animal studies, are lacking. As of this morning, PubMed lists no publications of any type from Bial on any FAAH inhibitor.
And even the clinical information on the patient death and severe adverse reactions in others is too general to make any significant interpretations. CHU Rennes neurology department head Dr. Pierre-Gilles Edan described MRIs of the patients' brains: The worst showing cerebral hemorrhage and necrosis, "with three others were suffering a 'handicap that could be irreversible' and another also had neurological problems."
Amsterdam-based writer for Science, Martin Enserink, has two reports with the greatest medical details to date. Update: Enserink corrected me on Tuesday morning that while the five surviving men in Rennes are hospitalized, one is without symptoms. That's perhaps the only bright spot in this tragedy and will certainly be valuable in the investigation.
The Bial website does feature the company's commitment to transparency of clinical trial information and data sharing upon completion of human studies. To their credit, they have registered 93 of their drug trials at clinicaltrials.gov, mostly those that have been completed or terminated. Many are with their drug for epilepsy, eslicarbazepine acetate (BIA 2-093), that was approved in the United States in November 2013 and sold under the brand name Aptiom by Sunovion Pharma. However, no trials with BIA 10-2474 are listed.
The Paris prosecutor's office has opened an investigation, according to Agence France-Presse, so any further technical information on the drug is unlikely to be made public.
What we do know can be gleaned from 1) the drug's likely chemical structure and 2) the development and testing of other FAAH inhibitors that have progressed to clinical trials.
Drug is an irreversible FAAH inhibitor
Chemical informatics expert Christopher Southan has been the best source of the most-likely chemical structure of BIA 10-2474, having examined patents and patent applications from Bial (although he notes that an anonymous Wikipedian posted the structure prior to him). Southan cites the comment thread at In the Pipeline, the highly-read blog by industrial medicinal chemist Derek Lowe, PhD, at Science Translational Medicine. One discussion point has been that the drug is likely to be an irreversible FAAH inhibitor, meaning that the drug makes chemical bonds with the enzyme to inactivate it.
Irreversible inhibitors are not necessarily a bad thing–that's how aspirin works. In fact, scientists writing in a 2011 paper in Nature Reviews Drug Discovery discussed a resurgence in this covalent binding class of drugs across therapeutic areas.
But any time a small molecule is designed to irreversibly bind to an enzyme, an opportunity arise for the immune system to recognize it as foreign. In rare cases, the immune system will mount an inflammatory response against it, but also against regions of the normal enzyme surrounding where the drug binds. This response can result in a hypersensitivity or autoimmune reaction in a subset of individuals whereby the immune system attacks everywhere the enzyme is present.
One example of an autoimmune drug reaction is the liver damage that can occur with the old inhaled anesthetic, halothane. The liver metabolizes halothane to a compound that forms covalent chemical bonds with liver proteins and, in a very small subset of patients, the immune system attacks the liver.
A similar reaction can occur with people who are hypersensitive to some types of penicillins.
But none of these reactions have been observed in the brain, as with BIA 10-2474. Most drug hypersensitivity reactions occur in the liver, the skin, and in the bone marrow, where some drugs can cause potentially-fatal aplastic anemia.
Scientists at Merck who developed their FAAH inhibitor, called MK-4409, made a case against developing an irreversible inhibitor in a 2014 ACS Medicinal Chemistry Letters publication,
More recently, however, several scaffolds have been disclosed as reversible noncovalent modifying inhibitors of FAAH. Aminopyrimidine and sulfonamide are chief among these novel classes of FAAH inhibitors. This approach, in our view, would decrease potential safety concerns over the creation of a long-lived covalent adduct between a compound and the FAAH enzyme. [emphasis mine]
"Off-target" effects
Several writers, including Forbes contributor Judy Stone, MD, have pointed to so-called off-target effects that may have led to brain injury in this group of trial volunteers. What's meant by this is that the drug may have had effects on an unintended biochemical process. For example, FAAH belongs to a family of 200 enzymes called serine hydrolases, some whose function is not yet understood. In addition to binding these other enzymes at higher doses, the drug might bind to other cellular proteins that affect blood flow in the brain.
The potential for these off-target effects has clearly been well-appreciated by Janssen/J&J and Pfizer research teams with their respective drugs. For example, Janssen tested their FAAH inhibitor against 50 enzymes, receptors, transporters, and ion-channels as well as the major drug-metabolizing enzymes of the cytochrome P450 (CYP) family.
Pfizer and their collaborators were even more comprehensive in making sure their drug didn't attack proteins other than FAAH at concentrations higher than would be expected in the body. PF-04457845 was examined against in the serine hydrolase superfamily, several CYP drug-metabolizing enzymes, and a panel of 68 receptors using an in vitro pharmacology screening service offered by the French company, CEREP.
Sean Ekins, PhD, who writes at Collaborative Chemistry, has run the BIA 10-2474 structure through two computational algorithms to identify potential protein targets other than FAAH to which the drug might bind. While he stresses that these are simply computational analyses that have not been confirmed experimentally, Ekins writes
"These high scores for many protein targets in humans could suggest the molecule is highly promiscuous and there may not be a single pathway interfered with. Vesicular acetylcholine transporter is slightly lower down in the list which also makes you wonder how many GPCRs might be impacted too. For now this is all idle speculation until we hear more about exactly what happened. Perhaps this compound could be profiled both computationally and experimentally to answer these questions of what the target/s of the toxicity are."
If Bial performed similar experiments for their drug–biochemical or computational–no similar data has been published or otherwise made available publicly.
Other possibilities: Drug impurities?
This speculation is mine entirely and derives from a close examination of the U.S. patent application that most commentators believe is for BIA 10-2474. The inventors at Bial describe that in the chemical synthesis of their top FAAH inhibitor (denoted "compound 23" and "Formula A" in the application), an additional purification step must be done to remove an "N-acetylated aniline impurity."
Aniline caught my eye because it was used in the manufacture of dyes and other industries. Occupational exposure to high concentrations of aniline can change hemoglobin to methemoglobin and impair oxygen delivery to tissues. As a result, acute aniline poisoning first affects organ systems with high oxygen requirements, such as the brain and the heart.
This is an intellectual leap, of course, but if a drug impurity were present in the batch of study drug used for this trial, those receiving the highest doses might have been exposed to an aniline relative capable of causing methemoglobinemia. One would expect that clinicians would have easily detected this since the patient's skin would take on a blue hue.
Longtime British scientist at Pfizer, Dennis A. Smith, PhD, wrote in his textbook, Metabolism, Pharmacokinetics, and Toxicity of Functional Groups that,
"Compounds containing aromatic amines (anilines) induce a variety of toxicological responses including carcinogenicity and hepatotoxicity. Several drugs containing an aniline moiety, which have been withdrawn from the market, have a black box warning on their labels. Therefore, anilines have been put on the blacklist of functional groups that the medicinal chemists generally avoid."
BIA 10-2474 does not itself contain an aniline group (it's a phenylurea for those interested), it's possible that an N-acetyl aniline impurity could produce aniline toxicity. Again, this is pure speculation based on what is thought to be the patent application for the study drug. I contacted Bial for confirmation of the chemical structure of the drug but a response had not been received at the time this commentary was posted.
Other questions: Why did the study have a placebo group?
Forbes contributor Judy Stone, MD, raised an interesting point as to why a placebo group was included in a Phase 1 trial, a stage at which drug efficacy is not among the intended endpoints. One possible reason is that modulating endocannabinoid levels could produce psychoactive side effects, although this has not occurred with the Merck or Pfizer compounds. But since effects like agitation, confusion, dizziness, and headache are observed in a percentage of placebo participants in trials for any therapeutic area, the investigators may have wanted to have a baseline against which the study drug's effects could be compared.
In summary, this rare Phase 1 clinical trial tragedy has generated substantial scientific and medical interest in the precise cause of these severe neurological reactions that have led one patient dying. Even with spotty details on the drug, its intended target, and other clues, it's clear that investigators will have many possibilities to explore in determining why this tragedy occurred.
