I took the report released yesterday by ANSM on BIA 10-2474 and Google translated the whole doc.. Its not perfect but here goes. All the usual caveats apply – that its far from a perfect job but should provide the key points. Would be great to get an official translation at some point.

I find comments in section 5 around the molecule structure of most interest – I would not say N-oxides are common in approved drugs

“Examination of the chemical structure of this molecule evokes a priori nothing
particular, especially regarding potential toxicity. The functional groups and
chemical it contains nuclei are commonly found in medicinal chemistry.
For example, the N-oxide function is found in the chordiazépoxide (benzodiazepine
anxiolytic), minoxidil (potassium channel agonist developed as
6
antihypertensive and secondarily used to develop the hair), and various
antiretrovirals.”

then discussions of selectivity suggesting they tested against very few off targets compared to whats known for other FAAH inhibitors :

An important difference BIA 10-2474 with the majority of previously known inhibitors
regards its low specificity for the target enzyme. Concentrations in inhibiting
50% of FAAH activity (IC50) are on average 1.7 (1.5 to 1.9) micromolar (M)
in mice and 1.1 (0.9 to 1.3) M rats. They are, according to the documents
provided by the Laboratory Bial at most 100 times greater for the various
others vis-à-vis enzymes which the BIA has been tested. The Laboratory has actually Bial
tested compound and its one of its metabolite (BIA 10-2445) and vis-à-vis three
serine hydrolases: monoacylglycerol lipase (MAGL), a carboxyl esterase and
acetylcholinesterase (10 selectivity for FAAH rat or 50 FAAH
human). The other enzymes tested were: the dopamine-beta-hydroxylase,
glutamic acid decarboxylase, the monoamine oxidases A and B and choline acetyl
transferase.
This contrasts with results obtained with other compounds such as PF
04457845 Pfizer (tested vis-a-vis 68 receptors) having an IC50 of 7.2
nanomolar (nM) for human FAAH (240 times lower than that of the BIA 10-
2474) and more than 100 M to a panel of twenty hydrolases. Ratio
specificity Pfizer compound is thus not more than 100, but of the order of 14 000. This
The same applies to the Janssen & Janssen JNJ-42165279 tested vis-a-vis 50 enzymes.
This low affinity / specificity of enzyme-BIA for its target to bring further
consider a “parasite” on fixing other serine hydrolases in the discussion
the mechanism of toxicity observed in humans. Recall that the superfamily
hydrolases serine approximately 300 members and is, therefore, recommended
to develop inhibitors having the greatest possible affinity for the target enzyme.

And then discussion on metabolism – what we do not have here is an idea of what metabolites are seen in the brain:

Nine metabolites supposed BIA have éésynthéisé 10-2474 (BIA compound 10-2639,
10-2583, 10-3258, 10-3827, 10-2445, 10-2631, 10-3844, 10-2580 and 10-3764). All
7
have a structure very close to that of the parent drug. They correspond,
is the reduction of N-oxide, or hydroxylation of the cyclohexane ring (which
results in the formation of more hydrophilic compounds), or the demethylation of
amine or to two concurrent changes. A priori, nothing in these
Metabolites is no indication, in the examination of their chemical structure, toxicity
possible. Three of them have the potential for inhibition of FAAH near
that of the parent molecule. These metabolites are, most of them found
in very small quantities, including after 14 days of administration of the BIA in 10-2474
the animal. The BIA 10-2631 (obtained by reduction of N-oxide and hydroxylation
cyclohexane), however, is found in larger amounts in primates.
During pharmacokinetic studies in man (see below), four
these metabolites have been identified, 2 undetectable and 2 measured at concentrations
Plasma much lower than those of the parent molecule (<3%).
Unless we assume the existence of a very different metabolite or a distribution very
important for tissue (tissue concentration ratio / concentration
very high plasma, explaining the low circulating concentrations), involvement
of these metabolites in the toxicity observed during the clinical trial Rennes seems little
likely.
On another level, remember that the imidazole ring of the molecule BIA 10-2474 in
adjacent the electrophilic carbon site binding to FAAH may be
considered a “leaving group” may give rise to an isocyanate on
which many brain proteins are capable of binding.

and toxicity issues – some confusion apparent in histology and it seems they dismiss the damage seen in monkeys:

However, in rats and mice, brain damage, particularly at
hippocampi with gliosis and infiltration by inflammatory cells were
rated in three animals treated with very high doses. This concerns a male and a
female mice in the study to 500 mg / Kg / 24h for 4 weeks and a rat study 150
mg / kg / 24 h over 4 weeks. These violations, discussed in CSST due to the context,
seem quite frequently observed in rodents in studies of this type and
were a priori not likely to generate a signal that such abuses do
appear to have been observed with other inhibitors of FAAH. For three
animals concerned, as their fellows of the same group reports
observation does not mention the existence of neurological disorders or
behavior.
Similarly, in primates and rats, damage to the brain and especially in
the autonomic nervous system have been observed in some animals treated with high dose.
Thus, in monkeys, in studies over 4 weeks at doses of,
respectively, 10, 50 and 100 mg / Kg / 24h damage of the medulla oblongata (bulb
spinal) type of “axonal dystrophy” have been observed in some animals
Group 100 mg / Kg / 24 and not in those receiving lower doses. He is difficult
to decide on the precise histological nature of the infringement since both
pathologists who reviewed sections of blades on behalf of Bial not
used the same terminology.

Here is the full document

1
Report of the Scientific Committee Specialized Temporary (CSST) “Inhibitors
FAAH (Fatty Acid Amide Hydrolase) “on the causes of the accident occurred at Rennes
in a clinical Phase 1 trial in January 2016.
1. Foreword
The CSST “inhibitors of FAAH” was established by the Director General
the National Health Security Agency of Medicines and Health Products
(ANSM), following the accident January 10, 2016 in Rennes when testing
Clinical Phase 1, first-in-human, BIA 10-2474 molecule.
The missions of the CSST, scientists were based on the data available to them
disposal and the expertise of its members:
– To analyze the mechanisms of action and possible toxicity of the substances,
as the BIA 10-2474, are expected to act, directly or indirectly, through the
endocannabinoid system.
– To formulate and, if possible, prioritize hypotheses to explain the
toxicity observed in several volunteers from the trial at Rennes by Biotrial.
– To adopt, if appropriate, general recommendations for strengthening
Volunteer security, particularly during the first administration of study
to humans (Phase 1).
The mission of the CSST has, since its establishment (January 25, 2016) to render
its report (18 April 2016), organized in three modes:
– A work of individual expertise conducted by each of its members, on
various documents available to them and the relevant literature.
– Two meetings “open” one day (15 February and 24 March 2016) during
which these appraisals were reported. A meeting with the Laboratory
Bial was also held on 18 March 2016. All three meetings were held in
local ANSM, with two inspectors from the Inspectorate General
Social Affairs, IGAS (Christine and Gilles Autume Duhamel). Otherwise,
a representative of the EMA (European Medicines Agency) attended as
an observer, meetings of February 15 (Hans-Georg Eichler) and 24 March
(Jean Marc Vidal); two representatives of the Portuguese Agency of Medicines
(Ana Catarina Fonseca and Isabel Vieira) also participated as
observers at this meeting.
– A phase of exchanges and restricted collective writing to members of the
CSST, which led to the approval of both versions (interim and final) of
this report.
This organization allowed to reserve the discussion of the key points of expertise,
its conclusions and recommendations arising to members of the CSST,
regardless of the organizer (ANSM) and observers (inspectors general
2
IGAS, representatives of European and Portuguese agencies). This restricted phase,
far accounted for the largest part of the labor CSST.
The mission of CSST experts, although it represented an amount of labor
large (estimated at more than 600 hours in total), was not intended to
substitute for an inspection. As such, the findings of this report do not
prejudging those administrative and judicial investigations currently underway.
Regarding the source documents used include in the preamble Brochure
BIA investigator 10-2474, drafted by the Laboratory Bial has
translation imperfections, transcription errors, particularly in the tables and
FIGS. This is, in many places, such as to generate ambiguities and difficulties
understanding, including vis-à-vis important information (see Chapter 6).
This deserved to be reported because the Investigator Brochure is a reference document
the evaluation of a health product as recalled inter alia by the Order of 19 May
2006 ( “on the content and presentation of the terms of a brochure
the investigator of biomedical research on a human drug “).
Finally, although the CSST was introduced by decision of the Director General of ANSM,
and has received logistical support from the Agency, the CSST conducted during the two
half months of its existence, its work and investigations in total independence,
especially vis-à-vis the ANSM, the Bial Laboratory, Biotrial the center, volunteers
who participated in the test, their families and their advocates.
All CSST experts, worked as volunteers during the entire
their mission.
The drafting of the various versions of this report have been subject to alone
Experts of the CSST and the many exchanges necessary for its completion and attainment
a consensus on the key points of the dossier have always been restricted to this
configuration.
2. Composition of the CSST
Bernard Bégaud (Medical Pharmacology. University and the University Hospital of Bordeaux. INSERM CR 1219), Marie
Germaine Bousser (Lariboisière Hospital, Assistance Publique Hôpitaux de Paris, Université Paris
Diderot), Pascal Cohen (Internal Medicine, Hôpital Cochin, Paris), Bertrand Diquet (Pharmacology
Medical and Toxicology. Health Department Medical Faculty. University and CHU d’Angers), Pierre Duprat
(DVM, PhD in toxicology, European College of Veterinary Pathologists) Walter
Janssens (Federal Agency for Medicines and Health Products, Belgium), Michel Mallaret
(Clinical Pharmacology, Pharmacovigilance Regional Centre and the Drug Information, CHU
Grenoble), Guy Mazué (DVM), Joëlle Micallef (Medical Pharmacology, Aix
Marseille University and the University Hospital of Marseille, CNRS UMR 7289 Institute of Neurosciences Timone), Claude
Monneret (Director Emeritus of Research at CNRS, President of the Academy of Pharmacy), Jean
Louis Montastruc (Medical Pharmacology and Clinic. Faculty of Medicine and University Hospital of Toulouse)
Laurent Venance (Centre for Interdisciplinary Research in Biology, College of France, INSERM
U1050, CNRS UMR7241, Labex Memolife, Paris).
3
3. Context
The molecule BIA 10-2474 of Bial Laboratories (Portela & Ca, Portugal) belongs to the
family inhibitors of FAAH, degrading enzyme anandamide, acting biolipid
as mediator in the endocannabinoid system said.
More than a dozen of such inhibitors are, or have been developed, no
that have, to date, been marketed; for many, due to a deemed effective
disappointing. These inhibitors belong to the structural chemistry plane, essentially
two families:
– Molecules with a urea function: URB 524 and 597 and compounds
especially developed by Sanofi-Aventis, Astellas, Bristol-Myers Squibb
Janssen & Janssen.
– The molecules having a carbamate function, developed in particular by
Sanofi-Aventis, Vernalis, Pfizer and Bial.
Research in the field of inhibitors of FAAH was driven by hopes
important and the prospects for a variety of therapeutic indications: pain,
vomiting, anxiety, mood disorders, Parkinson’s disease, Huntington’s
Huntington, various cardiovascular indications, to name a few.
For the product Bial, the Investigator Brochure states that the BIA was 10-2474
developed “for the treatment of conditions for which there would be advantage
increase the concentrations of endocannabinoids. ”
The indication seems to have been favored, at least initially, is the
neuropathic pain; this was confirmed by the laboratory during its Bial
hearing on March 18, 2016.
first-in-human studies were entrusted to Biotrial Research to
Rennes, specialized center for almost twenty years of investigations and research
of that type. The accident occurred in mid-January 2016 led to the suspension of
10-2474 clinical development of the BIA. Gravity and have spectacular
profoundly influenced the drug workplace, scientists and
public, both in France and in the World. Understand the circumstances and, if possible,
of this unprecedented accident occurred mechanisms is a collective priority and
justification of expertise of the work of the CSST.
This expert report, after a recall on the endocannabinoid system (prerequisite
needed to introduce the discussion on the mechanisms of action of the molecule and
assumptions about its toxicity) will address the analysis of the molecule, its
pharmacological properties and the toxicity studies conducted in animals, the
protocol implemented by Biotrial, symptoms observed in healthy volunteers
testing and pharmacodynamic and pharmacokinetic data. The second
part will explore the possible explanations for the accident at Rennes.
A conclusion will summarize the opinions and positions of the SSTC on key points of the record. The
report will conclude on recommendations relating to the conduct of studies
first administration to humans, the CSST wishes to bear at European level
and international.
4
4. Reminder of the endocannabinoid system
BIA 10-2474 is an inhibitor of FAAH, serine hydrolase degrading anandamide,
one of the major mediators of the endocannabinoid system said. This system, the
ambiguous name (it is actually much broader and complex than target
action of cannabis derivatives) are among a large number of species (vertebrates and
invertebrates, except insects) and, particularly, in mammals. He’s from
Recent knowledge (the first receptor was identified by cloning in 1990) and
still incomplete.
There are two types of receptors (CB1 and CB2), and coupled transmembrane
G proteins (inhibiting adenylyl cyclase).
– CB1 is a ubiquitous presynaptic receptor found on the surface of
several types of cells (neurons, astrocytes, pericytes, cells
endothelial) and in a large number of brain maps (ganglia
Basically, cerebellum, hippocampus, cerebral cortex, olfactory bulb, etc.). CB1 is one
receptors coupled to the G proteins whose expression level is the highest
in the central nervous system, with the notable exception of the brainstem. We
also found CB1 periphery (lung, intestine, testis, uterus,
etc.). The exogenous agonist type of this receptor is tetrahydrocannabinol (THC).
– CB2 is found primarily in the immune system cells
(Immunomodulatory effects).
To date, eight endocannabinoids have been identified. These are bioactive lipids acting
both as neurotransmitters and neuromodulators and synthesized and released “to the
demand ‘unlike conventional neurotransmitters that are released from
storage vesicles.
The three main endocannabinoids are:
– Anandamide (AEA); historically, it was the first endocannabinoid
characterized (1992),
– 2-arachidonylglycéol (2-AG) ester of arachidonic acid,
– 2-AG eher (eher arachidonic acid).
As THC, the AEA has affinitéprééentielle for CB1 and low réepteur trè
for CB2. Conversely, 2-AG has a strong affinitépour both types of réepteurs, it
era can therefore seen as the main méiateur of systèe then endocannabinoïe
that AEA has practically no action on the CB2 and is capable of interacting with
several other systèes. In addition, the 2-AG is retrouvéàdes rate 200 times A800
Eeve that anandamide in rodents.
Unlike 2-AG, anandamide is little spéifique of systèe
endocannabinoïe strict sense and can also be considered a
endovanilloïde. Indeed it is able to activate the TRPV1 receptor (transient
receptor potential vanilloid 1) which are non-selective cation channels in the family
TRP channels.
In addition, the AEA is on other systems:
– It is a good agonist PPAR (peroxisome proliferator-activated receptor) alpha
and gamma nuclear receptors involved in energy metabolism and
inflammation,
5
– Interacts at the NMDA-type glutamate (N-methyl D
aspartate), both as stimulator and by direct action as inhibitor
by indirect action via the CB1,
– Finally, as other endocannabinoids, it can, through the MAP kinase and
a chain reaction, leading to the activation of multiple factors
transcription involved in the phenomena of neuroprotection which
is a way of promising research.
The effects of stimulation of the endocannabinoid system are similar to those induced
by cannabis derivatives. Low to medium concentrations induce
behavioral responses and then mixing stimulatory effects as depressants
high doses, effects are always kind of depressing. Schematically, there
then in animals:
– Antinociception,
– Hypothermia,
– A hypolocomotion.
Working memory is impaired, no effect on the reference memory. The effect on the
anxiety level is biphasic, anxiolytic at low doses and the doses anxiogenesis
high.
At the synaptic transmission, endocannabinoids act so
retrograde (the neuronal postsynaptic element to the pre-synaptic) and
overall effect of reducing the transmission and at short (a few
seconds) or long term (hours or days). They modulate both
excitatory transmission (glutamate) and inhibitory (GABA).
After being synthesized and released by the postsynaptic compartment, the AEA is
normally degraded by FAAH (membrane hydrolase) which also degrades
Part 2-AG but also quite a few other bioactive lipids.
Unlike the animal, two isoforms of FAAH (FAAH FAAH-1 and-2) may exist
in the human species. The prevalence of carriers of both isoforms would be about
38% in the general population and that the holders of the only low isoform
activity (FAAH2) of 5%.
If inhibition of the activity of FAAH, AEA concentrations increase but
Schedule a degradation pathway takes over: that of cyclooxygenase. This results
the formation of eicosanoids: leukotrienes and prostanoids (prostaglandins,
thromboxanes, prostacyclins) which can act on the phenomena of apoptosis and
vasomotor; the vasoconstrictor in the brain of 20-HETE (acid 20-
hydroxyeicosatétraeinoïque) is, for example, known.
5. Review of the BIA molecule 10-2474
Examination of the chemical structure of this molecule evokes a priori nothing
particular, especially regarding potential toxicity. The functional groups and
chemical it contains nuclei are commonly found in medicinal chemistry.
For example, the N-oxide function is found in the chordiazépoxide (benzodiazepine
anxiolytic), minoxidil (potassium channel agonist developed as
6
antihypertensive and secondarily used to develop the hair), and various
antiretrovirals.
The originality of the BIA 10-2474 is, moreover, relative; it can be considered a
“Variation” around previously developed molecules as inhibitors of
FAAH. For example, Pfizer’s PF-3845 also contains a pyridine ring
directly adjacent to the urea function. This compound, effective in vivo and selective, was
proved to be a potent inhibitor of FAAH, well tolerated in clinical trials
Phase I, but without satisfactory therapeutic efficacy in Phase 2. those
Similarly, imidazole, current in pharmaceutical chemistry, is present in the
compounds developed by Bristol-Myers Squibb (carbamate inhibitors).
However, in the case of BIA 10-2474, and which ring is positioned adjacent to the site
electrophile of the molecule which (see below) it potentially a “group
runner “.
All inhibitors of FAAH and developed based on the formation of a bond
covalent between the serine 241 of the hydrolase and the electrophilic carbon of the carbamate or
urea. Inhibition of FAAH can thus be considered as irreversible. according to
Laboratory indications of Bial, the BIA 10-2474 would bind well to the way of FAAH
covalent (therefore irreversible) in vitro but partially reversible manner in vivo. This
has already been reported in the case of Janssen & Janssen inhibitor (JNJ-42165279) with
wherein a portion of the enzymatic activity is found after 8 hours.
An important difference BIA 10-2474 with the majority of previously known inhibitors
regards its low specificity for the target enzyme. Concentrations in inhibiting
50% of FAAH activity (IC50) are on average 1.7 (1.5 to 1.9) micromolar (M)
in mice and 1.1 (0.9 to 1.3) M rats. They are, according to the documents
provided by the Laboratory Bial at most 100 times greater for the various
others vis-à-vis enzymes which the BIA has been tested. The Laboratory has actually Bial
tested compound and its one of its metabolite (BIA 10-2445) and vis-à-vis three
serine hydrolases: monoacylglycerol lipase (MAGL), a carboxyl esterase and
acetylcholinesterase (10 selectivity for FAAH rat or 50 FAAH
human). The other enzymes tested were: the dopamine-beta-hydroxylase,
glutamic acid decarboxylase, the monoamine oxidases A and B and choline acetyl
transferase.
This contrasts with results obtained with other compounds such as PF
04457845 Pfizer (tested vis-a-vis 68 receptors) having an IC50 of 7.2
nanomolar (nM) for human FAAH (240 times lower than that of the BIA 10-
2474) and more than 100 M to a panel of twenty hydrolases. Ratio
specificity Pfizer compound is thus not more than 100, but of the order of 14 000. This
The same applies to the Janssen & Janssen JNJ-42165279 tested vis-a-vis 50 enzymes.
This low affinity / specificity of enzyme-BIA for its target to bring further
consider a “parasite” on fixing other serine hydrolases in the discussion
the mechanism of toxicity observed in humans. Recall that the superfamily
hydrolases serine approximately 300 members and is, therefore, recommended
to develop inhibitors having the greatest possible affinity for the target enzyme.
A proteomic screening would probably have brought relevant information on
this plan.
Nine metabolites supposed BIA have éésynthéisé 10-2474 (BIA compound 10-2639,
10-2583, 10-3258, 10-3827, 10-2445, 10-2631, 10-3844, 10-2580 and 10-3764). All
7
have a structure very close to that of the parent drug. They correspond,
is the reduction of N-oxide, or hydroxylation of the cyclohexane ring (which
results in the formation of more hydrophilic compounds), or the demethylation of
amine or to two concurrent changes. A priori, nothing in these
Metabolites is no indication, in the examination of their chemical structure, toxicity
possible. Three of them have the potential for inhibition of FAAH near
that of the parent molecule. These metabolites are, most of them found
in very small quantities, including after 14 days of administration of the BIA in 10-2474
the animal. The BIA 10-2631 (obtained by reduction of N-oxide and hydroxylation
cyclohexane), however, is found in larger amounts in primates.
During pharmacokinetic studies in man (see below), four
these metabolites have been identified, 2 undetectable and 2 measured at concentrations
Plasma much lower than those of the parent molecule (<3%).
Unless we assume the existence of a very different metabolite or a distribution very
important for tissue (tissue concentration ratio / concentration
very high plasma, explaining the low circulating concentrations), involvement
of these metabolites in the toxicity observed during the clinical trial Rennes seems little
likely.
On another level, remember that the imidazole ring of the molecule BIA 10-2474 in
adjacent the electrophilic carbon site binding to FAAH may be
considered a “leaving group” may give rise to an isocyanate on
which many brain proteins are capable of binding.
Contacted with liver microsomes, the BIA 10-2474, at least until the
concentration of 30 g / mL, only inhibits cytochrome P450 2D6 bit and 3A4 and
appear to inhibit cytochromes P450 1A2, 2A6, 2B6, 2C8, 2C9, 2C19 and 2E1. He … not
seem to act as inducer, at least on the cytochrome P450 2B series and
3A, with a doubt for the 1A2.
6. preclinical pharmacodynamic data
Vis-à-vis the inhibition of FAHH (presented as a central mechanism for its
pharmacological activity), we can consider it as a compound BIA 10-2474:
– On the relatively low activity. The 50% inhibition of FAAH in vitro application, by
example, concentrations in the medium micromolar while
for the majority of inhibitors developed to date, lies in the
nanomolar. As mentioned above, the IC50 BIA 10-2474 for FAAH
appeared 240 times higher than that of PF-04457845 Pfizer, some
inhibitors are characterized by even lower IC50.
– No specific. Inhibition (always in vitro) other enzymes is achieved
for concentrations 50 to 100 times those inhibiting FAAH. It is possible (this
have not been tested) this ratio is even lower vis-à-vis other
brain hydrolases. By comparison, this ratio is, as we have seen,
around 14 000 for the PF-04457845.
– On some progressive action. The curves dose-effect (by inhibiting the activity of
FAAH) of the BIA 10-2474 are unique in that we pass in the beach
unusually narrow concentrations of the absence of inhibition to an inhibition
8
almost total. Although it should take into account the experimental variability,
slope of the dose-response curve seems so high, very steep, when compared to
those of other enzyme inhibitors and, more generally, to other
drugs.
– On prolonged action. Even if we consider that the BIA 10-2474 is not
typically an irreversible inhibitor, inhibition of FAAH it entails is
extremely extended. It is still virtually complete after 8 hours.
In humans, this inhibition continues largely beyond 24 hours,
while the plasma concentrations of BIA have become lower than the
quantification limit of the analytical method used (ie not
measurable).
Regarding the dose at which the BIA inhibits the activity of FAAH, there
apparently a significant discrepancy between what could be extrapolated to
From animal studies and what has been observed in the trial volunteers
Rennes. Animal data show that from a dose of 0.3 mg / Kg
in monkeys, it reached the maximum effect; a dose of 1 mg / kg does not allow
to increase the inhibition of FAAH, or concentrations of anandamide. This
(Not detailed calculations in this report, but verified by the CSST) permitted
predict that total inhibition of FAAH would be obtained in humans
a dose between 10 and 40 mg. However, one observes the
Biotrial volunteers from the essay that about 50% inhibition is obtained for a
0.25 mg and nearly 100% for a dose of 5 mg. This corresponds to
a ratio of at least 10 between the estimated dose from animal data
that measured in humans.
As mentioned earlier, the inhibition obtained is extremely
prolonged because the recovery of activity is not complete after 72 hours
administration, while the product has almost completely disappeared at the
plasma.
As regards the analgesic activity BIA 10-2474 (therapeutic potential has
privileged priori), two tests typically used to test this potential, have
was performed in mice:
– The Formalin paw test. This test involves the injection of a formalin solution 5%
in the plantar posterior end of the dough. This aggression results in
Persistent pain motivating reflexes licking repeated by the rodent.
The efficacy of the molecule under study is evaluated by the decrease in
licking (licking score) during a period (here: 15 to 50 minutes) following
injection. Three doses of BIA 10-2474 (0.3 mg, 1 mg and 3 mg / Kg) were
compared, either administered alone or in combination with 5 mg / Kg of AEA
also administered alone. Gabapentin at a dose of 300 mg, has been used
as an analgesic reference for this comparison; this acid derivative
GABA is used as an anti-epileptic and in pain
chronic.
On this test, the effect of the BIA 10-2474 appears alone but extended amplitude
modest. Indeed, for three increasing doses and compared to the
inactive solvent, the score decreases of 29%, 28% and 41%. The effect of
AEA is the only of the same order of magnitude (35%). In contrast, it is observed
Indeed scored and dose dependent for the BIA + association AEA scores
9
decreasing by 42% for combination with 0.3 mg / Kg of BIA, 65% for
association with 1 mg / kg and 86% for the combination with 3 mg / kg. On this test,
gabapentin appears significantly more effective than the BIA since variation
is 76% against 41% for the BIA only to its higher dose.
Note that in the Investigator Brochure corresponding figure (4.6) is
doubly wrong with the document source provided by the investigator
(Porsolt & Partners, Report No. 09770/2, 2010): the vertical axis of the figure is
seconds then it is of score values and the corresponding column
score of gabapentin (though presented as reference for this test
Comparative) has been removed from the histogram.
– The Tail flick test (tail reflex retraction assaulted by a source
heat). For this test, a significantly higher dose of BIA 10-2474 (10 mg / Kg)
was used. The anti-nociceptive effect is greatest in the eighth hour, but as
in the previous test, modest amplitude (the withdrawal period passing by
average, about 4.8 seconds at almost 6, a difference of 1.2
second) but longer (a persistent difference in the 72nd time). On the base
the same test, the BIA and URB 597 (inhibitor of FAAH of the family
carbamates), both administered in a dose of 1 mg / Kg, were compared vis-à-vis
potentiation of the analgesic effect of AEA. this potentiation
was observed with the compound Bial.
The dose levels used in these tests differ greatly (from 0.3 to 10 mg / kg) without
be possible to trace dose-response curve or estimate an effective dose 50 (which is a
Surprisingly gap). We take note that :
– The anti-nociceptive activity of 10-2474 BIA remains modest when the molecule is
administered alone (assumed conditions for its future use
therapeutic). The BIA potentiates however strongly the action
antinociceptive AEA.
– The antinociceptive activity of the BIA increases slightly (score ranging from 29% to 41%)
when going from 0.3 to 3 mg / Kg, for the formalin test ie when
majorises the dose by a factor of 10. This could mean that the inhibition of
the enzyme involved in this effect is almost complete at a dose of 0.3 mg / kg and,
therefore, that the range of doses tested, too small and / or poorly chosen, not allowed
not to determine an effective dose accurately. This would be a
Another shortcoming rational vis-à-vis the issue of choice of doses during
subsequent development steps.
– The anti-nociceptive activity or potentiating the AEA by the BIA 10-2474 is, in
contrast, sharp and prolonged, whatever the tested dose.
The relative poverty of pharmacology data in relation to other studies
Preclinical justify a recommendation for the CSST regarding
development of new drugs.
7. animal toxicology data
7.1. preliminary remark
10
Data interpretation of the toxicology studies is always complex. These
studies were conducted with doses that can be very high, disproportionate
with those used in humans. Therefore, at higher doses, demonstrations
Toxic various, often non-specific, or visible (macroscopic or microscopic)
Only after sacrifice, were observed in the majority of animals.
In an accident of the type of Rennes, there exists a high probability that
found a posteriori in all animal data pretend elements
consistent with the desired type of toxicity. This does not mean that these
factors constituted predictive signals of toxicity of this type. To avoid this
through classical interpretation, the CSST has carefully considered the matter,
particularly vast, studies in animals; it must be seen in
whole and in context.
7.2. Toxicology file
Studies with the BIA 10-2474 appear to have been carried out according to the protocols
currently validated (including ICH guidelines), with a very pure product (more
99.9%), identical to that used for making capsules administered to
Volunteers Biotrial center.
The studies in question, which is very uncommon (especially for a non molecule
particularly innovative), four different species (rat, mouse, dog and monkey) and
were conducted in two centers reputable Harlan Laboratories SA
Spain (non-rodents in studies) and anapath GmbH in Switzerland (studies in
rodents).
Another species (rabbit) was, moreover, used for studies concerning the effects
potential BIA 10-2474 on fertility and reproduction.
During his hearing by the CSST March 18, 2016, the Laboratory Bial explained that
particularly developed toxicology program was linked to a shift in the
Start of clinical development, this has allowed for studies of
Additional toxicology or continued after the first administration to man
(Studies of carcinogenesis for example). In fact, the CSST has not recovered in
all the data he had to analyze, element can accept the hypothesis
it particularly expensive and full toxicology program was
undertaken because of doubts about the good tolerability of the molecule.
On the basis of data that have been analyzed to date and generally, and
to very high doses, there is no noticeable toxicity BIA 10-2474 targeted on a
particular organ and that should have been taken into account as a signal. One of the
toxic effects usually found in the treated animals, as several
other inhibitors of FAAH, affects sperm and a more
General, gametes. This point, perhaps the most net folder, should be taken into
account if the BIA 10-2474 was taken to use as a medicament.
Note that the Laboratory Bial, contrary to what has been done for several other
FAAH inhibitors, has not defined organ (s) Target (s) in its program
toxicology.
The sensitivity of the assay methods used in toxicology studies
possible to identify the periphery (plasma compartment) five metabolites
11
among which the new BIA 10-2474 can give birth. These metabolites are a
ante identical to those found in humans and also products in amounts very
low (of the order of 1% of the parent product), this for the four species. Thereby,
no specific toxicity study these metabolites was statutorily
mandatory and has in fact been conducted.
There is no noticeable accumulation of the product or its metabolites during the studies
with repeated doses (13 weeks).
Doses NOAEL (no observed adverse effect level) and NOEL (No Observable Effect
Level) appear to have been correctly specified. They vary according to the species
tested and interestingly, depending on length of administrations, especially in mice. The
NOAEL for studies 4 weeks and 3 months, are thus, respectively:
– 100 and 25 mg / kg / 24h in mice
– 30 and 10 mg / kg / 24h in rats,
– 50 and 20 mg / kg / 24h in dogs
– 100 and 75 mg / kg / 24h in monkeys.
On the basis of the NOAEL and calculated with reference to the procedures of the Food and
Drug Administration (FDA), it was a priori logic to test a dose of up to 100
mg in humans (96 mg according to the calculation made by the CSST).
As in any toxicology protocol, the organs of animals provided in Protocol
(40 elements) were systematically examined macroscopically and
microscopically without, according to reports from study centers, toxicity
notable for a particular organ, let alone common to all four species studied have
was observed. This is also true for the nervous system, both central and
device, including primates.
However, in rats and mice, brain damage, particularly at
hippocampi with gliosis and infiltration by inflammatory cells were
rated in three animals treated with very high doses. This concerns a male and a
female mice in the study to 500 mg / Kg / 24h for 4 weeks and a rat study 150
mg / kg / 24 h over 4 weeks. These violations, discussed in CSST due to the context,
seem quite frequently observed in rodents in studies of this type and
were a priori not likely to generate a signal that such abuses do
appear to have been observed with other inhibitors of FAAH. For three
animals concerned, as their fellows of the same group reports
observation does not mention the existence of neurological disorders or
behavior.
Similarly, in primates and rats, damage to the brain and especially in
the autonomic nervous system have been observed in some animals treated with high dose.
Thus, in monkeys, in studies over 4 weeks at doses of,
respectively, 10, 50 and 100 mg / Kg / 24h damage of the medulla oblongata (bulb
spinal) type of “axonal dystrophy” have been observed in some animals
Group 100 mg / Kg / 24 and not in those receiving lower doses. He is difficult
to decide on the precise histological nature of the infringement since both
pathologists who reviewed sections of blades on behalf of Bial not
used the same terminology.
12
Always in monkeys, a clear dose-dependent toxicity (appearing only
the group 100 mg / Kg / 24h), and sharper than the last, concerns an infringement
type of edema Meissner plexus cells of the gastrointestinal tract.
In the group of dogs (Beagles) treated 13 weeks, attention was drawn CSST
by the presence of lung changes clearly visible macroscopically,
confirmed by microscopy and labeled “pulmonary disease / pneumonitis acute focal
and multifocal “. These events seem surprising by their frequency. The
toxicology report submitted by the Laboratory Bial connects these lesions to inhalation
Bronchial BIA 10-2474. This hypothesis seemed implausible to most
CSST experts. Indeed, the envelope of the capsules administered was provided for
resistant to gastric juices making it very unlikely opening and
inhalation powder. The same reason and the method of administration and capsules
the absence of symptoms suggestive also make an implausible
by regurgitation “pulmonary aspiration” (explained for example by neurological toxicity
BIA 10-2474). No alternative hypothesis (infectious contamination,
particular susceptibility of this group of dogs or other) has been preferred by the
CSST for this event, which is one of the highlights of toxicology file
BIA 10-2474. The link with the existence of a high density of CB1 receptors in
lung, can not, without further investigation, be considered
an explanation as to remember. We note, furthermore, the absence of similar symptoms
in the other three species (mouse, rat, monkey) and apparently protests
similar in toxicology studies conducted with other inhibitors of FAAH, y
Included in dogs from the same source. This symptomatology clearly dosedépendante,
has significantly disrupted the planned study plan in dogs. This was to
test (versus controls) over 4 weeks at doses of 20, 50 and 100 mg / Kg / 24h. Because of
significant pulmonary events (two dogs, one male and one female, the
high dose group had to be sacrificed before the end of the study) and signs
of motor incoordination, the doses had to be reduced in the groups 50 and 100
mg / kg / 24h in order to conduct the study to completion. In fact, the dose of 20 mg / Kg / 24 has
was considered the NOAEL. This inability to test higher doses
could explain the use of another kind: the monkey was, a priori, a
model more suitable for the study of an inhibitor of FAAH (better sensitivity
stimulation of the endocannabinoid system) although this has not proved to be the case
the plane of the dose calculations. It seems however that the dog and monkey studies have
began in the same period
Studies have in fact been carried out in primates (cynomolgus macaque or). Can
be due to what was observed in dogs, these studies were preceded (in the
Unlike what has been practiced for other inhibitors of FAAH) of a phase
gradual increase of doses increments (up-titration). This was intended to obtain, according to
Laboratory Bial, a “stabilization of the endocannabinoid system” monitoring
the appearance of symptoms indicating such as hypothermia (the file review
however shows that it never exceeded one degree Celsius), a hypolocomotion
or the decrease in food intake, etc.
No mortality was observed in the long-term study (13 weeks at 75 mg / Kg
after rising dose increments). In contrast, another group is a female
died after escalating doses over 12 days (10, 25 and 50 mg / kg / 24h) followed by 9
13
days of administration of the BIA at 75 mg / kg / 24h. The record indicates nothing special
for this animal.
Similarly, several primates had to be sacrificed “for ethical Reasons” in studies
in ascending doses to test the tolerance of the product at very high doses, the two
animals in Group 1 on the fourth day of the terminal landing to 250 mg / kg, both
Animal Group 2 (125 mg / Kg / 24h) and a female of Group 3 after three
administrations of 60 mg / kg / 24h, other animals surviving to the end of
escalation to 110 mg / kg / 24h. These premature deaths are in primates
however occurred for very high repeated doses that would correspond in doses
equivalent human (Human Equivalent Dose or HED), respectively, to 78 mg / Kg,
39 mg / Kg and 34 mg / Kg. Purely as an HED 78 mg / Kg correspond
about 100 times the highest dose have been tested in humans
repeated dose (50 mg).
At the hearing of 18 March, the Laboratory Bial said it had set a project study
Toxicology BIA 10-2474 in the long term monkey (52 weeks).
The issue of animal studies of globally 10-2474 BIA seems good and
no evidence in the data that the CSST studied constituted a kind of signal
cons-indicate a passage in humans. This is particularly true for toxicity
term neurological damage to the central nervous system and the system
having autonomic affected a small number of animals treated with higher doses.
This character a priori not alarming observed neurological was confirmed
by examining sections of the tissues affected by the experts of the CSST (photographs
Bial provided by the Laboratory following a request by the CSST March 18
2016).
Several comments yet deserve to be made:
– The reasons that could have led Bial to use four different species (2
rodents), which is for a file of this type, unusual (no other
FAAH inhibitor appears to have been studied on this basis) were
discussed at length within the CSST. It is however possible that this may
be the result of a change in the study plan being
Development: Dog crossing the monkey because of poor tolerance
observed in the first (dog and monkey studies, however, seem to have
started virtually the same dates), offset development plan
clinical (response officially provided by the Laboratory Bial). It is
also credible that studies in mice have been, in fact,
designed to determine the doses to be used for the studies (long-term) of
carcinogenesis.
– There should be cautious regarding the conclusions that could be
from the linkage of the doses tested in animals to those
administered to the volunteers of Biotrial test (single administration or
repeated). To compare the densities of exposure between animal and man,
are the ratios of plasma concentrations and, more specifically, areas
under the curve that should be used. Nevertheless, the doses used
animal did not warrant a reservation as to the validity of the file
Preclinical. According to data provided by the Laboratory at the request of the CSST,
the reports of the areas under the curves animal / man, when values are
available, are close to 1 with the exception of monkey in which the margins are
14
higher (ratio of about 6). This confirms that the extrapolations of the animal to
man, in terms of margin of exposure and dose tested, must
consider all the species used and not only that
appearing to be the best model pharmacologically.
– If it is not possible to release signal on a specific toxicity,
tolerance BIA 10-2474, in toxicology studies, and generally seems
clearly worse than other inhibitors of FAAH
previously developed. Apart from the specific issue of
pulmonary symptoms in dogs, other inhibitors (despite
sometimes higher doses ratios than in the case of BIA) did not appear
cause toxic effects deserve to be commented on. In particular,
violations at the Meissner plexus in monkeys have not been found
and no premature sacrifice was necessary. This tolerance is better,
for example, demonstrated by the fact that for one of the products, the lack of effect
toxic observed led to use the highest dose to define
the NOAEL.
8. The clinical trial in Rennes by Biotrial
The trial, Phase 1 single-center, first-in-man (First in Humans
or FIH) expected to involve a total of 128 healthy volunteers, men and women aged
18 to 55 and included four parts:
– A study single ascending dose (SAD or Single Ascending Dose)
– A study in increasing repeat dose (MAD or Multiple Ascending Dose)
– An open study of interaction with food and
– A study of pharmacodynamics (unrealized).
Note that the dispersion of volunteers recruited ages (18-55 years) is high,
some being relatively old, compared to what is typically observed in
Phase 1 studies, the first administration to humans. The ages ranged from
example, from 27 to 49 for six patients hospitalized at the University Hospital of Rennes. One note of
Furthermore, the inclusion of several volunteers that can be categorized as
potential risk factor vis-à-vis certain ADRs. We
notes, inter alia, a PR interval measured at more than 240 milliseconds on several
electrocardiograms in pre-dose and a blood pressure higher than 140/90 mm
Hg in four steps in the pre-dose.
The choice of the first dose (0.25 mg) for the SAD part was cautious because
corresponding to about 1 / 400th of the highest no-effect dose (NOAEL)
in animals.
The SAD1 part involved 64 volunteers in eight cohorts of 8 volunteers (six receiving
verum and two placebo) corresponding to the 8 levels of tested doses (0.25 mg to
1 As a reminder, the 2006 recommendations of the French Medicines Agency
(AFSSAPS) for the first administration to man (page 4):
“In the same group:
• the number of volunteers receiving simultaneously the new active substance. It is necessary, except
substantiated justification, to limit the number of volunteers receiving simultaneously the new
15
100 mg); 48 subjects were therefore exposed to the verum. For each landing, two subjects (an
verum and placebo) were tested before administration to the other six.
Part MAD included six cohorts of 8 volunteers (six verum and placebo two) or
48 subjects. The six doses to be tested were: 2.5 mg; 5 mg; 10 mg; 20 mg; 50 mg
and 100 mg. Each dose should be administered 10 consecutive days. Topics
each cohort were to stay 15 days (and 14 nights) at Biotrial center. From the
10 mg dose, the administration has been conditioned to pharmacokinetic data
measured at N-2 (or, for example, those of the 10 mg cohort to start
administration of 50 mg). As is the rule in Phase 1, the passage to the bearing dose
conditioning was superior to no adverse events in volunteers
the previous level after consulting a monitoring committee. Under this condition,
the administration of the molecule is made, at the same time, to all volunteers
of the cohort. According to the protocol, a noticeable intolerance manifestation ( “severe”)
4 volunteers observed in the cohort (50% of the workforce) would result
the interruption of the test. Because of the accident in volunteers cohort
MAD No. 5, only 30 volunteers (of the 36 planned) received verum for this
part of the test in repeated administrations.
Looking for interaction with the food involved 12 volunteers in a dose of 40
mg.
It is therefore, in total, 90 subjects were exposed to BIA 10-2474 over this
test this very different doses (0.25 to 100 mg, a ratio of 400), it
either in a single dose or repeated daily administrations.
Part SAD (single dose) began July 9, 2015 and ends (8 cohort: 100
mg) October 9, 2015.
Part MAD (10 days of treatment) was launched on 6 October 2015. The penultimate
cohort expected (Cohort # 5, or 50 mg) began January 6, 2016 or 19 days after
the end of the Cohort # 4 (20 mg).
Thus an overlap of three days between the two programs. This was
no problem from the standpoint of safety because of the weak (2.5 mg)
the first dose tested in MAD.
The evening of the fifth day (10 January 2016) and therefore the fifth administration
in cohort MAD (50 mg cumulative dose of 250 mg), one of six volunteers who received
the verum is hospitalized at Rennes University Hospital in serious condition. The company has a Biotrial
priori considered the link between the acute symptoms presented by this and the
molecule tested as possible as the other volunteers received their sixth
dose the next morning, 11 January at 8:00 (cumulative dose: 300 mg) without waiting
exploration results (including MRI) during the voluntary hospitalized.
The 5 volunteers verum, not two placebo subjects were successively
active substance based on identified risk factors.
• the deadline between the administration in a voluntary and administration to the next. An observation time
sufficient should be allowed between administrations, including the characteristics of the product,
available data (pharmacokinetics, pharmacodynamics) and the risk factors identified,
• administrative criteria following voluntary
• the stop criteria administrations with previously untreated volunteers “.
16
hospitalized at the University Hospital of Rennes between 13 and 15 January, 2 to 4 days after this
last administration. The test suspension was effective since 11
administrations, which were to continue until 15, were suspended from
that date.
Several remarks on this crucial part deserve to be made:
– The trial of Rennes was conducted in a specialized center (Biotrial) Good
reputation. Apart from the reservation expressed above on recruitment
volunteers, the test is similar to preferred plans now for a
first-in-human / Phase 1. In particular, and contrary to
what has been said in error by several information channels, it
there was no temporal overlap between the different cohorts,
including MAD. On the contrary, a period of time was provided between the end of a
cohort and the beginning of the next.
– While this is not specific to this test, we can only wonder, especially
in the case of a molecule acting through the central nervous system, the balance
intended for the selection, inclusion and tracking of volunteers did not include a
neuropsychological evaluation with clinical interview and cognitive tests. Such a
Evaluation was performed by means of, inter alia, suitable scales for
other inhibitors of FAAH. This will motivate one of the recommendations of the
CSST.
– The dose escalation mode (arithmetic and geometric progressions and
Fibonacci, etc.) in so-called Phase 1 trials is based on neither a consensus nor
on clear recommendations internationally. In the case of the test
Rennes, the mode of progression doses clearly appears problematic because
too brutal end of progression while common sense would have advocated
reverse. For example, jumping between dose cohorts MAD 4 and 5 (20 to 50 mg)
corresponds to a ratio of 2.5 while it is only 2 (2.5 to 5 mg) at the
Cohorts 1 and 2, zone where the risk could be considered low
because of the small difference in absolute value between the two doses. Progression
Rennes is retained at the base of the geometric type (essentially less
conservative than the arithmetic Fibonacci progressions) with a reason
(Multiplier) of 2, more not maintained between 20 and 50 mg. This type of
progress seems to have been used in other trials of inhibitors
FAAH. It seemed important to the CSST that the scientific community and
if regulatory grasp this aspect, important for safety volunteers
(See recommendations at the end of report).
– A difficult issue to be decided in the case of Rennes test is the choice of
the maximum dose expected to be tested in volunteers. Thereof was fixed
100 mg, both single and repeated administration. We saw it
choice appears a priori logic if we extrapolate to humans measured data
Animal: NOAEL extrapolation leads to a very close estimate (96
mg) and that of the concentration inhibiting FAAH leads at a dose
between 10 to 40 mg. Being a non considered at risk by the criteria
usually used (including by the European Medicines Agency)
the use of the NOAEL in the most sensitive species and not of the MABEL
(Minimal Anticipated Biological Effect Level) was statutorily justified. The
problem is the apparently very different human response to this
last point. Indeed, it is observed that the inhibition of FAAH (mechanism
17
alleged the pharmacological activity of 10-2474 BIA) is obtained in humans,
to 1.25 mg and is almost complete to 5 mg. In those circumstances, choose 100 mg
returning to testing a dose 20 to 50 times greater than that assumed effective, which
seems decidedly excessive even if the rule, which means that one does not exceed
first administration to man a dose corresponding to the NOAEL was here
respected. The major point in terms of safety could not be anticipated at
the approval and initiation of the trial (only animal data being
known). However, it would have been logical and expected that the escalation plan
doses should be reviewed in light of pharmacokinetic data from the
volunteers as had been planned for other inhibitors of FAAH. This
was not the case in this study.
9. symptomatology observed in hospitalized volunteers
For obvious reasons of protection and right to respect for privacy
people and medical confidentiality, the information contained in this report were
limited to the useful elements in the context of the mission of the CSST and are presented
to respect the anonymity of volunteers.
One of the files of six volunteers cohort MAD No. 5 have been exposed to BIA 10-
2474 has not been not adopted because the person, considered in a systematic way, not
complained of no symptoms and MRI revealed only a
Image interpreted as “incidentaloma”.
9.1. clinical symptomatology
The first volunteer was hospitalized in the evening of 10 January 2016, the day of the
Fifth, the test product administration. Two other volunteers were
hospitalized January 11 (day of the sixth administration), two other 12 January
(After the last administration) and the last volunteer January 13 or
two days after the last administration.
The main clinical symptoms were noted:
– Headache, present in five volunteers, in very severe one but
however occurring “thunderclap”
– Cerebellar signs in three volunteers
– Disorders of consciousness (with three volunteers) from a slowdown
psychomotor coma (at the deceased voluntary)
– Memory disorders in two volunteers.
Other symptoms were noted only once: diplopia, paresthesia thighs,
hemiparesis with “tremors” of hémicorps without pyramidal syndrome, and pain
stiffness of the spine.
For three volunteers, the evolution of initial presentation was made to
worsening: the first subject passed hospitalized in a state of brain death three
days after the onset of disorders. For the other two, the picture worsened during
three to four days before a stabilization phase (two to three days), then
18
improvement. Both volunteers were guarding however disorders (mainly
cerebellar and memory) as they exit the Rennes University Hospital.
For the two volunteers whose disorders were minor or thereby,
difficult to interpret, no worsening was observed which justified their output
CHU no apparent ill effects.
The four volunteers (except the one who died) were treated from January 13,
methylprednisolone (Solumedrol °) at 1 g / 24 h, with no possibility of
whether the corticosteroid dose was instrumental in improving the picture
clinical.
The clinical presentation of the five subjects is of remarkable character due to
purely neurological symptoms resulting impairment of the single system
Central nervous (but without epilepsy) without any symptoms suggesting an attack
other organs and without any signs of infection. Apart episodes
bradycardia and hemodynamic instability in the deceased subject, no anomalies
heart or blood pressure were noted.
The changing profile of the neurological symptoms and isolated character go
against a vascular processes, tumoral or infectious and are hardly compatible
with an inflammatory process, metabolic or toxic.
9.2. Examinations of blood and cerebrospinal fluid (CSF)
No metabolic abnormalities or blood chemistry were observed in five
volunteers. The immunological tests performed in the most severely affected was found
negative.
CSF was examined in three subjects. It was normal in one of them, a seat
hyperprotéinorrachie isolated from another and quite abnormal, but non-specifically,
in the most severely affected with a hyperprotéinorrachie and the presence of
neutrophils can translate inflammation or necrosis
tissue.
9.3. Results of exploration MRI
In one of the volunteers, MRI, performed 24 h after administration of the test product,
allowed only to find a small hyperintense puncture body
right hippocampus was not present on MRI performed two and
four days later.
For the other four volunteers, MRI allowed to visualize abnormalities
very interesting variable intensity, bilateral and symmetrical way,
seahorses and bridge (protrusion) predominant in its front part
(Sometimes extending bulb or midbrain). In the subject died, and only
home, there was, on MRI performed two days after the initial examination, a breach of
thalamus and cerebral cortex.
The observed anomalies were of the same type but of varying severity, in
four volunteers with:
– 1 intensity on diffusion with a decrease in the diffusion coefficient
Apparent (ADC) reported once and indicating a restriction on broadcasting
water (which may reflect a cytotoxic edema, whatever the cause,
but also cellular infiltration, inflammation, for example). at the
19
voluntary worst hit, hyperintense the broadcast was accompanied,
in the posterior part of the brainstem, an increase of the ADC
could indicate the existence of vasogenic edema.
– 2. A low signal on SWI (Susceptibility Weighted Imaging), interpreted in the
context as reflecting the presence of blood (hemoglobin) in the form
multiple small rounded hyposignals may correspond to
microhemorrhages.
– 3. A Flair hyperintensity (Fluid Attenuated Inversion Recovery), non-specific, and
may be combined with an increase in water content, demyelination,
gliosis or necrosis.
These signal abnormalities are identical seahorses and the bridge in four
volunteers and in the cerebral cortex in the most severely affected. This
suggests that cortical involvement stems from the same mechanism and not related to anoxia
bradycardia episodes.
Signal abnormalities observed therefore reflect the presence of modifications
microstructural with a vascular component (microhemorrhages) that
are not specific to a given mechanism. However, bilateral and topography
symmetrical and the presence of very early in hyposignals SWI allow a priori
to exclude an inflammatory process. That same topography, quite singular,
makes it unlikely a primitive microvascular mechanism and is more compatible
with a toxic / metabolic process.
9.4. comments
The circumstances of onset, the stereotypical character and the evolving profile of achievement
Neurological strongly suggest the responsibility of the tested product. The anomalies
visualized by MRI, consistent with clinical symptoms, suggest
also strongly, due to their signal characteristics and their topography
bilaterally symmetrical, toxic or metabolic mechanism. Accountability of
Product testing in the onset of neurological involvement appears so certain.
The topography quite unusual (and perhaps even unique to this day) lesions
(Seahorses, bridge, thalamus, cortex) is a key element for the search
mechanism by which the test substance was able to generate such effects.
10. Search of toxicity in the other volunteers
One of the most striking elements of the BIA 10-2474 file is the absence of toxicity (from
adverse event of notable intensity, much less severe), especially type
neurological, observed in the trial volunteers other than those of the cohort MAD
# 5. Despite administrations in a single dose up to 100 mg or
administrations repeated 10 times up to 20 mg / 24h, a cumulative dose of 200
mg (reminder: the cumulative doses in volunteers hospitalized ranged from 250 to 300
mg).
20
Among the 76 volunteers (excluding cohorts MAD) who received verum, there was 18
adverse events, of which 11 (frequency: 14.5%) cardiovascular in nature
(Orthostatic hypotension, reflex tachycardia, prolongation of the PR space
QT or electrocardiogram, etc.), as well as cases of dizziness or headache little
severe.
Findings were similar in volunteers with cohorts MAD
the absence of gravity or notable event severity and prevalence
cardiovascular symptoms. Note however that two volunteers cohort
MAD 10 mg had submitted twice, blurred vision (blurred vision). These
episodes lasting, each time, between ten and thirty minutes can not be described,
in the state, diplopia, corresponding symptom, neurology, to a definition
semeiological says. Indeed, this finding was not considered a
relevant event by the investigator and the study and monitoring committee has, moreover,
not been found in volunteers cohorts exposed to higher doses.
Since the test of the suspension, all volunteers (other than those of the
MAD cohort # 5) to which the BIA 10-2474 was administered were contacted again for
a complete clinical assessment and exploration by brain MRI. To date, no
abnormality, or clinical or biological, or MRI, can be related
the administration of the molecule and the test conditions were detected.
Safety data in first-in-human studies
other inhibitors of FAAH appears to have qualitatively the same profile (sedation,
digestive disturbances, orthostatic hypotension, dizziness, etc.) with a yet
difference in frequency of adverse events, although quite varied
greatly from one product to another, appears to be generally weaker than in the case of
BIA 10-2474.
11. Pharmacokinetic Data
Generally, pharmacokinetic studies in animals
call for no particular comment even if, as is frequently the
case, pharmacokinetics appears to become non-proportional, at least in dogs
and monkeys, with higher doses. For example, the ratios of the areas under the curve
(AUC) and the doses administered, assumed to remain constant, is 960 at 0.1 mg / Kg and
1886-1 mg / Kg.
A study with administration (oral and intravenous) product shows marked
very good distribution of the BIA 10-2474 and its large volume of distribution.
Pharmacokinetic studies in volunteers cohort SAD
show that the half-life of BIA 10-2474 gradually lengthens
when the administered doses are becoming higher; AUC, a reflection of the exhibition,
thus increasing faster than the increase in doses. This, from a point of view
purely theoretical, could be explained by the acceleration of the absorption beyond a
certain threshold (type barrier disruption, crossing facilitation, induction
carriers, etc.) or, more likely, by saturation
elimination process for a dose of between 40 and 100 mg, without
21
it is possible to precisely locate the threshold dose at which the primer
disproportionality.
In studies in MAD, the same phenomenon is observed AUC increasing more
faster than the dose from 20 mg. especially we notice that:
– The dispersion of pharmacokinetic parameters between voluntary influences
highly individual AUC values, greater than 50 mg to 20 mg.
– Always to 50 mg, and as it was already, but less sharply observed
to 20 mg, trough plasma concentrations of BIA 10-2474
continue to increase slightly in the fifth and sixth administration. The
Tea equilibrium plasma concentrations do not appear to reach
some of the cohort MAD 5. This strongly suggests that the half-life
elimination BIA 10-2474 is longer in these subjects than
able to predict values of the half-lives of elimination calculated
previously for single doses or lower repeated doses (10
mg). For example, the elimination half-life calculated in the SAD 100 mg group
was on average about 10 hours, value a priori incompatible with what
was observed in cohorts MAD 20 and 50 mg.
– As SAD, non-proportionality is therefore likely from 50 mg
Repeat dose.
The four metabolites identified in animals are expected to be the same in
the man ; two of them (BIA BIA 10-2639 and 10-2445) reach concentrations
measurable plasma but remaining however very low (<3% of those of the product
parent). In the absence of direct administration of metabolites themselves, it is difficult
to determine individual characteristics. However, it seems that the variability
pharmacokinetic parameters is higher for these two metabolites that
observed in animals with, for example, an estimate (very rough due
very low concentrations) the elimination half-life ranging from 4 to 23 hours.
This variability also affected, but to a lesser extent, the pharmacokinetics of
molecule itself. This is commonly observed with drugs due to
variations, among other metabolism between individuals; some (qualified
of outliers) which, vis-à-vis certain parameters be located well away from other
group members. In the case of a Phase 1 study, this variability may become
problem if the calculation of the dose to be administered is based, as is the case here,
on average measures key parameters measured on a small number of
volunteers doses lower bearings. By definition, such an approach takes
not account for the extreme values of subjects expressing a particular response which
distribution may vary from one group to another. This may be such as to induce
quite significant prediction errors (see recommendations at the end of report).
12. Assumptions studied to explain the accident Rennes
The accident occurred during testing of the BIA 10-2474 Biotrial the center of Rennes is of
undeniably amazing and unprecedented nature by
– Seriousness (several volunteers from the same cohort of hospitalized,
one of them died within days of admission)
22
– The fact that apparently toxicology studies yet conducted on four
animal species up to very high doses showed no damage or
table likely to predict a particular neurological toxicity,
– The very unusual nature of the clinical presentation and radiological
the brain damage seen in several voluntary cohort MAD No. 5,
not amounting to anything known to date,
– The fact that no obvious signs, neurological or radiological, this type has been
found among the other volunteers (some having absorbed up to 100 mg
single dose or 200 mg cumulative dose over 10 days),
– Finally, the fact that this accident has occurred with a related molecule
Other compounds (ten) many of which have seen their development
abandoned due to lack of efficacy with no toxicity
neurological or another has been observed.
In terms of formal logic, being a toxicity occurred in one of 14
cohorts of volunteers who received the BIA-2474 could be explained by:
– An error of directors or procedure specifically affecting this
cohort,
– A feature common to the six subjects with signs of toxicity,
– An effect linked to the cumulative dose of IAB 10-2474 that these subjects have received.
The exploration of the first case did not fall within the tasks of the CSST but
this seems very unlikely. The procedures for control of the dose
administered are very strict in Phase 1 studies In addition, the product contained
into capsules administered to all volunteer groups was the same as that
used for toxicology studies and secondarily tested as being
high purity.
The ACST has therefore mainly focused on two assumptions.
12.1. Assuming a common feature to the volunteers of the fifth cohort
MAD
Several options were discussed:
12.1.1. Assumption of an infectious contamination
Volunteers proximity of life of a cohort and consumption
the same food several times a day may accredit this assumption, especially
some infections, possible in this context may have an expression
Central nervous. For example, there is a form of rhombencéphalitique
Listeriosis with location of lesions suggestive that observed in Rennes.
This hypothesis, however implausible. First, this is rare clinical form
humans (the most common neurological form of listeriosis is a
meningoencephalitis). He then took it hits that volunteers
exposed to the verum. These two conditions were making a highly unlikely scenario
statistically. Not least, the analysis of the symptoms presented by the volunteers
and the images seen on MRI are not consistent with this hypothesis.
12.1.2. Assumption of an interaction with other products
23
Mentioned several times in the days following the announcement of the accident Rennes
such interaction could involve drugs, foods (such as
chocolate) or “recreational” substances (alcohol, drugs including cannabis, etc.).
The “drugs” resolutely hypothesis seems unlikely in view of good practice
Phase 1; above all, it must be admitted taking one or more drug (s)
Common (s) by the six patients hospitalized when they were age (27 to 49)
and of very different profiles.
An interaction, pharmacodynamic or pharmacokinetic, BIA 10-2474
or a metabolite thereof with a food also seems implausible. A
analysis of the literature failed to find an example of neurological toxicity
Central, symptoms suggestive of that found in Rennes, attributed to
drug interaction / food, whatever the type. It’s the same for
consumption of chocolates by these volunteers (the accident having occurred in both
weeks after the Christmas holidays). Chocolate contains only very small amounts
anandamide. In addition, overstimulation of the endocannabinoid system is not
known as that can generate a table of the type observed in Rennes (see
far).
Altered metabolism of the BIA by a food or beverage (acting by
example, as an inhibitor or inducer of cytochrome P450 pathway) can be,
as such, excluded but seems implausible.
To date, nothing accredits the hypothesis of drug use by
volunteers immediately before or during their stay in Biotrial Centre. The
results of inspections and investigations will be relevant to that view to
reverse or reinforce this hypothesis. However, apart from the serious distortion
good practice Phase 1 (in terms of supervision of volunteers including)
such consumption, a priori joint, would (it would, in fact, the
consider joint voluntary hospitalized cohort MAD 5 and, further,
nonexistent for the subjects of the lower dose levels), this hypothesis faces the
same argument as before. It seems, in fact, acquired neuroscientist plan
as stimulation, direct or indirect, even massive, receptors
endocannabinoids, especially CB1, can not induce toxicity type
that observed in Rennes. Although they may in some patients induce effects
the serious psychiatric (eg, psychosis), or cannabis, or its
main active ingredient, tetrahydrocannabinol, does induce even
experimentally and high-dose, acute toxic brain lesion reached. In
Clearly, such a consumption could be, at best, but as a cofactor
certainly not as a triggering element and, even less, be the cause of
the accident.
12.1.3. Hypothesis of a genetic peculiarity, metabolism or response
common pharmacological subjects of the fifth cohort MAD
There are many factors, including genetic, capable of modulating
Individual susceptibility to administration of an inhibitor of FAAH. For example,
this enzyme exists, we have seen in two isoforms (FAAH-1 and FAAH-2) Work
different. Similarly, there are several levels of the cytochrome P450 system
whose activity can vary greatly from one individual to another. For attractive it
is this assumption clashes with the laws of statistics. For the example of FAAH, if
Considering that the prevalence of carriers of both isoforms was 38% in
24
general population, the probability of finding 5 among subjects exposed for 6
cohort was 3 out of 100 and 3 chances in 1000 to find her in six
subjects exposed. If one recalculates, which would be more relevant, not 38% but
with 5% (prevalence of carriers of the only isoform low activity), results
then totally exclude such a possibility. It is the same for the probability
have included, by chance, a majority of subjects in a metabolisers
previous cohort which could distort the pharmacokinetic predictions for the
calculate the dose for volunteers cohort MAD 5.
It is therefore imperative that the symptoms presented by the volunteers
Cohort # 5 can only be related to the dose of IAB 10-2474 that was their
administered daily and repeatedly.
12.2. Assumptions around a threshold effect related to the cumulative dose of IAB 10-2474
If this second group of hypotheses seems much more likely, mechanisms
Possible are particularly numerous, some purely remaining
hypothetical. They may involve the molecule itself and / or mediator
as anandamide.
Recall first the very unusual nature of this dose-dependent toxicity, has
priori without signs reported in volunteers who were exposed to
lower doses. Everything happens “as if something or let go suddenly toppled
at a specific threshold dose or concentration. ” Such expression can be compatible
with the fact that the BIA 10-2474 is characterized by a very steep dose-response curve (one
pass, within a fairly narrow range of concentrations, lack of inhibition of the
FAAH in a very complete and sustained inhibition) and by the fact that the pharmacokinetics
BIA becomes non-proportional in humans from a dose between 40 and
100 mg.
Remember also that the BIA has the characteristics of a molecule can bind
on (and thus potentially inhibit) hydrolases brain other than
constituting its pharmacological target. The specificity of the BIA is 10-2474 for FAAH
significantly lower than other previously developed inhibitors. She has,
moreover, was administered to the volunteers cohort MAD # 5 in doses about 10
times stronger than pretending inhibit completely and very prolonged,
FAAH humans.
It is very likely, if not almost certain that the mechanism of injury
occurred in Rennes is to look outside of the endocannabinoid system, especially
more than the stimulation of the receptors by anandamide endocannabinoids can has
priori induce toxicity of this type.
Several mechanisms may raised:
12.2.1. Inhibition other brain hydrolases by the BIA 10-2474
This track and the next (see 12.2.2) are, by their biological plausibility, those in
focus on first analysis. Recall that the BIA 10-2474 was administered in
volunteers cohort MAD # 5 at a dose (50 mg) probably more than 10 times more
stronger than completely inhibit the activity of FAAH. increase
Tissue concentrations beyond those already completely inhibiting the enzyme
may, in any case, increase the pharmacological effect but is likely to favor
25
(Especially with as little specific product) fixated on other hydrolases
serine. This could, for direct or indirect mechanism (known to date), to be
causing brain damage observed in Rennes.
It is interesting to remember that these violations have been observed or a priori at
the animal despite the use of very high doses, nor in six volunteers receiving
BIA 100 mg single dose. They arose from the fifth
day of a daily administration of 50 mg. This finding is in no way
inconsistent with the hypothesis:
– The mechanism involved could be affected, as is common enough
toxicology and pharmacovigilance (cumulative toxicity, for example), a
latency of expression,
– The pharmacokinetic characteristics of the molecule (not pharmacokinetic
proportional, large volume of distribution), detailed above, make
plausible a progressive accumulation of the molecule in the tissues
cerebral. The concentrations could reach there a trigger threshold
fifth or sixth day of administration.
Unlike animal susceptibility / man seems harder to explain. Such
features, in one direction or the other, have, however, been observed. It is not
not without interest to note that in terms of inhibition of FAAH, the
necessary doses were approximately 10 times lower in men than
allowed to await studies conducted in animals.
12.2.2. Toxicity BIA 10-2474 by another mechanism
The starting point would be, as in the previous scenario, an “overflow” of form
Free BIA 10-2474 in brain tissue bound (1) at a dose too strong
vis-à-vis the blocking endocannabinoid FAAH, and, no doubt, (2) to
pharmacokinetic characteristics of the molecule (kinetic becoming nonproportionnelle,
possible to tissue accumulation with repeated administration, etc.).
The difference here is that the pathogenic mechanism would not go through inhibition
other serine hydrolases but by in situ toxicity vis-a-vis molecule
cellular structures involved, for example, in trade at the
blood-brain barrier. A credible hypothesis target the group “runner”
imidazole-pyridine BIA 10-2474 which could either bind to proteins
brain or inhibit the system of cytochrome P450-epoxides resulting in a
vasospasm. Indeed, in the case of BIA 10-2474, and unlike other
inhibitors of FAAH, the imidazole ring is located adjacent to the carbon
electrophilic, seat liaison with FAAH. As before, this assumption
against the fact that no such toxicity was observed in animals despite
the administration of high doses; responses to this objection can be
the same as above.
12.2.3. Toxicity of a metabolite of BIA 10-2474
The toxicity of one of the four metabolites identified in the periphery (plasma) in
man and animal could also be considered. Their specific toxicity has not
was tested by Bial but these metabolites are produced in very small amounts (<3%
the circulating concentration BIA 10-2474) although pharmacokinetic variability
seems higher in men than in animals. All metabolites
identified BIA retain a structure similar to the parent drug and nothing in the
26
folder, only appears for a particular toxicity of one of them. It is, in addition,
as possible (as is frequently the case during the metabolism process) these
derivatives are more hydrophilic than the parent product type. This would tend to make the
crossing of the blood-brain barrier more difficult unless consider
the existence of a specific carrier and / or inhibiting an efflux pump when
rise of circulating levels in repeated outlet. This statement is
however weighting that Bial said he planned to develop
several of these metabolites as inhibitors of FAAH suggesting a
priori a significant crossing of the blood-brain barrier.
It remains to consider the case, purely theoretical, a metabolite with strong tropism
tissue and not quantified in humans due to a ratio tissue / plasma very high and
very low plasma concentrations that result.
Finally, the effect of a genetic polymorphism, making for example produce a metabolite
in larger quantities in some individuals, appears to be excluded for reasons
statistics (need this feature, if it exists, has affected all
volunteers with the cohort MAD No. 5).
12.2.4. Possibility of toxic effects related to anandamide
Blocking FAAH activity results we have seen, to the increase in
intracerebral concentrations of anandamide which allows to consider several
possibilities:
– Fixing of anandamide receptors on other
Anandamide is a mediator whose ubiquity well beyond the system
endocannabinoid. It is capable, especially as its concentrations
increase, interact with several types of receivers (at least TRPV1,
PPAR and NMDA) and the path of MAP kinase with possible consequences
in terms of apoptosis and neuroprotection.
– Toxicity of degradation products of anandamide
In case of complete and prolonged inhibition of FAAH, anandamide can be
gradient through cyclooxygenase which gives rise to various
compounds (prostanoids and leukotrienes) some of which have known effects on
cerebral vasomotor may be compatible with some damage
observed in volunteers cohort MAD 5.
The plausibility of these assumptions is however strongly questioned by a
series of counterarguments:
– The administration, including high dose of anandamide or its analogues nonmétabolisables
(Eg methanandamide) to animals is not known for
cause neurological toxicity, at least of the type examined by the
this report.
– This toxicity was not observed (or humans or animal) with
other inhibitors, let alone more specific, the FAAH, including those with
described as irreversible, even during this administration in high doses
and repeated.
– Inhibition, complete and lasting, of FAAH and therefore a priori, the plateau
intracerebral concentrations of anandamide, seem obtained in humans
from doses BIA 10-2474 of about 5 mg when no toxicity
27
Neurological was observed in the volunteers receiving in administrations
repeated (10 days) up to four times that dose (20 mg).
13. Conclusions of the CSST
The accident occurred at several volunteers in the trial conducted by Biotrial appears
clearly linked to the tested molecule.
It is very unlikely that this toxicity is related to the stimulation system
endocannabinoid via inhibition of FAAH (pharmacological target BIA 10-2474).
Toxic mechanism linked to increased intracerebral concentrations
anandamide also seems a priori excluded.
In the current state of knowledge, responsibility for one of the identified metabolites
10-2474 of the BIA does not seem to favor either. Note however that the shift
occurrence of toxic effects observed in the cohort MAD No. 5 could be
compatible with the production of a metabolite that, due to a half-life of
longer than that of the parent compound, might accumulate in the compartment
the tissue As administrations until its concentration therein
reaches a triggering threshold.
Notwithstanding, the assumption that the most likely day is that of a clean toxicity
the molecule via its binding to other brain cell structures facilitated by:
– Its low specificity for the target enzyme,
– The use of repeated doses well above those causing (at least in
humans) complete inhibition of FAAH and extended, and
– Its likely gradual accumulation in the brain, probably related to
pharmacokinetic characteristics of BIA 10-2474. This would, in
explain why the accident Rennes has occurred until the fifth
day administration of a dose of 50 mg and not in volunteers who received
a dose twice as strong in single administration.
It is at this stage difficult to favor a mechanism of toxicity of the two
most likely pretend: inhibition of other serine hydrolases or deleterious
the group “runner” imidazole-pyridine.
The fact that this toxicity was not observed in animals despite administration
very high doses remains to this day, unexplained. however it should be noted that, vis-à-vis
inhibition of FAAH, the BIA 10-2474 appears about 10 times more active in men
and in animals.
The sudden appearance of toxic symptoms could be related to the
character can be judged a posteriori “unwieldy” of the molecule BIA 10-
2474 characterized by relatively low efficiency (inhibitory concentrations of the order
micromolar), low specificity and concentration-response curve
particularly steep. Under these conditions, little understandable accelerating
escalating dose cohorts between MAD 20 and 50 mg likely played a
important role in triggering the crash. In this context, such
Increasing doses was the more risky that, due to the chronology
concatenation of cohorts and time required for analysis, latest
Pharmacokinetic data were those of subjects of the 10 mg cohort.
28
Such a sequence virtually prohibited to adjust the dose for the view
an emerging non-proportionality. This would be even more problematic for
originally scheduled last dose of either 100 mg (dose for which the
Pharmacokinetics is probably clearly non-proportional) since
the potential security adjustment would be based on data from the 20 mg cohort.
It was not within the powers of the CSST (unlike the two inspections
course) to decide on the merits of the authorization of the trial after the ANSM
opinion of the Committee to Protect People of Brest. Scientifically, the CSST
however, that:
– The BIA 10-2474 could not, a priori, be regarded as a product
risk according to the criteria listed in the recommendations in force,
especially the Guideline on strategies to Identify and Mitigate Risks in first in human
clinical trials with investigational medicinal products (Committee for Medicinal
Products for Human use, CHMP, EMA, 2007).
– The information transmitted, including the Investigator Brochure did not contain
no information, including toxicological, such as to fear
a particular risk during a first administration to humans. Rapellons
However, this brochure includes a fair number of errors,
inaccuracies, reversing numbers or erroneous translation of documents
sources that make comprehension difficult on several points. This is all
that surprising because of the regulatory importance of this document.
14. Recommendations of the CSST to conduct first test
drug delivery man
The severity of the accident at Rennes justify regulation and good
international practices regarding the first in-human studies
evolve on several points. While the BIA 10-2474 folder and protocol
the trial conducted by Biotrial comply with the provisions and recommendations in force
it’s in the rule that in mind. Meet regulatory requirements should
nor forget the basics of pharmacology and clinical or purpose
the therapeutic drug development. This pre-eminence of the rule on the
Common sense and scientific logic indicates a potentially dangerous development and
calls for collective awareness. The accident at the Rennes illustrates
dramatically.