Edaravone cocrystals: Synthesis, screening, and preliminary characterization

Article abstract of DOI:10.1007/s00706-013-1029-0

Cocrystals constructed of edaravone (a neuroprotective agent) and phenolic acids (phytochemicals) are investigated. Edaravone and cocrystal formers were coground, slurried, and cocrystallized from solvent at 1:1, 2:1, and 1:2 molar ratios. The physicochemical characteristics of the products were studied using elemental analysis, optical microscopy, Fourier transform infrared spectroscopy, differential scanning calorimetry, and X-ray powder diffraction. Edaravone cocrystals were prepared at various molar ratios. For three cocrystals, i.e., edaravone:gallic acid (2:1), edaravone:protocatechuic acid (1:1), and edaravone:trimesic acid (1:1), two polymorphs have been identified. Representative samples of the cocrystals were exposed to accelerated oxidative and thermal stress to investigate their stability. From the stability screening, protocatechuic acid and gallic acid cocrystals were identified as development candidates because they provide stable cocrystal forms.

Full text of DOI:10.1007/s00706-013-1029-0

Monatsh Chem (2013) 144:1335–1349
DOI 10.1007/s00706-013-1029-0
ORIGINAL PAPER
Edaravone cocrystals: synthesis, screening, and preliminary
characterization
•
Miroslav Veverka Tibor Dubaj Jan Gallovic
•
•
´
ˇ
ˇ
Emil Svajdlenka Beata Meluchova
ˇ
•
•
´
Vladimır Jorık Peter Simon
´
ˇ
•
´
´
Received: 22 October 2012 / Accepted: 31 May 2013 / Published online: 12 July 2013
Ó Springer-Verlag Wien 2013
Abstract Cocrystals constructed of edaravone (a neuro-
protective agent) and phenolic acids (phytochemicals) are
investigated. Edaravone and cocrystal formers were co-
ground, slurried, and cocrystallized from solvent at 1:1,
2:1, and 1:2 molar ratios. The physicochemical character-
istics of the products were studied using elemental analysis,
optical microscopy, Fourier transform infrared spectros-
copy, differential scanning calorimetry, and X-ray powder
diffraction. Edaravone cocrystals were prepared at various
molar ratios. For three cocrystals, i.e., edaravone:gallic
acid (2:1), edaravone:protocatechuic acid (1:1), and eda-
ravone:trimesic acid (1:1), two polymorphs have been
identified. Representative samples of the cocrystals were
exposed to accelerated oxidative and thermal stress to
investigate their stability. From the stability screening,
protocatechuic acid and gallic acid cocrystals were iden-
tified as development candidates because they provide
stable cocrystal forms.
Keywords Nutraceutical ꢀ Pharmaceutical cocrystal ꢀ
Polymorph ꢀ Cogrinding ꢀ Crystallization ꢀ
Characterization ꢀ Combination drug therapy ꢀ Stability
Introduction
Edaravone (ED), 3-methyl-1-phenyl-2-pyrazolin-5-one, a
strong free radical scavenger, was developed by Mitsubi-
shi-Tokyo Pharmaceuticals Inc. (Tokyo, Japan). It is used
in the treatment of patients with acute brain infarction. ED
is an electron donor and when it interacts with free radicals,
such as peroxyl radical, the reaction yields corresponding
3-methyl-1-phenylpyrazolin-2,5-dione and 2-oxo-3-phe-
nylhydrazonobutanoic acid. ED blocks both the water-
soluble and lipid-soluble peroxyl radical-induced peroxi-
dation systems [1–5]. At present, ED is commercially
available as the injection solution in a glass ampoule. From
in vitro experiments, Watanabe et al. reported that this
formulation is highly reactive toward active oxygen [6–8].
As a consequence of the high affinity for oxygen, an
injection formulation of ED in aqueous ethanolic solution
should be stabilized with thio-compounds such as pyro-
sulfate or cysteine [9].
ˇ
ˇ
ˇ
´
M. Veverka ꢀ J. Gallovic ꢀ E. Svajdlenka ꢀ B. Meluchova
Research and Development Department,
Eurofins Bel/Novamann Ltd.,
It is more or less accepted that antioxidant substances
might be valuable in mitigating neurodegeneration result-
ing from a transient ischemic event. Oxidative damage is
strongly implicated in the pathogenesis of neurodegenera-
tive diseases. Hence, free radical scavengers or chain-
terminating antioxidants could prevent the onset or slow
down the progression of some, if not most, neurodegen-
erative conditions. Examples of such antioxidant agents
include melatonin, lipoic acid, nonsteroidal anti-inflam-
matory agents, cysteine, glutathione monoethyl ester,
phenolic acids, vitamin E, resveratrol, and flavonoids [10].
820 04 Bratislava, Slovakia
ˇ
T. Dubaj ꢀ P. Simon (&)
Faculty of Chemical and Food Technology,
Institute of Physical Chemistry and Chemical Physics,
´
Slovak University of Technology, Radlinskeho 9,
812 37 Bratislava, Slovakia
e-mail: peter.simon@stuba.sk
´
V. Jorık
Faculty of Chemical and Food Technology,
Institute of Inorganic Chemistry, Slovak University
´
of Technology, Radlinskeho 9, 812 37 Bratislava, Slovakia
123

1336
M. Veverka et al.
Despite the high chance of finding synergy effects, the
combination therapy of ED and other drugs has been
studied rarely [11–15].
Results and discussion
In order to avoid the solubility limitation and formation of
solvates, SDG was the first method of cocrystal preparation.
At first, the molar ratio 1:1 was tested, but occasionally the
ratios 1:2 or 2:1 were applied for further evaluation and
verification of the ratio. Step 1 of screening included simple
analytical techniques, i.e., hot-stage microscopy and FTIR, to
provide a preliminary indication of potential cocrystal for-
mation. In this way, 14 compounds of three different families,
i.e., keto acids, sulfobenzoic acids, and phenolic acids, were
screened to cocrystallize with edaravone. The results are
shown in Table 1, where positive results (?) from a particular
stage of screening indicate the potential interactions. In order
to provide a more profound insight into the structural factors
that govern cocrystal formation with phenolic acids, sys-
tematic slurry preparation of cocrystals has been carried out;
the results are summarized in Table 2.
Pharmaceutical cocrystals represent the latest approach
to obtain new solid forms of an active pharmaceutical
ingredient (API) [16]. Examples of pharmaceutical co-
crystals have been demonstrated in the literature to provide
improved properties such as physical and chemical stability
and thus they can be used as alternative forms of drugs
[17, 18]. This paper describes the synthesis and charac-
terization of cocrystals of ED with phenolic acid
nutraceuticals as the cocrystal formers (CCFs). Phenolic
acids have been reported to possess high antioxidant
activity in various systems and some anticarcinogenic
activity [19–22]. Pharmaceutical cocrystals of phenolic
acids except salicylic acid are relatively marginally
explored in terms of solid-state chemistry despite the fact
that they contain phenolic groups and the carboxylic acid
group serving as a hydrogen bond donor/acceptor [23–28].
The ED:carboxylic acids cocrystals may represent an
alternative with the potential for development of combi-
nation drug therapy formulations.
The results of SDG experiments demonstrate the efficiency
of the preparation method in 11 out of 14 screening experi-
ments. However, SDG provides the cocrystal with
composition corresponding to the stoichiometric ratio of
starting materials (1:1) in six out of 11 experiments. This
number is difficult to compare with literature data as the
number of tested compounds is only occasionally mentioned
in the literature. It is noteworthy that in the relatively small
number of successful experiments, three new corresponding
polymorphs were identified. The solvent crystallization of a
1:1 ingoing ratio of material, sample no. 5, results in the same
cocrystallization product with excess unreacted gallic acid
(GA); hence, ED:GA (2:1) ratio was expected. The formation
of phase-pure ED:GA (2:1) cocrystals was achieved bythe SC
method in acetonitrile/acetone (1:1) mixture or by SDG when
an adequate stoichiometric ratio was applied. In addition, the
series of crystallizations from a number of solvents also
yielded different polymorphs in the material obtained; FTIR,
DSC, and XRPD did not correspond to the aforementioned
cocrystal yielded from the SC or SDG method. It was found
thatthecrystallizationfromethanol/acetonitrile(1:2)afforded
a phase-puresamplewith microanalysis datacorresponding to
the2:1cocrystal, buthavingadifferentFTIRandDSC records
and a new XRPD pattern. Thus, the XRPD pattern corre-
sponds to a new cocrystal ED:GA (2:1) termed here
polymorphII. Cocrystal ED:GA(2:1)polymorphIIcould also
be prepared by the slurry crystallization after 1 week of stir-
ring of ingoing components at ambient conditions in
cyclohexane or aqueous ethanol. It should be noted that the
polymorphism ofpharmaceutical cocrystalshasbeenreported
relatively rarely [37–44]. Zaworotko investigated the general
occurrence of polymorphism in existing cocrystals and found
11 examples [37]. When stoichiometric ratio variations were
tested in the SC experiments (samples no. 7, 9) these indeed
confirmed the stoichiometric ratios from the SDG results.
The aim of this study was the screening and identifica-
tion of suitable candidates and assessment of the
developability of cocrystal forms of ED in the early
development stages. The screening method involved three
steps. In the first step, the formation of a cocrystal was
verified by applying the simplest and most rapid tech-
niques, i.e., m.p. determination and Fourier transform
infrared spectroscopy (FTIR). If the presence of a cocrystal
was indicated, the material prepared passed the first step. In
the second step, differential scanning calorimetry (DSC)
and X-ray powder diffraction (XRPD) measurements
were carried out in order to prove the formation of co-
crystals. Finally, in the third step, preliminary solid-state
stability tests were performed [29–32]. In light of the
previously mentioned oxidative instability of the ED
injection formulation, oxidative stability of the represen-
tative cocrystals against hydrogen peroxide vapor at
30 °C with relative humidity (RH) control was further
evaluated [33]. This resulted in a rapid selection of
developable cocrystals.
From elemental analysis, FTIR, DSC, and XRPD, novel
cocrystals were identified and characterized. Formation of
cocrystals with different stoichiometry depends on the
experimental procedure [34]; thus, in this study solvent
drop grinding (SDG) was compared with solvent cocrys-
tallization (SC) experiments. The systematic study of series
of phenolic acids was conducted by the slurring method
(SL).
This is a follow-up paper to our previous works [35, 36]
dealing with the preparation and characterization of
imatinib mesylate cocrystals.
123

Edaravone cocrystals: synthesis, screening
1337
Table 1 List of CCFs subjected to screening and m.p., IR, DSC, and XRPD responses after SDG or crystallization with edaravone
No.
Cocrystal former
(molar ratio, ED:CCF)
Preparation
method
Response
m.p.
IR
DSC
XRPD
Stability
1
a-Resorcylic acid (2:1)
SDG, SC, SL
SDG
?
?
?
?
?
?
?
?
?
?
?
?
?
?
?
?
?
?
?
?
?
?
?
?
?
?
?
?
?
?
am
?
am
?
?
?
?
?
?
?
?
?
?
nc
nc
?
2
3a
Citric acid (1:1)
am
?
Camphanic acid (1:1)
SDG
4
Ketoglutaric acid (2:1)
SDG, SC
SDG, SC
SC, SL
SDG
am
?
nc
?
5a
6a
7
Gallic acid (2:1) polymorph I
Gallic acid (2:1) polymorph II
Trimesic acid (1:1) polymorph I
Trimesic acid (1:1) polymorph II
Protocatechuic acid (1:1) polymorph I
Protocatechuic acid (1:1) polymorph II
4-Sulfobenzoic acid (1:1)
?
?
?
nc
?
8a
SC, SL
SDG
?
9
?
nc
?
10a
11
12a
13
14
SDG, SL
SDG
?
im
?
nc
?
4-Sulfobenzoic acid (2:1)
SDG
4-Sulfobenzoic acid (1:2)
SDG
im
?
nc
nc
Pyrogallol-4-carboxylic acid (1:1)
SL
am amorphous phase, nc test not conducted, im impure sample
a
Data from these experiments indicate positive responses for all methods and are classified as leads
However, inthe courseofscale-upstudies ofthe primary SDG
methodusingtheSCmethodfrom1:1ethanol/acetonitrile, the
ED:trimesic acid (TMA) (1:1) cocrystal polymorph II was
identified.
fusion or by the infrared rule [45]. As stated by the heat of
fusion rule, a higher melting form has a higher heat of
fusion in a monotropic system and a lower heat of fusion in
an enantiotropic system. In the case of ED:PC cocrystal,
the polymorph I with higher melting temperature has a
higher heat of fusion; it can be concluded that the poly-
morph I is a monotrope. Likewise, in the case of ED:TMA
form II, higher melting forms have higher heats of fusion.
The relative stability of the two polymorphs of the ED:PC
(1:1) cocrystal was studied by placing a mixture of crystals
of form I and form II in open glass vials at 40 °C with
75 % RH; no transformation was observed after 2 months
by DSC and the 1:1 cocrystal mixture showed two
unchanged endothermic events at onset temperatures of
116.4 °C (128 J/g) and 114.4 °C (124 J/g) for the poly-
morphs I and II, respectively. This result gives evidence
that in this range of temperature and RH both polymorphs I
and II are stable.
Generally, SDG experiments were conducted in 5-cm3
reaction vessels using balls with a diameter of 5 mm. How-
ever, in the course of scale-up studies of the SDG method
affording the form I of edaravone:protocatechuic acid (PC)
(1:1) cocrystal, by applying 20-cm3 reaction vessels and balls
with a diameter of 9 mm another polymorphic form II was
prepared. Similar cases are reported, e.g., the effect of the
quantity of material on the polymorphic outcome and gener-
ation of thermodynamically unstable cocrystals during
substantially energetic grinding processes [42, 44].
The identified polymorphs were characterized using
FTIR, XRPD, and DSC. Each form exhibits a unique
XRPD pattern that differs from the patterns of starting
materials (Figs. 1, 2). The melting points of the poly-
morphs were determined by hot-stage microscopy and
verified from the onset temperatures of DSC endothermic
peaks (Fig. 3; Table 3). All the polymorphs show unique
melting points that differ from those of the starting mate-
rials and between them: ED:TMA (1:1) I (200.1 °C),
ED:TMA (1:1) II (208.4 °C), ED:PC (1:1) I (116.4 °C),
ED:PC (1:1) II (114.4 °C), ED:GA (2:1) I (154.4 °C),
ED:GA (2:1) II (146.4 °C). The fact that a higher number
of polymorphs were prepared in small series of cocrystals
suggests the importance of polymorph screening of co-
crystals and raw materials and their thorough
characterization. Polymorphs can usually be discriminated
using DSC data from the onset of melting and heats of
The thermodynamic relationships between polymorphs I
and II of ED:GA (2:1) respectively could be enantiotropi-
cally related. Relative polymorph stabilities for ED:GA
cocrystals were deduced from SL cocrystallization in two
solvents; where after 1 week form II was confirmed as
remaining solid, meanwhile the themodynamically unsta-
ble form I was obtained by energetic grinding [42]. The
FTIR spectra of forms I and II of ED:GA (2:1) cocrystals
are presented in Fig. 4. Absorption peaks in the
&1,670 cm-1 region correspond to carbonyl group vibra-
tions of gallic acid. Other differences between the IR
spectra of the polymorphs are present in the regions
740–760, 1,150–1,180, 1,400–1,420, 1,550–1,560, and
123

1338
M. Veverka et al.
Table 2 Overview of SDG, SC, and SL cocrystal formation experiments of edaravone with phenolic acids
CCF
Method
SDG (molar ratio)
SC (molar ratio)
SL (molar ratio)
2-Hydroxybenzoic acid
- (1:1, 2:1, 1:2) DMF
- (1:1, 2:1, 1:2) W
- (1:1, 2:1, 1:2) AN
- (1:1, 2:1, 1:2) DMF
- (1:1, 2:1, 1:2) W
- (1:1, 2:1, 1:2) AN
- (1:1, 2:1, 1:2) DMF
? (2:1) W
- (1:1, 2:1, 1:2) ET
- (1:1, 2:1, 1:2) AN
- (1:1, 2:1, 1:2) BU
- (1:1, 2:1, 1:2) AN
- (1:1, 2:1) ET, AN
- (1:1) AN, AE, W
- (1:2) AN, W, AE
- (2:1) AN, W, AE
- (1:2) CH, W, AE
- (2:1) W, AE
3-Hydroxybenzoic acid (3HB)
4-Hydroxybenzoic acid (4HB)
2,3-Dihydroxybenzoic acid
2,5-Dihydroxybenzoic acid
3,4-Dihydroxybenzoic acid (PC)
3,5-Dihydroxybenzoic acid (RA)
2,4,6-Trihydroxybenzoic acid
2,3,4-Trihydroxybenzoic acid (PG)
3,4,5-Trihydroxybenzoic acid (GA)
m-Coumaric acid
- (1:1) AE, W, CH
- (2:1) W, AE
- (1:1, 2:1, 1:2) AN
- (1:1, 2:1, 1:2) ET
- (1:1) AE, W, CH
- (1:1, 2:1, 1:2) AN
- (1:1, 2:1, 1:2) DMF
- (1:1, 2:1, 1:2) W
- (1:1, 2:1, 1:2) AN
- (1:1, 2:1, 1:2) DMF
- (1:1, 2:1, 1:2) W
- (1:1, 2:1, 1:2) AN
? (1:1), - (2:1, 1:2) DMF
? (1:1), - (2:1, 1:2) W
? (1:1), - (2:1, 1:2) AN
? (2:1), - (1:1, 1:2) DMF
? (2:1), - (1:1, 1:2) W
? (2:1), - (1:1, 1:2) AN
- (1:1, 2:1, 1:2) DMF
- (1:1, 2:1, 1:2) W
- (1:1, 2:1, 1:2) AN
- (1:1, 2:1, 1:2) DMF
- (1:1, 2:1, 1:2) W
- (1:1, 2:1, 1:2) AN
? (2:1), - (1:1, 1:2) DMF
? (2:1), - (1:1, 1:2) W
? (2:1), - (1:1, 1:2) AN
- (2:1, 1:1) W
- (2:1) W, AE
- (1:1) AE, W, CH
- (1:1, 2:1, 1:2) ET
- (1:1, 2:1, 1:2) AN
- (2:1) CH
? (1:1) AE
? (1:1) AC-AN
? (1:1) ET-AN
? (1:1) AE
- (1:1) CH
- (1:1) CH
? (2:1) W
? (2:1) AE
- (1:1, 1:2) AN,
- (2:1) AN
- (1:1, 2:1, 1:2) AN
? (1:1) AE
- (1:1) W
- (2:1) CH, W, AE
? (2:1) CH
? (2:1) AC-AN
? (2:1) ET-AN
? (2:1) AE
- (1:1) W, AE
- (2:1, 1:1) AN
- (1:1, 2:1, 1:2) DMF
- (2:1, 1:1, 1:2) AN
- (2:1, 1:1) W
Caffeic acid
- (1:1) W, AE
- (2:1) W, AE
- (1:1, 2:1, 1:2) DMF
1 sign denotes the cocrystal formation (stoichiometry highlighted in bold); - sign denotes a mixture of the starting materials without a cocrystal
formation
AC acetone, AE (1:1, v/v) aqueous ethanol solution, AN acetonitrile, BA butyl acetate, BU butanol, CH cyclohexane, Cl chloroform, DMF dimethyl-
formamide, ET ethanol, W water
3,200–3,360 cm-1. The absorption peak relation between
the two polymorphs located in the &1,670 cm-1 region
appears to be in accordance with Burger and Ramberger’s
infrared rule, which states that the phase I that absorbs at
higher frequencies is less stable at 0 K [45].
complexity, TMA with its threefold symmetrically dis-
posed carboxyl groups and SBA with two nonequivalent
acidic groups. This stimulated us to construct supramo-
lecular architectures of ED and SBA, which would be
expected to exhibit diversified hydrogen-bonding modes
and interesting architectures [46]. Only the SDG method
afforded a phase-pure cocrystal ED:SBA in the stoichi-
ometric ratio 2:1. DSC records also indicated the presence
We further investigated cocrystal formation of edarav-
one with SBA and TMA. From a crystal engineering view,
SBA and TMA are CCFs with higher molecular
123

Edaravone cocrystals: synthesis, screening
1339
Fig. 2 XRPD patterns for (from top to bottom) ED:PC (1:1)
polymorph I, ED:PC (1:1) polymorph II, ED:RA (2:1), ED:GA (2:1)
polymorph I, ED:GA (2:1) polymorph II, ED:PG (1:1) cocrystals
Fig. 1 XRPD patterns for (from top to bottom) ED:TMA (1:1)
polymorph II, ED:TMA (1:1) polymorph I, ED:CA (1:1), ED:4-
sulfobenzoic acid (SBA) (2:1), ED:SBA (1:1), ED:SBA (1:2), ED:KG
(2:1) cocrystals
amorphous by DSC and XRPD methods. This is observed
during grinding [51, 52] as is true for ED:citric acid (1:1)
cocrystal. Pure citric acid monohydrate (CIT) has bands at
1,721 and 1,685 cm-1 corresponding to the m(COOH) and
(C=O) stretch. During the formation of ED:CIT (1:1) co-
crystal these bands were shifted to 1,698 and 1,678 cm-1
as weak broad bands, which indicated the formation of
hydrogen bonding. However, the XRPD pattern revealed
that the material was obtained as an amorphous solid. On
the other hand, ED:ketoglutaric acid (2:1) cocrystals pre-
cipitated as a glassy (softening around 130 °C) material
from an ethanol/acetonitrile (9:1, v/v) system. To examine
the effect of solvents on the solid state of the aforemen-
tioned materials, we took the same measurements on the
samples from SC using various solvents including aceto-
nitrile, acetone, 2-methyl-2-butanol, methyl ethyl ketone,
dipropyl ether, butyl acetate, and chloroform [53]. In sol-
vent treatments mentioned above, if a solid material
precipitated, we found that we got the identical amorphous
material using different solvents. No peaks were observed
of new phase via the SDG method for 1:1 and 1:2 stoi-
chiometry; ED and SBA were consumed and a smaller
amount of impurity was detected alongside with traces of
solvent (Fig. 3). But, the different XRPD patterns of three
resulting structures (Table 1, samples no. 8, 9, 10) were
observed as presented in Fig. 1. However, these samples
were excluded from the next screening. The variations of
stoichiometry were unexpected and we hypothesized that
the sulfonic group can be complexed, too. Cocrystal sul-
fonates and their single-crystal structure determinations
have been reported rarely and mimicking of a carboxylic
group was confirmed [47–50]. Unfortunately, in our case a
single crystal of this structure for the single-crystal XRD
measurement was not obtained. As stated above the co-
crystal of ED:TMA 1:1 using SDG afforded polymorph I
and the SC method afforded polymorph II. These examples
illustrate another case in which the different methods
produce different cocrystal polymorphs.
The first stage of screening identified two materials by
melting point and FTIR that were then identified as
123

1340
M. Veverka et al.
Fig. 3 DSC curves of (from
top to bottom) ED:PG (1:1),
ED:PC (1:1) polymorph II,
ED:PC (1:1) polymorph I,
ED:RA (2:1), ED:GA (2:1)
polymorph I, ED:GA (2:1)
polymorph II, ED:TMA (1:1)
polymorph I, ED:TMA (1:1)
polymorph II, ED:CA (1:1),
ED:SBA (2:1), ED:SBA (1:1),
ED:SBA (1:2), ED:KG (2:1),
ED:CIT (1:1) cocrystals
in XRPD patterns except for a halo-like pattern at 2h
between 10° and 25° confirming its amorphous state
(Fig. 1). As we are not interested in amorphous forms,
ED:CIT (1:1) and ED:KG (2:1) materials were rejected as
development candidates and excluded from the next stage
of screening.
Unexpectedly, in the case of salicylic acid as CCF the
FTIR and DSC clearly show that no cocrystal is formed.
Following from these negative observations, the synthesis
of cocrystals was attempted by SDG, SC, or the SL method
from other structurally related phenolic acids. As non-
bonding interactions in cocrystals take place, a steric
123

Edaravone cocrystals: synthesis, screening
1341
Table 3 Melting points and
DSC data of the edaravone
cocrystals
Cocrystal former (molar ratio, ED:CCF)
m.p./°C
Tonset/°C
D
fusH/J g-1
CCF, Tonset/°C
Edaravone
127–129
126–127
*103
128.0
127.8
am
151.6
126.5
–
–
a-Resorcylic acid (2:1)
Citric acid (1:1)
235.1
151.6
204.7
114.2
259.1
259.1
361.5dec
361.5dec
202.2
202.2
&99
Camphanic acid (1:1)
Ketoglutaric acid (2:1)
Gallic acid (2:1) I
123–124
&130
127.9
am
129.4
–
154–155
144–146
[235dec
[215dec
114–115
111–113
163–170
155–158
200–202
158
154.4
146.4
200.1dec
208.4dec
116.4
114.4
im
124.8
135.1
78.0dec
93.3dec
128.3
123.7
–
Gallic acid (2:1) II
Trimesic acid (1:1) I
Trimesic acid (1:1) II
Protocatechuic acid (1:1) I
Protocatechuic acid (1:1) II
4-Sulfobenzoic acid (1:1)
4-Sulfobenzoic acid (2:1)
4-Sulfobenzoic acid (1:2)
Pyrogallol-4-carboxylic acid (1:1)
158.7
im
86.2
–
&99
&99
dec decomposition, im impure
sample, am amorphous phase
151.9
112.2
214.1
arrangement of hydroxyl moieties seems to be important.
From the results mentioned above we deduced that the
presence of the 3-hydroxyl moiety in benzoic acids leads to
interaction with ED; therefore 3-hydroxybenzoic acid
(3HB), 3,5-dihydroxybenzoic acid (a-resorcylic acid, RA),
and 2,3,4-trihydroxybenzoic acid (pyrogallol-4-carboxylic
acid, PG) were tested as CCFs. Indeed, the corresponding
cocrystal (sample 1, Table 1) was prepared by SDG, SL, or
the SC method at the stoichiometry 2:1 ED:RA. All the
cocrystals obtained by SDG are reproduced by SL; how-
ever, ED:PG (1:1) cocrystals were obtained solely by SL,
and thus slurry cocrystallization seems more effective. For
3-hydroxybenzoic acid the FTIR method indicated a new
phase and DSC showed a clear single melting peak
(92 °C). Additional mixtures were prepared to verify the
formation of the eutectic mixture. However, all of the
materials with the 2:1, 1:2, or 1:1 ratio show the same
eutectic point thereby indicating the absence of cocrystal
formation (Fig. 5). Also, the XRPD patterns of these
materials showed negligible differences against the pat-
terns of starting compounds; so the formation of cocrystal
did not take place.
the cocrystals prepared from hydroxybenzoic acids by the
slurry method. This behavior of hydroxyl benzoic acids has
already been reported. Broad screening of acetazolamide
with many structural isomers of hydroxybenzoic acids
yielded a cocrystal only for 4-hydroxy- and 2,3-dihy-
droxybenzoic acid [28].
The melting points were compared with those of the ED
and CCFs. The melting points of obtained samples are
diffused; however, they fell mostly between the values of
edaravone (127–129 °C) and the corresponding CCF. The
distinct melting points of cocrystals clearly indicate the
formation of a new crystalline phase. These cocrystals were
shown to alter the melting point of edaravone, suggesting
that physical properties of great pharmaceutical relevance,
e.g., solubility and thermal stability, have been altered. In
this series, however, no direct correlation between melting
points of either individual component and that of the co-
crystal could be found. The FTIR spectra of representative
cocrystals are depicted in Figs. 6 and 7. Use of FTIR was
possible through the inclusion of a set of physical mixtures
of edaravone and CCFs representing possible solid-state
outcomes from the cocrystal screening. The tendency of
CCFs to interact with edaravone is demonstrated in
Table 1. FTIR spectra of cocrystals compared with those of
the pure compounds or their physical mixture showed
numerous changes. It can be seen that for cocrystals, the
characteristic peaks either reduced their intensity, disap-
peared, shifted, or new peaks appeared (Fig. 4). Indeed,
most of the peaks of CCFs shifted in this way, e.g., the
C=O stretching vibration of the carboxylic group at
1,728 cm-1 for camphanic acid occurred at 1,674 cm-1 in
ED:CA (1:1) cocrystal. An increase in the frequency of the
keto group of camphanic acid was observed from 1,745 to
1,782 cm-1 in ED:CA (1:1) cocrystal (Table 4). An
Phenolic acids occasionally showed surprising results,
as two stoichiometrically different cocrystals were isolated
[18, 53, 54]. The cocrystals obtained from a pair of iso-
meric dihydroxybenzoic acids (3,5- vs. 3,4-dihydroxy) and
trihydroxybenzoic acids (3,4,5- vs. 2,3,4-trihydroxy) exhi-
bit different stoichiometric ratios of 2:1 vs. 1:1. It was
found that the SL method afforded a cocrystal with
appropriate stoichiometry (Table 2) regardless of the used
ongoing ratio (1:1, 1:2, 2:1) of starting materials.
According to the aforementioned idea we tested caffeic
acid and 3-hydroxycinnamic acids using the slurry co-
crystallization; no cocrystal formed. Table 2 summarizes
123

1342
M. Veverka et al.
Fig. 4 Example of FTIR
spectra used to identify potential
cocrystals (from top to bottom):
ED:GA (2:1) polymorph I,
prepared by SC; ED:GA (2:1)
polymorph I, prepared by SDG;
ED:GA 2:1 polymorph II,
prepared by SC; GA and ED
physical mixture
increase in the frequency of the keto group of camphanic
acid was observed from 1,745 to 1,782 cm-1. The IR
spectrum of pure ketoglutaric acid (KG) and CIT starting
material shows characteristic peaks attributed to the
m(COOH) and (C=O) stretch, respectively. During cocrys-
tallization these bands were shifted which indicates that the
carboxyl group is participating in hydrogen bonding.
However, the situation for these ED:polycarboxylic
acids (1:1) systems was somewhat unclear. The IR spec-
trum shows an indistinguishable broad multiplet at
&1,705 cm-1, and the new absorption bands attributable
to m(COOH) group were strongly overlapped with the
mostly unperturbed COOH absorption bands. For all other
tested acids, however, the carbonyl peak occurred at
1,715–1,682 cm-1 clearly in agreement with the structural
evidence for the cocrystal formation. Also, a new band at
1,596 cm-1 attributed to the C=N of the edaravone
skeleton appeared. This band appeared only in those cases
where edaravone was found to form the cocrystal.
Crystal morphology and particle size are important for
handling the drug substance and for reliable manufacture of
dosage forms with uniform drug content. The optical
micrographs of representative cocrystals are shown in
Fig. 8. The morphology of the cocrystals ranges from a
rhombic habit for 8a (mean value 450 lm, ED:GA (2:1),
polymorph II), to a hexagonal morphology with irregular
surface for 8b, 8d and tetragonal prism for 8c [mean value
270 lm, ED:TMA (1:1)]. Form ED:GA (2:1), polymorph
II is particularly easy to filter and dry to a free-flowing
powder.
DSC experiments were carried out to study the thermal
behavior of ED cocrystals in relation to the individual
components. The DSC records and m.p. for ED, individual
native CCFs, and cocrystals were compared. The DSC
123

Edaravone cocrystals: synthesis, screening
1343
Fig. 5 DSC records of (from
top to bottom) potential
cocrystal ED:3HB (1:1) by the
SL method, potential cocrystal
ED:3HB (1:1) by the SL method
doped with ED, and the physical
mixture ED:3HB (1:1)
Fig. 6 FTIR spectrum of (from
top to bottom) ED:CA (1:1),
ED:SBA (2:1), ED:SBA (1:1),
ED:RA (1:1), ED:SBA (1:2),
ED:KG (2:1) cocrystals
records are shown in Fig. 3; the melt endotherms obtained
from the DSC data are shown in Table 3, with the melting
points for the corresponding cocrystal components. The
DSC results revealed significant differences between the
values of the melting point onsets of pure ED and ED
cocrystals. This melting point mostly occurs between the
melting points of the starting materials (ED at 127–129 °C
and CCFs). Formation of the cocrystal then results in
increased thermal stability of ED, but a lower stability for
the CCFs molecules. A single endothermic peak for co-
crystals indicates the absence of any unbound or absorbed
solvent and demonstrates the stability of phase until the
123

1344
M. Veverka et al.
Fig. 7 FTIR spectrum of (from
top to bottom) ED:TMA (1:1)
polymorph II, ED:PC (1:1)
polymorph II, ED:PC (1:1)
polymorph I, ED:PG (1:1),
ED:CIT (1:1), ED:TMA (1:1)
polymorph I cocrystals
Table 4 Assignment of the major bands of the infrared spectra of cocrystals
No.
Cocrystal former (molar ratio, ED:CCF)
m(COOH)/cm-1
m(C=Nhet)/cm-1
m(OH) of COOH/cm-1
1
a-Resorcylic acid (2:1)
1,653
1,698
1,674
1,705
1,668
1,662
1,686
1,691
1,682
1,677
1,697
1,715
1,686
1,646
1,592
1,598
1,594
1,596
1,597
1,591
1,594
1,596
1,595
1,601
1,603
1,600
1,601
1,607
1,409
1,420
1,391
1,403
1,411
1,419
1,406
1,407
1,411
1,389
1,397
1,397
1,397
1,407
2
Citric acid (1:1)
3
Camphanic acid (1:1)
4
Ketoglutaric acid (2:1)
5
Gallic acid (2:1) polymorph I
Gallic acid (2:1) polymorph II
Trimesic acid (1:1) polymorph I
Trimesic acid (1:1) polymorph II
Protocatechuic acid polymorph I (1:1)
Protocatechuic acid polymorph II (1:1)
4-Sulfobenzoic acid (1:1)
6
7
8
9
10
11
12
13
14
4-Sulfobenzoic acid (2:1)
4-Sulfobenzoic acid (1:2)
Pyrogallol-4-carboxylic acid (1:1)
melting point. For ED:SBA (1:1, 1:2) cocrystals, however,
the DSC of the cocrystal indicated traces of CCF. When
CIT or KG was applied, the DSC traces clearly indicate an
amorphous phase. There are several suggestions found in
the literature that cogrinding induces an amorphization [51,
52]. The amorphous samples of ED:CIT and ED:KG co-
crystals were aged at 25 °C with 35 % RH, for 2 months,
to assess the effect of storage on their crystallizability.
Cocrystals ED:CIT and ED:KG solidified but the repeated
DSC measurements did not detect any crystalline phase.
A different XRPD pattern for each ED cocrystal con-
firms the formation of a new cocrystal phase; a unique
XRPD pattern enables unambiguous distinction between
the cocrystals from ED and CCFs. As stated above, only
the samples with indubitable XRPD patterns from the
screening experiments were classified as positive. The
XRPD patterns of cocrystals are shown in Figs. 1 and 2.
The low-angle region at 2h = 5°–30° was used to reliably
distinguish between various crystal forms of new cocrys-
tals. The XRPD patterns revealed that the formation of
cocrystals via SDG occurred with stoichiometric ratios
corresponding to the starting materials; no starting material
was detected except SBA. Total consumption of starting
ED and unused CCFs was revealed by XRPD in some
cases. Finally, using appropriate stoichiometric variation
(e.g., ED:GA 2:1), the desired cocrystals were obtained and
123

Edaravone cocrystals: synthesis, screening
1345
Fig. 8 Optical micrographs of
a ED:GA (2:1) polymorph II,
b ED:GA (2:1) polymorph I,
c ED:TMA (1:1) polymorph II,
d ED:TMA (1:1) polymorph I
prepared by SC (left) or SDG
(right)
confirmed by XRPD. The new XRPD patterns were found
to correspond to polymorphs of cocrystals ED:GA,
ED:TMA, and ED:PC.
Stability tests as a function of time versus humidity and
temperature are useful in assessing the shelf life of co-
crystals. Thus, cocrystals were subjected to temperature
40 °C and 75 % RH for 2 months. The solids were ana-
lyzed by m.p., FTIR, DSC, and XRPD and compared with
unstressed samples to confirm a physical stability. Co-
crystals displayed physical stability; no recognizable
differences in m.p. and the FTIR, DSC, and XRPD patterns
were found. The type of polymorphic form does not have a
significant influence on the stability under the testing
conditions. As an example, the XRPD patterns for ED:PC
(1:1) cocrystals before and after exposure are shown in
Fig. 9. It can be seen that any significant change indicating
decomposition in the patterns did not occur. The intensity
of diffraction peaks in the stressed samples was somewhat
different compared with the starting material owing to a
diverse measurement condition (sample volume). Thermal
stability/relative humidity results (TS) are summarized in
Table 5.
Fig. 9 XRPD patterns for (from top to bottom) cocrystal ED:PC
(1:1) polymorph II and cocrystal ED:PC (1:1) polymorph II after
exposure
parallel using the condition 25 °C/65 % RH for 3 days.
Generally, the tested cocrystals maintain a practically
unchanged content of ED. This preliminary study indicates
that stabilization is probably a result of the beneficial effect
of CCF. These experimental results demonstrated that the
proposed antioxidant protection can be used to generate
stable products and is useful for the development of drug
products; the results are summarized in Table 5.
The next stage, i.e., the oxidative-stability screening,
also provides additional information regarding the stability.
Edaravone is thermally stable; however, peroxide vapor-
mediated oxidation was confirmed under the test conditions
with unidentified degradants as monitored by HPLC. E-
daravone, as a blank, showed significant decrease of the
content to 89.8 % during exposure to hydrogen peroxide
vapor. Thus, the cocrystal solid samples were exposed in
In the final stage of stability screening, sorption of water
was conducted to identify ‘‘developable’’ cocrystal candi-
dates. Water sorption for selected cocrystals was
determined at 25 °C and 55, 80, and 90 % RH. The
123

1346
M. Veverka et al.
1.2
1.0
0.8
0.6
0.4
0.2
0.0
Table 5 HPLC data and stability results for ED and representative
cocrystals
Sample
OS
WS
TS
ED
89.8 (area %)
?
?
?
?
?
?
nc
?
?
?
?
?
ED:CA (1:1)
?
?
?
?
?
ED:PC (1:1) polymorph II
ED:SBA (2:1)
ED:TMA (1:1) polymorph II
ED:GA (2:1) polymorph I
ED:GA (2:1) polymorph I
ED:TMA (1:1) polymorph II
OS oxidation stability, ? no decrease in content with respect to ED,
WS water sorption, ? change in mass\0.1 %, TS solid-state stability
with respect to temperature and relative humidity, ? test passed, nc
test not conducted
ED
Eq. (1)
0
20
40
60
80
100
120
Time/min
cocrystal was regarded as non-hygroscopic if it sorbed less
than 0.1 % water even at high humidity; the results for
90 % RH are summarized in Table 5. It was evident that
ED cocrystals sorbed a negligible amount of water
(ꢁ0.1 %), practically the same as a ground sample of ED.
These preliminary results indicate that the ED cocrystals
are non-hygroscopic. Generally, all tested materials had
satisfactory thermal and oxidation stability and did not
undergo hydration under increased humidity.
Fig. 10 Representative dissolution curves of ED:GA (2:1) poly-
morph I, ED:TMA (1:1) polymorph II cocrystals, and ED. Solid line
represents non-linear fit of experimental data to Eq. (1)
Table 6 Solubility rate constants for ED and representative
cocrystals
Sample
k/min-1
ED
0.092
0.138
0.150
0.159
0.118
0.147
0.047
The dissolution curves almost perfectly obeyed first-
order kinetics (Fig. 10). Purely formally (without consid-
ering mechanistic details such as the solution-mediated
transformation and solute–solute interaction), the dissolved
concentration c can be expressed as
ED:PC (1:1) polymorph I
ED:PC (1:1) polymorph II
ED:GA (2:1) polymorph I
ED:CA (1:1)
ED:SBA (2:1)
c ¼ c1½1 ꢂ expðꢂktÞꢃ
ð1Þ
ED:TMA (1:1) polymorph II
where c1 is the final concentration, k is the dissolution rate
constant, and t stands for time.
Since the final concentration is unknown, for the treat-
ment of dissolution data non-linear regression was
employed; the results are presented in Table 6.
3. Cocrystals with significantly lower solubility rate
compared to ED [ED:TMA (1:1) polymorph II]
It is accepted that a substance with a high melting point
has strong self-interactions, and therefore low solubility
[55]. This trend is approximately observed in the tested
series: solubility rates of cocrystals ED:PC (1:1) poly-
morph I (Tonset = 116.4 °C), ED:PC (1:1) polymorph II
The dissolution tests were conducted to compare the
dissolution rate of ED and cocrystals. Table 6 shows the
dissolution rate constants for representative samples. The
comparison of rate constants of ED cocrystals with that of
ED itself reveals that the solubility rates of cocrystals are on
average 1.6 times higher than that of edaravone [except
ED:TMA (1:1) polymorph II]. Nevertheless, this pre-
liminary study indicates that the dissolution rate of ED can be
modified by cocrystallization. From the values of dissolution
rate constants it can be seen that the prepared cocrystals can
be roughly divided into the following three groups:
(Tonset = 114.4 °C), ED:GA (2:1) polymorph I (Tonset
=
154.4 °C), and ED:SBA (2:1) (Tonset = 158.7 °C) are
about three times higher than that of ED:TMA (1:1)
polymorph II (Tonset = 208.4 °C). As expected, the dif-
ferences in the obtained solubility rates reflect the different
CCFs solubilities [55], i.e., the saturated solubility of GA at
20 °C is 12 mg/cm3, for PC 18.2 mg/cm3, and TMA has
poor solubility in water (owing to intermolecular hydrogen
bonding). It may be expected that solubility of potential
cocrystals of ED with more soluble acids may be higher
than that of the applied phenolic acids. From this result,
ED:PC and ED:GA cocrystals appear the most robust
candidates; this demonstrates that using antioxidants
1. Cocrystals with significantly higher solubility rate
compared to ED [ED:PC (1:1) polymorph I, ED:PC
polymorph II, ED:SBA (2:1), and ED:GA (2:1)
polymorph I]
2. Cocrystals with practically unchanged solubility rate
[ED:CA (1:1)]
123

Edaravone cocrystals: synthesis, screening
1347
(phenolic acids) as CCFs yields cocrystals of ED as a
crystalline, nonsolvated, highly stable solid.
frequency of 30 Hz. External temperature of the grinding
jar did not exceed 30 °C.
Preparation of cocrystals by solvent crystallization was
carried out in round-bottomed flasks. ED (100–200 mg) was
added to the flasks in an oil bath. An adequate amount of
solvent (1.5–3.5 cm3) was added to the flask at reflux to form
a solution. Then an appropriate amount of CCF was added.
The solutions formed were quickly filtered through a nylon
filter (0.22 lm) and then the filtrates were transferred to a
10-cm3 flask. In some cases cosolvents (2–6.0 cm3) were
carefully added to the stirred solutions at a gentle reflux. The
solutions were gently refluxed for an additional 10 min with
stirring. Then the oil bath was removed and the temperature
was gradually reduced to 20 °C over 0.5 h while stirring.
The cooling rates for all flasks were the same. The flasks
were allowed to stand at 5–10 °C for 48 h. The precipitated
materials were filtered via a flask-to-flask filtration assembly
using a sintered glass no. 3 funnel with vacuum. The crystals
were dried under vacuum at 45 °C and stored in a desiccator.
For the stability testing the materials were scaled up fol-
lowing the same reaction procedures.
Conclusion
We have shown that phenolic acids are able to cocrystallize
with edaravone and also form polymorphs. Sixteen novel
cocrystals have been described, with supporting analysis of
FTIR and XRPD. The DSC technique was used for char-
acterization of cocrystal polymorphs. Formation of
cocrystals with different stoichiometry was observed. Each
of the cocrystals was stable upon exposition to increased
relative humidity and/or temperature over 8 weeks. Fur-
thermore, dissolution studies of representative cocrystals
show an alteration in solubility of edaravone with the
formation of cocrystal indicating the potential for further
development in this field. To summarize, cocrystal eda-
ravone:gallic acid (2:1) polymorph II can be prepared on a
large scale with robust stoichiometric control and accept-
able yield, volume, and purity (chemical and solid state).
Preparation of cocrystals via slurring
Experimental
The following procedure was applied only for phenolic
acids summarized in Table 2. Stoichiometric amounts of
starting materials were slurried (1 cm3 of solvent per
100 mg of starting material) to stir for 1 week at ambient
temperature. The undissolved solid was isolated by vacuum
filtration via a flask-to-flask filtration assembly using a
sintered glass no. 3 funnel. The solids were dried under
vacuum at 60 °C and stored in a desiccator. Solids were
characterized using IR spectroscopy and m.p. Table 2
summarizes the cocrystals obtained in this manner.
Edaravone and CCFs were purchased from Sigma-Aldrich
Ltd. and were used as received. All the samples gave
microanalysis results for C 0.38 %, H 0.35 %, and
N
0.25 %. The water content was measured by ther-
mogravimetric analysis. All the samples showed negligible
weight loss between 25 and 110 °C. Analytical grades of
solvents, i.e., methanol, ethanol, acetonitrile, acetone,
2-isopropoxyethanol, methyl isobutylketone, 3-methyl-1-
butanol, and N,N-dimethylformamide, were purchased
from Sigma-Aldrich Ltd. Whatman nylon syringe membrane
filters (0.2 lm) were used for filtration of crystallization
solutions.
Determination of melting points
Melting points were determined on a Boetius Nagema-type
¨
hot-stage apparatus (Rapido Wagetechnik, Radebeul, Ger-
Preparation of cocrystals by solvent drop grinding
and solvent crystallization
many). The melting point can be determined with an
accuracy of 1.0 °C. The heating rate was 5 °C min-1. The
melting points of potential cocrystals were compared with
those of ED and CCFs.
Cocrystals were prepared with guest compounds selected
on the basis of their solubility, representative differences in
structure, and range of functional groups as summarized in
Table 1. The preparation included SDG and SC with a molar
ratio 1:1, 2:1, or 1:2. Alternatively, these experiments were
used as a rapid method of obtaining material for seeding of SC
for the scale-up experiment no. 5 (Table 1).
Fourier transform infrared spectroscopy
The spectra were collected on an FTIR spectrometer
Impact 410-IR (Nicolet, USA). The ATR technique (dia-
mond ATR crystal) was used with Omnic software version
5.2. A total of 50 scans were collected over the range of
4,000–400 cm-1. The IR spectra of potential cocrystals
were compared with those of physical mixtures of parent
CCF and edaravone. The CCFs and ED were separately
For SDG experiments, a Retsch 200MM mixer mill,
equipped with 10-cm3 stainless steel grinding jars with two
stainless steel grinding balls with a diameter of 5 or 9 mm,
was used. All SDG experiments were performed with one
drop of acetonitrile, dimethylformamide, or water at a
123

1348
M. Veverka et al.
Optical micrographs
gently ground before mixing to achieve the same particle
size as ground cocrystal particles.
Optical micrographs of ED, CCFs, and cocrystals were
taken using the Nikon Eclipse LV 100POL polarizing
microscope and analyzed using NIS Elements. Morphology
of the solids was observed using a magnification 940. The
crystal diameter was measured and the number mean size
for a cocrystal was calculated.
Thermal analysis
A Perkin-Elmer DSC-7 differential scanning calorimeter
(DSC) with Pyris software was employed. The temperature
scale was calibrated to the fusion temperatures of In, Sn,
and Pb; the enthalpic scale was calibrated to the enthalpy
of fusion of In. The heating rate was 10 °C min-1 and
nitrogen was used as a purge gas.
Dissolution test
The dissolution tests were conducted using a Distek Evo-
lution 6300 dissolution system and were carried out for 1 h
in 250 cm3 of purified water at 37 °C with a paddle speed
of 50 rpm. All fine samples (crystalline samples were
carefully ground in an agate mortar) were agglomerated
and the samples showed different wettability. In order to
improve the agglomeration and wettability, the samples for
the powder dissolution tests were prepared from a physical
mixture of samples and a-lactose (1/20, w/w). For the
dissolution test, about 5 mg of each of the fine samples
(ED, cocrystals sample) was compressed into a 1.13-cm2
disk by a hydraulic press at 90 kN for 1 min using a die
with a hole 1.2 cm in diameter (Spectra-Tech, Inc., CT,
USA). Dissolution experiments were run in duplicate. The
stability of the solid samples after disk compression was
confirmed by IR of compressed neat samples under iden-
tical conditions. Drug concentrations were determined by
high-performance liquid chromatography (HPLC, Agilent
Technologies 1200 system) using a DAD UV detector and
a C18 column (ZORBAX Eclipse 4.6 9 100 mm 3.5-lm
particle size). The dissolution samples (1 cm3) were col-
lected for each tablet at 3, 5, 10, 15, 20, 25, 30, 40, 50, 60,
90, and 120 min. Concentrations of ED were determined
by HPLC using an external calibration curve measured
before and after dissolution measurements. The wavelength
for ED and cocrystals was set at 240 nm.
X-ray powder diffraction
The patterns of powder samples were collected within the
2h range 3°–51° in steps of 0.02° on a Bragg–Brentano
diffractometer Philips PW1730/1050, using b-filtered
CoKa radiation, 40 kV/35 mA and with a scintillation
detector. In order to facilitate the data interpretation, XRPD
patterns of ED and CCFs were measured and compared
with that of the material prepared.
Stability in solid-state and oxidative stability
Samples of cocrystals and ED were placed in open glass
vials at 40 °C/75 % RH and tested periodically during
2 months by m.p. and FTIR methods. At the end of the
testing period, the m.p, FTIR, and XRPD were measured
for stressed samples.
For the preliminary water sorption studies, the samples
and ED were milled in an agate mortar and then dried for
6 h under vacuum at 45 °C. The samples (approximately
60–120 mg) were then exposed to 55, 80, and 90 % RH for
24 h each and the weight differences for appropriate RH
were calculated for dry mass. The temperature was main-
tained at 20 2 °C; humidity was kept constant by use of
saturated aqueous salt baths. The humidity and temperature
were monitored using a GMH 3330 (Greisinger GmbH,
Germany) with a TFS0100 probe. The cocrystal is con-
sidered non-hygroscopic if it sorbs less than 0.1 % water
even at high humidities (90 % RH).
Acknowledgments This investigation was supported by the Slovak
Research and Development Agency, grant no. APVV-121-2009 and
The Agency of the Ministry of Education, Science, Research and
Sport of the Slovak Republic for the Structural Funds of EU, OP RaD
of ERDF by realization of the project ‘‘Evaluation of natural sub-
stances and their selection for prevention and treatment of lifestyle
To evaluate the oxidative stability, the crystalline solids
were exposed to hydrogen peroxide vapor, with RH control
(25 °C/60 % RH), using the reaction chamber as described
by Zhu [33]. The saturated NaCl solution was loaded into
the bottom of the glass chamber. Exposure to the hydrogen
peroxide vapor was from 3 days to 1 week. Samples were
taken at exposure intervals and analyzed for content of
edaravone using the HPLC method [56]. Degradation
products were not analyzed, except sample no. 10
(Table 1) for which the LC/MS was performed with an
MSD SL Ion Trap mass spectrometer with a 123 ESI and
APCI source coupled with an 1100 series HPLC system.
ˇ
ˇ
diseases’’, ITMS 26240220040 (M. Veverka, J. Gallovic, E. Sva-
jdlenka, B. Meluchova). Financial support from Slovak grant agency
ˇ
´
ˇ
(VEGA 1/0882/13) is gratefully acknowledged (T. Dubaj, P. Simon).
References
1. Higashi Y, Jitsuiki D, Chayama K, Yoshizumi M (2006) Recent
Pat Cardiovasc Drug Discov 1:85
2. Munakata A, Ohkuma H, Nakano T, Shimamura N, Asano K,
Naraoka M (2009) Neurosurgery 64:423
3. Lapchak PA (2010) Expert Opin Pharmacother 11:1753
123

Edaravone cocrystals: synthesis, screening
1349
ˇˇ ´
29. Friscic T, Trask VA, Jones W, Motherwell SDW (2006) Angew
Chem Int Ed 45:7546
4. Lee BJ, Egi Y, van Leyen K, Lo EH, Arai K (2010) Brain Res
1307:22
5. Kikuchi K, Takeshige N, Miura N, Morimoto Y, Ito T, Tancha-
roen S, Miyata K, Kikuchi C, Iida N, Uchikado H, Miyagi N,
Shiomi N, Kuramoto T, Maruyama I, Morioka M, Kawahara KI
(2012) Exp Ther Med 3:3
6. Yamamoto Y, Kuwahara T, Watanabe T, Watanabe K (1996)
Redox Rep 2:333
7. Watanabe K, Watanabe T, Kuwahara T, Yamamoto Y (1997) J
Jpn Oil Chem Soc 46:797
8. Kawai H, Nakai H, Suga M, Yuki S, Watanabe T, Saito KI (1997)
J Pharmacol Exp Ther 281:921
30. Zhang GG, Henry RF, Borchardt TB, Lou X (2007) J Pharm Sci
96:990
31. Childs SL, Rodriguez-Hornedo N, Reddy LS, Jayasankar A,
Maheshwari C, McCausland L, Shipplett R, Stahly BC (2008)
CrystEngComm 10:856
32. Li Z, Yang B-S, Jiang MO, Eriksson M, Spinelli E, Yee N,
Senanayake C (2009) Org Process Res Dev 13:1307
33. Zhu D, Zhang GG, George KL, Zhou D (2011) J Pharm Sci
100:3529
34. Jayasankar A, Reddy L, Bethune SJ, Rodriguez-Hornedo N
(2009) Cryst Growth Des 9:889
9. Yoshinori M, Naoki M (1988) Stable injection containing
3-methyl-1-phenyl-2-pyrazolon-5-one. Jpn Patent (Kokoku)
JP63-132833, 4 Jun 1988
ˇ
´
35. Veverka M, Simon P, Lokaj J, Veverkova E (2012) Monatsh
Chem 143:65
ˇ
36. Veverka M, Simon P, Gallovic J, Jorık V, Veverkova E, Dubaj T
ˇ
(2012) Monatsh Chem 143:1405
´
´
10. Kamat CD, Gadal S, Mhatre M, Williamson KS, Pye QN,
Hensley K (2008) J Alzheimers Dis 15:473
11. Jin Y, Mima T, Raicu V, Park CY, Shimizu K (2002) Neurosci
Res 43:75
12. Kokura S, Yoshida N (2005) Cancer Lett 229:223
13. Zhao X, Aronowski J, Sun G, Zhang J, Strong R, Zhang Y, Grotta
CJ (2005) J Neurol Sci 22:374
14. Zhang W, Sato K, Hayashi T, Omori N, Nagano I, Kato S,
Horiuchi S, Abe K (2004) Neurol Res 26:342
15. Shah D, Wader GR, Mehta T (2011) Novel injectable combina-
tion. WIPO Patent WO/2011/111070, 15 Sep 2011
37. Vishweshwar P, McMahon JA, Peterson ML, Hickey MB,
Shattock TR, Zaworotko MJ (2005) Chem Commun 4601
38. McMahon JA, Bis JA, Vishweshwar P, Shattock TR, McLaughlin
OL, Zaworotko MJ (2005) Z Kristallogr 220:340
39. Ueto T, Takata N, Muroyama N, Nedu A, Sasaki A, Tanida S,
Terada K (2012) Cryst Growth Des 12:485
40. Schultheiss N, Roe M, Boerrigter SXM (2011) CrystEngComm
13:611
41. Stahly GP (2007) Cryst Growth Des 7:1007
42. Trask AV, van de Streek J, Motherwell WDS, Jones W (2005)
Cryst Growth Des 5:2233
43. Childs SL, Hardcastle KI (2007) CrystEngComm 9:364
44. Limwikrant W, Nagai A, Hagiwara Y, Higashi K, Yamamoto K,
Moribe K (2012) Int J Pharm 431:237
45. Hilfiker R (ed) (2006) Polymorphism in the pharmaceutical
industry. Wiley-VCH, New York
¨
16. Almarsson O, Zaworotko MJ (2004) Chem Commun 17:1889
17. Schultheiss N, Newman A (2009) Cryst Growth Des 9:2950
18. Trask AV, Motherwell WDS, Jones W (2005) Cryst Growth Des
5:1013
19. Papadopoulos G, Boskou D (1991) J Am Oil Chem Soc 68:669
20. Rice-Evans CA, Miller NJ, Paganga G (1996) Free Radic Biol
Med 20:933
¨
21. Nakamura Y, Torikai K, Ohto Y, Murakami A, Tanaka T,
Ohigashi H (2000) Carcinogenesis 21:1899
22. Babich H, Sedletcaia A, Kenigsberg B (2002) Pharmacol Toxicol
91:245
23. Childs SL (2007) Metronidazole cocrystals and imipramine co-
crystals. US Patent WO2007/067727, 14 Jun 2007
24. Kavuru P, Aboarayes D, Arora KK, Clarke DH, Kennedy A,
Marshall L, Ong TT, Perman J, Pujari T, Wojtas Ł, Zaworotko
JM (2010) Cryst Growth Des 10:3568
25. Weyna DR, Cheney ML, Shan N, Hanna M, Wojtas L, Zawor-
otko MJ (2012) CrystEngComm 14:2377
26. Springuel G, Norberg B, Robeyns K, Wouters J, Leyssens T
(2012) Cryst Growth Des 12:475
46. Aakeroy CB, Salmon DJ (2005) CrystEngComm 7:439
47. Haynes DA, Chisholm JA, Jones W, Motherwell WDS (2004)
CrystEngComm 6:584
48. Balasubramani K, Muthiah PT, Lynch DE (2007) Chem Cent J
1:28
49. Kojima T, Tsutsumi S, Yamamoto K, Ikeda Y, Moriwaki T
(2010) Int J Pharm 399:52
50. Zhang X-L, Ye B-H, Chen X-M (2005) Cryst Growth Des 5:1609
51. Descamps M, Willart JF, Dudognon E, Caron V (2007) J Pharm
Sci 96:1398
52. Crowley KJ, Zografi G (2002) J Pharm Sci 91:492
53. Leyssens T, Springuel G, Montis R, Candoni N, Veesler S (2011)
Cryst Growth Des 12:1520
27. Viertelhaus M, Hilfiker R, Blatter F (2009) Cryst Growth Des
9:2220
54. Habgood M, Price SL (2010) Cryst Growth Des 10:3263
´
55. Good DJ, Rodrıguez-Hornedo N (2009) Cryst Growth Des 9:2252
56. Lanchote VL, Ping WC, Santos SR (1997) Ther Drug Monit
19:705
´
´
´
28. Arenas-Garcıa IJ, Herrera-Ruiz D, Mondragon-Vasquez K,
¨
Morales-Rojas H, Hopfel H (2012) Cryst Growth Des 12:811
123