Determination of quinacrine dihydrochloride dihydrate stability and characterization of its degradants

Article abstract of DOI:10.1002/jps.22543

Although quinacrine dihydrochloride dihydrate is a widely used drug substance, a comprehensive determination of its stability profile is lacking. In this work, an integrative approach is implemented to determine the drug stability both in the solid state and aqueous solutions, identify the impurities that can be found in the active pharmaceutical ingredient, and evaluate the associated toxicity risks. Thermal analyses pointed out a two-step dehydration of the solid state. This phenomenon seems to be consistent with the organization of the water molecules in the crystal structure and results in the destruction of the lattice. Seven related compounds of quinacrine have been identified by liquid chromatography-ion trap mass spectrometry. The main thermal degradant both in the solid state and the solution corresponds to the N-deethyl compound, whereas quinacrine tertiary amine oxyde appears to be a signal impurity of oxidative stress in solution. Moreover, two photolytic impurities can be formed in solution either by aromatic amine cleavage or via O-demethylation. Additionally, using computational approaches, the analysis of the potential toxicity of the impurities compared with the parent compound one shows that ketone and O-demethyl derivatives may exhibit specific toxicity profiles.

Full text of DOI:10.1002/jps.22543

Determination of Quinacrine Dihydrochloride Dihydrate
Stability and Characterization of Its Degradants
ROMAIN ROTIVAL,1,2 PHILIPPE ESPEAU,1 YOHANN CORVIS,1 FRANC¸ OIS GUYON,2 BERNARD DO2
1Laboratoire Physico-chimie Industrielle du Me´dicament, EA 4066, Faculte´ des Sciences Pharmaceutiques et Biologiques,
Universite´ Paris-Descartes, 75006 Paris, France
2De´partements Laboratoires et Innovation Pharmaceutique, Etablissement Pharmaceutique des Hoˆpitaux de Paris, 75005 Paris, France
Received 21 December 2010; accepted 21 February 2011
Published online 21 March 2011 in Wiley Online Library (wileyonlinelibrary.com). DOI 10.1002/jps.22543
ABSTRACT: Although quinacrine dihydrochloride dihydrate is a widely used drug substance,
a comprehensive determination of its stability profile is lacking. In this work, an integrative
approach is implemented to determine the drug stability both in the solid state and aqueous
solutions, identify the impurities that can be found in the active pharmaceutical ingredient,
and evaluate the associated toxicity risks. Thermal analyses pointed out a two-step dehydra-
tion of the solid state. This phenomenon seems to be consistent with the organization of the
water molecules in the crystal structure and results in the destruction of the lattice. Seven
related compounds of quinacrine have been identified by liquid chromatography–ion trap mass
spectrometry. The main thermal degradant both in the solid state and the solution corresponds
to the N-deethyl compound, whereas quinacrine tertiary amine oxyde appears to be a signal
impurity of oxidative stress in solution. Moreover, two photolytic impurities can be formed
in solution either by aromatic amine cleavage or via O-demethylation. Additionally, using
computational approaches, the analysis of the potential toxicity of the impurities compared
with the parent compound one shows that ketone and O-demethyl derivatives may exhibit
specific toxicity profiles. © 2011 Wiley-Liss, Inc. and the American Pharmacists Association
J Pharm Sci 100:3223–3232, 2011
Keywords: physical stability; chemical stability; dehydration; impurity profiling; thermal
analysis; mass spectrometry
erythematosus.4 It has also been tested in diseases
INTRODUCTION
such as pneumothorax,5 asthma,6 and hepatitis.7
The dihydrochloride dihydrate salt of quinacrine, 4-
Moreover, QDD has been considered in the treatment
N-(6-chloro-2-methoxyacridin-9-yl)-1-N,1-N-diethyl-
of Creutzfeldt–Jacob disease8–10 and continues to be
pentane-1,4-diamine, is the first synthetic anti-
the subject of various research.10–12
malarial drug (Fig. 1). Quinacrine dihydrochloride
The physicochemical stability profile of QDD and
dihydrate (QDD) was used in the oral treatment
its degradation scheme are poorly known. Besides, the
of malaria for over seven decades and was re-
research and characterization of drug-related com-
placed by other drugs such as chloroquine to face a
pounds such as degradants or synthesis impurities
spread of resistance.1 However, this drug remains
(SIs) is of major importance because they may not only
of great importance because of its broad spectrum
alter properties such as solubility, crystallization ki-
of pharmacological activity. The Food and Drug
netics, crystal size distribution, and morphology,13–15
Administration has approved the use of QDD not
but they may also exhibit their own pharmaceutical
only in the treatment of tapeworm infestation2
and toxicological activities.
and resistant giadiasis3 but also in systemic lupus
The aim of this study is to elucidate the QDD stabil-
ity profile both in solid state and solution, including
reactivity toward various stress agents and elucida-
tion of degradation pathways. In addition, the iden-
tification of the related compounds that may be en-
countered in the drug substance allows evaluating
their inherent toxicity risk.
Additional Supporting Information may be found in the online
version of this article. Supporting Information
Correspondence to: Romain Rotival (Telephone: +331-46691395;
Fax: +331-46691492; E-mail: romain.rotival@parisdescartes.fr)
Journal of Pharmaceutical Sciences, Vol. 100, 3223–3232 (2011)
© 2011 Wiley-Liss, Inc. and the American Pharmacists Association
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 100, NO. 8, AUGUST 2011
3223

3224
ROTIVAL ET AL.
tector [Debye–Scherrer geometry, transmission mode,
Co K"1 radiation (λ = 1.7890 Å)].
The Mercury 2.3 program (The Cambridge Crystal-
lographic Data Centre, Cambridge, UK) was used for
graphic representation.
Powder micrographs were taken using a JSM-
6510LV (Jeol Ltd., Tokyo, Japan) scanning electron
microscope (SEM) in low vacuum mode (40 Pa) with a
10 kV accelerating voltage. Powder samples were af-
fixed to aluminum stages using double adhesive car-
bon conducting tape. Images of 1280 × 960 pixels res-
olution were recorded.
Figure 1. Molecular structure of quinacrine with
numbering of the atoms.
EXPERIMENTAL
Stability Study of the Drug Substance
Reagents and Materials
Thermal stability of the solid state was assessed by
submitting samples to different temperature check-
points using the TGA apparatus.
Exhaustive stress tests in solution were performed
with aqueous solutions containing 500 :g·mL−1 of
QDD. Solutions submitted to different stress condi-
tions were placed in 5 mL hermetic vials. The vials
were heated using an oven. Each stress test was per-
formed in triplicate.
The QDD standard substance (assay: 99.0%) and
6-chloro-2-methoxy-9-acridone (assay: 98.0%) were
purchased from Seratec (Courville-sur-Eure, France).
For the mobile phase preparation, analytical grade
trifluoroacetic acid (TFA) was purchased from Sigma
(Steinheim, Germany) and high-performance liquid
chromatography (HPLC) grade acetonitrile (ACN)
was purchased from Prolabo-VWR International SAS
(Fontenay-sous-Bois, France). Ultrapure water was
obtained by deionization and filtration through a
Milli-Q water purification system (Millipore, Saint-
Quentin-en-Yvelines, France). For the degradation
studies, hydrochloric acid, sodium hydroxide, and hy-
drogen peroxide were of analytical grade and supplied
by Prolabo-VWR International SAS. Azobis-4-cyano-
valeric acid (ACVA) was purchased from Sigma.
The independent stock solutions of 500 :g·mL−1
QDD were prepared by dissolving appropriate
amounts of the substance in ultrapure water. Solu-
tions were prepared extemporaneously before use.
The impurity test solution of 3-chloro-7-methoxy-
9-acridone was prepared in an ultrapure water/
acetonitrile mixture (50:50, v/v).
Hydrolytic stress studies were set as follows: (a) in
acidic condition—1 mL HCl (1 M) was added to 1 mL
of QDD solution and placed at 80◦C for 2.5 h, this
solution was then neutralized by addition of 1 mL
NaOH (1 M) at room temperature; (b) in alkaline con-
dition—1 mL NaOH (1 M) was added to 1 mL QDD
solution and placed at 80◦C for 2.5 h, this solution
was then neutralized by addition of 1 mL of HCl (1 M)
at room temperature; and (c) at unmodified pH—1 mL
of ultrapure water was added to 1 mL of QDD solution
and placed at 80◦C for 4 h.
Three oxidizing conditions were tested: (a) hydro-
gen peroxide stress: 1 mL H2O2 (1%, v/v) was added
to 1 mL of QDD solution and left at 25◦C for 24
h, (b) molecular oxygen: QDD solution in methanol/
water 50:50 (v/v) was submitted to oxygen bubbling
in hermetic vials and stored at 40◦C for 7 days, and
(c) singlet oxygen: QDD solution in methanol/water
50:50 (v/v) containing 5% (mol/mol) of radical initiator
ACVA was submitted to oxygen bubbling in hermetic
vials and stored at 40◦C for 7 days. Radical initia-
tor ACVA was used to predict products of free radical
autoxidation.16,17
Solid-state Analysis
Thermogravimetric analyses (TGA) were performed
using a TGA 850 from Mettler–Toledo (Greifensee,
Switzerland) sensitive to 0.001 mg. Samples of ap-
proximately 10 mg were analyzed in open pans at
a heating rate of 10 K·min−1
.
Differential scanning calorimetry (DSC) experi-
ments were run with a DSC822e thermal analyzer
Photolysis was performed with a combination of
ultraviolet (UV) light at 350 nm and cool white fluo-
rescent light. The suitability of light exposition was
checked using quinine as chemical actinometer in ac-
cordance with the International Conference on Har-
monization (ICH) Q1B guideline.18
from Mettler–Toledo. Indium (Tfus = 429.75 K, ꢀfus
H
= 28.45 J·g−1) was used as a standard for temperature
and enthalpy calibration. Specimens were weighed
in aluminum pans using a microbalance sensitive to
0.01 mg.
X-ray powder diffraction profile was recorded us-
ing a high-resolution diffractometer (Inel, Artenay,
France) equipped with a CPS (curve position angu-
lar sensitive detector) 120 position-sensitive multide-
Degradation after each test was assessed by chro-
matographic peak area ratio determination using the
signal of a pure QDD sample as reference. Because
most of the impurities were unavailable as standards,
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 100, NO. 8, AUGUST 2011
DOI 10.1002/jps

QDD STABILITY AND DEGRADANTS
3225
this calculation was also used to determine the chem-
ical purity of the QDD raw material.
Toxicity Risk Analysis
A comparative study was led using quantitative
structure–activity relationship approaches. Derek for
Liquid Chromatography–Mass Spectrometry System
R
Windowsꢀ version 11.0.0 (Lhasa Ltd., Leeds, UK) and
A reversed-phase liquid chromatography (LC) method
with mass spectrometry (MS) detection was set up to
achieve the impurity profile of QDD.
R
ACD Tox suiteꢀ version 2.9 (ACD Labs) were used to
predict the potential toxicity of the impurities.
The LC system (Thermoseparations Products,
Fremont, California) consisted of a P4000 quater-
nary pump, a SCM400 vacuum degasser, and an
AS3000 autosampler. The selected analytical column
was a Varian Polaris C18-A (Ansys Technologies, Lake
Forest, California), 150 mm length, 4.6 mm internal
diameter, and 3 :m particle size. The flow rate was
set at 1.0 mL·min−1 and the sample injection volume
was 20 :L.
A mobile phase linear gradient was set up (solvent
A: TFA in ultrapure water 0.5% (v/v), solvent B: ACN).
At injection time, the volume ratio A:B was set at
75:25, it went to 60:40 between 15 min and 20 min
of analysis time, then a return to the initial condi-
tions was set from 30 to 35 min. Inter-run stabiliza-
tion lasted for 5 min. The gradient ensured absence of
impurity coelution and took into account late eluting
impurities.
Mass spectrometry detection consisted of a LCQ
quadrupole ion trap mass spectrometer (Thermo
Finnigan, San Jose, California) and an electrospray
ionization (ESI) source. A postcolumn split was set up
to obtain a resulting flow rate equal to 0.3 mL·min−1
upstream to the MS spectrometer. The instrumental
parameters of mass spectrometer were as follows: ESI
source operated in positive ionization mode, with a
4.5 kV spray voltage and an 80 :A intensity. High pu-
rity nitrogen was used as sheath/auxiliary gas (flow
rate 80:20). Capillary voltage was 3.0 V and its tem-
perature was set at 270◦C. The analysis of impurities
was achieved by scanning a mass range from 50 to
500 mass-to-charge ratio (m/z).
RESULTS AND DISCUSSION
Characterization of the Drug Substance
As previously determined by Courseille et al.,20 QDD
crystallizes in the triclinic system, space group Pi
¯
with two molecules per unit cell. Representation of
the environment of the water molecules in the struc-
ture is given in Figure 2. The acridine molecules
stack in the direction of the c-axis. As can be seen,
they make up channels, in which the W31 and W32
water molecules as well as the two Cl29 and Cl30
chlorine atoms are included. W31 is engaged in two
hydrogen bondings that imply Cl29 and Cl30, re-
spectively, and interacts with N18 belonging to the
secondary amine. As far as W32 is concerned, it is en-
gaged in two hydrogen bondings with the same chlo-
rine atom and, also, in a strong interaction with Cl16.
The lengths of the solvation bondings are given in
Table 1.
The TGA experiments, performed at 10 K·min−1
,
show two dehydration steps during the heating pro-
cess (Fig. 3). From 25◦C to 95◦C, a slow phenomenon
corresponding to the loss of one molecule of water
(weight loss equal to 3.5%) occurs. It is followed by
a sharper one from 95◦C to 150◦C (weight loss equal
to 3.5%). The total weight loss (7.5%), measured ex-
perimentally, agrees with the calculated weight loss
value (7.1%), assuming two moles of water per mole
of quinacrine. Moreover, the DSC profile shows two
overlapping endotherms in the 25◦C–150◦C temper-
ature range. The calculated molar enthalpy of dehy-
dration is 91 2 kJ·mol−1. This dehydration enthalpy
(45.5 kJ per mol of water) compares well with the liter-
ature values for other molecules presenting internal
cavities: ∼50 kJ per mol of water for $-cyclodextrin
hydrates21 and between 50 and 63 kJ per mol of wa-
ter for hydrated cucurbituril.22 It is also in agreement
with the literature values for aqua complexes with
Mg, Co, Ni, or Zn.23,24
Tandem mass spectrometry (MS2) studies were per-
formed by colliding specific precursor ion with the ni-
trogen damping gas present in the trap for 30 ms.
The isolation width was fixed at 2.0 mass unit (u).
The product ion spectra were obtained by scanning a
50–500 m/z range.
R
Data acquisition software was Xcaliburꢀ (Thermo
Finnigan), version 1.2. Processing of MS data was
R
performed using MS Managerꢀ software version 10
As a matter of fact, W31 presents a stronger binding
than W32 (Table 1). However, it is worth considering
that W32, contrary to W31, is interposed between Cl30
atoms along the b axis (Fig. 2). This suggests that the
interposed chlorine atoms could delay the departure
of W32 from the cavity, whereas the leaving of W31
would be less hindered. This is in good accordance
with the two dehydration steps noticed during the
TGA. As a consequence, the first and second dehydra-
tion steps following the increase of temperature may
(ACD Labs, Toronto, Ontario, Canada).
Liquid chromatography–mass spectrometry condi-
tions were set up to ensure detection of main impu-
rities. A 0.05% threshold was chosen in accordance
with ICH recommendations19 and concentrations of
the test samples were adjusted according to the de-
tection limit of the method. Therefore, each sample
was adequately diluted in ultrapure water to obtain
a final concentration of 20 :g·mL−1
.
DOI 10.1002/jps
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 100, NO. 8, AUGUST 2011

3226
ROTIVAL ET AL.
Table 1. Lengths of Solvation Bondings (Extracted from the
Crystalline Structure in Ref. 20)
Contacts
O31Cl29
Length (Å)
3.210
3.209
2.937
3.283
3.308
3.080
O31Cl30
O31N18
O32Cl29
O32Cl30
O32Cl16
Tandem mass spectrometry fragmentation pat-
terns of quinacrine and the impurities were correlated
in order to achieve their structural elucidations.
Protonated Quinacrine Peaks At m/z 400. A charac-
teristic 37Cl isotopic ion pattern at signal (M + 2)H+
with approximately 35% relative abundance outlines
the presence of a single chlorine atom in detected ions.
Table 2 shows the daughter ion mass spectrum of pro-
tonated quinacrine after 25% collision-induced disso-
ciation (CID).
The major route of quinacrine fragmentation takes
place through C N bond cleavage.25 Fragments at
m/z 142 and m/z 259 are reckoned to be the side
chain and the tricyclic moieties, respectively. The m/z
157 fragment represents the side chain including the
secondary amine group and the m/z 244 fragment is
generated through the loss of a NH2 group from the
acridine moiety. The m/z 327 signal corresponds to
quinacrine without the terminal tertiary amine.
Tandem mass spectrometry fragmentation pat-
terns of impurities similar to those underwent by
quinacrine were observed (Table 2). SI1 exhibits a
protonated molecule [M + H]+ at m/z 372. The sig-
nal difference between quinacrine and the impurity
is equal to 28 u and only occurs with the fragments of
the side chain. N-deethyl quinacrine has been thereby
Figure 2. Representation of the environment of the wa-
ter molecules in the quinacrine dihydrochloride dihydrate
crystal (top: b-axis vertical, bottom: projection along b).
be related to the departure of W31 and W32, respec-
tively.
The chemical purity of the raw material was as-
sessed by HPLC–MS analysis. Four main impurities
were pointed out. As these impurities were detected
in the raw material stored under recommended condi-
tions of conservation (i.e., tightly closed container pro-
tected from light), they were considered and denoted
as SIs. A typical chromatogram including impurities
SI1 to SI4 is shown in Figure 4a.
Figure 3. ATG (dashed line) and DSC (solid line) curves
of quinacrine dihydrochloride dihydrate (10 K·min-1, open
pan).
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 100, NO. 8, AUGUST 2011
DOI 10.1002/jps

QDD STABILITY AND DEGRADANTS
3227
identified as resulting from the loss of an ethyl group
on the tertiary amine.
CID showed a loss of 15 u, corresponding to the cleav-
age of a methyl. The presence of a chlorine isotopic
peak and an even number of nitrogen atoms in the
molecule accounts for the fact that the most plausible
structure is issued from an absence of side chain on
the secondary amine. Thus, the structure of 6-chloro-
2-methoxyacridin-9-amine has been attributed to SI2.
The assignment of SI1 and SI2 structures is in ac-
cordance with their previously reported identification
as human and rat metabolites, respectively.25
SI2 is characterized by a protonated molecule
[M + H]+ at m/z 259. The MS2 analysis with 40%
The molecular ion [M + H]+ of SI3 is m/z 345. De-
tection of fragments at m/z 327 and m/z 259 implies
an even number of nitrogen atoms on the molecule,
suggesting the presence of the aromatic nitrogen as
well as the secondary amine moiety. The 18 u differ-
ence between the molecular ion and the fragment at
m/z 327 suggests the neutral loss of water and thus
the presence of an OH moiety at the side chain. Hence,
this impurity probably corresponds to 4-[(6-chloro-2-
methoxyacridin-9-yl)amino]pentan-1-ol.
The MS spectrum of SI4 shows a protonated
molecule [M + H]+ at m/z 260. The structure of
6-chloro-2-methoxy-9-acridone was confirmed by cor-
relation with MS2 fragmentation pattern of synthe-
sized standard. This impurity was previously de-
scribed as a byproduct resulting from the hydrolysis
of 9-chloroacridines.26
Structures of the SIs detected in the raw material
are presented in Figure 5. As the signals obtained
from MS and UV detections (results not shown) were
very similar, a 99.5% chemical purity of the raw ma-
terial was admitted. Considering the low levels of SIs,
it is likely that they did not significantly hinder the
formation of the crystal lattice. Because of their signif-
icant structural analogy with quinacrine, they may be
included in the network, generating isolated defects
of the crystal.
Determination of the Drug Substance Stability
Stability in Solid State
Figure 4. LC–MS analysis of powder samples stored at
ambient temperature (a, purity ∼99 %) and heated to 250◦C
(b, purity ∼93 %), solution samples submitted to heating at
80◦C at native pH (c, purity ∼92 %) in acidic (d, purity
∼90 %) and alkaline (e, purity ∼91 %) conditions.
The influence of heating on both the physical state
and chemical purity of QDD was investigated.
As presented above, the TGA thermogram shows
that dehydration is complete when the temperature
Table 2. Origin of Detected Compounds and Major MS2 Fragments
Monoisotopic
Fragments Containing
Fragments Attributed
Origin
Notation
Signal (m/z)
the Acridine Ring (m/z)
to Side Chain (m/z)
Initially present
Initially present, increased in stress tests
Initially present
Initially present
Initially present, increased in stress tests
Stress tests
Quinacrine
SI1
400
372
259
345
260
416
244
386
244, 259, 327
244, 259, 327
216, 244
244, 259, 327
217, 225, 245
259, 327, 388
201, 216, 229
230, 245, 313
142, 157
114, 129
ND
ND
ND
158
ND
142, 157
SI2
SI3
SI4
DI1
DI2
DI3
Stress tests
Stress tests
ND, not detected.
DOI 10.1002/jps
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 100, NO. 8, AUGUST 2011

3228
ROTIVAL ET AL.
to the degradation of the anhydrous compound. This
is in agreement with previously reported tempera-
tures of chemical decomposition.27
Moreover, analysis by time interval shows that
in the experimental conditions, degradation impu-
rities were detected in the samples heated to the
temperature of 250◦C (Fig. 4b). The major thermal
degradant detected has been identified by MS2 anal-
ysis as deethyl quinacrine (SI1), pointing out the
fragility of the tertiary amine bonds.
Stability in Solution
Figure 5. Molecular structures of the synthesis impuri-
ties detected in the quinacrine dihydrochloride dihydrate
raw material.
Stability studies in solution were performed to inves-
tigate water-mediated degradations of the compound.
In order to obtain a realistic decomposition profile, the
stress tests were limited to a decrease from 5% to 20%
of the active drug concentration.28
Under hydrolytic conditions tested at native and
extreme pHs, chemical degradation appears to be
moderate (Fig. 4). The impurity profiles in acidic and
alkaline stress conditions are not significantly differ-
ent from the profile of the native sample. However, SI1
and SI4 signals are increased, and a novel degradation
impurity (DI) denoted DI1 with m/z 416 is detected at
trace levels. Considering the fact that DI1 is found
in each stress condition, the heating of the aqueous
solution appears to be the major factor leading to its
formation.
Interestingly, the behavior of the drug substance
under oxidative stress depends on the conditions ap-
plied to the samples. Indeed, exposition to hydrogen
peroxide, molecular oxygen, and oxygen singlet lead
to different results (Fig. 7).
Hydrogen peroxide induced drug degradation be-
yond targeted values (−35%). DI1 is detected in
high level concentrations and three other oxidation-
specific impurities (denoted OX1, OX2, OX3) exhibit
a protonated molecule [M + H]+ at m/z 416. Another
impurity OX4 with m/z 432 molecular ion is also ex-
clusively produced in this stress condition. The gain in
mass of 16 u suggests that the addition of one oxygen
atom took place. Analogously, the gain of 32 u implies
the addition of two oxygen atoms as in hydroperoxide
formation.
of 150◦C is reached. X-ray powder diffraction carried
out on a sample obtained at this temperature shows
an amorphous state (not presented here). Moreover,
dehydration causes a change in powder color from
yellow to orange. Observation using SEM shows that
the resulting particles are amorphous and probably
proceed from cracking phenomena (Fig. 6). This indi-
cates that the departure of the two water molecules
is destructive for the crystal lattice.
It has to be noticed that various heating rates and
isotherm segments were tested during thermal analy-
sis. In this case, the two dehydration steps remained
unseparated, indicating that the network remained
poorly stable after the departure of the W31 water
molecule.
The chemical purity of the samples dehydrated by
TGA at 150◦C was assessed by LC–MS and the ob-
tained profile is identical to the one of the nonheated
samples. These results indicate that dehydration
did not cause chemical degradation of quinacrine as
could be expected from the high energy applied to the
solid sample and the departure of water molecules.
Differential scanning calorimetry analysis shows
an endothermic signal at 254.8 0.5◦C. This signal is
correlated with powder browning and was attributed
The two other oxidative conditions produced a more
selective effect on drug degradation. A significant in-
crease of DI1 levels is noticed in samples submitted to
oxygen bubbling. In addition, this effect seems more
important when the samples are submitted to ACVA.
Hence, DI1 appears to be a signal impurity valid
for the three different types of oxidative stress. DI1
fragmentation is similar to the one of quinacrine. Two
main product ions are detected at m/z 388 and m/z
158. The first one corresponds to the loss of an ethyl
residue on the tertiary amine. The latter of m/z (142 +
16 u) implies that the molecule underwent oxidation
Figure 6. SEM images of the raw material prior (right)
and after (left) thermal dehydration.
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 100, NO. 8, AUGUST 2011
DOI 10.1002/jps

QDD STABILITY AND DEGRADANTS
3229
Thus, DI2 has been identified as being 6-chloro-2-
methoxyacridine. This impurity proceeds from a pho-
todeamination. Interestingly, this phenomenon has
also been reported in the case of tetracycline.31
The molecular ion [M + H]+ of DI3 with m/z 386
presents a 14 u difference in mass from quinacrine.
The MS2 fragmentation of this compound leads to
common product ions at m/z 157 and m/z 142, sug-
gesting that the side chain is fully preserved. A con-
stant difference of 14 u is found on the other frag-
ments in comparison with quinacrine fragmentation
pattern, that is, m/z 244, m/z 313, and m/z 230. The
latter fragment suggests a demethylation on the ether
group. The structure of 6-chloro-9-[(4-diethylamino-1-
methyl-butyl)amino]acridin-2-ol has thus been iden-
tified. This structure assignment is in accordance
with the previously reported identification of DI3 as a
rat metabolite.25
The proposed scheme for the in vitro formation of
DI3 implies a cation radical intermediate. The lat-
ter is produced under light exposition as it has been
described in the case of alkyl aryl ethers,32 addition
of H2O to the cation radical is followed by elimina-
tion of methanol with the formation of acrinol. Simi-
larly, acifluorfen shows photocleavage of ether bond-
ing and formation of a phenolic compound.33 This
degradation pathway is worth considering because
it has not been identified in the photodegradation
of primaquine, which holds a methoxy group on a
quinoline.34
Besides, previously performed laser flash pho-
tolytic studies and electron paramagnetic resonance
studies have shown that most antimalarials produce
active intermediates, and have highlighted the pres-
ence of an unidentified quinacrine nitroxide.35 The re-
sults obtained in our study show that DI1 is present
at significant levels after photolytic stress. As a con-
sequence, this compound probably corresponds to the
latter unidentified nitroxide.
Figure 7. LC–MS analysis of aqueous samples submitted
to oxygen bubbling (a, purity ∼94 %), oxygen bubbling with
ACVA (b, purity ∼92 %), oxygen peroxide (c, purity ∼65 %),
and light (d, purity ∼88 %).
at the side chain. The structure of N-oxyde quinacrine
has been characterized.
Although the quinacrine structure contains three
amine functions, the tertiary amine seems to be a
preferential site for oxidation as shown by hydrolytic
and oxidative stress tests. This can be explained by
the fact that the lone pair of electrons on the sec-
ondary amine is poorly available because it is engaged
in a tautomeric equilibrium with the heterocyclic ni-
trogen atom.29 Moreover, the pKa of the heterocyclic
nitrogen atom and the one of the tertiary amine on
the side chain (10.2 and 8.0, respectively)30 reflect a
greater availability of the electron pair of the latter.
Photostress studies in solution have underlined
a reactivity of the quinacrine molecule. Two main
impurities denoted DI2 and DI3 are specifically de-
tected in samples submitted to photolytic degradation
(Fig. 7).
DI2 Peaks At m/z 244. Fragmentation of the ion re-
quires a more important energy (50% CID). Major ion
fragments are detected at m/z 229 and m/z 215, cor-
responding to the losses of the methyl and methoxy
moieties, respectively. Other fragments (most abun-
dant m/z 201) are attributed to acridine ring fis-
sion, consecutively to the high energy of CID applied.
Considering the degradants structures, it has to be
noted that the main degradation pathways of QDD
are amine dealkylation, amine oxidation, and ether
hydrolysis (Fig. 8). Indeed, as the impurities iden-
tified in this study can be found in drug products,
further toxicity considerations are helpful.
Study of the Potential Toxicity of Impurities
The potential toxicity of the identified impurities was
evaluated and compared with the parent compound
properties using structure–activity relationships and
chemical databases.
By considering the fact that drug impurities can be
provided by oral administration, the Lipinski’s rule of
five was first used to estimate their absorption. The
application of this rule has proved relevant in drug
discovery36 but also in the field of toxin molecules.37
This rule states that poor absorption is more likely
DOI 10.1002/jps
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 100, NO. 8, AUGUST 2011

3230
ROTIVAL ET AL.
Figure 8. Proposed degradation pathways of quinacrine dihydrochloride dihydrate in solid
state and aqueous solution.
when the molecule has more than 5 H-bond donors
(HD), 10 H-bond acceptors (HA), a molecular weight
(MW) greater than 500 g·mol−1, and a calculated log
P (Clog P) greater than 5.38 Good absorption is likely
for all the impurities because they present no rule vi-
olations contrary to the parent compound (Table 3). A
more detailed analysis was thus necessary to evaluate
the systemic toxicity of the compounds.
Then, a research for hazardous fragments in the
parent compound retrieves three structural hits con-
sisting in the presence of (a) an acrinamide moiety,
(b) amphiphilic properties, and (c) the hERG (human
ether-a-go-go-related gene) pharmacophore.
The acridine ring is identified as responsible for
mutageneicity. This risk is increased in the pres-
ence of an aromatic amine as it is the case in the
acrinamide moiety.39,40 Moreover, polycyclic aromatic
amines generally exhibit greater mutagenic potency
than other aromatic amines, provided that the poly-
cyclic system does not exceed six rings in size.41 The
amine part of a polycyclic aromatic ring system can
also act as a genotoxic chemical carcinogen following
metabolic activation.42
to induce phospholipidosis in the rat. The halogen
atom improves the permeation through biological
membranes.43–45
The third hazardous fragment in quinacrine is
the hERG pharmacophore. Two matching molecule
sites are present, each consisting of a benzene ring
bonded to a nitrogen and a diamine chain in an ortho
position.46,47 hERG encodes for the alpha-subunit of a
potassium channel. The blocking of this channel can
lead to the alteration of the ventricular repolariza-
tion phase in the heart, with lengthening of the QT
interval.48,49
Finally, in terms of hazardous fragments, the ob-
servation of the structures of the impurities reveals
little uniqueness from quinacrine (Table 4).
SI4 is quite distinct from the other potential degra-
dation products. The diaryl ketone group, which is
limited to this particular impurity imparts photoal-
lergenicity and phototoxicity.50 Additionally, an anal-
ogy with acetaminophen is noted and can confer
hepatotoxicity.51,52
Besides, it is worth considering that the in vitro
photolysis of QDD yields DI3, whose structure holds
a phenol. As a matter of fact, the phenol moiety is
known to be photosensitive.40 DI3 could therein con-
tribute to the in vivo phototoxicity of QDD.
Additionally, quinacrine is a cationic amphiphilic
drug consisting of a hydrophilic cationic amine and
a hydrophobic ring. This property was observed
Table 3. Molecular Descriptors of the Detected Compounds
Quinacrine
SI1
SI2
SI3
SI4
DI1
DI2
DI3
CAS number
log P
83-89-6
5.59
399.96
908844-46-2 3548-09-2 916911-56-3 13161-87-0 83196-72-9 21332-86-5 854892-24-3
4.66
3.53
3.77
4.33
2.95
4.13
4.02
MW (g·mol−1
)
371.90
258.70
344.84
259.69
415.96
243.68
385.93
HD
HA
1
4
1
2
4
0
2
3
0
2
4
0
1
3
0
1
5
0
0
2
0
2
4
0
Lipinski violations
MW, molecular weight; HD, number of hydrogen bond donors; HA, number of hydrogen bond acceptors
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 100, NO. 8, AUGUST 2011
DOI 10.1002/jps

QDD STABILITY AND DEGRADANTS
Involved Compounds
3231
Table 4. Hazardous Fragments Identified in the Detected Compounds
Structural Hit
Toxicity Risk
Acrinamide/acridine alone
Hydrophilic cationic amine and hydrophobic ring region
hERG pharmacophore
Diaryl ketone group
Phenol
Mutagenicity, chromosome damage
Phospholipidosis
QT interval lengthening
Photoallergenicity and phototoxicity
Photosensitivity
QA, SI1, SI2, SI3, DI1, DI3 / SI4, DI2
QA, SI1, DI1, DI3
QA, SI1, DI1, DI3
SI4
DI3
10. Collinge J, Gorham M, Hudson F, Kennedy A, Keogh G, Pal S,
Rossor M, Rudge P, Siddique D, Spyer M, Thomas D, Walker
S, Webb T, Wroe S, Darbyshire J. 2009. Safety and efficacy
of quinacrine in human prion disease (PRION-1 study): A
patient-preference trial. Lancet Neurol 8:334–344.
11. Wang YX, Bi QW, Dong L, Li X, Ge XY, Zhang XX, Fu J,
Wu DC, Li SL. 2010. Quinacrine enhances cisplatin-induced
cytotoxicity in four cancer cell lines. Chemotherapy 56:127–
134.
12. Mollapour M, Neckers L. 2010. Quinacrine: New anti-
tumor application for an old anti-malaria drug. Cell Cycle
9:228.
13. Descamps G, Cartigny Y, Sanselme M, Petit MN, Petit S,
Aubin E, Coquerel G. 2009. Structural and physicochemical
characterization of a solid solution produced by antisolvent
crystallization of a new phosphoantigen. Cryst Growth Des
9:3910–3917.
CONCLUSIONS
The physicochemical stability study of QDD has been
performed. The main drug impurities have been iden-
tified and characterized regarding stress specificity
and potential toxicity. Indeed, various degradation
conditions applied in the solid state and aqueous so-
lutions allowed establishing a comprehensive drug
stability profile. The latter should define the synthe-
sis and storage conditions of the raw material, and
should guide preformulation studies. Early in drug
development, the likelihood and the conditions of ap-
pearance of impurities may be correlated with their
potential toxicity.
14. Fiebig A, Jones MJ, Ulrich J. 2007. Predicting the effect of
impurity adsorption on crystal morphology. Cryst Growth Des
7:1623–1627.
15. Ottens M, Lebreton B, Zomerdijk M, Rijkers MPWM,
Bruinsma OSL, van der Wielen LAM. 2004. Impurity effects
on the crystallization kinetics of ampicillin. Ind Eng Chem Res
43:7932–7938.
16. Hovorka SW, Schoneich C. 2001. Oxidative degradation
of pharmaceuticals: Theory, mechanisms and inhibition.
J Pharm Sci 90:253–269.
17. Nelson ED, Harmon PA, Szymanik RC, Teresk MG, Li L,
Seburg RA, Reed RA. 2006. Evaluation of solution oxygena-
tion requirements for azonitrile-based oxidative forced degra-
dation studies of pharmaceutical compounds. J Pharm Sci
95:1527–1539.
ACKNOWLEDGMENTS
The authors want to thank Philippe Negrier and
Yohann Cartigny for helpful discussions, Rene Lai-
Kuen for SEM images, Herve Graffard, Sandrine Roy
´
and Alice Richard for their support.
REFERENCES
1. Tanenbaum L, Tuffanelli DL. 1980. Antimalarial agents.
Chloroquine, hydroxychloroquine, and quinacrine. Arch
Dermatol 116:587–591.
2. Koul PA, Wahid A, Bhat MH, Wani JI, Sofi BA. 2000.
Mepacrine therapy in niclosamide resistant taeniasis. J Assoc
Physicians India 48:402–403.
18. ICH Q1B. 1997. Photostability testing of new drug substances
and products. Fed Reg 62:27115–27122.
19. ICH Q6A. 2000. Specifications: Test procedures and accep-
tance criteria for new drug substances and new drug products:
Chemical substances. Fed Reg 65:83041–83063.
3. Gardner TB, Hill DR. 2001. Treatment of giardiasis. Clin
Microbiol Rev 14:114–128.
4. Zuehlke RL, Lillis PJ, Tice A. 1981. Antimalarial therapy for
lupus erythematosus: An apparent advantage of quinacrine.
Int J Dermatol 20:57–61.
5. Janzing HMJ, Derom A, Derom E, Eeckhout C, Derom F,
Rosseel MT, Tyers GFO. 1993. Intrapleural quinacrine in-
stillation for recurrent pneumothorax or persistent air leak.
Ann Thorac Surg 55:368–371.
20. Courseille C, Busetta B, Hospital M.
1973. Crystal
and molecular-structure of quinacrine. Acta Crystallogr B
29:2349–2355.
21. Bilal M, Debrauer C, Claudy P, Germain P, Letoffe JM. 1995.
Beta-cyclodextrin hydration—A calorimetric and gravimetric
study. Thermochim Acta 249:63–73.
22. Germain P, Letoffe JM, Merlin MP, Buschmann HJ. 1998.
Thermal behaviour of hydrated and anhydrous Cucurbi-
turil—A DSC, TG and calorimetric study in temperature range
from 100 to 800 K. Thermochim Acta 315:87–92.
23. Masuda Y, Hatakeyama M. 1998. Measurement of equilib-
rium water vapor pressures for the thermal dehydrations
of some formate dihydrates by means of the transpiration
method. Thermochim Acta 308:165–170.
24. Stoilova D, Koleva V. 1995. Thermal dehydration of magne-
sium selenate hydrates. Thermochim Acta 255:33–38.
25. Huang Y, Okochi H, May BCH, Legname G, Prusiner SB,
Benet LZ, Guglielmo BJ, Lin ET. 2006. Quinacrine is mainly
metabolized to mono-desethyl quinacrine by CYP3A4/5 and its
brain accumulation is limited by P-glycoprotein. Drug Metab
Dispos 34:1136–1144.
6. Struhar D, Kivity S, Topilsky M. 1992. Quinacrine inhibits
oxygen radicals release from human alveolar macrophages.
Int J Immunopharmaco 14:275–277.
7. Bodenheimer HC, Schaffner F, Vernace S, Hirschman SZ,
Goldberg JD, Chalmers T. 1983. Randomized controlled trial
of quinacrine for the treatment of Hbsag-positive chronic hep-
atitis. Hepatology 3:936–938.
8. Korth C, May BCH, Cohen FE, Prusiner SB. 2001. Acridine
and phenothiazine derivatives as pharmacotherapeutics for
prion disease. Proc Natl Acad Sci USA 98:9836–9841.
9. Haik S, Brandel JP, Salomon D, Sazdovitch V, Delasnerie-
Laupretre N, Laplanche JL, Faucheux BA, Soubrie C, Boher
E, Belorgey C, Hauw JJ, Alperovitch A. 2004. Compassionate
use of quinacrine in Creutzfeldt-Jakob disease fails to show
significant effects. Neurology 63:2413–2415.
DOI 10.1002/jps
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 100, NO. 8, AUGUST 2011

3232
ROTIVAL ET AL.
26. Anderson MO, Sherrill J, Madrid PB, Liou AP, Weisman
JL, DeRisi JL, Guy RK. 2006. Parallel synthesis of 9-
aminoacridines and their evaluation against chloroquine-
resistant plasmodium falciparum. Bioorg Med Chem 14:334–
343.
27. O’Neil MJ. 2001. The Merck index—An encyclopedia of
chemicals, drugs, and biologicals. 13th ed.Whitehouse Station:
Merck and Co., Inc.
28. Jansen PJ, Smith WK, Baertschi SW. 2005. Stress testing:
Analytical considerations. In Pharmaceutical stress testing
predicting drug degradation;Baertschi SW, Ed. Boca Raton:
Taylor and Francis group, pp 141–172.
40. Cronin MT, Basketter DA. 1994. Multivariate QSAR anal-
ysis of a skin sensitization database. Sar Qsar Environ Res
2:159–179.
41. Benigni R, Passerini L, Gallo G, Giorgi F, Cotta-Ramusino M.
1998. QSAR models for discriminating between mutagenic and
nonmutagenic aromatic and heteroaromatic amines. Environ
Mol Mutagen 32:75–83.
42. Williams GM, Weisburger JH. 1991. Chemical carcinogene-
sis. In Casarett and Doull’s toxicology: The basic science of
poisons;Amdur MO, Doull J, Klaassen CD, Eds. 4th ed.New
York: McGraw-Hill, pp 127–200.
43. Reasor MJ, Ogle CL, Walker ER, Kacew S. 1988. Amiodarone-
induced phospholipidosis in rat alveolar macrophages. Am Rev
Respir Dis 137:510–518.
44. Lullmannrauch R, Stoermer B. 1982. Generalized lipidosis in
newborn rats and guinea-pigs induced during prenatal devel-
opment by administration of amphiphilic drugs to pregnant
animals. Virchows Arch B 39:59–73.
45. Lullmannrauch R, Nassberger L. 1983. Citalopram-induced
generalized lipidosis in rats. Acta Pharmacol Tox 52:161–
167.
46. Ekins S, Crumb WJ, Sarazan RD, Wikel JH, Wrighton SA.
2002. Three-dimensional quantitative structure-activity rela-
tionship for inhibition of human ether-a-go-go-related gene
potassium channel. J Pharmacol Exp Ther 301:427–434.
47. Cavalli A, Poluzzi E, De Ponti F, Recanatini M. 2002. Toward
a pharmacophore for drugs inducing the long QT syndrome:
Insights from a CoMFA study of HERG K+ channel blockers.
J Med Chem 45:3844–3853.
29. Rosenberg LS, Schulman SG.
1978. Tautomerism of
singly protonated chloroquine and quinacrine. J Pharm Sci
67:1770–1772.
30. Perrin DD. 1965. Dissociation constants of organic bases in
aqueous solution. London: Butterworths.
31. Davies AK, Mckellar JF, Phillips GO, Reid AG. 1979. Photo-
chemical oxidation of tetracycline in aqueous-solution. J Chem
Soc Perk Trans 2:369–375.
32. Musgrave OC. 1969. Oxidation of alkyl aryl ethers. Chem Rev
69:499–531.
33. Vialaton D, Baglio D, Paya-Perez A, Richard C. 2001. Photo-
chemical transformation of acifluorfen under laboratory and
natural conditions. Pest Manag Sci 57:372–379.
34. Kristensen S, Grislingaas AL, Greenhill JV, Skjetne T,
Karlsen J, Tonnesen HH. 1993. Photochemical stability of bi-
ologically active compounds.5. Photochemical degradation of
primaquine in an aqueous-medium. Int J Pharm 100:15–23.
35. Motten AG, Martinez LJ, Holt N, Sik RH, Reszka K, Chignell
CF, Tonnesen HH, Roberts JE. 1999. Photophysical studies
on antimalarial drugs. Photochem Photobiol 69:282–287.
36. Khan MTH. 2010. Predictions of the ADMET properties of
candidate drug molecules utilizing different QSAR/QSPR mod-
elling approaches. Curr Drug Metab 11:285–295.
48. Crumb W, Cavero II. 1999. QT interval prolongation by non-
cardiovascular drugs: Issues and solutions for novel drug de-
velopment. Pharm Sci Technol Today 2:270–280.
49. Mitcheson JS, Chen J, Lin M, Culberson C, Sanguinetti MC.
2000. A structural basis for drug-induced long QT syndrome.
Proc Natl Acad Sci USA 97:12329–12333.
37. Khanna V, Ranganathan S. 2009. Physiochemical property
space distribution among human metabolites, drugs and tox-
ins. BMC Bioinformatics 10.
38. Lipinski CA, Lombardo F, Dominy BW, Feeney PJ. 2001. Ex-
perimental and computational approaches to estimate solu-
bility and permeability in drug discovery and development
settings. Adv Drug Deliv Rev 46:3–26.
50. Pendlington RU, Barratt MD. 1990. Molecular basis of photo-
contact allergy. Int J Cosmet Sci 12:91–103.
51. Larson AM, Polson J, Fontana RJ, Davern TJ, Lalani E, Hynan
LS, Reisch JS, Schiodt FV, Ostapowicz G, Shakil AO, Lee WM.
2005. Acetaminophen-induced acute liver failure: Results of
a United States multicenter, prospective study. Hepatology
42:1364–1372.
39. Ferguson LR, Denny WA. 1990. Frameshift mutagenesis by
acridines and other reversibly-binding DNA ligands. Mutage-
nesis 5:529–540.
52. Zimmerman HJ. 1999. Hepatotoxicity: The adverse effects of
drugs and other chemicals on the liver. 2nd ed.Philadelphia:
Lippincott Williams & Wilkins.
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 100, NO. 8, AUGUST 2011
DOI 10.1002/jps