Note: Descriptions are shown in the official language in which they were submitted.
CA 02798615 2012-12-07
NOVEL COCRYSTALLIZATION OF HYDROCHLORIC ACID SALT
OF
AN ACTIVE AGENT
FIELD OF THE INVENTION
The present disclosure describes cocrystals comprising
active agents, especially active pharmaceutical ingredients
(APIs), and methods relating to cocrystals. in particular,
novel cocrystals are provided of a salt of an active-
pharmaceutical ingredient (such as a salt having chloride
as the counterion) and a guest that forms a relatively
strong interaction with the counterion. Methods are
provided for searching for possible solid state phases of a
sample and include solidifying the sample as a cocrystaa..
Methods are also provided for screening a sample for solid
state phases and include solidifying the sample as a
cocrystal.
,CXGROLJN.J OF THE INVENTION
Cocrystals are crystals that contain two or more non-
identical molecules. Examples of cocrystals may be found
in the Cambridge Structural Database. Examples of
cocrystals may also be found at Etter, Margaret C., and
Daniel A. Adsmond (1990) The use of cocrystallization as a
1
CA 02798615 2012-12-07
method of studying hydrogen bond preferences of 2-
aminopyridine" J. Chem. Sac., Chem. Commun. 1990 589-591,
Etter, Margaret C., John C.. MacDonald, and Joel Bernstein
(1990a) "Graph-set analysis of hydrogen-bond patterns in
organic crystals" Acta Crystallogr., Sect. B, Struct. Sci.
B46 256-262, Etter, Margaret C., Zofia UrbaAczyk-Lipkowska,
Mohammad Zia-Ebrahimi, and Thomas W. Panunto (1990b)
"Hydrogen bond directed cocrystallization and molecular
recognition properties of diarylureas" J. Am. Chem. Soc.
112 8415-8426,
The following articles are also referenced: Carl
Henrik Gcrbotz and Hans-Petter Hersleth, 2000, "On the
inclusion of solvent molecules in the crystal structures of
organic compounds" Acta Cryst. (2000), B56, 625-534; and
V.S. Senthil Kumar, Ashwini Nangia, Amy K. Katz and H.L.
Carrell, 2002, "Molecular Complexes of Some Mono- and
Dicarboxylic Acids with trans-l, 4,-Dithiane-1,4-dioxide"
American Chemical Society, Crystal Growth & Design, Vol. 2,
No. 4, 2002.
The identification of an optimal composition,
formulation, and/or solid state phase is important in the
pharmaceutical field, as well as in other fields including
nutraceuticals, agricultural chemicals, dyes, explosives,
polymer additives, lubricant additives, photographic
chemicals, and structural and electronic materials. The
new methods described herein may be useful in any of these
fields as well as others where solid materials are used.
2
CA 02798615 2012-12-07
SUMMARY OF THE INVENTION
As one aspect, novel cocrystals are provided. The
novel cocrystals comprise one or more active agents,
particularly of the salts of such active agents.
As another aspect, novel cocrystallization methods are
provided which have increased probability of successful
cocrystallization. A suitable method of cocrystallization
may include identifying a crystal. comprising a salt of an
active agent, wherein the salt comprises the active agent
and a negative counterion. One may identify coordination of
the negative counterion (for example, its hydrogen bond
interactions within that crystal.). One may then select a
guest to coordinate more strongly with the negative
counterion than the coordination within the crystal. Based
upon the evaluation of the nonbonded interactions involving
one component of an active agent and/or guest, one selects
another molecule or molecules or a salt that will
coordinate well, or interact strongly with a hydrogen bond
acceptor site that has been identified as being involved in
a weak hydrogen bond. If the acceptor site has the ability
to interact with stronger hydrogen bond donors, and thus
form a more energetically favorable interaction, yet it is
presently involved in a weak interaction, then the
opportunity exists to replace the weak donor with a
stronger one. For example, if a strong hydrogen bond
acceptor is interacting with a weak hydrogen bond donor in
a crystal, a cocrystal could be created by adding a strong
hydrogen bond donor molecule to the system which would
replace the weak donor and bond to the strong acceptor site
in the resulting cocrystal. After the selection of a
suitable guest, a solution, melt, or physical mixture
comprising the active agent, the counterion, and the guest
3
CA 02798615 2012-12-07
may be prepared. The solution or melt is subjected to a
crystallization process, such as evaporation, cooling, or
any of the many well-known processes for forming a crystal
from a solution or melt. The physical mixture can. be ground
to form the cocrystal. A cocrystal is formed comprising
the salt of the active agent and the guest.
As another aspect, the present disclosure provides a
cocrystallization method that produces a novel type of
chloride salt cocrystal structure- The method can be
useful for generating beneficial solid chloride salts of
APIs in cases where the chloride salt was previously
disfavored.
As yet another aspect, novel forms of salts of active
pharmaceutical ingredients are provided. For example, the
present disclosure provides novel cocrystals of fluoxetine
HC1 and benzoic acid; fluoxetine HC1 and succinic acid; and
fluoxetine HC1 and fumaric acid. Novel forms or solid
state phases of active pharmaceutical ingredients may be
prepared for which there are no known polymorphs, solvates
or hydrates, or where such polymorphs, solvates or hydrates
were disfavored.
As a further aspect, a method of modifying one or more
physical properties of a drug formulation or drug
composition which comprises an API, the method comprising
forming a series of cocrystals of the API with a plurality
of guests. The method may further comprise measuring a
physical property of the cocrystal and/or adjusting the
drug formulation or drug composition.
As yet another aspect, an improved method for
screening or selecting the optimal solid state phase for
active agents, particularly active pharmaceutical
ingredients, and salts thereof, is provided. The screening
4
CA 02798615 2012-12-07
method comprises crystallizing or attempting to crystallize
the free base of the active agent, a chloride salt of the
active agent, and optionally other salts of the active
agent, and cocrystallizing or attempting to cocrystallize
the free base of the active agent, a chloride salt of the
active agent, sand optionally other salts of the active
agent. The method may further comprise evaluating one or
more properties of the solid forms, such as one or more
physical properties.
BRIEF DESCRIPTION OF THE DRAWINGS
FIGS. 1(a) and (b) illustrate a crystal structure of
an active pharmaceutical ingredient and a cocrystal
structure containing the same API with a guest molecule.
FIGS. 2(a) and (b) are drawings of two-dimensional and
three-dimensional models of a cocrystal of fluoxetine HC1
and benzoic acid (1:1).
FIGS. 3(a) and (b) are drawings of two-dimensional and
three-dimensional models of a cocrystal of fluoxetine HCl
and succinic acid (2:1).
FIGS. 4(a) and (b) show a two-dimensional drawing of
nabumetone and 2,3-naphthalenediol and a three-dimensional
model of a cocrystal of nabumetone and 2,3-naphthalenediol
(1:1).
FIG. 5 shows examples of general classes of guests.
DETAILED DESCRIPTION OF THE INVENTION
The present disclosure provides a way of investigating
cocrystals and a way of creating new solid state phases in
which one or more active agents are cocrystallized with a
guest.' By cocrystallizing an active agent with a guest
such as a pharmaceutically acceptable compound, one can
CA 02798615 2012-12-07
create new solid state phases which may have improved
properties over existing solid state phases of that active
agent. For example, new drug formulations comprising
cocrystals of active pharmaceutical ingredients may have
superior properties over existing drug formulations. The
active agent and guest will vary depending on the industry.
For example, in the pharmaceutical field, the active agent
or guest may be an API, and the other component of the
cocrystal must be a pharmaceutically acceptable compound.
The present techniques are also applicable to active agents
from other fields including nutraceuticals, agricultural
chemicals, pigments, dyes, explosives, polymer additives,
lubricant additives, photographic chemicals, and structural
and electronic materials.
Broadly speaking, one aspect relates to the use of
undercoordinated counterions to facilitate
cocrystallization. While the inventor does not wish to be
bound to theory, the inventor believes excellent cocrystals
may be formed using hydrochloride salts and similar salts
which are strong hydrogen bond acceptors yet contain
relatively undercoordinated ions. "Undercoordinated" in
this case refers to ions, for example a chloride ion, that
are able to form a number of strong hydrogen bonds. An
undercoordinated counterion may have hydrogen bonds within
a crystal of that salt, but it could form additional
hydrogen bonds in a cocrystal and/or form relatively
stronger hydrogen bonds in a cocrystal with a guest. An
ion is "undercoordinated" when the system is limited in the
number of hydrogen bond donors that are available and
bonded to the ion. In these cases, the extra hydrogen bond
acceptor sites are typically filled by weakly interacting
donors such as C-H groups. Chloride ions are strong
6
CA 02798615 2012-12-07
hydrogen bond acceptors in a crystal structure. In a
crystal structure such as fluoxetine hydrochloride, the
chloride ion coordinates to the two strong hydrogen bond
donors available in the system, and the chloride ion also
has three weaker CH-Cl interactions resulting in a pseudo-
octahedral coordination environment. There is an
opportunity for banding with these coordination sites, by
displacing the weak CH donors that the chloride has
recruited to fill its coordination sphere with somewhat
stronger hydrogen bond donors from a guest such as benzoic
acid, succinic acid, fumaric acid, or another carboxylic
acid.
It is useful in forming cocrystals to recognize that
relatively weak interactions may be replaced by stronger
interactions, even though those stronger interactions may
be relatively weak themselves, compared to other
interactions. For example, an undercoordinated chloride
may have one strong hydrogen bond donor and several weak
hydrogen bond donors or two strong hydrogen bond donors and
several weak hydrogen bond donors. In a cocrystal, weaker
interactions may be replaced by stronger interactions,
although those stronger interactions may still be weaker
than the strong interactions (charge-assisted hydrogen
bonds) present in fluoxetine HC1 crystals. The strongest
interactions involving chloride ions in crystal structures
of organic salts are the charge assisted hydrogen bonds
that invariably form between the protonated nitrogen base
and the chloride ion. The strongest interactions between
neutral molecular groups and the chloride ion involve acids
and the chloride ion. Carboxylic acids, for instance, have
strong interactions with chloride ions. It can be seen
that a combination of carboxylic acids and hydrochloride
7
CA 02798615 2012-12-07
salts of nitrogen containing bases are especially well
suited to cocrystal formation (as demonstrated by the
examples included). Furthermore, it can be anticipated
that different combinations of these elements could lead to
other cocrystals. For example, the active molecule of
interest may contain either the neutral carboxylic acid
moiety or the protonated nitrogen. The potential exists to
cocrystallize an API having a neutral carboxylic acid
moiety with a guest that is a hydrochloride salt of a
nitrogen-containing organic base.
It is further contemplated that the nature of the
protonated nitrogen base will affect the potential for
cocrystallization. Numerous strong hydrogen bond donor
groups will compete with the carboxylic acid guest for the
open acceptor sites on the chloride ion. In order to favor
cocrystal formation, the nitrogen base is preferably a
tertiary amine because this presents a situation where only
one strong charged hydrogen bond donor exists and thus will
only occupy one site on the chloride acceptor.
Additionally, systems that have only this one tertiary
amine and no other strong donors present an especially
favorable system for potential cocrystallization.
Protonated secondary amines with two N-H donor groups are
also favored, although protonated primary amines may also
be used. Special consideration must be taken for systems
with additional strong hydrogen bond donor and acceptor
sites in order to determine the potential for
cocrystallization and the optimal guest molecule type for
cocrystallization. The potential for cocrystallization
involving a carboxylic acid and a hydrochloride salt may be
reduced as the number of available strong donors in the
system is increased. Additional guidance as to evaluating
8
CA 02798615 2012-12-07
undercoordination may be found in the inventor's prior work,
particularly in its discussion of nonbonded
motifs: Scott L. Childs, "Nonbonded Interactions In
Molecular Crystal Structures", Emory Univ., USA, available
from UMI, Order No. DA3009424 (288 pp.), Dissertation
Abstract Int. Ref. B2001, 62(3), 1394. In some
circumstances, the undercoordination can be determined by
measuring distances, comparing profiles in the Cambridge
Structural Database, measuring the pKa of the donors and
acceptors, or evaluating the ratio of strong hydrogen bond
donors to available acceptors. Other crystal engineering
theories may also be used.
The formation of cocrystals is very unpredictable. It
is difficult to foresee structural changes as a function of
changes in molecular substitution patterns or in molecular
geometry. However, the present disclosure provides greater
predictability and better probability of success in
designing and forming cocrystals.
The present techniques may be employed to generate a
wide variety of cocrystals of active agents and guests.
For example, the present techniques may be used to generate
cocrystals of a salt of an active agent, such as a salt of
an active pharmaceutical ingredient, with a neutral guest.
Alternatively, a cocrystal of a neutral or zwitterionic
active agent (or a salt of an active agent) may be
generated with a guest salt, which includes a positive ion
and a negative ion of its own. Where the active agent is
provided in a salt, it may be positively or negatively
charged and have a negative or positive counterion. As an
example, for fluoxetine HCI, the active agent fluoxetine is
positively charged by virtue of accepting a proton from HC1 '
9
CA 02798615 2012-12-07
to form a protonated amine, and chloride is present as a
negative counterion. Furthermore, some of the present
methods may be employed with a neutral or zwitterionic
active agent to forma cocrystal with a neutral guest or
ionic guest.
The present techniques provide an opportunity to
create a stable solid state phase of a hydrochloride salt
of an API that was previously found to have properties that
were unsuitable for development. Opportunities for
continued development in such a situation have often relied
on the fortuitous formation of a stable hydrate or solvate,
but the present techniques present the ability to
systematically examine alternative formulations of the
hydrochloride salt by cocrystallizing the hydrochloride
salt of the API with appropriate guest molecules.
Cocrystallization may be an attractive technique for
salts of APIs that have been rejected due to problems
relating to physical properties. Since cocrystals may have
different physical properties than the individual
components, APIs with unfavorable physical properties can
be cocrystallized with suitable guest molecules and the
physical properties of the resulting crystalline solids can
be evaluated.
The cocrystals of fluoxetine HC1 provide examples of
the modification of a physical property (solubility) of an
API salt. Cocrystals of fluoxetine HC1:benzoic acid are
less soluble and have a lower dissolution rate than
crystals of fluoxetine HC1, while cocrystals of fluoxetine
HC1:succinic acid are more soluble and have a faster
dissolution rate than crystals of fluoxetine HC1.
Other physical properties of APIs or their salts that
may be modified by forming a cocrystal include: melting
CA 02798615 2012-12-07
point, density, hygroscopicity, crystal morphology, loading
volume, compressibility, and shelf life. Furthermore,
other properties such as bioavailability, toxicity, taste,
physical stability, chemical stability, production costs,
and manufacturing method may be modified by the use of the
present cocrystallization techniques.
An active agent can be screened for possible
cocrystals where polymorphic forms, hydrates or solvates
are especially problematic. A neutral compound that can
only be isolated as amorphous material could be
cocrystallized. Forming a cocrystal may up-grade the
performance of a drug formulation of an active
pharmaceutical ingredient by changing physical properties.
Some APIs are problematic during wet granulation and
compression stages. A bioequivalent cocrystal could
rectify this problem.
A cocrystal can be used to isolate or purify a compound
during manufacturing. If it is desirable to identify all
of the solid state phases of an active pharmaceutical
ingredient, then cocrystallization may be particularly
desirable.
The present techniques provide new methods of
developing and screening active pharmaceutical ingredients.
Non-toxic cocrystalline forms of neutral active
pharmaceutical ingredients may be prepared, screened,
tested, and commercialized. Screening based on-cocrystal
formation is equivalent in many respects to a salt-screen
for neutral APIs. Furthermore, new types of HC1 salt
structures may be prepared. The properties of
hydrochloride salts can be tuned and perfected. New,
unique, stable, and marketable phase of hydrochloride salts
11
CA 02798615 2012-12-07
may be prepared. One can choose whether to make the
formulation more soluble or less soluble.
As another aspect, the present techniques may also be
used to remove or reduce the water of hydration, and/or to
prepare a cocrystal substantially free of water of
hydration. A hydrate may be viewed as a cocrystal having
water as the guest. Water and guest acids perform a
similar role in the stabilization of the crystal structure.
In fact, about 28% of the hydrochloride salts of API in the
Cambridge Structure Database are hydrates, compared to
about 8% of all other organic structures. This indicates
an affinity for hydration. The present techniques both
capitalize and rectify this affinity, in that an affinity
for cocrystallization (as evidence by hydration) is likely
indicated, and this affinity for cocrystallization may be
employed for the formation of cocrystals with a suitable
guest, such asan acid, for example a carboxylic acid.
Indeed, in many cocrystals, an acid may have stronger
interactions than water molecules and may displace the
water of hydration during the formation of the cocrystal.
Accordingly, the present techniques provide a method of
preparing a cocrystal from a hydrate. A hydrate of a salt
is provided, and the hydrate comprises water of hydration.
A guest is selected to coordinate with the counterion.
Preferably, the guest coordinates more strongly with,the
counterion than the solvent does. A solution, melt or
physical mixture is prepared which comprises the hydrate
and the guest. The solution or melt is subjected to a
crystallization process, or the physical mixture is
subjected to grinding, and a cocrystal comprising the salt
of the active agent and the guest is formed, and the salt
comprises the active agent and a counterion. Similarly,
12
CA 02798615 2012-12-07
the present techniques provide a method of preparing a
cocrystal from a solvate. A solvate of a salt is provided,
and the solvate comprises solvent molecules coordinated
with the salt. A guest is selected to coordinate with the
counterion. Preferably, the guest coordinates more
strongly with the counterion than the solvent does. A
solution, melt or physical mixture is prepared comprising
the solvate and the guest. The solution or melt is
subjected to a crystallization process, or the physical
mixture is subjected to grinding, and a cocrystal
comprising the salt of the active agent and the guest is
formed. The salt comprises the active agent and a
counterion.
FIGS. 2(a) and (b)-are drawings of two-dimensional and
three-dimensional models of a cocrystal of fluoxetine HC1
and benzoicacid (1:1). FIG. 2(a) shows a two-dimensional
model in which the chloride ion interacts with the
hydrogens of the amine group of fluoxetine and of the
hydroxyl group of benzoic acid. Through these
interactions, which may be characterized as hydrogen
bonding, fluoxetine hydrochloride and benzoic acid 'form a
supramolecular structure that may be the basis of a
cocrystal. FIG. 2(b) shows a three-dimensional model of
the supramolecular organization of fluoxetine hydrochloride
and benzoic acid.
FIG. 3(a) and (b) are drawings of two-dimensional and
three-dimensional models of a cocrystal of fluoxetine HC1
and succinic acid (2:1). FIG. 3(a) shows a two-dimensional
model in which the chloride ion interacts with the
hydrogens of the ammonium group of fluoxetine and of the
hydroxyl group of succinic acid. Through these
interactions, which may be characterized as hydrogen
13
CA 02798615 2012-12-07
bonding, two molecules of fluoxetine hydrochloride and one
molecule of succinic acid form a supramolecular structure
that may be the basis of a cocrystal. FIG. 3(b) shows a
three-dimensional model of the supramolecular organization
of the molecules of fluoxetine hydrochloride and succinic
acid.
FIGS. 4(a) and (b) show a two-dimensional drawing of
nabumetone and 2,3-naphthalenediol and a three-dimensional
model of a cocrystal of nabumetone and 2,3-naphthalenediol
(1:1) .
Active agent
The active agent is the molecule whose activity is
desirable or the object of interest. It is contemplated
that one or more active agents may be employed in a
cocrystal, according to any of the present techniques. For
example, where the active agent is an active pharmaceutical
ingredient, the pharmaceutical activity of the active agent
is desirable. Other active agents may be nutraceuticals,
agricultural chemicals, pigments, dyes, explosives, polymer
additives, lubricant additives, photographic chemicals, or
structural and electronic materials.
The active agent may be provided as a salt. It is
contemplated that one or more salts may be employed in a
cocrystal, according to any of the present techniques.
The salt may be prepared from the active agent or obtained
from a commercial source. Hydrochloride salts of active
pharmaceutical ingredients, especially of amine APIs, are
especially preferred in the pharmaceutical industry.
In general, it is contemplated that the present
techniques will have particularly good results as applied
to amine HCl salts as well as other ammonium salts as
14
CA 02798615 2012-12-07
described in more detail herein. In ammonium acid salts,
the active agent has at least one amine moiety which is
relatively basic (at least one relatively basic nitrogen),
and a salt is formed with an acid that -reacts with the
amine moiety. Cocrystals may be then formed between the
ammonium salts and guests which act as hydrogen-bond donors
to the salts. Cocrystals may be formed of chloride salts
of APIs, for example buspirone hydrochloride, fluoxetine
hydrochloride, and metformin hydrochloride.
The present cocrystals may comprise salts other than
chloride salts -- the hydrochloride API salts that are
listed above are only a sampling of the relevant compounds
because the starting material need not be a known
hydrochloride. Indeed, many relevant APIs are salts that
are not HCl salts because the'HCl salt was not believed to
be an appropriate material and a different salt was
commercialized instead. The present techniques may enable
one to employ an HCl salt of an API that is marketed as
another type of salt. Alternatively, it may be desirable
to employ a salt other than an HCl salt, by replacing the
HCI or by forming a salt comprising an active agent that
acts as a base with an acid other than HCl. The following
acids provide anionic counterions that would be used to
replace chlorine. These are relatively strong acids, and
include but are not limited to mineral acids, and the
carboxylic acid guest is expected to form one or'more
hydrogen bonds with a hydrogen bond acceptor on the anionic
counterion. The list is the conjugate acid that would
react with a basic active agent to form a salt:
sulfuric acid
phosphoric acid
hydrobromic acid
9.5
CA 02798615 2012-12-07
nitric acid
pyrophosphoric acid
methanesulfonic acid
thiocyanic acid
naphthalene-2-sulfonic acid
1,5--naphthalenedisulfonic acid
cyclamic acid
p-toluenesulfonic acid
maleic acid
L-aspartic acid
2-hydroxy-ethanesulfonic acid
glycerophosphoric acid
ethanesulfonic acid
hydroiodic acid
The present techniques also extend beyond salts as
starting materials and also include many weak bases that
may have been marketed as neutral forms because the known
salts did not have appropriate properties. These salts
could be revisited and attempts could be made to
cocrystallize the HC1 salt. For example, a drug
formulation marketed as a tartrate salt of an API could be
reformulated by cocrystallizing the HC1 salt of the active
molecule with an appropriate guest molecule. Thus,
cocrystallization could make a useful HC1 cocrystal out of
the API that is currently marketed as a tartrate, sulfate,
or other salt formulation. For this reason the present
disclosure includes APIs that are not HC1 salts as starting
materials.
Furthermore, the present techniques relate to salts
other than chloride salts. It is contemplated that
hydrobromide salts and sodium salts of APIs may especially
benefit from the present techniques, since they form
16
CA 02798615 2012-12-07
relatively strong nonbonded interactions. For example, the
hydrobromide salts citalopram hydrobromide and galantamine
hydrobromide are contemplated for cocrystallization with
benzoic acid, succinic acid, and other guests compatible
with hydrochloride salts.
The present techniques may be employed to form
cocrystals of sodium salts of APIs such as, for example,
naproxen sodium, tolmetin sodium, and warfarin sodium.
When a sodium salt (or other salt of an API having a
positive counterion) is employed, different guests are
expected to be suitable for cocrystallization than when a
hydrochloride salt (or other anionic salt) of an API is
employed.
Anions and Cations
As one aspect, the active agent is provided as a salt.
A salt of the active agent is formed. Alternatively or
additionally, the guest is provided as a salt or a salt of
the guest is formed. The salt may comprise the active
agent and a counterion that is either a cation or an anion.
Among the preferred cations (including cations as well as
compounds that can form cations) are aluminum, ammonium,
benzathine, calcium, diethanolamine, diethylamine,
dimeglumine, disodium, lithium, lysine, magnesium,
meglumine, potassium, sodium, and zinc. Among the
preferred anions are acetate, L-aspartate, besylate,
bicarbonate, carbonate, D-camsylate, L-camsylate, citrate,
edisylate, fumarate, gluconate, hydrobromide/bromide,
hydrochloride/chloride, D-lactate, L-lactate, DL-lactate,
D,L-malate, L-malate, mesylate, pamoate, phosphate,
succinate, sulfate, D-tartrate, L-tartrate, D,L-tartrate,
meso-tartrate, benzoate, gluceptate, D-glucuronate,
17
CA 02798615 2012-12-07
hybenzate, isethionate, malonate, rnethylsufate, 2-
napsylate, nicotinate, nitrate, orotate, stearate,
tosylate, acefyllinate, aceturate, aminosalicylate,
ascorbate, ascorbate,'borate, butyrate, camphorate,
camphocarbonate, decanoate, hexanoate, cholate, cypionate,
dichloroacetate, edentate, ethyl sulfate, furate, fusidate,
galactarate (mucate), galacturonate, gallate, gentisate,
glutamate, glutamate, glutarate, glycerophosphate,
heptanoate (enanthate), hydroxybenzoate, hippurate,
phenylpropionate, iodide, xinafoate, lactobionate, laurate,
maleate, mandelate, methane sufonate, myristate,
napadisilate, oleate, oxalate, palmitate, picrate,
pivalate, propionate, pyrophosphate, salicylate,
salicylsulfate, sulfosalicylate, sulfosalicylate, tannate,
terephthalate, thiosalicylate, tribrophenate, valerate,
valproate, adipate, 4-acetamidobenzoate, camsylate,
octanoate, estolate, esylate, glycolate, thiocyanate, and
undecylenate.
When a metal cation is employed as a counterion of the
active agent, the interaction between guest and cation is
not a hydrogen bond but rather is an intermolecular
interaction between an electron rich group such as a
carbonyl and the metal cation. This interaction is often
not as strong as a hydrogen bond, but is still a favorable
interaction and thus can contribute to the stabilization of
a cocrystal.
The HC1 salt of an active pharmaceutical ingredient is
especially preferred to create a new type of cocrystal. In
this type of solid state phase, one can cocrystallize the
HC1 salt with a neutral guest molecule. By doing this one
can create solid state phases with specific properties.
For instance one can make a solid comprising an active
is
CA 02798615 2012-12-07
pharmaceutical ingredient having greater or lesser
intrinsic solubility and/or a faster or slower dissolution
rate, depending on the guest compound that is chosen.
Guests
The guest is present in order to form the cocrystal
with the active agent. It is contemplated that one or more
guests may be employed in a cocrystal, according to any of
the present techniques. Accordingly, the guest is not
required to have an activity of its own, although it may
have some activity that does not overly derogate from the
desired activity of the active agent. In some situations,
the guest may have the same activity as or an activity
complementary to that of the active agent. The guest may
be another API. For example, some guests may facilitate
the therapeutic effect of an active pharmaceutical
ingredient. For pharmaceutical formulations, the guest may
be any pharmaceutically acceptable molecule(s) that forms a
cocrystal with the API or its salt. The RTECS database is
a useful source for toxicology information, and the GRAS
list contains about 2500 relevant compounds.
The guest may be neutral (such as benzoic acid and
succinic acid in the examples below) or ionic (such as
sodium benzoate or sodium succinate). Neutral guests are
nonionic guests. Ionic guests are compounds or complexes
having ionic bonds. FIG. 5 shows several general classes
of guests (organic bases, organic salts, alcohols &
aldehydes, amino acids, sugars, ionic inorganics, aliphatic
esters & ketones, organic acids, and aromatic esters &
ketones).
The guest may be an acid that forms hydrogen bonds
with the chloride (or other anion). For example, suitable
19
CA 02798615 2012-12-07
guests which are acids include (but not are not limited
to)
ascorbic acid
glucoheptonic acid
sebacic acid
alginic acid
cyclamic acid
ethane-l,2--disulfonic acid
2-hydroxyethanesulfonic acid
2-oxo-glutaric acid
naphthalene--1, 5--disulfonic acid
nicotinic acid
pyroglutamic acid
4-acetamidobenzoic acid
Table 1 sets forth a group of presently preferred
guests. It is contemplated that, the guests set forth in
the Table may be arranged in subgroups based upon molecular
structure and/or physiological effect. Furthermore, the
foregoing list is intended to provide a written description
of any sublist that omits one or more guests.
Table 2 sets forth another group of preferred guests.
It is contemplated that the guests set forth in the Table
may be arranged in subgroups based upon molecular structure
and/or physiological effect. Furthermore, the foregoing
list is intended to provide a written description of any
sublist that omits one or more guests.
Table 3 sets forth the group comprising molecules
believed at present to be suitable guests. It is
contemplated that the guests set forth in the Table may be
arranged in subgroups based upon molecular structure and/or
physiological effect. Furthermore, the foregoing list is
CA 02798615 2012-12-07
intended to provide a written description of any sublist
that omits one or more guests.
Ionic guests are salts themselves, and may be formed
from bases and acids prior to being used to form
cocrystals. For example, the following bases and acids may
be reacted to form ionic guests:
Bases
Ammonia
L-Arginine
Bene thamine
Benzathine
Betaine
Calcium Hydroxide
Choline
Deanol
Diethanolamine
Diethylamine
2-(Diethylamino)ethanol
2-Aminoethanol
Ethylenediamine
N-Methylglucamine
Hydrabamine
1H-Imidazole
Lysine
Magnesium Hydroxide
Morpholine
4-(2-Hydro--yethyl)Morpholine
Piperazine
Potassium Hydroxide
Pyrrolidine
1-(2-Hydroxyethyl)Pyrrolidine
Sodium Hydroxide
21
CA 02798615 2012-12-07
Triethanolamine
Troinethamine
Zinc Hydroxide
Acids
(4-)-L-Tartaric Acid
1,2,2-Trimethyl-1,3-cyclopentanedicarboxyli.c Acid
10-Undecylenic Acid
1-Hydroxy-2-naphthoic Acid
(+)-Camphor-l0-sulfonic Acid
2,5-Dihydroxybenzoic Acid
2-Furancarboxylic Acid
2-Mercaptobenzoic Acid
3-Cyclopentylpropionic Acid
3-Phenylpropionic Acid
4-Aminosalicylic Acid
4-Hydroxybenzoic Acid
Acetic Acid
Adipic Acid
alpha-Hydroxypropionic Acid
Benzenesulfonic Acid
Benzoic Acid
Carbonic Acid
Cholic Acid
Citric Acid
(-)-D-Tartaric Acid
(+)-D-Camphoric Acid
(+)-D-Malic Acid
(-+-) -L-Malic Acid
2,2-Dichloroacetic Acid
DL-10--Camphorsulfonic Acid
DL-Glutamic Acid
22
CA 02798615 2012-12-07
DL-Malic Acid
DL-Tartaric Acid
Dodecylsulfuric Acid
Ethanesulfonic Acid
Ethylenediaminetetraacetic Acid
Ethylsulfuric Acid
Fumari.c Acid
Galactaric Acid
Gallic Acid
Gluconic Acid
Glutaric Acid
Glycolic Acid
Hippuric Acid
Hydriodic Acid
Hydrobromic Acid
Hydrochloric Acid
(-)-L-Apple Acid
(+) -L--Lactic Acid
(+)-L-Tartaric Acid
D,L-Lactic Acid
Lactobionic Acid
L-Aspartic Acid
Lauric Acid
L-Glutamic Acid
Maleic Acid
(-)-L-Malic Acid
Malonic Acid
D,L-Mandelic Acid
Methanesulfonic Acid
Naphthalene--2-sulfonic acid
n-Butyric Acid
n-Decanoic Acid
23
CA 02798615 2012-12-07
n-Hexanoic Acid
Nitric acid
n-Tetradecanoic Acid
Octanoic Acid
Oleic Acid
Orotic Acid
Orthoboric Acid
Oxalic Acid
4-Acetamidobenzoic Acid
Palmitic Acid
Pamoic Acid
Phosphoric Acid
Picric Acid
Pivalic Acid
Propionic Acid
p-Toluenesulfonic Acid
Pyrophosphoric Acid
Salicylic Acid
Stearic Acid
Succinic Acid
Sulfosalicylic Acid
Sulfuric Acid
Terephthalic Acid
Thiocyanic Acid
Valeric Acid
Valproic Acid
Typically, suitable guests will have complementary
ability to noncoval ently'bond to the active agent or its
salt, for example the ability to form hydrogen bonds with
the active agent or its salt. Suitable guests for active
agents having negative counterions include, but are not
limited to, compounds having alcohol, ketone, ester, and/or
24
CA 02798615 2012-12-07
carboxylic acid functionalities. Suitable guests may
include organic acids, organic bases, organic salts,
alcohols, aldehydes, amino acids, sugars, ionic inorganic
compounds, aliphatic esters and ketones, and aromatic
esters and ketones.
Among the presently preferred neutral guests are those
which are not liquids at room temperature. Also among the
presently preferred neutral guests are carboxylic acids
having at least three carbon atoms, alternatively at least
four carbon atoms, and which do not form solvates. For
example, if the following acids were combined with active
agents, the combination would more properly be considered a
solvate than a cocrystal: acetic acid, propionic acid, and
butyric acid. However, in certain embodiments of the
present invention (for example, in certain cocrystals,
cocrystallization methods, and screening methods), the use
of solvents and solvates may still be desirable, and the
use of solvents and solvates is not excluded from the scope
of any cocrystal or method except where explicitly stated.
Detection of Cocrtals
Cocrystals may be detected by x-ray diffraction
analysis or other suitable techniques. The observation of
physical properties of a solid (particularly its melting
point) which differ from the physical properties of the
starting materials and the polymorphs and/or solvates
and/or hydrates of the starting materials, is an indicator
that a cocrystal has been formed.
A method of crystal engineering is described. An
active pharmaceutical ingredient such as fluoxetine
CA 02798615 2012-12-07
hydrochloride is recognized as possessing a strong hydrogen
bond acceptor. The API is screened against a library of
strong hydrogen bond donors or other possible guest
compounds. Such a library is selected and ordered based
upon nontoxicity, physical property, and the availability
and geometric arrangement of hydrogen bond donors that are
complementary to the API.
The results from a cocrystal screen of fluoxetine
hydrochloride demonstrate a new class of cocrystal that is
broadly applicable to a wide variety of hydrochloride salts
of APIs. This new approach is a general method that allows
creation of cocrystals starting with the hydrochloride salt
of the API. Starting with the hydrochloride retains the
advantages of the salt, yet one is still able to use the
cocrystal method to alter the physical properties of the
resulting solid by adding guest molecules.
Example 1: Cocrystallization of fluoxetine HC1 and
benzoic acid
Cocrystals of fluoxetine HC1:benzoic acid were formed
using the following procedures. In one preparation, a 505
mg sample of fluoxetine HC1 and 178 mg of benzoic acid were
dissolved with heating in 5 mL of acetonitrile. The
solution was allowed to crystallize in a small
crystallization dish. Well-formed crystalline material
formed within 7 minutes. This material was isolated on
filter paper and dried in the air to yield 546 (80%) of
fluoxetine HC1:ben.zoic acid (1:1) cocrystal.
In another preparation, a 5.00g sample of fluoxetine
HC1 and 1.76 g of benzoic acid were dissolved in 50 mL of
acetonitrile with heating. The solution was allowed to
crystallize in a large evaporating dish. The resulting
26
CA 02798615 2012-12-07
solid was isolated on filter paper and dried in the air to
yield 5.40 g (92a) of fluoxetine HC1:benzoic acid (1:1)
cocrystal.
The cocrystal had a relatively slow dissolution rate
and lower water solubility. The measured melting point was
1342C +/- 22C for the cocrystal. The cocrystal is expected
to have a good toxicology profile, since benzoic acid is
known to be safe and appears on the GRAS list from the U.S.
Food and Drug Administration.
The resulting cocrystal is a ternary system comprising
the protonated API base, the chloride ion, and the neutral
guest molecule.
The present inventor believes there are no known
solvates or hydrates of fluoxetine hydrochloride. Thus,
the formation of a cocrystal of fluoxetine hydrochloride
constitutes a surprising achievement and provides a unique
composition.
Example 2: Cocrystallization of fluoxetine HC1 and
succinic acid
Cocrystals of fluoxetine HC1 and succinic acid were
prepared as follows. In one preparation, a 458 mg sample
of fluoxetine HC1 was dissolved in 8 mL of acetonitrile by
heating the solution gently. A 78 mg sample of succinic
acid was added to the warm solution and dissolved. The
solution was allowed to evaporate rapidly in a
crystallization dish. Well-formed crystals as blocks
formed as the solvent evaporated over 8 minutes. The
product was collected on. filter paper and dried to yield
401 mg of fluoxetine HCl:succinic acid (2:1) cocrystal (75%
yield).
27
CA 02798615 2012-12-07
In another preparation, a 5.00 g sample of fluoxetine
HCl and 0.85 g of succinic acid were dissolved in
acetonitrile with heating. The solution was allowed to
crystallize in an open evaporating dish over a 15 minute
period. The solid material was isolated on filter paper
and dried to yield 5.40 g (92% yield.) of fluoxetine
HC1:succinic acid (2:1) cocrystal.
The measured melting points were 1582C for fluoxetine
HC1, 1842C for succinic acid, and 1372C for the cocrystal.
The cocrystal is expected to have a good toxicology
profile, since succinic acid is known to be safe and
appears on the Generally Recognized As Safe ("GRAS") list
from the U.S. Food and Drug Administration.
Example 3: Cocrystallization of nabuietone and 2,3-
naphthalenediol
As a demonstrative example, a cocrystal comprising a
neutral API is described in this example. Cocrystals of
nabumetone (a neutral API) and 2,3-naphthalenediol were
prepared as follows. A 4.01 g sample of 2,3-
naphthalenediol and 5.7 g of nabumetone were dissolved in
50 mL of nitromethane with heating. A solid was formed as
the solution cooled and was allowed to stand overnight.
The solid was filtered from the remaining solvent and dried
in the air to yield 6.61 g (68%) of nabumetone:2,3-
naphthalenediol (1:1) cocrystal.
The resulting cocrystal had a 1:1 molar ratio of
nabumetone to 2,3-naphthalienediol. The measured melting
points were 8020 for nabumetone, 16290 for 2,3-
naphthalienediol, and 98 -C for the cocrystal. The
cocrystal is expected to have a relatively poor toxicology
profile. However, this example demonstrates one basis for
28
CA 02798615 2012-12-07
the selection of guest molecules: molecular structural
similarities. In this case the molecular recognition of
the naphthalene moieties of the API and the guest
contribute to the stability of the cocrystal. In addition,
the stronger alcohol to ketone hydrogen bonds formed by the
cocrystal contribute to the stability of the cocrystal.
The only hydrogen bond donors available in the API crystal
structure are weak C-H groups. The stronger hydrogen bond
donors on the guest molecule are able to form stronger
intermolecular interaction between the API and guest,
compared to the interactions between molecules of the API.
Example 4: Crystal Structure Analysis of fluoxetine
HC1:benzoic acid cocrystal (1:1)
A suitable cocrystal of fluoxetine HC1:benzoic acid
(1:1) was coated with Paratone N oil, suspended in a small
fiber loop and placed in a cooled nitrogen gas stream at
100 K on a Bruker D8 SMART APEX CCD sealed tube
diffractometer with graphite monochromated MoKa (0.71073A)
radiation. Data were measured using a series of
combinations of phi and omega scans with 10 second frame
exposures and 0.3 frame widths. Data collection, indexing
and initial cell refinements were all carried out using
SMART software (SMART Version 5.624, 2000, Bruker AXS,
Inc., Analytical X-ray Systems, 5465 East Cheryl Parkway,
Madison WI 53711-5373) . Frame integration and final cell
refinements were done using SAINT software (SAINT Version
6.02, 2000, Bruker AXS, Inc., Analytical X-ray Systems,
5465 East Cheryl Parkway, Madison WI 53711-5373). The
final cell parameters were determined from least-squares
refinement on 5435 reflections. The SADABS program was used
29
CA 02798615 2012-12-07
to carry out absorption corrections (SADABS Version 2.03,
2001, George Sheldrick, University of Gottingen).
The structure was solved using Direct methods and
difference Fourier techniques (SHELXTL V5.10, 2000, Bruker
AXS, Inc., Analytical X-ray Systems, 5465 East Cheryl
Parkway, Madison WI 53711-5373). Hydrogen atoms were
placed their expected chemical positions using the HFIX
command and were included in the final cycles of least
squares with isotropic Ujj 's related to the atom's ridden
upon. The C-H distances were fixed at 0.93 A(aromatic and
amide), 0.98 A (methine), 0.97 A (methylene), or 0.96 A
(methyl). All non-hydrogen atoms were refined
anisotropically. Scattering factors and anomalous
dispersion corrections are taken from A. J. C. Wilson (ed),
International Tables for X-ray Crystallography, Volume C.
Kynoch, Academic Publishers, Dordrecht, 1992, Tables
6.1.1.4 (pp. 500-502) and 4.2.6.8 (pp. 219-222). the
International Tables for X-ray Crystallography. Structure
solution, refinement, graphics and generation of
publication materials were performed by using SHELXTL,
V5.10 software. Additional details of data collection and
structure refinement are given in Table 4 which follows.
Example 5: Crystal Structure Analysis of fluoxetina
HC1:succinic acid cocrystal (2:1)
A suitable cocrystal of fluoxetine HC1-succinic acid
(2:1) was coated with Paratone N oil, suspended in a small
fiber loop and placed in a cooled nitrogen gas stream at
100 K on a Bruker D8 SMART APEX CCD sealed tube
diffractometer with graphite monochromated MoKa. (0.710731)
radiation. Data were measured using a series of
CA 02798615 2012-12-07
combinations of phi and omega scans with 10 second frame
exposures and 0.3' frame widths. Data collection, indexing
and initial cell refinements were all carried out using
SMART software (SMART Version 5.624, 2000, Bruker AXS,
Inc., Analytical X-ray Systems, 5465 East Cheryl Parkway,
Madison WI 53711-5373) . Frame integration and final cell
refinements were done using SAINT software (SAINT Version
6.02, 2000, Bruker AXS, Inc., Analytical X-ray Systems,
5465 East Cheryl Parkway, Madison. WI 53711-5373). The
final cell parameters were determined from least-squares
refinement on 5435 reflections. The SADABS program was
used to carry out absorption corrections (SADABS Version
2.03, 2001, George Sheldrick, University of Gottingen).
The structure was solved using Direct methods and
difference Fourier techniques (SHELXTL V5.10, 2000, Bruker
AXS, Inc., Analytical X-ray Systems, 5465 East Cheryl
Parkway, Madison WI 53711-5373). Hydrogen atoms were
placed their expected chemical positions using the HFIX
command and were included in the final cycles of least
squares with isotropic Ujj 's related to the atom's ridden
upon. The C-H distances were fixed at 0.93 .(aromatic and
amide), 0.98 A (methine), 0.97 A (methylene), or 0.96 A
(methyl) . All non-hydrogen atoms were refined
anisotropically. Scattering factors and anomalous
dispersion corrections are taken from A. J. C. Wilson (ed),
International Tables for X-ray Crystallography, -Volume C.
Kynoch, Academic Publishers, Dordrecht, 1992, Tables
6.1.1.4 (pp. 500-502) and 4.2.6.8 (pp. 219-222). Structure
solution, refinement, graphics and generation of
publication materials were performed by using SHELXTL,
V5.10 software. Additional details of data collection and
structure refinement are given in Table 5 which follows.
31
CA 02798615 2012-12-07
Example 6: Crystal Structure Analysis of nabumetone:
2,3-naphthalenediol cocrystal (1:1)
A suitable cocrystal of nabumetone: 2,3-naphthalenediol
(1:1) was coated with Paratone N oil, suspended in a small
fiber loop and placed in a cooled nitrogen gas stream at
100 K on a Bruker D8 SMART 1000 CCD sealed tube
diffractometer with graphite monochromated CuKa (1.541.78
radiation. Data were measured using a series of
combinations of phi and omega scans with 10 second frame
exposures and 0.3 frame widths. Data collection, indexing
and initial cell refinements were all carried out using
SMART software (SMART Version 5.55, 2000, Bruker AXS, Inc.,
Analytical X-ray Systems, 5465 East Cheryl Parkway, Madison
WI 53711-5373). Frame integration and final cell
refinements were done using SAINT software (SAINT Version
6.02, 1999, Bruker AXS, Inc., Analytical X-ray Systems,
5465 East Cheryl Parkway, Madison WI 53711-5373). The
final cell parameters were determined from least-squares
refinement on 2869 reflections. ,
The structure was solved using Direct methods and
difference Fourier techniques (SHELXTL V5.10, 1997, Bruker
AXS, Inc., Analytical X-ray Systems, 5465 East Cheryl
Parkway, Madison WI 53711-5373). Hydrogen atoms were
placed their expected chemical positions using the HFIX
command and were included in the final cycles of least
squares with isotropic Ulf 's related to the atom's ridden
upon. The C-H distances were fixed at 0.93 A(aromatic and
amide), 0.98 A (methine), 0.97 A (methylene), or 0.96 A
(methyl). All non-hydrogen atoms were refined
anisotropically. Scattering factors and anomalous
32
CA 02798615 2012-12-07
dispersion corrections are taken from A. J. C. Wilson (ed),
International Tables for X-ray Crystallography, Volume C.
Kynoch, Academic Publishers, Dordrecht, 1992, Tables
6.1.1.4 (pp. 500-502) and 4.2.6.8 (pp. 219-222). Structure
solution, refinement, graphics and generation of
publication materials were performed by using SHELXTL,
V5.10 software. Additional details of data collection and
structure refinement are given in Table 6 which follows.
Example 7: Cocrystals of Active Agent salts and guests
The experiments of Examples 1 and 2 are repeated, using
conditions similar to those of those previous Examples, with
each possible combination of the salts of active agents and
guests identified earlier in this disclosure. Cocrystals are
formed which have utility according to the known activity of
the active agent.
Example 6: Cocrystals of Active Agents and guest salts
The experiments of Example 3 are repeated, using
conditions similar to those of those previous Examples, with
each possible combination of the neutral or zwitterionic
active agents and guest salts identified earlier in this
disclosure. Cocrystals are formed which have utility
according to the known activity of the active agent.
Example 9: Cocrystallization of fluoxetine HCI and
fumaric acid
Cocrystals of fluoxetine HC1 and succinic acid were
prepared as follows. A 6.00 g sample of fluoxetine HC1 and
1.01 g of fumaric acid were dissolved in 20 mL of ethanol
with heating. The solution was filtered through a 0.2 pm
nylon filter, concentrated to a volume of 8 mL, and cooled
33
CA 02798615 2012-12-07
in an ice bath for 6 hours. The solid material was
isolated on filter paper and allowed to dry in the air to
give 5.74 g (82% yield) of fluoxetine HC1:fumaric acid
(2:1) cocrystal. The measured melting points were 1582C
for fluoxetine HC1, >3002C (decomposes) for fumaric acid.,
and 1642C for the cocrystal. The cocrystal is expected to
have a good toxicology profile, since fumaric acid is known
to be safe and appears on the Generally Recognized As Safe
("GRAS") list from the U.S. Food and Drug Administration.
Example 10: Crystal Structure Analysis of fluoxetine
HC1:fumaric acid cocrystal (2:1)
A suitable cocrystal of fluoxetine HC1 and fumaric
acid was coated with Paratone N oil, suspended in a small
fiber loop and placed in a cooled nitrogen gas stream at
100 K on a Bruker D8 SMART 1000 CCD sealed tube
diffractometer with graphite monochromated CuKa (1.541.78.)
radiation. Data were measured using a series of
combinations of phi and omega scans with 10 second frame
exposures and 0.3 frame widths. Data collection, indexing
and initial cell refinements were all carried out using
SMART software (SMART Version 5.55, 2000, Bruker AXS, Inc.,
Analytical X-ray Systems, 5465 East Cheryl Parkway, Madison
WI 53711-5373). Frame integration and final cell
refinements were done using SAINT software (SAINT Version
6.02, 1.999, Bruker AXS, Inc., Analytical X-ray Systems,
5465 East Cheryl Parkway, Madison WI 53711-5373). The
final cell. parameters were determined from least-squares
refinement on 5625 reflections.
The structure was solved using Direct methods and
difference Fourier techniques (SHELXTL V5.10, 1997, Bruker.
AXS, Inc., Analytical X-ray Systems, 5465 East Cheryl
34
CA 02798615 2012-12-07
Parkway, Madison WI 53711-5373). All the hydrogen atoms
were located from difference Fouriers and included in the
final cycles of least squares with isotropic Uij `s. All
non-hydrogen atoms were refined anisotropically. Scattering
factors and anomalous dispersion corrections are taken from
A. J. C. Wilson (ed), International Tables for X-ray
Crystallography, Volume C. Kynoch, Academic Publishers,
Dordrecht, 1992, Tables 6.1.1.4 (pp. 500-502) and 4.2.6.8
(pp. 219-222). Structure solution, refinement, graphics
and generation of publication materials were performed by
using SHELXTL, V5.10 software. Additional details of data
collection and structure refinement are given in Table G
which follows.
While the present invention has been described and
illustrated by reference to particular embodiments, it will
be appreciated by those of ordinary skill in the art that
the invention lends itself to many different variations not
illustrated herein. For these reasons, then, reference
should be made solely to the appended claims for purposes
of determining the true scope of the present invention.
