Anda di halaman 1dari 29

Accepted Manuscript

Synthesis of a glibenclamide cocrystal: full spectroscopic and thermal characterization

Silvério Ferreira Silva Filho, Andreia Cardoso Pereira, Jorge M.G. Sarraguça, Mafalda
C. Sarraguça, João Lopes, Pedro Freitas Filho, Adenilson Oliveira dos Santos, Paulo
Roberto da Silva Ribeiro
PII: S0022-3549(18)30078-9
DOI: 10.1016/j.xphs.2018.01.029
Reference: XPHS 1075

To appear in: Journal of Pharmaceutical Sciences

Received Date: 18 September 2017


Revised Date: 19 December 2017
Accepted Date: 8 January 2018

Please cite this article as: Silva Filho SF, Pereira AC, Sarraguça JMG, Sarraguça MC, Lopes J, Filho
PF, dos Santos AO, da Silva Ribeiro PR, Synthesis of a glibenclamide cocrystal: full spectroscopic and
thermal characterization, Journal of Pharmaceutical Sciences (2018), doi: 10.1016/j.xphs.2018.01.029.

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT

Synthesis of a glibenclamide cocrystal: full

spectroscopic and thermal characterization


1 1 1
Silvério Ferreira Silva Filho , Andreia Cardoso Pereira , Jorge M. G. Sarraguça , Mafalda C.
2 3 1 1
Sarraguça, João Lopes , Pedro Freitas Filho , Adenilson Oliveira dos Santos , Paulo Roberto
1*
da Silva Ribeiro

PT
1
Centro de Ciências Sociais, Saúde e Tecnologia (CCSST), Universidade Federal do
Maranhão (UFMA), 65.900-410 Imperatriz, Maranhão, Brazil
2
LAQV/REQUIMTE, Laboratório de Química Aplicada, Faculdade de Farmácia, Universidade
do Porto, Portugal

RI
3
Research Institute for Medicines (iMed.ULisboa), Faculdade de Farmácia, Universidade de
Lisboa, 1649-003 Lisboa, Portugal

SC
E-mail: pauloufma@ufma.br

U
AN
M
D
TE

Corresponding author:
Paulo Roberto da Silva Ribeiro
Address for correspondence: Centro de Ciências Sociais, Saúde e Tecnologia (CCSST),
Universidade Federal do Maranhão (UFMA), 65.900-410 Imperatriz, Maranhão, Brazil
E-mail: pauloufma@ufma.br
EP

Phone: +55 (99)3529-6068


Fax: +55 (99)3529-6001
C
AC

1
ACCEPTED MANUSCRIPT

Abstract

A cocrystal of glibenclamide (GLB), an antidiabetic drug classified as type II compound

according to the Biopharmaceutics Classification System (BCS), has been synthesised using

tromethamine (TRIS) as coformer in 1:1 molar ratio, by slow solvent evaporation

cocrystalization. The cocrystal obtained was characterized by X-ray powder diffraction (XRPD),

PT
differential scanning calorimetry (DSC), Raman, mid infrared (MIR), and near-infrared (NIR)

spectroscopy. The results consistently show the formation of a cocrystal between API and

RI
conformer with the synthons corresponding to hydrogen bonding between hydrogen in amines

SC
of tromethamine and carbonyl and sulfonyl groups in glibenclamide.

U
Keywords: co-crystals; FTIR; calorimetry (DSC); X-ray powder diffractometry; Near-infrared
AN
spectroscopy; Raman spectroscopy;
M
D
TE
C EP
AC

2
ACCEPTED MANUSCRIPT

1 Introduction

The interest in the development of cocrystals by the pharmaceutical industry has gained a large

importance in the last few years, in the pursuit for enhanced physico-chemical properties and

PT
therapeutical performance of drugs. The synthesis of cocrystals is a very promising strategy for

increasing solubility, stability, dissolution rate and bioavailability for a large number of active

RI
1-8
pharmaceutical ingredients (API) . A cocrystal can be defined as a crystalline form of a
9
mixture of different molecules. In the simplest form, an API and a conformer , that are solid at

SC
room conditions when isolated, and, when mixed at an exact stoichiometric ratio, form
6 10,11
noncovalent bonds , usually hydrogen-bonds . The new crystalline structure is

U
12
characterized by a crystallographic lattice different from that of the original compounds .
AN
Glibenclamide (GLB),(5-chloro-N-[2-[4-(cyclohexylcarbamoylsulfamoyl) phenyl] ethyl]-2-

methoxybenzamide), [Figure 1a)] is an antidiabetic drug used for control of glicemia in type 2
13
diabetis patients. According to the Biopharmaceutics Classification System (BCS) ,
M

-1
glibenclamide is a type II compound, showing low solubility (<0,004 mg mL at 37°C and
D

14-17
neutral pH), and high membrane permeability . The low solubility may represent an

undesired characteristic, influencing the dissolution profile and consequently the drug
TE

13,18-23
bioavailability . It is usually described as an odorless white crystalline powder with a
24,25
melting temperature in the range Tfus=169-174ºC , although thermal studies have observed
EP

that crystallization, depending on the solvent used, may result in different polymorphic and
26-29
pseudo polymorphic forms, which may show different solubility and melting temperature .
C

The attempts for increasing solubility of GLB have spanned a set of different strategies such as
AC

molecular dispersion, micronization, inclusion complexes with cyclodextrin and use of


28-31
polymorphs .

Tromethamine (TRIS), (2-amino-2-(hydroxymethyl)propane-1,3-diol), [Figure 1b)] is a BCS

class I type compound, presenting a high solubility in water, described as a white (or almost
32
white) crystalline powder . With pKa=8.06 it presents an effective pH range that is compatible
33,34
with physiological pH . Thermodynamically, the compound is characterized by a melting

range Tfus=168-174 °C, and occurs in two forms, with a solid -solid transition occurring at

3
ACCEPTED MANUSCRIPT
T=134.3 °C, corresponding to a plastic deformation from a crystalline (orthorhombic) to a body
35
centered cubic structure . This transition renders the compound with possible applications as a
36
material for thermal energy storage in solar cell systems . This compound has been tested as
37
a possible coformer for indomethacin cocrystals , although no cocrystal was obtained, and for
38
artemisinin , an antimalaric drug, with a cocrystal being obtained resorting to a

mechanochemical method.

PT
39
McMahon et al have reported in a patent the synthesis of a GLB and TRIS cocrystal . The

novelty of this work is focused on the extensive vibrational spectroscopic characterization of the

RI
cocrystal between GLB and TRIS obtained by slow solvent evaporation. The GLB:TRIS

SC
cocrystal was characterized by X-ray powder diffraction (XRPD), differential scanning

calorimetry (DSC), Raman spectroscopy, mid infrared (MIR) spectroscopy, and near-infrared

(NIR) spectroscopy. The different vibrational spectroscopic techniques employed in the study

U
yielded consistent results, from where the most probable synthons responsible for the cocrystal
AN
formation were determined.
M

2 Experimental methods
D

2.1 Synthesis
TE

The GLB:TRIS (1:1) cocrystals (Figure 2) were synthesized by dissolving 59.07mg of

tromethamine (Sigma-Aldrich, ≥ 99.8% pure) and 240.92 mg of glibenclamide (Pharmanostra,


EP

≥99.8% pure) in 150 ml of methanol (Sigma-Aldrich, ≥99.8% pure). Dissolution was obtained

after 15 minutes under stirring. The solution was filtered through a 28 µm paper filter. Solvent
C

evaporation and cocrystallization occurred in a vessel covered with a PVC film with randomly
AC

distributed needle diameter sized holes, to slow down evaporation rate. The temperature at

which evaporation occurred was controlled (T=8.0±1.0 °C) in a Biothec BT 60 incubator, for 4-5

days. The product was dried at 50 °C for 8 hours, a ssuring total absence of solvent.

4
ACCEPTED MANUSCRIPT
2.2 Characterization

Pure compounds and final products of cocrystalization were characterized by X-ray powder

diffraction (XRPD), differential scanning calorimetry (DSC), Raman, mid-infrared (MIR), and

near-infrared (NIR) spectroscopy.

The DSC measurements were carried out on a thermal analyzer (DSC 60, Shimadzu

PT
Instruments, Columbia, USA). The samples were subjected to unique heating cycles with a
-1
temperature rate of 2 ºC min , in the 25-200ºC temperature range, where thermal stability was

RI
-1
verified, in a dry air atmosphere (50 ml min flow rate). Thermal stability was determined by

thermogravimetry analysis prior to DSC analysis (data not shown). Samples were enclosed in

SC
an aluminum crucible with a closed lit. All measurements were performed at atmospheric

pressure, using masses of samples in the range of 2 to 3 mg. An empty crucible with closed lit

U
was used as reference. Instrument calibration was performed using a metallic Indium standard
AN
-1
sample (Tfus= 156.4 °C; ∆Hfus = 28.5 J g ), 99.99% pure. Correction factors were calculated

according to manufacturer's specifications.


M

X-ray diffraction analyses were performed on powders obtained after submitting the co-

crystals to manual grinding in a mortar with an agate pestle. Measurements were performed
D

using a Rigaku Miniflex II (the Woodlands, Texas, USA) X-ray diffractometer, with a Cu Kα with
TE

λ = 1.5418 Å wavelength, (at 40 kV and 30 mA). The diffraction patterns were collected in the 2º

to 45º (2θ) range, with an angular step size of 0.02º and 2s acquisition time period.
EP

Raman spectra were acquired using a Trivista 557 spectrophotometer from Princeton

Instruments (Acton, MA, USA), operating in a subtractive configuration and equipped with a
C

thermometric cooler charged coupled device (CCD) detector system. The slits were setting for a
-1
spectral resolution of 2 cm . The scattered light from the powder were obtained at an angle of
AC

90º to the incident beam. It was used a He-Ne laser excitation source (632.8 nm).

MIR spectra were obtained using a Nicolet iS10 FTIR Fourier transform spectrometer

from Thermo Scientific (Waltham, MA USA), equipped with a germanium crystal attenuated total
-1
reflectance accessory (ATR). Spectra were acquired in the 4000-600 cm spectral range, with 4
-1
cm resolution as an average of 32 spectra. Instrument control and spectra acquisition were

done using the Thermo Scientific OMNIC software (version 8.0.342, 2015).

5
ACCEPTED MANUSCRIPT
A FTLA2000 Fourier-transform NIR analyser (ABB, Québec, Canada) was used to

analyse all samples (pure components, physical mixture and co-crystal). The spectrophotometer

is equipped with an indium-gallium-arsenide (InGaAs) detector and powder sampling accessory

(ACC101, ABB, Québec, Canada) with a 2 cm diameter window enabling diffuse reflectance
2
measurements on a 0.28 cm illumination area. The spectra were acquired with a resolution of 8
-1 -1
cm as an average of 64 scans per spectra in the wavenumber range between 10000 cm and

PT
-1
4000 cm . The instrument was controlled via the Grams LT software (version 7, ABB, Québec,

Canada). The background was acquired using a certified material (polytetrafluoroethylene,

RI
PTFE) (SKG8613G, ABB, Québec, Canada). For each sample, three spectral replicates were

SC
measured and the average spectrum considered.

3 Results and discussion

U
AN
3.1 X-ray diffraction

The powder X-ray diffraction (PXRD) results obtained for the pure components of the mixture,
M

the physical mixture and the 1:1 cocrystal are shown in Figure 3. The diffraction pattern of the
D

physical mixture is observed to be essentially composed by the superposition of the information

contained in the patterns of both components, as expected. On the other hand, the formation of
TE

the cocrystal clearly leads to the appearance of a new set of peaks not observed in the pure

components (5.030°, 7.950°, 10.438°, 12.028°, 13.56 0°, 15.137°, 15.657°, 16.590°, 17.056°,
EP

18.824°, 19.895°, 21.122°, 21.848°, 23.057°, 23.717 °, 24.503°, 25.504°, 26.148°, 27.435°,

29.177°, 30.198°, 32.016°, and 37.262° (2 θ)). Additionally, peaks characteristic to GLB are not
C

present in the cocrystal (10.615°, 11.487°, 19.338° , 20.877° and 30.268° (2 θ)). The same
AC

evidence applies to peaks characteristic of TRIS crystal structure (18.055°, 20.056°, 22.464°

and 33.963° (2 θ)). The results are thus conclusive about the formation of a new crystalline

phase.

3.2 Thermal analysis

The thermogram for GLB shows an endothermic transformation at Tonset = 173.96 ºC, with an
-1 24,32,40
enthalpy variation of ∆Hfus=110.14 J g , characteristic of a fusion event (Table 1 and

6
ACCEPTED MANUSCRIPT
Figure 4). The thermogram for TRIS shows two events. An endothermic peak at Tonset = 135.23

ºC, corresponding to a solid-solid phase transition, and a second endothermic peak at Tonset =
32,40
171.96 ºC corresponding to melting, as described in the literature . The thermogram relative

to the 1:1 molar ratio physical mixture presents two peaks partly superimposed which can be

ascribed to the superposition of the fusion events of both compounds.

A single, well defined peak attributed to a fusion event is observed in the thermogram of

PT
-1
the cocrystal, at Tonset = 138.84 °C. The enthalpy variation ( ∆Hfus=135.47J g ) evidences phase

stability and purity of the cocrystallization product.

RI
3.3 Vibrational spectroscopy

SC
3.3.1 Mid infrared

U
The formation of the 1:1 GLB:TRIS cocrystal is clearly observed by comparing the cocrystal
AN
FTIR spectrum to that of the pure components or physical mixture (Figure 5 and Table 2).

Vibrational frequencies corresponding to structural parts of the GLB molecule that could
M

in principle take part of cocrystal synthons have been identified in the pure GLB vibrational

spectra. The two carbonyl stretching ν(C=O) frequencies that, in the pure form, occur at a
D

-1 -1
1714.1 cm and 1616.5 cm , are observed to occur with lower frequency values in the cocrystal
TE

-1 -1
spectrum (1646.7 cm and 1559.7 cm ). Changes in GLB functional groups spectra from pure

to cocrystal form also occur in the symmetric and asymmetric stretching vibrations of SO2,
EP

-1
νs,as(SO2), although the change is markedly smaller than in the previous case (from 1341.6 cm
-1 -1 -1
to 1321.0 cm , and 1157.3cm to 1125cm ).
C

-1
The stretching frequency of NH, ν(NH), appears in pure form of GLB at 3313.3 cm and
-1
AC

3367.3 cm . On the other hand, in the coformer spectrum, the vibration frequencies of
-1
symmetric and asymmetric stretching for NH2, νs,as(NH2), are observed at 3348.0 cm and
-1
3288.9 cm . In the cocrystal spectrum, both vibrations occupy the same spectral region making

it difficult to discern a clear relation to the pure forms spectral information. Nevertheless, it is
-1
apparent that the peak at 3288.9 cm in GLB is not affected by the formation of the cocrystal.

Some vibrational frequencies corresponding to the amine group in are shifted to slightly

higher wavenumbers. This can be assumed to be relative to changes in hydrogen bonding.

Note that when in pure form, NH2 in TRIS may also be involved in hydrogen bonding, so this

7
ACCEPTED MANUSCRIPT
change in the frequency is not directly related to the simple formation of a bond but instead to a

change relative to a prior bonding situation, which may explain the blueshift. The NH2 in TRIS
-1
also presents a scissoring vibration, δ(NH2) to which the peak at 1588.2 cm is attributed, and a
-1
scissoring vibration δ(CH2) is visible at 1463.6 cm . The stretching for CO and CC vibrations,
-1 -1
ν(CO) and ν(CC), appear as very marked peaks at 1036.5cm and 1023.4cm . A scissoring
-1
vibration δ(HNC) is attributed to a vibrational frequency of 981.9 cm . The OH stretching

PT
-1
vibration, ν(OH), has a spectral representation in pure form as a broad band at 3189.1 cm . The

most marked difference in the TRIS spectral information upon change from pure to cocrystal

RI
corresponds to the NH2 symmetric and asymmetric vibrations, νa,as(NH2), as frequency
-1 -1

SC
increases to 3410.3 cm and 3384.0 cm , while the ν(OH) stretching maintains its position.

Thus, it is apparent that only hydrogen atoms in amine groups of TRIS are involved in the

hydrogen bonding leading to cocrystal formation between GLB and TRIS.

U
Results evidence that the most probable synthons leading to the cocrystal formation for
AN
this API:coformer pair involve the amine group in TRIS and carbonyl and sulfonyl groups in

GLB.
M

3.3.2 Raman spectroscopy


D

Raman spectra were also acquired from pure GLB, pure TRIS, the physical mixture and
TE

the cocrystal ( Figure 6 and Table 3). Results confirm the conclusions from PXRD, DSC and

MIR and provide additional conformational information.


EP

-1
The spectra for all cases in the 150-18 cm region are shown in Figure 6a). In this region,

it is observed that GLB presents a set of bands at lower wavenumbers. Usually, the region
C

-1 41
below 200 cm is attributed to the lattice modes . Changes in this spectral region can confirm
AC

the formation of the cocrystal, as new hydrogen bonds formed during the cocrystallization

process stabilize the new structure. Bands in the same region appear in the cocrystal spectrum

presenting wavenumbers and relative intensities very different from those of GLB. For higher
-1
wavenumbers, up to 300 cm , no more information appears in the spectrum of the cocrystal
-1
except for a broad band around 107 cm which appears to be a downshift from a similar band
-1
at 121 cm in GLB. These results represent a powerful proof of the synthesis of the cocrystal.

8
ACCEPTED MANUSCRIPT
-1
In the 1400-700 cm region [Figure 6b)] several other characteristic bands of the GLB
-
molecule are visible in the cocrystal spectrum, some of them downshifted. The bands at 818 cm
1
may be attributed to the angular deformation δ(CCC) or to the τ(HCCC) torsion, and the peak
-1 42
at 830 cm may be to NCN angular deformation plus CC stretching, δ(NCN) + ν(CC) . These

vibrational modes, which occur in the same spectral range of the NC stretching, ν(NC) in the

TRIS molecule, present a downshift, and appear in the cocrystal spectrum as a doublet band

PT
-1 -1
(803 and 809 cm ). On the other hand, the band at 850 cm is attributed to the C-C stretching,
-1
ν(CC), while the peak at 1027 cm may be attributed to ν(CC), δ(CCC) and/or τ(HCNC).

RI
Angular deformations ν(NC), δ(CNC) and/or δ(CCN) vibrations are thought to be responsible for
-1

SC
the peak at 1104 cm . In these cases, the intensity of the peaks is very much reduced up to not
29,43
being identifiable in the cocrystal spectrum . These results show that GLB molecules

involved in the cocrystal formation have suffered conformation changes due to the new

U
interactions with TRIS molecules. The peak attributed to the SO2 symmetric stretching, νs(SO2),
AN
-1 -1
in GLB at 1165 cm undergoes a strong downshift (1136 cm ) in the spectrum of the cocrystal

which is an evidence of the functional group being involved in hydrogen bonding with the API
M

29,31,43 -1 -1 -1 -1 -1
. The bands at 1191 cm , 1211 cm , (1257 cm and 1262 cm ) and 1316 cm may be
42
associated with deformation of HCN and HCC, δ(HCN) and δ(HCC) .
D

-1
Regarding TRIS, the peak at 900.9 cm may be attributed to δ(HCC) and δ(HCN) while the
TE

-1 -1 -1
peak at 925 cm may be attributed to δ(HCC) or τ(CO), while 1030 cm , 1044 cm and 1078
-1 -1
cm are attributed to the stretching ν(CO), and the peak at 1266.1 cm is assumed to be due to
EP

44
carbon-carbon stretching, ν(CC), present in the cocrystal spectrum with increased intensity .
-1 -1
Figure 6c) shows the Raman spectra in the 3400 cm to 2800 cm region. One of the
C

most significant changes from the spectrum of API to the spectrum of the cocrystal is observed
AC

-1
at 2853 cm (marked with an arrow), with the peak attributed to the C-H symmetric stretching,

νs(CH2) in GLB disappearing in cocrystal


42
. Another important change is related to band at

-1
3003 cm corresponding to CH stretching, νs(CH3), in GLB spectrum. The peak intensification

and removal of degeneracy of three bands is observed in the cocrystal spectrum in that same
-1
small region. The bands in the region 3030-3110 cm may be attributed to CH stretching, ν(CH)
42,45
. These changes involving CH, CH2 and CH3 demonstrate once again the conformational

changes undergone by molecules upon the cocrystal formation.

9
ACCEPTED MANUSCRIPT
The TRIS spectrum in this region presents only two visible peaks worth of mention in the
-1
pure form of the molecule, at 3296 and 3353 cm . These peaks are attributed to NH2 symmetric

and asymmetric stretching, νs(NH2) and νa(NH2) 44 and they are absent in the cocrystal form.

This behavior confirms the participation of the NH2 unit, belonging to the TRIS, in the hydrogen

bond interaction.

PT
3.3.3 Near- infrared spectroscopy

RI
The NIR spectrum obtained for the physical mixture is very similar to the spectrum of GLB, and

shows very little contribution from TRIS. However, the cocrystal spectrum shows some

SC
distinctions from the physical mixture that will be discussed below (Figure 7).
-1
At high wavenumbers, in the 9000-8000 cm , GLB and physical mixture show two very

U
broad bands corresponding to the second overtone region of CH, which are not changed with
-1
AN
the formation of the cocrystal. In the 1000-9000 cm range TRIS shows a clear peak located in

a region attributed to the third overtone of amine (RNH2), which appears very much broadened

in physical mixture and vanishes in the cocrystal spectrum. This is a clear indication that the
M

amine group of TRIS is involved in the cocrystal API:coformer hydrogen bonding pattern. In the
-1
D

6900-6100 cm range, there is a marked peak structure in GLB and physical mixture that
st
appears to broaden with the formation of the cocrystal. This region corresponds to the 1
TE

st
overtone of amide (present in GLB), and to the 1 overtone of RNH2 (present in TRIS). The
-1
increase in intensity in 6500-6100cm in a broad undefined band may reflect the changes in
EP

vibrational frequencies of TRIS RNH2 with the formation of hydrogen bonds. In the 6000-5500
-1
cm range GLB spectrum shows a characteristic peak structure that may be attributed to
C

aromatic CH vibrations and to carbonyl in GLB. This region is essentially unchanged in the
AC

cocrystal spectrum. The changes in the cocrystal spectrum when compared to the physical
-1
mixture and GLB spectra between 4900-4300 cm , are related with the amide groups of GLB

since this region of the NIR spectrum is related, among others, with the first overtone vibrations

of the amide group.

The NIR analysis is consistent with the evidences obtained with Raman and Mid infrared

spectroscopy, that the groups involved in the hydrogen bonds between GLB and TRIS are the

amide groups of GLB and the amine group of TRIS (Figure 8).

10
ACCEPTED MANUSCRIPT
3.4 Cocrystal or salt?

One aspect that is sometimes overlooked in the characterization of a product of a

cocrystalization process regarding new pharmaceutical material is the nature of the interaction,

if the product is a cocrystal or if, in the process of formation the compounds may undergo a

chemical transformation into a salt cocrystal. The challenges and the techniques used for

PT
46
addressing the problem have been recently reviewed . The rule of thumb is that one

component may donate a proton to the other component if the difference between the pKa

RI
values of the two is larger than 3. In the present case, glibenclamide has a pKa=5.3, while TRIS

as pK=8.02, which puts the mixture in the theoretical threshold for salt formation. However,

SC
compared with water as solvent, acids have a higher pKa value in organic solvents. The

variation of the pKa of the compound depends on the nature of the substances. Dissociations of

U
neutral acids is governed the solvation effect. In the dissociation of neutral acid (as
AN
glibenclamide) charges are created and the dissociation process is disturbed when the dielectric
47
constant of the medium decreases (Ԑwater=79.99, Ԑmethanol = 33.30) . Therefore, the pKa values
M

of neutral acids increase, when methanol is used as solvent comparing with water. On the other

hand, in the dissociation of bases (such as TRIS), there is no change in the number of charges,
D

so the change in the dielectric constant of the medium does not affect the dissociation process
TE

48
. Therefore, it can be concluded that in the cocrystallization medium used in this case

(methanol) the pKa difference will be much lower than the same different in water (ΔpKa=2.72).
EP

The formation of a salt between these two components, according to the pKa, would

imply the loss of the proton in the nitrogen located adjacent to the sulfonyl and the carbonyl
C

functional groups of glibenclamide, and the concomitant protonation of the amine group in TRIS.

In the ionic form, GLB would have a negative charge on the carbonyl oxygen or on the donor
AC

nitrogen.

It is also to be noted that, in the present case, as discussed in the vibrational

spectroscopy section, the hydrogen from TRIS involved in hydrogen bonding, most possibly

interacting with the sulfonyl and carbonyl of GLB. This corresponds to a situation where the

formation of a salt cocrystal would correspond to a larger change than the simple exchange

from the donor to acceptor at the hydrogen bond location.

11
ACCEPTED MANUSCRIPT
A direct way to access if the compounds are present in the salt form is by single crystal

X-ray diffraction (SCXRD), where the bonds formed of broken can directly be inferred from the
49
distances between the atoms of interest . However, the product of the synthesis not often

possesses the characteristics for this kind of analysis (e.g. mechanochemical synthesis

produces exclusively microcrystalline powders). Additionally NMR, if available, may be of great


50,51
value for addressing the problem . Vibrational spectroscopy may give some insight about

PT
the structure properties, although the interpretation of the data must be done very carefully.

As discussed in the results section of the vibrational spectroscopy techniques, there is

RI
no evidence that the nitrogen atom in the GLB molecule has donated the hydrogen, nor that the

SC
TRIS molecule changed its protonation state. Regarding the protonation of the amine in TRIS,

the vibrational spectroscopy doesn´t show any evidence. In the MIR spectroscopy the

frequencies attributed to NH2 show slight deviations from the ones observed in the physical

U
mixture spectra, but consistent with the involvement in hydrogen bonding, not with the
AN
-1 -1 -1
protonation of the group (frequency found at 3348cm shifted to 3384cm and 1588cm shifted
-1
to 1577cm ). Moreover, in the Raman spectra the symmetric and asymmetric stretching
M

+
vibration of NH2 are not visible in the cocrystal spectrum. If the formation of a group RNH3 in

TRIS had occurred, it could be expected that the vibrational spectra would present some
D

-1 -1
characteristic vibration e.g. in the range 2700cm - 3100cm , or that the symmetric and
TE

-1 -1
asymmetric stretching would be visible in the at 1660 cm – 1500 cm range. None of these

situations is observed.
EP

-1
The NH2 rocking at 1500cm visible in glibenclamide, is also present in the cocrystal
-1
spectrum, as the stretching modes for NH (approx. 3200cm ) continue to appear in the
C

cocrystal evidence that the hydrogen atom isn´t removed, although this spectral region is
AC

specially difficult to interpret for this particular system because of the amount of NH bonds

present in both molecules.

Furthermore, it would be expected that if the salt was formed, it might be observed a

fusion temperature higher than both formers for the synthesis product. The intermolecular

forces of the initial formers are weak (hydrogen bonding, van der Walls interactions, etc.),

however when a salt is formed the intermolecular forces are of ionic nature therefore stronger.

This will lead to a fusion temperature that is normally higher than the fusion temperature of the

12
ACCEPTED MANUSCRIPT
initial formers. In the case of the product formed in this work, the fusion temperature is lower

than the one from GLB and TRIS. A study showed that a survey of 50 cocrystal analyzed 51%

has melting points between those of the API and coformer, 39% were lower than either the API

and coformer (as is the case of the product formed in this study), and only 6% were higher and
52
4% have the same melting point as either API or conformer .

In conclusion, from the discussion presented in the above paragraphs, the evidences

PT
point to the formation of a cocrystal, and not a salt.

RI
4 Conclusion

SC
Using the slow evaporation method for cocrystal synthesis we were able to obtain

glibenclamide:tromethamine 1:1 cocrystal, using methanol as solvent. The XRPD and DSC

U
results show that a new pure and crystalline phase was obtained. The vibrational spectroscopic
AN
techniques show that the hydrogen bounds between the API and coformer to form the cocrystal

are between the amine group of TRIS and the carbonyl and sulfonyl groups in GLB.
M
D

Acknowledgements
TE

The authors would like to thank CAPES, CNPq, and FAPEMA (APCINTER-00284/14 and

UNIVERSAL-00290/15) for financial support. This work received financial support from the
EP

European Union (FEDER funds POCI/01/0145/FEDER/007265) and National Funds (FCT/MEC,

Fundação para a Ciência e Tecnologia and Ministério da Educação e Ciência) under the
C

Partnership Agreement PT2020 UID/QUI/50006/2013. Mafalda C. Sarraguça thanks FCT

(Fundação para a Ciência e Tecnologia) and POPH (Programa Operacional Potencial Humano)
AC

for her Post-Doc grant reference SFRH/ BPD/ 74788/2010. JMGS would like to thank FAPEMA

(Brazil) for the “visiting researcher” grant PV1-05545/15.

13
ACCEPTED MANUSCRIPT
REFERENCES

1. Good DJ, Rodriguez-Hornedo N 2009. Solubility Advantage of Pharmaceutical


Cocrystals. Cryst Growth Des 9(5):2252-2264.
2. Nanjwade VK, Manvi FV, Nanjwade BK, Maste MM 2011. New trends in the
co-crystallization of active pharmaceutical ingredients. J Applied Pharm Sci 1(8):01-
05.
3. Trask AV, Jones W 2005. Pharmaceutical cocrystals: an emerging approach to

PT
physical property enhancement. MRS Bull 31(11):875-879.
4. Thakuria R, Delori A, Jones W, Lipert MP, Roy L, Rodriguez-Hornedo N 2013.
Pharmaceutical cocrystals and poorly soluble drugs. Int J Pharm 453(1):101-125.

RI
5. Nehm SJ, Rodriguez-Spong B, Rodriguez-Hornedo N 2006. Phase solubility
diagrams of cocrystals are explained by solubility product and solution complexation.
Cryst Growth Des 6(2):592-600.

SC
6. Seefeldt K, Miller J, Alvarez-Nunez F, Rodriguez-Hornedo N 2007.
Crystallization pathways and kinetics of carbamazepine-nicotinamide cocrystals from
the amorphous state by in situ thermomicroscopy, spectroscopy, and calorimetry
studies. J Pharm Sci 96(5):1147-1158.

U
7. Jayasankar A, Good DJ, Rodriguez-Hornedo N 2007. Mechanisms by which
moisture generates cocrystals. Mol Pharm 4(3):360-372.
AN
8. Childs SL, Rodriguez-Hornedo N, Reddy LS, Jayasankar A, Maheshwari C,
McCausland L, Shipplett R, Stahly BC 2008. Screening strategies based on solubility
and solution composition generate pharmaceutically acceptable cocrystals of
M

carbamazepine. Crystengcomm 10(7):856-864.


9. Vishweshwar P, McMahon JA, Bis JA, Zaworotko MJ 2006. Pharmaceutical co-
crystals. J Pharm Sci 95(3):499-516.
D

10. Etter MC 1991. Hydrogen-Bonds as Design Elements in Organic-Chemistry. J


Phys Chem 95(12):4601-4610.
TE

11. Etter MC, Frankenbach GM 1989. Hydrogen-bond directed cocrystallization as a


tool for designing acentric organic solids. Chem Mat 1(1):10-12.
12. Herrmann M, Förter‐Barth U, Kröber H, Kempa PB, Juez‐Lorenzo MdM, Doyle
S 2009. Co‐crystallization and characterization of pharmaceutical ingredients. Part Part
EP

Syst Charact 26(3):151-156.


13. BCS. 2015. Biopharmaceutics Classification System, U.S. Food and Drug
Administration. ed.
C

14. Wei H, Lobenberg R 2006. Biorelevant dissolution media as a predictive tool for
glyburide a class II drug. Eur J Pharm Sci 29(1):45-52.
AC

15. Amidon GL, Lennernas H, Shah VP, Crison JR 1995. A theoretical basis for a
biopharmaceutic drug classification: the correlation of in vitro drug product dissolution
and in vivo bioavailability. Pharm Res 12(3):413-420.
16. Brayfield A. 2014. Martindale: The complete drug reference. Drug monographs.
ed.
17. Dhillon B, Goyal NK, Sharma PK 2014. Formulation and evaluation of
glibenclamide solid dispersion using different methods. Global J Pharmacol 8(4):551-
556.
18. Nery CGC, Pires M, Pianetti GA, Soares C 2008. Caracterização do fármaco
hipoglicemiante glibenclamida. Braz J Pharm Sci 44(1):61-73.
19. Al-Ajmi MF 2011. The effect of fenugreek on the bioavailability of
glibenclamide in normal beagle dogs. Afr J Pharm Pharmacol 5(6):671-677.

14
ACCEPTED MANUSCRIPT
20. Matsui R, Kiyota Y, Nakashima M, Kanayama T. 2001. Antidiabetic external
skin application composition. ed., USA.
21. Arnqvist H, Karlberg B, Melander A 1982. Pharmacokinetics and effects of
glibenclamide in two formulations, HB 419 and HB 420, in type 2 diabetes. Ann Clin
Res 15:21-25.
22. Borchert H, Muller H, Pfeifer S 1976. Zur biologischen verfugbarkeit von
glibenclamide in abhangigkeit von der teilchengrosse. Pharmazie 31:307-309.
23. Chalk J, Patterson M, Smith M, Eadie M 1986. Correlations between in vitro
dissolution, in vivo bioavailability and hypoglycaemic effect of oral glibenclamide. Eur

PT
J Clin Pharmacol 31(2):177-182.
24. FB. 2010. Farmacopeia Brasileira 2010. 5 ed., Brasil: Agência Nacional de
Vigilância Sanitária. Anvisa.

RI
25. BP. 2008. British Pharmacopoeia 2008. ed., London,UK: Office of the British
Pharmacopoeia Commission.
26. Blume H, Ali SL, Siewert M 1993. Pharmaceutical Quality of Glibenclamide

SC
Products a Multinational Postmarket Comparative-Study. Drug Dev Ind Pharm
19(20):2713-2741.
27. Hassan M, Sheikh SM, Sallam E, Al-Hindawi M 1996. Preparation and
characterization of a new polymorphic form and a solvate of glibenclamide. Acta Pharm

U
Hung 67(2-3):81-88.
28. Suleiman MS, Najib NM 1989. Isolation and Physicochemical Characterization
AN
of Solid Forms of Glibenclamide. Int J Pharm 50(2):103-109.
29. Hassan MA, Najib NM, Suleiman MS 1991. Characterization of Glibenclamide
Glassy State. Int J Pharm 67(2):131-137.
M

30. Ghosh LK, Thakur RS, Sharma PK, Ghosh NC, Gupta BK 1998. Development
and evaluation of an ideal formulation of glibenclamide by solid dispersion techniques.
Boll Chim Farm 137(1):26-29.
31. Mah PT, Laaksonen T, Rades T, Aaltonen J, Peltonen L, Strachan CJ 2013.
D

Unravelling the relationship between degree of disorder and the dissolution behavior of
milled glibenclamide. Mol Pharm 11(1):234-242.
TE

32. USP. 2008. US Pharmacopeia 2008. ed., USA: United States Pharmacopeia and
National Formulary
33. Bates RG, Robinson RA 1973. Tris(Hydroxymethyl)Aminomethane - a Useful
EP

Secondary Ph Standard. Anal Chem 45(2):420-420.


34. Roy RN, Swensson EE, LaCross Jr G, Krueger CW 1975. Standard buffer of N,
N-bis (2-hydroxyethyl)-2-aminoethanesulfonic acid (Bes) for use in the physiological
pH range 6.6 to 7.4. Anal Chem 47(8):1407-1410.
C

35. Eilerman D, Rudman R 1980. Polymorphism of crystalline poly


AC

(hydroxymethyl) compounds. III. The structures of crystalline and plastic tris


(hydroxymethyl) aminomethane. J Chem Phys 72(10):5656-5666.
36. Divi S, Chellappa R, Chandra D 2006. Heat capacity measurement of organic
thermal energy storage materials. J Chem Thermodyn 38(11):1312-1326.
37. Kojima T, Tsutsumi S, Yamamoto K, Ikeda Y, Moriwaki T 2010. High-
throughput cocrystal slurry screening by use of in situ Raman microscopy and multi-
well plate. Int J Pharm 399(1-2):52-59.
38. Karki S, Friscic T, Fabian L, Jones W 2010. New solid forms of artemisinin
obtained through cocrystallisation. Crystengcomm 12(12):4038-4041.
39. Mcmahon J, Peterson M, Zaworotko MJ, Shattock T, Bourghol HM. 2010.
Pharmaceutical co-crystal compositions and related methods of use. A61K 31/55
(2006.01) ed., USA.

15
ACCEPTED MANUSCRIPT
40. BP. 2009. British Pharmacopoeia 2009. ed.: The Stationery Office on behalf of
the Medicines and Healthcare products Regulatory Agency (MHRA).
41. Larkin PJ, Dabros M, Sarsfield B, Chan E, Carriere JT, Smith BC 2014.
Polymorph Characterization of Active Pharmaceutical Ingredients (APIs) Using Low-
Frequency Raman Spectroscopy. Appl Spectrosc 68(7):758-776.
42. Karakaya M, Kurekci M, Eskiyurt B, Sert Y, Cirak C 2015. Experimental and
computational study on molecular structure and vibrational analysis of an
antihyperglycemic biomolecule: Gliclazide. Spectrochim Acta A 135:137-146.
43. Panagopoulou-Kaplani A, Malamataris S 2000. Preparation and characterisation

PT
of a new insoluble polymorphic form of glibenclamide. Int J Pharm 195(1-2):239-246.
44. Schroetter S, Bougeard D 1987. The calculated and observed vibrational spectra
of the ordered phase of tris (hydroxymethyl) aminomethane. Ber Bunsenges Phys Chem

RI
91(11):1217-1221.
45. Colthup NB, Daly H, Wiberley SE. 1975. Introduction to Infrared and Raman
Spectroscopy. 2nd ed., USA: Academic Press.

SC
46. Cerreia Vioglio P, Chierotti MR, Gobetto R 2017. Pharmaceutical aspects of salt
and cocrystal forms of APIs and characterization challenges. Adv Drug Deliv Rev
117:86-110.
47. Mohsen-Nia M, Amiri H, Jazi B 2010. Dielectric Constants of Water, Methanol,

U
Ethanol, Butanol and Acetone: Measurement and Computational Study. J Solution
Chem 39(5):701-708.
AN
48. Grushka E, Grinberg N editors. 2012. ed., New York: CRC press.
49. Childs SL, Stahly GP, Park A 2007. The salt-cocrystal continuum: The influence
of crystal structure on ionization state. Mol Pharm 4(3):323-338.
M

50. Filip X, Borodi G, Filip C 2011. Testing the limits of sensitivity in a solid-state
structural investigation by combined X-ray powder diffraction, solid-state NMR, and
molecular modelling. Phys Chem Chem Phys 13(40):17978-17986.
51. Rajput L, Banik M, Yarava JR, Joseph S, Pandey MK, Nishiyama Y, Desiraju
D

GR 2017. Exploring the salt-cocrystal continuum with solid-state NMR using natural-
abundance samples: implications for crystal engineering. Iucrj 4:466-475.
TE

52. Schultheiss N, Newman A 2009. Pharmaceutical Cocrystals and Their


Physicochemical Properties. Cryst Growth Des 9(6):2950-2967.
C EP
AC

16
ACCEPTED MANUSCRIPT

Figure 1: Chemical structure of a) glibenclamide and b) tromethamine.

Figure 2: GLB:TRIS (1:1) monocrystal obtained by slow solvent evaporation.

Figure 3: Powder X-ray diffractograms for GLB (A), GLB+TRIS physical mixture

PT
(B), GLB:TRIS 1:1 cocrystal (C) and TRIS (D).

RI
Figure 4 : DSC results for GLB (A), GLB+TRIS physical mixture (B), cocrystal
(C) and TRIS (D).

SC
Figure 5: Mid-infrared spectra of: (A) GLB, (B) GLB+TRIS physical mixture (C)
cocrystal and (D) TRIS.

U
AN
Figure 6: Raman spectra for GLB (A), GLB+TRIS physical mixture (B),
cocrystal (C) and TRIS (D). Data shown in 3 separate panels for clarity: region
M

3400-2800 cm-1 (a), 1350-780 cm-1 region (b) and 320-18 cm-1 (c).

Figure 7 : NIR spectra preprocessed with standard normal variate (SNV) for
D

GLB (A), GLB+TRIS physical mixture (B), cocrystal (C) and TRIS (D).
TE

Figure 8: Chemical structures of GLB (A) and TRIS (B) with indications where
EP

vibrational spectroscopy showed significant changes in the frequency of the


vibrational modes. Panel C) shows schematically the most probable locations
for synthons.
C
AC

17
ACCEPTED MANUSCRIPT

Table 1: Thermal analysis (DSC) results for GLB, TRIS, physical mixture and cocrystal.

Temperature / °C
-1 -1 ° -1
Transitions ∆H / J g ∆S / J g C
Tonset Tpeak Tendset

PT
GLB 1° (endo) 173.96 174.90 175.99 110.14 0.62

1° (endo) 135.23 136.61 139.39 310.81 2.28


TRIS

RI
2° (endo) 171.96 172.93 173.53 28.61 0.17

1° (endo) 133.75 139.32 141.91 34.29 0.25


Physical mixture

SC
2° (endo) 151.54 162.54 173.40 39.29 0.24

Cocrystal 1° (endo) 138.84 142.89 146.53 135.74 0. 95

U
AN
M
D
TE
C EP
AC
ACCEPTED MANUSCRIPT
Table 2: MIR vibrational frequencies of GLB, TRIS, physical mixture and cocrystal. (n.p. (not
present) denotes attributed peaks that are not visible in the cocrystal; “n.a.” refers to vibrations
which were not possible to attribute a definite frequency in the cocrystal).

-1
Vibrational modes Wavenumber /cm
GLB TRIS Cocrystal
ν (NH) 3367 3313
3313 3410
νs,as (NH2) 3348 3384
3289

PT
ν(OH) --- 3189 n.p
ν(C=O) (amide) 1714 1646

ν(C=O) (benzoil) 1616 1559

RI
ρ(NH) (urea) 1521 1518
νs,as(SO2) 1341 1321
1157 1125

SC
δas,s(NH2) ---- 1588 1577
δ(HNC) --- 982 n.p
788 n.p
ρ(NH2) 915 n.a

U
AN
M
D
TE
C EP
AC
ACCEPTED MANUSCRIPT

Table 3 : Raman frequencies in GLB, physical mixture, cocrystal and TRIS. n.p. (not present)

denotes attributed peaks that are not visible in the cocrystal.


-1
Vibrational mode Wavenumber ( cm )
GLB TRIS Cocrystal
νas(NH2) 3354 n.p.
νs(NH2) 3296 n.p.
νasCH) 3000 3000

PT
νs(CC) 2860 2860
νs(CH) 2853 n.p.
ν(CC) 1266 1266

RI
ν(SO) 1257 1251
δ(HCN), δ(HCC); 1211 1205
νs(SO2) 1164 n.p.
ν(NC), 1104 n.p.

SC
δ(HCN) δ(CCN)
ν(CO) 1078 n.a.
1044 n.p
1030 n.a.

U
ν(CC), 1027 n.p.
δ(CCC) τ(HCNC);
AN
δ(HCN) 925 n.a
τ(CO)
δ(HCC), δ(HCN) 901 n.p
ν(CC) 850 n.a
ν(CC), 832 n.p.
M

δ(NCN)
δ(CCC) 818 816
τ(HCCC).
D

ν(CC) 809 809


TE
C EP
AC
ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC
ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC
ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC
ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC
ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC
ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC
ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC
ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC

Anda mungkin juga menyukai