Two mixtures of neutral cyclodextrins (CDs) were used in Electrokinetic Chromatography (EKC) to model and optimize the simultaneous enantiomeric separation of a group of seven drugs. Heptakis(2,6-di-Omethyl)-β-CD (DM-β-CD) combined with methyl-γ -CD (M-γ -CD) or with carboxyethyl-γ -CD (CE-γ -CD) was employed in a 25 mM formate buffer at pH 3.0 to have the drugs studied positively charged.
Trang 1Journal of Chromatography A 1681 (2022) 4634 4 4
Contents lists available at ScienceDirect
journal homepage: www.elsevier.com/locate/chroma
L García-Cansino a , # , J.M Saz a , # , M.A García a , b , M.L Marina a , b , ∗
a Universidad de Alcalá, Departamento de Química Analítica, Química Física e Ingeniería Química, Ctra Madrid-Barcelona Km 33.600, 28871 Alcalá de
Henares (Madrid), Spain
b Universidad de Alcalá, Instituto de Investigación Química Andrés M del Río, Ctra Madrid-Barcelona Km 33.600, 28871 Alcalá de Henares (Madrid), Spain
a r t i c l e i n f o
Article history:
Received 26 July 2022
Revised 18 August 2022
Accepted 22 August 2022
Available online 27 August 2022
Keywords:
cyclodextrin-electrokinetic chromatography/
chiral separation/ dual systems/ Dubsky’s
model/ drugs/
a b s t r a c t
Two mixtures of neutral cyclodextrins (CDs) were used in Electrokinetic Chromatography (EKC) to model and optimize the simultaneous enantiomeric separation of a group of seven drugs Heptakis(2,6-di-O-
was employed in a 25 mM formate buffer at pH 3.0 to have the drugs studied positively charged Dub- sky’s model was applied to calculate the enantiomer effective electrophoretic mobilities for each com- bination of CDs at different averaged molar fractions and total CDs concentrations The most adequate averaged molar fraction and total CDs concentration in terms of the simultaneous enantiomeric separa-
migration order for some compounds when changing the total CDs concentration was also predicted and the model showed its potential even at concentrations out of the experimental range of CD concentra-
taneous enantiomeric separation of six of the drugs studied (except verapamil) with resolutions ranging from 0.6 to 4.0
© 2022 The Author(s) Published by Elsevier B.V This is an open access article under the CC BY-NC-ND license ( http://creativecommons.org/licenses/by-nc-nd/4.0/)
1 Introduction
Chiral analysis is a very interesting area in analytical
chem-istry due to the different properties that enantiomers may have.
These differences are a very relevant issue in the pharmaceutical,
food, environmental, cosmetic or agrochemical fields, among
oth-ers [1] This interest has originated that chiral separation
meth-ods have been developed enabling the individual determination of
enantiomers Among the most employed chiral separation
tech-niques, Capillary Electrophoresis has attracted a great attention
due to its inherent characteristics such as high efficiency, the
pos-sibility of changing very easily the chiral selector and the low
con-sumption of reagents and samples, being considered an
environ-∗Corresponding author: Tel.: ( + 34) 918854935; fax: ( + 34) 918854971
E-mail address: mluisa.marina@uah.es (M.L Marina)
# These authors contributed equally to this work
mentally friendly technique [2–4] Numerous chiral selectors have been employed in the separation medium in the so-called Elec-trokinetic Chromatography (EKC) mode, such as cyclodextrins (CDs) [5] , macrocyclic antibiotics, chiral surfactants, etc Among all these chiral selectors, CDs have been the most employed due to their discrimination power and the big variety of derivatives commer-cially available However, even when using powerful chiral selec-tors as CDs, sometimes the separation of the enantiomers of a chi-ral compound can be difficult In these cases, one possibility that has demonstrated to be very useful can be the use of a mixture
of two CDs that are combined to produce an enhancement in the chiral separation From the first pioneering works dealing with the use of mixtures of CDs [6–12] for chiral separations by EKC, the combination of CDs has received an increasing attention [13–16] However, the use of these systems generally supposes a complex process for optimizing the most adequate experimental conditions
to achieve a given enantiomeric separation This complexity is even
https://doi.org/10.1016/j.chroma.2022.4634 4 4
0021-9673/© 2022 The Author(s) Published by Elsevier B.V This is an open access article under the CC BY-NC-ND license
( http://creativecommons.org/licenses/by-nc-nd/4.0/ )
Trang 2higher when multicomponent mixtures of chiral compounds have
to be enantiomerically separated From this point of view, the
pro-posal of physical-chemical models to facilitate the optimization of
the chiral separation conditions has demonstrated to be an
inter-esting tool.
For a single CD system, Eq (1) proposed by Wren and Rowe
[17] for an analyte interacting with a chiral selector can be
ap-plied This equation enables the calculation of the apparent
com-plexation constants ( KC) and the electrophoretic mobilities of the
enantiomer-CD complexes (μC) assuming a 1:1 stoichiometry for
these complexes:
μA,eff= μA,f + μCKC[ S ]
where μA,eff is the effective electrophoretic mobility of the
enan-tiomer, μA,fthe electrophoretic mobility of the free enantiomer (in
the absence of CD) and [S] is the concentration of the chiral
selec-tor (the concentration of CD that remains free in the complexation
equilibrium with the enantiomers).
When a mixture of CDs is employed as chiral selector, different
physical-chemical models were developed to describe and predict
the migration of enantiomeric compounds with different
character-istics in EKC [ 7 , 12 , 18-23 ] The fundamentals and potential of these
models were reviewed in some interesting articles [ 13 , 15 , 24 , 25 ].
Lurie et al. [7] proposed a model to describe the impact of
mix-tures of neutral and charged CDs on the migration behaviour and
chiral resolution of cationic analytes This model was based on
considering the two equilibria taking place between each analyte
and each CD present in the mixture A 1:1 stoichiometry has to
be assumed for enantiomer-CD complexes in this model and that
no mixed complexes were formed Also based on the consideration
of two independent equilibria between the enantiomers and each
CD, Surapaneni et al. [18] developed a model for the chiral
separa-tion of neutral analytes with mixtures of neutral and charged CDs.
Equations for selectivity and resolution were given from the
ex-pressions of the enantiomers mobility Kranack et al. [19] derived
equations describing the effect of derivatized CDs with different
substitution degrees on the migration behaviour of analytes These
multicomponent mixtures were considered single-component
ad-ditives if the fraction of each component remained constant These
equations were used in the study of mixtures of charged and
neu-tral CDs From the equilibrium constants and mobilities for each
chiral selector and analyte, migration behaviour of analytes could
be predicted at various concentrations of mixtures of two chiral
selectors Fillet et al. [12] proposed a mechanism to explain the
changes in selectivity observed with dual CDs systems based on
the effect of the chiral selector on the analyte mobility The
over-all mobility difference between the enantiomers with a
combina-tion of CDs was expressed as the addition of the mobility
differ-ence originated by each CD corrected by their respective statistical
weights Two ways to improve the separation selectivity could be
established If the two chiral selectors employed affect in opposite
ways the analyte mobility, then the affinity pattern of both
enan-tiomers for the two chiral selectors should also be opposite On the
contrary, if both chiral selectors affect similarly the analyte
mobil-ity, enantiomers should have a similar affinity pattern for the chiral
selectors Among the different models proposed, Dubsky’s model is
considered a very simple possibility to be applied from a practical
point of view [20] .
Dubsky et al. [20] proposed a theoretical model where the
system is considered to behave as a single hypothetical CD for
each ratio at which the two CDs are mixed (molar fraction, Xi).
In such case, the apparent complexation constant ( Kover
C ) for each enantiomer-CD complex would be calculated by Eq (2) :
Kover
i
being Ki the apparent complexation constant of each enantiomer with each CD separately ( Kiis equivalent to KC in Eq (1) ).
The electrophoretic mobility of the complex formed by each enantiomer with the hypothetical CD (resulting from the combi-nation of two CDs) ( μover
C ) would be obtained by Eq (3) :
μover
iχiμiKi
iχiKi =
iχiμiKi
Kover C
(3)
where μiis the electrophoretic mobility of the complex formed by each enantiomer with each CD separately ( μi is equivalent to μC
in Eq (1) ).
Once these parameters are obtained ( Kover
C and μover
C ) from Eqs (2) and (3) , the effective electrophoretic mobility of each enantiomer with the hypothetical CD ( μA,eff) would be obtained
by an equation similar to that of Wren and Rowe ( Eq (4) ):
μA,eff= μA,f + μover
C Kover
C ctot
1 + Kover
C ctot
(4)
being μA,f the free electrophoretic mobility of each enantiomer
in the absence of CDs and tot the total concentration of the CD mixture that would remain without forming a complex with the enantiomers.
This model assumes that: i) the rate of the complexation reac-tion between the enantiomers and each of the CDs is higher than the separation and interconversion rates, ii) the ratio at which each enantiomer interacts with each CD is 1:1, and iii) the concentration
of free CDs remains constant since the concentration of each enan-tiomer is very low compared to that of each CD [ 7 , 12 , 20 ] On the contrary, the use of pure chiral selectors is not required as theoret-ically justified and experimentally demonstrated by Dubsky’s et al who showed the validity of the model when a commercial mixture
of CDs was employed [ 20 , 21 ].
Although Dubsky’s model has demonstrated its usefulness to predict the non-enantiomeric separation of a mixture of ibuprofen and flurbiprofen using a combination of
heptakis(2,6-di-O-methyl)-β -CD (DM- β -CD) and β -CD or a combination of DM- β -CD and 6-O- α -maltosyl- β -CD [26] , its potential to model and optimize the enantiomeric separation of chiral compounds has scarcely been illustrated In fact, just two articles reported the modelling of chi-ral separations by EKC using mixtures of CDs Thus, the separa-tion of lorazepam enantiomers was modelled when using a com-mercially available mixture of highly sulphated- β -CDs [21] On the other hand, our research team successfully applied Dubsky’s model for the first time to model and rapidly optimize the si-multaneous enantiomeric separation of a multicomponent mix-ture of six chiral phenoxy acid herbicides using a combination
of 2-hydroxypropyl- β -CD and heptakis(2,3,6-tri-O-methyl)- β -CD [27]
The aim of this work was to apply Dubsky’s model to pre-dict and optimize the simultaneous enantiomeric separation of a multicomponent mixture of seven chiral drugs by EKC (sitagliptin, ivabradine, clopidogrel, ibrutinib, bupivacaine, terbutaline, and ve-rapamil) The enantiomers of some of these drugs were recently separated by EKC due to their novelty (ivabradine [28] and ibru-tinib [29] ) and a single CD system was employed in both cases The chiral separation of the other drugs was reported before: sitagliptin was separated using a mixture of CDs [30] , clopidogrel using a single CD system [ 31 , 32 ], and the separation of the enan-tiomers of bupivacaine [33–43] , terbutaline [ 35 , 44-63 ] and vera-pamil [ 33 , 58 , 60 , 61 , 64-68 ] was described in different works under
a variety of experimental conditions After a screening of different CDs, two combinations of CDs were employed and compared in this work to optimize the multicomponent mixture of the studied drugs using Dubsky’s model.
Trang 3L García-Cansino, J.M Saz, M.A García et al Journal of Chromatography A 1681 (2022) 4634 4 4
2 Materials and methods
2.1 Chemicals, reagents, and standards
Formic acid and sodium hydroxide (NaOH) were from
Sigma-Aldrich (St Louis, MO, USA) Dimethyl sulfoxide (DMSO) was from
Merck (Darmstadt, Germany) and hydrochloric acid (HCl) from
Scharlab S.L (Barcelona, Spain) The water employed was purified
in a Millipore Milli-Q system (Bedford, MA, USA) To obtain the
25 mM formate buffer solution, the required volume of formic acid
was diluted with Milli-Q water and the pH was adjusted to 3.0
with 1 M NaOH before to reach the final volume.
CDs used in this work were: 2-hydroxypropyl- β -CD (HP- β
-CD, average degree of substitution (DS) 0.6) from Fluka (Buchs,
Switzerland), heptakis(2,6-di-O-methyl)- β -CD (DM- β -CD) from
Sigma-Aldrich, and methyl- γ -CD (M- γ -CD, DS 12),
carboxyethyl-β -CD (CE- β -CD, DS 3.5), carboxyethyl- γ -CD (CE- γ -CD, DS 3.3),
carboxymethyl- β -CD (CM- β -CD, DS 3.5), carboxymethyl- γ -CD
(CM- γ -CD, DS 3.5), and heptakis(2,3,6-tri-O-methyl)- β -CD (TM- β
-CD) from Cyclolab (Budapest, Hungary) When working with single
CD systems, the adequate amount of CD, to obtain the desired CD
concentration, was dissolved in the buffer solution Regarding
mix-tures of CD systems, the corresponding amounts of the two CDs
to be combined were dissolved in the separation buffer to obtain
the desired individual and total CDs concentrations Averaged
mo-lar fractions of each CD in the CDs mixture were calculated as if
the second component of the mixture, which is M- γ -CD (apparent
and averaged molecular weight 1476.00 g mol−1) or CE- γ -CD
(ap-parent and averaged molecular weight 1535.13 g mol−1), were
sin-gle chiral selectors This was assumed since these CDs have a
sub-stitution degree as previously indicated In these cases, the
aver-aged molar mass indicated in the bottle of each CD was employed
for calculations.
(R)-ivabradine (505.05 g mol−1) was obtained from Toronto
Research Chemicals Canada (North York, ON, Canada);
(S/R)-bupivacaine (molecular weight 324.90 g mol−1), (S)- and
(R)-sitagliptin (molecular weight 505.31 g mol−1), (S)-ivabradine
(molecular weight 505.05 g mol−1), (S/R)-terbutaline (molecular
weight 274.30 g mol−1), (S/R)-verapamil (molecular weight 491.06
g mol−1), (S)-clopidogrel and (S/R)-clopidogrel (molecular weight
419.90 g mol−1) were purchased from Sigma-Aldrich; and
(R)-ibrutinib and (S/R)-ibrutinib (molecular weight 440.50 g mol−1)
were from MedChem Express (Monmouth Junction, NJ, USA) All
standard compounds had a purity > 96 % Stock standard solutions
(600 mg L−1, except for (R)-ivabradine that had a concentration
of 10 0 0 mg L−1) were prepared dissolving the required amount
of each one analyte in DMSO as electroosmotic flow (EOF) marker
and stored at 4 °C Working standard solutions were obtained by
mixing the necessary volumes of each stock standard solution with
Milli-Q water until the desired concentrations were reached.
2.2 Apparatus
An Agilent 7100 CE system from Agilent Technologies
(Wald-bronn, Germany) with a diode array detector (DAD) was employed.
The electrophoretic system was controlled with the HP3DCE
Chem-Station software that included data collection and analysis
Separa-tions were achieved in uncoated fused-silica capillaries of 50 μ m
I.D with a total length (Lt) of 58.5 cm (50 cm effective length (Ld))
from Polymicro Technologies (Phoenix, AZ, USA).
Reagents and standards were weighed using an OHAUS
Adven-turer Analytical Balance (Nänikon, Switzerland) pH measurements
were performed in a pHmeter model 744 from Metrohm (Herisau,
Switzerland) All solutions were sonicated with an ultrasonic bath
B200 from Branson Ultrasonic Corporation (Danbury, CO, USA).
2.3 CE conditions
All analyses were carried out at 25 °C in positive-polarity (20 kV) mode applying a pressure of 50 mbar for 5 and a de-tection wavelength of 200 nm (band width 5 nm) was used Con-ditioning of a new capillary was achieved by flushing 1 M NaOH for 30 min, Milli-Q water for 15 min and buffer solution (25 mM sodium formate at pH 3.0) for 60 min At the beginning of each working day, the capillary was flushed with 0.1 M NaOH for 5 min, Milli-Q water for 5 min, 0.1 M HCl for 5 min, Milli-Q water for
5 min, buffer solution for 5 min, and BGE for 5 min To en-sure repeatability between injections, the capillary was condi-tioned with 0.1 M NaOH for 2 min, Milli-Q water for 2 min, 0.1 M HCl for 2 min, Milli-Q water for 2 min, buffer solution for
2 min, and, finally, with BGE for 2 min.
2.4 Data treatment
The values of migration times, and resolution values (Rs) were obtained using the Chemstation software from Agilent Technolo-gies Excel Microsoft was employed for experimental data analysis and to calculate all required parameters Origin Pro8 was used for the composition of graphs and to obtain the values of the appar-ent and averaged complexation constant KCand the electrophoretic mobility μCfor each enantiomer-CD complex using Eq (1) The experimental effective electrophoretic mobility ( μA,eff) was calculated using Eq (5) :
μA , e f f= Ld Lt
V
1
tm− t 1 0
(5)
where Ld is the effective capillary length, Lt is the total capillary length, V is the voltage, tmis the migration time and t0 is the EOF time (determined with the EOF marker).
3 Results and discussion
Some preliminary experiments were carried out to select ade-quate CDs enabling the chiral separation of the compounds studied when used as the sole chiral selectors in the separation medium With this aim, a screening with eight neutral CDs at a 10 mM concentration was achieved (DM- β -CD, M- γ -CD, CM- β -CD,
CM-γ -CD, CE- β -CD, CE- γ -CD, HP- β -CD, TM- β -CD) A 25 mM formate buffer (pH 3.0) was employed in order to have the drugs studied positively charged A temperature of 25 °C and an applied voltage
of 20 kV were also chosen as initial experimental conditions The CDs presenting more advantages for the separation of a mixture of the drugs studied in terms of number of peaks, analysis time, and peak shape were DM- β -CD, M- γ -CD and CE- γ -CD As DM- β -CD enabled the enantiomeric separation of a higher number of com-pounds while M- γ -CD and CE- γ -CD had complementary selectivi-ties, two mixtures of CDs consisting of DM- β -CD/M- γ -CD and
DM-β -CD/CE- γ -CD were selected These mixtures were employed to evaluate the potential of the Dubsky’s model for the optimization
of the chiral separation of a mixture of the seven drugs as well as
to predict the individual separation of them under different exper-imental conditions The variation of the temperature (15 °C, 20 °C,
25 °C) and the applied voltage (20 kV, 25 kV) did not originate bet-ter results so a temperature of 25 °C and an applied voltage of 20
kV were selected for further experiments.
Each drug was individually injected in CE using a single CD sys-tem based on DM- β -CD or M- γ -CD or CE- γ -CD as the sole chiral selector in the separation buffer (25 mM formate buffer (pH 3.0)
at 25 °C and a separation voltage of 20 kV) For each single system, the CD concentration was varied from 5 to 25 mM The experi-mental effective electrophoretic mobilities obtained for the enan-tiomers of each compound under these conditions using Eq (5) are
Trang 4Table 1
Apparent and averaged association constants ( K C ) and averaged electrophoretic mobilities ( μC ) of the CD-enantiomer complexes for each individual CD obtained using
Eq [1]
K C (L mol −1 ) ± SD μC (m 2 s −1 V −1 ) ± SD K C (L mol −1 ) ± SD μC (m 2 s −1 V −1 ) ± SD K C (L mol −1 ) ± SD μC (m 2 s −1 V −1 ) ± SD
Bupivacaine 1 4 ± 6 -(4 ± 9) x 10 −8 0.04 ± 9 -(4 ± 1080) x 10 −6 4 ± 6 -(3 ± 6) x 10 −8
Bupivacaine 2 10 ± 6 -(1 ± 1) x 10 −8 0.06 ± 6 -(3 ± 261) x 10 −6 2 ± 6 -(6 ± 18) x 10 −8
(S)-Sitagliptin 4 ± 7 -(6 ± 14) x 10 −8 0.02 ± 14 -(4 ± 3400) x 10 −6 0.9 ± 1 -(1 ± 2) x 10 −7
(R)-Sitagliptin 2 ± 7 -(9 ± 28) x 10 −8 0.02 ± 14 -(4 ± 3400) x 10 −6 0.9 ± 1 -(1 ± 2) x 10 −7
(R)-Ivabradine 12 ± 3 -(4 ± 4) x 10 −9 0.03 ± 13 -(4 ± 2030) x 10 −6 10 ± 6 -(1 ± 1) x 10 −8
(S)-Ivabradine 12 ± 3 -(5 ± 5) x 10 −9 0.03 ± 13 -(4 ± 2030) x 10 −6 10 ± 6 -(1 ± 1) x 10 −8
Terbutaline 1 376 ± 135 (7.0 ± 0.8) x 10 −9 7 ± 1 -(1.4 ± 0.5) x 10 −8 - -
Terbutaline 2 419 ± 129 (6.8 ± 0.7) x 10 −9 8 ± 1 -(9 ± 2) x 10 −9 - -
Verapamil 1 442 ± 202 (5.3 ± 0.7) x 10 −9 56 ± 16 (4 ± 2) x 10 −9 72 ± 11 -(2 ± 1) x 10 −9
Verapamil 2 442 ± 202 (5.3 ± 0.7) x 10 −9 76 ± 21 (5 ± 1) x 10 −9 72 ± 11 -(2 ± 1) x 10 −9
(S)-Clopidogrel 137 ± 12 (2 ± 5) x 10 −10 0.3 ± 4 -(8 ± 121) x 10 −7 25 ± 5 -(6 ± 3) x 10 −9
(R)-Clopidogrel 150 ± 13 (2 ± 5) x 10 −10 0.2 ± 5 -(1 ± 20) x 10 −6 25 ± 5 -(6 ± 3) x 10 −9
(S)-Ibrutinib 1246 ± 182 (3.94 ± 0.09) x 10 −9 132 ± 11 (2.1 ± 0.3) x 10 −9 401 ± 26 -(3 ± 10) x 10 −11
(R)-Ibrutinib 1084 ± 193 (3.8 ± 0.1) x 10 −9 141 ± 12 (2.0 ± 0.3) x 10 −9 398 ± 25 -(1 ± 10) x 10 −11
grouped in Table S1 in Supplementary Material for the three CDs
selected.
Data grouped in Table S1 enabled the calculation of the
appar-ent and averaged association constants ( KC) for each enantiomer
and each CD as well as the averaged electrophoretic mobilities
for the enantiomer-chiral selector complexes ( μC) using Eq (1)
Results obtained are grouped in Table 1 From data included in
Table 1 , the effective electrophoretic mobilities were also
theo-retically calculated for all enantiomers with the three CDs using
Eq (1) (Table S2). Fig 1 shows, as an example, the good
agree-ment observed between the values corresponding to the
exper-imental effective electrophoretic mobilities and those calculated
by Eq (1) for (R)-ibrutinib with DM- β -CD, M- γ -CD and CE- γ -CD
when used as the sole chiral selectors in the separation medium.
From the apparent and averaged association constants values for
each enantiomer and each CD as well as the electrophoretic
mo-bilities for the enantiomer-chiral selector complexes included in
Table 1 , Dubsky’s model was applied to calculate the values of
the global apparent and averaged association constants ( KC over)
and global averaged electrophoretic mobilities ( μC over) of the
com-plexes Eqs (2) and ( (3) ) when using each combination of CDs at
different averaged molar fractions (from 0 to 1) relative to DM- β
-CD Results are shown in Table 2 Using these global apparent and
averaged association constants and averaged electrophoretic
mobil-ities for the complexes corresponding to the mixture of CDs,
effec-tive electrophoretic mobilities ( μA,eff) for the enantiomers of each
drug were calculated at different total CD concentrations (from 5
to 40 mM) and averaged molar fractions (from 0 to 1) values
rel-ative to DM- β -CD (Tables S3 and S4 in supplementary material)
( Eq (4) ).
3.1 DM- β -CD/M- γ -CD system
The results obtained for the DM- β -CD/M- γ -CD system allowed
to select the most appropriate averaged molar fraction and total CD
concentration as a compromise enabling the best possible
simulta-neous enantiomeric separation predicted by the model, based on
the calculated differences between the electrophoretic mobilities
for consecutive peaks in the mixture (Table S5) A value of 1 ×10−10
m2 −1 V−1 was established as the minimum difference between
the electrophoretic mobilities to experimentally observe some
chi-ral discrimination An averaged molar fraction of 0.8 for DM- β -CD
and total CD concentrations ranging from 20 to 40 mM were
con-sidered the best option allowing the individual enantiomeric
sep-aration of each drug (except verapamil and ivabradine) as well as
the simultaneous enantiomeric separation of the mixture In
ad-dition, the model predicted that, at an averaged molar fraction of
0.8 relative to DM- β -CD, a total CDs concentration of 23 mM was the optimum to achieve the simultaneous enantiomeric separation
of the drugs studied in the mixture In fact, although at concen-trations higher than 23 mM the model predicted an improvement
in the enantiomeric separation for sitagliptin and ivabradine, an approaching between the peaks corresponding to clopidogrel and ibrutinib was also predicted (Table S5) Even, a reversal in the mi-gration order for both compounds was predicted at concentrations higher than 30 mM (see Fig 2 A) This inversion could also be pre-dicted by the model for other values of the molar fraction relative
to DM- β -CD such as 0.7 and 0.9 ( Figs 2 B and 2 C) but at different total CDs concentrations (from 30 to 35 mM for a molar fraction
of 0.7 and from 25 to 30 mM for a molar fraction of 0.9).
In order to corroborate these predictions derived from the model, a mixture of the seven drugs studied was injected under the selected conditions (a molar fraction of 0.8 relative to DM- β
-CD and total CDs concentrations from 20 to 40 mM). Fig 3 and Table 3 show, as an example, the separations and resolutions ob-tained, respectively, at total CDs concentrations of 20, 23, 24, 25,
30 y 40 mM As shown in Fig 3 , a total CD concentration of 23 mM was observed to allow the best simultaneous enantiomeric separa-tion of six drugs (except verapamil) These results agreed with the predictions of the model including the fact that verapamil enan-tiomers were not separated at any of the total CD concentration values assayed. Fig 3 also shows that an inversion in the migration order for clopidogrel and ibrutinib was experimentally observed when increasing the total CDs concentration from 20 to 40 mM according to the model predictions.
In addition to the optimization of the simultaneous enan-tiomeric separation of the mixture of the seven drugs derived from the application of the model, some interesting effects could be ob-served at an individual level for some of the compounds investi-gated when using the mixture of CDs. Table 4 compares the differ-ences between the enantiomer effective electrophoretic mobilities calculated for the mixture of CDs ( μ3) with the sum of these differences experimentally obtained with each single CD system ( μ1+ μ2) for all the compounds studied As shown in Table 4 , the model predicted a loss in the chiral separation for bupivacaine and verapamil when using the mixture of CDs at the three con-centrations for which the results predicted by the model and the experimental values observed could be compared (20, 25 and 30 mM) At these three concentrations, this loss in the enantiomeric separation predicted by the model was experimentally corrobo-rated through a decrease in the enantiomeric resolutions The same effect was observed for other compounds for some total CDs con-centrations, e.g., terbutaline, clopidogrel and ibrutinib at 20 mM (no individual data were obtained for other concentrations due
Trang 5Gar
Table 2
Global association constants (K Cover ) and electrophoretic mobilities ( μCover x 10 8 ) of each enantiomer-CD complex for the DM- β-CD/M- γ-CD and DM- β-CD/CE- γ-CD systems at different DM- β-CD averaged molar fractions ( χDM ) using Eqs [2] and [3]
DM-β-CD/M-γ-CD
χDM Bupi 1 Bupi 2 (S)-Sitag (R)-Sitag (R)-Ivab (S)-Ivab Ter 1 Ter 2 Ver 1 Ver 2 (S)-Clop (R)-Clop (S)-Ibrut (R)-Ibrut
K C over 0.0 0.04 0.06 0.02 0.02 0.03 0.03 7 8 56 76 0.3 0.2 132 141
0.9 3 9 3 2 11 10 339 378 403 405 123 135 1135 990
1.0 4 10 3 2 12 11 376 419 442 442 137 150 1246 1084
χDM Bupi 1 Bupi 2 (S)-Sitag (R)-Sitag (R)-Ivab (S)-Ivab Ter 1 Ter 2 Ver 1 Ver 2 (S)-Clop (R)-Clop (S)-Ibrut (R)-Ibrut
μCover 0.0 -408.67 -259.65 -431.83 -431.83 -404.66 -404.66 -1.36 -0.90 0.40 0.47 -84.87 -105.13 0.21 0.20
0.1 -36.46 -15.77 -24.47 -34.81 -8.05 -8.74 0.42 0.45 0.46 0.50 -1.58 -1.51 0.30 0.28
0.2 -19.19 -7.86 -14.16 -20.62 -3.85 -4.23 0.56 0.57 0.49 0.51 -0.70 -0.66 0.34 0.32
0.3 -13.07 -5.11 -10.60 -15.66 -2.43 -2.70 0.62 0.61 0.50 0.52 -0.40 -0.38 0.36 0.34
0.4 -9.94 -3.71 -8.80 -13.14 -1.71 -1.93 0.65 0.63 0.51 0.52 -0.25 -0.23 0.37 0.35
0.5 -8.04 -2.86 -7.71 -11.61 -1.28 -1.47 0.66 0.65 0.52 0.52 -0.16 -0.15 0.38 0.36
0.6 -6.76 -2.29 -6.98 -10.58 -0.99 -1.16 0.67 0.66 0.52 0.53 -0.10 -0.09 0.38 0.36
0.7 -5.84 -1.89 -6.46 -9.85 -0.79 -0.94 0.68 0.66 0.53 0.53 -0.06 -0.05 0.39 0.37
0.8 -5.15 -1.58 -6.07 -9.29 -0.63 -0.78 0.69 0.67 0.53 0.53 -0.03 -0.02 0.39 0.37
0.9 -4.61 -1.34 -5.77 -8.86 -0.51 -0.65 0.69 0.67 0.53 0.53 -0.002 0.01 0.39 0.38
1.0 -4.18 -1.15 -5.52 -8.52 -0.42 -0.55 0.70 0.68 0.53 0.53 0.02 0.02 0.39 0.38
DM-β-CD/CE-γ-CD
χDM Bupi 1 Bupi 2 (S)-Sitag (R)-Sitag (R)-Ivab (S)-Ivab Ter 1 Ter 2 Ver 1 Ver 2 (S)-Clop (R)-Clop (S)-Ibrut (R)-Ibrut
χDM Bupi 1 Bupi 2 (S)-Sitag (R)-Sitag (R)-Ivab (S)-Ivab Ter 1 Ter 2 Ver 1 Ver 2 (S)-Clop (R)-Clop (S)-Ibrut (R)-Ibrut
μCover 0.0 -3.08 -6.35 -14.56 -14.56 -1.13 -1.13 - - -0.16 -0.16 -0.63 -0.63 0.0033 0.0013
0.1 -3.18 -4.78 -11.89 -13.19 -1.04 -1.06 - - 0.12 0.12 -0.39 -0.37 0.10 0.089
0.2 -3.29 -3.78 -10.18 -12.16 -0.96 -1.00 - - 0.26 0.26 -0.26 -0.24 0.17 0.15
0.3 -3.39 -3.10 -8.98 -11.35 -0.88 -0.93 - - 0.34 0.34 -0.18 -0.16 0.23 0.20
0.4 -3.50 -2.59 -8.10 -10.70 -0.80 -0.87 - - 0.40 0.40 -0.12 -0.11 0.27 0.24
0.5 -3.61 -2.21 -7.42 -10.17 -0.73 -0.81 - - 0.44 0.44 -0.083 -0.069 0.30 0.28
0.6 -3.72 -1.91 -6.88 -9.73 -0.66 -0.76 - - 0.46 0.46 -0.053 -0.041 0.33 0.30
0.7 -3.83 -1.66 -6.44 -9.36 -0.60 -0.70 - - 0.49 0.49 -0.029 -0.019 0.35 0.33
0.8 -3.94 -1.46 -6.08 -9.04 -0.54 -0.65 - - 0.51 0.51 -0.011 -0.002 0.37 0.35
0.9 -4.06 -1.29 -5.78 -8.76 -0.48 -0.60 - - 0.52 0.52 0.005 0.013 0.38 0.36
1.0 -4.18 -1.15 -5.52 -8.52 -0.42 -0.55 - - 0.53 0.53 0.018 0.025 0.39 0.38
Trang 6Fig 1 Comparison between the experimental and theoretical effective elec-
trophoretic mobilities of (R)-ibrutinib with each individual CD (DM- β-CD, M- γ-CD,
CE- γ-CD) Experimental ( ) and theoretical ( ο) values Experimental conditions:
uncoated fused-silica capillary, 58.5 cm (50 cm effective length) × 50 μm id; 25
mM formate buffer (pH 3.0); temperature: 25 °C; voltage: 20 kV; hydrodynamic in-
jection: 50 mbar × 5 s; λ: 200 nm ± 5 nm Theoretical values were obtained using
Eq (1)
Table 3
Enantiomeric resolution values obtained for all drugs studied with the DM- β-CD/M-
γ-CD system for an averaged molar fraction of DM- β-CD of 0.8 ( χDM-β-CD = 0.8) and
different total CDs concentrations
Drug Resolution values
20 mM 23 mM 24 mM 25 mM 30 mM 40 mM
Bupivacaine 3.21 3.28 3.89 4.00 3.66 4.48
Sitagliptine 0.00 0.24 0.53 0.57 0.51 –
Ivabradine 0.00 0.21 0.48 0.52 0.44 0.51
Terbutaline 2.74 2.37 2.02 2.07 1.89 1.67
Verapamil 0.00 0.00 0.00 0.00 0.00 0.00
Clopidogrel 1.95 1.44 – – – 1.26
Ibrutrinib 0.76 0.82 – – – 1.19
Fig 2 Theoretical prediction of the inversion in the migration order of clopidogrel
((S)-clop ( ●) and (R)-clop ( ο)) and ibrutinib ((S)-ibrut ( ) and (R)-ibrut ( )) when
a mixture of DM- β-CD/M- γ-CD was used at DM- β-CD averaged molar fractions of 0.8 (A), 0.7 (B) and 0.9 (C) Data were obtained from Table S3
to the comigration of ibrutinib with clopidogrel peaks) For these five compounds, a different behavior was observed for each of the CDs in the mixture which could justify the results obtained No chiral separation was observed for verapamil with DM- β -CD nor for terbutaline with M- γ -CD Likewise, bupivacaine and ibrutinib showed a higher chiral discrimination with one of the CDs in the mixture with respect to the other An improvement of the chiral separation was also predicted by the model and experimentally demonstrated for ivabradine and terbutaline at 25 and 30 mM to-tal CDs concentrations, in spite of the fact that terbutaline did not show enantiomeric separation with M- γ -CD and that ivabradine did not show chiral discrimination with any of the CDs in the mix-ture at 25 mM CDs concentration or with M- γ -CD at 30 mM total CDs concentration In the case of sitagliptine, the model only pre-dicted correctly its behavior at a total CDs concentration of 25 mM (a slight improvement was observed even if this drug did not show
Trang 7Gar
Table 4
Experimental and theoretical differences between the electrophoretic mobilities of enantiomers when using the DM- β-CD/M- γ-CD system ( μ3 ) compared to the use of each CD separately ( μ1 and μ2 ) (a)
ExperimentalμA,effx 10 10 TheoreticalμA,effx 10 10
μ1 μ2 μ1 + μ2 μ3 |μ3 | − (| μ1 + μ2 |) μ1 μ2 μ1 + μ2 μ3 |μ3 | − (| μ1 + μ2 |)
15 mM 5 mM 16 / 4 mM 15 mM 5 mM 16 / 4 mM
20 mM 5 mM 20 / 5 mM 20 mM 5 mM 20 / 5 mM
25 mM 5 mM 24 / 6 mM 25 mM 5 mM 24 / 6 mM
(a)μ3 = μA,eff (enantiomer 1) - μA,eff (enantiomer 2) when using the DM- β-CD/M- γ-CD system ;μ1 and μ2 = μA,eff (enantiomer 1) - μA,eff (enantiomer 2) for each CD separately .
Trang 8Fig 3 Electropherograms corresponding to the separation of the enantiomers of
the drugs mixture in a standard solution containing (1) racemic bupivacaine 10 mg
L −1 , (2) (S)-sitagliptine 5 mg L −1 and (R)-sitagliptine 10 mg L −1 , (3) (R)-ivabradine 5
mg L −1 and (S)-ivabradine 7 mg L −1 , (4) racemic terbutaline 10 mg L −1 , (5) racemic
verapamil 5 mg L −1 , (6) (R)-clopidogrel 5 mg L −1 and (S)-clopidogrel 7 mg L −1 , (7)
(S)-ibrutrinib 5 mg L −1 and (R)-ibrutrinib 7 mg L −1 , using 25 mM formate buffer
(pH 3.0) with DM- β-CD/M- γ-CD ( χDM-β-CD = 0.8) as chiral selector at different total
CDs concentration (A) 20 mM, (B) 23 mM, (C) 24 mM, (D) 25 mM, (E) 30 mM, and
(F) 40 mM Other experimental conditions as in Fig 1
chiral discrimination with M- γ -CD) As a result, out of a total of
17 cases ( Table 4 ), 15 were correctly predicted by the model at a
qualitative level.
3.2 DM- β -CD/CE- γ -CD system
As in the case of the DM- β -CD/M- γ -CD system, the application
of the Dubsky’s model for the DM- β -CD/CE- γ -CD mixture allowed
to select the most appropriate averaged molar fraction relative to
DM- β -CD and total CDs concentration based on the calculated
dif-ferences between the electrophoretic mobilities between
consecu-tive peaks in the mixture (Table S6) An averaged molar fraction
of 0.8 for DM- β -CD and total CD concentrations ranging from 20
to 40 mM were considered the best option allowing the
individ-ual enantiomeric separation of each drug as well as the
simulta-neous enantiomeric separation of the mixture The differences
be-tween the electrophoretic mobilities for the enantiomers could not
be calculated for terbutaline since it showed enantiomeric
separa-tion only with DM- β -CD as the sole chiral selector and no peaks
were observed for this drug with CE- γ -CD as the sole chiral
selec-tor For verapamil, no differences between the electrophoretic
mo-bilities for the enantiomers were predicted at any of the averaged
molar fraction and total CDs concentrations considered In
addi-tion, the model predicted that, at an averaged molar fraction of 0.8,
a total CDs concentration of 40 mM was the optimum to achieve
the simultaneous enantiomeric separation of the drugs studied in
the mixture.
A mixture of the seven drugs studied was injected under the
se-lected conditions (averaged molar fraction 0.8 relative to DM- β -CD
and total CDs concentration from 20 to 40 mM). Fig 4 shows, as
an example, the separations obtained at total CDs concentrations
of 20, 30 and 40 mM As shown in Fig 4 and Table 5 , a total CD
concentration of 40 mM was observed to allow the best
simultane-ous enantiomeric separation of six drugs (except verapamil) These
results agreed with those predicted by the model including the fact
that verapamil enantiomers were not separated at any of the total
CDs concentration values assayed. Fig 4 also shows that an
inver-Fig 4 Electropherograms corresponding to the separation of the enantiomers of
the drugs mixture in a standard solution containing (1) bupivacaine racemic 10 mg
L −1 , (2) (S)-sitagliptin 5 mg L −1 and (R)-sitagliptin 10 mg L −1 , (3) (R)-ivabradine
5 mg L −1 and (S)-ivabradine 7 mg L −1 , (4) racemic terbutaline 10 mg L −1 , (5) ve- rapamil racemic 5 mg L −1 , (6) (R)-clopidogrel 5 mg L −1 and (S)-clopidogrel 7 mg
L −1 , (7) (S)-ibrutrinib 5 mg L −1 and (R)-ibrutrinib 7 mg L −1 , using 25 mM formate buffer (pH 3.0) with a mixture of DM- β-CD/CE- γ-CD ( χDM-β-CD = 0.8) at different to- tal CDs concentration (A) 20 mM, (B) 30 mM, and (C) 40 mM Other experimental conditions as in Fig 1
Table 5
Enantiomeric resolution values obtained for all drugs studied with the DM- β- CD/CE- γ-CD system for an averaged molar fraction of DM- β-CD of 0.8 ( χDM-β-CD = 0.8) and different total CDs concentrations
Drug
Resolution values
20 mM 30 mM 40 mM
sion in the migration order for clopidogrel and ibrutinib was ex-perimentally observed when increasing the total CDs concentration from 20 to 40 mM This inversion in the migration order for these two compounds was predicted by the model as shown in Fig 5 It can be observed that the model predicted that the inversion in the migration order for clopidogrel and ibrutinib could be expected for
an averaged molar fraction of 0.8 at concentrations ranging from
30 to 35 mM ( Fig 5 A) This inversion could also be predicted by the model for other values of the averaged molar fraction such as 0.7 and 0.9 ( Figs 5 B and 5 C) but at different total CDs concentra-tions (from 35 to 40 mM for an averaged molar fraction of 0.7 and from 25 to 30 mM for an averaged molar fraction of 0.9).
In addition to the optimization of the simultaneous enan-tiomeric separation of the mixture of the seven drugs derived from the application of the model, some interesting effects could be ob-served at an individual level for some of the compounds investi-gated In fact, improvements in the chiral separation of some com-pounds could be observed when using the mixture of the two CDs with respect to the use of single CD systems For example, the model predicts for ivabradine, sitagliptin, and ibrutinib that the difference between the electrophoretic mobilities for enantiomers with the mixture of CDs should be higher than that obtained when using a single CD system, and this fact was experimentally demon-strated ( Table 6 ) In the case of clopidogrel and bupivacaine a
Trang 9Gar
Table 6
Experimental and theoretical differences between the electrophoretic mobilities of enantiomers when using the DM- β-CD/CE- γ-CD system ( μ3 ) compared to the use of each CD separately ( μ1 and μ2 ) (a)
ExperimentalμA,effx 10 10 TheoreticalμA,effx 10 10
μ1 μ2 μ1 + μ2 μ3 |μ3 | − (| μ1 + μ2 |) μ1 μ2 μ1 + μ2 μ3 |μ3 | − (| μ1 + μ2 |) DM-β-CD CE-γ-CD DM-β-CD/CE-γ-CD DM-β-CD CE-γ-CD DM-β-CD/CE-γ-CD
15 mM 5 mM 16 / 4 mM 15 mM 5 mM 16 / 4 mM
20 mM 5 mM 20 / 5 mM 20 mM 5 mM 20 / 5 mM
25 mM 5 mM 24 / 6 mM 25 mM 5 mM 24 / 6 mM
(a)μ3 = μA,eff (enantiomer 1) - μA,eff (enantiomer 2) when using the DM- β-CD/CE- γ-CD system ;μ1 and μ2 = μA,eff (enantiomer 1) - μA,eff (enantiomer 2) for each CD separately .
Trang 10Fig 5 Theoretical prediction of the inversion in the migration order of clopidogrel
((S)-clop ( ●) and (R)-clop ( ο)) and ibrutinib ((S)-ibrut ( ) and (R)-ibrut ( )) when
the mixture DM- β-CD/ CE- γ-CD was used at DM- β-CD averaged molar fractions of
0.8 (A), 0.7 (B) and 0.9 (C) Data were obtained from Table S4
crease in the difference between the electrophoretic mobilities for
enantiomers with the mixture of CDs was predicted by the model
(for 25 and 30 mM total CDs concentrations for bupivacaine and
for 20 mM total CDs concentration for clopidogrel) and this fact
was also experimentally corroborated Out of 14 cases, the model
correctly predicted 13 cases at a qualitative level ( Table 6 ).
4 Concluding remarks
The chiral separation of a mixture of seven drugs by EKC
us-ing two mixtures of CDs (DM- β -CD/M- γ -CD and DM- β -CD/CE- γ
-CD) was modelled using Dubsky’s model A good agreement
be-tween the experimental results obtained and those predicted by
the model was observed The model showed its potential to
opti-mize the simultaneous enantiomeric separation of the mixture of
drugs studied in this work even at concentrations out of the
exper-imental range of CD concentrations employed Some interesting ef-fects relative to the use of the CDs mixtures were also predicted by the model and experimentally corroborated In addition, the model predicted the reversal in the migration order of some compounds when changing the total CDs concentration according to the exper-imental observations The combination of DM- β -CD with CE- γ -CD
at a 40 mM total CDs concentration showed to be the most appro-priate conditions to achieve the simultaneous enantiomeric sepa-ration of the multicomponent mixture of drugs with resolutions values ranging from 0.6 to 4.0.
This is the first time that Dubsky’s model is applied to predict
a simultaneous chiral separation of a mixture by extrapolating the results to total CDs concentrations out of the experimental range established to obtain the parameters of the model Thus, it has been shown that the model enabled to find interesting separation conditions from a few initial experiments achieved with each pair
of enantiomers and each CD employed as the sole chiral selector Therefore, the model can be considered a powerful tool to help in the optimization of chiral separations when mixtures of CDs are employed in EKC.
Declaration of Competing Interest
The authors declare that they have no known competing finan-cial interests or personal relationships that could have appeared to influence the work reported in this paper.
CRediT authorship contribution statement
L García-Cansino: Investigation, Validation, Data curation, Vi-sualization, Writing – original draft J.M Saz: Investigation, Formal analysis, Validation, Data curation, Visualization, Writing – original draft M.A García: Methodology, Visualization, Data curation, Re-sources, Supervision, Writing – original draft, Writing – review & editing, Project administration, Funding acquisition M.L Marina:
Conceptualization, Methodology, Visualization, Data curation, Re-sources, Supervision, Writing – original draft, Writing – review & editing, Project administration, Funding acquisition.
Data availability
Data will be made available on request.
Acknowledgments
Authors thank financial support from the Spanish Min-istry of Science and Innovation (research project PID2019-104913GB-I00, Agencia Estatal de Investigación, Referencia del Proyecto/AEI/10.13039/50110 0 011033), and the University of Alcalá for research project CCG20/CC-023 L.G.C thanks the University of Alcalá for her predoctoral contract.
Supplementary materials
Supplementary material associated with this article can be found, in the online version, at doi: 10.1016/j.chroma.2022.4634 4 4
References
[1] G D’Orazio, C Fanali, C Dal Bosco, A Gentili, S Fanali, Chiral separation and analysis of antifungal drugs by chromatographic and electromigration tech- niques: Results achieved in 2010-2020, Reviews in Analytical Chemistry 40 (2021) 220–252, doi: 10.1515/revac- 2021- 0136
[2] E Sánchez-López, M Castro-Puyana, M.L Marina, A.L Crego, Chiral Separa- tions by Capillary Electrophoresis, in: J.L Anderson, A Berthod, V.P Estévez, A.M Stalcup (Eds.), Analytical Separation Science, Wiley-VCH-Books, Germany,
2015, pp 731–744