Enantiopurity determination of the enantiomers of the triple reuptake inhibitor indatraline

Article abstract of DOI:10.1002/chir.22235

The present study describes the development of two approaches for the determination of the enantiopurity of both enantiomers of indatraline. Initially, a method was developed using different chiral solvating agents (CSAs) for diastereomeric discrimination regarding signal separation in ¹H nuclear magnetic resonance (NMR) spectroscopy, revealing MTPA as a promising choice for the differentiation of the indatraline enantiomers. This CSA was also tested for its ideal molar ratio, temperature, and solvent. Optimized conditions could be achieved that made determination of enantiopurity for (1R,3S)-indatraline up to 98.9% enantiomeric excess (ee) possible. To quantify even higher enantiopurities, a high-performance liquid chromatography (HPLC) method based on a modified β-cyclodextrine phase was established. The influence of buffer type, concentration, pH value, percentage and kind of organic modifier, temperature, injection volume as well as sample solvent on chromatographic parameters was investigated. Afterwards, the reliability of the established HPLC method was demonstrated by validation according to the ICH guideline Q2(R1) regarding specificity, accuracy, precision, linearity, and quantitation limit. The developed method proved to be strictly linear within a concentration range of 1.25-1000 μM for the (1R,3S)-enantiomer and 1.25-750 μM for its mirror image that enables a reliable determination of enantiopurities up to 99.75% ee for the (1R,3S)-enantiomer and up to 99.67% ee for the (1S,3R)-enantiomer.

Full text of DOI:10.1002/chir.22235

CHIRALITY 25:923–933 (2013)
Enantiopurity Determination of the Enantiomers of the Triple
Reuptake Inhibitor Indatraline
STEFANIE H. GRIMM, LARS ALLMENDINGER, GEORG HÖFNER, AND KLAUS T. WANNER*
Department of Pharmacy – Center of Drug Research, Ludwig-Maximilians-Universität München, Munich, Germany
ABSTRACT
The present study describes the development of two approaches for the deter-
mination of the enantiopurity of both enantiomers of indatraline. Initially, a method was developed
using different chiral solvating agents (CSAs) for diastereomeric discrimination regarding signal
1
separation in H nuclear magnetic resonance (NMR) spectroscopy, revealing MTPA as a promising
choice for the differentiation of the indatraline enantiomers. This CSA was also tested for its ideal
molar ratio, temperature, and solvent. Optimized conditions could be achieved that made determina-
tion of enantiopurity for (1R,3S)-indatraline up to 98.9% enantiomeric excess (ee) possible. To quantify
even higher enantiopurities, a high-performance liquid chromatography (HPLC) method based on a
modified β-cyclodextrine phase was established. The influence of buffer type, concentration, pH
value, percentage and kind of organic modifier, temperature, injection volume as well as
sample solvent on chromatographic parameters was investigated. Afterwards, the reliability
of the established HPLC method was demonstrated by validation according to the ICH
guideline Q2(R1) regarding specificity, accuracy, precision, linearity, and quantitation limit.
The developed method proved to be strictly linear within a concentration range of 1.25–1000
μM for the (1R,3S)-enantiomer and 1.25-750 μM for its mirror image that enables a reliable
determination of enantiopurities up to 99.75% ee for the (1R,3S)-enantiomer and up to 99.67% ee
for the (1S,3R)-enantiomer. Chirality 25:923–933, 2013.
© 2013 Wiley Periodicals, Inc.
KEY WORDS: ICH Q2(R1); validation; HPLC; NMR; CSA; cyclodextrine; enantiomeric excess
INTRODUCTION
of depression, but about 35% of patients with depression
do not or only partially respond to these agents. That’s
3
Indatraline [trans-3-(3,4-dichlorophenyl)-N-methyl-indan-1-
amine, rac-1], also known as Lu 19-005 (Fig. 1), was reported
for the first time in 1982 in a study of K.P. Bøgesø that aimed
at the development of new psychopharmacologically active
drugs for the treatment of psychic disorders such as depression.
The trans-configured 3-phenyl-indan-1-amines with different
halogen substitution patterns that in addition to indatraline
rac-1) had been studied as part of this project demonstrated
nonselective inhibition of the neurotransmitter transporters for
dopamine (DAT), norepinephrine (NET), and serotonin
why current efforts include the development of triple reup-
take inhibitors, as so-called next-generation antidepres4-
sants, for a faster onset and better efficiency in therapy.
Due to its high and close affinities to DAT, NET, and
SERT and its long-acting effect, indatraline (rac-1), and
especially its (1R,3S)-enantiomer, (1R,3S)-1, could serve
as a model substance to test the potential of this new class
1
(
5
6
of agents, e.g., for the treatment of depression or even
7
cocaine abuse.
For the characterization of the pharmacological and
pharmacokinetic properties of the enantiomers of a chiral
compound, only highly enantioenriched samples should be
used, as the wrong isomer will falsify the data the more it is
present in the samples studied. In the study of the determination
of the eudismic ratio of neurotransmitter transporter inhibition
by the indatraline enantiomers, Bøgesø et al. used samples with
(
SERT), proteins that are responsible for signal termination
and recycling of the respective neurotransmitters, whereas most
of the cis-configured 3-phenyl-indan-1-amines selectively
inhibited SERT.
2
Being trans-configured, indatraline (rac-1) also displayed
similar and also very low IC50 values in uptake experiments for
each of the three above-mentioned neurotransmitter transporters
based on rat synaptosomes [dopamine (DA) = 0.99 nM,
norepinephrine (NE) = 0.26 nM, serotonin (5-HT) = 0.48 nM].
Also for the single enantiomers of indatraline (rac-1), low to
negligible selectivities as inhibitors of these transporters (DAT,
NET, and SERT) were found, although one enantiomer, the
2
enantiopurities of ≥ 95%. The eudismic ratio for DAT inhibition of
indatraline enantiomers was found to be 51 [(1R,3S)- as compared
to the (1S,3R)-enantiomer]. But considering the enantiopurity
of the samples used, which in the worst case could still
contain 2.5% of the wrong enantiomer, the true eudismic ratio
of the indatraline enantiomers as DAT inhibitors might possi-
bly be distinctly higher than that. As can be seen from this
(
1R,3S)-stereoisomer, appeared to be distinctly more potent than
its stereoisomeric counterpart. The eudismic ratio in favor of the
1R,3S)-enantiomer ranged from 5 (5-HT) to 51 (DAT).2
(
The neuronal systems of DA, NE, and 5-HT are involved in
different mental disorders such as depression, Parkinson’s
disease, schizophrenia, and are also associated with drug
abuse. Dual (NE and 5-HT, or NE and DA, respectively) as
well as selective NE and selective 5-HT reuptake inhibitors
Additional Supporting Information may be found in the online version of this
article.
*
Correspondence to: Klaus T. Wanner, Department of Pharmacy – Center of
Drug Research, Ludwig-Maximilians-Universität München, Butenandtstr.
5-13, 81377 Munich, Germany. E-mail: Klaus.Wanner@cup.uni-muenchen.de
Received for publication 16 May 2013; Accepted 22 July 2013
DOI: 10.1002/chir.22235
Published online 7 October 2013 in Wiley Online Library
(wileyonlinelibrary.com).
[
e.g., sertraline (2), Zoloft, a cis-configured structural analog
of indatraline; see Fig. 1] are widely used in the treatment
©
2013 Wiley Periodicals, Inc.

9
24
GRIMM ET AL.
High-resolution mass spectrometry (HRMS): electron ionization (EI+):
+
·
M
calcd. for C16
2
H15NCl : 291.0576; found: 291.0571 (error 1.7 ppm).
Chemicals
L-(+)-tartaric acid p.a. for the racemate resolution was provided by
Riedel de Haën.
HPLC-grade acetonitrile and methanol were purchased from VWR
International (Darmstadt, Germany). Water was obtained by distillation
and filtration (0.45 μm filter) of demineralized water, prepared by a reverse
osmosis system.
Fig. 1. (1R,3S)-indatraline [(1R,3S)-1] and sertraline (1S,4S)-4-(3,4-
dichlorophenyl)-N-methyl-1,2,3,4-tetrahydronaphthalen-1-amine) (2).
Ammonium acetate (AAc) p.a. was purchased from Fluka
(Taufkirchen, Germany), acetic acid p.a. from VWR International, and
triethylamine (TEA) pure ( ≥ 99%), which was distilled prior to use, from
AppliChem (Darmstadt, Germany).
(S)-(-)-1,1´-bi-naphthyl-2,2´-diol for synthesis 99.8% (99.7% ee) (Merck,
Darmstadt, Germany), (S)-(+)-6-methoxy-α-methyl-2-naphthaleneacetic acid
example, there is a substantial need for analytical methods that
enable a reliable and precise determination of the enantiopurity
even of highly enantioenriched indatraline samples. Of course,
such methods will also be valuable tools for the preparation of
the indatraline enantiomers, independent of whether this is
accomplished by a resolution via crystallization (by formation
98%, (S)-(+)-O-acetylmandelic acid 99% (98% ee) (both Sigma-Aldrich,
Steinheim, Germany), (R)-(+)-α-methoxy-α-trifluoromethylphenylacetic
acid 99% (Alfa Aesar, Karlsruhe, Germany), and (S)-(-)-α-methoxy-α-
trifluoromethylphenylacetic acid p.a. 99% (97% ee) (Fluka) were used as
2
of diastereomers) as described by Bøgesø et al., or by asym-
6
purchased. The deuterated solvent benzene-d (99.6% D) was purchased from
8–13
metric synthesis applying chiral catalysts
group transfer reagents.
or chiral functional
Sigma-Aldrich, acetonitrile-d (99.8% D) from euriso top (Saarbrücken,
3
14
Germany), and chloroform-d (99.8% D) from Sigma-Aldrich, and euriso top.
So far, only two methods for the characterization of the
enantiopurity of indatraline enantiomers have been described
Resolution of Indatraline Via Crystallization by Formation of
Diastereomers
1
,2,8
in the literature. One is based on polarimetry,
on H nuclear magnetic resonance NMR measurements
the other
1
The resolution of rac-1 was accomplished via crystallization
2
employing (R)-(-)-2,2,2-trifluoro-1-(9-anthryl)ethanol as shift
according to Bøgesø et al., thereby the enantiopurity was monitored by
2
1
reagent. The latter represents the above-mentioned method
the developed H NMR or HPLC method, respectively. Contrary to Bøgesø
2
Bøgesø et al. used for the determination of the enantiopurity
et al., the tartaric acid salt was further transferred into the free base by
an alkaline ether extraction, then finally treated with an excess of
aqueous HCl (1 M) and subsequent freeze-drying of the resulting
residue yielded (1R,3S)-indatraline · HCl [(1R,3S)-1 · HCl], which was
recrystallized from ethyl acetate to give colorless crystals of distinctly
reduced hygroscopicity. The specific rotation of this product (melting point
of indatraline enantiomers which, however, allows the quantifi-
cation of enantiomeric excesses (ee) of no higher than 95%.2
The aim of the present study was to develop a method
enabling the determination of enantiomeric excesses up to at
least 99% (ee) for the two indatraline enantiomers and to validate
this method to demonstrate its reliability.
2
1
1
2
72–173°C) amounted to½αꢀ +16.4 (c 0.39, methanol) using a Perkin Elmer
D
41 C polarimeter (Perkin Elmer, Rodgau, Germany).
1H NMR
MATERIALS AND METHODS
Synthesis of Indatraline · HCl
1
Instrumentation. All H NMR spectra were recorded on a JNM
Eclipse +500 (500 MHz) NMR spectrometer from JEOL equipped with a
direct 5 mm broadband probe operating at 500.16 MHz. The sample
temperature was set to 25°C except for the investigations concerning
the influence of variable temperature (VT) on the enantiomeric chemical
shift differences (ΔΔδ). To achieve optimal results concerning peak
shape, a gradient shim was applied using the gradient shim tool provided
by the instrument operating software. The pulse angle was set to 45° and
a relaxation delay of 1.3 s was used. In all, 64 transients over a frequency
width of 7003 Hz were applied with 32 K data points, giving an overall
digital resolution of 0.21 Hz per points. The spectra were processed using
NMR software MestReNova v5.2.5-4119 provided by Mestrelab Research.
No line broadening was applied prior to Fourier transformation (FT). The
peak integration was performed manually with the integration tool
included in the software. The comparison of the effect of the different
solvents was made by measuring the chemical shift and the full width
at half maximum (FWHM) of each N-methyl proton signal. The chem-
ical shift was manually measured with MestReNova, and in contrast
the FWHM values were measured with delta v4.3.3 software (JEOL).
A 4 times zero-filling prior to FT was applied to obtain a reasonable
number of data points suitable for the determination of FWHM.
Signal-to-noise ratio (S/N) calculations were performed using the
S/N tool in root mean square (RMS) mode provided by the delta
v4.3.3 software (JEOL).
The synthesis of indatraline · HCl (rac-1 · HCl) was accomplished in
five steps starting from commercially available 3,4-dichlorocinnamic acid
(see also Supporting Information). The latter was reacted with benzene in
the presence of AlCl to give 3-(3,4-dichlorophenyl)-3-phenylpropanoic
3
acid, which was transformed to the corresponding acid chloride utilizing
thionyl chloride and subsequently subjected to an intramolecular Friedel
Crafts acylation with AlCl
3
to yield 3-(3,4-dichlorophenyl)indan-1-one
8
(93%). The following steps were performed according to Davies et al. : re-
duction of the ketone with K-Selectride to afford the corresponding alco-
hol. In situ activation of the alcohol as mesilate followed by nucleophilic
substitution with methylamine provided indatraline as free base, which
was converted to its hydrochloride by treatment with aqueous HCl
followed by freeze-drying. All procedures for the synthesis of rac-1 as
well as the analytical data are shown in the Supporting Information.
Analytical data for the free base (rac-1), which is the test material for all
1
H NMR experiments, is listed below.
H NMR for rac-1 (chloroform-d, 500 MHz): δ = 2.24 (dt, J = 13.2/7.0 Hz,
1
1
H, NCHCH
NCH ), 4.25 (dd, J = 6.8/3.2 Hz, 1H, CH
NCH), 6.97 (m, 2H, Hindan), 7.22 (dd, J = 2.1 Hz, 1H, CHCCl),7.23-7.29
m, 2H, Hindan), 7.35 (d, J = 8.2 Hz, 1H, CHCHCCl), 7.39 (dd, J = 8.2/2.1
2
), 2.44 (ddd, J = 13.2/7.8/3.3 Hz, 1H, NCHCH
2
), 2.51 (s, 3H,
3
2
CH), 4.50 (t, J = 7.6 Hz, 1H,
(
1
3
Hz, 1H, CHCHCCl). C NMR (chloroform-d, 125 MHz): δ = 34.15,
4
1
3.21, 48.51, 63.64, 124.71, 125.25, 127.27, 127.45, 128.36, 129.88, 130.27,
30.44, 132.45, 144.74, 145.49, 145.53.
-1
1
Infrared spectroscopy (IR) (film): ev = 3326 cm , 3276, 3068, 3023, 2959,
Preparation of the stock solutions for the H NMR experiments and
1
2
933, 2848,1469.
method development. All H NMR experiments were measured using
+
Mass spectrometry (MS): chemical ionization (CI, CH
5
); m/z (%): 292
a total sample volume of 700 μl. To obtain stock solutions with a defined
concentration, rac-1 and the chiral solvating agents (CSAs) [(S)-(-)-1,1´-
+
(100) [M + H] .
Chirality DOI 10.1002/chir

EE DETERMINATION OF (1R,3S)-INDATRALINE
925
binaphthyl-2,2´-diol (3), (S)-(+)-6-methoxy-α-methyl-2-naphthaleneacetic
acid (4), (S)-(+)-O-acetylmandelic acid (5), (S)-(-)-α-methoxy-α-
trifluoromethylphenylacetic acid ((S)-6), and (R)-(+)-α-methoxy-α-
trifluoromethylphenylacetic acid ((R)-6)] (Fig. 2) were dissolved in an
appropriate volume of the required deuterated solvent (chloroform-d,
HPLC method development, data analysis, and validation. For
each modification in the development of the HPLC method, the samples
were measured in triplicates. Buffers (i.e., AAc and TEAA) were adjusted
to the required pH value by addition of glacial acetic acid.
Peak intensity during method development was defined as peak height
benzene-d
6
, and acetonitrile-d
3
, respectively) to yield final concentrations
[mAU] of the (1S,3R)-1 peak. Resolution (R ) and asymmetry factors (A )
were calculated according to Kuss et al.
s
s
15
of 0.1 M. The samples were prepared by mixing a defined volume of a 0.1
M stock solution of rac-1 (free base) and various volumes of 0.1 M stock
solutions of the respective CSAs in a final volume (700 μl) of the respective
deuterated solvent. When higher molar ratios (1:5 and 1:10, respectively) of
the CSA (S)-(-)-1,1´-binaphthyl-2,2´-diol (3) were examined, no addition of
the 0.1 M stock solution of 3 was possible without changing the final sample
volume. Due to the low solubility of 3 in the chosen deuterated solvents
For quantification, peak areas (y) of both indatraline enantiomers as a
function of their concentration (x) were investigated. Linearity was
investigated by measurement of a series of calibration standards (13
concentration levels, i.e., 2.5, 5, 10, 25, 50, 75, 100, 250, 500, 750, 1000,
1500, 2000, and 3000 μM rac-1 · HCl).
The final method based on a mobile phase composition of 7.3 mM
TEAA pH 4.0 and acetonitrile (85:15, v/v), at a flow rate of 1.0 ml/min
and a temperature of 20°C, an injection volume of 50 μl, a sample solvent
of 7.3 mM TEAA pH 4.0 and a detection wavelength of 230 nm was
validated for specificity (comparison of a sample solvent blank with a
rac-1 · HCl sample at the expected retention time), linearity (described
above), accuracy (defined as recovery obtained for QCs in five
concentrations each in five replicates), precision (calculated as RSD for
the same QCs used for accuracy), and quantitation limit (QL) according
6 3
(chloroform-d, benzene-d , and acetonitrile-d ) the preparation of a higher
concentrated stock solution, e.g., 1 M, was also impossible. Therefore, the
required amount of 3 was weighed and directly dissolved in the final sample
volume (700 μl) already containing the defined concentration of rac-1.
In the end, a 0.1 M stock solution of (1R,3S)-1 in chloroform-d was
made analogously to the rac-1 sample and analyzed under the conditions
established for the rac-1-CSA investigations.
1
6
UV-absorption measurements
to the ICH guidance Q2(R1). Validation of accuracy and precision was
extended to (1R,3S)-1 · HCl samples of high enantiopurity that were
spiked with defined amounts of rac-1 · HCl as described above.
The absorption spectrum of indatraline (rac-1 · HCl) was measured in a
6-well quartz UV plate (Hellma, Müllheim, Germany) by means of a
9
Spectra Max M2e plate reader (Molecular Devices, Ismaning, Germany).
Spectra were corrected for solvent absorption by subtraction of the
spectra of the pure solvents.
RESULTS AND DISCUSSION
1
H NMR Method Development and Optimization
1
In 1985 Bøgesø et al. published an H NMR method based
Chromatography
on the use of a shift reagent that enabled the determination o2f
Instrumentation. An Agilent 1100 HPLC system consisting of an
vacuum degasser, quaternary pump, column oven, diode array detector
the enantiopurity of (1R,3S)- and (1S,3R)-1 up to 95% ee.
Since this determination limit was not sufficient for our
purpose, we intended to develop an alternative NMR method
that allows the determination of enantiomeric excesses (ee)
up to 99% and higher. To realize this method, we wanted to
make use of a technique recently published by Claridge
et al. which had distinctly improved the sensitivity for the
quantification of diastereomeric excesses (de) by comparing
the integrals of the signals of the minor diastereomer with
(
(
Agilent, Waldbronn, Germany), and an SIL HTA Shimadzu autosampler
Shimadzu, Duisburg, Germany) in combination with an Astec
Cyclobond I 2000 DM (25 cm x 4.6 mm; 5 μm) stationary phase
Sigma-Aldrich) was used. Data analysis was performed by means of
(
Analyst 1.4.2 (Applied Biosystems, Darmstadt, Germany).
Preparation of stock solutions for the HPLC experiments. All
solutions and buffers were prepared in water unless stated otherwise.
For method development and validation we used 10 mM rac-1 · HCl stock
solutions. These stock solutions were prepared by dissolving rac-1 · HCl
in the appropriate volume of water or 7.3 mM TEAA (triethylammonium
acetate) pH 4.0, respectively. Analogously, a 10 mM stock solution of
13
1
the integral of C satellites of a suitable H NMR signal of
17
the major diastereomer. That way de of up to 99.8% could
be successfully determined. Our plan was to apply this
17
technique to the determination of ee of indatraline samples.
Actually, we considered this approach especially rewarding
as an 1H NMR-based method would, in addition, allow
detection of sample impurities, such as solvent remnants,
and recovery of the samples used.
(1R,3S)-1 · HCl in 7.3 mM TEAA pH 4.0 was prepared. All stock solutions
were frozen at -20°C, thawed on the day of the experiment, and finally
diluted in the corresponding sample solvent to obtain the samples for
method development and validation (calibration standards, QCs, spiked
samples).
Due to the fact that the N-methyl group gives rise to the
1
most intense signal in the H NMR spectrum of rac-1, it
was selected as the reference signal for the development of
the above-described 13C satellite method. As this secondary
methylamino moiety is directly attached to one of the two
chiral centers of 1, the chemical shift differentiation of this
group was expected to be especially sensitive to the different
nature of the diastereomeric ion pairs.
Various CSAs were tested to determine their effect on the
chemical shift differences between the signals of the
1
diastereomeric complexes of rac-1 in the respective H NMR
spectra.
As such, (S)-(-)-1,1´-binaphthyl-2,2´-diol (3) was selected as it
had been successfully applied to the determination of the
enantiopurity of the structurally closely related sertraline (2), a
1
8
method that had been established by Salsbury and Isbester
Fig. 1). Also, the well-established CSAs (S)-(+)-6-methoxy-α-
(
1
methyl-2-naphthaleneacetic acid (4), (S)-(+)-O-acetylmandelic
acid (5), (S)-(-)-α-methoxy-α-trifluoromethylphenylacetic acid
Chirality DOI 10.1002/chir
Fig. 2. Investigated CSAs for the H NMR spectroscopic enantiomeric
discrimination.

9
26
GRIMM ET AL.
[
(S)-6], and (R)-(+)-α-methoxy-α-trifluoromethylphenylacetic
6 3
TABLE 2. Influence of benzene-d and acetonitrile-d as
solvent on enantiomeric chemical shift difference
acid [(R)-6] were studied. With these CSAs, a series of
experiments was performed in which the molar ratios (CSA :
rac-1, Table 1), solvents (Table 2), and temperature (Table 3)
were varied to identify the best-suited reagent and optimal
conditions for the determination of the enantiopurity of the
enantiomers of 1.
Regarding the molar ratios between the individual CSA and
indatraline (1), the optimal value was defined as the ratio that
led to the largest enantiomeric chemical shift difference (ΔΔδ)
between the N-methyl signals of the diastereomeric complexes,
together with low line broadening of the respective signals,
which were measured as FWHM.
(
ΔΔδ) and FWHM
c
Molaar
ratio
ΔΔδ b
[ppm]
FWHM
[Hz]
No.
Solvent
CSA
1
2
3
4
5
6
benzene-d
6
3
3
5 : 1
1 : 1
1 : 1
5 : 1
1 : 1
1 : 1
0.065
0.013
0.023
0.005
0.006
0.024
1.7 1.8
1.2 1.2
1.2 1.2
1.7 1.7
1.0 1.0
0.9 0.9
5
(S)-6
3
acetonitrile-d
5
(S)-6
Standard conditions for these experiments were 25°C and 64 transients.
Molar ratio CSA: rac-1.
Chemical shift difference between the N-methyl signal of the diastereomeric
enantiomer-CSA complexes.
First value of the low frequency signal, second value of the high frequency
signal.
a
In brief, in chloroform-d at 25°C the best results were obtained
using CSA (S)-6 or (R)-6, at a molar ratio of 1:1 (Table 1 entries
b
1
4 and 20, respectively, 64 transients were taken) (Fig. 3). Under
these conditions, not only ΔΔδ values amounting to 0.046 to
.047 ppm were very high, but also the FWHM values in the
c
0
range of 1.3–1.5 Hz were very low.
Due to the fact that the diastereomeric complex of 1 and
TABLE 3. Influence of the temperature on enantiomeric
chemical shift difference (ΔΔδ) and FWHM using rac-1
and (S)-6 (1:1)
(
S)-6 is the mirror image of 1 and (R)-6, the ΔΔδ and FWHM
values should be the same; the slight difference of entry 14
compared to entry 20 (Table 1) can be explained by different
enantiopurities of the purchased CSAs used and a determination
imprecision originating from manual measurements of both
parameters. Actually, for (S)-6 and (R)-6, the ΔΔδ had reached
their maximum when 1.0 equivalent of the CSA was employed
and decreased when the molar ratio of the CSA was further
increased (entries 15–17 and 21–13, Table 1), which at the same
a
b
No.
T [°C]
ΔΔδ [ppm]
FWHM [Hz]
1
2
3
4
5
- 50
- 25
0
25
50
0.076
0.064
0.055
0.045
0.037
6.0
11.5
3.3
2.0
1.4
1.3
2.4
1.7
1.2
1.2
Standard conditions for these experiments were chloroform-d and 64
TABLE 1. Influence of different CSA and their molar ratio on
enantiomeric chemical shift difference (ΔΔδ) and
FWHM of rac-1
transients.
a
Chemical shift difference between the N-methyl signal of the diastereomeric
enantiomer-CSA complexes.
b
First value of the low frequency signal, second value of the high frequency
a
b
c
No.
CSA
Molar ratio
ΔΔδ [ppm]
FWHM [Hz]
signal.
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
3
0.5 : 1
1 : 1
2 : 1
0.014
0.026
0.044
0.068
0.078
0.007
0.013
0.013
0.019
0.027
0.033
0.027
0.039
0.046
0.043
0.044
0.036
0.024
0.035
0.047
0.042
0.042
0.036
0.9
1.0
1.1
1.5
2.2
1.1
1.2
1.2
1.6
1.3
1.7
1.7
1.7
1.3
1.5
1.8
1.5
1.5
1.6
1.4
1.5
1.8
1.6
0.9
1.0
1.1
1.4
1.9
1.1
1.2
1.2
1.8
1.3
1.8
2.1
1.9
1.4
1.5
1.8
1.5
1.9
1.9
1.5
1.5
1.8
1.6
time was accompanied by a worsening of the FWHM values. In
contrast, the ΔΔδ values either improved or reached a plateau
for the CSA 3-5 (see entries 7 and 8, Table 1) when the molar
ratios of the CSA as compared to rac-1 were increased. Using
(S)-6 and (R)-6 at a molar ratio of 1:1 (CSA : rac-1) gave the
highest ΔΔδ values for all tested CSAs in the respective molar
ratio range studied, except for CSA 3. Also, 3 when employed
in molar ratios of 2:1, 5:1, and 10:1, respectively (Table 1, entries
5 : 1
10 : 1
0.5 : 1
1 : 1
2 : 1
0.5 : 1
1 : 1
4
5
0
1
2
3
4
5
6
7
8
9
0
1
2
3
2 : 1
3
-5) gave rise to ΔΔδ values reaching or even surpassing the
(S)-6
0.5 : 1
0.75 : 1
1 : 1
1.25 : 1
1.5 : 1
2 : 1
0.5 : 1
0.75 : 1
1 : 1
1.25 : 1
1.5 : 1
2 : 1
optimal results observed for (S)-6 and (R)-6, but this approach
suffered from two major disadvantages. First, an overlap of the
relevant N-methyl signals with a signal resulting from one of
the methylene protons of rac-1 occurred which complicated
the analysis of these signals, but also the line broadening
become distinctly worse. Thus, compound 3 appeared to be less
suitable than (S)-6 and (R)-6 for use as a CSA for the analysis of
the enantiopurity of the enantiomers of rac-1.
(R)-6
In further experiments the aprotic solvents acetonitrile-d and
3
benzene-d were also tested for their suitability in chemical shift
6
experiments with 3, 5, and (S)-6 as CSA, but clearly turned out
to be less appropriate than chloroform-d (Table 2). Finally,
varying temperature experiments were performed for the CSA
Standard conditions for these experiments were chloroform-d, at 25°C and 64
transients.
a
Molar ratio CSA:rac-1.
(S)-6 in chloroform-d. According to the results of these
b
Chemical shift difference between the N-methyl signal of the diastereomeric
experiments, a temperature of 25°C represents a good
compromise, yielding a reasonably high ΔΔδ together with an
adequately low FWHM value (entry 4, Table 3).
enantiomer-CSA complexes.
c
First value of the low frequency signal, second value of the high frequency
signal.
Chirality DOI 10.1002/chir

EE DETERMINATION OF (1R,3S)-INDATRALINE
927
1
Fig. 3. Excerpt of the H NMR spectrum of rac-1 with (S)-6 in a molar ratio of 1:1 measured in chloroform-d at 25°C with 64 transients. Separated N-methyl sig-
nals for the determination of ΔΔδ as well as FWHM values are labeled.
In the end, the optimized method was subjected to testing
at a sample that had been obtained from multiple resolution
via crystallization of the racemic compound rac-1, which
was expected to be of exceptional high ee. Corresponding to
Parker and Taylor,19 the size of the ΔΔδ values in general
varies depending on the enantiomeric composition of the
analyte and the absolute configuration of the CSA employed.
Hence, differing ΔΔδ values had to be expected for this highly
enantioenriched sample depending on the enantiomer of the
CSA used [(S)-6 or (R)-6]. Therefore, the (1R,3S)-1 obtained
by resolution of rac-1 via crystallization (see Materials and
solutions could be used. But as we intended, for practical
reasons, to employ only low amounts of enantioenriched
samples obtained during the racemic resolution of rac-1,
this opportunity was rejected. Nevertheless, as the integral
of the N-methyl signal of the minor enantiomer was in any
case unequivocally lower than the integral of one of the 13
C
satellite signals of the N-methyl signal of the major enantiomer,
it can be stated that the ee of the analyzed sample must have
been higher than 98.9%. The ee being ≥ 98.9%, follows from
13
the fact that the natural occurrence of C amounts to 1.108%,
12
1
and for this reason the ratio between the integrals of the C- H
1
13
17
Methods) was analyzed by means of the optimized H NMR
NMR signal and a single C satellite signal is 178.5:1, which
compares to an integral of 0.56% (= major signal, equivalent to
method using (S)-6 and (R)-6 as CSA as well. The spectra
resulting from these studies revealed that (S)-6 is better
suited for the analysis of the enantiopurity of (1R,3S)-1, as
this CSA enantiomer gives higher values for ΔΔδ (0.080
ppm compared to 0.050 ppm for (R)-6). Moreover, when
13
100%, divided by 178.5) for a C satellite signal in relation to
the parent signal. Accordingly, for theoretical reasons, the ee must
12
1
amount to ≥ 98.9% if the integral of the C- H NMR signal of the
13
minor isomer is less than the integral of the C satellite signal of
the major isomer.
(
R)-6 was used as CSA the N-methyl signals of the two
enantiomers of 1 were overlapping, severely impeding their
separate integration. According to Claridge et al.,
In any case, the above-described method represents a
simple and quick tool for the analysis of the progress of
racemic resolutions of indatraline (1) up to even very high
enantiomeric excesses of ee ≤ 98.9%.
17
theoretically diastereomeric excesses of up to 99.8% de can
be determined provided an acceptable S/N ratio is reached.
Unfortunately, we were not able to reliably quantitate the
N-methyl signal of the (1S,3R)-enantiomer with 64
transients, as its integral was substantially lower than that
of the 13C satellite signal of the respective N-methyl signal
of the major isomer the (1R,3S)-1. Although an S/N of 12:1
HPLC Method Development and Optimization
With the aim of developing an analytical method suitable
for determination of ee ≥ 99% for both indatraline enantiomers,
we tried next to benefit from the capabilities of a chiral stationary
phase.
1
3
was found for one C satellite which did fulfill our key
criterion of S/N ≥10:1 for accurate quantification, the N-methyl
signal of the minor isomer almost vanished in the signal noise
for this real sample with high enantiopurity. Therefore, an S/N
ratio optimization was performed by incrementally increasing
the number of transients (128, 512, 2048, and 8192). But even
20
The method described by Rao et al. for the separation and
quantitation of stereoisomers and related enantiomeric impurities
of sertraline (2) by means of a dimethylated β-cyclodextrine
material as stationary phase was considered a suitable starting
point. The column employed in this study, an Astec Cylcobond I
2000 DM, seemed to be promising for our purpose due to the
structural analogy of indatraline (1) compared to sertraline (2).
The described mobile phase, however, employing 0.4% (v/v)
1
3
with 8192 transients [S/N 66:1 for one C satellite of the
N-methyl group of (1R,3S)-1], the enantiopurity of this
highly enantioenriched material of (1R,3S)-1 could not be
analyzed accurately, as slight changes of phase correction
performance had significant effects on the integration
values. To overcome this obstacle, more concentrated
20
trifluoroacetic acid (TFA) at pH 3.0 can be considered relatively
harsh and possibly destructive for the stationary phase as these
conditions are at or even not within the manufacturer’s advice
Chirality DOI 10.1002/chir

9
28
GRIMM ET AL.
for handling of the Astec Cylcobond I 2000 DM. Additionally, for
Several trends can be recognized from the results obtained
when varying the chromatographic conditions, which are shown
in Table 4a and 4b. First, the resolution of the enantiomers
21
TFA a memory effect was found by Ye et al. which was assumed
to impair a rapid switching between different kinds of buffers.
Hence, a gentle but powerful HPLC method avoiding TFA should
be developed for this stationary phase. First, in order to select an
appropriate wavelength for analyte detection, an absorption
spectrum of rac-1· HCl was recorded. Figure 4 shows the
absorption spectrum of a 25 μM aqueous solution of rac-
appeared to be sufficient for all conditions investigated (R ≥3,
s
allentriesTable4aand4b). Second, forbothbuffers(AAcandTEAA)
a decrease in buffer concentration led to shorter retention times
(Table 4a and 4b, entries 1–3, respectively) associated with
increased intensities (at least for AAc, Table 4a entries 1–3) and
lower peak widths (data not shown). Third, the opposite effect
(i.e., longer retention time associated with decreased intensities
and broader peaks) was observed when the pH value was
increased(Table4aand4b, entries 2and4–6, respectively). Fourth,
enhanced amounts of acetonitrile resulted in shorter retention
times, higher intensities, and improved peak shapes (i.e., lower As
Table 4a and 4b, entries 2 and 7–8, respectively). Fifth, decreased
temperatures caused a conspicuous peak broadening (data not
shown) and a marked loss of intensity due to longer retention
times (Table 4a and 4b, entries 2 and 9–11, respectively). Finally,
it should be mentioned that substitution of acetonitrile for
methanol led to markedly worse results regarding intensity,
resolution, run time, and peak shape (data not shown).
1
· HCl. It reveals a weak maximum at 265 nm and a steep
increase of absorption at a wavelength <240 nm with a
slight shoulder at about 230 nm.
Further spectra recorded in possible mobile phases such as
0 mM AAc pH 4.0, 7.3 mM TEAA pH 4.0, or 7.3 mM TEAA
1
pH 4.0 and acetonitrile (85:15, v/v), employing different
concentrations of rac-1 · HCl and varying pH values of the
buffers used as solvent (pH 3.0–7.0) covering the suggested
operation range of the Astec Cyclobond I 2000 DM confirmed
these absorption characteristics. These spectra were also
22
similar to those of 2 reported by Mandrioli et al., who
identified absorption maxima at 208 nm and 275 nm
(
conditions not specified), respectively. For the present
study, we decided to use 230 nm as the detection wavelength
Based on the conditions of entry 8 (Table 4b, i.e., 7.3 mM
TEAA pH 4.0 with acetonitrile in a ratio of 85:15 (v/v), flow
rate of 1.0 ml/min, temperature of 20°C) that were
considered a good compromise regarding intensity (at the
same noise level compared to AAc), resolution, and peak
shape, the effect of enhanced injection volumes was
investigated. Thus, to gain higher intensities, the original
injection volume of 10 μl of the 1 mM sample solution of
rac-1 · HCl in water was raised to 50 μl.
The results shown in Table 5 document a proportional
increase of intensities in relation to the injection volume at
acceptable resolution and asymmetry factors for both
enantiomers (Table 5, entries 1–3) the highest intensity thus
being observed for the injection volume of 50 μl. But as peak
splitting of the (1S,3R)-1 peak of rac-1·HCl samples in water
and in the mobile phase with concentrations > 1 mM occurred,
additional experiments to optimize the sample solvent were
performed. We found that by using 7.3 mM TEAA pH 4.0 as
sample solvent the situation distinctly improved, giving higher
intensities for (1S,3R)-1 and better peak shape for (1R,3S)-1
(Table 5, entry 4), with the remaining analytical characteristics
unchanged. Thus, in this way, peak splitting of the (1S,3R)-
enantiomer could be suppressed up to concentrations of 1.5
mM of rac-1· HCl. A representative chromatogram recorded
after injection of 50 μl of a 1 mM rac-1· HCl sample in 7.3 mM
TEAA pH 4.0 under these optimized conditions is shown in
Figure 5a. The chromatographic parameters calculated for this
sample are shown in Table 5.
2
2
for the HPLC analysis, just as Mandrioli et al. had done in
the case of 2, as it provides high analytical sensitivity and at
the same time is still sufficiently distant from the cutoff of
the mobile phase.
To develop the required HPLC separation method for the
enantiomers of indatraline (1) based on the selected Astec
Cylcobond I 2000 DM column, we started with the
chromatographic conditions suggested by the manufacturer
for separation of racemic hydrobenzoine. When these were
employed [i.e., 10 mM AAc buffer pH 4.0 and acetonitrile
9
0:10 (v/v)] injecting 10 μl of a 1 mM aqueous solution of
rac-1 · HCl, at a flow rate of 1.0 ml/min and a temperature
of 20°C, a reasonable separation for the two enantiomers of
indatraline (R = 3.95) was obtained (see Table 4a, entry 2).
s
As discrimination of enantiomers by means of cyclodextrine
phases is known to depend on the nature of the buffer, its
concentration and pH value, on the organic modifier as well
2
3
as on temperature, we investigated the influence of these
parameters on separation, run time, peak symmetry, and
peak intensity [defined as peak height of (1S,3R)-1]. Regarding
the latter, the peak resulting from the (1S,3R)-1 enantiomer
was considered to be crucial due to its longer retention time
and broader peak width.
These results were considered appropriate for a reliable
quantitation of high enantiomeric excesses (i.e., >99% ee)
for both enantiomers of indatraline. Therefore, we intended
to verify the validity of the method used based on an Astec
Cyclobond I 2000 DM as stationary phase (25 x 4.6 cm, 5 μm),
a mobile phase composition of 7.3 mM TEAA pH 4.0, and
acetonitrile (85:15, v/v) at a flow rate of 1.0 ml/min at 20°C
and an injection volume of 50 μl of the samples dissolved in
7.3 mM TEAA pH 4.0 in the next step.
Validation
Fig. 4. Absorption spectrum, corrected by subtraction of the correspond-
ing absorption spectrum of the sample solvent, of rac-1 · HCl (25 μM in water)
in the range from 200–300 nm with an enlargement of the slight shoulder at
about 230 nm.
When a single enantiomer is selected for development as a
new drug, the other enantiomer, i.e., the minor isomer,
should be analyzed in the same manner as required for other
Chirality DOI 10.1002/chir

EE DETERMINATION OF (1R,3S)-INDATRALINE
929
TABLE 4a. Effect of concentration and pH value of ammonium acetate buffer, organic modifier and temperature on intensity, reso-
lution, retention times, and asymmetry factors of the indatraline enantiomers
t
R
[min]
A
s
a
b
No.
Conditions
Peak height [mAU]
R
s
1R,3S
1S,3R
1R,3S
1S,3R
Buffer conc. [mM]
1
2
3
5
*10
15
135.00
93.83
78.10
3.51
3.95
4.24
7.64
11.00
13.33
9.55
13.90
16.97
2.43
2.32
2.22
2.40
2.40
2.72
pH
3.5
*4.0
4.5
4
2
5
6
115.33
93.83
89.00
3.87
3.95
3.99
4.06
9.53
11.00
12.40
11.94
13.90
15.70
1.71
2.32
2.60
1.59
2.40
2.74
3.03
5.0
70.33
15.10
19.23
2.96
c
A : B ratio
95:5
*90:10
85:15
T [°C]
5
7
2
8
37.80
93.83
148.33
4.57
3.95
3.17
19.30
11.00
6.96
26.33
13.90
8.20
3.31
2.32
1.69
2.94
2.40
1.62
9
1
1
63.83
60.30
79.27
3.90
3.98
4.02
3.95
12.33
12.43
11.53
16.14
16.13
14.80
2.19
2.33
2.38
2.32
1.88
2.23
2.14
2.40
0
1
10
15
*20
2
93.83
11.00
13.90
Highlighted entries represent starting conditions, during variation of each parameter (e.g., concentration, pH, A : B ratio, temperature) the three remaining
parameters were kept constant.
a
1
0 μl of an aqueous 1 mM rac-1 · HCl solution injected on an Astec Cyclobond I 2000 DM (25 cm x 4.6 mm, 5 μm).
b
(
1S,3R)-1.
c
Ammonium acetate buffer (A):acetonitrile (B) (v/v).
defined impurities.24 The ICH Q2(R1) guidance,16 which
addresses this topic, recommends quite generally “to
demonstrate that it (i.e., the developed method) is suitable
for its intended purpose.” However, the ICH Q2(R1) guidance
contains almost no exactly defined (i.e., quantitative)
requirements regarding validation parameters for methods
to determine enantiopurity. For the intended purpose of the
described method it is mandatory for the accurate and precise
TABLE 4b. Effect of concentration and pH value of triethylammonium acetate buffer, organic modifier and temperature on intensity,
resolution, retention times, and asymmetry factors of the indatraline enantiomers
t
R
[min]
A
s
a
b
No.
Conditions
Peak height [mAU]
R
s
1R,3S
1S,3R
1R,3S
1S,3R
d
Buffer conc. [mM]
1
2
3
3.65
*7.30
14.60
pH
133.00
134.67
105.33
3.28
3.66
4.08
6.94
8.66
12.63
8.65
10.87
16.2
2.20
2.49
2.63
2.16
2.50
2.63
4
2
5
6
3.5
*4.0
4.5
5.0
A: B ratio
95:5
*90:10
85:15
T [°C]
5
135.33
134.67
101.67
80.20
3.68
3.66
3.78
3.82
8.18
8.66
10.57
13.00
10.30
10.87
13.47
1.86
2.49
2.58
2.87
1.89
2.50
2.88
2.94
16.60
c
7
2
8
58.40
134.67
176.33
3.99
3.66
3.00
15.33
8.66
5.97
20.90
10.87
7.03
4.00
2.49
2.25
4.40
2.50
2.07
9
1
1
85.17
116.00
126.67
3.72
3.68
3.67
3.66
10.37
9.94
9.64
13.63
12.90
12.37
2.44
2.67
2.60
2.49
2.47
2.70
2.78
2.50
0
1
10
15
*20
2
134.67
8.66
10.87
Highlighted entries represent starting conditions, during variation of each parameter (e.g., concentration, pH, A : B ratio, temperature) the three remaining
parameters were kept constant.
a
1
0 μl of an aqueous 1 mM rac-1 · HCl solution injected on an Astec Cyclobond I 2000 DM (25 cm x 4.6 mm, 5 μm).
b
(
1S,3R)-1.
c
Triethylammonium acetate buffer (A):acetonitrile (B) (v/v).
d
Preparation of triethylammonium acetate buffer by diluting triethylamine in water and pH adjustment using glacial acetic acid (7.3 mM are equivalent to 0.1% (v/v)
triethylamine in water).
Chirality DOI 10.1002/chir

9
30
GRIMM ET AL.
TABLE 5. Effect of the injection volume on intensity, resolution, retention time, and asymmetry factors of the indatraline enantio-
*
mers based on the chromatographic conditions of entry 8 table 4b
Peak b
height
[mAU]
t
R
[min]
A
s
Sample
solvent
Injectiona
volume
R
s
No.
1R,3S
1S,3R
1R,3S
1S,3R
1
2
3
4
water
water
water
10 μl
20 μl
50 μl
50 μl
176.33
338.67
778.67
934.83
3.00
2.95
2.74
2.69
5.97
5.95
5.91
5.72
7.03
6.99
6.90
6.64
2.25
2.75
3.28
1.81
2.07
2.09
2.81
3.18
7.3 mM TEAA
a
1
mM rac-1 · HCl injected.
b
(
1S,3R)-1.
*
7.3 mM triethylammonium acetate pH 4.0 and acetonitrile (85:15, v/v) as mobile phase, at a flow rate of 1.0 ml/min at 20°C using an Astec Cylclobond I 2000 DM
(
25 cm x 4.6 mm, 5 μm) as stationary phase.
Fig. 5. Chromatograms of (a) 1 mM rac-1 · HCl, (b) sample solvent blank (7.3 mM triethylammonium acetate pH 4.0, (c) 500 μM of (1R,3S)-1 · HCl
(
R
obtained by racemate resolution, see Materials and Methods), an asterisk denotes t of the (1S,3R)-enantiomer. All chromatograms were recorded
employing a mobile phase of 7.3 mM triethylammonium acetate pH 4.0 and acetonitrile (85:15, v/v), at a flow rate of 1.0 ml/min at 20°C on an Astec
Cylclobond I 2000 DM (25 cm x 4.6 mm, 5 μm).
Chirality DOI 10.1002/chir

EE DETERMINATION OF (1R,3S)-INDATRALINE
931
determination of both enantiomers – even if one is present
in vast excess – that the range covered by the method is
as large as possible. Following these considerations, we
tried to demonstrate the performance of the established
method for both indatraline enantiomers examining rele-
vant validation parameters [i.e., range, quantitation limit
Accuracy – precision. Accuracy (i.e., recovery) and precision
(i.e., repeatability) were examined for quality control samples
(QCs) at five concentration levels (i.e., 2.5, 10, 50, 500, and
1500 μM) of rac-1 · HCl, each in replicates of five. Table 6
shows the recovery (ratio of experimentally determined
concentration to nominal concentration ± confidence interval,
α = 0.05) and the repeatability (given as relative standard
deviation, RSD) calculated from the experimentally obtained
areas by means of the respective calibration functions. The
developed method yielded accurate and precise results with
recoveries between 98.6 ± 0.4% and 101.0 ± 1.0% and RSDs
between 0.3% and 1.4% for (1R,3S)-1, and recoveries of
98.9 ± 0.3% to 100.5 ± 1.0% with RSDs of 0.2% to 1.1% for
(1S,3R)-1. As all recoveries and RSDs are distinctly within
the commonly accepted limits (recovery: 98.0%–102.0%, RSD
(
QL), linearity, precision, i.e., repeatability, accuracy, i.e.,
recovery, specificity] for both indatraline enantiomers by
investigation of rac-1 · HCl samples (due to the fact that
no certified enantiopure material was available) over a
broad concentration range.
Specificity. Injection of a blank sample (i.e., 7.3 mM TEAA
pH 4.0) led to the chromatogram shown in Figure 5b. The
absence of peaks at the retention times expected for the
indatraline enantiomers indicates that there are no
substances interfering with the indatraline enantiomers.
Also, the resolution for the separation of the enantiomers
25
≤ 2.0% ), the established method is clearly in agreement with
the ICH recommendations for accuracy (i.e., recovery) and
precision (i.e., repeatability).
(
Rs = 2.69, see Fig. 5a) could be considered sufficient, thus
Considering the results of the validation based on rac-1·HCl
samples, characterization of enantiopurities up to 99.75% ee [i.e.,
1000 μM (1R,3S)-1 in presence of 1.25 μM (1S,3R)-1] for
demonstrating that this method is specific for both
enantiomers.
(1R,3S)-1 and 99.67% ee [i.e., 750 μM (1S,3R)-1 in presence of
Quantitation limit (QL). The S/N recorded for five 2.5 μM
calibration standards of rac-1 · HCl was calculated by means
of Analyst 1.4.2 software. It was found to be ≥ 23 for (1R,3S)-1
and ≥ 18 (1S,3R)-1,. This means that the ICH recommendation
regarding QL (S/N >10) is fulfilled for both enantiomers at a
concentration of 1.25 μM.
1.25 μM (1R,3S)-1] for (1S,3R)-1 should be possible, when
enantioenriched samples are analyzed at the upper concentration
of the enantiomer representing the major isomer of the sample.
In the next step, we employed this method to study a
sample of (1R,3S)-1, that was expected to possess an
extremely high enantiopurity in the range of the analytical
limit of the aforementioned method that had become
available to us by resolution of rac-1 via crystallization using
L-(+)-tartaric acid as auxiliary reagent (see Materials and
Methods) (chromatogram shown in Fig. 5c). As a (1R,3S)-1
reference standard was still not available, the obtained
(1R,3S)-1 · HCl material was analyzed employing the
calibration function established for rac-1 · HCl.
The (1R,3S)-1· HCl samples were investigated in a nominal
concentration of 1000 μM in five replicates. The concentrations
of both enantiomers calculated by means of the aforementioned
calibration functions were 979 μM for (1R,3S)-1 and 2.28 μM for
(1S,3R)-1, indicating an ee of 99.5%. The concentrations obtained
in this experiment indicated a recovery of 98.1 ± 0.3%
(mean ± confidence interval, α = 0.05) related to the sum of
both enantiomers. The small but nevertheless distinctly
lower recovery determined for the (1R,3S)-1 · HCl sample
in comparison to the total recoveries of rac-1 · HCl samples
[e.g., 98.8% for 1500 μM rac-1 · HCl, i.e., 98.6% for 750 μM
(1R,3S)-1 and 98.9% 750 μM (1S,3R)-1, see Table 6] is
presumably due to a small amount of ethyl acetate present
Linearity – range. From the resulting peak areas obtained for
both enantiomers, calibration functions were generated by
2
means of linear regression. Employing a 1/x weighting to the
calibration functions gave best results regarding a concentration
range as broad as possible. In this way, a calibration function of
2
y = 29.55x - 1.731 was found for the (1R,3S)-1 (r = 0.9995 and a
negligible y-intercept corresponding to 4.38% of the area obtained
25
at QL comply with the requirements for linearity ) for a range of
.25–1000 μM (i.e., 2.5–2000 μM rac-1·HCl). For (1S,3R)-1, this
1
2
led to y=29.63x – 1.623 (r = 0.9999, y-intercept corresponds to
.62% of the area obtained at QL) as calibration curve for a range
4
of 1.25–750 μM (i.e., 2.5–1500 μM rac-1· HCl). Regarding the
upper concentrations of the range, it should be recalled that for
(
1S,3R)-1 a tendency to peak splitting had been observed when
samples of rac-1· HCl with concentrations ≥2000 μM had been
injected (see above). Furthermore, the injection of calibration stan-
dards containing 3000 μM rac-1· HCl (see above) appeared to
overload the column, resulting in peak broadening and
decreased resolution.
TABLE 6. Validation results (mean values, n = 5) regarding accuracy (recovery ± confidence interval, CI, α = 0.05) and precision
relative standard deviation, RSD) for five different QC levels of Indatraline
(
Concentration [μM]
Found
Recovery [%] ± CI
RSD [%]
Nominal
1R,3S
1S,3R
1R,3S
1S,3R
1R,3S
1S,3R
1
5
5
50
50
.25
1.26
5.04
25.1
252
1.26
5.02
25.0
251
101.0 ± 1.0
100.9 ± 1.3
100.3 ± 0.3
100.8 ± 0.2
98.6 ± 0.4
100.5 ± 1.0
100.3 ± 0.3
100.0 ± 0.3
100.5 ± 0.2
98.9 ± 0.3
1.1
1.4
0.29
0.25
0.46
1.1
0.32
0.28
0.19
0.34
2
2
7
740
742
Chirality DOI 10.1002/chir

9
32
GRIMM ET AL.
TABLE 7. Validation results (mean values, n = 5) regarding accuracy (recovery ± confidence interval, CI, α = 0.05) and precision
(
relative standard deviation, RSD) after spiking a 1 mM solution of highly enantiomerically enriched (1R,3S)-1 · HCl (99.5% ee) with
small volumes of a 1 mM rac-1 · HCl solution
Concentration [μM]
ee [%]
c
Expected
Found
Recovery [%] ± CI
RSD [%]
d
Spiking
1R,3S
1S,3R
1R,3S
1R,3S
1R,3S
1S,3R
1R,3S
1S,3R
expected
found
a
1b
2
977
975
4.77
6.26
970
967
4.72
6.21
99.3 ± 0.3
99.2 ± 0.4
99.0 ± 1.0
99.2 ± 0.9
0.28
0.41
1.1
1.0
99.0
98.7
99.0
98.7
a
Sample 1: spiking 995 μL of a 981 μM (1R,3S)-1 · HCl (99.5% ee) solution with 5 μL 1 mM rac-1 · HCl (both in 7.3 mM TEAA pH 4.0).
b
Sample 2: spiking 992 μL of a 981 μM (1R,3S)-1 · HCl (99.5% ee) with 8 μL 1 mM rac-1 · HCl to (both in 7.3 mM TEAA pH 4.0).
c
Expected concentration according to the experimentally determined concentration of 981 μM for the enantiomenriched (1R,3S)-1 · HCl sample (99.5% ee) and the
added amount of rac-1 · HCl.
d
Expected enantiomeric excesses calculated from the expected concentrations.
in the respective (1R,3S)-1 · HCl material. Unfortunately,
this small amount of ethyl acetate — unambiguously
identified by H NMR (≤1.0%) in the respective (1R,3S)-1·HCl
analysis of highly enantioenriched samples, an HPLC method
was developed that could be shown to enable accurate and pre-
cise quantitation of ee up to 99.75% for the (1R,3S)-enantiomer
of indatraline (eutomer) and 99.67% for the (1S,3R)-enantiomer
(distomer). Additionally, the results of HPLC method develop-
ment show that ESI-MS detection employing a mobile phase
based on ammonium acetate buffer is possible as well. The reli-
ability of the established method was confirmed by validation
according to the ICH guideline Q2(R1). As indatraline is a prom-
ising triple reuptake inhibitor, it is worth mentioning that the
presented HPLC method represents a prerequisite for a correct
pharmacological characterization of the pure enantiomers of
indatraline and their possible use as pharmacological tools.
1
sample — could not be completely removed even after intensive
high-vacuum treatment. That EtOAc has been found in this
sample, of (1R,3S)-1· HCl, but not in those of rac-1·HCl is
certainly a result of the different procedures used for the
preparation of these compounds, (1R,3S)-1 · HCl having
been recrystallized from EtOAc, whereas rac-1 · HCl has
been isolated by freeze-drying of an aqueous solution in the
final step (see Materials and Methods).
Since validation was initially based on rac-1·HCl samples, we
additionally tried to validate accuracy (i.e., recovery) and
precision (i.e., repeatability) of the method for their application
to highly enantioenriched (1R,3S)-1·HCl samples by means of
spiking experiments, which should ensure that even slight
changes in the ee of highly enriched (1R,3S)-1·HCl samples
can be reliably quantified. To this end, we spiked the above-
mentioned 981 μM (1R,3S)-1·HCl sample of 99.5% ee [979 μM
ACKNOWLEDGMENTS
We thank Gerd Bauschke for performing the synthesis and
resolution of indatraline via crystallization (LMU München,
Department of Pharmacy).
(1R,3S)-1 and 2.28 μM (1S,3R)-1], with different volumes of a 1
mM rac-1· HCl solution in replicates of five and determined
the resulting concentrations of (1R,3S)-1 and (1S,3R)-1, by
means of the calibration functions established for the racemic
sample. The results from this experiment are shown in Table 7.
For both of the spiked samples the expected concentrations
calculated from the concentrations of the original solutions [977
μM and 975 μM for (1R,3S)-1 and 4.77 μM and 6.26 μM for
LITERATURE CITED
1
. Bøgesø KP. Novel 3-phenyl-1-indamines, pharmaceutical compositions
and methods of preparation. 1982. European Patent Office (EP-A1-0 076
669).
2. Bøgesø KP, Christensen AV, Hyttel J, Liljefors T. 3-Phenyl-1-indanamines.
Potential antidepressant activity and potent inhibition of dopamine,
norepinephrine, and serotonin uptake. J Med Chem 1985;28:1817–1828.
3
. Richelson
E Pharmacology of antidepressants. Mayo Clin Proc
(1S,3R)-1] matched those found experimentally almost exactly
2001;76:511–527.
[
970 μM and 967 μM for (1R,3S)-1 and 4.72 μM and 6.21 μM
4
. Skolnick P, Popik P, Janowsky A, Beer B, Lippa AS. "Broad spectrum"
antidepressants: Is more better for the treatment of depression? Life Sci
for (1S,3R)-1]. To this end, the found recoveries of ≥99.0% and
RSDs ≤1.1% are also within the commonly accepted limits
2003;73:3175–3179.
25
(recovery: 98.0%–102.0%, RSD ≤2.0%) ; additionally, a perfect
5
6
7
. Guiard BP. A new class of antidepressant drugs in the treatment of psychiatric
disorders: the triple reuptake inhibitors. In: Uehara T, editor. Psychiatric
disorders. Intech Open; 2011. 291–316.
. Hyttel J, Larsen J-J. Neurochemical profile of Lu 19-005, a potent inhibitor
of uptake of dopamine, noradrenaline, and serotonin. J Neurochem
agreement between expected and found enantiomeric excesses
could be deduced from the performed spiking experiments. As
consequence, the developed method can be considered reliable
for the enantiopurity determination of the triple reuptake
inhibitor (1R,3S)-indatraline.
1985;44:1615–1622.
. Negus SS, Brandt MR, Mello NK. Effects of the long-acting monoamine
reuptake inhibitor indatraline on cocaine self-administration in rhesus
monkeys. J Pharmacol Exp Ther 1999;291:60–69.
8. Davies HML, Gregg TM. Asymmetric synthesis of (+)-indatraline using
rhodium-catalyzed C-H activation. Tetrahedron Lett 2002;43:4951–4953.
. Takatsu K, Shintani R, Hayashi T. Copper-catalyzed 1,4-addition of
organoboronates to alkylidene cyanoacetates: mechanistic insight and applica-
tion to asymmetric catalysis. Angew Chem Int Ed Engl 2011;50:5548–5552.
0. Taylor JG, Correia CRD. Stereoselective synthesis of unsymmetrical β,β-
diarylacrylates by a Heck – Matsuda reaction: versatile building blocks
for asymmetric synthesis of β,β-diphenylpropanoates, 3-aryl-indole, and
CONCLUSION
1
Even though enantiodiscrimination by H NMR spectroscopy
is a well-established approach, this work demonstrates its limits
for indatraline concerning the determination of low amounts of
the minor enantiomer in samples containing the major enantio-
9
1
mer in large excess. Nevertheless, the established H NMR
1
method may serve as tool to monitor the progress of a resolution
of indatraline enantiomers. For a more precise and sensitive
Chirality DOI 10.1002/chir

EE DETERMINATION OF (1R,3S)-INDATRALINE
933
1
4
-aryl-3,4-dihydro-quinolin-2-one and formal synthesis of (-)-indatraline. J
18. Salsbury JS, Isbester PK. Quantitative H NMR method for the routine
spectroscopic determination of enantiopurity of active pharmaceutical in-
gredients fenfluramine, sertraline, and paroxetine. Magn Reson Chem
Organ Chem 2011;76:857–869.
1. Turner HM, Patel J, Niljianskul N, Chong JM. Binaphthol-catalyzed asymmetric
conjugate arylboration of enones. Org Lett 2011;13:5796–5799.
2. Wei W-T, Yeh J-Y, Kuo T-S, Wu H-L. Highly enantioselective rhodium-
catalyzed asymmetric 1,4-addition reactions of arylboronic acids to acyclic
α,β-unsaturated compounds: the formal synthesis of (-)-indatraline. Chem
Eur J 2011;17:11405–11409.
3. Yoo K, Kim H, Yun J. Asymmetric synthesis of 1,1-diarylalkyl units by a
copper hydride catalyzed reduction: differentiation between two similar
aryl substituents. Chem Eur J 2009;15:11134–11138.
1
2005;43:910–917.
1
1
9. Parker D, Taylor RJ. Direct H NMR Assay of the enantiomeric composition of
amines and β-amino alcohols using o-acetyl mandelic acid as a chiral solvating
agent. Tetrahedron 1987;43:5451–5456.
20. Rao RN, Talluri MVN, Maurya PK. Separation of stereoisomers of sertraline
and its related enantiomeric impurities on a dimethylated β-cyclodextrin
stationary phase by HPLC. J Pharm Biomed Anal 2009;50:281–286.
1
1
21. Ye YK, Lord B, Stringham RW. Memory effect of mobile phase additives
in chiral separations on a Chiralpak AD column. J Chromatogr A
1
4. Roesner S, Casatejada JM, Elford TG, Sonawane RP, Aggarwal VK.
Enantioselective syntheses of (+)-sertraline and (+)-indatraline using
2002;945: 139–146.
lithiation/borylation-protodeboronation
011;13:5740–5743.
5. Kuss HJ, Stauffer D. Evaluating chromatograms. In: Kuss HJ, Kromidas S,
methodology.
Org
Lett
22. Mandrioli R, Saracino MA, Ferrari S, Berbardi D, Kenndler E, Raggi MA.
HPLC analysis of the second-generation antidepressant sertraline and its
main metabolite N-desmethylsertraline in human plasma. J Chromatogr
B Analyt Technol Biomed Life Sci 2006;836:116–119.
2
1
1
editors. Quantification in LC and GC. Hoboken, NJ: Wiley-VCH; 2009.
3
–8.
23. Astec. Cyclobond handbook: a guide to using cyclodextrin bonded phases
for chiral LC separation. 2002.
24. Hadley MR. Stereochemical impurities. In: Smith RJ, Webb ML, editors.
6. International Conference on Harmonisation of Technical Requirements for
Registration of Pharmaceuticals for Human Use; ICH Harmonised Tripartite
Guideline; Validation of Analytical Procedures: Text and Methodology Q2
Analysis of drug impurities. New York: Blackwell Publishers; 2007. p 47–81.
(
R1); Current Step 4 version (2005).
2
5. Ermer J. Fundamentals of validation in pharmaceutical analysis. In: Ermer J, M
JHMcB, editors. Method validation in pharmaceutical analysis. Hoboken, NJ:
Wiley-VCH; 2005. 15–16, 81.
1
7. Claridge TDW, Davies SG, Polywka MEC, Roberts PM, Russel AJ, Savory
ED, Smith AD. "Pure by NMR"? Org Lett 2008;10:5433–5436.
Chirality DOI 10.1002/chir