Identification, characterization and HPLC quantification of process-related impurities in Trelagliptin succinate bulk drug: Six identified as new compounds

Article abstract of DOI:10.1016/j.jpba.2016.04.041

A sensitive, selective and stability indicating reversed-phase LC method was developed for the determination of process related impurities of Trelagliptin succinate in bulk drug. Six impurities were identified by LC-MS. Further, their structures were char

Full text of DOI:10.1016/j.jpba.2016.04.041

Journal of Pharmaceutical and Biomedical Analysis 128 (2016) 18–27
Contents lists available at ScienceDirect
Journal of Pharmaceutical and Biomedical Analysis
journal homepage: www.elsevier.com/locate/jpba
Identification, characterization and HPLC quantification of
process-related impurities in Trelagliptin succinate bulk drug: Six
identified as new compounds
Hui Zhanga, Lili Sunb, Liang Zoua, Wenkai Huia, Lei Liua, Qiaogen Zoua,∗
,
Pingkai Ouyangb
a School of Pharmaceutical Sciences, Nanjing Tech. University, Nanjing 210009, PR China
b College of Biotechnology and Pharmaceutical Engineering, Nanjing Tech. University, Nanjing 210009, PR China
a r t i c l e i n f o
a b s t r a c t
Article history:
A sensitive, selective and stability indicating reversed-phase LC method was developed for the determi-
nation of process related impurities of Trelagliptin succinate in bulk drug. Six impurities were identified
by LC–MS. Further, their structures were characterized and confirmed utilizing LC–MS/MS, IR and NMR
spectral data. The most probable mechanisms for the formation of these impurities were also discussed. To
the best of our knowledge, six structures among these impurities are new compounds and have not been
reported previously. The superior separation was achieved on an InertSustain C18 (250 mm × 4.6 mm,
5 ␮m) column in a gradient mixture of acetonitrile and 20 mmol potassium dihydrogen phosphate with
0.25% triethylamine (pH adjusted to 3.5 with phosphate acid). The method was validated as per regulatory
guidelines to demonstrate system suitability, specificity, sensitivity, linearity, robustness, and stability.
© 2016 Elsevier B.V. All rights reserved.
Received 11 January 2016
Received in revised form 26 April 2016
Accepted 27 April 2016
Available online 10 May 2016
Keywords:
Trelagliptin succinate
Process related impurity
LC–MS
NMR
Quantification
Structural elucidation
1. Introduction
In literature, limited information is available interested in
the analysis of TRE and its process-related impurities. Bao et al.
Trelagliptin
1-piperidinyl]-3,4-dihydro-3
succinate
(TRE),
2-[[6-[(3R)-3-amino-
described an RPLC method for quantification of eight related sub-
stances in TRE and its preparation using gradient elution however
without qualitative identification [5]. Neither formation nor struc-
tural elucidation for process-related impurities could be found in
current literature survey. Thus, there is a need to elaborate the
impurity profiles of TRE for industry and research reference. Simul-
taneously, a sensitive, selective and stability indicating analytical
method has been developed to monitor the levels of impurities in
TRE bulk drug.
There are several synthesis routes of TRE in the previous liter-
ature [6–9]. Some of these methods [6–8] may generate positional
isomers as process-impurities, because (R) −3- aminopiperidine
dihydrochloride contains a primary amine and a secondary amine
which are arenucleophilic attack groups. In Ref. [9], when (R)-
3-Boc-aminopiperidine as starting material with primary amine
protected to avoid generating positional isomers impurities. So the
synthetic route shown in Fig. 1 was adopted in laboratory sam-
ple preparation, and after analysis of different laboratory batches
of TRE, six impurities were detected in the range of 0.04–0.13%.
Impurities including unreacted starting materials, intermediates,
and by-product inevitably form at the end of this process, and
−methyl-2,4-dioxo-1(2H)-
pyrimidinyl] methyl]-4-fluoro-benzonitrile, is a highly selective
and long-acting Dipeptidyl peptidase IV (DPP-4) inhibitor that is
used in the treatment of Type 2 Diabetes [1]. It controls blood
glucose levels by selectively and continually inhibiting DPP-4, an
enzyme that causes the inactivation of glucagon-like peptide-1,
glucagon dependent insulinotropic polypeptide and incretin hor-
mones that play an important role in blood glucose regulation. The
inhibition of DPP-4 increases insulin secretion depending on blood
glucose concentration, accordingly controlling blood sugar levels.
In addition, TRE marketed in 2015 is the first long-acting oral
hypoglycemic agents with single weekly dosing, and compared to
once-daily short-acting product on the market, it is expected to
improve the convenience and adherence of patients significantly
[2–4].
∗
Corresponding author at: Nanjing Tech. University, 5 Xinmofan Road, Nanjing
210009, Jiangsu Province, PR China.
E-mail address: qiaogenzou@hotmail.com (Q. Zou).
http://dx.doi.org/10.1016/j.jpba.2016.04.041
0731-7085/© 2016 Elsevier B.V. All rights reserved.

H. Zhang et al. / Journal of Pharmaceutical and Biomedical Analysis 128 (2016) 18–27
19
Fig. 1. The synthesis route of TRE.
the presence of impurities exceeding the accepting limit of 0.1%
2.2. HPLC instrumentation and methods
may have an impact on the quality and safety of the drug product
[10,11]. Therefore, it is necessary for identification and character-
ization of the process-related impurities in TRE via NMR, MS and
related techniques.
This study aims to: (1) optimize LC conditions and achieve a
reliable LC method for the quantitative determination and analysis
of the impurities in TRE bulk drug; (2) characterize and confirm
the structures of these impurities by MS, IR and NMR; (3) obtain
mechanisms for the origin and formation of these impurities with
the respect to the knowledge of chemical synthesis.
All chromatographic experiments were carried out on a Waters
2695 system (Waters Technologies, MA, USA) equipped with a
2996 photo diode array detector. The system control, data acquisi-
tion and processing was accomplished by Empower data-handling
system. The separation was achieved on an InertSustain C18
(250 mm × 4.6 mm, 5 ␮m) column (Tokyo, Japan). Mobile phase A
consisting of 20 mmol potassium dihydrogen phosphate and 0.25%
triethylamine (pH adjusted to 3.5 with phosphate acid) and mobile
phase B of ACN were pumped at a flow rate of 1.0 mL/min. The gra-
dient program was set as follows: Time (min)/A: B (v/v); T0 87/13,
T8 87/13, T35 60/40, T60 40/60, T65 40/60, T66 87/13, T75 87/13. The
injection volume and detection was fixed at 20 ␮L and 230 nm. The
column temperature was maintained at 30 ◦C.
2. Experimental
2.1. Chemicals and reagents
TRE and standards of 2-(6-chloro-3-methyl-2,4-dioxo-
2.3. LC–MS instrumentation and methods
3,4-dihydro-
2H-pyrimidin-1-ylmethyl)-4-fluoro-benzonitrile
(Imp-A), N-[(3R)-1-[3-[(2-cyano-5- fluorophenyl) methyl]-1,2,3,6-
tetrahydro-1-methyl −2,6-oxo-4-pyrimidinyl]-3- piperidinyl]
-carbamate-(1,1-dimethylethyl)ester (Imp-B), (R)-3-methyl-1-(3-
LC–MS analysis was performed on an API4000 mass spectrom-
eter (Milwaukee, WI, USA) coupled to an Agilent 1100-LC system
(Palo Alto, CA, USA) in positive or negative APCI mode. The Anal-
ysis of all compounds was carried out on an InertSustain C18
(250 mm × 4.6 mm, 5 ␮m) column (Tokyo, Japan) using 1 mL/min
flow rate. The gradient elution employed solution A and B as
mobile phase components. Mobile phase A was 20 mmol ammo-
nium acetate buffer (pH adjusted to 3.5 with glacial acetic acid),
while mobile phase B was ACN. The gradient program was the same
as the HPLC chromatographic conditions described in Section 2.2.
The product MS spectra were collected over the m/z range from
100 to 800 Da with the following conditions: Ion spray voltage,
4200 V; declustering potential, 70 V; entrance potential, 10 V; turbo
ion spray temperature, 400 ◦C; collision energy, 25 V; and inter-
face heater, on. All date were acquired and processed by Analyst
Software workstation, version 1.5.2.
fluorobenzyl)-6-(3-amino-
dihydro-2H-pyrimidine
piperidin-1-yl)-2,
(Imp-C), 2-(3-methyl-2,4,6-oxo
4-dioxo-3,4-
−tetrahydro-2H-pyrimidin-1-yl methyl)-4-fluoro- benzonitrile
(Imp-D), (R)-2-[6-(3-amino-piperidin-1-yl)-3-methyl-2,4-dioxo-
3,4- dihydro-2H-pyrimidin −1-yl methyl]-4-fluoro-benzoic
acid
(Imp-E),
2-(3-methyl-6-ethoxy-2,4-dioxo-3,4-dihydro-
2H-pyrimidin-1-ylmethyl)-4-fluoro-
(R)-2-[6-(3-amino-piperidin-1-yl)-3-methyl-2,4-dioxo-
benzonitrile
(Imp-F),
3,4-
dihydro-2H-pyrimidin-1-yl methyl]-4-fluoro-benzamide (Imp-G),
4-((R)-3-amino-piperidin- 1-yl]-2-[(6-((R)-3- amino-piperidin-1-
yl)- 3-methyl- 2,4-oxo-3,4- dihydro-pyrimidin −1(2H)-yl) methyl)
benzonitrile (Imp-H) were gained from our laboratory. Starting
material A (SMA; 3-methyl-6-chloro uracil), Starting material B
(SMB; 2-bromomethyl-4-fluoro-benzonitrile), Starting material C
(SMC; (R)-3-Boc-aminopiperidine), and Starting material D (SMD;
Succinic acid) were purchased from Nanjing Chemlin Chemical
Industry (Nanjing, China). The purity of all substances was >98%
by HPLC test. Acetonitrile (HPLC grade) was purchased from Merck
Ltd (Darmstadt, Germany). Water used for the preparation of
mobile phase was purified using a Milli-Q pure water system
(Millipore, MA, USA). Other chemicals were of analytical grade.
2.4. NMR instrumentation and methods
1H, 13C and DEPT NMR experiments were performed on 300 or
500 MHz NMR spectrometer using dimethyl sulfoxide (DMSO) as
solvent and concentration was 50 mg/mL.

20
H. Zhang et al. / Journal of Pharmaceutical and Biomedical Analysis 128 (2016) 18–27
Table 1
Forced degradation results.
Stress condition
% Assay of TRE
Obeservation
Mass balance
Acid hydrolysis
Oxidation
thermal
Base hydrolysis
Photolytic
97.9
95.2
96.4
89.8
99.8
Mild degradation product i.e. Imp-E & Imp-D formed
Mild degradation product i.e. Imp-G formed
3.6% unknown degradation product formed
Mild degradation product i.e. Imp-G formed, and 9.5% unknown major degradation product formed
No any known and unknown degradation product formed
96.5%
96.8%
95.8%
103.1%
100.1%
Fig. 2. The overlaid HPLC chromatogram of synthesized impurities with the impurities in TRE bulk drug.
and results of MS2 product ion analysis of impurities, we speculated
that the six impurities were (R)-3-methyl-1-(3-fluorobenzyl)-6-
(3-amino-piperidin- 1-yl)-2,4-dioxo-3,4-dihydro-2H-pyrimidine
2.5. FT-IR instrumentation and methods
The IR spectra were recorded in the solid state as a KBr powder
dispersion using a Thermo Scientific Nicolet iS5 FT-IR spectrom-
eter (Waltham, MA, USA). Data were collected between 400 and
(Imp-C),
1-yl
2-(3-methyl-2,4,6-oxo-
methyl)-4-fluoro-benzonitrile
tetrahydro-2H-pyrimidin-
(Imp-D), (R)-2-[6-
4000 cm−1, at a resolution of 4.0 cm−1
.
(3-amino-piperidin-1-yl)-3-methyl-2,4-dioxo-3,4-
dihydro-2H-pyrimidin-1-yl-methyl]-4-fluoro-benzoic
acid(Imp-E),
2H-
2-(3-methyl-6-ethoxy-2,4-dioxo-3,4-dihydro-
(Imp-F),
2.6. Sample preparation
pyrimidin-1-ylmethyl)-4-fluoro-benzonitrile
(R)-2-[6-(3-amino-piperidin-
dihydro-2H-pyrimidin-1-yl-methyl]-4-fluoro-
(Imp-G), 4-((R)-3-amino-piperidin-1-yl]-2-[(6-((R)-3-amino-
piperidin-1-yl)-3-methyl-2,4-oxo-3,4-dihydro-pyrimidin-1(2H)-
yl)methyl) benzonitrile (Imp-H), respectively (Fig. 4).
1-yl)-3-methyl-2,4-dioxo-3,4-
Stock solution of TRE was prepared at a concentration of
0.5 mg/mL. Diluent for sample preparation is the admixture of
water and acetonitrile (80:20, v/v). Spiking the 9 impurities with
TRE at 0.1% of the sample concentration was used to investigate the
system suitability.
benzamide
TRE samples subjected to base hydrolysis and oxidation showed
obvious degradation products as Imp-G. When exposed to oxida-
tion, Imp-E and Imp-D were found. TRE was found to be stable in
photolytic stress condition and labile in heat condition (Fig. 5). The
mass balance results were calculated for all of the stressed sam-
ples and were found to be more than 95%. The purity and assay of
TRE was unaffected by the presence of degradation products, which
confirms the stability-indicating power of the developed method.
The information of degradation studies is given in Table 1.
The forced degradation of TRE was carried out under hydrolytic
(acidic and alkaline), photo, hot and oxidative conditions. Acidic
and alkaline hydrolytic degradations were carried out in 1 N HCl
(5 mL) at 80 ◦C and 1 N NaOH (5 mL) at room temperature for 0.5 h.
After the heating, the samples of acidic and alkaline hydrolysis were
neutralized with NaOH and HCI, respectively. For oxidative degra-
dation, the drug was subjected to 3% H2O2 at 80 ◦C for 0.5 h. TRE
was also subjected to thermolytic (100 ◦C, 2 h) and photolytic (UV
light, 4500 lx, 24 h) stress. Finally, all of the stressed samples were
kept at a concentration of 0.5 mg/mL.
3.2. Structural elucidation of TRE and its impurities
Recently, a number of methods for structural elucidation have
emerged exuberantly and been widely applied, like MS, IR, 1H NMR,
13C NMR, and DEPT. In these ways, we confirm the structures of TRE,
Imp-C, Imp-D, Imp-E, Imp-F, Imp- G and Imp-H. The carbon atoms
were numbered in Table 2.
3. Results and discussion
3.1. Detection of process-related impurities and forced
degradation of TRE
Analysis of TRE samples using the newly developed method
described in Section 2.2 revealed the presence of six impuri-
ties (RT = 5.9, relative retention time (RRT) = 0.314; RT = 10.2,
RRT = 0.537; RT = 17.7, RRT = 0.935; RT = 21.6, RRT = 1.136; RT = 31.5,
RRT = 1.657; RT = 38.1, RRT = 2.005) consistently in several batches
(Fig. 2). The molecular weight of these impurities were detected
as 437.2, 375.4, 376.3, 332.3, 275.2 and 303.2, respectively, which
correspond to Imp-H, Imp-G, Imp-E, Imp-C, Imp-D and Imp-F by
analyzing by LC–MS. The typical chromatogram is shown in Fig. 3.
Then on the basis of the knowledge of the route of TRE synthesis
3.2.1. Structural elucidation of Imp-C
The on-line LC–MS and MS2 spectra of Imp-C in TRE suggested
that it might be (R)-3-methyl-1-(3-fluorobenzyl)-6-(3-amino-
piperidin-1-yl)-2,4-dioxo-3,4-dihydro- 2H- pyrimidine. The 1H
NMR spectrum of Imp-C showed 14 signals corresponding to 21
protons (except solvent peaks for ␦: 2.50 ppm), which accords with
the molecular structure of Imp-C. Imp-C misses one cyano group
and succinic acid compared to that of TRE, In 13C NMR, the chem-
ical shift of cyano group and succinic acid was disappeared. The

Table 2
Determination and structure analysis of TRE and its impurities.
Compound
Structure (number assigned for NMR characterization)
%DetectedvaluesinbulkdrugbatcheSstructure analysis
Source
No.1
No.2
No.3
MS
NMR and IR
[M+H]+ (m/z)
[M−H]− (m/z)
MS/MS
fragmentation
ions(m/z)
TRE
99.58
0.08
99.75
0.04
99.62
0.09
358.2
333.9
–
–
1H NMR, 13
NMR
C
Target compound
Imp-C
–
333.9, 317.2,
208.1, 194.1,
109.4
1H NMR, 13
NMR
C
C
Process
Imp-D
Imp-E
0.04
0.07
0.05
0.06
0.04
0.04
276.8
377.9
274.4
276.6, 208.1,
195.7, 134.1
1H NMR, 13
NMR, IR
Process or degradation
–
378.1, 295.1,
221.3, 208.0,
153.1 100.8
1H NMR
Degradation
Imp-F
Imp-G
Imp-H
0.13
0.04
0.04
0.06
0.04
–
0.11
0.05
0.05
304.4
376.8
438.6
–
–
–
304.5, 276.4,
219.0, 195.2,
151.5,
1H NMR, 13
NMR, IR
C
Process
376.8, 360.6,
293.3, 152.0,
101.4
1H NMR
Degradation
438.5, 421.5,
381.8, 225.3,
214.2, 208.3,
197.3, 181.9
1H NMR, 13
NMR
C
Process

22
H. Zhang et al. / Journal of Pharmaceutical and Biomedical Analysis 128 (2016) 18–27
Fig. 3. The LC–MS chromatogram of TRE with bulk drug.
(a) Total ion chromatography, (b) the LC chromatogram of TRE with bulk drug, (c) the MS spectrum of Imp-H; (d) the MS spectrum of Imp-G; (e) the MS spectrum of Imp-E;
(f) the MS spectrum of Imp-C; (g) the MS spectrum of Imp-D; (h) the MS spectrum of Imp-F.
detailed information of 1H NMR and 13C NMR spectra can be seen
in Tables 3 and 4.
protons, which further supports the proposed molecular structure.
The detailed information of 1H NMR and 13C NMR spectra can be
seen in Tables 3 and 4.
3.2.2. Structural elucidation of Imp-D
The on-line LC–MS and MS2 spectra of Imp-D suggested
that it might be 2-(3- methyl-2,4,6-oxo-tetrahydro-2H-pyrimidin-
1-yl-methyl)-4-fluoro-benzonitrile. The IR spectrum displays
characteristic absorptions at 2983.9, 2234.0, 1703.7, 1609.8 and
1187.7 cm−1, corresponding to the methyl C H stretching mode,
aromatic cyano CN stretching mode, polyimide C O stretch-
ing mode, benzene ring, and C F stretching mode, respectively.
To further confirm the structure, 1H NMR and 13C NMR spec-
tra were provided. The 1H NMR spectrum of Imp-D shows 6
signals corresponding to 10 protons, which accords with the pro-
posed molecular structure. The data 1H NMR are as follows: 1H
NMR (500 MHz, DMSO), ␦: 7.99–7.94 (m, 1H), 7.47–7.44 (m, 1H),
7.38–7.32(m, 1H), 5.08 (s, 2H), 3.84 (s, 2H), 3.14 (s, 3H). The 13C
NMR spectrum of Imp-D shows 13 signals corresponding to 13
3.2.3. Structural elucidation of Imp-E
The on-line LC–MS and MS2 spectra of Imp-E suggested that
it might be (R) −2- [6-(3-amino-piperidin-1-yl)-3-methyl-2,4-
dioxo-3,4-dihydro-2H-pyrimidin-1-yl methyl]-4-fluoro-benzoic
acid, which was a white powder, and the HPLC purity was found
to be 99.72%. The 1H NMR spectrum of Imp-E showed ten signals
corresponding to nineteen protons (two active hydrogens were
not detected in amino-group), which accords with the molecular
structure of Imp-E. Imp-E has one additional carboxylic acid group
but is missing one cyano group and succinic acid compared to that
of TRE. In 1H NMR, the chemical shift of the additional hydrogen
was deshielded to ı8.03-7.99 ppm. The detailed information of 1
NMR spectra can be seen in Table 3.
H

H. Zhang et al. / Journal of Pharmaceutical and Biomedical Analysis 128 (2016) 18–27
23
3.2.4. Structural elucidation of Imp-F
The on-line LC–MS and MS2 spectra of Imp-F suggested
that it might be 2-(3- methyl-6-ethoxy-2,4-dioxo-3,4-dihydro-2H-
pyrimidin-1-yl-methyl)-4-fluoro- benzonitrile. The IR spectrum
displays characteristic absorptions at 2986.7, 2227.0, 1709.0,
1609.4, 1238.5 and 1185.1 cm−1, corresponding to the methyl C
H
stretching mode, aromatic cyano CN stretching mode, polyimide
O stretching mode, benzene ring, ether C stretching
C
C O
mode, and C F stretching mode, respectively. To further confirm
the structure, 1H NMR and 13C NMR spectra were provided.
Imp-F has one additional ethoxy group but is missing one ami-
dogen group and succinic acid compared to that of TRE. In 1H NMR,
the additional ethoxy group was dehielded to ␦ 4.12–4.05 and
1.21–1.16, and the chemical shift of amidogen group and succinic
acid disappeared. The 13C NMR spectrum of Imp-F shows 21 signals
(six carbon splitting) corresponding to 15 protons, which further
supports the proposed molecular structure. The detailed informa-
tion of 1H NMR and 13C NMR spectra can be seen in Tables 3 and 4.
3.2.5. Structural elucidation of Imp-G
The on-line LC–MS and MS2 spectra of Imp-G suggested that it
might be (R)-2-[6-(3-amino-piperidin-1-yl)-3-methyl-2,4-dioxo-
3,4-dihydro-2H-pyrimidin-1-yl methyl]-4-fluoro-benzamide, and
the HPLC purity was found to be 98.89%. The 1H NMR spectrum
of Imp-G showed 13 signals corresponding to 22 protons, which
accords with the molecular structure of Imp-G. Imp-G has one addi-
tional amide group but is missing one cyano group and succinic acid
compared to that of TRE. In 1H NMR, the chemical shift of the addi-
tional hydrogen was deshielded to ı7.33 and 6.12 ppm. The detailed
information of 1H NMR spectra can be seen in Table 3.
3.2.6. Structural elucidation of Imp-H
The on-line LC–MS and MS2 spectra of Imp-H suggested that
it might be 4-((R)- 3-amino-piperidin-1-yl)-2-((6-((R)-3-amino-
piperidin-1-yl)-3-methyl-2,4-oxo-3,4- dihydro-pyrimidin-1(2H)-
yl) methyl) benzonitrile.
Imp-H has one additional 3-amino-piperidine group but is miss-
ing one fluorine group and succinic acid compared to that of TRE.
The 1H NMR spectrum of Imp-H shows 13 signals (except sol-
vent peaks for ␦2.50 ppm) corresponding to 31 protons. The 13C
NMR spectrum of Imp-H shows 23 signals corresponding to 23 car-
bons. Seven carbon signals which disappeared in DEPT-135 were
quaternary carbon atoms. There are nine negative carbon signals,
indicating that they are secondary carbon atoms. Six carbons sig-
nals that appeared in DEPT-90 spectrum were considered as six
tertiary carbon atoms. The remaining a positive carbon signal in
DEPT-135 was a primary carbon. which further supports the pro-
posed molecular structure. The detailed information of 1H NMR and
13C NMR spectra can be seen in Tables 3 and 4.
3.3. Possible mechanisms for the formation of impurities
According to the synthesis of TRE, it was surmised that five for-
mation routes for impurities would arise (Fig. 6). In route 1, SMB
may contain 3-fluorobenzyl bromide (SMB-1) as an impurity in its
synthesis process, and affords Imp-C following the same reaction
that yields TRE. In route 2, intermediate 1 and ethanol occurred
bimolecular nucleophilic-substitution reaction and formed Imp-F
in alkaline environment. In route 3, SMA may contain 1-Methyl bar-
bituric acid (SMA-1) as an impurity, and affords Imp-D following
the same reaction that gives TRE in the first step. Besides, Imp-D
was speculated as an acid degradation product of TRE, which was
confirmed in degradation test. In this two ways, Imp-D was formed.
In route 4, intermediate 3 and SMC reacted and then remove the Boc
Fig. 4. MS2 spectra and plausible fragmentation pathways for the impurities.
group to form Imp-H. In route 5, the cyano group of intermediate
3 may be hydrolyzed to form Imp-E and Imp-G.

24
H. Zhang et al. / Journal of Pharmaceutical and Biomedical Analysis 128 (2016) 18–27
Fig. 5. The chromatogram of TRE under stress conditions (a) acid hydrolysis, (b) oxidative degradation, (c) thermal degradation and (d) base hydrolysis.
Fig. 6. Five routes of producing impurities of TRE and their relationship.

H. Zhang et al. / Journal of Pharmaceutical and Biomedical Analysis 128 (2016) 18–27
25
Table 3
1H NMR assignment for TRE and its impurities.
position
TRE
Imp-C
Imp-D
Imp-E
Imp-F
Imp-G
Imp-H
2
3
4
7.36–7.32 (ma,1H)
7.96–7.93 (m,1H)
–
7.40–7.32 (m,1H)
7.09–7.01 (m,1H)
7.09–7.01 (m,1H)
7.47–7.44 (m,1H)
7.99–7.94 (m,1H)
–
7.22 (m,1H)
8.03–7.99 (m,1H)
–
7.39–7.36 (m,1H)
7.99–7.95 (m,1H)
–
7.08–7.04 (m,1H)
7.65–7.62 (m,1H)
–
6.62 (s,1H)
7.51–7.54 (db,1H)
–
6
7
8
9
7.18–7.16 (m,1H)
–
7.09–7.01 (m,1H)
5.26–4.96 (m,2H)
–
3.13 (s,3H)
–
5.74 (s,1H)
–
2.77–2.70 (m,1H)
2.59–2.56 (m,1H)
1.50–1.43 (m,1H)
1.17 (s,1H)
1.80–1.76 (m,1H)
1.67–1.63 (m,1H)
3.04–3.00 (m,1H)
2.93–2.89 (m,1H)
7.38–7.32 (m,1H)
–
5.08 (s,2H)
–
3.14 (s,3H)
–
3.84 (s,2H)
6.89–6.87 (m,1H)
8.03–7.99 (m,1H)
5.41–5.29 (m.2H)
–
3.22–3.14 (m,3H)
–
5.41–5.29 (m.1H)
–
7.33–7.25 (m,1H)
–
5.18 (s,2H)
–
3.16 (s,3H)
–
5.28 (s,1H)
–
6.67–6.64 (m,1H)
6.90–6.93 (d,1H)
–
5.07 (s,2H)
–
3.09 (s,3H)
–
5.32 (s,1H)
–
–
5.15–5.12 (ddc,2H)
–
5.39–5.28 (qd,2H)
–
10
11
12
13
13ꢀ
14
14ꢀ
15
15ꢀ
16
16ꢀ
17
18
18ꢀ
19
19ꢀ
20
21
22
23
23ꢀ
24
25
3.09 (se,3H)
–
5.39 (s,1H)
–
3.04 (s,3H)
–
5.45 (s,1H)
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
2.90 (s,1H)
2.71 (s,1H)
1.78 (s,1H)
1.52–1.50 (m,1H)
1.91 (s,1H)
1.52–1.50 (m,1H)
3.14 (s,1H)
3.22–3.20 (s,1H)
2.71 (s,1H)
–
3.05–2.90 (m,1H)
3.05–2.90 (m,1H)
1.67 (m,1H)
1.37 (m,1H)
1.91 (m,1H)
1.37 (m,1H)
3.22–3.14 (m,1H)
3.05–2.90 (m,1H)
4.12–4.05 (m,1H)
4.12–4.05 (m,1H)
1.21–1.16 (m,1H)
1.21–1.16 (m,2H)
1.87 (s,1H)
1.67 (s,1H)
1.67 (s,1H)
1.67 (s,1H)
1.67 (s,1H)
1.33–1.31 (m,1H)
3.19–3.16 (d,1H)
2.75 (m,1H)
2.0–1.98 (d,1H)
7.33 (s,1H)
6.12 (s,1H)
3.04 (s,2H)
–
2.86–2.94 (m,1H)
2.59–2.62 (m,1H)
1.81 (m,1H)
1.81 (m,1H)
1.81 (m,1H)
1.81 (m,1H)
2.86–2.94 (m,1H)
3.20–3.23 (m,1H)
2.59–2.62 (m,1H)
3.78–3.81 (m,1H)
2.86–2.94 (m,1H)
1.33–1.44 (m,2H)
1.33–1.44 (m,2H)
3.58–3.62 (m,1H)
1.81 (m,1H)
1.81 (m,1H)
1.81 (m,2H)
1.81 (m,2H)
–
–
–
–
–
–
–
–
–
–
–
–
–
–
2.36–2.30 (m,1H)
2.13 (s,1H)
2.13 (s,1H)
2.68 (m,1H)
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
2.30 (s,2H)
2.30 (s,2H)
–
9.47 (s.1H)
9.47 (s.1H)
9.47 (s.1H)
9.47 (s.1H)
–
–
–
–
–
a
Multiple.
Double.
Doublet of doublets.
Quartet.
Single.
b
c
d
e
Table 4
13C NMR assignment for TRE and its impurities.
position
TRE
Imp-C
Imp-D
Imp-F
Imp-H
␦
C
DEPT
␦
DEPT
␦
DEPT
␦
C
DEPT
␦
C
DEPT
C
C
1
2
3
4
5
6
7
8
165.59, 163.57
115.56, 115.38
135.97, 135.89
106.44
145.27, 145.20
115.01, 114.83
116.62
45.73
151.77
27.30
162.17
89.87
159.20
51.25
21.86
28.60
46.22
54.37
175.22
–
162.05, 163.69
122.34, 122.38
130.28, 130.39
113.22, 113.93
140.39, 140.49
113.51, 113.66
46.81, 46.83
152.22
27.28
160.46
88.75
159.71
50.91
22.87
32.66
58.97
47.09
–
–
166.52, 163.16
114.86–114.53
135.76–135.63
106.48–106.44
144.19–144.06
115.58–115.28
116.47
40.20
152.02
27.88
166.04
42.49
–
166.3, 163.00
115.03–114.72
135.76–135.63
106.61–106.57
144.57–144.46
115.95–115.65
116.45
43.29
151.02
27.52
162.33
78.29
159.47
66.52
13.55
–
–
–
–
–
–
–
–
–
159.52
113.31
142.51
111.87
152.03
118.39
134.34
51.03
152.83
29.87
173.51
96.73
162.18
51.91
22.60
30.03
46.89
63.04
55.53
22.39
46.11
46.38
89.27
–
CH
CH
–
–
CH
–
CH2
–
CH3
–
CH
–
CH2
CH2
CH2
CH
CH2
–
CH2
CH2
–
CH
CH
CH
–
CH
CH2
–
CH3
–
CH
–
CH3
CH2
CH2
CH2
CH
–
CH
CH
–
–
CH
–
CH2
–
CH3
–
CH2
–
–
–
–
–
–
–
–
–
–
CH
CH
–
–
CH
–
CH2
–
CH3
–
CH
–
CH2
CH3
–
–
–
–
–
–
–
–
CH
CH
–
–
CH
–
CH2
–
CH3
–
CH
–
CH2
CH2
CH2
CH
CH2
CH2
CH2
CH2
CH
CH2
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
165.89
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
31.40
31.40
175.22
–
–
–
–
–
3.4. HPLC method development and optimization
substances however without qualitative identification of the impu-
rities. In other words, the impurities in report are different from the
impurities in our study possibly. This meant separating result was
unpredictable when it’s directly applied for analyzing the impu-
rities discussed in our study. After primary trial with reported
HPLC method, it showed that the impurities could not be separated
As starting materials and intermediates with residual poten-
tial, SMA, Imp-A and Imp-B are in need to be involved in the
HPLC method for quantification. As mentioned above, Bao et al.
[6] developed an RPLC method to determine TRE and its related

26
H. Zhang et al. / Journal of Pharmaceutical and Biomedical Analysis 128 (2016) 18–27
Table 5
Summary of method validation.
Compound System suitability
RRTa PCb
Linearity
Sensitivity
Precision
Rc
SFd
Range(␮g/mL) re
Slope Intercept LODf(␮g/mL) LOQg(␮g/mL) Inter-day%RSD(n = 6) Intra-day%RSD(n = 12)
TRE
1.00
2.11
2.60
1.14
1.73
0.94
2.02
0.52
0.29
0.33
45,256
223,814 4.84
3.69
0.92 0.15–10.78
0.93 0.15–1.07
0.997 56.10 −8.87
0.992 52.03 −0.35
0.07
0.06
0.07
0.07
0.04
0.05
0.06
0.10
0.05
0.03
0.15
0.15
0.19
0.20
0.10
0.10
0.17
0.23
0.12
0.08
1.0
1.4
1.2
0.8
1.0
0.9
1.4
1.2
1.1
1.4
0.8
1.6
1.4
1.2
1.2
1.3
0.7
1.3
1.4
1.8
Imp-A
Imp-B
Imp-C
Imp-D
Imp-E
Imp-F
Imp-G
Imp-H
SMA
293,456 26.71 0.98 0.18–1.04
137,951 9.16 0.90 0.20–1.06
119,416 36.21 0.99 0.11–1.05
67,946 22.96 0.98 0.11–0.98
249,503 16.27 0.92 0.17–1.14
9058
7284
14,641
0.995 46.64
1.23
0.996 63.51 −1.77
0.998 54.43 −6.17
0.993 47.39 −0.73
0.996 57.90 −1.47
11.34 0.90 0.23–1.00
–
3.25
0.994 33.92
0.34
0.86 0.11–1.20
0.85 0.09–0.93
0.996 58.14 −1.68
0.992 52.03 −0.35
a
Relative retention time.
(USP) plate count.
(USP) resolution.
Symmetry factor.
Correlation factor.
(S/N = 3).
b
c
d
e
f
g
(S/N = 10).
Table 6
The summary of accuracy and calibration factor.
Compound
Accuracy
50% MRa
Calibration Response Factor
100% MR
150% MR
RSD% (n = 9)
Slope
Intercept
MCFb
MRRFc
%RSD (n = 6)
TRE
–
–
–
–
56.62
73.72
45.44
65.77
56.25
47.92
58.65
33.93
59.37
53.87
−6.31
−2.48
1.40
−8.31
−10.13
−1.09
−1.92
0.92
1.00
1.31
0.81
1.05
0.92
0.84
1.10
0.63
0.97
0.93
1.00
0.76
1.23
0.95
1.08
1.19
0.90
1.58
1.03
1.07
1.0
1.4
0.8
1.6
0.7
1.1
3.6
1.5
3.6
1.3
Imp-A
Imp-B
Imp-C
Imp-D
Imp-E
Imp-F
Imp-G
Imp-H
SMA
86.8
84.0
92.1
91.3
88.7
91.4
96.7
93.4
86.9
87.5
84.5
86.3
88.8
86.7
86.2
91.9
89.8
86.1
87.4
92.4
84.3
89.7
85.6
85.6
89.2
90.9
85.1
0.7
5.3
4.6
3.0
2.6
4.7
3.8
2.1
1.4
−1.46
−0.56
a
Mean recovery.
Mean calibration factor.
Mean relative response factor.
b
c
Fig. 7. The HPLC chromatogram of TRE spiked with its impurities.
well in this system, especially when focusing on the resolutions
(R) between SMA and Imp-H (R = 1.41) as well as Imp-E and TRE
(R = 1.05). For this reason, method optimization was performed as
follows:
Efforts were taken to optimize the reported HPLC method from
aspects of gradient profile, concentration and pH of potassium
dihydrogen phosphate buffer in the mobile phase. Mobile phase
A consisting of 20 mmol potassium dihydrogen phosphate with
0.25% triethylamine (pH adjusted with phosphate acid) and mobile
phase B of ACN were chosen preliminary. First, to modify the
performance, pH was adjusted to 3.0–6.0 with phosphate acid in
reportorial gradient mode. The results showed that all process-
related impurities and TRE had the preferable chromatographic
behavior on pH 3.5. Second, different gradient elution programs
were designed to optimize the HPLC condition with fixed pH 3.5.
With one of the programs, sufficient resolution between SMA and
Imp A (from R = 1.41 with original condition to R = 3.25 after opti-
mization) as well as Imp-E and TRE(R = 1.05 to 3.69) was observed.
Finally, the optimum separation was performed on an InertSus-
tain C18 (250 mm × 4.6 mm, 5 ␮m) column in a gradient mixture of

H. Zhang et al. / Journal of Pharmaceutical and Biomedical Analysis 128 (2016) 18–27
27
acetonitrile and 20 mmol potassium dihydrogen phosphate with
0.25% triethylamine (pH adjusted to 3.5 with phosphate acid) as
described in Section 2.2.
factor was calculated by ratio of the slope of principal components
and the impurities with the linear regression equation. The results
indicated that the correction factor of all impurities were within
1.31, which provided the reference to quantitative analysis of the
impurities in TRE.
3.5. Method validation
The HPLC method was validated according to the ICH Q2 (R1)
guidelines[12], and results of the validation study are given below.
4. Conclusions
Six process-related impurities in TRE bulk drug obtained from
our laboratory were identified and separated by an optimized HPLC
method. Their structures were proposed by the synthesis route of
TRE and MS2 product ions analyses. The structures of all impurities
of TRE were demonstratively confirmed and characterized using 1H
NMR, 13C NMR, IR, and DEPT, and probable mechanisms for their
formation were also discussed. To the best of our knowledge, six
of the impurities are new compounds and the structures have not
been reported previously. The LC method was validated as per ICH
guidelines, and was found to be simple, sensitive, selective, cost
effective and stability indicating. Thus, the method can be used in
the separation and quantification of impurities in TRE bulk drugs.
3.5.1. System suitability
System suitability studies were performed to detect the min-
imum number of theoretical plates, resolution, tailing factor. The
system suitability was conducted using a 0.5 mg/mL TRE solution
containing 0.5 ␮g/mL of 9 impurities by making six replicate injec-
tions. An efficient resolution (>3.25), high number of theoretical
plates (>5000/m), and good tailing factor was obtained as shown in
Table 5. The HPLC chromatograms of the 9 impurities spiked with
TRE is shown in Fig. 7.
3.5.2. Sensitivity and linearity
The LOD and LOQ for TRE and its 9 impurities were estimated at
a signal-to-noise ratio of 3:1 and 10:1, respectively, by injecting a
series of solutions diluted to known concentrations (Table 5).
A set of all 9 impurities and TRE solutions at the seven con-
centration levels ranging from the LOQ to 200% of the specification
limit (0.1% impurity level in a 0.5 mg/mL TRE sample solution) were
prepared. The calibration equations were calculated using least
squares linear regression, and the correlation coefficient of regres-
sion (r) was found to be >0.99, indicating excellent linearity for the
method.
Acknowledgement
The Project was funded by the Priority Academic Program Devel-
opment of Jiangsu Higher Education Institutions.
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.jpba.2016.04.041.
3.5.3. Precision and accuracy
References
The precision of the method, including repeatability (interday
precision) and intermediate precision (intraday precision), was
tested with six different system suitability solutions with differ-
ent analysts on different days using different instruments. The RSD
of TRE and its impurities was within 2.0. The low RSD values of the
peak areas for the principal peak and its impurities demonstrate
the good precision of the method (Table 5).
The accuracy was determined by using the recovery and RSD
values obtained from test TRE solutions (0.5 mg/mL) spiked with
impurity levels of 0.05%, 0.10%, and 0.15% in triplicate. The per-
centage recovery was calculated at each level and found to be in
the range of 84.0%–96.7% (Table 6).
[1] K. McKeage, Trelagliptin: first global approval, Drugs 75 (2015) 1161–1164.
[2] K. Kaku, First novel once-weekly DPP-4 inhibitor, trelagliptin, for the
treatment of type 2 diabetes mellitus, Expert Opin. Pharmacol. 16 (2015) 1–9.
[3] N. Inagaki, H. Onouchi, H. Maezawa, S. Kuroda, K. Kaku, Once-weekly
trelagliptin versus daily alogliptin in Japanese patients with type 2 diabetes: a
randomised, double-blind, phase 3 non-inferiority study, Lancet Diabetes
Endocrinol. 3 (2015) 191–197.
[4] T. Klein, M. Mark, H. Niessen, L. Thomas, DPP-IV inbibitors foruse in the
treatment of nafld, WO Patent 2009147125 A1 (2009).
[5] M.L. Bao, M. You, X.F. Zhao, A method for the quantification of related
substances in Trelagliptin succinate and its preparation, CN Patent
104237421 A (2014).
[6] J. Feng, S.L. Gwaltney, J.A. Stafford, Z.Y. Zhang, Dipeptidyl peptidase inhibitors,
WO Patent 2005095381 A1 (2005).
[7] R.C. Kelly, L.H. Koztecki, Polymorphs of succinate salt of
2-[6-(3-amino-piperldin-1-yl)3-
methyl-2,4-dioxo-3,4-dihydro-2H-pyrimidin-1-ylmethy-4-fluor-benzonitril
and methods of use therefor, US Patent 20080280931 A1 (2008).
[8] H. Fukuyama, K. Harada, M. Koike, Laminated tablet and manufacturing
method therefor, US Patent 20140023708 A1 (2014).
[9] J.P. Yu, W. Chen, X.Y. Zhu, X.Q. Zhang, X.H. Chang, X.K. Liu, Optimized synthetic
process of trelagliptin succinate, Chin. New Drug J. 24 (2015) 1061–1064.
[10] International Conference on Harmonization ICH Guidelines Q3A (R2),
Impurities in New Drug Substances: Step 4 Version (2006).
[11] S. Husain, R.N. Rao, Monitoring of process impurities in drugs, J. Chromatogr.
Libr. 60 (1998) 833–888.
3.5.4. Robustness and solution stability
To assess the robustness of the method, careful evaluation of
deliberately changed conditions: column temperature (30 5 ◦C),
flow rate (1.0 0.1 mL/min), the pH value of the buffer solution
(3.5 0.5), and trying different columns revealed that the reso-
lution between any two analytes is >1.5. The stability of the TRE
sample solution and the mixed impurities solution was monitored
at room temperature for 12 h, and the RSD values of the peak areas
of the analytes were found to be within 2.05.
[12] International Conference on Harmonization ICH Guidelines Q2A (R1),
Validation of Analytical Procedures: Step 4 Version (2005).
3.5.5. Correction factors
The correction factor experiments were carried out on two dif-
ferent chromatographic instruments with three chromatographic
columns injecting six concentration levels solution. The correction