Characterization of degradation products of regorafenib by LC-QTOF-MS and NMR spectroscopy: Investigation of rearrangement and odd-electron ion formation during collision-induced dissociations under ESI-MS/MS

Article abstract of DOI:10.1039/c7nj01440f

Regorafenib is an oral multikinase inhibitor, and it was subjected to stress conditions (hydrolysis, oxidative, thermal and photolytic) as per ICH specified conditions. The drug showed considerable degradation under hydrolysis (acidic, basic and neutral) and oxidative stress conditions, whereas it was stable under other stress conditions. A total of five degradation products (DPs) were observed and these were analyzed by using a UHPLC-DAD system. Chromatographic separation was achieved on an Acquity CSH C18 column (100 × 2.1 mm, 1.7 μ) using 0.1% formic acid and acetonitrile:methanol (80:20%, v/v) as the mobile phase in gradient mode. All DPs were characterized by LC-MS/MS, and the major degradation product (DP1) was isolated by using an HPLC preparative system from a degradation mixture and analyzed using NMR (1D and 2D NMR) and IR experiments. It was observed that protonated DP1 and DP3 undergo rearrangement reactions during collision-induced dissociations under positive electrospray ionization conditions. Additionally, in silico toxicity of the drug and its degradation products (DP1-DP5) was evaluated using TOPKAT and DEREK toxicity prediction software tools.

Full text of DOI:10.1039/c7nj01440f

View Article Online
View Journal
NJC
Accepted Manuscript
This article can be cited before page numbers have been issued, to do this please use: S. M. Baira, S.
Gannoju, R. D. Nimbalkar, P. Garg, S. R and M. V. N. K. Talluri, New J. Chem., 2017, DOI:
10.1039/C7NJ01440F.
This is an Accepted Manuscript, which has been through the
Volume 40 Number
1 January 2016 Pages 1–846
Royal Society of Chemistry peer review process and has been
accepted for publication.
NJC
Accepted Manuscripts are published online shortly after
New Journal of Chemistry
www.rsc.org/njc
A journal for new directions in chemistry
acceptance, before technical editing, formatting and proof reading.
Using this free service, authors can make their results available
to the community, in citable form, before we publish the edited
article. We will replace this Accepted Manuscript with the edited
and formatted Advance Article as soon as it is available.
You can find more information about Accepted Manuscripts in the
author guidelines.
Please note that technical editing may introduce minor changes
to the text and/or graphics, which may alter content. The journal’s
standard Terms & Conditions and the ethical guidelines, outlined
in our author and reviewer resource centre, still apply. In no
event shall the Royal Society of Chemistry be held responsible
for any errors or omissions in this Accepted Manuscript or any
consequences arising from the use of any information it contains.
ISSN 1144-0546
PAPER
Jason B. Benedict et al.
The role of atropisomers on the photo-reactivity and fatigue of
diarylethene-based metal–organic frameworks
rsc.li/njc

Page 1 of 32
New Journal of Chemistry
View Article Online
DOI: 10.1039/C7NJ01440F
1
2
3
4
5
6
Characterization of degradation products of regorafenib by LC-QTOF-MS
and NMR: Investigation of rearrangement and odd-electron ion formation
during collision-induced dissociations under ESI-MS/MS
Shandilya Mahamuni Baira,a G. Srinivasulu,a Rakesh Nimbalkar,b Prabha Garg,b
R. Srinivas,ac M.V.N Kumar Talluri*a
7
8
9
a Department of Pharmaceutical Analysis, National Institute of Pharmaceutical Education &
Research, Balanagar, Hyderabad, Telangana, India
b
10
Department of Pharmacoinformatics, National Institute of Pharmaceutical Education &
11 Research, S.A.S Nagar, Mohali, Punjab, India
12 c NCMS, CSIRꢀIndian Institute of Chemical Technology, Hyderabad, Telangana, India
13
14
15
16
17
18
19
20
21
22
23
24
25
26
*a Corresponding author: Dr. M.V.N. Kumar Talluri, Department of Pharmaceutical Analysis, National
Institute of Pharmaceutical Education & Research (NIPER), Hyderabad, Telangana, India
Tel: +91ꢀ40ꢀ23074750, Ext: 2012, Fax: +91ꢀ40ꢀ23073751.
Eꢀmail: narendra.talluri@gmail.com, narendra@niperhyd.a
1

New Journal of Chemistry
Page 2 of 32
View Article Online
DOI: 10.1039/C7NJ01440F
1
2
Graphical abstract
3
4
5
6
7
8
9
10
11
12
2

Page 3 of 32
New Journal of Chemistry
View Article Online
DOI: 10.1039/C7NJ01440F
1
2
3
4
5
6
7
8
9
Abstract
Regorafenib is an oral multikinase inhibitor, was subjected to stress conditions
(hydrolysis, oxidative, thermal and photolytic) as per ICH specified conditions.
The drug showed considerable degradation under hydrolysis (acidic, basic and
neutral) and oxidative stress conditions, whereas it was stable under other stress
conditions. A total of five degradation products (DPs) were observed and these
were analyzed by using UHPLCꢀDAD system. The chromatographic separation
was achieved on an Acquity CSH C18 (100×2.1 mm, 1.7µ) column using 0.1%
formic acid and acetonitrile: methanol (80: 20, %v/v) as mobile phase in
10 gradient mode. All DPs were characterized by LC–MS/MS and major
11 degradation product (DP1) was isolated by using preparative HPLC from
12 degradation mixture and analyzed using NMR (1D and 2D NMR) and IR
13 experiments. It was observed that protonated DP1 and DP3 undergo
14 rearrangement reactions during collision induced dissociations under positive
15 electrospray ionization conditions. Additionally, in silico toxicity of the drug
16 and its degradation products (DP1ꢀDP5) were evaluated using TOPKAT and
17 DEREK toxicity prediction software tools.
18 Keywords: Regorafenib; Stress study; Characterization; Rearrangement
19 reaction; Oddꢀelectron ion
20
21
22
3

New Journal of Chemistry
Page 4 of 32
View Article Online
DOI: 10.1039/C7NJ01440F
1
2
3
4
5
6
7
8
9
Introduction
According to the American Cancer Society, cancer is the second most common
cause of death and it accounts for nearly one out of every 4 deaths. Cancer is
defined as a disease caused by an uncontrolled division of abnormal cells in a
part of the body. In this condition, the cells of a specific tissue lose their ability
to respond to the signals within the tissue that regulate cellular differentiation,
survival, proliferation, and death. As a result, these cells accumulate causing
local damage and inflammation. There are over 200 different types of cancer,
and each is classified by the type of cell that is initially affected.1
10 Of all the medications that have come out to treat this condition, kinase
11 inhibitors have played an increasingly prominent role. Phosphorylation
12 regulates many biological processes of our body, and protein kinase inhibitors
13 can be used to treat diseases due to hyperactive protein kinases (including
14 mutant or over expressed kinases in cancer) or to modulate cell functions to
15 overcome other disease drivers.2
16 Regorafenib [4ꢀ[4ꢀ({[4ꢀChloroꢀ3ꢀ(trifluoromethyl)phenyl]carbamoyl}amino)ꢀ3ꢀ
17 fluorophenoxy]ꢀNꢀmethylpyridineꢀ2ꢀcarboxamide hydrate] is a novel oral
18 multikinase inhibitor that potently inhibits the endothelial cell kinases in
19 biochemical and cellular kinase phosphorylation assays. It also inhibits
20 additional angiogenic and the mutant oncogenic kinases (KIT, RET, and Bꢀ
21 RAF).3ꢀ5
22
4

Page 5 of 32
New Journal of Chemistry
View Article Online
DOI: 10.1039/C7NJ01440F
1
2
3
4
5
6
7
8
9
Forced degradation or stress degradation study include degradation of a new
drug substance and drug product under different stress conditions more harsher
than accelerated conditions and it provides an insight into determination of
degradation pathways, degradation products (DPs), intrinsic stability of a drug
substance, structure elucidation of DPs and also essential for demonstrating
specificity of stability indicating methods. These forced degradation studies are
also helpful for a demonstration of the chemical behavior of the drug molecule
which in turn used in the formulation development and final package. 6ꢀ8
The FDA and other regulatory agencies state the stability testing data as a
9,10
10 regulatory requirement and it has been extended to generic drugs.
The
11 degradation products formed under forced degradation studies may or may not
11ꢀ14
12 mimic process impurities
and real time storage impurities. However, the
13 developed stability indicating method can also be applied for the analysis of
14 samples generated under long term, accelerated stability studies and real time
15 storage conditions. 6,9
16 Recently there have been reports on synthesis of REG,15,16 HPLC method,17 and
17 LCꢀMS assay method for determination of regorafenib in rat plasma.18ꢀ20
18 However, no work has been published on forced degradation studies of
19 regorafenib and characterization of its degradation products. Hence, the purpose
20 of this study is to identify and characterize the degradation products of
21 regorafenib using LC/ESI/QTOF/MS/MS, NMR and IR experiments and to
22 evaluate in vitro toxicities of degradation products.
5

New Journal of Chemistry
Page 6 of 32
View Article Online
DOI: 10.1039/C7NJ01440F
1
Experimental
2
3
4
5
6
7
8
9
Chemicals and reagents
Pure regorafenib was obtained as gratis sample from MSN Laboratories
(Hyderabad, India). Highꢀperformance liquid chromatography (HPLC) grade
acetonitrile (ACN) and methanol (MeOH) were purchased from Merck (Merck,
Mumbai, India). Buffer salts and all other chemicals such as formic acid,
ammonium acetate, hydrochloric acid, sodium hydroxide and 30% (w/w)
hydrogen peroxide were purchased from SD Fine Chemicals Pvt. Ltd.,
(Mumbai, India). Ultra pure water (H2O) was obtained from a Millipore MilliꢀQ
10 plus system (Milford, MA, USA)
11 Instrumentation
12 UPLC (Ultra Performance Liquid Chromatography) studies were carried out on
13 Waters UPLC® Hꢀclass system (Waters Corp., Milford, MA, USA), which was
14 equipped with an auto sampler and quaternary gradient pump with an inꢀline
15 degasser, multiple column compartments and photodiode array (PDA) detector.
16 The chromatographic data were acquired using Empower 3 software (Waters
17 Corp.).
18 LCꢀHRMS (Liquid Chromatographyꢀ High Resolution Mass Spectrometry)
19 studies were carried out using an Agilent Infinity 1290 series instrument
20 (Agilent Technologies, Santa Clara, California, USA) comprising of
6

Page 7 of 32
New Journal of Chemistry
View Article Online
DOI: 10.1039/C7NJ01440F
1
2
3
4
5
6
7
8
9
quaternary pump (G4204A), auto sampler (G4226A), column oven (G1316C),
thermostat (G1330B) and PDA detector (G4212B). The HRMS system
consisted of quadrupole timeꢀofꢀflight (QꢀTOF LC–MS 6540) equipped with an
electrospray ionization (ESI) source. System control and data acquisition were
done by Mass Hunter Workstation software.
Isolation of selected DP was carried out by using semi preparative HPLC
instrument (Waters Corp.), which is equipped with a binary pump (515 and
515), a deꢀgasser, pump control module, manual injector, a diodeꢀarray detector
(2489 UVꢀVIS), and fraction collector. The data acquisition was done under the
10 control of Empower 3 software (Waters Corp.).
11 NMR (Nuclear magnetic resonance) experiments (1D and 2D NMR) were done
12 by AVANCE III HDꢀ500 MHz NMR, Bruker (Bruker, Billerica, Massachusetts,
1
13
13 USA). DMSOꢀd6 was used as solvent. H and C chemical shift values were
14 reported on the δ scale in ppm relative to TMS (δ = 0.00 ppm) as an internal
15 standard. The data acquisition and processing of NMR spectra was done using
16 Top spin software (3.2 version). The FTIR (Fourier transform infrared
17 spectroscopy) spectra were recorded by using PerkinElmer Spectrum (Two
18 spectrophotometer, PerkinElmer, Inc., Waltham, MA, USA). Photolytic studies
19 were performed in a photostability chamber (Osworld Scientific Pvt. Ltd.,
20 Mumbai, India) and thermal degradation studies were carried out in the Osworld
21 laboratory oven (Osworld Scientific Pvt. Ltd.). Eutech pH meter (Eutech
22 Instruments, Singapore) was used to evaluate the pH of all the solutions. To
7

New Journal of Chemistry
Page 8 of 32
View Article Online
DOI: 10.1039/C7NJ01440F
1
2
3
4
5
6
7
dissolve the samples, Power Sonicꢀ405, an ultraꢀsonicator (Hwashin
Technology Co., Seoul, South Korea) was used.
Sample preparation
The diluent (ACN: water (80: 20, % v/v)) was used to prepare the standard as
well as stress degradation samples. The acidic and basic stressed degradation
samples were neutralized and diluted to 100ppm with the diluent before UPLC
analysis.
8
9
Forced degradation studies on REG
Forced degradation of REG was performed on the bulk drug (API) as per ICH
10 guidelines.21 REG was subjected to different stress conditions such as
11 hydrolysis (acid, base and neutral), oxidative, photolytic and thermal stress
12 conditions. The drug solutions were prepared at 2.0 mg/mL concentration for all
13 stress samples and diluted with acid, base and 30% H2O2 in 1: 1 ratio. All the
14 optimized stress conditions are given in Table 1.
15
16 Table 1 Stress conditions for optimum degradation of REG
Stress condition
Concentration
stressor
of Exposed condition
Duration
Degradation products formed
Acid hydrolysis
Base hydrolysis
Neutral hydrolysis Diluent
2 N HCl
0.1 N NaOH
80 °C
80 °C
80 °C
5 h
24 h
3 d
DP1, DP2 and DP3
DP1, DP2 and DP4
DP1 and DP2
Oxidation
Thermal
Photolysis
30% H2O2
Solid state
Solid state
RT
6 d
15 d
4 d
DP3 and DP5
No degradation
No degradation
100 °C
40 ◦C, 75% RH
17
18 Enrichment, isolation and NMR studies on major degradation product DP1
8

Page 9 of 32
New Journal of Chemistry
View Article Online
DOI: 10.1039/C7NJ01440F
1
2
3
4
5
6
7
8
9
The major degradation product DP1 was observed in all the hydrolytic (acid,
alkaline and neutral) stress conditions. Therefore neutral hydrolytic condition
was tried to enrich major degradation product DP1 by dissolving 150 mg of
REG in 50 mL of diluent and refluxed for 120 h at 75 °C, where DP1 was
formed in major quantity. The degradation product (DP1) was separated on a
preparative column (Xꢀbridge C18 (250 mm × 19 mm, 5 ꢁm)), by using a
mobile phase composed of 0.1% formic acid (A) and acetonitrile (B) in an
isocratic mode (35: 65, % v/v). The optimized flow rate, injection volume, and
detection wavelength were 12 mL/min, 1.5 mL and 261 nm, respectively. The
10 fraction (DP1) was collected and concentrated by using rotavapor. The isolated
11 products were dissolved in deuterated DMSO and subjected to 1H and 13C NMR
12 experiments.
13 Fourier transform infrared spectroscopy (FTIR)
14 The FTIR spectra of REG and DP1were recorded in the range of 4000–400
15 cm−1. A small quantity of dried powder sample of REG and DP1 were placed on
16 ATR crystal and recorded FTIR spectra for functional group determination.
17 In silico toxicity prediction
18 In silico models offer an economical, fast and easy way of evaluating
19 toxicological properties of a drug molecule prior to synthesis. Several
20 commercial toxicity prediction tools available for toxicity prediction, covering a
21 range of toxicity endpoints such as carcinogenicity, reproductive effects,
22 irritation, hepatotoxicity, teratogenicity, and mutagenicity etc. Two types of
9

New Journal of Chemistry
Page 10 of 32
View Article Online
DOI: 10.1039/C7NJ01440F
1
2
3
4
5
6
7
8
9
approaches are available to predict the toxicity of molecules such as knowledgeꢀ
based and statistically based systems. A knowledgeꢀbased system (DEREK,
Deductive Estimation of Risk from Existing Knowledge) uses human expert
opinion to predict the potential toxicity of new molecules based on relationship
between structure and biological activity, whereas statistically based systems
(TOPKAT, Toxicity Prediction by Komputer Assisted Technology) use
application of various statistical methods, calculated parameters, and structural
connectivity to derive the toxicity data for molecules.22,23 In the present study
two computational tools (TOPKAT (Discovery Studio 2.5, Accelrys, Inc., San
10 Diego, CA, USA) and DEREK (Nexus v2.0, Lhasa Ltd., Leeds, UK)) were used
11 to evaluate the toxicity of REG and its degradation products.
12 Results and discussion
13 Optimization of chromatographic conditions
14 Initial trials were done with ammonium acetate buffer at different pH (3.5, 4.0,
15 5.0 and 6.5 pH) conditions on BEH C18 column for optimum separation of
16 REG and its DPs. It was observed that peak shapes were not good. Then moved
17 to 0.1% formic acid from ammonium acetate which gave better peak shapes. At
18 this stage DP3 coꢀeluted with DP5. To get better separation between the DP3
19 and DP5, methanol was added to organic phase B (acetonitrile). Finally REG
20 and its DPs were separated on an Acquity CSH C18 (100 mm × 2.1 mm, 1.7
21 ꢁm) by using mobile phase composed of aqueous phase A (0.1% formic acid)
22 and organic phase B (acetonitrile: methanol, 80: 20 % v/v) in a gradient mode
10

Page 11 of 32
New Journal of Chemistry
View Article Online
DOI: 10.1039/C7NJ01440F
1
(Tmin/B; T0/10; T1.5/10;T3/50; T8.5/80; T9.5/80; T11/10; T12 /10). The flow
rate, injection volume, and detection wavelength were 0.3 mL/min, 1µL and
261 nm, respectively. The same method was used for LCꢀHRMS studies.
Degradation behavior of REG
2
3
4
5
6
7
8
9
The degradation behavior of REG was monitored using UPLCꢀPDA and
UPLC–MS under various stress degradation conditions. Optimum degradation
was observed under all stress degradation conditions except photolysis and
thermal degradation, where the drug was found to be stable. A total of five DPs
(DP1 to DP5) were formed in hydrolysis (acid, base and neutral) and oxidative
10 stress conditions (Fig. 1). Among all the DPs, two major degradation products
11 (DP1 and DP2) were observed as common degradation products in all the
12 hydrolysis (acid, base and neutral) conditions. The degradation product DP3
13 was formed under both acid hydrolysis and oxidation stress conditions. The
14 drug was found to be more stable even after subjecting to thermal degradation at
15 100 °C for 15 days. The overlay chromatograms of all stress degradation
16 samples of REG are given in Fig.1
17
11

New Journal of Chemistry
Page 12 of 32
View Article Online
DOI: 10.1039/C7NJ01440F
1
2
3
4
Fig. 1. Overlaid chromatograms of (a) Blank (ACN: water, 80: 20, %v/v), (b) standard
(unstressed) of REG, (c) acid degradation of REG, (d) oxidative degradation of REG, (e)
neutral degradation of REG, (f) base degradation of REG.
5
6
7
8
9
Characterization of REG and its DPs by LC–MS/MS
REG and all the DPs (DP1ꢀ5) were well separated by LC and they show
abundant protonated molecular ions ([M+H]+) in ESI positive ionization mode.
The ESIꢀMS/MS spectra of all the five degradation products are shown in Fig.
2, and their elemental compositions and accurate masses are given in Table S1
10 (see the ESI, Electronic Supplementary Information).
11 MS/MS of [M+H]+ of REG (m/z 483)
12 To understand the fragmentation pathway of protonated REG (m/z 483), its
13 MS/MS spectrum was examined. The spectrum showed abundant product ions
14 at m/z 465, m/z 443, m/z 424, and m/z 404. Other abundant product ions were
15 observed at m/z 288, m/z 270, m/z 262, m/z 244, m/z 229, m/z 201 and m/z 193.
12

Page 13 of 32
New Journal of Chemistry
View Article Online
DOI: 10.1039/C7NJ01440F
1
The ion at m/z 288 can be formed due to cleavage of carbamide bond between
chlorotrifluoromethyl benzene ring and rest of the moiety of REG. The
elemental compositions of all above fragment ions (Scheme 1) have been
confirmed by accurate mass measurements (Table S1).
2
3
4
5
6
7
Fig. 2. LCꢀESI/MS/MS spectra of [M+H]+ ions of (a) REG, (b) DP1, (c) DP2, (d) DP3, (e)
DP4, (f) DP5
13

New Journal of Chemistry
Page 14 of 32
View Article Online
DOI: 10.1039/C7NJ01440F
1
2
3
Scheme 1.
Proposed fragmentation pathway of protonated REG
4
5
6
7
8
9
MS/MS of [M+H]+ of DP1 (m/z 262)
The MS/MS spectrum of [M+H]+ ions (m/z 262) of DP1 (Rt= 5.1 min) shows
the product ions of m/z 244, m/z 203, m/z 175, m/z 155, m/z 148, m/z 128, m/z
126, m/z 111, m/z 101, m/z 98, m/z 77 and m/z 51 (Fig. 2b). It was observed that
the [M+H]+ of REG gives the ion at m/z 262 with same elemental composition
10 that of DP1 (Scheme 2). However, the absence of abundant product ions at m/z
11 175, 155 and 148 in the MS/MS of protonated REG clearly points to structural
14

Page 15 of 32
New Journal of Chemistry
View Article Online
DOI: 10.1039/C7NJ01440F
1
changes in DP1. Thus the formation of these product ions from DP1 can be
attributed to rearrangement reactions following an orthoꢀeffect in 4ꢀ
phenoxypyridine ring moiety of DP1 (Scheme 3) during collisionꢀinduced
dissociation (CID). This type of rearrangement reaction was reported in
aromatic ethers, sulphides, and sulphoxides under EI conditions.24,25 Based on
the above data, the structure of DP1 could be proposed as 4ꢀ(4ꢀaminoꢀ3ꢀ
fluorophenoxy)ꢀNꢀmethylpicolinamide.
2
3
4
5
6
7
8
9
Scheme 2.
Proposed fragmentation pathway of protonated DP1
10
15

New Journal of Chemistry
Page 16 of 32
View Article Online
DOI: 10.1039/C7NJ01440F
1
2
3
Scheme 3. Proposed reaction pathway for the formation of rearranged fragment ions (m/z
175, 155, 148 and 101) in DP1 and DP3
4
5
6
7
8
9
MS/MS of [M+H]+ of DP2 (m/z 196)
The accurate mass of [M+H]+ of DP2 (Rt= 6.9 min) was determined to be m/z
196.0138 (Fig. 2(c)) corresponding to an elemental composition of C7H6ClF3N+.
The MS/MS spectrum of protonated DP2 is shown in Fig. 2(c). Scheme 4
includes the structure of DP2 and its product ions at m/z 178, m/z 176, m/z 161,
10 m/z 142, m/z 127, m/z 125, m/z 111, m/z 99, m/z 92, m/z 75, m/z 74, m/z 68 and
11 m/z 65. The diagnostic fragment ions at m/z 178, 161, 127 and 68 can be formed
12 from m/z 196 by the loss of NH3, Cl, CF3 and C6H6ClN, respectively. Formation
16

Page 17 of 32
New Journal of Chemistry
View Article Online
DOI: 10.1039/C7NJ01440F
1
of radical fragment ions at m/z 161, 127, 111 and 74 involve homolytic bond
cleavage during CIDs,26 which is generally observed in EIꢀMS, are less
common among CIDs under ESI and APCI conditions.27 The characteristic
product ions at m/z 127 and m/z 68 indicate that 4ꢀchloroaniline and fluoroform
moieties are intact in the structure of degradation product DP2. Based on
fragmentation of protonated DP2 and accurate mass measurements, the structure
of DP 2 could be proposed as 4ꢀchloroꢀ3ꢀ(trifluoromethyl)aniline.
2
3
4
5
6
7
8
9
10
Scheme 4.
Proposed fragmentation pathway of protonated DP2
11
12 MS/MS of [M+H]+ of DP3 (m/z 349)
13 The MS/MS spectrum (Fig. 2(d)) of protonated DP 3 (Rt 7.8 min, m/z 349)
14 shows the product ions at m/z 221, m/z 196, m/z 176, m/z 166, m/z 161, m/z 154,
17

New Journal of Chemistry
Page 18 of 32
View Article Online
DOI: 10.1039/C7NJ01440F
1
2
3
4
5
6
7
8
9
m/z 141, m/z 128, m/z 126, m/z 108, m/z 99, m/z 83 and m/z 53. The generation
of intense ions at m/z 196, 176 and 161 indicates the presence of 4ꢀchloroꢀ3ꢀ
(trifluoromethyl) aniline, whereas ions at m/z 128, 108 and 83 represents 4ꢀ
aminoꢀ3ꢀfluorophenol moieties are intact in the DP3. The formation of product
ions at m/z 221 and m/z 154 displayed urea bond linking between 1ꢀchloroꢀ2ꢀ
(trifluoromethyl)benzene and 3ꢀfluorophenol rings. Based on the fragmentation
pattern
(Scheme
5),
DP3
was
identified
as
1ꢀ(4ꢀchloroꢀ3ꢀ
(trifluoromethyl)phenyl)ꢀ3ꢀ(2ꢀfluoroꢀ4ꢀhydroxyphenyl)urea. The elemental
compositions have been confirmed for [M+H]+ of DP3 and its fragment ions by
10 accurate mass measurements.
11
12
Scheme 5.
Proposed fragmentation pathway of protonated DP3
18

Page 19 of 32
New Journal of Chemistry
View Article Online
DOI: 10.1039/C7NJ01440F
1
MS/MS of [M+H]+ of DP4 (m/z 304)
2
3
The MS/MS spectrum of protonated DP4((Rt= 4.9 min, m/z 304
+
C15H15FN3O3 ), displayed product ions at m/z 286, m/z 245, m/z 227, m/z 203,
4
m/z 183, m/z 175, m/z 155, m/z 148, m/z 128, m/z 126, m/z 111 and m/z 98. The
common fragment ions at m/z 175, 155, 148 (Scheme 3), 126 and 111 were
generated from 4ꢀaminoꢀ5ꢀfluoroꢀ2ꢀ(pyridinꢀ4ꢀyl)phenol moiety and it was
intact in both the DPs (DP1 and DP4). Based on the fragmentation pattern
(Scheme 6) and accurate measurements, DP4 was identified as 4ꢀ(4ꢀacetamidoꢀ
3ꢀfluorophenoxy)ꢀNꢀmethylpicolinamide.
5
6
7
8
9
10
11
12
Scheme 6.
Proposed fragmentation pathway of protonated DP4
19

New Journal of Chemistry
Page 20 of 32
View Article Online
DOI: 10.1039/C7NJ01440F
1
2
3
4
5
6
7
8
9
MS/MS of [M+H]+ of DP5 (m/z 499)
The degradation product, DP5 (Rt = 8.1min) was formed under oxidation stress
condition. The mass difference between the protonated drug (m/z 483) and DP5
(m/z 499) is 16 Da, which indicates an addition of oxygen from hydrogen
peroxide to REG as hydroxylated form or Nꢀoxide of the drug. The MS/MS
spectrum of DP5 displayed product ions at m/z 481, m/z 461, m/z 442, m/z 424,
m/z 404, m/z 388, m/z 368, m/z 304, m/z 278, m/z 247, m/z 229, m/z 201, m/z
174, m/z 126 and m/z 111 (Scheme 7). The product ion at m/z 424 indicated the
presence of 1ꢀ(4ꢀchloroꢀ3ꢀ(trifluoromethyl)phenyl)ꢀ3ꢀ(2ꢀfluoroꢀ4ꢀ(pyridinꢀ4ꢀ
10 yloxy)phenyl)urea moiety which is similar to that of the drug (Scheme 1).
11 However, the presence of product ions at m/z 304 and 278, with a mass of 16 Da
12 more than the corresponding fragment ions of the drug (m/z 288 and 262),
13 showed that a structural change occurred in 4ꢀ(4ꢀaminoꢀ3ꢀfluorophenoxy)ꢀNꢀ
14 methylpicolinamide moiety. The product ions at m/z 278 and 247 suggested that
15 hydroxylation had taken place presumably on the picolinamide ring.
16 Atmospheric pressure chemical ionization (APCI)ꢀMS experiment was
17 performed to distinguish between Nꢀoxide and hydroxylated product. It was
18 observed that there is no instant loss of OH radical, which is generally observed
19 in the case of Nꢀoxide by APCIꢀMS experiment. It was reported that Nꢀoxides
20 are thermally labile and undergo deoxygenation by thermal decomposition
21 (thermal energy activation in the vaporizer of the APCI source) in the APCI.28,29
22
20

Page 21 of 32
New Journal of Chemistry
View Article Online
DOI: 10.1039/C7NJ01440F
1
2
3
Scheme 7.
Proposed fragmentation pathway of protonated DP5
4
5
21

New Journal of Chemistry
Page 22 of 32
View Article Online
DOI: 10.1039/C7NJ01440F
1
2
3
4
5
6
7
8
9
The structure of degradation product DP1 was similar to one of the key
intermediate while DP2 mimic the starting material in the synthesis of REG 15,16
and DP3 matched with regorafenib urea impurity (CAS No 1333390ꢀ56ꢀ9). The
DP5 (hydroxylated DP) did not mimic with any of the process impurities or
metabolites but closely similar to regorafenibꢀNꢀOxide 19 (M2 Metabolite, CAS
No 835621ꢀ11ꢀ9) and hydroxyregorafenib (M3 Metabolite).
NMR structural characterization studies
The degradation product DP1 and REG (Fig. 3) were subjected to 1D NMR (1H,
10 13C and DEPT) and 2D NMR (COSY, ROESY, HSQC and HMBC)
11 experiments to verify the proposed structure of DP1. The data are compiled in
1
13
12 Table S2 (see the ESI). H and C NMR experiments in DMSOꢀd6 reveal that
13 REG contained 15 protons and 21 carbons while DP1 contained 12 protons and
13
14 13 carbons. Based on DEPTꢀ90, DEPTꢀ135, and C NMR experiments, DP1
15 has 1ꢀCH3 (primary), 6ꢀCH (tertiary) and 6ꢀC (quaternary) carbons, while REG
1
16 has 1ꢀCH3 (primary), 9ꢀCH (tertiary) and 10ꢀC (quaternary) carbons. In H
17 NMR spectrum of DP1, the protons at Hꢀ12, Hꢀ15, Hꢀ16, and Hꢀ17 were
18 shielded to 6.99, 6.84, 6.77 and 5.20 ppm when compared with REG this
19 revealed that absence of 1ꢀ(4ꢀchloroꢀ3ꢀ(trifluoromethyl)phenyl)urea moiety
20 which is present in the parent molecule (Fig. 3).
21 In addition, 13C NMR spectra were consistent with aboveꢀmentioned
22 observations, the carbon Cꢀ14 was shifted to deshielding as 142.2 ppm from
22

Page 23 of 32
New Journal of Chemistry
View Article Online
DOI: 10.1039/C7NJ01440F
1
124.9 ppm because of –NHCONH group attached to Cꢀ14 was replaced by the
simple NH2 group. The structure of DP1 was further confirmed by COSY and
ROESY experiments where all the correlations are in accordance with the
proposed structure.
2
3
4
5
6
7
8
9
Other 2D techniques such as HSQC and HMBC were also performed to verify
the structure of DP1. The HSQC spectra of DP1 and REG are shown in Fig. 4.
HSQC spectra of DP1 shows the presence of correlation of CH3 (9) and CH
(2,4,5,12,15,16) and absence of correlations of CH (21, 24, 25) which were
observed in REG HSQC spectra (Fig. 4). In addition, HMBC spectra of DP1
10 and REG were used to verify the correlations between the protons and carbons
11 in the structural framework, which were ≥2 bonds apart. Moreover, notable
12 correlations observed in HMBC spectra also helped to assign chemical shift
13 values corresponding to protons and carbons. Based on the observations from
14 1D NMR (1H, 13C, DEPTꢀ90, DEPTꢀ135) and 2D NMR (COSY, ROESY,
15 HSQC, and HMBC) spectra, the structure of DP1 was identified as 4ꢀ(4ꢀaminoꢀ
16 3ꢀfluorophenoxy)ꢀNꢀmethylpicolinamide (Fig. 3).
17
23

New Journal of Chemistry
Page 24 of 32
View Article Online
DOI: 10.1039/C7NJ01440F
1
2
Fig. 3. Schematic diagram of chemical structures of REG and DP1 with number depiction
3
4
5
6
Fig. 4. HSQC spectra of REG and DP1 (500 MHz, DMSOꢀd6)
24

Page 25 of 32
New Journal of Chemistry
View Article Online
DOI: 10.1039/C7NJ01440F
1
Mechanistic explanation of formation of degradation products (DP1-DP5)
The formation of degradation products (DP1ꢀDP5) was shown in Scheme 8.
DP1 and DP2 were formed as major degradation products under acid, base, and
neutral hydrolytic condition because of susceptibility of urea bond to all
hydrolytic (acid, base and neutral) conditions. DP1 and DP2 were formed by
hydrolytic cleavage of urea bond. The degradation product DP3 was formed by
cleavage of diphenyl ether linkage of REG under acid hydrolysis and oxidative
stress conditions. Formation of DP4 can be explained by hydrolysis of urea
bond to form carbamic (RꢀNHCOOH) acid followed by its conversion to
2
3
4
5
6
7
8
9
10 acetamide (RꢀNHCOCH3). The degradation product DP5 was formed only
11 under oxidative stress condition and it can be explained by an addition of
12 hydroxyl group to Nꢀmethylpicolinamide ring of REG.
13
25

New Journal of Chemistry
Page 26 of 32
View Article Online
DOI: 10.1039/C7NJ01440F
1
2
Scheme 8.
Probable mechanisms of formation of degradation products (DP1-DP5)
3
4
5
6
7
8
9
IR structural characterization studies
The major differences in the functional groups and similarities between REG
and DP1were evaluated by comparing the IR spectra of DP1 with that of REG.
The major differences found at 3388.92, 3349.84 and 1718.59 cmꢀ1 (IR spectra
of REG) corresponding to NH and C=O stretch of urea bond of REG which is
absent in the DP1 spectra. The peaks at 3394.96 and 1567.74 cmꢀ1 (IR spectra of
DP1) represented primary amine (ꢀNH2) group which is absent in the spectra of
10 REG. Group of peaks at 1140.05, 1128.94 and 1106.97 cmꢀ1 (IR spectra of
26

Page 27 of 32
New Journal of Chemistry
View Article Online
DOI: 10.1039/C7NJ01440F
1
REG) indicates the CꢀF bond in CF3 functional group while sharp peak was
observed at 1140.52 cmꢀ1 (IR spectra of DP1) represent the CꢀF bond. The CꢀCl
bond peak was observed at 650.69 cmꢀ1 (IR spectra of REG) which are absent in
the spectrum of DP1.
2
3
4
5
6
7
8
9
The IR spectra of DP1 compared with REG was utilized to find the similarities
at 1655.38 cmꢀ1 corresponding to C=O stretch of amide group which is present
at 1666.09 cmꢀ1 in the spectra of DP1. The peaks at 1595.21and 1539.27 cmꢀ1
(IR spectra of REG) indicated NH bending vibration of monoꢀsubstituted amide
bond while in the spectra of DP1 it was observed at 1535.88 cmꢀ1. Other peaks
10 can observe at 1504.48 cmꢀ1 (IR spectra of REG) and 1507.87 cmꢀ1 (IR spectra
11 of DP1) represented aromatic CH bond. The peaks were observed at 870.62,
12 835.41, 742.99 cmꢀ1 (IR spectrum of REG) and 890.08, 848.32, 803.03 cmꢀ1 (IR
13 spectra of DP1) indicates the 1,2 disubstitution or 1,3 disubstitution, 1,4
14 disubstitution and 1,2,4 trisubstitution respectively.
15 In silico toxicity prediction of REG and its DPs
16 The results of toxicity prediction by TOPKAT are shown in Table S3 (see the
17 ESI). Assessment of these results shows similarity in the toxicity profile of the
18 drug and its degradation products. But some of the animal models show the
19 variations for them. DP2 shows a high probability for rat male NTP (National
20 Toxicology Program). All compounds show very high and similar toxicity
21 profile for mouse male NTP, mouse female NTP and ocular irritancy mild vs
22 moderate severe (with the exception of DP2 shows nonꢀtoxic profile). Also for
27

New Journal of Chemistry
Page 28 of 32
View Article Online
DOI: 10.1039/C7NJ01440F
1
2
3
4
5
6
7
8
9
rat male FDA (Food and Drug Administration) single vs multiple prediction
DP4 exceptionally shows nonꢀtoxic profile. In a case of skin sensitization weak
vs strong model, DP2 and DP3 remarkably shows nonꢀtoxic endpoints whereas
other compounds show a high probability of toxicity. Skin sensitization none vs
sensitizer model shows extremely high toxicity endpoint for DP2. The DP1
shows extreme high toxicity probability for mouse female FDA none vs
carcinogen.
Variations have also been observed in the form low toxicity profile. DP1, DP2,
and DP4 show low probabilities for rat male FDA none vs carcinogen and skin
10 irritancy none vs irritant model. DP5 has shown in weight of evidence rodent
11 carcinogenicity model. DP2 and DP3 have been found to be least toxic for
12 developmental toxicity potential. However DP1 and parent drug show
13 intermediate probability for mouse male FDA none vs carcinogen. DP4 shows
14 for mouse female FDA none vs carcinogen.
15 The qualitative results from DEREK software are shown in Table S4 (see the
16 ESI). Qualitative toxicity profiling of regorafenib and its DPs shows a presence
17 of carcinogenicity, skin sensitization and peroxizome proliferation toxicity end
18 points. The parent drug regorafenib, DP1, DP4 and DP5 have shown
19 carcinogenicity and peroxizome proliferation toxic end points in mouse and rat
20 models. These toxicity end points have been observed due to a presence of
21 alkylaryl or bisaryl carboxylic acid or precursor moiety. DP1and DP2 have
22 shown skin sensitization toxicity end point due to the presence of aromatic
28

Page 29 of 32
New Journal of Chemistry
View Article Online
DOI: 10.1039/C7NJ01440F
1
primary or secondary amine. This DEREK toxicity profile for DP1 and DP2
further confirms the results obtained using TOPKAT. DP3 didn't show any
toxicity profile. Both studies (TOPKAT and DEREK) further confirm the
carcinogenicity toxicity profile of DP1, DP4, and DP5.
2
3
4
5
6
7
8
9
Conclusion
In conclusion, a formation of all DPs (DP1ꢀDP5) from REG under hydrolysis
(acidic, basic and neutral) and oxidation stress condition was discussed. The
structure of DPs (DP2ꢀDP5) was elucidated by using LC–MS/MS and the major
degradation product (DP1) was identified by using LC–MS/MS,
10 multidimensional NMR and FTIR analytical techniques. The LCꢀMS/MS and
11 NMR data supported the observed rearrangement reactions in DP1 and DP4.
12 Additionally in silico toxicity data were predicted for degradation products by
13 using DEREK and TOPKAT software tools. The parent drug (REG) and DPs
14 (DP1, DP4, and DP5) shows peroxizome proliferation toxic end points and
15 carcinogenicity in mouse and rat models. DP1and DP2 have shown skin
16 sensitization whereas DP3 has not shown any toxicity profile.
17 Conflicts of interest
18 There are no conflicts of interest to declare
19 Acknowledgements
20 The authors are thankful to Project Director of NIPER, Hyderabad for facilities
21 and support. B.S.M is thankful to the Department of Pharmaceuticals, Ministry
29