Large Scale Synthesis of an Ampakine-type Active Pharmaceutical Ingredient Based on a Telescoped Regioselective Double Amidation Reaction

Article abstract of DOI:10.1021/acs.oprd.9b00241

This work describes the process development and manufacture of an ampakine-type active pharmaceutical ingredient in clinical trials. A regioselective CDI-mediated amidation process was optimized for the subsequent couplings of two distinctive amines with

Full text of DOI:10.1021/acs.oprd.9b00241

Subscriber access provided by UNIVERSITY OF TOLEDO LIBRARIES
Full Paper
Large Scale Synthesis of an Ampakine-Type API Based on
a Telescoped Regioselective Double Amidation Reaction
Christophe HARDOUIN, Frédéric Pin, Jean-François Giffard, Yvon Hervouet, Julie Hublet, Solenn Janvier,
Christine Penloup, Julien Picard, Nathalie Pinault, Bruno Schiavi, Peng Zhang, weiwei Zhao, and Xue-yan Zhu
Org. Process Res. Dev., Just Accepted Manuscript • DOI: 10.1021/acs.oprd.9b00241 • Publication Date (Web): 11 Jul 2019
Downloaded from pubs.acs.org on July 11, 2019
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a service to the research community to expedite the dissemination
of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in
full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully
peer reviewed, but should not be considered the official version of record. They are citable by the
Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore,
the “Just Accepted” Web site may not include all articles that will be published in the journal. After
a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web
site and published as an ASAP article. Note that technical editing may introduce minor changes
to the manuscript text and/or graphics which could affect content, and all legal disclaimers and
ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or
consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W.,
Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 45
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
Large Scale Synthesis of an Ampakine-Type API
Based on a Telescoped Regioselective Double
Amidation Reaction
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Christophe Hardouin,§* Frédéric Pin,§ Jean-François Giffard,§ Yvon Hervouet,§ Julie Hublet,§
Solenn Janvier,§ Christine Penloup,§ Julien Picard,§ Nathalie Pinault,§ Bruno Schiavi,§ Peng
Zhang,† Weiwei Zhao,† Xueyan Zhu.†
§ Industrial Research Centre, Oril Industrie, 13 rue Desgenétais, 76210 Bolbec, France.
† Shanghai Institute of Pharmaceutical Industry, China State Institute of Pharmaceutical Industry,
285 Gebaini Road, Pudong District, Shanghai 201203, People’s Republic of China
Author for correspondence: christophe.hardouin@servier.com
† These authors developed the one-pot procedure involving the diazotization and annulation
sequence to synthesize the final product.
ACS Paragon Plus Environment

Organic Process Research & Development
Page 2 of 45
1
2
3
4
Table of Contents Graphic
5
6
7
8
9
N
NO2
Protecting
group free
F
NH NO2
F
N
N
O
HO2C
O
O
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
O
O
CO2H
On 127 kg scale
OH
1
OH H N
N
ACS Paragon Plus Environment

Page 3 of 45
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
ABSTRACT: This work describes the process development and manufacture of an ampakine-
type API in clinical trials. A regioselective CDI-mediated amidation process was optimized for
the subsequent couplings of two distinctive amines with a terephthalic acid substrate. Choice of
the synthetic route, key scale up issues, safety calorimetry and optimization of all steps for multi
kg production are discussed.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
KEYWORDS: regioselective CDI-mediated amidation, protecting-group-free synthesis,
telescoped process, AMPA receptors.
ACS Paragon Plus Environment

Organic Process Research & Development
Page 4 of 45
1
2
3
4
Introduction
5
6
7
8
Glutamate (Glu) is the major excitatory neurotransmitter in the central nervous system (CNS) and
is involved in the physiological regulation of processes such as synaptic plasticity, learning and
memory.1 Glutamate activates specific receptors that belong to the classes of metabotropic
receptors (coupled to G-proteins) and ionotropic receptors (ligand-gated ion channel), the latter
consisting of two primary families: N-methyl-D-aspartic acid (NMDA) receptor family and non-
NMDA family including the kainic acid (KA) receptor and the (R,S)-2-amino-3-(3-hydroxy-5-
methylisoxazol-4-yl)propionic acid (AMPA) receptor.2,3 AMPA-type receptors are found in high
concentration in the superficial layers of the neocortex, in each of the major synaptic zones of
hippocampus and in the striatal complex.4 They mediate fast excitatory postsynaptic currents at
the great majority of synapses and are involved in the expression of long-term potentiation, an
increase in the strength of synaptic contacts that follows repetitive physiological activity of a type
known to occur in the brain during learning.5,6 Thus a promising therapy under investigation
based on restoring diminished glutamatergic neurotransmission is the administration of AMPA
receptor potentiators for the treatment of Alzheimer’s disease, cognitive disorders, schizophrenia
or depression.7,8,9 Considerable advances have been made in solving the three-dimensional
structure of the AMPA receptor and X-ray structures have been reported.10,11 Based on strong
association with neurological disorders, extensive studies have been pursued to identify AMPA
receptor positive modulators as cognition enhancers.12 Among several distinct classes of
compounds investigated, Ampakines such as CX-516,13 CX-546,14 CX-61415 or CX-71716
(Figure 1) have shown to have very high activity in in vitro and in vivo models, acting as positive
allosteric modulators (PAM). After extensive research, compounds bearing a triazone moiety
increasing metabolic stability were disclosed and were considered as potential drugs to cure
neurodegenerated diseases by modulating AMPA receptors at nanomolar concentrations. A wide
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment

Page 5 of 45
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
range of Ampakines have been synthesized leading to 1 which entered in clinical trials.17,18
Unlike “low-impact” AMPA receptor potentiators such as CX-516, 1 is able to increase AMPA
receptor activation and is therefore considered as a “high-impact” AMPA receptor potentiator.19
Compound 1 has also been found to enhance cognition and memory in animals, to produce
antidepressant-, antianhedonic-, and anxiolytic-like effects, and to have neurotrophic and
neuroplasticity-promoting activities.20,21 Moreover, it has been found to increase levels of brain-
derived neurotrophic factor (BDNF) in the hippocampus and to stimulate hippocampal
neurogenesis.20,22
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
As part of a program aimed at identifying compounds that increase AMPA receptor-mediated
responses, 1 was identified as a good candidate to enter in clinical trials. In order to deliver large
quantities of 1, we embarked upon an effort to develop an optimal route for its long-term
manufacture. The successful completion of this goal is the subject of this paper.
Figure 1. Chemical structures of Ampakines
O
O
O
O
N
N
N
N
O
N
F
N
N
O
CX-516
CX-546
O
O
O
O
O
O
1
N
N
N
N
N
O
O
O
CX-614
CX-717
Low impact on the AMPA receptor
(type 1)
High impact on the AMPA receptor
(type 2)
Results and Discussion
ACS Paragon Plus Environment

Organic Process Research & Development
Page 6 of 45
1
2
3
4
5
6
7
8
9
The original synthesis of 1 is illustrated in Scheme 1 and starts with 2-hydroxy-4-methylbenzoic
acid 2.18 Amide formation with cyclopropylamine mediated with CDI and Et3N in methylene
chloride provided intermediate 3 as a solid. Insertion of the formaldehyde synthon was carried
out with trioxane and sulfuric acid in refluxed chloroform in the presence of sodium sulfate.
Chromatography of the crude on silica gel was required to obtain a satisfactory chemical purity.
Compound 4 was then subjected to nitric acid in a mixture of methylene chloride and acetic acid
to provide the desired 6-nitro isomer 5. The methyl group was then oxidized in a two-step
sequence: treatment with N,N-dimethylformamide dimethyl acetal in DMF at 125 °C during 16
hours followed by concentration to dryness and addition of sodium periodate in a mixture of THF
and water to provide the nitro aldehyde intermediate, which was smoothly converted into acid 6
with oxone in DMF. Coupling of 2-(3-fluorophenyl)ethanamine to the nitro acid derivative was
accomplished in methylene chloride using thionyl chloride and a catalytic amount of DMF.
Reduction of the nitro moiety was performed with a freshly prepared Zn/Cu reagent23 in a
mixture of THF, MeOH and formic acid. Filtration on silica gel followed by concentration to
dryness provided aniline 8 which was treated with an excess of isoamyl nitrite in DMF to
generate the triazinone ring. Addition of diethyl ether caused precipitation of the final product 1.
Although this route produced successfully 1 on gram scale, several issues had to be addressed to
manufacture larger quantities for clinical trials. Our main concern was the lack of reproducibility
of the nitration step and also the instability of the reaction mixture. Other drivers for the
development of a new route for the synthesis of 1 focused on operational issues. Besides some
low-yielding steps, eight isolations by concentration to dryness were necessary and are unsuitable
for a pilot-plant environment combined with a reprocess of the mother liquors after the nitration
step to isolate the remaining nitro intermediate 5. Chromatography on silica gel of 4 and 8 was
also required to obtain a satisfactory chemical purity which is a burden on scale. Finally
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment

Page 7 of 45
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
quantitative structure-activity relationship (QSAR) predictions (DEREK and Leadscope) showed
genotoxic alert for compounds 5, 6 and 7 that was confirmed since all intermediates gave a
positive Ames II test. Therefore, we were mindful of the potential difficulties engendered by the
presence of impurities in the active pharmaceutical ingredient (API). This latter driver prompted
us to reinvestigate the entire synthesis.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 1. Discovery route of 1a
NO2
a
b
c
O
O
O
O
OH OH
OH H N
O
N
O
N
2
3
4
5
Chromatography
Unsatisfactory reproducibility
Genotoxic


OH NO2
F
F
O
d
e
f
g
NH NO2
NH NH2
1
O
O
O
O
N
O
O
7
8
6
O
N
O
N
Low yield
Genotoxic


Genotoxic
Chromatography
aReagents and conditions: (a) CDI, CH2Cl2, 24 h at 20 °C then cyclopropylamine, Et3N, 3 days, 86%; (b) trioxane,
CHCl3, conc. H2SO4, Na2SO4, 2 h at reflux, 83% after chromatography; (c) conc. HNO3, AcOH, CH2Cl2, 0 °C, 41%;
(d) N,N-Dimethylformamide dimethyl acetal, DMF, 125 °C, 16 h, concentration then NaIO4, THF, water, 20 °C;
oxone, DMF, 20 °C overnight, 38% over 2 steps; (e) SOCl2, CH2Cl2, DMF cat., 20 °C overnight then 2-(3-
fluorophenyl)ethanamine, Et3N, 68%; (f) Zn/Cu, THF, MeOH, HCO2H, 20 °C overnight, chromatography; (g)
isoamyl nitrite, DMF, 2 h at 20 °C, 59% over 2 steps.
Early development synthesis of 1
One of our major goals was to define a robust and efficient method to synthesize 1. We reasoned
that desired product 1 could be disconnected into three components that required two distinctive
ACS Paragon Plus Environment

Organic Process Research & Development
Page 8 of 45
1
2
3
4
5
6
7
8
9
amide bond formations (see Figure 2). Our strategy relied on taking advantage of the non-
symmetrical terephtalic acid to selectively functionalize one carboxylic acid moiety over the
other one without any protecting group. This approach could be valuable in terms of both step
and atom economy.24 Therefore compound 9 was selected as key intermediate of the synthesis
and could be manufactured starting from brominated terephthalic acid 10.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 2. Proposed retrosynthesis to access 1
Diazotisation
F
NH2
N
NO2
Nitration
F
N
N
O
HO2C
Peptide
coupling
HO2C
O
CO2H
O
Aromatic
nucleophilic
substitution
CO2H
H2 N
Peptide coupling
Br
OH
N
1
10
9
The synthesis of 9 is depicted in Scheme 2. The first step of the sequence consisted of a
regioselective nitration performed on 10 in neat sulfuric acid at 40 °C by adding sequentially a
1:2 v/v mixture of nitric acid and sulfuric acid. For safety reasons this process was evaluated in a
reaction calorimeter (RC1). As shown in Figure 3, a steady heat flow of 15-20 W/kg was
observed over the entire course of the addition of HNO3 except when the product precipitates.25
The spike in heat flow was attributed to the onset of crystallization that continued throughout
addition of HNO3. The total amount of heat generated was 94 kJ/kg of reaction mass that
corresponded to a Tad of 54 °C. Despite some thermal accumulation (23% at stoichiometry), the
main concern was the ability to stir the thick reaction mixture. DSC (Differential scanning
calorimetry) showed an event with moderate energy (Tad of 36 °C) between 86 and 166 °C (see
Figure 4). Associated with a TMRad of 3 days at 40 °C, the probability to reach 166 °C where an
ACS Paragon Plus Environment

Page 9 of 45
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
exothermic decomposition occurred was low. Hydrolysis was also evaluated in a RC1 prior to the
pilot-plant batch. To ensure a safe control of the heat release (the total heat output was 312 kJ/kg
of reaction mass corresponding to a Tad of 90 °C), it was decided to add slowly 20% of the total
amount of water. In that case, the average heat release varied between 20 and 31 W/kg (Figure 5).
The remaining amount of water was then added and the average heat release was 10 W/kg as
shown on Figure 6. It is worth mentioning that the reaction of water with both sulfuric and nitric
acids was responsible for most of the heat release. This safety assessment study was also
consolidated by DSC where an estimated TMRad of 24 hours at 47 °C was associated to a
moderate event between 80 and 340 °C (Tad of 152 °C) including an exothermic decomposition
at 200 °C (see Figure 7). Following those observations, compound 11 was successfully
synthesized on 100 kg scale with 85% yield after precipitating out of the reaction mixture.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 2. Synthesis of 9a
NO2
NO2
HO2C
HO2C
HO2C
a
b
CO2H
CO2H
CO2H
Br
Br
OH
10
11
9
aReagents and conditions: (a) conc. H2SO4 (1 equiv), 65% aq. HNO3 (1.2 equiv), 3 h at 65 °C, 85%; (b) aq. NaOH
(10 equiv), 12 h at reflux, 75%.
Figure 3. Heatflow for nitration of 10 with a mixture HNO3 / H2SO4
ACS Paragon Plus Environment

Organic Process Research & Development
Page 10 of 45
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 4. DSC thermogram for nitration of 10 with a mixture HNO3 / H2SO4
ACS Paragon Plus Environment

Page 11 of 45
Organic Process Research & Development
1
2
3
4
Figure 5 . Heatflow for hydrolysis with 20% of total amount of water
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 6. Heatflow for hydrolysis with the 80% of remaining water
ACS Paragon Plus Environment

Organic Process Research & Development
Page 12 of 45
1
2
3
4
5
6
Figure 7. DSC thermogram of the reaction mixture after hydrolysis
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Substitution of the bromide was performed with an excess of sodium hydroxide at reflux. After
treatment with an aqueous solution of HCl and filtration of the reaction mixture, the remaining
NaCl was removed with several aqueous washes to provide anhydrous 9 (Kf < 0.4 m/m) with
75% yield and high purity (HPLC: > 99.0 area %) on 113 kg scale.
To overcome the issues associated with the medicinal route, we first developed an approach
based on the selective formation of the 1,3-benzodioxin-4-one intermediate 12 (Scheme 3).
Treatment of 9 with trioxane in a mixture of water and sulfuric acid provided 12 with 75% yield.
The remaining carboxylic acid function was then reacted with 2-(3-fluorophenyl)ethanamine in
acetonitrile in the presence of Et3N and T3P to provide amide 13. The nitro group was easily
ACS Paragon Plus Environment

Page 13 of 45
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
reduced with hydrogen and a catalytic amount of Pd/C in THF yielding aminobenzamide 14
(88% yield and 98.6% purity). Diazotization was initially performed with tert-butyl nitrite in
THF. The conversion was complete after 1 hour at 10 °C but the filtration of the slurry was
extremely slow after acidic hydrolysis. DSC showed that even if an estimated TMRad of 5 days at
20 °C was attributed for tert-butyl nitrite (Table 1, entry 1), a couple of highly exothermic
decompositions with onsets at respectively 94 and 195 °C prompted us to evaluate alternative
conditions in terms of process safety. The use of isoamyl nitrite in THF was rapidly eliminated
because of a decomposition exotherm with an onset at 74 °C and a TMRad of only 24 hours at 20
°C (entry 2). The results depicted in Table 1 showed that sodium nitrite in a mixture of water and
acetic acid was the best candidate (entry 3). It was selected because its degradation was moderate
and spread on a wide window of temperature associated with a TMRad > 7 days at 20 °C (see
Supporting Information).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Table 1. Evaluation of thermal stability of three nitrite derivatives
Temperature
(°C, onset/max/final)
Enthalpy
(J/g)
TMRad
at 20 °C
Tad
Entry
1a
Nitrite
(°C)
94/182/195
195/230/384
74/166/222
108/269/287
-234
-2455
-1000
-296
117d
1228
590e
115f
tert-Butyl
5 days
2b
3c
Isoamyl
Sodium
24 h
>7 days
a: neat. b: 1.4 g in 1.5 g of THF. c: 1 g in 5.26 g of water and 12.5 g of acetic acid. d: Cp evaluated at 2 kJ/kg.K. e: Cp
= 1.7 kJ/kg.K based on THF. f: Cp = 2.6 kJ/kg.K based on water and acetic acid.
When an aqueous solution of sodium nitrite was added on a solution of 14 in acetic acid,
triazinone 15 was obtained with 90% yield and 99.0% chemical purity. Although some
accumulation occurred, the heat release was easily manageable (Tad of 27 °C) and heat flow in
ACS Paragon Plus Environment

Organic Process Research & Development
Page 14 of 45
1
2
3
4
5
6
7
8
9
accordance with our cooling capacities.26 Some brown smokes were observed five minutes after
the end of the nitrite addition. It reflected the release of NO2 resulting from decomposition of
N2O3, the latter being generated from HNO2. Few minutes seemed to be necessary to initiate and
to go through the whole process of NO2 evolution.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Formation of the second amide bond was achieved by adding cyclopropanamine on 15 in THF to
afford intermediate 16, which was re-slurred in a mixture of acetonitrile and DMSO to improve
the purity (63% overall yield). The desired product was finally obtained after reacting 16 with
trioxane in CH2Cl2 under acidic conditions. Crude 1 was recrystallized in a mixture of EtOAc and
DMSO and was isolated in 77% yield with a satisfactory purity of 99.8%.
Scheme 3. Early development synthesis of 1a
O
NO2
NO2
NO2
HO2C
HO2C
a
b
F
N
H
O
O
CO2H
O
O
OH
O
O
9
12
13
Mutagenic cat 2
F
F
F
N
N
d
e
f
c
NH NH2
N
N
O
N
O
N
1
O
O
O
O
O
O
O
O
OH H N
14
15
16
Positive Ames assay for 12, 13, 14 and 15
aReagents and conditions: (a) trioxane, conc. H2SO4, water, 75%; (b) 2-(3-fluorophenyl)ethanamine, T3P, Et3N,
MeCN, 71%; (c) H2, THF, Pd/C, 88%; (d) NaNO2, AcOH, water, 90%; (e) cyclopropanamine, EtOAc, iPr2O, then
reslurry in a 9:1 MeCN/DMSO mixture, 63%; (f) trioxane, CH2Cl2, Na2SO4, conc. H2SO4, then recrystallization in a
8:2 DMSO/AcOEt mixture, 77%.
ACS Paragon Plus Environment

Page 15 of 45
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
At this stage, we had developed from 12 a 5 isolated step synthesis with 26% overall yield
including two recrystallizations that allowed us to manufacture batches of up to 30 kg of API.
However, the use of trioxane, considered as CMR (carcinogenic, mutagenic, reprotoxic) reagent,
in first and last steps was not ideal. Besides, intermediates 12, 13, 14 and 15 were positive in the
Ames assay (13 identified as mutagenic, category 2). We assumed it was linked to the relative
labile nature of the 1,3-benzodioxin-4-one moiety releasing formaldehyde after ring opening. It is
worth mentioning that during the synthesis we observed the corresponding hydroxy-benzoic acid
derivatives of compounds 12, 13, 14 and 15 at levels of up to few %. Since we anticipated that
dealing with genotoxic impurities could be a burden for the development of the API, our strategy
was to avoid the formation of the 1,3-benzodioxin-4-one moiety.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Final process development
Peptide couplings are one of the most common chemical reactions performed in the
pharmaceutical industry.27 Our goal was to generate selectively two amide bonds early in the
synthesis taking advantage of the difference in terms of reactivity of the carboxylic acid
functions. Acid activation by coupling agents was required to promote the condensation with an
amine.28 We selected several coupling agents described in the literature such as T3P, ethyl
chloroformate, oxalyl chloride, thionyl chloride, EDCI, HBTU, TBTU and CDI. After screening
the aforementioned coupling agents, we were pleased to discover that a stepwise process reacting
first CDI followed by addition of 2-(3-fluorophenyl)ethanamine in refluxing EtOAc provided
selectively amide 19 with 83% yield and 85% purity on lab scale (see Scheme 4).29 This superb
regioselectivity was attributed to the distinctive pKa between the two acidic functions (pKa of
respectively 1.1 and 2.2 for acids in ortho- and meta-position of the nitro group). The order of
ACS Paragon Plus Environment

Organic Process Research & Development
Page 16 of 45
1
2
3
4
5
6
addition was also crucial to get a satisfactory yield. Addition of 9 to a solution of 1.1 equiv of
CDI furnished up to 20% of impurity 31 (See structure on Scheme 9).
7
8
9
Gum formation after addition of CDI was the major issue and was attributed to intermediates 17
and/or 18. Since attempts to filter and purify 19 failed, significant effort was put into finding
alternate conditions to improve both process and purity.30 To identify scalable conditions, we
conducted a solvent survey and the results are illustrated in Figure 8. Ability to stir the reaction
mixture was the major driver: when CDI was dissolved in DMSO, DMF, sulfolane, NMP, DMPU
or DMA and added to the starting material in EtOAc solution, no gum formation was observed.
DMSO was finally selected because of its lower toxicity. Solubility of CDI in DMSO was also
evaluated for concentrations ranging from 5% w/w up to 43.8% w/w in conjunction with DSC
(see Supporting Information). For safety reasons, a solution of 28% w/w was finally selected
(heating the mixture at 37 °C was required to solubilize CDI instead 75 °C for a 43.8% w/w
solution).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 8. Coupling of 9 and 2-(3-fluorophenyl)ethanamine in EtOAc with CDI in additional
solvents (in red, no ability to stir; in yellow, formation of more or less gums; in green, solution).
ACS Paragon Plus Environment

Page 17 of 45
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
C4H6O3: propylene carbonate
Failure to efficiently isolate amide 19 prompted us to envision a telescoped process to install the
second amide bond. CDI was kept as coupling agent and scouting experiments were initiated on a
pure sample of 19. Among the solvents or mixtures of solvents we evaluated, MeTHF provided
the best results and advantageously dissolved CDI. When a solution of 19 in a mixture of EtOAc
and DMSO was sequentially reacted with CDI in MeTHF followed by a solution of
cyclopropanamine in MeTHF at 20 °C, desired 22 was obtained. Room temperature was
preferred because of the low boiling point of cyclopropanamine. It is noteworthy to point out that
this order of addition was crucial to get the best selectivity.31 Concentration followed by acidic
treatment allowed 22 to crystallize. After filtration, the wet cake was washed with MeCN to
provide 22 with 50-60% overall yield and a satisfactory 98% chemical purity on lab scale.
However, when this sequence was scaled up, we experienced a significant drop in terms of yield
(only 30 to 50% at best). Investigations showed that stirring the solution of CDI in DMSO at 40
°C for few hours before addition to terephthalic acid 9 was deleterious. Degradation of CDI was
ACS Paragon Plus Environment

Organic Process Research & Development
Page 18 of 45
1
2
3
4
5
6
suspected since the level of impurities in the reaction mixture increased resulting in a lower yield
of isolated 22.
7
8
9
In order to improve the robustness of this telescoped sequence, CDI was finally dissolved in
MeCN and added on 9 in DMSO. Implementing this procedure on large scale allowed us to
address the issue associated with the instability of CDI and we were pleased to isolate 22 with
65% yield on 127 kg scale with an adequate quality to be processed to the next step without
additional purification.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 4. Sequential regioselective amidationa
NO2
NO2
NO2
O
O
O
HO2C
a
N
O
N
N
N
O
O
CO2H
OH
OH OH
OH OH
17
18
9
ACS Paragon Plus Environment

Page 19 of 45
Organic Process Research & Development
1
2
3
4
5
6
F
F
7
8
b
c
NH NO2
NH NO2
9
N
O
O
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
O
O
N
OH OH
OH
O
O
20
19
F
F
d
NO2
OH
NO2
NH
NH
O
O
O
O
N
OH H N
21
22
N
aReagents and conditions: (a) 9 in DMSO/MeCN, then CDI (1.1 equiv) in MeCN, 4 h at 60 °C; (b) 2-(3-
fluorophenyl)ethanamine in MeCN, 18 h at 80 °C; (c) CDI (1.1 equiv) in MeCN, 9 h at 25 °C; (d) cyclopropanamine
(1.1 equiv) in MeCN, 1 h at 25 °C; 5 N HCl, 65% overall yield.
It should be mentioned that addition of 2.3 equivalents of CDI to diacid 9 yielded to intermediate
23. In that case, the regioselectivity of the peptide couplings was inverted. Treatment with
cyclopropanamine at room temperature provided amide 24 that was directly reacted at 80 °C with
2-(3-fluorophenyl)ethanamine to give desired compound 22 (see Scheme 5). However, since the
rate conversion was too slow (more than 48 hours for the second peptide coupling) and the
overall yield less than 30%, this approach was not further explored.
Scheme 5. Regioselective amidation with 2.3 equivalents of CDIa
ACS Paragon Plus Environment

Organic Process Research & Development
Page 20 of 45
1
2
3
4
5
6
7
8
9
NO2
NO2
OH
O
NO2
O
HO2C
N
b
c
a
N
N
22
N
N
O
N
CO2H
OH H N
24
OH
O
9
23
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
aReagents and conditions: (a) CDI (2.3 equiv) in DMSO, EtOAc, 25 °C, 1 h; (b) cyclopropanamine (1.2 equiv), 18 h;
(c) 2-(3-fluorophenyl)ethanamine (1.05 equiv), 65 °C, 48 h, 24% overall yield.
At this stage, two different routes could be envisioned to access 1 (see Scheme 6). The first one
consisted in reacting 22 with formaldehyde to generate 7. Since the reaction stalled at 50%
conversion in preliminary trials, we decided to reduce the nitro moiety first before conducting the
diazotization with sodium nitrite. Reduction in THF with Pd/C under 1 bar of hydrogen smoothly
provided compound 26 which was crystallized by addition of water after solvent switch from
THF to DMSO. 2-Aminobenzamide 26 was then re-slurred with 20v of EtOAc to obtain a
satisfactory purity but with inconsistent yields ranging from 50 to 70%. To improve both
efficiency and robustness, a solvent switch directly from THF to EtOAc (10 L/kg) resulted in a
smooth crystallization of 26 with 82% yield and with a quality greater than 99%.
Scheme 6. Envisioned routes and completion of the synthesis of 1a
ACS Paragon Plus Environment

Page 21 of 45
Organic Process Research & Development
1
2
3
4
5
F
6
7
a
NH NO2
8
9
O
O
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
F
O
N
7
NO2
NH
O
O
OH H N
22
F
F
N
N
N
NH NH2
c
b
d
1
O
O
O
O
OH H N
OH H N
16
26
aReagents and conditions: (a) Formaldehyde 37 wt % in water (2 equiv), 2 h at 90 °C, 40% after isolation; (b) Pd/C,
THF, H2, 1 bar, 2 h at 25 °C, 82%; (c) NaNO2 (1.3 equiv), AcOH, H2O, 3 h at 50 °C, 89%; (d) 1,3,5-trioxane (0.5
equiv), conc. H2SO4, EtOAc, AcOH, 92% then recrystallization in anisole/EtOH, 86%.
The triazinone ring was installed by reacting sodium nitrite in a mixture of AcOH and water. As
shown on Figure 9, safety aspect of this step was evaluated in a RC1 experiment. The heat output
was moderate (37 kJ/kg of reaction mass) but thermal accumulation not negligible. The former
was linked to a limited mass transfer because of the thickness of the reaction mixture. However
with an associated Tad of 12 °C, the average heat flow of 16 W/kg was in accordance with our
cooling capacities. Thermal stability of the reaction mixture at the end of the addition was also
evaluated by DSC and didn’t show any major issue (Figure 10). This is why we estimated that 50
°C was reasonable to run the reaction in safe conditions. Concerning intermediate 16, an
exothermic decomposition with an onset at 128 °C and an associated Tad of 563 °C (see Figure
ACS Paragon Plus Environment

Organic Process Research & Development
Page 22 of 45
1
2
3
4
5
6
7
8
9
11) prompted us to dry the powder at 45 °C (estimated TMRad of 8 days instead of 24 hours at 75
°C). Taking in consideration that efficiency of stirring was a key factor, the reaction proceeded
well in the pilot-plant and intermediate 16 was isolated with 89% yield and >98% purity on 24 kg
scale.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 9. Heat flow for triazinone 16 formation
Figure 10. DSC thermogram of the reaction mixture for diazotization step
ACS Paragon Plus Environment

Page 23 of 45
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 11. DSC profile of intermediate 16
ACS Paragon Plus Environment

Organic Process Research & Development
Page 24 of 45
1
2
3
4
5
6
7
8
9
In the early development synthesis, cyclisation was performed reacting 1,3,5-trioxane in a
mixture of concentrated H2SO4 and dichloromethane with Na2SO4 used as dehydrating agent.
However we encountered few incomplete conversions during the manufacturing campaigns
without any root cause clearly identified. In order to circumvent this lack of robustness, we
embarked into a survey of alternative conditions and the results are depicted in Table 2. We
focused on a mixture of AcOH and concentrated H2SO4 in CH2Cl2 that enabled us to generate the
desired 1 (entry 2).32 Even if the amount of trioxane was lowered, the conversion proceeded
faster at room temperature than in the original procedure (entry 1). Addition of water resulted in
precipitating 1 which was easily isolated by filtration. Removal of CH2Cl2 allowed us to get a
greener synthesis while improving the yield (entry 3). Replacement of AcOH with formic acid
proved to be deleterious with the formation of an unidentified impurity (entry 5). Reducing the
amount of trioxane to 0.3 equiv led to a slow conversion rate even at higher temperature (entry 6)
whereas 0.5 equiv seemed to be enough to get a full conversion (entry 7 and 8). Addition of
EtOAc was evaluated and proved to be beneficial in terms of mass transfer (entry 9). In that case
the reaction mixture turned out to be a solution and the rate conversion was improved. Further
experiments showed that the amount of trioxane could be lowered to 0.5 equiv. The combination
of H2SO4 and AcOH was the key factor to obtain a satisfactory outcome since the reaction stalled
in a EtOAc/H2SO4 mixture (entry 10) and degradation was observed in neat H2SO4 (entry 4).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Table 2. Initial studies to generate 1 in acidic conditions
Entry
1c
Trioxane
(equiv)
Conditions
Conv. Yield Purity
(%)
(%)a
(%)b
3
H2SO4 (1 equiv), Na2SO4 (5 equiv),
CH2Cl2, 6 h at reflux
98
70
98
2
1
H2SO4 (3.5 equiv), AcOH (2.5 mL/g),
54
51
78
ACS Paragon Plus Environment

Page 25 of 45
Organic Process Research & Development
1
2
3
4
CH2Cl2 (2.5 mL/g), 3 h at 20 °C
5
6
7
3
1
H2SO4 (3.5 equiv), AcOH (5 mL/g),
2 h at 20 °C
86
74
97
8
9
4
5
1
1
H2SO4 (35 equiv), 1 h at 20 °C
0
-
-
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
H2SO4 (3.5 equiv), HCO2H (5 mL/g),
1 h at 20 °C
40
32
59
6
7
0.3
0.3+0.2
0.5
H2SO4 (3.5 equiv), AcOH (5 mL/g),
2 h at 50 °C then 1 h at 90 °C
18
14
46
87
1
-
-
H2SO4 (3.5 equiv), AcOH (5 mL/g),
22 h at 20 °C then 2 h at 50 °C
58
57
74
-
97
98
98
-
8
H2SO4 (3.5 equiv), AcOH (5 mL/g),
2 h at 50 °C
9
1
H2SO4 (3.5 equiv), AcOH (2.5 mL/g),
EtOAc (2.5 mL/g), 2 h at 20 °C
10
0.5
H2SO4 (3.5 equiv), EtOAc (5 mL/g),
2 h at 20 °C then 4 h at 40 °C
aIsolated yields. b Determined by HPLC. c Solvent switch from CH2Cl2 to isopropyl ether and filtration.
The scouting experiments suggested that temperature, amount of H2SO4, ratio of EtOAc and
AcOH were variables potentially influencing both yield and purity. In order to map the reaction
parameters in terms of robustness and to identify the best reaction conditions for scale-up, we
initiated a Design of Experiment (DoE) optimization.33 The parameters chosen for this study and
their settings are shown in Table 3. The amount of trioxane (0.5 equiv) and the dilution (5 mL/g)
were held constant.
Table 3. Overview of DoE parameters and responses
Parameter studied (unit)
AcOH/EtOAc ratio (%)
Range
Responses measured
0 – 100
Yield (%, from w/w assay)
ACS Paragon Plus Environment

Organic Process Research & Development
Page 26 of 45
1
2
3
4
5
6
H2SO4 charge (equiv)
Temperature (°C)
0.5 – 4.0
20 – 40
Conversion (% area)
Purity (% area)
7
8
9
A 24 full factorial screening design was selected and 20 reactions were launched on 1 g scale. 1
was isolated by filtration after adding water to the reaction mixture. Analysis of the data revealed
that running the reaction at 40 °C gave better purity (see Figure 12). At the same time, the
amount of H2SO4 has to be lowered to avoid the formation of impurities. This was also consistent
with the lower yield of isolated product.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 12. Contour plots for purity as a function of AcOH/EtOAc ratio and H2SO4 charge at 20,
30 and 40 °C
On the basis of these data, we opted to run the reaction at 40 °C and finally adjusted the amount
of H2SO4 to 1.5 equiv and the ratio of AcOH/EtOAc to 1:1 (v/v). Following those conditions,
intermediate 16 was converted into desired 1 after 5 hours with a purity of 93.7 area % in the
reaction mixture. Water was used as anti-solvent and its amount had a dramatic impact on the
yield. A 7.5 mL/g dilution was selected to furnish 1 with 94% yield and with an excellent >99
area % purity by HPLC (identical purity with 5 and 10 mL/g with respectively 73 and 93%
ACS Paragon Plus Environment

Page 27 of 45
Organic Process Research & Development
1
2
3
4
yield). Upon scale-up to 48 kg, we were pleased to see that the reaction proceeded well and
5
6
provided crude 1 with 92% yield and >98% purity.
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Toward a telescoped process to generate the triazinone moiety and to access 1
In our quest for reducing batch cycle times, workup and purification unit operations as well,
process intensification via reaction telescoping was explored to synthesize 1. Recently, a method
to prepare 1,2,3-benzotriazine-4(3H)-ones was developed using the Nef reaction.34 In the
presence of Cs2CO3 and AcOH, nitromethane could generate HCHO and HNO. Then, HNO
could be further oxidized to HNO2 by KIO generated in situ from KI and TBHP. Finally, the
direct condensation of 27 with HNO2 affords the desired product 29 by removal of two molecules
of water (Scheme 7). As a minor pathway, 28 can be obtained via a tandem condensation-
addition-oxidation process of 27 with HCHO.
Scheme 7. In situ generating of HNO2 and HCHO by Nef reaction.
t-BuOH
TBHP
Cs2CO3
AcOH
KIO
KI
HNO
HNO2
CH3NO2
HCHO
O
O
N
O
N
HNO2
- 2 H2O
HCHO, TBHP
N
H
N
N
N
- 2 H2O, t-BuOH
NH2
28
29
27
According to the authors, HNO2 demonstrated more active reactivity than HCHO when subjected
to compound 27. On the basis of this study, we envisioned a one-pot reaction to access to 1. In
ACS Paragon Plus Environment

Organic Process Research & Development
Page 28 of 45
1
2
3
4
5
6
7
8
9
the scouting experiments, we initiated an experiment by treating 26 with NaNO2,
paraformaldehyde and AcOH under catalysis of concentrated H2SO4 to depolymerize
paraformaldehyde into formaldehyde. As outlined in Scheme 8, the reaction was initially
performed at room temperature and starting material 26 was fully converted to benzotriazinone
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
°
16. Increasing the reaction temperature to 80 C provided desired product 1. We therefore
envisioned that this one-pot reaction could be completed if the initial reaction temperature was
set at 80 °C. This assumption was confirmed on 2 g scale and desired product 1 was isolated with
99% yield and 95% purity starting from 26. These new conditions could therefore be an
opportunity to improve the cost of the synthesis.
Scheme 8. One-pot synthesis of 1
F
F
F
N
N
NH NH2
N
N
N
N
O
a
b
O
O
O
O
O
O
OH H N
OH H N
N
26
16
1
c
aReagents and conditions: (a) NaNO2 (1.2 equiv), paraformaldehyde excess, AcOH, H2SO4, 20 °C; (b) 80 °C; (c)
NaNO2 (1.2 equiv), paraformaldehyde excess, AcOH, H2SO4, 80 °C, 99%.
Control of impurities
In order to insure the safety of the patient, control of impurities generated in the steps discussed
before was carefully monitored. As shown on Scheme 9, compounds 30, 31, and 32 were
identified as the most prevalent impurities generated during the regioselective double amidation
step. Compound 30 arises from 22 undergoing a second peptide coupling with 2-(3-
fluorophenyl)ethanamine mediated with CDI. Compounds 31 and 32 arise respectively from
ACS Paragon Plus Environment

Page 29 of 45
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
condensation of CDI on 30 and 22. We assumed that condensation of acetic acid with 26 was
responsible of the formation of compound 39. Detailed studies revealed that the abovementioned
impurities reacted in the downstream process and were purged except compound 35 with a level
below 0.1%.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
To ensure that none of the impurities introduced in the process could impact the purity profile of
the drug substance, 1 was initially recrystallized in a mixture of EtOAc and DMSO (8:2, 11 L/kg)
with a moderate 77% yield. In order to improve this result, further studies were initiated and the
solubility profile of 1 was evaluated in various solvent systems. A mixture of EtOH and anisole
(3:1, 4.5v) heated at 80 °C was identified in which the polish filtered solution of 1 was stirred,
cooled at 10 °C and filtered. Gratifyingly, this new process produced 1 in 86% yield with a
satisfactory chemical purity of 99.8%. Levels of all the impurities were below 0.10% thanks to
the efficient control offered by the recrystallization. Comparison of the powder X-ray diffraction
(PXRD) traces of 1 revealed that the form was identical as the one obtained in the EtOAc/DMSO
mixture (see Figure 13).
Scheme 9. Fate of process related impurities 30, 31 and 32 found in intermediate 22 after
hydrogenation, sodium nitrite treatment and annulation mediated with paraformaldehyde
ACS Paragon Plus Environment