A More Sustainable Process for Preparation of the Muscarinic Acetylcholine Antagonist Umeclidinium Bromide

Article abstract of DOI:10.1002/cmdc.201800387

A more sustainable process for the synthesis of the long-acting muscarinic acetylcholine antagonist umeclidinium bromide is described. Specifically, we report the synthesis of ethyl 1-(2-chloroethyl)-4-piperidinecarboxylate, a key intermediate in the prep

Full text of DOI:10.1002/cmdc.201800387

Accepted Article
Title: A More Sustainable Process for the Preparation of the Muscarinic
Acetylcholine Antagonist Umeclidinium Bromide
Authors: Margarida Espadinha, Nuno M. T. Lourenço, Luis Sobral,
Rafael Antunes, and Maria M. M. Santos
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: ChemMedChem 10.1002/cmdc.201800387
Link to VoR: http://dx.doi.org/10.1002/cmdc.201800387
A Journal of
www.chemmedchem.org

10.1002/cmdc.201800387
ChemMedChem
FULL PAPER
A More Sustainable Process for the Preparation of the Muscarinic
Acetylcholine Antagonist Umeclidinium Bromide
Margarida Espadinha,[a] Nuno M. T. Lourenço,[b] Luis Sobral,[b] Rafael Antunes,[b] and Maria M. M.
Santos*[a]
[a]
M. Espadinha, M. M. M. Santos
Research Institute for Medicines (iMed.ULisboa)
Faculty of Pharmacy, Universidade de Lisboa
Av. Prof. Gama Pinto, 1649-003 Lisbon, Portugal
E-mail: mariasantos@ff.ulisboa.pt
[b]
N. M. T. Lourenço, L. Sobral, R. Antunes
Hovione Farmaciência SA
Sete Casas, 2674-506 Loures, Portugal
Supporting information for this article is given via a link at the end of the document.
There are several therapeutic options to relieve COPD
symptoms and provide a good quality of life for people suffering
from this disease. Synthetic corticosteroids, as fluticasone
propionate and fluticasone furoate, can be used in the
management of COPD. They act on glucocorticoid receptor
complexes, regulating C-reactive protein and inflammatory
cytokines and cells.[7] These corticosteroids can be used in
combination with long-acting muscarinic receptor antagonists
(LAMAs). The LAMAs also exhibit anti-inflammatory effects and
anti-remodeling effects such as inhibition of mucus gland
hypertrophy.[8] One example is the umeclidinium bromide (1)
Abstract: A more sustainable process for the synthesis of the long-
acting muscarinic acetylcholine antagonist umeclidinium bromide is
described. Specifically, we report the synthesis of ethyl 1-(2-
chloroethyl)-4-piperidinecarboxylate,
a key intermediate in the
preparation of umeclidinium bromide, in good yields using
triethylamine, as well as the identification and characterization of the
by-product formed in this reaction. This new method of synthesis
leads to an improvement of the yield, compared with the previous
reported protocols using potassium carbonate as base (65.6%
versus 38.6%). Moreover, in the last synthetic step of the process to
obtain umeclidinium bromide we were able to replace the use of
toxic solvents (acetonitrile/chloroform) by water. The use of this
green solvent allowed the precipitation of the API from the reaction
medium with high purity and in high yield. Overall, we have
developed a more efficient and green process for the synthesis of
the umeclidinium bromide active pharmaceutical ingredient with a
higher overall yield (37.8% versus previously reported overall yield of
9.7%).
(Scheme 1),
a potent muscarinic acetylcholine receptor
antagonist identified in 2009.[9] This molecule was approved by
the US FDA at the end of 2013 as a highly effective active
pharmaceutical ingredient (API) for the maintained treatment of
stable COPD.[10-12]
The first process to obtain umeclidinium bromide (1) was
disclosed by GlaxoSmithKline and comprises a synthetic route
with four steps starting from ethyl isonipecotate (2) (Scheme
1).[13] In the first step, ethyl 1-(2-chloroethyl)piperidine-4-
carboxylate (3) is obtained by a nucleophilic addition of ethyl
isonipecotate (2) to 1-bromo-2-chloroethane, in the presence of
potassium carbonate in acetone.[13] Under these mild conditions,
product 3 is isolated, by chromatography, in low yield (38.6%).
This low yield compromises the overall yield of the synthetic
process to obtain the API 1. More recently, two other synthetic
methods to obtain the intermediate 3 were reported (Scheme 1)
Introduction
Chronic obstructive pulmonary disease (COPD) is a multi-
component disease characterized to be preventable, however
there is no cure and is considered one of the major causes of
death worldwide.[1] An increase of this disease is expected due
to continued exposure to COPD risk factors, as cigarrete
smoking, indoor and outdoor pollution, chemicals, and an ageing
population.[2-4] According to World Health Organization (WHO),
globally, is estimated that the disease caused about 3 million
deaths in 2015 (that is, 5% of all deaths globally in that year).
WHO predicts that COPD will become the third leading cause of
death worldwide by 2030.[1]
[14-15]
but both of them present some security risks to be used in
a large scale synthesis (e.g. use of toxic reagents, use of high
temperatures, release of hydrogen). The second and third
synthetic steps involve the intramolecular cyclization of
compound 3 (yield of 95.7%), followed by reaction of 4 with
phenyl lithium (yield of 60.7%), to obtain compound 5.
umeclidinium bromide (1) is then obtained by reaction of
intermediate 5 with benzyl 2-bromoethyl ether, in a mixture of
chloroform and acetonitrile, in 43.3%.[13]
COPD is associated with an enhanced chronic inflammatory
response, which leads to airway abnormalities and also some
architectural distortion of the lung parenchyma. The cough,
sputum production, and dyspnea are very common symptoms in
individuals affected with this disease.[5-6] Due to the fast world
population growth, pharmaceuticals will have an increasingly
prominent role in the health care of the future generations.
Therefore, the development of efficient processes that can
deliver more safer, effective and affordable medicines is vital.
Although this process allows obtaining umeclidinium bromide (1)
with an overall yield of 9.7% it has some limitations for the
pharmaceutical industry in terms of toxicity of the
solvents/reagents used. Therefore, the search for more efficient
and green synthetic alternatives to this method of synthesis are
required.
1
This article is protected by copyright. All rights reserved.

10.1002/cmdc.201800387
ChemMedChem
FULL PAPER
Scheme 1. The first process described by GlaxoSmithKline for the preparation of umeclidinium bromide (1) and reported processes
for the preparation of intermediate 3.
In all cases, the yield of compound 3 reduced compared to the
reaction using acetone as solvent (Entries 3-6 versus Entry 1).
Finally, we studied the effect on the reaction of using other
bases. The yield of compound 3 was increased when the base
Results and Discussion
used was dipotassium phosphate (pKa = 12.4) (Entry 7) or
triethylamine (pKa = 10.7) (Entry 9), while it reduced when the
Herein we present our results on the optimization of this process
of synthesis of umeclidinium bromide (1) by improving the yields
of the first and last synthetic steps and applying greener
synthetic methods. In order to improve the yield of 3 we studied
the effect on the reaction of changing the temperature, the base
or the solvent (Table 1). Using the reaction conditions reported
by GlaxoSmithKline (acetone as solvent and potassium
carbonate as base)[13] we evaluated the effect of changing the
temperature from room temperature to 56 ºC (Entries 1-2). The
higher yield of compound 3 was obtained when the reaction was
performed at 25º C (Entry 1). Then, a screening of solvents was
performed (acetonitrile, THF, DCM and toluene).
base was potassium bicarbonate (pKa
= 10.3) (Entry 8).
Comparing with GlaxoSmithKline approach, the yield of
compound 3 was increased to 65.6% using triethylamine as
base. Using these reaction conditions, we isolated a by-product
in 14% yield (Scheme 2). We also tested the use of other bases
such as pyridine, DMAP, DIPEA and DBU. Unfortunately, the
amount of the crude obtained was too low, so the isolation of the
product was not carried out.
We then focused our attention on the optimization of the last
synthetic step described in Scheme 1. The reported yield for this
reaction is low and the solvents used,
a mixture of
acetonitrile/chloroform, are not ideal for industrial application.
These solvents should be limited in pharmaceutical products
because of their inherent toxicity. Using 1.5 equivalents of
benzyl 2-bromoethyl ether, we explored the effect of other
solvents and temperatures on the reaction (Table 2). Comparing
with the yields obtained by GlaxoSmithKline approach, in all
reactions the yields were improved to 53.4-82.2%. The highest
yield was obtained by using THF as solvent (82.2%, Entry 1).
The yields were lower when the reactions were performed in a
mixture of water/acetone (1:1) (53.4%, Entry 5) or ethanol
(62.9%, Entry 10). Quite interestingly, even when water was
used as solvent the final product was obtained in very good
yields (64.3-74.2%, Entries 6-9). While in water, the higher
yields were obtained with a temperature of 60ºC (Entry 6), in
toluene the reaction temperature had almost no impact in the
reaction yield (Entries 3-4). Moreover, the precipitation of
Table 1. Reaction conditions for the preparation of ethyl 1-(2-chloroethyl)-4-
piperidinecarboxylate (3).
Base
Solvent[a]
T [ºC] [b]
Yield [%][c]
38.6[13]
rt
K2CO3
K2CO3
K2CO3
Acetone
Acetone
56
rt
18.6
11.1
CH3CN
THF
K2CO3
K2CO3
K2CO3
K2HPO4
rt
rt
rt
rt
35.9
25
DCM
11.5
58.4
Toluene
Acetone
Acetone
Acetone
KHCO3
NEt3
rt
rt
33.8
65.6
umeclidinium bromide was induced at 2–4 ºC for the reactions
developed in water, while in the other cases it was required to
evaporate to dryness the solvent to obtain the final product in a
less pure form. This efficient process allowed obtaining directly
the umeclidinium bromide (1) from the reaction medium with
good purity (confirmed by HPLC).
[a] THF: tetrahydrofuran, DCM: dichloromethane. [b] rt – room temperature.
[c] Isolated yield after flash chromatography (silica gel, 1:1 n-
hexane/EtOAc).
2
This article is protected by copyright. All rights reserved.

10.1002/cmdc.201800387
ChemMedChem
FULL PAPER
Table 2. Reaction conditions for the preparation of umeclidinium
bromide (1).
Entry
Volume
[mL]
Solvent[a]
T [ºC]
Yield [%]
]
82.2[b
30
30
30
30
30
30
20
30
20
30
THF
60
60
1
2
]
75.7[b
acetone
toluene
toluene
]
79.6[b
60
3
]
81.0[b
reflux
60
4
]
53.4[b
water/acetone
(1:1)
5
]
74.2[c
water
60
6
]
71.6[c
water
60
7
]
64.3[c
water
reflux
reflux
60
8
]
68.3[c
water
9
]
62.9[b
ethanol
10
[a] THF: tetrahydrofuran. [b] Isolated yield after evaporation to dryness of
the solvent. [c] Isolated yield after precipitation from the reaction medium.
The reaction in water at a temperature of 60ºC (entry 6) was
also performed in a larger scale (starting from 2.56 mmol of
compound 5). In this scale the reaction outcome was almost
identical with a yield of 76.9%. Using the optimized steps (step 1
- entry 9, Table 1; step 4 - entry 6, Table 2) Umeclidium Bromide
(1) was obtained with an overall yield of 37.8%.
Scheme 2. Advantages of the proposed process comparing with
the GlaxoSmithKline approach reported in 2005.
Experimental Section
A comparison of the impurity profile of the final product obtained
from the two different routes (GSK and ours) showed that the
absolute purity obtained following GSK route is 95.9%, while
following our higher scale route is 98.9%. Moreover, the
amounts of toxic solvents present in each sample were
quantified. The sample prepared following GSK route presented:
CH3CN (662 ppm), AcOEt (44 ppm), and CHCl3 (407 ppm). The
sample prepared following our route only presented n-heptane
(39 ppm).
Finally, we evaluated if the optimized process could be
developed using directly the mixture of intermediate 3 and by-
product to synthesize the final product. Using this approach and
starting from 25.95 mmol of compound 2, intermediate 5 was
obtained in 23% yield and with high purity (99.3% by HPLC).
Reaction with benzyl 2-bromoethyl ether led to umeclidinium
bromide (1) in an overall yield of 19.3% and HPLC purity of
98.1%.
General Information: All reagents and solvents were obtained
from commercial suppliers and were used without further
purification. Thin layer chromatography was performed using
Merck Silica Gel 60 F254 plates and visualized by UV light. Merck
Silica Gel (230-400 mesh) was used for flash column
chromatography. 1H and 13C-NMR spectra were recorded on a
Bruker Fourier 300. 1H and 13C-NMR chemical shifts are
reportedꢀinꢀpartsꢀperꢀmillionꢀ(ppm,ꢀδ)ꢀreferencedꢀ to the solvent
used. Proton coupling constants (J) are expressed in hertz (Hz).
Multiplicities are given as: s (singlet), d (doublet), t (triplet), q
(quartet), and m (multiplet). MS experiment was performed on
Micromass® Quattro Micro triple quadrupole (Waters®, Ireland)
with an electrospray in positive ion mode (ESI+), ion source at
120 ºC, capillary voltage of 3.0 kV and source voltage of 30V, at
the Liquid Chromatography and Mass Spectrometry Laboratory,
Faculty of Pharmacy, University of Lisbon. LC-MS experiments
to determine the purity of the compounds were performed on a
Waters Alliance 2695 Separations Module, equipped with a PDA
detector set at 220 nm.
General procedure for the synthesis of ethyl 1-(2-
chloroethyl)piperidine-4-carboxylate (3): To a solution of ethyl
isonipecotate (2) (0.8 mL, 5.19 mmol) in the corresponding
solvent (8.6 mL) was added the appropriate base (7.79 mmol)
followed by 1-bromo-2-chloroethane (0.48 mL, 10.38 mmol). The
reaction mixture was stirred for 24h at the corresponding
temperature and then concentrated under vacuum. The resulting
residue was treated with water and extracted with diethyl ether.
The combined organic layers were dried with MgSO4, filtered
and concentrated under vacuum. The purification of the crude
was performed by flash chromatography on silica gel (1:1 n-
heptane/ethyl acetate) resulting in the desired compound
(colourless liquid). The 1H-NMR of ethyl 1-(2-chloroethyl)-4-
Conclusions
Overall, the new protocols developed in these two synthetic
steps allow a considerable improvement of the overall yield of
the process described in Scheme 1 (from 9.7%[13] to 37.8%).
Additionally, in the final synthetic step the solvents acetonitrile
and chloroform were replaced by water, the solvent of choice in
green chemistry. This change in the reaction conditions allowed
obtaining the API directly from the reaction medium by
precipitation. This optimized process (Scheme 2) represents a
clear advantage for the large scale synthesis of umeclidinium
bromide (1) by the pharmaceutical industry.
3
This article is protected by copyright. All rights reserved.

10.1002/cmdc.201800387
ChemMedChem
FULL PAPER
piperidinecarboxylate (3) was in accordance with the one
reported [13]. By-product: 1H-NMR (300 MHz, CDCl3)
4.11 (q,
J = 5.25 Hz, 4H), 2.99 2.76 (m, 4H), 2.47 (s, 4H), 2.30
(m, 2H), 2.07 1.99 (m, 6H), 1.90 1.84 (m, 4H), 1.79
was added ((2-bromoethoxy)methyl)benzene (1.40 mL, 9.01
mmol) and heated up to reflux. The solution was stirred during
24h. The reaction mixture was slowly cooled to room
temperature and stirred for 2h at a temperature between 2-4ºC.
The product was filtrated under vacuum and the excess of
bromide was removed by washing the compound with n-heptane
(20.0 mL). Umeclidinium bromide (1) was obtained as a white
solid and then dried under vacuum (2.55 g, 84.0%). HPLC
purity: 98.1%.
δ
–
–
–
2.20
1.66
–
–
(m, 4H), 1.23 (t, J = 6.0 Hz, 6H); 13C-NMR (75 MHz, CDCl3)
δ
175.3, 60.5, 56.4, 53.7, 41.3, 28.4, 14.4. MS (ESI) m/z calcd for
C18H32N2O2: 340, found 341 [M + H]+.
General Procedure for the synthesis of umeclidinium
bromide (1) [Entries 1-5 and 10, Table 2]: To a solution of
compound 5 (0.20 g, 0.68 mmol) in the corresponding solvent
(20 or 30 mL) was added benzyl 2-bromoethyl ether (0.16 mL,
1.02 mmol). The solution was stirred at the mentioned
temperature, during 24h. The reaction mixture was cooled down
and the solvent was removed under vacuum. The resulting solid
was washed with ethyl acetate (5x20 mL) and n-hexane (5x20
mL). The resulting white solid was then dried under vacuum to
afford umeclidinium bromide (1) as a white powder.
Acknowledgements
This work was supported by Hovione Farmaciência SA and, in
part, by FCT (Fundação para a Ciência e a Tecnologia) through
iMed.ULisboa
FEDER-022125,
(UID/DTP/04138/2013),
fellowship
LISBOA-01-0145-
SFRH/BD/117931/2016 (M.
Espadinha) and grant IF/00732/2013 (M. M. M. Santos).
General Procedure for the synthesis of umeclidinium
bromide (1) [Entries 6-9, Table 2]: To a solution of compound
5 (0.20 g, 0.68 mmol) in water (20 or 30 mL) was added benzyl
2-bromoethyl ether (0.16 mL, 1.02 mmol). The solution was
stirred at the temperature indicated in table 2, during 24h. The
reaction mixture was slowly cooled to 2-40C, forming a white
solid. The product was filtrated under vacuum and the excess of
bromide was removed by washing the compound with ethyl
acetate (20 mL) and n-hexane (5x20 mL). The white solid was
then dried under vacuum to afford umeclidinium bromide (1) as
Supplementary data
Supplementary data associated with this article can be found in
the online version. These data include: H and 13C NMR spectra
1
of by-product; HPLC chromatograms of compound 1.
Keywords: COPDꢀ•ꢀgreenꢀsolventsꢀ•ꢀindustrialꢀchemistry•ꢀ
triethylamineꢀ•ꢀumeclidiniumꢀbromide
1
a white powder. The H-NMR of umeclidinium bromide (1) was
in accordance with the one reported.[13]
References:
Experimental procedure for the preparation of umeclidinium
bromide (1) in large scale starting from ethyl isonipecotate,
and without removing the by-product: To a solution of ethyl
isonipecotate (4.0 mL, 25.95 mmol) in acetone (43.0 mL) was
added triethylamine (5.45 mL, 38.95 mmol) followed by 1-
bromo-2-chloroethane (4.32 mL, 52.14 mmol). The reaction
mixture was stirred for 17h at 25ºC. n-Heptane (43.0 mL) was
added and acetone was removed under vacuum. This procedure
was repeated twice. Then, water (43 mL) was added and the
resulting solution was extracted with n-heptane (2x43 mL). The
combined organic layers were dried with MgSO4, filtered and
concentrated under vacuum. After, n-heptane (11.50 mL) was
added, and the solution was placed at 0ºC during 1h and at -
20ºC for 16h, followed by filtration. The solution obtained was
concentrated under vacuum. The obtained residue (3.57 g) was
dissolved in THF (89.6 mL), under nitrogen atmosphere, and the
solution was cooled to -50ºC. LDA (1.0 M in hexanes/THF 20.72
mL, 20.72 mmol) was added at -50ºC during 25 min. The
reaction mixture was allowed to warm up to room temperature
over 16h. The reaction was quenched with saturated aqueous
solution of K2CO3 (74.4 mL) and extracted with ethyl acetate
(3x74.4 mL). The combined organic layers were dried with
MgSO4, filtered and concentrated under vacuum, to give an
orange oil (3.02 g).
[1]
[2]
http://www.who.int/respiratory/copd/en/ (Accessed July 30th, 2018).
P. J. Barnes, P. G. J. Burney, E. K. Silverman, B. R. Celli, J. Vestbo, J.
A. Wedzicha, F. M. Wouters, Nat. Rev. Dis. Primers 2015, 1, 15076.
E. Diaz-Guzman, D. M. Mannino, Clin. Chest. Med. 2014, 35, 7-16.
A. J Cohen, M. Brauer, R. Burnett, H. R. Anderson, J. Frostad, K. Estep,
K. Balakrishnan, B. Brunekreef, L. Dandona, R. Dandona, V. Feigin, G.
Freedman, B. Hubbell, A. Jobling, H. Kan, L. Knibbs, Y. Liu, R. Martin,
L. Morawska, C. A. Pope III, H. Shin, K. Straif, G. Shaddick, M. Thomas,
R. van Dingenen, A. van Donkelaar, T. Vos, C. J. L. Murray, M. H.
Forouzanfar, Lancet 2017, 389, 1907-1918.
[3]
[4]
[5]
S. Ejiofor, A. Turner, Clin. Med. Insights Circ. Respir. Pulm. Med. 2013,
7, 17-34.
[6]
[7]
P. J. Viviers, R. N. Zyl-Smital, S. Afr. Med. J. 2015, 105, 786–788.
J. A. Falk, O. A. Minai, Z. Mosenifar, Proc. Am. Thorac. Soc. 2008, 5,
506–512.
K. Alagha, A. Palot, T. Sofalvi, L. Pahus, M. Gouitaa, C. Tummino, S.
Martinez, D. Charpin, A. Bourdin, P. Chanez, Ther. Adv. Chronic Dis.
2014, 5, 85–98.
D. I. Laine, B. McCleland, S. Thomas, C. Neipp, B. Underwood, J.
Dufour, K. L. Widdowson, M. R. Palovich, F. E. Blaney, J. J. Foley, E. F.
Webb, M. A. Luttmann, M. Burman, K. Belmonte, M. Salmon, J. Med.
Chem. 2009, 52 (8), 2493-2505.
[8]
[9]
[10] A. C.Flick, H. X. Ding, C. A. Leverett, R. E. Kyne, K. K. C. Liu, S. J. Fink,
C. J. O'Donnell, Bioorg. Med. Chem. 2016, 24 (9), 1937-1980.
[11] R. Tal-Singer, A. Cahn, R. Mehta, A. Preece, G. Crater, D. Kelleher, I. J.
Pouliquen, Eur. J. Pharmacol. 2013, 701, 40-48.
The orange oil (3.02 g) was dissolved in THF (36.7 mL). The
resulting solution was slowly added to a solution of phenyllithium
(1.9 M in 70 cyclohexane/30 ether, 33.7 mL, 64.1 mmol), under
nitrogen atmosphere, at -30ºC during 25 min. The reaction
mixture was allowed to warm up to room temperature over 16h.
The reaction was quenched with water (15 mL) and then
evaporated to dryness under vacuum. Water (60.2 mL) and ethyl
acetate (60.2 mL) were added, causing a white solid to crash
out. This solid was filtered off under vacuum, to give 1-
[12] E. L Huisman, S. M. Cockle, A. S. Ismaila, A. Karabis, Y. S. Punekar,
Int. J. Chron. Obstruct. Pulmon. Dis. 2015, 10, 1863-1881.
[13] D. Laine, M. R. Palovich, B. McCleland, C. Neipp, S. Thomas, 2005:
US. a) WO 2005/104745, b) US 2007/0185155, c) US 2007/0249664.
[14] F. Hossner, J. B. Strachan, 2013: US. a) WO 2014/027045, b) US
2015/0232423.
[15] G. Bertolini, C. Colli, A. Bianchi, F. Colombo, S. Maiorana, F. Nissic,
2015: IT. WO 2016/071792.
azabicyclo[2.2.2]oct-4-yl(diphenyl)methanol (5) as
powder (1.76 g, three steps yield: 23.0%). HPLC purity: 99.3%.
To suspension of 1-azabicyclo[2.2.2]oct-4-
yl(diphenyl)methanol (5) (1.76 g, 5.83 mmol) in water (258.0 mL)
a
white
a
4
This article is protected by copyright. All rights reserved.