Development of a Multi-Kilogram-Scale Synthesis of AZD1283: A Selective and Reversible Antagonist of the P2Y12 Receptor

Article abstract of DOI:10.1021/op400288v

Ethyl 6-chloro-5-cyano-2-methylnicotinate (4) was coupled with 4-piperidinecarboxylic acid (isonipecotic acid) in 81% yield to pyridine acid 10. An amide coupling between 10 and benzylsulfonamide (6) afforded AZD1283 (1) in 79% yield using CDI as coupling reagent. The synthesis has been developed and scaled up to 20 kg batches of 1, supporting preclinical and clinical studies. Development work towards 2-chloropyridine 4 and benzylsulfonamide (6) is included.

Full text of DOI:10.1021/op400288v

Article
pubs.acs.org/OPRD
Development of a Multi-Kilogram-Scale Synthesis of AZD1283: A
Selective and Reversible Antagonist of the P2Y12 Receptor
Søren M. Andersen,*,† Carl-Johan Aurell,*,†,‡ Fredrik Zetterberg,† Martin Bollmark,‡ Robert Ehrl,‡
Peter Schuisky,‡ and Anette Witt‡
†AstraZeneca R&D Molndal, SE-43183 Molndal, Sweden
̈
̈
‡AstraZeneca Process R&D Sodertalje, SE-15185 Sodertalje, Sweden
̈
̈
̈
̈
S
* Supporting Information
ABSTRACT: Ethyl 6-chloro-5-cyano-2-methylnicotinate (4) was coupled with 4-piperidinecarboxylic acid (isonipecotic acid) in
81% yield to pyridine acid 10. An amide coupling between 10 and benzylsulfonamide (6) afforded AZD1283 (1) in 79% yield
using CDI as coupling reagent. The synthesis has been developed and scaled up to 20 kg batches of 1, supporting preclinical and
clinical studies. Development work towards 2-chloropyridine 4 and benzylsulfonamide (6) is included.
was performed using this approach and 0.5 kg of material for
preclinical studies delivered.
INTRODUCTION
■
Drug candidate AZD1283 (1) (Figure 1) has been identified as
a potent reversible P2Y12 antagonist.1 Larger amounts were
The left-hand fragment ethyl 6-chloro-5-cyano-2-methylni-
cotinate (4), was prepared according to published procedures2,3
after minor modifications: ethyl acetoacetate and DMF
dimethyl acetal was condensed neat without the need for a
catalytic amount of acid and instead of isolating sodium
cyanoacetamide and reacting it with enaminone 2; we formed
sodium cyanoacetamide with NaOEt/EtOH and then added a
solution of enaminone 2. A small amount of methyl ketone
byproduct was formed,4 but could be removed with an aqueous
bicarbonate wash. Pyridone 3 was converted into the
corresponding 2-chloropyridine 4 in a good yield under
Vilsmeier conditions using SOCl2/DMF instead of oxalyl
chloride/DMF5 and the crude product was used directly in the
downstream chemistry. It was necessary to add 1 equiv of DMF
in the chlorination of 3, since the reaction often stalled when
substoichiometric amounts were employed. Although the
pyridone 3 was a colorless solid, 2-chloropyridine 4 appeared
as a bright orange solid. Despite attempts to identify the source
of the color, we were unable to determine it in any way. In the
limited time frame available we were unable to find suitable
conditions for recrystallizing 2-chloropyridine 4 during
Campaign 1. Starting from Boc-protected isonipecotic acid 5
and benzylsulfonamide (6), we assembled the right-hand
fragment 7. A screen of coupling conditions indicated that a
TBTU coupling performed in THF was a high yielding
procedure. Especially in the presence of LiCl as a Lewis acid,
the yield was excellent. In order to avoid TBTU reacting with
benzylsulfonamide (6) forming guanidine adducts, the
protected isonipecotic acid 5 was allowed to react with
TBTU for 1−2 h before the addition of 6. During workup
the solvent was swapped to EtOAc and the crude product
washed with dilute aqueous acid (substantial loss under neutral
or basic conditions) yielding acyl sulfonamide 7 in a good yield,
after precipitation from 50% EtOH. Treatment of 7 with six
Figure 1. AZD1283.
needed to support both preclinical and clinical studies and a
synthetic route capable of delivering multikilo quantities as well
as being a good choice on a commercial level was sought.
The original in-house synthesis of 1 was performed by
coupling the left-hand pyridine fragment to the central
isonipecotic acid and then coupling the resulting pyridine
acid to benzylsulfonamide (6) (Figure 2, Pyridine Acid
Route).1 The last step was low yielding and chromatography
needed to isolate AZD1283 (1). In order to develop a scaleable
route that would generate material for preclinical studies in a
short time frame, it was critical to develop a crystallization
method for the final product to avoid chromatography. In the
original synthesis1 it was observed that the SNAr coupling
between ethyl 6-chloro-5-cyano-2-methylnicotinate (4) and
isonipecotic acid was high yielding and we envisaged using the
SNAr reaction between 2-chloropyridine 4 and acyl sulfonamide
8 as the final step to AZD1283 to increase the probability of
crystallizing 1 from the reaction mixture (Scheme 1).
RESULTS AND DISCUSSION
■
GLP Campaign, Campaign 1. Small scale experiments
confirmed the high yielding coupling between 4 and 8 (Scheme
1) and we were pleased to discover that the final compound 1
could be precipitated from the reaction mixture. Campaign 1
Received: October 14, 2013
Published: November 13, 2013
© 2013 American Chemical Society
1543
dx.doi.org/10.1021/op400288v | Org. Process Res. Dev. 2013, 17, 1543−1551

Organic Process Research & Development
Article
Figure 2. Tested routes to AZD1283.
Scheme 1. Campaign 1
volumes of formic acid afforded isonipecotic amide 8, which
was precipitated as a zwitterion at pH 6 from an aqueous
solution in high purity and yield. Significant foaming was
observed during deprotection of 7 and it was barely manageable
1544
dx.doi.org/10.1021/op400288v | Org. Process Res. Dev. 2013, 17, 1543−1551

Organic Process Research & Development
Article
Scheme 2. Campaign 2
on a 10 L scale. Finally, isonipecotic amide 8 was coupled with
1 equiv 2-chloropyridine 4 (in the presence of triethylamine)
by briefly heating up the reaction mixture to reflux. Upon
cooling to ambient temperature and addition of aqueous
potassium hydrogen sulfate, the crude product 1 oiled out and
then precipitated from oily droplets. The purity was increased
to >98% by treatment with charcoal and recrystallization from
EtOAc to afford AZD1283 (1) for preclinical studies
(polymorph I).6 The orange color from 2-chloropyridine 4
persisted. In spite of considerable efforts, we were unable to
identify the source of the color in any way. Attempts to further
telescope the sequence from the coupling between Boc-
isonipecotic acid 5 and benzylsulfonamide (6) to AZD1283,
resulted in a final product that could not be precipitated. The
chlorination could also be performed in EtOAc instead of
toluene, but the final precipitation of AZD1283 (1) was difficult
and generated larger amounts of byproducts.
First GMP Campaign, Campaign 2. The first GMP
Campaign was based on Campaign 1 to supply material in time
for initial clinical studies. Four kilograms of 1 was needed and
based on Campaign 1, some key changes were necessary: (1) A
recrystallization method was needed for chloropyridine 4 in
order to avoid a strongly colored final compound; (2) an
alternative to the Boc-deprotection in formic acid was essential
to avoid the generation of foam; (3) a better crystallization
method of the final product 1 was needed since it precipitated
from oily droplet producing large particles. Analogously to
Campaign 1, ethyl acetoacetate and DMF dimethyl acetal were
reacted neat on a 15 kg scale and the methanol formed was
distilled off under reduced pressure. After addition of
cyanoacetamide to the intermediate enaminone 2 (in the
presence of base), a better crystallization procedure was
developed, precipitating pyridone 3 from acetonitrile/water in
67% yield. The material was recrystallized from acetonitrile/
water affording 3 in 50% overall yield starting from ethyl
acetoacetate and DMF dimethylacetal. 2-Chloropyridine 4 was
prepared under Vilsmeier conditions, but contrary to Campaign
1, the product was crystallized and consequently was less
colored. When scaling up the Vilsmeier reaction to 6−8 kg
scale in a 100 L reactor, polymeric byproducts were formed.
After completion of the reaction, the mixture was concentrated
under reduced pressure and the viscous polymeric material
caused the agitator to stop. On small scale these polymeric
byproducts were only observed as a thin surface film. By
repeatedly extracting the polymeric byproduct with toluene, the
encapsulated product was dissolved leaving the polymeric
material in the reactor. The crude product 4 was dissolved in
ethanol and, upon controlled addition of water, crystallized in
67% yield. The material was dissolved in 95% ethanol, treated
with activated carbon, and subsequently recrystallized from
cyclohexane to give colorless 2-chloropyridine 4 in 39% yield
starting from pyridone 3 (Scheme 2).
In Campaign 2, the coupling of Boc-isonipecotic acid 5 and
benzylsulfonamide (6) was carried out in acetonitrile (THF
used in Campaign 1). After completion of reaction, the solvent
was swapped to isopropyl acetate and, after workup, the
product crystallized from ethanol/water to afford Boc protected
7 in 81% yield. Two batches were prepared affording 9 kg of 7
in total. Using formic acid for Boc removal of 7 became
increasingly problematic on scale due to foaming, but could be
replaced with 5 M HCl/iPrOH. It was observed that the
product 8 had a low solubility in 2-MeTHF and reaction
conditions were found where the deprotection of 7 took place
in a mixture of iPrOH and 2-MeTHF (3:1) at 55−60 °C and
upon cooling 8 crystallized in 98% yield on a 6 kg scale. During
Campaign 1, the final crystallization of 1 was sluggish. At first, 1
formed oily droplets together with inorganic salts and then
precipitated. As a result particles several millimeters across in
diameter were formed and it was challenging to get the material
out of the reactor. Like Campaign 1, the coupling of 4 and 8
was performed in refluxing ethanol in the presence of
triethylamine. After reaching 95% conversion, the mixture was
cooled to 50 °C and upon acidification with an aqueous
solution of ortho-phosphoric acid (instead of potassium
1545
dx.doi.org/10.1021/op400288v | Org. Process Res. Dev. 2013, 17, 1543−1551

Organic Process Research & Development
Article
hydrogen sulfate), the product 1 precipitated without initially
oiling out. Treatment with activated charcoal and recrystalliza-
tion from ethanol gave the stable polymorph II6 of AZD1283
(1) in a good yield and more than 99% purity. A total amount
of 8 kg of 1 was prepared in this manner to support phase I
clinical studies.
Second GMP Campaign, Campaign 3a, Prestudy, and
Common Intermediates. For Campaign 3a it was estimated
that more than 60 kg of API was needed. In parallel with the
Campaign 2 manufacture, a prestudy was carried out with the
aim to identify the best route for Campaign 3a and onwards. As
a result of the large amounts needed for clinical studies and
stressed timelines, the main criteria for the route to AZD1283
were as follows: robustness, manufacturability, and low cost of
API. When taking the criteria into account, 5 different synthetic
routes were identified and investigated (Figure 2): the
Campaigns 1 and 2 route, the pyridine acid route, the pyridine
amide route, the enamine acid route, and the enamine amide
route.
The selected routes to AZD1283 (1) have some
intermediates in common, such as the 2-chloropyridine 4 and
the benzylsulfonyl chloride or amide. A route to the pyridone 3
and 2-chloropyridine 4 was developed in collaboration with
Almac7 and was recently published.4 It was found that when
malononitrile was used as the starting material, instead of
cyanoacetamide, a robust and scaleable procedure emerged
producing colorless 2-chloropyridine 4 (Scheme 3). In this
manner, three 25 kg batches of 4 were produced.
methane, and acetonitrile with 2−12 equiv of 25% ammonium
hydroxide solution added at −10 to +10 °C. Both addition of
aqueous ammonia to the benzylsulfonyl chloride (9) and
reverse addition were tested and it was found that addition of 9
in a THF solution to excess aqueous ammonia not only
minimized the formation of benzylsulfonic acid from 3% to less
than 1%, but also reduced the level of other byproducts.
Using this approach, 139 kg of benzylsulfonyl chloride (9)
was converted to benzylsulfonamide (6). It was difficult to
control the crystallization of 6, which precipitated immediately
on cooling, but after diluting the solids with more toluene and
stirring the suspension overnight at 65 °C, a second cooling
step afforded an excellent product quality and yield.
Evaluation of the Campaigns 1 and 2 Route. The
Campaigns 1 and 2 route was improved by substituting
TBTU/LiCl/Acetontrile with EDC/2-MeTHF10 and as such
had the potential to become a robust route since all the
problematic steps could be improved. The conversions in all
steps were good and the synthesis short. On the negative side,
during Campaign 2 the pyridone 3 and the 2-chloropyridine 4
were flagged as potentially genotoxic intermediates and the
route also had the highest cost of API, since the two most
expensive starting materials, Boc-protected isonipecotic acid 5
and benzylsulfonamide (6), were used early in the synthesis.
Evaluation of the Pyridine Acid and Pyridine Amide
Routes. 2-Chloropyridine 4 is the starting point of both the
pyridine acid and the pyridine amide route. The pyridine acid
route is similar to the original Medicinal Chemistry route,1,5 but
with the improved route to the 2-chloropyridine 4,4 it was
found that this route could be dramatically improved and
AZD1283 (1) crystallized without the need for chromatog-
raphy. The SNAr reaction between 2-chloropyridine 4 and
isonipecotic acid could be performed in ethanol in the presence
of triethylamine and the subsequent coupling with benzylsulfo-
namide (6) could be accomplished using EDC/DMAP or CDI
in NMM or 2-MeTHF (Scheme 5). Alternatively, the coupling
could be performed via the acid chloride, although it was tricky
to reproduce.
Scheme 3. Malononitrile route to 2-chloropyridine (4)4
The pyridine amide route is very similar, except that
isonipecotic amide is used instead of isonipecotic acid and
the last step consists of sulfonylation of the resulting pyridine
amide. Similar to the pyridine acid route, chloropyridine 4
could be coupled with the isonipecotic amide in ethanol and
triethylamine. Sulfonylation with benzylsulfonyl chloride 911
was done in THF in the presence of sodium tert-amylate at low
temperature (it was found that the reaction temperature had to
be less than −25 °C). By comparison of the pyridine acid/
amide routes to the Campaigns 1 and 2 route, they are of
similar length, but with a lower cost of API (no protecting
group used). In the case of the pyridine acid route via CDI/
EDC coupling, potentially genotoxic 4 was used early in the
synthesis and the sequence is robust. Via the acid chloride, the
final step was unexpectedly tricky and in need of more
investigation. The pyridine amide route required low temper-
ature in the final step, reducing the manufacturability.
Benzylsulfonamide (6) for Campaign 3a was outsourced to
SAFC,8 who prepared 6 from benzylsulfonyl chloride (9) via
ammonolysis (Scheme 4). The reaction between benzylsulfonyl
chloride (9) and ammonia9 was tested in THF, dichloro-
Evaluation of the Enamine Acid and Enamine Amide
Routes. In the last two routes, which were considered, the
pyridine ring was build with the isonipecotic acid moiety in
place (Scheme 6).
Scheme 4. Synthesis of benzylsulfonamide (6)
Initial experiment showed that it was possible to build the
pyridine ring system with isonipecotic acid/amide in place. For
both routes ethanol was added to malononitrile affording the
corresponding enamine HCl salt.12 In the enamine acid route,
1546
dx.doi.org/10.1021/op400288v | Org. Process Res. Dev. 2013, 17, 1543−1551

Organic Process Research & Development
Article
Scheme 5. Pyridine acid and pyridine amide route
Scheme 6. Enamine acid and enamine amide routes
the enamine was substituted with isonipecotic acid in the
presence of potassium carbonate or DBU and in the enamine
amide route, with isonipecotic amide.13 These products in turn
were reacted with enaminone 2 to yield the final intermediates:
pyridine acid and pyridine amide. The pyridine acid could be
coupled with benzylsulfonamide (6) to afford AZD1283 (1)
using EDC/DMAP or CDI in NMM or 2-MeTHF or via the
acid chloride in pyridine/toluene. For the enamine amide route,
the pyridine amide could be reacted with benzylsulfonyl
chloride (9) in THF at low temperature. The lengths of the
enamine acid/amide routes are comparable to previously
discussed alternatives, but have lower cost of API than the
Campaigns 1 and 2 route (no protection group used). When
adding isonipecotic acid/amide to enamine 2, the use of
potassium carbonate as a base caused some concerns about the
huge amount of salt waste and the reaction was difficult to work
up when using DBU (byproducts). The enamine acid route
using thionyl chloride in the last step shared similar concerns
with the pyridine acid route, being tricky to reproduce. The
enamine amide route required low temperatures for the final
step. In general, the stability of the enamine intermediates, and
thereby the robustness of the sequences, was unknown, and
much more development work is needed.
Second GMP Campaign, Campaign 3a. The pyridine
acid route was selected for further development and scale-up
due to the low cost of API, but also as the most secure route
alternative with the highest probability for good manufactur-
ability and robustness in all steps. In addition, both of the
potential genotoxic compounds (pyridone 3 and 2-chloropyr-
idine 4) were used early in the synthesis. In Campaign 3, the
solvent for the SNAr substitution of 4 with isonipecotic acid was
changed from EtOH to iPrOH to minimize the formation of
1547
dx.doi.org/10.1021/op400288v | Org. Process Res. Dev. 2013, 17, 1543−1551

Organic Process Research & Development
Article
Scheme 7. Campaign 3a
the corresponding diester of 10 (Scheme 7). Using ethanol
resulted in 0.6% diethyl ester of 10, whereas using iPrOH only
resulted in 0.2% ethyl isopropyl diester of 10. It was also found
that 2 equiv of base were needed for the reaction to work well
(1 equiv for the formed HCl and 1 equiv to deprotonate the
carboxylic acid) and a small excess of isonipecotic acid was
necessary to ensure that all 2-chloropyridine 4 was consumed.
Triethylamine could be added at both ambient and high
temperature, but to have control over the energy release, the
base was added slightly below reflux temperature. In this
manner 18 kg of 2-chloropyridine 4 was reacted with 1.03 equiv
of isonipecotic acid and after refluxing for 4.5 h, all starting
material had been consumed. On lab scale it was observed that
rapid addition of 1.05 equiv of aqueous HCl caused the product
to crash out. The crystallization behaved much better if the
reaction mixture was seeded after 55% of the aqueous HCl had
been added (if less aqueous HCl added, the seed crystals
dissolved). After aging the crystallization mixture, the remaining
aqueous HCl could be added during 2.5 h. After cooling, the
pyridine acid 10 was filtered off and washed with iPrOH−water
mixtures. In order to get rid of triethylammonium chloride, the
first wash contained mainly iPrOH (5:1 wt/wt iPrOH−water),
whereas the second and third washes contained more water to
get rid of excess HCl (1:4 wt/wt iPrOH−water). Due to the
water in the wash, the filter cake was voluminous and the drying
time long. After Campaign 3a it was found that the third wash
could be performed with iPrOH at lower temperature (0 °C
instead of 10 °C), improving the yield and shortening the
drying time. It was also observed that the solids could be dried
at 60 °C instead of 40 °C, without affecting the purity of 10.
The largest observed impurity was the addition of isonipecotic
acid to 10 (added to the isonipecotic acid part, MW 428).
A screen of different coupling conditions was carried out for
the final coupling between pyridine acid 10 and benzylsulfo-
namide (6). Using thionyl chloride for the coupling was
abandoned since a huge excess of pyridine as base/solvent was
necessary and the subsequent coupling did not produce good
results. The coupling was also tested with EDC, CDI, and
isobutyl chloroformate. CDI14 was chosen since it gave the
highest yield and no base was needed (addition of a base such
as N-methyl morpholine or pyridine only resulted in lower
yields). A range of solvents were screened for the coupling,
such as 2-butanone, MIBK, THF, 2-MeTHF, acetonitrile,
DMSO, NMP, and DMI, and the best result were obtained with
2-butanone, NMP, and DMI2-butanone was chosen for
Campaign 3a since it had a lower boiling point and no potential
SHE issues. Both pyridine acid 10 and CDI forms slurries in 2-
butanone, but since the slurry of 10 in 2-butanone was more
homogeneous than CDI in 2-butanone, it was decided that the
slurry of 10 should be added to the slurry of CDI. Thus on a 20
kg scale, pyridine acid 10 was added to a mixture of CDI in 2-
butanone, affording the corresponding acyl imidazole. Initially
the reaction was performed at 45−50 °C, but was later lowered
to 20−25 °C, since it gave better results. For the final coupling
of the acyl imidazole of 10 with 6, the reaction rate increased
with increased temperature, but at temperatures above 110 °C
in THF/MIBK, considerable side reactions were seen and
decomposition of 1 started to occur. With 2-butanone and a
reflux temperature of 82 °C, a fast conversion was observed on
lab scale with limited byproduct formation. In the beginning of
the development work, an excess of 1.3 equiv of benzylsulfo-
namide (6) was used with a reaction time around 1 h. Lowering
the excess of 6 to 1.2 equiv gave similar results, but when 1.1
equiv was used a significantly longer reaction time was needed
(22 h). With 1.05 equiv of 6 the reaction did not progress
beyond 70% conversion. Both EtOH and iPrOH were tested as
antisolvents, but with EtOH the product oiled out in several
experiments. Changing to iPrOH improved the process, no
oiling out was observed, and the solubility properties improved.
For the development work, a ratio of 65:35 2-butanone−iPrOH
(v/v) was chosen and the reaction mixture was clear filtered at
80 °C. Upon cooling, AZD1283 (1) precipitates from this
solvent mixture. Generated imidazole also precipitates, either
by cocrystallization or incrustation in the formed crystals, but
by addition of HCl/iPrOH the formed imidazole hydrochloride
remains in the mother liquid. Only a minor excess of HCl can
be added since AZD1283 (1) is not stable below pH 1. It was
also noted that if the HCl/iPrOH mixture was added after the
mixture had been seeded, the crystallization was interrupted
and a clear solution obtained. From that point on, the
crystallization was uncontrollable. A possible reason for this
could be the higher solubility of 1 in slightly basic media. By
adding part of the HCl/iPrOH mixture prior to seeding and
reducing the crystallization temperature from 65 to 60 °C, the
kinetics for the nucleation was favored over those of
dissolution. After aging for 1 h, the remaining HCl/iPrOH
was added and the crystallization temperature lowered to 0 °C.
This protocol gave robust crystallization, affording AZD1283
(1) (polymorph II)6 with a purity above 99.5%, even in the
presence of a large excess of benzylsulfonamide (6) (1.45
equiv) and up to 20% of hydrolyzed 1 (decomposition product
of 1). The only impurity of concern from the reaction,
imidazole, could be removed by aqueous wash. Taking all these
details into account, we converted pyridine acid 10 into
AZD1283 (1) on a 20 kg scale in 83% yield.
CONCLUSION
■
Large scale synthesis routes to AZD1283 (1) have been
developed and material manufactured to support preclinical and
clinical studies in scale-up Campaigns 1, 2, and 3a. In Campaign
1, 0.5 kg non-GMP AZD1283 (1) was produced in six steps
starting from ethyl acetoacetate and DMF dimethyl acetal. The
1548
dx.doi.org/10.1021/op400288v | Org. Process Res. Dev. 2013, 17, 1543−1551

Organic Process Research & Development
Article
overall yield was 50% starting from 2-chloropyridine 4, Boc-
isonipecotic acid 5, and benzylsulfonamide (6). The Campaign
1 material appeared as an orange solid and the synthesis was in
need of better workup procedures. One synthetic step was not
scaleable (Boc-deprotection performed in formic acid). In
Campaign 2, using the same synthetic sequence, improved
workup procedures were developed as well as a scaleable Boc-
deprotection process using HCl/iPrOH. Campaign 2 generat-
ing 8 kg of colorless GMP material for phase I clinical studies in
58% overall yield starting from 4, 5, and 6. Prior to the next
round of scale-up, Campaign 3a, a thorough prestudy was
performed to identify the most suitable route based on
robustness, manufacturability, and low cost of API. Five routes
were selected and investigated and the pyridine acid route was
selected (Figure 2). In the pyridine acid route, isonipecotic acid
is coupled with 2-chloropyridine 4 to give pyridine acid 10,
which is coupled with benzylsulfonamide (6) to afford
AZD1283 (1). The route was developed and scaled up to
yield 20 kg batches of AZD1283 (1) in a 64% overall yield
starting from 4, 6, and isonipecotic acid. Campaign 1 afforded
polymorph I and Campaigns 2 and 3 yielded polymorph II.6
The synthesis of benzylsulfonamide (6) and 2-chloropyridine 4
were outsourced and optimized procedures are included for 6.
Optimized procedures for 4 have recently been published.4
During Campaign 3a it was found that pyridone 3 and 2-
chloropyridine 4 were not genotoxic after all and we also found
a thermodynamically more stable polymorph of AZD1283 (1),
being formed after 4−5 days at 0 °C in 2-butanone (more
stable at 0 to 65 °C, polymorph II more stable at 65 to 80 °C).
If seeded the transformation of polymorph II into the more
stable polymorph takes less than 24 h at 0 °C in 2-butanone. A
total of six polymorphs were observed.
off) and isopropyl acetate (23 L) was added. The solvents were
distilled off under reduced pressure at 40 °C (23 L distilled off,
all acetonitrile removed). Isopropyl acetate (22 L) was added
and then 1 M HCl (45 L). The organic phase was washed with
1 M HCl (3 × 45 L) and concentrated under reduced pressure
at 40 °C to a crude residue. EtOH (15 L) added and the
mixture concentrated under reduced pressure at 40 °C (13 L
distilled off). EtOH (26 L) was added and at 40 °C, water (25
L) was added until crystallization was induced. The product
mixture was cooled to 5 °C over 4 h, and after 12 h, the solids
were filtered off. The mother liquid was concentrated under
reduced pressure at 40 °C (12 L distilled off) and water (20 L)
added until crystallization was induced. Cooled to 5 °C and
after 2 h, the solids were filtered off. The combined solids were
washed with cold EtOH (0 °C, 7 L) and dried under reduced
1
pressure at 40 °C to 6.0 kg (81% yield). H NMR (400 MHz,
CDCl3) δ 1.43 (s, 9H), 1.43 − 1.76 (m, 4H), 2.26 (tt, J = 3.8,
11.3, 1H), 2.67 (s, 2H), 4.04 (d, J = 13.3, 2H), 4.6 (s, 2H), 7.21
− 7.45 (m, 5H), 8.69 (s, 1H). 13C NMR (151 MHz, CD3OD)
δ 28.7 (3C), 29.0 (2C), 43.8, 44.0 (2C), 59.1, 81.2, 129.8 (2C),
130.0, 130.4, 131.9 (2C), 156.3, 176.3. HRMS (ESI): [M]+ m/
z Calcd for C18H26N2O5S 382.1562; Found 382.1563.
4-[(Benzylsulfonyl)carbamoyl]piperidinium hydrochloride
(8). A reactor was charged with tert-butyl 4-[(benzylsulfonyl)-
carbamoyl]piperidin-1-carboxylate (7) (6.0 kg, 15.7 mol),
iPrOH (30 L) and 2-MeTHF (8.5 L). The temperature was
increased to 60 °C and 5 M HCl in iPrOH (8 L) was added
during 2 h. The addition vessel was washed with 2-MeTHF (2
L). After 2 h the reaction mixture was cooled to 10 °C over 5 h.
The product crystallized from the cold reaction mixture and
was filtered off and washed with 2-MeTHF (10 L). The
product was dried at 40 °C under reduced pressure to 4.78 kg
(96% yield). 1H NMR (400 MHz, DMSO-d6) δ 1.66−1.92 (m,
4H), 2.79 (d, J = 9.9, 2H), 3.22 (d, J = 12.6, 2H), 3.35 (s, 1H),
4.68 (s, 2H), 7.21−7.33 (m, 2H), 7.33−7.44 (m, 3H), 9.05 (s,
1H), 9.23 (s, 1H), 11.72 (s, 1H). 13C NMR (101 MHz,
DMSO-d6) δ 24.3 (2C), 39.1, 41.9 (2C), 57.5, 128.6 (2C),
128.7, 129.0, 130.7 (2C), 173.6. HRMS (ESI): [M]+ m/z
Calcd for C13H18N2O3S 282.1038; Found 282.1038.
EXPERIMENTAL SECTION
■
Commercially available solvents and reagents were used
without purification. All reaction and operations were
conducted under an inert nitrogen atmosphere and, unless
otherwise noted, temperature describes the mantle temperature
of the reactor. Thin-layer chromatography was made on silica
gel 60 F254 (Merck KGaA, Darmstadt). LC analyses were
performed using a Waters 600 instrument, C8 kromasil 100
column (50 × 4.6 mm, df = 5 μm), mobile phase 0.1 M
NH4OAc buffer/acetonitrile. HRMS analyses were performed
using a Micromass LCT mass spectrometer. 1H NMR
measurements were performed on a Varian Mercury VX 400
AZD1283: Ethyl 6-{4-[(benzylsulfonyl)carbamoyl]-
piperidin-1-yl}-5-cyano-2-methylpyridine-3-carboxylate (1).
A reactor was charged with ethyl 6-chloro-5-cyano-2-methyl-
nicotinate (4) (3.0 kg, 86 wt %, 11.5 mol), 4-[(benzylsulfonyl)-
carbamoyl]piperidinium chloride (8) (4.6 kg, 96 wt %, 1.2
equiv) and EtOH (30 L). Triethylamine (5.2 kg, 4.5 equiv) was
added and the reaction mixture was refluxed for 2 h. The
mixture was cooled to 50 °C and a solution of ortho-phosphoric
acid (4 kg in 15 L of water) was added to the reactor. The
resulting slurry was stirred for one hour and then cooled to 0
°C over 5 h. The crystalline product was filtered off and washed
with cold EtOH (0 °C, 20 L). The product was dried under
reduced pressure at 40 °C to 5.1 kg crude product (94%). The
crude product was treated with activated charcoal (350 g) in
THF (25 L) at reflux for 5 h. The filtrate was concentrated
under reduced pressure at 60 °C to a volume of 12.5 L. Cooled
to 10 °C and upon addition of EtOH (25 L) the product
precipitated. After 3 h, the solids were filtered off and washed
twice with cold EtOH (10 °C, 7.5 and 10 L). The product was
dried under reduced pressure at 40 °C to 4.1 kg of polymorph
1
spectrometer, operating at a H frequency of 400 MHz or on
Varian UNITY plus 400, 500, and 600 spectrometers, operating
1
at H frequencies of 400, 500, and 600 MHz, respectively.
Chemical shifts are given in ppm with the solvent as internal
standard. Coupling constants are given in Hz. Purity is
measured by %area in LC and %wt by Quantitative NMR.
When %wt is noted the yields have been corrected.
Campaign 2. tert-Butyl 4-[(benzylsulfonyl)carbamoyl]-
piperidine-1-carboxylate (7). A reactor was charged with 1-
(tert-butoxycarbonyl)piperidine-4-carboxylic acid (5) (4.5 kg,
98.6 wt %, 19.4 mol), 1-benzylsulfonamide (6) (3.44 kg, 1.05
equiv), LiCl (0.21 kg, 0.25 equiv), TBTU (6.56 kg, 1.05 equiv),
and then AcCN (27 L). The reaction mixture was cooled to 15
°C and triethylamine (8.78 kg, 4.5 equiv) was added. For the
coupling, the temperature was increased to 25 °C. In order to
reach full conversion more TBTU (1.25 kg, 0.2 equiv) was
added after 2.5 h. After 17 h, the reaction mixture was
concentrated under reduced pressure at 40 °C (25 L distilled
6
1
II (98 wt %, 74% yield). H NMR (400 MHz, CDCl3) δ 1.25
(t, 3H, J = 7.1), 1.59 − 1.8 (m, 4H), 2.26 − 2.37 (m, 1H), 2.60
(s, 3H,), 2.94 − 3.06 (m, 3H), 4.19 (q, 2H, J = 7.1), 4.48 −
4.57 (m, 4H), 7.11 − 7.31 (m, 5H), 8.03 (s, 1H), 8.22 (s, 1H).
1549
dx.doi.org/10.1021/op400288v | Org. Process Res. Dev. 2013, 17, 1543−1551

Organic Process Research & Development
Article
13C NMR (101 MHz, CDCl3) δ 14.3, 25.7, 27.7 (2C), 42.8,
46.4 (2C), 58.8, 61.0, 89.2, 115.0, 117.9, 127.8, 129.0 (2C),
129.5, 130.6 (2C), 147.7, 158.3, 164.4, 164.6, 173.0 . HRMS
(ESI): [M]+ m/z Calcd for C23H26N4O5S 470.1624; Found
470.1598 .
(33 kg) at 20 °C over 30 min (evolution of carbon dioxide).
The reactor was rinsed with 2-butanone (16 L). After 1 h, 90%
conversion. More 1,1-carbonyldiimidazole (1.72 kg, 0.17 equiv)
in 2-butanone (4 kg) was added. After a total of 7 h 30 min
more than 98% conversion. A solution of benzylsulfonamide
(6) (13.2 kg, 1.21 equiv) in 2-butanone (33 kg) was added over
10 min and the addition vessel was rinsed with 2-butanone (8
kg). The temperature was increased to 90 °C and 2-butanone
was distilled off until 8 relative volumes were left in the reactor
(170 L). Thereafter the reactor content was refluxed for 17 h
(full conversion). More 2-butanone (35 kg) was added to the
hot reaction mixture, keeping the internal temperature above
80 °C, and then iPrOH (64 kg) was added, keeping the internal
temperature above 70 °C. At 80 °C, the reaction mixture was
clear filtered and the filter was washed with solution of 2-
butanone (12 kg) and iPrOH (7 kg). The filtrate was cooled to
60 °C and a solution of 37% hydrochloric acid (9.8 kg) in
iPrOH (18.1 kg) was added over 40 min, keeping the internal
temperature at 60 °C. The mixture was seeded with AZD1283
(1) (0.24 kg of polymorph II) and after 1 h, a solution of 37%
hydrochloric acid (16.0 kg) in iPrOH (29.6 kg) was added over
2 h, keeping the temperature at 60 °C. The reaction mixture
was cooled to 0 °C over 10 h and stirred overnight (7 h). The
product slurry was filtered and the product was washed with a
cold 1:1 (w/w) mixture of 2-butanone and iPrOH (0 °C, 32
kg) and then with cold water (2 °C, 3 × 40 kg). The product
was dried under reduced pressure at 40 °C to 24.7 kg of
polymorph II6 (excl. seed, 83% yield).
Campaign 3. Benzylsulfonamide (6). A reactor was
charged with 25% aqueous ammonia (209 kg, 3068 mol) and
cooled to 1 °C. A solution of benzylsulfonyl chloride 9 (139 kg,
729 mol) in THF (312 kg) was slowly added over 1 h (max
reaction temperature 9 °C) and the addition vessel was washed
with THF (21 kg). The temperature was increased to 22 °C
and stirred for 8 h (less than 0.2% starting material left). The
lower aqueous phase was removed and MTBE (217 kg) added.
A small amount of additional aqueous phase was removed (5 L)
and the organic phase was washed twice with brine (116 and 83
kg) and then with water (84 L). The organic phase was
concentrated under reduced pressure at 45 °C (400 L
removed) and then cooled to 21 °C. The solution was filtered
and the temperature increased to 43 °C. Toluene (417 kg) was
added maintaining a temperature of 40−45 °C. The solution
was concentrated under reduced pressure at 80 °C (250 L
removed). The product precipitated immediately on cooling
instead of slowly crystallizing and therefore more toluene (40
L) was added and the mixture stirred at 65 °C overnight. The
solution was slowly cooled to 20 °C over 5 h and stirred at this
temperature for additional 3 h. The product was filtered off,
washed with toluene (249 kg), and dried under reduced
pressure at 38 °C to 91.8 kg (73.2%, UV purity 99.95% at 210
nm).
Analytical data as described under Campaign 2.
1H NMR (400 MHz, CDCl3) δ 4.32 (s, 2H), 4.66 (s, 2H)
1
ASSOCIATED CONTENT
* Supporting Information
Experimental methods. This material is available free of charge
via the Internet at http://pubs.acs.org.
and 7.36−7.47 (5H, m). C NMR (150.9 MHz, DMSO- d6) δ
■
60.2, 127.8, 128.2 (2C), 130.7 (2C), 130.8. HRMS (ESI): [M]−
m/z Calcd for C7H8NO2S 171.0354; Found 171.0337.
S
1-[3-Cyano-5-(ethoxycarbonyl)-6-methylpyridin-2-yl]-
piperidine-4-carboxylic acid (10). A reactor was charged ethyl
6-chloro-5-cyano-2-methylnicotinate (4) (17.6 kg, 78.1 mol),
isonipecotic acid (10.4 kg, 80.5 mol) and iPrOH (83 kg). The
temperature was increased to 80 °C and triethylamine (15.8 kg,
2.0 equiv) was added over 1 h. The reaction mixture was heated
to reflux for 4.5 h to reach completion and then cooled to 45
°C. 12% Hydrochloric acid (14.4 kg) was added over 30 min.
The mixture was seeded with 10 (124 g) and after 1 h, more
12% hydrochloric acid (11.3 kg) was added over 2.5 h. After
additional 45 min the mixture was cooled to 8 °C over 5.5 h.
After 4 h the product was filtered off and washed once with a
cold 5:1 mixture (w/w) of iPrOH and water (10 °C, 43 kg) and
then twice with a cold 1:4 (w/w) mixture of iPrOH and water
(10 °C, 2 × 50 kg). The product was dried at 40 °C under
reduced pressure to 21.4 kg (excl. seed, 94.1 wt %, 81% yield).
1H NMR (400 MHz, CDCl3) δ 1.38 (t, 3H, J = 7.1), 1.8 − 1.96
(m, 2H), 2.02 − 2.16 (m, 2H), 2.62 − 2.78 (m, 4H), 3.25 −
3.38 (m, 2H), 4.32 (q, 2H, J = 7.1), 4.53 − 4.66 (m, 2H), 8.35
(s, 1H). 13C NMR (101 MHz, CDCl3) δ 14.3, 25.7, 27.8 (2C),
40.6, 46.7 (2C), 60.9, 89.1, 114.8, 118.0, 147.7, 158.4, 164.4,
164.7, 180.0. HRMS (ESI): [M]+ m/z Calcd for C16H19N3O4
317.1376; Found 317.1375.
AUTHOR INFORMATION
Corresponding Authors
*E-mail: soren.andersen@astrazeneca.com.
*E-mail: carl-johan.aurell@astrazeneca.com.
Notes
■
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We would like to thank Dr. Thomas Antonsson, Astrazeneca
R&D for constructive criticism on the present paper. Jim
Brennan, Veronica Profir, Emma Carlsson, and Stefan Berlin at
Astrazeneca Process R&D are also acknowledged for their
contributions to this work. The authors also thank SAFC for
providing compound 6 on large scale.
REFERENCES
■
(1) Bach, P.; Antonsson, T.; Bylund, R.; Bjoerkman, J.; Oesterlund,
K.; Giordanetto, F.; van Giezen, J. J. J.; Andersen, S. M.; Zachrisson,
H.; Zetterberg, F. J. Med. Chem. 2013, 17, 7015−7024.
(2) Beck, J. R.; Lynch, M. P. J. Heterocycl. Chem. 1987, 3, 693−695.
(3) Mosti, L.; Menozzi, G.; Schenone, P.; Dorigo, P.; Gaion, R. M.;
Benetollo, F.; Bombieri, G. Eur. J. Med. Chem. 1989, 5, 517−529.
(4) Bell, S. J.; McIntyre, S.; Garcia, C. F.; Kitson, S. L.; Therkelsen,
F.; Andersen, S. M.; Zetterberg, F.; Aurell, C.-J.; Bollmark, M.; Ehrl, R.
Org. Process Res. Dev. 2012, 5, 819−823.
(5) Bach, P.; Bostroem, J.; Brickmann, K.; van Giezen, J. J. J.;
Groneberg, R. D.; Harvey, D. M.; O’Sullivan, M.; Zetterberg, F. Eur. J.
Med. Chem. 2013, 360−375.
AZD1283: Ethyl 6-{4-[(Benzylsulfonyl)carbamoyl]-
piperidin-1-yl}-5-cyano-2-methylpyridine-3-carboxylate (1).
A reactor at 25 °C was charged with 1-[3-cyano-5-
(ethoxycarbonyl)-6-methylpyridin-2-yl]piperidine-4-carboxylic
acid (10) (21.4 kg, 94.1 wt %, 63.5 mol) and 2-butanone (114
kg). After 45 min the solution of 10 was added to a mixture of
1,1-carbonyldiimidazol (12.4 kg, 1.20 equiv) and 2-butanone
1550
dx.doi.org/10.1021/op400288v | Org. Process Res. Dev. 2013, 17, 1543−1551

Organic Process Research & Development
Article
(6) Andersen, S.; Aurell, C.-J.; Eriksson, C.; Sigfridsson, C.;
Zetterberg, F. WO Patent WO2008004945.
(7) www.almacgroup.com
(8) www.safcglobal.com
(9) Davis, F. A.; Zhou, P.; Carroll, P. J. J. Org. Chem. 1993, 18, 4890−
4896.
(10) Sheehan, J. C.; Preston, J.; Cruickshank, P. A. J. Am. Chem. Soc.
1965, 11, 2492−2493.
(11) Ellis, D. Tetrahedron: Asymmetry 2001, 11, 1589−1593.
(12) McElvain, S. M.; Schroeder, J. P. J. Am. Chem. Soc. 1949, 40−46.
(13) Cocco, M. T.; Congiu, C.; Onnis, V.; Maccioni, A. Synthesis
1991, 7, 529−530.
(14) Anderson, G. W.; Paul, R. J. Am. Chem. Soc. 1958, 4423.
NOTE ADDED AFTER ASAP PUBLICATION
■
This paper was published ASAP on November 22, 2013. The
text “polymorph II” was corrected to “polymorph I” in the first
section of the Results and Discussion. The corrected version
was reposted on December 3, 2013.
1551
dx.doi.org/10.1021/op400288v | Org. Process Res. Dev. 2013, 17, 1543−1551