Development of an efficient and practical route for the multikilogram manufacture of ethyl 5-cyano-2-methyl-6-oxo-1,6-dihydropyridine-3-carboxylate and ethyl 6-chloro-5-cyano-2-methylnicotinate, key intermediates in the preparation of P2Y12 antagonists

Article abstract of DOI:10.1021/op200368m

Elucidation of the mechanism of formation of two major impurities in the synthetic route towards key intermediate ethyl 5-cyano-2-methyl-6-oxo-1,6- dihydropyridine-3-carboxylate 1, led directly to the development of a route with significant process improvements in terms of yield, purity, and operability. The overall process yield increased from 15% to 73% without the need for extra purification steps, giving the key intermediate ethyl 6-chloro-5-cyano-2- methylnicotinate, 2, in excess of 80 kg to support clinical development.

Full text of DOI:10.1021/op200368m

Communication
pubs.acs.org/OPRD
Development of an Efficient and Practical Route for the
Multikilogram Manufacture of Ethyl 5-Cyano-2-methyl-6-oxo-1,6-
dihydropyridine-3-carboxylate and Ethyl 6-Chloro-5-cyano-2-
methylnicotinate, Key Intermediates in the Preparation of P2Y12
Antagonists
,
†
†
†
†
†
Stephen J. Bell,* Steve McIntyre, Conchita F. Garcia, Sean L Kitson, Frans Therkelsen,
‡
‡
§
§
§
Søren M Andersen, Fredrik Zetterberg, Carl-Johan Aurell, Martin Bollmark, and Robert Ehrl
†
Almac, Almac House, 20 Seagoe Industrial Estate, Craigavon, Northern Ireland BT63 5QD, United Kingdom
‡
Astrazeneca R&D, Mo
̈
lndal, Sweden
§
Pharmaceutical Development, AstraZeneca R&D, So
̈
̈
dertalje, Sweden
using a similar chemistry approach to the above route. Thus, a
better understanding of the reaction and the generation of the
process impurities was believed to be of fundamental
importance. The results of this investigation are discussed in
this report.
ABSTRACT: Elucidation of the mechanism of formation
of two major impurities in the synthetic route towards key
intermediate ethyl 5-cyano-2-methyl-6-oxo-1,6-dihydro-
pyridine-3-carboxylate 1, led directly to the development
of a route with significant process improvements in terms
of yield, purity, and operability. The overall process yield
increased from 15% to 73% without the need for extra
purification steps, giving the key intermediate ethyl 6-
chloro-5-cyano-2-methylnicotinate, 2, in excess of 80 kg to
support clinical development.
RESULTS AND DISCUSSION
■
Two main impurities were observed in the preparation of 1,
and these are shown in Figure 2. We started our work by trying
to understand how these impurities were formed.
Impurity 7 was initially believed to be the acid from ester
hydrolysis of the ester product, due to residual water. However,
the revised assignment of 7 to 8 as the methyl ketone,
1
3
INTRODUCTION
supported by the carbonyl peak at 192 ppm in the C NMR
spectrum, made this impurity the product of an alternate
cyclisation (Scheme 2). The condensation of the amide
nitrogen with the carbonyl of the ester in intermediate 10
accounts for the formation of ketone 8. Thus, the required
mode of condensation to afford 1 is that between the amide
nitrogen and the carbonyl of the ketone in 10.
■
Ethyl 5-cyano-2-methyl-6-oxo-1,6-dihydropyridine-3-carboxy-
late, 1 (cyanopyridone), and ethyl 6-chloro-5-cyano-2-methyl-
nicotinate, 2 (chloropyridine), were key intermediates to an
1
AstraZeneca development candidate (Figure 1). The initial
kilogram-scale route to these compounds was problematic, and
the development of a more robust and scaleable route to
multikilogram quantities was needed.
This kilogram-scale route (Scheme 1), a three-step synthesis,
involved a condensation of ethyl acetoacetate 3, and N,N-
dimethylformamide dimethylacetal (DMFDMA) 4, to give the
enaminone 5, followed by a condensation with cyanoacetamide
Impurity 9, the formal product of nitrile hydrolysis, was
initially believed also to be due to water. However, the higher
concentration of impurity 9 formed at both lower temperatures
or when the weaker base K CO was used, was inconsistent
2
3
with a simple nitrile hydrolysis of 1. Rationally, a different
mechanism must be operating to afford impurity 9. A proposed
6
to afford 1, and a final chlorination to produce 2. The process
3
mechanism for the competing pathway to form impurity 9 is
involved lengthy workups and low yields for both steps 2 and 3,
mainly due to the instability of enaminone 5 in step 1.
Furthermore, formation of a thick slurry during step 2 reaction
made agitation difficult, causing the isolation of impurities with
the step 2 product. The colour of step 3 material (caused by
impurities) and the compatibility of the thionyl chloride/DMF
reagent-pair in step 3 also needed attention.
outlined in Scheme 3. Under mildly basic conditions the
enolate 10b is formed, use of a stronger base enables
deprotonation of 10b to give the dianion 10a which gave the
desired product 1. The intramolecular condensation between
the enolate and the cyano group of 10b results in the pyran
intermediate 11. Ring-opening of the pyran intermediate 11
can be easily accomplished by a nucleophile. Dimethylamine,
still present in the reaction mixture, could attack the pyran ring
to undergo ring-opening and generate the diamide intermediate
The major concerns were safety, poor yield and difficult
isolations experienced with the original route. The work plan
was to carry out a comprehensive literature review to
investigate any potential alternative routes to 1 and 2, followed
by testing in the laboratory of the most promising alternatives.
Finally, the preferred method of choice was to be developed to
12. The final ring closure is then achieved by an intramolecular
Michael addition-type reaction, which affords impurity 9.
2
give a process suitable for 400−600 L scale. Literature showed
Received: December 12, 2011
Published: April 20, 2012
that the vast majority of pyridinones such as 1 are synthesized
©
2012 American Chemical Society
819
dx.doi.org/10.1021/op200368m | Org. Process Res. Dev. 2012, 16, 819−823

Organic Process Research & Development
Communication
Figure 1. Key intermediates.
Scheme 1. Original route
with a substantially reduced level of impurity 8 still present due
to reaction between deprotonated amide and the ester group in
intermediate 16. The information obtained from the increased
understanding of the reaction process led directly to the
development of the new conditions for target 1.
The development work examined the use of a weaker base
and a change in the order of addition so that the malononitrile
was added slowly over time due to concerns of self-
condensation when strong bases in combination with
malononitrile are used. This work showed that triethylamine
could be used in place of stronger alkoxide bases, and since
dimethylamine is liberated as the reaction proceeds, only a
catalytic quantity was required. It was observed that the use of
the catalytic amounts of weaker base, triethylamine, gave an
easily stirred and almost homogeneous mixture up to the
quench with acetic acid. Ethanol was the solvent of choice for
steps 1 and 2, allowing the process to be telescoped with the
product of step 2 isolated by crystallisation simply on addition
of water.
Figure 2. Impurities of original route.
With step 2 product in hand, the chlorination of 1 to afford 2
was found to proceed cleanly in both neat POCl and POCl /
3
3
Assuming that the formation of impurity 9 follows the
pathway in Scheme 3, then the replacement of cyanoacetamide
acetonitrile, avoiding any concerns with the potential genotoxic
dimethylcarbamoyl chloride (DMCC). The optimised con-
4
6
1
(Scheme 1) by malononitrile 13 would still afford 1 via 14,
5, and 16 (Scheme 4). This pathway was demonstrated by an
ditions were 1.6 equiv of POCl in 3 vol of acetonitrile; an
3
easily stirred mixture was obtained, and when 1.6 equiv of
POCl3 was used, a clear solution was obtained after
approximately 2 h. Full conversion (<1% area by HPLC of
1) was observed after 20 h. The reaction did not go to
experiment using malononitrile in place of cyanoacetamide
under the initial reaction conditions.
Not only did this support the proposed mechanism of
formation of impurity 9 but also gratifyingly formed product 1
completion when 1.2 equiv of POCl was used.
3
8
20
dx.doi.org/10.1021/op200368m | Org. Process Res. Dev. 2012, 16, 819−823

Organic Process Research & Development
Communication
Scheme 2. Mechanism for the formation of impurity 8
Scheme 3. Mechanism for the formation of impurity 9
The optimised process included the addition of 6 vol of MTBE
5
at 0−15 °C and slow addition of 6 vol of water. The water
phase was back-extracted with MTBE and the combined
organic phases were subsequently washed with 4 vol of water, 2
×
5% K CO solution and water. Crystallisation was found to
2 3
be excellent from either methanol, ethanol or isopropanol in
:1 ratio with water. The recovery of 2 was found to be
1
approximately 90% in all three experiments. It was found that a
solvent exchange to ethanol proceeded very efficiently.
Concentrating the MTBE to approximately 2−3 vol and
charging 6 vol of ethanol and concentrating to 5 vol resulted in
1
<
1% (by H NMR) of MTBE and acetonitrile in the solution.
The precipitation of 2 was completed by addition of 5 vol of
water at 5−10 °C. The filtration was also found to proceed very
efficiently. Under optimised conditions (Scheme 5), 2 was
afforded in 89% in 500 g scale.
Scheme 4. New process to compound 1
Scheme 5. Developed process
The first scale-up campaign (1 kg) was performed in a
combination of 30 and 50 L vessels to test the chemistry at
scale. The formation of the pyridinone behaved as expected
with 1.25 kg (79%) of the pyridone obtained after
crystallisation from the reaction mixture via the addition of
water. However, the step 3 product from the chlorination
1
reaction, while pure by H NMR spectroscopy, was much
darker than usual (brown vs pale yellow). Following a brief
investigation, this was found to be due to charring of the
material on the jacket of the reactor. The temperature was thus
lowered in subsequent batches from acetonitrile at reflux to
70−75 °C. Although the reaction took longer to reach
completion (typically 20 h), the reaction profile remained
unchanged, and more importantly the colour of the isolated
chloropyridine was more acceptable. Thus, for the manufactur-
ing plant campaign, the jacket temperature was controlled (set
to 80−85 °C and contents of reactor monitored) during the
The workup was initially performed by concentrating to
dryness. Addition of ice and water was followed by
neutralisation with K CO or NaOH. Extraction into either
MTBE or DCM resulted in the isolation of crude brown
material of 2 in 90−95% yield. However, compound 2, even
under strongly acidic conditions (pH <1), is easily extracted
into an organic phase such as MTBE or ethyl acetate.
Consequently, the neutralisation with a vast amount of base
could be avoided and 2 easily obtained in a MTBE solution.
2
3
8
21
dx.doi.org/10.1021/op200368m | Org. Process Res. Dev. 2012, 16, 819−823

Organic Process Research & Development
Communication
formation of the chloropyridine. A check of the colour of the
product solution at the end of the aqueous workup was also
introduced prior to the solvent exchange into ethanol as a final
check prior to isolation of the chloropyridine where a charcoal
treatment on the solution could be performed in the event of
problems with colour. A total of 80.5 kg of chloropyridine 2
was manufactured in three independent batches with consistent
The product was filtered, washed three times with water (3 ×
1.25 L), and dried at 50 °C under vacuum to afford 539 g (82%
yield) of the desired product 1 as an orange solid pure by
HPLC at 99.5% and 100.8% w/w NMR assay (maleic acid
1
standard); H NMR (500 MHz, DMSO-d ) δ 13.0 (s, 1H),
6
8.44 (s, 1H), 4.21−4.25 (q, 2H, J = 5 Hz), 2.61 (s, 3H), 1.28−
1.31 (t, 3H, J = 5 Hz).
Ethyl 6-Chloro-5-cyano-2-methylnicotinate (2). Cya-
nopyridinone 1 (500 g, 1 equiv), acetonitrile (1.5 L, 3 vol), and
6
yields and purities obtained.
POCl (361 mL, 1.6 equiv) were charged in 20 L round-bottom
CONCLUSION
3
■
flask with an overhead mechanical stirrer and condenser, heated
to 80−85 °C (external temperature), and stirred for 20 h under
nitrogen until less than 1% area of cyanopyridinone 1 was
observed by HPLC (260 nm). The reaction mixture was cooled
to 0−5 °C, and MTBE (3 L, 6 vol) was charged; then purified
water (3 L, 6 vol) was charged over 1.5 h at a temperature
between 0 and 15 °C. CAUTION: Exothermic quench of POCl3.
Layers were stirred for 30 min at 10−20 °C and settled for 10
min. The layers were separated, the remaining aqueous layer
was recharged, and MTBE (1 L) was added to this layer. It was
stirred for 5 min and settled for 5 min at 20 °C. The layers were
separated, and the remaining organic layer was combined with
the previous one. Water (2 L) was charged to the combined
organic layer, stirred for 5 min, and settled for 5 min at 20 °C.
The layers were separated. A solution of potassium carbonate
The new process to the key intermediate cyanopyridone 1 has
led to significant process improvements in terms of yield,
purity, and operability. The overall yield of chloropyridine 2
from ethyl acetoacetate increased from 15% to 73% without the
need for extra purifications post-isolation. The successful
characterisation of the key impurities in the original route
and the drive to understand the mechanism of their formation
led directly to the development of the new route. These
processes scaled well into our manufacturing plant, enabling us
to produce >80 kg of chloropyridine 2 to the required purity
specification. This process was successfully applied to the
1c
manufacture of P2Y12 inhibitors development candidate.
EXPERIMENTAL SECTION
■
5
% w/w in water (2 L, 4 vol) was charged, stirred for 5 min,
All materials were purchased from commercial suppliers and
were used as supplied by manufactures. NMR spectra of
intermediates were recorded on a Bruker 500 MHz. HPLC
analysis results are described as area %. The analysis of
intermediates and reaction monitoring was performed by GC
for step 1 and HPLC for steps 2 and 3. GC analysis was
performed in an Agilent 6890 series or equivalent with FID
equipped with a Zebron ZB-5 ms, 30 m × 0.32 mm ID, 0.5 μm
FT. The method started with an initial temperature of 100 °C
and temperature ramp of 10 °C/min to 300 °C, run time 35
min, carrier gas He with a flow 2.0 mL/min. The FID detector
parameters were temperature 300 °C, hydrogen flow 45 mL/
min., air 450 mL/min, nitrogen 45 mL/min (make up gas) with
an injection volume of 1 μL. HPLC analysis was carried out on
a system with a binary/quaternary pump with a variable
wavelength or photodiode array detector equipped with a C-18
Onyx Monolithic 100 mm × 4.6 mm column. It used a gradient
method of 100% mobile phase A (20 mM sodium phosphate
pH 6.0) to 50:50 mobile phase A/mobile phase B (70:30
acetonitrile: mobile phase A), flow rate 2 mL/min injection
volume 10 μL.
and settled for 5 min at 20 °C; the layers were separated. The
previous wash was repeated a second time. Purified water (2 L)
was charged, stirred for 5 min, settled for 5 min and the layers
were separated. The remaining organic layer was concentrated
under vacuum at 40 °C to 2−3 volumes of solvent (∼1 L of
solvent, ∼2 mL of solvent/g of pyridinone). Ethanol (3 L) was
added and distilled under vacuum at 40 °C up to 5 volumes
(
∼2.5 L). The previous operation was repeated until residual
1
MTBE and acetonitrile was below 1 mol % by H NMR relative
to ethanol. The mixture was stirred and cooled down to 5−10
°
C. Purified water was charged (2.5 L) over 20 min and stirred
for 30 min at 5−10 °C. The product was filtered and washed
twice with purified water (2 × 1 L) to give 484 g (yield 89%) of
the desired product 2 as a yellow solid pure by HPLC at 99.7%
1
and 100.5% w/w NMR assay (maleic acid standard); H NMR
(
500 MHz, CDCl ) δ 8.49 (s, 1H), 4.40−4.44 (q, 2H, J = 5
3
Hz), 2.90 (s, 3H), 1.41−1.44 (t, 3H, J = 5 Hz).
AUTHOR INFORMATION
Corresponding Author
*Stephen.bell@almacgroup.com
■
Ethyl 5-cyano-2-methyl-6-oxo-1,6-dihydropyridine-3-
carboxylate (1). To a clean dry reactor equipped with
overheaded mechanical stirrer and condenser, ethyl acetoace-
tate (0.4 L, 1 equiv), ethanol (0.82 L, 2 vol) and DMFDMA
Notes
The authors declare no competing financial interest.
(
0.46 L, 1.03 equiv) were charged. The mixture was warmed to
ACKNOWLEDGMENTS
Almac greatly appreciate the supportive environment encour-
aged by AstraZeneca during this collaboration.
■
an internal temperature of 41−43 °C for 5 to 6 h until ≤2% of
ethyl acetoacetate remained by GC. The reaction mixture was
cooled to 20−25 °C and triethylamine (44.4 mL, 0.1 equiv)
was added followed by slow addition of a solution of
malononitrile (235 g in 1.77 L of ethanol, 1.1 equiv), cautiously
maintaining the temperature between 25 and 36 °C. The
mixture was stirred for 16 h at 20 °C until less than 5% area of
enaminone 5 remained by HPLC (315 nm). Acetic acid (217
mL, 1.2 equiv) was added slowly at 20−25 °C to pH 4−5, and
a dense precipitate was formed. The mixture was warmed to
REFERENCES
■
(
1) (a) Andersen, S.; Bach, P.r; Brickmann, K.; Giordanetto, F.;
Zetterberg, F.; Oesterlund, K. Pyridine analogues as P2Y12 inhibitors
and their preparation, pharmaceutical compositions and use in the
treatment of platelet aggregation disorders. PCT Int. Appl. WO/2007/
0
08140, 2007. (b) Antonsson, T.; Bach, P.; Brickmann, K.; Bylund, R.;
Giordanetto, F.; Johansson, J.; Zetterberg, F. PCT Int Appl. WO/
2008/085118, 2008. (c) Bylund, R.; Hovdal, D.; Johansson, J.; Sellen,
M.; Zetterberg, F. PCT Int Appl. WO/2010/005385, 2010.
7
0−75 °C, and water (5.6 L) was charged. The suspension was
cooled down to 10 °C over 3 h and stirred for 15 min at 0 °C.
8
22
dx.doi.org/10.1021/op200368m | Org. Process Res. Dev. 2012, 16, 819−823

Organic Process Research & Development
Communication
(
(
2) For a review see: Keller, P. A. Sci. Synth. 2004, 15, 285−387.
3) Abdelrazek, F. M.; Sharaf, M. F.; Metz, P.; Jaeger, A. J. Heterocycl.
Chem. 2010, 47, 528.
(
(
4) Robinson, D. I. Org. Process Res. Dev. 2010, 14, 946.
5) ARSST analysis of the reaction and quench mixtures was used to
establish the basis of safety for the process.
(
6) Analytical information: NMR spectra of intermediates were
recorded on a Bruker 500 MHz. HPLC analysis results are described
as area %. The analysis of intermediates and reaction monitoring was
performed by GC for step 1 and HPLC for steps 2 and 3. GC analysis
was performed in an Agilent 6890 series or equivalent with F.I.D
equipped with a Zebron ZB-5 ms, 30 m × 0.32 mm ID, 0.5 μm FT.
The method started with an initial temperature of 100 °C and
temperature ramp of 10 °C/min to 300 °C, run time 35 min, carrier
gas He with a flow 2.0 mL/min. The FID detector parameters were
temperature 300 °C, hydrogen flow 45 mL/min, air 450 mL/min,
nitrogen 45 mL/min (make up gas), with an injection volume of 1 μL.
HPLC analysis was carried out on a system with a binary/quaternary
pump with a variable wavelength or photodiode array detector
equipped with a C-18 Onyx Monolithic 100 mm × 4.6 mm column. It
used a gradient method of 100% mobile phase A (20 mM sodium
phosphate pH 6.0) to 50:50 mobile phase A/mobile phase B (70:30
acetonitrile/mobile phase A), flow rate 2 mL/min, injection volume 10
μL.
8
23
dx.doi.org/10.1021/op200368m | Org. Process Res. Dev. 2012, 16, 819−823