A morpholine-free process amenable convergent synthesis of apixaban: a potent factor Xa inhibitor

Article abstract of DOI:10.1007/s00706-017-1920-1

A convergent synthesis of the anti-coagulant drug apixaban has been efficiently demonstrated on a multi-gram scale. The synthetic route is noteworthy for its brevity and fact that it completely avoids the use of morpholine, a toxic and flammable reagent, in constructing the 5,6-dihydro-1H-pyrazolo[3,4-c]pyridin-7(4H)-one core present in apixaban. Graphical abstract: [Figure not available: see fulltext.].

Full text of DOI:10.1007/s00706-017-1920-1

Monatsh Chem
DOI 10.1007/s00706-017-1920-1
ORIGINAL PAPER
A morpholine-free process amenable convergent synthesis
of apixaban: a potent factor Xa inhibitor
1
1
1
•
Narasimha Rao Nevuluri
1
•
Rajesh Kumar Rapolu
1
•
Javed Iqbal
2
1
Bhaskar Kandagatla
•
Saikat Sen
•
Vilas H. Dahanukar
•
Srinivas Oruganti
Received: 13 December 2016 / Accepted: 10 January 2017
Ó Springer-Verlag Wien 2017
Abstract A convergent synthesis of the anti-coagulant
drug apixaban has been efficiently demonstrated on a
multi-gram scale. The synthetic route is noteworthy for its
brevity and fact that it completely avoids the use of mor-
pholine, a toxic and flammable reagent, in constructing the
Introduction
The current decade has witnessed the US FDA approval of
Ò
apixaban (1, Eliquis , Pfizer and Bristol–Myers Squibb),
an anti-clotting drug that is slated to reduce the risk of
stroke in people with abnormal heart rhythm due to non-
valvular atrial fibrillation (Fig. 1) [1, 2]. Importantly, 1
represents a new class of highly selective direct factor Xa
inhibitors, and its development and approval as an orally
administered drug highlight a significant breakthrough in
the treatment of deep vein thrombosis and pulmonary
embolism [3–5]. Indeed, with its annual revenue increasing
to $774 million in 2014 from $146 million in 2013, apix-
aban was estimated to account for roughly 25% of the
Bristol–Myers Squibb’s value in 2015 [6].
5
,6-dihydro-1H-pyrazolo[3,4-c]pyridin-7(4H)-one
core
present in apixaban.
Graphical abstract
Given the immense commercial and therapeutic poten-
tial of 1 as an anti-clotting agent, we decided to investigate
the existing synthetic routes to 1 with intent to redress
potential safety related concerns and arrive at a better
streamlined preparative sequence that is more process
amenable. We realized that the point of convergence in all
known syntheses of 1 is the acquisition of the 1,6-diary-
lated 5,6-dihydro-1H-pyrazolo[3,4-c]pyridin-7(4H)-one 2,
and that the most convenient method of constructing the
heterocyclic core, present in 2, is to condense the hydrazine
Keywords Anti-coagulant drug Á
Copper catalyzed N-arylation Á Eliminations Á
Halogenation Á Heterocycles Á Pyrazole ring formation
Electronic supplementary material The online version of this
article (doi:10.1007/s00706-017-1920-1) contains supplementary
material, which is available to authorized users.
3
with a 5,6-dihydropyridin-2(1H)-one 4, bearing a suit-
able leaving group L at the C3 position (Scheme 1). An
overview of the salient steps in all known synthetic
approaches to 1 has been provided in Schemes 2 and 3.
The original discovery chemistry route to 1, reported by
Bristol–Myers Squibb, represented an archetypical linear
synthesis of apixaban and employed the N-arylated d-
valerolactam 5a (conveniently prepared from p-iodoaniline
and 5-bromovaleryl chloride) as the starting point to
append together the key fragments of 2 (Scheme 2,
&
Srinivas Oruganti
soruganti@drils.org
1
Center for Process Research and Innovation, Dr. Reddy’s
Institute of Life Sciences, University of Hyderabad Campus,
Gachibowli, Hyderabad 500 046, Telangana, India
2
Integrated Product Development Organization, Dr. Reddy’s
Laboratories, Bachupally, Qutbullapur, Hyderabad 500 090,
Telangana, India
123

N. R. Nevuluri et al.
better yield (Scheme 2, approach D) [12]. It should be
noted at this point that all linear syntheses of 1 necessarily
involve introduction of the second d-valerolactam ring in
an advanced intermediate of the synthetic scheme, and
therefore, do present certain challenges in terms of con-
trolling the amount of impurities, arising from either
incomplete lactam ring formation (such as the potentially
genotoxic 5-chloro-N-phenyl pentanamides) or hydrolysis
of the d-valerolactam reactant, in the final apixaban.
As compared to the number of linear syntheses of
apixaban known in the literature or patent prior art,
examples of a convergent synthesis of 1 are rather limited
to the sole one reported by the innovator (Bristol–Myers
Squibb) in 2003 (Scheme 3) [11]. The route was demon-
strated on a small scale and commenced with a low
yielding (*18%) preparation of 3-morpholino-5,6-dihy-
dropyridin-2(1H)-one (12) from d-valerolactam via the
intermediacy of 10 and 11. Condensation of 12 with the
hydrazone 3 led to the acquisition of the 5,6-dihydro-1H-
pyrazolo[3,4-c]pyridin-7(4H)-one core in 14 (via 13),
which was then coupled with 5a to obtain the desired
apixaban precursor 2. We noted that the late stage Cu-
catalyzed C–N bond formation between 14 and 5a was
decidedly more efficacious in leading to 2 than the one
between 8a and d-valerolactam (see Scheme 2, approach
A), and decided therefore to include this step as the end-
game move in our efforts towards a convergent process
amenable synthesis of 1.
Fig. 1 Apixaban, a potent factor Xa inhibitor based oral anti-clotting
drug
approach A) [7, 8]. This synthesis not only required use of
a large excess of morpholine (a toxic and flammable
reagent) [9] to obtain the enamide 7a from the dichloro
precursor 6a, but also included an extremely low yielding
(
21%) late stage Cu-catalyzed C–N bond formation on the
advanced 5,6-dihydro-1H-pyrazolo[3,4-c]pyridin-7(4H)-
one 8a.
Alternate linear syntheses of 1 have focused on ame-
liorating these shortcomings by, for example, converting
the dichloro intermediate 6a to an N-arylated 3-chloro-5,6-
dihydropyridin-2(1H)-one 9, which may be employed for
condensation with 3 either directly or after conversion to
the known 3-morpholino enamide 7a (Scheme 2, approach
B) [10]. An early stage Cu-catalyzed C–N bond formation
between d-valerolactam and 7a was shown by the inno-
vator (Bristol–Myers Squibb) to proceed with a higher
yield (77%) than that between d-valerolactam and 8a
Results and discussion
(
Scheme 2, approach C) [11]. Another alternate linear
synthesis of 1 employed p-nitroaniline, instead of p-
iodoaniline, for the first d-valerolactam construction so that
the nitro group may be used as a handle at a later stage of
the synthesis to introduce the second d-valerolactam in
In particular, we were interested in improving the overall
yield of 14 by (a) increasing the overall efficiency of the two-
step conversion of d-valerolactam to the enamide 11, and
(b) eliminating the need to use morpholine by removing the
Scheme 1
1
23

A morpholine-free process amenable convergent synthesis of apixaban: a potent factor Xa…
Scheme 2
123

N. R. Nevuluri et al.
Scheme 3
Scheme 4
‘
middle-man’ 12 (and its purification by column chro-
Access to 10 also led us to examine if the enamide 12
may be obtained by a direct reaction of morpholine with
10. While this reaction could be successfully performed on
a small scale in neat refluxing morpholine and afforded the
desired product 12 in 74% yield (Scheme 5), safety con-
cerns, arising from the toxicity and flammability of
morpholine [9] and the possibility of obtaining a bis-mor-
pholino impurity [10] in the reaction led us to stick with
our initial decision to employ 11 as a precursor of 14.
Much to our delight, triethylamine mediatedcouplingof11
and 3 was uneventful even when carried out on a 60 g scale
matography) and employing 11 directly for coupling with 3
to obtain 14 [13]. Thus, d-valerolactam was converted to 10
in 77% yield (Scheme 4), following a variation of the syn-
thetic protocol reported by Bristol–Myers Squibb. Upon
subjecting 10 to a base mediated elimination in Li CO -LiCl
2
3
milieu in DMF, the enamide 11 could be obtained in excel-
lent yield even on a 90 g scale. Interestingly, reaction of 10 in
DMF with Li CO alone furnished a significantly lower yield
2
3
of 11, while no product formation was observed when DMF
was replaced with toluene, acetonitrile, THF, or acetone.
1
23

A morpholine-free process amenable convergent synthesis of apixaban: a potent factor Xa…
Scheme 5
into the reaction vessel while maintaining the temperature
of the reaction mass at 0–5 °C. The resulting mixture was
refluxed at 65–70 °C for 4 h. It was then allowed to cool to
3
0
–5 °C, slowly poured into 3.40 dm ice cold water and
carefully rendered basic (pH 9–10) by addition of 50%
aqueous NaOH solution at 0–5 °C. The product was then
extracted into dichloromethane (2 9 1.70 dm ). The
3
3
organic layers were combined, washed with 1.70 dm
brine, dried over anhydrous Na SO , and concentrated
4
2
(
Scheme 4) and directly afforded 14 with yields comparable
to that observed for the two-stage conversion of 12 to 14 (see
under reduced pressure at \50 °C. The crude material
obtained was stirred in 340 cm n-hexane at RT for
3
Scheme 3). With alternate bases such as DIPEA, K CO ,
2
3
5 min, filtered, washed with 100 cm n-hexane, and
3
1
Li CO , and Cs CO , reaction of 11 with 3 was typically less
3
2
3
2
finally dried at RT for 4 h to obtain 10 (221 g, 77%,
8.4% purity by HPLC analysis) as a white solid.
efficient and gave a low yield of 14. With the 5,6-dihydro-1H-
pyrazolo[3,4-c]pyridin-7(4H)-one 14 in hand, we proceeded
with the two known end-game steps in the synthesis of 1.
Thus, a CuI/L-proline catalyzed arylation of 14 with 5a
afforded 2 in good yield. The ethyl ester 2 smoothly under-
went ammonolysis in methanolic ammonia to furnish the
eventual apixaban in 79% yield and with an HPLC purity of
9
3-Chloro-5,6-dihydropyridin-2(1H)-one (11)
3
DMF (1.35 dm ), 90 g 10 (0.536 mol), 79.2 g Li
(1.071 mol), and 22.7 g LiCl (0.536 mol) were sequentially
CO
3
2
3
charged at RT into a 3 dm three-necked round-bottomed
flask, fitted with a condenser, a mechanical stirrer, and a
9
9.5%.
CaCl guard tube. The reaction mixture was then stirred at
2
3
130 °C for 4 h, cooled to RT and poured into 2.70 dm ice
cold water. The resulting suspension was filtered through a
3
Celite bed and the residue washed with 900 cm dichlor-
Conclusion
omethane. The combined filtrate was extracted with
3
dichloromethane (3 9 1.80 dm ). The organic layers were
In summary, we have demonstrated a convergent and
scalable synthesis of the anti-coagulant drug apixaban. The
synthetic route described herein is frugal, but efficient, in
its approach and bypasses the safety concerns associated
with the use of large quantities of morpholine.
combined, dried over anhydrous Na SO , concentrated
2 4
under reduced pressure, and finally dried under high vacuum
at 60 °C for 1 h to obtain 11 (63.6 g, 90%, 97.5% purity by
HPLC analysis) as a dark brown solid. H NMR spectral data
1
of 11 (see Supplementary Material for details) was found to
be consistent with the values reported in Ref. [14].
Experimental
3
-Morpholino-5,6-dihydropyridin-2(1H)-one (12)
3
Morpholine (14 cm ) and 7 g 10 (0.0417 mol) were
1 13
H and C NMR spectra were recorded on a Varian 400 MHz
3
charged into a 50 cm single neck round-bottomed flask,
spectrometer. LR-MS data were recorded on an Agilent 1200
Series liquid chromatography module hyphenated to a 6430
Triple Quad LC/MS system. HPLC was performed with a
Waters Alliance e2695 separation module, equipped with a
Waters 2998 photodiode array detector (column: Symmetry
equipped with a stir bar, a condenser, and a CaCl guard
2
tube. The resulting mixture was refluxed for 2 h, cooled to
3
RT, and poured into 50 cm ice cold water. The product
3
was extracted with dichloromethane (3 9 40 cm ). The
3
organic layers were combined, washed with 30 cm water
3
Shield RP 18, 250 9 4.6 mm, 5 lm; flow rate: 1.0 cm /min).
3
and 30 cm brine, dried over Na SO , and concentrated
2
4
Reactions were monitored by thin layerchromatography (TLC)
performed on Merck TLC silica gel 60 F254 aluminium plates.
The hydrazone 3 and N-arylated d-valerolactam 5a were syn-
thesized following procedures already known in the literature
and patent prior art [7, 8].
under reduced pressure at\45 °C. The crude material thus
obtained was stirred with 15 cm MTBE at RT for 10 min,
3
3
filtered, washed with 10 cm n-hexane, and dried at RT for
4
h to obtain 12 (5.6 g, 74%, 95.3% purity by HPLC
1
analysis) as a dark brown solid. H NMR spectral data of
2 (see Supplementary Material for details) was in
agreement with that reported in Ref. [14].
1
3
,3-Dichloropiperidin-2-one (10)
3
CHCl (2.89 dm ) was charged into a 5 dm three-necked
3
3
round-bottomed flask, fitted with a condenser, a mechan-
ical stirrer and a CaCl2 guard tube. PCl5 (1.071 kg,
Ethyl 1-(4-methoxyphenyl)-7-oxo-4,5,6,7-tetrahydro-1H-
pyrazolo[3,4-c]pyridine-3-carboxylate (14)
3
Toluene (1.55 dm ), 62 g 11 (0.471 mol), 242 g 3
(0.943 mol), and 286 g triethylamine (2.828 mol) were
5
.144 mol), followed by solution of 170 g d-valerolactam
1.715 mol) in 510 cm CHCl , was gradually introduced
3
(
3
123

N. R. Nevuluri et al.
3
sequentially charged at RT into a 3 dm three-necked
3
10 min, diluted with 2.55 dm dichloromethane, and then
stirred again for 10 min at RT. Thereafter, ammonia gas
was bubbled for 5 h into the resulting mixture which was
maintained at 0–5°C during this period. The reaction was
then allowed to stir at RT for 72 h and concentrated under
reduced pressure at \45 °C upon completion. The crude
round-bottomed flask, fitted with a condenser, a mechan-
ical stirrer, and a CaCl guard tube. The reaction mixture
2
was refluxed at 130 °C for 4 h, cooled to RT, poured into
3
30 cm ice cold water, and extracted with dichloro-
9
3
methane (4 9 930 cm ). The organic layers were
3
combined, washed with brine (2 9 930 cm ), dried over
3
material obtained was stirred with 1.275 dm methanol for
anhydrous Na SO , and concentrated under reduced pres-
2
15 min at 65 °C; the mixture was then cooled to RT and
stirred for 30 min. The separated product was filtered,
4
sure at \55 °C. The crude material thus obtained was
stirred with 620 cm MTBE for 1 h at RT, filtered, washed
3
3
washed with 510 cm methanol, and finally dried at 60 °C
for 8 h to obtain 1 (190 g, 79%, 99.5% purity by HPLC
1
analysis) as an off-white solid. H NMR spectral data of 1
3
with 310 cm MTBE, and finally dried at 45 °C for 2 h to
obtain 14 (83 g, 56%, 99.3% purity by HPLC analysis) as a
1
brown solid. H NMR spectral data of 14 (see Supplemen-
(see Supplementary Material for details) were found to be
consistent with the values reported in Ref. [8].
tary Material for details) was found to be consistent with
the values described in Ref. [15].
Ethyl 1-(4-methoxyphenyl)-7-oxo-6-[4-(2-oxopiperidin-1-
yl)phenyl]-4,5,6,7-tetrahydro-1H-pyrazolo[3,4-c]pyridine-
References
3
-carboxylate (2)
De-gassed toluene (1.968 dm ), 82 g 14 (0.260 mol),
10 g 5a (0.364 mol), 381 g Cs CO (1.171 mol), 24.8 g
1. Granger CB, Alexander JH, McMurray JJ, Lopes RD, Hylek EM,
Hanna M, Al-Khalidi HR, Ansell J, Atar D, Avezum A, Bahit
MC, Diaz R, Easton JD, Ezekowitz JA, Flaker G, Garcia D,
Geraldes M, Gersh BJ, Golitsyn S, Goto S, Hermosillo AG,
Hohnloser SH, Horowitz J, Mohan P, Jansky P, Lewis BS, Lopez-
Sendon JL, Pais P, Parkhomenko A, Verheugt FW, Zhu J, Wal-
lentin L (2011) N Engl J Med 365:981
3
1
2
3
CuI (0.130 mol), and 12.0 g L-proline (0.104 mol) were
3
sequentially charged at RT into a 2 dm four-necked
round-bottomed flask, fitted with a CaCl guard tube, a
2
2
3
4
5
. Greig SL, Garnock-Jones KP (2016) Drugs 76:1493
. Hull RD, Gersh MH (2015) Curr Med Res Opin 31:197
. Boey JP, Gallus A (2016) Drugs Aging 33:475
. Dobesh PP, Fanikos J (2015) Drugs 75:1627
mechanical stirrer, and a Dean-Stark condenser. The
resulting suspension was stirred and refluxed at 120 °C
for 24 h; thereafter, it was cooled to RT, diluted with
3
92 cm EtOAc, and filtered through a Celite bed. The
4
6. Why eliquis is so important for Bristol–Myers squibb forbes.
Available at: http://www.forbes.com/sites/greatspeculations/
3
residue on the Celite bed was washed with 328 cm
EtOAc. The organic layer was separated from the com-
2
#
015/09/15/why-eliquis-is-so-important-for-bristol-myers-squibb/
757f0478518c. Accessed 15 Sept 2015
3
bined filtrate, washed with water (2 9 656 cm ) and brine
7
8
. Pinto D, Quan M, Orwat M, Li Y-L, Han W, Qiao J, Lam P, Koch
S (2003) Lactam-containing compounds and derivatives thereof
as factor Xa inhibitors. PCT Patent Application WO 03/026652
A1, 03 Apr 2003; Chem Abstr 138:411267
. Pinto DJP, Orwat MJ, Koch S, Rossi KA, Alexander RS,
Smallwood A, Wong PC, Rendina AR, Luettgen JM, Knabb RM,
He K, Xin B, Wexler RR, Lam PYS (2007) J Med Chem 50:5339
3
2 9 656 cm ), dried over Na SO , and finally concen-
(
2
4
trated under reduced pressure at \60 °C. The crude
material thus obtained was diluted with 4.92 dm dichlor-
3
omethane and stirred with 41 g silica gel (230–400 mesh)
at RT for 10 min. The suspension was filtered through a
3
Celite bed and the residue washed with 820 cm dichlor-
9. Lewis RJ Sr (2008) Hazardous chemicals desk reference, 6th edn.
Wiley, New York, p 970
0. Rajan ST, Eswaraiah S, Venkatesh M (2015) Novel intermediate
and polymorphs of 1-(4-methoxyphenyl)-7-oxo-6-[4-(2-ox-
opiperidin-1-yl)phenyl]-4,5,6,7-tetrahydro-1H-pyrazolo[3,4-
c]pyridine-3-carboxamide and process thereof. US Patent
015/0353541 A1, 10 Dec 2015; Chem Abstr 2016:2037398
1. Zhou J, Oh LM, Ma P, Li H-Y, Confalone PN (2003) Synthesis of
,5-dihydro-pyrazolo[3,4-c]pyrid-2-ones. PCT Patent Application
omethane. The combined filtrate was concentrated under
1
1
reduced pressure at \40 °C. The crude material obtained
was stirred with 328 cm MTBE for 60 min, filtered,
3
3
washed with 164 cm MTBE and dried at 50 °C for 4 h to
obtain 2 (92 g, 72%, 98.4% purity by HPLC analysis) as an
2
1
off-white solid. H NMR spectral data of 2 (see Supple-
4
mentary Material for details) were found to be consistent
with the values reported in Ref. [8].
WO 03/049681 A2, 19 Jun 2003; Chem Abstr 139:53014
2. Jiang J, Ji Y (2013) Synth Commun 43:72
3. This work was disclosed in: Iqbal J, Oruganti S, Rapolu RK
1
1
(
2
2016) Preparation of Apixaban. Indian Patent Application IN
013CH00639 A, 27 May 2016; Chem Abstr 2016:881266
1
-(4-Methoxyphenyl)-7-oxo-6-[4-(2-oxopiperidin-1-yl)phe-
nyl]-4,5,6,7-tetrahydro-1H-pyrazolo[3,4-c]pyridine-3-car-
1
1
4. Wu Y, Min X, Zhuang C, Li J, Yu Z, Dong G, Yao J, Wang S,
Liu Y, Wu S, Zhu S, Sheng C, Wei Y, Zhang H, Zhang W, Miao
Z (2014) Eur J Med Chem 82:545
5. Ding Z, Lai G, Chen S, Yan X (2015) Hydrazine compound as
blood coagulation factor Xa inhibitor. PCT Patent Application
WO 2015/176625 A1, 26 Nov 2015; Chem Abstr 164:36952
boxamide (apixaban, 1)
3
Methanol (12.75 dm ) and 255 g 2 (0.522 mol) were
3
charged into a 20 dm four-necked round-bottomed flask,
equipped with a KOH guard tube, a mechanical stirrer, and
a gas inlet tube. The reaction mass was stirred at RT for
1
23