Scalable synthesis of the VEGF-R2 kinase inhibitor JNJ-17029259 using ultrasound-mediated addition of MeLi-CeCl3 to a nitrile

Article abstract of DOI:10.1021/jo7021372

(Chemical Equation Presented) The preparation of the selective VEGF-R2 kinase inhibitor 10 (JNJ-17029259) is described in which the key precursor, 4-(5-isoxazolyl)benzonitrile, undergoes clean transformation to the corresponding cumylamine derivative with

Full text of DOI:10.1021/jo7021372

SCHEME 1. Retrosynthetic Analysis
Scalable Synthesis of the VEGF-R2 Kinase
Inhibitor JNJ-17029259 Using
Ultrasound-Mediated Addition of MeLi-CeCl3 to
a Nitrile
Michael Reuman,*,† Sandra Beish,† Jeremy Davis,‡
Mark J. Batchelor,‡ Martin C. Hutchings,‡
David F. C. Moffat,‡ Peter J. Connolly,† and
Ronald K. Russell†
Johnson and Johnson Pharmaceutical Research & DeVelopment,
L.L.C., 1000 Route 202, Raritan, New Jersey 08869, and UCB
Pharma, 216 Bath Road, Slough, SL14EN, United Kingdom
mreuman@prdus.jnj.com
ReceiVed October 1, 2007
SCHEME 2. Preparation of 4-(5-Isoxazolyl)benzonitrile
SCHEME 3. Addition of MeLi-CeCl3 to
4-(5-Isoxazolyl)benzonitrile
priate guanidine and acrylonitrile derivatives; this acrylonitrile
is derived from an isoxazole.3 While the isoxazole is easily
prepared from commercially available 4-acetylbenzonitrile
(Scheme 2),4 the conversion of the nitrile to the required benzyl
amine is especially problematic because of the sensitivity of
the isoxazole to base.
The preparation of the selective VEGF-R2 kinase inhibitor
10 (JNJ-17029259) is described in which the key precursor,
4-(5-isoxazolyl)benzonitrile, undergoes clean transformation
to the corresponding cumylamine derivative with CeCl3-
MeLi in THF. This high-yielding cerium mediated transfor-
mation is robust, reproducible, and readily scalable based
on a requirement for the anhydrous CeCl3 to be milled and
subjected to ultrasound treatment prior to addition of
methyllithium.
The use of organocerium reagents appeared to be appropriate
as the double addition of these reagents to nitriles had been
reported.5 Under the standard conditions for preparing and using
these reagents, the use of MeLi-CeCl3 (Scheme 3) was
routinely proven to be inconsistent and variable yields (0 to
95% isolated crude product) were observed. The results were
independent of whether the anhydrous CeCl3 was generated from
the heptahydrate or the commercially available anhydrous
reagent. In cases where poor results were obtained, the reaction
mixture usually became a bright yellow color and the product
amine was often accompanied by unidentified solid impurities.
Pyrimidinecarbonitrile derivative (10) is a promising anti-
angiogenic lead as a potent inhibitor of VEGF-R2 kinase. This
compound exhibits significant oral activity in suppressing tumor
growth in murine models and is also a potent inhibitor of corneal
neovascularization in vivo.1,2 The synthesis of this target pre-
sented a unique challenge in that it contained a 2-aminopyri-
midine-5-carbonitrile bearing a dimethylbenzyl amine (cumyl-
amine) group. A retrosynthetic analysis (Scheme 1) shows the
construction of the pyrimidine unit by condensation of the appro-
The reason for the inconsistent results obtained in our initial
runs is not proven, but it is most likely due to incomplete
activation of CeCl3 and subsequent incomplete formation of the
MeLi-CeCl3 complex. Sonication has been described as an
alternative method to facilitate the addition of organocerium
reagents to carbonyl compounds.6 In the reported study, the
† Johnson and Johnson Pharmaceutical Research & Development.
‡ UCB Pharma.
(1) Emanuel, S.; Gruninger, R. H.; Fuentes-Pesquera, A.; Connolly, P.
J.; Seamon, J. A.; Hazel, S.; Tominovich, R.; Hollister, B.; Napier, C.;
D’Andrea, M. R.; Reuman, M.; Bignan, G.; Tuman, R.; Johnson, D.; Moffat,
D.; Batchelor, M.; Foley, A.; O’Connell, J.; Allen, R.; Perry, M.; Jolliffe,
L.; Middleton, S. A. Mol. Pharmacol. 2004, 66, 635.
(2) Hughes, T. V.; Emanuel, S. L.; Beck, A. K.; Wetter, S. K.; Connolly,
P. J.; Prabha, K.; Reuman, M.; Seraj, J.; Fuentes-Pesquera, A. R.; Gruninger,
R. H.; Middleton, S. A.; Lin, R.; Davis, J. M.; Moffat, D. F. C. Bioorg.
Med. Chem. Lett. 2007, 17, 3266.
(3) Batchelor, M. J.; Moffat, D. F. C.; Davis, J. M.; Hutchings, M. C.
U.S. Patent 6,579,983, 2003.
(4) (a) Garcia, E. E.; Benjamin, L. E.; Fryer, R. I. J. Heterocycl. Chem.
1974, 11, 275. (b) Lin, Y-I.; Lang, S. A., Jr. J. Heterocycl. Chem. 1977,
14, 345. (c) Lin, Y.-I.; Lang, S. A., Jr. J. Org. Chem. 1980, 45, 4857. (d)
Swain, C. J.; Teran, A.; Maroto, M.; Cabello, A. Bioorg. Med. Chem. Lett.
2006, 16, 6058.
(5) Ciganek, E. J. Org. Chem. 1992, 57, 4521.
10.1021/jo7021372 CCC: $40.75 © 2008 American Chemical Society
Published on Web 01/03/2008
J. Org. Chem. 2008, 73, 1121-1123
1121

TABLE 1. Addition of MeLi-CeCl3 to
SCHEME 4
4-(5-Isoxazolyl)benzonitrilea
isoxazole
(g)
sonication
time (h)
CeCl3
(g)
yield
(%)
entry
footnote
1
2
3
4
5
6
7
8
5.75
5.75
5.75
5.75
5.75
10.87
44.85
68.6
none
2
4
5.5
5
4
25
25
25
25
25
47.13
195
298.4
0
30
95
45
47
83
100
100
b
b
b
b
b
c
d
d, e
21
43
a CeCl3-MeLi complex was formed by addition of 0.85 mol of freshly
titrated MeLi-Et2O to each mol of CeCl3. The product purity was >80%
by NMR unless otherwise stated. b CeCl3 as 10 mesh beads with magnetic
stirring. c Milled CeCl3 with magnetic stirring. d Milled CeCl3 with soni-
cation and simultaneous overhead stirring. e For runs of this scale, the results
were unchanged with stirring and sonication times between 16 and 48 h.
2-5 and was encouraging. Magnetic stirring became impractical
as the reaction scale was increased; for a run attempted using
17 g of nitrile 2 and 74 g of CeCl3, the mixture could no longer
be stirred reliably using magnetic means even though the CeCl3
had been milled prior to the start of the run. Additionally,
commercial CeCl3 in THF is a heavy suspension that needs
strong stirring to prevent it from settling. For larger scale runs,
magnetic stirring was changed to overhead stirring to ensure
that the CeCl3 would remain well suspended during the reaction
(entries 7 and 8). This change also allowed for simultaneous
stirring and sonication operations.
Typically, large-scale runs were set up to sonicate and stir
for 16 h (overnight); stirring and sonication times longer than
this were not detrimental to the product yield or purity. In all
cases, when the mixture was stirred and sonicated, a quantitative
yield of the amine was obtained and product purity was high
(>80%, NMR estimate). The reaction noted in entry 8 was
performed several times on the scale indicated; the scale up
was limited by the size of the reaction flask (5 L) that could be
placed into the sonicator. The color of the reaction mixture
typically became faint yellow during or after the addition of
methyllithium. These runs invariably led to good conversion to
the desired product. In runs where the reaction picked up a
pronounced deep yellow color9 during the addition, the yield
was poor or the reaction failed completely. This observation
was more common for the initial magnetically stirred runs and
not typical for those done with combined overhead stirring and
sonication.
With a proven method for the preparation of the desired
benzylamine 3a, this compound was converted immediately into
the Boc derivative 3b (Scheme 4). The yield of 3b over both
steps was 82%. The isoxazole was opened with NaOC2H5 to
give cyanoketone 4 (92%); this was next converted into
acrylonitrile 5 (62%). Guanidine 8 was prepared from com-
mercially available 4-nitrophenylethyl alcohol (Scheme 5). This
alcohol was coupled to morpholine through the corresponding
mesylate intermediate.10 Reduction of the nitro group provided
aniline 7,11 which was finally converted to guanidine 8 by
treatment with 3,5-dimethylpyrazole-1-carboxamidine nitrate (90
°C) for 1 h (43%).12
CeCl3-THF mixture was sonicated for 1 h prior to cooling the
suspension and generating the organocerium complex leading
to 77% to 97% isolated yields of the addition product for the
examples reported. In a subsequent report, CeCl3 sonication was
used in the addition of MeLi-CeCl3 to a nitrile during the
preparation of the (R)-3-(1-amino-1-methylethyl)pyrrolidine
group of the antiinfective agent PD-138312.7 Sonication was
shown to be important to aid in the addition of organocerium
reagents to carbonyl groups and may also be important for
cerium-mediated additions to the nitrile group. While encour-
aged by these examples, there was no report of this addition in
the presence of the strongly sensitive isoxazole. Nevertheless,
it is clear that sonication should, at a minimum, help to keep
the CeCl3 more finely suspended in THF facilitating its
activation.
In our work, commercially available anhydrous CeCl3 (10
mesh beads) was preferred over the heptahydrate because the
drying step was eliminated.8 Examination of representative runs
(Table 1) indicates that little product was isolated when the
CeCl3 suspension was stirred magnetically prior to methyllithium
addition (entry 1). When the CeCl3 in THF was initially
sonicated prior to methyllithium addition, the yields improved,
but inconsistent results were still obtained (entries 2-5). It is
possible that these results were due to nonuniform or excessively
large CeCl3 particle size obtained even after the sonication of
the commercial reagent. The magnetic stirring used in these runs
may be inadequate to fully suspend the CeCl3. The net result
of this situation would be the presence of excess methyllithium
to the extent that a substantial portion of the CeCl3 was unavail-
able to react. To address this possibility, instead of using 10
mesh anhydrous CeCl3, this reagent was ground to a fine powder
using a commercial household coffee bean grinder. The grinding
operation was carried out in a nitrogen-flushed glovebag and
the CeCl3 powder was transferred to a flask in the glovebag.
The initial run (entry 6) with finely ground and sonicated
CeCl3 appeared to address the inconsistent results for entries
(6) Greeves, N.; Lyford, L. Tetrahedron Lett. 1992, 33, 4759.
(7) (a) Fedij, V.; Lenoir, E. A., III; Suto, M. J.; Zeller, J. R.; Wemple,
J. Tetrahedron: Asymmetry 1994, 5, 1131. To the best of our knowledge,
prior to this report, there were no examples of sonication-assisted addition
of MeLi-CeCl3 to nitriles. These authors did not indicate whether sonication
altered the yield, purity, or percent ee of their product. (b) After our work
was completed an alternate method for CeCl3 activation was reported:
Limanto, J.; Dorner, B.; Devine, P. N. Synthesis, 2006, 4143.
(9) Incomplete or improperly dried CeCl3 has been known to cause yellow
gummy reaction mixtures, see: Dimitrov, V.; Kostova, K.; Genov, M.
Tetrahedron Lett. 1996, 37, 6787.
(10) Palmer, B. D.; Lee, H. H.; Johnson, P.; Baguley, B. C.; Wickham,
G.; Wakelin, L. P. G.; McFadyen, W. D.; Denny, W. A. J. Med. Chem.
1990, 33, 3008.
(11) Leffler, M. T.; Volwiler, E. H. J. Am. Chem. Soc. 1938, 60, 896.
(12) An alternate route to guanidine 8 can be found in: Adams, J. L.;
Drewry, D. H.; Linn, J. A. WO 2007/018941 A2.
(8) Takeda, N.; Imamoto, T. In Organic Synthesis; Martin, S. F., Ed.;
John Wiley & Sons, Inc.: New York, 1999; Vol. 76, p 228.
1122 J. Org. Chem., Vol. 73, No. 3, 2008

SCHEME 5
reagent was ground in 100-150 g portions for 1 min and then
transferred to a dry, nitrogen-flushed 500-mL single-neck round-
bottom flask inside the glovebag. After the entire portion of the
reagent was ground, the flask was sealed with a septum, removed
from the glovebag, and weighed.
Preparation of the CeCl3-THF complex: The anhydrous CeCl3
powder (298.4 g, 1.21 mol) prepared above was transferred to a
dry, nitrogen-flushed 5-L 4-neck Morton flask equipped with an
overhead stirrer, nitrogen inlet adapter, thermocouple, and septum.
After the transfer, the flask was evacuated and filled with nitrogen
(3×) and cooled in an ice-water bath for 30 min, then anhydrous
THF (1.85 kg) was added to the stirred solid by cannula. After the
THF addition, the cooling bath was removed and replaced by a
sonicator. The mixture was sonicated continuously while stirring
typically for 16-48 h.
Representative run: After the sonication with stirring, the
sonicator was removed and the mixture cooled to an internal
temperature of less than -75 °C. After this point was reached,
freshly titrated MeLi (1.57 M, 655 mL, 1.03 mol) in ether was
added using a gastight 100-mL syringe. The resulting pale yellow
mixture was stirred for 1 h at <-75 °C. A solution of isoxazole 2
(68.6 g, 0.403 mol) in anhydrous THF (1.29 kg) was added to the
cold CeCl3-MeLi mixture over 0.75 h via cannula. After the
reaction was stirred at -75 °C for 1 h, the cooling bath was removed
and the mixture was allowed to warm to 0 °C over 1 h. This mixture
was re-cooled to -75 °C prior to workup.
SCHEME 6
Workup: A 22-L, 4-neck Morton flask equipped with an
1
overhead stirrer, nitrogen inlet adapter, stopper, and a /4 in. ID
Teflon tube was charged with EtOAc (3 L) and concentrated
NH4OH (1 L). The above reaction mixture at -75 °C was
transferred into the stirred flask containing EtOAc and NH4OH with
a slight nitrogen pressure and vacuum using the Teflon tube. After
the transfer was completed, the reaction flask was rinsed with
additional EtOAc (2 × 1.5 L). The combined EtOAc solution was
filtered and dried overnight with Na2SO4 (500 g). The solution
was filtered and concentrated in vacuo using a diaphragm pump
(20-25 IHg) with the bath maintained near room temperature. After
most of the solvent was removed, the pump was changed to a high-
vacuum pump (25-28 IHg) and the bath temperature was increased
to 30 °C. The mixture was allowed to cool to give 86 g (>100%)
As shown in Scheme 6, coupling of acrylonitrile 5 with
guanidine nitrate 8 was accomplished by heating the components
in 2-propanol in the presence of NaOH to give 9 (75%);
deprotection using TFA provided the target pyrimidine 10
(97%). This product was crystallized (EtOH) to give a 79%
isolated yield that was increased to 97% by additional material
provided from the mother liquors.
An efficient synthesis of VEGF-R2 kinase inhibitor 10 has
been described in which the pivotal step in the reaction sequence
was a stirred-ultrasound-assisted cerium-mediated preparation
of a key cumyl amine intermediate 3a. The ultrasound modi-
fication has made this step highly dependable and easily scaled
to provide approximately 80 g per run in a 5-L reaction flask.
The modification to Ciganek’s method5 described in this report
for the MeLi-CeCl3 addition to a nitrile requires no special
equipment and should have applicability to other substrates.
1
of 3a as a light yellow oil. H NMR (CDCl3) indicated that the
product was about 90+% pure. The product was used immediately
in the next step without any purification. MS: [MH]+ 203. 1H NMR
(CDCl3): δ 8.30 (d, 1H, J ) 1.9 Hz), 7.77 (d, 2H, J ) 8.6 Hz),
7.64 (d, 2H, J ) 8.6 Hz), 6.50 (d, 1H, J ) 1.9 Hz), 1.55 (br s, 2H),
1.50 (s, 6H).
Acknowledgment. The authors thank Stephen Stefanick for
guidance and safety studies on guanidine nitrate 8, Gilles Bignan
for helpful chemistry discussions, and Richard Dunphy for
HRMS analysis (all from Johnson and Johnson Pharmaceutical
Research & Development, L.L.C.). In addition, we would also
like to thank Prof. Thomas Livinghouse, Montana State
University, Bozeman, MT, for important discussions concerning
alternate methods for drying hydrated lanthanide reagents.
Experimental Section
Starting materials, reagents, and solvents obtained from com-
mercial suppliers were used without any purification. Methyllithium
in ether was assayed before use as described in the literature
(Kofron, W. G.; Baclawski, L. M. J. Org. Chem., 1976, 41, 1879).
General Procedure for MeLi-CeCl3 Addition to 4-(5-Isox-
azolyl)benzonitrile (2) To Give 3a. Preparation of cerium(III)
chloride: Anhydrous CeCl3 (99.9%), supplied as 10 mesh beads
(packaged in ampules), was ground in a coffee bean grinder in a
nitrogen-filled glovebag. For a run requiring 300 g of CeCl3, the
Supporting Information Available: Experimental details for
the preparation of compounds 1, 2, and 3b-10 and spectral
characterization data for compounds 3a-5 and 8-10. This material
is available free of charge via the Internet at http://pubs.acs.org.
JO7021372
J. Org. Chem, Vol. 73, No. 3, 2008 1123