Development an efficient route to the 5-lipoxygenase inhibitor PF-04191834

Article abstract of DOI:10.1021/op200173g

A convergent six-step process for the synthesis of PF-04191834 (1), a potent and selective 5-lipoxygenase inhibitor, has been developed and used to deliver over 20 kg of API. The process uses the same bond-forming steps as the initial medicinal chemistry

Full text of DOI:10.1021/op200173g

ARTICLE
pubs.acs.org/OPRD
Development an Efficient Route to the 5-Lipoxygenase Inhibitor
PF-04191834
Pieter D. de Koning* and Lorraine Murtagh
Pfizer Global Research and Development, Ramsgate Road, Sandwich, Kent CT13 9NJ, U.K.
Jon P. Lawson,*,† Richard A. Vonder Embse,‡ Sastry A. Kunda,§ and Wei Kong^
Pfizer Global Research and Development, St. Louis Laboratories, 700 Chesterfield Parkway West, Chesterfield, Missouri 63017,
United States
ABSTRACT: A convergent six-step process for the synthesis of
PF-04191834 (1), a potent and selective 5-lipoxygenase inhi-
bitor, has been developed and used to deliver over 20 kg of API.
The process uses the same bond-forming steps as the initial
medicinal chemistry route, including the use of two consecutive
Pd-catalyzed ArꢀS couplings to form the key diaryl thioether
linkage. The reaction conditions and downstream processing
have been optimized to eliminate column chromatography and
aqueous work-ups and to minimize disproportionation of 1, to
ensure successful scale-up.
’ INTRODUCTION
Another initial concern was the use of TIPS-thiol 7 as the
hydrogen sulfide surrogate.8 However, while recognizing the
poor atom economy9 and potential long-term challenges asso-
ciated with this reagent (unpleasant odor, limited number of
vendors, long leadtime), we were able to source adequate
supplies for early phase development,10 and thus no alternative
sources of sulfur were examined.
Leukotrienes are a family of highly potent lipid pro-inflam-
matory mediators that play a significant role in a number of
allergic and inflammatory diseases.1 These bioactive lipids are
generated by metabolism of arachidonic acid, initially to leuko-
triene A4 (LTA4), by the enzyme 5-lipoxygenase (5-LO)2 and its
accessory activating partner protein, 5-LO-activating protein
(FLAP).3 Subsequently LTA4 may be further metabolized by
LTA4 hydrolase to form LTB4, a potent chemoattractant for
neutrophils, macrophages and other inflammatory cells. Alter-
natively, metabolism of LTA4 by LTC4 synthase enzyme leads to
the cysteinyl leukotrienes, LTC4, LTD4 and LTE4, which in-
crease vascular permeability, contract smooth muscles cells and
increase mucus production. Inhibiting the production of these
pro-inflammatory mediators generated by the 5-LO pathway has
potential therapeutic benefit in numerous indications, including
asthma, COPD, atherosclerosis,4 pain5 and cancer.6 The key
role of 5-LO in catalyzing the first step of the metabolism of
arachidonic acid has therefore made it an important pharmaco-
logical target for a number of disease areas.
As part of a 5-LO drug discovery program, the potent and
selective 5-LO inhibitor PF-04191834 (1) was identified7 and
progressed into development. From an inspection of the medicinal
chemistry route (Scheme 1), it was evident that the convergent
synthesis developed would be an appropriate strategy to assemble the
active pharmaceutical ingredient (API); however, the possibility of
reversing the step order and/or telescoping the two Pd-catalyzed
steps were obvious areas for exploration. The use of chromatog-
raphy to separate the pyrazole regioisomers 4 and 5, as well as to
purify the API, was also addressed in our development program.
’ RESULTS AND DISCUSSION
The original synthesis of pyrazole 4 entailed heating a mixture
of 4-bromoacetophenone 2 and dimethylformamide dimethyl
acetal (DMFDMA) in DMF at 125 ꢀC for 3 h.7 The cooled
reaction mixture was concentrated to a thick oil, redissolved in
DMF and treated with methyl hydrazine (3 equiv) at 75 ꢀC. Crude
4 (containing 5% of the regioisomer 5) was isolated by concentra-
tion to dryness and purified by chromatography on silica gel. While
this was suitable for small-scale preparations, it was not appropriate
for our scale-up facilities and required some modification.
The purpose of the first concentration to dryness was to
remove excess DMFDMA to avoid reaction with methyl hydra-
zine. Postulating that the condensation product(s) of DMFDMA
and methyl hydrazine should be water-soluble and therefore
purge upon aqueous quench, we optimized the relative amounts
of both DMFDMA (2.0 equiv) and methyl hydrazine (4.0 equiv)
to ensure complete consumption of 3. As anticipated, any
DMFDMA-methyl hydrazine related byproducts were purged
in the aqueous work-up.11
Received: June 28, 2011
Published: August 18, 2011
r
2011 American Chemical Society
1046
dx.doi.org/10.1021/op200173g Org. Process Res. Dev. 2011, 15, 1046–1051
|

Organic Process Research & Development
ARTICLE
Scheme 1. Medicinal Chemistry Route to PF-04191834 (1)a
a Reagents and conditions: (a) N,N-dimethylformamide dimethylacetal, DMF, reflux, 4ꢀ5 h; (b) methylhydrazine, DMF, 75 ꢀC, 4 h, 68%; (c) NaOtBu,
Pd(OAc)2, 1,1-bis(diisopropylphosphino) ferrocene, 1,4-dioxane, reflux, 1 h, 84%; (d) (Ph3P)4Pd, bis[(2-diphenylphosphino)]phenylether, 1 M KOt-
Bu in THF, i-PrOH/5% H2O, 90 ꢀC, 4 h, 86%.
Scheme 2. Literature Synthesis of Amide 6 (from ref 13a)a
a Reagents and conditions: (a) aqueous NaOH, n-Bu4NHSO4, THF, reflux; then IPA/water cryst, 68%; (b) KOH, IPA, reflux; then water, 89%.
We then examined the effect of temperature on the condensa-
tion of 3 with methyl hydrazine, in the hope that lower tempera-
tures would reduce the level of regioisomer 5 formed. Fortunately
this proved to be the case; when the reaction was conducted at
25 ꢀC, the level of 5 was halved (to around 2.5%), although the
reaction took ∼16 h to reach completion (compared to 3 h in the
original procedure). Given the challenges of separating the regio-
isomer 5 from 4, this was an acceptable compromise. In order to
simplify the work-up, MTBE was used as solvent for the con-
densationreaction(in conjunctionwith DMF from thefirst stage),
and this had no effect on the regioselectivity or reaction rate.
The product solution was then washed with aqueous ammonium
chloride to remove DMF and any DMFDMA-methyl hydrazine
adducts. The resulting solution was filtered through silica gel (∼3 g
silica/g of 2; eluting with MTBE) to remove polar impurities that
adversely affected the crystallization. Solvent exchange to heptane
via distillation and controlled crystallization, seeding at 40ꢀ45 ꢀC,
efficiently removed the regioisomer, affording the desired pyrazole
4 in an acceptable 68% overall yield from bromoacetophenone 2.
This procedure was then successfully outsourced. All batches of 4
were analyzed to ensure that the level of residual methyl hydrazine
was below the threshold of toxicological concern (TTC).12
The other half of the API, amide 6, is a known compound,13
prepared in a relatively straightforward fashion by base-mediated
condensation of commercially available 2-(3-bromophenyl)aceto-
nitrile and bis(2-chloroethyl)ether, followed by nitrile hydrolysis
(Scheme 2). As an alternative to the literature procedure (KOH
in refluxing IPA), refluxing with KOH in tert-butanol or treat-
ment with sulfuric acid7 could also be used to hydrolyze the
nitrile. This process was outsourced to support all manufacturing
campaigns. Analysis of 6 showed that the level of residual bis(2-
chloroethyl)ether was below the TTC.
The crux of the synthesis was the two consecutive Pd-
catalyzed carbonꢀsulfur bond-forming steps (occasionally re-
ferred to as the Migita reaction).14 In addition to a careful evalu-
ation of the existing route, both the alternative step-reordered
sequence (first reacting bromopyrazole 4 with TIPS-thiol 7,
followed by amide 6) and the completely telescoped process
were examined. While in the long term a telescoped sequence
would offer the opportunity for streamlined processing, at this
early stage, the importance of having a crystalline late-stage
intermediate (namely, 8) to control impurities was paramount,
and therefore the telescoped process was not developed beyond
the initial, successful proof of concept experiments. Since the
intermediate S-TIPS pyrazole in the step-reordered sequence
was not crystalline, the route via 8 was selected.
The initial synthetic procedure for the coupling between
bromide 6 and TIPS thiol 7 (1.1 equiv) used the Buchwald
conditions,15 namely, sodium tert-butoxide (1.2 equiv) in refluxing
1,4-dioxane in the presence of palladium(II) acetate (2 mol %)
and 1,10-bis(diisopropylphosphino)ferrocene (2.4 mol %). After
an aqueous work-up, 8 was crystallized in excellent purity from
1047
dx.doi.org/10.1021/op200173g |Org. Process Res. Dev. 2011, 15, 1046–1051

Organic Process Research & Development
ARTICLE
Scheme 3. Palladium-Mediated Equilibration of 1
ethyl acetateꢀhexanes (80% yield). While this procedure had
been successfully demonstrated on a 300-g scale,7 the yield
dropped significantly on further scale-up, and the process there-
fore required modification prior to transfer into our kilo labora-
tory or pilot plant facilities. The main concern was the stability of
the product 8 towards the aqueous work-up employed. An
investigation showed that protracted work-up times, as would
be expected on scale-up, led to significant degradation of 8,
primarily through loss of the TIPS protecting group and sub-
sequent side reactions (e.g., disulfide formation). A detailed
investigation showed that the aqueous stability of 8 was linked
to pH and especially to temperature (with greatest stability
around pH 7 and at low temperature). A pH-controlled, low
temperature (10 ꢀC) aqueous work-up was developed, but not-
withstanding a successful demonstration thereof at laboratory
scale, the decision was made to develop a nonaqueous work-up
protocol to avoid the hydrolytic instability altogether.
In parallel with these initial work-up investigations, a small
catalyst/solvent screen for the reaction was conducted, from
which [1,10-bis(diphenylphosphino)ferrocene]dichloropalladium-
(II) [PdCl2(dppf)] in toluene emerged as a suitable alternative.16
Replacing the undesirable solvent, 1,4-dioxane, with toluene and
using a readily available, preformed metal complex resolved two
other concerns with the initial synthesis.
Having identified a suitable solvent/catalyst combination,
attention returned to the work-up process. Upon reaction
completion, the mixture consists of a solution containing the
product, excess thiol 7 (1.2 equiv used) and catalyst, plus sus-
pended inorganic salts (primarily sodium bromide). Since the
purpose of the aqueous work-up was largely to remove these
inorganic salts, this was readily replaced by a simple filtration
through Celite. The filtrate was then diluted with heptane to
induce crystallization of 8. To ensure that 8 remained in solution
during the filtration, ethyl acetate was added beforehand, and the
Celite was washed with ethyl acetate.
This process was then scaled up, and reproducibly delivered
good quality 8 in ∼80% yield (up to 20-kg scale). Key points
essential to success were to ensure that the water level was
<0.01% (easily achieved by azeotropic distillation with toluene)
and to rigorously exclude oxygen from the reactor by a thorough
degassing process prior to addition of the catalyst; high oxygen
levels led to catalyst poisoning and incomplete reaction. The
presence of water resulted in the formation of symmetrical sulfide
9 (Scheme 3), through TIPS cleavage and reaction of the
resulting thiol with bromide 6. The use of high quality sodium
tert-butoxide was essential for reliable conversion, with best
results obtained using material containing less than 0.5% NaOH
(as measured by KF titration).
Importantly, we found that addition of catalyst to the reaction
mixture at room temperature followed by a slow heating ramp to
mimic potential scale-up conditions resulted in incomplete con-
version. This was interpreted asresulting from catalyst degradation
by the TIPS thiolate prior to the mixture reaching a temperature
atwhichthecouplingproceeded. The coupling was not exothermic,
and optimal results were obtained if the catalyst was charged at
reaction temperature (75 ꢀC), rather than prior to heating. A small
excess of TIPS thiol (1.2 equiv) was used to ensure complete
consumption of bromide 6 as this did not purge well in the
crystallization of 8. Any residual 6 reacted in the subsequent step,
forming sulfide 9, which was also challenging to purge (vide infra).
In the initial synthesis, the second ArꢀS coupling between
S-TIPS amide 8 and bromo pyrazole 4 was conducted under
basic conditions in aqueous isopropyl alcohol (5% water) using a
mixture of tetrakis(triphenylphosphine)palladium (0) (10 mol %)
and bis[(2-diphenylphosphino)]phenyl ether (10 mol %). The
product (1) was crystallized by addition of aqueous acid and
purified by chromatography on silica gel. Our enabling work was
focused on optimizing the reaction conditions, reducing the
catalyst level and developing a suitable isolation that removed the
need for chromatography while still controlling the level of
residual palladium to below 50 ppm. One unexpected complica-
tion was that the oxidized ligand proved difficult to separate from
1, even via chromatography, and therefore an alternative catalyst
was highly desirable.
A small screen of catalyst/solvent/base combinations was con-
ducted, from which PdCl2(dppf), sodium methoxide (in methanol)
and 2-methyl THF (MeTHF) were identified as the optimal
conditions. The presence of a nucleophilic base and an alcohol
proved essential for the initial in situ deprotection of 8 to
generate the requisite thiolate. A slight excess of bromopyrazole
4 (1.1 equiv) was used to ensure complete consumption of 8.
The choice of MeTHF as solvent has the added advantage that
the product 1 is not particularly soluble and precipitates upon
cooling, with good purge of most impurities (apart from sulfide 9).
This initially isolated crude 1 still contains large quantities of
inorganic salts (notably sodium bromide) and significant levels of
palladium (>5000 ppm). The inorganic salts were removed by
slurrying crude 1 in water; however, removing the palladium was
more challenging. Since 1 has poor solubility in most organic
solvents, the selected approach was to dissolve crude 1 in THF
(70 mL/g), add a suitable scavenger resin (we chose to use
SiliaBond thiol, based on prior experience) and stir at reflux until
the palladium was removed. The palladium purge took up to 24 h,
and to ensure that acceptable levels were achieved, in-process
monitoring of the Pd level by ICP-MS analysis of samples was used
on scale. After filtration to remove the resin, 1 was crystallized by
1048
dx.doi.org/10.1021/op200173g |Org. Process Res. Dev. 2011, 15, 1046–1051

Organic Process Research & Development
ARTICLE
Scheme 4. Enabled Route to 1a
6 h about 3% of the symmetrical sulfides 9 and 10 was observed,
increasing to 5% after 18 h. A control experiment in which the Pd
catalyst was omitted showed no symmetrical sulfides. Repeat
experiments confirmed this result; however, there was some
variability in the levels of 9 and 10 observed, in some cases
requiring 24ꢀ48 h to reach 3%. This suggests that these
symmetrical sulfide impurities can also be formed in the
absence of an external thiolate, in contrast to the literature
mechanism, but due to time constraints we were unable to
conduct further investigations.
On the basis of this result and the literature evidence for the
generation of symmetrical sulfides during extended reaction
periods,17,18 the reaction conditions were modified to minimize
the reaction time (Scheme 4). In practice, the catalyst loading
was increased, in this case to 2 mol % (added in two 1 mol %
portions 1 h apart), and the catalyst was charged while the
reaction mixture was at the required reaction temperature
(70 ꢀC). As a further precaution, all in process analyses were
conducted promptly (within 15 min of the sample being taken).
Under these conditions, the reaction proceeded to completion in
around 2 h on scale. The work-up procedure used previously
was retained, and we were able to prepare 20 kg of suitable quality
1 in around 70% yield, without requiring chromatographic
purification.
a Reagents and conditions: (a) (i) NaOt-Bu, PdCl2(dppf), toluene,
reflux 1 h; (ii) filtration, EtOAc; (iii) heptane cryst, 80%; (b) (i)
PdCl2(dppf), NaOMe (25% w/w in MeOH), 2-MeTHF, 70 ꢀC; (ii)
water reslurry; (iii) SiliaBond thiol, THF, reflux; then heptane; (iii)
MeOH reflux; 70% overall.
In conclusion, a robust, convergent, four-step process to
prepare PF-04191834 (1) has been developed and has been
used to deliver over 20 kg of 1 in 37% overall yield from
4-bromoacetophenone. Optimal catalyst, base and solvent com-
binations were identified for the two consecutive palladium-
catalyzed carbonꢀsulfur couplings, and an effective strategy to
purge palladium residues and deliver the correct polymorph was
developed. Palladium-mediated disproportionation of 1 was
observed, but this could be suppressed by adding the catalyst
to the reaction at 70 ꢀC and minimizing the reaction time.
partial concentration and addition of heptane. However, this process
provided the wrong polymorph, and consequently an additional
reslurry in refluxing methanol was required to convert the material
to the correct polymorph. While this work-up process was labo-
rious and rather inefficient, we were unable to identify an alter-
native in the time available, and therefore it was used on scale.
The process above was scaled to provide 4 kg of suitable
quality 1 in 73% yield from 8. However, when the process was
scaled up to ∼15 kg, high levels of the sulfides 9 and 10 were
observed (around 5% of each, Scheme 3). The bis-pyrazole
sulfide 10 purged in the initial crystallization from MeTHF, but
the bis-amide sulfide 9 did not purge through the work-up,
resulting in API (1) that failed to meet purity specifications.
While we were unable to identify crystallization conditions to
purge this impurity, an efficient chromatographic procedure was
developed to purify the batch to an appropriate standard; see the
Experimental Section for details thereof.
The key difference between this batch and previous, successful
batches was that the reaction mixture had been held at tempera-
ture (70 ꢀC) for an extended period (40 h, instead of the more
normal 2ꢀ3 h). In addition, at this stage of development the
reaction protocol had been to charge everything to the vessel
prior to heating, and therefore the slower heat-up time on larger
scale might have had an effect. This suggested that the product
(1) was degrading under the reaction conditions, possibly
through a Pd-mediated CꢀS bond cleavage, and the resulting
species were recombining to generate the symmetrical sulfides
9 and 10 (Scheme 3). The formation of symmetrical sulfide
byproducts has been previously observed in both Ni-17 and Pd-
catalyzed18 processes, and a mechanism involving Ni-mediated
ArꢀS cleavage of a diaryl thioether by a thiolate has been
proposed for the Ni-catalyzed process.17 In the case of the Pd-
catalyzed process, formation of the symmetrical byproducts was
suppressed by increasing the catalyst loading and conducting the
reaction at a higher temperature.18
’ EXPERIMENTAL SECTION
General Procedures. Intermediates were analyzed by reverse
phase LCꢀMS on an Agilent 1100 series instrument, coupled to
a Waters Micromass ZQ mass spectrometer according to the
following conditions: column Extend-C18 3.0 mm ꢁ 50 mm i.d.,
1.8 μm; eluent A, 5% v/v acetonitrile in 10 mM aqueous
ammonium acetate; eluent B, acetonitrile; flow rate 1.2 mL/min;
wavelength, diode array (190ꢀ400 μm); column temperature,
50 ꢀC; injection volume, 10 μL; at t = 0 min, 5% eluent B; at t =
3.5 min, 100% eluent B; at t = 4.5 min, 100% eluent B; at t =
4.6 min, 5% eluent B. 1H and 13C NMR spectra were recorded on
a Bruker Ultrashield 400 Plus spectrometer at 400 and 100.6
MHz, respectively. The quoted melting points for all materials
are the onset temperatures observed by DSC.
5-(4-Bromophenyl)-1-methyl-1H-pyrazole 4. A reactor was
charged with DMF (12 mL), 4-bromoacetophenone 2 (20 g,
0.1 mol) and N,N-dimethylformamide dimethylacetal (29 mL,
0.2 mol). The mixture was heated to 110 ꢀC for 4 h, with the
methanol and water generated during the reaction being re-
moved by distillation. The mixture was then cooled to 25 ꢀC,
MTBE (100 mL) and methylhydrazine (21 mL, 0.4 mol) were
added, and the mixture was stirred overnight at room temperature.
The reaction mixture was then washed with 1 M aqueous
ammonium chloride (3 ꢁ 40 mL) and water (40 mL). The
organic phase was dried by azeotropic distillation using a
DeanꢀStark apparatus. The solution was filtered through a silica
A purified sample of 1 (>99% purity) was subjected to the
reaction conditions (NaOMe, PdCl2(dppf), 2-MeTHF), and after
1049
dx.doi.org/10.1021/op200173g |Org. Process Res. Dev. 2011, 15, 1046–1051

Organic Process Research & Development
ARTICLE
gel cartridge (60 g), eluting with MTBE. The fractions contain-
ing product were combined and concentrated by distillation
(to approximately 70 mL). Heptane (120 mL) was added, and
distillation was continued until the pot temperature reached
98 ꢀC (approximately 100 mL of distillate was collected). The
mixture was cooled to 40 ꢀC and seeded, and the temperature
was maintained at 40 ꢀC for 30 min while crystallization was
initiated and then slowly chilled to 0 ꢀC over 90 min. The mixture
was held at 0 ꢀC for 30 min, and the mixture was then filtered.
The solid was washed with chilled (0 ꢀC) heptane (3ꢁ) and
dried on the filter to give the product 4 as a cream-colored,
was washed with 2-methyltetrahydrofuran (3 ꢁ 71.1 L). The
filter cake was then suspended in water (569 L), stirred at 23 ꢀC
for a minimum of 2 h, and filtered, and the solid was washed with
water (2 ꢁ 142 L) to afford crude 1 as an off-white solid (23.2 kg)
after drying at 70 ꢀC under vacuum.
A reactor was charged with tetrahydrofuran (812 L), crude 1
(11.6 kg, 29.5 mol) and SiliaBond thiol (Silicycle Inc., 5.8 kg), the
slurry was heated to 60 ꢀC, and the mixture was stirred at this
temperature until ICP-MS analysis of a sample confirmed
adequate Pd purge, a minimum of 18 h. The mixture was cooled
to 22 ꢀC and filtered, and the filter cake was washed with
tetrahydrofuran (58 L). The combined filtrate was distilled under
vacuum to a volume of 10 L/kg and then cooled to 25 ꢀC.
Heptane (58 L) was added to the slurry at this temperature, and
the mixture was aged for 30 min before the solid was isolated by
filtration, washed with heptane (58 L), and dried under vacuum
at 45 ꢀC to afford partially purified 1 as a white solid (11.08 kg,
95% yield). This process was repeated on the remaining crude 1.
A reactor was charged with methanol (441 L) and partially
purified 1 (22.06 kg). The slurry was heated to reflux for 18 h,
cooled to 20 ꢀC, and stirred at this temperature for 2 h. The
product was then isolated by filtration, washed with methanol
(110 L), and dried under vacuum at 70 ꢀC to afford the title
compound 1 as a white solid (19.96 kg, 70%). Mp 173 ꢀC. 1H
NMR (400 MHz, DMSO-d6) δ: 7.52 (2H, m), 7.48 (2H, m),
7.42 (2H, m), 7.35 (2H, m), 7.29 (2H, m), 7.07 (1H, br. s), 6.42
(1H, d, J = 1.8 Hz), 3.85 (3H, s), 3.74 (2H, dt, J = 11.7, 3.7 Hz),
3.47 (2H, br. t, J = 11.7 Hz), 2.41 (2H, br. d, J = 13.3 Hz), 1.80
(2H, m). 13C NMR (100.6 MHz, DMSO-d6) δ: 174.6, 146.0,
141.9, 137.9, 136.0, 133.2, 130.1, 129.7, 129.4, 129.3, 128.6,
125.6, 105.9, 64.6, 47.8, 37.6, 33.9. LCꢀMS: found m/z 394.17
[M + H]+. Anal. Calcd for C22H23N3O2S: C, 67.15; H, 5.89; N,
10.68; S, 8.15. Found: C, 67.09; H, 5.93; N, 10.69; S, 8.16.
Optional Procedure for Chromatographic Purification of
1. Silica gel (J.T. Baker flash silica, 40ꢀ60 μm; 5 kg/kg crude 1)
was slurried in toluene (2.1 kg/kg silica), and the resulting slurry
was loaded onto a suitable column. The column was equilibrated
with toluene (4.6 kg/kg silica) at a flow rate of 4 L/min. A
solution of 1 in a mixture of CH2Cl2/EtOH (97:3; 22.6 kg/kg
crude 1) was loaded onto the equilibrated column at a rate of
3 L/min; the feed line was rinsed with additional CH2Cl2/EtOH
(97:3; 1 kg/kg crude 1). The product was then eluted from the
column with a toluene/i-PrOH mixture (10:1 mixture; 4 L/min
flow rate). Fractions containing 1 were combined and concen-
trated under reduced pressure to afford crude 1, which was taken
directly into the next step (Si-thiol purification); see previous
experimental procedure for details thereof.
1
crystalline solid (16.3 g, 68% yield). Mp 58 ꢀC. H NMR
(400 MHz, DMSO-d6) δ: 7.68 (2H, dt, J = 8.4, 1.8 Hz), 7.49
(3H, m), 6.43 (1H, d, J = 2.0 Hz), 3.86 (3H, s). 13C NMR (100.6
MHz, DMSO-d6) δ: 141.5, 137.7, 131.4, 130.6, 129.6, 121.8,
106.0, 37.5. LCꢀMS: found m/z 237.05/235.02 [M + H]+. Anal.
Calcd for C10H9BrN2: C, 50.66; H, 3.83; Br, 33.70; N, 11.82.
Found: C, 50.62; H, 3.78; Br, 33.94; N, 11.77.
4-(3-(Triisopropylsilylthio)phenyl)-tetrahydro-2H-pyran-
4-carboxamide 8. A nitrogen-purged reactor was charged with
toluene (349.6 L), amide 6 (26 kg, 91.5 mol) and sodium tert-
butoxide (10.55 kg, 109.8 mol), followed by tri-isopropylsila-
nethiol 7 (20.91 kg, 109.8 mol) and a toluene wash (10.4 L). The
mixture was heated to 75 ꢀC, [1,10-bis(diphenylphosphino)-
ferrocene]dichloropalladium(II) (0.67 kg, 0.92 mol) was added
under nitrogen, and the mixture was stirred at reflux until
reaction completion was confirmed by HPLC analysis (approx
2 h). The reaction mixture was cooled to 22 ꢀC, ethyl acetate
(780 L) was added, and the mixture was stirred at this tempera-
ture for at least 30 min. The slurry was filtered through Celite
(39kg), and theCelite cake was washed withethyl acetate (260 L).
The combined filtrates were distilled under vacuum to one-fifth
of the volume, heptane (520 L) was added to the slurry, and the
mixture was cooled to 5 ꢀC and stirred at this temperature for at
least 1 h. The product was isolated by filtration, and the cake was
washed with heptane (260 L), to afford the title compound 8 as a
white solid (28.45 kg, 79% yield) after drying at 50 ꢀC under
1
vacuum. Mp 141 ꢀC. H NMR (400 MHz, DMSO-d6) δ:
7.45ꢀ7.23 (4H, m), 7.05 (1H, m), 5.10 (1H, s), 3.72 (2H, m),
3.47 (2H, t, J = 10.8 Hz), 2.40 (2H, br, d, J = 13.3 Hz), 1.77 (2H,
m), 1.04ꢀ0.87 (21H, m). 13C NMR (100.6 MHz, DMSO-d6) δ:
174.5, 145.7, 135.9, 129.4, 125.8, 125.2, 125.0, 64.6, 47.9, 33.9,
17.8, 12.1. LCꢀMS: found m/z 394.24 [M + H]+. Anal. Calcd for
C21H35NO2SSi: C, 64.07; H, 8.96; N, 3.56; S, 8.15. Found: C,
64.13; H, 8.97; N, 3.55; S, 8.20.
4-(3-([4-(1-Methyl-1H-pyrazol-5-yl)phenyl]thio]phenyl)-
tetrahydro-2H-pyran-4-carboxamide 1. A reactor was charged
with 2-methyltetrahydrofuran (256 L), S-TIPS amide 8 (28.45 kg,
72.3 mol) and pyrazole 4 (18 kg, 75.9 mol), the reactor was
evacuated and repressurized with nitrogen, and the mixture was
heated to 60 ꢀC. Sodium methoxide in methanol (25 wt %; 32.8 kg,
151.8 mol) was added, the reactor was evacuated and repressur-
ized with nitrogen and stirred at 60 ꢀC for a minimum of 30 min.
The reaction mixture was then heated to 70 ꢀC, and [1,10-bis-
(diphenylphosphino)ferrocene] dichloropalladium(II) (0.530 kg,
0.73 mol) was added; the reaction mixture was stirred at this
temperature for 1 h, whereupon a further charge of [1,10-bis-
(diphenylphosphino)ferrocene] dichloropalladium(II) (0.530 kg,
0.73 mol) was added, and the mixture was stirred at 70 ꢀC until
reaction completion was confirmed by HPLC analysis (approx 1 h).
The mixture was cooled to 0 ꢀC and stirred at this temperature
for a minimum of 1 h; the slurry was filtered, and the filter cake
The major impurity removed through this chromatographic
procedure is sulfide 9. 1H NMR (400 MHz, DMSO-d6) δ: 7.35
(6H, m), 7.26 (2H, br. s), 7.16 (2H, dt, J = 7.0, 2.0 Hz), 7.04 (2H,
br. s), 3.73 (4H, dt, J = 11.7, 3.5 Hz), 3.46 (4H, t, J = 10.6 Hz),
2.38 (4H, br. d, J = 13.5 Hz), 1.77 (4H, ddd, J = 13.5, 10.6, 3.5 Hz).
13C NMR (100.6 MHz, DMSO-d6) δ: 174.6, 145.7, 134.5, 129.5,
128.5, 127.8, 124.9, 64.6, 47.8, 33.9. LCꢀMS: found m/z 441.12
[M + H]+. Anal. Calcd for C24H28N2O4S: C, 65.43; H, 6.41; N,
6.36; S, 7.28. Found: C, 65.01; H, 6.37; N, 6.41; S, 7.30.
’ AUTHOR INFORMATION
Corresponding Authors
*E-mail: pieter.de.koning@pfizer.com; jplawsonconsulting@
gmail.com.
1050
dx.doi.org/10.1021/op200173g |Org. Process Res. Dev. 2011, 15, 1046–1051

Organic Process Research & Development
ARTICLE
Present Addresses
†J. P. Lawson Consulting LLC, 18804 Melrose Rd., Wildwood,
MO 63038.
‡Novus International, 20 Research Park Dr., St. Charles,
MO 63304.
§1302 Still House Creek Rd., Chesterfield, MO 63017.
^Covidien, 385 Marshall Ave., Webster Groves, MO 63119.
’ ACKNOWLEDGMENT
We thank James Murphy for developing large-scale chromato-
graphic procedures; Dave Clifford and Marie Stephan for their
expertise in transferring the process to large-scale facilities;
Watcharee Cooper, Dan Basford, Mike Rance, David Edwards,
and Zakariya Hasson for analytical support; and all the shift
operators and analysts for their efforts.
’ REFERENCES
(1) (a) Werz, O; Steinhilber, D. Pharmacol. Ther. 2006, 112, 701.
(b) Rubin, P; Mollison, K. W. Prostaglandins Other Lipid Mediators 2007,
83, 188. (c) Rastogi, P; McHowat, J. Lett. Drug Des. Discovery 2006, 3.
(2) Gilbert, N. C; Barlett, S. G.; Waight, M. T; Neau, D. B; Boeglin,
W. E; Brash, A. R; Newcomer, M. E. Science 2011, 331, 217.
(3) Evans, J. F; Ferguson, A. D; Mosley, R. T.; Hutchinson, J. H.
Trends Pharmacol. Sci. 2008, 29, 72.
(4) Back, M. Curr. Pharm. Des. 2009, 15, 3116.
(5) Cortes-Burgos, L. A; Zweifel, B. S; Settle, S. L; Pufahl, R. A;
Anderson, G. D; Hardy, M. M; Weir, D. E; Hu, G; Happa, F. A; Stewart,
Z; Multian, S; Granetto, M. J; Masferrer, J. L. Eur. J. Pharmacol. 2009,
617, 59.
(6) Poff, C. D; Balazy., M. Curr. Drug Targets Inflammation Allergy
2004, 3, 19.
(7) Graneto, M. J.; Maddux, T. M.; Masferrer, J. L. U.S. Patent
application 2008/0125474 A1.
(8) Fernꢀandez-Rodríguez, M. A.; Shen, Q.; Hartwig, J. F. J. Am.
Chem. Soc. 2006, 128, 2180.
(9) The use of TIPS-thiol (MW 190.42) to add one sulfur atom
(MW 32.06) is clearly an inefficient approach.
(10) During early development, the primary focus was to deliver a
“fit-for-purpose” process to deliver material for clinical studies. Subse-
quent work to identify a more suitable long-term process will be the
subject of a future publication.
(11) We did not attempt to identify any byproducts removed in the
work-up.
(12) See Robinson, D. I. Org. Process Res. Dev. 2010, 14, 946 for a
more detailed overview of the staged TTC approach and how it is
applied during different phases of development.
(13) (a) Norris, T.; Leeman, K. Org. Process Res. Dev. 2008, 12, 869.
(b) Norris, T.; Hnatow, M. E. Eur. Pat. Appl., 2001, EP 1081144 A2.
(14) (a) Migita, T.; Shimizu, T.; Asami, Y.; Shiobara, J; Kato, Y.;
Kosugi, M. Bull. Chem. Soc. Jpn. 1980, 53, 1385. (b) Kosugi, M.; Shimizu,
T.; Migita, T. Chem. Lett. 1978, 13.
(15) Murata, M.; Buchwald, S. L. Tetrahedron 2004, 60, 7397.
(16) See Okauchi, T.; Kuramoto, K.; Kitamura, M. Synlett 2010,
2891 for a recent example of a Migita coupling using PdCl2(dppf).
(17) Takagi, K. Chem. Lett. 1987, 2221.
(18) (a) Fernꢀandez-Rodríguez, M. A.; Shen, Q.; Hartwig, J. F. Chem.
—Eur. J. 2006, 12, 7782. (b) Fernꢀandez-Rodríguez, M. A.; Hartwig, J. F.
J. Org. Chem. 2009, 74, 1663.
1051
dx.doi.org/10.1021/op200173g |Org. Process Res. Dev. 2011, 15, 1046–1051