Development of a Practical Synthesis of Functionalized Azaxanthene-Derived Nonsteroidal Glucocorticoid Receptor Modulators

Article abstract of DOI:10.1021/acs.oprd.6b00035

An efficient route to two functionalized 2-aryl-5H-chromeno[2,3-b]pyridines (azaxanthenes) is reported. The addition of lithiated 2,6-dichloropyridine to salicylaldehyde followed by cyclization was a key process improvement identified for the formation of the azaxanthene core. Further elaboration of 2-chloro-5H-chromeno[2,3-b]pyridin-5-ol at the 5 position was accomplished via Lewis acid-catalyzed coupling with commercially available ((1-methoxy-2-methylprop-1-en-1-yl)oxy)trimethylsilane. A partial classical resolution coupled with a preparative chiral supercritical fluid chromatography (SFC) separation was used to isolate the desired enantiomer of the azaxanthene carboxylic acid that is a common intermediate for both compounds 1 and 2. Suzuki-Miyaura cross-coupling with appropriately substituted boronic acids, followed by condensation with 2-amino-1,3,4-thiadiazole, provided the target compounds with an overall yield of approximately 10%. The use of stable, amorphous materials to support clinical comparison of functionalized azaxanthenes 1 and 2 is also discussed.

Full text of DOI:10.1021/acs.oprd.6b00035

Article
pubs.acs.org/OPRD
Development of a Practical Synthesis of Functionalized
Azaxanthene-Derived Nonsteroidal Glucocorticoid Receptor
Modulators
David A. Conlon,*,† Kenneth J. Natalie, Jr.,† Nicolas Cuniere,† Thomas M. Razler,† Jason Zhu,†
Nuria de Mas,†,∥ Steven Tymonko,† Kenneth J. Fraunhoffer,† Eric Sortore,†,⊥ Victor W. Rosso,†
Zhongmin Xu,† Monica L. Adams,§ Anisha Patel,§ Jun Huang,§ Hua Gong,‡ David S. Weinstein,‡
Fernando Quiroz,† and Doris C. Chen†
§
†Chemical & Synthetic Development and Drug Product Science and Technology, Bristol-Myers Squibb, One Squibb Drive, New
Brunswick, New Jersey 08903-0191, United States
‡Research and Development, Bristol-Myers Squibb, Princeton, New Jersey 08543-4000, United States
ABSTRACT: An efficient route to two functionalized 2-aryl-5H-chromeno[2,3-b]pyridines (azaxanthenes) is reported. The
addition of lithiated 2,6-dichloropyridine to salicylaldehyde followed by cyclization was a key process improvement identified for
the formation of the azaxanthene core. Further elaboration of 2-chloro-5H-chromeno[2,3-b]pyridin-5-ol at the 5 position was
accomplished via Lewis acid-catalyzed coupling with commercially available ((1-methoxy-2-methylprop-1-en-1-yl)oxy)-
trimethylsilane. A partial classical resolution coupled with a preparative chiral supercritical fluid chromatography (SFC)
separation was used to isolate the desired enantiomer of the azaxanthene carboxylic acid that is a common intermediate for both
compounds 1 and 2. Suzuki−Miyaura cross-coupling with appropriately substituted boronic acids, followed by condensation with
2-amino-1,3,4-thiadiazole, provided the target compounds with an overall yield of approximately 10%. The use of stable,
amorphous materials to support clinical comparison of functionalized azaxanthenes 1 and 2 is also discussed.
secondary alcohol 6. Installation of the isobutyrate fragment
was readily achieved by treatment of the highly electrophilic,
doubly benzylic alcohol 6 with ((1-methoxy-2-methylprop-1-
en-1-yl)oxy)trimethylsilane (25) under Lewis acid-promoted
Mukaiyama conditions to provide methyl ester 7 as a racemic
mixture.6 Treatment of 7 with m-CPBA followed by exposure
to phosphorus oxychloride effected a regioselective chlorination
to provide chloroazaxanthene 9, which was saponified to
provide a racemic mixture of carboxylic acids 10a and 10b in
high yield. The mixture of carboxylic acids was separated by
preparative chiral supercritical fluid chromatography (SFC) to
afford the homochiral enantiomeric acid 10a. An X-ray crystal
structure of the (R)-(+)-α-methylbenzylamine salt of 10a was
used to determine that the desired acid had the (R) absolute
configuration. Condensation of 10a with 2-amino-1,3,4-
thiadizaole gave chloride 11 in 99% yield, which was then
subjected to a Suzuki−Miyaura cross coupling with boronic
acid 13 to give penultimate 12 in 95% crude yield.7 Subsequent
condensation with dimethylamine or ethylmethylamine pro-
vided corresponding amides 1 and 2, respectively. For
removing impurities present in both dimethylamine and
ethylmethylamine, a second SFC purification was required.
Although the process described in Scheme 1 was suitable for
the preparation of multigram quantities of 1 and 2, further
scale-up to supply material for toxicological and clinical
evaluation presented several challenges. The linear nature of
this synthesis and our accelerated timeline prompted us to
INTRODUCTION
■
Agonists of the glucocorticoid receptor (GR), such as
prednisolone, have potent anti-inflammatory and immunosup-
pressive properties and have found broad utility in the
treatment of autoimmune, allergic, and anti-inflammatory
diseases.1 However, the unparalleled efficacy of steroidal
glucocorticoids (GC) is countered by serious side effects such
as glucose intolerance, muscle wasting, skin thinning, and
osteoporosis.2 Our efforts to identify selective nonsteroidal
glucocorticoid receptor modulators that possess anti-inflamma-
tory activity similar to GCs with reduced side-effects
culminated in the identification of two potent and selective
compounds (compounds 1 and 2, Scheme 1) with overall in
vitro biological profiles that are clearly distinct from those of
steroidal GCs and from each other. Although both in vitro and
in vivo models have been used to identify anti-inflammatory
activity, preclinical animal models are not good predictors of
GC side effects in humans; therefore, we needed to rapidly
prepare azaxanthene derivatives 1 and 2 for clinical evaluation.3
The results of our development activities that led to the
successful preparation of both functionalized azaxanthenes are
described.
Original and Revised Routes to 1 and 2. The initial
approach for the preparation of the 2-aryl-5H-chromeno[2,3-
b]pyridines (azaxanthenes) is shown in Scheme 1.4 The
procedure developed by Villani,5 which coupled 2-chloronico-
tinic acid with sodium phenoxide in an excess of the phenol at
180 °C followed by an intramolecular Friedel−Crafts reaction,
was used to prepare 1-azaxanthone 5 in 47% yield over the two
steps. Reduction of 5 with sodium borohydride generated
Received: February 5, 2016
© XXXX American Chemical Society
A
DOI: 10.1021/acs.oprd.6b00035
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
Scheme 1. Original Route to 1 and 2
Scheme 2. Revised Route to 1 and 2
pursue a more convergent approach for the scale-up.
Strategically, we sought to effect a late stage coupling of an
enantiopure azaxanthene core already suitably functionalized at
the 2-position, utilizing the fully elaborated aromatic amides as
their boronic acid derivatives 18 and 19 (Scheme 2). This
approach would provide late stage intermediates 20 and 21.
These intermediates could be readily transformed to clinical
candidates 1 and 2 by amide bond formation with 2-amino-
1,3,4-thiadiazole.
We selected 2,6-dichloropyridine (14) as our starting
material because it is an inexpensive commodity chemical.
We were able to demonstrate that reaction of the 3-lithio
species generated from 2,6-dichloropyridine with salicylalde-
hyde followed by cyclization under basic conditions provided a
rapid entry to azaxanthene 17 and furnished the requisite
B
DOI: 10.1021/acs.oprd.6b00035
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
Scheme 3. Formation of the Azaxanthol Core (17)
functionality at C-2 of the azaxanthene ring system for further
elaboration. This is the most expeditious route to the
azaxanthene core that we have identified.8 In addition, this
approach circumvented the harsh conditions inherent in the
reaction sequence reported by Villani that were utilized to
generate the azaxanthene core in the original route and which
were not amenable to scale-up. Performing the Suzuki−
Miyaura cross-coupling on this C-2 functionalized azaxanthene
17 prior to incorporation of 2-amino-1,3,4-thiadiazole
precluded the formation of impurities originating from
degradation of the thiadiazole moiety during the high
temperature (90 °C, 12 h) required for successful Suzuki−
Miyaura cross-coupling. In the original route, the late-stage
amidation with dimethylamine and ethylmethylamine neces-
sitated a second SFC purification of the API to remove
impurities arising from amine contaminants in these com-
pounds. Suzuki−Miyaura cross-coupling using 18 and 19 with
the dialkyl amide moieties already incorporated was envisaged
to provide greater control of impurities associated with
dimethylamine and ethylmethylamine at the arylboronic acid
stage without chromatography.
Synthesis of the Azaxanthene Core (17). Radinov et al.
reported that the lithiation of 2,6-dichloropyridine with LDA
gave the 3-substituted pyridine as the major product (90:10)
when chlorotrimethylsilane (TMSCl) was used as the electro-
phile; however, quenching with benzaldehyde was reported to
provide only a 2:1 mixture, still favoring the 3-substituted
product over the 4-substituted pyridine.9a Extending the time
for the lithiation prior to the addition of benzaldehyde resulted
in a significant increase in the selectivity for the 3-substituted
product up to 98:2. This result was purported to be a result of
an equilibration to the thermodynamically more stable 3-
lithiopyridine.9 Treatment of 2,6-dichloropyridine with an
excess of lithium diisopropylamide (LDA) at −60 °C followed
by the addition of salicylaldehyde provided a 4−5:1 mixture of
regioisomers 16 and 24 (Scheme 3). Surprisingly, all efforts to
equilibrate the lithiated species (22 and 23), including
prolonged holds and substoichiometric base charges, failed to
influence the product ratio. For determining if the modest
selectivity was due to an elevated reactivity of the 4-
lithiopyridine toward salicylaldehyde, TMSCl trapping experi-
ments were conducted. Trapping aliquots of the lithiopyridine
solution with TMSCl consistently gave the identical 5:1 ratio of
products favoring the 3-silylpyridine over the 4-silylpyridine
regardless of the equilibration time. This is in contrast to
previously reported experiments that gave a 90:10 mixture of
the 3-trimethylsilyl derivative.9 Likewise, trapping the lithiopyr-
idines with a substoichiometric amount of TMSCl gave a 5:1
ratio, indicating similar reactivity between TMSCl and
salicylaldehyde with the lithiopyridines.
Despite the modest regioselectivity of the alkylation, the
crude mixture of 16 and 24 could be treated with potassium
tert-butoxide at elevated temperature (65 °C) to successfully
effect the cyclization of 16 to azaxanthol 17 (Scheme 3). The
undesired byproduct 24, along with any remaining salicylalde-
hyde and 2,6-dichloropyridine, were removed during crystal-
lization of 17 from a mixture of ethyl acetate/n-heptane to
provide azaxanthol 17 in 56% yield over two telescope steps on
a laboratory scale.
Upon initial scale-up to a 20 L vessel, a dramatic drop in
reaction conversion to 16 was observed resulting in only a 19%
isolated yield of 17. Subsequent laboratory experiments
identified the addition rate of the salicylaldehyde as a critical
parameter for achieving high conversion. When salicylaldehyde
was added in a single, rapid charge, >90% conversion was
observed, whereas slower additions resulted in reaction stalling.
The likely cause of the observed stalling is competitive attack
on the aldehyde functionality in salicylaldehyde by LDA
resulting in the formation of a hemiaminal that masks the
salicylaldehyde under the reaction conditions. In fact, secondary
amines have been previously utilized in this fashion as in situ
aldehyde protecting groups during organolithium reactions.10
For testing this possibility, an inverse addition experiment was
conducted, where salicylaldehyde was incubated with two
equivalents of LDA followed by the addition of 2,6-
dichloropyridine. As anticipated, no reaction was observed,
consistent with the formation of the hemiaminal mask. Upon
treatment with TMSCl, no silylated pyridine species were
detected, confirming that both equivalents of LDA were
consumed by salicylaldehyde.
The preparative formation of 16 was carried out by charging
the salicylaldehyde at the maximum possible rate while
C
DOI: 10.1021/acs.oprd.6b00035
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
Scheme 4. Formation of Carboxylic Acids 10a and 10b
Scheme 5. Preparation of Suzuki Coupling Partners 18 and 19
maintaining a batch temperature below −40 °C to prevent
degradation of the pyridyllithium species. Utilizing this
procedure, typical reaction conversions of 80% were achieved
(0.5 kg scale), resulting in a 37−42% isolated yield of 17.
Despite the modest yield, this is an expeditious route to the
azaxanthene core. Indeed, we found that this approach provides
a general process for the rapid preparation of a variety of
substituted azaxanthene analogues.
Because of the exothermic nature of the reaction, any
additional scale-up (>0.5 kg) of the reaction in a batch mode
would require an extended salicylaldehyde charge to maintain
the required low reaction temperature, resulting in further
decreases in yield. To circumvent this issue, we considered
investigating a flow process to allow for rapid mixing at the
desired low temperature; however, an accelerated timeline
precluded our investigation of this approach. Therefore, we
limited the scale of the reaction and performed several batches
to secure the required amount of 17.
Having secured access to kilogram quantities of the
azaxanthene core through multiple batches, the requisite 2-
carbon homologation of 17 at the C-5 benzylic position was
accomplished in an analogous manner to the original route by a
Lewis acid-mediated alkylation with [(1-methoxy-2-methyl-
prop-1-ene-1-yl)oxy]trimethylsilane 25 (Scheme 4).11
A solubility screen suggested that dichloromethane was a
suitable solvent for the reaction of 17 with 25, and a screen of
Lewis acids determined that titanium tetrachloride delivered
the highest yield. Not surprisingly, residual water and
temperature were identified as critical parameters for this
reaction. At Karl Fischer (KF) values greater than 1.5% and
temperatures above 5 °C, we observed reaction stalling, and
kicker charges of 23 and titanium tetrachloride were required to
achieve the desired reaction end point. Dichloromethane
streams of 17 were distilled to azeotropically remove water to
a KF value below 500 ppm, which reproducibly provided
quantitative conversion to 9 upon treatment with 25 and
TiCl4.12 The removal of titanium byproducts proved to be
problematic as the direct aqueous quench of TiCl4 in the
reaction mixture formed a difficult-to-filter titanium hydrogel.13
After multiple attempts to prevent hydrogel formation, we
found that the titanium salt could be easily removed by polish
filtration after treatment of the reaction stream with wet ethyl
acetate at controlled pH (see Experimental Section), which
resulted in the precipitation of the titanium byproducts as
filterable solids. Crystallization from ethyl acetate provided 9 in
75−80% isolated yield. Ester hydrolysis was readily accom-
plished with potassium hydroxide in NMP, and the mixture of
10a and 10b was isolated directly from the crude reaction
stream upon addition of water and pH adjustment in 91−93%
isolated yield. We subsequently determined that a stream of 9
in ethyl acetate could be taken directly into the hydrolysis step
to provide 10a and 10b in 88% yield after solvent exchange into
NMP; however, this two-step telescope was not developed in
time to implement on scale.
Resolution of Carboxylic Acids 10a and 10b. In the
original route, carboxylic acid 10a was isolated from a mixture
of 10a and 10b by chiral preparatory SFC. Because of our
accelerated timeline and the amount of 10a required to support
multiple deliveries of both 1 and 2, we explored the option of
using a classical resolution for the purification of 10a.
Unfortunately, the only crystalline salt prepared from
cinchonidine14 was enriched with the undesired enantiomer
10b. The filtrate that was enriched in 10a was subjected to a
salt break and used as the feed for the chiral preparatory SFC.
Coupling of our optimized partial classical resolution with a
chiral SFC process using the 85:15 mixture of 10a:10b as
feedstock allowed the production of enough material to support
the initial development of both 1 and 2, and this work is
described in an accompanying paper.15
D
DOI: 10.1021/acs.oprd.6b00035
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
Suzuki−Miyaura Cross-Coupling. As previously men-
tioned, the original process utilized a Suzuki−Miyuara cross-
coupling of 11 and boronic acid 13 to provide compound 12
(Scheme 1). In the final step of the synthesis, compound 12
was reacted with either dimethylamine or ethylmethylamine to
provide 1 and 2, respectively. This synthetic sequence provided
flexibility in the preparation of different analogues and was
sufficient for the initial deliveries; however, the use of this
chemistry on a preparative scale was precluded by the following
observations: incorporation of the 2-amino-1,3,4-thiadiazole
prior to the Suzuki−Miyaura cross-coupling resulted in multiple
impurities arising from decomposition of the thiadiazole moiety
under the reaction conditions. In addition, multiple difficult to
reject impurities were formed in the final step from
contaminants in the corresponding amines used to form 1
and 2 that required a second SFC purification. As a result of
these issues, we investigated an alternative route that changed
the order of the reaction sequence. In an attempt to minimize
the formation of the impurities resulting from contaminants in
dimethylamine and ethylmethylamine, we turned our attention
to developing processes to prepare the requisite nucleophilic
Suzuki−Miyaura coupling partners 18 and 19 (Scheme 5). A
three-step telescoped process starting from commercially
available 4-bromo-2-fluorobenzoic acid 26 was developed.16
Coupling of the active ester of 26 with either dimethylamine or
ethylmethylamine gave the corresponding amides 27 and 28.
These were not isolated but rather converted directly to the
boronate esters 29 and 30 using B2(Pin)2. We were unable to
isolate the boronic esters due to the formation of an
uncontrollable mixture that also contained the boronic acid.
Therefore, acidic sodium periodate was used to convert free
pinacol to acetone, eliminating the boronate ester and boronic
acid equilibrium and providing exclusively arylboronic acids 18
and 19.
as the ligand in conjunction with monobasic potassium
phosphate. No conversion of 10a was observed, but rather,
arylboronic acid 18 was fully consumed to deliver dimeric
impurity 27 with no detectable cross-coupled product. In
contrast, when tricyclohexylphosphonium tetrafluoroborate was
employed with more basic potassium carbonate and tetrabu-
tylammonium bromide (TBABr), 99 LCAP (HPLC area
percent) of cross-coupled product 20 was observed with no
detectable 31.18 This remarkable flip in product selectivity may
be explained by the coordinative ligand saturation of the active
catalyst.19 The formation of the coordinatively unsaturated and
highly reactive PdL species is more likely active when bulky
ligands, such as PCy3, are employed, thus promoting the
oxidative addition of challenging substrates, such as the 2-
chloroazaxanthene system. With bidentate dppe, however,
oxidative addition into the carbon−Cl bond is slow when
compared to the homocoupling of compound 18 due to the
presence of a saturated PdL2 catalyst system.20 Interestingly,
when TBABr was omitted as a reagent, up to 5 LCAP of 10a
remained after 24 h of reaction.
With Suzuki−Miyaura cross-coupling conditions in hand, we
turned our attention to palladium remediation to less than 20
ppm at the penultimate stage so as to not burden the API
crystallization with palladium removal. We were unable to
achieve our target levels for palladium by treatment of MeTHF
streams containing 20 with carbon- or sulfur-based adsorbents,
such as thiourea or thiol-based scavengers, and the isolated
solid contained 100−200 ppm of palladium. We then turned
our attention to palladium removal via crystallization. We
hypothesized that the carboxylic acid of 20 would allow for
facile salt formation, and thus present an opportunity for a
crystallative palladium removal strategy. We quickly found that
the dicyclohexylamine (DCHA) salt of 20 was a highly
crystalline solid that, when isolated, reduced palladium levels to
200 ppm. Liberation of 20 from the DCHA salt with 3 N HCl,
followed by crystallization of the free carboxylic acid from n-
BuOAc, provided a white, fluffy solid with >99.7 LCAP purity
and <10 ppm of palladium.
The cross-coupling of highly functionalized 2-chloroazax-
anthene 10a and electron-deficient arylboronic acids 18 and 19,
two traditionally poor coupling partners, proved to be
challenging (Scheme 6).17 The initial Suzuki−Miyaura cross-
On a 0.5 kg scale reaction, the Suzuki−Miyaura cross-
coupling of 2-chloroazaxanthene 10a and [3-fluoro-4-
(dimethylaminocarbonyl)phenyl]boronic acid 18 delivered
enantiomerically pure cross-coupled product 20 with >99.9
LCAP and <10 ppm of residual palladium in 83% yield.
Similarly, the cross-coupling employing [3-fluoro-4-
(ethylmethylaminocarbonyl)phenyl]boronic acid provided
enantiomerically pure cross-coupled product 21 in 84% yield
and >99.9 LCAP with <20 ppm of residual palladium (Scheme
2). Interestingly, both 20 and 21 were found to contain up to
0.3 wt % of residual dicyclohexylamine due to partitioning of
the DCHA HCl salt byproduct into the organic stream during
the palladium removal, which was then carried into the API
step. For meeting purity specifications, DCHA was removed by
washing MeTHF solutions of 1 or 2 (vide infra) with 10%
aqueous citric acid, which effectively removed residual DCHA
to undetectable levels by LC/MS/MS.21
Preparation of 1 and 2. With high-quality penultimate
intermediates 20 and 21 in hand, we turned our attention to
the coupling with 2-amino-1,3,4-thiadiazole to complete the
synthesis of 1 and 2. Initial optimization studies of the coupling
of carboxylic acid 20 with 2-amino-1,3,4-thiadiazole to produce
1 centered on the conditions utilized in the original route to
prepare 11. Screening quickly identified 1-hydroxy-7-azabenzo-
triazole (HOAt) as a superior coupling reagent providing good
Scheme 6. Suzuki−Miyaura Cross-Coupling to Prepare
Amides 20 and 21
coupling of 18 or 19 with 10a led to low conversion of the 2-
chloroazaxanthene fragment or homocoupling of the arylbor-
onic acids under standard Suzuki−Miyaura coupling conditions.
A high-throughput screening approach was employed to
rapidly evaluate ligands, bases, solvents, and additives such as
inorganic salts and phase transfer catalysts. After screening 288
different reaction conditions, the cross-coupling of 10a and 3-
fluoro-4-N,N-dimethylbenzoylphenylboronic acid 18 was
shown to be very sensitive to the choice of ligand and base
(Scheme 7). The most dramatic example included using dppe
E
DOI: 10.1021/acs.oprd.6b00035
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
Scheme 7. High-Throughput Suzuki−Miyaura Screening
Scheme 8. Preparation of 1 and 2
conversion in a variety of solvents, such as dimethyformamide
and tetrahydrofuran. Other activators examined (HOBt and
HOPO) provided slow or incomplete reactions. 2-Methylte-
trahydrofuran was selected for further development to simplify
the workup procedure. In addition, we found that two full
equivalents of 2-amino-1,3,4-thiadiazole at 55 °C in MeTHF
were required to effect complete conversion of the intermediate
active esters to the desired target compounds 1 and 2 (Scheme
8). Treatment of 20 with HOAt, 1-(3-(dimethylamino)propyl)-
3-ethylcarbodiimide hydrochloride, and diisopropylethylamine
in MeTHF followed by reaction of the intermediate active ester
with 2-amino-1,3,4-thiadiazole furnished 1 as a solution in
MeTHF after aqueous workup. After solvent exchange, 1 was
crystallized from isopropyl acetate in 76% isolated yield.
In a similar manner, 2 was prepared by condensation of 21
with 2-amino-1,3,4-thiadiazole. After an aqueous workup, the
rich organic stream was solvent exchanged into isopropyl
acetate, and 2 was isolated as its crystalline isopropyl acetate
solvate in 80% yield. It is worth mentioning that identifying
crystallization conditions for 2 required additional studies and
that, unlike 1, compound 2 was isolated as an isopropyl acetate
solvate.
arises from the reaction of 2-amino-1,3,4-thiadiazole with
residual 10a in 20 or 21. Subsequent fate and tolerance studies
demonstrated that up to 0.2 LCAP of 11 could be purged by
crystallization from isopropyl acetate. The improved Suzuki−
Miyaura coupling conditions in the penultimate step, as well as
optimization of the crystallization of 20 and 21, provided
material with <0.2 LCAP of 10a on a preparative scale.
The solid-state form of the drug substance is an important
specification that impacts the stability, dissolution, and
bioavailability of the drug product. Crystallization afforded an
avenue for purification; however, both 1 and 2 had previously
been isolated as amorphous solids, and the amorphous solids
were used in the exploratory toxicological studies. Focus then
turned to the selection of the appropriate final form for both 1
and 2.
Final API Form and Formulation Development. As
previously mentioned during the original preparations of 1 and
2, both compounds were isolated as amorphous solids. In
addition, initially neither compound readily crystallized during
preparation nor was form conversion evident under conditions
expected during subsequent drug product processing and
manufacturing. Interestingly, the amorphous forms of both 1
and 2 were found to be chemically and physically stable (under
accelerated stability conditions of heat and humidity) as bulk
powders under the conditions evaluated, suggesting that the
amorphous materials possessed adequate stability for further
development. Given that amorphous phases are generally less
stable than crystalline forms, the potential risks associated with
the use of amorphous materials to support the clinical
comparison of 1 and 2 were assessed.
Crystallization of 1 and 2 was found to effectively purge the
majority of the known impurities except 11 (Scheme 9), which
Scheme 9. By-product Formation During the API Amidation
Our development efforts focused on assessment of the
suitability of these amorphous forms for use in phase-
appropriate formulations that could be prepared at a local
compounding site. As shown in Figure 1, the amorphous forms
F
DOI: 10.1021/acs.oprd.6b00035
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
the propensity of both molecules toward crystallization under
the expected clinical dosing conditions remained unknown.
Development of Solution and Suspension Formula-
tions for Clinical Dosing. For enabling rapid progression into
clinical trials, a drug-in-bottle (DIB) approach was selected to
support clinical administration as either a solution or
suspension across a dose range of 5−1200 mg. The equilibrium
aqueous solubility of amorphous 1 and 2 at 24 3 °C was
measured at 0.012 mg/mL (pH 5.6) and 0.005 mg/mL (pH
4.9), respectively, by quantitative reverse-phase HPLC analysis.
The addition of a surfactant increased the apparent aqueous
solubility of amorphous 1 from 0.012 to 0.092 mg/mL and that
of 2 from 0.005 to 0.075 mg/mL. For assessing the suitability of
the amorphous materials for use in suspension formulations,
the physical stability of aqueous suspensions with and without
0.1% w/v polysorbate 80 (added to aid wetting) were
monitored for 24 h using capillary PXRD. Although the
apparent aqueous solubility of both compounds increased by an
order of magnitude in the presence of surfactants, crystal-
lization did not occur over the duration of the studies (24 h).
Regardless of the exact mechanism of physical stability, it was
determined that both compounds exhibited suitable physical
stability as amorphous solids under the anticipated storage and
handling conditions that would be used in the clinical setting.
Suspension vehicles evaluated for clinical dosing included
citric acid anhydrous USP (25 mM) and Simple Syrup, NF
(25% v/v, to improve palatability), and additional screening
studies were then conducted to determine whether the method
used to prepare the amorphous materials (i.e., spray-drying vs
rotary evaporation) impacted suspension characteristics. Visual
observations indicated that spray-dried samples of 1 and 2
dispersed more readily than material prepared by rotary
evaporation. However, it was possible to prepare smooth
suspensions using material prepared by either process,
particularly after the addition of 1% (w/v) microcrystalline
cellulose PH101 to aid dispersion and hydroxypropyl cellulose
EXF to increase viscosity and improve the suspension
homogeneity. The addition of microcrystalline cellulose was
also advantageous in that the visual appearance of the
constitution vehicle closely resembled that of the active
suspensions, thereby ameliorating the need for development
of a separate placebo. Crystallization of 1 or 2 in aqueous
suspensions prepared at 10 or 30 mg/mL in 1% (w/v)
hydroxypropyl cellulose EXF, 1% (w/v) MCC PH101, 25 mM
citric acid, 25% (v/v) Simple Syrup, NF, and 0.1% (w/v)
Tween 80 at 5 or 25 °C for up to 24 h could not be detected by
slurry PXRD, indicating that both amorphous compounds had
suitable physical stabilities as amorphous materials to support
Phase I clinical dosing as oral suspensions.
Figure 1. DSC overlay of crystalline and amorphous materials for 1
(Tm = 176.8 °C, Tg = 147.8 °C) and 2 (Tm = 162.9 °C, Tg = 141.8
°C).
of both 1 and 2 are characterized by atypically high glass
transition temperatures (147.8 and 141.8 °C, respectively) that
are well above the usual temperatures required for processing
and manufacturing operations of the drug product. The same
glass transition (Tg) temperatures were observed for multiple
DSC scans when samples were run in pinhole pans. In addition,
similar results were observed when different batches of 1 and 2
(spray-dried or rotovaped) were tested. A slight decrease in
both the melting (Tm) and glass transition (Tg) temperatures
was observed in sealed pans using samples that contained
residual solvents (≤1.3 wt %).
Thermal studies were conducted on the corresponding
desolvated crystalline materials (confirmed by TGA analysis),
and these results are shown in Figure 1. For both compounds, a
relatively small difference (ΔT = Tm − Tg) was observed
between melting (Tm) and glass transition (Tg) temperatures.
For 1, ΔT = 29 °C and Tg (K)/Tm = 0.94, whereas for 2, ΔT =
21 °C and Tg/Tm = 0.95. This small ΔT suggests unusually
high stability for both compounds as amorphous materials.22
Powder X-ray Diffraction (PXRD). PXRD was used to
detect crystallinity in amorphous samples of 1 and 2. Following
exposure to a variety of temperature and relative humidity (R.
H.) conditions (i.e., 5 °C, 25 °C/60% R. H., 40 °C/75% R. H.,
and 50 °C), all of the PXRD patterns were consistent with
initial samples in that no crystalline reflections were observed.
In addition, the PXRD patterns collected from amorphous
samples stored at 5 and 40 °C/75% R. H. for six months
showed no signs of crystallinity. Hence, amorphous 1 and 2
were expected to remain physically stable under conditions
required for manufacturing, handling, and storage. However,
Given the promising stability profile obtained from the
aqueous suspension of amorphous materials, significant effort
was spent on identifying suitable conditions for the preparation
of amorphous 1 and 2. Although spray-drying reproducibly
yielded amorphous materials, concerns over the potential for
compromised yield led to additional solvent screening studies
for the rotary evaporation process. Our initial attempt to
prepare amorphous 1 by dissolution in methanol followed by
rotary evaporation to a solid residue was not completely
successful. Recovered 1 was found to contain residual methanol
that could not be removed even after extended drying at
elevated temperature. Subsequent PXRD analysis indicated that
we had prepared amorphous 1 contaminated with a significant
amount of a crystalline methanol solvate. Although bulk
G
DOI: 10.1021/acs.oprd.6b00035
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
preparation of completely amorphous 1 by rotary evaporation
from MeOH proved difficult, it was possible using EtOH as the
solvent. Conversely, ethanolic solutions of 2 were prone to
crystallization, but rotary evaporation of a MeOH solution
reproducibly yielded amorphous material even on a larger scale.
For preparing the clinical batches, solutions of either 1 or 2 in
MeTHF (unstabilized) were washed with a 10 wt % citric acid
solution to remove residual DCHA. After additional aqueous
washes, the solvent was removed, and the appropriate alcohol
was added. This solution was passed through a 1 μm line filter,
and the volatiles were removed by rotary evaporation (100
mmHg, 60 °C bath). The pressure was then reduced to 5
mmHg, and the batch was stripped to dryness. The solids were
collected, passed through a 16-mesh screen, and further dried in
a vacuum oven to give amorphous solids of both 1 and 2. The
isolated solids were jet milled to provide a particle size
distribution that readily formed a uniform suspension in the
aqueous vehicle used at the clinical compounding site for the
DIB products.
of 7. The organic phase was washed with 15 wt % of brine (1.25
L) and then concentrated under reduced pressure (300 mmHg,
75 °C jacket temperature) to give a mixture of 16 and 24 as a
light orange oil. This was dissolved in THF (4.0 L), and then
potassium tert-butoxide (395 g) was added all at once. An
exotherm from 20 to 33 °C was observed. The resulting
solution was heated to 65 °C for 5 h after which time HPLC
analysis indicated that <1 LCAP of intermediate 16 remained.
The reaction mixture was then cooled to 20 °C and partitioned
between water (3.0 L) and ethyl acetate (4.0 L). The organic
phase was washed with 15% brine (1.5 L) and then
concentrated via distillation (310 mmHg, 80 °C jacket
temperature) to a volume of 1 L. The KF was adjusted to
700 ppm via azeotropic distillation with ethyl acetate to a final
volume of 1.4 L. n-Heptane (2.0 L) was added, and the
suspension was heated to 75 °C for 2 h. The suspension was
cooled to 20 °C over 1 h, and the solids were collected via
filtration. The cake was slurry washed with 1 L of 4:1 n-
heptane:ethyl acetate, deliquored, and then dried at 40 °C in a
vacuum oven to provide 266 g (32.7%) of 17 as a tan solid. 1H
NMR (500 MHz, DMSO-d6) δ 8.09 (d, J = 8.2 Hz, 1H), 7.60
(dd, J = 7.7, 1.4 Hz, 1H), 7.50−7.37 (m, 2H), 7.32−7.22 (m,
2H), 5.78 (s, 1H). 13C NMR (100 MHz, DMSO-d6) δ 155.66,
149.35, 147.18, 143.04, 129.65, 129.39, 124.30, 123.58, 120.33,
118.27, 116.32, 61.22. HRMS (ESI): calcd for C12H8 O2NCl
(M+ + H) 234.0316. Found: 234.0314.
CONCLUSIONS
■
In conclusion, we have demonstrated an improved synthesis of
a new class of potential GR agonists. Notable features of this
work include the development of a novel and efficient route to
azaxanthols and the combination of a partial classical resolution
with a preparative chiral SFC that allowed us to meet our
accelerated delivery timelines. A telescoped procedure for the
preparation of the amide boronic acids (18 and 19) that
generated material with excellent purity was also developed and
implemented. In addition, the amorphous forms of 1 and 2
were shown to exhibit acceptable physical stabilities for oral
suspension development, storage, and use-time period required
to support the limited Phase I clinical studies for the program.
Development of amorphous APIs and their use in phase-
appropriate suspensions provided a means to accelerate the
development of the program and allowed rapid evaluation of
the clinical/biologic hypothesis for this chemotype. This
approach aided the rapid investigation of the proof of
mechanism for these dissociated GR agonists and may be
applicable to other clinical candidates.
Preparation of 2-(2-Chloro-5H-chromeno[2,3-b]-
pyridin-5-yl)-methyl-2-methylpropanoate 9. To a jack-
eted 20 L Chemglass reactor were charged 17 (500 g, 2.14 mol)
and dichloromethane (6.62 kg, 5 L). The mixture was cooled to
−5 °C, and titanium tetrachloride (1 M in dichloromethane,
2.35 L, 2.35 mol) was charged over a 30 min period. The
resulting solution was stirred for an additional 30 min at −5 °C,
and then 1-methoxy-1-trimethylsilyloxy-2-methyl-1-propene
(23, 1.02 L, 4.81 mol) was added over a 30 min period,
maintaining the batch temperature between 0 and −5 °C. The
reaction mixture was stirred at this temperature for an
additional 1 h, after which time HPLC analysis indicated
reaction completion. Ethyl acetate (2.5 L, 2.25 kg) was added
followed by a slow addition of ammonium hydroxide (7.19 M
in water, 0.5 L, 3.60 mol) to an apparent pH of 9−10. The
mixture was warmed to 20 °C and held at this temperature for
1 h, and then the titanium salts were removed via polish
filtration. Water (3 L) was added to the filtrate, and the biphasic
reaction mixture was stirred for 30 min. Agitation was stopped,
and the lower spent aqueous stream was drawn off for disposal.
The dichloromethane was removed via codistillation with ethyl
acetate, and the final volume was adjusted to 1 L. The solution
was cooled to 25 °C over 3 h and held at this temperature for 1
h to give a slurry. The crystals were collected via filtration, and
the cake was washed with 0.2 L of ethyl acetate. The cake was
dried under vacuum (30 Torr, 50 °C) to an LOD end point of
<0.5 wt % to give 9 as a white solid (370 g, 99.2 LCAP, 99.1 wt
EXPERIMENTAL SECTION
■
Preparation of 2-Chloro-5H-chromeno[2,3-b]pyridin-
5-ol (17). To a 20 L Hastelloy reactor were added THF (2.0
L), n-heptane (2.0 L), and diisopropylamine (1.15 L), and the
resulting solution was cooled to −40 °C. Then, n-butyl lithium
(3.15 L, 2.56 M) was added over 45 min. A 21 °C exotherm to
−19 °C was observed with active cooling during the course of
the addition. The batch temperature was adjusted to −55 °C,
and a solution of 2,6-dichloropyridine (500 g) in THF (1.5 L)
was added over 40 min, maintaining the temperature below
−50 °C. The resulting suspension was cooled to −60 °C and
held at this temperature for 2 h. Then, salicylaldehyde (367
mL) was added rapidly via an addition funnel in three equal
portions, maintaining the batch temperature below −40 °C.
After each addition, the batch temperature was readjusted to
−53 to −55 °C before the next fraction was added. After the
final portion was added, the reaction mixture was stirred at −50
°C for 1 h and then warmed to −20 °C. The reaction was
quenched with 10 wt % aqueous ammonium chloride (4.25 L).
The organic phase was separated and washed with 7.5 wt % of
acetic acid (2 × 2.5 L then 1.25 L) to an apparent pH endpoint
1
%, 61% corrected yield). H NMR (400 MHz, DMSO-d6) δ
7.77 (1H, d, J = 7.83 Hz), 7.39−7.36 (2H, m), 7.22−7.17 (3H,
m), 4.39 (1H, s), 3.60 (3H, s), 0.92 (3H, s), 0.90 (3H, s); 13C
NMR (100 MHz, DMSO-d6) δ 175.33, 158.47, 152.05, 146.68,
142.72, 129.63, 128.91, 127.27, 121.03, 120.07, 116.49, 115.48,
51.83, 49.02, 45.63, 21.33, 21.17. HRMS (ESI): calcd for
C17H17O3NCl (M+ + 1) 318.08981. Found: 318.08915.
Preparation of 2-(2-Chloro-5H-chromeno[2,3-b]-
pyridin-5-yl)-2-methylpropanoic Acid (10a, 10b). To a
20 L jacketed Chemglass reactor were charged 9 (380 g, 1.00
H
DOI: 10.1021/acs.oprd.6b00035
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
equiv, 1.20 mol) and N-methylpyrrolidone (2.81 L; 2.89 kg).
The mixture was warmed to 50−55 °C, and then an aqueous
solution of potassium hydroxide (4.78 mol; 409.6 mL; 596.4 g)
was charged over 15 min. The reaction mixture was stirred at
this temperature range for 1 h after which time HPLC analysis
indicated reaction completion. The mixture was adjusted to an
apparent pH of 5 with 3.8 L of 1 N HCl to give a cloudy
solution. The batch was warmed to 65 °C, an additional 2.18 L
of 1 N HCl was added, and then the reaction mixture was
warmed to 85−90 °C. Then, 3.80 L of water was added,
maintaining the reaction temperature between 85 and 90 °C.
The resulting slurry was cooled to 20 °C over 2 h and held at
this temperature for an additional 1 h. The crystals were
collected via filtration, washed with 3.8 L of water, deliquored
for 1 h, and dried on the filter overnight to give 339 g (93.4%
L) and 2 N HCl (3.5 L), and the mixture was agitated for 5
min. The lower spent aqueous stream was back extracted with
2-methyltetrahydrofuran (1.5 L). The combined rich organic
streams were used immediately in the preparation of 18
without further purification.
Preparation of [4-(Dimethylcarbamoyl)-3-
fluorophenyl]boronic Acid (18). The rich MeTHF stream
of 29 was charged to a 20 L reactor and cooled to 10 °C. To
this were charged sodium periodate (1.02 kg, 4.77 mol) and
water (7 L), and the mixture was agitated for 1 h. An exotherm
from 10 to 18 °C was observed. Then, 1 N HCl (5 L) was
charged to the reactor, and the resulting biphasic mixture was
stirred at room temperature overnight after which time HPLC
analysis indicated that <2 LCAP of 29 remained. Agitation was
stopped, and the layers were allowed to separate. The bottom
spent aqueous stream was back extracted with 2-methyltetrahy-
drofuran (2 L). The combined organic streams were washed
with a 15 wt % sodium thiosulfate solution (2 L) and then with
half-saturated brine (3 L). The organic stream was filtered
through a 0.9 sq ft R53SP carbon Cuno cartridge and
concentrated under vacuum to a volume of approximately 3
L. In a separate crystallizer were charged 18 seed crystals (7 g)
and 2 L of n-heptane. To this were added 3 L of the rich
MeTHF stream of 18 simultaneously with n-heptane (4.8 L)
over 2 h at 23 °C. The resulting thick slurry was agitated for 1.5
h at room temperature. The crystals were collected via
filtration, and the cake was washed with a 1:1.5 mixture of 2-
methyltetrahydrofuran and n-heptane (1.25 L) and then with n-
heptane (1 L). The wet cake was deliquored for 1 h and dried
under vacuum at room temperature overnight to give 443 g
1
yield) of a mixture of 10a and 10b as an off-white solid. H
NMR (400 MHz, DMSO-d6) δ 12.72 (1H, s), 7.82 (1H, d, J =
8.08 Hz), 7.39−7.35 (3H, m), 7.33−7.18 (2H, m), 4.42 (1H,
s), 0.87 (s, 3H), 0.87 (3H, s); 13C NMR (100 MHz, DMSO-d6)
δ 176.87, 158.55, 152.11, 146.57, 142.83, 129.89, 128.80,
124.20, 121.35, 120.00, 116.45, 115.82, 48.79, 45.16, 21.31,
21.27. HRMS (ESI): calcd for C16H15O3NCl (M+ + 1)
304.07350. Found: 304.07413.
Isolation of 10a. As described in the accompanying paper,
a combination of a partial classical resolution with cinchonidine
combined with preparative chiral SFC was used to isolate
10a.15
Preparation of 4-Bromo-2-fluoro-N,N-dimethylbenza-
mide (27). To a 20 L reactor were charged 4-bromo-2-
fluorobenzoic acid 26 (695 g, 3.17 mol), dimethylamine
hydrochloride (650 g, 7.97 mol), 1-(3-(dimethylamino)-
propyl)-3-ethylcarbodiimide hydrochloride (1.22 kg, 6.36
mol), 1-hydroxybenzotriazole hydrate (980 g, 6.4 mol),
acetonitrile (4.48 kg), and diisopropylethylamine (1050 g,
8.12 mol). The reaction mixture was heated to 69 °C and held
at this temperature until HPLC analysis indicated that <2%
LCAP of unreacted 26 remained. The reaction mixture was
concentrated at 70 °C and 60 mmHg to a volume of 5 L. Then,
2-methyltetrahydrofuran (7 L) and 1 N sodium hydroxide (7
L) were added to the reaction, and the mixture was stirred for
15 min. Agitation was stopped, and the bottom spent aqueous
stream was drawn off for disposal. The rich organic stream was
washed with half-saturated brine (4 L), and the KF was
adjusted to <0.2 wt % via azeotropic distillation at 70 °C and 60
mmHg.
Preparation of 2-Fluoro-N,N-dimethyl-4-(tetramethyl-
1,3,2-dioxaborolan-2-yl)benzamide (29). To the rich
MeTHF stream of 27 were charged bis(pinacolato)diboron
(970 g, 3.82 mol), potassium acetate (930 g, 9.48 mol), and
(1,1′-bis(diphenylphosphino)ferrocene) palladium(II) chloride
(26.0 g, 32.8 mmol). The reaction was heated to 80 °C and
held at this temperature for 16 h after which time <3% LCAP of
30 remained as determined by HPLC. The reaction was cooled
to 20 °C and washed with water (7.2 L) and then with 4% brine
(7 L). Palladium was removed via filtration of the rich organic
stream through a 0.9 sq ft R53SP carbon Cuno cartridge. This
was charged to a 20 L reactor followed by 1 N sodium
hydroxide (7 L) and water (0.8 L). The biphasic mixture was
agitated for 5 min and then allowed to settle. The lower rich
aqueous stream was drawn off, and the spent organic stream
was extracted again with 1 N NaOH (3 L). The combined rich
aqueous streams were charged back into a clean reactor and
cooled to 0 °C. To this were added 2-methyltetrahydrofuran (7
1
(66% yield from 26, 99.9 LCAP) of a white solid. H NMR
(400 MHz, DMSO-d6) δ 8.32 (br s, 2H), 7.79−7.54 (m, 2H),
7.41−7.29 (m, 1H), 3.03−2.96 (m, 3H), 2.85−2.78 (m, 3H);
13C NMR (101 MHz, DMSO-d6) δ 166.20−165.00 (m, 1C),
158.5, 156.0, 130.3 (d, J = 2.9 Hz, 1C), 127.9 (d, J = 3.7 Hz,
1C), 126.1 (d, J = 18.3 Hz, 1C), 120.4 (d, J = 19.0 Hz, 1C),
37.8 34.2; 19F NMR (376 MHz, DMSO-d6) δ −118.13 to
−118.25 (m, 1F); HRMS (+CI) m/z 211.0939 [(M)+; calcd for
C9H11BFNO3, 211.0931].
Preparation of 4-Bromo-N-ethyl-2-fluoro-N-methyl-
benzamide (28). To a jacketed 20 L Chemglass reactor fitted
with an overhead stirrer were charged 26 (700 g, 3.20 mol), 1-
hydroxybenzotriazole hydrate (987 g, 6.45 mol), 1-(3-
(dimethylamino)propyl)-3-ethylcarbodiimide hydrochloride
(1225 g, 6.39 mol), acetonitrile (5.62 L), diisopropylamine
(1.4 L, 8.01 mol), and ethylmethylamine (715 mL, 8.26 mol).
The reaction mixture was heated at 62 °C (bath temperature of
70 °C) for 1.5 h after which time HPLC analysis indicated the
complete consumption of 26. The volume was adjusted to
approximately 4 L via distillation under reduced pressure and
then cooled to 20 °C. Then, MeTHF (11.7 kg, 23.6 L) and 1 N
sodium hydroxide (3.64 kg, 3.5 L, 3.5 mol) were added, and the
biphasic mixture was stirred for 5 min. The lower spent
aqueous stream was drawn off for disposal. The rich organic
stream was washed with 1 N sodium hydroxide (3.63 kg, 3.49
L, 3.49 mol) followed by half-saturated brine (3.82 kg, 3.2 L)
and then transferred through a 0.9 sq ft R53SP carbon cartridge
followed by a 0.5 μm in-line filter to a 20 L Schott filter flask.
The rich organic stream was concentrated under reduced
pressure (100−60 mmHg) to a volume of 4 L. The volume was
adjusted to 7 L with MeTHF. The water content was 0.09 wt %
as determined by Karl Fischer titration.
I
DOI: 10.1021/acs.oprd.6b00035
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
Preparation of N-Ethyl-2-fluoro-N-methyl-4-(tetra-
methyl-1,3,2-dioxaborolan-2-yl)benzamide (30). To the
MeTHF solution of 28 were charged bis(pinacolato) diboron
(970 g, 3.82 mol), potassium acetate (920 g, 9.37 mol), and
(1,1′-bis(diphenylphosphino)ferrocene) palladium(II) chloride
(26 g, 31.8 mmol). The reactor was evacuated to 200 mmHg,
and the vacuum was broken slowly with nitrogen. The reaction
mixture was heated to 78 °C (jacket temperature of 80 °C) and
stirred at this temperature for 15 h after which time HPLC
analysis indicated complete consumption of 28. The reaction
was cooled to 20 °C. Water (7 L) was added, and the biphasic
mixture was stirred at room temperature for 10 min. Agitation
was stopped, and the lower spent aqueous stream was drawn off
for disposal. The rich organic stream was washed with water
followed by half-saturated brine (8 L) and then transferred
through a 0.9 sq ft R53SP carbon cartridge followed by a 0.5
μm in-line filter into a clean reactor. The lines were rinsed with
an additional 3.4 L of MeTHF. Then, 1 N sodium hydroxide (7
kg) was added, and the biphasic mixture was stirred for 30 min.
The upper spent organic stream was drawn off for disposal.
This was found to contain 2.6% of 30 as determined by HPLC
analysis. To the lower rich aqueous stream was added MeTHF
(7 L), and the mixture was cooled to 2 °C. Then, 2 N
hydrochloric acid (4.2 L) was added to an apparent pH end
point of 2. The biphasic mixture was stirred and allowed to
warm to room temperature over 0.5 h. Agitation was stopped,
and the layers were allowed to separate. The lower spent
aqueous stream was back extracted with MeTHF (3.8 L), and
the combined rich organic stream of 30 was used immediately
in the next reaction without further purification.
Preparation of [4-Ethyl(methyl)carbamoyl]-3-fluoro-
phenyl Boronic Acid (19). The rich MeTHF solution of 30
was cooled to 6.5 °C. Then, sodium periodate (960 g; 4.49
mol) and water (5.1 L) were added. The mixture was stirred for
1 h as the temperature was warmed to 20 °C. Then, 1 N
hydrochloric acid (4.46 L) was added, and the biphasic mixture
was stirred at room temperature overnight after which time
HPLC analysis showed complete consumption of 30. Agitation
was stopped, and the layers were allowed to separate. The
bottom spent aqueous stream was back extracted with MeTHF
(2.1 L). To the combined organic layers was added 20%
sodium thiosulfate (3 L), and the biphasic mixture was stirred
for 15 min. A 6 °C exotherm from 20 to 26 °C was observed as
the dark reaction mixture turned light yellow. Agitation was
stopped, and the layers were allowed to separate. The lower
spent aqueous stream was drawn off for disposal. The rich
organic stream was washed with half-saturated brine (3 L) and
then transferred through a 0.9 sq ft R53SP carbon cartridge
followed by a 0.5 μm in-line filter to a clean reactor. The
volume was adjusted to 3.8 L via distillation under reduced
pressure (90 mmHg, jacket temperature of 70 °C, batch
temperature of 25 °C). To a clean crystallizer were charged n-
heptane (5.7 kg) and 26 g of 19 seed crystals. To this were
simultaneously added 3.8 L of the rich MeTHF stream of 19
and n-heptane (4.3 kg) over a 2 h period. The thick white slurry
was stirred overnight at 20 °C. The crystals were collected via
filtration and washed with a 1:1 mixture of MeTHF:n-heptane
(3 L) and heptane (1 L). The cake was deliquored for 1 h and
then dried at 45 °C for 72 h to obtain 486 g (69% yield from
26) of a white crystalline solid. 1H NMR (400 MHz, DMSO-d6,
rotamers) δ 8.32 (br s, 2H), 7.80−7.50 (m, 4H), 7.43−7.26 (m,
2H), 3.53−3.37 (m, 2H), 3.21−3.06 (m, 2H), 3.02−2.88 (m,
3H), 2.84−2.71 (m, 3H), 1.16−1.09 (m, 3H), 1.00 (s, 3H); 13C
NMR (101 MHz, DMSO-d6) δ 165.46−164.98 (m, 1C), 158.4,
155.9, 130.3 (dd, J = 5.1, 2.9 Hz, 1C), 128.02−127.19 (m, 1C),
126.64−126.08 (m, 1C), 120.4 (d, J = 19.0 Hz, 1C), 44.9, 41.2,
35.2, 31.3, 13.2, 11.9; 19F NMR (376 MHz, DMSO-d6) δ
−118.52, −118.72. HRMS (+CI) m/z 225.1097 [(M)+; calcd
for C10H13BFNO3, 225.1087].
Preparation of 2-[(5S)-2-[4-(Dimethylcarbamoyl)-3-
fluorophenyl]-5H-chromeno[2,3-b]pyridin-5-yl]-2-meth-
ylpropanoic Acid (20). Compounds 10a (0.243 kg, 0.801
mol) and 18 (0.254 kg, 1.20 mol) were charged to a 10 L
Chemglass reactor under nitrogen. Then, solid potassium
carbonate (0.225 kg, 1.61 mol) and tetrabutylammonium
bromide (TBABr, 0.026 kg, 0.080 mol) were charged, followed
by tricyclohexylphosphonium tetrafluoroborate (0.012 kg,
0.032 mol), tetrahydrofuran (THF, 1.3 L, 5 mL/g), and
water (1.3 L, 5 mL/g). While the solution was agitated, a
nitrogen line was introduced directly into the biphasic solution
with vigorous bubbling. After 20 min, Pd(OAc)2 (3.64 g, 0.016
mol) was added directly to the solution, and nitrogen sparging
was continued for an additional 5 min. The nitrogen line was
removed from the solution, and the reaction mixture was
warmed to 70 °C. After 3 h, in-process HPLC analysis showed
1.19 relative area percent (RAP) (10a/20), and the reaction
was determined to be complete. The reaction mixture was
cooled to 20 °C. THF (1.3 L, 5 mL/g) and water (1.3 L, 5 mL/
g) were added, followed by 3 N hydrochloric acid (850 mL, 3.5
mL/g) to an apparent pH of 2.5. Agitation was stopped, and
the phases were allowed to separate. The lower spent aqueous
stream was drawn off for disposal. The upper rich organic
stream was washed with 14% aq sodium chloride (1.3 L, 5 mL/
g) and dried over sodium sulfate (0.3 kg). The dark rich THF
solution was passed through a Cuno zeta pad (R53SP). It was
observed during filtration through the zeta pad that a white
solid was precipitating from solution, later identified as product
20. The carbon-filtered solution was introduced back into the
clean reactor, followed by n-butylacetate (2.6 L, 10 mL/g). The
reactor jacket was set to 80 °C with vacuum (120 psig max),
and THF was removed via distillation. During the course of the
distillation, 20 began to precipitate from solution. After the
distillation was complete, the slurry was cooled to 20 °C. Then,
n-heptane (1.3 L, 5 mL/g) was added over 15 min, and the
thick white slurry was allowed to stir at room temperature for
12 h. The crystals were collected via filtration, washed with n-
heptane (1 L), deliquored for 5 h, and dried under vacuum at
50 °C for 24 h to give 20 (0.307 kg, 88.2%) as a white powder.
Preparation of 2-[(5S)-2-{4-[Ethyl(methyl)carbamoyl]-
3-fluorophenyl}-5H-chromeno[2,3-b]pyridin-5-yl]-2-
methylpropanoic Acid (21). Compounds 10a (0.100 kg,
0.033 mol) and 19 (0.096 kg, 0.427 mol) were charged to a 3-
neck, 2 L round-bottom flask equipped with an overhead
stirrer. Then, solid potassium carbonate (0.091 kg, 0.658 mol)
and tetrabutylammonium bromide (TBABr, 0.011 kg, 0.033
mol) were charged, followed by tricyclohexylphosphonium
tetrafluoroborate (4.85 g, 0.001 mol), MeTHF (0.5 L, 5 mL/g),
and water (0.5 L, 5 mL/g). The solution was agitated, and a
nitrogen line was introduced directly into the biphasic mixture
with vigorous bubbling. After 20 min, Pd(OAc)2 (1.48 g, 0.006
mol) was added directly to the solution, and nitrogen sparging
was continued for an additional 5 min. The nitrogen line was
removed from the solution; the flask was heated to 90 °C and
held at this temperature for 10 h after which time in-process
HPLC analysis showed <0.1 RAP (10a/21). The reaction
mixture was cooled to room temperature. Then, 3 N
J
DOI: 10.1021/acs.oprd.6b00035
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
hydrochloric acid (360 mL, 3.5 mL/g) was added to the
reaction mixture to an apparent pH end point of 2.5. Agitation
was stopped, and the layers were allowed to separate. The
lower spent aqueous stream was drawn off for disposal. The
upper rich organic stream was washed with 14 wt % aq sodium
chloride (0.5 L, 5 mL/g). The rich MeTHF layer was drained
from the reactor into an Erlenmeyer flask (3 L). With stirring,
dicyclohexylamine (98 mL, 0.492 mol) was added, followed by
n-BuOAc (1 L, 10 mL/g). After 10 min, a light gray solid
precipitated out of solution, and stirring was continued for an
additional 1 h. The slurry was filtered, washed with n-BuOAc
(0.2 L, 2 mL/g), and dried via passing air through the cake over
2 h. After 2 h of drying, the cake was transferred into a 4 L
Erlenmeyer flask, and THF (1 L, 10 mL/g) and water (0.75 L,
7.5 mL/g) were added. This mixture was agitated for 30 min,
and some solids were still not dissolved; however, 3 N HCl
(180 mL, 1.8 mL/g) was added until a thick precipitate
(dicyclohexylamine HCl salt) crashed out of solution. The gray
biphasic slurry was polish filtered (Celite), resulting in a clean
phase split. The filtrate was transferred to a 4 L separatory
funnel, and n-BuOAc (0.50 L, 5 mL/g) was added to assist in
the phase split. The bottom aqueous layer was removed, and
the upper organic layer that was a mixture of THF and n-
BuOAc layer was distilled to remove THF. After removal of
THF (0.80 L, 8 mL/g), a thick white precipitate formed. The
white slurry was added to a 4 L Erlenmeyer flask with overhead
stirring. Additional n-BuOAc (0.50 L, 5 mL/g) was added, and
the thick white slurry was stirred for 2 h. The slurry was filtered,
washed with n-heptane (0.50 L, 5 mL/g), and dried by passing
air through the cake for 1 h. The cake was loaded into a tray
and vacuum-dried at 50 °C for 15 h. Product 21 (0.120 kg,
81.3%) containing 20 ppm of residual palladium was isolated
with an LCAP purity of 99.9%.
was warmed to 75−80 °C and stirred at this temperature
overnight. The resulting slurry was cooled to 20 °C and held at
this temperature for an additional 24 h. The crystals were
collected via filtration and washed with 500 mL of isopropyl
acetate, and the cake was deliquored for 3 h. The cake was
dried at 20 °C for 18 h to give 281 g (86.7% yield) of crystalline
1
1 as a white powder. H NMR (500 MHz, DMSO-d6) δ 12.67
(1 H, br s), 9.24 (1 H, s), 8.03 (1 H, dd, J = 8.11, 1.51 Hz),
7.99 (1 H, dd, J = 11.13, 1.24 Hz), 7.90 (1 H, d, J = 7.70 Hz),
7.67 (1 H, d, J = 7.97 Hz), 7.51 (1 H, t, J = 7.56 Hz), 7.37 (1 H,
ddd, J = 8.25, 6.60, 2.20 Hz), 7.31 (1 H, d, J = 7.97 Hz), 7.15−
7.19 (1 H, m), 7.11−7.16 (1 H, m), 4.85 (1 H, s), 3.03 (3 H, s),
2.89 (3 H, s), 1.06 (3 H, s), 1.04 (3 H, s), 13C NMR (126
MHz, DMSO-d6) δ 174.17, 164.94, 158.92 (2 C, s), 157.99 (1
C, d, J = 259.40 Hz), 152.53, 151.48, 148.97, 140.34−140.44 (1
C, m), 140.27, 129.20 (2 C, br s), 128.79, 125.18 (1 C, d, J =
17.80 Hz), 123.88, 122.66, 120.92, 116.74, 116.58, 116.03,
113.37 (1 C, d, J = 22.89 Hz), 49.66, 44.61, 37.69, 34.19, 20.93,
20.69.
Preparation of Amorphous 1 from Crystalline Materi-
al. Crystalline 1 (358 g) was charged to a 10 L reactor. To this
was added 3.1 kg of nonstabilized 2-methyltetrahydrofuran, and
the resulting mixture was agitated until complete dissolution
was obtained. Then, 3.0 kg of a 10 wt % citric acid solution was
charged to the reactor, and the biphasic mixture was agitated
for 5 min. Agitation was stopped, and the layers were allowed
to separate. The lower spent aqueous stream was drawn off for
disposal. To the rich organic stream was charged 3 kg of water,
and the biphasic mixture was agitated for 5 min. Agitation was
stopped, and the layers were allowed to separate. The lower
spent aqueous stream was drawn off for disposal. Then, 3.1 kg
of a saturated aqueous sodium bicarbonate solution was
charged to the batch, and the biphasic solution was agitated
for 5 min. Agitation was stopped, and the layers were allowed
to separate. The lower spent aqueous stream was drawn off for
disposal. To the rich organic stream was charged 3.2 kg of half-
saturated brine, and the biphasic mixture was agitated for 5 min.
Agitation was stopped, and the layer were allowed to separate.
The lower spent aqueous stream was drawn off for disposal.
The rich organic stream was transferred to a rotary evaporator
through a 1 μm in-line filter, and the transfer line was rinsed
with 1 L of 2-methyltetrahydrofuran. The solvent was removed
via rotoevaporation at 60 mmHg with a bath temperature of 60
°C. The batch temperature was 22−25 °C. Then, 3 kg of
absolute ethanol was charged to the rotary evaporator. The
solvent was removed under the same conditions. The solids
were dried at 55 °C and 1 mmHg on the rotovap. The solids
were removed from the flask, passed through a 16 mesh screen,
and dried at 55 °C overnight to give 342 g (92.6% recovery
corrected for 3% residual ethanol) of 1 as a white solid. This
was determined to be amorphous by PXRD. A portion of the
Preparation of 2-Fluoro-N,N-dimethyl-4-[(5S)-5-{1-
methyl-1-[(1,3,4-thiadiazol-2-yl)carbamoyl]ethyl}-5H-
chromeno[2,3-b]pyridin-2-yl]benzamide (1). Compound
1 was prepared by coupling 20 with 2-amino-1,3,4-thiadiazole,
and the crude product crystallized from isopropyl acetate. The
crystalline isopropyl acetate solvate was then converted to the
amorphous material.
Preparation of Crystalline 1. To a 3 L 3-neck round-bottom
flask equipped with an overhead stirrer, temperature probe, and
nitrogen inlet were sequentially added 20 (274.0 g, 0.631 mol),
1-hydroxy-7-azabenzotriazole (HOAt, 103.0 g, 0.757 mol), 1-
(3-(dimethylamino)propyl)-3-ethylcarbodiimide hydrochloride
(146.0 g, 0.762 mol), MeTHF (2.77 L), and diisopropylethyl-
amine (134.9 mL, 0.773 mol). The resulting yellow suspension
was stirred under nitrogen at ambient temperature for 30 min.
2-Amino-1,3,4-thiadiazole (128.0 g, 1.266 mol) was added, and
the resulting mixture was heated to 55 °C for 16 h after which
time HPLC analysis indicated that <0.5 RAP of 20 as the
HOAT adduct remained. The solution was cooled to room
temperature. Then, 1 N hydrochloric acid (1.38) kg was added,
and the resulting biphasic mixture was agitated for 5 min.
Agitation was stopped, and the layers were allowed to separate.
The lower spent aqueous stream was drawn off for disposal.
The rich organic stream was washed with 5% aqueous sodium
bicarbonate (1.0 L) and half-saturated brine (0.5 L) and then
concentrated under reduced pressure to a volume of
approximately 2 L. Solvent exchange was accomplished by
intermittent addition of 4 kg of isopropyl acetate and
distillation between 100 and 300 mg of Hg to a final volume
of approximately 800 mL. The isopropyl acetate solution of 1
1
solids were passed through a jet mill to provide 319 g of 1. H
NMR (500 MHz, DMSO-d6) δ 12.67 (1 H, br s), 9.24 (1 H, s),
8.03 (1 H, dd, J = 8.11, 1.51 Hz), 7.99 (1 H, dd, J = 11.13, 1.24
Hz), 7.90 (1 H, d, J = 7.70 Hz), 7.67 (1 H, d, J = 7.97 Hz), 7.51
(1 H, t, J = 7.56 Hz), 7.37 (1 H, ddd, J = 8.25, 6.60, 2.20 Hz),
7.31 (1 H, d, J = 7.97 Hz), 7.15−7.19 (1 H, m), 7.11−7.16 (1
H, m), 4.85 (1 H, s), 3.03 (3 H, s), 2.89 (3 H, s), 1.06 (3 H, s),
1.04 (3 H, s). 13C NMR (126 MHz, DMSO-d6) δ 174.17,
164.94, 158.92 (2 C, s), 157.99 (1 C, d, J = 259.40 Hz), 152.53,
151.48, 148.97, 140.34−140.44 (1 C, m), 140.27, 129.20 (2 C,
br s), 128.79, 125.18 (1 C, d, J = 17.80 Hz), 123.88, 122.66,
120.92, 116.74, 116.58, 116.03, 113.37 (1 C, d, J = 22.89 Hz),
K
DOI: 10.1021/acs.oprd.6b00035
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
49.66, 44.61, 37.69, 34.19, 20.93, 20.69. 19F NMR (471 MHz,
DMSO-d6) δ −115.75 (1 F, t, J = 7.27 Hz). HRMS (ESI): calcd
for C27H24FN5O3S (M+ + 1) 518.16567. Found: 518.16642.
Preparation of N-Ethyl-2-fluoro-N-methyl-4-[(5S)-5-{1-
methyl-1-[(1,3,4-thiadiazol-2-yl)carbamoyl]ethyl}-5H-
chromeno[2,3-b]pyridin-2-yl]benzamide (2). Compound
2 was prepared by coupling 21 with 2-amino-1,3,4-thiadiazole,
and the crude product was crystallized from isopropyl acetate.
The two batches of crystalline material were combined and
processed into a single lot of amorphous material. The
procedure below details the preparation of the first batch and
the turnover of the combined material.
stream was charged 3.1 kg of half-saturated brine, and the
biphasic mixture was agitated for 5 min. Agitation was stopped,
and the layers were allowed to separate. The lower spent
aqueous stream was drawn off for disposal. The rich organic
stream was transferred to a rotary evaporator through a 1 μm
in-line filter, and the transfer line was rinsed with 1 L of
MeTHF. The solvent was removed via rotoevaporation at 60
mmHg with a bath temperature of 70 °C. The batch
temperature was 25 °C. Then, 3.2 kg of methanol was charged
to the rotary evaporator to completely dissolve the batch. The
solvent was removed under the same conditions. The solids
were dried at 55 °C and 1 mmHg on the rotary evaporator. The
solids were removed from the flask, passed through a 16 mesh
screen, and dried at 55 °C overnight to give 362 g (94.0%
recovery corrected for 0.3% residual methanol) of 2 as a white
solid. This was determined to be amorphous by PXRD. The
solids were passed through a jet mill to provide 352 g of 2.
At 27 °C in DMSO-d6, a 52:48 mixture of rotational isomers
Preparation of Crystalline 2. To a 20 L reactor equipped
with an overhead stirrer, temperature probe, and nitrogen inlet
were charged 21 (265 g, 0.591 mol), 1-(3-(dimethylamino)-
propyl)-3-ethylcarbodiimide hydrochloride (135.9 g, 0.709
mol), 1-hydroxy-7-azabenzotriazole (HOAt, 96.51 g, 0.709
mol), 2-methyltetrahydrofuran (3.09 kg, 3.5 L), and diisopro-
pylethylamine (91.6 g, 123.7 mL, 0.709 mol). The resulting
yellow suspension was stirred under nitrogen at 23 °C for 3 h
after which time HPLC analysis indicated complete conversion
to the HOAT adduct. Then, 2-amino-1,3,4-thiadiazole (126.0 g,
1.25 mol) was added, and the resulting mixture was heated to
60 °C for 16 h after which time HPLC analysis indicated that
<0.5 RAP of 21 as the HOAT adduct remained. The solution
was cooled to room temperature. Then, 1 N hydrochloric acid
(1.38) kg was added, and the resulting biphasic mixture was
agitated for 5 min. Agitation was stopped, and the layers were
allowed to separate. The lower spent aqueous stream was
drawn off for disposal. The rich organic stream was washed
with 5% aqueous sodium bicarbonate (1.0 L) followed by half-
saturated brine (0.5 L) and concentrated under reduced
pressure (275 mg Hg with jacket temperature set to 100 °C to a
volume of approximately 2 L. Solvent exchange was
accomplished by intermittent addition of 5.3 kg of isopropyl
acetate and distillation at 275 mg Hg to a final volume of
approximately 1 L. The resulting slurry was warmed to 60−65
°C and held at this temperature overnight. The slurry was then
cooled to 20 °C over a 2 h period and held at this temperature
for an additional 1 h. The crystals were collected via filtration
using a 5−10 μm polypropylene filter cloth with a Whatman #1
filter paper. The crystals were washed with 0.53 kg of isopropyl
acetate, and the cake was deliquored for 4 h after which time
the LOD end point had been reached and no further drying
was necessary. Compound 2 was obtained in 88.9% yield (306
g) as a crystalline IPAc solvate.
Preparation of Amorphous 2 from Crystalline Materi-
al. Crystalline isopropyl acetate solvate 2 (388 g) was charged
to a 10 L reactor. To this was added 4.4 kg of nonstabilized 2-
methyltetrahydrofuran, and the resulting mixture was agitated
until complete dissolution was obtained. Then, 3 kg of a 10 wt
% citric acid solution was charged to the reactor, and the
biphasic mixture was agitated for 5 min. Agitation was stopped,
and the layers were allowed to separate. The lower spent
aqueous stream was drawn off for disposal. To the rich organic
stream was charged 3 kg of water, and the biphasic mixture was
agitated for 5 min. Agitation was stopped, and the layers were
allowed to separate. The lower spent aqueous stream was
drawn off for disposal. Then, 3.0 kg of a saturated aqueous
sodium bicarbonate solution was charged to the batch, and the
biphasic solution was agitated for 5 min. Agitation was stopped,
and the layers were allowed to separate. The lower spent
aqueous stream was drawn off for disposal. To the rich organic
1
around the tertiary amide bond could be seen in both the H
and 13C NMR spectra. Upon heating to 120 °C in DMSO-d6,
1
the aromatic protons in the H NMR sharpened significantly.
In addition, the doubled methyl and methylene signals of the
tertiary amide each broadened and collapsed to a single broad
1
resonance. H NMR of 2 at 27 °C: mixture of rotational
isomers (500 MHz, DMSO-d6) δ 12.67 (2 H, br s), 9.24 (2 H,
s), 8.03 (2 H, ddd, J = 7.97, 3.16, 1.51 Hz), 7.99 (2 H, dt, J =
11.07, 1.75 Hz), 7.90 (2 H, dd, J = 7.70, 1.92 Hz), 7.67 (2 H, d,
J = 7.70 Hz), 7.50 (2 H, t, J = 7.56 Hz), 7.37 (2 H, ddd, J =
8.25, 6.60, 2.20 Hz), 7.31 (2 H, d, J = 7.97 Hz), 7.16−7.18 (2
H, m), 7.11−7.16 (2 H, m), 4.85 (2 H, s), 3.51 (2 H, q, J = 7.06
Hz), 3.19 (2 H, q, J = 7.06 Hz), 3.00 (3 H, s), 2.85 (3 H, s),
1.16 (3 H, t, J = 7.01 Hz), 1.06 (6 H, s), 1.04 (6 H, s), 1.02−
1.05 (3 H, m)
1H NMR of 2 at 120 °C (500 MHz, DMSO-d6) δ 12.09 (1
H, br s), 9.11 (1 H, s), 7.95 (1 H, d, J = 7.97 Hz), 7.89 (1 H, d,
J = 11.00 Hz), 7.73−7.78 (1 H, m), 7.69 (1 H, d, J = 7.42 Hz),
7.43 (1 H, t, J = 7.56 Hz), 7.34 (1 H, t, J = 7.70 Hz), 7.25 (1 H,
d, J = 8.25 Hz), 7.17−7.22 (1 H, m), 7.10 (1 H, t, J = 7.42 Hz),
4.81 (1 H, s), 3.22−3.51 (2 H, m), 2.83−2.95 (3 H, m), 1.10 (3
H, br s), 1.08 (3 H, br s), 1.03−1.15 (3 H, m). 13C NMR of 2
at 27 °C (126 MHz, DMSO-d6) δ 174.25 (2 C, br s), 164.80 (1
C, s), 164.56 (1 C, s), 159.10 (2 C, br s), 158.94 (2 C, s),
157.88 (2 C, d, J = 241.60 Hz), 152.57 (2 C, s), 151.52 (2 C, s),
148.93 (2 C, br s), 140.29 (2 C, s), 140.16 (2 C, d, J = 7.63
Hz), 129.25 (2 C, s), 129.07 (2 C, br s), 128.79 (2 C, br s),
125.47 (2 C, d, J = 20.34 Hz), 123.88 (2 C, br s), 122.70 (2 C,
br s), 120.96 (2 C, s), 116.72 (2 C, s), 116.60 (2 C, s), 116.03
(2 C, br s), 113.43 (1 C, d, J = 22.89 Hz), 113.39 (1 C, d, J =
22.89 Hz), 49.68 (2 C, s), 44.81 (1 C, br s), 44.65 (2 C, s),
41.17 (1 C, s), 35.12 (1 C, s), 31.32 (1 C, s), 20.97 (2 C, br s),
20.71 (2 C, s), 13.12 (1C,s), 11.85 (1 C, s). For the 19F
spectrum of 2 at 27 °C, the resonances are distinguishable
between the two rotational isomers. 19F NMR (471 MHz,
DMSO-d6) δ −116.12 (1 F, t, J = 7.27 Hz), −116.21 (1 F, t, J =
7.26 Hz). HRMS (ESI): calcd for C28H26FN5O3S (M+ + 1)
532.18132. Found: 532.18144.
Differential Scanning Calorimetry (DSC). DSC measure-
ments were performed using a TA Instruments DSCQ1000
under a 50 mL/min N2 purge. Samples in the weight range of
2−5 mg were scanned at 10 °C/min in hermetically sealed
aluminum pans that were pierced with a pin to allow escape of
residual solvent and other volatiles. The temperature and heat
flow were calibrated using indium. Both the melting (Tm) and
L
DOI: 10.1021/acs.oprd.6b00035
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
Gutier
́
rez, J. M.; Rodríguez, F. Org. Lett. 2007, 9, 2027−2030.
glass transition temperatures (Tg) were reported as the onsets
of the thermal transitions.
(d) Shirakawa, S.; Kobayashi, S. Org. Lett. 2007, 9, 311−314. Mayr, H.;
Bug, T.; Gotta, M. F.; Hering, N.; Irrgang, B.; Janker, B.; Loos, R.;
Ofial, A. R.; Remennikov, G.; Schimmel, H. J. Am. Chem. Soc. 2001,
123, 9500−9512.
Powder X-ray Diffraction (PXRD). PXRD data was
obtained using a Bruker C2 GADDS. The radiation was Cu
Kα (40 kV, 40 mA). The sample−detector distance was ∼15
cm. Powder samples were packed in sealed glass capillaries of 1
mm or less in diameter; suspension samples were centrifuged
into capillaries of this size. The capillary was rotated during data
collection. Data was collected in the range 2 ≤ 2θ ≤ 35° with a
sample exposure time of at least 1000 s. The resulting two-
dimensional diffraction arcs were integrated to create a
traditional 1-dimensional PXRD pattern with a step size of
0.05 degrees 2θ in the approximate range of 2−35 degrees 2θ.
(7) Miyaura, N.; Suzuki, A. Chem. Rev. 1995, 95, 2457−2483.
(8) This process was shown to be applicable to several 2-halogenated
pyridines in combination with multiple aromatic aldehydes.
(a) Donaldson, M. C.; Tymonko, S.; Conlon, D. A., unpublished
results. See the following for alternate methods: (b) Li, J.; Jin, C.; Su,
W. Heterocycles 2010, 81, 255−2567. (c) Marquise, N.; Harford, P. J.;
Chevallier, F.; Roisnel, T.; Dorcet, V.; Gagez, A.-L.; Sable,
́
S.; Picot, L.;
Thiery, V.; Wheatley, A. E. H.; Gros, P. C.; Mongin, F. Tetrahedron
́
2013, 69, 10123−10133. (d) Dineen, T. A.; Chen, K.; Cheng, A. C.;
Derakhchan, K.; Epstein, O.; Esmay, J.; Hickman, D.; Kreiman, C. E.;
Marx, I. E.; Wahl, R. C.; Wen, P. H.; Weiss, M. M.; Whittington, D. A.;
Wood, S.; Fremeau, R. T., Jr.; White, R. D.; Patel, V. F. J. Med. Chem.
2014, 57, 9811−9831.
AUTHOR INFORMATION
Corresponding Author
*E-mail: david.conlon@bms.com.
■
(9) (a) Radinov, R.; Chanev, C.; Khaimova, M. J. Org. Chem. 1991,
56, 4793−4796. (b) Marzi, E.; Bigi, A.; Schlosser, M. Eur. J. Org. Chem.
Present Address
∥N.d.M.: Biologics Development, Bristol-Myers Squibb, 38
Jackson Road, Devens, Massachusetts 01434, United States.
2001, 2001, 1371−1376. (c) Mongin, F.; Queg
́
uiner, G. Tetrahedron
ourt, F.;
uiner, G.; Marsais, F. Tetrahedron Lett. 2004, 45, 7873−7877.
2001, 57, 4059−4090. (d) Awad, H.; Mongin, F.; Trec
́
Queg
́
Notes
(10) (a) Comins, D. L.; Brown, J. D. Tetrahedron Lett. 1981, 22,
4213−4216. (b) Roschangar, F.; Brown, J. C.; Cooley, B. E., Jr.; Sharp,
M. J.; Matsuoka, R. T. Tetrahedron 2002, 58, 1657−1666.
(11) See ref 4.
⊥E.S.: deceased.
The authors declare no competing financial interest.
(12) Surprisingly, boron trifluoride etherate provided very low
conversion and multiple unidentified side products. HBF4 and lithium
perchlorate were also screened and failed to provide good conversion.
(13) Riabinin, A. I.; Doroshenko, G. A. Ukr. Khim. Zh. (Russ. Ed.)
1974, 40, 462−464.
ACKNOWLEDGMENTS
■
We thank Dr. Brian He for developing the HPLC analytical
methods and Dr. Chiajen Lai for performing the spray drying
studies to prepare amorphous materials. We also acknowledge
Madhushree Gokhale, Ling Gao, and Victor Guarino for helpful
discussions and the Chemical & Synthetic Development senior
management for support during the preparation of the
manuscript.
(14) A crystallization with cinconine failed to provide an
enantiomeric enrichment.
(15) de Mas, N.; Natalie, K. J., Jr.; Quiroz, F.; Rosso, V. W.; Chen, D.
C.; Conlon, D. A. Org. Proc. Res. Dev. 2016, 19, in press
DOI: 10.1021/acs.oprd.6b00036.
(16) Zhu, J.; Razler, T. M.; Xu, Z.; Conlon, D. A.; Sortore, E. W.;
Fritz, A. W.; Demerzhan, R.; Sweeney, J. T. Org. Process Res. Dev. 2011,
15, 438−442.
(17) (a) Itoh, T.; Mase, T. Tetrahedron Lett. 2005, 46, 3573−3577.
(b) Guram, A. S.; King, A. O.; Allen, J. G.; Wang, X.; Schenkel, L. B.;
Chan, J.; Bunel, E. E.; Faul, M. M.; Larsen, R. D.; Martinelli, M. J.;
Reider, P. J. Org. Lett. 2006, 8, 1787−1789. (c) Billingsley, K. L.;
Anderson, K. W.; Buchwald, S. L. Angew. Chem., Int. Ed. 2006, 45,
3484−3488.
(18) Kudo, N.; Perseghini, M.; Fu, G. C. Angew. Chem., Int. Ed. 2006,
45, 1282−1284.
(19) Christmann, U.; Vilar, R. Angew. Chem., Int. Ed. 2005, 44, 366−
374.
(20) Martin, R.; Buchwald, S. L. Acc. Chem. Res. 2008, 41, 1461−
1473.
(21) Camp, D.; Matthews, C. F.; Neville, S. T.; Rouns, M.; Scott, R.
W.; Truong, Y. Org. Process Res. Dev. 2006, 10, 814−821.
(22) Crystallization of a supercooled liquid (e.g., amorphous material
or glass) requires both molecular mobility and thermodynamic driving
force (i.e., supercooling). Molecular mobility is low at temperatures
below Tg. Given the small difference between Tg and Tm of 1 and 2,
there may not be sufficient supercooling when the temperature is
above Tg but below Tm.
REFERENCES
■
(1) (a) Schacke, H.; Berger, M.; Rehwinkel, H.; Asadullah, K. Mol.
̈
Cell. Endocrinol. 2007, 275, 109−117. (b) Stahn, C.; Lowenberg, M.;
̈
Hommes, D. W.; Buttgereit, F. Mol. Cell. Endocrinol. 2007, 275, 71−
78. (c) Schacke, H. Endocrine Abstracts 2010, 21, S4.4. (d) Joanny, E.;
̈
Ding, Q.; Gong, L.; Kong, P.; Saklatvala, J.; Clark, A. R. Br. J.
Pharmacol. 2012, 165, 1124−1136.
(2) (a) Schacke, H.; Docke, W.-D.; Asadullah, K. Pharmacol. Ther.
̈
̈
2002, 96, 23−43. (b) Schacke, H.; Schottelius, A.; Docke, W.-D.;
̈
̈
Strehlke, P.; Jaroch, S.; Schmees, N.; Rehwinkel, H.; Hennekes, H.;
Asadullah, K. Proc. Natl. Acad. Sci. U. S. A. 2004, 101, 227−232.
(c) Rosen, J.; Miner, J. N. Endocr. Rev. 2005, 26, 452−464. (d) Rhen,
T.; Cidlowski, J. A. N. Engl. J. Med. 2005, 353, 1711−1723. (e) Quax,
R. A. M.; Peeters, R. P.; Feelders, R. A. Endocrinology 2011, 152,
2927−2929.
̀
(3) (a) Vayssiere, B.; Dupont, S.; Choquart, A.; Petit, F.; Garcia, T.;
Marchandeau, C.; Gronemeyer, H.; Resche-Rigon, M. Mol. Endocrinol.
1997, 11, 1245−1255. (b) Song, I.-H.; Gold, R.; Straub, R. H.;
Burmester, G.-R.; Buttgereit, F. J. Rheumatol. 2005, 32, 1199−1207.
(c) Hu, X.; Du, S.; Tunca, C.; Braden, T.; Long, K. R.; Lee, J.; Webb,
E. G.; Dietz, J. D.; Hummert, S.; Rouw, S.; Hegde, S. G.; Webber, R.
K.; Obukowicz, M. G. Endocrinology 2011, 152, 3123−3134.
(4) Weinstein, D. S.; Gong, H.; Doweyko, A. M.; Cunningham, M.;
Habte, S.; Wang, J. H.; Holloway, D. A.; Burke, C.; Gao, L.; Guarino,
V.; Carman, J.; Somerville, J. E.; Shuster, D.; Salter-Cid, L.; Dodd, J.
H.; Nadler, S. G.; Barrish, J. C. J. Med. Chem. 2011, 54, 7318−7333.
(5) Villani, F. J.; Mann, T. A.; Wefer, E. A.; Hannon, J.; Larca, L. L.;
Landon, M. J.; Spivak, W.; Vashi, D.; Tozzi, S.; Danko, G.; del Prado,
M.; Lutz, R. J. Med. Chem. 1975, 18, 1−8.
(6) (a) Mukaiyama, T.; Banno, K.; Narasaka, K. J. Am. Chem. Soc.
1974, 96, 7503−7509. (b) Prashad, M.; Lu, Y.; Repic,
̌
O. J. Org. Chem.
́
2004, 69, 584−586. (c) Sanz, R.; Miguel, D.; Martinez, A.; Alvarez-
M
DOI: 10.1021/acs.oprd.6b00035
Org. Process Res. Dev. XXXX, XXX, XXX−XXX