Kilogram synthesis of a second-generation LFA-1/ICAM inhibitor

Article abstract of DOI:10.1021/op100225g

The process development and the kilogram-scale synthesis of BMS-688521 (1) are described. The synthesis features a highly efficient telescoped sequence which utilizes previously described spirocyclic hydantoin (4b) to produce the final intermediate via an

Full text of DOI:10.1021/op100225g

Organic Process Research & Development 2011, 15, 64–72
Kilogram Synthesis of a Second-Generation LFA-1/ICAM Inhibitor
Albert J. DelMonte,* Yu Fan, Kevin P. Girard, Gregory S. Jones, Robert E. Waltermire, Victor Rosso, and Xuebao Wang
Process Research and DeVelopment, Bristol-Myers Squibb Company, One Squibb DriVe, P.O. Box 191, New Brunswick,
New Jersey 08903-0191, United States
Abstract:
Results and Discussions
The original synthesis of 1 by our Discovery colleagues
provided the initial quantities required for some of the preclinical
development studies (Scheme 1).4 The generation of tert-butyl
nicotinic ester 3 was achieved by formation of the acid chloride
via the use of thionyl chloride and subsequent treatment with
tert-butanol. Following the workup, the reaction mixture was
dried with Na2SO4 and concentrated via rotovap to afford 3 as
a yellow solid. Treating the spirocyclic hydantoin 4a with a
slight excess of 3 at 112 °C for 18 h in the presence of DIPEA
afforded the desired N-alkylated product 5a. Addition of a
solution of 5a into cold H2O resulted in precipitation of a crude
solid which was subsequently chromatographed to remove
1-3% of the isopropyl ester analogue 6 (Figure 1). Impurity 6
resulted from low levels of isopropanol present in the tert-
butanol used to prepare 3.5 Purified 5a was then deprotected
utilizing a 1:1 trifluoroacetic acid:dichloromethane solvent
system, and after an extractive workup, was concentrated to
dryness. The crude material was subsequently dissolved in hot
chloroform and the desired product 1 was isolated in two crops.
Unfortunately residual chloroform (0.04%) made this material
unsuitable for toxicology studies.
A modified synthesis of 1 (Scheme 2) was developed by
our Discovery scale-up group colleagues to address the iso-
propyl ester impurity as well as the residual chloroform issue
in order to supply multigram quantities for preclinical develop-
ment studies.6 In this modified approach, the 6-chloronicotinic
acid 2 was protected in situ with TMSCl, and the resulting
intermediate 7 was coupled with hydantoin 4a. The reaction
mixture containing intermediate 8 was treated with MeOH, and
crude 1 was directly isolated from the reaction mixture upon
addition of water. Crude 1 was recrystallized from EtOH/H2O
to afford the desired 1 in high purity and yield (99.5 HPLC
area percent purity, >99.9% ee, and 89% overall yield).
While the modified approach was attractive for producing
multigram quantities of API quickly, we had concerns that the
extensive telescope which affords 1 directly provides little
control over any potential impurities that might be introduced
from the raw materials. From an impurity control perspective,
The process development and the kilogram-scale synthesis of BMS-
688521 (1) are described. The synthesis features a highly efficient
telescoped sequence which utilizes previously described spirocyclic
hydantoin (4b) to produce the final intermediate via an SNAR
reaction. A final deprotection step affords BMS-688521 (1) in high
quality with an overall yield of 65% from the key intermediate,
spirocyclic hydantoin (4b).
Introduction
A key feature of the intercellular immune response is that
cytokine stimulated cells express LFA-1 (leukocyte function-
associated antigen-1) on their surface. LFA-1 then interacts with
ICAM (intercellular adhesion molecule), which is found on the
surfaces of both leukocytes and endothelium. This interaction
facilitates T-cell adhesion and migration through the blood
vessel wall to the inflamed area.1 Small molecules which
successfully inhibit the LFA-1/ICAM interaction have potential
as drugs for the treatment of a variety of autoimmune and
inflammatory diseases such as rheumatoid arthritis and psoriasis.2,3
The LFA-1 receptor antagonist, BMS-688521, 1, was a second
generation molecule selected for clinical development and we
required a synthesis that would reliably generate kilogram
quantities of API. This contribution details the identification
and development of a synthesis which enabled the realization
of this goal.
* Author for correspondence. E-mail: albert.delmonte@bms.com.
(1) For a discussion on the inhibition of LFA-1/ICAM-1as an approach
to treating autoimmune diseases see: Yusuf-Makagiansar, H.; Ander-
son, M. E.; Yakovleva, T. V.; Murray, J. S.; Siahaan, T. J. Med. Res.
ReV. 2002, 22, 146.
(2) For a discussion of therapeutic options for treatment of psoriasis, see:
Gottlieb, A. B. J. Acad. Dermatol 2005, 53, S3. Larson, R. S.; Davis,
T.; Bologa, C.; Semenuk, G.; Vijayan, S.; Li, Y.; Oprea, T.; Chigaev,
A.; Buranda, T.; Wagner, C. R.; Sklar, L. A.
(3) For other small-molecule LFA-1/ICAM-1 antagonists as potential drugs
please see: (a) Pei, Z.; Xin, Z.; Liu, G.; Li, Y.; Reilly, E. B.; Lubbers,
N. L.; Huth, J. R.; Link, J. T.; von Geldern, T. W.; Cox, B. F.; Leitza,
S.; Gao, Y.; Marsh, K. C.; DeVries, P.; Okasinski, G. F. J. Med. Chem.
2001, 44, 2913. (b) Liu, G.; Huth, J. R.; Olejniczak, E. T.; Mendoza,
R.; DeVries, P.; Leitza, S.; Reilly, E. B.; Olasinski, G. F.; Fesik, S. W.;
von Geldern, T. W. J. Med. Chem. 2001, 44, 1202. (c) Wu, J.-P.;
Emeigh, J.; Gao, D. A.; Goldberg, D. R.; Kuzmich, D.; Miao, C.;
Potocki, I.; Qian, K. C.; Sorcek, R. J.; Jeanfavre, D. D.; Kishimoto,
K.; Mainolfi, E. A.; Nabozny, G.; Peng, C.; Reilly, P; Rothlein, R.;
Sellati, R. H.; Woska, J. R.; Chen, S.; Gunn, J. A.; O’Brien, D.; Norris,
S. H.; Kelly, T. A. J. Med. Chem. 2004, 47, 5356. (d) Last-Barney,
K.; Davidson, W.; Cardozo, M.; Frye, L. L.; Grygon, C. a.; Hopkins,
J. L.; Jeanfavre, D. D.; Pav, S.; Qian, C.; Stevenson, J. M.; Tong, L.;
Zindell, R.; Kelly, T. A. J. Am. Chem. Soc. 2001, 123, 5643. (e) Wang,
G. T.; Wang, S.; Gentles, R.; Sowin, T.; Leitza, S.; Reilly, E. B.; von
Geldern, T. W. Bioorg. Med. Chem. Lett. 2005, 15, 195. (f) Wattanasin,
S.; Albert, R.; Ehrhardt, C.; Roche, D.; Savio, M.; Hommel, U.;
Welzenbach, K.; Weitz-Schmidt, G. Bioorg. Med. Chem. Lett. 2003,
12, 499.
(4) Watterson, S. H.; Xiao, Z.; Dodd, D. S.; Tortolani, D. R.; Vaccaro,
W.; Potin, D.; Launay, M.; Stetsko, D. K.; Skala, S.; Davis, P. M.;
Lee, D.; Yang, X.; McIntyre, K. W.; Balimane, R.; Patel, K.; Yang,
Z.; Marathe, P.; Kadiyala, P.; Tebben, A. J.; Sheriff, S.; Chang,
C. Y. Y.; Ziemba, T.; Zhang, H.; Chen, B.-C.; DelMonte, A. J.;
Aranibar, N.; McKinnon, M.; Barrish, J. C.; Suchard, S. J.; Dhar,
T. G. M. J. Med. Chem. 2010, 53, 3814.
(5) Due to the higher level of reactivity of isopropanol relative to that of
tert-butanol, the isopropyl impurity 6 is enriched to a higher level
than the level of isopropanol observed in the tert-butanol.
(6) Zhang, H.; Watterson, S. H.; Xiao, Z.; Dhar, T. G. M.; Balasubra-
manian, B.; Barrish, J. C.; Chen, B.-C. Org. Process Res. DeV. 2010,
14, 936.
64
•
Vol. 15, No. 1, 2011 / Organic Process Research & Development
10.1021/op100225g  2011 American Chemical Society
Published on Web 11/29/2010

Scheme 1. Discovery synthesis of 1a
a Reagents and conditions: (a) SOCl2; t-BuOH, CH2Cl2, Et3N, DMAP, 70%; (b) 3, DIPEA, DMA, 112 °C, 77%; (c) TFA, DCM, 86%.
adiabatic temperature change of -37 °C. During the tert-butyl
ester formation there is potential for the generation of 4.5 equiv
of CO2 gas. Figure 2 shows the off-gassing profile for an
experiment utilizing 53.0 g of 2. A total of 32.6 L of CO2 was
measured, corresponding to 88% of the theoretical amount.
Interestingly, the gas evolution consisted of several phases. The
first was a non-dose-controlled release (∼0.6 L/min) with an
induction period, followed by a dose-controlled release (∼1.2
L/min) during addition of the first equivalent of Boc2O solution.9
The rate of gas evolution decreased during the addition of the
final 1.25 equiv of Boc2O. Based on this data, the maximum
off-gassing rate for a batch utilizing 1.8 kg of 2 as input was
calculated to be 41 L/min for a similar Boc2O solution feed
rate. This off-gassing rate was well within the design basis of
the process equipment planned for use in our pilot plant.10
To address the non-dose-controlled release of CO2 at the
beginning of the reaction, an initial charge of ∼10% of the
Boc2O/THF solution was performed. After the initial gas
evolution phase was over, the remainder of the Boc2O/THF
solution was added in an addition-controlled fashion. The higher
temperature of the reaction (55-60 °C) combined with the
subsequent aqueous alkaline washes ensured very little dissolved
CO2 would be present after the workup.
With confidence in the safety of the reaction, our efforts were
then focused on the isolation protocol. Our initial attempts to
isolate 3 as a solid only afforded amorphous materials. Eventu-
ally, through the use of high-throughput crystallization experi-
ments, we found that crystalline material could be obtained from
an IPA/water mixture. However, we quickly realized that drying
and handling of crystalline 3 on large scale would pose a
challenge due to its low melting point (54 °C). While much of
the early development work for the subsequent SNAR step was
Figure 1. Impurity observed during the esterification of 5a.
we considered the original Discovery approach, in which 5a
was isolated and deprotected in a final step, to be a more
attractive strategy. However, the pursuit of this approach
required addressing the specific issues of the isopropyl ester
impurity and the identification of a chloroform-free crystalliza-
tion process to the API. In general, the identification of a robust
and scalable process required additional efforts to determine
the optimal reagents,7 reaction conditions, and isolation proce-
dures, with focus on avoiding chromatography, isolation of
amorphous solids, recrystallizations, and two-crop isolations.
Preparation and Isolation of tert-Butyl Nicotinic Ester
(3). An alternative approach for the preparation of 3 which
completely circumvents the issue of the isopropanol contamina-
tion of tert-butanol was identified. A variant of the previously
reported procedure8 utilizing 2.25 equiv of di-tert-butyl dicar-
bonate (Boc2O) in THF with catalytic (7 mol %) dimethylami-
nopyridine (DMAP) afforded conversion of 6-chloronicotinic
acid 2 to the desired tert-butyl protected 3 with very low
impurity levels (Scheme 3). A basic extraction was utilized to
quench residual Boc2O and remove low levels of unreacted
6-chloronicotinic acid 2. To safely scale this process we
conducted a detailed study to fully understand both the heat
flow as well as the rate of CO2 off-gassing. No unusual
thermochemical properties were observed, and calorimetry
indicated the reaction was endothermic (155 KJ/mol) with an
(9) To the best of our knowledge, this mechanism has not been studied
in detail. Our hypothesis regarding the induction period is that at the
beginning of the reaction the majority of the DMAP is not initially
available to enter the catalytic cycle as it forms a salt with the nicotinic
acid.
(10) This calculation must be performed for any changes to scale or
equipment to ensure the equipment can tolerate the maximum off-
gassing rate. In practice, 2.3 equiv of BOC2O was utilized on scale-
up with a molar feed rate of 0.163 mol/min. Since the gas evolution
was addition controlled, the maximum potential offgassing rate was
∼7.4 L/min.
(7) Identification of alternative reagents for the tert-butyl ester formation
was desired to avoid the use of the very odiferous thionyl chloride.
Ideally, an alternative procedure would also obviate the need to source
ultra-pure tert-butanol which was free of isopropanol.
(8) Goossen, L. J.; Do¨hring, A. Synlett 2004, 2, 263.
Vol. 15, No. 1, 2011 / Organic Process Research & Development
•
65

Scheme 2. Discovery-modified synthesis of 1a
a Reagents and conditions: (a) (CH3)3SiNHSi(CH3)3, cat. TMSCl, 78 °C, 3 h, 100%; (b) 4a, DMAP, DMA, DIPEA, 95 °C, 18 h; (c) MeOH, 25 °C, 2 h, 98%.
Scheme 3. Synthesis of tert-butyl nicotinic ester
performed with isolated 3, its low melting point was an intrinsic
liability which could not be avoided and would complicate all
handling efforts. This suggested that the best approach ultimately
would involve a telescoped process in which 3 was not isolated,
but rather directly coupled with spirocyclic hydantoin 4a.
Early Development of the Penultimate Step. The free base
spirocyclic hydantoin 4a was initially utilized in the Discovery
synthesis for the synthesis of gram-scale quantities of API. We
planned to ultimately use the hemi (+)-di-p-toluyl-D-tartaric acid
((+)-DTTA) salt 4b directly as it was readily available from a
prior large-scale synthesis.11 However, to simplify our initial
lab development work, we initially used the free base spirocyclic
hydantoin 4a. To access gram quantities of the freebase 4a,
the DTTA salt 4b was dissolved in 10 mL/g THF/MTBE (1:1)
and extracted with a 5 mL/g of 1 M NaOH solution to remove
the 0.5 equiv of the DTTA.12 Following a 5 mL/g aqueous wash,
the organic layer was concentrated, and 4a was isolated as the
free base (Scheme 4).
of 4a on DMA was also observed (Figure 3).13 NMP was
identified as an ideal solvent as it afforded reactions free of
impurity 9. Further development work indicated that low water
content of the NMP (KF < 0.3%) was critical as higher water
levels afforded increased levels of impurity 10 resulting from
epimerization of the benzylic center (Figure 3).14 We also found
1.4 equiv of tert-butyl nicotinic ester 3 was optimal, as this
(11) The DTTA salt 4b was readily available as it was a common
intermediate with BMS-578101, a first-generation molecule. For a
description of its synthesis, see: DelMonte, A. J.; Waltermire, R. E.;
Fan, Y.; McLeod, D. D.; Gao, Z.; Gesenberg, K. D.; Girard, K. P.;
Rosingana, M.; Wang, X.; Kuehne-Willmore, J.; Braem, A. D.;
Castoro, J. Org. Process Res. DeV. 2010, 14, 553.
(12) Throughout this publication mL/g refers to mL of solvent to grams of
input material.
(13) The same impurity was also observed during the Discovery develop-
ment work (see ref. 6).
(14) With similar substrates (see ref 11), it was observed that the benzylic
center could epimerize under basic conditions. To test this, ∼50 mg
of the penultimate 5a (one diastereomer) was heated in 0.75 mL DMF-
d7 in the presence of 2 equiv of DBU. After 2 h, NMR and HPLC
indicated ∼15% formation of the undesired diastereomer 10. This
amount increased to ∼30% after 4 h of heating.
Addition of solid tert-butyl nicotinic ester 3 to 4a with 2.5
equiv of DIPEA and heating the reaction at 112 °C for 18 h in
dimethylacetamide (DMA) afforded the desired product. How-
ever, the acetyl impurity 9 which results from nucleophilic attack
Figure 2. Lab-scale off-gassing experiment.
66
•
Vol. 15, No. 1, 2011 / Organic Process Research & Development

Scheme 4. Synthesis of 5a from 4b
moderate excess helped drive the reaction to completion.
However, it was clear that we would require a strategy to
remove this excess reagent upon reaction completion in order
to facilitate isolation of the product. In our early development
efforts, a small sample of pure SNAR product 5a was isolated
by chromatography and subjected to a high-throughput salt
screen, resulting in the identification of a crystalline p-
toluenesulfonic acid (p-TSA) salt 5b. With these encouraging
results for each component of the overall reaction sequence,
free-basing of 4b, the SNAR reaction conditions, and the isolation
of product as a crystalline salt 5b, we focused on the develop-
ment of a telescoped process using a solution of 3 and the
DTTA salt 4b.
Development of a Telescoped Process for the Production
of 5b. Since the low melting point of the tert-butyl nicotinic
ester 3 was anticipated to complicate isolation and drying of
crystalline material on scale, it was highly desirable to develop
a telescoped process. One of the primary objectives in develop-
ing an efficient telescoped process is the careful integration of
the steps so the number of unit operations (i.e., extractions,
distillations, etc.) is minimized. We chose to perform the tert-
butyl ester protection under the Boc2O conditions as previously
described, followed by addition of MTBE and aqueous NaOH.
Since a freebasing of the spirocyclic hydantion DTTA salt 4b
to 4a was ultimately required, we saw an opportunity to reduce
the number of overall extractions by charging 4b directly to
the basic mixture.15 After a short period of stirring, we continued
with the extraction. This approach allowed us to quench any
excess Boc2O while simultaneously removing the DTTA and
any residual nicotinic acid 2. The next step was to take this
MTBE/THF solution and solvent exchange to NMP which was
the preferred solvent for the SNAR reaction. While solvent
exchanges can sometimes be time-consuming, this solvent
exchange to a much higher-boiling solvent was quite facile.
More importantly, it afforded a clear processing advantage as
it was effective at affording a very dry solution via azeotropic
water removal from the reaction mixture, thus reducing the
probability of forming impurity 10. The SNAR reaction per-
formed as expected, but we still required a strategy for removing
0.4 equiv of excess 3 from the reaction mixture. We found we
could capitalize on the NMP solvent system and the nonpolar
nature of 3 and simply employ a heptane/cyclohexane extraction
of the NMP solution to remove the majority of the residual
tert-butyl nicotinic ester 3. MTBE was then charged to the
reaction mixture, and the NMP was removed by several aqueous
extractions. After azeotropic drying, p-TSA was charged
followed by isolation of the desired compound 5b (Scheme 5).
This process was performed on a 6-kg scale in two batches,
affording material in good yield and quality (88.7% and 81.8%
yield, with both batches >98.0 HPLC area percent purity).16
The Hydrolysis of 5b to API. The first step was to remove
the p-TSA via a basic extraction of an MTBE/THF mixture
containing 5b. A variety of reagents and reaction conditions
for the ester hydrolysis of 5a were rapidly screened, and
ultimately a mixture of concentrated HCl (0.94 mL/g) and
AcOH (1.75 mL/g) in toluene was selected, affording the
hydrolyzed product in 3-4 h at 55-65 °C. The reaction mixture
was washed with aqueous sodium chloride and sodium hy-
droxide and subsequently solvent exchanged to EtOH. Seeding
and addition of heptane as an antisolvent afforded 1 as the
desired neat form (Scheme 6). Rapid development of this
process and execution on a 5-kg scale afforded 2.2 kg of API
in 66.2% yield, >99.9% ee and 98.0 HPLC area percent purity
with a D90 ) 36 µm (after milling).17
During the 5-kg scale-up we identified two opportunities for
improvement. The first was related to the purity of the product
1 (98.0 AP) which was lower than desired. While several
impurities (11-13) were previously observed in our lab runs
(Figure 4), the tert-butylamide impurity 11 was present at a
much higher level (1 AP) than desired. Our hypothesis is that
impurity 11 results from a Ritter reaction18 in which there is
nucleophilic attack of the cyano nitrogen on the isobutyl cation
byproduct from the acidic deprotection of the tert-butyl ester.
The second opportunity for improvement was related to the
crystal morphology and particle size of isolated 1. The current
procedure involved a solvent exchange to EtOH generating a
supersaturated solution. Uncontrolled crystallization afforded
a previously unseen crystal morphology which, after SEM
(Figure 5), was described as hard, round agglomerates with a
D90 of ∼200 µm. Either due to their hardness or spherical
shape, our attempts to significantly reduce the particle size by
(15) The successful quench of Boc2O was evidenced by absence of the
tert-butyl carboxylate of 4a which could be observed if 4b was charged
to the mixture immediately.
(16) Yields reported are corrected yields, and the HPLC area percent purity
reported is the combined HPLC area percent purity for the spirocyclic
hydantoin 5a and p-TSA.
(17) A D90 < 60 µm was desired.
(18) For a review on the Ritter reaction see: Krimen, L. I.; Cota, D. J.
Org. React. 1969, 17, 213.
(19) The high loss of 1 to mother liquor may have been due to the
uncontrolled crystallization, the new morphology or the higher than
previously observed level of impurities. It is notable that pXRD
indicated 1 had the correct polymorphic form with little to no
amorphous content.
Figure 3. Impurities formed in the SNAR step.
Vol. 15, No. 1, 2011 / Organic Process Research & Development
•
67

Scheme 5. Telescoped synthesis of 5b
Scheme 6. Hydrolysis of 5b to 1
passing through a wet mill were unsuccessful. Even after two
passes the D90 was only reduced to 130 µm. However, dry
milling with a hammer mill was successful and after two passes
a D90 of 36 µm was obtained. In addition to the unexpected
loss of product due to dry milling (4.9%), there were also higher
than anticipated quantities of 1 left in the mother liquor
(∼24.5%) resulting in the lower than expected yield.19 Before
proceeding with the next batch, we worked rapidly in the
laboratory to increase our process knowledge to enable greater
control of the process.
impurity, we utilized DOE experimentation to investigate the
reaction parameters and their interdependent effects on conver-
sion and purity levels. We identified temperature, HCl concen-
tration, and the two-way interaction between these factors as
playing the biggest roles on reaction rate with increased
temperature and increased HCl concentration affording the
highest rate of conversion to product. These results are easily
visualized through the use of a Pareto plot which uses a
histogram to rank the variables from most impactful to least
impactful (Figure 6). We then ran a central composite DOE
which highlighted the effect of water concentration and HCl
concentration on the impurity formation with focus on the tert-
Development Work on the Hydrolysis Reaction. To
address the higher than desired level of the tert-butyl amide
Figure 4. Impurities formed in the hydrolysis step.
Figure 5. SEM of premilled API (agglomerates and primary particles).
68
•
Vol. 15, No. 1, 2011 / Organic Process Research & Development

Figure 6. DOE on the hydrolysis of 5b.
Figure 7. Central composite DOE (effect of HCl and H2O on impurities 11, 12, and 13).
butyl amide impurity. A much cleaner reaction profile was
obtained with increased water concentration and decreased HCl
concentrations (Figure 7). Further lab work suggested modified
conditions using AcOH (3.6 mL/g), HCl (0.91 mL/g), and H2O
(0.63 mL/g) in toluene at 30 °C. We were pleased to observe
that under these new conditions, all of the impurities, including
the tert-butyl amide impurity 11, could be controlled to levels
below 0.25 HPLC area percent purity (Figure 8).
Optimization of Particle Size and Morphology of 1. To
better control the crystal morphology, we conducted further
crystallization studies. During our lab work a new morphology
was obtained in which the crystals formed fragile rods. We were
very pleased when analysis by pXRD indicated we had still
obtained the desired polymorphic form. By using the identical
crystallization conditions previously employed, but seeding with
crystals possessing this new morphology, we were able to
generate fragile rods which were easily wet-milled and filtered
well. This obviated the need for dry milling and the material
loss associated with the additional handling.
Figure 8. Hydrolysis reaction profile under optimized conditions.
with improved control. The second batch afforded 1 in 84.5%
yield and 99.3 HPLC area percent purity, demonstrating
significant improvements over batch 1 (66% yield and 98.0
HPLC area percent purity). In addition, the particle size was
well controlled by wet milling with a D90 of 28 µm with no
dry milling required, and chiral HPLC analysis indicated the
material was >99.9% ee.
Results from Batch 2. Incorporating all of these modifica-
tions allowed us to execute the second batch on 4.95 kg scale
Vol. 15, No. 1, 2011 / Organic Process Research & Development
•
69

Scheme 7. Efficient two-step synthesis of 1
Conclusions
at 60 °C to a volume of 25 L. MTBE (26.5 kg) was charged,
and the mixture was distilled to a volume of 25 L. If the KF >
2.7%, MTBE was charged, and the distillation was repeated.
NMP (21.4 kg), 16.1 kg of cyclohexane, and 1.2 kg of DIPEA
was charged. The reaction mixture was distilled until the
temperature reached 120 °C and no more THF or MTBE was
collected. The mixture was held at reflux, and 1.8 kg of DIPEA
was charged (total DIPEA charge of 3.0 kg). The mixture was
held at reflux until full conversion was achieved (<3% of 4a
remaining), and then the mixture was cooled to 25 °C. To the
reactor was charged 14.2 kg of heptane and 1.82 kg of
cyclohexane. The mixture was agitated 15 min and allowed to
split for 30 min, and the top heptane/cyclohexane layer was
discarded. To the mixture was charged 14.2 kg of heptane and
1.82 kg of cyclohexane. The mixture was agitated 15 min and
allowed to split 30 min, and the top heptane/cyclohexane layer
was discarded. To the mixture was charged 23.4 kg of MTBE
and 20.5 kg of H2O. The mixture was agitated 15 min and
allowed to split 30 min, and the bottom aqueous layer was
discarded. To the mixture was charged 5.2 kg of MTBE and
20.5 kg of H2O. The mixture was agitated 15 min and allowed
to split 30 min, and the bottom aqueous layer was discarded.
To the reactor was charged 20.5 kg of water, the mixture was
agitated 15 min, the phases were allowed to split for 30 min,
and the bottom aqueous layer was discarded. The mixture was
filtered through a ZetaCarbon Pad (04701-R53SP) to remove
color and particles. The pad and lines were rinsed with 8.1 kg
of MTBE. The mixture was distilled at 55-62 °C to a volume
of 30 L. (Note: If KF > 0.3%, recharge 25.1 kg of MTBE and
distill to a volume of 30 L). To the reaction was charged 16.8
kg of MTBE and 8.6 kg of IPA. A solution of 1.45 kg of p-TSA
and 17.3 kg of IPA was prepared and charged slowly to the
reaction mixture. A seed mixture (29 g of seeds suspended in
0.5 kg of MTBE) was charged and the slurry was stirred for
30 min. The cake was filtered and washed in two portions with
a solution of 7.1 kg of IPA and 6.8 kg of MTBE. A total of
5.59 kg of 5b was obtained in >98 HPLC area percent purity
and 88.7% yield as a white solid: [R] +58.1 (c 0.439, DMSO);
IR (KBr) 3434 (s), 2974 (s), 1725 (s), 1650 (s), 1455 (s), 1307
(s), 1220 (s), 1163 (s), 1123 (s), 1039 (m) cm-1; 1H NMR (400
MHz, DMF) δ 8.74 (d, J ) 1.8, 1H), 8.15-8.25 (m, 1H), 7.96
(d, J ) 8.3 Hz, 2H), 7.67 (dd, J ) 12.0, 8.3, 4H), 7.62 (t, J )
1.8, 1H), 7.18 (d, J ) 12.0, 2H), 7.06 (dd, J ) 8.3, 1H), 7.02
(d, J ) 1.9, 2H), 4.25-4.62 (m, 5H), 3.80-3.93 (m, 1H), 3.03
(s, 3H), 2.32 (s, 3H), 1.59 (s, 9H); 13C NMR (100 MHz, DMF)
δ 171.7, 164.14, 162.67, 162.57, 162.38, 162.08, 153.76, 145.36,
139.55, 139.08, 134.67, 134.05, 132.98, 129.74, 128.66, 128.23,
126.12, 125.02, 118.62, 116.91, 112.11, 81.30, 71.93, 62.87,
In summary, we have described an efficient, 2-step tele-
scoped synthesis of BMS-688521 (1) from an available
intermediate, spirocyclic hydantoin 4b, as shown in Scheme 7.
A highly efficient telescoped process was utilized to avoid
handling and drying of a low-melting point solid as well as
minimizing the overall number of unit operations. Only general
purpose processing equipment was required, providing versatil-
ity in future manufacturing site selection. The overall process
has been demonstrated on kilogram scale and is amenable for
commercial-scale manufacture.
Experimental Section
All reactions were performed under a nitrogen atmosphere.
All reagents purchased from vendors were used as received
unless otherwise indicated. Chiral analysis was conducted on a
Shimadzu HPLC with a Chiracel AD or AD-H column unless
stated otherwise. Proton and carbon NMR were run on a Bruker
AC-300 at 300 MHz for proton and 75 MHz for carbon or on
a Bruker AVANCE 400 at 400 MHz for proton and 100 MHz
for carbon. The melting points were obtained with a Mettler-
Toledo FP 62 melting point instrument by measurement of the
change of luminous intensity during the melting process.
Reported yields have not been corrected for impurity levels or
moisture content.
Preparation of tert-Butyl 6-((5S,9R)-9-(4-cyanophenyl)-
3-(3,5-dichlorophenyl)-1-methyl-2,4-dioxo-1,3,7-
triazaspiro[4.4]nonan-7-yl)nicotinate p-TSA Salt (5b). To a
reactor was charged 1.83 kg of 2 (11.62 mol, 1.00 equiv), 0.10
kg of DMAP (0.82 mol, 0.07 equiv) and 10.7 kg of THF. The
reactor was heated to 62.5 °C. In a separate vessel the Boc2O
solution was prepared by dissolving 5.82 kg of Boc2O (26.67
mol, 2.3 equiv) in 8.35 kg of THF. Approximately 10% of the
Boc2O solution was charged to the mixture of 2 at a rate of 44
g/min (addition time of ∼0.5 h). The mixture was held for 40
min at 62.5 °C, and the remaining Boc2O solution was charged
at a rate of 87 g/min (addition time of ∼2.5 h). The reaction
was held for 2 h at 62.5 °C, and the conversion of 2 to 3 was
monitored. Upon reaction completion, the mixture was cooled
to 21 °C, and a mixture of THF/H2O (3.9 kg:0.83 kg) was
charged slowly to control gas evolution. The mixture was stirred
1 h, 28.7 kg of 1 M NaOH was charged, and the mixture was
held for 20 min. To the reactor was charged 4.7 kg of 4b and
20.4 kg of MTBE. The mixture was agitated 15 min and
allowed to split 30 min, and the aqueous layer was discarded.
A charge of 27.7 kg of H2O was performed followed by
agitation for 15 min. The phases were allowed to split 30 min,
and the aqueous layer was discarded. The mixture was distilled
70
•
Vol. 15, No. 1, 2011 / Organic Process Research & Development

49.45, 48.59, 46.34, 27.85, 25.41, 25.16, 20.71; Anal. Calcd
for C37H35Cl2N5O7S·0.75H2O: C, 57.10; H, 4.82; N, 9.10;
Found: C, 57.08; H, 4.87; N, 8.91.
<1%. The isolate solids were then subjected to two passes
through a hammer mill and D90 ) 36 µm was achieved.
In total, 2.17 kg of 1 was obtained in 98.0 HPLC area
percent purity, >99.9% ee, and 66.2% yield (corrected for
potency of input and product). The discrepancy between
the observed yield (66.2%) and the expected yield (72%)
was due to losses to the mother liquor (24.5%) as well as
losses during the development and execution of the dry
Preparation of 6-((5S,9R)-9-(4-Cyanophenyl)-3-(3,5-dichlo-
rophenyl)-1-methyl-2,4-dioxo-1,3,7-triazaspiro[4.4]nonan-7-
yl)nicotinic Acid (1). To a reactor was charged 5.0 kg of
5b (6.54 mol), 12.5 kg of THF, 28.4 kg of MTBE, and
26.0 kg of 1 M NaOH. The mixture was agitated 15 min
and allowed to split for 30 min, and the bottom aqueous
layer was discarded. To the reactor was charged 25 kg of
H2O, the mixture was agitated 15 min and allowed to split
for 30 min, and the bottom aqueous layer was discarded.
To the reactor was charged 9.75 kg of toluene, and the
mixture was distilled until the batch temperature reached
85-90 °C. The reaction mixture was cooled to 25 °C and
9.2 kg of concentrated AcOH and 5.6 kg of concentrated
HCl were charged while maintaining the reaction tem-
perature <40 °C. The reaction was heated at 55-65 °C
until <1% of 5b remained. The reaction mixture was
cooled to 5 °C and 26.7 kg THF, 16.8 kg 10 wt % brine
solution was charged. A total of 6.5 kg of 10 N NaOH
was then charged to adjust the pH to 4.5 (Note: AcOH
can be used to lower pH if pH is accidentally adjusted
too high). The mixture was agitated 15 min and allowed
to split for 30 min, and the bottom aqueous layer was
discarded. To the reaction mixture was charged 16.8 kg
of 10 wt % brine solution and 26.0 kg of 2 N NaOH
solution. The mixture was agitated 15 min and allowed
to split for 30 min, and the bottom aqueous layer was
charged. To the mixture was charged 11.25 kg of H2O,
and the pH was adjusted to 4.5 using NaHCO3 solution
(made from 0.94 kg NaHCO3 and 17.8 kg H2O). The
mixture was agitated 15 min and allowed to split for 30
min, and the bottom aqueous layer was discarded. The
batch was distilled until the reaction mixture was 82-90
°C. To the reactor was charged 13.34 kg of THF, and the
KF was measured. (Note: If KF > 0.3%, the distillation
and THF could be repeated). The mixture was cooled to
25 °C and passed through a 10 µm filter into a clean
reactor. The reactor and filter were rinsed with 1 kg of
THF. The mixture was distilled under vacuum until
minimum volume or no more distillate was obtained. To
the reactor was charged 23.7 kg of 200 proof ethanol, and
the mixture was distilled. The ethanol charge and distil-
lation were repeated. (Note: If the toluene content was
>10%, the ethanol charge and distillation could be
repeated.) The total volume was adjusted to 40 L and
cooled to 60 °C. A seed solution (33 g of 1 in 0.8 kg of
heptane and 0.2 kg of 200 proof ethanol) was charged
and the line rinsed with 0.5 kg of 200 proof ethanol. The
mixture was held at 60 °C until a slurry formed and the
antisolvent solution (17.1 kg of heptane and 2.19 kg of
ethanol) was slowly charged. The batch was cooled to 20
°C over 4 h. The slurry was passed through a wet mill
with only minimal change to particle size. The cake was
filtered and washed with 2 portions of wash solution (2.37
kg of ethanol and 18.47 kg of heptane). The solids were
dried in a vacuum oven at 40-50 °C until the LOD was
1
milling step (4.9%). H NMR (400 MHz, DMF) δ 8.81
(d, J ) 2.3, 1H), 8.13 (dd, J ) 8.8, 2.3, 1H), 7.96 (d, J
) 8.3 Hz, 2H), 7.66 (dd, J ) 8.3, 2H), 7.61 (t, J ) 1.9,
1H), 7.02 (d, J ) 2.0, 2H), 6.78 (d, J ) 8.6, 1H), 4.5 (d,
J ) 9.1, 1H), 4.24-4.42 (m, 5H), 4.16 (d, J ) 12.1, 1H);
13C NMR (100 MHz, DMF) δ 171.91, 167.19, 162.66,
162.07, 159.06, 153.81, 151.29, 140.07, 138.49, 134.66,
132.92, 129.76, 128.17, 125.02, 118.66, 115.66, 111.98,
105.97, 72.01, 49.08, 47.85, 46.54, 25.10; Anal. Calcd
for C26H19Cl2N5O4S: C, 58.22; H, 3.57; N, 13.05; Found:
C, 58.02; H, 3.57; N, 12.81.
Modified Protocol for the Preparation of 6-((5S,9R)-9-
(4-cyanophenyl)-3-(3,5-dichlorophenyl)-1-methyl-2,4-dioxo-
1,3,7-triazaspiro[4.4]nonan-7-yl)nicotinic Acid (1). To a
reactor was charged 4.95 kg of 5b (6.47 mol), 12.4 kg of
THF, 28.1 kg of MTBE, and 25.7 kg of 1 M NaOH. The
mixture was agitated 15 min and allowed to split for 30
min, and the bottom aqueous layer was discarded. To the
reactor was charged 24.8 kg of H2O. The mixture was
agitated 15 min and allowed to split for 30 min, and
bottom aqueous layer was discarded. To the reactor was
charged 9.66 kg of toluene, and the solution was distilled
until the batch temperature reached 85-90 °C. The
mixture was cooled to 25 °C, and 3.12 kg of H2O, 18.46
kg of concentrated AcOH, and 5.44 kg of concentrated
HCl were charged, maintaining the reaction temperature
<30 °C. The mixture was heated at 30 °C until <1% of
5b remained. The mixture was cooled to 5 °C, and 26.4
kg of THF and 7.2 kg of 10 wt % brine solution was
charged. The pH was then adjusted to 4.5 with 16.3 kg of
10 N NaOH (Note: AcOH can be used to lower pH if pH
is accidentally adjusted too high). The mixture was
agitated 15 min and allowed to split for 30 min, and the
bottom aqueous layer was discarded. To the reaction
mixture 6.6 kg of 10 wt % brine solution and 25.7 kg of
2 N NaOH solution was charged. The mixture was agitated
15 min and allowed to split for 30 min, and the bottom
aqueous layer was discarded. To the reaction mixture was
charged 11.14 kg of H2O, and the pH was adjusted to 4.5
using NaHCO3 solution (0.928 kg NaHCO3 and 17.63 kg
H2O). The reaction mixture was agitated 15 min and
allowed to split for 30 min, and the bottom aqueous layer
was discarded. The solution was distilled until the reaction
mixture was 82-90 °C. To the reactor was charged 13.2
kg of THF, and the KF was measured. If KF > 0.3% the
distillation and THF charge could be repeated. The mixture
was cooled to 25 °C, passed through a 10 µm filter into
a clean reactor, and the reactor and lines were rinsed with
1 kg of THF. The mixture was distilled under vacuum
until minimum volume or no more distillate was obtained.
Vol. 15, No. 1, 2011 / Organic Process Research & Development
•
71

To the reactor was charged 23.4 kg of 200 proof ethanol
and the mixture was distilled. The ethanol charge and
distillation was repeated. If the toluene content was >10%,
the ethanol charge and distillation could be repeated. The
total volume was adjusted to 40 L, and the reaction
mixture was cooled to 60 °C. Seeds were charged (33 g
of 1 in 0.8 kg of heptane and 0.2 kg of 200 proof ethanol),
and the line was rinsed with 0.5 kg of 200 proof ethanol.
The mixture was held at 60 °C until a slurry formed and
the antisolvent solution (16.9 kg of heptane and 2.17 kg
of ethanol) was slowly charged. The batch was cooled to
20 °C over 4 h, and the slurry was wet-milled until the
desired particle size was obtained (D90 < 60 µm). The
cake was filtered and washed with two portions of wash
mixture (2.34 kg of ethanol and 18.28 kg of heptane).
The solids were dried in a vacuum oven at 40-50 °C until
the LOD was <1%. A total of 2.655 kg was obtained in
85.4% yield (corrected for potency of input and product),
>99.9% ee, and 99.8 HPLC area percent purity with a
D90 of 28 µm.
Acknowledgment
We thank the BMS plant operations and Analytical R&D
staff for their valuable support during our scale-up runs. In
addition, we thank the safety group for hazard assessment
studies and Beth Sarsfield for SEM analysis. We also thank
Dr. Rodney Parsons and Dr. Kenneth Fraunhoffer for helpful
discussions during the preparation of the manuscript.
Received for review August 13, 2010.
OP100225G
72
•
Vol. 15, No. 1, 2011 / Organic Process Research & Development