Development of a hydrogenative reductive amination for the synthesis of evacetrapib: Unexpected benefits of water

Article abstract of DOI:10.1021/op500025v

For the synthesis of cholesteryl ester transfer protein (CETP) inhibitor evacetrapib, a hydrogenative reductive amination was chosen to join the substituted cyclohexyl subunit to the benzazepine core. The addition of water, which suppressed undesired epimerization without affecting the rate of product formation, was key to the reactions success. The process was scaled to produce more than 1100 kg of material.

Full text of DOI:10.1021/op500025v

Communication
pubs.acs.org/OPRD
Development of a Hydrogenative Reductive Amination for the
Synthesis of Evacetrapib: Unexpected Benefits of Water
Michael O. Frederick,* Scott A. Frank, Jeffrey T. Vicenzi, Michael E. LeTourneau, K. Derek Berglund,
Adler W. Edward, and Charles A. Alt
Small Molecule Design and Development, Eli Lilly and Company, Indianapolis, Indiana 46285, United States
proved best with regards to lower impurity levels and anti:syn
ratios.
Selection of Solvent and Optimization. The choice of
ABSTRACT: For the synthesis of cholesteryl ester
transfer protein (CETP) inhibitor evacetrapib, a hydro-
genative reductive amination was chosen to join the
substituted cyclohexyl subunit to the benzazepine core.
The addition of water, which suppressed undesired
epimerization without affecting the rate of product
formation, was key to the reaction’s success. The process
was scaled to produce more than 1100 kg of material.
solvent was important, as we wanted a solvent that could be
used for both the aldehyde deprotection and the reductive
amination, without a solvent exchange. The aldehyde
deprotection was screened in toluene, IPAC, THF, and 2-
MeTHF12 (5 volumes, Table 2) with an aqueous solution of
Na2CO3 (3 equiv in 5 volumes H2O). The aldehyde was
obtained in yields greater than 96% for toluene, IPAC, and 2-
MeTHF, and 86% for THF. When these aldehyde solutions
were taken into the reductive amination, we noticed perform-
ance differences between drying the solutions with MgSO4
before use and using undried solutions directly.13 Conversion
and reaction rates were unaffected, but the anti:syn ratio was
impacted. To probe this effect, the aldehyde solutions in Table
2 were dried (MgSO4), and two reactions were performed with
each, one with 2% v/v water added and the other without water
added. With all eight reactions, conversion was 100%, but the
anti:syn ratios were variable, and the addition of water generally
proved beneficial. The anti:syn ratio with toluene was 83:17
with and without water, but the addition of water improved the
anti:syn ratio for IPAC, THF, and 2-MeTHF. THF and 2-
MeTHF were the most promising solvents for anti:syn ratio
(Table 2),14 but 2-MeTHF was chosen, as the aldehyde yield
was higher.
Examination of the Effect of Water on Reaction
Products. To further understand and optimize the effects of
water, a series of reactions in 2-MeTHF were conducted (Table
3). The anti:syn ratio improves as more water is added, up to
approximately 5% v/v. Using more than 5% v/v results in a
second aqueous layer (the solution is homogeneous with less
than 5%), and it seems extra water has no effect on the reaction,
positive or negative. To our knowledge, this is the first example
of water having a beneficial effect on reducing epimerization
during a reductive amination.
INTRODUCTION
■
Evacetrapib (5) is an experimental drug being investigated to
raise high-density lipoprotein cholesterol (HDL-C) via
inhibition of the cholesteryl ester transfer protein (CETP).1
Previously, a sodium triacetoxyborohydride (STAB)-2mediated
reductive amination was used to couple the substituted
cyclohexyl portion of the molecule to the benzazepine core
(Scheme 1).3 The aldehyde (3) used in the reductive amination
tends to epimerize and decompose upon prolonged storage,
leading us to protect it as the sodium bisulfite adduct (2) prior
to use.4 A biphasic mixture of aqueous sodium carbonate and
toluene converted the bisulfite adduct (2) to the intermediate
aldehyde (3), which was used directly in the reductive
amination with the benzazepine core (1). The reductive
amination with STAB proceeded smoothly, providing product
4 in 88% yield with 0.8% of the undesired syn-diastereomer 4′.
Saponification (aq NaOH in IPA) then provided evacetrapib
(5) in 92% yield along with 0.5% of the syn-diastereomer 5′.
Although the reductive amination proceeded smoothly,
concerns about cost and waste related to STAB5 led us to
investigate a hydrogenative reductive amination.6
RESULTS AND DISCUSSION
■
Selection of Catalyst. Reductive aminations utilizing a
heterogeneous catalyst are used for the manufacture of many
simple amines on scales exceeding thousands of metric tons a
year.7 The process has also found use in the manufacture of
more complex pharmaceutical targets,8 often using ketones,9
simple aldehydes (such as formaldehyde) in excess,10 or are
intramolecular reductive aminations.11 Many conditions were
explored for the optimization of the reductive amination (see
Table 1 for a few examples). Catalysis by Pd/C, Rh/C, and Pt/
C were all successful in giving full conversion; however, Pt/C
was chosen as it provided better anti:syn selectivity. Having a
pressure below 50 psi was important to minimize the
competitive reduction of the aldehyde. Many acids were
effective in catalyzing the reductive amination, but AcOH
In an attempt to understand the mechanistic basis for water
improving the anti:syn ratio, the epimerization and reaction
rates were monitored with and without 5% v/v water present.
Stirring a solution of aldehyde (3), amine (1), 5% Pt/C, and
AcOH in 2-MeTHF under a N2 atmosphere with and without
5% v/v water present, showed that water significantly decreases
the rate of epimerization (Figure 1). Surprisingly, when the
reaction is run with and without 5% v/v water present, the rate
of product formation is unaffected (Figure 2). Although a
hydrogenation of iminium 7 to product 4 cannot be ruled out,
Received: January 22, 2014
© XXXX American Chemical Society
A
dx.doi.org/10.1021/op500025v | Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Communication
a
Scheme 1. Synthesis of evacetrapib (5) via a STAB-mediated reductive amination.
a
Reagents and conditions: a) Na2CO3 (3.0 equiv), toluene, water, 25 °C, 3 h, 98% yield, 99.8:0.2 anti:syn; b) 3 (1.5 equiv), NaBH(OAc)3 (1.5
equiv), ACN, toluene, −10 °C, 2.5 h, 88% yield, 99.2:0.8 anti:syn; c) NaOH (3.0 equiv), water, IPA, 60 °C, 7 h, 92% yield, 99.5:0.5 anti:syn.
Pt from 19 ppm at the end of a reaction (which results in
isolated solids containing 250 ppm Pt), to 1.4 ppm (which
results in isolated solids containing less than 25 ppm Pt) after
2.5 h of heating (Table 4).20 The means by which heating
lowers the dissolved Pt levels is not completely understood.
One possibility is that the reaction is occurring through
dissolved Pt and the heating under a H2 atmosphere serves to
redeposit the Pt back on the support where it can be removed
by filtration.21
Table 1. Catalyst and acid screening results
Plant-Scale Reactions. Scheme 3 summarizes the process
utilized on manufacturing scale. Just 1.2 equiv of sodium
bisulfite adduct (2)22 was used to give greater than 99%
conversion, and the 5% Pt/C (5 wt %) was added as a slurry in
water, making the process safer than adding the catalyst dry or
in a flammable solvent. After the reaction was complete, the
solution was heated to 35 °C for 2 h and filtered, and the
product was crystallized from EtOH. Four batches were run at
250-kg scale with an average yield of 89% and an average level
of undesired syn-diastereomer 4′ of 0.5%,23 with Pt levels in the
product ranging from 16 ppm to less than 10 ppm
(representing our limit of detection). In total, more than
1100 kg of material were produced.
entry
catalyst
acid
AcOH
anti:syn
1
2
3
4
5
6
Pd/C
Rh/C
Pt/C
Pt/C
Pt/C
Pt C
95:5
93:7
98:2
90:10
91:9
89:11
AcOH
AcOH
TsOH
chloroacetic
salicylic
the seemingly counterintuitive effect whereby water is observed
to decrease the epimerization rate without affecting the rate of
product formation could more easily be explained through a
hydrogenolysis of hemiaminal 6 (Scheme 2).15 With a
hydrogenolysis, water would be expected to have little effect
on the rate of product formation but would be expected to slow
the epimerization through Le Chatelier’s principle.16
Platinum Residue Remediation. Adding 5% v/v water to
a reaction at 0 °C resulted in a very robust reaction giving high
yields (generally 85−90%) and less than 0.3% of the undesired
syn-isomer in the isolated product.17 Unfortunately, large
amounts of Pt were initially present in the isolated solids
(generally 250−1000 ppm). Various filter-aids and activated
carbons were successful in reducing the levels of Pt in solution,
but these were undesirable, due to extra manufacturing
burdens.18 A successful alternative was to heat the reaction to
35 °C under 50 psi H2 after the reaction was complete.19
Heating the solution rapidly decreases the amount of dissolved
CONCLUSIONS
■
A hydrogenative reductive amination for the synthesis of
evacetrapib was developed. Key to the success of the reaction
was the addition of water to improve the anti:syn ratio via a
slowing of the epimerization, and the thermal cycling to reduce
the amount of Pt in the product. Owing to the benefits of
water, the aldehyde solution could be directly used in the
reductive amination without drying, and the catalyst could be
added as a slurry in water. This method proved reliable for the
synthesis of over 1100 kg of product and could prove to be
general for other reductive amination processes.
EXPERIMENTAL SECTION
■
Amine 1 and sodium bisulfite adduct 2 were prepared
according to previously reported procedures.3 Volumes refer
B
dx.doi.org/10.1021/op500025v | Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Communication
Table 2. Solvent screening and water effects on the aldehyde deprotection and reductive amination
entry
solvent
aldehyde yield (%)
%H2O
4:4′
1
2
3
4
5
6
7
8
toluene
toluene
IPAC
99
99
96
96
86
86
97
97
0
2
0
2
0
2
0
2
83:17
83:17
85:15
92:8
IPAC
THF
95:5
THF
98:2
2-MeTHF
2-MeTHF
94:6
97:3
Table 3. Effects of water on the anti:syn ratio with the
reductive amination
entry
% water
anti:syn
Figure 1. Epimerization rate of aldehyde over time with 0% and 5% v/
v water added. Reaction carried out at 0 °C with 7 v 2-MeTHF, 2
equiv AcOH, 1.0 equiv amine 1, 1.7 equiv aldehyde 3, 5 wt % 5% Pt/
C, 1.0 equiv amine 1 under a N2 atmosphere.
1
2
3
4
5
6
7
8
0
1
94.7:5.3
97.6:2.4
97.7:2.3
98.2:1.8
98.4:1.6
98.6:1.4
98.8:1.2
98.5:1.5
2
3
4
5
10
20
to liters of solvent per kg of amine 1. Equivalents are based on
moles of amine 1.
Methyl(1S,4R)-4-(((S)-5-((3,5-bis(trifluoromethyl)-
benzyl)(2-methyl-2H-tetrazol-5-yl)amino)-7,9-dimethyl-
2,3,4,5-tetrahydro-1H-benzo[b]azepin-1-yl)methyl)-
cyclohexane-1-carboxylate (4). Sodium bisulfite adduct 2
(205.5 kg, 1.2 equiv corrected for 85% potency, 0.702 mol) and
2-MeTHF (1028 L, 4.11 volumes) were combined in a 1000-
gal glass-lined vessel with stirring under nitrogen at 20−25 °C.
In a separate 1000-gal glass-lined vessel, sodium carbonate (238
kg, 4.5 equiv, 2.25 mol) and water (1028 L, 4.11 volumes) were
combined with stirring at 20−25 °C until the sodium carbonate
had completely dissolved in the water. The sodium carbonate
solution was then charged to the sodium bisulfite adduct (2)/2-
MeTHF slurry. Once the addition was complete, the resulting
Figure 2. Reaction conversion over time with 0% and 5% v/v water
added. Reaction carried out at 0 °C with 7 v 2-MeTHF, 2 equiv
AcOH, 50 psi H2, 1.0 equiv amine 1, 1.7 equiv aldehyde 3, 5 wt % 5%
Pt/C, 1.0 equiv amine 1. Reaction with no water contained 3.2% of the
syn-isomer at the end of the reaction. Reaction with 5% water
contained 0.4% of the syn-isomer at the end of the reaction.
C
dx.doi.org/10.1021/op500025v | Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Communication
Scheme 2. Proposed hydrogenolysis mechanism of the reductive amination
with 2-MeTHF (725 L, 2.9 volumes). The resulting filtrate was
held until it was to be charged to the hydrogenation reactor
below.
Table 4. Effects of heating over time on dissolved Pt levels
entry
time at 35 °C (h)
dissolved Pt (mg/mL)
1
2
3
4
5
6
0
19
Five percent platinum on carbon catalyst (13.9 kg, ∼60%
water, 2.5 wt % of Pt) was slurried with water (63 L, 0.25
volumes) and charged into an empty 2000-gal glass-lined
hydrogenation reactor. The charge bottle was then rinsed with
water (25 L, 0.1 volumes) and again charged into the
hydrogenator. The solution of amine (1)/aldehyde (3) from
above was added to the hydrogenator, and the slurry was
cooled to 0−5 °C while the system was pressurized to 35−50
psig with nitrogen with vigorous stirring for 5 min. This
purging cycle was repeated three additional times to further
remove any oxygen. Once the system was cooled and purged,
AcOH (74.3 kg, 2.5 equiv, 1.24 mol) was charged, and the
hydrogenation reactor was then evacuated to 100−200 mmHg,
and pressurized to 50 psig with hydrogen starting from vacuum
for the reaction. The reaction was allowed to proceed for 16 h
while monitoring the hydrogen uptake to estimate reaction end
point. After the 16 h, the slurry was heated to 35 °C for 2 h
while still under hydrogen.
1.5
2.5
3.5
4.5
5.5
1.9
1.4
1.4
1.6
1.4
slurry was stirred vigorously at 20−25 °C, under nitrogen, for 2
h.24
After stirring for 2 h, the biphasic mixture was separated, and
the aqueous layer was discarded. The organic layer was then
washed with a solution of sodium chloride (75 kg, 2.5 equiv,
1.28 mol) in water (411 L), and the lower aqueous layer was
again discarded. Amine 1 (250 kg, 1.0 equiv, 0.502 mol) was
charged to the aldehyde (3)/2-MeTHF solution at 20−25 °C
and stirred until the amine (1) completely dissolved. The
resulting solution was then filtered through a 1−2-μm polish
filter to remove any potential insoluble material and then rinsed
Scheme 3. Optimized conditions for the scale-up of four 250-kg batches
D
dx.doi.org/10.1021/op500025v | Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Communication
Chem. Rev. 1980, 49, 14. (c) Rylander, P. A. Catalytic Hydrogenation
over Platinum Metals; Academic Press: New York, 1967; pp 291−303;
(d) Emerson, W. S. Org. React. 1948, 4, 174.
The reaction mixture was cooled to 20−25 °C and vented
and purged with nitrogen to remove all hydrogen. The slurry
was filtered to remove catalyst,25 passing through a final 1−2
μm polish filter and rinsed with 2-MeTHF (125 L, 0.5
volumes) into a 2000-gal glass-lined vessel. The filtered stream
was then washed with an aqueous solution of sodium
bicarbonate (92.8 kg, 2.2 equiv, 1.1 mol) and water (1000 L,
4 volumes), and the aqueous layer was discarded.
The organic layer was distilled down to 2−2.5 volumes
(500−600 L) at a temperature below 40 °C through the
application of a vacuum with heating through the jacketed
vessel. 3A EtOH26 (1750 L, 7 volumes) was then charged, and
the solution was distilled down through a vacuum distillation to
3−3.5 volumes (750−825 L) at a temperature below 40 °C. To
the reaction vessel, 3A EtOH (1250 L, 5 volumes) was again
added over 30 min and distilled down to 3−3.5 volumes (750−
825 L) at a temperature below 40 °C. 3A EtOH (1250 L, 5
volumes) was added a final time to the reaction mixture over 30
min. At this point, the slurry was cooled down to 0 °C over 1 h
and stirred for 2 h at 0 °C. The slurry was filtered and washed
twice with 3A EtOH (2 × 3 volumes, 1500 L total). The
isolated solid was then dried under vacuum between 45−55 °C.
The reductive amination product (4) was obtained as a white
solid (288 kg, 0.44 mol, 88% yield).
(7) Hayes, K. S. Appl. Catal., A 2001, 221, 187.
(8) (a) Tripathi, R. P.; Verma, S. S.; Pandey, J.; Tiwari, V. K. Curr.
Org. Chem. 2008, 12, 1093. (b) Cooper, C. G. F.; Lee, E. R.; Silva, R.
A.; Bourque, A. J.; Clark, S.; Katti, S.; Nivorozhkin, V. Org. Process Res.
Dev. 2012, 16, 1090.
(9) Allwein, S. P.; Roemmele, R. C.; Haley, J. J.; Mowrey, D. R.;
Petrillo, D. E.; Reif, J. J.; Gingrich, D. E.; Bakale, R. P. Org. Process Res.
Dev. 2012, 16, 148.
(10) Berliner, M. A.; Dubant, S. P. A.; Makowski, T.; Ng, K.; Sitter,
B.; Wager, C.; Zhang, Y. Org. Process Res. Dev. 2011, 15, 1052.
(11) Liang, J. T.; Deng, X.; Mani, N. S. Org. Process Res. Dev. 2011,
15, 876.
(12) NMP and MeOH were also good solvents for the reductive
amination, but their incompatibility with the bisulfite adduct
deprotection made them unusable.
(13) The impetus for exploring undried solutions was to simplify
large-scale production. These explorations led to the observed
differences in anti:syn ratios.
(14) Many of our screening reactions, such as those depicted in
Tables 1, 2, and 3, were done at 25 °C because our multiport reactors
did not have the ability to cool. Later studies such as the results
depicted in Figures 1 and 2 and later scale-up runs were conducted at
0 °C as it was found to provide better anti:syn ratios.
(15) For examples of reductive aminations believed to proceed
through reduction of the aminal, see: (a) Berdini, V.; Cesta, M. C.;
Curti, R.; D’Anniballe, G.; Di Bello, N.; Nano, G.; Nicolini, L.; Topai,
A.; Allegretti, M. Tetrahedron 2002, 58, 5669. (b) Tadanier, J.; Hallas,
R.; Martin, J. R.; Stanaszek, R. S. Tetrahedron 1981, 37, 1309.
(c) Smith, M. B.; March, J. MARCH’S Advanced Organic Chemistry, 5th
ed.; Wiley: New York, 2001; p 1188 and references cited therein.
(16) Atkins, P. W. The Elements of Physical Chemistry, 3rd ed.; Oxford
University Press: New York, 1993.
(17) Generally 50% of the undesired syn-isomer was rejected during
crystallization from EtOH. Reactions that contained 0.5% of the syn-
isomer at the end of reaction would contain 0.2−0.3% in the isolated
product.
(18) For methods to remove metals from process streams see: Bien,
J. T.; Lane, G. C.; Oberholzer, M. R. Top. Organomet. Chem. 2004, 6,
263.
(19) The relationship between temperature and Pt removal was not
explored; 35 °C was the first temperature tried, and it worked well so
that we did not explore alternate temperatures.
(20) Removing the Pt from the solution is beneficial for minimizing
the formation of an iminium impurity (dehydrogenation in the ring),
resulting from oxidative degradation of the product (4). The iminium
impurity could not be isolated, but was assigned a structure based on
HR/MS data.
(21) For discussion on transition-metal-catalyzed reactions being
homogeneous vs heterogeneous see: (a) Finney, E. E.; Finke, R. G.
Inorg. Chim. Acta 2006, 359, 2879. (b) Widegren, J. A.; Finke, R. G. J.
Mol. Catal A: Chem. 2003, 198, 317. (c) Conlon, D. A.; Pipik, B.;
Ferdinand, S.; LeBlond, C. R.; Sowa, J. R.; Izzo, B.; Collins, P.; Ho, G.-
J.; Williams, J. M.; Shi, Y.-J.; Sun, Y. Adv. Synth. Catal. 2003, 345, 931.
(d) Davies, I. W.; Matty, L.; Hughes, D. L.; Reider, P. J. J. Am. Chem.
Soc. 2001, 123, 10139.
(22) The equivalents of aldehyde were screened, and full conversion
was obtained with as few as 1.2 equiv. Less than 1.2 equiv resulted in
incomplete conversion.
(23) On plant scale, there was a one-hour delay from when the
AcOH was added to when H2 pressure was applied, resulting in a
slightly elevated level of the syn-diastereomer compared to lab trials.
(24) Homogeneity of the reaction solution is a good indicator of
reaction completion as the bisulfite adduct has limited solubility.
Generally, the solution becomes homogeneous within 30 min of
stirring; thus, we targeted 2 h on-scale to ensure completion.
AUTHOR INFORMATION
Corresponding Author
*frederickmo@lilly.com
Notes
■
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
The authors thank Lars Magnusson, Greg Rener, and Naomi
Miller for analytical support. The authors thank collaborators at
Evonik Industries for material production and helpful
discussions, especially around Pt removal via heating.
REFERENCES
■
(1) Nicholls, S. J.; Brewer, H. B.; Kastelein, J. J. P.; Krueger, K. A.;
Wang, M.; Shao, M.; Hu, B.; McErlean, E.; Nissen, S. E. J. Am. Med.
Assoc. 2011, 306, 2099.
(2) For reviews on STAB see: (a) Abdel-Magid, A. F.; Mehrman, S. J.
Org. Process Res. Dev. 2006, 10, 971. (b) Abdel-Magid, A. F.; Carson, K.
G.; Harris, B. D.; Maryanoff, C. A.; Shah, R. D. J. Org. Chem. 1996, 61,
3849. (c) Gribble, G. W.; Nutaitis, C. F. Org. Prep. Proced. Int. 1985, 17
(4−5), 317. (d) Gribble, G. W. Chem. Soc. Rev. 1998, 27, 395.
(3) For previous development work on the route to evacetrapib, see:
Chen, X.; Frank, S. A.; Remick, D. M.; Pederson, S. W. U. S. Patent
U.S. Patent 2010/0331309 A1, Dec. 30, 2010.
(4) (a) He, M.; Beahm, B. J.; Bode, J. W. Org. Lett. 2008, 10, 3817.
(b) Ragan, J. A.; am Ende, D. J.; Brenek, S. J.; Eisenbeis, S. A.; Singer,
R. A.; Tickner, D. L.; Teixeira, J. J., Jr.; Vanderplas, B. C.; Weston, N.
Org. Process Res. Dev. 2003, 7, 155. (c) Kjell, D. P.; Slattery, B. J.;
Semo, M. J. J. Org. Chem. 1999, 64, 5722. (d) Young, P. R.; Jencks, W.
P. J. Am. Chem. Soc. 1978, 100, 1228. (e) Chihara, T.; Wakabayashi, T.;
Taya, K. Chem. Lett. 1981, 1657. (f) Kissane, M. G.; Frank, S. A.;
Rener, G. A.; Ley, C. P.; Alt, C. A.; Stroud, P. A.; Vaid, R. K.; Boini, S.
K.; McKee, L. A.; Vicenzi, J. T.; Stephenson, G. A. Tetrahedron Lett.
2013, 54, 6587.
(5) Each kg of evacetrapib synthesized required roughly an equal
weight of STAB, for the addition of a single hydrogen to the final
product.
(6) For some reviews on reductive aminations see: (a) Mallat, T.;
Baiker, A. Handbook of Heterogeneous Catalysis; VCH: Weinheim,
1997; Vol. 5, pp 2334−2348. (b) Klyuev, M. V.; Khidekel, M. L. Russ.
E
dx.doi.org/10.1021/op500025v | Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Communication
(25) Dry Pt/C is pyrophoric, and care should be taken in handling
and disposal.
(26) 3A EtOH is EtOH denatured with MeOH and contains 5.0 gal
of MeOH for 100.0 gal of EtOH.
F
dx.doi.org/10.1021/op500025v | Org. Process Res. Dev. XXXX, XXX, XXX−XXX