Practical synthesis of capromorelin, a growth hormone secretagogue, via a crystallization-induced dynamic resolution

Article abstract of DOI:10.1016/j.bmc.2016.12.015

A practical synthesis of capromorelin (1), a growth hormone secretagogue, is described that utilizes as a key step a crystallization-induced dynamic resolution (CIDR) of (±)-3a-benzyl-2-methyl-4,5,6,7-tetrahydro-2H-pyrazolo[4,3-c]pyridin-3(3aH)-one [(±)-2] by L-tartaric acid salt formation, yielding (R)-2.L-tartaric acid in high chemical yield (>85%) and with diastereomeric excess (de) of ～98%. Treatment of (R)-2.L-tartaric acid with ammonium hydroxide provided (R)-2 without loss of chiral purity. In situ generated (R)-2 was coupled with (R)-3-(benzyloxy)-2-(2-(tert-butoxycarbonyl)-2-methylpropanamido)propanoic acid [(R)-3] to give predominantly a single diastereomer of N-Boc-protected capromorelin [(1R,3aR)-4]. This process was used to prepare bulk quantities of capromorelin from (±)-2 to support preclinical toxicology studies.

Full text of DOI:10.1016/j.bmc.2016.12.015

Accepted Manuscript
Practical synthesis of capromorelin, a growth hormone secretagogue, via a crys-
tallization-induced dynamic resolution
Colin R. Rose, Michael P. Zawistoski, Bruce A. Lefker, F. Michael Mangano,
Ann S. Wright, Philip A. Carpino
PII:
S0968-0896(16)31398-0
DOI:
http://dx.doi.org/10.1016/j.bmc.2016.12.015
BMC 13442
Reference:
Bioorganic & Medicinal Chemistry
To appear in:
Received Date:
Revised Date:
Accepted Date:
19 September 2016
7 December 2016
8 December 2016
Please cite this article as: Rose, C.R., Zawistoski, M.P., Lefker, B.A., Michael Mangano, F., Wright, A.S., Carpino,
P.A., Practical synthesis of capromorelin, a growth hormone secretagogue, via a crystallization-induced dynamic
resolution, Bioorganic & Medicinal Chemistry (2016), doi: http://dx.doi.org/10.1016/j.bmc.2016.12.015
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Practical synthesis of capromorelin, a growth hormone secretagogue, via a
crystallization-induced dynamic resolution
Colin R. Rose, Michael P. Zawistoski, Bruce A. Lefker, F. Michael Mangano, Ann S. Wright, Philip A.
Carpino*
*Department of Medicinal Sciences, Pfizer Inc., 610 Main Street, Cambridge, MA 02139 USA
Abstract
A practical synthesis of capromorelin (1), a growth hormone secretagogue, is described that utilizes as a
key step a crystallization-induced dynamic resolution (CIDR) of ()-3a-benzyl-2-methyl-4,5,6,7-
tetrahydro-2H-pyrazolo[4,3-c]pyridin-3(3aH)-one [()-2] by (L)-tartaric acid salt formation, yielding
(R)-2.L-tartaric acid in high chemical yield (>85%) and with diastereomeric excess (de) of ~98%.
Treatment of (R)-2.L-tartaric acid with ammonium hydroxide provided (R)-2 without loss of chiral
purity. In situ generated (R)-2 was coupled with (R)-3-(benzyloxy)-2-(2-(tert-butoxycarbonyl)-2-
methylpropanamido)propanoic acid [(R)-3] to give predominantly a single diastereomer of N-Boc-
protected capromorelin [(1R,3aR)-4]. This process was used to prepare bulk quantities of capromorelin
from ()-2 to support preclinical toxicology studies.
Keywords
Capromorelin
Crystallization-induced dynamic resolution
Growth hormone secretagogue
GHS-R1a
Ghrelin mimetic
1. Introduction
Capromorelin (2-amino-N-[(1R)-2-[(3aR)-2,3,3a,4,6,7-hexahydro-2-methyl-3-oxo-3a-
(phenylmethyl)-5H-pyrazolo[4,3-c]pyridin-5-yl]-2-oxo-1-[(phenylmethoxy)methyl]ethyl]-2-methyl-
propanamide; 1) is an orally active growth hormone secretagogue and small molecule mimetic of
ghrelin, a hormone that plays a critical role in the regulation of energy homeostasis.1 Capromorelin (in
the form of its (L)-tartrate salt) was initially studied in human clinical trials as a treatment for frailty in
elderly adults. Only modest improvements in various measures of physical performance were observed
after one year in a Phase II clinical trial, insufficient to justify further clinical investment.2

Capromorelin was later investigated in companion animals as an orexigenic agent and was recently
approved in 2016 by the Food and Drug Administration’s Center for Veterinary Medicine as a treatment
for inappetence in dogs with underlying pathophysiology that results in eating disorders (Entyce®).
In the original discovery synthesis of capromorelin, the piperidine heterocycle, ()-3a-benzyl-2-
methyl-4,5,6,7-tetrahydro-2H-pyrazolo[4,3-c]pyridin-3(3aH)-one [()-2], prepared in a straightforward
manner, was coupled with (R)-3-(benzyloxy)-2-(2-(tert-butoxycarbonyl)-2-
methylpropanamido)propanoic acid [(R)-3] in the presence of triethylamine, 1-ethyl-3-(3-
dimethylaminopropyl)carbodiimide hydrochloride (EDC), and 1-hydroxy-7-azabenzotriazole (HOAt)3
to yield a 1:1 mixture of two diastereomeric products, (1R,3aR)-4 and (1R,3aS)-4, which were separated
by silica gel column chromatography (see Scheme 1).1 The desired diastereomer, (1R,3aR)-4, was
treated with trifluoroacetic acid to remove the N-Boc protecting group and provide capromorelin of high
diastereomeric excess (de). While this route could be used to prepare small scale batches of drug to
support in vitro and in vivo pharmacology studies, it was not practical for synthesizing the larger
amounts of drug needed for toxicology and clinical studies. Herein, we describe the development of a
novel resolution of ()-2 by (L)-tartaric acid salt formation, yielding a single diastereomeric product,
(R)-2.L-tartaric acid, in high chemical yield. We also report conditions for coupling (R)-3 with
enantiomerically pure (R)-2, generated in situ from (R)-2.L-tartaric acid, to give N-Boc-protected
capromorelin, (1R,3aR)-4, without any need for chromatography.

Scheme 1. Discovery synthesis of capromorelin (1) starting from ()-2. Reagents: (a) EDC, NEt3,
HOAt; (b) silica gel column chromatography; (c) trifluoroacetic acid.
2. Results and Discussion
A diastereomeric crystallization of ()-2 was initially investigated using several different chiral acids
as resolving agents, the most effective of which was L-tartaric acid. Treatment of ()-2 with L-tartaric
acid in wet acetone at room temperature provided an insoluble salt containing 2 of modest ~44%
enantiomeric excess (ee). Interestingly, compound 2 obtained from the salt complex that remained
behind in the acetone mother liquor was racemic, showing no enrichment of the opposite enantiomer as
was expected. The racemization of 2 was proposed to proceed via a retro-Mannich reaction (Scheme 2).
A survey of the literature revealed several examples of C3-carboxy or C3-cyano substituted piperidine
derivatives that were found to epimerize via retro-Mannich reactions.4

Scheme 2. Proposed mechanism for the racemization of enantiomerically enriched 2 (stereochemistry
undefined) that remained in solution during the diastereomeric crystallization process.
The facile racemization of 3a-benzyl-2-methyl-4,5,6,7-tetrahydro-2H-pyrazolo[4,3-c]pyridin-
3(3aH)-ones such as 2 suggested the possibility of developing a dynamic kinetic resolution of ()-2. In
our work exploring diastereomeric crystallizations of ()-2, we discovered that one of the diastereomeric
L-tartrate salts of 2 was significantly less soluble in wet acetone compared to the other diastereomeric
salt form and could be used to remove one of the enantiomers of 2 that remained in the acetone solution,
potentially enabling a variation of a dynamic kinetic resolution called a crystallization-induced dynamic
resolution (CIDR). Starting with the protocol used in the diastereomeric crystallization and
experimenting with different combinations of acetone and water mixtures, we eventually found
conditions under which a CIDR of ()-2 by chiral (L)-tartrate salt formation proceeded in high yield. A
mixture of ()-2 and L-tartaric acid in acetone/water (25:1 ratio) was stirred at 50 C for 70 h, providing
an insoluble L-tartrate salt of 2 with diastereomeric excess (de) of ~98% and in >85% chemical yield,
significantly greater than the maximum theoretical yield of a classical chiral salt crystallization (Fig. S1-
2). The stereochemical configuration at the C3a position in 2.(L)-tartaric acid obtained from the CIDR
of ()-2 was assigned by first treating the compound with triethylamine to give the free amine 2.
Compound 2 was then coupled with (R)-3 in the presence of EDC, HOAt, and triethylamine as
previously described1 to yield predominantly a product that exactly matched the 1H spectral
characteristics and thin layer chromatography profile of N-Boc-protected capromorelin (Fig. S3),
thereby revealing the configuration of the C3a stereocenter in 2.(L)-tartaric acid to be R. The
configuration of the C3a stereocenter in N-Boc-protected capromorelin was established by x-ray
crystallographic analysis of capromorelin (Fig. S4). Elemental analysis of (R)-2.L-tartaric acid
prepared via a CIDR showed that a monotartrate 1.5 hydrate was formed. The CIDR of ()-2 was
shown to be water sensitive as overall yields of the diastereomeric salt product significantly decreased if
dry acetone was used as a solvent instead of an acetone:water mixture. The resolution of ()-2 was

successfully scaled up from 1 to >100 g to give (R)-2.L-tartaric acid without loss of chiral purity.
Stability studies conducted on (R)-2.L-tartaric acid did not reveal any significant issues or loss of de
upon standing for several months. In contrast, the free amine, (R)-2, was shown to readily racemize
upon standing at room temperature or at 0 C as an amorphous solid.
With an efficient CIDR process in hand to prepare multi-gram quantities of (R)-2.L-tartaric acid,
we next turned our attention to the bulk synthesis of N-Boc-protected capromorelin. Since (R)-2 was
found to undergo partial racemization when stored as a solid or when generated on scale-up in the
presence of triethylamine at 0 C, we developed a protocol for the in situ generation of (R)-2. A slurry
of (R)-2.L-tartaric acid in CH2Cl2 was treated at 0 C with 2 equivalents of NH4OH, yielding (R)-2 of
>98 ee (Fig. S5) and an insoluble ammonium tartrate salt which was removed by filtration to minimize
concerns that it might promote epimerization of (R)-2. Stability studies showed that a CH2Cl2 solution
of (R)-2 could be stored for at least 1 h at 0 C without (R)-2 undergoing any significant epimerization.
The CH2Cl2 solution containing (R)-2 was subsequently treated at 0 C with (R)-3 in the presence of
EDC and HOAt, providing N-Boc-protected capromorelin of ~94% de and in 95% chemical yield
(Scheme 3; Fig S6-7). A minor amount of epimerization of (R)-2 was found to occur during the
coupling reaction. The salt exchange of (R)-2.L-tartaric acid with NH4OH and the coupling of (R)-2
with (R)-3 were scaled from 1 to >100 g to give N-Boc-protected capromorelin without any significant
loss of diastereomeric purity or other issues. N-Boc-protected capromorelin was converted into
capromorelin as previously described, at which point the small amounts of the undesired capromorelin
diastereomers that formed during the key coupling reaction could be removed.1
3. Conclusions
A novel CIDR of ()-2 by chiral (L)-tartrate salt formation was developed to provide (R)-2.L-
tartaric acid in high yield and diastereomeric purity. An efficient and scalable salt exchange of (R)-
2.L-tartaric acid with NH4OH in CH2Cl2 provided (R)-2 with minimal epimerization. In situ generated
(R)-2 was coupled with (R)-3 in the presence of EDC and HOAt to give N-Boc-protected capromorelin
of ~94% de. This process enabled the bulk preparation of the penultimate intermediate in the synthesis
of capromorelin without a requirement for chromatography to separate diastereomers and facilitated the
practical synthesis of bulk lots that were needed to support pre-clinical toxicology studies (Scheme 3).

Scheme 3. Practical synthesis of N-Boc-protected capromorelin from ()-2. Reagents and conditions:
(a) L-tartaric acid, 25:1 acetone:H2O, 50 C, 70 h (b) NH4OH, CH2Cl2, 0 C; (c) (R)-3, EDC, HOAt,
CH2Cl2, 0 C.
4. Experimental
4.1 Chemical Synthesis
1H NMR spectra were recorded on either a Varian XL-300 or Varian VNMRS 600 spectrometer.
Chemical shifts are expressed in ppm relative to internal CDCl3 or CD3OD. Low resolution mass
spectra were obtained on either a Fisons PlatForm II mass spectrometer utilizing ApcI ionization or a
Fisons Thermospray TSP1000 mass spectrometer. Solvents and reagents were commercially available
and used directly without further purification. All reactions were monitored by thin-layer
chromatography on 0.25 mm x 5 cm x 10 cm silica gel 60 GF-254 (EM Science) plates using ultraviolet
light for visualization. Flash chromatography was performed using 40-63 m silica gel (EM Science).
HPLC methodology for ee determinations: Chiral Pak AS column, eluting with hexane/2-propanol
(65:35) at a flow rate of 0.6 ml/min, UV detection at 231 nm; HPLC methodology for de determination:
Chiral Tech IC 250 mm x 4.6 mm, 5 micron column using a gradient of CO2/MeOH + 0.2% NH3 in
MeOH.
4.1.1 Synthesis of 3aR-benzyl-2-methyl-4,5,6,7-tetrahydro-2H-pyrazolo[4,3-c]pyridin-3(3aH)-
one.L-tartrate [(R)-2.L-tartaric acid]
A mixture ()-2 (5.000 g, 20.6 mmol) and L-tartaric acid (3.092 g, 20.6 mmol) in acetone (80 mL)
and H2O (3.2 mL) was heated at 50 C under N2 for 70 h, during which time the reaction mixture
became a thick suspension. The reaction mixture was slowly cooled to 23 C and filtered to give a solid
that was washed with acetone (3 x 50 mL) and dried in vacuo to give 7.034 g (87% yield) of (R)-2.L-
tartaric acid as a white solid. An aliquot of the L-tartaric acid salt was treated with triethylamine in
CH2Cl2 at room temperature to provide (R)-2 which was subjected to chiral HPLC analysis for ee
1
determination (98% ee; Fig. S1-2). H NMR (CD3OD)  7.20 – 7.19 (m, 3H), 7.08 – 7.07 (d, J = 6.2
Hz, 2H), 4.48 (s, 2H), 3.63 – 3.60 (m, 1H), 3.54 – 3.43 (m, 2H), 3.30 – 3.28 (m, 2H), 3.07 (d, J = 13.5
Hz, 1H), 2.92 – 2.87 (m, 2H), 2.86 – 2.78 (m, 1H), 2.63 (d, 13.1 Hz, 1H); MS m/e 244 (M++1).

.
Elemental anal. calcd for C18H23N3O7 1.5H2O: C, 51.42; H, 6.23; N, 10.00. Found: C, 51.19; H, 6.04; N,
9.97.
4.1.2 Synthesis of 3aR-benzyl-2-methyl-4,5,6,7-tetrahydro-2H-pyrazolo[4,3-c]pyridin-3(3aH)-one
[(R)-2]
Concentrated NH4OH (0.34 mL, 5.1 mmol) was added to a suspension of (R)-2.L-tartaric acid (1.00 g,
2.5 mmol) in anhydrous CH2Cl2 (20 mL) at 0 C. The resulting mixture was stirred at 0 C for 0.25 h
and filtered. The filtrate containing free amine (R)-2 (>98% ee; Fig. S5) was used immediately in the
coupling reaction.
4.1.3 Synthesis of {1-[2-(3aR-Benzyl-2-methyl-3-oxo-2,3,3a,4,6,7-hexahydro-pyrazolo[4,3-
c]pyridin-5-yl)-1-(R)-benzyloxymethyl-2-oxo-ethylcarbamoyl]-1-methyl-ethyl}-carbamic acid tert-
butyl ester [(1R,3aR)-4]
A mixture of (R)-2 in CH2Cl2 from above [assumed to contain 3.09 g, 12.7 mmol of (R)-2], (R)-3
(4.84 g, 12.7 mmol), EDC (2.45 g, 12.8 mmol), and HOAt (2.60 g, 19.1 mmol) was stirred at 0 C for 1
h, then was warmed to 23 C and allowed to stir for 16 h. The reaction mixture was filtered, and the
filtrate was washed with saturated aq. NaHCO3 (100 mL) and H2O (100 mL), dried over MgSO4, and
1
evaporated in vacuo to give 7.35 g (95% yield) of (1R,3aR)-4 as a white solid (94% de; Fig. S6-7). H
NMR (600 MHz, CDCl3) δ 7.38 – 7.07 (m, 12H), 5.26 (d, J = 6.46 Hz, 1H), 5.07 (d, J = 12.91 Hz, 1H),
4.83 (br. s., 1H), 4.50 (d, J = 7.04 Hz, 2H), 3.72 (m, 1H), 3.58 – 3.67 (m, 1H), 3.16 (d, J = 13.50 Hz,
1H), 3.07 (s, 3H), 2.83 – 2.95 (m, 3H), 2.56 (d, J = 9.98 Hz, 1H), 2.49 (d, J = 12.91 Hz, 1H), 1.47 (d, J =
19.37 Hz, 6H), 1.32 – 1.42 (m, 9H) (Fig. S8); MS m/e 606 (M++1).
4.1.4 Synthesis of capromorelin (1) for x-ray crystallographic analysis
A solution of (1R,3aR)-4 (64.7 g, 106.8 mmol) in CH2Cl2 (300 mL) at 0 C was treated slowly with
ice-cold trifluoroacetic acid (300 mL) over 1.5 h. The resulting mixture was warmed to room
temperature, stirred for an additional 3 h, and filtered to yield a clear yellow solution which was
concentrated and re-dissolved in ethyl acetate. The ethyl acetate solution was washed with saturated aq.
NaHCO3 (3x) and saturated aq. NaCl (2x), dried over MgSO4, filtered, and concentrated in vacuo to give
49 g of an orange foam. A small aliquot of the orange foam was dissolved in isopropyl alcohol and
stirred overnight at 50 C to give white crystals of capromorelin which were removed by filtration and
found suitable for x-ray crystallography studies. The remaining foam was fully characterized after it
was transformed into capromorelin.(L)-tartaric acid as previously described.1
4.2 X-ray crystallography studies

A representative crystal of capromorelin was surveyed and a 1 Å data set was collected on a Siemens
R3RA/v diffractometer. Atomic scattering factors were taken from the International Tables for X-ray
Crystallography.5 All crystallographic calculations were facilitated by the SHETXTL system.6 All
diffractometer data were collected at room temperature.
Structure summary for capromorelin (1): empirical formula = C28H35N5O4; formula weight = 505.6;
calculated density (g/cm3) = 1.184; color, habit = colorless prismatic; crystal size (mm) = 0.16 x 0.22 x
0.44; crystal system = monoclinic; space group = C2; unit cell dimensions a = 24.035 (2) Å, b = 8.0450
(10) Å, c = 15.7380 (10) Å,  = 90.0,  = 111.180 (10); volume = 2837.6(5) Å3; Z = 4; absorption
coefficient (mm-1) = 0.651; F(000) = 1080.
Coordinates and structural factors can be obtained free of charge from The Cambridge
Crystallographic Data Centre (accession code: CCDC 1517678).
Acknowledgments
The authors thank Michael DeNinno for helpful discussions and Jon Bordner and Debra
DeCosta for generating and analyzing the x-ray crstallographic data.
Supplementary data
Supplementary data associated with this article can be found, in the online version, at:
Disclosures
C.R., B.L., A.W., and P.C. are employees of Pfizer, Inc.
References
1. Carpino, P. A.; Lefker, B. A.; Toler, S. M.; Pan, L. C.; Hadcock, J. R.; Cook, E. R.; DiBrino, J. N.;
Campeta, A. M.; DeNinno, S. L.; Chidsey-Frink, K. L.; Hada, W. A.; Inthavongsay, J.; Mangano, F.
M.; Mullins, M. A.; Nickerson, D. F.; Ng, O.; Pirie, C. M.; Ragan, J. A.; Rose, C. R.; Tess, D. A.;
Wright, A. S.; Yu, L.; Zawistoski, M. P.; DaSilva-Jardine, P. A.; Wilson, T. C.; Thompson, D. D.
Bioorg. Med. Chem. 2003, 11, 581.
2. White, H. K.; Petrie, C. D.; Landschulz, W.; MacLean, D.; Taylor, A.; Lyles, K.; Wei, J. Y.;
Hoffman, A. R.; Salvatori, R.; Ettinger, M. P.; Morey, M. C.; Blackman, M. R.; Merriam, G. R. J.
Clin. Endocrinol. Metab. 2009, 94, 1198.
3. Carpino, L. A. J. Am. Chem. Soc. 1993, 115, 4397.
4. Gmeiner, P.; Feldman, P. L.; Chu-Moyer, M. Y.; Rapoport, H. J. Org. Chem. 1990, 55, 3068.
5. Ibers, J. A.; Hamilton, W. C.; Editors, International Tables for X-Ray Crystallography, Vol. 4:
Revised and Supplementary Tables to Volumes 2 and 3. Kynoch Press: 1974.
6. Sheldrick, G. M. Acta Crystallogr., Sect. C: Struct. Chem. 2015, 71, 3.

Graphical Abstract
Practical synthesis of capromorelin, a growth hormone
Leave this area blank for abstract info.
secretagogue, via a crystallization-induced dynamic resolution
Colin R. Rose, Michael P. Zawistoski, Bruce A. Lefker, F. Michael Mangano, Ann S. Wright, and Philip A.
Carpino
Department of Medicinal Sciences, Pfizer Inc., 610 Main Street, Cambridge, MA 02139 USA