Synthesis of ximelagatran, melagatran, hydroxymelagatran, and ethylmelagatran in H-3 labeled form

Article abstract of DOI:10.1002/jlcr.3028

In support of a study designed to better understand the liver toxicity of ximelagatran, ximelagatran, and melagatran, hydroxymelagatran and ethylmelagatran were prepared in tritium labeled form. Incorporation of tritium was achieved by hydrogen isotope exchange using Crabtree's catalyst and later with N-heterocyclic containing Ir catalyst. The tritiated product was then converted into the four target compounds to afford them in high purity and specific activity. The syntheses of ximelegatran and melagatran is reported from a common intermediate which was prepared by HIE. A direct comparison of HIE using an N-heterocyclic carbene containing Ir catalyst and Crabtree's catalyst is made. Copyright

Full text of DOI:10.1002/jlcr.3028

Technical Note
Received 07 December 2012,
Accepted 13 January 2013
Published online 16 February 2013 in Wiley Online Library
(wileyonlinelibrary.com) DOI: 10.1002/jlcr.3028
Synthesis of ximelagatran, melagatran,
hydroxymelagatran, and ethylmelagatran in
H-3 labeled form
Roger Simonsson, Gunnar Stenhagen, Cecilia Ericsson, and
*
Charles S. Elmore
In support of a study designed to better understand the liver toxicity of ximelagatran, ximelagatran, and melagatran, hydro-
xymelagatran and ethylmelagatran were prepared in tritium labeled form. Incorporation of tritium was achieved by hydro-
gen isotope exchange using Crabtree’s catalyst and later with N-heterocyclic containing Ir catalyst. The tritiated product was
then converted into the four target compounds to afford them in high purity and specific activity.
Keywords: ximelagatran; melagatran; hydroxymelagatran; ethylmelagatran; Hydrogen-isotope exchange; N-heterocyclic carbene
lead to large non-volatile impurities arising from the catalyst in
Introduction
the crude reaction mixture. The radiochemical purity of the crude
reaction mixture was only 30%. An improved synthesis of [3H]-1
Warfarin is an extensively used anticoagulant that has a narrow
therapeutic window. As a result, the levels of warfarin in the
blood must be closely monitored and any compound, which
was effected by using an N-heterocyclic carbene containing Ir cata-
lyst (2)11 at a loading of approximately 10% (Scheme 2). Not only
interacts with a clearance mechanism of warfarin has a poten-
was the specific activity improved substantially (to 1560 MBq/mmol)
but also the radiochemical purity of the crude non-volatile reaction
mixture was 84%. Compound 1 is susceptible to hydrolysis to amide
3 and a 5% UV impurity of 3 was observed in the starting material
for the second generation synthesis. However, HIE reactions with 1
containing 5% of 2 and deuterium gas had shown amide 3 to give a
tially severe drug–drug interaction.1 As a result many drugs are
contraindicated for users of warfarin and the diet of these
patients must be controlled.
AstraZeneca investigated ximelagatran as an alternative anticoa-
gulant.2 Ximelagatran is a prodrug of melagatran that is a thrombin
inhibitor.3 A two-step process converts ximelagatran to melagatran:
much lower incorporation of isotopic hydrogen (Table 1) and no
impurity arising from 3 was observed in [3H]-1.
either reduction of the hydroxyamidine followed by hydrolysis of
the ester or hydrolysis followed by reduction (Scheme 1).4
As we worked with this compound, a challenge immediately pre-
Drug-induced liver injury (DILI) is one of the most common
sented itself. We noted large variability in our scintillation counting
adverse reactions leading to product withdrawal post-marketing.
Recently, genome-wide association studies have identified a
number of human leukocyte antigen (HLA) alleles associated
values and that led us to believe that we had lost most of the radi-
olabeled material. After a careful investigation, it was determined
that 1 bound reversibly to plastic, and we later determined that this
with DILI. In the efforts to understand the observed liver toxicity of
also extended to other materials such as the carbon in 10% Pd/C.
ximelagatran (raised ALAT/ASAT (aspartate aminotransferase/alanine
The binding to plastic was saturable and thus was only observed
aminotransferase) values in about 7.9% of the patients on prolonged
with high-specific activity material. A similar behavior was observed
for all other compounds reported in this manuscript; therefore, only
glass was used for sample manipulations. Of course, a simple way to
overcome this phenomenon was to dilute the specific activity, but
treatment) focus was aimed at interesting findings in the HLA-
system.5–8 These studies required the active drug melagatran
together with some of its prodrugs ximelagatran, hydroxymelagatran,
and ethylmelagatran with high isotopic incorporation of tritium.
in this case, the user required a specific activity of >500 MBq/mmol
which precluded dilution.
The subsequent conversion from [3H]-1 to 5 was initiated by
deprotection of the benzylcarbonate protection of the amidine
Results
A common H-3 precursor to all four of these target molecules
was envisioned, and thus, 1 was targeted for tritium labeling.
This molecule seemed ideally poised for Ir-catalyzed hydrogen
isotope exchange (HIE) and exchange with Crabtree’s catalyst9,10
afforded [3H]-1 with a specific activity of 900 GBq/mmol. Unfor-
tunately, to achieve this level of incorporation, a stoichiometric
amount of catalyst was required and the high levels of catalyst
Isotope Chemistry, DMPK, AstraZeneca Pharmaceuticals LP, 43183 Mölndal,
Sweden
*Correspondence to: Charles S. Elmore, Isotope Chemistry, DMPK, AstraZeneca
Pharmaceuticals LP, 43183 Mölndal Sweden.
E-mail: chad_elmore@yahoo.com
J. Label Compd. Radiopharm 2013, 56 334–337
Copyright © 2013 John Wiley & Sons, Ltd.

R. Simonsson et al.
O
H
O
O
N
N OH
H
N
HO
N
NH2
reduction
hydrolysis
hydroxymelagatran
O
H
O
O
N
H
N
NH
O
H
O
O
HO
N
N
N
OH
H
N
NH2
O
N
NH2
melagatran
reduction
ximelagatran
hydrolysis
O
H
O
O
O
N
NH
H
N
N
NH2
ethylmelagatran
Scheme 1. Proposed metabolic pathway of ximelagatran to melagatran.
-
PF6
P
O
O
Ir+
N
O
O
NH
N
H
N
OBn
EtO
N
N
NH2
O
T
T
O
O
H
O
1
NH
2
N
N
OBn
EtO
N
NH2
T
2, CH2Cl2
O
H
N
O
O
H
O
O
55%
N
[3H]-1
EtO
N
N
H
OBn
3
Pd(OH)2
H2, Water
EtOH
65%
O
T
T
T
O
O
O
H
O
O
NH
NH
NH
H
NH
NaOH, H2O, EtOH
87%
N
N
EtO
N
N
HO
NH2
NH2
T
4
5
Scheme 2. Synthesis of [3H]ethylmelagatran (4) and [3H]melagatran (5).
to give 4. Care must be taken to avoid reduction of the aromatic
ring of [3H]-1; an initial reaction using 10% Pd/C gave a 30%
impurity of the over reduced product. The hexahydro adduct
was very difficult to separate by preparative HPLC. Several depro-
tection methodologies were investigated in an attempt to
improve this reaction including iodotrimethylsilane, Pd/CaCO3
with NaO2CH or H2 and finally, wet Pd(OH)2 + NaO2CH or H2. The
reaction of wet Pd(OH)2 and H2 for 90 min reliably afforded the tar-
get molecule without concomitant reduction of the aromatic ring.
Table 1. Incorporation of deuterium in 1 and 3 with catalyst 2
using reported conditions for tritiation with D2 in place of T2
Compound
Deuterium incorporation
D0
6
90
D1
12
2
D2
82
8
1
Amide 3
J. Label Compd. Radiopharm 2013, 56 334–337
Copyright © 2013 John Wiley & Sons, Ltd.
www.jlcr.org

R. Simonsson et al.
field of 11.74 T. Preparative HPLC was performed using basic (Gemini 5u, C18,
21.2 ꢀ 100 mm column, using a gradient of CH3CN in 0.2% conc. NH4OH +
5% CH3CN) or acidic (Kromasil 5u, C8, 20 ꢀ 100 mm, a gradient of CH3CN
in 0.1% TFA (trifluoroacetic acid)) conditions at a flow rate of 10 mL/min.
Ethyl 2-((R)-2-((S)-2-(4-((Z)-N0-(benzyloxycarbonyl)carbamimidoyl)-3,5-
[3H2]benzylcarbamoyl)azetidin-1-yl)-1-cyclohexyl-2-oxoethylamino)
acetate ([3H]-1)
Maintaining a short reaction time and low catalyst loading were
keys to achieving good recovery of material and preventing over
reduction. No loss in specific activity was observed for any of the
reductions. Hydrolysis of 4 to give 5 proceeded in good yield to
give both compounds in good yields.
The reaction of [3H]-1 with hydroxylamine gave 6 in good yield
(Scheme 3); however, purification of this material proved to be
difficult. After two purifications by preparative HPLC, a radiochemi-
cal purity of only 94% was obtained. Therefore, a portion of the
material was subjected to hydrolysis, but this reaction gave incom-
plete conversion. The two products, 6 and 7, were isolated and the
radiochemical purity of both compounds was 99%.
A solution of 13 mg (22 mmol, approx 95% pure by UV with the
remainder consisting of amide 3) of 1 and 2.2 mg (2.4 mmol) of 2 in
0.6 mL CH2Cl2 was attached to a tritium manifold (RC Tritec, Switzerland),
and the solution was degassed with three freeze-thaw cycles. The solu-
tion was then pressurized with 370 hPa (360 GBq, 170 mmol) of tritium
gas and after thawing, the solution was stirred for 2 h. The tritium gas
was recovered on a ‘recovery’ U-bed via the manifold system and the
volatiles were removed under a stream of nitrogen. The residue was
taken up in 1 mL of EtOH and evaporated to dryness with nitrogen. This
dissolution–evaporation process was repeated twice more. The crude
residue was purified in four batches by preparative HPLC (basic condi-
tions, gradient from 35% to 85% over 20 min) and the product containing
fractions were combined and evaporated to dryness. The residue was
taken up in 40 mL of EtOH to give 18.8 GBq of [3H]-1. The material was
characterized by MS (mass spectroscopy) to have a specific activity of
be 1550 GBq/mmol (592 (20%), 594 (55%), and 596 (100%)) and an HRMS
monoisotopic ion of 592.3141 (Calc = 592.3135 for C32H42N5O6). The com-
bined liquid waste from the preparations contained 3.5 GBq and the
amount of HTO released from the fume hood via the stack was
53.8 GBq (determined by continuous monitoring of the stack effluent
using a MARC 7000 tritium sampler from SDEC, France).
Conclusion
Four high-specific activities, tritium-labeled compounds were pre-
pared for use in studies directed at understanding the liver toxicity
of ximelagatran. Two different catalysts were directly compared for
the initial tritiation of 1 by HIE and the reaction using N-heterocyclic
carbene catalyst 211 was shown to give a higher specific activity and
less non-volatile waste than did the reaction with Crabtree’s cata-
lyst.9,10 All the tritiated compounds reported in this manuscript were
observed to stick to plastic at high specific activities and thus plastic
was avoided during the preparation.
Ethyl 2-((R)-2-((S)-2-(4-carbamimidoyl-3,5-[3H2]benzylcarbamoyl)azetidin-
1-yl)-1-cyclohexyl-2-oxoethylamino)acetate ([3H]ethylmelagatran, 4)
A solution of 5.0 GBq (3.2 mmol) of [3H]-1 in 10 mL of EtOH was con-
centrated to dryness, and the residue was mixed under N2 with 0.5 mg
of 20% Pd(OH)2 on C containing 50% water and 1 mL of EtOH. The slurry
was purged with N2 and then with H2. The slurry was stirred for 90 min
under balloon pressure of H2. After filtering, the solution was concen-
trated to dryness and then purified by preparative HPLC (basic condi-
tions, 40–90% over 20 min). The product containing fractions were
combined and evaporated to give an impure product (the peak shape
was poor and broad). A second purification by preparative HPLC (acidic
conditions, 10–65% over 20 min) was performed. The product containing
fraction was evaporated to dryness at room temperature and the mate-
rial was dissolved in 36 mL of EtOH to give 3270 MBq of 4. Radio-HPLC
analysis showed a radiochemical purity of 99% (method 1). The specific
activity was determined by MS to be 1570 GBq/mmol (458 (18%), 460
(55%), and 462 (100%)) and an HRMS monoisotopic ion of 458.2772
(Calc = 458.2767 for C24H36N5O4). The compound decomposed to a radio-
chemical purity of 96.4% over 1 year (method 2).
2-((R)-2-((S)-2-(4-carbamimidoyl-3,5-[3H2]benzylcarbamoyl)azetidin-1-
yl)-1-cyclohexyl-2-oxoethylamino)acetic acid ([3H]melagatran, 5)
A solution of 1500 MBq (0.96 mmol) of [3H]ethylmelagatran (4) in
16.5 mL of EtOH was evaporated to approximately 1 mL and was treated
with 30 mL 0.5 M NaOH (15 mmol) and 200 mL water. After 4 h, a solution of
TFA in CH3CN was added to the reaction mixture, and the solution was
concentrated to dryness. The residue was purified by preparative HPLC
(acidic conditions, 10–60% CH3CN over 20 min). The product containing
fraction was concentrated to dryness and taken up in 17 mL of EtOH to
give 1300MBq of 5 with a radiochemical purity of 99.7% (method 1). The
specific activity was determined by MS to be 1532GBq/mmol (430 (18%),
Experimental
In general, all the substances reported in this manuscript have a ten-
dency to stick to plastic. Therefore, care must be taken to avoid the use
of plastics during all synthetic and analytical stages. Starting benzylcar-
bonate 1 and all reference compounds were obtained from medicinal
chemistry, AstraZeneca, Mölndal, Sweden. Catalyst 2 was obtained from
Billy Kerr, University of Strathclyde. All other reagents were obtained
from Sigma and were used without further purification.
Radiochemical purity was determined by HPLC on a Waters 2695
Separations module (Waters Corporation, MA, USA) with a radioactivity
flow monitor using a Perkin-Elmer Radiomatic 625 TR. Two HPLC meth-
ods were used for radiochemical purity determinations and used a
mobile phase consisting of A (water with 0.2% formic acid adjusted to
pH3) and B (95% MeCN/water 0.2% formic acid pH3): Method 1 (Waters
Xselect CSH C18, 3.0 ꢀ 100 mm, 3.5 mm with gradient elution (0% B for
3 min then ramp to 95% B over 25 min and hold at 95% B for 2 min)
and UV detection at 220 nm) and method 2 (Waters Xbridge C18
3.5 mm, 4.6 ꢀ 100 mm with gradient elution (5% B for 0–3 min, then ramp
to 95% B over 22 min and holds at 95% B for 5 min) with UV detection at
254 nm). The identity of the compounds was determined by co-elution
on HPLC and by LC/HRMS (Pos ESI, PenTOF 200-1500Da, four scans/s,
Poroshell EC-C18, 1 ꢀ 50 mm, 2.7 um, 50 mL/min, 5–95% MeCN-H20 in
4 min). All compounds had an observed mass within 5 ppm of the calcu-
lated mass. The specific activities of the products were determined by
LC/HRMS. Liquid scintillation counting was performed with a Beckman
3
LS 6500 scintillation counter. H NMR spectra were acquired at 533 MHz
on a Bruker Avance III instrument (Bruker BioSpin Corporation, Billerica,
MA, USA) equipped with a special 3H probe and operating at a magnetic
O
T
T
O
T
T
O
O
O
O
NH2OH·HCl,
Et3N, THF, 40 °C
83%
H
NH
N OH
NH2
H
NH
N OH
NH2
NaOH,
H2O, EtOH
87%
[3H]-1
N
N
EtO
N
HO
N
6
7
Scheme 3. Synthesis of [3H]ximelagatran (6) and [3H]hydroxymelagatran (7).
www.jlcr.org
Copyright © 2013 John Wiley & Sons, Ltd.
J. Label Compd. Radiopharm 2013, 56 334–337

R. Simonsson et al.
432 (65%), and 434 (100%)) and an HRMS monoisotopic ion of 434.2610 the analytical support, Professor William Kerr for providing
(Calc = 434.2613 for C22H29T2N5O4). 3H-NMR (CD3OD, 533 MHz): 7.80 ppm. catalyst 2, Göran Nilsson for the helpful discussions, and Tommy
(3H NMR 1 year later showed an additional 5% peak at 7.87). The compound Andersson and Karin Cederbrandt for the assistance in the pre-
decomposed to a radiochemical purity of 94.5% over 1year (method 2).
Ethyl 2-((R)-1-cyclohexyl-2-((S)-2-(4-((Z)-N0-hydroxycarbamimidoyl)-
3,5-[3H2]benzylcarbamoyl)azetidin-1-yl)-2-oxoethylamino)acetate ([3H]
ximelagatran, 6)
A solution of 5.0 GBq (3.2 mmol) of [3H]-1 was concentrated to dryness
and 7 mg (0.1 mmol) of crushed NH2OH ꢁ HCl and 33 mg (0.3 mmol) of
Et3N in 0.2 mL THF was added. The reaction was stirred at 40 ꢂC over
night and was then concentrated to dryness. The residue was purified
paration of the manuscript.
Conflict of Interest
The authors did not report any conflict of interest.
by preparative HPLC (basic conditions, 20–70% over 20 min). The product References
containing fraction was evaporated at room temperature to afford
[1] J. Hirsh, New England Journal of Medicine 1991 ; 324 , 1865–1875.
4170 MBq of 6. Analysis by HPLC showed the radiochemical purity to be
94% (method 1). A more pure sample of 6 was isolated during the prepara-
tive HPLC of 7 (radiochemical purity 98.6%, method 1). The specific activity
was determined to be 1585 GBq/mmol (474 (16%), 476 (54%), and 478
(100%)) and an HRMS monoisotopic ion of 478.2869 (Calc = 478.2875 for
C24H33T2N5O5). The compound decomposed to a radiochemical purity of
96.7% over 1year (method 2).
2-((R)-1-cyclohexyl-2-((S)-2-(4-((Z)-N0-hydroxycarbamimidoyl)-3,5-
[3H2]benzylcarbamoyl)azetidin-1-yl)-2-oxoethylamino)acetic acid ([3H]
hydroxymelagatran, 7)
A solution of 1500 MBq (0.94 mmol, 94% radiochemical purity) of 6 in
EtOH was evaporated to 1 mL and was treated with 20 mL of 0.5 M
(10 mmol) NaOH and 100 mL water. After 70 min, a few drops of TFA were
added to the mixture and the solution was evaporated to dryness. The
material was purified by preparative HPLC (basic conditions, 20–70%
over 20 min) to give two major peaks. The first peak consisted of
400 MBq of 7 and was determined to have a radiochemical purity of
99.3% (method 1). The specific activity was determined to be 1580 GBq/
mmol (446 (18%), 448 (54%), and 450 (100%)) and an HRMS monoisotopic
ion of 450.2562 (Calc = 450.2562 for C22H29T2N5O5). The second peak
consisted of 1000 MBq of 6 (data reported earlier). The compound decom-
posed to a radiochemical purity of 92.8% over 6months (method 2).
[2] J. Hirsh, M. O’Donnell, J. W. Eikelboom, Circulation 2007 ; 116 , 552–560.
[3] D. Gustafsson, J. Nyström, S. Carlsson, U. Bredberg, U. Eriksson, E.
Gyzander, M. Elg, T. Antonsson, K. Hoffmann, A. Ungell, H. Sörensen,
S. Någård, A. Abrahamsson, R. Bylund, Thrombosis Research 2001 ;
101 , 181.
[4] U. G. Eriksson, U. Bredberg, K. Hoffmann, A. thuresson, M. Gabrielsson,
H. Ericsson, M. Ahnoff, K. Gislen, G. Fager, D. Gustafsson, Drug Metabo-
lism and Disposition 2003 ; 294 , 294–305.
[5] Kindmark, A., Jawaid, A., Harbron, C.G., B. J. Barratt, O. F. Bengtsson,
T. B. Andersson, S. C. Carlsson, K. E. Cederbrant, N. J. Gibson,
M. Armstrong, M. E. L. Lagerström-Fermér, A. Dellsén, E. M. Brown,
M. Thornton, C. Dukes, S. C. Jenkins, M. A. Firth, G. O. Harrod, T. H. Pinel,
S. M. E. Billing-Clason, L. R. Cardon, R. E. March, Pharmacogenomics
Journal 2008 ; 8 , 186–195.
[6] U. Andersson, J. Lindberg, S. Wang, R. Balasubramanian, M. Marcusson-
Ståhl, M. Hannula, C. Zeng, P. J. Juhasz, J. Kolmert, J. Bäckström, L. Nord,
K. Nilsson, S. Martin, B. Glinghammar, K. Cederbrant, I. ISchuppe-
Koistinen, Biomarkers 2009 ; 14 , 572–586.
[7] P. T. Illing, J. P. Vivian, N. L. Dudek, L. Kostenko, Z. Chen,
M. Bharadwaj, J. J. Miles, l. Kjer-Nielsen, S. Gras, N. A. Williamson,
S. R. Burrows, A. W. Purcell, J. Rossjohn, J. James McCluskey, nature
2012 ; 486 , 554–558.
[8] M. Keisu, T. B. Andersson, in Drug induced liver injury in human: The case
of ximelagatran, (Ed: E. Uetrecht), Springer-Verlag, Berlin Heidelberg,
2010, pp. 407–418.
[9] R. H. Crabtree, Acc. Chem. Res. 1979 ; 12 , 331.
[10] D. Hesk, P. R. Das, B. Evans, J. Labelled Compd. Radiopharmaceut.
1995 ; 36 , 497–502.
[11] J. A. Brown, S. Irvine, A. R. Kennedy, W. J. Kerr, S. Andersson,
G. N. Nilsson, Chem. Comm. 2008 ; 1115.
Acknowledgements
We would like to thank Marie Rydén Landergren and
3
Göran Carlström for the H-NMR support, Lena von Sydow for
J. Label Compd. Radiopharm 2013, 56 334–337
Copyright © 2013 John Wiley & Sons, Ltd.
www.jlcr.org