Scale-up of trisodium [(3Β,5Β,12α)-3-[[4(S)-4-[bis[2- [bis[(carboxy-kO)methyl]aminokN] ethyl]amino-kN]-4-(carboxy-kO)-1-oxobutyl] amino]-12-hydroxycholan-24- oato(6-)]gadolinate(3-)], a Gd(III) complex under development as a contrast agent for MRI coronary angiography

Article abstract of DOI:10.1021/op900008a

Process chemistry involved in the discovery and development routes to trisodium [(3Β,5Β,12α)-3-[[4(S)-4-[bis[2-[bis[(carboxykO) methyl]amino-kN]ethyl]amino-kN]-4-(carboxy-kO)-1-oxobutyl]- amino]-12- hydroxycholan-24-oato(6-)]gadolinate(3-)] (B22956/1) starting from L-glutamic acid and (3α,5Β,12α)-3,12-dihydroxycholan-24-oic acid is described. The best process is based on seven chemical steps and overcomes difficult purification protocols. Such process has been successfully implemented to prepare multikilogram batches of the target compound in 20% overall yield from (3α,5Β,12α)-3,12-dihydroxycholan-24-oic acid.

Full text of DOI:10.1021/op900008a

Organic Process Research & Development 2009, 13, 739–746
Scale-Up of Trisodium [(3ꢀ,5ꢀ,12r)-3-[[4(S)-4-[Bis[2-[bis[(carboxy-kO)methyl]amino-
kN]ethyl]amino-kN]-4-(carboxy-kO)-1-oxobutyl]amino]-12-hydroxycholan-24-
oato(6-)]gadolinate(3-)], a Gd(III) Complex under Development As a Contrast Agent
for MRI Coronary Angiography
Pier Lucio Anelli,* Marino Brocchetta, Luciano Lattuada, Giuseppe Manfredi, Pierfrancesco Morosini, Marcella Murru,
Daniela Palano, Marco Sipioni, and Massimo Visigalli
Centro Ricerche Bracco, Bracco Imaging Spa, Via Ribes 5, 10010 Colleretto Giacosa (TO), d Italy
Abstract:
promising,5,6 and presently, B22956/1 has successfully com-
pleted phase 1 clinical studies.7
Process chemistry involved in the discovery and development
routes to trisodium [(3ꢀ,5ꢀ,12r)-3-[[4(S)-4-[bis[2-[bis[(carboxy-
kO)methyl]amino-kN]ethyl]amino-kN]-4-(carboxy-kO)-1-oxobutyl]-
amino]-12-hydroxycholan-24-oato(6-)]gadolinate(3-)] (B22956/1)
When B22956/1 was first prepared in the lab, it was
synthesized following the route previously reported for the
analogous complex featuring a subunit of cholic acid (Scheme
1).8
Indeed, compound 1, containing a DTPA skeleton and an
additional carboxylic group for the amidation, is obtained from
L-glutamic acid, a cheap starting material, and amino ester 2
can be obtained from the commercially available (3R,5ꢀ,12R)-
3,12-dihydroxycholan-24-oic acid.9 To introduce the amino
group in position 3 the different reactivity of the two hydroxy
groups, (i.e., the axial hydroxy group in position 12 is much
less reactive than the equatorial one in position 3) was
exploited.10
starting from -glutamic acid and (3r,5ꢀ,12r)-3,12-dihydroxy-
L
cholan-24-oic acid is described. The best process is based on seven
chemical steps and overcomes difficult purification protocols. Such
process has been successfully implemented to prepare multikilo-
gram batches of the target compound in 20% overall yield from
(3r,5ꢀ,12r)-3,12-dihydroxycholan-24-oic acid.
The amidation of 2 with 1 (DCC/HOBT/CH2Cl2, silica gel
flash-chromatography, 94% yield) afforded the hexaester 3
which was deprotected to the hexaacid 4 [(i) TFA, (ii) aq NaOH,
(iii) aq HCl; 98% yield]. The latter was complexed to B22956/1
(GdCl3/aq NaOH, 88% yield) in 81% overall yield (Scheme
1). As an alternative, the amidation was performed with diethyl
cyanophosphonate and triethylamine in DMF (83%).11
The main critical issues for the scale-up of this synthetic
approach were: (i) aminoester 2 was prepared following a too
expensive route, and (ii) acid 1 was obtained in too-low yields.
Furthermore, it slowly racemises on storage for long times.
Accordingly, alternative synthetic routes leading to 4 had to be
investigated.
Introduction
Over the past decade, several complexes of paramagnetic
metal ions have been tested as contrast agents for in ViVo
magnetic resonance imaging (MRI) and a few of them have
been brought into clinical practice. Gd(III) complexes of ligands
derived from diethylenetriaminepentaacetic acid (DTPA) have
played a major role in this field.1,2 Furthermore, the conjugation
of DTPA-like ligands to bile acids has been pursued to achieve
MRI contrast agents featuring improved hepatospecificity.3 In
addition, the quite strong binding to human serum albumin of
the cholanoic residue proved able to guarantee a prolonged
permanence in blood vessels, and accordingly, one of these
conjugates, B22956/1 (Scheme 1), has been selected for
development as contrast agent for MRI coronary angiography.4,5
Preliminary imaging results in animal models were very
Here, we report how the initial discovery route to B22956/1
was modified into a process suitable for kilogram-scale
preparations.
Results and Discussion
Synthesis of Aminoester 2. Aminoester 2 was a key
intermediate of the above-described process, and its synthesis
* pier.lucio.anelli@bracco.com.
(1) Caravan, P.; Ellison, J. J.; McMurray, T.; Lauffer, R. B. Chem. ReV.
1999, 99, 2293–2352.
(2) Zhang, Z.; Nair, S. A.; McMurry, T. J. Curr. Med. Chem. 2005, 12,
751–778.
(6) de Hae¨n, C.; Anelli, P. L.; Lorusso, V.; Morisetti, A.; Maggioni, F.;
Zheng, J.; Uggeri, F.; Cavagna, F. M. InVest. Radiol. 2006, 41, 279–
291.
(3) Anelli, P. L.; Lattuada, L.; Lorusso, V.; Lux, G.; Morisetti, A.;
Morosini, P.; Serleti, M.; Uggeri, F. J. Med. Chem. 2004, 47, 3629–
3641.
(7) La Noce, A.; Stoelben, S.; Scheffler, K.; Hennig, J.; Maier Lenz, H.;
La Ferla, R.; Lorusso, V.; Maggioni, F.; Cavagna, F. Acad. Radiol.
2002, 9, S404–406.
(4) Anelli, P. L.; Brocchetta, M.; de Hae¨n, C.; Gazzotti, O.; Lattuada, L.;
Lux, G.; Manfredi, G.; Morosini, P.; Palano, D.; Serleti, M.; Uggeri,
F.; Visigalli, M. U.S. Patent 6,461,588 B12002.
(8) Anelli, P. L.; Fedeli, F.; Gazzotti, O.; Lattuada, L.; Lux, G.; Rebasti,
F. Bioconjugate Chem. 1999, 10, 137–140.
(5) Anelli, P. L.; Brocchetta, M.; Lorusso, V.; Manfredi, G.; Morisetti,
A.; Morosini, P.; Murru, M.; Palano, D.; Visigalli, M. Bile Acids at
Work: DeVelopment of a New IntraVascular MRI Contrast Agent;
Sessler, J. L., Doctrow, S. R., McMurry, T. J. Lippard, S. J., Eds.; In
Medicinal Inorganic Chemistry; ACS Symposium Series 903. Ameri-
can Chemical Society: Washington, DC, 2005; pp 137-150.
(9) Anelli, P. L.; Lattuada, L.; Uggeri, F. Synth. Commun. 1998, 28, 109–
117.
(10) Gao, H.; Dias, J. R. Org. Prep. Proced. Int. 1999, 31, 145–166.
(11) Shioiri, T.; Yokoyama, Y.; Kasai, Y.; Yamada, S. Tetrahedron 1976,
32, 2211–2217.
10.1021/op900008a CCC: $40.75  2009 American Chemical Society
Published on Web 05/22/2009
Vol. 13, No. 4, 2009 / Organic Process Research & Development
•
739

Scheme 1
Scheme 2a
A reaction was needed which would transform the 3R-OH
group of 5 into the 3ꢀ-NH2 group of 2 with complete inversion
of configuration. Accordingly, we investigated both classical
SN2 and Mitsunobu reactions.12a,b
An alternative transformation of 5 into 2 was explored
reacting the methanesulfonate of 5 (MsCl in pyridine) with
ammonia as well as classical and modified Gabriel reagents
[potassium phthalimide, Boc2N-K+, CF3CONH2, NaN-
(CHO)2].16 With potassium phthalimide the yield was very low
(13%), while the other reagents gave complicated mixtures and
no desired product.
An additional route, not depicted in Scheme 2, to 2 was also
investigated starting with the oxidation of 5 under Oppenauer
conditions [Al(t-BuO)3/acetone or toluene]17 to afford a mixture
of the desired 3-keto ester and the corresponding 3-keto acid
in moderate yield. Pure 3-keto ester was obtained by oxidation
with massive amounts of Ag2CO3 over Celite.18 The reductive
amination of the 3-keto derivatives (both methyl ester and acid)
was performed as reported in the literature19 using AcONH4
and NaBH3CN in MeOH. However, this approach was aban-
doned as mixtures of the 3R and 3ꢀ-NH2 derivatives were
obtained.
a Reagents and yields: (i) (1) MsCl/toluene/Et3N followed by NaN3 in H2O/
toluene/n-Bu4NBr, 64%; or (2) MsCl/pyridine followed by NaN3 in DMF, 71%;
or (3) Ph3P/DEAD/DPPA/THF, 70%; (ii) (1) Ph3P/CH3CN then H2O, 86%;9,13
or (2) H2/Pd-C/THF, 30 °C, 50 bar, 80%; or (3) H2/Ni Raney/THF, 30 °C, 1
bar, 80%; (iii) Ph3P/DEAD/(diethoxyphosphinyl)carbamic acid 1,1-dimethylethyl
ester/THF, 44% 7a [R1 ) t-BuO2C, R2 ) (EtO)2OP];14 Ph3P/DEAD/imidocar-
bonic acid bis(phenylmethyl) ester/THF, 45% 7b (R1 ) R2 ) Cbz);15 (iv) HCl/
MeOH for 7a, 87%; H2/Pd-C/MeOH for 7b, 83%.
Scheme 3a
Despite the quite good overall yield for the conversion of 5
into 2 with a two-step synthesis through 6, the use of hazardous
sodium azide raised some concern. As an alternative, we
explored the routes depicted in Scheme 3.
The reduction with NaBH4 of one of the carbonyl groups
of the phthalimido residue to afford an intermediate like 9
which, in acidic medium, gives the desired amine and 1-isoben-
(13) Knouzi, N.; Vaultier, M.; Carrie, R. Bull. Soc. Chim. Fr. 1985, 5,
815–819.
(14) Zwierzak, A.; Pilichowska, S. Synthesis 1982, 922–924.
(15) Degerbeck, F.; Fransson, B.; Grehn, L.; Ragnarsson, U. J. Chem. Soc.
Perkin Trans. 1 1992, 245–253.
a Reagents and yields: (i) Ph3P/DIAD/phthalimide/THF, 80%; (ii) (1) N2H4,
84%;20 or (2) ethylenediamine/MeOH (76%);21 or (3) other amines i.e. 2-ami-
noethanol, aq CH3NH2,22 n-BuNH2,23 NH4OH or CH3NH2/EtOH,24 5-50%; (iii)
NaBH4/DMA/phosphate buffer pH 8 then H2O, 99%; (iv) Ph3P/DIAD/phthal-
imide/DMA followed by NaBH4/DMA/phosphate buffer pH 8 then H2O, 77%;
(v)HCl/MeOH then H2O/NaOH, 85%.
(16) Ragnarsson, U.; Grehn, L. Acc. Chem. Res. 1991, 24, 285–289.
(17) Jones, A. S.; Webb, M.; Smith, F. J. Chem. Soc. 1949, 2164–2168.
(18) Kou-yi, T. J. Lipid Res. 1978, 19, 501–504.
(19) Bellini, A. M.; Quaglio, M. P.; Guarnieri, M. Farmaco 1986, 41, 401–
407.
was explored in depth. (3R,5ꢀ,12R)-3,12-Dihydroxycholan-24-
oic acid was reacted in MeOH/PTSA ·H2O to afford 5 (93%),
then several possibilities to transform the latter into 2, e.g.
through intermediates 6, 7, or 8 (Schemes 2 and 3), were
investigated.
(20) Hensen, A.; Glombik, H.; Kramer, W.; Wess, G. U.S. Patent 5,486,626,
1996.
(21) Kanie, O.; Crawley, S. C.; Palcic, M. M.; Hindsgual, O. Carbohydr.
Res. 1993, 243, 139–169.
(22) Motawia, M. S.; Wengel, J.; Abdel-Megid, A.; Pedersen, E. B.
Synthesis 1989, 384.
(23) Durette, P. L.; Meirzner, E. P.; Shen, T. Y. Tetrahedron Lett. 1979,
42, 4013–4016.
(12) (a) Hughes, D. L. Org. React. 1992, 42, 335–656. (b) Hughes, D. L.
Org. Prep. Proced. Int. 1996, 28, 127–164.
(24) Greene, T. W.; Wutts, P. G. M. ProtectiVe Groups in Organic
Synthesis, 3rd ed.; John Wiley and Sons: New York, 1999.
740
•
Vol. 13, No. 4, 2009 / Organic Process Research & Development

Dimethylacetamide (DMA) and N-methylpyrrolidone were quite
stable, and the content of 9a was around 2%. All the
aforementioned solvents worked in concentrated solutions, and
DMA was selected as the solvent of choice; (2) the excess of
NaBH4 was lowered from the initial 5.0 to 0.97 mol equiv, still
observing a good reactivity. However, when the amount of
NaBH4 was further reduced to 0.85 or 0.7 mol equiv, the
reaction was incomplete and afforded significant amounts
(8-12%) of 9e; (3) the first trials were performed in organic
solvents/water mixtures. Under these conditions the final pH
value largely exceeded 13, and as a consequence, the formation
of 9b-d was not negligible at all. The use of a 2 M phosphate
buffer (pH 8) allowed a certain control of the pH in the earlier
stage of the reaction. Indeed, although the mixture is strongly
alkaline (pH 12.2) at the end of the reaction, hydrolytic cleavage
of the labile functions is very limited: about 3-5% of 9c is
formed, along with traces of 9b and 9d; (4) the concentration
of 8 was progressively increased up to 10%, which resulted in
the upper limit. In fact, when it was further increased to 13%,
8 started to precipitate, and the amount of 9a significantly
increased (3.6%) in the reaction crude; (5) the reaction was
performed at 45-48 °C in order to avoid the precipitation of
both 8 and 9; (6) the choice of the reaction time was quite crucial
and was tentatively found after the other parameters were
adjusted. After 55-60 min, the reaction was stopped by addition
of AcOH because the undesired formation of 9a started to
become significant. When we changed the conditions and
worked in DMA/phosphate buffer, 9 was conveniently precipi-
tated by addition of water and recovered by filtration, thus
leaving the inorganic salts in the mother liquor. The deprotection
of isolated 9 was performed with conc HCl/MeOH at 50 °C.
These conditions are advantageous since the byproduct 9b is
newly esterified, affording 9 which is the precursor of 2.
The process was further improved when the isolation of 8
was avoided. In this respect, NaBH4 was suspended in the
solution of crude 8 in DMA, and after obtaining a solution, the
reduction was performed at a temperature <50 °C. It must be
underlined that 5 started to precipitate as a crystal containing
0.5 mol of MeOH when the scale of the esterification reaction
was increased, working in more concentrated solutions. This
was beneficial for the purification of 5 and simplified the workup
but imposed the distillation of MeOH from the DMA solution
in the next step, as MeOH would have been a reagent in the
Mitsunobu reaction leading to 8.
Figure 1
zofuranone is reported in the literature25 as the mildest method
for the deprotection of the phthalimido group. It has been
especially applied to protected aminoacids, but has never been
used for cholanoic derivatives. The reaction was studied in depth
before reaching the final conditions (Scheme 3, reaction (iii))
as those reported in the literature (2-PrOH/H2O, 5 mol equiv
of NaBH4, 20 °C) were not suitable due to the scarce solubility
of 8. Indeed, the reaction had to be performed under very dilute
conditions (i.e., 2.5% concentration) which are not appealing
from an industrial perspective. Furthermore, after 24 h at 40
°C, 9 was recovered in only 45% yield. Interestingly, several
byproducts (9a-g, Figure 1) were isolated by silica gel
chromathography and characterized.
Most of the byproducts stem from the presence of the methyl
ester moiety, which is sensitive to both reduction and alkaline
hydrolysis. In particular, working with 5 mol equiv of NaBH4,
compound 9a was formed in about 25% yield (TLC) and could
not be eliminated from 9 by crystallization from several solvents.
The subsequent acidic hydrolysis converted 9a into 9f, whose
very low solubility prevented its elimination from crude 2.
Compound 9e, deriving from incomplete reduction of 8,
afforded during the following acidic hydrolysis compound 9g
that could be only partially eliminated from crude 2. Compounds
9b-d resulted from hydrolysis at alkaline pH values and
lowered the reaction yield. However, in the following step, they
afforded byproducts that could be eliminated to a great extent.
As a consequence of the above-mentioned results, the reaction
conditions were investigated to understand the effect of (1)
solvent, (2) stoichiometry, (3) pH, (4) concentration, (5)
temperature, and (6) reaction time, in reactions on the 200-g
scale. The findings can be summarized as follows: (1) DMF
proved a much better solvent, in comparison with 2-PrOH, to
lower the formation of 9a (about 3%) but was hydrolysed to
some extent, affording the undesired dimethylamine. N,N-
New Synthetic Approach to Ligand 4. Owing to the
problems linked to the use of acid 1 in the discovery route (Vide
infra), a different strategy to afford the target ligand 4 was
investigated. While in the discovery route the disconnection
approach involved at first the formation of bonds a and then of
bond b (Figure 2), in the development route the disconnection
was turned the other way around, i.e. first bond b and then bonds
a are formed.
This approach required the choice of a suitably protected
L-glutamic acid derivative. The results of the amidation of 2
with N-Cbz-L-glutamic acid 1-t-Bu ester and coupling agents
were discouraging because of the troublesome purification of
the crude product and the consequent very low yield. The
reaction with N-Cbz-L-glutamic acid anhydride, although oc-
(25) Osby, J. O.; Martin, M. G.; Ganem, B. Tetrahedron Lett. 1984, 20,
2093–2096.
Vol. 13, No. 4, 2009 / Organic Process Research & Development
•
741

tions,30 affords hexaester 19 which is a useful precursor for the
preparation of B22956/1.
We discovered that 14 could be obtained, although a silica
gel chromatography was needed for its purification. In parallel,
the analogous reaction of 2 and 12 (Scheme 5) performed in
toluene afforded 15, which spontaneously crystallized, as a
solvate with toluene, on cooling the reaction mixture. A
noticeable improvement was thus achieved as the chromato-
graphic purification could be avoided. Furthermore, the crystal-
line solid 15, unlike the waxy 14, could be easily stored. An
additional important advantage was related to the following
deprotection (Scheme 6).
Figure 2
Scheme 4a
For the new route we still needed as alkylating agent
mesylate 17, which in our hands proved a convenient replace-
ment to the corresponding bromide classically used in the
Rapoport approach to DTPA frameworks.30 The preparation of
the precursor 16 underwent several optimisations and was finally
carried out with a very simplified workup. Compound 18, which
was at first synthesised from 14 by hydrogenolysis of the Cbz
protecting group (H2/Pd-C/MeOH, 1 bar, rt, 95%), could be
obtained, in a simpler and safer way, through Boc deprotection
of 15. The acidic deprotection did not induce racemisation at
the stereogenic centre in the amino ester moiety. Several trials,
performed using different solvents and acids, showed that 18
could be isolated as hydrochloride or methanesulfonate and
obtained as free amine from MeOH/H2O mixtures, after
neutralisation with DIPEA.
a Reagents and yields: (i) SOCl2/MeOH; (ii) KOH/MeOH, 72% (steps i and
ii);28a,b (iii) Boc2O/EtOAc/DMAP, 91%;29 50%; (iv) CbzCl/CH2Cl2, 32%; (v)
MeI/DIPEA/CH2Cl2, 98%.
Scheme 5a
The bisalkylation of 18 with 17 needed acetonitrile/pH 8
phosphate buffer (or H2O maintained at pH 8.5 with NaOH)
and gave 19 in 68% yield after silica gel flash chromatography.
The use of H2O/EtOH or H2O/i-PrOH mixtures, as well as of
pure solvents (EtOAc, DMA), gave worse results. Subsequent
studies showed that the bisalkylation of 18 could also be
performed in EtOAc in the presence of DIPEA (50 °C for 18 h,
silica gel flash-chromatography, 74%) using crude 17 in its turn
prepared in EtOAc. Compound 19 was prepared in kilogram-
scale in EtOAc and needed a silica gel chromatography
purification using EtOAc/n-hexane. This procedure, not suitable
for scale-up, allowed the elimination of most of the byproducts,
mainly deriving from incomplete alkylation or quaternisation
of the nitrogen atoms and/or lactamisation. Alternative purifica-
tions of 19 such as liquid extraction with water/organic solvents
mixtures, did not prove suitable. Subsequent studies showed
that, using n-BuOAc instead of EtOAc as the reaction solvent,
the amount of some byproduct was reduced, and accordingly,
both preparations of 17 and of 19 were performed in n-BuOAc.
Moreover, we realised that a thermal treatment of the crude,
while not affecting the content of 19, transformed most of the
byproduct (e.g., quaternary onium salts) into others that, after
the following alkaline hydrolysis, could be to a great extent
eliminated by elution through a nonfunctionalized polystyrene
a Reagents and yields: (i) 14: (1) dioxane, 50 °C 24 h, 95 °C 29 h,
chromatography, 75%; or (2) toluene, 100 °C, 24 h, chromathography, 70%;
15: toluene, 90 °C, 24 h, 74%.
curring in good yields, was ruled out because of the formation
of the unwanted regioisomer, i.e. the one deriving from
amidation of the carboxy group in position 1. Eventually, the
ring-opening of a protected pyroglutamate with amine 2 was
taken into account. The only reference26 found in the literature
reports a reaction performed with a large excess of amine (6
mol equiv) and catalyzed by KCN under ultrasonic irradiation
in order to shorten the reaction time. Obviously, such conditions
are not suited for an industrial application. Conversely, several
cases of ring-opening of protected pyroglutamates and lactams
with other nucleophiles have been reported.27a-d
The preparation of two potential candidates, 11 and 12, was
performed as reported in Scheme 4.
We subsequently investigated the reaction of 11 with 2 to
afford intermediate 14 (Scheme 5). This after deprotection and
bisalkylation with 17 (Scheme 6) under Rapoport-like condi-
(27) (a) Flynn, D. L.; Zelle, R. E.; Grieco, P. A. J. Org. Chem. 1983, 48,
2424–2426. (b) Schoenfelder, A.; Mann, A. Synth. Commun. 1990,
20, 2585–2588. (c) Attwood, M. R.; Carr, M. G.; Jordan, S.
Tetrahedron Lett. 1990, 31, 283–284. (d) Dixit, A. N.; Tandel, S. K.;
Rajappa, S. Tetrahedron Lett. 1994, 35, 6133–6134.
(28) (a) Adkins, H.; Billica, H. R. J. Am. Chem. Soc. 1948, 70, 3121–
3125. (b) Silverman, R. B.; Levy, M. A. J. Org. Chem. 1980, 45,
815–818.
(26) Molina, M. T.; Del Valle, C.; Escribano, A. M.; Ezquerra, J.; Pedregal,
(29) Jain, R. Org. Prep. Proced. Int. 2001, 33, 405–409.
(30) Williams, M. A.; Rapoport, H. J. Org. Chem. 1993, 58, 1151–1158.
C. Tetrahedron 1993, 49, 3801–3808.
742
•
Vol. 13, No. 4, 2009 / Organic Process Research & Development

Scheme 6a
a Reagents and yields: (i) tert-butyl bromoacetate/DMA/Na2CO3 then H2O, 76.5%; (ii) MsCl/EtOAc/DIPEA, chromatography, 71%; (iii) 12/toluene, 77%; (iv) (1)
MeSO3H/MeOH, 85%; or (2) HCl/MeOH, 90%; (3) 37% aq HCl/MeOH, 80%; (v) 17/EtOAc/DIPEA, 75%; (vi) LiOH/dioxane, 90%; (vii) H2O/Gd2O3/NaOH, 83%.
resin. Consequently, the chromatographic purification was
postponed to the following step.
other stereogenic carbon atoms contained in the stereoidic
residue was deemed unlikely. Besides the compounds of
interest, i.e. the 3ꢀ(S) isomers reported in Scheme 5 (note: the
stereochemical configuration of the stereogenic carbon atom
present in the residue linked to the position 3 of the cholanoic
skeleton is indicated in parentheses), the 3ꢀ(R) and/or the 3ꢀ(RS)
epimeric mixtures were also prepared while for the 3R isomers
only the combination with the (S) products was taken into
account. Indeed, the formation of 3R(R) isomers during the
synthesis was considered very unlikely because they would
derive from reactions between impurities. Furthermore, it is
noteworthy that racemisation at the stereogenic centre in the
DTPA moiety can in principle occur under strongly acidic or
basic conditions. Conversely, racemisation at the stereogenic
carbon in position 3 of the cholanoic moiety, and therefore the
conversion of 3ꢀ into 3R isomers, is extremely unlikely. The
entry to 3R derivatives stemmed from the transformation of 5,
under Mitsunobu conditions, into the inverted formyloxy
derivative31 (Ph3P/DEAD/HCOOH/THF, 95%) which was
deprotected with HCl/MeOH (70%) to the 3ꢀ-OH epimer. From
there on, the path through the azide depicted in Scheme 3 was
followed so that the subsequent Mitsunobu reaction (Ph3P/
DEAD/DPPA/THF, 73%) again inverted the configuration,
affording the 3R-azide, in its turn reduced (Ph3P/CH3CN then
H2O, 66%) to (3R,5ꢀ,12R)-3-amino-12-hydroxycholan-24-oic
acid methyl ester. The latter was the starting material for the
3R(S) isomers of 15, 18, 19, 4, and B22956/1. The 3ꢀ(R)
isomers were prepared, according to Scheme 5, using compound
12 with R configuration.
For the deprotection of the t-butyl esters of 19, the use of
trifluoroacetic acid, although very efficient, was expensive and
needed a further alkaline deprotection step for the methyl esters.
Accordingly, we studied the possibility to deprotect both methyl
and tert-butyl esters with a single reagent, and among others,
LiOH proved the most effective. Compound 4 was at first
obtained by treatment of pure, isolated 19 with LiOH in dioxane
(rt, 72 h) followed by precipitation at pH 1.5 (90% yield). Later
on, dioxane was replaced with the safer i-PrOH, crude 19 was
directly used and NaOH employed as the base of choice.
Besides being cheaper than LiOH, NaOH was preferable
because sodium was the cation already present in the final
product. The racemisation at the stereogenic centre in the DTPA
moiety slightly increased, passing from LiOH to NaOH, but
the epimer content could be maintained below 2%. In particular,
we used at first i-PrOH/H2O/LiOH, 25 °C, 20 h then i-PrOH/
H2O/NaOH, 15-20 °C, 3 h then 25 °C, 20 h. The solution
containing crude 4 was then purified by elution through
Amberlite XAD 16.00 loading the crude at neutral pH. This
allowed to absorb on the resin the lipophilic impurities only
and to remove the hydrophilic ones in the early eluting fractions.
Compound 4 was recovered, in 60% yield (from 18), with a
total impurity content lower than 2%.
B22956/1 was prepared by complexation of 4 with Gd2O3
in the presence of NaOH, and then it was crystallized from
acetone/water (90% overall), a procedure which increased the
purity of the isolated compound, especially lowering the 3ꢀ(R)
epimer content.
Stereochemical Issues. Besides controlling the chemical and
stereochemical purity of the final product, we also needed to
check the stereochemical purity of most intermediates, focusing
on the two stereogenic carbon atoms which are likely to undergo
stereochemical inversion under the experimental conditions
employed. For the sake of clarity, we refer to the stereogenic
carbon atom of the DTPA residue and the carbon atom in
position 3 of the cholanoic moiety. Possible inversion at all the
Conclusions
The discovery route to B22956/1 through the versatile
intermediate 1 was totally revised during development. Indeed,
a different disconnection approach was pursued. L-Glutamic acid
was conveniently incorporated into the scaffold after conversion
(31) Bose, A. K.; Lal, B.; Hofman, W. A.; Manhas, M. S. Tetrahedron
Lett. 1973, 18, 1619–1622.
Vol. 13, No. 4, 2009 / Organic Process Research & Development
•
743

into an ester of N-Boc protected pyroglutamic acid. This species
proved very effective for a smooth coupling with aminodeoxy-
cholate 2. The construction of the DTPA moiety was then
achieved following a Rapoport-like approach. The new route
allowed to produce multikilogram batches of B22956/1 for the
clinical trials in 20% overall yield from (3R,5ꢀ,12R)-3,12-
dihydroxycholan-24-oic acid.
Transformation of 5 into 9 without Isolation of 8. A
solution of 5.0.5 MeOH (75 g; 0.177 mol) in DMA (240 mL)
was concentrated under reduced pressure to 230 g and then
diluted with DMA (60 mL) and cooled to rt. Triphenylphos-
phine (52.5 g; 0.2 mol) and phthalimide (28.5 g; 0.194 mol)
were added, the resulting solution was cooled to 15 °C, and
then a solution of diisopropyl azodicarboxylate (39.9 g; 0.197
mol) in DMA (70 mL) was dropped over 75 min maintaining
the temperature at 15-20 °C. After 24 h at rt the reaction
mixture was diluted with DMA (0.33 L), vigorously stirred at
35 °C, and diluted with 2 M phosphate buffer (pH 8, 0.25 L).
Solid NaBH4 (7.5 g; 0.198 mol) was added over 5 min, and
the suspension was stirred at 48 °C for 2 h. AcOH (22 mL)
was added to destroy the hydride excess, and in the meanwhile
a clear solid started to crystallize. The suspension was stirred
for 15 h at rt, and then the solid was filtered, washed at first
with 65/25 DMA/H2O (2 × 90 mL), and then with H2O (2 ×
100 mL). The wet solid was suspended in H2O (0.4 L),
vigorously stirred for 2 h, filtered, washed with H2O (3 × 50
mL), and dried to afford 9 (73.9 g; 0.137 mol) as a white solid.
Yield 77%. HPLC 95.7% (method A).
(3ꢀ,5ꢀ,12r)-3-Amino-12-hydroxycholan-24-oic Acid Meth-
yl Ester (2). A suspension of 9 (4.38 kg; 8.11 mol) in MeOH
(22 L) and 34% HCl (1.13 kg; 10.5 mol) was heated to 65 °C
under stirring. In about 1.5 h a solution was obtained, and after
an additional 1.5 h, the reaction mixture was cooled to 20 °C
and the pH adjusted to about 3.5 with 30% NaOH. The solvent
was evaporated at reduced pressure to about half volume;
n-BuOAc (31 L) was added, and distillation was continued to
eliminate MeOH. After 30 min at 50 °C, the solution was stirred
at 17-20 °C for 3 h, and then the precipitated hydrochloride
was filtered and washed with n-BuOAc. The wet product was
dissolved in MeOH (15.8 L), and then 30% NaOH was added
to completely neutralise the hydrochloric acid present in the
salt, which had been previously titrated, and to reach pH
9.8-10.1. Water (37.5 kg) was then added at 20 °C to obtain
crystallization, and after 2 h, the suspension was filtered. The
solid was washed first with 7/3 H2O/MeOH (11 L) and then
with H2O (2 × 31 L) and dried to afford 2 (2.8 kg; 6.9 mol) as
a white solid. Yield 85%. Mp 155-155.5 °C; TLC Rf 0.69
(eluent D); HPLC 96.5% (method B); GC 96.3% (method A).
(S)-5-Oxo-1,2-pyrrolidinedicarboxylic Acid 1-(1,1-Dim-
ethylethyl) 2-Methyl Ester (12). SOCl2 (732 g; 6.15 mol) was
added over 2 h to a suspension of L-glutamic acid hydrochloride
(551 g; 3 mol) in MeOH (3 L) stirred at 0-5 °C. After about
3.5 h at rt the reaction mixture turned into a clear solution that
was stirred for 20 h. The solvent was evaporated to give crude
L-glutamic acid dimethyl ester hydrochloride (650.8 g, TLC:
Rf 0.79 (eluent E); argentometric titer: 105.3%) as a thick oil.
Into a solution of the latter in MeOH (1 L), was added 3 M
KOH in MeOH (1.08 L; 3.24 mol) over 1.5 h, causing the
precipitation of KCl that was filtered off. The clear solution
was concentrated and again filtered and evaporated to a residue
(600 g) which was heated at 115 °C at atmospheric pressure
for about 1 h while distilling off MeOH produced by the
cyclization reaction. The residue containing 10 (bp 127-130
°C/0.2 kPa; TLC: Rf 0.71 (eluent E); HPLC: S/R ratio 99.0/1.0
with method C) was diluted with EtOAc (3 L) and DMAP (7.3
Experimental Section
(3r,5ꢀ,12r)-3,12-Dihydroxycholan-24-oic acid methyl
ester (5). Monohydrate p-toluenesulfonic acid (152.7 g; 0.8 mol)
was added to a suspension of (3R,5ꢀ,12R)-3,12-dihydroxy-
cholan-24-oic acid (1570.3 g; 4.0 mol) in MeOH (6.15 L) stirred
at rt. After 1 h a solution was obtained, and after additional
1 h, the desired product started to precipitate. The suspension
was stirred for 23 h and then filtered, and the crystalline solid
was suspended in H2O (3 L) and stirred for 1 h. After filtration,
washings with H2O (3 × 0.8 L), and drying, 5 (1205.9 g; 2.85
mol) was obtained as a white solid containing 0.5 mol MeOH.
The mother liquor was concentrated to 2.5 kg and stirred at rt
for 20 h. Using the above-described methodology a second crop
of 5.0.5 MeOH (383.3 g; 0.907 mol) was obtained. Overall yield
94%. Mp 100-108 °C (Mp 67-70 °C for the compound
devoid of MeOH) (in the literature,32 mp 70-108 °C is
reported); TLC: Rf 0.39 (eluent A).
(3ꢀ,5ꢀ,12r)-3-(1,3-Dihydro-1,3-dioxo-2H-isoindol-2-yl)-
12-hydroxycholan-24-oic Acid Methyl Ester (8). Into a
solution of 5 (512 g; 1.259 mol, as a product without MeOH
was used) and triphenylphosphine (372.7 g; 1.421 mol) in THF
(1.5 L), phthalimide (202.8 g; 1.378 mol) was suspended, and
then a solution of diisopropyl azodicarboxylate (284.7 g; 1.408
mol) in THF (0.49 L) was dropped over 1.5 h, maintaining the
temperature of the reaction mixture at 15 °C. The solution thus
obtained was left at rt for 18 h, and then the solvent was distilled,
and the oily crude (1518 g) was dissolved with MeOH (3.2 L)
and stirred for 20 h. The crystalline solid which precipitated
was filtered, washed with MeOH (700 + 2 × 350 mL), and
dried to afford 8 (481.3 g; 0.898 mol) as a white solid. The
mother liquor and the washings were concentrated to 2.35 kg,
and after stirring for 20 h at 3 °C, a second crop of 8 (61.6 g;
0.114 mol) was recovered. Overall yield 80%. Mp 161.1-161.8
°C; TLC Rf 0.90 (eluent D); HPLC 98.7% (method A).
(3ꢀ,5ꢀ,12r)-12-Hydroxy-3-[[2-(hydroxymethyl)benzoyl]-
amino]cholan-24-oic Acid Methyl Ester (9). A solution of 8
(225 g; 0.42 mol) in DMA (1.75 L), vigorously stirred at 34
°C, was diluted with 2 M phosphate buffer (pH 8, 0.5 L) and
a cloudy mixture at 43 °C was obtained. Solid NaBH4 (15.45
g; 0.408 mol) was added in 5 min, and the suspension was
stirred at 47-48 °C. After 55 min an almost clear solution was
obtained in which AcOH (50 mL) was added to neutralise the
hydride excess and to correct the pH, from 12.8 to about 8.
The mixture was poured into H2O (4.6 L) and stirred for 3 h,
the precipitate was filtered and suspended in H2O (4 L) then
stirred for 30 min. Finally, the precipitate was filtered, washed
with H2O (1.5 L), and dried to obtain 9 (224.6 g; 0.416 mol)
as a white solid. Yield 99%. Mp 190.2-192.7 °C; TLC Rf 0.80
(eluent D); HPLC 93% (method A).
(32) Chang, F. C.; Wood, N. F.; Holton, W. G. J. Org. Chem. 1965, 30,
1718–1723.
744
•
Vol. 13, No. 4, 2009 / Organic Process Research & Development

g; 0.06 mol) was added, then di-tert-butyl dicarbonate (673 g;
3.1 mol) was added, over 1 h, to the mixture stirred at 15-17
°C. After 3 h the solution was washed with 0.066 M phosphate
buffer (pH 5.8, 3 × 0.7 L) and water (2 × 0.5 L) and was
dried (Na2SO4). The filtered solution was evaporated to a thick
oil (670 g) which was taken with EtOAc (0.35 L), obtaining a
solution that was diluted with n-hexane (0.95 L). After 24 h at
rt the suspension was filtered, and the solid was washed with
n-hexane (3 × 0.25 L) and dried (320.8 g; 1.318 mol). The
mother liquor and the washings were combined and evaporated,
and then the residue was analogously crystallized to afford a
second crop of 12 having similar purity (160.8 g; 0.66 mol).
Overall yield 66%. Mp 70-71.5 °C (lit.33 72-72.5 °C); TLC
Rf 0.74 (eluent G); HPLC 99.6% with method A; HPLC S/R
ratio 100/0 with method D.
(3ꢀ,5ꢀ,12r)-12-Hydroxy-3-[[5-methoxy-1,5-dioxo-(4S)-4-
[[(1,1-dimethylethoxy)carbonyl]amino]pentyl]amino]cholan-
24-oic Acid Methyl Ester (15). A suspension of 2 (5.36 kg;
13.2 mol) and 12 (3.31 kg; 13.6 mol) in toluene (12.3 L) was
slowly heated to 90 °C to obtain a solution which was kept at
the same temperature for 24 h. The temperature was slowly
decreased over 4 h to 17 °C, and the suspension thus obtained
was stirred at the same temperature for 4 h. After filtration and
washing with toluene (6 L), the wet product was analogously
crystallized with toluene (12.5 L) to afford, after drying, 15
(7.35 kg) as a white solid. The compound contained toluene
(about 10%, roughly corresponding to 0.75 mol), so the yield
based on the dry product (6.6 kg) resulted 77%. Mp 172-173
°C; TLC Rf 0.76 (eluent D); HPLC 99.0% with method A.
(3ꢀ,5ꢀ,12r)-3-[[(4S)-4-Amino-5-methoxy-1,5-dioxopentyl]-
amino]-12-hydroxycholan-24-oic Acid Methyl Ester (18).
MeSO3H (1.12 kg; 11.7 mol) was added to a solution of 15
(6.5 kg corresponding to 5.85 kg without toluene; 9 mol) in
MeOH (17 L), and then the reaction mixture was stirred at 50
°C for 3 h. After cooling to 10 °C, DIPEA (1.51 kg; 11.7 mol)
was added, the temperature was further lowered to 0 °C, and
H2O (14.5 L) was dropped in, causing the precipitation of a
white solid. The suspension was stirred overnight at 0 °C and
filtered, and the solid, washed with 1/1 MeOH/H2O (3 L) and
H2O (3 L), was dried to afford 18 (4.2 kg; 7.65 mol). Yield
85%. Mp 121 °C; TLC Rf 0.64 (eluent D); HPLC 96.5%
(method A).
of 16 (5 g; 17.3 mmol) and DIPEA (2.89 mL; 20.7 mmol) in
EtOAc (12 mL). The resulting suspension was stirred at 0 °C
for 1.5 h, then the precipitate was filtered and washed with cold
EtOAc (2 × 20 mL). The clear solution was evaporated under
reduced pressure, and the crude was purified by silica gel flash-
chromatography (7/3 n-hexane/EtOAc) to afford 17 (4.5 g;
12.25 mmol) as a colourless oil. Yield 71%. TLC: Rf 0.27
(eluent C).
(3ꢀ,5ꢀ,12r)-3-[[(4S)-4-[Bis[2-[bis[2-(1,1-dimethylethoxy)-
2-oxoethyl]amino]ethyl]amino]-5-methoxy-1,5-dioxopentyl]-
amino]-12-hydroxycholan-24-oic Acid Methyl Ester (19).
MsCl (9 mL; 138 mmol) was dropped at -10 °C into a solution
of 16 (33.3 g; 115 mmol) and DIPEA (23.5 mL; 138 mmol) in
EtOAc (80 mL). The resulting suspension was stirred at 0 °C
for 1.5 h, and then the precipitate was filtered and washed with
cold EtOAc (2 × 50 mL). The clear solution was diluted with
EtOAc (40 mL), and then 18 (27.4 g; 50 mmol) and DIPEA
(23.5 mL; 138 mmol) were added. The mixture was stirred for
18 h at 50 °C, cooled to rt, and filtered, and the solid was
washed with EtOAc (2 × 50 mL). The mother liquor and the
washings were evaporated to dryness, and the crude was purified
by silica gel flash chromatography (gradient elution 7/3 f 2/8
n-hexane/EtOAc) to afford 19 (41.2 g; 38 mmol) as a white
solid. Yield 75%. Mp 56-58 °C; TLC Rf 0.45 (eluent A);
HPLC 99.3% (method A).
(3ꢀ,5ꢀ,12r)-3-[[(4S)-4-[Bis[2-[bis(carboxymethyl)amino]-
ethyl]amino]-4-carboxy-1-oxobutyl]amino]-12-hydroxycholan-
24-oic Acid (4) by Reaction of 17, Directly Obtained from
16, with 18 and Hydrolysis of Not Isolated 19. MsCl (1.19
kg; 10.4 mol) was dropped in 1 h into a solution of 16 (3.02
kg; 10.4 mol) and DIPEA (1.55 kg; 12 mol) in n-BuOAc (6.8
L), initially stirred at -5 °C, maintaining the temperature around
0 °C. The resulting suspension was stirred at 0-2 °C for 1.5 h,
and then the precipitate was filtered and washed with cold
n-BuOAc (2 × 2.3 L). The solution containing 17 was loaded
at 17 °C, washing the reactor with n-BuOAc (3.4 L), into a
solution of 18 (2.6 kg; 4.74 mol) and DIPEA (1.62 kg; 12.5
mol) in n-BuOAc (6 L). After stirring for 20 h at 50 °C and
7 h at 80 °C, the mixture was cooled to rt and filtered, and the
solid was washed with cold n-BuOAc (2 × 1.8 L). The mother
liquor and the washings were collected, and the solvent was
evaporated under reduced pressure to the minimum stirred
volume, and then 2-PrOH (8.5 L) was added. The mixture was
concentrated at reduced pressure, 2-PrOH (4.2 L) was added
and the procedure repeated. Finally, the residue was dissolved
with 2-PrOH (8.6 L) and the solution diluted with H2O (32 L).
Into the mixture containing crude 19, 30% NaOH (9.5 kg; 71
mol) was loaded in 0.5 h at 17 °C obtaining a complete
dissolution of the initially formed two phases after 3 h at the
same temperature. The solution was stirred at 25 °C for 22 h,
and then the pH was adjusted to 6.5-7.0 with 34% HCl, and
2-PrOH was evaporated until there was a concentration around
10% in the final product. The mixture was then loaded onto an
Amberlite XAD 16.00 column (118 L) which was eluted with
water at a flow rate of 1 BV/h. The eluate was partially desalted
and concentrated by nanofiltration to about 50 kg, then slowly
added at 20 °C into a 0.5 M aqueous solution of HCl (42.6 L;
21.3 mol). The suspension was centrifuged, and the solid was
N-[2-(1,1-Dimethylethoxy)-2-oxoethyl]-N-(2-hydroxyeth-
yl)glycine 1,1-Dimethylethyl Ester (16). Na2CO3 (0.667 kg;
6.29 mol) was added to a solution of tert-butyl bromoacetate
(1.17 kg; 5.99 mol) in DMA (0.77 kg), and then ethanolamine
(0.19 kg; 3.11 mol) was dropped in about 2 h, keeping the
temperature below 40 °C. After 22 h at 40 °C and cooling to
20 °C, H2O (7 kg) was added in order to dissolve the salts and
induce the precipitation of 16. The suspension was kept at 17
°C for 2 h and then filtered, and the solid was washed with
H2O (1.6 kg) and dried to afford 16 (6.9 kg; 2.38 mol). Yield
76.5%. Mp 55-60 °C; GC 99.8% (method B).
N-[2-(1,1-Dimethylethoxy)-2-oxoethyl]-N-[2-[(methylsul-
fonyl)oxy]ethyl]glycine 1,1-Dimethylethyl Ester (17). MsCl
(1.4 mL; 18.1 mmol) was dropped at -10 °C into a solution
(33) Yoshifuji, S.; Tanaka, K.; Kawai, T.; Nitta, Y. Chem. Pharm. Bull.
1986, 34, 3873–3878.
Vol. 13, No. 4, 2009 / Organic Process Research & Development
•
745

washed with H2O (2 × 9 L), suspended in H2O (40 L), and
again centrifuged and washed with H2O (9 + 4 L). After drying
at 35 °C under reduced pressure, 4 (2.39 kg; 2.84 mol) was
obtained as a white solid. Yield 60%. Mp 185-190 °C dec;
TLC Rf 0.16 (eluent F); HPLC 98.9% (method E).
°C which was subsequently cooled in about 6 h to 17 °C. After
12 h stirring at the same temperature, a second portion of
acetone (9.1 L) was added over 2 h, and the suspension stirred
for 1 h. The subsequent filtration followed by washings with
acetone (2 × 4 L) and drying afforded B22956/1 (2.7 kg; corr.
to 2.51 kg anhydrous; 2.37 mol) as a white solid. Yield 82.9%.
Mp > 300 °C; HPLC 99.8% (method F).
Trisodium [(3ꢀ,5ꢀ,12r)-3-[[4(S)-4-[Bis[2-[bis[(carboxy-kO)-
methyl]amino-kN]ethyl]amino-kN]-4-(carboxy-kO)-1-oxobutyl]-
amino]-12-hydroxycholan-24-oato(6-)]gadolinate(3-)] (B22956/
1). A suspension of 4 (2.4 kg; 2.86 mol) in H2O (15 L) was
dissolved with 30% NaOH (1.03 kg; 7.72 mol), adjusting the
pH to 4.8, and then Gd2O3 (0.526 kg; 1.45 mol) was added.
The suspension was stirred for 15 h at 50 °C until complete
dissolution, and then the pH was corrected to 6.7-7.2 with 30%
NaOH (0.115 kg; 0.86 mol) and the solution heated to 100 °C
for 1 h. After cooling to 50 °C, decolourisation with Carbopuron
4N and filtration, the solution was concentrated up to 28 ( 3%
of the initial weight by evaporation at reduced pressure. Acetone
(9.1 L) was added in about 0.5 h to the solution stirred at 48
Supporting Information Available
Experimental general information as well as NMR spectra,
MS spectra and elemental analyses of the compounds described
in the Experimental Section. This material is available free of
charge via the Internet at http://pubs.acs.org.
Received for review January 7, 2009.
OP900008A
746
•
Vol. 13, No. 4, 2009 / Organic Process Research & Development