Synthesis and Evaluation of Water-Soluble Prodrugs of Ursodeoxycholic Acid (UDCA), an Anti-apoptotic Bile Acid

Article abstract of DOI:10.1002/cmdc.201300059

Ursodeoxycholic acid (UDCA) is a bile acid with demonstrated anti-apoptotic activity in both invitro and invivo models. However, its utility is hampered by limited aqueous solubility. As such, water-soluble prodrugs of UDCA could have an advantage over the parent bile acid in indications where intravenous administration might be preferable, such as decreasing damage from stroke or acute kidney injury. Five phosphate prodrugs were synthesized, including one incorporating a novel phosphoryloxymethyl carboxylate (POMC) moiety. These prodrugs were highly water-soluble, but showed significant differences in chemical stability, with oxymethylphosphate prodrugs being the most unstable. In a series of NMR experiments, the POMC prodrug was bioactivated to UDCA by alkaline phosphatase (AP) faster than a prodrug containing a phosphate directly attached to the alcohol at the 3-position of UDCA. Both of these prodrugs showed significant anti-apoptotic activity in a series of invitro assays, although the POMC prodrug required the addition of AP for activity, while the other compound was active without exogenous AP.

Full text of DOI:10.1002/cmdc.201300059

CHEMMEDCHEM
FULL PAPERS
DOI: 10.1002/cmdc.201300059
Synthesis and Evaluation of Water-Soluble Prodrugs of
Ursodeoxycholic Acid (UDCA), an Anti-apoptotic Bile Acid
Peter I. Dosa,*[a] Tim Ward,[a] Rui E. Castro,[b] Cecꢀlia M. P. Rodrigues,[b] and Clifford J. Steer[c]
Ursodeoxycholic acid (UDCA) is a bile acid with demonstrated
anti-apoptotic activity in both in vitro and in vivo models.
However, its utility is hampered by limited aqueous solubility.
As such, water-soluble prodrugs of UDCA could have an ad-
vantage over the parent bile acid in indications where intrave-
nous administration might be preferable, such as decreasing
damage from stroke or acute kidney injury. Five phosphate
prodrugs were synthesized, including one incorporating
a novel phosphoryloxymethyl carboxylate (POMC) moiety.
These prodrugs were highly water-soluble, but showed signifi-
cant differences in chemical stability, with oxymethylphosphate
prodrugs being the most unstable. In a series of NMR experi-
ments, the POMC prodrug was bioactivated to UDCA by alka-
line phosphatase (AP) faster than a prodrug containing a phos-
phate directly attached to the alcohol at the 3-position of
UDCA. Both of these prodrugs showed significant anti-apop-
totic activity in a series of in vitro assays, although the POMC
prodrug required the addition of AP for activity, while the
other compound was active without exogenous AP.
Introduction
Ursodeoxycholic acid (UDCA) is a naturally occurring compo-
nent of bile used clinically to treat primary biliary cirrhosis. In
addition to treating various other liver diseases, UDCA and its
taurine-conjugated derivative tauroursodeoxycholic acid
(TUDCA) are anti-apoptotic agents[1,2] that are known to have
protective effects in animal models of multiple disorders, in-
cluding Huntington’s disease,[3] spinal cord injury,[4] cataracts,[5]
and acute pancreatitis.[6] The molecular mechanisms underlying
tion caused by endoplasmic reticulum (ER) stress, and by mod-
ulating gene regulation.[1,2,7,8]
Previous research from our groups has shown that TUDCA
has beneficial effects in animal models of ischemic and hemor-
rhagic stroke,[9,10] acute kidney injury,[11] and myocardial infarc-
tion.[12] However, we were unable to test UDCA in these
models because its limited aqueous solubility precluded the
administration, preferably intravenously (iv) for these indica-
tions, of doses large enough to have an effect. This is unfortu-
nate as UDCA is an FDA-approved drug in the U.S. with an ex-
cellent safety record in humans even when used for extended
periods of time at high doses,[13–16] making it a particularly at-
tractive potential therapy. With this in mind, we set out to syn-
thesize highly water-soluble prodrugs of UDCA. We chose to
focus on synthesizing phosphate ester prodrugs as this modifi-
cation often increases aqueous solubility by several orders of
magnitude. There are several examples of this class of pro-
drugs already in clinical use, including prednisolone phos-
phate, fosphenytoin, and fosamprenavir.[17–20] These prodrugs
are activated in vivo by ubiquitous endogenous phosphatases.
A particular advantage of this class of prodrugs is that they
show very little interspecies differences in rates of bioactiva-
tion. This stands in contrast to other potential classes of pro-
drugs of UDCA such as carboxylic acid esters and could make
a phosphate prodrug easier to develop.[17]
the cytoprotective activities of these molecules are believed to
engage a number of different pathways including preventing
Bax-induced membrane perturbation in mitochondria, blocking
caspase-3 activation, inhibiting calpain and caspase-12 activa-
[a] Dr. P. I. Dosa, Dr. T. Ward
Department of Medicinal Chemistry
Institute for Therapeutics Discovery and Development
University of Minnesota, 717 Delaware Street, Minneapolis, MN 55414 (USA)
E-mail: pidosa@umn.edu
Results and Discussion
[b] Dr. R. E. Castro, Prof. Dr. C. M. P. Rodrigues
Research Institute for Medicines and Pharmaceutical Sciences (iMed.UL)
Faculty of Pharmacy, University of Lisbon, 1649-003 Lisbon (Portugal)
Synthesis of prodrugs
[c] Prof. Dr. C. J. Steer
Ursodeoxycholic acid has two alcohol moieties to which
a phosphate can be directly attached, located at the 3- and 7-
positions. We began our synthesis of both of these potential
Departments of Medicine and Genetics, Cell Biology, and Development
University of Minnesota, VFW Cancer Research Center
406 Harvard Street, Minneapolis MN 55455 (USA)
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2013, 8, 1002 – 1011 1002

CHEMMEDCHEM
FULL PAPERS
www.chemmedchem.org
prodrugs by benzyl protecting the acid of UDCA, which pro-
ceeded in high yield using benzyl bromide (BnBr) as the alky-
lating agent (Scheme 1). Heating the resulting benzyl ester
1 with dibenzyl N,N-diethylphosphoramidite followed by oxida-
tion with H2O2 furnished a phosphate ester which was tenta-
tography and allowed us to unambiguously confirm the regio-
1
chemistry of our prodrugs by H NMR analysis, as the signal of
the proton next to the Cbz-protected alcohol (Ha in Scheme 2)
was a dddd in the major mono-substituted product (consistent
with structure 4) and a ddd in the minor mono-substituted
product (consistent with structure 5). Compound 4 was then
converted into 7-substituted phosphate ester prodrug 7 using
dibenzyl N,N-diethylphosphoramidite followed by oxidation
with H2O2 and then Pd/C catalyzed debenzylation. Similar stan-
dard conditions converted 5 into the same 3-substituted phos-
phate ester prodrug 3 that was obtained using Scheme 1. Both
phosphate prodrugs were highly water-soluble, rapidly dissolv-
ing at all concentrations tested (up to 20 mgmLꢀ1), and were
stable in solution for extended periods of time (>6 months)
without any apparent decomposition.
In addition to prodrugs 3 and 7, where the phosphate is di-
rectly linked to one of the alcohols in UDCA, we also set out to
synthesize 3- and 7-substituted oxymethylphosphate (OMP)
UDCA prodrugs. While often more difficult to synthesize, OMP
prodrugs (also referred to a phosphonooxymethyl or POM pro-
drugs) are typically bioactivated by alkaline phosphatase at
a significantly higher rate than their directly linked phosphate
ester analogues due to decreased steric hindrance, which
would be preferred for rapid treatment of stroke or myocardial
infarction.[19] Upon bioactivation, OMP prodrugs release parent
drug and formaldehyde in a two-step process.[23]
Scheme 1. Synthesis of 3-substituted phosphate prodrug 3. Reagents and
conditions: a) BnBr, K2CO3, CH3CN, 808C; b) 1. dibenzyl N,N-diethylphosphor-
amidite, 1,2,4-triazole, NaHCO3, 1,2-dichloroethane, 658C, 2. 30% H2O2, 08C;
c) 1. Pd/C, H2, MeOH, 2. Na2CO3.
tively assigned the structure 2 based on reports that similar
steroidal structures react more readily at the 3-position than at
the 7-position.[21,22] This assignment was later confirmed by
NMR spectroscopy (see below). Removal of the three benzyl
groups of 2 with hydrogen and Pd/C followed by treatment
with sodium carbonate yielded the desired 3-substituted phos-
phate ester prodrug of UDCA (3).
To obtain the 7-substituted phosphate ester prodrug of
UDCA we treated benzyl ester 1 with benzyl chloroformate
and pyridine in dichloromethane (Scheme 2). This led to a mix-
ture of products, including 3-Cbz-protected alcohol 4 (41%), 7-
Cbz-protected alcohol 5 (10%), recovered starting material
(41%) and a small amount of 3,7-diCbz-protected material.
These products could readily be separated by column chroma-
The synthesis of the 3- and 7-substituted oxymethylphos-
phate (OMP) prodrugs of UDCA did indeed prove to be consid-
erably more complicated than the synthesis of the directly
linked phosphate prodrugs 3 and 7. Attempts to directly alky-
late the UDCA scaffold at either the 3- or 7-positions with
either dibenzyl chloromethyl phosphate or chloroiodomethane
were unsuccessful. Instead, we turned to a synthetic scheme
that had previously be used to synthesize OMP prodrugs,
namely methylthiomethyl (MTM) ether formation followed by
reaction with N-iodosuccinimide (NIS) and a phosphate.[22,24,25]
We successfully synthesized the desired MTM ether intermedi-
ate 9 via a Pummerer rearrangement by stirring 4 in DMSO,
acetic anhydride (Ac2O), and acetic acid (AcOH) (Scheme 3).
Scheme 2. Synthesis of 7-substituted phosphate prodrug 7 and alternate route to prodrug 3. Reagents and conditions: a) benzyl chloroformate, pyridine,
CH2Cl2; b) 1. dibenzyl N,N-diethylphosphoramidite, 1,2,4-triazole, NaHCO3, D, 2. 30% H2O2, 08C; c) 1. Pd/C, H2, MeOH, 2. Na2CO3.
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2013, 8, 1002 – 1011 1003

CHEMMEDCHEM
FULL PAPERS
www.chemmedchem.org
Scheme 3. Synthesis of 7-substituted oxymethylphosphate prodrug 13. Reagents and conditions: a) DMSO, Ac2O, AcOH; b) dibenzyl phosphate, N-iodosuccini-
mide, molecular sieves; c) H3PO4, N-iodosuccinimide, molecular sieves d) SOCl2, CH2Cl2, 1008C; e) dibenzyl phosphate, K2CO3, [D3]CH3CN; f) H3PO4, NBu3,
CH3CN; g) Pd/C, H2, MeOH; h) ion exchange using Na+-Dowex resin, MeOH/H2O.
Unfortunately, treating 9 with NIS and either dibenzyl phos-
phate or H3PO4 did not lead to any isolable product. However,
we were able to convert the MTM ether 9 into chloroalkyl
ether 11 by heating it in CH2Cl2 and thionyl chloride. In an
NMR experiment, reaction of chloroalkyl ether 11 with dibenzyl
phosphate and K2CO3 in [D3]acetonitrile led to an impure prod-
uct (likely 10) which decomposed before it could be isolated.
Similarly, reaction of 11 with either K3PO4 or Na3PO4 failed to
lead to the desired OMP product. We were finally able to suc-
cessfully substitute 11 by following the example of Komatsu
and co-workers, who found that a tri(n-butyl)amine salt of
phosphate could be successfully reacted with a chloroalkyl
ether, presumably because of its improved solubility in organic
solvents.[26,27] Thus, stirring 11 with a tri(n-butyl)amine salt of
phosphate in acetonitrile led to 12, which was then deprotect-
ed using hydrogen and Pd/C in methanol. The crude material
was purified by C18 column to afford 7-substituted OMP pro-
drug 13a as an NBu3 salt. Similarly, 3-substituted OMP prodrug
16 could be obtained from compound 5 using the same se-
quence of synthetic steps (Scheme 4).
Scheme 4. Synthesis of 3-substituted oxymethylphosphate prodrug 16. Re-
agents and conditions: a) DMSO, Ac2O, AcOH; b) SOCl2, CH2Cl2, 1008C;
c) 1. H3PO4, NBu3, CH3CN, 2. Pd/C, H2, MeOH.
zymatic activation. We are aware of only one example of such
a phosphoryloxymethyl carboxylate (POMC) prodrug in the
chemical literature, in a recent patent application by Barnes
and co-workers.[29] However, no discussion of the properties of
the potential prodrug was presented other than to mention
that the material was not obtained cleanly. In addition, Stella
and co-workers explored related phosphoryloxymethyloxy car-
bonyl prodrugs of alcohols, aliphatic amines and aromatic
amines, but found their potential utility limited by chemical in-
stability.[30]
The 3- and 7-substituted oxymethylphosphate prodrugs 16
and 13a were poorly water-soluble as NBu3 salts. Therefore,
compound 13a was converted into a sodium salt 13b by ion-
exchange filtration through Dowex resin.[28] The resulting white
solid rapidly dissolved in water at all concentrations tested (up
to 10 mgmLꢀ1). Unfortunately, a significant portion of the ma-
terial decomposed when left in D2O solution overnight.
Due to the combination of chemical instability and the rela-
tively difficult synthesis of the 3- and 7-substituted OMP pro-
drugs, we instead decided to prepare a prodrug where the
OMP group is linked to the carboxylic acid of UDCA instead of
one of its alcohols. Such a prodrug could potentially be bioac-
tivated in vivo to parent drug both by alkaline phosphatase
and by esterases and has the additional advantage that the
phosphate moiety is sterically unhindered (relative to the
phosphate group in 3 or 7), which may increase the rate of en-
We began our synthesis of the POMC prodrug by reacting
UDCA with K2CO3 and dibenzyl chloromethyl phosphate (17)
to afford ester 18 (Scheme 5). Interestingly, this reaction pro-
ceeded in higher yield (81% instead of 22%) and at much
lower temperature (room temperature instead of 1208C) when
DMF was used as a solvent instead of acetonitrile. Using DMF
instead of acetonitrile also greatly minimized the formation of
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2013, 8, 1002 – 1011 1004

CHEMMEDCHEM
FULL PAPERS
www.chemmedchem.org
occurs primarily via an intramolecular general base or
intramolecular nucleophilic catalysis mechanism.[30]
This hypothesis is further supported by data showing
that adding an additional equivalent of Tris to 19b
has little effect on its stability in solution (Table 1,
entry 19e). The compound also showed similar stabil-
ity in pH 7.4 tris buffer, with 81% remaining after one
day at RT.
Alkaline phosphatase activation of prodrugs
To determine whether the POMC prodrug was
indeed activated under in vitro conditions faster than
a prodrug where the phosphate moiety is directly
linked to an alcohol, we conducted a series of experi-
ments where we monitored the alkaline phosphatase
catalyzed activation of prodrugs 19d and 3 by inverse-gated
decoupled 31P NMR (see Experimental Section for details). As
shown in Figure 1, UDCA is more rapidly released from pro-
drug 19d under in vitro conditions than prodrug 3.
Scheme 5. Synthesis of phosphoryloxymethyl carboxylate (POMC) prodrug. Reagents and
conditions: a) K2CO3, DMF; b) 1. Pd/C, H2, MeOH, 2. Na2CO3, isopropylamine or Tris.
benzyl ester 1 as a major side product. Next, benzyl deprotec-
tion of 18 using hydrogen and Pd/C followed by treatment
with sodium carbonate yielded the desired POMC prodrug
19a as a disodium salt. Unfortunately, NMR analysis showed
this material to contain a significant amount of impurities and
several attempts to synthesize this product cleanly failed. How-
ever, we were encouraged by a report from Farquhar and co-
workers, who isolated a similar compound that they were
using as a chemical intermediate as a dicyclohexylammonium
salt.[31] When we replaced sodium carbonate with two equiva-
lents of tris(hydroxymethyl)aminomethane (Tris), we able to
cleanly isolate the desired product as a diamine salt (19b). The
di-Tris salt of 19 was highly water-soluble, rapidly dissolving at
all concentrations tested (up to 20 mgmLꢀ1), and stable for ex-
tended periods of time when stored in a freezer. However, it
showed moderate chemical instability in solution at room tem-
perature (only 36% remained after one week in D2O solution,
Table 1). A diisopropylamine salt (19c) showed similar chemical
Figure 1. Alkaline phosphatase catalyzed activation of prodrugs 3 and 19d.
Each data point represents the meanꢁSEM of three separate experiments.
Evaluation of anti-apoptotic activity of prodrugs
Table 1. Chemical stability of different salt forms of 19.
We have previously shown that UDCA significantly inhibits cas-
pase-3 activation and apoptosis induced by transforming
growth factor b1 (TGFb1) in primary rat hepatocytes.[32] To
compare the in vitro cytoprotective effects of our new pro-
drugs with UDCA, we evaluated one of our alcohol-linked
phosphate prodrugs (the 3-substituted prodrug 3) and our
POMC prodrug 19d for their ability to modulate TGFb1-in-
duced cytotoxicity. Our preliminary results demonstrated that
the presence of alkaline phosphatase (AP) in the culture
medium does not significantly change TGFb1, UDCA or com-
pound 3 cytotoxic and cytoprotective properties. However,
compound 19d was cytoprotective only in the presence of AP
(data not shown). Therefore, AP was added to the culture
medium whenever assessing the cytoprotective potential of
19d. We first analyzed the ability of 3 and 19d to inhibit
TGFb1-induced general cell death, as compared with UDCA.
The results showed that UDCA, 3 and alkaline phosphatase-ac-
tivated 19d inhibited TGFb1-induced loss of cellular viability
by at least 50% (p<0.05) (Figure 2, upper panel). In addition,
TGFb1 also induced a 30% increase in the amount of LDH re-
Compd
Counter-ion
Tris
Isopropylamine
Tris
Tris
Equiv
[%] Remaining[a]
19b
19c
19d
19e
2
2
1
3
36
34
88
33
[a] Percent of prodrug remaining after seven days at RT in D2O solution;
n=3, SD<2%.
stability (34% remained after one week at room temperature
in D2O). However, we noticed that formulations of 19 contain-
ing less than two equivalents of amine proved to be signifi-
cantly more chemically stable in solution. This led us to synthe-
size the mono-Tris salt of our POMC prodrug, 19d, which was
highly water-soluble (>20 mgmLꢀ1), and decomposed relative-
ly slowly in solution (88% remained after one week at room
temperature in D2O). The increased aqueous stability of the
monoanionic prodrug relative to the dianionic prodrug is simi-
lar to that seen with Stella’s phosphoryloxymethyloxy carbonyl
prodrugs and is consistent with his hypothesis that hydrolysis
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2013, 8, 1002 – 1011 1005

CHEMMEDCHEM
FULL PAPERS
www.chemmedchem.org
Figure 2. Compounds 3 and 19d inhibit TGFb-induced cytotoxicity in pri-
mary rat hepatocytes. Primary rat hepatocytes were incubated with 100 mm
UDCA, compound 3, compound 19d, or no addition (control), in the pres-
ence or absence of alkaline phosphatase (3 UmLꢀ1) for 12 h. Cells were then
exposed to 1 nm TGFb for 24 h before processing for cell viability assays. Vi-
ability was assessed by live-cell protease cleavage of a fluorogenic peptide
substrate (top). Culture medium was used for LDH viability assays (bottom).
Results are expressed as meansꢁSEM of at least three different experi-
ments; *p<0.01 from respective control; §p<0.05 from respective TGFb.
Figure 3. Compounds 3 and 19d inhibit TGFb-induced nuclear fragmenta-
tion in primary rat hepatocytes. Primary rat hepatocytes were incubated
with 100 mm UDCA, compound 3, compound 19d, or no addition (control),
in the presence or absence of alkaline phosphatase (3 UmLꢀ1) for 12 h. Cells
were then exposed to 1 nm TGFb for 24 h before processing for apoptosis
assays. Fluorescence microscopy of Hoechst staining in a) control hepato-
cytes and in cells exposed to either b) compound 3, c) TGFb, d) TGFb+com-
pound 3, e) AP, f) compound 19d+AP, g) TGFb+AP, or h) compound
19d+TGFb+AP (top). Normal nuclei showed noncondensed chromatin dis-
persed over the entire nucleus. Apoptotic nuclei were identified by con-
densed chromatin, contiguous to the nuclear membrane, as well as nuclear
fragmentation of condensed chromatin. Percentage of apoptotic of cells
(bottom). Results are expressed as meansꢁSEM of at least three different
experiments; *p<0.01 from respective control; §p<0.05 from respective
TGFb.
leased from cells, an indicator of cellular toxicity (p<0.01)
(Figure 2, lower panel). Similarly as before, UDCA, 3 and 19d,
inhibited TGFb1-induced cytotoxicity by ~40, 30 and 60%, re-
spectively (p<0.05).
We then specifically evaluated the effects of 3 and 19d on
modulation of TGFb1-induced apoptosis by changes in nuclear
morphology (Figure 3) and by caspase activity (Figure 4). We
confirmed that significant levels of apoptosis occurred in cul-
tured primary rat hepatocytes after incubation with TGFb1 (p<
0.01), with a concomitant increase in caspase-3-like activity
(p<0.01). Notably, UDCA, 3 and 19d protected against TGFb1-
induced nuclear fragmentation by 50–80% (p<0.05) and cas-
pase-3-like activation by 40–70% (p<0.05). Altogether, these
results show that much like UDCA, the newly synthesized
UDCA prodrugs 3 and 19d display significant cytoprotective
properties in vitro.
Figure 4. TGFb-induced caspase-3 activity is inhibited by compound 3 and
19d. Primary rat hepatocytes were incubated with 100 mm UDCA, com-
pound 3, compound 19d, or no addition (control), in the presence or ab-
sence of alkaline phosphatase (3 UmLꢀ1) for 12 h. Cells were then exposed
to 1 nm TGFb for 24 h before processing for apoptosis assays. Caspase-3-like
activity was analyzed by enzymatic cleavage of the proluminescent caspase-
3/7 DEVD-aminoluciferin substrate. Results are expressed as meansꢁSEM of
at least three different experiments; *p<0.01 from respective control;
§p<0.05 from respective TGFb.
Conclusions
We have prepared five highly water-soluble prodrugs of the
anti-apoptotic bile acid UDCA from three distinct classes: di-
rectly linked phosphate esters, oxymethylphosphate (OMP)
prodrugs and
a novel phosphoryloxymethyl carboxylate
(POMC) prodrug. As the OMP prodrugs of UDCA were both dif-
ficult to synthesize and chemically unstable, they were not
tested in any biological assays. Compound 3, a directly linked
phosphate ester, proved to have similar anti-apoptotic potency
to UDCA in our in vitro assays, even without prior bioactivation
by alkaline phosphatase. Our POMC prodrug compound 19, in
contrast, was also highly active in these assays, but required
activation by exogenous alkaline phosphatase to have an
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2013, 8, 1002 – 1011 1006

CHEMMEDCHEM
FULL PAPERS
www.chemmedchem.org
to 0.950 mL of a 0.100m tris glycine buffer containing 1.0 mm
ZnCl2 and 1.0 mm MgCl2; 10.0 mL of this diluted AP solution was
added to the prodrug solution by syringe. A series of 42 inverse-
gated decoupled 31P NMRs were taken (24 scans each, ~1 min ac-
quisition time). Conversion (%) was determined from the relative
integration of the starting material and product peaks, NMR time
stamps were used to determine time. Each experiment was repeat-
ed three times.
effect. In our future studies, we intend to evaluate the ability
of prodrugs 3 and 19 to decrease apoptosis in an in vivo
animal model.
The novel POMC prodrug 19 was bioactivated by alkaline
phosphatase to UDCA faster than prodrug 3, in which the
phosphate ester is directly linked to an alcohol. We were
unable to isolate 19 cleanly as a sodium salt, but pure mono
and diamine salts of 19 could be readily obtained on a large
scale (>5 g) in just two steps from the parent carboxylic acid
(UDCA). Diamine salts of 19 were somewhat unstable in solu-
tion over long periods of time at room temperature, but the
mono-Tris salt of 19 decomposed at a much lower rate and
was stable for extended periods when stored cold. While the
moderate chemical instability of the POMC prodrugs will defi-
nitely limit their use, they might still find potential application
as prodrugs of carboxylic acids for situations in which cold
storage is readily available and high aqueous solubility and
rapid release are crucial, such as in acute medical conditions in
a hospital environment.
Animal experiments: All experiments involving animals were per-
formed by an Investigator accredited for directing animal experi-
ments (FELASA level C), in conformity with the Public Health Ser-
vice (PHS) Policy on Humane Care and Use of Laboratory Animals,
incorporated in the Institute for Laboratory Animal Research (ILAR)
Guide for Care and Use of Laboratory Animals. Experiments re-
ceived prior approval from the Portuguese National Authority for
Animal Health (DGAV).
Cell culture and treatments: Primary rat hepatocytes were isolated
from male rats (100–150 g) by collagenase perfusion.[33] Briefly, rats
were anesthetized with phenobarbital sodium (100 mgkgꢀ1 body
weight) injected into the peritoneal cavity. After administration of
heparin (200 Ukgꢀ1 body weight) in the tail vein, the animals’ ab-
domen was opened and the portal vein exposed and cannulated.
The liver was then perfused at 378C in situ with a calcium-free
Hank’s Balanced Salt Solution (HBSS) for ~10 min, and then with
0.05% collagenase type IV in calcium-present HBSS for another
10 min. Hepatocyte suspensions were obtained by passing collage-
nase-digested livers through 125 mm gauze and washing cells in
Complete William’s E medium (Sigma–Aldrich) supplemented with
26 mm sodium bicarbonate, 23 mm HEPES, 0.01 UmLꢀ1 insulin,
2 mm l-glutamine, 10 nm dexamethasone, 100 UmLꢀ1 penicillin,
and 10% heat-inactivated fetal bovine serum (Invitrogen). Viable
primary rat hepatocytes were enriched by low-speed centrifugation
at 200 g for 3 min. Cell viability was determined by trypan blue ex-
clusion and was typically 80–85%. After isolation, hepatocytes
were resuspended in Complete William’s E medium and plated on
Experimental Section
Chemical stability
In D2O: Prodrug 19b, 19c, or 19d (4.0 mg) was dissolved in D2O
(1.0 mL) and placed in an NMR tube with a sealed capillary tube
containing a solution of phenylphosphonic acid in D2O as a stan-
1
dard. At t=0, a H NMR spectrum was obtained, and the proton
signal at d=5.51 ppm was integrated (It=0) relative to the aromatic
signals of phenylphosphonic acid. The tube was then stored at RT
for 7 days at RT. Another 1H NMR spectrum was taken, and the
proton signal at d=5.51 ppm was again integrated (It=7) as before.
The proportion (%) of prodrug remaining was determined as It=7
/
It=0 ꢂ100. Each experiment was repeated three times. Chemical sta-
Primaria tissue culture dishes (BD Biosciences) at 5ꢂ104 cellscmꢀ2
.
bility results obtained using 31P NMR were similar to those ob-
1
Cells were maintained at 378C in a humidified atmosphere of 5%
CO2 for 6 h, to allow attachment. Plates were then washed with
medium to remove dead cells and incubated in Complete William’s
E medium supplemented with either 100 mm UDCA, compound 3,
compound 19d or no addition (control), in the presence or ab-
sence of 3 UmLꢀ1 of alkaline phosphatase (Invitrogen) for 12 h.
Cells were then exposed to 1 nmolLꢀ1 recombinant human TGF-b1
(R&D Systems) for 24 h before processing for cell viability and
apoptosis assays.
tained using H NMR. For 19e, 4.0 mg 19b was dissolved in D2O
and an additional equivalent of Tris added (the stoichiometry was
1
confirmed by H NMR) and the experiment conducted as above.
In pH 7.4 tris-buffered saline: Prodrug 19d (4.0 mg) was dissolved
in H2O (0.9 mL) and treated with tris-buffered saline (0.1 mL; BM-
300 from Boston BioProducts), containing tris (250 mm), KCl
(27 mm), and NaCl (1.37m). A sealed capillary tube containing phe-
nylphosphonic acid dissolved in D2O was used in NMR experiments
as a standard. Chemical stability results were obtained by measur-
ing the disappearance of prodrug relative to the internal standard
by 31P NMR. The experiment was repeated three times. The stan-
dard deviation (SD) was ꢁ3%.
Cell viability assays: LDH, a stable cytosolic enzyme, is released to
cell culture media following cell lysis, and can be used as a marker
of cytotoxicity. Briefly, to assess LDH release, supernatants taken
from a gentle centrifugation of cell culture media at 250 g, were
combined in microplates with lactate (substrate), tetrazolium salt
(coloring solution), and NAD (co-factor), previously mixed in equal
proportions, following the manufacturer’s instructions (Sigma–Al-
drich). Multi-well plates were protected from light and incubated
for 10 min at room temperature. Finally, absorbance was measured
at 490 nm, with 690 nm as reference, using a Bio-Rad model 680-
microplate reader (Bio-Rad Laboratories, Hercules, CA, USA).
Biological evaluation
Activation of prodrugs 3 and 19d: Alkaline phosphatase from
bovine intestinal mucosa (Sigma–Aldrich, P5521-2KU) was dis-
solved in 2.0 mL of a 0.100m sodium glycine buffer containing
1.0 mm ZnCl2 and 1.0 mm MgCl2. This stock solution was stored at
48C between uses. Compound 19d (10.0 mg, 0.016 mmol) or com-
pound
3
(8.7 mg, 0.016 mmol) were dissolved in 0.6 mL of
To assess cellular viability, the CellTiter-Fluor viability assay was
used (Promega, Madison, WI, USA). Briefly, viable cells are mea-
sured using a fluorogenic, cell-permeant, peptide substrate (Gly-
Phe-AFC), which is cleaved by the live-cell protease activity to gen-
erate a fluorescent signal proportional to the number of living
a 0.100m tris glycine buffer solution containing 1.0 mm ZnCl2 and
1.0 mm MgCl2. Neither compound showed decomposition by
31P NMR when left in this buffer solution for 1 h; 50 mL of the previ-
ously prepared AP stock solution was further diluted by addition
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2013, 8, 1002 – 1011 1007

CHEMMEDCHEM
FULL PAPERS
www.chemmedchem.org
cells. Cells were incubated with an equal volume of CellTiter-Fluor
Reagent for 30 min at 378C and resulting fluorescence (380–
400 nmEx/505 nmEm) measured using a GloMax+ Multi Detection
System (Promega).
vacuo. Purification by flash chromatography (30!100% EtOAc/
hexanes) on silica gel furnished the desired product as a white
solid (4.72 g, 95%): 1H NMR (400 MHz, CD3OD): d=7.39–7.28 (m,
5H), 5.13 and 5.10 (ABq, JAB =12.3 Hz, 2H), 3.56–3.42 (m, 2H), 2.46–
2.36 (m, 1H), 2.36–2.25 (m, 1H), 2.08–1.97 (m, 1H), 1.94–1.76 (m,
5H), 1.67–0.98 (m, 18H), 0.97 (s, 3H), 0.94 (d, J=6.4 Hz, 3H),
0.67 ppm (s, 3H); 13C NMR (100 MHz, CD3OD): d=12.7, 18.9, 22.4,
23.9, 27.9, 29.6, 31.1, 32.2, 32.3, 35.2, 36.1, 36.6, 38.0, 38.6, 40.7,
41.5, 44.0, 44.5, 44.8, 56.5, 57.5, 67.2, 71.9, 72.1, 129.2, 129.3, 129.6,
137.7, 175.7 ppm.
Apoptosis assays: General caspase-3/7 activity was evaluated using
the Caspase-Glo 3/7 Assay (Promega). Briefly, the assay provides
a proluminescent caspase-3/7 DEVD-aminoluciferin substrate and
a proprietary thermostable luciferase in a reagent optimized for
caspase-3/7 activity, luciferase activity and cell lysis. Cells were in-
cubated with an equal volume of Caspase-Glo 3/7Reagent for
30 min at 378C and resulting luminescence measured using
a GloMax+ Multi Detection System (Promega).
3-(Bis(benzyloxy)phosphoryloxy)ursodeoxycholic acid benzyl
ester (2): A stirred suspension of 1 (1.497 g, 3.10 mmol), 1,2,4-tria-
zole (450 mg, 6.52 mmol), and NaHCO3 (1.906 g, 22.69 mmol) in
1,2-dichloroethane (30 mL) was treated with dibenzyl N,N-diethyl-
phosphoramidite (1.00 mL, 3.15 mmol). The reaction mixture was
heated overnight to 658C. After cooling in an ice bath, THF
(12 mL) was added to the mixture, followed by dropwise addition
of 30% H2O2 (6 mL). After stirring for 5 min., saturated aq Na2S2O3
(30 mL) was added slowly (CAUTION: highly exothermic reaction).
The mixture was diluted with water (200 mL) and extracted with
CH2Cl2 (2ꢂ200 mL). The combined organic layers were dried
(MgSO4), filtered, and concentrated in vacuo. Purification by flash
chromatography (30!100% EtOAc/hexanes) on silica gel followed
by a second flash chromatography (0!10% MeOH/CH2Cl2) on
silica gel furnished the desired product as a clear colorless oil
In addition, Hoechst labeling of cells was used to detect apoptotic
nuclei by morphological analysis. Briefly, culture medium was
gently removed to prevent detachment of cells. Attached primary
rat hepatocytes were fixed with 4% paraformaldehyde in phos-
phate-buffered saline (PBS), pH 7.4, for 10 min at RT, washed with
PBS, incubated with Hoechst dye 33258 (Sigma–Aldrich) at
5 mgmLꢀ1 in PBS for 5 min, washed with PBS, and mounted using
Fluoromount-G (SouthernBiotech). Fluorescence was visualized
using an Axioskop fluorescence microscope (Carl Zeiss GmbH).
Blue-fluorescent nuclei were scored blindly and categorized ac-
cording to the condensation and staining characteristics of chro-
matin. Normal nuclei showed non-condensed chromatin disperse
over the entire nucleus. Apoptotic nuclei were identified by con-
densed chromatin, contiguous to the nuclear membrane, as well
as by nuclear fragmentation of condensed chromatin. Five random
microscopic fields per sample containing ~150 nuclei were count-
ed, and mean values expressed as the percentage of apoptotic
nuclei.
1
(1.1158 g, 48%): H NMR (400 MHz, CDCl3): d=7.43–7.27 (m, 15H),
5.12 and 5.09 (ABq, JAB =12.4 Hz, 2H), 5.07–4.96 (m, 4H), 4.29–4.15
(m, 1H), 3.58–3.44 (m, 1H), 2.46–2.33 (m, 1H), 2.33–2.21 (m, 1H),
2.01–1.93 (m, 1H), 1.93–0.93 (m, 23H), 0.91 (s, 3H), 0.91 (d, J=
6.1 Hz, 3H), 0.64 ppm (s, 3H); 13C NMR (100 MHz, CDCl3): d=12.1,
18.4, 21.2, 23.2, 26.8, 28.2, 28.6, 31.0, 31.3, 33.9, 34.5, 34.8, 35.2,
36.5, 39.1, 40.1, 42.3, 43.66, 43.73, 54.9, 55.7, 66.1, 69.04, 69.07,
69.10, 69.12, 71.1, 78.6 (d, J31P =6.0 Hz), 127.88, 127.90, 128.2,
128.3, 128.45, 128.54, 135.98, 136.05, 136.10, 174.0 ppm; HRMS: m/
z [M+H]+ calcd for C45H59O7P: 743.4077, found: 743.4092.
Statistical analysis: Statistical analysis was performed using Graph-
Pad InStat version 3.00 (GraphPad Software, San Diego, CA, USA)
for the analysis of variance and Bonferroni’s multiple comparison
tests. Values of p<0.05 were considered significant.
3-(Phosphonatooxy)ursodeoxycholic acid sodium salt (3): A solu-
tion of 2 (2.22 g, 2.99 mmol) in MeOH (100 mL) was treated with
10% Pd/C (291 mg). The reaction mixture was stirred at RT under
an atmosphere of H2 maintained using a gas-filled balloon for 2 h
and filtered through Celite. Na2CO3 (476 mg, 4.49 mmol) dissolved
in water (25 mL) was added to the filtrate, and the solution was
concentrated in vacuo until most of the MeOH was removed. The
remaining solution was lyophilized to afford the desired product as
a white solid (1.627 g, quant.): 1H NMR (400 MHz, D2O): d=4.04–
3.91 (m, 1H), 3.72–3.62 (m, 1H), 2.29–2.17 (m, 1H), 2.17–2.06 (m,
1H), 2.07–1.97 (m, 1H), 1.95–1.00 (m, 23H), 0.96 (s, 3H), 0.95 (d, J=
6.0 Hz, 3H), 0.70 ppm (s, 3H); 13C NMR (100 MHz, D2O): d=12.1,
18.6, 21.5, 23.3, 27.2, 28.8, 29.0, 33.1, 34.1, 35.2, 35.3, 35.7, 35.8,
Chemistry
General: Microwave experiments were performed on a Biotage Ini-
tiator Microwave Synthesizer. 1H NMR and 13C NMR spectra were re-
corded on a Bruker 400 spectrometer. H NMR data are reported as
1
follows: chemical shift in parts per million (ppm) downfield of the
internal reference tetramethylsilane (TMS), multiplicity (s=singlet,
bs=broad singlet, d=doublet, t=triplet, q=quartet, quint=quin-
tet and m=multiplet), coupling constant (J) in Hertz (Hz), and inte-
grated value. Coupling constant defined as J31P disappeared when
1H NMR spectra were taken with 31P decoupling. 13C NMR spectra
were measured with complete proton decoupling. 31P NMR spectra
taken for compound characterization were measured with com-
plete proton decoupling and were referenced to 85% phosphoric
acid, which was added to the NMR tube in a sealed capillary tube.
LC/MS analysis was carried out using a BEH C18 column (2.1 mmꢂ
50 mm, 5 mm) on a Waters Acquity UPLC system with a Waters ZQ
mass detector. HRMS analyses were performed on a Waters LCT-
TOF High-resolution mass spectrometer. Ursodeoxycholic acid
(UDCA) was obtained from Sigma–Aldrich. Dibenzyl chloromethyl
phosphate was synthesized by the method of Mꢃntylꢃ,[34] but is
also commercially available from Sigma–Aldrich.
37.1, 39.6, 40.4, 42.9, 43.5, 44.0, 55.1, 55.7, 71.9, 75.9 (d, J31P
=
5.0 Hz), 185.5 ppm; 31P NMR (162 MHz, D2O): 2.51 ppm; HRMS: m/z
[M+H]+ calcd for C24H41O7P: 473.2668, found: 473.2670.
3-(Benzyloxycarbonyloxy)ursodeoxycholic acid benzyl ester (4)
and 7-(benzyloxycarbonyloxy)ursodeoxycholic acid benzyl ester
(5): A stirred solution of 1 (1.465 g, 3.04 mmol) in dry CH2Cl2
(50 mL) was treated with pyridine (0.600 mL, 7.42 mmol) followed
by slow addition of benzyl chloroformate (1.00 mL, 7.03 mmol).
After stirring for 1 h, additional pyridine (0.300 mL, 3.71 mmol) and
benzyl chloroformate (0.600 mL, 4.22 mmol) were added. After
30 min., the reaction mixture was extracted with 1m aq HCl
(50 mL). The organic layer was dried (Na2SO4), filtered, and concen-
trated in vacuo. Purification by flash chromatography (10!100%
EtOAc/hexanes) on silica gel first furnished 4 as a slightly yellow
Ursodeoxycholic acid benzyl ester (1): A suspension of UDCA
(4.03 g, 10.3 mmol) and K2CO3 (4.88 g, 35.3 mmol) in CH3CN
(100 mL) was treated with BnBr (6.00 mL, 50.5 mmol). The reaction
mixture was heated to 808C for 3 h, filtered, and concentrated in
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2013, 8, 1002 – 1011 1008

CHEMMEDCHEM
FULL PAPERS
www.chemmedchem.org
foam (0.7769 g, 41%), followed by 5 as a slightly yellow foam
(181 mg, 10%), which was then followed by recovered 1 as
a white solid (603.2 mg, 41%).
5.9 Hz), 185.6 ppm; 31P NMR (162 MHz, D2O): 0.93 ppm; LC/MS
(ESIꢀ): m/z=471.4 [MꢀH]ꢀ.
3-(Benzyloxycarbonyloxy)-7-(methylthiomethoxy)ursodeoxychol-
ic acid benzyl ester (9): A solution of 4 (2.71 g, 4.39 mmol) in
DMSO (34 mL) was treated with Ac2O (21 mL) followed by AcOH
(34 mL). After stirring at RT for 24 h, the reaction mixture was dilut-
ed with water (500 mL) and neutralized with solid NaHCO3. The
mixture was extracted with EtOAc (500 mL). The organic layer was
then further washed with water (5ꢂ500 mL), dried (Na2SO4), fil-
tered, and concentrated in vacuo. Purification by flash chromatog-
raphy (5!30% EtOAc/hexanes) on silica gel furnished the desired
3-(Benzyloxycarbonyloxy)ursodeoxycholic acid benzyl ester (4):
1H NMR (400 MHz, CDCl3): d=7.39–7.30 (m, 10H), 5.14 (s, 2H), 5.12
and 5.09 (ABq, JAB =12.3 Hz, 2H), 4.56 (dddd, J=5, 5, 11, 11 Hz,
1H), 3.60–3.50 (m, 1H), 2.45–2.34 (m, 1H), 2.33–2.22 (m, 1H), 2.02–
1.94 (m, 1H), 1.94–0.98 (m, 23H), 0.95 (s, 3H), 0.91 (d, J=6.2 Hz,
3H), 0.65 ppm (s, 3H); 13C NMR (100 MHz, CDCl3): d=12.1, 18.3,
21.2, 23.3, 26.4, 26.9, 28.6, 31.0, 31.3, 33.0, 34.1, 34.5, 35.2, 36.6,
39.1, 40.1, 42.2, 43.7, 43.8, 54.9, 55.7, 66.1, 69.3, 71.2, 77.9, 128.18,
128.24, 128.3, 128.46, 128.54, 128.6, 135.4, 136.1, 154.5, 174.0 ppm.
1
product as a slightly yellow oil (1.3966 g, 47%): H NMR (400 MHz,
CDCl3): d=7.40–7.28 (m, 10H), 5.14 (s, 2H), 5.12 and 5.09 (ABq,
J
AB =12.4 Hz, 2H), 4.61–4.50 (m, 1H), 4.59 and 4.52 (ABq, JAB =
7-(Benzyloxycarbonyloxy)ursodeoxycholic acid benzyl ester (5):
1H NMR (400 MHz, CDCl3): d=7.39–7.30 (m, 10H), 5.16 and 5.12
(ABq, JAB =12.2 Hz, 2H), 5.12 and 5.10 (ABq, JAB =12.3 Hz, 2H), 4.64
(ddd, J=5, 11, 11 Hz, 1H), 3.63–3.52 (m, 1H), 2.44–2.34 (m, 1H),
2.32–2.22 (m, 1H), 2.01–1.93 (m, 1H), 1.91–0.96 (m, 23H), 0.94 (s,
3H), 0.90 (d, J=6.3 Hz, 3H), 0.62 ppm (s, 3H); 13C NMR (100 MHz,
CDCl3): d=12.0, 18.3, 21.2, 23.2, 25.6, 28.6, 30.2, 31.0, 31.3, 33.0,
33.9, 34.7, 35.2, 37.1, 39.4, 39.9, 40.0, 42.2, 43.6, 55.0, 55.2, 66.1,
69.3, 71.3, 78.5, 128.1, 128.19, 128.24, 128.4, 128.5, 128.6, 135.6,
136.1, 154.6, 174.0 ppm.
11.2 Hz, 2H), 3.33 (ddd, J=5, 11, 11 Hz, 1H), 2.45–2.35 (m, 1H),
2.32–2.22 (m, 1H), 2.17 (s, 3H), 2.00–1.93 (m, 1H), 1.92–0.97 (m,
23H), 0.95 (s, 3H), 0.90 (d, J=6.2 Hz, 3H), 0.63 ppm (s, 3H);
13C NMR (100 MHz, CDCl3): d=12.2, 15.3, 18.4, 21.3, 23.3, 26.3, 26.6,
28.5, 31.0, 31.3, 32.5, 33.0, 34.1, 34.5, 35.2, 39.4, 40.1, 41.5, 42.0,
43.8, 55.0, 55.8, 66.1, 69.4, 73.0, 77.9, 78.1, 128.16, 128.23, 128.3,
128.46, 128.54, 128.6 135.4, 136.2,154.6, 174.1 ppm.
3-(Benzyloxycarbonyloxy)-7-(chloromethoxy)ursodeoxycholic
acid benzyl ester (11): A solution of 9 (847 mg, 1.25 mmol) in dry
CH2Cl2 (20 mL) was treated with 2m SOCl2 in CH2Cl2 (1.9 mL,
3.8 mmol). The reaction mixture was heated in a microwave to
1008C for 30 min and then concentrated in vacuo. The 1H NMR
spectrum of the crude material in CDCl3 showed a new set of dou-
blets d=5.56 and 5.47 ppm (J=5.4 Hz, 1H each)[35] and the disap-
pearance of the AB pattern at d=4.59 and 4.52 ppm as well as the
3-(Benzyloxycarbonyloxy)-7-(bis(benzyloxy)phosphoryloxy)urso-
deoxycholic acid benzyl ester (6): A stirred suspension of 4
(374 mg, 0.61 mmol), 1,2,4-triazole (89.8 mg, 1.30 mmol), and
NaHCO3 (263 mg, 3.13 mmol) in CH2Cl2 was treated with dibenzyl
N,N-diethylphosphoramidite (0.900 mL, 3.00 mmol). The reaction
mixture was heated overnight to 408C. After cooling in an ice
bath, THF (5 mL) was added to the mixture, followed by dropwise
addition of 30% aq H2O2 (3 mL). After stirring for 5 min, saturated
aq Na2S2O3 (20 mL) was added slowly (CAUTION: highly exothermic
reaction). The mixture was diluted with CH2Cl2 and extracted with
water (100 mL). The organic layer was dried (Na2SO4), filtered, and
concentrated in vacuo. Purification by flash chromatography (30%
EtOAc/hexanes) on silica gel furnished the desired product as
1
SMe peak present in the H NMR spectrum of 9 at d=2.17 ppm.
The crude material was used in the next reaction without further
purification.
7-(Phosphonooxymethoxy)ursodeoxycholic acid tributylamine
salt (13a): A suspension of H3PO4 (586 mg, 5.98 mmol) and 4 ꢄ
molecular sieves (2.023 g) in CH3CN (40 mL) was treated with Bu3N
(5.4 mL, 22.7 mmol). The mixture was stirred overnight and then
added to a flask containing crude 11. After stirring for 24 h, the
mixture was filtered through Celite and concentrated in vacuo. The
residue was dissolved in MeOH (50 mL) and concentrated in vacuo
again. Next, the residue was dissolved in MeOH (50 mL), 10% Pd/C
(2.369 g) was added, and the reaction mixture stirred at RT under
an atmosphere of H2 maintained using a gas-filled balloon for 2 h
and then filtered through Celite. Additional 10% Pd/C (2.14 g) was
added to the filtrate, and the reaction mixture was again stirred
under H2 for an additional 72 h. The reaction mixture was filtered
through Celite and concentrated in vacuo. The resulting residue
was purified by liquid chromatography (5% CH3CN/water to 100%
CH3CN, C18 column) to yield, after lyophilization, the desired prod-
uct as a white solid (116.7 mg). There are ~1.4 equiv of NBu3 pres-
ent for every equiv of bile acid based on H NMR analysis (c.f. inte-
gration of CH3 peak at d=0.70 ppm and multiplet at d=3.12–
3.02 ppm). H NMR (400 MHz, CD3OD): d=5.18 (dd, J=6 Hz, J31P
6 Hz, 1H), 4.99 (dd, J=6 Hz, J31P =8 Hz, 1H), 3.66–3.55 (m, 1H),
3.53–3.42 (m, 1H), 3.13–3.02 (m, 8.2H), 2.35–2.24 (m, 1H), 2.20–
2.10 (m, 1H), 2.08–1.98 (m, 1H), 1.94–0.90 (m, 57H), 0.70 ppm (s,
3H); LC/MS (ESIꢀ): m/z=501.3 [MꢀH]ꢀ.
1
a slightly yellow oil (350.1 mg, 66%): H NMR (400 MHz, CDCl3): d=
7.43–7.27 (m, 20H), 5.16 and 5.15 (ABq, JAB =12.4 Hz, 2H), 5.12 and
5.10 (ABq, JAB =12.4 Hz, 2H), 5.05–4.91 (m, 4H), 4.51 (dddd, J=5, 5,
10, 10 Hz, 1H), 4.30–4.17 (m, 1H), 2.45–2.34 (m, 1H), 2.32–2.21 (m,
1H), 1.99–0.99 (m, 24H), 0.93 (s, 3H), 0.90 (d, J=6.1 Hz, 3H),
0.61 ppm (s, 3H); 13C NMR (100 MHz, CDCl3): d=12.1, 18.4, 21.2,
23.2, 26.2, 28.4, 31.0, 31.3, 32.7, 33.8, 34.3, 34.4, 35.2, 39.2, 39.8,
41.8, 41.9, 42.0, 43.7, 54.9, 55.0, 66.1, 68.86, 68.92, 69.0, 69.1, 69.4,
77.5, 79.7, 79.8, 127.86, 127.90, 128.17, 128.22, 128.3, 128.4, 128.47,
128.54, 128.6, 135.4, 136.09, 136.11, 136.15, 136.18, 154.5,
174.0 ppm; LC/MS (ESI+): m/z=877.7 [M+H]+.
7-(Phosphonatooxy)ursodeoxycholic acid sodium salt (7): A sus-
pension of 6 (1.1984 g, 1.37 mmol) in MeOH (200 mL) was treated
with 10% Pd/C (322 mg). The reaction mixture was stirred at RT
under an atmosphere of H2 maintained using a gas-filled balloon
for 2 h and then filtered through Celite. Na2CO3 (216.2 mg,
2.04 mmol) dissolved in water (25 mL) was added, and the solution
was concentrated in vacuo until most of the MeOH was removed.
The remaining solution was lyophilized to afford the desired prod-
1
1
=
1
uct as a white solid (762.7 mg, quant.): H NMR (400 MHz, D2O): d=
4.13–3.99 (m, 1H), 3.69–3.55 (m, 1H), 2.30–2.17 (m, 1H), 2.17–2.07
(m, 1H), 2.07–1.92 (m, 3H), 1.92–0.99 (m, 21H), 0.97 (s, 3H), 0.95 (d,
J=6.5 Hz, 3H), 0.69 ppm (s, 3H).13C NMR (100 MHz, D2O): d=12.1,
18.6, 21.5, 23.4, 27.0, 28.8, 29.6, 33.1, 34.1, 35.0, 35.2, 35.4, 35.9,
7-(Phosphonooxymethoxy)ursodeoxycholic acid sodium salt
(13b): A 1 cm wide column was filled with 12 cm of DOWEX 50W2
(50–100 mesh, strongly acidic) ion exchange resin.[28] The column
was prepared by sequentially washing with MeOH/water (1:1), 1m
aq NaHCO3 (gas evolution), water, and then finally MeOH/water
36.4, 39.5, 40.2, 42.6, 42.7, 44.0, 55.1, 55.3, 72.0, 76.4 (d, J31P
=
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2013, 8, 1002 – 1011 1009

CHEMMEDCHEM
FULL PAPERS
www.chemmedchem.org
(1:1). Compound 13a (115 mg) was dissolved in MeOH/water (1:1)
and loaded onto the column, which was eluted with MeOH/water
(1:1). The product-containing fractions were lyophilized to furnish
the desired product as a white solid (76.4 mg): H NMR (400 MHz,
D2O): d=5.18 (dd, J=5.7 Hz, J31P =6.8 Hz, 1H), 4.99 (dd, J=5.7 Hz,
Ursodeoxycholic acid (bis(benzyloxy)phosphoryloxy)methyl
ester (18): A suspension of UDCA (1.46 g, 3.72 mmol) and K2CO3
(984 mg, 7.12 mmol) in DMF (10 mL) was treated with dibenzyl
chloromethyl phosphate (1.23 g, 3.76 mmol). The mixture was
stirred at RT overnight, diluted with water (250 mL), and extracted
with EtOAc (3ꢂ250 mL) and CH2Cl2 (1ꢂ250 mL). The combined or-
ganic layers were dried (MgSO4), filtered, and concentrated in va-
cuo. Purification by flash chromatography (40!100% EtOAc/hex-
anes) on silica gel furnished the desired product as a clear, color-
1
J
31P =9.4 Hz, 1H), 3.74–3.56 (m, 2H), 2.40–2.28 (m, 1H), 2.27–2.14
(m, 1H), 2.08 (m, 24H), 1.00–0.92 (m, 6H), 0.70 ppm (s, 3H).
3-(Methylthiomethoxy)-7-(benzyloxycarbonyloxy)ursodeoxychol-
ic acid benzyl ester (14): A solution of 5 (1.113 g, 1.80 mmol) in
DMSO (17 mL) was treated with Ac2O (10.5 mL) followed by AcOH
(17 mL). After stirring at RT for 24 h, the reaction mixture was dilut-
ed with water (500 mL) and neutralized with solid NaHCO3. The
mixture was extracted with EtOAc (500 mL). The organic layer was
then further extracted with water (5ꢂ500 mL), dried (Na2SO4), fil-
tered, and concentrated in vacuo. Purification by flash chromatog-
raphy (5!50% EtOAc/hexanes) on silica gel furnished the desired
1
less foamy oil (2.05 g, 81%): H NMR (400 MHz, CD3OD): d=7.42–
7.38 (m, 10H), 5.64 (d, J31P =13.8 Hz, 2H), 5.10 (d, J31P =8.3 Hz, 4H),
3.58–3.44 (m, 2H), 2.41–2.31 (m, 1H), 2.29–2.18 (m, 1H), 2.08–1.98
(m, 1H), 1.96–1.72 (m, 5H), 1.70–1.00 (m, 18H), 0.99 (s, 3H), 0.93 (d,
J=6.5 Hz, 3H), 0.71 ppm (s, 3H); 13C NMR (100 MHz, CD3OD): d=
12.7, 18.9, 22.4, 23.9, 27.9, 29.6, 31.1, 31.6, 31.8, 35.2, 36.1, 36.5,
38.0, 38.6, 40.7, 41.5, 44.0, 44.5, 44.8, 56.4, 57.5, 71.1 (d, J31P
=
5.9 Hz), 71.9, 72.1, 83.9 (d, J31P =5.7 Hz), 129.2, 129.7, 129.8, 136.9,
137.0, 173.8 ppm. 31P NMR (162 MHz, CD3OD): d=ꢀ1.59 ppm;
HRMS: m/z [M+H]+ calcd for C39H55O8P: 683.3713, found:
683.3735.
1
product as a slightly yellow oil (364 mg, 30%): H NMR (400 MHz,
CDCl3): d=7.41–7.28 (m, 10H), 5.16 and 5.12 (ABq, JAB =12.0 Hz,
2H), 5.12 and 5.10 (ABq, JAB =12.4 Hz, 2H), 4.67–4.58 (m, 1H), 4.65
(s, 2H), 3.57 (dddd, J=5, 5, 10, 10 Hz, 1H), 2.44–2.34 (m, 1H), 2.31–
2.22 (m, 1H), 2.15 (s, 3H), 2.00–1.93 (m, 1H), 1.92–0.95 (m, 23H),
0.94 (s, 3H), 0.89 (d, J=6.3 Hz, 3H), 0.62 ppm (s, 3H); 13C NMR
(100 MHz, CDCl3): d=12.2, 13.7, 18.3, 21.2, 23.2, 25.6, 26.9, 28.4,
31.0, 31.3, 33.1, 33.4, 34.2, 34.7, 35.2, 39.2, 39.9, 40.0, 42.2, 43.6,
55.0, 55.2, 66.1, 69.3, 72.0, 75.2, 78.5, 128.10, 128.18, 128.24, 128.39,
128.64, 135.6, 136.1, 154.6, 174.0 ppm.
General procedure for the synthesis ursodeoxycholic acid phos-
phonooxymethoxy ester salts (19b–e): A solution of 18 in MeOH
was treated with 10% Pd/C. The reaction mixture was stirred at RT
under an atmosphere of H2 maintained using a gas-filled balloon
for 45 min and then filtered through Celite. The appropriate amine
(1–2 equiv) was added to the filtrate, and the solution concentrat-
ed in vacuo.
3-(Chloromethoxy)-7-(benzyloxycarbonyloxy)ursodeoxycholic
acid benzyl ester (15): A solution of 14 (360 mg, 0.53 mmol) in dry
CH2Cl2 (20 mL) was treated with 2m SOCl2 in CH2Cl2 (0.8 mL,
1.6 mmol). The reaction mixture was heated in a microwave to
1008C for 30 min and then concentrated in vacuo. The 1H NMR
spectrum of the crude material in CDCl3 showed a new AB pattern
at d=5.55 and 5.54 ppm (J=5.4 Hz, 2H total) and the disappear-
ance of the singlet at d=4.65 ppm as well as the SMe peak pres-
Mono-Tris salt 19d: White solid, 99% yield: 1H NMR (400 MHz,
D2O): d=5.51 (d, J31P =12.8 Hz, 2H), 3.74 (s, 6H), 3.60–3.54 (m, 2H),
2.58–2.46 (m, 1H), 2.44–2.32 (m, 1H), 2.09–1.98 (m, 1H), 1.96–1.74
(m, 5H), 1.72–1.02 (m, 18H), 1.01–0.95 (m, 6H), 0.72 ppm (s, 3H);
13C NMR (100 MHz, D2O): d=12.6, 19.0, 22.0, 23.9, 27.2, 29.0, 30.2,
31.1, 31.4, 34.4, 35.4, 35.7, 36.6, 37.3, 39.9, 40.8, 42.8, 43.7, 44.1,
55.3, 56.1, 60.0, 62.1, 71.6, 71.7, 83.6, 176.4 ppm; 31P NMR
(162 MHz, D2O): d=ꢀ0.30 ppm; HRMS: m/z [MꢀH]ꢀ calcd for
C25H43O8P: 501.2617, found: 501.2585.
1
ent in the H NMR spectrum of 14 at d=2.15 ppm. The crude ma-
terial was used without further purification in the next reaction.
3-(Phosphonooxymethoxy)ursodeoxycholic acid tributylamine
salt (16): A suspension of H3PO4 (248 mg, 2.53 mmol) and 4 ꢄ mo-
lecular sieves (0.760 g) in CH3CN (15 mL) was treated with Bu3N
(2.3 mL, 9.68 mmol). The mixture was stirred overnight and then
added to a flask containing crude 15. After stirring for 72 h, the
mixture was filtered and concentrated in vacuo. The residue was
dissolved in MeOH (25 mL) and again concentrated in vacuo. Next,
the residue was dissolved in MeOH (40 mL), 10% Pd/C (656 mg)
was added, and the reaction mixture was stirred under a balloon
filled with H2 for 2 h and then filtered through Celite. A crude NMR
of an aliquot showed no reaction. Additional 10% Pd/C (744 mg)
was added, and the reaction mixture was stirred under a balloon
filled with H2 for 2 h and filtered through Celite. A crude NMR of
an aliquot again showed no reaction. An additional 10% Pd/C
(1901 mg) was added, and the reaction mixture was stirred under
a balloon filled with H2 overnight, filtered through Celite, and con-
centrated under reduced pressure. The resulting residue was puri-
fied by liquid chromatography (5% CH3CN/H2O to 100% CH3CN,
C18 column) to yield 80.7 mg of a white solid after lyophilization.
There are ~1.7 equiv of NBu3 present for every equiv of bile acid
based on 1H NMR analysis (c.f. integration of CH3 peak at d=
0.71 ppm and multiplet at d=3.12–3.02 ppm). 1H NMR (400 MHz,
CD3OD): d=5.08 (d, J31P =8.4 Hz, 2H), 3.75–3.63 (m, 1H), 3.54–3.43
(m, 1H), 3.12–2.98 (m, 10H), 2.34–2.23 (m, 1H), 2.19–2.09 (m, 1H),
2.08–20 (m, 1H), 1.94–0.90 (m, 68H), 0.71 ppm (s, 3H); LC/MS
(ESIꢀ): m/z=501.3 [MꢀH]ꢀ.
Acknowledgements
This research was supported by a Faculty Research Development
Grant from the University of Minnesota Academic Health Center
and by the Fundażo para a CiÞncia e a Tecnologia, Portugal
(PTDC/SAU-ORG/119842/2010). The authors thank Drs. Gunda
Georg, Sandeep Gupta, Shunan Li, and Henry Wong for helpful
discussions.
Keywords: apoptosis
· bile acids · phosphoryloxymethyl
carboxylate · prodrugs · ursodeoxycholic acid
[1] T. Ikegami, Y. Matsuzaki, Hepatol. Res. 2008, 38, 123–131.
[2] J. D. Amaral, R. J. S. Viana, R. M. Ramalho, C. J. Steer, C. M. P. Rodrigues,
J. Lipid Res. 2009, 50, 1721–1734.
[3] C. D. Keene, C. M. P. Rodrigues, T. Eich, C. Linehan-Stieers, A. Abt, B. T.
Kren, C. J. Steer, W. C. Low, Exp. Neurol. 2001, 171, 351–360.
˘
˘
[4] A. C¸olak, B. Kelten, A. Sagmanligil, O. Akdemir, A. Karaoglan, E. S¸ahan,
ꢅ. C¸elik, S¸. Barut, J. Clin. Neurosci. 2008, 15, 665–671.
[5] H.-P. Qi, S.-Q. Wei, X.-C. Gao, N.-N. Yu, W.-Z. Hu, S. Bi, H. Cui, Mol. Vis.
2012, 18, 151–160.
[6] E. Seyhun, A. Malo, C. Schꢃfer, C. A. Moskaluk, R.-T. Hoffmann, B. Gçke,
C. H. Kubisch, Am. J. Physiol.: Gastrointest. Liver Physiol. 2011, 301,
G773–G782.
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2013, 8, 1002 – 1011 1010

CHEMMEDCHEM
FULL PAPERS
www.chemmedchem.org
[7] R. E. Castro, D. M. S. Ferreira, X. Zhang, P. M. Borralho, A. L. Sarver, Y.
Zeng, C. J. Steer, B. T. Kren, C. M. P. Rodrigues, Am. J. Physiol.: Gastroint-
est. Liver Physiol. 2010, 299, G887–G897.
[22] V. del Amo, L. Siracusa, T. Markidis, B. Baragana, K. M. Bhattarai, M. Galo-
bardes, G. Naredo, M. N. Perez-Payan, A. P. Davis, Org. Biomol. Chem.
2004, 2, 3320–3328.
[8] C. M. P. Rodrigues, S. Solꢆ, J. C. Sharpe, J. J. G. Moura, C. J. Steer, Bio-
chemistry 2003, 42, 3070–3080.
[23] S. S. Dhareshwar, V. J. Stella, J. Pharm. Sci. 2008, 97, 4184–4193.
[24] I. G. Georg, S. P. Patil, A. K. Saluja, R. Chugh, S. M. Vickers (University of
Minnesota, USA), 2010-US34117, 2010.
[25] D. A. DeGoey, D. J. Grampovnik, W. J. Flosi, K. C. Marsh, X. C. Wang, L. L.
Klein, K. F. McDaniel, Y. Liu, M. A. Long, W. M. Kati, A. Molla, D. J. Kempf,
J. Med. Chem. 2009, 52, 2964–2970.
[26] H. Komatsu, H. Awano, J. Org. Chem. 2002, 67, 5419–5421.
[27] H. Komatsu, I. Ikeda, Nucleosides Nucleotides Nucleic Acids 2003, 22,
1685–1686.
[28] L. Bialy, H. Waldmann, Chem. Eur. J. 2004, 10, 2759–2780.
[29] D. Barnes, R. Chopra, S. L. Cohen, J. Fu, M. Kato, P. Lu, M. Seepersaud,
W. Zhang, (Novartis AG), 2011-EP56304, 2011.
[30] M. Safadi, R. Oliyai, V. Stella, Pharm. Res. 1993, 10, 1350–1355.
[31] D. Farquhar, S. Khan, D. N. Srivastva, P. P. Saunders, J. Med. Chem. 1994,
37, 3902–3909.
[32] S. Solꢆ, X. Ma, R. E. Castro, B. T. Kren, C. J. Steer, C. M. P. Rodrigues, J.
Biol. Chem. 2003, 278, 48831–48838.
[9] C. M. P. Rodrigues, S. Solꢆ, Z. Nan, R. E. Castro, P. S. Ribeiro, W. C. Low,
C. J. Steer, Proc. Natl. Acad. Sci. USA 2003, 100, 6087–6092.
[10] C. M. P. Rodrigues, S. R. Spellman, S. Sola, A. W. Grande, C. Linehan-Sti-
eers, W. C. Low, C. J. Steer, J. Cereb. Blood Flow Metab. 2002, 22, 463–
471.
[11] S. Gupta, S. Li, M. J. Abedin, K. Noppakun, L. Wang, T. Kaur, B. Najafian,
C. M. P. Rodrigues, C. J. Steer, PLoS One 2012, 7, e48950.
[12] A. L. Rivard, C. J. Steer, B. T. Kren, C. M. P. Rodrigues, R. E. Castro, R. W.
Bianco, W. C. Low, Am. J. Chin. Med. 2007, 35, 279–295.
[13] G. Paumgartner, U. Beuers, Hepatology 2002, 36, 525–531.
[14] M. Arat, R. Idilman, E. A. Soydan, I. Soykan, E. Erden, S. KarayalÅın, H.
Akan, Clin. Transplant. 2005, 19, 798–803.
[15] D. Festi, M. Montagnani, F. Azzaroli, F. Lodato, G. Mazzella, A. Roda, A. R.
Di Biase, E. Roda, P. Simoni, A. Colecchia, Curr. Clin. Pharmacol. 2007, 2,
155–177.
[16] M. Omata, H. Yoshida, J. Toyota, E. Tomita, S. Nishiguchi, N. Hayashi, S.
Iino, I. Makino, K. Okita, G. Toda, K. Tanikawa, H. Kumada, Gut 2007, 56,
1747–1753.
[33] S. Solꢆ, J. D. Amaral, R. E. Castro, R. M. Ramalho, P. M. Borralho, B. T.
Kren, H. Tanaka, C. J. Steer, C. M. P. Rodrigues, Hepatology 2005, 42,
925–934.
[17] J. Rautio, H. Kumpulainen, T. Heimbach, R. Oliyai, D. Oh, T. Jarvinen, J.
Savolainen, Nat. Rev. Drug Discovery 2008, 7, 255–270.
[18] K. M. Huttunen, H. Raunio, J. Rautio, Pharmacol. Rev. 2011, 63, 750–771.
[19] Prodrugs: Challenges and Rewards, Part 2, (Eds.: V. J. Stella, R. T. Borch-
ardt, M. J. Hageman, R. Oliyai, H. Maag, J. W. Tilley), Springer, New York,
2007.
[34] A. Mꢃntylꢃ, J. Vepsꢃlꢃinen, T. Jꢃrvinen, T. Nevalainen, Tetrahedron Lett.
2002, 43, 3793–3794.
[35] T. Benneche, P. Strande, K. Undheim, Synthesis 1983, 762–763.
[20] V. J. Stella, K. W. Nti-Addae, Adv. Drug Delivery Rev. 2007, 59, 677–694.
[21] D. Kritchevsky, G. Poli, C. Scolastico, C. R. Sirtori, Steroids 1986, 47, 41–
48.
Received: February 7, 2013
Published online on May 2, 2013
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2013, 8, 1002 – 1011 1011