The cytostatic activity of NUC-3073, a phosphoramidate prodrug of 5-fluoro-2′-deoxyuridine, is independent of activation by thymidine kinase and insensitive to degradation by phosphorolytic enzymes

Article abstract of DOI:10.1016/j.bcp.2011.05.024

A novel phosphoramidate nucleotide prodrug of the anticancer nucleoside analogue 5-fluoro-2′-deoxyuridine (5-FdUrd) was synthesized and evaluated for its cytostatic activity. Whereas 5-FdUrd substantially lost its cytostatic potential in thymidine kinase (TK)-deficient murine leukaemia L1210 and human lymphocyte CEM cell cultures, NUC-3073 markedly kept its antiproliferative activity in TK-deficient tumour cells, and thus is largely independent of intracellular TK activity to exert its cytostatic action. NUC-3073 was found to inhibit thymidylate synthase (TS) in the TK-deficient and wild-type cell lines at drug concentrations that correlated well with its cytostatic activity in these cells. NUC-3073 does not seem to be susceptible to inactivation by catabolic enzymes such as thymidine phosphorylase (TP) and uridine phosphorylase (UP). These findings are in line with our observations that 5-FdUrd, but not NUC-3073, substantially loses its cytostatic potential in the presence of TP-expressing mycoplasmas in the tumour cell cultures. Therefore, we propose NUC-3073 as a novel 5-FdUrd phosphoramidate prodrug that (i) may circumvent potential resistance mechanisms of tumour cells (e.g. decreased TK activity) and (ii) is not degraded by catabolic enzymes such as TP which is often upregulated in tumour cells or expressed in mycoplasma-infected tumour tissue.

Full text of DOI:10.1016/j.bcp.2011.05.024

Biochemical Pharmacology 82 (2011) 441–452
Contents lists available at ScienceDirect
Biochemical Pharmacology
journal homepage: www.elsevier.com/locate/biochempharm
The cytostatic activity of NUC-3073, a phosphoramidate prodrug of
5-fluoro-20-deoxyuridine, is independent of activation by thymidine
kinase and insensitive to degradation by phosphorolytic enzymes
Johan Vande Voorde a, Sandra Liekens a, Christopher McGuigan b, Paola G.S. Murziani b,
Magdalena Slusarczyk b, Jan Balzarini a,
*
a Rega Institute for Medical Research, K.U.Leuven, Minderbroedersstraat 10, B-3000 Leuven, Belgium
b Welsh School of Pharmacy, Cardiff University, King Edward VII Avenue, Cardiff, CF10 3NB, United Kingdom
A R T I C L E I N F O
A B S T R A C T
A novel phosphoramidate nucleotide prodrug of the anticancer nucleoside analogue 5-fluoro-20-
deoxyuridine (5-FdUrd) was synthesized and evaluated for its cytostatic activity. Whereas 5-FdUrd
substantially lost its cytostatic potential in thymidine kinase (TK)-deficient murine leukaemia L1210 and
human lymphocyte CEM cell cultures, NUC-3073 markedly kept its antiproliferative activity in TK-
deficient tumour cells, and thus is largely independent of intracellular TK activity to exert its cytostatic
action. NUC-3073 was found to inhibit thymidylate synthase (TS) in the TK-deficient and wild-type cell
lines at drug concentrations that correlated well with its cytostatic activity in these cells. NUC-3073 does
not seem to be susceptible to inactivation by catabolic enzymes such as thymidine phosphorylase (TP)
and uridine phosphorylase (UP). These findings are in line with our observations that 5-FdUrd, but not
NUC-3073, substantially loses its cytostatic potential in the presence of TP-expressing mycoplasmas in
the tumour cell cultures. Therefore, we propose NUC-3073 as a novel 5-FdUrd phosphoramidate prodrug
that (i) may circumvent potential resistance mechanisms of tumour cells (e.g. decreased TK activity) and
(ii) is not degraded by catabolic enzymes such as TP which is often upregulated in tumour cells or
expressed in mycoplasma-infected tumour tissue.
Article history:
Received 5 April 2011
Accepted 23 May 2011
Available online 31 May 2011
Keywords:
Nucleoside analogues
Cancer
Phosphoramidate prodrugs
Mycoplasma
Thymidine phosphorylase
5-Fluoro-20-deoxyuridine (5-FdUrd)
ß 2011 Elsevier Inc. All rights reserved.
1. Introduction
substrate dUMP to the enzyme [6–8]. TS is the enzyme responsible
for the conversion of dUMP to TMP and is thereforeindispensablefor
Current treatment of cancer using chemotherapeutics is largely
based on the use of nucleoside analogues. These molecules are
designed to mimic natural pyrimidine and purine nucleosides.
After uptake by the cell, they are phosphorylated by cellular
enzymes such as (deoxy)cytidine kinase (dCK), thymidine kinase
(TK) and/or nucleo(s)(t)ide kinases. These antimetabolites can
subsequently interfere with the de novo synthesis of DNA/RNA
precursors to eventually inhibit DNA/RNA synthesis resulting in
cytotoxic/static activity [1,2].
cell proliferation, making it a crucial target for drug action. Among
the fluoropyrimidines mentioned above, 5-FdUrd requires only one
metabolicconversion,aphosphorylationcatalysedbyTKtogenerate
5-FdUMP [8]. This obligatory phosphorylation is often the rate-
limiting step in the metabolism of many anti-cancer drugs
(including 5-FdUrd), and is therefore still one of the limiting factors
for the therapeutic use of nucleoside analogues. Hence, different
strategies to improve the antitumour efficacy of nucleoside
analogues are being investigated [2].
Fluoropyrimidine-based antimetabolites such as fluorouracil (5-
FU), capecitabine and 5-fluoro-20-deoxyuridine (5-FdUrd) are
mainly used in the treatment of colon, breast and ovarian carcinoma
[3–5]. Intracellularly, these drugs are metabolized to the monopho-
sphate 5-FdUMP, which forms a stable inhibitory complex with
thymidylate synthase (TS) and the reduced co-substrate 5,10-
methylenetetrahydrofolate, thereby blocking binding of the normal
The charged nature of nucleoside monophosphates under
physiological conditions results in poor, if any, penetration across
the cell membrane [9]. They are also subject to extracellular
dephosphorylation. Therefore, the direct administration of phos-
phorylated molecules to circumvent the first phosphorylation step
has little therapeutic advantage. Hence, different strategies for
bypassing the rate-limiting phosphorylation using various types of
nucleoside 50-monophosphate prodrugs for more efficient drug-
delivery have been explored [10]. The administration of lipophilic
phosphoramidate nucleotide prodrugs (ProTides) has proved
successful for several molecules with anti-viral/cancer activity
* Corresponding author. Tel.: +32 16 337352; fax: +32 16 337340.
E-mail address: jan.balzarini@rega.kuleuven.be (J. Balzarini).
0006-2952/$ – see front matter ß 2011 Elsevier Inc. All rights reserved.
doi:10.1016/j.bcp.2011.05.024

442
J. Vande Voorde et al. / Biochemical Pharmacology 82 (2011) 441–452
[11–13]. By masking the charges of the phosphate motif, good
passive membrane diffusion of the prodrugs can be accomplished
after which the prodrug is rapidly converted intracellularly into the
nucleoside monophosphate by enzymatic cleavage [9].
Mycoplasmas are the smallest known self-replicating organ-
isms and are characterized by the lack of a cell wall and a strongly
reduced genome (600–1200 kb). Many of these bacteria have a
parasitic lifestyle and reside in the human body causing
asymptomatic infections [14]. It was shown that these prokaryotes
tend to preferentially colonize tumour tissue: Huang et al. reported
that 40–56% of human gastric, colon, oesophageal, lung and breast
cancers are infected with mycoplasmas compared to 21–30% in
non-tumourigenic tissue [15]. Pehlivan et al. found 22% of small
cell lung cancer tissue samples and >80% kidney tissue samples of
patients suffering renal cell carcinoma to be infected with
mycoplasmas compared to 5% and 14%, respectively, in control
tissue samples [16,17]. Chan et al. reported a 59% mycoplasma
infection rate in ovarian cancer tissues [18] and other studies also
reported a high infection rate in gastric [19,20] and cervical
condyloma tissues [21]. Due to their reduced set of genes,
mycoplasmas lack the pathway for de novo pyrimidine and
purine synthesis and therefore express a wide array of salvage
nucleo(s)(t)ide-metabolizing enzymes, such as thymidine phos-
phorylase (TP), deoxycytidine deaminase, etc. [22–25]. Already in
1985 it was observed that mycoplasma-encoded enzymes (e.g. TP),
present in contaminated cell cultures, lead to decreased dTTP
incorporation in lymphocytes [26]. Recently, it has been demon-
strated that these enzymes, in particular the mycoplasma-encoded
TP, can also interfere with the cytostatic activity of several
chemotherapeutics, including 5-trifluorothymidine, in vitro [27–
29]. Therefore we hypothesized that the elimination of mycoplas-
mas by antibiotics, suppression of mycoplasma-encoded enzymes
in human tumour tissue or the development of mycoplasma-
insensitive nucleoside analogue prodrugs may optimize treatment
of cancer patients using purine and pyrimidine antimetabolites
[29]. In the absence of such approaches, cancer patients may
receive sub-optimal chemotherapeutic treatment.
Scheme1. Reagents and conditions: (i) dry Et2O, dryEt3N, ꢀ78 8C, 30 minthen R.T., 3 h.
McGuigan et al. [30–32]. Arylphosphorodichlorophosphate (2) has
been prepared by coupling commercially available 1-naphthol
(Sigma–Aldrich, Dorset, UK) (3) with phosphorus oxychloride
(Sigma–Aldrich) (4) in the presence of Et3N (Sigma–Aldrich)
(Scheme 1) and this was allowed to react with L-alanine benzyl
ester tosylate (NovaBiochem (now Merck Chemicals Ltd, Darm-
stadt, Germany)) (5) in the presence of Et3N to generate the
phosphorochloridate derivative (6) (Scheme 2).
The nucleoside 5-FdUrd (Fig. 1) (Carbosynth Ltd, Berkshire, UK)
(7) was converted to the 50-ProTide derivative by coupling with the
phosphorochloridate derivative (6) in THF (Sigma–Aldrich) in the
presence of N-methyl imidazole (NMI) (Sigma–Aldrich) to give the
target compound NUC-3073 (1) (Scheme 3). The product was
obtained as a mixture of two diastereoisomers as confirmed by the
presence of two peaks in the 31P and 19F NMR spectra, and two
closely spaced peaks detectable by HPLC.
Alternatively, the compound 1 was prepared using tBuMgCl
(Sigma–Aldrich) (1 M solution in THF). Due to the lack of selectivity
towards the primary hydroxyl group, formation of 30-O-phosphor-
ylated derivative 8 was also observed (Scheme 4).
2.1.1. General
Anhydrous solvents were obtained from Sigma–Aldrich and
used without further purification. All reactions were carried out
under an argon atmosphere. Reactions were monitored with
analytical TLC on Silica Gel 60-F254 precoated aluminium plates
(Merck Chemicals Ltd) and visualized under UV (254 nm) and/or
with 31P NMR spectra. Column chromatography was performed on
This study was aimed at the development and assessment of TK-
independent phosphoramidate prodrugs of 5-FdUrd that would also
be insensitive to the TP-dependent inactivation of its free nucleoside
analogue. From a wide variety of newly synthesized phosphorami-
date prodrugs of 5-FdUrd, NUC-3073 (Fig. 1) was chosen for further
in depth studies. This molecule contains a naphthyl and benzyla-
laninyl group to mask the charged 50-phosphate on 5-FdUMP.
silica gel (35–70
m
M). Proton (1H), carbon (13C), phosphorus (31P)
and fluorine (19F) NMR spectra were recorded on a Bruker Avance
500 spectrometer (Bruker UK, Coventry, UK) at 25 8C. Spectra were
auto-calibrated to the deuterated solvent peak and all 13C NMR and
31P NMR were proton-decoupled. Analytical HPLC was conducted
by a Varian Prostar (LC Workstation-Varian prostar 335 LC
detector) using Varian Polaris C18-A (Agilent Technologies,
Edinburgh, UK) (10
mM) as an analytic column.
2. Materials and methods
Low and high resolution mass spectra were performed as a
service by Birmingham University, using electrospray (ES). CHN
microanalysis was performed as a service by MEDAC Ltd, Surrey.
2.1. Compound synthesis
The compound NUC-3073 (Fig. 1) (1) has been synthesized
2.1.2. Synthesis of 1-naphthyl dichlorophosphate (2)
using phosphorochloridate chemistry as previously reported by
Phosphorus oxychloride (1.94 mL, 20.81 mmol) was added to a
solution of 1-naphthol (3.00 g, 20.81 mmol) in diethyl ether
(Sigma–Aldrich) (70 mL) under argon atmosphere, then anhydrous
triethylamine (2.9 mL, 20.81 mmol) was added dropwise at ꢀ78 8C
and the resulting reaction mixture was stirred for 30 min.
Subsequently the reaction mixture was allowed to slowly warm
up to room temperature for 3 h. A formation of the desired
compound (2) was monitored by 31P NMR. After the reaction was
completed, the resulting mixture was filtered and then evaporated
in vacuo under nitrogen to afford the crude colorless oil as product,
which was used without further purification in the next step
(4.59 g, 84%) [Rf = 0.93 (hexane–EtOAc (Fisher Scientific, Leicester-
shire, UK), 1:1)], 31P NMR (202 MHz, CDCl3): dP 5.07; 1H NMR
(500 MHz, CDCl3): dH 7.52–7.71 (m, 4H, ArH), 7.86–7.89 (m, 1H,
ArH), 7.95–7.98 (m, 1H, ArH), 8.16–8.19 (m, 1H, ArH).
Fig. 1. Structural formula of 5-FdUrd and its phosphoramidate prodrug NUC-3073.

J. Vande Voorde et al. / Biochemical Pharmacology 82 (2011) 441–452
443
Scheme 2. Reagents and conditions: (i) dry Et3N, dry CH2CI2, ꢀ78 8C, 30 min then R.T., 1 h.
Scheme 3. Reagents and conditions: (i) NMI, dry THF, then phosphorochloridate (6), R.T., 16 h.
2.1.3. Synthesis of 1-naphthyl-(benzyl-
L-alaninyl)
NMI (0.40 mL, 5.07 mmol) was added dropwise. The reaction
mixture was allowed to stir for 30 min, and then a solution of 1-
phosphorochloridate (6)
Anhydrous triethylamine (2.66 mL, 19.14 mmol) was added
naphthyl-(benzyl-L-alaninyl) phosphorochloridate (6) (0.82 g,
dropwise under argon atmosphere at ꢀ78 8C to a stirred solution of
3.04 mmol) dissolved in anhydrous THF (3 mL) was added
dropwise. The reaction mixture was stirred at room temperature
for 16 h and then evaporated in vacuo to give a residue that was re-
dissolved in CH2Cl2 (Fisher Scientific), washed twice with 0.5 M
HCl (Sigma–Aldrich) (2 ꢁ 5 mL). The organic phase was purified by
column chromatography on silica gel, eluting with CH2Cl2–MeOH
(Fisher Scientific) as a gradient (0–5% MeOH) to afford the product
1 as colorless solid (47.0 mg, 8%) [Rf = 0.19 (CH2Cl2–MeOH, 95:5)],
(Found: MNa+, 636.1520. C29H29N3O9FNaP requires [MNa+],
636.1523); 31P NMR (202 MHz, MeOD, mixture of diastereoi-
somers 43%:57%): dP 4.61, 4.25; 19F NMR (470 MHz, MeOD): dF
ꢀ167.45, ꢀ167.25; 1H NMR (500 MHz, MeOD): dH 8.18–8.12 (m,
1H, ArH), 7.90–7.86 (m, 1H, ArH), 7.72–7.67 (m, 2H, ArH, H-6),
7.55–7.47 (m, 3H, ArH), 7.45–7.27 (m, 6H, ArH), 6.16–6.06 (m, 1H,
H-10), 5.13, 5.08 (2 ꢁ AB system, 2H, J = 12.0 Hz, OCH2Ph), 4.36–
4.24 (m, 3H, 2 ꢁ H-50, H-30), 4.15–4.03 (m, 2H, CHCH3, H-40), 2.17-
2.08 (m, 1H, H-20), 1.79–1.67 (m, 1H, H-20), 1.38–1.34 (m, 3H,
1-naphthyl dichlorophosphate (2) (2.50 g, 9.57 mmol) and
L-
alanine benzyl ester p-tosylate salt (NovaBiochem) (3.36 g,
9.57 mmol), in anhydrous DCM (Sigma–Aldrich) (50 mL). The
reaction mixture was stirred at ꢀ78 8C for 30 min and then allowed
to slowly warm to room temperature and stirred for 1 h. A
formation of the desired compound was monitored by 31P NMR.
After the reaction was completed, the solvent was removed under
reduced pressure and the crude residue was purified by flash
chromatography (hexane/EtOAc 7/3) to give the product (6) as an
oil (1.82 g, 47%) [Rf = 0.90 (hexane–EtOAc, 7:3)], 31P NMR
(202 MHz, CDCl3, mixture of diastereoisomers): dP 7.92, 8.14
(Int.: 1.00:1.00); 1H NMR (500 MHz, CDCl3, mixture of diastereoi-
somers with a ratio of 1:1): dH 1.42–1.45 (m, 3H, CHCH3), 4.20–4.23
(m, 1H, CHCH3), 4.78–4.81 (m, 1H, NH), 5.09 (s, 2H, OCH2Ph), 7.09–
7.73 (m, 11H, ArH), 7.97–8.12 (m, 1H, ArH).
3
2.1.4. Synthesis of 5-fluoro-20-deoxyuridine-50-O-[
naphthyl(benzyl-
a
-
CHCH3); 13C NMR (125 MHz, MeOD): dC 174.9 (d, JC-P = 4.3 Hz,
3
2
L
-alaninyl)] phosphate (1)
C55O, ester), 174.6 (d, JC-P = 5.0 Hz, C55O, ester), 159.3 (d, JC-
F = 26.1 Hz, CH-base), 150.5 (d, 4JC-F = 4.0 Hz, C55O, base), 147.9 (d,
To a solution of 5-fluoro-20-deoxyuridine (Carbosynth) (0.25 g,
2
1.01 mmol) in dry THF (10 mL) at 0 8C under argon atmosphere,
2JC-P = 7.4 Hz, C-Ar, Naph), 147.8 (d, JC-P = 7.7 Hz, C-Ar, Naph),
Scheme 4. Reagents and conditions: (i) tBuMgCl, dry THF, then phosphorochloridate (6), R.T., 16 h.

444
J. Vande Voorde et al. / Biochemical Pharmacology 82 (2011) 441–452
141.7 (d, 1JC-F = 233.9 Hz, CF), 141.6 (d, 1JC-F = 233.8 Hz, CF), 137.2,
137.1, 136.2 (C-Ar), 129.7, 129.6, 129.5, 129.4, 129.0, 128.9, 128.1,
128.0 (CH-Ar), 127.9, 127.8 (C-Ar), 127.7, 127.6, 126.6, 126.5,
(CHCH3), 39.9 (C-20), 20.4, 20.3 (CHCH3); m/z (ES) 613 (M+, 100%);
reverse-phase HPLC eluting with H2O/MeOH from 100/0 to 0/100
in 45 min, 1 ml/min, l = 275 nm, showed two peaks of the
diastereoisomers with tR 26.69 min and tR 27.41 min (23%:72%).
3
126.2, 125.7, 125.6, 125.4, 125.3, 122.6 (CH-Ar), 116.5 (d, JC-
P = 3.5 Hz, CH-Ar), 116.2 (d, 3JC-P = 3.3 Hz, CH-Ar), 87.0, 86.9 (C-10),
3
86.8, 86.7 (2 ꢁ d, JC-P = 8.1 Hz, C-40), 72.1, 72.0 (C-30), 68.1, 68.0
2.1.6. Synthesis of 5-fluoro-20-deoxyuridine-50-O-[(
L-alaninyl)]
2
(CH2Ph), 67.8, 67.6 (2 ꢁ d, JC-P = 5.2 Hz, C-50), 51.9, 51.8 (CHCH3),
phosphate ammonium salt (11)
3
3
40.9, 40.8 (C-20), 20.5 (d, JC-P = 6.5 Hz, CHCH3), 20.3 (d, JC-
P = 7.6 Hz, CHCH3); m/z (ES) 636 (MNa+, 100%), reverse HPLC
eluting with (H2O/MeOH from 100/0 to 0/100) in 45 min, showed
two peaks of the diastereoisomers with tR 34.23 min and tR
34.59 min. Analytically calculated for C29H29FN3O9P: C, 56.77; H,
4.76; N, 6.85. Found: C, 56.57; H, 5.06; N, 6.72; UV (0.05 M
5-Fluoro-20-deoxyuridine-50-O-[(
L
-alaninyl)] phosphate am-
monium salt was prepared by dissolving 5-fluoro-20-deoxyuri-
dine-50-O-[
-naphthyl(benzyl- -alaninyl)] phosphate (1) (0.08 g,
a
L
0.130 mmol) in a solution of triethylamine (5 mL) and water (5 mL)
(Scheme 5). The reaction mixture was stirred at 35 8C for 16 h and
then the solvents were removed under reduced pressure. The
residue was treated with water and extracted with dichloro-
methane (Fisher Scientific). The aqueous layer was concentrated
and evaporated under reduced pressure, then the resultant crude
material was purified by column chromatography on silica, eluting
with 2-propanol–H2O–NH3 (Fisher Scientific) (8:1:1) to afford the
title compound 11 as a colorless solid (15.0 mg, 30%) [Rf = 0.04 (2-
propanol–H2O–NH3 (8:1:1))]; 31P NMR (202 MHz, D2O): dP 7.13;
19F NMR (470 MHz, D2O): dF ꢀ168.00; 1H NMR (500 MHz, D2O): dH
7.93 (d, 1H, 3JH-F = 6.1 Hz, H-6), 6.25–6.30 (m, 1H, H-10), 4.44–4.49
(m, 1H, H-30), 4.06–4.11 (m, 1H, H-40), 3.94–3.83 (m, 2H, CH2OPh),
3.53 (q, 1H, J = 7.5 Hz, CHCH3), 2.28–2.37 (m, 2H, H-20), 1.19–1.25
phosphate buffer (Sigma–Aldrich), pH 7.4)
lmax = 271 nm
(
e
max = 7050); log P (octanol (Sigma–Aldrich):water) measured:
1.7, calculated 3.53 (Chemoffice 11.0).
2.1.5. Synthesis of 5-fluoro-20-deoxyuridine-50-O-[
a-
naphthyl(benzyl-L-alaninyl)] phosphate (1) – alternative method
To a solution of 5-fluoro-20-deoxyuridine (0.35 g, 1.42 mmol)
in dry THF (15 mL), tBuMgCl (1 M in THF, 0.40 mL, 1.56 mmol) was
added dropwise. The reaction mixture was stirred for 30 min and
then a solution of 1-naphthyl-(benzyl-L-alaninyl) phosphoro-
chloridate (6) (1.14 g, 2.84 mmol) dissolved in anhydrous THF
(3 mL) was added dropwise. The reaction mixture was stirred at
room temperature for 16 h and then evaporated under reduced
3
(m, 3H, CHCH3); 13C NMR (125 MHz, MeOD): dC 174.8 (d, JC-
2
4
P = 4.6 Hz, C55O), 159.2 (d, JC-F = 26.2 Hz, CH-base), 150.3 (d, JC-
F = 4.0 Hz, C55O, base), 141.8 (d, 1JC-F = 233.8 Hz, CF-base), 87.0 (C-
10), 86.7 (d, 3JC-P = 7.5 Hz, C-40), 71.1 (C-30), 67.2 (d, 2JC-P = 5.5 Hz, C-
pressure to give
a crude residue. Purification by column
chromatography on silica gel (Merck Chemicals) eluting with
gradient of MeOH (0–5%) in CH2Cl2 afforded the product 1 as a
colorless solid (78 mg, 9%). (All the spectroscopic data of 1 are in
agreement with data reported in Section 2.1.4.) Alongside with
the desired product 1, the 30-ProTide analogue (8) was isolated
and characterized by NMR and HPLC.
3
50), 51.0 (CHCH3), 40.2 (C-20), 20.3 (d, JC-P = 7.2 Hz, CHCH3); m/z
(ES) 396.1 (Mꢀ2NH4ꢀ+H]ꢀ, 100%).
2.2. Radioactive pyrimidine deoxynucleosides
5-Fluoro-20-deoxyuridine-30-O-[
a
-naphthyl(benzyl-L-alani-
[5-3H]dCyd (radiospecificity: 22 Ci/mmol) and [5-3H]dUrd
(radiospecificity: 15.9 Ci/mmol) were obtained from Moravek
Biochemicals Inc. (Brea, CA).
nyl)] phosphate (8) was obtained as a colorless solid (15.0 mg, 2%)
[Rf = 0.50 (CH2Cl2–MeOH, 95:5)], (Found: MNa+, 636.1525.
C
29H29N3O9NaPF requires [MNa+], 636.1523); 31P NMR
(202 MHz, MeOD): dP 3.79, 3.11; 19F NMR (470 MHz, MeOD): dF
ꢀ168.62, ꢀ168.53; 1H NMR (500 MHz, MeOD): dH 8.24–8.15 (m,
2.3. Cell cultures
3
2H, ArH), 7.94–7.88 (m, 1H, ArH), 7.73 (d, 1H, JC-F = 7.9 Hz, H-6),
Murine leukaemia L1210/0 and human T-lymphocyte CEM/0
cells were obtained from the American Type Culture Collection
(ATCC) (Rockville, MD). Human glioblastoma U87 cells were kindly
provided by Dr. E. Menue (Institut Pasteur, Paris, France).
Thymidine kinase-deficient CEM/TKꢀ cells were a kind gift from
Prof. S. Eriksson (currently at Uppsala University, Uppsala,
Sweden) and Prof. A. Karlsson (Karolinska Institute, Stockholm,
Sweden). Thymidine kinase-deficient L1210/TKꢀ were derived
from L1210/0 cells after selection for resistance against 5-bromo-
20-dUrd [33]. Infection of relevant cell lines with Mycoplasma
hyorhinis (ATCC) resulted in chronically infected cell lines further
referred to as L1210.Hyor and U87.Hyor. All cells were maintained
in Dulbecco’s modified Eagle’s medium (DMEM) (Invitrogen,
Carlsbad, CA) with 10% foetal bovine serum (FBS) (Biochrom AG,
7.60–7.40 (m, 3H, ArH), 7.38–7.24 (m, 6H, ArH), 6.28–6.17 (m, 1H,
H-10), 5.14–5.11 (m, 2H, CH2Ph), 4.21–4.03 (m, 2H, CHCH3, H-30),
3.93–3.63 (m, 3H, 2 ꢁ H-50, H-40), 2.51–2.43 (m, 1H, H-20), 2.32–
2.23 (m, 1H, H-20), 1.37 (d, 3H, J = 7.2 Hz, CHCH3, one diast.), 1.35
(d, 3H, J = 7.2 Hz, CHCH3, one diast.); 13C NMR (125 MHz, MeOD):
3
3
174.6 (d, JC-P = 4.8 Hz, C55O ester), 174.4 (d, JC-P = 5.1 Hz, C55O
2
4
ester), 159.1 (d, JC-F = 26.0 Hz, C-base), 150.2 (d, JC-F = 4.1 Hz,
2
2
C55O, base), 148.0 (d, JC-P = 7.2 Hz, C-Ar, Naph), 147.9 (d, JC-
P = 7.2 Hz, C-Ar, Naph), 141.4 (d, 1JC-F = 233.0 Hz, CF), 141.2 (d, 1JC-
F = 233.0 Hz, CF), 137.5, 137.4, 136.3 (C-Ar), 130.8, 129.6, 129.5,
129.3, 129.2, 129.0, 128.9, 128.1, 128.0, 127.8, 127.6, 126.2, 126.1,
125.7, 122.6, 116.7, 116.3, 106.5 (CH-Ar), 87.5, 87.4 (C-10), 86.6 (C-
4
40), 79.7 (d, JC-P = 4.9 Hz, C-30), 68.0 (C-50), 62.6 (CH2Ph), 51.8
Scheme 5. Reagents and conditions: (i) Et3N:H2O (1:1), 35 8C, 16 h.

J. Vande Voorde et al. / Biochemical Pharmacology 82 (2011) 441–452
445
Berlin, Germany), 10 mM Hepes and 1 mM sodium pyruvate
(Invitrogen). Cells were grown at 37 8C in a humidified incubator
with a gas phase of 5% CO2.
and appropriate amounts of the test compounds (5-FdUrd and
NUC-3073). After 30 min, 2 h and 4 h pre-incubation of the cells
with the compounds at 37 8C, 1
was added to the cell cultures. After 30 min incubation, 100
the cell suspensions were withdrawn and added to a cold
suspension of 500 activated charcoal (VWR, Haasrode,
m
Ci of [5-3H]dUrd or [5-3H]dCyd
L of
m
2.4. Cytostatic assays
m
L
Monolayer cells (U87 and U87.Hyor) were seeded in 48-well
microtiter plates (NuncTM, Roskilde, Denmark) at 10 000 cells/
well. After 24 h, an equal volume of fresh medium containing
the test compounds was added. On day 5, cells were trypsinized
Belgium) (100 mg/ml in TCA (Merck Chemicals Ltd) 5%). After
10 min, the suspension was centrifuged at 13 000 rpm for 10 min,
after which the radioactivity in 400
mL supernatant was counted in
a
liquid scintillator using OptiPhase HiSafe (Perkin Elmer,
´
and counted in a Coulter counter (Analis, Suarlee, Belgium).
Waltham, MA).
Suspension cells (L1210/0, L1210/TKꢀ, L1210.Hyor, CEM/0, CEM/
TKꢀ) were seeded in 96-well microtiter plates (NuncTM) at
60 000 cells/well in the presence of a given amount of the test
compounds. The cells were allowed to proliferate for 48 h
(L1210) or 72 h (CEM) and were then counted in a Coulter
counter. The 50% inhibitory concentration (IC50) was defined as
the compound concentration required to reduce the number of
viable cells by 50%.
2.7. Stability assays
2.7.1. Carboxypeptidase Y (EC 3.4.16.1) assay
The enzymatic stability of the prodrug NUC-3073 towards
carboxypeptidase Y (Sigma–Aldrich) was studied using in situ
31P NMR (202 MHz). The experiment was carried out by
dissolving NUC-3073 (3.0 mg) in d6-acetone (Cambridge Isotope
Laboratories, Andover, MA) (150
(Sigma–Aldrich) buffer pH 7.6 (300
m
L) and adding TRIZMA
2.5. Exposure of 5-FdUrd and NUC-3073 to Escherichia coli-encoded
TP and human-encoded TP and UP
mL). The resulting solution
was placed in an NMR tube and a 31P NMR experiment at 25 8C
was recorded as the blank experiment. The enzyme carboxy-
The substrate specificity of TP towards natural thymidine
(dThd) (Sigma–Aldrich), uridine (Urd) (Sigma–Aldrich), 5-FdUrd
(Sigma–Aldrich) and NUC-3073 was investigated by high
pressure liquid chromatography (HPLC) (Waters, Milford, MA).
peptidase Y (0.2 mg) was dissolved in TRIZMA (150 mL) and
added to the solution of the phosphoramidate derivative in the
NMR tube. Next, the tube was placed in the NMR machine,
which was set to run a 31P NMR experiment (64 scans) every
4 min for 14 h at 25 8C. Data were processed and analyzed with
the Bruker Topspin 2.1 program.
Reaction mixtures containing 100
recombinant TP or uridine phosphorylase (UP) (human TP:
8.6 ng/ L; E. coli TP: 3.0 ng/ L; human UP: 4.0 ng/ L) in a total
volume of 500 L reaction buffer (10 mM Tris–HCl (Sigma–
Aldrich); NaCl (Sigma–Aldrich) (150 M for TP and 300 M for
mM test compound and
m
m
m
m
2.7.2. Human serum
m
m
The stability of the prodrug NUC-3073 in the presence of human
serum was studied using in situ 31P NMR (202 MHz). The ProTide
NUC-3073 (1) (5.0 mg) was dissolved in a mixture of DMSO
(Sigma–Aldrich)/D2O (Cambridge Isotope Laboratories) (0.05 mL/
0.15 mL) and then subjected to 31P NMR analysis at 37 8C. After the
first spectrum (control) was recorded, the NMR sample was treated
with human serum (Sigma–Aldrich) (0.3 mL) and immediately
subjected to further 31P NMR experiments at 37 8C. 31P NMR data
were recorded every 15 min over 12 h and 30 min. In order to
improve the visualization of the results all the spectra were further
processed using the Lorentz–Gauss deconvolution method. Data
recorded were processed and analyzed with the Bruker TopSpin 2.1
programme.
UP); 1 mM EDTA (Sigma–Aldrich); 2 mM KH2PO4 (Merck
Chemicals Ltd)/K2HPO4 (Acros Organics, Geel, Belgium)) were
incubated at room temperature. At different time points (i.e. 0,
20, 40 min) 100
mL aliquots of the reaction mixtures were
withdrawn and heated at 95 8C for 3 min to inactivate the
enzyme. The resulting reaction products were separated on a
reverse-phase RP-8 column (Merck Chemicals Ltd) and quanti-
fied by HPLC analysis (Alliance 2690, Waters). The separation of
dThd from thymine was performed by a linear gradient from
98% separation buffer (50 mM NaH2PO4 (Acros Organics) and
5 mM heptane sulfonic acid (Sigma–Aldrich), pH 3.2) and 2%
acetonitrile (BioSolve BV, Valkenswaard, the Netherlands), to
20% separation buffer + 80% acetonitrile (8 min 98% separation
buffer + 2% acetonitrile; 5 min linear gradient of 98% separation
buffer + 2% acetonitrile to 20% separation buffer + 80% acetoni-
trile; 10 min 20% separation buffer + 80% acetonitrile, followed
by equilibration at 98% separation buffer + 2% acetonitrile). UV-
based detection was performed at 267 nm. The separation of Urd
from uracil was performed by a linear gradient from 100%
separation buffer (see above) to 60% separation buffer + 40%
acetonitrile (3 min 100% separation buffer; 6 min linear gradient
of 100% separation buffer to 60% separation buffer + 40%
acetonitrile; 6 min 60% separation buffer + 40% acetonitrile,
followed by equilibration at 100% separation buffer). UV-based
detection was performed at 258 nm.
2.7.3. Buffer at pH 1
The stability of the prodrug NUC-3073 towards hydrolysis at
pH 1 was studied using in situ 31P NMR (202 MHz). The ProTide
NUC-3073 (1) (2.6 mg) was dissolved in MeOD (Cambridge
Isotope Laboratories) (0.1 mL) after which 0.5 mL buffer (pH 1,
prepared from equal parts of 0.2 M HCl and 0.2 M KCl (Sigma–
Aldrich)) was added. Then, on this sample 31P NMR experiments
were carried out at 37 8C recording the data every 12 min for
14 h. Data were processed and analyzed with the Bruker
TopSpin 2.1 programme.
2.7.4. Buffer at pH 8
The stability of the prodrug NUC-3073 towards hydrolysis at
pH 8 was studied using in situ 31P NMR (202 MHz). The ProTide
NUC-3073 (1) (4.9 mg) was dissolved in MeOD (0.1 mL) after
which 0.5 mL buffer (pH 8, prepared from a solution of 0.1 M
Na2HPO4 (Sigma–Aldrich), which was adjusted by 0.1 M HCl)
was added. Next, on this sample 31P NMR experiments were
carried out at 37 8C recording the data every 12 min for 14 h.
Data were processed and analyzed with the Bruker TopSpin 2.1
programme.
2.6. Thymidylate synthase (TS) activity measurements
The activity of TS in intact L1210/0 and L1210/TKꢀ cells was
measured by evaluation of tritium release from [5-3H]dUMP
(formed in the cells from [5-3H]dUrd or [5-3H]dCyd) in the reaction
catalysed by TS. This method has been described in detail by
Balzarini and De Clercq [34]. Briefly, cell cultures (500
culture medium) were prepared containing ꢂ3 ꢁ 106 L1210 cells
mL DMEM

446
J. Vande Voorde et al. / Biochemical Pharmacology 82 (2011) 441–452
Table 1
human TP. When both compounds were exposed to UP, dThd and
NUC-3073 were not affected by the enzyme, whereas 5-FdUrd was
slightly hydrolyzed (Fig. 2, panel C).
Cytostatic activity of 5-FdUrd and NUC-3073 as represented by the IC50 value in
different cell lines.
a
Cell lines
IC50
(m
M)
3.4. Inhibition of thymidylate synthase (TS) by 5-FdUrd and NUC-
3073
5-FdUrd
NUC-3073
L1210/0
0.0008 ꢃ 0.000095
3.1 ꢃ 0.14
0.011 ꢃ 0.0065
0.027 ꢃ 0.0028
0.025 ꢃ 0.0073
L1210/TK-
L1210.Hyor
The major target for the cytostatic activity of 5-FdUrd is TS. The
activity of TS in intact tumour cells can be directly monitored by
measuring the tritium release in intact L1210/0 cell cultures that
were exposed to [5-3H]deoxyuridine ([5-3H]dUrd) or [5-3H]deox-
ycytidine ([5-3H]dCyd). Indeed, after intracellular conversion of
[5-3H]dUrd or [5-3H]dCyd to [5-3H]dUMP, the C-5 tritium atom on
the pyrimidine base is released during the TS-catalysed reductive
methylation. The ability of 5-FdUrd and its prodrug NUC-3073 to
inhibit tritium release from [5-3H]dUrd and [5-3H]dCyd was
therefore evaluated in L1210/0 cell cultures at a variety of
0.24 ꢃ 0.054
CEM/0
CEM/TK-
0.028 ꢃ 0.0014
1.5 ꢃ 0.071
0.089 ꢃ 0.030
0.32 ꢃ 0.049
U87
0.007 ꢃ 0.001
3.0 ꢃ 0.55
0.035 ꢃ 0.0005
0.039 ꢃ 0.0025
U87.Hyor
a
50% inhibitory concentration or compound concentration required to inhibit
tumour cell proliferation by 50%.
3. Results
compound concentrations. 5-FdUrd proved to be
inhibitor of TS in situ. Its IC50 for tritium release from [5-3H]dCyd
and [5-3H]dUrd was around 0.0007–0.0009
M (Table 2). The
inhibitory activity of NUC-3073 on tritium release was much less
a potent
3.1. Cytostatic activity of 5-FdUrd and its prodrug NUC-3073 against
TK-competent and TK-deficient tumour cell lines
m
The cytostatic activity of 5-FdUrd and NUC-3073 was deter-
mined in different TK-expressing and TK-deficient tumour cell
lines. As shown in Table 1, 5-FdUrd is strongly dependent on the
expression of TK for its cytostatic activity. Its IC50 increased 4000-
E.coli TP
A
B
C
100
50
0
dThd
fold for L1210/TKꢀ cells (IC50: 3.1
cells (IC50: 0.0008
M) and 50-fold for CEM/TKꢀ cells (IC50
1.5 M) versus CEM/0 cells (IC50: 0.028 M). In contrast, the
mM) versus wild-type L1210/0
5-FdUrd
NUC-3073
Urd
m
:
m
m
cytostatic activity of the 5-FdUrd prodrug NUC-3073 remained
virtually unchanged in TK-deficient cells when compared with
wild-type cells (IC50: 0.027 and 0.011
L1210/0, and 0.32 and 0.089
M for CEM/TKꢀ and CEM/0 cells,
m
M for L1210/TKꢀ and
m
respectively). Although the cytostatic activity of NUC-3073 was 3–
10-fold inferior to 5-FdUrd against wild-type L1210/0 and CEM/0
cells, it proved 5–100-fold superior to 5-FdUrd in the TK-deficient
tumour cell lines (Table 1).
0
0
0
10
20
30
40
40
40
Incubation time (min)
Human TP
3.2. Effect of mycoplasma infection of tumour cell cultures on the
cytostatic activity of 5-FdUrd and its prodrug NUC-3073
100
50
0
dThd
5-FdUrd
NUC-3073
Urd
The L1210/0 cell cultures were infected with the mycoplasma
species M. hyorhinis (cells designated: L1210.Hyor). 5-FdUrd lost
its cytostatic activity against the mycoplasma-infected L1210.Hyor
cells by 300-fold (IC50: 0.24
mM). Also, 5-FdUrd lost its cytostatic
activity by 400-fold in U87.Hyor cell cultures when compared with
uninfected U87 cells (Table 1). In sharp contrast, the 5-FdUrd
prodrug NUC-3073 kept a similar cytostatic potential in both
L1210/0 and L1210.Hyor cell cultures (IC50: 0.011 and 0.025
mM,
10
20
30
respectively). A similar observation was made for this prodrug
when evaluated for its cytostatic activity in U87 and U87.Hyor cell
Incubation time (min)
Human UP
cultures (IC50: 0.035 and 0.039
mM, respectively). Thus, whereas
100
50
0
the free nucleoside 5-FdUrd markedly lost its cytostatic potential
against M. hyorhinis-infected tumour cell lines, the antiprolifera-
tive potential of its prodrug NUC-3073 was independent of the
mycoplasma infection.
dThd
5-FdUrd
NUC-3073
Urd
3.3. Phosphorolysis of 5-FdUrd and NUC-3073 by thymidine and
uridine phosphorylases
5-FdUrd and its prodrug NUC-3073 were exposed to purified TP
derived from E. coli or human erythrocytes, and to purified UP
derived from human tumour tissue. Whereas E. coli and human TP
rapidly converted dThd and 5-FdUrd to their free bases, NUC-3073
was completely stable in the presence of these enzymes (Fig. 2,
panels A and B). Under similar experimental conditions, uridine
was converted to uracil by human UP, but not by E. coli TP, or
10
20
30
Incubation time (min)
Fig. 2. Effect of thymidine phosphorylase from E. coli (panel A) and human (panel B)
source and human uridine phosphorylase (panel C) on dThd, Urd, 5-FdUrd and NUC-
3073. Data are the mean of at least 2 independent experiments (ꢃS.D.).

J. Vande Voorde et al. / Biochemical Pharmacology 82 (2011) 441–452
447
Table 2
(ꢂ200-fold) pronounced than that of 5-FdUrd, especially after
IC50 values of 5-FdUrd and NUC-3073 against TS in intact L1210/0 tumour cells (as
determined by tritium release from [5-3H]dUrd and [5-3H]dCyd after 30 min
exposure to the drugs).
only 30 min preincubation of the cells with the drugs (IC50: 0.16–
0.19
mM). However, longer pre-incubation times of the cells (up to
4 h) with 5-FdUrd and NUC-3073 before measuring TS activity in
the intact tumour cells revealed a much more pronounced
inhibitory activity of the prodrug against TS in situ (Fig. 3).
Indeed, whereas the inhibition of 3H release was only 2-fold
increased upon longer pre-incubation times of 5-FdUrd, the
inhibitory potential of NUC-3073 increased 10-fold (Fig. 3, panels
A and B, and C and D). Interestingly, pre-incubation of the tumour
cells with 5-FdUrd and NUC-3073 for at least 4 h results in TS
inhibition in the intact tumour cells at drug concentrations that
are very comparable with the 50% cytostatic activity concentra-
tions of these drugs. Taken all data together, our observations
indicate that the 5-FdUrd prodrug needs several metabolic
a
Compound
IC50
(
mM)
Tritium release from
[5-3H]dUrd
Tritium release from
[5-3H]dCyd
5-FdUrd
0.0009 ꢃ 0.0002
0.16 ꢃ 0.05
0.0007 ꢃ 0.003
0.19 ꢃ 0.08
NUC-3073
a
50% inhibitory concentration or compound concentration required to inhibit
tritium release from [5-3H]dUrd or [5-3H]dCyd in drug-exposed L1210/0 cell
cultures by 50%.
Inhibition of TS in L1210/0 cells by 5-FdUrd
(substrate: [5-³H]dUrd)
Inhibition of TS in L1210/0 cells by NUC-3073
A
B
(substrate: [5-³H]dUrd)
0.0015
0.3
0.0010
0.0005
0.0000
0.2
0.1
0.0
30 min
2 h
4 h
30 min
2 h
4 h
Preincubation time
Preincubation time
Inhibition of TS in L1210/0 cells by 5-FdUrd
(substrate: [5-³H]dCyd)
Inhibition of TS in L1210/0 cells by NUC-3073
(substrate: [5-³H]dCyd)
C
D
0.0015
0.3
0.0010
0.0005
0.0000
0.2
0.1
0.0
30 min
2 h
4 h
30 min
2 h
4 h
Preincubation time
Preincubation time
Inhibition of TS in L1210/TK-cells by 5-FdUrd
(substrate: [5-³H]dCyd)
Inhibition of TS in L1210/TK-cells by NUC-3073
(substrate: [5-³H]dCyd)
E
F
8
6
4
2
0
0.8
0.6
0.4
0.2
0.0
30 min
2 h
4 h
30 min
2 h
4 h
Preincubation time
Preincubation time
Fig. 3. Inhibition of TS by 5-FdUrd and NUC-3073 as measured by tritium release from [5-3H]dUrd (panels A and B) and [5-3H]dCyd (panels C and D) in L1210/0 cell cultures
and by tritium release from [5-3H]dCyd (panels E and F) in L1210/TKꢀ cell cultures. Data are the mean of 2 independent experiments (ꢃS.E.M.).

448
J. Vande Voorde et al. / Biochemical Pharmacology 82 (2011) 441–452
against TS remained virtually unchanged in L1210/TKꢀ cells (IC50
0.053 M in L1210/0 cells).
M in L1210/TKꢀ cells versus 0.013
:
Table 3
IC50 values of 5-FdUrd and NUC-3073 against TS in intact L1210/0 and L1210/TKꢀ
cells (as determined by tritium release from [5-3H]dCyd after 4 h of preincubation
with the products).
m
m
3.5. Metabolism of 5-FdUrd phosphoramidates
a
Compound
IC50
(m
M)
L1210/0
L1210/TKꢀ
As shown in Fig. 4, the putative mechanism of activation of
ProTides inside the cell, after uptake, involves a first enzymatic
activation step mediated by a carboxypeptidase-type enzyme that
hydrolyzes the ester of the aminoacyl moiety (step a, compound
9). Next, a spontaneous cyclization followed by subsequent,
spontaneous displacement of the aryl group (step b) and opening
of the unstable ring mediated by water (step c) can give
compounds 10 and 11, respectively. The last step involves a
hydrolysis of the P–N bond mediated by a phosphoramidase-type
enzyme (step d) with release of the nucleoside monophosphate
(12) in the intact cell [13,35].
To prove the proposed metabolic scheme for NUC-3073 (1) and
to determine whether the ester motif of the 5-FdUrd phosphor-
amidate derivative is cleaved-off or not, we carried out an enzyme
incubation experiment designed to mimic the first stages of the
putative activation in the intact tumour cells. The compound 1 was
incubated with carboxypeptidase Y (also known as cathepsin A) in
TRIZMA buffer and its conversion was monitored by 31P NMR.
5-FdUrd
0.0003 ꢃ 0.00003
0.013 ꢃ 0.008
1.42ꢃ 0.09
NUC-3073
0.053 ꢃ 0.0009
a
50% inhibitory concentration or compound concentration required to inhibit
tritium release from [5-3H]dCyd in drug-exposed L1210 cells by 50% upon pre-
exposure of the tumour cells for 4 h to 5-FdUrd or its prodrug NUC-3073.
conversion steps before reaching TS as the target enzyme for
inhibition, and support the view that NUC-3073 acts as an efficient
prodrug of 5-FdUrd to exert its eventual cytostatic activity.
The activity of TS in the presence of 5-FdUrd and NUC-3073 was
also measured in intact L1210/TKꢀ cells using [5-3H]dCyd as an
externally supplied substrate (due to TK deficiency, [5-3H]dUrd
cannot be used). As demonstrated in Table 3 and Fig. 3 (panels E
and F), the concentration of 5-FdUrd required to cause 50%
inhibition of TS increased by a factor 5700 in TK-deficient L1210/
TKꢀ cells (IC50: 1.4
(IC50: 0.0003 M). In contrast, the inhibitory activity of NUC-3073
mM) when compared to wild-type L1210/0 cells
m
Fig. 4. Proposed putative mechanism of activation of the 5-FdUrd ProTide NUC-3073.

J. Vande Voorde et al. / Biochemical Pharmacology 82 (2011) 441–452
449
Fig. 5. Carboxypeptidase-mediated cleavage of prodrug NUC-3073 (1) monitored by 31P NMR.
Spectra were recorded for 14 h acquiring scans at the periodic
intervals every 4 min as shown in Fig. 5. For a better resolution
original spectra (lower graphs) and deconvoluted ones (upper
graphs) are shown.
monophosphate was not detected in the enzyme experiment, as
expected. This experiment indicates that the first activation step of
ProTide NUC-3073 (1) may be sufficiently efficient, and therefore,
may allow the eventual delivery of the nucleoside monophosphate
metabolite in the intact tumour cells, following subsequent
phosphoramidase-mediated cleavage of the amino acid to liberate
FdUMP.
Two peaks at
d
4.07; 4.23 ppm in the 31P NMR spectrum
correspond to two diastereoisomers of the parent compound 1.
Upon the addition of cathepsin A, during the first 4 min of the
experiment the compound 1 was hydrolyzed to the corresponding
In order to support both, the putative mechanism proposed as
well as the results obtained in the enzymatic assay, we decided to
prepare the intermediate 11 via a synthetic route. Thus chemical
hydrolysis of the compound 1 in the presence of triethylamine and
water was performed (Scheme 5). The product 11 obtained as a
diammonium salt was next added to the assay’s final sample
intermediate(s) 9 (d 4.95; 5.16 ppm), which lack the ester motif
(Fig. 4). These intermediates did not persist as they were quickly
converted to the aminoacyl phosphoramidate derivative (the final
product in this assay) via the loss of the aryl group (steps a to c in
Fig. 4). Due to the achirality at the phosphate the metabolite 11
appeared as a singlet (
d
6.85 ppm). Thus, the enzymatic assay
4.00 with
(containing only the enzymatic metabolite 11 in TRIZMA), and 31
P
spectra showed a fast metabolism of the parent ꢂ
d
NMR analysis was investigated. The 31P NMR spectrum has shown
exclusively one peak at 6.85 ppm (data not shown); therefore our
results have proved the proposed metabolic pathway and
activation of ProTides.
complete conversion to the putative intermediate within 26 min,
which further stayed consistently present throughout the 14 h of
the assay. The cleavage of the P–N bond releasing the nucleoside
Fig. 6. 31P NMR spectrum of compound NUC-3073 (1) in human serum.

450
J. Vande Voorde et al. / Biochemical Pharmacology 82 (2011) 441–452
Fig. 7. 31P NMR spectrum of compound NUC-3073 (1) in buffer pH 1.
3.6. Stability studies
constant two peaks of diastereoisomers at
throughout the time of the assay (Fig. 7).
Also, at pH 8.0 the spectra displayed a persistent peak of the
prodrug (1) at 4.48 and a single peak at 2.55 corresponding to a
d 4.35; 4.50
In the current work also chemical stability studies on the
prodrug NUC-3073 (1) have been performed by exposing the
compound to human serum and to aqueous buffers (pH 1.0 and
8.0) using in situ 31P NMR. Each experiment has been carried out
dissolving the ProTide in the suitable deuterated solvent and
analyzing the samples at 37 8C for about 14 h, acquiring scans at
the regular time intervals. For a better resolution original spectra
(lower graphs) and deconvoluted ones (upper graphs) are reported.
The stability assay of the phosphoramidate NUC-3073 (1), after
incubation in human serum, showed 73% of unchanged compound
after 12 h and 30 min as shown in Fig. 6.
d
d
buffer peak (Fig. 8).
4. Discussion
Various mechanisms of tumour cell resistance towards
fluoropyrimidines such as 5-FU, 5-FdUrd and trifluorothymidine
(TFT) have been described, including a decreased activity of
crucial drug-activating enzymes (e.g. TK and orotate phosphor-
ibosyltransferase), an increased activity of drug-inactivating
enzymes (i.e. TP), an upregulation of the target enzyme (e.g. TS)
and/or defective facilitated diffusion of nucleosides [36–40].
Also, high TP levels found in several types of cancer tissue were
reported to be predictive for a poorer prognosis upon treatment
with fluoropyrimidines [41–43], although other studies have not
confirmed these findings [43,44]. The study presented here was
The spectra displayed a single peak inherent to the human
serum at ꢂ
which after 4 h and 15 min was partly hydrolyzed to the aminoacyl
phosphoramidate intermediate shown as a single peak at 7.20.
d
2.00 and the double peak of the parent at ꢂ
d 4.50
d
When chemical hydrolysis was evaluated at extreme
experimental conditions, i.e. at pH 1.0 and pH 8.0 at 37 8C, a
full stability of prodrug NUC-3073 (1) in both acidic and basic
buffer conditions was observed. Spectra were recorded for 14 h
acquiring scans every 12 min at regular intervals as shown in
Figs. 7 and 8. The ProTide (1) examined at pH 1.0 showed
aimed at the development of
a prodrug for 5-FdUrd, to
circumvent possible resistance mechanisms and susceptibility
to degradation by catabolic enzymes, present in the tumour
micro-environment.
Fig. 8. 31P NMR spectrum of compound NUC-3073 (1) in buffer pH 8.

J. Vande Voorde et al. / Biochemical Pharmacology 82 (2011) 441–452
451
We describe herein NUC-3073, a phosphoramidate prodrug of
5-FdUrd that may fulfil these aims. After uptake into the tumour
cells, NUC-3073 generates 5-FdUMP intracellularly upon enzy-
matic cleavage. Our stability studies and enzyme/serum studies by
31P NMR technology revealed that the prodrug NUC-3073 is fully
stable in acid and alkaline conditions, but subject to hydrolysis in
the presence of carboxypeptidase Y, or to some extent by serum
resulting in the formation of the nucleoside 50-phosphoramidate
derivative. Whereas TK is a key enzyme in the activation of 5-
FdUrd, NUC-3073 was found to be much less dependent on TK to
exert its cytostatic action in both murine (L1210) and human
(CEM) cell cultures. Due to the lipophilic nature of ProTides, these
molecules can deliver nucleoside-monophosphates directly into
the intact tumour cell after conversion to their nucleoside
phosphoramidate derivative by enzymes such as carboxyesterases
or carboxypeptidases (i.e. carboxypeptidase Y), eliminating the
need for an initial phosphorylation by specific nucleoside kinases
such as cytosolic TK-1. In this regard, NUC-3073 may be an
adequate tool for the treatment of tumour cells with a modified TK-
1 activity (either acquired or inherent). Also, since TK-1 expression
is S-phase-dependent, it is expected that NUC-3073 can also
efficiently deliver 5-FdUMP in tumour cells that are not in the S-
phase of their replication cycle. Our TS activity studies revealed
that NUC-3073 was able to inhibit TS in both wild-type and TK-
deficient tumour cell lines, pointing again to an efficient
intracellular delivery of the 50-monophosphate of 5-FdUrd, and
its virtual independence of cellular TK-1 for metabolic activation.
Normal tissue toxicity is always a concern in the development
of anticancer drugs and in particular it might be a potential concern
whether the 5-FdUrd prodrug may show an increased toxicity in
non-proliferating cells or other cell types with low TK-1 activity. In
such cell tissues, thymidine nucleotide pools are rather low, and
thus, the released 5-FdUMP derivative and the di- and triphosphate
drug metabolites may have a competitive advantage and exert a
more pronounced cytotoxic activity. However, it should be kept in
mind that in non-proliferating cells, TS activity is also markedly
decreased due to lack of ongoing DNA synthesis. Thus, due to a
lower degree of target enzyme (TS) activity in resting cells, it is
expected that the released 5-FdUMP should not necessarily afford
higher cell toxicity in such cells. It was indeed found that both 5-
FdUrd and the 5-FdUrd prodrug NUC-3073 were markedly less
toxic in confluent human cervix carcinoma HeLa/0 cell cultures
compared with the toxicity values observed for proliferating L1210
and CEM cell cultures [minimal cytotoxic concentration (MCC) of
these observations. Therefore, the administration of a TP-insensi-
tive anti-cancer prodrug such as NUC-3073, demonstrated to be
chemically stable at extreme pH conditions, may further improve
cancer chemotherapy.
It should be mentioned that the isopropylalaninyl monoamidate
phenyl monoester prodrug of tenofovir (GS 7340) has been
investigated for its pharmacokinetics in dogs [46]. This compound,
being rather similar to NUC-3073 in terms of the prodrug moiety,
had a half-life of 90 min in human plasma, and reached an oral
bioavailabilityof ꢂ17% after a single oral dose of GS 7340 (10 mg/kg)
to beagle dogs. There was an increased distribution of the parent
drug to tissues of lymphatic origin compared to the commercially
available prodrug tenofovir-DF (Viread) [46]. No increased side-
effectsfor GS7340 werereported whencompared withtheclinically
used Viread. These observations revealed that there may be a
potential clinical application for prodrugs such as NUC-3073.
In conclusion, we believe that ProTides, such as NUC-3073, may
represent an interesting new approach towards the development
of more resilient anti-cancer drugs. NUC-3073 may have at least
several advantages over its parent drug 5-FdUrd: it exerts its
cytostatic activity independently of TK and it is resistant to
metabolic breakdown by TP, an enzyme that is often upregulated
in tumours or may be externally expressed by mycoplasma
infection of the tumour tissue.
Acknowledgements
We wish to thank Ria Van Berwaer, Lizette van Berckelaer,
Kristien Minner and Christiane Callebaut for their excellent
technical assistance and Dr. Tarmo Roosild for generously
providing human uridine phosphorylase hUPP1. Johan Vande
Voorde acknowledges a PhD grant from the Institute for the
Promotion of Innovation through Science and Technology in
Flanders (IWT-Vlaanderen). This work is supported by a grant of
NuCana BioMed, the Concerted Actions of the K.U.Leuven (GOA 10/
014) and the ‘‘Fonds voor Wetenschappelijk Onderzoek (FWO)
Vlaanderen’’ (Credit n8 G. 0486.08).
References
[1] Hatse S, De Clercq E, Balzarini J. Role of antimetabolites of purine and
pyrimidine nucleotide metabolism in tumor cell differentiation. Biochem
Pharmacol 1999;58:539–55.
[2] Galmarini CM, Mackey JR, Dumontet C. Nucleoside analogues and nucleobases
in cancer treatment. Lancet Oncol 2002;3:415–24.
[3] de Bruin M, van Capel T, Smid K, van der Born K, Fukushima M, Hoekman K,
et al. Role of platelet derived endothelial cell growth factor/thymidine phos-
phorylase in fluoropyrimidine sensitivity and potential role of deoxyribose-1-
phosphate. Nucleosides Nucleotides Nucleic Acids 2004;23:1485–90.
[4] Ishikawa T, Utoh M, Sawada N, Nishida M, Fukase Y, Sekiguchi F, et al. Tumor
selective delivery of 5-fluorouracil by capecitabine, a new oral fluoropyrimi-
dine carbamate, in human cancer xenografts. Biochem Pharmacol
1998;55:1091–7.
NUC-3073 for HeLa: 2
mM]. Also, NUC-3073 was not found more
toxic in confluent HeLa/TKꢀ cells than in wild-type HeLa/0 cells
(data not shown). Thus, evidence for increased toxicity of a prodrug
such as NUC-3073 against non-proliferating tissue was not
observed.
We also proved that NUC-3073 is unlikely to be inactivated by
catabolic enzymes involved in nucleoside metabolism. Indeed,
whereas 5-FdUrd is highly susceptible to enzymatic hydrolysis by
TP resulting in the formation of 5-FU and 20-deoxyribose-1-
phosphate, we found that its prodrug NUC-3073 is not a substrate
for prokaryotic (i.e. E. coli) or mammalian (i.e. human erythrocyte)
TP. Also, UP (derived from human tumour tissue) does not
recognize NUC-3073 as a substrate, whereas 5-FdUrd is (poorly,
but measurably) hydrolyzed by this enzyme. Several studies
revealed that many tumour cells have elevated levels of TP, which
also acts as an angiogenic factor [43,45]. Moreover, there are
several reports on the preferential colonization of tumour tissue by
mycoplasmas [15–17] which interfere with the cytostatic activity
of several conventional chemotherapeutics in vitro through its
encoded TP [27–29]. Our findings that 5-FdUrd, but not NUC-3073,
markedly loses cytostatic activity when the tumour cells are
infected by (TP-expressing) mycoplasmas, is in full agreement with
[5] Walko CM, Lindley C. Capecitabine: a review. Clin Ther 2005;27:23–44.
[6] Beck A, Etienne MC, Cheradame S, Fischel JL, Formento P, Renee N, et al. A role
for dihydropyrimidine dehydrogenase and thymidylate synthase in tumour
sensitivity to fluorouracil. Eur J Cancer 1994;30A:1517–22.
[7] Tanaka F, Fukuse T, Wada H, Fukushima M. The history, mechanism and
clinical use of oral 5-fluorouracil derivative chemotherapeutic agents. Curr
Pharm Biotechnol 2000;1:137–64.
[8] Longley DB, Harkin DP, Johnston PG. 5-Fluorouracil: mechanisms of action and
clinical strategies. Nat Rev Cancer 2003;3:330–8.
[9] Mehellou Y, Balzarini J, McGuigan C. Aryloxy phosphoramidate triesters: a
technology for delivering monophosphorylated nucleosides and sugars into
cells. ChemMedChem 2009;4:1779–91.
[10] Hecker SJ, Erion MD. Prodrugs of phosphates and phosphonates. J Med Chem
2008;51:2328–45.
[11] Harris SA, McGuigan C, Andrei G, Snoeck R, De Clercq E, Balzarini J. Synthesis
and antiviral evaluation of phosphoramidate derivatives of (E)-5-(2-bromo-
vinyl)-20-deoxyuridine. Antivir Chem Chemother 2001;12:293–300.
[12] Congiatu C, Brancale A, Mason MD, Jiang WG, McGuigan C. Novel potential
anticancer naphthyl phosphoramidates of BVdU: separation of diastereoi-
somers and assignment of the absolute configuration of the phosphorus
center. J Med Chem 2006;49:452–5.

452
J. Vande Voorde et al. / Biochemical Pharmacology 82 (2011) 441–452
[13] Derudas M, Carta D, Brancale A, Vanpouille C, Lisco A, Margolis L, et al. The
application of phosphoramidate protide technology to acyclovir confers anti-
HIV inhibition. J Med Chem 2009;52:5520–30.
[14] Razin S, Yogev D, Naot Y. Molecular biology and pathogenicity of mycoplas-
mas. Microbiol Mol Biol Rev 1998;62:1094–156.
and may act by the generation of a novel intracellular metabolite. J Med Chem
1996;39:1748–53.
[32] McGuigan C, Tsang HW, Cahard D, Turner K, Velazquez S, Salgado A, et al.
Phosphoramidate derivatives of d4T as inhibitors of HIV: the effect of amino
acid variation. Antiviral Res 1997;35:195–204.
[15] Huang S, Li JY, Wu J, Meng L, Shou CC. Mycoplasma infections and different
human carcinomas. World J Gastroenterol 2001;7:266–9.
[16] Pehlivan M, Itirli G, Onay H, Bulut H, Koyuncuoglu M, Pehlivan S. Does
Mycoplasma sp. play role in small cell lung cancer? Lung Cancer
2004;45:129–30.
[17] Pehlivan M, Pehlivan S, Onay H, Koyuncuoglu M, Kirkali Z. Can mycoplasma-
mediated oncogenesis be responsible for formation of conventional renal cell
carcinoma? Urology 2005;65:411–4.
[18] Chan PJ, Seraj IM, Kalugdan TH, King A. Prevalence of mycoplasma conserved
DNA in malignant ovarian cancer detected using sensitive PCR-ELISA. Gynecol
Oncol 1996;63:258–60.
[33] Balzarini J, De Clercq E, Torrence PF, Mertes MP, Park JS, Schmidt CL, et al. Role
of thymidine kinase in the inhibitory activity of 5-substituted-20-deoxyur-
idines on the growth of human and murine tumor cell lines. Biochem Phar-
macol 1982;31:1089–95.
[34] Balzarini J, De Clercq E. Strategies for the measurement of the inhibitory effects
of thymidine analogs on the activity of thymidylate synthase in intact murine
leukemia L1210 cells. Biochim Biophys Acta 1984;785:36–45.
[35] Mehellou Y, Valente R, Mottram H, Walsby E, Mills KI, Balzarini J, et al.
Phosphoramidates of 20-beta-
D-arabinouridine (AraU) as phosphate prodrugs;
design, synthesis, in vitro activity and metabolism. Bioorg Med Chem
2010;18:2439–46.
[19] Sasaki H, Igaki H, Ishizuka T, Kogoma Y, Sugimura T, Terada M. Presence of
Streptococcus DNA sequence in surgical specimens of gastric cancer. Jpn J
Cancer Res 1995;86:791–4.
[36] Agarwal RP, Han T, Fernandez M. Collateral resistance of a dideoxycytidine-
resistant cell line to 5-fluoro-20-deoxyuridine. Biochem Biophys Res Commun
1999;262:657–60.
[20] Yang H, Qu L, Ma H, Chen L, Liu W, Liu C, et al. Mycoplasma hyorhinis infection in
gastric carcinoma and its effects on the malignant phenotypes of gastric cancer
cells. BMC Gastroenterol 2010;10:132.
[37] Murakami Y, Kazuno H, Emura T, Tsujimoto H, Suzuki N, Fukushima M.
Different mechanisms of acquired resistance to fluorinated pyrimidines in
human colorectal cancer cells. Int J Oncol 2000;17:277–83.
[21] Kidder M, Chan PJ, Seraj IM, Patton WC, King A. Assessment of archived
paraffin-embedded cervical condyloma tissues for mycoplasma-conserved
DNA using sensitive PCR-ELISA. Gynecol Oncol 1998;71:254–7.
[22] Razin S. The mycoplasmas. Microbiol Rev 1978;42:414–70.
[23] Charron J, Langelier Y. Analysis of deoxycytidine (dC) deaminase activity in
herpes simplex virus-infected or HSV TK-transformed cells: association with
[38] Kosaka T, Usami K, Ueshige N, Sugaya J, Nakano Y, Takashima S. Effects of
thymidine phosphorylase levels in cancer, background mucosa, and regional
lymph nodes on survival of advanced gastric cancer patients receiving post-
operative fluoropyrimidine therapy. Oncol Rep 2004;12:1279–86.
[39] Zhang ZG, Harstrick A, Rustum YM. Mechanisms of resistance to fluoropyr-
imidines. Semin Oncol 1992;19(2 Suppl. 3):4–9.
mycoplasma contamination but not with virus infection.
1981;57:245–50.
[24] Neale GA, Mitchell A, Finch LR. Enzymes of pyrimidine deoxyribonucleotide
J
Gen Virol
[40] Sobrero AF, Moir RD, Bertino JR, Handschumacher RE. Defective facilitated
diffusion of nucleosides, a primary mechanism of resistance to 5-fluoro-20-
deoxyuridine in the HCT-8 human carcinoma line. Cancer Res 1985;45:
3155–60.
metabolism in Mycoplasma mycoides subsp. mycoides.
J
Bacteriol
1983;156:1001–5.
[41] Kamoshida S, Shiogama K, Shimomura R, Inada K, Sakurai Y, Ochiai M, et al.
Immunohistochemical demonstration of fluoropyrimidine-metabolizing
enzymes in various types of cancer. Oncol Rep 2005;14:1223–30.
[42] Ciaparrone M, Quirino M, Schinzari G, Zannoni G, Corsi DC, Vecchio FM, et al.
Predictive role of thymidylate synthase, dihydropyrimidine dehydrogenase
and thymidine phosphorylase expression in colorectal cancer patients receiv-
ing adjuvant 5-fluorouracil. Oncology 2006;70:366–77.
[25] Tham TN, Ferris S, Kovacic R, Montagnier L, Blanchard A. Identification of
Mycoplasma pirum genes involved in the salvage pathways for nucleosides. J
Bacteriol 1993;175:5281–5.
[26] Sinigaglia F, Talmadge KW. Inhibition of [3H]thymidine incorporation by
Mycoplasma arginini-infected cells due to enzymatic cleavage of the nucleo-
side. Eur J Immunol 1985;15:692–6.
[27] Bronckaers A, Balzarini J, Liekens S. The cytostatic activity of pyrimidine
nucleosides is strongly modulated by Mycoplasma hyorhinis infection: impli-
cations for cancer therapy. Biochem Pharmacol 2008;76:188–97.
[43] Koopman M, Venderbosch S, Nagtegaal ID, van Krieken JH, Punt CJ. A review on
the use of molecular markers of cytotoxic therapy for colorectal cancer, what
have we learned? Eur J Cancer 2009;45:1935–49.
´
[28] Jette L, Bissoon-Haqqani S, Le Franc¸ois B, Maroun JA, Birnboim HC. Resistance
[44] Ciccolini J, Evrard A, Cuq P. Thymidine phosphorylase and fluoropyrimidines
of colorectal cancer cells to 5-FUdR and 5-FU caused by Mycoplasma infection.
Anticancer Res 2008;28:2175–80.
efficacy: a Jekyll and Hyde story. Curr Med Chem Anticancer Agents
2004;4:71–81.
[29] Liekens S, Bronckaers A, Balzarini J. Improvement of purine and pyrimidine
antimetabolite-based anticancer treatment by selective suppression of myco-
plasma-encoded catabolic enzymes. Lancet Oncol 2009;10:628–35.
[30] McGuigan C, Pathirana RN, Balzarini J, De Clercq E. Intracellular delivery of
bioactive AZT nucleotides by aryl phosphate derivatives of AZT. J Med Chem
1993;36:1048–52.
[45] Bronckaers A, Gago F, Balzarini J, Liekens S. The dual role of thymidine
phosphorylase in cancer development and chemotherapy. Med Res Rev
2009;29:903–53.
[46] Lee WA, He GX, Eisenberg E, Cihlar T, Swaminathan S, Mulato A, et al. Selective
intracellular activation of a novel prodrug of the human immunodeficiency
virus reverse transcriptase inhibitor tenofovir leads to preferential distribu-
tion and accumulation in lymphatic tissue. Antimicrob Agents Chemother
2005;49:1898–906.
[31] McGuigan C, Cahard D, Sheeka HM, De Clercq E, Balzarini J. Aryl phosphor-
amidate derivatives of d4T have improved anti-HIV efficacy in tissue culture