L-Aspartic and l-glutamic acid ester-based ProTides of anticancer nucleosides: Synthesis and antitumoral evaluation

Article abstract of DOI:10.1016/j.bmcl.2016.03.076

A series of novel aryloxyphosphoramidate nucleoside prodrugs based on l-aspartic acid and l-glutamic acid as amino acid motif has been synthesized and evaluated for antitumoral activity. Depending on the cancer cell line studied and on the nature of the parent nucleoside compound (gemcitabine, 5-iodo-2′-deoxy-uridine, floxuridine or brivudin), the corresponding ProTides are endowed with an improved or decreased cytotoxic activity.

Full text of DOI:10.1016/j.bmcl.2016.03.076

Bioorganic & Medicinal Chemistry Letters 26 (2016) 2142–2146
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
L-Aspartic and L-glutamic acid ester-based ProTides of anticancer
nucleosides: Synthesis and antitumoral evaluation q
Ling-Jie Gao a, Steven De Jonghe a, Dirk Daelemans b, Piet Herdewijn a,
⇑
a KU Leuven, Rega Institute for Medical Research, Laboratory of Medicinal Chemistry, Minderbroedersstraat 10, 3000 Leuven, Belgium
b KU Leuven, Rega Institute for Medical Research, Laboratory of Virology and Chemotherapy, Minderbroedersstraat 10, 3000 Leuven, Belgium
a r t i c l e i n f o
a b s t r a c t
Article history:
A series of novel aryloxyphosphoramidate nucleoside prodrugs based on L-aspartic acid and L-glutamic
Received 1 March 2016
Revised 18 March 2016
Accepted 20 March 2016
Available online 22 March 2016
acid as amino acid motif has been synthesized and evaluated for antitumoral activity. Depending on
the cancer cell line studied and on the nature of the parent nucleoside compound (gemcitabine, 5-
iodo-20-deoxy-uridine, floxuridine or brivudin), the corresponding ProTides are endowed with an
improved or decreased cytotoxic activity.
Ó 2016 Elsevier Ltd. All rights reserved.
Keywords:
ProTides
Nucleosides
Prodrugs
Antitumoral
Nucleoside analogues represent important drugs for the treat-
ment of different cancers. Currently approved antitumoral nucle-
oside analogues include Cytarabine, Fludarabine, Cladribine,
Gemcitabine, Clofarabine, Capecitabine, Floxuridine, Deoxyco-
formycin, Azacitidine and Decitabine. These agents have been
licensed for a variety of oncology indications, including leukaemia,
lung cancer, pancreatic cancer and colon cancer.1 The mechanism
of action for most of these compounds is very similar. They are
intracellularly converted to their respective nucleotide analogues,
which inhibit DNA synthesis by inhibition of DNA polymerase
and/or ribonucleotide reductase. Decitabine is incorporated into
DNA as its triphosphate metabolite, resulting in inhibition of
DNA methylation. The cytotoxic activity of azacitidine is mainly
due to its conversion to deoxynucleotides via ribonucleotide
reductase and its incorporation into DNA. Floxuridine is intracellu-
larly phosphorylated to its active form, inhibiting thymidylate syn-
thase. Deoxycoformycin (pentostatin) is the only antimetabolite
active without metabolism. It acts as a potent inhibitor of adeno-
sine deaminase.2
first step of phosphorylation represents the rate-limiting step of
the bioactivation and its inefficiency may limit the therapeutic
potential of the nucleoside analogues. Another reason for the
decreased activity of the nucleoside analogue is the limited uptake
by the tumor cells, due to a decreased expression of nucleoside
transporter proteins.
Various prodrug or ‘pronucleotide’ approaches have been
devised and investigated in order to tackle these issues.5 In general,
the goal of these approaches has been to promote passive diffusion
through the lipophilic cell membranes and to release the parent
nucleotide intracellularly, where it can be further phosphorylated
to the pharmacologically active species. One of the most promising
approaches are the aryloxyphosphoramidates (also known as the
ProTides).6 Originally, this approach was focused on improving
the biological activity of antiviral agents. The ultimate success
came with the marketing of Sofosbuvir, which is currently used
for the treatment of HCV infected patients.7 Aryloxyphosphono-
amidates, such as GS-9131,8 moved into clinical trials to be evalu-
ated as anti-HIV medication. GS-7340 (also known as tenofovir
alafenamide) is an example of another phosphono-amidate pro-
drug that received recently marketing approval as anti-HIV drug.9
More recently, the aryloxyphosphoramidate prodrug strategy
has also been applied to antitumoral nucleosides. Phosphorami-
date ProTides of the anticancer agents gemcitabine,3 floxuridine,10
cladribine,11 and brivudine12 have been synthesized and evaluated
for antitumoral activity. NUC-1031 (Fig. 1) is the phosphoramidate
prodrug of gemcitabine, which is currently undergoing
clinical trials in patients with solid tumors. Thymectacin, an
A common problem in the treatment of cancer patients is the
resistance to nucleoside analogues.3,4 Resistance can be caused
by poor conversion of the parent nucleoside to its active nucleoside
monophosphate, diphosphate and triphosphate. In most cases, the
q
In memory of Professor Christopher McGuigan.
⇑
Corresponding author. Tel.: +32 16 33 73 87.
E-mail address: piet.herdewijn@kuleuven.be (P. Herdewijn).
http://dx.doi.org/10.1016/j.bmcl.2016.03.076
0960-894X/Ó 2016 Elsevier Ltd. All rights reserved.

L.-J. Gao et al. / Bioorg. Med. Chem. Lett. 26 (2016) 2142–2146
2143
Figure 1. Known ProTides.
L
-aspartic acid 1a using thionyl chloride and isoamyl alcohol,
according to a known method.14 The corresponding diisoamyl
glutamic acid 3b, was prepared similarly in a good yield, starting
from -glutamic acid 1b. The asymmetric esters of -aspartic acid
(compounds 3c and 3d) were prepared from the commercially
available Boc-protected -aspartic acid analogues 2a and 2b in a
two step procedure. Reaction of the carboxylic acid moiety with
isoamyl alcohol using N,N,N0,N0-tetramethyl-O-(6-chloro-1H-ben-
zotriazol-1-yl)uronium hexafluorophosphate (HCTU) as coupling
agent yielded the corresponding ester derivative. Acidic cleavage
aryloxyphosphoramidate prodrug of brivudine, is currently
undergoing clinical trials in colon cancer.
L
-
In all phosphoramidate ProTides studied so far (antiviral, as
L
L
well as, antitumoral congeners),
L-alanine is the preferred amino
acid motif. Other amino acids, such as the
L
-aspartic acid dimethy-
L
lester have been evaluated and were found to be less active and
have therefore not been further pursued.13 However, we recently
reported the synthesis and antiviral evaluation of phosphorami-
date prodrugs of 20-C-Me-cytidine, 20-C-Me-uridine and 20-C-Me-
20-F-uridine using -aspartic acid as amino acid.14 These com-
L
of the Boc group afforded the desired di-esterified L-aspartic acid
derivatives (3c–d), as their hydrochloride salts.
pounds were endowed with very potent antiviral activity when
tested against the hepatitis C virus. As their activity is higher than
The ProTides of 5-Iodo-20-deoxy-uridine (IDU), Floxuridine (20-
deoxy-5-fluoro-uridine, FDU) and Brivudin ((E)-5-(2-bromovinyl)-
20-deoxyuridine, BVDU) were prepared from the unprotected, par-
ent nucleosides 4–6 (Scheme 2). The key reagent 7 was prepared
in situ by reacting phenyl phosphorodichloridate with the esteri-
the classical
tic acid should be considered as amino acid alternative in prodrug
design. These findings prompted us to synthesize a series of
L-alanine containing ProTides, it suggests that L-aspar-
L
-
aspartic acid phosphoramidate analogues as potential prodrugs
of antitumoral nucleosides.
fied L-aspartic acid 3a, after which the nucleosides 4–6 were cou-
Although different methods are available for the synthesis of
the aryloxyphosphoramidate prodrugs, we selected the classical
procedure in which the (protected) nucleoside is coupled with a
suitable chlorophosphoramidate reagent, which is activated as its
imidazolium intermediate using N-methylimidazole (NMI).15
The synthesis of the amino acid ester building blocks is shown
pled in the presence of N-methylimidazole (NMI), in a one pot
procedure, according to a known method.16 Literature precedents
suggest that these reactions preferentially take place in dichloro-
methane. However, in our hands, the reaction did not work well,
due to the poor solubility of the reactants in dichloromethane.
The problem could be easily resolved by running the reaction in
in Scheme 1. Diisoamyl L-aspartic acid 3a was prepared from
Scheme 1. Reagents and conditions: (a) SOCl2, isoamyl alcohol, 0 °C to rt, 12 h, then reflux, 3 h, 80%; (b) HCTU, isoamyl alcohol, Et3N, CH2Cl2, rt, 4 h, 95%; (c) 5 N HCl in i-PrOH,
CH2Cl2, rt, 2 h, 95%.

2144
L.-J. Gao et al. / Bioorg. Med. Chem. Lett. 26 (2016) 2142–2146
Scheme 2. Reagents and conditions: (a) phenyl dichlorophosphate, N-methylimidazole, THF, ꢀ40 °C to rt, 12 h; (b) ꢀ40 °C to rt, 6 h, 46–50%.
THF. A major byproduct was the 30,50-biphosphorylated product,
which could be easily separated from the desired products 8–10
by column chromatography on silica gel. Although the yield was
only moderate (40–50%), at this stage of the research program, this
procedure was still acceptable, as only small amounts of material
were needed for a first in vitro screening.
glioblastoma (U87), as well as hematological tumors such as T-
ALL (Jurkat) and Burkitt’s lymphoma (Ramos), was selected for bio-
logical profiling of the antitumoral prodrugs (Table 1).
In a first round of synthesis, di-isoamyl esters of L-aspartic acid
were used as amino acid moiety in the aryloxyphosphoramidate
part of the ProTides. The choice of these ester moieties is based
on the fact that these ester groups were found to be optimal to con-
fer antiviral activity, as reported by us earlier.14 Although IDU
(compound 4) lacked completely a cytotoxic effect against the dif-
ferent tested cell lines, its prodrug (compound 8) is endowed with
an increased cytotoxicity in all cell lines, when compared to the
parent nucleoside 4. The same pattern is observed for the prodrugs
of BVDU. Whereas BVDU (compound 6) lacks activity against this
panel of tumor cell lines, the corresponding ProTides do show an
improved activity against the different cell lines, when compared
to the parent nucleoside. Noteworthy to mention is that the BVDU
prodrug 10 is the most potent congener of all compounds evalu-
ated in the A549 cell line.
Floxuridine 5 is extremely cytotoxic when tested against the
Jurkat and Ramos cell line, with EC50 values in the low nM range.
The prodrug 9 was significantly reduced in cytotoxic activity
against these two cell lines (ꢁ 300-fold). Only moderate or very
weak activity was observed for 5 in the A549, PC-3 and U-87
cancer cell line. The FDU prodrug 9 showed a better efficacy
against these three cell lines, indicating that poor phosphoryla-
tion might be the cause of the poor biological effect of the parent
nucleoside.
For the synthesis of gemcitabine ProTides, the 30-OH group of
gemcitabine 11 was selectively protected by reaction with di-
tert-butyl carbonate in a mixture of dioxane and water, affording
the 30-Boc-protected gemcitabine 12 in good yield (Scheme 3).17
The phenyl aminoacyl phosphorochloridates active intermedi-
ates 13a–e were prepared in situ by reacting phenyl dichlorophos-
phate with the esters of amino acids 3a–e in presence of N-
methylimidazole (NMI) in anhydrous dichloromethane. Then, the
30-protected gemcitabine (12) was added to the unpurified inter-
mediates at ꢀ40 °C and slowly warmed to ambient temperature.
The 30-Boc-protected gemcitabine phosphoramidates 14a–e were
isolated in good yields (75–85%). Finally, acidic deprotection of
the Boc group afforded the final products 15a–e in excellent yields.
The final aryloxyphosphoramidate prodrugs were isolated as a
diastereomeric mixture due to the chirality of the phosphorus
atom and were tested as such.
Biological evaluation and SAR studies: In order to evaluate L-
aspartic acid as amino acid motif in antitumoral prodrug design,
4 known antitumoral nucleosides were selected. Gemcitabine
(20,20-difluoro-20-deoxycytidine, compound 11) is licensed for the
treatment of pancreatic cancer, non-small lung cancer, bladder
cancer and breast cancer.18 5-Iodo-20-deoxy-uridine (IDU, com-
pound 4) is a well-known antiviral drug used topically for the
treatment of herpes simplex infections, but it has also been
reported to be endowed with antitumoral activity.19 Floxuridine
(20-deoxy-5-fluoro-uridine, compound 5) is often used when col-
orectal cancer is advanced and has metastasized to the liver.20 Bri-
vudin ((E)-5-(2-bromovinyl)-20-deoxyuridine, compound 6) is
highly potent and selective inhibitor of herpes simplex and vari-
cella-zoster viral infections. Brivudine-monophosphate is a potent
inhibitor of thymidylate synthase, which explains its antitumoral
effects.21 A panel of five cancer cell lines, representing solid tumor
types including prostate cancer (PC-3), lung cancer (A549), and
Gemcitabine (compound 11) displayed strong cytotoxic activity
against the Jurkat and PC-3 cell line (EC50 values in the low nM
range), and especially the Ramos cell line is extremely sensitive
to gemcitabine with an EC50 value in the picomolar range. The L-
aspartic acid diisoamyl prodrug 15a showed a decreased activity
in these three cell lines: a 10-fold drop in potency in the Jurkat cell
line, a 50-fold decreased activity in the PC-3 cells, and a 9000-fold
drop in cytotoxicity for the Ramos cell line. A two-fold boost in
activity was observed for prodrug 15a versus the U87 cell line. In
contrast, for the A549 cell line, there was no beneficial effect of
the prodrug when compared to gemcitabine. A number of ana-
logues was also prepared. Instead of an iso-amyl ester, we opted

L.-J. Gao et al. / Bioorg. Med. Chem. Lett. 26 (2016) 2142–2146
2145
Scheme 3. Reagents and conditions: (a) phenyl dichlorophosphate, N-methylimidazole, THF, ꢀ40 °C to rt, 12 h; (b) di-tert-butyl dicarbonate, Na2CO3, 1,4-dioxane/water
(4:1), rt, 72 h, 88%; (c) 3a–e, phenyl dichlorophosphate, N-methylimidazole, ꢀ40 °C to rt, 12 h. then 12, ꢀ40 °C to rt, 6 h, 75–86%; (d) trifluoroacetic acid/CH2Cl2 (1/1), rt, 84–
91%.
Table 1
to the L-aspartic acid based prodrugs. Although we observed before
Cytotoxic activity of nucleosides and their prodrugs against tumor cell lines
a clear correlation between lipophilicity and antiviral activity (i.e.
the more lipophilic esters were endowed with potent antiviral
activity),14 lipophilicity is not a determining factor in the antitu-
moral activity of the ProTides synthesized here.
In summary, a novel series of phosphoramidate nucleoside pro-
drugs based on L-aspartic acid and L-glutamic acid as amino acid
motif has been synthesized and evaluated for their cytotoxic activ-
ity against a panel of cancer cell lines. When the parent nucleoside
(e.g. IDU, BVDU and FDU in some cell lines) displays only weak to
moderate activity, the corresponding ProTides do show an
improved cytotoxic activity. It suggests that poor phosphorylation
of the parent nucleoside analogue in these tumor cell lines is the
reason for their weak antitumoral activity. FDU and gemcitabine
are very potent in the Jurkat and Ramos cell line. In addition, gem-
citabine is endowed with very strong cytotoxicity in the prostate
cancer cell line. In these cases, the aryloxyphosphoramidate pro-
drugs show a pronounced decreased cytotoxicity, when compared
to the parent nucleoside. It indicates that, in these cell lines, the
intracellular conversion of the phosphoramidate prodrugs to the
nucleoside monophosphate is problematic.
a
Compound
EC50
(lM)
Jurkat
A549
PC-3
Ramos
U87
4 (IDU)
8
5 (FDU)
9
6 (BVDU)
10
11 (gemcitabine)
15a
15b
15c
15d
15e
79.7
14.7
0.00333
1.1
>100
5.2
0.023
0.294
1.2
0.210
0.209
0.129
>100
27.0
60.8
43.4
>100
9.9
>100
>100
>100
>100
>100
>100
>100
33.3
86.9
22.7
>100
12.5
0.065
3.5
10.9
3.6
47.9
13.6
>100
38.0
18.2
10.1
>100
12.8
>100
43.3
>100
>100
43.7
>100
0.00751
0.724
>100
6.3
0.00003
0.294
1.2
0.813
0.700
0.365
3.8
2.3
a
EC50 is the compound concentration required to inhibit tumor cell viability by
50%.
for benzyl esters, as it’s well known that the introduction of
benzylesters leads to an enhancement in the antiviral potency of
dideoxynucleoside phosphoramidates and to a more rapid esterase
mediated hydrolysis. Changing of the isoamyl moiety into a benzyl
ester afforded compounds 15c and 15d, which roughly spoken
have the same level of cytotoxicity and antitumoral profile as the
Supplementary data
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.bmcl.2016.03.
076.
L
-aspartic acid diisoamyl ProTide 15a. Also two ProTides (com-
pounds 15b and 15e) bearing a -glutamic acid as amino acid motif
were prepared. Here again, the antitumoral activity is very similar
L

2146
L.-J. Gao et al. / Bioorg. Med. Chem. Lett. 26 (2016) 2142–2146
10. McGuigan, C.; Murziani, P.; Slusarczyk, M.; Gonczy, B.; Vande Voorde, J.;
References and notes
Liekens, S.; Balzarini, J. J. Med. Chem. 2011, 54, 7247.
11. Walsby, E.; Congiatu, C.; Schwappach, A.; Walsh, V.; Burnett, A. K.; McGuigan,
C.; Mills, K. I. Blood 2005, 106, 941A. Abstract 3369.
12. Congiatu, C.; Brancale, A.; Mason, M. D.; Jiang, W. G.; McGuigan, C. J. Med.
Chem. 2006, 49, 452.
13. Cahard, D.; McGuigan, C.; Balzarini, J. Mini-Rev. Med. Chem. 2004, 4, 371.
14. Maiti, M.; Maiti, M.; Rozenski, J.; De Jonghe, S.; Herdewijn, P. Org. Biomol. Chem.
2015, 13, 5158.
15. Pradere, U.; Garnier-Amblard, E. C.; Coats, S. J.; Amblard, F.; Schinazi, R. F.
Chem. Rev. 2014, 114, 9154.
16. Lehsten, D. M.; Baehr, D. N.; Lobl, T. J.; Vaino, A. R. Org. Process Res. Dev. 2002, 6,
819.
17. Guo, Z.; Gallo, J. M. J. Org. Chem. 1999, 64, 8319.
18. Noble, S.; Goa, K. L. Drugs 1997, 54, 447.
19. Prusoff, W. H. Cancer Res. 1960, 20, 92.
20. Power, D. G.; Kemeny, N. E. Mol. Cancer Ther. 2009, 8, 1015.
21. De Clercq, E. Med. Res. Rev. 2005, 25, 1.
1. Jordheim, L. P.; Durantel, D.; Zoulim, F.; Dumontet, C. Nat. Rev. Drug Disc. 2013,
12, 447.
2. Parker, W. B. Chem. Rev. 2009, 109, 2880.
3. Slusarczyk, M.; Lopez, M. H.; Balzarini, J.; Mason, M.; Jiang, W. G.; Blagden, S.;
Thompson, E.; Ghazaly, E.; McGuigan, C. J. Med. Chem. 2014, 57, 1531.
4. Galmarini, C. M.; Mackey, J. R.; Dumontet, C. Leukemia 2001, 15, 875.
5. Wagner, C. R.; Iyer, V. V.; McIntee, E. J. Med. Res. Rev. 2000, 20, 417.
6. Mehellou, Y.; Balzarini, J.; McGuigan, C. ChemMedChem 2009, 4, 1779.
7. Sofia, M. J.; Bao, D.; Chang, W.; Du, J.; Nagarathnam, D.; Rachakonda, S.; Reddy,
G.; Ross, B. S.; Wang, P.; Zhang, H.; Bansal, S.; Espiritu, C.; Keilman, M.;
Micolochick Steuer, H. M.; Niu, C.; Otto, M. J.; Furman, P. A. J. Med. Chem. 2010,
53, 7202.
8. Mackman, R. L.; Ray, A. S.; Hui, H. C.; Zhang, L.; Birkus, G.; Boojamra, C. G.;
Desai, M. C.; Douglas, J. L.; Gao, Y.; Grant, D.; Laflamme, G.; Lin, K. Y.;
Markevitch, D. Y.; Mishra, R.; McDermott, M.; Pakdaman, R.; Petrakovsky, O. V.;
Vela, J. E.; Cihlar, T. Bioorg. Med. Chem. 2010, 15, 3606.
9. Ray, A. S.; Fordyce, M. W.; Hitchcock, M. J. Antiviral Res. 2016, 125, 63.