Synthesis and in vitro anticancer activity of new gemcitabine-nucleoside analogue dimers containing methyltriazole or ester-methyltriazole linker

Article abstract of DOI:10.1016/j.bmcl.2019.08.003

Two series of novel gemcitabine-nucleoside analogue dimers were synthesized using the ‘click’ chemistry approach. In the first series of dimers (21–30), the nucleoside units were connected with a stable methyltriazole 4N-3′(or 5′)C linker whereas in the second series (31–40) with a cleavable ester-methyltriazole 4N-3′(or 5′)C linker. Dimers 21–40 were evaluated for their cytotoxic activity in five human cancer cell lines such as cervical (HeLa), nasopharyngeal (KB), lung (A549), brain (U87), liver (HepG2) and normal dermal fibroblast cell line (HDF) using the sulforhodamine B (SRB) assay. Compound 29 comprising two gemcitabine (dFdC) units exhibited the highest activity among dimers 21–30. The activity of compound 29 was higher than that of dFdC in all the studied cancer cell lines. A similar order of activity was observed for compounds 25, 28, and 30. The best activity among all the dimers synthesized was displayed by compound 39, comprising two gemcitabine units with a cleavable linker. The activity of compound 39 was 5 to 9 times higher than that of dFdC, depending on the cell line. In addition, marked cytotoxic activity was shown by compounds 31, 36, 38, and 40.

Full text of DOI:10.1016/j.bmcl.2019.08.003

Journal Pre-Proof
Synthesis and in vitro anticancer activity of new gemcitabine-nucleoside ana-
logue dimers containing methyltriazole or ester-methyltriazole linker
Roksana Trznadel, Aleksandra Singh, Natalia Kleczewska, Joanna Liberska,
Piotr Ruszkowski, Lech Celewicz
PII:
S0960-894X(19)30529-3
https://doi.org/10.1016/j.bmcl.2019.08.003
BMCL 26599
DOI:
Reference:
Bioorganic & Medicinal Chemistry Letters
To appear in:
Received Date:
Revised Date:
Accepted Date:
10 June 2019
31 July 2019
1 August 2019
Please cite this article as: Trznadel, R., Singh, A., Kleczewska, N., Liberska, J., Ruszkowski, P., Celewicz, L.,
Synthesis and in vitro anticancer activity of new gemcitabine-nucleoside analogue dimers containing methyltriazole
or ester-methyltriazole linker, Bioorganic & Medicinal Chemistry Letters (2019), doi: https://doi.org/10.1016/
j.bmcl.2019.08.003
This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover
page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will
undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing
this version to give early visibility of the article. Please note that, during the production process, errors may be
discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
© 2019 Published by Elsevier Ltd.

JOURNAL PRE-PROOF
Synthesis and in vitro anticancer activity of new gemcitabine-nucleoside analogue dimers
containing methyltriazole or ester-methyltriazole linker
Roksana Trznadela, Aleksandra Singha, Natalia Kleczewskaa, Joanna Liberskab,
Piotr Ruszkowskib, Lech Celewicza,*
aFaculty of Chemistry, Adam Mickiewicz University, Uniwersytetu Poznańskiego St 8, 61-614
Poznań, Poland
bDepartment of Pharmacology, Poznań University of Medical Sciences, Rokietnicka St 5 a,
60-806 Poznań, Poland
* Corresponding author. Tel.: +48 61 829 1690; fax: +48 61 829 1555. E-mail address:
celewicz@amu.edu.pl (L. Celewicz).
Abstract
Two series of novel gemcitabine-nucleoside analogue dimers were synthesized using the
‘click’ chemistry approach. In the first series of dimers (21–30), the nucleoside units were
connected with a stable methyltriazole 4N-3’(or 5’)C linker whereas in the second series (31–
40) with a cleavable ester-methyltriazole 4N-3’(or 5’)C linker. Dimers 21–40 were evaluated
for their cytotoxic activity in five human cancer cell lines such as cervical (HeLa),
nasopharyngeal (KB), lung (A549), brain (U87), liver (HepG2) and normal dermal fibroblast
cell line (HDF) using the sulforhodamine B (SRB) assay. Compound 29 comprising two
gemcitabine (dFdC) units exhibited the highest activity among dimers 21–30. The activity of
compound 29 was higher than that of dFdC in all the studied cancer cell lines. A similar order
of activity was observed for compounds 25, 28, and 30. The best activity among all the dimers
synthesized was displayed by compound 39, comprising two gemcitabine units with a
cleavable linker. The activity of compound 39 was 5 to 9 times higher than that of dFdC,
depending on the cell line. In addition, marked cytotoxic activity was shown by compounds
31, 36, 38, and 40.

JOURNAL PRE-PROOF
Keywords: Gemcitabine, 5-fluoro-2’-deoxyuridine, 3’-azido-3’-deoxythymidine, nucleoside
dimers, ‘click’ chemistry, cytotoxic activity, human cancer cell lines: HeLa, KB, A549, U87,
HepG2
A number of pyrimidine and purine derivatives, including nucleoside analogues, have
found widespread use as antiviral1 and anticancer2 therapeutics. In particular, gemcitabine
(2’,2’-difluoro-2’-deoxycytidine, dFdC, 1), a pyrimidine nucleoside analogue, is widely used
as the hydrochloride salt to treat pancreatic, non-small cell lung, ovarian, bladder, and breast
cancers (Fig. 1).3,4 Gemcitabine is converted intracellularly to the active metabolites such as
gemcitabine-5’-diphosphate (dFdCDP) and gemcitabine-5’-triphosphate (dFdCTP). dFdCDP
is a potent inhibitor of ribonucleotide reductase (RNR), thereby decreasing the
deoxyribonucleotide pools available for DNA synthesis.2 dFdCTP is incorporated into DNA,
resulting in DNA strand elongation termination after the insertion of one additional
nucleotide.5 A number of deactivation processes take place during gemcitabine’s metabolic
transformations into its active 5’-triphosphate form. dFdC is susceptible to catabolic effects of
cytidine deaminase (CDA), deoxycytidylate deaminase (dCTD) and 5’-nucleotidases (5’-
NTs).6 Importantly, CDA displays two fold higher affinity for gemcitabine than for
deoxycytidine (dC). Hence, dFdC derivatives are designed with an N-substituted exoamine
group which renders them resistant to cytidine deaminase7 thus extending the biological half-
life of gemcitabine due to their increased metabolic stability.8 Also, some 5-fluoropyrimidine
derivatives including 5-fluorouracil (FUra), 5-fluoro-2’-deoxyuridine (floxuridine, FdU, 2)
and 5’-deoxy-5-fluoro-4-N-(pentyloxycarbonyl)cytidine (capecitabine, CAP) play an
important role among anticancer drugs (Fig. 1).2,9 These agents are primarily used in the
treatment of colorectal, head and neck, and breast cancers.10 It was established that the main
mechanism of action of FUra, FdU and CAP involves their intracellular conversion to 5-
fluoro-2’-deoxyuridine 5’-monophosphate (FdUMP), which in turn acts as a suicide inhibitor
of thymidylate synthase (TS), a key enzyme in DNA synthesis.11 Another pyrimidine
nucleoside analogue, 3’-azido-3’-deoxythymidine (zidovudine, AZT, 3) plays a crucial role
among anti-HIV drugs for the treatment of acquired immunodeficiency syndrome (AIDS)
(Fig. 1).1 In addition, there are reports on the use of AZT as an antitumor agent in
combination with either 5-fluorouracil or methotrexate in the therapy of advanced colon
cancer.12–14 Furthermore, Wagner reported the potent inhibitory activity of AZT in cultured
human breast cancer cells.15

JOURNAL PRE-PROOF
O
NH2
N
HN
O
O
N
O
F
O
N
F
F
CH3
N
HN
HN
N
HN
O
O
N
H
O
O
N
O
HO
HO
HO
O
F
O
O
O
F
HO
OH
OH
N3
1
2
dFdC,
FUra
FdU,
3
CAP
AZT,
Fig. 1. 5-Fluoropyrimidine-based drugs (FUra, FdU and CAP), gemcitabine (dFdC) and 3’-
azido-3’-deoxythymidine (AZT).
An alternative to combination therapy in the treatment of cancer or multiple cancers is
the concept of molecular hybridization of anticancer drugs.16 In a hybrid compound two or
more drugs are covalently linked and available as a single entity.17 Drug hybrids with
improved affinity and efficacy relative to those of the parent drugs are of great value. Various
methods of synthesis of bioconjugates are known.18,19 In recent years methods using azides
have become particularly popular: Staudinger ligation of azides and phosphine20 as well as the
copper(I) catalyzed Huisgen azide-alkyne cycloaddition.21 An important role is played by the
linker between two components of a hybrid molecule.22 Conjugation using the copper(I)
catalyzed Huisgen reaction can yield, for example, an amide-methyltriazole linker or an ester-
methyltriazole linker that are considered biolabile, unlike a simple triazole or ether-
methyltriazole linker.23,24
The aim of our research was to employ the ‘click’ chemistry approach in the synthesis
of novel anticancer nucleoside dimers with a cleavable ester-methyltriazole linker, having
potentially better cytotoxic properties than the parent nucleosides. In this report, we describe
the synthesis of new gemcitabine-nucleoside analogue dimers containing methyltriazole (21–
30) or ester-methyltriazole 4N-3’(or 5’)C linker (31–40) as well as the evaluation of their
cytotoxic activity in five human cancer cell lines: cervical (HeLa), nasopharyngeal (KB), lung
(A549), brain (U87), liver (HepG2) and normal dermal fibroblast cell line (HDF).
The synthesis of two series of nucleoside dimers 21–30 and 31–40 by the ‘click’
chemistry method using the copper(I) catalyzed Huisgen azide-alkyne cycloaddition
reaction,7,25,26 required the prior preparation of the appropriate alkyne and azide components.

JOURNAL PRE-PROOF
The alkyne components: 4-N-propargyl-gemcitabine (5) and 4-N-propargyloxycarbonyl-
gemcitabine (6) were prepared as shown in Scheme 1. In the first step of the synthesis of 4-N-
propargyl-gemcitabine (5), gemcitabine (1) was treated with trimethylsilyl chloride in
pyridine for transient protection of the hydroxyl groups27 and next in situ with p-
toluenesulfonyl chloride. The trimethylsilyl groups were then removed from the 3’ and 5’
positions of the intermediate compound using methanolic ammonia to give 4-N-tosyl-
gemcitabine (4).28 The latter compound was reacted with propargylamine in acetonitrile at
120 oC (in a sealed glass tube) to afford 4-N-propargyl-gemcitabine (5). 4-N-
Propargyloxycarbonyl-gemcitabine (6) was prepared starting from gemcitabine (1) which was
treated with trimethylsilyl chloride in pyridine and next in situ with propargyl chloroformate
(propargyloxycarbonyl chloride, PocCl). Finally, the trimethylsilyl groups were removed
from the 3’ and 5’ positions of the intermediate compound using methanol.
NH2
N
HNTs
HN
N
N
N
O
O
N
O
N
a
b
HO
HO
HO
O
F
O
F
O
F
F
F
F
HO
HO
HO
1
4
5
c
O
O
S
HN
O
Ts:
O
N
O
O
N
HO
Poc:
O
O
F
F
HO
6
Scheme 1. Synthesis of 4-N-propargyl-gemcitabine (5) and 4-N-propargyloxycarbonyl-
gemcitabine (6). Reagents and conditions: (a) (i) TMSCl, Py, under Ar, rt, 90 min; (ii) TsCl,
60 oC, 20 h; (iii) CH3OH/NH3, rt, 24 h; (b) propargylamine, CH3CN, 120 oC, 8 h, sealed glass
tube; (c) (i) TMSCl, Py, under Ar, rt, 90 min; (ii) PocCl, under Ar, rt, 12 h; (iii) CH3OH, rt,
15 h.

JOURNAL PRE-PROOF
The azide components: 5’-azido-2’,5’-dideoxythymidine (10) and 5’-azido-2’,5’-
dideoxy-5-fluorouridine (11) as well as their 3’-O-t-butoxycarbonyl derivatives 12 and 13
were prepared as presented in Scheme 2. In the first step of the synthesis, thymidine (7) or 5-
o
fluoro-2’-deoxyuridne (2) was tosylated at low temperature (about 0 C) in pyridine to yield
the appropriate 5’-O-tosyl derivatives 8 and 9, which in turn were reacted with sodium azide
o
in DMF at 90 C to afford 10 and 11, respectively.29,30 Compounds 10 and 11 were
transformed into their corresponding 3’-O-t-butoxycarbonyl (3’-O-Boc) derivatives 12 and 13
by means of di-t-butyl dicarbonate (DBDC) in the presence of DMAP and triethylamine in
dioxane.31
O
N
O
N
O
N
O
N
R
R
R
R
HN
HN
HN
HN
O
O
O
O
b
c
a
HO
TsO
N3
N3
O
O
O
O
OH
OH
OH
OBoc
12
13
7
2
8
9
10
11
: R = CH3
: R = F
: R = CH3
: R = F
: R = CH3
: R = F
: R = CH3
: R = F
O
O
S
Boc:
Ts:
O
O
Scheme 2. Synthesis of 5’-azido-2’,5’-dideoxythymidine (10) and 5’-azido-2’,5’-dideoxy-5-
fluorouridine (11) and their 3’-O-t-butoxycarbonyl derivatives 12 and 13. Reagents and
conditions: (a) TsCl, Py, 0 oC, 48 h; (b) NaN3, DMF, 90 oC, 5 h; (c) DBDC, DMAP, dioxane,
Et3N, rt, 1 h.
The synthesis of 5’-azido-gemcitabine (15) and its derivatives 16 and 17 used in the
preparation of nucleoside dimers is outlined in Scheme 3. 5’-Azido-4-N,3’-O-dibenzoyl-
gemcitabine (14) was synthesized by the known method of Dwyer et al.31 Compound 14 was
then treated with methanolic ammonia to yield 5’-azido-gemcitabine (15) which was reacted
with di-t-butyl dicarbonate (DBDC) in the presence of DMAP and triethylamine in dioxane to
afford 5’-azido-3’-O-t-butoxycarbonyl-gemcitabine (16). The latter compound was converted
into 5’-azido-3’-O-t-butoxycarbonyl-4-N-benzyloxycarbonyl-gemcitabine (17) using benzyl
chloroformate (benzyloxycarbonyl chloride, CbzCl) in the presence of DMAP in pyridine.

JOURNAL PRE-PROOF
HNBz
NH2
N
NH2
HNCbz
N
N
N
N
O
N
O
O
N
O
N
a
b
c
N3
N3
N3
N3
O
F
O
F
O
F
O
F
F
F
F
F
BzO
14
HO
15
BocO
16
BocO
17
O
O
O
O
Cbz:
Bz:
Boc:
O
Scheme 3. Synthesis of 5’-azido-gemcitabine (15) and its derivatives 16 and 17. Reagents and
conditions: (a) CH3OH/NH3, rt, 12 h; (b) DBDC, DMAP, dioxane, Et3N, rt, 12 h; (c) CbzCl,
DMAP, Py, rt, 12 h.
The synthesis of 5’-azido-4-N-benzyloxycarbonyl-gemcitabine (20) and its
transformation to 5’-azido-3’-O-t-butoxycarbonyl-4-N-benzyloxycarbonyl-gemcitabine (17)
is depicted in Scheme 4. Gemcitabine (1) was treated with trimethylsilyl chloride in pyridine
and next in situ with benzyl chloroformate, trimethylsilyl groups were then removed from the
3’ and 5’ positions of the intermediate compound using methanol to yield 4-N-
benzyloxycarbonyl-gemcitabine (18). The latter compound was mesylated at low temperature
(about 0 oC) in pyridine to afford 5’-O-mesyl-4-N-benzyloxycarbonyl-gemcitabine (19) which
was then reacted with sodium azide in DMF at 70 oC to give 5’-azido-4-N-
benzyloxycarbonyl-gemcitabine (20). Compound 20 was finally converted into 3’-O-t-
butoxycarbonyl (3’-O-Boc) derivative 17 by means of di-t-butyl dicarbonate (DBDC) in the
presence of DMAP and triethylamine in dioxane.31

JOURNAL PRE-PROOF
NH2
N
H
H
NCbz
N
NCbz
N
HNCbz
HNCbz
N
N
N
N
N
O
O
O
O
N
O
N
a
b
c
d
HO
HO
MsO
N3
N3
O
F
O
F
O
F
O
F
O
F
F
F
F
F
HO
HO
18
HO
19
HO
F
BocO
1
20
17
O
O
O
S
O
Ms:
Cbz:
Boc:
O
O
Scheme 4. Synthesis of 5’-azido-4-N-benzyloxycarbonyl-gemcitabine (20) and 5’-azido-3’-
O-t-butoxycarbonyl-4-N-benzyloxycarbonyl-gemcitabine (17). Reagents and conditions: (a)
(i) TMSCl, Py, under Ar, rt, 90 min; (ii) CbzCl, Py, rt, 4 h; (iii) CH3OH, rt, 15 h; (b) MsCl,
Et3N, Py, 0 oC, 60 h; (c) NaN3, DMF, 70 oC, 5 h; (d) DBDC, DMAP, dioxane, Et3N, rt, 1 h.
Dimers 21–30 were synthesized by the Huisgen reaction of 4-N-propargyl-
gemcitabine (5) with appropriate azides 3, 10–17, 20 in the presence of copper(I) cations
generated in situ from copper(II) sulfate and sodium ascorbate in THF-water (3:1, v/v)
(Scheme 5).

JOURNAL PRE-PROOF
HN
HN
N
R1
N
N
N
N
O
N
O
N
a
R1 N3
+
HO
HO
O
F
O
F
3 10 17 20
,
–
,
F
HO
F
HO
21 30
–
5
O
N
O
O
O
N
O
N
CH3
F
CH3
F
CH3
HN
HN
HN
HN
HN
O
O
N
O
N
O
O
R1
=
HO
O
O
O
O
O
OBoc
OBoc
OH
OH
25
21
22
24
23
HNBz
NH2
N
NH2
N
HNCbz
HNCbz
N
N
N
N
N
O
N
O
O
O
N
O
N
O
F
O
F
O
F
O
F
O
F
F
F
F
F
BzO
26
HO
27
BocO
BocO
F
HO
30
29
28
Scheme 5. Synthesis of nucleoside dimers 21–30 via the copper(I) catalyzed Huisgen
reaction. Reagents and conditions: (a) CuSO4, sodium ascorbate, THF/H2O (3:1, v/v), rt, 12 h.
Dimers 31–40 (except compound 35) were also synthesized by the Huisgen reaction of
4-N-propargyloxycarbonyl-gemcitabine (6) with appropriate azides 3, 10–12, 14–17, 20 in the
presence of copper(I) cations generated in situ from copper(II) sulfate and sodium ascorbate
in THF-water (3:1, v/v) (Scheme 6). Unfortunately, it was not possible to synthesize
compound 35 despite numerous attempts including the use of Meldal's copper(I) iodide in the
presence of diisopropylethylamine (DIPEA)26 and microwave methods.33

JOURNAL PRE-PROOF
O
O
HN
O
HN
O
N
R1
N
N
N
N
O
N
O
N
a
R1 N3
+
HO
HO
O
F
O
F
3 10 17 20
,
–
,
F
HO
F
HO
31 40
–
6
O
N
O
O
O
N
O
N
CH3
F
CH3
F
CH3
HN
HN
HN
HN
HN
O
O
N
O
N
O
O
R1
=
HO
O
O
O
O
O
OBoc
OBoc
OH
OH
35
31
32
34
33
HNBz
NH2
N
NH2
N
HNCbz
HNCbz
N
N
N
N
N
O
N
O
O
O
N
O
N
O
F
O
F
O
F
O
F
O
F
F
F
F
F
BzO
36
HO
37
BocO
BocO
F
HO
40
39
38
Scheme 6. Synthesis of nucleoside dimers 31–40 via the copper(I) catalyzed Huisgen
reaction. Reagents and conditions: (a) CuSO4, sodium ascorbate, THF/H2O (3:1, v/v), rt, 12 h.
A number of triazole-linked nucleoside dimers and oligomers have been previously
obtained,23,24 including those with an ether-methyltriazole linkage instead of a phosphodiester
bridge.33-36 Some of the synthesized nucleoside dimers have been tested for their anticancer
activity.35,36 On the other hand, it is known that dinucleotide 5-aza-2’-deoxycytidylyl-(3’-5’)-
2’-deoxyguanosine (guadecitabine, SGI-110) exhibits anticancer properties.2
The synthesized nucleoside dimers 21–40 were evaluated for their cytotoxic activity in
five human cancer cell lines such as cervical (HeLa), nasopharyngeal (KB), lung (A549),
brain (U87), liver (HepG2) and normal dermal fibroblast cell line (HDF) using the
sulforhodamine B (SRB) assay.37 The resulting cytotoxic activity data of compounds 21–40
are presented in Table 1. Data for the reference compounds such as gemcitabine (dFdC, 1), 5-
fluoro-2’-deoxyuridine (FdU, 2), 3’-azido-3’-deoxythymidine (AZT, 3) and cytarabine (ara-
C) are also included.

JOURNAL PRE-PROOF
Table 1
In vitro cytotoxic activity of nucleoside dimers 21–40 in five human cancer cell lines: cervical
(HeLa), nasopharyngeal (KB), lung (A549), brain (U87), liver (HepG2) and normal human
dermal fibroblast cell line (HDF)
Cytotoxicity (IC50, M)a ± SDb
Compd
log Pc
HeLa
KB
A549
U87
HepG2
12.16 ± 0.11
130.10 ± 0.07
78.91 ± 0.45
6.34 ± 0.09
5.71 ± 0.29
5.82 ± 0.13
15.12 ± 0.25
5.19 ± 0.19
3.93 ± 0.16
4.64 ± 0.14
1.53 ± 0.08
14.55 ± 0.85
7.66 ± 0.25
11.10 ± 0.11
-
HDF
21
10.89 ± 0.05
128.74 ± 1.32
78.92 ± 0.38
6.30 ± 0.06
4.68 ± 0.12
5.41 ± 0.14
16.23 ± 0.24
4.11 ± 0.09
3.12 ± 0.08
4.23 ± 0.07
1.52 ± 0.07
13.53 ± 0.24
6.70 ± 0.31
9.65 ± 0.21
-
12.21 ± 0.61
130.61 ± 1.04
81.51 ± 0.85
7.07 ± 0.27
4.62 ± 0.06
5.34 ± 0.15
15.91 ± 0.27
4.09 ± 0.13
3.09 ± 0.07
4.14 ± 0.09
1.55 ± 0.06
13.13 ± 0.15
6.52 ± 0.28
10.74 ± 0.14
-
10.57 ± 0.07
125.51 ± 1.02
79.29 ± 0.84
6.07 ± 0.25
5.86 ± 0.32
6.01 ± 0.21
16.11 ± 0.11
4.23 ± 0.14
3.24 ± 0.09
4.31 ± 0.10
1.67 ± 0.08
13.44 ± 0.67
8.01 ± 0.15
8.50 ± 0.23
-
10.77 ± 0.75
127.50 ± 0.86
80.30 ± 0.47
6.97 ± 0.33
4.73 ± 0.28
5.50 ± 0.12
16.47 ± 0.28
4.54 ± 0.21
3.81 ± 0.11
4.45 ± 0.12
1.89 ± 0.05
13.53 ± 0.44
7.56 ± 0.36
9.68 ± 0.27
-
17.81 ± 0.02
151.81 ± 1.25
105.72 ± 0.66
14.22 ± 0.21
13.74 ± 0.31
14.22 ± 0.23
29.29 ± 0.31
10.90 ± 0.25
8.12 ± 0.21
10.10 ± 0.21
6.70 ± 0.10
16.87 ± 0.34
10.20 ± 0.61
15.04 ± 0.07
-
-2.79
-2.79
-3.07
-1.14
-1.42
0.88
22
23
24
25
26
27
-2.96
-1.31
0.51
28
29
30
-1.14
-2.74
-2.74
-3.03
-1.09
-1.38
0.92
31
32
33
34
35
36
3.22 ± 0.06
12.44 ± 0.25
2.01 ± 0.06
0.97 ± 0.05
1.98 ± 0.03
9.42 ± 0.19
6.50 ± 0.24
10.77 ± 0.66
3.54 ± 0.16
3.13 ± 0.07
11.35 ± 0.28
2.12 ± 0.05
0.95 ± 0.08
2.01 ± 0.05
8.21 ±0.07
8.69 ± 1.18
9.77 ± 0.57
4.07 ± 0.08
3.32 ± 0.08
11.43 ± 0.12
2.17 ± 0.06
0.83 ± 0.07
2.11 ± 0.07
5.05 ±0.30
5.91 ± 0.23
11.86 ± 0.82
3.47 ± 0.07
3.41 ± 0.06
12.58 ± 0.18
2.21 ± 0.09
1.09 ± 0.08
2.14 ± 0.09
5.21 ± 0.46
6.14 ± 0.14
11.88 ± 0.16
2.88 ± 0.19
3.51 ± 0.09
10.81 ± 0.21
2.34 ± 0.10
1.01 ± 0.09
2.21 ± 0.08
4.90 ± 0.79
8.91 ± 0.34
13.03 ± 0.23
2.86 ± 0.09
9.12 ± 0.11
15.82 ± 0.17
8.81 ± 0.12
4.69 ± 0.06
8.12 ± 0.12
19.45 ± 0.03
13.05 ± 0.74
14.41 ± 0.58
4.99 ± 0.84
37
-2.91
-1.26
0.56
38
39
40
-1.09
-1.60
-1.72
-0.10
-1.93
dFdCd
FdUd
AZTd
ara-Cd
a IC50 is the compound concentration required to inhibit cell growth by 50%.
b SD (standard deviation) of three independent experiments.
c log P (logarithm of partition coefficient) was calculated using “log PKnowwin” method.38
d Standards: 2’,2’-difluoro-2’-deoxycytidine (gemcitabine, dFdC), 5-fluoro-2’-deoxyuridine
(FdU), 3’-azido-3’-deoxythymidine (AZT) and cytarabine (ara-C).

JOURNAL PRE-PROOF
Among dimers 21–30 with methyltriazole 4N-3’(or 5’)C linker, the highest activity
(IC50 in the range of 3.09–3.93 M), in the all studied cancer cell lines, was exhibited by
compound 29 comprising two gemcitabine units, with the second unit having a
benzyloxycarbonyl group (Cbz) on the exoamine function and a t-butoxycarbonyl group
(Boc) on the 3'-hydroxyl function. The activity of compound 29 was 1.3 to 3 times higher
than that of the parent dFdC depending on the cell line. A similar order of activity was
observed for compounds 25, 28 and 30. It should be noted that the most active compound 29
within the 21–30 series has a partition coefficient (log P) value of 0.51 and could potentially
penetrate the cell membranes. Once inside the cells, the benzyloxycarbonyl and t-
butoxycarbonyl groups could be hydrolyzed by esterase-type enzymes. We have previously
demonstrated that the introduction of the t-butoxycarbonyl group into the 3’-position of some
5-fluoro-2'-deoxyuridine 5'-phosphoramidates results in a significant increase of their
cytotoxic activity, which might be associated with their enhanced ability to penetrate the cell
membranes.39
As for dimers 31–40 with ester-methyltriazole 4N-3’(or 5’)C linker, the highest
activity (IC50 in the range of 0.83–1.09 M), in all the studied cancer cell lines, was observed
for compound 39 also comprising of two gemcitabine units, with the second unit having a
benzyloxycarbonyl group (Cbz) on the exoamine function and a t-butoxycarbonyl group
(Boc) on the 3'-hydroxyl function. The activity of compound 39 was 5 to 9 times higher than
that of gemcitabine (dFdC) depending on the cell line. Moreover, considerable activity, higher
than that of dFdC, was displayed by compounds 31, 36, 38, and 40 (IC50 in the range of 1.52–
3.51 M). The most active compound 39 in the series 31–40 has a partition coefficient (log P)
value of 0.56 which suggests that it could easily penetrate the cell membranes.
Partition coefficients (log P) values of dimers 21–30 and 31–40 were calculated38 to
determine a possible correlation between their cytotoxicity data and lipophilicity (Table 1).
Linear regression analysis revealed some correlations between log P values and the
cytotoxicity expressed as negative logarithm of IC50 and this is shown in Figure 2 for two
cancer cell lines: HeLa and KB.

JOURNAL PRE-PROOF
(a)
(b)
6
5.5
5
6
28
28
30
24
29
26
30
29
26
5.5
245
4.5
4
25
21
25
21
27
27
4.5
4
23
23
22
22
3.5
3.5
3
HeLa
KB
3
-4
-3
-2
-1
0
1
2
log P
0
-5
5
log P
(c)
(d)
6.5
6
6.5
6
39
39
31
40
31
40
38
38
36
5.5
36
5.5
33
37
33
5
34
34
5
32
37
32
4.5
4
4.5
4
HeLa
KB
3.5
3
3.5
3
-4
-3
-2
-1
0
1
2
-4
-3
-2
-1
0
1
2
log P
log P
Fig. 2. Correlation between log P and –log IC50 for compounds 21–30 for two cancer cell
lines: (a) HeLa and (b) KB and for compounds 31–40 for two cancer cell lines: (c) HeLa and
(d) KB.
In general, compounds 31–40 having the ester-methyltriazole linker showed greater
cytotoxic activity than compounds 21–30 with the methyltriazole linker. The ester-
methyltriazole linker can be considered as a fragment of the (1H-1,2,3-triazol-4-yl)methyl
carbamate structure. A similar carbamate (pentyl carbamate) motif appears in the capecitabine
structure (Fig. 1) and it was shown to be hydrolyzed by carboxylesterases.40,41 In contrast, the
methyltriazole linker present in compounds 21–30 is a fragment of the (1H-1,2,3-triazol-4-
yl)methyl amine structure and is much more stable.
The cytotoxic effect was also studied in the normal human dermal fibroblasts (HDF)
to assess the toxicity of nucleoside dimers 21–40 (Table 1). The selectivity index (SI) was

JOURNAL PRE-PROOF
calculated for the most active nucleoside dimers as the ratio of the IC50 for the normal cell line
(HDF) to the IC50 for a respective cancer cell line (Table 2). Higher values of SI indicate
greater anticancer selectivity and compounds exhibiting SI’s values higher than 3 are
considered to be highly selective.42 Compounds 31, 38–40 had not only high cytotoxic
activity but also displayed low toxicity against normal fibroblast cells and their SI was higher
than 3 (Table 2). The highest selectivity index (SI = 5.65) was observed for compound 39 in
A549 cancer cell line.
Table 2
The calculated values of the selectivity index (SI) of the most active nucleoside dimers
Compd
SIa
HeLa
2.94
2.65
2.60
2.39
4.41
2.83
4.38
4.83
4.10
2.07
2.01
1.34
1.41
KB
A549
U87
2.91
2.40
2.13
2.27
3.55
2.67
3.99
4.30
3.79
3.73
2.13
1.21
1.73
HepG2
2.41
2.10
2.07
2.18
4.38
2.60
3.77
4.64
3.67
3.97
1.47
1.11
1.75
25
2.97
2.67
2.63
2.44
4.32
2.91
4.16
4.94
4.04
2.37
1.50
1.48
1.23
2.35
2.58
2.51
2.34
4.01
2.75
4.06
5.65
3.85
3.85
2.21
1.22
1.44
28
29
30
31
36
38
39
40
dFdCb
FdUb
AZTb
ara-Cb
a The selectivity index (SI) was calculated for each compound using the formula: SI = IC50 for
normal cell line HDF/IC50 for respective cancer cell line. A beneficial SI > 1.0 indicates a
compound with efficacy against tumor cells greater than the toxicity against normal cells.
b Standards: 2’,2’-difluoro-2’-deoxycytidine (gemcitabine, dFdC), 5-fluoro-2’-deoxyuridine
(FdU), 3’-azido-3’-deoxythymidine (AZT) and cytarabine (ara-C).

JOURNAL PRE-PROOF
For studies on the metabolism of nucleoside dimers, compound 21 from the series with
methyltriazole linker and compound 31 from the series with ester-methyltriazole linker were
selected. The studies were conducted using the HeLa cell line and were described in detail in
supporting information. The lysate of HeLa cells was analyzed after 12 hours of incubation
with compound 21 or 31 by the HPLC/MS method. We have shown the presence of
gemcitabine in the lysate of HeLa cells incubated with compound 21, but unfortunately, the
thymidine fragment could not be detected. It may have been converted into a range of
products and hence there is a problem with their detection because not present in sufficient
concentration. We assumed that compound 21 could be hydroxylated on the 4N-CH2 group by
cytochrome P450 and decomposed into gemcitabine and 3’-[4-formyl-(1,2,3-triazol-1-yl)]-3’-
deoxythymidyne which could be further oxidized by the aldehyde dehydrogenase to 3’-[4-
carboxy-(1,2,3-triazol-1-yl)]-3’-deoxythymidyne.
Similarly,
the
metabolism
of
cyclophosphamide begins with the hydroxylation of the NH-CH2 group by cytochrome
P450.43 Nevertheless, the release of gemcitabine by compound 21 may indicate that it acts as
a gemcitabine prodrug. In turn, we showed the presence of gemcitabine and 3’-[4-
(hydroxymethyl)-(1,2,3-triazol-1-yl)]-3’-deoxythymidine (41) in the lysate of HeLa cells
incubated with compound 31, which demonstrates that the carbamate ester bond is hydrolyzed
(Scheme 7). Initially released 4-N-carboxy-gemcitabine rapidly decarboxylates to give
gemcitabine.41

JOURNAL PRE-PROOF
O
CH3
HN
O
N
HO
O
O
N
NH2
CH3
HN
N
N
O
O
N
a
N
N
HO
HO
+
O
O
O
F
HN
O
F
HO
N
N
N
N
1
O
N
HO
O
F
OH
41
F
HO
31
b
O
N
CH3
HN
O
HO
+
O
OH
N3
3
Scheme 7. Hydrolysis of compound 31 in HeLa cells and independent synthesis of compound
41 via the copper(I) catalyzed Huisgen reaction. Reagents and conditions: (a) incubation with
HeLa cells at 37 oC for 12 h; (b) CuSO4, sodium ascorbate, THF/H2O (3:1, v/v), rt, 12 h.
An authentic sample of compound 4144 was synthesized by the Huisgen reaction of 3'-
azido-3'-deoxythymidine (3) with propargyl alcohol in the presence of copper(I) cations
generated in situ from copper(II) sulfate and sodium ascorbate in THF-water (3:1, v/v)
(Scheme 7). The synthesized compound 41 served as a reference for analysis of the lysate of
HeLa cells incubated with compound 31 by HPLC/MS method.
It should be mentioned that in the lysate of HeLa cells incubated with compound 21
and 31 no gemcitabine deamination product (2',2'-difluoro-2'-deoxyuridine, dFdU) was
detected.
In conclusion, we have synthesized two series of novel gemcitabine nucleoside
analogue dimers employing the ‘click’ chemistry approach. Each of the two series of dimers
has a different type of the linker connecting gemcitabine to the other nucleoside unit. In the
first series of dimers 21–30 a stable methyltriazole linker connects the exoamine group

JOURNAL PRE-PROOF
nitrogen atom (4N) of gemcitabine with the 3’ or 5’ carbon atom of the second nucleoside
unit. Whereas, in the second series of dimers 31–40, there is a labile ester-methyltriazole 4N-
3’(or 5’)C linker which can be cleaved by carboxylesterases. Dimers 21–40 were tested for
their cytotoxic activity in five human cancer cell lines: cervical (HeLa), nasopharyngeal (KB),
lung (A549), brain (U87), liver (HepG2) and normal dermal fibroblast cell line (HDF) using
the sulforhodamine B (SRB) assay. Within dimers 21–30 the highest activity was exhibited by
compound 29 incorporating two gemcitabine units, with the second unit having a
benzyloxycarbonyl group (Cbz) on the exoamine function and a t-butoxycarbonyl group
(Boc) on the 3'-hydroxyl function. The activity of compound 29 was 1.3 to 3 times higher
than that of the parent gemcitabine, depending on the cell line. A similar order of activity was
observed for compounds 25, 28, and 30. The highest activity, among all nucleoside dimers
21–40, was demonstrated by compound 39 also incorporating two gemcitabine units joined by
a bio-cleavable linker. The activity of compound 39 was 5 to 9 times higher than that of
gemcitabine, depending on the cell line. Moreover, marked activity was shown by compounds
31, 36, 38, and 40. Compounds 31, 38–40 demonstrated not only high cytotoxic activity, but
also low toxicity against normal fibroblast cell line and their SI was higher than 3.
Metabolism studies revealed the presence of gemcitabine in the lysate of HeLa cells incubated
with compound 21 and gemcitabine and compound 41 in the lysate of HeLa cells incubated
with compound 31.
Acknowledgments
The work was supported by grant no. POWR.03.02.00-00-I023/17 co-financed by the
European Union through the European Social Fund under the Operational Program
Knowledge Education Development.
Appendix A. Supplementary data
Experimental details and NMR data (PDF).
References
1. De Clercq E, Li G. Clin. Microbiol. Rev. 2016;29:695–747.
2. Shelton J, Lu X, Hollenbaugh JA, Cho JH, Amblard F, Schinazi RF. Chem. Rev.
2016;116:14379–14455.

JOURNAL PRE-PROOF
3. Eckel F, Schneider G, Schmid RM. Expert Opin. Invest. Drugs. 2006;15:1395–1410.
4. Toschi L, Finocchiaro G, Bartolini S, Gioia V, Cappuzzo F. Future Oncol. 2005;1:7–
17.
5. Huang P, Chubb S, Hertel LW, Grindey GB, Plunkett W. Cancer Res. 1991;51:6110–
6117.
6. Gesto DS, Cerqueira NMFSA, Fernandes PA, Ramos MJ. Curr. Med. Chem.
2012;19:1076–1087.
7. Gonzales C, Sanchez A, Collins J, Lisova K, Lee JT, van Dam RM, Barbieri MA,
Ramachandran C, Wnuk SF. Eur. J. Med. Chem. 2018;148:314–324.
8. Moysan E., Bastiat G., Benoit JP. Mol. Pharm. 2013;10:430–444.
9. Walko CM, Lindley C. Clin. Ther. 2005;27:23–44.
10. Grem JL. Invest. New Drugs. 2000;18:299–313.
11. Longley DB, Harkin DP, Johnston PG. Nat. Rev. Cancer. 2003;3:330–338.
12. Brunetti I, Falcone A, Calabresi P, Goulette FA, Darnowski JW. Cancer Res.
1990;50:4026–4031.
13. Tosi P, Calabresi P, Goulette FA, Renaud CA, Darnowski JW. Cancer Res.
1992;52:4069–4073.
14. Darnowski JW, Goulette FA. Biochem. Pharmacol. 1994;48:1797–1805.
15. Wagner CR, Ballato G, Akanni AO, McIntee EJ, Larson RS, Chang S, Abul-Hajj YJ.
Cancer Res. 1997;57:2341–2345.
16. Nepali K, Sharma S, Sharma M, Bedi PMS, Dhar KL. Eur. J. Med. Chem.
2014;77:422–487.
17. Mishra SS, Singh P. Eur. J. Med. Chem. 2016;124:500–536.
18. Hermanson GT. Bioconjugate Techniques. third ed., Academic Press, London, 2013.
19. Spicer CD, Pashuck ET, Stevens MM. Chem. Rev. 2018;118:7702–7743.
20. Schilling CI, Jung N, Biskup M, Schepers U, Brase S. Chem. Soc. Rev. 2011;40:4840–
4871.
21. Singh MS, Chowdhury S, Koley S. Tetrahedron. 2016;72:5257–5283.
22. Jaracz S, Chen J, Kuznetsova LV, Ojima I. Bioorg. Med. Chem. 2005;13:5043–5054.
23. Amblard F, Cho JH, Schinazi RF. Chem. Rev. 2009;109:4207–4220.
24. Efthymiou T, Gong W, Desaulniers JP. Molecules. 2012;17:12665–12703.
25. Rostovtsev VV, Green LG, Fokin VV, Sharpless KB. Angew. Chem. Int. Ed.
2002;41:2596–2599.
26. Meldal M, Tornoe CW. Chem. Rev. 2008;108:2952–3015.

JOURNAL PRE-PROOF
27. Ti GS, Gaffney BL, Jones RA. J. Am. Chem. Soc. 1982;104:1316–1319.
28. Pulido J, Sobczak AJ, Balzarini J, Wnuk SF. J. Med. Chem. 2014;57:191–203.
29. Lin TS, Prusoff WH. J. Med. Chem. 1978;21:109–112.
30. Henn TFG, Garnett MC, Chhabra SR, Bycroft BW, Baldwin RW. J. Med. Chem.
1993;36:1570–1579.
31. Bazzanini R, Gouy MH, Peyrottes S, Gosselin G, Périgaud C, Manfredini S.
Nucleosides Nucleotides Nucleic Acids. 2005;24:1635–1649.
32. Labroli MA, Dwyer MP, Shen R, Popovici-Muller J, Pu Q, Richard J, Rosner K,
Paruch K, Guzi TJ. Tetrahedron Lett. 2014;55:598–602.
33. Lucas R, Neto V, Bouazza AH, Zerrouki R, Granet R, Krausz P, Champavier Y.
Tetrahedron Lett. 2008;49:1004–1007.
34. Lucas R, Zerrouki R, Granet R, Krausz P, Champavier Y. Tetrahedron.
2008;64:5467–5471.
35. Michalska L, Wawrzyniak D, Szymanska-Michalak A, Barciszewski J, Boryski J,
Baraniak D. Nucleosides Nucleotides Nucleic Acids. 2019;38:218–235.
36. Baraniak D, Baranowski D, Ruszkowski P, Boryski J. Nucleosides Nucleotides
Nucleic Acids. https://doi.org/10.1080/15257770.2019.1610891
37. Skehan P, Storeng R, Scudiero D, Monks A, McMahon J, Vistica D, Warren JT,
Bokesch H, Kenney S, Boyd MR. J. Natl. Cancer Inst. 1990;82:1107–1112.
38. Pyka A, Babuska M, Zachariasz M. Acta Pol. Pharm. Drug Res. 2006;63:159–167.
39. Lewandowska M, Ruszkowski P, Chojnacka K, Kleczewska N, Hoffmann M,
Kacprzak K, Celewicz L. Bioorg. Med. Chem. 2016;24:2330–3241.
40. Quinney SK, Sanghani SP, Davis WI, Hurley TD, Sun Z, Murry DJ, and Bosron WF.
J. Pharmacol. Exp. Ther. 2005;313:1011–1016.
41. Ghosh AK, Brindisi M. J. Med. Chem. 2015;58:2895–2940.
42. Machna S, Weerapreeyakul N, Barusrux S, Nonpunya A, Sripanidkulchai B,
Thitimethaaroch T. Chin. Med. 2011;6:39–46.
43. Emadi A, Jones RJ, Brodsky RA. Nat. Rev. Clin. Oncol. 2009;6:638–647.
44. Zhou L, Amer A, Korn M, Burda R, Balzarini J, De Clercq E, Kern ER, Torrence PF.
Antivir. Chem. Chemother. 2005;16:375–383.