Design, synthesis, and evaluation of liver-specific gemcitabine prodrugs for potential treatment of hepatitis C virus infection and hepatocellular carcinoma

Article abstract of DOI:10.1016/j.ejmech.2020.113135

Many successful anti-viral and anti-cancer drugs are nucleoside analogs, which disrupt RNA and/or DNA synthesis. Here, we present liver-specific prodrugs of the chemotherapy drug gemcitabine (2′,2′-difluorodeoxycytidine) for the treatment of hepatitis C v

Full text of DOI:10.1016/j.ejmech.2020.113135

European Journal of Medicinal Chemistry 213 (2021) 113135
Contents lists available at ScienceDirect
European Journal of Medicinal Chemistry
journal homepage: http://www.elsevier.com/locate/ejmech
Design, synthesis, and evaluation of liver-specific gemcitabine
prodrugs for potential treatment of hepatitis C virus infection and
hepatocellular carcinoma
,
b
,
1
,
,
2
,
, 3, 5
Anthony A. Stephenson a
5, Sheng Cao a 5, David J. Taggart a
,
Vinod P. Vyavahare a 4, Zucai Suo a
,
, b, c, *
a Department of Biochemistry, The Ohio State University, Columbus, OH, 43210, USA
b The Ohio State Biochemistry Program, The Ohio State University, Columbus, OH, 43210, USA
c Department of Biomedical Sciences, Florida State University College of Medicine, Tallahassee, FL, 32306, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
Many successful anti-viral and anti-cancer drugs are nucleoside analogs, which disrupt RNA and/or DNA
synthesis. Here, we present liver-specific prodrugs of the chemotherapy drug gemcitabine (20,20-
difluorodeoxycytidine) for the treatment of hepatitis C virus (HCV) infection and hepatocellular carci-
noma. The prodrugs were synthesized by introducing aromatic functional moieties to the cytosine 4-NH2
group of gemcitabine via amide bonds. The chemical modification was designed to i) enable passive
diffusion across cellular membrane, ii) protect the prodrugs from inactivating deamination by cellular
enzymes, and iii) allow release of active gemcitabine after amide hydrolysis by high levels of carbox-
ylesterases in the liver. We found that many of our prodrugs exhibited similar toxicity as gemcitabine
toward liver- and kidney-derived cancer cell lines but were 24- to 620-fold less cytotoxic than gemci-
tabine in breast- and pancreas-derived cancer cells, respectively. The prodrugs also inhibited an HCV
Received 28 September 2020
Received in revised form
16 December 2020
Accepted 21 December 2020
Available online 5 January 2021
Keywords:
Gemcitabine prodrugs
Hepatitis C virus
Liver-specific prodrugs
Hepatocellular carcinoma
Carboxylesterase
replicon with IC50 values ranging from 10 nMe1.7 mM. Moreover, many of the prodrugs had therapeutic
index values of >10,000 and have synergetic effects when combined with other Food and Drug
Administration-approved anti-HCV small molecule drugs. These characteristics support the development
of gemcitabine prodrugs as liver-specific therapeutics.
Therapeutic index
© 2020 Elsevier Masson SAS. All rights reserved.
1. Introduction
cases arise each year [1,2]. Chronic HCV infection can cause both
liver cirrhosis and hepatocellular carcinoma [3]. As the fifth most
Approximately 170 million individuals suffer from chronic
hepatitis C virus (HCV) infection world-wide and 3e4 million new
common cancer and the third leading cause of cancer-related
deaths, hepatocellular carcinoma is a tremendous socioeconomic
Abbreviations: BNPP, bis-p-nitrophenyl phosphate; BxPC3, biopsy xenograft of human pancreatic carcinoma 3 cell line; CNTs, concentrative nucleoside transporters; dCK,
deoxycytidine kinase; dFdC, 20,20-difluorodeoxycytidine or gemcitabine; dFdCDP, gemcitabine diphosphate; dFdCMP, gemcitabine monophosphate; dFdCTP, gemcitabine
triphosphate; dFdU, 20,20-difluorodeoxyuridine; DMEM, Dulbecco’s modified Eagle’s medium; DMSO, dimethyl sulfoxide; ENTs, equilibrative nucleoside transporters; ENT1,
Equilibrative Nucleoside Transporter 1; FBS, fetal bovine serum; HCV, hepatitis C virus; HPLC, high performance liquid chromatography; HepG2, human hepatocarcinoma G2
cell line; Huh7, human hepatocarcinoma 7 cell line; Huh7HCVꢀRep-Luc, human hepatocarcinoma 7 cell line harboring an HCV replicon containing a luciferase reporter gene;
IC50, compound concentration that induces 50% inhibition; MCF-7, Michigan Cancer Foundation-7 breast cancer cell line; NBTI, nitrobenzylthioinosine; TC50, compound
concentration that induces 50% cytotoxicity.
* Corresponding author. Department of Biomedical Sciences, College of Medicine, Florida State University, Tallahassee, FL, 32306, USA.
E-mail address: zucai.suo@med.fsu.edu (Z. Suo).
1
Present address: Post-doctoral Scientist in the Center for Gene Therapy, Abigail Wexner Research Institute of Nationwide Children’s Hospital, 700 Children’s Drive,
Columbus, OH, 43205, USA.
2
Present address: Scientist III at Milliporesigma, 3858 Benner Rd, Miamisburg, OH, 45342, USA.
Present address: Laboratory Director at Laboratory for Advanced Medicine, Inc., 1201 Cumberland Ave, West Lafayette, IN, 47906, USA.
Present address: Research Scientist at Lupin Limited, Lupin Research Park, Survey No. 46 A/47 A, Village Nande, Taluka Mulshi, Pune, 411042, India.
Contributed equally.
3
4
5
https://doi.org/10.1016/j.ejmech.2020.113135
0223-5234/© 2020 Elsevier Masson SAS. All rights reserved.

A.A. Stephenson, S. Cao, D.J. Taggart et al.
European Journal of Medicinal Chemistry 213 (2021) 113135
burden [4]. Thus, novel treatments for these liver-specific diseases
are needed.
difluorodeoxyuridine (dFdU) caused by deoxycytidine deaminase
in cells and serum (Fig. 1) [27e31].
Gemcitabine (20,20-difluorodeoxycytidine, dFdC) is a clinically
approved chemotherapeutic drug used for the treatment of a wide
spectrum of cancers including liver, pancreatic, non-small cell lung,
breast, bladder, and ovarian cancers [5e9]. Gemcitabine is a polar
deoxycytidine analog that requires nucleoside transporters to
translocate across the cellular membrane [10]. Transport of gem-
citabine across the plasma membrane can occur by both equili-
brative nucleoside transporters (ENTs) and concentrative
nucleoside transporters (CNTs), of which human equilibrative
nucleoside transporter 1 (hENT1) is the major transporter (Fig. 1)
[11,12]. After cell entry, gemcitabine is phosphorylated by deoxy-
cytidine kinase (dCK) to generate gemcitabine monophosphate
(dFdCMP), and dFdCMP is then subsequently phosphorylated by
nucleotide kinases to form the active di- and tri-phosphorylated
gemcitabine metabolites: dFdCDP and dFdCTP, respectively
(Fig. 1). dFdCTP has been shown to compete efficiently against dCTP
and CTP for incorporation into genomic DNA and cellular RNA,
respectively [13,14]. Once incorporated, dFdCMP reduces the
incorporation efficiencies of the next several nucleotides into a
nascent polynucleotide chain, thus causing premature termination
of DNA or RNA synthesis [13,15,16]. Furthermore, dFdCMP in DNA
can evade the proofreading exonuclease activity of human DNA
Here, we generated 11 monocyclic and two bicyclic prodrugs of
gemcitabine by introducing aromatic functional groups to the N4
amine group of the cytosine of gemcitabine via an amide bond
linkage (Fig. 2). We rationalized that these aromatic moieties would
protect gemcitabine from inactivating deamination within blood
and tissues as well as reduce the dependence of ENT1-mediated
cellular import. Additionally, the prodrugs are theorized to be
cleaved and activated by carboxylesterases, which are highly
expressed in liver tissue [32]. Thus, these prodrugs were rationally
designed to deliver gemcitabine to liver tissues for the treatment of
liver-specific diseases, such as HCV infection or hepatocellular
carcinoma. We performed in vitro assays to determine cytotoxicity,
HCV inhibition, and developed a preliminary mechanism of action
using select prodrugs to build a model to describe their effects in
cells. We concluded the study by synthesizing an additional 14
bicyclic compounds as well as one tricyclic compound and provided
additional cytotoxicity and HCV-inhibition data on these novel
compounds.
2. Experimental section
2.1. General information
polymerases ε and
g [15,16]. Notably, dFdCDP and dFdCTP can
inhibit ribonucleotide reductase and decrease cellular dCTP con-
centrations, leading to increased phosphorylation of dFdCDP
[17,18]. Consequently, high concentrations of dFdCTP inhibit CTP
synthetase and lower cellular dCTP and CTP pools [19,20]. Dimin-
ished dCTP pools results in reduced competition and increases the
probability of dFdCTP incorporation thus causing cell cycle arrest
and apoptosis as well as inhibition of viral infection [17]. Lastly,
dFdCMP and dFdCTP inhibit dCMP deaminase thereby reducing
inactivation of dFdCMP through dFdUMP [21]. These collective ef-
fects are termed “self-potentiation” (Fig.1) [22,23]. However, tumor
cells can acquire gemcitabine resistance via loss of ENT1 activity,
overexpression of ribonucleotide reductase, and/or reduced dCK
activity [24e26]. Additionally, the therapeutic impact of gemcita-
bine is reduced by deamination to the inactive metabolite 20,20-
Reagents and anhydrous solvents were obtained from com-
mercial sources and used without further purification. Reactions
were monitored by thin layer chromatography with 250 mm silica
gel plates (E. Merck silica gel 60 F254 aluminum-backed plates) and
visualized using UV light or by charring with 5% sulfuric acid in
methanol. Preparative TLC was carried out on silica gel 60 F254
plates (20 ꢁ 20 cm, 1 mm) from EMD Chemicals, Inc. Silica gel
matrices (230e400 mesh size) were utilized for all chromato-
graphic purifications. 1H NMR, 13C NMR, and electrospray ionization
mass spectrometry (ESI-MS) were performed at The Ohio State
University Campus Chemical Instrumentation Center. Chemical
shifts are given in ppm and J values in Hz. Multiplicity is indicated
using the following abbreviations: d for doublet, t for a triplet, q for
a quintet, m for a multiplet. Preparative HPLC was performed on a
Beckman HPLC system using a gradient of water and acetonitrile
with 0.1% TFA at a flow rate of 10 mL/min over 50 min on a C-18
reverse phase column (10
m
m particle size; 21 mm ꢁ 150 mm).
Compound names were generated using ChemBioDraw Ultra,
version 14.0.0.117. Compounds were screened for potential assay
interference activity prior to synthesis using online tools (http://
zinc15.docking.org/patterns/home/and http://advisor.docking.org/
) [33].
2.2. Preparation of 30,50-O-(1,1,3,3-Tetraisopropyldisiloxane-1,3-
dily) gemcitabine (compound 2)
To the solution of gemcitabine (compound 1) (136.6 mg,
0.52 mmol) in dry pyridine (40 mL), 1,3-dichloro-1,1,3,3-
tetraisopropyldisloxane (0.17 mL) was added slowly with stirring
to silylate and protect the 30 and 50 hydroxyl groups. The mixture
was stirred at room temperature for 48 h. Pyridine was then
removed under vacuum and the residue was subjected to a silica gel
column chromatography, with a gradient of methanol (1e2.5%) in
dichloromethane to give compound 2 as a white foam (189.7 mg,
72%). 1H NMR (600 MHz, DMSO‑d6)
d 7.46e7.51 (m, 2H), 7.41 (s, 1H),
6.01e6.20 (m, 1H), 5.81 (d, J ¼ 7.8 Hz, 1H), 4.23e4.45 (m, 1H),
4.14e4.19 (m, 1H), 3.93e4.0 (m, 2H), 0.97e1.09 (m, 24H). 13C NMR
Fig. 1. Gemcitabine activation and self-potentiation pathways. Dashed lines indicate
inhibition of the indicated enzyme or process. Gemcitabine (1) is denoted as dFdC.
Gemcitabine monophosphate (dFdCMP), gemcitabine diphosphate (dFdCDP), gemci-
tabine triphosphate (dFdCTP), deaminated gemcitabine (dFdU) and deaminated
gemcitabine monophosphate (dFdCMP) are also indicated.
(600 MHz, DMSO‑d6)
95.1, 83.9, 78.1, 70.4, 60.4, 17.4, 17.3, 17.2, 16.9, 16.8, 16.7, 12.8, 12.5,
d
165.8, 154.7, 139.5, 122.5 (t, J ¼ 255.0 Hz),
12.2, 12.1. LCMS m/z 528.2 [MþNa]þ.
2

A.A. Stephenson, S. Cao, D.J. Taggart et al.
European Journal of Medicinal Chemistry 213 (2021) 113135
Fig. 2. Synthesis of gemcitabine prodrugs.
2.3. General procedures for synthesis of gemcitabine derivatives
activity detected in each well was then plotted as a function of
compound concentration, and the data were then fit to the four-
parameter logistic equation (1) using Kaleidagraph (Synergy
Software),
For synthesis of each of compounds 3e30, the following general
procedure was used individually. To a solution of compound 2
(0.054 mmol) in dry pyridine (7 mL), an acid chloride derivative of
the respective functional group (1.5 eq.) was added slowly with
stirring. The mixture was stirred at room temperature overnight.
The solvent was removed and the silylated N4-substituted 50-
deoxy-5-fluocytidine derivative (product A, Fig. 2) was used in the
next reaction without further purification. The 30 and 50 hydroxyl
groups were deprotected by addition of tetrabutylammonium
fluoride (1 M in tetrahydrofuran, 0.3 mL) followed by stirring at
room temperature for 1.5 h. After removal of the solvent, the res-
idue was subjected to silica gel column chromatography, with a
stepwise gradient of methanol (1e5%) in dichloromethane to give
the raw N4-substituted 50-deoxy-5-fluocytidine derivative (product
B, Fig. 2). Compounds 3e30 were purified from their respective raw
product B solutions by C-18 reverse phase preparative HPLC using a
gradient of water and acetonitrile with 0.1% TFA. HPLC fractions
containing pure gemcitabine prodrugs (compounds 3e30) were
lyophilized overnight to yield dried products in 60e80% yield with
a purity of >99%.
A
y ¼
(1)
ꢀ ꢁ
C
x
1 þ
B
where y is the relative percent luciferase activity, x is the compound
concentration, A is the relative percent amplitude, B is the con-
centration of the compound at which the relative percent luciferase
activity is reduced to 50% (IC50), and C is the Hill coefficient. To test
for synergy between compound 13 and daclatasvir, a previously
described method was used [34]. Briefly, luciferase activity of
Huh7HCVꢀRep-Luc under a wide range of concentrations for com-
pound 13 (0.001e10,000 nM) while holding a fixed concentration
of daclatasvir was measured as described above. Measurements
were performed over
a series of sub-IC50 concentrations
(0.005e0.025 nM) of daclatasvir. The resulting data sets were each
fit to equation (1), where y is the relative percent luciferase activity,
x is the compound concentration, A is the relative percent ampli-
tude, B is the concentration of the compound at which relative
luciferase activity is reduced to 50% of the corresponding value in
the absence of compound 13, and C is the Hill coefficient. As the B
values derived from data collected in these assays were dependent
upon activity of both inhibitory compounds, the A, B, and C values
were used with equation (2) to calculate IC50 values of compound
13 for each concentration of daclatasvir [34]. Calculated IC50 values
were then plotted against the isobologram of daclatasvir and
compound 13 (Fig. 4).
2.4. Structural and mass analysis of gemcitabine prodrugs
Purified compounds were dissolved in deuterated DMSO and
the 1H and 13C NMR spectra collected on Bruker Avance NMR
spectrometers (300 MHz, 400 MHz, or 600 MHz). For electro-spray
ionization mass spectrometry (ESI-MS) compounds were dissolved
in methanol and measured using electrospray ionization source on
an Agilent 6224 LC-TOF/MS system. Lists of NMR peaks and the ESI-
MS results are provided in Supplemental Data.
ꢂꢂ
ꢃ
1
ꢃ
IC50 ¼ B ꢁ
C ꢀ 1
(2)
A
50
2.5. HCV replicon inhibition assays
Human hepatocarcinoma 7 (Huh7) cells containing an HCV
replicon with a luciferase reporter gene (Huh7HCVꢀRep-Luc) were
maintained at 37 ꢂC in a humidified atmosphere at 5% CO2 in
Dulbecco’s modified Eagle’s medium (DMEM) containing 10% fetal
2.5.1. Cell cytotoxicity assays
Cells were maintained at 37 ꢂC in a humidified atmosphere at 5%
CO2 in DMEM containing 10% FBS. To determine the TC50 of each
compound, cells were first seeded in 96-well plates at 5,000 cells/
well, and the plates were incubated at 37 ꢂC for 24 h to allow for cell
attachment. Three biological replicates of cells were then treated
with each compound solubilized in DMSO at concentrations
bovine serum (FBS) and 250 mg/mL G418. To determine the IC50 of
each compound for the inhibition of the HCV luciferase replicon,
the Huh7HCVꢀRep-Luc cells were seeded in 96-well plates at
10,000 cells/well, and the plates were incubated at 37 ꢂC for 24 h to
allow for cell attachment. Three biological replicates of cells were
then treated with each compound at final concentrations ranging
ranging from 0 mM to 500 mM in DMEM without phenol red, con-
taining 10% FBS and 0.5% DMSO, final. For assays with the ENT1
from 0
mM to 5,000
mM in DMEM medium containing 10% FBS and
inhibitor nitrobenzylthioinosine (NBTI), cells were pretreated with
0.5% DMSO. Cells were treated with compounds and assayed at
timepoints indicated in the text. Luciferase activity in each well was
determined by using the Bright-Glo Luciferase Assay System
(Promega) and a Glowmax 96 microplate luminometer (Promega)
according to the manufacturer’s instructions. The relative luciferase
DMEM containing 10% FBS, 0.5% DMSO and 10 mM NBTI for 1 h prior
to the addition of each compound diluted in the same medium.
After 48 or 96 h of treatment at 37 ꢂC, the relative number of living
cells in each well was determined by using the CellTiter 96 system
(Promega) according to the manufacturer’s instructions. The
3

A.A. Stephenson, S. Cao, D.J. Taggart et al.
European Journal of Medicinal Chemistry 213 (2021) 113135
absorbance in each well was then determined by using a Flex-
station 3 microplate reader (Molecular Dynamics). Absorbance
values representing the relative number of live cells in each con-
dition were then plotted as a function of compound concentration,
and the data were then fit to the sigmoidal equation (3) using
Kaleidagraph (Synergy Software),
prodrugs with good yields of 60e80% (see EXPERIMENTAL SEC-
TION). After chemical synthesis, each prodrug was purified through
C-18 reverse phase preparative HPLC with a purity of >99%. Sub-
sequently, the structural identity of each purified prodrug was
analyzed and verified through both electro-spray ionization mass
spectrometry and NMR (1H and 13C) (see EXPERIMENTAL SECTION).
AxC
BC þ xC
y ¼
(3)
3.2. Inhibitory effect of gemcitabine prodrugs on HCV replicon
where y is the relative corrected absorbance value, x is the com-
pound concentration, A is the relative percent amplitude (maximal
relative inhibition), B is the concentration of the compound at
which the total live cell population is reduced to 50% (TC50), and C is
the Hill coefficient.
Inhibition of HCV replication by the gemcitabine prodrugs was
tested in human hepatocellular carcinoma (Huh7) cells harboring
an HCV replicon containing
a
luciferase reporter gene
(Huh7HCVꢀRep-Luc) [37]. Briefly, the Huh7HCVꢀRep-Luc cells were
treated with increasing concentrations of either gemcitabine (1) or
one of the prodrugs (3e30) for 48 h and then assayed for luciferase
activity. Compound 2, the intermediate in Fig. 2, was not tested. The
concentration of gemcitabine required to inhibit the HCV replicon
activity to 50% of its activity relative to untreated cells (IC50) was
found to be 58 6 nM (Table 1). Gemcitabine is hypothesized to
inhibit HCV replication through direct inhibition of HCV
nonstructural protein 5B (NS5B), an RNA-dependent RNA poly-
merase, and/or inhibition of CTP synthetase (Fig. 1). Consistent with
the mechanism of cleavage and activation of the prodrugs in the
liver-derived Huh7HCVꢀRep-Luc cells, the majority of our prodrugs
inhibited the HCV replicon with IC50 values ranging from 10 to
100 nM (Table 1). To investigate if these inhibitory prodrugs were
hydrolyzed and converted into gemcitabine by carboxylesterases
within the Huh7HCVꢀRep-Luc cells, we remeasured the IC50 value of a
2.5.2. Determining prodrug stability in aqueous solutions
To assess the pH stability of the gemcitabine prodrugs, select
prodrugs were diluted to 200 mM in aqueous solutions at pH values
of 8.0, 6.0, 4.0, 2.0, and 1.0 and incubated at 40 ꢂC for 4 h. As the IC50
and TC50 values of only compounds 1e15 had been determined
when this assay was conducted, we arbitrarily selected several
compounds to test. The intact prodrug was separated from any
potential degradation products using reverse-phase high perfor-
mance liquid chromatography (HPLC). Remaining prodrug after the
incubation was quantified relative to 200 mM untreated prodrug
using UV spectrophotometry by comparing peak area of the 260 nm
peak before and after the incubation. Remaining prodrug was
calculated using equation (4) and expressed as “% compound
remaining after 4 h”.
representative potent prodrug, 13, in the presence of either 100
mM
benzil, a specific inhibitor of human carboxyesterases 1 and 2, or
260 nm peak area without incubation
260 nm peak area after incubation
%compound remaining after 4 hrs ¼
x100%
(4)
3. Results
100
mM
bis-p-nitrophenyl phosphate (BNPP),
a
non-specific
esterase inhibitor [38e42]. The IC50 value of compound 13 was
increased from 49 9 nM (Table 1) to 125 10 and 81 6 nM in the
presence of benzil and BNPP, respectively (data not shown). This
indicates that these carboxyesterase and esterase inhibitors slowed
the activation of compound 13 in the Huh7HCVꢀRep-Luc cells, leading
to 2.5- and 1.6-fold lower HCV inhibition potency of compound 13,
respectively. Furthermore, we also incubated compound 13 with
either human carboxyesterase 1 or 2 [42] for an hour at 37 ꢂC and
detected significant amounts of gemcitabine formation by HPLC
(data not shown). Together, these abovementioned assays
demonstrate that compound 13 was efficiently metabolized into
active gemcitabine by liver carboxyesterases in the Huh7HCVꢀRep-Luc
cells. Unlike compound 13, nine out of the 28 prodrugs (6, 15, 17, 18,
19, 22, 23, 24, and 26) exhibited poor inhibition of the HCV replicon
with IC50 values of >100 nM (Table 1). This suggests that these
compounds were not good substrates for liver carboxylesterases
and were thus inefficiently converted to gemcitabine in the
Huh7HCVꢀRep-Luc cells.
3.1. Design and preparation of gemcitabine prodrugs
Human carboxylesterases are expressed at much higher levels in
liver tissue than most other tissues including intestine [32,35].
These enzymes can efficiently cleave the amide linkages in N4-
substituted 50-deoxy-5-fluocytidine derivatives including N4-aro-
matic derivatives [32,36]. Based on similar chemical structures of
50-deoxy-5-fluocytidine and gemcitabine, we designed 28 gemci-
tabine derivatives by introducing a series of aromatic moieties to
the N4 primary amine of the cytosine base ring in gemcitabine
through an amide linkage (Table 1). We predicted that the aromatic
moieties in those prodrugs would be cleaved from the parent
compound by carboxylesterases, leading to the release of active
gemcitabine predominantly in liver cells. Other than targeted drug
delivery, this prodrug approach may also protect gemcitabine from
inactivating deamination (Fig. 1). Considering that the active me-
tabolites of gemcitabine perturb cellular CTP pool and RNA/DNA
synthesis (Fig. 1), this prodrug strategy should be beneficial to the
treatment of liver maladies, such as HCV infection or hepatocellular
carcinoma.
3.3. Cytotoxicity of gemcitabine prodrugs in human liver cell line
Huh7
To synthesize the gemcitabine prodrugs (Table 1), we designed a
three-step synthetic route based on the starting material gemci-
tabine (Fig. 2). This route was used to efficiently generate the
Cytotoxicity of the gemcitabine prodrugs was determined in
Huh7 cells by measuring the concentration of each compound
4

A.A. Stephenson, S. Cao, D.J. Taggart et al.
European Journal of Medicinal Chemistry 213 (2021) 113135
Table 1
IC50 concentrations for inhibition of the HCV replicon in the Huh7HCVꢀRep-Luc cells, TC50 values in Huh7 cells, and Therapeutic Index values of gemcitabine and gemcitabine
prodrugs after 48 h treatment. Errors reported with IC50 and TC50 (i.e. the values) are fitting errors of equation (1) and equation (3), respectively.
Compound
Number
Structure
MW (g/mol)
IC50 (nM)
TC50
(
m
M)
Therapeutic Index
597
1 (gemcitabine)
263.2
58
6
35 12
2
505.7
ND
ND
ND
3
381.33
43.1 4.4
329 57
7,650
4
395.36
58.4 9.9
492 58
8,430
5
397.33
60.0 8.2
572 73
9,530
6
427.36
451 37
4580 1100
10,155
7
8
401.75
436.19
37.6 5.4
138 28
3670
77.5 7.0
>100
>1290
(continued on next page)
5

A.A. Stephenson, S. Cao, D.J. Taggart et al.
European Journal of Medicinal Chemistry 213 (2021) 113135
Table 1 (continued)
Compound
Number
Structure
MW (g/mol)
436.19
IC50 (nM)
17.4 1.8
TC50
(m
M)
Therapeutic Index
9
>100
>5750
10
11
12
436.19
466.22
427.36
25.5 2.2
19.3 1.6
73.4 12.0
110 36
4310
>100
>5180
13,100
962 110
13
457.38
49.3 8.8
495 97
10,040
14
15
16
411.31
425.34
439.11
59.8 8.4
126.7 8.6
91.5 7.6
473 74
855 166
881 287
7,910
6,790
9,628
6

A.A. Stephenson, S. Cao, D.J. Taggart et al.
European Journal of Medicinal Chemistry 213 (2021) 113135
Table 1 (continued)
Compound
Number
Structure
MW (g/mol)
437.14
IC50 (nM)
177 45
TC50
(
m
M)
Therapeutic Index
6,565
17
18
19
1162 204
954 247
2214 821
409.11
145 17
6,579
489.00
340 45
6,512
20
21
22
421.15
447.07
439.12
86.0 26.5
46.4 27.0
1690 1060
132 40
188 50
626 191
1,535
4,051
370
23
495.02
185 15
74 43
400
24
417.11
126 16
125 27
992
(continued on next page)
7

A.A. Stephenson, S. Cao, D.J. Taggart et al.
European Journal of Medicinal Chemistry 213 (2021) 113135
Table 1 (continued)
Compound
Number
Structure
MW (g/mol)
497.10
IC50 (nM)
12.8 0.7
TC50
(
m
M)
Therapeutic Index
19,450
25
249 86
26
421.1
161 46
92 39
571
27
456.1
46.7 3.9
>500
>10,700
28
29
30
447.12
407.13
407.13
93.5 18.2
10.0 0.9
58.6 24.5
168 31
1796
>500
>50,000
378 119
6710
required to reduce cellular proliferation to 50% compared to un-
treated cells (TC50). Briefly, cells were treated with various con-
centrations of either gemcitabine or one of the prodrugs for 48 h
before being assayed for proliferation. While gemcitabine inhibited
3.4. Cytotoxicity of gemcitabine prodrugs in other human cancer
cell lines
To investigate the cytotoxicity of the prodrugs in cell lines
derived from other tissues, the TC50 of gemcitabine and select
prodrugs were determined in cell lines derived from liver (Huh7
and HepG2), kidney (HEK293), breast (MCF-7), and pancreatic
(BxPC3) tissues after 96 h continuous treatment (Table 2). As the
IC50 and TC50 values of only compounds 1e15 had been determined
when this assay was conducted, we arbitrarily selected several
prodrugs to compare. In liver- and kidney-derived cell lines, which
express higher levels of carboxylesterases, the select prodrugs
exhibited similar (within one order of magnitude) cytotoxicity as
the proliferation of the Huh7 cells with a TC50 of 35 12
majority of the prodrugs inhibited the proliferation of the Huh7
cells with a TC50 values > 300 M (Table 1). Thus, although many of
mM, the
m
the prodrugs retained similar or lower IC50 values when compared
to the parent compound, most of the prodrugs were less cytotoxic
than gemcitabine alone. This reduced cytotoxicity of the gemcita-
bine prodrugs resulted in calculated therapeutic index values of
370 to >50,000 (Table 1) for inhibition of the HCV replicon within
the Huh7HCVꢀRep-Luc cells.
8

A.A. Stephenson, S. Cao, D.J. Taggart et al.
European Journal of Medicinal Chemistry 213 (2021) 113135
Table 2
Comparison of the cytotoxicity of gemcitabine and gemcitabine prodrugs after 96 h of treatment in various human cell lines. Errors (i.e. the values) are fitting errors of
equation (3).
TC50 (mM) after 96 h continuous treatment
Compound
Huh7 (Liver)
HepG2 (Liver)
HEK293 (Kidney)
MCF-7 (Breast)
BxPC3 (Pancreas)
Gemcitabine (1)
3.4 1.5
4.2 1.1
18.5 5.3
5.1 1.2
7.0 1.4
3.3 1.0
0.31 0.055
0.60 0.20
3.0 1.4
1.7 0.40
2.8 0.50
1.1 0.30
0.56 0.13
1.2 0.18
4.0 0.79
1.3 0.26
4.1 0.73
5.6 0.80
0.067 0.014
0.33 0.088
1.6 0.34
0.57 0.12
1.2 0.33
0.0029 0.00030
ND
ND
1.8 0.43
0.72 0.18
1.2 0.23
10
12
13
14
15
1.4 0.40
gemcitabine after 96 h continuous treatment (Table 2). However,
the prodrugs were less cytotoxic than gemcitabine in cell lines
derived from breast (up to 24-fold) and pancreatic (up to 620-fold)
tissues, which produce less carboxylesterases than liver and kidney
cells (Table 2). These findings are consistent with the above
conclusion that the prodrugs are mainly hydrolyzed by carbox-
ylesterases to intracellularly produce gemcitabine [32,43].
3.5. Inhibitory mechanism of gemcitabine prodrugs on HCV
replicon
To explore the mechanism by which the gemcitabine prodrugs
inhibited the HCV replicon, the Huh7HCVꢀRep-Luc cells were treated
with either 1,000, 100, or 25 nM of gemcitabine or a prodrug either
Fig. 4. Isobologram of compound 13 and daclatasvir. IC50 values of compound 13
(circles) were measured at various sub-IC50 concentrations of daclatasvir. Individual
IC50 values of compound 13 and daclatasvir alone are plotted as squares on the axes
and the expected IC50 values of compound 13 after combined treatment with the
corresponding daclatasvir concentration without synergy between the two com-
pounds is given by the straight line. Error bars are fitting errors of equation (1).
alone or in the presence of 50 mM cytidine or deoxycytidine and
then assayed for luciferase activity after 96 h continuous treatment
(Fig. 3). As the IC50 and TC50 values of only compounds 1e15 had
been determined when this assay was conducted, we arbitrarily
selected several prodrugs to test. Because gemcitabine is a cytidine
analog, excess exogenous cytidine or deoxycytidine was predicted
to reduce the inhibitory effects of gemcitabine in cells (Fig. 1).
Consistently, the HCV inhibition activity of gemcitabine and the
gemcitabine prodrugs was reduced significantly by exogenous
cytidine and less significantly by deoxycytidine (Fig. 3). We
concluded that the gemcitabine prodrugs inhibited replication of
the HCV replicon mainly through the perturbation of the cellular
CTP pools as well as direct inhibition of HCV RNA synthesis as the
parent compound gemcitabine (Fig. 1).
As the gemcitabine prodrugs likely act through inhibition of
NS5B, we next asked whether the prodrugs could have synergistic
effects with anti-HCV drugs that target other HCV proteins.
Daclatasvir is an approved drug for HCV treatment and acts through
inhibiting hyperphosphorylation of the non-enzymatic HCV
Fig. 3. The anti-HCV activities of gemcitabine and gemcitabine prodrugs can be reversed by exogenous nucleosides. The Huh7HCVꢀRep-Luc cells were mock treated (gray) or treated
with various concentrations (25e1,000 nM) of gemcitabine (1, black), 10 (red), 13 (blue) or 12 (green) alone (open bars), in the presence of 50 M cytidine (C, striped bars), or in the
presence of 50
M 20-deoxycytidine (dC, solid bars). After 48 h of continuous treatment, the cells assayed for luciferase activity. Error bars are standard deviation values calculated
from three biological replicates.
m
m
9

A.A. Stephenson, S. Cao, D.J. Taggart et al.
European Journal of Medicinal Chemistry 213 (2021) 113135
protein NS5A [44]. Because NS5A directly interacts with NS5B and
regulates its activity, we hypothesized that daclatasvir and the
gemcitabine prodrugs would exhibit synergistic inhibition of HCV
replication [45]. To test this hypothesis, we first measured the IC50
value of daclatasvir against the HCV replicon after 48 h of treatment
to be 0.031 0.003 nM. Next, IC50 values of compound 13 were
measured in the Huh7HCVꢀRep-Luc cells via 48 h of treatment in the
presence of a series of sub-IC50 concentrations of daclatasvir and
the resulting data were plotted as an isobologram (Fig. 4) [46]. We
selected compound 13 for this synergism assay because it exhibited
both potent inhibition of HCV replicon and a high therapeutic index
value (Table 1) among compounds 1e15, while compounds 16e30
had not yet been synthesized. We found that the combined effect of
daclatasvir and compound 13 was more potent than the additive
effect predicted by individual IC50 values of the compounds alone
(Fig. 4, straight line) with experimental combination indexes
(experimental IC50 divided by predicted IC50 without synergy) of
<1, indicating synergism between the gemcitabine prodrug and
daclatasvir [34]. Interestingly, similar isobologram analysis shows
synergistic effect between compound 13 and PSI-6130 (data not
shown), a known HCV NS5B-inhibiting nucleoside analogue [47].
To further investigate inhibition of the HCV replicon by the
gemcitabine prodrugs, a time course of HCV replicon inhibition by
gemcitabine or a prodrug was conducted. The Huh7HCVꢀRep-Luc cells
were either mock treated or treated with 500 nM gemcitabine or a
prodrug for up to 48 h. We chose to only compare compound 10 or
13 to gemcitabine as we felt that they were representative of two
different types of aryl modifications among our library of com-
pounds 1e15, while compounds 16e30 had not yet been synthe-
sized when this assay was conducted. Gemcitabine treatment
resulted in 50% inhibition of the HCV replicon within 6 h with
maximal inhibition at 36 h (Fig. 5). In contrast, the select prodrugs
inhibited the luciferase activity of the HCV replicon to 50% between
24 and 30 h and did not reach maximum inhibition until 48 h of
treatment (Fig. 5). We concluded that this delay in the inhibitory
effect of the prodrugs relative to gemcitabine was due to slower
cellular import and/or the time needed to hydrolyze the prodrugs
to produce active gemcitabine by cellular carboxyesterases.
To evaluate cellular import of the gemcitabine prodrugs, the IC50
of gemcitabine or select prodrugs were determined by either
continuously treating the Huh7HCVꢀRep-Luc cells with each com-
pound for 48 h or treating the Huh7HCVꢀRep-Luc cells for only 2 h
followed by replacement with fresh growth medium for 46 h. We
chose to only compare compounds 13 and 14 to gemcitabine as we
felt that they were potent HCV inhibitors with high therapeutic
index values (Table 1) among compounds 1e15, while compounds
16e30 had not yet been synthesized when this assay was con-
ducted. Treatment for 2 h versus 48 h with gemcitabine had only a
small impact (2.9-fold) on the IC50 value, suggesting that gemci-
tabine was rapidly imported into the Huh7HCVꢀRep-Luc cells. How-
ever, treatment for only 2 h versus 48 h had a strong negative
impact (51- to 56-fold) on the IC50 values of the select prodrugs
(Table 3). We concluded that the select gemcitabine prodrugs enter
cells more slowly than the parent compound gemcitabine.
Gemcitabine primarily enters cells through the action of ENT1
(Fig. 1) [11,12]. However, it is possible that the gemcitabine pro-
drugs, which are conjugated to large hydrophobic moieties, cannot
be recognized and imported by ENT1, which would account for
their slower cellular import and higher IC50 values than gemcita-
bine (Table 3). To test this hypothesis, Huh7 cells were treated with
gemcitabine or select prodrugs either alone or in the presence of
the ENT1 inhibitor nitrobenzylthioinosine (NBTI) and then assayed
for cytotoxicity. We chose to only compare compounds 3, 4, 13, and
14 to gemcitabine as we felt that they were representative of major
types of aryl modifications among our library of compounds 1e15,
while compounds 16e30 had not yet been synthesized when this
assay was conducted. Consistent with the importance of ENT1 for
the cellular import of gemcitabine, the TC50 of gemcitabine
decreased more than 14-fold in the presence of NBTI (Table 4). In
contrast, the TC50 values of the select prodrugs were insignificantly
changed (Table 4). These results strongly suggest that the gemci-
tabine prodrugs do not mainly utilize ENT1 for cellular import.
3.6. Stability test of select gemcitabine prodrugs
Stability over a wide range of pH values is a requirement for
bioavailability of a drug with oral administration. We initially
investigated the pH stability of gemcitabine and four representative
prodrugs (10, 13, 14, and 15) out of compounds 1e15 (Table 1) while
compounds 16e30 had not yet been synthesized. To assess the pH
stability of each of the select compounds, the prodrug (200 mM) was
incubated at 40 ꢂC for 4 h in buffered solutions with pH values of
1.0e8.0 to simulate the pH values of the human digestive system.
The relative amount of each intact prodrug was then determined by
separating the intact prodrug from any degradation products using
reverse-phase HPLC and subsequently quantifying the relative
amount of remaining prodrug by spectrophotometry (see section
2.5.2).
Compounds 13, 14, and 15 exhibited similar stability as gemci-
tabine with less than 20% degradation over 4 h even at very low pH
(Table 5). In comparison, compound 10 was mostly degraded at pH
values of 1.0 and 8.0 (Table 5). Taken together, several prodrugs in
Fig. 5. Time courses of HCV luciferase replicon inhibition by gemcitabine and select
gemcitabine prodrugs. The Huh7HCVꢀRep-Luc cells were mock treated (black) or treated
with gemcitabine (1, blue), 10 (red), 13 (green). At various times, the cells were lysed
and the cell lysates were immediately frozen in liquid nitrogen. The cell lysates were
then assayed for luciferase activity simultaneously.
Table 3
Comparison of HCV replicon inhibition in the Huh7HCVꢀRep-Luc cells by gemcitabine and gemcitabine prodrugs during continuous treatment or 2 h of treatment. Errors values
(
) are fitting errors of equation (1).
Compound
IC50 (nM), treated for 2 h, assayed at 48 h
IC50 (nM), 48 h continuous treatment
Ratio 2 h treatment to 48 h treatment
Gemcitabine (1)
13
14
170 33
2,800 270
3,100 560
58.4 6.1
49.3 8.8
59.8 8.4
2.9
56
51
10

A.A. Stephenson, S. Cao, D.J. Taggart et al.
European Journal of Medicinal Chemistry 213 (2021) 113135
Table 4
a well-characterized downregulation mechanism that depletes
cellular expression of hENT1 and hENT2, which are required for
gemcitabine membrane transport. Modification of nucleoside an-
alogs to generate rationally designed prodrugs remains a highly
active area of research as it can confer tissue-specificity, evasion of
known resistance mechanisms, and other functions [48].
Comparison of the cytotoxicity of gemcitabine and gemcitabine prodrugs in Huh7
cells after 48 h of treatment in the presence or absence of the ENT1 inhibitor NBTI.
Errors ( ) are fitting errors of equation (3).
TC50 (mM) after 48h continuous treatment
Compound
No NBTI
10 mM NBTI
Here, liver-targeting gemcitabine prodrugs were generated by
conjugation of the primary amine on its nitrogenous base to various
aromatic groups through an amide linkage (Fig. 2). We showed that
several of the gemcitabine aryl-amide prodrugs were stable at a
broad range of buffer pH (Table 5) and bypassed the requirement of
hENT1-mediated membrane transport of gemcitabine (Table 4).
These results support the conclusion that some of the prodrugs in
Table 1 are promising leads for chemotherapeutic development
considering that they evade cellular resistance to gemcitabine in
liver cancer cells with low hENT1 and hENT2 expression. Similarly,
other chemical modifications of the 4-NH2, 30-OH, and/or 50-OH of
Gemcitabine (1)
3
4
13
14
35 12
>500
330 60
490 50
490 100
470 70
240 40
360 80
398 44
440 190
Table 5
Degradation of gemcitabine and gemcitabine prodrugs at 200
at 40 ꢂC for 4 h.
m
M in aqueous buffer
% compound remaining after 4 h
gemcitabine,
e.g.
20-deoxy-20,20-difluoro-N-(1-oxo-2-
Condition
Gemcitabine (1)
10
13
14
15
propylpentyl)-cytidine (LY2334737), have generated promising
gemcitabine prodrugs which exhibit enhanced efficacy, stability,
and oral bioavailability than gemcitabine in vitro and in vivo
[43,48e53].
Before treatment
100
79
88
98
100
100
100
35
82
98.4
99.6
35.4
100
86
87
98.6
99.5
82.7
100
64
100
100
100
92
100
71
100
100
100
100
pH 1
pH 2
pH 4
pH 6
pH 8
We demonstrated that several gemcitabine prodrugs exhibited
much lower cytotoxicity (up to 620-fold) when compared to
gemcitabine with breast- and pancreas-derived cells lines (Table 2).
However, these prodrugs had comparable (<10-fold different)
cytotoxicity in liver- and kidney-derived cell lines (Table 2). These
differences in cytotoxicity are likely related to higher carbox-
ylesterase expression levels in liver and kidney cells relative to
breast and pancreas cells, suggesting that the cytotoxic effects of
the prodrugs require their hydrolysis in vivo. These differences also
support the postulation that the gemcitabine prodrugs likely
exhibit similar inhibitory efficacy against liver cancer with less
systemic toxicity in vivo. Future translational studies will further
evaluate the therapeutic potential of these gemcitabine prodrugs
for treatment of liver cancer.
Our direct observation of gemcitabine formation from the hy-
drolysis of compound 13 catalyzed by recombinant human car-
boxylesterases and the 2.5-fold higher IC50 value of compound 13 in
the presence than in the absence of the specific carboxylesterase
inhibitor benzil clearly indicate that cellular carboxylesterases play
a major role in cellular activation of gemcitabine prodrugs. This
conclusion is consistent with previous findings that liver carbox-
ylesterases are responsible for activating another gemcitabine
prodrug LY2334737 [54] and N4-alkoxy-50-deoxy-5-fluorocytidine
prodrugs [36] in the liver. Interestingly, several prodrugs (6, 19,
22, and 27), containing bulky substitution groups near the amide
linkage, consistently exhibited greatly reduced cytotoxicity in liver-
derived Huh7HCVꢀRep-Luc cells. Likely, these bulky groups decreased
cellular activation efficiency of compounds 6, 19, 22, and 27 relative
to the other prodrugs in Table 1 by hindering the binding and hy-
drolysis of the four prodrugs at the active sites of liver carbox-
ylesterases. Indeed, carboxylesterases require access to the
carbonyl center of the amide linkage to initiate a nucleophilic attack
via a catalytic serine residue [55]. Thus, the access was hampered
by bulky methoxy, methyl, nitro, bromo, or dioxole groups near the
carbonyl center of compounds 6,19, 22, and 27. Further mechanistic
investigation of carboxylesterase-catalyzed prodrug hydrolysis by
using in vitro enzymatic assays and metabolite analysis is needed to
unveil the structure-activity relationship in the gemcitabine pro-
drugs described in Table 1.
Table 1 were stable under a wide range of pH values and thus are
expected to exhibit good oral bioavailability.
4. Discussion and conclusions
Liver cancer and HCV-driven chronic liver disease are leading
causes of liver-related illness and death world-wide. Resistance to
anti-viral and chemotherapeutic drugs is considered the greatest
challenge in the development of therapies to treat liver cancer and
HCV infection. Gemcitabine is among the first-line chemotherapies
used for treatment of metastatic or non-operable, solid-tumor liver
cancers and is cytotoxic to rapidly growing cell types like cancer.
Gemcitabine exhibits a dual mechanism of cytotoxicity through i)
direct incorporation into nascent DNA or RNA chains, causing pre-
mature termination of nucleic acid chain elongation, and ii)
depletion of nucleotide triphosphate pools through competitive
enzyme inhibition, leading to more gemcitabine triphosphate
incorporation. The self-potentiation mechanism of replication in-
hibition (Fig. 1) contributes to the observed anti-cancer and anti-
viral activities of gemcitabine in vitro and in vivo. Investigation
into the structure-activity relationship of gemcitabine has led to a
multi-mechanistic mode of action involving two different nucleo-
tide phosphate metabolites of gemcitabine, dFdCDP and dFdCTP.
Competitive binding to ribonucleotide reductase by dFdCDP re-
duces conversion of CDP to dCDP and thus leads to depletion in
cellular dCTP for DNA synthesis (Fig. 1). dFdCTP can both directly
mediate premature termination of DNA and RNA synthesis via
incorporation into nascent nucleic acid chains by many DNA and
RNA polymerases as well as deplete the cellular CTP pool by
inhibiting CTP synthetase which catalyzes the conversion of UTP to
CTP (Fig. 1).
Due to the ubiquitous roles of nucleotides and DNA and RNA
synthesis in cellular metabolism and genomic replication, gemci-
tabine phosphates exhibit broad toxic effects in patients including
hair loss, mouth sores, fatigue, loss of appetite, nausea, diarrhea, as
well as low white and red blood cell counts. These side effects
negatively impact patients’ health and quality of life and should be
considered in future development of chemotherapeutics. Moreover,
cancer cells can rapidly develop resistance to gemcitabine through
Notably, many of the prodrugs (6, 12, 13, 25, 27, and 29) were
shown to be potent inhibitors of HCV replication with therapeutic
index values > 10,000 as demonstrated by HCV replicon inhibition
assays and Huh7-based cytotoxicity assays (Table 1). The HCV
11

A.A. Stephenson, S. Cao, D.J. Taggart et al.
European Journal of Medicinal Chemistry 213 (2021) 113135
[2] Global Hepatitis Report 2017, World Health Organization, Geneva, 2017.
Licence: CC BY-NC-SA 3.0 IGO.
[3] G. Chander, M.S. Sulkowski, M.W. Jenckes, M.S. Torbenson, H.F. Herlong,
E.B. Bass, K.A. Gebo, Treatment of chronic hepatitis C: a systematic review,
Hepatology 36 (2002) S135eS144.
replicon inhibition assay used here is a robust and rapid read-out of
the anti-viral activities of gemcitabine prodrugs in vitro as the HCV
replicon can be inhibited at much lower concentrations and thus
avoid some of the confounding effects of cytotoxicity and enable
more robust evaluation of their anti-viral mechanisms in compar-
ison with gemcitabine. The anti-viral effects of gemcitabine and its
prodrugs were reversed by abundance of extracellular CTP or dCTP,
suggesting that the prodrugs have a similar mechanism of action as
outlined for gemcitabine in Fig. 1. However, the dramatic increases
in the therapeutic index values for 25 out of 28 prodrugs compared
to gemcitabine (Table 1) cannot be entirely explained by differences
in their cell permeability and carboxylesterase hydrolysis efficiency
as cellular uptake and the activating cleavage reaction are both
required for cytotoxicity and HCV replicon inhibition based on the
mechanisms in Fig. 1. This means that the gemcitabine prodrugs
may have utilized other anti-viral mechanisms such as direct in-
hibition of HCV proteins and/or conversion to alternative anti-viral
metabolites. For example, the most potent prodrug 29 contains a
dihydroindenyl modification linked via the pentyl ring, which
enhanced the anti-HCV activity by ~5-fold compared to the parent
gemcitabine but reduced cytotoxicity by more than 10-fold
(Table 1). The very similar dihydroindenyl modification in com-
pound 30, linked via the aromatic moiety instead, also reduced
cytotoxicity by ~10-fold but did not enhance the anti-HCV activity
relative to gemcitabine (Table 1). It is possible that certain aspects
of the anti-HCV activity of compound 29 were mediated by direct
binding of the prodrug and/or its unknown metabolite(s) to HCV
replication proteins, such as NS3 (RNA helicase/protease), NS5A
(interacting with NS5B), or NS5B (RNA-dependent RNA polymer-
ase). Future in vitro enzyme inhibition investigations, molecular
docking simulations, and prodrug metabolite studies may help
divulge the inhibitory mechanism(s) of compound 29 and guide
rational design of more potent anti-HCV gemcitabine prodrugs.
Taken together, our data in this paper provide proof-of-concept for
future development of gemcitabine aryl-amide prodrugs for the
treatment of HCV infection and hepatocellular carcinoma.
[4] R.L. Siegel, K.D. Miller, A. Jemal, Cancer statistics, CA A Cancer J. Clin. 65 (2015)
5e29.
[5] H.A. Burris 3rd, M.J. Moore, J. Andersen, M.R. Green, M.L. Rothenberg,
M.R. Modiano, M.C. Cripps, R.K. Portenoy, A.M. Storniolo, P. Tarassoff,
R. Nelson, F.A. Dorr, C.D. Stephens, D.D. Von Hoff, Improvements in survival
and clinical benefit with gemcitabine as first-line therapy for patients with
advanced pancreas cancer: a randomized trial, J. Clin. Oncol. : Off. J. Am. Soc.
Clin. Oncol. 15 (1997) 2403e2413.
[6] A. Sandler, D.S. Ettinger, Gemcitabine: single-agent and combination therapy
in non-small cell lung cancer, Oncol. 4 (1999) 241e251.
[7] D.A. Yardley, Gemcitabine plus paclitaxel in breast cancer, Semin. Oncol. 32
(2005) S14eS21.
[8] M.D. Shelley, G. Jones, A. Cleves, T.J. Wilt, M.D. Mason, H.G. Kynaston, Intra-
vesical gemcitabine therapy for non-muscle invasive bladder cancer (NMIBC):
a systematic review, BJU Int. 109 (2012) 496e505.
[9] R.F. Ozols, Gemcitabine and carboplatin in second-line ovarian cancer, Semin.
Oncol. 32 (2005) S4eS8.
[10] J. Zhang, F. Visser, K.M. King, S.A. Baldwin, J.D. Young, C.E. Cass, The role of
nucleoside transporters in cancer chemotherapy with nucleoside drugs, Canc.
Metastasis Rev. 26 (2007) 85e110.
[11] R. Mori, T. Ishikawa, Y. Ichikawa, K. Taniguchi, R. Matsuyama, M. Ueda, Y. Fujii,
I. Endo, S. Togo, P.V. Danenberg, H. Shimada, Human equilibrative nucleoside
transporter 1 is associated with the chemosensitivity of gemcitabine in hu-
man pancreatic adenocarcinoma and biliary tract carcinoma cells, Oncol. Rep.
17 (2007) 1201e1205.
[12] J. Garcia-Manteiga, M. Molina-Arcas, F.J. Casado, A. Mazo, M. Pastor-Anglada,
Nucleoside transporter profiles in human pancreatic cancer cells: role of
hCNT1 in 2’,2’-difluorodeoxycytidine- induced cytotoxicity, Clin. Canc. Res. :
Off. J. Am. Assoc. Canc. Res. 9 (2003) 5000e5008.
[13] K.A. Richardson, T.P. Vega, F.C. Richardson, C.L. Moore, J.C. Rohloff,
B. Tomkinson, R.A. Bendele, R.D. Kuchta, Polymerization of the triphosphates
of AraC, 2’,2’-difluorodeoxycytidine (dFdC) and OSI-7836 (T-araC) by human
DNA polymerase alpha and DNA primase, Biochem. Pharmacol. 68 (2004)
2337e2346.
[14] V.W. Ruiz van Haperen, G. Veerman, J.B. Vermorken, G.J. Peters, 2’,2’-Difluoro-
deoxycytidine (gemcitabine) incorporation into RNA and DNA of tumour cell
lines, Biochem. Pharmacol. 46 (1993) 762e766.
[15] P. Huang, S. Chubb, L.W. Hertel, G.B. Grindey, W. Plunkett, Action of 2’,2’-
difluorodeoxycytidine on DNA synthesis, Canc. Res. 51 (1991) 6110e6117.
[16] J.D. Fowler, J.A. Brown, K.A. Johnson, Z. Suo, Kinetic investigation of the
inhibitory effect of gemcitabine on DNA polymerization catalyzed by human
mitochondrial DNA polymerase, J. Biol. Chem. 283 (2008) 15339e15348.
[17] J. Shao, B. Zhou, B. Chu, Y. Yen, Ribonucleotide reductase inhibitors and future
drug design, Curr. Cancer Drug Targets 6 (2006) 409e431.
[18] W. Plunkett, P. Huang, C.E. Searcy, V. Gandhi, Gemcitabine: preclinical phar-
macology and mechanisms of action, Semin. Oncol. 23 (1996) 3e15.
[19] J.D. Davidson, L. Ma, M. Flagella, S. Geeganage, L.M. Gelbert, C.A. Slapak, An
increase in the expression of ribonucleotide reductase large subunit 1 is
associated with gemcitabine resistance in non-small cell lung cancer cell lines,
Canc. Res. 64 (2004) 3761e3766.
[20] V. Heinemann, L. Schulz, R.D. Issels, W. Plunkett, Gemcitabine: a modulator of
intracellular nucleotide and deoxynucleotide metabolism, Semin. Oncol. 22
(1995) 11e18.
Author contributions
zS.C., A.A.S., and D.J.T. performed chemical synthesis and all
other experiments in the manuscript and contributed equally to the
manuscript. V.P.V. chemically synthesized several gemcitabine
prodrugs in Table 1. A.A.S., D.J.T., and Z.S. wrote the paper. All au-
thors have approved the final article.
[21] T. Neff, C.A. Blau, Forced expression of cytidine deaminase confers resistance
to cytosine arabinoside and gemcitabine, Exp. Hematol. 24 (1996)
1340e1346.
Funding sources
[22] W. Plunkett, P. Huang, Y.Z. Xu, V. Heinemann, R. Grunewald, V. Gandhi,
Gemcitabine: metabolism, mechanisms of action, and self-potentiation,
Semin. Oncol. 22 (1995) 3e10.
[23] V. Heinemann, Y.Z. Xu, S. Chubb, A. Sen, L.W. Hertel, G.B. Grindey, W. Plunkett,
Cellular elimination of 2’,2’-difluorodeoxycytidine 5’-triphosphate: a mecha-
nism of self-potentiation, Canc. Res. 52 (1992) 533e539.
[24] J.R. Kroep, W.J. Loves, C.L. van der Wilt, E. Alvarez, I. Talianidis, E. Boven,
B.J. Braakhuis, C.J. van Groeningen, H.M. Pinedo, G.J. Peters, Pretreatment
deoxycytidine kinase levels predict in vivo gemcitabine sensitivity, Mol. Canc.
Therapeut. 1 (2002) 371e376.
[25] Y. Saiki, Y. Yoshino, H. Fujimura, T. Manabe, Y. Kudo, M. Shimada, N. Mano,
T. Nakano, Y. Lee, S. Shimizu, S. Oba, S. Fujiwara, H. Shimizu, N. Chen,
Z.K. Nezhad, G. Jin, S. Fukushige, M. Sunamura, M. Ishida, F. Motoi, S. Egawa,
M. Unno, A. Horii, DCK is frequently inactivated in acquired gemcitabine-
resistant human cancer cells, Biochem. Biophys. Res. Commun. 421 (2012)
98e104.
This work was supported by a research grant from the Drug
Discovery Institute at The Ohio State University to Z.S.
Declaration of competing interest
The authors declare that they have no known competing
financial interests or personal relationships that could have
appeared to influence the work reported in this paper.
Appendix A. Supplementary data
Supplementary data to this article can be found online at
https://doi.org/10.1016/j.ejmech.2020.113135.
[26] E. Moysan, G. Bastiat, J.P. Benoit, Gemcitabine versus Modified Gemcitabine: a
review of several promising chemical modifications, Mol. Pharm. 10 (2012)
430e444.
[27] R. Cohen, L.H. Preta, V. Joste, E. Curis, O. Huillard, A. Jouinot, C. Narjoz,
A. Thomas-Schoemann, A. Bellesoeur, M. Tiako Meyo, J. Quilichini, D. Desaulle,
I. Nicolis, A. Cessot, M. Vidal, F. Goldwasser, J. Alexandre, B. Blanchet, De-
terminants of the interindividual variability in serum cytidine deaminase
activity of patients with solid tumours, Br. J. Clin. Pharmacol. 85 (2019)
References
[1] R. Bartenschlager, F. Penin, V. Lohmann, P. Andre, Assembly of infectious
hepatitis C virus particles, Trends Microbiol. 19 (2011) 95e103.
12

A.A. Stephenson, S. Cao, D.J. Taggart et al.
European Journal of Medicinal Chemistry 213 (2021) 113135
1227e1238.
J. Warmus, A.S. Kalgutkar, Utility of the carboxylesterase inhibitor bis-para-
nitrophenylphosphate (BNPP) in the plasma unbound fraction determina-
tion for a hydrolytically unstable amide derivative and agonist of the TGR5
receptor, Xenobiotica 40 (2010) 369e380.
[28] N. Simon, O. Romano, P. Michel, C. Pincon, M. Vasseur, N. Lemahieu,
C. Barthelemy, M. Hebbar, B. Decaudin, P. Odou, Influence of infusion method
on gemcitabine pharmacokinetics: a controlled randomized multicenter trial,
Canc. Chemother. Pharmacol. 76 (2015) 865e871.
[29] J.H. Beumer, J.L. Eiseman, R.A. Parise, E. Joseph, J.M. Covey, M.J. Egorin, Mod-
ulation of gemcitabine (2’,2’-difluoro-2’-deoxycytidine) pharmacokinetics,
metabolism, and bioavailability in mice by 3,4,5,6-tetrahydrouridine, Clin.
Canc. Res. : Off. J. Am. Assoc. Canc. Res. 14 (2008) 3529e3535.
[30] J.R. Kroep, C.J. van Moorsel, G. Veerman, D.A. Voorn, R.M. Schultz, J.F. Worzalla,
L.R. Tanzer, R.L. Merriman, H.M. Pinedo, G.J. Peters, Role of deoxycytidine
kinase (dCK), thymidine kinase 2 (TK2), and deoxycytidine deaminase (dCDA)
in the antitumor activity of gemcitabine (dFdC), Adv. Exp. Med. Biol. 431
(1998) 657e660.
[42] E.V. Pindel, N.Y. Kedishvili, T.L. Abraham, M.R. Brzezinski, J. Zhang, R.A. Dean,
W.F. Bosron, Purification and cloning of a broad substrate specificity human
liver carboxylesterase that catalyzes the hydrolysis of cocaine and heroin,
J. Biol. Chem. 272 (1997) 14769e14775.
[43] S. Hong, Z. Fang, H.Y. Jung, J.H. Yoon, S.S. Hong, H.J. Maeng, Synthesis of
gemcitabine-threonine amide prodrug effective on pancreatic cancer cells
with improved pharmacokinetic properties, Molecules 23 (2018).
[44] C. Lee, Daclatasvir: potential role in hepatitis C, Drug Des. Dev. Ther. 7 (2013)
1223e1233.
[45] Y. Shirota, H. Luo, W. Qin, S. Kaneko, T. Yamashita, K. Kobayashi, S. Murakami,
Hepatitis C virus (HCV) NS5A binds RNA-dependent RNA polymerase (RdRP)
NS5B and modulates RNA-dependent RNA polymerase activity, J. Biol. Chem.
277 (2002) 11149e11155.
[46] R.J. Tallarida, Quantitative methods for assessing drug synergism, Genes Canc.
2 (2011) 1003e1008.
[31] M. Gusella, F. Pasini, C. Bolzonella, S. Meneghetti, C. Barile, A. Bononi, S. Toso,
D. Menon, G. Crepaldi, Y. Modena, L. Stievano, R. Padrini, Equilibrative
nucleoside transporter 1 genotype, cytidine deaminase activity and age pre-
dict gemcitabine plasma clearance in patients with solid tumours, Br. J. Clin.
Pharmacol. 71 (2011) 437e444.
[32] S.C. Laizure, V. Herring, Z. Hu, K. Witbrodt, R.B. Parker, The role of human
carboxylesterases in drug metabolism: have we overlooked their importance?
Pharmacotherapy 33 (2013) 210e222.
[33] C. Aldrich, C. Bertozzi, G.I. Georg, L. Kiessling, C. Lindsley, D. Liotta,
K.M. Merz Jr., A. Schepartz, S. Wang, The ecstasy and agony of assay inter-
ference compounds, ACS Cent. Sci. 3 (2017) 143e147.
[34] T.C. Chou, Theoretical basis, experimental design, and computerized simula-
tion of synergism and antagonism in drug combination studies, Pharmacol.
Rev. 58 (2006) 621e681.
[35] T. Satoh, M. Hosokawa, The mammalian carboxylesterases: from molecules to
functions, Annu. Rev. Pharmacol. Toxicol. 38 (1998) 257e288.
[36] N. Shimma, I. Umeda, M. Arasaki, C. Murasaki, K. Masubuchi, Y. Kohchi,
M. Miwa, M. Ura, N. Sawada, H. Tahara, I. Kuruma, I. Horii, H. Ishitsuka, The
design and synthesis of a new tumor-selective fluoropyrimidine carbamate,
capecitabine, Bioorg. Med. Chem. 8 (2000) 1697e1706.
[37] J.M. Vrolijk, A. Kaul, B.E. Hansen, V. Lohmann, B.L. Haagmans, S.W. Schalm,
R. Bartenschlager, A replicon-based bioassay for the measurement of in-
terferons in patients with chronic hepatitis C, J. Virol Methods 110 (2003)
201e209.
[38] J.L. Hyatt, L. Tsurkan, M. Wierdl, C.C. Edwards, M.K. Danks, P.M. Potter,
Intracellular inhibition of carboxylesterases by benzil: modulation of CPT-11
cytotoxicity, Mol. Canc. Therapeut. 5 (2006) 2281e2288.
[47] J.L. Clark, L. Hollecker, J.C. Mason, L.J. Stuyver, P.M. Tharnish, S. Lostia,
T.R. McBrayer, R.F. Schinazi, K.A. Watanabe, M.J. Otto, P.A. Furman, W.J. Stec,
S.E. Patterson, K.W. Pankiewicz, Design, synthesis, and antiviral activity of 2’-
deoxy-2’-fluoro-2’-C-methylcytidine, a potent inhibitor of hepatitis C virus
replication, J. Med. Chem. 48 (2005) 5504e5508.
[48] H. Miao, X. Chen, Y. Luan, Small molecular gemcitabine prodrugs for cancer
therapy, Curr. Med. Chem. 27 (2020) 5562e5582.
[49] E. Wickremsinhe, J. Bao, R. Smith, R. Burton, S. Dow, E. Perkins, Preclinical
absorption, distribution, metabolism, and excretion of an oral amide prodrug
of gemcitabine designed to deliver prolonged systemic exposure, Pharma-
ceutics 5 (2013) 261e276.
[50] Y.X. Zhang, Y.K. Gao, X.J. Wen, H.Y. Ma, Current prodrug strategies for
improving oral absorption of nucleoside analogues, Asian J. Pharm. Sci. 9
(2014) 65e74.
[51] Y. Tsume, T. Incecayir, X. Song, J.M. Hilfinger, G.L. Amidon, The development of
orally administrable gemcitabine prodrugs with D-enantiomer amino acids:
enhanced membrane permeability and enzymatic stability, Eur. J. Pharm.
Biopharm. 86 (2014) 514e523.
[52] X. Song, P.L. Lorenzi, C.P. Landowski, B.S. Vig, J.M. Hilfinger, G.L. Amidon,
Amino acid ester prodrugs of the anticancer agent gemcitabine: synthesis,
bioconversion, metabolic bioevasion, and hPEPT1-mediated transport, Mol.
Pharm. 2 (2005) 157e167.
[39] R.M. Wadkins, J.L. Hyatt, X. Wei, K.J. Yoon, M. Wierdl, C.C. Edwards,
C.L. Morton, J.C. Obenauer, K. Damodaran, P. Beroza, M.K. Danks, P.M. Potter,
Identification and characterization of novel benzil (diphenylethane-1,2-dione)
analogues as inhibitors of mammalian carboxylesterases, J. Med. Chem. 48
(2005) 2906e2915.
[40] H. Buch, W. Buzello, E. Heymann, K. Krisch, Inhibition of phenacetin- and
acetanilide-induced methemoglobinemia in the rat by the carboxylesterase
inhibitor bis-[p-nitrophenyl] phosphate, Biochem. Pharmacol. 18 (1969)
801e811.
[53] G. Wang, H. Chen, D. Zhao, D. Ding, M. Sun, L. Kou, C. Luo, D. Zhang, X. Yi,
J. Dong, J. Wang, X. Liu, Z. He, J. Sun, Combination of l-carnitine with lipophilic
linkage-donating gemcitabine derivatives as intestinal novel organic cation
transporter 2-targeting oral prodrugs, J. Med. Chem. 60 (2017) 2552e2561.
[54] S.E. Pratt, S. Durland-Busbice, R.L. Shepard, K. Heinz-Taheny, P.W. Iversen,
A.H. Dantzig, Human carboxylesterase-2 hydrolyzes the prodrug of gemcita-
bine (LY2334737) and confers prodrug sensitivity to cancer cells, Clinical
cancer research, Off. J. Am. Assoc. Canc. Res. 19 (2013) 1159e1168.
[55] M. Hosokawa, Structure and catalytic properties of carboxylesterase isozymes
involved in metabolic activation of prodrugs, Molecules 13 (2008) 412e431.
[41] H. Eng, M. Niosi, T.S. McDonald, A. Wolford, Y. Chen, S.T. Simila, J.N. Bauman,
13