Synthesis and in vitro evaluation of novel lipophilic monophosphorylated gemcitabine derivatives and their nanoparticles

Article abstract of DOI:10.1016/j.ijpharm.2012.03.014

Gemcitabine hydrochloride (HCl) is approved for the treatment of a wide spectrum of solid tumors. However, the rapid development of resistance often makes gemcitabine less efficacious. In the present study, we synthesized several novel lipophilic monophosphorylated gemcitabine derivatives, incorporated them into solid lipid nanoparticles, and then evaluated their ability to overcome major known gemcitabine resistance mechanisms by evaluating their in vitro cytotoxicities in cancer cells that are deficient in deoxycytidine kinase (dCK), deficient in human equilibrative nucleoside transporter (hENT1), over-expressing ribonucleotide reductase M1 subunit (RRM1), or over-expressing RRM2. In dCK deficient cells, the monophosphorylated gemcitabine derivatives and their nanoparticles were up to 86-fold more cytotoxic than gemcitabine HCl. The majority of the gemcitabine derivatives and their nanoparticles were more cytotoxic than gemcitabine HCl in cells that over-expressing RRM1 or RRM2, and the gemcitabine derivatives in nanoparticles were also resistant to deamination by deoxycytidine deaminase. The gemcitabine derivatives (in nanoparticles) hold a great potential in overcoming gemcitabine resistance.

Full text of DOI:10.1016/j.ijpharm.2012.03.014

International Journal of Pharmaceutics 429 (2012) 123–134
Contents lists available at SciVerse ScienceDirect
International Journal of Pharmaceutics
journal homepage: www.elsevier.com/locate/ijpharm
Pharmaceutical nanotechnology
Synthesis and in vitro evaluation of novel lipophilic monophosphorylated
gemcitabine derivatives and their nanoparticles
Dharmika S.P. Lansakara-P., B. Leticia Rodriguez, Zhengrong Cui∗
The University of Texas at Austin, College of Pharmacy, Pharmaceutics Division, Austin, TX 78712, United States
a r t i c l e i n f o
a b s t r a c t
Article history:
Gemcitabine hydrochloride (HCl) is approved for the treatment of a wide spectrum of solid tumors.
However, the rapid development of resistance often makes gemcitabine less efficacious. In the present
study, we synthesized several novel lipophilic monophosphorylated gemcitabine derivatives, incorpo-
rated them into solid lipid nanoparticles, and then evaluated their ability to overcome major known
gemcitabine resistance mechanisms by evaluating their in vitro cytotoxicities in cancer cells that are
deficient in deoxycytidine kinase (dCK), deficient in human equilibrative nucleoside transporter (hENT1),
over-expressing ribonucleotide reductase M1 subunit (RRM1), or over-expressing RRM2. In dCK deficient
cells, the monophosphorylated gemcitabine derivatives and their nanoparticles were up to 86-fold more
cytotoxic than gemcitabine HCl. The majority of the gemcitabine derivatives and their nanoparticles
were more cytotoxic than gemcitabine HCl in cells that over-expressing RRM1 or RRM2, and the gemc-
itabine derivatives in nanoparticles were also resistant to deamination by deoxycytidine deaminase. The
gemcitabine derivatives (in nanoparticles) hold a great potential in overcoming gemcitabine resistance.
© 2012 Elsevier B.V. All rights reserved.
Received 12 November 2011
Received in revised form 6 March 2012
Accepted 7 March 2012
Available online 16 March 2012
Keywords:
Gemcitabine resistance
Nanoparticles
In vitro cytotoxicity
Cancer cells
1. Introduction
Gemcitabine is a prodrug, which needs to be phosphorylated to
gemcitabine monophosphate (dFdCMP) by dCK (Kroep et al., 2002;
Gemcitabine (2ꢀ,2ꢀ-difluorodeoxycytidine, dFdC) HCl is
approved for the treatment of a wide spectrum of solid tumors
including pancreatic, non-small-cell lung cancer, breast, and
bladder cancers (Candelaria et al., 2007, 2010; Barton-Burke, 1999;
Kalykaki et al., 2008; Sandler et al., 2000; Zucali et al., 2008; Cetina
et al., 2004). However, tumor cells often acquire resistance during
or after gemcitabine treatment (Bergman et al., 2002; Andersson
et al., 2009; Sezgin et al., 2005). Gemcitabine is a polar deoxycy-
tidine analogue. It requires nucleoside transporters to translocate
across the cellular membrane (Heinemann et al., 1995). Clinical
data showed that patients with tumors with a decreased expres-
sion of hENT1, a nucleoside transporter, have a significantly lower
survival rate after gemcitabine treatment than those with tumors
that express a higher level of hENT1 (Giovannetti et al., 2006;
Mey et al., 2006; Spratlin et al., 2004). More than 90% gemcitabine
that are internalized into cells are deaminated by deoxycytidine
deaminase (dCDA) to form inactive 2ꢀ-deoxy-2ꢀ,2ꢀ-difluorouridine
(dFdU) (Immordino et al., 2004). Therefore, deamination affects
the efficacy of gemcitabine adversely (Immordino et al., 2004).
Bouffard et al., 1993). Subsequently, dFdCMP is phosphorylated
by nucleotide kinases to di- and tri-phosphorylated gemcitabine
(dFdCDP and dFdCTP, respectively) that are active metabolites of
gemcitabine (Bergman et al., 2002; Mini et al., 2006; Ueno et al.,
2007). Therefore, tumor cells deficient in dCK are resistant to
gemcitabine. Clinical studies in patients with resected pancreatic
adenocarcinoma showed a strong correlation between low level
of dCK expression and poor clinical outcomes after gemcitabine-
based adjuvant therapy (Maréchal et al., 2010). Disease-free
survival was significantly longer in patients having high levels
of dCK expression (38.6–77.5 months) than in patients having
low levels of dCK expression (2.9–9.6 months) (Maréchal et al.,
2010). The anti-proliferative activity of gemcitabine is known to be
exerted mainly through the inhibition of DNA synthesis by masked
chain termination and inhibition of DNA polymerase by dFdCTP
(Huang and Plunkett, 1995). RR is required to convert ribonu-
cleotides to deoxyribonucleotides. dFdCDP inhibits RR, leading to
the depletion of deoxynucleotide triphosphate (dNTP) pool and
the enhancement of the activity of gemcitabine (Heinemann et al.,
1992). Active RR composes of two homodimers of non-identical
subunits, the large RRM1 subunit and the small RRM2 subunit
(Candelaria et al., 2010). Both pre-clinical and clinical data have
shown that tumor cells that over-express of RRM1 or RRM2
are resistant to gemcitabine treatment (Jordheim et al., 2005;
Boukovinas et al., 2008).
∗
Corresponding author at: The University of Texas at Austin, The Dell Pediatric
Research Institute, 1400 Barbara Jordan Boulevard, Austin, TX 78723, United States.
Tel.: +1 512 495 4758, fax: +1 512 471 7474.
E-mail address: zhengrong.cui@austin.utexas.edu (Z. Cui).
0378-5173/$ – see front matter © 2012 Elsevier B.V. All rights reserved.
doi:10.1016/j.ijpharm.2012.03.014

124
D.S.P. Lansakara-P. et al. / International Journal of Pharmaceutics 429 (2012) 123–134
The most commonly reported acquired resistance to gemc-
(Pittsburgh, PA) and were used as received unless noted. Gem-
citabine HCl was from U.S. Pharmacopeia (Rockville, MD). Soy
lecithin was from Alfa Aesar (Ward Hill, MA). The 1-hydroxy-7-
azabenzotriazole (HOAt) was from CreoSalus, Inc. (Louisville, KY).
Air or moisture-sensitive reactions were performed under an atmo-
sphere of argon. Thin-layer chromatography (TLC) on Whatman
silica gel plates (UV254) from Fisher Scientific was used to moni-
tor the reaction progress. Silica gel–grade 60 (230–400 mesh) from
Fisher Scientific was used for column chromatography to purify
reaction products. The chemical structures of final compounds
were confirmed using NMR and high resolution mass spectrometry.
itabine is dCK deficiency (Bergman et al., 2002; Gregoire et al.,
2002). Inefficient intracellular monophosphorylation of gem-
citabine may reduce the efficacy of gemcitabine drastically.
However, only a few gemcitabine analogues with a phosphate
moiety have been reported to address the problem. The first syn-
thesized compound to address dCK deficiency was a gemcitabine
derivative linked with a C-1 thioether, C-2 oxyether phospholipid
(Alexander et al., 2003, 2005; Alexander and Kucera, 2005). Wu
et al. (2007) then introduced the gemcitabine phosphoramidate
prodrug, which was shown to be 4 times more cytotoxic than
gemcitabine in dCK deficient cell lines. In the present study, sev-
eral new nucleoside analogues were designed and synthesized to
overcome major gemcitabine resistance mechanisms. All the com-
pounds were linked with a monophosphate group at the 5ꢀC of the
deoxyribofuranose ring, and thus, have the potential to bypass the
rate limiting step of dCK-dependent gemcitabine activation. More-
over, by attaching a long chain stearoyl group on gemcitabine, it is
expected that nucleoside transporters may no longer be required
for the lipophilized gemcitabine to enter cells. Furthermore, it
is expected that the lipophilic monophosphorylated gemcitabine
derivatives may also become resistant to deamination by dCDA
(Song et al., 2005).
Finally, we previously reported that the incorporation of
another lipophilic gemcitabine derivative, 4-N-stearoyl gemc-
itabine (GemC18), into solid lipid nanoparticles engineered from
lecithin/glycerol monostearate-in-water emulsions rendered the
gemcitabine less sensitive to resistance caused by the over-
expression of RRM1 (Chung et al., 2012; Sloat et al., 2011).
Therefore, the solid lipid nanoparticle formulation was extended to
the novel lipophilic monophosphorylated gemcitabine derivatives
as well. In order to evaluate the extent to which the novel gemc-
itabine derivatives, alone or in nanoparticles, can overcome various
gemcitabine resistance mechanisms, their in vitro cytotoxicities
were determined in cancer cells that are dCK deficient, hENT1 defi-
cient, over-expressing RRM1, or over-expressing RRM2. In addition,
the ability of selected gemcitabine derivative-containing nanopar-
ticles to competitively inhibit the deamination activity of dCDA was
evaluated and compared to that of gemcitabine HCl to understand
the extent to which the gemcitabine derivatives in nanoparticles
may become less sensitive to deamination.
2.2. Chemical syntheses (Scheme 1)
2.2.1. 4-N-stearoyl gemcitabine (4)
Compound 4 or GemC18 was synthesized as previously reported
(Guo and Gallo 1999; Immordino et al., 2004). Gemcitabine HCl
salt (1) (200 mg, 0.67 mmol) in 13.3 mL of 1 N potassium hydroxide
(KOH) was cooled to 4 ◦C. To this solution, di-tert-butyl dicarbon-
ate (Boc2O, 1.483 g, 6.8 mmol) in anhydrous dioxane (13.3 mL) was
added over 10 min under argon atmosphere. The reaction mixture
was stirred at 22 ◦C for 1 h, followed by an extraction procedure
(i.e. The mixture was extracted with ethyl acetate (EtOAc)) and
a workup protocol (i.e. the organic layer was washed with brine,
dried over anhydrous sodium sulfate (Na2SO4), filtered and solvent
was removed under reduced pressure). The residue was added to
Boc2O (1.483 g, 6.8 mmol) in anhydrous dioxane (13.3 mL) and 1 N
KOH (13.3 mL) at 22 ◦C. The reaction progress was monitored by
TLC. After 1 h, the extraction and workup protocols were repeated,
and the crude product was purified by column chromatography
(dichloromethane (DCM): acetone, 1:1). The desired product frac-
tions were pooled and dried to yield 219 mg of 2 (yield of 71%).
1H NMR (500 MHz, acetone-d6) ı 7.60 (1H, d, J = 7.6 Hz, 6-CH), 6.34
(1H, brs, 1ꢀ-CH), 5.97 (1H, d, J = 7.6 Hz, 5-CH), 5.29 (1H, brs, 3ꢀ-CH),
4.53–4.39 (3H, m, 4ꢀ-CH, 5ꢀA-CH, 5ꢀB-CH), 2.82 (2H, s, NH2) 1.50,
1.47 (18H, two s, (CH3)3C). A solution of 2 (219 mg, 0.47 mmol),
stearic acid (149 mg, 0.52 mmol) and HOAt (70 mg, 0.52 mmol)
in anhydrous DCM was pre-cooled to 4 ◦C, and 1-ethyl-3-(3-
dimethylaminopropyl)carbodiimide (EDCI) (109 mg, 0.57 mmol)
was added. The mixture was de-gassed by vacuum sonication and
then stirred at room temperature under argon for about 40 h. Water
(15 mL) was added to the reaction mixture and extracted with the
mixture of EtOAc and hexane (2:1). The combined organic phase
was washed with saturated ammonium chloride (NH4Cl) followed
by the workup protocol as aforementioned, and the residue was
purified by column chromatography (EtOAc:hexane, 3:7). The con-
jugated amide 3 was isolated as a white powder (319 mg, 92%).
1H NMR (300 MHz, acetone-d6) ı 9.90 (1H, s, NHCO), 8.03 (1H, d,
J = 7.8 Hz, 6-CH), 7.45(1H, d, J = 7.5 Hz, 5-CH), 6.38 (1H, t, J = 8.7 Hz, 1ꢀ-
CH), 5.40–5.30 (1H, m, 3ꢀ-CH), 4.56–4.44 (3H, m, 4ꢀ-CH and 5ꢀ-CH2),
2.57 (2H, t, J = 7.5 Hz, CO CH2), 1.71–1.65 (2H, m, CO CH2 CH2),
1.50, 1.47 (18H, two s, (CH3)3C), 1.40–1.20 (28H, m, CH2), 0.90–0.87
(3H, m, terminal CH3). To a stirred solution of the compound 3
(319 mg, 0.44 mmol) in DCM (7 mL), about 1.5 mL of trifluroacetic
acid (TFA) was added. After 2 h, excess TFA was removed under
reduced pressure. The concentrated sample was co-distilled with
DCM for 5 times. The crude sample was chromatographed on silica
gel (DCM:ethanol, 94:6) (Immordino et al., 2004). The final product
4 was a white powder (162 mg, 70%). 1H NMR (500 MHz, pyridine-
d5) ı 11.97 (1H, s, NHCO), 8.75 (1H, d, J = 7.6 Hz, 6-CH), 7.74 (1H,
d, J = 7.6 Hz, 5-CH), 6.99 (1H, t, J = 7.2 Hz, 1ꢀ-CH), 5.18–5.11 (1H, m,
3ꢀ-CH), 4.47–4.28 (3H, overlapping m, 4ꢀ-CH and 5ꢀ-CH2), 2.67 (2H,
t, J = 7.4 Hz, CO CH2), 1.83–1.76 (2H, m, CO CH2 CH2), 1.34–1.20
(28H, m, CH2), 0.87 (3H, t, J = 6.9 Hz, terminal CH3). ESI-HRMS [M+H]
2. Materials and methods
2.1. General materials and methods
Proton NMR spectra were recorded on a 300 MHz Varian UNITY
Plus or a 500 MHz Varian INOVA. Chemical shifts (ı) of 1H NMR
spectra were recorded in parts per million (ppm) relative to tetram-
ethylsilane (TMS), which was the reference (ı = 0 ppm). 1H NMR
data are reported according to the following order: chemical shift,
integration (i.e. number of hydrogen atoms), multiplicity (s, sin-
glet; d, doublet; t, triplet; q, quartet; m, multiplet; br, broad; brs,
broad singlet), and coupling constant in Hertz (Hz). High reso-
lution mass spectra were acquired in electronspray positive and
negative ionization modes by direct injection onto an IonSpec 9.4T
QFT-FTMS system in the mass spectrometry facility of the Depart-
ment of Chemistry and Biochemistry at the University of Texas at
Austin. The concentrations of deoxycytidine (dCyd) and deoxyuri-
dine (dUrd) in the dCDA assay were determined using an Agilent
1260 Infinity high performance liquid chromatography (HPLC) with
an Agilent ZORBAX Eclipse Plus C18 column (4.6 × 150 mm, 5 ␮m)
attached to a ZORBAX Eclipse Plus guard column (Agilent Technolo-
gies, Inc., Santa Clara, CA).
All commercially available chemical reagents were purchased
from Sigma–Aldrich (St. Louis, MO) or Thermo Fisher Scientific Inc.
+
m/z calculated for C27H46F2N3O5: 530.3406, found: 530.3401.

D.S.P. Lansakara-P. et al. / International Journal of Pharmaceutics 429 (2012) 123–134
125
2.2.2. 2ꢀ-2ꢀ-difluoro-cytosine-5ꢀ-octadecylphosphate (8) (Bligh
and Dyer, 1959; Guo and Gallo 1999; Perie et al., 1990)
DCM, about 0.9 mL of TFA was added. This solution was stirred at
room temperature for 2 h. Excess TFA was removed under reduced
pressure, and the concentrated sample was co-distilled with DCM
for 5 times. The crude sample was column-purified on silica gel
by eluting with 20%, 40%, and 50% CH3OH in CHCl3, sequentially.
The desired fractions with the Rf value of 0.25 (CHCl3:CH3OH,
3:2) were pooled and evaporated to dryness to yield 41 mg of
8 (58%). 1H NMR (300 MHz, CDCl3:CD3OD, 4:1) ı 7.88 (1H, d,
J = 8.1 Hz, 6-CH), 6.17 (1H, t, J = 6.9 Hz, 1ꢀ-CH), 6.06 (1H, d, J = 7.2 Hz,
5-CH), 4.42–4.00 (4H, overlapping m, 3ꢀ-CH, 4ꢀ-CH, 5ꢀA-CH, 5ꢀB-
CH), 3.94–3.71 (2H, m, CH2OP), 1.62 (2H, t, J = 6.6 Hz, CH2CH2OP),
1.39–1.19 (30H, m, CH2 from C18 chain), 0.88 (3H, t, J = 6.6 Hz, ter-
minal CH3). ESI-HRMS [M-H]− m/z calculated for C27H47F2N3O7P−:
594.3125, found: 594.3125.
The mixture of gemcitabine HCl (200 mg, 0.67 mmol) and
Na2CO3 (354 mg, 3.3 mmol) in water (3.3 mL) and dioxane (13.3 mL)
was added to Boc2O (147 mg, 0.67 mmol). After 48 h of stir-
ring at room temperature, 15 mL of water was added, followed
by the extraction and workup procedure mentioned above. The
crude sample was chromatographed on silica gel (DCM:acetone,
1:2) to obtain 212 mg of 3ꢀ-O-(tert-butoxycarbonyl)-2ꢀ-2ꢀ-difluoro-
cytidine (5A) (87%). 1H NMR (300 MHz, acetone-d6) ı 7.75 (1H, d,
J = 7.5 Hz, 6-CH), 6.34 (1H, t, J = 9.1 Hz, 1ꢀ-CH), 6.04 (1H, d, J = 7.8 Hz,
5-CH), 5.40-5.31 (1H, m, 3ꢀ-CH), 4.23–4.19 (1H, m, 4ꢀ-CH), 4.00–3.82
(2H, overlapping m, 5ꢀA-CH, 5ꢀB-CH), 1.51 (9H, s, (CH3)3C). Boc2O
(1.26 g, 5.8 mmol) was added to a stirred solution of 5A (212 mg,
0.58 mmol) in 10 mL of dioxane. The resultant mixture was main-
tained at 37 ◦C in a rotary shaker at 250 rpm for 3 days. Water
(25 mL) was added to the sample and followed by the extraction
and workup protocol mentioned above. The concentrated sample
was chromatographed on silica gel (EtOAc: hexanes, 1:1) to obtain
the desired product of 4-N-3ꢀ-O-bis(tert-butoxycarbonyl)-2ꢀ-2ꢀ-
difluoro-cytidine (5B) (196 mg). 1H NMR (300 MHz, acetone-d6) ı
8.25 (1H, d, J = 7.5 Hz, 6-CH), 7.30 (1H, d, J = 7.5 Hz, 5-CH), 6.36 (1H, t,
J = 8.6 Hz, 1ꢀ-CH), 5.37–5.28 (1H, m, 3ꢀ-CH), 4.34–4.28 (1H, m, 4ꢀ-CH),
4.08–3.98 (1H, m, 5ꢀA-CH), 3.87 (1H, m, 5ꢀB-CH), 1.52, 1.50 (18H,
two s, (CH3)3C). Octadecanol (5 g, 18.48 mmol) and triethylamine
(4.8 g, 47.52 mmol) were partially dissolved in 50 mL of DCM under
argon. Phosphorus oxychloride (POCl3) (2.8 g, 18.48 mmol) was
added drop-wise and heated to reflux for 2 h. The reaction mixture
was filtered to remove triethylamine hydrochloride, and the filtrate
was added to 0.2 N sodium bicarbonate (NaHCO3) (370 mL). After
15 h stirring at room temperature, 370 mL of acetone was added,
and the white precipitate was recovered by filtration. The precipi-
tate was washed with acetone and re-dissolved in 400 mL of water.
Another 260 mL of acetone was added, and the precipitate was
recovered, washed with acetone, dissolved in a homogeneous mix-
ture of 200 mL of chloroform (CHCl3), 400 mL of methanol (CH3OH),
and 200 mL of 0.1 N HCl, and stirred for 1 h at room temperature. A
mixture of 200 mL of CHCl3 and 200 mL of water was added, and the
organic layer was isolated (Bligh and Dyer, 1959; Perie et al., 1990).
The aqueous phase was extracted with CHCl3 (2 × 100 mL). The
combined organic layer was evaporated to dryness and lyophilized
to obtain desired product of octadecylphosphate (6) (2.2 g, 34%).
1H NMR (300 MHz, CDCl3:CD3OD, 4:1) ı 3.98 (2H, q, J = 6.8 Hz,
CH2OP), 1.70–1.62 (2H, m, CH2CH2OP), 1.39–1.19 (30H, m, CH2 from
C18 chain), 0.88 (3H, t, J = 6.8 Hz, terminal CH3). ESI-HRMS [M-H]−
m/z calculated for C18H38O4P−: 349.2513, found: 349.2515. The
powders of compound 5 (100 mg, 0.22 mmol) and octadecylphos-
phate (220 mg, 0.62 mmol) were mixed and lyophilized for 15 h.
To the lyophilized powder, 2,4,6-triisopropylbenzenesulfonyl chlo-
ride (TPS) (154 mg, 0.5 mmol) and 2 mL of pyridine were added
under argon environment, and the reaction was stirred at 38–40 ◦C
for 24 h. A few drops of water were added, and solvent was removed
under reduced pressure. The crude oil was chromatographed
on silica gel, eluting first with CHCl3:CH3OH (24:1) and then
with CHCl3:CH3OH (9:1) (Alexander et al., 2003). The fractions of
the desired product (Rf = 0.2 in CHCl3:CH3OH, 9:1) were pooled,
and 4-N-3ꢀ-O-Bis(tert-butoxycarbonyl)-2ꢀ-2ꢀ-difluoro-cytidine-5ꢀ-
octadecylphosphate (7) was isolated as a white powder (156 mg,
89%). 1H NMR (300 MHz, CDCl3:CD3OD, 4:1) ı 7.94 (1H, d, J = 7.2 Hz,
6-CH), 7.30 (1H, d, J = 7.5 Hz, 5-CH), 6.35 (1H, t, J = 8.6 Hz, 1ꢀ-CH),
5.29–5.18 (1H, m, 3ꢀ-CH), 4.29–4.10 (3H, overlapping m, 4ꢀ-CH,
5ꢀA-CH, 5ꢀB-CH), 3.89–3.78 (2H, m, CH2OP), 1.60 (2H, t, J = 6.6 Hz,
CH2CH2OP), 1.52, 1.50 (18H, two s, (CH3)3C), 1.39–1.19 (30H, m,
CH2 from C18 chain), 0.88 (3H, t, J = 6.6 Hz, terminal CH3). ESI-HRMS
[M-H]− m/z calculated for C37H63F2N3O11P−: 794.4174, found:
794.4162. To a stirred solution of 7 (95 mg, 0.12 mmol) in 6 mL of
2.2.3. 4-N-acetyl-2ꢀ-2ꢀ-difluoro-cytosine-5ꢀ-octadecylphosphate
(9) (Watanabe and Fo, 1966)
Acetic anhydride (150 ␮L) and 8 (16.5 mg, 0.03 mmol) in 2 mL
of CH3OH was refluxed for 15 h, and the resultant mixture was
co-distilled with CH3OH five times. The resultant sample was vac-
uum dried overnight to obtain 14 mg of 9 (73%). 1H NMR (300 MHz,
CDCl3: CD3OD, 4:1) ı 7.90–8.10 (1H, m, 6-CH), 6.30–6.05 (2H,
two m, 1ꢀ-CH, 5-CH), 4.42–3.80 (6H, overlapping m, 3ꢀ-CH, 4ꢀ-CH,
5ꢀA-CH, 5ꢀB-CH, CH2OP), 2.04 (3H, s, NHCOCH3) 1.65–1.50 (2H, m,
CH2CH2OP), 1.39–1.19 (30H, m, CH2 from C18 chain), 0.88 (3H,
t, J = 6.3 Hz, terminal CH3). ESI-HRMS [M-H]− m/z calculated for
C
29H49F2N3O8P−: 636.3231, found: 636.3220.
2.2.4. 2ꢀ-2ꢀ-difluoro-5ꢀ-monophosphate(1-stearate-2-chloro-
propyl)cytosine (12) (Alexander et al., 2003; Bligh and Dyer,
1959; Perie et al., 1990)
Glycerol monostearate (1 g, 2.79 mmol) and triethylamine
(0.726 g, 7.17 mmol) were partially dissolved in 12 mL of DCM
under argon. POCl3 (0.428 g, 2.79 mmol) was added drop-wise
and heated to reflux for 2 h. The reaction mixture was filtered to
remove triethylamine hydrochloride, and the filtrate was added
to 0.2 N NaHCO3 (56 mL). After 15 h of stirring at room tempera-
ture, 56 mL of acetone was added, and the white precipitate was
recovered by filtration. The precipitate was washed with acetone
and re-dissolved in 60 mL of water. Another 40 mL of acetone was
added, and the precipitate was recovered. It was then washed
with acetone, dissolved in a homogeneous mixture of 100 mL of
CHCl3, 200 mL of CH3OH, and 100 mL of 0.1 N HCl, and stirred
for 1 h at room temperature. A mixture of 100 mL of CHCl3 and
100 mL of water was added, and the organic layer was isolated. The
aqueous phase was extracted with CHCl3 (2 × 50 mL). The com-
bined organic layer was evaporated to dryness and lyophilized
to obtain 1-stearate-2-hydroxy-3-phosphatidic acid (10) (720 mg,
59%). 1H NMR (300 MHz, CDCl3: CD3OD, 4:1) ı 4.10–4.03 (2H,
m, COOCH2), 4.00–3.90 (3H, overlapping m, CH2OP, CHOH), 2.27
(2H, t, J = 7.4 Hz, COCH2), 1.60–1.48 (2H, m, COCH2CH2), 1.30–1.10
(28H, m, CH2 from C18 chain), 0.80 (3H, t, J = 6.0 Hz, terminal
CH3). ESI-HRMS [M-H]− m/z calculated for C21H42O7P−: 437.2674,
found: 437.2673. The powders of 5 (50 mg, 0.11 mmol) and 10
(110 mg, 0.25 mmol) were mixed and lyophilized for 15 h. To
the lyophilized powder, TPS (74 mg, 0.25 mmol) and 1 mL of
pyridine was added under argon environment, and the reaction
was stirred at 38–40 ◦C for 24 h. A few drops of water were
added, and the solvent was removed under reduced pressure. The
crude oil was chromatographed on silica gel, eluting first with
CHCl3:CH3OH (24:1) and then with CHCl3:CH3OH (9:1) (Alexander
et al., 2003). 4-N-3ꢀ-O-Bis(tert-butoxycarbonyl)-2ꢀ-2ꢀ-difluoro-
5ꢀ-monophosphate(1-stearate-2-chloro-propyl)cytosine (11) was
isolated as a white powder (56 mg, 62%). 1H NMR (300 MHz,
CDCl3:CD3OD, 4:1) ı 7.95 (1H, d, J = 7.8 Hz, 6-CH), 7.32 (1H, d,
J = 6.9 Hz, 5-CH), 6.34 (1H, t, J = 7.6 Hz, 1ꢀ-CH), 5.18–5.31 (1H,

126
D.S.P. Lansakara-P. et al. / International Journal of Pharmaceutics 429 (2012) 123–134
m, 3ꢀ-CH), 4.60–4.10 (7H, overlapping m, 4ꢀ-CH, 5ꢀA-CH, 5ꢀB-
CH, COCH2, CH2OP), 3.82–3.71 (1H, m, CHCl), 2.39–2.29 (2H, m,
COCH2), 1.68–1.56 (2H, m, CH2CH2CO), 1.53,1.51 (18H, two s,
(CH3)3C), 1.38–1.19 (28H, m, CH2 from C18 chain), 0.88 (3H,
t, J = 6.6 Hz, terminal CH3). ESI-HRMS [M-H]− m/z calculated for
2.4. Preparation of nanoparticles
All gemcitabine derivative-containing nanoparticles were pre-
pared as previously described (Sloat et al., 2011, 2010). Briefly,
3.5 mg of soy lecithin, 0.5 mg of glycerol monostearate, and 1 mg
of gemcitabine derivative were added to water (1 mL). The mixture
was maintained at 50–62 ◦C while stirring until the formation of
homogenous slurry. Tween 20 was added to a final concentration
of 1% (v/v). The resultant emulsions were allowed to cool to room
temperature while stirring to form nanoparticles. Particle size and
zeta potential were determined using a Malvern Zetasizer Nano ZS
(Westborough, MA).
C
40H66ClF2N3O13P−: 900.3995, found: 900.3997, [M + H]+ m/z cal-
culated for C40H68ClF2N3O13P+: 902.4141, found: 902.4150. To a
stirred solution of 11 (50 mg, 0.055 mmol) in 4 mL of DCM, about
0.5 mL of TFA was added. This solution was stirred at room temper-
ature for 2 h. The excess TFA was removed under reduced pressure,
and the concentrated sample was co-distilled with DCM for 5 times.
The crude sample was column-purified on silica gel by eluting with
5%, 10%, and 15% CH3OH in CHCl3, sequentially (Alexander et al.,
2003). The desired fractions with the Rf value of 0.2 (CH3OH:CHCl3,
2:1) were pooled and evaporated to dryness to yield 18 mg of com-
pound 12 (47%). 1H NMR (500 MHz, CDCl3:CD3OD, 4:1) ı 7.78 (1H,
d, J = 7.6 Hz, 6-CH), 6.22 (1H, t, J = 6.9 Hz, 1ꢀ-CH), 5.95–5.92 (1H, m,
5-CH), 4.24–3.66 (9H, overlapping m, 3ꢀ-CH, 4ꢀ-CH, 5ꢀA-CH, 5ꢀB-CH,
COCH2, CH2OP, CHCl), 2.36–2.29 (2H, m, COCH2), 1.62–1.58 (2H,
m, CH2CH2CO), 1.31–1.19 (28H, m, CH2 from C18 chain), 0.88 (3H,
t, J = 7.0 Hz, terminal CH3). ESI-HRMS [M-H]− m/z calculated for
2.5. Gel permeation chromatography
Gel permeation chromatography (GPC) was performed to sepa-
rate un-incorporated 8 from nanoparticles using a 6 mm × 150 mm
Sepharose® 4B column, which was equilibrated with phosphate
buffered saline (PBS, pH 7.4). Samples (100 ␮L) were applied into
the column and eluted with PBS (Sloat et al., 2011). Elution fractions
of 250 ␮L were collected, and their absorbances at 234 nm were
measured using a BioTek SynergyTM HT Multi-Mode Microplate
Reader (Winooski, VT). As a control, compound 8 in Tween 20
micelles (100 ␮g/mL of 8 in 1% (v/v) of Tween 20 in water) were
applied to the GPC column.
C
20H50ClF2N3O9P−: 700.2947, found: 700.2958, [M + H]+ m/z cal-
culated for C30H52ClF2N3O9P+: 702.3092, found: 702.3089.
2.2.5. 4-N-heptadecylcarbonyl-2ꢀ-2ꢀ-difluoro-cytosine-5ꢀ-
2.6. In vitro release of 8 from 8-NPs
sodiummonophosphate (13) (Perie et al., 1990)
Compound 4 (50 mg, 0.094 mmol) and triethylamine (0.072 g,
0.72 mmol) were partially dissolved in 2 mL of DCM under argon.
POCl3 (0.0329 g, 0.21 mmol) was added drop-wise, and the same
phosphorylation reaction protocol as mentioned in Section 2.2.4
was followed. The precipitate was washed with acetone several
times and dried under vacuum to obtain 7 mg of pale yellow
powder of 13 (12% yields). 1H NMR (300 MHz, CD3OD) ı 8.4–8.3
(1H, m, 6-CH), 7.6–7.5 (1H, m, 5-CH), 6.3–6.2 (1H, m, 1ꢀ-CH),
4.2–3.8 (4H, overlapping m, 3ꢀ-CH, 4ꢀ-CH and 5ꢀ-CH2), 2.44 (2H, m,
CO CH2), 1.8–1.7 (2H, m, CO CH2 CH2), 1.34–1.20 (28H, m, CH2),
0.89 (3H, m, terminal CH3). ESI-HRMS [M-H]− m/z calculated for
In vitro release of 8 from 8-NPs was performed as described pre-
viously (Sloat et al., 2011). 8-NPs (100 ␮g/mL) were placed into a
1 mL cellulose ester dialysis tube (MWC 50,000) from Spectrum
Chemicals & Laboratory Products (New Brunswick, NJ). The dial-
ysis tube was placed into a conical tube containing 12 mL of PBS
with 0.05% (w/v) of sodium dodecyl sulfate (SDS) and incubated
in a 37 ◦C shaker incubator. At predetermined time points, 200 ␮L
of the release medium was withdrawn and replaced with 200 ␮L
of fresh release medium. As a control, the diffusion of 8-in-Tween
20 micelles across the dialysis membrane was also measured. The
absorbance was measured at 234 nm.
C
27H45F2N3O8P+: 608.2918, found: 608.2933, [M+H]+ m/z calcu-
lated for C27H47F2N3O8P+: 610.3063, found: 610.3064.
2.7. In vitro cytotoxicity and apoptosis assay
Cells (5000/well) were seeded in 96-well plates. After overnight
incubation, they were treated with various concentrations of
gemcitabine HCl, gemcitabine derivatives, or gemcitabine deriva-
tives in nanoparticles at 37 ◦C, 5% CO2. TC-1 and TC-1-GR cells
were incubated for 48 h. CCRF-CEM, CCRF-CEM-AraC-8C, CCRF-
CEM/dCK−/−, L1210 wt, L1210 10K, and MCF-7 cells were incubated
for 72 h, and MIA PaCa-2 and PANC-1 for 96 h. Gemcitabine HCl
was dissolved in PBS, and gemcitabine derivatives were dissolved
in dimethyl sulfoxide (DMSO). The maximum amount of DMSO
added per well was 1 ␮L, which was found non-toxic. Compound
13 was not used in the in vitro cytotoxicity assay because it
was not sufficiently solubilized in DMSO. The number of viable
cells after the incubation was determined using an MTT assay.
Briefly, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bro-
mide (20 ␮L, 5 mg/mL) was added in each well and incubated for
3 h. Formazan crystals were solubilized with acidified isopropanol
2.3. Cell lines and cell culture
Human leukemia cell line CCRF-CEM (CCL-119), human pancre-
atic cancer cell lines PANC-1 (CRL-1469), MIA PaCa-2 (CRL-1420),
and BxPC-3 (CRL-1687), human breast adenocarcinoma cell line
MCF-7 (HTB-22), and mouse lung cancer cell line TC-1 (CRL-2785)
were from the American Type Culture Collection (Rockville, MD).
CCRF-CEM-AraC-8C cells (hENT1 deficient) and CCRF-CEM/dCK−/−
cells (dCK deficient) were kindly provided by Dr. Buddy Ull-
mann (Oregon Health & Science University, Portland, OR) and
Dr. Margaret Black (Washington State University, Pullman, WA),
respectively. L1210 wt and L1210 10K were kindly supplied by
Dr. Lars Petter Jordheim (Université Claude Bernard Lyon I, Lyon,
France). TC-1-GR cells were previously developed in our lab (Chung
et al., 2012). CCRF-CEM, CCRF-CEM-AraC-8C, CCRF-CEM/dCK−/−
,
(150 ␮L) (for CCRF-CEM, CCRF-CEM-AraC-8 C, CCRF-CEM/dCK−/−
,
L1210 wt, L1210 10K, TC-1, and TC-1-GR cells were cultured in
RPMI 1640 medium. MCF-7 and PANC-1 cells were cultured in
Dulbecco’s modified Eagle medium (DMEM), and MIA PaCa-2 cells
were cultured in DMEM medium with 2.5% horse serum. All media
were supplemented with 10% fetal bovine serum (FBS), 100 U/mL
of penicillin, and 100 ␮g/mL of streptomycin (all from Invitrogen,
Carlsbad, CA).
L1210 wt, and L1210 10K cells) or DMSO (150 ␮L) (for TC-1, TC-1-
GR, MCF-7, PANC-1, and MIA PaCa-2). Absorbance was measured
using a BioTek SynergyTM HT Multi-Mode Microplate Reader at
570 nm and 630 nm. The fraction of affected (dead) cells (Fa) and
the fraction of unaffected (live) cells (Fu) at every dose were cal-
culated, and the Log (Fa/Fu) values were plotted against the Log
(concentration of gemcitabine) (Chou and Talalay, 1984). IC50 was

D.S.P. Lansakara-P. et al. / International Journal of Pharmaceutics 429 (2012) 123–134
127
the dose at Log (Fa/Fu) = 0. The experiment was repeated at least
three times. To understand whether the cells underwent apoptosis
after treatment with the gemcitabine derivatives in nanoparticles,
CCRF-CEM-dCK−/− cells (5 × 104) were incubated at 37 ◦C, 5% CO2
with gemcitabine, 8 in solution, or 8-NPs (10 ␮M). After 72 h, cells
were resuspended in PBS with 2% FBS, stained using a Guava Nexin
kit that contains annexin V and 7-amino actinomycin D (7-AAD)
according to the manufacturer’s protocol, and analyzed using a
Guava Easycyte 8HT Flow Cytometry System (Millipore, Hayward,
CA). Flow cytometry data were analyzed using the Guava Analysis
Software.
nanoparticles, with molar equivalent concentrations of gemc-
itabine derivatives, were included in the reaction mixture. Controls
include a reaction with substrate (dCyd) but without an inhibitor
and a reaction without substrate and inhibitors, but with blank
nanoparticles.
2.10. Statistics
Statistical analyses were completed using ANOVA followed by
Fisher’s protected less significant procedure. A p value of ≤0.05 was
considered significant.
2.8. In vitro cellular uptake
PANC-1 cells (2.5 × 105/well) were seeded in a 6-well plate
and incubated overnight at 37 ◦C, 5% CO2. The medium was then
replaced with 1 mL of medium containing 8-NPs (40 ␮M), 8 in
solution (40 ␮M), or gemcitabine HCl (40 ␮M and 400 ␮M) and
incubated for predetermined time points at 37 ◦C, 5% CO2. Com-
pound 8 in solution was prepared by diluting 8 in DMSO with
culture media (DMSO concentration at 0.5%, v/v). The culture
medium was removed; cells were washed three times with cold
PBS and then lysed with 1% SDS. The cell lysates were lyophilized
and re-dissolved in CH3OH for gemcitabine and CHCl3:CH3OH (4:1)
for 8-NPs and 8 in solution. The concentrations of gemcitabine
and 8 were determined using HPLC. For gemcitabine analysis, the
mobile phase was CH3OH in 5 mM sodium acetate (15%, v/v) at the
detection wave length of 266 nm, and the flow rate was 1 mL/min.
Arabinofuranosyluracil (AraU) was used as an internal standard,
and the extraction efficiency of AraU was about 100% with the
aforementioned extraction procedure. For compound 8, the mobile
phase was 1 mM PBS in acetonitrile (45%, v/v) at the detection wave
length of 270 nm, and the flow rate was 1 mL/min in reverse phase
HPLC.
3. Results and discussion
3.1. Syntheses of novel lipophilic monophosphorylated
gemcitabine derivatives
GemC18 was synthesized as previously reported with slight
modification (Scheme 1) (Guo and Gallo, 1999; Immordino et al.,
2004; Sloat et al., 2011). Briefly, the primary and secondary alcohols
of deoxyribofuranose ring of gemcitabine (1) were Boc-protected
to prevent potential side reactions. The stearoyl group was con-
jugated to 4-amino group, and the Boc groups were removed to
obtain 4 or GemC18 as a white crystalline powder (Steps a–c;
Scheme 1). To facilitate the direct conjugation to the 5ꢀ-OH, the
primary alcohol and 4-amino groups of gemcitabine were Boc
protected (5B) (Guo and Gallo, 1999). Octadecanol was phospho-
rylated to give the desired product of 6 (Bligh and Dyer, 1959;
Perie et al., 1990). The mixture of lyophilized powders of 5B and
6 were conjugated at the 5ꢀ-OH. After removing Boc groups, the
crude sample was chromatographed to obtain compound 8. Com-
pound 9 was obtained by the acetylation of 8 on the 4-amino
group (Steps e–i; Scheme 1) (Watanabe and Fo, 1966). Glycerol
monostearate was phosphorylated to obtain compound 10 (Bligh
and Dyer, 1959; Perie et al., 1990). Compounds 5B and 10 were
conjugated (Alexander et al., 2003), and deprotection resulted in
compound 12 (Steps j–k; Scheme 1). Finally GemC18 was phos-
phorylated to obtain compound 13 (Step d; Scheme 1) (Perie
et al., 1990). Purities of synthesized compounds 4, 8, 9, 12, and 13
(Scheme 1) were ≥95.0% based on NMR data.
2.9. Partial purification dCDA and dCDA activity assay
Deoxycytidine deaminase was partially purified from BxPC-
3 cells as previously described (Laliberté and Momparler, 1994;
Bergman et al., 2004). The pellet of 1 × 108 cells was suspended
in 4 mL of 20 mM Tris buffer (pH 7.5) containing 5 mM potas-
sium chloride (KCl), 1 mM dithiothreitol, 40 ␮L of streptomycin
sulfate (12.74 mg/mL), and 50 ␮L of protease inhibitor cocktail.
The suspended cells were sonicated and centrifuged for 30 min
at 20,000 × g. Ammonium sulfate was added to reach 40% satu-
ration, stirred for 1 h, and centrifuged at 36,000 × g for 20 min at
4 ◦C. Ammonium sulfate was added to the supernatant to reach
55% saturation, mixed for 1 h and centrifuged at 36,000 × g for
20 min at 4 ◦C. The pellet was resuspended in 1 mL of 20 mM
Tris buffer (pH 7.5) and desalted by overnight dialysis against
water. Protein concentration was measured using Bradford reagent
from Sigma–Aldrich. The dCDA activity assay was carried out as
described previously with slight modifications (Bergman et al.,
2004; Ruiz van Haperen et al., 1993). Briefly, 55 ␮L of partially puri-
fied dCDA (3.2 mg/mL) and 0.5 mM dCyd (20 ␮L) in a total volume
of 200 ␮L of 20 mM of Tris buffer (pH 7.5) were incubated at 37 ◦C
for 15 min. The reaction was terminated by the addition of 50 ␮L of
trichloroacetic acid (40%, w/v) and chilling on ice for 20 min. Pro-
tein was precipitated by centrifugation at 10,000 × g for 10 min,
and the supernatant was neutralized with 500 ␮L of trioctylamine
and 1,1,2-trichloro-trifluoroethane (1:4). The mixture was cen-
trifuged at 10,000 × g for 1 min, and the upper layer was analyzed
using HPLC (detection wavelength, 260 nm; mobile phase, 10%
CH3OH in water). The relevant peaks were quantified to determine
the concentrations of dCyd and dUrd. For the competition assay,
20 ␮L of gemcitabine HCl or gemcitabine derivative-containing
3.2. Preparation and characterization of gemcitabine derivatives
containing nanoparticles
GemC18 and other newly synthesized lipophilic monophospho-
rylated gemcitabine derivatives, 8, 9, 12, and 13, were incorporated
into solid lipid nanoparticles prepared from lecithin/glycerol
monostearate-in-water emulsions to prepare GemC18-NPs, 8-NPs,
9-NPs, 12-NPs, and 13-NPs, respectively (Sloat et al., 2010, 2011).
The sizes of the resultant nanoparticles were 150–175 nm (Fig. 1A),
with a polydispersity index of 0.2–0.3 and zeta potentials of −27 mV
to −40 mV (Fig. 1A). It was likely that all the gemcitabine deriva-
tives were incorporated into the nanoparticles, as supported by
the lack of a micelle peak in the dynamic light scattering spectra
(Fig. 1B, 8-NPs are shown) and the GPC graph of the nanoparticles
(Fig. 1C). In fact, in our previous study, when 5 times more GemC18
(i.e. 5 mg/mL) was used to prepare nanoparticles, the incorpora-
tion efficacy remained at close to 100% (Sloat et al., 2011). Shown
in Fig. 1D is the in vitro release profile of compound 8 from 8-NPs.
Compound 8 was used in Fig. 1B–D because data that will be pre-
sented in the following sections showed that it had the highest
cytotoxicity against most of the gemcitabine resistant tumor cells.

128
D.S.P. Lansakara-P. et al. / International Journal of Pharmaceutics 429 (2012) 123–134
NHBoC
NH2
N
N
O
O
N
O
N
F
F
k
F
F
O
O
O
P
O
O
P
OH
O
OBoC
O
O
O
16
OH
O
O
16
OH
Cl
Cl
11
12
j
10
O
F
O
F
NH2 HCl
5
NH2
N
NHBOC
HN
(CH2)16CH3
NH2
N
HN
(CH2)16CH3
N
N
O
N
O
N
N
N
6
O
N
O
O
N
F
F
c
f
b
O
N
O
e
a
F
F
F
O
N
O1'
F
F
2'
3'
F
F
O
F
O
O
4'
HO
OH
BocO
OBoc
OBOC
HO
5'
BocO
HO
OBoc
OBoc
HO
OH
2
3
1
4
5A
5B
g
6
d. i, ii
O
NHBOC
N
O
O
N
F
F
HN
(CH2)16CH3
N
O
P
O
N
OBOC
O
OH
F
O
16
F
O
7
h
HO
O
OH
P
O
-O
Na+
13
O
NH2
CH3
HN
N
N
O
i
N
O
N
O
F
F
F
F
O
O
P
O
OH
O
OH
9OH
P
O
O
16
O
16
OH
8
Scheme 1. Reagents and conditions: (a) Boc2O, KOH, 1,4 dioxane, 22 ◦C; (b) CH3(CH2)16COOH, EDCI, HOAt, DCM, rt; (c) TFA, DCM, rt; (d) i. POCl3, triethylamine, DCM, reflux
2 h; ii. NaHCO3, rt, 15 h; (e) Boc2O, Na2CO3, dioxane, H2O; (f) Boc2O, dioxane, 37 ◦C, 250 rpm, 72 h; (g) 6, TPS, pyridine, 38–40 ◦C, 24 h; (h) TFA, DCM; (i) acetic anhydride,
CH3OH, reflux; (j) 10, TPS, pyridine, 38–40 ◦C, 24 h; (k) TFA, DCM.
3.3. Cytotoxicities of gemcitabine derivatives and their
nanoparticles in cancer cells
agreement with data from our previous studies, which showed that
the GemC18-NPs were less cytotoxic than gemcitabine HCl in vari-
ous cancer cells including the CCRF-CEM (Chung et al., 2012), likely
because the gemcitabine needs to be hydrolyzed from the GemC18
or GemC18-NPs to be effective (Sloat et al., 2011). In fact, doubling
the incubation time of the GemC18-NPs with the TC-1 lung can-
cer cells enabled the GemC18-NPs to kill the same proportion of
the cancer cells as gemcitabine HCl (Sloat et al., 2011). Finally, it
appears that the IC50 values of GemC18, 8, 9, 12 were not signifi-
cantly different from that of their corresponding nanoparticles in
CCRF-CEM cells (Table 1), indicating that the incorporation of the
gemcitabine derivatives into nanoparticles did not improve their
cytotoxicities against the CCRF-CEM cells.
To evaluate the antitumor activity of the gemcitabine deriva-
tives and the extent to which the gemcitabine derivatives and their
corresponding nanoparticles can overcome various mechanisms
of gemcitabine resistance, the cytotoxicities of them in cancer
cells that are dCK deficient, hENT1 deficient, or over-expressing
RRM1 or RRM2 were determined. In addition, the ability of selected
gemcitabine derivative-containing nanoparticles to competitively
inhibit the deamination activity of partially purified dCDA was eval-
uated and compared to that of gemcitabine HCl as well.
In the CCRF-CEM/dCK−/− cells, the IC50 value of gemcitabine HCl
was 240.4 29.0 ␮M, which was 82,897-fold greater than that in
the parent CCRF-CEM cells (Table 1). In contrast, the IC50 values of
GemC18, 8, 9, 12, GemC18-NPs, 8-NPs, 9-NPs, 12-NPs, and 13-NPs
in the CCRF-CEM/dCK−/− cells were only 25 to 2438-fold greater
than their IC50 values in the parent CCRF-CEM cells (Table 1). In the
dCK deficient CCRF-CEM-dCK−/− cells, the IC50 values of the gem-
citabine derivatives and their corresponding nanoparticles were
3–86-fold smaller than that of gemcitabine HCl. In other words,
the gemcitabine derivatives, alone or in nanoparticles, were more
cytotoxic to the CCRF-CEM/dCK−/− cells than gemcitabine HCl. To
further confirm this finding, an apoptosis assay was performed in
3.3.1. Lipophilic gemcitabine derivatives and their nanoparticles
can overcome dCK deficiency
The in vitro cytotoxicities of the gemcitabine derivatives and
their nanoparticles in human leukemia cell line, CCRF-CEM, and its
derivative line, CCRF-CEM/dCK−/−, were evaluated and compared
to that of gemcitabine HCl. The IC50 value of gemcitabine HCl in the
parent CCRF-CEM cells was 2.9 1.8 nM (Table 1), which was 6–77-
fold smaller than that of the gemcitabine derivatives, GemC18, 8,
9, 12, and the derivatives in nanoparticles, GemC18-NPs, 8-NPs, 9-
NPs, 12-NPs, and 13-NPs (Table 1), demonstrating that in CCRF-CEM
cells, gemcitabine HCl was more cytotoxic than the gemcitabine
derivatives, alone or in nanoparticles. Overall, this finding is in

D.S.P. Lansakara-P. et al. / International Journal of Pharmaceutics 429 (2012) 123–134
129
Fig. 1. Characterization of gemcitabine derivative-containing nanoparticles. (A) The size and zeta potential of the nanoparticles, blank NPs, GemC18-NPs, 8-NPs, 9-NPs,
12-NPs, and 13-NPs. (B) The dynamic light scattering spectra of the 8-in-Tween 20 micelles (left peak) and 8-NPs (right peak) overlaid. (C) In GPC, 8-NPs (ꢁ) eluted in much
earlier fractions (fractions 8–11) than 8 in Tween 20 micelles (᭹) (fractions 17–24). The concentrations of the 8 in Tween 20 micelles and in 8-NPs were 100 ␮g/mL and
1 mg/mL, respectively. (D) The release profile of 8 from 8-NPs. As a control, the diffusion of compound 8 (8 in micelles) through the dialysis membrane was also measured.
Data shown are mean S.D (n = 3 for A and C, 6 for D). Standard deviations were not shown in C and D for clarity. (For interpretation of the references to color in this figure
legend, the reader is referred to the web version of this article.)
CCRF-CEM/dCK−/− cellsby stainingthemwithannexinVand7-AAD
after they were incubated for 72 h with gemcitabine HCl (10 ␮M),
8-NPs (10 ␮M), 8-free nanoparticles, or PBS. Similar to gemcitabine
HCl, 8-NPs induced tumor cells to undergo apoptosis (Fig. 2), but
the 8-NPs were significantly more effective than gemcitabine HCl in
inducing apoptosis (Fig. 2). Therefore, the gemcitabine derivatives
and their nanoparticles are less dependent on dCK to be active than
gemcitabine HCl. The finding with the gemcitabine derivatives in
nanoparticles is new, and the finding with the derivatives alone is
consistent with previous data generated in dCK over-expressing or
dCK deficient cells using other phospholipid gemcitabine deriva-
tives and gemcitabine phosphoramidate (Alexander et al., 2005;
Wu et al., 2007).
to show that the monophosphorylated gemcitabine derivatives (i.e.
8, 9, and 12) were less dependent on the dCK to be active than
GemC18, which is not monophosphorylated (Table 1). Importantly,
it appears that the incorporation of the gemcitabine derivatives into
nanoparticles made the monophosphorylated gemcitabine deriva-
tives further less dependent on the dCK to be active (Table 1). For
example, for the gemcitabine derivatives alone, the ratio of the
IC50 value of CCRF-CEM/dCK−/− to CCRF-CEM was 2438 for the
GemC18, 196–1755 for the other monophosphorylated derivatives
(Table 1). However, for the gemcitabine derivatives in nanoparti-
cles, the ratio was 816 for the GemC18-NPs, but only 25–68 for
8-NPs, 9-NPs, 12-NPs, and 13-NPs (Table 1). Gemcitabine is phos-
phorylated by dCK, and the monophosphorylation of gemcitabine
is the rate limiting step in the activation of gemcitabine (Ueno
et al., 2007; Mini et al., 2006). Therefore, it was expected that the
Moreover, the ratios of the IC50 value of the same compound in
the CCRF-CEM/dCK−/− to that in the parent CCRF-CEM cells seem
Table 1
The IC50 values of gemcitabine, gemcitabine derivatives, and the derivatives in nanoparticles in CCRF-CEM and CCRF-CEM/dCK−/− cells.
CCRF-CEM (nM)
CCRF-CEM/dCK−/− (␮M)
Ratio of IC50 in CCRF-CEM/dCK−/−
to that in CCRF-CEM
Gemcitabine HCl
2.9 1.8
240.4 29.0
82,897
GemC18
8
9
12
19.4 13.3
195.9 28.1
58.1 18.5
49.8 5.9
47.3 8.7
38.3 1.2
62.3 5.3
87.4 10.1
2438
196
1072
1755
GemC18-NPs
8-NPs
9-NPs
12-NPs
13-NPs
16.3 4.5
152.1 15.7
86.5 9.4
58.8 15.6
222.2 68.7
13.3 2.6
3.8 0.2
5.9 1.4
2.8 0.2
11.3 3.5
816
25
68
48
51

130
D.S.P. Lansakara-P. et al. / International Journal of Pharmaceutics 429 (2012) 123–134
Fig. 2. Flow cytometric graphs of CCRF-CEM-dCK−/− cells after 72 h of incubation with gemcitabine HCl (10 ␮M) or 8-NPs followed by staining with Annexin V and 7-AAD.
Numbers in the quadrates are % of cells (mean from 3 replicates). Upper right quadrant represents cells in late apoptotic stage; lower right, cells in early apoptotic stage;
lower left, viable cells.
monophosphorylated gemcitabine derivatives are less dependent
on dCK to be active than the GemC18. Interestingly, it appears that
the combination of monophosphorylation of gemcitabine and the
incorporation of the lipophilic monophosphorylated gemcitabine
derivative into nanoparticles can more effectively bypass the rate
limiting step of phosphorylation in gemcitabine activation.
To further validate this finding, the cytotoxicities of gemcitabine
HCl and selected gemcitabine derivatives in nanoparticles were
evaluated in another dCK deficient cell line, the murine leukemia
cells L1210 10K. The IC50 value of gemcitabine HCl in the parent
L1210 wt cells was 1.3 0.3 nM, which was 17,046-fold smaller
than the IC50 value of gemcitabine HCl in the dCK deficient L1210
10K cells (22.2 3.7 ␮M) (Table 2). Interestingly, in the L1210 10K
cells, GemC18-NPs and 8-NPs were 4- and 8-fold more cytotoxic
than gemcitabine HCl, respectively (Table 2). In addition, the IC50
values of GemC18-NPs and 8-NPs in the L1210 10K cells were only
17–431-fold greater than that in the L1210 wt cells (Table 2), further
confirming that the incorporation of the gemcitabine derivatives in
nanoparticles makes them less dependent on dCK to be active. We
did not investigate the mechanism of hydrolysis of the monophos-
phorylated gemcitabine derivatives, but it is likely that they were
hydrolyzed between the lipophilic chain and the phosphate group,
similar to the hydrolysis of 1-␤-d-arabinofuranosylcytosine (ara-C)
or other gemcitabine phospholipid derivatives (Raetz et al., 1977;
Alexander and Kucera, 2005).
3.3.2. Lipophilic gemcitabine derivatives and their nanoparticles
can overcome gemcitabine resistance related to RRM1
over-expression
RRM1 plays a substantial role in DNA synthesis and gemcitabine
resistance (Bergman et al., 2005; Ceppi et al., 2006; Davidson et al.,
2004; Goan et al., 1999; Ohtaka et al., 2008; Rosell et al., 2004;
Yen, 2003). Previously, we developed a tumor cell line that over-
expresses RRM1 (TC-1-GR) (Chung et al., 2012). In TC-1-GR cells,
GemC18-NPs were significantly more toxic than gemcitabine HCl,
although in the parent TC-1 cells, GemC18-NPs were significantly
less toxic than gemcitabine HCl (Chung et al., 2012). Importantly,
in mice with pre-established TC-1-GR tumors, GemC18-NPs signif-
icantly inhibited the tumor growth, but gemcitabine HCl did not
show any significant anti-tumor activity (Chung et al., 2012). In
the present study, the IC50 values of the new lipophilic monophos-
phorylated gemcitabine derivatives and their nanoparticles in
both TC-1 and TC-1-GR cells were determined to evaluate their
ability to overcome gemcitabine resistance caused by RRM1 over-
expression. As expected, in TC-1 cells, gemcitabine HCl was more
cytotoxic (IC50, 14.7 2.8 nM) than the gemcitabine derivatives and
their nanoparticles (Table 3). However, in TC-1-GR cells, the major-
ity of gemcitabine derivatives (except 12) and all the gemcitabine
derivatives in nanoparticles were more cytotoxic than gemc-
itabine HCl (2- to 10-fold) (Table 3). Importantly, in TC-1-GR cells,
the IC50 value of gemcitabine HCl was 36.7 5.1 ␮M, which was

D.S.P. Lansakara-P. et al. / International Journal of Pharmaceutics 429 (2012) 123–134
131
Table 2
The IC50 values of gemcitabine HCl, gemC18-NPs, and 8-NPs in L1210 wt and L1210 10K cells.
L1210 wt (nM)
L1210 10K (␮M)
Ratio of IC50 values in
L1210 10K cellsa
Ratio of IC50 in L1210 10K
to that in L1210 wt
Gemcitabine HCl
GemC18-NPs
8-NPs
1.3 0.3
13.1 0.3
172.5 55.2
22.2 3.7
5.6 0.1
2.9 0.3
1
4
8
17,046
431
17
a
Ratio is the IC50 values of gemcitabine HCl divided by that of the nanoparticles.
2497-fold greater than that in TC-1 cells. In contrast, the IC50 val-
ues of GemC18, 8, 9, 12 and the nanoparticles, GemC18-NPs, 8-NPs,
9-NPs, 12-NPs, and 13-NPs, in TC-1-GR cells were only 23- to 177-
fold greater than that in TC-1 cells (Table 3), demonstrating that
the gemcitabine derivatives and their nanoparticles are less sen-
sitive to gemcitabine resistance caused by RRM1-over-expression
than gemcitabine HCl. Incorporation of the gemcitabine derivatives
into nanoparticles tended to make the derivatives, particularly the
GemC18, more cytotoxic to the RRM1-over-expressing TC-1-GR
cells. However, unlike what was observed in the dCK deficient cells
(Tables 1 and 2), it seemed that in RRM1-over-expressing TC-1-GR
cells, monophosphorylation of gemcitabine did not add significant
additional benefits compare with GemC18.
them more cytotoxic cannot be drawn. Previously, Duxbury et al.
reported that the IC50 values of gemcitabine in MIA PaCa-2 cells and
PANC-1 cells were 40 nM and 50 nM, respectively. The IC50 value
of gemcitabine HCl in MIA PaCa-2 cells determined in the present
study is comparable to what was reported (Duxbury et al., 2003),
but the PANC-1 cells were significantly more resistant to gemc-
itabine HCl in our study. Data in Fig. 3B showed that after 0.5 h
of incubation, only about 0.15% of gemcitabine was taken up by
PANC-1 cells, in contrast to 10% and 6% of 8 in nanoparticles or in
solution. Therefore the low uptake of gemcitabine by PANC-1 cells
was probably related to its resistance to gemcitabine. However, it
needs to be noted that the low percentage of gemcitabine detected
in the PANC-1 cells could also be due to the rapid deamination of
gemcitabine. Moreover, it is known that PANC-1 cells overexpress
RRM1 and RRM2, which are known determinants of gemcitabine
resistance (Jordheim et al., 2005; Boukovinas et al., 2008).
Finally, similar results were obtained in the MCF-7 human breast
adenocarcinoma cell line. Gemcitabine HCl, at a concentration as
high as 400 ␮M, was not sufficient to kill half of the MCF-7 cells
after 72 h of incubation (see Supplementary data, Table S1), and
data from a trypan blue exclusion assay showed that more than
50% of MCF-7 cells were still alive, even after 116 h of incubation
with 400 ␮M of gemcitabine HCl. However, all the gemcitabine
derivatives, except 12, and all gemcitabine derivatives in nanopar-
ticles were significantly cytotoxic to the MCF-7 cells (IC50 values,
2–31 ␮M) (Table S1). The mechanism of gemcitabine resistance in
the MCF-7 cells is unknown, but the data obtained in the MCF-7 cells
clearly showed that the gemcitabine derivatives and the derivatives
in nanoparticles can overcome that resistance mechanism.
3.3.3. Cytotoxicities of gemcitabine derivatives and their
nanoparticles in cancer cells over-expressing different levels of
RRM2
It was reported that MIA PaCa-2 and PANC-1 cells both over-
expressed RRM2, but PANC-1 cells express ∼70% more RRM2 than
MIA PaCa-2 (Duxbury et al., 2003). In MIA PaCa-2 cells, the IC50
value of gemcitabine HCl and GemC18 NPs were 49.7 17.7 nM and
40.6 8.2 nM, respectively (Table 4), all other gemcitabine deriva-
tives and their nanoparticles were less toxic than gemcitabine HCl
(Table 4). However, more than 50% of PANC-1 cells were still alive
after 96 h of incubation with 400 ␮M of gemcitabine HCl (Table 1
and Fig. 3A). Data from a trypan blue exclusion assay showed that it
took 116 h to kill 50% of PANC-1 cells with 400 ␮M of gemcitabine
HCl (Fig. 3A). The IC50 values of the gemcitabine derivatives and
their nanoparticles in PANC-1 cells were 5.8–58.7 ␮M (Table 4).
The IC50 values of gemcitabine HCl in PANC-1 cells was more
than 8000-fold greater than that in the MIA PaCa-2 cells, but the
IC50 values of gemcitabine derivatives and their nanoparticles in
PANC-1 cells were only 24- to 239-fold greater than those in MIA
PaCa-2 cells (Table 4). In other words, the gemcitabine deriva-
tives and their nanoparticles were less sensitive than gemcitabine
HCl to gemcitabine resistance caused by RRM2 over-expression.
Again, monophosphorylation of gemcitabine did not add addi-
tional benefits in its cytotoxicity against the RRM2-over-expressing
PANC-1 cells, and except for the GemC18-NPs, a conclusion that the
incorporation of gemcitabine derivatives into nanoparticles makes
3.3.4. Cytotoxicities of gemcitabine derivatives and their
nanoparticles in nucleoside transporter deficient cancer cells
It is known that nucleoside transporters are a prerequisite for
the cellular uptake of gemcitabine (Damaraju et al., 2003). There-
fore, in the hENT1 deficient CCRF-CEM-AraC-8C cells, the IC50 value
of the gemcitabine HCl was 998.8 9.4 nM, 344 times greater than
that in the parent CCRF-CEM cells (Table 5). However, the IC50
values of gemcitabine derivatives and their nanoparticles in CCRF-
CEM-AraC-8C cells were only 4- to 35-fold greater than that in the
parent CCRF-CEM cells (Table 5), indicating that the gemcitabine
Table 3
The IC50 values of gemcitabine HCl, gemcitabine derivatives, and the derivatives in nanoparticles in TC-1 and TC-1-GR cells.
TC-1 (nM)
TC-1-GR (␮M)
Ratio of IC50 values in
TC-1-GR cellsa
Ratio of IC50 in TC-1-GR
to that in TC-1
Gemcitabine HCl
14.7 2.8
36.7 5.1
1
2497
GemC18
8
9
12
132.1 17.2
245.5 39.9
210.7 85.0
430.0 52.4
7.7 2.4
21.1 1.7
9.3 2.5
5
2
4
0.5
58
86
44
76.3 10.5
177
GemC18-NPs
8-NPs
9-NPs
12-NPs
13-NPs
59.8 18.4
371.3 27.3
258.6 60.9
405.7 114.1
395.5 40.7
3.6 0.2
8.4 0.5
10.1 0.9
10.5 2.0
9.0 2.6
10
60
23
39
26
23
4
4
3
4
a
Ratio is the IC50 values of gemcitabine HCl divided by that of the derivatives or derivatives in nanoparticles.

132
D.S.P. Lansakara-P. et al. / International Journal of Pharmaceutics 429 (2012) 123–134
Table 4
The IC50 values of gemcitabine HCl, gemcitabine derivatives, and the derivatives in nanoparticles in MIA PaCa-2 and PANC-1 cells.
MIA PaCa-2 (nM)
PANC-1 (␮M)
Ratio of IC50 in PANC-1 to
that in MIA PaCa-2
Gemcitabine HCl
49.7 17.7
> 400
>8000
GemC18
8
9
12
133.0 60.4
835.9 163.8
204.6 39.5
245.1 38.0
6.0 1.1
50.4 4.3
38.5 5.7
58.7 14.2
45
60
188
239
GemC18-NPs
8-NPs
9-NPs
12-NPs
13-NPs
40.6 8.2
5.8 0.6
6.2 1.5
8.8 1.8
22.0 3.5
8.7 2.9
143
31
30
58
24
201.7 50.2
290.0 84.6
380.0 98.3
357.6 172.2
12
120
80
40
0
A
B
10
8
6
4
Gemcitabine (0 µM)
Gemcitabine (400 µM)
2
0
0
24
48
72
96
120
8-NPs
8
GEM
Time (h)
Fig. 3. (A) Percentage of PANC-1 cells alive after incubation with gemcitabine HCl at 400 ␮M for up to 116 h. Cell viability was determined using a trypan blue exclusion
assay. The experiment was repeated twice. (B) In vitro uptake of 8 in nanoparticles, 8 in solution, or gemcitabine HCl (GEM) by PANC-1 cells. The concentration of compound
8 in the 8-NPs and 8 in solution was 40 ␮M, and the gemcitabine HCl concentration was 400 ␮M. Gemcitabine uptake was not detected when cells were incubated with
40 ␮M of gemcitabine HCl. The incubation time was 0.5 h. Data shown are mean S.D (n = 3).
derivatives, alone or in nanoparticles, are less sensitive to hENT1
deficient than gemcitabine HCl, possibly because the lipophilized
gemcitabine derivatives can diffuse into cells without the help of
the nucleoside transporters, and the gemcitabine derivatives in
nanoparticles can be taken up by cells via endocytosis. However,
the cytotoxicities of the monophosphorylated gemcitabine deriva-
tives and their nanoparticles in the CCRF-CEM-AraC-8C cells are not
significantly different from that of gemcitabine HCl (Table 5). Only
the GemC18 and GemC18-NPs were 2.7- and 12.3-fold more cyto-
toxic than gemcitabine HCl, respectively, in the hENT1-deficient
CCRF-CEM-AraC-8C cells (Table 5), which was consistent with our
previous data (Chung et al., 2012). It is likely that the phosphate
group on the gemcitabine derivatives made them less effective in
entering the hENT1 deficient cells.
3.4. Gemcitabine HCl, but not gemcitabine derivatives in
nanoparticles, inhibits dCDA activity
Previously, Bouffard et al. reported that gemcitabine HCl, as
a substrate to dCDA, competitively inhibits the deamination of
dCyd by dCDA (Bouffard et al., 1993). In order to test whether
the gemcitabine derivatives are still good substrates of dCDA and
inhibit its activity, dCDA was partially purified from BxPC-3 human
pancreatic cancer cells, and its deamination activity against dCyd
was determined in the presence or absence of gemcitabine HCl
or selected gemcitabine derivatives in nanoparticles (Bergman
et al., 2004; Laliberté and Momparler, 1994; Ruiz van Haperen
et al., 1993). The nanoparticles, but not the gemcitabine derivatives
alone, were used due to the poor water solubility of the gemcitabine
Table 5
The IC50 values of gemcitabine HCl, gemcitabine derivatives, and the derivatives in nanoparticles in CCRF-CEM and CCRF-CEM-AraC-8C cells.
CCRF-CEM (nM)
CCRF-CEM-AraC-8C (nM)
Ratio of IC50 values in
CEM-AraC-8C cellsa
Ratio of IC50 in CCRF-CEM-AraC-8C
to that in CCRF-CEM
Gemcitabine HCl
2.9 1.8
998.8 9.4
1
333
GemC18
8
9
12
19.4 13.3
195.9 28.1
58.1 18.5
49.8 5.9
369.4 235.1
806.5 111.8
575.3 97.1
2.7
1.2
1.7
0.6
19
4
10
35
1766.9 532.8
GemC18-NPs
8-NPs
9-NPs
12-NPs
13-NPs
16.3 4.5
152.1 15.7
86.5 9.4
58.8 15.6
222.2 68.7
81.0 17.9
942.6 163.8
1325.1 167.2
712.9 342.9
1166.6 293.1
12.3
1.1
0.8
1.4
0.9
5
4
13
12
5
a
Ratio is the IC50 values of gemcitabine HCl divided by that of the derivatives or derivatives in nanoparticles.

D.S.P. Lansakara-P. et al. / International Journal of Pharmaceutics 429 (2012) 123–134
133
Fig. 4. dCDA assay. (A) Effect of gemcitabine concentration on the conversion of dCyd to dUrd (ratio, w/w) by dCDA; (B) effect of GemC18-NPs and 8-NPs on the conversion
of dCyd to dUrd by dCD (No, no inhibitor; GEM, gemcitabine HCl, 1.7 mM). The molar concentration of GemC18 and 8 in the nanoparticles was 1.7 mM. Blank nanoparticles
did not inhibit the conversion. Data shown are mean S.D. The experiment was repeated at least twice with similar results.
derivatives. GemC18-NPs and 8-NPs were chosen because of their
greater cytotoxicity in previous in vitro cytotoxicity assays. dUrd
was not detected in the control reaction (with dCyd, but no dCDA)
indicating that any observed dUrd would be due to the deamina-
tion of dCyd by dCDA. As shown in Fig. 4A, dCyd was converted to
dUrd in the presence of the partially purified dCDA. Gemcitabine
HCl competitively inhibited the conversion of dCyd to dUrd, and
the extent of the inhibition was increased by increasing the con-
centration of gemcitabine HCl (Fig. 4A). However, GemC18-NPs and
8-NPs did not significantly inhibit the deamination activity of dCDA
(Fig. 4B), confirming that gemcitabine HCl, but not GemC18-NPs
and 8-NPs, can competitively inhibit the deamination of dCyd by
dCDA. Therefore, the gemcitabine derivatives in nanoparticles are
no longer good substrates of dCDA, and it is expected that they can
potentially overcome gemcitabine resistance caused by deamina-
tion. This finding is in agreement with data from previous studies,
showing that other gemcitabine derivatives were no longer good
substrates of dCDA as well (Bergman et al., 2004; Song et al., 2005).
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.ijpharm.2012.03.014.
References
Alexander, R.L., Greene, B.T., Torti, S.V., Kucera, G.L., 2005. A novel phospholipid gem-
citabine conjugate is able to bypass three drug-resistance mechanisms. Cancer
Chemother. Pharmacol. 56, 15–21.
Alexander, R.L., Kucera, G.L., 2005. Lipid nucleoside conjugates for the treatment of
cancer. Curr. Pharm. Des. 11, 1079–1089.
Alexander,
Kucera, G.L., 2003. Synthesis and cytotoxic activity of two novel
1-dodecylthio-2-decyloxypropyl-3-phosphatidic acid conjugates with
gemcitabine and cytosine arabinoside. J. Med. Chem. 46, 4205–4208.
R.L.,
Morris-Natschke,
S.L.,
Ishaq,
K.S.,
Fleming,
R.A.,
Andersson, R., Aho, U., Nilsson, B.I., Peters, G.J., Pastor-Anglada, M., Rasch, W.,
Sandvold, M.L., 2009. Gemcitabine chemoresistance in pancreatic cancer:
molecular mechanisms and potential solutions. Scand. J. Gastroenterol. 44,
782–786.
Barton-Burke, M., 1999. Gemcitabine: a pharmacologic and clinical overview. Cancer
Nurs. 22, 176–183.
Bergman, A.M., Eijk, P.P., Ruiz van Haperen, V.W.T., Smid, K., Veerman, G., Hubeek,
I., van den IJssel, P., Ylstra, B., Peters, G.J., 2005. In vivo induction of resistance to
gemcitabine results in increased expression of ribonucleotide reductase subunit
M1 as the major determinant. Cancer Res. 65, 9510–9516.
Bergman, A.M., Kuiper, C.M., Voorn, D.A., Comijn, E.M., Myhren, F., Sandvold, M.L.,
Hendriks, H.R., Peters, G.J., 2004. Antiproliferative activity and mechanism of
action of fatty acid derivatives of arabinofuranosylcytosine in leukemia and solid
tumor cell lines. Biochem. Pharmacol. 67, 503–511.
Bergman, A.M., Pinedo, H.M., Peters, G.J., 2002. Determinants of resistance to
2ꢀ,2ꢀ-difluorodeoxycytidine (gemcitabine). Drug Resist. Updat. 5, 19–33.
Bligh, E.G., Dyer, W.J., 1959. Lipid extraction of C. elegans. Can. J. Biochem. Physiol.
37, 911–917.
Bouffard, D.Y., Laliberté, J., Momparler, R.L., 1993. Kinetic studies on
2ꢀ,2ꢀ-difluorodeoxycytidine (gemcitabine) with purified human deoxycytidine
kinase and cytidine deaminase. Biochem. Pharmacol. 45, 1857–1861.
Boukovinas, I., Papadaki, C., Mendez, P., Taron, M., Mavroudis, D., Koutsopoulos, A.,
Sanchez-Ronco, M., Sanchez, J.J., Trypaki, M., Staphopoulos, E., Georgoulias, V.,
Rosell, R., Souglakos, J., 2008. Tumor BRCA1, RRM1 and RRM2 mRNA expres-
sion levels and clinical response to first-line gemcitabine plus docetaxel in
non-small-cell lung cancer patients. PLoS ONE 3, e3695.
4. Conclusions
In the present study, four novel lipophilic monophosphory-
lated gemcitabine derivatives were synthesized and incorporated
into solid lipid nanoparticles. All the gemcitabine derivatives
and their nanoparticles showed a significantly higher cytotoxic-
ity than gemcitabine HCl in cells that are deficient in dCK, and
the gemcitabine derivatives in nanoparticles were more cytotoxic
than the corresponding gemcitabine derivatives. The majority of
the gemcitabine derivatives and all nanoparticles are also more
cytotoxic than gemcitabine HCl to cancer cells that over-express
RRM1 or RRM2. Finally, the gemcitabine derivatives in nanopar-
ticles were no longer good substrates to dCDA and thus became
resistance to deamination. Collectively, the 2ꢀ-2ꢀ-difluoro-cytosine-
5ꢀ-octadecylphosphate (8) in nanoparticles showed the highest
cytotoxicity to cells that are deficient in dCK, over-expressing
RRM1, or over-expressing RRM2, and were resistant to deami-
nation. Future in vivo studies to evaluate its ability to overcome
multiple gemcitabine resistant mechanisms are warranted.
Candelaria, M., Cetina, L., de la Garza, J., Duen˜as-Gonzálz, A., 2007. Clinical implica-
tions of gemcitabine in the treatment of cervical cancer. Eur. J. Cancer Suppl. 5,
37–43.
Candelaria, M., de la Cruz-Hernández, E., Pérez-Cárdenas, E., Trejo-Becerril, C.,
Gutiérrez-Hernández, O., Duen˜as-González, A., 2010. Pharmacogenetics and
pharmacoepigenetics of gemcitabine. Med. Oncol. 27, 1133–1143.
Ceppi, P., Volante, M., Novello, S., Rapa, I., Danenberg, K., Danenberg, P., Cambieri,
A., Selvaggi, G., Saviozzi, S., Calogero, R., Papotti, M., Scagliotti, G., 2006. ERCC1
and RRM1 gene expressions but not EGFR are predictive of shorter survival in
advanced non-small-cell lung cancer treated with cisplatin and gemcitabine.
Ann. Oncol. 17, 1818–1825.
Acknowledgment
Cetina, L., Rivera, L., Candelaria, M., de la Garza, J., Duen˜as-González, A., 2004.
Chemoradiation with gemcitabine for cervical cancer in patients with renal
failure. Anticancer Drugs 15, 761–766.
This work was supported in part by a National Cancer Institute
grant (CA135274) to Z.C.

134
D.S.P. Lansakara-P. et al. / International Journal of Pharmaceutics 429 (2012) 123–134
Chou, T.C., Talalay, P., 1984. Quantitative analysis of dose–effect relationships: the
combined effects of multiple drugs or enzyme inhibitors. Adv. Enzyme Regul.
22, 27–55.
Chung, W.-G., Sloat, B.R., Sandoval, M.A., Lansakara-P, D.S.P., Cui, Z., 2012. Stearoyl
gemcitabine nanoparticles overcome resistance related to the over-expression
of ribonucleotide reductase subunit M1. J. Control. Release 157, 132–140.
Damaraju, V.L., Damaraju, S., Young, J.D., Baldwin, S.A., Mackey, J., Sawyer, M.B., Cass,
C.E., 2003. Nucleoside anticancer drugs: the role of nucleoside transporters in
resistance to cancer chemotherapy. Oncogene 22, 7524–7536.
Davidson, J.D., Ma, L., Flagella, M., Geeganage, S., Gelbert, L.M., Slapak, C.A., 2004.
An Increase in the expression of ribonucleotide reductase large subunit 1 is
associated with gemcitabine resistance in non-small cell lung cancer cell lines.
Cancer Res. 64, 3761–3766.
Duxbury, M.S., Ito, H., Zinner, M.J., Ashley, S.W., Whang, E.E., 2003. RNA interfer-
ence targeting the M2 subunit of ribonucleotide reductase enhances pancreatic
adenocarcinoma chemosensitivity to gemcitabine. Oncogene 23, 1539–1548.
Giovannetti, E., Del Tacca, M., Mey, V., Funel, N., Nannizzi, S., Ricci, S., Orlandini,
C., Boggi, U., Campani, D., Del Chiaro, M., Iannopollo, M., Bevilacqua, G., Mosca,
F., Danesi, R., 2006. Transcription analysis of human equilibrative nucleoside
transporter-1 predicts survival in pancreas cancer patients treated with gemc-
itabine. Cancer Res. 66, 3928–3935.
Mey, V., Giovannetti, E., Braud, F.D., Nannizzi, S., Curigliano, G., Verweij, F., Cobelli,
O.D., Pece, S., Tacca, M.D., Danesi, R., 2006. In vitro synergistic cytotoxicity of
gemcitabine and pemetrexed and pharmacogenetic evaluation of response to
gemcitabine in bladder cancer patients. Br. J. Cancer 95, 289–297.
Mini, E., Nobili, S., Caciagli, B., Landin, I.I., Mazzei, T., 2006. Cellular pharmacology of
gemcitabine. Ann. Oncol. 17, v7–v12.
Ohtaka, K., Kohya, N., Sato, K., Kitajima, Y., Ide, T., Mitsuno, M., Miyazaki, K., 2008.
Ribonucleotide reductase subunit M1 is a possible chemoresistance marker to
gemcitabine in biliary tract carcinoma. Oncol. Rep. 20, 279–286.
Perie, J., Monsan, P.F., Willson, M.J., Klaebe, A., Lauth, N.J., Thibault, P.A., 1990. Process
for the preparation of lysophosphatidic acids and of salts of the latter FR. Patent
2636331.
Raetz, C.R., Chu, M.Y., Srivastava, S., Turcotte, J.G., 1977. A phospholipid derivative
of cytosine arabinoside and its conversion to phosphatidylinositol by animal
tissue. Science 196, 303–305.
Rosell, R., Danenberg, K.D., Alberola, V., Bepler, G., Sanchez, J.J., Camps, C., Proven-
cio, M., Isla, D., Taron, M., Diz, P., Artal, A., 2004. Ribonucleotide reductase
messenger RNA expression and survival in gemcitabine/cisplatin-treated
advanced non-small cell lung cancer patients. Clin. Cancer Res 10,
1318–1325.
Ruiz van Haperen, V.W., Veerman, G., Braakhuis, B.J., Vermorken, J.B., Boven,
E., Leyva, A., Peters, G.J., 1993. Deoxycytidine kinase and deoxycytidine
deaminase activities in human tumour xenografts. Eur. J. Cancer 29A,
2132–2137.
Goan, Y.-G., Zhou, B., Hu, E., Mi, S., Yen, Y., 1999. Overexpression of ribonucleotide
reductase as a mechanism of resistance to 2,2-difluorodeoxycytidine in the
human KB cancer cell line. Cancer Res. 59, 4204–4207.
Gregoire, V., Rosier, J.F., De Bast, M., Bruniaux, M., De Coster, B., Octave-Prignot, M.,
Scalliet, P., 2002. Role of deoxycytidine kinase (dCK) activity in gemcitabine’s
radioenhancement in mice and human cell lines in vitro. Radiother. Oncol. 63,
329–338.
Sandler, A.B., Nemunaitis, J., Denham, C., von Pawel, J., Cormier, Y., Gatzemeier, U.,
Mattson, K., Manegold, C., Palmer, M.C., Gregor, A., Nguyen, B., Niyikiza, C., Ein-
horn, L.H., 2000. Phase III trial of gemcitabine plus cisplatin versus cisplatin alone
in patients with locally advanced or metastatic non-small-cell lung cancer. J.
Clin. Oncol. 18, 122.
Sezgin, C., Karabulut, B., Uslu, R., Sanli, U.A., Goksel, G., Yuzer, Y., Goker, E., 2005.
Gemcitabine treatment in patients with inoperable locally advanced/metastatic
pancreatic cancer and prognostic factors. Scand. J. Gastroenterol. 40, 1486–1492.
Sloat, B.R., Sandoval, M.A., Hau, A.M., He, Y., Cui, Z., 2010. Strong antibody responses
induced by protein antigens conjugated onto the surface of lecithin-based
nanoparticles. J. Control. Release 141, 93–100.
Guo, Z.-w., Gallo, J.M., 1999. Selective protection of 2ꢀ,2ꢀ-difluorodeoxycytidine
(gemcitabine). J. Org. Chem. 64, 8319–8322.
Heinemann, V., Schulz, L., Issels, R.D., W., P., 1995. Gemcitabine: a modulator of intra-
cellular nucleotide and deoxynucleotide metabolism. Semin. Oncol. 22, 11–18.
Heinemann, V., Xu, Y.-Z., Chubb, S., Sen, A., Hertel, L.W., Grindey, G.B., Plunkett,
W., 1992. Cellular elimination of 2ꢀ,2ꢀ-difluorodeoxycytidine 5ꢀ-triphosphate: a
mechanism of self-potentiation. Cancer Res. 52, 533–539.
Huang, P., Plunkett, W., 1995. Induction of apoptosis by gemcitabine. Semin. Oncol.
4, 19–25.
Sloat, B.R., Sandoval, M.A., Li, D., Chung, W.-G., Lansakara-P, D.S.P., Proteau, P.J.,
Kiguchi, K., DiGiovanni, J., Cui, Z., 2011. In vitro and in vivo anti-tumor activ-
ities of a gemcitabine derivative carried by nanoparticles. Int. J. Pharm. 409,
278–288.
Song, X., Lorenzi, P.L., Landowski, C.P., Vig, B.S., Hilfinger, J.M., Amidon, G.L., 2005.
Amino acid ester prodrugs of the anticancer agent gemcitabine: synthesis, bio-
conversion, metabolic bioevasion, and hPEPT1-mediated transport. Mol. Pharm.
2, 157–167.
Immordino, M.L., Brusa, P., Rocco, F., Arpicco, S., Ceruti, M., Cattel, L., 2004.
Preparation, characterization, cytotoxicity and pharmacokinetics of liposomes
containing lipophilic gemcitabine prodrugs. J. Control. Release 100, 331–346.
Jordheim, L.P., Guittet, O., Lepoivre, M., Galmarini, C.M., Dumontet, C., 2005.
Increased expression of the large subunit of ribonucleotide reductase is involved
in resistance to gemcitabine in human mammary adenocarcinoma cells. Mol.
Cancer Ther. 4, 1268–1276.
Kalykaki, A., Papakotoulas, P., Tsousis, S., Boukovinas, I., Kalbakis, K., Vamvakas, L.,
Kotsakis, A., Vardakis, N., Papadopoulou, P., Georgoulias, V., Mavroudis, D., 2008.
Gemcitabine plus oxaliplatin (GEMOX) in pretreated patients with advanced
ovarian cancer: a multicenter Phase II Study of the Hellenic Oncology Research
Group (HORG). Anticancer Res. 28, 495–500.
Kroep, J.R., Loves, W.J.P., van der Wilt, C.L., Alvarez, E., Talianidis, I., Boven, E.,
Braakhuis, B.J.M., van Groeningen, C.J., Pinedo, H.M., Peters, G.J., 2002. Pre-
treatment deoxycytidine kinase levels predict in vivo gemcitabine sensitivity.
Supported by Eli Lilly & Co., International, The Netherlands. Mol. Cancer Ther. 1,
371–376.
Laliberté, J., Momparler, R.L., 1994. Human cytidine deaminase: purification of
enzyme, cloning, and expression of its complementary DNA. Cancer Res. 54,
5401–5407.
Spratlin, J., Sangha, R., Glubrecht, D., Dabbagh, L., Young, J.D., Dumontet, C.,
Cass, C., Lai, R., Mackey, J.R., 2004. The absence of human equilibrative
nucleoside transporter
1 is associated with reduced survival in patients
with gemcitabine-treated pancreas adenocarcinoma. Clin. Cancer Res. 10,
6956–6961.
Ueno, H., Kiyosawa, K., Kaniwa, N., 2007. Pharmacogenomics of gemcitabine: can
genetic studies lead to tailor-made therapy? Br. J. Cancer 97, 145–151.
Watanabe, K.A., Fo, J.J., 1966. A simple method for selective acylation of cytidine of
the 4-amino group. J. Angew. Chem. Int. Ed. 5, 579–580.
Wu, W., Sigmond, J., Peters, G.J., Borch, R.F., 2007. Synthesis and biological activity
of a gemcitabine phosphoramidate prodrug. J. Med. Chem. 50, 3743–3746.
Yen, Y., 2003. Ribonucleotide reductase subunit one as gene therapy target. Clin.
Cancer Res. 9, 4304–4308.
Zucali, P.A., Ceresoli, G.L., Garassino, I., De Vincenzo, F., Cavina, R., Campagnoli,
E., Cappuzzo, F., Salamina, S., Soto Parra, H.J., Santoro, A., 2008. Gemcitabine
and vinorelbine in pemetrexed-pretreated patients with malignant pleural
mesothelioma. Cancer 112, 1555–1561.
Maréchal, R., Mackey, J.R., Lai, R., Demetter, P., Peeters, M., Polus, M., Cass, C.E.,
Salmon, I., Devière, J., Van Laethem, J.-L., 2010. Deoxycitidine kinase is associ-
ated with prolonged survival after adjuvant gemcitabine for resected pancreatic
adenocarcinoma. Cancer 116, 5200–5206.