A novel method for quantification of gemcitabine and its metabolites 2',2'-difluorodeoxyuridine and gemcitabine triphosphate in tumour tissue by LC-MS/MS: comparison with (19)F NMR spectroscopy.

Abstract

To develop a sensitive analytical method to quantify gemcitabine (2',2'-difluorodeoxycytidine, dFdC) and its metabolites 2',2'-difluorodeoxyuridine (dFdU) and 2',2'-difluorodeoxycytidine-5'-triphosphate (dFdCTP) simultaneously from tumour tissue.
Pancreatic ductal adenocarcinoma tumour tissue from genetically engineered mouse models of pancreatic cancer (KP ( FL/FL ) C and KP ( R172H/+) C) was collected after dosing the mice with gemcitabine. (19)F NMR spectroscopy and LC-MS/MS protocols were optimised to detect gemcitabine and its metabolites in homogenates of the tumour tissue.
A (19)F NMR protocol was developed, which was capable of distinguishing the three analytes in tumour homogenates. However, it required at least 100 mg of the tissue in question and a long acquisition time per sample, making it impractical for use in large PK/PD studies or clinical trials. The LC-MS/MS protocol was developed using porous graphitic carbon to separate the analytes, enabling simultaneous detection of all three analytes from as little as 10 mg of tissue, with a sensitivity for dFdCTP of 0.2 ng/mg tissue. Multiple pieces of tissue from single tumours were analysed, showing little intra-tumour variation in the concentrations of dFdC or dFdU (both intra- and extra-cellular). Intra-tumoural variation was observed in the concentration of dFdCTP, an intra-cellular metabolite, which may reflect regions of different cellularity within a tumour.
We have developed a sensitive LC-MS/MS method capable of quantifying gemcitabine, dFdU and dFdCTP in pancreatic tumour tissue. The requirement for only 10 mg of tissue enables this protocol to be used to analyse multiple areas from a single tumour and to spare tissue for additional pharmacodynamic assays.

Full text

Page 1

Cancer Chemother Pharmacol (2011) 68:1243–1253
DOI 10.1007/s00280-011-1613-0
123
ORIGINAL ARTICLE
A novel method for quantiWcation of gemcitabine and its
metabolites 2?,2?-diXuorodeoxyuridine and gemcitabine
triphosphate in tumour tissue by LC–MS/MS: comparison
with19F NMR spectroscopy
Tashinga E. Bapiro · Frances M. Richards · Mae A. Goldgraben · Kenneth P. Olive · Basetti Madhu ·
Kristopher K. Frese · Natalie Cook · Michael A. Jacobetz · Donna-Michelle Smith · David A. Tuveson ·
John R. GriYths · Duncan I. Jodrell
Received: 17 January 2011 / Accepted: 4 March 2011 / Published online: 23 March 2011
© The Author(s) 2011. This article is published with open access at Springerlink.com
Abstract
Purpose
quantify gemcitabine (2?,2?-diXuorodeoxycytidine, dFdC)
and its metabolites 2?,2?-diXuorodeoxyuridine (dFdU)
and 2?,2?-diXuorodeoxycytidine-5?-triphosphate (dFdCTP)
simultaneously from tumour tissue.
Methods
Pancreatic ductal adenocarcinoma tumour tissue
from genetically engineered mouse models of pancreatic
cancer (KPFL/FLC and KPR172H/+C) was collected after dos-
ing the mice with gemcitabine. 19F NMR spectroscopy and
LC–MS/MS protocols were optimised to detect gemcita-
bine and its metabolites in homogenates of the tumour
tissue.
Results
A 19F NMR protocol was developed, which was
capable of distinguishing the three analytes in tumour
homogenates. However, it required at least 100 mg of the
tissue in question and a long acquisition time per sample,
making it impractical for use in large PK/PD studies or
To develop a sensitive analytical method to
clinical trials. The LC–MS/MS protocol was developed
using porous graphitic carbon to separate the analytes,
enabling simultaneous detection of all three analytes from
as little as 10 mg of tissue, with a sensitivity for dFdCTP of
0.2 ng/mg tissue. Multiple pieces of tissue from single
tumours were analysed, showing little intra-tumour varia-
tion in the concentrations of dFdC or dFdU (both intra- and
extra-cellular). Intra-tumoural variation was observed in
the concentration of dFdCTP, an intra-cellular metabolite,
which may reXect regions of diVerent cellularity within a
tumour.
Conclusion
We have developed a sensitive LC–MS/MS
method capable of quantifying gemcitabine, dFdU and
dFdCTP in pancreatic tumour tissue. The requirement for
only 10 mg of tissue enables this protocol to be used to ana-
lyse multiple areas from a single tumour and to spare tissue
for additional pharmacodynamic assays.
Keywords
spectroscopy · Mass spectrometry
Gemcitabine · Pancreatic cancer · NMR
Introduction
Gemcitabine, 2?,2?-diXuorodeoxycytidine, dFdC, is an anti-
cancer drug licensed for the treatment of a number of solid
tumour types including pancreatic, non-small cell lung,
ovary, bladder and breast cancer.
Gemcitabine is a prodrug that is phosphorylated intra-
cellularly by deoxycytidine kinase (dCK), thymidine kinase
(TK2), dCMP kinase and nucleoside diphosphate kinase
into the active metabolites 2?,2?-diXuorodeoxycytidine-5?-
diphosphate (dFdCDP) and triphosphate (dFdCTP). The
cytotoxic eVects of gemcitabine have been attributed to a
combination of two mechanisms of action: (1) inhibition of
T. E. Bapiro · F. M. Richards (&) · M. A. Goldgraben ·
K. P. Olive · B. Madhu · K. K. Frese · N. Cook · M. A. Jacobetz ·
D.-M. Smith · D. A. Tuveson · J. R. GriYths · D. I. Jodrell
Cancer Research UK, Cambridge Research Institute,
Li Ka Shing Centre, Robinson Way, Box 278,
Cambridge CB2 0RE, UK
e-mail: Frances.Richards@cancer.org.uk
T. E. Bapiro · F. M. Richards · M. A. Goldgraben · N. Cook ·
M. A. Jacobetz · D. A. Tuveson · D. I. Jodrell
Department of Oncology, University of Cambridge,
Cambridge, UK
Present Address:
K. P. Olive
Herbert Irving Comprehensive Cancer Center and
Departments of Medicine and Pathology,
Columbia University, New York, NY 10032, USA

Page 2

1244Cancer Chemother Pharmacol (2011) 68:1243–1253
123
ribonucleotide reductase, which reduces the pool of dNTPs
available for DNA synthesis, and (2) competition with
dCTP for utilisation by DNA polymerase, leading to incor-
poration into DNA and chain termination [14, 26].
Gemcitabine is metabolically inactivated by cytidine
deaminase (CDA) into the metabolite 2?,2?-diXuorodeoxy-
uridine (dFdU) [1]. CDA enzyme is present in normal and
tumour cells [24, 30] as well as in blood [3]. The cytotoxic
eVects of gemcitabine are also modulated by the cell mem-
brane nucleoside transporters ENT1, ENT2, CNT1 and
CNT3, without which gemcitabine cannot enter cells and be
activated by phosphorylation [6, 20, 21]. Survival after
gemcitabine treatment has been linked to expression of
hENT1 and hCNT3 [21, 33].
Despite being licenced for use in pancreatic adenocarci-
noma, gemcitabine exhibits only a modest improvement in
patient survival [7], and it has been suggested that its activ-
ity may be limited by poor drug delivery, particularly in the
case of pancreatic adenocarcinoma (PDA) tumours, which
tend to be hypovascular with extensive desmoplastic
stroma [23, 32]. Previous studies have demonstrated that
the combination treatment of gemcitabine with an agent
that depletes the tumour stroma by inhibition of the Hedgehog
signalling pathway enhances drug delivery and eYcacy in a
mouse model of pancreatic cancer [23].
Fluorine-19 nuclear magnetic resonance spectroscopy
(19F NMR) is used commonly to detect Xuorine-containing
compounds and their metabolites that are non-naturally
occurring in biological tissues and Xuids. Among NMR
methods, 19F NMR has advantages due to the characteris-
tics of the natural Xuorine nucleus, which includes a nuclear
spin (I) of ½, 100% natural abundance, high sensitivity
(83% of 1H NMR), a large chemical shift range (500 ppm)
and short longitudinal relaxation times (T1) [22]. It oVers a
unique method of observing Xuorinated drugs, their phar-
macokinetics and metabolism either in vivo or in vitro
(including ex vivo tissue samples [4, 9, 39].
We have now improved a previously published 19F NMR
method for gemcitabine [9, 23] and used it to quantify gem-
citabine in mouse pancreatic tumour tissue. However, large
tissue samples were required for processing and the sample
run-time required to achieve acceptable signal to noise for
metabolite quantiWcation became limiting factors. To per-
form a detailed analysis of gemcitabine tumour pharmacoki-
netics, further improved analytical methods are required.
Several LC–MS/MS methods to measure gemcitabine
and dFdU in plasma have been described [5, 34, 41].
Honeywell et al. [13] described a method for analysis of
gemcitabine and dFdU in plasma and tissue. However,
intra-tumoural levels of the active metabolite dFdCTP
would clearly be the preferred correlant with pharmacody-
namics of the drug. Levels of dFdCTP measured in periph-
eral blood mononuclear cells have been used as a substitute
for tissue concentrations [36]. Recently, Jansen et al. [16]
described a method for quantifying gemcitabine, dFdU and
their phosphorylated metabolites including dFdCTP in
PBMCs. To the best of our knowledge, Wve studies have
reported tissue concentrations of dFdCTP but these either
measured only dFdCTP and not tissue dFdC and dFdU
[10, 26, 29], or used radioisotope-labelled dFdC [37], or
used enzymatic dephosphorylation of dFdCTP and then
measured the resulting dFdC [31]. An analytical method
capable of simultaneous quantiWcation of dFdC, dFdU and
dFdCTP in tumour tissue would be very useful.
We now describe a novel LC–MS/MS protocol for
simultaneous quantiWcation of gemcitabine, the active
metabolite dFdCTP and the inactive metabolite dFdU in
tumour tissue from a mouse model of pancreatic cancer.
This LC–MS/MS protocol was also used to quantify gem-
citabine and dFdU in plasma. We have assessed the perfor-
mance of the method and believe that it is faster and at least
as sensitive as previously published methods [10, 37].
This study provides insights into the intra-tumoural
distribution of the three analytes. We compare the 19F NMR
and LC–MS/MS protocols and discuss the relative
merits of each method for detection of gemcitabine and
its metabolites.
Methods
Chemicals
dFdCTP was generously provided by InWnity Pharmaceuti-
cal Co. (MA, USA) or was purchased from Sierra Biotech.
Gemcitabine hydrochloride was obtained from Tocris Bio-
science (UK), and dFdU was purchased from Carbosynth
(UK). GemzarTM (Eli Lilly), a »48% preparation of dFdC,
was obtained from Hannas (Delaware) or Addenbrooke’s
Hospital Pharmacy (Cambridge, UK) and was used at
10.2 mg/ml in sterile normal saline to dose the mice. Stable
isotope-labelled 13C9, 15N3-cytidine 5?-triphosphate, dFUR,
2-Xuoro-2?-deoxyadenosine (2F2dA) and tetrahydrofuran
(THF) were obtained from Sigma–Aldrich (UK). Tetrahy-
drouridine (THU) was purchased from Calbiochem (Merck
Chemicals, UK). Methanol, acetonitrile triXuoroacetic acid
(TFA), sodium hydroxide and ammonia solution were
obtained from Fisher ScientiWc (UK).
Tissue extraction for19F NMR
The snap-frozen tissues collected at necropsy were
weighed. Four volumes of ice-cold acetonitrile were added
to homogenise the sample using the Qiagen TissueLyser
with a 5-mm ball bearing for two rounds of 6 min at
25 kHz. An equal volume of ice-cold water (to acetonitrile)

Page 3

Cancer Chemother Pharmacol (2011) 68:1243–12531245
123
was added, and samples were incubated on ice for 10 min.
Samples were centrifuged at 14,000 rpm for 10 min at 4°C.
Supernatants were transferred to cold-resistant vials and
snap-frozen in liquid nitrogen. Holes were punctured in
the tops once samples were frozen. Vials were transferred
into a Heto PowerDry LL1500 Freeze-Dryer (Thermo) and
allowed to freeze-dry for at least 24 h.
19F NMR spectroscopy on tumour/tissue extracts
Freeze-dried tumour tissue extracts (from at least 100 mg
tissue wet weight) were re-suspended in 600 ?l of D2O,
spiked with 40 nmoles of internal standard 2F2dA, adjusted
to pH5 using 20 ?l of Universal pH Indicator solution
(Sigma) and diluted HCl and then transferred to a 5-mm
standard NMR tube (Wilmad) for 19F NMR analysis. All
19F NMR measurements were carried out on a Bruker 600
MHZ (14.1T) Avance NMR spectrometer, and a QNP
probe was used for acquisition of 19F NMR spectra. Acqui-
sition parameters included a 1-D pulse sequence of 19F
observation and inverse-gated 1H decoupling, spectral
sweep width of 177 ppm (100,000 Hz), 4,096 scans and
1.65 s of repetition time; total acquisition time was about
1 h 55 min. Chemical shift assignments of reference 2F2dA
at ¡52.06 ppm, gemcitabine (dFdC) at ¡116.33 ppm,
dFdU distributed over 4 peaks at ¡115.70, ¡116.12,
¡116.95, ¡117.08 ppm and dFdCTP at ¡116.92 ppm in
19F NMR spectra were determined by spiking the samples
with appropriate standards. A broad hump observed in the
baseline of 19F NMR spectra was removed by application
of linear prediction (LP) back-projection to the time
domain data by using 2,000 (number of LP) coeYcients and
128 back-prediction points prior to Fourier transformation
and phase correction. dFdC, dFdU and dFdCTP peak areas
were integrated using the Bruker Topspin software process-
ing package. The drug (dFdC) and its metabolite absolute
concentrations were estimated in terms of nmoles by using
the internal standard reference 2F2dA, and normalised to
the tissue wet weights, and 19F metabolite concentrations
are shown in ng/mg tissue.
Sample preparation for LC–MS/MS
In order to minimise enzyme-mediated degradation of the
analytes ex vivo, all manipulations were done on ice. A
minimum of 10 mg of tumour tissue was required, and
extraction of gemcitabine, dFdU and dFdCTP was achieved
by homogenising the tissue in ice-cold acetonitrile (50% v/v)
containing tetrahydrouridine (25 ?g/ml) in a Precellys 24
tissue homogeniser to give a Wnal concentration of 0.05 mg/?l
of tissue homogenate. An aliquot (50 ?l) of the homoge-
nate was added to a microfuge tube with 200 ?l of ice-
cold acetonitrile (50% v/v) containing 50 ng/ml each of
5?-deoxy-5-Xuorouridine (dFUR) and 13C9, 15N3-cytidine
triphosphate as internal standards (previously used as inter-
nal standards [36, 41]). Vortex mixing was followed by
centrifugation at 20,000£g for 25 min, and the resultant
supernatant was evaporated to dryness in a Speedvac. The
residue was reconstituted in water (100 ?l), and 20 ?l was
injected into the mass spectrometer. With respect to
plasma, 25 ?l was processed in the same way as the 50 ?l
of tumour homogenate, by adding 200 ?l of ice-cold aceto-
nitrile (85% v/v) containing the internal standards.
Preparation of calibration standards for LC–MS/MS
One milligram per millilitre stock solutions of gemcitabine
and dFdCTP were made in water, while dFdU was dis-
solved in methanol. The appropriate working solutions
were used to spike tumour homogenate prepared from
untreated mice as described above to give calibration stan-
dards with the following ranges: 20–5,000 ng/ml (0.4–
100 ng/mg) for dFdU, 10–2,500 ng/ml (0.2–50 ng/mg) for
gemcitabine and dFdCTP. Quality control samples were
prepared in the same way to give the appropriate concentra-
tions.
Chromatography
The analysis of adenosine phosphates using porous gra-
phitic carbon (PGC) by Wang et al. [38] served as an
important reference for the assay development work in this
study. Chromatography was performed on an Accela pump
and Accela autosampler (Thermo Fisher ScientiWc, USA).
The analytes were separated on a PGC Hypercarb column
(100 £ 2.1 ID, 5 ?m; Thermo Fisher ScientiWc) Wtted with
a guard column (Hypercarb 10 £ 2.1, 5 ?m; Thermo Fisher
ScientiWc) with (A) 10 mM ammonium acetate, pH 10 and
(B) acetonitrile as mobile phase. The autosampler and col-
umn temperatures were maintained at 4 and 30°C, respec-
tively. The gradient program at a Xow rate of 300 ?l/min
started with 95% A for 2 min, a decrease to 80% in 0.2 min
and held for 5.6 min, back to 95% over 0.2 min and held at
95% for 7 min to give a total run-time of 15 min. In order to
minimise carry-over between injections, the needle and
injection path were Xushed using the external wash proce-
dure (Thermo ScientiWc, surveyor autosampler plus hard-
ware manual: Post injection events) Wrst with 6 ml of water,
150 ?l of 100% acetonitrile followed by 3 £ 150 ?l of 50%
acetonitrile and Wnally with 400 ?l water.
Column regeneration
Regeneration of the column was sometimes necessary to
avoid loss of retention capacity, as has been reported by
other groups [12, 15, 28]. The manufacturer’s instructions

Page 4

1246Cancer Chemother Pharmacol (2011) 68:1243–1253
123
were used to restore the column to its original state. BrieXy,
the column was inverted and Xushed at 0.2 ml/min with
25 ml of THF/water (1:1) containing 0.1% TFA, followed
by 25 ml of THF/water containing 0.1% sodium hydroxide
and a further Xush with 25 ml of THF/water. The column
was then re-equilibrated with methanol/water (95:5) before
use.
Mass spectrometry
A TSQ Vantage triple stage quadrupole mass spectrometer
(Thermo ScientiWc, USA) Wtted with a heated electrospray
ionisation (HESI-II) probe operated in positive and nega-
tive mode at a spray voltage of 2.5 kV, capillary tempera-
ture of 150°C and vaporizer temperature of 250°C. Sheath
and auxiliary gas pressures were set at 50 and 20 units,
respectively. Compound optimisation was done manually
using Thermo TSQ Tune Master 2.1.0.1028 (Thermo
Fisher ScientiWc, USA) by infusion into the mass spectrom-
eter using a T-connector, and the scan parameters are
shown in Table 1. Quantitative data acquisition was done
using LC Quan2.5.6 (Thermo Fisher ScientiWc, USA).
Assessment of LC–MS/MS assay performance
Assessment of assay performance was performed using the
FDA Guidance for Industry 2001 document “Bioanalytical
Method Validation” as a guide, determining the linearity,
precision, accuracy, recovery, matrix eVects and stability of
the analytes. The linearity of the assay was tested using
eight non-zero standards with the back-calculated concen-
tration of each standard values not exceeding §15% of the
spiked value (§20% at the lower limit of quantiWcation).
The precision and accuracy were assessed by the replicate
analysis (n = 5) of QC samples at four diVerent concentra-
tions. A blank sample injected after the highest calibration
standard was used to determine carry-over. The short-term
stability of the analytes in homogenate and plasma contain-
ing THU on ice for 4 h and three freeze–thaw cycles at
¡80°C were assessed using three aliquots of QC samples at
two concentrations. The matrix eVects of tumour homoge-
nate were determined by comparing peak areas of analytes
spiked after extraction with neat standards in triplicate at
10 ng/mg (dFdU) and 5 ng/mg (dFdC and dFdCTP). The
recovery was assessed by comparing the peak response
ratios of analytes spiked before extraction with those spiked
after extraction at the same concentrations used for matrix
eVect determination.
Use of the assays to determine analyte levels
in mouse tumour tissue and plasma
The assays were used to study the levels of gemcitabine,
dFdU and dFdCTP in mouse tumour tissue and plasma.
Mouse studies were performed in accordance with the UK
Animals (ScientiWc Procedures) Act 1986 and the NCRI
2010 Guidelines for the welfare and use of animals in can-
cer research [40] with approval from the local Animal Eth-
ics Committee. Mice were of either KrasG12D; p53Xox/Xox;
Pdx1-Cre (KPFL/FLC) or KrasG12D; p53R172H; Pdx1-Cre
(KPR172H/+C) strains as previously described [2, 11]. Pan-
creatic ductal adenocarcinoma development was monitored
by ultrasound and/or abdominal palpation for KPR172H/
+C and KPFL/FLC mice. Three of the mice (MH1011,
MH1015 and MH1019) were found to have haemorrhagic
ascites as a result of their tumour.
Mice were dosed with 50 or 100 mg/kg gemcitabine by
IP injection in saline approximately 1 h before collecting
blood and tissues. These doses of gemcitabine are the
same as was used previously for eYcacy studies in the
KPR172H/+C mouse model [23] and are estimated to be
equivalent to 150 and 300 mg/m2 (using a mass-to-body
surface area factor of 3 for mouse [27]). Tumours were
rapidly excised at necropsy, soft surrounding tissue was
trimmed oV, and the tissue was snap-frozen in liquid
nitrogen then stored at ¡80°C until required. Blood was
taken into tubes containing EDTA (1.75 mg/ml) and THU
(25 ?g/ml) on ice and centrifuged, and plasma was stored
at ¡80°C.
The amount of compound was measured by LC–MS/MS
in ng/ml of tissue homogenate or plasma. The tumour
homogenate contained 50 mg of tissue per ml, and ng/ml
values were converted to ng/mg tissue (e.g. 50 ng/ml
measured concentration is equivalent to 1 ng/mg of tumour
tissue).
Results
19F NMR
Previous laboratory experience suggested that in order to
quantify minute amounts of Xuorinated species in a poorly
perfused tissue sample, the 19F NMR method required fur-
ther optimisation to increase the sensitivity. The chemical
Table 1 Mass spectrometry scan parameters
AnalyteParent
ion (m/z)
Product
ion (m/z)
Collision
energy
Polarity
dFdC
dFdU
dFdCTP
dFUR
13C15N CTP
264.03
263.00
504.00
245.00
496.00
112.04
202.13
326.07
202.11
119.01
18
14
23
14
23
Positive
Negative
Positive
Negative
Positive

Page 5

Cancer Chemother Pharmacol (2011) 68:1243–12531247
123
environment of the sample is responsible for the coupling
and de-coupling of the spins as well as the chemical shift of
peaks. To identify the typical spectra of gemcitabine
(dFdC), dFdU and dFdCTP, standards were spiked into an
untreated homogenised mouse tumour sample.
Under the conditions in our recently published paper
[23], dFdCTP produced one broad and shallow peak, mak-
ing it diYcult to quantify. The most optimal condition iden-
tiWed was an adjustment of the sample to pH 5, where the
standards appeared the sharpest in their resulting spectra.
Figure 1 shows a cumulative representation of peaks of the
three Xuorinated standards when run individually on frac-
tions of the same tumour tissue homogenate. Here, the Xuo-
rine signals for dFdC and dFdCTP each show as one
distinct peak, whereas the signal for dFdU appears spread
over 4 peaks. The relative ratios between the 4 peaks of
dFdU always remain the same, as the two Xuorine atoms
remain decoupled (i.e. peak 1:2 is 0.32:1; peak 2:3 is
1:0.67; peak 3:4 is 1:0.29). Using these ratios, the overlap
between the smallest peak of dFdU and dFdCTP can be
deduced from the integral of the combined peak and the
respective values derived.
These optimised conditions were then used to measure
gemcitabine and its metabolite concentrations in tumours
from mice dosed with gemcitabine. Representative calcu-
lated values are shown in Table 2. There was considerable
inter-mouse variation both in the absolute analyte concen-
trations and in the relative ratios of the analytes (e.g. the
dFdCTP/dFdC ratio ranged from 0.33 for TB7798 to 18 for
MH733).
Despite eVorts to optimise this method for analysing
tumour samples, some restricting factors still remained. At
least 100 mg of tissue was necessary for processing, which
limits the amount of tissue available from any one tumour
for other assays (such as histology and pharmacodynamic
assays). Secondly, the time required for each sample run
to gain an acceptable signal to noise ratio (1 h 55 min)
hindered the possibility of high-throughput analysis.
LC–MS/MS
In light of the limitations of the NMR method (100 mg of
tissue required and 1 h 55 min acquisition time), the goal
was set of developing a sensitive LC–MS/MS assay, suit-
able for high throughput, to measure dFdC, dFdU and
dFdCTP, using a minimum of 10 mg of mouse tumour tis-
sue. Typical chromatograms from this new method, at the
limit of quantiWcation, are shown in Fig. 2. There were no
signiWcant interfering peaks for all the analytes in blank
tumour homogenate and plasma. With respect to plasma,
only 25 ?l of plasma was required for quantiWcation of
gemcitabine and dFdU. dFdCTP was not assayed for in
plasma because negligible amounts were expected in the
plasma and initial studies showed poor recovery of spiked
dFdCTP after acetonitrile precipitation from plasma. The
recoveries of all analytes from tumour tissue were greater
than 90%. A study of matrix eVects in tumour homogenate
showed signal reduction for dFdU (21%) and enhancement
for gemcitabine (7.3%) and dFdCTP (1.4%).The calibration
ranges for gemcitabine, dFdCTP and dFdU were from 0.2
to 50, 0.2–50 and 0.4–100 ng/mg tissue, respectively. The
calibration ranges in plasma were from 10 to 2,500 and
Fig. 1 Overlaid spectra of 19F NMR peaks for 30 nmoles of dFdC
(green), dFdU (blue), dFdCTP (red), spiked and quantiWed separately
into an untreated KPFL/FLC tumour sample
[ppm] -115.5 -116.0 -116.5 -117.0 -117.5
[rel]
-5
0
5
10
15
dFdC
dFdU
dFdCTP
Table 2 Concentrations of gemcitabine (dFdC) and metabolites
(dFdU and dFdCTP) in tumour samples from 14 KPFL/FLC mice, col-
lected 1 h after dosing with 100 mg/kg gemcitabine, measured by 19F
NMR
Mouse ID SexTumour concentration
(ng/mg tissue)
dFdCdFdUdFdCTP
TB7798
TB7799
TB7801
TB7802
TB7803
TB9343
TB9393
TB9399
MH729
MH731
MH732
MH733
MH735
MH737
M
F
F
F
F
M
F
F
M
F
F
F
M
M
Mean (st dev)
1.86
1.62
2.53
3.19
2.3
6.65
5.48
2.66
9.98
2.47
2.51
1.56
2.47
9.95
3.95 (2.92)
4.61
3.2
9.95
12.0
11.7
10.1
6.81
19.8
34
15.3
24.1
17.8
13.9
8.98
13.7 (8.20)
0.61
0.87
17.7
14.6
0.86
24.4
25.2
16.0
28.8
26.1
30.4
28.6
20.7
7.15
17.3 (11.0)

Page 6

1248Cancer Chemother Pharmacol (2011) 68:1243–1253
123
20–5,000 ng/ml for gemcitabine and dFdU, respectively.
Linear regression and 1/x weighting were used for gemcita-
bine and dFdCTP while a quadratic Wt was used for dFdU.
The correlation coeYcients (r2) of the curves were greater
than 0.99. The intra-assay precision and accuracy values for
tissue samples were well within recommended levels with
accuracy values of 91.7–103.9% and precision (% CV) less
than 6% for all analytes in tumour tissue homogenate
(Table 3). For plasma, the intra-assay precision and accu-
racy for gemcitabine and dFdU were also within recom-
mended levels with accuracy values of 80.4–112.7%
(80.4% at the LLOQ where §20% is allowed, all other con-
centrations were within §15%), and the % CV was less
than 6% (Table 3). The freeze–thaw stability was assessed
using Wve replicates at each of the concentrations shown in
Table 4. After each freeze cycle, the samples were allowed
to thaw on ice. All analytes were stable in tumour tissue
homogenate and plasma containing THU after three freeze–
thaw cycles at ¡80°C. Long-term stability during storage at
¡80°C is ongoing. In order to assess short-term stability,
three replicates at each of the concentrations were kept on
ice for 4 h. All analytes were stable on ice for the 4 h. The
quality control samples were always within §15% of nomi-
nal concentrations for gemcitabine. The carry-over was
assessed by injecting a blank sample after the highest stan-
dard with values in both tumour tissue homogenate and
plasma less than 15%.
The LC–MS/MS assay was then used to measure the
analytes in tumours and plasma from mice dosed with gem-
citabine at 50 or 100 mg/kg (Table 5). Samples were col-
lected 75 min after dosing mice MH1015, MH1019 and
MH1014, 70 min after dosing mouse MH1011, 65 min
after dosing mouse MH963 and 60 min after dosing mouse
MH959. QuantiWcation of gemcitabine and dFdU in the
plasma samples required dilution by a factor of hundred
and a dilution QC of the same factor was run together with
the samples.
As with 19F-NMR, there was inter-mouse variation in
absolute concentrations and ratios of the analytes. The
amounts of dFdU observed in tumour tissue in these mice
were similar to the plasma concentration (assuming tissue
density is 1 g/ml, 1 ng/mg tissue = 1 ?g/ml). The plasma
concentrations of gemcitabine and dFdU are comparable to
those previously reported, using an HPLC method [23].
In order to investigate the distribution of analytes within
a tumour, multiple pieces from diVerent regions of each of
Fig. 2 Typical LC–MS/MS
chromatogram for tumour tissue
homogenate spiked with concen-
trations of analytes at the lower
limit of quantiWcation,
a gemcitabine (0.2 ng/mg),
b dFdU (0.4 ng/mg) and
d dFdCTP (0.2 ng/mg). Internal
standards: c dFUR and e13C,
N15CTP

Page 7

Cancer Chemother Pharmacol (2011) 68:1243–12531249
123
Table 3 LC–MS/MS assay precision and accuracy in tumour tissue and plasma calculated from replicates at each concentration
CompoundNominal
conc. (ng/mg)
Mean measured
conc. (ng/mg)
Precision (%)Accuracy (%)
Tumour homogenate
dFdC (gemcitabine)0.2
0.6
5.0
35.0
0.40
1.20
10.0
70.0
0.2
0.6
5.0
35.0
0.20
0.57
4.82
36.36
0.40
1.22
9.74
70.14
0.18
0.56
4.78
33.78
2.1
5.9
2.0
3.5
5.1
5.0
3.8
3.7
5.1
5.1
4.8
1.4
0.3
¡5.0
¡3.4
3.9
0.8
1.3
¡2.6
0.2
¡8.3
¡7.5
¡4.3
¡3.5
Accuracy (%)
dFdU
dFdCTP
CompoundNominal
conc. (ng/ml)
Mean (ng/ml)Precision (%)
Plasma
dFdC10
30
8.0
27.1
263
3.9
3.0
8.2
4.1
8.6
4.2
7.5
4.9
¡19.6
¡9.7
5.3
¡2.1
¡11.4
4.8
12.7
2.4
250
1,7501,713
dFdU20
60
500
17.7
62.9
563
3,5843,500
Table 4 Stability of analytes in tumour homogenate and plasma measured by LC–MS/MS
CompoundConc. (ng/mg)
¡80°C 3 freeze–thaw cycles
% stability
4 h on ice
% CV% stability% CV
Tumour homogenate
dFdC5110.2
113.1
102.4
98.8
94.1
97.4
2.7
3.3
2.7
2.6
1.1
1.3
119.0
125.3
107.3
105.3
92.0
92.9
2.5
2.2
1.3
0.5
0.9
0.8
50
10
100
dFdU
dFdCTP5
50
¡80°C 3 freeze–thaw cycles
% stability
4 h on ice
CompoundConc. (ng/ml)% CV% stability % CV
Plasma
dFdC 250
2,500
500
5,000
111.2
121.3
111.1
104.3
2.6
4.5
2.2
3.2
99.7
118.2
107.8
106.2
2.4
12.7
2.2
6.4
dFdU

Page 8

1250Cancer Chemother Pharmacol (2011) 68:1243–1253
123
one KPR172H/+C and one KPFL/FLC tumour were assayed for
the 3 analytes (Table 6). These samples were taken 60 min
after a single dose of 100 mg/kg gemcitabine. There was
variation in the total amount of analytes and the ratios of
analytes in diVerent regions of the same tumour. The
KPR172H/+C tumour appeared to have generally higher lev-
els of dFdC and dFdU than KPFL/FLC tumour, but lower
dFdCTP concentrations.
Discussion
Comparing 19F NMR with LC–MS/MS
The advantage of 19F NMR is the detection of all Xuori-
nated species and metabolites of a compound in any tissue
type simultaneously. Once optimised, the 19F NMR proto-
col can be used directly whereas the LC–MS/MS method
requires optimising for each type of tissue sample and com-
pound metabolite analysed. However, the sensitivity of 19F
NMR is generally quite low. In principle, sensitivity can be
improved by increasing the sample size (processing enough
tissue with very low concentrations of drug to have quanti-
Wable signal), the magnetic Weld (deWned by the instrument
used) and the number of transients accumulated (run-time).
Even with a prolonged acquisition time (1 h 55 min in this
study), 19F NMR required more tissue (100 mg) than the
LC–MS/MS protocol which can quantify with greater reso-
lution down to 0.2 ng/mg (0.4 pmol/mg) of dFdCTP from
as little as 10 mg of tissue in only 15 min per sample.
If the signal from a Xuorinated compound is high enough,
it is possible to do non-invasive in vivo 19F MRS spectros-
copy, as demonstrated with 5-FU and capecitabine in cancer
patients [17, 18, 35]. Unfortunately, the sensitivity of MRS in
vivo is usually lower than for ex vivo NMR due to the lower
power of whole body magnets and the transient accumulation
time that is limited by patient tolerance. Another major prob-
lem of MRS in vivo is poor peak resolution, making it
extremely diYcult to resolve the gemcitabine metabolites
detected in the present ex vivo studies. Three in vivo 19F
NMR studies of gemcitabine in mouse xenograft tumours
have been reported; two studies used very high-dose gemcit-
abine (500–800 mg/kg), and none were able to distinguish
between the parent compound and its metabolites [4, 8, 19],
which is key for gemcitabine because of signiWcant amounts
of the inactive metabolite dFdU. Another attractive ex vivo
method to detect Xuorinated drugs and their metabolites in
intact tissue (samples are not destroyed during analysis) is
high-resolution magic angle spinning (HRMAS) 19F NMR,
but in general this has lower peak resolution than the extrac-
tion method employed in the present study, since spinning
the samples may not eliminate all the line-broadening dipolar
coupling mechanisms. To our knowledge, HRMAS NMR
has not been reported for gemcitabine.
A shortcoming of the LC–MS/MS method is that it
requires excised tissue samples for analysis. However, the
LC–MS/MS protocol is higher throughput than the stated
NMR method, because analysis is faster and it can be auto-
mated for sample preparation and autoloading of multiple
samples sequentially. The major advantage of the LC–MS/
MS method over NMR is the requirement for a small sam-
ple, of as little as 10 mg (discussed further below).
Chromatography
We have overcome the well-documented analytical chal-
lenges (such as simultaneous chromatographic resolution)
involving nucleotides and nucleosides to develop an attrac-
tive LC–MS/MS method. The generally poor separation of
nucleotides and nucleosides on many stationary phases
including C-18 has resulted in the use of mass spectrome-
try-incompatible methods such as ion-exchange chroma-
tography. Porous graphitic carbon (PGC) has, however,
been shown by others to obviate the need for ion-exchange
with excellent peak shapes. The analysis of adenosine
phosphates using PGC by Wang et al. [38] was an impor-
Table 5 Concentrations of dFdC, dFdU and dFdCTP in one piece of tumour, and dFdC and dFdU in plasma, from 6 KPFL/FLC mice 60–75 min
after gemcitabine dose, measured by LC–MS/MS
Mouse ID Sex Dose (mg/kg)Tumour concentration (ng/mg tissue) Plasma (?g/ml)
dFdC dFdUdFdCTP dFdCdFdU
MH959
MH963
MH1011
M
F
M
100
100
100
Mean (st dev)
50
50
50
Mean (st dev)
6.4
5.1
19.2
10.2 (7.8)
3.4
3.2
5.4
4.0 (1.2)
10.1
24.7
19.9
18.2 (7.4)
11
9.4
13.3
11.2 (2.0)
13.4
8.3
3.1
8.3 (5.2)
4.2
1
2.1
2.4 (1.6)
9.7
5.1
89.7
34.8 (47.6)
1.5
4.2
15.5
7.1 (7.4)
8.6
26.4
22.4
19.1 (9.3)
8.6
12.3
17.9
12.9 (4.7)
MH1014
MH1015
MH1019
F
F
M

Page 9

Cancer Chemother Pharmacol (2011) 68:1243–12531251
123
tant reference for the method development in this study.
We, however, observed shifts in retention times with con-
tinued column usage, as reported by others [12, 15, 16]. In
our view, reproducibility of retention on PGC represents
the most important challenge. Regeneration of the column
was suYcient to restore the column to its original condi-
tions and reproducibility of retentions times within each run
appeared to be dependent on the cleanness of the samples.
The samples prepared using procedures described in this
study were suYciently clean to give good quantitative data
for the analytes investigated.
Analyte concentrations in tumour tissue and plasma
measured by LC–MS/MS
The method was used to measure concentrations of the ana-
lytes in pancreatic tumour tissue and plasma from mice
treated with gemcitabine. There were inter-mouse diVer-
ences in absolute concentrations and ratios of analytes,
which were not obviously due to the range in sampling
times from 60 to 75 min after dosing. There was also no
obvious diVerence between those mice with haemorrhagic
ascites and those without. Similar variation was observed in
the tumour samples measured by 19F NMR. Concentrations
of the 3 analytes were generally higher in those dosed with
100 mg/kg gemcitabine than with 50 mg/kg. It is clear that
dFdCTP concentrations in tumour cannot be predicted from
either tumour or plasma gemcitabine or dFdU concentra-
tions, underlining the value of this analytical assay for
investigation of tumour PK. The concentrations of dFdCTP
detected in the tumours (ranging from <0.2 to 1.54 ng/mg
in KPR172H/+C tumours, and 0.5–13.4 ng/mg in KPFL/
FLC tumours) are similar to those reported in a human glio-
blastoma biopsy 1 h after a gemcitabine dose (3,000 pmol/
g, which is equivalent to 1.5 ng/mg)[31], and to head and
neck tumour concentrations 2 h after a 300 mg/m2 dose of
gemcitabine (2.13 pmol dFdCTP/mg, equivalent to 1.07 ng/
mg) [10, 31].
Analysis of multiple areas of a tumour is made feasible
by the small sample size requirements of this analytical
assay. We analysed multiple samples from diVerent regions
of one KPR172H/+C and one KPFL/FLC tumour. These
showed little intra-tumour variation in gemcitabine and
dFdU, but signiWcant variation in dFdCTP, which may reX-
ect diVerent cellularity in diVerent parts of the tumour,
because dFdCTP is formed intra-cellularly, whereas gem-
citabine and dFdU are expected to be present both intra-cel-
lularly and extra-cellularly in the interstitial Xuid. The
KPR172H/+C tumour showed higher concentrations of gem-
citabine and dFdU than the KPFL/FLC tumours, but the
dFdCTP concentrations were lower in the KPR172H/+C than
the KPFL/FLC. The vascular function and density of the
KPFL/FLC tumours in comparison to the hypovascular
KPR172H/+C tumours is currently under investigation (M.
Jacobetz, personal communication).
The requirement for only 10 mg of tumour tissue raises
the possibility of measuring tumour PK in biopsy samples
from human tumours, such as those obtained from a Tru-
cut biopsy (as used for liver tumours, for example) or an
endoscopic microcore biopsy (being developed for pancre-
atic adenocarcinoma [R. Brais et al. manuscript in prepara-
tion]). It may be possible to use the LC–MS/MS method for
multiple biopsies or repeat biopsies. Also, using only a
small piece for PK analysis allows tissue to be spared for
additional pharmacodynamic assays.
Conclusions
We have improved a 19F NMR protocol and described a
sensitive HPLC–MS/MS method to quantify gemcitabine,
Table 6 Concentrations of dFdC, dFdU and dFdCTP in ten pieces of
a tumour from one KPR172H/+C mouse and ten pieces of one tumour
from a KPFL/FLC mouse, 1 h post-gemcitabine (100 mg/kg)
The lower limit of quantiWcation (LLOQ) is 0.2 ng/mg for dFdC and
dFdCTP and 0.4 ng/mg for dFdU
aBLQ: Below the limit of quantiWcation
bExcludes the 3 which were BLQ therefore true mean value is lower
PieceWeight (mg)Tumour concentration (ng/mg tissue)
dFdC dFdUdFdCTP
KPR172H/+C (MH2106) tumour pieces
132.5
229.7
3 25.8
4 38.0
5 36.7
632.0
732.7
844.0
933.7
10 22.6
Mean (st dev)
KPFL/FLC (MH959) tumour pieces
1 33.9
2 17.2
3 23.7
420.5
534.8
619.2
722.1
8 19.4
914.5
10 19.7
Mean (st dev)
21.2
23.5
22.6
15.4
29.5
19.1
20.0
25.4
29.8
22.1
22.9 (4.5)
46.0
47.4
47.8
46.6
56.1
42.2
48.8
50.6
49.9
44.3
48.0 (3.8)
0.42
BLQa
0.54
0.95
BLQ
0.50
0.46
BLQ
0.73
1.54
0.73 (0.40)b
6.9
8.4
5.4
7.2
5.4
5.1
5.5
7.1
5.1
4.5
6.2 (1.1)
20.7
24.1
8.8
10.2
12.6
8.7
11.8
13.5
9.9
9.7
13.0 (5.3)
2.7
2.6
1.2
4.2
1.0
1.8
0.5
2.7
2.1
1.6
2.1 (1.1)

Page 10

1252Cancer Chemother Pharmacol (2011) 68:1243–1253
123
dFdU and dFdCTP in mouse pancreatic tissue. As little as
10 mg of tissue is required for LC–MS/MS, which makes it
ideal when tissue specimens are also required for other
assessments. The method can be applied to preclinical and
clinical studies. The protocol will now be used to investi-
gate tumour pharmacokinetics of gemcitabine in more
detail and to correlate results with tumour histology and
pharmacodynamic endpoints. We are also developing com-
bination therapy strategies, using agents targeting the
tumour stroma, which may increase the delivery of gemcit-
abine to tumour tissue. This has already been demonstrated
in studies with the Hedgehog pathway inhibitor IPI-926
[23], and this assay will be instrumental in the development
of this combination approach.
Acknowledgments
advice on LC–MS/MS protocol, and Paul Mackin and staV of the CRI
BRU, especially Steve Kupzac, for animal colony maintenance and
assistance with animal studies. We thank InWnity Pharmaceuticals for
providing dFdCTP. Funding for this work was provided by Cancer
Research UK. The authors would like to particularly acknowledge the
advice, support and encouragement given by Dr Merrill Egorin, who
passed away earlier this year. Without his outstanding mentorship of a
young investigator, this Pharmacology and Drug Development
Group would not exist. He is sadly missed.
We thank Michael Williams (CRUK CRI) for
Open Access
tive Commons Attribution Noncommercial License which permits any
noncommercial use, distribution, and reproduction in any medium,
provided the original author(s) and source are credited.
This article is distributed under the terms of the Crea-
References
1. Abbruzzese JL, Grunewald R, Weeks EA, Gravel D, Adams T,
Nowak B, Mineishi S, TarassoV P, Satterlee W, Raber MN et al
(1991) A phase I clinical, plasma, and cellular pharmacology study
of gemcitabine. J Clin Oncol 9:491–498
2. Bardeesy N, Aguirre AJ, Chu GC, Cheng KH, Lopez LV, Hezel
AF, Feng B, Brennan C, Weissleder R, Mahmood U, Hanahan D,
Redston MS, Chin L, Depinho RA (2006) Both p16(Ink4a) and the
p19(Arf)-p53 pathway constrain progression of pancreatic adeno-
carcinoma in the mouse. Proc Natl Acad Sci USA 103:5947–5952
3. Besnard T, Renee N, Etienne-Grimaldi MC, Francois E, Milano G
(2008) Optimized blood sampling with cytidine deaminase inhibitor
for improved analysis of capecitabine metabolites. J Chromatogr
B Analyt Technol Biomed Life Sci 870:117–120
4. Blackstock AW, Lightfoot H, Case LD, Tepper JE, Mukherji SK,
Mitchell BS, Swarts SG, Hess SM (2001) Tumor uptake and elim-
ination of 2?,2?-diXuoro-2?-deoxycytidine (gemcitabine) after
deoxycytidine kinase gene transfer: correlation with in vivo tumor
response. Clin Cancer Res 7:3263–3268
5. Bowen C, Wang S, Licea-Perez H (2009) Development of a sensi-
tive and selective LC-MS/MS method for simultaneous determina-
tion of gemcitabine and 2,2-diXuoro-2-deoxyuridine in human
plasma. J Chromatogr B Analyt Technol Biomed Life Sci
877:2123–2129
6. Burke T, Lee S, Ferguson PJ, Hammond JR (1998) Interaction of
2?,2?-diXuorodeoxycytidine (gemcitabine) and formycin B with
the Na+-dependent and -independent nucleoside transporters of
Ehrlich ascites tumor cells. J Pharmacol Exp Ther 286:1333–1340
7. Burris H, Storniolo AM (1997) Assessing clinical beneWt in the
treatment of pancreas cancer: gemcitabine compared to 5-Xuoro-
uracil. Eur J Cancer 33(Suppl 1):S18–S22
8. Cron GO, Beghein N, Ansiaux R, Martinive P, Feron O, Gallez B
(2008) 19F NMR in vivo spectroscopy reXects the eVectiveness of
perfusion-enhancing vascular modiWers for improving gemcita-
bine chemotherapy. Magn Reson Med 59:19–27
9. Edzes HT, Peters GJ, Noordhuis P, Vermorken JB (1993) Deter-
mination of the antimetabolite Gemcitabine (2?,2?-diXuoro-2?-
deoxycytidine) and of 2?,2?-diXuoro-2?-deoxyuridine by 19F
nuclear magnetic resonance spectroscopy. Anal Biochem 214:25–30
10. Eisbruch A, Shewach DS, Bradford CR, Littles JF, Teknos TN,
Chepeha DB, Marentette LJ, Terrell JE, Hogikyan ND, Dawson
LA, Urba S, Wolf GT, Lawrence TS (2001) Radiation concurrent
with gemcitabine for locally advanced head and neck cancer: a
phase I trial and intracellular drug incorporation study. J Clin
Oncol 19:792–799
11. Hingorani SR, Wang L, Multani AS, Combs C, Deramaudt TB,
Hruban RH, Rustgi AK, Chang S, Tuveson DA (2005)
Trp53R172H and KrasG12D cooperate to promote chromosomal
instability and widely metastatic pancreatic ductal adenocarci-
noma in mice. Cancer Cell 7:469–483
12. Holmgren E, Carlsson H, Goede P, Crescenzi C (2005) Determi-
nation and characterization of organic explosives using porous
graphitic carbon and liquid chromatography-atmospheric pressure
chemical ionization mass spectrometry. J Chromatogr A
1099:127–135
13. Honeywell R, Laan AC, van Groeningen CJ, Strocchi E, Ruiter R,
Giaccone G, Peters GJ (2007) The determination of gemcitabine
and 2?-deoxycytidine in human plasma and tissue by APCI tandem
mass spectrometry. J Chromatogr B Analyt Technol Biomed Life
Sci 847:142–152
14. Huang P, Chubb S, Hertel LW, Grindey GB, Plunkett W (1991)
Action of 2?,2?-diXuorodeoxycytidine on DNA synthesis. Cancer
Res 51:6110–6117
15. Jansen RS, Rosing H, Schellens JH, Beijnen JH (2009) Retention
studies of 2?-2?-diXuorodeoxycytidine and 2?-2?-diXuorodeoxy-
uridine nucleosides and nucleotides on porous graphitic carbon:
development of a liquid chromatography-tandem mass spectrom-
etry method. J Chromatogr A 1216:3168–3174
16. Jansen RS, Rosing H, Schellens JH, Beijnen JH (2009) Simulta-
neous quantiWcation of 2?,2?-diXuorodeoxycytidine and 2?,2?-di-
Xuorodeoxyuridine nucleosides and nucleotides in white blood
cells using porous graphitic carbon chromatography coupled with
tandem mass spectrometry. Rapid Commun Mass Spectrom
23:3040–3050
17. Kamm YJ, Heerschap A, van den Bergh EJ, Wagener DJ (2004)
19F-magnetic resonance spectroscopy in patients with liver metas-
tases of colorectal cancer treated with 5-Xuorouracil. Anticancer
Drugs 15:229–233
18. Klomp D, van Laarhoven H, Scheenen T, Kamm Y, Heerschap A
(2007) Quantitative 19F MR spectroscopy at 3 T to detect hetero-
geneous capecitabine metabolism in human liver. NMR Biomed
20:485–492
19. Kristjansen PE, QuistorV B, Spang-Thomsen M, Hansen HH (1993)
Intratumoral pharmacokinetic analysis by 19F-magnetic resonance
spectroscopy and cytostatic in vivo activity of gemcitabine (dFdC)
in two small cell lung cancer xenografts. Ann Oncol 4:157–160
20. Mackey JR, Yao SY, Smith KM, Karpinski E, Baldwin SA, Cass
CE, Young JD (1999) Gemcitabine transport in xenopus oocytes
expressing recombinant plasma membrane mammalian nucleoside
transporters. J Natl Cancer Inst 91:1876–1881
21. Marechal R, Mackey JR, Lai R, Demetter P, Peeters M, Polus M,
Cass CE, Young J, Salmon I, Deviere J, Van Laethem JL (2009)
Human equilibrative nucleoside transporter 1 and human concen-
trative nucleoside transporter 3 predict survival after adjuvant

Page 11

Cancer Chemother Pharmacol (2011) 68:1243–12531253
123
gemcitabine therapy in resected pancreatic adenocarcinoma. Clin
Cancer Res 15:2913–2919
22. Martino R, Malet-Martino M, Gilard V (2000) Fluorine nuclear
magnetic resonance, a privileged tool for metabolic studies of
Xuoropyrimidine drugs. Curr Drug Metab 1:271–303
23. Olive KP, Jacobetz MA, Davidson CJ, Gopinathan A, McIntyre D,
Honess D, Madhu B, Goldgraben MA, Caldwell ME, Allard D,
Frese KK, Denicola G, Feig C, Combs C, Winter SP, Ireland-
Zecchini H, Reichelt S, Howat WJ, Chang A, Dhara M, Wang L,
Ruckert F, Grutzmann R, Pilarsky C, Izeradjene K, Hingorani SR,
Huang P, Davies SE, Plunkett W, Egorin M, Hruban RH, White-
bread N, McGovern K, Adams J, Iacobuzio-Donahue C, GriYths
J, Tuveson DA (2009) Inhibition of Hedgehog signaling enhances
delivery of chemotherapy in a mouse model of pancreatic cancer.
Science 324:1457–1461
24. Onodera H, Kuruma I, Ishitsuka H, Horii I (2000) Pharmacokintic
study of capecitabine in monkeys and mice: species diVerences in
distribution of the enzymes responsible for its activation to 5-FU.
Xenobio Metabol and Dispos 15:439–451
25. Peters GJ, Clavel M, Noordhuis P, Geyssen GJ, Laan AC,
Guastalla J, Edzes HT, Vermorken JB (2007) Clinical phase I and
pharmacology study of gemcitabine (2?,2?-diXuorodeoxycytidine)
administered in a two-weekly schedule. J Chemother 19:212–221
26. Plunkett W, Huang P, Gandhi V (1995) Preclinical characteristics
of gemcitabine. Anticancer Drugs 6(Suppl 6):7–13
27. Reagan-Shaw S, Nihal M, Ahmad N (2008) Dose translation from
animal to human studies revisited. FASEB J 22:659–661
28. Reepmeyer JC, Brower JF, Ye H (2005) Separation and detection
of the isomeric equine conjugated estrogens, equilin sulfate and
delta 8,9-dehydroestrone sulfate, by liquid chromatography–
electrospray-mass spectrometry using carbon-coated zirconia and
porous graphitic carbon stationary phases. J Chromatogr A
1083:42–51
29. Ruiz van Haperen VW, Veerman G, Boven E, Noordhuis P,
Vermorken JB, Peters GJ (1994) Schedule dependence of sensitiv-
ity to 2?,2?-diXuorodeoxycytidine (Gemcitabine) in relation to
accumulation and retention of its triphosphate in solid tumour cell
lines and solid tumours. Biochem Pharmacol 48:1327–1339
30. Ruiz van Haperen VW, Veerman G, Braakhuis BJ, Vermorken JB,
Boven E, Leyva A, Peters GJ (1993) Deoxycytidine kinase and
deoxycytidine deaminase activities in human tumour xenografts.
Eur J Cancer 29A:2132–2137
31. Sigmond J, Honeywell RJ, Postma TJ, Dirven CM, de Lange SM,
van der Born K, Laan AC, Baayen JC, Van Groeningen CJ, Berg-
man AM, Giaccone G, Peters GJ (2009) Gemcitabine uptake in
glioblastoma multiforme: potential as a radiosensitizer. Ann Oncol
20:182–187
32. Sofuni A, Iijima H, Moriyasu F, Nakayama D, Shimizu M,
Nakamura K, Itokawa F, Itoi T (2005) DiVerential diagnosis
of pancreatic tumors using ultrasound contrast imaging.
J Gastroenterol 40:518–525
33. Tanaka M, Javle M, Dong X, Eng C, Abbruzzese JL, Li D (2010)
Gemcitabine metabolic and transporter gene polymorphisms are
associated with drug toxicity and eYcacy in patients with locally
advanced pancreatic cancer. Cancer 116:5325–5335
34. Vainchtein LD, Rosing H, Thijssen B, Schellens JH, Beijnen JH
(2007) Validated assay for the simultaneous determination of the
anti-cancer agent gemcitabine and its metabolite 2?,2?-diXuorode-
oxyuridine in human plasma by high-performance liquid chroma-
tography with tandem mass spectrometry. Rapid Commun Mass
Spectrom 21:2312–2322
35. van Laarhoven HW, Klomp DW, Kamm YJ, Punt CJ, Heerschap
A (2003) In vivo monitoring of capecitabine metabolism in human
liver by 19Xuorine magnetic resonance spectroscopy at 1.5 and 3
Tesla Weld strength. Cancer Res 63:7609–7612
36. Veltkamp SA, Hillebrand MJ, Rosing H, Jansen RS,
Wickremsinhe ER, Perkins EJ, Schellens JH, Beijnen JH (2006)
Quantitative analysis of gemcitabine triphosphate in human
peripheral blood mononuclear cells using weak anion-exchange
liquid chromatography coupled with tandem mass spectrometry.
J Mass Spectrom 41:1633–1642
37. Veltkamp SA, Pluim D, van Tellingen O, Beijnen JH, Schellens
JH (2008) Extensive metabolism and hepatic accumulation of
gemcitabine after multiple oral and intravenous administration in
mice. Drug Metab Dispos 36:1606–1615
38. Wang J, Lin T, Lai J, Cai Z, Yang MS (2009) Analysis of adeno-
sine phosphates in HepG-2 cell by a HPLC-ESI-MS system with
porous graphitic carbon as stationary phase. J Chromatogr B Analyt
Technol Biomed Life Sci 877:2019–2024
39. Wolf W, Presant CA, Servis KL, el-Tahtawy A, Albright MJ,
Barker PB, Ring R 3rd, Atkinson D, Ong R, King M et al (1990)
Tumor trapping of 5-Xuorouracil: in vivo 19F NMR spectroscopic
pharmacokinetics in tumor-bearing humans and rabbits. Proc Natl
Acad Sci USA 87:492–496
40. Workman P, Aboagye EO, Balkwill F, Balmain A, Bruder G,
Chaplin DJ, Double JA, Everitt J, Farningham DA, Glennie MJ,
Kelland LR, Robinson V, Stratford IJ, Tozer GM, Watson S,
Wedge SR, Eccles SA, Committee of the National Cancer
Research I (2010) Guidelines for the welfare and use of animals in
cancer research. Br J Cancer 102:1555–1577
41. Xu Y, Keith B, Grem JL (2004) Measurement of the anticancer
agent gemcitabine and its deaminated metabolite at low concentra-
tions in human plasma by liquid chromatography-mass spectrome-
try. J Chromatogr B Analyt Technol Biomed Life Sci 802:263–270