We have previously shown that RL-G cells were 2,092-fold more resistant than RL-7 cells to gemcitabine (RL-G IC50: 25.1 ± 5.0 μM; RL-7 IC50: 0.015 ± 0.0006 μM) 23. RL-G cells were also cross resistant to ara-C.

We determined the effect of gemcitabine on the growth of the RL-7 and RL-G tumors when grown subcutaneously as xenografts in athymic nude mice. Median tumor weights in mg of RL-7 and RL-G xenografts in Swiss nude female mice treated or not with gemcitabine are shown in Figure 1. At day 60 after tumor injection, the RL-7 xenografts treated with gemcitabine were significantly smaller than the control tumors (Student t test, p = 0.04) (Fig. 1). The T/C value for RL-7 xenografts was 0.24. Conversely, no significant differences in tumor size were observed between the treated and control groups of mice xenotransplanted with the RL-G cells. The T/C value for RL-G xenografted mice was 0.98. These results indicate that RL-7 tumors were significantly more sensitive to gemcitabine treatment than the RL-G tumors.

To assess whether a nucleoside transport defect was a contributing factor to the gemcitabine resistance of RL-G cells, the initial uptake rates (2–10 s) of 10 μM [³H]gemcitabine by RL-7 and RL-G cells were compared, either in the absence or presence of excess (100-fold) non-radioactive gemcitabine. Both RL-7 and RL-G cells were able to transport [³H]gemcitabine with comparable initial uptake rates of 0.81 ± 0.28 and 0.5 ± 0.22 pmol/10⁶ cells, respectively (Fig. 2A). In the presence of excess non-radioactive gemcitabine, initial uptake rates were decreased by 72% (RL-7) and 65% (RL-G), indicating that the uptake of [³H]gemcitabine was mediated by a transporter in both cell lines. The initial uptake rates of 10 μM [³H]uridine, a physiologic nucleoside, by RL-7 and RL-G cells were also measured and found to be comparable to those measured with gemcitabine (data not shown). These data suggested that RL-G cells were capable of transporting nucleosides intracellularly and that a transport defect was not involved in the gemcitabine-resistance mechanism.

Although RL-G cells were capable of inward nucleoside transport, the cells were only able to metabolize [³H]gemcitabine in trace amounts over a 4 h exposure period as shown by the lack of measurable phosphorylated metabolites detected by HPLC analysis. Trace amounts of [³H]dFdCTP could be detected in some samples if the scale was greatly amplified. In contrast, RL-7 cells accumulated 16.4-fold more [³H]gemcitabine-derived compounds than RL-G cells (Fig. 2B). Approximately 90% of the total radioactivity measured in acid extracts of RL-7 cells was recovered as the cytotoxic triphosphate (2.2 μmol/mg protein) and diphosphate (0.98 μmol/mg protein) derivatives with a [³H]dFdCTP/[³H]dFdCDP ratio of 2.3. ATP/ADP ratios for untreated vs. gemcitabine-treated cells, were 3.6 vs. 2.8 for RL-7, and 3.1 vs. 2.1 for RL-G, respectively. Since gemcitabine has the ability to be incorporated into both DNA and RNA, the amount of [³H]gemcitabine present in the nucleic acid extractions was determined. No [³H]gemcitabine was detected in the DNA or RNA of RL-G cells as expected (Fig. 2C). In RL-7 cells, the incorporation of [³H]gemcitabine was higher in genomic DNA (11.9-fold) than in total RNA (Fig. 2C).

Flow cytometry of RL-7 and RL-G cells demonstrated a diploid cell population. After 24 h of continuous exposure to gemcitabine, a substantial accumulation of RL-7 cells in G0/G1 phase was observed with a decrease in the G2/M fraction (Table 2). At 72 h, a maximal arrest of cells in the S phase fraction (84%) was detected. The G0/G1 fraction dropped dramatically as the cells moved into S phase. Conversely, in RL-G cells gemcitabine treatment slightly arrested the cell cycle in S-phase at the expense of the G2/M fraction (Table 2). No G1 phase arrest was observed in this cell line (Table 2).

Exposure of cells for 24 h and 72 h with gemcitabine caused a proportional increase of annexin-V staining typical of apoptosis in RL-7 cells. However, the amount of apoptotic cells did not increase in the RL-G cell line (Fig. 3A). Poly-(ADP-ribose) polymerase (PARP) cleavage was detected after 72 h of exposure to gemcitabine of RL-7 cells (Fig. 3B). As shown in Figure 3C, no significant modifications of the anti-apoptotic protein Bcl-2 and the pro-apoptotic protein Bax levels were detected after gemcitabine exposure.

To test whether resistance of RL-G cells to gemcitabine was due to multiple mechanisms, the mRNA levels of several biological parameters (and potential drug targets) involved in gemcitabine metabolism, were analyzed by real time-PCR. According to quantitative RT-PCR results shown in Figure 4A, RL-G cells had no detectable dCK expression compared to RL-7 cells. In contrast, the 5'-nucleotidases, cN-II and dNT-2, were slightly increased in RL-G cells (2-fold). Similarly, expression levels of hENT1 and RNR subunits 1 and 2 were increased in 2.7-, 3.4- and 1.8-fold in RL-G cells, respectively (Figure 4B). hENT2, 5'-nucleotidase dNT1 and DNA POL mRNA were expressed at similar levels in both cell lines. Neither RL-7 cells nor RL-G cells expressed detectable CDD (data not shown).

To determine the cellular modification responsible for dCK mRNA down-regulation, we performed a qualitative study of the dCK gene. For this purpose, exons of the dCK gene as well as 5'-untranslated regions were amplified by PCR as described in the methods section. Figure 5 shows PCR products obtained using genomic DNA from RL-7 and RL-G cells. In resistant RL-G cells, no amplicons were obtained for regions located before the fourth exon, suggesting a partial deletion of the dCK gene in this cell line. Exons 4, 5, 6 and 7 of the human dCK gene were amplified in both cell lines. These results suggest a genetic modification of the dCK gene in RL-G cells.

Gemcitabine (Gemzar^®) was a kind gift from Eli Lilly (Indianapolis, IN, USA). Stock solutions were prepared in distilled water and fresh dilutions were prepared before each experiment. Dilazep, tri-N-octylamine, 1,1,2-trichlorotrifluoroethane, propidium iodide and MTT were purchased from Sigma-Aldrich (St. Louis, MI, USA). Bradford protein assay solution was purchased from Bio-Rad Laboratories (Ontario, Canada). [³H]Gemcitabine (14 Ci/mmol) and [³H]uridine (17.7 Ci/mmol) were purchased from Moravek Biochemicals Inc. (Brea, CA, USA). Ecolite™ was purchased from ICN (Costa Mesa, CA, USA). Antibodies against Bax were purchased from Santa Cruz Biotechnology Inc. (Santa Cruz, CA, USA); antibodies against Bcl-2 were purchased from DAKO (Glostrup, Denmark). Peroxidase-conjugated secondary antibodies were purchased from Covalab (Oullins, France). Enhanced chemiluminescence western blot detection reagents (ECL system) were purchased from Amersham (Amersham Corp, Buckinghamshire, UK). Other reagents used were analytical or HPLC grade and commercially available.

The RL-7 cells are derived from a human follicular lymphoma 43. The resistant variant RL-G was developed using step-wise increases of concentration of gemcitabine over a 12 month period. The maximum concentration used during selection was 2 μM. All cells were grown on 25 cm² flasks at 37°C in RPMI containing 10% fetal calf serum, 1% L-glutamine and 2% penicillin-streptomycin in a humidified atmosphere containing 5% CO₂.

Cellular uptake assays using suspensed cells have been described previously 244. Assays (10⁶ cells/tube) were conducted at room temperature in transport buffer (20 mM Tris/HCl, 3 mM K₂HPO₄, 1 mM MgCl₂.6H₂O, 2 mM CaCl₂, 5 mM glucose and 130 mM NaCl, pH 7.4, 300 ± 15 mOsm) at various time intervals. Cell pellets were lysed with 5% Triton X-100 and mixed with Ecolite™ scintillation fluid to measure the cell-associated radioactivity (Beckman LS 6500 scintillation counter; Beckman-Coulter Canada, Mississauga, ON). Initial uptake rates were measured over 2–10 s and drug accumulation was determined after exposure of cells to 10 μM [³H]gemcitabine for 4 h.

Exponentially growing RL-7 and RL-G cells were exposed to 10 μM [³H]gemcitabine for 4 h at 37°C. Cells were harvested by centrifugation and washed with phosphate-buffered saline. Aliquots (5 × 10⁶ cells) were snap-frozen in liquid nitrogen and stored at -70°C before HPLC analysis and nucleic acid extraction. DNA was isolated using a Wizard genomic DNA kit (Promega Corp., Madison, WI, USA). Total RNA was extracted using an RNA easy kit (Qiagen, Valencia, CA, USA) and treated with RNase-free DNase 1 (Pharmacia Biotech) to remove any DNA contamination. Radioactivity incorporation was determined by scintillation counting. Experiments were performed in duplicate.

The level of mRNA expression of metabolic factors involved in gemcitabine resistance was assessed by quantitative real time RT-PCR, performed in a LightCycler detection system (Roche, Mannheim, Germany) as previously described 22. Briefly, cDNA (5 μl) was mixed with primers (300 nM each), LightCycler-FastStart DNA Master SYBR Green I (Roche) (hENT1, hENT2, CDD, RNR-M1, RNR-M2 and DNA POL) or LightCycler-FastStart DNA master hybridization probes (Roche) (18S, dCK, cNII, dNT-1 and dNT-2), and probes (130 nM; if necessary) in a total volume of 20 μl. These reactions were prepared in duplicate. Primers and probes sequences for hENT1, dCK, CDD, cNII, dNT-1, RNR-M2 and DNA POL are published elsewhere 22. Primers for hENT2 were as follows: for: atgagaacgggattcccagtag ; rev: gctctgattccggctcctt. Primers for RNR-M1 were as follows: for: gcagctgagagaggtgcttt; rev: aatggttgtagaattaagaatagc. Primers and probes for dNT-2 were as follows: for: catcagcatttgggagtcaa; rev: cgacacaatctgctccagaa; probe: 5'-fam-cgtcttcatctgcacaagcccca-tamra. All samples were analyzed in three separate experiments.

For each sample (RL-7 or RL-G cell lines) and the calibrator (K562 cell line), the relative amount of a target gene and a reference gene (18S) were determined. Crossing point values (C_t), which are the PCR cycle numbers at which the accumulated fluorescent signal in each reaction crosses a threshold above background, were obtained with the LightCycler software 3.5 (Roche) using the second derivative maximum method. (C_t) values are a function of the amplification efficiency of the respective PCR. These data were then exported into the RelQuant software (Roche) as *.txt files. This software provides efficiency-corrected, calibrator-normalized quantification results. Results are expressed as the target/reference ratio of the sample divided by the target/reference ratio of the calibrator and therefore are corrected for sample inhomogenities and detection-caused variances. The efficiency-corrected quantification performed by Relquant is based on relative standard curves describing the PCR efficiencies of each target and the reference gene. The relative standard curves are determined and are used for each analysis.

Ratio results obtained with the RelQuant software were considered as final relative PCR arbitrary units. Results were then expressed as % of PCR arbitrary units in RL-G cells related to RL-7 cells PCR arbitrary unit expression.

Genomic DNA was prepared with a phenol/chloroform extraction method 46. The coding sequence of the 7 last exons of the human dCK gene were amplified in a final reaction volume of 25 μl using PCR primers (Table 1), Taq DNA polymerase (Invitrogen) and 250 ng of genomic DNA. The PCR program profile was: 10 min at 94 °C followed by 50 cycles of 30 sec at 94 °C, 30 sec at 60 °C and 30 sec at 72 °C. The PCR products were separated on agarose gel 1,5% using low molecular DNA weight (Invitrogen) as marker of size.

For analysis of DNA content and cell cycle distribution, RL-7 and RL-G cells were treated with gemcitabine 0.3 μM for 24 h and 72 h. After drug-exposure, 10⁶ cells/ml were resuspended in 2 ml of propidium iodide solution (50 μl/ml), incubated at 4°C overnight and then analyzed by flow cytometry. For apoptosis determination, we used the Annexin-V-Fluos Staining Kit (Boehringer Mannheim, Gmbh, Germany) as recommended by manufacturers. Flow cytometry was performed on a FACScalibur (Becton Dickinson, San Jose, CA, USA). For determination of the apoptotic fraction, analysis was performed using CellQuest™ software (Becton Dickinson, San Jose, CA, USA). Cell cycle distribution and DNA ploidy status were calculated after exclusion of cell doublets and aggregates on a FL2-area/FL2-width dot plot using Modfit LT 2.0™ software (Verity Software Inc, Topsham, ME, USA).

Protein expression was determined by western blot analysis in untreated RL-7 and RL-G cells and after 24 h incubation with gemcitabine 0.3 μM. Briefly, protein was extracted from cells with cold lysis buffer (20 mM Tris-HCL pH 6.8, 1 mM MgCl2, 2 mM EGTA, 0.5% NP40, STI 1 mg/ml, leupeptin 100 μg/ml, aprotinin 100 μg/ml, benzamidine 30 mg/ml, TPCK 1 mg/ml and PMSK 5 mg/ml). Cell lysates were resolved by 12% SDS-PAGE, and transferred onto a nitro-cellulose membrane (Hybond-ECL, Amersham Corp, Buckinghamshire, UK). The blots were incubated with the appropriate dilution of primary antibody, followed by incubation with peroxidase-conjugated secondary antibody. Protein signals were detected by chemiluminiscence and exposure to Kodak film (Eastman Kodak Company, Rochester, NY, USA). Horizontal scanning densitometry was performed on western blots by utilizing acquisition into Adobe PhotoShop (Apple, Cupertino, CA, USA).

Experiments were performed in 4 week-old female athymic nude mice (IFFA-CREDO, L'Arbresle, France). The animals were kept in conventional housing. Access to food and water was not restricted. RL-7 and RL-G xenografts were developed in groups of five mice. Each animal (average body weight 25 g before injection) received a subcutaneous injection in the right flank area containing 10⁶ RL-7 or RL-G cells in serum-free RPMI 1640 (day 0). After 24 h, gemcitabine was given intraperitoneally at a dose of 200 mg/kg once a week (6 doses; 3 consecutive weeks followed by 1 resting week) in the treated group. Initial drug toxicity studies were performed in non-tumor bearing mice at these doses and no major toxicities were observed (weight loss <10%).

The mice were inspected daily for s.c. tumor development and evaluation of clinical condition, and once tumor development occurred, tumor size was determined by caliper measurements while monitoring changes in animal weight twice a week. Animals were euthanized when their total tumor burden reached approximately 5,000 mg (15% of body weigth) to avoid animal discomfort or if clinical conditions may anticipate the potential for animal suffering.

The endpoints for assessing tumor activity were as follows: (a) tumor weight (mg) = (A × B²)/2, where A and B are the tumor length and width (in mm), respectively; (b) tumor growth inhibition = T/C, where T is the weight of treated tumor and C is the weight of control group, both evaluated at 60 days after injection. All studies involving mice were performed under Institutional Review Board-approved protocol.