Role of deoxycytidine kinase (dCK) activity in gemcitabine's radioenhancement in mice and human cell lines in vitro

Radiotherapy and Oncology 63 (2002) 329–338 www.elsevier.com/locate/radonline

Role of deoxycytidine kinase (dCK) activity in gemcitabine’s radioenhancement in mice and human cell lines in vitro Vincent Gre´goire a,*, Jean-Franc¸ois Rosier a,b, Marc De Bast a, Monique Bruniaux a,b, Blanche De Coster a, Michelle Octave-Prignot a, Pierre Scalliet a a

Department of Radiation Oncology and Laboratory of Radiobiology, Universite´ Catholique de Louvain, St-Luc University Hospital, 10 Ave Hippocrate, 1200 Brussels, Belgium b Tumoral Metabolism Research Unit, C.H. Jolimont-Lobbes, 7161 Haine-St-Paul, Belgium Received 10 December 2001; received in revised form 8 April 2002; accepted 7 May 2002

Abstract Background: Gemcitabine (dFdC, 2 0 ,2 0 -difluorodeoxycytidine) is a deoxycytidine nucleoside analog which has a marked effect on several enzymes involved in DNA synthesis and repair. Gemcitabine has been tested as a radiosensitizer in various biological models, and radiation dose modification factors (DMF) have been reported in the range between 1.1 and 2.4. Gemcitabine is a prodrug that requires intracellular activation by phosphorylation into its active triphosphate dFdCTP form. Deoxycytidine kinase (dCK) is the enzyme involved in the first phosphorylation cascade, and several observations have suggested that dCK was a limiting factor for the cytotoxic activity of gemcitabine. Objective: In the present article, we investigated the relationship between dCK activity and gemcitabine’s radiosensitization in four mice and two human cell lines. Materials and methods: Four mice and two human tumor cell lines were investigated. Radiosensitization was assessed on confluent cell incubated with 5 mM gemcitabine for 3 h prior to a single radiation dose. Enzymatic activity was assessed using deoxycytidine as substrate with (specific activity) or without (total activity) inhibition of thymidine kinase 2 activity. dCK protein level was assessed by immunoblotting using a rabbit anti-human dCK antibody. mRNA expression was assessed with Northern blot using b-actin as internal control. Results: Gemcitabine’s radiosensitization was heterogeneous with DMF ranging from 0.8 to 1.5. A good correlation was observed between the specific dCK activity and the protein level or the mRNA expression indicating that in our cell systems no post-transcriptional or posttranslational activation occurred. An excellent correlation ðr ¼ 0:99Þ was observed between the specific enzymatic activity and gemcitabine’s radiosensitization. Cell lines that expressed a high enzymatic activity were the more radiosensitized by gemcitabine. This correlation holds when radiosensitization was plotted against the dCK mRNA expression and protein level. Conclusions: The present study has suggested the role of dCK activity in gemcitabine’s radioenhancement in human and mice cell lines. The study suggests that determination of the enzymatic activity prior to a concurrent gemcitabine and radiotherapy treatment might represent a good predictive assay for tumor response. Such concept should deserve further testing in pre-clinical and clinical settings. q 2002 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Radioenhencement; Deoxycytidine kinase; Cell lines; Gemcitabine

1. Introduction Gemcitabine (dFdC, 2 0 ,2 0 -difluorodeoxycytidine) is a deoxycytidine nucleoside analog which has a marked effect on several enzymes involved in DNA synthesis and repair [32,34]. Like other nucleoside analogs, gemcitabine is a prodrug that requires intracellular activation by phosphorylation into its active triphosphate dFdCTP form. dFdCTP is incorporated into DNA at the penultimate position and blocks further elongation of DNA strand. An array of selfpotentiation mechanisms have been identified and they

likely contribute to both the high accumulation and the low elimination of the intracellular dFdCTP [17]. Gemcitabine can also be incorporated into RNA [42] and can also induce apoptosis [4,14,20]. Gemcitabine has been tested in various phase I–II trials and promising clinical activity has been reported in non small cell lung cancer, as in pancreatic, ovarian, breast, bladder and head and neck tumors [1,5]. Gemcitabine has been tested as a radiosensitizer in various biological models. In vitro, radioenhancement was observed in several human and mice tumor cell lines with radiation dose modification factors (DMF) ranging from 1.1 to 2.4, depending on both drug incubation time and concen-

* Corresponding author. 0167-8140/02/$ - see front matter q 2002 Elsevier Science Ireland Ltd. All rights reserved. PII: S 0167-814 0(02)00106-8

330

V. Gre´ goire et al. / Radiotherapy and Oncology 63 (2002) 329–338

tration [11,23,24,25,36,38,47,48]. No preferential radioenhancement was observed in more radioresistant cell lines [38]. In vivo, dFdC was reported to enhance radiationinduced tumor regrowth delay and tumor cure with a more pronounced effect in a fractionated setting of irradiation and drug administration [22,29]. Studies on normal tissue radiotolerance have shown that dFdC sensitized early reacting tissues, but has minimal or no effect on late responding tissues [10,12,27]. A phase I clinical trial combining gemcitabine and radiotherapy for head and neck squamous cell carcinomas has been recently completed [9]. Despite extensive studies, the molecular mechanism of interaction between gemcitabine and ionizing radiation remains partially unresolved. In several in vitro models, radioenhancement was associated with dATP depletion [24,47]. However, an absence of radioenhancement was observed in a glioblastoma cell line despite a profound dATP depletion [30,35,39]. Cell cycle synchronization has been reported after gemcitabine treatment both in vitro and in vivo [10,25,29,37,48]. Accumulation of cells in a more radiosensitive phase of the cycle could explain some degree of radioenhancement. The interaction of gemcitabine with ionizing radiation at the genomic level has been studied, but conflicting data have been observed. Whereas gemcitabine was reported to inhibit the repair of chromosome breaks, however, no effect on the repair of DNA DSBs after a single dose of irradiation could be demonstrated [11,21,25,40]. In addition, it has been recently reported that functional nonhomologous end-joining pathways were not required for gemcitabine’s radiosensitization [35]. As a prodrug, gemcitabine requires intracellular phosphorylation into its active triphosphate form to exhibit its biological activity. Deoxycytidine kinase (dCK) is the enzyme involved in the first phosphorylation cascade to the monophosphate form, dFdCMP [3]. There are observations suggesting that dCK is a limiting factor for the cytotoxic activity of gemcitabine. Several cell lines, which lack dCK activity, have been reported to be resistant to nucleoside analogs such as ara-C or dFdC [2,31,44,45,50]. It has also been shown that sensitivity to nucleoside analogs could be restored by transfection of a wild type dCK cDNA [16,26,51]. The critical role of dCK in nucleoside analog’s radioenhancement is not known but can be hypothesized on the basis of a number of observations. It has been reported that gemcitabine’s radioenhancement in vitro increased with drug concentration in various cell lines [38]. These observations which have been reported by others, suggest that radiosensitization by nucleoside analogs also depends on the rate of drug phosphorylation which in turn depends on dCK activity [24,38,48]. In the present article, we investigated the relationship between the dCK activity and gemcitabine’s radioenhancement in mice and human cell lines. We reported that cells expressing the highest basal enzymatic activity were also more radiosensitized by gemcitabine.

2. Materials and methods 2.1. Cell culture The C3H mice fibrosarcoma cell lines designated FSA, NFSA and SA-NH were generously supplied by L. Milas from M.D. Anderson Cancer Center in Houston, TX, USA. The mice fibroblast NIH-3T3 and the leukemic L1210 cell lines were a gift from P. Coulie from the Ludwig Institute for Cancer Research in Brussels, Belgium. The human head and neck squamous cell carcinomas designated SCC-61 and SQD-9 were a gift from A.C. Begg from the Netherlands Cancer Institute, The Netherlands. The MCF-7 mammary carcinoma cells were a gift from M. Symann from the Oncology Laboratory of the Catholic University of Louvain, Brussels, Belgium. The fibrosarcoma cells were cultured in McCoy’s 5A medium (Gibco BRL, Paisley, UK) supplemented with 10% fetal calf serum (Gibco BRL, Paisley, UK) and 2% penicillin/streptomycin (10.000 UI/ml– 10.000 mg/ml, Gibco BRL, Paisley, UK). The SCC-61, SQD-9 and NIH-3T3 cells were cultured in Dulbecco MEM Medium (Gibco BRL, Paisley, UK) supplemented with 10% fetal bovine serum, 1% penicillin/streptomycin and 0.1% hydrocortisone. The MCF-7 and L1210 cells were cultured in Dulbecco MEM Medium (Gibco BRL, Paisley, UK) supplemented with 10% fetal bovine serum and 1% penicillin/streptomycin. All cells were cultured at 378C in a humidified 5% CO2 atmosphere. All experiments were performed with confluent cells having on average less than 11% (SCC-61, SQD-9 and NIH-3T3 cells) or 15–20% (FSA, NFSA and SA-NH cells) replicating S-phase cells (as measured by flow cytometry after a pulse-labeling of BrdUrd). 2.2. Nucleoside analogs DFdC was generously supplied by ELI LILLY and COMPANY (Indianapolis, IN, USA). The drug was reconstituted in 1 £ PBS to a concentration of 1 mM, pH adjusted to 7.0 ^ 0.2 with NaOH solution, filtered through a 0.45 mM Acrodiscw filter and stored in small aliquots at 2208C. Before each experiment, the stock solution was thawed and diluted to various concentrations in cultured conditioned medium. 2.3. Clonogenic cell survival assay The medium of confluent cells was sucked up, filtered to remove debris and floating cells and used to dilute dFdC at a chosen concentration of 5 mM. Cells were then washed with serum-free medium and reincubated in the conditioned medium containing dFdC or PBS only (control dish) in a humidified 5% CO2 atmosphere at 378C. After a 3 h incubation time, cells were washed twice with ice-cold serum-free medium, trypsinized and kept on ice at a concentration of 2.10 5 cells/ml in small sterile cryotubes. Control and dFdC treated cells were irradiated in suspension in cryotubes with

V. Gre´ goire et al. / Radiotherapy and Oncology 63 (2002) 329–338

graded doses of 250 kV X-rays at a dose rate of 1.30 Gy/ min. After irradiation, cells were diluted at appropriate concentrations and plated in Petri dishes. After a 7 day (mice cells) or 13 day (human cells) incubation period in a humidified 5% CO2 atmosphere at 378C, the dishes were stained with crystal violet and colonies with more than 50 cells were counted. 2.4. dCK assay dCK activity was measured by an enzymatic assay as already described [41,49]. Briefly, a minimum of 10 7 confluent cells were trypsinized, washed twice in cold 1 £ PBS, centrifuged at 13,000g for 20 min at 48C and the pellets were frozen-thawed in liquid nitrogen three times to disrupt the cell membranes. Disrupted cell pellets were kept in liquid nitrogen at 2808C until enzymatic reaction was performed. Cells were then thawed and homogenized in 100 ml cold dCK buffer containing 0.3 M Tris at pH 8.0 (Sigma Chemical Co., St Louis, MO, USA) and 50 mM bmercaptoethanol (Sigma Chemical Co., St-Louis, MO, USA). The suspension was centrifuged at 13,000g for 20 min at 48C. Twenty-five microliters of the supernatant was used to determine the protein content. To 25 ml of the supernatant containing the dCK, 25 ml of the substrate mixture was added. The substrate mixture was prepared with one volume of dCK buffer, two volumes of Mg-ATP solution containing 50 mM Na2-ATP (Sigma Chemical Co., St-Louis, MO, USA) and 25 mM MgCl2 (Sigma Chemical Co., St Louis, MO, USA) at pH 7.4, and two volumes of 3HCdR (Amersham Pharmacia Biotech, Belgium) with a specific activity of 681 GBq/mmol and a concentration of 1.14 mM. The reaction mixture was incubated at 378C. At various incubation times from 0 to 20 min, 5 ml of the reaction mixture was removed, dropped on polyethylene imine (PEI)-cellulose layers (Merck Belgolabo, Belgium) preloaded with 5 ml of 10 mM cold CdR, and allowed to migrate for about 15 min using distillated water as eluant. Spots were then visualized under 254 nm UV light and marked out. A solution of cold dCMP was dropped on the PEI-cellulose layer to visualize the separation between dCMP, dCDP, dCTP, and CdR. PEI-cellulose layers were cut in small strips to separate CdR from the phosphorylated nucleosides, and incubated with 0.75 ml of a 0.1 M HCl and 0.2 M KCl solution for at least 2 h at room temperature in scintillation vials. Ultima gold (4.5 ml, Canberra Packard Benelux) was added, and after agitation and stabilization for at least 2 h each, the radioactivity was measured with a Wallacw scintillation counter. The percentage of phosphorylated nucleosides formed was plotted as a function of the incubation time. A linear fit was applied after subtraction of the background level (incubation of 0 min) to calculate the percentage of product formed per unit of time. The dCK activity was expressed as pmoles of product formed per 30 min/mg of protein. For each cell line, 3–9 independent experiments were performed.

331

2.5. RNA extraction and Northern blot analysis RNA was extracted by a non-ionic detergent method as previously described [8]. Briefly, 10–20 £ 10 6 cells were trypsinized, washed and lyzed with the non-ionic detergent Nonidet P-40 (NP40, Sigma Chemical Co., St Louis, MO, USA). RNA was extracted with SS-phenol and chloroformisoamyl at a ratio of 24:1. RNA was precipitated in cold 100% ethanol. Contaminant DNA was eliminated by digestion with RNAse-free DNAse (Promega, Madison, WI, USA) for 30 min at 378C. RNA was extracted again with SS-phenol and chloroform, and precipitated with cold 100% ethanol. After centrifugation, the pellet was resuspended in 50 ml of Tris–EDTA at pH 8 and RNA concentration was measured with a spectrophotometer at OD of 260 and 280 nm. RNA was stored at 2808C until electrophoresis. For RNA electrophoresis, a 1% formaldehyde agarose gel containing 0.15 mg/ml ethidium bromide was prepared. Denatured RNA was loaded and the electrophoresis was run at 125 V and 250 mA for 3 h. Gels were washed with 10 £ SSC and blotted overnight on nylon GeneScreen Plus (Dupont-NEN). cDNA probes for human and mice dCK were kindly provided by S. Eriksson from the University of Uppsala, Sweden. The probes were denatured at 1008C and labeled for 20 min at 378C with 32P-dCTP using the Prime-It II kit (Stratagene, La Jolla, CA, USA). Labeled probes were collected using the QIAquick nucleotide removal kit (Quiagen Inc, Santa Clarita, CA, USA). Hybridization was performed at 708C for 90 min under mild agitation in ExpressHyb solution (ClonTech, Palo Alto, CA, USA). Blots were washed three times in 2 £ SCC and 0.05% SDS, then twice in 0.1 £ SCC and 0.1% SDS. Blots were autoradiographed using X-OMat films (Eastman Kodak Co., Rochester, NY, USA) at 2808C for various exposure times. After exposure, blots were stripped and hybridized again with b-actin used as internal control. For both dCK and b-actin probes, various exposure times were used to insure a linear blackening of the film with exposure time. Films were scanned and analyzed with the public domain NIH software. For each blot, the optic density value of dCK was related to the OD value of b-actin to take into account possible variation in the amount of RNA loaded from cell line to cell line. Variations in exposure from blot to blot were taken into account by expressing the dCK over b-actin ratio of the cell of interest to the dCK over bactin ratio of control cells loaded in every blot. MCF-7 and L1210 cells were used for the human and the mice dCK, respectively. Data are average values (^SEM) of six independent experiments. 2.6. Quantitation of dCK by immunoblotting Quantitation of the dCK protein was performed as already described [49]. Briefly, proteins were extracted as described in Section 2.4. Proteins were dissolved in 375 mM Tris–Cl buffer containing 0.9% SDS, 2.2 M b-mercaptoethanol, 3%

332

V. Gre´ goire et al. / Radiotherapy and Oncology 63 (2002) 329–338

glycerol and 0.02% bromophenol blue. After homogenization, samples were boiled at 958C for 5 min, centrifuged at 13,000g and 30 mg was loaded into a 12.5% polyacrylamid gel. Electrophoresis was performed using a Mini-P II cell (Biorad, Hercules, CA, USA) at 100 V and 225 mA for 2 h. Electrotransfert on nitrocellulose sheet was then carried out overnight at 30 V and 40 mA in Towbin blotting buffer (25 mM Tris–base buffer containing 192 mM glycine and 20% methanol). Nitrocellulose sheets were then blocked for 1 h at room temperature with 5% milk in TBST buffer (TBS 1 0.05% Tween 20). After blocking, sheets were rinsed several times with TBST buffer and incubated for 2 h at room temperature with the rabbit anti-human dCK antibody diluted (1/1000, v/v) in TBST with 1% milk. Anti-dCK antibody was kindly provided by S. Eriksson from the University of Uppsala, Sweden. The sheets were washed several times with TBST buffer and incubated for 1 h at room temperature with the peroxydase-conjugated goat anti-rabbit IgG antibody (Jackson Laboratories, West Grove, PA, USA) diluted (1/20.000 v/v) in TBST with 1% milk. Sheets were then washed several times in TBST and incubated in 1 ml ECL solution (NEN Life Science, Boston, MA, USA) for 1 min. Sheets were then autoradiographed using hyperfilms ECL (Amersham Pharmacia Biotech, Belgium). Various exposure times from 0.5 to 5 min were used to insure a linear blackening of the film with exposure time. Films were scanned and analyzed with the public domain NIH software. For each blot, the optical density value of dCK was related to the OD value of the L1210 control cells used as external control to take into account variation in exposure from blot to blot. Data are average values (^SEM) of four independent experiments. 2.7. Analysis of the clonogenic cell survival data Cell survival curves were fitted through the linear-quadratic model. A least square regression analysis was performed on the mean value of the logarithm of the surviving fractions for each radiation dose. A normal distribution of the logarithm of the surviving fraction was assumed. For each radiation dose, at least four separate experiments were performed. For cells treated with dFdC, surviving fractions were corrected for the toxicity of the drug alone. DMFs were calculated at different levels of survival from the fitted curves. Ninety five percent confidence limits on the DMF were calculated by the method described by McLarty and Thames [28]. 3. Results 3.1. Radiosensitization by dFdC The radiosensitization effect of dFdC was studied in vitro by its ability to increase the radiation-induced lethality on various mice and human tumor cell lines. To minimize the cytotoxic effect of dFdC alone, a low drug concentration of

5 mM was used on confluent cells. Such drug concentration is within achievable range in the clinical setting [15,32]. A short drug incubation time of 3 h was chosen before irradiation to allow adequate intracellular dFdCTP concentration and to minimize other mechanisms of radiosensitization, e.g. cell synchronization in a more radiosensitive phase of the cycle [15,43]. Under these experimental conditions, the cytotoxic effect of dFdC alone resulted in a mean ^ SEM surviving fraction of 0.71 ^ 0.10, 0.53 ^ 0.07, 0.74 ^ 0.08, 0.68 ^ 0.15, 0.53 ^ 0.02 and 0.72 ^ 0.09 for SA-NH, FSA, NFSA, NIH-3T3, SCC-61 and SQD-9 cells, respectively. As illustrated in Fig. 1, radiosensitization by dFdC was observed in three out of the six cell lines investigated, FSA, SQD-9 and SCC-61 cells. In FSA cells, however, radiosensitization appeared very modest especially at low surviving fractions. All the cell survival curves were fitted through the linear-quadratic model, except SCC-61 where a linear model was used. When occurred, radiosensitization was associated with an increase in the a term (Table 1). From the resulting fits, DMFs were derived for a survival fraction of 0.50 corresponding roughly to a dose of 2 Gy in the cells treated by irradiation alone, except for the more radiosensitive SCC-61 cells for which it corresponded roughly to a dose of 1 Gy. In radiosensitized cells, DMFs ranged from 1.1 to 1.5 (Table 1). In FSA cells, the DMF of 1.1 ^ 0.2 did not reach, however, a level of significance. 3.2. dCK assay To adequately assess the dCK activity in the various cell lines, it was necessary to perform the enzymatic reaction at saturation of the enzyme, i.e. with large concentration of the substrate CdR. The Km values of the enzymatic reaction were first assessed in SA-NH and NIH-3T3 cell lines. Km values reached 2.44 and 2.64 mM in SA-NH and NIH-3T3, respectively. Consequently, a final CdR concentration of 230 mM was used for all enzymatic reactions. For each cell line, the linearity of the CdR phosphorylation with time was also carefully assessed. Linearity was always observed for the first 15–30 min of reaction. dCK activities of the various cell lines are presented in Table 2. The two human lines consistently exhibited a much higher activity than the four mice lines. As it is known that phosphorylation of CdR can also be performed by thymidine kinase 2 (TK2), the activities reported might thus overestimate the true activity of the dCK enzyme. In this framework, the dCK activities were also assessed in the presence of excess of thymidine (1.2 mM) to saturate the TK2, which has a much better affinity to thymidine than to CdR. As shown in Table 2, a lower dCK specific activity was measured in all cell lines, with a more dramatic decrease in the two human lines. 3.3. Semi-quantitative assessment of dCK mRNA expression and protein content The previous experiments have shown that determination

V. Gre´ goire et al. / Radiotherapy and Oncology 63 (2002) 329–338

333

Fig. 1. Radiosensitization by dFdC in various mice (SA-NH, FSA, NFSA, NIH-3T3) and human (SCC-61 and SQD-9) cells by dFdC. Confluent cells were incubated with 5 mM of dFdC for 3 h prior to irradiation. For the combined treatment, the cell survival curves were corrected for the lethality induced by the drug alone. Each data point is the average of four to seven separate experiments ^ SEM.

V. Gre´ goire et al. / Radiotherapy and Oncology 63 (2002) 329–338

334 Table 1 Results of the cell survival curve fitting Cell line

Linear-quadratic parameters of the cell survival curve fit (mean ^ SD) RT 1 dFdC (3 mM for 3h)

RT alone 21

SA-NH FSA NFSA NIH-3T3 SCC-61 SQD9 a

a (Gy ) 0.14 ^ 0.04 0.26 ^ 0.03 0.18 ^ 0.07 0.51 ^ 0.06 0.91 ^ 0.01 0.30 ^ 0.02

DMF (SF: 0.5) (mean ^ 95% CI)

22

b (Gy ) 0.056 ^ 0.001 0.030 ^ 0.001 0.051 ^ 0.014 0.078 ^ 0.020 –a 0.024 ^ 0.001

a (Gy 21) 0.08 ^ 0.07 0.28 ^ 0.04 0.05 ^ 0.03 0.22 ^ 0.01 1.36 ^ 0.02 0.38 ^ 0.04

b (Gy 22) 0.083 ^ 0.011 0.045 ^ 0.005 0.077 ^ 0.007 0.21 ^ 0.001 –a 0.043 ^ 0.005

0.9 (0.7–1.4) 1.1 (0.9–1.4) 0.8 (0.6–1.0) 0.9 (0.7–1.3) 1.5 (1.4–1.53) 1.3 (1.1–1.6)

Fitted through a linear model.

of the true enzymatic activity of dCK was not so obvious as the substrate CdR could also be phosphorylated by TK2. Although this effect could be by-passed by using an excess of thymidine in the reaction mixture, it may well be that activity of other enzymes, e.g. cytidine deaminase, nucleoside phosphorylase, also comes into play. In this framework, the dCK mRNA expression and protein content were assessed by Northern blot and Western immunoblotting, respectively. Results are presented in Table 2. For mRNA expression, results are expressed in arbitrary units relative to b-actin used as an internal standard. For protein content, as no purified dCK was used in the blots, the results are also reported in arbitrary units after normalization to the dCK from a fixed amount of L1210 cells loaded in every blot as standard. Large differences in both mRNA expression and protein content were observed among the various cell lines. As for enzymatic activity, the two human cell lines exhibited a much higher mRNA expression and protein content than the mice counterparts. Overall, cell lines that expressed a high enzymatic activity also expressed a high mRNA expression and a high level of protein (Fig. 2). 3.4. Correlation between dCK and dFdC’s radiosensitization The correlation between dCK activity and dFdC’s radiosensitization is presented in Fig. 3. A good correlation was

observed between dCK activity, mRNA expression, protein content, and dFdC’s radiosensitization. Cell lines that expressed a high enzymatic activity were the more radiosensitized by dFdC. The correlation was somehow lower with the protein content and the mRNA expression reflecting the more heterogeneous spread of the data points for these two assays.

4. Discussion Among the nucleoside analogs, gemcitabine is widely used for the treatment of solid malignancies, especially pancreatic and lung tumors. In those tumors, as in head and neck squamous cell carcinomas, gemcitabine has been combined with ionizing radiation, taking profit of the radiosensitizing properties of the drug. Improvement in local tumor response has been suspected from phase I trials, and gemcitabine appears to be a promising agent to be combined with radiotherapy in further phase II studies [6,54]. However, increased local toxicities, both acute and late, have been typically reported, so that routine use of the combined modality treatment requires great caution. Furthermore, in experimental models, large variations in radiosensitization have been reported, addressing thus the issue of pre-treatment selection of tumors that might benefit the most from the combined modality treatment. Gemcitabine is a prodrug that requires successive intra-

Table 2 dCK activities of the various cell lines Cell line

SA-NH FSA NFSA NIH 3T3 SCC-61 SQD9 a

dCK activity (mean ^ SE) Total activity (nmol/30 min/mg protein)

Specific activity a (nmol/30 min/mg protein)

1.77 ^ 0.17 3.93 ^ 0.25 1.14 ^ 0.28 1.35 ^ 0.38 9.12 ^ 1.66 13.11 ^ 1.50

1.41 ^ 0.03 1.75 ^ 0.22 0.81 ^ 0.20 1.10 ^ 0.18 2.99 ^ 0.50 2.43 ^ 0.49

Saturation of TK2 by an excess of thymidine.

dCK protein content (mean ^ SE) (arbitrary unit)

dCK mRNA expression (mean ^ SE) (arbitrary unit)

45.3 ^ 6.9 32.9 ^ 4.3 4.0 ^ 1.5 6.3 ^ 1.4 100.8 ^ 16.7 155.5 ^ 36.7

16.4 ^ 3.5 10.4 ^ 3.4 8.8 ^ 5.3 33.1 ^ 20 71.6 ^ 20.0 53.4 ^ 13.7

V. Gre´ goire et al. / Radiotherapy and Oncology 63 (2002) 329–338

335

and possibly animal studies might be required to further validate the correlation. On the other hand, in the present study no correlation was observed between cytotoxicity of the drug alone and the various dCK parameters (see Table 2 and first paragraph of Section 3). As is known, dFdC is mainly active on Sphase cells, and variations of cell cycle distribution at the time of drug treatment is likely to translate into different cell kill. Although all experiments were performed on confluent cells, S-phase cell fraction typically varied between less than 11% for SCC-61, SQD-9 and NIH-3T3 and 15–20% for FSA, NFSA and SA-NH cells. The influence of such difference on dFdC-induced cell kill may have occulted any impact of the dCK activity. Among the six cell lines analyzed in the present study,

Fig. 2. Relationship between dCK specific activity and mRNA expression (top panel) or protein content (bottom panel) in various mice (SA-NH, FSA, NFSA, NIH-3T3) and human (SCC-61 and SQD-9) confluent cell lines. Each data point is the average (^SEM) of four to six separate experiments.

cellular phosphorylations into its active triphosphate form [33]. The enzyme dCK is required for the first phosphorylation step into dFdCMP, non-specific kinases being involved in the further phosphorylation steps. As it has been reported that the level of enzymatic activity of dCK could have a profound influence on cellular resistance to gemcitabine cytotoxicity, the present investigations were designed to address the relationship between gemcitabine’s radiosensitization and the activity of dCK in various mice and human cell lines. It shows that under the experimental conditions used in the study, a significant correlation was observed between dCK activity and gemcitabine’s radiosensitization. This correlation holds when radiosensitization was plotted against the dCK mRNA expression and protein content. The correlation is, however, based on only six cell lines expressing a rather limited range of dFdC’s radiosensitization, limiting thus the power of the data. Additional cell lines

Fig. 3. Relationship between the dCK activity (top panel), mRNA expression (middle panel), protein content (bottom panel), and dFdC’s radioenhencement in various mice (SA-NH, FSA, NFSA, NIH-3T3) and human (SCC-61 and SQD-9) confluent cell lines. Each data point is the average of four to seven separate experiments.

336

V. Gre´ goire et al. / Radiotherapy and Oncology 63 (2002) 329–338

four were of mice origin and they consistently expressed lower dCK activity than the two human counterparts. Similar observation has been reported between human and mice normal tissues [19]. Large variations in dCK activities have been, however, reported in a series of 28 human tumor xenografts and four human tumor cell lines in culture [41]. In this series, the tumors expressing the lowest dCK activity were in the range of activities reported in the present study with the mice cell lines, so that definite conclusions regarding tumor strain specificity for dCK activity cannot be formulated. The proliferation index of the various cells of interest is another factor that might influence the level of dCK activity. Inconsistent data have been reported regarding the cell cycle regulation of dCK activity; some studies reporting an increase in S-phase, others not (see review in Ref. [49]). In few of our cell lines, we indeed observed a higher enzymatic activity in exponentially growing cells (data not shown). In this context, in our study, all cell lines used for gemcitabine’s radiosensitization or for enzymatic activity measurements were cultured to reach the same level of confluence to minimize possible enzymatic variation in the cell cycle phases. Confluent cells were preferred to exponentially growing cells to minimize the effect of gemcitabine on proliferating S-phase cells. Other factors such as UV or X-ray irradiation have been shown to regulate dCK activity in mammalian cells [53]. Gemcitabine itself was shown to regulate dCK activity likely through a decrease in the pool of intracellular dCTP [46]. In our study, it is unlikely that one of these factors could have played a role to explain the various patterns of dCK activity among the cell lines, and if it did, all cell lines would have been subject to the same modifications as all experiments were performed under similar conditions. In the present study, CdR was used as a substrate for dCK. It is, however, known that CdR is also a good substrate for other enzymes such as TK2, so that the presence of TK2 in the cellular extract could lead to an overestimation of the dCK activity. Indeed, in the present study, the use of an excess of thymidine that is preferentially taken as a substrate by TK2, lead to a lower dCK activity in all six cell lines, but without modification of the ranking between them. Others have proposed to use CdA, which is considered to be a specific substrate for dCK [49]. They reported enzymatic activities of five human tumor samples in the range of what is reported in the present study. But, even using CdA, overestimation of the dCK activity as been reported due to phosphorylation by the mitochondrial deoxyguanosine kinase [52]. In this framework, it was of interest to also investigate the level of protein content and mRNA expression. In the cell lines investigated in the present study, a good correlation was observed between the protein content or the mRNA expression and the enzymatic activity (Fig. 2). These data are in good agreement with data reported in mononuclear cells from patients with different type of leukemia, where an excellent correlation ðr ¼ 0:88Þ was observed between CdA phosphorylation and dCK protein

level [49]. In our experiment, however, the data are a little more scattered in the relationship with the protein content than with the mRNA. One likely explanation is the fact that both the protein content and the mRNA expression were only assessed semi-quantitatively using appropriate controls to relate the various data. Another explanation is that these differences might reflect the existence of some degree of post-transcriptional activation of the protein as accounted for the cyclic variation of the enzyme activity in some cell systems [7,18]. Gemcitabine’s radiosensitization is a complex phenomenon whose mechanisms are still not fully understood. Cell cycle regulation, processing of DNA damages, and modification of the intracellular deoxynucleotide pool have been reported to be involved in gemcitabine’s radiosensitization (see review in Ref. [13]). Thus, in addition to the intracellular metabolism of gemcitabine for which dCK is a key factor, intrinsic variation in the cellular response to the analog after insults with ionizing radiation is likely to also play a role in gemcitabine’s radiosensitization. The relative importance of dCK activity among these various parameters is, however, unknown. In conclusion, the present study has suggested the role of dCK activity in gemcitabine’s radiosensitization in human and mice cell lines. Although it is not the only factor involved, the study suggests that determination of the enzymatic activity prior to a concurrent gemcitabine and radiotherapy treatment might represent a predictive factor for tumor response. Such concept should deserve further testing in pre-clinical and clinical settings.

Acknowledgements The authors thank Professor S Eriksson for providing the human and mice dCK cDNA probes and the anti-dCK antibody, and Professor G. Peters for helpful comments during the realization of the experiments. This investigation was supported by grant #3.4538.98 and #3.4574.00 from the ‘Fonds de la Recherche Scientifique Me´ dicale (FRSM)’, by the ‘Fonds Joseph Maisin’ and by a research grant from ELI LILLY and COMPANY.

References [1] Barton-Burke M. Gemcitabine: a pharmacologic and clinical overview. Cancer Nurs 1999;22:176–183. [2] Bergman AM, Pinedo HM, Jongsma AP, et al. Decreased resistance to gemcitabine (2 0 ,2 0 -difluorodeoxycytidine) of cytosine arabinosideresistant myeloblastic murine and rat leukemia cell lines: role of altered activity and substrate specificity of deoxycytidine kinase. Biochem Pharmacol 1999;57:397–406. [3] Bouffard DY, Laliberte J, Momparler RL. Kinetic studies on 2 0 ,2 0 difluorodeoxycytidine (Gemcitabine) with purified human deoxycytidine kinase and cytidine deaminase. Biochem Pharmacol 1993;45: 1857–1861. [4] Bouffard DY, Momparler RL. Comparison of the induction of apop-

V. Gre´ goire et al. / Radiotherapy and Oncology 63 (2002) 329–338

[5] [6] [7]

[8]

[9]

[10]

[11]

[12]

[13]

[14]

[15]

[16]

[17]

[18]

[19]

[20] [21]

[22]

[23]

[24]

tosis in human leukemic cell lines by 2 0 ,2 0 -difluorodeoxycytidine (gemcitabine) and cytosine arabinoside. Leuk Res 1995;19:849–856. Carmichael J. The role of gemcitabine in the treatment of other tumours. Br J Cancer 1998;78(Suppl 3):21–25. Choy H. Combination chemoradiotherapy with gemcitabine: potential applications. Oncology (Huntingt) 2000;14(Suppl 4):20–25. Csapo Z, Sasvari-Szekely M, Spasokoukotskaja T, Talianidis I, Eriksson S, Staub M. Activation of deoxycytidine kinase by inhibition of DNA synthesis in human lymphocytes. Biochem Pharmacol 2001;61:191–197. Davis L, Kuehl M, Battey J. Nonionic detergent for isolation of cytoplasmic (and nuclear) RNA, 2nd ed.. Basic methods in molecular biologyNorwalk: Appleton and Lange, 1994. pp. 329–334. Eisbruch A, Shewach DS, Bradford CR, et al. Radiation concurrent with gemcitabine for locally advanced head and neck cancer: a phase I trial and intracellular drug incorporation study. J Clin Oncol 2001;19:792–799. Gregoire V, Beauduin M, Rosier JF, et al. Kinetics of mouse jejunum radiosensitization by 2 0 -2 0 difluorodeoxycytidine (Gemcitabine) and its relationship with pharmacodynamics of DNA synthesis inhibition and cell cycle redistribution in crypt cells. Br J Cancer 1997;76:1315– 1321. Gregoire V, Beauduin M, Bruniaux M, Decoster B, Octave-Prignot M, Scalliet P. Radiosensitization of mouse sarcoma cells by Fludarabine or Gemcitabine, two nucleoside analogs, is not mediated by an increased induction or a repair inhibition of DNA double-strand breaks as measured by pulsed-field gel electrophoresis. Int J Radiat Biol 1998;73:511–520. Gregoire V, Cvilic S, Beauduin M, et al. Effect of Gemcitabine (dFdC) on the tolerance of the lung to single dose irradiation in C3H mice. Radiat Res 1999;151:747–749. Gregoire V, Hittelman WN, Rosier JF, Milas L. Chemo-radiotherapy: radiosensitizing nucleoside analogues (review). Oncol Rep 1999;6:949–957. Gruber J, Geisen F, Sgonc R, et al. 2 0 ,2 0 -Difluorodeoxycytidine (gemcitabine) induces apoptosis in myeloma cell lines resistant to steroids and 2-chlorodeoxyadenosine (2-CdA). Stem Cells 1996;14:351–362. Grunewald R, Kantarjian H, Du M, Faucher K, Tarassoff P, Plunkett W. Gemcitabine in leukemia: a phase I clinical, plasma, and cellular pharmacology study. J Clin Oncol 1992;10:406–413. Hapke DM, Stegmann AP, Mitchell BS. Retroviral transfer of deoxycytidine kinase into tumor cell lines enhances nucleoside toxicity. Cancer Res 1996;56:2343–2347. Heinemann V, Xu Y-Z, Chubb S, et al. Cellular elimination of 2 0 ,2 0 difluorodeoxycytidine 5 0 -triphosphate: a mechanism of self-potentiation. Cancer Res 1992;52:533–539. Hengstschlager M, Denk C, Wawra E. Cell cycle regulation of deoxycytidine kinase. Evidence for post-transcriptional control. FEBS Lett 1993;321:237–240. Ho DH. Distribution of kinase and deaminase of 1-beta-d-arabinofuranosylcytosine in tissues of man and mouse. Cancer Res 1973;33:2816–2820. Huang P, Plunkett W. A quantitative assay for fragmented DNA in apoptotic cells. Anal Biochem 1992;207:163–167. Huang NJ, Hittelman WN. Transcient inhibition of chromosome damage repair after ionizing radiation by gemcitabine (Abstract). Proc Am Assoc Cancer Res 1995;36:115 (see also p. 612). Joschko MO, Webster LK, Groves J, et al. Enhancement of radiationinduced regrowth delay by Gemcitabine in a human tumor xenograft model. Radiat Oncol Invest 1997;5:62–71. Latz D, Fleckenstein K, Eble M, Blatter J, Wannenmacher M, Weber KJ. Radiosensitizing potential of Gemcitabine within the cell cycle in vitro. Int J Radiat Oncol Biol Phys 1998;41:875–882. Lawrence TS, Chang E, Hahn TM, Hertel HW, Shewach DS. Radiosensitization of pancreatic cancer cells by 2 0 -2 0 difluorodeoxycytidine. Int J Radiat Oncol Biol Phys 1996;41:867–872.

337

[25] Lawrence TS, Chang E, Hahn, Shewach DS. Delayed radiosensitization of human colon carcinoma cells after a brief exposure to 2 0 -2 0 difluorodeoxycytidine (Gemcitabine). Clin Cancer Res 1997;3:777– 782. [26] Manome Y, Wen PY, Dong Y, et al. Viral vector transduction of the human deoxycytidine kinase cDNA sensitizes glioma cells to the cytotoxic effects of cytosine arabinoside in vitro and in vivo. Nat Med 1996;5:567–573. [27] Mason KA, Milas L, Hunter NR, et al. Maximizing therapeutic gain with gemcitabine and fractionated radiation. Int J Radiat Oncol Biol Phys 1999;44:1125–1135. [28] McLarty JW, Thames HD. A parametric method for estimation of dose-modifying factors: low-dose extrapolation limit. Radiat Res 1977;69:1–15. [29] Milas L, Fujii T, Hunter N, et al. Enhancement of tumor radioresponse in vivo by Gemcitabine. Cancer Res 1999;59:107–114. [30] Ostruszka LJ, Shewach DS. Cytotoxicity and radiosensitization of Gemcitabine in human glioblastoma cell lines (Abstract). Proc Am Assoc Cancer Res 1997;38:102 (see also p. 683). [31] Owens JK, Shewach DS, Ullman B, Mitchell BS. Resistance to 1beta-d-arabinofuranosylcytosine in human T-lymphoblasts mediated by mutations within the deoxycytidine kinase gene. Cancer Res 1992;52:2389–2393. [32] Peters GJ, Ackland SP. New antimetabolites in preclinical and clinical development. Exp Opin Invest Drugs 1996;5:637–679. [33] Plunkett W, Huang P, Gandhi V. Preclinical characteristics of gemcitabine. Anticancer Drugs 1995;6(Suppl 6):7–13 (Review). [34] Plunkett W, Huang P, Xu Y-Z, Heineman V, Grunewald R, Gandhi V. Gemcitabine: metabolism, mechanisms of action, and self-potentiation. Semin Oncol 1995;22(Suppl 11):3–10. [35] van Putten JWG, Groen HJM, Smid K, Peters GJ, Kampinga HH. End-joining deficiency and radiosensitization induced by gemcitabine. Cancer Res 2001;61:1585–1591. [36] Rockwell S, Grindey GB. Effect of 2 0 -2 0 difluorodeoxycytidine on the viability and radiosensitivity of EMT6 cells in vitro. Oncol Res 1992;4:151–155. [37] Rosier JF, Beauduin M, Bruniaux M, et al. The role of Gemcitabineinduced cell cycle synchronization in radiosensitization of a SCC61 human head and neck squamous carcinoma cell line (Abstract). Proc Eur Cancer Conf 1997;33:S23–S24. [38] Rosier JF, Beauduin M, Bruniaux M, et al. The effect of 2 0 -2 0 difluorodeoxycytidine (dFdC, Gemcitabine) on radiation-induced cell lethality in two human head and neck squamous carcinoma cell lines differing in intrinsic radiosensitivity. Int J Radiat Biol 1999;75:245– 251. [39] Rosier JF, Hondekijn JC, Bruniaux M, Beauduin M, Gre´ goire V. Intracellular metabolism of dFdC (gemcitabine) as a mechanism of radioenhancement in two human head and neck squamous cell carcinoma cell lines. Radiother Oncol 2000;56(Suppl 1):S114. [40] Rosier JF, Michaux L, Libouton JM, Verellen Ch, Scalliet P, Gre´ goire V. Repair inhibition of radation-induced chromosome aberrations but not DNA DSBs is involved in gemcitabines’s (dFdC) radio-emhancement. Radiother Oncol 1998;48(Suppl 1):S184. [41] Ruiz van Haperen VW, Veerman G, Braakhuis BJ, et al. Deoxycytidine kinase and deoxycytidine deaminase activities in human tumour xenografts. Eur J Cancer 1993;29A:2132–2137. [42] Ruiz van Haperen VW, Veerman G, Vermorken JB, Peters GJ. 2 0 ,2 0 Difluoro-deoxycytidine (gemcitabine) incorporation into RNA and DNA of tumour cell lines. Biochem Pharmacol 1993;46:762–766. [43] Ruiz Van Happeren VWT, Veerman G, Boven E, Noordhuis P, Vermoken JB, Peters GJ. Schedule dependence of sensitivity to 2 0 ,2 0 -Difluorodeoxycytidine (gemcitabine) in relation to accumulation and retention of its triphosphate in solid tumour cell lines and solid tumours. Biochem Pharmacol 1994;48:1327–1339. [44] Ruiz van Haperen VW, Veerman G, Eriksson S, et al. Development and molecular characterization of a 2 0 ,2 0 -difluorodeoxycytidine-resis-

338

[45]

[46]

[47]

[48] [49]

V. Gre´ goire et al. / Radiotherapy and Oncology 63 (2002) 329–338 tant variant of the human ovarian carcinoma cell line A2780. Cancer Res 1994;54:4138–4143. Ruiz van Haperen VW, Veerman G, Eriksson S, Stegmann AP, Peters GJ. Induction of resistance to 2 0 ,2 0 -difluorodeoxycytidine in the human ovarian cancer cell line A2780. Semin Oncol 1995;4(Suppl 11):35–41. Ruis van Haperen VW, Veerman G, Vermorken JB, Pinedo HM, Peters G. Regulation of phosphorylation of deoxycytidine and 2 0 ,2 0 -difluorodeoxycytidine (gemcitabine); effects of cytidine 5 0 -triphosphate and uridine 5 0 -triphosphate in relation to chemosensitivity for 2 0 ,2 0 -difluorodeoxycytidine. Biochem Pharmacol 1996;51:911–918. Shewach DS, Hahn TM, Chang E, Hertel HW, Lawrence TS. Metabolism of 2 0 -2 0 difluorodeoxycytidine and radiation sensitization of human colon carcinoma cells. Cancer Res 1994;54:3218–3223. Shewach DS, Lawrence TS. Radiosensitization of human tumor cells by Gemcitabine in vitro. Semin Oncol 1995;22:68–71. Spasokoukotskaja T, Arne´ r ESJ, Brosjo¨ O, et al. Expression of deoxycytidine kinase and phosphorylation of 2-chlorodeoxyadenosine in human normal and tumour cells and tissues. Eur J Cancer 1995;31A:202–208.

[50] Stegmann AP, Honders MW, Hagemeijer A, Hoebee B, Willemze R, Landegent JE. In vitro-induced resistance to the deoxycytidine analogues cytarabine (AraC) and 5-aza-2 0 -deoxycytidine (DAC) in a rat model for acute myeloid leukemia is mediated by mutations in the deoxycytidine kinase (dck) gene. Ann Hematol 1995;71:41–47. [51] Stegmann AP, Honders WH, Willemze R, Ruiz van Haperen VW, Landegent JE. Transfection of wild-type deoxycytidine kinase (dck) cDNA into an AraC- and DAC-resistant rat leukemic cell line of clonal origin fully restores drug sensitivity. Blood 1995;85:1188– 1194. [52] Wang L, Karlsson A, Arner ES, Eriksson S. Substrate specificity of mitochondrial 2 0 -deoxyguanosine kinase. Efficient phosphorylation of 2-chlorodeoxyadenosine. J Biol Chem 1993;268:22847–22852. [53] Wei S, Ageron-Blanc A, Petridis F, Beaumatin J, Bonnet S, Luccioni C. Radiation-induced changes in nucleotide metabolism of two colon cancer cell lines with different radiosensitivities. Int J Radiat Biol 1999;75:1005–1013. [54] Wolff RA, Evans DB, Gravel DM, et al. Phase I trial of gemcitabine combined with radiation for the treatment of locally advanced pancreatic adenocarcinoma. Clin Cancer Res 2001;7:2246–2253.