Cytotoxicity of 2-chlorodeoxyadenosine and arabinosylcytosine in leukaemic lymphoblasts from paediatric patients: significance of cellular nucleoside transporter content

Authors


Dr Wendy P. Gati, Department of Pharmacology, University of Alberta, 9–70 Medical Sciences Building, Edmonton, Alberta, T6G 2H7 Canada. E-mail: wendy.gati@ualberta.ca

Abstract

Summary. 2-chlorodeoxyadenosine (2-CdA) and arabinosylcytosine (araC) are nucleoside drugs that are used to treat various leukaemias, although 2-CdA has not been tested extensively in children with acute lymphoblastic leukaemia (ALL). Nucleoside cytotoxicity depends on the conversion of these agents to 5′-phosphate derivatives, following drug entry into cells via nucleoside transport (NT) processes. This study compared es nucleoside transporter content, determined using a flow cytometric assay with SAENTA [5′-S-(2-aminoethyl)-N6-(4-nitrobenzyl)-5′-thioadenosine] fluorescein, and cytotoxicities of 2-CdA and araC in fresh lymphoblasts from previously untreated paediatric ALL patients and the human T-lymphoblast cell line, CCRF-CEM. Lymphoblast samples from individual patients ranged widely in sensitivity to both 2-CdA (IC50, 6 nmol/l to > 5 μmol/l; mean = 418 nmol/l; n = 8) and araC (IC50, 59 nmol/l to > 5 μmol/l; mean = 1050 nmol/l; n = 7), although IC50 values for the two drugs were correlated (r = 0·78, P = 0·032, n = 7). Cellular es nucleoside transporter content varied more than 35-fold among samples from 10 patients. The correlation between es nucleoside transporter content and drug sensitivity was statistically significant for araC (r = −0·93, P = 0·023, n = 5), but not for 2-CdA (r = −0·57, P = 0·23, n = 6). Exposure of CCRF-CEM cells to araC resulted in a substantial araC concentration-dependent increase in the relative survival of es transporter-deficient cells, whereas the increase was slight following exposure to 2-CdA. We conclude that, in ALL lymphoblasts, es nucleoside transporter content is a determinant of araC sensitivity and that a deficiency in NT may impart resistance to araC.

In the current treatment of acute lymphoblastic leukaemia (ALL), the most common form of childhood cancer, approximately 25% of patients fail to achieve long-term, disease-free survival despite intensive combination chemotherapy (Rivera et al, 1991), an outcome attributed to the emergence of drug-resistant disease. 2-chlorodeoxyadenosine (2-CdA), a purine nucleoside analogue with important antileukaemic activity in both resting and proliferating lymphocytes (Piro, 1992), is currently used in the treatment of a number of indolent lymphoid malignancies (Piro et al, 1988; Kay et al, 1989; Kurzrock et al, 1991). 2-CdA has also shown notable antileukaemic activity in previously untreated paediatric acute myeloid leukaemia (AML) patients (Santana et al, 1991, 1992; Kearns et al, 1994). The success of 2-CdA in treating these malignancies, along with its low clinical toxicity (Carson et al, 1984; Santana et al, 1991, 1992; Kearns et al, 1994), has made 2-CdA a candidate for cytotoxic evaluation in the treatment of paediatric ALL.

Phase I and Phase II studies by Santana et al (1991, 1992) reported some antileukaemic activity of 2-CdA treatment in a small number of paediatric ALL patients, although those studies included patients in their second or third relapse, poor candidates for reinduction with any single agent. In vitro studies of freshly isolated ALL cells by Kumagai et al (1994) showed that sensitivity to 2-CdA appeared to be independent of either a high-risk karyotype (Philadelphia chromosome or 11q23 abnormalities) or a previous relapse history, factors that are usually predictive of an unfavourable response to chemotherapy. Furthermore, potentially cytotoxic concentrations of 2-CdA have been attained in cerebrospinal fluid (CSF) with a well-tolerated, conventional treatment regimen, suggesting a potential role for systemic 2-CdA in the treatment of meningeal leukaemia, a major cause of relapse in this disease (Kearns et al, 1994).

An aim of the present study was to compare the cytotoxic potential of 2-CdA and arabinosylcytosine (araC; cytarabine) towards leukaemic lymphoblasts obtained from the bone marrow of newly diagnosed paediatric patients with ALL, using an in vitro tetrazolium dye reduction assay [3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay] of cell viability. AraC was chosen for this comparison because of its current use in remission induction and consolidation protocols in ALL therapy; araC has also been used in combinations prophylactically, to reduce the possibility of central nervous system (CNS) involvement in childhood ALL (Morra et al, 1993). Pieters and colleagues (Pieters et al, 1994; Klumper et al, 1995) showed that the MTT chemosensitivity assay may be a good predictor of clinical outcome in acute leukaemia of childhood, and a retrospective study in adult AML also demonstrated the prognostic value of the MTT assay (Sargent & Taylor, 1989).

The cytotoxicity of 2-CdA and araC depends upon the formation, inside cells, of 2-CdATP and araCTP respectively (Griffig et al, 1989; Rustum & Raymakers, 1992). Thus, a prerequisite for the antileukaemic activity of the nucleoside drugs is entry into target cells, which is mediated by nucleoside transporters (Belt et al, 1993; Cass, 1995). In mammalian cells, several nucleoside transporter subtypes may be expressed, including equilibrative (facilitated diffusion) and concentrative (Na-dependent) transporters. Both 2-CdA and araC are substrates for the equilibrative nucleoside transport (NT) processes, es and ei (Crawford et al, 1988), which are distinguishable by their differential sensitivity to nitrobenzylthioinosine (NBMPR), a potent inhibitor of the es nucleoside transporter (Paterson & Oliver, 1971). 2-CdA has also been shown to be a substrate for two of five Na-linked, concentrative transporters, namely cif (N1) (King & Cass, 1994) and cs (N5) (Paterson et al, 1993). Of the five concentrative nucleoside transporters expressed in mammalian cells, only cs is sensitive to NBMPR (Paterson et al, 1993).

Although several NT processes have been identified in human leukaemia cells, the es transporter appears to account largely for inward fluxes of nucleosides (Belt et al, 1993; Alessi-Severini et al, 1995). Nevertheless, the cellular content of es nucleoside transporters in leukaemic blasts varies widely among samples from adult patients with acute leukaemia (White et al, 1987; Jamieson et al, 1993; Gati et al, 1997), suggesting that the cellular es transporter content may account in part for differences in clinical response to nucleoside drugs. It has been reported that es nucleoside transporter abundance in human leukaemia cells correlated with both araC influx and levels of araCTP formed (Wiley et al, 1983; Young et al, 1985). This laboratory has reported a correlation between es nucleoside transporter abundance and sensitivity to araC in fresh leukaemia cells from adult patients with AML and ALL (Gati et al, 1996, 1997, 1998). Thus, we hypothesized that the es-mediated, inward flux of 2-CdA and araC in lymphoblasts from paediatric ALL patients may be a determinant of cytotoxicity and that es nucleoside transporter deficiency may be a factor in the emergence of resistance to these agents.

The present study examined fresh lymphoblasts from paediatric patients with ALL and a human T-lymphoblast cell line to test relationships between the cellular content of es nucleoside transporters and cell sensitivity to 2-CdA and araC. The study used SAENTA [5′-S-(2-aminoethyl)-N6-(4-nitrobenzyl)-5′-thioadenosine] fluorescein, a fluorescent ligand for the es nucleoside transporter, in the flow cytometric measurement of transporter abundance in cells (Jamieson et al, 1993; Buolamwini et al, 1994; Gati et al, 1996). This study also reports, for the first time, the es nucleoside transporter content of fresh lymphoblasts from children with ALL.

Materials and methods

Cell preparation.  Fresh leukaemic cells were obtained, with informed consent, from 11 bone marrow aspirates and one peripheral blood sample from paediatric ALL patients. All patients were newly diagnosed and immunophenotyping was performed at the Department of Laboratory Medicine, University of Alberta Hospital, Edmonton, AB, Canada. Leukaemic blasts were isolated using a conventional, density-gradient procedure in which the heparinized bone marrow or blood sample was diluted 1:1 with Roswell Park Memorial Institute (RPMI)-1640 medium (Sigma, St. Louis, MO, USA) containing 2 mmol/l Hepes buffer and 10% fetal bovine serum (RH10F). The cell suspension was layered over Ficoll-Paque and centrifuged (30 min, 400 g). Cells from the Ficoll-Paque interface were washed twice (10 min, 150 g) with RH10F medium and resuspended (a) in RH10F medium at a concentration of 1 × 107 cells/ml for the MTT cytotoxicity assay, and (b) in phosphate-buffered saline (PBS) of the following composition: 137 mmol/l NaCl, 2·7 mmol/l KCl, 8·1 mmol/l Na2HPO4 and 1·1 mmol/l KH2PO4 (pH 7·4) at a concentration of 6 × 106 cells/ml for the flow cytometric SAENTA fluorescein-binding assay.

CCRF-CEM human T-lymphoblasts (hereafter, CEM cells) were maintained in logarithmic growth in RPMI-1640 medium containing 10% fetal bovine serum in a humidified atmosphere of 5% CO2 in air at 37°C.

Chemicals. 2-CdA was prepared (Kazimierczuk et al, 1984) by Dr J. S. Wilson, Department of Medical Microbiology and Immunology, University of Alberta, Canada. AraC and MTT were obtained from Sigma Chemical Co., and NBMPR (Paul et al, 1975) and 5-(SAENTA-×8)-fluorescein (Buolamwini et al, 1994) were provided by Alberta Nucleoside Therapeutics, Department of Pharmacology, University of Alberta. Phycoerythrin-labelled annexin V (annexin V-PE) was from Pharmingen Canada (Mississauga, ON, Canada) and 7-aminoactinomycin D (7-AAD) was from Molecular Probes, (Eugene, OR, USA). Ficoll-Paque was purchased from Pharmacia Biotech (Baie d'Urfé, PQ, Canada). Cell culture materials were from Gibco BRL (Mississauga, ON, Canada).

MTT assay of drug sensitivity.  The in vitro sensitivity of fresh leukaemic lymphoblasts to 2-CdA and araC was determined using the MTT dye reduction assay (Alley et al, 1988). Briefly, 100-μl portions of cell suspensions (1 × 106 cells) were added to 100 μl RH10F medium in 96-well microculture plates containing graded concentrations of either 2-CdA or araC (1–5000 nmol/l); between three and six replicate microcultures were prepared for each drug concentration. Control assays lacked only drugs. Plates were incubated in a humidified atmosphere of 5% CO2 in air at 37°C for 72 h. After incubation, 30 μl of MTT solution prepared in RH10F (1 mg/ml) was added to each well and plates were incubated for a further 2 h. The plates were then centrifuged (7 min, 50 g) and the supernatant fluid aspirated from each well. Following this, the crystalline formazan product was dissolved by the addition to each well of 150 μl dimethylsulphoxide (spectrophotometric grade, Aldrich, Milwaukee, WI, USA) and the absorbance of the well solutions was measured in a microplate reader (Titertek Multiskan Plus, ICN Biomedicals, Mississauga, ON, Canada) at 540 nm (A540). The A540 of each microwell solution was proportional to cell concentration. Cell viabilities in the test wells were expressed as percentages of those in control wells, which lacked the cytotoxic test agent.

The cytotoxic effects of the nucleoside drugs in fresh lymphoblasts were quantified by determining the concentration of antileukaemic agent producing a half-maximal response in the MTT assay. A computer program (PRISM, Version 1·0, GraphPad Software Inc., San Diego, CA, USA) was used to fit the following logistic equation (1) to plots of cell viability (% of control) versus the log of the drug concentration:

image(1)

in which Y is the response (cell viability,% of control), A1 and A2 are the upper and lower limits for Y, X is the drug concentration, IC50 is the concentration of drug producing a half-maximal response, and b is a constant related to the steepness of the plot.

Flow cytometric enumeration of es nucleoside transporter sites on primary leukaemic blasts.  Freshly isolated leukaemic lymphoblasts from patients were washed twice (5 min, 150 g) and resuspended at 3 × 106 cells/ml in PBS for determination of the total binding of SAENTA fluorescein, or in PBS containing 5 μmol/l NBMPR for measuring non-specific binding of SAENTA fluorescein. Prior treatment of cells with NBMPR under these conditions (30 min, 22°C) prevented the site-specific binding of SAENTA fluorescein to the es transporter sites. To determine total binding of SAENTA fluorescein, cells suspended in PBS were incubated with graded concentrations (0·1–20 nmol/l) of SAENTA fluorescein for at least 15 min at 22°C before flow cytometric analysis of cell-associated fluorescence. Non-specific binding was determined by incubating NBMPR-treated cells with the graded concentrations of SAENTA fluorescein in medium containing 5 μmol/l NBMPR. Cell suspensions were protected from light during these incubations. The autofluorescence of leukaemic cells was measured in PBS in the presence and absence of 5 μmol/l NBMPR.

Cell-associated fluorescence and light scatter signals were analysed using a FACScan flow cytometer [Becton Dickinson Immunocytometry Systems (BDIS), San Jose, CA, USA]. SAENTA fluorescein molecules that were cell associated were excited with 488 nm light from an argon laser and the emitted fluorescence was collected at 530 nm (FL1). Fluorescence signals from 10 000 cells for each condition were acquired and analysed using BDIS Lysis II software. Analysis of SAENTA fluorescein fluorescence was restricted to signals that fell within an electronic gate encompassing the blast/lymphocyte region of the light scatter dot plot (Sutherland et al, 1994). The mean relative fluorescence intensity (RFI) obtained from histograms of SAENTA fluorescein fluorescence was used as a measure of cell-associated SAENTA fluorescein, and the es transporter site-bound content of SAENTA fluorescein was calculated for each concentration of the fluorescent ligand as the difference between the RFI values for total SAENTA fluorescein binding and non-specific SAENTA fluorescein binding. The latter (non-specific binding) values were determined from plots of SAENTA fluorescein fluorescence versus concentration of SAENTA fluorescein measured in the presence of 5 μmol/l NBMPR.

The equilibrium binding constants, KD and Bmax for SAENTA fluorescein were determined by non-linear regression analysis (Microcal ORIGIN, Version 4·1) of the equilibrium binding data by using the following equation (2):

image(2)

in which Y is the specific binding (RFI) of SAENTA fluorescein, X is the total concentration of SAENTA fluorescein, Bmax is the maximum specifically bound SAENTA fluorescein (RFI) per cell, and KD is the equilibrium dissociation constant for SAENTA fluorescein at es nucleoside transporter sites.

Flow cytometric assay of drug sensitivity. In some experiments, nucleoside cytotoxicity was determined by flow cytometric analysis of CEM T-lymphoblast populations. Cell suspensions (7·5 × 104 cells in 1·5 ml RH10F) were incubated in 24-well microculture plates for 96 h in the absence (controls) or presence of graded concentrations of 2-CdA or araC. The microcultures were centrifuged, washed in HSC buffer (140 mmol/l NaCl, 2·5 mmol/l CaCl2 and 10 mmol/l Hepes, pH 7·4) and stained with 20 nmol/l SAENTA fluorescein (or 20 nmol/l SAENTA fluorescein in the presence of 5 μmol/l NBMPR for non-specific binding), a ligand concentration that yields an estimate of the Bmax value and, thus, a measure of the cellular content of es nucleoside transporters. Cells were counterstained with annexin V-PE and 7-AAD, to identify apoptotic and necrotic cells respectively (Schmid et al, 1992; Koopman et al, 1994). Cell-associated fluorescence signals were recorded using a FACScan flow cytometer, with excitation at 488 nm and simultaneous collection of fluorescence signals emitted from SAENTA fluorescein, annexin V-PE and 7-AAD. CellQuest software was used to enumerate viable cells in populations that survived each concentration of 2-CdA or araC, by setting a gate that excluded apoptotic (annexin V-PE-positive) and necrotic (7-AAD-positive) cells, and to simultaneously determine the cellular content of es nucleoside transporters (specific binding of SAENTA fluorescein) in the electronically isolated, viable cell subsets.

Statistical analysis.  Wilcoxon's signed rank test was used to determine levels of significance, using P < 0·05 to indicate a significant difference between groups.

Results

Cytotoxicity of 2-CdA and araC in fresh leukaemic lymphoblasts

The MTT dye reduction assay was used to determine the sensitivity of freshly isolated ALL cells to 2-CdA and araC. A linear relationship was demonstrated between cell number per well and the reduced dye product (formazan) formed, measured colorimetrically as A540 (data not shown). That linearity occurred over a range of cell concentrations (104−106 cells per well) in four patient samples.

Figure 1 and Table I illustrate the variability in response to both agents among patient samples. The IC50 values for araC toxicity towards the preB-cell ALL lymphoblasts from Patients 4 and 8 were 372 nmol/l and 374 nmol/l, respectively, and, although several-fold greater than those for 2-CdA in the same experiment, these values are within a concentration range (60–800 nmol/l) that is achievable in plasma with conventional araC chemotherapy (Rustum & Raymakers, 1992). However, cells from preB-cell ALL Patient 6 were less sensitive to both 2-CdA and araC (IC50 values of 1660 nmol/l and 2650 nmol/l respectively, Fig 1). Similarly, resistance to both 2-CdA and araC (Patient 9) was shown in lymphoblasts from T-cell ALL (Table I).

Figure 1.

Sensitivity of leukaemic lymphoblasts from acute lymphoblastic leukaemia (ALL) Patient 5 (T-cell, top) and Patient 6 (pre-B-cell, bottom) to 2-chlorodeoxyadenosine (2-CdA) (▿) and arabinosylcytosine (araC) (●). Cells were exposed to graded concentrations of each drug for 72 h. Cell viability was measured using the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay at the end of the incubation period. Logistic curves were fitted to the data to obtain IC50 values from the concentration-effect relationships, as described in Materials and methods. Each data point represents the mean ± SD of six replicate measurements in a single experiment.

Table I.  Immunophenotype and in vitro sensitivity to 2-CdA and araC *of leukaemic lymphoblasts from children with ALL.
Patient numberImmunophenotypeBone marrow blasts (%)IC50, 2-CdA (nmol/l)IC50, araC (nmol/l)
  • *IC50 values were determined from concentration-effect plots following 72-h exposure of lymphoblasts to the nucleoside analogues. The MTT dye reduction assay was used to measure cell viability in the drug-treated cultures.

  • †Not determined. All samples used in the present study are listed in the table. Although some samples were not tested for drug sensitivity, SAENTA fluorescein binding data for those samples are shown in Fig 4.

  • Peripheral blood sample.

1preB-cell937992,290
2preB-cell99ndnd
3T-cellnd408nd
4preB-cell9579372
5T-cell77665
6preB-cell9516602,650
7preB-cellndndnd
8preB-cell9584374
9T-cell97> 5000> 5000
10preB-cellndndnd
11preB-cell771711,540
12T-cellnd13659

Table I shows that interpatient variation in sensitivity to the nucleoside analogues was considerable, with IC50 values for 2-CdA ranging from 6 nmol/l to > 5 μmol/l, and for araC from 59 nmol/l to > 5 μmol/l. The leukaemic cells showed significantly (P < 0·05) greater sensitivity to 2-CdA (n = 8) than to araC (n = 7). In the present study, IC50 values for 2-CdA were < 100 nmol/l in one-third of the patient samples tested; concentrations of 2-CdA achievable in the plasma were in the 23–85 nmol/l range (Kearns et al, 1994).

Figure 2 tested the correlation between IC50 values for 2-CdA and araC cytotoxicity in leukaemic lymphoblasts samples from seven paediatric patients with ALL in which both parameters were determined (Table I), yielding a statistically significant relationship (r = 0·78, P = 0·032).

Figure 2.

Correlation of sensitivity to 2-chlorodeoxyadenosine (2-CdA) and to arabinosylcytosine (araC) in leukaemic lymphoblasts from acute lymphoblastic leukaemia (ALL) patients. The coordinates of each data point are logarithms of the IC50 values for 2-CdA and araC in individual patient samples, determined according to the methods of Fig 1. A statistically significant (P = 0·032) correlation of sensitivity to the two nucleoside drugs was found in the lymphoblast samples tested.

Cellular content of es nucleoside transporters in fresh leukaemic lymphoblasts

The data of Fig 3 illustrate the equilibrium binding of graded concentrations of SAENTA fluorescein to cells from preB-cell ALL Patient 8 and T-cell ALL Patient 9, as determined using flow cytometry. The RFI value obtained from each fluorescence histogram (RFI versus frequency of events) was a quantitative measure of bound SAENTA fluorescein. Specific binding was calculated as the difference between SAENTA fluorescein binding in the absence (total binding) and presence (non-specific binding) of an excess of NBMPR. These experiments showed that SAENTA fluorescein binding was saturable and es transporter-selective in that bound SAENTA fluorescein was reduced by prior treatment of cells with NBMPR. Non-specific binding of 20 nmol/l SAENTA fluorescein was 8·1% and 22% of the total binding in cells from Patients 8 and 9, respectively, after correction for autofluorescence. Mass law analysis of the binding data for lymphoblasts from Patient 8 yielded the constants Bmax = 57·8 ± 1·6 RFI units/cell and KD = 2·06 ± 0·24 nmol/l. The binding constants for Patient 9 were Bmax = 11·4 ± 0·4 RFI units/cell and KD = 2·85 ± 0·36 nmol/l. Conversion of RFI values to es nucleoside transporter sites (NBMPR binding sites) using a factor (200) determined in human cell lines (Gati et al, 1997) would yield Bmax = 11 600 ± 320 sites/cell and Bmax = 2280 ± 80 sites/cell for lymphoblasts from Patients 8 and 9 respectively.

Figure 3.

Equilibrium binding of SAENTA [5′-S-(2-aminoethyl)- N6-(4-nitrobenzyl)-5′-thioadenosine] fluorescein to leukaemic lymphoblasts from preB-cell acute lymphoblastic leukaemia (ALL) Patient 8 (top) and T-cell ALL Patient 9 (bottom). Cells were incubated with graded concentrations of SAENTA fluorescein for 30 min at 22°C in the absence (▪) or presence (▵) of 5 µmol/l nitrobenzylthioinosine (NBMPR). Cell-associated fluorescence was determined by measuring the mean fluorescence intensity of 10 000 cells for each data point. Specific binding (○) was determined as the difference between total (▪) and non-specific (▵) fluorescence. Hyperbolas were fitted to the specific binding data to yield Bmax and KD values, as described in Materials and methods. The constants so determined were: Patient 8, Bmax = 57·8 ± 1·6 RFI, KD = 2·06 ± 0·24 nmol/l; Patient 9, Bmax = 11·4 ± 0·4 RFI, KD = 2·85 ± 0·36 nmol/l; The inset figures show Scatchard analyses of the specific binding data.

Marked patient to patient differences were apparent in the maximum binding of SAENTA fluorescein in leukaemic lymphoblasts from 10 paediatric patients with ALL, as shown in Fig 4. The Bmax values ranged from 4·00 to 147 RFI units/cell (800–29 400 es transporter sites/cell). The KD values for SAENTA fluorescein binding to the es transporter ranged from 0·663 to 11·1 nmol/l (not shown), indicating that SAENTA fluorescein was bound with high affinity to the es transporter in leukaemic lymphoblasts.

Figure 4.

Interpatient variation in the cellular content of es nucleoside transporters in leukaemic lymphoblasts from children with acute lymphoblastic leukaemia (ALL). Bmax values [in relative fluorescence intensity (RFI) units] were determined by measuring the equilibrium binding of SAENTA [5′-S-(2-aminoethyl)- N6-(4-nitrobenzyl)-5′-thioadenosine] fluorescein, as described for the experiments of Fig 3. Open bars, preB-cell ALL; hatched bars, T-cell ALL.

Correlation of chemosensitivity and cellular content of es nucleoside transporters

The present study investigated whether the abundance of es nucleoside transporters in fresh leukaemic lymphoblasts was a determinant of 2-CdA or araC cytotoxicity. Figure 5 shows tests of the correlation of 2-CdA and araC sensitivity with es nucleoside transporter content of the blasts. The correlation between 2-CdA sensitivity and es transporter abundance was weak (Fig 5, top), and did not reach statistical significance (r = −0·57, P = 0·23). In contrast, a strong, statistically significant correlation between araC sensitivity and es transporter abundance was found (Fig 5, bottom), that is, the IC50 value for araC was inversely correlated with the Bmax value for SAENTA fluorescein binding (r = −0·93, P = 0·023).

Figure 5.

Correlation of 2-chlorodeoxyadenosine (2-CdA) (top) and arabinosylcytosine (araC) (bottom) sensitivity with es nucleoside transporter content of lymphoblasts from paediatric acute lymphoblastic leukaemia (ALL) patients. Es transporter content is expressed as the Bmax (in RFI units) for SAENTA [5′-S-(2-aminoethyl)-N6-(4-nitrobenzyl)-5′-thioadenosine] fluorescein binding, determined as in the experiments of Fig 3. IC50 values are the concentrations of nucleoside analogues that reduced cell viability by 50%, determined by the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay as in the experiments of Fig 1.

Proportional increases in es nucleoside transporter-deficient CEM cell subsets in cultures exposed to araC, but not in those exposed to 2-CdA

Replicate samples of CEM cells were exposed to graded concentrations of 2-CdA or araC for 96 h, then stained with SAENTA fluorescein, annexin V-PE and 7-AAD for the measurement of es nucleoside transporter content, and the recognition of apoptotic and necrotic cells, respectively, as described in Materials and methods. IC50 values for 2-CdA and araC under these conditions were 28 nmol/l and 4·3 nmol/l respectively (not shown).

Figure 6 shows examples of flow cytometry histograms of SAENTA fluorescein-stained control and drug-treated CEM cell populations that contained only viable cells (non-apoptotic and non-necrotic according to annexin V-PE and 7-AAD staining). In these experiments, cells that were stained with SAENTA fluorescein, or SAENTA fluorescein in the presence of 5 μmol/l NBMPR, yielded histograms that specified the fluorescence intensity values of total binding (solid line) or non-specific binding (overlay, dotted line, marker M1) respectively (Fig 6). Fluorescence intensity values for specific binding were recognized (marker M2) as those of the total binding histogram that did not fall within the range of non-specific binding fluorescence signals (marker M1). In control samples (no drug, Fig 6A), a small number of cells (5·8%) were dimly fluorescent after staining for total SAENTA fluorescein binding (solid line), emitting signals that fell within the range of marker M1 (non-specific binding). As these cells lacked measurable specific binding of SAENTA fluorescein, they were regarded as ‘es transporter-deficient’ cells. Similarly, Fig 6B shows a SAENTA fluorescein histogram of CEM cells that survived 40 nmol/l 2-CdA, a concentration that killed 64% of the original cell population. Under these conditions, the transporter-deficient subset (7·8%) was only slightly greater than that of the control sample in Fig 6A. In contrast, cells that survived 10 nmol/l araC (Fig 6C), a concentration that killed 97% of the cells, included a large (40%) proportion of transporter-deficient cells. These results suggested that cells with a high es nucleoside transporter content were selectively eliminated by araC.

Figure 6.

Increases in es nucleoside transporter-deficient cell subsets in arabinosylcytosine (araC)-treated CEM cultures, but not in those treated with 2-chlorodeoxyadenosine (2-CdA). CEM cells were exposed to graded concentrations of the nucleoside drugs for 96 h, then stained with SAENTA [5′-S-(2-aminoethyl)-N6-(4-nitrobenzyl)-5′-thioadenosine] fluorescein in the absence (total binding) or presence of nitrobenzylthioinosine (NBMPR) (non-specific binding), and with phycoerythrin-labelled annexin V (annexin V-PE) and 7-aminoactinomycin D (7-AAD). The samples were analysed by flow cytometry as described in Materials and methods. Representative samples from four replicate data sets are shown. (A) SAENTA fluorescein binding in control cells (no drug), showing fluorescence intensity ranges for non-specific binding (overlay: dotted line, marker M1) and total binding (solid line). Marker M2 (specific binding) designates fluorescence signals emitted from the total binding sample that exceeded values identified as non-specific binding. Signals from the total binding sample that fell within the range of marker M1 were calculated as a percentage of total events (M1 + M2) to determine the proportion of cells that did not bind SAENTA fluorescein specifically and were therefore regarded as es transporter-deficient (control value shown, 5·8%). (B) Histogram of total SAENTA fluorescein binding in cells surviving exposure to 40 nmol/l 2-CdA, indicating that 7·8% of cells were transporter-deficient. (C) Histogram of total SAENTA fluorescein binding in cells surviving exposure to 10 nmol/l araC, indicating that 40% of cells were transporter-deficient.

CEM cells that were exposed to graded concentrations of 2-CdA or araC, and analysed according to the procedure of Fig 6, yielded the data shown in Fig 7. The increase in es transporter-deficient CEM cell subsets in araC-surviving cell populations was dependent on araC concentration (Fig 7A) and on araC cytotoxicity (Fig 7B). In contrast, the proportion of transporter-deficient cells in CEM cultures exposed to 2-CdA changed only slightly with graded increases in 2-CdA concentration or the extent of cell kill (Fig 7). In the experiments with 2-CdA, a 100-nmol/l concentration of the nucleoside was also tested, but the number of surviving cells was insufficient to allow a precise analysis of flow cytometry data (not shown).

Figure 7.

The proportion of es transporter-deficient cells in CEM cell populations after exposure to 2-chlorodeoxyadenosine (2-CdA) (□) or arabinosylcytosine (araC) (●). CEM cells were exposed to graded concentrations of the nucleoside drugs for 96 h, stained with SAENTA [5′-S-(2-aminoethyl)-N6-(4-nitrobenzyl)-5′-thioadenosine] fluorescein, phycoerythrin-labelled annexin V (annexin V-PE) and 7-aminoactinomycin D (7-AAD), and analysed by flow cytometry as described in Materials and methods. The proportions of es transporter-deficient cells were determined as described in the examples of Fig 6. (A) Nucleoside concentration-dependence of changes in the proportion of es transporter-deficient cells. (B) Changes in the proportion of es transporter-deficient cells as a function of nucleoside cytotoxicity. The data are means ± SEM (n = 4).

Discussion

The present study examined the role of nucleoside transport in the cytotoxicity of 2-CdA and araC in leukaemic lymphoblasts from previously untreated paediatric patients with ALL, by comparing the cellular es nucleoside transporter content and nucleoside drug sensitivity of freshly isolated lymphoblasts.

A wide interpatient variation in sensitivity of the leukaemic blast samples to 2-CdA and araC was reflected in IC50 values that ranged more than 800-fold and 80-fold respectively. Although the leukaemic cell sample that was most resistant to 2-CdA and araC was from a patient with T-cell ALL, an immunophenotype often associated with an unfavourable outcome (Mahony, 1994), two T-cell ALL samples in this study were also among those that were most sensitive to the nucleosides. Pieters et al (1994) have shown a wide variation in sensitivity to a number of antitumour agents in leukaemic blasts from patients with ALL. Campana et al (1993) also observed considerable interpatient variation in lymphoblast sensitivity to vincristine, 6-thioguanine, araC and teniposide, using a novel stroma-supported immunocytometric assay (SIA) to evaluate blasts from patients with ALL. In contrast to results reported here, in vitro studies by Kumagai et al (1994) showed little interpatient variation in cytotoxicity of 100 nmol/l 2-CdA to the 20 samples of paediatric ALL lymphoblasts examined. A possible explanation for the difference may be found in the design of the study by Kumagai et al (1994), in which the SIA assay was used to assess cell viability after exposure of cells to 2-CdA for periods of 96–120 h. Those results also showed that 2-CdA was effective in cells from patients who had relapsed after initial treatment with multiagent chemotherapy and in cells from patients who had an unfavourable karyotype.

In the present study, IC50 values for 2-CdA were lower than those for araC in most lymphoblast samples, and a statistically significant linear correlation between IC50 values for 2-CdA and araC was observed. The latter finding may indicate a common cytotoxic mechanism or resistance mechanism for these nucleoside drugs in ALL blasts. Although the leukaemic lymphoblasts were generally more sensitive to 2-CdA than to araC, 2-CdA may have little advantage over araC in the treatment of ALL of childhood, as higher plasma concentrations of araC are obtained clinically.

Previous studies have shown that the cellular content of es nucleoside transporters was a determinant of araC influx (Wiley et al, 1982), araCTP formation (Wiley et al, 1985) and araC cytotoxicity (Gati et al, 1996, 1997, 1998) in freshly isolated blast samples from adult patients with acute leukaemia. Furthermore, es nucleoside transporters have been implicated in the cellular efflux of 2-CdA in cultured lines of leukaemic lymphoblasts, enabling manipulation of 2-CdA cytotoxicity in cells exposed sequentially to 2-CdA and NBMPR (Wright et al, 2000). The present study compared the cell sensitivity to 2-CdA and araC with the es transporter content of lymphoblasts from paediatric patients with ALL. It was hypothesized that es transporter-mediated inward flux of 2-CdA and araC may be a requisite of nucleoside cytotoxicity in the leukaemic lymphoblasts, and that a low es transporter abundance might be a predictor of resistance to these agents. A rapid and sensitive flow cytometric assay, using the es transporter-specific ligand, SAENTA fluorescein, was used to determine the cellular content of es transporters in the ALL blasts. Bmax values for SAENTA fluorescein binding ranged > 35-fold among samples from 10 ALL patients, reflecting a wide variation in cellular es transporter abundance.

A statistically significant correlation between sensitivity to araC and es nucleoside transporter abundance was found in this study, suggesting that the es transporter is a determinant of araC cytotoxicity in ALL blasts. Accordingly, lymphoblast samples that were relatively deficient in es transporters were resistant to araC, and araC-sensitive samples with low es transporter abundance were not found. Thus, the data (Fig 5, bottom) suggest that the IC50 value for araC would be > 1 μmol/l in ALL lymphoblasts in which the Bmax value for SAENTA fluorescein binding (es nucleoside transporter content) is less than about 10 RFI, and that such cells are likely to be resistant to araC at concentrations of the drug achieved in plasma during therapy at conventional doses. In contrast, the corresponding correlation with 2-CdA was not statistically significant, suggesting that, in fresh leukaemic lymphoblasts, nucleoside transport is a less important factor in the cytotoxicity of 2-CdA than in the cytotoxicity of araC. These ideas are consistent with our findings in the CEM lymphoblast cell line, in which nucleoside transporter expression is limited to the es transporter (Belt et al, 1993). The present experiments showed a substantial, araC concentration-dependent increase in the proportion of es transporter-deficient cells within CEM cell populations that survived exposure to araC. Those results suggest that nucleoside transport-proficient CEM cells were selectively killed by araC, and that transporter-deficient cells were araC-resistant. In contrast, only a slight increase in the proportion of transporter-deficient cells was found in cell populations surviving exposure to 2-CdA, indicating that es nucleoside transporter abundance was apparently a less important determinant of sensitivity to 2-CdA in CEM cells.

Although the present experiments support the hypothesis that es nucleoside transporter deficiency may impart resistance to araC in fresh ALL blasts, these studies suggest that mechanisms unrelated to the es nucleoside transporter probably account for 2-CdA resistance in lymphoblast samples from some patients. Correlations of es transporter content and sensitivity to nucleoside drugs may be confounded by the expression, in some cells, of other nucleoside transporters, such as the cs, cif or ei nucleoside transporters that mediate the uptake of 2-CdA, but are not recognized by SAENTA fluorescein. Because of the limited size of lymphoblast populations obtained from bone marrow samples in the present study, no attempt was made to detect other nucleoside transporters in these cells, or to measure other mechanisms of 2-CdA resistance, such as a deficiency in deoxycytidine kinase (Kearns et al, 1994). A comparison of the es nucleoside transporter content of leukaemic lymphoblasts and clinical response to therapy with nucleoside drugs in the controlled setting of a clinical trial will be needed to determine the clinical significance of nucleoside transport deficiency in resistance to 2-CdA and araC in ALL of childhood.

Acknowledgments

We are indebted to the paediatric healthcare staff at the University of Alberta Hospital and to John Chan for help in the provision of bone marrow samples. We thank Carmen Harris for technical assistance. This study was supported by the Alberta Cancer Board, the Alberta Heritage Foundation for Medical Research and the Peter Jang Memorial Fund. A.M.P.W. was the recipient of a Studentship Award from the Alberta Heritage Foundation for Medical Research during the conduct of this study.

Ancillary