Abstract
Background/Aim: An active metabolite of the anti-leukemia agent clofarabine (Cl-F-ara-A) is an intracellular triphosphate form, Cl-F-ara-ATP. Monitoring this active form could provide crucial information for optimizing treatment regimens based on Cl-F-ara-A. A simple, isocratic HPLC method was developed. Materials and Methods: Samples (500 μl) from leukemic cells were loaded onto an anion-exchange column and eluted with a phosphate-acetonitrile buffer (flow rate: 0.7 ml/min) at ambient temperature. The Cl-F-ara-ATP concentration was determined by measuring absorbance at 254 nm. Results: The standard curve was linear, with minimal within-day and inter-day variability. Recovery was excellent; low and high quantitation limits were 10 pmol and 5,000 pmol, respectively. Cl-F-ara-ATP eluted independently of ATP, GTP, UTP, and CTP. Production of Cl-F-ara-ATP was successfully measured in cultured leukemia HL-60 cells treated in vitro with Cl-F-ara-A. Conclusion: This method is simple, sensitive and applicable for determination of the Cl-F-ara-ATP content of biological materials.
Clofarabine, 2-chloro-2’-arabinofluoro-2’-deoxyadenosine (Cl-F-ara-A), is a second-generation purine nucleoside analog, notable for having halogenated purine and ribose rings (1). In 2004, Cl-F-ara-A received accelerated approval from the US Food and Drug Administration for the treatment of pediatric patients with refractory acute lymphoblastic leukemia (2-4). The agent is currently undegoing several clinical trials targeting other malignancies, including acute myeloid leukemia.
Cl-F-ara-A enters leukemia cells by way of several transporters, including equilibrative nucleoside transporter 1 (ENT1) and concentrative nucleoside transporter 3 (CNT3). Inside the cell, Cl-F-ara-A is phosphorylated to Cl-F-ara-A monophosphate by cytosolic deoxycytidine kinase (dCK) and mitochondrial deoxyguanosine kinase (dGK). The monophosphate is further phosphorylated to Cl-F-ara-A triphosphate (Cl-F-ara-ATP), which is the active intracellular metabolite (1). Cl-F-ara-ATP competes potently with dATP for binding to DNA polymerase-α and -ε (5). In addition, incorporation of Cl-F-ara-ATP into DNA in monophosphate form impairs DNA synthesis by causing chain termination and strand breaks (6). Furthermore, Cl-F-ara-ATP potently inhibits ribonucleotide reductase activity (5). Cl-F-ara-ATP is therefore a metabolite critical to the cytotoxic activity of Cl-F-ara-A.
Clinically, the most useful parameter for predicting the efficacy of a nucleoside analog is not its plasma concentration but rather the level of its triphosphate form in leukemia cells, as previously demonstrated for cytarabine (7-10). Previous reports have suggested a correlation between clinical efficacy of Cl-F-ara-A and intracellular Cl-F-ara-ATP concentration (11). Thus, pharmacokinetic evaluation of intracellular Cl-F-ara-ATP may provide crucial information when attempting to optimize treatment with Cl-F-ara-A.
The most widely used assay for monitoring Cl-F-ara-ATP levels is the gradient elution ion-exchange HPLC method (6, 12, 13). This method was shown to be sensitive and applicable for determining the clinical pharmacology of Cl-F-ara-ATP. In general, however, use of a gradient mode requires a complicated computer-based system to control differing concentrations of elution buffers, a fact that effectively limits its use to hospitals with adequately equipped laboratories. In addition, a gradient elution sometimes results in base-line drift, thereby leading to unexpected errors when measuring peak heights or areas resulting from application of crude, biological materials.
Here we present an isocratic HPLC method for measuring Cl-F-ara-ATP in leukemic cells. The isocratic mode is simple and inexpensive to perform, as it can be run using a single, mechanical pump. This method is also sensitive and applicable to biological samples.
Materials and Methods
Chemicals. ATP, CTP, UTP, and GTP were purchased from Sigma Chemical Co. (St. Louis, MO, USA). Cl-F-ara-A and Cl-F-ara-ATP were prepared by Genzyme Japan (Tokyo, Japan). Water was purified in our laboratories by a combination of reverse osmosis and ion exchange. All other chemicals used were of analytical grade.
Preparation of standard Cl-F-ara-ATP in aqueous solution. Cl-F-ara-ATP was dissolved in water to make a stock solution having a final concentration 10 mM. Serially diluted standards were prepared from this solution for generating a standard curve, and were then stored at 4°C until used during sample analysis.
HPLC apparatus and chromatographic conditions. The HPLC system consisted of a pump, (CCPM-II), an autosampler (AS-8020), an anion-exchange column (length x inside diameter, 250 × 4.6 mm) packed with TSK DEAE-2SW gel, (particle size 5 μl) (Tosoh Corp.), an in-line degasser (SD-8022), and a variable-wavelength detector (UV-8020). The HPLC was controlled and monitored using a personal computer equipped with LC-8020 software. All components of the HPLC system were obtained from Tosoh Corp., Tokyo, Japan. Elution was isocratic, using 0.06 M Na2HPO4 (pH 6.9) with 20% acetonitrile at a constant flow rate of 0.7 ml/min and at ambient temperature. The eluate was monitored at 254 nm for detection of Cl-F-ara-ATP (14).
Linearity, precision and accuracy. To validate the method, seven different concentrations of the standard Cl-F-ara-ATP stock solution were made by serial dilution in purified water. Then 500-μl aliquots of each diluted standard, containing amounts of Cl-F-ara-ATP ranging from 10 pmol to 5,000 pmol, were loaded onto the HPLC column. Measurements for this standard curve were made in triplicate on three separate days. Data were combined and plotted to determine linear correlation between amounts of Cl-F-ara-ATP and corresponding peak areas calculated automatically by software. The standard curve was fitted by the weighed least-squares linear regression analysis method using the equation y=ax, where a represents the slope that was forced through 0. To assess precision, within-day and inter-day variation was determined as the percentage coefficient of variation (CV) values. To determine accuracy, triplicate samples of 3 different solutions having known concentrations of Cl-F-ara-ATP (10, 100, 1,000 pmol) were applied to the HPLC column, and resulting peak areas were plotted on the standard curve to verify amounts applied.
Cell culture and treatment. To validate the applicability of this method to biological materials, human leukemia HL-60 cells were grown in RPMI-1640 culture medium (Sigma Chemical Co.) supplemented with 10% heat-inactivated fetal bovine serum (Sigma Chemical Co.) in a 5% CO2 humidified atmosphere at 37°C. Cells in logarithmic growth phase (1×106/ml, 20 ml) were incubated with Cl-F-ara-A at different concentrations for 0.5, 1, 2, and 6 h at 37°C. Treated and untreated cells were then washed twice with fresh medium. Cells were centrifuged (400 × g, 10 min, 4°C) in micro test tubes and pellets were resuspended in 200 μl 0.3 M cold perchloric acid. Mixtures were vortexed for 10 s, and then allowed to stand for 15 min at 4°C. Acidic supernatants were isolated by centrifugation of samples (15,600 × g, 20 s, 4°C), and then were mixed with 100 μl of 0.5 N potassium hydroxide for neutralization. After another centrifugation (15,600 × g, 20 s, 4°C), neutralized supernatants (containing pooled intracellular nucleotides) were collected (15). The sample loop space for the HPLC was 500 μl, and the dead volume for aspiration of the sample by the autosampler was 170 μl. The volume of each sample was therefore adjusted to 700 μl by the addition of water, and a 500 μl aliquot was applied to the column. Intracellular concentrations of Cl-F-ara-ATP were expressed as pmol/107 cells.
Cl-F-ara-ATP standard curve. Aqueous solutions of Cl-F-ara-ATP standards (10, 20, 50, 100, 500, 1,000, and 5,000 pmol) were run in triplicate on three separate days. A linear relationship was obtained between amounts of Cl-F-ara-ATP and the corresponding peak areas. The parameters for the linear regression line were: y=0.78x (r2=0.999, p<0.0001).
Calculations and statistical analyses. The standard curve was generated using GraphPad Prism software (GraphPad Software, Inc., San Diego, CA, USA). The % CV values and standard deviations for Cl-F-ara-ATP measurements were calculated using Microsoft Excel software (Microsoft Corporation, Redmond, WA, USA).
Results
Validation. To generate the standard curve, seven different concentrations of Cl-F-ara-ATP were applied in triplicate to the HPLC column on three separate days. As shown in Figure 1, using all the data sets, a linear relationship was demonstrated between the amounts of Cl-F-ara-ATP applied (10-5,000 pmol) and the corresponding peak areas (r2=0.999, p<0.00001). For each concentration, within-day and inter-day variability was minimal; all % CV values were <10% (Table I). Accuracy assessment results are shown in Table II. Equality was demonstrated between each of 3 different concentrations of Cl-F-ara-ATP applied to the HPLC and the corresponding determinant extrapolated from the standard curve, indicating recovery of over 90%. The low and high quantitation limits, defined as the lowest or highest concentration to give a %CV <10%, were found to be 10 pmol and 5,000 pmol, respectively. The low limit of detection reported here is similar to that using the previous HPLC method (13 pmol) (6, 12, 13). Therefore, this isocratic HPLC method was shown to provide a linear, precise, and sensitive measurement of Cl-F-ara-ATP concentrations.
Within-day and inter-day variation for Cl-F-ara-ATP. Various amounts of standard Cl-F-ara-ATP in aqueous solution in triplicate were applied to HPLC to determine peak area counts on three separate days. The within-day variation was the value for day 1. The value of the inter-day variation was calculated from the mean value of the peak area count on each separate day.
Chromatography. To assess column performance, a mixture of nucleotides (CTP, UTP, ATP, GTP) and Cl-F-ara-ATP was run under study conditions. The Cl-F-ara-ATP peak was quite clear; the retention time was 81 min and it eluted independently of the other nucleotides (Figure 2A). Under study conditions, retention times of deoxyribonucleotides (dATP, dCTP, dTTP, dGTP) were similar to those of the corresponding ribonucleotides (data not shown). To assess applicability of our method to biological samples, the mixture of nucleotides extracted from cultured untreated leukemia cells was passed through the column; the results are shown in Figure 2B. It can be seen that endogenous nucleoside triphosphates (CTP, UTP, ATP, and GTP) separated in a pattern identical to that seen when using standard ribonucleotides. Furthermore, when cell supernatants were co-eluted with a Cl-F-ara-ATP standard, all peaks were distinctly isolated (Figure 2C). In addition, resulting chromatograms clearly demonstrated intracellular production of Cl-F-ara-ATP in Cl-F-ara-A-treated leukemia cells (Figure 2D).
Accuracy of Cl-F-ara-ATP measurement. Different amounts of standard Cl-F-ara-ATP in aqueous solution in triplicate were applied to the HPLC and the concentrations were determined in the assay procedure.
Cl-F-ara-ATP concentrations in leukemia cells treated in vitro. When HL-60 cells were incubated with different concentrations of Cl-F-ara-A for various time periods, Cl-F-ara-ATP concentrations increased in a dose- and time-dependent manner (Figure 3A and B).
Discussion
In previously reported studies, Cl-F-ara-ATP was separated using a Partisil 10-SAX (Whatman, Clifton, NJ, USA) column and a linear gradient over a period of 60 min at 1.5 ml/min, starting with 100% buffer A (0.005 M NH4H2PO4, pH 2.8) and ending with 100% buffer B (0.75 M NH4H2PO4, pH 3.6) (6, 12, 13). While this method requires a computerized system to regulate two pumps, the isocratic method presented here requires a single buffer that can be controlled using either a computerized system or a single, conventional, mechanical pump. Furthermore, isocratic elution does not induce base-line drift (Figure 2), thereby allowing accurate and sensitive measurement of Cl-F-ara-ATP in biologic samples (Figure 3). Therefore, this method may prove advantageous for conducting pharmacokinetic evaluation in hospitals lacking state-of-the-art facilities.
There have been several studies that examined plasma and cellular pharmacology during Cl-F-ara-A therapy (11, 13, 16). Patients with acute myeloid leukemia, acute lymphoblastic leukemia, or blast crisis of chronic myeloid leukemia received 40 mg/m2 of Cl-F-ara-A in a phase I clinical trial (16). The median plasma Cl-F-ara-A level reached 1.5 μM (0.42-3.2 μM). Cl-F-ara-ATP concentrations in circulating leukemia blasts were determined at the end of Cl-F-ara-A infusion, and the median concentration was 19 μM (3-52 μM) (16). In a phase II trial in patients with acute leukemia, myelodysplastic syndrome, or blast crisis of chronic myeloid leukemia, 40 mg/m2 Cl-F-ara-A was given daily for 5 days per course. The resulting median plasma Cl-F-ara-A concentration was 1.0 μM (0.26-1.94 μM). The median concentration of Cl-F-ara-ATP in circulating leukemia blasts was 15 μM (1-44 μM) (11). In addition, circulating chronic lymphocytic leukemia lymphocytes from two patients taking Cl-F-ara-A 15 mg/m2/d for 5 days were evaluated; accumulated Cl-F-ara-ATP concentrations were 2.8 and 8.3 μM, respectively (13). In all of these previous studies, 1 pmol/2×107 cells corresponded to a cellular concentration of 0.2 μM. Measured values (1-52 μM) corresponded to 2.5-130 pmol/107 cells; most were therefore within the range of detection demonstrated for our assay.
HPLC chromatograms. A: Standard nucleotides (ATP, CTP, UTP, GTP) mixed with Cl-F-ara-ATP and dissolved in water. B: Pooled nucleotides extracted from untreated human leukemic HL-60 cells (2×107 cells). C: Same supernatants as used in (B) co-eluted with standard Cl-F-ara-ATP. D: Supernatants extracted from HL-60 cells after incubation with 20 μM Cl-F-ara-A for 4 h.
Quantitating Cl-F-ara-ATP in leukemia cells in vitro. Human leukemia HL-60 cells (1×106 cells/ml, 20 ml) were incubated with different concentrations of Cl-F-ara-A for 6 h (A) or 10 μM Cl-F-ara-A for the indicated time periods (B). After washing with fresh medium, cells were extracted and supernatants were applied to the HPLC.
Cl-F-ara-A has been included in several clinical studies now in progress (17, 18). Pharmacokinetic evaluation of intracellular Cl-F-ara-ATP may provide information crucial for scheduling and dosing, since it varies widely among patients and subtypes of leukemia and is not predictable from plasma analysis. We have demonstrated that our simple, inexpensive, isocratic HPLC method has clinical utility, and suggest that it may eventually prove to be of great value on a daily basis for treating patients in hospitals having only standard laboratory facilities.
Acknowledgements
We are grateful for Genzyme Japan for kindly providing us Cl-F-ara-A and Cl-F-ara-ATP.
- Received May 13, 2011.
- Revision received June 21, 2011.
- Accepted June 22, 2011.
- Copyright© 2011 International Institute of Anticancer Research (Dr. John G. Delinassios), All rights reserved
