Pharmacokinetics of the 13C Labeled Anticancer Agent Temozolomide Detected zyx in zyxw Vivo by Selective Cross-Polarization Transfer zyxw Dmitri Artemov, Zaver M. Bhujwalla, Ross J. Maxwell, John R. Griffiths, Ian R. Judson, Martin 0. Leach, Jerry D. Glickson zyxwv The anticancer agent temozolomide labeled with I3C (8-Car- bam0y1-3-'~C-methylimidazo-[5,1 -d]-l,2,3,5-tetrazin-4-(3H)- one), was noninvasively detected in subcutaneous RIF-1 tu- mors by a selective cross polarization 13C NMR method, at a field strength of 9.4T. Pharmacokinetics of the drug, at a dose of 150 mg/kg, were determined for intravenous and intraper- itoneal modes of administration (three animals per mode). The half-life of the drug in the tumors was approximately 60 min. The uptake and clearance of the drug, however, varied signif- icantly between individual hosts, for both modes of adminis- tration. These results demonstratethe feasibility of obtaining pharmacokinetics of anticancer agents for individual tumors without the need for a label that might modify drug activity (e.g., fluorine). The variability of the zyxwvutsr in vivo measurements, even within the same tumor model, demonstratesthe neces- sity of directly monitoring the tumor to evaluate drug pharma- cokinetics. Key words: I3C NMR; temozolomide; tumors; in vivo detec- tion. INTRODUCTION Pharmacokinetic measurements are important determi- nants of both the activity and toxicity of anticancer agents (1). Thus, absorption, distribution, metabolism, and excretion can influence the levels of active drug and/or toxic metabolites. This can have a major impact on the outcome of chemotherapy. The conventional ap- proach to pharmacokinetic measurements is to deter- mine the concentration of a drug and its metabolites in accessible body fluids (plasma, urine) and use these data to model the exposure of relevant compartments such as tumor cells and sites of toxicity. Needless to say, the noninvasive detection of the anticancer agent directly in the tissue of interest would rule out uncertainties in- volved in using computer-modeled drug pharmacokinet- ics. The feasibility and advantages of using lgF NMR zyxwvu MRM zyxwvutsrqp 34:338-342 (1995) From the Department of Radiology, The Johns Hopkins University School of Medicine, Baltimore, Maryland (D.A., Z.M.B., J.D.G.), The St. George's HospitalMedicalSchool, London, England (R.J.M., J.R.G.), and the Institute of Cancer Research, Sutton, Surrey, England (R.J.M., I.R.J., M.O.L.). Address correspondence to: Zaver M. Bhujwalla, Ph.D., Department of Radiology, The Johns Hopkins University School of Medicine, Room 217, Traylor Building, 720 Rutland Avenue, Baltimore, MD 21205-2195. Received January 3, 1995; revised May 3, 1995; accepted May 22, 1995. This work was supported by National Institutes of Health Grants 1 R01 CA 51950 and CA 51935, Cancer Research Campaign Grant SP1971/0501, and North Atlantic Treaty Organization Collaborative Grant CRG No. 921346. Copyright 0 1995 by Williams & Wilkins All rights of reproduction in any form reserved. 0740-31 94/95 $3.00 spectroscopy to detect tissue pharmacokinetics of the anticancer agent 5-fluorouracil have been demonstrated for both animal (2) and human (3-5) tumors. NMR pharmacokinetic measurements of tumors in vivo, have, however, been restricted mainly to fluori- nated drugs detected by 19F NMR (6). Although 19F NMR has the advantage of relatively high sensitivity and no background signal, the physicochemical and pharmaco- logical properties of the drug may be altered by fluorine labeling. Even where a fluorinated compound is of inter- est, it is often desirable to compare the behavior of a range of related compounds (of which only one or a few may be fluorinated). Techniques have been developed for the detection of biochemical compounds by water sup- pressed, localized 'H NMR spectroscopy and spectro- scopic imaging (7, 8). Recently, He et zyx al. (9) have used a gradient filtering method to detect an unlabeled drug, iproplatin, in animal tumors by 'H NMR spectroscopy. 13C-labeled compounds (particularly glucose) have also been detected in vivo, both directly (lo) and with en- hanced sensitivity, by indirect detection using the adja- cent 'H nuclei (11). The main disadvantage of I3C NMR is its low sensitiv- ity (32-fold lower than 'H for high resolution NMR, and about eight-fold lower for in vivo studies (12, 13)). This can be partly overcome by (i) NOE (nuclear Overhauser enhancement): up to three-fold improvement; (ii) polar- ization transfer (e.g., INEPT (insensitive nuclei enhanced by polarization transfer) or DEPT (distortionless en- hancement by polarization transfer)): four-fold improve- ment; (iii) indirect detection of coupled protons: a factor of 16 (4 in vivo) for a C-H group using gradient-enhanced heteronuclear multiple quantum coherence (geHMQC) (11,13). The relative sensitivities for indirect detection of coupled protons differs by a factor of four for high reso- lution NMR and in vivo NMR studies due to inhomoge- neous line broadening and the loading of the RF coil; both of these properties are proportional to the frequency of the nucleus. However, each of these methods has its limitations: the full NOE may not be realized under con- ditions of restricted mobility; the long TI of 13Climits the sensitivity enhancement of the NOE; INEPT-based polar- ization transfer and ge-HMQC are sensitive to pulse im- perfections, motion artefacts, and T, losses. Selective cross polarization (CP) transfer has been used for detecting '%-labeled glucose in animal tumors (13). It has the advantages of higher sensitivity with lower power deposition (13). It is also less susceptible to mo- tion artifacts and T, relaxation losses and experimental imperfections. In the present study we have applied this method to detect a %-labeled anticancer agent, temozo- 338