NOTES Annals of Nuclear Medicine Vol. 7, No. 3, 199-205, 1993 Brain tumor accumulation and plasma pharmacokinetic parameters of 2¥-deoxy-5-18F-fluorouridine Kiichi ISHIWATA,* Yuji TSURUMI,** Motonobu KAMEYAMA,** Kiyotaka SATO,** Ren IWATA,* Toshihiro TAKAHASHI,* Tatsuo IDo* and Takashi YOSHIMOTO** *Division of Radiopharmaceutical Chemistry, Cyclotron and Radioisotope Center, Tohoku University * * Division of Neurosurgery, Institute of Brain Diseases, Tohoku University School of Medicine Using positron emission tomography and radio-high performance liquid chromatography, the accumulation of 2'-deoxy-5-18F-fluorouridine in the brain tumors and plasma pharma-cokinetic parameters were investigated in 20 patients. High accumulation of the tracer in high grade gliomas and meningiomas and very rapid degradation of the tracer in the plasma were found. Very large variations were observed in both tumor accumulation and pharma-cokinetic data. The tumor accumulation, however, did not correlate with any of the plasma pharmacokinetic parameters : area under the plasma concentration-time curve, mean re-sidence time, total body clearance and steady-state volume of distribution. The results suggest that the accumulation of the tracer reflects the metabolic activity of the brain tumor tissues and that the effect of the rapid metabolic change in the tracer in the p]asma on the tumor accumulation may be minor. Key words: brain tumor, 2'-deoxy-5-18h~-fiuorouridine, positron emission nuc]eic acid metabolism, pharmacokinetics tomography, INTRODUCTION FLUORlNATED PYRIMIDINE derivatives are antineo-plastic agents widely used in the treatment. of gastrointestinal and breast cancers.1 The cytotoxicity is most likely mediated by the formation of an ir-reversible complex with thymidylate synthase and in part by incorporation into RNA via fluorinated nucleotides. Because the fluorinated pyrimidines are converted to nucleotide forms by several enzymes in nucleic acid metabolism, we have proposed 18F-labeled 2'-deoxy-5-fluorouridine (_18F-FdUrd) as a radiopharmaceutical for the nucleic acid metabolism to assess cancer viability in vivo by positron emission tomography (PET).2-13 High grade gliomas were clearly visualized and the accumulation of 18F-FdUrd was higher than that in the low grade glioma or sur- rounding normal brain tissues.7,8 However, a large variation in the accumulation in high grade gliomas was found. Several reasons are considered to explain the variation ; 1) the heterogeneity of high grade gliomas and the difference in proliferative potential in the same grade gliomas, 2) the effect of treatment on the tumor metabolism and the drug metabolism in the whole body, 3) Iimitations in reliable biopsies to assess the histological grade and 4) the individual difference in the catabolism of the tracer. In a pre-liminary metabolic study of human plasma, degrada-tion of the 18F-FdUrd ; 18F-FdUrd -> 5-18F-fluo-rouracil (18F-FUra) H> 5,6-dihydr0-5-18F-fluorouracil (18F-DHFU) -> ce-18F-fluoro-P-ureidopropionic acid (18F-FUPA) -> ce-18F-fluoro-P-alanine (18F-FBAL), was very rapid and seemed to be dose-dependent.ro The rapid degradation is probably controlled by the individual catabolic activity in the whole body, especially in the liver. The plasma pharmacokinetic parameters of 18F-FdUrd may affect the accumula-tion of the tracer in the high grade gliomas in which integrity of the blood-brain barrier cannot be pre-served. This study investigated whether the different potential of drug metabolism affects on the accumu-lation of 18F-FdUrd in brain tumors. The pharma-cokinetic parameters were measured by means of a radio-high performance liquid chromatographic (radio-HPLC) technique. MATERIALS AND METHODS Chemicals Uracil, FUra and FdUrd were purchased from Wako Pure Chemical Industry (Tokyo), 5-fluorouridine from Sigma Chemical Company (St. Louis), FBPA from Tokyo Kasei Kogyo (Tokyo), and 2-deoxy-uridine from Seikagaku Kogyo (Tokyo). DHFU was generously supplied by Hoffmann La Roche, Basel, Switzerland. All other reagents were of the highest grade available. Preparation of 18F-FdUrd 18F-FdUrd was prepared by a previously described method,14 or the following modified method. Acetyl 18F-hypofluorite was bubbled into 6 mL of acetic acid containing 1 5 mg of 2-deoxyuridine at room temperature. After drying the solution by evapora-tion, the residue was dissolved in 2 m/ of triethyl-amine and heated at 90'C for 5 min. After drying the solution, the residue was dissolved in 1.0 ml of 0.1 % acetic acid and applied to HPLC with a Delta-Pak C.18 cartridge (25 mm i.d. x 100 mm length) with a pre-column cartridge (25 mm i.d. x 10 mm) equipped with an RCM 25 x 10 compression module (Waters). The mobile phase was 0.1 % acetic acid containing 2 % ethanol and the fiow rate was 20 m/l min. 18F-FdUrd eluted 8.5 to 10.5 min was collected and evaporated to dryness. The residue was dis-solved in physiological saline, and the solution was passed through a O.22 pm membrane filter for injec-tion. The specific activity was 26-45 GBq/mmol. PET studies Twenty patients with brain tumors listed in Table I , were examined by PET. Eleven patients were studied before radio-chemotherapy, and the remaining 9 patients were studied during or after the therapy. All of the patients showed normal liver functions which were investigated by routine blood biochemical examinations. Some patients were treated with radial surgery after the PET examination. Histo-logical diagnoses were made from CT-guided stereotaxic biopsies or operative specimens. Before the PET study. CT scanning was carried out. The PET scanner used was a PT-931 model with four detector rings and a spatial reso]ution of 8 mm full width at half maximum (CTI, Knoxville, Ten-nessee). The patients were injected intravenously with 120 to 410 MBq (0.9 to 4.0 mg). of 18F-FdUrd as a bolus, and sequential images with 5 min data acquisition were obtained over a period of 40 to 60 min. Before the emission scan, a transmission scan with a 68Ge/68Ga external ring source was performed. The emission data were corrected for attenuation by means of the transmission data. Each pixel count was converted to a radioactivity concentration (nCi/m/) by applying the cross-calibration factor for the dav between the well scintillation counter and the positron camera. The radioactivity in the tumor was measured in the region of interest (volume : 2.2 m/), including the highest radioactivity point, and was expressed as the differential absorption ratio {DAR-PET, [tissue radioactivity/total injected radio-activity]f[tissue volume (m/)/body weight (g)]} . The mass concentrations of the tracer (equivalent to FdUrd) in the tumor were also represented as ng/ml tissue, calculated from the uptake of radioactivity and specific activity of 18F-FdUrd injected. This project was approved by the Committee for Clinical PET Study of Tohoku University and informed consent was obtained from every patient. Measurement of labeled metabolites in pla,snla During the PET scanning, arterial blood sample** were taken at 0.33, 0.67, 1.0, 1.5, 2.0, 3.0, 4.0, 5.0, 7.5, 10, 15, 20, 30, 40, 50 and 60 min after the i.v. administration. The blood volumes were 3 ml at l, 3, 5, 10, 1 5, 30 and 60 min, and I ml at other times. About 0.5 m/ of plasma was weighed and measured for radioactivity with an Nal(Tl) gamma scintillation counter. The radioactivity level was expressed as the differential absorption ratio {DAR, [plasma radio-activity/total injected radioactivity]/[plasma wei**ht (g)/body weight (g)]} . Plasma samples taken at l, 3, 5, 10, 15, 30 and 60 min were assayed for measurement of radio-active metabolites by a previously described methodro with slight modification. Briefly, 0.5 to 1.5 ml of the plasma was treated with I M HCI04. After centrif-ugation, the acid-soluble supernatant was neutralized with 1M KOH and the precipitated KCI04 was removed by centrifugation. The supernatant was applied on a reverse phase column (ERC-ODS-2121, 8 mm i.d. x250 mm length, Erma Optical Industry, Tokyo). The column was eluted with an ion pair solution with a gradient modifier at room tempera-ture at a constant flow rate of 1.5 ml/min. For the first 10 min the initial so]vent (buffer A: 5 mM sodi-um phosphate buffer, pH 6.8, containing 10 mM tetrabutylammonium hydrogensulfate) was eluted. For the next 10 min, the column was eluted with a solvent with a liner gradient of the following compo-sition : 0% to 100% of the stronger elution solvent {buffer B : a mixture of methanol and 10 mM sodium phosphate, pH 6.8, containing 10mM tetrabuty-lammonium hydrogensulfate (50: 50, v/v)} . For the last 7.5min, the buffer B was eluted, and then the column was reequilibrated with the initial buffer A. The elution profile was detected with a radioactivity monitor (Ramona-D equipped with an IM-2020X fiow cell for 3H/14C measurement, Raytest), and the radioactivity in each 1.0 ml was measured with an auto-gamma scintillation counter. The radioactivity was corrected for the half-life of 18F. A portion of the sample applied was also measured to calculate the total applied radioactivity, and the percentage of radioactivity in each peak of the total applied radio-activity was calculated. Recovery of the radioactivity was essentially quantitative. Pharnracokinetic cakulations The area under the plasma concentration-time curve (AUC) or 18F-FdUrd for the interval 0-60 min and the mean residence time (MRT), the steady-state volume of distribution (Vd) and the total body clearance (CL) were calculated according to the method of Yamaoka et al.15 The AUC for 18F-FdUrd was written as the AUC1, and AUC2 and AUC3 were also calculated for 18F-FUra and 18F-DHFU. RESULTS AND DISCUSSION By PET with 18F-FdUrd, all brain tumors investi-gated were clearly visualized as a positive spot in the region which was shown as the high density by post contrast CT. Time-radioactivity curves of the tumor and plasma are shown in Fig. I . Radioactivity in the plasma was rapidly cleared for the first 20 min, while it decreased slightly in the tumor. After this time, both radioactivity levels remained constant. These patterns were similar in all 20 patients, and DAR-PET values at 30 to 40 min after the injection are summarized in Table I . The mean DAR-PET values in patients before radiochemotherapy and in patients during and after treatment were 0.78+-0.29 and 0.62 +-0.30, respectively, but the difference did not reach statistical significance. The DAR-PET values in the contralateral brain regions were O. 1 7 +-0.05. Although the radioactivity represents 18F-FdUrd and its metabolites, the mass concentration in the tumors was also expressed as the mass equivalent of FdUrd (ng)/tissue volume (m/) (Table 1). Radioactive metabolites in the plasma were analyzed by radio-HPLC. In a typical analysis, six radioactive peaks were detected : their retention times were 6.1, 7.8, 9.0, 11.8, 13.8 and 21.2 min. As pre-viously described,10 metabolites were identified as follows. Radioactive peaks 4, 5 and 6 were 18F-DHFU, 18F-FUra and 18F-FdLTrd, respectively, compared with retention times for the authentic samples. When peak I in the elution front was col-lected and analyzed by means of a cation-exchange column,3 over 90 % of the radioactivity was eluted as 18F-FBPA. Peak 2 was not identified. Peak 3 was assumed to be 18F-FUPA, because P-ureidopropionic acid showed a corresponding retention time. Peaks 2 and 3 overlapped when their ratios were high. 5-Fluorouridine was eluted in the retention time of 19.5 min, where the corresponding radioactivity was detected in very small amounts (<1%) in some cases. Radioactive nucleotides were negligible. Aver-age percentages for the labeled metabolites in 20 patients are summarized in Table 2. Very rapid degradation of 18F-FdUrd occurred. Subsequently, proportions of 18F-FUra and 18F-DHFU initially increased with time and then decreased. The pro-portion of peak I was predominant later. In 60 min-samples, the proportions of 18F-FdUrd and 18F-FUra were lower than 1%. In three typical cases, plasma clearance of 18F-FdLTrd and its metabolites ex-pressed as mass concentrations (,ug/m/) is shown in Fig. 2 (A, No. 5; B, No. 10; and C, No. 19). Plasma pharmacokinetic data for 20 patients are summarized in Table 3. No clear relationship between the tumor uptake (DAR-PET values and ng/ml) and any plasma pharmacokinetic parameters was observed (Fig. 3). During the time-scale of PF_T scanning, very large individual variations and some interesting aspects were found in the pharmacokinetic parameters in plasma, although this study was carried out in patients with normal liver functions which were investigated by routine blood biochemical ex-aminations. The following individual differences were found in the three catabolic reactions. The first reaction from FdUrd to FUra was catalyzed by thymidine phosphorylase, and the MRT values were in the 1.9 to 5.1 min range. Further degradation steps catalyzed by dihydrouracil dehydrogenase and dihydropyrimidinase did not necessarily parallel the MRT of 18F-FdUrd. In two patients {Nos. 5 (Fig. 2A) and 7} who showed large AUC2/AUC1 ratios, the reaction from FUra to DHFU was slow. In another group {Nos. 4, 10 (Fig. 2B), 14 and 18} , the degradation of DHFLT assessed by the AUC3/ AUC2 ratios was slow. Among 20 patients patient No. 19 (Fig. 2C) showed signs of very rapid elimina-tion of all FdUrd, FUra and DHFU. The lowest radioactivity accumulation in the tumor was ob-served in this case, whereas patient No. 12 who had similar MRT and ALJCI values to No. 19 had the largest DAR-PET. In experimental tumor tissues, a considerable amount of the total radioactivity was present as 18F-FdUrd and 18F-FUra as well as their biologically active forms which were activated by such enzymes as thymidine kinase, uridine phosphorylase, uridine kinase, thymidylate synthase and so on.10 And catabolites of the 18F-FdUrd were found in the tumor tissues.10 Because AUC shows the bioavailability of drugs or input function in PET studies, AUCI and AUC2 may be major parameters affecting the tumor uptake of the 18F-FdUrd. DAR-PET values in the brain tumor, however, were not necessarily correlated with these parameters (Fig. 3). In animals2~4 and humans ,5 most of the 18F-FdUrd was degraded in the liver and!or excreted into urine through the kidneys, and total tumor uptake was low. MRT, CL and Vd values are controlled by the whole body metabolism. The effect of the catabolites of the 18F-FdUrd on the tumor uptake may therefore be repre-sented by these three parameters and total plasma radioactivity. And the DAR-PET values in the brain tumor were not correlated with these parameters (Fig. 3), with total radioactivity levels or with integrated plasma radioactivity in individual patients (data not shown). The secluential studies before and after radio-chemotherapy in the same patients may more clearly represent the relationship between the DAR-PET values and pharmacokinetic parameters. However, these studies were carried out in a limited number of patients, because some patients were treated with radical surgery after the first PET study and because metabolites analyses could not be carried out to obtain the pharmacokinetic parameters in some patients. Several reasons were considered to explain a large variation in the DAR-PET values in previous and this studies. As discussed above, it was found that the individual difference in the catabolism of the tracer may be a minor contribution. In none of the patients investigated, was the blood-brain barrier in the tumor region reserved, because enhancement was observed in the region by CT scanning. Because the tracer does not essentially pass through the blood-brain barrier, the status of the barrier may be another factor. However, if the 18F-FdUrd accumulation in the tumor is dependent on the breakdown of the barrier, it should have been decreased with time due to the concentration gradient between the plasma and brain tumor tissue. This was not true in our studies. On the other hand, the analysis of Patlak's plotl6 was done, and the active uptake pattern of 18F-FdUrd was observed in high grade tumors (un-pub]ished data). Radio-chemotherapy may also influence the 18F-FdUrd accumulation in the tumor. In this study, no significant difference was found between the DAR-PET values for the patient during and after the radio-chemotherapy or those in patients before the treatment. However. the effect in the same patients in the process of treatment needs to be dealt with, and this work is in now progress. For the reasons given above and the experimental studies with tumors on the metabolic and autoradiographic comparison of 18F-FdUrd and radio-thymidine4,9,10 and radio- and chemotherapeutic effects on the 18F-FdUrd accumulation,11,13 we consider that the ac-cumulation of 18F-FdUrd in brain tumors is mainly dependent on the tumor metabolism and that the effect of rapid drug metabolism on tumor ac-cumulaton may be only minor. ACKNOWLEDGMENTS This work was supported, in part, by Grant-in-Aid for Scientific Research No. 01571193 from the Ministry of Education, Science and Culture, Japan. The authors wish to thank the staff at the Cyclotron and Radioisotope Center for their cooperation and Hoffmann-La Roche, Basel, Switzerland, for supplying DHFU. REFERENCES 1 . Heiderberger C, Danengerg PV, Moran R : Fluori-nated pyrimidines and their nucleoside. Adv Enzymol 54: 57-119, 1983 2. Abe Y, Fukuda H, Ishiwata K, et al : Studies on 18F-labeled pyrimidines. Turnor uptake of 18F-5-fiuoro-uracil, 18F-5-fiuorouridines and 18F-5-fluorodeoxy-uridine in animals. Eur J Nucl Med 8 : 258-261, 1983 3. Ishiwata K, Ido T, Kawashima K, et al : Studies on 18F-labeled pyrimidines II. Metabolic investigation of 18F-5-fluorouracil, 18F-5-fiuor0-2'-deoxyuridine and 18F-5-fluorouridine in animals. Eur J Nucl Med 9: 185-189, 1984 4. Ishiwata K, Ido T. Abe Y, et al: Studies on 18F-labeled pyrimidines 111. Biological investigation of 18F-labeled pyrimidines and comparison with 3H-deoxythimidine in tumor-bearing rats and mice. Eur J Nucl Med 10: 39~'4, 1985 5. Abe Y, Matsuzawa T, Itoh M, et al : Cancer detection with 18F-5-fiuorodeoxyuridine-a new cancer diag-nosis agent reflecting the nucleic acid metabolism. Jpn J Nucl Med 22 : 583-586, 1985 (in Japanese) 6. Kiyosawa M, Mizumo K, Ishiwata K, et al : Orbital tumor diagnosis by positron emission tomography using 18F-fiuorodeoxyuridine. Ophthalmic Res 18: 292-298, 1986 7. Kameyama M, Tsurumi Y, Itoh J, et al : Biological malignancy and 18FdUrd uptake in glioma patients -PET study of nucleic acid metabolism. CYRIC Annual Report, Tohoku University, 1988 : 215-227, 1988 8. Kameyama M, Tsurumi Y, Shirane R, et al : Nucleic acid metabolism in glioma studied with 18FdUrd and PET. J Cereb Blood Flow Metab 9 (suppl 1): S221, 1989 9. Tsurumi Y, Kameyama M. Ishiwata K, et al : 18F-fiuor0-2'-deoxyuridine as a tracer of nucleic acid metabolisrn in brain tumors. J Neurosurg 72 : 1 10-113, 1990 10. Ishiwata K, Sato K, Kameyama M, et al : Metabolic fates of 2'-deoxy-5-[18F]fluorouridine in tumor-bcaring mice and human plasma. Nucl Med Biol 1 8 : 539-545, 1991 11. Kubota K, Ishiwata K, Kubota R, et al: Tracer feasibility for monitoring tumor radiotherapy: A quadruple tracer study with fluorine-18-fluoro-deoxyglucose or fluorine-18-fluorodeoxyuridine, L-[methyl-14C]methionine, [6-3H]thymidine, and gal-lium-67. J Nucl Med 32: 2118-2123, 1991 12. Ishiwata K, Takahashi T, Iwata R, et al : Tumor diagnosis by PET : Potential of seven tracers examined in five experimental tumors including an artificial metastasis model. Nucl Med Biol 19: 611-618, 1992 13. Sato K, Kameyama M, Ishiwata K, et al : Metabolic changes of glioma following chemotherapy : An ex-perimental study using PET tracers. J Neuro-Oncol 14: 81-89 1992 14. Ishiwata K, Monma M, Iwata R, et al : Automated synthesis of 5- [18F]fluor0-2'-deoxyuridine. Appl Radiat Isot 38: 467-473, 1987 15. Yamaoka K, Tanigawara Y, Nakagawa T, et al : Capacity-limited elirnination of cefmetazole in rat. Int J Pharmaceu 10: 291-300, 1978 1 6. Parlak CS, Blasberg RG, Fenstermacher JD : Graphi-cal evaluation of blood-to-brain transfer constants from multi-time uptake data. J Cereb Blood Flow Metab 3 : 1-7, 1983