Simplified Labeling Approach for Synthesizing 3'-Deoxy-3'-[¹⁸F]Fluorothymidine ([¹⁸F]FLT)

[¹⁸F]FLT (3'-deoxy-3'-[¹⁸F]fluorothymidine) turned out to be a tracer particularly suitable for PET imaging of tumor proliferation because of lacking degradation in vivo. To facilitate clinical studies with [¹⁸F]FLT, we investigated two new easily accessible precursors, 2,3'-anhydrothymidine (AThy) and 5'-O-(4,4'-dimethoxytriphenylmethyl)-2,3'-anhydro-thymidine (DMTThy), using a common approach for introducing the label with nucleophilic [¹⁸F]fluoride. Radiochemical yields were determined in dependence on substrate concentration, reaction time and temperature. In the case of AThy (10 mg), best FLT yields were 5.3% ± 1.2 (130° C, 30 min). Labeling of DMTThy (10 mg) gave 14.3% ± 3.3 at 160° C within 10 min. Starting with an aqueous solution of 20 GBq [¹⁸F]fluoride the new method allows to produce 1.3 GBq [¹⁸F]FLT within 90 min ready for intravenous injection. The new labelling procedures allow [¹⁸F]FLT synthesis without lengthy preparation of the precursor and with high reproducibility mandatory for clinical application.

For PET imaging of cell proliferation in tumors, Grierson and coworkers [1,2,3] developed 3'-deoxy-3'-[¹⁸F]fluorothymidine ([¹⁸F]FLT) as a tracer which appears particularly promising. In a recent imaging study Shields et al. [4] demonstrated the biochemical rationale of FLT for clinical application in PET assessment of tumor proliferation. Compared to ¹¹C-labeled thymidine, the advantages of the new tracer are the longer half life of fluorine-18 and the lack of metabolic degradation in vivo.

Until now, the clinical applicability of [¹⁸F]FLT has been limited by the lengthy synthesis of the precursor, the low reproducibility and extent of labeling in large scale reactions. Therefore, in an alternative synthetic route, 2,3'-anhydro-thymidine (AThy, (1)) was chosen as a precursor into which the [¹⁸F]-label can be introduced by nucleophilic substitution to form FLT (2). Since hydroxylic groups are thought to decrease the reactivity (nucleophilicity) of [¹⁸F]fluoride, a second route was used with 5'-O-(4,4'-dimethoxytriphenylmethyl)-2,3'-anhydro-thymidine (DMTThy, (3)), a precursor in which the OH-group is protected by the 4,4'-dimethoxytriphenylmethyl substituent (Fig.1).

Precursor for the labelling reaction
Precursors for the labeling reaction For the first synthetic approach 2,3'-anhydrothymidine (1) was purchased from Sigma and used as supplied. The precursor for the second route, 5'-O-(4,4'-dimethoxytriphenylmethyl)-2,3'-anhydrothymidine (3), was synthesized by a modified procedure of Siegmund et. al [5]. Briefly, this procedure started from commercially available 5'-O-(4,4'-dimethoxytriphenyl-methyl)thymidine (Sigma) using methanesulfonylchloride, then the isolated crude 3'-O-(methylsulfonyl)-5'-O-(4,4'-dimethoxytriphenylmethyl)-2,3'-anhydrothymidine was brought to reaction in dry CH₂Cl₂ using 1,8-diazabicyclo[5.4.0]undec-7-en (DBU) and molecular sieve to obtain (3), which was purified by means of a preparative silicagel (70-230 mesh) column. Identification of the structure was assured by MS and ¹H-NMR. The precursor can be stored at 0° C for several weeks, no alteration has been observed.

Radiochemistry
[¹⁸F]Fluoride was produced at the cyclotron (PETtrace, GE Medical Systems, Uppsala) in the PET-Center Tuebingen via the ¹⁸O(p,n)¹⁸F reaction by irradiating >94% enriched ¹⁸O-water with 16.5 MeV protons. ¹⁸F-activity (100 µL to 1300 µL) was transferred to a 5 mL vial fitted with a septum. Kryptofix [2.2.2] (Merck, 15 mg, 40 µmol) and KHCO₃ (4% aqueous solution, 75 µL, 30 µmol) were added to the [¹⁸F]fluoride solution. Water was removed from the solution by azeotropic distillation with acetonitrile (3 x 1 mL) at 140°C with the aid of an argon sparge line. DMSO solutions (1 mL, <0.01% water, Fluka) with varying amounts (1-20 mg) of either precursor (1) or (3) were added to the dry residues and then stirred. Labeling reactions were carried out for various times (5-60 min) at temperatures between 100°C - 180°C. With the optimum amount of precursor the effect of temperature (100°C to 180°C) and reaction time (5-60 min) was studied.

For removal of the DMT protecting group from (4), 350 µL of 1 N HCl were added and kept at 50°C for 10 min. In case of production runs the acidity of the solution was decreased by 1.5 mL 0.5 N sodium acetate before hplc purification. Complete hydrolysis was assessed by tlc. In production runs with precursor (3), a purification step was added before hydrolysis to increase the efficiency of the hplc separation. In that case, the reaction solution was first diluted with water (4 mL), then the solution was passed through two preconditioned (ethanol/water) Sep-Pak Plus C-18 cartridges (Waters). After washing with water (20 mL), the product was eluted with 2mL DMSO.

As indicated by the activity balance, decomposition of [¹⁸F]FLT did not occur. Therefore the reaction parameters were determined by assessing the yield of DMT-FLT representing the yield of FLT (2) directly. The radiochemical yields were decay corrected and based on the amount of fluorine-18 initially solubilized into DMSO with an efficiency of more than 85 % and did not account for losses on glass surfaces and handling.

Analyses
Product solutions were analyzed by thin layer chromatography using silica gel plates (Polygram® Sil G/UV254, Macherey-Nagel, Germany) and eluted with ethyl acetate/heptane (2/1, v/v). Non-radioactive FLT or DMT-FLT were used as standards on the same tlc plate. The radioactive spots were assessed quantitatively by means of an InstantImager (Canberra Packard, electronic autoradiography). The Rf-values on silica gel for fluoride, FLT and DMT-FLT were 0.0, 0.18 and 0.66, respectively.

In addition, hplc equipped with a NaI(Tl)-scintillation detector and an UV detector (254 nm) was used for identification of the labeled product. DMT-FLT (4) was separated on a C₁₈ phase (Phenomenex, Luna 5µ C₁₈ (2), 250 mm x 10 mm (i.d.) column) eluted with aqueous acetonitrile (35/75 v/v, k'= 4.0). The same hplc column was used for isolating FLT from the reaction mixture with isotonic sodium chloride solution and ethanol (90/10, v/v) as eluent at a flow rate of 5 mL/min (k'= 3.0).

RESULTS AND DISCUSSION

The amount of precursor clearly affected the labeling yield. As shown in Table 1 with 1 mg of (1) no product was observed but 5 mg precursor yielded 2.7% ± 0.1 of product and further increase was found with 10 mg (2.8% ± 0.5) or 20 mg (3.7% ± 1.7). In the case of (3), labeling yields of 9.0% ± 1.5 and 11% ± 0.4 were found with 10 mg and 20 mg precursor, respectively. As a compromise 10 mg were used for further studies of the labeling reaction with (3) in order to have a sufficiently good hplc purification of [¹⁸F]FLT and, yet, to use the column described above. A clear effect was found in the dependence on the reaction temperature (Fig. 2). At 130°C the labeling of (1) yielded 5.3% ± 1.2 FLT. Above 150°C the radiochemical yield clearly decreased down to 0.9% ± 0.3 at 170°C. In case of (3), on the other hand, even higher yields were found above 140°C, with the best results (14.3% ± 3.3) at 160°C. The optimized labeling conditions gleaned from these pilot experiments (10 mg of precursor (3), 160°C, 10 min) were used for large-scale labeling syntheses of [¹⁸F]FLT. Losses of [¹⁸F]-activity due to manual manipulations and adsorption on the glass reaction vessel were observed. Yet, after hplc purification 1.3 GBq [¹⁸F]FLT could be prepared from 20 GBq of [¹⁸F]fluoride within 90 min as a sterile solution ready for intravenous injection.

CONCLUSION

Two easily accessible precursors were successfully investigated using the common method of introducing the [¹⁸F]-label by a nucleophilic substitution reaction. Thus, [¹⁸F]FLT can be prepared without lengthy synthesis of the precursor and with a high reliability which allows to start the clinical application of [¹⁸F]FLT in a routine scale. Moreover, the labeling reaction is well suited for automation facilitating routine production for clinical studies.

ACKNOWLEDGMENTS

This work was supported by the fortüne research program of the University Hospital of Tuebingen (611-0-0).

REFERENCES

1. Grierson JR, Shields AF, Eary JF; Development of a radiosynthesis for 3'-[F-18]fluoro-3'-deoxynucleosides. J. Labelled Comp. Radiopharm. 1997; 40:60-62.
2. Grierson JR, Shields AF; Optimizing labeling substrate structure for 3'-deoxy-3'-[F-18]Fluorothymidine: J. Nucl. Med. 1999; 40:83P.
3. Grierson JR, Shields AF; An improved radiosynthesis of [F-18]FLT. J. Labelled Comp. Radiopharm. 1999; 42(Suppl.1):S525-S526.
4. Shields AF, Grierson JR, Dohmen BM et al.; Imaging proliferation in vivo with [F-18]FLT and positron emission tomography. Nature Medicine 1998; 4:1334-1336.
5. Siegmund H, Pfleiderer W, The synthesis of phospholipid conjugates of antivirally active nucleosides by the improved phosphoramidite methodology. Helv. Chim. Acta 1996; 79:426-438.