Synthesis of 2′-Deoxy-2′-[18F]Fluoro-9-β-D-Arabinofuranosylguanine: a Novel Agent for Imaging T-Cell Activation with PET
9-(β-D-Arabinofuranosyl)guanine (AraG) is a guanosine analog that has a proven efficacy in the treatment of T-cell lymphoblastic disease. To test the possibility of using a radiofluorinated AraG as an imaging agent, we have synthesized 2′-deoxy-2′-[18F]fluoro-9-β-D-arabinofuranosylguanine ([18F]F-AraG) and investigated its uptake in T cells.
We have synthesized [18F]F-AraG via a direct fluorination of 2-N-acetyl-6-O-((4-nitrophenyl)ethyl)-9-(3′,5′-di-O-trityl-2′-O-trifyl-β-D-ribofuranosyl)guanine with [18F]KF/K.2.2.2 in DMSO at 85°C for 45 min. [18F]F-AraG uptake in both a CCRF-CEM leukemia cell line (unactivated) and activated primary thymocytes was evaluated.
We have successfully prepared [18F]F-AraG in 7–10% radiochemical yield (decay corrected) with a specific activity of 0.8–1.3 Ci/μmol. Preliminary cell uptake experiments showed that both a CCRF-CEM leukemia cell line and activated primary thymocytes take up the [18F]F-AraG.
For the first time to the best of our knowledge, [18F]F-AraG has been successfully synthesized by direct fluorination of an appropriate precursor of a guanosine nucleoside. This approach maybe also useful for the synthesis of other important positron emission tomography (PET) probes such as [18F]FEAU, [18F]FMAU, and [18F]FBAU which are currently synthesized by multiple steps and involve lengthy purification. The cell uptake studies support future studies to investigate the use of [18F]F-AraG as a PET imaging agent of T cells.
Key wordsImaging T cells 9-(β-D-Arabinofuranosyl) guanine (AraG) Positron emission tomography (PET)
The fluorinated purine and pyrimidine derivatives have attracted attention in the past few decades due to their strong bioactive properties . Several 2′-deoxy-2′-fluoro-arabino nucleosides have been reported as antiviral agents [2, 3, 4]. Also, many 2′-β-fluoropurines dideoxynucleosides have potent inhibitory activities against the HIV-1 virus [5, 6, 7, 8]. Likewise, 2′-deoxy-2′-fluoro-9-β-d-arabinofuranosylguanine (F-AraG) is known to exhibit selective T-cell toxicity. It has been established that the strong electron-withdrawing effect of fluorine of a fluorinated substrate has a major effect on the chemical stability  and enzymatic activity of the substrate .
Some 20 years ago, the selective in vitro toxicity of the 2-deoxyguanosine analog 9-β-d-arabinofuranosylguanine (AraG) for T lymphoblasts was noted by several investigators [10, 11]. AraG is metabolized in a unique fashion by deoxyguanosine kinase and incorporated into mitochondrial DNA . These observations led to the synthesis of a water-soluble AraG prodrug, 2-amino-6-methoxypurine arabinoside (506U, Nelarabine) for potential clinical application in the treatment of T lymphoblastic diseases. This compound, developed over a number of years by Glaxo Smith Kline, is now FDA-approved for the treatment of relapsed T-cell acute lymphoblastic leukemia and T-cell lymphoblastic lymphomas [13, 14]. Radu et al. have recently published data on the molecular imaging of lymphoid tissues using 1-(2′-deoxy-2′-[18F]fluoroarabinofuranosyl)cytosine ([18F]FAC) . The striking selectivity of this compound for the immune system immediately raised the question of whether AraG, which has undergone extensive testing to demonstrate T lymphoblast selectivity, might be a more specific imaging agent for activated T lymphocytes as well as for T lymphoblasts. In addition, this work has raised the question of whether in vivo imaging with a labeled AraG compound might also be useful in the imaging and/or treatment of acute graft versus host disease. Despite much speculation, this question has never been addressed experimentally.
We have been exploring the synthesis of 8-fluoroguanine and radio-labeled 8-[18F]fluoroguanine derivatives in search for in vivo probe for imaging reporter gene expression with positron emission tomography (PET) . We have developed a method for the preparation of 8-[18F]fluoroguanine derivatives based on a direct radiofluorination reaction and were able to synthesize 8-[18F]fluoroguanosine from guanosine [1, 17, 18]. Recently, the synthesis of 2′-deoxy-2′-[18F]fluoro-9-β-d-arabinofuranosyladenine ([18F]FAA]) via a direct fluorination of a protected adenosine triflate has been reported . Here, we report the direct fluorination method for the preparation of the novel PET imaging agent 2′-deoxy-2′-[18F]fluoro-9-β-d-arabinofuranosylguanine ([18F]F-AraG). Additionally, the probe’s applicability as a T-cell imaging agent was preliminarily investigated via T-cell uptake studies.
Materials and Methods
No carrier-added [18F]fluoride was prepared by the 18O(p, n)18F nuclear reaction on a GE PET tracer cyclotron. [18F]Fluoride processing and synthesis of crude [18F]-labeled guanosine derivative 6 were completed in the GE TRACER lab FX-FN synthesis module.
Cell Uptake Studies
CCRF-CEM (acute lymphoblastic T leukemia cell line procured from ATCC) cells were maintained in RPMI 1640 (Cellgro) supplemented with 100 U penicillin/100 μg streptomycin/mL (Cellgro) and 10% fetal bovine serum (Gibco). CCRF-CEM cells, 5 × 105, were plated in 12 well dishes. Cells were allowed to settle for 1 h then [18F]F-AraG was added to each well and incubated at 37°C for indicated times. Primary T cells were isolated from spleens and thymus of outbred mice. Briefly, tissues were minced and single-cell suspensions made in RPMI supplemented as above. Cells were centrifuged at 300 ×g for 8 min at 4°C, and cell pellets were resuspended in phosphate-buffered saline (PBS) supplemented with 0.5% bovine serum albumin. T cells were purified from the suspensions using Miltenyi Pan T antibodies and columns (Miltenyi Biotec) per manufacturer’s directions. Purified T cells were either stimulated with 100 U/mL IL-2 (eBiosciences) for 24 or 48 h, or 50 nM phorbol myristate acetate (PMA; Fluka) with 1 μg/mL ionomycin (Sigma)  for 48 h or left unstimulated then exposed to 1 μCi of [18F]F-AraG/106 cells for 60 min. All cells were then washed in 1× PBS, lysed, and radioactivity of each well was determined by gamma counter for [18F] samples. All results were done in triplicate and are expressed in counts per minute per microgram of protein, standard errors were determined, and student t tests were performed.
Trifluoromethanesulfonyl chloride (44 μl, 0.42 mmol) in CH2Cl2 (1 mL) was added to a solution of 1 (0.20 g, 0.21 mmol) and DMAP (25 mg, 0.21 mmol) in CH2Cl2 (4.1 mL) containing triethylamine (58 μl, 0.21 mmol). The mixture was stirred for 1 h at room temperature and concentrated under vacuum. The crude triflate was purified by column chromatography (silica gel) using 1:1 chloroform and ethyl acetate as the eluent to afford 156 mg (68%) of 2 as a colorless viscous oil. 1H NMR (400 MHz, CDCl3) δ ppm: 2.04 (3H, s, Ac), 2.64 (1H, d, H5′, J5′5′′ = 11.2 Hz), 3.1 (1H, bs, H4′), 3.23 (1H, d, H5′′), 3.34 (2H, t, (nitrophenyl)ethyl, J = 6.9 Hz), 4.40 (1H, d, H3′, J2′3′ = 4.7 Hz), 4.81(2H, t, (nitrophenyl)ethyl), 6.03(1H, m, H2′), 6.64 (1H, d, H1′, J1′,2′ = 7.8 Hz), 7.16–7.35 (30H, m, 2 x trityl), 7.52 (2H, d, (nitrophenyl)ethyl, J = 8.7 Hz), 7.68 (1H, s, NH), 8.09 (1H, s, H8), 8.18 (2H, d, (nitrophenyl)ethyl). 19F NMR (CDCl3) δ ppm:−75.0 (s). High resolution MS: Calcd. MH+ for C59H50N6O10F3S: 1091.3261; Found 1091.3279.
Compound 1 (277 mg, 0.289 mmol) in dry CH2Cl2 (3 mL) was added to a solution of diethylaminosulfur trifluoride (DAST; 231 μl, 6 equiv) in dry CH2Cl2 (4 mL) containing pyridine (231 μl, 0.21 mmol). The mixture was stirred at room temperature overnight and diluted with CH2Cl2 (30 mL). The solution was subsequently washed with 5% NaHCO3 (6 mL) and H2O (6 mL). After drying the organic fraction over with MgSO4, the solvent was evaporated under vacuum. Crude product was purified by column chromatography (silica gel) using 4:1 chloroform and ethyl acetate as the eluent to afford 79 mg (30%) of 3 as foam.
1H NMR (400 MHz, CDCl3) δ ppm: 3.16 (1H, dd, H5′, J5′,4′ = 2.6 Hz, J5′,5′′ = 10.0 Hz), 3.25–3.30 (3H, m, H5′′,(nitrophenyl)ethyl), 3.72 (1H, dd, H2′, J1′,2′ = 2.3 Hz, J2′,F = 50.1 Hz), 4.24 (1H, dd, H3′, J3′,4′ = 2.3 Hz, J3′,F = 15.4 Hz), 4.49 (1H, m, H4′), 4.67–4.77 (2H, m, (nitrophenyl)ethyl), 4.88 (2H, bs, NH2), 6.28 (1H, dd, H1′, J1′,2′ = 2.3 Hz J1′,F = 25 Hz), 7.20–7.41 (30 H, m, 2xTr), 7.46 (2H, d, (nitrophenyl)ethyl, J = 8.7 Hz), 7.71 (1H, d, H8, J8,F = 3.7 Hz), 8.15 (2H, d, (nitrophenyl)ethyl). 19F NMR (CDCl3) δ ppm:−196.11 (m).
A solution of compound 3 (55 mg, 0.06 mmol) in dry pyridine (1.2 mL) containing DBU (91 mg, 0.6 mmol) was kept at room temperature for 15 h. The reaction mixture was neutralized to pH 6 with acetic acid and evaporated under vacuum. The residue was co-evaporated with toluene, dissolved in CH2Cl2 (4 mL), and the solution was washed with H2O (2 × 1 mL). After drying the organic fraction over with MgSO4, the solvent was evaporated under vacuum, and the crude product was purified by column chromatography (silica gel) using 95/5 CHCl3 and C2H5OH as the eluent to afford 39 mg (85%) of 4 as a foam. 1H NMR (400 MHz, CD3OD/CDCl3) δ ppm: 3.09 (1H, dd, H5′, J5′,4′ = 2.6 Hz, J5′,5′′ = 10.1 Hz), 3.19 (1H, broad t, H5′′), 3.65 (1H, d, H2′ J2′,F = 52.5 Hz), 4.15 (1H, d, H3′, J3′,F = 15.3 Hz), 4.39–4.41 (1H, m, H4′), 6.09 (1H, dd, H1′, J1′,2′ = 1.7 Hz, J1′,F = 24.8 Hz), 7.15–7.33 (30 H, m, 2xTr), 7.51 (1H, d, H8, J8,F = 3.2 Hz). 19F NMR (CD3OD/CDCl3) δ ppm:−196.91.
2′-Deoxy-2′-fluoro-9-β-d-arabinofuranosylguanine (5, F-AraG)
A solution of compound 4 (35 mg, 0.046 mmol) in CF3COOH–CHCl3 (1:9, v/v, 0.45 mL) was kept at room temperature for 3 h. The residue was co-evaporated with toluene (350 μL), partitioned between CHCl3 and water (2 mL: 2 mL). The aqueous layer was separated, neutralized with 5% NaHCO3, and concentrated under vacuum. Product 5 (9.5 mg, 72%) was collected by filtration. 1H NMR (400 MHz, D2O) δ ppm: 3.66 (1H, dd, H5′, J5′,4′ = 5.7 Hz, J5′,5′′ = 12.4 Hz), 3.73 (1H, dd, H5′′, J5′′,4′ = 3.7 Hz,), 3.91 (1H, approximately q, H4′, J = 5.0 Hz), 4.39 (1H, dm, H3′, J3′,F = 16.6 Hz), 5.07 (1H, dt, H2′, J1′,2′ = 3.2 Hz, J2′,F = 51.4 Hz), 6.13 (1H, dd, H1′, J1′,2′ = 4.2 Hz, J1′,F = 17.5 Hz), 7.78 (1H, d, H8, J8,F = 2.6 Hz), 19F NMR (D2O) δ ppm:−198.55 (1F, dt, F2′, J2′,F = 50.4 Hz, J1′,F = J3′,F = 17.2 Hz). High resolution MS: Calcd. MNa+ for C10H12N5O4FNa : 308.0771; Found 308.0783.
2′-Deoxy-2′-[18F]fluoro-9-β-d-arabinofuranosylguanine (7, [18F]F-AraG)
No carrier-added [18F]fluoride trapped on a QMA cartridge was eluted with a solution of K2CO3 (3.5 mg) and kryptofix 2.2.2 (15 mg) in water (0.9 mL) and acetonitrile (0.1 mL). The solvent was removed under vacuum at 65°C and to the anhydrous residue was added a solution of triflate precursor (compound 2, 4–6 mg, Scheme 3) in dimethyl sulfoxide (DMSO; 0.5 mL). The mixture was heated for 45 min at 85°C. After cooling to room temperature, the reaction mixture was passed through a silica gel cartridge and eluted with 3 mL of ethyl acetate. After ethyl acetate was removed under vacuum at 35°C, the residue was diluted to 1.5 mL with acetonitrile/water (80/20, v/v), and the resulting solution was injected into a semipreparative HPLC column (Phenomenex Gemini, C18, 5 μ, 10 mm × 250 mm, 4 mL/min flow rate). The [18F]6 was collected at 32 min, and it was de-protected first by base (0.5 mL of 0.5 M NaOCH3) at 100°C for 10 min and then by acid (0.5 mL of 1 N HCl) at 100°C for 10 min. After cooling to room temperature, the resulting solution was neutralized and injected into a C18 reverse phase semipreparative HPLC column. The product [18F]F-AraG 7 was collected at 11.5 min and concentrated to dryness under vacuum at 45°C. Finally, [18F]F-AraG 7 was reconstituted in saline and passed through a 0.22-μm Millipore filter into a sterile multidose vial for biological experiments. The radiochemical yield was 7–10% (decay corrected, n = 10). The chemical and radiochemical purities of [18F]F-AraG 7 were determined by reverse phase analytical HPLC method (Phenomenex Gemini C18, 5 μ, 4.6 × 250 mm) and was more than 95% pure. The radio synthesis time was 140–160 min, and the specific activity was 0.8–1.3 Ci/μmol.
Scheme 1 shows the synthesis of 2-N-Acetyl-6-O-((4-nitrophenyl)ethyl)-9-(3,5-di-O-trityl-2-trifyl-β-d-ribiofuranosyl)guanine (2), the [18F]F-AraG precursor. Treatment of 2′,5′-di-O-trityl guanosine derivative 1 with CF3SO2Cl/DMAP afforded 2-N-acetyl-6-O-((4-nitrophenyl)ethyl)-9-(3′,5′-di-O-trityl-2′-O-trifyl-β-d-ribofuranosyl)guanine (2) in 65% yield. Scheme 2 shows the synthesis of cold F-AraG standard which was prepared according to the literature procedure . 3′, 5′-Di-O-trityl guanosine derivative 1 was converted to 6-O-((4-nitrophenyl)ethyl)-9-(3′,5′-di-O-trityl-2′-fluoro-β-d-arabinofuranosyl)guanine 3 with DAST reagent. De-protection of 3 with DBU afforded 9-(3′, 5′-di-O-trityl-2′-fluoro-β-d-arabinofuranosyl)guanine (4). Finally, de-protection of 4 by TFA afforded 2′-deoxy-2′-fluoro-9-β-d-arabinofuranosylguanine 5 (F-AraG).
Cell Uptake Assays
There are many reports on the indirect synthesis of 2′-deoxy-2′-fluoro-9-β-d-arabinofuranosylguanine (F-AraG) in which fluorine is first introduced in the arabino position at C-2, and then fluorinated sugar reacted with the purine base [22, 23, 24]. The same methodology was applied to the synthesis of PET tracers such as [18F]FEAU, [18F]FMAU, and [18F]FBAU . However, there is only one report on the direct synthesis of cold F-AraG in which fluorine is incorporated into arabino position at C-2 of the sugar by direct fluorination of an appropriately protected guanosine derivative with DAST . The synthesis of [18F]labeled F-AraG has not been reported to date. Due to the difficult synthesis of [18F]-labeled DAST and long reaction times of DAST-mediated fluorinations, it is not practical to synthesize [18F]F-AraG via the [18F]DAST method. Compound 2, the [18F]F-AraG precursor, was prepared (Scheme 1) and characterized by 1H and 19F NMR spectroscopy and high-resolution mass spectrometry. Chemical shift of H2′ changed from 4.73 ppm in 1 to 6.03 ppm in 2 due to the electronegativity of the trifyl group at C2′ position. 19F NMR showed a singlet at −75.00 ppm which is consistent with the chemical shift of sugar triflates. Similar chemical shift trends were observed for the synthesis of adenosine triflate . To the best of our knowledge, precursor 2 is new and has been synthesized for the first time in our laboratory (provisional patent filed). Also, for the first time, we have synthesized [18F]F-AraG (7, Scheme 3) via a direct fluorination of 2 with [18F]KF/K.2.2.2 in DMSO in 7–10% radiochemical yield (decay corrected) with a specific activity of 0.8–1.3 Ci/μmol. The identity and radiochemical purity of 7 was confirmed by co-injection with an authentic standard compound 5 on an analytical HPLC column (Fig. 1).
To evaluate the performance of [18F]F-AraG 7 in cell culture, we performed several assays. To ascertain the ability of cells to uptake [18F]F-AraG, we exposed the CCRF-CEM cell line (acute lymphoblastic T leukemia cells, unactivated) and primary T cells to [18F]F-AraG. Fig. 2 shows the uptake of [18F]F-AraG by CCRF-CEM cell and indicates that [18F]F-AraG uptake is dose-dependent. These data also support the finding that a majority of the [18F]F-AraG is taken up by cells within the first hour of exposure. The rapid uptake is necessary if [18F]F-AraG, with an isotope half-life of 110 min, is eventually going to prove efficacious as a PET tracer. Having ascertained that lymphoblastic T cell lines will take up [18F]F-AraG, we then looked to see if primary T cells, non-neoplastic T cells derived from normal mouse tissue, would also uptake [18F]F-AraG. Fig. 3 represents the data of two independent experiments indicating that non-neoplastic but activated primary T cells will take up [18F]F-AraG to an appreciable level. The increased uptake of [18F]F-AraG by activated T cells may enable one to utilize [18F]F-AraG as a PET tracer in the detection of graft versus host disease (GVHD). GVHD is predominantly a T cell-driven disease, and the ability to detect aberrantly activated T cells by PET may facilitate an early diagnosis of GVHD in patients without invasive procedures. As AraG has been reported to induce neurotoxic side effects in some patients at therapeutic serum levels (~150 μM) , we chose to utilize doses lower than reported therapeutic levels of AraG in our assays to optimize [18F]F-AraG as a tracer for PET while avoiding the potential neurotoxicity in PET patients.
For the first time to the best of our knowledge, [18F]F-AraG has been successfully synthesized. This was accomplished by a direct fluorination method. This approach can potentially be used for the synthesis of other important PET tracers such as [18F]FEAU, [18F]FMAU, and [18F]FBAU which are currently synthesized by multiple steps and involve lengthy purification processes. Preliminary cell uptake experiments done in CCRF-CEM cells (unactivated) and activated primary T cells suggest possible application of the [18F]F-AraG as a new PET imaging agent for detection of disease of T-cell origin. In the future, we will further examine tracer uptake and metabolism in cell lines and determine the efficacy of this compound in imaging T lymphoblasts in mouse models. These data should lay a foundation for the use of this compound as an imaging agent in human disease. Finally, since AraGTP is incorporated into mitochondrial, as opposed to nuclear DNA, in future work, we would like to study whether an imaging agent targeted to mitochondrial DNA might have distinguishing characteristics from [18F]FAC that is presumably incorporated primarily into nuclear DNA.
This work was supported in part by NCI In Vivo Cellular Molecular Imaging Center grant P50 CA114747 (SSG). We also thank Dr. David Dick for the [18F] production, Dr. Frederick T. Chin for modification of a GE TRACERlab FX-FN synthetic module for radiosynthesis, and Dr. Jelena Levi for her review of the manuscript.
Conflict of interest disclosure
The authors declare that they have no conflict of interest.
- 8.Hitchcock MJ, Woods K, DeBoeck H, Ho HT (1990) Biochemical pharmacology of 2-flouro-2′, 3′-dideoxyarabinofuranosyladenine, an inhibitor of HIV with improved metabolic and chemical stability over 2′, 3′-dideoxyadenosine. Antiviral Chem Chemother 1:319–327Google Scholar
- 9.Silverman RB (1988) In mechanism-based enzyme inactivation: chemistry and enzymology, vol 1. CRC, Boca Raton, p 59Google Scholar
- 13.Kurtzberg J, Ernst TJ, keating MJ, Gandhi V, Hodge JP, Kisor DF, Lager JJ, Stephens C, Levin J, Kerenitsky T, Elion G, Mitchell BS (2005) Phase I study of 506U78 administered on a consecutive 5-day schedule in children and adults with refractory hematologic malignancies. J Clin Oncol 23:3396–3403PubMedCrossRefGoogle Scholar
- 16.Lyer M, Barrio JR, Namavari M, Bauer E, Satyamurthy N, Nguyen K, Toyokuni T, Phelps ME, Herschman HR, Gambhir SS (2001) 8-[18F]Fluoropenciclovir: an improved reporter probe for imaging HSV1-tk reporter gene expression in vivo using PET. J Nucl Med 42:96–105Google Scholar
- 24.Chu CK, Matulic-Aamic J, Huang JT, Chou TC, Bruchenal JH, Fox JJ, Watanabe KA (1989) Synthesis of 9-(-2-Deoxy-2-fluoro-b-d-arabinofuranosy)-9-H-purines and their biological activites. Chem Pharm Bull 37:336–339Google Scholar
- 26.Cooper TM (2007) Role of nelarabine in the treatment of T-cell acute lymphoblastic leukemia and T-cell lymphoblastic lymphoma. Ther Clin Risk Manage 3(6):1135–1141Google Scholar