Abstract
The aim of this study was to develop the radiosyntheses of diastereomerically pure 6R- and 6S-3ʹ-aza-2ʹ-18F-fluoro-5-methyltetrahydrofolate (MTHF) (6R-18F-1 and 6S-18F-1) using the integrated approach and to compare the in vitro and in vivo performance characteristics of both radioligands with the previously reported 3ʹ-aza-2ʹ-18F-fluorofolic acid tracer (18F-2), the oxidized form. Methods: 6R-18F-1, 6S-18F-1, and 18F-2 were radiolabeled with 18F using aromatic nucleophilic substitution reaction. In vitro cell uptake studies and binding affinity assays were performed using folate receptor (FR)-α–expressing KB cells. PET/CT imaging and biodistribution experiments were performed with KB tumor–bearing mice. Results: Reference compounds 6R-1 and 6S-1 were obtained after acidic hydrolysis of the corresponding protected intermediates 6R-3 and 6S-3 in high chemical yields (81%–87%) and chemical purities of more than 95%. 6R-18F-1, 6S-18F-1, and 18F-2 were obtained after a 2-step radiosynthetic procedure in a decay-corrected radiochemical yield of up to 5% and molar radioactivities ranging from 20 to 250 GBq/μmol. In vitro binding affinity studies using FR-α–positive KB cells gave half-maximal inhibitory concentrations of 27.1 ± 3.7 and 23.8 ± 4.0 nM for 6R-1 and 6S-1, respectively, which were higher than for the previously reported 3ʹ-aza-2ʹ-fluorofolic acid 2 (1.4 ± 0.5 nM). Comparably high cell uptake values in FR-α–expressing KB cells were found for all 3 radiofolates. In biodistribution studies, exceptionally high KB tumor uptake value of over 32% injected activity per gram of tissue for both 6R-18F-1 and 6S-18F-1 was observed at 180 min after injection, whereas for 18F-2 only 15% injected activity per gram was found in the KB tumors. Radioactivity uptake in the kidneys, liver, salivary glands, and spleen was substantially different for the 6R- and 6S-diastereoisomers and 18F-2. Excellent KB tumor visualization was found in PET/CT images with 6R-18F-1 and 6S-18F-1, both of which outperformed the corresponding oxidized 18F-2. Conclusion: We have successfully radiolabeled 6R- and 6S-3ʹ-aza-2ʹ-18F-fluoro-5-MTHF with 18F using the integrated approach. Our results suggest that both 6R- and 6S-3ʹ-aza-2ʹ-18F-fluoro-5-MTHF are promising reduced radiofolates for imaging FR-α–expressing cancers.
The folate receptor (FR)-α represents a promising target for tumor imaging since it is overexpressed on various epithelial tumor types such as ovarian, endometrial, renal, breast, lung, and colorectal cancer but shows only limited expression in healthy tissues, primarily in the kidneys (1). In the past 2 decades, a large number of folic acid–based radiopharmaceuticals have been investigated for imaging FR-positive tumor tissues. However, folic acid–based radiotracers seem to have a limitation regarding their uptake into FR-positive tumors given that only slightly increased tumor uptake values were observed when modifications were made to the chemical moieties linked to folic acid (2–5).
Our group was the first to report on the syntheses and biologic evaluation of 18F-labeled 6S- and 6R-5-methyltetrahydrofolate (MTHF) conjugates as an alternative to folic acid derivatives for targeting FR-positive tumors (6). The pendant approach, which involves the reaction of a radiolabeled prosthetic group with 5-MTHF, resulting in a radioconjugate, was used for the 18F labeling of 5-MTHF, and stable radiotracers were obtained in the presence of antioxidants. The biologic results of the study demonstrated differences in the in vivo behavior of the diastereoisomers of 5-MTHF, and the conclusion drawn from that study was that 5-MTHF–based derivatives can be used as alternative targeting molecules to image FR-positive tumor tissues.
Another strategy that can be adapted for the 18F labeling of folic acid derivatives is the integrated approach, as previously demonstrated by Ross et al. (7). This approach allows the incorporation of the 18F radiolabel directly into the folic acid–targeting molecule without the need for a prosthetic group and has the further advantage that the obtained folate derivative is structurally close to folic acid. However, it was unclear whether the integrated approach, which involved harsher reaction conditions for the 18F-incorporation step, could be used for the radiosynthesis of the chemically less stable 5-MTHF derivatives (8–10).
The aim of the present study was to assess in a first step whether the radiosyntheses of 6R- and 6S-3ʹ-aza-2ʹ-18F-fluoro-5-MTHF (6R-18F-1 and 6S-18F-1, Fig. 1) would be feasible using the integrated approach. In a second step, and more importantly, we aimed to address the question of whether 6R- and 6S-3ʹ-aza-2ʹ-18F-fluoro-5-MTHF would show comparable in vivo performance characteristics when compared with the previously reported 3ʹ-aza-2ʹ-18F-fluorofolic acid (18F-2, Fig. 1) (11). 6R-18F-1 and 6S-18F-1 are the corresponding reduced forms of 18F-2, which was synthesized using the integrated approach by our group and is currently being evaluated in a clinical trial.
The results of the study show that the radiolabeling with 18F of both diastereomers, 6R-18F-1 and 6S-18F-1, can be accomplished using the integrated approach. Furthermore, 6R-18F-1 and 6S-18F-1 outperform 18F-2 and show the highest tumor uptake ever obtained for a radiofolate.
MATERIALS AND METHODS
General
General information can be found in a previous publication (6).
Chiral High-Performance Liquid Chromatography (HPLC) Separation
The diastereomeric 1:1 mixtures of 6R- and 6S-N2-acetyl-3ʹ-aza-2ʹ-chloro-5-MTHF-di-tert-butylester (6R-4 and 6S-4) and 6R- and 6S-N2-acetyl-3ʹ-aza-2ʹ-fluoro-5-MTHF-di-tert-butylester (6R-3 and 6S-3) were provided by Merck & Cie (Supplemental Fig. 1; supplemental materials are available at http://jnm.snmjournals.org). Then the diastereomers were separated by chiral HPLC using a Reprosil 100 Chiral-NR HPLC column (Dr. Maisch GmbH) running on normal phase (isocratic, hexane/isopropanol 1:1, Supplemental Fig. 2).
Synthesis of 6R- and 6S-3ʹ-Aza-2ʹ-Fluoro-5-MTHF
6R-3 (7.00 mg, 11.1 μmoL) or 6S-3 (8.00 mg, 12.6 μmoL) was dissolved in acetone (200 μL) and 4 M HCl (500 μL) and stirred for 1 h at 70°C. After completion of the reaction, 4 M NaOH (500 μL) was added and the product was purified by semipreparative HPLC. Lyophilization of the product fractions afforded 6R-1 or 6S-1 as white solids in 87% (4.62 mg) and 81% yield (4.96 mg), respectively, in high chemical and diastereomeric purity of more than 95%. 6R-1: 1H nuclear MR (500 MHz, D2O) δ 7.85 (t, J = 8.8 Hz, 1H), 6.39 (d, J = 8.8 Hz, 1H), 4.35 (dd, J = 8.2, 4.6 Hz, 1H), 3.68 (d, J = 13.0 Hz, 1H), 3.64–3.56 (m, 1H), 3.49 (d, J = 13.0 Hz, 2H), 3.43–3.32 (m, 1H), 2.86 (s, 3H), 2.42–2.28 (m, 2H), 2.25–2.14 (m, 1H), 2.08–1.99 (m, 1H). High-resolution mass spectrometry (matrix-assisted laser desorption/ionization) calculated for C19H23FN8O6: 478.1719; found: 478.1718. 6S-1: 1H nuclear MR (500 MHz, D2O) δ 7.85 (t, J = 8.8 Hz, 1H), 6.39 (d, J = 8.8 Hz, 1H), 4.36 (dd, J = 8.0, 4.6 Hz, 1H), 3.77–3.62 (m, 2H), 3.58–3.44 (m, 3H), 2.91 (s, 3H), 2.42–2.31 (m, 2H), 2.27–2.15 (m, 1H), 2.09–1.98 (m, 1H). High-resolution mass spectrometry (matrix-assisted laser desorption/ionization) calculated for C19H23FN8O6: 478.1719; found: 478.1719.
Preparation of 6R- and 6S-3ʹ-Aza-2ʹ-18F-Fluoro-5-MTHF
Radiosyntheses of 6R- and 6S-3ʹ-aza-2ʹ-18F-fluoro-5-MTHF were performed in analogy to the radiosynthesis of 3ʹ-aza-2ʹ-18F-fluorofolic acid previously described in the literature (11). The precursor 6R- or 6S-N2-acetyl-3ʹ-aza-2ʹ-chloro-5-MTHF di-tert-butylester (6R-4 or 6S-4, 2.50 mg, 3.85 μmol) was dissolved in anhydrous dimethyl sulfoxide (400 μL) and added to the azeotropically dried 18F-fluoride-cryptate complex (35–42 GBq) (11). The reaction mixture was heated at 150°C for 17 min (Fig. 2). Then, the solution was allowed to cool to 75°C and H2O (3 mL) was added. The solution was passed through a Sep-Pak Plus tC18 cartridge (Waters, preconditioned with 5 mL of MeOH, followed by 10 mL of H2O) for trapping the intermediate 6R-18F-3 or 6S-18F-3. Unreacted 18F-fluoride was removed by rinsing the cartridge with H2O (5 mL). The labeled intermediate 6R-18F-3 or 6S-18F-3 was eluted by passing MeCN (2 mL) through the cartridge into another sealed Wheaton reactor (3 mL). MeCN was evaporated to near dryness under reduced pressure and a nitrogen stream at 105°C. The protecting groups were cleaved by adding 4 M HCl solution (1 mL) to the reactor and heating for 10 min at 60°C to afford 6R-18F-1 or 6S-18F-1. A 10 mM sodium phosphate buffer, pH 7.4, containing a 50 mg/mL solution of Na-(+)-l-ascorbate (1 mL) and 4 M NaOH (0.9 mL) was added for stabilizing the product and for neutralizing the acidic solution, respectively. The product was purified by semipreparative HPLC. The product fraction was collected and passed through a sterile filter into a sterile vial. At the end of synthesis, 350–1,600 MBq (1%–5% decay-corrected radiochemical yield) of the final radiotracers were obtained. Molar activity ranged from 20 to 250 GBq/μmol, and radiochemical purity was greater than 95%.
Circular Dichroism (CD) Spectroscopy
CD spectra were recorded with a Chirascan-Plus CD spectrometer from AppliedPhotophysics and at a wavelength ranging from 200 to 400 nm. The samples were measured with a concentration of 2 mM in CH3OH. Background was corrected using pure CH3OH before the experiment.
Determination of Distribution Coefficient
The shake-flask method was used for the determination of the distribution coefficients (logD7.4) of 6R-18F-1 and 6S-18F-1 according to a previously published procedure (12).
In Vitro Binding Affinity
The binding affinities of reference compounds 6R-1 and 6S-1, 6R- and 6S-5-MTHF, and folic acid to the FR-α were determined in a competitive in vitro binding assay on FR-positive KB cells according to a previously published procedure (11). The binding curves are available in the supplemental materials.
Cell Uptake and Internalization
Cell uptake and internalization experiments were performed with both 6R-18F-1 or 6S-18F-1 as previously described (13,14).
Preparation of Tumor Mice
Animal experiments were performed in compliance with Swiss and local laws on animal protection and approved by the Veterinary Office of Switzerland. Female CD-1 nude mice were purchased from Charles River and kept on a folate-deficient rodent diet (ssniff Spezialdiäten GmbH) starting 1 wk before the KB tumor cell inoculation. Experiments were performed 2 wk after the KB tumor cell inoculation. A cell suspension (5 × 106 cells in 100 μL of phosphate-buffered saline, pH 7.4) was inoculated into the subcutis of each shoulder.
Biodistribution Studies
Animals were injected with approximately 5 MBq (∼0.2 nmol, 100 μL) of the corresponding radiotracer via a lateral tail vein (n = 4 per time point). Blocking studies (n = 3) were performed with excess folic acid dissolved in phosphate-buffered saline, pH 7.4 (1 mg/mL), which was intravenously injected (100 μL per mouse) 2–3 min before injection of the radiotracer. Animals were sacrificed at 0.5, 1, 1.5, 2, and 3 h after injection. Organs and tissues were collected and measured in a γ-counter. The radioactivity that accumulated in organs and tissues was expressed as percentage injected activity per gram of tissue (%IA/g) and decay-corrected.
PET/CT Imaging Studies
PET/CT scans were performed using a small-animal benchtop PET/CT scanner (G8; Perkin Elmer). The energy window was set to 150–650 keV. Mice were injected intravenously with the 18F radiotracers (∼5 MBq in 100 μL, ∼0.2–0.3 nmol). For blocking studies, an excess of folic acid in phosphate buffered saline (100 μg/100 μL) was injected 2–3 min before the injection of the radiotracers.
During the scans, which lasted 10 min, the mice were anesthetized using a mixture of isoflurane and oxygen. Scans were performed at 1, 2, and 3 h after injection of the 18F radiotracers using G8 acquisition software (version 2.0.0.10). The PET scans were followed by a CT scan of 1.5 min. The images were reconstructed with maximum-likelihood expectation maximization using the software of the scanner. All images were prepared using VivoQuant postprocessing software (version 2.10; inviCRO Imaging Services and Software). A Gauss postreconstruction filter (full width at half maximum, 1 mm) was applied to the images. For presenting the PET/CT images, the scale of the images was adjusted allowing optimal visualization of the tumor tissue und kidneys.
Metabolite Studies
6R-18F-1 or 6S-18F-1 (56.8–62.2 MBq; ∼2.1–3.5 nmol) was injected intravenously into mice, and the animals were sacrificed 30 min after injection of the radiotracer. Blood, liver, and urine were collected and analyzed. Ice-cold methanol containing a 10 mg/mL concentration of 2-mercaptoethanol and 0.025% (v/v) ammonium hydroxide was added to the blood plasma, homogenized liver, and urine to precipitate the proteins (15). After centrifugation at 5,000g for 5 min at 4°C, the supernatants of the blood plasma, liver, and urine samples were analyzed by radio–ultra-performance liquid chromatography.
RESULTS
Synthesis of Nonradioactive Reference Compounds
Pure diastereomeric 6R-3 and 6S-3 were obtained after chiral separation and used as starting materials for the synthesis of 6R-1 and 6S-1. Acidic deprotection of 6R-3 and 6S-3 in acetone and purification afforded 6R- and 6S-3ʹ-aza-2ʹ-fluoro-5-MTHF in chemical yields of 81% and 87%, respectively (Fig. 3).
Determination of the Absolute Configuration of the 2 Reference Diastereoisomers 6R-1 and 6S-1 by CD
CD spectra of 6R- and 6S-5-MTHF (Supplemental Fig. 3A) and the 2 diastereomeric pure reference compounds 6R- and 6S-3ʹ-aza-2ʹ-fluoro-5-MTHF (6R-1 and 6S-1, Supplemental Fig. 3B) were measured from 200 to 400 nm (16). The resulting spectra of 6R- and 6S-5-MTHF served as reference spectra, since the stereochemistry of these 2 folates at position 6 of the pterin entity is known. Pterin was reported to have 2 typical absorption bands at around 280 nm and at 345 nm (8,17,18). In all 4 recorded CD spectra, these 2 typical absorption bands were apparent. Correlation of the elution profile and the reference CD spectra of 6R- and 6S-5-MTHF showed that the R-isomers were the first to elute from the chiral column and showed negative and positive peaks at 280 and 310 nm, respectively (Supplemental Fig. 3C). The S-isomer eluted second and exhibited positive and negative peaks at 280 and 310 nm (Supplemental Fig. 3D). The same correlation was done with 6S-5-MTHF whereby the S-diastereomer eluted as the second peak from the chiral HPLC column. As depicted in Supplemental Fig. 3D, the peak at 280 nm was positive and the peak at 310 nm was negative for both folate derivatives. On the basis of these results, we unambiguously determined the absolute configurations of 6R- and 6S-3ʹ-aza-2ʹ-fluoro-5-MTHF.
Radiochemistry
6R- and 6S-18F-1 were radiosynthesized in a 2-step reaction sequence starting from the diastereomerically pure chlorinated precursors 6R-4 or 6S-4 (Fig. 2) obtained via normal-phase chiral HPLC. The reaction sequence involved a nucleophilic aromatic substitution of the chloro leaving group by 18F-fluoride followed by cleavage of the protecting groups under acidic conditions. After semipreparative HPLC, the radiolabeled diastereoisomers 6R-18F-1 and 6S-18F-1 were obtained in a total synthesis time of 100 min, decay-corrected radiochemical yields of 1%–5%, and radiochemical purities of more than 95% with molar radioactivities ranging from 20 to 250 GBq/μmol. Coinjection of the nonradioactive reference compound 6R-1 or 6S-1 confirmed the identity of 6R- and 6S-18F-1. The distribution coefficient (logD7.4) of both 6R-18F-1 and 6S-18F-1 was −4.8 ± 0.1 (n = 3), which is in the same range as the logD7.4 value, −4.2 ± 0.1, for 18F-2 (11).
In Vitro Characterization
The binding affinities of the nonradioactive reference compounds 6R-1, 6S-1, and 2 to the FR-α was determined in a displacement assay with 3H-folic acid using KB tumor cells. The binding curves are available in the supplemental materials. The two reduced aza-folates 6R-1 and 6S-1 exhibited affinities that were similar to FR-α and comparable to 6R- and 6S-5-MTHFs (Table 1). The half-maximal inhibitory concentration values of 3ʹ-aza-2ʹ-fluorofolic acid 2 and folic acid were considerably higher than those of the 5-MTHF compounds.
Cell Uptake and Internalization with KB Cells
Cell uptake and internalization of 6R-18F-1, 6S-18F-1, and 18F-2 were investigated using FR-α–expressing KB cells (Fig. 4). A constant increase in cell uptake over time was observed for all radiofolates, resulting in a similar uptake in the range of 50%–60% of total added radioactivity after 3 h of incubation at 37°C. The internalized fraction was 22% of total added activity for 18F-2, whereas 16% and 8% were internalized for 6S-18F-1 and 6R-18F-1, respectively. FR-α–specific binding was confirmed by coincubating cells with the radiotracer and an excess of folic acid, resulting in a high inhibition of uptake to less than 1% for all 3 radiofolates.
Biodistribution
Biodistribution studies were performed with KB tumor–bearing mice at different time points after injection of 6R-18F-1, 6S-18F-1, or 18F-2, and the results are summarized in Figures 5 and 6 and Supplemental Tables 1–3.
A high tumor uptake of 13.3 ± 1.80 %IA/g for 6R-18F-1 and 11.13 ± 0.90 %IA/g for 6S-18F-1 was already observed at 0.5 h after injection (Fig. 5). Tumor uptake constantly increased over time, resulting in exceptionally high values of 32.3 ± 6.1 %IA/g for 6R-18F-1 and 34.8 ± 6.0 %IA/g for 6S-18F-1 at 3 h after injection. In contrast, tumor uptake of 18F-2 increased only slightly over time to only half the amount of radioactivity (15.0 ± 2.3 %IA/g) of the reduced folates. Injection of folic acid reduced the uptake in the tumors by 85 and 89 %IA/g for the 6R- and 6S-isomers, respectively, and by 80 %IA/g for 18F-2 at 1 h after injection. Kidney uptake was already lowest for 6R-18F-1 at 0.5 h after injection and considerably decreased over time to below 20 %IA/g. In contrast, radioactivity accumulation in the kidneys was nearly constant for 6S-18F-1 and 18F-2 over the investigated 3 h. Liver uptake was 2- and 3-fold higher for 6R-18F-1 than for 18F-2 and 6S-18F-1, respectively, and was constant over time for all radiofolates. Liver uptake could not be blocked with folic acid for any of the 3 radiofolates, suggesting unspecific uptake. FR-α–specific salivary gland uptake was approximately 3-fold higher for 6S-18F-1 and 18F-2 than for 6R-18F-1.
6R-18F-1 exhibited the highest tumor-to-kidney ratio of the 3 radiofolates at all investigated time points because of the decreased kidney uptake (Fig. 6). An exceptionally high tumor-to-kidney ratio of 1.63 ± 0.32 was determined for 6R-18F-1 at 3 h after injection. The tumor-to-liver ratio was the highest for 6S-18F-1, as can be explained by the decreased liver uptake. The tumor-to-blood ratios constantly increased over time for all 3 radiofolates, resulting in similar values 3 h after injection.
In Vivo PET Imaging
PET/CT scans of KB tumor–bearing mice were performed at 1 and 3 h after injection of 6R-18F-1, 6S-18F-1, or 18F-2 (Fig. 7). A considerably higher uptake of both reduced radiotracers was already observed in KB tumor xenografts 1 h after injection and the tumors were well visualized 3 h after injection for both 6R-18F-1 and 6S-18F-1. FR-positive kidneys were clearly visualized 1 h after injection with 6R-18F-1 but nearly invisible 3 h after injection. In contrast, high kidney accumulation was observed for 6S-18F-1 and 18F-2 at all investigated time points. Radioactivity uptake in the liver was found mainly for 18F-2 and the 6R-isomer; however, after 3 h after injection the liver was only barely visible for both radiotracers. Radioactivity uptake in the salivary glands, gallbladder, and choroid plexus was higher for the 6S-isomer than for the 6R-isomer. Blocking studies were performed by injecting an excess of folic acid before administration of the radiotracers and resulted in a remarkably reduced uptake of 6R-18F-1, 6S-18F-1, and 18F-2 in all FR-positive tissues (KB tumor xenografts, kidneys, salivary glands, and choroid plexus).
Metabolite Studies
The in vivo stability experiments revealed no radiometabolites in blood plasma, urine, or liver samples for either radiofolate. Only intact parent radiotracers were detected. The same finding was previously reported for the oxidized version of 18F-2 (11).
DISCUSSION
The nonradioactive reference compounds 6R-1 and 6S-1 could not be separated on either a normal C18 or a chiral HPLC column; therefore, the 6R- and 6S-diastereoisomers were separated using the protected intermediates 6R-3 and 6S-3, as these were well separated on the chiral HPLC column. After acidic deprotection of the protecting groups of 6R-3 and 6S-3, elucidation and assignment of the absolute configuration of the nonradioactive reference compounds 6R-1 and 6S-1 were performed using CD spectroscopy.
6R-1 and 6S-1 showed similar IC50 values to the FR-α as did the nonderivatized 6R- and 6S-5-MTHF, indicating that modifications on the pteroate moiety have no influence on the binding affinity to the FR-α (Table 1). The half-maximal inhibitory concentrations for compound 2 and folic acid were considerably lower than those for the 5-MTHF derivatives and agreed with values reported in the literature (18).
The radiosynthesis of 6R-18F-1 and 6S-18F-1 via the integrated approach was accomplished using a 2-step synthetic procedure starting with the protected chlorinated, diastereomerically pure precursor 6R-4 or 6S-4, which were obtained after chiral HPLC separation of the diastereomeric 1:1 mixture. The varying radiochemical yield of 6R-18F-1 and 6S-18F-1 can be explained by the aromatic substitution reaction, which was found to be very water-sensitive. However, a critical issue during the radiosynthesis was the chemical stability of the radiolabeled reduced folate derivatives, similar to the findings for the previously reported 18F-labeled 5-MTHF conjugates (6). High chemical decomposition of 6R-18F-1 and 6S-18F-1 was observed during the radiolabeling, especially after the cleavage of the protecting groups. This chemical instability of the 3ʹ-aza-5-MTHF tracers was revealed as a major disadvantage compared with the corresponding folic acid product, 18F-2, which was stable under the same reaction conditions. However, in the presence of the antioxidant sodium ascorbate, 6R-18F-1 and 6S-18F-1 were found to be stable over 3 h after isolation. No additional thiol-bearing antioxidant was required for stabilization as opposed to 18F-labeled 5-MTHF conjugates, where a combination of both sodium ascorbate and l-cysteine was crucial for their stabilization (6). Optimization of the reaction conditions has been planned for the future. The high variation of the molar activity can be explained by the challenging separation of the final radioproduct and the chlorinated side-product.
Cell experiments with FR-α–expressing KB cells revealed high specific uptake and internalization of all 3 aza-folate radiotracers, suggesting that the lower binding affinity of the 5-MTHF derivatives has no influence on the binding to FR-α (Fig. 4).
6R-18F-1 and 6S-18F-1 showed high in vivo stability, which was critical for in vivo experiments, including biodistribution and PET studies. Because 5-MTHF is the predominant folate circulating in the body, accounting for approximately 98% of folates in human plasma, high in vivo stability was expected (15). In biodistribution studies, both 6R-18F-1 and 6S-18F-1 showed exceptionally high KB tumor uptake of over 32 %IA/g at 180 min after injection, representing the highest tumor uptake value for a folate-based radiopharmaceutical ever reported in the literature (Supplemental Tables 1 and 2). Excellent tumor visualization for 6R-18F-1 and 6S-18F-1 was also evident in the PET imaging studies, whereas background activity was low, suggesting that the 3ʹ-aza-5-MTHF tracers were not transported by the ubiquitously expressed reduced folate carrier (19). For 18F-2, only 15 %IA/g was found in the tumors, resulting in a poor visualization of the tumors in the PET images compared with both reduced radiofolates (Fig. 7). A surprising finding of this study was that the accumulation of 6R-18F-1 and 6S-18F-1 in the KB tumor xenografts considerably increased over time, whereas for 18F-2 the uptake was nearly constant over time (Fig. 5). This was not the case for 18F-2, as radioactivity uptake in the kidneys and liver was nearly constant over time. The different binding affinities of the 5-MTHF and the folic acid derivatives to the FR, or different transport characteristics by other transporters expressed in the kidneys and the liver, could be a potential reason for these considerable differences in radioactivity accumulation. However, further experiments are needed to investigate the transport characteristics of the aza-folates. 6R-18F-1 exhibited the highest tumor-to-kidney ratio (1.63) of all 3 radiofolates because of its low kidney uptake—a major advantage compared with the folic acid–based radiotracer (0.35). Low kidney uptake of 6R-18F-1 was also visible in the PET images, suggesting a fast washout from the kidneys. In contrast, liver uptake was higher for 6R-18F-1 than for 6S-18F-1 or 18F-2, and these results agreed with the PET images. A possible explanation might be the higher hepatobiliary clearance of the R-isomer than the S-isomer or 18F-2. From an imaging point of view, the lower uptake of the S-isomer in the liver is advantageous because it would permit imaging of FR-positive tumors near the liver.
CONCLUSION
In this study, we showed that 6R-18F-1 and 6S-18F-1 can be directly radiolabeled with 18F using the integrated approach. Both 18F-labeled 5-MTHF derivatives exhibited a tumor uptake of over 32 %IA/g, which is the highest tumor uptake ever obtained for a radiofolate. These results demonstrate that both 6R- and 6S-3ʹ-aza-2ʹ-18F-fluoro-5-MTHF are highly promising radiopharmaceuticals for clinical trials to image FR-α–positive tumors.
DISCLOSURE
This project was financially supported by Merck & Cie (Schaffhausen, Switzerland) and by the Swiss National Science Foundation (grant 310030_156803). No other potential conflict of interest relevant to this article was reported.
Acknowledgments
We thank Susan Cohrs, Claudia Keller, Bruno Mancosu, Patrycja Guzik, and Jasmin Egli for their support and technical assistance.
Footnotes
Published online Jul. 24, 2018.
- © 2019 by the Society of Nuclear Medicine and Molecular Imaging.
REFERENCES
- Received for publication April 19, 2018.
- Accepted for publication June 14, 2018.