Abstract
The aim of this study was to identify a folate receptor-α (FRα)–selective PET agent potentially suitable for the selection of patients who might profit from FRα-targeted therapies. The 6R and 6S isomers of 18F-aza-5-methyltetrahydrofolate (MTHF) were assessed regarding their binding to FRα and FRβ, expressed on cancer and inflammatory cells, respectively, and compared with 18F-AzaFol, the folic acid–based analog. Methods: FR selectivity was investigated using FRα-transfected (RT16) and FRβ-transfected (D4) CHO cells. The cell uptake of 18F-folate tracers was investigated, and receptor-binding affinities were determined with the nonradioactive analogs. In vitro autoradiography of the 18F-folate tracers was performed using RT16 and D4 tissue sections. Biodistribution studies and PET/CT imaging of the radiotracers were performed on mice bearing RT16 and D4 xenografts. Results: The uptake of 18F-6R-aza-5-MTHF was high when using RT16 cells (62% ± 10% of added activity) but much lower when using D4 cells (5% ± 2%). The FRα selectivity of 18F-6R-aza-5-MTHF was further demonstrated by its approximately 43-fold higher binding affinity to FRα (half-maximal inhibitory concentration [IC50], 1.8 ± 0.1 nM) than to FRβ (IC50, 77 ± 27 nM). The uptake of 18F-6S-aza-5-MTHF and 18F-AzaFol was equal in both cell lines (52%–70%), with similar affinities to FRα (IC50, 2.1 ± 0.4 nM and 0.6 ± 0.3 nM, respectively) and FRβ (0.8 ± 0.2 nM and 0.3 ± 0.1 nM, respectively). The autoradiography signal obtained with 18F-6R-aza-5-MTHF was 11-fold more intense for RT16 than for D4 tissue sections. Biodistribution data showed high uptake of 18F-6R-aza-5-MTHF in RT16 xenografts (81% ± 20% injected activity per gram [IA]/g 1 h after injection) but significantly lower accumulation in D4 xenografts (7.3% ± 2.1% IA/g 1 h after injection), which was also visualized using PET. The uptake of 18F-6S-aza-5-MTHF and 18F-AzaFol was similar in RT16 (53% ± 10% IA/g and 45% ± 2% IA/g, respectively) and D4 xenografts (77% ± 10% IA/g and 52% ± 7% IA/g, respectively). Conclusion: This study demonstrated FRα selectivity for 18F-6R-aza-5-MTHF but not for 18F-6S-aza-5-MTHF or 18F-AzaFol. This characteristic, together with its favorable tissue distribution, makes 18F-6R-aza-5-MTHF attractive for clinical translation to enable detection of FRα-positive cancer while preventing undesired accumulation in FRβ-expressing inflammatory cells.
The folate receptor-α (FRα) is a cell membrane–associated protein that has been used for targeted therapies in oncology (1). The FRα-expressing malignancies are mainly gynecologic cancers, such as ovarian, endometrial, and cervical tumors, but non–small cell lung cancer, triple-negative breast cancer, and kidney cancer were also reported to be frequently positive for FRα (2–6).
The use of folic acid–based radiotracers for nuclear imaging was proposed for diagnosis of FR-positive cancer and for the selection of patients who would profit from FR-targeted tumor therapies (7–10). 111In-DTPA-folate and 99mTc-EC20 suitable for SPECT were the first two folate radioconjugates tested in patients (8,11,12). Folic acid–based radiotracers bind to both FRα and FRβ, which have distinct tissue expression profiles (13,14). The FRα is present in malignant tissue (15), whereas FRβ is expressed mainly on activated macrophages involved in inflammatory diseases (16). As a result, folic acid–based radiotracers accumulate not only in tumors but also at sites of inflammation, which may result in false-positive findings due to coexisting inflammatory conditions in cancer patients.
6S/6R-5-methyltetrahydrofolates (MTHF) are reduced folate forms, in contrast to folic acid, which is the synthetic, oxidized version of folate vitamins. It was previously reported that the physiologic 6S-5-MTHF, but not the 6R-5-MTHF, binds with approximately 50-fold higher affinity to FRα than to FRβ (17,18). In a proof-of-concept study published by Vaitilingam et al., a dimethylated reduced version of 99mTc-EC20 was prepared to achieve FRα selectivity (19). Indeed, it was experimentally demonstrated that 99mTc-DMTHF, a reduced form of 99mTc-EC20, accumulated much more in tumors of mice than at sites of inflammation (19).
In view of a clinical application, folate radiotracers for PET imaging would be clearly favored (20,21). 18F-based radiotracers profit from the favorable decay characteristics of 18F (half-life, 110 min; average Eβ+, 250 keV) and the option of quantifying the accumulated activity using standard protocols that are also used for interventions with 18F-FDG (22).
Among a large number of developed 18F-folate tracers (21) only two were used in clinics. Verweij et al. reported on the use of 18F-polyethylene glycol folate for macrophage imaging in patients with rheumatoid arthritis (23), whereas 18F-AzaFol has recently been tested in a clinical phase I trial in ovarian and lung cancer patients in Switzerland (NCT0342993) (24,25). A tumor-selective 18F-folate tracer for PET that targets solely FRα and not FRβ would be essential to unambiguously identify patients who could profit from FRα-targeted therapies.
We have previously developed radiotracers based on 6S- and 6R-5-MTHF as a targeting agent (26), in which the 18F-label was integrated in the folate backbone as was the case for 18F-AzaFol (Fig. 1) (26). These 5-MTHF–based 18F-tracers accumulated to a higher extent in tumor xenografts of mice than did 18F-AzaFol, and the 6R isomer even showed a favorable excretion profile. The goal of the present study was, therefore, to investigate 18F-6R-aza-5-MTHF and 18F-6S-aza-5-MTHF with regard to their binding affinity to FRα and FRβ, in order to assess the potential of using them for tumor-selective PET imaging.
Chemical structures of 18F-AzaFol, 18F-6R-aza-5-MTHF, and 18F-6S-aza-5-MTHF, as well as folic acid, 6S-5-MTHF, and 6R-5-MTHF (26). Stereochemical nomenclature of corresponding isomers of nonfluorinated and fluorinated 5-MTHFs are inversed because of change in substituents’ priority at stereogenic center.
MATERIALS AND METHODS
Folate Derivatives
The precursors (6R- or 6S-N2-acetyl-3'-aza-2'-chloro-5-MTHF di-tert-butylester and N2-acetyl-3'-aza-2'-chlorofolic acid di-tert-butylester) for the radiofluorination and the nonradioactive 6S-5-MTHF (physiologic form) and 6R-5-MTHF (nonphysiologic form) were provided by Merck & Cie. Folic acid was obtained from Sigma-Aldrich. The nonradioactive fluoro-folates, as well as 6R-3'-aza-2'-18F-5-MTHF (18F-6R-aza-5-MTHF), 6S-3'-aza-2'-18F-5-MTHF (18F-6S-aza-5-MTHF), and 3'-aza-2'-18F-folic acid (18F-AzaFol), were synthesized at ETH Zurich according to a previously reported method (24,26). The molar activity of the radiotracers was in the range of 20–250 GBq/μmol, with commonly higher values for 18F-AzaFol than for the 5-MTHF–based radiotracers (24,26). 3H-folic acid was obtained from Moravek Biochemicals, Inc.
Cell Culture
CHO cells transfected with FRα (designated as RT16) or FRβ (designated as D4) were kindly provided by Prof. Larry H. Matherly of Wayne State University (27). The cells were cultured in folate-free minimal essential medium-α. KB cells (FR-positive human cervical carcinoma cell line, ACC-136) were obtained from the German Collection of Microorganisms and Cell Cultures (DSMZ GmbH) and cultured in folate-deficient RPMI medium. PC-3 cells (FR-negative human prostate cancer cell line, ACC-465) were also obtained from DSMZ but cultured in normal RPMI 1640 medium. The cell culture media were supplemented with 10% fetal calf serum, l-glutamine, and antibiotics.
Western Blot
Expression of FRα and FRβ in RT16 and D4 cells, respectively, was verified by Western blot analysis, whereas FRα-expressing KB and FR-negative PC-3 tumor cells were used as positive and negative controls, respectively. Cell protein extracts (30 μg/well) were separated by sodium dodecyl sulfate–polyacrylamide gel electrophoresis and transferred to a polyvinylidene difluoride membrane. After blocking with 5% bovine serum albumin (BSA) solution, the membrane was incubated with a primary anti-FRα antibody (1:625 rabbit monoclonal antibody PA5-42004; Invitrogen) or anti-FRβ antibody (1:1,000 rabbit monoclonal antibody GTX105822; GeneTex) overnight at 4°C. For signal detection, a secondary antirabbit IgG antibody functionalized with horseradish peroxidase was used together with chemiluminescent substrate. Detection of β-actin served as a loading control (anti-β-actin antibody: 1:2,000 mouse monoclonal antibody 3700 [Cell Signaling Technology] and 1:5,000 horseradish peroxidase–conjugated antimouse IgG 7076S [Cell Signaling Technology]). The expression of FRα and FRβ on RT16 and D4 cells, respectively, was assessed by comparison of the signal with that of KB cells and shown on one representative Western blot.
Cell Internalization of 18F-6R/6S-Aza-5-MTHF and 18F-Azafol
Cell uptake of 18F-6R-aza-5-MTHF, 18F-6S-aza-5-MTHF, and 18F-AzaFol was determined as previously reported (26,28). In brief, RT16 and D4 cells were seeded in poly-l-lysine–coated 12-well plates to form confluent monolayers overnight. The cells were incubated with the 18F-folate radiotracer (∼200 kBq; 25 μL) with or without excess folic acid (∼100 μM) for 3 h at 37°C. The results were expressed as percentage of total added activity. The statistical significance was assessed using a two-way ANOVA with a Tukey multiple comparisons posttest using GraphPad Prism software (version 7.0). A P value of less than 0.05 was considered statistically significant.
FRα- and FRβ-Binding Affinity (Half-Maximal Inhibitory Concentration [IC50] Values)
The IC50 values for FRα and FRβ were determined in displacement experiments using nonradioactive fluoro-folates (19F-6R-aza-5-MTHF, 19F-6S-aza-5-MTHF, and 19F-AzaFol) and 3H-folic acid as previously reported (24). The binding affinities of the corresponding nonfluorinated analogs (6R-5-MTHF, 6S-5-MTHF, and folic acid) were determined for comparison. The folate derivatives of interest were applied in a concentration range of 5 pM–50 μM (Supplemental Table 1; supplemental materials are available at http://jnm.snmjournals.org). The FR-bound fraction of 3H-folic acid was measured using a β-counter (Packard Bioscience Cobra II). The IC50 values were determined by nonlinear regression analysis of displacement curves obtained from at least three independent experiments, using GraphPad Prism software. For comparison, the relative affinities of 19F-6S-aza-5-MTHF and 19F-6R-aza-5-MTHF were presented as a percentage of the binding affinity determined for 19F-AzaFol (set as 100%). The receptor-binding affinities of 6R-5-MTHF and 6S-5-MTHF were expressed relative to the determined binding affinity of folic acid (set as 100%).
Autoradiography Studies
Autoradiography studies were performed using tissue sections of RT16, D4, KB, and PC-3 xenografts as previously reported (29). The sections were incubated in Tris-HCl buffer containing 0.25% BSA. After removal of the buffer, the sections were incubated for 1 h at room temperature with the 18F-folate radiotracers (150 kBq/100 μL) in Tris-buffer containing 1% BSA with or without addition of excess folic acid (100 μM). Autoradiographic images were obtained using a storage phosphor system (Cyclone Plus; Perkin Elmer) and quantified using OptiQuant software (version 5.0). The signals obtained from RT16, D4, and PC-3 xenograft sections were normalized to the signal obtained from a KB xenograft section (set as 100%). The resulting values of 18F-6R-aza-5-MTHF and 18F-6S-aza-5-MTHF were expressed relative to the signal of 18F-AzaFol, which was set as 100%. Representative images were prepared using ImageJ (version 1.52d).
Immunohistochemistry
After deparaffinization, rehydration, and antigen retrieval of the xenograft sections (RT16, D4, and KB), the slides were treated with 3.5% hydrogen peroxide, followed by endogenous biotin blockade using avidin solution (avidin/biotin blocking kit SP-2001; Vector Laboratories) in an aqueous solution of 3% BSA and normal 5% goat serum. The sections were incubated with an anti-FRα antibody (PA5-42004; Invitrogen) and an anti-FRβ antibody (GTX105822; GeneTex) diluted 1:800 and 1:400, respectively, in biotin solution mixed with 3% BSA at 4°C overnight. Afterward, the sections were incubated with the secondary antibody (goat antirabbit IgG, BA-1000; Vector Laboratories) diluted 1:1,000 and 1:500 in phosphate-buffered saline containing 3% BSA. Signal was visualized using reagents of commercial kits (Vectostain Elite ABC-horseradish peroxidase kit, peroxidase PK-6100, and 3,3′-diaminobenzidine substrate kit peroxidase (horseradish peroxidase) SK-4100; Vector Laboratories), followed by counterstaining using hematoxylin. After tissue dehydration, the sections were fixed with xylene and images were taken using a light microscope (Axio Lab.A1; Zeiss).
In Vivo Studies
All applicable international, national, and institutional guidelines for the care and use of laboratory animals were followed. The studies were performed according to the guidelines of the Swiss Regulations for Animal Welfare after ethical approval by the Cantonal Committee of Animal Experimentation and permission by the responsible authorities. Female severe combined immunodeficient CB17 mice were purchased from Charles River Laboratories. All animals were fed with a folate-deficient rodent diet (ssniff Spezialdiäten GmbH). The mice were inoculated with RT16 cells (6 × 106 cells in 100 μL of phosphate-buffered saline) on the right shoulder and D4 cells (6 × 106 cells in 100 μL of phosphate-buffered saline) on the left shoulder. Biodistribution and PET/CT imaging studies were performed 8–10 d later.
Biodistribution Studies
The respective 18F-folate radiotracer (∼5 MBq, 100 μL, ∼0.2 nmol) was injected into the lateral tail vein, and the mice were sacrificed at 1 or 3 h after injection. Selected tissues and organs were collected and weighed, and the activity was measured using a γ-counter (Wallac Wizard 1480; Perkin Elmer). The results for 3–4 mice per time point were listed as a percentage of the injected activity (IA) per gram of tissue mass, using counts of a standard solution (5% IA) measured at the same time. The datasets were analyzed for significance using a one-way ANOVA with a Tukey multiple comparison posttest using GraphPad Prism software. A P value of less than 0.05 was considered statistically significant.
PET/CT Imaging Studies
PET/CT scans were performed using a small-animal PET/CT scanner (G8; Perkin Elmer (30)) as previously reported (26). After emptying the urinary bladder of the mice, they were anesthetized with a mixture of isoflurane (1.5%–2.0%) and oxygen for PET/CT acquisitions. Static whole-body PET scans of 10-min duration were obtained at 1 and 3 h after intravenous injection of 18F-folate radiotracers (5 MBq, ∼0.2 nmol, 100 μL), followed by a CT scan of 1.5 min. The acquisition of the data and their reconstruction was performed using the G8 PET/CT scanner software (version 2.0.0.10). The images show one representative example for each radiotracer, prepared using VivoQuant postprocessing software (version 3.5; inviCRO Imaging Services and Software).
RESULTS
Uptake of 18F Radiotracers in FRα-Positive RT16 and FRβ-Positive D4 Cells
Western blot analysis unambiguously confirmed FRα expression on RT16 cells and FRβ expression on D4 cells by detection of the bands at 38 and 29 kDa, respectively (Fig. 2A). The quantified signal for FRα on RT16 cells and FRβ on D4 cells was in the range of 20%–45% of the FRα signal of KB cells. Only a weak signal of 6%–7% was obtained for FRα on D4 cells and FRβ on RT16 cells. This was in the same range as the signal for PC-3 cells (5%–6%), which are FR-negative (Supplemental Fig. 1) (31).
(A) Western blot analysis of FRα expression in RT16 cell lysates and FRβ expression in D4 cell lysates (top panel). β-Actin staining as loading control (bottom panel). (B) Cell uptake and internalization of 18F-6R-aza-5-MTHF, 18F-6S-aza-5-MTHF, and 18F-AzaFol in RT16 and D4 cells after 3 h of incubation at 37°C. Unspecific binding of radiotracers (blocked) was determined by coincubation of cells with folic acid (100 μM). Results are presented as average ± SD (n = 3).
The uptake of 18F-6R-aza-5-MTHF (62% ± 10% of total added activity) and 18F-6S-aza-5-MTHF (64% ± 15%) into RT16 cells was in the same range as for 18F-AzaFol (52% ± 4%) after a 3-h incubation period (P > 0.05). The internalized fractions of all three radiotracers were in the range of 11%–13% of total added activity (P > 0.05). FRα-specific uptake was confirmed by complete blockade of radiotracer uptake in cell samples preincubated with excess folic acid (Fig. 2B). Experiments performed on D4 cells showed significantly lower uptake of 18F-6R-aza-5-MTHF (5% ± 2%; P < 0.05) than of 18F-6S-aza-5-MTHF (70% ± 7%) or 18F-AzaFol (61% ± 14%), which were both in the same range (P > 0.05). Coincubation of D4 cells with excess folic acid reduced the uptake of all radiotracers to background levels (<0.1%) (Fig. 2B).
FRα and FRβ Binding Affinities
The binding affinity of 19F-6R-aza-5-MTHF to FRα was approximately 40-fold higher (IC50 value, 1.8 ± 0.1 nM) than that to FRβ (77 ± 27 nM). In contrast, the binding affinity of 19F-6S-aza-5-MTHF to both FR isoforms was similar (2.1 ± 0.4 nM and 0.8 ± 0.2 nM, respectively), and the same held true also for 19F-AzaFol (0.6 ± 0.3 nM and 0.3 ± 0.1 nM, respectively). These findings corresponded well with the determined values of the respective nonfluorinated analogs, demonstrating an approximately 70-fold higher binding of 6S-5-MTHF to FRα than to FRβ but equal binding affinities to both FR isoforms for 6R-5-MTHF and folic acid (Table 1; Supplemental Fig. 2).
FR-Binding Affinities of Fluorinated and Nonfluorinated (Aza)-5-MTHF Derivatives Relative to 19F-Azafol or Folic Acid, Respectively (Set as 100%), Determined with FRα-Expressing RT16 Cells and FRβ-Expressing D4 Cells
Determination of the FRα-binding affinity of 19F-6R-aza-5-MTHF and 19F-6S-aza-5-MTHF relative to 19F-AzaFol (set as 100%) revealed a lower value for the reduced folates (29%–34%), which was in line with the results obtained for nonfluorinated 6S-5-MTHF and 6R-5-MTHF relative to folic acid (28%–44%). The relative binding affinity to FRβ was 300-fold lower for 19F-6R-aza-5-MTHF than for 19F-AzaFol, whereas the FRβ-binding affinity of 19F-6S-aza-5-MTHF was only slightly reduced, which corresponded well with the binding affinities of the nonfluorinated folates to FRβ relative to folic acid (Table 1).
Autoradiography Studies Using 18F-Folate Tracers
In vitro autoradiography studies revealed a similar signal intensity on RT16 xenograft sections for 18F-6R-aza-5-MTHF (74% ± 28%) and 18F-6S-aza-5-MTHF (89% ± 33%), which was comparable to the signal of 18F-AzaFol (set as 100%) (Fig. 3). A similar binding pattern was obtained with FRα-positive KB xenograft sections (Supplemental Fig. 3). The images obtained with D4 xenograft sections demonstrated, however, an over 10-fold lower signal for 18F-6R-aza-5-MTHF (7% ± 4%; P < 0.05) than for 18F-6S-aza-5-MTHF (81% ± 36%) relative to 18F-AzaFol (set as 100%). Coincubation of 18F-folates with excess folic acid to block the FR-specific binding resulted in only background signals, comparable to those obtained for FR-negative PC-3 tissue (Supplemental Fig. 3).
Signal intensities (normalized digital light units [DLU] per area [mm2]) of 18F-6R-aza-5-MTHF, 18F-6S-aza-5-MTHF, and 18F-AzaFol (set as 100%) quantified on the basis of RT16 (left) or D4 (right) autoradiography images. Bottom panel demonstrates FR blockade performed with excess folic acid.
The expression of FRα on RT16 (and KB as a positive control) and FRβ on D4 xenograft sections was verified by a positive immunohistochemical staining result using an anti-FRα antibody and anti-FRβ antibody. The nonspecific binding was determined in the absence of the primary antibody (Supplemental Fig. 4).
Biodistribution of 18F-Folate Tracers
Biodistribution studies were performed on RT16/D4 xenograft-bearing mice. At 1 and 3 h after injection of 18F-6R-aza-5-MTHF, the uptake into FRα-positive RT16 xenografts ranged from 81% to 94% IA/g and was significantly higher than the uptake into FRβ-positive D4 xenografts (7.3%–7.6% IA/g; P < 0.05) (Fig. 4). The accumulation of 18F-6S-aza-5-MTHF was in the same range for both xenograft types (RT16: 53%–122% IA/g; D4: 77%–149% IA/g), which was also the case for 18F-AzaFol (RT16: 26%–45% IA/g; D4: 28%–52% IA/g) (Fig. 4). The increased uptake of 5-MTHF–based 18F-folate tracers in FRα-positive RT16 xenografts, compared with the uptake of 18F-AzaFol, was in line with previous data obtained from KB tumor–bearing nude mice (26).
Graph representing uptake of 18F-6R-aza-5-MTHF (red), 18F-6S-aza-5-MTHF (blue), and 18F-AzaFol (green) in RT16 and D4 xenografts and in kidneys at 1 and 3 h after injection of 18F-folate radiotracers (5 MBq/mouse). p.i. = after injection.
The presented data, obtained for severe combined immunodeficient CB17 mice, showed a 3-fold lower renal retention of 18F-6R-aza-5-MTHF (11%–25% IA/g) than of 18F-6S-aza-5-MTHF (31%–41% IA/g) or 18F-AzaFol (35%–58% IA/g), as previously observed in nude mice (Supplemental Tables 2–4) (26).
PET/CT Imaging Studies Using 18F-Folate Tracers
Mice bearing RT16 and D4 xenografts on the right and left shoulders, respectively, were imaged at 1 and 3 h after injection of the 18F-folate radiotracers using preclinical PET/CT (Fig. 5). In agreement with the biodistribution data, a selective accumulation of 18F-6R-aza-5-MTHF in the RT16 xenograft was readily visualized on PET/CT images. This finding was in clear contrast to the images obtained after injection of 18F-6S-aza-5-MTHF and 18F-AzaFol, which accumulated equally in RT16 and D4 xenografts. The PET images further confirmed the previously published data that showed much lower kidney retention of 18F-6R-aza-5-MTHF than of 18F-6S-aza-5-MTHF or 18F-AzaFol (26). The RT16 xenograft-to-kidney ratios were, thus, considerably higher for the 6R isomer than for the 6S isomer, whereas the opposite held true for the xenograft-to-liver ratios.
PET/CT images of mice bearing RT16 and D4 xenografts 1 h (A–C) and 3 h (D–F) after injection of 18F-folate radiotracers (5 MBq/mouse) shown as maximum-intensity projections.
DISCUSSION
FR-targeted cancer therapies, including folic acid–drug conjugates (e.g., vintafolide (32,33)), FRα-targeted antibodies (e.g., farletuzumab (34)), or FRα antibody–drug conjugates (e.g., mirvetuximab soravtansine (35)), hold promise for the treatment of patients with FRα-positive tumors (1,8). The response to these therapies will, however, critically depend on the patient inclusion criteria, which should consider only those cases in which most lesions are FRα-positive (36). Although 18F-AzaFol visualizes FR-positive tissue on PET images (25), false-positive results may occur because of concomitant accumulation of the radiotracer in FRβ-expressing macrophages involved at sites of inflammation. Having a means at hand to provide a full picture of FRα-positive lesions in an individual patient would thus present an essential step toward the success of any FRα-targeted cancer therapy concept (1,8).
In this study, we have demonstrated that 19F-6R-aza-5-MTHF displayed significantly higher affinity to FRα than to FRβ, indicating the anticipated FRα selectivity. These findings agree with those obtained from the corresponding nonfluorinated versions of 5-MTHF, confirming the original observation of Wang et al., who reported 6S-5-MTHF to have an approximately 50-fold higher affinity to FRα than to FRβ (17,18). However, folic acid and the nonphysiologic 6R-5-MTHF, as well as their fluorinated analogs, showed equal binding to both FR isoforms.
The fact that only the 6R isomer—not the 6S isomer—of the novel aza-5-MTHF–based 18F-folate tracers bound with higher affinity to FRα than to FRβ is an essential finding. It indicates the need for diastereomerically pure folate radiotracers other than previously proposed by the development of a racemic mixture of 99mTc-DMTHF (19).
The in vivo evaluation of tumor-specific folate radiotracers in the presence of inflammation is challenging in mice, since the number of activated macrophages is commonly low and, hence, the expected signal from inflammatory sites is lower than the signal from the tumor tissue. This situation may complicate the interpretation of the results with regard to FRα selectivity. To unambiguously determine whether the 18F-folate radiotracers accumulated specifically in FRα-expressing tissue, we have established a mouse model using RT16 and D4 cells to grow xenografts of comparable volumes. Using these mice enabled the determination of FRα-selective uptake of 18F-6R-aza-5-MTHF in RT16 xenografts, whereas the accumulation in D4 xenografts was significantly lower. In contrast, 18F-6S-aza-5-MTHF and 18F-AzaFol showed comparable accumulation in both xenografts.
It is important to recognize that in addition to the FRα-selective accumulation of 18F-6R-aza-5-MTHF, this radiotracer also showed the most favorable clearance from background tissues, including the kidneys as previously demonstrated by Boss et al., who used KB tumor–bearing nude mice (26). The somewhat decreased FRα-binding affinity of the 5-MTHF–based 18F radiotracers, as compared with the affinity of 18F-AzaFol, is in line with the common knowledge that reduced folates display lower FR-binding affinity than does folic acid (15,17,18). Although the high FR-binding affinity of folic acid was postulated as a particular advantage of FR-targeting agents, we believe that 5-MTHF–based 18F radiotracers may be favorably used for this purpose as they may be more efficiently released from FR after internalization (37). This could explain the higher uptake of 18F-6R-aza-5-MTHF and 18F-6S-aza-5-MTHF than of 18F-AzaFol in RT16 and KB xenografts (26).
CONCLUSION
In this study, we have identified 18F-6R-aza-5-MTHF as a novel PET agent that accumulated in FRα-positive RT16 xenografts but not in FRβ-positive D4 xenografts grown in the same mouse. The favorable tissue distribution profile of 18F-6R-aza-5-MTHF, together with the herein-determined FRα selectivity, means a breakthrough in the field. PET imaging using 18F-6R-aza-5-MTHF might serve as a way to unambiguously identify patients who can profit from FRα-targeted therapies; thus, 18F-6R-aza-5-MTHF's potential for clinical translation deserves the highest attention.
DISCLOSURE
A patent application for 18F-based folate radiotracers has been filed by Merck & Cie, Inc., whereby Roger Schibli, Simon M. Ametamey, Silvan D. Boss, and Cristina Müller are coinventors. The project was financially supported by Merck & Cie, Inc. (Schaffhausen, Switzerland), and by the Swiss National Science Foundation (grants 310030_156803 and 310030_188968). Patrycja Guzik was financially supported by a Swiss Government Excellence Scholarship. No other potential conflict of interest relevant to this article was reported.
KEY POINTS
QUESTION: Are 5-MTHF–based PET agents specific for the tumor-associated FRα?
PERTINENT FINDINGS: The data from this preclinical study confirm that 18F-6R-aza-5-MTHF binds specifically to FRα, which makes this radiotracer useful as a tumor-specific PET agent.
IMPLICATIONS FOR PATIENT CARE: A PET imaging agent such as 18F-6R-aza-5MTHF for the unambiguous identification of FRα-expressing tumor types will be essential for the selection of patients who might profit from FRα-targeted therapies.
ACKNOWLEDGMENTS
We thank Anna Becker, Raffaella Schmid, and Fan Sozzi-Guo for technical assistance with the experiments, and we thank Annette Krämer for producing the 18F-based radiotracers at ETH Zurich.
Footnotes
Published online January 15, 2021.
- © 2021 by the Society of Nuclear Medicine and Molecular Imaging.
REFERENCES
- Received for publication October 17, 2020.
- Accepted for publication January 13, 2021.