Radiobrominated benzimidazole-quinoline derivatives as Platelet-derived growth factor receptor beta (PDGFRβ) imaging probes

Platelet-derived growth factor receptor beta (PDGFRβ) affects in numerous human cancers and has been recognized as a promising molecular target for cancer therapies. The overexpression of PDGFRβ could be a biomarker for cancer diagnosis. Radiolabeled ligands having high affinity for the molecular target could be useful tools for the imaging of overexpressed receptors in tumors. In this study, we aimed to develop radiobrominated PDGFRβ ligands and evaluate their effectiveness as PDGFRβ imaging probes. The radiolabeled ligands were designed by modification of 1-{2-[5-(2-methoxyethoxy)-1H- benzo[d]imidazol-1-yl]quinolin-8-yl}piperidin-4-amine (1), which shows selective inhibition profile toward PDGFRβ. The bromine atom was introduced directly into C-5 of the quinoline group of 1, or indirectly by the conjugation of 1 with the 3-bromo benzoyl group. [77Br]1-{5-Bromo-2-[5-(2-methoxyethoxy)-1H-benzo[d]imidazol-1-yl]quinoline-8-yl}piperidin-4-amine ([77Br]2) and [77Br]-N-3-bromobenzoyl-1-{2-[5-(2-methoxyethoxy)-1H-benzo[d]imidazol-1-yl]quinolin-8-yl}-piperidin-4-amine ([77Br]3) were prepared using a bromodestannylation reaction. In a cellular uptake study, [77Br]2 and [77Br]3 more highly accumulatd in BxPC3-luc cells (PDGFRβ-positive) than in MCF7 cells (PDGFRβ-negative), and their accumulation was significantly reduced by pretreatment with inhibitors. In biodistribution experiments, [77Br]2 accumulation was higher than [77Br]3 accumulation at 1 h postinjection. These findings suggest that [76Br]2 is more promising for positron emission tomography (PET) imaging of PDGFRβ than [76Br]3.

Br (3.4 MeV) could be a disadvantage in terms of image resolution and absorbed radiation dose 16 , the relatively longer half-life (t 1/2 = 16.1 h) of 76 Br than those of 11 C, 13 N, 15 O, and 18 F could be advantageous. Values of physical properties of the brominated derivatives, such as stability, molecular size, lipophilicity, and solubility, range between those of fluorinated and iodinated derivatives 11 . Thus, the steric effect of radiobromine might be smaller than that of radioiodine 17 .
To date, several types of probes targeting PDGFRβ have been developed for cancer imaging in nuclear medicine. Askoxylakis et al. reported a 125 I-labeled dodecapeptide targeting PDGFRβ with an inhibitory concentration 50 (IC 50 ) value of 1.4 µM 9 . Tolmachev et al. reported PDGFRβ specific affibodies labeled with 111 In or 68 Ga, which exhibited a high tumor-to-blood ratio and good IC 50 . The labeled affibodies clearly visualized PDGFRβ-expressing U-87 MG xenografts in mice 18,19 . The peptide and affibodies bind to the extracellular part of PDGFRβ. Contrastingly, intracellular domains, especially adenosine triphosphate (ATP)-binding sites, could also be a promising target for PDGFRβ imaging. Radiolabeled tyrosine kinase inhibitors (TKIs), such as the 11 C-labeled imatinib 20 , 18 F-labeled dasatinib 21 , 11 C-labeled sorafenib (tumor uptake in RXF393 xenograft mice about 2.52 ± 0.33%ID/g) 22 , and 125 I-labeled sunitinib 23 , have been synthesized and evaluated in order to develop probes targeting the ATP-binding site in nuclear medicine. However, not only do they all bind PDGFRs, but also bind other RTKs, such as vascular endothelial growth factor receptors 2 (VEGFR2), BCR-ABL1, and c-KIT.  (Fig. 1), as PDGFRβ imaging probes 24 . In both in vitro and in vivo experiments, [ 125 I]4 showed a significantly higher tumor uptake with high PDGFRβ expression than [ 125 I] 5.

methoxyethoxy)-1H-benzo[d]imidazol-1-yl]quinolin-8-yl}-piperidin-4-amine
(3), were designed and synthesized. Bromine was incorporated into 1 instead of iodine in 4 and 5, and their affinities for PDGFRβ were examined. Although we are interested in developing 76 Br-labeled PDGFRβ PET imaging  probes, 77 Br was used in this initial study because it has a relatively longer half-life (t 1/2 = 57 h) and using a different radioisotope of the same element in the chemical compound does not alter its overall biological profile. Radiobrominated compounds, [ 77 Br]2 and [ 77 Br]3 ( Fig. 1), were synthesized and the in vitro and in vivo experiments were performed.

Results
Synthesis of the reference compounds and their precursors. The nonradioactive brominated reference compounds, 2 and 3, were synthesized as described in Figs 2 and 3. Compound 2 was synthesized by direct bromination at C-5 position of the quinoline group of 1 using N-bromosuccinimide (NBS) (Fig. 2). Compound 3 was obtained by acylation of the amino group of 1 using SBrB (Fig. 3). Tributyltin precursors (6 and 7) were synthesized as described previously 24 .
Cell viability assays. The binding affinity of 2 and 3 to the ATP-binding site in PDGFRβ was evaluated using PDGFRβ overexpressed TR-PCT1 cells. Cells were treated with 1-1000 nM of synthesized ligands, 1, 2, or 3. As seen in Fig. 4, 3 showed the similar effects compared to 1. 2 was more effective in decreasing the viability of TR-PCT1 cells than 1.
Radiolabeling. Radiobrominated  , and the precursors (6 and 7) were completely separated using RP-HPLC. N-chlorosuccinimide (NCS) was used as an oxidizing agent in these syntheses. The   Cell uptake experiments. Figure  Blocking studies using 1, 2, and 3 were performed (Fig. 6). Pretreatment of an excess amount of 1 or 2 decreased uptakes of [ 77 Br]2 in BxPC3-luc cells. The uptake of [ 77 Br]3 was inhibited by a pretreatment using 1 or 3 in the same way.
Biodistribution experiments. In this study, to minimize both the number of mice consumed in the experiment and the potential for experimental errors, we co-injected radiobrominated and radioiodinated compounds into mice 30,31 . The biodistribution of [ 77 Br]2 and [ 125 I]4 in ddY mice were summarized in Table 1. Table 2 lists the biodistribution of [ 77 Br]3 and [ 125 I]5 in ddY mice. High radioactivity in the liver, small intestine, and large intestine was observed. At 24 h postinjection of the radiobrominated and radioiodinated compounds, radioactivity in feces was much higher than that in urine, suggesting hepatobiliary excretion as the main excretion pathway for both radiotracers.
We investigated the biodistribution of the radiotracers in  Table 4). Accumulation of [ 77 Br]2 in the tumor at 1 h after administraion was 1.61% injected dose (ID)/g, which was significantly higher than that of [ 77 Br]3 (1.15%ID/ gram). Figure 7 displays the blocking effect of the pretreatment with 1 on the tumor accumulation of [ 77 Br]2 at 1 h postinjection. The tumor uptake of both the control and blocking groups was 1.61 ± 0.24 and 0.96 ± 0.20% ID/g, respectively. Thus, the figures are shown as not only % ID/g but also as tumor-to-blood ratio. The pretreatment with 1 significantly reduced the tumor uptake and the tumor-to-blood ratio of [ 77 Br]2 at 1 h postinjection.

Discussion
In our study, we performed cell viability assays of 1, 2, and 3, to evaluate the effect of the structural changes on the affinity between the ligand and the molecular target, PDGFRβ. Similar affinity was exhibited by 1 and 3, whereas 2 displayed a higher affinity than 1 (Fig. 4). The larger size of the bromine compared with hydrogen may have contributed to this result. In accordance with 4, incorporating bromine into 1 could increase the affinity of 2 for PDGFRβ. Although the results of the cell viability assays showed the in vitro affinity of the iodinated compound 4 Figure 4. Cell viability after exposure 1, 2, and 3 by WST-8 assay. Data were presented as mean ± SD for three samples. Significance was determined using a one-way ANOVA followed by Tukey's post hoc test ( * p < 0.01, ** p < 0.001).
The comparison of chloramine-T, peracetic acid, and NCS in this study showed NCS was the best oxidizing agent for the bromination of 1 through an oxidative bromodestannylation reaction under non-carrier added condition (data not shown). When NCS was used, undesired radioactive peaks had almost disappeared. Previously, we reported the preparation of a 77 Br-labeled sigma-1 receptor ligand, (+)-[ 77 Br]pBrV, by using oxidative bromodestannylation with chloramine-T; its radiochemical yield was 53% 30 . Hanaoka et al. performed 77 Br labeling of α-methyl-phenylalanine by bromodestannylation reaction with NCS, and its radiochemical yield was approximately 60% 32 . In these reports, using the same corresponding precursors, the radiochemical yields for radiobromine labeling were lower than those for radioiodine labeling. In this study, we obtained [ 77 Br]2 and [ 77 Br]3 with prominently high radiochemical yields (95% and 83%, respectively), and the radiochemical yields were comparable to those of the corresponding radioiodinated compounds at 95% for [ 125 I]4 and 85% for [ 125 I]5 24 .
In cellular uptake experiments for [ 77 Br]2 and [ 77 Br]3, both radiolabeled compounds more highly accumulated in BxPC3-luc cells than in MCF7 cells, and [ 77 Br]2 showed higher accumulation in BxPC3-luc cells (PDGFRβ-positive) than [ 77 Br]3 (Fig. 5). This result was consistent with the cell viability assay, in which 2 showed a higher affinity for PDGFRβ than 1 and 3. This result also agreed with the in vivo experiment in which [ 77 Br]2 showed higher accumulation in the BxPC3-luc tumor than [ 77 Br]3 ( Table 2). The difference in the lipophilicity may be an important factor that contributed to this result. Moreover, the excess amount of PDGFRβ ligand can reduce [ 77 Br]2 uptake in PDGFRβ-positive tumor cells (Fig. 6) and in the in vivo blocking experiments using tumor-bearing mice (Fig. 7). [ 77 Br]2 uptake in tumors should be PDGFRβ specific, and [ 77 Br]2 should bind to the ATP-binding site of PDGFRβ in the tumor cell.
Although free iodide ions generated by the deiodination of the radioiodine-labeled compounds highly accumulate in the stomach and thyroid. However, the biodistribution of free bromide ions is much different. Because the free bromide ions accumulate in blood and are retained for a long time 32,33 , the radioactivity in the blood can be used as an in vivo stability index for radiobromine-labeled compounds. As summarized in Tables 1 and 2 5. Lower lipophilicity and/or smaller molecule size of the radiobrominated compounds compared with the corresponding radioiodinated compounds might contribute to these results. The tumor-to-blood ratio of radioactivity at 1 h postinjection was 2.8 for [ 77 Br]2 and 1.9 for [ 125 I]4, indicating that the radiobrominated 1 derivatives are more promising than radioiodinated 1 derivatives. However, the tumor uptake of a radiobrominated compound was not high enough as an appropriate probe for PDGFRβ imaging, and further modification is still needed.
In conclusion, [ 77 Br]2 and [ 77 Br]3 were easily prepared using a bromodestannylation reaction without carrier addition in excellent radiochemical yields and high radiochemical purities. Furthermore, 76 Br could be incorporated into 1 instead of 77 Br. Although this study suggests that radiobrominated 2 has more promising property for PET imaging of PDGFRβ than radioiodinated 4, in clinical application of the radiobrominated compound as a PDGFRβ-targeted PET imaging agent, structural modification would be required to improve tumor uptake and tumor-to-background ratios.

Synthesis of reference compounds and precursors.
Intermediate, reference compounds, 1, and corresponding tin precursors were synthesized according to the reported studies, with a slight modification 24 .
Radiolabeling. Radiotracers, [ 77 Br]2 and [ 77 Br]3, were prepared by a bromodestannylation reaction using the corresponding tin precursors (6 or 7) and NCS as an oxidizing agent. The radiolabeled compounds were purified by reversed phase (RP)-HPLC performed with a Cosmosil 5C 18 -MS-II column (4.6 × 150 mm; Nacalai Tesque) at the flow rate of 1 mL/min with a gradient mobile phase of 70% methanol in water with 0.05% TEA to 90% methanol in water with 0.05% TEA for 20 min. The column temperature was 40 °C. Radiochemical yield and radiochemical purity were calculated by counting radioactivity using an auto well gamma counter.  24 . Cells were washed with ice-cold 1 mL of PBS and dissolved using 0.5 mL of 1 M NaOH and wells were washed with 1 M NaOH aqueous solution (0.5 mL). The radioactivity of pooled basic fractions was counted by a gamma counter. A range between 16 and 71 keV was used for measuring 125 I and between 95 and 700 for 77 Br. When radioactivity of 77 Br was counted, the crossover of 125 I activity into the 77 Br channel was negligible. More than one month after the experiment, the radioactivity of 125 I was determined because at that time 77 Br has been decayed and its radioactivity was negligible. The protein in the cell was quantified by a BCA Protein Assay Kit. All data were expressed as %dose/µg protein.
For in vitro blocking experiment, inhibitor (1, 2, or 3 with final concentration 10 µM) in 1 mL of medium without FBS was added to wells containing 2 × 10 5 cells/well. After 10 min incubation, [ 77 Br]2 or [ 77 Br]3 (3.7 kBq/ well) in 1 mL of medium without FBS was added to each well. Radioactivity and protein concentration in the cells were determined by the same method above-mentioned.
Competitive binding assay using BxPC3-luc cells. BxPC3-luc cells in medium containing 10% FBS and antibiotics were seeded on 24-well plates (50,000 cells/wells) and incubated for 24 h in a 5% CO 2 incubator at 37 °C. Nine concentrations of displacing nonradiolabeled ligands (1, 2 and 3) (ranging from 1 pM to 1 mM) and [ 125 I]4 in medium without FBS were incubated at 37 °C for 4 h. After washing the cells twice using 250 µL of ice-cold PBS, the unbound radioligand was removed. The cells were dissolved using 250 µL of 1 M NaOH and wells were washed with 250 µL of 1 M NaOH. The bound radioactivity was determined using a gamma counter. Tissues in mice were resected and weighed. The radioactivity of the tissues was counted by a gamma counter and counts were corrected for background radiation. The data were expressed as percent injection dose per gram tissue (%ID/g). Statistical analysis. All data were statisticaly analyzed using GraphPad 5.0 software (La Jolla, CA, USA) and expressed as mean ± standard deviation (SD). Significance for in vitro blocking experiments was determined by a one-way analysis of variance (ANOVA) followed by Dunnett's post hoc test compared to the control group. Significance in cell viability assays was determined by ANOVA followed by Tukey's post hoc test. IC 50 values for the binding assay were calculated by nonlinear regression. Significant differences in biodistribution experiments between [ 77 Br]2 and [ 77 Br]3 groups were determined using unpaired Student's t-test. Significance for in vivo blocking studies between control and blocking groups were determined by unpaired Student's t-test. Results were considered statistically significant at p < 0.05.