2022 Volume 45 Issue 1 Pages 94-103
Our previous studies identified that nimesulide analogs which bear a methoxy substituent at the para-position of the phenyl ring could be potential radiotracer candidates for detecting disorders related to cyclooxygenase-2 (COX-2) expression and activity in vivo using positron emission tomography (PET) in the brain. The present study was conducted to evaluate the in vivo characteristics of 11C-labeled para-methoxy nimesulide ([11C]1d) as a brain COX-2-targeted imaging agent compared to other isomeric methoxy analogs of nimesulide ([11C]1b and [11C]1c). [11C]1b–d were synthesized with reasonable yield and purity by the methylation of the O-desmethyl precursor with [11C]methyl triflate in the presence of NaOH at room temperature. We performed in vivo biodistribution analysis, brain PET imaging, ex vivo autoradiography, and metabolite analysis in mice. The uptake of [11C]1b–d was lower in the brain than in other tissues, including in the blood, and both [11C]1c and [11C]1d were rapidly metabolized. However, [11C]1d showed a small, but significant, specific signal and heterogeneous distribution in the brain. In vivo evaluation suggested that [11C]1d might correlate with COX-2 expression in the brain. Given its instability in vivo, [11C]1d seems unsuitable as a brain-COX-2 radioimaging agent. Further structural refinement of these radiotracers is necessary to enhance their uptake in the brain and to achieve sufficient metabolic stability.
Cyclooxygenase-2 (COX-2) is an isoform of enzymes that catalyze the conversion of arachidonic acid to prostaglandins. COX-2 is induced by inflammatory stimuli and is not expressed normally in most tissues. However, recent studies have shown that COX-2 is expressed in various tumors, such as head and neck cancers, colorectal cancers, and melanomas.1–3) Furthermore, in the brain, COX-2 modulates pathological conditions, such as neuroinflammation, Alzheimer’s disease, and Parkinson’s disease.4–6) To detect disorders related to COX-2 expression and activity in vivo using positron emission tomography (PET) and single-photon emission computed tomography, several radiotracers have been developed and evaluated.7–9) Specifically, COX-2 has been evaluated as a brain biomarker, and multiple radiotracers have been investigated for imaging brain COX-2 expression.10,11) However, many of the radiotracer candidates evaluated for COX-2 imaging have exhibited poor brain uptake, lack of specific binding to brain COX-2, and rapid metabolism.8,9) Nevertheless, in 2018 and 2020, Kim et al. reported that [11C]MC1 is a potential COX-2-imaging radioligand that facilitates imaging and quantification of COX-2 upregulation in the monkey brain following lipopolysaccharide-induced neuroinflammation and in inflamed human peripheral tissues, and that it seems promising as a PET radioligand for COX-2 imaging in vivo.12,13)
Nimesulide, with the chemical name N-(4-nitro-2-phenoxyphenyl)methanesulfonamide (Fig. 1), is a well-known and clinically used selective COX-2 inhibitor. Nimesulide mitigates inflammation and pain by inhibiting COX-2 and phosphodiesterase IV and activating CD73.14,15) Nimesulide has exhibited antitumor activity as well as neuroprotectant activity in the brain. Various nimesulide analogs have been synthesized and investigated in the treatment of disorders associated with pain and inflammation.16,17)
COX-2, cyclooxygenase-2.
In our previous study, we designed, synthesized, and evaluated through in vitro studies novel isomeric nimesulide analogs bearing methoxy substituents attached to the phenyl ring as brain COX-2-targeted imaging candidates18,19) (Fig. 1). The methoxy moiety is a potential site for labeling with 11C, a positron emitter with a half-life of 20.4 min. According to in vitro evaluation of isomeric methoxy nimesulide analogs, lipophilicity analysis, COX-inhibitory potency evaluation, transport studies in Caco-2 cells, docking simulations of COX-2, and stability studies in mouse plasma, the para-methoxy analog 1d is an attractive candidate as a brain COX-2 imaging agent. Although the IC50 of 1d (2.31 µM) for COX-2 appears to be higher than that of MC1 (1 nM),12,13) the IC50 of the parent structure, nimesulide, for COX-2 varies in a time-dependent manner, with values ranging from 50 nM to 70 µM.20) Based on the results of docking simulations of 1a and 1d with COX-2, introducing a methoxy group at the para-position of the benzene ring does not change the two binding poses that contribute to the inhibitory potency of nimesulide to COX-2 and is assumed not to be involved in the variable IC50 to COX-2.18) 1d not only has higher selectivity for COX-2 (43 times higher than COX-1), but may also decrease the IC50 for COX-2 to the nanomolar scale, similarly as with nimesulide, and is expected to have a high affinity for COX-2. Therefore, even if the IC50 of 1d for COX-2 is on the µM scale, 1d remains a potential candidate as a brain COX-2 imaging agent.
This article describes the radiosynthesis and in vivo and ex vivo evaluation of isomeric [11C]methoxy analogs of nimesulide. In this study, we performed biodistribution studies, PET imaging, ex vivo autoradiography, and metabolite analysis of 11C-labeled methoxy nimesulide in mice to clarify the in vivo characteristics of the para-methoxy analog [11C]1d, using the isomeric methoxy analogs [11C]1b and [11C]1c as negative controls, which exhibited no COX-2 inhibitory potency. In our previous study, we only evaluated non-radiolabeled nimesulide methoxy analogs using in vitro studies. The present study is the first attempt to radiolabel and evaluate a nimesulide derivative as a COX-2 imaging agent in vivo. This study would provide additional insights into COX-2 expression and activity, as well as COX-2-related disorders of the brain, such as neuroinflammation, Alzheimer’s disease, and Parkinson’s disease, and aid in the diagnosis of these diseases.
All chemicals, reagents, and solvents were of the highest purity available and were used without further purification. The progress of the reactions was monitored using TLC on silica gel 60 F254 glass plates (Merck Millipore, Darmstadt, Germany), and the spots were visualized under UV light. Column chromatography was performed using silica gel 60 Å, 200–400 mesh (Sigma-Aldrich Inc., St. Louis, MO, U.S.A.). All melting points (mp) were determined on a Yanaco melting point apparatus (Yanagimoto Ind. Co., Kyoto, Japan) and were uncorrected. 1H-NMR spectra were recorded on a JNM-LA600, JNM-ECZ400S/L1, or JNM-ECZ600R/S1 (JEOL Ltd., Tokyo, Japan) NMR spectrometer using tetramethylsilane as an internal standard in dimethyl sulfoxide (DMSO)-d6, and the chemical shifts are reported in δ (ppm). The IR spectra were recorded with a Spectrum One FT-IR Spectrometer (PerkinElmer, Inc., Waltham, MA, U.S.A.). Electron ionization (EI)-MS and high-resolution EI-MS spectra were recorded on a JMS-700 (JEOL Ltd.).
[11C]CO2 was produced by proton irradiation of nitrogen gas at 50 µA for 15–20 min using the HM-20 cyclotron (Sumitomo Heavy Industries Ltd., Tokyo, Japan). The preparation of [11C]methyl triflate from [11C]CO2via [11C]methyl iodide and subsequent 11C-methylation were performed using either the CFN-MPS100 or C-11-BII automated synthesis system (Sumitomo Heavy Industries Ltd.). Radioactivity was quantified using the auto-well gamma counter (2480 WIZARD2; PerkinElmer, Inc., or Hidex AMG; Hidex, Turku, Finland).
HPLC for metabolite analysis was performed using a Shimadzu liquid chromatography system (DGU20A3, LC-20AD, and SPD-20 A, Kyoto, Japan) with a UV-detector set at 254 nm and the Frac-920 fraction collector (GE Healthcare Life Sciences, Uppsala, Sweden).
Male ddY mice (7–8 weeks old) and BALB/cCrSlc mice (6 weeks old) were purchased from Japan SLC Inc. (Shizuoka, Japan). The animals were acclimatized to the laboratory environment for at least 1 week prior to the experiments. The protocols for the animal studies were approved by the Animal Care and Use Committee of the Tokyo Metropolitan Institute of Gerontology (Approval No. 190003).
ChemistryN-(4-Nitro-2-phenoxy-phenyl)methanesulfonamide (1a), N-(2-(2-Methoxyphenoxy)-4-nitrophenyl)methanesulfonamide (1b), N-(2-(3-Methoxyphenoxy)-4-nitrophenyl)methanesulfonamide (1c), and N-(2-(4-Methoxyphenoxy)-4-nitrophenyl)methanesulfonamide (1d)Compounds 1a–d were prepared as previously described.18)
N-(2-(2-Hydroxyphenoxy)-4-nitrophenyl)methanesulfonamide (2b)Boron tribromide (BBr3) (1 M in CH2Cl2, 0.9 mL, 0.9 mmol, 3.0 equivalent (equiv.)) was added dropwise to an ice-cooled solution of 1b (100 mg, 0.3 mmol) in anhydrous CH2Cl2 (3 mL). The reaction mixture was mechanically stirred at 5 °C for 30 min, followed by stirring at room temperature for 1.5 h; the reaction mixture was diluted with water and extracted with CHCl3. The combined organic phase was washed with water, dried, and filtered, and the solvent was removed in vacuo. The residue was purified using chromatography on silica gel with CHCl3 : hexane : acetone in a 1 : 1 : 1 proportion, which yielded 2b (58.9 mg, 61.1%) as a yellow solid: mp: 155–160 °C; 1H-NMR (600 MHz, DMSO-d6) δ ppm: 9.99 (s, 1H), 9.82 (s, 1H), 7.95 (dd, J = 2.6 Hz, 9.2 Hz, 1H), 7.68 (d, J = 8.8 Hz, 1H), 7.29 (d, J = 2.6 Hz, 1H), 7.18 (t, J = 7.9 Hz, 2H), 7.06 (d, J = 7.5 Hz, 1H), 6.93 (t, J = 7.7, 1H), 3.23 (s, 3H); Fourier transform (FT)IR (KBr) cm−1; 3445, 3259, 1596; EI-MS m/z: 324 [M]+; HR-EI-MS: Calcd for C13H12N2O6S [M]+: 324.0416. Found 324.0421.
N-(2-(3-Hydroxyphenoxy)-4-nitrophenyl)methanesulfonamide (2c)Compound 2c was produced using a similar synthesis process as that of 2b from 1c; the reaction mixture was stirred at 5 °C for 30 min and subsequently stirred for 2.0 h at room temperature. The crude product was purified using chromatography on silica gel with a CHCl3 : hexane : acetone mixture in a 1 : 8 : 1 ratio, which yielded 2c (95.1 mg, 99.2%) as a yellow solid: mp: 150–152 °C; 1H-NMR (600 MHz, DMSO-d6) δ ppm: 10.14 (s, 1H), 9.81 (s, 1H), 8.02 (dd, J = 2.7 Hz, 9.0 Hz, 1H), 7.72 (d, J = 9.2 Hz, 1H), 7.57 (d, J = 2.6 Hz, 1H), 7.25 (t, J = 8.2 Hz, 1H), 6.67 (dd, J = 2.2 Hz, 8.4 Hz, 1H), 6.57 (dd, J = 2.6 Hz, 8.1 Hz, 1H), 6.53 (t, J = 2.4 Hz, 1H), 3.20 (s, 3H); FTIR (KBr) cm−1; 3272, 1594, 1523; EI-MS m/z: 324 [M]+; HR-EI-MS: Calcd for C13H12N2O6S [M]+: 324.0416. Found 324.0421.
N-(2-(4-Hydroxyphenoxy)-4-nitrophenyl)methanesulfonamide (2d)Compound 2d was produced similarly as was 2b from 1d. The crude product was purified using chromatography on silica gel with a CHCl3 : hexane : acetone mixture in a 1 : 3 : 1 ratio, which yielded 2d (100.4 mg, quant) as a yellow solid: mp: 53–58 °C; 1H-NMR (400 MHz, DMSO-d6) δ ppm: 10.07 (s, 1H), 9.58 (s, 1H), 7.95 (dd, J = 2.5 Hz, 8.9 Hz, 1H), 7.68(d, J = 8.7, 1H), 7.37 (d, J = 2.3 Hz, 1H), 7.04 (d, J = 8.7 Hz, 2H), 6.86 (d, J = 8.7 Hz, 2H), 3.21 (s, 3H); FTIR (KBr) cm−1; 3283, 1507; EI-MS m/z: 324 [M]+; HR-EI-MS: Calcd for C13H12N2O6S [M]+: 324.0416. Found 324.0417.
RadiochemistryMethylation of the O-desmethyl precursor with [11C]methyl triflate was performed using a previously described procedure with slight modifications.21) [11C]Methyl triflate was trapped in acetone (250 µL) containing O-desmethyl precursor 2b–d (0.25 mg, 0.77 µmol) and 1 M aqueous NaOH (5 µL, 5 µmol) as the base. After trapping [11C]methyl triflate, 1.5 mL of H2O was added, and the reaction mixture was applied to a semi-preparative HPLC column (YMC Pack Pro C18 RS, 10 ×250 mm, S-5 µm, 8 nm, YMC, Kyoto, Japan) comprising a UV absorbance detector (λ = 260 nm) and a radiation detector. The mobile phase was a mixture of acetonitrile (MeCN) : 50 mM aqueous AcOH : 50 mM aqueous AcONH4 = 55 : 22.5 : 22.5 (v/v/v) at a flow rate of 5 mL/min. Each [11C]product fraction (retention time: 8.9 min for [11C]1b, 9.3 min for [11C]1c, and 9.0 min for [11C]1d) was collected in a flask containing a solution of ascorbate injection (0.2 mL, 250 mg/mL; Nipro Pharma, Osaka, Japan) and 0.5 mL of ethanol and evaporated to dryness. The residue was dissolved in physiological saline containing 0.125% (v/v) Tween 80. Radiochemical purity was analyzed using HPLC: YMC Pack ODS-A column (4.6 ×150 mm, 3 µm, 120 Å, YMC); mobile phase, MeCN : 50 mM aqueous AcOH : 50 mM aqueous AcONH4 mixture in a 60 : 20 : 20 (v/v/v) ratio; and flow rate, 1.0 mL/min. The retention times were 4.4 min for [11C]1b, 4.6 min for [11C]1c, and 4.6 min for [11C]1d. To determine molar activity, the mass (nmol) of the labeled compound with known radioactivity (MBq) was determined for the labeled compound with HPLC comparison of the UV absorbance at 245 nm with that of the known concentrations of the corresponding non-radioactive compound.
Biodistribution of Radioactivity Following the Administration of Radiotracers in MiceEach radiotracer was intravenously injected. The 11C-radioactivity in the samples obtained was counted with the auto-well gamma counter, and the tissues were weighed. The tissue uptake of 11C was expressed as the standardized uptake value (SUV): cpm measured per gram of tissue/cpm injected per gram body weight.
Tissue Distribution in ddY MiceMale ddY mice (9 weeks old, body weight 36–43 g) were sacrificed by cervical dislocation 1, 5, 15, 30, and 60 min after the radiotracer (3.4–5.0 MBq/0.2 mL/head, 21–50 pmol) injection (n = 4).
Blocking Study in ddY MiceMale ddY mice (9 weeks old, body weight 33–42 g) received blocker solutions (1 mg/kg body weight, 0.1 mL) of celecoxib, indomethacin, and nimesulide dissolved in DMSO, which were co-injected with each radiotracer (n = 5). In the control mice, the same amount of DMSO was co-injected. The mice were sacrificed by cervical dislocation 15 min after the radiotracer (3.4–4.5 MBq/0.2 mL/head, 16–44 pmol) injection, and the radioactivity levels of the brain and blood were measured. The Dunnett’s test was used to compare the groups in the pharmacological experiments; p < 0.05 was considered statistically significant.
Tissue Distribution of [11C]1d in a Mouse Model of InflammationA mouse model of inflammation was established using a previously described procedure with slight modifications.22) For developing the inflammation model, male BALB/cCrSlc mice (7 weeks old) were anesthetized with a 2% isoflurane air mixture and a paper disc (for antibiotic, ϕ = 8 mm, Advantec Toyo Kaisha, Ltd., Tokyo, Japan) soaked in turpentine was subcutaneously inserted into the thigh of the right hind leg. Seven days after the turpentine treatment, [11C]1d (9.5 MBq/0.2 mL/head, 74 pmol) was intravenously injected into the mouse model of inflammation (8 weeks old, body weight 18–23 g). The mice were sacrificed by cervical dislocation 15, 30, 60, and 90 min after radiotracer injection, and the radioactivity levels of the blood, muscle, and inflammation were evaluated (n = 5).
PET Imaging with [11C]1b–d in ddY MiceA PET study in mice was performed using a small animal PET scanner (MIP-100, Sumitomo Heavy Industries Ltd.).23) [11C]1b–d (27.1–30.6 MBq, 0.2–0.8 nmol, n = 1) were injected into the tail vein of ddY mice (9 weeks old, body weight 41–42 g) with 1.5% (v/v) isoflurane. A 60 min dynamic scan was performed immediately after the injection. The resulting sinograms were reconstructed into 21 frames (8 × 30, 3 × 60, 2 × 120, 2 × 180, 3 × 300, 2 × 540, and 1 × 600 s) using an interactive reconstruction algorithm (three-dimensional ordered-subset expectation maximization, provided by Sumitomo Heavy Industries; one iteration, 32 subsets). The final data sets consisted of 31 slices, with a slice thickness of 0.85 mm and an in-plane image matrix of 256 × 256 pixels (0.3 × 0.3 mm pixel size). The data sets were fully corrected for random coincidences and scatter. We obtained SUV images by normalizing tissue radioactivity concentrations based on the injected dose and body weight using the PMOD software, version 3.409 (PMOD Technologies, Zurich, Switzerland).
Ex Vivo Autoradiography in ddY Mice[11C]1b–d (131–392 MBq, 0.6–2.7 nmol, n = 1) were intravenously injected into ddY mice (8–9 weeks old, body weight 35–40 g). The mice were sacrificed by cervical dislocation, at 15 and 30 min after radiotracer injection. The brain was rapidly dissected, frozen, and cut into 20 µm-thick coronal sections by using a cryotome (Bright Instrument Co., Ltd., Huntingdon, U.K.). The sections were dried on a hot plate at 60 °C and exposed to a storage phosphor screen (GE Healthcare Life Sciences). After the decay of radioactivity, the 11C distribution was visualized using a phosphorimager system (Storm 820; GE Healthcare Life Sciences).
Metabolite AnalysisEach of the radiotracers (108–166 MBq, 0.7–1.5 nmol, n = 3) was intravenously injected into ddY mice (8–9 weeks old, body weight 35–41 g). The mice were sacrificed by cervical dislocation after 15 and 30 min of the radiotracer injection. The blood was removed by cardiac puncture using a heparinized syringe at both time points, and the brain was removed at the 15-min time point and stored on ice until further processing.
The blood was centrifuged at 4000 × g for 1 min at 4 °C to obtain the plasma, which (50 µL) was denatured with 200 µL of 50 mM AcOH–AcONH4 aqueous buffer (50 : 50) and 250 µL of MeCN. The mixture was centrifuged under the same conditions, and the precipitate was resuspended in 500 µL of a 1 : 1 mixture of MeCN and 50 mM AcOH–AcONH4 aqueous buffer (50 : 50), followed by centrifugation. This procedure was repeated twice. The cerebral cortex (approximately 50 mg) was homogenized in 500 µL of a 1 : 1 mixture of MeCN and 50 mM AcOH–AcONH4 aqueous buffer (50 : 50). The homogenate was then processed using the same procedure as was followed for processing the plasma sample. After protein elimination, the recovery of the radioactivity in the soluble fractions was 77–99%.
The combined supernatant was analyzed using the HPLC: YMC Pack ODS-A column (10 × 250 mm, 5 µm, YMC). The mobile phase comprised a mixture of MeCN : 50 mM aqueous AcOH : 50 mM aqueous AcONH4 in a 70 : 15 : 15 (v/v/v) ratio at a flow rate of 2 mL/min. Fractions were collected at 30-s intervals, and their radioactivity was measured using the auto-well gamma counter.
Nimesulide (1a) and its isomeric methoxy derivatives 1b–d were prepared as previously described.18) The phenolic precursors 2b–d for 11C-radiolabeling were synthesized through demethylation of the corresponding methoxy derivatives 1b–d using BBr3 as a dealkylating agent, as depicted in Chart 1. The structures of all compounds were characterized using 1H-NMR and high-resolution mass spectrometry.
Radiosynthesis of [11C]1b–d was accomplished as shown in Chart 2. [11C]1b–d were synthesized through O-methylation of the O-desmethyl precursors 2b–d in acetone with [11C]methyl triflate in the presence of NaOH as the base at room temperature. The presence of NaOH was essential for efficiently achieving the desired methylation, and the excess NaOH (6.5 equiv.) successfully produced the required products. When the precursors 2b–d were dissolved in acetone and stored at −20 °C for 2 weeks, poor radiochemical yields were observed ([11C]1b: 1.9%, [11C]1c: 25.8%, and [11C]1d: not detected). Since no change was observed in the 1H-NMR spectrum and HPLC chart of the precursors, it is difficult to know why the radiochemical yields were reduced. For better radiochemical yields, the precursors were dissolved and used rapidly at every synthesis. The total synthesis time was within 30 min from the end of the bombardment. Table 1 summarizes the results of the radiosynthesis.
| Compound | Radiochemical yield (%) a) | Radiochemical purity (%) | Molar activity (MBq/nmol) |
|---|---|---|---|
| [11C]1b | 47–52 | >95 | 177–240 |
| [11C]1c | 26–46 | >95 | 94–237 |
| [11C]1d | 29–35 | >96 | 204–300 |
a) Radiochemical yield is based on [11C]CH3I and decay corrected.
The tissue distribution of the radioactivity after the injection of [11C]1b–d into normal male ddY mice is summarized in Tables 2a–c. [11C]1b–d exhibited a relatively fast clearance of radioactivity in the blood within 60 min, with [11C]1d showing the fastest blood clearance. The liver showed the highest initial uptake (SUV). The whole-brain uptake of all tracers was low, with values of 0.53, 0.37, and 0.34 for [11C]1b, [11C]1c, and [11C]1d, respectively, at 1 min after administration. The brain uptake of [11C]1b and [11C]1c at 60 min after administration was 0.28 and 0.35, respectively; in contrast, the uptake of [11C]1d decreased to 0.08.
| Tissue | Uptake (SUV) | ||||
|---|---|---|---|---|---|
| 1 min | 5 min | 15 min | 30 min | 60 min | |
| Blood | 2.50 ± 0.4 | 1.64 ± 0.1 | 1.20 ± 0.1 | 0.95 ± 0.2 | 0.54 ± 0.2 |
| Heart | 1.92 ± 0.1 | 1.21 ± 0.1 | 0.91 ± 0.1 | 0.74 ± 0.1 | 0.38 ± 0.1 |
| Lung | 1.84 ± 0.2 | 1.06 ± 0.1 | 0.84 ± 0.0 | 0.64 ± 0.1 | 0.38 ± 0.1 |
| Liver | 2.88 ± 0.2 | 2.46 ± 0.2 | 2.74 ± 0.2 | 2.49 ± 0.2 | 2.10 ± 0.1 |
| Pancreas | 0.94 ± 0.1 | 0.71 ± 0.1 | 0.77 ± 0.1 | 0.85 ± 0.1 | 0.95 ± 0.2 |
| Spleen | 0.68 ± 0.2 | 0.55 ± 0.0 | 0.47 ± 0.0 | 0.43 ± 0.1 | 0.39 ± 0.0 |
| Kidney | 1.31 ± 0.0 | 1.09 ± 0.1 | 1.04 ± 0.0 | 0.87 ± 0.2 | 0.61 ± 0.1 |
| Small intestine | 1.16 ± 0.1 | 1.13 ± 0.1 | 1.83 ± 0.4 | 1.45 ± 0.4 | 1.15 ± 0.1 |
| Muscle | 0.63 ± 0.1 | 0.42 ± 0.1 | 0.36 ± 0.0 | 0.23 ± 0.1 | 0.18 ± 0.0 |
| Brain | 0.53 ± 0.1 | 0.40 ± 0.1 | 0.36 ± 0.0 | 0.34 ± 0.1 | 0.28 ± 0.1 |
SUV, standardized uptake value; S.D., standard deviation.
| Tissue | Uptake (SUV) | ||||
|---|---|---|---|---|---|
| 1 min | 5 min | 15 min | 30 min | 60 min | |
| Blood | 2.51 ± 0.1 | 0.92 ± 0.1 | 0.43 ± 0.1 | 0.46 ± 0.1 | 0.48 ± 0.0 |
| Heart | 1.61 ± 0.1 | 0.75 ± 0.1 | 0.21 ± 0.1 | 0.27 ± 0.1 | 0.33 ± 0.0 |
| Lung | 1.51 ± 0.3 | 0.58 ± 0.0 | 0.35 ± 0.2 | 0.29 ± 0.1 | 0.38 ± 0.0 |
| Liver | 3.23 ± 0.2 | 2.96 ± 0.1 | 3.36 ± 1.4 | 3.86 ± 0.3 | 2.96 ± 0.3 |
| Pancreas | 0.90 ± 0.0 | 0.70 ± 0.2 | 1.42 ± 0.3 | 2.15 ± 0.3 | 1.85 ± 0.2 |
| Spleen | 0.57 ± 0.2 | 0.49 ± 0.1 | 0.36 ± 0.2 | 0.38 ± 0.1 | 0.49 ± 0.0 |
| Kidney | 1.14 ± 0.1 | 0.79 ± 0.1 | 0.56 ± 0.1 | 0.62 ± 0.1 | 0.66 ± 0.1 |
| Small intestine | 1.07 ± 0.1 | 0.82 ± 0.2 | 0.72 ± 0.2 | 0.79 ± 0.1 | 0.87 ± 0.1 |
| Muscle | 0.51 ± 0.1 | 0.20 ± 0.0 | 0.19 ± 0.1 | 0.20 ± 0.1 | 0.15 ± 0.0 |
| Brain | 0.37 ± 0.0 | 0.24 ± 0.0 | 0.29 ± 0.0 | 0.50 ± 0.1 | 0.35 ± 0.1 |
SUV, standardized uptake value; S.D., standard deviation.
| Tissue | Uptake (SUV) | ||||
|---|---|---|---|---|---|
| 1 min | 5 min | 15 min | 30 min | 60 min | |
| Blood | 1.89 ± 0.1 | 0.67 ± 0.0 | 0.25 ± 0.0 | 0.24 ± 0.0 | 0.23 ± 0.0 |
| Heart | 1.08 ± 0.1 | 0.57 ± 0.0 | 0.26 ± 0.0 | 0.22 ± 0.0 | 0.18 ± 0.0 |
| Lung | 1.40 ± 0.1 | 0.48 ± 0.0 | 0.28 ± 0.1 | 0.21 ± 0.0 | 0.21 ± 0.1 |
| Liver | 4.67 ± 0.8 | 3.97 ± 0.8 | 2.47 ± 0.5 | 1.73 ± 0.2 | 1.43 ± 0.1 |
| Pancreas | 0.69 ± 0.1 | 0.81 ± 0.1 | 1.08 ± 0.1 | 1.01 ± 0.2 | 0.95 ± 0.1 |
| Spleen | 0.82 ± 0.1 | 0.42 ± 0.0 | 0.34 ± 0.1 | 0.23 ± 0.0 | 0.26 ± 0.1 |
| Kidney | 1.02 ± 0.0 | 0.77 ± 0.1 | 0.45 ± 0.1 | 0.34 ± 0.1 | 0.29 ± 0.0 |
| Small intestine | 1.02 ± 0.1 | 0.63 ± 0.0 | 0.41 ± 0.1 | 0.32 ± 0.1 | 0.36 ± 0.1 |
| Muscle | 0.45 ± 0.0 | 0.27 ± 0.0 | 0.18 ± 0.0 | 0.11 ± 0.0 | 0.09 ± 0.0 |
| Brain | 0.34 ± 0.0 | 0.28 ± 0.0 | 0.21 ± 0.1 | 0.15 ± 0.0 | 0.08 ± 0.0 |
SUV, standardized uptake value; S.D., standard deviation.
The brain uptake and brain-to-blood ratio of [11C]1b–d with or without COX inhibitors 15 min after the injection are summarized in Tables 3a and b. Celecoxib and indomethacin were used with [11C]1b–d, and nimesulide (1a) was used with [11C]1b and [11C]1d. None of the COX inhibitors caused significant inhibition in the brain uptake and brain-to-blood ratio of [11C]1b and [11C]1c. In contrast, in the case of [11C]1d, celecoxib, indomethacin, and nimesulide inhibited its brain uptake significantly. Although the COX inhibitors increased the brain uptake and brain-to-blood ratio of [11C]1b and [11C]1c, they decreased the brain uptake and brain-to-blood ratio of [11C]1d (compared to the control).
| Compound | Brain uptake (SUV) | |||
|---|---|---|---|---|
| Control | Celecoxib | Indomethacin | Nimesulide | |
| [11C]1b | 0.31 ± 0.0 | 0.42 ± 0.0 | 0.46 ± 0.0 | 0.33 ± 0.0 |
| [11C]1c | 0.35 ± 0.1 | 0.48 ± 0.1 | 0.48 ± 0.0 | — |
| [11C]1d | 0.47 ± 0.1 | 0.37 ± 0.0* | 0.37 ± 0.1* | 0.30 ± 0.0* |
*Significant decrease (p < 0.05) compared to the control (Dunnett’s test). SUV, standardized uptake value; S.D., standard deviation.
| Compound | Brain-to-blood ratio | |||
|---|---|---|---|---|
| Control | Celecoxib | Indomethacin | Nimesulide | |
| [11C]1b | 0.27 ± 0.0 | 0.36 ± 0.0 | 0.37 ± 0.0 | 0.27 ± 0.1 |
| [11C]1c | 0.61 ± 0.1 | 0.75 ± 0.2 | 0.67 ± 0.0 | — |
| [11C]1d | 0.92 ± 0.1 | 0.90 ± 0.1 | 0.82 ± 0.1 | 0.83 ± 0.1 |
S.D., standard deviation.
The biodistribution of radioactivity in the mouse model of inflammation is summarized in Table 4. The uptake of [11C]1d in the inflammatory region was low as opposed to the high levels of [11C]1d-associated radioactivity in the blood throughout the observation period. The ratio of inflammation-to-muscle was considerably less than 2.5 units (1.2–2.1 units) at all time points.
| Tissue | Uptake (SUV) | |||
|---|---|---|---|---|
| 15 min | 30 min | 60 min | 90 min | |
| Blood | 0.33 ± 0.1 | 0.32 ± 0.1 | 0.33 ± 0.1 | 0.40 ± 0.1 |
| Muscle | 0.19 ± 0.0 | 0.14 ± 0.0 | 0.12 ± 0.0 | 0.09 ± 0.0 |
| Inflammation | 0.27 ± 0.0 | 0.22 ± 0.1 | 0.14 ± 0.0 | 0.16 ± 0.0 |
SUV, standardized uptake value; S.D., standard deviation.
Representative 11C-PET images and time–activity curves with radiotracers are shown in Fig. 2. The distribution of radioactivity of [11C]1b–d in the head decreased 1.5–2 min after the administration of the compound, and radioactivity was barely visible at 50–60 min after administration.
Dynamic PET undertaken 1.5–2.0 min and 50–60 min after administration ([11C]1b: 27.1 MBq, [11C]1c: 30.6 MBq, and [11C]1d: 30.5 MBq) under 1.5% isoflurane anesthesia. Coronal (upper left), sagittal (upper right), and horizontal (lower left) sections. SUV, standardized uptake value; PET, positron emission tomography. (Color figure can be accessed in the online version.)
Figure 3 presents the ex vivo autoradiographic images of brain coronal sections. Both [11C]1b and [11C]1c showed a uniform distribution of 11C radioactivity. In contrast, [11C]1d was heterogeneously distributed in the cerebral cortex with a slightly high density.
Autoradiographic images that were obtained 15 and 30 min after administration ([11C]1b: 132 MBq, [11C]1c: 146 MBq, and [11C]1d: 392 MBq). The online version contains pseudo-color images. (Color figure can be accessed in the online version.)
The in vivo metabolic behavior of each radiotracer was studied in the mice at 15 and 30 min after the injection of the compounds. As summarized in Table 5, both [11C]1c and [11C]1d were rapidly metabolized. The unchanged forms of [11C]1c and [11C]1d accounted for 16.4 and 8.6% of the total radioactivity in the plasma, respectively, at 15 min after injection; these values reduced to 9.4 and 5.9%, respectively, at 30 min after injection. Moreover, 15 min after the injection, the levels of the unchanged forms of [11C]1c and [11C]1d were both 2.5% of the total radioactivity in the brain, which was considerably lower than the corresponding levels in the plasma. In contrast, [11C]1b retained its parent structure compared to other radiotracers. The percentage of unchanged [11C]1b in the brain was less than that in the plasma, which was consistent with the results for the other compounds. [11C]1b showed two polar peaks, with retention times of 5–6 and 9 min, respectively, and [11C]1c showed one polar peak, with a retention time of 5–6 min. [11C]1d showed one polar broad peak with a retention time of 5–9 min.
| Compound | Sampling interval and specimen | Percentage of metabolite (%) | ||
|---|---|---|---|---|
| Parent | Metabolite 1 | Metabolite 2 | ||
| 13 min b) | 5–6 min b) | 9 min b) | ||
| [11C]1b | Plasma (15 min) | 86.4 ± 4 | 9.2 ± 3 | 2.7 ± 1 |
| Plasma (30 min) | 78.7 ± 7 | 16.1 ± 7 | 3.0 ± 1 | |
| Brain (15 min) | 62.8 ± 6 | 30.5 ± 8 | 4.7 ± 1 | |
| [11C]1c | Plasma (15 min) | 16.4 ± 4 | 72.4 ± 6 | N.D. |
| Plasma (30 min) | 9.4 ± 2 | 52.9 ± 2 | N.D. | |
| Brain (15 min) | 2.5 ± 1 | 92.1 ± 2 | N.D. | |
| [11C]1d | Plasma (15 min) | 8.6 ± 5 | 74.0 ± 9 (broad, 5–9 min b)) | |
| Plasma (30 min) a) | 5.94 | 73.1 (broad, 5–9 min b)) | ||
| Brain (15 min) | 2.5 ± 1 | 90.7 ± 5 (broad, 5–9 min b)) | ||
a) n = 1. b) Retention time, N.D., not detected; S.D., standard deviation.
The goal of this study was to evaluate isomeric [11C]methoxy analogs of nimesulide as brain-COX-2-targeted imaging agents. In our previous study, we designed and synthesized three isomeric methoxy analogs of nimesulide, 1b–d; moreover, we found that the para-methoxy analog 1d exhibited appropriate inhibitory potency and selectivity for COX-2.18) Furthermore, 1b–d showed good plasma stability in vitro, with more than 95% of the parent compound remaining after 120 min of incubation and with no other detectable metabolites.19) In our previous study, we only evaluated the non-radiolabeled nimesulide methoxy analogs by using in vitro studies. The present study constitutes the first attempt to radiolabel and evaluate a nimesulide derivative as a COX-2 imaging agent in vivo.
We synthesized the phenolic precursors 2b–d for 11C radiolabeling with satisfactory yields. All the synthetic compounds produced satisfactory spectroscopic data, which were in complete accordance with their depicted structures. [11C]-O-methylation was successfully achieved with reasonable radiochemical yield, radiochemical purity, and molar activity.
According to the results of the biodistribution study in mice, [11C]1b–d exhibited fast blood clearance; for all radiotracers, that is, [11C]1b–d, the initial liver uptake was the highest, and brain uptake was the lowest, when compared to uptake by other tissues. We performed blocking experiments to determine the specific binding of radiotracers to COX-2 by using celecoxib and nimesulide as the COX-2 selective inhibitors and indomethacin as the nonselective COX inhibitor. Both [11C]1b and [11C]1c were not significantly inhibited by any COX inhibitor. In contrast, the brain uptake of [11C]1d was significantly inhibited by COX inhibitors. In our previous study, 1b and 1c showed no inhibitory potency for both COX enzymes.18) The results of the blocking study appeared to indicate a small fraction of displaceable binding in the brain. Although our target is COX-2 in the brain, we performed a biodistribution study by using a peripheral inflammation model to more clearly determine the specific binding of the compounds to COX-2. Turpentine-induced inflammation causes COX-2 expression in activated neutrophils and macrophages, and it is used for the evaluation of COX-2-targeted radiotracers.24–26) In our previous study, we observed that a paper disc soaked with turpentine induced clearer regions with inflammation than when turpentine was injected directly; 67Ga-citrate, as an inflammatory diagnostic agent, exhibited higher uptake in the inflammatory region (inflammation-to-blood ratio: 1.59 ± 0.3 and inflammation-to-muscle ratio: 4.16 ± 0.5, when administered after 24 h, n = 3) in the same model. In our preliminary in vivo data, 125I-labeled para-iodo nimesulide accumulated in inflammatory regions with gradual increase and slow clearance, and the inflammation-to-blood ratio and inflammation-to-muscle ratio were significantly inhibited by COX inhibitors. However, in the case of [11C]1d, the uptake in the inflammatory region was low, and this uptake further decreased with no peak; the inflammation-to-blood ratio was less than 1.0 units, and the inflammation-to-muscle ratio was considerably less than 2.5 units at all time points. Here, the inflammation-to-muscle ratio was greater than 1.0 units at all time points, which appears to be a positive result in that [11C]1d may potentially bind to the inflammatory region, but [11C]1d did not exhibit remarkable uptake in the inflammatory region in the mouse model of inflammation. As para-iodo nimesulide has higher lipophilicity and COX-2 inhibitory activity than 1d (logP7.4 of 1d and para-iodo nimesulide is 1.41 and 2.77, respectively, whereas the IC50 of 1d and para-iodo nimesulide for COX-2 is 2.31 and 0.47 µM, respectively),18,19) the lack of significant [11C]1d uptake in the inflammatory region in the inflammation model may be attributed to the differences in lipophilicity and COX-2 inhibitory activity (which may occur individually or together).
The 11C-PET images with radiotracer [11C]1b–d were explained by the results of the biodistribution studies: lower brain uptake and relatively fast clearance. Ex vivo autoradiographic images indicate the difference in COX-2 inhibitory potency between 1b–c and 1d. Both [11C]1b and [11C]1c, with no COX-inhibitory potency, showed uniform distribution, whereas [11C]1d which had potent COX-2 inhibitory potency exhibited heterogeneous distribution. In the rat brain, high basal levels of COX-2 were observed in the neocortex and hippocampus.27,28) However, immunohistochemistry using C57BL/6N mouse brains showed COX-2 expression in the white matter and nerve fibers, and there was almost no detectable immunoreactivity in the cortex.29) Thus, we need to further investigate whether the observed heterogeneous distribution of [11C]1d correlates with the COX-2 enzyme distribution.
The results of the in vivo metabolite analysis demonstrated that [11C]1b was relatively stable and [11C]1c and [11C]1d were rapidly metabolized; however, 1b–d showed good plasma stability in vitro.19) The percentages of unchanged [11C]1b–d in the brain were lower than that in the plasma, and especially in the brain, the metabolites of [11C]1c and [11C]1d were present in greater proportion than that of the unchanged forms. Considering the possibility that blood contamination influences the results of the metabolite analysis in the brain and assuming that approximately 5% of the whole brain contains blood, the percentage of blood-derived radioactivity of [11C]1b–d in the brain was calculated to be 16.5, 7.5, and 6.0%, respectively. This calculated result did not completely explain the percentage of the [11C]1b–d metabolites in the brain (35.2, 92.1, and 90.7%, respectively); however, we cannot infer that the results of the metabolite analysis in the brain were unaffected by contamination with blood components. Para-hydroxy nimesulide (2d), amino des-nitro nimesulide (and/or its para-hydroxylated agents), and N-acetylated nimesulide (and/or its para-hydroxylated agents) are representative metabolites of nimesulide.30–32) If the isomeric methoxy analogs of nimesulide 1b–d were metabolized similarly as was nimesulide, then the unidentified metabolites 1 and 2 might be para-hydroxylate, amino des-nitro, or N-acetylate metabolites, or demethylated 11C-methoxy moiety. In addition, the rapid metabolism of [11C]1d may be caused by the 4-position of the phenoxy ring as a potential reactive site. Among the putative metabolites, para-hydroxy nimesulide has been reported to exhibit a pharmacokinetic behavior similar to that of the parent compound.33,34) It is possible that the expected metabolites in the blood could be transferred to the brain, considering the chemical structure and lipophilicity of para-hydroxylate, amino des-nitro, and N-acetylate metabolites.30) The involvement of the metabolic enzymes CYP2C9, 2C19, and 1A2 in the metabolism of nimesulide was previously reported.34,35) CYP2C9, 2C19, and 1A2 have been detected in mouse, rat, and human brain microsomes; however, CYPs in the brain are regulated differently from those in the liver, and their expression levels are significantly lower than the hepatic expression levels.36,37) In contrast, CYP proteins in the brain have been reported to reach levels similar to that observed in the liver in certain cell types.38) The activity of CYPs against [11C]1b–d in the brain is unknown; however, as long as CYP2C9, 2C19, and 1A2 exist in the brain, the possibility that [11C]1b–d was metabolized in the brain cannot be denied. The metabolites of nimesulide cause hepatotoxicity.35,39) The highest initial liver uptake and the moderate small intestinal uptake of [11C]1b–d indicates the level of hepatic metabolism, and these uptakes may be explained by their metabolites.40,41) P-glycoprotein has been reported to play an important role in restricting the brain uptake of radioligands.42,43) In our previous study, transport studies using Caco-2 cells showed that P-glycoprotein was not involved in the cellular transport of 1b–d, and the experimental logP7.4 and pKa values for 1b–d were all within the acceptable range that was required for passive brain penetration.18,19) The lower brain uptake of [11C]1b–d in vivo may be explained by the metabolism of these compounds rather than by their lipophilicity and the effect of P-glycoprotein. Even though only 2.5% of the unchanged form of [11C]1d was present in the brain 15 min after injection, the reason why the brain uptake of the radioactivity was significantly reduced by the COX inhibitor is unclear. Hypothetically, this metabolite also has COX-2 specific binding and could have been inhibited by COX inhibitors. Although in the case of [11C]1d, the para-hydroxylated metabolite is excluded because it contains no radioactivity, the possibility that amino des-nitro and N-acetylated metabolites may interact with COX-2 cannot be denied. Another hypothesis is the effect of DMSO in blocking experiments. As shown in Table 2c, the brain uptake of radioactivity (SUV) at 15 min after [11C]1d injection is 0.21, while in the control of the blocking experiment, it is 0.47, as shown in Table 3a. The only difference between them is the presence or absence of DMSO, indicating that the presence of DMSO affects the amount of brain uptake of radioactivity. This is not a difference seen only in the brain, for example, in the blood there is an increase from 0.25 to 0.51, indicating an overall increase in the uptake of radioactivity (SUV) at 15 min after [11C]1d injection with the presence of DMSO. The mechanism by which DMSO enhanced the distribution of radioactivity to the various tissues is unclear, but increased brain uptake (SUV) at 15 min after [11C]1d injection may have made COX-2-specific binding more detectable.
This paper describes the radiosynthesis and in vivo and ex vivo evaluation of isomeric [11C]methoxy analogs of nimesulide for identifying a suitable radiotracer candidate for imaging brain COX-2 expression. The potential COX-2-targeted radiotracer candidate [11C]1d showed lower brain uptake than the negative controls ([11C]1b and [11C]1c); both [11C]1d and [11C]1c were metabolized rapidly in ddY mice. Furthermore, [11C]1b and [11C]1c, which have been previously demonstrated as lacking COX-2 inhibitory potency, did not exhibit specific binding affinity toward the COX-2 enzyme in the brain. In contrast, [11C]1d, which exhibited potent COX-2 inhibitory potency and selectivity in our previous study, appeared to show a small but significant specific signal. The results of biodistribution using the peripheral inflammation model and 11C-PET study showed no remarkable COX-2-dependent uptake. However, the heterogeneous ex vivo distribution of [11C]1d might correlate with the distribution of the COX-2 enzyme. We investigated [11C]1d as a potential COX-2-targeted radiotracer candidate and found that it appears to be unsuitable as a COX-2 imaging agent because of the insufficient brain uptake and instability of this compound in vivo. However, the design of the present study is subject to some limitations; we performed in vivo studies using mice, which exhibit a relatively faster metabolism compared to that of other animals. Therefore, the results might be different in animals with slower metabolisms than that of mice. Further structural refinement of the radiotracers is necessary to enhance the brain uptake and to achieve sufficient metabolic stability.
The authors thank Mr. Masanari Sakai, SHI Accelerator Service Ltd., Tokyo, Japan, for technical assistance with the cyclotron operation and radiosynthesis. The authors greatly appreciate Ms. Asuka Okayasu and Ms. Akiyo Oiwa, Tohoku Medical and Pharmaceutical University, Sendai, Japan, for their technical assistance with the precursor synthesis.
This research was partially supported by a Grant-in-Aid for Scientific Research (C) from the Japan Society for the Promotion of Science [JSPS KAKENHI Grant Numbers JP17K10372 and JP20K08088]. The funder had no role in the study design, data collection, data analysis and interpretation, preparation of the manuscript, or decision to publish.
The authors declare no conflict of interest.
The online version of this article contains supplementary materials.
