Abstract
The serotonin (5-hydroxytryptamine [5-ΗΤ]) 7 receptor (5-HT7R) is the most recently discovered 5-HT receptor, and its physiologic and possible pathophysiologic roles are not fully elucidated. So far, no suitable 5-HT7R PET radioligand is available, thus limiting the investigation of this receptor in the living brain. Here, we present the radiosynthesis and in vivo evaluation of Cimbi-712 (3-{4-[4-(4-methylphenyl)piperazine-1-yl]butyl}p-1,3-dihydro-2H-indol-2-one) and Cimbi-717 (3-{4-[4-(3-methoxyphenyl)piperazine-1-yl]butyl}-1,3-dihydro-2H-indol-2-one) as selective 5-HT7R PET radioligands in the pig brain. The 5-HT7R distribution in the postmortem pig brain is also assessed. Methods: In vitro autoradiography with the 5-HT7R selective radioligand 3H-labeled (R)-3-(2-(2-(4-methylpiperidin-1-yl)ethyl)pyrrolidine-1-sulfonyl)phenol (SB-269970) was performed on pig brain sections to establish the 5-HT7R binding distribution. Radiolabeling of 5-HT7R selective compounds was performed in an automated synthesis module in which we conducted either palladium-mediated cross coupling (11C-Cimbi-712) or conventional O-methylation (11C-Cimbi-717) using 11C-MeI and 11C-MeOTf, respectively. After intravenous injection of the radioligands in Danish Landrace pigs, the in vivo brain distribution of the ligands was studied. Specific binding of 11C-Cimbi-712 and 11C-Cimbi717 to 5-HT7R was investigated by intravenous administration of SB-269970 before a second PET scan. Results: High 5-HT7R density was found in the thalamus and cortical regions of the pig brain by autoradiography. The radiosynthesis of both radioligands succeeded after optimization efforts (radiochemical yield, ∼20%–30% at the end of synthesis). Time–activity curves of 11C-Cimbi-712 and 11C-Cimbi-717 showed high brain uptake and distribution according to 5-HT7R distribution, but the tracer kinetics of 11C-Cimbi-717 were faster than 11C-Cimbi-712. Both radioligands were specific for 5-HT7R, as binding could be blocked by pretreatment with SB-269970 for 11C-Cimbi-717 in a dose-dependent fashion. For 11C-Cimbi-717, nondisplaceable binding potentials of 6.4 ± 1.2 (n = 6) were calculated in the thalamus. Conclusion: Both 11C-Cimbi-712 and 11C-Cimbi-717 generated a specific binding in accordance with 5-HT7R distribution and are potential PET radioligands for 5-HT7R. 11C-Cimbi-717 is the better candidate because of the more reversible tracer kinetics, and this radioligand showed a dose-dependent decline in cerebral binding after receptor blockade. Thus, 11C-Cimbi-717 is currently the most promising radioligand for investigation of 5-HT7R binding in the living human brain.
The serotonergic system plays a key modulatory role in the brain and is a target for many drug treatments for brain disorders either through reuptake blockade or interactions with one or more of the 14 subtypes of 5-hydroxytryptamine (5-HT) receptors. Our knowledge about the behavior of the 5-HT system in vivo is still scattered, and most of the understanding is derived from animal models. However, the use of imaging techniques such as positron emission tomography (PET) and the increasing number of radioligands for the 5-HT receptors enable in vivo investigation of the 5-HT system in the human brain.
The 5-HT7 receptor (5-HT7R) is the most recently discovered 5-HT receptor, and its biologic functions are not fully elucidated. However, its implications in brain disorders such as depression and schizophrenia (1) make it an interesting target for both drug discovery and radioligand development. Both pharmacologic blockade of 5-HT7R and inactivation of the receptor gene led to an antidepressant-like behavioral profile in the forced swim test and in the tail suspension test (2–5). Pharmacologic blockade of 5-HT7R also presented anxiolytic effects in animal models of anxiety (5,6).
Currently, no well-validated radioligand is available for in vivo imaging of 5-HT7R. Interestingly, the 5-HT7R and the 5-HT1AR are the receptors for which 5-HT has the highest affinity. This is relevant for the aim of discovering a radioligand that is sensitive to changes in cerebral levels of endogenous 5-HT. If the competition model applies, the probability of measuring a signal change in response to a pharmacologic challenge that changes 5-HT levels will depend solely on the affinity of 5-HT to the target receptor (7).
Several potent and selective ligands for 5-HT7R have been discovered, but so far only a limited number of PET radioligands have been evaluated in vivo. 11C-DR4446 had good blood–brain barrier permeability and was metabolically stable but showed only a minimal specific binding component (8). The 5-HT7R antagonist (R)-3-(2-(2-(4-methylpiperidin-1-yl)ethyl)pyrrolidine-1-sulfonyl)phenol (SB-269970) has been used as a lead structure for discovering 18F-labeled radioligands. 18F-1-{2-[(2S)-1-(phenylsulfonyl)pyrrolidin-2-yl]ethyl}piperidin-4-yl 4-fluorobenzoate and 1-(2-{(2R)-1-[(2-18F-fluorophenyl)sulfonyl]pyrrolidin-2-yl}ethyl)-4-methylpiperidine (18F-2FP3) were evaluated ex vivo in rats and in vivo in cats (9,10). However, no input function was obtained while evaluating 18F-2FP3. Furthermore, the 5-HT7R binding distribution was not evaluated in cats, thus making it difficult to verify if the binding of 18F-2FP3 was specific to 5-HT7R. We recently also reported the evaluation of a 5-HT7R PET radioligand, 11C-Cimbi-806, that displayed selectivity in vitro, but the lack of blocking effect by SB-269970 in vivo led us to conclude that this compound does not selectively image the 5-HT7R in vivo (11).
Further work with the oxindole compound class led to the synthesis of a group of compounds including Cimbi-712 (3-{4-[4-(4-methylphenyl)piperazine-1-yl]butyl}p-1,3-dihydro-2H-indol-2-one) and Cimbi-717 (3-{4-[4-(3-methoxyphenyl)piperazine-1-yl]butyl}-1,3-dihydro-2H-indol-2-one). The affinity of both Cimbi-712 and Cimbi-717 is in the lower nanomolar range for 5-HT7R, with an inhibition constant (Ki) of 1.1 and 2.6 nM, respectively (12). Cimbi-712 is 2,191-fold selective for 5-HT7R over 5-HT1AR (Ki for the 5-HT1AR is 2,410 nM), whereas Cimbi-717 is 130-fold selective (Ki for the 5-HT1AR is 261 nM). From the outcome of the affinity testing on a range of receptors, Cimbi-712 displayed slightly higher selectivity for 5-HT7R than did Cimbi-717 (12).
Here, we report the 11C-labeling and in vivo evaluation, including receptor occupancy measurements, of two novel 5-HT7R selective PET radioligands in Danish Landrace pigs. For comparison between in vivo and postmortem receptor distribution, we also investigated the distribution of 5-HT7R in the pig brain.
MATERIALS AND METHODS
In Vitro Autoradiography
One brain hemisphere of a 30-kg Danish Landrace pig was sliced on a HM500OM cryostat (Microm International GmbH) in 20-μm coronal sections except for the cerebellum, which was sliced in the sagittal plane. Sections were thaw-mounted on Superfrost Plus glass slides (Thermo Scientific) and stored at −80°C until use. Autoradiography was performed at room temperature in 50 mM Tris-HCl buffer (pH 7.4) with 5 nM 3H-SB-269970 (1,476.3 MBq [39.9 mCi]/μmol at synthesis [January 2012]; PerkinElmer). Nonspecific binding was determined by adding 10 μM SB-258719 (3-methyl-N-[(1R)-1-methyl-3-(4-methyl-1-piperidinyl)propyl]-N-methylbenzenesulfonamide hydrochloride; Tocris Bioscience). Sections were preincubated in buffer for approximately 30 min and then incubated with radioactive buffer for 2 h. Sections were then washed for 1 × 5 min and for 2 × 10 min in 50 mM Tris-HCl buffer, rinsed in distilled H2O for 20 s, and dried before exposure to tritium-sensitive plates (Fujifilm Europe GmbH) for 19 d.
Calibration, quantification, and data evaluation of all autoradiography images were done with ImageJ analysis software (http://rsb.info.nih.gov/ij/). Regions of interest were hand-drawn around anatomic landmarks—for example, borders of sections—for each brain region, and the mean pixel density was measured in each brain region as outcome. A third-degree exponential calibration function of decay-corrected 3H-microscales (GE Healthcare) was used to convert mean pixel density to receptor binding measured in kBq/mg tissue equivalents (TE). Finally, the decay-corrected specific activity of 3H-SB-269970 was used to convert kBq/mg TE to fmol/mg TE.
Organic Synthesis
The precursors and reference compounds were synthesized as racemates as previously described (12,13).
Radioligand Preparation
11C-Cimbi-717
11C-methyl trifluoromethanesulfonate, produced using a fully automated system, was transferred in a stream of helium to a 1.1-mL vial containing the labeling precursor (0.3–0.4 mg, 0.8–1.0 μmol), 0.5 M K2CO3 (14 μL, 7 μmol), and MeCN (300 μL). The resulting sealed mixture was heated at 60°C for 5 min and then separated by high-performance liquid chromatography (HPLC) on a Luna 5-μm C18(2) 100-Å column (Phenomenex Inc.) (250 × 10 mm, 50:50 acetonitrile:0.01 M sodium borate buffer, at a flow rate of 6 mL/min). Retention times were 610 s for 11C-Cimbi-717 and 300 s for the precursor. The fraction corresponding to the labeled product was collected in sterile water (150 mL), and the resulting solution was passed through a solid-phase C18 Sep-Pak extraction column (Waters Corp.), which had been preconditioned with ethanol (10 mL), followed by isotonic sodium chloride solution (20 mL). The column was flushed with sterile water (3 mL). Then, the trapped radioactivity was eluted through a sterile filter with ethanol (3 mL), followed by isotonic sodium chloride solution (3 mL), into a 20-mL vial containing sodium phosphate–buffered saline (9 mL, 100 mM, pH 7), giving a 15-mL solution of racemic 11C-Cimbi-717 with a pH of approximately 7.
11C-Cimbi-712
11C-Cimbi-712 was produced as described previously (13). The HPLC fraction corresponding to the labeled product was collected in sterile water (150 mL), and the resulting solution was passed through a solid-phase C18 Sep-Pak extraction column, which had been preconditioned with ethanol (10 mL), followed by water (20 mL). The trapped radioactivity was eluted through a sterile filter with ethanol (3 mL) into a 20-mL vial containing sodium phosphate–buffered saline (9 mL, 100 mM, pH ∼7), giving a 12-mL solution of racemic 11C-Cimbi-712 with a pH of approximately 7.
Determination of Radiochemical Purity and Specific Radioactivity
Chemical and radiochemical purities were assessed on the same sample by HPLC analysis. Specific activity of the radioligands was calculated from 3 consecutive HPLC analyses (average) and determined by comparing the area of the ultraviolet absorbance peak corresponding to the radiolabeled product on the HPLC chromatogram with a standard curve relating mass to ultraviolet absorbance (λ = 225 nm). 11C-Cimbi-712 and 11C-Cimbi-717 were analyzed with HPLC using a Luna 5-μm C18(2) 100-Å column (150 × 4.6 mm, 50:50 MeCN:0.01 M borax buffer, at a flow rate of 2 mL/min). Retention times were 180 s for 11C-Cimbi-717 and 290 s for 11C-Cimbi-712).
Animal Procedure
Eight female Danish Landrace pigs (mean weight ± SD, 19 ± 2.0 kg) were used for in vivo PET imaging. All animal procedures were approved by the Danish Council for Animal Ethics (journal no. 2012-15-2934-00156).
PET Protocol
11C-Cimbi-712 was given as an intravenous bolus injection over 20 s, and the injected doses were 155 and 110 MBq for baseline scans (n = 2) and 123 and 73 MBq for scans for which SB-269970 was preadministered (n = 2). 11C-Cimbi-717 was also given as an intravenous bolus over 20 s, and the injected dose (mean ± SD) was 236 ± 117 MBq for baseline scans (n = 6; range, 115–435 MBq) and 147 ± 82.1 MBq for scans for which a preblocking agent was administered (n = 6; range, 40–277 MBq). The pigs were subsequently scanned for 90 min in list mode with a high-resolution research tomograph (HRRT; Siemens AG), with scanning starting at the time of injection (0 min). Immediately after the baseline scan (90 min), SB-269970 (Tocris Bioscience), a selective 5-HT7R antagonist (14), was given intravenously as a bolus infusion (0.2, 1.0, or 4.2 mg/kg/h), and rescanning started after 30 min of pretreatment with SB-269970. In one pig, a 0.5 mg/kg dose of prazosin (Sigma Aldrich), an α1 adrenergic receptor antagonist (15,16), was given as an intravenous infusion before injection of 11C-Cimbi-717.
Whole-Blood and Plasma Input Functions
During the first 30 min of the scans, radioactivity in arterial whole blood was continuously measured using an ABSS autosampler (Allogg Technology) counting coincidences in a lead-shielded detector. Concurrently, arterial whole blood was sampled manually at 2.5, 5, 10, 20, 30, 40, 50, 70, and 90 min after injection and radioactivity was measured in whole blood and plasma using a well counter (Cobra 5003; PerkinElmer). Cross calibration between the HRRT scanner, the autosampler, and the well counter allowed for the determination of plasma input functions.
Metabolite Analysis
Radiolabeled parent compound and metabolites were measured in plasma using HPLC with online radioactivity detection. In short, 11C-Cimbi-712 and 11C-Cimbi-717 were separated from their respective radiolabeled metabolites by direct injection of plasma in a column-switching HPLC system. Whole-blood samples were centrifuged (3,500 rpm, 7 min), and the supernatant plasma fraction was collected and filtered through a 0.45-μM syringe filter before analysis with online radioactive detection, as previously described (17,18).
Determination of Free Fraction
The free fraction of 11C-Cimbi-717 in pig plasma was measured using an equilibrium dialysis method as previously described (19) and calculated as the ratio between radioactivity in a buffer and plasma compartment after equilibrium between the chambers had been reached after 3 h.
Quantification of PET Data
Ninety-minute list-mode PET data were reconstructed into 38 dynamic frames of increasing length (6 × 10, 6 × 20, 4 × 30, 9 × 60, 2 × 180, 8 × 300, and 3 × 600 s). Images consisted of 207 planes of 256 × 256 voxels of 1.22 × 1.22 × 1.22 mm. A summed picture of all counts in the 90-min scan was reconstructed for each pig and used for coregistration to a standardized MR imaging–based atlas of the Danish Landrace pig brain, similar to that previously published (18,19). The time–activity curves were calculated for the following volumes of interest: cerebellum, cortex, hippocampus, lateral and medial thalamus, caudate nucleus, and putamen. The activity in the striatum is defined as the mean radioactivity in the caudate nucleus and putamen. The activity in the thalamus is calculated as the mean radioactivity in the lateral and medial thalamus. Radioactivity in all volumes of interest was calculated as the average of radioactive concentration (Bq/mL) in the left and right sides. The outcome measure in the time–activity curves was calculated as radioactive concentration in the volume of interest (in kBq/mL) normalized to the injected dose corrected for animal weight (in kBq/kg), yielding standardized uptake values (g/mL).
Distribution volumes (VT) for the volumes of interest were calculated on the basis of 3 different models: the 1-tissue-compartment (1-TC), 2-tissue-compartment (2-TC), and Logan linearization models. A single parent fraction curve fitted to a biexponential function, obtained from the average across all scans in the study, was used to generate individual metabolite-corrected plasma concentration curves. The VT of baseline and blocked conditions was used to determine the nondisplaceable distribution volume (VND) by use of the Lassen plot and thereby allow for calculation of the nondisplaceable binding potential BPND (20). All modeling was initiated with the same starting parameters, and SD coefficient variance was below 15% for macroparameters (the rate constant K1 and VT). Datasets that did not fulfill this criterion were not included in the results. Of the 176 fittings for 11C-Cimbi-717, 2 failed for the 1-TC, 31 failed for the 2-TC, and 7 failed for the Logan linearization model. Of the 40 attempts for 11C-Cimbi-712, 2 failed for the 1-TC, 12 failed for the 2-TC, and 11 failed for the Logan linearization model.
RESULTS
5-HT7R Distribution in the Pig Brain
The highest 3H-SB-269970 binding was found in small subregions of the thalamus (62.2 fmol/mg TE), amygdala (48.3 fmol/mg TE), and cingulate cortex (44.9 fmol/mg TE) (Table 1; Fig. 1). Low binding was found in the striatum and especially in the cerebellum (13.4 fmol/mg TE). In the cortical regions, binding increased toward the posterior parts of the pig brain, with the highest cortical binding found in the cingulate lobe (Supplemental Fig. 1; supplemental materials are available at http://jnm.snmjournals.org).
5-HT7R Distribution in Different Brain Regions Determined by 3H-SB-269970 Autoradiography
Representative sections of 5-HT7R autoradiography. Sections were incubated with 5 nM 3H-SB-269970 to determine total binding (upper row) and with 10 μM SB-258719 to determine nonspecific binding (lower row).
Radiochemistry
Radiolabeling of 11C-Cimbi-712 has previously been described (13). In brief, 11C-Cimbi-712 was labeled with a radiochemical yield (RCY) of 0.3–0.5 GBq and with a specific activity of 106 ± 68 GBq/μmol (n = 4) in a total synthesis time of 40–50 min.
Radiolabeling of 11C-Cimbi-717 underwent extensive optimization (Supplemental Table 1) as compared with 11C-Cimbi-712 (13), now using 0.3 mg of precursor, 14 μL of 0.5 M K2CO3 (1 equivalent), and 11C-CH3OTf on a fully automated system (decay-corrected RCY, ∼20%) (Fig. 2). Two major radioactive peaks were detected. The minor peak corresponded to 11C-Cimbi-717 (∼600 s) (further information is in the supplemental materials). The precursor eluted approximately 300 s before the radiolabeled product as indicated by the ultraviolet-absorption chromatogram (Supplemental Fig. 2). The radiosynthesis, including HPLC purification and formulation, generated an injectable solution of 11C-Cimbi-717 (radiochemical purity > 97%, chemical purity > 98%) within 40–50 min. Typically, 0.2–0.4 GBq of 11C-Cimbi-717 were isolated, with a specific activity of 69.7 ± 41.0 GBq/μmol (n = 12) at the end of synthesis. No precursor could be detected in the final formulation.
Radiosyntheses of 11C-Cimbi-712 and 11C-Cimbi-717. Reagents and conditions: 11C-CH3I, K2CO3, Pd2(dba)3, P(o-tolyl)3, 60°C, DMF:H2O (v:v 9:1) (A) and 11C-CH3OTf, 0.3 mg precursor, MeCN, K2CO3 (1 equivalent), 60°C, 5 min (B).
In Vivo Distribution
With 11C-Cimbi-712, the highest brain uptake was observed in the thalamus and the lowest uptake in the cerebellum. Time–activity curves displayed slow kinetics, with peak uptake after approximately 50 min (Fig. 3A). Pretreatment of the animals with 1.0 mg/kg/h SB-269970 decreased binding by approximately 50% in all regions. Like 11C-Cimbi-712, 11C-Cimbi-717 showed the highest brain uptake in the thalamus and the lowest uptake in the cerebellum (Fig. 3D). The 11C-Cimbi-717 time–activity curves indicated reversible binding in the pig brain, with fast brain uptake and a more pronounced decline in tissue time–activity curves than was found for 11C-Cimbi-712 (Fig. 3B). Pretreatment of the animal with SB-269970 (1.0 mg/kg/h and 4.2 mg/kg/h) decreased 11C-Cimbi-717 uptake and increased the rate of washout in all brain regions.
(A) Time–activity curves for 11C-Cimbi-712 at baseline (● and ▲, n = 2) and after blocking with 1.0 mg/kg/h SB-269970 (○ and △, n = 2). (B) Time–activity curves for 11C-Cimbi-717 at baseline (● and ▲, n = 6) and after blocking with 1.0 mg/kg/h SB-269970 (○ and △, n = 3). (C) MR-based atlas of pig brain. (D) 11C-Cimbi-717 baseline summed PET images from 0 to 90 min. (E) SB-269970–pretreated 11C-Cimbi-717 summed PET images from 0 to 90 min. (F) Color bar of standardized uptake value (SUV) (g/mL). Error bars = SEM.
Because no significant changes were observed between the baseline and blocking experiments in the composition of parent compound and other metabolites over the time course of the scan (data not shown), a single parent fraction obtained from the average over all scans with 11C-Cimbi-717 was computed and used to generate arterial plasma input time–activity curves. No changes in arterial input function were observed between the baseline and blocking experiments. Receptor binding of 11C-Cimbi-717 quantified with the 1-TC model showed the VT to be highest in the thalamus (16 ± 2.5 mL/cm3) and lowest in the cerebellum (6.6 ± 1.2 mL/cm3), indicating good regional separation, as also confirmed by the time–activity curves (Fig. 3B). VT was also calculated for all doses of SB-269970 and prazosin (Fig. 4C). A dose-dependent reduction in VT was observed with SB-269970, with significant reductions with the doses 1.0 mg/kg/h (P < 0.01) and 4.2 mg/kg/h (P < 0.001) in all regions except the cerebellum. The slope of the occupancy curve for prazosin was not significantly different from 0, supporting the likelihood that prazosin does not alter the binding of 11C-Cimbi-717 in vivo.
(A) VTs for 11C-Cimbi-712 quantified by 1-TC modeling. (B) Occupancy plots for 11C-Cimbi-712, where VTs for individual pigs are shown. (C) VTs for 11C-Cimbi-717 quantified by 1-TC modeling. (D) Occupancy plots for 11C-Cimbi-717 with 3 different doses of SB-269970. In A and C, bars represent mean ± SD; in B and D, bars represent mean ± SEM. **P < 0.01 and ***P < 0.001 in comparison to baseline data within each volume of interest on statistical analysis with 2-way ANOVA and Bonferroni posttest. Tha = thalamus; Str = striatum; Ctx = cortex; Hip = hippocampus; Cb = cerebellum.
Receptor binding of 11C-Cimbi-712 was quantified with the 1-TC model. VTs were—as with 11C-Cimbi-717—highest in the thalamus (87.5 and 54.9 mL/cm3) and lowest in the cerebellum (35.2 and 32.1 mL/cm3). The VTs of the pretreated scans revealed a decrease in binding, although the decrease could not be reliably statistically assessed (Fig. 4A). The occupancy slopes were, however, significantly different from 0, verifying that SB-269970 did block the binding of 11C-Cimbi-712.
Based on occupancy plots of 11C-Cimbi-717 (Fig. 4D), the VND and occupancy were extracted from the linear regression. The 11C-Cimbi-717 VND was on average 2.1 ± 0.8 mL/cm3 (mean ± SD, n = 5), and intravenous pretreatment with 0.2 mg/kg/h SB-269970 resulted in 25.4% occupancy, whereas pretreatment with the higher doses of 1.0 and 4.2 mg/kg/h resulted in 59.9% and 75.4% occupancy, respectively. Consequently, the BPND of 11C-Cimbi-717 at baseline in the thalamus and cerebellum were calculated to be 6.4 ± 1.2 and 2.1 ± 0.6 (n = 6), respectively.
Treatment with 1.0 mg/kg/h SB-269970 before 11C-Cimbi-712 injection resulted in an occupancy of 75.3% and a VND of 8.57 mL/cm3 (Fig. 4B). Consequently, the BPND of 11C-Cimbi-712 in the thalamus is 7.3 based on average VTs and VNDs.
Comparison of the in vitro and in vivo binding data revealed significant positive correlations between 3H-SB-269970 binding in vitro and 11C-Cimbi-712 (P = 0.016) and 11C-Cimbi-717 (P < 0.001) binding in vivo (Fig. 5).
Comparison of 5-HT7R brain distribution determined by 3H-SB-269970 autoradiography and by in vivo PET experiments using baseline VTs of 11C-Cimbi-712 (n = 2) (A) and 11C-Cimbi-717 (n = 6) (B). Cb = cerebellum; Ctx = cortex; Hip = hippocampus; Str = striatum; Tha = thalamus. Error bars represent mean ± SEM.
The free fraction of 11C-Cimbi-717 in plasma was on average 6.7% at equilibrium.
Metabolism
Radio-HPLC analysis of pig plasma revealed that after intravenous injection, 11C-Cimbi-717 was relatively slowly metabolized (50% remaining after 30 min), and two radiometabolites were observed (Fig. 6). The polar metabolite fraction increased during the 90-min scanning time. The other more lipophilic metabolite was detected in relatively low amounts (∼15%). However, this metabolite was less lipophilic than 11C-Cimbi-717.
Metabolism of 11C-Cimbi-717. Three radioactive compounds could be detected with radio-HPLC: 11C-Cimbi-717 (parent compound) and two radiolabeled metabolites, both of which had lower retention time than parent compound. Data are presented as mean ± SEM.
DISCUSSION
Here we have presented the radiosyntheses and in vivo evaluation of two novel 5-HT7R radioligands, 11C-Cimbi-712 and 11C-Cimbi-717. To enable comparisons between in vitro and in vivo data, we have also assessed for the first time the cerebral 5-HT7R distribution in the pig brain.
Consistent with observations in the human brain (21,22) and brain membrane binding assays in different species (23), in vitro data with 3H-SB-269970 showed that 5-HT7R distribution in the pig brain was fairly homogeneous in the neocortex but differed more across the remaining brain regions. Our data also showed that the hypothalamus, thalamus, and amygdala are areas with high 5-HT7R density, as is in line with the involvement of 5-HT7R in circadian rhythm controlled by the suprachiasmatic nucleus (24,25). As described for human brain tissue, 3H-SB-269970 binding in the pig striatum was low. We found low 5-HT7R binding in the pig cerebellum, corresponding to a study showing low amounts of 5-HT7R messenger RNA in cerebellum in pigs (26). We found a general high binding in the cerebral cortex of the pig brain as reported earlier for membrane binding assays (23). Although the cortical binding pattern did not resemble that of 5-HT2AR or 5-HT1AR (27), we confirmed the specificity of 3H-SB-269970 through autoradiographic blocking experiments with the specific 5-HT1AR and 5-HT2AR antagonists WAY-100635 and MDL-100907 (data not shown). No displacement was found with WAY-100635, but in accordance with the affinity of MDL-100907 (Ki ∼ 50 nM) toward 5-HT7R (28), a 40% displacement was observed.
Although 11C-Cimbi-717 was successfully labeled in sufficient RCY for the PET experiments, we did not succeed in further optimizing the RCY. We speculate that the major radioactive side product is due to methylation of the oxindole moiety, and if so, the introduction of a protecting group at the N-1 position of the oxindole may reduce the formation of the observed side product. Although the racemic material could be separated on chirale HPLC, the stereogenic center epimerizes rapidly. Thus, all compounds in this study have been investigated as the racemate. The quantification is therefore an average of binding of two enantiomers, and it is possible that one of the enantiomers is superior to the other.
In vivo evaluation of 11C-Cimbi-717 demonstrated a high brain uptake and a binding distribution similar to that of 5-HT7R found in vitro; thalamus has the highest and cerebellum the lowest VT, consistent with the 3H-SB-269970 autoradiography. 11C-Cimbi-717 binding in the striatum was not quite in line with the correlation between in vivo binding in other regions and in vitro autoradiography binding. Further analysis of these differences revealed that this discrepancy between 5-HT7R binding in vivo and in vitro was larger for the putamen than for the caudate nucleus; thus, the discrepancy could be due to binding to an unknown target, for which either of the enantiomers of 11C-Cimbi-717 has affinity. However, the occupancy plot did not reveal lower displacement in the striatal regions. We suspect that the difference observed between the in vitro and in vivo results are due to inhomogeneity in the autoradiography experiments. With autoradiography, only a few cross sections were evaluated, whereas in PET, binding in the whole region is evaluated. The putamen is near the amygdala in the pig brain, and partial-volume effects may affect the signal from this high-binding region, leading to an overestimation of the 11C-Cimbi-712 and 11C-Cimbi-717 binding in the putamen. The correlation between in vitro and in vivo binding was not as strong for 11C-Cimbi-712 as for 11C-Cimbi-717. Along with the difference in binding in the striatum, 11C-Cimbi-712 also displayed equal uptake in the hippocampus and cerebellum, a finding that is not in line with what we see with 3H-SB-269970 autoradiography. This discrepancy in binding between 11C-Cimbi-712 and 11C-Cimbi-717 in the hippocampus could be due to off-target binding. Highest binding is, however, still observed in the thalamus, consistent with the autoradiography results.
Pretreatment with SB-269970 resulted in a dose-dependent decrease in 11C-Cimbi-717 binding in the pig brain supporting 5-HT7R selectivity in vivo for 11C-Cimbi-717. VT decreased in all regions examined, including the cerebellum. Although this decrease in binding in the cerebellum was not statistically significant, it confirms the in vitro autoradiography data and the literature data that have found 5-HT7R to be present in the cerebellum (26). This decrease in binding in the cerebellum also invalidates the cerebellum as a reference region for a reference tissue model analysis of the PET data. In the absence of a reference region, we determined the nondisplaceable binding using the occupancy plot. Because of the affinity of Cimbi-717 for the α1 adrenergic receptor (Ki = 47 nM) (12), we ensured that pretreatment with prazosin, an α1 adrenergic receptor antagonist, did not result in any significant decrease in 11C-Cimbi-717 VT. 11C-Cimbi-712 binding was blocked with SB-269970, supporting the possibility that this radioligand also labels 5-HT7R specifically. The slow kinetics of 11C-Cimbi-712, however, complicate modeling of the binding, resulting in large variations in the outcome measures (VT) and consequently also larger uncertainties in the calculated occupancy and VND.
The average nonspecific binding in the pig brain, VND, of 11C-Cimbi-717 as determined by the occupancy plot was 2.1 mL/cm3 and thus comprised about 15% of the VT in the high-binding regions. This ratio of specific-to-nonspecific binding is larger than what is obtained by other PET ligands evaluated in the same species, for example, approximately 50% for 11C-NS14492 and approximately 35% for 11C-SB2047145 (19,29).
Modeling of the 11C-Cimbi-717 data was done with the 1-TC model first because of the simplicity of the model, second because more regions converged with this model compared with 2-TC, and finally because the Akaike information criterion was generally lower for the 1-TC than for the 2-TC (Supplemental Table 2). The Logan linearization model underestimated the VT values by 10%–15% compared with both the 1-TC and the 2-TC.
The systemic metabolism of 11C-Cimbi-717 was relatively slow compared with other radioligands evaluated in pigs (11,18,19), with approximately 60% of the total plasma activity arising from the parent compound left after 20 min. Metabolism was nonsignificantly faster after blockade with SB-269970, as could be explained by an increased availability in the blood and thus an increased availability to enzymatic degradation. No effects on metabolism were observed with prazosin.
CONCLUSION
11C-Cimbi-712 and 11C-Cimbi-717 were both successfully radiolabeled in sufficient RCYs for in vivo evaluation in the pig. Of the two novel radioligands for brain imaging of 5-HT7R, 11C-Cimbi-717 generated the highest brain uptake and showed more reversible tracer kinetics than 11C-Cimbi-712—benefits that are important for quantification. Both 11C-Cimbi-712 and 11C-Cimbi-717 had a regional distribution pattern compatible with 5-HT7R distribution in the pig brain, as assessed independently by autoradiography. Finally, 11C-Cimbi-717 showed a dose-dependent decrease in binding after pretreatment with the 5-HT7R–specific antagonist SB-269970. We conclude, on the basis of these preclinical data, that 11C-Cimbi-717 may be a useful radioligand for in vivo imaging of 5-HT7R binding sites in the human brain.
DISCLOSURE
The costs of publication of this article were defrayed in part by the payment of page charges. Therefore, and solely to indicate this fact, this article is hereby marked “advertisement” in accordance with 18 USC section 1734. Financial support by the Intra European Fellowship (MC-IEF-275329), the Faculty of Health at the University of Copenhagen, and particularly the Lundbeck Foundation is gratefully acknowledged. The John and Birthe Meyer Foundation and the Toyota Foundation are acknowledged for financial support for the HRRT scanner and HPLC system, respectively. No other potential conflict of interest relevant to this article was reported.
Acknowledgments
We thank the staff at the PET and Cyclotron Units for expert technical assistance. We also thank Mette Værum Olesen for excellent technical assistance with animal preparation.
Footnotes
Published online Feb. 24, 2014.
- © 2014 by the Society of Nuclear Medicine and Molecular Imaging, Inc.
REFERENCES
- Received for publication April 3, 2013.
- Accepted for publication December 4, 2013.
