Abstract
Fatty acid amide hydrolase (FAAH) is a key enzyme in the endocannabinoid system. N-(3,4-Dimethylisoxazol-5-yl)piperazine-4-[4-(2-fluoro-4-[11C]methylphenyl)thiazol-2-yl]-1-carboxamide ([11C]DFMC) was developed as an irreversible-type positron emission tomography (PET) tracer for FAAH. Here, we attempted to noninvasively estimate rate constant k3 (rate of transfer to the specifically-bound compartment) as a direct index for FAAH in the rat brain. First, the two-tissue compartment model analysis including three parameters [K1−k3, two-tissue compartment model for the irreversible-type radiotracer (2TCMi)] in PET study with [11C]DFMC was conducted, which provided 0.21 ± 0.04 ml·cm−3·min−1 of the net uptake value (Ki), an indirect index for FAAH, in the FAAH-richest region (the cingulate cortex). Subsequently, to noninvasively estimate Ki value, the reference model analysis (Patlak graphical analysis reference model) was tried using a time-activity curve of the spinal cord. In that result, the noninvasive Ki value (KREF) was concisely estimated with high correlation (r > 0.95) to Ki values based on 2TCMi. Using estimated KREF value, we tried to obtain calculated-k3 based on previously defined equations. The calculated k3 was successfully estimated with high correlation (r = 0.95) to direct k3 in 2TCMi. Finally, the dose relationship study using calculated k3 demonstrated that in vivo ED50 value of [3-(3-carbamoylphenyl)phenyl] N-cyclohexylcarbamate, a major inhibitor of FAAH, was 66.4 µg/kg in rat brain. In conclusion, we proposed the calculated k3 as an alternative index corresponding to regional FAAH concentrations and suggested that PET with [11C]DFMC enables occupancy study for new pharmaceuticals targeting FAAH.
SIGNIFICANCE STATEMENT In the present study, we proposed calculated k3 as an alternative index corresponding with fatty acid amide hydrolase concentration. By using calculated k3, in vivo ED50 of [3-(3-carbamoylphenyl)phenyl] N-cyclohexylcarbamate was successfully estimated to be 66.4 µg/kg for rats. Thus, we demonstrated the pharmacological utility of positron emission tomography with N-(3,4-dimethylisoxazol-5-yl)piperazine-4-[4-(2-fluoro-4-[11C]methylphenyl)thiazol-2-yl]-1-carboxamide.
Introduction
The endocannabinoid system is known as a key biologic system having retrograde neurotransmission in the central nervous system (Devane et al., 1992; Bayewitch et al., 1995) and has been reported to regulate a broad range of physiologic processes in multiple disorders, such as pain, neuroinflammation, anxiety, neurodegenerative disorders, cancer, epilepsy, and metabolic syndrome (Pacher et al., 2006). Endocannabinoids (anandamide and 2-arachidonoyl glycerol) are synthesized by several enzymes, depending on the intracellular Ca2+ concentration, on postsynaptic neurons and metabolized by fatty acid amide hydrolase (FAAH) and monoacylglycerol lipase (Piomelli, 2003; Ahn et al., 2008). Among the endocannabinoids, anandamide has important roles for the activation of microglia in neuroinflammation (Raboune et al., 2014; Malek et al., 2015). Moreover, a great deal of evidence from preclinical studies indicates that elevating anandamide concentrations through the inhibition of FAAH can mitigate pain and neuroinflammation (Schlosburg et al., 2009). Several FAAH inhibitors, classified as urea, carbamate, and keto-heterocycle derivatives, have been developed (Seierstad and Breitenbucher, 2008) and progressed to clinical trials to treat inflammatory pain, cannabis dependence, and schizophrenia (Kathuria et al., 2003; Li et al., 2012). Subsequently, to further understand the function of FAAH and to research drug kinetics in vivo, several positron emission tomography (PET) tracers for FAAH were synthesized based on the inhibitors (Wilson et al., 2011; Rotstein et al., 2014; Kumata et al., 2015; Shimoda et al., 2015, 2016) (Fig. 1).
Current PET tracers for FAAH imaging. (A) Reversible-type PET tracers. (B) Irreversible-type PET tracers.
PET is frequently used as an imaging modality or quantification tool for basic and clinical research to elucidate drug kinetics, molecular density, and distribution in vivo. In general, PET studies with reversible-type radiotracers can acquire the nondisplaceable binding potential (BPND) as a reasonable index to estimate receptor density (Innis et al., 2007), which permits pharmacological applications, such as the measurement of dose-occupancy relationships for drugs (Saijo et al., 2009). To the best of our knowledge, although two reversible-type PET tracers for FAAH (Fig. 1A) have been developed, both tracers failed to estimate sufficient BPND values (Liu et al., 2013; Wang et al., 2016). In contrast, PET studies with the irreversible-type radiotracer were generally conducted using two-tissue compartment model analysis with three parameters [k4 = 0, two-tissue compartment model for the irreversible-type radiotracer (2TCMi); Fig. 2]. Of the three parameters, the k3 consisted of Bmax multiplied by the association rate constant (kon) to the target molecule. Thus, estimated k3 gives the most important information for target molecules. Unfortunately, a directly estimated k3 is usually unstable because of including moderate stochastic varieties. Therefore, the macro parameter [e.g., net uptake value (Ki)] is often estimated as the stable quantitative index for target molecules in place of BPND value (Egerton et al., 2010; Carter et al., 2012; Rusjan et al., 2013; Frick et al., 2015).
Schematic of two-tissue compartment model for irreversible-type radiotracer. CP compartment represents free radiotracer in the plasma. C1 compartment expresses free and nonspecific binding of radiotracer in brain tissue past the blood-brain barrier (BBB). C2 compartment displays specific binding of radiotracer with target molecule. K1 and k2 describe the influx and efflux rates of radiotracer between CP and C1. k3 and k4 exhibit the association and dissociation rates of radiotracer to target molecule.
Recently, we have developed N-(3,4-dimethylisoxazol-5-yl)piperazine-4-[4-(2-fluoro-4-[11C] methylphenyl)thiazol-2-yl]-1-carboxamide ([11C]DFMC) (Fig. 1B), which has higher affinity for FAAH (IC50 = 6.1 nM) than the primary PET tracer [11C-carbonyl]-6-hydroxy-(1,1’-biphenyl)-3-yl cyclohexylcarbamate ([11C]CURB) (IC50 = 30 nM). Moreover, [11C]DFMC showed high uptake in rat brain, and the radioactivity was irreversibly trapped (Shimoda et al., 2016).
In our study, we first conducted a quantitative PET analysis with blood sampling to estimate the kinetic parameters of [11C]DFMC in various brain regions of the rat. Subsequently, we attempted to noninvasively estimate the net uptake value (KREF) using a reference tissue model and aimed to establish an alternative index for the direct relationship with FAAH concentrations. Finally, to demonstrate the pharmacological utility of PET with [11C]DFMC, we attempted to measure the ED50 of [3-(3-carbamoylphenyl)phenyl] N-cyclohexylcarbamate (URB597), a clinically used inhibitor of FAAH, using PET with [11C]DFMC in vivo.
Materials and Methods
General
All chemical reagents and organic solvents were purchased from Sigma-Aldrich (St. Louis, MO), FUJIFILM Wako Pure Chem. (Osaka, Japan), or Nacalai Tesque (Kyoto, Japan) and used without further purification. The commercially available compound URB597 (≥98% purity) was purchased from Sigma-Aldrich and dissolved in saline containing 10% ethanol and 5% Tween 20 for animal experiments. 11C was produced using a cyclotron (CYPRIS HM-18; Sumitomo Heavy Industries, Tokyo, Japan). All radioactive values were used with decay correction (a half-life of 11C: 20.4 minutes) (Lederer et al., 1967).
Subjects
Male Sprague-Dawley rats (7–10 weeks old, n = 30) were purchased from Japan SLC (Shizuoka, Japan), housed in a temperature-controlled environment with a 12-hour light/dark cycle, and fed a standard diet. All animal experiments were performed according to the recommendations specified by the Committee for the Care and Use of Laboratory Animals of the National Institutes for Quantum and Radiologic Science and Technology and Animal Research: Reporting of In Vivo Experiments guidelines.
Radiotracer
[11C]DFMC was synthesized according to a previous report (Shimoda et al., 2016). Briefly, [11C]DFMC was synthesized using a C-11C coupling reaction of an aryl boronic ester precursor with [11C]methyl iodide in the presence of a Pd catalyst. Over 370 MBq of [11C]DFMC was obtained with radiochemical purity of >99% and molar activity of >37 GBq/µmol.
PET Study
PET Analysis with Blood Sampling.
Prior to the PET scan, a rat (n = 4; 309 ± 14 g) was implanted with a polyethylene catheter (FR2; Imamura, Tokyo, Japan) inserted into the left femoral artery for blood sampling. Subsequently, the rat was secured in a custom-designed chamber and placed in a small-animal PET scanner (Inveon; Siemens Medical Solutions, Knoxville, TN). Body temperature was maintained using a 40°C water circulation system (T/Pump TP401; Gaymar Industries, Orchard Park, NY). A 24-gauge intravenous catheter (Terumo Medical Products, Tokyo, Japan) was placed into the tail vein of the rat. A bolus of [11C]DFMC (1 ml, 52–57 MBq, 0.3–0.9 nmol) was injected at a flow rate of 0.5 ml/min via a tail vein catheter. Dynamic emission scans in three-dimensional list mode were performed for 90 minutes (10 second × 12 frames, 20 seconds × 3 frames, 30 seconds × 3 frames, 60 seconds × 3 frames, 150 seconds × 3 frames, and 300 seconds × 15 frames). The acquired PET dynamic images were reconstructed by filtered back projection using a Hanning’s filter with a Nyquist cutoff of 0.5 cycle/pixel. The time-activity curves (TACs) of [11C]DFMC were acquired from volumes of interest in the cingulate cortex, striatum (caudate/putamen), hippocampus, thalamus, hypothalamus, pons, and cerebellum by referring to a rat brain magnetic resonance imaging (MRI) template using PMOD software (version 3.4; PMOD technology, Zurich, Switzerland). The radioactivity was decay corrected to the injection time and is expressed as the standardized uptake value (SUV).
For the blocking study, a rat (281 g) cannulated as described above was intravenously injected with URB597 at a concentration of 3 mg/kg (0.28 ml vehicle) via the tail vein catheter while under anesthesia. After 30 minutes of anesthesia, a PET assessment with [11C]DFMC (56 MBq; 1.6 nmol) was conducted as described above. volumes of interest were drawn on the spinal cord in addition to general regions.
For counting radioactivity, blood samples were manually collected into microtubes containing heparin (1 µl) at intervals of 20 seconds (0.05 ml for 120 seconds) and 0.5 (0.05 ml for 1 minute), 1 (0.05 ml for 2 minutes), 5 (0.08 ml for 10 minutes), 30 (0.3 ml), 60 (0.4 ml), and 90 minutes (0.5 ml) after initiation of the PET scan. Blood samples were centrifuged at 15,000g at 4°C to separate the plasma. The radioactivity in the whole blood and plasma was measured by a 1480 Wizard auto-gamma scintillation counter (PerkinElmer, Waltham, MA). The radioactivity was corrected for decay. For metabolite analysis, six plasma samples were separated at 1 (0.02 ml), 5 (0.02 ml), 15 (0.05 ml), 30 (0.1 ml), 60 (0.2 ml), and 90 minutes (0.3 ml) after the injection.
Metabolite analysis was performed as described previously (Yamasaki et al., 2014). Briefly, whole blood samples were treated to separate the plasma, which was deproteinized with an equivalent amount of acetonitrile. An aliquot of the supernatant obtained from the plasma was analyzed using a high-performance liquid chromatography system with a radiation detector (Takei et al., 2001). Plasma protein binding was not determined in our study. The time curves for a fraction of unchanged [11C]DFMC in the plasma were fitted using three exponential equations and subsequently used for kinetic analyses.
Test-Retest PET Studies.
Four rats were used twice within 7 days (285 ± 7 g at first and 323 ± 9 g at the second scan) for PET assessments with [11C]DFMC (47−61 MBq; 0.5−0.8 nmol), and the reliability of the data was assessed using the intraclass correlation coefficient (ICC). The parameters were calculated as follows:
Relative difference (%) = (scan 2 − scan 1)/scan 1 × 100
Test-retest variability (%) = |scan 2 − scan 1|/[(scan 2 + scan 1)/2] × 100
Percentage of coefficients of variation (%COV) = S.D./mean × 100
ICC with BSMSS as “mean sum of squares between subjects” and WSMSS as “mean sum of squares within subjects”: ICC = (BSMSS − WSMSS)/(BSMSS + WSMSS). An ICC value of −1 denotes no reliability, whereas a value of 1 indicates maximum reliability (Elmenhorst et al., 2012).
Theory
Compartment Model Analysis for Irreversible-Type PET Tracers.
To estimate kinetic parameters in PET with [11C]DFMC, 2TCMi (Fig. 2) was conducted. Each rate constant was derived from the following equations:
(1)
(2)
(3)in which K1 describes the influx rate of radiotracer from the plasma compartment (CP) to the free and nonspecific compartment (C1); k2 represents the efflux rate of radioligand from C1 to CP; k3 describes the transfer from C1 to the specific-bound compartment (C2). F is the blood flow, E is the first pass extraction factor, Vd is the distribution volume of the radiotracer in the C1 compartment, fND is the tissue-free fraction, kon is the [11C]DFMC-FAAH association rate constant, and Bmax is the concentration of FAAH. In addition, the Ki as the quantitative index for the net uptake volume of [11C]DFMC with FAAH was determined as follows:
(4)
To compare the accuracy of Ki values based on 2TCMi (Ki2TCMi), Patlak graphical analysis (PGA) (Patlak et al., 1983) with linear regression was also performed (the slope of a regression line in PGA theoretically equals the Ki value).
A Reference Tissue Model for Irreversible-Type PET Tracers.
When reference tissue can be employed, the application of PGA as a reference method [PGA reference model (PGAREF)] is possible (Patlak and Blasberg, 1985). In this case, the procedure merely replaces CP(t) by TAC in the reference tissue. In accordance with a previous report (Patlak and Blasberg, 1985), the slope in graphical analysis reflects the following relation:
(5)in which, K1′, k2′, and Keq′ indicate input rate, output rate, and equilibrium constant in the reference area, respectively. In the reference region without irreversible binding, it may be reasonable to assume that Keq′ = 0 (Patlak and Blasberg, 1985).
Here, 
is replaced as A,
(6)
Noninvasive Estimation of the Alternative k3 Value Based on the KREF Value.
In this study, we attempted to noninvasively estimate the alternative k3 (defined as calculated k3) values, since the KREF value is an indirect index for FAAH concentration. Calculated k3 is induced by modifying eq. 6.
(7)Here, the regional K1 and k2 values were fixed by averaged values in 2TCMi analyses (n = 4). The constant A was also displaced as follows:
(8)
Conditions for the Use of Calculated k3 Values
In this study, we proposed a calculated k3 as a new quantitative index for FAAH concentrations. However, there are several conditions for the use of calculated k3, which are as follows:
To estimate averaged K1, k2, and constant A, several repeated PET assessments with blood sampling and compartment model analyses are essential in advance of these estimates.
It is required that there are no differences in influx (K1) and efflux (k2) rates of radiotracer between the research subject and the baseline subject.
The calculated k3 remains an unstable value and therefore includes a specific variation based on individual differences. To surmount this disadvantage, several samples sizes (n ≥ 3, at least) should be considered.
Although there are several limitations, it would be valuable to consider whether calculated k3 can be adapted because calculated k3 in multidose response assays using PET could be more easily obtained than direct k3 with blood sampling.
Multidose URB597 Treatment-Response Assays
A series of dynamic PET scans ([11C]DFMC: 38−60 MBq; 0.3−1.0 nmol) without blood sampling were performed for each rat (n = 3 for each dose; 241−319 g) 30 minutes after administration with different doses of URB597 (0.003, 0.01, 0.03, 0.1, 0.3, 1, and 3 mg/kg in 0.3-ml vehicle).
The results of the inhibitory experiments were subjected to nonlinear regression analysis using Prism 5 (GraphPad Software, La Jolla, CA), and ED50 values of URB597 were calculated using each averaged calculated k3 value of several brain regions (cingulate cortex, striatum, hippocampus, thalamus, and cerebellum).
Statistical Methods
Goodness of fit was evaluated using the Akaike information criterion (Akaike, 1974) and the model selection criterion (Handbook, 1995). Values are given as mean ± S.D. The %COV was estimated from the diagonal of the covariance matrix of the fitting. All data analyses were performed using GraphPad Prism v5.0 (GraphPad Software).
Results
Invasive Quantitative PET Analysis Using [11C]DFMC.
Figure 3 show representative averaged PET/MRI images (A) and TACs (B) in brain regions. High uptake of radioactivity was detected in the cingulate cortex and striatum (caudate/putamen), and moderate radioactivity was detected in the hippocampus, thalamus, cerebral cortex, and cerebellum. Meanwhile, radioactivity in the hypothalamus and pons was relatively low. These radioactive distribution patterns corresponded with the distribution of FAAH concentrations (Thomas, et al., 1997). TACs in FAAH-rich brain regions constantly increased without clearance after the injection of [11C]DFMC.
Representative PET/MRI-fused images (A) and time-activity curves (B) of [11C]DFMC in brain regions (n = 4). PET images were summed between 0 to 90 minutes of acquisition data. The radioactivity was expressed by SUV. Ce, cerebellum; Ci, cingulate cortex; Hi, hippocampus; Hy, hypothalamus; Po, pons; St, striatum; Th, thalamus.
Figure 4A shows the metabolite-corrected plasma input function of [11C]DFMC. The unchanged [11C]DFMC in the arterial plasma peaked at 4.19 ± 0.72 SUV 2 minutes after the injection and declined to 0.35 ± 0.03 SUV 5 minutes after the injection and 0.04 ± 0.01 SUV 90 minutes after the injection. The metabolic rate of [11C]DFMC was relatively slow: >65% of the parent compound remained 30 minutes after the injection, and roughly 30% of the parent compound remained 90 minutes after the injection.
Curve-fitting analysis using input function in PET with [11C]DFMC. (A) Plasma input curves of [11C]DFMC. (B) Comparison between 1TCM and 2TCMi curve fittings of data from the cingulate cortex. (C) Correlation plots between Ki values based on 2TCMi and Ki values based on PGA in rat brains (n = 4).
Figure 4B shows TACs with a one-tissue compartment model (1TCM) containing CP and C1 compartments only, and 2TCMi fitting curves in the cingulate cortex. The 1TCM showed a poorly fitting curve for TACs. Conversely, the fitting curve for 2TCMi showed a good shape and indicated good scores (see Supplemental Table 1) regarding the goodness of fit. Thus, the 2TCMi is an adequate kinetic model for this radiotracer. Detailed full kinetic parameters in brain regions are shown in Table 1. Of the rate constants, the directly estimated k3 values were acquired as 0.08–0.18 minutes−1 with 4.6%–9.5%COV in the regions of interest. The macroparameter Ki values for the quantitative uptake value of [11C]DFMC were obtained as 0.12–0.21 ml·cm−3·min−1 with 0.0%–0.8%COV in the investigated brain regions.
Kinetic rate constants estimated with 2TCMi in PET with [11C]DFMC (n = 4, mean ± S.D.) The %COV was expressed within parentheses.
Next, to validate the accuracy of the estimated Ki values based on the compartmental analysis (Ki2TCMi), we compared these values with those based on graphical analysis (KiPGA), as shown in Fig. 4C. The slope of the resulting regression line was almost 1 (0.967) and the r-value was 0.995. These results indicate a high correlation between KiPGA and Ki2TCMi values, which suggests that Ki2TCMi values were estimated with high reliability.
Validation Studies for the PGA Reference Model.
Figure 5, A and B show representative PET-averaged images (A) and TACs (B) in the cingulate cortex and spinal cord of rats pretreated with or without URB597 (3 mg/kg). Radioactivity in the cingulate cortex of the baseline subject was accumulated at a high level without clearance during the PET scan, which was significantly decreased by pretreatment with URB597. Meanwhile, uptake of radioactivity between the baseline and blocking subjects showed no significant differences (P = 0.106) in the spinal cord (Fig. 5B), which suggests that the spinal cord adequately serves as the reference region.
Validation study for the use of the reference model. (A) Representative PET images of the baseline (top) and blocking (bottom, pretreatment with URB597 of 3 mg/kg) rats. (B) Time-activity curves of [11C]DFMC in the cingulate cortex (Ci) and spinal cord (Sc) of the baseline (n = 4) and blocking (n = 3) rats. (C) Correlation plots between averaged KREF value based on PGAREF and the averaged Ki value based on 2TCMi in rat brains (n = 4). (D) The relationship between an averaged calculated k3 value (estimating from measured KREF, constant A, averaged K1, k2, and the bias of reference region) and averaged direct k3 values in the compartment model analysis.
Subsequently, to noninvasively estimate net uptake values (defined as KREF), we performed the PGAREF using the TAC of the spinal cord as a reference region. The validity of KREF values was evaluated by comparing with Ki2TCMi values. Figure 5C shows the relationship between the averaged KREF and Ki2TCMi values in baseline subjects (n = 4). The slope (k2′/K1′, defined as constant A, see eq. 8) of the regression line was 0.075 with high correlation (r = 0.981, P < 0.001) and a small intercept (−0.004).
Additionally, to support the accuracy of KREF estimations, the reproducibility of the KREF values was evaluated by a test-retest study using PET with [11C]DFMC. Table 2 shows the reproducibility of the test-retest PET study for the estimation of KREF. In the cingulate cortex, the FAAH-richest region in the brain, the percentage of variability, ICC, and Pearson’s r were 8.8, 0.836, and 0.900, respectively. Additionally, the correlation (Pearson’s r) between test and retest outcomes in all regions of interest was 0.891, indicating high reproducibility of KREF values after PET with [11C]DFMC.
Reliability of outcome parameters in test-retest PET studies with [11C]DFMC (n = 4)
Estimation for Calculated k3 Values.
The KREF value is an indirect index of FAAH concentration due to the macroparameter containing K1, k2, and k3 rate constants. We attempted to noninvasively estimate an alternative k3 value (defined as calculated k3) by inserting measured KREF values, constant A, and fixed values (averaged K1 and k2 values in 2TCMi). Here, because KREF values included a small bias (−0.004) compared with Ki2TCMi values as described above (Fig. 5C), we modified eq. 7 as follows:
(9)
Figure 5D exhibits the relationship between averaged calculated k3 and the directly estimated k3 values. Although both k3 values were robust, the slope of the regression line was 1.177 and showed high correlation (r = 0.946, P = 0.001). This result suggested that the calculated k3 value would be useful as a direct index of FAAH concentration.
Estimation for ED50 of URB597 Using Calculated k3.
Figure 6 shows representative PET/MRI images (A) of rat brains treated with multiple doses of URB597 and dose responses of calculated k3 values (B) in the cingulate cortex, striatum, hippocampus, thalamus, and cerebellum. Radioactivity in all brain regions was gradually decreased by increasing the URB597 doses (Fig. 6A). Treatment with 1 mg/kg URB597 almost completely blocked the accumulation of radioactivity in the brain, and the calculated k3 was close to zero. Thus, the ED50 value for URB597 was estimated to be 66.4 µg/kg in rat brain (Fig. 6B).
Dose-response assay using URB597. (A) Representative PET/MRI images of [11C]DFMC in the brains of rats treated with different doses (0.01, 0.1, 0.3, and 1 mg/kg) of URB597. (B) The relationship between calculated k3 values (minute−1) and doses of URB597 in various brain regions. Averaged values of calculated k3 in the cingulate cortex, striatum, hippocampus, thalamus, and cerebellum are plotted against the dose of URB597. Ce, cerebellum; Ci, cingulate cortex; Hi, hippocampus; St, striatum; Th, thalamus.
Discussion
In our study, [11C]DFMC, the most recently developed PET tracer for FAAH imaging, was used for the index measurement reflecting FAAH concentrations in the brain. The first quantitative PET study for FAAH was performed using [11C]CURB, a primary PET tracer for FAAH. In that report, the rate constant k3 for [11C]CURB in PET with 2TCMi analysis was under 0.06 minutes−1 in the human brain (Rusjan et al., 2013). 6-Hydroxy-[1,1'-biphenyl]-3-yl cyclohexylcarbamate is a URB597 derivative and has lower affinity (IC50 = 30 nM) for FAAH than URB597 (IC50 = 7.7 nM) (Clapper et al., 2009). N-(3,4-Dimethylisoxazol-5-yl)-4-(4-(2-fluoro-4-methylphenyl)thiazol-2-yl)piperazine-1-carboxamide has been recently developed as an inhibitor possessing a 3-fold higher affinity for FAAH than URB597 (Shimoda et al., 2016). Although the study subjects were different in this study, FAAH between the human and the rat is known to have high biologic homology and similar Vmax values in brains (Desarnaud et al., 1995; Giang and Cravatt, 1997; Maccarrone et al., 1998). Therefore, an estimated k3 with over 0.1 minutes−1 in every FAAH-rich region of rat brain in the compartmental analysis of [11C]DFMC (Table 1) would reflect a high affinity for FAAH, as described in eq. 3. Thus, [11C]DFMC showing a much k3 value would be a reasonable radiotracer for PET assessments.
To promote applications using PET with [11C]DFMC, we tried to noninvasively estimate the macroparameter Ki value, including k3, using the reference tissue method. Prior to kinetic analysis, a blocking PET study using URB597 (3 mg/kg) was conducted to determine the reference region. The heterogeneous uptake of radioactivity in all brain regions of control subjects showed significant displaceability by pretreatment with URB597, whereas radioactive uptake in the spinal cord exhibited nondisplaceable uptake (Fig. 5). In a pathologic report, FAAH-containing neurons were detected in the cerebral cortex, hippocampus, and Purkinje cells of the cerebellar cortex but not in the spinal cord (Tsou et al., 1998). However, to our knowledge, there are no reports regarding quantitative PET analysis using the spinal cord as a reference region. Therefore, we performed a test-retest PET study to validate the use of the spinal cord as a reference region. The ICC of the KREF value in FAAH-rich regions was 0.681–0.836 (Table 2), which supported relatively high reliability for KREF values in the PET study using [11C]DFMC with PGAREF. Meanwhile, a high variability (>40%) was detected in the low-FAAH regions (KREF < 0.005 minutes−1), such as the hypothalamus and pons. The KREF value was concisely estimated with high reproducibility in FAAH-rich regions.
Another concern is the differences in distribution volumes (K1/k2) between the region of interest and the reference region. Theoretically, the KREF value equals k2k3/(k2+k3) in the case of K1/k2 = K1′/k2′ (see eq. 5). However, in this study, the distribution volume in the spinal cord (K1′/k2′) did not equal K1/k2 in other brain regions (see in Supplemental Table 2). Therefore, the KREF values are modified to K1k3/(k2+k3)(k2′/K1′). In the results of the PGAREF analysis, the KREF values in brain regions were estimated with a range of 0.004–0.011 minutes−1. Compared with the Ki2TCMi values, the KREF values showed high correlation (r > 0.95; Fig. 5C). In this regression, the slope showed 0.075 (= k2’/K1’), which is defined as constant A in eq. 8. However, a small negative bias (−0.004) of the KREF value was recognized, although PGAREF and Ki2TCMi are proportional in theory. This small negative bias may be caused by the slight accumulation of the radiotracer at nondisplaceable sites in the spinal cord, which would cause an undesirable increase in the value of TAC in the spinal cord. Nevertheless, the spinal cord would be an adequate reference region for the estimation of regional KREF values in the present PGAREF analysis. Moreover, estimated KREF values would be a reasonable index for alternative net uptake values of [11C]DFMC with FAAH despite including a small bias.
Finally, to propose a pharmacological application of [11C]DFMC-PET using KREF values, we estimated ED50 values of URB597 in the brain in vivo. URB597 has been developed a decade before (Fegley et al., 2005), and widely tested as a treatment of neuroinflammation and pain (Murphy et al., 2012; Lomazzo et al., 2015). Here, since the KREF value is not proportional to the regional FAAH concentrations in theory (eqs. 3 and 6), we tentatively estimated the k3 value (defined as calculated k3) from KREF value (eq. 9). In a comparison of the directly estimated k3 obtained by the compartmental analysis with blood sampling, the calculated k3 showed high correlation (r = 0.946), although a slight overestimation was exhibited (Fig. 5D). This result suggested that the calculated k3 would be favorably used as an alternative parameter of the directly estimated k3 value, which motivated us to progress our pharmacological application study. However, there is a considerable limitation in the use of calculated k3. Since the equation for calculated k3 includes two variable parameters, K1 and k2, it is important that administration of URB597 does not affect blood flow. Fortunately, administration of under 1 mg/kg (the maximum dose used in this assessment) of URB597 did not produce any effects on the initial uptake of radioactivity (Supplemental Fig. 1). Therefore, the present dose-response assay using the calculated k3 would be expected to give reasonable ED50 values of URB597 in brain regions without effects on blood flow.
Although the assay used a calculated-k3, the ED50 values of URB597 were estimated to be 66.4 µg/kg (0.2 µmol/kg ≈ 60 nmol/head) in the rat brain (Fig. 6B). Previously, we reported that the in vitro IC50 value of URB597 using a membrane fraction of rat brain was 19.6 ± 3.5 nM (Shimoda et al., 2016). It was previously reported that an 11C-labeled URB597 analog remained at 0.6 percentage of injection dose per gram tissue (%ID/g) in the brain of mice 60 minutes after injection (Wyffels et al., 2010). In general, the %ID/g value of rat brain is roughly 10-fold lower than that of the mouse brain because of the differences in body weights. Therefore, the present ED50 value of URB597 was supposed to be approximately 36 pmol/g brain ([injection dose of URB597 (60 nmol/head)] × 0.06%ID/g brain) in PET with [11C]DFMC. The present in vivo ED50 value (36 nmol/kg brain) could be concisely estimated since it was close to the in vitro measurement (19.6 nmol/l).
In summary, we demonstrated the noninvasive estimation of the KREF value of [11C]DFMC in FAAH-containing brain regions using the reference tissue model. Moreover, by using KREF value, we proposed calculated k3 as an alternative index for the quantification of FAAH. Finally, using calculated k3, we estimated ED50 value of URB597 responsible for FAAH inhibition in rat brain as one of the pharmacological applications in vivo. Thus, our technique using PET with [11C]DFMC could contribute to occupancy studies using rats for new pharmaceuticals for the treatment of central nervous system disorders targeting FAAH.
Acknowledgments
We thank the staff of the National Institute of Radiological Sciences for their support with the following: cyclotron operation, radioisotope production, radiosynthesis, and animal experiments. We thank Trent Rogers from Edanz Group (www.edanzediting.com/ac) for editing a draft of this manuscript.
Authorship Contributions
Participated in research design: Yamasaki T, Ohya T.
Conducted experiments: Yamasaki T, Zhang Y, Wakizaka H.
Contributed new reagents or analytic tools: Mori W, Nengaki N, Fujinaga M.
Performed data analysis: Yamasaki T.
Wrote or contributed to the writing of the manuscript: Yamasaki T, Ohya T, Kikuchi T, Zhang MR.
Footnotes
- Received November 6, 2019.
- Accepted March 27, 2020.
Declaration of conflicting interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
↵
This article has supplemental material available at jpet.aspetjournals.org.
Abbreviations
- BPND
- nondisplaceable binding potential
- [11C]CURB
- [11C-carbonyl]-6-hydroxy-(1,1’-biphenyl)-3-yl cyclohexylcarbamate
- [11C]DFMC
- N-(3,4-dimethylisoxazol-5-yl)piperazine-4-[4-(2-fluoro-4-[11C]methylphenyl)thiazol-2-yl]-1-carboxamide
- %COV
- percentage of coefficients of variationa
- CP
- plasma compartment
- FAAH
- fatty acid amide hydrolase
- ICC
- intraclass correlation coefficient
- %ID
- percent injected dose
- MRI
- magnetic resonance imaging
- PGA
- Patlak graphical analysis
- PGAREF
- PGA reference model
- PET
- positron emission tomography
- SUV
- standardized uptake value
- TAC
- time-activity curve
- 1TCM
- one-tissue compartment model
- 2TCMi
- two-tissue compartment model for the irreversible-type radiotracer
- URB597
- [3-(3-carbamoylphenyl)phenyl] N-cyclohexylcarbamate
- Copyright © 2020 by The American Society for Pharmacology and Experimental Therapeutics
