|Home | About | Journals | Submit | Contact Us | Français|
Radioligand binding studies indicate a down-regulation of myocardial β1-adrenoceptors (β1-AR) in cardiac disease which may or may not be associated with a decrease in β2-ARs. We have chosen ICI 89,406, a β1-selective AR antagonist, as the lead structure to develop new β1-AR radioligands for PET and have synthesised a fluoro-ethoxy derivative (F-ICI).
(S)-N-[2-[3-(2-Cyano-phenoxy)-2-hydroxy-propylamino]-ethyl]-N′-[4-(2-[18F]fluoro-ethoxy)-phenyl]-urea ((S)-[18F]F-ICI) was synthesised. Myocardial uptake of radioactivity after intravenous injection of (S)-[18F]F-ICI into adult CD1 mice or Wistar rats was assessed with positron emission tomography (PET) and postmortem dissection. Metabolism was assessed by high-performance liquid chromatography analysis of plasma and urine.
The heart was visualised with PET after injection of (S)-[18F]F-ICI but neither unlabelled F-ICI nor propranolol (non-selective β-AR antagonist) injected 15 min after (S)-[18F]F-ICI affected myocardial radioactivity. Ex vivo dissection demonstrated that predosing with propranolol or CGP 20712 (β1-selective AR-antagonist) did not affect myocardial radioactivity. Radiometabolites rapidly appeared in plasma and both (S)-[18F]F-ICI and radiometabolites accumulated in urine.
Myocardial uptake of (S)-[18F]F-ICI after intravenous injection was mainly at sites unrelated to β1-ARs. (S)-[18F]F-ICI is not a suitable β1-selective-AR radioligand for PET.
It is generally accepted that the sympathetic nervous system plays a role in the development of heart failure. Overactivation of the cardiac sympathetic nerves causes an increased release of noradrenaline from the heart to plasma [1–3]. Myocardial β-adrenoceptor (β-AR) density is reduced and downstream mechanisms altered. Clinical trials have shown that β-AR antagonists, such as metropolol, carvedilol and bisoprolol, reduce morbidity and mortality of human heart failure [1,3].
Many studies of myocardial membranes obtained at autopsy or surgery have demonstrated a preferential down regulation of β1-AR [4–7]. Decreases in both β1- and β2-ARs, however, have been reported [5–8], and it is not clear whether or not there is a clinical advantage of β1-AR antagonist therapy [1,3]. In addition genetic polymorphisms of β1- and β2-AR may be risk factors which influence the progression of heart failure . A better understanding of the roles of each sub-type may help to improve therapy.
Positron emission tomography (PET) with the non-selective β-AR radioligand (S)-[11C]CGP 12177 has been used to demonstrate down-regulation of myocardial β-ARs in vivo in patients with some types of heart disease, including hypertrophic cardiomyopathy, idiopathic right ventricular outflow tract tachycardia and arrhythmogenic right ventricular cardiomyopathy [10–12] but not Brugada syndrome , and a newer more easily prepared radioligand [11C]CGP 12388 is now available [14,15]. Although several groups have been searching for a β1-selective-AR radioligand for PET, none has succeeded .
We have chosen ICI 89,406 (Fig. 1A), a β1-AR-selective antagonist [17–19], as our lead compound for developing β1-AR-selective ligands for molecular imaging. Twenty derivatives were prepared by modifying the aromatic substituents and lengths of the aliphatic chains. Of these, 16 derivatives were selective for β1-ARs (high affinity inhibition constant for β1-ARs Ki1, 0.02–35 nM; low affinity inhibition constant for β2-ARs Ki2, 12–16,000 nM; β1-selectivity Ki1/Ki2, 35–16,800) . One derivative, I-ICI-H , iodinated at ring A (Fig. 1A) by exchanging the cyano moiety for iodine, showed especially high affinity and selectivity in vitro but the radioactive compound [123/125I]I-ICI-H was not suitable for in vivo studies due to high nonspecific binding . A more hydrophilic iodinated derivative, I-ICI-COOH, was synthesised, but although this compound again showed high affinity and selectivity in vitro, it was rapidly de-iodinated in vivo . A third derivative, (S)-[11C]ICI-OMe, with a [11C] methoxy moiety at C4 of ring B (Fig. 1A) was taken up by the heart in mice and rats but uptake was not displaced or blocked by established β-AR antagonists, indicating that uptake was not specific to β-ARs. In addition, there was metabolism to [11C]CO2 .
Other radiolabelled analogues of ICI 89,406 may be less susceptible to metabolism. An 18F-labelled compound would have the advantage of a longer radioactive half life than 11C so that biochemical pathways could be followed for longer times. In addition, the spacial resolution for 18F-labelled ligands is superior due to the shorter positron range for 18F compared to 11C. Substitution of the CN moiety in ring A (Fig. 1) by fluorine is theoretically feasible. However, by analogy with I-ICI-H, the resulting compound may show high non-specific binding. Introduction of [18F]fluorine in ring B (Fig. 1B) was an alternative option. Although the [11C]-OMe derivative was metabolised to 11CO2, presumably via [11C]methanol and the Krebs cycle , the fluorinated derivative may be metabolised less rapidly by an alternative pathway enabling sufficient radiolabelled ligand to be available for β1-AR binding. We have, therefore, synthesised the fluorinated analogues (R)- and (S)-N-[2-[3-(2-cyano-phenoxy)-2-hydroxy-propylamino]-ethyl]-N′-[4-(2-fluoro-ethoxy)-phenyl]-urea (Fig. 1B) . In vitro studies showed that the (S)-enantiomer had higher β1-AR selectivity and affinity than the (R)-enantiomer (affinity Ki1=0.049 nM vs Ki1=0.297 nM; selectivity 40,800 vs. 1580) . The affinity and selectivity of the (S)-enantiomer were also higher than those for (S)-ICI-OMe (Ki1=0.07, selectivity=1240, respectively) . The 18F-labelled (S)-analogue [(S)-[18F]F-ICI] was synthesised at a higher specific activity [40 GBq·µmol−1 at end of synthesis (EOS)]  than was achieved for (S)-ICI-OMe (6.5 GBq·µmol−1 at EOS) .
The present article describes experiments to evaluate the suitability of (S)-[18F]F-ICI for PET in mice and rats using small animal PET, e dissection and high-performance liquid chromatography (HPLC) analysis of radioactive metabolites.
(S)-N-[2-[3-(2-Cyano-phenoxy)-2-hydroxy-propylamino]-ethyl]-N′-[4-(2-[18F]fluoro-ethoxy)-phenyl]-urea ((S)-[18F]F-ICI) was prepared by a two-step synthesis from ethylene glycol ditosylate, [18F]fluoride ion and the phenol precursor, (S)-N-[2-[3-(2-cyano-phenoxy)-2-hydroxy-propylamino]-ethyl]-N′-(4-hydroxy-phenyl)-urea, as described elsewhere .
Unlabelled (S)-F-ICI was prepared as described previously . It was dissolved in ethanol/dimethylsulphoxide (DMSO)/water (25/1/50, v/v/v) at 2 µmol·ml−1 for injection into animals.
The non-selective β-AR antagonists, propranolol HCl and CGP 12177 HCl, and the β1-AR selective antagonist, CGP 20712A, were purchased from Sigma Aldrich Chemie (Germany). These antagonists were dissolved in saline at concentrations of 5 µmol·ml−1 for injection.
Studies were approved by local and federal ethics committees and were performed in accordance with institutional guidelines for health and care of experimental animals.
Male CD1 mice (30–40 g) or Wistar rats (270–430 g) which have had free access to food were anaesthetised by inhalation (isoflurane 2%, oxygen 0.5 l·min−1) for insertion of indwelling intravenous catheters. For PET scanning, each mouse or rat had a catheter in one lateral tail vein. For ex vivo biodistribution studies each rat had two catheters, one in the ventral tail artery and one in a lateral tail vein. Animals were allowed to recover for 1–2 h under light restraint before being re-anaesthetised for PET scanning. Ex vivo biodistribution and metabolite studies were carried out with conscious animals [21,23,26].
PET was carried out using a sub-millimeter high resolution (0.7-mm full width at half maximum) dedicated small animal scanner (32 module quadHIDAC, Oxford Positron Systems, UK). This scanner uses wire-chamber detectors and offers uniform spatial resolution (<1 mm) over a large cylindrical field (165-mm diameter, 280-mm axial length) [27,28].
Two anaesthetised mice or rats (isoflurane 2%, oxygen 0.2 l·min−1 per mouse or 0.5 l·min−1 per rat) were placed on a heating pad to maintain body temperature during the scan. Animals were positioned on the scanner and 30 s after the start of data acquisition (S)-[18F]F-ICI (1–3 MBq in 100µl for mice or 4.8–8.7 MBq in 200 µl for rats) was injected simultaneously into each animal in the scanner via the tail vein catheters. For each scan, one animal was given unlabelled (S)-F-ICI (2 µmol·kg−1) or the non-selective β-AR blocker propranolol (5 µmol·kg−1) 15 min after injection of (S)-[18F]F-ICI. To confirm the location of the heart, 2-[18F]fluoro-2-deoxy-d-glucose ([18F]FDG) (3–7.5 MBq in 50 µl for mice or 9 MBq in 200 µl for rats) was injected via the tail vein 5–10 min after completion of the (S)-[18F]F-ICI scan and data were acquired for 30 min. To improve myocardial uptake of [18F]FDG in rats, 0.5 ml glucose (20%) was injected 5 min before [18F]FDG.
List-mode data were acquired for 60 min [(S)-[18F]F-ICI] or 30 min ([18F]FDG) and subsequently reconstructed using an iterative reconstruction algorithm . PET images were analysed using in-house software programmes in MATLAB (The Math Works Company) and C programming languages . Details of the analysis are published elsewhere . In brief, the total radioactivity in each animal was assessed by drawing a cube encompassing the body, excluding the tail and paws, on the reconstructed [18F]FDG image (15–30 min time frame). The parameters defining this cube were saved and used to compute the whole body radioactivities (cps mouse or rat) for each frame of the (S)-[18F]F-ICI scan. The parameters required to create images of each heart and to compute time-activity curves for (S)-[18F]F-ICI were also defined using the [18F]FDG scan. The average radioactivity in a given volume of interest (cps·ml−1) was calculated for each time frame and expressed as a percentage of the whole body activity defined as:
Ex vivo biodistribution studies were carried out in rats (as described previously [21,23,25]. (S)-[18F]F-ICI or unlabelled antagonists were injected as a bolus (1 µl·g−1 body weight) via the tail vein. Aliquots of each (S)-[18F]F-ICI injectate were diluted in ethanol/saline and measured to determine the radioactivity injected into each rat. Six sequential arterial blood samples (ca. 100 µl) were taken from some rats. Rats were sacrificed by intravenous injection of sodium pentobarbitone (Euthatal) at 200 mg·(kg body weight)−1 at predetermined times after injection and tissues rapidly removed. Tissue radioactivity was measured by an automated gamma counter (Wallac Wizard 3, Perkin Elmer Life Sciences, Boston, MA, USA). Radioactivity was expressed as an uptake index, defined as:
Radiometabolites in plasma and urine were assessed by gradient RP-HPLC . Rats (280–350 g) were injected with 20–110 MBq (140–300 nmol·kg−1) (S)-[18F]F-ICI via the tail vein catheters. Up to three blood samples (1 ml) were collected from the tail artery catheter at 2–25 min after injection of radioligand. Cell-free plasma was prepared by centrifuging each blood sample at 2000 g for 2 min. Ice-cold acetonitrile (0.7 ml) was added to the plasma (0.5 ml) and the precipitated proteins were removed by centrifugation. A sample of the resulting supernatant (200 µl) was spiked with non-radioactive (S)-F-ICI (10µg) for HPLC analysis using a gradient RP-HPLC system composed of two Knauer K-1800 pumps, a Knauer S-2500 UV detector, a Raytest Gabi Star γ-detector and a RP-HPLC Nucleosil Eurosphere 100-10 C-18 column (250×8 mm). The following conditions were applied: eluent A: water (0.1% TFA), eluent B: acetonitrile (0.1% TFA), gradient from 85% A to 30% A in 18 min, halt at 30% A for 5 min, gradient from 30% A to 85% A in 1 min, flow: 5.5 ml·min−1. The eluate was monitored for radioactivity and UV absorbance (λ=254 nm).
The specific activities of the [18F]F-ICI used in this study, ranged from 0.8 to 16.9 GBq·µmol−1 (mean 4.7±1.8 GBq·µmol−1, n=10) at the end of synthesis. These are lower than the value previously published (40±8 GBq/µmol)  because the activities of [18F]fluoride at the start of radiosynthesis were lower than those used in the earlier paper. In our experience, the amount of carrier [19F]fluoride that is introduced in our specific synthesis procedure, using our specific synthesiser, O-18-water and chemicals, is constant and independent of the [18F]fluoride radioactivity at the start of synthesis. Consequently, the specific activity should be lower in this paper. Furthermore, the preparations were performed in a synthesiser with old TEFZEL (a fluoropolymer) tubes, which potentially increase the fluoride carrier and decrease the specific activity.
Fig. 2 illustrates the biodistribution of radioactivity after intravenous injection of (S)-[18F]F-ICI into mice. The hearts can be seen but there was high uptake of radioactivity in abdominal regions (liver, intestine, kidney, and bladder). Fig. 2 illustrates radioactivity in coronal planes encompassing the heart chambers but excluding most of the brain. Examination of more dorsal planes indicted that there was no appreciable uptake in brain. The heart was barely visible after the injection of vehicle or unlabelled (S)-F-ICI (Fig. 2B). The positions of the hearts were confirmed by the [18F]FDG scan carried out after completion of the (S)-[18F]F-ICI scan (Fig. 2C).
Fig. 3 shows coronal images of the thoraces of the mice. These images have been rotated in the sagittal plane so that the hearts are shown as in clinical studies, i.e. as if the mice were on their backs with their hearts on the right (Fig. 3A and D). As in Fig. 2, unlabelled (S)-F-ICI injected 20 min after (S)-[18F]F-ICI had no obvious effect on the later images (Fig. 3B and E). One-minute time frames were reconstructed for the (S)-[18F]F-ICI scan and time-activity curves were computed for regions of interest drawn round the left ventricle wall and septum. Myocardial radioactivity was detected immediately after injection and reached a maximum during the first minute. The early phase (<5min) of rapid loss from the left ventricle wall and septum was followed by a slow loss of myocardial radioactivity. Injection of unlabelled (S)-F-ICI had no significant effect on this loss.
Fig. 4 illustrates the biodistribution of radioactivity after intravenous injection of (S)-[18F]F-ICI into rats. As for mice, the hearts can be seen (Fig. 4A), and there was high uptake of radioactivity in liver, intestine, kidney, and bladder but little in brain. In Fig. 4, saline or the non-selective β-AR antagonist propranolol was injected at 20 min and in the rats neither heart was visible at late times (Fig. 4B). The positions of both hearts, however, were confirmed by the [18F]FDG scan carried out after completion of the (S)-[18F]F-ICI scan (Fig. 4C).
Fig. 5 shows coronal images of the thoraces of the rats shown in Fig. 4. Again, the hearts are shown as in clinical studies, i.e., as if the rats were on their backs with their hearts on the right (Fig. 5A and D). Propranolol had no visible effect (Fig. 5B and E). Time-activity curves were similar to those for mice. Myocardial radioactivity cleared rapidly during the first 5 min to a low level and injection of propranolol at 20 min had no significant effect on the slow phase of loss.
The nanomolar doses of (S)-F-ICI injected for PET scanning ranged from 3–10 (data not shown) nmol·kg−1. Unlabelled (S)-F-ICI in the dispensed material competes with radioactively labelled (S)-[18F]F-ICI for any binding sites and these doses may preclude the detection of specific binding by PET. Lower amounts of radioactivity can be used for ex vivo studies which use well-counter gamma detection techniques.
Clearance of radioactivity from plasma after intravenous injection of (S)-[18F]F-ICI is shown in Fig. 6. There was a rapid loss of plasma radioactivity during the first 5 min after radiotracer injection, followed by a slow clearance. Comparable results were obtained for injected doses ranging from 0.5 to 8 nmol·kg−1. Data for all doses, therefore, were combined and a single biexponential fit is shown in Fig. 6.
Fig. 7 illustrates uptake of radioactivity in various tissues as a function of time after injection of (S)-[18F]F-ICI at doses ranging from 0.5 to 8 nmol·kg−1. In all tissues, maximum radioactivity was observed at the first dissection time (2 min). The very high uptake observed for kidney at 2 min (UI 25, equivalent to 8.4 cpm·g−1 % injected cpm for a 300 g rat) decreased exponentially (at 30 min UI=1.9, cpm·g−1 % injected cpm=0.64), consistent with rapid clearance by the kidneys. After the initial peaks, radioactivity cleared more slowly from other tissues. For example, at 30 min after injection, uptake indices for liver and heart were 3.9 and 1.7 (1.3 and 0.56 cpm·g−1 % injected cpm), respectively, compared to 0.53 (0.18, % injected·g−1) for plasma. Similar results were obtained for low (0.5–1 nmol·kg−1) and high (8 nmol·kg−1) nanomolar doses. Radioactivity was detected in the urine at all times after injection of tracer (data not shown).
Two experiments (12 rats) were carried out to compare the effects of predosing with the β1-AR selective antagonist CGP 20712A to those of the non-selective antagonist propranolol. Fig. 8 shows the results with data from the time activity study (Fig. 7, 30 min, 0.5 and 1.0 nmol·kg−1) used for the case designated as (S)-[18F]F-ICI given alone. Uptake indices for heart and lung were significantly higher in rats which received vehicle before (S)-[18F]F-ICI than in those without pretreatment (t test, df=4, P<.05). When compared to vehicle controls, there was a small but significant reduction in radioactivity in myocardium and lung after predosing with CGP 20712A or propranolol (t test, df=5, P<.02) but not in the other tissues studied. When compared to (S)-[18F]F-ICI alone, however, the reduction was only significant in lung after CGP 20712 A (t test, df=5; P<.02).
The results shown in Figs. 7 and and88 suggest that myocardial uptake of (S)-[18F]F-ICI radioactivity does not depend on injected dose (>0.5 nmol·kg−1). In vitro competition studies, however, indicate a very high affinity of (S)-F-ICI for β1-ARs which may compromise detection of specific binding by PET. The effect of injected dose, therefore, was investigated by injecting serial dilutions of dispensed (S)-[18F]F-ICI and assessing tissue radioactivity at 30 min after injection. There was a linear relationship between myocardial uptake, expressed as cpm·(g wet tissue)−1, and injected radioactivity [cpm·(g body weight)−1] which was independent of the specific activity of the dispensed (S)-[18F]F-ICI (0.8 or 2.5 GBq·µmol−1). In other words, the uptake index was not dependent on injected dose between 0.04 and 200 nmol·kg−1 (Fig. 9). Similar results were obtained for other tissues (e.g., liver and plasma in Fig. 9).
Radioligands which have proved useful for studying myocardial neuroreceptors by PET show a decrease in myocardial uptake index as the dose of injected ligand is increased [21,26,30,31]. In vivo dose effect curves for the non-selective β-AR radioligand (S)-[3H]CGP 12177 have been published for rat myocardium, giving a “Bmax” of 4.7 pmol·(g myocardium)−1 and a “KD” of 0.58 nmol·(kg body weight)−1 . Using these values to calculate myocardial uptake of CGP 12177 as a function of injected dose gives the dotted line shown in Fig. 9. For CGP 12177 “non specific” uptake (injected dose >10 nmol·kg−1) was ~1 and as injected dose was decreased the uptake index increased to >5 at <0.5 nmol·kg−1. Uptake indices for (S)-[18F]F-ICI were 2 at comparable doses, consistent with nonspecific binding of radioligand.
Fig. 10 shows HPLC analyses of samples of plasma and urine taken after intravenous injection of (S)-[18F]F-ICI into three rats studied on different days. (S)-[18F]F-ICI eluted at 11 min. (S)-[18F]F-ICI and metabolites at ~2 min and possibly 9 min were detected in plasma at early times (2.5 and 5 min after injection) but only polar metabolites were seen later (20 and 60 min after injection). On the third day, the radiochemical purity of the synthesised (S)-[18F]F-ICI was poor (88%) compared with the usual value (>95%) and the HPLC trace for the injectate given to Rat 3 is shown in Fig. 10. The impurity (~12 min) was not detectable in the plasma (2.5, 5 and 20 min) but was seen in the urine (20 min) together with parent (S)-[18F]F-ICI and polar metabolites. Both (S)-[18F]F-ICI and polar metabolites were detected in urine at 60 min after injection.
The heart was clearly visible in quadHIDAC PET scans of both mice (Fig. 2) and rats (Fig. 4) following intravenous injection of (S)-[18F]F-ICI. The apparent loss of myocardial radioactivity observed during 0.5–5 min (Figs. 3 and and5),5), is due to a combination of decreases in radioactivity in the myocardial microcirculation and of reduced spillover as radioactivity in the heart chambers decreases. After vascular mixing and the initial rapid phase of loss, a higher proportion of the injected (S)-[18F]F-ICI radioactivity was observed in the myocardium of mice (2–3 cps·ml−1 % mouse, Fig. 3) than of rats (0.1–0.3 cps·ml−1 % rat, Fig. 5). Myocardial uptake of radioactivity, however, was low compared to the uptake of radioactivity in liver, kidney and bladder. There was a low level of radioactivity in the brain in both mice and rats, an observation which is consistent with the low lipophilicity of F-ICI (logP=1.44, log D=0.17) , similar to that of the parent compound ICI 89,406 (logP=1.57, logD=0.13) . Injection of unlabelled (S)-F-ICI (2 µmol·kg−1) (Fig. 3) or propranolol (5 µmol·kg−1) (Fig. 5) at 20 min had no effect on the rate of loss of radioactivity from the myocardium.
To demonstrate specific binding of a radioligand to a receptor by predosing or displacement using unlabelled ligands, the amount of radioligand used should give low levels of receptor occupancy (<1%) . It may not be possible to achieve the specific activity required to allow the injection of sufficient radioactivity to acquire statistically significant PET data while maintaining low receptor occupancy in small animals. If the radioactivity required for PET is equivalent to a ligand dose which saturates the receptor of interest, it will not be possible to demonstrate specific receptor binding using PET.
The doses of (S)-[18F]F-ICI used in the PET studies were 6 nmol·kg−1 for mice and 3 nmol·kg−1 for rats. For the non-selective β-AR ligand (S)-[3H]CGP 12177 injected intravenously into rats, a dose of 3 nmol·kg−1 reduced myocardial “specific” uptake of radioactivity to 20% of that observed for 1 nmol and 15 nmol·kg−1 completely blocked it . Using current 18F-labelling methods, the dispensed solution of (S)-[18F]F-ICI includes unlabelled (S)-F-ICI. The 3–6 nmol·kg−1 used in the PET scan may give an in vivo occupancy by unlabelled (S)-F-ICI which is too high for the detection of specific binding of (S)-[18F]F-ICI to β1-AR by predosing or displacement experiments.
Lower doses of (S)-[18F]F-ICI (<1 nmol·kg−1) could be used in the ex vivo dissection studies. These studies showed that after vascular mixing (0–30 s), plasma radioactivity decreased to ~10% during the first 5 min after injection of (S)-[18F]F-ICI (Fig. 6). In all tissues studied, maximum radioactivity was observed at the first dissection time (2 min). After the initial peaks, radioactivity cleared more slowly from tissues than from plasma, consistent with a small uptake in extravascular compartments. Radioactivity in the kidney was especially high immediately after injection; subsequently, there was a rapid exponential loss consistent with high kidney extraction of the tracer. Similar results were obtained for low (0.5–1 nmol·kg−1) and high (8 nmol·kg−1) nanomolar doses suggesting that either there was no specific uptake or that the in vivo “Ki” was 0.5 nmol·kg−1.
The ex vivo biodistribution studies, however, failed to demonstrate high specific binding using low doses of (S)-[18F]F-ICI (0.5–1.0 nmol·kg−1). Although predosing with the selective β1-AR antagonist CGP 20712A or the non-selective antagonist propranolol at high dose (5 µmol·kg−1) had small effects on myocardial and lung radioactivity (Fig. 8), whether or not these differences are physiologically relevant is uncertain. When compared to vehicle controls, there was a small but statistically significant reduction in radioactivity in myocardium and lung after predosing with CGP 20712A or propranolol (t test, df=5, P<.02) but when compared to (S)-[18F]F-ICI alone, the reduction was only significant in lung after CGP 20712 A (t test, df=5, P<.02). Considering the different β1 to β2 ratios in myocardium (β1>β2) and lung (β2>β1) , CGP 20712A might be expected to have the greater effect in myocardium. Moreover, radioactivity in plasma was not effected by the nanomolar dose of (S)-F-ICI (Fig. 6) or by predosing (Fig. 8), whereas increases are observed after β-AR blockade for β-AR radioligands such as CGP12177  and fluorocarazol .
Injected doses >0.5 nmol·kg−1 may include sufficient unlabelled (S)-F-ICI to reduce the proportion of (S)-[18F]F-ICI bound to β1-AR to undetectable levels. Reducing injected (S)-[18F]F-ICI to 0.04 nmol·kg−1, however, did not increase the percentage of the injected radioactivity taken up by the myocardium. There was a linear relationship between myocardial uptake, expressed as cpm·(g wet tissue)−1, and injected radioactivity expressed as cpm·(g body weight)−1. Consequently, the uptake index was not dependent on injected dose between 0.04 and 200 nmol·kg−1 (Fig. 9). Similar results were obtained for other tissues (eg, liver and plasma) in which specific binding to β1-AR was not expected. A radioligand specific for β-AR, for example. (S)-CGP 12177, with an ex vivo Ki of 0.58 nmol·(kg body weight)−1 , would give a sigmoidal dose effect curve (Fig. 9). For (S)-[3H]CGP-12177 “non-specific” uptake (injected dose >10 nmol·kg−1) was ~1, and as injected nanomolar dose was decreased, the uptake index increased to >5 at <0.5 nmol·kg−1. Comparable dose effect curves have been reported for the β2-AR selective radioligand (S)-1′-[18F]fluorocarazol . Uptake indices for doses >10 nmol·kg−1 were ~0.5 whereas that for 0.5 nmol·kg−1 was ~2.2. For a subtype selective ligand, the uptake index at low doses would depend on the relative numbers of each subtype. For rat myocardium, 50–80% of β-ARs are of the β1-subtype [31–37] from which it might be expected that the uptake index at 0.5 nmol·kg−1 would be 3–4 for the β1-subtype and 1–2 for the β2-subtype. Uptake indices for (S)-[18F]F-ICI were 2 for a wide range of doses (Fig. 9), consistent with no specific binding to β-ARs.
For the non-selective β-AR ligand (S)-[11C]CGP 12177 in rats, a dose of 0.58 nmol·kg−1 halved myocardial uptake . In vitro membrane binding studies show that the KD for CGP 12177 (0.19–0.67 nM [34–38]) is 4–14 times greater than the Ki for (S)-F-ICI (0.049 nM ). If metabolism were similar, radioligands with higher affinity to β1-AR would be expected to give sigmoid dose effect curves shifted to lower doses (in vivo Ki0.5 mol·(kg body weight)−1 with higher uptake indices at low doses. Considered together, Figs. 3, ,5,5, and and88–10 indicate that there is no detectable specific binding of (S)-[18F]F-ICI to β1-ARs.
The HPLC studies showed rapid breakdown of (S)-[18F]F-ICI (Fig. 10). (S)-[18F]F-ICI and polar metabolites were detected in plasma at early times (2.5 and 5 min) but only polar metabolites were seen later (20 and 60 min). Both (S)-[18F]F-ICI and polar metabolites were detected at 20 and 60 min in urine. Rapid metabolism and excretion of (S)-[18F]F-ICI would result in a low availability of radioligand to myocytes and consequently a very low occupancy of myocardial β1-AR which precludes the detection of specific binding by dissection techniques. An alternative explanation for the failure to demonstrate an effect of predosing with unlabelled antagonists could be that labelled metabolites rather than parent compound were present in myocardium. Studies to assess radioactive metabolites in myocardium were not carried out as myocardial radioactivity at later times after injection, when radioactivity had cleared form the blood, was below detection by HPLC. The nonspecific binding observed in the blocking studies, however, is greater than that observed for the hydrophilic non-subtype-selective ligand CGP 12177, suggesting that (S)-F-ICI has a higher non-specific binding to myocardial membranes.
The biodistribution, low myocardial uptake and failure to demonstrate specific binding of [18F]F-ICI to myocardial β1-ARs are comparable to the results for [11C]ICI-OMe. The HPLC profiles of radioactive metabolites, however, differed. Two polar radioactive metabolites were observed in plasma after injection of [11C]ICI-OMe , whereas only one was observed after [18F]F-ICI. 11CO2 was observed in exhaled air after [11C]ICI-OMe . Parent compound was detected in urine after injection of either derivative but radioactive metabolites in urine were only detected after [18F]F-ICI. Unfortunately, both derivatives were rapidly metabolised and excreted.
Other novel compounds have been considered for radiolabelling. The β-AR antagonist, (R)-2-(2-hydroxy-3-(isopropylamino)propoxy)benzonitrile  is amenable to labelling with fluorine-18. We have synthesised this compound and have developed a fluorination method, based on a published procedure , in which the hydroxyl group is replaced by [18F]fluoride. Unfortunately, the fluorinated derivative is unstable . Derivatives of the lead structure LK204545 , that represents a β1-selective AR ligand, are other candidates for labelling with fluorine-18. Work is underway to synthesize these compounds and to develop labelling methods.
A fluoroethoxy derivative of the β1-adrenoceptor antagonist ICI 89,406 was labelled with fluorine-18 [(S)-[18F]F-ICI]. Although in vitro membrane studies showed that (S)-F-ICI has high affinity and selectivity for β1-ARs, the present study in mice and rats failed to demonstrate high specific binding of (S)-[18F]F-ICI to myocardial β1-AR following intravenous injection. Similar results were previously obtained for three other derivatives of ICI 89,406. New lead compounds for a β1-selective AR radioligand for PET are, therefore, being investigated.
This work was supported by grants from the Deutsche Forschungsgemeinschaft (DFG), Sonderforschungsbereich 656 MoBil Münster, Germany (projects A5 and C6), by the Interdisciplinary Clinical Research Centre, Münster (IZKF, core unit SmAP), Germany, and by the EC - FP6-project DiMI (WP 11.1), LSHB-CT-2005-512146. Dr. V.W. Pike is supported by the Intramural Research Program of the National Institutes of Health (the National Institute of Mental Health), USA.
The authors would like to thank Dr. Klaus Schäfers for advice on HIDAC scanning and software for image analysis and Ms Christine Bätza, Mr Daniel Burkert, Ms Irmgard Hoppe and Ms Sandra Schröer for technical assistance.