![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
|
|
|
|
|
||
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Institut für Nuklearchemie and Brain Imaging Centre West, Forschungszentrum Jülich GmbH, Jülich, Germany
(Received July 4, 2005; accepted January 12, 2006)
 |
Abstract |
|---|
 Top Abstract Materials and MethodsResultsDiscussionReferences |
|---|
) and pharmacological properties of receptors in vivo (Merlet et al., 1993; Beugel et al., 1999). However, receptor-PET studies favor radioligands that are either stable in the body or that give rise to metabolites that do not interfere with the imaging of specifically bound ligand.
The ligand [18F]8-cyclopentyl-3-(3-fluoropropyl)-1-propylxanthine ([18F]CPFPX) (Holschbach et al., 2002) is used to image the A1 adenosine receptor (A1AR) in human brain (Bauer et al., 2003). Because this ligand does not undergo degradation in the central nervous system, specifically bound ligand accounts for a very large fraction of brain radioactivity. However, such is not the case in peripheral tissues. The metabolism of [18F]CPFPX in primates (Boy et al., 1998) and humans gives rise to at least three polar metabolites in blood, and studies illustrated the confounding effects of these metabolites. For example, the intravenous administration of [18F]CPFPX to experimental animals caused intense labeling of the heart that was unaffected by the administration of unlabeled ligand, evidence for high unspecific binding of a metabolite (Holschbach et al., 1998).
The physiological importance of the A1AR, its wide tissue distribution, and the success of PET-imaging A1ARs in the central nervous system urge extension of this technique to other organs. The design of more stable radioligands to achieve that end requires the kind of information about the metabolism of [18F]CPFPX provided by this study.
Measurements of receptor density by PET depend on compartmental analysis by mathematical models that are very sensitive to the concentration of native radioligand in blood perfusing the organ (the "input function"). Such measurements on the plasma of human subjects (Meyer et al., 2004) identified several [18F]CPFPX metabolites in addition to unchanged ligand. Because the radiotracers for PET studies are prepared under no-carrier-added conditions, the amount of compound administered is in the low to subnanomolar range, making direct spectrometric identification of metabolites impossible. As this report describes, incubating CPFPX with human liver microsomes generated compounds that by HPLC had the same mobilities as the metabolites in plasma, and LC-MS tentatively identified them by measuring m/z of the [M + H]+ ions.
The literature contains little information about the metabolism of synthetic xanthines. The use of CPFPX in humans for diagnostic and research PET imaging necessitates the knowledge of its metabolism in vivo. To our knowledge, the present study of the biotransformation of CPFPX in humans is the first of its kind.
|
Materials and Methods |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
; Meyer et al., 2004) reviewed and approved by the Ethics Committee of the Medical Faculty of the University of Düsseldorf and the German Federal Office for Radiation Protection. Sterile, pyrogen-free physiological saline (Braun, Melsungen, Germany) was the solvent for solutions administered intravenously. Blood collection from rodents was by cardiac puncture into syringes moistened with heparin. The in-house syntheses of CPFPX (Holschbach et al., 1998b), [3H]CPFPX (Holschbach et al., 2003), and [18F]CPFPX (Holschbach et al., 2002) followed published methods. Solvents were HPLC grade and were used as supplied by Sigma-Aldrich Chemie GmbH (Steinheim, Germany). TLC of blood extracts used glass-backed silica gel sheets (G 25 UV254; Macherey-Nagel, Düren, Germany); scanning of the plates by an InstantImager (Canberra-Packard GmbH, Dreieich, Germany) measured radioactivity. Calibration showed that the response of the device to fluorine-18 was linear over the range 23 to 5900 Bq (0.63–160 nCi). Human liver microsomes pooled from different human donors were obtained from Sigma-Aldrich Chemie GmbH, BD Biosciences (Woburn, MA), XenoTech (Kansas City, KS), and In Vitro Technologies (Baltimore, MD). All biochemicals were obtained from Fluka/Sigma-Aldrich Chemie GmbH (Steinheim, Germany).
Microsomes from three rats and three mice, respectively, were prepared by differential centrifugation using a procedure slightly modified from the method described by Sugiura et al. (1974). Rats were sacrificed by decapitation; livers were removed rapidly and homogenized with 3 volumes of 1.5% (isotonic) KCl solution in a Potter homogenizer on ice. The homogenate was centrifuged (10,000g, 20 min, 4°C). The lipophilic layer on top was carefully removed. The pellet was discarded and the microsomes sedimented by ultracentrifugation (105,000g, 60 min, 4°C). The pellet was washed three times with ice-cold KCl solution and finally suspended in 1 volume of 1.5% KCl. The homogenate was stored in 50-µl portions at –70°C. Protein estimation used a commercial assay (Bio-Rad DC Protein Assay; Bio-Rad, Hercules, CA) after solubilization in 15% NH4OH containing 2% SDS (w/v); human serum albumin served as a standard.
Extraction and Metabolite Profile of [18F]CPFPX in Mice. Preliminary experiments used mice to determine an extraction solvent. Groups of three mice each were bled 1, 2, 5, 10, 15, 30, or 60 min after the tail vein injection of [18F]CPFPX (3700 KBq, 100 µCi) in a volume of 50 µl. The addition of [3H]CPFPX (74 KBq, 2 µCi) to the blood samples served for evaluating the effect of extraction solvents to determine whether residual protein interfered with chromatography (Blanchard, 1981) and for calculation of tracer recovery. Centrifugation (1250g, 3 min, 18°C) separated the plasma. Triplicate aliquots (20 µl each) were mixed with an equal volume of either ethanol, methanol, acetonitrile, methanol/acetonitrile, 50:50 (v/v), or a mixture of methanol/dichloromethane, 80:20 (v/v) and vigorously shaken for 2 min on a vortex mixer. After centrifugation (20,800 g, 2 min, 18°C), triplicate samples (2 µl each) of the supernatant were spotted on each of two TLC sheets. One sheet was developed with ethyl acetate/hexane, 75:25 (v/v). The other sheet, which also contained samples (2 µl each) of whole blood and native plasma, was not developed, and served for measuring total radioactivity in each sample. Scanning both sheets immediately with the InstantImager measured fluorine-18 activity. After 48 h (27 half-lives of fluorine-18, when <10–6% of the original radioactivity of fluorine-18 remained), the sheets were exposed for 5 days to an imaging plate (TR2025; Fuji, Tokyo, Japan), which is sensitive to the low-energy ß–-radiation of tritium, and were scanned with a Fujifilm BAS 5000 system (Fuji) to measure tritium activity. Recovery of radiotracer by the extraction solvent was calculated as the ratio of radioactivity in an aliquot of extract to that in an aliquot of plasma of an equivalent volume, expressed as a percentage.
Metabolite Profile of [18F]CPFPX in Rats. [18F]CPFPX (3700 KBq, 100 µCi in 300 µl) was injected into rats (n = 7) via the tail vein, 1, 2, 5, 10, 15, 30, or 60 min (one rat per time point) before exsanguination. Centrifugation of the blood samples (1250g, 3 min, 18°C) yielded ca. 2 ml of plasma per rat. Aliquots (2 µl) were spotted on TLC sheets as standards for calculation of the recovery of radioactivity from plasma extracts. Aliquots of plasma (100 ml) were then mixed with 100 µl of methanol/acetonitrile, 50:50 (v/v), and a sample (2 µl) of the supernatant after centrifugation (20,800g, 2 min, 18°C) was applied to a TLC sheet for subsequent chromatography. Triplicate samples (2 µl each) were spotted on a second TLC sheet for measurement of radioactivity. The first TLC sheet was developed with ethyl acetate/hexane, 75:25 (v/v). Scanning both sheets immediately with the InstantImager measured fluorine-18 activity.
Pharmacokinetics and Metabolite Profile of [18F]CPFPX in Human Volunteers. The PET studies were as described previously (Bauer et al., 2003; Meyer et al., 2004). The dose of [18F]CPFPX was 273 ± 12 GBq (7.4 ± 0.3 mCi, 0.5–5 nmol) in saline (10 ml), injected i.v. over approximately 20 s. Blood sampling (
4 ml) into heparinized syringes was at every 6 s for 90 s, then at 1, 2, 3, 4, 6, 8, 10, 15, 20, 30, 45, 60, 75, and 90 min after the beginning of tracer injection. Assays of metabolites in plasma were made on samples collected at 1 and 2 min and on each sample collected thereafter. The blood samples were centrifuged (1000g, 3 min, 18°C) and the plasma was collected. Samples of plasma (100 µl each) were extracted with 100 µl of acetonitrile/methanol, 50:50 (v/v). Aliquots (5 µl each) of both plasma and, separately, plasma extract were spotted directly above the TLC lane beyond the zone of TLC development (see Fig. 1A). TLC and analyses were as described above for rodent plasma.
|
Mass Spectrometry. For LC-MS measurements, the outlet of the UV-detector was coupled via an electrospray interface to a mass spectrometer (Surveyor MSQ; Thermo Electron Corporation, San Jose, CA). Nebulizer gas pressure was 4 bar and desolvation temperature was 450°C. Positive ion electrospray ionization detected CPFPX and its metabolites. The sprayer voltage and the cone voltage were 3000 and either 60 or 110 V, respectively. Positive ion spectra were recorded over an m/z range of 150 to 450 at a scan time of 1 s. The Xcalibur software, version 1.3, provided with the instrument permitted scans of the chromatograms for ions of a desired m/z over the range m/z ± 0.5.
Oxidation of [18F]CPFPX by Liver Microsomes. Liver microsomes (0.8 mg of protein) were dispersed in 0.1 M phosphate buffer (pH 7.4) containing 3.3 mM MgCl2 and a NADPH-generating system consisting of 1.3 mM NADP, 3.3 mM glucose 6-phosphate, and 0.4 U of glucose-6-phosphate dehydrogenase in a final volume of 1 ml. Incubation was at 37°C. The addition of [18F]CPFPX, alone or with carrier CPFPX (3 µg, 9 nmol) in dimethyl sulfoxide (1 µl), initiated the reaction, which was stopped after 20 min (rat and mouse microsomes) or 4 h (human microsomes) by the addition of acetonitrile (1 ml). After 2 min of vortex mixing and centrifugation (20,800g, 1 min, 4°C), the supernatant was vacuum evaporated to dryness at ambient temperature. A solution of the residue in HPLC eluent (100 µl) was centrifuged (20,800g, 1 min, 4°C) to remove sediment. Control experiments testing the possibility that the addition of carrier affected the metabolism of [18F]CPFPX were identical except that the reaction mixture contained no carrier CPFPX.
Preparation of Human Blood Samples for HPLC. HPLC analysis of the [18F]CPFPX metabolites in human blood required an 8-ml sample of whole blood. The samples, which contained 30 KBq (0.8 µCi) of fluorine-18 radioactivity, were centrifuged (1000g, 5 min, 18°C) to obtain approximately 3 ml of plasma. After the addition of 2 volumes of acetonitrile, the mixture was shaken (vortex mixer) for 1 min. The supernatant separated from precipitated protein by centrifugation (5000g, 5 min) was evaporated to dryness in vacuo at ambient temperature. A solution of the residue in HPLC eluent (500 µl) was filtered through a 2-µm membrane filter before injection.
|
Results |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
Species Differences in Metabolism of [18F]CPFPX in HPLC Experiments. Figure 2 compares HPLC analyses of mouse, rat, and human plasma samples after i.v. injection of [18F]CPFPX and the product profiles obtained by microsomal oxidation. The radioactive trace reflects the quantitative metabolic composition, assuming that every molecule of the metabolites contains one radioactive fluorine-18 atom. Native tracer had a retention time of 41.9 min; the substances in minor concentrations having retention times less than 10 min were disregarded. Human microsomes from different sources generated identical metabolite profiles. Plasma and microsomal preparations contained seven well defined peaks representing radiofluorinated metabolites ranging from 11 to 22 min (P1–P7; Table 1, columns 1 and 2).
|
|
LC-MS Studies of Microsomal CPFPX Metabolites. In LC-MS studies, no fragmentation of CPFPX was seen at a cone voltage of 60 V. None of the [M + H]+ ions in the mass spectra of microsomal extracts reported in Table 1 (columns 1–3) were present in control incubations containing all reactants except CPFPX. Augmenting substrate concentration by the addition of carrier CPFPX caused no qualitative changes in the metabolite profile determined by HPLC.
Table 1 shows the results obtained by scanning the chromatograms for the [M + H]+ ions having m/z values equivalent to those of monohydroxylated derivatives (339, P1, P2, P4, P5, P7), an oxo derivative (337, P3), and a difunctionalized oxo-desaturation product (335, P6). Scanning for [M + H]+ ions of masses equivalent to those of despropyl CPFPX, m/z 281, or desfluoropropyl CPFPX, m/z 263, detected none, evidence that the eliminations expected from oxidations of C1 of the propyl or fluoropropyl substituents did not occur. Likewise, a scan for the [M + H]+ ion of a descyclopentyl derivative, m/z 255, detected no such metabolite, evidence that cleavage of the xanthine-C8/cyclopentane-C1 bond did not occur.
Identities of [18F]CPFPX Metabolites in Plasma by LC-ESI-In Source Fragmentation Experiments. Fragmentation of CPFPX and its desfluoro analog DPCPX occurred at a cone voltage of 110 V. Therefore, to clarify the fragmentation processes, the peak fractions (P1–P7) obtained by LC-MS were subjected to in source fragmentation. Figure 3 shows spectra obtained at a cone voltage of 110 V. From these data, it is obvious that the fluorine atom in CPFPX leads to dramatic changes in the fragmentation reactions compared with its unfluorinated counterpart, DPCPX. Figure 4 shows plausible fragmentation routes for CPFPX and DPCPX.
|
|
Table 1, column 4, lists fragments obtained by this method from microsomal LC-MS experiments. It is conspicuous that all but one hydroxylated derivative, P2, of m/z 339, had fragments of m/z 321, corresponding to a desaturated species formed by the loss of a molecule of water. The fragmentation reactions mainly reflect eliminations of HF and H2O, C-N bond cleavage, and ring closures according to those proposed for theophylline (Beaudry et al., 2005). Assuming that the fragmentation of in vitro formed metabolites follows the same main pathway as that for CPFPX, it is reasonable to suppose that P1, P2, and P7 are hydroxylated and P3 oxidized at the cyclopentyl moiety, P4 and P5 are hydroxylated at the N1-propyl chain, and P6 is difunctionalized at the cyclopentyl ring. A detailed interpretation of the various fragments obtained is shown in column 4 of Table 2. The fragmentation scheme depicted in Fig. 4 accounts for the m/z values of fragments found during the in source fragmentation of P1 to P7. Table 1, column 4, also lists fragments not explained by the fragmentation scheme, probably due to either species-specific fragmentation pathways or impurities.
|
|
Discussion |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
It is not surprising that [18F]CPFPX undergoes oxidation in humans; alkylxanthines such as caffeine and theophylline are substrates for oxidation and demethylation by cytochrome P450 in liver microsomes (Lohmann and Miech, 1976). Although it was possible to identify radioactive metabolites of [18F]CPFPX in the plasma of subjects receiving this radioligand for PET, the mass of tracer given, 0.5 to 5 nmol/subject, was too low to permit their identification. Accordingly, the use of human liver microsomes to obtain measurable amounts of these metabolites was a reasonable first step toward establishing their identities.
The profile of [18F]CPFPX metabolites in rodent plasma was qualitatively similar to that in human plasma and, therefore, tends to reflect metabolism in humans as well as in animal models. Alternatively, this study demonstrated that an in vitro liver microsome system is appropriate to validate chromatographic procedures and might reflect human metabolism as well as animal models. Furthermore, use of this in vitro technique allows fewer animal experiments, particularly for the evaluation of chromatographic techniques.
Quantitative receptor PET studies require analytical methods that accurately measure unchanged radioligand in plasma over the time of observation. In the present study, the recovery of radioactivity in the low protein extract was an equally important requirement because it ensured that a chromatographic profile reflected the composition of all the labeled compounds in plasma at the time of sampling. In other words, a solvent that extracts the unchanged ligand quantitatively might not be suitable for extracting the metabolites that become the main components at later times. Methanol/acetonitrile, 50:50 (v/v) fulfilled both criteria. Solvent interfered negligibly with the TLC system used in this study, as shown by the small amounts of radioactivity remaining at the origin in thin-layer chromatograms of mouse and human plasma extracts. In Fig. 1A, for example, the chromatogram of human plasma sampled at 1 min had a detectable spot at the origin. However, the activity in that spot was less than 1% of the radioactivity in the only other spot present at that time, that of [18F]CPFPX.
Since, in LC-MS experiments, no fragmentation of CPFPX occurred at a cone voltage of 60 V, it was a reasonable assumption that scanning the LC-MS chromatograms at the m/z values of the [M + H]+ ions of possible metabolites would detect the intact metabolites rather than fragments generated during spectrometry. The observation that the number of metabolites detected by MS and by radioactivity monitoring agreed validates that assumption (data not shown). Moreover, such a result implies that all main metabolites from the microsomal oxidation have been identified. Those measurements were consistent with hydroxy and oxo derivatives and one difunctionalized oxo-desaturation product as the main metabolites in human plasma.
Chromatographic comparison of blood samples taken 30 min p.i. with the metabolites generated by microsomes showed five hydroxylated metabolites. Plasma seems to contain fewer metabolites than generated by microsomes and in different proportions. Whereas the microsome preparations are closed systems from which metabolites cannot escape, plasma is an open compartment that loses tracer by distribution to other tissue compartments and by excretion.
Finally, there were no ions corresponding to a despropyl or desfluoropropyl derivative, evidence against hydroxylation, or oxidation of C1 of either the 1-propyl or 3-fluoropropyl groups. Likewise, no [M + H]+ ion had the mass of a descyclopentyl derivative.
The fact that the 3-fluoropropyl moiety resisted dealkylation and defluorination is consistent with experiments in mice (Holschbach et al., 1998a) showing that fluorine does not accumulate in bone and with PET scans of humans receiving [18F]CPFPX (Bauer et al., 2003; Meyer et al., 2004), which likewise show little or no deposition in bone. Those observations could be an important lead for the development of ligands more stable than [18F]CPFPX and thus more suited to imaging peripheral receptors.
Oxidation of C1 of the 1-propyl- or 3-(3-fluoropropyl) group would cause dealkylation. No such ions were found, eliminating that possibility. The strong electronegativity of the fluorine substituent prevents oxidation of C3 of the 3-(3-fluoropropyl) substituent. Accordingly, there are six theoretical positions of potential functionalization. We found that only five positions hydroxylated. Scans for fragments consistent with a hydroxylated fluoropropyl derivative were negative, meaning that the fluoropropyl moiety does not undergo metabolism.
LC-MS with in source fragmentation identified the cyclopentyl moiety as the part of the molecule most functionalized by liver microsomes and in vivo oxidations. Except for two metabolites hydroxylated at the N1 propyl chain, which are formed only in minor quantities in vivo, all oxidative modifications found took place at the cyclopentyl ring.
|
Footnotes |
|---|
ABBREVIATIONS: CPFPX, 8-cyclopentyl-3-(3-fluoropropyl)-1-propylxanthine; A1AR, A1 adenosine receptor; DPCPX, 8-cyclopentyl-1,3-dipropylxanthine; PET, positron emission tomography; HPLC, high-performance liquid chromatography; LC-MS, liquid chromatography-mass spectrometry; TLC, thin-layer chromatography; p.i., postinjection.
Address correspondence to: Dr. Dirk Bier, Institut für Nuklearchemie, Forschungszentrum Jülich GmbH, D-52425 Jülich, Germany. E-mail: d.bier{at}fz-juelich.de
|
References |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
 |
 |
|
 |
 |
 |
