Abstract
Accelerator mass spectrometry (AMS) has been used in a human mass balance and metabolism study to analyze samples taken from four healthy male adult subjects administered nanoCurie doses of the farnesyl transferase inhibitor 14C-labeled (R)-6-[amino(4-chlorophenyl)(1-methyl-1H-imidazol-5-yl)methyl]-4-(3-chlorophenyl)-1-methyl-2(1H)-quinolinone ([14C]R115777). Plasma, urine, and feces samples were collected at fixed timepoints after oral administration of 50 mg [14C]R115777 (25.4 Bq/mg or 687 pCi/mg i.e., equivalent to 76.257 × 103 dpm) per subject. AMS analysis showed that drug-related 14C was present in the plasma samples with Cmax values ranging from 1.6055 to 2.9074 dpm/ml (1.0525–1.9047 μg/ml) attmax = 2 to 3 h. TheCmax values for acetonitrile extracts of plasma samples ranged from 0.3724 to 0.7490 dpm/ml in the four male subjects. Drug-related 14C was eliminated from the body both in the urine and the feces, with a mean total recovery of 79.8 ± 12.9% in the feces and 13.7 ± 6.2% in the urine. The majority of drug-related radioactivity in urine and feces was excreted within the first 48 h. High-performance liquid chromatography (HPLC)-AMS profiles were generated from radioactive parent drug plus metabolites from pooled diluted urine, plasma, and methanolic feces extracts and matched to retention times of synthetic reference substances, postulated as metabolites. All HPLC separations used no more than 5 dpm injected on-column. The radioactive metabolite profiles obtained compared well with those obtained using liquid chromatography/tandem mass spectometry. This study demonstrates the use of AMS in a human phase I study in which the administered radioactive dose was at least 1000-fold lower than that used for conventional radioactive studies.
Radiolabeled organic xenobiotics are widely used in biomedical science for bioavailability and biotransformation studies in both animals and humans (Dain et al., 1994; Reith et al., 1998; Argenti et al., 2000). The isotope of choice for these studies is 14C, a low energy β-emitter usually detected by decay counting.14C is a useful labeling agent for drugs because it is often possible to incorporate 14C in a chemically and metabolically stable position of the xenobiotic molecule. However, decay counting as a method of analyzing14C is an inefficient process owing to the long radioactive half-life of this isotope of 5740 years. In any one year, only 0.012% of 14C atoms decay in a sample; to detect 1 dpm requires approximately 109 atoms to be present in a sample. 14C-labeled drugs are widely used in human mass balance studies, in which an inventory of the parent drug and metabolites can be measured in urine, feces, and plasma (Garner et al., 2000; Young et al., 2001). In such studies, an appropriate dose of the 14C-labeled drug is administered to healthy subjects, and the concentrations of radioactivity in plasma, urine, and feces are determined by decay counting.
In the mid-70s, accelerator mass spectrometry (AMS1), a nuclear physics technique, was developed primarily for use in radiocarbon dating (Bennett et al., 1977; Nelson et al., 1977). AMS uses a Van de Graaff accelerator operating at millions of volts to provide the energy potential to permit the outer valency electrons to be stripped away from negatively ionized atoms. These positive-charged atoms can then be separated using conventional mass spectrometry techniques. AMS separates elemental isotopes contained within a sample through differences in mass, charge, and energy. For the measurement of the isotopes of carbon viz. 12C,13C, and 14C, AMS separates these from each other and quantitates the two former using Faraday cups and the latter with a gas ionization detector (Vogel et al., 1995;Fifield, 1999). The output of an AMS instrument is an isotope ratio resulting from the sample being analyzed. As AMS analyses measure the number of 14C atoms present, rather than relying on decay counting, AMS measurements can be up to 1 × 106 times more sensitive than decay counting methods for 14C.
Preliminary studies have demonstrated the potential utility of AMS in pharmaceutical research, particularly for dose reduction of radioactivity administered to humans (Kaye et al., 1997). In addition, the value of AMS has been demonstrated in many toxicology and cancer research studies, in particular for measuring macromolecular adducts and especially DNA adducts (Mauthe et al., 1999; Lightfoot et al., 2000).
In this study, we describe for the first time the use of AMS analysis in a human mass balance and HPLC metabolite-profiling study for a novel inhibitor of farnesyl transferase, R115777. This inhibitor exerts its antitumor activity by blocking post-translational modification of p21ras. The study was conducted using only 50 nCi of [14C]R115777 admixed with 50 mg of unlabeled R115777. Based on extrapolation of the absorption, distribution, metabolism, and excretion characteristics of [14C]R115777 in rats, the radioactive dose administered to the healthy male subjects participating in this study was estimated to be less than 1 μSv (0.1 millirem). This radiation dose is exempt from regulatory approval in the United Kingdom as being comparable with everyday radiation exposures.
Materials and Methods
Chemicals and Glassware.
[14C]R115777 was radiosynthesized by labeling the asymmetric carbon atom (Fig. 1) and had a radiochemical purity of 99.2%. This material was mixed with unlabeled R115777 to provide material of the appropriate-specific radioactivity for human administration. Methanol (HPLC grade), methanol (Analar), HPLC grade water, HPLC grade acetonitrile, HPLC grade dimethylsulphoxide, HPLC grade ammonium acetate, and concentrated HCl were purchased from Fisher Scientific (Loughborough, UK). Kimax culture tubes, tributyrin, copper oxide wire (ACS), cobalt powder (100 mesh, 99.9% w/w), zinc powder (100 mesh, 98.98% w/w), titanium (II) hydride (325 mesh, 98% w/w), and cobalt powder (100 mesh, 99.9% w/w) were all purchased from Aldrich Chemical Co. (Poole, Dorset, UK). All glassware for sample graphitization was purchased from York Glassware Ltd. (York, UK) and was baked at 500°C before use. Australian National University sugar (certificated value = 1.5061 fraction modern) was kindly provided by The Quaternary Dating Research Center, Australian National University, Canberra, Australia. Synthetic graphite (200 mesh, 99.9999% w/w) was from Alfa Aesar (Karlsruhe, Germany) and contained aluminum powder (99.99%) from Acros Organics/Fisher Scientific. Poco graphite rod (1 mm diameter) and aluminum cathodes were purchased from National Electrostatics Corporation (Middleton, Wisconsin). Urea used as a standard for carbon, hydrogen, nitrogen analysis, tin capsules, and Chromosorb W were all purchased from Elemental Microanalysis Ltd. (Okehampton, Devon, UK).
Human Dosing with [14C]R115777.
Four healthy male subjects received a single oral dose of 50 mg of [14C]R115777 (34.35 nCi/subject equivalent to 76.257 × 103 dpm/subject) as a suspension in 20 ml of water. Blood was collected from each subject once before dosing and at 1, 2, 3, 4, 6, 8, 24, 32, 48, 72, 96, and 168 h after dosing. Plasma was obtained from the blood samples by centrifugation. Urine was collected in intervals of 0 to 4, 4 to 8, 8 to 24, 24 to 32, 32 to 48, 48 to 72, 72 to 96, 96 to 120, 120 to 144, and 144 to 168 h after dosing. Feces were collected per stool during the 0- to 168-h period after dosing. Feces samples were homogenized in and extracted with methanol. Methanolic feces extracts and solid fecal residues were separated by filtration. Fecal residues were dried and ground to a fine powder. All samples were frozen prior to shipment to York, UK for AMS analysis.
AMS Analysis of Plasma, Urine, Fecal Residues, Fecal Extracts, and HPLC Fractions.
All samples prior to being graphitized were analyzed by liquid scintillation counting (model number Packard Tri-Carb TR/SL 2770; Packard Bioscience Co., Pangbourne, UK) to ensure that samples above 20 dpm/ml or g were not inserted into the AMS instrument.
Full details of the AMS analysis procedure are described by Garner et al. (2000). In brief, duplicate aliquots of plasma, urine, fecal residues, and extracts and HPLC fractions were graphitized according to the method described by Vogel (1992). The resulting graphite/cobalt mix was pressed into aluminum cathodes and each sample analyzed using a 5 MV 15SDH-2 Pelletron AMS system (National Electrostatics Corporation). Samples containing minimal amounts of carbon were bulked up using tributyrin. Each sample was counted in the AMS instrument for a minimum of 100 s, and this was repeated for each sample at least three times. AMS data were converted to either disintegrations per minute per milliliter or grams taking into account the carbon content of the sample and any added carrier. Carbon contents were measured using a carbon, hydrogen, nitrogen analyser (Elemental Microanalysis Ltd.).
HPLC Fractionation of Plasma, Urine, and Fecal Extracts.
A combined 0- to 24-h-diluted urine pool, a combined pool of the methanolic extracts of a selection of the feces samples, and a combined 3-h plasma pool from the four subjects were fractionated by HPLC. In addition, aliquots of the 3-h plasma pool and the 0- to 24-h-diluted urine pool were incubated for16 h at 37°C with Escherichia coli β-glucuronidase (Boehringer Ingelheim GmbH, Ingelheim, Germany).
The sample preparation was as follows. The combined 0- to 24-h-diluted urine pool (1.5 ml) was centrifuged and an aliquot of 1000 to 1500 μl of the supernatant was taken for injection directly onto the HPLC column (two 500 μl injections were performed). Two milliliters of the combined pool of the methanolic feces extract were evaporated under a stream of nitrogen gas. The residue was dissolved in 500 μl of DMSO. An aliquot of 200 to 300 μl of the supernatant solution in DMSO was taken for injection directly onto the HPLC column (two 100-μl injections were performed). One milliliter of the combined pool of the 3-h plasma sample was deproteinized by addition of 1.5-ml acetonitrile. The precipitated proteins were removed by centrifugation, and the supernatant was collected and evaporated to dryness. The evaporation residue was redissolved in 500 μl of DMSO and 500 μl of distilled water and vortexed. The supernatant of the DMSO/water solution was taken for injection directly onto the HPLC column (two 300-μl injections were performed). Each injection also contained a mixture of reference metabolites to permit UV detection of metabolite peaks.
HPLC-AMS was conducted as follows. The HPLC apparatus (Model HP 1100; Agilent Technologies Inc., Wilmington, DE) consisted of a HPLC-gradient pump, an automatic injector, and a stainless steel Luna C18 reverse phase column (4.6 mm i.d. × 250 mm with a particle size of 5 μm). The column was eluted as follows: a linear gradient over 5 min from 100% of an aqueous solution of 0.1 M ammonium acetate, adjusted to pH 5.5 (solvent system A), to 40% v/v of solvent system A and 60% v/v of a mixture of an aqueous solution of 1.0 M ammonium acetate, adjusted to pH 5.5/methanol/acetonitrile (10/10/80) (solvent system B). The flow composition was then held for 15 min, followed by a linear gradient over 10 min from 40% v/v solvent system A and 60% v/v solvent system B to 20% v/v solvent system A and 80% v/v solvent system B. Finally, a linear gradient over 3 min to 100% v/v solvent system B was run and held for a further 27 min. A linear flow profile was also included from an initial flow of 1.0 ml/min for 10 min to 0.5 ml/min at 10.01 min and held at 0.5 ml/min until 30 min into the run. At this point, the flow was increased from 0.5 to 1.0 ml/min at 30.01 min and held until completion of the fraction collection procedure. The diode array detector was set at 230 nm. During the HPLC run, postcolumn eluent was collected manually, with the aid of a stopwatch in 1-min fractions from 0 to 60 min (urine and plasma) or in one fraction from 0 to 15 min, 1-min fractions from 15 to 45 min and one fraction from 45 to 60 min (fecal extract).
Data Analysis.
The amounts of 14C excreted in urine or feces (as determined by AMS) were expressed as a percentage of the administered dose of 14C. The levels of14C in plasma (as determined by AMS) are presented as microgram-equivalent R115777 per milliliter.
Results
In this low radioactive dose study, 50 mg of [14C]R115777 (34.35 nCi/subject) as a suspension in 20 ml water was administered to four healthy male subjects. Figure 2 shows the elimination of radioactivity from plasma for all subjects expressed as either radioactive content (Fig. 2A) or drug plus metabolite content (Fig.2B). The radioactive Cmax for all subjects was between 1.6055 and 2.9074 dpm/ml, which equated to a drug equivalent concentration of R115777 between 1.0525 and 1.9047 μg/ml. Plasma samples obtained at the various timepoints from this study were precipitated with acetonitrile, the resulting supernatant taken and analyzed by AMS. Approximately 40% of the plasma radioactivity was recovered by acetonitrile extraction.
Urine and feces were recovered from the dosed individuals over the study period to establish the percentage of the administered dose excreted. Figure 3 shows the cumulative total dose excreted via both routes. The mean percentage excretion in all four subjects was 93.5 ± 6.7% (range 83.7–98.5%) after 7 days. The cumulative urine radioactivity accounted for 13.7 ± 6.2% (range 9.3–22.6%) and for the feces 79.8 ± 12.9% (range 61.1–89.2%).
In addition to obtaining drug mass balance and plasma concentration data, it is valuable to also obtain HPLC profiles of both parent drug and metabolites in urine, fecal extracts, and plasma. For large doses of radioactivity, this often involves on-line detection using a radiodetector coupled to a HPLC. It is not possible to couple a HPLC to an AMS instrument because of the high vacuum (10−9 torr) at which these instruments operate. Hence analysis must be conducted off-line by collecting timed fractions after sample injection onto a HPLC.
We have examined metabolite profiles of pooled urine sample before and after β-glucuronidase treatment, a pooled methanolic fecal extract, and a pooled plasma sample before and after β-glucuronidase treatment (Fig. 4). Each profile was obtained along with coinjection of synthetic reference substances monitored by UV absorbance. Table 1 shows the relative abundance of unchanged R115777 and some of its metabolites in the samples detailed in Fig. 4. Detailed pharmacokinetic analysis of the plasma radioactivity is presented in Table 2for total radioactivity in deproteinized samples and in Table3 for unchanged R11577. Figure5 shows the structure of the synthetic reference substances that were used for the HPLC cochromatography studies.
It should be emphasized that this paper focuses on the radioactive metabolite profiles of R115777 and not the characterization of these metabolites using LC/MS/MS methods.
Discussion
The introduction of new enabling technologies in drug discovery or research permits new ways of approaching scientific problems which could not be addressed prior to the new technology being developed. AMS is just such an enabling technology, but its use has been limited in biomedical research through lack of access to the technology by biomedical researchers. Recently a number of papers have been published on using AMS for pharmaceutical research that indicate the considerable potential of AMS to analyze drugs either at pharmacological or subpharmacological doses (Barker and Garner, 1999; Turteltaub and Vogel, 2000).
In this clinical study, we have examined the use of AMS to quantitate ultralow levels of drug radioactivity that fall far below the limit of detection of liquid scintillation counting. Healthy male subjects were administered approximately 34 nCi of 14C-labeled drug admixed with 50 mg of unlabeled drug (R115777). The radiation exposure for this study was below 1 μSv, a dose which in the United Kingdom does not require regulatory approval. Blood, urine, and fecal samples were collected over 7 days to obtain mass balance and plasma pharmacokinetic information.
Over 90% of the drug was excreted over the 7-day period with the majority being excreted in the feces. Since no mass balance from any subject exceeded 100%, this gave us confidence that no contamination was being introduced during sample collection, processing, or analysis. During the 0- to 72-h period after dose administration, the radioactivity was fairly rapidly excreted in the urine with an excretion half-life of 23.3 +/− 3.6 h. During the 72- to 168-h period after dosing, the urinary excretion was much slower. The radioactivity in the feces was mostly excreted during the 0- to 72-h period after dose administration.
Plasma radioactivity was determined without and with a protein precipitation step. Protein precipitation was necessary in this study because the blood concentration of [14C]R115777 was low in comparison to the 14C content of plasma proteins. Concentrations of total radioactivity in nonprecipitated plasma samples were nearly 4 times higher than those in deproteinized plasma samples. However, the drug-related radioactivity concentrations in the nonprecipitated plasma samples were considered to be potentially inaccurate, because they were maximally only about 5 times higher than the background radioactivity concentrations in blank plasma samples. The radioactivity concentration in some nonprecipitated plasma samples collected after dose administration was sometimes lower than the background radioactivity concentration in the nonprecipitated blank plasma sample, collected from the same subject before dose administration. The high background level of radioactivity was almost certainly due to naturally occurring 14C in plasma proteins. The background radioactivity levels in plasma samples were reduced approximately 25-fold by deproteinization. Thus, the determination of total radioactivity concentrations in deproteinized plasma samples was much less affected by the background radioactivity and therefore more accurate than the determination of total radioactivity concentrations in nonpretreated plasma samples. However, one major disadvantage of the determination of total radioactivity concentrations in deproteinized plasma samples is that one may underestimate total radioactivity concentrations, due to occlusion of drug-related radioactivity in the precipitated proteins. To minimize the risk of occlusion, it is advisable to use a mild-protein precipitation method (e.g., addition of acetonitrile). In the present study, the differences in the total radioactivity concentrations in nonpretreated and deproteinized plasma samples may be partially due to occlusion of drug-related radioactivity in precipitated proteins.
Concentrations of unchanged R115777 in plasma samples were on average approximately 42 times lower than the total radioactivity concentrations in nonprecipitated plasma samples and approximately 8 times lower than the total radioactivity concentrations in deproteinized plasma. This finding confirms the metabolite profiles, demonstrating that unchanged R115777 is only a relatively minor plasma constituent. This finding was confirmed by LC/MS/MS (data not shown).
HPLC followed by fraction collection and AMS analysis provides a way of separating out the various components that account for total radioactivity in the sample being analyzed. In this study, we have obtained radiochromatograms of a pooled urine sample, a methanolic feces extract, and pooled plasma extract. For both the urine and plasma samples, a major polar peak was seen with a retention time of 12.5 min that shifted to a longer retention time after treatment with β-glucuronidase. The post-β-glucuronidase-treatment peak corresponded to the parent drug R115777 (retention time circa 34 min). Confirmation of these assignments came from cochromatography with authentic reference substances and from LC/MS/MS analysis of these samples.
In the HPLC studies of the overall urine pool (0–24 h), only one major radioactive peak was seen with a retention time of 12 to 18 min. This accounted for 96.0% of the radioactivity; glucuronidase treatment shifted this peak to a retention time of 30 to 33 min, which corresponded to unchanged parent drug. Hence, this peak was tentatively assigned as the glucuronide of unchanged parent drug. This was confirmed by LC/MS/MS (data not shown) in which only one major peak could be detected.
In the HPLC analysis of the overall pool of methanolic fecal extracts, four major and one minor radioactive peak was observed. Based on cochromatography, the parent drug and metabolites were assigned. Some discrepancies were observed between these results and the LC/MS/MS analysis. Loss of the methyl-imidazole moiety, resulting in the formation of R101763 and R104209 appeared to be a minor metabolic pathway based on LC/MS/MS analysis but appeared to be a major metabolic pathway based on AMS analysis. This may be due to the instability of these metabolites in methanolic feces extracts. The LC/MS/MS analyses were conducted several months later than the AMS analyses. Unless a direct comparison using the two analytical techniques is made on the same methanolic fecal extract at approximately the same time, this discrepancy cannot be resolved.
Using AMS analysis of collected fractions, the HPLC analysis of a pooled plasma sample showed R115777-glucuronide was a major metabolite with a concentration 2 to 8 times higher than unchanged R115777. R130525 and R104209 were minor metabolites with concentrations that were 3- to 12-times lower than those of unchanged R115777. The metabolite profile determined by LC/MS/MS analysis corresponded well with that determined by AMS analysis both from a qualitative and quantitative point of view (data not presented).
The data reported here are of significance not only in terms of the overall metabolism of R115777 but as a demonstration that AMS can be used to analyze radioactivity levels far below the limit of detection by liquid scintillation counting. The dose of radioactivity used in this study was at least 1000-times lower than would be used for a conventional high-dose radioactivity study. Nevertheless, despite the lower quantities of radioactivity used, all the standard pharmacokinetic parameters can be obtained using the AMS technique (see Tables 2 and 3).
We and others have previously shown that AMS data compares well with liquid scintillation counting data but that the detection limits of AMS can be many thousands of fold lower (Freeman and Vogel, 1995). As with all other analytical techniques, it is the background level that will determine the limit of detection for the procedure. In the case of plasma, removing the plasma proteins by precipitation reduces the background by some 25-fold thus enhancing the signal-to-noise ratio. HPLC fractionation of a plasma extract increases the signal still further by removing interfering substances that contain14C. In another clinical study with AMS detection, the limit of quantification for a HPLC fraction of serum extract was 0.009 dpm/fraction compared with a limit of quantification of 0.43 dpm/ml for neat serum (Young et, 2001).
For liquid scintillation counting, the precision of the analysis will be determined by the number of events (disintegrations) collected during the counting period. For this study, samples were analyzed in the AMS instrument for either 100 or 250 s. Each single cathode was run at least three times and the number of counts collected was used to compute the radioactivity content. For a predose plasma sample, sufficient counts can be collected in 100 s to obtain the necessary precision (below 5%). To get the same precision using liquid scintillation counting for a predose sample would require the sample to be counted for approximately 100 h. The main disadvantage of using low-dose radioactivity and AMS detection for human studies compared with high-dose studies is the inability to use on-line detection rather than fraction collection for chromatographic analysis. We have found for all other applications that AMS does not slow the analysis of samples compared with liquid scintillation counting. In addition, as samples have to be combusted for AMS analysis, issues such as color quenching, volume of sample required to obtain a signal, or matrix effects are negated compared with liquid scintillation counting.
The potential applications of AMS in pharmaceutical research have been reviewed by a number of authors (Garner, 2000; Garner and Leong, 2000;Papac and Shahrokh, 2001). This study demonstrates that the same data can be obtained with AMS as can be obtained by decay counting but that the administered human radioactive doses can be reduced by at least 1000-fold. From an ethical and environmental standpoint, we believe the use of AMS will become routine for human radioactive studies. In addition, this ultrasensitive analysis procedure permits new areas of drug development to be investigated that cannot be conducted using other analytical methods. These include human microdosing for early absorption, distribution, metabolism, and excretion and pharmacokinetic information; analysis of biotechnology products; and the use of low-dose radioactivity for nonoral routes of administration.
Footnotes
-
This research was financially supported by Janssen Pharmaceutica, Beerse, Belgium.
- Abbreviations used are::
- AMS
- accelerator mass spectrometry
- HPLC
- high-performance liquid chromatography
- DMSO
- dimethyl sulfoxide
- LC/MS/MS
- liquid chromatography/tandem mass spectometry
- R115777
- (R)-6-[amino(4-chlorophenyl)(1-methyl-1H-imidazol-5-yl)methyl]-4-(3-chlorophenyl)-1-methyl-2(1H)-quinolinone
- Received January 22, 2002.
- Accepted March 29, 2002.
- The American Society for Pharmacology and Experimental Therapeutics