Introduction

Next to AIDS, tuberculosis (TB) is one of the most important infectious diseases in the world. Although drugs are available that can prevent, treat, and cure TB [1], newer and better drugs are urgently needed to shorten the duration of TB treatment and to reduce the emergence of drug resistance. In 2005, Andries et al. [2] reported the potent antimycobacterial properties of a diarylquinoline, R207910 (also known as TMC207) (Fig. 1). Owing to its novel mechanism of action, the new compound is active against all multidrug-resistant strains of TB tested so far, and hence, may have the potential to improve and shorten the treatment of TB.

Fig. 1
figure 1

Molecular structure and exact mass (79Br isotope) of the new anti-TB compound R207910 and its metabolites identified in rat and dog faeces extracts (0–24 h post-dose)

Full size image

Although pharmacokinetic studies of R207910 in mice are promising, a profound study of the compound’s metabolism in humans is required before the prospective drug can enter definitive clinical testing. Since the development of atmospheric pressure ionization (API) sources, mass spectrometry has become the preferred analytical tool for the detection and identification of metabolites [3–6]. Because electrospray ionization (ESI) is the most “soft” ionization technique, limiting fragmentation of analyte ions, it is generally preferred for metabolite identification [7, 8]. The intensity of the MS signal obtained with the API sources strongly depends on the chemical structure analyzed. Therefore, authentic standards should be used for MS quantification, but these are generally not available in metabolism studies. Radiotracer technology (14C or 3H) is most often used instead and is still the method of choice to study the in vivo disposition of a new drug [9]. There can however be ethical reasons or cost concerns that hinder the use of radiotracers. To circumvent the disadvantages of radioactivity detection, the combination of HPLC and inductively coupled plasma mass spectrometry (ICP-MS) was recently introduced as a promising technique for the analysis of drug compounds containing an element detectable by ICP-MS [10–14]. As the anti-TB compound under investigation contains a Br atom in the molecule, it was possible to use HPLC with ICP-MS detection for the metabolite profiling in faeces samples after dosing the compound to dogs and rats. In a previous paper [10], the technical aspects of the method and the modifications to the standard experimental conditions were described and some first (qualitative) data obtained with the method developed were presented.

The first aim of this work is to present some information on the quantitative aspects of the HPLC-ICP-MS method. All results were compared with those obtained via radio-HPLC for faeces samples from the same dog and rat, dosed with [14C]R207910. Secondly, the value of specific tools helpful in the identification of metabolites will be demonstrated. Since ICP-MS detection does not provide structural information, HPLC-ESI-MS is used instead. Ideally, a combination of ICP-MS and ESI-MS detection coupled to one single HPLC system can be deployed for simultaneous quantification and structure identification of bromine-containing compounds [15–18]. However, since both techniques are rather expensive and generally used for quite different applications, the combination of both ICP-MS and ESI-MS on a single HPLC system is not very common. Different configurations of HPLC systems are often used, with the disadvantage that even the smallest retention time shifts can jeopardize the link between the structure characterized with ESI-MS and the quantitative ICP-MS data. This work demonstrates that in both ICP-MS and ESI-MS, the 79Br/81Br isotope ratio can be used to obtain qualitative “bromatograms”, ensuring the correct link with the quantitative data obtained by means of ICP-MS, even if both instruments are coupled to different HPLC devices.

Experimental

Instrumentation

HPLC-ICP-MS

All measurements were carried out using a Perkin-Elmer SCIEX DRCPlus quadrupole-based ICP mass spectrometer, equipped with a dynamic reaction cell (DRC). The chromatographic system used was an Äkta™ purifier HPLC system (Amersham Pharmacia Biotech, UK). To enable the coupling of HPLC with ICP-MS, several modifications were made, e.g., the use of an alternative nebulizer, spray chamber, and injector tube [10]. The sample introduction system consisted of a PFA-LC nebulizer and a PC3 Peltier cooled inlet system (both from Elemental Scientific, Inc., Omaha, USA). The PC3 unit incorporates a cyclonic spray chamber and was operated at a temperature of +2 °C. The standard injector tube (2-mm internal diameter) was replaced by a quartz injector with an internal diameter of only 1 mm. In order to avoid carbon buildup on the ICP torch and the interface cones, the nebulizer gas was admixed with 6% v/v oxygen using an external mass flow controller (MFC 5876/controller 5850E) (Brooks Instrument, Veenendaal, The Netherlands). Sampler and skimmer cones were made of platinum. A T-piece was used to mix the effluent of the HPLC column with the species-unspecific spike solution that was delivered by an additional HPLC pump (HP1050 Series) at a flow rate of 100 μL min−1. The T-piece was connected to the liquid sample inlet of the nebulizer using PEEK tubing. The instrument settings and data acquisition parameters used throughout this work are summarized in Table 1.

Table 1 Chromatographic conditions and experimental conditions used for the ICP-MS measurements throughout this work
Full size table

Radio-HPLC

The online radioactivity measurements were performed on a Waters Alliance 2695 HPLC system (Waters, Milford, Massachusetts, USA), equipped with a Waters PDA996 UV photodiode array detector and a Berthold LB509 liquid scintillation radiodetector (Berthold, Bad Wildbad, Germany). The radiodetector was equipped with a 1-mL-volume Z-1000-4 cell (Berthold, Bad Wildbad, Germany). Prior to detection in the cell, 4 mL min−1 of Ultima Flow M scintillation liquid (Perkin Elmer, Boston, MA, USA) was admixed with the HPLC eluent via a low death volume T-piece by a Berthold scintillator pump. The chromatographic conditions used are shown in Table 1.

HPLC-ESI-MS

Two different configurations were used for the HPLC-ESI-MS measurements: a Q-Tof Ultima mass spectrometer (Waters, Milford, Massachusetts, USA), equipped with a Waters Alliance 2795 HPLC system, and an LTQ-Orbitrap mass spectrometer (Thermo, Bremen, Germany), equipped with a Waters Alliance 2695 HPLC system. The chromatographic conditions used on both systems are shown in Table 1.

The Q-Tof Ultima mass spectrometer was equipped with a dual electrospray ionization probe and was operated in the positive ion mode at a resolution of 8,000 (FWHM at m/z 409). The source temperature was 100 °C, desolvation temperature 250 °C, and the cone voltage was set at 40 V. The second LockSprayTM ESI probe provided an independent source of the lock mass calibrant H3PO4·NH4 +. The cluster ion at m/z 409.94184 was used as the calibrant in full MS, whereas the ion at m/z 392.91534 (daughter ion of m/z 490.9) is used as the lock mass in MS/MS mode. Data were acquired in the centroid mode with a scan time of 1 s and processed using Masslynx software 4.1.

The LTQ-Orbitrap mass spectrometer was equipped with an ESI source operated in the positive ion mode. Accurate mass measurements were obtained using external calibration. The LTQ-Orbitrap was operated at 60,000 (FWHM) resolution in MS and 7,500 in MSn. The source parameters were tuned for maximum sensitivity using a 10 ng μL−1 R207910 standard solution. The same solution was used to define the optimal collision energy deployed during MSn fragmentation. Data were acquired in the centroid mode and processed using Xcalibur 2.0 software.

Reagents

During the gradient elution, two solvents, denoted as A and B, were admixed in varying proportions. Solvent A consisted of 0.1 M ammonium acetate solution (pH 7.5) and solvent B was a mixture of 10% of 1 M ammonium acetate solution (pH 7.5), 45% methanol, and 45% acetonitrile. Ammonia (25%) (Merck, Germany) and/or acetic acid (ICP-MS: Panreac Quimica SA, Spain; other analyses: Merck, Germany) were used to bring the pH to a value of 7.5. Methanol and acetonitrile (both HPLC grade) were supplied by Panreac (Panreac Quimica SA, Spain) for the ICP-MS measurements and by Sigma-Aldrich (Sigma-Aldrich, Swiss) for the other analyses. Ammonium acetate was purchased from UCB (UCB, Belgium) for the ICP-MS measurements and from Acros (Acros organics, USA) for the other analyses. For preparation of the real-life samples, dimethylsulfoxide (DMSO) (ICP-MS: Merck, Germany; other analyses: JT Baker, The Netherlands) was used. The isotopic spike solution enriched in 81Br was prepared by dissolving NaBr (99.62% Na81Br, CK Gas Products Ltd, UK) in 0.14 M HNO3 and diluting to a concentration of 1 g L−1 NaBr. As the enrichment of the spike was too high for an accurate and precise characterization, 10.4 mL of the spike solution (1 g L−1 NaBr) was mixed with 1.9 mL of a Br solution with a natural isotopic composition (1 g L−1 Br) and diluted to 1 L with 0.14 M HNO3 to obtain a spike solution with a concentration of 10 mg L−1 and an isotopic abundance of approximately 90% for 81Br. The molar concentration of 81Br present in the spike solution was determined by means of reverse ID-MS using a Br standard solution of natural isotopic composition (1 g L−1 Br standard solution obtained from Inorganic Ventures, Spain) for the isotope dilution step. Unless mentioned otherwise, all reagents were of analytical grade. High-purity water was used throughout and was obtained by purifying doubly distilled water in a Milli-Q system (Millipore, USA). Single element standard solutions were diluted with HNO3, purified by sub-boiling distillation in quartz equipment.

Procedures

Animal dosing and sample pretreatment

An adult male Sprague-Dawley rat and male Beagle dog were dosed orally with an aqueous 10% hydroxypropyl-β-cyclodextrin (HP-β-CD) solution of a mixture of [14C]R207910 (specific activity 1.93 GBq mmol−1) and R207910 at 6 mg kg−1 (rat) and 10 mg kg−1 (dog) body weight, respectively. Faeces were collected from the animals at various post-dose intervals, e.g. 0–24, 24–48, and 48–96 h after dosing of R207910. Aliquots of the faeces samples were extracted with 10 mL methanol. The extracts were evaporated to dryness under nitrogen. The remaining extract was redissolved in 1 mL DMSO and, after vortexing and sonication, centrifuged at 13,000 rpm for 10 min. Subsequently, the supernatant was transferred into an injection vial and 100 μL of this supernatant (containing the same amount of compound-related material as 1 mL of the original faeces extract) was injected onto the column for radio-HPLC and HPLC-ICP-MS analyses. Only 20 μL of the same supernatant was injected onto the HPLC-ESI-LTQ-Orbitrap, while 10 and 60 μL were injected onto the HPLC-ESI-Q-Tof instrument for the analysis of the rat and dog faeces samples, respectively.

Post-column online isotope dilution analysis

To allow accurate quantification of the Br content of the metabolites, the effluent of the HPLC column was continuously mixed with the spike solution, enriched in 81Br [19]. The mixture was nebulized into the ICP and the signal intensities for 79Br and 81Br were measured, whereafter the 81Br/79Br ratio was calculated as a function of time. Subsequently, the 79Br and 81Br mass flow for the samples was calculated using Eqs. 1 and 2, respectively.

$$ {\text{M}}^{{{\text{79}}}}_{{{\text{sample}}}} {\left( t \right)} = \frac{{{\text{R}}_{{{\text{spike}}}} - {\text{R}}_{{{\text{mix}}}} {\left( t \right)}}} {{{\text{R}}_{{{\text{mix}}}} {\left( t \right)} - {\text{R}}_{{{\text{sample}}}} }} \cdot {\text{M}}^{{{\text{79}}}}_{{{\text{spike}}}} $$
(1)
$$ {\text{M}}^{{{\text{81}}}}_{{{\text{sample}}}} {\left( t \right)} = \frac{{{\text{R}}_{{{\text{sample}}}} }} {{{\text{R}}_{{{\text{spike}}}} }} \cdot \frac{{{\text{R}}_{{{\text{spike}}}} - {\text{R}}_{{{\text{mix}}}} {\left( t \right)}}} {{{\text{R}}_{{{\text{mix}}}} {\left( t \right)} - {\text{R}}_{{{\text{sample}}}} }} \cdot {\text{M}}^{{{\text{81}}}}_{{{\text{spike}}}} $$
(2)

where Rmix is the isotope ratio measured online during a complete chromatographic run of the mixture, and Rspike and Rsample are the measured isotope ratios for the spike and the sample, respectively. Msample and Mspike denote the mass flow of, respectively, the sample and the spike solution. Mspike was calculated from the flow rate (FR) and the concentration (c) of the spike.

$${\text{M}}_{{\text{spike}}} = {\text{FR}}_{{\text{spike}}} \cdot c_{{\text{spike}}}$$
(3)

Plotting these sample mass flows (sum of M79 sample and M81 sample) as a function of elution time, results in mass flow chromatograms for which integration of the peak area for each peak directly gives the mass of Br in the metabolites.

Results and discussion

HPLC-ICP-ID-MS

As described previously, an HPLC-ICP-MS method was developed [10] and used successfully for the qualitative metabolite study of faeces samples, after dosing the bromine-containing anti-TB compound R207910 to dogs and rats. However, to be able to fully assess and interpret the results of a metabolite study, quantitative information on the metabolites is required in addition to qualitative information. One of the major difficulties expected in obtaining accurate, quantitative results with the proposed method could be the use of a gradient elution required to allow separation of the metabolites, as it is well known that variations in the concentration of organic solvents introduced into the ICP, may have a substantial influence on the sensitivity [20, 21]. To tackle this problem, species-unspecific isotope dilution was used as a calibration technique, as first described by Rottmann and Heumann [19] in 1994. Following the experimental setup and the calculations described in the Experimental, the mass flow of Br in the samples was calculated on the basis of the results obtained for the measurement of the 81Br/79Br isotope ratio in the samples (Rsample), the spike (Rspike), and the mixture of sample and spike (Rmix), and the mass flow of the spike (Mspike). Figure 2a shows the 79Br and 81Br signal intensities detected online by ICP-MS during the chromatographic analysis of an isotope-diluted rat faeces sample, collected in the interval of 0–24 h after dosing with R207910. The corresponding 81Br/79Br isotope ratio and the Br mass flow are shown in Fig. 2b and c, respectively. It is clear that the individual signals of 79Br+ and 81Br+ are severely influenced by the varying concentrations of organic solvents in the column effluent during the linear gradient (increase of the signals between 20 and 45 min). However, the corresponding isotope ratio in this time interval is stable, which demonstrates the benefit of using isotope ratios instead of the single isotope intensities when gradient elution has to be used, especially for metabolites eluting from the column in the first 45 min of the analysis. In terms of accuracy, it can be mentioned that the isotope ratios do not deviate from the expected value by more than is expected on the basis of mass discrimination. From the mass flow chromatogram, the Br concentration in each metabolite can be calculated by integrating the corresponding mass flow peak. Table 2 summarizes the results obtained for the most important peaks that were observed in the chromatogram. To validate the method under investigation, these results were compared with those obtained with 14C radiochemical detection. There is good agreement between both techniques, especially taking into account that the sample preparation for the HPLC-ICP-MS analysis and for the radio-HPLC analysis were carried out in two different laboratories, and that two different HPLC systems were used. Calculation of the limits of detection (LODs) of both methods resulted in absolute values of 5 ng R207910 and 1 ng R207910 for HPLC-ICP-MS and radio-HPLC, respectively. Although at first sight, the absolute LOD for radio-HPLC is superior to that obtained for HPLC-ICP-MS, it should be mentioned that in vivo metabolism studies are never carried out with the pure radioactive compound, but with the 14C-labeled compound ‘diluted’ with nonradioactive compound. Therefore, only a few percent of the administered dose of R207910, i.e., the radioactive part of the dose, can be detected in radio-HPLC, while all molecules can be detected by means of ICP-MS. The higher the ‘dilution’ of the radioactive compound, the more the LOD as provided by radio-HPLC deteriorates. Table 2 provides additional chromatographic data for the faeces extract of a second rat for which enough sample was available to perform two independent analyses. In this way, some insight can be obtained into the repeatability of the quantification.

Fig. 2
figure 2

Chromatogram obtained by means of HPLC-ICP-MS for a an isotope-diluted rat faeces extract, b the corresponding Br isotope ratio chromatogram, and c Br mass flow chromatogram (numbers correspond to metabolites illustrated in Fig. 1)

Full size image
Table 2 Chromatographic data obtained for the major bromine-containing compounds in faeces of rats, collected in the period of 0–24 h after dosing with [14C]R207910, detected by means of HPLC-ICP-ID-MS and radio-HPLC
Full size table

A similar analysis was carried out for a methanolic faeces extract of a dog, collected in the 0–24 h period after dosing with R207910. Good agreement was also found between the results obtained for this sample by means of ICP-MS detection and radiochemical detection, as can be seen from Table 3. It should be mentioned that in order to obtain relevant information concerning the metabolism of R207910 in the animal body, the total amount of Br should be determined for all samples. In this way, comparison of the sum of the Br concentrations in the different peaks in the HPLC-ICP-MS chromatogram with the total Br concentration in the sample allows one to obtain an insight into the recovery. Total Br determinations in this work were performed by means of flow injection ICP-MS with single standard addition for calibration. In all cases, the recovery was found to be approximately 100% (e.g., for sample 2, reported on in Table 2, the total Br concentration was 2.30 mg L−1, whereas summing up the Br concentrations in the peaks results in a Br concentration of 2.39 mg L−1—average of the two replicate measurements). From these results, it is clear that the use of HPLC-ICP-ID-MS for quantitative metabolite studies is a promising alternative to the more traditional radiolabeling approach.

Table 3 Chromatographic data obtained for the major bromine-containing compounds in faeces of a dog, collected in the period of 0–24 h after dosing with [14C]R207910, detected by means of HPLC-ICP-ID-MS and radio-HPLC
Full size table

HPLC-ESI-MS

In the previous section, it was shown that HPLC-ICP-MS allows for accurate quantification of bromine-containing metabolites in faeces from rats and dogs, collected following dosing with R207910. However, as this approach affords no information concerning the metabolite structures, an additional analysis was performed by ESI-MS. The Br isotope ratios can also be useful in the search for metabolites in the complex background of endogenous compounds obtained using HPLC-ESI-MS analyses. The typical 1:1 ratio of the 79Br and 81Br isotopes makes it easier for the mass spectrometrist to distinguish between bromine-containing ions and the background. By employing the isotopic-data-dependent scanning settings in the ion trap software in combination with a well-chosen minimum signal threshold, the ratio can also be used to trigger the mass spectrometer so that only the ions of interest are selected for MSn fragmentation. In this way, all necessary structural information can be gathered in a single run. An additional advantage of this approach is the higher selectivity of the MS2 total ion current (TIC) chromatogram, where only peak signals are recorded for bromine-containing ions or other ions that show the same mass shift in a 1:1 ratio (e.g., by coincidence or in the background noise). Figure 3e illustrates the MS2 TIC chromatogram for the 0–24 h dog faeces sample, obtained by deploying isotopic-data-dependent scanning (using the following Xcalibur settings: mass difference 2, expected ratio 1, match tolerance 0.1). A quite selective bromatogram is obtained, almost exclusively composed of metabolite peaks as can be derived from the striking similarity between the qualitative ESI-MS2 chromatogram and the quantitative ICP-MS (Fig. 3a) and radiochromatogram (Fig. 3b). This allows accurate linkage of the structures characterized to the quantitative HPLC-ICP-MS or radio-HPLC data. The same holds for the MS2 TIC trace obtained for the 0–24 h rat faeces sample illustrated in Fig. 4e where the same settings were used as for dog. Whenever the dose level or the MS sensitivity of the compounds of interest is lower and/or the interference of background or endogenous ions is higher, the selectivity gain in the MS2 TIC chromatogram is insufficient to arrive at such selective bromatograms. This is illustrated in Fig. 5 where the (a) MS, (b) MS2, (c) MS3, and (d) MS4 TIC traces acquired on an LTQ ion trap with isotopic-data-dependent scanning for a 0–24 h rat plasma sample (no radioactive and ICP-MS data available) are shown. Whereas no compound-related peaks can be found in the MS TIC chromatogram, an obvious peak for the unchanged drug R207910 is apparent in the background of the MS2 trace but is still accompanied with a large number of background ions. For every additional MSn step, the number of false positive background ions decreases so that a more selective and cleaner chromatogram is obtained, finally resulting in a qualitative ICP-MS-like bromatogram in the MS4 trace where even the smallest bromine-containing metabolites are filtered out of the background. Essential in this approach is the selection of a broad isolation width so that both 79Br and 81Br isotope peaks are isolated and fragmented every next MSn step.

Both rat and dog faeces extracts were also analyzed on a Q-Tof Ultima mass spectrometer. Although a similar method can be deployed on such an instrument, such that online MS2 fragmentation is only triggered by isotope ratio recognition in the MS total ion current, a different and more effective approach was chosen to transform the rather unselective MS TIC trace into a very clean and selective bromatogram. The Q-Tof mass spectrometer operational software (Masslynx) provides different tools to process the data obtained. The software allows stripping of the data with a cluster analysis function, such that all bromine-containing ions can be filtered out of a data file. Figures 3 and 4 show the MS TIC chromatogram of the 0–24 h dog and rat faeces extract with (Figs. 3d and 4d) and without (Figs. 3c and 4c) stripping after analysis (using the following Masslynx settings: first mass difference 1.9979, first mass ratio 1, mass tolerance 0.01 amu, ratio tolerance 10%, threshold 0.8%). Again a selective ICP-MS-like, but qualitative, bromatogram is obtained, allowing linkage of the structural characterization with ESI-MS and the corresponding quantitative data gathered with ICP-MS. The advantage of this post-processing tool is that one can “play” with the different parameters, especially the threshold used, to discard the majority of the background noise in order to obtain the best result without the need for any re-analysis. Another benefit of this approach is that all mass spectra extracted from this newly generated chromatogram only contain those ions that show the bromine isotope ratio and so almost all background and endogenous ions are discarded.

The main metabolites identified in the 0–24 h dog and rat faeces samples are shown in Fig. 1. Chemical formulas for the metabolites were calculated from the accurate mass measurements acquired and illustrated in Table 4. It should be noted that the mass accuracy of the current Q-Tof instrumentation is better than the Q-Tof Ultima employed in this work. Structures were elucidated using the accurate mass measurements in combination with the MS2 (Q-Tof) and MSn (LTQ-Orbitrap) fragmentation obtained. The fragmentation observed on both systems is very similar. All major product ions detected for R207910 (Fig. 6) are identical on both MS systems, except for the ion at m/z 229 that is, due to its radical nature, only formed at the higher collision energies applied in the Q-Tof. Different LTQ-Orbitrap low energy CID MS2 and MS3 spectra have to be acquired to gather the same amount of structural information as in one higher energy CID Q-Tof MS2 spectrum. The Q-Tof mass spectrometer allows for rapid assessment of structural data, especially when a so-called all-in-one (MSE) approach is applied [22, 23], whereas the MSn fragmentation on the LTQ-Orbitrap often allows more in-depth structure assignment [24, 25]. The Q-Tof MS2 spectra of metabolites 1–4 are illustrated in Fig. 7. Similar information can be derived from the LTQ-Orbitrap MSn spectra (data not shown). All bromine-containing product ions are easily recognized in the different spectra by their typical isotope ratio since a broad isolation width was used, transferring both the 79Br and the 81Br isotope peak to be fragmented in MS2. Although most of the fragments have been characterized, we will only focus on the most diagnostic ions that allowed the structure elucidation of the metabolites. The protonated molecular ion of metabolite 1 (Fig. 7a) shows a shift of −28 amu relative to the drug substance R207910. A successive neutral loss of water (−18 amu) or methanol (−32 amu) and methanimine (–CH2=NH; −29 u) results in the fragments at m/z 466/468 (79Br/81Br) and 480/482 (79Br/81Br), respectively. The product ions at m/z 141 and 328/330 (79Br/81Br) are unchanged relative to the drug substance (Fig. 6a). Therefore, metabolite 1 is identified as the N-didesmethyl metabolite of R207910. Metabolite 2 (Fig. 7b) shows a shift of −14 u relative to the drug substance. The same shift can be noticed for the ions at m/z 214 (228 – 14) and 215 (229 – 14), while the ion at m/z 141 is unchanged. The fragments at m/z 466/468 (79Br/81Br) and 480/482 (79Br/81Br) originate from a successive neutral loss of water (−18 amu) or methanol (−32 amu) and N-methylenemethanamine (–CH2=N–CH3; −43 amu), respectively. Therefore, metabolite 2 is assigned as the N-desmethyl metabolite of R207910. Metabolites 3 and 4 are oxidation products of R207910 (Table 4). The position of the hydroxylation of metabolite 3 (Fig. 7c) can be derived from the product ion at m/z 344/346 (m/z 328/330 + 16) that has shifted 16 amu, while the ion at m/z 229 is unchanged relative to the drug substance fragmentation. The limited fragmentation observed in the Q-Tof MS2 spectrum of metabolite 3 is caused by the low intensity of the m/z 571 ion signal. The position of the hydroxylation on metabolite 4 (Fig. 7d) can be defined from the ions at m/z 157 (141 + 16), 244 (228 + 16), and 245 (229 + 16) that have shifted by 16 amu, while the ion at m/z 328/330 (79Br/81Br) is unchanged relative to the fragmentation seen for R207910.

Fig. 3
figure 3

a ICP-MS bromatogram, b radiochromatogram, c ESI-MS TIC (Q-Tof), d ESI-MS TIC after isotopic stripping (Q-Tof), and e ESI-MS2 TIC obtained with isotopic-data-dependent scanning (LTQ-Orbitrap) acquired for a 0–24 h dog methanolic faeces extract; numbers correspond to metabolites illustrated in Fig. 1

Full size image
Fig. 4
figure 4

a ICP-MS bromatogram, b radiochromatogram, c ESI-MS TIC (Q-Tof), d ESI-MS TIC after isotopic stripping (Q-Tof), and e ESI-MS2 TIC obtained with isotopic-data-dependent scanning (LTQ-Orbitrap) acquired for a 0–24 h rat methanolic faeces extract; numbers correspond to metabolites illustrated in Fig. 1

Full size image
Fig. 5
figure 5

a MS, b MS2, c MS3, and d MS4 TIC chromatograms obtained for a 0–24 h rat plasma sample. The sample was protein precipitated adding three volumes of acetonitrile, the supernatant was dried and reconstituted in DMSO prior to analysis

Full size image
Table 4 List of chemical formulas and accurate mass measurements obtained for R207910 and the identified metabolites
Full size table
Fig. 6
figure 6

a MS2 and b MSn spectra obtained for R207910 on a Q-Tof and LTQ-Orbitrap mass spectrometer, respectively. c Possible explanation for the fragmentation behavior of R207910

Full size image
Fig. 7
figure 7

Q-Tof MS2 spectra acquired for a metabolite 1 in rat faeces, b metabolite 2 in rat faeces, c metabolite 3 in dog faeces, and d metabolite 4 in rat faeces

Full size image

Conclusion

The combination of inductively coupled plasma mass spectrometry (ICP-MS) and electrospray mass spectrometry (ESI-MS) is shown to be a very powerful tool in the metabolite profiling and metabolite identification of bromine-containing compounds. Unlike ESI-MS, the intensity of the signal obtained with ICP-MS is independent of the chemical structure analyzed. Therefore, it can be used as an alternative for radiochemical detection for metabolite profiling and quantification without the need for a radiolabeled compound. Rat and dog faeces samples, containing the majority of the excreted dose, were analyzed for R207910 and its major metabolites. The bromine atom in R207910 allowed the selective detection and quantification of the drug substance and its metabolites by ICP-MS. Fluctuations in the Br signal caused by the HPLC gradient were counteracted by the use of species-unspecific isotope dilution. The results obtained were in very good accordance with those acquired with radioactivity detection.

Ideally the same chromatographic system is equipped with both ICP-MS and ESI-MS detection so that quantitative results and structure identification data are simultaneously acquired. However, these combined systems are not readily available and it was shown that qualitative bromatograms could also be obtained in ESI-MS by the use of the typical 79Br/81Br isotope ratio so that the right link with the quantitative results is ensured. Data-dependent scanning using isotope recognition on an ion trap or orbitrap mass spectrometer or isotope-triggered processing of Q-Tof data provided qualitative bromatograms that are very similar to those obtained with ICP-MS. The combination of accurate mass measurements and the fragmentation behavior in the MS2 spectra obtained on the Q-Tof Ultima mass spectrometer or MSn spectra acquired on the LTQ-Orbitrap allowed the structural elucidation of the metabolites.

The HPLC-ICP-MS methodology presented can be used for any bromine-containing compound examined in discovery or development. The HPLC-ESI-MS methodology presented can also be deployed for compounds with other specific isotope ratios, e.g., 35Cl/37Cl, 12C/14C, etc.