- Email: [email protected]
Identification and Measurement of Oxindole (2-Indolinone) in the Mammalian Brain and Other Rat Organs
ANALYTICAL BIOCHEMISTRY ARTICLE NO.
244, 74–79 (1997)
AB969854
Identification and Measurement of Oxindole (2-Indolinone) in the Mammalian Brain and Other Rat Organs Raffaella Carpenedo, Vincenzo Carla`, Gloriano Moneti, Alberto Chiarugi, and Flavio Moroni Department of Preclinical and Clinical Pharmacology, University of Florence, Viale Morgagni 65, 50134 Florence, Italy
Received July 22, 1996
Oxindole, a putative tryptophan metabolite able to cause profound sedation when administered in relatively low doses to mammals, has been identified and measured in the brains of mice, rats, and guinea pigs using HPLC and GC/MS with a quadrupole ion trap and a collision-induced dissociation mass spectrometer. The identification and measurement of the compound required a protein precipitation step with HClO4 , extraction into chloroform, an HPLC separation on a reverse-phase column, and detection by UV or coulometry. The definitive identification of the oxindole peak was obtained with a Saturn 4D GC/MS quadrupole ion trap operated under GC/MS, GC/MS/MS, and GC/MS/MS/MS modes. The HPLC methods we used had a low interassay variability, easily allowing the identification and measurement of the compound in 1 g of tissue. The oxindole concentrations in rat brain, blood, liver, and kidney were each approximately 100 pmol/g wt. Interestingly, the content of oxindole in the guinea pig brain was found to be significantly lower than that in the mouse and rat brains, possibly reflecting a lower dietary intake of tryptophan in the guinea pigs. q 1997 Academic Press, Inc.
Tryptophan metabolites such as 5OH-tryptamine have been studied for many years as classical neurotransmitters, while others such as quinolinic and kynurenic acids have more recently been described as molecules able to affect neuronal function by acting as either agonists or antagonists of the excitatory amino acid receptors (see (1, 2) for reviews). In the past several years our group has been particularly interested in understanding the role that these last metabolites may have in physiology or pathology (3–6), and we have been investigating the regulation of the activities of a number of enzymes responsible for quinolinate or kynurenate formation in the mammalian brain (7–9). In the course of these studies, in order to understand the metabolic pathways leading to the synthesis of the
excitotoxin quinolinic acid, we administered several possible precursors, including a few indole derivatives, to rats and mice, and we noted that administration of 5-hydroxyindole (20–200 mg/kg ip) caused convulsions, while similar doses of an isomer, oxindole (2-indolinone), had profound neurodepressant effects. Library research and a discussion with a chemist who has been involved in studies on tryptophan metabolism for several decades (Dr. Allegri, Padua, Italy) revealed that in the early sixties, oxindole had been administered to mice, rats, hamsters, rabbits, cats, dogs, and humans, and its potent neurodepressant action was observed and carefully described (10). The convulsant actions of 5-hydroxyindole were not previously described, but it has been recently demonstrated that this compound may modulate the function of 5HT3 receptors (11, 12). None of these studies, however, discussed the possibility that either oxindole or 5-hydroxyindole could be present in mammalian tissues because both indoles were usually considered by-products of the activities of bacterial enzymes (13, 14). We thought it important to clarify whether such neuroactive molecules are present in mammalian tissues. We describe a simple method which allows quantification of oxindole (the neurodepressant molecule) with HPLC separation and UV or EC detection in the blood, brain, and other organs of the mammals. The oxindole HPLC peak was then collected and injected in a quadrupole ion trap operating in GC/MS/MS/MS mode for its definitive identification. MATERIALS AND METHODS
Chemicals Oxindole, 4-hydroxyindole, and 5-hydroxyindole were obtained from Aldrich (Milan, Italy). Solvents (acetonitrile and chloroform) came from Merck (Darmstadt, Germany). Concentrated solutions of the chemicals were prepared by dissolving adequate amounts of standards in 0.4 N perchloric acid (oxindole) or distilled water (4-hydroxyindole and 5-hydroxyindole), kept at 0207C, and then diluted before each experiment.
74
AID
0003-2697/97 $25.00 Copyright q 1997 by Academic Press, Inc. All rights of reproduction in any form reserved.
AB 9854
/
6m24$$$181
11-26-96 01:42:10
abal
75
CHROMATOGRAPHIC ANALYSIS OF OXINDOLE TABLE 1
HPLC Retention Times and Peak Heights of Synthetic Oxindole and of Material Extracted from Rat Organs Mobile phase 1
Retention time Oxindole Brain Liver Blood
7 7 7 7
min min min min
0 0 0 0
s s s s
Peak height (cm) 8.5 2 7.1 7.6
Mobile phase 2
Mobile phase 3
Ratio
Retention time
Peak height (cm)
Ratio
1 0.23 0.83 0.89
9 min 0 s 9 min 0 s — —
10.2 2.5 — —
1 0.24 — —
Retention time 13 13 13 13
min min min min
25 25 25 25
s s s s
Peak height (cm)
Ratio
9 2.15 7.5 8
1 0.24 0.83 0.89
Note. Mobile phase 1 was 0.05 M acetate buffer, pH 3.2, plus 15% (v/v) acetonitrile. Mobile phase 2 was 0.5 N acetic acid plus 10% (v/v) acetonitrile. Mobile phase 3 was 0.2 M phosphate buffer, pH 4.1, plus 8% (v/v) acetonitrile. The flow rate of the mobile phases was 1.5 ml/ min.
Extraction and Purification of Oxindole Male Wistar rats (Harlan–Nossan, Correzzana, Italy) weighing 200–250 g, male white Swiss mice (Morini, San Polo D’Enza, Italy) weighing 20–25 g, and male guinea pigs (Rodentia, Torre Pallavicina, Italy) weighing 450–500 g were used. The animals were decapitated and the blood and other organs (brain, liver, and kidney) were rapidly removed, weighed, and frozen. Approximately 1 g of each tissue was then homogenized in 2 vol of 0.4 N HClO4 (or 1 N HClO4 for the blood). The mixture was then centrifuged at 18,000g for 20 min and the procedure was repeated twice. The supernatants collected by the two successive centrifugations were mixed with 8 ml of chloroform and agitated for 5 min. The chloroform layers were collected and evaporated under a stream of nitrogen. The residues were resuspended in 0.4 N HClO4 and aliquots of them were injected into the HPLC apparatus. The recovery of a known amount of oxindole which was passed through the entire procedure was 75 { 2% (mean { SE of 10 determinations). HPLC Methodology The HPLC apparatus consisted of a Perkin–Elmer LC pump (Model 250), a syringe loading sample injection valve (Rheodyne Model 7125), a C18 reverse-phase precolumn filter (0.5 cm long; Waters, Milford, MA), and a 25-cm reverse-phase 18 SpheriSorb ODS-2 10 U column (Alltech, Deerfield, IL). The detection was performed either with a spectrophotometer (Perkin– Elmer Model LC 90 UV) or with a ESA 5100 coulometric detector. The following mobile phases were used: (1) 0.05 M acetate buffer, pH 3.24, and 15% acetonitrile; (2) 0.5 M acetic acid and 10% acetonitrile; and (3) 0.2 M phosphate buffer, pH 4.01, and 8% acetonitrile.
peak, was dried down in a Savant speed vac concentrator centrifuge and then resuspended in acetonitrile (40 ml). A portion of it was injected into a Saturn 4D GC/MS quadrupole ion trap equipped with waveform generator (Varian, Walnut Creek, CA). The gas chromatograph was a Varian Star 3400 CX equipped with a temperature programmable injector (Varian 1078) which permits large volume injection. A fused-silica capillary column (Supelcowax 10, 30 m 1 0.25 mm i.d. 1 0.5 mm film thickness; Supelco, Bellefonte, PA) was used. Helium was the carrier gas at a head pressure of 12 psi and a linear flow velocity of 44 cm/s at 507C. Initial injector temperature was 957C. It was maintained for 1 min and then rapidly increased at a velocity of 1007C/ min until 2507C; the injector splitter valve was initially open (split mode), closed at 1 min (splitless mode), and reopened at 2.55 min. The initial oven temperature was 707C for 2.55 min and then rapidly increased at a rate of 357C/min until 2407C followed by a 16-min isothermal step. Transfer line and manifold were maintained at 240 and 1707C, respectively. The quadrupole ion trap mass spectrometer was operating in the scan mode in the range 50–140 m/z at 1 scan/s. The ‘‘Toolkit’’ parameters used for MS/MS experiments were parent ion m/z 133.05 with a collision-induced dissociation (CID)1 time 30 ms. In the MS/MS/ MS experiments parent and daughter ions were m/z 133.05 and 104.03 and were collided with a CID time of 5 ms. Other parameters used were CID amplitude, 0.6 V; isolation window, 1; CID type, resonant. RESULTS
HPLC Identification and Determination of Oxindole A first identification of oxindole was obtained by comparing the retention times of the synthetic compound with those of one of the peaks present when purified
GC/MS Methodology In selected experiments, 1.5 ml of HPLC eluate, manually collected at the retention time of the oxindole
AID
AB 9854
/
6m24$$$182
11-26-96 01:42:10
1 Abbreviations used: CID, collision-induced dissociation; TIC, total ion chromatogram.
abal
76
CARPENEDO ET AL.
/0.4 V for the first electrode and /0.8 V for the second one; mobile phase 2). The figure shows the presence of one peak in the biological material, detected with either UV or coulometry, which had a retention time and detection properties identical to those of the standard. In most of our determinations, we used UV detection and mobile phase 1 because the method is simpler and allows sufficient sensitivity for the determination of oxindole content in one rat brain. Standard curves of authentic oxindole, passed through the purification procedures, gave linear responses in the range of 4– 400 pmol per injection. The interassay variability was 3%; the quantitative detection limit of the method (signal to noise ratio ú3) was 2 pmol per injection (80 ml). Under the above-mentioned conditions, oxindole had a retention time of 7 min, while other isomers had the following retention times: 4-hydroxyindole, 5 min and 45 s; 5-hydroxyindole, 6 min and 30 s. GC/MS Identification of Oxindole
FIG. 1. HPLC analysis of synthetic oxindole (50 pmol/100 ml; traces A and B) and of material extracted from the rat brain and treated as described in the text (C and D). In A and C mobile phase 1 and an UV detector operated at 255 nm were used. In B and D mobile phase 2 and a coulometric detector were used (see text for details).
homogenates of mammalian organs were injected into the HPLC, and the UV absorbance at 255 nm was recorded. The retention times of one of the peaks found in the biological material and of the standard peak were identical, regardless of the composition of the mobile phases utilized (see Table 1). Spiking experiments were performed using each of the three mobile phases reported in the table, confirming that the peaks present in both the homogenates and the standard had identical chromatographic properties. Table 1 also indicates that the ratio between the peak height of synthetic oxindole and that of the material present in brain, blood, liver, and kidney did not change when different mobile phases were used. Similar results were also obtained when a coulometric approach was used as a detector system. Figure 1 shows the chromatographic pattern obtained by using the UV (at 255 nm; mobile phase 1) and the coulometric detector (oxidation potentials
AID
AB 9854
/
6m24$$$182
11-26-96 01:42:10
Figure 2A shows the proposed fragmentation pattern of authentic oxindole (see Porter (15)) and Fig. 2B shows the EI positive-ion mass spectrum we obtained in full scan acquisition mode. The mass spectrum shown in Fig. 2C was obtained by recording in MS/MS condition mode after CID on isolated m/z 133 ion; Fig. 2D shows the mass spectrum obtained in MS/MS/MS mode with CID first on m/z 133 and then on m/z 104. When the collected HPLC eluate fractions from rat brain homogenates were dried down, resuspended in acetonitrile, and injected into the GC/MS, the presence of the oxindole peak was confirmed. Figure 3A shows the total ion chromatogram (TIC) and the extracted ion chromatogram of m/z 133 obtained from rat brain homogenates: a peak at the same gas chromatographic retention time of authentic oxindole was recorded on the m/z 133 trace. The corresponding EI mass spectrum was identical to that of oxindole (Figs. 2B and 3A). It is interesting to observe the reduction in chemical noise due to the matrix under the different analytical conditions. Although in the TIC recording trace of the full scan MS acquisition (Fig. 3A) a peak at the retention time of oxindole is not clearly evident, in the MS/MS trace (Fig. 3C) and in the MS/MS/MS traces (Fig. 3E) a distinct peak appears. Comparing the mass spectra obtained in the three different acquisition modes with those of authentic oxindole, we confirmed the presence of this molecule in the brain homogenates. In a further series of experiments, we compared the peak areas of a standard passed through the entire procedure with that of brain homogenates in order to have semiquantitative confirmation of the data obtained in the HPLC. The results are in line with the quantitative data we report in Table 2. However, the variability of the biochemical and chromatographic procedures does not allow precise quantitative mea-
abal
CHROMATOGRAPHIC ANALYSIS OF OXINDOLE
77
FIG. 2. The proposed fragmentation pattern of authentic oxindole (A) and the EI positive ion mass spectrum (background subtracted) recorded using a Saturn 4D GC/MS quadrupole ion trap in a full scan acquisition mode (B). In C the mass spectrum was recorded under MS/MS conditions after collision-induced dissociation (CID) on the ion m/z 133, and in D the mass spectrum was recorded under MS/MS/ MS conditions and CID first on the ion m/z 133 and then on the ion m/z 104.
surements of oxindole using the GC/MS/MS/MS approach without suitable internal standards. Oxindole Content in the Mammalian Brain and Other Rat Organs As reported in Table 2, the HPLC method using UV detection allowed the quantitative determination of oxindole in several rat organs and in the brains of other commonly used laboratory animals. In the rat brain, the concentrations of oxindole were similar to those in the blood, while larger concentrations were found in the liver and the kidney. Mouse brain contained relatively elevated concentrations of oxindole, while guinea pig had brain oxindole concentrations one order of magnitude lower than those of the rat. DISCUSSION
Oxindole, a possible tryptophan metabolite with strong sedative action, has been identified and measured in mammalian tissues by the use of a simple extraction, HPLC separation, and UV detection. The putative oxindole peak was first identified by utilizing the retention times using different mobile phases and
AID
AB 9854
/
6m24$$$182
11-26-96 01:42:10
standard spiking procedures. We then used a coulometric detector operated at different oxidation potentials and obtained identical results. Finally, we used a GC/ MS apparatus which allows identification of a compound not only on the basis of the gas chromatographic retention time and the mass spectrum, but also on the basis of the fragmentation pattern of parent ions for two further generations (GC/MS/MS/MS). This approach was used for selected samples in which the eluate corresponding to the peak of oxindole in the HPLC– UV procedure was collected, dried, resuspended in acetonitrile, and injected into the GC/MS. The results of these experiments allowed a nonequivocal identification of oxindole. The GC/MS, GC/MS/MS, and GC/MS/ MS/MS approaches we used to identify oxindole offer interesting indications of the versatility of this method for analysis of complex biological material. Both MS/ MS and MS/MS/MS modes achieved a higher signalto-noise ratio than the full scan mode of the GC/MS. The very low background signals in the reconstructed daughter and granddaughter ion chromatograms (Figs. 3B and 3C) are indicative of a very high methodology selectivity. These observations confirm the hypothesis put forth by Bush and Cooks several years ago (16).
abal
78
CARPENEDO ET AL.
FIG. 3. A, C, and E are total ion chromatograms (above tracing) and extracted ion chromatograms (bottom tracing) obtained by injecting the biological sample (extracted and purified as described in the text) into a Saturn 4D quadrupole ion trap and recorded in GC/MS, GC/ MS/MS, and GC/MS/MS/MS mode, respectively (see text for details). B, D, and F are the mass spectra corresponding to the peak indicated by the arrows.
The concentrations of oxindole in the liver and kidney were approximately twice those in the brain and measured approximately 1007 M. Several other functionally active tryptophan metabolites such as trypt-
TABLE 2
Content of oxindole in mammalian brain Mice Rats Guinea pigs
90 { 8.0 54 { 4.7 4.5 { 0.3
Content of oxindole in other rat tissues Liver Kidney Blood
113 110 78
{ 8.2 { 13 { 8.7
Note. Values are pmol/g wt or pmol/ml and are means { SE of at least 10 determinations. These values were obtained using HPLC with UV detection.
AID
AB 9854
/
6m24$$$182
11-26-96 01:42:10
amine, kynurenic acid, and quinolinic acid are present in mammalian tissue in this range of concentrations. Since these compounds became electrophysiologically active in vitro in a concentration range of 1005 –1004, it is not clear whether they play a functional role in brain physiology. Interestingly, the concentrations of oxindole in the brains of mice and rats were significantly higher than those in the guinea pig. It is not unreasonable to ascribe these differences to tryptophan content in the respective diets: the guinea pigs are herbivores, while rat and mice are omnivores. In conclusion, we report a simple method which may be used to evaluate the content of oxindole, another active tryptophan metabolite present in the mammalian tissues. This method has been substantiated by a gas chromatographic–mass spectrometric identification of the compound and may be useful in studying the neosynthesis and the origin of oxindole in mammalian tissues. Finally, considering that an increase in the oxindole level in blood may have profound sedative ef-
abal
CHROMATOGRAPHIC ANALYSIS OF OXINDOLE
fects leading to a true coma, it may be useful to monitor its concentration in human blood or other biological fluids in several pathological conditions characterized by loss of consciousness. ACKNOWLEDGMENTS This work was supported by the University of Florence (MURST Funds), by CNR, and by the EU (Biomed 1 CT93-1033). We thank Dr. G. Pieraccini and the Mass-Spectrometry Centre of the University of Florence for excellent technical support.
REFERENCES 1. Zifa, E., and Fillion, G. (1992) Pharmacol. Rev. 44, 401–458. 2. Stone, T. W. (1993) Pharmacol. Rev. 45, 309–379. 3. Moroni, F., Lombardi, G., Carla`, V., and Moneti, G. (1984) Brain Res. 295, 352–355. 4. Moroni, F., Lombardi, G., Carla`, V., Pellegrini, D., Carassale, G. L., and Cortesini, C. (1986) J. Neurochem. 46, 869–874. 5. Moroni, F., Russi, P., Lombardi, G., Beni, M., and Carla`, V. (1988) J. Neurochem. 51, 177–181.
AID
AB 9854
/
6m24$$$183
11-26-96 01:42:10
79
6. Moroni, F., Russi, P., Gallo-Mezo, M. A., Moneti, G., and Pellicciari, R. (1991) J. Neurochem. 57, 1630–1635. 7. Carpenedo, R., Chiarugi, A., Russi, P., Lombardi, G., Carla`, V., Pellicciari, R., Mattoli, L., and Moroni, F. (1994) Neuroscience 61, 237–244. 8. Chiarugi, A., Carpenedo, R., Molina, M. T., Mattoli, L., Pellicciari, R., and Moroni, F. (1995) J. Neurochem. 65, 1176–1183. 9. Chiarugi, A., Carpenedo, R., and Moroni, F. (1995) Soc. Neurosci. Abstr. 21, 1587. [Abstract] 10. Orcutt, J. A., Prytherch, J. P., Konicov, M., and Michaelson, S. M. (1964) Arch. Int. Pharmacodyn. 152, 121–131. 11. Kooyman, R. A., van Hooft, J. A., and Vijverberg, H. P. M. (1993) Br. J. Pharmacol. 108, 287–289. 12. Kooyman, R. A., van Hooft, J. A., Vanderheijden, P. M. L., and Vijverberg, H. P. M. (1994) Br. J. Pharmacol. 112, 541–546. 13. van Pe´e, K. H., and Lingens, F. (1984) in Progress in Tryptophan and Serotonin Research (Schlossberger, H. G., Kochen, W., Linzen, B., and Steinhart, H. Eds.), pp. 753–760, de Gruyter, Berlin. 14. King, L. J., Parke, D. V., and Williams, R. T. (1966) Biochem. J. 98, 266–277. 15. Porter, Q. N. (1985) Mass-Spectrometry of Heterocyclic Compounds, Wiley, New York. 16. Busch, K. L., and Cooks, G. (1983) in ‘‘Tandem Mass Spectrometry’’ (F. McLafferty, Ed.), pp. 11–39, Wiley, New York.
abal
244, 74–79 (1997)
AB969854
Identification and Measurement of Oxindole (2-Indolinone) in the Mammalian Brain and Other Rat Organs Raffaella Carpenedo, Vincenzo Carla`, Gloriano Moneti, Alberto Chiarugi, and Flavio Moroni Department of Preclinical and Clinical Pharmacology, University of Florence, Viale Morgagni 65, 50134 Florence, Italy
Received July 22, 1996
Oxindole, a putative tryptophan metabolite able to cause profound sedation when administered in relatively low doses to mammals, has been identified and measured in the brains of mice, rats, and guinea pigs using HPLC and GC/MS with a quadrupole ion trap and a collision-induced dissociation mass spectrometer. The identification and measurement of the compound required a protein precipitation step with HClO4 , extraction into chloroform, an HPLC separation on a reverse-phase column, and detection by UV or coulometry. The definitive identification of the oxindole peak was obtained with a Saturn 4D GC/MS quadrupole ion trap operated under GC/MS, GC/MS/MS, and GC/MS/MS/MS modes. The HPLC methods we used had a low interassay variability, easily allowing the identification and measurement of the compound in 1 g of tissue. The oxindole concentrations in rat brain, blood, liver, and kidney were each approximately 100 pmol/g wt. Interestingly, the content of oxindole in the guinea pig brain was found to be significantly lower than that in the mouse and rat brains, possibly reflecting a lower dietary intake of tryptophan in the guinea pigs. q 1997 Academic Press, Inc.
Tryptophan metabolites such as 5OH-tryptamine have been studied for many years as classical neurotransmitters, while others such as quinolinic and kynurenic acids have more recently been described as molecules able to affect neuronal function by acting as either agonists or antagonists of the excitatory amino acid receptors (see (1, 2) for reviews). In the past several years our group has been particularly interested in understanding the role that these last metabolites may have in physiology or pathology (3–6), and we have been investigating the regulation of the activities of a number of enzymes responsible for quinolinate or kynurenate formation in the mammalian brain (7–9). In the course of these studies, in order to understand the metabolic pathways leading to the synthesis of the
excitotoxin quinolinic acid, we administered several possible precursors, including a few indole derivatives, to rats and mice, and we noted that administration of 5-hydroxyindole (20–200 mg/kg ip) caused convulsions, while similar doses of an isomer, oxindole (2-indolinone), had profound neurodepressant effects. Library research and a discussion with a chemist who has been involved in studies on tryptophan metabolism for several decades (Dr. Allegri, Padua, Italy) revealed that in the early sixties, oxindole had been administered to mice, rats, hamsters, rabbits, cats, dogs, and humans, and its potent neurodepressant action was observed and carefully described (10). The convulsant actions of 5-hydroxyindole were not previously described, but it has been recently demonstrated that this compound may modulate the function of 5HT3 receptors (11, 12). None of these studies, however, discussed the possibility that either oxindole or 5-hydroxyindole could be present in mammalian tissues because both indoles were usually considered by-products of the activities of bacterial enzymes (13, 14). We thought it important to clarify whether such neuroactive molecules are present in mammalian tissues. We describe a simple method which allows quantification of oxindole (the neurodepressant molecule) with HPLC separation and UV or EC detection in the blood, brain, and other organs of the mammals. The oxindole HPLC peak was then collected and injected in a quadrupole ion trap operating in GC/MS/MS/MS mode for its definitive identification. MATERIALS AND METHODS
Chemicals Oxindole, 4-hydroxyindole, and 5-hydroxyindole were obtained from Aldrich (Milan, Italy). Solvents (acetonitrile and chloroform) came from Merck (Darmstadt, Germany). Concentrated solutions of the chemicals were prepared by dissolving adequate amounts of standards in 0.4 N perchloric acid (oxindole) or distilled water (4-hydroxyindole and 5-hydroxyindole), kept at 0207C, and then diluted before each experiment.
74
AID
0003-2697/97 $25.00 Copyright q 1997 by Academic Press, Inc. All rights of reproduction in any form reserved.
AB 9854
/
6m24$$$181
11-26-96 01:42:10
abal
75
CHROMATOGRAPHIC ANALYSIS OF OXINDOLE TABLE 1
HPLC Retention Times and Peak Heights of Synthetic Oxindole and of Material Extracted from Rat Organs Mobile phase 1
Retention time Oxindole Brain Liver Blood
7 7 7 7
min min min min
0 0 0 0
s s s s
Peak height (cm) 8.5 2 7.1 7.6
Mobile phase 2
Mobile phase 3
Ratio
Retention time
Peak height (cm)
Ratio
1 0.23 0.83 0.89
9 min 0 s 9 min 0 s — —
10.2 2.5 — —
1 0.24 — —
Retention time 13 13 13 13
min min min min
25 25 25 25
s s s s
Peak height (cm)
Ratio
9 2.15 7.5 8
1 0.24 0.83 0.89
Note. Mobile phase 1 was 0.05 M acetate buffer, pH 3.2, plus 15% (v/v) acetonitrile. Mobile phase 2 was 0.5 N acetic acid plus 10% (v/v) acetonitrile. Mobile phase 3 was 0.2 M phosphate buffer, pH 4.1, plus 8% (v/v) acetonitrile. The flow rate of the mobile phases was 1.5 ml/ min.
Extraction and Purification of Oxindole Male Wistar rats (Harlan–Nossan, Correzzana, Italy) weighing 200–250 g, male white Swiss mice (Morini, San Polo D’Enza, Italy) weighing 20–25 g, and male guinea pigs (Rodentia, Torre Pallavicina, Italy) weighing 450–500 g were used. The animals were decapitated and the blood and other organs (brain, liver, and kidney) were rapidly removed, weighed, and frozen. Approximately 1 g of each tissue was then homogenized in 2 vol of 0.4 N HClO4 (or 1 N HClO4 for the blood). The mixture was then centrifuged at 18,000g for 20 min and the procedure was repeated twice. The supernatants collected by the two successive centrifugations were mixed with 8 ml of chloroform and agitated for 5 min. The chloroform layers were collected and evaporated under a stream of nitrogen. The residues were resuspended in 0.4 N HClO4 and aliquots of them were injected into the HPLC apparatus. The recovery of a known amount of oxindole which was passed through the entire procedure was 75 { 2% (mean { SE of 10 determinations). HPLC Methodology The HPLC apparatus consisted of a Perkin–Elmer LC pump (Model 250), a syringe loading sample injection valve (Rheodyne Model 7125), a C18 reverse-phase precolumn filter (0.5 cm long; Waters, Milford, MA), and a 25-cm reverse-phase 18 SpheriSorb ODS-2 10 U column (Alltech, Deerfield, IL). The detection was performed either with a spectrophotometer (Perkin– Elmer Model LC 90 UV) or with a ESA 5100 coulometric detector. The following mobile phases were used: (1) 0.05 M acetate buffer, pH 3.24, and 15% acetonitrile; (2) 0.5 M acetic acid and 10% acetonitrile; and (3) 0.2 M phosphate buffer, pH 4.01, and 8% acetonitrile.
peak, was dried down in a Savant speed vac concentrator centrifuge and then resuspended in acetonitrile (40 ml). A portion of it was injected into a Saturn 4D GC/MS quadrupole ion trap equipped with waveform generator (Varian, Walnut Creek, CA). The gas chromatograph was a Varian Star 3400 CX equipped with a temperature programmable injector (Varian 1078) which permits large volume injection. A fused-silica capillary column (Supelcowax 10, 30 m 1 0.25 mm i.d. 1 0.5 mm film thickness; Supelco, Bellefonte, PA) was used. Helium was the carrier gas at a head pressure of 12 psi and a linear flow velocity of 44 cm/s at 507C. Initial injector temperature was 957C. It was maintained for 1 min and then rapidly increased at a velocity of 1007C/ min until 2507C; the injector splitter valve was initially open (split mode), closed at 1 min (splitless mode), and reopened at 2.55 min. The initial oven temperature was 707C for 2.55 min and then rapidly increased at a rate of 357C/min until 2407C followed by a 16-min isothermal step. Transfer line and manifold were maintained at 240 and 1707C, respectively. The quadrupole ion trap mass spectrometer was operating in the scan mode in the range 50–140 m/z at 1 scan/s. The ‘‘Toolkit’’ parameters used for MS/MS experiments were parent ion m/z 133.05 with a collision-induced dissociation (CID)1 time 30 ms. In the MS/MS/ MS experiments parent and daughter ions were m/z 133.05 and 104.03 and were collided with a CID time of 5 ms. Other parameters used were CID amplitude, 0.6 V; isolation window, 1; CID type, resonant. RESULTS
HPLC Identification and Determination of Oxindole A first identification of oxindole was obtained by comparing the retention times of the synthetic compound with those of one of the peaks present when purified
GC/MS Methodology In selected experiments, 1.5 ml of HPLC eluate, manually collected at the retention time of the oxindole
AID
AB 9854
/
6m24$$$182
11-26-96 01:42:10
1 Abbreviations used: CID, collision-induced dissociation; TIC, total ion chromatogram.
abal
76
CARPENEDO ET AL.
/0.4 V for the first electrode and /0.8 V for the second one; mobile phase 2). The figure shows the presence of one peak in the biological material, detected with either UV or coulometry, which had a retention time and detection properties identical to those of the standard. In most of our determinations, we used UV detection and mobile phase 1 because the method is simpler and allows sufficient sensitivity for the determination of oxindole content in one rat brain. Standard curves of authentic oxindole, passed through the purification procedures, gave linear responses in the range of 4– 400 pmol per injection. The interassay variability was 3%; the quantitative detection limit of the method (signal to noise ratio ú3) was 2 pmol per injection (80 ml). Under the above-mentioned conditions, oxindole had a retention time of 7 min, while other isomers had the following retention times: 4-hydroxyindole, 5 min and 45 s; 5-hydroxyindole, 6 min and 30 s. GC/MS Identification of Oxindole
FIG. 1. HPLC analysis of synthetic oxindole (50 pmol/100 ml; traces A and B) and of material extracted from the rat brain and treated as described in the text (C and D). In A and C mobile phase 1 and an UV detector operated at 255 nm were used. In B and D mobile phase 2 and a coulometric detector were used (see text for details).
homogenates of mammalian organs were injected into the HPLC, and the UV absorbance at 255 nm was recorded. The retention times of one of the peaks found in the biological material and of the standard peak were identical, regardless of the composition of the mobile phases utilized (see Table 1). Spiking experiments were performed using each of the three mobile phases reported in the table, confirming that the peaks present in both the homogenates and the standard had identical chromatographic properties. Table 1 also indicates that the ratio between the peak height of synthetic oxindole and that of the material present in brain, blood, liver, and kidney did not change when different mobile phases were used. Similar results were also obtained when a coulometric approach was used as a detector system. Figure 1 shows the chromatographic pattern obtained by using the UV (at 255 nm; mobile phase 1) and the coulometric detector (oxidation potentials
AID
AB 9854
/
6m24$$$182
11-26-96 01:42:10
Figure 2A shows the proposed fragmentation pattern of authentic oxindole (see Porter (15)) and Fig. 2B shows the EI positive-ion mass spectrum we obtained in full scan acquisition mode. The mass spectrum shown in Fig. 2C was obtained by recording in MS/MS condition mode after CID on isolated m/z 133 ion; Fig. 2D shows the mass spectrum obtained in MS/MS/MS mode with CID first on m/z 133 and then on m/z 104. When the collected HPLC eluate fractions from rat brain homogenates were dried down, resuspended in acetonitrile, and injected into the GC/MS, the presence of the oxindole peak was confirmed. Figure 3A shows the total ion chromatogram (TIC) and the extracted ion chromatogram of m/z 133 obtained from rat brain homogenates: a peak at the same gas chromatographic retention time of authentic oxindole was recorded on the m/z 133 trace. The corresponding EI mass spectrum was identical to that of oxindole (Figs. 2B and 3A). It is interesting to observe the reduction in chemical noise due to the matrix under the different analytical conditions. Although in the TIC recording trace of the full scan MS acquisition (Fig. 3A) a peak at the retention time of oxindole is not clearly evident, in the MS/MS trace (Fig. 3C) and in the MS/MS/MS traces (Fig. 3E) a distinct peak appears. Comparing the mass spectra obtained in the three different acquisition modes with those of authentic oxindole, we confirmed the presence of this molecule in the brain homogenates. In a further series of experiments, we compared the peak areas of a standard passed through the entire procedure with that of brain homogenates in order to have semiquantitative confirmation of the data obtained in the HPLC. The results are in line with the quantitative data we report in Table 2. However, the variability of the biochemical and chromatographic procedures does not allow precise quantitative mea-
abal
CHROMATOGRAPHIC ANALYSIS OF OXINDOLE
77
FIG. 2. The proposed fragmentation pattern of authentic oxindole (A) and the EI positive ion mass spectrum (background subtracted) recorded using a Saturn 4D GC/MS quadrupole ion trap in a full scan acquisition mode (B). In C the mass spectrum was recorded under MS/MS conditions after collision-induced dissociation (CID) on the ion m/z 133, and in D the mass spectrum was recorded under MS/MS/ MS conditions and CID first on the ion m/z 133 and then on the ion m/z 104.
surements of oxindole using the GC/MS/MS/MS approach without suitable internal standards. Oxindole Content in the Mammalian Brain and Other Rat Organs As reported in Table 2, the HPLC method using UV detection allowed the quantitative determination of oxindole in several rat organs and in the brains of other commonly used laboratory animals. In the rat brain, the concentrations of oxindole were similar to those in the blood, while larger concentrations were found in the liver and the kidney. Mouse brain contained relatively elevated concentrations of oxindole, while guinea pig had brain oxindole concentrations one order of magnitude lower than those of the rat. DISCUSSION
Oxindole, a possible tryptophan metabolite with strong sedative action, has been identified and measured in mammalian tissues by the use of a simple extraction, HPLC separation, and UV detection. The putative oxindole peak was first identified by utilizing the retention times using different mobile phases and
AID
AB 9854
/
6m24$$$182
11-26-96 01:42:10
standard spiking procedures. We then used a coulometric detector operated at different oxidation potentials and obtained identical results. Finally, we used a GC/ MS apparatus which allows identification of a compound not only on the basis of the gas chromatographic retention time and the mass spectrum, but also on the basis of the fragmentation pattern of parent ions for two further generations (GC/MS/MS/MS). This approach was used for selected samples in which the eluate corresponding to the peak of oxindole in the HPLC– UV procedure was collected, dried, resuspended in acetonitrile, and injected into the GC/MS. The results of these experiments allowed a nonequivocal identification of oxindole. The GC/MS, GC/MS/MS, and GC/MS/ MS/MS approaches we used to identify oxindole offer interesting indications of the versatility of this method for analysis of complex biological material. Both MS/ MS and MS/MS/MS modes achieved a higher signalto-noise ratio than the full scan mode of the GC/MS. The very low background signals in the reconstructed daughter and granddaughter ion chromatograms (Figs. 3B and 3C) are indicative of a very high methodology selectivity. These observations confirm the hypothesis put forth by Bush and Cooks several years ago (16).
abal
78
CARPENEDO ET AL.
FIG. 3. A, C, and E are total ion chromatograms (above tracing) and extracted ion chromatograms (bottom tracing) obtained by injecting the biological sample (extracted and purified as described in the text) into a Saturn 4D quadrupole ion trap and recorded in GC/MS, GC/ MS/MS, and GC/MS/MS/MS mode, respectively (see text for details). B, D, and F are the mass spectra corresponding to the peak indicated by the arrows.
The concentrations of oxindole in the liver and kidney were approximately twice those in the brain and measured approximately 1007 M. Several other functionally active tryptophan metabolites such as trypt-
TABLE 2
Content of oxindole in mammalian brain Mice Rats Guinea pigs
90 { 8.0 54 { 4.7 4.5 { 0.3
Content of oxindole in other rat tissues Liver Kidney Blood
113 110 78
{ 8.2 { 13 { 8.7
Note. Values are pmol/g wt or pmol/ml and are means { SE of at least 10 determinations. These values were obtained using HPLC with UV detection.
AID
AB 9854
/
6m24$$$182
11-26-96 01:42:10
amine, kynurenic acid, and quinolinic acid are present in mammalian tissue in this range of concentrations. Since these compounds became electrophysiologically active in vitro in a concentration range of 1005 –1004, it is not clear whether they play a functional role in brain physiology. Interestingly, the concentrations of oxindole in the brains of mice and rats were significantly higher than those in the guinea pig. It is not unreasonable to ascribe these differences to tryptophan content in the respective diets: the guinea pigs are herbivores, while rat and mice are omnivores. In conclusion, we report a simple method which may be used to evaluate the content of oxindole, another active tryptophan metabolite present in the mammalian tissues. This method has been substantiated by a gas chromatographic–mass spectrometric identification of the compound and may be useful in studying the neosynthesis and the origin of oxindole in mammalian tissues. Finally, considering that an increase in the oxindole level in blood may have profound sedative ef-
abal
CHROMATOGRAPHIC ANALYSIS OF OXINDOLE
fects leading to a true coma, it may be useful to monitor its concentration in human blood or other biological fluids in several pathological conditions characterized by loss of consciousness. ACKNOWLEDGMENTS This work was supported by the University of Florence (MURST Funds), by CNR, and by the EU (Biomed 1 CT93-1033). We thank Dr. G. Pieraccini and the Mass-Spectrometry Centre of the University of Florence for excellent technical support.
REFERENCES 1. Zifa, E., and Fillion, G. (1992) Pharmacol. Rev. 44, 401–458. 2. Stone, T. W. (1993) Pharmacol. Rev. 45, 309–379. 3. Moroni, F., Lombardi, G., Carla`, V., and Moneti, G. (1984) Brain Res. 295, 352–355. 4. Moroni, F., Lombardi, G., Carla`, V., Pellegrini, D., Carassale, G. L., and Cortesini, C. (1986) J. Neurochem. 46, 869–874. 5. Moroni, F., Russi, P., Lombardi, G., Beni, M., and Carla`, V. (1988) J. Neurochem. 51, 177–181.
AID
AB 9854
/
6m24$$$183
11-26-96 01:42:10
79
6. Moroni, F., Russi, P., Gallo-Mezo, M. A., Moneti, G., and Pellicciari, R. (1991) J. Neurochem. 57, 1630–1635. 7. Carpenedo, R., Chiarugi, A., Russi, P., Lombardi, G., Carla`, V., Pellicciari, R., Mattoli, L., and Moroni, F. (1994) Neuroscience 61, 237–244. 8. Chiarugi, A., Carpenedo, R., Molina, M. T., Mattoli, L., Pellicciari, R., and Moroni, F. (1995) J. Neurochem. 65, 1176–1183. 9. Chiarugi, A., Carpenedo, R., and Moroni, F. (1995) Soc. Neurosci. Abstr. 21, 1587. [Abstract] 10. Orcutt, J. A., Prytherch, J. P., Konicov, M., and Michaelson, S. M. (1964) Arch. Int. Pharmacodyn. 152, 121–131. 11. Kooyman, R. A., van Hooft, J. A., and Vijverberg, H. P. M. (1993) Br. J. Pharmacol. 108, 287–289. 12. Kooyman, R. A., van Hooft, J. A., Vanderheijden, P. M. L., and Vijverberg, H. P. M. (1994) Br. J. Pharmacol. 112, 541–546. 13. van Pe´e, K. H., and Lingens, F. (1984) in Progress in Tryptophan and Serotonin Research (Schlossberger, H. G., Kochen, W., Linzen, B., and Steinhart, H. Eds.), pp. 753–760, de Gruyter, Berlin. 14. King, L. J., Parke, D. V., and Williams, R. T. (1966) Biochem. J. 98, 266–277. 15. Porter, Q. N. (1985) Mass-Spectrometry of Heterocyclic Compounds, Wiley, New York. 16. Busch, K. L., and Cooks, G. (1983) in ‘‘Tandem Mass Spectrometry’’ (F. McLafferty, Ed.), pp. 11–39, Wiley, New York.
abal