Determination of Boron-Containing Compounds in Urine and Blood

BNCT patients were infused with p-boronophenylalanine (BPA)−fructose complex before the therapy started. Urine and blood plasma samples were collect...
0 downloads 0 Views 78KB Size
Anal. Chem. 2002, 74, 5358-5363

Determination of Boron-Containing Compounds in Urine and Blood Plasma from Boron Neutron Capture Therapy Patients. The Importance of Using Coupled Techniques Eva Svantesson, Jacek Capala,†,‡ Karin E. Markides, and Jean Pettersson*

Department of Analytical Chemistry, Uppsala University, Box 531, SE-751 21 Uppsala, Sweden

The necessity of using coupled techniques to analyze samples from boron neutron capture therapy (BNCT) patients prior to element-specific detection has been demonstrated. BNCT patients were infused with p-boronophenylalanine (BPA)-fructose complex before the therapy started. Urine and blood plasma samples were collected at different times after the start of the BPA administration and were run on a porous graphitic carbon column coupled on-line to an inductively coupled plasmaatomic emission spectrometer (ICP-AES) and an ICP time-of-flight mass spectrometer (TOF-MS). In addition to BPA, a possible metabolite to BPA and some minor boron-containing compounds, eluting close to the front, were also found in the urine and plasma samples. Because only the total concentration of boron has been measured so far in earlier studies, the suspected metabolite could not be detected, and this is the first report indicating its presence in urine and plasma of BNCT patients. The abundance of 10B in urine was about the same for BPA and its possible metabolite (98-99%). The ratio between the possible metabolite and BPA was found to differ in the urine from different patients. Most of the patients had a metabolite concentration of ∼10 mol % of the BPA content in their urine 5-11 h after the start of the BPA administration. This ratio increased to between 30 and 80% when 24 h had passed. The ratio of metabolite to BPA was found to be lower in the plasma than in the urine samples at comparable time after the start of BPA infusion. Preliminary results from micro-LC-electrospray ionization (ESI)-MS/MS measurements on four urine samples indicate that the metabolite has a higher mass than BPA. Clinical trials of boron neutron capture therapy (BNCT)-applied treatment of brain tumors have been carried out in Japan, Europe, and the United States.1-6 In BNCT, a boron-containing compound, for example, p-boronophenylalanine (p-BPA), Figure 1a, is given intravenously to the patients and is accumulated preferably in the * Corresponding author. Fax: +46 (0)18 471 36 92. E-mail: jean.pettersson@ kemi.uu.se. † Unit of Biomedical Radiation Sciences, Rudbeck Laboratory, Uppsala University. Box 535, SE-751 85 Uppsala, Sweden. ‡ Studsvik Medical, SE-611 82 Nyko ¨ping, Sweden.

5358 Analytical Chemistry, Vol. 74, No. 20, October 15, 2002

Figure 1. (a) p-Boronophenylalanine (p-BPA), (b) phenylboric acid (PB), (c) tyramine analogue of p-BPA, (d) proposed structure for the BPA fragment of m/z 163 (M + H+), and (e, f) suggested structures for the 251 mass fragment of BPA-X (M + H+).

fast-growing tumor cells. The irradiation of the tumor-invaded tissues with neutrons causes the 10B isotope to split into an R particle and a lithium ion, which deposit a relatively large amount of energy (2.3-2.8 MeV) within a range of about one cell diameter. The cancer cells are thus damaged, while the surrounding tissue is kept intact. Since only 10B (comprising 20% of elemental boron in nature) has the large cross-section for neutron capture reaction (1) Chanana, A. D.; Capala, J.; Chadha, M.; Coderre, J. A.; Diaz, A. Z.; Elowitz, E. H.; Iwai, J.; Joel, D. D.; Liu, H. B.; Ma, R. M.; Pendzick, N.; Peress, N. S.; Shady, M. S.; Slatkin, D. N.; Tyson, G. W.; Wielopolski, L. Neurosurgery 1999, 44, 1182-1193. (2) Capala, J.; Henriksson, R.; Stensam, B. H.; Salford, L.; Carlsson, J.; Sko ¨ld, K. Proceedings of the Ninth International Symposium on Neutron Capture Therapy for Cancer, Osaka, Japan, October 2000. (3) Kankaanranta, L.; Seppa¨la¨, T.; Kallio, M.; Karila, J.; Aschan, C.; Sere´n, T.; Kortesniemi, M.; Kotiluoto, P.; Ja¨rviluoma, E.; Kulvik, M.; Laakso, J.; Brander, A.; Ja¨a¨skela¨inen, J.; Va¨ha¨talo, J.; Auterinen, I.; Savolainen, S.; Fa¨rkkila¨, M.; Joensuu, H. Proceedings of the Ninth International Symposium on Neutron Capture Therapy for Cancer, Osaka, Japan, October 2000. (4) Busse, P. M.; Zamenhof, R. G.; Harling, O. K.; Kaplan, I.; Kaplan, J.; Chuang, C. F.; Goorley, J. T.; Kiger, W. S., III; Riley, K. J.; Tang, L.; Solares, G. R.; Palmer, M. R. Frontiers in Neutron Capture Therapy; Hawthorne, M. F., Shelly, K., Wiersema, R. J., Eds.; Kulwer Academic/Plenum Publishers: New York, 2001; pp 37-60. (5) Nakagawa, Y.; Hatanaka, H. J. Neuro-Oncol. 1997, 33, 105-115. (6) Diaz, A. Z.; Coderre, J. A.; Chanana, A. D.; Ma, R. Ann. Med. 2000, 32, 81-85. 10.1021/ac025798e CCC: $22.00

© 2002 American Chemical Society Published on Web 09/12/2002

and is thereby useful for BNCT, the boron in the BPA transport molecules used in the Studsvik therapy was enriched with the 10B isotope to nearly 100%. Generally, BPA and other BNCT agents in blood and tissues are determined by the total boron content using, for example, plasma-atomic emission spectrometric (AES) techniques. Boron isotope ratios can be obtained with inductively coupled plasma mass spectrometry (ICPMS).7 For more than 20 years, the ICP techniques have been employed in speciation analysis8 studies, that is, where different species of the same element are separated from each other prior to detection. Together with a separation technique, the ICP spectrometer quickly gives very selective information on the number of compounds in the sample that contain a certain element. Since the molecular structure is totally destroyed in the hot plasma of the ICP, no structural information can be obtained. However, the retention time through the analytical column may give a hint of, for example, the polarity or size of unknown analytes. To obtain more information on the identity of the compounds in speciation analysis, the ICP measurements are often complemented with electrospray ionization (ESI)MS or -MS/MS. Separation of intact BPA from endogenous substances in biological tissues9 and blood10,11 and from trace impurities in synthetic mixtures12 using C8 and C18 materials with UV9,10,12 and ESI-MS detection11 has been reported. In the present study, porous graphitic carbon (PGC) was chosen as separation media, because this material can selectively retain not only hydrofobic analytes but also polar compounds in, for example, biomedical and pharmaceutical analyses.13-16 Another advantage with the PGC material is that it can be used over the entire pH range, so that acidic samples and mobile phases can be used without any special precautions. A few mechanistic studies of p-BPA have been performed.10,17,18 Belkouh et al. suggest that borate is lost from BPA and takes part in the BNCT process, but that the phenylalanine moiety of BPA is not incorporated into proteins or metabolized into melaninlike phenylalanine and tyrosine do.17 Yoshino et al. investigated the stability of BPA in blood in vitro and found no degradation of p-BPA within 18 h.10 However, they did not exclude the possibility (7) Probst, T. U. Fresenius’ J. Anal. Chem. 1999, 364, 391-403. (8) Templeton, D. M.; Ariese, F.; Cornelis, R.; Danielsson, L.-G.; Muntau, H.; van Leeuwen, H. P.; Lobin´ski, R. Pure Appl. Chem. 2000, 72, 1453-1470. (9) Yoshino, K.; Takasoh, T.; Hatanaka, H.; Komada, F.; Okumura, K.; Mishima, Y.; Honda, C.; Mori, Y. Advances in Neutron Capture Therapy; Plenum Press: New York, 1993; pp 465-468. (10) Yoshino, K.; Koike, N.; Kuroda, Y.; Mori, Y.; Ichihashi, M.; Kakihana, H.; Mishima, Y. Cancer Neutron Capture Therapy; Plenum Press: New York, 1996; pp 115-120. (11) Di Pierro, D.; Lazzarino, G.; Pastore, F. S.; Tavazzi, B.; Del Bolgia, F.; Amorini, A. M.; Fazzina, G.; Giuffre`, R. Anal. Biochem. 2000, 284, 301306. (12) Va¨ha¨talo, J.; Tuominen, J.; Kokkonen, J.; Krˇ´ızˇ, O.; Karonen, S.-L.; Kallio, M. Rapid Commun. Mass Spectrom. 1998, 12, 1118-1122. (13) Gu, G.; Lim, C. K. J. Chromatogr. 1990, 515, 183-192. (14) Ayrton, J.; Evans, M. B.; Harris, A. J.; Plumb, R. S. J. Chromatogr., B 1995, 667, 173-178. (15) Koivisto, P.; To¨rnkvist, A.; Heldin, E.; Markides, K. E. Chromatographia 2002, 55, 39-42. (16) Mama, J. E.; Fell, A. F.; Clark, B. J. Anal. Proceed. 1989, 26, 71-73. (17) Belkhou, R.; Abbe´, J.-Ch.; Pham, P.; Jasner, N.; Sahel, J.; Dreyfus, H.; Moutaouakkil, M.; Massarelli, R. Amino Acids 1995, 8, 217-229. (18) Yoshino, K.; Morita, M.; Mori, Y.; Kakihana, H.; Mishima, Y.; Ichihashi, M. Advances in Neutron Capture Therapy; Elsevier Science: New York, 1997; pp 239-242.

that enzymes may cause a metabolisation of p-BPA in vivo. The same group has performed an enzymatic investigation of the effect of L-tyrosine decarboxylase on p-BPA.18 Their results indicate that p-BPA is slowly metabolized by the enzyme. Gibson et al. have recently shown that BSH, sodium undecahydromercapto-closododecaborate (Na2B12H11SH), another drug used clinically in BNCT of malignant brain tumors, gives rise to urinary metabolites.19 The primary objective of this study was to couple LC on-line to ICP-AES and to ICP-time-of-flight(TOF)-MS in order to investigate the existence of degradation products or metabolites in urine samples from glioblastoma patients treated with BNCT at Studsvik Medical in Sweden.2 EXPERIMENTAL SECTION Materials. All chemicals were of analytical grade and were obtained from Merck (Darmstadt, Germany). A 1000 µg/mL certified boric acid standard was purchased from BDH Chemicals (Poole, U.K.) and diluted to 2-10 ppm (µg/g) of boron in 0.1 M hydrochloric acid. DL-4-boronophenylalanine, 95%, was obtained from Aldrich Chem. Co., Milwakee, WI. Phenylboric acid (PB), Fluka, Buchs, Switzerland, was used in the optimization of the chromatographic system as a possible degradation product. Urine, blood plasma samples, and infusion solutions of BPA were obtained from patients being treated by BNCT at Studsvik Medical, Sweden. A urine sample from a voluntary nonpatient was collected at the department. ICP-AES Instrumentation. A Spectroflame P ICP-AES instrument (Spectro Analytical Instruments, Kleve, Germany) was equipped with an MCN-100 nebulizer (model M-2, Cetac Technologies, Omaha, NE). Boron was monitored at the 249.678-nm atomic emission line. The nebulizer gas flow rate was set to 0.70 L min-1, and the power was 1290 W, 27 MHz. For new types of samples or sample pretreatments, the iron 259.940 line was also monitored along with the boron 182.640-nm line to check for possible iron interferences on the B 249.678-nm line. Collection and integration of the emission signal was performed over 0.4-s intervals with the Spectro Software 2.10 provided by the manufacturer. Transient peaks were integrated by curve-fitting in Origin (Microcal Software, Northampton, MA). ICP-TOFMS Instrumentation. For boron isotope ratio measurements, a Renaissance ICP-TOFMS instrument (Leco, St. Jospeh, MI) equipped with an MCN-100 nebulizer and a Glass Expansion cyclonic spray chamber (Glass Expansion SARL, Romainmoˆtier, Switzerland) was used. The spray chamber had to be slightly modified to fit the outer dimensions of the MCN100 nebulizer. Details of the Renaissance instrument have been described by Tian et al.20 The boron isotope signals were optimized by pumping a 1 ppm boron standard solution in 0.5% trifluoroacetic acid (TFA) in 30% aq. acetonitrile (ACN) to the plasma at a flow rate of 150 µL min-1. ESI-MS/MS Instrumentation. An API 365 triple quadrupole (Q) mass spectrometer (PE-Sciex, Concord, Canada) equipped with an IonSpray pneumatically assisted electrospray interface was (19) Gibson, C. R.; Staubus, A. E.; Barth, R. F.; Yang, W.; Kleinholz, N. M.; Jones, R. B.; Green-Church, K.; Tjarks, W.; Soloway, A. H. Drug Metab. Dispos. 2001, 29, 1588-1598. (20) Tian, X.; Emteborg, H.; Adams, F. C. J. Anal. At. Spectrom. 1999, 14, 18071814.

Analytical Chemistry, Vol. 74, No. 20, October 15, 2002

5359

used for structural investigation of one possible metabolite to BPA in the urine samples. The electrospray voltage was 5 kV, and the orifice potential was kept at 30 V. The mass spectrometer was run in positive Q1 scan mode to select masses that belong to BPA. The ions at the retention time of the possible metabolite that could be fragmented into those of BPA, m/z 163 and 209, were detected in precursor ion scans. Fragmentation in Q2 was performed by collision-activated dissociation (CAD) with nitrogen as the collision gas. Multiple-reaction monitoring was used to secure those ions that belonged to the chromatographic peak of the possible metabolite. LC Systems. The LC system coupled to the ICP instruments consisted of a pump (Jasco PU180i, Tokyo, Japan) and a 6-port injector (Cheminert Valco C2-2346, Houston, TX) with a 20-µL sample loop. The column (100 × 2 mm i.d.; particle size, 5 µm) of porous graphitic carbon (Hybercarb) was obtained from Shandon Scientific, Runcorn, U.K. For cleanup of the plasma samples, the column-switching methodology21 was applied with a Hypercarb guard column (10 × 2 mm i.d.). Deproteinized plasma (100 or 150 µL) was loaded onto the precolumn by means of a Rheodyne 7120 sample injector (Berkeley, CA) with 500 µL min-1 0.5% TFA in 5% aq. ACN fed by a Schimadzu LC-6A LC pump (Kyoto, Japan). After 80 s, the flow was reversed through the precolumn, and the analytes were transferred to the analytical column with 0.5% TFA in 30% aq. ACN at a flow rate of 150 µL min-1 using the Jasco pump. This was the same mobile phase and flow rate as was used for the urine samples. All tubing within the LC system was made of PEEK and had an i.d. of 130 µm. The LC system was connected to the spectrometer by a 470-mm-long and 175-µm-i.d. PEEK tube encapsulated in FEP TEFLON provided by the manufacturer of the nebulizer. The micro-LC system coupled to ESI-MS/MS instrument consisted of a PU-980 Jasco pump (Jasco, Tokyo, Japan), a 200µm-i.d. × 12.7-cm-long fused-silica capillary (Polymicro Technologies, Phoenix, AZ) slurry-packed with 5-µm PGC particles (Hypercarb, Hypersil, U.K.) using a slightly different procedure, as described in Koivisto et.al.15 Injections of 500 nL were performed using a manual 4-port Valco injection valve with internal loop (Valco Instruments, Houston, TX). Formic acid (0.1%) in 50% ACN was pumped through the LC-MS system at a rate of ∼2 µL min-1. The salts of the urine samples were prevented from entering the mass spectrometer by simply decoupling the LC system from the interface during the first minutes after sample injection. The retention times of salts, BPA, and the possible metabolite had been investigated in advance using a UV detector (µPEAK Monitor, Amersham Pharmacia Biotech, Uppsala, Sweden) showing wavelengths of 220 or 254 nm. BPA Standard and Sample Preparation. Stock solutions of 4 mg/mL of DL-4-boronophenylalanine were prepared in 0.5 M hydrochloric acid and diluted to concentrations ranging between 0 and 40 ppm of boron. Urine samples and infusion solutions were diluted 5-fold with 0.1 M hydrochloric acid. Both BPA standard solutions and urine samples were centrifuged at 20000g for 15 min at 5 °C and filtered through a 0.2-µm nylon membrane (Lida Manufacturing Corp., Kenosha, WI). The urine samples and BPA infusion solutions were further diluted with 0.1 M hydrochloric (21) Campı´ns-Falco´, P.; Herra´ez-Herna´ndez, R.; Sevillano-Cabeza, A. J. Chromatogr. 1993, 619, 177-190.

5360 Analytical Chemistry, Vol. 74, No. 20, October 15, 2002

Figure 2. LC-ICP-AES chromatogram of a urine sample from a BNCT-treated patient. The boron signal at 249.678 nm was measured. Approximate boron concentrations in milligrams per gram of urine are shown in brackets. Peak identification: (1, 2) unknown possible degradation products (0.05 altogether), (3) BPA (0.28), and (4) unknown (BPA-X), probably a metabolite to BPA (0.13).

acid to appropriate BPA concentrations before the LC-ICP measurements. Blood plasma samples were deproteinized using a method similar to that described by Di Pierro et al.11 To the plasma samples, a double volume of 1.0 M perchloric acid was added. The mixtures were shaken vigorously and centrifuged at 20000g for 20 min at 5 °C. After filtration through a 0.2-µm nylon membrane, the supernantant was injected into the column switching system. To determine the repeatability and efficiency of the plasma sample pretreatment and the column-switching system for BPA and the possible metabolite, although the structure of the possible metabolite is unknown, plasma deficient in BPA from one patient was spiked with BPA-containing urine from another patient. Safety Considerations. All acids and bases should be handled with great care. TFA is toxic and corrosive when in contact with the skin and eyes and when ingested and inhaled. Perchloric acid is highly corrosive to all tissues and reacts violently with many oxidizable substances. The anhydrous form and certain salts are highly explosive. Acetonitrile is combustible, volatile, and toxic if exposed to the lungs or skin. Urine and plasma blood samples should be considered as infected and should be handled only in places intended for handling biological materials using protective gloves. RESULTS AND DISCUSSION The collection times for urine and plasma samples reported refer to the time from the start of the BPA-fructose administration to the patients that suffered from glioblastoma multiforme. The BPA was administrated for 6 h and to a total concentration of 900 mg BPA/kg body weight. The infusion solutions consisted of BPA-fructose to make the BPA more soluble at physiological pH. Three unknown boron-containing compounds in the LC-ICP chromatograms were detected (Figures 2-3), that is, two possible degradation products and one suspected metabolite to BPA. The latter one is hereafter called BPA-X for the sake of simplicity. The first two boron peaks also appeared in a blank urine from a healthy person that had not received the BPA administration, but the peak heights were only about one-tenth or less of the corresponding

Table 1. BPA and Total Boron Concentrations in Urine Samples Collected 5-11 h after the Start of BPA Administration

Figure 3. LC-ICP-TOFMS chromatogram of a urine sample from a BNCT-treated patient. The 10B and 11B isotopes were monitored. Peak assignment as in Figure 2.

peaks in urine from the BNCT patients. One or both of these boron peaks in the blank urine could be a disturbance in the signal from the urinary salts eluting at the same retention time. Methodology To Investigate the Genesis of BPA-X. The unknown BPA-X was first seen when a porous graphitic carbon column was coupled on-line to an ICP-AES instrument (see Figure 2). The retention time of BPA-X was close to that of PB (see the structure in Figure 1b) used as a standard for a potential degradation product of BPA. PB was, however, ruled out as the unknown compound, since standard addition of PB to a BNCT urine sample resulted in an additional peak in the chromatogram. Urine samples from 10 patients were analyzed with LC-ICPAES. The samples were taken between 5 and 10 hours after the start of the BPA-fructose infusion. In addition, urine samples after 24 h were analyzed for seven of the patients. BPA-X was found in all urine samples. The BPA-fructose infusion solutions given to each patient were also analyzed chromatographically, as well as the BPA starting material. BPA-fructose in physiological buffer eluted at the same retention time with the TFA/ACN buffer as the single BPA units or BPA-fructose diluted with 0.1 M hydrochloric acid. Furthermore, all of these solutions resulted in just a single BPA peak. The hypothesis that BPA-X could be a byproduct formed during the synthesis of BPA was thereby rejected. A study was also performed to exclude metabolic degradation products formed either in the body or during storage of the urine samples. Three urine samples, that is, a urine sample, a BPAfructose-spiked urine, and a pure BPA solution, were heated at 37 °C for 8 h and analyzed with LC-ICP-AES. The BPA-X-to-BPA peak area ratios remained constant after heating the urine, and no peaks at the same retention time as BPA-X were found for the other heated solutions. It is, therefore, suggested that BPA-X is a metabolite of BPA. Isotope ratio measurements of the urine samples with LCICP-TOFMS (Figure 3) gave nearly the same abundance of 10B for BPA and BPA-X, that is, 99 and 98%, respectively. Although isotope ratio measurements with a separation technique coupled to an ICPMS generally suffer from less precise results than direct ICPMS measurements,22 there should be no doubt that BPA-X (22) Seubert, A. Trends Anal. Chem. 2001, 20, 274-287.

patient

time after start of BPA administration; h

BPA concn; mg B g-1 urine

total boron concn; mg B g-1 urine

A B C D E F G H I

9.5 10.5 6.7 8.7 7.5 5.7 7.6 6.5 7.3

0.31 0.41 0.32 0.32 0.33 0.30 0.57 0.26 0.23

0.51 0.56 0.34 0.53 0.43 0.43 0.73 0.37 0.35

originates from BPA. The first two peaks in the chromatogram give a 10B content of ∼80%, which probably cannot be explained by a dilution with the very low concentration of endogenous boron, which should have the natural 10B content of ∼20%. A more probable explanation for the relatively low abundance of 10B in the two first peaks is that the urinary salts, which elute at the same time, affect the 10B and 11B signals differently. The retention time of BPA-X increased from day to day, as shown in Figures 2 and 3. This was probably due to a commonly experienced change in retention properties of the PGC material with an increasing number of complex urine injections. Simple cleaning of the column by back-flushing also restored the original retention times. The day-to-day difference in retention time for BPA-X did not influence the BPA-X-to-BPA peak area ratios. Total Boron and BPA Concentrations in Urine. Since creatinine is excreted at a constant rate from the kidneys at normal kidney function, it can be used as a concentration marker for urine. However, no determination of the creatinine content of the urine samples was performed, which prevented an absolute comparison between the BPA concentrations excreted by the same patient at different times after the start of BPA administration and between different patients. Reported here are only the BPA concentrations and total boron concentrations determined by LC-ICP-AES and ICP-AES, respectively, in the urine samples collected after 5 to 11 h (see Table 1). The total boron concentrations in the urine samples and the BPA-fructose solutions were determined using the same sample pretreatment as was used prior to the LC-ICPAES measurements but with boron elemental standards instead of pure BPA standards. Depending on the degree of metabolization (Figure 4), the BPA concentration differed more or less from the total boron concentration (Table 1), and the concentration of BPA generally was 2-5 times less than in the BPA-infusion solutions. The relative standard deviation (RSD) for double injections of the urine samples generally was within 4%. The sum of the concentrations of the boron-containing compounds in the LC runs was ∼10% less than that given by the total boron determination with external calibration. To check for matrix interferences, three of the urine samples were subjected to standard additions of boric acid standard. The same boron content was, however, obtained with both the boron 182.640 and 249.678 emission line as with external standard calibration, showing no significant spectral overlap or influence from the matrix on the boron response. The reason lower concentrations are obtained with LC could be that some boron compounds are bound irreversibly to the stationary phase. Analytical Chemistry, Vol. 74, No. 20, October 15, 2002

5361

Figure 4. BPA-X-to-BPA ratios in urine at different times after the start of BPA-fructose administration. Patients A-H were exposed to a complete BNCT treatment, but patients I-J received a negligible dose or no BNCT.

No liquid certified reference material exists for boron at a suitable concentration level. To get a hint of the accuracy of the BPA determination, a spinach sample, IAEA-331, with a certified boron content of (37.6 ( 1.0) mg kg-1 was digested in nitric acid in a steel bomb and determined with ICP-AES. The result obtained for boron measured at its 182.640 nm emission line was (38.1 ( 0.5) mg kg-1 (n ) 3, 95% confidence level). BPA and BPA-X Relative Concentrations in Urine and Blood Samples. Urine samples from 10 patients and plasma samples from 2 patients were analyzed with respect to their BPA and BPA-X relative concentrations, calculated as the peak area of BPA-X divided by the peak area of BPA in the same chromatogram. It is likely that the boron response for these two compounds are nearly the same.23 The BPA-X-to-BPA ratio increased from the first urination to that collected at 24 h after start of the administration. The metabolization curves of BPA for different patients show, as expected, somewhat different rates of BPA-X formation (see Figure 4), probably reflecting the individual variations due to, for example, age, health status, and genetics. The patients were between 41 and 63 years old, except for patients A and J, who were 26 and 17 years old, respectively. The young patient A had the fastest rate of metabolism. The two raw urine samples from patient B consisted of just the supernatant, which could possibly result in somewhat false BPA-X-to-BPA ratios and a fairly large rate of metabolism. Two patients (J and I) were subject to a preliminary study with BPA-fructose infusion for one and 6 h, respectively, but without or with a negligible 4% dose of BNCT treatment afterward. Patient J was not suffering from glioblastome multiforme but had a primitive neuroectodermal tumor. Also these urine samples showed an increasing BPA-X concentration, with a rate of formation close to that of the BNCT treated patients (see Figure 4). The BNCT process in itself is therefore not considered to be the cause of the BPA-X formation. More likely, the BPA is altered into a more polar compound by the body to facilitate its (23) Svantesson, E.; Pettersson, J.; Markides, K. E. J. Anal. At. Spectrom. 2002, 17, 491-496.

5362

Analytical Chemistry, Vol. 74, No. 20, October 15, 2002

excretion. The more of the BPA that had been modified, the larger the two first peaks seen in the chromatograms in Figures 2, 3 were in relation to the BPA peak. The total boron concentration was lower in the plasma than in the urine. Therefore, larger sample volumes of the plasma samples had to be injected to be quantifiable. The BPA concentration in the plasma of one of the patients (31 years old) was found to agree with the total boron concentration normally found in the blood after 6 h BPA administration (typically, ∼40 ppm of boron). The 24 h plasma sample from the young patient denoted A in the urine study was so low in BPA and BPA-X concentration, probably due to a fast clearance of BPA and BPA-X from the blood, that BPA-X was not detectable at all with the present sample pretreatment and instrumental setup. The repeatability in BPA-X-to-BPA ratios was found to be within 2% (RSD) for four consecutive injections of urine, whereas the urine-spiked plasma obtained an RSD of 5% for four replicate sample pretreatments. The BPA-X was found to be extracted from the plasma 1.3 times more efficiently than BPA when compared to the efficiencies of the simple pretreatment of the urine samples. The quotients of BPA-X and BPA in plasma was corrected for this difference in extraction efficiency to be more comparable with the ratios reported for urine. For patient A, the BPA-X- to-BPA quotient in the plasma 7 h after the start of the BPA administration was 0.02, whereas the urine sample collected after 8.5 h had a ratio of 0.42. The other patient’s BPA-X-to-BPA quotient in plasma increased from 0.01 to 0.11 between 6 and 24 h. The urine BPAX-to-BPA ratio was not measured for this patient. The concentration of BPA in the plasma after 24 h, compared to that of 7 h after the start of the administration, had decreased by a factor of ∼100 for the young patient A. The other patient’s plasma concentration of BPA had declined about 20 times between 6 and 24 h. Micro-LC-ESI-MS/MS on BPA-X. As indicated by the work of Yoshino et al.,18 one metabolite of BPA could be the boron analogue of tyramine, that is, the BPA (M ) 208 g mole-1) without its carboxyl group (M ) 164 g mole-1; see Figure 1c). Preliminary results with micro-LC-ESI-MS/MS on urine from four of the

patients suggest that BPA-X is a larger compound than BPA. The most abundant BPA-X ion was found to be m/z 251 (M + H+), but even higher masses were detected in precursor ion scans of m/z 209 (BPA + H+) and 251. This could also be an indication that BPA is conjugated with one or several polar groups in the human body to facilitate its urinary excretion. A boron-containing fragment of m/z 163 (M + H+) was found both in the BPA and BPA-X chromatographic peaks, which might be BPA minus the carboxyl croup and with a double bond between the alanine carbons, as shown in Figure 1d. The 251 mass ion could be fragmented into both the ions of BPA and the decarboxylated and unsaturated BPA. Thus, the ion of m/z ) 163 might just be a fragment formed during the electrospray process. The 251 mass ion is proposed to be BPA acetylated either on the amine or on one of the hydroxyl groups bound to the boron atom (see Figure 1e-f). CONCLUSIONS The importance of using hyphenated techniques in medicine analysis has been shown with the application of urinary and blood plasma samples from BNCT treated patients. The analyzed body

fluids were found to contain a possible metabolite to BPA, which could not be created when incubating BPA-fructose, BPAfructose-spiked urine, or a urine sample at 37 °C from a BNCTtreated patient. Further work will include positive identification of the structure of the suggested metabolite in urine and plasma. Preliminary results with LC/ESI-MS/MS suggest that the unknown compound has a higher mass than BPA, which probably indicates that BPA undergoes some conjugation in the human body. ACKNOWLEDGMENT Dr. Paul Ross and Hypersil, Runcorn, U.K., are gratefully acknowledged for granting the LC porous graphitic carbon column. The Swedish National Research Council and The Swedish Foundation for Strategic Research are thanked for financial support.

Received for review May 24, 2002. Accepted August 13, 2002. AC025798E

Analytical Chemistry, Vol. 74, No. 20, October 15, 2002

5363