Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Boron neutron capture therapy (BNCT²) is based on a nuclear capture reaction that occurs when boron-10, a nonradioactive constituent of natural elemental boron, is irradiated with low-energy neutrons to produce high-energy alpha particles and recoiling lithium-7 nuclei. For BNCT to be successful, a sufficient amount of ¹⁰B-containing compound must be delivered to the tumor, and enough neutrons must be absorbed by the ¹⁰B atoms to produce a lethal ¹⁰B(n,)⁷Li capture reaction. More detailed information on various aspects of BNCT are provided in recent reviews by Barth et al. (1999), Soloway et al. (1998), and Coderre and Morris (1999).

Determining the metabolic fate of a drug is a major focus of drug development efforts since the identification of metabolites can help predict or explain possible toxic side effects, pharmacokinetics, and other potential biologically active agents. Elucidation of the metabolic scheme of a drug also may lead to identification of new potential synthetic drug candidates, which may possess improved efficacy relative to the reference drug. Sodium undecahydromercapto-closo-dodecaborate (Na₂B₁₂H₁₁SH), referred to as sodium borocaptate (BSH), is a drug that has been used clinically for BNCT of malignant brain tumors in Japan since 1967 and in Europe since 1998. Very little, however, is known about its metabolic fate. This, in part, may be due to the stability of the polyhedral borane anion and its resistance to analytical methods that usually are used to determine drug metabolism. A recent report has concluded that the chemical state of boron in tumor samples from patients who received BSH was altered, and this may have been due to metabolism, although no specific metabolites were suggested other than to state that they might have been attributable to possible chemical opening of the polyhedral boron cage (Gilbert et al., 2000).

BSH contains a sulfhydryl group as part of its chemical structure, and the biotransformation of xenobiotics, via the sulfhydryl group, has many well characterized pathways including: glutathione conjugation, S-glucuronidation, S-methylation, S-oxidation, disulfide formation (dimerization), and mercapturic acid formation (Silbernagl and Heuner, 1993; Testa, 1995; Rettie and Fisher, 1999). Due to the chemical, biological, and thermal stability and of the (B₁₂H₁₂)² molecule (Muetterties et al., 1964), it is unlikely that any significant biotransformation of BSH would occur on the boron cage portion of the BSH molecule. A similar polyhedral borane anion, (B₁₀H₁₀)², was administered to brain tumor patients in the early 60s, and based on chemical and NMR data, it was reported to be excreted unchanged in the urine (Sweet et al., 1962). Efforts to determine biotransformation of BSH have focused on possible metabolic modification of the sulfhydryl function by means of the aforementioned biotransformation pathways.

Although there have been several reports on the pharmacokinetics of BSH (Haselsberger et al., 1994; Stragliotto and Fankhauser, 1995; Gabel et al., 1997; Kageji et al., 2001; Horn et al., 1998), the present study is the first to identify metabolites of this drug following administration to brain tumor patients. Using electrospray ionization mass spectrometry (ESI-MS), we have attempted to identify what urinary metabolites, if any, could be detected following administration of BSH to patients who had surgical resection of their gliomas. Boronated ions were found in patient urine samples that appeared to be consistent with the following chemical structures: BSH sulfenic acid (BSOH), BSH sulfinic acid (BSO₂H), BSH disulfide (BSSB), BSH thiosulfinate (BSOSB), and a BSH-S-cysteine conjugate (BSH-CYS). These finding are presented in detail in the following article.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Chemicals. Tetrabutylammonium acetate (TBA), polyalanine, sodium iodide, alanine, and tetrapentylammonium bromide (TPA) all were purchased from Sigma Chemical Co. (St. Louis, MO). Natural abundance BSH was purchased as a drug substance from Centronic, Ltd. (Croydon, UK). The drug substance was converted to a drug product by the Pharmaceutical Service Division, University of Iowa (Iowa City, IA) and stored in sterile vials, which had been flushed with nitrogen to reduce the oxygen content to <5%, thereby minimizing oxidation. Oxidation products of BSH were <0.25% of the total drug product and included BSSB (B₁₂H₁₁SSB₁₂H₁₁) and BSOSB (B₁₂H₁₁SOSB₁₂H₁₁), both of which were measured using an high-pressure liquid chromatography assay (Goodman et al., 2000). Cesium salts of BSSB and ¹⁰BSOSB¹⁰ were obtained as a generous gift from Boron Biologicals, Inc. (Fancy Gap, VA). High-pressure liquid chromatography grade methanol was purchased from Fisher Scientific (Pittsburgh, PA). Ultrapure (18 M) laboratory deionized water was obtained from a Barnstead (Dubuque, IA) Nanopure Diamond ultrafiltration unit.

Sample Collection and Preparation. As part of the protocol for the phase I clinical study (Goodman et al., 2000), urine samples were collected from patients before dosing and during the following time intervals after the start of the 60-min intravenous infusion of BSH: 0 to 3, 3 to 6, 6 to 9, 9 to 12, 12 to 24, 24 to 48, 48 to 72, 72 to 96, and 96 to 120 h. Urine samples from two patients (patient 14 and 20), who had both received an i.v. dose of 88.2 mg/kg of BSH, were evaluated for the presence of boronated ions. A series of scans (two positive ion and one negative ion) were performed to correlate empirically observed boronated ions to putative chemical structures of possible BSH metabolites. For positive ion scans, patient urine samples were diluted using a solution of (1:1) methanol/5 mM aqueous TBA and (1:1) methanol/5 mM aqueous TPA for separate positive ion scans. The patient urine samples were diluted with methanol for negative ion scans. In all cases, the urine samples were centrifuged after dilution at 10,000g for 5 min, and the supernatant was then infused into the electrospray source at a rate of 5 to 10 µl min¹.

Instrumentation. All experiments were performed using a Micromass Q-TOF II (Micromass, Wythenshawe, UK) mass spectrometer equipped with an orthogonal electrospray source (Z-spray), operated in both positive and negative ion modes. Polyalanine and alanine was used for a positive ion mass calibration range of 100 to 2000 m/z. Other instruments had been tried, but the Q-TOF II was the only one found to give reproducible results. Sodium iodide was used for a negative ion mass calibration range of 20 to 2000 m/z. Optimal ESI conditions were determined to be: a capillary voltage of 3050 V, a source temperature 110°C, and a cone voltage of 60 V. The ESI nebulization and drying gases were nitrogen. To increase the ion intensity, the collision gas flow was set to zero. Although setting the collision gas flow to zero does slightly reduce mass resolution and accuracy, it was necessary to improve the instrument sensitivity for some of the metabolite ions. The linear quadrupole (Q1) was set to pass ions from 50 to 2000 m/z into the pusher region of the time of flight (TOF) mass analyzer where they were scanned from 500 to 2000 m/z for positive ion scans and from 50 to 1200 m/z for negative ion scans, both using a 1-s integration time. Data were acquired in continuum mode until acceptable averaged spectra were obtained (10-20 min). MassLynx v.3.4. software (Waters, Milford, MA) was used to control all settings of the Q-TOF II and was also used for calculation of theoretical isotope patterns. The averaged spectra were evaluated for the presence of mass clusters that had an isotope pattern identifiable to molecules containing at least 12 boron atoms. Typical mass resolution for the positive and negative ion scans was between 8,000 and 10,000. All boronated ions detected in the urine samples using this ESI method were assumed to have originated from the administration of BSH. Although direct-current plasma atomic emission spectroscopic boron measurements were able to detect boronated compounds in all collected urine samples (Goodman et al., 2000), due to lower sensitivity, ESI-MS analysis was limited to those patients' urine samples that were collected for the first 24 h following termination of the infusion of BSH for both positive ion scans and for the first 48 h for the negative ion scan.

Correlation of Observed Boronated Ions to BSH Metabolite Chemical Structures. The boron cage (polyhedral borane) portion of BSH has a di-negative charge at all physiological pH values (Muetterties et al., 1964; Gruner et al., 1991). As such, we were able detect BSH and its metabolites as multication, pseudomolecular ions during positive ion scans. BSH and its metabolites formed detectable positive ions by complexing with both sodium ions and the quaternary ammonium compound TBA. Reference spectra for BSH, ¹⁰BSOSB¹⁰, and BSSB were used to identify these respective compounds in the patients' urine samples. Based on known thiol biotransformation reactions, chemical structures for putative BSH metabolites were proposed. After the metabolite chemical structures were proposed, pseudomolecular weights were calculated for each metabolite using all possible perturbations of positive ion formation, taking into account the chemical structures of the proposed BSH metabolites, the charge of the proposed metabolite, and the molecular weights of sodium and TBA. After comparing the boronated ions that were detected in the urine samples with all the calculated values, chemical structures of several metabolites were found to be consistent with the boron ion clusters that were observed. The positive ion scans were repeated using a different quaternary ammonium compound (TPA). The rationale for this was that since BSH and its metabolites were using TBA cations to form positively charged complexes, the observed ions should shift m/z, corresponding to the difference in mass of the cations, if TPA is substituted for TBA. A third independent scan was performed with the urine samples using the negative ion mode. Samples from patient 20 were tested using all three experimental conditions (TBA, TPA, and negative ions), whereas samples from patient 14 were not tested using TPA.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Twenty-four hours following the administration of BSH, between 73 and 89% of the parent compound was excreted unchanged in the urine (manuscript in preparation). The remainder was in a chemical form other than BSH. A summary of the proposed BSH metabolites, which were correlated to observed ESI-MS data, is shown in Table 1, and summaries of the ESI-MS ion data for each scan can be found in Tables 2 to 4. ESI-MS analysis was limited to the urine samples that were collected during the first 24 h after BSH administration. The urine samples that were collected from 48 to 120 h after BSH administration had no detectable boron ions.

BSH. The negative ion data showed a double negatively charged boron ion cluster at 87.0 m/z that correlated with BSH (B₁₂H₁₁SH)², which had a calculated m/z of 86.9 and was identical to spectra produced from BSH reference material. The ion cluster at 87.0 m/z was detectable in all urine samples from patient 14 and patient 20, which were collected over the first 48 h after the start of the BSH infusion. There also was a singly charged negative boron ion cluster at 197.1 m/z that corresponded with BSH [(B₁₂H₁₁SH)² + Na⁺¹]¹, which had a calculated m/z of 196.9 and was identical to spectra produced from BSH reference material (data not shown). The ion cluster at 197.1 m/z was detectable in all urine samples from patient 14 and patient 20, which were collected over the first 48 h.

2

2

2

2

2

View larger version (17K): [in this window] [in a new window]   Fig. 1.   The singly charged positive boron ion cluster at 901.0 m/z. A, computer-predicted isotope pattern for the singly charged positive ion corresponding to [(B12H11SH)2 + 3TBA+1]+1. B, positive ion smoothed ESI mass spectrum of BSH obtained from the 0- to 3-h urine sample from patient 20 using TBA to form positive ions. The cluster at 901.0 m/z corresponded to the singly charged positive ion [(B12H11SH)2 + 3TBA+1]+1.

BSOH. The mass spectrum shown in Fig. 2 was attributable to the presence of BSOH in the patients' urine samples. The negative ion data had a singly charged negative boron ion cluster at 213.1 m/z that corresponded to BSOH [(B₁₂H₁₁SOH)² + Na⁺¹]¹, which had a calculated m/z of 212.9. The ion cluster at 213.1 m/z was detectable in all urine samples from patient 20, which were collected over the first 24 h, and in all urine samples from patient 14, which were collected over the first 12 h. There was also a doubly charged negative boron ion cluster at 94.0 m/z that approximately correlated with BSOH (B₁₂H₁₁SOH)², which has a calculated m/z of 94.8. The ion cluster at 94.0 m/z was detectable in all urine samples from patient 14 patient 20, which were collected over the first 24 h.

View larger version (13K): [in this window] [in a new window]   Fig. 2.   Negative ion smoothed ESI mass spectrum, which was correlated to BSOH, obtained from the 0-to 3-h urine sample from patient 20.  The cluster at 231.1 m/z corresponded to the singly charged negative ion [(B12H11SOH)2 + Na+1]1.

2

2

2

2

BSO₂H. The mass spectrum in Fig. 3 was determined to be attributable to the presence of BSO₂H in the patients' urine samples. The doubly charged negative boron ion cluster at 103.0 m/z corresponded to BSO₂H (B₁₂H₁₁SO₂H)², which had a calculated m/z of 102.9. The ion cluster at 103.0 m/z was detectable in all urine samples from patient 14 and patient 20, which were collected over the first 48 h. There was a singly charged positive boron ion cluster at 1100.3 m/z that approximately corresponded to BSO₂H [(B₁₂H₁₁SO₂H)² + 3TPA⁺¹]⁺¹, which had a calculated m/z of 1101.4. The ion cluster at 1100.3 m/z was detectable in all urine samples from patient 20, which were collected over the first 24 h.

View larger version (15K): [in this window] [in a new window]   Fig. 3.   Negative ion smoothed ESI mass spectrum, which was correlated to BSO2H, obtained from the 0- to 3-h urine sample from patient 14.  The cluster at 103.0 m/z corresponded to the doubly charged negative ion (B12H11SO2H)2.

2

2

BSH-CYS. The negative ion data had a singly charged negative boron ion cluster at 338.2 m/z that corresponded to BSH-CYS [(B₁₂H₁₁SSC₃H₅NO₂)³ + 2Na⁺¹]¹, which had a calculated m/z of 338.1 (data not shown). The ion at cluster 338.2 m/z was detectable in the 0- to 3-h urine sample from patient 14 and in the urine samples from patient 20, which were collected between 3 and 9 h.

3

3

3

View larger version (26K): [in this window] [in a new window]   Fig. 4.   Positive ion smoothed ESI mass spectrum, which was correlated to BSH-CYS, obtained from the 0-to 3-h urine sample from patient 20.  The cluster at 890.0 m/z correlated to the singly charged positive ion [(B12H11SSC3H5NO2)3 + 2TPA+1 + 2H+1]+1.

3

3

BSSB. The mass spectra shown in Figs. 5 and 6 were attributed to the presence of BSSB in the patients' urine samples. There was a singly charged negative boron ion cluster at 416.4 m/z that approximately correlated to BSSB [(B₁₂H₁₁SSB₁₂H₁₁)⁴ + 3Na⁺¹]¹, which had a calculated m/z of 414.7. The ion cluster at 416.4 m/z was detectable in all urine samples from patient 14 and patient 20, which were collected over the first 6 h. There was a singly charged positive boron ion cluster at 1841.3 m/z that approximately correlated to BSSB [(B₁₂H₁₁SSB₁₂H₁₁)⁴ + 5TPA⁺¹]⁺¹, which had a calculated m/z of 1839.1. The ion cluster at 1841.3 m/z was detectable in all urine samples from patient 20, which were collected for the first 12 h.

View larger version (25K): [in this window] [in a new window]   Fig. 5.   Positive ion smoothed ESI mass spectra. A, positive ion smoothed ESI mass spectrum obtained from BSSB reference material. The cluster at 1315.6 m/z corresponded to the singly charged positive ion [(B12H11SSB12H11)4 + 4TBA+1 + H+1]+1. B, positive ion smoothed ESI mass spectrum, which was correlated to BSSB, obtained from the 3-to 6-h urine sample from patient 20.

View larger version (22K): [in this window] [in a new window]   Fig. 6.   Positive ion smoothed ESI mass spectra. A, positive ion smoothed ESI mass spectrum obtained from BSSB reference material. The cluster at 1558.8 m/z corresponded to the singly charged positive ion [(B12H11SSB12H11)4 + 5TBA+1]+1. B, positive ion smoothed ESI mass spectrum, which was correlated to BSSB, obtained from the 6- to 9-h urine sample from patient 14.

4

4

4

4

View larger version (16K): [in this window] [in a new window]   Fig. 7.   Positive ion smoothed ESI-MS-MS daughter ion spectra. A, positive ion smoothed ESI-MS-MS daughter ion spectra from the 1557.8 m/z ion from BSSB reference standard, using TBA to form positive ions. B, positive ion smoothed ESI-MS-MS daughter ion spectra from the 1557.8 m/z ion in the 0- to 3-h urine sample from patient 14, using TBA to form positive ions.

BSOSB. The mass spectra shown in Fig. 8 were attributable to the presence of BSOSB in the patients' urine samples. There was a singly charged positive boron ion cluster at 1580.7 m/z that corresponded to BSOSB [(B₁₂H₁₁SOSB₁₂H₁₁)⁴ + 4TPA⁺¹ + Na⁺¹]⁺¹, which had a calculated m/z of 1578.7 (data not shown). The ion at cluster 1580.7 m/z was detectable in urine samples from patient 20, which were collected between 3 and 9 h.

View larger version (21K): [in this window] [in a new window]   Fig. 8.   Positive ion smoothed ESI mass spectra of BSOSB obtained from the 9- to 12-h urine sample from patient 20, using TBA to form positive ions. The cluster at 1573.9 m/z correlated to the singly charged positive ion [(B12H11SOSB12H11)4 + 5TBA+1]+1.

4

4

4

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

The application of various counter-ions to produce stable detectable pseudomolecular ions for ESI-MS analysis has been described for various organic compounds (Suwan et al., 2000; Law and Temesi, 2000; Wu, 2000; Yin et al., 2001) and specifically for boron ESI-MS analysis (Moraes et al., 2000). Using this type of approach, the chemical structures for several putative metabolites of BSH were correlated to those boron ion clusters detected in the urine samples of patients that had received the drug. Since boron has two stable isotopes of atomic masses 10 and 11, which have approximate natural abundances of 19 and 81%, respectively (Porteous et al., 1995), molecules containing multiple boron atoms in their chemical structures will have readily identifiable isotope patterns resulting from statistical permutations of isotopic abundances (Fig. 1). As such, molecules containing multiple boron atoms could be identified, based on their isotope pattern, from electrospray mass spectra and then using multiple independent scans, correlated to possible chemical structures.

Some of the observed apparent base mass peaks for the reported boron ion clusters did not exactly match the calculated molecular weights of the pseudomolecular or molecular ions for the proposed BSH metabolites. However, the calculated molecular weight of the proposed BSH metabolite was found within 2 m/z (usually 1 m/z) of the observed apparent base molecular ion, and the calculated pseudomolecular or molecular ion was contained in the observed boron ion cluster. For both BSOH and BSO₂H, the negative ion data yielded boronated ion clusters that had apparent base mass peaks that were indistinguishable from the predicted calculated molecular weights of the metabolites. The same was true of BSH-CYS. There were both positive and negative ion mass spectral data that was practically indistinguishable from the predicted pseudomolecular ions of BSH-CYS. Reference material of BSSB resulted in one boron mass cluster at 1315.5 m/z, which was indistinguishable from the mass cluster observed in the urine samples (Fig. 5), and another mass cluster at 1557.8, which had an apparent base peak mass difference of 2 m/z when compared with the urine data (Fig. 6). Ions were chosen for ESI-MS-MS determinations, from the 1557.8 boron ion mass cluster, and yielded identical CID daughter ion data from both patients' urine samples and BSSB reference material. This suggested that, although there was an apparent base peak mass discrepancy for this particular ion cluster (1557.8), they were the same molecule. When accurate mass measurements could be made, there was agreement between the measured mass and the calculated mass to suggest that the ions, which were being measured, originated from the proposed structures. Accurate mass determinations were made for ions resulting from BSOH, BSSB, and BSOSB. Scans of ¹⁰BSOSB¹⁰ reference material did not produce ions that would have been consistent with ¹⁰BSOH, indicating that the presence of ions correlated to BSOH in the patient urine samples did not originate from fragmentation of BSOSB in the ESI source.

The reason for the apparent mass discrepancy in some of the ESI spectra may have been due to various factors, such as the boron isotope distribution, possible free-radical formation inside the ESI source, and low ion abundance. Although the absence of collision gas during the positive and negative ion scans certainly reduced mass accuracy, it was not sufficient to cause the deviations that were observed in certain instances. These reasons are not offered as explanations but rather to provide insight into the problems that have been reported when performing mass spectral analysis on boron compounds. It has been recognized that mass spectral analysis of heavily chlorinated compounds can lead to apparent molecular ions, which differ from the predicted molecular weight value, and that the difference can be as great as 2 m/z (Law and Temesi, 2000). Although boron has only two naturally occurring isotopic forms, ¹¹B and ¹⁰B (80.4 and 19.6%, respectively), the exact proportion of these may vary significantly depending on the source of the boron containing raw material (Vanderpool et al., 1994; Porteous et al., 1995). Therefore, it is not unreasonable to assume that a similar phenomenon may have occurred with mass spectral analysis of heavily boronated compounds. The synthesis of BSSB has been characterized (Wellum et al., 1977), and an exceptionally stable free-radical (BS^·)² can be formed from BSSB, which also could complicate the mass spectral analysis of such compounds.

Boron ion clusters consistent to both BSH-S-oxide (B₁₂H₁₁SOH) (BSOH) and BSH-S-dioxide (B₁₂H₁₁SO₂H) (BSO₂H) were detected in patients' urine samples by direct infusion ESI-MS. Oxidation of thiol groups to sulfenic acids (XSOH) and sulfinic acids (XSO₂H) can be catalyzed in vivo by a number of enzyme systems (Silbernagl and Heuner, 1993; Testa, 1995; Rettie and Fisher, 1999). Ions consistent to both the BSSB and BSOSB were identified in the patients' urine samples. Both BSSB and BSOSB can be formed from spontaneous oxidation of BSH in aqueous media (Soloway et al., 1998). The sulfur atom of a sulfenic acid (BSOH) can become the site of nucleophilic substitution by a thiol group (BSH), yielding the dimerized disulfide (BSSB) and 1 mol of water (Testa, 1995). It is not clear from this ESI-MS data whether the formation of BSSB and BSOSB was a spontaneous chemical reaction or whether it was enzymatically mediated.

Perhaps the most interesting biotransformation product that was correlated to the observed boron ions was the BSH-S-cysteine conjugate. The formation of cysteine-S-conjugates has been characterized for many drugs as a multistep, multienzyme process that usually leads to mercapturic acid formation (Silbernagl and Heuner, 1993; Testa, 1995; Rettie and Fisher, 1999). Neither the glutathione-S-conjugate of BSH or the -glutamyl hydrolysis intermediate was detected in the patient urine samples. The absence of detectable amounts of these chemical species in the urine can be explained by the biochemical reactions that occurred. Glutathione-S-conjugates are readily excreted into the bile by carrier-mediated transport (Ketterer and Mulder, 1990; Silbernagl and Heuner, 1993). The glutathione-S-conjugates may have been subject to enzymatic hydrolytic reactions or, without any appreciable enterohepatic recycling, excreted into the feces completely bypassing the systemic circulation. This explanation is plausible considering that, according the principle of mass-balance, 14.63% of the boron dose was unaccounted for in the urine (Goodman et al., 2000). The absence of detectable -glutamyl hydrolysis intermediate may be explained by considering the high activity and colocalization of the hydrolyzing enzymes (Silbernagl and Heuner, 1993). Once the glutamic acid residue was hydrolyzed, the glycine residue will then be hydrolyzed before the intermediate can re-enter the systemic circulation.

ESI-MS scans were used to correlate observed boron ion clusters to chemical structures of putative BSH metabolites, including those that contained disulfide bonds and a cysteine-S-conjugate. Although the patient urine samples were stored at 20°C until analyzed, there was a significant time delay (approximately 2 years) between collecting the patient samples and having the ability to use ESI-MS to scan for boronated compounds. Because of this time delay, it is possible that some of the compounds that were identified as apparent BSH metabolites actually could have been products of chemical degradation of either the parent drug or other metabolites. ESI-MS scans using ¹⁰BSOSB¹⁰ standard material suggested that the boronated ions, which were consistent with BSOH and BSO₂H, were not products of molecular fragmentation of BSOSB in the ESI source. However, fragmentation of larger molecules also could have occurred and produced the ions detected. Without reference material for all the suspected metabolites and their precursors, ESI molecular fragmentation cannot be ruled out. The metabolic and oxidative products, which have been correlated to the ESI-MS, ESI-MS-MS, and accurate mass data, are reasonable biotransformations and reaction products for sulfhydryl containing compounds.

In conclusion, until the present study, there was very little data indicating that BSH underwent metabolic transformation following administration to brain tumor patients. Although 73 to 89% of the BSH dose was excreted unchanged in the urine by 24 h following administration (manuscript in preparation), our data conclusively establish that the drug did indeed undergo metabolism after administration to brain tumor patients. A liquid chromatography-MS assay for BSH has been developed in our laboratory and will be focus of a future article. Using this liquid chromatography-MS assay, we will be able to relate the plasma concentration of BSH to the direct current plasma atom emission spectroscopy plasma boron values (Goodman et al., 2000) so that preliminary information about the extent of metabolism of BSH can be obtained. Further studies, possibly using various in vitro metabolic screening techniques, will be needed to conclusively establish the presence and biological significance of these proposed BSH metabolites as tumor-targeting agents.