Formation of Hemoglobin and Albumin Adducts of Benzene Oxide in

1,8-Diazabicyclo[5.4.0]undec-7-ene (DBU) and methanesulfonic acid were ..... Data were analyzed with statistical software provided by either Microsoft...
0 downloads 0 Views 160KB Size
Download PDF
302

Chem. Res. Toxicol. 1998, 11, 302-310

Formation of Hemoglobin and Albumin Adducts of Benzene Oxide in Mouse, Rat, and Human Blood Andrew B. Lindstrom,†,‡ Karen Yeowell-O’Connell,† Suramya Waidyanatha,† Thomas A. McDonald,†,∇ Bernard T. Golding,§ and Stephen M. Rappaport*,† Department of Environmental Sciences and Engineering, School of Public Health, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina 27599-7400, National Exposure Assessment Research Laboratory, United States Environmental Protection Agency, MD-56, Research Triangle Park, North Carolina 27711, and Department of Chemistry, Bedson Building, The University of Newcastle upon Tyne, Newcastle upon Tyne NE1 7RU, U.K. Received October 6, 1997

Little is known about the formation and disposition of benzene oxide (BO), the initial metabolite arising from oxidation of benzene by cytochrome P450. In this study, reactions of BO with hemoglobin (Hb) and albumin (Alb) were investigated in blood from B6C3F1 mice, F344 rats, and humans in vitro. The estimated half-lives of BO in blood were 6.6 min (mice), 7.9 min (rats), and 7.2 min (humans). The following second-order rate constants were estimated for reactions between BO and cysteinyl residues of Hb and Alb [in units of L (g of Hb- or Alb-h)-1]: mouse Hb ) 1.16 × 10-4, rat Hb ) 15.4 × 10-4, human Hb ) 0.177 × 10-4, mouse Alb ) 2.68 × 10-4, rat Alb ) 4.96 × 10-4, and human Alb ) 5.19 × 10-4. These rate constants were used with BO-adduct measurements to assess the systemic doses of BO arising from benzene in vivo in published animal and human studies. Among rats receiving a single gavage dose of 400 mg of benzene/kg of body weight, the BO dose of 2.62 × 103 nM BO-h, predicted from Alb adducts, was quite similar to the reported AUC0-∞ ) 1.09 × 103 nM BO-h of BO in blood. Interestingly, assays of Hb adducts in the same rats predicted a much higher dose of 14.7 × 103 nM BO-h, suggesting possible in situ generation of adducts within the erythrocyte. Doses of BO predicted from Alb adducts were similar in workers exposed to benzene [13.3 nM BO-h (mg of benzene/kg of body weight)-1] and in rats following a single gavage dose of benzene [8.42 nM BO-h (mg of benzene/kg of body weight)-1]. Additional experiments indicated that crude isolates of Hb and Alb had significantly higher levels of BO adducts than dialyzed proteins, suggesting that conjugates of low-molecular-weight species were abundant in these isolates.

Introduction Benzene is known to cause leukemias in humans (1) and various other cancers in rodents (2). While the specific mechanism by which benzene exerts its carcinogenicity remains unclear, there is strong evidence that metabolic activation of benzene in the liver plays an important role (3, 4). The accepted metabolic scheme assumes an initial reaction in which benzene is oxidized by P450 2E1 (and possibly other P450 enzymes) to form the reactive species benzene oxide (BO)1 (5, 6), which is in equilibrium with its oxepin form (7). We recently * To whom correspondence should be addressed. Tel: (919) 9665017. Fax: (919) 966-4711. † University of North Carolina at Chapel Hill. ‡ U.S. Environmental Protection Agency. § The University of Newcastle upon Tyne. ∇ Current address: California Environmental Protection Agency, OEHHA/RCHAS, 2151 Berkeley Way Annex 11, Berkeley, CA 94704. 1 Abbreviations: Alb, albumin; AUC, area under the curve; BO, benzene oxide; BO-Alb, benzene oxide adduct of albumin; BO-Hb, benzene oxide adduct of hemoglobin; bw, body weight; DBU, 1,8diazabicyclo[5.4.0]undec-7-ene; EI, electron ionization; Hb, hemoglobin; kBOY, the second-order rate constant for the reaction of benzene oxide with a cysteine residue of either Alb or Hb (nucleophile Y); MWCO, molecular weight cutoff; NICI, negative-ion chemical ionization; PTTA phenyl trifluorothioacetate; SPC, S-phenylcysteine; TFAA, trifluoroacetic anhydride; Y, nucleophilic site (cysteine residue) on either Alb or Hb.

showed that BO was present in the blood of F344 rats to which benzene had been administered (8). Reactions of BO are thought to give rise to all other metabolites, notably, phenol, benzene-1,2-dihydrodiol, and muconaldehydes (9). Phenol and benzene-1,2-dihydrodiol can be further oxidized to hydroquinone and 1,2-dihydroxybenzene (catechol), which can be converted to 1,4-benzoquinone and 1,2-benzoquinone, respectively (9). While the identity of the ultimate carcinogen(s) associated with benzene exposure remains elusive, much recent research has focused on 1,2- and 1,4-benzoquinone (and their respective semiquinones) and upon muconaldehydes (10-13). Although these quinones and (E,E)-muconaldehyde demonstrate significant myelotoxicity and/or genotoxicity when administered alone, various combinations of these and other benzene metabolites lead to enhanced toxicity, suggesting that two or more substances may act in concert to produce the deleterious effects (14-16). Notably lacking from any of the proposed mechanisms of benzene carcinogenicity is a direct role for BO, despite the fact that this compound produced lung tumors in newborn mice in the only bioassay which has been reported (17). Furthermore, putative adducts of BO with proteins [i.e., S-phenylcysteine from hemoglobin (Hb) and

S0893-228x(97)00178-1 CCC: $15.00 © 1998 American Chemical Society Published on Web 03/06/1998

Benzene Oxide Hemoglobin and Albumin Adducts

Chem. Res. Toxicol., Vol. 11, No. 4, 1998 303

Figure 1. Reaction of BO with the cysteine residues in Hb and Alb to produce S-(2-hydroxycylcohexa-3,5-dien-1-yl)cysteine derivatives. These may be persistent or spontaneously dehydrate to give S-phenylcysteine residues. Subsequent treatment with methanesulfonic acid and TFAA converts the S-phenylcysteine residues [and any remaining S-(2-hydroxycyclohexa-3,5-dien-1-yl)cysteine residues] into PTTA.

albumin (Alb)] (18-21) and with DNA (22) have been detected in animals and workers exposed to large amounts of benzene. And while there is some conflicting evidence concerning BO’s mutagenicity (23), at least two separate groups have reported that BO is mutagenic in the Ames bioassay (24, 25). Given the reactivity, as well as the difficulties associated with synthesis and measurement of BO, it would be useful to employ protein adducts to estimate the tissue dose of BO in vivo. Toward this end, we developed an assay for measuring adducts of BO with cysteinyl residues of Hb (20) and Alb (these adducts will henceforth be referred to as BO-Hb and BO-Alb, respectively) which is considerably simpler and more sensitive than that developed previously by Bechtold et al. (18). In this assay we assume that reactions of cysteine residues in Hb and Alb with BO give S-(2-hydroxycyclohexa-3,5-dien-1-yl)cysteine derivatives (Figure 1) which may be persistent or spontaneously dehydrate to give S-phenylcysteine residues. Treatment of the derivatized protein with methanesulfonic acid and trifluoroacetic anhydride (TFAA) converts the S-phenylcysteine residues into phenyl trifluorothioacetate (PTTA). Any remaining S-(2-hydroxycyclohexa-3,5-dien-1-yl)cysteine residues are also expected to be converted into PTTA. The assay therefore measures total S-adducts in the protein derived from BO. We have applied this assay to observe dose-related increases of BO-Hb from both rats (20) and humans (26) exposed to benzene. Unfortunately, the lack of critical kinetic data has prevented the use of BO-Hb and BO-Alb levels to estimate the doses of BO to the blood. Likewise, initial applications of our simplified assay (20, 26) focused upon Hb, and it has not been demonstrated heretofore that BOAlb can also be reliably measured. In the current investigation, we verify the suitability of the assay for BO-Alb and generate the key kinetic parameters for reactions of BO with Hb and Alb in vitro with blood from B6C3F1 mice, F344 rats, and humans. In each case, the overall rate of reaction is estimated and the second-order reaction rate constants, leading to formation of cysteinyl BO-Hb and BO-Alb, are determined. These rate constants are then used with measurements of BO-Hb and BO-Alb to estimate the BO-blood doses associated with exposures to benzene in both animals and humans. We also examine the extent to which dialysis of isolated Hb and Alb, a step which removes low-molecular-weight conjugates, affects the observed levels of BO-Hb and BOAlb in the assay.

Materials and Methods Chemicals. Acetone, diethyl ether (anhydrous), and hexane (nanograde) were obtained from Mallinckrodt (Paris, KY). Benzene was purchased from EM Science (Gibbstown, NJ). 1,8Diazabicyclo[5.4.0]undec-7-ene (DBU) and methanesulfonic acid were purchased from Aldrich Chemical Co. (Milwaukee, WI). Ethyl acetate was obtained from Fisher Scientific (Pittsburgh, PA). Ethyl alcohol was purchased from Aaper Alcohol and Chemical Co. (Shelbyville, KY). Human Hb and Alb and sodium acetate were purchased from Sigma (St. Louis, MO). Methoxyflurane was obtained from Pitman-Moore (Mundelein, IL). S-Phenylcysteine (SPC) and [2H5]SPC were kindly provided by Drs. A. Gold and R. Sangaiah of the University of North Carolina, Chapel Hill. [2H8]Styrene was purchased from Cambridge Isotope Laboratories (Andover, MA). Toluene was purchased from Baxter Healthcare Corp. (Muskegon, MI). Trifluoroacetic anhydride (TFAA) was purchased from Pierce (Rockford, IL) and was distilled once before use. BO was synthesized by dehydrohalogenation of the precursor 4,5-dibromo-1,2-epoxycyclohexane, using DBU in anhydrous ether at a 1:4 molar ratio, in a manner similar to that described in Gillard et al. (27). The structure and purity of the final product were confirmed by GC/ MS and proton NMR (500 MHz, acetone-d6) [δ 6.20 (2H, m), 5.84 (2H, m), 5.26 (2H, br d, J ) 4.5 Hz)] (8). Caution: Please note that TFAA reacts violently with water and should only be used to derivatize protein that is completely dry as described below. Also note that benzene and BO are carcinogens that should be handled with great care. All of the reactions and processes involving volatile compounds described below should only be conducted in a certified laboratory fume hood. Kinetic Basis for the Use of Protein Adducts as Tissue Dosimeters. The underlying kinetic basis for using protein adducts to estimate tissue doses of reactive intermediates can be gleaned largely from the work of Ehrenberg and co-workers (28-30). Let [BO] represent the blood concentration of BO in vivo following exposure to benzene. Our goal is to investigate the blood dose of BO, which is the integrated blood level of this substance over time t. We designate this dose DBO, which is given by:

DBO )

∫ [BO](t) dt ) t

0

[BO]o (1 - e-ket) ke

(1)

where [BO]o represents the level of BO at time t ) 0 and ke is the pseudo-first-order elimination rate constant representing loss of BO by all reactions. Within the blood, BO can react with any particular nucleophile, designated Y (at concentration [Y]), to produce adduct BO-Y (in our work, Y represents a free cysteine thiol residue of Hb or Alb) at rate kBO-Y, representing the second-order reaction rate constant. Estimates of kBO-Y can be obtained from in vitro experiments where virtually all BO has been allowed to react (i.e., where t . 1/ke) and where

304 Chem. Res. Toxicol., Vol. 11, No. 4, 1998

Lindstrom et al.

adducts have been measured prior to loss due to chemical instability or protein turnover. Under such circumstances,

kBO-Y =

[BO-Y] k [Y][BO]o e

(2)

and the blood dose of BO (nM BO-h) to protein Y, designated YDBO, can be predicted from the adduct level as follows: YDBO

=

[BO-Y] [Y]kBO-Y

(3)

We will apply eq 2 to estimate the second-order reaction rate constants between BO and cysteinyl residues of Hb and Alb in vitro and eq 3 to predict the doses of BO in animals following a single administration of benzene. In each case we will substitute the sample estimate of [BO-Y]/[Y] (based upon measurements of adduct level per g of Y) and of kBO-Y for the corresponding true values in the above expressions. For humans chronically exposed to benzene, [BO-Y] rises and falls over time due to production of [BO] (associated with exposure to benzene) and to loss of adduct due to protein turnover and/or adduct instability. Let µBO and µBO-Y represent the mean values of [BO] and [BO-Y] in the blood of a person following long-term exposure to benzene according to a consistent regimen, such as 8 h/day, 5 days/week. If BO-Hb and BOAlb are chemically stable, then the relationship between µBO-Y and µBO depends on the process of protein turnover, which differs between Hb and Alb. Hb turns over with the removal of erythrocytes from the blood, which is a fixed time (designated ter) for each animal species; e.g., ter = 120 days in humans (31). In this case,

µBO-Hb [Hb]

) kBO-Hb µBO

ter 2

(4)

from which the dose of BO to Hb can be predicted from BO-Hb levels as follows: HbDBO

)

µBO-Hb [Hb]kBO-Hb

) µBO

ter 2

(5)

The turnover of Alb, on the other hand, is a random process governed by first-order kinetics, so that

µBO-Alb [Alb]

)

kBO-Alb µBO kAlb

(6)

where kAlb represents the first-order rate of turnover of Alb which is approximately 0.0014 h-1 in humans (32). Thus, the dose of BO to Alb can be predicted from levels of BO-Alb as AlbDBO

)

µBO-Alb [Alb]kBO-Alb

)

µBO kAlb

(7)

We will employ eq 7 to predict the blood dose of BO, based upon BO-Alb measured in the blood of workers exposed to benzene by substituting the sample estimate of µBO-Y for the true value in this expression. The following units will be employed unless indicated otherwise: [BO-Y] in nM, [Y] in g/L, t in h, ke in h-1, kBO-Y in L (g of Y)-1 h-1, and YDBO in nM-h. We also assume a molecular weight of 66 000 Da for both Hb and Alb and values of [Hb] and [Alb] in mouse, rat, and human blood of 137, 153, and 152 g/L for Hb and 16.9, 21.1, and 24.3 g/L for Alb (33-36). Animals and Blood Collection. Male F344 rats (200-250 g) and B6C3F1 mice (20-25 g) were obtained from Charles River Breeding Laboratories (Raleigh, NC) and were housed in polycarbonate cages maintained on a 12-h light/dark cycle for at least 2 weeks before use. Food and water were provided ad libitum. Blood for in vitro experiments was collected from 5

rats and 40 mice by cardiac puncture into a heparinized syringe following anesthesia with methoxyflurane. Blood was stored immediately on ice and used in the experiments described below within 5 h. For the in vivo portion of this study, BO adduct levels were determined in blood samples from a study which had been previously conducted in this laboratory (21). Briefly, groups of four male F344 rats [average weight 308 g (SE ) 7.5)] had been administered [14C/13C6]benzene in corn oil at 50, 100, 200, and 400 mg/kg of body weight (bw) via gastric intubation and had been sacrificed after 24 h. Portions of red blood cell lysate and plasma from that study had been stored at -80 °C for approximately 2 years prior to the current investigation. Human Samples. Human blood samples for in vitro studies were collected by venous puncture from two volunteers with informed consent. Blood was stored immediately on ice and used in the experiments described below within 5 h. Elimination of BO from Blood in Vitro. Whole blood was obtained from humans (n ) 2), rats (n ) 3), and mice (n ) 20; initially pooled and then divided into three portions) as described above; 1-mL aliquots were placed in 4-mL vials, and sufficient BO (0.368 µg/µL in acetone) was added to provide a final concentration of 36.8 µM. Vials were mixed by gentle inversion and incubated at 37 °C for a period of 1-40 min. Then the samples were vigorously shaken with a solution of 1 mL of ethyl acetate containing 1 µg of toluene (internal standard) for 30 s and centrifuged for 3 min, and the organic layers were transferred to clean vials. The levels of BO in the extracts were determined via GC/MS in the positive chemical ionization mode with a Hewlett-Packard 5890 series II gas chromatograph equipped with a Hewlett-Packard 5989 A MS engine. The chemical ionization reagent gas, methane, was at a pressure of 2.0 Torr. The ion source and the detector temperatures were 150 and 280 °C, respectively. He was used as the carrier gas at a flow rate of 1 mL/min. Portions (2 µL) were injected (on column) into a phenylmethyl-deactivated guard column (5 m, 0.32-mm i.d.; Restek, Bellefonte, PA) connected to a DB-5 fused silica gel capillary column (30 m, 0.32-mm i.d., 1-µm phase thickness; J & W Scientific, Inc., Folsom, CA) at 50 °C. The oven was kept at 50 °C for 6 min and then increased at 50 °C/ min to 250 °C. The molecular ions (M + H) of BO (m/z 95) and toluene (m/z 93) were monitored in the selected ion monitoring mode. Production of Adducts of BO with Blood Proteins in Vitro. Aliquots (2 mL) of fresh blood from rats (n ) 2), mice (n ) 20; initially pooled and then separated into two portions), and humans (n ) 2) were placed in 4-mL vials, and a sufficient quantity of a solution of 0.368 µg/µL BO in acetone was added to provide a final concentration of 0, 9.20, 36.8, 110, or 184 µM BO. Samples were incubated at 37 °C for 3 h with gentle mixing once every 15 min. Isolation and Purification of Hb and Alb. After reaction with BO (in vitro experiments) blood samples were centrifuged at 800g for 15 min, and the plasma layer was removed via pipet. The red blood cell layer was then washed three times with equal volumes of 0.9% saline, with the first two washes being added to the initial plasma isolate. An equal volume of deionized water was added to each specimen of red blood cells which was then frozen at -20 °C overnight to ensure maximal cell lysis. Hb and Alb were isolated following the procedure described in Rappaport et al. (37) with some modifications. Briefly, Hb was obtained from thawed red blood cells by centrifuging at 30000g to remove cell membranes, purifying the supernatant by exhaustive dialysis (Spectra-Pore 1, 6000-8000 MWCO) against 4 × 3.5 L of deionized water at 4 °C over 24 h (rather than by Sephadex chromatography), and precipitating the globin with the addition of cold acidified acetone (0.1% HCl). The precipitated globin was washed with ice-cold acetone and dried to constant weight in a vacuum oven at 37 °C. For the in vivo experiment, previously prepared samples of membrane-free Hb from each dose group (21) were thawed and purified as described above.

Benzene Oxide Hemoglobin and Albumin Adducts Alb was obtained from thawed plasma (for both in vitro and in vivo studies) by adding a solution of saturated (NH4)2SO4 dropwise until a final concentration of 50% was achieved. This mixture was then centrifuged at 900g to remove the immunoglobulins. The supernatant was removed, placed in dialysis tubing (Spectra-Pore, 12 000-14 000 MWCO), exhaustively dialyzed against 4 × 3.5 L of deionized water at 4 °C over 24 h, and then lyophilized to a constant weight. Effect of Protein Isolation and Purification upon Adduct Levels. In preliminary studies we observed that Hb, which had been purified previously in our laboratory by Sephadex size-exclusion chromatography (21) rather than by dialysis, and Alb, which had been purified by the method described in Bechtold et al. (19), appeared to contain higher levels of BO adducts than those of comparable proteins purified in our current assay. To test the hypothesis that the methods of protein isolation and purification affected the apparent levels of BO-Hb and BO-Alb, we conducted two experiments. In the first, BO-Hb levels were determined in parallel samples of rat Hb which had been purified by both Sephadex chromatography (original data presented graphically in ref 20) and the exhaustive dialysis technique described above. In the second experiment, three sets of 2-mL aliquots (in triplicate) of human blood from a single donor were modified with 36.6 µM BO as outlined above. Then Alb was isolated from the plasma and purified by one of three procedures, i.e., by our current assay, by the method of Bechtold et al. as reported (19), and by the Bechtold method with an additional step of exhaustive dialysis against deionized water (Spectra-Pore 12 00014 000 MWCO) prior to measurement of BO-Alb. To determine whether dialysis itself contributed to loss of BO-Alb, possibly due to hydrolysis, portions of the dried dialyzed Alb (isolated by our current procedure) were redissolved in water and exhaustively dialyzed a second time against deionized water (total ) 8 × 3.5 L over 48 h) prior to analysis of BO-Alb. Storage Stability of BO-Protein Adducts. The long-term storage stability of BO-protein adducts was assessed by statistical analysis of data which have been previously presented (graphically) in ref 20. Briefly, BO-Hb adduct levels from the rats dosed by McDonald et al. (21) (described above) were determined shortly after dosing and then reanalyzed after storage of isolated Hb for 1 year at -80 °C. Measurement of BO Adducts with Hb and Alb. The cysteinyl adducts of BO with Hb and Alb were assayed with the method of Yeowell-O’Connell et al. (20). Briefly, 100 µL of a 50 mg/mL solution of protein was added to a known amount of isotopic internal standard ([2H5]-S-phenylcysteine, typically 5 pmol for Hb and 150 pmol for Alb) and then dried in a vacuum oven at 60 °C. After addition of 800 µL of TFAA and 20 µL of methanesulfonic acid, the proteins were incubated at 100 °C for 40 min to produce PTTA. The samples were then allowed to cool to room temperature, and the excess TFAA was removed under a stream of N2. Hexane and 0.1 M Tris buffer (pH 7.5) (1 mL each) were added, and the mixture was vigorously agitated for 30 s and then centrifuged. The resulting organic extract was washed with 1 mL of 0.1 M Tris buffer followed by 1 mL of deionized water. Samples were then transferred to vials for analysis via GC/NICI-MS. Standard curves for the assay were prepared by adding a range of S-phenylcysteine to 5-mg aliquots of blank Sigma Hb and Alb which were mixed with a known amount of isotopic internal standard and assayed as described above for the experimental samples. To determine how the mass of Alb used in the derivatization might affect the linear range of this assay, a solution of 20 µL of human BO-Alb (containing approximately 35 pmol of BOAlb) was added to 1, 2, 3, 4, 5, 7, 8, 10, 15, 20, and 30 mg of commercial human Alb containing 30 pmol of the isotopic internal standard. These proteins were then assayed for adducts as described above. GC/NICI-MS Analysis of PTTA. PTTA was analyzed using a Hewlett-Packard 5890 series II plus gas chromatograph equipped with a Hewlett-Packard 5989B MS engine employing

Chem. Res. Toxicol., Vol. 11, No. 4, 1998 305

Figure 2. In vitro elimination of benzene oxide (BO) from B6C3F1 mouse (O), F344 rat (4), and human (0) blood. Initial BO concentration was 36.8 µM. Each point represents the mean of three measurements from mouse and rat blood and four measurements from human blood. a DB-5 fused silica gel capillary column (60 m, 0.25-mm i.d., 0.25-µm phase thickness; J & W Scientific, Inc.); 3-µL samples were injected in the splitless mode with a He carrier gas flow of 1.5 mL/min. The injection port and source temperatures were 250 and 100 °C, respectively. The oven temperature was held at 50 °C for 3 min, then increased at 2.3 °C/min to 140 °C, followed by 50 °C/min to 250 °C where it was held for 10 min. Methane at a source pressure of 2 Torr was used as the chemical ionization reagent gas. The molecular ions of PTTA (m/z 206), [2H5]PTTA (m/z 211), and [13C6]PTTA (m/z 212) were monitored using the selected ion monitoring mode. Data Analysis. Data were analyzed with statistical software provided by either Microsoft Excel (Redmond, WA) or GraphPad Prism (San Diego, CA) with a p-value of 0.05 indicating statistical significance. The effects of Hb purification method (Sephadex vs dialysis) and long-term cold storage upon adduct levels were assessed using two-way ANOVA. The effect of Alb isolation method [that of Bechtold et al. (19) vs (NH4)2SO4 method (37)] was determined using a two-sample t-test. Analysis of the effect of prolonged dialysis was determined using a paired t-test. Standard errors of the second-order reaction rate constants (kBO-Y) and the adduct-derived dose (YDBO) were estimated from the following relationships:

x( x(

SEkBO-Y ) kBO-Y

SEYDBO ) YDBO-Y

) ( ) ) ( )

SE[BO-Y]/[Y] [BO-Y]/[Y]

SE[BO-Y]/[Y] [BO-Y]/[Y]

+

SEke

+

SEkBO-Y

2

2

2

ke

2

kBO-Y

where kBO-Y, [BO-Y]/[Y], and YDBO refer to the estimated values. Differences between HbDBO and AlbDBO were tested with a z-test because, as shown above, SEYDBO is comprised of components related to estimation of both [BO-Y]/[Y] and kBO-Y and it is not possible to determine degrees of freedom with which to perform a standard t-test.

Results Elimination of BO from Whole Blood. To determine the first-order elimination constant (ke) and the corresponding half-life [t1/2 ) ln(2)/ke] of BO in blood, the natural logarithm of the BO/toluene peak ratio was plotted against time; the resulting slope (ke) should represent the sum of all first-order reaction rate constants occurring as BO was eliminated from the specimen. As shown in Figure 2, the slopes of the relationships were very similar for blood from the three species,

306 Chem. Res. Toxicol., Vol. 11, No. 4, 1998

Lindstrom et al.

Table 1. First-Order Elimination Constants (ke) and Corresponding Half-Times (t1/2) for BO Incubated with Whole Blood from B6C3F1 Mice, F344 Rats, and Humans at 37 °C in Vitro species

ke (h-1) (SE)

t1/2 (h)

r2

mouse

6.32 (0.11) 5.24 (0.10) 5.82 (0.13)

0.11

0.999

0.13

0.993

0.12

0.994

rat human

Table 2. Linear Regression of [BO-Y]/[Y] on [BO]o for BO Incubated with Whole Blood from B6C3F1 Mice, F344 Rats, and Humans at 37 °C in Vitroa

protein

species

βY (SE) [pmol of BO-Y (g of Y)-1 (µM BO)-1]

Hb Hb Hb Alb Alb Alb

mouse rat human mouse rat human

18.4 (0.71) 294 (6.12) 3.03 (0.12) 42.5 (1.74) 94.6 (9.44) 89.1 (13.3)

r2

linear range (µM [BO]o)

0.988 0.997 0.988 0.993 0.962 0.919

0-184 0-184 0-184 0-36.8 0-36.8 0-36.8

a [BO-Y] represents the concentration of a cysteinyl adduct of BO with Y ) Hb or Alb; [BO]o represents the initial concentration of BO; βY represents the estimated slope of the linear regression of [BO-Y]/[Y] on [BO]o.

Table 3. Second-Order Reaction Rate Constants (kBO-Y) for Reaction of BO with Y in Whole Blood from B6C3F1 Mice, F344 Rats, and Humans at 37 °C in Vitroa kBO-Y (SE) protein species Hb Hb Hb Alb Alb Alb a

Figure 3. In vitro production of (A) BO-Hb (pmol of BO-Hb/g of Hb) and (B) BO-Alb (pmol of BO-Alb/g of Alb) in fresh whole B6C3F1 mouse (O), F344 rat (4), and human (0) blood during a 3-h incubation with increasing concentrations of BO (0-184 µM). Each point represents the mean of two measurements; lines correspond to the linear range of the relationship.

indicating the same rate of elimination. Table 1 lists the estimated values of ke and the corresponding half-times as well as coefficients of determination (r2). In all cases t1/2 was about 7-8 min and r2 was greater than 0.993, indicating good fit of the linear model. Production of Adducts of BO with Blood Proteins in Vitro. The formation of BO adducts with Hb and Alb was investigated using mouse, rat, and human blood which had been incubated with BO for 3 h at 37 °C at initial concentrations of 0-184 µM. As shown in Figure 3A, the production of BO-Hb was proportional to [BO]0 over the entire range. However, the production of BOAlb showed limited evidence of being greater than proportional to concentration for [BO]0 > 36.8 µM (Figure 3B), particularly for rat and human blood. The estimated slopes and standard errors from the linear regressions of [BO-Y]/[Y] on [BO]0 are shown in Table 2 over the

mouse rat human mouse rat human

L (g of

Y)-1

h-1

1.16 × 10-4 (0.049 × 10-4) 15.4 × 10-4 (0.439 × 10-4) 0.177 × 10-4 (0.008 × 10-4) 2.68 × 10-4 (0.119 × 10-4) 4.96 × 10-4 (0.504 × 10-4) 5.19 × 10-4 (0.783 × 10-4)

L (mol of Y)-1 h-1 7.67 (0.32) 102 (2.90) 1.16 (0.05) 17.7 (0.79) 32.7 (3.33) 34.2 (5.16)

Y ) Hb or Alb. Rate constants were estimated under eq 2.

indicated ranges, along with the coefficients of determination. The relative rate of adduct formation for each protein can be inferred from the second-order rate constant (kBO-Y), which was estimated under eq 2, using the estimates of [BO-Y]/[Y], [BO]0, and ke from Tables 1 and 2. These estimated values of kBO-Y are indicated in Table 3. For convenience, the estimates are given for nucleophile concentrations on both a weight basis [i.e., for kBO-Y in L (g of Y)-1 h-1] and (after multiplying by the molecular weight of Hb or Alb) a molar basis [i.e., for kBO-Y in L (mol of Y)-1 h-1]. These rate constants indicate the following order of reactivity of BO with Hb, rat . mouse > human, and with Alb, human ≈ rat > mouse. Production of Adducts of BO with Blood Proteins in Vivo. Table 4 shows the levels of BO-Hb and BOAlb from the animals dosed by McDonald et al. (21) determined in this study with our standard protein isolation techniques (based on exhaustive dialysis). Clearly, the adduct levels increased with the dosage of benzene over the whole range of 50-400 mg/kg of bw. However, the level of BO-Hb/[Hb] per unit of dosage showed evidence of less than proportional production of BO with increasing benzene dosage; i.e., the level dropped from 0.141 nmol of BO-Hb (g of Hb)-1 (mg/kg of bw)-1 at 100 of mg/kg bw to 0.074 nmol of BO-Hb (g of Hb)-1 (mg/ kg of bw)-1 at 200 mg/kg of bw, a difference that was statistically significant (p < 0.05, t-test). Since this effect was not prominent for the albumin adducts, where levels remained constant at 0.0042 nmol of BO-Alb (g of Alb)-1 (mg/kg of bw)-1] from 50 to 200 mg/kg of bw, and diminished only slightly to 0.0033 nmol of BO-Alb (g of Alb)-1 (mg/kg of bw)-1 at 400 mg/kg of bw (a difference that was not statistically significant), and given the small number of data points, we cannot speculate further as to the linearity of the dose-response relationship in vivo. Range of the Assay for Alb. To determine how the mass of Alb might affect the linear range of the assay,

Benzene Oxide Hemoglobin and Albumin Adducts

Chem. Res. Toxicol., Vol. 11, No. 4, 1998 307

Table 4. Levels of BO-Hb and BO-Alb Measured in F344 Ratsa to Which Benzene Had Been Administered According to McDonald et al. (21)

a

benzene dosage (mg/kg of bw)

BO-Hb (SE) (nmol/g of Hb)

BO-Hb dosage (SE) [nmol (g of Hb)-1 (mg/kg of bw)-1]

BO-Alb (SE) (nmol/g of Alb)

BO-Alb dosage (SE) [nmol (g of Alb)-1 (mg/kg of bw)-1]

0 50 100 200 400

0.41 (0.26) 7.74 (0.64) 14.1 (1.59) 14.8 (1.88) 22.6 (1.68)

0.155 (0.013) 0.141 (0.016) 0.074 (0.009) 0.057 (0.004)

0.013 (0.013) 0.208 (0.026) 0.421 (0.041) 0.831 (0.073) 1.30 (0.137)

0.0042 (5.2 × 10-4) 0.0042 (4.1 × 10-4) 0.0042 (3.7 × 10-4) 0.0033 (3.4 × 10-4)

Adduct levels represent the estimated mean for 3-4 animals/group.

Table 5. Levels of BO-Hb Adductsa Measured in F344 Ratsb to Which Benzene Had Been Administered According to McDonald et al. (21) benzene dosage (mg/kg of bw) 0 50 100 200 400

Table 6. Levels of BO-Alb Adducts (nmol/g of Alb) Assayed after 3 h of Incubation of Human Blood with BO at an Initial Concentration of 36.8 µM

BO-Hb (SE) (nmol/g of Hb) sephadex dialysis 0.40 (0.23) 9.66 (0.71) 28.8 (3.50) 29.9 (2.79) 37.8 (2.01)

0.41 (0.26) 7.74 (0.64) 14.1 (1.59) 14.8 (1.88) 22.6 (1.68)

a Proteins were purified by either Sephadex size-exclusion chromatography or exhaustive dialysis. b Adduct levels represent the mean and standard errors for 3-4 animals/group.

we combined a constant amount of BO-Alb (approximately 35 pmol) with increasing amounts of human Alb (1-30 mg) prior to measuring the adducts. Results from the experiment were unremarkable, with all measured adduct levels being within 23% of the overall mean (average deviation ) 14.5%) and with no apparent trends toward higher or lower yields with increasing amounts of Alb. Effect of Protein Isolation and Purification upon Adduct Levels. The possible influence upon the assay of different methods of protein isolation and/or purification was investigated with blood samples modified with BO either in vivo or in vitro. Regarding purification of Hb, blood samples from the animals dosed by McDonald et al. (21) were purified using both exhaustive dialysis (removing compounds < 6000-8000 MW) and Sephadex size-exclusion chromatography (37). The results, shown in Table 5, indicate the levels of BO-Hb were about 40% lower following dialysis compared to Sephadex chromatography, a difference that was statistically significant (p < 0.0001, two-way ANOVA). Because the two techniques differed only in the final purification step, it seems likely that dialysis was able to remove more of the lowmolecular-weight compounds (including glutathione conjugates) that otherwise inflated the adduct determinations using this assay. Regarding isolation and purification of Alb, we compared BO-Alb levels in proteins which had been prepared by either our standard (NH4)2SO4 method (which includes dialysis to remove compounds less than 12 000-14 000 MW) or that described by Bechtold et al., either as published (19) or with an additional dialysis step (n ) 3 for each method). Results, summarized in Table 6, indicate that the levels of BO-Alb in proteins isolated by the Bechtold et al. method were 1.87-fold higher than those obtained by our procedure (p < 0.0001, two-sample t-test). Alb isolated by the Bechtold et al. procedure and then purified by dialysis produced adduct levels which were quite similar to those observed in our assay (p ) 0.14, two-sample t-test). Finally, the levels of BO-Alb observed in the specimens of Alb which had been dialyzed

(NH4)2SO4 method with 24-h dialysis (SE)

Bechtolda method (SE)

Bechtolda method with 24-h dialysis (SE)

(NH4)2SO4 method with 48-h dialysis (SE)

4.79 (0.07)

8.97 (0.17)b

4.02 (0.41)

4.75 (0.33)

a

Refers to the protein isolation method of Bechtold et al. (19). Indicates significant difference from (NH4)2SO4 method (p < 0.0001). Assays were performed in triplicate.

b

twice were not significantly different from those dialyzed only once (p ) 0.89, paired t-test), indicating that the adducts were stable during the dialysis procedure. Taken together, these results suggest that low-molecular-weight compounds contributed to the higher values observed in the Bechtold et al. procedure and that much of the BOAlb reported by Bechtold et al. (19) resulted from lowmolecular-weight cysteinyl conjugates. Storage stability of BO-protein adducts was assessed by analysis of BO-Hb samples from animals dosed by McDonald et al. (21) and then reanalyzed after storage for 1 year at -80 °C. The results (data not presented) indicate that adduct levels remained nearly constant at all dose levels with no significant changes noted (p ) 0.62, two-way ANOVA).

Discussion The relationships between [BO-Y] (for Y representing cysteinyl residues of Hb or Alb) and [BO]0, shown in Figure 3A (for BO-Hb) and 3B (for BO-Alb), provide conclusive evidence that BO reacts with Hb and Alb in blood to produce cysteinyl adducts which can subsequently be measured as PTTA. This finding, coupled with our report of BO in the bloodstream of benzenedosed rats (8), supports the use of these adducts as biomarkers of benzene exposure as reported in recent studies (20, 21, 26). Taken together, these results indicate that the dose of BO to the systemic circulation accounts for a significant fraction of the metabolized benzene dose. Indeed, the relatively long half-life of BO in blood (≈7-8 min) suggests that BO can react with proteins and DNA in many different tissues throughout the body following metabolism of benzene. An examination of the estimated second-order rate constants (Table 3) reveals that BO is much more reactive with rat Hb than with any of the other nucleophiles examined and that the rank order of BO’s reactivity toward Hb was rat > mouse > human. These findings are consistent with those observed for styrene 7,8-oxide (37, 38) and ethylene oxide (39) and presumably reflect the presence of an additional reactive free cysteine residue in rat Hb (at position 125 on the β chain) which

308 Chem. Res. Toxicol., Vol. 11, No. 4, 1998

Lindstrom et al.

Table 7. Predictions of the Dose of BO in the Blood of F344 Rats to Which Benzene Had Been Administered from Measurements of BO-Hb and BO-Alb dosage (mg/kg of bw)

BO-Hb (nmol/g of Hb)

a HbDBO (SE) (nM BO-h)

BO-Alb (nmol/g of Alb)

50 100 200 400

7.74 14.1 14.8 22.6

5.03 × 103 (440) 9.16 × 103 (1060) 9.61 × 103 (1250) 14.7 × 103 (1170)

0.208 0.421 0.831 1.30

a AlbDBO

(SE) (nM BO-h)

HbDBO/AlbDBO

419 (68) 849 (119) 1.68 × 103 (225) 2.62 × 103 (384)

12.0 10.8 5.74 5.60

a HbDBO and AlbDBO represent the blood doses predicted for BO under eq 3 based upon estimates of BO-Hb and BO-Alb, respectively, and the corresponding second-order rate constants given in Table 3.

is not present in mice or humans (40). Likewise, our results indicate that BO is about equally reactive with Alb from humans and rats, a finding which is consistent with the reactivity of 1,4-benzoquinone demonstrated by McDonald et al. (41). This similar reactivity is probably related to the structural similarities of Alb molecules from both species, which contain a single free cysteine at position 34 (42, 43). Overall, the reactivity of BO toward cysteinyl residues of Hb and Alb is comparable to that of styrene 7,8-oxide (37, 38) and is substantially less than that of 1,4-benzoquinone (41). Although the reactivity of BO toward human Alb is about 30 times greater than with human Hb, the turnover rate of human Hb is slower than that of Alb. Thus, assuming that BO-Alb and BO-Hb are both stable in vivo, then from eq 4 and 6 the mean level of BO-Alb per unit mass of protein should be about 14.5 times that of BOHb [i.e., (µBO-Alb/[Alb])/(µBO-Hb/[Hb]) ) 14.5] following long-term exposure to benzene. This suggests that BOAlb would be more easily detected than BO-Hb and thus would be a more useful biomarker of human exposure to benzene. Such a difference in the relative abundance of adducts probably explains, in part, why Bechtold et al. were able to detect BO-Alb but not BO-Hb in Chinese workers exposed to benzene (18, 19). Given the potential utility of BO-Alb as a possible biomarker of human exposure to benzene, it is gratifying to see that our use of TFAA and methanesulfonic acid to produce PTTA can be applied to BO-Alb as well as to BOHb. Our results indicate that the assay can be used with 1-30 mg of purified Alb with a coefficient of variation of 8% (n ) 6) when 5 mg of protein is used (our standard procedure). We further estimate the limit of detection for BO-Alb to be approximately 270 fmol/5 mg of Alb (54 pmol/g of Alb) based on analysis of 5 mg of blank Sigma Alb combined with 2 pmol of S-phenylcysteine in our standard assay at a 3:1 signal-to-noise ratio. This study provides evidence that, when cysteinebound BO-Hb and BO-Alb are measured, protein purification should include dialysis to remove low-molecularweight compounds, the most important of which may be glutathione conjugates. Our results, shown in Tables 5 and 6, indicate that neither purification of Hb by Sephadex chromatography (21) nor purification of Alb without dialysis (19) satisfactorily removes these low-molecularweight compounds. Indeed, dialysis resulted in apparent adduct levels which were about one-half of those without dialysis. Considering the ample evidence that S-phenylmercapturic acid arises from benzene exposure (44), it is reasonable to postulate that much of the low-molecularweight material removed by dialysis could include glutathione conjugates. Moreover, exhaustive dialysis itself does not appear to affect BO-Alb levels since samples which had been dialyzed twice were indistinguishable from those which had been dialyzed one time.

The suggestion of a greater than proportional increase in the formation of BO-Alb adducts for [BO]0 > 36.8 µM BO (Figure 3B) provides indirect evidence of significant depletion of reactive nucleophiles (including glutathione) at high BO concentrations which resulted in increased rates of reaction of BO with Alb. The fact that the production of BO-Hb in the same blood samples was linear with [BO]0 (Figure 3A) presumably reflects the high concentration of glutathione in red blood cells relative to that in the plasma. [One recent study (45) determined that glutathione levels in normal human adult red blood cells (2.6 mM) were 2000 times higher than the corresponding plasma levels (1.2 µM).] In a study of naphthalene-dosed mice, Cho et al. (46) reported similar findings regarding production of Alb and Hb adducts, which they also attributed to depletion of glutathione. As noted in the Introduction, a major impetus for this work was to predict protein doses (values of YDBO) under eq 3, based upon measurements of BO-Hb or BO-Alb. Such applications require the appropriate second-order reaction rate constants (kBO-Y), estimates of which are given in Table 3 for Hb and Alb from rats, mice, and humans. To illustrate the use of this relationship, we first compare values of YDBO predicted from measurements of BO-Hb and BO-Alb following administration of a single oral dosage of 50-400 mg of benzene/kg of body weight to F344 rats in the study of McDonald et al. (21). The estimated levels of BO-Hb and BO-Alb, shown in Table 4, were used to predict the values of YDBO. For BO-Hb measured at 7.74 nmol/g of Hb in animals to which 50 mg of benzene/kg of body weight had been administered, the predicted dose of BO to Hb would have been:

[BO-Hb] ) [Hb]kBO-Y 7.74 nmol g of Hb-h × ) 5.03 × 103 nM BO-h -4 g of Hb 15.4 × 10 L

HbDBO

=

The results of these calculations, shown in Table 7, indicate that estimates of HbDBO were between about 6 and 12 times those of AlbDBO at a given benzene dosage, a difference that was statistically significant (z-test, p < 0.0001). Since the blood from this experiment had been obtained from the rats 24 h after administration of benzene, the loss of adducts due to protein turnover should have been negligible [ter ) 66 days (47), kAlb ) 0.0154-0.008 h-1 (48)] and, therefore, the doses of BO to Hb and Alb should have been the same. The fact that HbDBO was greater than AlbDBO suggests that, in fact, the dose of BO was higher to Hb (in the erythrocyte) than to Alb (in the serum). This opens the possibility that adducts could have been generated in situ within the

Benzene Oxide Hemoglobin and Albumin Adducts

erythrocyte due to oxidation of benzene, as has been reported previously for Hb-mediated oxidation of styrene to styrene 7,8-oxide (49) and aniline to p- and o-aminophenol (50). The above values of the blood dose of BO, predicted from eq 3, can be compared with a conventional estimate of the dose (i.e., that of area under curve0-∞, AUC0-∞), obtained from direct measurements of BO in the blood of 19 F344 rats at various times from 0 to 24 h after administering 400 mg of benzene/kg of body weight (8). In that experiment, the AUC0-∞ was estimated to have been 59.0 nmol of BO-h/kg of body weight, which is equivalent to 1.09 × 103 nM BO-h assuming a blood volume of 54.2 mL/kg of body weight for a F344 rat (51). This value of the AUC0-∞ is about one-half of the predicted value of AlbDBO for rats to which 400 mg of benzene/kg of body weight had been administered (2.62 × 103 nM BO-h, Table 7). Such a difference seems reasonable given a likely negative bias in the estimation of the AUC0-∞, due to losses of BO during collection and processing of the blood. However, the value of the AUC0-∞ is 14-fold lower than that of HbDBO for rats to which 400 mg of benzene/kg of body weight had been administered (14.7 × 103 nM BO-h, Table 7), a difference which is difficult to rationalize based upon our current knowledge. Again, we suspect that direct oxidation of benzene within the erythrocyte may have played some role in this disparity. Similar calculations can be used to predict the systemic dose of BO in humans based upon BO-Alb levels reported by Bechtold et al. (19) in 12 female workers with benzene exposures ranging from 4.4 to 23 ppm (mean ) 13.2 ppm or 42.1 mg/m3) and 9 unexposed controls in China. After measuring S-phenylcysteine in the Alb of these subjects, the authors reported a regression coefficient of 0.044 nmol of BO-Alb (g of Alb)-1 (ppm benzene)-1 and an intercept of 0.135 nmol of BO-Alb/g of Alb. Adjusting for the likely 1.87-fold overestimation of BO-Alb in these samples (since the Alb was not dialyzed, see Table 6), we estimate that an average of 0.238 nmol of BO-Alb/g of Alb in these exposed workers can be attributed to benzene exposure, that is,

(

0.044 nmol of BO-Alb (g of Alb)(ppm benzene) 1 0.135 nmol of BO-Alb × 1.87 0.238 nmol of BO-Alb ) g of Alb

13.2 ppm benzene ×

)

which corresponds to a point estimate of µBO-Alb/[Alb]. From eq 7 we predict the blood dose of BO during the 1/kAlb ) 714, h ) 4.25 weeks preceding blood collection to have been: AlbDBO

)

µBO-Alb

) [Alb]kBO-Alb 0.238 nmol of BO-Alb (g of Alb)[5.19 × 10-4 L (g of Alb-h)-1] ) 459 nM BO-h

During the same 4.25-week period, we predict the absorbed dose of inhaled benzene to have been 34.4 mg/ kg of body weight, that is,

Chem. Res. Toxicol., Vol. 11, No. 4, 1998 309

42.1 mg of benzene 0.5 m3 × × h m3 0.48 (uptake) 8 h 5 days × × 4.25 weeks × d week 50 kg of bw 34.4 mg of benzene ) kg of bw

benzene dose )

assuming exposure for 8 h/day, 5 day/week for a 50-kg Asian female and an uptake factor of inhaled benzene of 0.48 in Asian females (52). Thus, the blood dose of BO received by these workers per unit of benzene dose would have been (459 nM BO-h)/(34.4 mg of benzene/kg of body weight) ) 13.3 nM BO-h (mg of benzene/kg of body weight)-1. Using data provided in Table 7, the dose of BO in F344 rats was 8.42 nM BO-h (mg of benzene/kg of body weight)-1, over the linear range for production of BO-Alb between 50 and 200 mg of benzene/kg of body weight. Therefore, the human dose of BO per unit of absorbed benzene, predicted from levels of BO-Alb reported by Bechtold et al., is similar to that predicted from adduct measurements in F344 rats. This finding seems reasonable given the comparable rates of metabolism of benzene [estimated via P450 2E1 activities/kg of bw (53)]. These two applications of the dosimetric relationships involving Alb adducts of BO provide plausible predictions of the tissue dose of BO in F344 rats and, arguably, in workers exposed to benzene. Such information will be useful for assessing exposures to benzene on the basis of BO-Alb levels as well as for elucidating the potential role which BO might play in benzene-induced carcinogenesis. The anomalous behavior of Hb adducts of BO suggests possible Hb-mediated oxidation of benzene that is currently being investigated in this laboratory.

Acknowledgment. The authors thank Dorthy Thompson for her help with blood collection. This work was supported by the National Institute of Environmental Health Sciences through Grant P42ES05948. The information in this document has been funded in part by the United States Environmental Protection Agency. It has been subjected to Agency review and approved for publication.

References (1) International Agency for Research on Cancer (1982) IARC Monographs on the Evaluation of the Carcinogenic Risk of Chemicals to Humans. Vol 29: Some Industrial Chemicals and Dyestuffs, IARC, Lyon, France. (2) National Toxicology Program, U.S. Department of Health and Human Services. (1986) Toxicology and Carcinogenesis Studies of Benzene in F344/N Rats and B6C3F1 Mice, National Toxicology Program Technology Report Series No. 289, National Institutes of Health, Washington, DC. (3) Sammett, D., Lee, E. W., Kocsis, J. J., and Snyder, R. (1979) Partial hepatectomy reduces both metabolism and toxicity of benzene. J. Toxicol. Environ. Health 5, 785-792. (4) Kenyon, E. M., Kraichely, R. E., Hudson, K. T., and Medinsky, M. A. (1996) Differences in rates of benzene metabolism correlate with observed genotoxicity. Toxicol. Appl. Pharmacol. 136, 4956. (5) Snyder, R., Witz, G., and Goldstein, B. D. (1993) The toxicology of benzene. Environ. Health Perspect. 100, 293-306. (6) Kalf, G. F. (1987) Recent advances in the metabolism and toxicity of benzene. CRC Crit. Rev. Toxicol. 18, 141-159. (7) Jerina, D. M., and Daly, J. W. (1974) Arene oxides: a new aspect of drug metabolism. Science 185, 573-582. (8) Lindstrom, A. B., Yeowell-O’Connell, K., Waidyanatha, S., Tornero-Velez, R., Golding, B. T., and Rappaport, S. M. (1997) Measurement of benzene oxide in the blood of rats following administration of benzene. Carcinogenesis 18, 1637-1641.

310 Chem. Res. Toxicol., Vol. 11, No. 4, 1998 (9) Snyder, R., and Hedli, C. C. (1996) An overview of benzene metabolism. Environ. Health Perspect. 104 (Suppl. 6), 1165-1171. (10) Bleasdale, C., Kennedy, G., McGregor, J. O., Nieschalk, J., Pearce, K., Watson, W. P., and Golding, B. T. (1996) Chemistry of muconaldehyde of possible relevance to the toxicology of benzene. Environ. Health Perspect. 104 (Suppl. 6), 1201-1209. (11) Chen, H., and Eastmond, D. (1995) Topoisomerase inhibition by phenolic metabolites: a potential mechanism for benzene’s clastogenic effects. Carcinogenesis 16, 1963-1969. (12) Snyder, R., and Kalf, G. F. (1994) A perspective on benzene leukemogenesis. Crit. Rev. Toxicol. 24, 177-209. (13) Smith, M. T. (1996) The mechanism of benzene-induced leukemia: a hypothesis and speculations on the causes of leukemia. Environ. Health Perspect. 104 (Suppl. 6), 1219-1225. (14) Eastmond, D. A., Smith, M. T., and Irons, R. D. (1987) An interaction on benzene metabolites reproduces the myelotoxicity observed with benzene exposure. Toxicol. Appl. Pharmacol. 74, 85. (15) Barale, R., Marrazzini, A., Betti, C., Vangelisti, V., Loprieno, N., and Barrai, I. (1990) Genotoxicity of two metabolites of benzene: phenol and hydroquinone show strong synergistic effects in vivo. Mutat. Res. 244, 15. (16) Guy, R. L., Hu, P., Witz, G., and Goldstein, B. D. (1991) Depression of iron uptake into erythrocytes in mice by treatment with the combined benzene metabolites p-benzoquinone, muconaldehyde, and hydroquinone. J. Appl. Toxicol. 11, 443. (17) Busby, W. F., Wang, J., Stevens, E. K., Padykula, R. E., Aleksejczyk, R. A., and Berchtold, G. A. (1990) Lung tumorigenicity of benzene oxide, benzene dihydrodiols and benzene diolepoxides in the BLU:Ha newborn mouse assay. Carcinogenesis 11, 14731478. (18) Bechtold, W. E., Sun, J. D., Birnbaum, L. S., Yin, S. N., Li, G. L., Kasicki, S., Lucier, G., and Henderson, R. F. (1992) S-Phenylcysteine formation in hemoglobin as a biological exposure index to benzene. Arch. Toxicol. 66, 303-309. (19) Bechtold, W. E., Willis, J. K., Sun, J. D., Griffith, W. C., and Reddy, T. V. (1992) Biological markers of exposure to benzene: S-phenylcysteine in albumin. Carcinogenesis 13, 1217-1220. (20) Yeowell-O’Connell, K., McDonald, T. A., and Rappaport, S. M. (1996) Analysis of hemoglobin adducts of benzene oxide by gas chromatography-mass spectrometry. Anal. Biochem. 237, 4955. (21) McDonald, T. M., Yeowell-O’Connell, K., and Rappaport, S. M. (1994) Comparison of protein adducts of benzene oxide and benzoquinone in the blood and bone marrow of rats and mice exposed to [14C/13C6]benzene. Cancer Res. 54, 4907-4914. (22) Norpoth, K. H., Mu¨ller, G., Schell, C., and Jorg, E. (1996) Phenylguanine found in urine after benzene exposure. Environ. Health Perspect. 104 (Suppl. 6), 1159-1163. (23) Glatt, H., Padyluka, R., Berchtold, G. A., Ludewig, G., Platt, K. L., Klein, J., and Oesch, F. (1989) Multiple activation pathways of benzene leading to products with varying genotoxic characteristics. Environ. Health Perspect. 82, 81-89. (24) Kinoshita, T., Santella, R., Pulkrabek, P., and Jeffery, A. M. (1981) Benzene oxide: genetic toxicity. Mutat. Res. 91, 99-102. (25) Stark, A.-A., and Rastetter, W. H. (1996) Structure-activity relationships in the mutagenicity and cytotoxicity of putative metabolites and related analogues of benzene derived from the valence tautomers benzene oxide and oxepin. Environ. Mol. Mutagen. 28, 284-293. (26) Yeowell-O’Connell, K., Rothman, N., Smith, M. T., Hayes, R. B., Gui-Lin, L., Bechtold, W., Dosemeci, M., Zhang, L., Yin, S. Y., and Rappaport, S. M. (1996) Hemoglobin adducts of benzene oxide in workers exposed to high levels of benzene in China. Proc. Am. Assoc. Cancer Res. 37, 251. (27) Gillard, J. R., Newlands, M. J., Bridson, J. N., and Burnell, D. J. (1991) Pi-Facial stereoselectivity in the Diels-Alder reactions of benzene oxides. Can. J. Chem. 69, 1337-1343. (28) Ehrenberg, L., and Hussain, S. (1981) Genetic toxicity of some important epoxides. Mutat. Res. 86, 1-113. (29) Osterman-Golkar, S., Ehrenberg, L., Segerba¨ck, D., and Ha¨llstro¨m, I. (1976) Evaluation of genetic risks of alkylating agents. II. Haemoglobin as a dose monitor. Mutat. Res. 34, 1-10. (30) Ehrenberg, L., and To¨rnqvist, M. (1991) Human health risk assessment and biological reactive intermediates: hemoglobin binding. Adv. Exp. Med. Biol. 283, 641-647. (31) Guyton, A. C. (1977) Red blood cells, white blood cells, and the resistance of the body to infection. Basic Human Physiology: Normal Function and the Mechanisms of Disease, p 60, W. B.

Lindstrom et al. Saunders Co., Philadelphia, PA. (32) Peters, T. (1970) Serum albumin. Adv. Clin. Chem. 13, 37-111. (33) The Jackson Laboratory. (1966) In Biology of the Laboratory Mouse. Physiological Characteristics (Green, E. L., Ed.) McGrawHill, New York. (34) International Commision on Radiological Protection. (1975) Reports of the Task Group on Reference Man, pp 37-40, Pergamon Press, Oxford, U.K. (35) Ringler, D. H., and Dabich, L. (1979) Chapter 5-Hematology and Clinical Biochemistry. In The Laboratory Rat Volume I Biology and Diseases (Baker, H. J., Lindsey, J. R., and Weisbroth, S. H., Eds.) pp 105-121, Academic Press, New York. (36) The Charles River Breeding Laboratories. (1984) Baseline Hematology and Clinical Chemistry Values for Charles River Fischer344 Rats-CDF (F344) CrlBR as a Function of Sex and Age, Technical Bulletin 3(1), The Charles River Breeding Laboratories, Wilmington, MA. (37) Rappaport, S. M., Ting, D., Jin, Z., Yeowell-O’Connell, K., Waidyanatha, S., and McDonald, T. (1993) Application of Raney nickel to measure adducts of styrene oxide with hemoglobin and albumin. Chem. Res. Toxicol. 6, 238-244. (38) Yeowell-O’Connell, K., Pauwels, W., Severi, M., Jin, Z., Walker, M. R., and Rappaport, S. M. (1997) Comparison of styrene-7,8oxide adducts formed via reaction with cysteine, N-terminal valine and carboxylic acid residues in human, mouse and rat blood. Chem.-Biol. Interact. 106, 67-85. (39) Segerba¨ck, D. (1990) Reaction products in hemoglobin and DNA after in vitro treatment with ethylene oxide and N-(2-hydroxyethyl)-N-nitrosourea. Carcinogenesis 11, 307-312. (40) Hamboek, H., Fischer, R. W., Di Iorio, E. E., and Winterhalter, K. H. (1981) The binding of s-triazine metabolites to rodent hemoglobins appears irrelevant to other species. Mol. Pharmacol. 20, 579-584. (41) McDonald, T. A., Waidyanatha, S., and Rappaport, S. M. (1993) Measurement of adducts of benzoquinone with hemoglobin and albumin. Carcinogenesis 14, 1927-1932. (42) Rothschild, M. A., Oratz, M., and Schreiber, S. S. (1988) Serum albumin. Hepatology 8, 385-401. (43) Sargent, T. D., Yang, M., and Bonner, J. (1981) Nucleotide sequence of cloned rat serum albumin messanger RNA. Proc. Natl. Acad. Sci. U.S.A. 78, 243-246. (44) Van Sittert, N. J., Boogaard, P. J., and Beulink, G. D. J. (1993) Application of the urinary S-phenylmercapturic acid test as a biomarker for low levels of exposure to benzene in industry. Br. J. Indust. Med. 50, 460-469. (45) Hogarth, L., English, M., Price, L., Wyllie, R., Pearson, A. D. J., and Hall, A. G. (1996) The effect of treatment with high dose melphalan, cisplatin or carboplatin on levels of glutathione in plasma, erythrocytes, mononuclear cells and urine. Cancer Chemother. Pharmacol. 37, 479-485. (46) Cho, M., Jedrychowski, R., Hammock, B., and Buckpitt, A. (1994) Reactive naphthalene metabolite binding to hemoglobin and albumin. Fundam. Appl. Toxicol. 22, 26-33. (47) Derelanko, M. J. (1987) Determination of erythrocyte life span in F-344, Wistar, and Sprague-Dawley rats using a modification of the [3H] diisopropylfluorophosphate ([3H]DFP) method. Fundam. Appl. Toxicol., 9, 271-276. (48) Schultze, H. E., and Heremans, J. F. (1966) In Molecular Biology of Human Proteins with Special Reference to Plasma Proteins (Schultze, H.E., Heremans, J. F., Eds.) p 479, Elsevier, Amsterdam. (49) Belvedere, G., and Samaja, M. (1994) Oxidation of olefins catalyzed by hemoglobin. Methods Enzymol. 231, 598-621. (50) Barton, H. A., and Marletta, M. M. (1994) Comparison of aniline hydroxylation by hemoglobin and microsomal cytochrome P450 using stable isotopes. Toxicol. Lett. 70, 147-153. (51) Donaldson, H. H. (1924) The Rat: Data and Reference Tables for the Albino Rat (Mus Norvegicus Albinus) and the Norway Rat (Mus Norvegicus), Henry H. Donaldson, Philadelphia, PA. (52) Nomiyama, K., and Nomiyama, H. (1974) Respiratory retention, uptake and excretion of organic solvents in man. Int. Arch. Arbeitsmed. 32, 75-83. (53) Seaton, M. J., Schlosser, P. M., and Medinsky, M. A. (1995) In vitro conjugation of benzene metabolites by human liver: potential influence of interindividual variability on benzene toxicity. Carcinogenesis 16, 1519-1527.

TX9701788