Immunochemical analysis of quinol-thioether-derived covalent protein adducts in rodent species sensitive and resistant to quinol-thioether-mediated nephrotoxicity

Quinone-thioether metabolites of hydroquinone play a dual role in promoting a vicious cycle of ROS generation: in vitro and in silico insights

Humans are exposed to hydroquinone (HQ) via diet, smoking, occupation, and even via inhalation of polluted air. Given its preferential distribution in kidney and liver, the impact of biotransformation on the nephrotoxicity and hepatotoxicity of HQ was evaluated. Indeed, HQ and its metabolites, benzoquinone, and quinone-thioethers (50, 100, 200, and 400 μM) all induced ROS-dependent cell death in both HK-2, a human kidney proximal epithelial cell line, and THLE-2, a human liver epithelial cell line, in a concentration-dependent manner. For a deeper insight into the biological mechanism of ROS stimulation, the bioinformatics database was reviewed. Intriguingly, 163 proteins were currently reported to form co-crystal complex with benzoquinone analogs, a large proportion of which are closely related to ROS generation. After a thorough assessment of the interaction affinity and binding energy, three key mitochondrial proteins that are particularly involved in electric transport, namely, cytochrome BC1, succinate dehydrogenase, and sulfide:quinone oxidoreductase, were highlighted for further verification. Their binding affinity and the action pattern were explored and validated by molecular docking and molecular dynamics simulations. Remarkably, quinone-thioether metabolites of HQ afforded high affinity to the above proteins that purportedly cause a surge in the generation of ROS. Therefore, HQ can be further converted into quinone-thioethers, which on one hand can function as substrates for redox cycling, and on the other hand may afford high affinity with key proteins evolved in mitochondrial electron transport system, leading to a vicious cycle of ROS generation. The combined data provide a prospective insight into the mechanisms of ROS motivation, expanding HQ-mediated toxicology profiles.

One-dimensional western blotting coupled to LC-MS/MS analysis to identify chemical-adducted proteins in rat urine

The environmental toxicant hydroquinone (HQ) and its glutathione conjugates (GSHQs) cause renal cell necrosis via a combination of redox cycling and the covalent adduction of proteins within the S₃ segment of the renal proximal tubules in the outer stripe of the outer medulla (OSOM). Following administration of 2-(glutathion-S-yl)HQ (MGHQ) (400 μmol/kg, i.v., 2 h) to Long Evans (wild-type Eker) rats, Western analysis utilizing an antibody specific for quinol-thioether metabolites of HQ revealed the presence of large amounts of chemical-protein adducts in both the OSOM and urine. By aligning the Western blot film with a parallel gel stained for protein, we can isolate the adducted proteins for LC-MS/MS analysis. Subsequent database searching can identify the specific site(s) of chemical adduction within these proteins. Finally, a combination of software programs can validate the identity of the adducted peptides. The site-specific identification of covalently adducted and oxidized proteins is a prerequisite for understanding the biological significance of chemical-induced posttranslational modifications (PTMs) and their toxicological significance.

Identification of chemical-adducted proteins in urine by multi-dimensional protein identification technology (LC/LC-MS/MS)

Recent technological advancements in mass spectrometry facilitate the detection of chemical-induced posttranslational modifications (PTMs) that may alter cell signaling pathways or alter the structure and function of the modified proteins. To identify such protein adducts (Kleiner et al., Chem Res Toxicol 11:1283-1290, 1998), multi-dimensional protein identification technology (MuDPIT) has been utilized. MuDPIT was first described by Link et al. as a new technique useful for protein identification from a complex mixture of proteins (Link et al., Nat Biotechnol 17:676-682, 1999). MuDPIT utilizes two different HPLC columns to further enhance peptide separation, increasing the number of peptide hits and protein coverage. The technology is extremely useful for proteomes, such as the urine proteome, samples from immunoprecipitations, and 1D gel bands resolved from a tissue homogenate or lysate. In particular, MuDPIT has enhanced the field of adduct hunting for adducted peptides, since it is more capable of identifying lesser abundant peptides, such as those that are adducted, than the more standard LC-MS/MS. The site-specific identification of covalently adducted proteins is a prerequisite for understanding the biological significance of chemical-induced PTMs and the subsequent toxicological response they elicit.

Utilization of MALDI-TOF to determine chemical-protein adduct formation in vitro

Biological reactive intermediates can be created via metabolism of xenobiotics during the process of chemical elimination. They can also be formed as by-products of cellular metabolism, which produces reactive oxygen and nitrogen species. These reactive intermediates tend to be electrophilic in nature, which enables them to interact with tissue macromolecules, disrupting cellular signaling processes and often producing acute and chronic toxicities. Quinones are a well-known class of electrophilic species. Many natural products contain quinones as active constituents, and the quinone moiety exists in a number of chemotherapeutic agents. Quinones are also frequently formed as electrophilic metabolites from a variety of xeno- and endobiotics. Hydroquinone (HQ) is present in the environment from various sources, and it is also a known metabolite of benzene. HQ is converted in the body to 1,4-benzoquinone, which subsequently gives rise to hematotoxic and nephrotoxic quinone-thioether metabolites. The toxicity of these metabolites is dependent upon their ability to arylate proteins and to produce oxidative stress. Protein tertiary structure and protein amino acid sequence combine to determine which proteins are targets of these electrophilic quinone-thioether metabolites. We have used cytochrome c and model peptides to view adduction profiles of quinone-thioether metabolites, and have determined by MALDI-TOF analysis that these electrophiles target specific residues within these model systems.

The cytoprotective effect of N-acetyl-L-cysteine against ROS-induced cytotoxicity is independent of its ability to enhance glutathione synthesis

2,3,5-Tris(glutathion-S-yl)-hydroquinone (TGHQ), a metabolite of hydroquinone, is toxic to renal proximal tubule epithelial cells. TGHQ retains the ability to redox cycle and create an oxidative stress. To assist in elucidating the contribution of reactive oxygen species (ROS) to TGHQ-induced toxicity, we determined whether the antioxidant, N-acetyl-L-cysteine (NAC), could protect human kidney proximal tubule epithelial cells (HK-2 cell line) against TGHQ-induced toxicity. NAC provided remarkable protection against TGHQ-induced toxicity to HK-2 cells. NAC almost completely inhibited TGHQ-induced cell death, mitochondrial membrane potential collapse, as well as ROS production. NAC also attenuated TGHQ-induced DNA damage and the subsequent activation of poly (ADP-ribose) polymerase and ATP depletion. Moreover, NAC significantly attenuated c-Jun NH2-terminal kinase and p38 mitogen-activated protein kinase phosphorylation induced by TGHQ. In contrast, NAC itself markedly increased extracellular regulated kinase1/2 (ERK1/2) activation, and the upstream mitogen-activated protein/extracellular signal-regulated kinase kinase inhibitor, PD-98059, only partially inhibited this activation, suggesting that NAC can directly activate ERK1/2 activity. However, although NAC is frequently utilized as a glutathione (GSH) precursor, the cytoprotection afforded by NAC in HK-2 cells was not a consequence of increased GSH levels. We speculate that NAC exerts its protective effect in part by directly scavenging ROS and in part via ERK1/2 activation.

Role of hydroquinone-thiol conjugates in benzene-mediated toxicity

Hydroquinone (HQ) is a metabolite of benzene, and in combination with phenol (PHE), reproduces benzene myelotoxicity. HQ readily oxidizes to 1,4-benzoquinone (1,4-BQ) followed by the reductive addition of glutathione (GSH). Subsequent cycles of oxidation and GSH addition give rise to a variety of mono-, and multi-GSH substituted conjugates. Following administration of PHE/HQ (1.1 mmol/kg/0.9 mmol/kg, ip) to male Sprague-Dawley (SD) rats, 2-(glutathion-S-yl)HQ [GS-HQ], 2,5-bis-(glutathion-S-yl)HQ [2,5-GS-HQ], 2,6-bis-(glutathion-S-yl)HQ [2,6-GS-HQ], and 2,3,5-tris-(glutathion-S-yl)HQ [2,3,5-GS-HQ] were all identified in bone marrow. 2-(Cystein-S-ylglycine)HQ [2-(CysGly)HQ], 2-(cystein-S-yl)HQ [2-(Cys)HQ], and 2-(N-acetylcystein-S-yl)HQ [2-(NACys)HQ] were also found in the bone marrow of PHE/HQ and benzene treated rats and mice, indicating the presence of an active mercapturic acid pathway within bone marrow. Moreover, 2,6-GS-HQ and 2,3,5-GS-HQ were hematotoxic when administered to rats. All of the HQ-GSH conjugates retain the ability to redox cycle and generate reactive oxygen species (ROS), and to arylate target proteins. Recent in vitro and in vivo studies in our laboratory revealed lysine and arginine residues as primary targets of 1,4-BQ, GS-HQ and 2-(NACys)HQ adduction. In contrast 1,4-BQ-adduction of cysteine residues may be a transient interaction, where physiological conditions dictate adduct stability. The generation of ROS and alkylation of proteins may both contribute to benzene-mediated myelotoxicity, and the two processes may be inter-dependent. However, the precise molecular mechanism by which benzene and HQ-GSH conjugates induce hematotoxicity remains to be determined. Within 18h of administration of PHE/HQ to SD rats a significant decrease in blood lymphocyte count was observed. At this early time point, erythrocyte counts and hemoglobin concentrations remained within the normal range. Concomitant with the decrease in lymphocyte count, western blot analysis of bone marrow lysate, using HQ-GSH and 4-hydroxy-2-nonenal (4HNE) specific antibodies, revealed the presence of HQ-GSH- and 4HNE-derived protein adducts. Identification of these adducts is required before the functional significance of such protein modifications can be determined.

Protein electrophile-binding motifs: lysine-rich proteins are preferential targets of quinones

Quinones represent an important class of endogenous compounds such as neurotransmitters and coenzyme Q10, electrophilic xenobiotics, and environmental toxicants that have known reactivity based on their ability to redox cycle and generate oxidative stress, as well as to alkylate target proteins. It is likely that topological, chemical, and physical features combine to determine which proteins become targets for chemical adduction. Chemical-induced post-translational modification of certain critical proteins causes a change in structure/function that contributes to the toxicological response to chemical exposure. In this study, we have identified a number of proteins that are modified by quinone-thioethers after administration of 2-(glutathion-S-yl)HQ. Parallel one-dimensional gel electrophoresis was performed, and the Coomassie-stained gel was aligned with the corresponding Western blot, which was probed for adductions. Immunopositive bands were then subjected to trypsin digestion and analyzed via liquid chromatography/tandem mass spectrometry. The proteins that were subsequently identified contained a higher than average (9.7 versus 5.5%) lysine content and numerous stretches of lysine run-ons, which is a presumed electrophile binding motif. Approximately 50% of these proteins have also been identified as targets for electrophilic adduction by a diverse group of chemicals by other investigators, implying overlapping electrophile adductomes. By identifying a motif targeted by electrophiles it becomes possible to make predictions of proteins that may be targeted for adduction and possible sites on these proteins that are adducted. An understanding of proteins targeted for adduction is essential to unraveling the toxicity produced by these electrophiles.

Protein damage by reactive electrophiles: targets and consequences

It has been 60 years since the Millers first described the covalent binding of carcinogens to tissue proteins. Protein covalent binding was gradually overshadowed by the emergence of DNA adduct formation as the dominant paradigm in chemical carcinogenesis but re-emerged in the early 1970s as a critical mechanism of drug and chemical toxicity. Technology limitations hampered the characterization of protein adducts until the emergence of mass spectrometry-based proteomics in the late 1990s. The time since then has seen rapid progress in the characterization of the protein targets of electrophiles and the consequences of protein damage. Recent integration of novel affinity chemistries for electrophile probes, shotgun proteomics methods, and systems modeling tools has led to the identification of hundreds of protein targets of electrophiles in mammalian systems. The technology now exists to map the targets of damage to critical components of signaling pathways and metabolic networks and to understand mechanisms of damage at a systems level. The implementation of sensitive, specific analyses for protein adducts from both xenobiotic-derived and endogenous electrophiles offers a means to link protein damage to clinically relevant health effects of both chemical exposures and disease processes.

Accumulation of Neurotoxic Thioether Metabolites of 3,4-(±)-Methylenedioxymethamphetamine in Rat Brain

The serotonergic neurotoxicity of 3,4-(+/-)-methylenedioxymethamphetamine (MDMA) appears dependent upon systemic metabolism because direct injection of MDMA into the brain fails to reproduce the neurotoxicity. MDMA is demethylenated to the catechol metabolite N-methyl-alpha-methyldopamine (N-Me-alpha-MeDA). Thioether (glutathione and N-acetylcysteine) metabolites of N-Me-alpha-MeDA are neurotoxic and are present in rat brain following s.c. injection of MDMA. Because multidose administration of MDMA is typical of drug intake during rave parties, the present study was designed to determine the effects of multiple doses of MDMA on the concentration of neurotoxic thioether metabolites in rat brain. Administration of MDMA (20 mg/kg s.c.) at 12-h intervals for a total of four injections led to a significant accumulation of the N-Me-alpha-MeDA thioether metabolites in striatal dialysate. The area under the curve (AUC)(0-300 min) for 5-(glutathion-S-yl)-N-Me-alpha-MeDA increased 33% between the first and fourth injections and essentially doubled for 2,5-bis-(glutathion-S-yl)-N-Me-alpha-MeDA. Likewise, accumulation of the mercapturic acid metabolites was reflected by increases in the AUC(0-300 min) for both 5-(N-acetylcystein-S-yl)-N-Me-alpha-MeDA (35%) and 2,5-bis-(N-acetylcystein-S-yl)-N-Me-alpha-MeDA (85%), probably because processes for their elimination become saturated. Indeed, the elimination half-life of 5-(N-acetylcystein-S-yl)-N-Me-alpha-MeDA and 2,5-bis-(N-acetylcystein-S-yl)-N-Me-alpha-MeDA increased by 53 and 28%, respectively, between the first and third doses. Finally, although the C(max) values for the monothioether conjugates were essentially unchanged after each injection, the values increased by 38 and approximately 50% for 2,5-bis-(glutathion-S-yl)-N-Me-alpha-MeDA and 2,5-bis-(N-acetylcystein-S-yl)-N-Me-alpha-MeDA, respectively, between the first and fourth injections. The data indicate that neurotoxic metabolites of MDMA may accumulate in brain after multiple dosing.

Cytosolic and nuclear protein targets of thiol-reactive electrophiles

Reactive electrophiles formed from toxic drugs and chemicals and by endogenous oxidative stress covalently modify proteins. Although protein covalent binding is thought to initiate a variety of adaptive and toxic responses, the identities of the protein targets are generally unknown, as are protein structural features that confer susceptibility to modification. We have analyzed the protein targets in nuclear and cytoplasmic proteomes from HEK293 cells treated in vitro with two biotin-tagged, thiol-reactive electrophiles, (+)-biotinyl-iodoacetamidyl-3, 6-dioxaoctanediamine (PEO-IAB) and 1-biotinamido-4-(4'-[maleimidoethylcyclohexane]-carboxamido)butane (BMCC). Biotinylated peptides were captured by affinity enrichment using neutravidin beads, and the adducted peptides were then analyzed by multidimensional liquid chromatography-tandem mass spectrometry. A total of 897 adducts were mapped to different cysteine residues in 539 proteins. Adduction was selective and reproducible, and > 90% of all adducted proteins were modified at only one or two sites. A core group of 125 cysteines (14% of the total) was consistently modified by both electrophiles. Selective modification of several protein domain structures and motifs indicates that certain protein families are particularly susceptible to alkylation. This approach can be extended to studies of other protein-damaging oxidants and electrophiles and can provide new insights into targets and consequences of protein damage in toxicity and disease.

Changes in gene expression during chemical-induced nephrocarcinogenicity in the Eker rat

Hydroquinone (HQ) is a rodent carcinogen and a potential human carcinogen. Glutathione conjugation of HQ enhances its biological reactivity, and 2,3,5-tris-(glutathion-S-yl)hydroquinone (TGHQ) is a potent nephrotoxicant and nephrocarcinogen in the Eker rat. Moreover, a single exposure of primary epithelial cells derived from Eker rat kidneys to TGHQ transforms these cells into an immortalized phenotype (quinol-thioether transformed rat renal epithelial (QT-RRE) cells). The Eker rat bears a mutation in one allele of the tuberous sclerosis-2 (Tsc-2) tumor suppressor gene, which predisposes the animals to the development of spontaneous and chemical-induced renal cell carcinoma. Thus, the Eker rat provides a unique model for elucidating the mechanisms of renal tubular epithelial carcinogeneisis. cDNA microarray analysis of QT-RRE3 cells and of tumor tissue derived from the kidneys of Eker rats treated with TGHQ revealed alterations (by threefold or greater) in the expression of a total of 80 genes. Fifteen percent of these genes exhibited similar expression patterns in both QT-RRE cells and tumor tissue. The differentially expressed genes primarily participate in three major areas: (1) signal transduction or in the regulation of signal transduction (extracellular signal regulated kinase 2 (ERK2); protein kinase CK2; protein kinase B; c-jun; NF-kappaB; ras-related GTPases; annexins), (2) stress response, tissue remodeling, and DNA repair (glutathione-S-transferases; procollagen c proteinase enhancer; plasminogen activator; tissue inhibitor of metalloprotease 3; apurinic/apyrimidic endonuclease), and (3) electron transport and energy homeostasis (cytochrome c oxidase subunits). The changes in the expression of many of these genes was confirmed by reverse transcription (RT)-polymerase chain reactions (PCR) using primers specific for the differentially expressed genes. As an example, the annexin I and II genes, implicated in signal transduction, were highly induced in tumor tissue and also in dysplastic lesions isolated from the kidneys of rats treated chronically with TGHQ. The annexin I and II proteins were also upregulated in tumor tissue, which probably play an important role in TGHQ-induced nephrocarcinogenesis. Moreover, in the present study, a tumorigenicity assay using athymic nude mice revealed that QT-RRE cell lines formed tumors when injected in the subcutis of nude mice, providing evidence that the cells are malignantly transformed. Histopathological analysis further indicated that the tumors were composed of neoplastic cells, resembling renal carcinoma cells with varying degrees of atypia, with the presence of apoptotic and mitotic figures.

Cell proliferation is insufficient, but loss of tuberin is necessary, for chemically induced nephrocarcinogenicity

Although 2,3,5-tris-(glutathion-S-yl)hydroquinone (TGHQ; 2.5 micromol/kg ip) markedly increased cell proliferation within the outer stripe of the outer medulla (OSOM) of the kidney in both wild-type (Tsc2(+/+)) and mutant Eker rats (Tsc2(EK/+)), only TGHQ-treated Tsc2(EK/+) rats developed renal tumors, indicating that cell proliferation per se was not sufficient for tumor development. Tuberin expression was initially induced within the OSOM after TGHQ treatment but was lost within TGHQ-induced renal tumors. High extracellular signal-regulated kinase (ERK) activity occurred in the OSOM of Tsc2(EK/+) rats at 4 mo and in TGHQ-induced renal tumors. Cyclin D1 was also highly expressed in TGHQ-induced renal tumors. Reexpression of Tsc2 in tuberin-negative cells decreased ERK activity, consistent with the growth-suppressive effects of this tumor suppressor gene. Thus 1) stimulation of cell proliferation after toxicant insult is insufficient for tumor formation; 2) tuberin induction after acute tissue injury suggests that Tsc2 is an acute-phase response gene, limiting the proliferative response after injury; and 3) loss of Tsc2 gene function is associated with cell cycle deregulation.

The interplay of glutathione-related processes in antioxidant defense

This review summarizes current knowledge on glutathione (GSH) associated cellular processes that play a central role in defense against oxidative stress. GSH itself is a critical factor in maintaining the cellular redox balance and has been demonstrated to be involved in regulation of cell signalling and repair pathways. Enhanced expression of various enzymes involved in GSH metabolism, including glutathione peroxidases, γ-glutamyl cysteinyl synthetase (γ-GCS), glutathione S-transferases (GST) and membrane proteins belonging to the ATP-binding cassette family, such as the multidrug resistance associated protein, have all been demonstrated to play a prominent role in cellular resistance towards oxidative stress. This review stresses the fact that aco-ordinateinterplay between these systems is essential for efficient protection against oxidative stress.

Transformation of kidney epithelial cells by a quinol thioether via inactivation of the tuberous sclerosis-2 tumor suppressor gene

Although hydroquinone (HQ) is a rodent carcinogen, because of its lack of mutagenicity in standard bacterial mutagenicity assays it is generally considered a nongenotoxic carcinogen. 2,3,5-Tris-(glutathion-S-yl)HQ (TGHQ) is a potent nephrotoxic metabolite of HQ that may play an important role in HQ-mediated nephrocarcinogenicity. TGHQ mediates cell injury by generating reactive oxygen species and covalently binding to tissue macromolecules. We determined the ability of HQ and TGHQ to induce cell transformation in primary renal epithelial cells derived from the Eker rat. Eker rats possess a germline inactivation of one allele of the tuberous sclerosis-2 (Tsc-2) tumor suppressor gene that predisposes the animals to renal cell carcinoma. Treatment of primary Eker rat renal epithelial cells with HQ (25 and 50 microM) or TGHQ (100 and 300 microM) induced 2- to 4-fold and 6- to 20-fold increases in cell transformation, respectively. Subsequently, three cell lines (The QT-RRE 1, 2, and 3) were established from TGHQ-induced transformed colonies. The QT-RRE cell lines exhibited a broad range of numerical cytogenetic alterations, loss of heterozygosity at the Tsc-2 gene locus, and loss of expression of tuberin, the protein encoded by the Tsc-2 gene. Only heterozygous (Tsc-2(EK/+)) kidney epithelial cells were susceptible to transformation by HQ and TGHQ, as wild-type cells (Tsc-2(+/+)) showed no increase in transformation frequency over background levels following chemical exposure. These data indicate that TGHQ and HQ are capable of directly transforming rat renal epithelial cells and that the Tsc-2 tumor suppressor gene is an important target of TGHQ-mediated renal epithelial cell transformation.

Paracetamol (acetaminophen)-induced toxicity: molecular and biochemical mechanisms, analogues and protective approaches

An overview is presented on the molecular aspects of toxicity due to paracetamol (acetaminophen) and structural analogues. The emphasis is on four main topics, that is, bioactivation, detoxication, chemoprevention, and chemoprotection. In addition, some pharmacological and clinical aspects are discussed briefly. A general introduction is presented on the biokinetics, biotransformation, and structural modification of paracetamol. Phase II biotransformation in relation to marked species differences and interorgan transport of metabolites are described in detail, as are bioactivation by cytochrome P450 and peroxidases, two important phase I enzyme families. Hepatotoxicity is described in depth, as it is the most frequent clinical observation after paracetamol-intoxication. In this context, covalent protein binding and oxidative stress are two important initial (Stage I) events highlighted. In addition, the more recently reported nuclear effects are discussed as well as secondary events (Stage II) that spread over the whole liver and may be relevant targets for clinical treatment. The second most frequent clinical observation, renal toxicity, is described with respect to the involvement of prostaglandin synthase, N-deacetylase, cytochrome P450 and glutathione S-transferase. Lastly, mechanism-based developments of chemoprotective agents and progress in the development of structural analogues with an improved therapeutic index are outlined.

Bioreactivity of glutathionyl hydroquinone with implications to benzene toxicity

Glutathionyl hydroquinone (GHQ), a highly reactive metabolite of benzene, has been implicated as a causative intermediate of benzene toxicity. To substantiate, the bioreactivity of GHQ was investigated under in vitro and in vivo conditions using end points, characteristic of benzene toxicity. Under in vitro conditions, the presence of GHQ: (a) linearly increased the release of aldehydic products from L-glutamate or deoxyuridine at GHQ concentrations of 5-25 microM and from rat liver homogenates at GHQ concentrations of 50-250 microM; (b) cleaved plasmid pUC 18 supercoiled DNA through a single strand nick to yield open circular relaxed DNA, and through a double strand cut to give out linear DNA at GHQ concentrations of 25-200 microM, with evidence of protection by catalase and superoxide dismutase; and (c) induced cross-linking and polymerization of lymphocyte nuclear DNA through in situ generation of GHQ, which was protected by pretreatment of lymphocytes with N-ethylmaleimide. In vivo exposure of Swiss albino mice to GHQ (100 mg/kg, intraperitoneally once daily for 30 days) resulted in significant increase of liver weight and inhibition of mitotic index in the bone marrow. The other test parameters, namely spleen weight, hematological indices, hepatic sulphahydryl content and nonenzymatic lipid peroxidation, and chromosomal aberrations in the bone marrow were, however, unaffected by GHQ treatment. The observations indicate pro-oxidant and cytotoxic potential of GHQ, mediated by the reactive oxygen species generated during the course of its auto-oxidation. Bioreactivity of GHQ with cellular macromolecules in vitro and inhibition of mitotic index of bone marrow on in vivo exposure have relevance to benzene toxicity, although in situ generation of GHQ at the site of action appears critical in bringing about hematological and chromosomal effects that were probably spared due to rapid metabolic disposition and, consequently, poor bioavailability of intraperitoneally administered GHQ.