Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Chloroform is a ubiquitous atmospheric and water contaminant. Beside its extensive use as a solvent in industrial processes, it is formed as a by-product during the chlorination of water intended for human use and paper bleaching. Due to its volatility, chloroform can be easily released from waste or chlorinated waters into the atmosphere or in the ambient air. Therefore, a large part of the human population may be chronically exposed to chloroform from different sources, although drinking water has been considered the main one. Recently, routes of exposure other than oral consumption of chlorinated water have been evaluated as relevant. Indeed, some indoor activities, such as showering or bathing as well as cooking and housekeeping, may significantly contribute to total chloroform body burden through dermal and inhalation exposure (Wallace, 1997; Backer et al., 2000). The main concern for public health arose with the carcinogenic potential of chloroform in experimental animals (NCI, 1976; Jorgenson et al., 1985), which is strain-, species- and gender-specific (IPCS, 1994).

No studies of chronic toxicity or cancer incidence in humans exposed exclusively to chloroform are available. Nevertheless, there are a number of epidemiological studies on populations exposed to chlorinated drinking water, some of which evidenced a weak association between water consumption and cancer of the bladder and lower gastrointestinal tract (Hogan et al., 1979; Gottlieb et al., 1981; Cantor et al., 1998; Hidelsheim et al., 1998). Moreover, some adverse reproductive outcomes have been recently reported in humans exposed to chloroform via drinking water (Gallagher et al., 1998; Waller et al., 1998). However, the poor assessment of exposure and the concomitant presence of many water contaminants, including other trihalomethanes and disinfection by-products, makes it difficult to establish a causal link between chloroform itself and adverse effects in humans.

The required step for CHCl₃-induced toxicity is the cytochrome P450 (P450¹)-mediated bioactivation to reactive metabolites (IPCS, 1994; Constan et al., 1999) (Fig. 1). Extensive in vitro and in vivo studies on rodents have demonstrated that chloroform may be metabolized oxidatively to trichloromethanol, which spontaneously decomposes to the electrophilic phosgene (Mansuy et al., 1977; Pohl et al., 1977). COCl₂ is highly reactive and binds covalently to cell components containing nucleophilic groups, including proteins, phospholipid (PL) polar heads, and reduced glutathione (Pohl et al., 1981; Testai et al., 1990; De Biasi et al., 1992; Gemma et al., 1996; Vittozzi et al., 2000). Alternatively, phosgene may be hydrolyzed by reacting with water, yielding carbon dioxide and hydrochloric acid (Fig. 1). In anoxic or hypoxic conditions, chloroform may be reduced to dichloromethyl radical (Tomasi et al., 1985; Testai et al., 1995), which is able to bind to PL-fatty acyl chains (De Biasi et al., 1992; Gemma et al., 1996; Vittozzi et al., 2000) or to abstract a hydrogen atom from the biological environment, leading to dichloromethane (Testai et al., 1995) (Fig. 1). The relative ratio between the two pathways depends on the oxygen partial pressure, on chloroform concentration and is specie-, organ- and gender-specific (Ade et al., 1994; Gemma et al., 1996; Vittozzi et al., 2000).

View larger version (15K): [in this window] [in a new window]   Fig. 1.   The two pathways of chloroform bioactivation.

In rodents, it has been evidenced that at low concentration chloroform is oxidized by CYP2E1 (Testai et al., 1996; Constan et al., 1999). At higher CHCl₃ concentration, in the presence of oxygen, phosgene formation is catalyzed by CYP2B1/2, while in anoxic conditions dichloromethyl radical production seemed to be mediated by constitutive P450s (Testai et al., 1996). These direct observations supported a number of studies, showing the potentiation effect of a variety of CYP2E1 and 2B1/2 inducers on chloroform-induced toxicity (Sato et al., 1980; Stevens and Anders, 1981; Branchflower et al., 1983).

Only rare data on chloroform metabolism in human tissues have been published (Fry et al., 1972; Testai et al., 1991). No direct information on the human P450(s) involved in the reaction is available at present, although CYP2E1 is suspected as the major catalyst of chloroform metabolism. The knowledge of the P450(s) responsible for CHCl₃ bioactivation can provide useful information to identify the population group characterized by higher susceptibility to chloroform toxicity. Moreover, it would be possible to evaluate the environmental influence on susceptibility to chloroform-induced effects, through the study of metabolic interaction due to the combined exposure with other xenobiotics, affecting the rate of chloroform metabolism. Therefore, we have undertaken this study to characterize oxidative and reductive chloroform metabolism in human liver microsomes (HLM). In addition, we have identified the human P450(s) responsible for COCl₂ and ^·CHCl₂ formation, by using different experimental approaches, including cDNA-expressed enzymes, correlation studies in a panel of phenotyped HLM, and inhibition by either chemical-selective inhibitors or anti-human P450 antibodies (Abs).

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Products. [¹⁴C]Chloroform (3.1 mCi/mmol, radiochemical purity 99%) was obtained from PerkinElmer Life Sciences (Boston, MA); [¹⁴C]CHCl₃ was diluted with unlabeled chloroform (infrared purity) from Merck (Darmstadt, Germany) to achieve the specific radioactivity required. Liquid scintillation cocktails Lipoluma and Optiphase Hisafe were purchased from Lumac Systems A.G. (Basel, Switzerland) and PerkinElmer Wallac (Turku, Finland), respectively. NADP, glucose-6-phosphate (G6P), G6P-dehydrogenase, glucose oxidase, catalase were from Roche Ltd. (Basel, Switzerland). The P450 inhibitors, 4-methylpyrazole (4MPYR), orphenadrine (ORP), coumarin (COU), and troleandomycin (TAO) were purchased from Sigma-Aldrich (St. Louis, MO), whereas sulfaphenazole (SFN) was from Ultrafine (Manchester, UK). Monoclonal anti-human CYP2A6, CYP2B6, CYP2C8, CYP2E1, CYP3A4, and polyclonal anti-human CYP2C for immunoinhibition studies were purchased from BD Gentest Corp. (Woburn, MA). All other analytical grade chemicals were obtained from available commercial sources.

Recombinant Human P450s and Human Liver Samples. cDNA-derived human P450s (1A1, 1A2, 2A6, 2B6, 2C8, 2C9, 2C19, 2D6, 2E1, and 3A4) coexpressed with NADPH-cytochrome P450 reductase from human B-lymphoblastoid cell line AHH-1 TK+/ and control microsomes from untransfected cells were purchased from BD Gentest Corp. HLM were prepared from a single liver sample, kindly gifted by the University of Washington (Seattle, WA) (Kharasch and Thummel, 1993). Protein concentration (23.2 mg/ml) and P450 content (0.56 nmol/mg of protein) were measured according to the methods of Oyama and Eagle (1956) and Omura and Sato (1964), respectively. All the other HLM from hepatic biopsies either pooled from five donors or individual were purchased from Human Biologics International (Scottsdale, AZ). Protein concentration was 20 mg/ml; P450 content ranged between 0.14 and 0.82 nmol/mg of protein. Each sample was characterized by the supplier for monooxygenase activities by using selective model substrates for single P450. The range and mean value of monooxygenase activities of tested HLM is reported in Table 1.

In Vitro Bioactivation of [¹⁴C]chloroform and Covalent Binding of Its Metabolites to Microsomal PL. The standard incubation mixture (final volume 0.6 ml) contained microsomal protein coming from either HLM or lymphoblast-expressed-P450 preparations (2 mg/ml), EDTA (1 mM), G6P (2 mM), MgCl₂ (2 mM), G6P-dehydrogenase (1 U/ml), NADP (0.2 mM), and [¹⁴C]chloroform (0.033, 0.1, 0.5, 1, 2.5, 5 mM final concentration) in 50 mM Tris-HCl buffer, pH 7.4. The specific radioactivity (range 0.34-3.4 mCi/mmol) was adjusted, depending on CHCl₃ final concentration, to have the required sensitivity. Hypoxic conditions (1 and 5% pO₂) were obtained by flushing the ice-cold incubation mixture in sealed vessels with ultrapure N₂ or with a mixture N₂/O₂ (95:5) for 20 min, respectively, then kept for 3 min at 37°C in a shaking bath before starting the reaction by injecting NADP and [¹⁴C]chloroform (as a methanolic solution) at different concentrations. When incubations have to be carried out in anoxic conditions (0% pO₂), the ice-cold mixture was added of an oxygen-scavenging enzyme system, consisting of D-glucose (60 mM), glucose oxidase (12.5 U/ml) and catalase (300 U/ml), flushed with ultrapure nitrogen for 20 min and warmed for 10 min at 37°C in a shaking bath. Then reaction was started by the injection of NADP and [¹⁴C]chloroform dissolved in methanol at different concentrations. When cDNA-expressed CYP2E1 has to be tested, due to the strong inhibitory effect exerted by very low levels of organic solvents on this isoform, neat [¹⁴C]chloroform was injected: the lowest concentration obtainable was 15 mM. Further details of incubation procedure have been previously described (Testai et al., 1990).

Correlation Studies. PL-PHOS and PL-RAD levels were determined in standard incubations with 12 individual HLM, fully characterized by the supplier for single P450 activities (Table 1). In aerobic conditions, chloroform was tested both at low (0.1 mM) and high (5 mM) concentrations, chosen on the basis of K_m values obtained for the two components of chloroform oxidative metabolism; at 0% pO₂, only 5 mM CHCl₃ was used. Correlation coefficients were calculated by plotting the rate of PL-PHOS and PL-RAD formation versus the marker activity of each isoform.

Chemical Inhibition Assays. Isoform-selective P450 inhibitors were added to the standard incubation mixtures carried out with four different individual HLM at the following final concentrations: 50 µM COU, 10 µM ORP, 100 µM 4MPYR, 5 µM SFN, and 100 µM TAO. Chemical inhibitors were diluted in the incubation buffer and added to the mixture, except for COU, SFN, and TAO, which were added as methanolic solution (final methanol concentration 1%); in these latter cases, control incubations (set as 100% metabolic activity) contained the same amount of methanol. The presence of methanol caused <5% decrease with respect to activity measured in the absence of the solvent. Competitive inhibitors such as COU, 4MPYR, and SFN were added to the standard incubation mixture just before [¹⁴C]chloroform injection, then the reaction was started by the addition of NADP. Incubations containing mechanism-based inhibitors such as TAO and ORP were first preincubated with NADPH-generating system for 15 min at 37°C under aerobic conditions. Then [¹⁴C]chloroform was added to start the reaction for chloroform metabolism at 20% pO₂. Alternatively, to test the reductive metabolism in anoxic conditions, the mixture was placed on ice, the oxygen-scavenging system was added, the reaction tubes were sealed and flushed with nitrogen for 20 min, warmed at 37°C for 10 min, and injected with [¹⁴C]chloroform to start the reaction.

Immunoinhibition Assays. In immunoinhibition experiments, monoclonal Abs (both single and mixtures) or antiserum were preincubated at 4°C for 20 min and at room temperature for 30 min, respectively, with 1.2 mg of microsomal protein from HLM pooled from five donors. Following the protocol indicated by the supplier, the amounts of Abs or antiserum used were expected to produce maximal inhibition (>80%) of the specific enzymatic activity: 3 µl/100 µg of protein anti-CYP1A2 Ab and anti-CYP2B6 Ab; 5 µl/100 µg of protein anti-CYP2A6 Ab, anti-CYP2C8 Ab, and anti-CYP2E1 Ab; 10 µl/100 µg of protein anti-CYP3A4 Ab; 25 µl/100 µg of protein anti-CYP2C antiserum. In control incubations, the presence of monoclonal Abs or antiserum was replaced by Tris buffer or normal goat serum, respectively, and was set as 100% metabolic activity. Immunoinhibited microsomes were then added to the incubation mixture to test the in vitro bioactivation of [¹⁴C]chloroform following the standard operating procedure.

Calculation. The kinetic parameters K_mapp and V_maxapp were obtained from the Eadie-Hofstee plot. Linear regression was carried out by using the least-squares method to obtain the best fitting of single experimental data. In correlation studies, P450 activities measured in 12 HLM tested toward both chloroform and model substrates, exceeding ± 2 S.D. of the mean, have been considered outliers and were not included in the calculation of the Pearson's regression coefficient (r). Student's t test was used to assess the significance of results of inhibition studies; the cut-off of statistical significance was set at p < 0.05.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

In Vitro Chloroform Metabolism by HLM.

Substrate and time dependence In vitro characterization of chloroform metabolism was carried out by using three individual HLM, characterized by different relative P450 content. The ratio of CYP2E1-to-CYP2A6 was 14.4, 0.33, and 0.15 in HLM₁, HLM₂, and HLM₃, respectively, the latter HLM being characterized also by high levels of CYP1A2- and 2B6-related activities. The marker activities corresponding to the other P450s were in the average range expressed within the panel of individual HLM tested during the study. Results showed that all HLM samples can bioactivate [¹⁴C]CHCl₃ both oxidatively and reductively, as illustrated in Fig. 2 with HLM₃. The Eadie-Hoffstee plots clearly identified two components, characterized by different affinity for chloroform, suggesting multiple P450 involvement (Fig. 2A inset; Table 2). The K_mapp1 values ranged between 20 and 650 µM, being the lowest value obtained in HLM₁, characterized by a high level of CYP2E1 and low levels of CYP2A6. In HLM₂, which is characterized by an extremely low content of CYP2A6, the low affinity phase was less defined; indeed, both K_mapp2 and V_maxapp2 were one order of magnitude lower than kinetic parameters obtained with the other two samples. Although the HLM₂ and HLM₃ differed for their content in CYP1A2 and 2B6, their K_mapp2 and V_maxapp2 values resulted very similar, in line with their almost equal levels of CYP2A6.

View larger version (15K): [in this window] [in a new window]   Fig. 2.   Bioactivation of chloroform in individual human liver microsomes illustrated with HLM3: dependence on substrate concentration (A) and incubation time (B). PL-PHOS (squares) and PL-RAD (circles) formation were measured in aerobic and anoxic conditions, respectively; they represent the mean ± S.D. of three independent determinations and are expressed as nmol [14C] · mg PL1 · nmol P4501. The inset in panel A shows the Eadie-Hoffstee plot of PL-PHOS formation in aerobic conditions. When substrate concentration dependence was studied, incubation time was 20 min. Time dependence of chloroform oxidation and reduction was tested at 5 mM chloroform. Details are described under Materials and Methods. CYP, cytochrome P450.

1

1

Oxygen concentration dependence. When the dependence on pO₂ of 5 mM CHCl₃ metabolism was studied, PL-PHOS production was detected at similar high levels both in aerobic and in hypoxic conditions (1-5% pO₂), whereas it was negligible at 0% pO₂ (Fig. 3). On the other hand, PL-RAD formation was very efficiently inhibited by oxygen, attaining at 1 and 5% pO₂ only 14 and 2%, respectively, the level measured in anoxic conditions (Fig. 3).

View larger version (34K): [in this window] [in a new window]   Fig. 3.   Oxidative and reductive bioactivation of chloroform in human liver microsomes: dependence on oxygen partial pressure. Formation of PL-PHOS (dotted bars) and PL-RAD (hatched bars) was measured at any pO2 after 20 min standard incubations; they represent the mean ± S.D. of three independent determinations and are expressed as nmol [14C] · mg PL1 · nmol P4501. CYP, cytochrome P450.

Interindividual variability. A quantitative analysis of our data, obtained in a panel of 12 individual HLM showed that PL-PHOS levels varied by 8- and 58-fold at 0.1 and 5 mM CHCl₃, respectively (Table 1). Among the tested samples PL-RAD amounts, which were quantitatively higher than PL-PHOS ones, showed a level of variation by about 10-fold (Table 1). The average production of the two types of PL-adducts was very similar when measured at 5 mM chloroform in the oxygenation conditions favoring their respective formation (Table 1).

Correlation Studies. To correlate PL-PHOS and PL-RAD levels with the activity of each single isoform measured with marker substrates, CHCl₃ metabolism by microsomes from a panel of 12 phenotyped individual human livers was measured. Different chloroform concentrations were chosen to investigate the possible contribution of high affinity (0.1 mM) and low affinity component (5 mM) of the reaction. A strong positive correlation between PL-PHOS formation and Ch6OH (r = 0.837; p < 0.001) was found at 0.1 mM CHCl₃ in aerobic conditions (Table 1). A correlation with a weaker level of significance (p < 0.05) was evidenced also with CYP1A2- and 2C9-marker activities; however, this result seemed to be misleading, since the content of these two P450s paralleled that one of CYP2E1, showing a Pearson's coefficient r = 0.68 and 0.70, respectively (p < 0.01). On increasing substrate concentration up to 5 mM, PL-PHOS significantly correlated (r = 0.777; p < 0.01) with Cou7OH (Table 1). No significant correlation was observed between PL-PHOS formation and all of the other tested activities at any CHCl₃ concentration tested (Table 1). In anoxic conditions, a strong positive correlation was found between PL-RAD production at 5 mM chloroform and Ch6OH (r = 0.744; p < 0.01). Due to the involvement of CYP2E1 in PL-RAD formation, results were affected by the same misleading correlation of PL-RAD with CYP1A2 and 2C9 activities, previously evidenced in aerobic incubations at low CHCl₃ concentration. All other isoform marker activities did not significantly correlate with CHCl₃ reductive biotransformation (Table 1).

In Vitro Chloroform Metabolism by cDNA-Expressed Human P450s. Microsomes from human B-lymphoblastoid cell lines expressing single P450s were tested to characterize their abilities to metabolize [¹⁴C]CHCl₃ both oxidatively and reductively. At 5 mM CHCl₃ in aerobic conditions, CYP2A6, CYP2B6, and to a lesser extent CYP2C19 were found to catalyze CHCl₃ oxidation (Fig. 4A). At lower CHCl₃ concentrations (1 mM) only CYP2A6 showed a significant activity in the oxidative reaction, which was no more significant at 0.1 mM. The levels of PL-PHOS produced by CYP2B6 and 2C19 were negligible at 1 mM chloroform yet (Fig. 4A). CYP2E1, tested only at 15 mM CHCl₃, due to its sensitivity to solvent inhibition, resulted as a poor catalyst for PL-PHOS production in these experimental conditions (Fig. 4A). When the oxygen concentration was lowered down to 5% pO₂, the activity of CYP2A6 and 2B6 in forming PL-PHOS resulted in 82 and 52% of their production measured in aerobic conditions.

View larger version (27K): [in this window] [in a new window]   Fig. 4.   Oxidative and reductive bioactivation of chloroform by cDNA-expressed P450s. PL-PHOS (A) and PL-RAD (B) were measured, respectively, after 20-min aerobic and anoxic standard incubations, as described under Materials and Methods. Incubations contained a single heterologously expressed P450 and 0.1, 1, 5, or 15 mM chloroform, as indicated. Results are expressed as nmol [14C] · mg PL1 · nmol P4501, and represent the mean ± S.D. of three independent determinations. n.d., not detectable. CYP, cytochrome P450.

1

1

Chemical Inhibition Studies. The effect of isoform-selective chemical inhibitors on oxidative and reductive chloroform metabolism has been studied in a panel of four individual HLM, characterized by quite different single P450 content, to take into account the individual variability in the response (Table 3). The addition of 4MPYR to aerobic incubations containing 0.1 mM CHCl₃ caused the higher inhibitory effect in all the tested samples. Depending on CYP2E1 content, the level of inhibition in the presence of 4MPYR varied between 98 to 61% in HLM showing the highest and the lowest Ch6OH activity, respectively. TAO, COU, and ORP were also able to decrease to a lesser extent the production of PL-PHOS. Although the four HLM tested were characterized by very different Cou7OH, 3A4-, and 2B6-related activities, the inhibition of PL-PHOS formation due to COU, TAO, and ORP was not proportionally variable and did not parallel the related P450 content. Therefore the inhibition by TAO, COU, and ORP can be considered unspecific, very likely due to the inhibitor concentrations used. SFN addition did not affect the high affinity phase of chloroform oxidative metabolism (Table 3). At 5 mM chloroform and 20% pO₂, COU was the most efficient inhibitor: indeed, inhibition of chloroform oxidation ranged between 94 and 45%, in HLM showing the highest and the lowest CYP2A6 content, respectively. 4MPYR inhibition of PL-PHOS production attained about 30% in all the tested samples, independently on the CYP2E1 content (Table 3). The presence of ORP and TAO evidenced some inhibition on PL-PHOS production, totally unrelated to the corresponding P450 level, whereas SFN was completely ineffective (Table 3).

Immunoinhibition Studies. The effects of human-selective Abs were studied in a combinatorial immunoinhibition assay with HLM pooled from five donors. Data showed that at low CHCl₃ concentration (0.1 mM), PL-PHOS production was not affected until anti-CYP2E1 was added to the Abs mixture, when 90% inhibition was measured (Fig. 5A). Anti-CYP3A4 tested singularly did not produce any decrease in chloroform oxidation (data not shown). At 5 mM CHCl₃, the presence of anti-CYP2A6 as the first Ab in the mixture caused a strong decrease of PL-PHOS levels (84%); all further addition did not cause any change (Fig. 5B). To pick up other significant inhibition, anti-CYP2E1 and anti-CYP2B6 Ab were singularly added to the incubation mixture, producing a decrease in PL-PHOS formation as low as 20.2 ± 2.1% and 15 ± 1.6%, respectively. In anoxic conditions and 5 mM CHCl₃, PL-RAD levels decreased by 93% only in the presence of anti-CYP2E1 in the Ab mixture (Fig. 5C).

View larger version (20K): [in this window] [in a new window]   Fig. 5.   Combinatorial Abs analysis of individual P450 contribution to 0.1 mM (A) and 5 mM (B, C) chloroform bioactivation. PL-PHOS and PL-RAD were measured after 20-min aerobic (A, B) and anoxic (C) standard incubations, respectively. Incubations were carried out with HLM, pooled from five donors. Details of experimental conditions were described under Materials and Methods. The order of Abs addition is indicated under each bar, representing the mean ± S.D. of duplicate determinations. Results are expressed as nmol [14C] · mg PL1 · nmol P4501. Stars indicate the statistical significance of results (p < 0.001). CYP, cytochrome P450.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

The characterization of the P450-dependent oxidative and reductive metabolism of chloroform in HLM is reported for the first time. The study has been carried out with a multifaceted experimental strategy: the comparison and integration of the obtained results have allowed the identification of P450s involved in chloroform bioactivation in the human liver.

It has been clearly demonstrated that, depending on substrate and oxygen concentration, HLM are able to bioactivate chloroform to its oxidative and reductive metabolites, COCl₂ and dichloromethyl radical. When the oxidative pathway was studied, by using a relative broad range of chloroform concentrations (33 µM-5 mM), two phases with different affinity for the substrate were evident, suggesting the involvement of multiple P450s. The presence of two components in chloroform oxidation is not peculiar of HLM; in fact, they were previously reported in hepatic microsomes from different rodent species, showing quite similar chloroform concentration dependence (Testai et al., 1990, 1996). The differences between kinetic parameters among samples reflected the variability in the relative single P450 content.

The high affinity phase appeared to be readily saturated and catalyzed almost exclusively by CYP2E1, as suggested by the highest intrinsic clearance, expressed as V_maxapp1/K_mapp1 ratio (3.5 nmol PL-PHOS · mg PL¹ · nmol P450¹ · µM¹) obtained in HLM₁, showing the highest content of CYP2E1, associated to a low CYP2A6 level. The contribution of CYP2A6, 2B6, and 1A2 could be ruled out based on the lower intrinsic clearance values (1.22 and 0.37 nmol PL-PHOS · mg PL¹ · nmol P450¹ · µM¹) measured, respectively, in HLM₂, endowed with a high 2A6 content and in HLM₃, characterized by high levels of CYP1A2- and 2B6-related activities. In addition, at 0.1 mM chloroform PL-PHOS formation correlated with Ch6OH was almost canceled only by anti-CYP2E1 Ab and strongly inhibited by 4MPYR. The high sensitivity to organic solvents of heterologously expressed CYP2E1 did not allow to test its metabolic efficiency at low CHCl₃ concentration. The low activity shown by CYP2E1 when tested at 15 mM chloroform is compatible with our observation that the high affinity phase quickly undergoes saturation. However, no other cDNA-expressed P450 is able to catalyze the oxidative reaction at concentration <1 mM. The involvement of CYP2E1 in chloroform oxidation is in line with data on halothane-oxidative pathway (Kharasch et al., 1996).

The low affinity component of oxidative metabolism was not saturated up to 5 mM CHCl₃, in line with the high K_mapp2 values obtained, and showed high affinity for oxygen. Indeed, even in hypoxic conditions (1-5% pO₂), the oxidative metabolism seemed to be perfectly efficient; only the totally anoxic conditions did not allow, as expected, any chloroform oxidation. Coherent results from different approaches indicated CYP2A6 as the major isoform responsible for phosgene formation in this phase. Beside results from the kinetic analysis in HLM characterized by different P450 content, the positive correlation with Cou7OH and the inhibition by COU and anti-CYP2A6 Ab, cDNA-expressed CYP2A6 was one of the most active isoforms in aerobic conditions, maintaining a significant activity when oxygen concentration was decreased to 5% pO₂. The ORP inhibitory effect and the activity of heterologously expressed CYP2B6 could suggest some role for this isoform, too. However, the relevance of this isoform in chloroform metabolism could be ruled out based on the lack of correlation with the 2B6-marker activity, on the very similar V_maxapp2 and K_mapp2 obtained in HLM₂ and HLM₃, essentially differing for their content in CYP2B6, on the very low inhibition (about 15%) due to anti-CYP2B6 Ab, and on the relatively unspecific inhibition properties of ORP, acting also with other P450s including CYP2A6 (Guo et al., 1997).

The human reductive metabolism was confirmed as a low affinity process, as it has been demonstrated in rodent hepatic microsomes (Testai et al.,1990, 1996; Gemma et al., 1996). It increased linearly with substrate concentration, at least up to 5 mM chloroform and was strongly inhibited by the presence of O₂. By using cDNA-expressed isoforms, reduction appeared as a relative unspecific reaction, catalyzed by many P450s with low affinity for the haloform. However, considering results from the multifaceted approach with HLM, the reductive process appeared to be mediated by a single P450, identified as CYP2E1: Ch6OH strongly correlated with PL-RAD formation, which was decreased only by 4MPYR and anti-CYP2E1 Ab. In line with this result, CYP2E1 has been indicated as the primary isoform responsible for CCl₄ reduction in the human liver (Zangar et al., 2000). On the contrary, although the quite similar chemical structure, halothane reductive metabolism has been reported to be catalyzed by two different P450s, namely 2A6 and 3A4 (Spracklin et al., 1996).

The amount of chloroform-derived adducts produced by HLM in strictly anaerobic or in room air-equilibrated incubations were similar and even higher, when compared with those obtained in the same experimental conditions with rodent hepatic microsomes (Gemma et al., 1996). A strong similarity was found also between rodents and human hepatic P450s involved in low dose-chloroform metabolism, which is exclusively oxidative. Indeed, similar to the human data, CYP2E1 has been reported as the major catalyst of chloroform oxidation in vitro by rat liver microsomes (Testai et al., 1996) and in vivo in mice (Constan et al., 1999). Therefore, the commonly used rodent species appear to make reasonable models for studying metabolism and toxicity of low levels of CHCl₃, very likely the more appropriate to extrapolate data to the actual human exposure condition.

However, some differences exist, when high CHCl₃ concentrations have to be considered, starting with the pO₂ dependence of chloroform metabolism. The rate of both chloroform oxidation and reduction in the two limit oxygenation conditions (anoxic and aerobic) are comparable both in human and in rodent liver microsomes (Gemma et al., 1996). The differences arise when comparing results obtained in hypoxic conditions (5 and 1% pO₂) (Gemma et al., 1996), better representing the physiological situation of hepatic centrilobular region, where most of drug metabolism takes place. In HLM the oxidative pathway strongly prevailed on radical formation; the oxidation/reduction (ox/red) ratio was 34 and 5 at 5 and 1% pO₂, respectively. In the same conditions, rat and mouse liver microsomes were characterized by much lower values of the ox/red ratio: 1.5 to 6 at 5% pO₂ and 0.2 to 0.5 at 1% pO₂ (Gemma et al., 1996), clearly suggesting a more efficient hepatic reductive metabolism. Therefore, in the absence of investigations able to quantitate the actual proportion of each pathway in vivo, these data suggest that, within a broad range of chloroform concentrations, the major toxicologically relevant metabolite produced by the human liver is phosgene.

Secondly, at variance with the human CYP2A6 involvement, the low affinity phase of chloroform oxidation in rat liver microsomes is mediated mainly by CYP2B1/2 (Testai et al., 1996), as supported also by the potentiation of chloroform toxicity due to animal pretreatment with CYP2B inducers (Lavigne and Marchand, 1974; Stevens and Anders, 1981). Regarding reductive metabolism, in the rat constitutive isoforms and CYP2B1/2 at high CHCl₃ concentrations were identified as active, and no contribution has been attributed to CYP2E1 (Testai et al., 1996). A similar situation was reported for CCl₄: CYP2B has been implicated in its reductive metabolism in rodent (Frank et al., 1982), whereas in HLM the only isoform active in the reaction was CYP2E1 (Zangar et al., 2000).

Concluding, chloroform can undergo both oxidative and reductive metabolism in the human liver, depending on oxygen and substrate concentration. At the low levels, typical of actual human exposure through the use of chlorinated waters, CHCl₃ is metabolized primarily to phosgene by CYP2E1, a typical high affinity-low capacity enzyme. When the CYP2E1-mediated reaction is saturated, the predominant role in phosgene production is for CYP2A6, efficient even in highly hypoxic conditions (1% pO₂). CYP2E1 is a unique isoform, in that beside its role in chloroform oxidation, it is able to catalyze also dichloromethyl radical formation. This latter pathway seems to be scarcely relevant in the human liver, since it is active only at high substrate concentrations and in strictly anaerobic conditions.

The relevance of CYP2E1 in chloroform metabolism in humans provides useful information to identify eventual differences in susceptibility to chloroform-induced adverse effects. Variations in the level of expression of CYP2E1 is not uncommon, being determined either by genetic features, by pathophysiological conditions such as diabetes, or by environmental factors, such as alcoholic beverages consumption or exposure to other xenobiotics known to affect CYP2E1 expression. A peculiar situation may occur in the human fetus, between gestational days 50 and 113, when the fetal brain CYP2E1 mRNA is expressed at relatively high amounts corresponding to a fairly constant level of enzymatic activity of ethanol metabolism (Brzezinski et al., 1999). The eventual effects related to chloroform exposure during this stage of development might be of interest, providing some biological basis for reproductive outcomes in humans exposed to chloroform via drinking water.