Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

FBM¹ (Felbatol; 2-phenyl-1,3-propanediol dicarbamate) is a broad-spectrum antiepileptic drug. It was approved in 1993 by the United States Food and Drug Administration for the treatment of several forms of epilepsy, including Lennox-Gastaut syndrome and related disorders, for which there have been very few therapeutic options. After its release, several cases of severe idiosyncratic adverse reactions (aplastic anemia and hepatotoxicity) with the use of FBM were reported. This prompted a recommendation by the United States Food and Drug Administration that FBM therapy be discontinued unless the benefit of seizure control outweighs the risk of toxicity. It is estimated that approximately 15,000 patients worldwide are still receiving FBM therapy. In the neurological community, FBM is thought to be too important to be discarded as a therapeutic agent, but it should primarily be used for patients with severe epilepsy that is refractory to other therapy (Pellock and Brodie, 1997).

The reasons for these idiosyncratic untoward reactions remain unclear. The possibility of metabolic activation to toxic products has been proposed (Thompson et al., 1996). Initially, four metabolites of FBM, i.e. pOH-FBM, 2OH-FBM, MCF, and CPPA, were identified in humans (Adusumalli et al., 1993) and experimental animals (Yang et al., 1992; Romanyshyn et al., 1993). None of these compounds is thought to have significant toxicological potential. More recently, mercapturic acid metabolites of FBM were identified in human and rat urine (Thompson et al., 1997). This was suggestive of reactive metabolite formation, and a metabolic scheme for the disposition of FBM (fig. 1) was proposed (Thompson et al., 1997).

View larger version (46K): [in this window] [in a new window]   Fig. 1.   Proposed metabolic scheme for FBM. Chemical structures in boxes, known in vivo metabolites of FBM in humans. NAc, N-acetyl.

A reactive aldehyde was proposed as an intermediate to the mercapturates. MCF could be oxidized to CBMA, which could undergo reversible cyclization to CCMF, oxidation to CPPA (the major metabolite in humans in vivo), or -elimination (spontaneous or catalyzed) to ATPAL. ATPAL, an ,-unsaturated aldehyde, is a potent electrophile that undergoes rapid conjugation with glutathione and could represent the toxic reactive metabolite of FBM (Thompson et al., 1996). CCMF, on the other hand, is relatively stable and could represent a metabolic depot and part of a mechanism of transport from the liver to distal sites for the putative ultimate reactive metabolite, ATPAL. The present study was designed to investigate the validity of the proposed pathways leading to reactive metabolites in humans, starting with human liver tissue in vitro. These studies were performed primarily with the intermediate metabolite MCF, because neither the identity nor the tissue and subcellular distribution of the enzymes responsible for its formation from FBM are known.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Chemicals. FBM, MCF, 2OH-FBM, pOH-FBM, and CPPA were provided by Wallace Laboratories (Cranbury, NJ), and CCMF and ATPAL were synthesized as described previously (Thompson et al., 1996). Methanol and acetonitrile were obtained from Burdick & Jackson (Muskegon, MI). EDTA (disodium salt), D-glucose-6-phosphate (monosodium salt), glucose-6-phosphate dehydrogenase (from Torula yeast), NADP⁺ (sodium salt), 4MP, and DDC were purchased from Sigma Chemical Co. (St. Louis, MO), PFBH from Aldrich Chemical Co. (Milwaukee, WI), and NAD⁺ from Fluka Chemie AG (Buchs, Switzerland).

Liver Tissue. Human liver specimens (medically unsuitable for transplantation) were acquired under the auspices of the Washington Regional Transplant Consortium (Washington, DC) and were stored at 80°C until use. S9 fractions and microsomes were prepared by tissue homogenization and differential centrifugation as described previously (Jamis-Dow et al., 1995), divided into aliquots, and stored at 80°C.

Metabolic Incubations. Incubations were carried out in 1 ml of phosphate buffer (31 mM monobasic potassium phosphate, 69 mM dibasic sodium phosphate, 5 mM magnesium chloride, 1 mM EDTA, pH 7.4) at 37°C for 60 min, in the presence of human liver S9 fractions or microsomes (1 mg of protein). FBM (250 µM) or MCF (250 µM) was added to the medium. For the inhibition experiments, MCF (250 µM) in the presence of either DDC (100 µM) or 4MP (100 µM) was used for the incubations. After a 3-min pre-equilibration at 37°C, the reactions were started by the addition of an NADPH-generating system (100 µM D-glucose-6-phosphate, 1 Sigma unit of glucose-6-phosphate dehydrogenase, and 10 µM NADP⁺). In some cases (as indicated), 10 µM NAD⁺ was also added because it is the commonly preferred cofactor for alcohol and aldehyde dehydrogenases (Sladek et al., 1989), which are expected to be involved in the metabolism of FBM. Control blank samples were incubated in the absence of one of the following: microsomes or S9 fraction, substrate, or NADP⁺. The omitted component was added at the end of the reaction, before the extraction. With the exception of trapping experiments (see below), reactions were terminated by the placement of tubes on ice and the addition of 2 ml of ethyl acetate (with the addition of 100 µl of 1 N hydrochloric acid necessary for the extraction of CPPA). For analysis, the ethyl acetate extract was evaporated to dryness under nitrogen at room temperature and reconstituted in 50-100 µl of the mobile phase for HPLC or 50-100 µl of acetonitrile for GC/MS analysis.

Analysis. HPLC (Hewlett Packard Series 1100 system) separation was achieved using a Waters NovaPak C₁₈ column (3.9 × 150 mm, 4 µm), at 40°C. The mobile phase was composed of 30 mM monobasic potassium phosphate/acetonitrile/methanol (82:12:6), at a flow rate of 0.5 ml/min. Monitoring was performed using a diode-array detector, at 214 and 220 nm.

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

Metabolism of MCF. The formation of both CCMF and CPPA from MCF by human liver S9 fractions and microsomes was demonstrated after incubation of MCF in the presence of both NADP⁺ and NAD⁺ (fig. 2). HPLC chromatograms for a S9 incubation medium extract (fig. 2, top) revealed two metabolic products, with retention times of 7.061 and 9.398 min, that coeluted with synthetic reference CCMF and CPPA, respectively, and were not present in the postreaction control medium (fig. 2, bottom). Similarly, HPLC analysis of medium extracts from CCMF S9 incubations confirmed the production of CPPA from CCMF (fig. 3).

View larger version (21K): [in this window] [in a new window]   Fig. 2.   Representative HPLC tracings for MCF metabolism by human liver S9 fractions. Top, extract of an active reaction sample; bottom, corresponding control (S9 fraction was added at the end of the reaction). Incubations were performed in the presence of an NADPH-generating system, NADP+, and NAD+.

View larger version (20K): [in this window] [in a new window]   Fig. 3.   Representative HPLC tracings for CCMF metabolism by human liver S9 fractions. Top, extract of an active reaction sample; bottom, corresponding control (S9 fraction was added at the end of the reaction). Incubations were performed in the presence of an NADPH-generating system, NADP+, and NAD+.

View larger version (12K): [in this window] [in a new window]   Fig. 4.   Representative SIM tracings for CCMF formation from MCF by human liver microsomes. Top, extract of an active reaction sample; middle and bottom, corresponding controls, with microsomes added at the end of the reaction (middle) and with the absence of cofactors (bottom). Incubations were performed in the presence of an NADPH-generating system. Samples were analyzed using GC/MS in the chemical ionization mode, with methane as the reagent gas.

Cofactor Dependence. MCF was incubated with human liver S9 fractions in the absence or presence of NADP⁺ or NAD⁺. GC/MS SIM or HPLC analyses of the medium extracts were used to monitor the formation of CCMF from MCF in the presence of NADP⁺ and/or NAD⁺. The amount of CCMF in the S9 incubation extracts, as analyzed by HPLC or GC/MS, was approximately 2-fold greater in the presence of NADP⁺ than in the presence of NAD⁺ and was intermediate between these values when both cofactors were added (table 1). The difference between NADP⁺ and NAD⁺ was even more pronounced (approximately 10-fold) with microsomes instead of S9 fractions. This finding was independently confirmed with trapping experiments (see below). On the other hand, the opposite observation was made for CPPA. The amount of CPPA detected was approximately 20-fold greater after incubation in the presence of NAD⁺ than in the presence of NADP⁺ and was intermediate between these values when both cofactors were present (table 1).

Adduct Formation. The proposed aldehyde metabolites (CBMA and ATPAL) may be reactive intermediates requiring a trapping procedure for identification. Therefore, we used a method analogous to that described for 4-hydroxycyclophosphamide/aldophosphamide (Anderson et al., 1995). The proposed aldehyde intermediate metabolites were reacted with PFBH to yield relatively stable Schiff base adducts (oxime derivatives), according to the scheme shown in fig. 5.

View larger version (14K): [in this window] [in a new window]   Fig. 5.   Proposed scheme for trapping the intermediate aldehyde metabolites. Adduct formation using PFBH was used to generate stable oxime derivatives.

View larger version (18K): [in this window] [in a new window]   Fig. 6.   Total-ion chromatogram and mass spectra for the ATPAL adduct. Chemical ionization with methane reagent gas was used for analysis. Top, reconstructed total-ion current; middle and bottom, mass spectra for the E- and Z-isomeric oxime derivatives.

View larger version (18K): [in this window] [in a new window]   Fig. 7.   Total-ion chromatogram and mass spectra for the CBMA adduct. Chemical ionization with methane reagent gas was used for analysis. Top, reconstructed total-ion current; middle and bottom, mass spectra for the E- and Z-isomeric oxime derivatives.

View larger version (20K): [in this window] [in a new window]   Fig. 8.   SIM tracings for the ATPAL and CBMA adducts in microsomal reactions. Microsomal reactions were performed using MCF (top) or FBM (bottom) as the substrate, in the presence of an NADPH-generating system, NADP+, and NAD+. Samples were analyzed using GC/MS in the chemical ionization mode, with methane as the reagent gas.

Metabolism of FBM. The formation of adducts from FBM itself could also be shown, although to a much lesser extent than from MCF (fig. 8). Both ATPAL and CBMA adducts were seen, with the ATPAL adduct peaks being larger than the CBMA adduct peaks. This adduct peak ratio is opposite that seen with MCF as the substrate. However, as discussed above, the relative amount of the ATPAL adduct resulting from an enzymatic reaction is unclear.

Inhibition. The effects of several P450 inhibitors on adduct formation from MCF were assessed. The most pronounced effect was seen in the presence of the P4502E1-selective inhibitors DDC and 4MP. An approximately 90% decrease in the CBMA adduct level was observed in the presence of 100 µM DDC or 4MP (fig. 9). Ketoconazole (3 µM) (P4503A4) or sulfaphenazole (100 µM) (P4502C9) did not show significant effects, whereas coumarin (100 µM) (P4502A6) caused an approximately 40% decrease in the CBMA adduct level.

View larger version (15K): [in this window] [in a new window]   Fig. 9.   SIM tracings for the ATPAL and CBMA adducts in microsomal reactions in the presence of inhibitors, DDC, and 4MP. Microsomal reactions were conducted using MCF as the substrate, in the absence of inhibitors (top) and in the presence of DDC (middle) or 4MP (bottom). Incubations were carried out in the presence of an NADPH-generating system, NADP+, and NAD+. Samples were analyzed using GC/MS in the chemical ionization mode, with methane as the reagent gas.

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

Although FBM was marketed in 1993, its metabolic profile and the enzymes responsible for its metabolism have not been fully elucidated. The only enzymes clearly implicated in FBM metabolism are P4503A4 and P4502E1, which are involved in formation of the hydroxy metabolites pOH-FBM and 2OH-FBM (Racha et al., 1996; Glue et al., 1997). Other enzymes possibly involved in FBM metabolism, especially via the MCF route, are esterases, amidases, hydrolases, and alcohol and aldehyde dehydrogenases. However, their identities and tissue and subcellular localization remain to be determined. Recently, microsomal carboxylesterase (hydrolase A) was implicated in the metabolism of an another carbamate, ethyl carbamate (Lee et al., 1998). In general, P450 enzymes have been the focus of attention, and the roles of other enzymes in drug metabolism have not been investigated as extensively. Recently, mercapturic acid metabolites of FBM were demonstrated in humans in vivo (Thompson et al., 1997). Their presence is suggestive of reactive metabolite formation and has led to the postulated metabolic scheme (fig. 1). This study concentrated on further elucidating the metabolic pathways involving the proposed reactive metabolites (i.e. CBMA and ATPAL) in humans in vitro.

Formation of the proposed cyclic carbamate intermediate metabolite CCMF from MCF and its metabolism to CPPA were demonstrated using human liver S9 fractions and microsomes. This provides the first direct evidence for the proposed metabolic scheme. This cyclic intermediate may serve as a depot and part of a transport mechanism for the reactive and potentially toxic reactive aldehydes, especially the highly electrophilic and cytotoxic ATPAL. CCMF could be distributed within the body and revert to the reactive species at distal sites. It would be worthwhile to examine whether the relatively stable CCMF can also be detected in vivo. Corresponding to the metabolism of CCMF to ATPAL is the biotransformation of 4-hydroxycyclophosphamide, the activated form of cyclophosphamide, to acrolein (Anderson et al., 1995).

Also in analogy to cyclophosphamide biotransformation, relatively unstable aldehyde intermediates were trapped by formation of stable oxime derivatives. These Schiff base adducts of CBMA (the open form of CCMF) and ATPAL were demonstrated in microsomal reactions with MCF as a substrate. Adduct formation was also seen in preliminary human liver slice experiments (data not shown). Subsequent experiments also showed adduct formation in the microsomal reactions with FBM as the substrate. However, the adduct formation from FBM was much less than that from MCF. This could be because MCF formation might be a rate-limiting step for the subsequent metabolism and/or extramicrosomal enzymes might be involved. It should be pointed out that there is some decomposition of the CBMA adduct to the ATPAL adduct, via a typical loss of the relatively labile carbamoyl moiety, under the GC/MS conditions. This currently makes it difficult to distinguish between metabolic and degradative contributions to ATPAL adduct levels. However, because nonenzymatic -elimination can readily take place, the pathway for ATPAL production is available and would be expected to play a role, as was the case for cyclophosphamide; this is supported by the identification of mercapturic acids derived from ATPAL in vivo.

It interesting to note that P4502E1-selective inhibitors were very effective in inhibiting adduct formation. This P450 isozyme is known to oxidize ethanol (Itoga et al., 1998; Ingelman-Sundberg et al., 1994) and has also been implicated in the metabolism of other smaller carbamates, i.e. ethyl carbamate and vinyl carbamate (Guengerich and Kim, 1991). Our results shown in fig. 9 are consistent with the involvement of P4502E1 in the conversion of MCF to CBMA. However, although these two inhibitors are selective for P4502E1, they have other biological activities. DDC inhibits P4502E1 and is an aldehyde dehydrogenase inhibitor, and 4MP is an alcohol dehydrogenase inhibitor; both of these enzymes could be involved in the metabolism of MCF.

Another interesting observation concerns the selectivity of metabolic pathways for the cofactors NAD⁺ and NADP⁺ (table 1). Because the conversion of MCF to CBMA and that of CBMA to CPPA preferentially utilize the cofactors NADP⁺ and NAD⁺, respectively, the NADH⁺/NAD⁺ ratio would be expected to affect the relative amounts of ATPAL and CPPA formed. This selective cofactor dependence may be an important determinant for the occurrence of toxicity, because CCMF can serve as a precursor to the putative reactive species CBMA, which is spontaneously converted to ATPAL. Factors within the body that alter the balance between these two pathways could affect the expression of toxicity. Numerous examples that demonstrate that drugs or xenobiotics can affect the NADH⁺/NAD⁺ energy state are available in the literature. For example, Ramasamy et al. (1997) reported that zopolrestat maintains a lower cytosolic NADH⁺/NAD⁺ ratio in diabetic animal models and markedly protects both nondiabetic and diabetic hearts during ischemia and reperfusion. The elevated NADH⁺/NAD⁺ ratio resulting from ethanol consumption has been implicated as contributing to hepatic toxicity (Seitz and Csomos, 1992).

Further understanding of putative toxic metabolic pathways may provide means for minimizing their contributions or may suggest structural drug modifications to preclude them. In addition to investigations of possible toxic pathways of FBM metabolism, these approaches may have more general applicability. The use of adduct formation to trap reactive metabolic aldehydes may be practical for other drugs, as was shown for cyclophosphamide and FBM. The use of data from human in vitro systems to predict the in vivo generation of reactive metabolites in humans also warrants further exploration.