Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Felbamate (1, Fig. 1, 2-phenyl-1,3-propanediol dicarbamate) is an antiepileptic drug approved in 1993 for the treatment of several forms of epilepsy (Leppik, 1996). Previously, we have proposed and provided evidence for a metabolic scheme leading to a reactive metabolite from felbamate (Thompson et al., 1996). 3-Carbamoyl-2-phenylpropionaldehyde 3 (Fig. 1) was proposed as a reactive intermediate in the oxidation of 2-phenyl-1,3-propanediol monocarbamate 2 to the major human metabolite 3-carbamoyl-2-phenylpropionic acid 4. The aldehyde carbamate 3 was found to undergo spontaneous elimination in vitro to the cytotoxic ,-unsaturated aldehyde 2-phenylpropenal 5, as well as reversible cyclization to the more stable cyclic carbamate 4-hydroxy-5-phenyl-tetrahydro-1,3-oxazin-2-one 6. In vitro studies using human liver tissue have provided further support for this metabolic pathway (Kapetanovic et al., 1998). 2-Phenylpropenal 5 may play a role in the development of the idiosyncratic toxicities observed during felbamate therapy (Pennell et al., 1995). Evidence for the formation of 5 in vivo has been reported with the identification of the modified N-acetylcysteine conjugates 7 and 8 of this ,-unsaturated aldehyde in both human and rat urine after felbamate administration (Thompson et al., 1997).

View larger version (19K): [in this window] [in a new window]   Fig. 1.   The metabolism of felbamate.

There is an intriguing parallelism between the transformation of 6 to 5 and the known pathway from 4-hydroxycyclophosphamide 12 to acrolein 14 (Fig. 2) (Takamizawa et al., 1972). The similarity of these transformations led us to consider 5-phenyl-1,3-oxazinane-2,4-dione 9 as a potential metabolite of felbamate. The corresponding species, 4-ketocyclophosphamide 15, has been reported to form and is excreted in urine after cyclophosphamide treatment (Takamizawa et al., 1972; Torkelson et al., 1974). These two "keto"-metabolites appear to offer an alternative mechanism involving selective hydrolysis for the formation of the corresponding carboxylic acid metabolites [i.e., 3-carbamoyl-2-phenylpropionic acid 4 and 2-carboxyethyl-N,N-bis-(2-chloroethyl)phosphorodiamidate 16]. However, Takamizawa et al. (1972) have reported that, after the administration of 4-ketocyclophosphamide to rabbits, 16 was not observed in urine, and 15 was recovered unchanged. Whether this would also be true for the felbamate-derived metabolites formed in humans remains to be determined.

View larger version (14K): [in this window] [in a new window]   Fig. 2.   Metabolic transformations of 4-hydroxycyclophosphamide 12 and the cyclic carbamate 6 derived from felbamate.

As the formation of 9 may have both potential pharmacologic and toxicologic implications for felbamate therapy, we wished to study its reactivity. To this end, we have developed a synthesis of 5-phenyl-1,3-oxazinane-2,4-dione 9 and evaluated its reactivity in vitro. We report that this species undergoes a base-catalyzed decomposition to three products, one of which is the major human metabolite 3-carbamoyl-2-phenylpropionic acid 4. Additionally, we have found evidence of the presence of oxazinane-dione in human urine after felbamate treatment through the identification of its major in vitro decomposition product, 2-phenylacrylamide 11.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Chemicals and Instruments. All reagents were purchased from either Aldrich Chemical Co. or Sigma Chemical Co. and were of the highest quality available. HPLC was performed on a Waters 2690 separations module with a Waters 484 tunable absorbance detector. Mass spectra were obtained by coupling this liquid chromatography (LC)¹ system to a Finnigan MAT LCQ ion trap mass spectrometer equipped with an electrospray ionization source. NMR spectra were recorded on a General Electric QE300 spectrometer at 300 MHz, and chemical shifts are reported in ppm. Melting points were determined on a Thomas-Hoover UNI-MELT apparatus and are uncorrected.

Synthesis. 3-Carbamoyl-2-phenylpropionic acid (4).

5-Phenyl-1,3-oxazinane-2,4-dione (9). 3-Carbamoyl-2-phenyl propionic acid (4, 0.1 mmol) and 1,1'-carbonyldiimidizole (0.25 mmol) were dissolved in dichloromethane (5 ml), and the resultant solution was magnetically stirred for 12 h in a sealed vessel (9). The mixture was purified by passage of the crude reaction mixture through a pad of silica gel (10 g) in a fritted funnel, eluting with ethyl ether, affording 9 as a white powder in 53% yield (R_{f (ethyl ether)} = 0.65, mp = 91-93°C). ¹H NMR (CDCl₃):  7.68 (bs, 1H), 7.44-7.38 (m, 3H), 7.29-7.22 (m, 2H), 4.56 (dd, 1H, J = 11.55, 5.77 Hz), 4.64 (dd, 2H, J = 11.55, 8.48 Hz), 4.0 (dd, 1H, J = 8.47, 5.77) LC/ESI-MS: MH⁺ = 192.

2-Phenylacrylamide (11). 2-Phenylacrylic acid (Zadrozna et al., 1993) (0.5 mmol) was dissolved in dichloromethane (5 ml). N-Methylmorphiline (1.1 mmol) and 1,3-dicyclohexylcarbodiimide (0.65 mmol) were added at 0°C with magnetic stirring under an argon atmosphere. After 30 min, anhydrous ammonia was bubbled through the solution for 10 min at 0°C. Concentration of the reaction mixture and flash chromatography of the resultant powder, eluting with ethyl ether, afforded crystalline 11 as white flakes in 82% yield (R_{f (ethyl ether)} = 0.2, mp = 121-122°C). ¹H NMR (CDCl₃):  7.40-7.34 (m, 5H), 6.37 (bs, 2H), 6.17 (s, 1H), 5.7 (s, 1H). ¹³C NMR (CDCl₃):  169.9, 144.7, 137.6, 129.2, 129.2, 129.1, 128.6, 128.6, 123.6. (C₉H₉NO  FW = 147) LC/ESI-MS: MH⁺ = 148.1.

Kinetic and Product Studies of the Decomposition of Tetrahydro-1,3-Oxazinane-2,4-Dione. Tetrahydro-1,3-oxazinane-2,4-dione 9 (200 µM) was added to 20 mM KPO₄ buffer controlled at various pH values. The alcohol carbamate 2 (200 µM) was also added for use as an internal standard. Each solution was incubated at 37°C, and aliquots were removed at appropriate time points. The remaining concentration of the oxazinane-dione was determined by HPLC [C₁₈, 2.1 × 150 mm, 200 µl/min isocratic 20% CH₃CN:80% (0.1%) HOAc]. The half-life (t_1/2) was determined using a first-order approximation by plotting the natural logarithm of the remaining concentration of oxazinane-dione versus time to reveal a slope equal to k. The t_1/2 is equal to 0.693/k. The identity of the three decomposition products was determined using LC/MS. The retention time and m/z of each decomposition product was compared with that of its corresponding synthetic standard except in the case of 1-hydroxy-2-phenylpropamide for which a synthetic standard was not available.

Identification of 2-Phenylacrylamide from the Decomposition of Oxazinane-Dione Excreted in Urine from Patients Undergoing Felbamate Therapy. Urine samples were obtained from patient volunteers undergoing felbamate therapy for control of epileptic seizures as part of an ongoing urinary metabolite study (Thompson et al., 1999). Two 20-ml portions of each urine sample were placed in 50-ml conical tubes. The pH value of one of the samples was adjusted to 9 by the addition of 1 N NaOH. The pH value of the other sample was not adjusted and remained ~5.5. Both tubes were incubated for 3 h at 37°C in an orbital shaking water bath.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

The stability of 5-phenyl-1,3-oxazinane-2,4-dione 9 was assessed at various pH values in phosphate buffer, and the results show that 9 readily decomposes in a pH dependent fashion to three products. Figure 3 shows a graph of the decomposition of the oxazinane-dione at each pH evaluated. The half-lives for this decomposition were very long at low to neutral pH (t_1/2 = 3.5 h at pH 7.4) but decreased at higher pH (t_1/2 = 0.8 h at pH 9).

View larger version (25K): [in this window] [in a new window]   Fig. 3.   pH dependent decomposition of the oxazinane-dione.

Figure 4 shows three chromatograms (UV,  = 214 nm) obtained during the decomposition of the oxazinane-dione at pH 9. The chromatogram obtained at time = 0 min is shown in Fig. 4A. The oxazinane-dione eluted at 10.5 min and the monocarbamate alcohol, which was used as an internal standard, eluted at 5.9 min. Figure 4B shows the chromatogram after 30 min of incubation at 37°C. The peak corresponding to 9 decreased relative to the internal standard peak, and three new peaks (retention times: 3.5, 7.4, 8.8) appeared. The results after 60 min of incubation are shown in Fig. 4C. The oxazinane-dione peak continued to decrease, although the three new peaks became larger. These same three decomposition products were observed at lower pH values. Using LC/MS, parent molecular ion data for the three decomposition products were obtained (retention times, 3.5, 7.4, and 8.8 min; MH⁺, 166, 148, and 210, respectively).

View larger version (31K): [in this window] [in a new window]   Fig. 4.   UV chromatograms from HPLC separation of the decomposition of the oxazinane-dione at pH 9.

The m/z and retention time of the peak observed at 8.8 min matches that of the acid carbamate metabolite of felbamate, and this analyte coeluted with a synthetic standard of 4 (not shown). The identity of the remaining two decomposition products were predicted as shown in Fig. 5. Similar to 4, 1-hydroxy-2-phenylpropamide 10 could arise from attack of water at one of the carbonyl carbons. Such a nucleophilic addition by water could give rise to an intermediate that would spontaneously give off CO₂ to result in 1-hydroxy-2-phenylpropamide 10 [MH⁺ = 166, retention time = 3.5 min (Fig. 4)].

View larger version (15K): [in this window] [in a new window]   Fig. 5.   Decomposition products of the oxazinane-dione.

The formation of 2-phenylacrylamide 11 would presumably proceed via the enolate. This intermediate would also extrude CO₂ to give rise to the observed product (perhaps via a retro-Diels-Alder mechanism). The identity of the decomposition product with a retention time of 7.4 min (Fig. 4) was confirmed as 2-phenylacrylamide by coelution with a synthetic standard (not shown). Based on comparison of the data from these decomposition studies and data obtained from standard curves generated with synthetic standards, it was determined that ~4-fold more 2-phenylacrylamide is formed than acid carbamate under the in vitro decomposition conditions. Unfortunately, a synthetic standard of 1-hydroxy-2-phenylpropamide was not available so exact values for the decomposition to all three products cannot be determined. However, based on the relative area of the peaks corresponding to 10, the acid carbamate in Fig. 4C, and the similarity of the chromophores for these two molecules, it can be conservatively estimated that 1-hydroxy-2-phenylpropamide represents 5% of the in vitro decomposition products. Making this assumption, we can approximate that 11 accounts for ~75% of the in vitro decomposition products with 4 constituting the remaining ~20%.

Despite extensive efforts at the direct identification of the oxazinane-dione from postdose patient urine samples, we have been unable to observe this species directly. Our efforts have been confounded by the lack of a strong chromophore for this molecule, its poor ionization by the electrospray process, its apparent low abundance in the urine samples, and the complexity of the biological matrix. However, we were able to indirectly demonstrate the presence of 9 in urine samples from patients undergoing felbamate therapy. Based on the in vitro decomposition studies, we anticipated that, if present in a urine sample, 9 would decompose to 2-phenylacrylamide 11 with time. Additionally, this decomposition could be accelerated if the pH value of the urine were made more basic. The advantage to this approach is that 11 possesses a much stronger chromophore than 9 and demonstrates excellent ionization using electrospray.

A postdose urine sample from a patient undergoing felbamate therapy was obtained, made basic (pH 9) by addition of 1 N NaOH, and incubated at 37°C for 3 h. This sample was then repeatedly extracted with chloroform. The organics were combined, evaporated, and the resulting residue was redissolved in 20% CH₃CN, 80% (0.1%) acetic acid. This reconstituted material was then analyzed for the presence of 2-phenylacrylamide by LC/MS. Figure 6 shows the results obtained from this experiment.

View larger version (34K): [in this window] [in a new window]   Fig. 6.   2-Phenylacrylamide formation from oxazinane-dione present in postdose urine. A, chloroform-extracted urine sample incubated at pH ~ 5.5. B, chloroform extracted urine sample incubated at pH 9. C, coelution of synthetic 2-phenylacrylamide with analyte from B.

Figure 6A shows the data obtained from a sample that underwent the above procedure except the pH was not adjusted and remained at ~5.5 during the incubation. The large peak present in panels 1, 2, and 3 is from felbamate present in the postdose urine sample. Figure 6B shows the data obtained for the same urine sample after adjusting the pH to 9 and incubating at 37°C. A new peak is observed (retention time = 7.4 min) in both the UV chromatogram (Fig. 6B, panel 2) and the m/z = 148 chromatogram (Fig. 6B, panel 4) for the pH adjusted sample. No new signal is observed from a control urine sample (i.e., no felbamate therapy) treated to the same protocol (not shown). The m/z of this new peak corresponds to the m/z of 2-phenylacrylamide 11. Figure 6C shows that the addition of a synthetic standard of 2-phenylacrylamide to this pH adjusted sample results in an increase in the peak at 7.4 min. This increase is evident both for the UV and m/z = 148 chromatograms. Thus, this new peak, which appears after incubation under basic conditions, has the same m/z as and coelutes with 2-phenylacrylamide 11.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

We expected that, if formed in vivo, felbamate-derived tetrahydro-1,3-oxazinane-2,4-dione could be excreted in urine. Although we demonstrated that this compound undergoes several decomposition pathways, the decomposition occurs slowly (t_1/2 = 15 h, pH 6.5) at neutral to acidic pH. The pH of human urine being slightly acidic, we anticipated that 9 might survive the process of excretion. Although we were unable to detect the oxazinane-dione directly, we successfully detected its major in vitro decomposition product in all felbamate patient urine samples examined (n = 7) after subjecting the urine samples to mild base.

In addition to establishing the presence of the oxazinane-dione in patient urine after felbamate treatment, these results indicate that 2-phenylacrylamide is not directly excreted as evidenced by the absence of peaks corresponding to 11 in Fig. 6A. If generated in vivo, 2-phenylacrylamide could have toxic potential. The toxicity of acrylamide has been well documented and this 2-phenyl analog could be expected to demonstrate comparable effects in vivo. Additionally, like acrylamide, it may be bioactivated in vivo to the epoxide. This species could demonstrate greater reactivity and consequent toxic potential. The epoxide of acrylamide is thought to play a prominent role in the toxicity of this molecule (Tilson, 1981). The absence of 2-phenylacrylamide in urine samples before pH adjustment and incubation suggests that either any 2-phenylacrylamide formed reacts before excretion or the uncatalyzed decomposition pathways that we elucidated in vitro are not occurring in vivo to a significant extent. This does not rule out the possibility that the oxazinane-dione 9 is a/the metabolic precursor to the acid carbamate 4 in vivo as this specific conversion may be enzymatically catalyzed.

In addition to the implications for the metabolic pathway leading to the acid carbamate 4, the formation of the oxazinane-dione 9 in vivo may have pharmacological consequences with respect to the antiepileptic activity of felbamate. The structure of the oxazinane-dione bears intriguing similarity to several established antiepileptic drugs as illustrated in Fig. 7. It is possible that the oxazinane-dione is responsible for some aspects of the efficacy of felbamate in vivo. As patients undergoing felbamate therapy for seizure control ingest large quantities of felbamate (grams per day), even a 1 to 2% conversion to a pharmacologically active metabolite could have significant effects. This could play an important role in the seizure control observed with felbamate, particularly if the metabolite (i.e., the oxazinane-dione) was a more potent compound. In light of the possibility that the oxazinane-dione could be a/the metabolic precursor to the major human metabolite, it may be formed in significant quantities [the acid carbamate was reported to represent ~12% of a dose (Adusumalli et al., 1993)]. We are in the process of evaluating isosteric analogs of the oxazinane-dione for their pharmacological effects.

View larger version (15K): [in this window] [in a new window]   Fig. 7.   Structural similarity of the oxazinane-dione and some established antiepileptic drugs.