Species-differences in the in vitro biotransformation of trifluoroethene (HFO-1123)

Abstract

1,1,2-Trifluoroethene (HFO-1123) is anticipated for use as a refrigerant with low global warming potential. Inhalation studies on HFO-1123 in rats indicated a low potential for toxicity (NOAELs ≥ 20,000 ppm). In contrast, single inhalation exposure of Goettingen^® minipigs (≥ 500 ppm) and New Zealand white rabbits (≥ 1250 ppm) resulted in severe toxicity. It has been suggested that these pronounced species-differences in toxicity may be attributable to species-differences in biotransformation of HFO-1123 via the mercapturic acid pathway. Therefore, the overall objective of this study was to evaluate species-differences in glutathione (GSH) dependent in vitro metabolism of HFO-1123 in susceptible versus less susceptible species and humans as a basis for human risk assessment. Biotransformation of HFO-1123 to S-(1,1,2-trifluoroethyl)-L-glutathione (1123-GSH) and subsequent cysteine S-conjugate β-lyase-mediated cleavage of the corresponding cysteine conjugate (1123-CYS) was monitored in hepatic and renal subcellular fractions of mice, rats, minipigs, rabbits, and humans. While 1123-GSH formation occurred at higher rates in rat and rabbit liver S9 compared to minipig and human S9, increased β-lyase cleavage of 1123-CYS was observed in minipig kidney cytosol as compared to cytosolic fractions of other species. Increased β-lyase activity in minipig cytosol was accompanied by time-dependent formation of monofluoroacetic acid (MFA), a highly toxic compound that interferes with cellular energy production via inhibition of aconitase. Consistent with the significantly lower β-lyase activity in human cytosols, the intensity of the MFA signal in human cytosols was only a fraction of the signal obtained in minipig subcellular fractions. Even though the inconsistencies between GSH and β-lyase-dependent metabolism do not allow to draw a firm conclusion on the overall contribution of the mercapturic acid pathway to HFO-1123 biotransformation and toxicity in vivo, the β-lyase data suggest that humans may be less susceptible to HFO-1123 toxicity compared to minipigs.

Keywords: 19F-NMR; HFO-1123; In vitro biotransformation; LC–MS/MS; Mercapturic acid pathway; Trifluoroethene; β-Lyase.

Fig. 1

Proposed biotransformation of HFO-1123 1 to account for its toxicity in sensitive species. Cytochrome P450 (CYP-450)-mediated biotransformation may result in hydrolysis of the 1,2,2-trifluorooxirane 6 into F⁻ and oxalic acid 8 or subsequent conjugation with GSH 7 and urinary excretion of a mercapturic acid 9. The mercapturic acid pathway (lower pathway) is expected to be responsible for HFO-1123 toxicity. HFO-1123 may be conjugated with GSH, likely mediated by glutathione S-transferases (GST), to form S-(1,1,2-trifluorethyl)-L-glutathione (1123-GSH) 2. This conjugate may be cleaved by γ-glutamyltransferases (γ-GT) and aminopeptidase (APN)/dipeptidase (DPP) to form the corresponding cysteine S-conjugate 3 (S-(1,1,2-trifluoroethyl)-L-cysteine; 1123-CYS). 1123-CYS can either be eliminated by N-acetyltransferases (NAT) or bioactivated by β-lyases. The latter reaction may give rise to monofluoroacetic acid 4 (MFA) as the ultimate toxic agent expected to be responsible for HFO-1123 toxicity

Fig. 2

a Time-dependent formation of 1123-GSH in hepatic S9 fractions of mice, rats, rabbits, minipigs, and humans in the presence or absence of an NADPH regenerating system. b Velocity of 1123-GSH formed (ng 1123-GSH/mL/h) in hepatic S9 fractions of mice, rats, rabbits, minipigs, and humans after incubation with HFO-1123. Results are presented as mean ± standard deviation (n = 3)

Fig. 3.

¹⁹F-NMR spectra of rat, human, minipig, and rabbit hepatic S9 subcellular fractions incubated with HFO-1123 in the presence of either GSH or the NADPH regenerating system. The black spectra show fluorine-containing metabolites present at the beginning of the incubation (T_0h) and the gray spectra reveal fluorine-containing metabolites after 6 h of incubation (T_6h). ¹⁹F-NMR signals corresponding to 1123-GSH (δ = − 85.1 ppm, m) were detected in hepatic rat and rabbit S9 fractions in the presence of GSH, but not in S9 of other species. Compared to human and rat S9, increased F⁻ formation was detected in minipig and rabbit hepatic S9 in the presence of an NADPH regenerating system. In contrast to the synthesized 1123-GSH reference compound, which is characterized by two signals (δ = − 85.0 ppm, m, 2F; δ = − 227.9 ppm, m, 1F), only the signal corresponding to the C₁ bound fluorine atoms (δ = − 85.1 ppm, m, 2F) was detected in rat and rabbit incubations close to the limit of detection. Absence of the additional signal (δ = − 227.9 ppm, m, 1F) and signals in other S9 fractions can be explained by the lower sensitivity of ¹⁹F-NMR compared to LC–MS/MS

Fig. 4

a Protein concentration-dependent formation of pyruvate in incubations of minipig liver and human kidney cytosol with 1123-CYS and b inhibition of pyruvate formation in incubations of cytosols and 1123-CYS/TCVC in the presence of AOAA. c In vitro kinetics of β-lyase-mediated cleavage of 1123-CYS in renal and hepatic cytosols (Mouse, rat, minipig, rabbit, and human). Data are presented as mean ± standard deviation of three independent experiments performed in triplicates (n = 3)

Fig. 5

Identification of species-differences in β-lyase-dependent formation of MFA and metabolites containing a CH₂F unit in cytosolic fractions of minipigs, rabbits and humans. Resonances were assigned to MFA (δ = − 217.1 ppm; t, J_HF = 48.4 Hz); unknown CH₂F unit-containing metabolite (I) (δ = − 202.0 ppm; t, J_HF = 48.5 Hz); unknown CH₂F unit-containing metabolite (II) ((δ = − 208.2 ppm, t, J_HF = 48.5 Hz); unknown CH₂F unit-containing metabolite (III) (δ = − 220.2 ppm; t, J_HF = 48.5 Hz)

Fig. 6

Inhibition of the formation of MFA and unknown metabolites containing a CH₂F unit in the presence of AOAA in renal minipig cytosol. Resonances were assigned MFA (δ = − 217.1 ppm; t, J_HF = 48.4 Hz); unknown CH₂F unit-containing metabolite (I) (δ = − 202.0 ppm; t, J_HF = 48.5 Hz) and unknown CH₂F unit-containing metabolite (III) (δ = − 220.2 ppm; t, J_HF = 48.5 Hz)

Fig. 7

Cytotoxicity of 1123-CYS in human (HK-2), rat (NRK-52E) and porcine (LLC-PK1) proximal tubular cell lines after 48 h treatment with a 1123-CYS and b 1123-CYS in the absence and presence of the β-lyase inhibitor AOAA (NRK-52E and HK-2: 125 µM AOAA; LLC-PK1: 125 µM/250 µM AOAA). Results are presented as mean ± standard deviation of three independent experiments, each performed in triplicates