(1E)-1-chloro-3,3,3-trifluoroprop-1-ene

Endpoint summary

Link to relevant study record(s)

Reference

Reference 1

Endpoint:

basic toxicokinetics

Type of information:

experimental study

Adequacy of study:

key study

Reliability:

2 (reliable with restrictions)

Rationale for reliability incl. deficiencies:

comparable to guideline study with acceptable restrictions

Objective of study:

toxicokinetics

Qualifier:

no guideline followed

GLP compliance:

no

Radiolabelling:

no

Species:

other: Rat and rabbit in vivo and in vitro; human in vitro

Strain:

other: Sprague Dawley, New Zealand, and human liver microsomes

Sex:

male/female

Details on test animals or test system and environmental conditions:

Male Sprague Dawley rats (Harlan, Venray, Netherlands) and female albino New Zealand rabbits (Charles River, Sulzfeld, Germany) were used for all inhalation exposures. Animal weights ranged from 212 g to 244 g for male rats and from 2.4 kg to 2.8 kg for female rabbits. All animals were stabilized for at least 7 days and examined daily (general conditions, skin and fur, eyes, respiration, nose, oral cavity and external genitalia) during the stabilization period to confirm suitability for the study. Signs of toxic effects were recorded during treatment and post treatment period.

Route of administration:

inhalation

Details on exposure:

Inhalation exposure of rats and rabbits to trans-HCFO-1233zd: Male rats (n=5/concentration) and female rabbits (n=3/concentration) were exposed to target concentrations of 0, 2,000, 5,000 and 10,000 ppm of trans-HCFO-1233zd for 6 h in open dynamic exposure chambers (Bayer et al., 2002; Schuster et al., 2008, 2010). trans-HCFO-1233zd chamber concentrations were determined by GC/MS (HP-Plot Q column). Chamber air samples (100 μL) were injected (split 5:1). The oven temperature was increased from 60 to 180°C within 7 min and a mass fragment characteristic for trans-HCFO-1233zd was recorded (trans-HCFO-1233zd, 95 [m/z], Rt=7.7 min; internal standard—oxygen, 32 [m/z], Rt=1.8 min). Quantification of trans-HCFO-1233zd was based on calibration curves of air samples containing known amounts of trans-HCFO-1233zd.

Duration and frequency of treatment / exposure:

6 hours

Dose / conc.:

2 000 ppm

Dose / conc.:

5 000 ppm

Dose / conc.:

10 000 ppm

No. of animals per sex per dose / concentration:

5 male rats and 3 female rabbits

Control animals:

yes, plain diet

Details on study design:

- Chemicals: trans-1-Chloro-3,3,3-trifluoropropene (trans-HCFO-1233zd, purity of 99,99%) was supplied by Honeywell (Morristown, New Jersey). 3,3,3-Trifluorolactic acid (96%) was purchased from Alfa Aesar (Karlsruhe, Germany) and ammonia (purity≥99.98%) was purchased fromLinde (Pullach, Germany). All other chemicalswere obtained from Sigma-Aldrich (Taufkirchen, Germany) in the highest available purity. Synthesis and characterization of S-(1,2-dichlorovinyl)-glutathione (Vamvakas et al., 1988) and S-(3,3,3-trifluoro-trans-propenyl)- mercaptolactic acid were reported previously (Schuster et al., 2009). N-acetyl-S-(3,3,3-trifluoro-trans-propenyl)-L-cysteine, N-acetyl-d3-S-(3,3,3-trifluoro-trans-propenyl)-L-cysteine, and S-(3,3,3- trifluoro-trans-propenyl)-glutathione were synthesized.
- Animal treatment: To induce cytochrome P-450 2E1, Sprague Dawley rats were treated with pyridine (100 mg/kg ip, dissolved in isotonic chloride solution) once daily for 5 days before euthanized (Urban et al.,1994). Liver microsomes were prepared according to standard protocols (Herbst et al., 1994; Koster et al., 1994; Urban et al., 1994). Cytochrome P-450 2E1 activity in the liver microsomes was determined as 3.33±0.05 [nmol/min×mg (protein)] by analysis of the oxidation of p-nitrophenol (Koop, 1986). Urine was collected for approximately 12 h before initiation of the inhalation exposure and for approximately 48 h after exposure, at 6 and 12 h intervals. Urine was collected at 4°C and volume was measured for each collection interval. Blood samples were taken from rats approximately 12 h before initiation of the inhalation exposure and 0, 24 and 48 h after end of inhalation exposure. All samples were stored at −20°C until analysis.
- In vitro incubations: To study biotransformation, pooled human liver microsomes (Becton Dickinson, Heidelberg, Germany), rabbit liver S9 (Becton Dickinson, Heidelberg, Germany), or rat liver microsomes from pyridine-induced male rats (3 mg protein/mL) were incubated in sealed GC vials with either 10 mM glutathione or a NADPH regenerating system (10 mM NADP, 10 mM glucose-6-phosphate and 0.25 units/mL glucose-6-phosphate dehydrogenase) or 10 mM glutathione and a NADPH regenerating system in potassium phosphate buffer pH 7.4 containing 1 mM EDTA (final volume of 1 mL). Liquid trans-HCFO-1233zd (20 μL) was added through a septum with a microsyringe and the incubations were performed for up to 12 h at 37°C. Reactions were stopped by placing the opened GC vials on ice. For in vitro kinetic experiments either pooled human, male rat, or female rabbit S9 liver fractions (4 mg protein/mL) were pre-incubated with 10 mM glutathione in 100 mM potassium phosphate buffer pH 7.4 containing 1 mM EDTA (end volume 0.5 mL) in sealed GC vials for 5 min at 37 °C. Reaction mixtures were incubated for another 30 min, after adding gaseous trans-HCFO-1233zd (100, 250, 500, 750, 1,000 or 2,000 μL) through the septum with a gastight microliter syringe (equal amount of air was removed). Reactions were stopped by placing the opened GC vials on ice and adding 50 μL of acetonitrile.
- 19F NMR analyses: 19F NMR spectra were recorded with a Bruker DRX 300 NMR spectrometer with a 5 mm fluoride probe operating at 282.4 MHz. 19F chemical shifts were referenced to external CFCl3. Spectra were recorded with a 30° pulse, a pulse length of 13.6 μs and a cycle delay of 6 μs. The acquisition time was 3 s and 10,240 scans were recorded to obtain a good signal to noise ratio (S/N). The 19F spectra were acquired with proton coupling and a spectral width of 200 ppm (+10 to−190 ppm). Before integration of peak areas, an accurate baseline correction and spectra phasing were performed between chemical shifts of −60 and −85 ppm.
- Sample preparation for 19F NMR analysis: After thawing at 4 °C, urine samples (1 mL) were vortexed and then centrifuged for 15 min at 14,000 rpm and at 4 °C. Aliquots (720 μL) of the obtained supernatants were added to 80 μL of deuterium oxide. NMR spectra were analyzed with TopSpin 3.0 (Bruker) and trans-HCFO-1233zd metabolites were identified by comparison of their chemical shifts,multiplicities and coupling constants with spectra of authentic standards (Schuster, 2009; Schuster et al., 2008, 2010).
- LC-MS/MS analyses were performed with an API 3000 mass spectrometer (Applied Biosystems, Darmstadt, Germany) or an API/Q-Trap 2000 (Applied Biosystems-MDS Sciex, Darmstadt, Germany) both coupled to an Agilent 1100 HPLC-pump equipped with an Agilent 1100 autosampler (Agilent, Waldbronn, Germany).
- Sample preparation for LC-MS/MS analysis: Before analysis, frozen urine samples were thawed at 4 °C. Aliquots (1 mL) of urine or incubation mixtures were vortexed and then centrifuged for 15 min at 14,000 rpm and 4°C. Plasma (20 μL) was added to 40 μL of methanol. After keeping samples on ice for 30 min followed by centrifugation (14,000 rpm, 4°C, 10 min), obtained supernatants were mixed with 40 μL of acetonitrile and subjected to a second centrifugation at 14,000 rpm (4°C, 10 min). Obtained urine and plasma supernatants were diluted with water (1:10). If analyte responses were outside the linear range of the calibration curve, an additional dilution by 10 or 100 fold was performed. GC/MS analysis was performed with an Agilent 5973 mass spectrometer and electron impact ionization. The MS was coupled to an Agilent 6890 GC (Agilent, Santa Clara, United States). Analytical columns were either a HP-Plot Q column (30 m∗0.32 mm i.d., 20 μm film thickness; Agilent, Santa Clara, United States) or an Agilent DB-WAX column (30 m∗0.25 mm i.d., 25 μm film thickness; Agilent, Santa Clara, United States). Helium served as carrier gas at a flow rate of 1 mL/min.
- Quantification of N-acetyl-S-(3,3,3-trifluoro-trans-propenyl)-L-cysteine: Diluted samples (10 μL) were added to 90 μL of a 90 pg/μL internal standard solution (N-acetyl-d3-S-(3,3,3-trifluoro-trans-propenyl)-Lcysteine), vortexed, transferred into glass autosampler vials and sealed. Aliquots (10 μL) were injected into the LC. Separation was achieved using a Reprosil Pur C18 AQ column (2∗150 mm, 5 μm; Maisch, Ammerbuch, Germany) with a gradient elution using water containing 0.1% formic acid (solvent A) and methanol containing 0.1% formic acid (solvent B). The elution conditions were: flow-rate of 300 μL/min; 2% solvent B for 1 min followed by a linear increase to 90% solvent B within 5 min, then 90% solvent B for another 3 min. R2-values of the calibration curves were >0.996 and deviations between repeatedly measured samples were 5.4±5.0%. The LOQ for N-acetyl-S-(3,3,3-trifluoro-trans-propenyl)-L-cysteine was between
0.21 and 0.26 μg/L (0.81–1.00 nmol/L) in rat and rabbit urine, with LODs between 0.05 and 0.07 μg/L (0.19–0.26 nmol/L). In plasma, the quantitation method for N-acetyl-S-(3,3,3-trifluoro-trans-propenyl)-L-cysteine achieved a LOQ of 0.62 μg/L (2.38 nmol/L) and a LOD of 0.17 μg/L (0.65 nmol/L). R2-values of the calibration curves in plasma were >0.998 and deviations between repeatedly measured samples were 6.6±4.5%.
- Quantification of 3,3,3-trifluorolactic acid. An ion pair chromatography combined with mass spectrometry was developed based on a published procedure (Loos and Barcelo, 2001). Samples (7.5 μL) were added to 92.5 μL of a 150 pg/μL internal standard solution (dichloroacetic acid), vortexed and transferred into glass autosampler vials. Aliquots (10 μL) were analyzed. Quantification was performed using a Reprosil Pur C18 column (2∗150 mm, 5 μm; Maisch, Ammerbuch, Germany) with an isocratic elution by water containing 5 mM triethylamine and 5% acetonitrile adjusted to pH 6.2 with acetic acid at a flow-rate of 200 μL/min for 10 min. R2-values of the calibration curves were >0.997 and deviations between repeatedly measured samples were 0.5±11.4%. The limit of quantification (LOQ) for 3,3,3-trifluorolactic acid was 0.43 μg/L (2.99 nmol/L) with a limit of detection (LOD) of 0.14 μg/L (0.97 nmol/L) in urine.
- Quantification of S-(3,3,3-trifluoro-trans-propenyl)-glutathione. Incubation supernatants (50 μL) were added to 50 μL of a 500 pg/μL solution of the internal standard S-(1,2-dichlorovinyl)-glutathione, vortexed and transferred into glass autosampler vials. Aliquots (10 μL) were injected. Analytes were separated on a Reprosil Pur C18 column (2∗150 mm, 5 μm; Maisch, Ammerbuch, Germany) by gradient elution with water containing 0.1% formic acid (solvent A) and acetonitrile containing 0.1% formic acid (solvent B) at a flow-rate of 200 μL/min. Conditions were set to 10% solvent B for 1 min, increasing to 90% solvent Bwithin 9 min followed by solvent B at 90% for 3 min. R2-values of the calibration curves were >0.999 and deviations between repeatedly measured samples were 7.7±3.7%. The limit of quantification (LOQ) for S-(3,3,3-trifluoro-trans-propenyl)-glutathione was 5.5 μg/L (13.7 nmol/L) with a limit of detection (LOD) of 1.7 μg/L (4.2 nmol/L) in matrix.
- Identification of minor metabolites. Trifluoroacetic acid, 3,3,3-trifluoropropionic acid. and 3,3,3-trifluoro-1,2-propanediol were identified by GC/MS (Schuster et al., 2008). 3,3,3-Trifluoropropionic acid was identified by its characteristic mass fragments m/z 142, 111, 83 and 69 (Rt=7.5 min), trifluoroacetic acid by m/z 59 and 69 (Rt=4.6 min) and 3,3,3-trifluoro-1,2-propanediol by m/z 80, 69 and 31 (Rt=9.6 min). Analysis of 3,3,3-trifluoro-1-propanol involved heating of 200 μL urine at 85 °C for 30 min. A 1 mL headspace sample was injected splitless on a HP-Plot Q column with a carrier gas flow rate of 2 mL/min. Oven temperature was kept at 40 °C for 1 min, followed by an increase to 160 °C within 8 min holding the final temperature for 1 min. The intensities of m/z 95 and 69 were recorded with dwell times of 100 ms during analysis. 3,3,3-Trifluoro-1-propanol was eluted at 9.1 min.
- Determination of S-(3,3,3-trifluoro-trans-propenyl)-mercaptolactic acid, S-(3,3,3-trifluoro-trans-propenyl)-L-cysteine and screening for the presence of other conceivable urinary metabolites was performed by LC-MS/MS analysis on a Reprosil Pur C18 AQ column (2∗150 mm, 5 μm; Maisch, Ammerbuch, Germany) using gradient elution with 0.1% formic acid in ACN or methanol and 0.1% formic acid in water at a flow rate of 200 μL/min. Organic mobile phase content was increased from10% to 90% within 30 min and held for further 2 min. Scans to detect conceivable metabolites were performed with negative ionization either in the sensitive multiple reaction monitoring (MRM)/enhanced product ion (EPI), enhanced mass spectrometry (EMS)/EPI or constant neutral loss (CNL)/EPI modes monitoring loss of 20 Da (hydrogen fluoride), 73 Da (glycine), 80 Da (sulfate), 86 Da (mercaptopyruvic acid), 87 Da (cysteine), 88 Da (mercaptolactic acid), 129 Da (mercapturic acid or glutathione), and 176 Da (glucuronide).

Statistics:

All data were analyzed by Prism 5.01 (GraphPad Software, Inc, La Jolla, CA 92037 USA).

Metabolites identified:

yes

Details on metabolites:

- Incubation of trans-HCFO-1233zd with liver microsomes: These incubations were performed both in the presence or absence of glutathione and/or NADPH regenerating system and the incubation mixtures were analyzed by 19F NMR and LC-MS/MS. The major signal in the 19F NMR spectra observed in rat, rabbit, and human was allocated to S-(3,3,3-trifluoro-trans-propenyl)-glutathione (δ=−62.0 ppm; dd, 3JHF=6.5 Hz, 4JHF=2.1 Hz). The structure of S-(3,3,3-trifluoro-trans-propenyl)-glutathione was confirmed by LC-MS/MS analysis. Minor metabolites observed in 19F NMR spectra of incubations of trans-HCFO-1233zd in rat liver microsomes were identified as 3,3,3-trifluorolactic acid (δ=−75.4 ppm; d, 3JHF=8.2 Hz), 3,3,3-trifluoro-1,2-propanediol (δ=−77.2 ppm; d, 3JHF=7.3 Hz) and 3,3,3-trifluoro-1-hydroxyacetone (δ=−83.4 ppm; s). 19F NMR spectra of rabbit microsomal incubations also showed several minor metabolites and two signals were attributed to 3,3,3-trifluoro-1,2-propanediol (δ=−77.2 ppm; d, 3JHF=7.3 Hz) and 3,3,3-trifluoro-1-propanol (δ=−64.3 ppm; t, 3JHF=11.2 Hz). The minor biotransformation products were present in both incubations containing a NADPH regenerating system and glutathione or only a NADPH regenerating system. In order to identify the additional signals in the 19F NMR spectra, LC-MS/MS experiments were performed in the CNL/EPI, EMS/EPI and MRM/EPI scan mode with negative ionization. Four minor metabolites were thus characterized in incubations in the presence of glutathione. All EPI spectra showed the characteristic fragmentation pattern of glutathione adducts (m/z 272, m/z 254 and m/z 179) (Farkas et al., 2007). Two of the EPI spectra suggested the presence of 2-S-(3,3,3-trifluoro-1, 1-propanediol)-glutathione, and 1-S-(1-chloro-3,3,3-trifluoropropyl)-glutathione or 2-S-(1-chloro-3,3,3-trifluoropropyl)-glutathione. EPI spectra indicative of 2-S-(3,3,3-trifluoro-1-propanol)-glutathione and 2-S-(3,3,3-trifluoropropionic acid)-glutathione were only obtained in supernatants of rat liver microsomes. The latter metabolites will give doublets in 19F NMR spectra and the unidentified signals observed in the 19F NMR spectra of rat microsomal incubations at −73.5 ppm (d, 3JHF=4.9 Hz) and at −84.8 ppm (d, 3JHF=4.0 Hz) may be assigned to these structures. Kinetics of the formation of S-(3,3,3-trifluoro-trans-propenyl)-glutathione were characterized in rabbit, rat and human subcellular liver fractions. Kinetic parameters were determined by nonlinear regression analysis as Km (trans-HCFO-1233zd headspace gas concentration) — 249,000 ppm (rabbit), 213,400 ppm (rat) and 163,400 ppm (human) and Vmax as 21.8 [pmol/min ∗mg protein] (rabbit), 13.1 [pmol/min ∗mg protein] (rat), and 5.0 [pmol/min∗mg protein] (human). Kinetic analyses of the oxidative biotransformation of trans-HCFO-1233zd in liver subcellular fractions could not be performed due to the low sensitivity of the quantitation procedures available for the highly polar products formed.
- Inhalation exposures of rats and rabbits to trans-HCFO-1233zd: Mean chamber concentrations of trans-HCFO-1233zd were 1,961±261, 5,084±549 and 10,089±818 ppm for rabbit and 1,977±256, 5,220±1147 and 10,128±714 ppm for rat exposures. Predominant metabolites in 19F NMR spectra of rat urine were 3,3,3-trifluorolactic acid (δ=−75.3 ppm; d, 3JHF=8.2 Hz) and N-acetyl-S-(3,3,3-trifluoro-trans-propenyl)-L-cysteine (δ=−61.8 ppm; dd, 3JHF=6.6 Hz, 4JHF=2.1 Hz), representing 32% and 40%, of the total fluorine signal area (0–6 h post inhalation urine fraction), respectively. Additional identified metabolites were S-(3,3,3-trifluoro-trans-propenyl)-mercaptolactic acid (δ=−61.7 ppm; dd, 3JHF=6.6 Hz, 4JHF=2.1 Hz), trifluoroacetic acid (δ=−75.4 ppm; s), 3,3,3-trifuoro-1,2-dihydroxypropane (δ=−77.2 ppm; d, 3JHF=7.1 Hz), and 3,3,3-trifluoropropionic acid (δ=−63.4 ppm; t, 3JHF=11.2 Hz). The structures of the chemical entities giving the doublets at −71.3 and −77.1 ppm as well as the doublet of doublets at −61.5 ppm could not be elucidated. In 19F NMR spectra of rabbit urine, the major metabolite was N-acetyl-S-(3,3,3-trifluoro-trans-propenyl)-L-cysteine (δ=−61.8 ppm; dd, 3JHF=6.6 Hz, 4JHF=2.1 Hz) representing 46% of total 19F NMR signal area (0–6 h post inhalation urine fraction). S-(3,3,3-trifluorotrans-propenyl)-mercaptolactic acid (δ=−62.1 ppm; dd, 3JHF=6.5 Hz, 4JHF=2.1 Hz), trifluoroacetic acid (δ=−75.4 ppm; s) and 3,3,3-trifuoro-1,2-dihydroxypropane (δ=−77.2 ppm; d, 3JHF=7.1 Hz) were also present in rabbit urine. Besides, additional signals were observed at −72.3 (d, 3JHF=6.8), −72.4 (d, 3JHF=7.0 Hz) and −62.1 ppm dd, 3JHF=6.5 Hz, 4JHF=2.1 Hz), the latter was assigned to S-(3,3,3-trifluoro-trans-propenyl)-L-cysteine. A weak signal for 3,3,3-trifluoro-1-propanol was seen at a chemical shift of −64.3 ppm (t, 3JHF=11.2 Hz). Again, the chemical entity giving the doublets appearing at −72.3, −72.4 and −77.1 ppm could not be identified, signals indicative of the presence of 3,3,3-trifluorolactic acid were not present. The structures of the major metabolites N-acetyl-S-(3,3,3-trifluoro-trans-propenyl)-L-cysteine, 3,3,3-trifluorolactic acid, S-(3,3,3-trifluoro-trans-propenyl)-mercaptolactic acid, and S-(3,3,3-trifluoro-trans-propenyl)-L-cysteine were confirmed by LC-MS/MS in the sensitiveMRM/EPI scanmode. Metabolite retention times and fragmentation patterns in the negative scan mode were compared with synthesized or purchased standards. More detailed analysis by LC-MS/MS revealed an EPI mass spectrum assigned to the S-oxide of N-acetyl-S-(3,3,3-trifluoro-trans-propenyl)-L-cysteine. The fragmentation showed the characteristic pattern of trans-HCFO-1233zd conjugates (loss of two molecules of hydrogen fluoride). This signal was only observed in rabbit urine. Minor volatile metabolites in urine were detected by GC/MS analysis. Trifluoroacetic acid, 3,3,3-trifluoropropionic acid and 3,3,3-trifluoro-1,2-propanediol were confirmed in rat and rabbit urine. 3,3,3-Trifluoro-1-propanol was only present in rabbit urine. While the presence of trifluoroacetic acid was verified by 19F NMR and GC/MS analysis, no reasonable mechanisms of formation can be developed and the detection of trifluoroacetic acid may be an artifact (Bayer et al., 2002; Schuster et al., 2008). Excretion kinetics of metabolites were determined by quantitation of N-acetyl-S-(3,3,3-trifluoro-trans-propenyl)-L-cysteine in rat/rabbit urine and in rat plasma and of 3,3,3-trifluorolactic acid in rat urine. Due to low amounts of rat blood, 3,3,3-trifluorolactic acid kinetics could not be determined in plasma. The urinary excretion profiles of N-acetyl-S-(3,3,3-trifluoro-transpropenyl)-L-cysteine showed the highest concentrations in urine fractions collected between 0 and 6 h after the end of the inhalation exposure in both rats and rabbits. In contrast, 3,3,3-trifluorolactic acid peaked in urine samples collected between 6 and 12 h after the end of the inhalation exposures. Within the first 24 h after termination of the inhalation exposures,>98.2±3.0% (rats) and >97.1±3.5% (rabbits) of the N-acetyl-S-(3,3,3-trifluoro-trans-propenyl)-L-cysteine and >87.2±4.4% of 3,3,3-trifluorolactic acid were recovered. Elimination half-life times were determined as 5.6±3.8 h for 3,3,3-trifluorolactic acid in rats and 4.3±4.4 h and 2.9±2.1 h for N-acetyl-S-(3,3,3-trifluoro-trans-propenyl)-L-cysteine in rats and rabbits, respectively. Mean peak plasma levels of N-acetyl-S-(3,3,3-trifluoro-trans-propenyl)-L-cysteine were determined in all blood samples taken before and after the inhalation exposure in all three exposure groups. In general, concentration levels of N-acetyl-S-(3,3,3-trifluoro-trans-propenyl)-L-cysteine were higher in the 10,000 ppm exposure group compared to 5,000 and 2,000 ppm, but large standard deviations, in part due to the low concentrations of the analyte present, did not permit meaningful analysis.

Discussion

In analogy to published data for trans-1,1,1,3-tetrafluoropropene (Schuster et al., 2009), trans-HCFO-1233zd is metabolized by both glutathione conjugation and by oxidative metabolism by cytochrome P-450 2E1. This is consistent with the information available on the biotransformation of structurally related compounds and the formation of all identified products of the biotransformation of trans-HCFO-1233zd can be readily explained as formed by these two pathways. The identified metabolites in intact animals show that a number of downstream processing reactions result in a complex metabolite pattern.

The biotransformation of trans-HCFO-1233zd mediated by glutathione S-transferase catalyzed addition–elimination reaction of glutathione leads to N-acetyl-S-(3,3,3-trifluoro-trans-propenyl)-L-cysteine as final excretory product. Consistent with this observation, S-(3,3,3-trifluoro-trans-propenyl)-glutathione is the major metabolite formed from trans-HCFO-1233zd in liver subcellular fractions. Renal processing of S-(3,3,3-trifluoro-trans-propenyl)-glutathione by γ-glutamyl-transpeptidase and dipeptidases (Anders and Dekant, 1998) results in S-(3,3,3-trifluoro-trans-propenyl)-Lcysteine. This cysteine S-conjugate is further processed by three different pathways. i) Cysteine S-conjugate transaminase to give S-(3,3,3-trifluoro-trans-propenyl)-mercaptopyruvic acid followed by reduction to S-(3,3,3-trifluoro-trans-propenyl)-mercaptolactic acid; ii) renal cysteine S-conjugate β-lyase dependent biotransformation (Anders et al., 1992; Monks et al., 1990) to finally give 3,3,3-trifluoropropionic acid and 3,3,3-trifluoro-1-propanol; and iii) N-acetyl-transferase to form N-acetyl-S-(3,3,3-trifluoro-trans-propenyl)-L-cysteine which may be further oxidized to N-acetyl-S-(3,3,3-trifluoro-trans-propenyl)-L-cysteine S-oxide.

The second metabolic pathway for trans-HCFO-1233zd is catalyzed by cytochrome P-450, initially resulting in the formation of 1-chloro-3,3,3-trifluoro-1,2-epoxypropane. This epoxide is hydrolyzed to 3,3,3-trifluoro-2-hydroxypropanal which may be oxidized to 3,3,3-trifluorolactic acid or reduced to 3,3,3 -trifluoro-1,2 -propanediol. The latter may be subject to further phase two reactions. Alternatively, ring opening of 1-chloro-3,3,3 -trifluoro-1,2 -epoxypropane may occur by a reaction with glutathione to give 2-S-(3,3,3-trifluoropropanal)-glutathione, which may be oxidized to 2-S-(3,3,3-trifluoropropionic acid)-glutathione or reduced to 2-S-(3,3,3-trifluoro-1-propanol)-glutathione. The glutathione S-conjugates might be subject to further processing by the enzymes of the mercapturic acid pathway. However, the compounds likely are only formed in very low yields and thus could not be detected in urine samples.

The structures of the minor metabolites proposed to be formed in the liver subcellular fractions are consistent with the reaction schemes developed. Tentative minor in vitro metabolites detected by LC-MS/MS in all species were the glutathione conjugate 1-S-(1-chloro-3,3,3-trifluoro-propyl)-glutathione or 2-S-(1-chloro-3,3,3 -trifluoro-propyl)-glutathione and the hydrate of 2-S-(3,3,3-trifluoropropanal)-glutathione. Addition of glutathione to double bonds may not necessarily be followed by elimination of HCl since addition reactions of glutathione to the haloalkene have been described (Dekant, 2003). 2-S-(3,3,3-trifluoropropionic acid)-glutathione and 2-S-(3,3,3-trifluoro-1-propanol)-glutathione, products of the glutathione-mediated ring opening of the epoxide were only detected in supernatants of rat microsomes. Both are formed by oxidation or reduction of the intermediate 2-S-(3,3,3-trifluoropropanal)-glutathione. These products were only detected using highly sensitive LC-MS/MS methods. Related 19F NMR signals were not observed, likely concentrations were below the much higher limit of detection of the 19F NMR method.

The 19F NMR spectra of rat and rabbit urine indicated the presence of several unidentified metabolites. The signal (δ=−61.5 ppm) close to the major metabolite N-acetyl-S-(3,3,3-trifluoro-transpropenyl)-L-cysteine (9, Scheme 1) is likely a processing product of S-(3,3,3-trifluoro-trans-propenyl)-glutathione pathway because of its similar multiplicity, 19F chemical shift and coupling constants. The unidentified doublets observed at 19F chemical shifts below −71 ppm possibly have a hydroxyl group in alpha position to the trifluoromethyl group, also indicated by their similar 19F chemical shift, multiplicity and coupling constants compared to 3,3,3-trifluorolactic acid and 3,3,3-trifluoro-1,2-propanediol. Traces of N-acetyl-S-(3,3,3-trifluoro-trans-propenyl)-L-cysteine S-oxide, observed by LC-MS/MS analysis in rabbit urine, were below the limit of detection of the 19F NMR method.

The metabolite pattern observed indicates species differences between rats and rabbits and the two pathways of trans-HCFO-1233zd biotransformation may be competitive. Whereas 3,3,3-trifluorolactic acid was the main urinary metabolite of trans-HCFO-1233zd in the rat, N-acetyl-S-(3,3,3-trifluoro-trans-propenyl)-L-cysteine was the main urinary biotransformation product in the rabbit. In the rat, urinary metabolite quantification showed an increase of N-acetyl-S-(3,3,3-trifluoro-trans-propenyl)-L-cysteine excretion compared to 3,3,3-trifluorolactic acid with increasing trans-HCFO-1233zd exposure concentrations. The ratio of N-acetyl-S-(3,3,3-trifluoro-transpropenyl)- L-cysteine to 3,3,3-trifluorolactic acid was 0.09 at 2,000 ppm, 0.15 at 5,000 ppm and 0.43 at 10,000 ppm. The increased contribution of the glutathione S-transferase catalyzed pathway also explains that the oxidation of trans-HCFO-1233zd in in vitro incubations at very high trans-HCFO-1233zd concentrations is only a minor pathway.

Quantification of the major metabolites 3,3,3 -trifluorolactic acid and N-acetyl-S-(3,3,3-trifluoro-trans-propenyl)-L-cysteine revealed rapid excretions, with elimination half-life times of less than 6 h for both substances in both species. Calculations based on total urinary recoveries for N-acetyl-S-(3,3,3-trifluoro-trans-propenyl)-L-cysteine or 3,3,3-trifluorolactic acid and estimations of trans-HCFO-1233zd uptake, revealed very low overall biotransformation amounts. The extent of biotransformation of trans-HCFO-1233zd is within the same order of magnitude as trans-1,1,1,3-tetrafluoropropene (Schuster, 2009) with app. 0.007% of the received dose undergoing biotransformation in the rat, the remaining material is likely exhaled due to its high volatility. In vitro enzyme kinetics confirmed the low in vivo biotransformation amount and showed a lower glutathione S-transferase affinity to trans-HCFO-1233zd in humans compared to rodents.

The potential reactive metabolite 1-chloro-3,3,3-trifluoro-1,2 -epoxypropane formed by cytochrome P-450 is formed only in very low amounts due to low overall biotransformation and is apparently rapidly detoxified by epoxide hydrolase. Therefore, covalent binding with tissue nucleophiles may be efficiently prevented by hydrolysis or by glutathione conjugation. Formation of reactive metabolites (thioketenes or thioacylhalodies) by the glutathione S-transferase/renal β-lyase pathway observed with other haloolefins is unlikely for trans-HCFO-1233zd due to the lack of appropriate leaving groups (Dekant et al., 1990; Monks et al., 1990; Vamvakas et al., 1989; Volkel et al., 1998).

In summary, the low biotransformation amount and the rapid excretion of metabolites as well as the absence of potent toxic biotransformation products in both species, support the results of prior studies that trans-HCFO-1233zd has only a low potential for toxicity in mammals. The data on metabolism and toxicokinetics of trans-HCFO-1233zd are in good agreement with previously published data for the related compound trans-1,3,3,3-tetrafluoropropene by Schuster et al. (2009).

Conclusions:

No bioaccumulation potential based on study results. The test substance undergoes both oxidative biotransformation and glutathione conjugation at very low rates. The very low extent of biotransformation and the rapid excretion of metabolites formed are consistent with the very low potential for toxicity of the test substance in mammals.

Executive summary:

The biotransformation of trans-HCFO-1233zd and kinetics of metabolite excretion with urine were assessed in vitro and in animals after inhalation exposures. To characterise biotransformation in vitro, liver microsomes from rats, rabbits and humans were incubated with trans-HCFO-1233zd. Male Sprague Dawley rats and female New Zealand White rabbits were exposed by inhalation to 2,000, 5,000 and 10,000 ppm of trans-HCFO-1233zd for 6 hours and urine was collected for 48 hours after the end of the exposure. Study samples (urine, plasma, microsomal incubation supernatants) were analysed for metabolites using 19F-NMR, LC-MS/MS and GC/MS. S-(3,3,3-trifluoro-trans-propenyl)-glutathione was identified as predominant metabolite of trans-HCFO-1233zd in rat, rabbit and human liver microsomes in the presence of glutathione. Products of the oxidative biotransformation of trans-HCFO-1233zd were only minor metabolites when glutathione was present. In rats, both 3,3,3-trifluorolactic acid and N-acetyl-(3,3,3 -trifluoro-trans-propenyl)-L-cysteine were observed as major urinary metabolites. 3,3,3-trifluorolactic acid was not detected in the urine of rabbits exposed to trans-HCFO-1233zd. Several minor metabolites formed by alternative reactions of S-(3,3,3-trifluoro-trans-propenyl)-glutathione were also excreted. Quantitation showed rapid excretion of both metabolites with urine t1/2 of less than 6 hours in both species. Based on metabolite recovery in urine and estimated doses of trans-HCFO-1233zd received by inhalation, the extent of biotransformation of trans-HCFO-1233zd was determined as approximately 0.01 % of received dose in rabbits and approximately 0.002 % in rats. The remaining material is likely exhaled due to its high volatility. The metabolite structures show that trans-HCFO-1233zd undergoes both oxidative biotransformation and glutathione conjugation at very low rates. The very low extent of biotransformation and the rapid excretion of metabolites formed are consistent with the very low potential for toxicity of trans-HCFO-1233zd in mammals.

Reference 2

Endpoint:

basic toxicokinetics in vivo

Type of information:

other: physiologically based toxicokinetic modeling

Adequacy of study:

key study

Reliability:

2 (reliable with restrictions)

Rationale for reliability incl. deficiencies:

comparable to guideline study with acceptable restrictions

Objective of study:

other: physiologically based toxicokinetic modeling

Qualifier:

no guideline available

GLP compliance:

no

Radiolabelling:

no

Details on study design:

Exposure simulation using the BBB PBPK model
The breath-by-breath physiologically-based toxicokinetic (BBB PBPK) inhalation model used in this assessment contains an extensive mathematical representation of the human respiratory tract that includes a pulmonary region containing the gas volume of the lung and pulmonary tissue and capillaries, and proximal and distal dead space in the upper respiratory and tracheobronchial region, as taken from the work of Vinegar et al, 2000. The model also contains physiological compartments for fat, liver, gastrointestinal tissue (gut), and richly and slowly perfused tissues. The mathematical construction of the model allows for the evaluation of kinetic processes occurring during the first few breaths of exposure. The model simulates sinusoidal inhalation and exhalation breathing patterns, time delay for air and gases to travel through the dead-space volume, absorption of inhaled gases within the pulmonary region and subsequent exchange of the gas with blood. Acute exposures were assessed for a female human using the BBB PBPK. The use of the BBB PBPK model is computationally intensive and is appropriately suited for simulation of short term exposure periods or in situations where exposure is highly transient. A modified version of the model which is more computationally efficient (CF PBPK model) was employed for longer term simulations where it is computationally identical to the BBB PBPK model.
Exposure simulation using the CF PBPK model
The CF PBPK model assumes instantaneous equilibrium between exhaled air and arterial blood flow, and constant inhalation and exhalation flow rates. At exposures of more than a few minutes, the CF PBPK model provides peak concentrations that are essentially identical to the BBB PBPK model and in a much more computationally efficient manner. The CF PBPK differs from the BBB PBPK primarily in its absence of the sinusoidal breath-by-breath variation in the arterial blood concentration during the first few minutes of exposure. Whereas the BBB PBPK model was used to simulate very brief exposures, the faster CF PBPK code was used for modeling of more extensive exposures.
Summary of physiological input parameters
The uptake and distribution of the inhaled chemical are described via the application of physiological input parameters which include body weight, tissue volumes, cardiac output, alveolar ventilation rate and additional pulmonary parameters. Values for each of the physiological parameters were taken from the published literature.
Summary of chemical specific input parameters
The only chemical specific input parameters for the test substance were the experimentally determined partition coefficients. Briefly, tissue/air partition coefficients were measured at equilibrium for rat and rabbit blood, liver, muscle and fat as well as human (female and mixed male and female) blood at 37°C. There is no appreciable metabolic activity towards the test substance. As such, the only route of elimination in the PBPK model is via exhalation.
Sensitivity analysis of the BBB and CF PBPK models
A sensitivity analysis of the BBB and CF PBPK model predicted blood concentration at steady state in relation to changes in model parameters was calculated for a 400 ppm exposure scenario in a female human. Normalized sensitivity coefficients were calculated with the forward difference method and normalized to both the parameter value and response variable.

INHALATION EXPOSURE SCENARIOS
A series of exposure scenarios were evaluated using BBB and CF PBPK models for humans, rabbits and rats. The results of the simulations are expressed in peak blood concentrations and AUC.
- Scenario 1: Acute exposure of adult female human: This scenario represents the brief inhalation exposure of an adult female human to a concentration of 25,000 ppm v/v test substance. The concentration is based on the NOEL for cardiac sensitization in dogs. This scenario is considered a worst case as it is unlikely to occur except in the case of catastrophic failure. This scenario would be short term as personnel would be evacuated under these conditions. This scenario was simulated by the BBB PBPK model. The specific exposure conditions of this scenario were as follows: Concentration of test substance in air: 25,000 ppm v/v (ERPG-3); Duration of exposure: 0.5, 1, 5, 15, 30 and 60 minutes; Duration of simulation: 8 hours (to evaluate clearance from body). The results summarized from this exposure are the peak concentrations in arterial blood and the AUC which represents the dose received.
- Scenario 2: Occupational exposure of adult female human: This scenario represents a situation in which an individual in the workplace is occupationally exposed to an ambient concentration of the test substance. A variety of exposure time frames are assessed to allow for the evaluation of repetitive daily exposures ranging from 1 day up to 4 weeks of workplace exposure. The exposure concentration of the test substance selected for this scenario is the proposed OEL of 400 ppm v/v 8-hr TWA which was based on a LOAEL of 4,000 ppm from a 13 week exposure study in rats. The specific exposure conditions used in the CF PBPK model for this scenario are as follows: Concentration of the test substance in air: 400 ppm v/v; Duration of exposure and simulation: (a) 6 hours/day for 1 day, (b) 6 hours/day for 1 “work week” *, (c) 6 hours/day for 2 “work weeks” *. (d) 6 hours/day for 4 “work weeks” *, (e) 8 hours/day for 1 day, (f) 8 hours/day for 1 “work week” *, (g) 8 hours/day for 2 “work weeks” * (* A “work week” is defined as 5 consecutive days of exposure followed by 2 days without exposure”). The results summarized from this exposure are the peak concentrations in arterial blood and the AUC which represents the dose received.
- Scenario 3: Repetitive exposure of pregnant rabbit: Pregnant rabbits were exposed to various concentrations of the test substance, for 6 hours per day from day 6 through day 28 of gestation. Based on the responses of the rabbits to this repetitive exposure, a NOEL 15,000 ppm v/v was reported. To permit comparison of the arterial blood concentrations of the rabbits with humans, the experimental exposure scenario for the test substance was simulated with the CF PBPK model and applied to the test substance. The specific exposure conditions for this scenario are as follows: Concentration of test substance in air: 15,000 ppm v/v; Duration of exposure and simulation (a) 1 hour/day for 1 day, (b) 6 hours/day for 1 day; (c) 6 hours/day for 14 days; (d) 6 hours/day for 28 days The results summarized from this exposure are the peak concentrations in arterial blood and the AUC which represents the dose received.
- Scenario 4: Repetitive exposure of rat: Rats were exposed to various concentrations of the test substance in air 6 hours per day and 5 days per week for up to thirteen weeks. The NOEL was 4,000 ppm based on increased potassium levels in plasma. To allow comparison of the resulting arterial blood concentrations of the rats with that of humans, the experimental exposure scenario used for the rats was simulated with the CF PBPK model and applied to the test substance. The specific exposure conditions for this scenario are as follows: Concentration of the test substance in air: 4,000 ppm v/v; Duration of exposure and simulation (a) 1 hour/day for 1 day, (b) 6 hours/day for 1 day, (c) 6 hours/day for 14 days, (d) 6 hours/day for 28 days. The results summarized from this exposure are the peak concentrations in arterial blood and the AUC which represents the dose received.
- Monte Carlo Simulation: MC analysis with the CF PBPK model was conducted for scenario 2 (human occupational exposure) in order to reduce uncertainty in our safety assessment of the test substance. The MC simulations were performed by randomly selecting parameter values from a distribution and repeating simulations to produce a population of possible dosimetrics. Two types of distributions are applied in this analysis: lognormal and normal distributions. All rates and ratios including cardiac output, pulmonary ventilation and partition coefficients are simulates with lognormal distributions. Body weight, tissue fractional volumes, and tissue fractional perfusion rates are simulated as normally distributed. The published values for the parameters were taken as the mean values, and the other MC parameters for each, standard deviation, lower truncation, and upper truncation, were calculated based on a standard deviation set to ½ the mean value. The lower truncation was limited to 1.95 standard deviations and the upper truncation was set to two standard deviations. A random draw of parameter values is repeatedly simulated until a population of output values is achieved. After a total of 500 runs, the results are summarized as mean, standard deviation, 5th percentile and 95th percentile for each scenario for arterial blood peak concentration and AUC.

SIMULATION RESULTS

The simulation results for each scenario are summarised below with a primary focus on determination of the peak concentration of the test substance in arterial blood as well as the AUC of the arterial blood concentration-time curve for each exposure scenario.

- Sensitivity analysis: As the only route for uptake and elimination in the model is via inhalation and exhalation, the most sensitive parameter in the model is the blood/air partition coefficient. For the BBB PBPK model, blood concentration was also sensitive to the larger volume tissues such as such as gut, slowly perfused and fat, which have an impact on acute uptake. As would be expected, the vast majority of model parameters had little to no impact on the steady state blood concentration.

- Scenario 1: Acute exposure of adult female human: The exposure of an adult female human to a concentration of 25,000 ppm v/v test substance resulted in a peak concentration of 33.7 to 64.6 mg/L in arterial blood after exposure times of 0.5 to 60 minutes, respectively. A steady-state concentration of 67.0 mg/L is reached after approximately 14 hours of continuous exposure with over half of this concentration being reached after 30 seconds of exposure, indicating very rapid uptake and distribution of the test substance in humans. The slow increase in blood concentration that occurs between 30 minutes and 11 hours is the result of tissue loading where tissue solubility is greatest for fat which represents approximately 33% of female body mass. Combining rapid uptake distribution with a lack of saturable metabolic clearance for the test substance, the AUC for each exposure period is approximately linear with the time of exposure (i.e. a 5 minute exposure results in an AUC of 5.53 mg/L*hr compared to 16.6 mg/L*hr for a 15 minute exposure).

- Scenario 2: Occupational exposure of adult female human to the test substance: Two scenarios were used with the CF PBPK model for the workplace exposure of an adult female human to the OEL concentration of 400 ppm v/v test substance. The first assumes an exposure of 6 hours per day for periods of one day, one week, 2 weeks and 4 weeks. The second series extended the daily exposure period to 8 hours per day for the same exposure durations. Each work week was simulated as a 5 day period of exposure with 2 days of no exposure representing the weekend. The peak concentration in blood was approximately 1.1 mg/L regardless of exposure time. This would indicate that the blood is very close to steady state as an additional two hours of exposure did not result in an increased maximum concentration. As with the acute exposure simulation, the AUC for the different exposure lengths was linear with length of exposure.The model simulation of a human female at 400 ppm v/v for 8 hours/day over 4 weeks indicates the rapid uptake and clearance from blood for the test substance with the concentration falling to near 0.0 before the start of a subsequent day’s exposure. MC analysis for the human female under occupational exposure: The mean maximum blood concentration of approximately 0.98 mg/L was similar for all exposure times and lengths as were the 5th and 95th percentiles (i.e., 0.9 and 1.7 mg/L). A similar result was seen with the model predicted AUC where length of exposure was linear with AUC with a single 8 hour exposure resulting in an AUC of 8.0 mg/L*hr AUC versus 40.3 mg/L*hr for a 5 day exposure (8 hours per day). The 5th and 95th percentiles also shared a similar patter with regards to AUC. A simulation of 50 random draws from the parameter distributions for a female worker exposed for 8 hours over a single day shows that the time slope of the approach to the maximum blood concentration varied slightly, the analysis was more widely influenced by the change in blood/air partitioning as would be expected given the high sensitivity of this parameter.

- Scenario 3: Repetitive exposure of pregnant rabbit: In response to various lengths of daily exposure to 15,000 ppm v/v of test substance in air, the steady-state peak concentration of this chemical in the arterial blood of a pregnant rabbit was predicted to be 62.7 mg/L. The AUC for a two week exposure (6 hours per day, 7 days per week) was 5,270 mg/L*hr while a four week exposure resulted in an AUC of 10,539 mg/L*hr.

- Scenario 4: Repetitive exposure of rat to the test substance: In response to various lengths of daily exposure to 4,000 ppm v/v of test substance in air, the steady-state peak concentration of this chemical in the arterial blood of a rat was 27.8 mg/L with 6 hour per day exposure. The AUC after two weeks of exposure was 1,667 mg/L*hr (6 hours per day, 5 days per week) with a 4 week exposure giving a model predicted AUC of 3,333 mg/L*hr.

 

Ratios of animal NOEL to human OEL dosimetrics

To provide an evaluation of relative risk of human exposure to the test substance, AUC values were calculated for the pregnant rabbit and rat at inhalation concentrations corresponding to the NOEL. The AUC values for potential human exposure at the OEL were then compared with the animal AUC values at the NOEL to determine a corresponding margin of safety. The ratio of animal AUC to the human AUC thus represents the ratio of the dose received by an animal with no observable effect and the dose received by a human at the 8 hr-TWA maximally allowed concentration for the workplace. Taking this approach one step further by incorporating Monte Carlo analysis of the human CF PBPK model allows one to address uncertainty in the predicted safety margins. This is accomplished by taking the ratio of the comparable exposure animal AUC to the 95th percentile for the human at the OEL.

The ratios of animal/human AUC for rabbit/human and rat/human: The rabbit AUC for 6 hours (one study day) of exposure at the NOEL was 34.6 times higher than the human AUC for 6 hours of exposure and 48.2 times for the 2 week and 4 week exposure scenarios. The margins were reduced to 26.0 and 36.2 when comparing the 6 hour/day exposure in the rabbit to the 8 hour/day exposure in the human. For the rat, the lower NOEL combined with a 5 day per week study design resulted in lower margins of exposure with one day exposure margins of 15.7 when the 6 hour exposure was used for human and 11.8 when the 8 hour exposure scenario was used for the calculation. Adding variability to the human model had the expected effect of reducing the calculated safety margins given the 95th percentile would be greater than safety margins calculated with the mean model parameter values. However, these estimates include our uncertainty in model parameters as well as an assessment of population variability and thus, increase confidence in the predicted safety margins. With the conservative approach to the analysis, the safety margin was 36.2 for the rabbit (6 hour - 14 day exposure in rabbit as compared to an 8 hour - 14 day exposure in human). For the rat, the margin of safety was approximately 11.4 for the same exposure comparison.

The CF PBPK model can also be used to estimate the human equivalent concentration (HEC) or the concentration at which the human blood dosimetric matches that of the animal at the NOEL. This is accomplished by adjusting the exposure concentration in the human model until the dosimetric, Cmax and AUC in blood for 28 day exposure in this scenario, is equivalent to the animal at the NOEL. The animal is simulated as the study design. For the rabbit this was 6 hr per day and 7 days per week while the rat was exposed 6 hr per day and 5 days per week. The human scenario used for comparison was an 8 hr per day exposure over 5 days each week. Using the rabbit Cmax of 62.7 mg/L and AUC of 10,539 mg/L*hr, the HEC would be 23,560 or 32,150 ppm. When the rat dosimetrics were used (Cmax = 27.8 mg/L and AUC = 3,333 mg/L*hr) the HEC was 10,040 ppm based upon Cmax and 10,180 ppm when AUC was the basis. The higher exposure required in the human to reach the animal dosimetric is based on several factors. The most important factor is differences in the solubility in blood of the test substance between the species as near steady state conditions are reached with each day of exposure and virtually the entire amount of absorbed compound is cleared before the subsequent day’s start of exposure. That is, the model reaches a stable periodic behavior within the two days of exposure. Other factors include differences in the study design compared to the normal human workday as well as differences in physiology.

Conclusions:

Physiologically based pharmacokinetic (PBPK) models were used to evaluate concentrations of the test substance in arterial blood (peak concentration and area under the time-course curve) in adult female humans, pregnant rabbits and rats in response to inhalation at various concentrations and for various periods of time.

Executive summary:

This report describes the use of physiologically based pharmacokinetic (PBPK) modelling techniques to assess the toxicokinetic relationship across species. To specifically model acute accidental exposure, a breath-by-breath (BBB) PBPK model was applied; a modification assuming constant inhalation flow rates (CF) PBPK was used to address long-term occupational exposure. These PBPK models were used to evaluate concentrations of the test substance in arterial blood in adult female humans, pregnant rabbits and rats in response to inhalation at various concentrations and for various periods of time. The models simulated two scenarios of inhalation exposure for adult female humans: acute exposure to the test substance in a confined space, and repetitive inhalation exposure at the recommended Occupational Exposure Limit (OEL) in the workplace. The human model results were compared to equivalent scenarios using models developed for the two experimental species including pregnant rabbit and rat. The animal scenarios were run at experimentally derived NOELs (no observed effect level) for both species. In addition, a probabilistic approach known as Monte Carlo (MC) PBPK was used to generate a population based assessment of the model and reduces uncertainty in our overall assessment of the relationship between the experimental animal exposures and occupational exposure at the OEL. Comparisons between human and experimental animal were made based on total dose received in blood (Maximum Concentration or Cmax or Area Under the Time-course Curve or AUC).

Both with the acute and the occupational exposure simulation in an adult female human, the AUC for the different exposure lengths was approximately linear with length of exposure. Further, these simulations indicated rapid uptake and distribution of the test substance with the concentration falling to near zero before the start of a subsequent day's exposure.

The rabbit AUC for 6 hours (one study day) of exposure at the NOEL was 34.6 times higher than the human AUC for 6 hours of exposure and 48.2 times for the 2 week and 4 week exposure scenarios. The margins were reduced to 26.0 and 36.2 when comparing the 6 hour/day exposure in the rabbit to the 8 hour/day exposure in the human. For the rat, the lower NOEL combined with a 5 day per week study design resulted in lower margins of exposure with one day exposure margins of 15.7 when the 6 hour exposure was used for human and 11.8 when the 8 hour exposure scenario was used for the calculation. Adding variability to the human model had the expected effect of reducing the calculated safety margins given the 95th percentile would be greater than safety margins calculated with the mean model parameter values. However, these estimates include our uncertainty in model parameters as well as an assessment of population variability and thus, increase confidence in the predicted safety margins. With the conservative approach to the analysis, the safety margin was 36.2 for the rabbit (6 hour - 14 day exposure in rabbit as compared to an 8 hour - 14 day exposure in human). For the rat, the margin of safety was approximately 11.4 for the same exposure comparison.

The CF PBPK model can also be used to estimate the human equivalent concentration (HEC) or the concentration at which the human blood dosimetric matches that of the animal at the NOEL. This is accomplished by adjusting the exposure concentration in the human model until the dosimetric, Cmax and AUC in blood for 28 day exposure in this scenario, is equivalent to the animal at the NOEL. The animal is simulated as the study design. For the rabbit this was 6 hr per day and 7 days per week while the rat was exposed 6 hr per day and 5 days per week. The human scenario used for comparison was an 8 hr per day exposure over 5 days each week. Using the rabbit Cmax of 62.7 mg/L and AUC of 10,539 mg/L*hr, the HEC would be 23,560 or 32,150 ppm. When the rat dosimetrics were used (Cmax = 27.8 mg/L and AUC = 3,333 mg/L*hr) the HEC was 10,040 ppm based upon Cmax and 10,180 ppm when AUC was the basis. The higher exposure required in the human to reach the animal dosimetric is based on several factors. The most important factor is differences in the solubility in blood of 1233zd between the species as near steady state conditions are reached with each day of exposure and virtually the entire amount of absorbed compound is cleared before the subsequent day’s start of exposure. That is, the model reaches a stable periodic behavior within the two days of exposure. Other factors include differences in the study design compared to the normal human workday as well as differences in physiology.

Description of key information

Since it is likely that the substance will be absorbed by inhalation and in the absence of substance-specific absorption data for the oral and dermal route, the default absorption values from the REACH guidance (Chapter 8, R.8.4.2) are used for DNEL derivation, namely: 100 % for inhalation, 50 % for oral and 50 % for dermal absorption. Due to the rapid excretion of metabolites formed no bioaccumulation potential is to be expected.

Key value for chemical safety assessment

Bioaccumulation potential:

no bioaccumulation potential

Absorption rate - oral (%):

50

Absorption rate - dermal (%):

50

Absorption rate - inhalation (%):

100

Additional information

Introduction

The substance is a colorless, liquified gas with a boiling point of 19°C. The substance has a high vapour pressure of 1065 hPa, an octanol/water partition coefficient of 2.2 and the substance is highly soluble in water.

 

Absorption:

Oral and dermal: Substance-specific absorption data for the oral and dermal route are not available. However, the substance is a gas and therefore dermal and oral exposure is unlikely

 

Inhalation: The substance has a vapour pressure of 1065 hPa and therefore inhalation is the major route of exposure. Experiments in rats indicated absorption as systemic effects were observed in a 14-day inhalation study at 7500 and 20000 ppm and in a 90-day inhalation study at 10000 and 15000 ppm. Using physiologically based pharmacokinetic (PBPK) modeling techniques, occupational exposure simulation in an adult female human indicated rapid uptake and clearance from blood for the substance with the concentration falling to near zero before the start of a subsequent day's exposure. The blood/air (BA) partition coefficient also indicates absorption. The blood/air partition coefficient in the rat was at least twice as high as in humans. It can therefore be concluded that the substance is likely to be absorbed via the inhalation route.

 

Metabolism

Metabolism of the substance has been investigated in Sprague Dawley rats, New Zealand rabbits, and in human liver microsomes. S-(3,3,3-trifluoro-trans-propenyl)-glutathione was identified as predominant metabolite of the test substance in rat, rabbit and human liver microsomes in the presence of glutathione. Products of the oxidative biotransformation of the test substance were only minor metabolites when glutathione was present. In rats, both 3,3,3-trifluorolactic acid and N-acetyl-(3,3,3 -trifluoro-trans-propenyl)-L-cysteine were observed as major urinary metabolites. 3,3,3-trifluorolactic acid was not detected in the urine of rabbits exposed the test substance. Several minor metabolites formed by alternative reactions of S-(3,3,3-trifluoro-trans-propenyl)-glutathione were also excreted. Quantitation showed rapid excretion of both metabolites with urine t1/2 of less than 6 hours in both species. Based on metabolite recovery in urine and estimated doses of the test substance received by inhalation, the extent of biotransformation of the test substance was determined as approximately 0.01% of received dose in rabbits and approximately 0.002% in rats. The remaining material is likely exhaled due to its high volatility. The metabolite structures show that the test substance undergoes both oxidative biotransformation and glutathione conjugation at very low rates. The very low extent of biotransformation and the rapid excretion of metabolites formed are consistent with the very low potential for toxicity of the test substance in mammals.

 

Conclusion

The test substance is absorbed via the inhalation route. The extent of biotransformation is very low and metabolites formed are excreted rapidly. Untransformed test substance is likely exhaled due to the high volatility of the substance. As, moreover, the log P value of the substance is lower than 3, the substance is unlikely to accumulate with the repeated exposure patterns normally encountered in the workplace. In the absence of substance-specific absorption data for the oral and dermal route, the default absorption values from the REACH guidance (Chapter 8, R.8.4.2) are used for DNEL derivation, namely: 100% for inhalation, 50% for oral and 50% for dermal absorption