Page 9: Guidelines for Canadian Drinking Water Quality: Guideline Technical Document – Vinyl Chloride

Part II. Science and Technical Considerations (continued)

8.0 Kinetics and Metabolism

8.1 Absorption

8.1.1 Experimental Animals

In rats, vinyl chloride is readily absorbed via oral and inhalation routes of exposure and is rapidly distributed throughout the body. After gastric intubation of male Wistar rats with 10 mL of aqueous solutions containing vinyl chloride at concentrations of 2.26-2.82 mg/mL (total dose 22.6-28.2 mg/animal), peak blood concentrations (6 to > 40 µg/mL) were found in less than 10 minutes, and absorption was almost complete (Withey, 1976; Feron et al., 1981). One dose of 450 mg/kg body weight (bw) administered orally to rats resulted in 98.7% absorption from the gastrointestinal tract (Green and Hathway, 1977).

Dermal absorption of vinyl chloride gas is not likely to be significant, as only 0.031% and 0.023% of the total vinyl chloride doses (7000 ppm and 800 ppm, respectively) were absorbed by male Rhesus monkeys, and the majority of the dose was exhaled (Hefner et al., 1975a).

8.1.2 Humans

Inhalation of vinyl chloride at concentrations of 7.5, 15, 30 and 60 mg/m3 for six hours resulted in rapid pulmonary absorption and retention of 40% in five young, healthy adult males voluntarily exposed, independent of the concentration (Krajewski et al., 1980).

8.2 Metabolism

The metabolism of vinyl chloride has been quantitatively estimated in humans from gas uptake experiments whereby after initial absorption of vinyl chloride, continued absorption is mainly attributed to metabolism. When exposing young men to vinyl chloride at concentrations of 7.5, 15, 30 and 60 mg/m3 by gas mask for six hours, Krajewski et al. (1980) reported that the retention of vinyl chloride was found to be independent of the inhaled concentration and did not change with increasing vinyl chloride concentrations, suggesting that exposure to vinyl chloride up to 60 mg/m3 did not cause saturation of the major metabolic pathway.

The fate of vinyl chloride in the animals appears to be dose dependent (Watanabe and Gehring, 1976), which is primarily due to the linear kinetics of the metabolic pathway for vinyl chloride at low doses. The metabolism of vinyl chloride has been shown to follow Michaelis-Menten kinetics in rats, with enzyme saturation occurring following exposure to approximately 100 ppm in air or between 1 and 100 mg/kg bw/day for a single gavage dose (Hefner et al. 1975b; Watanabe et al. 1976a).

The vinyl chloride metabolic pathway occurs through mixed-function oxidase cytochrome P450 2E1 (CYP2E1) both in vitro, with cell isolates of the microsomal fraction (S9) from rats (Kappus et al., 1976; Guengerich and Shimada, 1991; el Ghissassi et al., 1998) and humans (Guengerich et al., 1991) and with the human B-lymphoblastoid cell line (Chiang et al., 1997), and in vivo, with rats (Watanabe et al., 1976a, 1976b) and Rhesus monkeys (Buchter et al., 1980). Liver toxicity (vacuolization of the centrilobular parenchyma) increases in parallel with P450 liver content in rats after vinyl chloride inhalation exposure to 50 000 ppm for 6 hours (Reynolds et al., 1975).

Inhibition of vinyl chloride metabolism in rat and human microsomes was observed when vinyl chloride exposure occurred in the presence of diethyldithiocarbamate, a CYP2E1 selective inhibitor (Guengerich et al., 1991; el Ghissassi et al., 1998). Pretreatment of Sprague-Dawley rats with a general mixed-function oxidase blocker (SKF-525A) impeded the metabolism of inhaled vinyl chloride at a concentration of 1000 ppm (Hefner et al., 1975b).

Enzymatic transformation of vinyl chloride for excretion results in the generation of polar intermediates (Antweiler, 1976; Kappus et al., 1976; Rannug et al., 1976). The two major metabolites of vinyl chloride in the liver are the highly reactive and short-lived chloroethylene oxide (CEO) and CAA (Whysner et al., 1996; el Ghissassi et al., 1998), a highly reactive α-halocarbonyl compound which results from a rapid rearrangement of CEO (Pessayre et al.; 1979; Whysner et al., 1996; el Ghissassi et al., 1998). These two metabolites are detoxified mainly through glutathione (GSH) conjugation (Leibman, 1977; Tarkowski et al., 1980; Jedrychowski et al., 1985) which is supported by the observation of decreased non-protein sulfhydryl concentrations when exposed to high concentrations of vinyl chloride (Tarkowski et al., 1980; Jedrychowski et al., 1985), and by the excretion of GSH-conjugated metabolites in the urine of rats following exposure to vinyl chloride (Hefner et al., 1975b; Watanabe et al., 1976c). CAA may also combine directly or enzymatically with GSH by way of glutathione transferase (GST) to form S-formylmethylglutathione. S-formylmethylglutathione can act directly with GSH-derived cysteine to form N-acetyl-S-(2-hydroxyethyl)cysteine which is another major urinary metabolite of vinyl chloride (Green and Hathway, 1975). The GSH conjugates are then hydrolysed, resulting in excretion of cysteine conjugates in the urine (Hefner et al., 1975b). The two major metabolites identified in the urine of rats following exposure to vinyl chloride are N-acetyl-S-(2-hydroxyethyl)cysteine and thiodiglycolic acid (Watanabe et al., 1976b).

8.3 Distribution and Excretion

Vinyl chloride is distributed rapidly throughout the body, but poorly retained due to rapid metabolism and elimination (OEHHA, 2000; ATSDR, 2006).

In rats exposed to 4.5-70 ppm 14C-labelled vinyl chloride for five hours, vinyl chloride was found to bind in the largest proportion in the liver, followed by the small intestine, kidney, lung and spleen; this organ distribution is supported by the lipophilicity of vinyl chloride, as indicated by its low octanol:water partition coefficient of 1.62 (U.S. EPA, 2000a). Monkeys acutely exposed by the skin to 14C-labelled vinyl chloride gas also had detectable 14C activity in the bile, liver and kidney, however, no appreciable amount of activity was detected in other tissues (Hefner et al., 1975a; Bolt et al., 1980).

In an experiment in which rats were given single oral doses (by gavage) of 0.05, 1.0, 20.0 or 100 mg/kg bw of 14C-labelled vinyl chloride dissolved in corn oil, the percentage of the dose exhaled within 72 hours as unchanged vinyl chloride was 1.4%, 2.1%, 41.4 and 66.6%, respectively. Excretion in urine was 68.3%, 59.3%, 22.6 and 10.8%, respectively. 14CO2 in expired air accounted for 9.0%, 13.3%, 4.8% and 2.5%, respectively. A proportion of 0.47-2.39% was unabsorbed and excreted unchanged in the faeces. The liver was found to retain the maximum percentage of activity at all dose levels, 3-5 times the percentage found in muscle, lungs or fat (Watanabe and Gehring, 1976).

Alderley Park rats (Wistar-derived strain) exposed to vinyl chloride at 1-450 mg/kg bw pulmonarily excreted unchanged vinyl chloride after 3-4 hours, whereas polar metabolites and labelled carbon dioxide continued to be excreted for 3 days (Green and Hathway, 1977).

When Sprague-Dawley rats were exposed to > 100 ppm (260 mg/m3) vinyl chloride in air for 5 hours, 69% of the absorbed dose was excreted as metabolites within 24 hours in urine via the kidney (Watanabe et al., 1976b; OEHHA, 2000). An additional 1.7% was found in the urine 24-48 hours later (Bolt et al., 1976). The half-life for urinary excretion in rats was about 4 hours. The blood concentration fell rapidly after removal of vinyl chloride from the air (Withey, 1976).

Intravenous injection of vinyl chloride (250 µg of 14C-labelled vinyl chloride per kg in N-(β-hydroxyethyl) lactamide) was reportedly almost entirely exhaled (99%) after 1 hour by rats (Green and Hathway, 1975).

Deoxyribonucleic acid (DNA) adducts were found in various tissues (testis, kidney, spleen, lung, liver, lymphocytes) after exposing rats to 500 ppm vinyl chloride in air, suggesting the migration of CEO via the bloodstream after liver metabolism (Guichard et al., 1996; Barbin, 1999). No ethenoguanine or 7-(oxoethyl)-guanine adducts were found in brains of rats exposed to 1100 ppm vinyl chloride for 4 weeks (Morinello et al., 2002a). However, in this same study, an increase of adducts in weanling rats was found (Morinello et al., 2002a).

Regardless of route of administration, as the dose of vinyl chloride increases, so does the proportion of the dose (unmetabolized) that is exhaled; however, the proportion of the dose that is excreted in the urine and faeces decreases with increasing dose (Green and Hathway, 1975; Hopkins, 1979).

8.4 PBPK Modeling

Several PBPK models have been developed for vinyl chloride. Clewell et al. (1995) developed a PBPK model to refine the vinyl chloride risk calculation for animal to human extrapolations based on pharmacokinetic information. To illustrate the pharmacokinetics of vinyl chloride, four compartments were necessary for the model: liver, fat, and highly and poorly vascularized tissues, with the liver being the site of metabolism and the target tissue in which angiosarcoma of the liver (ASL) arises (Clewell et al., 1995). Further biotransformation of CEO, resulting in carbon dioxide generation, glutathione depletion and binding of reactive metabolic products to macromolecules in the liver, was also included in the model. The risk estimate for vinyl chloride exposure was found to be consistent across species, routes and media of exposure, showing agreement between the results and risk evaluation methods.

Clewell et al. (2001, 2004) further refined their PBPK model to include the glutathione dynamic and age extrapolations, based on CYP2E1 enzyme maturation and body weight-clearance capacity; the refined model also takes into account each life stage with cumulative exposure to 1 µg/kg bw per day over a lifetime. Blood concentrations of the parent compound were found to be the highest very early in life (< 6 months old) and then decrease in early childhood. The plateau reached in adulthood can be explained by rapid maturation of CYP2E1 after 6 months of age and changes in body weight-clearance capacity (Clewell et al., 2004).

Other PBPK models for vinyl chloride include those by Gentry et al. (2003), Chiu and White (2006) and Yoon et al. (2007). The Gentry et al. (2003) model examines differences in blood and tissue dose metrics between a mother and foetus (human) during pregnancy and lactation. This model adds fetal liver, fetal blood and other fetal tissue compartments to the adult model to account for transfers through the placenta and breast milk, as well as differences between adult and fetal enzyme activity. The vinyl chloride concentrations in fetal blood were predicted to be similar to concentrations in the mother's blood during gestation, however, during lactation, blood concentrations in the neonate were predicted to be much lower. Concentrations of vinyl chloride and CEO in the liver were not predicted. Owing to the lack of a developed CYP2E1 enzyme system, Gentry et al. (2003) concluded that fetal exposure to reactive metabolites of vinyl chloride is expected to be negligible. However, there is some evidence that transplacental exposure can result in carcinogenicity in rat studies, and the half-life of CEO may be long enough for this reactive metabolite to cross the placenta after being generated in maternal tissues (Rice, 1981). The Chiu and White (2006) human PBPK model incorporates fewer parameters and assumes that all vinyl chloride metabolism occurs in the liver via first-order kinetics. This model was found to duplicate the Clewell et al. (2001) model for extrapolation between oral and inhalation exposure to vinyl chloride; however, as it assumes first-order kinetics for metabolism, it is appropriate for use only in the dose range in which metabolism is approximately linear. A model by Yoon et al. (2007), validated in Sprague-Dawley rats, considered the effect of extrahepatic CYP2E1 on PBPK modeling results. The results of the modeling indicated that extrahepatic CYP2E1 metabolism did not contribute significantly to the results considering only liver metabolism, indicating that PBPK models only considering liver metabolism are appropriate for the risk assessment of vinyl chloride.

Health Canada (2011) developed a PBPK model based on the one developed by Clewell et al. (2001, 2004) to facilitate animal to human and high dose to low dose extrapolations for vinyl chloride. The model was updated to include a dermal component so that it could be used to estimate the contribution of showering and bathing to vinyl chloride exposure. The model was validated using pharmacokinetic data from animals (Feron et al., 1981; Til et al., 1983) and humans (Buchter et al., 1978) exposed to vinyl chloride. Non-cancer effects (liver toxicity) and tumour progression (DNA adducts, mutations) have been shown to result from CEO generation; therefore, the appropriate dose metric with which to estimate cancer and non-cancer risks corresponds to the quantity of vinyl chloride metabolites produced divided by the liver volume.

Data from both Feron et al. (1981) and Til et al. (1991) were used for PBPK modeling by Health Canada. The external doses from these studies were inputted into the rat PBPK model to determine the cumulative lifetime internal dose of vinyl chloride metabolites generated per litre of liver for several liver tumour endpoints in both male and female rats. During the modeling, vinyl chloride oral uptake was designated as zero-order (independent of concentration) and spread out over a 24-hr period; this ensured that saturation of the metabolic pathways was avoided, thus generating a maximum value of the dose metric which is conservative with respect to what may actually occur during an oral dose. This approach, coupled with the use of the same liver metabolic processes for both inhalation and oral inputs, increases the confidence in dose metrics derived from oral inputs. It also addresses concerns of reduced confidence in dose metrics derived from oral studies compared to those derived from inhalation studies (as indicated experimental data). Using the multistage cancer model from the U.S. EPA's benchmark dose (BMD) software, BMDS (U.S. EPA, 2010), the internal concentrations of vinyl chloride metabolites associated with an excess lifetime cancer risk of 10−4, 10−5 and 10−5 in male and female rats were estimated for several cancer endpoints in order to determine the most appropriate point of departure (POD) for use in the human PBPK model. The cumulative lifetime internal dose of vinyl chloride metabolites associated with combined liver tumours (neoplastic nodules, hepatocellular carcinoma [HCC] and ASL) was determined as the most appropriate POD to use in the human PBPK model to estimate the external human doses associated with each risk level for a lifetime (70 years) exposure. The external human doses generated by the model represent the levels in drinking water that would be associated with an excess lifetime cancer risk in humans of 10−4, 10−5 and 10−5 , when daily exposure to drinking water occurs through ingestion (1.5 L-eq/day) as well as through inhalation (0.4 L-eq/day) and dermal (1.9 L-eq/day) exposures from a 30-minute bathing scenario.

The estimated L-eq contribution from dermal exposure likely represents an over estimate since the PBPK model accounts for exposure to both the liquid and gas phases of vinyl chloride during a bathing event. Based on the physical/chemical properties of vinyl chloride and human physiological characteristics, the PBPK model estimated an inhalation absorption of 33%; a dermal absorption of 94% was estimated for the liquid phase (that is water containing solubilised vinyl chloride) which represents the majority of the L-eqs for the dermal route and the gas phase was determined as contributing a negligible amount of L-eqs given its very low absorption by the skin. This is a conservative approach used in the absence of published data on the proportion of solubilised vinyl chloride that could be present during showering or bathing; Health Canada determined that this conservative approach is appropriate to estimate the L-eq contribution from dermal exposure.

Page details

Date modified: