Methyloxirane

Carcinogenicity

Description of key information

Propylene oxide is a rodent carcinogen selectively producing tumors only at sites of contact.  There are no traditional epidemiology studies available with cohorts exposed only to propylene oxide, without other potentially confounding chemicals; however, there is no evidence that low dose exposures constitute health risks to humans.  Inhalation studies in animals have shown that propylene oxide produces a spectrum of upper respiratory tract changes, from inflammation and degeneration to metaplasia and neoplasia. In mice the development of squamous cell carcinoma and adenocarcinoma as well as haemangioma and haemangiosarcoma in the nasal cavity occurred following exposure to 400 ppm for 2 years. In similarly exposed rats, there was evidence of papillary adenoma development in the nasal cavity. A similar study in a second strain of rats exposed to 300 ppm showed degenerative and hyperplastic changes of the nasal mucosal epithelium but no nasal tumors, and a non-statistically significant, slightly increased incidence of carcinoma at slightly more distal sites in the respiratory tract including the larynx, pharynx, trachea, and lung only at this highest exposure. Repeated oral administration by gavage in rats induced carcinoma in the epithelium of the forestomach. The mode of action for propylene oxide-induced rodent nasal tumors has been extensively investigated and the evidence supports a complex mode of action with a practical threshold. Direct genotoxicity may be necessary, but it requires augmentation by propylene oxide’s associated toxicities for it to be made manifest.  Cell proliferation and glutathione depletion (perhaps with resulting indirect genotoxicity), seem to be required.   These associated toxicities appear to be the rate-limiting steps in tumor induction.  Since these are also threshold effects, as clearly documented in the studies discussed above, the process of carcinogenesis overall will have a threshold.

Justification for classification or non-classification

According to EU Classification, Labelling and Packaging of Substances and Mixtures (CLP) Regulation (EC) No. 1272/2008 Annex VI - 9th ATP 19 July 2016, the classification is: H350, Cat. 1B.

Additional information

There are no reports of human cancer associated with propylene oxide exposure, however the available information is limited.

Propylene oxide has been evaluated in rodent inhalation and oral cancer bioassays and demonstrated to be a site of contact carcinogen in animals.

Inhalation Carcinogenicity

In the 2-year study (National Toxicology Program, 1985) groups of 50 of each species and sex were exposed whole-body to 0, 200, or 400 ppm (0, 474 and 948 mg/m³) for 6 hours/day, 5 days/week, for 24 months. Routine observations included clinical signs of toxicity, body weight, and macroscopic and microscopic pathology. In mice and rats exposed to 400 ppm, mean body weight gain was reduced, compared to controls, and other treatment-related effects were noted as well, as discussed below.

At the end of the exposure period for rats, mortality rates were similar (60 - 70%) between exposed and non-exposed animals. Slightly decreased bodyweights (<9%) were recorded from week 20 in males, and week 40 in females. Increases were observed in the incidences of suppurative inflammation of the mucosa, epithelial hyperplasia, squamous cell metaplasia of the respiratory epithelium of the nasal turbinates, and papillary adenomas involving the respiratory epithelium and underlying submucosal glands of the nasal turbinates.

In mice propylene oxide caused increased incidence and severity of inflammation of the respiratory epithelium of the nasal turbinates. Squamous cell metaplasia in the nasal cavity was observed in one low-dose male and two high-dose females. One squamous cell carcinoma and one papilloma occurred in the nasal cavity of two high-dose males, and two high-dose females (4.0%) had adenocarcinoma of the nasal cavity. None of these tumors was observed in controls or low-dose animals in this study, and furthermore had not been observed in historical controls. Vascular neoplasms in the anterior nasal cavities, haemangiomas and haemangiosarcomas, were observed in mice receiving 400 ppm propylene oxide. In the high-dose group, haemangiomas developed in 5/50 males and 3/50 females, and haemangiosarcomas developed in 5/50 males and 2/50 females.

The conclusion from this study is that there were carcinogenic responses associated with exposure to propylene oxide in both rats and mice. In rats there was an increased incidence of papillary adenomas of the nasal epithelium in rats exposed to 400 ppm but not at 200 ppm; similarly in mice there was an increased incidence of hemangiomas and hemangiosarcomas in the anterior nasal cavity in mice exposed to 400 ppm but not at 200 ppm.

In the carcinogenicity study reported by TNO (TNO, 1984; Kuper et al., 1988), groups of 100 male and 100 female rats were exposed, whole-body, to propylene oxide vapour at 0, 30, 100, or 300 ppm (0, 71, 237, or 711 mg/m³) for 6 hours/day, 5 days/week, for 123-124 weeks. No adverse effects of treatment were observed on behaviour, food consumption, serum biochemistry, urinalysis, or haematology, compared to controls. Mortality was increased in both sexes at 300 ppm (55/100 males and 55/100 females) compared to controls (32/100 males and 30/100 females) and a similar tendency appeared for females at 100 ppm (43/100) contributed by the occurrence of mammary gland tumors. Most of these mammary tumors were benign, although the incidence of malignant mammary tumors was higher at 300 ppm. However these incidences were within the range of historical controls and were determined to be not treatment-related by the study authors.

Increased incidence of degenerative changes (slight to moderate “nest”- like infolds) of the nasal mucosa was observed in all exposed groups. Hyperplasia of the nasal epithelium was observed at 300 ppm. There were no tumors in the nasal turbinates, though there was a dose-related increase in focal hyperplasia. Overall, the results of this study provide evidence for the ability of propylene oxide to produce site-of-contact nasal effects in rats exposed to propylene oxide.

The third study evaluated the chronic inhalation toxicity and carcinogenicity of propylene oxide [and ethylene oxide] in a 2-year study conducted by Lynch et al. (1984a). Groups of 80 male F344 rats were exposed whole-body to 0, 100, or 300 ppm propylene oxide for 7 hours/day, 5 days/week, for 104 weeks. There was a dose-related increase in the incidence of complex epithelial hyperplasia in the nasal passages, and two adenomas occurred in the nasal passages in animals receiving 300 ppm. However, all rats in this study were affected byMycoplasma pulmonisinfection from about 16 months through termination, and this intercurrent disease, alone or in combination with exposure to propylene oxide, undoubtedly affected survival of the rats and influenced the development of proliferative lesions in the nasal mucosa. In addition, no tumor incidence was statistically significantly increased in this study.

All three inhalation cancer studies demonstrate NOAECs for nasal tumor formation as follows: 100 ppm (Lynchet al., 1984a); 200 ppm (NTP, 1985), and a NOEC of 300 ppm (TNO, 1984; Kuper et al., 1988), although the overall NOEC for Kuper et al. (1988) is 100 ppm.. These results fit very well with the threshold mode of action for induction of nasal tumors, and the key events approach, with identification of severe, sustained GSH depletion in target nasal tissue leading to induction of cell proliferation in the target nasal tissue as the initiating key events. 

As indicated in the repeated dose section, the DNEL is derived from the OEL and mode of action; with the local carcinogenicity within the nasal epithelium as a primary target.

Oral carcinogenicity

Dunkelberg (1982) administered 0, 15 or 60 mg/kgbwpropylene oxide by gavage, twice weekly, to groups of 50 female Sprague-Dawley rats for 150 weeks. Survival of treated rats was comparable to controls. Treatment with propylene oxide resulted in a dose-related increased incidence of epithelial hyperplasia, papilloma, and squamous cell carcinoma of the forestomach. In the groups receiving 0, 15 and 60 mg/kg bwthe combined incidence of hyperkeratosis, hyperplasia, and papilloma was 0/50, 7/50, and 17/50 respectively, and of squamous cell carcinoma was 0/100, 2/50, and 19/50. No increase in the incidence of tumors at other sites was observed. A positive control group, receiving 30 mg/kgbeta-propiolactone had a high incidence of forestomach tumors (46/50). The conclusion from this study is that propylene oxide can produce neoplasms at the site of application.

Mode of action of propylene oxide induction of chronic toxicity/carcinogenicity

Propylene oxide has been evaluated in three chronic inhalation studies in rats, one of which also included mice (National Toxicology Program, 1985; Lynch et al., 1984a; TNO, 1984; Kuper et al., 1988), and has been shown to be a site-of-contact carcinogen in rodents. The most notable effects (cancer and non-cancer) were found in the nasal cavities, the portal of entry for inhaled propylene oxide in rodents. Although rat and mouse demonstrated nasal tumors of different kinds, histopathological examination demonstrated that tumor formation was preceded and accompanied by severe inflammatory and hyperplastic changes in the nasal cavity of both species. In mice exposed by inhalation, propylene oxide produced haemangiomas and haemangiosarcomas of the nasal cavity and a few malignant nasal epithelial tumors, while in rats the tumors were papillomas of the nasal respiratory epithelium. Thus nasal tumors in both species were similar in location, observed in anterior regions of nasal respiratory epithelium.

The mode of action (MOA) is complex and characterized by a practical threshold, as described in Sweeney et al.,2009. The threshold corresponds with the requirement for a propylene oxide-induced severe, sustained depletion of GSH in target nasal respiratory epithelium, which then induces cell proliferation in the target nasal respiratory epithelium (Khan et al., 2009; Dow Chemical Company, 2009; Harkema ,2006). This is accompanied by the additional genotoxic (DNA reactive and /or non-DNA reactive) and non-genotoxic effects of propylene oxide, such as an increase in reactive oxygen species and decreased time for DNA repair due to increased cell proliferation.

Rats exposed to 50, 100, 200, or 300 ppm by inhalation (6 h/day for 3 days) had statistically significant decreases in nasal GSH levels (24-28% remaining) for exposures>100 ppm propylene oxide; similar levels of GSH depletion in nasal respiratory epithelium were found following treatments with DEM (500 + 150 mg/kg) or BSO (500 mg/kg) (Khan et al., 2009). The treatments that resulted in these severe, sustained reductions in target nasal mucosal GSH (>60% depletion compared with controls), whether due to propylene oxide or DEM or BSO, all triggered induction of cell proliferation in the nasal respiratory epithelium. A study conducted with mice and rats exposed to 50, 200, or 400 ppm propylene oxide (6 h/day; 5 d/wk; 4 wks)viawhole body inhalation also demonstrated a threshold for induction of cell proliferation in nasal respiratory epithelium in both species; only exposures to 400 ppm propylene oxide increased the cell proliferation in mice, while >200 ppm propylene oxide exposures resulted in an increase in target nasal respiratory epithelium cell proliferation in rats (Dow Chemical Company, 2009; Harkema, 2006).

GSH depletion is generally considered an adaptive/protective response to injury from toxicant exposure; typically it is not considered biologically significant until depletion is>60% of control values. Based on the observed depletion of GSH in target cells of nasal epithelium reported for rats at>200 ppm propylene oxide and the similar depletion from either DEM or BSO treatment, combined with the subsequent induction of cell proliferation in the target tissue of nasal respiratory epithelium, it is clear that there is a threshold for induction of the target nasal cell proliferation effects. This threshold underpins the mode of action for induction of nasal tumors in rodents by propylene oxide.

Significant GSH depletion is thought to leave cells vulnerable to oxidative damage that can produce cytotoxicity and indirect genotoxic effects by reduction of the inherent buffering capacity of cells, likely resulting in an increased formation of reactive oxygen species (ROS)-related pro-mutagenic adducts (e. g., 8-oxodeoxvguanosine or 8-oxo-dG). This is important, because GSH is key in the detoxification of propylene oxide. In all three chronic inhalation bioassays conducted on propylene oxide (TNO, 1984; Kuper et al., 1988; National Toxicology Program, 1985; Lynch et al., 1984a) focal hyperplasia of the nasal turbinates and degenerative changes and proliferative hyperplasia of the nasal epithelium were noted, particularly at the highest concentrations tested (300 and 400 ppm propylene oxide). At 30 ppm these responses were not present at 24-months, and only rated “slight”, of low incidence, and only identified in the 28-month treatment group (considerably longer duration than a standard bioassay) (TNO, 1984; Kuper et al., 1988). They were not identified in the 13-wk exposures to rats or mice, even at 500 ppm propylene oxide (NTP, 1985). The TNO study therefore points to a NOAEC of 30 ppm for chronic nasal effects in rats.

There is an extensive genotoxicity database on propylene oxide, includingin vitroandin vivostudies. This database is useful to inform the potential role of genotoxicity in the threshold mode of action of propylene oxide-induced carcinogenesis; thein vivogenotoxicity studies using a relevant exposure route are most appropriate. As reviewed in the publication by Albertini and Sweeney (2007), although propylene oxide reacts directly with DNA and is positive in everyin vitrotest reported, mutational events were not induced using relevant physiological routes of exposure,e.g., inhalation or oral. In a 4-week repeated dose inhalation study in rats (6 hours/day; 5 d/wk) up to 400 ppm propylene oxide, no increases in chromosomal aberrations or micronuclei in peripheral blood were found (Dow Chemical Company, 2009).  Chronic (2-yr) inhalation exposures to monkeys did not result in any increase in chromosomal aberrations in peripheral blood (Lynch et al., 1984b).

DNA adduct data serve as a sensitive dosimeter for exposure. For rats, a linear dose-response was observed for N7-HPG adducts; previous work had determined N7-HPG adducts were detectable in target nasal tissue following exposure to 5 ppm propylene oxide for 3 days or 4 weeks (Rios-Blanco et al., 2000). DNA adduct data were collected in the recentin vivostudy in rats and mice. The predominant DNA adduct formed from propylene oxide is the N7-hydroxpropylguanine (N7-HPG), estimated to account for up to 95% of all DNA adducts from propylene oxide. However, N7alkylguanine DNA adducts are not located at coding sites and are not considered mutagenic (Albertini and Sweeney, 2007; Boysen et al., 2009). In fact, Albertini and Sweeney describe all DNA adducts and apurinic (AP) sites as non-mutational genotoxic effects. This is because an adduct is not equivalent to a mutation; there are many steps required between formation of a DNA adduct and expression of an heritable mutation. Indeed, N7-HPG adducts are unstable and are eliminated by spontaneous depurination, which can result in AP sites that are mutagenic if replicated before repaired. These AP sites however are repaired quite efficiently. In fact there were no significant differences in AP site levels measured in tissues from mice and rats exposed for 4 weeks to 400 ppm propylene oxide, compared with tissues from unexposed control animals (Dow Chemical Company, 2009); the tissues examined were nasal respiratory epithelium, liver, spleen, and lung. The fact that AP sites were not increased with increasing exposure demonstrates that AP sites resulting from such chemical depurination do not lead to unbalanced DNA repair. The dose-response relationships for the predominant DNA adduct do not match the tumor dose-response, where significant chronic exposures (100 or 200 ppm propylene oxide) did not result in any increased incidence of tumors. This lack of dose-response concordance does not support a key role in the threshold mode of action for the formation of DNA adducts. Indeed, the threshold demonstrated for induction of GSH depletion, which appears to be the trigger for induction of cell proliferation in the target tissue, provides a point of departure for assessment of a safe exposure.

The experimental work reported by Khan et al. (2009) demonstrated that a 3-d inhalation exposure to>200 ppm propylene oxide would induce cell proliferation in the tumor target region of the nasal respiratory epithelium in male F344 rats, as demonstrated with BrdU incorporation. Their work also demonstrated that this exposure resulted in significant (>60%) depletion of GSH in nasal respiratory epithelium. When a similar significant GSH depletion was induced in rat nasal epithelium by treatment with either DEM or BSO, there was a similar induction of cell proliferation in nasal target tissue. These data provide critical supporting information to demonstrate the key role of GSH depletion as a key event in the threshold mode of action leading to propylene oxide-induced nasal tumors in rodents.

The primary aspect to be considered in deriving an OEL for propylene oxide is its local carcinogenicity with the nasal epithelium as primary target, which is well established experimentally in rats and mice. There is no evidence of carcinogenicity of propylene oxide from studies in humans. Indeed, the threshold demonstrated for induction of GSH depletion, which appears to be the trigger for induction of cell proliferation in the target tissue, provides a point of departure for assessment of a safe exposure. If there is no severe, sustained GSH depletion, and thus no induction of cell proliferation, then the risk will be negligible.Given the well-established mode of action for the local carcinogenicity that demonstrates a practical threshold (Sweeney et al.,2009), a DNEL rather than a DMEL is appropriate for propylene oxide.