Registration Dossier

Administrative data

Key value for chemical safety assessment

Additional information

There is a large database of genotoxicity studies for propylene oxide comprising 50 years of study. The available human information is limited but extensivein vitroandin vivotests in bacteria, mammalian cells, and animals have been conducted and are summarized by Albertini and Sweeney (2007)

Several reliable Ames tests are available, with conduct performed to or comparable with OECD guideline 471 or the method described by Haworth et al. 1983 (Shell Research Ltd., 1976; National Toxicology Program, 1980, 1985; Canter et al., 1985). In the key dataset (National Toxicology Program, 1980, 1985; Canter et al., 1985),S. typhimuriumstrains TA 97, TA100, TA1535 and TA1537 were all positive (or equivocal) with or without S9 for metabolic activation. The results were equivocal forS. typhimiriumTA98.In another study (Shell Research Ltd., 1976), propylene oxide was tested at 0-500 µg/ml with and without S9; results were positive forE. coliWP2 uvrA andE. coliWP2 and negative forS. typhimuriumTA 1538, the only strain tested. Positive results were obtained usingin vitrochromosome aberration tests (similar to OECD guideline 473, or related) using an epithelial cell line (RL1) and CHO cells (Shell Research Ltd., 1976; Gulati et al., 1989). Propylene oxide was also tested in anin vitromammalian cell gene mutation assay (McGregor et al., 1991). In this L5178Y mouse lymphoma assay (comparable to OECD guideline 476), cultured mouse lymphoma cells were exposed to propylene oxide vapours at concentrations of 0.04-1.25% (>1.25% was lethal to the cells). Propylene oxide demonstrated dose-related mutagenic activity at concentrations between 0.04% and 1.25%.

Propylene oxide was tested for genotoxic effectsin vivoin rats and mice. Rats were exposed by inhalation at 50, 100, 200, or 400 ppm, and chromosome aberrations (GLP and similar to OECD guideline 475) and induction of micronuclei (GLP, and similar to OECD guideline 474) were assessed. The results of both these assays were negative, with no statistically significant increases (Dow Chemical Company, 2009). The inhalation route is considered to be relevant to humans and is considered to be a physiologically relevant route of exposure. A study reported by Bootman et al. (1979) exposed mice to propylene oxide by oral gavage at (2 x each) 100, 250, or 500 mg/kg or via intraperitoneal injection at (2 x each) 75, 150, or 300 mg/kg. Positive control animals received either cyclophosphamide or chlorambucil, and vehicle controls received 0.5% gum tragacanth. Genotoxic effects were determined by a micronucleus assay (Bootman et al., 1979), and the results of the oral route—again, a physiological route--were negative, whereas the results of the intraperitoneal route were positive, although only at the highest doses. Micronucleated cells per 1,000 polychromatic erythrocytes were scored. No increase in the number of micronucleated cells was seen after oral gavage administration of up to 1000 mg/kg (2 x 500 mg/kg) propylene oxide. Indeed, 14-d repeated oral administration of up to 250 mg/kg was also negative for genotoxic effects in a dominant lethal assay.

However, there was a significant increase in micronucleated cells in mice receiving 600 mg/kg (as 2 x 300 mg/kg) by the intraperitoneal route (6.5/1,000 cells) compared to the vehicle controls (3/1,000 cells). For comparison, chlorambucil produced a much higher response (43/1,000 cells). Chronic (2-yr) inhalation exposure of monkeys to up to 300 ppm propylene oxide did not result in any increases in chromosomal aberrations in peripheral blood (Lynch et al.,1984b); again, a physiological route relevant to humans.

Two dominant lethal studies are reported for propylene oxide. In the mouse study, Bootman et al. (1979) administered 50 or 150 mg/kg bwt propylene oxide for 14 days and recorded implants, early and late deaths. Pregnancy rate, total implants/sire, and post-implantation loss did not differ significantly between controls and propylene oxide-treated groups. The positive control group (EMS) exhibited a significant reduction in pregnancy rate in weeks 2 and 5, and an increase in implant deaths after weeks 1 and 2 of mating. Hardin et al. (1983) evaluated pregnancy rate,corpora lutea, implantations, and early and deaths in rats that receivedrepeat (5-d) inhalation administration of 300 ppm 7 h/d propylene oxide.There was no difference found between controls and treated animals regarding any of the reproductive parameters to indicate a genotoxic effect of propylene oxide.

The genotoxic effects on germ cells of inhaled propylene oxide were determined using thesex-linked recessive lethal(SLRL) multiple-locus assay (which measures forward mutations in any locus on the X-chromosome) inDrosophila .melanogaster.When exposed to propylene oxide by inhalation males were mated withnucleotide excision repair (NER) negative- females, induction of mutations could be demonstrated for all 8 doses( 0- 48000 ppm), proving DNA adduct formation by Propylene oxide over the entire dose range. With NER+ females, however, significant frequencies of RL only occurred at the three highest doses. Both studies showed that propylene oxide does not induce high SLRL frequencies when mutagenized males are mated with wild-type females (Vogel and Nivard 1997, 1998).

Overall the data available on propylene oxide demonstrate that, althoughin vitrodata demonstrate genotoxic effects,in vivoonly non-physiological, irrelevant exposure routes (e.g., intraperitoneal injection) have resulted in genotoxic effects (micronucleus induction). Several significant datasets demonstrate no increases in genotoxic effects at high doses of propylene oxide administeredviaphysiological routes relevant to human exposure scenarios,viainhalation or oral routes. The positive result following intraperitoneal injection of high doses is considered to indicate the potential for mutagenicity at sites of initial contact in the body.

A review by Albertini and Sweeney (2007) considers the DNA–reactive genotoxic property of propylene oxide and is summarized below. By conventional definition, this includes all of the genotoxic events that include and follow its reactions with DNA, that is, the kinds and relative amounts of DNA adducts it produces. Propylene oxide’s extensive genotoxicity profile is then reviewed according to the test system employed, that is,in vitroorin vivoand, for the latter, the route of exposure-whether physiological or non-physiological.

A variety of studies have shown that propylene oxide reacts primarily with cyclic ring nitrogens in DNA via the Sy2 mechanism, thereby producing hydroxypropyl (HP) adducts primarily at N7G, N3 and N1A, and N3C and N3T (reviewed in Solomon, 1999; Koskinen and Plna, 2000). The N7G and N3A HP adducts neither distort DNA structure nor allow for cyclic or ring-open transformations. Therefore, they are non-pro-mutagenic lesions. Being unstable, these adducts are usually eliminated spontaneously, leaving behind the depurinated sites that can result in mutations only if not repaired before DNA replication occurs (Singer and Glunberger, 1983; Randell et al., 1987). The N1HPA adduct may be spontaneously converted to an N6HPA adduct by Dimroth rearrangement. Both are at coding sites and therefore are potentially pro-mutagenic. N3HPC can be rapidly converted to N3HP uracil (U) by hydrolytic deamination. Although both adducts, being at coding sites, are potentially pro-mutagenic, 3-HPU may be the more potent in this regard as there is no known pathway for its repair in DNA. The only adducts actually measured in tissues from animals exposed to propylene oxidein vivoinclude N7-HPG, N1-HPA/N6-HPA, and N3HPU (only in nasal tissues). Studies in rats have used the N7HPG DNA adducts as probes to define tissue amounts and distributions of PO DNA lesions following inhalation exposures (Rios-Blanco et al., 1997; Segerback et al., 1998; Plna et al,, 1999). By far the N7HPG adduct was most abundant (estimated to represent >95% of total adducts) and found in all tissues evaluated, including testes (although at much lower levels than systemic or site-of-contact tissues), followed by N1HPA and N6HPA together (~2.0% relative to N7HPG), and then N3HPU (~0.02% relative to N7HPG) identified only in the nasal respiratory tissues, where all adducts were most abundant. N3HPC adducts were not observed.

Primary lesions in the DNA, such as adducts or strand breaks, are early events in the mutation process and are considered to be non-mutational genotoxic events (shortened hereafter to non-mutations) (Albertini and Sweeney, 2007). Although adducts may be precursor events, progression to mutations is by no means inevitable; in fact, mutations are quite rare or often do not occur at all, depending on the reactive agent, the kind of interaction between the agent and DNA, the DNA repair capacity, and other factors. There are other kinds of non-mutations, such as sister-chromatid exchanges (SCE) or cell death, that, although reflecting genotoxic events, are neither part of the mutation process nor precursors to mutations. They too are defined here as non-mutations. Non-mutations may be tolerated by the cells without change or may be reversed; mutations, once formed, are irreversible. It follows, therefore that non-mutations, even though they are genotoxic endpoints, are not surrogates for the mutations that can initiate the carcinogenic process. Rather, they represent exposure dosimeters and may signify potential hazards. DNA-reactive genotoxic chemicals such as propylene oxide that are also rodent carcinogens are frequently assumed to induce cancer by this DNA-reactive mutagenic mode of action alone. Consideration of the following points raise questions as to whether this is a reasonable default assumption for PO in light of its genotoxicity profile, its pattern of tumor induction in rodents and its tumor-associated toxicities. Points of relevance, considered in detail in this review, may be summarized as follows:

  • Propylene oxide is a monofunctional alkylating agent that is a weak genotoxin. The most abundant DNA adduct formed is N7HPG, which is lost by spontaneous depurination, with the resulting AP sites being efficiently repaired. N7HPG is not itself a promutagenic DNA adduct. Other minor adducts, which may have pro-mutagenic potential are not formed in great numbers. Although propylene oxide has given positive responses in a multiplicity of genotoxicity tests, bothin vitroandin vivo, for both non-mutational and mutational endpoints, most studies have been at high exposure concentrations and/or have employed repair deficient organisms, such as in the Ames (as Salmonella is deficient in Nucleotide excision repair ). Noin vivogenotoxicity study in mammals, administering Propylene oxide by a physiological route, has been positive, including a 2-yr inhalation exposure of monkeys to 300 ppm Propylene oxide.

  • Propylene oxide induces tumors in rodents only at portals of entry, and only at relatively high exposure concentrations.

  • Propylene oxide produces cell death, cell proliferation/hyperplasia, and inflammation, also at the same sites of contact where tumor induction occurs. Careful investigations have shown these non-neoplastic cell changes to be threshold events. The neoplasms, seen in rodents in the long-term bioassays, appear to have arisen in the context of these non-neoplastic changes.

  • Propylene oxide is associated with a high degree of tissue GSH depletion that is most marked and sustained at sites of contact (portals of entry). This depletion is most marked in the anatomical regions of cell death, cell proliferation/hyperplasia, and inflammation.

  • As expected from its wide tissue distribution, Propylene oxide-induced N7HPG adducts are present in all tissues, but at lower concentrations in non-target tissues (including testicular tissue) than those found in nasal respiratory epithelium. As the presence of N7HPG is an indicator of other PO-induced DNA adducts, its presence in all tissues, including non-target ones, indicates that there is a threshold for carcinogenicity.

  • Extensive cell proliferation accompanying GSH depletion is likely to augment Propylene oxide's otherwise weak DNA-reactive genotoxicity by reducing DNA repair capacity.
  • Cell proliferation alone will have an effect of increasing the background mutations (but not the mutation rate) and the numbers of mutants in an affected tissue. To the extent that these mutants harbor cancer relevant mutations, the probability of evolution to malignancy increases.

  • Cell proliferation, with induced DNA adducts, will result in an increase total mutations over any elevations due to the adducts alone.

  • GSH depletion has the potential of producing genotoxic effects by mechanisms other than the direct DNA-reactivity of the reactive agent, likely mediated in part by reactive oxygen species (ROS) that are not adequately buffered due to the severe, sustained GSH depletion.

Based on these points, as documented in the descriptions given earlier, Albertini and Sweeney (2007) conclude that the mode of action for propylene oxide-induced cancer in rodents is not simply due to its DNA-reactive genotoxicity alone but is complex, and has a practical threshold.

Short description of key information:
Propylene oxide has been tested for genetic toxicity in a number of in vitro and in vivo assays. Propylene oxide is a monofunctional alkylating agent that is a weak genotoxin. The most abundant DNA adduct formed is N7HPG, which is lost by spontaneous depurination, with the resulting AP sites being efficiently repaired. N7HPG is not itself a pro-mutagenic DNA adduct. Other minor adducts, that may have promutagenic potential, are not formed in great numbers. Propylene oxide has given positive responses in a multiplicity of genotoxicity tests, both in vitro and in vivo, for both non-mutational and mutational endpoints, however, most studies have been at high exposure concentrations and/or have employed repair deficient organisms, e.g., in the Ames test. No in vivo genotoxicity study in rodents, administering propylene oxide by a physiological route, has been positive including a 2 year inhalation exposure study of monkeys to 300 ppm propylene oxide and a 90-day study in rats exposed to 400 ppm propylene oxide. The general toxicological profile for propylene oxide suggests that its potential to produce genetic damage might be expressed only at sites of initial contact. In relation to the potential of propylene oxide to induce heritable mutations in germ cells, dominant lethal tests involving inhalation exposure of rats and oral exposure of mice have given negative results. There is no evidence that propylene oxide causes heritable mutations in germ cells. Studies of DNA adduct formation indicate that N7-HPG DNA adducts, a marker of exposure not effect, were observed at very low levels (lower than systemic and site-of-contact tissues) in whole testis homogenate following repeated (4-wk) inhalation exposure to 500 ppm propylene oxide vapor. According to the requirements of classification, propylene oxide is classified as a germ cell mutagen but there is no evidence that it interacts with germ cell DNA.

Endpoint Conclusion: Adverse effect observed (positive)

Justification for classification or non-classification

According to EU Classification, Labelling and Packaging of Substances and Mixtures (CLP) Regulation (EC) No. 1272/2008 - 9th ATP 19 July 2016, the classification is H340, Cat. 1B.