Registration Dossier

Diss Factsheets

Toxicological information

Carcinogenicity

Currently viewing:

Administrative data

Description of key information

In summary, the recent data and reviews have strengthened the proposed mode of action in respect to the carcinogenic activity of styrene indicating that humans are not comparable to mice and should even be less susceptible than the non-responding rat. Specifically the data supported that 
- rats are resistant to styrene induced lung toxicity
- side chain oxidized metabolites (apart from styrene oxide) do not contribute to lung toxicity
- styrene and its toxic metabolites lead to glutathione depletion in the lung of mice and much less in rats
- styrene and its toxic metabolites may lead to oxidative stress in the lung (Clara cells) although this is less pronounced than glutathione depletion
- styrene and its toxic metabolites lead to Clara cell toxicity as demonstrated by changes in CC10 concentrations on the mRNA and protein level
- CYP2F2 is primarily responsible for metabolic activation of styrene in the lung of mice while the human analog (CYP2F1) has at most a very low activity and CYP2E1 does not play an important role
The mode of action proposed not only relates to styrene but is also operative for other lung carcinogens acting specifically in mice
The most important detoxification pathway of styrene oxide via epoxide hydrolase is highly efficient in humans.

Key value for chemical safety assessment

Justification for classification or non-classification

Overall, there is no convincing evidence that styrene possesses significant carcinogenic potential in humans.

Therefore, a classification for carcinogenicity according to EU-criteria (67/548/EEC) and to GHS-criteria (1272/2008/EC) is not warranted for styrene.

Additional information

In the UK RAR (June 2008) the data on carcinogenicity were summarized as follows (the references are given in brackets):

Studies in humans

In relation to human studies, several cohort and case-control studies covering workers exposed to styrene are available. In large, well-conducted studies, cancer mortality was investigated in the GRP industry with relatively high exposure to styrene and no significant exposures to other chemicals (Coggon et al., 1987; Kogevinas et al., 1993; 1994; Wong et al., 1994; Ruder et al., 2004). In these studies, and in studies in styrene production workers (Ott et al., 1980; Bond et al., 1992; Anttila et al., 1998), there was no clear and consistent evidence for a causal link between specific cancer mortality and exposure to styrene. The increased risks for lymphatic and haematopoietic neoplasms observed in some of these studies (Kogevinas et al., 1993; 1994; Kolstad et al., 1994) are generally small, statistically unstable and often based on subgroup analyses. These findings are not very robust and the possibility that the observations are the results of chance, bias or confounding by other occupational exposures cannot be ruled out. In the styrene-butadiene rubber industry, several studies have pointed to an increased risk of cancer of the lymphatic and haematopoietic systems (McMichael et al., 1976; Meinhardt et al., 1982; Matanoski et al., 1990; 1993; Macaluso et al., 1996; Delzell et al., 1996; 2001). However, detailed analysis of these data, together with the general toxicological picture for styrene and butadiene (see butadiene EU RAR), suggests that where increases are due to occupational exposure, it is butadiene, not styrene, that is the more likely causative agent (Santos-Burgoa et al., 1992; Matanoski et al., 1993; Macaluso et al., 1996; Delzell et al., 2001). In conclusion, based on human studies, there is no clear and consistent evidence for a causal link between specific cancer mortality and exposure to styrene.

A literature search from 1998 to January 2010 overlapping the date of the UK RAR (June 2008) was conducted. Overall, this did not lead to a modification of the conclusions reached by the UK RAR and reinforced some of the aspects elaborated above.

Scelo et al. (2004) investigated the lung cancer risk by occupational exposure to vinyl chloride, acylonitrile, and styrene in a large case control study (2861 cases, 3118 controls) with full adjustment for smoking. The subjects were recruited from seven European countries. For ever exposure to vinyl chloride the odds ratio was 1.05 (CI 0.68-1.62) and for ever exposure to acrylonitrile (39 cases, 20 controls) the odds ratio of 2.20 was significantly increased (CI 1.11-4.36). No association between exposure to styrene (51 cases, 47 controls) and lung cancer risk was found: odds ratio 0.70 (CI 0.42-1.18). 20 exposed and 16 controls were assigned to the highest exposure group (>12.5 ppm x years).

In another case control study 710 persons with malignant lymphomas were compared to 710 matched controls (Seidler et al., 2007). There was a significant association between exposure to chlorinated hydrocarbons and the risk for malignant lymphomas. For styrene (161 exposed, 169 controls) the results did not support the hypothesis of an association or a dose-response relationship between styrene and lymphoma risk. The highest exposure group (>67.1 ppm x years) comprised 12 exposed and 17 control persons.

A report of Delzell et al. (2006) to the Health Effects Institute (HEI) and Graff et al. (2005) supported the assumption that the increased incidence of lymphohematopoetic cancers in synthetic rubber industry workers has a strong association with butadiene but not with styrene.

 

In a follow up of previous investigations Sathiakumar et al. (2005) evaluated the mortality experience of 17924 men employed in thesynthetic rubber industry. There was a 16% leukemia increase (71 observed/61 expected) that was concentrated in hourly paid subjects with 20-29 years since hire and 10 or more years of employment (SMR 258; 156-403) and in workers employed in polymerization (SMR 204; 121-322), maintenance labor (SMR 326; 178-456), and laboratory operations (SMR 326; 178-546). No clear association was found between employment in this industry and other forms of hematopoetic cancers. More than expected deaths for colorectal and prostate cancers did not appear to be related to exposure in rubber industry. The authors point to the uncertainty about the specific agents responsible for the observed excesses.

Sathiakumar et al. (2009) examined lung cancer risk in male and female rubber industry workers in relation to butadiene and styrene exposure. Quantitative exposure estimated were used and a total of 104 female and 551 male decedents with lung cancer as underlying or contributing cause of death were identified out of a total population of 4101 female and 15958 male workers. No exposure response trends for styrene were seen among women or men. Another investigation was carried out for cancer mortality among 4863 female rubber industry workers (Sathiakumar and Delzell, 2009). The observed number of deaths was approximately equal to that expected for leukemia, Hodgkin lymphoma, multiple myeloma, non-Hodgkin lymphoma, and cancers of the breast and ovaries. More than expected deaths were found for lung and bladder cancers. Exposure response analysis, done only for lung cancer, indicated no trend for butadiene or styrene. The authors concluded that the observed excess for lung and bladder cancers may be attributable to non-occupational factors rather than to workplace exposure.

Boffetta et al. (2009) reviewed systematically studies of workers exposed to styrene in the manufacture and polymerization of the chemical, in the reinforced plastics industry and in styrene-butadiene rubber production. In addition, studies of workers monitored for styrene exposure were reviewed as well as studies of environmental exposure to styrene, community-based case-control studies of lymphoma and leukemia, and studies of DNA adducts. No consistently increased risk of any form of cancer was found among workers exposed to styrene. Thus the available epidemiologic evidence did not support a causal relationship between styrene exposure and any type of human cancer.

In the literature update until Oct 01, 2015, the following new information was found:

Coggon et al. (2015) extended the former study of 1987 (see above) by follow-up from 1946 through Dec. 2012 involving 7970 workers at 8 companies from England manufacturing glass-reinforced plastics. Mortality was compared with that for England and Wales. 3121 workers had died since the last follow-up. In addition, a nested case-controll study compared 122 cases of lymphohematopoetic cancers with 1138 controls. No elevation from lymphohematopoetic cancer was observed for the full cohort (SMR 0.90; 95% CI 0.69-1.15) or in those with more than background exposure to styrene (SMR 0.82; 95% CI 0.58-1.14) and no association was found in the case-control part. For workers with estimated high exposures (40-100 pp, 8h TWA, for >1year) the OR was 0.54 (95% CI 0.23-1.27). Mortality from lung cancer was significantly increased (SMR 1.44; 95% CI 1.10-1.86). To the cconclusion of the authotrs, this latter finding, as not being supported by other epidemiology studies, may have been confounded by smoking what would be worth further checking.

In summary, this study strengthens the decision in the EU for non-classification of styrene as a human carcinogen.

In a literature update until April 20th2017, the following new information was found:

The study by Christensen et al. (2017) is a major expansion and update of an investigation of cancer incidence among workers exposed to styrene in the Danish reinforced plastics industry by Kolstad et al. (Occup Environ Med 52:320-7 (1995) and Scand J Work Environ Health 20:272-8 (1994)). These earlier studies included workers employed between 1964 and 1988 at 386 companies. Christensen et al. expanded the cohort to include workers employed between 1964 and 2007 at 443 companies and extended followup to identify incident cancer cases diagnosed from 1968 through the end of 2012. Christensen et al. analyzed the incidence of cancers in the overall cohort and in relation to three surrogates of styrene exposure, including duration of employment at the study companies, year of first employment and probability of exposure to styrene at the companies where workers were employed and evaluated potential confounding by tobacco smoking. The estimates of styrene exposure probability came from a worker survey conducted in 2013 at the study companies. Christensen et al. considered an association between styrene exposure and the presence of a tumor to be “demonstrated” if the association with a particular form of cancer showed “consistent trends of increased risk with increasing duration of employment, early year of first employment, and exposure probability”. The study has several notable strengths, including large size, long follow-up and use of cancer incidence rather than cancer mortality as the endpoint of interest. The main conclusion of Christensen et al. was that the a priori specified criterion for an association of cancer with styrene exposure, i.e. increasing occurrence of cancer at a given site with duration of employment, early year of first employment and exposure probability, was not fulfilled for any cancer site. They noted that Hodgkin lymphoma, myeloid leukemia and cancer of the nasal cavities and sinuses were associated with some, but not all, of the three proxies of styrene exposure, and they did not interpret the observed associations as causal. Christensen et al. also noted that the excess of lung cancer in their cohort was not associated with proxies of styrene exposure and that their results did not confirm previous reports of associations between styrene exposure and cancers of the esophagus, pancreas, kidney and bladder or “other” lymphohematopoietic malignancies. They finally concluded that “Further studies are needed before firm conclusions about the human carcinogenicity of styrene can be made.” This study’s contribution to the epidemiological literature on styrene and cancer in humans is substantial but will not lead to change in classifications of styrene with regard to human carcinogenicity.

On July 10th 2017, another new publication was found: Ruder and Bertke (2017) analyzed cancer incidences on NIOSH boat-builders cohort exposed to styrene to test the hypothesis that leukemia and lymphoma might be associated with exposure to styrene. However, no notable excess of lymphatic and hematopoietic cancers was found. Evidence was found for a link between styrene exposure and overall cancer incidence, but no associations with tumor types observed in other works were found. The study has several limitations, inlcuding the imprecise exposure metric and the lack of data regarding confounding exposures. Future follow-up studies may provide more insight.

The German MAK commission(MAK, 2003) estimated the carcinogenic risk of styrene to humans on the basis of the internal exposure to styrene oxide or the levels of its hemoglobin or DNA adducts. The internal exposure levels were based on the toxicokinetics of styrene and styrene oxide in mice, rats, and humans. For a 40-year workplace exposure to 20 ppm styrene a cancer risk of 1.7-7.5 per 100000 exposed persons was calculated. This risk was smaller than the unavoidable risk of 1 per 10000 persons which has been estimated for the endogenously formed ethylene oxide. Therefore, styrene was assigned to the MAK category 5 for carcinogens: “Substances with carcinogenic and genotoxic effects, the potency of which is considered to be so low, that provided the MAK and BAT values are observed, no significant contribution to human cancer risk is to be expected”.

Filser et al. (2002) calculated the excess human lifetime risk that could result for lung and other possible systemic tumors from the daily intake of styrene via food and ambient air. Their considerations were based on the body and tissue burden of styrene oxide as calculated by the models of Csanady et al. (1994; 2003). Lung tumor incidences were taken from the inhalation bioassays. For other possible systemic tumor sites for which no exposure related increases had become apparent, the statistical uncertainties of the negative animal studies were taken into consideration to derive a hypothetical upper bound for the tumor incidences. For a daily oral intake of 12 µg styrene, the excess human lifetime risk for lung tumors was obtained to be between 5×10−9 and 2×10−8 and the highest possible risk for other systemic tumors to be between 7×10−9 and 2×10−8. Lifetime risks calculated for continuous exposure to 3 µg/m³ styrene in ambient air were between 8×10−7 and 3×10−6 (excess lung tumor risk) and between 2×10−8 and 4×10−8 (highest possible risk for other tumors).

 

In summary, the most recent data do not give any indication for an increased cancer risk due to styrene exposure, including lymphohematopoetic cancers and lung tumors.

A new literature search up to March 2013 was carried out. The following papers were identified; the new data did not change the former conclusions.

Cocco et al. (2010) investigated the incidence of B-cell non-Hodgkin's Iymphoma (B-NHL) and its major subtypes, as well as Hodgkin's Iymphoma and T-cell lymphoma in 2348 cases and 2462 controls. Exposure was assessed by industrial hygienists and occupational experts. For styrene the risk of follicular Iymphoma was significantly increased as was a trend test with three independent metrics of exposure. As regards follicular lymphoma the authors point to the small number of exposed subjects that did not allow a more detailed analysis. Overall they conclude that this large dataset confirms a role of occupational exposure to solvents in the etiology of B-NHL and chronic lymphocytic leukemia. Benzene was most likely to be implicated but a role for other solvents in relation to other Iymphoma subtypes such as follicular Iymphoma could not be excluded. In summary, in this case control study some indication was obtained for an increased risk for follicular lymphoma and exposure to styrene. But the small number of exposed subjects did not allow a detailed analysis.

Budroni et al. (2010) studied a cohort of 5350 male petrochemical workers with a follow-up period from 1990 up to 2006. Cohort members were employed for six months or more and overall, a total of 81,392 person-years at risk were accumulated. The standardized incidence ratios were calculated for the total cohort and for sub-cohorts. No significant increase in cancer risk was observed among workers potentially exposed to acrylonitrile, butadiene, or styrene. There was an increase in risk for all cancers, and particularly non-Hodgkin lymphoma, apparently concentrated among workers potentially exposed to vinyl chloride. In summary, in this cohort study of petrochemical workers there was no evidence for an increased cancer risk among workers potentially exposed to styrene.

Karami et al. (2011) investigated in a case control study whether occupational exposure to certain plastic monomers increased renal cell carcinoma (RCC) risk. For styrene there was no association between RCC and having ever been exposed or duration and average exposure. But a positive association was found with cumulative exposure (p-trend=0.02). In the discussion the authors refer to the studies of Ruder et al. (2004) and Wong et al. (1994) who also observed elevated kidney cancer mortality in styrene exposed workers. On the other hand it is pointed out that no such an association was found for occupations likely to entail styrene exposure (Sathiakumar et al., 2005, 2009; Gerin et al., 1998). The authors conclude that the results indicate a possible association between occupational styrene exposure and RCC risk. In this study occupational styrene exposure was observed primarily among styrene-manufacturing operators, tank cleaners and tank operators of copolymers manufacturers, auto body repairmen who utilized polyester resins, and plastic boat manufacturers who processed unsaturated polyesters. Additional studies are required to replicate the results in other populations. In summary, in this study some indications were obtained for an association between styrene exposure and renal cell carcinoma. But a clear conclusion cannot be drawn when taking all epidemiological evidence together.

Collins et al. (2013) investigated cancer mortality in an update of a large study of reinforced plastic industry workers with relatively high exposures to styrene. The study included 15826 workers exposed between 1948 and 1977 with a follow-up from 1948 to 2008. Mortality rates were associated with cumulative exposure, duration of exposure, peak exposures, average exposure, and time since first exposure. For cancer types associated with styrene exposure in former studies the following standardized mortality rates were found:

-      Lymphatic and hematopoietic cancers combined: 0.84 (0.69-1.02)

-      Non-Hodgkin lymphoma: 0.72 (0.50-1.00)

-      Leukemia: 0.84 (0.60-1.14)

-      Pancreatic cancer: 0.96 (0.73-1.22)

-      Lung cancer: 1.34 (1.23-1.46)

More lung cancers were observed than expected. But there was no increase in risk with increasing cumulative, peak, duration, or average exposure or with increasing latency. In fact there was an inverse association with cumulative and duration of exposure indicating that the excess lung cancer deaths occurred mostly among short term workers. By taking into account bronchitis, emphysema, and asthma rates to adjust for confounding by smoking, the authors concluded that the increased lung cancer risk was most likely related to smoking. An unexpected finding was an association of cumulative and peak exposure with kidney cancer. According to the authors this finding is likely due to chance because the exposure-response was weak, no increased risk was observed in other styrene related studies and there was a potential confounding with smoking. Overall, there was no coherent evidence that styrene exposure increases the risk of cancer in exposed workers. In conclusion, this epidemiological study did not find any indication for an association of styrene with lympho-hematopoetic or pancreatic tumors. Possible associations with lung or kidney tumors were most probably related to smoking or a chance finding.

Rhomberg et al. (2013) assess the evidence for carcinogenicity in humans based on the US NTP criteria for “limited” or “sufficient” evidence. They review the epidemiology, experimental animal and mode of action data. They conclude that the epidemiology studies show no consistent increased incidence of any type of cancer. In animal studies, increased incidence rates of tumors have been observed only in mice and at one tissue site (lung). The lack of concordance of tumor incidence and tumor type among animals and humans indicates that there has been no particular cancer consistently observed among all available studies. The only plausible mechanism for styrene-induced carcinogenesis - a non-genotoxic mode of action that is specific to the mouse lung - is not relevant to humans. In summary, by this review it is concluded that a characterization of styrene as “reasonably anticipated to be a human carcinogen,” according to the criteria of the US NTP, is not justified.

Studies in animals

The carcinogenic potential of styrene has been explored in rats (Cruzan et al., 1998; NTP, 1979) and mice (Cruzan et al., 2001; NTP, 1979), using the inhalation (Cruzan et al., 1998; 2001) and oral (NTP, 1979) routes of exposure. A carcinogenic effect of styrene towards the lung is evident in the mouse. This has been shown in a well-conducted lifetime inhalation study in CD1 mice (Cruzan et al., 2001) at exposure concentrations of³20 ppm styrene and, somewhat less convincingly, in an oral study in mice of the B6C3F1 strain (NTP, 1979). The inhalation study, which included extensive histopathological examination, showed that the tumors (prevalently adenomas) were preceded by cytotoxicity characterized by early Clara cell toxicity followed by progressive bronchiolar epithelial hyperplasia and bronchiolar-alveolar hyperplasia.

In the rat, styrene has not exhibited any clear evidence of carcinogenic potential by the inhalation (Cruzan et al., 1998) or oral route (NTP, 1979). In individual studies (Conti et al., 1988; Jersey et al., 1978) there have been isolated findings of statistically significantly higher incidences of various particular tumor types in particular groups of styrene-treated animals, compared with the in-study controls. However, the findings have been within historical background ranges, not reproducible between studies, in some cases have not shown an upward trend with increasing dose, and have not been associated with evidence of underlying styrene-induced changes at the site in question.

On the question of the relevance of the mouse lung tumours for human health, consideration of the available toxicokinetic information (see also section on metabolism) and data from single and repeated inhalation exposure studies in experimental rodents suggests the following as the most plausible toxicological mechanism for the mouse lung tumours. Styrene is metabolised by cytochrome P450 enzymes in the metabolically active Clara cells (non-ciliated bronchiolar epithelial cells involved in the metabolism of xenobiotics, but also in the secretion of surfactants and in the renewal process of the bronchiolar epithelium) of the bronchiolar epithelium of the mouse, producing cytotoxic metabolites of styrene including styrene 7,8 oxide (SO) and oxidative metabolites of 4-vinylphenol (4-VP). These metabolites cause early Clara cell toxicity/death and sustained regenerative bronchiolar cell proliferation which, in turn, leads to compensatory bronchiolar epithelial hyperplasia and ultimately tumour formation (Gadberry et al., 1996;Carlson et al., 2002a;Cruzan et al., 2005;Kaufmann et al.,2005). Clara cell toxicity could also be a consequence of the long term depletion of glutathione, because of conjugation with SO (Kaufmann et al.,2005[B3] ). Genotoxicity of SO (an EU-category 2 and IARC group 2A carcinogen) or other reactive styrene metabolites is unlikely to be involved in tumour development as minimal binding of styrene metabolites to DNA has been detected in mouse lung with no species- or tissue-specificity.

All of the key events of this postulated mode of action are less operative in the non-responsive rat (Cruzan et al., 2005) (which does not develop lung tumours at exposure concentrations up to 1000 ppm) and even less operative in humans.

The number of Clara cells (being responsible for both the formation of toxic metabolites and the target for their toxic action) is very low in humans, even less than in rats. While Clara cells comprise about 85% of bronchiolar epithelium in mice and 25% in rats, in humans such cells are rare.

Although the enzymes CYP2E1 and CYP2F2 required for the formation of the Clara cell toxicants such as SO (including the highly pneumotoxic R-enantiomer) and the downstream metabolites of 4-VP have been detected in human lung, their activities are low (at least 400 times lower than in the mouse) and metabolic activation of styrene to SO is minimal or undetectable.

In human lung, detoxification of SO (if formed at all in human pulmonary tissue) takes place predominantly via epoxide hydrolase (located on the endoplasmatic reticulum in close proximity to the toxifying cytochrome P450s). The close proximity of the “detoxifying” enzymes to any “toxifying” enzymes ensures the efficient removal of any toxic metabolites. Rodents use both epoxide hydrolase and glutathione-S-transferase as detoxification pathways with the mouse relying on glutathione conjugation more so than the rat. As glutathione S-transferase is located in the cytosol, this makes this detoxification pathway less efficient than the epoxide hydrolase pathway. In comparison to the rodent species, in humans, SO detoxification proceeds nearly exclusively via epoxide hydrolase and glutathione S-transferase accounts for only 0.1% of SO detoxification (Vodicka et al.,2006a).

Taking account both of the toxification to SO and its detoxification, PBPK-modelling has shown that the SO content of human lungs is very small, if there is any.

Formation of 4-VP and its downstream metabolites occurs at a far higher extent in mouse lung than in rat (14-79% of the mouse concentrations) or human lung (1.5-5% of the mouse concentrations). Although it cannot be ascertained whether or not these species differences in the formation of 4-VP metabolites in the lung may be a reflection of the different numbers of Clara cells (the metabolically active lung cells) present in the different species, since 4-VP metabolites are produced by the same cytochrome P450 enzymes involved in the production of SO, it is most likely that the species differences in the formation of 4-VP metabolites observed reflect species differences in metabolic capability.

As indicated by PBPK-modelling, glutathione depletion caused by SO does not occur in humans. Also, as reactive downstream metabolites of 4-VP are formed in human lung only to a very small extent, the 4-VP metabolic pathway is not expected to cause any glutathione depletion in human pulmonary tissue.

There is no evidence from extensive epidemiological investigations that long term exposure to styrene has produced lung damage or lung cancer in humans.

Hence, overall, the weight of evidence appears to indicate that the consequences of long term exposure to styrene in mouse lung cannot be replicated in the human situation at relevant levels of exposure. Although there are still some uncertainties in this postulated mode of action and in its relevance to humans, namely the lack of data on the relative rates of 4-VP metabolites detoxification in different species, no alternative modes of action that logically present themselves can be supported by as significant a body of evidence as the one presented in this assessment. Consequently, it is felt that the level of confidence in the postulated mode of action can be reasonably high and that, in view of the extensive negative lung epidemiology, it is reasonable to conclude that the lung tumours seen in mice are unlikely to be of any relevance for human health. A more detailed analysis (according to the IPCS framework for evaluating a mode of action in chemical carcinogenesis) of the evidence in support of the proposed mode of action and of its relevance for human health is presented in Annex A to this document.

The carcinogenicity of styrene was evaluated by IARC in 2002. Styrene was considered possibly carcinogenic to humans (Group 2B). The Working Group concluded that based on metabolic considerations, it is likely that the proposed mechanism involving metabolism of styrene to styrene 7,8-oxide in mouse Clara cells is not operative in human lungs to a biologically significant extent. However, based on the observations in human workers regarding blood styrene 7,8-oxide, DNA adducts and chromosomal damage, it cannot be excluded that this and other mechanisms are important for other organs.

In the Rapporteur’s view, pointing to a possible carcinogenic potential of styrene in other organs is highly speculative as: a) Several large cohort and case-control studies of workers exposed to styrene have shown no evidence for a causative association between styrene exposure and cancer in humans at any site; b) No consistent evidence for styrene-induced toxicity in any organ has emerged from studies of exposed workers; c) The level of DNA damage found in workers exposed to styrene is very low (10-fold lower than that produced by endogenously-generated genotoxic substances such as ethylene oxide) and thus cannot be considered to be of any relevance for subsequent tumour formation. Mechanistic studies have shown that styrene-oxide (SO) and its genotoxicity are not the driving force for lung tumour formation in mice, the only experimental tumour site observed so far. Furthermore, DNA adducts in animals after styrene exposure do not show any specific species or target organ relationship. For example, there is no excess of SO-adduct formation in tissues where SO is formed (e.g. in the liver) at high levels; d) Chromosomal damage caused by styrene exposure in humans is far away from being conclusive. Although 5 studies appear to present evidence that styrene may be weakly clastogenic in humans, there are 11 robust negative studies also. Together with a lack of evidence of a dose-response relationship and the negative response for induction of micronuclei when studied concurrently in two of the positive chromosome aberration studies, no clear conclusion on in vivo clastogenicity of styrene in humans can be made. Furthermore, at much higher exposures such effects were not observed in experimental animals.

 

In the following the above mentioned “IPCS Conceptual Framework for Evaluating a Mode of Action” for Styrene Induced Mouse Lung Tumors is given (Annex A to the UK Risk Assessment Report):

Introduction

Styrene inhalation, at exposures ranging from 20 ppm up to 160 ppm, for up to 24 months induced lung tumours in mice (Cruzan et al, 2001). Statistically increased incidences of bronchioloalveolar adenomas, but not of carcinomas, were seen in male mice exposed to 40, 80 or 160 ppm styrene for 24 months. No statistically significant increases in lung adenomas were seen in the 20 ppm treatment group and there was no statistically significant dose response relationship. In females, (exposure 22.5 months), the incidence of bronchioloalveolar adenomas was increased significantly in the 20, 40 and 160 ppm treatment groups but with no significant increase at 80 ppm. Only females in the 160 ppm treatment group had a statistically significant increase in bronchioloalveolar carcinomas. No other tissues were affected. In rats however no tumours were observed with inhalation concentrations up to 1000 ppm (Cruzan et al, 1998).

In addition to the Cruzan et al (2001) study, older investigations in the mouse, but not meeting contemporary standards, have given some indication for possible lung tumour formation after oral application. A critical examination of results from these 4 mouse oral studies has been provided by McConnell and Swenberg (1993), who concluded because of methodological deficiencies and equivocal results the overall data were inadequate to reach any firm conclusions.

Results of eight long term rat studies, in which styrene was given by various dose routes, revealed no tumorigenic effects including no increase in lung tumors. This included an inhalation study, using contemporary protocol design, in which rats were exposed to levels up to 1000 ppm styrene for 24 months (Cruzan et al., 1998).

In addition to animal investigations there have been a variety of human epidemiological studies examining cancer incidence in the reinforced plastics industry, the styrene manufacturing industry and the styrene-butadiene rubber industry. These studies have been the subject of several reviews. For example Coggon (1994) concluded that despite the large size of published data it was not possible to rule out a hazard from long term exposures to high exposures (> 50 ppm) styrene. However the data indicated any risk from lower exposure levels was extremely small. In 2002 the Harvard Centre for Risk Analysis also conducted a comprehensive review of potential health risks associated with exposure to styrene (Cohen et al, 2002). The science panel put together by theconcluded the human epidemiological data did not support a hazard for lung cancer in humans exposed to styrene. The panel also calculated lifetime risk in highly exposed workers and concluded, based on the statistical power of the studies, the balance of epidemiological evidence does not suggest a causal relationship between styrene and any form of cancer. Similarly, after reviewing the available cohort and case-control studies of workers in the GRP industry, the styrene production and styrene-butadiene rubber industry the draft Risk Assessment Report of theconcludes that these mortality studies provide no evidence for a causative association between styrene exposure and cancer in humans.

The Mode of Action analysis given below relates to mouse lung tumors, the only tumor type clearly induced by styrene in carcinogenicity studies in rats and mice. To assess the significance of the mouse lung tumor data for human health a variety of mechanistic investigations have been carried out over the past 4 years. These investigations have been directed towards an understanding of the mode of action and aetiology of the mouse lung tumors. The data available up to 2002 have been reviewed by Cruzan et al (2002) and an explanation provided for the high sensitivity of mice for lung tumor development as compared to rats and to humans. In view of the complexity of the data from these investigations and the requirement to provide transparency and structure to analyzing the mode of action the approach defined in the IPCS framework for evaluating mode of action in chemical carcinogenesis has been used (Sonich-Mullin et al, 2001).

 

Postulated Mode of Action

In the following proposed mode of action data for mice, (species responding to styrene exposure with lung tumours), are compared to those for rats, (species without a carcinogenic effect). This interspecies comparison provides a valuable countercheck whether the proposed mode of action leading to lung tumour formation in the mouse actually differentiates between responding and non-responding species. In the concluding section an extrapolation from mice and rats to humans is presented.

The postulated mode of action is explained by the special cellular and biochemical characteristics of mouse lung which results in the mouse Clara cells producing cytotoxic metabolites of styrene including styrene oxide (SO) and oxidative metabolites of 4-vinyl phenol (4-VP). These metabolites cause early Clara cell toxicity and bronchiolar cell proliferation followed by progressive bronchiolar epithelial hyperplasia and focal crowding of cells. These cellular changes will finally lead to tumour formation.

There are several predisposing factors explaining the species differences in susceptibility to lung tumour formation. These include:

physiological differences with the mouse having much higher numbers of Clara cells as compared to rats and especially to humans.

pharmacokinetic differences at the cellular level with mouse Clara cells being more efficient than the lung of rats and humans in oxidative toxification of styrene to SO (including specifically the highly pneumotoxic R-enantiomer) and toxic metabolites of 4-VP.

detoxification of SO which takes place in rodents both via the microsomal epoxide hydrolase and the cytosolic glutathione S transferase. In humans the latter accounts for only 0.1% while detoxification of SO (if formed at all) is nearly exclusively mediated by epoxide hydrolase. Species differences in these detoxification pathways probably contribute to the higher sensitivity of mice in comparison to rats and humans

pharmacodynamic differences such as glutathione depletion, (i.e. glutathione depletion is more prominent in mice than in rats), also probably play a role at the cellular level making mouse Clara cells more susceptible to damage. Glutathione depletion is more prominent in mice than in rats and is not expected to occur in humans.

The course of pulmonary effects observed in the chronic mouse styrene inhalation study, i.e. Clara cell toxicity caused by production of styrene metabolites via the CYP 2F2 metabolic pathway, cell proliferation, cell tolerance accompanied by changes in Clara cell morphology and biochemistry, (e.g. decreased eosinophilia, loss of apical cytoplasm, focal crowding and reduced CC10 protein), followed sequentially by progressive Clara Cell hyperplasia is consistent with a non-genotoxic multistage mode of action. 

When looking at these relative susceptibilities it is clear that the mouse carries all of the factors for tumour production whereas humans and rats do not share these predisposing factors.

 

Key Events

The key events considered with respect to styrene lung tumourgenesis in the mouse include:

Clara cells are the target cells for lung toxicity and proliferation. Species susceptibility correlates with the number of Clara cells. For example an examination of distribution and numbers of Clara cells showed that in the mouse these cells are distributed from the terminal bronchioles to the trachea while in rat they are found only in the terminal and distal bronchioles. In the mouse Clara cells comprise about 85% of bronchiolar epithelium whereas in the rat only about 25% of the bronchiolar epithelium are Clara cells (Plopper et al, 1980a).

Formation of the toxic metabolites in Clara cells is mediated via Cytochrome CYP 2F2 a pathway more pronounced in mice than in rats. A variety of investigations have demonstrated that Clara cell toxicity in the mouse is mediated by CYP 2F2 generated metabolites such as SO (including specifically the highly pneumotoxic R-enantiomer) and oxidative metabolites of 4-vinyl phenol (4-VP) (Cruzan et al, 2002, 2005). With regard to the formation of the different SO enantiomers Gadberry et al (1996) demonstrated that the R-enantiomer is a more potent pneumotoxicant and hepatotoxicant than the S-enantiomer. Various investigations have been undertaken examining the formation of the different isomers in different species and tissues. For example, Hynes et al (1999) demonstrated that mouse Clara cells produce about 3-times more of R-SO than of S-SO, while the rat produces more of the S-enantiomer. Overall, mouse Clara cells produce 15-times more of R-SO than rat Clara cells. In addition to species differences in pulmonary SO formation there is also a clear inter-species difference in the formation downstream metabolites of 4-VP, which are highly toxic to mouse Clara cells (see below). Studies in which styrene was incubated with lung microsomes of mice and rats in the presence of glutathione showed a clear species difference in the formation of the toxic 4-VP ”downstream” metabolites in the order of mice > rats (Bartels, 2004). This further “downstream” metabolism of 4-VP will occur either from further side chain oxidation and, perhaps more importantly, from further aromatic hydroxylation to a catechol derivative which can easily undergo autoxidation to highly reactive o-Quinones.

Cytochrome CYP 2F2 mediated metabolites of styrene, i.e. styrene oxide and the downstream metabolites of 4-VP, are highly toxic to mouse Clara cells but the effect on rat lung is minimal. To investigate whether the cell damage and proliferation was associated with formation of oxidative metabolites rather than parent styrene Green et al, (2001a) treated mice with 5-phenyl-1-pentyne, a cytochrome P450 inhibitor prior to exposing to styrene (40 or 160 ppm, 6 hours per day for 3 days). The results of the study showed without the inhibitor cellular damage and increases in cell division in the lung of mice. In rats such effects were not observed at exposure levels up to 500 ppm. Pretreatment with the inhibitor protected against the toxic effects in mice indicating that the pulmonary pathological changes are caused by a toxic metabolite(s) formed by the cytochrome P-450 metabolism of styrene. The role of individual styrene metabolites in producing Clara cell toxicity has been examined by Kaufmann et al (2005) after i.p. application over 3 days. In summary, treatment of mice with styrene oxide and 4-VP caused up to 19-fold increases in cell proliferation in the large, medium and terminal bronchioles. Cell proliferation at the high dose was associated with toxic injury to Clara cells and regeneration, while at a lower dose level only cell proliferation without toxicity was observed. Treatment also caused glutathione depletion with an up to a 50% reduction in the number of Clara cells staining for glutathione (see below). Histopathological changes were characteristic of those seen following styrene exposure, i.e. flattened cells with loss of apical buldging into the bronchial lumina (Kaufmann et al, 2005). Other experiments with 4-VP substantiated the pronounced effects of this metabolite on bronchiolar cell proliferation. For example, Cruzan et al, (2005) reported hyperplasia of the terminal bronchioles in mice treated intraperitoneally at 6, 20 and 60 mg/kg/day for 14 consecutive days. The NOAEL was 2 mg/kg/day. No effects were however seen in the lungs of rats treated at 60 mg/kg/day or in the liver of mice or rats. A single high i.p. dose of 4-VP (50-200 mg/kg) in mice produced prolonged lung damage as demonstrated by effects on the broncho-alveolar lavage fluid (BALF) (Carlson, 2002b; Carlson et al, 2002a). At the single high i.p. dose of 150 mg/kg lung toxicity was found by BALF in rats, too (Carlson, 2002b). Pre-treatment of mice with 5-phenyl-1-pentyne, (inhibitor of CYP 2F2), or with Diethyldithiocarbamate prevented the toxicity indicating that a further metabolism is required to produce the toxic species (Carlson, 2002b). Diethyldithiocarbamate was primarily used as an inhibitor of CYP 2E1 (Carlson, 2002b) but subsequently it was shown that it also acts on other cytochromes (Carlson 2003, 2004a). Studies with other styrene metabolites specifically following the ß-oxidation pathway failed to cause Clara cell toxicity (see Other Modes of Action).

 

Differences in the detoxification pathways may contribute to the higher sensitivity of mice. The detoxifiation pathways of SO via epoxide hydrolase and glutathione S-transferase are qualitatively similar in mice and rats. A detailed analysis of the metabolic constants shows that epoxide hydrolase mediated detoxification is less effective in the lungs of mice as compared to rats. On the other hand, glutathione S-transferase activity is higher in mice than in rats (Filser et al.,2000).

 

Glutathione depletion in the lung as a contributing factor to cytoxicity and lung tumour development is more pronounced in mice than in. Studies of glutathione conjugation in the mouse lung have shown that mouse pulmonary tissue is susceptible to styrene induced glutathione depletion (Dhawan-Robi et al., 2000). Filser et al (2000; 2002) using measured data and PBPK modelling reported lung glutathione levels of mice exposed to styrene, (300ppm), for 6 hours/day for up to 3 days. After a single 6-hour exposure glutathione levels were reduced by 20%, which by day 3 had declined to about 60% of that seen in untreated mice. Similar treatments produced only minor effects on glutathione levels in the rat. Furthermore, Kaufmann et al (2005) showed histopathologically by a specific glutathione stain that SO and 4-VP caused glutathione depletion in the bronchiolar epithelium of mice after intraperitoneal application over 3 days. Overall, it is highly likely that the combination of both SO and 4-VP conjugation results in glutathione depletion at the cellular level in the mouse. Based on experience with other mouse Clara cell toxicants such as Coumarin (Vassallo et al, 2000) it is likely that glutathione depletion plays an important contributing role in lung tumour formation in the mouse. The biological relevance of glutathione depletion is decreasing from mice to rats.

 

Tumour formation is a late event (i.e. not observed at 12 or 18 months) associated with progressive hyperplasia in terminal bronchioles. As has been discussed earlier Clara cell toxicity, followed by extensive cell replication and subsequent hyperplasia with focal crowding appear to play key roles in development of lung tumours in the mouse. No lung pathology has been seen in the rat. While Clara cell toxicity and cell replication moderated with continued styrene exposure (probably as the Clara cells became more tolerant to styrene), clear evidence of hyperplasia was seen in mice following 12 and 18 months of exposure (Cruzan et al, 2001). As no changes were ever observed in the alveoli the Clara cell pathology is both a species (mouse) and cell (Clara cell) specific phenomenon.

 

Dose-Response Relationship

The chronic inhalation study in the mouse while indicating that exposure of mice to styrene via inhalation increases the incidence of lung tumours did not show a statistically significant dose-response relationship. i.e. % female and male mice with lung tumours being 12, 32, 34, 22 and 54% and 34, 48, 72, 60 and 72% respectively at doses 0, 20, 40, 80 and 160 ppm (Cruzan et al, 2001). This lack of a clear dose-response relationship is reflected in the Key Events leading to tumour formation.

Clara cells are the target cells for lung toxicity and cell proliferation. The lack of a clear dose response relationship in tumour development is mirrored in the lack of a clear dose response for the effects on Clara cells in the mouse. For example, Green et al (2001a) exposed mice for up to 10 days with styrene at 0, 40 or 160 ppm. At various points pulmonary labelling indices (using BrdU labelling technique) were measured in large and terminal bronchioles of exposed mice. Although treatment at 160 ppm produced a slightly larger labelling index than was seen at 40 ppm both treatments produced significant increases as compared to controls. Similarly, Cruzan et al (1997) in a study in which mice were exposed to styrene for 2, 5 or 13 weeks at 0, 50, 100, 150 or 200 ppm found histopathological changes, (i.e. decreased eosinophilia and focal crowding) in bronchial epithelium in the majority of mice in all treatment groups. Studies examining labelling indices found higher labelling index at 150 or 200 ppm as compared with lower doses after 2 weeks of exposure and in occasional mice after 5 weeks. Similar to the tumour incidence there appears to be only a weak dose response relationship. Differences and variability in individual susceptibility to styrene in mice explains the lack of a clear dose response relationship to Clara cell toxicity and resulting tumour formation. No lung effects were seen in rats up to 1500 ppm after 13 weeks.

 

Formation of the toxic metabolites in Clara cells is mediated via Cytochrome CYP 2F2. As described above CYP 2F2 present in the mouse Clara cells is responsible for metabolising styrene to cytotoxic metabolites. Pre-dosing mice with the inhibitor of CYP 2F2, 5-phenyl-1-pentyne, is effective at preventing Clara cell toxicity and proliferation. There is evidence that the levels of the mouse Clara cell toxicant, SO, increases proportionally in mouse lung tissue with increasing styrene exposure. In a species comparison with ventilated perfused lungs of mice and rats Filser (2004) and Hofmann et al. (2006) concluded that at styrene exposures up to 500 ppm the levels of SO in mouse pulmonary tissues are at least 3-fold higher than those in rat lung. However the lung burden of SO in mice at styrene exposures producing lung tumours, e.g. 160 ppm are similar to SO lung burden in rats at 1000 ppm styrene exposure which was without effect. While such data support the fact that tumour induction in the mouse is not a genotoxic phenomenon it raises question about why mouse Clara cells are so sensitive to oxidative metabolites of styrene including SO. A possible reason again relates to the fact that the measures of SO formation are an aggregate assessment based on the total lung burden whereas in fact the interest is at the Clara cell level. The importance of considering events at the cellular level rather than the tissue level has been highlighted by Green (2000b) who exposed mice to 0, 40 or 160 ppm styrene for up to 6 hours after which lungs were excised and one lobe used to estimate CYP 2F2 levels with Western blotting while another lobe was subjected histopathological examination. Pulmonary CYP 2F2 was found to be remarkably variable with up to 5-fold differences between animals. There was also no correlation between histopathological changes and levels of CYP 2F2 in individual animals. Thus while the presence of CYP 2F2 is necessary to generate toxic metabolites there is no simple relationship between metabolism and morphological damage which explains why some mice in the 40 ppm treatment group showed more severe histopathological effects than did mice in the 160 ppm treatment group. An explanation for the variability in histopathology probably lies not only in the presence of the enzymes for styrene activation but also in the ability of the cell to detoxify the toxic metabolites. This complexity of toxification and detoxification in the mouse Clara cell again explains the lack of a clear dose-response relationship in tumour development.  

 

Glutathione depletion in the lungs is a contributing factor to cytotoxicity and lung tumour formation. Filser et al (2000; 2002) and Dhawan-Robi et al. (2000) reported both measured and modelled glutathione concentrations in the lungs of mice and rats exposed to 300 ppm styrene for 6 hours. After a single exposure glutathione levels in mouse was reduced by 20% and after 3 days exposure reductions had increased to 60 %. At 40 ppm there was however no statistical differences between glutathione levels in the lungs of styrene treated and control mice. At 80 ppm styrene treatment there was a statistically significant decrease in glutathione levels on the second day of exposure. At 160 ppm greater reductions in mouse pulmonary glutathione levels were reported. In rats glutathione reductions in lungs only became statistically significant at exposure concentrations of 300 ppm and above. The lack of aggregate glutathione depletion in pulmonary tissue at styrene exposures associated with pulmonary toxicity has raised a question about the biological relevance of glutathione depletion in mouse lung toxicity and development of pulmonary tumours. It is important to remember that all measurements were of aggregate glutathione levels in the lung as a total. Examination of lung from mice exposed to styrene have however shown large variations in glutathione levels at the cellular level, (Kaufmann et al 2005) with, in the same animal, some Clara cells being devoid of glutathione while other cells showed normal levels. Because of variations in individual animals and individual cells within the same animals it is difficult to show a clear dose response effect for glutathione depletion especially at lower dose levels. Again the lack of a clear dose-response in glutathione depletion is consistent with the lack of a clear dose response in tumour formation.

 

Tumour formation is a late event, (i.e. not observed at 12 or 18 months), associated with progressive hyperplasia in terminal bronchioles. The chronic inhalation study in the mouse reported by Cruzan et al (2001) showed a time related and dose-related increase in bronchiolar hyperplasia in the terminal bronchioles and bronchiolar epithelial hyperplasia extending into the alveolar ducts. Results of interim sacrifices indicated the presence of these hyperplasias in mice exposed to 160 ppm styrene for 12 months and in mice exposed to 40 ppm styrene for 18 months. These hyperplastic changes finally transformed into lung tumors at the end of the exposure period. Again these data show no clear dose-response but the data do indicate a concentration*time dependence.

 

Temporal Associations

To develop an understanding of the temporal changes in lung pathology following long term inhalation exposure to styrene groups of mice were killed 2, 5, 13, 52 and 78 weeks after start of treatment. Histopathological examination of the lungs showed a steady progression of the lesions, starting with decreased eosinophilia and focal crowding of nonciliated cells in the bronchiolar epithelium continuing to bronchiolar hyperplasia in the terminal bronchioles extending finally into the alveolar ducts. At the end of the 2 years of treatment all mice at all doses levels were affected. Only at the study termination increases in the incidence of lung tumors were observed (Cruzan et al, 2001).

Various investigations have characterized the time course of cellular changes in mouse lung following repeated inhalation exposures to styrene. For example Cruzan et al (1997) conducted a study in which in which CD-1 mice were exposed to 0, 50, 100, 150 and 200 ppm styrene for up to 13 weeks. Some mice were also killed after 2 and 5 weeks of exposure. Histopathological changes to the bronchiolar epithelium were characterized by decreased eosinophilia and focal crowding of Clara cells in the bronchiolar epithelium considered to represent cell proliferation. A study of cell proliferation using BrdU labelling identified a statistically significant increase in Clara cell replication (doses 150 and 200 ppm) after 2 weeks of treatments with an approximate 4-fold increase in percentage of labelled cells as compared to untreated controls. After 5 weeks of treatment a non significant increase (isolated animals) was seen in the treated mice while at 13 weeks no increase in the percentage of labelled cells was seen. The large variations in the labelling index among animals seen in the investigation are consistent with the poor dose response relationship for tumour formation as mentioned above. The cell proliferation was only seen in Clara cells with no effects in toxicity or increased BrdU labelling being seen in the Type II pneumocytes (Cruzan et al, 1997). These findings are similar to those reported by Roycroft et al (1992) who studied pulmonary changes in B6C3F1 mice exposed to up to 500 ppm styrene.

 

Biological Plausibility & Coherance

The postulated mode of action for the mouse lung tumours induced by inhalation exposure to styrene is consistent with what is known about the metabolic competence of mouse Clara cells and their capacity to form reactive intermediates by the CYP 2F2 metabolic pathway (Cruzan et al, 2002). It is widely accepted that toxicity and mitogenesis are of critical importance in expression of non-genotoxic carcinogenicity and sustained regenerating cell division secondary to cytotoxicity is accepted as a mode of action for many non-genotoxic carcinogens. The time course of pulmonary effects observed in the chronic mouse styrene inhalation study, i.e. Clara cell toxicity, cell proliferation, changes in Clara cell morphology, (e.g. decreased eosinophilia, loss of apical cytoplasm and focal crowding), followed sequentially by Clara Cell hyperplasia is consistent with a non-genotoxic multistage mode of action. 

The data show that mouse Clara cells develop a tolerance to cytotoxicity and cell replication with repeated styrene exposure. This tolerance is not only associated with a change in Clara cell morphology as described above but also with a change in the biochemistry of the cells. For example Gamer et al. (2001; 2004) reported changes in both glutathione and CC16 protein levels in mice treated with styrene for up to 28 days. Similar results have been seen with the mouse Clara cell toxicant coumarin which has also been reported to produce reductions in glutathione levels, loss of cell secretory protein CC10 (CC16) and increases in GGT, (gamma glutamyl transpeptidase), activity (Vassallo et al, 2000). It has been proposed that the development of Clara cell tolerance to cytotoxic chemicals such as coumarin is accompanied by both morphological and biochemical changes in the cell and may be an important step in the multi-step progression to tumour development.

 

Other Modes of Action

Genotoxicity as a Mode of Action

A series of experiments have been directed to the question of whether or not genotoxicity may play a role in tumour development. Specifically investigations were carried out to determine if:

any adduct formation is high enough to explain tumour formation

there is a difference in adduct formation between rats and mice or

there is a difference in adduct formation between target (lung) and non-target tissues (liver).

A post-labelling experiment with an inhalation exposure of two weeks did not lead to DNA-binding (detection limit 3-5 adducts / 10exp+7) in mice at 40 and 160 ppm and in rats at 500 ppm while post-labelling studies in the livers and lungs of mice and rats failed to find any correlation between tumour development and DNA adduct formation (Otteneder et al, 2002). In a second experiment with higher sensitivity rats and mice were exposed to 14C-Styrene, (with the highest possible specific radioactivity). Exposure was over a 6-hour period at 160 ppm (Boogaard et al, 2000). In this latter investigation minimal DNA-binding (1-5 adducts /10exp+8) was detected. This low level of adduct formation could not account for lung tumour formation. In addition there was no difference in DNA-binding of rats and mice (i.e. no species specificity) and DNA binding was less in the lung as compared to the liver (i.e. no tissue specificity).

Such data indicate that genotoxicity is unlikely to be involved in lung tumour formation.

Pneumotoxicity of Metabolites from ß–Oxidation Metabolic Pathway as a Mode of Action

An examination of on urinary metabolites (Johanson et al, 2000 and Sumner et al, 2001) showed that the ß-oxidation pathway (leading to 2-phenylethanol and via the aldehyde to phenylacetic acid) is a more prominent metabolic pathway in mice as compared to rats. Studies were therefore undertaken to explore the possibility that metabolites from the ß-oxidation pathway could be pneumotoxic to Clara cells thus helping to explain species differences. A series of investigations in mice were carried out by carried out by Kaufmann et al (2005) to measure lung cell proliferation after intraperitoneal application of the metabolites from the ß-oxidation pathway 2-phenylethanol, phenylacetaldehyde and phenylacetic acid. In addition, 1-phenylethanol and acetophenone were given. None of the metabolites produced Clara cell damage or cellular proliferation in the bronchiolar epithelium. Although treatment with the metabolites did produce an increased proliferation of the alveolar cells the response is not considered relevant to tumour induction as no evidence of styrene related pathological effects have been seen in the alveoli of mice (or rats) at any exposure concentration or at any time point. Thus, this pathway cannot account for lung tumour formation starting from the bronchiolar epithelium.

 

Oxidative Stress as a Mode of Action

To examine the possibility that oxidative stress may be playing a role in tumour development female mice were exposed to styrene at concentrations of 40 or 160 ppm (Gamer et al., 2001; 2004). Animals were killed after 1, 5 or 20 exposures and the lungs examined by light and electron microscopy for histopathological effects and lung lavage fluid was examined for cell counts, protein, lactate dehydrogenase, alkaline phosphatase, N-acetyl-ß-D-glucuronidase, glutamyl transferase, glutathione, CC 16 protein (marker for Clara cells) and lysozyme. Lung homogenates were analysed for 8-OH-deoxyguanosine in lung DNA, malondialdehyde, catalase, glutathione reductase, superoxide dismutase, glutathione peroxidase and glutathione as markers of oxidative stress. Finally CC16 protein was measured in blood serum.

Although a depletion of glutathione was seen in lung homogenates after 20 exposures there was no evidence of substantial oxidative stress as indicated by unchanged levels of 8-OH-deoxyguanosine, superoxide dismutase, malondialdehyde, catalase, glutathione reductase, and glutathione peroxidase. Examination of lavage fluid provided no indication of inflammatory responses. A concentration-time related increase in activity of glutamyltransferase indicating cytotoxicity to Clara cells was present over the study period with a maximum level after 5 exposures. Clara cell involvement was substantiated by reductions in concentrations of CC16 protein in the lavage fluid and in blood serum throughout the study.

Histopathological changes in the lung after a single treatment were characterized by vacuolation and desquamation of secretory cells in the large and medium bronchioles. After 5 and 20 exposures cellular crowding, indicative of early hyperplasia was seen in the large and medium bronchioles. Electron microscopy showed Clara cells were the main target cells. After 20 days of damage phenotypic changes were noted in the Clara cell population with some of the cells having normal features of Clara cells but with others exhibiting slightly different features such as absence of apical blebs and reduction in secretory granules. Overall, the data do not support the hypothesis that substantial or sustained oxidative stress plays a role in styrene induced tumour development.

In summary three other potential modes of action were investigated. The results of these investigations show genotoxicity is not a key factor in tumour development. Styrene metabolites generated via the ß-oxidation pathway also play no role in Clara cell toxicity and oxidative stress can be excluded as a mode of action.

Assessment of Postulated Mode of Action

Overall the results of investigative studies on styrene indicate that the susceptibility of mice to development of lung tumours is linked to the following factors:

A greater number of Clara cells in mouse pulmonary tissues as compared to rats.

Mouse Clara cells contain high levels of the cytochrome P-450 isoform, CYP 2F2 responsible for metabolising styrene to metabolites highly toxic to mouse Clara cells, (i.e. styrene oxide and oxidative metabolites of 4-VP). This metabolic toxification occurs much less in rats. These metabolites have been shown to lead to early lung toxicity and cell proliferation in mice, but such effects have not been found in rats.

The detoxification pathways involved may not be efficient enough in mice and thereby can lead at the cellular level to relatively high levels of the toxic species described above as compared to the rat.

Pharmacodynamic differences at the cellular level, probably related to glutathione depletion, may account for the susceptibility of mouse Clara cells to cellular injury whereas rats are more resistant.

 

These factors are consistent with styrene selectively targeting mouse Clara cells causing cytotoxicity and increased cell replication rates. While these early effects moderate with continued exposure, (presumably as Clara cells develop some tolerance to styrene toxicity), clear evidence of bronchiolar epithelial hyperplasia is seen in mouse lung following 12 and 18 months of exposure (Cruzan et al, 2001). 

The lack of histpathological effects in the alveoli of the mouse is consistent with the fact that Type II pneumocytes do not possess CYP 2F2 required to generate the toxic metabolites of styrene.

The morphological effects, (including the change in phenotype and the absence of tumours at 18 months indicate a late developing tumour consistent with progressive hyperplasia), can be linked directly to the CYP 2F2 mediated metabolism of styrene producing metabolites toxic to mouse Clara cells. Thus both the histomorphological and biochemical data provide a high level of support for the postulated mode of action.

 

Uncertainties, Inconsistencies and Data Gaps

Measured kinetic data has enabled the development of PBPK-models based on SO formation (Sarangapani et al, 2002 and Csanardy et al, 2003). The model of Sarangapani et al (2002) specifically estimated styrene and SO levels in the terminal bronchioles vs. whole lung. It predicts a 2-fold higher level of the R-SO isomer in mouse lung at 40 ppm than in rat lung at 1000 ppm. But it seems improbable whether this difference in one specific isomer solely would explain the marked species difference.

Overall, while the models have slight differences they have led to qualitatively similar results:

The exposure of lung cells to SO predominantly results from pulmonary metabolism of styrene with minimal contribution from SO formed in the liver.

At relatively low styrene inhalation exposures the local concentration of SO (including the R-isomer) in the lung is higher in mice than in rats.

 

Overall SO lung burden (specifically R-SO lung burden) is higher in mice than in rats.

While the mouse may have greater metabolic capacity for conversion of styrene to styrene oxide and its R-isomer the pharmacokinetic data shows that the pulmonary concentrations of SO in mouse exposed to 40 ppm, (exposure producing lung tumours), is not much different from pulmonary concentrations of SO in rats exposed to 1000 ppm where no excess of lung tumours were seen. Thus it has been proposed that the aggregate measures of SO exposure in the lung do not appear to sufficiently explain the relative susceptibilities of the mouse and rat for lung tumour development (Cohen et al, 2002 and Filser, 2004 and Hofmann et al. 2006).

The proposal that glutathione depletion may play an important role in mouse Clara cell toxicity and subsequent carcinogenesis has also been questioned. For example Cohen et al (2002) pointed out that exposures to 40 ppm of styrene did not produce significant glutathione depletion in mouse lung but such treatment did produce hyperplasia and even tumours. As described earlier it is however important to differentiate effects at the tissue level and effects at the cellular level. At the tissue level lung surfactant contains quite high levels of glutathione especially for “scavenging” reactive species while at the cellular levels glutathione can vary 4-fold between individual mouse Clara cells. This heterogeneity helps explain why in studies with 4-VP approximately 50% of mouse Clara cells did not stain for glutathione while the other 50% stained with a similar intensity to untreated controls. Thus at the tissue level there may be no apparent change in glutathione whereas at the cellular level individual cells may have low if any ability to conjugate reactive species making them extremely susceptible to damage.

To date the formation and metabolism of 4-VP, including the striking semi-quantitative species differences, has only been investigated in vitro. How this may relate in quantitative terms to the in vivo situation and especially to the levels in the target cells, the Clara cells, requires study. Taking account of the metabolic instability of the 4-VP downstream metabolites and the predominant importance of their concentrations at the cellular levels as compared to the “aggregate” total lung or even the whole body, it is not expected that such studies will be accomplished easily. Nevertheless, such studies are of importance in refining the mode of action.

Even in the absence of data from such studies there is compelling evidence that mouse Clara cells possess unique properties, which come together or act in combination for lung tumour development in mice. These unique properties include:

1. sensitivity to the cytotoxic effects of SO, including specifically R-SO,

2. sensitivity to cytotoxic effects of 4-VP metabolites,

3. glutathione depletion in individual Clara cells

4. cellular regeneration and development of tolerance resulting in morphological and biochemical changes to the Clara cell, and

5. long term hyperplasia with cellular/focal crowding.

None of these properties are operational in the rat and hence the rat does not develop lung tumours.

 

Conclusion and extrapolation to humans

Sensitive in vivo DNA-binding studies in mice and rats have shown that genotoxicity is most unlikely to be the driving force for lung tumour development in mice. If a non-genotoxic mechanism is operative, a threshold (or a highly non-linear) dose response relationship can be assumed and three questions required answers:

what mechanism is leading to lung tumors in mice?

why do lung tumors do not develop in rats?

what sensitivity is to be assumed for humans in relation to mice and rats?

 

Answers to questions 1. and 2. have been provided in the preceding sections. Basically the high sensitivity of mice in comparison to the non-responsiveness of rats for lung tumour induction depends on multiple factors:

1) Clara cells in mice both generate the toxic metabolites as well as being the target cells for these metabolites. The number of Clara cells is much higher in mice than rats.

2) There are two oxidative pathways for formation of toxic metabolites:

formation of SO is more pronounced in the lung of mice while the detoxification of SO may be more effective in rats. In total the local concentration of SO is higher in mice than in rats

4-VP and its downstream metabolites, which are cytotoxic for mouse Clara cells, are formed to a greater extent in the lungs of mice as compared to the rat.

3) These metabolites lead to greater glutathione depletion in mice than in rats: glutathione depletion may be a contributing factor for tumour formation.

4) These factors lead to an increase in cell proliferation and histopathological changes as pre-stages of lung tumour development in the mouse. Inhibition of oxidative metabolism in mice prevents Clara cell cytotoxicity and proliferation in this species as the first steps in tumour formation.

The lack of tumour formation in the rat is explained by the fact that rats do not produce the necessary cellular concentrations of toxic metabolites to cause Clara cell toxicity and replication resulting in the morphological and biochemical changes in Clara cells as a first stage in the multi-stage process to tumour development. The resistance of the rat and the sensitivity of the mouse to Clara cell toxicity are explained, as described above, by the physiological, pharmacokinetic and pharmacodynamic differences between the species.

The answer to the question about human sensitivity is found in the fact that the Key Events for the postulated mode of action are even less operative in humans as compared with the non-responsive rat. For example:

The number of Clara cells, (being both responsible for the formation of toxic metabolites and the target for their toxic action), is very low in humans, even less than in rats. While Clara cells comprise about 85% of bronchiolar epithelium in mice and 25% in rats, in humans such cells are rare (Plopper et al, 1980a).

The level of the enzyme CYP 2F2 required for the formation of the Clara cell toxicants such as SO, (including the highly pneumotoxic R-enantiomer), and the downstream metabolites of 4-VP occurs at best only to a negligible extent in humans. Studies with human lung tissues have shown a general lack of CYP 2F2 activity in human pulmonary tissue leading to the formation of SO. In a study of 38 human lung samples, Nakajima et al (1994b) found only low levels of cytochrome P450 dependent monoxygenase activity in some of the samples; the highest level measured was however approximately 400-fold lower than the average measurement in rat lung. Carlson et al (2000) reported only a slight styrene metabolising capability in 1 of 8 human lung samples while Filser et al (2002) and Oberste-Frielinghaus (1999) failed to find any evidence of CYP 2F2 activity in human lung samples and formation of SO.

In human lung, detoxification of SO (if formed at all in human pulmonary tissue) takes place predominantly via epoxide hydrolase (located on the endoplasmatic reticulum in close proximity to the toxifying cytochrome P450s). This close proximity of the “detoxifying” enzymes to any “toxifying” enzymes ensures the efficient removal of any toxic metabolites. Rodents use both epoxide hydrolase and glutathione-S-transferase as detoxification pathways with the mouse relying on glutathione conjugation more so than the rat. As glutathione-S-transferase is located in the cytosol this makes this detoxification pathway less efficient than the epoxide hydrolase pathway. In comparison to the rodent species, humans represent an extreme as SO detoxification nearly exclusively proceeds via epoxide hydrolase and glutathione S-transferase accounts for only approx. 0.1% of SO detoxification.

Taking account both of the toxification to SO and its detoxification, PBPK-modelling by two approaches (Sarangapani et al, 2002 and Csanardy et al, 2003) have shown that the SO content of human lungs must be very small, if there is any.

Formation of 4-VP and its downstream metabolites in humans is less than in rats and much less than in mice (Bartels, 2004). This has been demonstrated by a comparative incubation of lung microsomes from mice, rats and humans using glutathione as a “trapping” reagent for the reactive, unstable downstream metabolites of 4-VP. Such metabolites are formed either from further side chain oxidation or more importantly from further aromatic hydroxylation to a highly reactive catechol derivative.

Glutathione depletion caused by SO does not to occur in humans. This has been demonstrated by PBPK-modelling by Filser et al (1999 final report is 2000, 2002) using the biokinetic data for the toxification of styrene to SO and the detoxification of the latter via epoxide hydrolase and glutathione-S-transferase. As reactive downstream metabolites of 4-VP are formed in human lung only to a very small extent (Bartels 2004), the 4-VP metabolic pathway is not expected to cause any glutathione depletion in human pulmonary tissue.

 

As all of the Key Events are not, (or at best only to a negligible extent), operative in humans, the early histopathological and biochemical changes leading ultimately to tumours are not to be expected to occur in human lung.

 

A literature search from 1998 to January 2010 overlapping the date of the UK RAR (June 2008) was conducted. Overall, this did not lead to a modification of the mode of action and reinforced some of the aspects elaborated above.

 

In contrast to the findings in mice mentioned above (Gamer et al., 2001; 2004), no lung toxicity was observed in female rats exposed to styrene at 160 and 500 ppm over 5 days. The investigation included histopathology of the lung, cytological and humural constituents of the lung lavage fluid, and determination of CC16 in lavage fluid and blood (Gamer et al., 2002).

 

Carlson et al. (2002c) studied lung and liver toxicity in mice after a single intraperitoneal application of 1- and 2-phenylethanol (100 mg/kg bw each) and phenylacetaldehyde (100, 250, and 500 mg/kg bw). The treatments did not lead to lung toxicity (investigation of lung lavage fluid) or to liver toxicity (sorbitol dehydrogenase in blood) apart from phenyacetaldehyde at 500 mg/kg bw showing that these side chain oxidized metabolites are not operative in styrene induced toxicity.

Dhawan-Robi et al. (2000) exposed rats and mice by inhalation for two days (6 h/day) up to 300 ppm. Immediately after the second exposure the glutathione concentrations in the lungs were determined. A significant decrease of lung glutathione was observed in mice exposed at and above 80 ppm, while a significant decrease was found in rats only at 300 ppm.

 

Carlson et al. (2006) investigated the effect of a single intraperitoneal application of styrene (600 mg/kg bw) and the R- and S-isomers of styrene oxide (300 mg/kg bw) on the glutathione content and the antioxidant status in the lung and liver of mice. All treatments led to a decrease of glutathione in lung and liver and the effect was more pronounced by the R- isomer as compared to the S-isomer of styrene oxide. A decrease of oxidized glutathione (GSSG) was also found in lung and liver after styrene application. There was no treatment related effect on lipid peroxidation as measured by determination of malondialdehyde and only minimal effects on glutathione reductase and glutathione peroxidase. Styrene treatment increased gamma-glutamylcysteine synthetase activity. The authors conclude that in response to decreased glutathione levels there is an increase of its synthesis. In a similar experiment Turner et al. (2005) observed significant glutathione depletion after single ip applications of styrene starting at 200 mg/kg bw and at 250 mg/kg bw for styrene oxide. 100 mg/kg bw of 4-vinylphenol also led to decreased levels of glutathione in lung and liver but the degree of depletion was less as compared to styrene and styrene oxide.

Meszka-Jordan et al. (2009) investigated oxidative stress by the R-isomer of styrene oxide and the role of antioxidants on the toxicity to lung (bronchoalveolar lavage) and liver (sorbitol dehydrogenase in serum). Mice were dosed once with 300 mg/kg bw R-styrene oxide intraperitoneally. The antioxidants (glutathione, N-acetylcysteine, 4-methoxy-L-tyrosinyl-gamma-L-glutamyl-L-cyseinyl-glycine) did not decrease lung toxicity, while glutathione and N-acetylcysteine protected against liver toxicity. The relation of this finding to the carcinogenic mode of action remains unclear.

Harvilchuck and Carlson (2006) incubated Clara cells isolated from mice and rats with styrene, styrene oxide (racemic, R-, and S-), and 4-vinylphenol. In vitro the cytotoxicity of styrene was greater than that of its metabolites in contrast to the observations in vivo. Susceptibility of rat Cara cells was 4-fold less than that of mice Clara cells. In vitro treatment led to a decrease of glutathione in mouse Clara cells. After in vivo ip application a rebound effect of the glutathione concentration in the Clara cells of mice was observed after 12 h. By comparing the effects observed in vitro and in vivo the authors concluded that in vitro cytotoxicity of styrene and its metabolites does not strictly follow the in vivo effects.

The decrease of the Clara cell specific protein CC10 described by Gamer et al. (2001; 2004) was substantiated by Harvilchuck et al. (2008) after a single intraperitoneal application of styrene (600 mg/kg bw), styrene oxide (racemic, R-, and S-, each 300 mg/kg bw), and 4-vinylphenol (100 mg/kg bw). Clara cells were isolated and CC10 mRNA and protein were measured. The largest decrease of CC10 mRNA was seen with R- and racemic styrene oxide (SO). Styrene caused a significant decrease in CC10 protein after 24 h, rebounding through 240 h. Thus, acute changes in lung CC10 protein and mRNA expression occurred following in vivo treatment with styrene and its metabolites. After intraperitoneal exposure of mice over 5 consecutive days to styrene (600 mg/kg bw) and R-styrene oxide (300 mg/kg bw) CC10 mRNA was decreased directly after exposure (Harvilchuck and Carlson, 2009). At the protein level CC10 remained decreased 10 days following the last dose of R-SO.

 

While decreased GSH levels generally were observed after treatment with styrene, SO, or 4-vinylphenol (Gamer et al., 2001; 2004; Turner et al., 2005; Carlson et al., 2006; Harvilchuck and Carlson, 2006) there was no clear effect on parameters of oxidative stress and the antioxidant status (Gamer et al., 2001; 2004; Carlson et al., 2006; Meszka-Jordan et al., 2009). Harvilchuck et al. (2009) investigated early indicators of oxidative stress and apoptosis caused by styrene, SO (racemic, R-, and S-) and 4-vinylphenol. After in vitro incubation of Clara cells styrene and SO led to an increase of reactive oxygen species (ROS). For the in vivo experiments mice were exposed once intraperitoneally to styrene (600 mg/kg bw), the different enantiomers of SO (300 mg/kg bw), or 4-vinylphenol (100 mg/kg bw) and Clara cells were isolated. Increases in ROS, superoxide dismutase, and 8-hydroxydeoxyguanosine were observed after treatment with styrene and SO. The ratio of bax/bcl-2 was further investigated as an indicator of the susceptibility of a cell to apoptosis. The bax/bcl-2 mRNA ratio was increased 12 and 24 h following R-SO and 120 h following styrene administration. The bax/bcl-2 protein ratio was not increased until 240 h following R-SO, and 24 and 240 h following styrene treatment. Caspase 3 was only slightly increased. The activation of this protein leads to the morphological changes associated with apoptosis. These results indicated that oxidative stress occurred early after styrene or styrene oxide treatment while there was only limited apoptosis in Clara cells following acute exposure.

 

Carlson (2008) reviewed the expression of cytochrome P450 enzymes in human lung and assessed the possibility whether these enzymes might be involved in styrene tumorigenicity in humans. Metabolic activation rates were compared between liver and lung, and for the lung, between mice rats and humans. In general, pulmonary metabolism is very slow compared to hepatic metabolism. Furthermore, metabolic rates in humans are slow compared to those in rats and mice. Pharmacokinetic and PBPK models have indicated that very little styrene oxide produced in liver would be expected to be found in the lung by distribution via blood circulation. Therefore focus is on the in situ generation of active metabolites in human lung. The ability of human lung to metabolize styrene to styrene oxide is extremely limited and has only been demonstrated in rare cases. While the complete picture of the different cytochromes P450 involved in lung metabolism of styrene is not yet completely understood, CYP2F2 has been identified in mice as being primarily responsible for styrene metabolism in the lung. CYP 2F1, the corresponding human homolog of CYP 2F2, has a very low activity in the human lung. According to the authors, this argues against the hypothesis that human lung would produce enough styrene oxide (or other reactive metabolites) to damage pulmonary epithelial cells leading to cell death, increased cell replication and ultimately to tumors, in analogy to the mode of action proposed for lung tumorigenicity in mice.

 

Cruzan et al. (2009) evaluated the possible mode of action for several chemicals leading to lung tumors specifically in B6C3F1 mice but not in rats. Tumors were found at the same locations for these substances, namely in the outer layer of the lung where the terminal bronchioles and alveoli intersect. The compounds considered were coumarin, naphthalene, ethylbenzene, alpha-methylstyrene, cumene, divinylbenzene, and benzofurane besides styrene. Genotoxicity was of no or only low relevance for this site of action. Although a complete picture of the mode of action had not been developed for any one of these chemicals, the data from the individual substances were synthesized and a model was developed concluding that lung tumor development in mice is driven by lung toxicity mediated by CYP 2F2 metabolism.

Lung toxicity by these chemicals occurs in the terminal bronchioles of mice but not of rats. The target cells are the Clara cells (but not the alveolar cells) and increased cell proliferation has been demonstrated by BrdU labeling after short-term exposure for several of these chemicals, including styrene.

For coumarin, naphthalene, and styrene it was demonstrated that CYP 2F2 inhibition led to an inhibition of lung toxicity. For styrene, the ring oxidized metabolite 4-vinylphenol was found to be a potent lung toxicant (Carlson et al., 2002a) and toxicity was most probably mediated by further metabolism of 4-vinylphenol (Carlson, 2002). In this context it is important to note that no detectable lung toxicity was observed from exposure to side chain hydroxylated metabolites of styrene (1- and 2-phenylethanol, phenylacetaldehyde, phenylacetc acid, acetophenone). On the other hand, ip treatment with 4-vinylphenol led to a significant increase of cell proliferation in the large and medium bronchi and terminal bronchioles.

This mechanism of lung tumor development is mouse specific and is due to preferential and lung-mediated metabolism by CYP 2F2 located in mouse Clara cells. Although CYP 2F4, the isoenzyme in rats, appears to be equally active, it occurs to a much lower extent in rat Clara cells and therefore the levels of metabolites produced are not sufficient to cause lung toxicity. Human lungs contain far fewer Clara cells and thus much lower amounts of CYP 2F1 (the isoenzyme in humans) than rats or mice. Furthermore, human lung microsomes failed to, or only marginally, metabolize these compounds. In addition, human Clara cells differ markedly from mouse Clara cells. These differences make humans much less sensitive than mice to lung toxicity due to these reactive metabolites. Thus, while lung tumors from bronchiolar cell cytotoxicity are theoretically possible in humans, it is unlikely that metabolism by CYP 2F1 would produce levels of cytotoxic metabolites in human lungs sufficient to result in lung cytotoxic responses and thus in tumors (Cruzan et al., 2009).

Vodicka et al. (2006a) reviewed the metabolism and genotoxicity of styrene in relation to its potential carcinogenicity for humans. They conclude that styrene oxide that is formed by more than 95% by metabolic activation contributes quantitatively by far the most to the genotoxicity, while the minor ring oxidation products are potent pulmonary cytotoxins and contribute to local toxicities. The authors point to the importance of metabolic detoxification of styrene oxide that is efficiently carried out by the microsomal epoxide hydrolase. Mechanistic considerations and epidemiological studies have shown that only a very small fraction of the styrene oxide formed from styrene in the metabolically most active organ, the liver, reaches the systemic circulation, whereas the vast majority is immediately enzymatically hydrolyzed. As long as the local concentration of styrene oxide is lower than the local “concentration” of the epoxide hydrolase, detoxification of styrene oxide is practically instantaneous. In liver epoxide hydrolase is highly abundant. Only at higher local styrene oxide concentrations the slow regeneration of free epoxide hydrolase becomes important. In all species, including humans, the lung has lower epoxide hydrolase expression than the liver. However, the styrene activating activity of lung microsomes is very low, especially in humans. In contrast to rodents, detoxification of styrene oxide by glutathione conjugation is of minor importance. The efficiency of this cytosolic detoxification pathway is less than that of epoxide hydrolase.

Overall the author conclude that for equal exposures to styrene steady state concentrations of styrene oxide in blood are highest in mice, lower in rats, and still lower in humans. Based on toxicokinetic differences humans can be predicted with reasonable reliability to be still better protected against styrene induced carcinogenicity than the non-responding rat.

A new literature search up to March 2013 was carried out. The new studies identified supported the proposed mode of action and were related to

               GSH depletion in mice after styrene exposure

               Aromatic hydroxylation

               Relevance of CYP 2F2 in mouse lung

               Other aspects of styrene metabolism

GSH depletion in mice after styrene exposure:

Carlson (2010a) investigated the effect of styrene and of the two enantiomers of styrene oxide (R-SO and S-SO) on glutathione levels in plasma and bronchioalveolar lavage fluid (BAL) of mice after a single intraperitoneal application. Styrene (600 mg/kg bw) caused a significant fall in GSH levels in both BAL and plasma within 3 h. These returned to control levels by 12 h. R-SO (300 mg/kg bw), the more toxic and genotoxic enantiomer, also produced significant decreases of GSH in both body fluids, but S-SO was without marked effect. The author concludes that the reduced GSH concentrations may significantly impair the ability of the lung to buffer oxidative damage. This study supports that styrene and the more toxic isomer R-SO lead to a reversible depression of GSH.

However, these findings are not supported by a 5 -day study in which wild type mice showed a similar degree of lung toxicity after ip injection of R-SO or S-SO (Cruzan et al., 2012).

Richter et al. (2011) developed a histochemical / fluorimetric method to measure GSH in the lung of mice after inhalation exposure (40 and 60 ppm, up to 6 h/d over up to 3 d). GSH was determined in the whole lung and in the bronchial system in parallel to cell proliferation. Exposures to 40 and 160 ppm styrene resulted in a decrease of GSH in lung homogenates during each exposure period (6 h) of 0 - 29% and 31 - 47%, respectively. At both exposures, a rebound of GSH to 110% and 135% of the control concentrations was observed at the beginning of the exposures on days 2 and 3, respectively. There was a gradient of the GSH concentration within the airway epithelium decreasing from the proximal bronchi to the terminal bronchioles. In the latter, cell proliferation was significantly increased at both styrene concentrations. Generally, GSH was reduced in proliferating cells. The authors concluded that the imbalance of the GSH status seems to be a major determinant in styrene-induced cell proliferation in the mouse lung. In summary, this study shows a temporary GSH depletion in the bronchial/bronchiolar epithelium of mice after styrene exposure in parallel with increased cell proliferation.

Aromatic hydroxylation:

Linhart et al. (2010)analyzed urine from mice exposed to styrene vapors (600 and 1200 mg/m3, 6 h) for ring-oxidized metabolites, namely 2-, 3-, and 4-vinylphenol (2-, 3-, and 4-VP), 4-vinylpyrocatechol, and 2-, 3-, and 4-vinylphenylmercapturic acid (2-, 3-, and 4-VPMA). Three isomers, 2-, 3-, and 4-VP, were found and three novel minor urinary metabolites, the arylmercapturic acids 2-, 3-, and 4-VPMA. Excretion of the most abundant isomer, 4-VPMA represented approximately 0.047 and 0.043% of the absorbed dose. In model reactions of styrene 3,4-oxide (3,4-STO) withN-acetylcysteine in aqueous solutions, 4-VP was always the main product, while no 3-VP or mercapturic acids were found. Theinvivoformation of 2- and 3-isomers of both VP and VPMA, neither of which was formed from 3,4-STOin vitro, strongly suggests that another arene oxide, styrene 2,3-oxide, might be a minor metabolic intermediate of styrene. In summary some new urinary metabolites in mice were identified stemming from aromatic hydroxylation. There is an indication that styrene 2,3-oxide is formed as an intermediary metabolite to a minor extent.

Linhart et al. (2012)identified 4-VPMA together with traces of 2- and 3-VPMA in the urine of workers exposed to styrene vapors at concentrations ranging from 23 to 244 mg/m3. The excretion of 4-VPMA accounted for only about 3.5×10−4% of the absorbed dose of styrene. Despite this very low metabolic yield, formation of VPMAs clearly indicates ring oxidation of styrene. In summary, several metabolites derived from aromatic hydroxylation were isolated from urine of exposed workers.

Shenet al. (2010) compared aromatic and side chain metabolism of styrene in mice lung and liver microsomes and compared lung toxicity in mice evoked by some of the metabolites. After incubation of styrene with liver microsomes styrene glycol (SG), 2-, 3-, and 4-vinylphenol (VP) were identified as well as in incubation with lung microsomes. The amounts of these metabolites formed by lung microsomes was less that those obtained by liver microsomes. After large scale incubations, in addition vinyl-1,4-hydroquinone and 4- and 2-hydroxystyrene glycol were found with liver microsomes, and only 2-hydroxystyrene glycol with lung microsomes. The ratios of VPs/SG production were calculated to determine the relative contribution of vinyl/aromatic oxidation: larger VPs/SG ratios were obtained for lung as compared to liver microsomes.Among the phenolic metabolites, 2-VP was formed at a rate 10-fold faster than that of 3-VP and 4-VP in both mouse liver and lung microsomes and the production rate of SG was by a factor of 100 greater than that of 2-VP. Metabolism of styrene was slower in mouse lung than in liver microsomes. Disulfiram and 5P1P inhibited VPs and SG production both in liver and lung microsomes. Importantly, the formation of VPs and SG was almost completely inhibited by 5P1P in lung microsomal incubations. Incubation of VPs with liver microsomes led to a substantial disappearance of all of them and 4-VP was lost the fastest. The following downstream metabolites (vinylcatechols and vinyl-1,4-hydroquinone) were identified: 3-vinylcatechol and vinyl-1,4-hydroquinone from 2-VP; 3-vinylcatechol and 4-vinylcatechol from 3-VP; 4-vinylcatechol from 4-VP. In addition the 3 VPs were metabolized to the corresponding glycols by side chain oxidation. Thus, VPs undergo further aromatic hydroxylation as well as vinyl oxidation. In addition, the toxicity of styrene, SO and of the 3 VPs to the lung of mice (as determined by LDH activity and cell number in bronchioalveolar lavage) was determined after a single ip dose of 100 mg/kg bw. Significant increases of cell number and LDH activity in the lavage fluid were only found after application of 4-VP.

In summanry, this study extends the knowledge of aromatic hydroxylation of styrene by microsomes from mouse lung and liver. New styrene metabolites, namely 2-VP, 3-VP, vinyl-1,4-hydroquinone,and 2-hydroxystyrene glycol were identified in mouse liver microsomal incubations. CYP2F2 and CYP2E1 were found to catalyze the formation of VPs and SO from styrene in mouse liver and lung microsomes. The detected 2-VP, 3-VP, and 4-VP were further oxidized to the corresponding catechols and/or hydroquinone by aromatic hydroxylation and to the corresponding oxides by vinyl epoxidation. 2- and 3-VP were not as toxic as 4-VP to the pulmonary system in mice. However, according to the authors, the potential toxicity of 10-fold higher production of 2-VP than that of 4-VP in mouse liver and lung microsomal incubations of styrene should not be neglected.

Shenet al. (2011) studied oxidative metabolism of styrene in CYP2F2-null mouse liver and lung microsomes.An approximately 50% reduction was observed in the rates of the formation of SO (styrene glycol was monitored) in liver and lung microsomes prepared fromCYP2F2-null mice relative to the wild-type mice. Substantial decreases in the formation of vinylphenols (VP), vinyl-1,4-hydroquinone, and 2-hydroxystyrene glycol were also found in theCYP2F2-null microsomal incubations with styrene. In addition, incubation of 2-, 3-, and 4-VPs with liver and lung microsomes showed a tremendously decreased metabolism of the VPs to the corresponding catechols/hydroquinone and oxides (vinyl epoxides) inCYP2F2-null mouse lung microsomes, compared with that of the wild-type mice. In summary, these results show that CYP2F2 is not only involved in the primary aromatic and side chain metabolism of styrene but also in further downstream toxification of the aromatic hydroxylated metabolites.

Relevance of CYP 2F2 in mouse lung:

Carlson (2012) studied lung and liver metabolism and toxicity in knockout mice. Mice deficient in cytochrome P450 reductase in the liver (that attenuates the total cytochrome P450 activity) but not in other tissues were used as well as CYP2F2 deficient mice. Mice deficient in hepatic cytochrome P450 reductase had much less metabolism of styrene to SO by liver microsomes but not in lung microsomes. Styrene induced hepatotoxicity after 600 mg/kg bw (once ip) was markedly attenuated (as measured by sorbitol dehydrogenase activity) in the deficient mice as was the reduction of glutathione in the liver in comparison to the wild-type mice. Unexpectedly this enzyme deletion also led to some protection against pneumotoxicity (measured by different parameters in bronchioalveolar lavage fluid). According to the author, the protection against pneumotoxicity in hepatic cytochrome P450 reductase deficient mice may need to be clarified by further pharmacokinetic studies. When comparing CYP2F2 knockout mice with the wild-type the deficiency led to a large decrease in metabolism of styrene to SO in lung microsomes but only to a small decrease in liver microsomes. No difference in hepatotoxicity and a decrease in pneumotoxicity were observed in the knockout mice as compared to the wild-type mice. Finally, pneumotoxicity of SO was markedly reduced in CYP2F2 knockout mice supporting a further toxification pathway via aromatic ring hydroxylation by CYP2F2. This study underlines the importance of CYP2F2 for styrene induced lung toxicity. Liver toxicity depends on CYPs located in the liver. There is indication that CYP2F2 lung toxicity is also mediated by downstream metabolites of SO, possibly by aromatic hydroxylation.

Cruzan et al. (2012) studied the toxicity of styrene and SO in CYP2F2(-/-) knock out in comparison to wild-type mice. Mice were treated with styrene (400 mg/kg bw/d by gavage or 200 or 400 mg/kg bw/d ip) or with S- or R-SO (200 mg/kg bw/d ip) for 5 days. Wild-type mice displayed a significant necrosis and exfoliation of Clara cells after treatment with styrene. The cumulative BrdU-labelling index of S-phase cells was markedly increased in the terminal bronchioles after treatment with styrene or S- or R-SO. In contrast, no toxicity to Clara cells or terminal bronchioles was observed in CYP2F2(-/-) knockout mice exposed to styrene or S- or R-SO. This study clearly demonstrates that lung toxicity of styrene and also of SO depends on the metabolism by CYP2F2, most probably by production of ring hydroxylated downstream metabolites. Importantly, the human isoform of CYP2F2, namely CYP2F1, is expressed at much lower levels in humans. Therefore, styrene-induced mouse lung tumors do not predict lung tumor potential in humans. In conclusion, this study shows that lung toxicity of styrene and SO depends on CYP2F2 that is specific for the mouse lung. Most probably lung toxicity is mediated by ring hydroxylated metabolites of styrene and SO.

Zhang et al. (2011) investigated the metabolism of 4-vinylphenol (4VP), the glutathione (GSH) conjugation of the metabolites of 4VP and the cytochrome P450 (CYP) specificity in epoxidation in different microsomes(mouse lung and liver, rat lung, human lung).Aromatic ring cleavage of 4VP, as measured by formation of14C02was negligible. Incubations of 4VP with mouse lung microsomes afforded two major and several minor metabolites. The major metabolites resulted from ring hydroxylation and epoxidation of 4VP to 4VP catechol and the side chain epoxide (4VPO). By trapping with GSH, GSH conjugates of 4VP quinone and 4VPO were obtained. 4VPO was the most abundant metabolite of 4VP and formed two major GSH conjugates at the C-1 position and two minor ones at the C-2 position. The formation of the C-1- and C-2-GSH conjugates was inhibited with specific CYP inhibitors for 2F2 and 2E1 (5P1P and DDTC, respectively). The inhibitor 5P1P had a greater inhibitory effect in the mouse lung than DDTC (85% vs. 57%), suggesting that 2F2 was primarily responsible for 4VPO formation in the mouse lung. It is concluded that 4VP can be metabolized to the electrophilic metabolites, 4VPO and 4VP quinone. These metabolites could can form adducts with nucleophilic macromolecules and this would contribute to the toxicity of 4VP. The formation of GSH conjugates is much higher in rodent microsomal incubation than in those from humans, indicating that the possible toxicity caused by 4VP and styrene in humans will be much lower than in rodents. In summary, this study showed that downstream metabolism of 4-vinyphenol by aromatic and side chain oxidation is mainly effected by CYP2F2 and most prominent in mouse lung microsomes. These toxification pathways are much more effective in rodents than in humans.

Other aspects of styrene metabolism:

Carlson (2010b) studied the metabolism and toxicity of styrene in microsomal epoxide hydrolase-deficient mice (mEH-/-) in comparison to wild-type mice. mEH-/-mice metabolised styrene to SO at the same rate as wild-type mice, but there was minimal metabolism of SO to the glycol. mEH-/-mice were more susceptible to the lethal effects of styrene. After ip application of 200 mg/kg styrene hepatotoxicity (as measured by sorbitol dehydrogenase activity) and pneumotoxicity (by different parameters in bronchioalveolar lavage fluid) was higher in mEH-/-mice as compared to the wild-type. mEH-/-mice were also more susceptible to oxidative stress as indicated by greater decreases in hepatic glutathione levels. On the other hand, SO produced pneumotoxicity was similar in both strains. This was explained by the high exogenous SO dose overwhelming the detoxification capacity of mEH. Since the turnover number for mEH is low it can efficiently detoxify SO as long as it is locally produced in the endoplasmatic reticulum, but exogenously administered SO must be absorbed into the systemic circulation and distributed to the target tissue (cf Vodicka et al., 2006a). Overall, the author concludes that mEH plays an important role in the detoxification of styrene but not exogenously administered SO. This study demonstrates the importance of epoxide hydrolase for detoxification of styrene metabolites (SO) especially as long as SO is produced locally and not administered exogenously.

In the literature update up to Oct. 01, 2015, some investigations with genetically modified mice were identified that are summarized here:

Carlson (2011) exposed mice deficient in glutathione-transferase (P1P2-/-) in comparison to wild-type mice once by the intraperitoneal route to styrene (600 mg/kg), styrene oxide (300 mg/kg), and 4-vinylphenol (100 mg/kg). Hepatoxicity was determined by serum sorbitol dehydrogenase and pneumotoxicity by parameters in lung lavage (proteins, cells, lactate dehydrogenase). Similar changes were observed after application of these 3 substances in both strains of mice for hepato- and pneumotoxicity. It is concluded that either the glutathione pathway contributes little to styrene detoxification or that other isoforms that the P1P2 form are more important. Thus, glutathione detoxification plays a minor role in mice as compaered to epoxide hydrolase (Carlson, 2010b).

Cruzan et al. (2013) demonstrated that styrene and styrene oxide are not lung toxic in CYP2F2(-/-) knockout mice in contrast to the wild type strain indicating that styrene mediated lung tumors in mice are mediated through mouse specific CYP2F2 ring oxidized metabolites in the bronchioles. In addition, a human CYP2F1,2A13,2B6 transgene (humanized CYP2F1) was inserted into the CYP2F2(-/-) mice, and again no indication of cytotoxicity or increased cell proliferation was seen in the lung after treatment with styrene or styrene oxide. In contrast, 4-hydroxystyrene increased cell proliferation in the bronchioles of all three mice strains. In comparison to the genetically modified mice, the highest increase was seen of wild type mice which may result from intrinsic toxicity or from further metabolism. 4-hydroxystyrene toxicity was substantially reduced in the knockout and transgenic mice as compared to the wild type. The test substances were given intraperitoneally over 4 days at doses of 200 mg/kg (for styrene and styrene oxide) and 105 mg/kg (for 4-hydroxystyrene). This data support a mode of action for lung tumor induction in mice by the mouse specific CYP2F2, possibly via 4-hydroxystyrene and not styrene oxide, a pathway that is not operative in humans.

 

Shen et al. (2014) studied formation of styrene metabolites (i.e. styrene oxide and 4-vinylphenol) by lung microsmes and pulmonary toxicity of styrene usingCYP2E1- and CYP2F2-knockout mouse models. A dramatic decrease in the formation of styrene glycol and 4-vinylphenol was found inCYP2F2-null mouse lung microsomes relative to that in the wild-type mouse lung microsomes; however, no significant difference in the production of the styrene metabolites was observed between lung microsomes obtained fromCYP2E1-null and the wild-type mice. The knockout and wild-type mice were treated with styrene (600 mg/kg, ip, once), and cell counts and LDH activity in bronchioalveolar lavagefluids were monitored to evaluate the pulmonary toxicity induced by styrene.CYP2E1-null mice displayed a susceptibility to lung toxicity of styrene similar to that of the wild-type animals; however,CYP2F2-null mice were resistant to styrene-induced pulmonary toxicity. This data show that both CYP2E1 and CYP2F2 are responsible for the metabolic activation of styrene. The latter enzyme plays an important role in styrene-induced pulmonary toxicity.

In summary, these studies strongly support the mouse specific metabolism of styrene responsible for mouse lung tumor formation that may not be extrapolated to humans.