Distillates (petroleum), sweetened middle

Carcinogenicity

Description of key information

A straight run hydrotreated gas oil (a member of this category) gave a low tumour incidence of 12%, with latency exceeding 18 months. This finding together with the high skin irritancy potential of the material on repeated contact would be in keeping with tumour induction by a non-genotoxic mechanism. Further support for a non-genotoxic mode of action for "other" gas oils comes from the observation that when the straight run hydrotreated sample was diluted to 50% and 25% with highly refined mineral oil and applied at a frequency that ensured the total amount applied per week was the same as with the neat oil, skin irritancy was reduced and the test material failed to produce a significant number of tumours.
However, because other gas oils might contain cracked components, a worst-case scenario approach is taken. Catalytically cracked light cycle oil with 8.7% 3 to 7 ring polycyclic aromatic compounds was carcinogenic, but the 50% dilution in mineral oil applied 4 times a week produced more tumours than the compound applied neat twice a week. The evidence suggests that cracked gas oils are carcinogenic and act via a genotoxic mechanism, and if other gas oils have an unknown refining history that might include cracked components, other gas oils must be classified as carcinogens.

Justification for classification or non-classification

Other Gas Oils have the potential to produce tumours via both genotoxic and non-genotoxic mechanisms. Utilising a worst case approach, Other Gas Oils should be classsfied as Category 1B carcinogens (H350) according to EU CLP Regulation (EC No. 1272/2008) based on the information available on cracked gas oils. However, Note N has been assigned to this category in Annex VI of the CLP Regulation and the classification as a carcinogen need not apply if the full refining history is known and it can be shown that the substance from which it is produced is not carcinogenic.

Additional information

The carcinogenicity of gas oils have been investigated in a number of different studies. The American Petroleum Instiute (API) reported on skin cancer studies on crude oil fractions which showed that atmospheric distillates boiling between 49-371°C (light and middle distillates) could produce skin tumours in mice after repeated application (King R.W. et al, 1984; Lewis S.C. et al, 1984). Some of the fractions tested contained polycyclic aromatic comounds (PACs) and it was thought that these constitutents might have been the cause of the tumour development, acting via a genotoxic mechanism. Other fractions however e.g. straight-run kerosine and naphtha, had either very low or undetectable amounts of PACs and it was considered unlikely that a genotoxic mechanism was responsible for the tumour development with these samples. Severe skin irritation was noted in the studies reported by the API and it was thought that this was likely to have influenced skin tumour development.  Additional studies have been conducted on gas oils,, but in most of the studies interpretation of the results has been confounded by the extent of skin irritation that has occurred during the studies. Therefore studies have been carried out to investigate the influence of skin irritation on the development of tumours in mouse skin painting studies. The results of these studies have been summarised below.

 

In order to help elucidate the interaction between PAC content, skin irritation and skin carcinogenicity, CONCAWE initiated a middle distillate research program described below.

 

The middle distillate programme was undertaken to investigate the role of skin irritation in tumour development during long term dermal studies in mice. The programme was conducted in three phases.

 

Phase 1 involved screening studies on 10 middle distillates. These included the determination of physico-chemical properties and PAC (polycyclic aromatic compound) concentrations, short-term in vitro tests mainly investigating mutagenicity and detailed histopathological examination of mouse skin following repeated dermal application of the distillates. These studies suggest that mutagenic activity is related to PAC content. In particular, relatively high mutagenic activity was evident in a cracked gas oil sample, whereas, all of the kerosine samples which contained low levels of PACs were not mutagenic. Further it was shown that the middle distillates were irritant when applied repeatedly to mouse skin.

 

In phase 2, five middle distillates were evaluated in dermal studies of up to 90 days duration in mice to establish dilution conditions that would enable repeated application of middle distillates without causing significant skin irritation. It was found that use of a mineral oil diluent enabled daily application of a ‘standard’ amount of middle distillates with no more than slight irritation. Experiments with radiolabelled hydrocarbons showed that skin penetration was not reduced under these conditions. A regime was devised which enabled animals to be tested under conditions producing different degrees of skin irritation but at the same total weekly dose of middle distillate over the course of the study.

 

Once the preliminary investigations had been conducted to determine the physico-chemical properties of the materials and the doses that could be used while avoiding irritation, three middle distillates were evaluated in a two year mouse dermal carcinogenicity study (Dally, 1996). Selected for the investigation were a straight run kerosine, and two gas oils: a straight run hydrotreated gas oil (from Other Gas Oils) (CAS No. 64742-46-7; MD6) and a catalytically cracked light cycle oil (from Cracked Gas Oils) (CAS No. 64741-59-9; MD7). These were tested alongside a negative control (highly refined, non-irritating mineral oil) and a positive control treatment (heavy clarified oil (HCO), 5% in mineral oil). For the study, groups of 50 male C3H/HeNCrlBR VAF/Plus® mice were administered various test (28.5%, 50%, or 100%) or control materials applied regularly to clipped dorsal skin. The treatment regime established in phase 2 was designed to deliver a constant total weekly dose of each distillate while varying the extent of any local skin irritation by manipulating test material concentration and frequency of application.

In these studies: undiluted straight run kerosine and straight run gas oil caused marked irritation and a number of tumours at the treatment site.The magnitude of response (ranked as slight, moderate or marked on a numerical scale) was negligible-slight in the negative controls (mean 0.06) and also for animals treated with 28.5% or 50% MD6 (0.02 or 0.09, respectively), with moderate irritation present after application of 100% MD6 group (score 2.0). In contrast, concentration-dependent increases in dermal irritation were observed in all of the MD7 treatment groups (0.28, 1.59, 2.40).However, the cracked gas oil (LCO) produced tumours at all three application rates. Undiluted LCO produced marked skin irritation whereas there was moderate irritation with the 50% dilution and slight irritation with the 28.5% dilution. It is concluded that in long-term dermal studies in mice, cracked gas oils containing significant concentrations of 3 to 7 fused ring PACs probably produce skin tumours by a genotoxic mechanism, independent of irritation. Under the same conditions, undiluted straight-run hydrotreated gas oils and kerosines containing undetectable or low concentrations of PACs eventually produce tumours, but only when moderate to severe skin irritation is also present. These tumours are probably the consequence of a continuous cycle of cell damage and repair caused by chronic skin irritation.

 

For the straight run kerosine, skin tumours only developed in the group of animals in which substantial skin irritation occurred during the study. Since no PACs were detected in the straight run kerosine it is concluded that the occurrence of tumours is likely to have been caused by a non-genotoxic mechanism. This conclusion is consistent with reports by others (Ingram, A.J. and Grasso, P., 1991) that lighter middle distillates are tumour promotors but not initiators and furthermore that skin irritation plays an important role in skin tumour development.

 

For the straight run hydrotreated gas oil(MD6), which had a low polycyclic aromatic compound content (1.3%), skin tumours only developed in the animals in which substantial skin irritation occurred during the study. This material caused a significant increase in the number of tumours only in the presence of skin irritation, and it is thus concluded that the mechanism is likely to be non-genotoxic. Again, this finding is consistent with results reported previously (Jungen, H. et al, 1993) that gas oils which contain undetectable or low levels of PACs may be tumour promotors but do not possess tumour initiating activity. This is further supported by the results of the supporting IP study. This contrasts with the positive control material, diluted HCO, which had a high polycyclic aromatic compound content and where a relatively short time to tumour formation was recorded.

 

The catalytically cracked light cycle oil LCO (MD7) contained measurable quantities of PACs and caused the development of skin tumours even at dose levels which did not cause appreciable skin irritation. For this material it is concluded that tumour development was due to a genotoxic mechanism. It was surprising that fewer tumours developed in the group of mice receiving the highest concentration of cracked gas oil. However, this group also showed the most severe skin irritation and it is possible that undiluted cracked gas oil caused a sufficient degree of cytotoxicity to inhibit cell repair and possibly epidermal necrosis destroying initiated cells therefore inhibiting skin tumour formation.

 

It is concluded that frequent application of middle distillates that contain undetectable or low levels of PACs to mouse skin will produce skin tumours if accompanied by moderate to marked skin irritation. The mechanism is not considered to involve a direct genotoxic process, but rather may result from frequent cell damage and repair. When irritation does not occur, these same middle distillates do not give rise to tumours.  

There were no significant treatment-related gross findings at necropsy other than skin irritation and masses (tumours) at the treatment site. Benign and/or malignant skin tumours were diagnosed at the application site in several of the groups, and were frequent in the positive control group where a total of 164 tumours (88 benign, 73 malignant) developed in 47 mice with 217 days to first tumour. This contrasts with the negative control group which was tumour free. Tumours (predominately papillomas or squamous cell carcinomas) occurred in all groups treated with MD7, affecting 1, 17 and 7 mice from the high, intermediate or low dose groups respectively with a time to first tumour of approximately 300-650 days. In contrast, tumours were relatively infrequent in mice treated with MD6 with a total of 6 tumours in 4 mice treated with undiluted distillate, and a single squamous cell carcinoma in the low dose group.

The findings from the above study are further supported by Nessel et al 1999a who state that middle distillate fuels containing limited amounts of PNAs result in tumor formation secondary to chronic skin irritation via a nongenotoxic mechanism (Nessel, 1999a). 

The overall data available on gas oils presented in the dossier(API, 1989; Broddle et al., 1996; Dally et al., 1996; Nessel et al., 1998) indicate that they produce severe dermal irritation when applied repeatedly to the skin in neat form; observations made during most of the studies reportedly draw attention to this. Histological examination confirms epidermal necrosis, ulceration and dermal damage with inflammatory changes are often seen together with marked acanthosis (epidermal hyperplasia). While evidence suggests that dilution with a volatile organic solvent does not reduce skin irritation appreciably, the irritancy properties of gas oils and other middle distillates can be reduced substantially by dilution with highly refined mineral oil. It is evident that the level of skin irritation produced by neat gas oils would probably be sufficient to produce skin tumours in mice by a non-genotoxic mechanism, even in the absence of significant amounts of genotoxic constituents. However in order to demonstrate that a non-genotoxic process was responsible for skin tumour induction, it is essential to show that the gas oils concerned do not contain significant levels of genotoxic constituents and produce a tumour response consistent with that of a non-genotoxic substance (e. g. a tumour incidence of less than 25% with a latent period in excess of 1 year).

 Although the majority of the skin painting studies described above include only a limited evaluation of the internal organs, others (including investigations performed using samples containing cracked stocks) were more comprehensive, with histological evaluation of a range of tissues for neoplastic effects. Although not conforming to current guidelines, the results from these investigations provide no evidence for a carcinogenic effect at sites distal to the skin. Furthermore, polycyclic aromatic hydrocarbons are known to cause tumors at the site of contact .

In deriving an overview of the carcinogenicity of the "other" gas oils it is important to consider carcinogenicity studies on gas oil samples together with information on their irritancy potential, and available evidence on polycyclic aromatic hydrocarbon levels, genotoxicity data and other mechanistic studies (e. g. initiation/promotion testing; modified Ames test). Information available on the group "other" gas oils is limited. However, it can be concluded that gas oils of this type are generally severely irritating to the skin following repeated skin contact but show little, if any genotoxic potential and hence may give rise to skin tumours by a non-genotoxic mechanism. From this it can be concluded that tumour induction in mice with these gas oils probably requires a threshold level of skin irritancy and as a consequence is unlikely to be relevant to humans. Data from cracked gas oils indicate that if an other gas oils contain cracked components, they are likely to be carcinogenic. Since Other Gas Oils can be derived from various pre-cursor streams, it is very important to know the refining history of the streams.