Trichloroethylene

Carcinogenicity

Description of key information

Based on the available data, trichloroethylene is considered a threshold carcinogen. The threshold for renal toxicity is considered an appropriate starting point for the cancer risk characterisation. Therefore, the carcinogenic effects of trichloroethylene are not critical relative to the non-neoplastic effects.

Key value for chemical safety assessment

Carcinogenicity: via oral route

Link to relevant study records

Reference

Endpoint:

carcinogenicity: oral

Type of information:

experimental study

Adequacy of study:

weight of evidence

Reliability:

2 (reliable with restrictions)

Rationale for reliability incl. deficiencies:

other: Non-GLP, non-guideline study. Only one treatment dose level.

GLP compliance:

no

Species:

mouse

Strain:

B6C3F1

Sex:

male/female

Details on test animals or test system and environmental conditions:

The mice used in this study were from Harlan Industries (Indianapolis, IN), were quarantined for 2.5 weeks prior to the start of the test, and were 8 weeks of age when placed on study. The animals were assigned to cages according to a table of random numbers and then to groups according to another table of random numbers. They were fed Purina@ Rodent Chow 5001 ad libitum (Distributed by O.K. Feed Store, Miami, FL). Bedding provided was Sani Chip@ hardwood, Pinewood Products Co. (Distributed by O.K. Feed Store, Miami, FL). Tap water was provided via Edstrom Automatic Watering System (Waterford, WI). Polycarbonate cages were used (Lab Products, Rochelle Park, NJ) and were changed and sanitized twice weekly; racks sanitized every 2 weeks. 10 mice per cage for the first 8 months and then 5 per cage thereafter. Cage filters used were Cerex spun nylon (Florida Filters, Miami, FL) and changed every 2 weeks. The animal room environmental conditions were: 22-24C; 40%-60% relative humidity; 12 hours fluorescent light per day; room air changed 10-15 per hour.

Route of administration:

oral: gavage

Vehicle:

corn oil

Details on exposure:

The mice were dosed with 0, or 1,000 mg trichloroethylene/kg body weight in corn oil by gavage. 0.5 ml dose vol.

Analytical verification of doses or concentrations:

not specified

Duration of treatment / exposure:

103 weeks

Frequency of treatment:

5 days/week

Post exposure period:

until spontaneous death

Remarks:

Doses / Concentrations:
0 or 1000 mg/kg
Basis:
actual ingested
target concentrations

No. of animals per sex per dose:

50 males and 50 females of each strain.

Control animals:

yes

Details on study design:

The animals were allowed to live until spontaneous death.

Positive control:

None

Observations and examinations performed and frequency:

All animals were observed twice daily, and clinical signs were recorded once per week. Body weights by cage were recorded once per week for the first 12-15 weeks of the studies and once per month thereafter. Mean body weights were calculated for each group.

Sacrifice and pathology:

Animals found moribund and those surviving to the'end of the studies were humanely killed. A necropsy was performed on all animals including those found dead, unless they were excessively autolyzed or cannibalized, missexed, or found missing. Thus, the number of animals from which particular organs or tissues were examined microscopically varies and is not necessarily equal to the number of animals that were placed on study.
During necropsy, all organs and tissues were examined for grossly visible lesions. Tissues were preserved in 10% neutral buffered formalin, embedded in paraffin, sectioned, and stained with hematoxylin and eosin. Histologic exam performed on all animals; tissues examined include: gross lesions and tissue masses, skin, mesenteric lymph nodes, mammary gland, salivary glands, thigh muscle, lungs and mainstem bronchi, heart, thyroid gland, parathyroids, esophagus, stomach, duodenum, ileum, colon, liver, vertebrae with bone marrow, thymus, larynx, trachea, pancreas, spleen, kidneys, adrenal glands, urinary bladder, brain, pituitary gland, spinal cord, eyes, seminal vesicle/prostate/testes or ovaries/uterus.

Other examinations:

Not specified.

Statistics:

Probabilities of survival were estimated by the product-limit procedure of Kaplan and Meier (1958). Animals were statistically censored as of the time that they died of other than natural causes or were found to be missing. Statistical analyses for a possible dose-related effect on survival used the method of Cox (1972) for testing two groups for equality and Tarone’s (1975) extensions of Cox’s methods for testing for a dose-related trend. The incidence of neoplastic or nonneoplastic lesions has been given as the ratio of the number of animals bearing such lesions at a specific anatomic site to the number of animals in which that site was examined. For the statistical analysis of tumor incidence data, two different methods of adjusting for intercurrent mortality were employed. Each used the classical methods for combining contingency tables developed by Mantel and Haenszel (1959). Tests of significance included pairwise comparisons of high and low dose groups with controls and tests for overall dose-response trends. The first method of analysis assumed that all tumors of a given type observed in animals dying before the end of the study were "fatal”. This method of adjusting for intercurrent mortality is the life table method of Cox (1972) and of Tarone (1975). The second method of analysis assumed that all tumors of a given type observed in animals dying before the end of the study were “accidental". See Peto et al., 1980, for the computational details of both methods. In addition to these tests, one other set of statistical analyses was carried out and reported in the tables analyzing primary tumors: the Fisher's exact test for pairwise comparisons and the Cochran-Armitage linear trend test for dose- response trends (Armitage, 1971; Gart et al., 1979). These tests were based on the overall pro- portion of tumor-bearing animals. All reported P values are one-sided.

Clinical signs:

effects observed, treatment-related

Description (incidence and severity):

Male survival rate was reduced in comparison with the controls (probability of survival in the treated group was roughly half that of the control group).

Mortality:

mortality observed, treatment-related

Description (incidence):

Male survival rate was reduced in comparison with the controls (probability of survival in the treated group was roughly half that of the control group).

Body weight and weight changes:

effects observed, treatment-related

Description (incidence and severity):

Male bodyweight was reduced in the treated group, to 90% of the vehicle control value.

Food consumption and compound intake (if feeding study):

not examined

Food efficiency:

not examined

Water consumption and compound intake (if drinking water study):

not examined

Ophthalmological findings:

not examined

Haematological findings:

not examined

Clinical biochemistry findings:

not examined

Urinalysis findings:

not examined

Behaviour (functional findings):

not examined

Organ weight findings including organ / body weight ratios:

not examined

Gross pathological findings:

no effects observed

Histopathological findings: non-neoplastic:

effects observed, treatment-related

Description (incidence and severity):

Renal cytomegaly was observed in most treated animals of both sexes, although the lesions were not as severe as those seen in the rats.

Histopathological findings: neoplastic:

effects observed, treatment-related

Dose descriptor:

NOAEL

Effect level:

1 000 mg/kg bw/day

Sex:

male

Basis for effect level:

other: see 'Remark'

Remarks on result:

not determinable

Remarks:

no NOAEL identified. Effect type:carcinogenicity (migrated information)

Dose descriptor:

NOAEL

Effect level:

1 000 mg/kg bw/day

Sex:

female

Basis for effect level:

other: see 'Remark'

Remarks on result:

not determinable

Remarks:

no NOAEL identified. Effect type:carcinogenicity (migrated information)

Endpoint conclusion

Endpoint conclusion:

adverse effect observed

Dose descriptor:

LOAEL

1 000 mg/kg bw/day

Study duration:

chronic

Species:

mouse

Quality of whole database:

acceptable

Carcinogenicity: via inhalation route

Link to relevant study records

Reference

Endpoint:

carcinogenicity: inhalation

Type of information:

experimental study

Adequacy of study:

weight of evidence

Reliability:

2 (reliable with restrictions)

Rationale for reliability incl. deficiencies:

other: see 'Remark'

Remarks:

Non-GLP, non-guideline study, however well documented. Some deviations from the current guideline, including the late start in the dosing of the animals, and the continuation of the study to the natural death of the animals rather than terminatio at a defined time point such as 110 weeks as per the guideline

Principles of method if other than guideline:

Although the project was started in 1976, and most of the experiments were performed from the beginning of 1979, the methodological protocol was considered acceptable.

GLP compliance:

no

Species:

rat

Strain:

Sprague-Dawley

Sex:

male/female

Details on test animals or test system and environmental conditions:

Sprague-Dawly rats were of the breed routinely employed in the Bentivoglio Laboratory. Male and female Sprague-Dawley rats, 12-13 weeks old at the start of the experiment. The room temperature varied from 19-22 degrees and was checked 3 times daily. Animals were fed an adequate commercial diet and recieved water, ad libitum.

Route of administration:

inhalation

Type of inhalation exposure (if applicable):

whole body

Vehicle:

unchanged (no vehicle)

Details on exposure:

The chambers for inhalation exposure are made of stainless steel (210 x 206 x 200 cm, the volume of each chamber was 8652 liters) , for exposure of 10 animals simulaneously. Continuous air flow provided 12-15 air changes per hour. Before its introduction the air was filtered, and the chamber arrangement was such that air flowed from the top of the chamber to its bottom without recirculation. The internal pressure was about 1 mm Hg less than that of the room where the chamber was situated to avoid any possible contamination of the outside environment. Lighting was provided by room light. Exposure chambers were equipped with 5 fixed-point matrixes for checking the distribution of the test substance. The fallout from the chambers was decontaminated before being dispersed into the external atmosphere to avoid general pollution and, as far as the experiments are concerned, to avoid any remote possibility of reintroducing into exposure chambers air with any trace of the test compound.

Analytical verification of doses or concentrations:

yes

Details on analytical verification of doses or concentrations:

The concentrations in air were checked by continuous gas-chromatographic monitoring.

Duration of treatment / exposure:

8 weeks (experiment 302) or 104 weeks (experiment 304 and 304 bis)

Frequency of treatment:

5 days/week, 7 hours/day

Post exposure period:

until spontaneous death

Remarks:

Doses / Concentrations:
8 week exposure (experiment 302): 100, 600 ppm; 104 week exposure (experiment 304 and 304bis): 100, 300, 600 ppm
Basis:
other: target concentrations

No. of animals per sex per dose:

8 week exposure (experiment 302): 60-90 of each sex; 104 week exposure (experiment 304 and 304bis): at least 130 of each sex

Control animals:

yes

Details on study design:

The animals were allowed to live until spontaneous death.

Positive control:

None

Observations and examinations performed and frequency:

The status and behavior of the animals were examined at least three times daily. Every two weeks, the animals were submitted to a clinical examination for the detection of the gross changes, which were registeredd in the experimental records. The animals were weighed every two weeks during the treatment period and then every eight weeks.

Sacrifice and pathology:

A complete necropsy was performed on each animal. All of the different parts of the body were explored, including the central nervous system. Specimens for histology included: skin, mammary gland, subcutaneous lymph nodes, brain, pituitary gland, Zymbal glands, salivary glands, Harderian glands, eyeballs, thyroid, tongue, thymus and mediastinal lymph nodes, larynx, lungs, heart, aorta, esophagus, diaphragm, liver, kidneys, adrenals, spleen, pancreas, mesenteric lymph nodes, stomach, various segments of intestine (3 levels), urinary bladder, uterus, ovaries, seminal vesicles, prostate gland, testes and epididymes, right thigh muscle, interscapular brown fat, bone marrow (femur) and any other organ or tissue with gross pathological lesions.
The histological specimens were fixed in 70% ethyl alcohol. Once fixed, they were trimmed in a highly standardized way. A higher number of samples was taken when particular pathological lesions were seen. Sections were routinely stained with hematoxilin-eosin, and, when necessary, with other techniques. The bone marrow smears were stained with May-Grunwald-Giemsa and with the Papanicolaou technique. All slides were screened by a junior pathologist, and then reviewed by a senior pathologist.

Other examinations:

Not specified.

Statistics:

When necessary, data from the experiments were submitted to statistical analysis. The following statistical methods are routinely employed:
- Analysis of variance is used for the statistical evaluation of body weights
- For different survival rates the Log rank test has been used
- The non-neoplastic, pre-neoplastic and neoplastic lesions were evaluated by using the Chi-square or Fishers exact test
- The effect of different doses is evaluated by using the Cochran-Armitage test for linear trends in proportions and frequencies.

Clinical signs:

no effects observed

Mortality:

no mortality observed

Body weight and weight changes:

no effects observed

Food consumption and compound intake (if feeding study):

not specified

Food efficiency:

not specified

Water consumption and compound intake (if drinking water study):

not specified

Ophthalmological findings:

not specified

Haematological findings:

not specified

Clinical biochemistry findings:

not specified

Urinalysis findings:

not specified

Behaviour (functional findings):

not specified

Organ weight findings including organ / body weight ratios:

not specified

Gross pathological findings:

no effects observed

Histopathological findings: non-neoplastic:

effects observed, treatment-related

Histopathological findings: neoplastic:

effects observed, treatment-related

Details on results:

Evidence of general toxicity was limited to the observation of kidney tubule meganucleocytosis in rats exposed to concentrations of 300 and 600 ppm for 104 weeks. 
A slightly higher, no dose-related incidence of leukemias was observed in treated rats, mainly at 600 and 100 ppm (not significant). A dose-response related increase in testis Leydig cell tumors was observed in treated groups, the incidences being 31/130 at the highest dose (p<0.01), 30/130 at the mid-dose (p<0.01), 16/130 in the low-dose group (p<0.05), and 6/135 in the controls.
Five kidney adenocarcinomas (4 in males and 1 in females) were observed in the highest treatment group (600 ppm).
Since the occurence of Leydig cell tumors could not be reproduced in other studies, kidney tumors were considered critical and have been taken into account in the risk assessment.

After 8 weeks exposure, there were no clear differences in the incidence of tumours between the exposed and control groups. For rats receiving long-term exposure, the overall incidence of tumours was not affected by treatment. There was, however, a dose-related and statistically significant increase in the incidence of Leydig cell tumours in rats in all trichloroethylene exposed groups; the numbers of affected animals were 6/135 (4%), 16/130 (12%), 30/135 (22%) and 31/130 (24%) in the control, 100, 300 and 600 ppm groups, respectively. Also, four males and one female from the highest exposure group had kidney tubular adenocarcinomas, a tumour type not previously seen at the testing laboratory in Sprague-Dawley rats.

To summarise the findings of this study, increased incidences of Leydig cell tumours and, in association with other renal pathological changes, renal tubular adenocarcinomas in rats were observed in epoxide-free trichloroethylene treated groups. The increased incidence of Leydig cell tumours was not reproduced in other studies and has consequently been disregarded.

Dose descriptor:

NOAEL

Effect level:

300 ppm

Sex:

male/female

Basis for effect level:

other: 104 week exposure (experiment 304) Renal tubular adenomas. It should be noted that general toxicity was observed from 300 ppm onwards (kidney tubule meganucleocytosis).

Remarks on result:

other: Effect type: carcinogenicity (migrated information)

Dose descriptor:

NOAEL

Effect level:

600 ppm

Sex:

male/female

Basis for effect level:

other: 8 week exposure (experiment 302): No effect on the incidence of expected tumors and no unexpected type of tumor observed.

Remarks on result:

other: Effect type: carcinogenicity (migrated information)

Conclusions:

Limited evidence of kidney cancer in rats. However the differences between the study design and the ciurrent guideline bring the findings into question, particularly since the first observed kidney tumour occurred after a standard guideline study woudl have been terminated.

Endpoint conclusion

Endpoint conclusion:

adverse effect observed

Dose descriptor:

NOAEC

1 641 mg/m³

Study duration:

chronic

Species:

rat

Quality of whole database:

acceptable - see discussion

Carcinogenicity: via dermal route

Endpoint conclusion

Endpoint conclusion:

no study available

Justification for classification or non-classification

In rats, following high dose exposure (inhalation or oral gavage) kidney tumours (adenocarcinomas predominantly) were observed. As indicated the incidence of these tumours was very low, had a late-onset, and was accompanied by nephrotoxicity. Tumours were not observed at dose levels not associated with nephrotoxicity. It is therefore possible that in spite of other potential modes of action, the cytotoxicity observed in the kidney is a contributing factor in the ultimate development of renal cell carcinoma. The epidemiology data indicate that exposure to trichloroethylene, particularly at high exposure levels (>50 to 100ppm) is associated with a slight increased risk of kidney cancer. The evidence is however far from clear cut. There have been several large cohort studies in the US and Northern Europe that failed to identify an association between trichloroethylene exposure and risk of renal cell carcinoma. However some smaller studies, including case control studies did identify an association with odds rations typically falling in between 1.5 and 2.8 and varying in statistical significance. As a consequence of the available epidemiological data, animal carcinogenicity data and mode of action hypotheses a recent review by IARC upgraded the cancer classification of trichloroethylene from 2A (probable human carcinogen) to category 1, known human carcinogen.

With respect to the overall conclusions regarding classification for cancer in Europe, although there is still uncertainty about the mode-of-action, relevant human tumour types, and epidemiological cancer evidence in humans, based on the available animal and human cancer data the EC’s Working Group on the Classification and Labelling of Dangerous Substances decided to classify trichloroethylene for carcinogenicity. Therefore, trichloroethylene is classified for carcinogenicity (according to Annex I of Directive 67/548/EEC: Carc. Cat. 2; R45; according to EU Classification, Labelling and Packaging of Substances and Mixtures (CLP) Regulation (EC) No. 1272/2008: Carc. 1B; H350).

Additional information

Animal Data

The carcinogenicity of trichloroethylene has been investigated in several cancer bioassays via the oral and inhalation routes and using several strains of mice, rats and other species such as hamsters. A persistent issue with the oral studies using rats is that due to the high doses used, there was a high mortality rate in the top and, in some cases, the middle dose levels. This significantly reduced the ability of the oral rat studies to assess the carcinogenic potential of trichloroethylene, and as a consequence most of the rat studies are considered inadequate for assessing the carcinogenicity of trichloroethylene. The available inhalation studies typically had lower mortality rates, but in several cases were of limited value in predicting carcinogenicity due to the number of animals used in each dose group. The guideline requires a minimum of 50 animals per sex per dose, whereas some studies used only 30 animals per sex, per dose group or less. More importantly, in Maltoni et al (1986) [regarded by some to be the key study for kidney tumours in rats] although animal numbers were sufficient, the study deviates in several key areas from an OECD guideline compliant carcinogenicity study, with the onset of the observed kidney adenocarcinoma findings occurring significantly beyond the indicated termination time point of a guideline compliant GLP study. Thus this study cannot be considered to be clear evidence of kidney carcinogenicity in rats, and consequently should not be the critical study for risk assessment.

In addition to these issues, some of the studies tested trichloroethylene stabilized using epichlorhydrin and/or epoxy butane, both genotoxic and both known to be carcinogenic in animals. Although in the past some commercial grades of trichloroethylene did use these as stabilizers (at low levels, <0.1 ppm), current grades use non-genotoxic, non-carcinogenic stabilizers such as diisopropanolamine. Therefore carcinogenicity studies performed using grades of trichloroethylene stabilized using epichlorhydrin or epoxybutane may over estimate the carcinogenic potential of current trichloroethylene grades. This is discussed below where appropriate.

Oral Studies:

Mice

TCE (purity > 99%; containing 0.19% epoxybutane and 0.09% epichlorohydrin as stabilizers) was administered in corn oil to 6-week-old B6C3F1 mice (50/sex/group in treated groups, 20/sex in vehicle control group) by gavage on 5 days/week for 78 weeks, giving time weighted average TCE doses of 0, 1169 and 2339 mg/kg bw/day for males and 0, 869 and 1739 mg/kg bw per day for females. Mice were killed 90 weeks after treatment started and a “complete” necropsy and histopathological evaluation was undertaken. Survival figures were 8/20, 36/50 and 22/48 in the control, low-dose and high-dose males; the equivalent figures for females were 20/20, 42/50 and 39/47. The survival-adjusted (Cox and Tarone test) incidences of hepatocellular carcinomas were increased in animals of each sex in relation to dose; males: l/20 in controls, 26/50 (p = 0.004) at the low dose, 31/48 (p < 0.001) at the high dose; females: 0/20 in controls, 4/50 at the low dose, 11/47 (p = 0.008) at the high dose. One male at the high dose developed a forestomach papilloma (NCI, 1976).

In a subsequent study, TCE (purity > 99.9%; 8 ppm amine, no epichlorohydrin) was administered in corn oil to 8-week-old B6C3FI mice (50/sex/group) at 0 or 1000 mg/kg bw/day by gavage on 5 days/week for up to 103 weeks. Survival of treated males was significantly reduced (p = 0.004); 33 control and 16 treated males and 32 control and 23 treated females were alive at the end of the study. Treated mice had increased incidences (incidental tumour test) of liver tumours. In males, incidences in control and treated groups were, respectively, 7/48 and 14/50 (p = 0.048) for hepatocellular adenomas; 8/48 and 31/50 (p < 0.001) for hepatocellular carcinomas; and 14/48 and 39/50 (p < 0.001) for combined hepatocellular adenomas and/or carcinomas. The equivalent figures in females were 4/48 and 16/49 (p = 0.001) [adenomas], 2/48 and 13/49 (p = 0.002) [carcinomas], and 6/48 and 22/49 (p < 0.001) [adenoma/carcinoma combined]. No significant increases in tumour incidence were found at other tissue sites. Toxic nephrosis (cytomegaly) was seen in 90% of treated males and in 98% of treated females (NTP, 1990).

In a third study, the potential effect of epoxide stabilisers on the carcinogenicity of TCE was assessed. ICR:Ha Swiss mice (50/sex/group) were treated with purified TCE (purity > 99.9%; 0.0015% triethanolamine; no epoxy stabilisers), industrial grade TCE (purity 99.4%; 0.11% epichlorohydrin and 0.20% 1,2-epoxybutane), purified TCE with 0.8% added epichlorohydrin, purified TCE with 0.8% added epoxybutane, or purified TCE with 0.25% added of each stabiliser. The TCE grades were given initially at 1.8 and 2.4 g/kg bw/day (females and males respectively) by gavage in corn oil, on 5 days/week. Due to severe toxicity, dosing was interrupted for several weeks and all doses were reduced by one-half from week 40, up until 18 months, which was followed by a 6-month observation period without further treatment. Survival was reduced in males of all treated groups and females in the groups given purified TCE and purified TCE plus epichlorohydrin. Forestomach tumours were seen, but only in the groups given TCE stabilised with epoxides, which were thus held responsible for these local tumours. Hepatocellular tumours (adenomas and carcinomas combined) occurred in 3/50 controls, 6/50 treated with purified TCE and 9/50 treated with industrial grade TCE however there was no statistically significant difference to the controls reported by the investigators (Henschler et al. 1984).

Although there are various methodological issues with NCI (1976) and Henschler et al. (1984) studies (dosing duration, high mortality, interruptions in dosing), the oral studies in mice demonstrate that high oral gavage doses of trichloroethylene produces a treatment related increase in the incidence of liver tumours. These studies failed to identify any additional tumour sites that were associated with TCE exposure. In particular there was no evidence that lung, kidney or hematopoietic tumour incidence was increased. It is also noteworthy that kidney toxicity was not reported in the NCI (1976) and Henschler et al. (1984) studies even those these used higher doses that were more than double that used in the NTP 1990 study.

Rats

In an NCI study TCE (purity > 99%; 0.19% epoxybutane and 0.09% epichlorohydrin as stabilizers) was given to Osborne-Mendel rats (50/sex/treated group; 20/sex vehicle controls; age 6 weeks) in corn oil at time-weighted average TCE doses of 0, 549 or 1097 mg/kg bw/day, by gavage, on 5 days/week for 78 weeks. Rats were killed after 110 weeks and underwent a “complete” necropsy. Survival was low in all groups (3/20, 8/50 and 3/50 control, low- and high-dose males; 8/20, 13/48 and 13/50 control, low- and high-dose females). There were no significant differences in tumour incidence at any tissue site (NCI, 1976) however due to the low survival rate and reduced dosing duration (compared to a current guideline study) this study is considered to be inadequate to assess the carcinogenicity of TCE in rats.

There are two NTP studies (1988 and 1990) where the carcinogenicity of TCE was assessed in multiple strains of rats. In the NTP (1988) study, TCE (purity > 99.9%; 8 ppm amine, no epoxide) was given at 0, 500 or 1000 mg/kg bw/day, on 5 days/week for 103 weeks, by gavage in corn oil to rats of four strains (ACI, August, Marshall and Osborne-Mendel; 6.5-8 weeks of age). Additional groups (50/sex/strain) served as untreated controls. Untreated and vehicle-treated controls showed no kidney toxicity, conversely the incidence of renal cytomegaly was > 80% in all treated groups (both sexes), and toxic nephropathy (described as dilated tubules lined by elongated and flattened epithelial cells) occurred at rates of 17-80% in the treated groups. In the Osborne-Mendel rats, relative to control, survival was unaffected in the treated males (at 18, 22, 17 and 14 of the untreated control, vehicle controls, low-dose and high-dose groups) but reduced in the treated females (19, 18, 10 and 7 survivors in these groups, respectively). The low-dose males showed increased incidences of renal tubular-cell hyperplasia and tubular-cell adenoma (hyperplasia: 0/50, 0/50, 5/50 and 3/50 in untreated control, vehicle control, low-dose and high-dose groups; adenoma: 0/50, 0/50, 6/50 (p = 0.007; survival-adjusted incidental tumour test) and 1/50, respectively). One renal tubular cell adenocarcinoma occurred in a high-dose male Osborne-Mendel rat. In the Marshall rats, survival was reduced in treated groups (males: 32, 26, 12 and 6 survived in the untreated control, vehicle control, low- and high-dose groups; females: 31, 30, 12 and 10, respectively). The incidences of interstitial-cell tumours of the testis were increased in TCE-exposed Marshall rats: 16/46, 17/46, 21/48 (p < 0.001; survival-adjusted incidental tumour test) and 32/48 (p < 0.001) in the untreated control, vehicle control, low- and high-dose groups, respectively. No significant increases in tumour incidence were reported for ACI or August rats, but survival was generally poor (NTP, 1988).

In the NTP (1990) study TCE (purity > 99.9%; 8 ppm amine, no epichlorohydrin) was given at 0, 500 or 1000 mg/kg bw/day in corn oil by gavage on 5 days/week for up to 103 weeks to Fischer 344/N rats (50/sex/group; 8 weeks of age). Survival of treated males was reduced (p < 0.005). At the end of the study, there were 35, 20 and 16 male survivors in the control, low- and high-dose groups; equivalent numbers for females were 37, 33 and 26. An increased incidence of renal tubular-cell adenocarcinomas was seen in males: 0/49 untreated controls, 0/48 vehicle controls, 0/49 low dose and 3/49 high dose (p = 0.028; incidental tumour test). Two males at the low dose had renal tubular-cell adenomas. There was no evidence of carcinogenicity in the females. Toxic nephrosis of the kidney occurred in 96/98 treated males and in all of the treated females but not in vehicle control rats of either sex (NTP, 1990).

Both these studies identified kidney toxicity and an increase in the incidence of kidney tumours. Although the incidence of kidney tumours in these studies was very low, they were considered to be treatment related findings due to the rarity of the tumours in control animals. However due to the high rate of mortality, the high degree of treatment associated toxicity (particularly in the kidney) and deficiencies in the conduct of the studies (specifically the NTP 1988 study), it was concluded that they are inadequate to assess the carcinogenicity (presence OR absence) of TCE. As such the weight assigned to these studies when considering the overall carcinogenicity profile of TCE is reduced, particularly for tumour types not observed in other, better conducted studies.

In a smaller study of short duration, Sprague- Dawley rats (30/sex/group; aged 12-13 weeks), were treated by gavage with with TCE (purity 99.9%; no epoxide) at 0, 50 or 250 mg/kg bw/day in olive oil, 4-5 days/week, for 52 weeks followed by observation for life (Maltoni et al. 1986). Survival data were not provided, but the authors reported a non-significant increase in mortality among treated females. Renal tubular-cell cytokaryomegaly was observed only in male rats at the high dose (14/30; p < 0.01). A non-significant increase in the incidence of leukaemia was observed in males: (0/30, 2/30 and 3/30 in controls, low-dose and high-dose groups, respectively. Although this study was not afflicted by the same high mortality rate as other oral gavage studies, it is still limited by the short dose period and low number of animals per dose group and as such cannot be used to conclusively assess the carcinogenicity potential in rats. In addition, the study was performed at the Ramazzini Institute in Italy, and there is some question about the quality of studies performed during the 1970s and 1980s, particularly with regard to the confounding of the correct identification of leukemias due to respiratory infections that were very common in the study animals during this period.

Inhalation studies

Mice

There was evidence of a treatment-related increase in lung tumours in a study where ICR mice (49-50 females/group; 7 weeks of age) were exposed to air containing TCE (purity 99.8%; 0.13% carbon tetrachloride, 0.02% benzene and 0.019% epichlorohydrin) at 0, 50, 150 or 450 ppm (0, 270, 810 or 2430 mg/m3), 7 hours/day, 5 days/week, for up to 104 weeks (Fukuda et al. 1983). Survival was unaffected and a “complete” necropsy was carried out on all animals. Histopathological evaluation revealed a significant increase (Fisher's exact test) in the incidence of lung adenocarcinomas: 1/49, 3/50, 8/50 (p < 0.05) and 7/46 (p < 0.05) in the control, low-, mid- and high-dose groups, respectively. IARC found a significant dose-response trend (p = 0.034) when applying a Cochran-Mantel-Haenszel test. The mean number of lung tumours/mouse was increased in the mid- and high-dose groups (0.12, 0.10, 0.46 and 0.39 in control, low-, mid- and high-dose groups). However, the incidences of combined lung adenomas and adenocarcinomas at the mid (13/50) and high dose (11/46) were not significantly increased (6/49 in controls). Although this study used epoxide stabilised TCE and only included females, the increase in lung tumours is consistent with that identified in other studies using both sexes and epoxide-free TCE. Therefore the results can be considered indicative of carcinogenic potential in the mouse lung.

A dose-related increase in lung tumour incidence was also reported in females when B6C3F1 mice (90/sex/group; aged 12 weeks) were exposed to TCE (purity 99.9%; no epoxide) at 0, 100, 300 or 600 ppm (0, 540, 1620 or 3240 mg/m3) in air, 7 hours/day, 5 days/week for 78 weeks and then observed for life (Maltoni et al. 1986, 1988). Survival data were not provided, but mortality was reported to be higher (p < 0.05) in the treated males. Lung tumour incidence figures in females were: 4/90, 6/90, 7/90 and 15/90 (p < 0.05 for high-dose group) in control, low-, mid- and high-dose groups, respectively. There was also a slight increase in hepatoma incidence at the high dose; this increase was statistically significant when males and female were analysed together (combined incidences were 2, 3, 4 and 8% in the control, low-, mid- and high-dose groups, respectively).

A similar study in Swiss mice gave evidence of lung and liver cancer activity although, in this strain, the males were the susceptible sex. The mice (90/sex/group; 11 weeks old), inhaled TCE (purity, 99.9%; no epoxide) at 0, 100, 300 or 600 ppm (0, 540, 1620 or 3240 mg/m3), 7 hours/day, 5 days/week for 78 weeks and were then observed for life (Maltoni et al. 1986, 1988). Survival data were not provided. Dose-related increases in lung and liver tumour incidences were observed in males (Fisher's exact test or Cochran-Armitage linear trend test). Lung tumour incidences in males were 10/90, 11/90, 23/90 (p < 0.05) and 27/90 (p < 0.01) in the control, low-, mid- and high-dose groups, respectively. The equivalent figures for liver adenoma and carcinoma combined in the males were 4/90, 2/90, 8/90 and 13/90 (p < 0.05 for high-dose group), respectively.

In a limited study, NMRI mice (30/sex/group; age unspecified) were exposed to TCE (purity> 99.9%; 0.0015% triethanolamine; epoxide-free) at 0, 100 or 500 ppm (0, 540 or 2700 mg/m3) in air, 6 hours/day, 5 days/week for 18 months, followed by a period without exposure until 30 months (Henschler et al. 1980). At 18 months, treatment had not affected survival of females but, in males, survival was reduced from 83% in controls to 63% in low-dose and 56% in high-dose groups. Histopathological examination of spleen, liver, kidney, lung, heart, stomach, central nervous system and all tumours indicated statistically significant (p = 0.01 or better) increases in age-adjusted incidences of lymphomas in treated females: 9/29, 17/30 and 18/28 in controls, low-dose and high-dose groups, respectively. However, the NMRI strain has a high background incidence of lymphomas and as such the investigators concluded that the assay could not inform whether the lymphomas were related to TCE treatment.

Rats

An increase in testicular tumours and a marginal increase in kidney tumours were reported when Sprague-Dawley rats (130-145/sex/group; 12 weeks of age), were exposed to TCE (purity 99.9%; no epoxide) at 0, 100, 300 or 600 ppm (0, 540, 1620 or 3240 mg/m3) in air, 7 hours/day, 5 days/week for 104 weeks, followed by observation for lifetime (Maltoni et al. 1986, 1988). Survival data were not provided. There was a significant, dose-related increase in the incidence of Leydig cell (interstitial) tumours of the testis [p < 0.001; Cochran-Mantel-Haenszel test]. The percentages [incidences] of male rats bearing these tumours were 4.4% [6/135], 12.3% [16/130; p < 0.05; Fisher's exact test], 23.1% [30/130; p < 0.01; Fisher's exact test] and 23.8% [31/130; p < 0.01; Fisher's exact test] in the control, low-, mid- and high-dose groups, respectively. The occurrence of Leydig cell tumors have not been reproduced in other studies and therefore this finding was considered non-critical. Renal tubular adenocarcinomas were seen in four (3.1%) high-dose male rats, compared with none in the lower dose groups, in controls or in the historical control database for Sprague-Dawley rats at the study laboratory. Cytokaryomegaly of renal tubular cells was observed at the mid- and high-dose (in 17 and 78% of the male rats, respectively), but not in control or low-dose rats. Due to the prominence of the findings of this study in various hazard assessments further comments on this study are given separately below.

Gross and histopathological examination revealed no evidence of carcinogenic activity when Sprague-Dawley rats (49-51 females/group; aged 7 weeks) were exposed to TCE (purity 99.8%; 0.13% carbon tetrachloride, 0.02% benzene and 0.019% epichlorohydrin) at 0, 50, 150 or 450 ppm (0, 270, 810 or 2430 mg/m3) in air, 7 hours/day, 5 days/week for 104 weeks (Fukuda et al. 1983). At 100 weeks, about 50% of controls were alive compared with about 75% of the rats in the treated groups.

There was also no evidence of carcinogenic activity in a limited study where Wistar rats (30/sex/group; age unspecified) inhaled TCE (purity > 99.9%; 0.0015% triethanolamine; epoxide-free) at 0, 100 or 500 ppm (0, 540 or 2700 mg/m3) in air, 6 hours/day, 5 days/week for 18 months, with study termination after 36 months (Henschler et al. 1980). No differences in survival were reported (47, 23 and 37% of males, and 17, 13 and 17% of females in the control, low- and high-dose groups respectively). Histopathological and gross examination was carried out for the spleen, liver, kidneys, lungs, heart, stomach, central nervous system and “all tumours”.

Hamsters

Survival was unaffected and histopathological examination of spleen, liver, kidneys, lungs, heart, stomach, central nervous system and “all tumours” revealed no evidence of carcinogenic activity when Syrian hamsters (30/sex/group; age unspecified) were exposed to TCE (purity > 99.9%; 0.0015% triethanolamine; epoxide-free) at 0, 100 or 500 ppm (0, 540 or 2700 mg/m3) in air, 6 hours/day, 5 days/week for 18 months, with study termination at 30 months (Henschler et al. 1980).

Dermal Studies:

Mice

In a very limited study, 1 mg TCE was applied to the skin of 30 female ICR:Ha Swiss mice (6-8 weeks old), three times/week for 83 weeks (Van Duuren et al. 1979). No tumours were observed at the application site.

Summary of the animal carcinogenicity data

The carcinogenicity of trichloroethylene has been investigated in a number of long-term animal studies, using the oral and inhalation routes, and involving hamsters and a variety of strains of rat and mouse. These studies provide clear evidence that trichloroethylene is carcinogenic in mice and an indication that it is carcinogenic in rats.

Mouse summary

In the mouse, trichloroethylene induced hepatocellular tumours by either the inhalation or oral routes. This effect was seen in Swiss and B6C3F1 strains, but not in NMRI or Ha:ICR strains. The hepatocellular tumours were observed at high oral dose levels, of 1,000 mg/kg/day and above (lower exposure levels were not investigated), and by inhalation at 600 ppm, but not at 300 ppm. Also, by the inhalation route only, an increased incidence of lung adenomas or adenocarcinomas was observed in three strains of mice, namely the ICR, Swiss and B6C3F1 strains. These lung tumours occurred at exposure levels as low as 150 ppm in one study but not at 100 ppm in another.

With respect to the tumours observed in mice, in the review of the animal carcinogenicity data and potential modes of action, the EU risk assessment concluded that the mouse liver and lung tumours were unlikely to be of significance to humans. In both cases the conclusion was based on species differences in metabolism, and lack of sensitivity to the effects of the metabolites. 

Liver tumours:

Within the EU risk assessment the mechanism was considered likely to be trichloroacetic acid induced peroxisome proliferation. Humans and rats do not metabolise trichloroethylene to trichloroacetic acid (TCA) to the same extent that mice do, and human hepatocytes are resistant to peroxisome proliferation induced by trichloroactic acid (EU RAR, 2004). Since the EU risk assessment, the EU Scientific Committee on Occupational Exposure Limits review indicated that the mechanism of action is potentially more complex than TCA induced peroxisome proliferation alone. Given the complexity of the metabolism of trichloroethylene including the differences between species it is likely that there may be several parallel ‘key events’ occurring in mode of action leading to tumour formation in the mouse liver and it may be possible that genotoxicity is also playing some role. However given the fact that there are more similarities in the oxidative metabolism of trichloroethylene between rats and humans (EU RAR) versus mice it seems likely that the absence of hepatic tumours in rats would tend to support the conclusion that the mice hepatic tumours are of less relevance to humans.

It is noted that there is some evidence of an increase in liver tumour incidence in humans exposed to trichloroethylene (refer to epidemiology section). However, the increases are small, typically non-significant, and therefore the likelihood that the association with trichloroethylene exposure is due to chance rather than a causal relationship cannot be excluded (refer to epidemiology summary).

When developing the OEL for trichloroethylene, SCOEL concluded that there was no clear evidence for TCE-induced human liver tumours from epidemiology and that in mice, genotoxicity is unlikely to be the determining mode of action for tumours in the liver. Subsequently, it was assumed that protection from the risk for kidney toxicity would equally protect from the risk for liver tumours. This is based on a similar cytotoxic potency of TCE in the liver and the kidney and on the assumption that a strongly sub-linear dose response relationship would be the appropriate and still conservative risk extrapolation method for both tumour locations (SCOEL 2010)

Lung tumours:

The EU risk assessment concluded that there is evidence that the lung tumours seen in mice are related to the conversion of trichloroethylene to chloral hydrate in the Clara cells resulting in cytotoxicity and repeatedcycles of cell destruction and replication leading to tumour formation. Lung tumours were not seen in rats, which have a lower capacity than mice to metabolise trichloroethylene to chloral hydrate. Human lung tissue appears to have minimal capacity to metabolise trichloroethylene to chloral hydrate suggesting that the lung tumours seen in mice are of no relevance to humans. Given the absence of any observed association of trichloroethylene with lung tumour incidence in humans in the many epidemiology studies, SCOEL also came to the conclusion that the lung tumours observed in mice were of minimal significance to humans.

Rat summary

In the rat, trichloroethylene induced renal tubular adenomas or adenocarcinomas, albeit at a low incidence, in association with other pathological changes. This effect was observed consistently across a number of studies although the majority of the studies (particularly the oral gavage studies) were considered inadequate for assessing the presence or absence of cancer in rats. Given the various problems with the rat cancer database for TCE the assessment of carcinogenicity is not as clear cut as insinuated in previous assessments. Of the available studies, one study has been consistently considered as a key study in several risk assessments including the EU risk assessment (EU 2004); the Maltoni et al (1986) inhalation study in mice and rats. The study was considered to be well performed and it identified the presence of renal tumours that had previously been noted in oral gavage studies performed by NCI and the NTP groups. However, irrespective of its prominence in previous risk assessments, this study is not without its flaws that would bring into serious question the findings and their relevance to the human risk assessment. In this study, Sprague-Dawley rats were exposed by inhalation to epoxide-free trichloroethylene for 104 weeks and then monitored until spontaneous death. For rats receiving long-term exposure, the overall incidence of tumours was not affected by treatment.

Evidence of treatment-related general toxicity was limited to the finding of dose-related renal tubule meganucleocytosis observed in male rats exposed to concentrations of 300 (23% incidence) or 600 ppm (82% incidence) trichloroethylene. The average latency time for the detection of this finding at necropsy was 114.5 and 107.1 weeks after the initiation of treatment for the 300 and 600 ppm groups, respectively. This latency corresponds to 126.5 and 119.1 weeks of age for the 300 and 600 ppm groups, respectively. This male-specific finding was similarly observed in Sprague Dawley rats orally treated with 250 mg/kg epoxide-free trichloroethylene (Maltoni, 1986) and therefore was considered a clear effect of trichloroethylene exposure.

Four kidney adenocarcinomas (3 in males (3.5%) and 1 in females (1.1%)) were observed in the highest treatment group (600 ppm). The adenocarcinomas in males could be considered related to the concurrent kidney toxicity in this sex, however no other kidney findings were detected in females. The average latency time for the detection of the adenocarcinomas in 600 ppm males at necropsy was 115.7 weeks after the initiation of treatment, corresponding to a latency period of 127.7 weeks of age. The authors of this study considered the evidence of this effect borderline (Table 55 of the study report). However they report in a separate publication (Maltoni et al., 1988) that due to the very low incidence of renal tumours in their control population and the fact that similar tumours were observed in the oral gavage studies performed by the NTP (contrary to the oral gavage study performed by their own lab, Maltoni et al., 1986), the kidney tumors observed following inhalation exposure were treatment related findings. To assess the validity of this determination, a further examination was made into the conduct of this non-GLP study and the conditions under which the kidney adenocarcinomas were detected.

According to the OECD draft test guideline 451 for the conduct of carcinogenicity studies, carcinogenicity studies should initiate in rats no later than eight weeks of age and the duration of the study should normally be 24 months. The Maltoni 1986 inhalation study deviated from both of these requirements, initiating with animals twelve weeks of age and continuing longer than 148 weeks of age (or 136 week after initiation of treatment) until all animals had spontaneously died. Furthermore, the OECD draft guidance document (No 116) indicates that termination of the study should be considered when the number of survivors in the lower dose groups or the control group falls below 25 percent. This timepoint in the Maltoni 1986 inhalation study would have corresponded to 124 weeks of age (or 112 weeks after the initiation of treatment). Importantly, if the Maltoni 1996 inhalation study were conducted in an OECD GLP guideline compliant manner it is likely that the low incidence finding of kidney adenocarcinomas would not have been observed as the study would have been terminated well before the average time to onset of the effect (Table 1).

Study Type                            Duration of study              Kidney adenocarcinomas (Y/N)

OECD TG 451                     102 weeks (24 months)        NA

Maltoni, 1986 Inhalation        136 weeks                           Y (Avg onset = 115.7 weeks)

Maltoni, 1986 Oral Gavage   140 weeks                           N

To summarize, the authors of the Maltoni 1986 rat inhalation carcinogenicity study considered the low incidence finding of renal tubular adenocarcinomas to be borderline evidence of a treatment-related effect, and subsequently this finding was considered by various risk assessments as clear evidence of trichloroethylene induced kidney carcinogenicity. However, further investigation into this study suggests that it deviates in several key areas from an OECD guideline compliant carcinogenicity study, with the onset of the observed kidney adenocarcinoma findings occurring significantly beyond the indicated termination time point of a guideline compliant GLP study. Thus this study cannot be considered to be clear evidence of kidney carcinogenicity in rats, and consequently should not be the critical study for risk assessment.

Weight of evidence for rat kidney carcinogenesis

Considering the weight of evidence for kidney carcinogenicity in rats; overall, the in vivo genotoxicity studies demonstrate the absence of genotoxicity of TCE, of particular note is the lack of genotoxicity in well conducted comet assays (of the kidney) and transgenic mouse assay. TCE is clearly nephrotoxic with the observation of meganucleocytosis, renal tubular-cell cytokaryomegaly and toxic nephrosis in several oral and inhalation studies in the rat. Although renal tumours have been observed, the incidence is very low, the latency period is long (in some cases longer than the OECD standard guideline for carcinogenicity), not every strain of rat demonstrated tumour formation, and in almost all cases the tumours were associated with doses that were clearly toxic to the rats with survival rates far lower than controls. As such it is questionable whether these findings should be considered as clear evidence of trichloroethylene induced kidney carcinogenicity, and the very late onset of tumours (in the presence of kidney toxicity) coupled with the lack of clear evidence of genotoxicity in vivo raises some concern about attributing the formation of tumours to a genotoxic mode of action that could be relevant to humans. In spite of the equivocal nature of the weight of evidence in animals, several epidemiological investigations have identified a weak association between occupational exposure to TCE and development of kidney tumours. As with the animal database, the human database is also far from clear, with several large cohort studies failing to identify any causal association between TCE and kidney cancer(refer to Epidemiology summary). As a consequence of the available human and animal data, several reviews of the animal carcinogenicity of TCE (EU risk assessment, SCOEL, EPA IRIS review, IARC) have concluded that kidney tumours in rats are relevant treatment related effects, and are relevant to humans and should therefore be considered in the risk assessment.

Mode of action and threshold for rat kidney carcinogenesis

Mode of Action Summary

Multiple reviews of the human and animal data investigating the toxicity and carcinogenicity of TCE (EU risk assessment, SCOEL, EPA IRIS review, IARC) have concluded that kidney tumours in rats are relevant treatment related effects (see Cancer Summary). In addition, those reviews have concluded that supplementary mode of action information indicates kidney effects should be regarded as relevant to humans and therefore be considered in the risk assessment. However, based on interpretation of mode of action information, there are differing conclusions regarding the dose-dependent nature of TCE renal toxicity and potential for human cancer.

The likelihood of high dose-specific kidney toxicity of TCE has been particularly emphasized in the SCOEL (2009) assessment, which has concluded that TCE is a “genotoxic carcinogen, for which a practical threshold is supported by studies on mechanisms and/or toxicokinetics.” The SCOEL justified its conclusion in significant part based on mode of action information addressing the primary hypothesis that TCE-induced renal toxicity in rodents and humans results from generation of glutathione (GSH)-conjugate derived metabolite(s) known to be weakin vitro, but notin vivogenotoxicants, as well as only weakly active renal cell cytotoxicant(s). The observation that both animal and human kidney toxicity and potentially human kidney cancer (see Epidemiology and Cancer Summaries) have only been reported following high TCE doses or exposures was consistent with mode of action data indicating that the GSH-mediated metabolic pathway largely functions only at exposure doses saturating oxidative metabolism of TCE, the major metabolic pathway for TCE operating at lower exposures of approximately 50 ppm. Such lower exposures have not been identified as renal toxic or tumourigenic in multiple chronic rodent bioassays (see Cancer Summary). SCOEL states that although weak GSH-conjugate mediated genotoxicity may contribute to high-dose specific tumourigenicity, “a linear extrapolation of kidney tumour risks should be limited to clearly nephrotoxic concentrations.” Thus, absent doses sufficient to result in renal cell cytotoxicity and subsequent reparative cell proliferation, it is anticipated that the weakly genotoxicity of TCE GSH-conjugate metabolites are without toxicological or human cancer consequence.  

The overall routes of TCE metabolism have been well characterized, and have provided key insights into the postulated mode of action of kidney toxicity (Lock and Reed, 2006; Figure 1a, b). The oxidative pathway of TCE metabolism resulting in formation of TCE-epoxide, chloral, trichloroethanol, trichloroacetic acid (TCA), dichloroacetic acid (DCA) and other terminal metabolites is the primary route of TCE metabolism. Both TCE and DCA have been postulated as mediating the mode of action of mouse liver tumour responses by their ability to activate the PPARα receptor with resulting peroxisome proliferation. This mode of action is not regarded as relevant to human cancer risk (SCOEL, 2009; see Cancer Summary). A secondary route of TCE metabolism involves conjugation with GSH primarily in liver via action of a GSH-transferase, resulting in formation of isomeric S-(1,2-dichlorovinyl)glutathione and S-(2,2-dichlorovinyl)glutathione (DCVG). Once formed in liver, DCVG is either systemically delivered to kidney (humans) or excreted into bile (rodents). DCVG is metabolized by enzymes of the mercapturic acid pathway either in kidney or the biliary tract to S-(1,2-dichlorovinyl)-cysteine and S-(2,2-dichlorovinyl)-cysteine (DCVC). DCVC formed in rodent bile is reabsorbed via enterohepatic circulation and distributed to kidney. While DCVG itself is weakly cytotoxic, mutagenic in bacteria and nephrotoxic in rodents, the primary renal toxicity of GSH-conjugate derived metabolites is postulated to be mediated through subsequent metabolism of DCVC by renal-specific cysteine conjugate β-lyase to putative reactive and renal cytotoxic metabolites chlorothioketene and/or chlorothionoacetyl chloride (Lock and Reed, 2006; Figure 1b). DCVC may also be metabolized (detoxified) by N-acetylation to N-acetyl-DCVC). 

In contrast to the SCOEL analysis which emphasized the high-dose specific and weakly active cytotoxic and genotoxic activity of GSH-conjugate derived metabolites in modulating potential human kidney toxicity and cancer, other reviews such as the EPA IRIS have postulated that the renal toxicity and tumourigenicity potential of TCE extends to far lower TCE doses and exposures. This differing position is foundationally based on reliance on an analytical method for detection of GSH-conjugate pathway activity which provides data indicating that this pathway is highly active in both rodent and humans, and further, predicts significant of generation of GSH-derived toxic metabolites even at low exposures. Importantly, however, the indirect and non-specific analytical method relied upon by EPA and others to quantify activity of the GSH metabolism pathway has been significantly challenged. More technologically advanced and metabolite specific alternative analytical methods for detection of GSH-derived metabolites result in dramatically different estimates of the functional activity of the GSH pathway in both rodents and humans (Lock and Reed, 2006; Dekant, 2010). Thus, a primary issue associated with the differing conclusions regarding the functionality of GSH-conjugate TCE metabolism lies not in biology, but rather largely rests on differing analytical methods used to measure formation of DCVG or DCVC as the source of downstream toxic metabolites generated by the GSH pathway. The quantitative impact of the differing analytical methods for informing the importance of GSH-derived metabolites as dose-dependent drivers of kidney toxicity is highly significant to understanding the mode of action of kidney effects, and results in estimates of GSH-conjugate metabolite formation that vary by at least three to four orders of magnitude.

The differing outputs of the analytical methods are attributable to the specificity of the respective methods to detect DCVG and DCVC. Analytical findings relied upon by EPA IRIS used a complex and indirect method, the “Reed-Method”, to measure DCVG and DCVC formation. That method involves a complex multi-step procedure in which DCVC is derivatized with iodoacetamide and chlorodinitrobenzene and then subjected to ion exchange chromatography with UV detection of the dinitrophenyl chromophore. However, this method has been clearly shown to be subject to chromatographic separation and column performance challenges that are further compounded by fact that the derivatization process is not specific for DCVG or DCVC (Dekant, 2010). Given the lack of specificity of the Reed-Method and its vulnerability to chromatographic performance issues, it is far more appropriate to rely on analytical results generated from more technological advanced methods capable of specifically measuring DCVG/DCVC formation. Such methods have indeed been employed, and involve use of¹⁴C-TCE coupled to HPLC with radiochemical detection and LC/MS or GC/MS confirmation of metabolite identity. The impact of the differing analytical methods on estimation of GSH pathway activity can be readily visualized by comparing the rates of DCVC formation determined in variousin vitroliver and kidney cytosol and microsomal fractions of rat, mouse and humans (Table 1, from Dekant, 2010). Formation rates of DCVC determined with DCVC-specific detection methods found extremely low activity in rodent liver and approximately 10-fold lower levels in human liver; formation of DCVC was not detected in rat kidney, the primary target of concern in rats. In contrast, the indirect derivatization method reported far higher rates (> 3-4 orders of magnitude) of DCVC formation in both liver and kidney rodent fractions, and importantly also found that both human liver and kidney samples had even higher rates of DCVC formation relative to rat or mouse. These rate differentials and their impact on formation of downstream putative toxic metabolites are further magnified by data indicating that DCVC flux through the β-lyase pathway is one to two orders of magnitude less than that through the N-acetylation detoxification pathway, suggesting that low rates of DCVC formation determined with DCVC-specific analytical methods are likely to be preferentially funneled through the N-acetylation detoxification pathway. The significantly lower rate of DCVC formation using DCVC-specific analytical methodology also is consistent observations that isolated GSH S-transferases have a very low capability for metabolizing TCE to DCVC (Dekant, 2010).

The dramatic impact of the differing analytical methods on estimation of DCVG and DCVC formation also extends toin vivostudies in rodents and humans. Using the Reed-Method, rats administered a single oral dose of TCE up 1,971 mg/kg had blood DCVG concentrations of up to 100 nM that were neither dose nor time dependent (Lashet al., 2006), a finding inconsistent with a primary role in target-organ toxicity. The finding of significant DCVG formation in rats was distinctly different from those reported by Kimet al.(2009) in acute mouse studies. Using HPLC-ESI-MS/MS measurement of GSH pathway metabolites, these investigators found that blood concentrations of GSH conjugate metabolites were at least 3,600-fold below those of oxidative metabolites following a single 2,100 mg/kg oral dose of TCE, and the total systemic AUC for TCA and DCA oxidative metabolites was 40,000-fold higher than that of DCVG and DCVC. These dramatically lower findings of GSH-conjugate metabolite formation are despite the observation that mice generate DCVC at slightly higher rates than rats and greater than 10-fold higher in humans. In addition, the Lashet al.findings also were inconsistent with HPLC GS/MS analyses of biliary excretion of DCVG in rats orally dosed with 2,200 mg/kg and in which total DCVG collected in bile-cannulated animals over seven hours post TCE dosing accounted for less that 0.01% of the administered dose, even though excretion into bile represents the primary route of hepatic clearance of TCE GSH-conjugate metabolite (Dekantet al., 1990; Dekant, 2010). Finally, similar large analytical-dependent differences in GSH metabolite quantification also were apparent in human studies. Using the derivatization method, peak DCVG blood concentrations of 46 µM and 13 µM were found in volunteers 2 hours after the start of a 100 ppm inhalation exposure. These concentrations were dramatically higher than the approximate 40 nM peak blood concentrations detected in mice given single very high oral dose of 2,100 mg/kg (Kimet al., 2009) and also are counter toin vitroevidence that that human liver and kidney GSH-conjugate metabolism is very low (Table 1). These results again serve to illustrate the significant impact of non-specific versus specific analytical methods on estimation of GSH-conjugate pathway function and its implications as a dose-dependent driver of kidney toxicity and tumours.

Overall, a weight-of-evidence evaluation of the postulated GSH-conjugate mediated mode of action of TCE induced rodent and human kidney toxicity must include particular attention to the significant limitations and impact of DCVG and DCVC measurements using the non-specific Reed-Method versus those conducted by state-of-art HPLC MS-based methods. When such considerations are factored into a weight-of-evidence evaluation, the mode of action information is fully consistent with the SCOEL (2009) determination that exposures to TCE below the SCOEL or DNEL value (see DNEL summary) are unlikely to result in kidney toxicity or tumours. Importantly, such an analysis supports the SCOEL position that despite the weak genotoxic and tumourigenic activity of TCE, the possibility of human kidney cancer is not expected below the threshold exposure represented by the SCOEL/DNEL value.

References:

Dekant, W, Koob, M, and Henschler, D (1990a). Metabolism of trichloroethene--in vivo and in vitro evidence for activation by glutathione conjugation. Chem Biol Interact 73: 89-101.

Dekant, W. (2010). In “Toxicological Review of Trichloroethylene In Support of the IRIS Database (draft of October 2009); Comments of the Halogenated Solvents Industry Alliance, Inc., submitted by Paul Dugard to EPA IRIS.

Kim, S, Kim, D, Pollack, GM, Collins, LB, and Rusyn, I (2009). Pharmacokinetic analysis of trichloroethylene metabolism in male B6C3F1 mice: Formation and disposition of trichloroacetic acid, dichloroacetic acid, S-(1,2-dichlorovinyl)glutathione and S-(1,2-dichlorovinyl)-L-cysteine. Toxicol Appl Pharmacol 238: 90-99.

Lash, LH, Putt, DA, and Parker, JC (2006). Metabolism and tissue distribution of orally administered trichloroethylene in male and female rats: identification of glutathione- and cytochrome P-450-derived metabolites in liver, kidney, blood, and urine. J Toxicol Environ Health A 69: 1285-1309.

Lock, EA, and Reed, CJ (2006). Trichloroethylene: Mechanisms of Renal Toxicity and Renal Cancer and Relevance to Risk Assessment. Toxicol Sci.91: 313-331.

SCOEL/SUM/142 (April 2009). Recommendation from the Scientific Committee on Occupational Exposure Limits for Trichloroethylene.

.

Justification for selection of carcinogenicity via oral route endpoint:
only reliable and well conducted oral cancer study in rodents.

Justification for selection of carcinogenicity via inhalation route endpoint:
adequate study - although did not confirm to current guideline and was not performed according to GLP. Findings in the rat are equivocal but considered to be treatment related by most assessors. Supporting study in mice also relevant for this endpoint, with lung and liver tumours identified.

Carcinogenicity: via oral route (target organ): digestive: liver

Carcinogenicity: via inhalation route (target organ): digestive: liver; respiratory: lung; urogenital: kidneys