Guidelines for Canadian Drinking Water Quality: Guideline Technical Document - Haloacetic Acids

10.0 Health effects in laboratory animals and in vitro test systems

10.3 Long-term exposure and carcinogenicity

Long-term studies with MCA failed to produce tumours in rodents. In contrast, DCA, TCA and DBA produced liver tumours in laboratory animals. However, these liver tumours were seen only in mice, and not in rats, exposed to TCA and DBA, whereas both rats and mice were affected when exposed to DCA. However in rats treated with DBA, tumours in other organs were detected. No adequate long-term/carcinogenicity studies on MBA were identified. The results of these carcinogenicity studies relating to liver tumours are summarized in Table 9, while the details of these studies are described in the sections below.

Table 9: Summary table: carcinogenicity studies relating to liver tumours for MCA, DCA and TCA^a

HAA	Dosing route	Doses (mg/kg bw per day)	Duration	Acid or salt	Strain/ species	Sex	Liver tumours found	Study author
a A: acid; DW: drinking water; F: female; Gav: gavage; HA = hepatocellular adenomas; HC = hepatocellular carcinomas; M: male; S: salt; ? = form unknown. b Although no liver tumours were found, tumours in other organs were detected.
MCA	Gav	0, 15, 30	2 years	A	F344 rats	M & F	None	NTP, 1992
	Gav	0, 50, 100	2 years	A	B6C3F1 mice	M & F	None	NTP, 1992
	DW	0, 3.5, 26, 59.9	2 years	S	F344 rats	M	None	DeAngelo et al., 1997
DCA	DW	0, 140, 280	52 weeks	S	B6C3F1 mice	M	HA, HC	Bull et al., 1990
	DW	0, 7.6, 77, 410, 486	60-75 weeks	?b	B6C3F1 mice	M	HA, HC	DeAngelo et al., 1991; U.S. EPA, 1991
	DW	0, 93	104 weeks	S	B6C3F1 mice	M	HA, HC	Daniel et al., 1992
	DW	0, 8, 84, 168, 315, 429	90-100 weeks	S	B6C3F1 mice	M	HA, HC	DeAngelo et al., 1999
	DW	0, 40, 120, 330	360 or 576 days	S	B6C3F1 mice	F	HA, HC	Pereira, 1996
	DW	0, 3.6, 40.2, 139.1	100 weeks	S	F344 rats	M	HA, HC	DeAngelo et al., 1996
TCA	DW	0, 178, 319	Up to 52 weeks	S	B6C3F1 mice	M	HA, HC	Bull et al., 1990
	DW	0, 7, 71, 595	60 weeks	possible S	B6C3F1 mice	M	HA, HC	DeAngelo and Daniel, 1990; U.S. EPA, 1991
	DW	0, 64, 212, 640	360 or 576 days	S	B6C3F1 mice	F	HA, HC	Pereira, 1995, 1996; Pereira and Phelps, 1996
	DW	0, 71, 583	104 weeks	?b	B6C3F1 mice	F	HA, HC	U.S. EPA, 1991
	DW	0, 3.6, 32.5, 364	104 weeks	S	F344 rats	M	None	DeAngelo and Daniel, 1992; DeAngelo et al., 1997
DBA	DW	0, 4, 35, 65	104 weeks	S	B6C3F1 mice	F	HA, HC	Melnick et al. 2007; NTP, 2007
	DW	0, 4, 45, 87	104 weeks	S	B6C3F1 mice	M	HA, HC	Melnick et al. 2007; NTP, 2007
	DW	0, 2, 25 45	104 weeks	S	F344 rats	F	Noneb	Melnick et al. 2007; NTP, 2007
	DW	0, 2, 20, 40	104 weeks	S	F344 rats	M	Noneb	Melnick et al. 2007; NTP, 2007

10.3.1 Monochloroacetic acid

No tumours were observed in two different strains of mice (18 per sex per strain) gavaged with MCA (acid) in distilled water at 46 mg/kg bw per day from the age of 7 days to 4 weeks, after which MCA was mixed directly in the diet at 149 mg/kg for a total of 18 months (Innes et al., 1969).

In a 2-year carcinogenicity bioassay, F344/N rats (70 per sex per dose) were dosed with MCA (acid) in deionized water at 0, 15 or 30 mg/kg bw per day by gavage, 5 days per week (NTP, 1992). The mean body weight of high-dose male rats was reduced after 30 weeks. Survival rates were significantly reduced for the high-dose male and low- and high-dose female rats, but were not treatment related. Overall, there were no treatment-related increases for nonneoplastic lesions or neoplasia reported. The authors concluded that "there was no evidence of carcinogenic activity" for MCA in male and female F344/N rats at the dose levels used in the study.

In the same 2-year bioassay, B6C3F1 mice (60 per sex per dose) were dosed with MCA (acid) in deionized water by gavage at 0, 50 or 100 mg/kg bw per day, 5 days per week (NTP, 1992). A significant reduction was noted for mean body weight in high-dose females (after 52 weeks) and in survival rates for high-dose males. The incidences of acute nasal inflammation ranged from mild to minimal in severity but were not treatment related. The incidence of metaplasia of the olfactory epithelium was significantly greater in high-dose females. No other significant increases in non-neoplastic lesions were observed in either sex. No treatment-related increases for any neoplasia were reported. The authors stated that "there was no evidence of carcinogenic activity" for MCA in B6C3F1 mice (both sexes) at the dose levels used in the study.

In a 104-week drinking water study, male F344/N rats (50 per dose) were exposed to MCA (neutralized) at 0, 50, 500 or 1100 mg/L* (0, 3.5, 26 or 59.9 mg/kg bw per day) (DeAngelo et al., 1997). MCA treatment had no significant effect on survival, but drinking water consumption and final mean body weights were significantly reduced in the two highest dose groups. Absolute and relative spleen weights were significantly increased (74% and 80% over controls, respectively) for rats consuming 3.5 mg MCA/kg bw per day; however, this increase was seen in the absence of gross or microscopic lesions. Although decreases in absolute and relative spleen weights were observed at the two highest doses, only the absolute weight was statistically lower at the highest dose (59.9 mg/kg bw per day). Exposure to 26 mg/kg bw per day significantly decreased absolute liver and kidney weights and relative liver weights and increased relative testes weights. At the 59.9 mg/kg bw per day dose, absolute kidney and absolute and relative liver weights were reduced, while relative testes weights were increased. Serum enzyme analysis revealed no treatment-related effects on either AST or ALT, and MCA treatment did not enhance peroxisome proliferation or hepatocyte proliferation. No MCA-induced liver pathology was observed during the study, nor were there any significant lesions found in any non-hepatic tissues. Most common age-related spontaneous changes present in rat tissues were within historical controls except for myocardial degeneration and chronic/active inflammation of the nasal cavities, which were seen at an increased incidence at 104 weeks but not at earlier sacrifice periods or in the lower dose groups. There were no significant increases in either the prevalence or multiplicity of hepatocellular adenomas, carcinomas or hyperplastic nodules in any of the MCA-treated animals compared with controls. Similarly, neoplastic changes in non-hepatic tissues were consistent with those reported in historical controls. The authors concluded that the decreased body weights and other pathology observed in the mid-dose groups were not significant and set the NOEL for carcinogenicity at 500 mg/L (26 mg/kg bw per day) for MCA (DeAngelo and Daniel, 1992; DeAngelo et al., 1997).

However, Health Canada did not agree with the decision to ignore statistically significant changes in body weight and liver, kidney and testes weights, particularly since these changes are consistent with a dose-related trend. At the next lower dose of 50 mg/L (3.5 mg/kg bw per day), the only significant change was an increase in absolute and relative spleen weights. This change can be considered idiosyncratic (a physiological peculiarity), since there is no dose-related trend towards an increase in spleen weight. At the next dose, 500 mg/L (26 mg/kg bw per day), the absolute and relative spleen weights were lower than control, while the absolute weight was significantly lower at the highest dose. Based on these considerations, Health Canada derived a NOAEL of 3.5 mg/kg bw per day.

10.3.2 Dichloroacetic acid

Several studies on the potential carcinogenicity of DCA have been conducted and have been summarized in Table 9. Increased incidences of either liver tumours (adenomas and carcinomas) or preneoplastic lesions (e.g., hyperplastic nodules and altered hepatic foci) were reported when mice and rats were exposed to drinking water containing 0.5-5 g DCA/L for periods ranging from 52 to 104 weeks. The details of the studies are reported below.

In a 52-week drinking water study (Bull et al., 1990), groups of male B6C3F1 mice (n = 61, 11, 50) were exposed to DCA (neutralized) at 0, 1000 or 2000 mg/L (0, 140 or 280 mg/kg bw per day; WHO, 2005), and a group of 10 females was exposed to 2000 mg DCA/L for up to 52 weeks. Interim sacrifices of five male mice from the 2000 mg/L group were carried out at 15, 24 and 37 weeks. All other animals were sacrificed on week 52, including 11 males that had their exposure terminated at 37 weeks, followed by 15 weeks of recovery before sacrifice. A significant dose-related increase in the absolute and relative liver weights was noted in treated males at 37 or 52 weeks. Livers from all DCA-treated animals were enlarged (hepatomegaly), with a uniform distribution of marked cytomegaly and extensive accumulation of glycogen in hepatocytes as well as multi-focal areas of necrosis with lymphocyte infiltration throughout the liver. Glycogen-poor basophilic foci of cellular alteration were observed at 24 and 37 weeks in the central lobes of livers from males treated with 2000 mg DCA/L. After 52 weeks of treatment, hyperplastic nodules, adenomas and carcinomas were observed in the livers of males in the 2000 mg/L group. In contrast, the only hepatoproliferative lesions observed in the other groups were one hyperplastic nodule in a male mouse from each of the control and 1000 mg/L dose groups and hyperplastic nodules in three females. No hepatocellular carcinomas were reported in males where DCA treatment was terminated at 37 weeks; by 52 weeks, two animals had adenomas, and six had hyperplastic nodules. The relationship between mean total dose of DCA and total number of hepatic lesions (nodules + adenomas + carcinomas) per mouse was nonlinear, with the number of lesions increasing sharply as the dose increased from 1000 to 2000 mg/L (Bull et al., 1990).

Male B6C3F1 mice (30 per dose) were administered DCA in their drinking water at 50, 500 or 5000 mg/L for 60 and/or 75 weeks (DeAngelo et al., 1991; U.S. EPA, 1991). The control group received 2 g sodium chloride/L. In a concurrent study, mice were exposed to 3500 mg DCA/L and killed after 60 weeks. In this experiment, the control group received acetic acid. Time-weighted mean daily doses of 7.6, 77, 410 and 486 mg/kg bw per day were calculated for the 50, 500, 3500 and 5000 mg/L concentrations, respectively. Mice exposed to 3500 and 5000 mg/L had their final body weights reduced compared with the control group: 87% and 83% of the control value, respectively. The relative liver weights were increased at the three highest dose groups compared with the control value: 118%, 230% and 351% for 500, 3500 and 5000 mg/L, respectively.

Hyperplastic liver nodules (a non-neoplastic nodular lesion) were observed primarily at the two highest doses: 58% and 83% incidence at 3500 and 5000 mg/L, respectively. Mice receiving 5000 mg/L (after 60 weeks) had a 90% prevalence of liver neoplasia (carcinomas and adenomas), with a mean multiplicity of 4.50 tumours per animal; those receiving 3500 mg/L had a 100% prevalence, with a tumour multiplicity of 4.0 tumours per animal. The tumour prevalence and multiplicity in the two lowest dose groups at 75 weeks did not significantly differ from the control value. The control group had no liver tumours. The authors concluded that there was a significant positive dose-related trend in age-adjusted prevalence of liver tumours. The authors concluded that DCA exhibited a threshold of at least 500 mg/L (77 mg/kg bw per day) for tumour response in mice. A steep increase in tumour incidence occurred at 3500 mg/L (410 mg/kg bw per day), which also represents the maximum incidence attained (DeAngelo et al., 1991; U.S. EPA, 1991).

In a 104-week drinking water study (Daniel et al., 1992), DCA (neutralized) was administered to a group of 33 male B6C3F1 mice at a concentration of 500 mg/L (93 mg/kg bw per day), with one interim sacrifice (n = 5) at 30 weeks. There were no significant treatment-related effects on drinking water consumption, relative body weight gain or relative or absolute spleen, kidney or testes weight. Absolute and relative liver weights were increased at both 30 and 104 weeks. The most notable treatment-related non-neoplastic hepatic effects observed were cytomegaly, necrosis and chronic active inflammation. After 104 weeks, there was a significant increase in the incidence of hepatocellular carcinomas and hepatocellular adenomas. The combined incidences of carcinoma + adenoma and carcinoma + adenoma + nodule were significantly increased in treated animals compared with controls. Mice killed at the end of the study had hepatocellular carcinomas (15/24, compared with 2/20 in the control groups); hepatocellular adenomas (10/24, compared with 1/20 in the controls); and carcinomas or adenomas (18/24, compared with 3/20 in the controls). Other effects observed were hyperplastic nodules in 2/24 treated mice, hepatocellular necrosis in 8/24 treated mice (compared with 1/20 in the controls) and cytomegaly in 22/24 treated mice (compared with 1/20 in the controls) (Daniel et al., 1992). (Data were not provided for interim necropsies at week 30.)

In a 90- to 100-week drinking water study, male B6C3F1 mice (n = 35-71) were exposed to DCA (neutralized) at 50, 500, 1000, 2000 or 3500 mg/L (8, 84, 168, 315 and 429 mg/kg bw per day) (DeAngelo et al., 1999). The control group consisted of 88 mice. Interim sacrifices were performed throughout the study, except at the lowest dose. At final sacrifice, the body weights were decreased at the two highest doses. A dose-dependent increase in liver weight at 84 mg/kg bw per day and above was seen at 26 and 52 weeks and at 315 mg/kg bw per day and above at 100 weeks. Liver toxicity was demonstrated at 168 mg/kg bw per day and above, as indicated by an increase in serum liver enzymes and histopathology. Mortality at the two highest doses was statistically significant when compared with the control group. There was a significant increase in the incidence of hepatocellular carcinoma in male mice. At 26 weeks, no tumours were seen in the livers of any of the mice, but at 52 and 78 weeks, the incidence of hepatocellular carcinoma was significantly (P < 0.05) elevated in the highest dose group (Table 10). At the terminal sacrifice, the incidence of hepatocellular carcinoma was significantly elevated (P < 0.05) at the three highest dose groups. Hepatocellular adenomas were first observed at the high dose at 26 weeks; however, only at 100 weeks was the incidence of hepatocellular adenomas significantly (P < 0.05) increased (no dose-response) in the three highest dose groups (Table 10).

Table 10: Prevalence of hepacellular carcinomas and adenomas in mice (DeAngelo et al., 1999)

Dose (mg/kg bw per day)	Prevalence of male mice with hepatocellular carcinomas (%)			Prevalence of male mice with hepatocellular adenomas (%)
	Week 52	Week 78	Week 100	Week 52	Week 78	Week 100
a n/r = not reported. b Statistically significant (P # 0.05) when compared with the water control.
0 (water control)	0	10	26	0	10	10
8	n/ra	n/r	33	n/r	n/r	n/r
84	0	0	48	10	10	20
168	0	20	71b	10	20	51.4b
315	20	50	95b	0	50	42.9b
429	50	70b	100b	50	50	45b

Liver peroxisome proliferation was significantly elevated only in the high-dose group after 26 weeks, but not at 52 weeks. In contrast, hepatocyte proliferation was not significantly different from the control rates at any of the doses that produced tumours. The authors could not determine a NOEL for hepatocarcinogenicity based on the significant increase in hepatocellular carcinoma multiplicity at the lowest dose (0.58 compared with 0.28 in the control), but concluded that DCA produced a dose-related increase in the incidence of hepatocellular carcinoma that was not associated with either liver peroxisome or hepatocyte proliferation. Endocrine disruption and liver cell necrosis were proposed as playing important roles in the carcinogenesis (DeAngelo et al., 1999).

Groups of 7- to 8-week-old B6C3F1 female mice (n = 40-90 and n = 134 in the control) were administered DCA (neutralized) in drinking water for either 360 or 576 days at concentrations of 0, 2.0, 6.67 or 20.0 mmol/L (0, 40, 120 or 330 mg/kg bw per day) (Pereira, 1996; IPCS, 2000). An additional group of animals (n = 50) was exposed intermittently to 20.0 mmol DCA/L in drinking water in a 72-day cycle, 24 days with exposure and 48 days without. The 72-day cycle was repeated until the mice were sacrificed, so that the total dose was the same as for those continuously exposed to 6.67 mmol DCA/L. Body weights were reduced in the high-dose group after 35 weeks. There was a dose-related increase in relative liver weight and vacuolated hepatocytes. For the 576-day exposure groups, there was a significant increase in hepatic foci, hepatocellular adenomas and hepatocellular carcinomas at the high dose and an increased incidence of foci and adenomas (but not carcinomas) at the middle dose, compared with controls. The intermittent DCA group also displayed a significantly increased incidence of altered hepatic foci at 576 days, but no significant increases in neoplastic response (Pereira, 1996).

In a modified carcinogenesis bioassay (DeAngelo et al., 1996), 28-day-old male F344 rats (n = 50-78) were exposed to DCA (neutralized) at 0, 50, 500 or 1600* mg/L (estimated time-weighted mean daily doses: 0, 3.6, 40.2 and 139.1 mg/kg bw) in drinking water for 100 weeks. There were no significant differences in water consumption or survival for any of the treatment groups when compared with controls. Terminal body weights and relative liver and kidney weights were reduced only in the 1600 mg/L group. Testicular effects are reported in Section 10.5.2. The only non-neoplastic treatment-related hepatic change was hepatocellular cytoplasmic vacuolization, attributed to DCA-induced increases in glycogen deposition. The combined hepatocellular adenoma and hepatocellular carcinoma prevalence was significantly increased (P < 0.05) in the 500 mg/L group compared with controls, as was the total hepatoproliferative lesion prevalence (hyperplastic nodules, adenomas and carcinomas). Significant dose-related trends were observed for hepatocellular adenoma and hepatocellular carcinoma (P < 0.05 for each) prevalence, combined hepatocellular adenoma and hepatocellular carcinoma prevalence and total hepatoproliferative lesions. In the 1600 mg/L group, increased prevalences of hepatocellular carcinoma, combined hepatocellular carcinoma and hepatocellular adenoma and total hepatoproliferative lesions were observed. At the high dose, DCA induced hepatocyte peroxisome proliferation. DCA treatment depressed hepatocyte proliferation at 14 weeks; at the other time periods, it remained depressed, but did not differ significantly from the control group. The authors concluded that DCA is a hepatocarcinogen in male F344 rats. Male F344 rats were found to be more sensitive to DCA exposure than male B6C3F1 mice based on a previous study by DeAngelo et al. (1991). The authors set a NOEL of 3.6 mg/kg bw per day (DeAngelo et al., 1996).

The ability of DCA to act as a promoter of carcinogenesis in the liver was observed in several studies. According to Pereira (1995), "promotion is defined as the enhancement of the progression of initiated cells to precancerous lesions and tumors."

A tumour promotion study was conducted in male B6C3F1 mice with DCA (neutralized) (Herren-Freund and Pereira, 1986; Herren-Freund et al., 1987). Tumour incidences (hepatocellular adenomas and carcinomas) were significantly elevated for mice with and without initiation compared with the controls. The authors concluded that DCA is a complete hepatocarcinogen in B6C3F1 mice (Herren-Freund et al., 1987).

The potential of DCA (neutralized) to promote tumours was also investigated in female B6C3F1 mice (Pereira, 1995; Pereira and Phelps, 1996). An increased incidence of hepatic foci and adenomas was seen in initiated mice exposed to the highest dose of DCA for either 31 or 52 weeks compared with the corresponding animals receiving initiation or DCA alone. A second-order relationship (i.e., exponential relationship) was observed between the concentration of DCA and the yield of total lesions (foci + adenomas + carcinomas) at both 31 and 52 weeks. Upon termination of treatment, the foci of altered liver hepatocytes and adenomas regressed.

Two separate initiation-promotion studies, conducted by Bull et al. (2004) and Pereira et al. (1997), looked at the interaction between DCA and TCA (both tumour promoters involving different mechanisms) in B6C3F1 mice. Different results between the two studies may be related to doses used.

Bull et al. (2004) demonstrated that the interactions between TCA and DCA were limited by additivity: the lowest effective doses with TCA and DCA showed additivity, whereas the effects tended to be inhibitory (suppression of overall growth rate) as the doses increased. DCA and TCA selectively promote the growth of different types of initiated cells, by stimulating the clonal expansion in one cell type while inhibiting it in another cell type. Such a process would explain the presence of a mixed phenotype of tumours as well as the suppression of the overall tumour growth rate in the present study (Bull et al., 2004).

In contrast, Pereira et al. (1997) observed synergistic effects with a mixture of TCA and DCA at high doses; the yields of foci of altered hepatocytes and total lesions produced were greater than the sums of the yields produced by the two HAAs administered alone, whereas additive or inhibitory effects were seen with the mixtures at lower doses. The mixtures of DCA-TCA produced more foci of altered hepatocytes than adenomas, similar to the effects seen when DCA was administered alone. In addition, the phenotype of these lesions resembled those of DCA-induced tumours, suggesting that DCA may predominate in determining the characteristics of the proliferative lesions.

10.3.2.1 Mechanisms of carcinogenicity

Based on the results of the studies described above and summarized in Table 9, DCA has been shown to be a liver carcinogen in two species of rodents, mice and rats. DCA is thought to be a "complete carcinogen," since it has induced tumours at both low doses in long-term assays and high doses in shorter-term assays when administered alone; doses from these studies ranged from 50 to 5000 mg/L. More than one mode of action may explain DCA-induced carcinogenicity, and several hypotheses have been put forward, as described below. In all likelihood, a number of events would be significant to tumour development in the rodent under bioassay conditions. Uncertainty exists, however, as to which events may be relevant to human exposure to DCA at environmental levels.

10.3.2.1.1 Genotoxicity

Mixed reviews were seen with regard to DCA-induced carcinogenicity mediated through a genotoxic mechanism. Previous reviews by IARC (1995) and ILSI (1997) reported that DCA was not genotoxic, whereas a more recent review by IPCS (2000) reported that DCA had some ability to induce genotoxic effects but only at high concentrations; it concluded that at low doses, DCA was either not acting or acting minimally through a genotoxic mechanism. The U.S. National Center for Environmental Assessment (U.S. EPA, 2003c) reported that more recent literature indicated that DCA is a direct-acting genotoxic agent. IARC (2004) recently reported that DCA is genotoxic (in vivo and in vitro) but may be acting indirectly via an epigenetic mechanisn. Owing to the lack of causal data, U.S. EPA (2003c) took the cautious position that DCA might be genotoxic at high doses while uncertainty remained for the lower doses.

10.3.2.1.2 Peroxisome proliferation

Some hepatocarcinogens have been shown to increase the number and/or size of liver peroxisomes (peroxisome proliferation) in rodents (U.S. EPA, 2003c). Peroxisome proliferatoractivated receptors (PPARs) are a class of nuclear receptors that regulate this proliferation and are believed to be responsible for the initiation of certain cellular events leading to transformation (as well as other reported effects) for liver carcinogens. Although peroxisome proliferation has been correlated with carcinogenesis, the actual mechanism of carcinogenesis as it is related to peroxisome proliferation is unknown (Bull, 2000). Species differences are seen with relation to the expression of various PPARs; humans seem less responsive to a variety of peroxisome proliferators than rodents, and, as a result, controversy exists as to whether these compounds are even carcinogenic to humans (Bull, 2000; U.S. EPA, 2003c).

Studies conducted by DeAngelo et al. (1989, 1999), Daniel et al. (1992), Mather et al. (1990) and Pereira (1995, 1996) showed that DCA was a weak peroxisome proliferator in rodents.

Seemingly, much lower doses of DCA are needed to induce liver tumours compared with those that cause a significant increase in peroxisome proliferation (U.S. EPA, 2003c). In conclusion, U.S. EPA, (2003c), Thai et al. (2003) and Bull (2000) do not consider this mechanism an important route in DCA tumorigenesis.

10.3.2.1.3 Down-regulation of insulin

The glycogen content of liver cells of DCA-treated mice can vary depending on the state of the cell; altered hepatic foci and tumours are glycogen poor, whereas normal liver cells accumulate large amounts of glycogen (Lingohr et al., 2001), thus suggesting a possible link between glycogen and hepatic tumours. Glycogen levels are mediated primarily by insulin via metabolism in the liver (Lingohr et al., 2001). Insulin may have other roles, such as acting as an agent that triggers mitosis (mitogen) for normal and malignant liver cells and suppressing apoptosis (Lingohr et al., 2001).

Lingohr et al. (2001) investigated the effects of DCA on insulin levels and expression of insulin-controlled signalling proteins in normal liver tissue and DCA-induced liver tumour tissues of male B6C3F1 mice treated with DCA (neutralized) at 100-2000 mg/L in drinking water for 2-10 weeks. Decreases were seen in insulin receptor proteins, serum insulin levels and protein kinase B expression after 2 weeks in DCA-treated mice. In contrast, increases in glycogen (which preceded these effects) were seen as early as 1 week, suggesting that DCA-induced alterations in insulin, insulin receptors and possibly protein kinase B resulted from a compensatory down-regulation of the insulin pathway triggered by high glycogen levels in the liver. An in vitro study by Lingohr et al. (2002) using isolated hepatocytes showed similar results: increased glycogen levels (independent of insulin) and effects on insulin signalling proteins.

These studies show an apparent down-regulation of insulin and insulin receptor activity after relatively short durations of treatment.

10.3.2.1.4 Tumour promotion, alterations in cell replication and death

Stauber and Bull (1997) studied the differences in phenotype and cell replicative behaviour of liver tumours induced by DCA and TCA in male B6C3F1 mice. Clear differences in phenotype were seen between DCA- and TCA-induced tumours. The study also showed the extent to which changes in cell replication within tumours and normal hepatocytes were influenced by DCA to increase tumour formation. Stauber et al. (1998) also conducted in vitro studies, which duplicated these results.

Several mechanisms of action are proposed for tumour promoters. Bull et al. (2004) suggested that promoters should influence mainly tumour size rather than tumour numbers. This behaviour can be observed with DCA. Miller et al. (2000) investigated, by means of magnetic resonance imaging, the growth rates of liver tumours in male B6C3F1 mice given DCA (neutralized) at 2000 mg/L in their drinking water for 48 weeks - a treatment period ensuring induction of small liver tumours. Results showed that continued growth of the tumours was entirely dependent upon DCA treatment. The authors suggested that DCA's main effect in tumour induction is mediated through accelerated growth of spontaneously initiated cells and that it may be due to the suppression of apoptosis and modification of cell replication rates.

Another hypothesis for promoters is the suppression of apoptosis. Snyder et al. (1995) studied the frequency of spontaneous apoptosis in liver hepatocytes of male B6C3F1 mice when given DCA at dose levels of 0, 500 or 5000 mg/L in their drinking water for 5-30 days. DCA significantly reduced apoptosis in treated mice in a dose-dependent manner relative to the untreated controls. The authors suggest that disrupting the apoptosis process can result in the outgrowth of initiated cells (by suppressing the ability of the liver to remove initiated cells, preneoplastic cells) and thus leading to the formation of tumours, rather than by induction of selective proliferation of initiated cells.

In a review of TCE and its metabolites (Bull, 2000), the data suggest that the modification of cell signalling pathways resulting in cell replication, selection and apoptosis may be an important contributor to the hepatocarcinogenicity of DCA.

In a hepatic cell proliferation study (Pereira, 1995, 1996), groups of 10 female B6C3F1 mice were exposed to DCA (neutralized) in drinking water at concentrations of 0.26, 0.86 or 2.6 g/L (52, 172 or 520 mg/kg bw per day) until sacrifice at day 5, 12 or 33. A concentration-related increase in cell proliferation was seen after 5 days of treatment with DCA but not at 12 and 33 days, except for an apparent increase at the high dose at 12 days. Therefore, DCA treatment caused a transient increase in hepatic cell proliferation (Pereira, 1995, 1996).

Increases in hepatocyte proliferation were observed in other short-term studies (Carter et al., 1995; Stauber and Bull (1997). Decreased cell proliferation was seen at higher doses and with chronic dosing periods (Bull, 2000; U.S. EPA, 2003c).

Carter et al. (2003) conducted a histopathological analysis of hepatic lesions from male mice from the carcinogencity study by DeAngelo et al. (1999) in order to identify and quantify the different phenotypes of hepatocellular lesions, such as altered hepatic foci, large foci of cellular alteration, adenomas and carcinomas. As a result of the analysis, three different lesion sequences were proposed during mouse liver carcinogenesis - altered hepatic foci, large foci of cellular alteration and adenomas - which demonstrated neoplastic progression with time. The analysis also demonstrated that some toxic adaptive changes in non-involved liver were related to this neoplastic progression, regardless of the dose and length of exposure. According to Carter et al. (2003), the homeostasis of liver cells is altered by DCA, which leads to negative selection of cells (by suppressing apoptosis, a natural process for eliminating initiated cells), with a new state of differentiation resistant to DCA toxicity, thereby allowing the growth and/or survival of these initiated cells (i.e., preneoplastic cells and lesions). These dose-related effects occur at less than 1 g/L, which the authors considered as the inflection point of the dose-response curve for carcinogenesis.

10.3.2.1.5 Other mechanisms: hypomethylation

Another hypothesis for non-genotoxic carcinogens may be through an epigenetic mechanism involving DNA methylation (Pereira et al., 2004). The reduction in the level of DNA methylation (hypomethylation) is a common event in most cancers, including liver cancer (Pereira et al., 2004).

DNA methylation, a DNA modification that occurs naturally, takes place when a methyl group is added to the 5-position carbon of the cytosine ring to form 5-methylcytosine; the methyl group is supplied by S-adenosylmethionione, while the reaction is catalysed by DNA methyl-transferase (Ge et al., 2001; Pereira et al., 2004). Disruption of these different processes of DNA methylation may lead to hypomethylation (Tao et al., 2000).

DCA (neutralized) reduced the level of DNA methylation (hypomethylation) in the liver and as a result induced foci of altered hepatocytes and hepatocellular adenomas when given to female B6C3F1 mice in drinking water (Pereira et al., 2004). The authors suggested that there may be a correlation between the prevention of carcinogen-induced DNA methylation and prevention of liver tumours and that DNA hypomethylation may be critical for the carcinogenic effects of DCA.

Tao et al. (1998) showed in a promotion study that DCA reduced the levels of 5-methylcytosine in DNA of liver tumours and that the neoplastic progression of the liver lesions, from adenomas to carcinomas, seemed associated with a decrease in the level of 5-methylcytosine in DNA. In another study, Tao et al. (2000) reported that DCA decreased the methylation in the promoter regions for 2 proto-oncogenes, c-jun and c-myc, and increased the expression of their mRNA and proteins, while the addition of methionine prevented these changes. Both these proto-oncogenes are known to participate in the control of cell proliferation (U.S. EPA, 2003c).

Stauber and Bull (1997) identified that DCA-induced lesions had a phenotype that was cjun immunoreactive.

Uncertainty exists regarding the actual importance that decreased methylation associated with an increase in the expression of the mRNA bears on the mediation of tumorigenicity of DCA (U.S. EPA, 2003c).

10.3.3 Trichloroacetic acid

Several studies on the potential carcinogenicity of TCA have been conducted and have been summarized in Table 9. Increased incidences of liver tumours (adenomas and carcinomas) were reported when mice were exposed to drinking water containing 500-5000 mg TCA/L for periods ranging from 52 to 104 weeks. No liver tumours were found in rats. The details of the studies are reported below.

In a 1-year chronic drinking water study (Bull et al., 1990), groups of male B6C3F1 mice (n = 61, 11 and 50) were exposed to TCA (neutralized) at 0, 1000 or 2000 mg/L (0, 178 and 319 mg/kg bw per day, respectively; WHO, 2004b), and a group of 10 females was exposed to 2000 mg/L in drinking water for up to 52 weeks. Interim sacrifices were made throughout the study, but all animals were sacrificed by week 52. Dose-related accumulation of lipofuscin (indicative of intracellular lipid peroxidation) in the liver was seen in mice at 52 weeks but not in males terminated at 37 weeks. Dose-related increases in the incidence of hepatoproliferative lesions - namely, hyperplastic nodules, adenomas and hepatocellular carcinomas - were seen in mice at 1000 mg/L and above at 52 weeks. Large concentrations of lipofuscin were found in areas surrounding the hepatoproliferative lesions but were absent from the lesions themselves. There was a linear relationship between the mean total dose of TCA consumed over 52 weeks and the number of hepatic lesions (nodules + adenomas + carcinomas) per mouse, although the incidence of lesions per mouse for animals in the 37-week exposure group was less than what would have been predicted on the basis of the total dose administered. No hyperplastic or neoplastic lesions were observed in any of the female mice treated with TCA. In rats treated with TCA, the liver showed only small increases in hepatic cell size, modest accumulation of glycogen, absence of focal necrosis and only marginal induction of cell proliferation and organ hypertrophy (Bull et al., 1990).

In a 60-week drinking water study (DeAngelo and Daniel, 1990), groups of male B6C3F1 mice (number not specified) were administered TCA (possibly neutralized based on a similar study with DCA by DeAngelo et al., 1991) at 0, 50, 500 or 5000 mg/L (0, 7, 71 and 595 mg/kg bw per day). A dose-related increase in the incidences of hyperplastic nodules, adenomas and carcinomas was seen. The hepatocellular tumour incidence was 55.2% at the high dose, 37.9% at the middle dose and 13.37% in the untreated control group. At the lowest dose (50 mg/L), no significant difference in tumour prevalence or multiplicity was observed compared with the control group. Hyperplastic nodules of the liver were seen in the highest dose only (prevalence: 24.1%), while none was seen at the lower dose or in the control group. Chronic inflammation and liver necrosis were observed in the two highest dose groups near the end of the study.

In another chronic study, female B6C3F1 mice (n = 10-40) received TCA (neutralized) at 2.0, 6.67 or 20.0 mmol/L (64, 212 and 640 mg/kg bw per day) in their drinking water for 360 or 576 days (Pereira, 1996; Pereira and Phelps, 1996). The control group received 640 mg sodium chloride/kg bw per day. At the high dose, after 360 days of exposure, a significant increase in hepatocellular carcinoma incidence and multiplicity was seen; after 576 days of exposure, significant increases in altered hepatic foci, hepatocellular adenoma and hepatocellular carcinoma incidences and yields were seen. After 576 days, the middle dose also had increased incidences and yields of foci and carcinomas. The combined numbers of total lesions (foci + adenomas + carcinomas) and total tumours (adenomas + carcinomas) appeared to be linearly related to TCA dose. The authors concluded that TCA is a peroxisome proliferator and that the basophilic staining of the tumours is consistent with other peroxisome proliferators (Pereira, 1995, 1996; Pereira and Phelps, 1996).

In a review of the carcinogenicity of TCA in rodents by the U.S. EPA (1991), the results from a carcinogenicity study of TCA (form unknown) in female B6C3F1 mice (number unknown) were reported. Groups of female mice were administered 0, 500 or 4500 mg/L (0, 71 or 583 mg TCA/kg bw per day; WHO, 2004b) in drinking water for 104 weeks. The combined liver tumour incidence (adenomas + carcinomas) was significantly increased in the high-dose group. The high dose had a 64% hepatocellular tumour incidence, and the lowest dose had a tumour incidence of 16.7%; the incidence was 7.7% in the untreated control group. The LOAEL as derived by U.S. EPA (1991) in female mice is 71 mg/kg bw per day based on the presence of tumours.

There was no evidence of increased liver tumours in male F344 rats in a 2-year carcinogenicity study in which groups (50 per dose) were administered TCA (neutralized) at 0, 50, 500 or 5000 mg/L (0, 3.6, 32.5 or 364 mg/kg bw per day) in drinking water (DeAngelo and Daniel, 1992; DeAngelo et al., 1997). In the high-dose group, significant decreases were seen in body weights and in absolute liver weights, while increases in serum levels of ALT and palmitoyl conenzyme A activity (a marker of hepatic peroxisome proliferation) were seen. The authors set a NOEL of 364 mg/kg bw per day. A NOAEL based on non-neoplastic effects is derived at 32.5 mg/kg bw per day.

The ability of TCA to act as a promoter of carcinogenesis in the liver was also examined in several studies. Parnell et al. (1986, 1988) studied the initiation and promotion abilities of TCA (neutralized) using rat hepatic enzyme-altered foci bioassays. The authors concluded that TCA was a weak peroxisome proliferator and a liver tumour promoter, but not an initiator of liver tumours in Sprague-Dawley rats (Parnell et al., 1986, 1988).

The potential of TCA (neutralized) to promote tumours was also investigated in male B6C3F1 mice (Herren-Freund and Pereira, 1986; Herren-Freund et al., 1987). The incidence of hepatocellular adenomas and carcinomas was significantly increased for both types of treatment: those pretreated with an initiator or not. The authors concluded that TCA was hepatocarcinogenic, regardless of pretreatment with an initiator (Herren-Freund and Pereira, 1986; Herren-Freund et al., 1987).

In a second promotion study, Pereira (1995); Pereira and Phelps (1996) used groups of female B6C3F1 mice. A small increase in the incidence of neoplastic changes was noted at the highest dose in TCA groups without prior initiation. In contrast, when mice were pretreated with an initiator and then exposed to TCA, a significant increase in the incidence and multiplicity of adenomas and carcinomas was seen at the higher doses. The total number of neoplastic lesions per mouse was linearly related to the TCA dose with time.

10.3.4 Monobromoacetic acid

No long-term studies on MBA were identified.

10.3.5 Dibromoacetic acid

Although an abstract published by Bull (1995) referred to a 2-year limited drinking water study on the carcinogenic potential of DBA and other brominated haloacetates in B6C3F1 mice and F344 rats, the final version has yet to be published.

However, a more recent 2-year carcinogenicity drinking water study on DBA was published by Melnick et al. (2007) and with greater detail by NTP (2007). Groups of F344/N rats (50 per sex per dose) and B6C3F1 mice (50 per sex per dose) were administered DBA (neutralized to pH 5) in drinking water at 0, 50, 500 or 1000 mg/L (equivalent to 0, 2, 20 and 40 mg/kg bw per day for male rats and 0, 2, 25 and 45 mg/kg bw per day for female rats; 0, 4, 45 and 87 mg/kg bw per day for male mice and 0, 4, 35 and 65 mg/kg bw per day for female mice). No effect on survival was seen in either species, nor were there any changes in body weight in mice. However, in rats (both sexes), mean body weight was decreased at the two highest doses compared with controls, whereas water consumption was decreased at the highest dose (both sexes). Neoplastic lesions were observed at multiple sites in both rats and mice.

In male rats, a significant increase in malignant mesotheliomas of the abdominal cavity was observed at the highest dose. In female rats, a positive increasing trend in the incidence of monocellular cell leukaemia, a haematopoietic (involved in the formation of blood cells) neoplasm, was noted that was significant at the highest dose. In contrast, male rats showed a significant increase in monocellular cell leukaemia at the lowest dose and no increase at the highest dose (which was similar to concurrent controls and historical controls), but incidences seen at the low and middle doses exceeded historical controls. Other non-cancer effects included significant increased incidence of lesions in the liver (cystic degeneration, minimal to mild) of all exposed groups of males, in the lung (alveolar epithelial hyperplasia) of the two highest dose groups in females, and in the kidney (nephropathy) of all exposed groups of females.

In mice, neoplasms were observed in both the liver and lung. A significant increase in the incidence of multiple hepatocellular adenoma and hepatocellular adenoma or carcinoma (combined) was observed in all treated males and in females at 500 mg/L and above, whereas a significant increase in the incidence of hepatocellular carcinoma was observed only in high-dose males and in mid-dose females. A significant increase in the incidence of hepatoblastoma was observed in male mice at 500 mg/L and above. The incidence of alveolar/bronchiolar adenoma exceeded the historical controls in males at 500 mg/L and 1000 mg/L, although the increase was significant only in males at 500 mg/L. In females, a non-significant increase in lung neoplasms was observed. Non-cancer effects, non-significant increases in the incidence of alveolar epithelial hyperplasia, were observed at all doses in male mice. An increased incidence of splenic haematopoieisis was also observed in high-dose male mice (NTP, 2007).

NTP (2007) indicates that there was clear evidence of carcinogenic activity of DBA in mice, based on increased incidences of hepatocellular neoplasm (both sexes) and hepatoblastomas (male only). It also finds some evidence of carcinogenic activity in rats based on an increased incidence of malignant mesothelioma in males and an increased incidence and positive trend of mononuclear cell leukaemia in females.