3.1.1. Oral administration

Two 2-year studies of carcinogenicity had been conducted with PFOA (specifically, ammonium perfluorooctanoate, or C8) in Sprague-Dawley rats.

The first study was conducted by a pharmaceutical company in the USA. Original reports of this study were submitted as regulatory documents to the EPA in 1983, and were not publicly available until Butenhoff et al. (2012a) published a report of this study. In this study, male and female Sprague-Dawley rats [Crl:COBS@ CD(SD)BR] (age, 39–41 days) were given diets containing PFOA at a concentration of 0, 30, or 300 ppm, corresponding to an average daily dose of approximately 0, 1.3, and 14.2 mg/kg bw) in males, and 0, 1.6, and 16.1 mg/kg bw in females. At 2 years, there was a significant treatment-related increase in the incidence of testicular Leydig cell adenoma in males at 300 ppm compared with concurrent controls, but not at 30 ppm. There was an increase in the incidence of fibroadenoma of the mammary gland in females at 30 and 300 ppm, but only the increase in the group at 300 ppm was significant compared with concurrent controls. There was an increase in the incidence of hepatocellular hypertrophy in males and females at the highest dose, and an increase in the incidence of liver cystic degeneration and portal mononuclear cell infiltrate in males at the highest dose (Butenhoff et al., 2012a). In 2005, a pathology working group was convened to review the original slides of the mammary glands and to provide a consensus diagnosis for the neoplasms of the mammary gland using current diagnostic criteria. The pathology working group concluded that several lesions originally diagnosed as lobular hyperplasia had features consistent with fibroadenoma of the mammary gland (mainly in slides from the control group), and that, consequently, PFOA did not induce neoplasms of the mammary gland (Hardisty et al., 2010). In a review of the pancreatic lesions from the male rats, using the same diagnostic criteria as those applied in the study by Biegel et al. (2001) (see below), a significant increase in the incidence of pancreatic acinar cell hyperplasia was identified at the highest dose (3/46, 1/46, 10/47) (Caverly-Rae et al., 2014). These hyperplastic lesions were considered to be proliferative lesions similar to the pancreatic acinar adenomas seen in the study by Biegel et al. (2001), and this supported the conclusion that the pancreas is a target of PFOA in male rats.

In the second study, designed to compare the carcinogenic effects of Wyeth-14643 with those of PFOA (specifically, ammonium perfluorooctanoate) (Biegel et al., 2001), there was a treatment group in which male Sprague-Dawley rats [Crl:CD BR (CD)] (age, 6 weeks) were given diet containing PFOA at a concentration of 300 ppm for 2 years. There was also a control group that was fed ad libitum, and a control group that received the same amount of food as the PFOA-treated group (pair-fed control group). The average daily doses of PFOA were 0, 0, and 13.6 mg/kg bw in the control group fed ad libitum, the pair-fed control group, and the treated group, respectively. There were initially 156 animals per group, but rats were killed at various interim time-points for measurements of cell proliferation, peroxisome proliferation, and hormone levels. [It was unclear how many rats were designated for pathological evaluation at the 2-year time-point. Survival data were provided in graphic form only (the actual numbers were not reported); the Working Group estimated survival percentages from the graph presented.] At 2 years, exposure to PFOA significantly increased the incidence of hepatocellular adenoma, testicular Leydig cell adenoma, pancreatic acinar cell adenoma, and pancreatic acinar cell adenoma or carcinoma (combined). In the testis, there was also an increase in the incidence of Leydig cell hyperplasia in the treated group compared with concurrent controls (Biegel et al., 2001).

3.1.2. Initiation–promotion

In an initiation–promotion study, male Wistar rats were given N-nitrosodiethylamine (NDEA) at a dose of 200 mg/kg bw as a single intraperitoneal injection (initiation), followed 2 weeks later by diet containing 0.05% phenobarbital, 0.015% PFOA [analytical grade, purity not reported], or 0.02% PFOA, for 46 weeks (Abdellatif et al., 1991). A control group was initiated with NDEA and was fed untreated diet. There were 10 rats per group. The average daily doses of phenobarbital and PFOA were not reported. Survival in the initiated group was 7/10, 7/10, 7/10, and 9/10 in the control group, the phenobarbital-treated group, and the groups treated with 0.015% PFOA, and 0.02% PFOA, respectively. No tumours were identified in rats that died at an early stage of the experiment, all within the first 8 months of the study, with the cause of death reported to be pneumonia in all cases. At 12 months, there was a significant increase in the incidence of NDEA-induced hepatocellular carcinoma in the rats receiving 0.02% PFOA compared with the control group. No organs other than the liver were evaluated in this study. [The Working Group noted the small number of animals and the absence of liver tumours in the control group.]

In an initiation–selection–promotion study, male Wistar rats were initiated with NDEA at a dose of 200 mg/kg bw as a single intraperitoneal injection (Abdellatif et al., 1990, 1991). After 2 weeks, they were given diet containing 0.03% 2-acetylaminofluorene (2-AAF) for 2 weeks. After 1 week of treatment with 2-AAF, the rats received a single necrogenic dose of carbon tetrachloride (2 mL/kg bw) by gavage. One week after the cessation of treatment with 2-AAF, the rats were given diet containing 0.05% phenobarbital or 0.015% PFOA for 23 weeks. A control group were initiated with NDEA then received 2-AAF plus carbon tetrachloride, but was fed untreated diet. The average daily doses of 2-AAF, phenobarbital, or PFOA were not reported. There were 7 rats in the control group, 8 rats in the phenobarbital-treated group, and 12 rats in the PFOA-treated group. Survival was 100% in all groups. The incidences of hepatic cancers were 0/7, 6/8, and 4/12 in the control, phenobarbital-treated, and PFOA-treated groups, respectively. The incidences in the phenobarbital-treated and PFOA-treated groups were significantly increased compared with controls. The cancers reported were hepatocellular carcinomas in all cases, except for one in the PFOA-treated group, that was reported as “other histologic type” and not further classified. [The Working Group noted the small number of animals, the absence of liver tumours in the control group, and the large amount of 2-AAF and chloroform administered.]

Initiation–promotion

Rainbow trout have been used as a model of hepatic carcinogenesis for many years and are sensitive to several suspected human carcinogens, including the hepatic carcinogens aflatoxin B₁ (AFB₁) and polycyclic aromatic hydrocarbons (Williams et al., 2003, Williams, 2012). The background incidence of hepatic neoplasms is reported to be approximately 0.1% at age 9–12 months (Williams et al., 2003).

In an initiation–promotion study in rainbow trout (Mount Shasta strain), one cohort of four groups (with approximately 250 fish per group) was exposed to either 0.01% ethanol (non-initiated sham control) or 10 ppb AFB₁ for 30 minutes by aqueous exposure at 10 weeks post-spawn. Another cohort of three groups (with approximately 167 trout per group) was exposed to either 0.01% dimethylsulfoxide (non-initiated sham control) or N-methyl-N′-nitro-N-nitrosoguanidine (MNNG) at 35 ppm for 30 minutes by aqueous exposure at 10 weeks post-spawn. For the subsequent 4 weeks, the trout were fed untreated feed (Oregon Test Diet, or OTD, a semipurified, casein-based diet). Beginning at 14 weeks post-spawn (4 weeks after initiation), the trout were given feed containing PFOA at 2000 ppm for six months after which the trout were held for 3 months before necropsy. Control trout were fed untreated OTD. The average daily dose of PFOA was not reported.

In the first cohort, there were four groups: non-initiated sham control/untreated feed control, AFB₁/untreated feed control, non-initiated sham control/PFOA, and AFB₁/PFOA. Neither non-initiated group developed hepatic neoplasms. The group initiated with AFB₁ had an incidence of hepatic neoplasms of 13%, while the group initiated with AFB₁ and promoted with PFOA had an incidence of hepatic neoplasms of 62%, which was significant compared with the AFB₁/control group. In the second cohort, there were three treatment groups as follows: non-initiated sham control/untreated feed control, MNNG/untreated feed control, and MNNG/PFOA. While the control/control group did not develop hepatic neoplasms, both MNNG-initiated groups developed hepatic neoplasms. There was, however, a significant increase in the incidence of hepatic neoplasms in the MNNG/PFOA group (86%) compared with the MNNG/control group (51%) (Benninghoff et al., 2012).