Ecological screening assessment report on perfluorooctane sulfonate, salts and precursors: chapter 4


4. Key Toxicological Studies

The toxicity of PFOS has been studied in a variety of aquatic and terrestrial species, including aquatic plants, invertebrates and vertebrates and terrestrial invertebrates, birds and mammals. Effects in laboratory mammals include: histopathological effects, increased tumor incidence, hepatocellular adenomas, hepatocellular hypertrophy, increased liver, kidney, brain and testes weight, reduced body weight, change in estrous cycling, changes in levels of neurotransmitters, decreased serum cholesterol, decreased bilirubin, and decreased triiodothynine. In mammalian reproduction studies, effects include: decreased body weight of dams, reduced gestation time, delivery time and live litter size, transfer of PFOS to fetus and neonate via placenta and ingestion of maternal milk, and reduced survival, body weight gain and development of lactation in offspring of exposed females. These effects are more fully reported in Health Canada (2004). Previous studies have shown that perfluorinated compounds are peroxisome proliferators (Berthiaume and Wallace 2002) and tumor promoters and may inhibit gap junction intercellular communication at environmentally relevant concentrations (Hu et al., 2002).

The following is a summary of the key studies used to identify the Critical Toxicity Value (CTV) for PFOS. A more complete review of effects is given in the OECD hazard review of PFOS, which discusses effects on fish, invertebrates, aquatic plants (algae and higher plants), amphibians and microorganisms (OECD 2002a). Additional studies by Boudreau et al. (2003a,b) and Sanderson et al. (2002) not available in OECD (2002a) are also summarized.

Aquatic

A flow-through bioconcentration study with bluegill (Lepomis macrochirus) using PFOS potassium salt saw no significant mortality at an exposure concentration of 0.086 mg.L-1 over a 62-day uptake phase; however, significant mortality was observed after a 35-day exposure to 0.87 mg.L-1. The study was stopped because all the fish either had died or had been sampled (US EPA OPPT AR226-1030a042).

Results have been published from a laboratory evaluation of the toxicity of PFOS to five aquatic organisms: green algae (S. capricornutum and C. vulgaris), duckweed (L. gibba) and water flea (D. magna and D. pulicaria) (Boudreau et al. 2003a). NOEC values were generated from the most sensitive endpoints for all organisms. The most sensitive of the organisms in this study was D. magna, with a 48-hour immobility NOEC of 0.8 mg.L-1; the accompanying LC50 was 112 mg.L-1, and the 48-hour IC50 for growth inhibition was 130 mg.L-1. The 21-day NOEC for lethality for D. magna was 5.3 mg.L-1. Autotroph inhibition of growth NOEC values were 5.3 mg.L-1, 6.6 mg.L-1 and 8.2 mg.L-1 for S. capricornutum, L. gibba and C. vulgaris, respectively.

In an aquatic microcosm study (Boudreau et al. 2003a), a field evaluation assessed the toxicological risk associated with PFOS across levels of biological organization. The zooplankton community was significantly affected by the treatment for all sampling times. A community-level NOEC of 3.0 mg.L-1 was determined for the 35-day study. The most sensitive taxonomic groups, Cladocera and Copepoda, were virtually eliminated in the 30 mg.L-1 treatments after 7 days, although specific survival rates were not quantified.

In a laboratory microcosm study that examined impacts to zooplankton following exposure to PFOS, adverse effects were observed at 10 mg.L-1 over 14 days; several species were significantly reduced or eliminated (Sanderson et al. 2002). In comparison with controls, exposures of 10 mg.L-1 and 30 mg.L-1 resulted in an average 70% change in species diversity and total zooplankton. The most sensitive species in the study was Cyclops diaptomus. The statistically significant effect concentrations for all species endpoints (abundance) were above 1 mg.L-1.

A fathead minnow (Pimephales promelas) embryo-juvenile flow-through chronic study determined a NOEC of 0.3 mg.L-1 over a 42-day exposure period. This value was for both survival and growth (US EPA OPPT AR226-0097). In acute tests, the lowest 96-hour LC50 for freshwater fish species was 4.7 mg.L-1 for the fathead minnow (P. promelas). In salt water, a 96-hour LC50 of 13.7 mg.L-1 was reported for rainbow trout (O. mykiss) (OECD 2002a). In a 96-hour static acute study using the freshwater mussel (Unio complamatus), the NOEC for mortality was 20 mg.L-1 and the LC50 was 59 mg.L-1 (US EPA OPPT AR226-0091, AR226-1030a047). The most sensitive saltwater invertebrate studied was the saltwater mysid (Mysidopsis bahia). Survival, growth and reproduction were assessed over an exposure period of 35 days. The NOEC determined for growth and reproduction were both 0.25 mg.L-1 (US EPA OPPT AR226-0101). In acute toxicity testing, a 96-hour LC50 of 3.6 mg.L-1 was reported for mysid shrimp (OECD 2002a). There was one study reported for embryo teratogenesis in aquatic organisms, which involved a 96-hour static renewal study on the frog, Xenopus laevis (US EPA OPPT AR226-1030a057). The minimum concentration that inhibited growth was 7.97 mg.L-1. The LC50 for mortality was 13.8 mg.L-1, the EC50 for malformed embryos was 12.1 mg.L-1 and the NOEC for embryo malformation was 5.2 mg.L-1. Calculated teratogenic indices ranged from 0.9 to 1.1, indicating that PFOS has a low potential to be a developmental hazard in this species.

The fathead minnow early life stages study has one of the lowest NOEC values (0.3 mg.L-1 , Klimisch ranking of 1) (OECD, 2002a). However, a recent study by Macdonald et al. (2004), although ranked 2 on the Klimisch scale, calculated NOEC values that are lower. MacDonald et al. (2004) reported a 10 day NOEC of 0.0491 mg.L-1 for the growth and survival of the aquatic midge (Chironomus tentans). The Klimisch ranking of 2 was determined for this study for two main reasons: (i) the use of static renewal exposures every 48 hrs and, (ii) the measurement of concentrations at the end of the study period (as opposed to after each 48 hour renewal). However, there was good agreement between the nominal and measured concentrations for the 10-day study. Also since PFOS is not a volatile substance, losses due to volatilization are considered negligible. Therefore, there is high confidence in the 10-day exposure values while the 60-day exposures should be treated with caution. As such, the 10-day NOEC from the MacDonald et al (2004) study was chosen as the most appropriate CTV for aquatic organisms.

Terrestrial Invertebrates

The OECD (2002a) review summarizes data indicating moderate to high toxicity of PFOS to honey bees (Apis mellifera). In an acute oral test, a 72-hour LD50 for ingestion of PFOS was 0.40 µg/bee, and a 72-hour No-Observed-Effect Level (NOEL) was 0.21 µg/bee. A contact test found a 96-hour LD50 of 4.78 µg/bee and a 96-hour NOEL of 1.93 µg/bee.

Results have been reported for an acute toxicity study with the earthworm in an artificial soil substrate (US EPA OPPT AR226-1106). The PFOS potassium salt 14-day LC50 was determined to be 373 mg.kg-1 body weight (bw), with a 95% confidence interval of 316-440 mg.kg-1 bw. The 14-day No Observed Effect Concentration (NOEC) for burrowing behaviour, body weight and clinical signs of toxicity was 77 mg.kg-1 bw, and the 14-day LOEC for the same endpoints was 141 mg.kg-1 bw.

Wildlife

Avian

Studies on the effect of PFOS on birds include chronic studies on mallard (Anas platyrhynchos) and bobwhite quail (Colinus virginianus) (US EPA OPPT AR 226-1738 and AR226-1831) and acute studies on mallard, bobwhite quail and Japanese quail (Coturnix coturnix japonica) (US EPA OPPT AR226-0103 and 104, McNabb et al. 2005). Given the persistent nature of PFOS, effects from chronic exposure are of particular interest in this assessment and are detailed in Appendix 3.

Mallard and bobwhite quail were exposed to PFOS in feed for 21 weeks and a variety of endpoints examined including changes in: adult body and organ weights, feed consumption rate, fertility, hatchability, and offspring survival. Mallards were exposed to PFOS at dietary concentrations of 0, 10, 50 and 150 ppm for up to 21 weeks (US EPA OPPT AR 226-1735). Due to signs of overt toxicity, adult mallards in the 50 and 150 ppm treatments were euthanized at the end of 7 and 5 weeks, respectively. At 10 ppm, there was an increase in the incidence of small testes size and decreased spermatogenesis in adult males. At 10 ppm, there were no statistically significant treatment-related effects on adult body weight, feed consumption, fertility, hatchability or offspring health or survival compared to controls. Concentrations in serum and liver at the for the 10 ppm treatment group were 87.3 µg.mL-1 and 60.9 µg.g-1 wet weight livers, respectively. No effects were observed for female mallards and offspring for the 10 ppm treatment group. For the 10 ppm treatment group, concentrations in serum and liver of adult females were 76.9 µg.mL-1 serum (at 5 weeks), 16.6 µg.mL-1 (at 21 weeks) and 10.8 µg.g-1 wet weight liver, respectively. The difference in concentration of PFOS in serum of females between 5 weeks and 21 weeks may likely reflect maternal transfer of PFOS to egg or hatchling. Concentrations in liver and serum in males are appropriate basis for developing liver-based ENEVs.

Reduction in testes size is commonly mediated by reduced circulating testosterone which is also responsible for the development and maintenance of the testicular ultrastructure, seminiferous tubule differentiation, the excurrent ducts and numerous secondary sexual characteristics including reproductive behaviour, territorial defense, and courtship singing (Mineau and Shutt, 2005). None of these potential effects were included in the AR226-1738 study. Also in this study, male mallards were housed in groups with multiple females greatly simplifying the process of attaining a female and copulating compared to birds in wild condition. As a result, effects of reduced testosterone production in exposed birds may be masked in this experimental design. While post-reproductive testicular regression does occur in male birds, there is some uncertainty about the ecological significance of this effect. From the study summary, it is difficult to ascertain how long after reproduction was completed that the birds were sacrificed, which would have an effect on the level of testicular regression. Therefore, changes in testes size and testicular regression are considered endpoints of interest, albeit with some uncertainty as to the impact on bird population.

Northern bobwhite quails were also exposed to PFOS at dietary concentrations of 0, 10, 50 and 150 ppm for up to 21 weeks (US EPA OPPT 226-1831). As in mallard, signs of overt toxicity were observed at the 50 and 150 ppm treatments and those tests were terminated at the end of 7 and 5 weeks, respectively. At 10 ppm in diet, minor overt signs of toxicity were observed in adults, there was a statistically significant increase in liver weight (females) an increase in the incidence of small testes size (males), and a statistically significant reduction in survivability in quail chicks as a percentage of eggs set (p < 0.05). The increase in the number of adult males in the 10 ppm treatment group with reduced testes size was not accompanied by any morphological change in spermatogenesis. Additionally, there were slight, but not statistically significant treatment-related reductions in fertility, and hatchability. At 10 ppm in diet, there were no PFOS-related effects on adult body weight or feed consumption. Based on these effects, reproductive effects in bobwhite quail was determined to be 10 ppm PFOS in feed, based on 21 weeks of exposure. Concentrations in serum and liver of adult females was 84 µg.mL-1 serum (at 5 weeks) 8.7 µg.mL-1 (at 21 weeks) and 4.9 µg.g-1 wet weight liver, respectively and in adult males 141 µg.mL-1 and 88.5 µg.g-1, respectively. As in mallard, the difference in concentration of PFOS in serum of females between 5 weeks and 21 weeks may likely reflect maternal transfer of PFOS to eggs.

The effect of 10 ppm in diet of birds includes effects of smaller testes size and decreased spermatogenesis in mallard and on survivability of hatchlings, increased liver weight in females and decreased testes size in males in quails. The CTV of 10 ppm is associated with concentration in serum and liver of mallard males at the end of the test of 87.3 µg.mL-1 and 60.9 µg.g-1, respectively. Concentrations in male quail serum and liver at the end of the test are comparable.

McNabb et al. (2005) studied the acute effect of PFOS on thyroid function in bobwhite and Japanese quail. Adult quail were dosed orally with 5 mg.kg-1 body weight and sampled at 7 days (bobwhite quail) or at 7 and 14 days (Japanese quail). At these sampling times in both species, plasma thyroid hormones (both T4 and T3) were decreased indicating organismal level hypothyroidism. Body weights tended to be decreased and relative thyroid weights tended to be increased, the latter effect suggesting some hypothalamic- pituitary-thyroid axis response to decreased circulating thyroid hormones. The authors commented that thyroid gland-thyroid hormone content was much less affected by PFOS treatment in bobwhite quail than would have been expected based on the degree of circulating thyroid hormone depression that was observed. In Japanese quail, thyroid gland-thyroid hormone content was decreased at 7 days of exposure but showed some recovery by 14 days of exposure compared to controls for these same times.

Acute effects in dietary studies of PFOS in juvenile mallard (Anas platyrhynchos) and bobwhite quail (Colinus virginianus) are available (US EPA OPPT AR 226-0103, AR 226-0953) and examined mortality, growth, behaviour, and feed consumption. For mallards, the 8 day dietary LC50 was 603 mg.kg-1 feed. The No-Observed-Adverse-Effect-Level (NOAEL) based on lethality was 141 mg.kg-1 feed. The NOAEL based on reductions in body weight and feed consumption was 35 mg.kg-1 feed. For bobwhite quail, the 8 day dietary LC50 was 212 mg.kg-1 feed. The NOAEL based on lethality was 70.3 mg.kg-1 feed.

Mammalian

Given the lack of ecotoxicological studies using wild species, studies on laboratory mammals were used in this assessment as surrogates for wildlife mammals. Key mammalian toxicity studies are described by Health Canada (2004). A set of CTVs for mammal (liver) and bird (serum and liver) have been selected, as summarized in Appendices 2 and 3. A CTV for mammals was selected from a 2-year dietary rat study in which histopathological effects in the liver were seen in males and females at intakes as low as 0.06-0.23 mg PFOS/kg bw per day and 0.07-0.21 mg PFOS/kg bw per day, respectively (Covance Laboratories, Inc. 2002). Average values were determined for males and females, to establish Lowest-Observed-Effect Levels (LOELs) of 40.8 mg/kg in liver and 13.9 mg/L in serum.

Further supporting evidence for a No-Observed-Effect-Concentration (NOEC) in the low mg.kg-1 or mg.L-1 range in liver and sera includes results from a two-generation rat study (US EPA OPPT AR226-0569). A 2 generation rat study with PFOS administered by oral gavage reported a NOEC at the dose 0.1 mg.kg-1 bw/day at which concentration in liver and sera were 14.4 mg.kg-1 and 5.3 mg.L-1, respectively (US EPA OPPT AR226-0569). The Lowest-Observed Effect-Concentration (LOEC) at dose 0.4 mg.kg-1 bw/day was associated with reduced dam body mass. At the LOEC the concentration in liver and sera were 58 mg.kg-1 and 19 mg.L-1, respectively.

Cynomolgus monkeys administered PFOS for 26 weeks at 3 dose levels were observed to have thymic atrophy (females), and reduced high density lipoprotein, cholesterol, triiodothyronine, total bilirubin levels (males) (Covance Labs 2002a). The LOEL dose was 0.03 mg.kg-1 bw/day at which average mean female and male concentrations in sera and liver were 19.8 µg.g-1 and 14.5 µg.mL-1, respectively.

Mode of Action

Body mass reduction or poor food efficiency was seen in most toxicity studies and species (Haughom and Spydevold 1992; Campbell et al. 1993a, 1993b; US EPA OPPT AR226-0137, AR226-0139, AR226-0144, AR226-0949, AR226-0953, AR226-0956, AR226-0957, AR226-0958, AR226-0967). This is consistent with the mechanism of toxicity being the uncoupling of oxidative phosphorylation (US EPA OPPT AR226-0167, AR226-0169, AR226-0240). This mode of action, however, is not known with certainty to explain PFOS toxicity. There are other mechanisms that can be hypothesized. A study with rats (Luebker et al. 2002) tested the hypothesis that PFOS, PFOA and other perfluorinated chemicals can interfere with the binding affinity and capacity of liver binding proteins for fatty acids; the results revealed that the most potent competitor is PFOS. A study with common carp (Cyprinus carpio) by Hoff et al. (2003) has suggested that PFOS induces inflammation-independent enzyme leakage through liver cell membranes that might be related to cell necrosis. It was also suggested that PFOS might interfere with homeostasis of DNA metabolism.

Page details

Date modified: