Progress toward understanding the bioaccumulation of perfluorinated alkyl acids


Address correspondence to

The discovery of the global distribution and biomagnification of perfluorooctane sulfonate (PFOS) in wildlife [1] as well as the presence of perfluorooctanoate (PFOA) and PFOS in humans [2] came as a surprise to many bioaccumulation scientists [3]. These chemicals were nonvolatile anions at neutral pH, and the prevailing wisdom was that apart from a few organometallics such as methylmercury and tributyltin, only recalcitrant neutral halogenated organics such as polychlorinated biphenyls (PCBs) would biomagnify.

Further monitoring revealed that, while PFOS was the predominant perfluoroalkyl substance in most biotic samples, there were also quantifiable perfluoroalkyl sulfonates (PFSAs) with 4 to 10 carbons [4, 5], and PFOA was generally only detected near detection limits. Unexpectedly, a study on the fate of polyfluoroalkyl and perfluoroalkyl substances (PFASs) after a spill of aqueous film forming foam on a small creek revealed the presence of a range of perfluorocarboxylates (PFCAs) with 5 to 14 carbons in fish liver, with perfluorodecanoate predominating [6].

The source of all of the PFCA/PFSAs being measured in environmental and human samples at the time was puzzling. The production of PFOS and PFOA was relatively small compared with the total production of PFASs in commerce such as those based on perfluoroalkylsulfonamide, perfluoroalkyl phosphates/phosphonates, and corresponding polyfluoro-telomer–based compounds and their polymers (e.g., fluorotelomer-based polymer and phosphate surfactants) [7].

Therefore, the bioaccumulation properties of the PFASs, and the potential contribution to body burdens of precursor compounds that might degrade to form PFCAs and PFSAs, emerged as important research and risk assessment questions [3].

We set out to study the bioconcentration and dietary bioaccumulation of a suite of PFCAs and PFSAs to address the question of how their bioaccumulation potential varied with fluorinated chain length. The results were reported in 2 papers published in Environmental Toxicology and Chemistry in 2003 [8, 9]. No published studies existed on fish bioaccumulation of PFCAs and PFSAs, and we were concerned about the unusual properties of these chemicals such as the ability to bind to glass surfaces as well as possible contamination from polytetrafluoroethylene (e.g., Teflon) polymers. We chose a relatively standard experimental design for both studies involving flow-through exposures of juvenile rainbow trout (Oncorhynchus mykiss), a well-known test species. The bioconcentration study used plastic-lined aquaria to limit glass adsorption of the test chemicals. Fish were exposed for 12 d to a mixture of 8 PFCAs (C5–C14) and 3 PFSAs (C4–C8), followed by a 33-d depuration period in clean water. Dietary accumulation was studied separately in a 34-d uptake period on food spiked with the same chemicals followed by a 41-d depuration phase using clean food. Analysis of the PFCAs and PFSAs in the fish tissues was also challenging because very few electrospray liquid chromatography/mass spectrometry instruments were available in academic or government laboratories at the time, and also because no mass-labeled analytical standards existed to control for matrix effects. We used perfluorononanoate (PFNA) as an internal standard, which worked well but sacrificed bioaccumulation data for this key PFCA.

We found that PFCAs and PFSAs with perfluoroalkyl chain lengths shorter than 7 and 6 carbons, respectively, had insignificant bioconcentration factors (BCFs) and low bioaccumulation factors (BAFs) in rainbow trout; that is, levels were below detection limits in all tissues (liver, muscle, blood). However, both BCFs and BAFs increased with increasing length of the perfluoroalkyl chain. Bioconcentration factors increased by a factor of 8 for each additional carbon in the perfluoroalkylchain between 8 and 12 carbons, but this relationship deviated from linearity for the longest PFCA tested, possibly because of decreased gill permeability. Sulfonates had greater BCFs and BAFs, half-lives, and rates of uptake than the corresponding carboxylate of equal perfluoroalkyl chain length, indicating that chain length was not the only determinant of PFA bioaccumulation potential and that the acid functional group must also be considered.

The BCF data identified perfluorotetradecanoate as the most bioaccumulative of the suite of PFCA/PFSAs with BCF of 23 000 and BAF = 1 based on concentrations in the fish carcass (liver and gastrointestinal tract removed from the fish). None of the chemicals had BAFs statistically greater than 1, which demonstrates the limitations of short-term laboratory bioaccumulation exposures with small fish for assessing biomagnification. Many subsequent field studies have shown that C9–C14 PFCAs as well as C6–C10 PFSAs have whole-body predator/prey biomagnification greater than 1 [10, 11]. Also, many studies of marine and freshwater food webs have shown trophic magnification factors, which represent an average biomagnification factor, were greater than 1 for C8 to C12 PFCAs, PFOS, and perfluorooctanesulfonamide (FOSA) in food webs that include homeotherms as top predators [11]. Trophic magnification factors greater than 1 for C8 to C12 PFCAs and PFOS have also been demonstrated in the lichen-caribou-wolf terrestrial food web [12].

Recent bioconcentration measurements of PFCAs in carp (Cyprinus carpio) [13] have largely confirmed the BCFs reported by Martin et al [9] and demonstrated that long chain PFCAs are highly bioaccumulative. The assessment of aquatic bioaccumulation of PFASs using laboratory exposures has moved on to the precursor chemicals, including 8:2 fluorotelomer alcohol and FOSA [14], 8:2 fluorotelomer acrylate [15], and perfluorophosphonates and perfluorophosphinates [16]. The results of these studies have underlined the rapid biotransformation of precursors and the persistence of the PFCA and PFSA terminal metabolites in both fish and mammals (Butt et al., Duke University, Durham, NC, USA, unpublished manuscript). For example, studies with the 8:2 fluorotelomer alcohol universally show metabolism to PFOA, and, to a more limited extent, PFNA and lower-chain-length PFCAs.

Progress has also been made on understanding the fundamental aspects of bioaccumulation of perfluoroalkyl anions. The inclusion of protein interactions has been shown to predict tissue-specific PFCA/PFSA bioconcentration [17] and food web biomagnification [18]. Association of PFCA/PFSAs with phospholipids rather than just proteins and carbohydrates has been identified as a potentially important contributor to bioaccumulation of these anionics [19].

Did this research have an impact on the regulatory and industry decisions related to PFASs? Decisions on PFOS-related chemistry had largely been made by the time our bioaccumulation papers were published because the 3M Company announced a phaseout of PFOS and PFOA chemistry in 2001, and the US Environmental Protection Agency (USEPA) issued a significant new use ruling in 2002 banning future uses of PFOS-related chemistry in the United States [20]. However, the observation of low aquatic bioaccumulation potential for perfluorobutane sulfonate in our studies contributed to the assessment of the replacement chemistry adopted by 3M. The work was also cited in the Stockholm Convention risk profile for PFOS [21]. The many exemptions for PFOS under the Stockholm Convention (listed in Part III of Annex B of the Convention) resulted in PFOS and related compounds remaining in global commerce with significant production in China and Brazil.

Our work on PFCAs, along with that of many other research groups, contributed to the science underlying decisions related to long-chain polyfluorinated chemicals. In 2006, the USEPA announced a stewardship agreement with 8 leading fluorotelomer manufacturers to reduce emissions and product content of PFOA and related chemicals by 95% by 2010 and to work toward their elimination by 2015 [22]. Fluorotelomer manufacturers have announced products based on short-chain (C4, C6) perfluoro chains [23]. The United Nations Environment Program Strategic Approach for International Chemicals Management includes an initiative to reduce emissions of long-chain “fluorochemicals” and encourages use of short-chain perfluoro chemistry globally [24].

Many gaps in our knowledge of bioaccumulation of perfluoro- and polyfluoroalkyl substances remain. Limited laboratory or field bioaccumulation data are available for a wide range of PFASs, particularly for terrestrial food webs. The bioaccumulation and biotransformation of PFAS polymers is also of interest given the large quantities produced. For modeling of PFCA/PFSA bioaccumulation, additional information is needed on the apparent octanol–water and membrane–water partition coefficients, as well as protein–water and protein–air partition coefficients [17-19].


Table S1. (49 KB PDF).