We obtained a specific waiver from the Institutional Review Board of Kyungpook National University for the experimental use of amphibians or reptiles in Korea (no. KNU-2013-222). All the members of our research group attended educational and training courses on the appropriate care and usage of experimental animals. Adult Xenopus laevis were entrained in 12-h light/dark cycles at 18°C in containers provided by the Institutional Review Board of Kyungpook National University constructed to specifications for laboratory animal maintenance. There were no unexpected deaths of adult Xenopus during the study.

PFHxA and PFHpA were purchased from Sigma-Aldrich (St. Louis, MO, USA) and dissolved in dimethyl sulfoxide (DMSO) to produce 1 M stock solutions. The FETAX medium used in the experiment contained 10.7 mM NaCl, 1.14 mM NaHCO₃, 0.4 mM KCl, 0.1 mM CaCl₂, 0.35 mM CaSO₄.2H₂O, and 0.3 mM MgSO₄ and the levels of DMSO did not exceed 0.15%, which was below the permissible level for this assay. PFC stocks were freshly made and diluted in FETAX medium.

Experiments on Xenopus laevis were conducted in accordance with documented standards of the Animal Care and Use Committee, in agreement with international laws and policies [National Institutes of Health (NIH) Guide for the Care and Use of Laboratory Animals, NIH publication no. 85–23, 1985]. Adult Xenopus were purchased from Nasco (Fort Atkinson, WI, USA) and kept in climate-controlled clear plastic aquariums with dechlorinated tap water at 18±2°C, with 12-h light/dark cycles. The housed frogs were fed only three times a week. For the induction of ovulation, human chorionic gonadotropin (1,000 IU) (Sigma-Aldrich) was injected under the skin of a female Xenopus in the evening. The following day, the females were made to lay eggs in 60-mm plastic dishes. The eggs were immediately fertilized in 0.1X modified Barth's saline (MBS) containing 88 mM NaCl, 5 mM Hepes, 2.5 mM NaHCO₃, 1 mM KCl, 1 mM MgSO₄, and 0.7 mM CaCl₂ (pH 7.8) after washing three times with 0.1X MBS. The frogs were sacrificed as follows: The anesthetic of choice for male frogs, tricaine methanesulfonate (MS222) was dissolved in dechlorinated water in an induction tank at a dose of 500 mg of MS222/l of water. Xenopus testes were obtained from the sacrificed males and minced in 1 ml of cold 1X MBS. Following successful fertilization, the jelly coat was removed from the embryos by gentle swirling in a 2% L-cysteine solution. The embryos were then transferred to 1X MBS containing 3% Ficoll 400 (GE Healthcare, Little Chalfont, UK). Unfertilized eggs and dead embryos were removed and the remaining embryos were maintained at 22±0.5°C until they reached blastula stage 8.5.

FETAX assays were conducted to assess the developmental toxicity and teratogenic effects of the PFCs based on the American Society of Testing Material (ASTM) guide (ASTM E1439-98). Finely cleaved embryos in the blastula stage 8.5 were selected and used to exclude the effects of spontaneous embryonic developmental problems. Embryos were impartially allocated to 100-mm petri dishes (20–25 embryos/dish) and exposed to different concentrations of PFHxA (0.1, 0.5, 1, 1.5 and 2 mM) and PFHpA (0.25, 0.5, 0.75, 1 and 1.25 mM). DMSO (0.1%) and FETAX medium alone were used as the controls. Embryos were incubated at 23°C until the end of the assay, and the media were changed daily and dead embryos were removed. At the end of the experiment, embryo mortality was recorded and the surviving embryos were fixed in 4% formaldehyde to determine whether there was any malformation. Head-tail lengths and malformations were evaluated under a light microscope and images were analyzed using Axiovision software version 4.8 (Carl Zeiss, Munich, Germany) to measure growth inhibition.

Xenopus embryos were prepared for whole-mount in situ hybridization using a previously described method (20), with minor modification. Whole-mount in situ hybridization was conducted at the tadpole stage (stage 34–36) using digoxigenin-labeled RNA probes and BM purple staining. The probes were prepared in the following manner: RT-PCR products of phosphotyrosine-binding protein (xPTB; a liver marker) and NKX2.5 (heart marker) were sub-cloned into a T-easy vector (Promega, Fitchburg, WI, USA), linearized with ApaI and NcoI, respectively, and transcribed using a SP6 RNA polymerase kit (Invitrogen, Carlsbad, CA, USA).

RNA was isolated from stage 34–46 embryos using TRIzol reagent, according to the manufacturer's instructions (Invitrogen), and quantified using an Optizen 320 UV spectrometer (Mecacys, Deajeon, Korea). An aliquot of RNA (1 g) was reverse transcribed using a PrimeScript First-Strand cDNA Synthesis kit, according to the manufacturer's instructions (Takara, Shiga, Japan). Reactions containing template RNA without reverse transcriptase were prepared as the negative controls (no-RT) and carrier was used as the loading control. cDNA was amplified using specific PCR primer sets (Table I) and 2X Emerald Amp PCR Master mix (Takara). The PCR products were separated in a 1–1.2% agarose gel and visualized on a transilluminator (Vilber Lourmat, Marne-la-Vallée, France) following ethidium bromide staining. The data were representative of at least three experiments. The Primer3 program was used to design primers and ODC was used as an internal control.

Twenty-five embryos from each group were exposed to PFHxA and PFHpA, respectively, for 10 days. The surviving tadpoles were fixed in 4% formaldehyde for 2 h at room temperature. Tadpoles (3–5) were randomly selected from each group and their livers removed under a light microscope (Zeiss Stemi 2000-c; Carl Zeiss). Images were captured and analyzed to monitor changes in the liver size.

Data for the three endpoints from each FETAX assay were analyzed using GraphPad Prism software (GraphPad Software, Inc., La Jolla, CA, USA). Images from the RT-PCR and whole-mount in situ hybridization assays were analyzed by ImageJ software (NIH). Statistical analyses were conducted using SPSS 13.0 software (version 18; SPSS, Inc., Chicago, IL, USA).

FETAX assays were performed to determine the potential teratogenicity and developmental toxicity of PFCs. FETAX assay results were expressed relatively in terms of the teratogenic index (TI)=LC50/EC50, where LC50 was the concentration that at which 50% of the embryos were killed and EC50 was the concentration at which 50% of the embryos were malformed. The minimum concentration that inhibits growth (MCIG) was also determined. Compounds were considered toxic when TI was ≥1.2 (17).

We performed an initial FETAX assay to determine whether PFHxA and PFHpA had embryonic lethal or terato-genic effects. This result confirmed that PFHxA and PFHpA were potential teratogens and developmental toxicants. The LC50 value for PFHxA in Xenopus embryos was 1523.5 µM, while the EC50 value was 1046.5 µM, and the MCIG value was 412.3 µM, which produced a TI value of 1.43, indicating that this compound was a potential teratogen and developmental toxicant. The LC50 value for PFHpA was 942.4 µM in the Xenopus embryos, with an EC50 of 754.9 µM and an MCIG value of 297.4 µM, producing a TI value of 1.25. Thus, this compound was also considered a teratogen and developmental toxicant (Table II). PFHxA had a significantly higher TI value (TI=1.43) than PFHpA (TI=1.25). Furthermore, the LC50 and EC50 values were higher for PFHxA, which had a shorter fluorinated carbon chain than PFHpA.

Our FETAX assay results confirmed that PFHxA and PFHpA are potential teratogens and developmental toxicants. We then investigated the effects of different concentrations of PFHxA and PFHpA on mortality and malformation in developing embryos. As shown in Fig. 1B, exposure to PFHxA or PFHpA greatly increased the mortality and general malformation rates in the developing embryos. These concentration-dependent effects on mortality and malformation were more severe as the number of carbons increased. PFHpA significantly affected mortality and malformation rates within a narrow concentration range (Fig. 1B).

To investigate the mechanism underlying this toxicity, we assessed the phenotypes of embryos exposed to PFHxA and PFHpA. PFHxA and PFHpA were not highly toxic until early stage 26, as no significant differences were observed in the external phenotypes. PFC exposure induced certain embryo nic abnormalities that were observed at the end of the study, such as pericardial edema, dorsal fin blisters, a curved body axis, lack of eye pigment, and disturbed facial cells (Fig. 2A). PFHxA-treated embryos developed mild and general malformations including gut mis-coiling and improper eye shape, whereas PFHpA induced more severe and diverse malformations (Fig. 2A). In agreement with this, the whole body length was reduced by 14 and 24% in tadpoles treated with 1,500 µM of PFHxA or 1,000 µM of PFHpA, respectively (Fig. 2B). These results indicated that embryos exposed to PFHxA or PFHpA show a range of malformations and a reduced whole body length, which is a clear indicator of toxicity while PFHpA has slightly more severe defects compared to PFHxA.

We examined stage 36 tadpoles that had been continuously exposed to 130 µM PFHxA or PFHpA using whole-mount in situ hybridization. The liver is a major metabolizing organ that detoxifies many compounds and is therefore vulnerable to the toxic effects of PFCs. We also measured a liver-specific marker, xPTB, which is expressed in the diverticulum region of the embryonic liver and plays an important role in low-density lipoprotein (LDL) receptor endocytosis in the liver (21). Livers of embryos treated with PFHxA developed abnormally and were unusually large (Fig. 3A). We also observed enlarged livers in tadpoles exposed to PFHxA although these effects were more pronounced in the presence of PFHpA (Fig. 2A). These results indicated that PFHxA and PFHpA induced liver defects and should be regarded as developmental toxicants and teratogens.

The heart-specific marker, NKX2.5, was used to examine the cardiovascular effects of PFC exposure. The amphibian heart is composed of one ventricle, surrounded by atria on both sides, separated by a septum (22). Embryos treated with PFHxA and PFHpA showed enlarged atria and a loss of the atrial septum (Fig. 3B). Hearts from embryos exposed to either PFHxA or PFHpA were smaller and contained no trabeculae. Additionally, the pericardiac cavity was expanded and the atrial and ventricular walls were thinner than those observed in the control embryo hearts, which contained prominent trabeculae and the atria and ventricle were separated (Fig. 3B). This effect was more prominent in embryos exposed to PFHpA, which reduced trabeculation of the ventricle as well as the thickness of the atrial and ventricular walls (Fig. 3B). The results showed that PFHxA and PFHpA treatment caused abnormal heart development.

To investigate the effects of PFC exposure on the liver and heart further, the mRNA expression levels of three liver, heart, and intestine markers were determined by RT-PCR analysis. Previously, PFCs were shown to be tissue-specific toxicants (14,23,24). For this study, we exposed stage 36 embryos to 130 µM PFHxA or PFHpA, a concentration that induced the development of various malformations, rather than embryonic lethality. As shown in Fig. 2C, mRNA expression of xPTB was considerably reduced in PFHpA-treated embryos, as compared to the control or PFHxA-treated embryos (Fig. 3A and C). Previous findings have also shown reduced xPTB expression following exposure to various PFCs (16,25,26). Furthermore, the heart-specific marker, NKX2.5, was used to examine the effects of PFCs on the heart. As shown in Fig. 2B and C, mRNA expression of the heart-specific marker, NKX2.5, was reduced in the PFHpA-treated embryos (Fig. 3B and C). NKX2.5 is critical for vasculogenesis and angiogenesis (27). We also assessed the mRNA levels of an intestinal marker, Cyl18, which is mainly expressed in the presumptive small intestine of embryos at the tailbud stage. As shown in Fig. 2C, we observed no significant effects of either PFHxA or PFHpA exposure on Cy118 expression. These results suggested that PFHpA negatively affected liver and heart development (Fig. 3C).

The roles of peroxisome proliferator-activated receptor-γ (PPARγ), ERK, and JNK pathways in mediating the developmental toxicity associated with exposure to PFCs remain to be elucidated. To investigate the effects of PFCs on PPARγ, ERK, and JNK pathways, stage 36 embryos were exposed to 130 µM PFHxA or PFHpA and then analyzed by western blotting with anti-PPARγ, phosphorylated-ERK, ERK, p-JNK, JNK, and β-actin antibodies. PFHpA strongly induced ERK and JNK phosphorylation while PPARγ expression remained the same, which in turn activated signaling that caused liver toxicity, known as hepatocarcinogenesis (Fig. 4). These results clearly demonstrated that the observed developmental defects in the liver were associated with the PFHpA-mediated activation of ERK/JNK pathways.

Hepatomegaly is often associated with chronic inflammatory liver disease. The liver is the major metabolizing organ and it is targeted following PFC exposure (28). Previous findings have shown that exposure to PFCs caused various chronic abnormalities, including hepato-cellular hypertrophy and hepatomegaly (29,30). To investigate the effects of PFHxA and PFHpA on the liver, livers from 10-day old tadpoles that had been continuously exposed to 130 µM PFHxA or PFHpA exhibited hepatomegaly (Fig. 5). These results suggested that high levels of exposure to PFHpA or PFHxA may induce hepatic disorders.