|Home | About | Journals | Submit | Contact Us | Français|
Perfluorooctanoic acid (PFOA) is a ligand for peroxisome proliferator-activated receptor (PPAR) α, which exhibits marked species differences in expression and function, especially between rodents and humans. We investigated the functional difference in PFOA response between mice and humans, using a humanized PPARα transgenic mouse line. Three genotyped mice, 129/Sv wild-type (mPPARα), Pparα-null mice and humanized PPARα (hPPARα) mice (8-week-old males) were divided into three groups: the first was treated with water daily for 2 weeks by gavage (control group), and the remaining two groups were treated with 0.1 and 0.3 mg/kg ammonium perflurooctanate (APFO), respectively, for 2 weeks by gavage. The APFO dosages used did not influence the plasma triglyceride or total cholesterol levels in any mouse line, but the high dose increased both hepatic lipid levels only in mPPARα mice. APFO increased mRNA and/or protein levels of PPARα target genes cytochrome P450 Cyp4a10, peroxisomal thiolase and bifunctional protein only in the liver of mPPARα mice, but not in Pparα-null or hPPARα mice. This chemical also increased expression of mitochondrial very long chain acyl-CoA dehydrogenase only in the liver of mPPARα mice. Taken together, human PPARα may be less responsive to PFOA than that of mice when a relatively low dose is applied. This information may be very valuable in considering whether PFOA influences the lipid metabolism in humans.
Perfluorooctanoic acid (PFOA) (CAS, 335-67-1), an organofluoro compound, is used in industrial surfactants, emulsifiers and numerous consumer products (U.S. Environmental Protection Agency, U.S. EPA, 2006). This chemical is very stable and incombustible at normal temperature, and when released into the environment it is not decomposed in either the ecosystem or the body. Repeated exposure to PFOA may result in accumulation in humans and wildlife (Prevedouros et al., 2006). Thus, PFOA is scheduled to be added to the list of Persistent Organic Pollutants (POPs) in the near future (World Wildlife Fund, 2005).
PFOA is thought to be resistant to clearance in humans, with a biological half-life of 3.5–4.4 years (U.S. EPA, 2002; Olsen et al., 2007). Since there are dramatic sex and species clearance (Kennedy et al., 2004), it is very difficult to evaluate the level of human risk. Several studies of the potential health risk of exposure to PFOA have been conducted. Ammonium perfluorooctanoate (APFO) causes hepatic peroxisome proliferation (Pastoor et al., 1987), liver hyperplasia, liver adenoma, pancreatic acinar cell adenomas (Biegel et al., 2001) and Leydig cell adenomas (Cook et al., 1992; Biegel et al., 2001) in rats and mice. Since PFOA is not mutagenic, the induction of these tumors most likely occurs through nongenotoxic mechanisms (Kennedy et al., 2004).
PFOA is a peroxisome proliferator-activated receptor α (PPARα) agonist in mice and rats (Ikeda et al., 1985; Rosen et al., 2007, 2008a,b), and through PPARα increases peroxisomal and mitochondrial β-oxidation of fatty acids, inhibits the secretion of very low-density lipoproteins and cholesterol from the liver, as well as reduces cholesterol and triglyceride in serum and the accumulation of lipids in the liver (Maloney and Waxman, 1999; Xie et al., 2003; Kennedy et al., 2004). Epidemiological studies have resulted in disparate findings. Olsen et al. (2003a), reported a positive association between PFOA and serum cholesterol (T-Cho) and triglycerides (TG) among participants in a fluorochemical medical surveillance program in some factories. In contrast, no such association was reported in another factory (Olsen et al., 2003b). Since there are species differences in the expression levels of PPARα, it is very difficult to accurately evaluate the exact effects of PFOA on the lipid levels of human from the experiments using mice or rats. To investigate species-specific effects and mechanisms of PPARα, PPARα-humanized (hPPARα) mice were recently established (Cheung et al., 2004). The mice express human PPARα specifically in the liver under the control of the doxycycline (Tet-OFF) system in a PPARα-null background. High-level expression of human PPARα protein achieved in this hPPARα mouse model proved comparable to murine PPARα expression in wild-type mice. This mouse model was shown to be a useful tool in determining human PPARα function in the effects of Wy-14,643 on both the serum and hepatic lipid levels (Cheung et al., 2004).
In this study, we compared the effects of relatively low doses of APFO (0, 0.1, and 0.3 mg/kg) on the PPARα and target gene expression as well as on lipid levels between mouse and human PPARα using wild-type, Pparα-null and hPPARα mice, in order to determine the possible effects of this chemical or its principal metabolite, PFOA, on the lipid metabolism in humans.
APFO was purchased from Tokyo Kasei Kogyo (Tokyo, Japan).
This study was conducted according to the Guidelines for Animal Experiments of the Nagoya University Animal Center. Three genotyped male and female mice, i.e., wild-type (mPPARα), Pparα-null (Lee et al., 1995; Akiyama et al., 2001) and hPPARαTet-Off (Cheung et al., 2004) with a Sv/129 genetic background were provided by Dr. Gonzalez (U.S. National Institutes of Health), and each genotyped male and female was mated in our animal center. All mice were housed in a temperature- and light-controlled environment (25 °C, 12 h light/dark cycle), and maintained on stock rodent chow and tap water ad libitum. At 8 weeks old, the offspring male mice of each strain were assigned to treatment groups as follows: treated with distilled water daily for 2 weeks by gavage (control group); treated with 0.1 or 0.3 mg/kg APFO, respectively, for 2 weeks by gavage (Table 1). On the day following the last dose (18–20 h later), all mice were killed by decapitation, and the blood and livers were removed. The liver samples were stored at −80 °C until use. Plasma was collected after centrifuging blood 3500×g for 10 min and stored at −80 °C until use. All samples were used in all measurements, and the measurements were done in duplicate or triplicate except for the histopathology.
A nuclear fraction was extracted from a part of the frozen liver using a CelLytic® NuCLEAR® Extraction Kit (SIGMA, Tokyo, Japan).
Each tissue was homogenized with a threefold volume of 10 mM phosphate buffer (pH 7.4) containing 0.25 M sucrose. Protein concentrations of the homogenate and nuclear fraction samples were measured using a Protein Assay Kit (Bio-Rad, Tokyo, Japan).
Liver homogenate or nuclear fraction samples for electrophoresis were subjected to 10% SDS-PAGE and transferred to nitrocellulose membranes. After blocking with 3% skim milk, each membrane was incubated with the respective primary antibody, followed by incubation with alkaline phosphatase-conjugated goat anti-rabbit IgG (Jackson Immuno Research, West Grove, PA). The primary polyclonal antibody was prepared using purified medium-chain acyl-CoA dehydrogenase (MCAD) (Furuta et al., 1981), very long chain acyl-CoA dehydrogenase (VLCAD) (Izai et al., 1992), peroxisomal thiolase (PT) (Furuta et al., 1980) and peroxisomal bifunctional protein (PH) (Osumi and Hashimoto, 1980). These antibodies have been described earlier (Aoyama et al., 1998; Nakajima et al., 2000). The primary polyclonal antibodies of PPARα and RXRα were purchased from Santa Cruz Biotechnology, Inc. (CA). Each band was quantified using the densitometry with the Lane & Spot Analyzer version 5.0 (ATTO Corporation, Tokyo, Japan).
mPPARα (Accession No. NM_011144) -, hPPARα (BC000052) -, RXRα (NM_011305) -, CYP4A10 (AB018421) -, MCAD (NM_007382) -, VLCAD (NM_017366) -, PT (NM_130864) -, PH (NM_023737) -, diacylglycerol acyltransferase (DGAT) 1 (NM_010046) - and 2 (NM_02638), p65 (XM_342346) -, p50 (AY521463) - and p52 (AF155372) -, tumor necrosis factor (TNF) α (NM_013693) -, TNF receptor (TNFR) 1 (NM_011609) -, and TNFR2 (NM_011610) -mRNA levels were monitored by the ABI PRISM 7000 Sequence Detection system (Applied Biosystems, Foster City, CA). Total RNA was isolated using a RNeasy Protect Mini Kit (QIAGEN, Tokyo, Japan). Complementary DNA (cDNA) was synthesized from 1 μg of total RNA using Oligo(dT)20 primer. RNA quantity and quality were checked by a GeneQuant II RNA/DNA Calculator (Pharmacia Biotech, Framingham, MA). The sequences for the forward and reverse primers were as follows: hPPARα forward: 5′-GCGATCTAGAGAGCCCGTTATC-3′ and reverse: 5′-GCCAAAGCTTCCAGAACTATCC-3′; RXRα forward: 5′-CAGTTGCTGCGCATCCAA-3′ and reverse: 5′-GTGCTGCTAAATAACTCTTGCATGA-3′; DGAT1 forward: 5′-GGCGGTCCCCAACCAT-3′ and reverse 5′-GCTCTGCCACAGCATTGAGA-3′; DGAT2 forward: 5′-TGGCCTGCAGTGTCATCCT-3′ and reverse: 5′-GGGCGTGTTCCAGTCAAATG-3′. The sequences of other genes have been described elsewhere (Ito et al., 2007; Ramdhan et al., 2008).
The PCR mixture (25 μl) contained 1×SYBR Green PCR master Mix (Applied Biosystems), 0.1 μM of each primer, and cDNA diluted by TE including 1 μg/ml transfer-RNA. PCR amplification was as follows: an initial step for 2 min at 50 °C, then 10 min at 95 °C, followed by 50 cycles at 95 °C for 15 s and at 60 °C for 1 min. Each mRNA level was normalized to the level of GAPDH (BC096440) mRNA expression in the same preparation.
Lipid from livers was extracted using the method of Folch. TG and T-Cho in the liver and plasma were measured using kits of TG-IE and T-Cho IE (Wako, Osaka, Japan), respectively.
The organs fixed in 10% neutral buffered formalin were embedded in paraffin and sliced into 4 μm sections. Tissue sections of the livers were stained with hematoxylin and eosin, and examined under a light microscope using the BZ-8000 (Keyence, Osaka, Japan). Necrotic cells were measured in a blinded fashion according to the method of Soni et al. (1998). Inflammatory cells, and lipid accumulation were measured in a blinded fashion referring to the methods of Brunt et al. (1999) and Okiyama et al. (2009): (1) grade of steatosis: 0, none; 1, mild (5–33% of parenchymal involvement by steatosis); 2, moderate (33–66%); 3, severe (>66%); (2) intra-acinar (lobular) and portal tract inflammations: 0, none; 1, mild; 2, moderate; 3, severe.
Plasma ALT was measured using a Transaminase C II Test kit purchased from Wako (Osaka, Japan).
Comparisons were made using two-way analysis of variance (ANOVA) and the Tukey–Kramer HSD post hoc test among each treated group of each kind of genotyped mice, or also between mPPARα or Pparα-null mice and hPparα mice with the same dose. Values of P < 0.05 were considered to indicate statistical significance.
Body and liver weights of hPPARα mice were lower than those of mPPARα and Pparα-null mice, but APFO treatment did not influence the increase in the body weight of all genotyped mice (Table 1). The low dose (0.1 mg/kg) of APFO did not influence either absolute or relative liver weights in any mouse line, while the APFO high dose (0.3 mg/kg) increased the liver weight and the ratio of liver/body weight only in mPPARα mice.
In all groups, the expressions of PPARα protein and mRNA were higher in hPPARα mice than in mPPARα mice (Figs. (Figs.11 and and2).2). APFO treatments did not influence either PPARα mRNA or protein levels in any of the mice regardless of genotype, nor did they influence RXRα, which forms a heterodimer with PPARα in the nucleus, protein and mRNA levels, though a significant increase was observed in RXR-mRNA in hPPARα mice between 0.1 and 0.3 mg/kg APFO exposure.
We then analyzed the effects of APFO treatment on the induction of the PPARα target genes, PH, PT, VLCAD, MCAD and CYP4A10. Among these genes, PH was the most sensitive to APFO treatment; high dose APFO significantly increased PH mRNA and protein expression in mPPARα mice (Figs. (Figs.11 and and2).2). However, APFO increased neither PH mRNA nor protein levels in Pparα-null or hPPARα mice. APFO treatment influenced PT-mRNA and protein levels similarly to PH in mPPARα mice. A high dose of APFO increased mitochondrial VLCAD-mRNA and protein only in mPPARα mice, but the increases seemed to be less than those of the peroxisomal enzymes, PH and PT. MCAD-mRNA and protein levels were slightly increased only in hPPARα mice. CYP4A10-mRNA levels were not detectable in Pparα-null mice. In contrast, CYP4A10 expression was clearly detected in the livers of mPPARα and hPPARα mice (Fig. 1). APFO, however, clearly induced mRNA in the former, but not in the latter.
When the above mentioned gene expressions were compared at the same dosages, there were no significant differences in PH- and MCAD-mRNA and protein expressions among controls of three genotyped mice. In contrast, a significant difference was observed in the expression of VLCAD-mRNA and protein between the control mPPARα or hPparα mice and Pparα-null mice, whereas that of CYP4a10-mRNA was higher in hPparα mice than in mPPARα or Pparα-null mice. PT-mRNA expression was significantly higher in hPparα mice than mPPARα and Pparα-null mice, while the protein expression was higher in mPPARα and hPparα mice than Pparα-null mice.
APFO treatments did not increase hepatic NF-κB subunits, p65-, p50- and p52-mRNAs, TNFα and the receptors, suggesting that the dose used could not cause inflammation (data not shown).
TG and T-Cho are typically decreased upon PPARα activation (Peters et al., 1997). TG and T-Cho in the liver and plasma were measured to investigate whether enzymes involved in lipid metabolism were induced by APFO treatment (Fig. 3). Although plasma TG levels in Pparα-null mice and T-Cho levels in Pparα-null and hPparα mice were significantly higher than those in mPPARα mice, APFO did not influence the levels in any mouse lines (Fig. 3(A)). In hPparα mice, plasma T-Cho levels of all dosage groups were higher than those of the mPPARα group.
In the liver, TG and T-Cho levels were also significantly higher in Pparα-null mice than in mPPARα mice (Fig. 3(B)). Interestingly, a high dose of APFO significantly increased hepatic TG and T-Cho levels only in mPPARα mice. In Pparα-null mice, hepatic TG levels of all dosage group were higher than those of the mPPARα group.
We also measured mRNA levels of DGAT1 and 2 that are involved in TG synthesis. APFO treatments did not influence expression of these enzymes in any group (data not shown).
In the control animals, there were no obvious changes in the grade of steatosis, and the inflammatory cells in the lobular and portal area of liver in all genotyped mice, though little or no lipid accumulation was seen in the liver of PPARα-null mice (Fig. 4, score grade of data not shown). A high APFO dose did not cause any histopathological changes in liver except for the quite mild hepatocellular enlargements seen in mPPARα mice and hPPARα mice, and appeared to increase cytoplasmic vacuoles morphologically consistent with the lipid accumulation in zone 2 of hepatic lobule from Pparα-null mice as pointed out by Wolf et al. (2008a).
APFO treatment did not affect the ALT levels in any group (data not shown).
In general, administration of peroxisome proliferators to rodents increases liver weight and the levels of PPARα target genes such as PT, PH, VLCAD (Aoyama et al., 1998; Nakajima et al., 2000) and CYP4A10 (Ramdhan et al., 2008). The present study clearly shows that a relatively low dose of APFO increased liver weight and induced these genes only in mPPARα, but not in Pparα-null or hPPARα mice. These results were quite distinct from an earlier report that another mPPARα ligand, Wy-14,643, induced some of these genes not only in mPPARα mice but also in hPPARα mice (Cheung et al., 2004). One reason for this discrepancy may lie in the different dosage used; the dose of Wy-14,643 was roughly estimated to be 0.37 mmol/kg when mice weighed 25 g and were consuming 3 g feed a day, whereas the dose of APFO (0.3 mg/kg) was calculated to be 0.70 μmol/kg in our study. Indeed, in another study using hPPARα mice, 1 mg/kg APFO, which is about 3 times higher than that in the current study, clearly activated human PPARα judging from the induction of the target genes (in preparation). Recently, Abbott and colleagues measured the lowest effective concentrations of PFOA at which mouse and human PPARα were activated in an in vitro system and showed them to be 10 μM and 30 μM, respectively (Takacs and Abbott, 2007). Wolf et al. (2008b) also reported that activation of mouse PPARα by APFO was generally higher compared to that of human PPARα. Taken together, hPPARα may have a weaker affinity for PFOA than does mPPARα. Foreman et al. investigated the activation of mouse PPARα and human PPARα, by perfluorobutyrate, which has a similar structure to PFOA and is a PPARα agonist, and showed that dosages higher than 35 mg/kg caused activation in both wild-type and humanized mouse strains, and this was the same hPPARα strain as used in the present study. Thus, when a relative high dosage of APFO or perfluorobutyrate is used, PPARα activation may be observed not only in mPPARα mice but also in hPPARα mice.
Expression of PPARα protein was thought not to differ between mPPARα and hPPARα mice, although the mRNA level was greater in the latter than in the former (Cheung et al., 2004). In the current experiment, expression of PPARα protein and mRNA appeared to be greater in hPPARα than in mPPARα mice. Nevertheless, the fact that PFOA-activated mPPARα but not hPPARα suggested that the functional activation of hPPARα itself may also be weak compared to that of mPPARα.
As noted above, although APFO-induced peroxisomal and mitochondrial lipid β-oxidation enzymes in mPPARα mice, it did not reduce TG and T-Cho levels in plasma, and only slightly increased levels in the liver. In contrast, APFO treatment did not increase either of these enzyme levels or plasma and hepatic lipid levels in hPPARα mice. Loveless et al. (2006) reported that the same dose of APFO as the one used in the current study did not affect the plasma TG levels. On the other hand, Kudo et al. (1999) showed that 0.02% APFO in the diet for a week (which corresponded to 20 mg/kg/day for rats with 200 g body weight) reduced plasma TG levels, but increased hepatic TG levels of rats. Similar results were also obtained in our recent study using 1 and 5 mg/kg APFO in mice. In such case, APFO exposure dose-dependently reduced plasma TG level but slightly increased TG in livers of mPPARα mice. Thus, a relatively high dose or long exposure duration might influence both plasma and hepatic TG levels. In the study by Kudo et al., APFO increased the expression levels of glycerol 3-phosphotransferase and DGAT, which are involved in TG synthesis and are thought to induce an increase in hepatic TG levels (Kudo et al., 1999). The expression of these genes may be important for their influence on hepatic TG levels. In the current study, however, APFO affected neither DGAT1 nor DGAT2 expression. After APFO treatment, TG increase in the liver was seen only in mPPARα mice, suggesting that some PPARα target gene(s) might also be involved in the increase. Since fatty acid transport protein is a PPARα target gene (Martin et al., 1997) involved in lipid transport to the liver, the involvement of this protein in TG accumulation should be also be investigated.
In control animals, plasma T-Cho levels were reported to be greater in Pparα-null mice than in mPPARα mice (Peters et al., 1997), a finding which was reconfirmed in the present study. It is thought that endogenous PPARα ligands such as leukotriene B4 and palmitic acid could activate PPARα (Desvergne and Wahli, 1999), which might result in increased cholesterol metabolism, followed by a reduction of T-Cho levels in mPPARα mice. Since human PPARα may be less responsive to these endogenous fatty acids, similar to PFOA, it might contribute to higher levels of plasma T-Cho in hPPARα mice.
Although we did not measure the serum PFOA concentration in mice, Loveless et al. (2006) reported that the serum concentration in mice after a 0.3 mg/kg administration of APFO for 14 days was 13 ± 2.4 μg/ml. In our experiment, we administered the same concentration of APFO for the same period, so that the serum PFOA level might be comparable with those reported by Loveless et al. On the other hand, the mean blood concentration of PFOA among workers engaged in its production was 2.21 μg/ml (95% confidence interval, 1.66–2.77 μg/ml) (Olsen and Zobel, 2007). Because durations of exposure to PFOA differ considerably between animal and epidemiological studies, and half-life of PFOA is 12 days in ddY male mice (Kudo and Kawashima, 2003), which is significantly lower than that of human (U.S. EPA, 2002; Olsen et al., 2007), mouse and human PFOA levels could not be directly compared. Nevertheless, both PFOA levels were not markedly different. Since 0.3 mg/kg APFO did not influence either plasma TG or T-Cho levels and did not affect expression of PPARα and its target genes in hPPARα mice, blood PFOA levels in the workers may not induce the functional activation of PPARα nor alter lipid levels in their blood. The finding that the expression level of PPARα was 10 times lower in humans than in mice or rats (Tugwood et al., 1996; Palmer et al., 1998) may further support our conclusion as noted above. Indeed, although Olsen et al. (2003a) found a relationship between serum PFOA and the TG or T-Cho levels in the workers in two factories, they did not find such an association in another factory (2003b). Then, they re-analyzed the data obtained in two studies (2003a, b) excluding subjects who took antilipemic agents, and reported that serum PFOA concentrations were not associated with T-Cho in the workers (2007). On the other hand, a positive association was seen in the analysis between serum PFOA concentration and TG levels, but they concluded that this association might have been due, at least partially, to residual confounding as a consequence of the dispro-portionate number of employees in the two factories with higher PFOA concentrations than in a remaining factory (Olsen and Zobel, 2007).
Here, we focused on an interaction between PFOA and PPARα in relation to the lipid metabolism. However, Rosen et al. (2008b) demonstrated that PFOA could also activate constitutive androstane receptor, which also has an important role in the lipid metabolism (Maglich et al., 2009). Further study may be required to determine whether or not the dosages used in the current study activate CAR.
In conclusion, the 0.3 mg/kg dose of APFO significantly activated the function of PPARα in wild-type mice, but not in humanized PPARα mice. In addition, APFO increased the hepatic TG and T-Cho in mPPARα mice, but not in hPPARα mice. Since human PPARα expression in hPPARα mice was equal to or even higher than that of mice, the function of human PPARα was significantly lower than that of mice. Thus, PFOA levels such as 1.66–2.77 μg/ml in serum may not influence serum TG or T-Cho levels in human.
We wish to thank Mr. Nakagawa, Department of Occupational and Environmental Health, Nagoya University Graduate School of Medicine for his valuable assistance. This study was supported in part by Grants-in-Aid for Scientific Research from the Japan Society for the Promotion of Science (B. 14370121 and B. 17390169).
Conflict of interest None.