Chem. Res. Toxicol. 1996, 9, 689-695
689
Influence of Carbon Chain Length on the Hepatic Effects of Perfluorinated Fatty Acids. A 19F- and 31P-NMR Investigation Carol M. Goecke-Flora and Nicholas V. Reo* Department of Biochemistry and Molecular Biology, Department of Physics/Wright State University Magnetic Resonance Laboratory, Wright State University, Cox Institute, Dayton, Ohio 45429 Received December 29, 1995X
Using nuclear magnetic resonance (NMR) spectroscopy, we investigated the importance of carbon chain length with regard to the hepatic effects associated with perfluoro-n-carboxylic acids. Male F-344 rats were administered a single intraperitoneal dose of either perfluoron-heptanoic acid (C7-PFA), perfluoro-n-nonanoic acid (C9-PFA), or perfluoro-n-undecanoic acid (C11-PFA). Data from previous studies involving perfluoro-n-octanoic acid (C8-PFA) and perfluoro-n-decanoic acid (C10-PFA) are included for comparison. Food consumption/body weight was monitored daily for all groups. C9- and C11-PFA treatment yields a prolonged hypophagic response while C7-PFA shows a more acute response. Fluorine-19 NMR spectra of urine and bile samples show no evidence of fluorometabolites and suggest that the distribution of perfluorocarbons into urine or bile is dependent upon carbon chain length. The aqueous solubility of C7-PFA appears to facilitate rapid urinary excretion, similar to that observed for C8-PFA. The relative hydrophobicity of C9- and C11-PFA appears to favor biliary enterohepatic recirculation, yielding a more protracted toxicity, similar to C10-PFA. Phosphorus-31 NMR studies of liver in vivo and liver extracts show that perfluorocarbons of gC9 carbons produce a significant increase in liver phosphocholine concentration. These data are discussed with regard to the impact of these chemicals on hepatic phospholipid metabolism. Hepatic peroxisomal fatty acyl CoA-oxidase activity (FAO) was measured to determine if C7-, C9-, and C11-PFA are peroxisome proliferators. Data indicate that the induction of peroxisomal enzyme activity by perfluorocarbons requires a chain length greater than seven carbons. In general, these results demonstrate the significance of carbon chain length in the hepatotoxic response and provide clues toward understanding the processes involved in the biological activities associated with exposure to these compounds.
Introduction Peroxisome proliferators (PPs)1 are a structurally diverse class of compounds that have therapeutic and commercial utility (1). These include hypolipidemic drugs, phthalate ester plasticizers, and various halocarbon chemicals used as lubricants, solvents, and surfactants. Included in this latter group are perfluorocarboxylic acids and related neutral compounds which are known to be metabolized to analogous n-carboxylic acids (2). Such chemicals have been shown to cause a variety of hepatic toxicological effects in rodents. Most notably, these include hepatomegaly (3), pronounced hypophagia, marked body weight loss (4), induction of P450 enzymes (5-7), and proliferation of liver peroxisomes (3, 8-10). Additionally, chronic exposure to PPs results in hepatocellular neoplasms in rodents (3, 10-14). The relationship between these toxicological effects, peroxisome * Correspondences and request for reprints should be addressed to: Nicholas V. Reo, Ph.D., Wright State University Magnetic Resonance Laboratory, Cox Institute, 3525 Southern Blvd., Dayton, OH 45429. Phone: (513) 297-8046; Fax: (513) 294-7412; Email: nreo@ discover.wright.edu. X Abstract published in Advance ACS Abstracts, April 15, 1996. 1 Abbreviations: PPs, peroxisome proliferators; C -PFA, perfluoro7 n-heptanoic acid; C8-PFA, perfluoro-n-octanoic acid; C9-PFA, perfluoron-nonanocid acid; C10-PFA, perfluoro-n-decanoic acid; C11-PFA, perfluoron-undecanoic acid; FAO, fatty acyl CoA-oxidase; T1, spin-lattice relaxation time; PCr, phosphocreatine; PCho, phosphocholine; PME, phosphomonoester; PLC, phospholipase C; DAG, sn-1,2-diacylglycerol; PKC, protein kinase C.
S0893-228x(95)00217-7 CCC: $12.00
proliferation, and hepatocarcinogenesis is not understood and represents an active area of research. Using nuclear magnetic resonance (NMR) spectroscopy, we have previously investigated the toxicological effects of two PPs in rodents: perfluoro-n-octanoic acid (C8-PFA) and perfluoro-n-decanoic acid (C10-PFA) (4, 15). These are straight-chain eight- and ten-carbon perfluorinated carboxylic acids. In general, our studies have shown that neither compound is metabolized in vivo, suggesting that their toxicological effects can be attributed to the parent compounds (4). Surprisingly, although C8- and C10-PFA are structurally very similar, distinct differences are observed in their hepatotoxic effects: (i) C10-PFA causes a more delayed lethality and protracted toxicity, while C8-PFA causes an acute and transient toxicity (4); (ii) C10-PFA activates a phosphatidylcholine-specific phospholipase C (PLC) causing an increase in liver phosphocholine (PCho) and sn-1,2diacylglycerol (DAG), while C8-PFA does not (16). This increase in PCho can readily be observed by 31P-NMR, and appears as an enhanced signal intensity in the phosphomonoester (PME) region of the liver spectrum in vivo. Studies from our laboratory also indicate that C10PFA inhibits hepatic glucose uptake and glycogen deposition (17, 18). The results from these previous studies prompted us to evaluate the importance of carbon chain length with regard to the hepatic effects induced by these perfluorocarbons. Therefore, the toxicological effects of perfluoro© 1996 American Chemical Society
690 Chem. Res. Toxicol., Vol. 9, No. 4, 1996
Goecke-Flora and Reo
Table 1. Experimental Protocol Outlining Doses for Each Treatment and the Days Post-Dose That in Vivo Experiments Were Conducted days postdose
C7-PFA
31P-NMR
control (C7-PFA)
C9-PFA
control (C9-PFA)
C11-PFA
control (C11-PFA)
vehicle (n ) 4)
50 mg/kg (n ) 5)
vehicle (n ) 4)
vehicle (n ) 4)
50 mg/kg (n ) 5)
vehicle (n ) 4)
day 3
150 mg/kg (n ) 4)
vehicle (n ) 5)
50 mg/kg (n ) 7) 100 mg/kg (n ) 4) 150 mg/kg (n ) 4)
day 5
150 mg/kg (n ) 4)
vehicle (n ) 3)
50 mg/kg (n ) 5) 100 mg/kg (n ) 3)
n-heptanoic acid (C7-PFA), perfluoro-n-nonanocid acid (C9-PFA), and perfluoro-n-undecanoic acid (C11-PFA) were investigated as follows: (i) hepatic peroxisomal fatty acyl CoA-oxidase activity (FAO) was measured to determine if C7-, C9-, and C11-PFA are PPs; (ii) the impact of these perfluorocarbons on food consumption/body weight was monitored; (iii) 19F-NMR was used to determine the potential metabolism, biodistribution and excretion of these odd-chain perfluorocarbons; and (iv) 31P-NMR was used to determine the effects of these compounds on hepatic ATP levels and phospholipid metabolism. Together with results from our previous work involving C8and C10-PFA, these data demonstrate the significance of carbon chain length in the hepatic response.
Materials and Methods Materials. Perfluorocarboxylic acid compounds were purchased with a stated purity of 99%. C7-PFA [CF3(CF2)5CO2H] was obtained from Aldrich Chemical Co. (Milwaukee, WI), while C9-PFA [CF3(CF2)7CO2H] and C11-PFA [CF3(CF2)9CO2H] were obtained from Technolube Products Co. (Los Angeles, CA). Leuco-2,7-dichlorofluorescein was purchased from Molecular Probes (Eugene, OR), and palmitoyl coenzyme-A was from Fluka Biochemika (Ronkonkoma, NY). All other chemicals were of reagent grade from standard sources. Animals and Dosing Protocol. Male Fischer-344 rats (250-320 g) were obtained from Harlan (St. Louis, MO). Treated animals received a single intraperitoneal (ip) injection of either C7-, C9-, or C11-PFA, dissolved in 1:1 (v/v) propylene glycol/water. Weight-paired control animals and ad libitum fed control animals received a single ip injection of 1:1 propylene glycol/water (vehicle). Stock solutions were prepared such that the ip dose did not exceed 0.6 mL. Previous studies from our laboratory indicate that this volume of vehicle solution causes no distress or peritoneal reaction. Details relating to doses and experimental time points for FAO enzyme activity measurements and 19F-NMR experiments will be discussed in the following sections. Doses and time points examined in 31P-NMR experiments are outlined in Table 1. Doses used in this study were in the range from 50-150 mg/ kg. The experimental treatment protocols were chosen to match previous studies (4, 15) in which the hepatic toxicological effects of C8- and C10-PFA were investigated at doses approximating their ip 30-day LD50 (i.e., 150 and 50 mg/kg, respectively). To the authors’ knowledge, however, the 30-day LD50 for C7-, C9-, and C11-PFA are unknown. Following exposure to either fluorocarbon or vehicle, animals were individually housed in metabolism cages. The animal housing area was maintained at 22 °C with a 12 h light/dark cycle. Animals were fed powdered Teklad MRH 22/5 rodent diet. Food consumption and body weight were monitored daily for all animal groups. Pair-fed control animals were given the same amount of food that their treated partners had consumed on the previous day. Pair-feeding simulates the hypophagic effects associated with these perfluorocarbon treatments. A second group of control rats received ad libitum access to food. As shown in Table 1, each perfluorocarbon treatment group possesses a corresponding pair-fed control group, except for the C9PFA-treated group. This particular treatment group required only one pair-fed control group for all experimental dose levels and days post-treatment examined. This single control group was sufficient since no significant differences were observed in
the food consumption/body weight ratios among the various C9PFA-treated experimental groups (p e0.05). All animals were fasted 16-24 h prior to FAO activity measurements or NMR experiments. FAO Enzyme Methods. In a parallel study, hepatic FAO activity was measured in rats treated with either C7-PFA (150 mg/kg), C9-PFA (50 mg/kg), C11-PFA (50 mg/kg), or vehicle on days 3 and 5 post-treatment. Livers were freeze-clamped using aluminum tongs chilled in liquid N2 and were stored under liquid N2 for subsequent FAO analyses. Livers were homogenized in 10% (w/w) sucrose containing 3 mM imidazole, pH 7.4. A peroxisome-enriched fraction was prepared by repeated centrifugation, and FAO activity was measured using a spectrophotometric assay described by Small et al. (19). FAO activity was determined using palmitoyl-CoA as substrate and leucodichlorofluorescin as a chromophore coupled to H2O2 production. Following initiation of the reaction, spectrophotometric absorbances (506 nm) were recorded at 25 °C in 3 s intervals for a total of 5 min. Absorbances were corrected by subtracting the values attributable to substrate blank. The first 60 s of data were fit by linear regression (r2 g0.99), and FAO activities were determined from the slope of the regression line for each liver sample. Protein concentrations of liver homogenates were measured by bicinchoninic acid precipitation (Pierce kit #23225), and enzyme activities are reported in nmol/(min‚ mg of protein). Statistical Analyses. All data were analyzed using either the Student’s t-test for unpaired data or a one-way analysis of variance (ANOVA) and Bonferroni statistics. Post hoc statistical tests utilized Dunnett’s statistics. Error estimates are given as the standard error of the mean (SE) and are considered to be statistically significant at a value of p e0.05. Some p values are explicitly stated in the text. NMR Studies. All NMR experiments were performed on a Bruker AM 360 high-resolution spectrometer (8.5 T). 19F-NMR of Bodily Fluids in Vitro. In this portion of the study, animals were dosed with either C7-PFA (150 mg/kg), C9PFA (50 mg/kg), C11-PFA (50 mg/kg), or vehicle. Bile was collected from treated and control rats on day 2 post-treatment (n ) 2/group). Rats were anesthetized with an ip dose of sodium pentobarbital (1 mL/kg for induction and 0.1 mL/30 min for maintenance). Bile ducts were surgically isolated and cannulated with 22 gauge catheters, and bile was collected over a 3-4 h period (bile flow ∼1 mL/h). Urine samples were collected daily into plastic cups, centrifuged to remove particulate contamination, and stored frozen at -20 °C for subsequent NMR analyses. Fluorine-19 NMR spectra were acquired from urine samples collected on days 1, 2, and 3 post-dose (n ) 2/group/day). These particular time points were chosen based upon our previous work involving C8- and C10-PFA, in which the elimination of such fluorocarbons into these bodily fluids was determined to be near maximal (4). The in vitro 19F-NMR analyses of rat urine and bile were acquired using a commercial Bruker 5 mm fluorine probe operating at a centerband frequency of 338.86 MHz. Samples were placed into 5 mm NMR tubes (Wilmad Glass Co.; catalog #507PP8), and several drops of D2O were added to provide a deuterium signal for the magnetic field lock. The initial spectrum of each sample was acquired over the entire chemical shift range (ca. 400 ppm) to ensure detection of all possible metabolites. The spectral sweep width was subsequently decreased to provide better resolution over the range of interest. Spin-lattice relaxation times (T1) were measured previously in a sample of urine from a C8-PFA-treated rat and found to range
Perfluorocarbon Chain Length and Hepatic Effects from 0.9 to 1.1 s (4). Optimal acquisition pulsing parameters were determined based upon these T1 values (4) and are given in the legend to Figure 3. Since the range in T1’s is quite small, the degree of signal saturation is similar for the various fluorine spectral resonances. Except where noted, all spectra were acquired at 25 °C. Chemical shifts for all spectra are relative to the CF3 resonance of the parent perfluorocarboxylic acid, which was set at zero ppm. In reference, trifluoroacetic acid resonates at 5.4 ppm on this chemical shift axis. For all NMR experiments, the IUPAC convention for the chemical shift scale was used. 31P-NMR. As shown in Table 1, 31P-NMR experiments in vivo were performed on days 3 and 5 post-treatment. Rats were anesthetized with halothane (5% for induction, 1% for maintenance). The liver was surgically exposed, and data were acquired using a 1 cm diameter surface coil NMR probe operating at 145.8 MHz as detailed previously (20). Data were acquired using a pulse width of 20 µs (at 100 W), sweep width of 10 kHz, interpulse delay of 7.05 s, and 240 transients. Data were processed using 2K total points, exponential multiplication yielding 20 Hz line broadening, and Fourier transformation. Spectral peak intensities were obtained by integration. Hepatic cellular viability was assessed by measuring the integrated signal intensities for the β-ATP and inorganic phosphate (Pi) resonances and calculating the ratio β-ATP/Pi. The PME region of the spectrum was integrated and normalized relative to the total integrated intensity for the entire phosphorus spectrum. All NMR data processing utilized a Macintosh IIci computer and MacFID software (Tecmag, Inc., Houston, TX). Upon completion of the NMR experiment, livers were freeze-clamped using aluminum tongs chilled in liquid N2. Livers were then stored under liquid N2 until subsequent preparation of perchloric acid extracts and FAO activity measurements. High-resolution 31P-NMR spectra of perchloric acid liver extracts were obtained at 37 °C using a 10 mm broadband commercial probe. Data were acquired using a 45° pulse, sweep width of 6024 Hz, an interpulse delay of 2.68 s, and approximately 10 h of signal averaging. Data were processed with 16K total points and an exponential filter producing 2 Hz line broadening. Phosphocreatine (PCr) was added to the extract solutions (0.5 mM) to serve as a chemical shift reference (set at 0 ppm) and concentration standard for absolute quantitation of metabolite concentrations (15).
Chem. Res. Toxicol., Vol. 9, No. 4, 1996 691
Figure 1. Liver peroxisomal fatty acyl CoA-oxidase (FAO) activities (mean ( SE; n ) 3-4) in C7-PFA (150 mg/kg), C9PFA (50 mg/kg), and C11-PFA (50 mg/kg) treated rats and corresponding pair-fed controls (vehicle treated) measured at 3 and 5 days post-treatment. FAO activities (nmol/(min‚mg of protein)) were determined from peroxisome-enriched fractions of liver homogenates at 25 °C as described in Materials and Methods. The asterisks denote statistical significance in comparison to corresponding control groups (p e0.05).
Results FAO Activity. To determine if C7-, C9-, and C11-PFA are peroxisome proliferators, their effects on hepatic peroxisomal FAO activity were measured, and the data are displayed in Figure 1. Rats treated with C9- or C11PFA show ca. 4-fold greater FAO activity as compared to controls on days 3 and 5 post-treatment. In contrast, C7-PFA-treated rats show no significant difference in FAO activity from controls on days 3 or 5 post-treatment. Food Consumption/Body Weight. The ratio of food consumption/100 g body weight (measured daily) is shown in Figure 2. Ad libitum-fed control animals were monitored for 5 consecutive days prior to treatment with vehicle solution (days -4 to 0). Data acquisition for all other groups was initiated following treatment on day 0. At this time, control animals were treated with vehicle solution, and experimental groups were given the appropriate fluorocarbon compound. Data were collected for up to 5 days post-treatment for each group; however, since animals were fasted the day prior to experimentation, data for day 5 post-treatment are near zero. Within the first 24 h postdose, all fluorocarbon-treated animals show a significant decrease in the food consumption/body weight ratio relative to ad libitum-fed control animals. On day 1 post-treatment, C7-PFA-treated rats showed a pronounced decrease in food consumption/body weight ratio followed by recovery toward control levels
Figure 2. Mean daily food consumption/100 g body weight ratios for C7-PFA (150 mg/kg; squares; n ) 4), C9-PFA (50 mg/ kg; diamonds; n ) 8), C11-PFA (50 mg/kg; circles; n ) 5) and ad libitum-fed controls (inverted triangles; n ) 6). Control animals were monitored for 5 consecutive days prior to treatment (days -4 to 0) and for 4 days post-treatment with vehicle solution. Data acquisition for all other groups was initiated following treatment with the appropriate compounds on day 0. For some data points, the error bar ((SE) is contained within the size of the symbol and is not shown.
on days 2-4. This response is similar to that previously observed with the same dose of C8-PFA (4). Additionally, the dose used for C7-PFA (150 mg/kg) produced other signs of toxicity that were similar to those seen with C8PFA treatment: namely, pronounced urine output, lethargy, and hepatomegaly. In contrast, rats treated with either C9- or C11-PFA showed a continual decrease in food consumption/body weight ratio on days 2-4 post-treatment. This response is similar to that previously observed for C10-PFA treatment (4). Ad libitum-fed control rats show a relatively steady level of food consumption on all days examined. 19 F-NMR. In order to ascertain if any of these oddchain perfluorocarbons are metabolized, high-resolution 19F-NMR spectra were acquired from samples of bile and urine from treated rats. These spectra were compared with the high-resolution 19F-NMR spectra of each dosing
692 Chem. Res. Toxicol., Vol. 9, No. 4, 1996
Figure 3. High-resolution 19F-NMR spectra of rat bile (A) and urine (B) at 8.5 T and 27 °C, using a 30° pulse, an interpulse repetition time of 0.151 s, and a sweep width of 20 kHz. All spectra shown are from samples of bodily fluids collected on day 2 post-dose. Data were processed using 4K total points and an exponential filter producing 10 Hz line broadening. (IA) Bile from a C7-PFA-treated rat (150 mg/kg) with 12.6 h of signal averaging (300 070 transients); (IB) urine from a C7-PFA-treated rat (150 mg/kg) with 14.4 h of signal averaging (344 030 transients); (IIA) bile from a C9-PFA-treated rat (50 mg/kg) with 14.3 h of signal averaging (340 645 transients); (IIB) urine from a C9-PFA-treated rat (50 mg/kg) with 13.1 h of signal averaging (312 373 transients); (IIIA) bile from a C11-PFA-treated rat (50 mg/kg) with 14.9 h of signal averaging (353 961 transients); (IIIB) urine from a C11-PFA-treated rat (50 mg/kg) with 8.4 h of signal averaging (200 856 transients). Chemical shifts are relative to the CF3 resonance which was set at zero ppm.
solution (data not shown). Spectra of these parent compounds are similar to that of C8-PFA (see ref 4). Figure 3 shows high-resolution 19F-NMR spectra of bile (A) and urine (B) from C7-, C9-, and C11-PFA-treated rats on day 2 post-treatment (spectra of urine samples collected on days 1 and 3 post-treatment yield similar results to those shown for day 2). Note that the level of detection (i.e., sensitivity) in these 19F-NMR experiments is estimated to be 100 µM as determined in previous studies (4). The in vitro bile 19F spectra from C7-PFAtreated rats (IA) show no detectable signals. In contrast, in vitro bile 19F spectra from rats treated with either C9PFA (IIA) or C11-PFA (IIIA) reveal resonances attributable to the parent dosing compound. In vitro urine 19F spectra from rats treated with either C7-PFA (IB) or C9PFA (IIB) display resonances attributable to the respective parent compound, while urine spectra acquired from C11-PFA treated rats (IIIB) reveal no resonances. Thus, C7-PFA is predominantly found in urine, C11-PFA is in bile, and C9-PFA is found in both urine and bile. The signal-to-noise ratio for C9-PFA is substantially greater in bile (IIA) than urine (IIB), indicating that C9-PFA is more concentrated in bile. It should be noted that, in several of the urine or bile spectra obtained from C7-, C9-, or C11-PFA-treated rats, a very weak signal is observed in the range from -37 to -39 ppm. This peak was determined to arise from a dietary source of inorganic fluoride as discussed in a prior study (4). In general, the spectra show only signals from the parent compounds and no evidence of fluoro metabolites. 31P-NMR. Figure 4 shows representative 31P liver spectra from a C9-PFA treated rat (50 mg/kg) and its corresponding control in vivo at day 5 post-dose. Note the enhanced signal intensity in the PME region for the treated rat as compared to control. This effect was also
Goecke-Flora and Reo
31P-NMR spectra of liver in vivo at 8.5 T from a C9-PFA-treated rat (50 mg/kg) at 5 days post-dose (A) and its pair-fed control (B). Signals observed are from adenosine triphosphate (ATP), inorganic phosphate (Pi), and phosphomonoesters (PME). Data acquisition and processing parameters are given in Materials and Methods. Chemical shifts are referenced to phosphocreatine at 0.0 ppm by setting the R-ATP peak to -7.6 ppm.
Figure 4. Surface coil
observed for C10-PFA (15) and C11-PFA treatments, but not for C7-PFA (data not shown). The ratio of the integrated intensities of the β-ATP/Pi resonances serves as a monitor of ATP concentration and indicator of overall tissue viability. This ratio was calculated for all treated and control rats on days 3 and 5 post-treatment. Figure 5 shows a plot of the mean β-ATP/Pi ratios for C7-, C8-, C9-, C10 -, and C11-PFA-treated rats and their corresponding controls at days 3 (Figure 5A) and 5 (Figure 5B) posttreatment. Data from C8- and C10-PFA-treated rats were from a previous study (15). No differences are observed in the mean β-ATP/Pi ratios between treated and control groups for C7-, C8-, C9-, and C10-PFA. In contrast, C11PFA-treated rats show mean β-ATP/Pi ratios that are 27% and 31% lower than corresponding controls on days 3 and 5 post-treatment, respectively (p e0.05). For the higher doses of C9-PFA (100 and 150 mg/kg), the mean β-ATP/Pi ratios are not different from corresponding controls on day 3 post-treatment (data not shown). On day 5 post-treatment, however, the 100 mg/kg dose of C9PFA causes a significant decrease in the mean β-ATP/Pi ratio (0.76 ( 0.17 versus 1.49 ( 0.19). Figure 6 shows the integrated signal intensities of the PME region for C7-, C8-, C9-, C10-, and C11-PFA-treated and their corresponding control rats at days 3 (Figure 6A) and 5 (Figure 6B) post-treatment. Data from C8- and C10-PFA-treated rats were from a previous study (15). Rats treated with C9-, C10-, or C11-PFA (50 mg/kg) show a 2- to 2.5-fold increase in signal intensity in the PME
Perfluorocarbon Chain Length and Hepatic Effects
Figure 5. Integrated intensity ratios of β-ATP/Pi (means
( SE; n ) 3-8) from 31P-NMR rat liver spectra in vivo for C7-, C8-, C9-, C10-, and C11-PFA-treated (solid bars) and corresponding control groups (hatched bars). Data are shown for each group at day 3 (A) and day 5 (B) posttreatment. Data from C7- and C8-PFA-treated groups were obtained from rats treated at 150 mg/kg while data from C9-, C10-, and C11-PFA-treated groups were obtained from rats treated at 50 mg/kg. Integrated signal intensities are relative to the integrated area for the entire 31P spectrum set equal to 100. The asterisks denote statistical significance in comparison to corresponding control groups (p e0.05). region as compared to control on day 5 post-treatment (p
