Chem. Res. Toxicol. 1994, 7, 15-22
15
Effects of the Peroxisome Proliferator Perfluoro-n-decanoic Acid on Hepatic Gluconeogenesis and Glycogenesis: A 13C NMR Investigation Carol M. Goecke,? Bruce M. Jarnot,tJ and Nicholas V. Reo*’? Department of Biochemistry and Molecular BiologylKettering-Scott Magnetic Resonance Laboratory, Wright State University and Kettering Medical Center, Dayton, Ohio 45429, and Toxicology Division, Armstrong Laboratory, Wright-Patterson Air Force Base, Dayton, Ohio 45433 Received August 6, 1993’ Carbon-13 NMR spectroscopy was used to study the effects of the peroxisome proliferator perfluoro-n-decanoic acid (PFDA) on hepatic carbohydrate metabolism in male Fischer-344 rats. The data indicate that PFDA-treated rats display an inhibition in hepatic [l-13C]glucose and [3-l3CIalanine utilization on day 5 posttreatment. PFDA rats show hepatic mean glucose and alanine intensities which are significantly greater (ca. 10-100 % ) than controls. With [1J3C]glucose as substrate, PFDA rats show severe to complete inhibition in glycogenesis on days 3 and 5 posttreatment. With [3-l3C1alanine as substrate, both groups show functional gluconeogenesis and glycogenesis; however, treated rats show a more transient and less intense C1glycogen resonance relative to control. These data suggest that PFDA inhibits either the hepatocellular transport of glucose and/or its phosphorylation by glucokinase. The effect of PFDA on TCA cycle activity was determined by monitoring the flow of [3-13C]alanine into glutamate. The relative activity of pyruvate carboxylase (PC) uersus pyruvate dehydrogenase (PDH) is represented by the ratio of the glutamate NMR signal intensities (C2 C3)/C4. PFDA rats show a lower (C2 + C3)/C4 glutamate ratio, suggesting greater relative activity of PDH uersus PC in PFDA rats relative to controls. Differences in P D H activity may arise from differences in lipolytic activity. Our data suggest a dysfunction in fatty acid metabolism in PFDA rats and corroborate other studies which show that PFDA inhibits fatty acid oxidation.
+
Introduction Peroxisome proliferators constitute a structurally diverse class of compounds which alter hepatic lipid metabolism (1-3). Previous studies have shown a high correlation between peroxisome proliferation and hepatocellular neoplasms in rodents ( 1 , 2, 4 , 5). This has generated considerable interest regarding the underlying toxicological mechanisms of these compounds. Perfluoron-decanoic acid (PFDA)l is a perfluorinated carboxylic acid and known peroxisome proliferator in rodents. Previous studies ( 4 , 5 ) have shown that a single intraperitoneal (ip) injection of PFDA in rats results in a characteristic “wasting syndrome” and delayed lethal toxicity. Additionally, PFDA-treated rata require a higher caloric intake to maintain their body fat content. Toxic symptoms include pronounced hypophagia, marked body weight loss, and hepatomegaly. Hepatic studies reveal changes in the ratio of saturated to unsaturated fatty acids ( 4 , 5 ) decreased , ketogenesis,2elevated hepatic triglycerides (6,7),peroxisome proliferation, mitochondrial disruption
* Correspondence and requests for reprints should be addressed to this author at KSMRL, Cox Institute, Wright State University, 3525 Southern Blvd., Dayton, OH 45429. Phone: (513) 873-5327;Fax: (513) 294-7412. t Wright State University and Kettering Medical Center. 1 Armstrong Laboratory. a Abstract published in Aduance ACS Abstracts, December 1, 1993. 1 Abbreviations: PFOA, perfluoro-n-octanoicacid; PFDA, perfluoron-decanoic acid; NMR, nuclear magnetic resonance; NOE, nuclear Overhaueer enhancemenc G6P, glucose &phosphate; GGPase, glucose 6-phoephataae; GK, glucokinase; 2-DG, 2-deoxy-~-glucose;2-DG6P, 2-deoxy-~-glucose6-phosphate ; PDH, pyruvate dehydrogenase; PC, pyruvate carboxylase;FA, fatty acid; Ts,triiodothyronine;T,,thyroxine. 2 G. D. Pilcher, unpublished results.
(8), decreased mitochondrial and peroxisomal @-oxidation (7, 9-12),2 and increased w-oxidase a ~ t i v i t y . ~ Recently, we used lSFNMR to identify possible fluorinated metabolites of perfluoro-n-octanoic acid (PFOA), a straight-chain eight-carbon carboxylic acid, and PFDA, a straight-chain ten-carbon carboxylic acid, in rat liver in uiuo and in various bodily fluids (13). The data suggest that PFOA and PFDA are not metabolized in vivo and that differences in their toxicity may be related to differences in the excretory processes which clear them from the body. The toxicologic mechanism associated with PFDA treatment is not understood. The goal of this study was to investigate the effects of PFDA on hepatic gluconeogenesis and glycogenesis using 13CNMR spectroscopy in conjunction with [l-13Clglucose and [3J3C1alanine. Previous studies have shown that PFDA alters hepatic mitochondrial and peroxisomal @-oxidationof endogenous fatty acids; however, to the authors’ knowledge,no studies have been reported which evaluate the effects of PFDA on carbohydrate metabolism. Such an assessment may provide a clearer understanding of the biochemical basis for the toxicity associated with PFDA and the interrelationship between hepatic carbohydrate and lipid metabolism. NMR spectroscopy offers a unique advantage over conventional biochemical approaches in that metabolism can be followed in a living animal over real time. Distinct advantages of the 1%-nuclide include the following: (i) a low natural abundance (l.l%), which enables the use of isotopically enriched compounds to investigate specific B. M. Jarnot, J. R. Okita, and W. J. Scmidt, unpublished results.
0893-228x/94/2707-0015$04.50/00 1994 American Chemical Society
16 Chem. Res. Toxicol., Vol. 7, No. 1, 1994
groupb 1C 2T 2C 3T 3C 4T 4C 5T 5C
Table 1. ExDerimental Protocola no. of treatanimals (n) ment day3 day5 challenge compound control
4
}
[l-'3C]-~-glucose(600 mg/kg)
control PFDA control PFDA control PFDA control
5
}
unlabeled D-glUCOSe (600 mg/kg)
}
2-deoxy-~-glucose (600 mg/kg)
}
[3-l3C1-~-alanine (600 mg/kg)
;1
[3-13C]-~-alanine (600 mg/kg) and unlabeled D-glucose f6.39mg/(kg.min)l
Experimental protocol outlining the concentration of the challenged compound administered,the days postdose that each experiment was conducted, and the number of animals in each treatment group. PFDA-treatedand correspondingcontrols are indicated with a T and C, respectively. All challenge compounds were administered to treated and control groups as an iv bolus with the exception of Group 5, which received a continuous iv infusion of unlabeled D-glucose in addition to the iv bolus of alanine. NMR studies were performed on Groups 1, 4, and 5 while benchtop studies were performed on Groups 2 and 3.
metabolic pathways, and (ii) a broad chemical shift range (over 200 ppm), which offers great spectral dispersion and well-resolved individual '3C-metabolite resonances.
Experimental Procedures Materials. Ultrapure (99+%) PFDA was obtained from Technolube Products Co. (Los Angeles, CA). Isotopically enriched [1-13C]-~-glucose(99 atom % ) and [3-W]-~-alanine (99atom % ) were purchased from CambridgeIsotopesLaboratory (Woburn, MA). Animals. Male Fischer-344 rats (200-310 g) were obtained from Charles River Breeding Laboratories (Wilmington, MA) and Harlan (St. Louis, MO). The animal housing area was maintained at 22 OC with a 12-hlight/dark cycle, and the animals were fed powdered Tekland MRH 22/5 rodent diet no. 8640. Treated animals received a single ip injection of 50 mg PFDA/kg dissolved in vehicle, 1:l v/v propylene glycol/water. Weightpaired control animals received a single ip injection of vehicle. Stock solutionswere prepared such that the ip dose did not exceed 0.5 mL; previous studies from our laboratory indicate that this volume of vehicle solution causesno distress or peritoneal reaction. Following exposure to either PFDA or vehicle, animals were individuallyhoused in metabolism cages. PFDA animals received ad libitum access to food and water. Control animals received ad libitum access to water but were fed the same amount of food as their PFDA-treated partners had consumed during the previous 24-h period (pair-feeding). Pair-feeding simulates the prolonged hypophagic effects associated with PFDA. Animals were randomly placed into experimental groups as shown in Table 1. Experimental protocol and number of animals in each group are as outlined. Data were obtained on day 3 and/or day 5 posttreatment, depending on the experimental group. Groups 1 and 2 are identical except that Group 1 rats received [1-13C]glucosewhile Group 2 rats received unlabeled glucose. Groups 4 and 5 are identical except that Group 4 rats received only [3-l3C1alanine while Group 5 rats received a continuous infusion of unlabeled glucose in addition to [3-l3C1alanine. NMRStudies (Groups lY4,and5).Prior toexperimentation, animals were fasted for approximately 14 h to deplete hepatic glycogen stores. Rats were anesthetized with halothane (5 % for induction, 1% for maintenance), and a 24-gaugeQuik-CathTeflon catheter (Baxter, IL) was inserted into the femoral vein. A solution of heparinized saline (2.48 units of heparin/mL of 0.9% NaC1) was infused (0.024 mL/min) to maintain the iv line patent.
Goecke et al. The liver was surgically exposed by a transverse subcostalincision, the xiphoid process and falciform ligament were resected, and the abdominal area was covered with plastic wrap to reduce fluid loss and help maintain body heat. The animal was placed vertically in a plastic cradle (NMR probe), a 1 mm thick glass plate placed over the liver, and the surface coil (described below) positioned against this plate. Temperature-regulated water was circulated through tubing placed beneath the animal while in the NMR probe in order to maintain body temperature. Rectal temperature was measured using a digital thermometer (Model 500, VWR Scientific) and was maintained at 37 1 OC during data acquisition. Proton-decoupled 13C NMR spectra were obtained using a double-resonance W-{lH) surface coil (14), on a Bruker AM 360 spectrometer (Billerica, MA) a t 8.5 T (360.13 MHz for lH, 90.56 MHz for W). A gated bilevel (5 W; 0.5 W) composite pulse decoupling sequence (Waltz-16) was used to minimize sample heating. Data were collected using the following parameters: pulse width = 16 ps; acquisition time = 82 ms; sweep width = 25 kHz; and a pulse repetition time = 132ms. Data were collected with 5-min time resolution (2196 transients). During 20 min of base-line data accumulation, animals in Groups 1and 4 received an iv infusion of heparinized saline while animals in Group 5 received an iv infusion of unlabeled D-glucose. Animals then received an iv bolus injection of either [l-W]glucose or [3-W]alanine, and subsequent NMR spectra were collected for 90 min. Data were processed using 8K total points, exponential multiplication yielding 3 0 - H line ~ broadening, Fourier transformation, and a fourth-order polynomialbase-line correction. The IUPAC convention for the chemical shift scale was used. Benchtop Studies (Groups 2 and 3). Two complementary benchtop studies were performed. In both studies, animal surgery, positioning in the probe, and infusion of heparinized saline were performed in an identical fashion as discussed above in order to simulate the NMR experimental conditions. In the first study, blood and urine glucose levels were monitored in PFDA and control rats (Groups 2T and 2C) following an iv challenge of unlabeled D-glUCOSe. Blood samples (0.15 mL) were drawn for serum glucose determinations prior to and at 5,15,30, 60,90, and 120 min post-glucose. Bladder urine samples were collected after the final blood draw. Serum and urine samples were stored frozen at -14 "C (up to 1 week), and glucose determinations were performed on a Kodak Ektachem (Model 700 XR, Rochester, NY). In the second study, hepatic 2-deoxypglucose 6-phosphate (2-DG6P)levels were determined in PFDA and control rats (Groups 3T and 3C) following an iv bolus of 2-deoxy-~-glucose(2-DG). At 5 min post-2-DG,livers were freezeclamped using aluminum tongs chilled in liquid Nz, a perchloric acid extract was prepared (15),and the concentration of 2-DG6P was spectrophotometrically determined (16). Data Analyses. Statisticalanalysesemployed the appropriate repeated measures ANOVA using SAS-PC Version 6.04, Proc GLM. Significance was determined using the sequential Bonferroni method (17). Datawere consideredstatistically significant at a value of p 5 0.05. Error estimates are given as the standard error of the mean (SE).
*
Results Glucose Metabolism. A natural-abundance W-liver spectrum from a control rat on day 5 postexposure is shown in Figure 1A. The low-frequency region of overlapping signals between 15 and 50 ppm is due to the methyl and methylene carbons primarily from phospholipids, free fatty acids, and triacylglycerols. The two higher frequency regions constitute the single- and double-bonded allylic carbons associated with unsaturated fatty acids at ca. 130 ppm and the carbonyl carbons at ca. 172 ppm. Figure 1B is a 13C-liver spectrum obtained from the same control rat at 22.5 min post [1-13C]glucose. Resonances attributable to the C1 a-and p-anomers of glucose are observed at 92.8 and 96.6 ppm, respectively. An additional resonance
Chem. Res. Toxicol., Vol. ?, No. 1, 1994 17
Effect of PFDA on Gluconeogenesis and Glycogenesis -CHz-
100
1
Chollne
C -glucose
1
t
w".;,
B
0
I
10
20
30
40
50
40
50
TIME [min)
I
0 I
.
180
.
'
,
'
160
.
I
140
'
~
,
120
"
'
I
"
100
'
l
80
~
~
'
60
I
"
'
40
I
'
~
20
'
I
'
~
0
PPM
Figure 1. Proton-decoupled l3C NMR liver spectra in uiuo obtained 5 days posttreatment at 8.5 T, representing 5 min of
signal averaging. Data acquisition and processing parameters are aa given in ExperimentalProcedures. (A) Base-line spectrum from a pair-fed control rat; (B) Pair-fed control rat at 22.5 min post iv [l-l~Clglucosebolus; (C) PFDA rat at 22.5 min post iv [l-lWlglucosebolus. Note the absence of a resonance attributable to the C1-glycogenin the PFDA-treated rat. The peak attributable to the methyl carbons of the choline head groups served as an internal chemical shift reference and was set to 54.6 ppm relative to TMS at 0 ppm. attributable to the C1-glycogen is observed at 100.5 ppm, indicating that [l-13Clglucoseis incorporated into glycogen. In general, the C1-glycogenresonance is observed in all control rats (Group 1C) throughout the course of the experiment (90min),startingfromabout 5minpost [l-l3C]glucose. A 13C-liverspectrum obtained from a PFDA rat (Group 1T) on day 5 postdose at 22.5 min post [ l - W ] glucose is shown in Figure 1C. This spectrum shows resonances attributable to the C1 a- and p-anomers of glucose; however, unlike the spectrum in Figure lB, no C1-glycogen resonance is observed. Day 3 PFDA rats show severe inhibition in the incorporation of [l-13Clglucose into glycogen,with the C1-glycogen resonance being barely detectable above the noise, while day 5 PFDA rats show complete inhibition in which the glycogen resonance is not detectable. Hepatic glucose utilization was assessed for all rats by monitoring the C1-8signal intensity, which was normalized to 100% in the first spectrum following glucose infusion. The percent change in mean hepatic glucose intensity from 2.5 to 47.5 min post-glucose for PFDA and control rats (Groups 1Tand 1C)on day 3 (Figure 2A) and day 5 (Figure 2B) posttreatment is shown in Figure 2. PFDA rats show a mean glucose intensity which is a significant 1096 greater compared to controls at 7.5 min post-glucose on day 3 and approximately 20% greater at 7.5,12.5, and 17.5 min post-
'
I
10
20
30
TIME (min)
Figure 2. Percent change in the mean NMR intensity of the W1-/3-anomerof glucose at time post iv [l-lsClglucosebolus for PFDA (closed circles) and control rats (open circles),at 3 days (A) and 5 days (B)posttreatment. Data are plotted at the
midpoint in time for each 5-min spectrum. Error bars represent S E . The asterisk (*) denotes a significantdifference between groups at p I0.05. glucose on day 5. Later time points reveal no significant differences between groups. The apparent rates of hepatic [l-13Clglucoseutilization (mean f SE) were calculated for PFDA and control rats in Group 1 on days 3 and 5 postdose by fitting the curve of C1-@-glucoseintensity uersus time (from 2.5 to 47.5 min post-glucose) to a third-order polynomial for each individual rat. All curves yield a value of r2 1 0.97. The first derivative of this polynomial expression was used to calculate the slopes of each curve at the specifictime points shown in Figure 2. These values represent apparent rates of hepatic glucose utilization. On day 3, PFDA rats show a significant 15% lower apparent rate of glucose utilization compared to control at 2.5 min post-glucose (4.62 f 0.34 uerszm 5.50 f 0.49). On day5, PFDA rata show a significant 2 5 3 5 % lower apparent rate of glucose utilization compared to control at 2.5 and 7.5 min post-glucose (3.59 f 0.64 versus 5.75 f 0.29 and 3.30 f0.33 versus 4.30 f 0.15). In a parallel benchtop study, serum glucose concentrations were determined for PFDA and control rats (Groups 2T and 2C) on day 3 and 5 posttreatment (Figure 3). No significant difference in serum glucose levels is observed between groups on either day. Urine glucose levels are elevated above normal in both groups on days 3 and 5, but no significant difference is observed between groups (data not shown). In the second parallel benchtop study, the concentration of hepatic 2-deoxy-~-glucose6-phosphate was determined for PFDA and control rats (Groups 3T and 3C) on day 5 posttreatment. Control rats yield a ca. 2.6-fold greater
18 Chem. Res. Toxicol., Vol. 7, No. 1, 1994
Goecke et al.
Ik
A
A
PFDA
-
CONTROL
=
v
10
30
50 70 TIME (minl
90
110
130
10
20
30 TIME (min]
40
50
10
20
30 TIME (min)
40
50
B
B
10
30
50
70
90
110
130
TIME (minl
Figure 3. Mean serum glucose concentrationsfor PFDA (closed circles) and control rats (open circles),at 3 days (A) and 5 days (B)posttreatment. Error bars represent hSE. concentration of 2-DG6P (mean f SE) than PFDA rats, 1.90 f 0.11 versus 0.74 f 0.07 (mmol/kg of protein), respectively. Alanine Metabolism. In a parallel NMR study, the effect of PFDA on hepatic alanine metabolism was monitored in PFDA and control rats (Groups 4 and 5). Following [3 - W l alanine infusion, W-liver spectra from PFDA and control rats reveal a resonance of large intensity attributable to the Cs-alanine (17.2 ppm). Subsequent spectra reveal resonances attributable to glucose,glycogen, and other metabolites. Due to metabolic randomization of the 13C-label,glucose and glycogen are labeled at various carbon positions. The incorporation of alanine into glucose and glycogen was monitored in both groups. The C1glycogen resonance at ca. 100.5 ppm was used to monitor glycogen production since it is well resolved and contains no overlap from other glucose or glycogen resonances. The data demonstrate that all control rats (Groups 4 and 5) show resonances attributable to both glucoseand glycogen. In general, the C1-glycogenresonance is observed in these groups from approximately 10 to 15min post-alanine and persists throughout the experiment (90 min). Similar to controls, all PFDA rats (Groups 4 and 5) show resonances attributable to glucose; however, only some (n= 3, Group 4T; and n = 3, Group 5T) demonstrate incorporation of alanine into glycogen. In PFDA rats showing glycogen production, the C1-glycogen resonance was transient, lasting only for approximately 5-10 min. Hepatic alanine utilization was measured in PFDA and control rats by normalizing the C3-alaninesignal intensity to 100% in the second spectrum acquired post [3-l3C1alanine bolus and monitoring its intensity over time. Figure 4 shows the percent change in mean hepatic alanine intensity from 7.5 to 52.5 min post-alanine for both PFDA
01
1
Figure 4. Percent change in the NMR intensity of the [WS]alanine at times post iv [3-W]alaninebolus. (A) Control and PFDA rats (Groups 4C and 4T) receiving only [3-l3C1alanine; (B)controland PFDA rats (Groups5C and 5T)receiving [3-W]alanine and unlabeled glucose. Data are plotted at the midpoint in time for each 5-min spectrum. Error bars represent hSE. The asterisk (*) denotes a significant difference between treated and corresponding control groups at p I0.05. and control rats in Group 4 (Figure4A) and Group 5 (Figure 4B). PFDA rats receiving only alanine (Group 4T) show a mean alanine intensity which is ca. 30% higher than controls (Group 4C) from 27.5 to 37.5 min post-alanine, while PFDA rats receiving alanine and unlabeled glucose (Group 5T) show a mean alanine intensity which ranges from 10% to 100% greater than controls (Group 5C) from 17.5 to 52.5 min post-alanine. The apparent rates of hepatic alanine utilization (mean f SE) were determined for PFDA and control rats by fitting the curve of alanine intensity versus time (from 7.5 to 42.5 min post-alanine) to a third-order polynomial for each individual rat. All curves yielded a value of r2I 0.98. The first derivative of this polynomial expression was used to calculate the slopes of each curve at the specific time points shown in Figure 4. PFDA rats show no significant difference in the apparent rate of alanine utilization relative to their corresponding control; however, PFDA rats receiving alanine and glucose (Group 5T) show a significant 2- to 4-fold decrease in this rate relative to PFDA rats receiving only alanine (Group 4T) at 7.5 (0.77 f 0.77 versus 3.34 f 0.78) and 12.5min (1.57 f 0.48 versus 3.24 f 0.40). In contrast, control rats receiving alanine and glucose (Group 5C) show no significant decreaae in the apparent rate of hepatic alanine utilization as compared to control rata receiving only alanine (Group 4C). Glutamate Labeling. The fate of C13Clalaninein the TCA cycle was determined by measuring the percentage of ['3C]pyruvate which enters the TCA cycle as oxaloac-
Effect of PFDA on Gluconeogenesis and Glycogenesis
Chem. Res. Toxicol., Vol. 7,No. 1, 1994 19 Table 2. Ratios of (Cs + Cs)/Cd Glutamate Intensities (Mean f SE) for PFDA and Pair-Fed Control Rats Measured in '42 NMR Liver Spectra Group 4 C Group 4 T Group 5C: Group 5 T time (min) control, PFDA, control PFDA, post-alanine0 (-) glucose (-)-glucose (+)-glucose (+)-glucose 15 6.6 i 1.ObpC 3.9 f 0.5bpd 2.9 f 0 3 2.1 f 0.5d 25 7.9 f l.lb*c 3.6 i 0.lb~d 2.5 f 0.3c 2.9 f 1.0" 35 7.1 f 1.3bbc 3.7 i 0.4b~~3.4 f 0.8bsC 2.1 f O.qbpd Time points represent the midpoint in time of each 10-min difference spectrumacquired post-alanine administration. b Denotes a significant difference (p I0.05) between PFDA rata and their corresponding pair-fed control. Denotes a significant difference (p I0.05) between control rats receiving only alanine (Group 4C) and control rats receiving alanine and unlabeled glucose (Group 5C). Denotes a significant difference (p 2 0.05) between PFDA rata receiving only alanine (Group 4T) and PFDA rata receiving alanine and unlabeled glucose (Group 5T).
I/
i
L
1 ' ' ' ~ 1 ' ' ~ ~ 1 ' ' ~ ' 1 ' ~ ~ ~ I ' ' ~ ' l ~ ' ' ' I " '
SO
70
60
40
50
PPM
30
20
Figure 5. Expanded region of proton-decoupled 13CNMR liver difference spectra, obtained 15min post iv [3-13C]alanine bolus, representing 10 min of signal averaging. Data acquisition and processing parameters are as given in Experimental Procedures except that an exponential filter producing 20-Hz line broadening was used. (A) Pair-fed control rat receiving only [3-'3C]alanine (Group 4C); (B)PFDA-treated rat receiving only [3-'3C]alanine (Group 4T). The labeled 13C NMR peaks include those due to the a-and 4-anomers for D-glucose: @C-2,@C-3,4C-5,@C-S,aC2, (uC-3,and aC-6. Abbreviations: Gly, glycogen; Glu, glutamate; Gln, glutamine;Ala, alanine;Asp, aspartate; Lac, lactate. Carbon13chemical shifts are given relative to TMS at 0 ppm by setting the C3 of alanine at 17.2 ppm.
etate uersus that which enters as acetyl-coA (see Discussion). Difference spectra representing 10 min of data acquisition were obtained by subtracting the 13C-natural abundance background spectra from subsequent spectra acquired after the addition of [3-l3C]alanine. Difference spectra were obtained from 10to 40 min post-alanine, and the ratio of glutamate intensities (C2 + C3)/C4 was calculated and compared between PFDA and control rats in Groups 4 and 5. Figure 5 shows difference spectra from a PFDA and control rat receiving only alanine (Group 4T and 4C) at 15 min post-alanine. Note the presence of resonances attributable to the C2, C3, and C4 of glutamate. Also present are resonances attributable to other metabolites including glucose, glycogen, alanine, aspartate, and lactate. Resonance assignments were based on literature values (15, 18, 19). The (C2 C3)/C4glutamate intensity ratios for PFDA and control rats are shown in Table 2 (Groups 4 and 5). Signal intensities were not corrected for TI saturation nor nuclear Overhauser enhancement (NOE) effects since relative intensities were compared among the various noncarboxyl carbons of the glutamate molecule. On the basis of the small size of this molecule, it is reasonable to assume that any conditions which change the TI relaxation times and NOE factors will affect all the non-carboxyl carbons equally.
+
+
Within each group, the (C2 C3)/C4 glutamate ratio remains relatively constant through time. Comparison of the glutamate ratios between PFDA and control animals, however, reveals a significantly lower ratio in PFDA rats receiving only alanine (Group 4T) relative to control (Group 4C) from 15 to 35 min post-alanine. In addition, PFDA rats receiving alanine and glucose (Groups 5T) show a significantly lower ratio compared to control (Group 5C) at 35 min post-alanine. At all time points examined, ' / PFDA rats receiving alanine and unlabeled glucose (Group 5T) show a significantly lower ratio than PFDA rats receiving only alanine (Group 4T). In addition, control rats receiving alanine and unlabeled glucose (Group 5C) show a significantly lower ratio than control rats receiving only alanine (Group 4C). The magnitude of the decrease in this ratio, however, is greater between control groups than between treated groups.
Discussion Carbon-13 NMR liver spectra indicate that PFDA alters hepatic glucose, alanine, and glycogen metabolism. In general, PFDA rats show higher mean hepatic glucose and alanine NMR intensities compared to controls (Figures 2 and 4). The data suggest that hepatic glucose and alanine utilization is inhibited by PFDA treatment. Liver spectra also show that the administration of supplementary glucose in conjunction with alanine inhibits the apparent rate of alanine utilization in PFDA in comparison to control rats. PFDA rats receiving alanine and glucose (Group 5T) show an initial 2- to 4-fold decrease in the apparent rate of alanine utilization relative to PFDA rats receiving only alanine (Group 4T). Data from this study also show that PFDA affects glycogen synthesis from glucose and alanine. In general, liver NMR spectra from PFDA-treated rats show severe to complete inhibition in glycogenesisfrom glucose on days 3 and 5 posttreatment. With alanine as substrate, some PFDA rats show functional glycogenesis, while all rats show functional gluconeogenesis. Unlike controls, the appearance of liver glycogen (as monitored from the C1glycogen resonance) in PFDA rats receiving alanine alone is transient. This transience is believed attributable to differences in the metabolic status between PFDA and control rats. Though food consumption is identical for both groups, mitochondrial and peroxisomal fatty acid @-oxidation is inhibited in PFDA rats (10-12).2 Since PFDA rats are metabolically compromised, they are thought to be more "starved" than controls, and therefore,
Goecke et al.
20 Chem. Res. Toxicol., Vol. 7, No. 1, 1994
glycogen may be rapidly degraded for utilization by other tissues. To investigate this possibility, supplementary glucose was given to PFDA and control rats (Group 5) in order to promote the incorporation of a greater percentage of the 13C-label from alanine into hepatic glycogen. Previous studies have shown that glucose diverts gluconeogenic precursors into glycogen rather than being used itself for glycogenesis (20-22). Nevertheless, even in this parallel study, the C1-glycogen resonance remains transient in PFDA rats. Thus, the transient nature of the glycogen resonance in PFDA rats is not completely understood. Glycogen synthesis from glucose can occur either via a direct pathway, in which glucose is converted directly to glycogen (glucose glucose 6-phosphate (G6P) glycogen), or via an indirect pathway, in which glucose is first catabolized to gluconeogenicthree-carbon precursors (i.e., triose phosphates, pyruvate, or lactate) and then G6P C3 compounds converted to glycogen (glucose-, G6P glycogen). Glycogen synthesis can also occur from alanine via gluconeogenesis (alanine G6P glycogen). In fact, studies have shown that alanine is a more favorable glycogenic substrate than glucose (20, 23,24). Note that the glycogenic pathways of glucose and alanine converge at G6P and proceed via a common pathway to glycogen. The fact that some PFDA rats (5 days postdose) show functional glycogenesis from alanine but not from glucose suggests that the glycogenic pathway from G6P to glycogen remains functional in PFDA rats and that PFDA may inhibit hepatocellular glucose transport and/or glucokinase (GK). This hypothesis was further tested by administering 2-DG to PFDA and control rats and measuring 2-DG6P concentrations. Like glucose, 2-DG is transported into hepatocytes via the glucose transporter and is phosphorylated by GK to form 2-DG6P; however, 2-DG6P is not a substrate for glycolysis nor glycogenesis (25, 26). PFDA rats (Group 3T) show significantly less 2-DG6P than corresponding controls (Group 3C), which is consistent with the hypothesis that PFDA inhibits either the transport of 2-DG into the hepatocyte and/or its phosphorylation by GK. Since 2-DG6P can be dephosphorylated by glucose-6-phosphatase (G6-Pase), it is plausible that PFDA rats may have lower 2-DG6P levels due to increased G6Pase activity; however, Newgard et al. (27) have shown, using rats in uiuo, that hepatic G6Pase activity is inhibited by an intragastric infusion of glucose. Since supraphysiological concentrations of 2-DG were administered in this study, it appears unlikely that G6Pase activity is responsible for the lower levels of 2-DG6P in PFDA rats. PFDA may indirectly alter hepatocellular glucose transport by influencing the nature of the plasma membrane. It has been suggested that PFDA treatment changes membrane composition and fluidity ( 4 , 5 ) . Carruthers et d . have shown that glucose transport activity ( Vmm)is influenced by the nature of the plasma membrane, being altered by bilayer physical state, lipid acyl chain length and degree of unsaturation, and bilayer cholesterol content (28-30). Our data show improved 13C-liver spectral resolution in PFDA rats along with an increase in the intensity of the methylene resonance at 30 ppm relative to control (Figure 1). This may reflect an increase in membrane fluidity; however, it is more likely suggestive of increased liver triglycerides and/or free fatty acids which have also been reported by others (6, 7).
-
- -
-
-
-
With regard to the possibility that PFDA may affect GK activity, it is interesting to note that GK activity is affected by the following: fasting and refeeding ( 3 1 , 3 2 ) ; decreased levels of triiodothyronine (T3) and thyroxine (T4) (33,341;and changes in the concentration of hepatic long-chain fatty acyl-CoAs (35, 36). PFDA may inhibit GK activity since PFDA treatment results in hypophagia ( 4 , 5 ) ,decreased levels of T3 and T4 (37-39), and increased hepatic triglycerides (6, 7). Another means by which PFDA treatment may inhibit glycogenesis is by altering the specific metabolic zone of the liver involved in glycogen synthesis from glucose. Zone 1 hepatocytes (periportal) are most active in terms of gluconeogenesis and glycogen synthesis from lactate, while zone 3 cells (perivenous or centrilobular) are most active in glucose uptake, glycolysis, and glycogen synthesis from glucose ( 4 0 , 4 2 ) . PFDA treatment causes cellular hypertrophy, which may produce partial ischemia by decreasing blood flow in the capillaries (42). Since zone 3 cells are more severely affected by ischemia than zone 1 cells, glycogenesis may remain functional from alanine but not from glucose in PFDA rats due to ischemic changes in the liver. Recent studies from our laboratory using 31PNMR spectroscopy, however, have shown no difference in ATP levels between PFDA and control rats on days 3 or 5 postdose at 50 mg/kg. This result indicates that PFDA does not alter tissue energetic viability nor does it cause significant tissue necrosis at this particular dose and times posttreatment (43). The effect of PFDA on TCA cycle activity was determined by monitoring the flow of label from [3-l3C1alanine into glutamate. The transamination of [ 3 - W alanine involves the conversion of a-ketoglutarate to glutamate and yields [3-13Clpyruvate.Pyruvate enters the TCA cycle either as oxaloacetate via the pyruvate carboxylase (PC) pathway or as acetyl-coA via the pyruvate dehydrogenase (PDH) pathway. To a first approximation, pyruvate entering the TCA cycle as oxaloacetate yields a mixture of [2J3C1- and [3-l3C1-a-ketoglutarate.Due to the equilibria among oxaloacetate, malate, and the symmetric TCA cycle intermediate, fumarate, the 13C-labelis scrambled between carbons 2 and 3. Further transamination of alanine involving [2-l3C1-or [3-13Cl-a-ketoglutarateyields a mixture of glutamate molecules labeled at the CZor C3 positions. Pyruvate entering the TCA cycle as acetylCoA via PDH yields [4J3CI -a-ketoglutarate. Disregarding dilution from unlabeled pools, subsequent turns of the TCA cycle result in a-ketoglutarate labeled at multiple carbon sites (Le.,a second turn of the cycle yields C2,4-and C3,4-a-ketoglutarate). The C4 label is retained through all subsequent turns of the cycle. Further transamination of alanine involving singly labeled or multilabeled a-ketoglutarate yields glutamate labeled at identical carbon positions. Thus, labeling at C4-glutamate occurs only when pyruvate enters the TCA cycle as acetyl-coA and serves as a specific marker for this pathway (19, 44, 45). The relative activity of PC uersus PDH is represented by the ratio of the glutamate NMR signal intensities (C2 + C3)/ C4. PFDA rats show a significantly lower ratio than controls, suggesting a greater relative PDH activity in PFDA rats as compared to controls. PDH activity has been shown to be affected by the ratios of acetyl-CoA/CoA, NADH/NAD+, and the levels of pyruvate and ADP (46,47). Differences in PDH activity between PFDA and control rats may arise from differences
Effect of PFDA on Gluconeogenesis and Glycogenesis
in lipolytic activity. Others have shown a decrease in PDH activity upon starvation due to increased fatty acid (FA) oxidation which results in elevated acetyl-coA levels (48, 49). Therefore, PDH may be less active in control rats due to food deprivation and increased FA oxidation. In PFDA rats, however, PDH activity may be greater because FA oxidation and ketogenesis are inhibited (7, 9).2 This greater PDH activity also suggests that pyruvate serves as the predominant source of acetyl-coA in treated rats. Comparison of glutamate ratios for PFDAgroups (Group 4T versus 5T) and between control groups (Group 4C versus 5C) suggests that the addition of supplementary glucose results in a greater relative activity of PDH uersus PC. A more pronounced decrease in the (C2 + C3)/C4 glutamate ratio is observed in the control groups than between PFDA groups. The reason for this is not known. In conclusion, data from this study indicate that PFDA treatment affects hepatic carbohydrate metabolism in rats by inhibiting glycogen synthesis from glucose and altering both glucose and alanine metabolism. This glycogenic inhibition is probably attributed to an inhibitory effect of PFDA on glucose transport and/or GK activity. The data also suggest that hepatic PDH activity is increased in PFDA-treated rats relative to controls. This is thought to reflect a dysfunction in FA metabolism in treated rats. Further studies are currently in progress in our laboratory to investigate the effects of PFDA on other hepatocellular processes.
Acknowledgment. The authors would like to thank Latha Narayanan for her assistance with 2-DG6P analyses and the Department of Mathematics and Statistics, Wright State University, for assistance with statistical analyses of the data. This work was supported by the Air Force Office of ScientificResearch, Air Force Systems Command, USAF, under grant or cooperative agreement number AFOSR-90 0148.
References (1) Gibson, G. G., Orton, T. C., and Tamburini, P. P. (1982) Cytochrome P-450 induction by clofibrate. Purification and properties of a hepatic cytochrome P-450 relatively specific for the 12- and 11hydroxylation of dodecanoic acid (lauric acid). Biochem. J. 203, 161-168. (2) Sharma, R., Lake, B. G., Foster, J., and Gibson, G. G. (1988) Microsomal cytochrome P-452 induction and peroxisome proliferation by hypolipidaemic agents in rat liver. A mechanistic interrelationship. Biochem. Pharmacol. 37, 1193-1201. (3) Sharma, R., Lake, B. G., and Gibson, G. G. (1988) Co-induction of microsomal cytochrome P-452 and the peroxisomal fatty acid @-oxidationpathway in the rat by clofibrate and di-(2-ethylhexyl) phthalate. Biochem. Pharmacol. 37, 1203-1206. (4) George, M. E., and Andersen, M. E. (1986) Toxic effects of nonadecafluoro-n-decanoic acid in rats. Toxicol.Appl. Pharmacol. 85, 169-180. (5) Olson, C. T., and Andersen, M. E. (1983) The acute toxicity of perfluormtanoic and perfluorodecanoicacids in male rats and effects on tissue fatty acids. Toxicol. Appl. Pharmacol. 70,362-372. (6) Davis, J. W., VandenHeuvel,J. P., andPeterson,R. E. (1991) Effects of perfluorodecanoic acid on de nouo fatty acid and cholesterol synthesis in the rat. Lipids 26, 857-859. (7) Van Rafelghem, M. J., Vanden Heuvel, J. P., Menahan, L. A., and Peterson, R. E. (1988) Perfluorodecanoic acid and lipid metabolism in rats. Lipids 23, 671-678. (8) Van Rafelghem, M. J., Mattie, D. R., Bruner, R. H., and Andersen, M. E. (1987) Pathological and hepatic ultrastructural effects of a single dose of perfluoro-n-decanoic acid in the rat, hamster, mouse, and guinea pig. Fundam. Appl. Toxicol. 9, 522-540. (9) Vanden Heuvel, J. P., Kuslikis, B. I., Shrago, E., and Peterson, R. E. (1991) Inhibition of long-chain acyl-CoA synthetase by the peroxisome proliferator perfluorodecanoic acid in rat hepatocytes. Biochem. Pharmacol. 42, 295-302.
Chem. Res. Toxicol., Vol. 7, No. 1, 1994 21 (10) Singer, S. S., Andersen, M. E., and George, M. E. (1990) PerfluoroN-decanoic acid effects on enzymes of fatty acid metabolism. Toxicol. Lett. 54, 39-46. (11) Borges, T., Glauert, H. P., and Robertson, L. W. (1993) Perfluorodecanoic acid noncompetitively inhibits the peroxisomal enzymes enoyl-CoA hydratase and 3-hydroxyacyl-CoA dehydrogenase. Toxicol. Appl. Pharmacol. 118, 8-15. (12) Borges, T., Glauert, H. P., Chen, L. C., Chow, C. K., and Robertson, L. W. (1990) Effect of the peroxisome proliferator perfluorodecanoic acid on growth and lipid metabolism in Sprague Dawley rata fed three dietary levels of selenium. Arch. Toxicol. 64, 26-30. (13) Goecke, C. M., Jarnot, B. M., and Reo, N. V. (1992) A comparative toxicological investigation of perfluorocarboxylic acids in rats by fluorine-I9 NMR spectroscopy. Chem. Res. Toxicol. 5, 512-519. (14) Reo, N. V.,Ewy, C. S., Siegfried, B. A., and Ackerman, J. J. H. (1984) High-field 1sC NMR spectroscopy of tissue in uiuo. A doubleresonance surface-coil probe. J.Magn. Reson. 58, 76-84. (15) Cohen, S. M. (1983) Simultaneous l3C and 3lP NMR studies of perfused rat liver. Effects of insulin and glucagon and a 1*C NMR assay of free Mgz+. J.Biol. Chem. 258, 14294-14308. (16) Manchester, J. K., Chi, M. M., Carter, J. G., Pusateri, M. E., McDougal, D. B., and Lowry, 0. H. (1990) Measurement of 2-deoxyglucose and 2-deoxyglucose 6-phosphate in tissues. Anal. Biochem. 185,118-124. (17) Rice, W. R. (1989) Analyzing tables of statistical tests. Evaluation 43, 223-225. (18) Canioni, P., Alger, J. R., and Shulman, R. G. (1983) Natural abundance carbon-13 nuclear magnetic resonance spectroscopy of liver and adipose tissue of the living rat. Biochemistry 22,49744980. (19) Cohen, S. M., Shulman, R. G., and McLaughlin, A. C. (1979) Effects of ethanol on alanine metabolism in perfused mouse liver studied by NMR. h o c . Natl. Acad. Sci. U.S.A. 76, 4808-4812. (20) Katz, J., Kuwajima, M., Foster, D. W., and McGarry, J. D. (1986) The glucose paradox: new perspectives on hepatic carbohydrate metabolism. Trends Biochem. Sci. 11, 136-140. (21) Shikama, H., and Ui, M. (1978) Glucose load diverts hepatic gluconeogenic product from glucose to glycogen in uiuo. Am. J. Physiol. 235, E354-E360. (22) Youn, J. H., Kaslow, H. R., and Bergman, R. N. (1987) Fructose effect to suppress hepatic glycogen degradation. J. Biol. Chem. 262, 11470-11477. (23) Shalwitz, R. A., Reo, N. V., Becker, N. N., Hill, A. C., Ewy, C. S., and Ackerman, J. J. H. (1989) Hepatic glycogen synthesis from duodenal glucose and alanine. An in situ l3C NMR study. J. Biol. Chem. 264, 3930-3934. (24) Youn, J. H., and Bergman, R. N. (1990) Enhancement of hepatic glycogen by gluconeogenic precursors: substrate flux or metabolic control? Am. J. Physiol. 258, E899-E906. (25) Jacobs, A. E. M., Oosterhof, A.,andVeerkamp, J. H. (1990) 2-DeoxyD-glucoseuptake in cultured human muscle cells. Biochim.Biophys. Acta 1051, 230-236. (26) Hawkins, R. A., and Miller, A. L. (1987) Deoxyglucose-6-phosphate stability in uiuo and the deoxyglucose method. J. Neurochem. 49, 1941-1949. (27) Newgard, C. B., Foster, D. W., and McGarry, J. D. (1984) Evidence for suppression of hepatic glucose-6-phosphatasewith carbohydrate feeding. Diabetes 33, 192-195. (28) Carruthers,A., andMelchior,D. L. (1984)Human erythrocyte hexose transporter activity is governed by bilayer lipid composition in reconstituted vesicles. Biochemistry 23, 6901-6911. (29) Carruthers, A., and Melchior, D. L. (1986) How bilayer lipids affect membrane protein activity. Trends Biochem. Sci. 11, 331-335. (30) Carruthers, A., and Melchior, D. L. (1988) Role of bilayer lipids in governing membrane transport processes. In Lipid Domains and the Relationship to Membrane Function (Aloia, R. C., Curtain, C. C., and Gordon, L. M., Eds.) pp 201-225, Alan R. Lisa, New York. (31) Zahner, D., and Malaisse, W. J. (1990) Kinetic behaviour of liver glucokinase in insulinopenic situations: effect of fructose l-phosphate in fed and starved rats. Biochimie 72,715-718. (32) Newgard, C. B., Quaade, C., Hughes, S. D., and Milburn, J. L. (1990) Glucokinase and glucose transporter expression in liver and islets: Trans. implications for control of glucose homeostasis. Biochem. SOC. 18,851-853. (33) Sibrowski,W.,Muller,M.J.,andSeitz,H. J. (1981)Effectofdifferent thyroid states on rat liver glucokinase synthesis and degradation in uiuo. J. Biol. Chem. 256, 9490-9494. (34) Minderop, R. H., Hoeppner, W., and Seitz, H. J. (1987) Regulation of hepatic glucokinase gene expression. Role of carbohydrates, and glucocorticoid and thyroid hormones. Eur. J. Biochem. 164, 181187. (35) Dawson, C. M., and Hales, C. N. (1969) The inhibition of rat liver glucokinase by palmitoyl-CoA. Biochim. Biophys. Acta 176,657659.
22 Chem. Res. Toxicol., Vol. 7,No.1, 1994 (36) Tippett, P. S.,and Neet, K. E. (1982) Specific inhibition of glucokinaseby long chain acyl coenzymesA below the critical micelle concentration. J. Biol. Chem. 267, 12839-12845. (37) Gutshall,D. M., Pilcher, G. D., and Langley,A. E. (1989)Mechanism of the serum thyroid hormone lowering effect of perfluoro-n-decanoic acid (PFDA) in rats. J. Toricol. Environ. Health 28, 53-65. (38) Kelling, C. K., Van Rafelghem, M. J., Menahan, L. A., and Peterson, R. E. (1987)Effects of perfluorodecanoic acid on hepatic indices of thyroid status in the rat. Biochem. Pharmacol. 36,1337-1344. (39)Van Rafelghem, M. J., Inhorn, S. L., and Peterson, R. E. (1987) Effects of perfluorodecanoic acid on thyroid status in the rat. Toxicol. Appl. Pharmacol. 87,430-439. (40)Jungermann, K. (1987)Metabolic zonation of liver parenchyma: Significance for the regulation of glycogen metabolism, gluconeogenesis, and glycolysis. Diabetes Metab. Rev. 3,269-293. (41) Gumucio, J. J., and Miller, D. L. (1981)Functional implications of liver cell heterogeneity. Gastroenterology 80, 393-403. (42) Dianzani, M. U. (1991)Biochemical Aspects of Fatty Liver. In Hepatotoxicology (Meeke, R. G., Harrison, S. D., and Bull, R. J., Eds.) pp 327-400,CRC Press, Ann Arbor. (43) Reo, N. V., Goecke, C. M., Narayanan, L., and Jarnot, B. M. Effects of perfluoro-n-octanoic acid, perfluoro-n-decanoic acid, and clofibrate on hepatic phosphorus metabolism in rats and guinea pigs in vivo. Toricol. Appl. Pharmacol. 124 (in press).
Goecke et al. (44) Malloy, C. R., Sherry, A. D., and Jeffrey, F. M. H. (1988)Evaluation of carbon flux and substrate selection through alternate pathways involvingthe citric acid cycle of the heart by 1sCNMRspectroscopy. J. Biol. Chem. 263, 6964-6971. (45) Cohen, S.M. (1987)I3C NMR study of effects of fastingand diabetes on the metabolism of pyruvate in the tricarboxylic acid cycle and of the utilization of pyruvate and ethanol in lipogenesis in perfused rat liver. Biochemistry 26, 581-589. (46) McCormack, J. G., Longo, E. A,, and Corkey, B. E. (1990)Glucoseinduced activation of pyruvate dehydrogenase in isolated rat pancreatic islets. Biochem. J. 267,527-530. (47) Reed, L. J., Damuni, Z., and Merryfield, M. L. (1985)Regulation of mammalian pyruvate and branched-chain a-keto acid dehydrogenase complexes by phosphorylation-dephosphorylation. In Current Topics in Cellular Regulation. Modulation by Covalent Modification (Shaltiel, S., and Chock, P. B., Eds.) pp 41-49,Academic Press, New York. (48) Randle, P. J. (1986)Fuel selection in animals. Biochem. SOC. Trans. 14, 799-806. (49)Wieland, 0.H., Patzelt, C., and Loffler,G. (1972)Active and inactive forms of pyruvate dehydrogenase in rat liver. Effect of starvation and refeeding and of insulin treatment on pyruvate-dehydrogenase interconversion. Eur. J.Biochem. 26,426-433.
