- Email: [email protected]
Metabolic effects and distribution space of flufenamic acid in the isolated perfused rat liver
Chemico-Biological Interactions 116 (1998) 105 – 122
Metabolic effects and distribution space of flufenamic acid in the isolated perfused rat liver Carlos H. Lopez, Adelar Bracht *, Nair S. Yamamoto, Mariana D. dos Santos Laboratory of Li6er Metabolism, Department of Biochemistry, Uni6ersity of Maringa´, Maringa´, 87020900, Brazil Received 26 March 1998; received in revised form 1 June 1998; accepted 24 August 1998
Abstract The following aspects were investigated in the present work: (a) the action of flufenamic acid on hepatic metabolism (oxygen uptake, glycolysis, gluconeogenesis, uricogenesis and glycogenolysis), (b) the action of flufenamic acid on the cellular adenine nucleotide levels, and (c) the transport and distribution space of flufenamic acid in the liver parenchyma. The experimental system was the isolated perfused rat liver. Perfusion was accomplished in an open, non-recirculating system. The perfusion fluid was Krebs/Henseleit-bicarbonate buffer (pH 7.4), saturated with a mixture of oxygen and carbon dioxide (95:5) by means of a membrane oxygenator and heated to 37°C. The distribution space of flufenamic acid was measured by means of the multiple-indicator dilution technique with constant infusion (step input) of [3H]water plus flufenamic acid. The results of the present work indicate that the metabolic effects of flufenamic acid are the consequence of an uncoupling of oxidative phosphorylation, a conclusion based on the following observations: (a) flufenamic acid increased oxygen uptake, a common property of all uncouplers; (b) the drug also increased glycolysis and glycogenolysis in livers from fed rats (these are expected compensatory phenomena for the decreased mitochondrial ATP formation); (c) flufenamic acid inhibited glucose production from fructose, an energy-dependent process; (d) the cellular ATP levels were decreased by flufenamic acid whereas the AMP levels were increased; and (e) the total adenine nucleotide content was decreased by flufenamic acid and uric acid production was stimulated. Indicator-dilution experiments with flufenamic acid revealed that this substance undergoes flow-limited distribution in the liver and that its apparent distribution space * Corresponding author. Fax: + 55 44 2614396/2633655; e-mail: [email protected] 0009-2797/98/$ - see front matter © 1998 Published by Elsevier Science Ireland Ltd. All rights reserved. PII S0009-2797(98)00084-2
106
C.H. Lopez et al. / Chemico-Biological Interactions 116 (1998) 105–122
greatly exceeds the aqueous space of the liver. Flufenamic acid changed its behaviour when the portal concentration was increased from 25 to 50 mM. At 25 mM the initial upslope of the outflow profile clearly preceded that of all other concentrations. From the trend of the curves obtained with 50, 100 and 250 mM, one would expect an initial upslope situated at the right of the 50-mM curve. Furthermore, the time of appearance of flufenamic acid in the outflowing perfusate was practically the same irrespective of the portal concentration. For theoretical reasons one would expect progressively longer appearance times when the portal concentration was decreased. It is possible that the amount of flufenamic acid bound to the cell membranes during the early stages of the infusion produced changes that enabled these structures to bind a larger quantity of the drug than originally possible. © 1998 Published by Elsevier Science Ireland Ltd. All rights reserved. Keywords: Rat liver perfusion; Flufenamic acid; Metabolism; ATP; Distribution space
1. Introduction It has been shown that several non-steroidal anti-inflammatory agents like diflunisal, mefenamate, flufenamate and niflumate, act as uncouplers of oxidative phosphorylation in mitochondria isolated from rat liver [1–3]. Drugs that affect mitochondrial energy metabolism also produce several changes in cellular metabolism. Brass and Garrity [4] demonstrated that ibuprofen, indomethacin and meclofenamate are able to increase glycogenolysis in isolated hepatocytes by a mechanism that is independent of cyclooxygenase inhibition. Brass and Garrity [4] interpreted this action as a consequence of cytosolic calcium redistributions. Work from our laboratory, performed with the isolated perfused rat liver, has shown that mefenamate [5] and niflumate [6] are able to stimulate oxygen consumption, glycogenolysis, fructolysis and glycolysis. Gluconeogenesis, however, was inhibited [5,7,8]. These effects were attributed to the uncoupling action, which should cause a reduction in the ATP levels. A Ca2 + -mediated mechanism, on the other hand, was shown to be improbable by Nascimento et al. [9,10] in two studies performed with diclofenac and niflumic acid. It has been demonstrated in these studies that all metabolic effects of these drugs are exerted to the same extent either in the presence or in the absence of Ca2 + . As an extension of previous work performed in our laboratory, the present investigation was undertaken to investigate the metabolic effects of flufenamic acid, its distribution space and uptake rates in the perfused rat liver. All these aspects have not yet been investigated with flufenamic acid; the results should thus be useful for comparative purposes. The present study also offers the opportunity for investigating the influence of a non-steroidal anti-inflammatory agent on the cellular levels of adenine nucleotides, which should be decreased if these compounds really act as uncoupling agents. Up to now only indirect evidence is available concerning this phenomenon, which certainly deserves to be precisely quantified.
C.H. Lopez et al. / Chemico-Biological Interactions 116 (1998) 105–122
107
2. Materials and methods
2.1. Li6er perfusion Male albino rats (Wistar), weighing 180–220 g, were fed ad libitum with a standard laboratory diet (Purina®, Sa˜o Paulo, Brazil). For the surgical procedure the rats were anesthetized by intraperitoneal injection of sodium pentobarbital (50 mg/kg). Hemoglobin-free, non-recirculating perfusion was performed. The surgical technique was that described by Scholz and Bu¨cher [11]. After cannulation of the portal and cava veins the liver was positioned in a plexiglass chamber. The flow was maintained constant by a peristaltic pump. The perfusion fluid was Krebs/ Henseleit-bicarbonate buffer (pH 7.4), saturated with a mixture of oxygen and carbon dioxide (95:5) by means of a membrane oxygenator with simultaneous temperature adjustment at 37°C.
2.2. Analytical Samples of the effluent perfusion fluid were collected according to the experimental protocol and analyzed for their metabolite contents. The following compounds were assayed by means of standard enzymatic procedures: glucose [12], L-lactate [13] and pyruvate [14], ATP [15], ADP and AMP [16], and uric acid [17]. The lactate, pyruvate, AMP, ADP and ATP concentrations were calculated from spectrophotometric changes at 340 nm due to NAD + (or NADP + ) reduction or NADH oxidation. The extinction coefficient of NADH/NADPH (6.22 mM − 1 cm − 1) was used in the calculations. Glucose and uric acid were calculated from standard curves. A new standard curve was constructed for each assay. Linearity between concentration and absorbance was up to 1.0 absorbance units for glucose and up to 0.3 for uric acid. The oxygen concentration in the outflowing perfusate was monitored continuously, employing a Teflon-shielded platinum electrode adequately positioned in a Plexiglass chamber at the exit of the perfusate [18]. Metabolic rates were calculated from input-output differences and the total flow rates and were referred to the wet weight of the liver.
2.3. Ca 2 + -free perfusion For performing Ca2 + -free perfusion, the intracellular Ca2 + pools were exhausted. The following procedure was adopted. Livers were pre-perfused with Ca2 + -free Krebs/Henseleit-bicarbonate buffer containing 0.2 mM ethylenediamine tetraacetate (EDTA), 5 mM glucose, 1 mM lactate and 0.1 mM pyruvate. In order to ensure maximal depletion of the intracellular Ca2 + pools, phenylephrine (2 mM) was infused repeatedly (three times) during short periods of 2 min, with intervals of 5 min. According to Reinhart et al. [19], this procedure depletes the intracellular Ca2 + -pools which are normally mobilized when hormones are infused.
C.H. Lopez et al. / Chemico-Biological Interactions 116 (1998) 105–122
108
2.4. Measurement of the distribution space of flufenamic acid Indicator-dilution experiments were performed employing the constant infusion method (step input). [3H]Water and flufenamic acid were infused simultaneously into the liver prior to entering the portal vein. During the first 90 s after the onset of the infusion, the perfusate was fractionated in 0.5–2.0-s fractions by means of a specially designed fraction collector [18]. After the first 90 s, samples were taken manually in intervals ranging from 30 to 600 s. Aliquots of these samples were used for the assay of [3H]water and for the extraction and assay of flufenamic acid. Flufenamic acid was extracted from the perfusate samples with ethyl acetate and assayed spectrophotometrically at 285 nm (extinction coefficient= 17 mM − 1 cm − 1 in ethyl acetate) essentially as described previously for niflumic acid [20].
2.5. Treatment of data and calculations The statistical significance of the differences between parameters was evaluated by means of Student’s t-test or Student’s paired t-test, according to the context. Different steady-state values were compared and the analysis was performed by means of the Primer program (version 1.0, McGraw-Hill, 1988). PB 0.05 was adopted as a criterion of significance. Mean transit times of flufenamic acid (t( flu) and [3H]water (t( water) were calculated employing the formula of Meier and Zieler [21], i.e. t( =
&
[1 −Q(t)]dt
(1)
t0
where Q(t) is the normalized experimental outflow profile (effluent concentration/portal concentration), t is the time and t0 the transit time in the large vessels and catheters. The latter can be evaluated as the time of appearance of label in the outflowing perfusate. The amount of non-transformed flufenamic acid retained by the liver in excess to the aqueous space (R) was calculated as R=
&
FCp/P[Qwater(t) − Qflu(t)]dt
(2)
t0
where F is the perfusate flow through the liver (volume/time), Cp the portal flufenamic acid concentration, P the weight of the liver, Qwater(t) the normalized outflow profile of [3H]water and Qflu(t) the normalized outflow profile of flufenamic acid. Numerical integration was performed by means of the trapezoid rule [22].
2.6. Materials The liver perfusion apparatus and the rapid sampling apparatus were built in the workshops of the University of Maringa´. All enzymes and coenzymes used in the enzymatic assays were purchased from Sigma (St. Louis, USA). [3H]Water (NET001C, 25 mCi g − 1) was purchased from E.I. du Pont de Nemours (Boston, MA). All chemicals were from the best available grade.
C.H. Lopez et al. / Chemico-Biological Interactions 116 (1998) 105–122
109
3. Results
3.1. Effects of flufenamic acid on glycogen catabolism and oxygen uptake The effects of flufenamic acid on glycogen catabolism are illustrated in Figs. 1 and 2. Fig. 1 also illustrates the experimental protocol. Livers from fed rats were perfused with substrate-free Krebs/Henseleit-bicarbonate buffer (pH 7.4) saturated with a mixture of oxygen/carbon dioxide (95/5%). Glucose, lactate and pyruvate
Fig. 1. Time-course of the effects of flufenamic acid on glycogen catabolism and related parameters. Livers from fed rats were perfused with Krebs/Henseleit-bicarbonate buffer (pH 7.4) in an open system. Flufenamic acid (100 mM) was infused during the time intervals indicated by the horizontal bars. Samples of the effluent perfusate were taken at 2-min intervals for the measurement of glucose, lactate and pyruvate and for monitoring absorbance at 285 nm. Oxygen uptake was followed polarographically. The rates of glucose, lactate and pyruvate production and of oxygen consumption were expressed as mmol min − 1 (g liver wet weight) − 1 and represented against the perfusion time. Zero in the time scale is the instant at which the first sample of the effluent perfusate was collected. Values are the means of three liver perfusion experiments with identical protocol.
110
C.H. Lopez et al. / Chemico-Biological Interactions 116 (1998) 105–122
Fig. 2. Concentration dependence of the effects of flufenamic acid on glycogen catabolism and related parameters. Livers from fed rats were perfused with Krebs/Henseleit-bicarbonate buffer (pH 7.4) with different flufenamic acid concentrations in the range between 25 and 250 mM. The experimental protocol was the same illustrated in Fig. 1. The changes in oxygen uptake, lactate production and glucose production caused by flufenamic acid after 20-min infusion were evaluated and represented against the flufenamic acid concentration in the portal perfusate. Each data point represents the mean of three liver perfusion experiments. Bars are standard errors of the mean.
release were monitored as indicators of glycogen catabolism. Flufenamic acid was infused at a final portal concentration of 100 mM during 24 min. The introduction of flufenamic acid produced great changes in all parameters: glucose and lactate release were increased and remained elevated during the whole infusion time. Pyruvate production was only transiently increased with a peak value at 6 min after the onset of the infusion. Oxygen uptake was also increased with a maximum at 24 min and a small decline during the rest of the infusion period. The effects of flufenamic acid were reversible, i.e. cessation of the infusion restored the original levels of the variables. Flufenamic acid absorbs in the ultra-violet range with a peak around 285 nm. For this reason, absorbance at 285 nm was also monitored as a measure for the appearance of the drug in the venous perfusate. Flufenamic acid was infused at a constant rate, but its appearance in the outflowing perfusate was considerably delayed. The response of the liver to other concentrations of flufenamic acid can be inferred from Fig. 2 which shows the changes produced by several flufenamic acid concentrations at the end of the infusion period. Oxygen uptake, glucose release and lactate release stimulations are all presenting very clear and similar dose-effect relationships. Oxygen uptake stimulation varied between 0.22 mmol min − 1 g − 1 at 25 mM flufenamic acid and 0.87 mmol min − 1 g − 1 at 250 mM. Glucose and lactate release stimulations were relatively small at 25 mM flufenamic acid. At 250 mM, however, both parameters were stimulated by approximately 4 mmol min − 1 g − 1. In
C.H. Lopez et al. / Chemico-Biological Interactions 116 (1998) 105–122
111
the case of lactate release, this means an excess of ATP production in the glycolytic pathway of at least 6 mmol min − 1 g − 1. In order to see if Ca2 + influences the metabolic effects of flufenamic acid, experiments were performed in which the intracellular Ca2 + pools were exhausted by repeated phenylephrine pulses in the presence of 0.2 mM EDTA, as described in Section 2. After this procedure 100 mM flufenamic acid was infused during 14 min in the absence of Ca2 + . After this time the Ca2 + -free Krebs/Henseleit bicarbonate buffer was replaced by a normal perfusion fluid containing 2.5 mM CaCl2. Oxygen uptake, glucose release and lactate production were measured. The changes produced by flufenamic acid in the absence and presence of Ca2 + are shown in Table 1. It is clear from the results that Ca2 + has no influence on the metabolic effects of flufenamic acid.
3.2. Effects of flufenamic acid on fructose metabolism Fructose is rapidly transformed into glucose and lactate plus pyruvate in the liver. The first process is energy-dependent and the second one produces ATP. Fructose is, consequently, a suitable substrate for investigating the possible effect of flufenamic acid on energy metabolism. Fig. 3 shows the time-course of the effects of 100 mM flufenamic acid on fructose metabolism and Fig. 4 summarizes the effects of several concentrations of the drug. The experiments with fructose were performed with livers from fasted animals in order to avoid interference by glycogen catabolism. The introduction of fructose produced the known changes in oxygen uptake, and glucose, lactate and pyruvate production [23,24]. The subsequent introduction of flufenamic acid further increased oxygen uptake, but glucose production was considerably decreased. Pyruvate production was also decreased, but the effect on lactate production was more complex. Initially lactate production suffered a rapid increase which was followed by a
Table 1 Influence of Ca2+ on the metabolic effects of flufenamic acid Perfusion condition
Increment in glucose release (mmol min−1 g−1)
Increment in lactate production (mmol min−1 g−1)
Increment in oxygen consumption (mmol min−1 g−1)
2.5 mM Ca2+ Ca2+-free
2.2190.44 (n=4) 2.229 0.36 (n= 4)
2.5390.31 (n =4) 2.469 0.26 (n = 4)
0.6290.07 (n = 4) 0.649 0.05 (n = 4)
The intracellular Ca2+ pools were exhausted by repeated phenylephrine pulses in the presence of 0.2 mM EDTA, as described in Section 2. After this procedure 100 mM flufenamic acid was infused during 14 min in the absence of Ca2. After this time the Ca2+-free Krebs/Henseleit bicarbonate buffer was replaced by a normal perfusion fluid containing 2.5 mM CaCl2. Oxygen uptake, glucose release and lactate production were measured. The basal values (before flufenamic acid infusion) were subtracted from the corresponding values found in the presence of flufenamic acid.
112
C.H. Lopez et al. / Chemico-Biological Interactions 116 (1998) 105–122
Fig. 3. Time-course of the effects of flufenamic acid on fructose metabolism. Livers from 24-h fasted rats were perfused with Krebs/Henseleit-bicarbonate buffer (pH 7.4) in an open system. Fructose (5.0 mM) and flufenamic acid (100 mM) were infused during the times indicated by the horizontal bars. Samples of the effluent perfusate were taken at 2-min intervals for the measurement of glucose, lactate and pyruvate and for monitoring absorbance at 285 nm. Oxygen uptake was followed polarographically. The rates of glucose, lactate and pyruvate production and of oxygen consumption were expressed as mmol min − 1 (g liver wet weight) − 1 and represented against the perfusion time. Zero in the time scale is the instant at which the first sample of the effluent perfusate was collected. Values are the means of three liver perfusion experiments with identical protocol.
progressive decrease to lower values. The latter were still above the basal levels (i.e. in the absence of flufenamic acid). All these changes were perfectly reversible. The concentration dependence in Fig. 4 reveals that glucose production was already substantially inhibited by 25 mM flufenamic acid (− 0.6 mmol min − 1 g − 1). No oxygen uptake stimulation was found at this concentration. Further increases in concentration, however, were stimulative. Maximal stimulation, nearly 0.5 mmol min − 1 g − 1, was below the maximal stimulation found with substrate free-perfused livers from fed rats (0.87 mmol min − 1 g − 1; Fig. 2). At 250 mM flufenamic acid, glucose production was decreased by approximately 2.0 mmol min − 1 g − 1; this decrease means more than 90% inhibition.
C.H. Lopez et al. / Chemico-Biological Interactions 116 (1998) 105–122
113
3.3. The action of flufenamic acid on adenine nucleotide le6els and uric acid metabolism The metabolic effects of flufenamic acid are consistent with an action on the mitochondrial energy metabolism. If this conclusion is correct, flufenamic acid should have a profound influence on the adenine nucleotide levels. For this reason, the cellular levels of AMP, ADP and ATP were measured in livers from fed and fasted rats in the absence and after 20 min of 100 mM flufenamic acid infusion. The results are shown in Table 2. For comparative purposes, the effects of cyanide, a compound that blocks the respiratory chain, are also shown. Table 2 reveals that flufenamic acid diminished the levels of ATP and increased those of AMP in livers from both ad libitum fed and 24-h fasted rats. The levels of ADP were not changed, but the ATP to ADP and ATP to AMP ratios were substantially decreased. The decrease in the ATP levels was less pronounced in livers from fed rats, a phenomenon that can be attributed to the increased glycolysis. In livers from fasted rats glycolysis was minimal due to the fact that substrate-free perfusion fluid was used. The sum of the contents of AMP, ADP and ATP was diminished by flufenamic acid in both livers from fed and fasted rats. Here again, the change was more pronounced in livers from fasted rats. Concerning the adenine nucleotide levels, the effects of flufenamic acid were similar to those of 1.0 mM cyanide, but with some significant differences: (a) contrary to flufenamic acid, cyanide did not
Fig. 4. Concentration dependence of the effects of flufenamic acid on glucose production from fructose and on oxygen uptake. Livers from fasted rats were perfused with Krebs/Henseleit-bicarbonate buffer (pH 7.4) containing 5 mM fructose and with different flufenamic acid concentrations in the range between 25 and 250 mM. The experimental protocol was the same illustrated in Fig. 3. The changes in glucose production and oxygen uptake caused by flufenamic acid after 20-min infusion were evaluated and represented against the flufenamic acid concentration in the portal perfusate. Each data point represents the mean of three liver perfusion experiments. Bars are standard errors of the mean.
Fasted Fasted Fasted Fed Fed Fed
1 (n= 3) 2 (n=5) 3 (n =5) 4 (n=3) 5 (n=6) 6 (n= 3)
— 100 —
— 100 —
Flufenamic acid (mM)
— — 1.0
— — 1.0
Cyanide (mM)
ADP (mmol g−1)
0.85 9 0.06 0.93 90.03 0.95 90.14 0.73 90.02 0.829 0.03 0.96 90.02*
ATP (mmol g−1)
2.14 90.05 0.56 90.10* 0.06 90.01* 2.13 9 0.04 1.49 90.08* 1.88 90.07*
0.17 90.02 0.339 0.04* 0.26 9 0.06
0.22 9 0.02 0.459 0.05* 0.98 9 0.15*
AMP (mmol g−1)
3.049 0.02 2.6390.09* 3.099 0.04
3.229 0.12 1.9490.07* 1.999 0.16*
Total adenine nucleotides (mmol g−1)
2.92 9 0.12 1.82 90.13* 1.96 90.10*
2.52 90.15 0.60 9 0.1* 0.06 90.01*
ATP/ADP
9 0.007* 12.3891.38 4.529 0.77* 7.239 2.35
9.739 0.71 1.2490.35* 0.061
ATP/AMP
Livers were perfused in an open system as described in Section 2. Flufenamic acid or cyanide was infused during 20 min as illustrated in Fig. 5. For the determination of the adenine nucleotide levels the livers were freeze-clamped in liquid nitrogen and extracted with cold perchloric acid. Determination was accomplished by standard enzymatic procedures. * Significantly different from the corresponding controls according to Student’s t-test (PB0.05).
Metabolic condition of the rat
Experiment no.
Table 2 Influence of flufenamic acid and cyanide on ATP, ADP and AMP levels in livers from fed and 24-h fasted rats
114 C.H. Lopez et al. / Chemico-Biological Interactions 116 (1998) 105–122
C.H. Lopez et al. / Chemico-Biological Interactions 116 (1998) 105–122
115
Fig. 5. Time-course of the effects of flufenamic acid on uric acid production. Livers from 24-h fasted rats were perfused with Krebs/Henseleit-bicarbonate buffer (pH 7.4) in an open system. Flufenamic acid (100 mM) was infused as indicated. Samples of the effluent perfusate were taken at 2-min intervals for the measurement of uric acid by means of an enzymatic procedure. The rate of uric acid production was expressed as nmol min − 1 (g liver wet weight) − 1 and represented against the perfusion time. Zero in the time scale is the instant at which the first sample of the effluent perfusate was collected. Values are the means of three liver perfusion experiments with identical protocol.
affect the sum of the contents of AMP, ADP and ATP in livers from fed rats; (b) flufenamic acid had a more pronounced effect than cyanide on the ATP levels in livers from fed rats; the opposite occurred, however, in livers from 24-h fasted rats. The diminution of the total adenine nucleotide levels could be the consequence of an enhancement of purine nucleotide metabolism, which usually occurs when the degree of phosphorylation is decreased. Dephosphorylation of the adenine nucleotides usually generates multiple forms (inosine, hypoxanthine, etc.) which, contrary to the phosphorylated forms, can permeate the cell membrane and, consequently, escape from the hepatic cellular space. If this occurs when flufenamic acid is given to the liver, uric acid release should be increased when the drug is introduced, because this compound is one of the end products of purine nucleotides metabolism. In order to test this hypothesis, uric acid production was measured in livers from fasted rats, before and after the introduction of flufenamic acid. The results are shown in Fig. 5 and they reveal that flufenamic acid caused a great increase (in relative terms) in uric acid release, which passed from practically zero to a maximum of 11 nmol min − 1 g − 1 at the sixth minute of infusion. At the end of the infusion uric acid production stabilized at the much lower value of 2 nmol
116
C.H. Lopez et al. / Chemico-Biological Interactions 116 (1998) 105–122
min − 1 g − 1. The total extra amount of uric acid released during the infusion of flufenamic acid was equal to 90 nmol g − 1, as evaluated from the area under the time – response curve in Fig. 5.
3.4. Distribution space of flufenamic acid in the li6er The absorbance measurements at 285 nm during the infusion of flufenamic acid (Figs. 1 and 3) suggest that this compound undergoes flow-limited distribution in the liver [25]. In order to obtain more precise information about this topic, we have performed indicator-dilution experiments of the kind illustrated in Fig. 6. In this series of experiments flufenamic acid was simultaneously infused with [3H]water as an indicator for the total aqueous space of the liver. Flufenamic acid was extracted with ethyl acetate. The curves in Fig. 6 were normalized, i.e. the concentration in the effluent perfusate at the different times was divided by the concentration in the portal vein. Fig. 6 reveals that the time elapsed between the onset of the infusion and the time of appearance of the substance in the outflowing perfusate (appearance time) was not the same for both substances. No significant amount of flufenamic acid appeared in the effluent perfusion fluid before [3H]water. In the range between 250
Fig. 6. Indicator-dilution experiments with various portal concentrations of flufenamic acid. Livers from fed rats were perfused with Krebs/Henseleit-bicarbonate buffer (pH 7.4). Flufenamic acid, at various concentrations, and [3H]water (2.5 mCi min − 1) were infused simultaneously. The venous perfusate was fractionated in 0.5- to 600-s fractions. [3H]Water was determined by liquid scintillation spectrometry. Flufenamic acid was extracted with ethyl acetate and determined spectrophotometrically at 285 nm. The venous/portal concentration ratios of each substance were plotted versus the time after onset of the infusion. The portal flufenamic acid concentrations were: 250 mM ( – ); 100 mM (– ); 50 mM ( –); and 25 mM ("–"). For clarity, only the [3H]water curve obtained with 250 mM flufenamic acid is shown ( – ).
C.H. Lopez et al. / Chemico-Biological Interactions 116 (1998) 105–122
117
Table 3 Mean transit times of flufenamic acid in the isolated perfused rat liver Portal flufenamic acid concentration (mM)
Flufenamic acid mean transit t( flu/t( water time (t( flu) (s)
Amount of flufenamic acid retained by the liver (mmol g−1)
100 (n = 3) 250 (n = 4)
720.69 26.7 419.29 12.2
4.09 90.45 5.379 0.24
48.7 28.3
The mean transit times of flufenamic acid were calculated from indicator dilution experiments of the kind illustrated in Fig. 6. The mean transit time of tritiated water of 19 liver perfusion experiments, already corrected for catheter delay, was equal to 14.8 9 0.7 s. The mean transit times of flufenamic acid were also corrected for catheter delay. The amount of flufenamic acid retained by the liver was calculated according to Eq. (2).
and 50 mM the flufenamic acid curve was progressively shifted to the right with decreasing concentrations. At 25 mM, however, this trend was interrupted, i.e. the initial upslope of the 25-mM curve preceded in time all other curves. Normalized dilution curves resulting from step inputs tend asymptotically to unity if no significant metabolic transformation occurs, as in the case of the [3H]water curve. When metabolic transformation (extraction) occurs the curve tends asymptotically to steady-state values below unity. As revealed in Fig. 6, with 25 mM portal flufenamic acid, steady-state conditions were attained with quite a pronounced single pass extraction; the mean value of three perfusion experiments (n= 3) was 42.396.3%. Steady-state single pass extraction decreased considerably when the concentration was raised to 50 mM, namely 6.469 2.26% (n =3). At higher concentrations the curves tended asymptotically to values near unity and the single pass extraction could no longer be determined with confidence. This fact, on the other hand, allowed the determination of the mean transit times, which are listed in Table 3. The mean transit times are directly proportional to the real or apparent distribution spaces [21]. The difference between the mean transit times of 250 or 100 mM flufenamic acid and that of tritiated water was quite pronounced, 28.3- and 48.7-fold, respectively. The volume of the liver is only 20 or 30% superior to the aqueous space and the high mean transit times are evidently revealing that flufenamic acid was taken up in excess to the aqueous space of the liver. The amount that was retained by the liver can be calculated as the area between the [3H]water curve and the flufenamic acid curve, employing Eq. (2) in Section 2. The results of these calculations are shown in Table 3, expressed as mmol flufenamic acid per g liver. The results reveal that the concentrations of flufenamic acid in the liver after periods of 10 – 20 min of infusion became very high. For portal flufenamic acid equal to 100 mM, for example, the ratio of intra- to extracellular concentration is higher than 40.
118
C.H. Lopez et al. / Chemico-Biological Interactions 116 (1998) 105–122
4. Discussion The results of the present work indicate that the metabolic effects of flufenamic acid are the consequence of an uncoupling of oxidative phosphorylation [2,3]. The following observations support this conclusion: (a) flufenamic acid increased oxygen uptake, a common property of all uncouplers [9,10]; (b) the drug also increased glycolysis and glycogenolysis in livers from fed rats, an expected compensatory phenomenon for the decreased mitochondrial ATP formation; (c) flufenamic acid inhibited glucose production from fructose, an energy-dependent process; more significantly indeed, glucose production inhibition occurred in parallel with oxygen uptake stimulation, demonstrating that the latter phenomenon was actually deleterious to the liver cells; (d) the cellular ATP levels were decreased by flufenamic acid whereas the AMP levels were increased; this action was more pronounced in substrate-free perfused livers from fasted rats, which did not possess the possibility of compensatory increases in glycolysis; (e) the total adenine nucleotide content was decreased by flufenamic acid and uric acid production was stimulated; this is consistent with the observation that a decrease in the ratio of ATP/AMP accelerates uric acid production and purine nucleotides turnover [26,27]. It should be mentioned that the decreases in the ATP/ADP and ATP/AMP ratios may be an important cause for the changes in glycogenolysis and glycolysis [9]. As demonstrated by Nascimento et al. [9,10], calcium fluxes and redistributions seem not to play any significant role in glycogenolysis and glycolysis stimulation caused by uncouplers including non-steroidal anti-inflammatories such as niflumic acid and diclofenac. The same was observed in the present work with flufenamic acid. It should be remarked that when a Ca2 + -dependent agonist is infused according to the same experimental protocol employed in the present work with flufenamic acid, its action on glycogenolysis is enormously enhanced when Ca2 + is introduced in the system, as demonstrated by Nascimento et al. [9]. Oxygen uptake stimulation is one of the most characteristic effects of uncouplers. Frequently, however, uncouplers also produce inhibition at high concentrations [5]. This seems also to occur with flufenamic acid, but the effect is not very pronounced in the concentration range up to 250 mM. An indication that inhibition may be superimposing on stimulation is the finding that at high flufenamic acid concentrations oxygen uptake increased during the first half of the infusion until a maximum value and decreased thereafter, to a certain extent still remaining, however, significantly above the basal levels (Fig. 1). Another indication that inhibition may be superimposing on the uncoupling effect is the observation that in the presence of fructose as the gluconeogenic substrate flufenamic acid produced a smaller increase in oxygen uptake at 250 than at 100 mM (Fig. 4). Pyruvate production stimulation also presented a maximum during the first minute of flufenamic acid infusion with a subsequent decline (Fig. 1). This decline is more likely to reflect the progressive increase in the NADH/NAD + ratios in both the cytosol and mitochondria, which usually accompanies uncoupling of oxidative phosphorylation. Since the L-lactate dehydrogenase operates under near-equilibrium conditions, it rapidly equilibrates lactate and pyruvate and the lactate/pyruvate ratio actually reflects the cytosolic
C.H. Lopez et al. / Chemico-Biological Interactions 116 (1998) 105–122
119
NADH/NAD + ratio [28]. The fact that pyruvate production increased in parallel with lactate production during the first minute of flufenamic acid infusion means also that the changes in the NADH/NAD + ratio were relatively small during this time. Subsequently, however, the NADH/NAD + ratio increased leading to a substantial increase in the lactate/pyruvate ratio. A comparison of the potency of flufenamic acid as an inhibitor of energy metabolism with other non-steroidal anti-inflammatories reveals that it ranges among the most active ones. Most of its effects are exerted in the concentration range between 10 − 5 and 10 − 4 M; this potency is comparable to that of niflumic acid [6 – 8] and nimesulide [29] and slightly superior to that of diclofenac [9,30]. All these compounds are more potent than mefenamic acid [5] and piroxicam [31] which are more active in the concentration range between 10 − 4 and 10 − 3 M. A still less potent group of non-steroidal anti-inflammatories comprises naproxen and aspirin, which are not active or at least poorly active on energy metabolism at concentrations below 10 − 3 M [30,32]. Some general structure-effect relationships can be established. The carboxylic acids flufenamate, niflumate and mefenamate are very similar in structure. Mefenamate, which is less potent, possesses two methyl groups ( – CH3) in place of the trifluoromethyl group (–CF3) found in both flufenamate and niflumate. The trifluoromethyl group seems thus to be important in enhancing the uncoupling activity of these compounds. A carboxyl group linked to aromatic structures, on the other hand, seems not to be essential, because it is absent in nimesulide, which possesses a sulfoanilide group (active between 10 − 5 and 10 − 4 M) and piroxicam (active between 10 − 4 and 10 − 3 M), but present in naproxen and aspirin (active only at concentrations above 10 − 3 M). It seems that the activity of non-steroidal anti-inflammatories on energy metabolism depends on anionic groups linked to hydrophobic aromatic structures; special groups, as for example the trifluoromethyl group, can eventually enhance this activity, as it happens with both flufenamic and niflumic acid. The indicator-dilution experiments revealed that flufenamic acid undergoes flowlimited distribution in the liver, i.e. equilibration between the extra and intracellular spaces is almost complete along the hepatic acinus during a single passage. This is indicated by the observation that the whole dilution curve of flufenamic acid is situated on the right side of the [3H]water curve. As shown by Goresky et al. [25], the fraction of a substance that does not exchange at least once with the cellular space during a single passage through the liver must appear at an earlier time than [3H]water. Furthermore, the indicator-dilution experiments also revealed very high apparent distribution spaces for flufenamic acid. Actually this is an indication that the intracellular concentration of the compound under steady-state conditions greatly exceeds the extracellular concentration. This phenomenon, however, is unlikely to be the consequence of active transport for two reasons. In the first place, flufenamic acid undergoes flow-limited distribution. This means that the compound permeates the cell membrane at rates [25] that are certainly orders of magnitude superior to the rate of ATP production. In the second place, flufenamic acid itself decreases energy metabolism, so that the ATP supply is reduced in its presence. If active transport is improbable, the explanation that remains for the high intracellu-
120
C.H. Lopez et al. / Chemico-Biological Interactions 116 (1998) 105–122
lar concentration is binding to membrane components or proteins. In a previous study with niflumic acid we arrived at a similar conclusion: niflumic acid also inhibits energy metabolism, undergoes flow-limited distribution and has an apparent distribution space in excess to the aqueous space [20]. Intracellular binding of niflumic acid could indeed be quantified in terms of groups of binding sites with high and low affinities with a total binding capacity of 8.17 mmol (ml intracellular space) − 1 or approximately 5.5 mmol g − 1 liver [33]. This value is very similar to those found for 250 mM portal flufenamic acid, as revealed in Table 3, i.e. 5.37 mmol g − 1. The analysis of intracellular binding of niflumic acid in our previous work [20,33] was greatly facilitated by the fact that it was possible to analyze the outflow profiles of this compound in terms of the space-distributed variable transit time model of Goresky [25]. More specifically, the outflow profiles of niflumic acid were found to be in conformity with an equation describing a mechanism of flow-limited distribution in a single cellular compartment where metabolic transformation occurs. The latter could be quantified by a single first order rate parameter [20]. Flufenamic acid, however, shows a different behaviour and the equation used with niflumic acid could not be fitted to the outflow profiles measured in the present work. In relation to niflumic acid the behaviour of flufenamic acid presents two main differences. The first one is the change in behaviour when the portal concentration of flufenamic acid was increased from 25 to 50 mM. At 25 mM the initial upslope clearly preceded that of all other concentrations. From the trend of the curves obtained with 50, 100 and 250 mM, and also for theoretical reasons, one would expect an initial upslope situated at the right of the 50-mM curve. This was indeed what happened with niflumic acid [20]. It should be mentioned that the higher metabolic steady-state extraction observed with 25 mM flufenamic acid is an expected phenomenon and can be considered a normal consequence of the fact that metabolic transformation is a saturable and enzyme-catalyzed event or sum of events. The second difference between the niflumic acid and flufenamic acid is the appearance time. In the case of niflumic acid the appearance times were progressively delayed when the portal concentration was decreased. This is indeed a requirement of the model that was used to describe the niflumic acid outflow profiles [20]. In the case of flufenamic acid, however, the time of appearance is practically the same for all concentrations (Fig. 6). The behaviour of flufenamic acid in the perfused rat liver resembles that of octanoate [34]. This medium-chain fatty acid undergoes flow-limited distribution into the cell membrane and its surroundings. Its distribution into the rest of the cell space, however, occurs at relatively low rates. The relation between the free (not bound to albumin) extracellular concentration and the pool size into which octanoate undergoes flow-limited distribution is parabolic. This means that the apparent, or even real, space into which octanoate undergoes flow limited distribution can be increased by octanoate itself when its availability is also increased. This seems also to have happened with flufenamic acid because the change in the outflow profile which occurred when its portal concentration was increased to values above 25 mM is actually reflecting an increase in the apparent distribution space. In other
C.H. Lopez et al. / Chemico-Biological Interactions 116 (1998) 105–122
121
words, it is possible that the amount of flufenamic acid bound to the cell membranes during the early stages of the infusion produced changes that enabled these structures to bind a larger quantity of the drug than originally possible. These changes must have occurred during the early stages of the infusion of flufenamic acid. The transition may take several minutes and it could be the cause of the observation that the appearance time was practically the same with all infusion rates of flufenamic acid (Fig. 6). More detailed investigations on this subject could add useful data about the interactions of non-steroidal anti-inflammatories with the cellular membranes or other intracellular structures, which are believed to play some role in the biological activity of these compounds [30,35].
Acknowledgements This work was supported by grants from the Conselho Nacional de Desenvolvimento Cientı´fico e Tecnolo´gico (CNPq) and from the Programa Nacional de Apoio a Nu´cleos de Exceleˆncia (PRONEX). Carlos Henrique Lopez was a recipient of a fellowship from the Coordenac¸a˜o de Aperfeic¸oamento de Pessoal do Ensino Superior (CAPES).
References [1] E.L. Ishii, M. Alvarez, Efeito do a´cido niflu´mico sobre o metabolismo oxidativo de mitocoˆndrias de fı´gado de rato, Arq. Biol. Tecnol. 26 (1983) 234. [2] P. McDougall, A. Markham, I. Cameron, A.J. Sweetman, The mechanism of inhibition of mitochondrial oxidative phosphorylation by the non-steroidal anti-inflammatory agent diflunisal, Biochem. Pharmacol. 32 (1983) 2595 – 2598. [3] P. McDougall, A. Markham, I. Cameron, A.J. Sweetman, Action of the non-steroidal anti-inflammatory agent, flufenamic acid, on calcium movements in isolated mitochondria, Biochem. Pharmacol. 37 (1987) 1327–1330. [4] E.P. Brass, M.J. Garrity, Effect of non-steroidal anti-inflammatory drugs on glycogenolysis in isolated hepatocytes, Br. J. Pharmacol. 86 (1985) 491 – 496. [5] F.S. Kemmelmeier, A. Bracht, Effects of the non-steroidal anti-inflammatory mefenamic acid on energy metabolism in the perfused rat liver, Biochem. Pharmacol. 38 (1989) 823 – 830. [6] A.M. Kelmer-Bracht, M.M. Otani, A. Bracht, Ac¸a˜o do a´cido niflu´mico sobre alguns paraˆmetros metabo´licos do fı´gado perfundido isoladamente, Rev. Unimar 7 (1985a) 71 – 84. [7] A.M. Kelmer-Bracht, A. Bracht, Ac¸a˜o do a´cido niflu´mico sobre o metabolismo do glicerol no fı´gado de rato perfundido isoladamente, Rev. Unimar 7 (1985b) 123 – 132. [8] A.M. Kelmer-Bracht, A. Bracht, Inhibition of hepatic gluconeogenesis by niflumic acid correlates with the concentration of the free form, Res. Commun. Chem. Pathol. Pharmacol. 80 (1993) 241 – 248. [9] E.A. Nascimento, A.M. Kelmer-Bracht, A. Bracht, E.L. Ishii-Iwamoto, Activation of hepatic glycogenolysis by non-steroidal anti-inflammatories is independent of Ca2 + , Pharmacol. Commun. 3 (1993) 129–138. [10] E.A. Nascimento, F.A. Tonon, A.M. Kelmer-Bracht, A. Bracht, E.L. Ishii-Iwamoto, Glycogenolysis stimulation by non-steroidal anti-inflammatories in the perfused rat liver is not accompanied by Ca2 + release, Res. Commun. Mol. Pathol. Pharmacol. 96 (1997) 193 – 201. [11] R. Scholz, T. Bu¨cher, Hemoglobin-free perfusion of rat liver, in: B. Chance, R.W. Estabrook, J.R. Williamson (Eds.), Control of Energy Metabolism, Academic Press, New York, 1965, pp. 393 – 414.
122
C.H. Lopez et al. / Chemico-Biological Interactions 116 (1998) 105–122
[12] H.U. Bergmeyer, E. Bernt, Glucose determination with glucose oxidase and peroxidase, in: H.U. Bergmeyer (Ed.), Methods of Enzymatic Analysis, Academic Press, New York, 1974, pp. 1205 – 1215. [13] I. Gutman, A.W. Wahlefeld, ( + )Lactate determination with lactate dehydrogenase and NAD + , in: H.U. Bergmeyer (Ed.), Methods of Enzymatic Analysis, Academic Press, New York, 1974, pp. 1464–1468. [14] R. Czok, W. Lamprecht, Pyruvate, phosphoenolpyruvate and glycerate-2-phosphate, in: H.U. Bergmeyer (Ed.), Methods of Enzymatic Analysis, Academic Press, New York, 1974, pp. 1446 – 1451. [15] W. Lamprecht, I. Trautschold, Adenosine-5%-triphosphate. Determination with hexokinase and glucose-6-phosphate dehydrogenase, in: H.U. Bergmeyer (Ed.), Methods of Enzymatic Analysis, Verlag Chemie-Academic Press, Weinheim-London, 1974, pp. 2101 – 2110. [16] D. Jaworek, W. Gruber, H.U. Bergmeyer, Adenosine 5%-diphosphate and adenosine 5%-monophosphate, in: H.U. Bergmeyer (Ed.), Methods of Enzymatic Analysis, Verlag Chemie-Academic Press, Weinheim-London, 1974, pp. 2127– 2131. [17] P. Scheibe, E. Bernt, H.U. Bergmeyer, Uric acid, in: H.U. Bergmeyer (Ed.), Methods of Enzymatic Analysis, Verlag Chemie-Academic Press, Weinheim-London, 1974, pp. 1951 – 1958. [18] A.M. Kelmer-Bracht, E.L. Ishii, P.V.M. de Andrade, A. Bracht, Construction of a liver perfusion apparatus for studies on metabolic regulation and mechanisms of drug action, Arq. Bio. Tecnol. 27 (1984) 419–438. [19] P.H. Reinhart, W.M. Taylor, F.L. Bygrave, Calcium ion fluxes induced by the action of a-adrenergic agonists in perfused rat liver, Biochem. J. 208 (1982) 619 – 630. [20] A.M. Kelmer-Bracht, E.L. Ishii-Iwamoto, A. Bracht, Transport and distribution space of the anti-inflammatory drug niflumic acid in the perfused rat liver, Biochem. Pharmacol. 45 (1993) 1863–1871. [21] P. Meier, K.L. Zierler, On the theory of the indicator dilution method for measurement of blood flow and volume, J. Appl. Physiol. 6 (1954) 731 – 744. ˚ . Bjo¨rck, G. Dahlquist, Numerische Methoden, Oldenburg, Munich, 1972. [22] A [23] E.L. Ishii, A.J. Schwab, A. Bracht, Inhibition of monosaccharide transport in the intact rat liver by stevioside, Biochem. Pharmacol. 36 (1987) 1417 – 1433. [24] F. Suzuki-Kemmelmeier, E.L. Ishii-Iwamoto, A. Bracht, The metabolism of fructose in the bivascularly perfused rat liver, Biochim. Biophys. Acta 1116 (1992) 275 – 282. [25] C.A. Goresky, W.H. Ziegler, G.G. Bach, Capillary exchange modelling. Barrier-limited and flow-limited distribution, Circ. Res. 27 (1970) 739 – 764. [26] J. Faller, I.H. Fox, Ethanol-induced hyperuricemia. Evidence for increased urate production by activation of adenine nucleotide turnover, N. Engl. J. Med. 307 (1982) 1598 – 1602. [27] R.A. Zager, Adenine nucleotide changes in kidney, liver, and small intestine during different forms of ischemic injury, Circ. Res. 68 (1991) 185 – 196. [28] H. Sies, Nicotinamide nucleotide compartmentation, in: H. Sies (Ed.), Metabolic Compartmentation, Academic Press, New York, 1982, pp. 205 – 231. [29] S.M. Caparroz-Assef, C.A. Bersani-Amado, E.A. Nascimento, A.M. Kelmer-Bracht, E.L. IshiiIwamoto, Effects of the non-steroidal anti-inflammatory drug nimesulide on energy metabolism in livers from adjuvant-induced arthritic rats, Res. Commun. Mol. Pathol. Pharmacol. 99 (1998) 93 – 116. [30] I. Petrescu, C. Tarba, Uncoupling effects of diclofenac and aspirin in the perfused liver and isolated hepatic mitochondria of rat, Biochim. Biophys. Acta 1318 (1997) 385 – 394. [31] C.L. Salgueiro-Pagadigorria, J. Constantin, A. Bracht, E.A. Nascimento, E.L. Ishii-Iwamoto, Effects of the non-steroidal anti-inflammatory drug piroxicam on energy metabolism in the perfused rat liver, Comp. Biochem. Physiol. 113C (1996) 93 – 98. [32] E.A. Nascimento, N.S. Yamamoto, A. Bracht, E.L. Ishii-Iwamoto, Naproxen inhibits hepatic glycogenolysis induced by Ca2 + -dependent agents, Gen. Pharmacol. 253 (1995) 211 – 218. [33] A.M. Kelmer-Bracht, E.L. Ishii-Iwamoto, A. Bracht, Intracellular binding of the anti-inflammatory drug niflumic acid in the liver, Res. Commun. Mol. Pathol. Pharmacol. 89 (1995) 411 – 419. [34] O. Ferraresi-Filho, E.L. Ishii-Iwamoto, A. Bracht, Transport, metabolism and distribution space of octanoate in the perfused rat liver, Cell Biochem. Funct. 15 (1997) 69 – 80. [35] G. Weissmann, Aspirin, Sci. Am. 264 (1991) 58 – 64.
Metabolic effects and distribution space of flufenamic acid in the isolated perfused rat liver Carlos H. Lopez, Adelar Bracht *, Nair S. Yamamoto, Mariana D. dos Santos Laboratory of Li6er Metabolism, Department of Biochemistry, Uni6ersity of Maringa´, Maringa´, 87020900, Brazil Received 26 March 1998; received in revised form 1 June 1998; accepted 24 August 1998
Abstract The following aspects were investigated in the present work: (a) the action of flufenamic acid on hepatic metabolism (oxygen uptake, glycolysis, gluconeogenesis, uricogenesis and glycogenolysis), (b) the action of flufenamic acid on the cellular adenine nucleotide levels, and (c) the transport and distribution space of flufenamic acid in the liver parenchyma. The experimental system was the isolated perfused rat liver. Perfusion was accomplished in an open, non-recirculating system. The perfusion fluid was Krebs/Henseleit-bicarbonate buffer (pH 7.4), saturated with a mixture of oxygen and carbon dioxide (95:5) by means of a membrane oxygenator and heated to 37°C. The distribution space of flufenamic acid was measured by means of the multiple-indicator dilution technique with constant infusion (step input) of [3H]water plus flufenamic acid. The results of the present work indicate that the metabolic effects of flufenamic acid are the consequence of an uncoupling of oxidative phosphorylation, a conclusion based on the following observations: (a) flufenamic acid increased oxygen uptake, a common property of all uncouplers; (b) the drug also increased glycolysis and glycogenolysis in livers from fed rats (these are expected compensatory phenomena for the decreased mitochondrial ATP formation); (c) flufenamic acid inhibited glucose production from fructose, an energy-dependent process; (d) the cellular ATP levels were decreased by flufenamic acid whereas the AMP levels were increased; and (e) the total adenine nucleotide content was decreased by flufenamic acid and uric acid production was stimulated. Indicator-dilution experiments with flufenamic acid revealed that this substance undergoes flow-limited distribution in the liver and that its apparent distribution space * Corresponding author. Fax: + 55 44 2614396/2633655; e-mail: [email protected] 0009-2797/98/$ - see front matter © 1998 Published by Elsevier Science Ireland Ltd. All rights reserved. PII S0009-2797(98)00084-2
106
C.H. Lopez et al. / Chemico-Biological Interactions 116 (1998) 105–122
greatly exceeds the aqueous space of the liver. Flufenamic acid changed its behaviour when the portal concentration was increased from 25 to 50 mM. At 25 mM the initial upslope of the outflow profile clearly preceded that of all other concentrations. From the trend of the curves obtained with 50, 100 and 250 mM, one would expect an initial upslope situated at the right of the 50-mM curve. Furthermore, the time of appearance of flufenamic acid in the outflowing perfusate was practically the same irrespective of the portal concentration. For theoretical reasons one would expect progressively longer appearance times when the portal concentration was decreased. It is possible that the amount of flufenamic acid bound to the cell membranes during the early stages of the infusion produced changes that enabled these structures to bind a larger quantity of the drug than originally possible. © 1998 Published by Elsevier Science Ireland Ltd. All rights reserved. Keywords: Rat liver perfusion; Flufenamic acid; Metabolism; ATP; Distribution space
1. Introduction It has been shown that several non-steroidal anti-inflammatory agents like diflunisal, mefenamate, flufenamate and niflumate, act as uncouplers of oxidative phosphorylation in mitochondria isolated from rat liver [1–3]. Drugs that affect mitochondrial energy metabolism also produce several changes in cellular metabolism. Brass and Garrity [4] demonstrated that ibuprofen, indomethacin and meclofenamate are able to increase glycogenolysis in isolated hepatocytes by a mechanism that is independent of cyclooxygenase inhibition. Brass and Garrity [4] interpreted this action as a consequence of cytosolic calcium redistributions. Work from our laboratory, performed with the isolated perfused rat liver, has shown that mefenamate [5] and niflumate [6] are able to stimulate oxygen consumption, glycogenolysis, fructolysis and glycolysis. Gluconeogenesis, however, was inhibited [5,7,8]. These effects were attributed to the uncoupling action, which should cause a reduction in the ATP levels. A Ca2 + -mediated mechanism, on the other hand, was shown to be improbable by Nascimento et al. [9,10] in two studies performed with diclofenac and niflumic acid. It has been demonstrated in these studies that all metabolic effects of these drugs are exerted to the same extent either in the presence or in the absence of Ca2 + . As an extension of previous work performed in our laboratory, the present investigation was undertaken to investigate the metabolic effects of flufenamic acid, its distribution space and uptake rates in the perfused rat liver. All these aspects have not yet been investigated with flufenamic acid; the results should thus be useful for comparative purposes. The present study also offers the opportunity for investigating the influence of a non-steroidal anti-inflammatory agent on the cellular levels of adenine nucleotides, which should be decreased if these compounds really act as uncoupling agents. Up to now only indirect evidence is available concerning this phenomenon, which certainly deserves to be precisely quantified.
C.H. Lopez et al. / Chemico-Biological Interactions 116 (1998) 105–122
107
2. Materials and methods
2.1. Li6er perfusion Male albino rats (Wistar), weighing 180–220 g, were fed ad libitum with a standard laboratory diet (Purina®, Sa˜o Paulo, Brazil). For the surgical procedure the rats were anesthetized by intraperitoneal injection of sodium pentobarbital (50 mg/kg). Hemoglobin-free, non-recirculating perfusion was performed. The surgical technique was that described by Scholz and Bu¨cher [11]. After cannulation of the portal and cava veins the liver was positioned in a plexiglass chamber. The flow was maintained constant by a peristaltic pump. The perfusion fluid was Krebs/ Henseleit-bicarbonate buffer (pH 7.4), saturated with a mixture of oxygen and carbon dioxide (95:5) by means of a membrane oxygenator with simultaneous temperature adjustment at 37°C.
2.2. Analytical Samples of the effluent perfusion fluid were collected according to the experimental protocol and analyzed for their metabolite contents. The following compounds were assayed by means of standard enzymatic procedures: glucose [12], L-lactate [13] and pyruvate [14], ATP [15], ADP and AMP [16], and uric acid [17]. The lactate, pyruvate, AMP, ADP and ATP concentrations were calculated from spectrophotometric changes at 340 nm due to NAD + (or NADP + ) reduction or NADH oxidation. The extinction coefficient of NADH/NADPH (6.22 mM − 1 cm − 1) was used in the calculations. Glucose and uric acid were calculated from standard curves. A new standard curve was constructed for each assay. Linearity between concentration and absorbance was up to 1.0 absorbance units for glucose and up to 0.3 for uric acid. The oxygen concentration in the outflowing perfusate was monitored continuously, employing a Teflon-shielded platinum electrode adequately positioned in a Plexiglass chamber at the exit of the perfusate [18]. Metabolic rates were calculated from input-output differences and the total flow rates and were referred to the wet weight of the liver.
2.3. Ca 2 + -free perfusion For performing Ca2 + -free perfusion, the intracellular Ca2 + pools were exhausted. The following procedure was adopted. Livers were pre-perfused with Ca2 + -free Krebs/Henseleit-bicarbonate buffer containing 0.2 mM ethylenediamine tetraacetate (EDTA), 5 mM glucose, 1 mM lactate and 0.1 mM pyruvate. In order to ensure maximal depletion of the intracellular Ca2 + pools, phenylephrine (2 mM) was infused repeatedly (three times) during short periods of 2 min, with intervals of 5 min. According to Reinhart et al. [19], this procedure depletes the intracellular Ca2 + -pools which are normally mobilized when hormones are infused.
C.H. Lopez et al. / Chemico-Biological Interactions 116 (1998) 105–122
108
2.4. Measurement of the distribution space of flufenamic acid Indicator-dilution experiments were performed employing the constant infusion method (step input). [3H]Water and flufenamic acid were infused simultaneously into the liver prior to entering the portal vein. During the first 90 s after the onset of the infusion, the perfusate was fractionated in 0.5–2.0-s fractions by means of a specially designed fraction collector [18]. After the first 90 s, samples were taken manually in intervals ranging from 30 to 600 s. Aliquots of these samples were used for the assay of [3H]water and for the extraction and assay of flufenamic acid. Flufenamic acid was extracted from the perfusate samples with ethyl acetate and assayed spectrophotometrically at 285 nm (extinction coefficient= 17 mM − 1 cm − 1 in ethyl acetate) essentially as described previously for niflumic acid [20].
2.5. Treatment of data and calculations The statistical significance of the differences between parameters was evaluated by means of Student’s t-test or Student’s paired t-test, according to the context. Different steady-state values were compared and the analysis was performed by means of the Primer program (version 1.0, McGraw-Hill, 1988). PB 0.05 was adopted as a criterion of significance. Mean transit times of flufenamic acid (t( flu) and [3H]water (t( water) were calculated employing the formula of Meier and Zieler [21], i.e. t( =
&
[1 −Q(t)]dt
(1)
t0
where Q(t) is the normalized experimental outflow profile (effluent concentration/portal concentration), t is the time and t0 the transit time in the large vessels and catheters. The latter can be evaluated as the time of appearance of label in the outflowing perfusate. The amount of non-transformed flufenamic acid retained by the liver in excess to the aqueous space (R) was calculated as R=
&
FCp/P[Qwater(t) − Qflu(t)]dt
(2)
t0
where F is the perfusate flow through the liver (volume/time), Cp the portal flufenamic acid concentration, P the weight of the liver, Qwater(t) the normalized outflow profile of [3H]water and Qflu(t) the normalized outflow profile of flufenamic acid. Numerical integration was performed by means of the trapezoid rule [22].
2.6. Materials The liver perfusion apparatus and the rapid sampling apparatus were built in the workshops of the University of Maringa´. All enzymes and coenzymes used in the enzymatic assays were purchased from Sigma (St. Louis, USA). [3H]Water (NET001C, 25 mCi g − 1) was purchased from E.I. du Pont de Nemours (Boston, MA). All chemicals were from the best available grade.
C.H. Lopez et al. / Chemico-Biological Interactions 116 (1998) 105–122
109
3. Results
3.1. Effects of flufenamic acid on glycogen catabolism and oxygen uptake The effects of flufenamic acid on glycogen catabolism are illustrated in Figs. 1 and 2. Fig. 1 also illustrates the experimental protocol. Livers from fed rats were perfused with substrate-free Krebs/Henseleit-bicarbonate buffer (pH 7.4) saturated with a mixture of oxygen/carbon dioxide (95/5%). Glucose, lactate and pyruvate
Fig. 1. Time-course of the effects of flufenamic acid on glycogen catabolism and related parameters. Livers from fed rats were perfused with Krebs/Henseleit-bicarbonate buffer (pH 7.4) in an open system. Flufenamic acid (100 mM) was infused during the time intervals indicated by the horizontal bars. Samples of the effluent perfusate were taken at 2-min intervals for the measurement of glucose, lactate and pyruvate and for monitoring absorbance at 285 nm. Oxygen uptake was followed polarographically. The rates of glucose, lactate and pyruvate production and of oxygen consumption were expressed as mmol min − 1 (g liver wet weight) − 1 and represented against the perfusion time. Zero in the time scale is the instant at which the first sample of the effluent perfusate was collected. Values are the means of three liver perfusion experiments with identical protocol.
110
C.H. Lopez et al. / Chemico-Biological Interactions 116 (1998) 105–122
Fig. 2. Concentration dependence of the effects of flufenamic acid on glycogen catabolism and related parameters. Livers from fed rats were perfused with Krebs/Henseleit-bicarbonate buffer (pH 7.4) with different flufenamic acid concentrations in the range between 25 and 250 mM. The experimental protocol was the same illustrated in Fig. 1. The changes in oxygen uptake, lactate production and glucose production caused by flufenamic acid after 20-min infusion were evaluated and represented against the flufenamic acid concentration in the portal perfusate. Each data point represents the mean of three liver perfusion experiments. Bars are standard errors of the mean.
release were monitored as indicators of glycogen catabolism. Flufenamic acid was infused at a final portal concentration of 100 mM during 24 min. The introduction of flufenamic acid produced great changes in all parameters: glucose and lactate release were increased and remained elevated during the whole infusion time. Pyruvate production was only transiently increased with a peak value at 6 min after the onset of the infusion. Oxygen uptake was also increased with a maximum at 24 min and a small decline during the rest of the infusion period. The effects of flufenamic acid were reversible, i.e. cessation of the infusion restored the original levels of the variables. Flufenamic acid absorbs in the ultra-violet range with a peak around 285 nm. For this reason, absorbance at 285 nm was also monitored as a measure for the appearance of the drug in the venous perfusate. Flufenamic acid was infused at a constant rate, but its appearance in the outflowing perfusate was considerably delayed. The response of the liver to other concentrations of flufenamic acid can be inferred from Fig. 2 which shows the changes produced by several flufenamic acid concentrations at the end of the infusion period. Oxygen uptake, glucose release and lactate release stimulations are all presenting very clear and similar dose-effect relationships. Oxygen uptake stimulation varied between 0.22 mmol min − 1 g − 1 at 25 mM flufenamic acid and 0.87 mmol min − 1 g − 1 at 250 mM. Glucose and lactate release stimulations were relatively small at 25 mM flufenamic acid. At 250 mM, however, both parameters were stimulated by approximately 4 mmol min − 1 g − 1. In
C.H. Lopez et al. / Chemico-Biological Interactions 116 (1998) 105–122
111
the case of lactate release, this means an excess of ATP production in the glycolytic pathway of at least 6 mmol min − 1 g − 1. In order to see if Ca2 + influences the metabolic effects of flufenamic acid, experiments were performed in which the intracellular Ca2 + pools were exhausted by repeated phenylephrine pulses in the presence of 0.2 mM EDTA, as described in Section 2. After this procedure 100 mM flufenamic acid was infused during 14 min in the absence of Ca2 + . After this time the Ca2 + -free Krebs/Henseleit bicarbonate buffer was replaced by a normal perfusion fluid containing 2.5 mM CaCl2. Oxygen uptake, glucose release and lactate production were measured. The changes produced by flufenamic acid in the absence and presence of Ca2 + are shown in Table 1. It is clear from the results that Ca2 + has no influence on the metabolic effects of flufenamic acid.
3.2. Effects of flufenamic acid on fructose metabolism Fructose is rapidly transformed into glucose and lactate plus pyruvate in the liver. The first process is energy-dependent and the second one produces ATP. Fructose is, consequently, a suitable substrate for investigating the possible effect of flufenamic acid on energy metabolism. Fig. 3 shows the time-course of the effects of 100 mM flufenamic acid on fructose metabolism and Fig. 4 summarizes the effects of several concentrations of the drug. The experiments with fructose were performed with livers from fasted animals in order to avoid interference by glycogen catabolism. The introduction of fructose produced the known changes in oxygen uptake, and glucose, lactate and pyruvate production [23,24]. The subsequent introduction of flufenamic acid further increased oxygen uptake, but glucose production was considerably decreased. Pyruvate production was also decreased, but the effect on lactate production was more complex. Initially lactate production suffered a rapid increase which was followed by a
Table 1 Influence of Ca2+ on the metabolic effects of flufenamic acid Perfusion condition
Increment in glucose release (mmol min−1 g−1)
Increment in lactate production (mmol min−1 g−1)
Increment in oxygen consumption (mmol min−1 g−1)
2.5 mM Ca2+ Ca2+-free
2.2190.44 (n=4) 2.229 0.36 (n= 4)
2.5390.31 (n =4) 2.469 0.26 (n = 4)
0.6290.07 (n = 4) 0.649 0.05 (n = 4)
The intracellular Ca2+ pools were exhausted by repeated phenylephrine pulses in the presence of 0.2 mM EDTA, as described in Section 2. After this procedure 100 mM flufenamic acid was infused during 14 min in the absence of Ca2. After this time the Ca2+-free Krebs/Henseleit bicarbonate buffer was replaced by a normal perfusion fluid containing 2.5 mM CaCl2. Oxygen uptake, glucose release and lactate production were measured. The basal values (before flufenamic acid infusion) were subtracted from the corresponding values found in the presence of flufenamic acid.
112
C.H. Lopez et al. / Chemico-Biological Interactions 116 (1998) 105–122
Fig. 3. Time-course of the effects of flufenamic acid on fructose metabolism. Livers from 24-h fasted rats were perfused with Krebs/Henseleit-bicarbonate buffer (pH 7.4) in an open system. Fructose (5.0 mM) and flufenamic acid (100 mM) were infused during the times indicated by the horizontal bars. Samples of the effluent perfusate were taken at 2-min intervals for the measurement of glucose, lactate and pyruvate and for monitoring absorbance at 285 nm. Oxygen uptake was followed polarographically. The rates of glucose, lactate and pyruvate production and of oxygen consumption were expressed as mmol min − 1 (g liver wet weight) − 1 and represented against the perfusion time. Zero in the time scale is the instant at which the first sample of the effluent perfusate was collected. Values are the means of three liver perfusion experiments with identical protocol.
progressive decrease to lower values. The latter were still above the basal levels (i.e. in the absence of flufenamic acid). All these changes were perfectly reversible. The concentration dependence in Fig. 4 reveals that glucose production was already substantially inhibited by 25 mM flufenamic acid (− 0.6 mmol min − 1 g − 1). No oxygen uptake stimulation was found at this concentration. Further increases in concentration, however, were stimulative. Maximal stimulation, nearly 0.5 mmol min − 1 g − 1, was below the maximal stimulation found with substrate free-perfused livers from fed rats (0.87 mmol min − 1 g − 1; Fig. 2). At 250 mM flufenamic acid, glucose production was decreased by approximately 2.0 mmol min − 1 g − 1; this decrease means more than 90% inhibition.
C.H. Lopez et al. / Chemico-Biological Interactions 116 (1998) 105–122
113
3.3. The action of flufenamic acid on adenine nucleotide le6els and uric acid metabolism The metabolic effects of flufenamic acid are consistent with an action on the mitochondrial energy metabolism. If this conclusion is correct, flufenamic acid should have a profound influence on the adenine nucleotide levels. For this reason, the cellular levels of AMP, ADP and ATP were measured in livers from fed and fasted rats in the absence and after 20 min of 100 mM flufenamic acid infusion. The results are shown in Table 2. For comparative purposes, the effects of cyanide, a compound that blocks the respiratory chain, are also shown. Table 2 reveals that flufenamic acid diminished the levels of ATP and increased those of AMP in livers from both ad libitum fed and 24-h fasted rats. The levels of ADP were not changed, but the ATP to ADP and ATP to AMP ratios were substantially decreased. The decrease in the ATP levels was less pronounced in livers from fed rats, a phenomenon that can be attributed to the increased glycolysis. In livers from fasted rats glycolysis was minimal due to the fact that substrate-free perfusion fluid was used. The sum of the contents of AMP, ADP and ATP was diminished by flufenamic acid in both livers from fed and fasted rats. Here again, the change was more pronounced in livers from fasted rats. Concerning the adenine nucleotide levels, the effects of flufenamic acid were similar to those of 1.0 mM cyanide, but with some significant differences: (a) contrary to flufenamic acid, cyanide did not
Fig. 4. Concentration dependence of the effects of flufenamic acid on glucose production from fructose and on oxygen uptake. Livers from fasted rats were perfused with Krebs/Henseleit-bicarbonate buffer (pH 7.4) containing 5 mM fructose and with different flufenamic acid concentrations in the range between 25 and 250 mM. The experimental protocol was the same illustrated in Fig. 3. The changes in glucose production and oxygen uptake caused by flufenamic acid after 20-min infusion were evaluated and represented against the flufenamic acid concentration in the portal perfusate. Each data point represents the mean of three liver perfusion experiments. Bars are standard errors of the mean.
Fasted Fasted Fasted Fed Fed Fed
1 (n= 3) 2 (n=5) 3 (n =5) 4 (n=3) 5 (n=6) 6 (n= 3)
— 100 —
— 100 —
Flufenamic acid (mM)
— — 1.0
— — 1.0
Cyanide (mM)
ADP (mmol g−1)
0.85 9 0.06 0.93 90.03 0.95 90.14 0.73 90.02 0.829 0.03 0.96 90.02*
ATP (mmol g−1)
2.14 90.05 0.56 90.10* 0.06 90.01* 2.13 9 0.04 1.49 90.08* 1.88 90.07*
0.17 90.02 0.339 0.04* 0.26 9 0.06
0.22 9 0.02 0.459 0.05* 0.98 9 0.15*
AMP (mmol g−1)
3.049 0.02 2.6390.09* 3.099 0.04
3.229 0.12 1.9490.07* 1.999 0.16*
Total adenine nucleotides (mmol g−1)
2.92 9 0.12 1.82 90.13* 1.96 90.10*
2.52 90.15 0.60 9 0.1* 0.06 90.01*
ATP/ADP
9 0.007* 12.3891.38 4.529 0.77* 7.239 2.35
9.739 0.71 1.2490.35* 0.061
ATP/AMP
Livers were perfused in an open system as described in Section 2. Flufenamic acid or cyanide was infused during 20 min as illustrated in Fig. 5. For the determination of the adenine nucleotide levels the livers were freeze-clamped in liquid nitrogen and extracted with cold perchloric acid. Determination was accomplished by standard enzymatic procedures. * Significantly different from the corresponding controls according to Student’s t-test (PB0.05).
Metabolic condition of the rat
Experiment no.
Table 2 Influence of flufenamic acid and cyanide on ATP, ADP and AMP levels in livers from fed and 24-h fasted rats
114 C.H. Lopez et al. / Chemico-Biological Interactions 116 (1998) 105–122
C.H. Lopez et al. / Chemico-Biological Interactions 116 (1998) 105–122
115
Fig. 5. Time-course of the effects of flufenamic acid on uric acid production. Livers from 24-h fasted rats were perfused with Krebs/Henseleit-bicarbonate buffer (pH 7.4) in an open system. Flufenamic acid (100 mM) was infused as indicated. Samples of the effluent perfusate were taken at 2-min intervals for the measurement of uric acid by means of an enzymatic procedure. The rate of uric acid production was expressed as nmol min − 1 (g liver wet weight) − 1 and represented against the perfusion time. Zero in the time scale is the instant at which the first sample of the effluent perfusate was collected. Values are the means of three liver perfusion experiments with identical protocol.
affect the sum of the contents of AMP, ADP and ATP in livers from fed rats; (b) flufenamic acid had a more pronounced effect than cyanide on the ATP levels in livers from fed rats; the opposite occurred, however, in livers from 24-h fasted rats. The diminution of the total adenine nucleotide levels could be the consequence of an enhancement of purine nucleotide metabolism, which usually occurs when the degree of phosphorylation is decreased. Dephosphorylation of the adenine nucleotides usually generates multiple forms (inosine, hypoxanthine, etc.) which, contrary to the phosphorylated forms, can permeate the cell membrane and, consequently, escape from the hepatic cellular space. If this occurs when flufenamic acid is given to the liver, uric acid release should be increased when the drug is introduced, because this compound is one of the end products of purine nucleotides metabolism. In order to test this hypothesis, uric acid production was measured in livers from fasted rats, before and after the introduction of flufenamic acid. The results are shown in Fig. 5 and they reveal that flufenamic acid caused a great increase (in relative terms) in uric acid release, which passed from practically zero to a maximum of 11 nmol min − 1 g − 1 at the sixth minute of infusion. At the end of the infusion uric acid production stabilized at the much lower value of 2 nmol
116
C.H. Lopez et al. / Chemico-Biological Interactions 116 (1998) 105–122
min − 1 g − 1. The total extra amount of uric acid released during the infusion of flufenamic acid was equal to 90 nmol g − 1, as evaluated from the area under the time – response curve in Fig. 5.
3.4. Distribution space of flufenamic acid in the li6er The absorbance measurements at 285 nm during the infusion of flufenamic acid (Figs. 1 and 3) suggest that this compound undergoes flow-limited distribution in the liver [25]. In order to obtain more precise information about this topic, we have performed indicator-dilution experiments of the kind illustrated in Fig. 6. In this series of experiments flufenamic acid was simultaneously infused with [3H]water as an indicator for the total aqueous space of the liver. Flufenamic acid was extracted with ethyl acetate. The curves in Fig. 6 were normalized, i.e. the concentration in the effluent perfusate at the different times was divided by the concentration in the portal vein. Fig. 6 reveals that the time elapsed between the onset of the infusion and the time of appearance of the substance in the outflowing perfusate (appearance time) was not the same for both substances. No significant amount of flufenamic acid appeared in the effluent perfusion fluid before [3H]water. In the range between 250
Fig. 6. Indicator-dilution experiments with various portal concentrations of flufenamic acid. Livers from fed rats were perfused with Krebs/Henseleit-bicarbonate buffer (pH 7.4). Flufenamic acid, at various concentrations, and [3H]water (2.5 mCi min − 1) were infused simultaneously. The venous perfusate was fractionated in 0.5- to 600-s fractions. [3H]Water was determined by liquid scintillation spectrometry. Flufenamic acid was extracted with ethyl acetate and determined spectrophotometrically at 285 nm. The venous/portal concentration ratios of each substance were plotted versus the time after onset of the infusion. The portal flufenamic acid concentrations were: 250 mM ( – ); 100 mM (– ); 50 mM ( –); and 25 mM ("–"). For clarity, only the [3H]water curve obtained with 250 mM flufenamic acid is shown ( – ).
C.H. Lopez et al. / Chemico-Biological Interactions 116 (1998) 105–122
117
Table 3 Mean transit times of flufenamic acid in the isolated perfused rat liver Portal flufenamic acid concentration (mM)
Flufenamic acid mean transit t( flu/t( water time (t( flu) (s)
Amount of flufenamic acid retained by the liver (mmol g−1)
100 (n = 3) 250 (n = 4)
720.69 26.7 419.29 12.2
4.09 90.45 5.379 0.24
48.7 28.3
The mean transit times of flufenamic acid were calculated from indicator dilution experiments of the kind illustrated in Fig. 6. The mean transit time of tritiated water of 19 liver perfusion experiments, already corrected for catheter delay, was equal to 14.8 9 0.7 s. The mean transit times of flufenamic acid were also corrected for catheter delay. The amount of flufenamic acid retained by the liver was calculated according to Eq. (2).
and 50 mM the flufenamic acid curve was progressively shifted to the right with decreasing concentrations. At 25 mM, however, this trend was interrupted, i.e. the initial upslope of the 25-mM curve preceded in time all other curves. Normalized dilution curves resulting from step inputs tend asymptotically to unity if no significant metabolic transformation occurs, as in the case of the [3H]water curve. When metabolic transformation (extraction) occurs the curve tends asymptotically to steady-state values below unity. As revealed in Fig. 6, with 25 mM portal flufenamic acid, steady-state conditions were attained with quite a pronounced single pass extraction; the mean value of three perfusion experiments (n= 3) was 42.396.3%. Steady-state single pass extraction decreased considerably when the concentration was raised to 50 mM, namely 6.469 2.26% (n =3). At higher concentrations the curves tended asymptotically to values near unity and the single pass extraction could no longer be determined with confidence. This fact, on the other hand, allowed the determination of the mean transit times, which are listed in Table 3. The mean transit times are directly proportional to the real or apparent distribution spaces [21]. The difference between the mean transit times of 250 or 100 mM flufenamic acid and that of tritiated water was quite pronounced, 28.3- and 48.7-fold, respectively. The volume of the liver is only 20 or 30% superior to the aqueous space and the high mean transit times are evidently revealing that flufenamic acid was taken up in excess to the aqueous space of the liver. The amount that was retained by the liver can be calculated as the area between the [3H]water curve and the flufenamic acid curve, employing Eq. (2) in Section 2. The results of these calculations are shown in Table 3, expressed as mmol flufenamic acid per g liver. The results reveal that the concentrations of flufenamic acid in the liver after periods of 10 – 20 min of infusion became very high. For portal flufenamic acid equal to 100 mM, for example, the ratio of intra- to extracellular concentration is higher than 40.
118
C.H. Lopez et al. / Chemico-Biological Interactions 116 (1998) 105–122
4. Discussion The results of the present work indicate that the metabolic effects of flufenamic acid are the consequence of an uncoupling of oxidative phosphorylation [2,3]. The following observations support this conclusion: (a) flufenamic acid increased oxygen uptake, a common property of all uncouplers [9,10]; (b) the drug also increased glycolysis and glycogenolysis in livers from fed rats, an expected compensatory phenomenon for the decreased mitochondrial ATP formation; (c) flufenamic acid inhibited glucose production from fructose, an energy-dependent process; more significantly indeed, glucose production inhibition occurred in parallel with oxygen uptake stimulation, demonstrating that the latter phenomenon was actually deleterious to the liver cells; (d) the cellular ATP levels were decreased by flufenamic acid whereas the AMP levels were increased; this action was more pronounced in substrate-free perfused livers from fasted rats, which did not possess the possibility of compensatory increases in glycolysis; (e) the total adenine nucleotide content was decreased by flufenamic acid and uric acid production was stimulated; this is consistent with the observation that a decrease in the ratio of ATP/AMP accelerates uric acid production and purine nucleotides turnover [26,27]. It should be mentioned that the decreases in the ATP/ADP and ATP/AMP ratios may be an important cause for the changes in glycogenolysis and glycolysis [9]. As demonstrated by Nascimento et al. [9,10], calcium fluxes and redistributions seem not to play any significant role in glycogenolysis and glycolysis stimulation caused by uncouplers including non-steroidal anti-inflammatories such as niflumic acid and diclofenac. The same was observed in the present work with flufenamic acid. It should be remarked that when a Ca2 + -dependent agonist is infused according to the same experimental protocol employed in the present work with flufenamic acid, its action on glycogenolysis is enormously enhanced when Ca2 + is introduced in the system, as demonstrated by Nascimento et al. [9]. Oxygen uptake stimulation is one of the most characteristic effects of uncouplers. Frequently, however, uncouplers also produce inhibition at high concentrations [5]. This seems also to occur with flufenamic acid, but the effect is not very pronounced in the concentration range up to 250 mM. An indication that inhibition may be superimposing on stimulation is the finding that at high flufenamic acid concentrations oxygen uptake increased during the first half of the infusion until a maximum value and decreased thereafter, to a certain extent still remaining, however, significantly above the basal levels (Fig. 1). Another indication that inhibition may be superimposing on the uncoupling effect is the observation that in the presence of fructose as the gluconeogenic substrate flufenamic acid produced a smaller increase in oxygen uptake at 250 than at 100 mM (Fig. 4). Pyruvate production stimulation also presented a maximum during the first minute of flufenamic acid infusion with a subsequent decline (Fig. 1). This decline is more likely to reflect the progressive increase in the NADH/NAD + ratios in both the cytosol and mitochondria, which usually accompanies uncoupling of oxidative phosphorylation. Since the L-lactate dehydrogenase operates under near-equilibrium conditions, it rapidly equilibrates lactate and pyruvate and the lactate/pyruvate ratio actually reflects the cytosolic
C.H. Lopez et al. / Chemico-Biological Interactions 116 (1998) 105–122
119
NADH/NAD + ratio [28]. The fact that pyruvate production increased in parallel with lactate production during the first minute of flufenamic acid infusion means also that the changes in the NADH/NAD + ratio were relatively small during this time. Subsequently, however, the NADH/NAD + ratio increased leading to a substantial increase in the lactate/pyruvate ratio. A comparison of the potency of flufenamic acid as an inhibitor of energy metabolism with other non-steroidal anti-inflammatories reveals that it ranges among the most active ones. Most of its effects are exerted in the concentration range between 10 − 5 and 10 − 4 M; this potency is comparable to that of niflumic acid [6 – 8] and nimesulide [29] and slightly superior to that of diclofenac [9,30]. All these compounds are more potent than mefenamic acid [5] and piroxicam [31] which are more active in the concentration range between 10 − 4 and 10 − 3 M. A still less potent group of non-steroidal anti-inflammatories comprises naproxen and aspirin, which are not active or at least poorly active on energy metabolism at concentrations below 10 − 3 M [30,32]. Some general structure-effect relationships can be established. The carboxylic acids flufenamate, niflumate and mefenamate are very similar in structure. Mefenamate, which is less potent, possesses two methyl groups ( – CH3) in place of the trifluoromethyl group (–CF3) found in both flufenamate and niflumate. The trifluoromethyl group seems thus to be important in enhancing the uncoupling activity of these compounds. A carboxyl group linked to aromatic structures, on the other hand, seems not to be essential, because it is absent in nimesulide, which possesses a sulfoanilide group (active between 10 − 5 and 10 − 4 M) and piroxicam (active between 10 − 4 and 10 − 3 M), but present in naproxen and aspirin (active only at concentrations above 10 − 3 M). It seems that the activity of non-steroidal anti-inflammatories on energy metabolism depends on anionic groups linked to hydrophobic aromatic structures; special groups, as for example the trifluoromethyl group, can eventually enhance this activity, as it happens with both flufenamic and niflumic acid. The indicator-dilution experiments revealed that flufenamic acid undergoes flowlimited distribution in the liver, i.e. equilibration between the extra and intracellular spaces is almost complete along the hepatic acinus during a single passage. This is indicated by the observation that the whole dilution curve of flufenamic acid is situated on the right side of the [3H]water curve. As shown by Goresky et al. [25], the fraction of a substance that does not exchange at least once with the cellular space during a single passage through the liver must appear at an earlier time than [3H]water. Furthermore, the indicator-dilution experiments also revealed very high apparent distribution spaces for flufenamic acid. Actually this is an indication that the intracellular concentration of the compound under steady-state conditions greatly exceeds the extracellular concentration. This phenomenon, however, is unlikely to be the consequence of active transport for two reasons. In the first place, flufenamic acid undergoes flow-limited distribution. This means that the compound permeates the cell membrane at rates [25] that are certainly orders of magnitude superior to the rate of ATP production. In the second place, flufenamic acid itself decreases energy metabolism, so that the ATP supply is reduced in its presence. If active transport is improbable, the explanation that remains for the high intracellu-
120
C.H. Lopez et al. / Chemico-Biological Interactions 116 (1998) 105–122
lar concentration is binding to membrane components or proteins. In a previous study with niflumic acid we arrived at a similar conclusion: niflumic acid also inhibits energy metabolism, undergoes flow-limited distribution and has an apparent distribution space in excess to the aqueous space [20]. Intracellular binding of niflumic acid could indeed be quantified in terms of groups of binding sites with high and low affinities with a total binding capacity of 8.17 mmol (ml intracellular space) − 1 or approximately 5.5 mmol g − 1 liver [33]. This value is very similar to those found for 250 mM portal flufenamic acid, as revealed in Table 3, i.e. 5.37 mmol g − 1. The analysis of intracellular binding of niflumic acid in our previous work [20,33] was greatly facilitated by the fact that it was possible to analyze the outflow profiles of this compound in terms of the space-distributed variable transit time model of Goresky [25]. More specifically, the outflow profiles of niflumic acid were found to be in conformity with an equation describing a mechanism of flow-limited distribution in a single cellular compartment where metabolic transformation occurs. The latter could be quantified by a single first order rate parameter [20]. Flufenamic acid, however, shows a different behaviour and the equation used with niflumic acid could not be fitted to the outflow profiles measured in the present work. In relation to niflumic acid the behaviour of flufenamic acid presents two main differences. The first one is the change in behaviour when the portal concentration of flufenamic acid was increased from 25 to 50 mM. At 25 mM the initial upslope clearly preceded that of all other concentrations. From the trend of the curves obtained with 50, 100 and 250 mM, and also for theoretical reasons, one would expect an initial upslope situated at the right of the 50-mM curve. This was indeed what happened with niflumic acid [20]. It should be mentioned that the higher metabolic steady-state extraction observed with 25 mM flufenamic acid is an expected phenomenon and can be considered a normal consequence of the fact that metabolic transformation is a saturable and enzyme-catalyzed event or sum of events. The second difference between the niflumic acid and flufenamic acid is the appearance time. In the case of niflumic acid the appearance times were progressively delayed when the portal concentration was decreased. This is indeed a requirement of the model that was used to describe the niflumic acid outflow profiles [20]. In the case of flufenamic acid, however, the time of appearance is practically the same for all concentrations (Fig. 6). The behaviour of flufenamic acid in the perfused rat liver resembles that of octanoate [34]. This medium-chain fatty acid undergoes flow-limited distribution into the cell membrane and its surroundings. Its distribution into the rest of the cell space, however, occurs at relatively low rates. The relation between the free (not bound to albumin) extracellular concentration and the pool size into which octanoate undergoes flow-limited distribution is parabolic. This means that the apparent, or even real, space into which octanoate undergoes flow limited distribution can be increased by octanoate itself when its availability is also increased. This seems also to have happened with flufenamic acid because the change in the outflow profile which occurred when its portal concentration was increased to values above 25 mM is actually reflecting an increase in the apparent distribution space. In other
C.H. Lopez et al. / Chemico-Biological Interactions 116 (1998) 105–122
121
words, it is possible that the amount of flufenamic acid bound to the cell membranes during the early stages of the infusion produced changes that enabled these structures to bind a larger quantity of the drug than originally possible. These changes must have occurred during the early stages of the infusion of flufenamic acid. The transition may take several minutes and it could be the cause of the observation that the appearance time was practically the same with all infusion rates of flufenamic acid (Fig. 6). More detailed investigations on this subject could add useful data about the interactions of non-steroidal anti-inflammatories with the cellular membranes or other intracellular structures, which are believed to play some role in the biological activity of these compounds [30,35].
Acknowledgements This work was supported by grants from the Conselho Nacional de Desenvolvimento Cientı´fico e Tecnolo´gico (CNPq) and from the Programa Nacional de Apoio a Nu´cleos de Exceleˆncia (PRONEX). Carlos Henrique Lopez was a recipient of a fellowship from the Coordenac¸a˜o de Aperfeic¸oamento de Pessoal do Ensino Superior (CAPES).
References [1] E.L. Ishii, M. Alvarez, Efeito do a´cido niflu´mico sobre o metabolismo oxidativo de mitocoˆndrias de fı´gado de rato, Arq. Biol. Tecnol. 26 (1983) 234. [2] P. McDougall, A. Markham, I. Cameron, A.J. Sweetman, The mechanism of inhibition of mitochondrial oxidative phosphorylation by the non-steroidal anti-inflammatory agent diflunisal, Biochem. Pharmacol. 32 (1983) 2595 – 2598. [3] P. McDougall, A. Markham, I. Cameron, A.J. Sweetman, Action of the non-steroidal anti-inflammatory agent, flufenamic acid, on calcium movements in isolated mitochondria, Biochem. Pharmacol. 37 (1987) 1327–1330. [4] E.P. Brass, M.J. Garrity, Effect of non-steroidal anti-inflammatory drugs on glycogenolysis in isolated hepatocytes, Br. J. Pharmacol. 86 (1985) 491 – 496. [5] F.S. Kemmelmeier, A. Bracht, Effects of the non-steroidal anti-inflammatory mefenamic acid on energy metabolism in the perfused rat liver, Biochem. Pharmacol. 38 (1989) 823 – 830. [6] A.M. Kelmer-Bracht, M.M. Otani, A. Bracht, Ac¸a˜o do a´cido niflu´mico sobre alguns paraˆmetros metabo´licos do fı´gado perfundido isoladamente, Rev. Unimar 7 (1985a) 71 – 84. [7] A.M. Kelmer-Bracht, A. Bracht, Ac¸a˜o do a´cido niflu´mico sobre o metabolismo do glicerol no fı´gado de rato perfundido isoladamente, Rev. Unimar 7 (1985b) 123 – 132. [8] A.M. Kelmer-Bracht, A. Bracht, Inhibition of hepatic gluconeogenesis by niflumic acid correlates with the concentration of the free form, Res. Commun. Chem. Pathol. Pharmacol. 80 (1993) 241 – 248. [9] E.A. Nascimento, A.M. Kelmer-Bracht, A. Bracht, E.L. Ishii-Iwamoto, Activation of hepatic glycogenolysis by non-steroidal anti-inflammatories is independent of Ca2 + , Pharmacol. Commun. 3 (1993) 129–138. [10] E.A. Nascimento, F.A. Tonon, A.M. Kelmer-Bracht, A. Bracht, E.L. Ishii-Iwamoto, Glycogenolysis stimulation by non-steroidal anti-inflammatories in the perfused rat liver is not accompanied by Ca2 + release, Res. Commun. Mol. Pathol. Pharmacol. 96 (1997) 193 – 201. [11] R. Scholz, T. Bu¨cher, Hemoglobin-free perfusion of rat liver, in: B. Chance, R.W. Estabrook, J.R. Williamson (Eds.), Control of Energy Metabolism, Academic Press, New York, 1965, pp. 393 – 414.
122
C.H. Lopez et al. / Chemico-Biological Interactions 116 (1998) 105–122
[12] H.U. Bergmeyer, E. Bernt, Glucose determination with glucose oxidase and peroxidase, in: H.U. Bergmeyer (Ed.), Methods of Enzymatic Analysis, Academic Press, New York, 1974, pp. 1205 – 1215. [13] I. Gutman, A.W. Wahlefeld, ( + )Lactate determination with lactate dehydrogenase and NAD + , in: H.U. Bergmeyer (Ed.), Methods of Enzymatic Analysis, Academic Press, New York, 1974, pp. 1464–1468. [14] R. Czok, W. Lamprecht, Pyruvate, phosphoenolpyruvate and glycerate-2-phosphate, in: H.U. Bergmeyer (Ed.), Methods of Enzymatic Analysis, Academic Press, New York, 1974, pp. 1446 – 1451. [15] W. Lamprecht, I. Trautschold, Adenosine-5%-triphosphate. Determination with hexokinase and glucose-6-phosphate dehydrogenase, in: H.U. Bergmeyer (Ed.), Methods of Enzymatic Analysis, Verlag Chemie-Academic Press, Weinheim-London, 1974, pp. 2101 – 2110. [16] D. Jaworek, W. Gruber, H.U. Bergmeyer, Adenosine 5%-diphosphate and adenosine 5%-monophosphate, in: H.U. Bergmeyer (Ed.), Methods of Enzymatic Analysis, Verlag Chemie-Academic Press, Weinheim-London, 1974, pp. 2127– 2131. [17] P. Scheibe, E. Bernt, H.U. Bergmeyer, Uric acid, in: H.U. Bergmeyer (Ed.), Methods of Enzymatic Analysis, Verlag Chemie-Academic Press, Weinheim-London, 1974, pp. 1951 – 1958. [18] A.M. Kelmer-Bracht, E.L. Ishii, P.V.M. de Andrade, A. Bracht, Construction of a liver perfusion apparatus for studies on metabolic regulation and mechanisms of drug action, Arq. Bio. Tecnol. 27 (1984) 419–438. [19] P.H. Reinhart, W.M. Taylor, F.L. Bygrave, Calcium ion fluxes induced by the action of a-adrenergic agonists in perfused rat liver, Biochem. J. 208 (1982) 619 – 630. [20] A.M. Kelmer-Bracht, E.L. Ishii-Iwamoto, A. Bracht, Transport and distribution space of the anti-inflammatory drug niflumic acid in the perfused rat liver, Biochem. Pharmacol. 45 (1993) 1863–1871. [21] P. Meier, K.L. Zierler, On the theory of the indicator dilution method for measurement of blood flow and volume, J. Appl. Physiol. 6 (1954) 731 – 744. ˚ . Bjo¨rck, G. Dahlquist, Numerische Methoden, Oldenburg, Munich, 1972. [22] A [23] E.L. Ishii, A.J. Schwab, A. Bracht, Inhibition of monosaccharide transport in the intact rat liver by stevioside, Biochem. Pharmacol. 36 (1987) 1417 – 1433. [24] F. Suzuki-Kemmelmeier, E.L. Ishii-Iwamoto, A. Bracht, The metabolism of fructose in the bivascularly perfused rat liver, Biochim. Biophys. Acta 1116 (1992) 275 – 282. [25] C.A. Goresky, W.H. Ziegler, G.G. Bach, Capillary exchange modelling. Barrier-limited and flow-limited distribution, Circ. Res. 27 (1970) 739 – 764. [26] J. Faller, I.H. Fox, Ethanol-induced hyperuricemia. Evidence for increased urate production by activation of adenine nucleotide turnover, N. Engl. J. Med. 307 (1982) 1598 – 1602. [27] R.A. Zager, Adenine nucleotide changes in kidney, liver, and small intestine during different forms of ischemic injury, Circ. Res. 68 (1991) 185 – 196. [28] H. Sies, Nicotinamide nucleotide compartmentation, in: H. Sies (Ed.), Metabolic Compartmentation, Academic Press, New York, 1982, pp. 205 – 231. [29] S.M. Caparroz-Assef, C.A. Bersani-Amado, E.A. Nascimento, A.M. Kelmer-Bracht, E.L. IshiiIwamoto, Effects of the non-steroidal anti-inflammatory drug nimesulide on energy metabolism in livers from adjuvant-induced arthritic rats, Res. Commun. Mol. Pathol. Pharmacol. 99 (1998) 93 – 116. [30] I. Petrescu, C. Tarba, Uncoupling effects of diclofenac and aspirin in the perfused liver and isolated hepatic mitochondria of rat, Biochim. Biophys. Acta 1318 (1997) 385 – 394. [31] C.L. Salgueiro-Pagadigorria, J. Constantin, A. Bracht, E.A. Nascimento, E.L. Ishii-Iwamoto, Effects of the non-steroidal anti-inflammatory drug piroxicam on energy metabolism in the perfused rat liver, Comp. Biochem. Physiol. 113C (1996) 93 – 98. [32] E.A. Nascimento, N.S. Yamamoto, A. Bracht, E.L. Ishii-Iwamoto, Naproxen inhibits hepatic glycogenolysis induced by Ca2 + -dependent agents, Gen. Pharmacol. 253 (1995) 211 – 218. [33] A.M. Kelmer-Bracht, E.L. Ishii-Iwamoto, A. Bracht, Intracellular binding of the anti-inflammatory drug niflumic acid in the liver, Res. Commun. Mol. Pathol. Pharmacol. 89 (1995) 411 – 419. [34] O. Ferraresi-Filho, E.L. Ishii-Iwamoto, A. Bracht, Transport, metabolism and distribution space of octanoate in the perfused rat liver, Cell Biochem. Funct. 15 (1997) 69 – 80. [35] G. Weissmann, Aspirin, Sci. Am. 264 (1991) 58 – 64.