261
Biochimica et Biophysica Acta, 1033 (1990) 261-266 Elsevier
BBAGEN 23262
Effect of the antiepileptic drug sodium valproate on glutamine and glutamate metabolism in isolated human kidney tubules Guy Martin, Daniel Durozard, Jean Besson and Gabriel Baverel INSERM U 80 and CNRS UA 1177, Laboratoire de Physiologie Rknale et M~tabolique, Facult~ de Mbdeeine Alexis Carrel Lyon (France)
(Received 28 September 1989)
Key words: Glutamine; Glutamate; Ammoniagenesis; Valproate; (Human kidney)
We studied the effects of sodium valproate, a widely used antiepileptic drug and a hyperammonemic agent, on L-[1-14C]glutamine and L-[l-~4Clglutamate metabolism in isolated human kidney-cortex tubules. Valproate markedly stimulated glutamine removal as well as the formation of ammonia, t4CO2, pyruvate, lactate and alanine, but it inhibited glucose synthesis; the increase in ammonia formation was explained by a stimulation by vaiproate mainly of flux through glutaminase (EC 3.5.1.2) and to a much lesser extent of flux through glutamate dehydrogenase (EC 1.4.1.3). By contrast, vaiproate did not stimulate glutamate removal or ammonia formation, suggesting that the increase in flux through glutamate dehydrogenase observed with glutamine as substrate was secondary to the increase in flux through glutaminase. Accumulation of pyruvate, alanine and lactate in the presence of valproate was less from glutamate than from glutamine. Inhibition by aminooxyacetate of accumulation of alanine from glutamine caused by valproate did not prevent the acceleration of glutamine utilization and the subsequent stimulation of ammonia formation. It is concluded from these data, which are the first concerning the in vitro metabolism of glutamine and glutamate in human kidney-cortex tubules, that the stimulatory effect of valproate is primarily exerted at the level of glutaminase in human renal cortex.
Introduction Sodium valproate (n-dipropylacetic acid, sodium salt), a branched-chain fatty acid, is a widely used and very effective antiepileptic drug in adults and children [1]. Estimates (Internal communication, Sanofi Pharma International) indicate that about 1 million epileptic patients throughout the world are currently undergoing valproate therapy. Drug administration often induces a moderate hyperammonemia especially when other anticonvulsants are taken in addition to valproate [2-7]. Since this hyperammonemia is in some rare eases accompanied by clinical manifestations of liver dysfunction or of encephalopathy [7], it is of clinical importance to identify the origin of the valproate-induced elevation of blood ammonia levels. This adverse effect of valproate was first attributed to an inhibition of the hepatic synthesis of urea leading
Correspondence: G. Baverel, Laboratoire de Physiologie, Rrnale et M&abolique, Facult6 de Mrdecine Alexis Carrel, Rue Guillaume Paradin, 69008 Lyon, France.
to a diminished removal of circulating ammonia by the liver [8,9]; this conclusion arose from studies performed in vitro with isolated rat hepatocytes demonstrating that urea synthesis from alanine or glutamine was decreased by valproate [10,11]. Such an action has been ascribed to a primary reduction caused by valproate of the hepatic concentration of acetylglutamate, an activator of carbamylphosphate synthetase, a key enzyme of the urea cycle [10]. However, measurements on human and rat kidney have recently shown that valproate increases glutamine uptake and ammonia release [12,13]; although the renal blood flow was not measured in the latter studies [12,13], these observations suggest that valproate may be hyperammonemic by enhancing the renal venous release of ammonia. In a more recent study, we have clearly demonstrated that in an intact functioning rat kidney preparation valproate considerably increased the renal production, excretion and venous release of ammonia [14]. The purpose of the present study was to examine the possible mechanisms by which valproate stimulates the renal production of ammonia in the human kidney. For this, human kidney-cortex tubules were isolated and
0304-4163/90/$03.50 © 1990 Elsevier Science Publishers B.V. (Biomedical Division)
262 incubated with glutamine or glutamate as substrate in the absence or the presence of valproate. In addition to providing the first information available on the in vitro metabolism of glutamine and glutamate in adult human kidney tubules, our data establish that in this experimental model valproate stimulates renal ammoniagenesis by increasing primarily flux through glutaminase. Methods
Preparation of isolated human kidney-cortex tubules Fresh normal kidney cortex was obtained from the uninvolved pool of kidneys removed for neoplasm from 18 h fasted adult patients. Specimens of cortex were immediately dissected and placed in ice-cold KrebsHenseleit buffer gassed with a mixture of O2/CO 2 (19 : 1) until the beginning of the tubule-isolation period (usually within 10 min). Kidney tubules were prepared by collagenase treatment of renal cortex slices as described previously [15,16].
Incubations Incubations were performed at 3 7 ° C in a shaking water bath in 25 ml stoppered Erlenmeyer flasks, each with a center well, in an atmosphere of O2/CO 2 (19 : 1). The tubules obtained were incubated for 60 min in 4 ml of Krebs-Henseleit medium (pH 7.40) [17] with 1 mM L-[1Jnc]glutamine or 1 mM L-[1-14C]glutamate as substrate in the absence or the presence of valproate (sodium salt). The flasks were prepared in duplicate for all experimental conditions. Incubations were terminated by adding perchloric acid (2%, v / v , final concentration). In all experiments, zero-time flasks were prepared with and without substrate by adding perchloric acid before the tubules. Collection and measurement of the 14CO2 formed from radioactive glutamine or glutamate were carried out as previously described [18]. After removal of the denaturated protein by centrifugation (4000 × g for 10 min), the supernatant was neutralized with 20% ( w / v ) K O H for metabolite determination.
Analytical methods Glutamine, glutamate, ammonia, alanine, aspartate, pyruvate, lactate, a-ketoglutarate, fumarate, malate, citrate, glucose, glycogen and also the dry weight of the amount of tubules added to the flasks were determined as described in previous papers [15,18].
Calculations Net substrate utilization and product formation were calculated as the difference between the total contents of the flask (tissue + medium) at the start (zero-time flasks) and after 60 min of incubation. 14CO2 production from [1-14C]glutamine or [1-14C]glutamate were calculated by dividing the radioactivity in 14CO2 by the
N.; 1 14C_GLUTAMINE.t..114C_GLUTAMATE
PYRUVATE
SUCCINYL_CoA ~ I_14C_2_OXOGLUTARATE ALANINE 1" I
I
I
C I TRATE
OXALOACETATE
PHOSPHOENOLPYRUVATE I
,I, GLUCOSE
ACETYL .CoA PYRUVATE
1 LACTATE
Fig. 1. Schematic representation of pathways of glutamine metabolism in isolated human kidney-cortex tubules. (Dotted lines indicate that intermediates have been omitted.)
specific radioactivity of the labeled glutamine or glutamate determined in the zero-time samples for each experiment. The metabolic rates are expressed in /~mol of substance removed or produced per h per g dry wt. tubule fragments. They are reported as means + S.E. The results were analyzed by Student's t-test for paired data, comparing values obtained in the presence and the absence of valproate. The metabolic fate of the glutamine or glutamate metabolized by the tubules was established by calculating carbon and nitrogen balances. For these calculations, it was assumed that glutamine or glutamate was the sole source of carbon and nitrogen measured in the form of the various products measured. Complete oxidation of [1J4C]glutamine or [1-14C]glutamate was calculated as the difference between the release of 14CO2, which measures flux through a-ketoglutarate dehydrogenase (see Fig. 1), and the sum of glucose (expressed in C 3 units, because two molecules of glutamine or glutamate are needed for the synthesis of each glucose molecule), alanine, pyruvate and lactate formed. With [1-14C]glutamine or [1J4C]glutamate as substrate, flux through glutamate dehydrogenase was taken as the difference between the 14CO2 released and the alanine found (see Fig. 1).
Reagents L-Glutamine, L-glutamate, glutaminase (grade V), aminooxyacetate and rotenone were supplied by Sigma Chemical Co. (St Louis, MO, U.S.A.). Other enzymes and coenzymes came from Boehringer (Meylan, France). Sodium valproate was supplied by Sanofi Recherche (Service de Toxicologie, Montpellier, France). L-[1JaC]Glutamate (50 m C i / m m o l ) was obtained from Amersham International (Amersham, Bucks., U.K.). L-[1-1aC]Glutamine was synthesized by the method of Baverel and Lund [18]. The other chemicals used were of analytical grade.
263 Results At 0.01-10 raM, valproate greatly accelerated glutamine removal by isolated human kidney-cortex tubules; this effect was accompanied by a marked stimulation of the formation of ammonia, 14CO2, lactate and alanine as well as of the accumulation of pyruvate (Table I). The stimulation by valproate (0.1 mM) of both glutamine removal and ammonia formation was found to be linear with time when human kidney tubules were incubated for 30 and 60 min in the presence of 1 mM glutamine as substrate (n = 2; not shown). A dose-dependent stimulation of glutamine metabolism was also observed with 5 mM glutamine as substrate (n = 2; not shown). By contrast, increasing concentrations of valproate did not alter the accumulation of glutamate, but progressively inhibited the formation of glucose (see Table I). No significant accumulation of intermediates of the tricarboxylic acid cycle or of aspartate or glycogen was observed and no glycogen was present in the tubules at the start of the incubation. Carbon-balance calculations from the data of Table I indicate that, in the absence of valproate, the carbon skeleton of the glutamine metabolized was accounted for mainly by the formation of glucose (74.5%) and to a lesser extent by the formation of glutamate, alanine, lactate and pyruvate; no room was left for complete oxidation of glutamine (see Table I). With increasing concentrations of valproate, the proportion of the glutamine carbon fragment removed that was accounted for by pyruvate, lactate and alanine progressively in-
creased, whereas that accounted for by glucose dramatically decreased (Table I); although the synthesis and accumulation of pyruvate was markedly increased, no flux of pyruvate through pyruvate dehydrogenase occurred as indicated by the absence of complete oxidation of glutamine in the presence of valproate (Table I). Nitrogen-balance calculations reveal that, in the absence or the presence of valproate, most of the nitrogen atoms of glutamine were released as ammonia, but the amount of nitrogen found mainly as ammonia and to a lesser extent as alanine exceeded the amount of nitrogen removed as glutamine (two nitrogen atoms per glutamine molecule), indicating that part of the ammonia found came from endogenous sources, agreeing with the large amount of ammonia formation observed without added glutamine (see Table I). Calculations from the data of Table I also show that the increase in ammonia release caused by valproate was greater than the increase in the glutamine removed, which means that valproate stimulated not only the flux through glutaminase, which releases as ammonia the amide nitrogen of glutamine, but also the flux through glutamate dehydrogenase, which releases the amino nitrogen of glutamate formed from glutamine as ammonia. As can be seen in Table I, the flux through glutamate dehydrogenase was indeed moderately but significantly increased by 0.1 and 1 mM valproate; however, the acceleration of the flux through glutamate dehydrogenase under these conditions was about one order of magnitude less than that through glutaminase (taken as the glutamine removed). With increasing con-
TABLE I
Effect of oalproate on the metabolism of I m M L-[I-HC]glutamine in human kidney tubules K i d n e y tubules (7.8_+1.5 m g dry wt. per flask) were i n c u b a t e d for 60 min as described in the M e t h o d s section. Results ( / x m o l / g dry wt.) for m e t a b o l i t e r e m o v a l ( - ) or p r o d u c t i o n ( + ) are reported as m e a n s + S.E. for four e x p e r i m e n t s p e r f o r m e d in duplicate. Statistical difference was m e a s u r e d by the paired S t u d e n t ' s t-test against the control w i t h o u t valproate: * P < 0.01; * * P < 0.02; * * * P < 0.05. Experimental
M e t a b o l i t e r e m o v a l or p r o d u c t i o n
condition
glutamine
glutamate
ammonia
alanine
glucose
pyruvate
G l u t a m i n e (1 m M )
- 142.1 +25.7 -175.0 _+ 16.0 -217.8 +_30.9 * -262.3 -+42.4 * -275.9 +_35.4 * -267.5 +_50.9 - 5.5 +_1.8 + 5.0 +_1.3
+ 37.6 +8.5 +38.6 _+ 8.6 +34.4 _+6.4 +30.6 -t-5.1 +41.0 _+4.7 +33.6 +3.7 - 15.6 +_4.2 - 48.2 +_5.9 * * *
+ 350.4 +_61.1 +383.0 _+ 58.7 +457.1 -+86.3 +498.8 +_87.0 +527.7 +_85.8 +483.3 +_69.1 + 124.1 +16.3 + 152.6 +_25.2
+ 17.5 +_2.2 +37.3 +_ 9.8 +66.9 +_13.4 * * * +86.3 +_8.9 * +81.4 +13.2 ** +107.5 +_24.0 * * * + 1.8 +_5.3 + 20.6 -+4.8
+ 52.9 _+5.6 +53.8 +_ 5.6 +50.1 +_6.5 +38.2 _+2.8 +19.2 +4.3 +9.1 +_2.3 + 38.0 +2.9 - 3.4 -+2.2
+ 3.2 _+0.6 +4.9 -+ 1.4 +12.5 _+4.1 +16.9 -+2.5 +21.7 5:3.8 +30.5 -+6.9 + 1.2 _+1.1 + 9.7 +_1.4
G l u t a m i n e (1 m M ) + v a l p r o a t e (0.001 m M ) G l u t a m i n e (1 r a M ) + valproate0.01 mM G l u t a m i n e (1 r a M ) + v a l p r o a t e (0.1 m M ) G l u t a m i n e (1 m M ) + v a | p r o a t e (1 m M ) G l u t a m i n e (1 m M ) + valproate(10mM) No substrate No substrate + v a l p r o a t e (0.1 m M )
*** ** * *
Flux through
*** *** *
*
lactate
+ 6.4 +_5.7 +9.0 +_ 7.6 +23.8 +_12.4 +47.4 ** _+4.3 +68.1 ** +-4.9 +56.6 * * * +_3.2 - 5.3 +_16.4 + 37.2 * +_5.8
** ** *
~4CO2
glutamate dehydrogenase
+ 118.2 +_16.2 +135.3 +_ 17.2 +176.3 _+24.2 +201.7 -+27.7 +198.7 +-28.4 +207.9 -+28.9
100.7 +16.7 98.0 _+ 16.3 109.4 +16.3 115.4 +_19.8 * * * 117.3 +_16.2 * * 100.4 +13.1 -
***
* * * *
-
264 TABLE II Effect of valproate on the metabolism of I m M l.-[l. 14C]glutamate in human kidney tubules
Kidney tubules (7.6-+3.5 mg dry wt. per flask) were incubated for 60 rain as described in the Methods section. Results (/.tmol/g dry wt.) for metabolite removal ( - ) or production (+) are reported as means+ S.E. for four experiments performed in duplicate. Statistical difference was measured by the paired Student's t-test against the control without valproate: *P < 0.01; * *P < 0.02; * **P < 0.05. Experimental
Metabolite removal or production
condition
glutamate
Glutamate (1 mM)
-133.7 +_17.1 -130.7 -+17.5 - 152.7 -+22.8 - 164.8 -+27.6 - 149.2 -+17.3 - 145.4 -+18.4 - 1.2
Glutamate (1 mM)+ valproate (0.001 mM) Glutamate (1 mM) + valproate (0.01 mM) Glutamate (1 raM) + valproate (0.1 mM) Glutamate (1 raM) ÷ valproate (1 mM) Glutamate (1 raM) ÷ valproate (10 raM) No substrate No substrate valproate (1 mM)
ammonia
+137.8 +4.3 +128.2 _+8.9 + 140.3 -+11.6 _+143.9 -+11.6 + 135.0 -+10.3 + 133.1 -+5.4 + 79.8 -+ 1 . 6 -+8.0 - 38.9 + 93.7 -+9.4 *** +-9.1
Flux through
alanine
glucose
pyruvate
lactate
14CO2
glutamate dehydrogenase
+31.7 +9.4 +35.1 -+10.7 + 51.2 -+13.9 *** + 55.6 +_8.9 * + 47.4 -+9.4 *** + 65.9 -+8.0 * + 6.2 -+9.3 + 25.4 -+7.3 *
+48.9 -+6.1 +46.1 -+4.8 + 44.1 -+6.4 * + 25.1 -+5.8 * + 10.3 -+2.9 * + 2.7 -+2.1 * + 27.4 -+3.2 + 2.7 -+1.1 *
+3.4 -+1.1 +3.9 +_1.5 + 5.2 -+1.6 + 11.8 +-2.1 *** + 13.3 -+3.4 *** + 10.3 -+0.8 ** + 0.8 -+0.7 + 7.0 -+1.8 ***
+7.6 -+0.9 +8.5 -+0.6 + 13.8 +_3.6 + 38.1 -+5.0 * + 41.2 -+6.8 ** + 36.1 -+6.0 ** + 3.7 -+4.9 + 27.7 -+4.3 **
+148.5 -+31.4 +148.3 +33.2 + 161.5 -+35.2 + 161.9 -+35.5 + 145.1 -+25.8 + 151.3 +-26.4
116.8 -+24.7 113.2 -+25.9 110.3 -+26.8 106.3 -+25.0 97.7 +_17.4 85.4 -+22.3
centrations of valproate the p r o p o r t i o n of the a m i n o nitrogen of the g l u t a m i n e removed which was explained by the formation of a l a n i n e progressively increased at the expense of the f o r m a t i o n of a m m o n i a ; as a n example, the nitrogen f o u n d as a l a n i n e accounted for a b o u t half the nitrogen of the glutamate derived from glutam i n e a n d further metabolized in the presence of 10 m M valproate (Table I). In contrast with the finding with g l u t a m i n e as substrate, no significant change in glutamate removal or in a m m o n i a a n d t 4 c o 2 p r o d u c t i o n was observed in the presence of valproate (Table II); b u t a d d i t i o n of valproate caused a d o s e - d e p e n d e n t i n h i b i t i o n of glucose synthesis a n d a significant increase in pyruvate a c c u m u lation and in lactate a n d a l a n i n e formation; however, this increase was less t h a n that observed with g l u t a m i n e as substrate (see T a b l e I). Again n o a c c u m u l a t i o n of glycogen, asparate or intermediates of the tricarboxylic acid cycle was observed in the absence or the presence of valproate; complete oxidation of glutamate, which represented a negligible fraction (6.0%) of the glutamate metabolized in the absence of valproate, did n o t statistically change in the presence of valproate. In agreement with our d e m o n s t r a t i o n that h u m a n kidney cortex is devoid of g l u t a m i n e synthetase activity [19], n o glutamine synthesis occurred from glutamate. C a r b o n - b a l a n c e calculations from the data of T a b l e II reveal that glucose a n d a l a n i n e represented 73.1% a n d 23.7%, respectively, of the glutamate c a r b o n skeletons used in the absence of valproate; in the presence of
-
-
-
increasing c o n c e n t r a t i o n s of valproate; in the presence of increasing c o n c e n t r a t i o n s of valproate, a l a n i n e a n d lactate progressively b e c a m e the m a i n c a r b o n p r o d u c t s of g l u t a m a t e metabolism. N i t r o g e n - b a l a n c e calculations indicate that, despite the large a c c u m u l a t i o n of a l a n i n e caused b y valproate, u n d e r all c o n d i t i o n s a m m o n i a r e m a i n e d the m a i n n i t r o g e n o u s p r o d u c t of g l u t a m a t e metabolism, which m e a n s that g l u t a m a t e d e h y d r o g e n a s e was the m a i n m e t a b o l i c p a t h w a y by which g l u t a m a t e was degraded in h u m a n k i d n e y tubules. W i t h g l u t a m i n e (1 m M ) as substrate, the a d d i t i o n of valproate (0.1 m M ) plus a m i n o o x y a c e t a t e (1 mM), a n i n h i b i t o r of t r a n s a m i n a s e s [20], did n o t suppress the s t i m u l a t i o n of g l u t a m i n e removal a n d a m m o n i a p r o d u c tion despite a complete i n h i b i t i o n of the increase in a l a n i n e synthesis caused by valproate alone ( n = 4 ; results not shown). Discussion
T o our knowledge, the m e t a b o l i s m of glutamine, the m a i n a m i n o acid r e m o v e d from circulating b l o o d b y the h u m a n kidney [21-25], a n d of glutamate, the first p r o d uct of the g l u t a m i n a s e reaction, has not b e e n previously d o c u m e n t e d in a n y in vitro p r e p a r a t i o n of h u m a n kidney. O u r results, which show that most of the n i t r o g e n of g l u t a m i n e a n d g l u t a m a t e was released as a m m o n i a , together with the high a m o u n t of a m m o n i a p r o d u c t i o n d e m o n s t r a t e d in h u m a n k i d n e y in vivo [13,21,23-25], strongly suggest that in this tissue g l u t a m i n e is m a i n l y
265 metabolized by the combined action of phosphate-dependent glutaminase and glutamate dehydrogenase, two enzymes known to be present in human kidney [26,27] (see Fig. 1). The data of Table I also indicate that, like the intact human kidney, which contains alanine aminotransferase activity [27] and releases alanine into the renal vein [21,22,24,25], isolated human kidney tubules synthesize alanine via the alanine aminotransferase reaction thanks to the formation of both glutamate and pyruvate from glutamine (Fig. 1). Another interesting finding is that at near-physiological concentration of glutamine (1 raM) the carbon skeleton of this amino acid was mainly converted into glucose, whereas complete oxidation was absent; if this also occurs in vivo this would mean that glutamine does not represent an important source of energy for the human kidney. The present study clearly demonstrates that, at therapeutic blood concentrations [28], valproate stimulated the removal of glutamine and the production of ammonia by human kidney tubules; this increase was in the same range as that found by Wafter et al. [13] assuming that, in the patients studied by these authors, the renal blood flow was not affected by valproate. Increased glutamine removal may have resulted from a stimulatory effect of valproate on glutaminase and/or from an augmented removal of glutamate (synthesized from glutamine via glutaminase), a potent end-product inhibitor of renal glutaminase [29]. The fact that the increase in alanine synthesis was less than the increase in glutamine removal (see Table I) strongly suggests that this compound stimulated glutamine removal not only by increasing glutamate metabolism mainly via alanine aminotransferase, but also by stimulating glutaminase. Additional evidence that a stimulatory effect of valproate was exerted primarily on glutaminase was obtained by the addition of aminooxyacetate which suppressed the accumulation of alanine caused by valproate without abolishing the large stimulation of glutamine removal (see the Results section). The latter observation also rules out the possible involvement of glutamine-pyruvate transaminase, whose role in human renal glutamine metabolism has been proposed [30], in the stimulation of glutamine utilization observed in the presence of valproate. Flux through glutamate dehydrogenase was augmented by valproate when glutamine was the substrate, but remained unaltered when glutamate was the substrate (Tables I and II). This clearly indicates that, with glutamine as substrate (Table I), the increased flux through glutamate dehydrogenase observed with 0.1 and 1 mM valproate was secondary to the stimulation of flux through glutaminase. It should be stressed that the stimulation of ammonia synthesis from glutamine was due, to a great extent and under certain conditions, exclusively to the stimulation by valproate of the flux
through glutaminase which under all conditions was higher than the stimulation of the flux through glutamate dehydrogenase (see Table I). Increased synthesis of alanine from both glutamine and glutamate in the presence of valproate (Tables I and II) clearly means that valproate increased the availability of pyruvate for the alanine aminotransferase reaction; increased availability of pyruvate was also revealed by increased accumulation of pyruvate and lactate in the presence of valproate (Tables I and II). Such an accumulation of pyruvate may result from an inhibition by valproate either of pyruvate transport into the mitochondria or of pyruvate metabolism by pyruvate dehydrogenase; in this respect, preliminary experiments by Turnbull et al. [11] have shown that purified pig heart pyruvate dehydrogenase is inhibited by 0.4 mM valproyl-CoA. With the same concentration of valproate, the fact that the stimulation of alanine accumulation was always much greater by glutamine than by glutamate may seem surprising, since glutamate is a substrate of the alanine aminotransferase reaction. This is possibly due to the very large stimulation of pyruvate synthesis from glutamine (calculated as the pyruvate, lactate and alanine formed from glutamine) which resuited from the stimulation of flux through glutaminase. It should be stressed that the synthesis of pyruvate from glutamine observed in the present study might be much more limited in vivo in the presence of circulating hormones, such as catecholamines and parathyroid hormone, which might convert the pyruvate kinase type L present in the renal proximal tubule into its high g m form [31]. Inhibition of glucose synthesis from both glutamine and glutamate observed in the presence of valproate cannot be attributed to an inhibition of the utilization of the oxaloacetate formed from glutamine or glutamate by phosphoenolpyruvate carboxykinase, because no malate or aspartate accumulation was observed; this inhibition may be due, at least in part, to the increase in lactate synthesis (secondary to pyruvate accumulation), which competed with the gluconeogenic pathway for the reducing equivalents available in the cytosol. Finally, it should be pointed out that valproate had no direct stimulatory effect on the activity of glutaminase taken as the glutamate accumulated by isolated human kidney tubules in the presence of 1 or 5 mM glutamine plus 12.5 ~M rotenone (n = 2, results not shown) despite the very large increase in flux through glutaminase demonstrated in the present study (Table I). This suggests that valproate exerted the latter effect by some indirect mechanism which was probably related to the tubular metabolism of valproate; when human tubules were incubated for 60 rain in the presence of 1 mM glutamine and 0.1 mM valproate, valproyl-CoA accumulated and caused a dramatic decrease in the cellular concentrations of both CoA and
266 acetyl-CoA assayed by an HPLC procedure (n = 2; not shown). Since CoA and acetyl-CoA are well-established regulators of renal glutaminase activity [32], it is possible that the changes in their tubular content observed in the presence of valproate stimulated the flux through glutaminase found in the present study. Alternatively, valproyl-CoA might be a stimulator of renal glutaminase. Further studies are necessary to prove these hypotheses. In summary, this study represents the first biochemical characterization of glutamine and glutamate metabolism in isolated human kidney-cortex tubules. Additionally, it provides basic information demonstrating that the antiepileptic drug sodium valproate exerts a stimulatory effect on renal ammoniagenesis at an enzymatic step specific to glutamine metabolism, namely glutaminase. Thus, it adds further support to the view that valproate-induced hyperammonemia is, at least in part, of renal origin [12-14].
Acknowledgments We are grateful to Professor J.M. Dubernard and his co-workers who provided human kidneys. Excellent technical assistance was provided by Vrronique Prlamatti. This work was supported by grants from M.R.T. (85-C-1091) and from the Centre de Recherche CLIN-MIDY (Service de Toxicologie).
References 1 2 3 4
Simon, D. and Penry, J.K. (1975) Epilepsia 16, 549-573. Coulter, D.L. and Allen, R.J. (1980) Lancet i, 1310-1311. Coulter, D.L. and Allen, R.J. (1981) J. Pediatr. 99, 317-319. Warter, J.M., Marescaux, C., Rumbach, L., Galmich, J., Micheletti, G., Koehl, C. and Imler, M. (1981) La Nouv. Presse Med. 10, 3406-3407. 5 Rawat, S., Borkowski, W.J. and Swick, H.M. (1981) Neurology (N.Y.) 31, 1173-1174. 6 Ettlin, Th. and Scollo-Lavizzari, G. (1986) Schweiz, Rundschau Med. 75, 881-885. 7 Powell-Jackson, P.R., Tredger, J.M. and Williams, R. (1984) Gut 25, 673-681.
8 Coud6, F.X., Rabier, D., Cathelineau, L., Grimber, G., Parvy, P. and Kamoun, P. (1981) Pediatr. Res. 15, 974-975. 9 Coud6, F.X., Rabier, D., Cathelineau, L., Grimber, G., Parvy, P. and Kamoun, P. (1982) Adv. Exp. Med. Biol. 153, 153-161. 10 Coud6, F.X., Grimber, G., Parvy, P, Rabier, D. and Petit, F. (1983) Biochem. J. 216, 233-236. 11 Turnbull, D.M., Bone, A.J., Bartlett, K., Koundakjian, P.P. and Sherratt, H.S.A. (1983) Biochem. Pharmacol. 32, 1887-1892. 12 Warter, J.M., Imler, M., Marescaux, C., Chabrier, G., Rumbach, L., Micheletti, G. and Krieger, J. (1983) Eur. J. Pharmacol. 87, 177-182. 13 Warter, J.M., Marescaux, C., Chabrier, G., Rumbach, L., Micheletti, G. and Imler, M. (1984) Rev. Neurol. (Paris) 140, 370-371. 14 Ferrier, B., Martin, M. and Baverel, G. (1988) J. Clin. Chem. Clin. Biochem. 26, 65-67. 15 Baverel, G., Bonnard, M., d'Armagnac de Castanet, E. and Pellet, M. (1978) Kidney Int. 14, 567-575. 16 Baverel, G., Bonnard, M., d'Armagnac de Castanet, E. and Pellet, M. (1979) FEBS Lett. 101,282-286. 17 Krebs, H.A. and Henseleit, K. (1932) Hoppe-Seyler's Z. Physiol. Chem. 210, 33-66. 18 Baverel, G. and Lund, P. (1979) Biochem. J. 184, 599-606. 19 Lemieux, G., Baverel, G., Vinay, P. and Wadoux, P. (1976) Am. J. Physiol. 231 1068-1073. 20 Braunstein, A.E. (1964) Vitam. Horm. (N.Y.) 22, 451-484. 21 Owen, E.E. and Robinson, R.R. (1963) J. Clin. Invest. 42 263-276. 22 Wahren, J. and Felig, P. (1975) Diabetes 24, 730-734. 23 Tizianello, A., De Ferrari, G., Garibotto, G. and Gurreri, G. (1978) Clin. Sci. Mol. Med. 55, 391-397. 24 Tizianello, A., De Ferrari, G., Garibotto, G., Gurreri, G. and Robaudo, C. (1980) J. Clin. Invest. 65, 1162-1173. 25 Tizianello, A. De Ferrari, G., Garibotto, G., Robaudo, C., Acquarone, N. and Ghiggeri, G.M. (1982) J. Clin. Invest. 69, 240-250. 26 Mattenheimer, H. and De Bruin, H. (1964) Enzymol. Biol. Clin. 4, 65-83. 27 Mattenheimer, H., Pollack, V.E. and Muehrcke, R.C. (1970) Nephron 7, 144-154. 28 Turnbull, D.M., Dick, D.J., Wilson, L., Sherratt, H.S.A. and Alberti, K.G.M.M. (1986) J. Neurol. Neurosurg. Psychiatr. 49, 405-410. 29 Goldstein, L. (1966) Am. J. Physiol. 210, 661-666. 30 Fine, A., Scott, J. and Bourke, J. (1972) J. Lab. Clin. Med. 80, 591-597. 31 Schering, B., Reinacher, M. and Schoner, W. (1986) Biochim. Biophys. Acta 881, 62-71. 32 Kvamme, E. and Torgner, I.M. (1974) Biochem. J. 137, 525-530.
Biochimica et Biophysica Acta, 1033 (1990) 261-266 Elsevier
BBAGEN 23262
Effect of the antiepileptic drug sodium valproate on glutamine and glutamate metabolism in isolated human kidney tubules Guy Martin, Daniel Durozard, Jean Besson and Gabriel Baverel INSERM U 80 and CNRS UA 1177, Laboratoire de Physiologie Rknale et M~tabolique, Facult~ de Mbdeeine Alexis Carrel Lyon (France)
(Received 28 September 1989)
Key words: Glutamine; Glutamate; Ammoniagenesis; Valproate; (Human kidney)
We studied the effects of sodium valproate, a widely used antiepileptic drug and a hyperammonemic agent, on L-[1-14C]glutamine and L-[l-~4Clglutamate metabolism in isolated human kidney-cortex tubules. Valproate markedly stimulated glutamine removal as well as the formation of ammonia, t4CO2, pyruvate, lactate and alanine, but it inhibited glucose synthesis; the increase in ammonia formation was explained by a stimulation by vaiproate mainly of flux through glutaminase (EC 3.5.1.2) and to a much lesser extent of flux through glutamate dehydrogenase (EC 1.4.1.3). By contrast, vaiproate did not stimulate glutamate removal or ammonia formation, suggesting that the increase in flux through glutamate dehydrogenase observed with glutamine as substrate was secondary to the increase in flux through glutaminase. Accumulation of pyruvate, alanine and lactate in the presence of valproate was less from glutamate than from glutamine. Inhibition by aminooxyacetate of accumulation of alanine from glutamine caused by valproate did not prevent the acceleration of glutamine utilization and the subsequent stimulation of ammonia formation. It is concluded from these data, which are the first concerning the in vitro metabolism of glutamine and glutamate in human kidney-cortex tubules, that the stimulatory effect of valproate is primarily exerted at the level of glutaminase in human renal cortex.
Introduction Sodium valproate (n-dipropylacetic acid, sodium salt), a branched-chain fatty acid, is a widely used and very effective antiepileptic drug in adults and children [1]. Estimates (Internal communication, Sanofi Pharma International) indicate that about 1 million epileptic patients throughout the world are currently undergoing valproate therapy. Drug administration often induces a moderate hyperammonemia especially when other anticonvulsants are taken in addition to valproate [2-7]. Since this hyperammonemia is in some rare eases accompanied by clinical manifestations of liver dysfunction or of encephalopathy [7], it is of clinical importance to identify the origin of the valproate-induced elevation of blood ammonia levels. This adverse effect of valproate was first attributed to an inhibition of the hepatic synthesis of urea leading
Correspondence: G. Baverel, Laboratoire de Physiologie, Rrnale et M&abolique, Facult6 de Mrdecine Alexis Carrel, Rue Guillaume Paradin, 69008 Lyon, France.
to a diminished removal of circulating ammonia by the liver [8,9]; this conclusion arose from studies performed in vitro with isolated rat hepatocytes demonstrating that urea synthesis from alanine or glutamine was decreased by valproate [10,11]. Such an action has been ascribed to a primary reduction caused by valproate of the hepatic concentration of acetylglutamate, an activator of carbamylphosphate synthetase, a key enzyme of the urea cycle [10]. However, measurements on human and rat kidney have recently shown that valproate increases glutamine uptake and ammonia release [12,13]; although the renal blood flow was not measured in the latter studies [12,13], these observations suggest that valproate may be hyperammonemic by enhancing the renal venous release of ammonia. In a more recent study, we have clearly demonstrated that in an intact functioning rat kidney preparation valproate considerably increased the renal production, excretion and venous release of ammonia [14]. The purpose of the present study was to examine the possible mechanisms by which valproate stimulates the renal production of ammonia in the human kidney. For this, human kidney-cortex tubules were isolated and
0304-4163/90/$03.50 © 1990 Elsevier Science Publishers B.V. (Biomedical Division)
262 incubated with glutamine or glutamate as substrate in the absence or the presence of valproate. In addition to providing the first information available on the in vitro metabolism of glutamine and glutamate in adult human kidney tubules, our data establish that in this experimental model valproate stimulates renal ammoniagenesis by increasing primarily flux through glutaminase. Methods
Preparation of isolated human kidney-cortex tubules Fresh normal kidney cortex was obtained from the uninvolved pool of kidneys removed for neoplasm from 18 h fasted adult patients. Specimens of cortex were immediately dissected and placed in ice-cold KrebsHenseleit buffer gassed with a mixture of O2/CO 2 (19 : 1) until the beginning of the tubule-isolation period (usually within 10 min). Kidney tubules were prepared by collagenase treatment of renal cortex slices as described previously [15,16].
Incubations Incubations were performed at 3 7 ° C in a shaking water bath in 25 ml stoppered Erlenmeyer flasks, each with a center well, in an atmosphere of O2/CO 2 (19 : 1). The tubules obtained were incubated for 60 min in 4 ml of Krebs-Henseleit medium (pH 7.40) [17] with 1 mM L-[1Jnc]glutamine or 1 mM L-[1-14C]glutamate as substrate in the absence or the presence of valproate (sodium salt). The flasks were prepared in duplicate for all experimental conditions. Incubations were terminated by adding perchloric acid (2%, v / v , final concentration). In all experiments, zero-time flasks were prepared with and without substrate by adding perchloric acid before the tubules. Collection and measurement of the 14CO2 formed from radioactive glutamine or glutamate were carried out as previously described [18]. After removal of the denaturated protein by centrifugation (4000 × g for 10 min), the supernatant was neutralized with 20% ( w / v ) K O H for metabolite determination.
Analytical methods Glutamine, glutamate, ammonia, alanine, aspartate, pyruvate, lactate, a-ketoglutarate, fumarate, malate, citrate, glucose, glycogen and also the dry weight of the amount of tubules added to the flasks were determined as described in previous papers [15,18].
Calculations Net substrate utilization and product formation were calculated as the difference between the total contents of the flask (tissue + medium) at the start (zero-time flasks) and after 60 min of incubation. 14CO2 production from [1-14C]glutamine or [1-14C]glutamate were calculated by dividing the radioactivity in 14CO2 by the
N.; 1 14C_GLUTAMINE.t..114C_GLUTAMATE
PYRUVATE
SUCCINYL_CoA ~ I_14C_2_OXOGLUTARATE ALANINE 1" I
I
I
C I TRATE
OXALOACETATE
PHOSPHOENOLPYRUVATE I
,I, GLUCOSE
ACETYL .CoA PYRUVATE
1 LACTATE
Fig. 1. Schematic representation of pathways of glutamine metabolism in isolated human kidney-cortex tubules. (Dotted lines indicate that intermediates have been omitted.)
specific radioactivity of the labeled glutamine or glutamate determined in the zero-time samples for each experiment. The metabolic rates are expressed in /~mol of substance removed or produced per h per g dry wt. tubule fragments. They are reported as means + S.E. The results were analyzed by Student's t-test for paired data, comparing values obtained in the presence and the absence of valproate. The metabolic fate of the glutamine or glutamate metabolized by the tubules was established by calculating carbon and nitrogen balances. For these calculations, it was assumed that glutamine or glutamate was the sole source of carbon and nitrogen measured in the form of the various products measured. Complete oxidation of [1J4C]glutamine or [1-14C]glutamate was calculated as the difference between the release of 14CO2, which measures flux through a-ketoglutarate dehydrogenase (see Fig. 1), and the sum of glucose (expressed in C 3 units, because two molecules of glutamine or glutamate are needed for the synthesis of each glucose molecule), alanine, pyruvate and lactate formed. With [1-14C]glutamine or [1J4C]glutamate as substrate, flux through glutamate dehydrogenase was taken as the difference between the 14CO2 released and the alanine found (see Fig. 1).
Reagents L-Glutamine, L-glutamate, glutaminase (grade V), aminooxyacetate and rotenone were supplied by Sigma Chemical Co. (St Louis, MO, U.S.A.). Other enzymes and coenzymes came from Boehringer (Meylan, France). Sodium valproate was supplied by Sanofi Recherche (Service de Toxicologie, Montpellier, France). L-[1JaC]Glutamate (50 m C i / m m o l ) was obtained from Amersham International (Amersham, Bucks., U.K.). L-[1-1aC]Glutamine was synthesized by the method of Baverel and Lund [18]. The other chemicals used were of analytical grade.
263 Results At 0.01-10 raM, valproate greatly accelerated glutamine removal by isolated human kidney-cortex tubules; this effect was accompanied by a marked stimulation of the formation of ammonia, 14CO2, lactate and alanine as well as of the accumulation of pyruvate (Table I). The stimulation by valproate (0.1 mM) of both glutamine removal and ammonia formation was found to be linear with time when human kidney tubules were incubated for 30 and 60 min in the presence of 1 mM glutamine as substrate (n = 2; not shown). A dose-dependent stimulation of glutamine metabolism was also observed with 5 mM glutamine as substrate (n = 2; not shown). By contrast, increasing concentrations of valproate did not alter the accumulation of glutamate, but progressively inhibited the formation of glucose (see Table I). No significant accumulation of intermediates of the tricarboxylic acid cycle or of aspartate or glycogen was observed and no glycogen was present in the tubules at the start of the incubation. Carbon-balance calculations from the data of Table I indicate that, in the absence of valproate, the carbon skeleton of the glutamine metabolized was accounted for mainly by the formation of glucose (74.5%) and to a lesser extent by the formation of glutamate, alanine, lactate and pyruvate; no room was left for complete oxidation of glutamine (see Table I). With increasing concentrations of valproate, the proportion of the glutamine carbon fragment removed that was accounted for by pyruvate, lactate and alanine progressively in-
creased, whereas that accounted for by glucose dramatically decreased (Table I); although the synthesis and accumulation of pyruvate was markedly increased, no flux of pyruvate through pyruvate dehydrogenase occurred as indicated by the absence of complete oxidation of glutamine in the presence of valproate (Table I). Nitrogen-balance calculations reveal that, in the absence or the presence of valproate, most of the nitrogen atoms of glutamine were released as ammonia, but the amount of nitrogen found mainly as ammonia and to a lesser extent as alanine exceeded the amount of nitrogen removed as glutamine (two nitrogen atoms per glutamine molecule), indicating that part of the ammonia found came from endogenous sources, agreeing with the large amount of ammonia formation observed without added glutamine (see Table I). Calculations from the data of Table I also show that the increase in ammonia release caused by valproate was greater than the increase in the glutamine removed, which means that valproate stimulated not only the flux through glutaminase, which releases as ammonia the amide nitrogen of glutamine, but also the flux through glutamate dehydrogenase, which releases the amino nitrogen of glutamate formed from glutamine as ammonia. As can be seen in Table I, the flux through glutamate dehydrogenase was indeed moderately but significantly increased by 0.1 and 1 mM valproate; however, the acceleration of the flux through glutamate dehydrogenase under these conditions was about one order of magnitude less than that through glutaminase (taken as the glutamine removed). With increasing con-
TABLE I
Effect of oalproate on the metabolism of I m M L-[I-HC]glutamine in human kidney tubules K i d n e y tubules (7.8_+1.5 m g dry wt. per flask) were i n c u b a t e d for 60 min as described in the M e t h o d s section. Results ( / x m o l / g dry wt.) for m e t a b o l i t e r e m o v a l ( - ) or p r o d u c t i o n ( + ) are reported as m e a n s + S.E. for four e x p e r i m e n t s p e r f o r m e d in duplicate. Statistical difference was m e a s u r e d by the paired S t u d e n t ' s t-test against the control w i t h o u t valproate: * P < 0.01; * * P < 0.02; * * * P < 0.05. Experimental
M e t a b o l i t e r e m o v a l or p r o d u c t i o n
condition
glutamine
glutamate
ammonia
alanine
glucose
pyruvate
G l u t a m i n e (1 m M )
- 142.1 +25.7 -175.0 _+ 16.0 -217.8 +_30.9 * -262.3 -+42.4 * -275.9 +_35.4 * -267.5 +_50.9 - 5.5 +_1.8 + 5.0 +_1.3
+ 37.6 +8.5 +38.6 _+ 8.6 +34.4 _+6.4 +30.6 -t-5.1 +41.0 _+4.7 +33.6 +3.7 - 15.6 +_4.2 - 48.2 +_5.9 * * *
+ 350.4 +_61.1 +383.0 _+ 58.7 +457.1 -+86.3 +498.8 +_87.0 +527.7 +_85.8 +483.3 +_69.1 + 124.1 +16.3 + 152.6 +_25.2
+ 17.5 +_2.2 +37.3 +_ 9.8 +66.9 +_13.4 * * * +86.3 +_8.9 * +81.4 +13.2 ** +107.5 +_24.0 * * * + 1.8 +_5.3 + 20.6 -+4.8
+ 52.9 _+5.6 +53.8 +_ 5.6 +50.1 +_6.5 +38.2 _+2.8 +19.2 +4.3 +9.1 +_2.3 + 38.0 +2.9 - 3.4 -+2.2
+ 3.2 _+0.6 +4.9 -+ 1.4 +12.5 _+4.1 +16.9 -+2.5 +21.7 5:3.8 +30.5 -+6.9 + 1.2 _+1.1 + 9.7 +_1.4
G l u t a m i n e (1 m M ) + v a l p r o a t e (0.001 m M ) G l u t a m i n e (1 r a M ) + valproate0.01 mM G l u t a m i n e (1 r a M ) + v a l p r o a t e (0.1 m M ) G l u t a m i n e (1 m M ) + v a | p r o a t e (1 m M ) G l u t a m i n e (1 m M ) + valproate(10mM) No substrate No substrate + v a l p r o a t e (0.1 m M )
*** ** * *
Flux through
*** *** *
*
lactate
+ 6.4 +_5.7 +9.0 +_ 7.6 +23.8 +_12.4 +47.4 ** _+4.3 +68.1 ** +-4.9 +56.6 * * * +_3.2 - 5.3 +_16.4 + 37.2 * +_5.8
** ** *
~4CO2
glutamate dehydrogenase
+ 118.2 +_16.2 +135.3 +_ 17.2 +176.3 _+24.2 +201.7 -+27.7 +198.7 +-28.4 +207.9 -+28.9
100.7 +16.7 98.0 _+ 16.3 109.4 +16.3 115.4 +_19.8 * * * 117.3 +_16.2 * * 100.4 +13.1 -
***
* * * *
-
264 TABLE II Effect of valproate on the metabolism of I m M l.-[l. 14C]glutamate in human kidney tubules
Kidney tubules (7.6-+3.5 mg dry wt. per flask) were incubated for 60 rain as described in the Methods section. Results (/.tmol/g dry wt.) for metabolite removal ( - ) or production (+) are reported as means+ S.E. for four experiments performed in duplicate. Statistical difference was measured by the paired Student's t-test against the control without valproate: *P < 0.01; * *P < 0.02; * **P < 0.05. Experimental
Metabolite removal or production
condition
glutamate
Glutamate (1 mM)
-133.7 +_17.1 -130.7 -+17.5 - 152.7 -+22.8 - 164.8 -+27.6 - 149.2 -+17.3 - 145.4 -+18.4 - 1.2
Glutamate (1 mM)+ valproate (0.001 mM) Glutamate (1 mM) + valproate (0.01 mM) Glutamate (1 raM) + valproate (0.1 mM) Glutamate (1 raM) ÷ valproate (1 mM) Glutamate (1 raM) ÷ valproate (10 raM) No substrate No substrate valproate (1 mM)
ammonia
+137.8 +4.3 +128.2 _+8.9 + 140.3 -+11.6 _+143.9 -+11.6 + 135.0 -+10.3 + 133.1 -+5.4 + 79.8 -+ 1 . 6 -+8.0 - 38.9 + 93.7 -+9.4 *** +-9.1
Flux through
alanine
glucose
pyruvate
lactate
14CO2
glutamate dehydrogenase
+31.7 +9.4 +35.1 -+10.7 + 51.2 -+13.9 *** + 55.6 +_8.9 * + 47.4 -+9.4 *** + 65.9 -+8.0 * + 6.2 -+9.3 + 25.4 -+7.3 *
+48.9 -+6.1 +46.1 -+4.8 + 44.1 -+6.4 * + 25.1 -+5.8 * + 10.3 -+2.9 * + 2.7 -+2.1 * + 27.4 -+3.2 + 2.7 -+1.1 *
+3.4 -+1.1 +3.9 +_1.5 + 5.2 -+1.6 + 11.8 +-2.1 *** + 13.3 -+3.4 *** + 10.3 -+0.8 ** + 0.8 -+0.7 + 7.0 -+1.8 ***
+7.6 -+0.9 +8.5 -+0.6 + 13.8 +_3.6 + 38.1 -+5.0 * + 41.2 -+6.8 ** + 36.1 -+6.0 ** + 3.7 -+4.9 + 27.7 -+4.3 **
+148.5 -+31.4 +148.3 +33.2 + 161.5 -+35.2 + 161.9 -+35.5 + 145.1 -+25.8 + 151.3 +-26.4
116.8 -+24.7 113.2 -+25.9 110.3 -+26.8 106.3 -+25.0 97.7 +_17.4 85.4 -+22.3
centrations of valproate the p r o p o r t i o n of the a m i n o nitrogen of the g l u t a m i n e removed which was explained by the formation of a l a n i n e progressively increased at the expense of the f o r m a t i o n of a m m o n i a ; as a n example, the nitrogen f o u n d as a l a n i n e accounted for a b o u t half the nitrogen of the glutamate derived from glutam i n e a n d further metabolized in the presence of 10 m M valproate (Table I). In contrast with the finding with g l u t a m i n e as substrate, no significant change in glutamate removal or in a m m o n i a a n d t 4 c o 2 p r o d u c t i o n was observed in the presence of valproate (Table II); b u t a d d i t i o n of valproate caused a d o s e - d e p e n d e n t i n h i b i t i o n of glucose synthesis a n d a significant increase in pyruvate a c c u m u lation and in lactate a n d a l a n i n e formation; however, this increase was less t h a n that observed with g l u t a m i n e as substrate (see T a b l e I). Again n o a c c u m u l a t i o n of glycogen, asparate or intermediates of the tricarboxylic acid cycle was observed in the absence or the presence of valproate; complete oxidation of glutamate, which represented a negligible fraction (6.0%) of the glutamate metabolized in the absence of valproate, did n o t statistically change in the presence of valproate. In agreement with our d e m o n s t r a t i o n that h u m a n kidney cortex is devoid of g l u t a m i n e synthetase activity [19], n o glutamine synthesis occurred from glutamate. C a r b o n - b a l a n c e calculations from the data of T a b l e II reveal that glucose a n d a l a n i n e represented 73.1% a n d 23.7%, respectively, of the glutamate c a r b o n skeletons used in the absence of valproate; in the presence of
-
-
-
increasing c o n c e n t r a t i o n s of valproate; in the presence of increasing c o n c e n t r a t i o n s of valproate, a l a n i n e a n d lactate progressively b e c a m e the m a i n c a r b o n p r o d u c t s of g l u t a m a t e metabolism. N i t r o g e n - b a l a n c e calculations indicate that, despite the large a c c u m u l a t i o n of a l a n i n e caused b y valproate, u n d e r all c o n d i t i o n s a m m o n i a r e m a i n e d the m a i n n i t r o g e n o u s p r o d u c t of g l u t a m a t e metabolism, which m e a n s that g l u t a m a t e d e h y d r o g e n a s e was the m a i n m e t a b o l i c p a t h w a y by which g l u t a m a t e was degraded in h u m a n k i d n e y tubules. W i t h g l u t a m i n e (1 m M ) as substrate, the a d d i t i o n of valproate (0.1 m M ) plus a m i n o o x y a c e t a t e (1 mM), a n i n h i b i t o r of t r a n s a m i n a s e s [20], did n o t suppress the s t i m u l a t i o n of g l u t a m i n e removal a n d a m m o n i a p r o d u c tion despite a complete i n h i b i t i o n of the increase in a l a n i n e synthesis caused by valproate alone ( n = 4 ; results not shown). Discussion
T o our knowledge, the m e t a b o l i s m of glutamine, the m a i n a m i n o acid r e m o v e d from circulating b l o o d b y the h u m a n kidney [21-25], a n d of glutamate, the first p r o d uct of the g l u t a m i n a s e reaction, has not b e e n previously d o c u m e n t e d in a n y in vitro p r e p a r a t i o n of h u m a n kidney. O u r results, which show that most of the n i t r o g e n of g l u t a m i n e a n d g l u t a m a t e was released as a m m o n i a , together with the high a m o u n t of a m m o n i a p r o d u c t i o n d e m o n s t r a t e d in h u m a n k i d n e y in vivo [13,21,23-25], strongly suggest that in this tissue g l u t a m i n e is m a i n l y
265 metabolized by the combined action of phosphate-dependent glutaminase and glutamate dehydrogenase, two enzymes known to be present in human kidney [26,27] (see Fig. 1). The data of Table I also indicate that, like the intact human kidney, which contains alanine aminotransferase activity [27] and releases alanine into the renal vein [21,22,24,25], isolated human kidney tubules synthesize alanine via the alanine aminotransferase reaction thanks to the formation of both glutamate and pyruvate from glutamine (Fig. 1). Another interesting finding is that at near-physiological concentration of glutamine (1 raM) the carbon skeleton of this amino acid was mainly converted into glucose, whereas complete oxidation was absent; if this also occurs in vivo this would mean that glutamine does not represent an important source of energy for the human kidney. The present study clearly demonstrates that, at therapeutic blood concentrations [28], valproate stimulated the removal of glutamine and the production of ammonia by human kidney tubules; this increase was in the same range as that found by Wafter et al. [13] assuming that, in the patients studied by these authors, the renal blood flow was not affected by valproate. Increased glutamine removal may have resulted from a stimulatory effect of valproate on glutaminase and/or from an augmented removal of glutamate (synthesized from glutamine via glutaminase), a potent end-product inhibitor of renal glutaminase [29]. The fact that the increase in alanine synthesis was less than the increase in glutamine removal (see Table I) strongly suggests that this compound stimulated glutamine removal not only by increasing glutamate metabolism mainly via alanine aminotransferase, but also by stimulating glutaminase. Additional evidence that a stimulatory effect of valproate was exerted primarily on glutaminase was obtained by the addition of aminooxyacetate which suppressed the accumulation of alanine caused by valproate without abolishing the large stimulation of glutamine removal (see the Results section). The latter observation also rules out the possible involvement of glutamine-pyruvate transaminase, whose role in human renal glutamine metabolism has been proposed [30], in the stimulation of glutamine utilization observed in the presence of valproate. Flux through glutamate dehydrogenase was augmented by valproate when glutamine was the substrate, but remained unaltered when glutamate was the substrate (Tables I and II). This clearly indicates that, with glutamine as substrate (Table I), the increased flux through glutamate dehydrogenase observed with 0.1 and 1 mM valproate was secondary to the stimulation of flux through glutaminase. It should be stressed that the stimulation of ammonia synthesis from glutamine was due, to a great extent and under certain conditions, exclusively to the stimulation by valproate of the flux
through glutaminase which under all conditions was higher than the stimulation of the flux through glutamate dehydrogenase (see Table I). Increased synthesis of alanine from both glutamine and glutamate in the presence of valproate (Tables I and II) clearly means that valproate increased the availability of pyruvate for the alanine aminotransferase reaction; increased availability of pyruvate was also revealed by increased accumulation of pyruvate and lactate in the presence of valproate (Tables I and II). Such an accumulation of pyruvate may result from an inhibition by valproate either of pyruvate transport into the mitochondria or of pyruvate metabolism by pyruvate dehydrogenase; in this respect, preliminary experiments by Turnbull et al. [11] have shown that purified pig heart pyruvate dehydrogenase is inhibited by 0.4 mM valproyl-CoA. With the same concentration of valproate, the fact that the stimulation of alanine accumulation was always much greater by glutamine than by glutamate may seem surprising, since glutamate is a substrate of the alanine aminotransferase reaction. This is possibly due to the very large stimulation of pyruvate synthesis from glutamine (calculated as the pyruvate, lactate and alanine formed from glutamine) which resuited from the stimulation of flux through glutaminase. It should be stressed that the synthesis of pyruvate from glutamine observed in the present study might be much more limited in vivo in the presence of circulating hormones, such as catecholamines and parathyroid hormone, which might convert the pyruvate kinase type L present in the renal proximal tubule into its high g m form [31]. Inhibition of glucose synthesis from both glutamine and glutamate observed in the presence of valproate cannot be attributed to an inhibition of the utilization of the oxaloacetate formed from glutamine or glutamate by phosphoenolpyruvate carboxykinase, because no malate or aspartate accumulation was observed; this inhibition may be due, at least in part, to the increase in lactate synthesis (secondary to pyruvate accumulation), which competed with the gluconeogenic pathway for the reducing equivalents available in the cytosol. Finally, it should be pointed out that valproate had no direct stimulatory effect on the activity of glutaminase taken as the glutamate accumulated by isolated human kidney tubules in the presence of 1 or 5 mM glutamine plus 12.5 ~M rotenone (n = 2, results not shown) despite the very large increase in flux through glutaminase demonstrated in the present study (Table I). This suggests that valproate exerted the latter effect by some indirect mechanism which was probably related to the tubular metabolism of valproate; when human tubules were incubated for 60 rain in the presence of 1 mM glutamine and 0.1 mM valproate, valproyl-CoA accumulated and caused a dramatic decrease in the cellular concentrations of both CoA and
266 acetyl-CoA assayed by an HPLC procedure (n = 2; not shown). Since CoA and acetyl-CoA are well-established regulators of renal glutaminase activity [32], it is possible that the changes in their tubular content observed in the presence of valproate stimulated the flux through glutaminase found in the present study. Alternatively, valproyl-CoA might be a stimulator of renal glutaminase. Further studies are necessary to prove these hypotheses. In summary, this study represents the first biochemical characterization of glutamine and glutamate metabolism in isolated human kidney-cortex tubules. Additionally, it provides basic information demonstrating that the antiepileptic drug sodium valproate exerts a stimulatory effect on renal ammoniagenesis at an enzymatic step specific to glutamine metabolism, namely glutaminase. Thus, it adds further support to the view that valproate-induced hyperammonemia is, at least in part, of renal origin [12-14].
Acknowledgments We are grateful to Professor J.M. Dubernard and his co-workers who provided human kidneys. Excellent technical assistance was provided by Vrronique Prlamatti. This work was supported by grants from M.R.T. (85-C-1091) and from the Centre de Recherche CLIN-MIDY (Service de Toxicologie).
References 1 2 3 4
Simon, D. and Penry, J.K. (1975) Epilepsia 16, 549-573. Coulter, D.L. and Allen, R.J. (1980) Lancet i, 1310-1311. Coulter, D.L. and Allen, R.J. (1981) J. Pediatr. 99, 317-319. Warter, J.M., Marescaux, C., Rumbach, L., Galmich, J., Micheletti, G., Koehl, C. and Imler, M. (1981) La Nouv. Presse Med. 10, 3406-3407. 5 Rawat, S., Borkowski, W.J. and Swick, H.M. (1981) Neurology (N.Y.) 31, 1173-1174. 6 Ettlin, Th. and Scollo-Lavizzari, G. (1986) Schweiz, Rundschau Med. 75, 881-885. 7 Powell-Jackson, P.R., Tredger, J.M. and Williams, R. (1984) Gut 25, 673-681.
8 Coud6, F.X., Rabier, D., Cathelineau, L., Grimber, G., Parvy, P. and Kamoun, P. (1981) Pediatr. Res. 15, 974-975. 9 Coud6, F.X., Rabier, D., Cathelineau, L., Grimber, G., Parvy, P. and Kamoun, P. (1982) Adv. Exp. Med. Biol. 153, 153-161. 10 Coud6, F.X., Grimber, G., Parvy, P, Rabier, D. and Petit, F. (1983) Biochem. J. 216, 233-236. 11 Turnbull, D.M., Bone, A.J., Bartlett, K., Koundakjian, P.P. and Sherratt, H.S.A. (1983) Biochem. Pharmacol. 32, 1887-1892. 12 Warter, J.M., Imler, M., Marescaux, C., Chabrier, G., Rumbach, L., Micheletti, G. and Krieger, J. (1983) Eur. J. Pharmacol. 87, 177-182. 13 Warter, J.M., Marescaux, C., Chabrier, G., Rumbach, L., Micheletti, G. and Imler, M. (1984) Rev. Neurol. (Paris) 140, 370-371. 14 Ferrier, B., Martin, M. and Baverel, G. (1988) J. Clin. Chem. Clin. Biochem. 26, 65-67. 15 Baverel, G., Bonnard, M., d'Armagnac de Castanet, E. and Pellet, M. (1978) Kidney Int. 14, 567-575. 16 Baverel, G., Bonnard, M., d'Armagnac de Castanet, E. and Pellet, M. (1979) FEBS Lett. 101,282-286. 17 Krebs, H.A. and Henseleit, K. (1932) Hoppe-Seyler's Z. Physiol. Chem. 210, 33-66. 18 Baverel, G. and Lund, P. (1979) Biochem. J. 184, 599-606. 19 Lemieux, G., Baverel, G., Vinay, P. and Wadoux, P. (1976) Am. J. Physiol. 231 1068-1073. 20 Braunstein, A.E. (1964) Vitam. Horm. (N.Y.) 22, 451-484. 21 Owen, E.E. and Robinson, R.R. (1963) J. Clin. Invest. 42 263-276. 22 Wahren, J. and Felig, P. (1975) Diabetes 24, 730-734. 23 Tizianello, A., De Ferrari, G., Garibotto, G. and Gurreri, G. (1978) Clin. Sci. Mol. Med. 55, 391-397. 24 Tizianello, A., De Ferrari, G., Garibotto, G., Gurreri, G. and Robaudo, C. (1980) J. Clin. Invest. 65, 1162-1173. 25 Tizianello, A. De Ferrari, G., Garibotto, G., Robaudo, C., Acquarone, N. and Ghiggeri, G.M. (1982) J. Clin. Invest. 69, 240-250. 26 Mattenheimer, H. and De Bruin, H. (1964) Enzymol. Biol. Clin. 4, 65-83. 27 Mattenheimer, H., Pollack, V.E. and Muehrcke, R.C. (1970) Nephron 7, 144-154. 28 Turnbull, D.M., Dick, D.J., Wilson, L., Sherratt, H.S.A. and Alberti, K.G.M.M. (1986) J. Neurol. Neurosurg. Psychiatr. 49, 405-410. 29 Goldstein, L. (1966) Am. J. Physiol. 210, 661-666. 30 Fine, A., Scott, J. and Bourke, J. (1972) J. Lab. Clin. Med. 80, 591-597. 31 Schering, B., Reinacher, M. and Schoner, W. (1986) Biochim. Biophys. Acta 881, 62-71. 32 Kvamme, E. and Torgner, I.M. (1974) Biochem. J. 137, 525-530.
