Keith Veitch

In vitro effects of valproate and valproate metabolites on mitochondrial oxidations. Relevance of CoA sequestration to the observed inhibitions

The inhibitory effects of valproate (VPA) and nine of its metabolites on mitochondrial oxidations have been investigated. Valproate, 4-ene-VPE, 2,4-diene-VPA and 2-propylglutaric acid inhibited the rate of oxygen consumption by rat liver mitochondrial fractions with long- and medium-chain fatty acids, glutamate (+/- malate), succinate, alpha-ketoglutarate (+ malate) and pyruvate (+ malate) as substrates. Sequestration of intramitochondrial free CoA by valproate and these three metabolites has been demonstrated and quantified. However, CoA trapping could not account for all the inhibitions observed. 2-ene-VPA and 3-oxo-VPA, metabolites formed during the beta-oxidation of valproate, were not capable of trapping intramitochondrial CoA although they were inhibitors of the beta-oxidation of decanoate, probably by inhibition of the medium-chain acyl-CoA synthetase.

The antianginal agent ranolazine is a weak inhibitor of the respiratory Complex I, but with greater potency in broken or uncoupled than in coupled mitochondria

Ranolazine (RS-43285) has shown antianginal effects in clinical trials and cardiac anti-ischaemic activity in several in vivo and in vitro animal models, but without affecting haemodynamics. Its mechanism is thought to mainly involve a switch in substrate utilisation from fatty acids to glucose to, thus, improve efficiency of O2 use; however, its precise molecular target(s) are unknown. In studies to investigate its action further, using isolated rat heart mitochondria, ranolazine was found to weakly inhibit (pIC50 values > 300 microM) respiration by coupled mitochondria provided with NAD(+)-linked substrates but not with succinate. With broken mitochondrial membranes or submitochondrial particles, ranolazine inhibited NADH but not succinate oxidation and with pIC50 values in the lower range of 3-50 microM. Studies with different electron acceptors and respiratory inhibitors indicated that it inhibits respiratory Complex I at a site between ferricyanide and menadione and ubiquinone-1 reduction (i.e. at a similar locus to rotenone). However, unlike rotenone, ranolazine was an uncompetitive inhibitor with respect to ubiquinone-1. Ranolazine inhibition of Complex I was reversible and occurred also with mitochondria from pig, guinea pig, and human heart, and rat liver. Further studies using rat heart mitochondria in different energisation states (i.e. broken, uncoupled, or coupled) showed a 50-100-fold shift to greater potency of ranolazine in the broken compared to the coupled; with the uncoupled it was about 2-fold less potent than the broken. These shifts in potency were not found with rotenone or amytal. Studies with radiolabelled ranolazine showed that it bound to mitochondrial membranes with greater affinity in the broken compared to the coupled or uncoupled conditions. Rotenone displaced radiolabelled ranolazine from its binding site. This property of ranolazine may play some role in its anti-ischaemic activity.

Cytochrome aa3 depletion is the cause of the deficient mitochondrial respiration induced by chronic valproate administration

Liver mitochondria from rats fed 1% (w/w) valproic acid for 75 days displayed an approximate 30% decrease in coupled respiration rates with substrates entering the respiratory chain at complexes I and II. Uncoupling the respiration from proton-pumping, or measuring the respiration without complex IV removed this inhibition. The treatment induced a loss of activity of cytochrome oxidase consistent with a decrease in the mitochondrial content of cytochrome aa3. The inhibition induced by long lasting administration of valproate is apparently located at the site of the proton-pumping activity of complex IV. Furthermore, the capacity of electron transport through complex IV, being far in excess of that required for normal functioning in coupled mitochondria, seems to be controlled by the coupling to proton-pumping in which cytochrome aa3 appears to play a major role.

Altered acyl-CoA metabolism in riboflavin deficiency.

We have recently described the effects of riboflavin deficiency on the metabolism of dicarboxylic acids (Draye et al. (1988) Eur. J. Biochem. 178, 183-189). As both mitochondria and peroxisomes are thought to be involved, we have examined the activities of various enzymes in these organelles in the livers of riboflavin-deficient rats. Mitochondrial beta-oxidation of fatty acids was severely depressed due to loss of activity of the three fatty acyl-CoA dehydrogenases, whereas there was an enhancement of peroxisomal beta-oxidation due to an increased activity of the FAD-dependent fatty acyl-CoA oxidase, although the activities of other peroxisomal flavoproteins, D-amino acid oxidase and glycolate oxidase, were lowered. Hepatocyte morphometry revealed an increase in the numbers of peroxisomes, indicating a proliferation induced by the deficiency. The mitochondrial acyl-CoA dehydrogenases involved in branched-chain amino acid metabolism were also severely decreased leading to characteristic organic acidurias. There was some loss of activity of the flavin-dependent sections of the electron transport chain (complexes I and II), but these were probably not sufficient to affect normal function in vivo. The specificity of these effects allows the use of the riboflavin-deficient rat as a model for the study of dicarboxylate metabolism.

Influence of chronic administration of valproate on ultrastructure and enzyme content of peroxisomes in rat liver and kidney. Oxidation of valproate by liver peroxisomes.

Chronic administration to rats of the anticonvulsant drug, valproate, induced proliferation of liver peroxisomes and selectively increased the activity of the enzymes involved in beta-oxidation in these organelles. In kidney cortex, only a moderate increase in enzyme activity could be recorded. Valproate (1% w/w in the diet for 25 to 100 days) caused the appearance on electron micrographs of unusual tubular inclusions in the matrix of liver peroxisomes. SDS-PAGE analysis of purified peroxisomal fractions from treated rats demonstrated an increase in the content of five polypeptides; four of which most likely correspond to enzymes of the peroxisomal beta-oxidation. It is suggested that the peroxisomal inclusions correspond to the accumulation of these polypeptides in the matrix of the organelle. An in vivo evaluation of the peroxisomal hydrogen peroxide production suggested that valproate itself or one of its metabolites is substrate for peroxisomal beta-oxidation. This was confirmed by in vitro studies. Activation of valproate or its metabolites by liver acyl-CoA synthetase could be demonstrated, although it was 50 times slower than that of octanoate. This reaction further led to a small, but significant production of H2O2 by the action of peroxisomal acyl-CoA oxidase.

In vitro effects of eight-carbon fatty acids on oxidations in rat liver mitochondria

Sodium valproate, a commonly used anticonvulsant agent, is a simple branched-chain fatty acid which interferes with beta-oxidation and ammonia metabolism in most patients, with hepatotoxic consequences in some cases. Rat liver mitochondria incubated with valproate displayed time-dependent inhibitions of state 3 oxidation rates with all the substrates tested, but most markedly with glutamate, pyruvate, alpha-ketoglutarate and acylcarnitines (Ki = 125 microM with glutamate and palmitoylcarnitine, and 24 microM with pyruvate). The inhibition of glutamate appeared to be specifically directed against the glutamate dehydrogenase pathway of this oxidation. Valproate was less effective when added to uncoupled mitochondria, suggesting the formation of an inhibitory species by an ATP-dependent mechanism. Mitochondria from clofibrate-treated rats were less sensitive to valproate inhibition. Neither fasting nor the presence of 1 mM L-carnitine affected the inhibition of beta-oxidation. The branched-chain isomer, 2-ethylhexanoic acid, had similar effects to valproate, but the straight-chain octanoic acid was totally different in its spectrum of actions on mitochondria. The data support the theory that valproate may inhibit by sequestration of CoA as valproyl-CoA, but also suggest that there are other mechanisms responsible for some of the inhibitions. Furthermore, it argued that while mitochondrial respiration is decreased, valproate is not an inhibitor of oxidative phosphorylation per se.

Analysis by fast-atom bombardment tandem mass spectrometry of phosphatidylcholine isolated from heart mitochondrial fractions: Evidence of incorporation of monohydroxylated fatty acyl moieties

Phosphatidylcholine (PC) is one of the main phospholipids present in mitochondrial membranes. According to current knowledge, the predominant fatty acyl moieties in this phospholipid are 16, 18, 20, or 22 carbon atoms long with chains that contain only carbon and hydrogen atoms. We have conducted a detailed analysis of the fatty acid substituents of the phospholipids present in mitochondrial fractions by using fast-atom bombardment tandem PC extracted from mitochondrial fractions of rat heart. The structure of one of these monohydroxylated fatty acids has been elucidated and corresponded to 12-hydroxy 9-octadecenoic acid. Indications that concern the structure of the five other monohydroxylated fatty acids are presented. These monohydroxylated fatty acyl groups are preferentially associated in the PC molecule with C-18 and C-20 fatty acyl moieties. We present arguments to suggest that the formation of these compounds is probably not due to a free-radical initiated mechanism. The potential implication of these monohydroxylated fatty acids in several physiological functions is suggested by the fact that free hydroxylated fatty acids that are identical or closely related to those found in the mitochondrial fractions possess various biological activities.

The Metabolism of Dicarboxylic Acids in Rat Liver

In mammals, the beta-oxidation of fatty acids was thought to be the prerogative of mitochondria until the demonstration by Lazarow and de Duve (1) that rat liver peroxisomes also contain beta-oxidizing enzymes. The distinction between mitochondrial and peroxisomal fatty acid oxidation includes the specialisation of the latter system in the shortening of very long-chain acids (2–4), the oxidation of polyunsaturated fatty acids (5–7), cleavage of the cholesterol side chain (8–10) and the catabolism of prostaglandins (11,12). Conversely to beta-oxidation, the omega-oxidation of fatty acids does not require coenzyme A (13). It consists of the omega-hydroxylation of fatty acids followed by the oxidation of the resulting alcohol group to a carboxyl group (14–16). In vivo studies disclosed that omega-oxidation of long-chain monocarboxylic acids and beta-oxidation of the corresponding dicarboxylic acids are the basis of medium-chain dicarboxylicaciduria (17–18). We have previously demonstrated that rat liver contains a dicarboxylyl-CoA synthetase active on substrates having more than five carbon atoms (19). This paper deals with the respective contributions of peroxisomes and mitochondria to the catabolism of long-chain saturated dicarboxylyl-CoA esters. Arguments for the involvement of both organelles in the breakdown of long-chain dicarboxylates are presented.