Abstracts

Rationale: Glutamine is critical for numerous brain functions such as ammonia detoxification and synthesis of the neurotransmitters glutamate and gamma-aminobutyric acid (GABA).  Glutamine synthetase (GS) is the only enzyme capable of forming significant amounts of glutamine in mammals, and in patients with drug-resistant mesial temporal lobe epilepsy (MTLE) the enzyme is severely deficient in the seizure focus of the brain. However, glutamine levels are not significantly different between GS-deficient and -intact regions of the brain.  To explore possible mechanisms of the “glutamine paradox,” we investigated the relationship among several interconnected metabolites in the brain in patients with focal epilepsies, including MTLE.     Methods: Twenty-eight patients with drug-resistant focal epilepsies of different types were implanted with intracranial depth electrodes for seizure localization. A microdialysis catheter was inserted into the lumen of the depth electrodes. Sterile artificial cerebrospinal fluid was perfused through microdialysis catheter, and the efflux was collected in 1- or 2-h aliquots. All depth electrodes were classified into epileptic (i.e. seizure onset or propagation) or non-epileptic by an expert team. Glutamine, leucine, isoleucine, valine, and glutamate were quantified in the dialysis samples using liquid chromatography-tandem mass spectrometry. A linear mixed effect model (R software, version 3.6.0, lme4 package) was employed to assess the interaction of glutamine with the other amino acids. A chi-squared test was implemented to compare different models of both epileptic and non-epileptic regions. Results: All probes (n= 67) were classified as either epileptic (n= 40) or non-epileptic (n= 27). In the non-epileptic brain regions, glutamine interacted significantly only with glutamate (χ2(1)=6.4, p=0.01), but not with leucine (χ2(1)=1.4, p=0.2),valine (χ2(1)=1.2, p=0.2), or isoleucine (χ2(1)=0.7, p=0.3). In the epileptic regions, glutamine did not interact significantly with glutamate (χ2(1)=0.13, p=0.7). However, glutamine interacted significantly with leucine (χ2(1)=7.7, p=0.005), valine (χ2(1)=7.5, p=0.006), and isoleucine (χ2(1)=4.9, p=0.02).  Conclusions: Because glutamine is a synthetic precursor for glutamate, the correlation between the two amino acids in non-epileptogenic brain regions is expected. However, the remarkably different metabolic relationships in epileptogenic brain regions suggest that the branched-chain amino acids leucine, isoleucine and valine play a novel role in maintaining brain glutamine levels when GS is deficient. Funding: Swebilius Family Trust and National Institutes of Health