Our analytical approach using high-resolution in vitro NMR spectroscopy study confirmed that glutamate, which acts as an excitatory neurotransmitter and important metabolic and nitrogen metabolite, is highly concentrated in all intact brain regions. We found the level of glutamate as about twice that of glutamine in rats’ brain substructures, and significant differences were indicated also between the hippocampus, cerebellum and cortex (Table 1). Inhibitory neurotransmitter GABA (Cooper and Jeitner 2016) showed around seven to eight times smaller levels in brain substructures compared to glutamate, and, likewise, the ratios significantly differed among individual brain regions. The recent literature data provides similar results on methanol/water extract from the brain hippocampal homogenate using an alternative high-performance/pressure liquid chromatography (HPLC) (Zieminska et al. 2018), which indicates the applicability of both methodologies (in vitro NMR as well as HPLC) for determination of neurotransmitters levels in brain tissue homogenates. It should be also mentioned that the method of calculation can slightly affect the results, where the median of the ratios (as used in our work) does not have to be the same as the ratio of the medians, or the ratio of the average values.

The variability in metabolite abundance among brain structures could be linked to the different densities of respective neurotransmitter receptors, specific metabolic enzymes and the general role of metabolites in the neuronal and glial cell types within the brain structures. In the previous work, Duarte et al. showed a relatively strong correlation between regional glutamate and NAA (which is believed to be localized exclusively in neurons (Moffett et al. 2007) levels across mouse age and gender, presenting NAA and glutamate as relevant markers for neuronal density or function (Duarte et al. 2012). We calculated ratios of both neurotransmitter levels and glutamine to NAA. In our study, the ratios Glu/NAA, GABA/NAA, as well as Gln/NAA increased in line: cortex, hippocampus, and cerebellum. If we use data from mouse neuronal density, the order in which the neural density increases (Keller et al. 2018) decreased the values of ratios of metabolites with NAA value in the denominator (Table 2; Fig. 3). However, from this comparison, we can conclude rather the relation of neural density to the level of NAA than to the amino acids (neurotransmitters) levels.

The fact that the ratios between metabolites Glu/Gln and GABA/Gln cannot be linked with the neural densities, (when compared with the data from mice (Keller et al. 2018), is based likely on the multiple metabolic functions of glutamate, GABA and glutamine among neurons, astrocytes and microglia (microglia represents 5–12% of all glial cells (Lawson et al. 1990) in the brain substructures. Glutamate is stored mainly in neuronal vesicles prepared for its neurotransmitter function, however, a significant proportion of glutamate is oxidatively metabolized by conversion to oxoglutarate to the tricarboxylic acid cycle (TCA) in neurons (Hassel 2001), as well as glial cells (Gondáš et al. 2023), and also used as a substrate of glutamine synthetase to detoxify ammonia and produce glutamine (Schousboe et al. 2014). The main source of brain parenchymal GABA are GABAergic neurons, however, a substantial GABA synthesis also occurs in glial cells (Angulo et al. 2008; Héja et al. 2012; Serrano-Regal et al. 2020). Maintaining the availability of both neurotransmitters in brain tissues, their mutual conversion with glutamine is secured by constantly ongoing Glu-GABA/Gln shuttle between neurons and astrocytes, a pathway resulting in about 85% of the glutamine synthesis in rat cortex (Rothman et al. 2011). It was presented that approximately 30% of total glutamate entering astrocytes is subjected to oxidative processes and 70% contributes to the Glu-Gln cycle, which indicates that biosynthesis of glutamate continuously occurs in the neurons, supplying nerve terminals with sufficient neurotransmitters (Nieoullon 2009). The carbon skeleton of glutamate via entering the Krebs cycle is converted to malate, which is ultimately turned to pyruvate by the malic enzyme (ME). As an alternative, pyruvate kinase (PK) and phosphoenolpyruvate carboxykinase (PEPCK) work together to convert oxaloacetate (OAA) to pyruvate. Interestingly, for the complete oxidation of glutamate, all enzymes must operate in the direction to produce pyruvate, which is a substrate of pyruvate dehydrogenase that re/enters the Krebs cycle and is entirely oxidized to CO2.

The distribution of Glu and GABA over brain regions is surely linked with a proportion between excitatory glutamatergic and inhibitory GABAergic neurons. In intact rat brains, the cerebellum showed almost 20% higher Glu/GABA ratio compared to the cortex, and the hippocampus manifested almost 30% higher proportion of GABA to glutamine compared to the cerebellum and about 10% higher Glu/Gln ratio than the cortex (Table 1). Quantitative immunohistochemistry studies and genetic markers have estimated the ratio of approximately 4 to 1 of glutamatergic neurons for every GABAergic neurons in the cortex (Swanson and Maffei 2019), and it is even generally accepted that 20%of all neurons in rodents’ cortex are GABAergic interneurons (Sahara et al. 2012). The literature provides also data about the hippocampal circuit of GABAergic inhibitory interneurons which represents 10–15% of the total neuronal population (Pelkey et al. 2017). These data, showing a higher portion of glutamatergic to GABAergic neurons in the hippocampus than in the cortex could be indicative for the explanation of higher Glu/GABA levels in the hippocampus than in the cortex. In general, the determination of precise numbers of GABAergic and glutamatergic neurons in brain structures faces generally several methodological problems due to the complexity of the brain’s regional cellular and layer composition. Moreover, this number is influenced by genetic factors, developmental changes and as result of time-dependent neural plasticity changes. As detected in mice and also in other species, the levels of glutamate and GABA in the hippocampus and cortex show an age-dependent decline (Duarte et al. 2012). Unfortunately, due to methodological constraints, there is a persisting lack of detailed data about the exact quantities of glutamate and GABA pools in brain substructures, their precise cellular localization and the contributions of non-neuronal cells to total brain regional glutamate and GABA levels. The quantitative data may shed deeper light on aspects such as the availability of glutamine as a precursor for neurotransmitters, its role as a metabolic substrate, and nitrogen carrier or its involvement in the activation/proliferation of glial cells dependent on glutamine. Additionally, knowing the exact level of glutamine could be valuable in understanding the dynamics of ammonia transfer within the glutamate, GABA, and glutamine cycle and other amino acid metabolism.

Our work provides the first comprehensive study to identify and measure levels of the most important amino acids acting as excitatory and inhibitory neurotransmitters in the brain region parenchyma. As presented by Pulsinelly et al., who introduced the animal model of global cerebral ischemia also used in this study (Pulsinelli and Buchan 1988), the regional cerebral blood flow, as well as glucose metabolism in the brain regions, are differently affected by severe but transient global ischemia (Pulsinelli et al. 1983). These are the first original data which utilizes a reliable and novel methodological approach for the quantification of amino acid levels in brain homogenates.

Brain tissue extracts in 3 h reperfusion after the ischemic event were characterized by a generally decreased glutamate/glutamine ratio, which was most pronounced in the hippocampus–of about 40%, then in the cortex of about 35% and the smallest change was observed in the cerebellum, of about 20% (Fig. 3; Table 2). This is in line with the results of blood flow rate measurements in the 4VO model of global ischemia (Pulsinelli et al. 1983), where brain regions exposed to severe ischemia such as the hippocampus and cortex showed greater metabolic alterations than the cerebellum. Observed changes are paralleled with the excessive glutamate release from neurons and glial cells, (Belov Kirdajova et al. 2020) which is a cell death mechanism known as glutamate excitotoxicity, and is considered one of the major causes of damage after the ischemic event in the brain (Choi 2020). Brain glutamine synthetase, localized in all cells of the astroglial family (Anlauf and Derouiche 2013), increases after the ischemic event (Petito et al. 1992), and extracellular glutamate is converted to glutamine. As our results indicate, within three days of reperfusion (IR72), the ratio glutamate/glutamine is able to almost recover to the values observed for non-ischemic tissue (Fig. 4). Moreover, it is important to note, that both the ratio as well as the relative aminoacids pools themselves are almost recovered to the preischemic level, as we have shown in the previous work on the hippocampus (Baranovicova et al. 2022) and cortex (Baranovicova et al. 2021).

The results by Globus et al. (Globus et al. 1988) suggest that not the quantitative release of glutamate by itself accounts for the pattern of selective vulnerability of the organs, but, instead, the imbalance between excitation and inhibition may play a major role in the mediation of neuronal damage (Globus et al. 1991). In fact, we observed a decrease in GABA/Gln ratio for about 60% 3 h after ischemia (3IR) in the hippocampus, then in the cortex of about 50%, and the lowest change was detected in the cerebellum, of about 25%, which correlates with the proportion of observed compromised blood flow rate in 4VO model (Pulsinelli et al. 1983). These ratios did not fully recover in 72 h reperfusion after the ischemic event (72IR), remaining at the portion of 83% for the hippocampus, 75% for the cortex and 92% for cerebellum related to those found in the intact brain. Throughout cerebral I/R, the inhibitory neurotransmitter GABA undergoes complex pathological variations (Chen et al. 2019). GABA is, similarly to glutamate, after cerebral ischemia, released from neurons and also microglia, and either reabsorbed by neurons or uptaken by surrounding glial cells for further metabolism (Fig. 5). To compare the metabolic prioritization in restoring levels of glutamate and GABA after ischemia, we calculated the Glu/GABA ratio, which was increased over the first 72 h reperfusion time in all ischemic brain substructures (Fig. 4). Not only in tissue homogenates, as used in our study, including a contribution of both, extracellular and intracellular components, but increased Glu/GABA ratio was also observed by HPLC when monitoring neurotransmitters in interstitial fluid by analyzing microdialysate by Zeng et al., in rats after middle cerebral artery occlusion (MCAO) in the first 130 minutes after onset of ischemia (Zeng et al. 2007) as well as by Wang et al. in 90 min reperfusion after MCAO in mice (Wang et al. 2001). The results of this study and results of our previous works (Baranovicova et al. 2021, 2022) indicate the recovery of the relative glutamate, but not GABA (Baranovicova et al. 2022) pools in brain tissues in 72 h reperfusion. The Glu/GABA ratio in hippocampal tissue in 72 reperfusion time found in this study suggests that the ischemic balance between glutamate and GABA will be established later than 72 h after an ischemic event.

Simplified turnover of glutamate/GABA and glutamine, enhanced after ischemia-induced excessive neurotransmitter release, with proposed postischemically accelerated metabolic oxidation in the reperfusion/recirculation period

It is interesting to note that despite the postischemic damage or even death of neurons, a large part of the decreased glutamate amount seen after the ischemic event in the brain is replaced in the 72-hour reperfusion period. This occurs relatively quickly in comparison to the re-synthesis of the neuron-characteristic metabolite N-acetyl aspartate (NAA), considered to be an indicator of neural health and fitness (Demougeot et al. 2004). NAA levels were observed to elevate only 8 days after ischemia and continued to increase until 30 days (Demougeot et al. 2004). It can be assumed that the intact, not damaged, neurons and glia would efficiently synthesize glutamate to substitute its deprivation due to an ischemic event and therefore compensate for the altered synthetic function of damaged neural cells. On the other hand, not fully recovered levels of GABA three days after ischemia, may be the reason for the persistently limited feasibility of reestablishing GABAergic transmission after ischemia. Reduced GABA transmission with increased glutamate signals occurring in cerebral I/R injury was documented in previous studies (Chen et al. 2019), which also correlates with the increased glutamate/GABA ratio in affected brain regions after ischemia. In this context, an enhancement of the GABA actions and decreased glutamate can be a potential preventive strategy for cerebral I/R injury, which may be achieved by different enhancers of GABA transmission that could partly restore the balance in GABA and glutamate transmissions (Chen et al. 2019). GABA also affects functions of astroglial and microglial cells as well as peripheral immune cell populations accumulating in the ischemic territory and brain regions remote to the lesion (Michalettos and Ruscher 2022), however, the complex interactions between inflammatory cascades and neuronal functions are still not comprehensively understood.

A crucial role in cell defence against oxidative stress and the disruption of Zn2+ homeostasis plays glutathione, which is the most abundant low molecular weight thiol compound in the brain. Glutathione is an antioxidant with cellular protective functions, including reactive oxygen species scavenging in the brain, and it was found to preserve the disruption of BBB after ischemic injury and improve the survival of brain endothelial cells (Song et al. 2015). Decreased levels of glutathione after transient ischemia in the hippocampus were described previously (Higashi et al. 2021), and similarly, we observed decreased intensity of glutathione peaks in NMR spectra in brain tissue homogenates after cerebral ischemia, however, the quantitative evaluation of the NMR signals were not reliable. As just glutamate is an essential part of glutathione synthesis, the disturbances in glutamate level after cerebral ischemia may unfavourably affect glutathione synthesis and therewith also suppress its function of redox protection.

In addition to complex damaging glutamate excitotoxic mechanisms, excitotoxic glutamate inhibits mTOR signalling and causes elevated neuronal insulin resistance (Pomytkin et al. 2019), making neurons more vulnerable to metabolic stress induced by the ischemic event. On the other hand, this would be a sign of metabolic demand alternations to use alternative energy substrates to preserve cell viability. Under physiological conditions, the use of glutamate and GABA by astrocytes and glutamine by neurons as useful and actual energy substrates is not essential. However, after ischemia, when glucose oxidative metabolism in the brain is compromised, the metabolic conversion of these amino acids as alternative energy substrates was observed to increase (Pascual et al. 1998). It seems that cells in/after oxygen restriction might alternatively use molecules with primary other functions such as neurotransmitters and/or alternative energy substrates. As a consequence, this can further shift the balance between glutamate, GABA and glutamine in the affected brain parenchyma.

As documented above, the studies on tissue homogenates via in vitro metabolomic methods cannot reproducibly discriminate the distribution of glutamine among neurons, astrocytes and microglia, verify its presence in glial cells reuptake by neurons for neurotransmitter re-synthesis. When translating our study performed on animal models of the global cerebral ischemia, it should be also considered that the extent of the metabolic response may change depending on ischemia severity, as well as the postischemic profiles of glutamate and GABA vary in core versus peripheral zones (Wang et al. 2001).

The alteration in metabolic response in the animals with induced ischemic tolerance/neuro-protection was already shown by Dave et al., which found a higher GABA extracellular concentration within 80 min reperfusion in IPC animals, as a consequence of proposed neuroprotective preconditioning (Dave et al. 2005). In our work, animals subjected to ischemic preconditioning before ischemia showed, that all tested parameters manifested smaller postischemic responses compared to animals with single severe cerebral ischemia (Fig. 4). This weaker metabolic response in the levels of Glu/GABA/Gln in IPC animals provides further evidence for documenting the protective effect of sub-lethal ischemia to ischemic injury/damage. In this context, we would like to stress an exceeding Glu/Gln ratio compared to control animals at prolonged 72 h reperfusion time in IPC animals, which suggests accelerated glutamate metabolism recovery. The higher tissue glutamate level supports both the glutamatergic neurotransmission and also GABA resynthesis and other processes such as e.g. immunomodulatory actions of GABA (Michalettos and Ruscher 2022), and it could be considered as one of the important metabolic features in postischemic recovery.