**4. Discussion**

In this work, cerebellar metabolism of amino acids and its dependence on the TCA-cycle-limiting OGDHC are shown to be affected by physiological settings (pregnancy) and metabolic stress (acute hypobaric hypoxia). Pregnancy increases cerebellar levels of glutamate and tryptophan (Table 1) and decreases overall interdependence of the studied components of amino acid metabolism, compared to their interdependence in non-pregnan<sup>t</sup> rats (Table 4).

Hormonal changes during pregnancy are known to coordinate a broad range of physiological adaptations, from the supply of nutrients and oxygen for the fetus growth in utero to specific patterns of parental behavior [32,33]. Changed metabolism of amino acids in the brain of the pregnan<sup>t</sup> females may be involved in these adaptations, as many of amino acids are neurotransmitters or their precursors. The increases in cerebellar glutamate and tryptophan in pregnan<sup>t</sup> rats, observed in this work, are consistent with independent studies on the pregnancy-imposed changes in glutamatergic and serotonergic signaling. In fact, pregnancy is known to change expression of glutamate receptors to address specific physiological challenges faced by pregnan<sup>t</sup> females [34,35]. Our data on the pregnancy-increased cerebellar content of tryptophan are in line with the activation of serotonergic signaling observed during pregnancy, because tryptophan is the serotonin precursor [36]. The interaction between the levels of serotonin and estrogens also underlies sex-dimorphic prevalence of the serotonin-linked diseases, including migraine, depression, eating disorders and pregnancy-associated pathologies [36,37].

Our previous studies pointed to certain relationship between the amino acid levels in the blood plasma and brain [23], which may be used for translation of findings on the brain samples in animal models to human studies. For instance, increased content of tryptophan in cerebellum of pregnan<sup>t</sup> rats, shown in our study (Table 1) corresponds to the findings in humans, which indicate that in maternal plasma, tryptophan catabolites and related compounds change in pregnancy [38,39]. This study suggests the neurotransmitters-dependent adaptation of maternal body to the fetus growth. As considered above, such adaptation is obvious also from our findings (Table 1). Biomarker significance of the amino acid levels in human plasma and urine has also been explored in other studies [40–42]. Their findings point to the potential diagnostic significance of the changes in multivariate metabolic profiles, including amino acids, for prediction of gestational diabetes [42]. Besides, the pregnancy-dependent changes in the content of amino acids and their metabolites in plasma and urine point to variation of these parameters, dependent on the increased amino acid demands for the growing fetus [40,41].

The pregnancy-induced changes in cerebellar pool of amino acids (Table 1) are associated with different responses of cerebellar amino acid metabolism to hypoxia in the pregnan<sup>t</sup> and non-pregnan<sup>t</sup> rats (Tables 2–4). Strong negative correlations between the amino acid levels and OGDHC activity, which are a hallmark of cerebellar metabolism after hypoxia of non-pregnan<sup>t</sup> rats, are not induced by hypoxia in the pregnan<sup>t</sup> rats (Tables 3 and 4). Simultaneously, hypoxia significantly increases average levels of cerebellar amino acids in the non-pregnan<sup>t</sup> rats only (Table 2). Because the observed biochemical changes in the cerebellar metabolism of amino acids are associated with a higher resistance to hypoxia in the non-pregnan<sup>t</sup> vs. pregnan<sup>t</sup> rats (Figure 1), the metabolic rearrangemen<sup>t</sup> in the cerebellum of non-pregnan<sup>t</sup> rats is of adaptive significance. In particular, increased flux of the amino acids degraded through OGDHC (Glu, Gln, Arg, His, Pro) and of the branched-chain amino acids Val and Ile may generate succinyl-CoA for the substrate level phosphorylation in mitochondria. Generation of ATP at the expense of succinyl-CoA may help overcoming energy deficits upon hypoxia which impairs oxidative phosphorylation. Increased degradation of amino acids in the TCA cycle whose flux is limited by OGDHC, is in good accordance with the negative correlations between the OGDHC activity and the levels of cerebellar amino acids after hypoxia, in contrast to normal metabolism (Table 3). It is worth noting in this regard that hypoxic tolerance is associated with the mTOR-dependent autophagy [43]. Increased autophagy may generate the higher amino acid levels after hypoxia (Table 2) to use them for the substrate-level ADP phosphorylation in the hypoxic brain of non-pregnan<sup>t</sup> rats. Autophagy is also coupled to pro-survival function of mitochondrial fission under energy stress [44–47]. However, the homeostatic and pro-survival functions of mTOR, autophagy and mitochondrial fission are highly conditional, with overactivation of these processes also mediating the brain damage by hypoxia [48–50].

Different levels of activation of mTOR, autophagy and mitochondrial fission may be required for their pro-survival and death-inducing outcomes in different physiological states. In this regard, no increase in the brain amino acid levels by hypoxic exposure of pregnan<sup>t</sup> rats (Table 2) was due to a higher damaging potential of the autophagy stimulation in this physiological state. No adaptation to hypoxia by increased degradation of amino acids through OGDHC in pregnan<sup>t</sup> rats is also evident from the hypoxia-induced decrease in the interdependences between the levels of OGDHC activity and/or amino acids, whereas in the non-pregnan<sup>t</sup> rats hypoxia increases these interdependences (Table 4). The different action of hypoxia in the two physiological states is obvious from comparison of the hypoxia-induced changes in the summarized and average correlation coefficients or in the number of significant correlations in the pregnan<sup>t</sup> and non-pregnan<sup>t</sup> rats (Table 4). In view of the increased sensitivity of pregnan<sup>t</sup> vs. non-pregnan<sup>t</sup> rats to hypoxia (Figure 1), the stability of the biochemical parameters in the cerebellum of the pregnan<sup>t</sup> rats exposed to hypoxia (Table 2) along with the absence of negative correlations between the OGDHC activity and levels of amino acids (Tables 3 and 4) manifest limitations of the hypoxic adaptation through increased degradation of amino acids in the pregnan<sup>t</sup> vs. non-pregnan<sup>t</sup> rats. These findings indicate that the pregnancy-imposed changes in the cerebellar amino acid pool (Table 2) and metabolism (Tables 3 and 4) are associated with decreased stress adaptability, in line with other studies showing decreased perception of stress in pregnan<sup>t</sup> females [51–53].

It is worth noting that the response of the metabolic network of pregnan<sup>t</sup> rats to hypoxia, undetectable from the average levels of amino acids (Table 2), is evident from the correlation analysis (Table 4). Hence, average levels of cerebellar amino acids are not as sensitive indicators of metabolic

changes, as the cumulative parameters characterizing the interdependence of components of the corresponding metabolic network.

It has been noted previously that correlating metabolites may not only be of diagnostic significance [23], but also help deciphering the ye<sup>t</sup> unknown or poorly characterized synthetic and regulatory pathways [20]. In this regard, the pregnancy-induced changes in metabolic correlations of cerebellar lysine are of interest, because gestational diabetes is associated with plasma levels of lysine and tyrosine [54], both of them correlating to each other much stronger in cerebellum of the pregnan<sup>t</sup> than non-pregnan<sup>t</sup> rats (Table 3). Besides, lysine catabolism is related to biosynthesis of homoarginine which is elevated in normal pregnancy [13], with some studies linking its elevation to pregnancy disorders, including preeclampsia [55]. Physiological manifestations of pre-eclampsia and eclampsia, associated with nearly one-tenth of all maternal deaths [56], involve changed cerebral hemodynamics and hypertensive encephalopathy [57–59], potentially linked to impaired signaling by homoarginine, known as a predictor of cardiovascular risk and mortality [3]. Alternatively, homoarginine is synthesized from arginine and glycine. Because these amino acids are highly correlated independent of pregnancy and/or hypoxia (Table 3), metabolism of cerebellar lysine appears to be a more likely contributor to specific adaptations imposed by pregnancy. This is further supported by the fact that lysine is known as an antagonist of a serotonin receptor [60], with serotonergic signaling increased in pregnancy [36,37] and depressed upon increased synthesis of homoarginine from lysine [61].
