Circadian Alterations in Brain Metabolism Linked to Cognitive Deficits During Hepatic Ischemia-Reperfusion Injury Using [1H-13C]-NMR Metabolomics

Male C57BL/6 mice, aged 6–8 weeks and weighing 20 ± 3 g, were housed at the Experimental Animal Center of Tongji Hospital, Huazhong University of Science and Technology. The housing environment was maintained at a temperature of 23 ± 2 °C and a humidity of 55 ± 10%. Mice were kept on a standard 12-h light/dark cycle (lights on at 8:00 AM, lights off at 8:00 PM) with ad libitum access to food and water. Hepatic ischemia-reperfusion injury surgery was performed at Zeitgeber Time (ZT) 0 (8:00 AM) or ZT12 (8:00 PM). All mice were obtained from Hubei Bent Biological Technology Co., Ltd, Wuhan, China. All experimental protocols were approved by the Institutional Animal Care and Use Committee at Tongji Hospital, Huazhong University of Science and Technology. The experiment was reported according to the ARRIVE guidelines.

Sample size estimation was conducted using the online SPSSAU software (Version 24.0) (https://spssau.com/indexs.html↗, accessed on 4 November 2024) by selecting the Power analysis for two-factor ANOVA. The calculations were based on an alpha value of 0.05 and a 1-beta value of 0.8. The estimated sample size was 20.673. Considering the 3R principles and the HIRI modeling mortality rate, the initial total sample size for mice was n = 26, leading to a final total sample size of n = 52 for the two infusion substances. Mice infused with [2-¹³C]-acetate were divided into groups as follows: ZT0 Control (CTR) group (n = 6), ZT12 CTR group (n = 6), ZT0 HIRI group (n = 7), and ZT12 HIRI group (n = 7). Similarly, mice infused with [1-¹³C]-glucose were divided into the following groups: ZT0 CTR group (n = 6), ZT12 CTR group (n = 6), ZT0 HIRI group (n = 7), and ZT12 HIRI group (n = 7). The acclimated mice were numbered from lightest to heaviest, and random assignment to the four groups was conducted using a random number table: ZT0 CTR, ZT0 HIRI, ZT12 CTR, and ZT12 HIRI groups. At ZT0 and ZT12, partial hepatic blood flow occlusion was performed on the HIRI group mice to establish the model, while the CTR groups underwent sham operations involving only incision and suture at the same time. All mice subjected to surgery were tested behaviorally on the third day post-operation (72 h of reperfusion) [4,15]. After testing, glucose and acetate were administered, and hippocampal samples were collected for NMR analysis. Except for the technician responsible for HIRI modeling, the researchers conducting behavioral testing, glucose and acetate administration, tissue collection, and NMR data analysis were blinded to group allocations. Post-HIRI modeling, one mouse from the [1-¹³C]-glucose ZT0 HIRI group and two mice from the ZT12 HIRI group died. Additionally, one mouse from the [2-¹³C]-acetate ZT12 HIRI group died. During NMR analysis, due to environmental technical issues, one sample from the [2-¹³C]-acetate ZT12 CTR group and one sample from the ZT0 HIRI group were lost. The experimental flow is illustrated in Figure 1.

The HIRI model was established using partial ischemia-reperfusion (involving the left and middle lobes of the liver, with approximately 70% of the total liver area subjected to ischemia) [22,23]. After a 7-day acclimatization period, male C57BL/6 mice underwent surgical procedures. The ZT0 Control and ZT0 HIRI groups were modeled at 8:00 AM, while the ZT12 Control and ZT12 HIRI groups were operated on at 8:00 PM. All mice undergoing modeling were anesthetized preoperatively with 1% sodium pentobarbital administered via intraperitoneal injection (100 mg/kg). Once the mice reached an adequate surgical depth of anesthesia, their limbs were secured to the surgical platform, and the abdominal incision area was thoroughly disinfected. The surgical incision was approximately 2 cm long, initiated 0.5 cm below the xiphoid process, and meticulously opened layer by layer to access the abdominal cavity. Using a retractor to adequately expose the liver, a cotton swab moistened with saline was used to lift the liver, freeing the left lobe, middle lobe, and associated blood vessels while keeping the remaining part of the liver within the abdominal cavity. Micro hemostatic clamps were applied to occlude the respective hepatic artery and vein, and within 5 s, the ischemic liver area changed color from bright red to yellowish-tan, while the liver in the abdominal cavity remained bright red, confirming successful clamping. Sterile saline-soaked gauze was placed over the abdominal incision to ensure tissue hydration, and the mice were placed on a warming blanket to prevent hypothermia. After 90 min, the hemostatic clamps were released, and the ischemic liver rapidly regained its normal color, indicating successful reperfusion. The abdominal cavity was then closed layer by layer with sutures, the wound site was disinfected again, and lidocaine gel was applied to alleviate post-surgical pain. Surgical maneuvers were performed gently to avoid damaging essential abdominal organs and blood vessels. Mice in the CTR group underwent only incision and suture procedures, with postoperative care identical to that of the HIRI group.

Before the experiment, mice were placed in a quiet, dark behavioral laboratory for a 2-h acclimation period. The Y-maze apparatus consists of three arms (dimensions: 30 × 8 × 15 cm each) arranged at 120° angles to each other. Distinctive patterns are affixed at the end of each arm to help mice memorize spatial locations. The experiment comprises two phases, the learning phase and the testing phase, each lasting five minutes, with a 2-h interval between them. In the learning phase, one arm of the maze (the novel arm) was randomly closed before the experiment. A mouse was placed in one of the other arms (the starting arm) facing the experimenter to prevent pre-exposure to the Y-maze from above. During the learning phase, the mouse was allowed to explore freely between the starting arm and the other open arm for 10 min. Prior to the testing phase, the novel arm was opened, and the mouse was placed back in the Y-maze via the starting arm, allowing it to explore all three arms freely. The apparatus was thoroughly cleaned with medical alcohol between phases to eliminate odor interference. Throughout the entire experiment, data were recorded and analyzed using ANYMAZE software. We recorded the time spent exploring the novel arm to evaluate the mouse’s spatial memory performance [16,27].

On the day of the experiment, mice were weighed and then immobilized using a restrainer. A PE-10 catheter with a metal needle tip was inserted into the tail vein of the mouse. Once blood was successfully aspirated into the syringe to confirm patency, the catheter was secured with tape. The syringe end of the PE-10 catheter was connected to a Fusion100 micro-infusion pump, and [1-¹³C]-glucose (enriched at 99%, 0.75 M) was administered intravenously at a volume of 2.2 mL/kg over 2 min. After the infusion was complete, the needle was kept in place for 1 min before being removed, and the mouse was returned to a clean housing cage to move freely for 30 min. Following the same infusion process, [2-¹³C]-acetate (enriched at 99%,1.5 M) was administered at a volume of 3.6 mL/kg [17,28]. Afterward, the mice were deeply anesthetized with 2% isoflurane, followed by swift decapitation. The head was then placed in a microwave oven at 100% power for 14 s. The hippocampus was dissected, weighed, and stored at −80 °C for further analysis [29].

The method for tissue extraction replicated that of our earlier research [30]. In brief, to the frozen sample, HCl/methanol (0.1 mol/L, 100 μL) was introduced and the mixture was homogenized for 1.5 min at 20 Hz using a Tissuelyser II (Qiagen, Hilden, Germany). Subsequently, ice-cold 60% ethanol (800 μL) was added, and the homogenization process was repeated, followed by centrifugation at 14,000× g for 10 min. The supernatant was collected. This extraction procedure was performed twice more with 800 μL of 60% ethanol to ensure complete metabolite recovery from the residue. All collected supernatants were then dried using a centrifugal evaporator (Thermo Scientific 2010, Erlangen, Germany) and a freeze vacuum dryer (Thermo Scientific). The resultant dried material was stored for subsequent NMR spectroscopy analysis.

For NMR analysis, the dried sample was reconstituted in 60 μL of D2O (which included an internal standard, 3-(trimethylsilyl) propionic-2, 2, 3, 3-d4 acid sodium salt (TSP, 120 mg/L; 269913-1G, Sigma-Aldrich, Shanghai, China)) and 540 μL of phosphate buffer (pH 7.2). The solution was vigorously vortexed and then centrifuged at 14,000× g for 15 min. The clear supernatant was carefully extracted and placed into an NMR tube for analysis.

The samples under investigation were analyzed using a Bruker Avance III 600 MHz NMR spectrometer (operating at 298 K) outfitted with an inverse cryoprobe (Bruker BioSpin, Ettlingen, Germany) [31]. Spectra collection was performed utilizing a conventional Watergate pulse sequence. Acquisition parameters for each sample were as follows: p1 (90° pulse length), 8.35 μs; total scans, 256; spectral bandwidth, 20 ppm; pre-scans (dummy scans), 8; data points for free induction decay, 32 K.

All ¹H-NMR spectra underwent processing and examination using TopSpin software (Version 2.1, Bruker BioSpin) alongside a proprietary software, NMRSpec [32]. Initially, adjustments for phase correction and baseline distortion were manually conducted in TopSpin [31,33]. Subsequently, these adjusted spectra were transferred to NMRSpec for procedures such as aligning spectra, extracting peaks, integrating spectra, and integrating peaks associated with specific chemicals. This software has been applied in numerous metabolomics investigations.

The chemical shift values for key amino acids were found within the 1.20–4.46 ppm range, thus, this interval was selected for detailed analysis. Initially, the peak areas (area under the curve) within this range were automatically quantified for subsequent statistical evaluations. To adjust for varying concentrations, the area of each peak was normalized against the total peak area within this interval in the respective spectrum before proceeding with discriminant analysis [34,35]. The infusion of [1-¹³C]-glucose and [2-¹³C]-acetate in neuronal and astrocytic metabolic pathways is shown in Figure 2 and Figure 3, respectively.

All measurement results are expressed as mean ± standard error of the mean (Mean ± SEM). A significance level of p < 0.05 was considered statistically significant. Statistical analyses and graphing were conducted using SPSS17.0 (IBM, USA), R4.2.1, and GraphPad Prism8.0. Behavioral analyses were performed using the Mann–Whitney U test. Comparisons between the ZT0 and ZT12 CTR groups, as well as between the respective CTR and HIRI groups at ZT0 and ZT12, were conducted using a t-test. Comparisons of [1-¹³C]-glucose infusion among the four groups (ZT0 CTR, ZT0 HIRI, ZT12 CTR, ZT12 HIRI) were analyzed using two-way ANOVA. Similarly, the comparison of [2-¹³C]-acetate infusion among the four groups was also analyzed using two-way ANOVA. Correlation analyses between metabolic and behavioral data were conducted using Pearson correlation analysis.

In this study, we observed diurnal variations in cognitive impairments in HIRI mice in comparison to the control group of mice (Figure 4). The HIRI mice exhibited compromised spatial and learning memory. In the NORT, the HIRI mice displayed a substantial decline in the novel object RI at both ZT0 and ZT12 compared to CTR mice (Figure 4A, p < 0.001, p < 0.001, respectively). Additionally, the HIRI mice showed reduced exploration time of the new arm than the CTR mice, regardless of whether they were in the ZT0 or ZT12 group (Figure 4B, p < 0.001, p < 0.001, respectively). Furthermore, during the OFT, the HIRI mice in both the ZT0 and ZT12 groups spent significantly less time in the center of the open field, indicating anxiety-like behavior (Figure 4C, p < 0.001, p = 0.001, respectively). Taken together, the behavioral test results suggest impaired learning as well as memory in HIRI mice. Moreover, compared to the ZT0 group of HIRI mice, the spatial learning (p < 0.001) and memory abilities (p < 0.001) of the ZT12 HIRI group were impaired (Figure 4A,B). In this study, all HIRI mice exhibited a decline in cognitive abilities. Therefore, we utilized ¹³C nuclear magnetic resonance spectroscopy combined with intravenous injection of [2-¹³C]-acetate and [1-¹³C]-glucose in order to analyze the metabolic collaboration between astrocytes and neurons in the brains of HIRI mice.

Mice were administered [2-¹³C] acetate infusions at time points ZT0 and ZT12. Subsequently, the concentration of ¹³C-labeled amino acids in brain tissue extracts was measured using [¹H-¹³C]-NMR spectroscopy. A t-test was conducted to compare the differences in metabolite levels of astrocytes in mice under physiological conditions at time points ZT0 and ZT12 (Figure 5A, Supplementary Table S1), which reveals a notable difference only in the enrichment levels of Glu₄, which decreased in the ZT12 CTR group compared to the ZT0 CTR group (p = 0.046). Further comparison of metabolic changes in hippocampal astrocytes of HIRI mice at ZT0 and ZT12 was made employing the t-test. At ZT0, the [2-¹³C]-acetate metabolic results between the CTR group and the HIRI group demonstrated a significant decrease in Glu₄ (p = 0.007) and Gln₄ (p = 0.042) in the HIRI group, along with a notable increase in Myo₅ (p = 0.046, Figure 5B, Supplementary Table S2). Other metabolites showed no discernible changes. At ZT12, the HIRI group exhibited a significant decrease in Ala₃ (p = 0.026), whereas levels of Glx₃ (p = 0.021), Myo₅ (p = 0.016), and Myo₃ (p = 0.019) were higher in contrast to the ZT0 CTR group (Figure 5C, Supplementary Table S3). Other metabolites showed no significant differences. The effects of time, HIRI, and their interaction on the metabolism of hippocampus astrocytes in mice were examined using two-way ANOVA (Supplementary Table S4, Supplementary Figure S1). There was a marginally significant interaction between time and HIRI on Lac₃ (F = 4.338, p = 0.051, Supplementary Table S4), with mean comparisons indicating a decrease in Lac₃ enrichment in the HIRI group compared to the control at ZT0, but a significant increase at ZT12 post-HIRI (Supplementary Figure S1A). The level of Glx₃ was impacted by the interaction between time and HIRI (F = 5.873, p = 0.026, Supplementary Table S4), showing an increase in the HIRI group at ZT0 but a decline compared to the control group at ZT12 (Supplementary Figure S1B).

Our study analyzed the metabolic changes in neurons under physiological conditions in mice at ZT0 and ZT12 using t-tests, as illustrated in Figure 6 and Supplementary Tables S5–S7. At ZT12, compared to the ZT0 CTR group, mice in the CTR group exhibited lower enrichment levels of metabolites such as Ala₃ (p = 0.003), Glu₄ (p = 0.037), Creatine (p = 0.039), and Tau₂ (p = 0.020), but higher Gln₄ (p = 0.035) and Glx₂ (p < 0.001) levels (Figure 6A, Supplementary Table S5). For HIRI mice, their hippocampal neuron metabolism changes were assessed using T-tests at both ZT0 and ZT12. In the ZT0 CTR versus HIRI group, HIRI mice showed elevated levels of GABA₂ (p = 0.024) and Gln₄ (p = 0.017), with decreased Creatine (p = 0.028) enrichment (Figure 6B, Supplementary Table S6). The enrichment levels of Ala₃ (p < 0.001), NAA₃ (p = 0.050), GABA₂ (p = 0.030), Glu₄ (p < 0.001), Gln₄ (p = 0.003), GABA₄ (p = 0.002), Creatine (p < 0.001), and Tau₂ (p = 0.025) showed significant increases at ZT12 post-HIRI (Figure 6C, Supplementary Table S7). A two-factor ANOVA examined the effects of time and HIRI, along with their interaction, on hippocampal neuron metabolism (Figure 6D, Supplementary Table S8). Significant interactions were observed for some metabolites, highlighted through mean comparison analyses. The interaction effects on Ala₃, NAA₃, Glu₄, GABA₄, and Creatine were particularly significant, with distinct changes noted in their levels at ZT0 and ZT12 depending on HIRI influence. Specifically, the interaction between time and HIRI for Ala₃ showed significance (F = 10.143, p = 0.005, Supplementary Table S8). At ZT0, there was no significant change in Ala₃ enrichment, whereas at ZT12, Ala₃ levels decreased in the HIRI group (Supplementary Figure S2A). The interaction for NAA₃ was also significant (F = 4.753, p = 0.042, Supplementary Table S8). At ZT0, there was no significant difference in NAA₃ enrichment between the CTR and HIRI groups, but at ZT12, NAA₃ levels were significantly reduced in the HIRI group (Supplementary Figure S2B). For Glu₄ (F = 17.561, p < 0.001, Supplementary Table S8), there was no significant difference between the control and HIRI groups at ZT0, but at ZT12, the enrichment level significantly increased under the influence of HIRI (Supplementary Figure S2C). GABA₄ (F = 7.896, p = 0.011, Supplementary Table S8) was not affected by HIRI at ZT0, but at ZT12, the levels in the HIRI group were significantly higher than those in the control group (Supplementary Figure S2D). Creatine was influenced by the interaction of time and HIRI (F = 33.169, p < 0.001, Supplementary Table S8). At ZT0, the enrichment level in the HIRI group was lower compared to the control group, whereas at ZT12, the level increased in the HIRI group (Supplementary Figure S2E).

Metabolites labeled with [2-¹³C] acetate, including Lac₃ (r = 0.677), NAA₃ (r = 0.651), Glx₃ (r = 0.733), and GABA₂ (r = 0.643), have been found to positively correlate with the recognitive index in mice. Conversely, Myo₄ (r = −0.855) shows a negative correlation with RI. Asp₃ (r = 0.671), Tau₁ (r = 0.749), and Glx₂ (r = 0.744) are positively correlated with the exploration time of the novel arm, while Myo₃ (r = −0.641) is negatively correlated with the exploration time of the novel arm (Figure 7). Similarly, metabolites labeled with [1-¹³C] glucose such as Creatine (r = 0.755) are positively associated with the RI of mice, while Glu₄ (r = −0.631) and Asp₃ (r = −0.957) show negative correlations with RI. Additionally, Gln₄ (r = −0.604), Glx₂ (r = −0.615), Asp₃ (r = −0.817), and Myo₅ (r = −0.837) are negatively correlated with the exploration time of the novel arm. Furthermore, when labeled with [1-¹³C] glucose, GABA₄ (r = 0.643) is associated with anxiety-related behaviors in mice (Figure 8). On the other hand, Asp₃ (r = 0.741) and Myo₄ (r = 0.857) labeled with [2-¹³C] acetate show positive correlations with the time spent in the center of the open field (Figure 7), while Glx₃ (r = −0.640) and Gln₄ (r = −0.719) are negatively correlated with the time spent in the center of the open field (Figure 7 and Figure 8).

In clinical research, alterations in the glutamate-glutamine cycle are regarded as potential biomarkers for several neurodegenerative diseases. In patients with Alzheimer’s disease, metabolic abnormalities in glutamate and glutamine may be associated with cognitive decline [37,38]. Recent evidence from Magnetic Resonance Imaging (MRI) reveals significantly elevated glutamate concentrations in areas such as the basal ganglia and prefrontal cortex of patients with schizophrenia, supporting the hypothesis of excessive glutamatergic neurotransmission [39]. Glutamate, the most prominent excitatory neurotransmitter in the central nervous system, and its precursor, glutamine, are primarily synthesized by astrocytes and transported back to neurons through specific transporters. Under normal brain function, the release and reuptake of glutamate is a dynamic equilibrium. After being released into the synapse by neurons, glutamate is taken up by astrocytes via EAAT1 (glutamate/aspartate transporter, GLAST) and EAAT2 (Glu transporter 1, GLT-1) to prevent excitotoxicity [40]. Subsequently, glutamate is converted into glutamine within astrocytes by glutamine synthetase (GS), which is a crucial process in the glutamate-glutamine cycle [41]. Glutamine is then released extracellularly, allowing neurons to reutilize it, thus completing the cycle [42]. Excessive accumulation of glutamate can impair normal brain function, leading to excitotoxicity and neuronal death. According to studies, high glutamate concentrations activate NMDA receptors, causing a significant influx of calcium ions. Excessive calcium can lead to intracellular calcium overload, triggering apoptosis and necrosis [43]. Under normal circumstances, astrocytes efficiently uptake and metabolize glutamate, maintaining its proper concentration within the synaptic cleft. However, the ability to remove glutamate is reduced when astrocytic function is compromised, leading to increased synaptic concentrations and excitotoxicity [44]. Research suggests that during cerebral ischemia-reperfusion, the dysfunction or downregulation of astrocytic GLT-1 leads to an increased concentration of extracellular glutamate, exacerbating excitotoxicity [45]. In the hippocampal CA1 region of gerbils, GLT-1 and GLAST levels are reduced in the later stages of ischemia-reperfusion events, resulting in extrasynaptic accumulation of glutamate [46]. Our findings indicate that irrespective of time points ZT0 and ZT12, an increase in the concentration enrichment of Gln and GABA in neurons and astrocytes is observed post-HIRI, as depicted in the metabolic diagram. This implies a GABA-Gln cycle between GABAergic neurons and astrocytes that is analogous to the Glu-Gln cycle [47]. The dysfunction of GABAergic neurons disrupts the cortical excitatory/inhibitory balance, leading to core pathological mechanisms of cognitive dysfunction in schizophrenia [48]. Social cognition and spatial learning abilities are hampered when β7 nicotinic acetylcholine receptors in GABAergic neurons are lost [49]. The findings imply that excitatory activation of GABAergic neurons is important in HIRI-induced cognitive impairment. Further variance analysis indicates that neuronal glutamate Glu₄ enrichment in HIRI mice at time point ZT0 does not significantly differ from the control group. Its level, however, substantially rises at ZT12. On the other hand, no significant effect on Glu₄ was found in the acetate-labelled astrocyte metabolism at the corresponding ZT12 time point. This demonstrates that nocturnal HIRI surgery may promote Glu cycle disruption between neurons and astrocytes, increase glutamate synthesis, and decrease astrocytic glutamate uptake capacity, which could worsen neuronal excitotoxicity and cause synaptic glutamate accumulation.

Alanine primarily functions in neurotransmitter metabolism as a carrier of ammonia and carbon. It participates in the glutamate-glutamine cycle and plays a role in the lactate-alanine shuttle between astrocytes and neurons. In this shuttle, during the pathway of glutamate synthesis from glutamine in neurons, pyruvate receives NH⁴⁺ from glutamine via alanine aminotransferase (ALT) to form alanine, which is subsequently taken up by astrocytes. Within astrocytes, alanine is converted back to pyruvate and then reduced to lactate by lactate dehydrogenase (LDH). This lactate is subsequently transported back to neurons, where the ammonia produced in the process contributes to glutamine synthesis in astrocytes, thereby supplementing the brain’s glutamine-glutamate cycle and regulating nitrogen exchange [50]. Our results indicate that the concentration of alanine in hippocampal neurons is affected by time and HIRI. Although no significant change was observed in alanine levels in HIRI mice compared to controls at ZT0 following [1-¹³C]-glucose infusion, there was a notable increase at ZT12. Correspondingly, [2-¹³C] acetate-labelled astrocytes in the hippocampus of HIRI mice showed elevated lactate levels at ZT12. A t-test comparison between the HIRI and control groups at ZT12 indicated reduced alanine enrichment post-HIRI. This illustrates that liver injury at ZT12 may drive neuronal alanine to be transferred to astrocytes and converted to lactate, which then accumulates. However, the unchanged lactate enrichment in neurons at ZT12 despite HIRI suggests a potential disruption in the lactate-alanine cycle between neurons and astrocytes, possibly as a result of the reduced rate of lactate transportation from astrocytes to neurons.

Glutamate is essential for maintaining neuronal energy metabolism and cellular functions in addition to the synthesis and metabolism of neurotransmitters. In terms of energy metabolism, glutamate serves as both a neurotransmitter and a carbon source for the tricarboxylic acid (TCA) cycle. It is converted into α-ketoglutarate by glutamate dehydrogenase (GDH) [51,52], a process important for oxidative metabolism in neurons, especially when energy demands are high and where glutamate oxidation provides necessary energy support [53]. Both adult and aged rats have decreased GDH activity during ischemia recovery in their brains, which impacts mitochondrial energy metabolism and raises the possibility that glutamate accumulation is the result of impaired neuronal energy metabolism [54]. Additionally, the regulation of the glutamate-glutamine cycle is influenced by nitrogen oxides and inflammatory factors. Studies have shown that nitric oxide (NO) can regulate glutamate metabolism, affecting the conversion rates of glutamate and glutamine as well as neuronal excitability and energy metabolism [55,56]. Lipopolysaccharide (LPS)-induced neuroinflammation can suppress the expression of EAAT1 and EAAT2 and reduce GS activity, affecting glutamate transport [57]. As previously noted, HIRI at ZT12 dramatically increases glutamate levels in mouse hippocampal neurons, potentially influencing the glutamate-glutamine cycle between neurons and astrocytes through systemic inflammation, leading to excitotoxicity and metabolic dysregulation [58]. Moreover, lactate, which is mostly produced by astrocytes and transported to neurons, serves as an energy substrate and signaling molecule, supporting neuronal metabolism and memory consolidation [59,60]. Disruption of the lactate-alanine shuttle, as previously mentioned, may result in decreased lactate transfer from astrocytes to neurons, causing metabolic disturbances when neurons rely on lactate as an energy substrate.

The study highlights significant changes in the concentrations of various low-molecular-weight metabolites, placing a great emphasis on their critical roles in neurological processes. Inositol is an intracellular component whose elevated levels are considered markers of astrocyte proliferation and changes in cell permeability [61]. Astrocytes play a pivotal role in controlling the growth and differentiation of neurons, and they are involved in the clearance of neurotransmitters and the protection of neurons. Clinical research has shown that cognitive impairment in Alzheimer’s Disease patients is positively correlated with levels of myo-inositol in the brain [62]. Our study found significant changes in the content of inositol within astrocytes after HIRI. At ZT12 time points, the inositol content in the HIRI group was higher than that in the control group, indicating an increase in astrocyte proliferation within the brain tissue of mice in the HIRI group. The alteration in the homeostasis of taurine could potentially affect various biological processes, such as osmoregulation, calcium homeostasis, and inhibitory neurotransmitter transmission [63]. Models of chronic neurodegenerative diseases have shown changes in the concentration of taurine in the brain [64,65]. On the other hand, models of insulin-dependent diabetes, insulin resistance, and diet-induced obesity have shown an accumulation of taurine in the hippocampal region [66,67]. Given the potential cytoprotective role of taurine, the accumulation of taurine in the brain may represent a passive activation in response to cognitive impairment events. The increase in neuronal taurine concentrations at the ZT12 time point in HIRI mice suggests that cognitive impairment caused by HIRI may activate the neuroprotective effects of taurine. Creatine is a nitrogenous organic acid that serves as a phosphate energy buffer. Creatine is also one of the main substances regulating osmotic pressure in the brain and is crucial in tissues with high energy demands, such as muscles or the brain. The specific knockout of the creatine transporter gene Slc6a8 results in cognitive deficits [68]. Compared to the control group at ZT0, the concentration of creatine in HIRI mice at the ZT0 time point decreased, while at the ZT12 time point, the neuronal creatine concentration in HIRI mice showed increased, possibly indicating temporal variations in the degree of cognitive impairment caused by HIRI.