This study used C57BL/6 mice (male, 8 weeks, 23 $\pm$ 2 g) purchased from Gempharmatech (Chengdu, China). Mice were housed in a pathogen-free environment with a 12-hour/12-hour light/dark cycle, 25 °C ambient temperature, and 45–55% relative humidity. They had unrestricted access to food and water. The Animal Experimentation Ethics Committee of Chengdu University of Traditional Chinese Medicine (Chengdu, China) granted approval for all animal experiments in accordance with established procedures. All experiments were conducted in accordance with the National Academy of Sciences and National Institutes of Health’s Guide for the Care and Use of Laboratory Animals. No abuse or maltreatment occurred during our study.

We constructed the ICH mouse models according to our previously developed methods [21, 22]. Briefly, a stereotaxic apparatus (RWD Life Science Co., Shenzhen, Guangdong, China) was applied to immobilize mice after they were anesthetized with 3% and 1.5% isoflurane (Cat# R510-22-10, RWD Life Science Co., Shenzhen, China), respectively, for anesthesia induction and maintenance. A total volume of 0.5 µL type VII collagenase (0.075 IU, C0773, Sigma Aldrich, St. Louis, MO, USA) was continuously injected into the mouse left striatum (0.8 mm anterior, 2 mm lateral to bregma, and at a depth of 3.5 mm) using a syringe pump (Legato 130, RWD Life Science Co., Shenzhen, China). Control mice received saline injection (0.5 µL). Mouse body temperature was maintained at 37 °C throughout surgery and anesthesia resuscitation. After recovery from anesthesia, a Longa score $\ge$1 indicated the success of ICH model construction. The death rate and success rate of ICH mice was 3.41% and 92.04%, respectively. We discarded all ICH model mice that were deemed unsuccessful, either due to being asymptomatic or that resulted in death. TEPP-46 (HY-18657, MedChemExpress, Middlesex County, NJ, USA) at a dosage of 50 mg/kg was intraperitoneally injected into ICH mice once a day for 5 consecutive days, starting 6 hours after ICH.

According to our previously reported method [23], western blots were performed for assessing the PKM2 levels in the perihematomal brain tissues of ICH mice. Briefly, the ipsilateral brain tissues and control brain tissues of mice were collected and homogenized to extract proteins using a prepared radio immunoprecipitation assay (RIPA) lysis buffer (Beyotime, Shanghai, China), and a Bicinchoninic Acid (BCA) protein assay (Cat# 23227/23225, Thermo Scientific, Waltham, MA, USA) was used to determine the protein concentration. Proteins were resolved by sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE) and electroblotted onto polyvinylidene fluoride (PVDF; Millipore, Burlington, MA, USA) membranes after adjustment to the same final concentration. Next, after blocking the membranes for 2 hours at room temperature with 5% non-fat milk in tris-buffered saline containing 0.1% Tween-20 (TBST; Biosharp, Guangzhou, Guangdong, China), they were incubated with anti-PKM2 (#4053, 1:50, Cell Signaling Technology, Danvers, MA, USA) antibody overnight at 4 °C. The membranes were then washed four times with TBST and incubated with a secondary antibody conjugated with horseradish peroxidase (HRP) (1:2500; OR03L, Sigma-Aldrich, St. Louis, MO, USA) at room temperature for 2 hours. An enhanced chemiluminescent (ECL) substrate was used to view the immunoreactive proteins with an imaging system (GelView 6000Plus, Guangzhou Biolight Biotechnology Co., Ltd., Guangzhou, Guangdong, China). Following stripping, the membranes were incubated again with glyceraldehyde-3-phosphate dehydrogenase (GAPDH; Cat# 60004-1-Ig, Proteintech, Wuhan, Hubei, China) acting as the loading control. The band intensities were normalized to the values for GAPDH and quantified using Image J software (version 1.46J, National Institutes of Health, Bethesda, MD, USA).

Based on our previously reported procedures [23], we performed immunohistochemical staining to investigate the PKM2 expression profiles in the brains of ICH mice. The mice were sedated and their brains were rapidly removed and preserved in 4% paraformaldehyde (PFA; Cat# P1110, Solarbio, Beijing, China) at 4 °C overnight, after being intracardially perfused with PBS and 4% paraformaldehyde. Brain sections of 15 µm were then incubated with the following antibodies: anti-PKM2 (#4053, 1:50, Cell Signaling Technology, USA), Iba-1 (ab178846, 1:200, Abcam, Cambridge, MA, USA), NeuN (#24307, 1:200, Cell Signaling Technology, USA), and glial fibrillary acidic protein (GFAP; ab7260, 1:300, Abcam, USA). Secondary antibodies, including Alexa Fluor 594 (1:200, donkey anti-rabbit), Alexa Fluor TRITC (1:200, rabbit anti-goat), and Alexa Fluor 488 (1:200, donkey anti-mouse), were purchased from Proteintech Group (Wuhan, Hubei, China). The brain slices were imaged using confocal microscopy (Olympus IXplore SpinSR10, Olympus, Tokyo, Japan). Quantification of double-labeled cells was used to calculate the ratio of co-labeled cells. Three brain slices were selected from each mouse brain at random; double-labeled cells were counted automatically, and the average data obtained by three independent researchers who were unaware of group information was used to determine the value for each brain.

After the mice were euthanized, their brains were rapidly removed and fixed in a 10% neutral-buffered formalin solution (Cat# G2161, Solarbio, Beijing, China). Following dehydration in ethanol solutions, the brain tissues were embedded in paraffin. The brain tissues were then sectioned at 5 µm thickness using a rotary microtome. Lastly, three slices of each brain tissue were selected at random, then stained with H&E, and viewed under an Olympus microscope (Olympus SLIDEVIEW VS200, Olympus, Tokyo, Japan) to evaluate the damage in these perihematomal brain tissues.

Based on the manufacturer’s instructions of a commercial One-step TUNEL In Situ Apoptosis Kit (E-CK-A320, Elabscience, Wuhan, Hubei, China), TUNEL staining was performed. Briefly, 10 µL 1 $\times$ protease K working solution was added to the three slices of each mouse brain selected randomly and incubated at 37 °C for 10 min. The brain slices were then washed three times with PBS for 5 min each. The sections were then incubated with 100 µL 1 $\times$ DNase I Buffer at room temperature for 5 min. After removing excess liquid from the sample, 100 µL diluted DNase I working solution (200 U/mL) was added to each brain section, then incubated at 37 °C for 10–30 min. After washing with PBS, 100 µL TdT Equilibration Buffer was added to each brain slice and incubated at 37 °C for 10~30 min. After removing the TdT Equilibration Buffer, 50 µL marker working solution was added and incubated at 37 °C for 60 min in a wet box in the dark. Finally, 4,6-diamino-2-phenyl indole (DAPI; Cat# G1012-100ML, Servicebio, Wuhan, Hubei, China) solution was added to stain the nucleus. The TUNEL-stained brain slices were observed under a Lycra microscope (Leica DM6B, Leica AG, Wetzlar, Germany). For further analysis of the TUNEL results, three researchers who were unaware of the group information independently evaluated the number of positive cells in each standardized microscopic field in the perihematomal brain tissues using ImageJ software (version 1.46J, National Institutes of Health, USA).

The corner test was conducted using previously published procedures [24, 25]. ICH mice were first placed between two cardboard pieces with a 30${}^{\circ }$ angle corner. The operator progressively moved the board toward the mouse while preserving a 30${}^{\circ }$ angle, until the mouse was near the corner, at which point the board was lifted and rotated 180${}^{\circ }$ to face the open end. Each mouse was evaluated 10 times, and the direction in which it turned during each trial (left or right) was recorded. Behavioral testing was performed on 10 independent mice in each group, and the number of right (R) and left (L) turns was recorded. This was used to calculate an overall score = [(R)/(R+L)] $\times$ 100%. Subsequently, the mouse was placed in a fresh cage in order to evaluate its sensory neglect using the adhesive tape removal task. The adhesive tape was affixed to the planter area of the left and right front paws of the mouse, and the time it took for each paw to successfully remove the tape was recorded. Up to 300 seconds of testing was permitted for each paw.

Perihematomal brain tissue of ICH mice, after treatment with or without TEPP-46 at 5 days post ICH, was collected for sequencing. Brain tissue total RNA was extracted, and cDNA libraries were constructed using an Oxford Nanopore Technologies (ONT)-supplied Ligation Sequencing Kit 1D (PM) (SQK-LSK110, Oxford Nanopore Technology, Oxford, England) in accordance with the company’s protocol. Full-length cDNA was enriched using reverse transcriptase and the addition of defined PCR adapters to both ends of the first strand of cDNA, then 14-circle cDNA PCR (8 min extension time) was performed using LongAmp Tag DNA polymerase (New England Biolabs, Ipswich, MA, USA). T4 DNA ligase (New England Biolabs, Ipswich, MA, USA) was then used to bind the PCR products to the ONT adaptor ligation. DNA purification was performed using Agencourt XP beads (Beckman Coulter, Brea, CA, USA). A PromethION platform of the Biomarker Technology Company (Beijing, China) was used to analyze the final cDNA libraries after being processed by FLO-MIN109 flow cells (FLO-PRO002, Oxford Nanopore Technology, Oxford, England).

Raw reads with an average quality score of less than 7 and a length of less than 500 bp were eliminated. After the rRNA was mapped to the rRNA database, it was discarded. Next, primers at both ends of the reads were searched to identify the full-length, non-chimeric (FLNC) transcripts. The Mimimap2 alignment program [26]was applied to the reference genome library in order to obtain clusters of FLNC transcripts. Full-length reads were mapped to the reference genome (Rnor_5.0). Reads with match quality more than 5 were further used for quantitative analysis. Counts per million (CPM) [27] was used as an indicator to measure the levels of transcript or gene expression. Using the DESeq R package (DESeq2 1.6.3, Germany; http://www.bioconductor.org/packages/release/bioc/html/DESeq2.html), DEGs between the two groups were analyzed. The p-values were adjusted using Benjamini and Hochberg’s approach, and the DEGs induced by TEPP-46 treatment were defined as those with a p-value 0.05 and a fold change (FC) $\ge$1.5. The subcellular location information of the proteins encoded by down-regulated DEGs induced by TEPP-46 was collected from the UniProt database (http://www.uniprot.org/database/) [28]. The Goseq R package based on Wallenius noncentral hypergeometric distribution was used for Gene Ontology (GO) enrichment analysis of the down-regulated DEGs, which can adjust for the bias in gene length of DEGs. Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis was performed to analyze the statistical enrichment of down-regulated DEGs using KOBAS software (http://kobas.cbi.pku.edu.cn/index.php).

Conjoint analysis was performed using our published method [21, 22] by matching the down-regulated DEGs induced by TEPP-46 with previously published transcriptome data [29]. Considering the extensive research conducted on brain cells in the context of stroke, including neurons, glial cells (e.g., astrocytes and microglia), and endothelial cells, these four cell types were selected for further analysis. Firstly, the relatively high expression of genes in the four cells were defined using previous methods [21, 22]. For example, gene A is defined as being highly specific to astrocytes when the ratio of the fragments Per Kilobase of exon model per Million fragments mapped (FPKM) of gene A in astrocytes to that in the other three neural cells (neurons, microglia, and endothelial cells) is $>$10.0 and being highly expressed in neurons is defined as having a ratio $>$1.0. To further identify close relationships between down-regulated DEGs and neurons, glial cells, and endothelial cells, an association matching analysis was performed by matching the down-regulated DEGs with the highly expressed genes in each type of neural cell.

The data were presented as the mean $\pm$ standard deviation (SD), and were analyzed using SPSS 26.0 statistical software (IBM Corp., Armonk, NY, USA). For pairwise comparisons, nonparametric Wilcoxon-Mann-Whitney tests were conducted, while the nonparametric Kruskal-Wallis test was applied for comparisons among more than two groups. Values of p 0.05 were considered statistically significant.

To determine the role of PKM2 in ICH-induced brain damage, we analyzed the expression profiles of PKM2 in the perihematomal brain tissues following ICH. First, we performed a western blot and observed that PKM2 levels were significantly elevated beginning 3 days after ICH, peaked at 5 and 7 days, and remained elevated at day 14 compared with sham groups (Fig. 1A,B). The expression levels of PKM2 and the cellular sources of PKM2 in ICH mouse brains were further analyzed by the immunofluorescent staining. In the sham groups, the total PKM2 expression level was low, and PKM2 expression was mostly localized in glial cells, particularly in the IBA-1-positive microglia, even though these cells were not activated (Fig. 1C). In contrast, in the perihematomal brain tissues of mice at 5 days after ICH, we found that PKM2 expression and glial activation were both significantly increased, and PKM2 was mainly co-localized with GFAP-positive astrocytes, followed by IBA-1 positive microglia, and rarely co-localized with NeuN-positive neurons (Fig. 1D). These data revealed a dynamic change feature of PKM2 expression after ICH, in which low and microglial PKM2 expression gradually switches to high and astrocytic expression in response to the brain damage caused by ICH.

Fig. 1.

PKM2 expression profiles in the perihematomal brain tissues of ICH mice. (A,B) Representative image of western blot (A) and statistical analysis (B) of PKM2 expression in the perihematomal brain tissues at different time points after ICH (each time point, n = 5 for each group). (C,D) Immunofluorescent staining of cells expressing PKM2 with NeuN+ neurons, IBA-1+ microglia, and GFAP+ astrocytes in sham (C) and ICH mice (D), respectively (n = 4 for each group). (E) Representative immunofluorescent image of PKM2 nuclear translocation in the perihematomal brain tissues at 5 days following ICH. (F) Statistical results of immunofluorescent staining. n = 4 for each group. *p 0.05, **p 0.01, and ***p 0.001 vs sham controls; n.s. indicates not statistically significant. PKM2, pyruvate kinase M2; ICH, intracerebral hemorrhage; IBA-1, ionized calcium binding adaptor molecule-1; GFAP, glial fibrillary acidic protein; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; DAPI, 4,6-diamino-2-phenyl indole.

The enzymatic activity of PKM2 is closely related to its oligomerization status, including active tetramer, less active dimer, and inactive monomer [18]. PKM2 detetramerization directly leads to the nuclear translocation of its inactive monomer [20], which has been shown to regulate gene transcription by functioning as a protein kinase in the nucleus. [18, 19, 30]. Our immunofluorescent staining data was analyzed to determine the PKM2 oligomerization status after ICH by evaluating the nuclear expression levels. In the sham groups, low PKM2 expression was observed mostly in the cytoplasm, whereas 30% of PKM2-positive cells exhibited nuclear translocation in the ICH mice (Fig. 1E,F), suggesting an ICH-induced shift to PKM2 monomers. TEPP-46, a small-molecule activator that stabilizes PKM2 in its tetramer form [31], was then intraperitoneally injected into ICH mice for 5 consecutive days. The results demonstrated that TEPP-46 treatment notably prevented the nuclear translocation of PKM2 in the perihematomal brain tissues of ICH mice, compared with ICH mice treated with a vehicle (Fig. 2A,B). This indicates that increased PKM2 nuclear translocation in glial cells resulting from the PKM2 detetramerization induced by ICH may participate in the regulation of glial cell activation and brain injury, as PKM2 nuclear translocation has been shown to promote the activation of immune cells [13].

Fig. 2.

TEPP-46 treatment markedly reduced the PKM2 nuclear translocation after ICH. (A) Representative immunofluorescent image showing that TEPP-46 induced a significant reduction of PKM2 nuclear translocation in the perihematomal brain tissues at 5 days after ICH. (B) Statistical results of immunofluorescent staining. n = 4 for each group, **p 0.01 vs sham controls.

Through the evaluation of glial markers, it was shown that the mice in the vehicle-treated ICH group had a significant increase in the presence of activated astrocytes (Fig. 3A) and microglia (Fig. 3B) inside the perihematomal brain tissue, as compared with the sham group. In contrast, the TEPP-46 treatment significantly reduced both the activation of astrocytes and microglia after ICH (Fig. 3), suggesting that PKM2 detetramerization promoted glial activation after ICH. H&E staining was conducted to assess the general brain injury and we observed that TEPP-46 treatment resulted in a reduction of brain injury in the perihematomal brain tissues at 5 days post ICH. (Fig. 4A). Similarly, TEPP-46 treatment reduced the increased TUNEL+ cells in the brain tissue surrounding the hematoma (Fig. 4B). In the corner tests, the TEPP-46-treated ICH mice almost recovered from their physical disability within 5 days (Fig. 4C), which was consistent with the results of the PKM2 expression profile. However, the sensory neglect of mice with ICH was not significantly improved after TEPP-46 treatment, although there was a trend for improvement (Fig. 4D). These results strongly indicate that TEPP-46 treatment reduces brain injury and partly improves functional recovery of ICH that may be related to the inhibition of glial activation by preventing PKM2 nuclear translocation in glial cells.

Fig. 3.

Glial activation was significantly inhibited after ICH treated by TEPP-46. The activation of GFAP${}^{+}$ astrocytes (A) and IBA-1${}^{+}$ microglia (B) were inhibited in ICH mice at 5 days after TEPP-46 treatment (n = 4 for each group).

Fig. 4.

TEPP-46 treatment markedly reduced brain injury and improved functional recovery of mice with ICH. (A) H&E staining shows that brain injury and infiltration of immune cells were markedly reduced in ICH mice treated with TEPP-46 (n = 4 for each group). (B) TUNEL+ cells were obviously reduced in ICH mice treated with TEPP-46 (n = 4 for each group, **p 0.01, ***p 0.001). (C) TEPP-46 treatment markedly enhanced performance on the corner test (a measure of spatial neglect, n = 10 for each group; *p 0.05 vs vehicle-treated ICH mice) at 4 and 5 days post ICH. (D) TEPP-46 treatment did not markedly enhance performance on the tape removal task (a measure of sensory neglect, n = 10 for each group; p $>$ 0.05). Scale bar = 100 µm for (A) and = 50 µm for (B). TUNEL, Terminal Deoxynucleotidyl Transferase Deoxyurinary Triphosphate (dUTP) Nick End Labeling.

RNA-seq was then performed on perihematomal brain tissue from ICH mice treated with or without TEPP-46 to obtain insight into the molecular consequences of glial inhibition induced by PKM2 tetramerization (Fig. 5A). Surprisingly, 91.1% (205/225) of DEGs were down-regulated in the ICH mouse brains treated by TEPP-46 compared with those treated with vehicle (Fig. 5B), indicating that most of the upregulated DEGs induced by ICH may have been inhibited when PKM2 nuclear translocation was blocked. The proteins encoded by these down-regulated DEGs were widely distributed throughout the cells, including in the nucleus and cytoplasm (Fig. 5C). An analysis of these down-regulated DEGs using GO term enrichment analyses revealed that TEPP-46-treated mice, when compared with vehicle controls, had decreased responses to autophagy and its related biological processes, metabolic processes, receptor complex, and membrane-related events (Fig. 5D). In addition, KEGG pathway analysis found that the majority of these downregulated DEGs were enriched in the PI3K-Akt signaling pathway, autophagy, Ras signaling pathway, chemokine signaling pathway, and Wnt signaling pathway (Fig. 5E); all of which have been shown to be closely related to ICH brain damage.

Fig. 5.

Analysis of RNA-seq data of ICH brains. (A) PCA analysis reveals the dispersion degree of the transcriptome between the TEPP-46- and vehicle-treated ICH brains. (B) The volcano plot shows the down-regulated and up-regulated DEGs induced by TEPP-46 treatment for ICH brains compared with the vehicle-treated brains. (C) The subcellular localization of proteins encoded by down-regulated DEGs in (B). (D,E) GO (D) and KEGG (E) analysis of the down-regulated DEGs in (B). (n = 3 for each group). PCA, Principal Component Analysis; FC, fold change; DEGs, differentially expressed genes; GO, Gene Ontology; KEGG, Kyoto Encyclopedia of Genes and Genomes.

As the assessed samples were perihematomal brain tissues, cellular information regarding changes at the transcriptome levels was lost. To demonstrate that TEPP-46 treatment induced alterations in the transcriptome of neural cells in response to ICH injury [29], we performed a conjoint analysis of previously published RNA-seq data of brain cells and our down-regulated DEGs using our previously published methodologies [21, 22]. After comparing the highly expressed genes of brain cells [21, 22] with the down-regulated DEGs, we showed that the majority of the DEGs were relatively highly expressed in neurons, microglia, endothelial cells, and astrocytes (Fig. 6A); 37.56% of total down-regulated DEGs were identified in neurons, 17.56% in microglia, 17.07% in endothelial cells, and 13.17% in astrocytes (Fig. 6A). Considering that the levels of mRNA expression in cells vary greatly, and that higher mRNA expression may be important for cell function, the specificity ratio of the DEGs was further analyzed to evaluate whether these down-regulated DEGs induced by TEPP-46 treatment could potentially influence the biofunction of neural cells. The results depicted by pie charts indicate that 12 down-regulated DEGs (Gng4, Larp6, Bhlhe22, Fam19a2, Efna3, Npas1, Hs6st3, Cnksr2, Ank1, Crmp1, Vstm2l, and Ccdc92) with a specificity ratio $>$10 in neurons (Fig. 6B, upper right). For astrocytes and microglia, there were two DEGs with specificity ratios $>$10 (Igfbp2 and Btbd17 for astrocytes, Gns and Limd2 for microglia; Fig. 6B, left), respectively. Only a single DEG with a specificity ratio $>$10 was detected in endothelial cells (Sox17; Fig. 6B, lower right). Furthermore, the specificity ratios of DEGs in the four brain cell types exhibited a range of 1–3 (Fig. 6B). Notably, endothelial cells and microglia had a very high proportion; more than 72% of the specificity ratios were 2 (Fig. 6B). These findings demonstrate that TEPP-46 treatment may have a direct impact on the functioning of glial cells, as well as an indirect influence on other cellular processes inside the brain, hence providing a neuroprotective effect against ICH.

Fig. 6.

Combined analysis of the down-regulated DEGs of ICH brains induced by TEPP-46 and transcriptomics data. (A) The major cellular distributions of down-regulated DEGs of ICH brains after TEPP-46 treatment compared with vehicle treatment. (B) The pie charts show the specificity ratio distribution of the down-regulated DEGs in neurons (upper left), astrocytes (upper right), microglia (lower right), and endothelial cells (lower left).