Lactate promotes microglial scar formation and facilitates locomotor function recovery by enhancing histone H4 lysine 12 lactylation after spinal cord injury

Abstract

Lactate-derived histone lactylation is involved in multiple pathological processes through transcriptional regulation. The role of lactate-derived histone lactylation in the repair of spinal cord injury (SCI) remains unclear. Here we report that overall lactate levels and lactylation are upregulated in the spinal cord after SCI. Notably, H4K12la was significantly elevated in the microglia of the injured spinal cord, whereas exogenous lactate treatment further elevated H4K12la in microglia after SCI. Functionally, lactate treatment promoted microglial proliferation, scar formation, axon regeneration, and locomotor function recovery after SCI. Mechanically, lactate-mediated H4K12la elevation promoted PD-1 transcription in microglia, thereby facilitating SCI repair. Furthermore, a series of rescue experiments confirmed that a PD-1 inhibitor or microglia-specific AAV-sh-PD-1 significantly reversed the therapeutic effects of lactate following SCI. This study illustrates the function and mechanism of lactate/H4K12la/PD-1 signaling in microglia-mediated tissue repair and provides a novel target for SCI therapy.

Graphical Abstract

Background

Spinal cord injury (SCI) is a serious disorder of the central nervous system (CNS), causing autonomic, sensory, and motor dysfunction [1, 2]. SCI triggers a complicated range of inflammatory processes that culminate in scar formation involving astrocytic, fibrotic, and microglial scarring [3,4,5]. Previous research has suggested that the secretion of chondroitin sulfate proteoglycans (CSPGs) by astrocytic scars is responsible for failure of axon regeneration [6]. However, another study has challenged this belief by proposing that astrocytic scars support axon regeneration [7]. The fibrotic scar clearly leads to failure of axon regeneration, as axon terminals generate retraction bulbs upon contact with fibroblasts and cannot pass through the scar [8]. Microglial scars have received limited attention until recently. In the first week after SCI, microglia are activated into a proliferative state and quickly migrate to the lesion core, where they phagocytose myelin debris and participate in microglial scar formation [9, 10]. However, the mechanism underlying microglial scar formation remains unclear.

SCI is known to cause microvascular damage and inflammation, resulting in an ischemic and hypoxic microenvironment in the metabolically active spinal cord, thereby disrupting the coupling of glycolysis and oxidative phosphorylation [11, 12]. Lactate, produced as a byproduct of glycolytic metabolism under hypoxia plays a crucial role in the flexibility of metabolism and is emerging as a potential bioenergetic fuel and cellular function regulator [13]. Lactate accumulation has been suggested to result from an initial inflammatory response or insufficient blood perfusion [14]. Both exogenous and endogenous lactate can contribute to histone lactylation and subsequently activate downstream gene transcription [15, 16]. The pivotal roles of histone lactylation in cell fate determination, embryonic development, and neuropsychiatric disorders have been confirmed [17]. Nevertheless, the role of lactate and histone lactylation after SCI remains unclear.

In this study, we report a significant elevation in lactate and lactylation levels in the spinal cord following SCI. Notably, lactylation and H4K12la levels were significantly increased in microglia after SCI. Functionally, lactate treatment was found to promote microglial proliferation, scar formation, axon regeneration, and locomotor function recovery after SCI. Mechanically, lactate-mediated H4K12la elevation was found to promote the transcription of PD-1 in microglia. Our findings elucidate the function and mechanism of lactate/H4K12la/PD-1 signaling in microglia after SCI, providing a novel approach for treating SCI.

Materials and methods

Animals and SCI model

The relevant animal procedures complied with the Animal Committee of Anhui Medical University (Approval No. LLSC20220381). Eight-week-old C57BL/6 J mice (18–22 g) were purchased from the Experimental Animal Center of Anhui Medical University and, and 8-week-old CCR2^RFPCX3CR1^GFP dual-reporter mice (Catalog No. 032127) were purchased from Shanghai Model Organism. All mice were housed in an environment with controlled temperature and humidity and a 12/12 h light/dark cycle.

Establishment of the spinal cord compression injury model has been described in detail in our previous study [18]. In brief, mice were injected intraperitoneally with pentobarbitone (50 mg/kg). After laminectomy, the exposed thoracic level 10 (T10) spinal cord was exposed. A 5-s moderate crush injury was performed using Dumont forceps No. 5 (Fine Science Tools, 11251–20, Germany) with a spacer resulting in a tip width of 0.2 mm and a spacing of 0.25 mm upon closure [19]. In addition, a forcep imprint can be observed after immediate crush injury and is not accompanied by hemorrhagic erythema of the adjacent spinal cord. Auxiliary urination nursing was performed twice a day, and a daily subcutaneous injection of Baytril (10 mg/kg) was used for anti-infection for 1 week.

Primary microglial culture

Primary microglia were extracted from the brains of mice (postnatal 0–3 days old). Briefly, brain tissues were divided into pieces and digested with 0.25% trypsin (Biological Industries) for 25 min following removal of the blood arteries and meninges, before adding 3 ml of culture medium to stop digestion. After gently pipetting the cells up and down ten times using a 1-ml pipette, the cell suspension was filtered with a 40-μm-mesh filter and centrifuged at 1200 rpm for 5 min to obtain mixed glial cells. For primary microglia, the mixture of cells was inoculated into poly-D-lysine-precoated T25 flasks and cultured in DMEM/F12 supplemented with fetal bovine serum (FBS; Thermo-Fisher Scientific), penicillin, and streptomycin. After 14 days of in vitro culture, the flasks were shaken at 250 rpm for 6 h to remove mature microglia. The collected microglia were seeded in culture plates and maintained at 37 °C with 5% CO₂. For all assays, microglia were seeded at a density of approximately 1 × 10⁶/ml for 24 h before each treatment.

Western blotting

Spinal cord tissues were lysed using lysis buffer containing RIPA, protease inhibitor and phosphatase inhibitor. The protein concentration of the lysate was tested using the BCA assay and was set to the same final concentration. After heat denaturation for 10 min, identical proteins were separated by SDS-PAGE transferred to PVDF membranes, and incubated in blocking buffer (5% non-fat dry milk in TBST) for 2 h. The membranes were incubated overnight in the following primary antibodies: rabbit anti-Pan Kla (1:1000, PTM-1401, PTM BIO), rabbit anti-H4K5la (1:1000, PTM-1409, PTM BIO), rabbit anti-H4K8la (1:1000, PTM-1405, PTM BIO), rabbit anti-H4K12la (1:1000, PTM-1411, PTM BIO), rabbit anti-H3K18la (1:1000, PTM-1406RM, PTM BIO), rabbit anti-H4 (1:1000, 16047–1-AP, Proteintech), rabbit anti-β-Tubulin (1:1000, 10094–1-AP, Proteintech), rabbit anti-PD-1 (1:1000, ab214421, Abcam), and mouse anti-GAPDH (1:1000, 60004–1-Ig, Proteintech). After three rounds of washing, the membranes were incubated with the corresponding secondary antibodies β-Tubulin, H4, and GAPDH used as loading controls. Immunoreactive bands were detected using enhanced chemiluminescence (ECL) and quantified using ImageJ software.

Immunofluorescence

For in vivo analysis, 16 μm-thick sagittal spinal cord sections encompassing the lesion core from mice at pre-operation (Pre), and 7, 14, and 28 days post-injury (dpi) were applied for immunofluorescence staining followed procedures described previously [20]. For in vitro analysis, primary microglia cultured on coverslips were fixed in 4% PFA for 30 min at room temperature. For BrdU staining, the sections and coverslips were pretreated with 2 N hydrochloric acid (HCl, GEMIC, China) for 30 min at 37 °C, followed by 0.1 M boric acid buffer (KGR0101, Key-GEN BioTECH, China) for 15 min at room temperature. The sections and coverslips were then subjected to immunofluorescence staining.

The sections and coverslips were cleaned thrice with phosphate buffered saline (PBS) and incubated with blocking buffer (10% donkey serum containing 0.3% Triton X-100) for 2 h, before incubating overnight at 4 °C with the following primary antibodies: rabbit anti-Pan-lysine lactylation (Pan Kla, 1:100, PTM-1401, PTM BIO), rabbit anti-H4K12la (1:100, PTM-1411, PTM BIO), goat anti-Iba1 (1:500, NB100-1028, Novus Biologicals), chicken anti-GFP (1:500, GFP1010, Aves Labs), goat anti-RFP (1:200, 200-101-379, Rockland), rabbit anti-PD-1 (1:100, ab214421, Abcam), mouse anti-CX3CR1 (1:100, sc-377227, Santa Cruz), rabbit anti-CX3CR1 (1:100, Abcam, ab8021), goat anti-PDGFRβ (1:100, AF1042-SP, R&D Systems) rat anti-GFAP (1:400, 13–0300, Invitrogen), goat anti-5-hydroxytryptamine (5-HT) (1:100, 20079, Immunostar), rat anti-BrdU (1:100, ab6326, Abcam), and mouse anti-NeuN (1:100, MAB377, Millipore). After washing, sections and cell coverslips were incubated with secondary antibodies conjugated with Alexa Fluor 488 and Alexa Fluor 555, or Alexa Fluor 594 (1:500, Invitrogen). The sections and cell coverslips were stained with 4′, 6-diamidino-2-phenylindole (DAPI; D9542, Sigma) to label the cell nuclei. Zen software (Zeiss) was used to capture the images.

Pharmacological agents

For the in vivo experiments, mice received daily intraperitoneal injections of the following drugs: Sodium lactate (lactate, 1 g/kg, L7022, Sigma), with a final concentration selected according to previous reports [16]; alpha-cyano-4-hydroxycinnamate (4-CIN, 200 mg/g, C2020, Sigma) [21]; and sodium oxamate (OX, 300 mg/kg, O2751, Sigma) [22]. To block PD-1/PD-L signaling in vivo, InVivoPlus anti-mouse PD-1 (CD279) (P0146, Bio X Cell) was administered intraperitoneally at 5 mg/kg at 1 and 14 dpi, and mice were sacrificed at 28 dpi [23].

For in vitro experiments, microglia were treated with lactate (20 mM), PBS, AAV-sh-PD-1 (1–2 × 10¹⁰ vg/ml), or a combination of both as appropriate for another 24 h, before collecting microglia for subsequent experiments.

To assess microglial proliferation in vivo or in vitro, the mice were intraperitoneally injected with BrdU (50 mg/kg, Biosharp) daily from 1 to 7 dpi, and the primary microglia were cultured with BrdU (l0 μg/ml) for 24 h.

Image acquisition and quantitative analysis

Representative images of the sections were acquired using a Zeiss LSM 900 confocal microscope and a Zeiss Axio Scope A1 fluorescence microscope. Immunofluorescent images were quantified using ImageJ version 2.0 (NIH, United States) by observers who were blinded to the experimental conditions. For each individual analysis, three sagittal sections (spaced 160 μm apart) encompassing the lesion core in each sample were selected, with three animals per group. To quantify the intensity of Pan Kla⁺, H4K12la⁺, and PD-1⁺ in microglia (Iba1⁺, GFP⁺RFP⁻, or CX3CR1⁺), three random 40 × images encompassing the lesion core per section were analyzed, and the average was used as the final data for each sample. The data were expressed as relative fluorescence intensities between different groups, as previously described [16].

To identify the spatial expression pattern of H4K12la after SCI, three sections in each sample were stained for H4K12la and CX3CR1, PDGFRβ (pericytes), and GFAP (astrocytes). Three random 40 × images were counted per section, and the average was used as the final data for each sample. The data were expressed as a percentage of H4K12la⁺CX3CR1⁺ cells, H4K12la⁺PDGFRβ⁺ cells, or H4K12la⁺GFAP⁺ cells relative to the total number of H4K12la⁺ cells.

For the quantification of the proliferating microglia (BrdU⁺ CX3CR1⁺) in vivo, three sections in each sample were stained with BrdU and CX3CR1. And the number of BrdU⁺ CX3CR1⁺ cells per mm² was counted on three random 40 × images around the lesion core. To quantify proliferating microglia in vitro, the number of BrdU⁺ cells per mm² was counted on three random 40 × images, with three coverslips per culture, using the average as the final data for each culture.

To quantify the microglial scar (CX3CR1⁺) [5]. 100 μm-square grids were generated over the 10 × images encompassing the lesion core. The number of CX3CR1⁺ cells was counted on every six squares, and only DAPI⁺ cells were counted, using the average as the final data for each sample.

To quantify the number of NeuN⁺ cells, three sagittal sections of each sample were stained with GFAP and NeuN. The sections were divided into three zones based on their proximity to the lesion core. These zones were composed of Z1, Z2, and Z3 as previously described [24]. The number of NeuN⁺ cells was counted on 10 × images, with the average used as the final data for each sample.

To quantify the regenerated 5-HT axons, three sagittal sections of each sample were stained for GFAP and 5-HT. The immunoreactivity of 5-HT was normalized to the area of the GFAP⁻ region on 10 × images, with average used as the final data for each sample.

Adeno-associated virus infection

AAV-CX3CR1-eGFP-5′miR30-shRNA-3′miR30-WPRE (AAV-sh-PD-1) (Genomeditech, Shanghai, China) was used to target CX3CR1⁺ cells and interrupt PD-1 expression in microglia, with AAV-CX3CR1-eGFP-5′miR30-MCS-3′miR30-WPRE used as a control vector (AAV-Con). AAV9 vectors expressing a short hairpin RNA (shRNA) directed at PD-1 (AAV-sh-PD-1) or a control hairpin (AAV-Con) were generated to induce local depletion of PD-1 in vivo. Following SCI, the mice were placed in a stereotaxic frame, and a 10 μl Hamilton Neuros syringe connected to a micro-syringe pump controller (neMESYS OEM, Cetoni, Germany) was inserted to inject the viral vector (5 × 10⁹ vg; 1 μl) into the spinal cord at a speed of 200 nl/min into the region of the injured core at a depth of 1 mm per corner.

Cleavage under targets and tagmentation (CUT&Tag)

The CUT&Tag assay was conducted using the Hyperactive^™ In-Situ Chromatin immunoprecipitation (ChIP) Library Prep Kit for Illumina (TD901-TD902, Vazyme Biotech, China) according to the manufacturer’s instructions [25]. Briefly, Concanavalin A-attached magnetic beads were used to bind the Pre and 14 dpi tissue samples, followed by resuspension in antibody buffer and incubation with rabbit anti-H4K12la primary antibody at 4 °C overnight with slow rotation. After incubation with goat anti-rabbit IgG H&L (AB206-01-AA) at room temperature for 1 h, the samples were incubated with protein-A-Tn5 (pA-Tn5) transposase before resuspending in tagmentation buffer for 1 h. Following transposon activation and tagmentation, fragmented DNA was isolated, expanded, and purified to build a library, which was subsequently sequenced on the Illumina NovaSeq6000 platform to produce 150-bp paired-end reads for further analysis. DiffBind was employed to detect differential peaks from peak sets. The statistically significant differential binding sites were identified based on binding affinity, with a threshold of p-value < 0.05, and |log2(Foldchange)|> 1. Benjamini and Hochberg FDR procedure was used for multiple testing of significance. The raw sequence data were analyzed by (OE Biotech, Shanghai China), and the target genes were studied. Furthermore, differential peaks in promoter regions were screened using Gene Ontology (GO) enrichment and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analyses based on the hypergeometric distribution.

ChIP-qPCR

The ChIP assay was performed using a ChIP-IT^® Express Enzymatic Magnetic Chromatin Immunoprecipitation Kit & Enzymatic Shearing Kit (Active Motif), as previously described [26]. Briefly, 1% formaldehyde solution was used to crosslink isolated Pre and 14 dpi tissues or cultured primary microglia for 30 min at 37 °C, followed by the addition of glycine for 10 min. Tissues or cells were cleaned three times with PBS and centrifuged at 2000 rpm for 10 min, followed by lysis, sonication, and immunoprecipitation in accordance with the manufacturer’s guidelines. Diluted chromatin was immunoprecipitated with H4K12la or PD-1 antibody, while normal rabbit IgG was used as a nonspecific control. The mixture was continuously rotated while incubating at 4 °C overnight. Subsequently, 50 μl protein A/G magnetic beads were added and incubated at 4 °C for 2 h. Wash buffer was used to clean the beads, and crosslinking was reversed by incubating them for 12 h at 50 °C in 1% SDS and 0.1 M NaHCO_3, as provided in the kit. DNA fragments were then purified using a QIAquick PCR purification Kit (QIAGEN). DNA collected via ChIP was detected via qPCR. The primer sequences for PD-1 in ChIP were as follows: forward primer (TGCATGTGTGTTGTGGGATA); reverse primer (TGCATGTGTGTTGTGGGATA). The primer sequences for GAPDH in ChIP were as follows: forward primer (AGCCTCCTCCAATTCAACCCTT); reverse primer (TGCATGTGTGTTGTGG GATA).

qPCR

Total RNA was extracted from tissues or cells using TRIzol Reagent (Invitrogen) according to the manufacturer’s protocol. Then, cDNA was synthesized using a one-step first-strand cDNA synthesis kit (Transgen). Gene expression was detected using SYBR Green-based qPCR. The primer sequences for PD-1 in qPCR were as follows: forward primer (TGCCCTAGTGGGTATC CCTG); reverse primer (GAAGGCTCCTCCTTCAGAGTGT). The primer sequences for GAPDH in qPCR were as follows: forward primer (AAGAGGGATGCTGCCCTTACC); reverse primer (CCCAATACGGCCAAATCCGT). Finally, the mRNA expression were standardized to those of GAPDH.

Detection of lactate levels

The lactate levels in spinal cord tissues were detected as previously described [27]. Briefly, spinal cord tissues were sonicated and homogenized using lysis buffer at 500 W (4 s on and 8 s off) for 4 min, followed by centrifugation at 15,000 rpm for 15 min to collect the supernatant. According to the manufacturer’s guidelines, the tissue lactate levels were tested using the ab65330 Lactate Assay Kit (Colorimetric/Fluorometric), with the lactate concentration shown in mM.

Basso mouse scale (BMS) score assessment

According to the protocol developed by Basso and colleagues, each mouse received a BMS score at 0, 3, 7, 14 and 28 dpi to evaluate the recovery of locomotor function [28]. Three experienced evaluators scored the mice in a double-blind manner. Utilizing the criteria for scoring (mobility of the posterior ankle joint, coordination, paw posture, trunk stability, and tail posture), the score ranged from 0 to 9, with a higher score indicating improved hindlimb function.

Footprint analysis

Footprint analysis was used to assess motor function recovery at 28 dpi. The mice walked along a paper-paved track, as previously mentioned [29]. The front paws and hindpaws are colored green, and red respectively. Qualitative analysis considered the stride length, width, and paw rotation.

Quantification and statistical analysis

All experimental data were expressed as the mean ± standard error of the mean (SEM) unless stated otherwise. Data analysis was performed using GraphPad Prism software (version 8.0). For multiple comparisons, one- or two-way analysis of variance (ANOVA) was performed using Tukey’s post hoc test. Student’s t-test was used for pairwise comparisons between the two groups. A p-value ≤ 0.05 was considered statistically significant and expressed as *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001.

Results

Lactate and lactylation levels elevate after SCI

After SCI, the hypoxic microenvironment regulates lactate levels at the injury site [12]. We detected changes in lactate levels in the injured spinal cord. The colorimetric assay revealed a significant increase in lactate levels at the lesion site from 7 to 28 dpi (Fig. 1A). Because lactate, as a precursor factor, may cause histone lactylation [30], we detected changes of in lactylation after SCI using western blotting. The results showed an increase in Pan Kla levels at 7, 14, and 28 dpi (Fig. 1B). It has been shown that microglia play a key role in the regulation of SCI pathobiology and recovery. Given that microglia are metabolically flexible and can be controlled by lactylation [31], we hypothesized that lactylation may be altered in microglia after SCI. Compared to Pre levels, the Pan Kla signal intensity was considerably increased in microglia from 7 to 28 dpi (Fig. 1C, D). These results revealed that overall lactate level was elevated at the lesion site, while Pan Kla was upregulated in the entire spinal cord and microglia after SCI.

Fig. 1

Changes in lactate and lactylation modification levels after SCI in vivo. A Lactate levels of the spinal cord at Pre, and 3, 7, 14, 28 dpi in vivo. n = 3 mice per group, *p < 0.05; ****p < 0.0001; N.S., not significant. B Western blot analysis of the Pan histone lactylation levels in the spinal cord at Pre, and 3, 7, 14, and 28 dpi. C Representative images of Pan Kla (green) co-stained with microglia (Iba1, magenta) at Pre, and 7, 14, and 28 dpi in vivo. White arrowheads indicate the colocalization of Pan Kla and microglia. Asterisks indicate the lesion site. Scale bars: low magnification, 100 μm; high magnification, 20 μm. D Quantification of Pan Kla intensity in Iba1⁺ microglia in (C). All data were presented as the mean ± SEM, n > 100 cells per group, ****p < 0.0001

H4K12la levels are elevated in microglia after SCI

Previous studies have highlighted the importance of site-specific histone lactylation in regulating macrophage polarization, tumor immunity, and antiviral responses [30, 32,33,34]. To investigate site-specific histone lactylation after SCI, we performed western blotting using antibodies against distinct types of histone lactylation. The results revealed an increase of H3K18la and H4K12la levels in the injured spinal cord (Fig. 2A–J). Importantly, the H4K12la level increased early at 7 dpi. To study the spatial expression pattern of H4K12la after SCI, we used co-immunostaining with H4K12la and spinal cord cell markers including CX3CR1 (microglia), PDGFRβ (pericyte), and GFAP (astrocyte) at 7 dpi. The results indicated that H4K12la was highly expressed in CX3CR1⁺ microglia and slightly expressed in GFAP⁺ astrocytes or PDGFRβ⁺ pericytes around the lesion site. H4K12la protein exhibited nuclear distribution in CX3CR1⁺ microglia at 7 dpi (Fig. 2K–N). Pearson’s correlation coefficient was used to validate the co-localization (Fig. S1). Microglia can undergo rapid proliferation with a peak at 7 dpi, before forming a protective microglial scar at 14 dpi [5]. Interestingly, H4K12la expression was notably elevated at 7 and 14 dpi, and was temporally correlated with microglial proliferation and scar formation. To verify whether H4K12la modification was altered in microglia after SCI, we performed immunofluorescence co-staining of microglial markers (CX3CR1 and Iba1) and H4K12la. Our results suggested that H4K12la signal intensity was significantly increased in microglia after SCI (Fig. 2O–R). Consistently, using CCR2^RFPCX3CR1^GFP dual-reporter mice, in which macrophages were labeled with RFP and microglia were labeled with GFP, we found that the intensity of H4K12la was significantly increased in microglia after SCI (Fig. S2). These results demonstrate the upregulation of H4K12la in microglia after SCI.

Fig. 2

The changes of site-specific histones lactylation after SCI in vivo. A–J Western blotting analysis of site-specific histones lactylation in spinal cord in vivo at Pre, and 3, 7, 14, and 28 dpi with quantification of protein levels. All data were presented as the mean ± SEM, n = 3 mice per group, *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001; N.S., not significant. K–M Representative immunofluorescence images for expression of H4K12la (green) and CX3CR1 (microglia, magenta) (K), H4K12la and PDGFRβ (pericyte, magenta) (L), or H4K12la and GFAP (astrocyte, magenta) (M) in injured spinal cords at 7 dpi in vivo. White arrowheads indicated the colocalization of H4K12la and CX3CR1. Asterisks indicate the lesion site, scale bars: low magnification, 100 μm; higher magnification, 20 μm. N Quantitative analysis of the percentage of H4K12la⁺CX3CR1⁺, H4K12la⁺PDGFRβ⁺ and H4K12la⁺GFAP⁺ cells in total H4K12la⁺ cells in injured spinal cords. All data are mean ± SEM, n = 3 mice per group, ****p < 0.0001 (H4K12la⁺PDGFRβ⁺ or H4K12la⁺GFAP⁺ vs. H4K12la⁺ CX3CR1⁺). O Representative images of H4K12la (green) co-stained with microglia (CX3CR1, magenta) at Pre, and 7, 14, and 28 dpi in vivo. White arrowheads indicated the colocalization of H4K12la and CX3CR1. Asterisks indicate the lesion site, scale bars: low magnification, 100 μm; high magnification, 20 μm. P Quantification of H4K12la intensity in (O). All data were presented as the mean ± SEM, n > 100 cells per group, ****p < 0.0001. Q Representative images of H4K12la (green) co-stained with microglia (Iba1, magenta) at Pre, and 7, 14, and 28 dpi in vivo. White arrowheads indicated the colocalization of H4K12la and Iba1. Asterisks indicate the lesion site, scale bars: low magnification, 100 μm; high magnification, 20 μm. R Quantification of H4K12la intensity in (Q). All data were presented as the mean ± SEM, n > 100 cells per group, ****p < 0.0001

Exogenous lactate treatment upregulates H4K12la levels in microglia after SCI

It has been reported that exogenous lactate stimulates histone lactylation in cells [30]. Hence, we next investigated whether exogenous lactate could elevate H4K12la levels in microglia after SCI. The injured mice were treated with lactate, 4-CIN (a monocarboxylate transporter 2 inhibitor that can obstruct lactate transport into cells), and OX (an inhibitor of lactate dehydrogenase that can inhibit lactate production) after SCI (Fig. 3A). We found that lactate treatment increased lactate levels in spinal cord, whereas 4-CIN treatment showed no significant changes in lactate levels, while OX treatment reduced lactate levels at 14 and 28 dpi (Fig. 3B). Western blotting analysis indicated that the H4K12la levels were significantly increased after lactate treatment at 14 and 28 dpi (Fig. 3C, D), whereas there were insignificant changes in H4K12la levels after 4-CIN treatment (Fig. 3E, F). Notably, OX treatment reduced H4K12la levels at 14 and 28 dpi (Fig. 3G, H). Furthermore, immunofluorescence assays confirmed that the H4K12la signal intensity in microglia was significantly increased after lactate treatment, whereas an insignificant difference was observed after 4-CIN treatment. OX treatment suppressed H4K12la signal intensity at 14 and 28 dpi (Fig. 3I, J). Collectively, these results demonstrated that exogenous lactate treatment increased the overall lactate level in the spinal cord and elevated H4K12la levels in the spinal cord and microglia after SCI.

Fig. 3

Exogenous lactate treatment upregulates H4K12la level in microglia after SCI in vivo. A Schedule described that mice were injected intraperitoneally (i. p.) with lactate, 4-CIN, or OX daily until 28 dpi. B Lactate levels of the spinal cord at 14 and 28 dpi in PBS (control), lactate, 4-CIN and OX groups in vivo. All data were presented as the mean ± SEM, n = 3 mice per group, ****p < 0.0001; N.S., not significant. C Western blotting analysis detected H4K12la expression levels of the spinal cord at 14 and 28 dpi in lactate group in vivo. D Quantification of H4K12la expression levels in (C). All data were presented as the mean ± SEM, n = 3 mice per group, **p < 0.01; ***p < 0.001. E Western blotting analysis detected the H4K12la expression levels of the spinal cord at 14 and 28 dpi in 4-CIN group in vivo. F Quantification of H4K12la expression levels in (E). All data were presented as the mean ± SEM, n = 3 mice per group, N.S. meant control compared with 4-CIN, not significant. G Western blotting analysis detected H4K12la expression levels of the spinal cord at 14 and 28 dpi in OX group in vivo. H Quantification of H4K12la expression levels in (G). All data were presented as the mean ± SEM, n = 3 mice per group, *p < 0.05; ***p < 0.001. I Representative images of H4K12la (green) co-stained with microglia (Iba1, magenta) at 14 and 28 dpi in lactate, 4-CIN, and OX groups in vivo. White arrowheads indicated the colocalization of H4K12la and microglia. Asterisks indicate the lesion site, scale bars: low magnification, 100 μm; high magnification, 20 μm. J Quantification of H4K12la intensity in (I). All data were presented as the mean ± SEM, n > 100 cells per group, ****p < 0.0001; N.S., not significant

Lactate treatment promotes microglial proliferation and scar formation after SCI

Microglia proliferate rapidly at 7 dpi, which contributes to microglial scar formation after SCI [5]. Considering the upregulation of H4K12la levels in microglia after SCI and the temporal and spatial correlation of this change with microglial proliferation and scar formation, we further investigated the effect of exogenous lactate on microglial proliferation and scarring. Our results showed that the number of BrdU⁺CX3CR1⁺ cells was considerably augmented following lactate treatment, did not change significantly following 4-CIN treatment, and reduced following OX treatment at 7 dpi (Fig. 4A, B). Immunofluorescence analysis revealed an increase in the CX3CR1⁺ microglial scar after lactate treatment, a non-significant difference after 4-CIN treatment, and a decrease after OX treatment at 28 dpi (Fig. 4C, D). These results confirmed that lactate treatment promoted microglial proliferation, thereby contributing to microglial scar formation after SCI.

Fig. 4

Lactate treatment promotes microglia proliferation and scar formation after SCI in vivo. A Representative immunofluorescence images of microglia proliferation by double-staining BrdU (magenta) and CX3CR1 (green) in control, lactate, 4-CIN, and OX groups at 7 dpi in vivo. Scale bar: 20 μm. B Quantification of the density of BrdU⁺CX3CR1⁺ cells in (A). All data were presented as the mean ± SEM, n = 3 mice per group, ****p < 0.0001; N.S., not significant. C Immunofluorescence staining of CX3CR1 (red) in control, lactate, 4-CIN, and OX groups at 28 dpi in vivo. Asterisks indicate the lesion site, scale bar: 100 μm. D Quantification of the density of CX3CR1⁺ cells in (C). All data were presented as the mean ± SEM, n = 3 mice per group, ***p < 0.001; N.S., not significant

Lactate treatment promotes axon regeneration and locomotor function recovery after SCI

Previous studies have demonstrated that fibrotic scars hinder axon regeneration, astrocytic scars may either hinder or aid axon regeneration depending on their physical orientation, and microglial scars promote axon regeneration after SCI [3, 4, 7, 35, 36]. In addition, proliferated microglia prevent the diffusion of inflammatory cells and maintain residual viable neurons away from the lesion site [5, 9, 37]. After confirming that lactate treatment facilitated microglial scarring following SCI, we focused on the effects of lactate on residual neurons. We used the neuron marker NeuN to detect residual neurons in specific regions (Z1_Z3) of viable neurons located in the epicenter and edges after SCI [24]. The results illustrated that lactate treatment significantly increased the number of residual NeuN⁺ neurons in the Z1_Z3 regions, whereas 4-CIN treatment slightly reduced the counts of residual neurons. In contrast, OX treatment significantly reduced the counts of residual neurons in the Z1_Z3 regions at 28 dpi (Fig. 5A, B). We then detected descending serotonergic (5-HT) axon regeneration via immunofluorescence at 28 dpi. Lactate treatment obviously increased the number of 5-HT axons penetrating into the lesion site compared to 4-CIN and OX treatment (Fig. 5C, D).

Fig. 5

Lactate treatment promotes axon regeneration and locomotor function recovery after SCI in vivo. A Images of astrocytes (GFAP, green) stained with neurons (NeuN, magenta) in Zone1-Zone3 (Z1_Z3) at 28 dpi in control, lactate, 4-CIN, and OX groups in vivo. Z1 (0–250 μm) was in the lesion site, Z2 (250–500 μm) was started immediately adjacent to Z1, and Z3 (500–750 μm) was adjacent to Z2. Scale bars: magnification, 100 μm. Compass: D, dorsal; V, ventral; R, rostral; C, caudal. B Quantification of NeuN⁺ neurons in three zones (Z1_Z3) in control, lactate, 4-CIN, and OX groups. All data were presented as the mean ± SEM, n = 3 mice per group, *p < 0.05; **p < 0.01. C Images of GFAP (green) co-stained with axons (5-HT, magenta) at 28 dpi in control, lactate, 4-CIN, and OX groups in vivo, and the boxed region indicated the lesion center magnified in right panels. Asterisks indicate the lesion site, scale bars: low magnification, 100 μm; high magnification, 20 μm. C Quantification of the percentage of 5-HT⁺ area in the area of the spinal cord segment penetrating into the lesion core at 28 dpi in vivo. All data are presented as the mean ± SEM, n = 3 mice per group, *p < 0.05; ***p < 0.001; N.S., not significant. D Schedule of locomotor function analysis and the intraperitoneal injection with lactate, 4-CIN or OX daily until 28 dpi. E Locomotor function was evaluated by BMS at the indicated time points. All data are presented as the mean ± SEM, n = 8 mice per group, *p < 0.05; ***p < 0.001; ****p < 0.0001. F Representative images of footprint analysis in control, lactate, 4-CIN, and OX groups at 28 dpi. H–J Quantification of the stride length and width and paw rotation at 28 dpi. All data are presented as the mean ± SEM, n = 8 mice per group; *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001; N.S., not significant

To examine whether exogenous lactate-mediated axon regeneration promoted locomotor function recovery, we conducted behavioral tests following SCI. In a chronic experiment, mice were administered daily intraperitoneal lactate, 4-CIN, and OX injections for 28 consecutive days after SCI (Fig. 5E). Lactate treatment significantly improved locomotor function recovery, whereas 4-CIN treatment had no visible effect. Conversely, OX treatment hindered functional recovery as indicated by the BMS scores for behavioral analysis from 0 to 28 dpi (Fig. 5F). The footprint analysis yielded similar outcomes (Fig. 5G–J). According to the aforementioned results, lactate treatment could promote axon regeneration and locomotor function recovery after SCI.

H4K12la modifications activate PD-1 transcription in microglia

It has been demonstrated that histone lactylation modification contributes to the transcriptional regulation of downstream genes [30]. To identify the H4K12la-target gene after SCI, we used genome-wide CUT&Tag analysis to identify candidate genes regulated by H4K12la. CUT&Tag analysis utilizing anti-H4K12la antibodies showed that H4K12la peaks were enriched at 14 dpi (Fig. 6A). H4K12la was enriched in promoter domains, and the Pre sample showed 14,463 H4K12la binding peaks, with 60.51% of the peaks situated within promoter sequences (≤ 3 Kb). At 14 dpi, 16043 H4K12la binding peaks were detected, with 50.32% situated within the promoter sequences (Fig. 6B). To further identify potential downstream targets of H4K12la, we screened 37 genes with upregulated H4K12la expression in the promoter region for 37 genes at 14 dpi. Among these upregulated genes, PD-1 was upregulated after SCI and predominantly expressed in Iba1⁺ cells [38]. PD-1 has also been shown to promote microglial proliferation in the tumor microenvironment [39]. Importantly, a previous study also preliminarily confirmed the pro-repair effect of PD-1 on SCI [40, 41]. These findings suggested that lactate-mediated lactylation partly exerts a beneficial effect by regulating PD-1 gene transcription in Iba1⁺ microglia after SCI. Further analysis showed that the H4K12la level at the PD-1 promoter region was significantly elevated in the spinal cord at 14 dpi (Fig. 6C). To further confirm that PD-1 transcription was activated by H4K12la, we performed ChIP analysis using antibodies against H4K12la and found that the H4K12la level on the PD-1 promoter was significantly elevated in the spinal cord at 14 dpi (Fig. 6D). Consistently, qPCR analysis showed that PD-1 expression was elevated at 14 dpi (Fig. 6E). Taken together, these results indicate that H4K12la modifications in promoter areas activate PD-1 transcription following SCI.

Fig. 6

H4K12la activates PD-1 transcription in the injured spinal cord in vivo and microglia in vitro. A The binding density of H4K12la was visualized by deep Tools: the heatmap presents the CUT&Tag tag counted on the H4K12la binding peaks in the spinal cord at Pre and 14 dpi in vivo, ordered by signal strength. B Genome-wide distribution of H4K12la-binding peaks in the spinal cord at Pre and 14 dpi in vivo. C Genome browser tracks of CUT&Tag signal at the PD-1 loci. D ChIP analysis of the indicated promoters was performed using antibodies against H4K12la in the spinal cord at Pre and 14 dpi in vivo. All data are presented as the mean ± SEM, n = 3 mice per group, ****p < 0.0001. E qPCR assays monitoring PD-1 expression in the spinal cord at Pre and 14 dpi in vivo. All data are presented as the mean ± SEM, n = 3 mice per group, **p < 0.01. F ChIP analysis of the indicated promoters was performed using antibodies against H4K12la in primary microglia treated with PBS (control) and lactate in vitro. All data are presented as the mean ± SEM, n = 3 per group, ***p < 0.001. G qPCR assays monitoring expression of the PD-1 in primary microglia treated with PBS (control) and lactate in vitro. All data are presented as the mean ± SEM, n = 3 per group, **p < 0.01

Lactate promotes H4K12la expression in microglia, and H4K12la activates the transcription of PD-1 after SCI. We further explored the impact of H4K12la on the transcription of PD-1 in microglia. The results of ChIP analysis illustrated that the H4K12la level on the PD-1 promoter was increased in microglia treated with lactate (Fig. 6F). Consistent with H4K12la-mediated transcriptional activation in the SCI model, qPCR revealed that the expression of PD-1 was significantly elevated in microglia treated with lactate (Fig. 6G). Collectively, these results demonstrated that H4K12la modification activated PD-1 transcription in microglia. Additional analyses of genes with differential promoter lactylation are presented in Fig. S3.

Intraperitoneal injection of a PD-1 inhibitor suppresses microglial proliferation and scar formation, hinders axon regeneration, and prevents locomotor function recovery after SCI

Our results showed that lactate promoted H4K12la microglial proliferation and scar formation after SCI. Considering PD-1 activation by H4K12la, we next explored the influence of PD-1 on microglial proliferation and scar formation after SCI. The western blotting results showed that PD-1 expression was elevated at 7, 14 and 28 dpi (Additional file 1: Fig. S4A, B), while the immunofluorescence assays revealed that the fluorescence intensity of PD-1 was increased in microglia after SCI (Additional file 1: Fig. S4C–F). These results are consistent with previous findings [40]. We then explored the effect of PD-1 on microglial proliferation and scar formation after SCI by inhibiting PD-1/PD-L signaling with a PD-1 inhibitor [23]. Mice were intraperitoneally injected with PD-1 inhibitor at 1 and 14 dpi and sacrificed at 28 dpi (Fig. 7A). The effect of PD-1 inhibitors on microglial proliferation was assessed at 7 dpi, when the peak microglial proliferation was observed [5]. CX3CR1 cells were co-stained with BrdU to identify microglial proliferation. The results showed that the number of BrdU⁺CX3CR1⁺ cells was lower in PD-1 inhibitor group than in control group (Fig. 7B, C). Moreover, there was a significant decrease in microglial scar formation in PD-1 inhibitor group at 28 dpi (Fig. 7D, E).

Fig. 7

Intraperitoneal injection of PD-1 inhibitor suppresses microglial proliferation and scar formation, hinders axon regeneration, and prevents locomotor function recovery after SCI in vivo. A The schedule described that the mice were treated with PD-1 inhibitor at 1 and 14 dpi via intraperitoneal (i. p.) injection and were sacrificed at 28 dpi. B Immunofluorescence staining of BrdU (magenta), CX3CR1 (green) in sagittal sections of the PBS (control) and PD-1 inhibitor groups at 7 dpi in vivo. Scale bar: 20 μm. C Quantification of the density of BrdU⁺CX3CR1⁺ cells in (B). All data are presented as the mean ± SEM, n = 3 mice per group, **p < 0.01. D Immunofluorescence staining of CX3CR1 (red) in control and PD-1 inhibitor groups at 28 dpi in vivo. Scale bar: 100 μm. E Quantification of the density of CX3CR1⁺ cells in (D). All data are presented as the mean ± SEM, n = 3 mice per group, **p < 0.01. F Images of astrocytes (GFAP⁺, green) stained with neurons (NeuN⁺, magenta) in Zone1-Zone3 (Z1_Z3) at 28 dpi in PD-1 inhibitor and control groups in vivo. Scale bar: 100 μm. G Quantification of NeuN⁺ neurons in three zones (Z1_Z3) in (F). All data are presented as the mean ± SEM, n = 3 mice per group, **p < 0.01. H Images of GFAP (green) co-stained with axons (5-HT, magenta) at 28 dpi in PD-1 inhibitor and control groups in vivo. Asterisks indicate the lesion site, scale bars: low magnification, 100 μm; high magnification, 20 μm. I Quantification of the percentage of 5-HT⁺ area in the area of the spinal cord segment penetrating into the lesion core at 28 dpi in (H). All data are presented as the mean ± SEM, n = 3 mice per group, **p < 0.01. J Locomotor function was evaluated by BMS at the indicated time points in PD-1 inhibitor and control groups. All data are presented as the mean ± SEM, n = 8 mice per group, *p < 0.05; ***p < 0.001; ****p < 0.0001. K Representative images of footprint analysis in PD-1 inhibitor and control groups at 28 dpi. L, N Quantification of the paw rotation, stride width and stride length at 28 dpi. All data are presented as the mean ± SEM, n = 8 mice per group, **p < 0.01; ****p < 0.0001

To investigate the influence of PD-1 inhibition on residual neurons, we next used NeuN immunostaining to assess neuronal survival in particular zones (Z1_Z3) surrounding the lesion core at 28 dpi. The number of residual NeuN⁺ neurons in the Z1_Z3 region was considerably lower in the PD-1 inhibitor group than in the control group (Fig. 7F, G). We then detected descending 5-HT axon regeneration at 28 dpi. In the lesion site, significantly fewer 5-HT axons penetrating intothe lesion site in the PD-1 inhibitor group than in the control group (Fig. 7H, I). Next, BMS score and footprint analyses were performed to identify locomotor function recovery following SCI. The mice injected with the PD-1 inhibitor exhibited worse hindlimb locomotor function than those in the control group, corresponding to the results of BMS score (Fig. 7J). Accordingly, footprint analysis showed that mice in the PD-1 inhibitor group had worse locomotor function than those in the control group after SCI (Fig. 7K–N). These results confirmed positive role of PD-1 in microglial proliferation, scar formation, axon regeneration, and locomotor functional recovery after SCI.

PD-1 inhibitor reverses the advantageous effects of lactate on SCI repair

To further verify whether PD-1 participates in the lactate/H4K12la signaling pathway after SCI, rescue assays were performed. We first detected microglial proliferation with BrdU via immunofluorescence staining. The results showed that lactate-mediated microglial proliferation was dramatically reversed by PD-1 inhibitor at 7 dpi (Fig. 8A, B). The PD-1 inhibitor abolished the positive effect of lactate on microglial scar formation at 28 dpi (Fig. 8C, D). The number of residual NeuN⁺ neurons in the Z1_Z3 region of the spinal cord was increased in the lactate group, whereas the PD-1 inhibitor reversed the lactate-mediated increase in residual neurons (Fig. 8E, F). Additionally, the PD-1 inhibitor reversed the lactate-induced increase in 5-HT axons penetrating into the lesion site (Fig. 8G, H). Behavioral analysis experiments were conducted to observe improvement in locomotor function after SCI. Lactate treatment promoted locomotor function recovery, and PD-1 inhibition reversed lactate-mediated locomotor function recovery as shown in the BMS score and footprint analysis after SCI (Fig. 8I–M). These results confirmed that PD-1 inhibitors reversed the promotion of lactate on microglial proliferation, microglial scar formation, axon regeneration, and locomotor function recovery, supporting the role of PD-1 as a downstream regulator of lactate/H4K12la signaling after SCI.

Fig. 8

PD-1 inhibitor reverses the advantageous effects of lactate on SCI repair in vivo. A Immunofluorescence staining of BrdU (green) and CX3CR1 (magenta) in sagittal sections of the PBS (control), lactate, and lactate + PD-1 inhibitor groups at 7 dpi in vivo. Scale bar: 20 μm. B Quantification of the density of BrdU⁺CX3CR1⁺ cells in (A). All data are presented as the mean ± SEM, n = 3 mice per group, **p < 0.01; ***p < 0.001. C Immunofluorescence staining of CX3CR1 (red) in sagittal sections of the control, lactate, and lactate + PD-1 inhibitor groups at 28 dpi in vivo. Scale bar: 100 μm. D Quantification of the density of CX3CR1⁺ cells in (C). All data are presented as the mean ± SEM, n = 3 mice per group, ***p < 0.001; ****p < 0.0001. E Images of astrocytes (GFAP, green) stained with neurons (NeuN, magenta) in Z1_Z3 at 28 dpi in control, lactate, and lactate + PD-1 inhibitor groups in vivo. Scale bar: 100 μm. F Quantification of NeuN⁺ neurons in Z1_Z3 in (E). All data are presented as the mean ± SEM, n = 3 mice per group, **p < 0.01; ***p < 0.001; ****p < 0.0001. G Images of GFAP (green) co-stained with axons (5-HT, magenta) at 28 dpi in control, lactate, and lactate + PD-1 inhibitor groups in vivo. Asterisks indicate the lesion site, scale bars: low magnification, 100 μm; high magnification, 20 μm. H Quantification of the percentage of 5-HT⁺ area in the area of the spinal cord segment penetrating into the lesion core at 28 dpi in (G). All data are presented as the mean ± SEM, n = 3 mice per group, ****p < 0.0001. I The locomotor function was evaluated by BMS at the indicated time points. All data are presented as the mean ± SEM, n = 8 mice per group, **p < 0.01; ****p < 0.0001. J Representative images of footprint analysis in mice in control, lactate, and lactate + PD-1 inhibitor groups at 28 dpi. K–M Quantification of the stride length and width and paw rotation at 28 dpi. All data are presented as the mean ± SEM, n = 8 mice per group, **p < 0.01; ***p < 0.001; ****p < 0.0001

PD-1 suppression in microglia hinders axon regeneration, and inhibits functional recovery after SCI

To investigate the effects of microglial PD-1 inhibition on axonal regeneration and function recovery after SCI, we injected microglial-specific AAV-sh-PD-1 into the injured spinal cord immediately after SCI [42]. Western blotting revealed that the expression of PD-1 in the AAV-sh-PD-1 group was significantly lower than that in the AAV-Con group at 14 and 28 dpi (Additional file 1: Fig. S5A, B). Immunofluorescence staining was also performed to measure the intensity of PD-1 in microglia in mice administered AAV-sh-PD-1. The results illustrated that the intensity of PD-1 in microglia in mice injected with AAV-sh-PD-1 was significantly lower than that in mice injected with AAV-Con at 28 dpi (Additional file 1: Fig. S5C–F). Importantly, we observed reduced penetration of 5-HT axons into the lesion site in mice injected with AAV-sh-PD-1 compared to those injected with AAV-Con at 28 dpi (Fig. 9A, B). To investigate the influence of microglial PD-1 suppression on viable neurons, we stained the neuronal marker NeuN in specific regions (Z1_Z3) around the lesion core. The results revealed that mice injected with AAV-sh-PD-1 substantially reduced viable neurons compared to those injected with AAV-Con at 28 dpi (Fig. 9C, D). Footprint analysis and BMS scoring showed that the mice injected with AAV-sh-PD-1 exhibited reduced locomotor function recovery (Fig. 9E–I). These results indicate that interruption of the lactate/H4K12la/PD-1 signaling pathway via suppression of PD-1 in microglia hinders axon regeneration and functional recovery after SCI.

Fig. 9

PD-1 suppression in microglia hinders axon regeneration and inhibits functional recovery after SCI in vivo. A Images of GFAP (green) co-stained with axons (5-HT, magenta) in AAV-Con and AAV-sh-PD-1 groups at 28 dpi in vivo. Asterisks indicate the lesion site, scale bars: low magnification, 100 μm; high magnification, 20 μm. B Quantification of the percentage of 5-HT⁺ area in the area of the spinal cord segment spanning the lesion site in (A). All data are presented as the mean ± SEM, n = 3 mice per group, **p < 0.01. C Representative immunofluorescence images of GFAP (green) and NeuN (magenta) neurons in Z1-Z3 zones adjacent to central lesion site in AAV-Con and AAV-sh-PD-1 mice at 28 dpi in vivo. Scale bar: magnification, 100 μm. D Quantification of NeuN⁺ neurons in Z1_Z3 zones adjacent to lesion site in AAV-Con and AAV-sh-PD-1 groups in (C). All data are presented as the mean ± SEM, n = 3 mice per group, **p < 0.01; ***p < 0.001; ****p < 0.0001. E Representative images of footprint analysis in mice treated with AAV-Con or AAV-sh-PD-1 at 28 dpi in vivo. F–H Quantification of the stride length and width and paw rotation. All data are presented as the mean ± SEM, n = 8 mice per group, **p < 0.01; ***p < 0.001. I Locomotor function was evaluated by BMS at the indicated time points in mice treated with AAV-Con or AAV-sh-PD-1. All data were presented as the mean ± SEM, n = 8 mice per group, *p < 0.05; **p < 0.01; ****p < 0.0001

Suppression of PD-1 expression in microglia reverses the advantageous effects of lactate on SCI repair

To further explore the effect of microglial PD-1 suppression on lactate/H4K12la signaling, a series of rescue experiments were designed. We performed immunofluorescence staining with BrdU to detect proliferation, and the results showed that microglial proliferation was increased in mice injected with lactate, whereas AAV-sh-PD-1 obviously reversed the promotion of microglial proliferation induced by lactate at 7 dpi (Fig. 10A, B). We next assessed whether AAV-sh-PD-1 could abolish the effect of lactate on microglial proliferation in vitro. The lactate-mediated primary microglial proliferation was reversed by AAV-sh-PD-1 infection (Fig. 10C, D). Additionally, the microglial scarring in mice injected with lactate was significantly reversed by AAV-sh-PD-1 administration at 28 dpi (Fig. 10E, F). The results of immunofluorescence staining showed that the lactate-mediated increase in residual neurons was reversed by AAV-sh-PD-1 administration at 28 dpi (Fig. 10G, H). AAV-sh-PD-1 administration reversed the lactate-mediated increase in 5-HT axon at 28 dpi (Fig. 10I, J). The results of footprint analysis and BMS scoring showed that AAV-sh-PD-1 administration significantly reversed lactate-mediated locomotor functional recovery after SCI (Fig. 10K–O). Taken together, these results demonstrated that AAV-sh-PD-1 reversed the beneficial effects of lactate on microglial proliferation, scar formation, axon regeneration, and locomotor function recovery after SCI, supporting PD-1 as a target of lactate/H4K12la signaling in microglia after SCI.

Fig. 10

Inhibition of PD-1 expression in microglia rescues the beneficial effects of lactate on microglial proliferation, scar formation, axon regeneration, and locomotor function recovery after SCI in vivo. A Immunofluorescence staining of BrdU (magenta) and CX3CR1 (green) in sagittal sections of the control, lactate, and lactate + AAV-sh-PD-1 groups at 7 dpi in vivo. Scale bar: 20 μm. B Quantification of the density of BrdU⁺CX3CR1⁺ cells in (A). All data are presented as the mean ± SEM, n = 3 mice per group, **p < 0.01; N.S., not significant. C Immunofluorescence staining of BrdU (red) in microglia treated with PBS (control), lactate, and lactate + AAV-sh-PD-1 in vitro. Scale bar: 20 μm. D Quantification of the density of BrdU⁺CX3CR1⁺ cells in (C). n = 3 per group, **p < 0.01; N.S., not significant. E Immunofluorescence staining of CX3CR1 (red) in sagittal sections from control, lactate, and lactate + AAV-sh-PD-1 groups at 28 dpi in vivo. Scale bar: 100 μm. F Quantification of the density of CX3CR1⁺ cells in (E). All data are presented as the mean ± SEM, n = 3 mice per group, **p < 0.01; ***p < 0.001; N.S., not significant. G Representative immunofluorescence images of GFAP (green) NeuN (magenta) neurons in Z1_Z3 zones adjacent to central lesion core in the control, lactate, and lactate + AAV-sh-PD-1 groups at 28 dpi in vivo. Scale bar: 100 μm. H Quantification of NeuN⁺ neurons in Z1_Z3 in (G). All data were presented as the mean ± SEM, n = 3 mice per group, *p < 0.05; **p < 0.01; ****p < 0.0001; N.S., not significant. I Images of GFAP (green) co-stained with axons (5-HT, magenta) at 28 dpi in control, lactate, and lactate + AAV-sh-PD-1 groups in vivo. Asterisks indicate the lesion site, scale bars: low magnification, 100 μm; high magnification, 20 μm. J Quantification of the percentage of 5-HT⁺ area in the area of the spinal cord segment spanning the lesion core at 28 dpi in (I). All data are presented as the mean ± SEM, n = 3 mice per group, **p < 0.01; N.S., not significant. K Representative images of footprint analysis in control, lactate, and lactate + AAV-sh-PD-1 groups at 28 dpi. (L–N Quantification of the stride length and width and paw rotation at 28 dpi. All data are presented as the mean ± SEM, n = 8 mice per group, *p < 0.05; **p < 0.01; ***p < 0.001; N.S., not significant. O BMS scores during 28 days of recovery after SCI demonstrate better functional recovery in control, lactate and lactate + AAV-sh-PD-1 groups. All data are presented as the mean ± SEM, n = 8 mice per group, ^###meant lactate compared to lactate + AAV-sh-PD-1, ^###P < 0.001; ^####meant lactate compared to lactate + AAV-sh-PD-1, ^####p < 0.0001; ***meant lactate compared to the control, *p < 0.05; ***p < 0.001

Discussion

Histone lactylation has been widely reported to be closely associated with disease onset and development, such as AD, lung fibrosis, and post-myocardial infarction [16, 33, 42, 43]. SCI generates a hypoxic microenvironment in the injured area, resulting in active glycolysis and the generation of substantial amounts of lactate as a substrate for histone lactylation [44, 45]. We therefore proposed that histone lactylation was aberrant in the injured spinal cord and focused on the role and mechanisms of histone lactylation after SCI. In the present study, we discovered an elevation in the overall lactate level in the spinal cord together with an increase in globle lactylation in both the spinal cord and microglia after SCI. Our results demonstrated that H4K12la levels were significantly upregulated in microglia after SCI. Lactate treatment increased overall lactate level in the spinal cord and upregulated H4K12la expression in both the spinal cord and microglia. Functionally, lactate promoted microglial proliferation, scar formation, axon regeneration, and locomotor function recovery after SCI. Mechanically, lactate-mediated H4K12la promoted transcription of PD-1 in microglia, which was further demonstrated to induce microglial proliferation and thereby contribute to microglial scar formation and functional recovery after SCI. Our study illustrates the function and mechanism of the lactate/H4K12la/PD-1 signaling pathway in microglia-mediated SCI repair and identifies a novel target for SCI treatment.

Scars play complex roles in injury repair and axon regeneration after SCI. Fibrotic scars serve as barriers that inhibit axon regeneration and can result in permanent functional deficits [40]. Astrocytic scars have been a much-debated topic for SCI and repair. Traditionally, astrocytic scars were regarded as obstacles to axon regeneration [46], yet they can also act as bridges to promote axon regeneration. Indeed, regenerating 5-HT or corticospinal axons often closely connect with astrocyte processes when crossing the lesion core after SCI [47,48,49,50]. A previous study reported that microglia proliferate and gather in large numbers around the lesion site, forming a protective microglial scar between the fibrotic and astrocytic scars at the lesion edge [5]. The microglia depletion results in enhanced immune cell infiltration, decreased neuronal survival, and impaired functional recovery after SCI [5, 37]. In contrast, local application of colony-stimulating factor at the lesion site to promote microglial proliferation also facilitated axon regeneration and functional recovery after SCI [51, 52]. Consistent with these findings, our results showed that lactate-mediated lactylation promoted microglial proliferation and accumulation leading to microglial scarring, thereby facilitating axon regeneration and functional recovery following SCI. However, it remains unclear whether this injury repair was a direct effect of lactate or an indirect effect via microglia. Additionally, each type of fibrotic, astrocytic, and microglial scar cell interacts with each other, contributing to permanently remodeled tissue [3]. Particularly, microglia are considered the vital cell type orchestrating intercellular crosstalk and have been demonstrated to regulate border formation of astrocytic scars, potentially by controlling astrocyte proliferation, adhesion, and gliosis [51]. Microglial depletion disrupts the density and contiguous boundary of fibrotic and astrocytic scars, leading to widespread inflammation in the lesion core [5]. Moreover, microglia can contribute to the absence of fibrotic and astrocytic scars, thereby coordinating scar-free healing after SCI in neonates [53]. In this context, lactate-induced regulation of microglial scar formation may be involved in the interconnection among different scar types. To clarify changes in lactate expression in microglia, the widely reported markers Iba1 and CX3CR1, as well as CCR2^RFPCX3CR1^GFP dual-reporter mice, were used to distinguish microglia from macrophages. The differences in the series of cellular functional changes induced by variations in lactylation between microglia and macrophages need to be elucidated using transgenic methods in subsequent studies. Nevertheless, our results supported the positive role of lactate in SCI repair.

Growing evidence suggests that lactate is a neuroprotective energy substrate with a protective effect against excitotoxicity [54, 55]. Exogenous lactate injection also improved neurological function in CNS injury [56,57,58]. In the CNS, microglia are crucial early responders to injury, constantly monitor the microenvironment for changes resulting from injury or disease, and are essential in tissue remodeling [59, 60]. Lactate-induced regulation of microglial clearance ability may expedite the elimination of cellular debris and enhance tissue repair while reducing inflammation in the ischemic area [13]. Lactate treatment may reduce neuroinflammation by altering the polarization of microglia [61]. Our results suggested that lactate promoted microglial proliferation after SCI; however, we did not explore the effects of lactate on phagocytosis and polarization of microglia, pending further testing in subsequent studies, which represents a shortcoming of the present study. Lactate production is determined by the equilibrium between glycolysis and mitochondrial metabolism; enzymes activity in these pathways can regulate lactate levels, which modulate histone lactylation [62]. It has been reported that a lactate dehydrogenase inhibitor (OX) reduced lactate production and histone lactylation induced by hypoxia [30]. Similarly, we found that OX treatment reduced lactate levels and histone lactylation after SCI. Furthermore, monocarboxylate transporters (MCTs) are known to mediate lactate release and uptake and are widely expressed in various tissues, including the liver, brain, and heart [63]. A previous study revealed that 4-CIN blocked lactate transport into neurons [34]. However, treatment with 4-CIN (an MCT2 inhibitor) had no obvious effects on microglia, which is to be expected given that MCT2 is almost exclusively present in neurons [13]. Moreover, the expression of lactate transporter proteins MCT1 and MCT4 was specifically increased in LPS-stimulated primary and BV2 microglia. We speculated that the 4-CIN target MCT2 might not be expressed in microglia after SCI. Thus, the key role of MCTs in microglia following SCI requires further investigation.

Lactate-derived lysine lactylation in histones serves as a novel histone mark, and at least 28 lactylation spots on core histones have been identified, including H4K12, H3K18, H3K23, H4K5, and H4K8 [30]. Lactate affects gene transcription through histone lactylation,a unique posttranslational modification [64]. For example, in the late phase of proinflammatory macrophage polarization, H3K18la in promoter areas triggered the expression of homeostatic genes involved in wound healing [30]. H3K18la in promoter areas hastened tumorigenesis by activating YTHDF2 transcription, thereby representing potential histone lactylation targets for ocular melanoma treatment [33]. In this study, we detected enhanced H3K18la expression at 14 and 28 dpi; however, we did not investigate the function and mechanism of H3K18la after SCI, which can be further explored in subsequent studies. Furthermore, H4K12la in promoter areas activated PKM transcription in AD microglia, demonstrating that the positive feedback loop of glycolysis/H4K12la/PKM2 in microglia accelerates the onset of AD and illustrated that inhibiting this loop may provide a novel therapeutic strategy for AD treatment [16]. Our study represents the first to demonstrate that H4K12la in promoter areas induces PD-1 expression in microglia and promotes SCI repair. However, additional research is required to determine whether other histone modifications contributed to PD-1 transcription.

PD-1, a 288-amino acid type-I transmembrane protein belonging to the CD28 superfamily, is expressed in immune cells, glial cells, and neurons, enabling several levels of immunomodulatory and neuromodulatory activity in the CNS [65, 66]. PD-1 signaling may be crucial in microglia, peripherally recruited immune cells, and neurons within the CNS under both normal and pathological conditions [65]. Previous study showed that PD-1 protein was expressed mainly on macrophages/microglia after SCI, and PD-1 modulated macrophage and microglial phenotypes after SCI [67]. Moreover, dexmedetomidine has been shown to mitigate microglia-mediated neuroinflammation through PD-1 upregulation in an SCI model [41]. Furthermore, PD-1 inhibitors have been shown to inhibit microglial proliferation and induce glioma apoptosis, suggesting that PD-1 promotes microglial proliferation [68]. Consistently, our research revealed that PD-1 expression was elevated in microglia after SCI, and PD-1 inhibition suppressed microglial proliferation, scar formation, axon regeneration, and functional recovery. Certainly, we cannot completely rule out the potential role of the lactate/H4K12la/PD-1 signaling pathway in other cell types in the SCI microenvironment. Because of its expression in many cell types, the effects of the PD-1 inhibitor observed in this study cannot be specifically attributed to microglia alone. Therefore, we next used AAV-sh-PD-1 to specifically suppress PD-1 in microglia, which blocked axon regeneration and function recovery after SCI. Importantly, we designed and conducted rescue assays, and the results indicated that both PD-1 inhibitor and AAV-sh-PD-1 administration reversed the beneficial effects of lactate on SCI, supporting that lactate-mediated SCI repair may be involved in the regulation of H4K12la/PD-1 in microglia. However, our AAV transduction faces possible neuronal trophism and the potential of increasing immune activation [69]. In addition, the specific mechanisms by which PD-1 affects microglial proliferation were not further clarified in this study. We speculate that PD-1 promotes microglial proliferation by directly activating proliferation-related pathways, as previous studies have shown that directly blocking PD-1 in microglia in vitro can inhibit their proliferation [70].

In this study, we used female mice for all experiments to avoid increasing the risk of urinary tract infections due to difficulty with urination, which does represent a limitation of our study. Although previous research on SCI, including studies on microglia, has commonly used female animal models [71,72,73], it is well known that microglia are sexually dimorphic [74]. Moreover, Andrew N. Stewart et al. found differences in the number of macrophages and individual gene expression levels in macrophages/microglia between sexes after SCI [75], while another clinical study reported that the numbers of male and female patients with SCI is comparable [76]. These findings suggest that the role of sex in pathology and treatment should be considered in future research. In the future, we will use PD-1-deficient mice to further confirm our findings. Given the mature clinical application of targeting PD-1, future studies should investigate the relationship between targeting PD-1 and the functional recovery in patients with SCI.

Conclusion

We describe an unrecognized lactate/H4K12la/PD-1 signaling pathway involved in microglia-mediated SCI repair. Lactate induces H4K12la levels, thereby promoting PD-1 transcription in microglia, further resulting in enhanced microglial scar formation and functional recovery after SCI. Our study provides a previously unknown insight by demonstrating that post-translational protein modifications derived from lactate exert vital functions and contribute to microglia-mediated tissue repair after CNS injury.