Integrative Analysis of m⁶A Methylation and Transcriptomic Profiles in Cortical Tissues of Mice with Sepsis-Associated Encephalopathy

Abstract

Background

Sepsis-associated encephalopathy (SAE) is a common and severe complication in septic patients, characterized by the increased mortality and cognition dysfunction. N6-Methyladenosine (m⁶A), one of the most prevalent RNA modifications, plays crucial roles in the central nervous system by regulating RNA metabolism. However, its involvement in the pathogenesis of SAE remains unexplored.

Methods

RNA sequencing (RNA-seq) and methylated RNA immunoprecipitation (MeRIP)-sequencing were performed to analyze the transcriptomic profiles and m⁶A methylation in the brain cortex of LPS-induced septic mice. Bioinformatics analysis was carried out to identify genes with altered m⁶A methylation and differential expression. Target genes and m⁶A regulators were further validated using quantitative reverse transcription polymerase chain reaction (RT-qPCR).

Results

Numerous genes exhibited differential expression and m⁶A modification in the cortical tissues of septic mice. Gene function enrichment analysis revealed that genes with differential expression were enriched in pathways related to immune and inflammatory responses as well as neuroactive ligand-receptor interaction, and genes with differential m⁶A modification showed significant associations with pathways related to neurodegeneration, synapse function and metabolism. Comprehensive analysis of genes with both altered m⁶A methylation and transcription identified several key genes highly associated with inflammatory processes. Validation experiments confirmed significant dysregulation of mRNA levels for several target genes (Ahsg, Alb, Apoa1, Serpina1b, Serpinf2, and Serpina3K) and m⁶A regulators (Mettl3, Mettl16, Igfbp1 and Igfbp3) in the cortex of septic mice.

Conclusion

This study provided a transcriptome-wide analysis of m⁶A methylation in the cortical tissues of septic mice, suggesting that m⁶A-mediated regulation of inflammatory gene expression is associated with the pathogenesis of SAE. The findings presented valuable insights into the mechanisms of sepsis-associated brain injury and potential therapeutic targets.

Introduction

Sepsis poses a significant threat to hospitalized patients worldwide, and it is the leading cause of death in the intensive care units.¹ Sepsis originates from a systemic immune response, leading to multiple organ damage, especially to the brain. About 70% of patients with sepsis experience brain damage, ranging from mild cognitive impairment to coma.² This brain damage significantly increases patient mortality and, even after treatment, can result in long-term cognitive decline and an increased risk of neurodegenerative diseases.³ Despite its prevalence, the pathophysiology underlying sepsis-associated brain injury and resulting cognitive dysfunction remains elusive. It is broadly accepted that the occurrence of sepsis-associated encephalopathy (SAE) arises from the excessive activation of the inflammatory response. While inflammation is essential for pathogen elimination, overactivation of the immune response in sepsis causes a remarkable rise in serum inflammatory mediators and complement system components. This leads to blood-brain barrier disruption and excessive neuroinflammation. Microglia, the main immune cells of the central nervous system (CNS), play a pivotal role in mediating neuroinflammation during SAE. Overactivated microglia in SAE can release large amounts of inflammatory factors, leading to neuronal damage and cognitive impairment.^4,5

N6-Methyladenosine (m⁶A) is one of the most abundant RNA modifications, regulating post-transcriptional gene expression through modulating RNA splicing, transport, degradation, and translation. Emerging evidence confirmed the essential roles of m⁶A in the neural development and function, particularly in neuronal plasticity and cognition, by regulating genes associated with neuronal plasticity.⁶ Clinical evidence demonstrated a correlation between m⁶A and sepsis. For instance, numerous m⁶A-eQTLs linked to sepsis have been identified, as well as three m⁶A subtypes associated with patient heterogeneity.⁷ Animal studies suggest that m⁶A modifications regulate immune responses and immune-mediated organ damage in sepsis, including cardiovascular dysfunction, acute kidney injury, and acute lung injury.^8–10 A limited number of studies have explored the role of m⁶A methylation in SAE pathophysiology. Elevated Mettl3 and decreased FTO levels have been reported in the serum of SAE patients.¹¹ The upregulated m⁶A level in NLRP3, driven by elevated expression of Mettl3, has exhibited to inhibit the inflammation and pyroptosis in sepsis-induced lung and brain injury models.¹² In addition, knockdown of the m⁶A RNA methylation reader Ythdf1 in macrophages reduces m⁶A methylation of JAK2/STAT3 and alleviates the brain injury during sepsis.¹³ Moreover, Mettl3, FTO, and Ythdf1 play essential roles in the CNS and regulate cognitive function.^14–16 Mettl3 depletion promotes migration of monocyte-derived macrophages and Aβ clearance by reducing the m⁶A modification of Dnmt3a mRNAs and attenuating Ythdf1-mediated translation of DNMT3A, ameliorating the cognition in a mouse model of Alzheimer’s disease.¹⁴ Ythdf1-mediated m⁶A facilitates learning and memory by promoting protein synthesis of target transcripts in response to neuronal stimuli in the mouse hippocampus.¹⁵ The overexpression of FTO in the hippocampus reverses the synaptic and cognitive impairment in hypoxic-ischemic neonatal rats by promoting the degradation of PTEN mRNA.¹⁶ However, the role of m⁶A in SAE-related cognitive dysfunction has not yet been reported.

The cerebral cortex plays a pivotal role in the regulation of cognition, and is particularly critical for memory consolidation.^17,18 Although most studies on SAE have focused on hippocampal dysfunction, emerging evidence demonstrates that the cortical tissue also plays a critical role during sepsis pathogenesis. Clinical studies have revealed elevated peptidoglycan signaling and increased microinfarcts in the cortical tissue of sepsis patients.^19,20 Furthermore, animal model studies have demonstrated that the cortex is mechanistically involved in the pathogenesis of neuroinflammation, neurological dysfunction, cognitive impairment, anxiety, and depression-like behaviors during SAE.^21–24 In this study, m⁶A epitranscriptomic and transcriptomic profiles were analyzed in the cortical tissues of mice with LPS-induced endotoxemia, a widely used experimental model of SAE. Numerous genes exhibited differential expression and m⁶A methylation. Comprehensive analysis of genes with altered m⁶A methylation and differential transcription revealed several key genes closely associated with inflammatory processes. These findings provide novel insights into the molecular mechanisms of SAE-related brain dysfunction, advancing the understanding of SAE pathogenesis and identification of potential therapeutic targets.

Materials and Methods

Animals and SAE Models

This study utilized 8–12-week-old male C57BL/6 mice (Jicui Pharmachem Biotechnology Co., Ltd., China), with body weights ranging from 18 to 22 grams. Mice were housed under a standard 12:12 h light-dark cycle. SAE was induced according to previously published protocols.²⁵ Lipopolysaccharide (LPS, Escherichia coli 0111:B4; Sigma-Aldrich, St. Louis, MO, USA) was intraperitoneally injected at a dose of 10 mg/kg to establish th e SAE model, while control mice received an equivalent volume of PBS solution. Cortical tissues were collected 24 h post-injection for further analysis. For euthanasia, mice were anesthetized with 1.25% tribromoethanol (Beijing Lab Animal Technology Develop Co., LTD., China) and then subjected to cervical dislocation.

Novel Object Recognition (NOR)

The NOR test was conducted using an open field box measuring 50 × 50×50 cm³ (n=12 mice per group). Mice were first acclimated to the testing environment by exploring the empty box for 5 min. After acclimation, the mouse was placed in the center of the box and allowed to explore two identical objects during a 5-min familiarization session, followed by return to its cage. For short-term memory testing, a 30-min interval was set between the familiarization and test sessions. During the test session, mice were presented with one familiar object and one novel object, and exploration behavior was recorded for 5 min. Exploration time for each object was analyzed using ViewPoint Video Tracking software (ViewPoint Behavior Technology, Lyon, France). Recognition memory was quantified as the exploration rate, which was calculated as the percentage of time spent exploring the novel object relative to the total exploration time.

Forced Swimming Test (FST)

Mice (n=12 per group) underwent the FST 24 h after LPS treatment. Mice were placed individually in a transparent cylinder (35 cm high, 20 cm diameter) filled with 25 cm of water maintained at 25 ± 1 °C. The test consisted of a 15-min training session, followed by a 5-min test session conducted one h later. During the 5-min test, an observer, who was blinded to the treatment groups, recorded the immobility time, defined as minimal movements required to keep the head above water. Increased immobility duration was used as an indicator of depression-like behavior.

MeRIP-Seq Analysis and RNA-Seq Analysis

The sequencing of m⁶A-methylated mRNAs was performed by Novogene Co., Ltd. (Beijing, China). For MeRIP-seq, 30 µg of RNA extracted from brain cortical tissues (three mice per group) was used. RNA integrity was assessed using an Agilent 2100 Bioanalyzer (Agilent Technologies Inc., Santa Clara, CA, USA). Fragmented mRNA (~100 nt) was incubated at 4 °C for 2 h with an anti-m⁶A polyclonal antibody (Synaptic Systems, Berlin, Germany). Both input RNA and immunoprecipitated RNA were utilized for RNA-seq library preparation using the NEBNext^® Ultra™ II RNA Library Prep kit for Illumina^® (New England Biolabs, Ipswich, MA, USA). Sequencing was conducted on the Illumina NovaSeq 6000 platform, generating paired-end reads of 150 bp.

For RNA-seq, 1 µg of total RNA per sample was sequenced on the Illumina platform following standard protocols (three mice per group). Reference genome and gene model annotation files were downloaded from genome repositories, and Hisat2 (v2.0.5) was utilized to construct the reference genome index. Clean paired-end reads were aligned to the reference genome using Hisat2, constructing a splice junction database to enhance mapping accuracy. m⁶A peaks were identified using the exomePeak R package (ver. 2.16.0), with input samples serving as controls. Peak enrichment was determined using a q-value threshold of 0.05. Besides, m⁶A-enriched motifs were identified using HOMER (ver. 4.9.1).

Bioinformatics Analysis

RNA integrity (RIN ≥ 8.0, 28S/18S ≥ 1.5), mapping rate (> 90%), and IP reproducibility (Spearman r ≥ 0.90) met standard requirements for MeRIP-seq. For RNA-seq, the same RIN and mapping rate criteria were applied. Differential expression analysis was conducted using the DESeq2 R package (ver. 1.20.0). The Benjamini-Hochberg method was used to adjust P-values for false discovery rate control. Genes with an absolute log2 fold change (|log2FC|) ≥ 1 and an adjusted P-value < 0.05 were classified as differentially expressed. The Gene Ontology (GO) enrichment analysis of differentially expressed genes (DEGs) was performed using the clusterProfiler R package, with corrections for gene length bias. GO terms with a corrected P-value < 0.05 were considered as significantly enriched. The Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis was undertaken using the clusterProfiler R package to interpret molecular-level data in the context of high-level biological functions. The STRING database was utilized to analyze protein-protein interaction (PPI) networks, which were visualized using Cytoscape software.

Quantitative Reverse Transcription Polymerase Chain Reaction (RT-qPCR) and MeRIP-qPCR

Total RNA was extracted from the tissues using the TRIzol reagent (Thermo Fisher Scientific, Waltham, MA, USA). RNA was reversely transcribed to complementary DNA using the HiScript IV 1st Strand cDNA Synthesis kit (Vanyme, Beijing, China) according to the manufacturer’s protocol. RT-qPCR was conducted using the 2X SYBR Green Pro Taq HS Premix qPCR Kit (Accurate Biology, Hunan, China) on a Bio-Rad Real-Time PCR System. To validate the MeRIP-seq findings, RNA was extracted from the cortical tissues of three control mice and three septic mice. The RNA samples were subjected to immunoprecipitation with an m⁶A-specific antibody (Sigma, St. Louis, MO, USA). The immunoprecipitated RNA was then reverse-transcribed into cDNA using the Thermo Scientific RevertAid First Strand cDNA Synthesis Kit (Thermo Scientific, Waltham, MA, USA). The primers used for RT-qPCR and MeRIP-qPCR are listed in Supplementary Table S1 and S2.

Enzyme-Linked Immunosorbent Assay (ELISA)

Protein concentrations were measured using pre-coated ELISA plates (R&D Systems, Minneapolis, MN, USA). Standards or samples were incubated on the plates for 2 h, washed, and incubated with an enzyme-conjugated secondary antibody for 1 h. After additional washing, a substrate solution was added to develop color. The reaction was stopped using a stop solution, and absorbance was measured at 450 nm. Protein concentrations were determined using a standard curve.

Statistical Analysis

Data were derived from at least three independent experiments and were presented as mean ± standard error. Statistical analysis was conducted using GraphPad Prism 8.0 software (GraphPad Software Inc., San Diego, CA, USA). Comparisons between two groups were performed using unpaired Student’s t-test. P-value < 0.05 was considered statistically significant. Differential peaks and transcripts between control and LPS-treated groups were identified based on an absolute log2 fold change (|log2FC|) ≥ 1 and an adjusted P-value < 0.05.

Results

Neuroinflammation and Behavioral Impairment Induced by LPS

Sepsis was induced in mice via intraperitoneal injection of LPS. As sepsis is known to cause significant brain damage and cognitive impairment alongside elevated neuroinflammation, inflammatory markers and cognitive functions were assessed in septic mice. Cortical tissues were collected 24 h post-injection to measure levels of inflammatory cytokines. The mRNA levels of key pro-inflammatory cytokines, including interleukin-6 (IL-6) and tumor necrosis factor-α (TNF-α), were significantly elevated in the cortical tissues of septic mice, while IL-1β level remained unchanged. The anti-inflammatory cytokine IL-10 also exhibited an increased expression level (Figure 1A). ELISA confirmed significantly elevated TNF-α protein level in the cortical tissues (Figure 1B). Behavioral assessments were conducted to evaluate cognitive function. The NOR test revealed impaired memory in SAE mice (Figure 1C). Furthermore, depressive-like behavior was observed in SAE mice during the FST, as evidenced by the prolonged immobility time (Figure 1D). The locomotor activity assessed by open-field test showed no significant difference between the LPS and control groups (Figure 1E). These results highlight the influences of sepsis on brain injury.

Figure 1.

Neuroinflammation and behaviors of LPS-induced SAE mice and control mice. (A) mRNA expression levels of IL-1β, IL-10, IL-6, and TNF-α in the cortical tissues in the LPS group and control group (n=4). (B) Protein levels of IL-1β, IL-10, IL-6, and TNF-α detected by ELISA in the LPS group and control group (n=4). (C) Cognition function evaluated by new object recognition analysis in the LPS group and control group (n=12). (D) Depression-like behaviors measured by the forced swimming test in the LPS group and control group (n=12). (E) Locomotor activity measured by the open-field test in the LPS group and control group (n=12). * P < 0.05 compared with the control; ** P < 0.01 compared with the control. Comparisons between two groups were performed using unpaired Student’s t-test.

Functional Enrichment Analysis of DEGs in the Cortex of Mice

RNA-seq identified 3128 DEGs in the cortex of septic mice compared with control mice, including 1,451 upregulated and 1,677 downregulated genes (Figure 2A). Table 1 presents the top 10 upregulated and downregulated genes. GO analysis showed that the upregulated genes were most enriched in biological processes including defense response to other organism, positive regulation of immune response and response to bacterium (Figure 2B). The KEGG pathway analysis revealed that upregulated genes were enriched in pathways including cytokine−cytokine receptor interaction, Epstein-Barr virus infection and herpes simplex infection (Figure 2C). In contrast, GO analysis revealed that downregulated genes were enriched in biological processes including smoothened signaling pathway, regulation of smoothened signaling pathway and G-protein coupled receptor signaling pathway (Figure 2D). KEGG pathway analysis of these genes revealed significant association with neuroactive ligand-receptor interaction, PI3K−Akt signaling pathway and calcium signaling pathway (Figure 2E). The leading-edge genes for significantly enriched KEGG pathways identified by GSEA were listed in Supplementary Table S3 and S4. PPI network analysis further identified Tnf, Il6 and Cxcl10 as central upregulated genes which are associated with inflammatory responses (Figure 3A), while Col1a1, Fos, and H3c8 were central downregulated genes (Figure 3B).

Figure 2.

Transcriptome profiling and functional enrichment analysis of differentially expressed genes in the cortex in the SAE and control groups. (A) Volcano map showing DEGs in SAE mice (|log2FC|>1, adjusted P-value<0.05, n=3 in each group). (B) Major gene ontology terms of upregulated genes. (C) Top 20 significantly enriched pathways of upregulated genes. (D) Major gene ontology terms of downregulated genes. (E) Top 20 significantly enriched pathways of downregulated genes.

Table 1.

Top 10 Upregulated and Downregulated Genes in the Cortex of Septic Mice

Chr	Peak_Length	Transcript ID	Strand	Adjusted P-value	Log2FC	Gene Name
9	1801	ENSMUSG00000043613	+	0.00	11.30	Mmp3
5	2252	ENSMUSG00000034855	-	0.00	10.73	Cxcl10
7	2606	ENSMUSG00000040026	-	0.00	10.15	Saa3
5	3016	ENSMUSG00000029417	-	0.00	9.77	Cxcl9
11	1413	ENSMUSG00000038067	+	0.00	9.60	Csf3
18	2838	ENSMUSG00000068606	-	0.00	9.48	Gm4841
1	1925	ENSMUSG00000026390	-	0.00	9.33	Marco
9	7237	ENSMUSG00000049723	+	0.00	9.15	Mmp12
12	1719	ENSMUSG00000079012	+	0.00	8.54	Serpina3m
18	1216	ENSMUSG00000085977	+	0.00	8.16	Gm5970
8	1901	ENSMUSG00000036151	+	0.00	−8.13	Tm6sf2
9	1480	ENSMUSG00000111844	-	0.00	−7.33	Gm47097
1	1843	ENSMUSG00000103685	+	0.00	−6.91	Gm37074
7	2303	ENSMUSG00000108832	-	0.00	−6.61	Gm44832
12	540	ENSMUSG00000085539	-	0.00	−6.53	Gm15561
X	2090	ENSMUSG00000040229	+	0.00	−6.45	Gpr34
1	386	ENSMUSG00000101840	+	0.00	−6.25	Gm28294
X	2241	ENSMUSG00000031170	+	0.00	−6.12	Slc38a5
12	120	ENSMUSG00000080406	+	0.00	−6.09	Mir1188
6	844	ENSMUSG00000106965	-	0.00	−5.99	Gm18290

Figure 3.

PPI network analysis of differentially expressed genes in the cortex in the SAE and control groups. (A) PPI network of upregulated genes. (B) PPI network of downregulated genes.

Differential m⁶A Methylation Profiles Between Septic and Control Mice

To investigate the role of m⁶A modifications in SAE, cortical m⁶A modifications in septic and control mice were analyzed using MeRIP-seq. A total of 5,930 common m⁶A peaks corresponding to 5,543 gene transcripts were identified. The control group exhibited 1,286 unique m⁶A peaks associated with 1,096 genes, while the sepsis group displayed 1,021 unique peaks corresponding to 820 genes (Figure 4A). Analysis of m⁶A distribution along mRNAs revealed that m⁶A modifications predominantly occurred in the coding sequence (CDS) and the 3′ untranslated region (3′ UTR) (Figure 4B and C). Slightly higher methylation levels were found in the CDS of septic mice (46.94% vs 44.99%), with a corresponding decrease in 3′ UTR methylation (36.49% vs 37.68%) (Figure 4B). The GGACU motif, the binding site motif of m⁶A writer Mettl3, was identified as the core m⁶A consensus motif in both groups (Figure 4D). Differential methylation analysis revealed 924 differentially methylated peaks, involving 683 upregulated and 241 downregulated peaks, corresponding to 571 hyper-methylated and 209 hypo-methylated genes in septic mice (Figure 4E). Table 2 presents the top 10 upregulated and downregulated m⁶A peaks.

Figure 4.

m⁶A distribution profiles in the cortical tissues of SAE and control mice. (A) Overlap of m⁶A peaks in the two groups. Venn diagrams presenting number of unique m⁶A peaks in control and LPS groups, as well as shared peaks, including genes containing m⁶A modifications. (B) Distribution of m⁶A peaks across five transcriptional regions: CDS, 5′ UTR, 3′ UTR, TSS, and stop codon. (C) Distribution of m⁶A peaks in mRNA regions: CDS, 5′ UTR, and 3′ UTR. (D) Top five enriched m⁶A motifs identified in SAE and control groups. (E) Volcano plot of differentially modified m⁶A peaks, exhibiting 683 hyper-m⁶A and 241 hypo-m⁶A peaks in the LPS group compared with the control group (|log2FC| > 1, adjusted P-value < 0.05, n = 3 in each group).

Table 2.

Top 10 Upregulated and Downregulated m⁶A Peaks in the Cortex of Septic Mice

Chr	Peak_Length	Transcript ID	Strand	Adjusted P-value	DiffMod Log2FC	Gene Name
11	570	ENSMUSG00000020411	-	0.00	4.18	Nipal4
8	271	ENSMUSG00000079038	+	0.00	3.91	D130040H23Rik
9	300	ENSMUSG00000047766	-	0.00	2.91	Lrrc49
2	748	ENSMUSG00000026778	+	0.00	2.79	Prkcq
10	270	ENSMUSG00000020198	-	0.00	2.58	Ap3d1
7	478	ENSMUSG00000035004	-	0.00	2.52	Igsf6
6	22276	ENSMUSG00000051331	-	0.00	2.25	Cacna1c
16	209	ENSMUSG00000038168	-	0.00	2.24	P3h2
7	211	ENSMUSG00000049728	-	0.00	2.21	Zfp668
6	361	ENSMUSG00000062519	+	0.00	2.12	Zfp398
8	7632	ENSMUSG00000055730	+	0.00	−18.3	Ces2a
3	2476	ENSMUSG00000004038	-	0.00	−17.4	Gstm3
5	19474	ENSMUSG00000056035	-	0.00	−14.4	Cyp3a11
5	1707	ENSMUSG00000029260	-	0.00	−14.1	Ugt2b34
9	6678	ENSMUSG00000032310	-	0.00	−13.9	Cyp1a2
19	21773	ENSMUSG00000054827	+	0.00	−13.6	Cyp2c50
10	4670	ENSMUSG00000040134	-	0.00	−12.5	Rdh7
19	42245	ENSMUSG00000003053	+	0.00	−12.2	Cyp2c29
7	7918	ENSMUSG00000005547	+	0.00	−11.8	Cyp2a5
5	9863	ENSMUSG00000035836	-	0.00	−11.5	Ugt2b1

Functional Enrichment Analysis of Differentially Modified Genes (DMGs)

Functional enrichment analysis of DMGs was carried out to elucidate the potential roles of m⁶A methylation in SAE pathology. GO analysis revealed that hypermethylated genes were significantly enriched in biological processes related to regulation of transcription by RNA polymerase II, positive regulation of transcription by RNA polymerase II and chromatin remodeling (Figure 5A). KEGG pathway analysis indicated that these genes were involved in pathways, such as pathways of neurodegeneration − multiple diseases, Alzheimer disease and spinocerebellar ataxia (Figure 5B). In contrast, hypomethylated genes were enriched in processes related to cell differentiation, protein ubiquitination and response to lipopolysaccharide (Figure 5C). The enriched KEGG pathways included metabolic pathways, PI3K−Akt signaling pathway and retinol metabolism (Figure 5D). The leading-edge genes for significantly enriched KEGG pathways identified by GSEA were listed in Supplementary Table S5 and S6. PPI network analysis highlighted key genes with altered m⁶A methylation. Prkacb, Grin2b and Htt were identified as central genes with upregulated m⁶A (Figure 5E), while Alb, Ugt2b1 and Cyp3a11 were central among downregulated genes (Figure 5F).

Figure 5.

Functional analysis of genes associated with differentially modified m⁶A peaks. (A) Major gene ontology terms of genes with upregulated m⁶A modification. (B) Top 20 significantly enriched pathways of genes with upregulated m⁶A modification. (C) Major gene ontology terms of genes with downregulated m⁶A modification. (D) Top 20 significantly enriched pathways of genes with downregulated m⁶A modification. (E) PPI network of genes with upregulated m⁶A modification. (F) PPI network of genes with downregulated m⁶A modification. * P < 0.05.

Combined Analysis of DMGs and DEGs in the Cortex of Mice

A comprehensive analysis of DEGs and DMGs identified 63 genes that exhibited concurrent changes in both mRNA expression levels and m⁶A methylation (Figure 6A and B). To clarify the relationship between m⁶A methylation and mRNA expression levels, Pearson’s correlation analysis revealed a significant negative correlation between m⁶A modification level and mRNA expression level in the septic mice (Figure 6C), suggesting that m⁶A modification may play a repressive role in gene expression in the pathogenesis of SAE. The KEGG enrichment analysis of these genes revealed distinct pathways associated with mRNAs, exhibiting altered methylation and transcriptional upregulation and those with altered methylation and transcriptional downregulation (Table 3). The PPI network analysis revealed a single interaction network comprising seven key genes: Ahsg, Alb, Apoa1, Apoa2, Serpina1b, Serpinaf2, and Serpina3K (Figure 6D). Notably, these genes were linked to inflammatory processes. Analysis of m⁶A methylation levels and modification sites using the Integrative Genomics Viewer (IGV) demonstrated decreased methylation associated with elevated transcription levels in septic mice (Figure 6E and Figure S1).

Figure 6.

Combined analysis of genes with altered methylation and transcription levels. (A) Overlap of DMGs and DEGs. (B) Venn diagrams illustrating overlap between hyper-/hypo-methylated genes and upregulated/downregulated DEGs. (C) Pearson’s correlation analysis of the levels of DMGs and DEGs. (D) PPI network of overlapped DMGs and DEGs. (E) m⁶A methylation levels and modification sites on Ahsg mRNA.

Table 3.

Enriched KEGG Pathways of Genes with Altered Methlylation and mRNAs

	KEGG_Pathway	Official_Gene_Symbol	Gene Name
Altered DMGs/upregulated DEGs	PPAR signaling pathway	Apoa1	Apolipoprotein A-I (Apoa1)
		Apoa2	Apolipoprotein A-II (Apoa2)
		Plin4	Perilipin 4(Plin4)
		Pck1	Phosphoenolpyruvate carboxykinase 1, cytosolic (Pck1)
	Cholesterol metabolism	Apoa1	Apolipoprotein A-I (Apoa1)
		Apoa2	Apolipoprotein A-II (Apoa2)
Altered DMGs/downregulated DEGs	Cysteine and methionine metabolism	Amd1	S-adenosylmethionine decarboxylase 1(Amd1)
		Tst	Thiosulfate sulfurtransferase, mitochondrial(Tst)
	Morphine addiction	Pde10a	Phosphodiesterase 10A (Pde10a)
		Kcnj9	Potassium inwardly-rectifying channel, subfamily J, member 9(Kcnj9)

Altered Expression Levels of the Key Genes and m⁶A Regulators in the Cortex of Septic Mice

To validate the m⁶A methylation levels of the key genes identified in the PPI network, MeRIP-qPCR was performed. Decreased methylation levels of Ahsg, Alb, Apoa1, Serpina1b and Serpina3k were observed in the cortical tissue of septic mice (Figure 7A). To validate the expression of the key genes identified in the PPI network, RT-qPCR was conducted. The results confirmed significantly elevated mRNA levels of Ahsg, Alb, Apoa1, Serpina1b, Serpinf2, and Serpina3K in the cortical tissues of septic mice (Figure 7B).

Figure 7.

m⁶A methylationand expression levels of PPI network core genes. (A) MeRIP-qPCR analysis of PPI network core genes m⁶A methylation levels (n = 3). (B) mRNA expression levels of core genes identified by the PPI network (n = 4). *P < 0.05, ***P < 0.001 compared with control group. Comparisons between two groups were performed using unpaired Student’s t-test.

Given the critical role of m⁶A modifications in RNA regulation, the expression levels of m⁶A-related “writer,” “eraser,” and “reader” genes were analyzed in the transcriptome data. The analysis revealed significant downregulation of writer genes Mettl3 and Mettl16, along with the reader gene Igf2bp3. Conversely, reader genes Igfbp1 and Igfbp3 were upregulated (Figure 8A). RT-qPCR was performed to examine the expression levels of m⁶A related genes. Mettl3 and Mettl16 were downregulated whereas Igfbp1 and Igfbp3 were upregulated (Figure 8B).

Figure 8.

Expression levels m⁶A regulators. (A) Clustered heatmap of m⁶A-related genes from the cortical tissues in the LPS group and control group. (B) qPCR analysis of m⁶A-related genes from the cortical tissues in the LPS group and control group (n = 3). *P < 0.05, **P < 0.01, ***P < 0.001 compared with control group. Comparisons between two groups were performed using unpaired Student’s t-test.

Discussion

Sepsis is a severe and potentially fatal condition that frequently involves multiple organ systems, including the inflammatory processes in CNS, resulting in neurological injury and cognitive dysfunction. This inflammatory environment ultimately results in neuronal damage and cognitive impairments in affected cases.²⁶ Notably, cognitive deficits in septic patients can persist long after resolution of the acute phase of sepsis, and sepsis significantly increases the risk of developing neurodegenerative disorders.²⁷ Therefore, elucidating the mechanisms underlying sepsis-associated neurological damage is crucial to identify novel therapeutic targets and improve patient outcomes. Notably, m⁶A is a crucial post-transcriptional modification of RNA, regulating RNA stability and quality control.²⁸ Recent studies have highlighted the involvement of m⁶A modifications in sepsis-induced multi-organ dysfunction,²⁹ including cardiovascular impairment, acute lung injury (ALI), and acute kidney injury (AKI).^30,31 Zhang et al identified distinct sepsis subtypes through m⁶A methylation profiling in septic patients, revealing unique RNA epigenetic landscapes, immune profiles, biological processes, and clinical outcomes.³² Additionally, analysis of peripheral blood mononuclear cells (PBMCs) from SAE patients indicated an upregulation of the m⁶A methyltransferase Melttl3 and a downregulation of the m⁶A demethylase FTO.¹¹ However, no research has concentrated on m⁶A modifications in brain tissues affected by sepsis.

We performed RNA-seq and MeRIP-seq on the cerebral cortices of septic mice and control mice. Our RNA-seq analysis revealed that the upregulated mRNAs in the cortex of SAE mice were mainly enriched in pathways related to immune and inflammatory responses, including cytokine-cytokine receptor interaction, Epstein-Barr virus infection and herpes simplex infection. Pro-inflammatory cytokines exacerbate the systemic inflammatory response in sepsis, and induce neuroinflammation and blood-brain barrier disruption, eventually contributing to sepsis-associated brain injury.^2,33 Besides the most common bacterial infections, viral infections are also significant triggers of sepsis. Patients in the chronic phase of sepsis show increased reactivation of latent viruses, including Epstein–Barr virus and herpes simplex virus (HSV).³⁴ Epstein-Barr virus seropositivity has also been demonstrated to be highly associated with the immune dysregulation and increased mortality in pediatric sepsis.^35,36 In contrast, downregulated mRNAs were enriched in pathways including neuroactive ligand-receptor interaction, PI3K−Akt signaling pathway and calcium signaling pathway, all of which play important roles in the regulation of neural function. A recent study identified key pathways involved in the progression in septic patients, revealing significant enrichment of differentially expressed genes in neuroactive ligand-receptor interaction and the PI3K-Akt signaling pathway.³⁷ The neuroactive ligand-receptor interaction pathway is crucial for regulating neuronal activity, synaptic function and cognition.^38,39 Circulating inflammatory mediators in sepsis can lead to neuroendocrine, behavioral and cognitive dysfunction through a range of neuroactive substances such as acetylcholine and γ-aminobutyric acid.⁴⁰ Downregulated PI3K−Akt signaling pathway is closely related to the pathogenesis of sepsis, and activation of this pathway can ameliorate sepsis-induced organ injury in the brain, skeletal muscle, liver, lung and other organs.^41–44 Moreover, the PI3K−Akt signaling pathway plays a significant role in regulating synaptic plasticity and cognitive function, and activation of PI3K−Akt signaling can protect the blood-brain barrier and cognitive function in septic mice.^41,45 The calcium signaling pathway also plays a key role in synaptic plasticity and cognitive regulation, and calcium dynamics are involved in the pathogenesis of sepsis-associated delirium.^46–48

MeRIP-seq analysis examined the m⁶A methylation profiles in the cortices of septic mice. The functional enrichment analysis indicated that genes with upregulated m⁶A modification were mainly enriched in pathways including neurodegeneration−multiple diseases, Alzheimer’s disease, spinocerebellar ataxia, as well as several synapse-related pathways. Studies have shown that chronic neuroinflammation associated with sepsis can lead to neurodegeneration and cognitive dysfunction.^49,50 In particular, sepsis survivors have an increased risk of dementia, and sepsis-related systemic infection exacerbates the pathological progression of Alzheimer’s disease (AD), accelerating the cerebrovascular dysfunction and cognitive decline.^51–53 m⁶A methylation has also been shown to be involved in synaptic protein expression and neural function regulation, and it plays an important role in cognition and the pathogenesis of various neurodegenerative diseases, such as AD, amyotrophic lateral sclerosis (ALS), and Parkinson’s disease (PD).^54–56 Genes with downregulated m⁶A modification were mainly enriched in metabolic pathways, PI3K−Akt signaling pathway, and retinol metabolism. Dysregulation of metabolic processes plays a significant role in the pathogenesis of sepsis. Metabolic reprogramming induced by sepsis occurs in almost all types of cells.⁵⁷ The vigorous immune response initiated by the body to eliminate the pathogens leads to disruption of metabolic homeostasis and insufficient energy, which may ultimately result in multi-organ dysfunction (including brain injury), a core pathological feature of sepsis.^58,59 Taken together, our results indicate a potential role of m⁶A modification in crucial pathways involved in SAE, including pathways related to neurodegeneration, metabolic pathways and PI3K−Akt signaling pathway.

As m⁶A methylation participates in the RNA metabolism including transcription as well as post-transcriptional regulation of gene expression, a combined analysis of transcriptome and m⁶A epitranscriptome data, alongside PPI network analysis, was conducted to identify potential core genes. There were 63 genes that exhibited concurrent changes in both mRNA expression levels and m⁶A methylation. Specifically, there were 18 genes with increased transcription and methylation, 13 genes with increased transcription and decreased methylation, 19 genes with decreased transcription and increased methylation, and 13 genes with decreased transcription and methylation. m⁶A methylation, as the most prevalent epitranscriptomic modification, exerts multilayered control over protein synthesis through comprehensive regulation of RNA metabolism, including transcription, splicing, transport, degradation, and translation. m⁶A modification exhibits cell type-specificity,⁶⁰ with its regulatory complexity arising not only from dynamic modulation across all mRNA metabolic stages but also from the functional diversity of m⁶A writers, erasers and readers. Furthermore, m⁶A-associated regulatory proteins undergo various post-translational modifications (eg, phosphorylation, ubiquitination, SUMOylation), which alter their stability, activity and functionality.⁶¹ Multiple m⁶A reader proteins are important determinants mediating the functional effects of m⁶A modifications which are recognized by specific readers differently under different cellular contexts. The activities and functions of m⁶A readers are modulated by the expression levels, post-translational modifications as well as the cellular locations of reader proteins.⁶² And m⁶A readers recognize m⁶A sites selectively depending on the density and sequence contexts of m⁶A sites and function differently according to their binding sites.⁶³ For example, YTHDF2 and IGF2BP1/2/3 bind to different sites on transcripts with m⁶A modifications, and either increase or decrease mRNA stability in an m⁶A-dependent manner.⁶⁴ In addition to the m⁶A mediators, the regulatory effects of m⁶A methylation can be modulated by specific cellular and molecular contexts, including cell types, external stimuli and cellular signaling, and positions of m⁶A modification sites along the mRNA. m⁶A modifications have been shown to by dynamically modulated in response to cell differentiation and external signals from stimuli or stress.⁶⁵ Cell type is also an important determinant, as m⁶A effectors may target different groups of transcripts in different cells and tissues.⁶² Furthermore, m⁶A-mediated regulation also depends on the exact locations of m⁶A modification sites. m⁶A modification within the 5′ UTR can promote cap-independent translation by directly binding eIF3 and 5′ UTR methylation can also regulate alternative translation by controlling ribosomal scanning and start codon selection.^66,67 m⁶A in the coding sequence has been shown to promote translation via the m⁶A reader YTHDC2, and the position and context of m⁶A within codons affect translational dynamics via tRNA accommodation and translation elongation.^68,69 In addition, m⁶A in the 3′ UTR can promote translation initiation by binding reader proteins YTHDF1 and YTHDF3.^70,71 Such multifaceted regulatory mechanisms may ultimately lead to discordant effects on mRNA abundance and protein expression levels in m⁶A-modified targets.

PPI network analysis of these genes revealed a unique interaction network of seven genes associated with inflammatory process. The expression and methylation levels of these genes were detected, and mRNA levels of five hypomethylated genes were verified to be upregulated in the cortical tissues of septic mice, including Ahsg, Alb, Apoa1, Serpina1b and Serpina3K. These molecular changes may reflect the dynamic inflammatory responses in SAE-affected brains. The anti-inflammatory roles of Fetuin-A protein (encoded by Ahsg), Albumin (encoded by Alb), Apoa1, Serpina1b and Serpina3K have been proved, and Albumin, Serpina1b and Serpina3K also exhibit neuroprotective functions. Albumin possesses neuroprotective effects, protecting neurons against oxidative stress, reducing neuroinflammation, and improving cognitive behaviors in AD.^72–74 Serpina1b and Serpina3K, members of serine protease inhibitors, may have anti-inflammatory and neuroprotective roles in sepsis. Increased α-1 antitrypsin protein level (AAT, encoded by six Serpina1 genes in mice) can inhibit inflammation and improve the survival of septic mice,^75,76 and AAT has a protective effect on the CNS, inhibiting neuroinflammation associated with AD, as well as reducing neuronal cell death and alleviating brain damage following cerebral ischemia.^77,78 Serpina3K plays an anti-inflammatory and anti-angiogenic role by blocking Wnt signaling pathway.⁷⁹ And Serpina3K can inhibit oxidative stress and neuronal apoptosis in the injured areas, thereby alleviating the neural dysfunction of mice with traumatic brain injury.⁸⁰ In summary, our findings suggest that m⁶A methylation may modulate SAE brain injury by regulating the expression of these anti-inflammatory genes.

Limitation

There are still several limitations in our study. First, LPS and CLP-induced septic models are the most well-recognized animal models of sepsis, and it has been shown that the neuroinflammatory response to LPS exposure is affected by the animal’s sex. Given that all studies were conducted in LPS-induced septic male mice, it’s necessary to validate our findings in CLP-induced SAE mice and to take sex-mediated differences into consideration to increase the reliability and generalizability of conclusions in the upcoming studies. Second, all studies were performed in septic mice 24 hours post LPS injection. In the future, we will continue to analyze the m⁶A methylation and the long-term neurocognitive decline as well as other neural dysfunctions in the late phase of SAE. Lastly, further research is needed to determine the specific m⁶A regulators and target genes involved in the pathogenesis of SAE and explore the potential regulatory mechanism by which m⁶A regulators mediate the expression of targeted genes.

Conclusion

In conclusion, the present study revealed differential m⁶A methylation patterns and altered transcriptomic profiles in the cortices of LPS-induced SAE mice compared with control mice. Genes with concurrent alterations in m⁶A methylation and transcriptional levels were involved in inflammatory processes. The findings suggest that m⁶A-mediated potential link of inflammatory gene expression may play a role in the pathogenesis of SAE, providing new insights into the pathogenic mechanisms of sepsis-associated brain injury.

Funding Statement

This work was supported by the Guangdong Basic and Applied Basic Research Foundation (2022A1515111190) and Shenzhen Key Laboratory of Maternal and Child Health and Diseases (ZDSYS20230626091559006).

Data Sharing Statement

The datasets generated during the current study are available in the Sequence Read Archive database (BioProject ID PRJNA1163743).

Ethics Statement

The animal study was reviewed and approved by the Ethics Committee of Shenzhen Maternity and Child Healthcare Hospital (Shenzhen, China, approval number SZMCHH-21-2405-10) and adhered to the ARRIVE guidelines, and all procedures complied with the national guidelines for laboratory animal welfare of the People’s Republic of China (GB/T 35892–2018).

Author Contributions

Juan Zhao: Conceptualization, Data curation, Formal analysis, Investigation, Project administration, Visualization, Writing-original draft, Writing-review and editing.

Tengfei Zhu: Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Project administration, Visualization, Writing-original draft, Writing-review and editing.

Chao Zhao: Data curation, Formal analysis, Investigation, Visualization, Writing-review and editing.

Xu Huang: Data curation, Formal analysis, Investigation, Visualization, Writing-review and editing.

Fang Yao: Data curation, Formal analysis, Investigation, Visualization, Writing-review and editing. Yueyuan Qin: Formal analysis, Writing-review and editing.

Baojiang Wang: Formal analysis, Writing-review and editing.

Gehui Li: Formal analysis, Writing-review and editing.

Shan Duan: Conceptualization, Resources, Supervision, Writing-review and editing.

All authors gave final approval of the version to be published; have agreed on the journal to which the article has been submitted; and agree to be accountable for all aspects of the work.

Disclosure

The authors report no conflicts of interest in this work.