High-Throughput Sequencing and Co-Expression Network Analysis of lncRNAs and mRNAs in Early Brain Injury Following Experimental Subarachnoid Haemorrhage

Abstract

Subarachnoid haemorrhage (SAH) is a fatal neurovascular disease following cerebral aneurysm rupture with high morbidity and mortality rates. Long non-coding RNAs (lncRNAs) are a type of mammalian genome transcript, are abundantly expressed in the brain and are involved in many nervous system diseases. However, little is currently known regarding the influence of lncRNAs in early brain injury (EBI) after SAH. This study analysed the expression profiles of lncRNAs and mRNAs in SAH brain tissues of mice using high-throughput sequencing. The results showed a remarkable difference in lncRNA and mRNA transcripts between SAH and control brains. Approximately 617 lncRNA transcripts and 441 mRNA transcripts were aberrantly expressed at 24 hours after SAH. Gene ontology (GO) enrichment and Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis indicated that the differentially expressed mRNAs were mostly involved in inflammation. Based on the lncRNA/mRNA co-expression network, knockdown of fantom3_F730004F19 reduced the mRNA and protein levels of CD14 and toll-like receptor 4 (TLR4) and attenuated inflammation in BV-2 microglia cells. These results indicate that lncRNA fantom3_F730004F19 may be associated with microglia induced inflammation via the TLR signaling pathway in EBI following SAH. LncRNA represent a potential therapeutic target for the prognosis, diagnosis, and treatment of SAH.

Introduction

Subarachnoid haemorrhage (SAH) is a devastating cerebrovascular disease. Although SAH accounts for only 5% of all strokes, this type of haemorrhage affects a significant percentage of the population worldwide with high morbidity and mortality^1,2. Early brain injury (EBI) refers to the acute pathophysiological event that occurs within the first 72 h of SAH. Differ from the fact that the reversal of cerebral vasospasm (CVS) does not appear to improve outcome, EBI has been demonstrated to be the main cause of death or disability in SAH. Therefore, to a large extent, EBI determines the prognosis of SAH patients³. Despite the vast efforts made in EBI research following SAH, the mechanisms of the pathological process in the acute phase of SAH require further investigation⁴.

Long non-coding RNAs (lncRNAs) are a type of RNA transcript lacking protein-coding potential (>200 nucleotides). For some lncRNAs, a high degree of tissue specificity has been identified. Remarkably, lncRNAs are expressed abundantly in the nervous system, and approximately 40% lncRNAs are detected specifically in the brain^5,6,7. Previous studies have demonstrated that lncRNAs play important roles in regulating the pathological processes of neurological and psychiatric diseases^8,9,10. To date, however, lncRNA expression signatures and the co-expression network of lncRNAs and mRNAs in EBI after SAH remain poorly understood. Furthermore, the function of lncRNAs in EBI following SAH also requires further study.

It is imperative that appropriate animal models of SAH are used for research. The current SAH models mainly utilize endovascular perforation and blood injection techniques. However, the endovascular perforation model better mimics EBI, while the blood injection model better mimics vasospasm^11,

Figure 2: Representative map** of sequence reads in the Control and SAH groups.

Overall coverage of the detected reads was mapped to mouse chromosomes.

Table 1 Map** of sequence reads in the control and SAH groups.

Differentially expressed lncRNAs and mRNAs

All RNA samples were subjected to high-throughput sequencing analysis of lncRNA and mRNA expression. All the sequencing data were filtered using the volcano plot to illustrate the differentially expressed coding genes and non-coding genes (log2fold change >1; q < 0.05) (Fig. 3A). All of the significant differentially expressed genes were distributed among almost all of the chromosomes (Fig. 3B). The average lncRNA transcript was 2,289 bp, and the average mRNA transcript was 2,497 bp (Fig. 3C; Supplementary Table S2; Supplementary Table S3). Among all significant differentially expressed genes in SAH and control samples (log2fold change > 1; q < 0.05), there were 103 upregulated and 514 downregulated lncRNA transcripts (Fig. 3D) and 387 upregulated and 54 downregulated mRNA transcripts (Fig. 3E), respectively. In addition, fantom3_F730004F19, one of the upregulated lncRNAs in SAH brains compared with levels in the controls (fold change: 15.82, log2fold change: 3.98; P < 0.05), showed the closest correlation with a significant differentially expressed mRNA (CD14, fold change: 2.51, log2fold change: 1.33; P < 0.05) in the co-expression network of lncRNA transcripts and mRNA transcripts. Moreover, many transcripts did not align with any known mRNA or lncRNA, and their coding potential was analysed. Those transcripts longer than 200 bp but without coding potential were identified as lncRNAs. A total of 150 newly identified lncRNAs were differentially expressed between the control group and SAH group. The top differentially expressed transcripts are shown in Table 2.

Figure 3: Differentially expressed lncRNAs and mRNAs between mouse brains at 24 h after SAH and in control groups.

(A) The volcano plot illustrates the differentially expressed genes (including lncRNAs and mRNAs) between the control and SAH groups, with red dots representing differentially expressed genes (log2fold change > 1; q < 0.05). (B, C) Chromosome distribution (B) and average length (C) of the differentially expressed lncRNAs and mRNAs. (D,E) The hierarchical cluster analysis generated heat maps of the top 100 significant differential expressed lncRNAs (D) and mRNAs (E) in the mouse brains at 24 h following SAH or in control groups (log2fold change >1.5; q < 0.05). Colour represents the log10 (RPKM + 1) value, with red indicating upregulated genes and green indicating downregulated genes. The legend represents the standard normal distribution of log10 (RPKM + 1) values, and the blue line in the legend indicates the gene amounts in this colour area. SAH, subarachnoid haemorrhage; lncRNAs, long non-coding RNAs; mRNAs, messenger RNAs; RPKM, expected number of Reads Per Kilobase of transcript sequence per Millions base pairs Sequenced.

Table 2 Top 10 up-regulated and 10 down-regulated new lncRNAs.

GO and KEGG analysis

The function of lncRNAs is thought to be reflected in their associated protein-coding genes¹³. To clarify the potential function of these differentially expressed mRNA-related lncRNAs, gene ontology (GO) enrichment analysis of differentially expressed mRNAs was used. Three domains (biological processes, cellular components and molecular functions) of GO enrichment analysis were investigated. The significance of GO analysis in each domain was denoted by the false discovery rate (FDR) (FDR < 0.05 is recommended) (Fig. 4A–C). The response to wounding, immune system process, and inflammatory response were the most enriched terms in biological processes (BP) (Fig. 4A). Extracellular region, extracellular region part and extracellular space were the most enriched terms in cellular components (CC) (Fig. 4B). Protein binding, binding and chemokine activity were the most enriched terms in molecular functions (MF) (Fig. 4C). Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis indicated that the most enriched pathways involving significant differential expressed mRNAs were the tumour necrosis factor (TNF) signaling pathway, osteoclast differentiation and chemokine signaling pathway (Fig. 4D).

Figure 4: GO enrichment and KEGG pathway analysis of differentially expressed genes in SAH brains.

(A) BP, (B) CC and (C) MF presented the top 10 significance terms of GO enrichment analysis (P < 0.05 and FDR < 0.05). (D) The top 20 KEGG pathways of significantly differentially expressed genes between the control and SAH groups (P < 0.05 and FDR < 0.05). GO, Gene Ontology; KEGG, Kyoto Encyclopedia of Genes and Genomes; BP, biological processes; CC, cellular components; MF, molecular functions; FDR, false discovery rate; DE, differentially expressed.

LncRNA/mRNA transcripts co-expression networks

To reveal the key lncRNAs and their potential functions in SAH, we constructed a lncRNA/mRNA co-expression network, and investigated the potential interactions between the lncRNA transcripts and mRNA transcripts. More than 1,800 network nodes were composed. Ten network nodes with high correlation (COR)-values were selected, and the co-expression network was established using Cytoscape software (Fig. 5A). Most of the lncRNA-coexpressed mRNA transcripts were involved in the inflammatory response (biological process), extracellular region (cellular component) and protein binding (molecular function) according to GO enrichment. Moreover, we found that CD14, a ligand for toll-like receptor 4 (TLR4), which is involved in regulation of the TLR signaling pathway, showed the closest correlation with one lncRNA (fantom3_F730004F19) among all significant differentially expressed mRNA transcripts with the highest COR-value (COR: 0.999, P < 0.05) (Fig. 5B; Supplementary Table S4).

Figure 5: LncRNA/mRNA transcripts co-expression network after SAH.

(A) The network represents the co-expression correlations between the significantly differentially expressed mRNA and lncRNA transcripts (|COR| > 0.95, P < 0.05). Circles indicate lncRNA transcripts and triangles indicate mRNA transcripts. Solid lines indicate positive correlations, and dashed lines indicate negative correlations. Red represents upregulated, and green represents downregulated. (B) KEGG analysis suggested that lncRNA-coexpressed mRNAs were mainly targeted to the Toll-like receptor signaling pathway (http://www.genome.jp/kegg)^54,55. Blue represents significant genes in this pathway of SAH, and pentagrams represent significant lncRNA-coexpressed genes.

qRT-PCR validation

To validate the RNA-seq data, qRT-PCR was performed using 5 randomly selected lncRNAs (3 upregulated and 2 downregulated). qRT-PCR analysis revealed that the expression levels of fantom3_E430008K20, fantom3_C730003K16 and fantom3_I830129C17 were upregulated, while those of fantom3_A430024L20 and fantom3_C330006P03 were downregulated in SAH brains. The qRT-PCR results confirmed the accuracy of RNA-seq (Fig. 6D).

Figure 6: Gene expression profile.

(A,B) Subcellular localization pattern of fantom3_F730004F19 by RNA FISH staining. A significantly higher expression of fantom3_F730004F19 in treated BV-2 microglia cells was estimated, and fantom3_F730004F19 was predominantly expressed in the nucleus. (C) Cellular localization pattern of fantom3_F730004F19 by qRT-PCR. Although there was no statistical difference between BV2 and HT22 cells, the level of fantom3_F730004F19 present a higher in BV2 cell line, lesser extent in HT22 cells and MOPC cells, and not detected in MA cells. (D) Comparison between sequencing and qRT-PCR analysis of 5 randomly selected lncRNAs. Positive values refer to upregulation, and negative values refer to downregulation. β-actin was used to normalize the expression of samples. Red representes RNA-seq data, and blue representes qRT-PCR. The bars represent standard error of the mean (SEM). The qRT-PCR results were closely correlated with the sequencing data (P < 0.05). (E) The silencing efficiency of the lentivirused on fantom3_F730004F19 in BV-2 microglia cells. Lentivirus-50305 (KD1) and lentivirus-50307 (KD3) significantly inhibited fantom3_F730004F19 expression (P < 0.05 versus Blank group). (F) Inhibition of fantom3_F730004F19 reduced the expression of CD14 and TLR4 in BV-2 microglia cells following LPS treatment. The expression of fantom3_F730004F19, CD14 and TLR4 were overexpressed in BV-2 cells after LPS treatment. Knockdown of fantom3_F730004F19 abated the increase of CD14 and TLR4 in LPS-treated BV-2 cells. ^@P < 0.05 and ^@@P < 0.01; *P < 0.05 and **P < 0.01 versus the NC normal group; ^#P < 0.05 and ^##P < 0.01 versus the NC LPS-treated group. BV-2: mouse microglial cell line; HT22: mouse neuronal cell line; MA: mouse astrocyte cell line; MOPC: mouse oligodendrocyte precursor cell line; N: normal condition group; LPS: lipopolysaccharide treated; Blank: non-transfected cells; NC: negative control lentivirus transfected cells; KD: the lncRNA fantom3_F730004F19 lentivirus transfected cells.

Cellular and subcellular localization pattern of fantom3_F730004F19

Mouse microglial cell line BV-2 cells, neuronal cell line HT22 cells, astrocyte cell line MA cells and oligodendrocyte precursor cell line MOPC cells were used to identify the distribution of fantom3_F730004F19 in cells of the central nervous system (CNS). Although there was no statistical difference between BV2 and HT22 cells, qRT-PCR indicated BV2 cell line presented a higher level of fantom3_F730004F19 (Fig. 6C). After treatment with lipopolysaccharide (LPS) in BV-2 microglia cells, a significantly higher expression level of fantom3_F730004F19 was identified with RNA FISH staining. Additionally, fantom3_F730004F19 was predominantly expressed in the microglial nuclei (Fig. 6A,B).

Knockdown of fantom3_F730004F19 reduced the expression of inflammation-related genes

Three different lentiviruses were transfected into BV-2 cells to inhibit fantom3_F730004F19. The expression of fantom3_F730004F19 was significantly reduced after transfection with lentiviral vectors. Compared with the levels in the negative control (nonsense lentivirus, NC), lentivirus-50305 (KD1) and lentivirus-50307 (KD3) reduced the fantom3_F730004F19 levels by > 50% (P < 0.05) (Fig. 6E). Lentivirus-50307 (KD3) presented a specific lentiviral-mediated knockdown of fantom3_F730004F19 (Supplementary Figure S3). We next employed lentivirus-50307 (KD3) to inhibit fantom3_F730004F19 expression for subsequent functional analysis. After treatment of BV-2 microglia cells with LPS, the levels of CD14 and TLR4 mRNA were sharply increased, and fantom3_F730004F19 expression was dramatically upregulated (P < 0.05). Lentivirus (KD3) treatment significantly suppressed the expression of fantom3_F730004F19 in both the normal group and the LPS group, and the elevation in CD14 and TLR4 mRNA was also inhibited (P < 0.05) (Fig. 6F). These findings suggested that fantom3_F730004F19 regulated the levels of CD14 and TLR4 mRNA.

Knockdown of fantom3_F730004F19 attenuated the inflammatory response

Compared with the levels in the Blank and NC groups, the expression levels of CD14 and TLR4 proteins were dramatically suppressed after lentivirus treatment in LPS-treated BV-2 microglia cells (P < 0.05), but, this effect was not observed in normal conditions (Fig. 7A–F). We further quantified the protein levels of TNF-α, IL-1β and IL-6 in the BV-2 supernatants. After lentivirus treatment, the protein levels of IL-1β (Fig. 7G), IL-6 (Fig. 7H) and TNF-α (Fig. 7I) were significantly suppressed in the LPS group (P < 0.05). These findings indicated that knockdown of fantom3_F730004F19 attenuated microglia-related inflammation.

Figure 7: Silencing fantom3_F730004F19 attenuated the inflammatory response.

(A–F) Inhibition of fantom3_F730004F19 reduced the expression of CD14 and TLR4 in BV-2 cells following LPS treatment. Protein expression was shown to be completely depleted by both western blot analysis (A–C) and immunofluorescence staining (D–F) (full-length blots/gels are presented in Supplementary Figure S5). Although the inhibitory effect was not obvious in the normal condition, silencing fantom3_F730004F19 significantly reduced the expression of CD14 and TLR4 protein levels after LPS treatment (^#P < 0.05 and ^##P < 0.01 versus the NC LPS-treated group). (G–I) Inhibition of fantom3_F730004F19 attenuated inflammatory cytokines release. Silence fantom3_F730004F19 significantly reduced the release of IL-1β (G), IL-6 (H) and TNF-α (I) after LPS treatment (^#P < 0.05 and ^##P < 0.01 versus the NC LPS-treated group). N: normal condition group; LPS: lipopolysaccharide treated; Blank: non-transfected cells; NC: negative control lentivirus transfected cells; KD: the lncRNA fantom3_F730004F19 lentivirus transfected cells.

Discussion

In summary, significantly differentially expressed lncRNA and mRNA transcripts were examined in the present study. We identified 617 lncRNAs and 441 mRNAs that were aberrantly expressed at 24 h after SAH. KEGG and GO analysis revealed that these differentially expressed mRNA genes involved in many pathophysiological processes, including the inflammatory responses. Meanwhile, the vast majority of lncRNA-coexpressed mRNAs were also associated with inflammation. Specifically, we found that among all lncRNAs, lncRNA fantom3_F730004F19 has the closest relationship with differentially expressed mRNA transcripts, which was correlated with CD14 at the highest COR value (COR: 0.999). Silencing of fantom3_F730004F19 resulted in reduced expression of CD14 and TLR4 at both the mRNA and protein levels. Moreover, knockdown of fantom3_F730004F19 attenuated inflammation in BV-2 microglia cells. These results provide initial experimental evidence that lncRNAs may have some specific effects on the pathological processes of EBI in SAH by regulating inflammation.

The neuroinflammatory response in EBI after SAH has been identified in both clinical and experimental studies^14,15,16. Accordingly, anti-inflammatory intervention is a novel promising area of research for SAH treatment¹¹. Inflammatory cytokines, such as TNF-α and IL1β, are strongly associated with brain injury after SAH and have been demonstrated to exert inflammatory and excitotoxic effects¹⁷, initiate neuronal apoptosis¹⁸, activate matrix metalloprote 9 (MMP-9) and cause blood brain barrier (BBB) disruption^19,20. We systematically analysed the functions of differentially expressed mRNAs by GO annotation and pathway analysis. Our experimental findings revealed inflammatory responses in the mouse brain at 24 h after SAH. Significantly differentially expressed genes related to cytokine pathways were observed in our experiment, including Il1b and Tnf (gene names). The genes of chemokine family, including Ccl2 (also known as MCP1), Ccl3 (also known as MIP-1α), Ccl4 (also known as MIP-1β), which have been shown to be involved in inflammatory responses^21,22, were also differentially expressed at 24 h after SAH. These results are consistent with the common view that neuroinflammation contributes to EBI after SAH.

Based on recent advances in genome sequencing techniques, studies targeting the roles of lncRNAs in inflammation and the immune response has been conducted^23,24. In particular, aberrant lncRNA expression has been shown to be involved in neurological disorders, such as neurodegenerative diseases²⁵, schizophrenia²⁶, autism²⁷, intellectual disability and developmental delay^28,29. Zheng et al. elucidated the expression signatures of lncRNAs in a blood injection model using microarray assays³⁰. However, the discovery rate of high-throughput sequencing is higher than that of microarray technology. To better mimic EBI after SAH^11,51. The gene-level RPKM values were then normalized using the log10 values (RPKM + 1) for further analyses. Differential expression was determined with DEGseq software, and the q-value was used to denote the significance of the P-value (q-value < 0.05 is recommended)⁵². The data were deposited in Gene Expression Omnibus (GEO) with accession number GSE79416.

Bioinformatics analysis

All differentially expressed mRNAs were selected for GO and KEGG pathway analyses to investigate the potential role of the lncRNAs-coexpressed with mRNAs. GO was performed with KOBAS2.0 software. GO provides label classification of gene function and gene product attributes (http://www.geneontology.org). GO analysis covers three domains: cellular component (CC), molecular function (MF) and biological process (BP)⁵³. The false discovery rate (FDR) was used to denote the significance of the P-value (FDR < 0.05 is recommended). The differentially expressed mRNAs and the enrichment of different pathways were mapped using the KEGG pathways with KOBAS2.0 software (http://www.genome.jp/kegg)^54,55. The significance of the KEGG pathways among differentially expressed genes was denoted by the FDR (FDR < 0.05 is recommended).

Co-expression network of differentially expressed lncRNAs/mRNAs

To investigate the potential functions of differentially expressed lncRNAs and the interactions between mRNAs and lncRNAs, we constructed a lncRNA/mRNA transcripts co-expression network. We calculated the Pearson correlation coefficient (PCC), and the COR-value was used to calculate the correlation coefficient of the PCC between lncRNA and mRNA transcripts (not including lncRNA-lncRNA transcripts or mRNA-mRNA transcripts PCC). Considering the small sample (n < 8) and stochastic factors, the permutation test was performed. The Z score values were used to normalize the PCC, and then the P-value was calculated⁵⁶ (|COR| > 0.95 and p-value < 0.05 is recommended). The co-expression network was illustrated using Cytoscape software⁵⁷.

Cell culture and treatments

The BV-2 mouse microglial cell line was cultured in Dulbecco’s modified Eagle’s medium (DMEM, Gibco, Grand Island, CA, USA) supplemented with 10% foetal bovine serum (FBS, Gibco, Grand Island, CA, USA) and 50 mg/mL of penicillin-streptomycin and maintained in ahumidified incubator at 37 °C, with 5% CO₂. Seventy-two hours after the lentiviral vectors were transfected into BV-2 cells, LPS (Sigma, St. Louis, MO, USA) at a concentration of 1 μg/mL was added to cells of the LPS group. Then, the cells were cultured for 24~48 h. The HT22 mouse neuronal cell line and MA astrocyte cell line were cultured in DMEM-F12 (Gibco, Grand Island, CA, USA), and the MOPC oligodendrocyte precursor cell line was cultured in MM (Gibco, Grand Island, CA, USA), supplemented with 10% FBS (Gibco, Grand Island, CA, USA) and 50 mg/mL of penicillin-streptomycin. Cells were maintained in a humidified incubator at 37 °C, with 5% CO₂.

Lentiviral vector production and infection

Three different lentiviral vectors (termed 50305, 50306 and 50307; Genechem, Shanghai, China) targeting lncRNA fantom3_F730004F19 were transfected into BV-2 cells with the polybrene lentiviral vector transfection reagent (Roche, Mannheim, Germany). Non-transfected cells were used as the Blank control, negative control lentivirus was transfected as the negative control (NC), and the lncRNA fantom3_F730004F19 shRNA lentiviruses were transfected for the knock down (KD) groups. The sequences of the 3 lentiviral vectors were as follows:

Lentivirus-50305(KD1): TTCCTAAGGACTGGAAACATA

Lentivirus-50306(KD2): GAGGACAAGTCTGGAAGTCAA

Lentivirus-50307(KD3): TGACACAGGGCTACAGGGTAT

The efficacy of lentivirus transfection was evaluated by qRT-PCR.

Validation of gene expression by qRT-PCR

To identify the veracity of sequencing, we randomly selected 5 lncRNA transcripts for validation. Total RNA was extracted from the control and SAH brain samples using TRIzol reagent (Invitrogen Life Technologies, Carlsbad, CA, USA) and then reverse transcribed to cDNA. Five randomly selected lncRNAs and their expression levels were further assessed by qRT-PCR using iQ^TM SYBR^® Green Supermix (Bio-Rad Laboratories, Inc. Hercules, CA, USA) with a CFX96 Touch™ Real-Time PCR instrument (Bio-Rad Laboratories, Inc. Hercules, CA, USA). To identify the silencing efficiency of the lentivirus and the influence of target genes, the total RNA was extracted from each group of BV-2 cells. All samples were normalized to the expression of β-actin, and the experiment was repeated 3 times. The detailed primer information is available as supplemental data (Supplementary Table S1).

Fluorescence in situ hybridization experiments (FISH)

FISH staining was performed as previously described⁵⁸. Briefly, BV-2 cells were washed twice with 1X PBS and then fixed with 4% paraformaldehyde for 10 min. After washed 3 times with 1X PBS, the cells were permeabilized with 0.5% Triton X-100 for 10 min. The cells were treated with pre-hybridization buffer for 30 min at 37 °C and then the appropriate amount of probe in a hybridization solution was applied overnight in a humidified chamber at 37 °C. Cells were then washed twice for 30 min at 42 °C with 0.1% Tween-20 in 4X SSC. DAPI was applied during the second wash. Cells were then rinsed twice with 1X PBS before imaging. Images were captured with afluorescence microscope (Eclipse Ti-S; Nikon, Tokyo, Japan).

Western blotting analyses

Cells were washed twice with chilled PBS and lysed directly in wells by incubating with RIPA lysis buffer supplemented with a protease inhibitor (Roche, Basel, Switzerland) for 120 h post-transfection. The primary antibodies included mouse monoclonal antibody against CD14 (60253-1-Ig, Proteintech, Wuhan, China), mouse monoclonal antibody against TLR4 (ab 8376, Abcam, Cambridge, UK), and mouse monoclonal anti-beta-actin (#SC-47778, Santa Cruz, CA, USA).

Immunofluorescence staining

Immunofluorescence staining was conducted as previously described³⁵. Primary antibodies included mouse monoclonal antibody against CD14 (1:200, 60253-1-Ig, Proteintech, Wuhan, China), mouse monoclonal antibody against TLR4 (1:200, ab 8376, Abcam, Cambridge, UK). Secondary antibodies included a DyLight 488, goat anti-mouse IgG (A23210, Abbkin, California, USA). Nuclei were stained with 40′, 6-diamidino-2-phenylindole (C1006; Beyotime, Nan**g, China). Images were captured with afluorescence microscope (Eclipse Ti-S; Nikon, Tokyo, Japan). Image-pro plus (IPP) 6.0 software was used for immunofluorescence staining analysis.

Enzyme-linked immunosorbent assay (ELISA)

The protein levels of IL-1β, IL-6 and TNF-α in the BV-2 supernatants were quantified by ELISA according to the manufacturer’s instructions (Boster, Wuhan, China). The relative protein content of IL-1β, IL-6 and TNF-α was shown as picogram per milligram of total protein. The protein content of each supernatants sample was detected with a BCA kit (Beyotime, Shanghai, China).

Statistical analysis

All data are presented as the mean ± SEM unless otherwise stated. Comparisons between two groups were analysed using the Student’s t-test. P < 0.05 was considered to indicate a statistically significant difference.