Lipopolysaccharide-induced miR-1224 negatively regulates tumour necrosis factor-α gene expression by modulating Sp1

Abstract

The innate immune response provides the initial defence mechanism against infection by other organisms. However, an excessive immune response will cause damage to host tissues. In an attempt to identify microRNAs (miRNAs) that regulate the innate immune response in inflammation and homeostasis, we examined the differential expression of miRNAs using microarray analysis in the spleens of mice injected intraperitoneally with lipopolysaccharide (LPS) and saline, respectively. Following challenge, we observed 19 miRNAs up-regulated (1·5-fold) in response to LPS. Among these miRNAs, miR-1224, whose expression level increased 5·7-fold 6 hr after LPS injection and 2·3-fold after 24 hr, was selected for further study. Tissue expression patterns showed that mouse miR-1224 is highly expressed in mouse spleen, kidney and lung. Transfection of miR-1224 mimics resulted in a decrease in basal tumour necrosis factor-α (TNF-α) promoter reporter gene activity and a down-regulation of LPS-induced TNF-α mRNA in RAW264.7 cells. With public databases of miRNA target prediction, miR-1224 was shown to bind to the 3′ untranslated region (UTR) of Sp1 mRNA, whose coding product controls TNF-α expression at the transcriptional level. Furthermore, we found that in HEK-293 cells, the activity of the luciferase reporter bearing Sp1 mRNA 3′ UTR was down-regulated significantly when transfected with miR-1224 mimics. After transfection of miR-1224 in RAW264.7 cells, nucleus Sp1 protein level decreased, and when endogenous miR-1224 was blocked, the decrease was abolished. Therefore, we initially speculated that miR-1224 was a negative regulator of TNF-α in an Sp1-dependent manner, which was confirmed in vivo by chromatin immunoprecipitation assay, and might be involved in regulating the LPS-mediated inflammatory responses.

Introduction

The innate immunity with which we are born is the first line defending us from infection by other organisms.¹ Inflammation is one of the most common defence mechanisms in the body responding to infection. Invasion of micro-organisms into the host body triggers the release of inflammatory molecules including cytokines, chemokines, interferons, reactive oxygen species and nitrogen intermediates.^2–4 The presence of these molecules is essential to the host for bacterial killing and clearance. However, an excessive inflammatory response can cause severe tissue injury, and may even lead to death. So, when inflammatory molecules and pathways involved in innate immune response are activated, anti-inflammatory molecules and inhibitory pathways are triggered simultaneously to protect the host from inflammatory damage.⁵

Lipopolysaccharide (LPS) is an endotoxin and is a major component of the outer membrane of Gram-negative bacteria.⁶ Exposure of host cells to LPS can induce an inflammatory response.⁷ Upon stimulation, LPS binding to the CD14–Toll-like receptor 4 (TLR4) complex triggers dimerization and structural changes in the receptor that lead to the recruitment of adaptor proteins that activate downstream signalling. Signalling from TLR4 is transduced through the mitogen-activated protein kinase (MAPK) and nuclear factor-κB (NF-κB) pathway, finally leading to activation of transcription factors,^8,9 which in turn initiates the transcription of a range of pro-inflammatory cytokines as well as anti-inflammatory molecules. These anti-inflammatory molecules comprise molecules that are required for the response to LPS, and that promote the expression of genes that inhibit the response.¹⁰ The GC box-binding protein, Sp1, is a member of a family of zinc finger transcription factors. Sp1 is a ubiquitous transcription factor that has been implicated in the regulation of a large number of genes by binding promoter or as co-regulator interacting with other transcription factors.¹¹ In response to LPS, Sp1 is activated via the p38/MAPK pathway, which in turn regulates the expression of genes involved in the process.

In mammals, microRNAs (miRNAs) are a class of non-coding RNA that are 19–25 nucleotides in length, generated from endogenous hairpin-shaped transcripts.¹² The miRNAs function as regulators of the protein-coding gene by pairing with the 3′ untranslated region (UTR) of these genes to direct degradation or translation repression.¹³ Studies have revealed that miRNAs play important roles in a series of processes, including cell proliferation and differentiation, apoptosis, insulin secretion, cardiac and skeletal muscle development and immune response.^14,15 The first investigation on miRNA-associated immunity was performed by Chen et al. In this study, over-expression of miR-181 in haematopoietic precursor cells promoted lymphoid differentiation toward a B-cell lineage.¹⁶ Except for two miRNAs related to innate immune function that have been described, the involvement of miRNAs in the innate immune response is unclear.^17,18 Recently, miR-147 was identified as a negative regulator of TLR-associated signalling events in murine macrophages.¹⁹ However, all of these studies were performed in vitro, and their results may differ greatly from those obtained in vivo. As one of the largest lymphoid organs in the body, the spleen is a home for various immune cells, it plays an important role in scavenging dead tissue and other worn out cells, and helps to produce inflammation and immune homeostasis. To identify other miRNAs that regulate the innate immune response and homeostasis in vivo, we analysed the miRNA transcriptome using high-throughput miRNA microarray technology in the spleens of mice injected with LPS or saline. We found 19 miRNAs that were expressed more in spleens of LPS-treated mice than of control mice. Among these, miRNAs, miR-1224 was selected for further study. Both miR-146b and miR-155, which have the opposite regulatory effect on the LPS-mediated inflammatory response, were taken as experimental controls in this study.

Materials and methods

Sample collection

All animal procedures were performed according to guidelines developed by the China Council on Animal Care and protocol approved by the Animal Care and Use Committee of Guangdong Province, China.

Ten-week-old female specific pathogen-free BALB/c mice were purchased from Guangdong Laboratory Animal Centre (Guangdong, China). Mice were divided into three groups (five mice per group), LPS-I group, LPS-II group and saline group. Two mice from each group were used for microarray analysis, and the remaining three were taken for microarray data confirmation using real-time PCR. In the LPS-I and LPS-II groups, mice were intraperitoneally injected with LPS (Escherichia coli strain 0111:B4, Sigma, Saint Louis, MO, USA; 4 mg/kg). Control mice in the saline group were intraperitoneally injected with saline. Mice of the LPS-I and saline groups were killed 6 hr after injection, those in the LPS-II group were killed 12 hr after injection. All mouse spleens were dissected and total RNAs were isolated using the mirVana™ miRNA isolation kit (Ambion, Austin, TX) for microarray analysis. Tissue from the heart, liver, spleen, lung, kidney and muscle from untreated mice were also collected. Total RNAs isolated using TRIzol reagent (Invitrogen, Carlsbad, CA) were used for tissue expression pattern analysis.

miRNA microarray analysis

Agilent miRNA microarray was used to measure the expression level of miRNAs in spleen tissues from saline/LPS-treated mice. Briefly, 100 ng pCp-Cy3-labelled total RNA per sample was used for hybridization on each Agilent miRNA microarray chip, which contained 627 mouse miRNAs and 39 mouse γ-herpesvirus miRNAs. After washing, slides were scanned on an Agilent microarray scanner at 100% and 5% sensitivity settings. Image analysis was performed using Agilent Feature Extraction software version 8.1 (Agilent, CA, USA). Data analysis was performed in GeneSpring software version X. The quantified signals were normalized using a quantile normalization method to make the distribution of probe intensities for each array in a set of arrays the same. Then, normalized data were filtered using flag value = 6, making sure that only the miRNAs with acceptable values in all samples were taken to the next step for analysis. Differentially expressed miRNAs between LPS treatment and control were identified using a fold-change method with cut-off of 1·5.

Cell culture and treatment

Mouse macrophage-like RAW264.7 and HEK293 cells were cultured in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum and penicillin/streptomycin, and incubated at 37° in a humidified atmosphere of 5% CO₂. For cell stimulation, LPS was added to the medium at a final concentration of 1 μg/ml. Cells were harvested and total RNAs were extracted at different time-points after addition of LPS. The supernatant of RAW264.7 cells was collected for tumour necrosis factor-α (TNF-α) protein detection using an ELISA kit (Invitrogen, Camarillo, CA). Real-time PCR was used to detect the changes in expression level of miRNAs, and of TNF-α and TLR4 mRNA.

SYRB Green-based real-time PCR

Relative expression of miRNA was measured on total RNA extracts using an SYBR Green-based real-time PCR. The sequences for miRNAs, mRNA and reference gene were obtained from miRBase (http://www.mirbase.org/) and GenBank (http://www.ncbi.nlm.nih.gov/), and miRNA-specific stem-loop reverse transcription (RT) primers were designed as previously described.²⁰ The reverse transcriptase reaction was performed using a Moloney-murine leukaemia virus (M-MLV) reverse transcriptase kit (Promega, Madison, WI, USA). Each reaction was carried out as follows: 2 μg total RNA, 100 nm of each of miRNAs and Mus musculus U6 small nuclear RNA-specific stem-loop reverse transcript primers mix, 0·5 μg oligo-d(T) were mixed and nuclease-free water was added up to a volume of 13 μl; samples were then incubated at 65° for 5 min before being chilled on ice, and 7 μl RT mix containing the following components was added: 4 μl M-MLV 5 × RT Reaction Buffer, 2 μl dNTP mix (10 mm each), 0·5 μl RNase inhibitor (40 unit/μl) and 0·5 μl M-MLV. The samples were then incubated at 42° for 60 min, 85° for 5 min to inactivate the transcriptase, and then held at 4°. The RT products were stored at − 20° for use. Real-time PCR was performed using an SYBR Green qPCR mix kit (Takara, Kyoto, Japan) according to the manufacturer's protocol on a LightCycler480 system (Roche Diagnostics, Indianapolis, IN). U6 small nuclear RNA and glyceraldehyde 3-phosphate dehydrogenase were selected as internal control genes for the miRNAs and mRNA relative expression assays, respectively.

Evaluation of the role of miR-1224 in TNF mRNA transcription

RAW264.7 cells were cultured in six-well plates, transiently transfected with miR-1224 mimics (Qiagen, GmbH, Hilden, Germany) with a final concentration of 100 nm using Lipofectamine LTX and PLUS reagent (Invitrogen). Sixteen hours after transfection, LPS was added to the medium at a final concentration of 100 ng/ml per well. Three hours later, cells were harvested and the TNF-α, interleukin-1b (IL-1b) and Il6 mRNA levels were measured using real-time PCR.

To test the effect of miR-1224 on the activity of TNF mRNA promoter, a mouse TNF promoter-driven luciferase reporter assay was performed. RAW264.7 cells were co-transfected with miR-1224 mimics at a final concentration of 100 nm, luciferase reporter plasmid (containing the promoter of mouse TNF before luciferase gene in pGL3-basic vector) and Renilla luciferase control vector (pRL-TK-Promema) using Lipofectamine™ LTX and Plus™ Reagents (Invitrogen) according to the manufacturer's protocol. Twenty-four hours after transfection, cells were washed with cooled PBS and luciferase reporter assay was performed using the dual luciferase reporter assay system (Promega). Firefly luciferase activity was normalized to Renilla luciferase activity and the experiments were performed three times in triplicate. Statistical differences were determined using Student's t-test.

Bioinformatics analysis of miRNA–mRNA interaction

To understand the miRNA–mRNA network that was associated with the LPS response, the prediction of LPS-induced miRNA putative target genes was performed using TargetScan software in GeneSpring, into which the TargetScan database (Version 5.0, http://www.targetscan.org/ver50/) was integrated. Based on the context percentile = 50, putative conserved target genes of each specific miRNA were identified. All target genes were then put into GeneSpring software to organize the genes and miRNAs into a network. RNAhybrid software (http://bibiserv.techfak.uni-bielefeld.de/rnahybrid/submission.html) was used to confirm the prediction results from TargetScan.

miRNA-3′ UTR-mediated repression assay

Two single strands of 3′ UTR of the mouse SP1 gene containing miR-1224 potential binding site were synthesized and annealed to double strands. There are two restriction enzyme sites, XhoI and NotI, on the two ends of the double strand, respectively. Then the double strand was constructed into the psiCHECK-2 vector. HEK293 cells were cultured in 24-well plates and co-transfected with the psiCHECK-2-SP1-3UTR plasmid and miR-1224 mimics at a final concentration of 50 nm using Lipofectamine 2000 (Invitrogen). Twenty-four hours after transfection, cells were washed with cooled PBS and luciferase reporter assay was performed as previously described in 2·5. Renilla luciferase activity was normalized to firefly luciferase activity. The experiments were performed three times in triplicate. Statistical differences were determined using Student's t-test.

Western blotting

The miR-1224 mimics, inhibitor and control were transfected into RAW264.7 cells to estimate the role of miR-1224 on the activity of NF-κB and the expression of Sp1 protein using Lipofectamine™ LTX and Plus™ reagents. After 36 hr of transfection, nuclear proteins were extracted from RAW264.7 cells using Nuclear and Cytoplasmic Protein Extraction Kits according to the manufacturer's protocol (Beyotime Institute of Biotechnology, China). Bicinchoninic acid assay was used to determine the protein concentration as previously described.²¹ Equal amounts of nuclear protein (20 μg/lane) were separated on 12% SDS–polyacrylamide gel under denaturing conditions, and transferred onto nitrocellulose membrane (Pall Corporation, Washington, NY, USA). After incubation in blocking solution (5% BSA in Tris-buffered saline–Tween-20) at room temperature for 1 hr, the membrane was incubated with primary antibody rabbit anti-NF-κB p65, mouse anti-Sp1 at 4° overnight. The membrane was washed and incubated with secondary antibody conjugated to horseradish peroxidase (HRP) at temperature (25°) for 1 hr. The HRP labelling was detected using an enhanced chemiluminescence Western blot kit (Pierce).

miR-1224 did not inhibit mutated 3′ UTR of Sp1 mRNA

To confirm whether miR-1224 negatively regulates TNF-α by targeting Sp1, we constructed two recombined plasmids using pcDNA3.1 vector: Sp1-wt3UTR, which contains 3′ UTR of Sp1 mRNA downstream of the Sp1 open reading frame (ORF), and Sp1-mut3UTR, which contains 3′ UTR-mutated Sp1 mRNA downstream of Sp1 ORF. RAW264.7 cells were co-transfected with miR-1224 mimics and Sp1-wt3UTR and Sp1-mut3UTR, respectively. Basal TNF-α mRNA level was measured by real-time PCR.

Chromatin immunoprecipitation assay

We performed a chromatin immunoprecipitation assay in RAW264.7 cells to confirm in vivo the inhibitory effect of miR-1224 on the binding of Sp1 to TNF-α promoter as previously described.^22,23 RAW264.7 cells which transfected with miR-1224 mimics or control were used for cross-linking with paraformaldehyde in PBS at a final concentration of 1% for 10 min. Cells were lysed in lysis buffer with protease inhibitor and sonicated three times for 20 seconds each time. Mouse anti-Sp1 antibody was added and the mixture was incubated overnight at 4°. Mock samples which contained 1 × lysis buffer instead of chromatin (no antibody) were also incubated, and the supernatant from the mock sample was taken as total input chromatin. Antibody/protein/DNA complexes were extracted three times with immunoprecipitation elution buffer, and cross-linking was reversed by incubation at 65° overnight. The mock sample was also processed in the same manner. Samples were then digested with DNase-free and RNase-free proteinase K at 50° for 4 hr, and DNA was extracted and purified using a Cycle Pure Kit (Omega, Norcross, GA, USA). Two primer pairs were designed for real-time PCR to analyse the TNF promoter region. One pair was located at about 100bp of 5’ flanking sequence and 3’ flanking sequence from Sp1 binding site, respectively. The other pair which was taken as negative control, which located at about 500bp of 5’ flanking sequence and 3’ flanking sequence from Sp1 binding site, respectively.

Results

miRNA expression profile within the spleen of LPS-treated and untreated mice

To identify miRNA involved in the regulation of the innate immune response, we have employed a commercial microarray chip that can examine global expression level for each known miRNA. Measuring the levels of mature miRNAs induced by LPS, makes it possible to determine a role for miRNAs in the LPS-induced immune response. This chip contains 627 mouse miRNAs and 39 mouse γ-herpesvirus miRNAs. Individual spleen samples from treated mice were analysed. Selection of differential expression miRNAs was based on the fold-change method with cut-off of 1·5. From microarray data, we found that 19 miRNAs were up-regulated in spleen upon LPS challenge. Among these miRNAs, miR-146b, miR-155 and miR-132 have previously been found to be up-regulated in response to LPS.¹⁸ The list of over-expressed miRNAs is shown in Table 1.

Table 1.

The list of differentially induced microRNAs (miRNAs) in spleens of mice intraperitoneally injected with lipopolysaccharide (LPS)

	Fold change
miRNA name	LPS-I versus saline	LPS-II versus saline
mmu-miR-532-3p	1·10	1·72
mmu-miR-146b	1·13	1·61
mmu-miR-34a	1·15	1·67
mmu-miR-21	1·48	1·89
mmu-miR-712	1·60	1·59
mmu-miR-223	1·67	1·10
mmu-miR-705	1·86	1·26
mmu-miR-706	2·11	0·69
mmu-miR-801	2·21	0·92
mmu-miR-709	2·34	1·76
mmu-miR-155	2·75	2·12
mmu-miR-188-5p	2·79	1·86
mmu-miR-671-5p	3·09	2·14
mmu-miR-689	3·95	1·75
mmu-miR-494	4·17	2·22
mmu-miR-877	5·25	3·22
mmu-miR-135a*	5·59	3·23
mmu-miR-1224	5·75	2·34
mmu-miR-721	6·50	2·93

Validations of microarray results using quantitative real-time PCR

To confirm microarray results, the same RNA samples from mouse spleens of both experiment and control groups were also analysed by stem-loop quantitative RT-PCR for three differentially expressed miRNAs: miR-146b, miR-155 and miR-1224. The relative amount of miRNA was normalized against U6 snRNA, and fold change for each miRNA was calculated by the 2-ΔΔCt method. The primers used for stem-loop RT-PCR are shown in Table 2. As shown in Fig. 1, the relative levels of miR-146b, miR-155 and miR-1224 were higher in spleens treated with LPS than in control, which is in agreement with the results of microarray hybridization. Macrophages play a central role in the innate immune response of the spleen, which led us to test the expression pattern of these three miRNAs in macrophages stimulated with LPS. One of the most important mediators in innate immune response is TNF-α, so we believe that TNF-α can be seen as a marker gene reflecting the degree of LPS response.

Table 2.

Primers sequence for PCR

Gene	Reverse transcription primer	PCR primer
miR-146b	CTCAACTGGTGTCGTGGAGTCGGCAATTCAGAGCCTATG;	F: ACACTCCAGCTGGGTGAGAACTGAATTC;
		R: TGGTGTCGTGGAGTCG;
miR-155	CTCAACTGGTGTCGTGGAGTCGGCAATTCAGTTGAGACCCCTATC;	F: ACACTCCAGCTGGGTTAATGCTAATCGTGAT;
		R: TGGTGTCGTGGAGTCG;
miR-1224	CTCAACTGGTGTCGTGGAGTCGGCAATTCAGTTGAGCTCCACCT;	F: ACACTCCAGCTGGGGTGAGGACTGGGG;
		R: TGGTGTCGTGGAGTCG;
U6	CTCAACTGGTGTCGTGGAGTCGGCAATTCAGTTGAGAAAAATATGG;	F: CTCGCTTCGGCAGCACA;
		R: AACGCTTCACGAATTTGCGT;
TNF-α	F: CGAAGTGGTGGTCTTGTTG; R: CGAGTCTGGGCAGGTCTA;
TLR4	F: GCTTTCACCTCTGCCTTCAC; R: AGGCGATACAATTCCACCTG;
GAPDH	F: GGTGAAGGTCGGTGTGAACG; R: CTCGCTCCTGGAAGATGGTG;
TNF-Α promoter	F1: CGGGGAGGAGATTCCTTGATG; R1: CTGGCTGGCTGTGCAGACG
	F2: ATTATTTATTTATTTGCTTATGAATG; R2: CAGTGATGTAGCGACAGCCTG

RT primer is stem-loop primer used for reverse transcription of a specific miRNA or U6.

GAPDH, glyceraldehyde 3-phosphate dehydrogenase; TLR, Toll-like receptor; TNF, tumour necrosis factor.

Figure 1.

Up-regulated microRNAs (miRNAs) obtained for microarray scanning were confirmed by real-time PCR. Total RNA extracted from spleen of mouse injected with saline or lipopolysaccharide (LPS) was analysed by using stem-loop real-time PCR. Relative expression levels of miRNAs were normalized to the level of U6 in each sample, and data are shown with miRNAs expression in control set as one.

Upon stimulation, the TNF-α protein was dramatically increased, reaching a peak 12 hr after addition of LPS (Fig. 2a), this change also happened at the mRNA level (Fig. 2c). As time went by, the strength of LPS stimulation became less, which is reflected in the changing ratio of concentration of TNF-α protein produced by LPS-treated and untreated RAW264.7 (Fig. 2b). In addition, the level of TLR4, membrane receptor of LPS, was peaked at 12 hr (Fig. 2d). Because of these, we selected three time-points for detecting changes of miRNA induced by LPS. Real-time PCR results showed that miR-146b, miR-155 and miR-1224 were up-regulated in response to LPS (Fig. 2e–g). All the results from real-time PCR were reconfirmed by semi-quantitative PCR (Fig. 2h).

Figure 2.

MicroRNAs (miRNAs) were up-regulated in RAW264.7 cells upon stimulation with lipopolysaccharide (LPS). (a) Tumour necrosis factor-α (TNF-α) protein secretion was dramatically increased and reached a peak at 12 hr after addition of LPS. RAW264.7 cells were cultured in medium and treated with LPS at a final concentration of 1 μg/ml. TNF-α protein levels in supernatant were measured by ELISA. (b) The concentration ratio of TNF-α protein secreted from LPS-treated RAW264.7 cells to untreated RAW264.7 cells. (c, d) Relative expression levels of TNF and TLR4 mRNAs in response to LPS. (e–g) Relative expression levels of miR-146b, miR-155 and miR-1224 in RAW264.7 cells treated with LPS. (h) Semi-quantitative PCR for miRNAs and mRNAs.

Tissue expression pattern of miRNAs

Understanding the distribution of a gene in different tissues is a fundamental question, and may provide some clue to reveal the function of unknown genes. We have analysed the expression patterns of miR-146b, miR-155 and miR-1224 in six tissues. Total RNAs were extracted from normal mouse heart, liver, spleen, lung, kidney and skeletal muscle with TRIzo reagent. The levels of miR-146b, miR-155 and miR-1224 in all tested tissues were detected using real-time PCR (Fig. 3). These analyses showed that miR-146b was predominantly expressed in lung, miR-155 and miR-1224 had high expression levels in spleen, both of these tissues are essential for host defence against infection by Gram-negative bacteria.

Figure 3.

Quantitative real-time PCR analysis of microRNA (miRNA) tissue expression pattern. Stem-loop reverse transcription-PCR primers sequences to mature miRNAs and U6 are shown in Table 1. The results were normalized against U6, and the relative expression is shown with miRNA expression in the heart set as one.

miR-1224 negatively regulates transcription of TNF gene

Tumour necrosis factor-α is one of the main cytokines involved in the response to LPS. The fact that miR-1224 was greatly increased in LPS-treated spleen suggested that this small uncoding RNA may play a role in the response to LPS. The examination of miR-1224 on TNF-α transcription was performed to check this hypothesis. We inserted the upstream sequence of mouse TNF-α transcript (− 991 to + 1 bp) into the pGL3 luciferase vector upstream of the luciferase. TNF-α promoter-driven reporter activity in miR-1224-transfected RAW264.7 cells was performed using a dual luciferase reporter assay system. The co-transfection of RAW264.7 cells with a pGL3-TNF-α-promoter construct, along with mmu-miR-1224 mimics, resulted in a decrease of the luciferase activity (about 60%) compared with the co-transfection with a control oligoribonucleotide (Fig. 4a). RAW264.7 cells were transfected with miR-1224 mimics and 18 hr later LPS (100 ng/ml) was added to the cells. Total RNAs were extracted to assess the differential expression of TNF. The differential expression of the other two pre-inflammatory cytokine genes Il1b and Il6 were also assessed. Results indicated that transfection of miR-1224 into RAW264.7 cells would lead to a decrease in LPS-induced expression of TNF, Il1b and Il6 (Fig. 4b, C). From these data, it can be seen that when cells were transected with miR-1224, TNF-α mRNA levels declined both with and without LPS. The result indicated that miR-1224 could not entirely inhibit LPS-induced production of TNF-α, but miR-1224 could weaken the production yield.

Figure 4.

miR-1224 decreases tumour necrosis factor-α (TNF-α) promoter activity. Luciferase reporters driven by TNF-α core promoter were co-transfected with pRL-TK and miRNA-1224 mimics into RAW264.7 cells. Firefly and Renilla luciferase activities were quantified using the dual-luciferase reporter assay system. The results are shown as the mean ± SD of three individual samples and the graphs are representative of three independent experiments.

miRNA-1224 mimcs increase basal NF-κB/p65 activity

Nuclear factor-κB is one of the most important transcription factors controlling expression of transcription of TNF-α mRNA. To test whether the negative effect of miR-1224 on TNF-α transcription is achieved by inhibiting NF-κB activity, we constructed the NF-κB/p65 reporter, containing NF-κB/p65 binding element firefly upstream of the luciferase coding sequence. Co-transfection of miR-1224 into RAW264.7 cells with the pGL3-NF-κB-reporter plasmid resulted in a sharp increase of luciferase activity (Fig. 5a), the increase could be abolished by the locked nucleic acid/DNA-based antisense inhibitor of miR-1224 (Fig. 5b). In addition, Western blot analysis showed that over-expression of miR-1224 decreased the NF-κB/p65 level in nuclear protein isolated from RAW264.7 cells. It is shown that miR-1224 has a positive role in the activation of NF-κB/p65 activity in RAW264.7 cells (Fig. 5c), which indicates that the negative effect of miR-1224 on TNF transcription does not modulate NF-κB. Both miR-146b and miR-155 are known to have opposite effects on innate immune responses.^17,24,25 As experimental control, the effect of miR-146b and miR-155 upon the activation of NF-κB/p65 was also tested. Luciferase reporter analysis showed that miR-146b and miR-155 had negative and positive effects on NF-κB/p65 activity, respectively.

Figure 5.

Effect of microRNA (miRNA) mimics and miRNA inhibitors on nuclear factor-κB (NF-κB) activity. Luciferase reporters driven by NF-κB response elements were co-transfected with pRL-TK and either mimics or inhibitors into RAW264.7 cells. Firefly and Renilla luciferase activities were quantified using the dual-luciferase reporter assay system. (a, b) Results demonstrate that miR-146b decreased the basal activity of NF-κB; miR-155 and miR-1224 increased the basal activity of NF-κB in RAW264.7 cells. (c) miR-1224 decreased NF-κB (p65) protein level in nucleus. Nucleoproteins were extracted from RAW264.7 cells transfected with miR-1224 mimics and miR-1224 inhibitor, respectively, and NF-κB (p65) protein levels were evaluated by Western blot analysis.

Bioinformatics analysis of putative miRNA targets

An miRNA-gene network related to LPS signalling (Fig. 6), including 10 over-expressed miRNAs (miR-21, miR-146b, miR-155, miR-188-5p, miR-223, miR-494, miR-671-5p, miR-706, miR-709 and miR-1224) and 22 target genes from MAPK and NF-κB signalling pathways, was obtained by applying the combination of TargetScan and DAVID databases. Although several miR-155 targets had been identified, none of them could directly affect TNF-α production. So the mechanism by which over-expressing of miR-155 both in vivo and in vitro can increase the production of TNF-α is still unclear. From the network it can be seen that miR-1224 directly targets the transcript encoding protein Sp1. The GC box-binding protein, Sp1, has been implicated in the transcription of a large number of genes.

Figure 6.

MicroRNA (miRNA) -gene regulatory network. Prediction of miRNA putative target genes was performed using TargetScan software in GeneSpring. Predicted target genes that associated with lipopolysaccharide (LPS) signalling target and miRNAs were integrated into a network map.

miRNA-1224 directly targets Sp1

RNAhybrid and TargetScan softwares were used to predict miRNA targets, and the prediction results showed that miR-1224 might target the 3′ UTR of both mouse and human Sp1 transcripts. Alignment of miR-1224 binding sequence in 3′ UTR of Sp1 mRNA from 13 other species showed that the binding site sequence was highly conserved from tenrec to human (Fig. 7a). We hypothesized that miR-1224 might bind and regulate Sp1 transcripts in vivo. To investigate this possibility, we inserted 59 bp of 3′ UTR of mouse Sp1 gene (Fig. 7b), which contains the putative miR-1224 interactive site, into psicheck2 report vector downstream of the luciferase coding sequence. At the same time, we also constructed a recombinant plasmid psicheck2-Sp1mut3UTR, which contained a mutated 3′ UTR of Sp1 upstream of the luciferase gene (Fig. 7c). Co-transfection of HEK-293 cells with this psicheck2-3′-UTR-Sp1 construct, along with miR-1224, resulted in a sharp decrease of luciferase activity compared with the co-transfection of a control. But when 3′ UTR was mutated, there was no decrease of luciferase activity (Fig. 7d). The facts that transfection of miR-1224 into HEK293 cells repressed the activity of the Sp1 3′ UTR-fused report gene but failed to affect the activity of mutated Sp1 3′ UTR-fused report gene supported the possibility that miR-1224, involved in the post-transcriptional repression of Sp1 mRNAs, further down-regulates the production of Sp1 protein. Western blot was performed to analyse Sp1 protein expression in RAW264.7 cells that were transfected with miR-1224 mimics or control. After 36 hr of transfection, LPS was added for 3 hr, nuclear protein was extracted and the effect upon Sp1 protein expression was determined using anti-Sp1 antibody. The immune blots were visualized with enhanced chemiluminescence. The results indicated that over-expressed miR-1224 decreased Sp1 protein levels compared with its level in RAW264.7 transfected with inhibitor (Fig. 7e). Although it could not completely inhibit LPS-induced Sp1 production, over-expression was able to reduce production. If the miR-1224-mediated inhibitory role on TNF expression acts by regulating Sp1, then over-expressed Sp1 in cells can rescue this inhibition. To test this hypothesis, we constructed three recombined plasmids based on pcDNA3.1 vector. The first was pSp1-wt3UTR, which contains 3′ UTR of Sp1 mRNA downstream of the Sp1 ORF, and the second one is pSp1-mut3UTR, which contains a 3′ UTR mutation of Sp1 mRNA downstream of the Sp1 ORF, the last one only contains pSp1 ORF. RAW264.7 cells were co-transfected with miR-1224 mimics and pSp1, pSp1-wt3UTR, pSp1-mut3UTR, respectively. Basal TNF mRNA level was measured by real-time PCR. Results showed that miR-1224 decreased the TNF-α mRNA level in RAW264.7, which was transfected with pcDNA3.1 empty vector; re-transfection of pSp1 or Sp1-mut3UTR into RAW264.7 cells could rescue the inhibitory effect of miR-1224 over-expression on TNF, but this effect was weakened when transfected with pSp1-wt3UTR instead (Fig. 7f). All of these results showed that the inhibitory effect of miR-1224 on TNF gene was mediated by Sp1. The conclusion that over-expression of miR-1224 led to a decrease of Sp1 in our study presents a novel post-transcriptional modification for regulation of Sp1 activity.

Figure 7.

Micro(mi)R-1224 silences Sp1 by repressing its translation. (a) Alignment of potential miR-1224 binding sites in 3′ untranslated region (UTR) of Sp1 mRNA of different species. (b) The target sequence for miR-1224 at Sp1 3′ UTR. A 59-bp fragment bearing miR-1224 target binding site was chemically synthesized and inserted into psicheck-2 vector. (c) Mutant of the target sequence for miR-1224 at Sp1 3′ UTR. * indicates mutation sites. (d) Verification of interaction between miR-1224 and the 3′ UTR of Sp1 in HEK293 cells, determined by luciferase reporter activity. Ctl, control without transfected with miR-1224 mimics; Mimics Ctl, mimics negative control. Error bars indicate SD. *P < 0·05 versus mimics Ctl. (e) Western blot analysis of Sp1 protein expression in RAW264.7 cells transfected with miR-1224 mimics or control. Inhibitor-1224, single stranded nucleic acids designed to specifically bind to and functionally inhibit miR-1224. (f) Inhibitory role of miR-1224 on tumour necrosis factor-α (TNF-α).can be rescued by Sp1. Vector, pcDNA3.1 empty vector. Error bars indicates SD 5% level of significance of difference marked with lowercase letters of a, b, c. Significant difference between the two averages indicated by different letters, and the average difference between the two was not significant where letters are the same.

Conformation of the inhibitory effect of miR-1224 on TNF-α by examining Sp1 binding in vivo

We performed a chromatin immunoprecipitation assay in RAW264.7 cells to confirm in vivo binding of Sp1 to TNF promoter. The DNA–Sp1 complex was immunoprecipitated with antibody (shown in Fig. 8), followed by a reversal of cross-linking and real-time PCR using primers flanking the Sp1 binding site in the TNF-α promoter region. As shown in Fig. 8, transfection of miR-1224 significantly suppressed the binding of Sp1 to chromatin in RAW264.7 cells, which further confirmed the conclusion that miR-1224 negatively regulated expression of TNF-α by modulating Sp1 protein.

Figure 8.

Conformation of inhibitory effect of micro(mi)R-1224 on tumour necrosis factor-α (TNF-α) promoter by examining Sp1 binding in vivo. DNA-Sp1 complexes from RAW264.7 cells transfected with mimics control or miR-1224 were immunoprecipitated with anti-Sp1 antibody. DNA purified from each sample was used for real-time PCR and semi-RT-PCR. (a) Real-time PCR analysis of Sp1 binding to in TNF-α promoter region. 1: Mimics control transfection input. 2: Mimics control transfection. 3: miR-1224 transfection input. 4: miR-1224 transfection. (b) Semi-RT-PCR analysis of Sp1 binding to in TNF-α promoter region. Lane 1: one-tenth of total input for sample transfected with mimics control; lane 2: one-tenth of total input for sample transfected with miR-1224 mimics; lanes 3 and 5: sample transfected with mimics-Ctl; lanes 4 and 6: sample transfected with miR-1224; lane 7: no template control; F1/R1: TNF-α promoter specific primers; F2/R2: negative primer, flanking far away from Sp1 binding site.

Discussion

The aim of this study is to identify miRNAs that might be involved in the regulation of the innate immune response and to assess their potential roles. The miRNA microarray analysis found 19 miRNAs that were up-regulated in spleens of mice treated with LPS. The same trends in mmu-mir-146b, mmu-mir-155 and mmu-mir-1224 expression levels were recorded when macrophage-like RAW264.7 cells were stimulated with LPS. Although it has been reported that miR-146a inhibits signalling proteins of innate immune responses, up-regulation of miR-146a was not recorded in our study; a difference that may be the result of using different research material.

It was thought that miR-146b was a new negative regulator of inflammation or innate immunity by targeting interleukin receptor-associated kinase 1 (IRAK1) and TNF receptor-associated factor 6 (TRAF6), two proteins involved in the TLR pathway.¹⁸ Interestingly, the inhibitory effect of miR-146a/b on IL-1β-induced IL-8 and RANTES release were not mediated through IRAK1 and TRAF6,²⁵ suggesting that the actions of miR-146b upon the inflammatory response were more than the result of negative regulation of the classical TLR pathway. miR-155 is a multifunctional gene, playing a crucial role in various physiological and pathological processes, such as haematopoietic lineage differentiation, immunity, inflammation and cancer. Moreover, miR-155 is also implicated in anti-viral responses.²⁶ Studies on miR-155 transgenetic/deficient mice suggest that this small RNA has a role in adaptive or innate immune responses.^27–29 Up-regulation of miR-155 induced by LPS enhances production of TNF-αin vitro and in vivo, which is consistent with the finding in our study that miR-155 could enhance the promoter activity of TNF-α. Promoter analysis revealed expression of miR-146a in monocytes or macrophages and miR-155 in Epstein–Barr virus-infected cells was activated by NF-κB and AP-1 trans-factor,^18,30 and inhibition of NF-κB activity using dexamethasone can attenuate the actions both miR-146a and miR-155. In a previous study, miR-155 attenuated NF-κB signalling in Epstein–Barr virus-positive B lymphocytes by targeting I-κB kinase ε (IKKε).³¹ The miR-155 transgenic mice had higher TNF-α production and died sooner when challenged with LPS, which implied that miRNA-155 had a positive effect on the innate immune response.²⁴ These results showed that miR-155 may exert opposite effects on the activation of NF-κB under different conditions. In contrast, miR-146b is a negative regulator of the LPS/NF-κB pathway by targeting IRAK1 and TRAF6 transcripts, which matches our observation that over-expression of miR-146b caused a dramatic decrease in NF-κB activity. Despite their important roles in immunity, except for studies of transfection of miR-155 mimic and inhibitor in Epstein–Barr virus-positive cell lines, which was shown to inhibit the alternative NF-κB pathway by targeting IKKε, there is little direct experimental evidence to evaluate the effect of miR-146b and miR-155 on the NF-κB pathway. miR-1224 is a novel miRNA that has been identified by extensive cloning.³² Although miR-1224 was up-regulated in both the Lieber–Decarli diet-induced alcoholic model and the methionine choline-deficient diet-induced non-alcoholic steatohepatitis model in mice, its potential functions are unknown.³³

Nuclear factor-κB is one of the final targets of the LPS/TLR4 signal transduction cascades, and is thought to be a primary effector of inflammatory responses.^34,35 The NF-κB family consist of five proteins: RelA (p65), RelB, c-Rel (REL), NF-κB1 (p50/p105) and NF-κB2 (p52/p100), and NF-κB signalling occurs through the classical pathway, which mainly applies to RelA(p65)-p50 dimers, or the alternative pathway mediated by the RelB-p52 complex.³⁶ The classical pathway is the main mechanism controlling innate immunity and inflammation.³⁷ Under non-stimulated conditions, NF-κB complexes are held in the cytoplasm by interaction with the inhibitory protein IκB. Following stimulation with LPS, the IκB are phosphorylated and dissociated from NF-κB, which results in translocation of NF-κB into the nucleus and initiates transcription of target genes.³⁸ Hence, once NF-κB is activated, the level of p65 is often increased in the nucleus. Nuclear factor-κB is one of the most important transcription factors mediated by LPS signalling, whose activation leads to the expression of many genes associated with the innate immune response. So, the control of innate immune response is achieved often through modulation of the nuclear transcription factor activity. To study the involvement of miR-1224 in the regulation of the LPS response, we tested the effect of over-expressed mir-1224 on NF-κB activity. Our results showed that miR-146b, miR-155 and miR-1224 have opposite effects on NF-κB activity: miR-146b decreases the basal activity of NF-κB but miR-155 and miR-1224 increase the basal activity of NF-κB in RAW264.7 cells. The fact that transfection of miR-1224 mimics enhancement of NF-κB activity showed that the negative effect of miR-1224 on the induction of TNF-α di not result from the down-regulation of the NF-κB pathway.

Tumour necrosis factor-α is one of the main cytokine in regulation of LPS-induced response. Expression of TNF gene is meditated by MAPK and NF-κB pathway via activation transcription factor and co-transcription factor AP-1, such as NF-κB, Sp1. Over-expression of miR-1224 in RAW264.7 cells led to a decrease in TNF gene transcription in vitro and in vivo. Although it could not completely inhibit LPS-induced Sp1 production, over-expression of miR-1224 was able to reduce the production. The results correspond with the observation that miR-1224 could not entirely inhibit LPS-induced production of TNF, but miR-1224 could weaken the production yield. Furthermore, miR-1224 could result in a decline in the other two pre-inflammatory genes, Il1b and Il6. We believe that the target gene of miR-1224 was a factor not only controlling transcription of the TNF-α gene, but also related to expression of Il1b and Il6. Based on all the above data, we hypothesize that the target of miR-1224 might be a transcription factor. With bioinformatics we speculated that the down-regulation of TNF gene transcription mediated by miR-1224 was achieved by the genes included in the LPS signalling pathway. Sp1 is one of the predicted miR-1224 target genes, which is confirmed by our results. It is documented that Sp1 takes part in the regulation of many cellular processes including the immune response. Interleukin-10 is a key anti-inflammatory factor produced by macrophages and other immune cells that plays important roles in immune homeostasis. In the human monocytic cell line THP-1 and alveolar macrophages (HAM), LPS-induced IL-10 production is mediated by the Sp1-dependent MAPK pathway.^39,40 In addition, Sp1 activates transcription of many pro-inflammatory factors. Interleukin-1β in keratinocytes is induced by Sp1 protein upon TNF-α stimulation.⁴¹ There is an Sp1 binding site in the 5′ UTR of TNF-α mRNA, and Sp1 binding triggers the transcription of TNF-α in human leucocytes in response to nitric oxide.⁴² Inhibition of Sp1 by injecting decoy oligodeoxynucleotides down-regulates TNF-α in mouse melanoma tumours.⁴³ The IL-6 promoter contains two Sp1-responsive elements. Sp1 was shown to bind to these sites in vitro and is required for basal and LPS-induced IL-6 promoter activity.⁴⁴ Upon infection, micro-organisms are first recognized by TLRs, which are expressed as either member-bound or soluble proteins in host immune cells. Existing studies have shown that the expression of four members of TLR family, TLR2, TLR4, TLR9 and TLR13, is regulated by Sp1 to various degrees.^45–48 Despite the fact that the transcription of many immune genes is required for activation of the Sp1 protein, the mechanism of down-regulation in Sp1 binding activity induced by LPS is still not clear.⁴⁹ Sp1 activity is regulated by a variety of stimuli, which usually occurs at the post-transcriptional level. Glycosylation and phosphorylation are two major types of post-translational modifications thought to be involved in regulation of Sp1 activity. Our study showed that miR-1224 has a negative effect on the transcription of TNF-α mRNA by targeting Sp1 transcription factor.

Conclusions

In summary, we have used an miRNA microarray-based approach to profile the changes in the mature miRNA expression patterns in the spleen of mice treated with LPS. Nineteen miRNAs were found to be up-regulated. Stem-loop real-time PCR was used to confirm the results of the miRNA microarray. miR-1224 was selected for further study and was shown to increase NF-κB activity in cells through an unknown mechanism. Further investigation into the regulatory role of miR-1224 in vitro and in vivo demonstrated that miR-1224 decreased TNF-α by targeting Sp1 protein, and provided a new type of post-translational modification for Sp1, so providing a novel insight into regulatory mechanisms of the innate immune response.

Disclosures

The authors declare no financial conflicts of interests.