miR-155 and miR-146a collectively regulate meningitic Escherichia coli infection-mediated neuroinflammatory responses

The meningitic E. coli strain PCN033, which is a highly virulent cerebrospinal fluid isolate isolated in China in 2006 [14], was cultured aerobically in Luria-Bertani (LB) broth or on LB plates at 37 °C. This strain was shown to be able to cause host blood–brain barrier (BBB) disruption and severe neuroinflammation in vitro and in vivo [6].

The human astrocyte cell line U251 (a kind gift from Prof. Shengbo Cao in Huazhong Agricultural University, Wuhan, China) and HEK-293T cells (ATCC® CRL-3216™) were cultured in Dulbecco’s modified Eagle’s medium containing 10% heat-inactivated fetal bovine serum in a 37 °C incubator with a 5% CO₂ atmosphere. U251 cells were cultured in 10-cm dishes until monolayer confluence. Confluent cells were washed three times with phosphate-buffered saline (PBS) and starved in serum-free medium for 16–18 h prior to infection. PCN033 overnight cultures were resuspended in serum-free medium and added to the starved U251 monolayer cells at a multiplicity of infection (MOI) of 10.

The anti-TAB2 antibody was purchased from Proteintech (Chicago, IL, USA). Anti-IRAK1 and anti-TRAF6 antibodies were from Sigma-Aldrich (St. Louis, MO, USA). Anti-β-actin antibody was from HuaAn Biotechnology Co., Ltd. (Hangzhou, China). Anti-NF-κB p65, anti-phospho-p65, HRP-conjugated anti-rabbit IgG, and anti-mouse IgG were obtained from Cell Signaling Technology (Danvers, MA, USA). The transfection reagent jetPRIME was purchased from Polyplus Transfection (Illkirch, France). has-miR-155 and has-miR-146a mimics (double-stranded RNA oligonucleotides), inhibitors (single-stranded chemically modified oligonucleotides), and control oligonucleotides were commercially synthesized by GenePharma (Suzhou, China). Cholesterol-conjugated and chemically modified mmu–miR-155 inhibitors (antagomir-155) and mmu–miR-146a inhibitors (antagomir-146a) were synthesized by GenePharma. The super electrochemiluminescence (ECL) kit was obtained from US Everbright Inc (Suzhou, China).

The 3’ UTR sequences of target genes were amplified from U251 cDNA and cloned into the psiCheck-2 dual-luciferase reporter vector (Promega Madison, WI, USA). The promoter sequences of miR-155 and miR-146a were cloned into pGL3-basic (Promega Madison). The site-specific mutation plasmids were constructed by overlapping extension PCR. To construct the transcription factor expression vector, the coding regions of p65 were amplified from U251 cDNA and cloned into pCDNA3.1. All constructs were verified by sequencing.

Small interfering RNAs (siRNAs) siRNA-TAB2, siRNA-IRAK1, siRNA-TRAF6, and the control siRNA were purchased from Gene Pharma. Transfection was performed with jetPRIME according to the manufacturer’s instructions. Cells were transfected with 50 nM of each siRNA.

U251 cells and HEK-293T cells were plated in six-well plates and grown to 70% confluence. Cells were subsequently transfected with plasmids and/or RNAs using jetPRIME according to the manufacturer’s instructions, followed by 24–48 h of incubation at 37 °C with 5% CO₂.

HEK-293T cells were co-transfected with 200 ng psiCheck-2 target genes 3’ UTR luciferase reporter plasmid or psiCheck-2 mutant target genes 3’ UTR luciferase reporter plasmid, along with miRNA mimics or control (final concentration, 50 nM). After 24 h, cells were collected and assayed for both firefly and Renilla luciferase activities using the Dual-Luciferase Reporter System following the manufacturer’s instructions (Promega, Madison, WI, USA). The results were calculated as the ratio of Renilla luciferase activity to firefly luciferase activity and recorded as the mean + SD from three replicate wells.

HEK-293T cells were co-transfected with 200 ng of full-length or mutant promoter firefly luciferase reporter constructs and 10 ng of Renilla luciferase vector (pRL-TK), along with the transcription factor expression vector or control. Luciferase activities were determined with the Dual-Luciferase Reporter Assay System according to the manufacturer’s protocol. The results were expressed as relative luciferase activity by normalizing firefly luciferase activity against Renilla luciferase activity and recorded as the mean + SD from three replicate wells.

The 28-day-old SPF KM mice were obtained from the experimental animal center at China Three Gorges University (Hubei Province, China). Mice were randomly assigned to four groups as follows: control (PBS); antagomir control-treated and E. coli-infected group (antagomir-ctrl + PCN033); antagomir-155-treated and E. coli-infected (antagomir-155 + PCN033); and antagomir-146a-treated and E. coli-infected (antagomir-146a + PCN033). For the antagomir, 60 mg/kg body weight of the antagomir was injected through the tail vein. After 24 h, mice were intravenously challenged with 1 × 10⁷ CFUs of E. coli strain PCN033 or an equal volume of PBS. At 6 h post-infection, mice were sacrificed and blood was collected for quantitative circulating bacterial cultures. Subsequently, mice were perfused, and brain samples were collected and processed for further assays. For the in vivo colonization assay, brains were weighed, homogenized, and plated to determine the bacterial counts.

Total RNA was extracted using TRIzol® Reagent according to the manufacturer’s protocol. Following RNA extraction, 1 μg of RNA was reverse transcribed into cDNA using the HiScript II Q RT SuperMix (Vazyme, Nanjing, China). Real-time PCR was performed with the MonAmp™ SYBR Green qPCR Mix (RN04005M, Monad Biotech Co., Ltd, Wuhan, China) in accordance with the manufacturer’s instructions. The PCR conditions included an initial step at 50 °C for 2 min and 95 °C for 10 min, followed by 40 cycles of amplification and quantification (95 °C for 15 s, 60 °C for 1 min). The expression of all mRNA targets was normalized to GAPDH or β-actin levels, and the miRNA expression was normalized to U6. Primers used for qPCR in this study are listed in Table 1. The relative expression level was calculated by the 2^−ΔΔCT method, and the results are presented as the mean + SD. Each assay was performed independently three times in triplicate.

Challenged U251 cells were lysed in RIPA buffer with a protease inhibitor cocktail (Sigma-Aldrich, USA), sonicated, and centrifuged at 12,000 rpm for 10 min at 4 °C. The protein concentration in the supernatant was measured using a BCA protein assay kit (Beyotime, China). Sodium dodecyl sulphate-polyacrylamide gel electrophoresis was performed, followed by protein transfer to polyvinylidene fluoride membranes using a Mini Trans-Blot Cell (Bio-Rad). Blots were probed with relevant antibodies, and the detection of proteins was conducted using ECL reagent. The densitometric analysis was performed using ImageJ software.

Brain tissue samples from challenged mice were lysed in RIPA buffer (supplemented with protease inhibitor) and centrifuged at 12,000 rpm for 10 min to eliminate tissue debris. The supernatant was stored at – 80 °C and later used for the measurement of preselected cytokines and chemokines, including IL-1β, IL-6, TNF-α, MCP-1, and MIP-2 using the ECL V-Plex Proinflammatory Panel (mouse) kit (Meso Scale Discovery, Meso Scale Diagnostics, Rockville, MD, USA), following the manufacturer’s instructions.

U251 cells were transfected with miR-155, miR-146a, or miR-ctrl, and 24 h after transfection, cells were challenged with PCN033 at an MOI of 10 for 3 h, and cells were washed and fixed with 4% paraformaldehyde. The fixed cells were subsequently incubated with primary anti-p65 antibody and then with CY3-labeled goat anti-mouse IgG antibody. The plate was mounted and visualized using fluorescence microscopy.

Brain samples were collected, fixed in 4% formaldehyde solution for over 24 h, and embedded in paraffin. Sections were incubated overnight at 4 °C with primary antibodies against GFAP (glial fibrillary acidic protein) or IBA1 (ionized calcium-binding adaptor molecule 1). After washing, sections were incubated with appropriate secondary antibodies. Immunostainings were examined with a fluorescence microscopy. GFAP and IBA1 image analyses were performed using the ImageJ software (NIH, Bethesda, MD, USA).

The brain samples were collected, fixed in 4% formaldehyde solution, and embedded in paraffin. Individual 6-μm sections were mounted on adhesive glass slides, dewaxed in xylene, and rehydrated in descending graded ethanol concentration for hematoxylin and eosin (H&E) staining according to the previous protocol [6].

Data were expressed as the mean + SD unless otherwise specified. Statistical significance of the differences between each group was analyzed by Student’s t test or one-way analysis of variance (ANOVA) embedded in GraphPad Prism, version 7.0 (GraphPad Software Inc., La Jolla, CA, USA); p < 0.05 was considered statistically significant, p < 0.01 and p < 0.001 indicated extremely significant differences.

Our previous miRNA transcriptional profiling of meningitic E. coli-infected astrocytes U251 showed that a number of miRNAs was differentially expressed upon E. coli infection. Among these miRNAs, miR-155 and miR-146a were found to be most significantly upregulated. Since miR-155 and miR-146a have been reported to be inflammation-associated molecules and E. coli infection can cause strong neuroinflammation, we therefore sought to explore their roles in meningitic E. coli-induced neuroinflammatory responses. First, we validated the expression of miR-155 and miR-146a in U251 cells stimulated by meningitic E. coli PCN033 for different times through real-time qPCR. The results revealed that miR-155 and miR-146a levels were significantly upregulated at 3 h, which was the later period of E. coli infection (Fig. 1a, b). Expression of the pri-miRNA MIR155HG and pri-mir-146a were in accordance with that of miR-155 and miR-146a (Fig. 1c, d).

It has been widely reported that NF-κB and MAPK signaling pathways can regulate numerous important biological processes, including the biogenesis of miRNAs [15]. Considering that miR-155 and miR-146a are inflammation-associated miRNAs, we investigated whether these signaling pathways played roles in the regulation of miR-155 and miR-146a expression. We pretreated U251 cells with a set of specific inhibitors, including NF-κB (BAY-11072), JNK (SP600125), ERK1/2 (U0126), and p38 (SB202190), and then infected them with PCN033 for 3 h. As shown in Figs. 2a and b, the E. coli infection-induced upregulation of miR-155 and miR-146a was completely prevented by BAY-11072, but SP600125 and U0126 had no inhibitory effect on the expression of these miRNAs. SB202190 only had a limited inhibitory effect on the expression of miR-146a but had no effect on miR-155 levels. In addition, we noticed above that the pri-miRNA MIR155HG and pri-mir-146a exhibited the synchronous changes with the mature miR-155 and miR-146a (Fig. 1), suggesting that the upregulation of miR-155 and miR-146a might occur at the transcriptional level. To evaluate the transcriptional regulation of miR-155, we analyzed the promoter region of miR-155 to predict the potential transcription factor binding sites by using PROMO [16] (http://​alggen.​lsi.​upc.​es/​cgi-bin/​promo_​v3/​promo/​promoinit.​cgi?​dirDB=​TF_​8.​3). Combined with the results showing that miR-155 was modulated by NF-κB signaling pathways, we found an NF-κB binding site with high probability scores (Fig. 2c). As shown in Fig. 2d, the overexpression of p65 resulted in a concentration-dependent elevation of luciferase activity from the miR-155 promoter, but mutation of the NF-κB binding site significantly decreased miR-155 promoter activity (Fig. 2e). Similarly, the promoter region of miR-146a was analyzed and two NF-κB sites were found (Fig. 2f). It was observed from Figs. 2g and h that the overexpression of p65 enhanced the miR-146a promoter activity in a dose-dependent manner and the effect was abrogated by a mutation disturbing the two NF-κB element binding sites. Taken together, these results indicate that NF-κB signaling plays critical roles in the regulation of miR-155 and miR-146a by directly binding their promoters.

To investigate the potential role of miR-155 and miR-146a in the response to E. coli infection, we assessed their effects on E. coli-induced inflammatory gene expression in astrocytes. Using the chemically synthesized mimics or inhibitors, we succeeded in up or downregulating the levels of miR-155 and miR-146a (Fig. 3a, b). We found that transfecting mimics and inhibitors of miR-155 and miR-146a had no influence on bacterial loads (Fig. 3c). Results showed that expression of the inflammatory cytokines IL-1β, IL-6, and TNF-α, as well as chemokines MCP-1 and MIP-2, were decreased after miR-155 or miR-146a mimics treatment (Fig. 3d). In contrast, the inhibitor of miR-155 or miR-146a increased these cytokines and chemokines (Fig. 3e). NF-κB-mediated pro-inflammatory gene expression plays an important role in the innate immune response against bacterial infection. E. coli-infected astrocytes exhibited the time-dependent upregulation of NF-κB p65 subunit phosphorylation, whereas p-p65 was decreased at 3 h, which was the timepoint that miR-155 and miR-146a were significantly increased (Fig. 4a). To determine whether miR-155 and miR-146a affect the NF-κB signaling pathway, we investigated their effects on phosphorylation of the p65 subunit in U251 cells. As shown in Figs. 4b and c, the infection-caused phosphorylation p65 in astrocytes transfected with miR-155 or miR-146a mimics was significantly decreased compared with that in cells transfected with miR-ctrl. Moreover, transfection of the miR-155 or miR-146a inhibitor promoted the phosphorylation p65. In addition, using immunofluorescence microscopy, we observed p65 translocation from the cytoplasm to the nucleus upon E. coli infection, which could be partly prevented by the transfection of miR-155 or miR-146a (Fig. 4d). These results showed that miR-155 and miR-146a play an inhibitory role in E. coli-induced pro-inflammatory factor production by negatively modulating NF-κB signaling.

Generally, miRNAs exert their regulatory functions by modulating translation of their target genes by binding the 3’ UTR of mRNA. Considering the results indicating that miR-155 can inhibit NF-κB activity, we first searched potential targets that are associated with inflammation. MiRNA target prediction software TargetScan showed that TAB2 had potential seed matches for miR-155. TAB2 is an adaptor molecule of the TLR/IL-1 signaling cascade that facilitates IL-1-dependent TNF receptor-associated factor 6 (TRAF6) ubiquitination and assembles the IL-1 signaling complex; thus, we further investigated the relationship between miR-155 and TAB2. A luciferase reporter carrying a putative binding site of the TAB2 3’ UTR along with a mutant construct with 7-bp mutations in the seed region were generated (Fig. 5a), and the results shown in Fig. 5b indicated that the transfection of miR-155 significantly inhibited the luciferase activity of the wild-type TAB2 3’ UTR construct but had no effect on the luciferase activity from the mutant construct. To further authenticate the negative correlation between miR-155 and TAB2, the expression of TAB2 was measured in U251 cells transfected with miR-155 mimics or inhibitors. As shown in Figs. 5c and d, the overexpression of miR-155 significantly suppressed both mRNA and protein levels of TAB2, whereas the miR-155 inhibitor led to upregulated TAB2 mRNA and protein levels. In addition, the mRNA and protein expression of TAB2 were significantly downregulated in U251 cells upon E. coli infection (Fig. 5e, f), which was the opposite trend compared with that of miR-155. Moreover, we measured the protein level of TAB2 in U251 cells transfected with anti-miR-155 after E. coli infection. As shown in Fig. 5g, the E. coli infection-induced downregulation of TAB2 was restored by inhibiting miR-155, demonstrating that miR-155 can regulate TAB2 during E. coli infection. To further confirm that miR-155 exerts its effect through TAB2, we subsequently knocked down TAB2 in U251 cells using siRNA. As seen in Fig. 5h, TAB2 was successfully inhibited. As expected, the transfection of TAB2 siRNA resulted in decreased phosphorylation of NF-κB p65 upon E. coli infection compared with that with the transfection of siRNA control (Fig. 5i). Importantly, the augmented expression of inflammatory genes (IL-1β, IL-6, TNF-α, MCP-1, and MIP-2) induced by the inhibition of miR-155 was decreased by suppressing TAB2 (Fig. 5j). Collectively, these results demonstrate that TAB2 is the functional target of miR-155.

IRAK1 and TRAF6 have been reported to be important targets of miR-146a in monocytes and macrophages during the innate immune response [12, 17]. Since miR-146a was experimentally proven to be associated with meningitic E. coli infection, whether it also targeted IRAK1 and TRAF6 in astrocytes needed further investigation. Bioinformatic analysis showed that there are two binding sites in the IRAK1 3’ UTR and three binding sites in the TRAF6 3’ UTR (Fig. 6a). We found a significant decrease in luciferase activity in HEK-293T cells co-transfected with miR-146a mimics and the wild-type IRAK1 3’ UTR luciferase reporter plasmid, and miR-146a mimics still significantly downregulated the relative luciferase activity from the IRAK1-3’ UTR-MUT2 construct but had mild effects on the IRAK1-3’ UTR-MUT1 construct (Fig. 6b). This result showed that binding site 1 in the 3’ UTR of IRAK1 is a vital targeting site of miR-146a. Similarly, we constructed a wild-type TRAF6 3’ UTR luciferase reporter plasmid with a single site mutation, a double site mutation, and a triple site mutation. As shown in Fig. 6c, the overexpression of miR-146a markedly inhibited luciferase activity from the wild-type TRAF6 3’ UTR luciferase reporter plasmid, whereas the suppression of luciferase activity was restored gradually in the presence of site mutations, revealing that all three binding sites play important roles in combination with miR-146a. In addition, overexpressing miR-146a significantly suppressed IRAK1 and TRAF6 expression in U251 cells, both at the mRNA and protein levels, whereas miR-146a inhibition led to the increased expression of IRAK1 and TRAF6 (Fig. 6d, e). We also measured the expression of IRAK1 and TRAF6 along with the E. coli infection and found that they were decreased in U251 cells upon infection (Fig. 6f–h). Moreover, miR-146a was determined to regulate the expression of IRAK1 and TRAF6 during E. coli infection (Fig. 6i). To further verify the function of miR-146a targets, we subsequently knocked down the expression of IRAK1 and TRAF6 at the mRNA and the protein levels (Fig. 6j, k). As shown in Fig. 6l, the inhibition of both IRAK1 and TRAF6 reduced the phosphorylation of p65 induced by PCN033, as compared with that with their controls. Furthermore, inhibiting miR-146a promoted the expression of PCN033-induced inflammatory genes, including IL-1β, IL-6, TNF-α, MCP-1, and MIP-2, and these genes were notably reduced by interfering with IRAK1 or TRAF6 (Fig. 6m). Taken together, miR-146a exerts an inhibitory effect on inflammation by targeting IRAK1 and TRAF6. miR-155 and miR-146a targeted different molecules in the signaling cascade that downstream of TLR, including TAB2, IRAK1, and TRAF6, thereby collectively modulating the TLR-mediated inflammatory response.

When we used bioinformatics tools to predict the targets of miR-155 and miR-146a, interestingly, we found that EGFR possibly combines with both miR-155 and miR-146a. Moreover, our previous studies showed that EGFR contributes to bacterial-induced neuroinflammation by triggering the MAPK-ERK1/2 and NF-κB signaling pathways [18]. Thus, we investigated the regulative relationship between EGFR and miRNAs. Bioinformatics prediction showed that miR-155 and miR-146a contained respective binding sites for EFGR (Fig. 7a, c), and the overexpression of miR-155 or miR-146a inhibited luciferase expression from the wild-type EGFR 3’ UTR. In contrast, the activity of luciferase constructs containing the mutant 3’ UTR of EGFR was not inhibited by overexpressing miR-155 or miR-146a (Fig. 7b, d). In addition, miR-155 and miR-146a significantly decreased mRNA and protein levels of EGFR, whereas the inhibition of miR-155 or miR-146a promoted the expression of EGFR (Fig. 7e, f). Besides, U251 cells exhibited the significant downregulation of EGFR in response to meningitic E. coli infection (Fig. 7g, h). All these results suggested that EGFR is a potential target of miR-155 and miR-146a. To further investigate whether EGFR participates in the inflammatory process in response to E. coli infection, we pretreated U251 cells with 5, 10, or 20 μM of the EGFR inhibitor AG1478 or DMSO as a control, and explored whether NF-κB signaling pathways were involved in the EGFR-mediated inflammatory response. Results showed that the p65 phosphorylation induced by E. coli infection was notably attenuated by AG1478 in a dose-dependent manner (Fig. 7i), and correspondingly, as shown in Fig. 7 j, E. coli-induced IL-1β, IL-6, TNF-α, MCP-1, and MIP-2 expression was significantly suppressed by treatment with AG1478 in a dose-dependent manner. Taken together, these results indicated that EGFR is a target of miR-155 and miR-146a and that blocking EGFR activity could markedly alleviate downstream NF-κB signaling pathways, thus suppressing the production of pro-inflammatory cytokines and chemokines.

We further determined the effects of miR-155 and miR-146a in E. coli-infected mice using chemically modified antisense oligonucleotides specific for miR-155 (antagomir-155) and miR-146a (antagomir-146a), which can cross the BBB into the CNS via i.v. routes [19‐21]. Antagomir-155 and antagomir-146a were delivered to mice via the tail vein 24 h prior to PCN033 infection, and mice showed no signs of discomfort. At 6 h post-challenge, mice exhibited typical neurological signs including trembling, circling, paddling, and opisthotonos [6], at which time mice were sacrificed and brain tissues were collected. As shown in Figs. 8a and b, the injection of antagomir-155 significantly downregulated the expression of miR-155 induced by E. coli infection in mouse brains, and antagomir-146a also suppressed the level of miR-146a. Decreasing the expression of miR-155 and miR-146a did not affect the bacterial loads in blood and brains of challenged mice (Fig. 8c). Compared with levels in mice injected with antagomir-ctrl, inflammatory cytokines, and chemokines including IL-1β, IL-6, TNF-α, MCP-1, and MIP-2 were further enhanced at the transcription and protein levels by antagomir-155 and antagomir-146a treatment (Fig. 8d, e). These results indicated that miR-155 and miR-146a altered the inflammatory responses in respond to E. coli infection, which is not relevant to bacterial growth or survival. In addition, H&E staining results showed that E. coli infection could induce meningeal thickening and neutrophil infiltration in the meninges, and these pathological phenomena were further exacerbated by treatment with antagomir-155 and antagomir-146a (Fig. 9a). Moreover, E. coli infection induced astrocytosis and microgliosis in mice, and these pathological phenomena were further exacerbated by antagomir-155 and antagomir-146a treatment (Fig. 9b–d). Based on these observations, we evaluated the effects of mir-155 and mir-146a on mouse lethality. As shown in Fig. 9e, the administration of antagomir-155 and antagomir-146a caused no significant changes in mouse mortality; however, the survival time was significantly decreased. In conclusion, inhibiting miR-155 and miR-146a in mice can further aggravate pro-inflammatory responses and lead to the reduced survival duration.