Postnatal 3–4-day-old suckling BALB/C mice were infected with P3 strain of JEV. Upon onset of clinical manifestation of JEV such as poor brain response, limb paralysis, and whole body tremor, mice were killed and their brains were excised. Brain homogenate prepared in Dulbecco’s modified Eagle’s medium (DMEM) was centrifuged at 10,000g to remove cellular debris. The resultant suspension was filtered through 0.22 µm sterile filters to obtain viral suspension. Immediately, aliquots of filtered virus suspension were stored at −80°C until further use. The virus titer was determined by plaque formation assay on baby hamster Syrian kidney (BHK-21) cells as described previously (10).

Adult BALB/C mice (6 weeks old) were purchased from Hubei Provincial Center for Disease Control and Prevention, Wuhan, China. Mice were randomly assigned to two groups (n = 3 for each group): the JEV-infected group and the control group. For the infection group, mice were inoculated intracranially with 200 plaque-forming units (PFU) JEV (P3 strain) in 20 µl DMEM, whereas mice belonging to the control group were injected intracranially with equal volume of DMEM. On day 5 postinfection, mice infected with JEV showed signs of acute encephalitis. Mice from both groups were euthanized, and brain samples were collected for further studies. All animal experiments were performed following the National Institute of Health Guide for the Care and Use of Laboratory Animals, and the experimental protocols were approved by the Research Ethics Committee of College of Veterinary Medicine, Huazhong Agricultural University, Hubei, Wuhan, China (No. S02914040M). All of our virus experiments are under Biosafety Level 2 containment. We performed double-blind procedure in our in vivo experiments, blinding investigators, participants, and outcome assessors.

The RAW264.7 cells were purchased from the American Type Culture Collection, and the BV2 cells were kindly provided by Professor Yuanan Lu from University of Hawaii, Manoa, USA (27). The cells passage numbers were below 20. BV2 and RAW264.7 cells were cultured and maintained in DMEM that was supplemented with 10% (v/v) heat-inactivated fetal bovine serum, 100 U/ml penicillin, and 100 mg/ml streptomycin sulfate at 37°C in 5% CO₂ atmosphere. Cells were seeded in multiwell plates, and grown to 80% confluency. Non-adherent cells were removed by washing with non-supplemented DMEM prior to further treatment. All of our virus experiments are under Biosafety Level 2 containment.

The cells were mock-infected or infected with JEV P3 strain at multiplicity of infection (MOI) of 5 for 2 h. At 12, 24, and 36 h postinfection, cells were fixed and blocked with 10% bovine serum albumin in phosphate-buffered saline (PBS, pH 7.2) for 30 min. Then, cells were incubated with monoclonal antibody recognizing JEV NS5 (5 ng/ml for immunofluorescence assay, prepared by our laboratory) for 1 h. After washing three times with PBS, cells were incubated with Alexa Fluor 488-conjugated secondary antibody (Invitrogen) for 30 min. Cells nuclei were stained with 4′,6-diamidino-2-phenylindole dihydrochloride (Invitrogen). The staining was observed using a fluorescence microscope (Zeiss) with 20× magnification.

Small-interfering RNAs (siRNAs) corresponding to the sequence of lncRNA-E52329 and lncRNA-N54010, which were used to inhibit endogenous expression of lncRNA-E52329 and lncRNA-N54010, and the negative control siRNAs, which exhibited no downregulation of any mouse genes, were synthesized by Gene pharma. Transfection was performed with Lipofectamine 2000 (Invitrogen). Cells were transfected with 50 nM of each siRNA. The sequence of all siRNAs used in this study is listed in Table S6 in Supplementary Material.

Total RNA was extracted from mouse brain or treated BV2 or RAW264.7 cells using TRIzol reagent (Invitrogen), and subsequently, was reverse transcribed into cDNA using First Strand cDNA Synthesis Kit (TOYOBO) following the manufacturer’s instructions. Quantitative real-time PCR was performed using a ViiA™ 7 Real-time PCR System (Applied Biosystems) and SYBR Green Real-time PCR Master Mix (TOYOBO). Amplification was performed for 2 min at 50°C and 10 min at 95°C, followed by 40 cycles of 95°C for 15 s, 60°C for 15 s, and 72°C for 30 s. The relative expression levels of lncRNAs, TNF-α, IL-6, IL-1β, and IFN-β were normalized to that of β-actin within each sample using the 2^−ΔΔCt method. The specific primers for lncRNAs, TNF-α, IL-6, IL-1β, IFN-β, and β-actin are listed in Table S6 in Supplementary Material.

Total RNA was extracted and purified using miRNeasy Mini Kit and RNase-Free Dnase Set. Biotinylated cDNA was prepared according to the standard Affymetrix protocol from 250 ng total RNA by using GeneChip^® WT PLUS Reagent Kit. The resultant cDNA was then transcribed to generate cRNA, which is reverse transcripted to yield second-cycle cDNA, and then fragmented. Following labeling, 5.5 μg of cDNA was hybridized for 16 h at 45°C on GeneChip Mouse EG 1.0 ST Array in GeneChip^® Hybridization Oven 645. GeneChips were washed and stained in the Affymetrix Fluidics Station 450. The array was scanned by GeneChip2 Scanner 3000 7G. The probe array was scanned by Affymetrix^® GeneChip Command Console. The gene expression levels were normalized by Expression Console. Expression data were generated by Affymetrix Expression Console software and normalized by the Robust Multi-chip Average (RMA) method. RVM t-test performed by BRB-Array Tools (v4.3.2) was applied to filter the differentially expressed genes for the control and experimental group. After significant analysis and false discovery rate (FDR) analysis, we selected the differentially expressed genes according to the P-value and fold change threshold (absolute ratio). P-value <0.05 and fold change >2 for mRNA or fold change >3 for lncRNA, were considered as significant difference. The lncRNAs and mRNAs were annotated through the Affymetrix Power Tools from the database of Refseq, Ensembl, and Non-code. The microarray data are available on NCBI GEO database, and GEO accession number is GSE94789.

The statistical heat map of the two groups was constructed using software Cluster (Version 3.0) to show the difference of gene fold changes between both groups.

To gain insight into the functions of differentially expressed genes, associated gene ontology (GO) terms were identified using DAVID Bioinformatics Resources (version 6.7, http://david.abcc.ncifcrf.gov/). It can organize genes into hierarchical categories and uncovers the gene regulatory networks on the basis of biological process and molecular function (28, 29). Specifically, we used two-side Fisher’s exact test and χ² test to classify the GO category, and the FDR (30) was calculated to correct the P-value. Smaller the value of the FDR, smaller will be the error in judging the P-value. We computed P-values for the GOs of all differentially regulated genes. Enrichment provides a measure of significance of the function: as the enrichment increases, the corresponding function is more specific, which helps us to find those GOs with more concrete function description in the experiment.

We also used DAVID software to analyze the signaling pathways of differentially expressed genes according to KEGG, Biocarta, and Reatome. The significance was determined using the Fisher’s exact test and χ² test, and the threshold of significance was defined by P-value and FDR (31–33).

The coexpression networks were built according to normalized signal intensity of specific gene expressions. For each pair of genes, the Pearson correlation was calculated, and the significant correlation pairs were chosen to construct the network (34).

The gene coexpression network was constructed in experimental group and control group, respectively. Within the network analysis, a degree is the simplest and most important measure of centrality of a gene within a network and determines the relative importance. A degree is defined as the number of directly linked neighbors.

Degree in experimental group was recorded as exp_degree, whereas in control group was recorded as con_degree. A clustering coefficient is a measure of the degree to which nodes in a graph tend to cluster together. It was calculated by the local measure (35). To exclude other genes’ impact in each coexpression network, we further performed normalization of the degree, i.e., divided by maximum value of the gene degree in each network [Normalized degree(i) = Degree(i)/Degree(Max)]. Then, the difference value of a gene’s normalized degree (delta normalized degree, represented as |diffK|) was calculated between the two coexpression networks.

While considering different networks, core regulatory factors were determined based on degree differences between the two class samples (36). Considering our pathway analyses, we chose some genes of our interest, and functions of most of the selected genes were related to inflammation. The cytoscape software (version 5.3) was used to build the interaction map.

The lncRNA E52329 and N54010 were amplified by polymerase chain reaction. The lncRNAs were transcribed (T7 MEGAscript™, Ambion) and biotinylated (Pierce™ RNA 3′ End Desthiobiotinylation, Thermo Scientific) according to the manufacturer’s protocol. The JEV-infected BV2 cells were lysed in radioimmunoprecipitation assay buffer (Sigma) containing protease and phosphatase inhibitors (Roche). For pull-down experiments, biotin-labeled RNA or antisense-RNA were incubated with streptavidin magnetic beads on a rotator at room temperature for 30 min in the cellular extract solutions, then washed with PBS for five times, and the resultant pellet was extracted RNA for qRT-PCR (37).

Total mice brain tissues lysates were generated using radioimmunoprecipitation assay buffer (Sigma) containing protease and phosphatase inhibitors (Roche). Protein concentrations were measured with a BCA Protein Assay Kit (Thermo Scientific). Equal protein quantities were separated by SDS-PAGE and transferred to a polyvinylidene fluoride membrane (Millipore) using a Mini Trans-Blot Cell (Bio-Rad). Blots were probed with the relevant antibodies, and proteins were detected using the ECL reagent (Thermo Scientific). Mouse monoclonal antibodies against JEV NS5 were generated in our laboratory. Commercially obtained antibodies used were: mouse monoclonal antibody against GAPDH, rabbit polyclonal antibodies against β-tubulin, MKK4, IκBα, phosphor-MKK4-Thr261 (ABclonal Technology), phosphor-IκBα-Ser32, phosphor-SAPK/JNK-Thr183/Tyr185 (CST technology), and horseradish peroxidase-conjugated antimouse secondary antibodies (Boster).

BV2 and RAW264.7 cells were transfected with siE52329, siN54010, or their controls (final concentration, 50 nM) for 24 h, and subsequently infected with JEV at MOI of 5. At 12, 24, and 36 h postinfection, cell supernatants were harvested, serially diluted, and then used to inoculate monolayers of BHK-21 cells. After removal of unbound JEV virus particles, BHK-21 cells were further incubated for 3–5 days, and plaques identified. The visible plaques were counted and viral titers calculated. All data are expressed as the mean of triplicate samples. All of our virus experiments are under Biosafety Level 2 containment.

The culture supernatants were collected from treated cells at the indicated time points and stored at −80°C. The protein levels of TNF-α, IL-1β, and IL-6 in cell cultures or mouse brain tissue lysates were measured by ELISA kits (eBioscience) following the manufacturer’s instructions.

Mice were anesthetized with ketamine-xylazine (0.1 ml/10 g of body weight) and perfused with PBS, followed by 4% paraformaldehyde. Brain tissues were removed and embedded in paraffin for coronal sections. The sections were used for H&E staining, IHC, and the TUNEL assays as described previously (9). For IHC experiment, tissue sections were incubated overnight at 4°C with primary antibodies against ionized calcium-binding adapter molecule 1 (IBA-1) (Wako), glial fibrillary acidic protein (GFAP) (Dako), and neuron-specific nuclear protein NeuN (Chemi-Con) at concentrations indicated in the manufacturer’s guidelines. After washing, slides were incubated with antimouse horseradish peroxidase-conjugated secondary antibodies, washed, and 3,3′-diaminobenzidine (Vector Laboratories) was used for color development. For TUNEL assay, an In Situ Cell Death Detection Kit (Roche) was used according to the manufacturer’s instructions.

All results are expressed as mean ± SEM. Statistical analysis was performed by GraphPad Prism 5 (GraphPad Software, San Diego, CA, USA). Differences were analyzed for statistical significance using two-sided unpaired t test for two groups or multiple comparison one-way of variance (ANOVA) for more than two groups (Bonferroni’s multiple comparison test, P < 0.05).

To start our analysis, we first established a mouse model for JEV infection and then collected samples for the microarray analysis (Figure S1 in Supplementary Material). To confirm whether the mice have successfully acquired the JEV infection, we examined the virus replication in mice brain samples using plaque assay and Western blot analysis. The results showed that the infected mice had a high virus titer in the brain tissues (Figure 1A), and these tissues exhibited an increased expression of JEV NS5 (Figure 1B). To visualize the effect of JEV infection in mice brain, we examined the brain tissue sections processed for H&E staining and IHC analysis. The data demonstrated that JEV-infected mice presented the emblematic histopathological features of encephalitis, whereas no pathological changes were noticed in the mice belonging to control group (Figure 1C). Furthermore, an aberrant increase in number of astrocytes and microglia was detected by IHC using anti-GFAP and -IBA-1 antibody, respectively (Figure 1D). Taken together, these findings indicate the establishment of a productive JEV infection in the mice brain.

To identify changes in the expression levels of lncRNAs and mRNAs in mice brain infected with JEV, total RNA was extracted from JEV-infected or mock-infected mice brain. The array which covers 30,772 lncRNAs and 16,417 protein-coding transcripts from data sources such as Ensembl, fRNAdb, lncRNAdb, NOCODE4.0, and UCSC known gene, was used to detect lncRNAs and mRNAs in the mice brain samples (Table S1 in Supplementary Material). Scatter plot was used to show variation in lncRNAs and mRNAs expressions between JEV-infected and mock-infected mice brain samples (Figures 2A,B). RVM t-test revealed that 518 lncRNAs and 1,007 mRNAs were differentially expressed in a significant manner between the two groups. Among the differentially expressed lncRNAs and mRNAs, 116 lncRNAs and 795 mRNAs were found to be upregulated, whereas 402 lncRNAs and 212 mRNAs exhibited a downregulated pattern (P < 0.05; fold change >2 for mRNA or fold change >3 for lncRNA, Table 1; Table S2 in Supplementary Material). The hierarchical clustering analysis was also employed to show the differential expressions of lncRNAs and mRNAs between both groups of samples (Figures 2C,D).

To better understand the functions of differentially regulated mRNAs and lncRNAs in JEV-induced inflammation, GO term analysis of these mRNAs and lncRNAs was performed. We found that highly enriched GO categories for mRNAs included, but were not confined to, phosphoproteins, nuclear functions, signal transduction, acetylation, and nucleotide binding (Figure S2A in Supplementary Material). The most significant GOs associated with upregulated lncRNAs were related to innate immune response, apoptotic process, and inflammatory response (Figure S2C in Supplementary Material). The top 10 significantly upregulated and downregulated GO terms linked with differentially regulated mRNAs and lncRNAs are presented in Figures S2A–D and Tables S3 and S4 in Supplementary Material.

Pathways analysis of differentially expressed 1,007 mRNAs and 518 lncRNAs was performed using the KEGG (Kyoto Encyclopedia of Genes and Genomes). Several important signaling pathways, i.e., those of Toll-like receptor, Jak-STAT, MAPK, chemokines, Herpes simplex and influenza A infection, and PI3K-Akt, were markedly activated by differentially expressed mRNAs and lncRNAs (Figures S2A,C in Supplementary Material). The top 10 upregulated and downregulated pathways associated with these mRNAs and lncRNAs are shown in Figures S2A–D and Tables S3 and S4 in Supplementary Material.

To investigate the correlation of lncRNAs and mRNAs, we constructed the lncRNA-mRNA coexpression network in each of the two separate groups (Table S5 in Supplementary Material). We used biostatistics to calculate the possibility of interaction between these genes. The solid line between the two nodes strands for positive correlation, whereas the dotted line indicates negative correlation (Figure S3 in Supplementary Material). The clustering coefficient represents the density of the gene in the network. Higher value of the clustering coefficient indicates more dependent interaction with other genes in the network or vice versa. The genes with higher degree and higher clustering coefficient in both groups are listed in Table 2.

In order to interrogate interactions between the genes related to inflammation, we selected some of the lncRNA genes that may participate in the inflammatory pathways, and then built the lncRNA-mRNA and lncRNA-lncRNA coexpression networks (Figures 3A,B). To confirm our coexpression network analysis, we chose lncRNA E52329 and N54010 and performed RNA pull down experiments. Our results showed that Nod1, Tap2, and Col4a1 were pulled down by indicated bio-lncRNAs compared with bio-anti-lncRNAs (Figures 3C,D). In order to determine that which pathway plays an important role in our lncRNA network, we used pathway studio (version 10) and KEGG pathway to predict the possibility of pathway enrichment. The findings showed that MKK4/7 is the most enriched pathway.

An altered gene status in the coexpression network between JEV-infected and mock-infected groups suggest that the gene may perform a very important function during the JEV infection. Therefore, we compared the status of genes in both groups, and the genes exhibiting the most significant difference are enlisted in Table 2.

The differential expression levels of lncRNAs found in our microarray analysis were validated by quantitative real-time PCR. From the list of differentially regulated lncRNAs, we randomly selected five upregulated and five downregulated lncRNAs for their expression validation in mice brain tissues. Our results obtained from quantitative real-time PCR were analogous to those observed in microarray data (Figures 4A,B). We also detected the expression levels of lncRNAs in JEV-infected BV2 cells. Most of the lncRNAs results showed the same tendency as were observed in mice brain samples after JEV infection (data not shown). However, some of the lncRNAs were not detected in BV2 cells, which may happen due to tissue or cell specificity (38).

To discern a possible role of differentially regulated lncRNAs in JEV-induced inflammation, we selected two lncRNAs (lncRNA E52329, lncRNA N54010) form the data validated using mice brain samples. Since JEV infection is associated with marked activation of glial cells and exorbitant release of inflammatory cytokines (39), we chose mouse microglial cells (BV2 cells) for further analysis. To determine whether BV2 cells were permissive to JEV, the successful infection of BV2 cells was verified by plaque assay (Figure 5A) and immunofluorescence analysis (Figure 5B). Next, we examined the expression levels of selected lncRNAs in JEV-infected BV2 cells by quantitative real-time PCR. Our results were concordant to those as observed in JEV-infected mice brain samples (Figure 5C). To examine whether lncRNA E52329 and N54010 are involved in JEV-mediated inflammatory process, the effect of these lncRNAs on the regulation of inflammatory cytokine production was determined. We analyzed the outcomes of silencing of these two lncRNAs using lncRNA-specific siRNAs in BV2 cells. First, we confirmed that the siRNAs significantly inhibited the expressions of lncRNAs in BV2 cells (Figure 6A). To determine the role of lncRNAs in inflammatory cytokine production, the cells were transfected with siE52329, siN54010 or non-specific control siRNA, and then infected with JEV. The results revealed that knockdown of these lncRNAs significantly decreased the production of inflammatory cytokines at mRNA and protein levels (Figures 6B,C). Furthermore, we found that treatment of cells with siE52329 or siN54010 did not exhibit any effect on virus replication in JEV-infected BV2 cells, as viral titers were similar to those in control cells (Figure 6D). Furthermore, to examine whether these two lncRNAs are microglia specific, we selected another immune cell line (RAW264.7) as a control. The expression pattern of lncRNAs and the effects of lncRNAs-specific siRNAs on inflammatory cytokine production and virus replication were analogous to those observed in BV2 cells (Figures S4A–D in Supplementary Material). Thus, these data indicate that lncRNAs participates in regulating JEV-mediated inflammation.

Our previous study demonstrated that JNK plays an important role in JEV-mediated inflammation (12), and we found that MKK4/7, the upstream kinase of JNK, is the most enriched pathway in this study. Therefore, we wondered whether the selected lncRNAs (E52329 and N54010) can regulate the MKK/JNK pathway. To validate it, we first examined the phosphorylation of MKK4 at Thr²⁶¹ and JNK at Thr¹⁸³ and Tyr¹⁸⁵ in JEV-infected mouse brain samples (Figure 7B). Each of these phosphorylation events is associated with increased kinase activity (40, 41). To examine the role of lncRNA E52329 and N54010 in regulating the kinase activity of MKK4/JNK pathway, BV2 cells were transfected with siE52329, siN54010 or non-specific control siRNA, and then infected with JEV. Our findings showed that reduction of these lncRNAs significantly decreased the kinase activity of MKK4 and JNK after JEV infection at 6 and 12 h (Figures 7A,B). On the other hand, NFκB is known as one of the key transcriptional factors of inflammatory cytokines in addition to AP-1 which can be activated by JNK. We also measured whether the lncRNA E52329 and N54010 can modify the NFκB pathway. We found that the phosphorylation of IκBα, a direct inhibitor of NFκB, was increased and IκBα protein was degraded upon JEV infection, but blocking the lncRNAs did not alter the phosphorylation level and protein amount of IκBα (Figures 7A,B). Thus, we consider that the selected lncRNAs may regulate JEV-induced inflammation through activation of MKK4/JNK pathway but not NFκB pathway.