The expression profile of human peripheral blood mononuclear cell miRNA is altered by antibody-dependent enhancement of infection with dengue virus serotype 3

PBMCs were separated from EDTA anticoagulated whole blood samples from healthy adult males with no acknowledged diseases or infections and were isolated by Ficoll-Hypaque gradient centrifugation as adapted in the experimental operating instructions. The level of DENV IgG antibodies detection of donors serum were commissioned by Xishuangbanna Dai Autonomous Prefecture People’s Hospital, the test results showed that the DENV IgG antibodies of donors was negative. Stated, PBMCs plated in T25 flasks at 4.2 × 10⁵ cells per ml were cultivated in RPMI 1640 medium (BI, China) supplemented with 10% fetal bovine serum (FBS, SJQ, China) plus penicillin and streptomycin and incubated 2 h at 37 °C in 5% CO₂ in a humidified incubator. Two hours later, the DENV-3 virus strainthat originated from an epidemic in Guangdong, China in 2014 and DENV-2 anti-prM monoclonal antibody (abcam, AB41473) were added in PBMCs at a multiplicity of infection (MOI) of 5.Cells were infected in triplicate and collected at 0, 8 and 24 h post infection (hpi). Cells infected with DENV-3 and DENV-3 plus DENV-2 anti-prM complex at 0 hpi were used as controls. We defined the different experimental groups as DENV-3–0 h, DENV-3–8 h, DENV-3–24 h, DENV-3-ADE–0 h, DENV-3-ADE–8 h and DENV-3-ADE–24 h. Additionally, a subset of the DENV-3-0 h and DENV-3 ADE–0 h groups were subjected to normalization (the normalization value was set to 1), and these two groups were then designated Con.

miRNA was isolated from cultured DENV-3 and DENV-3-ADE infected PBMCs according to standard miRNeasy Mini Kit (Agilent technologies Santa Clara, US)protocols. The integrity and quality of RNA were evaluated using RNA 6000 Nano Lab Chips on an Agilent 2100 Bioanalyzer (Agilent technologies Santa Clara, US) and estimated by reviewing electropherograms and the RNA integrity number (RIN) of each sample (Table 1). Qualified miRNA samples from three independent experiments of each group were pooled and used for subsequent deep sequencing and library construction.

High-throughput sequencing technology (transcriptome analysis tool), not only detects known transcripts but also promote the discovery of novel transcripts [23]. MicroRNA library construction and sequencing was accomplish by the National Engineering Center for Biochip in Shanghai on an Illumina HiSeq 2000 system.The sequencing data were submitted to the Gene Expression Omnibus (GEO) database (www.​ncbi.​nlm.​nih.​gov/​geo/​) under the accession number GSE98859.

RT-qPCR was used to investigate the relative levels of gene expression among DENV nad ADE infection PBMCs during 48–96 h. Briefly, PBMCs cellular RNA was extracted using standard miRNeasy Mini Kit (Agilent technologies Santa Clara, US)protocolsand then subjected to reverse transcription with a first-strand cDNA synthesis kit before amplification by RT-qPCR. The amplifications were performed using SYBR Advantage qPCR Premix ((TIANGEN, CHINA) with a Bio-Rad CFX96 detection instrument.

The quantitative reverse transcription-polymerase chain reaction was used to validate miRNA expression. For further confirmation, we randomly selected 6 differentially expressed miRNAs for RT-qPCR analysis. miRNA expression was tested by poly(A)-tailed RT-qPCR. For each sample, 1μg of total RNA was polyadenylated and reverse transcribed using poly(A)polymerase with a miScript II RT Kit (QIAGEN, USA), in accordance with the manufacturer’s instructions. Subsequently, each cDNA was amplified on a 7900 HT Sequence Detection System (ABI, USA) with an mRQ 3primer and miRNA-specific 5 primers to quantify specific miRNA sequences; the amplifications were performed using SYBR Advantage qPCR Premix ((TIANGEN, CHINA). All miRNA-specific 5 primers used in the qPCR experiments are shown in the Additional files.

For sequencing data, raw reads achieved from each library were normalized to TPM. For RT-qPCR, the data are expressed as the mean ± standard error of the mean (SEM). Statistical analysis was measured using STATA 11.0 software (stata Corp, College Station, TX, USA). P - value of less than 0.05 was considered to indicated a statistically significant difference.

To determine the changes in miRNA expression patterns in PBMCs during direct infection and ADE infection with DENV-3, global cellular miRNA expression patterns following infection were compared with those of the controls. In this study, only those differentially expressed miRNAs with a p value < 0.05 and a fold change ≥2 or ≤ 0.5 are described. The results showed that, compared with the DENV-3–0 h group, there were 8 up-regulated and 7 down-regulated known miRNAs in the DENV-3–8 h group, as well as 13 up-regulated and 1 down-regulated known miRNAs and 1 up-regulated novel miRNA in the DENV-3–24 h group. Moreover, there were 10 up-regulated and 1 down-regulated known miRNAs in the ADE infection–8 h group, as well as 37 up-regulated and 4 down-regulated known miRNAs in the ADE infection–24 h group relative to the ADE infection-0 h group (Additional file 1: Table S1, Additional file 2: Figure S2, Table 2).

Subsequently, the sum aggregates of differentially expressed known miRNAs were taken from four groups. Then, we used the R language heatmap package to draw an miRNA Heatmap diagram; 50 differentially expressed miRNAs were used to perform hierarchical clustering to construct a heat map based on differential expression patterns with log 2 values (infected/control) and fold changes (Fig. 1a). A positive log 2 value indicated up-regulation, and a negative log 2 value indicated down-regulation. The data for all the differentially expressed miRNAs were also graphed as a Venn diagram (Fig. 1b). Common and distinct differentially expressed known microRNAs in response to direct infection and ADE infection with DENV-3 across all time points were revealed. We found that the patterns induced by DENV-3 differed from those induced by ADE infection at 8 h and 24 h. In addition, the miRNA expression patterns induced by DENV-3 at 8 h also differed from those induced by DENV-3 at 24 h, and the same was true for ADE infection. Therefore, these results demonstrated that miRNA expression patterns in PBMCs following infection with DENV-3 and ADE of DENV-3 are mode-and time-specific and the ADE infection at 24 h group differed the most from the other three groups.

To further confirm our sequencing data, six significantly differentially expressed miRNAs, including hsa-miR-184, hsa-let-7e-5p, hsa-miR-132-3p, hsa-miR-155-5p, and hsa-miR-1246, were chosen for RT-qPCR analysis. The results showed a general consistency between RT-qPCR and high-throughput sequencing results (Additional file 3: Figure S1).

To explore the possible regulatory roles of the 50 known differentially expressed miRNAs during DENV-3ADE and direct infection, we used the TargetScan and miRDB databases to predict possible targets genes. There were 12,509 genes found with TargetScan and 129,810 potential genes found with miRDB. Then, the genes that were identified by both programs were selected as the final target genes for subsequent GO and pathway analysis, and 6494 target genes were screened out (Additional file 4: Table S2). Seventy GO terms were enriched based on a P value < 0.0001 and FDR < 0.0001. The results are shown in Additional file 5: Table S3. The most obvious GO terms for the miRNA targets included DNA-dependent regulation of transcription, multi-cellular organism development, DNA-dependent negative regulation of transcription, DNA-dependent positive regulation of transcription, cell differentiation, protein transport, cell adhesion, cell cycle, apoptotic process, positive regulation of transcription from the RNA polymerase II promoter, nervous system development, chromatin modification, interspecies interaction between organisms, negative regulation of transcription from the RNA polymerase II promoter, ion transport, cell division, protein phosphorylation, biological-process, protein ubiquitination and mitosis (Fig. 2a). Then, the predicted targets genes were subjected to KEGG pathway enrichment analysis using the PANTHER Classification System. We obtained 122 important pathways based on P value <1E^− 7 and FDR < 1 E^− 07, and the results are shown in Additional file 6: Table S4. As with the GO analysis We have displayed the notable pathways of the target genes using a histogram (Fig. 2b). The notable pathways included the pathways in cancer, MAPK signaling pathway, the cGMP-PKG signaling pathway, the cAMP signaling pathway, endocytosis, the neurotrophin signaling pathway, the PI3K-Akt signaling pathway, regulation of the actin cytoskeleton, focal adhesion, axon guidance, proteoglycans in cancer, the longevity-regulating pathway, the Ras signaling pathway, HTL-I infection, the phospholipase D signaling pathway, metabolic pathways, the Rap1 signaling pathway, the Hippo signaling pathway, and the insulin signaling pathway. These results showed that numerous biological processes participate in DENV-3ADE and direct infection of PBMCs.

According to the results of miRNA high-throughput sequencing and target gene prediction analysis, autophagy and innate immunity pathways were also targeted by miRNA. In the process of DENV ADE infection, the expression levels of the SOCS1, SOCS3, RIG-1 and ISG56 genes was found to be different (≥2) comparing to the levels during DENV-3 direct infection. In addition, they showed expression trend that were the opposite of their related miRNAs, hsa-miR-330-3P, hsa-miR-342-3P, hsa-miR-501-3P, hsa-miR-331-3P, hsa-miR-340-3P, hsa-miR-3607-3P, hsa-miR-3614-3P and hsa-miR-374a-3P. It was worth noting that the expression levels of innate immunity and autophagy-related genes RIG-I, MDA5 and ATG5 significantly increased 48 h after DENV-ADE infection, while the expression levels of SOCS1, SOCS3, INF-a and ISG15 decreased. Meanwhile, the expression levels of SOCS1, INF-a, ISG15, and ISG56 greatly increased 96 h after DENV-ADE infection (Fig. 3).

To further profile the function of miRNAs, complex network systems were built, including a network of miRNA genes, a network of important GO terms and a network of pathways. Firstly, the intersections of notable target genes in the GO enrichment and KEGG analysis were selected for constructing the regulatory network of miRNA genes (Fig. 4a). The target genes that were most highly regulated by miRNAs were FZD3, SLC38A2, CCNG2, EIF4E, FRS2, MAP3K2, MAPK8, MEF2C, PCGF5, PRKAA2, SCN1A, SCN7A, SEMA6A, SIX4, SMAD2, SMG1, SORT1, SS18, UBE2D1, and UBE2G1. Secondly, we built a network of significant gene functions based on the Gene Ontology database (Fig. 4b), and the results showed that the GO classifications were mainly related to cell division, cell growth, protein ubiquitination, signal transduction, stress response, etc. Furthermore, the pathway regulation network was constructed and is shown in Fig. 4c. According to the interaction network of significant pathways, the main signaling pathways were the MAPK signaling pathway, the PI3K-Akt signaling pathway, the calcium signaling pathway, apoptosis, etc. Of these, the MAPK signaling pathway plays an important role in the whole signal pathway network.