Mir-331-3p Inhibits PRRSV-2 Replication and Lung Injury by Targeting PRRSV-2 ORF1b and Porcine TNF-α

Porcine reproductive and respiratory syndrome (PRRS) caused by a single-stranded RNA virus (PRRSV) is a highly infectious respiratory disease and leads to huge economic losses to the swine industry worldwide. To investigate the role of miRNAs in the infection and lung injury induced by PRRSV, the differentially expressed miRNAs (DE-miRs) were isolated from PRRSV-2 infected/mock-infected PAMs of Meishan, Landrace, Pietrain, and Qingping pigs at 9, 36, and 60 hpi. Mir-331-3p was the only common DE-miR in each set of miRNA expression profile at 36 hpi. Mir-210 was one of 7 common DE-miRs between PRRSV infected and mock-infected PAMs of Meishan, Pietrain, and Qingping pigs at 60 hpi. Mir-331-3p/mir-210 could target PRRSV-2 ORF1b, bind and downregulate porcine TNF-α/STAT1 expression, and inhibit PRRSV-2 replication, respectively. Furthermore, STAT1 and TNF-α could mediate the transcriptional activation of MCP-1, VCAM-1, and ICAM-1. STAT1 could also upregulate the expression of TNF-α by binding to its promoter region. In vivo, pEGFP-N1-mir-331-3p could significantly reduce viral replication and pathological changes in PRRSV-2 infected piglets. Taken together, Mir-331-3p/mir-210 have significant roles in the infection and lung injury caused by PRRSV-2, and they may be promising therapeutic targets for PRRS and lung injury/inflammation.


INTRODUCTION
Porcine reproductive and respiratory syndrome (PRRS) is characterized by severe reproductive failure and respiratory distress in pigs (1,2). The disease was first reported in the late 1980s in the United States and first found in Beijing, China, in 1995 (3)(4)(5)(6). At present, there are still many kinds of PRRS virus (PRRSV) sublineages in China, which poses a great threat to swine industry (7,8).
PRRS is caused by a small, enveloped positive-sense, single-stranded RNA virus, PRRSV, which belongs to the family arteriviridae. The PRRSV genome is ∼15 kb in length and contains at least

Ethics Statement
All animal procedures were approved by the Scientific Ethic Committee of Huazhong Agricultural University, Wuhan, China.

Cell Line and Virus
Marc-145 cells were obtained from China Center for Type Culture Collection (CCTCC) and cultured in RPMI 1640 (Hyclone, Logan, UT, USA) supplemented with 10% fetal bovine serum (CLARK Bioscience, Virginia, USA) in an incubator at 37 • C with 5% CO 2 . PRRSV-2 strain WuH3 (GenBank accession No.HM853673) was kindly provided by Dr. Xiao Shaobo.

DE-miRs Analyses
PAMs were isolated from Pietrain (P), Qingping (QP), Meishan (MS), and Landrace (L) pig breeds by bronchoalveolar lavage under aseptic conditions (28,29). Bronchoalveolar lavage was performed using pre-chilled PBS containing 200 µg of penicillin and 200 U of streptomycin per mL, after which bronchoalveolar lavage cells were collected, filtered, and centrifuged. PAMs were washed three times with PBS, after which they were suspended and cultured in RFMI 1640 medium with 10% FBS containing 100 µg of penicillin and 100 U of streptomycin per mL. After incubation for 2 h at 37 • C, the culture medium was changed to further purify PAMs by plastic adherence of PAMs on cell culture flasks. Then, the PAMs of 5 pigs of each breed were infected with PRRSV-2 strain WuH3 at multiplicity of infection (MOI) of 0.1 PFU/cell. The PRRSV-2 infected PAMs were collected at 9, 36, and 60 hpi and mixed evenly. The control group (mock-infected) PAMs were infected with culture medium and collected at 9, 36, and 60 h. Total cellular RNA was isolated using the Trizol reagent (Invitrogen, Cashman, CA, USA) to analyze the differential expression of miRNAs. Deep sequencing was performed by the Illumina/solexa Genome Analyzer (BGI, Shenzhen, China). Twenty-four miRNA libraries were constructed. The expression of miRNAs was normalized and analyzed by calculating fold-change and p-value (30,31). A miRNA was labeled as differentially expressed, when |log2(fold change)| ≥1 and p ≤ 0.01.

Bioinformatics and Luciferase Reporter Assay
JASPAR software was used to analyze the TNF-α 5 ′ flanking sequence. TargetScan, miRbase, RNAhybrid, and ViTa software were used to predict target genes of miRNAs. The 3 ′ UTR of TNF-α containing putative mir-331-3p binding site and the 3 ′ UTR of STAT1 containing putative mir-210 binding site were amplified by PCR and cloned into pmirGLO vector (Promega, Madison, Wisconsin, USA), respectively. Mutation sites were identified in the predicted target sites of mir-331-3p/mir-210 in the 3 ′ UTR of TNF-α/STAT1. The ORF1b sequence containing the mir-331-3p binding site and the ORF1b sequence containing the mir-210 binding site were amplified, respectively. Subsequently, these two amplified sequences were connected to the pmirGLO vector and named ORF1b-331-WT and ORF1b-210-WT, respectively. Fusion PCR was used to construct the binding site-specific mutant plasmids, and the two resultant plasmids were named ORF1b-331-MUT and ORF1b-210-MUT, respectively. The primer sequences for plasmid construction were listed in Supplementary Table 1. For luciferase reporter  assay, mir-331-3p or mir-210 were co-transfected with the  corresponding dual-fluorescence reporter plasmid into Marc-145 cells in 24-well plates by using Lipofectamine 2000 TM  (Invitrogen, Cashman, CA, USA). At 48 h post-transfection, the dual-luciferase reporter assay system (Promega, Madison, Wisconsin, USA) was used to measure the luciferase activity.

Cell Transfection and Viral Infection
Marc-145 cells were seeded into 6-well plate and cultured at 37 • C in humidified 5% CO 2 atmosphere. After 24 h incubation, miRNA or siRNA were transfected into cells of each well by using lipofectamine 2000 (Invitrogen, Cashman, CA, USA). The sequences of miRNA mimics/inhibitors and siRNA were listed in Supplementary Table 2. After 5 h, 3% of the cell maintenance medium was added, and the normal growth medium was added after 1 h PRRSV-2 infection at MOI = 0.1. The cells were collected for the extraction of total protein and total RNA at 36 h post viral infection.

RNA Extraction, Reverse Transcription, and qRT-PCR
Total RNA was extracted from tissues or cells with TRIzol reagent (Invitrogen, Cashman, CA, USA). RNA (500 ng) was reversely transcribed with RevertAid First Strand cDNA Synthesis Kit (Thermo Fisher Scientific, Waltham, MA, USA). Quantitative real-time PCR (qRT-PCR) assay was performed by using the SYBR Green real-time PCR Master Mix regents in the Roche LightCyler 480 system (Roche, Mannheim, Germany) according to the manufacturer's protocol. Primers used for qRT-PCR were shown in Supplementary Table 3. U6 or β-actin was applied as the internal control, while the fold changes were calculated by 2 − Ct method. Absolute quantification was used to detect the PRRSV copy number, and the primer ORF7-F/R and ORF7probe were shown in Supplementary Table 3. All experiments were performed at least three times in triplicate.

Animals
Six, 4-week-old piglets were randomly selected and divided into two groups, pEGFP-N1-mir-331-3p treatment group and control group. The pEGFP-N1-mir-331-3p or pEGFP-N1 (at the dose of 2.5 mg/kg body weight) were mixed with D5W solution to finally obtain 3 mL mixture solution, respectively. This 3 mL solution was administered to piglets through intramuscular injection. At 5 h post intramuscular injection, the 1.5 mL of PRRSV-2 strain WuH3 (10 5.2 TCID50) was administered to piglets. The rectal temperature of piglets was measured twice a day. On day 14, we performed pathological dissection and collected all the lungs and PAMs of the piglets. The sacrificed pigs were taken out and the animal experiments were carried out by random and blind method, in vivo experiment, three replicates for each of the two groups. All experiments were performed at least three times in triplicate, excluding average rectal temperatures.

Histological Assay
After being fixed in 4% paraformaldehyde, lung tissues were embedded in paraffin. Lung tissues were analyzed by Hematoxylin-Eosin staining (H&E). The experiments followed the procedures previously reported (32). Finally, these sections were observed under an optical microscope (Olympus, Tokyo, Japan) to detect morphological changes of lung tissues. To further detect the expression of TNF-α, immunohistochemistry (IHC) experiments were performed using specific polyclonal anti-TNF-α (A0277, ABclonal, Wuhan, China). Briefly, tissue blocks were cut into sections. The sections were deparaffinized and dehydrated through xylene and graded alcohols, and then these sections were rehydrated with demineralized water. The sections were blocked with 3% hydrogen peroxide for 30 min, boiled in a 0.01 M sodium citrate buffer for 10 min at high temperature, and boiled for 10 min at low setting to expose the antigen, cooled naturally, and washed 3 times with PBS for 3 min each time. Then, the sections were blocked with goat serum at room temperature for 20 min. Primary antibody was incubated overnight at 4 • C. The sections were washed with PBS and incubated with secondary HRPconjugated antibodies (Servicebio, Wuhan, China) for 30 min. Diaminobenzidine staining and hematoxylin staining were performed. Differentiation was conducted with 1% hydrochloric acerbic for 30 s. The sections were then dehydrated with ethanol series, washed in xylene, embedded in paraffin wax, and photographed with a microscope (Olympus, Tokyo, Japan).

Statistical Analysis
All experiments were performed at least three times in triplicate. The differences were assessed using two-tailed ttest or one-way ANOVA in vitro experiment. Non-parametric Mann-Whitney Statistical test was used in vivo experiment due to the few number of animals available (33,34). Data were presented as mean ± SD. Statistically significant difference was presented at the level of * p < 0.05 and * * p < 0.01. All the histograms and graphs were generated with GraphPad Prism version 5.0 and Adobe Photoshop CS5 software, respectively.

Mir-331-3p and Mir-210 Were Differentially Expressed Between PRRSV-Infected and Mock-Infected PAMs
DE-miRs were analyzed using previously reported methods (30,35). Four sets of miRNA expression profile were synthesized by Venn diagram analysis (Figures 1A-C). The Venn diagrams showed there were the most overlapped DE-miRs (67) between Pietrain and Landrace pigs in the early stage of PRRSV infection (9 hpi). There were the most overlapped DE-miRs Frontiers in Immunology | www.frontiersin.org (8 and 20) between Pietrain and Meishan pigs at 36 and 60 hpi. Among them, only mir-331-3p was significantly differentially expressed in each set of miRNA expression profile at 36 hpi ( Figure 1B), and mir-331-3p was also one of 4 common DE-miRs between PRRSV infected and mock-infected PAMs of Meishan, Pietrain, and Landrace pigs at 60 hpi ( Figure 1C). In addition, mir-210 was one of 7 common DE-miRs between PRRSV infected and mock-infected PAMs of Meishan, Pietrain, and Qingping pigs at 60 hpi. Mir-331-3p and mir-210 were both the common DE-miRs between PRRSV infected and mock-infected PAMs of Pietrain and Landrace pigs at 9 hpi (Figures 1D-I). These results implied that mir-331-3p and mir-210 might contribute to specific responses to PRRSV infection.

Mir-331-3p and Mir-210 Directly Targeted PRRSV-2 ORF1b
The RNAhybrid and ViTa softwares were used to predict the target sites of mir-331-3p and mir-210 in PRRSV genome.
To investigate whether mir-331-3p and mir-210 inhibit PRRSV-2 replication by binding to ORF1b, Marc-145 cells were transfected with the mimics or inhibitor of each miRNA (10 nM), followed by infection with PRRSV-2 strain WuH3 at an MOI of 0.1. The results showed that overexpression of mir-331-3p significantly inhibited PRRSV-2 copy number (p = 0.017, Figure 2H). Meanwhile, a similar tendency was observed at N protein (PRRSV-2 ORF7 encoded) level ( Figure 2I). Additionally, the inhibition of mir-331-3p led to the significantly increased the number of PRRSV-2 copies (p = 0.03, Figure 2J) and expression level of N protein (Figure 2K) in Marc-145 cells. In addition, mir-210 also significantly inhibited PRRSV-2 replication and N protein expression (p = 0.02, p = 0.008, Figures 2L-O). The above-mentioned findings provided evidence that mir-331-3p and mir-210 could inhibit the viral replication by targeting PRRSV-2 genome.

Mir-210 Directly Targeted Porcine STAT1
The 9 potential target genes of mir-210 were selected by bioinformatics and were quantitatively analyzed by qRT-PCR. The results showed that mir-210 could reduce the mRNA expression level of STAT1 gene about 3 times (p = 0.012), and significantly inhibit the expression of SCARA3 and DIMT1 (p = 0.026, p = 0.018, Figure 3G). Additionally, Western blot assay revealed that mir-210 mimics significantly decreased the protein level of STAT1 (Figure 3H), whereas mir-210 inhibitor significantly increased the protein level of STAT1 ( Figure 3I). To assess whether mir-210 directly targets STAT1 gene (Figure 3F), mir-210 mimics and luciferase reporter plasmid STAT1-WT (containing a portion of the STAT1 3 ′ UTR) were co-transfected into Marc-145 cells, the luciferase activity of STAT1-WT was significantly down-regulated (p = 0.01, p = 0.011, Figure 3J). No change was observed when the putative mir-210 binding sites were mutated (STAT1-MUT) ( Figure 3K). These results confirmed that mir-210 could downregulate the expression of STAT1 by directly targeting it.
Then, six 4-week-old landrace piglets were divided into two groups (n = 3 per group). These piglets were infected with 1.5 mL 10 5.2 TCID50 PRRSV-2 strain (WuH3) at 5 h post injection (2.5 mg/kg) of plasmid pEGFP-N1-mir-331-3p for the experimental group and pEGFP-N1 for the negative control group. The results of qRT-PCR showed that the expression of mir-331-3p was higher in the experimental group (p = 0.05, Figure 5C). The piglets were monitored for clinical signs, including anorexia, lethargy, fever, and weight. Two days after PRRSV-2 infection, pigs in the negative control group showed significant clinical symptoms such as abdominal breathing and increased rectal temperature, while the experimental group pigs presented clinical symptoms at 3 dpi. The average rectal temperature in the experimental group was significantly (p = 0.01) lower than that in the negative control group (Figure 5D). Average daily weight gain in piglets from the negative control group was lower compared with that from the experiment groups after PRRSV-2 infection (Figure 5E). At 14 dpi, all the piglets were euthanized and lungs were collected. The cycle threshold (Ct) of PRRSV ORF7 in lung tissues from the experimental group was found significantly (p = 0.05) higher than that from the negative control group (Figure 5F). The expression of the target gene TNF-α was also significantly (p = 0.05) inhibited by pEGFP-N1-mir-331-3p at both mRNA and protein level (Figures 5G,H). The expressions of inflammationassociated genes MCP-1, ICAM-1, and VCAM-1 in experimental groups were also significantly decreased, compared with those in the negative control groups (p = 0.05, p = 0.05, p = 0.05, Figure 5I).
Afterwards, we also assessed the extent of macroscopic lung lesions and histopathological damage. As shown in Figure 6, the lungs of the negative control group were dark red due to congestion, while those of the experimental group were lighter in color. The interstitial pneumonia was more severe in the lungs of the negative control group than that of the experiment group. Hematoxylin-eosin staining showed that interstitial enlargement and congestion was more prominent in the negative control group than that in the experiment group (Figures 7A,B). Localization of TNF-α was detected in the lung tissues of pEGFP-N1-mir-331-3p group, pEGFP-N1 group by immunohistochemical staining. The result showed that TNFα was mainly localized in lung epithelial cells and PAMs in the lung tissues of PRRSV-2 infected pigs ( Figure 7C). The high expression of pro-inflammatory factor TNF-α induces lung

DISCUSSION
Numerous studies have shown that the host miRNAs are involved in host-pathogen interactions and the regulation of immune responses and inflammation. Previous reports demonstrated that mir-181 inhibited viral replication by targeting PRRSV 3 ′ UTR (39), and that mir-30b-5p played an important role in lung injury in children (27). In our study, total cellular RNA was isolated from PRRSV-2 infected/mock-infected PAMs of Meishan, Landrace, Pietrain, and Qingping pigs at 9, 36, and 60 hpi to analyze the differential expression of miRNAs. To avoid the influence of mixed-leukocyte reactions (MLRs) on miRNA expression caused by mixed PAMs, an improved method was applied to increase the purity of the obtained PAMs (28,29). In addition, PAMs from each animal have been infected by PRRSV before mixing. The unmixed PAMs were used in the qRT-PCR verification of differentially expressed miRNAs. Mir-331-3p was the only common DE-miR between PRRSV-infected and mockinfected PAMs of 4 pig breeds at 36 hpi and one of 4 common DE-miRs between PRRSV-infected and mock-infected PAMs of Meishan, Pietrain, and Landrace pigs at 60 hpi. Mir-210 was one of 7 common DE-miRs between PRRSV-infected and mockinfected PAMs of Meishan, Pietrain, and Qingping pigs at 60 hpi. It is consistent with the previous study that found mir-210 and mir-331 was differentially expressed in PBMCs from HIV-1-infected and uninfected individuals (40).
In addition, mir-331-3p and mir-210 were predicted to directly target PRRSV-2 ORF1b, and verified by double luciferase assay. ORF1b encoded multiple proteins that were further  processed into multiple small protein products including Nsp9, Nsp10, Nsp11, and Nsp12, which were called non-structural proteins (Nsp). Of them, Nsp9 and Nsp10 were key enzymes for RNA synthesis of arterial virus, and closely related to the replication efficiency in vitro and in vivo and related to the increased pathogenicity and fatal virulence for piglets (12). Li et al. (41) showed that PRRSV-specific cytopathic effect (CPE) could be inhibited in the cells by shRNA targeting ORF1b, and that cellular virus titers were decreased by ∼100-folds compared with those of control cells. Li et al. (42) reported that two recombinant adenoviruses expressing shRNA could effectively inhibit PRRSV replication in vitro and in vivo by targeting ORF1b of PRRSV. Our study also revealed mir-210 and mir-331-3p could both significantly inhibit PRRSV replication.
Moreover, bioinformatics analysis and experiment results confirmed that TNF-α is the target gene of mir-331-3p. It is consistent with a previous study that found mir-331-3p targeted TNF-α and notably weakened its expression in VSMC (43). TNF-α is a pleiotropic cytokine that mediates host response to infections and play decisive roles in the outcome of a number of viral infections, contributing to virus control or immune mediated pathology. TNF-α inhibitors have been successfully used in the clinic to treat these immune-mediated diseases (44,45). TNF-α has also been implicated in a variety of pulmonary diseases and plays a crucial role in the occurrence and development of lung injury and fibrosis (46,47). Gomez-Laguna et al. (48) reported the expression of TNF-α in the lungs of pigs infected with PRRSV-1 was correlated with the development of the interstitial pneumonia typical of this disease. Nukuzuma et al. (49) reported TNF-α stimulation could induce JC polyomavirus (JCV) replication through the NF-κB pathway in IMR-32 cells transfected with JCV DNA. Han et al. (15) reported that the pigs infected with HP-PRRSV showed the higher levels of TNF-α and exhibited severe pathological changes of lungs, which were in part responsible for the additional morbidity and mortality observed in HP-PRRSV infection (42). Sun et al. (50) reported that matrine possesses activity against PRRSV/PCV2 co-infection in vitro and suppression of the TLR3,4/NF-κB/TNF-α pathway as an important underlying molecular mechanism. Ge et al. (51) reported that PRRSV replication was suppressed in Marc-145 cells treated with EGCG post-infection, likely because of down-regulation of pro-inflammatory cytokines, such as TNFα. Yang et al. (52) demonstrated that TNF-α might be a major contributor in ii/r-induced lung injury, and that the knockdown of TNF-α alleviated the severity of lung injury. Yu et al. (53) reported that downregulation of TNF-α signals by AT-Lipoxin A4 might be a significant mechanism in the attenuation in severe acute pancreatitis-associated lung injury. Our study also indicated that mir-331-3p inhibited the expression of TNF-α by directly targeting its 3 ′ UTR. Mir-331-3p could attenuate lung injury and significantly inhibit viral replication by intramuscular injection of the expression plasmid of mir-331-3p in vivo. To further illuminate the mechanisms that underlie the impact of TNF-α on lung inflammation/injury, we also examined and found mir-331-3p suppressed the expression of inflammationassociated genes MCP-1, VCAM-1, and ICAM-1 in vitro and in vivo. On the other hand, down-regulation of TNF-a might be beneficial for the early phase of PRRSV infection, and it might also cause a prolongation of PRRSV infection. Activation of NF-κB signaling is one of the most important canonical responses to the stimulation of TNF-α (44). When NF-κB is activated, NF-κB will rapidly transfer from the cytoplasm to the nucleus, acting as transcription factor for several adhesion molecules and inflammatory cytokines, such as VCAM-1, ICAM-1, and MCP-1, which play vital roles in inflammatory diseases such as lung inflammation/injury (54)(55)(56). Thus, we inferred that mir-331-3p suppressed the expression of MCP-1, VCAM-1, and ICAM-1 through inhibiting TNF-α-induced NF-κB activation. Of course, it needs further study.
Bioinformatics analysis and experiment results confirmed that STAT1 is the target gene of mir-210. STAT1 has been identified as a transcription factor, which is the important part of the cell signal pathway JAK/STAT, and plays a key role in lung injury and other inflammatory diseases. STAT1 is positioned as the trigger for an entire set of immune-response genes with antiviral function (57). The activated STAT1 (phosphorylation STAT1) is transported into the nucleus and then promotes the upregulation of pro-inflammatory factors including TNF-α, MCP-1, VCAM-1, and ICAM-1 (58)(59)(60)(61). STAT1 antisense oligonucleotides (ASON) could inhibit the secretion of TNF-α and ICAM-1 in alveolar macrophages (AMs), and STAT1 could become a target of treating pulmonary fibrosis (62,63). In our study, siSTAT1 also suppressed the expression of TNF-α, MCP-1, VCAM-1, and ICAM-1. Thus, we inferred that mir-210 suppressed the expression of TNF-α, MCP-1, VCAM-1, and ICAM-1 through inhibiting STAT1-mediated transcriptional activation. Of course, it needs further study.
In summary, our study demonstrates that mir-331-3p and mir-210 could inhibit PRRSV-2 replication and lung injury by directly targeting PRRSV-2 ORF1b and porcine TNF-α and STAT1, which mediated the transcriptional activation of MCP-1, VCAM-1, and ICAM-1. STAT1 could also upregulate the expression of TNF-α by binding to the promoter region of TNFα (Figure 8). These insights may be applicable to PRRSV and helpful for the treatment of lung inflammation/injury.

DATA AVAILABILITY STATEMENT
The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found in the article/Supplementary Material.

ETHICS STATEMENT
The animal study was reviewed and approved by the Scientific Ethic Committee of Huazhong Agricultural University, Wuhan, China.

AUTHOR CONTRIBUTIONS
XY, YQ, YZ, XG, CH, GL, and QL carried out the experiments. XY, ML, DX, YQ, YZ, and JH analyzed the data. XY, ML, and DX wrote the manuscript. DX and ML designed the experiments and revised the manuscript. All authors contributed to the article and approved the submitted version.