Small RNA sequencing and profiling of serum-derived exosomes from African swine fever virus-infected pigs

Abstract

African swine fever (ASF) virus (ASFV) is responsible for one of the most severe swine diseases worldwide, with a morbidity rate of up to 100%; no vaccines or antiviral medicines are available against the virus. Exosomal miRNAs from individual cells can regulate the immune response to infectious diseases. In this study, pigs were infected with an ASFV Pig/HN/07 strain that was classified as acute form, and exosomal miRNA expression in the serum of infected pigs was analyzed using small RNA sequencing (small RNA-seq). Twenty-seven differentially expressed (DE) miRNAs were identified in the ASFV-infected pigs compared to that in the uninfected controls. Of these, 10 were upregulated and 17 were downregulated in the infected pigs. All DE miRNAs were analyzed using gene ontology (GO) terms and the Kyoto Encyclopedia of Genes and Genomes (KEGG) database, and the DE miRNAs were found to be highly involved in T-cell receptor signaling, cGMP-PKG signaling, Toll-like receptor, MAPK signaling, and mTOR signaling pathways. Furthermore, the Cytoscape network analysis identified the network of interactions between DE miRNAs and target genes. Finally, the transcription levels of four miRNA genes (ssc-miR-24-3p, ssc-miR-130b-3p, ssc-let-7a, and ssc-let-7c) were examined using quantitative real-time PCR (qRT-PCR) and were found to be consistent with the small RNA-seq data. These DE miRNAs were associated with cellular genes involved in the pathways related to immune response, virus–host interactions, and several viral genes. Overall, our findings provide an important reference and improve our understanding of ASF pathogenesis and the immune or protective responses during an acute infection in the host.

Total 27 miRNAs were differentially expressed in serum-derived exosomes from African swine fever virus infected pigs and results showed that 21 differentially expressed miRNAs may regulate the expression of genes which involved in several immune-related pathways. This study provides information that is helpful to improve the understanding of pathogenesis of African swine fever virus.

Introduction

African swine fever (ASF) is the most severe swine disease worldwide, afflicting pigs of all ages and breeds; its occurrence is to be reported to the World Health Organization for Animal Health (OIE). The ASF virus (ASFV) is highly virulent and remains a global threat because of the lack of a vaccine and the ability of the virus to survive in varying environmental conditions (Cubillos et al., 2013; Gallardo et al., 2015; Galindo and Alonso, 2017; Kapoor et al., 2017). In addition, as a highly contagious virus, the exposure of pigs to ASFV results in up to 100% morbidity, and the mortality rate due to ASF depends on the virulence of the virus, the host, and transmission cycles (Kapoor et al., 2017; Quembo et al., 2018). Since the first reported outbreak of ASFV in China in 2018, the disease has spread to Mongolia, Vietnam, Cambodia, Hong Kong, Korea (Democratic People’s Republic of), Laos, Myanmar, Philippines, Korea (Republic of), Timor-Leste, Indonesia, Papua New Guinea, India, and Malaysia (Vergne et al., 2020). An ASF outbreak in Vietnam was reported officially in early February 2019, which spread across the country within 7 mo. In 2019, approximately six million pigs on infected farms and households were culled in Vietnam, indicating the ­socio-economic impact of the virus on the pig industry (Truong et al., 2020; Tran et al., 2021). According to the report of the Department of Animal Health in Vietnam, in total 3,029 outbreaks at 405 districts in 59/63 provinces/cities, approximately 280,000 pigs (0.99% of the total pig population in Vietnam) have been culled in 2021.

Exosomes are cup-shaped extracellular vesicles with a diameter of 30–150 nm that are released by most cell types such as macrophages that the host cell target of ASFV and are present in body fluids, such as blood, urine, saliva, and breast milk (Simons and Raposo, 2009; Vlassov et al., 2012; Hofmann et al., 2020). Exosomes are intraluminal vesicles (ILVs) that are released into the extracellular environment after the fusion of multivesicular bodies (MVBs) and plasma membranes (Simons and Raposo, 2009; Hessvik and Llorente, 2018). Exosomes contain lipids, proteins, and nucleic acids, including DNA, mRNA, and microRNA (miRNA) (Bobrie et al., 2011; Sato-Kuwabara et al., 2015). Specifically, certain miRNAs enter exosomes and can be transferred to target cells as exosomes circulate in body fluids (Zhang et al., 2015). Thus, exosomes play an important role in cell–cell communication during the immune response (Mathivanan et al., 2010). For example, exosomes released from B lymphocytes infected with Epstein-Barr virus may transfer the MHC II molecules and activate certain CD4⁺ T cell clones (Raposo, 1996). Also, exosomes from the herpes simplex virus 1-infected cells are able to deliver the viral miRNAs, mRNAs, and STING (a stimulator of IFN response) to uninfected cells (Kalamvoki et al., 2014).

miRNAs are noncoding small RNAs with an average length of 22 nucleotides (nt) (Kloosterman and Plasterk, 2006); they can regulate gene expression via two mechanisms: mRNA cleavage and translational repression (Valencia-Sanchez et al., 2006). In addition, miRNAs are involved in apoptosis, metabolism, development, and viral infection (Kloosterman and Plasterk, 2006). In addition, exosomal miRNA profiles of healthy and cancer samples are different (Sato-Kuwabara et al., 2015). Therefore, exosomal miRNAs can be used as indicators of disease status.

Exosomes have multiple functions and are particularly applicable in vaccine and biomarker development for diseases. Exosomes transfer their genetic materials to recipient cells and those exosomes are protected from degradation, have high biocompatibility as well as long circulation times, so exosomes show great potential as therapeutic tools (Andaloussi et al., 2013; Zhu et al., 2018) therefore, researchers have been interested in miRNA profiling of exosomes for developing diagnostic markers (Mathivanan et al., 2010). Since ASFV outbreaks have a significant impact on the swine industry and their incidence is increasing, information on exosomal small RNA-seq of ASFV-infected pigs is important. Recently, some studies have performed transcriptomic analysis of ASFV-infected porcine cells or tissues (Nunez-Hernandez et al., 2017; Ju et al., 2021), but there have been no similar studies on exosomes in ASFV infection. In this study, we infected pigs with ASFV belonging to the p72 genotype II and CD2v serotype 8. We evaluated the virulence of ASFV strain Vietnam/Pig/HN/07 (Pig/HN/07) and analyzed exosomal miRNAs in pig serum-derived exosomes against ASFV infection.

Materials and Methods

The study was conducted in compliance with the institutional rules for the care and use of laboratory animals and the protocols were reviewed by Department of Animal Health (DAH) and approved by the Ministry of Agriculture and Rural Development (MARD) of Vietnam (TCVN 8402:2010 and TCVN 8400-41:2019).

Cell culture, virus isolation, and HAD assay

Primary porcine alveolar macrophages (PAMs) were collected from the lungs of 2-week-old pigs that tested negative for ASFV, using conventional PCR, according to the OIE recommendation and real-time PCR as previously described (Tran et al., 2019, 2020; OIE, 2021). The cells were maintained in RPMI 1640 medium with 10% FBS and 1% antibiotics (Thermo Scientific, Waltham, MA, USA) at 37 °C with 5% CO₂, according to the recommendations of the World Organization for Animal Health (OIE, 2021).

In this study, we used the ASFV/Vietnam/Pig/HN/07 virus strain that was isolated in Hanoi in 2019; it belonged to the p72 genotype II and CD2v serotype eight based on the sequence of the intergenic region between the I73R and I329L genes I (IGR I) (Tran et al., 2021). We performed the hemadsorption (HAD) assay as described previously, with minor modifications (Thanh et al., 2021). Primary PAM cells (1 × 10⁴ cells/well) were seeded in 96-well plates, and the virus was added to the plates and titrated in triplicate using ten-fold dilution. ASFV load was determined by identifying the characteristic rosette formation, which represents HAD of erythrocytes around infected cells. HAD was observed for 5 d, and 50% HAD doses (HAD50) were calculated using the method described by Reed and Muench (1938).

DNA extraction from ASFV and real-time PCR

Genomic DNA of the ASFV was extracted using the QIAamp DNA Mini Kit (QIAgen, Hilden, Germany). The DNA of the ASFV in the supernatant was detected by real-time PCR using viral p72 gene-specific primers as recommended by the OIE (OIE, 2021). Real-time PCR was carried out an Agilent AriaMx Real-Time PCR System (Agilent, SC, CA, USA) according to the OIE-recommended procedure described in a previous study (King et al., 2003); DNA of ASFV-positive samples was used as the positive amplification control, and nuclease-free sterile water was used as the negative amplification control.

Animal experiment

Six 8-week-old pigs (Yorkshire × Landrace) were obtained from the Thuy Phuong’s Pig Research Centre of Vietnam National Institute of Animal Science (NIAS), and all tested negative for the following five pathogens: foot-and-mouth disease virus, porcine circovirus 2, PRRS virus, CSF virus, and ASFV. Three pigs were intramuscularly inoculated with the ASFV at a dose of 10⁴ HAD50/mL, and three pigs were used as negative controls. The pigs were monitored daily for temperature and clinical signs (anorexia, depression, fever, purple skin discoloration, staggering gait, diarrhea, and cough). Following this challenge, the clinical signs were recorded daily according to the protocol described in a previous study (Galindo-Cardiel et al., 2013). Cumulative clinical scores (CS) were calculated over time for individual pigs and their respective groups. In addition, body temperature was measured daily from 1 d before infection until the end of the experiment. A body temperature of >40.0 °C for at least two consecutive days was recorded as fever. Blood and swab (oral and rectal) samples were collected daily from each pig to assess the viral DNA load using real-time PCR. Necropsy was performed immediately if the animal died during the day or the next morning if the animal died at night, and tissues were collected for further analysis.

Exosome purification and characterization

Exosomes were purified from serum using the Total Exosome Isolation Reagent (Invitrogen, Carlsbad, CA, USA), according to the manufacturer’s protocol. Particle size was measured to characterize the exosomes using a nanoparticle analyzer (HORIBA, SZ-100, Kyoto, Japan). To further characterize the exosomes, western blotting was performed with pooled exosome samples using three exosomal markers, CD9 (#13174; Cell Signaling Technology, Danvers, MA, USA), CD63 (sc-5275, Santa Cruz Biotechnology, Heidelberg, Germany), and CD81 (#56039; Cell Signalling Technology, Danvers, MA, USA), according to previously described methods (Hong et al., 2020).

Exosomal RNA isolation and small RNA sequencing

Exosomal RNA was isolated separately using the miRNeasy Serum/Plasma Kit (Qiagen, Hilden, Germany) according to the manufacturer’s protocol. Three samples from the control and infected groups at 5 d post infection (dpi) were chosen based on results of clinical assessment that all ASFV-infected pigs at 5 dpi showed clinical, digestive, or respiratory symptoms and were used for library construction and small RNA-seq. A small RNA-seq library was constructed using the SMARTer smRNA-Seq Kit for Illumina (TAKARA Bio Inc., Otsu, Shiga, Japan), according to the manufacturer’s protocol. Small RNA-seq was performed using the Illumina platform (Illumina Inc., San Diego, CA, USA) by Macrogen (Seoul, Republic of Korea).

Sequencing data analysis

FastQC v0.11.7 (http://www.bioinformatics.babraham.ac.uk/projects/fastqc/) was used to filter the raw reads based on quality after sequencing. Cutadapt 2.8 (https://cutadapt.readthedocs.org/en/stable/) was used to remove adapter sequences. Trimmed reads should be at the minimum of 18 bp to be considered reliable for analysis. Therefore, short reads that below 17 bp in read length after adapter trimming were not used for analysis. The processed reads, including trimmed and non-adaptor reads, were used to analyze long targets (≥50 bp). And the trimmed reads were used as processed reads to analyze short targets (<50 bp). Processed reads were gathered to form a cluster containing non-redundant reads that were 100% matched to the sequence identity with processed reads as well as read length. The alignment of the cluster to the swine reference genome (Sscrofa11.1), miRBase v22.1 (http://www.mirbase.org/), and noncoding RNA database, RNAcentral 14.0 (https://rnacentral.org/) was performed to classify the known miRNA, novel miRNA, snoRNA, snRNA, rRNA, tRNA, Y RNA, scRNA, vault RNA, genome, mycoplasma, TruSeqIndexedAdapter (Kozomara et al., 2019). TruSeqIndexedAdapter was added to check whether the normal TruSeq adapter was mixed or not. Novel miRNAs were predicted using miRDeep2 (https://www.mdc-berlin.de/content/mirdeep2-documentation). The read counts of each miRNA extracted from the mapped miRNAs represent the abundance of each miRNA in the sample. To determine the differentially expressed (DE) miRNAs, statistical analysis was performed using fold change, hierarchical clustering, and exactTest using edgeR (Empirical Analysis of Digital Gene Expression Data in R).

Target genes of DE miRNAs were predicted using miRDB (http://www.mirdb.org/) (Liu and Wang, 2019; Chen and Wang, 2020). However, a pig database is not currently available in miRDB; therefore, we identify human miRNAs with sequences similar to those of DE miRNAs using miRBase V22 (Shi et al., 2012) based on several papers showed that the pig genome is similar to the human genome (Wernersson et al., 2005; Dawson, 2011). Gene ontology (GO) enrichment analysis was performed with target genes whose target score was over 80 recommend by miRDB using the gene ontology resource (http://geneontology.org/). David 2021 (https://david.ncifcrf.gov/) was used for KEGG (Kyoto Encyclopedia of Genes and Genomes) enrichment analysis (Huang da et al., 2009; Huang et al., 2009; Mi et al., 2019). To build the network with DE miRNAs and target genes related to the immune response in GO analysis, Cytoscape 3.9.0 software was used (Shannon et al., 2003).

Quantitative real-time PCR (qRT-PCR)

Only forward primers were designed for quantitative real-time PCR (qRT-PCR) to verify the small RNA-seq results. The mature miRNA sequence that was obtained from the miRBase V22 was used as the forward primer. The mRQ 3ʹ-primer from the Mir-X miRNA qRT-PCR TB Green kit (TaKaRa Bio) was used as the reverse primer. The primers were synthesized by Genotech Co. (Daejeon, South Korea) (Table 1).

Table 1.

Primers for qRT-PCR

miRNAs	Sequences (5ʹ→3ʹ)
ssc-let-7a	TGAGGTAGTAGGTTGTATAGTT
ssc-let-7c	TGAGGTAGTAGGTTGTATGGTT
ssc-miR-24-3p	TGGCTCAGTTCAGCAGGAACAG
Ssc-miR-6529	CAGTGCAATGATGAAAGGGCAT
U6	F: CTCGCTTCGGCAGCACA
	R: AACGCTTCACGAATTTGCGT

The Mir-X miRNA First-Strand Synthesis Kit (TaKaRa Bio) was used for cDNA synthesis. One microgram of RNA from each exosome sample was mixed with 5 μL mRQ buffer (2×), 1.25 μL mRQ enzyme, and water was added to make up the volume to 10 μL. The mixture was incubated for 1 h at 37 °C, then the enzyme was inactivated at 85 °C for 5 min. Next, 90 μL nuclease-free water was added to the template cDNA. miRNA expression was determined using the Mir-X miRNA qRT-PCR TB Green kit (TaKaRa Bio) in a CFX Connect Real-Time PCR Detection System (Bio-Rad, Hercules, California, USA). One microliter of cDNA, 12.5 μL TB Green Advantage Premix (2×), 0.5 μL ROX dye (50×), 0.5 μL forward primer (10 μM), 0.5 μL mRQ 3ʹ-primer (10 μM), and 10 μL ddH₂O were mixed to obtain a final volume of 25 μL. The following cycling conditions were used for qRT-PCR: 10 s at 95 °C for denaturation; 40 cycles of 5 s at 95 °C and 20 s at 60 °C for qRT-PCR; and 60 s at 95 °C, 30 s at 55 °C, and 30 s at 95 °C for the dissociation curve. qRT-PCR was performed in triplicate. Porcine U6 gene was used as the control, and miRNA expression was calculated using the 2^−ΔΔCt method (Livak and Schmittgen, 2001; Devhare et al., 2017; Gallo et al., 2018; Zhen et al., 2018).

Statistical analysis

SPSS software (version 25.0; IBM, Chicago, IL, USA) was used for statistical analysis. Data are expressed as mean ± SEM values. Comparisons between groups were conducted using two-tailed Student’s t-test, and statistical significance was set at P < 0.05.

Results

Clinical assessment

All experimental pigs infected with the Pig/HN/07 strain died 5–8 dpi (Supplementary Figure S1A); they exhibited clear clinical signs, including fever (>41 °C), loss of appetite, depression, lethargy, redness of the skin, cyanosis on the edges of the ears and ends of the tail and legs, respiratory distress, and vomiting. In the infected group, the body temperature highly increased in two out of the three pigs (>40.5 °C) at 3 dpi and all infected pigs developed high fever (40.1–41.2 °C) at 4 dpi, which peaked on days 5 and 6 (40.7–41.8 °C). In contrast, the control group did not develop a fever during the experimental period (Supplementary Figure S1B).

In the pigs infected with the ASFV Pig/HN/07 strain showed typical clinical symptoms such as high fever, anorexia, fatigue, and shortness of breath 3–5 dpi, and skin bleeding 5–6 dpi (Supplementary Figure S1C). However, digestive or respiratory symptoms were observed in all the pigs in the ASFV infection group at 5 dpi. At 5 dpi, all pigs (100%) in the ASFV infection group reached a clinical sign score >10 and they were dead by 8 dpi. These results demonstrated that the Pig/HN/07 virus is highly lethal to domestic pigs.

Virus replication and lesion score in organs

All pigs infected with the virus strain Pig/HN/07 died between 5 and 8 dpi with macroscopic lesions characterized by hemorrhage and permeation of internal organs; the combined results are shown in Supplementary Figure S2. The results indicated that the infected pigs showed typical gross lesions, such as 100% hemorrhage in the submandibular lymph nodes, stomach, liver, kidneys, lungs, and mesenteric membrane.

The ASFV was detected in the blood samples of all infected pigs after 48 h by real-time PCR, and C_q values ranged from 31.16 to 33.45 (Supplementary Table S1). We observed that the C_q values in the blood samples of all pigs increased after infection and reached high threshold values at 4 and 5 dpi, with C_q values ranging from 17.16 to 21.32 and reached high thresholds at 7–8 dpi with C_q values ranging from 16.34 to 16.79. Oral and rectal swabs were collected from virus-inoculated pigs and assessed by real-time PCR for the p72 gene. Viral genomic DNA was detectable from the oral swabs at 3–4 dpi and from the rectal swabs at 4–5 dpi (Supplementary Table S1).

Exosomal small RNA sequencing

Exosomes were extracted from serum samples from the control and infected groups at 5 dpi. The size of exosomes and exosomal markers were determined using a nanoparticle analyzer and western blotting, respectively (Figure 1). The mean size distribution was measured using a nanoparticle analyzer (Figure 1A). The mean size of the exosomes was 160.6 nm and 162 nm in the control and infection groups, respectively. The exosomal markers CD9, CD63, and CD81 were identified by western blotting (Figure 1B). In particular, the CD9 and CD81 protein bands had an apparent molecular mass of approximately 22 kDa, and the CD63 protein band had an apparent molecular mass of approximately 60 kDa (Figure 1B).

Figure 1.

Characterization of the exosomes. (A) Size distribution of the exosomes. Red and blue indicate serum-derived exosomes from control and ASFV-infected groups, respectively. (B) Western blotting analysis. Western blotting was performed using exosomes from the control and infection groups and exosomal markers (CD9, CD63, and CD81).

Small RNA-seq was performed to analyze exosomal RNA expression in the infected and control pigs. The raw and processed data of small noncoding RNA sequencing in the serum of the control and ASFV-infected pigs are shown in Table 2. The total reads ranged from 71.7 to 72.5 million reads in the control group and 41.1–47.8 million reads in the ASFV-infected group (Table 2). In the control group, after processing, the mapped reads ranged from 3.8 to 5.5 million reads, and the ratio of bases with a Phred quality score ≥30 (Q30) was more than 91%. In the ASFV-infected group, the mapped reads ranged from 11.9 to 20.0 million reads, and the ratio of bases with a Phred quality score ≥30 (Q30) was more than 85.9% (Table 2).

Table 2.

Raw and processed small noncoding RNA sequencing data from the serum samples of the control and ASFV-infected pigs

Sample ID	Total read bases	Total reads	Processed reads	Mapped reads	Q30 (%)	Known miRNA
Control-1	3,671,274,015	71,985,765	10,979,125	5,490,966 (50.01%)	92.7	80
Control-2	3,659,622,555	71,757,305	12,246,121	4,342,750 (35.46%)	91.47	114
Control-3	3,697,051,200	72,491,200	11,832,294	3,848,516 (32.53%)	90.98	113
Infection-1	2,437,043,058	47,785,158	26,691,136	20,018,396 (75.0%)	89.68	150
Infection-2	2,128,221,840	41,729,840	28,464,103	20,033,695 (70.38%)	90.51	140
Infection-3	2,097,263,310	41,122,810	19,586,914	11,973,682 (61.13%)	85.88	148

Total read bases = total reads × read lengths. Total read bases means the total number of bases sequenced, and total reads means the total number of reads. Mapped reads indicate that reads were mapped to the reference genome. Q30 (%) means ratio of bases that have a Phred quality score greater than or equal to 30.

The read length distribution of each control and infected sample is shown in Supplementary Figure S3. In particular, the small nuclear RNA (snRNA) length ranged from 100 to 300 nt, with that of snoRNA, tRNA, piRNA, and miRNA being 90, 70–90, approximately 27, and approximately 22 nt, respectively. The RNA compositions of the control and infected samples are shown in Supplementary Figure S4. In both samples, the ratio of known miRNAs and novel miRNAs accounted for 0.02–0.06% and 0.4–3.06% in the control, and 0.04–0.06% and 0.95–2.66% in the infected samples, respectively. Our results show that a large number of novel miRNAs in the serum-derived exosomes in pigs infected with virulent ASFV have remained undetected, and further research that focuses on the characterization and function of novel miRNAs in pigs is required.

Unique clustered reads were separately aligned against the reference genome and precursor miRNAs to predict known and novel miRNAs. Novel miRNAs were predicted from the mature, star, and loop sequences according to the RNA-fold algorithm using miRDeep2. Precursor miRNAs (Pre-miRNAs) are the hairpin precursors of miRNAs that have stem-loop structure (Zeng et al., 2005). The miRNA duplex is composed of two strands, one strand formed RNA induced silencing complex (RISC) and the other strand is called star strand (Medley et al., 2021). There were 457 mature miRNAs, and after processing, 51 mature miRNAs that mapped perfectly to one or more precursor miRNA candidates that have been excised from the genome by miRDeep2 in total were identified in both the control and infected samples. In total, 27 mature miRNAs were DE in infected samples with fold change ≥2 and P < 0.05 (Figure 2A). A total of 10 miRNAs were upregulated and 17 were downregulated in the infection group compared to those in the control group (Figure 2A and B). In addition, hierarchical clustering analysis using the Euclidean method and complete linkage was based on the expression level (normalized value) of the DE miRNAs, which is presented in Table 3 and Figure 2. Our results showed the differential expression between upregulated and downregulated miRNAs in the infected samples compared to that in the control group. The ssc-miR-146a-5p showed the highest fold change (20.6-fold), and ssc-miR-7d-5p showed the lowest fold change (2.3-fold) (P < 0.05) (Table 3, Figure 3). To the best of our knowledge, this is the first study to identify exosomal miRNAs against ASFV infection in pig serum and analyze the miRNA profile successfully.

Figure 2.

Volcano plot and heatmap of the expression level of the control and infection groups. (A) Volcano plot. The log₂ fold change and P-value were obtained from comparing the two groups; then, they were plotted as a volcano plot. X-axis: log₂ fold change; Y-axis: −log₁₀  P-value. Yellow dots show fold change ≥2 and raw P-value < 0.05 and blue dots show fold change ≤−2 and raw P-value < 0.05; (B) heatmap of hierarchical clustering analysis. The hierarchical clustering was performed using the Euclidean method and complete linkage. Similarity of 27 mature miRNAs and 6 samples by expression level (normalized value) using clustering analysis. The red box shows the control group, and the blue box shows the infection group.

Table 3.

Mature miRNA genes differentially expressed in the serum samples of ASFV-infected pigs compared to their expression in the non-infected control

No	Mature miRNA	Fold change (infected/control)	P value*
1	ssc-miR-150	–10.65	8.8424E−08
2	ssc-miR-223	–9.96	2.17906E−09
3	ssc-miR-6516	–9.80	0.001431482
4	ssc-let-7e	–7.07	2.91164E−08
5	ssc-let-7a	–5.83	2.8353E−09
6	ssc-miR-1	–5.05	0.002033109
7	ssc-let-7f-5p	–4.50	2.37681E−05
8	ssc-miR-7-5p	–4.26	0.002985752
9	ssc-let-7c	–4.20	2.41556E−09
10	ssc-miR-4332	–3.66	0.035189543
11	ssc-miR-6529	–3.60	3.3647E−05
12	ssc-miR-151-3p	–3.57	0.011909837
13	ssc-miR-4331-3p	–3.33	0.000251562
14	ssc-miR-451	–3.22	0.02777215
15	ssc-miR-1285	–3.11	0.006556144
16	ssc-let-7g	–2.36	0.009080681
17	ssc-let-7d-5p	–2.36	0.008626252
18	ssc-miR-23b	2.26	0.032426035
19	ssc-miR-574-3p	2.37	0.015518691
20	ssc-miR-142-5p	2.67	0.000403573
21	ssc-let-7d-3p	2.7	0.009640287
22	ssc-miR-185	2.87	0.002903744
23	ssc-miR-296-3p	3.05	0.009575366
24	ssc-let-7i-5p	3.06	0.010844226
25	ssc-miR-24-3p	3.17	0.000367316
26	ssc-miR-130b-3p	3.18	0.001659652
27	ssc-miR-146a-5p	20.62	3.1616E−11

Statistical analysis was performed using fold-change, and significant results were selected on conditions of |fold change| ≥2 and exactTest raw P-value < 0.05.

*The fold change was calculated using normalized value for three control and infected samples. The normalized value that used for calculating fold change are not much difference between each control and infection samples and the fold change is consistent across the three samples.

Figure 3.

Fold change of each day using five exosomal RNA samples. In total, 27 miRNAs were identified with DE. Statistical analysis was conducted using fold change, exactTest using edgeR. |fold change| ≥2 and exactTest raw P-value < 0.05 were used for selecting significant results.

miRNA target prediction and biological pathways analysis

The target genes of DE miRNAs were predicted using miRDB (http://www.mirdb.org/). However, a pig gene database was not available in the miRDB. Therefore, we selected human miRNAs based on their sequence similarities (Supplementary Table S2). Our results showed that no human miRNAs had a sequence similar to that of porcine ssc-miR-4331-3p and ssc-miR-4332. Then, 25 DE miRNAs were used to identify immune-related target genes with scores over 80 and were selected based on the human gene database using miRDB. In total, 3,191 genes were targeted by 25 DE miRNAs. The immune-related target genes are listed in Supplementary Table S3. ssc-miR-23b targeted 132 immune-related genes; and ssc-miR-451 targeted with two immune-related genes (Supplementary Table S3). All target genes were used for GO and KEGG pathway analysis, and the results are shown in Figure 4 and Supplementary Table S4. In particular, the biological process focused on the negative regulation of epithelial cell migration, ventricular septum development, and macromolecule diacylation (Figure 4A). The molecular function mainly focused on sequence-specific DNA binding and protein kinase activity (Figure 4B). The cellular component was focused on chromatin and axons (Figure 4C).

Figure 4.

DE miRNAs in infection/control group and gene ontology (GO) enrichment analysis. (A–C) GO analysis in biological process, molecular function, and cellular component. The target genes of DE miRNAs were used.

In addition, we mapped DE Genes (DEG)s using the KEGG database for signaling pathway analysis based on DAVID Bioinformatics Resources version 2021 (P < 0.05). Of the 25 DE miRNAs, 3,191 target genes were identified. Cellular target genes were functionally analyzed using the KEGG pathway database (Supplementary Table S4). In the case of ssc-miR-6529, ssc-let-7d-3p, ssc-miR-296-3p, ssc-miR-451, ssc-miR-574-3p, and ssc-miR-151-3p miRNAs, the related pathways were not identified, but some of the target genes were related to ASFV infection (Supplementary Table S4). Twenty-one DE miRNAs were significantly related to immune response pathways (signaling pathways regulating pluripotency of stem cells, mTOR signaling pathway, cytokine-cytokine receptor interaction, T cell receptor signaling pathway, cGMP-PKG signaling pathway, PI3K-Akt signaling pathway, or MAPK signaling pathway) and with some processes related to pathogenesis and virus-host interaction, neurotrophin, apoptosis, and autophagy (Supplementary Table S4), which represented the largest group and may have an important role in regulating pig response to ASFV infection.

Furthermore, we analyzed the network of miRNAs and target genes related to immune response (Figure 5). The results demonstrated that 24 DE miRNAs significantly interacted with 262 target genes. In particular, the let-7 miRNA family included ssc-let-7f-5p, ssc-let-7i-5p, ssc-let-7a, ssc-let-7c, and ssc-let-7d-5p; ssc-let-7e and ssc-let-7g interacted with more than 30 target genes related to immune response, including several important immune genes, such as IL-13, CD164, MAPK8, CD59, MAP3K1TNFSF9, and CCL7 (Figure 5). These results suggest that let-7 miRNA family may play an important role in regulating immune responses in pigs against ASFV.

Figure 5.

Network of interaction between miRNAs and target genes using Cytoscape. Red color indicates upregulated DE miRNAs, green indicates downregulated, and blue indicates target genes related to the immune response.

Validation of DE miRNAs

We performed qRT-PCR for four miRNAs (ssc-miR-24-3p, ssc-miR-130b-3p, ssc-let-7a, and ssc-let-7c) to validate the small RNA-seq results based on the significant difference of >2.0- or <2.0-fold change between control and ASFV-infected pigs. The four miRNAs were selected based on the read counts, number of immune-related genes, pathway analysis, and function in the immune system. The expression levels of ssc-miR-24-3p and ssc-miR-130b-3p were upregulated by 3.174- and 3.175-fold, respectively, in the serum of pigs infected with ASFV (Table 3, Figure 6). In contrast, the expression levels of ssc-let-7a and ssc-let-7c were significantly downregulated in infected samples compared to those in the control group by 5.83- and 4.20-fold, respectively (Table 3, Figure 6). The qRT-PCR results of the four miRNAs were highly correlated with the small RNA-seq results. In particular, the expression levels of ssc-miR-24-3p and ssc-miR-130b dramatically increased by 9.26- and 4.06-fold, respectively (P < 0.001). In contrast, ssc-let-7a and ssc-let-7c miRNAs were significantly decreased by 0.19-and 0.13-fold change (P < 0.001), respectively, in pigs infected with ASFV compared to those in the control group (Figure 6).

Figure 6.

Validation of DE miRNAs by qRT-PCR. The bar graphs indicate the average of fold change from individual samples. The expression of miRNAs was normalized using the expression of pig U6. Error bar indicates the SEM (***P < 0.001). Experiments were performed on individual samples in triplicate.

Discussion

The pathogenicity of ASFV varies from acute to chronic disease depending on the strain. This pathogenicity plays an important role in controlling and managing the ASF epidemic (Dixon et al., 2019b; Malogolovkin and Kolbasov, 2019; Sanchez et al., 2019). In the current study, we aimed to obtain information regarding the pathogenicity of the Pig/HN/07 ASFV strain, by profiling total exosomal miRNAs in serum samples from ASFV-infected pigs. To the best of our knowledge, this is the first study to evaluate the pathogenicity and global profile of exosomal miRNA expression in the serum of pigs inoculated with an ASFV strain belonging to the p72 genotype II and CD2v serotype 8, with the intergenic region between the I73R and I329L genes I (IGR I), isolated from the field in Vietnam.

Previous studies have demonstrated that the incubation period of ASF ranges from 3 to 19 d, depending on the virulence of the strain, dosage, and host characteristics (Pietschmann et al., 2015; Gallardo et al., 2017; Niederwerder et al., 2019; Pikalo et al., 2019; Zhao et al., 2019). In our study, the pigs infected with 10⁴ HAD50 doses of the Pig/HN/07 ASFV strain exhibited early disease signs by 3–4 dpi, and all infected pigs died between 5 and 9 dpi. Recent reports indicated the acute form of ASF based on the short incubation and early death period being 5–9 dpi for the Pig/HLJ/18 ASF strain from China (Zhao et al., 2019) and 5–8 dpi for the Georgia 2007/1 ASFV strain (O’Donnell et al., 2016; Niederwerder et al., 2019). Notably, the disease signs and necropsy observations made in this study in the Pig/HN/07 ASFV strain-infected group are similar to those caused by other acute pig diseases. Overall, the Pig/HN/07 ASFV strain in this study caused an acute form of ASF based on the short incubation period, early death, and various clinical and gross pathological findings.

Exosomes circulate through biological fluids, communicate with neighboring cells, and are especially involved in intercellular communication. Exosomal miRNAs can be transferred to other cells and regulate target genes, contributing to the immune response (Ramachandran and Palanisamy, 2012). Exosomal miRNAs can be used as diagnostic biomarkers; exosomal miRNA-373 is a biomarker for breast cancer, and exosomal miR-21 is a biomarker for colorectal cancer diagnosis (Sun et al., 2018).

To date, vaccines, drugs, and effective methods for controlling ASFV outbreaks in the pig industry are limited. ASFV outbreaks in Asia, particularly in Vietnam and China, which are the world’s largest pig producers, have significantly decreased food and environmental security in the region. Therefore, improving current anti-ASFV drugs, developing novel ASFV vaccines, and understanding host–pathogen interactions using miRNA and immune-related genes that regulate the immune system should be prioritized. Recently, several studies have focused on global miRNA expression in the tissues or blood samples of ASFV-infected pigs (Nunez-Hernandez et al., 2017; Ju et al., 2021), but there have been no studies on miRNAs isolated from serum-derived exosomes from ASFV-infected pigs. Previous studies have demonstrated that miRNAs regulate the expression of immune-related genes and the immune system by controlling immune or signaling pathways in ASFV-infected pigs (Nunez-Hernandez et al., 2017; Ju et al., 2021). Our results revealed the differential expression of 27 miRNAs in serum-derived pig exosomes of post infection with a virulent ASFV strain (fold change ≥ 2 and P < 0.05). A total of 10 miRNAs were upregulated and 17 miRNAs were downregulated in serum-derived exosomes of pigs infected with the virulent ASFV strain compared to their expression in the control at 5 dpi. Moreover, the downregulation of miRNA may be related to the spike in the upregulation of immune-related genes and cytokines, with a dramatic downregulation of leukocytes, hemorrhage, and death of ASFV-infected pigs (Nunez-Hernandez et al., 2017; Ju et al., 2021). Recent studies have demonstrated that the downregulation of miRNAs as key regulators of immune or immune-related genes involved in regulating the immune response, called a “cytokine storm,” during infection of pigs with a virulent ASFV strain included the B or T-cell receptor signaling pathway, Toll-like receptor pathway, NOD-Like/RIG-I-like receptor signaling pathway, nature killer cell mediated cytotoxicity, chemokine signaling, Fc epsilon RI signaling pathway, Fc gamma R-mediated phagocytosis, and leukocyte transendothelial migration (Davidson-Moncada et al., 2010; Ferretti and La Cava, 2014; Wu et al., 2015). Our results also showed that most of the miRNAs downregulated in the serum-derived exosomes of the virulent ASFV strain-infected pigs were involved in the immune system. This included signaling pathways regulating pluripotency of stem cells, mTOR signaling pathway, cytokine–cytokine receptor interaction, T cell receptor signaling pathway, cGMP-PKG signaling pathway, PI3K-Akt signaling pathway, MAPK signaling pathway, or apoptosis. Therefore, it is suggested that exosomal DE miRNAs could be a key regulator of the immune system in pigs infected with virulent ASFV strain.

Previous studies have indicated that the cAMP, and cGMP-PKG, Toll-like receptor (TLR), and PI3K-Akt signaling pathways play an important role in the progression of ASFV entry via biological processes (Zhang et al., 2017; Li et al., 2018). The cAMP signaling pathway regulates pivotal physiological processes, including metabolism, calcium homeostasis, gene transcription, muscle contraction, and cell programming (Zhang et al., 2017; Li et al., 2018), and cGMP regulates the production of nitric oxide (NO) and natriuretic peptide (NPs) and affects physiological processes (Zhang et al., 2017; Li et al., 2018). The cGMP/PKG signaling pathway is associated with and regulates the replication of some viruses, such as porcine reproductive and respiratory syndrome virus, carp virus, and ASFV (Zhang et al., 2017; Li et al., 2018). The TLR signaling pathway plays an important role in the ASFV pI329 L protein expressed in the cell membrane and it inhibits TLR3-mediated induction of IFN-β and activation of both NF-κB and IRF3 (Dixon et al., 2019a). On the other hand, the PI3K-Akt signaling pathway regulates ASFV replication through its association with the P112, P54, or P72 genes of ASFV, which play important roles in the virulence and replication of ASFV (Wang et al., 2021). Our results indicated that miRNAs in serum-derived exosomes might affect the regulation of ASFV replication through immune or signaling pathways, such as the cGMP/cAMP/ PI3K-Akt signaling pathways.

We identified eight let-7 family miRNAs, of which six were downregulated and two were upregulated in serum-derived exosomes in ASFV-infected cells compared to their expression in the control group. Several reports have indicated that miRNA-let-7 plays an important role in the replication and regulation of infectious diseases caused due to several viruses, including HIV, PRRS, high pathogenic avian influenza virus, swine influenza virus, SARS-CoV-2, hepatitis B and C viruses, Ebola virus, Enterovirus 71, and Pestivirus, through the regulation of the immune system such as Toll-like receptor signaling, cGMP/cAMP/PI3K-Akt signaling pathways, mTOR signaling, cytokine–cytokine receptor interaction, PI3K-Akt signaling, and MAPK signaling pathways (Banerjee et al., 2013; Chafin et al., 2014; Letafati et al., 2022). Furthermore, miR-let-7 can inhibit protease activity and increase the antiviral IFN response by regulating the expression of Bach1 or BCL-2 (Chen et al., 2019). Our results show that ­miRNA-let-7 closely interacts with immune-related genes such as cytokine, TNF family, MAPK family, STAT, cytotoxic, or regulatory genes. Therefore, it may be closely associated with the regulators, control of the immune system, and replication of ASFV in pigs.

Our results showed that miR-24-3p was upregulated in infected pigs compared with the control samples. A previous study reported that herpes simplex virus 1 infection activates MAPK and upregulates miR-24-3p, which targets STING related to antiviral response. Therefore, miR-24-3p plays an important role in herpes simplex virus 1 replication (Sharma et al., 2021). In addition, miR-24-3p may induce porcine reproductive and respiratory syndrome virus by regulating the expression of heme oxygenase-1, which suppresses the replication of several viruses, such as influenza virus and hepatitis C virus (Xiao et al., 2015).

In our sequencing results, miR-130b-3p was also upregulated in ASFV-infected exosome samples compared to its expression in the control samples. Some studies have revealed that miR-130b-3p may inhibit the replication of the hepatitis C virus by regulating DDX6, LDLR, HCCS, INTS6, NPAT, and E2FS expression (Li et al., 2017). Furthermore, miR-130b-3p might inhibit extracellular cod-inducible RNA-binding protein (eCIRP) through the TLR4 pathway, inducing inflammation (Gurien et al., 2020).

miRNA-23b was significantly upregulated in serum-derived exosomes of pigs infected with the virulent strain Pig/HN/07 compared to that in the control. Target prediction and KEGG pathway analysis showed that 132 genes were involved in many immune responses, such as MAPK signaling pathway (Figure 5, Supplementary Table S4). In addition, these target genes were involved in pathways associated with ASFV-cell interaction and replication, such as endocytosis, the cGMP-PKG signaling pathway, focal adhesion, apoptosis, and signaling pathways regulating pluripotency of stem cells (Alonso et al., 2013). Moreover, the miRNA-23 family includes miRNA-23a and miRNA-23b, which are significantly downregulated in the spleen and submandibular lymph nodes 7 d after infection with the virulent ASFV E75 strain and are involved in apoptosis and ER stress processes through a target or are associated with Bcl2, CASP3, and EIF2a (Nunez-Hernandez et al., 2017; Ju et al., 2021). Therefore, the miRNA-23 family, including miRNA-23b, may play an important role in the apoptotic immune response against ASFV infection.

To our knowledge, this result is the first finding using the small RNA-seq approach to identify the global DE miRNAs in serum-derived exosomes of pigs infected with virulent ASFV. However, different virulent, sub-genotypes of ASFV and a larger number of animals are needed to better understand between miRNAs with target genes and genes of ASFV.

Conclusions

In summary, our results demonstrated that pigs infected with ASFV showed acute disease with fever and hemorrhagic signs. We identified global DE miRNAs in serum-derived exosomes of pigs infected with virulent ASFV. In total, 27 differentially expressed (DE) miRNAs were identified in the ASFV-infected pigs compared to that in the uninfected controls. These DE miRNAs were associated with cellular genes involved in the pathways related to immune response, virus–host interactions, and several viral genes. Our results will provide a more accurate understanding of the relationship of miRNA with target genes and ASFV genes, and the role of miRNAs in virulent ASFV infection as well as host–pathogen interactions in pigs.

Glossary

Abbreviations

ASF

African swine fever

ASFV

African swine fever virus

CAM

cell adhesion molecule

CN

copy number

DE

differentially expressed

DPI

days post infection

KEGG

Kyoto Encyclopedia of Genes and Genomes

miRNA

microRNA

HAD

hemadsorption

PAM

primary porcine alveolar macrophages

Small RNA-seq

Small RNA sequencing

Conflict of Interest

The authors have no competing financial interest to declare.

Author Contributions

ADT, SK, HVD, and YHH conceived and designed the study. ADT, SK, YH, JH, THV, NCT, TVH, HTN, HTTT, HVD, and YHH performed experiments. ADT, HTTT, and SK analyzed the data. HVD and YHH contributed to the preparation of reagents and materials, and using analytical tools. ADT, HTTT, SK, HVD, and YHH wrote the manuscript. All the authors have read and approved the final manuscript.