Background
African swine fever (ASF) is an infectious disease affecting domestic and wild pigs caused by African swine fever virus (ASFV) belonging to the
Asfarviridae family. ASFV is an icosahedral, enveloped, dsDNA virus ranging from 170 to 193 kb [
1]. ASF genomes contain between 150 and 175 ORFs with 34 of them encoding for structural viral proteins and so far, half of them do not have any known function [
2]. Sixteen different genomes have been fully sequenced corresponding to virulent, attenuated and non-pathogenic ASFV strains isolated from pigs, wild boars, ticks and tissue culture cells. So far, 22 genotypes have been identified with several of them currently co-circulating in Sub-Saharan Africa. ASF was first described in 1921 in Kenya [
3] spreading rapidly throughout other African countries. The disease is endemic in most of Sub-Saharan Africa and in Sardinia [
4]. ASFV entered Europe for the first time in the early fifties remaining on the continent until its eradication from Spain and Portugal in 1995 [
5]. In 2007 the virus re-entered Europe through Georgia and since then it has spread to East-Europe, becoming a real threat for all Europe. Confirming the menace, in 2014 four countries from the EU declared ASF outbreaks [
6]. From 2014 to date, the presence of the disease and more outbreaks have been reported by the World Organization for Animal Health (OIE) in Estonia, Latvia, Lithuania, Poland, Russia, Belarus, Moldova and Ukraine while recently (June 2017), Czech Republic has reported its first case of ASFV [
7].
ASFV transmission is produced by direct contact between infected and susceptible animals and alternatively, following indirect transmission routes through contaminated pork, fomites, vehicles, people and ticks. The disease is characterized by an acute form that leads to fever, hemorrhages and several lesions causing death in a short period of time. High mortality leads to huge economical losses in affected areas [
7,
8]. Experimental infections occur with the development of clinical signs at 7–10 days post inoculation that can be considered late times of infection. At that time, histopathological evaluation shows typical alterations of the disease such as macrophage infiltrations and apoptotic cells together with a cytokine storm characteristic of the terminal phase in ASF infections according to previous studies [
9].
No commercial vaccine is available in spite of the efforts carried out to achieve it. Engineered live attenuated strains and cell culture adapted virus have been the two most common approaches used to obtain a candidate vaccine [
10,
11]. Cell culture adaptation leads to attenuation that generally ends rendering non-pathogenic viruses invalid for vaccine purposes. The genomic changes responsible of the phenotypic modification and the mechanisms involved are not well understood [
12].
MicroRNAs (miRNAs) are small RNA molecules that can regulate gene expression post-transcriptionally. miRNAs derive from the primary miRNA (pri-miRNA) [
13], which is transcribed in the nucleus and processed into a precursor (pre-miRNAs) by the endonuclease Drosha and DCGR8 [
14]. The pre-miRNA molecule is transported to the cytoplasm by exportin-5 [
15,
16]. Another endonuclease, Dicer [
17], processes the pre-miRNA to generate the double-stranded 18–23 nt miRNA duplex. The mature miRNA strand in which the first 2–8 nucleotides (seed region) are essential for recognizing the target mRNA, is incorporated into the RNA-induced silencing complex [
18,
19] in order to regulate gene expression by decreasing mRNA stability and/or block translation [
20].
miRNAs have an enormous capacity of regulation and have been shown to play key regulatory roles in many biological processes like development, differentiation, homeostasis, carcinogenesis, apoptosis, regulation of immunity as well as in viral infections [
21]. miRNAs have been identified in many species, from mammals to metazoans, funguses, plants and viruses. Studies on differential expression patterns of miRNAs have been carried out after virus infections, most of them in cell culture meanwhile a few of them have been performed in the natural host. In this proof-of-concept study, miRNA expression has been analyzed by using high throughput sequencing (HTS) in two tissues, spleen and submandibular lymph node (SLN), recovered at early times post-infection from pigs infected with two ASFV strains of different virulence and, on the other hand, the miRNA expression in these two tissues at different times post-infection in animals inoculated with the virulent strain E75 have been compared. The changes found in miRNA expression due to the virus infection could lead to a better understanding of the virus-host interactions and provide useful information to find new targets in order to control the infection. The information provided from this proof-of-concept study reinforces the application of this approach for future analysis using a higher number of animals and conditions.
Discussion
This study has been carried out in order to decipher the presence and the potential role of the miRNAs during in vivo ASFV infections and virus pathogenesis. The miRNA DE pattern has been studied for many viruses affecting pigs, like TGEV, PPV, Porcine cytomegalovirus, ADV, PCV2 and PRRSV [
25,
36‐
40], principally by using cell culture. Only a few studies have been carried out analyzing the expression pattern in the natural host. This in vivo expression pattern analysis could be considered a more reliable method to replicate the natural conditions in which infection and host-pathogen interactions take place [
25,
28,
41]. This proof-of-concept study has allowed the analysis of differences in miRNA expression at different times post-infection with a virulent strain and the expression pattern induced by virulent and attenuated strains at an early time post-infection. However, further functional analysis is needed to support and confirm these predictions. On the other hand, the use of a high number of animals could provide a more robust results. Nevertheless, we consider that the selected samples are representative of each group, taking into account that the well characterized virological, clinical and immunological parameters of each group were similar [
10].
As previously described [
10], animals inoculated with virulent strain E75 developed clinical signs compatible with acute ASF detectable from 4 dpi and reaching its peak at 7 dpi, while animals inoculated with the attenuated strain E75CV1 developed no clinical signs and no viral genome was detected at any time post-infection in the selected tissues.
When comparing different times post-infection with the E75-virulent strain, more than half of the spleen miRNAs expressed at higher levels (highest copy number), were DE (64.3%), when considering as DE those miRNAs with a FC > 5. Interestingly, two were up-regulated at 7 dpi (miR-451 and miR-145-5p), while the rest were down-regulated. In SLN, a minor proportion (24%) of the miRNAs were DE. This matches with the differential transcription and protein profile observed in gastrohepatical lymph node at these two times post-infection [
10].
Regarding the differential expression pattern between the virulent E75 and the attenuated E75CV1 at early times post-infection (3 dpi), again, a higher differential expression has been detected in spleen while a more conserved pattern was observed in SLN. In spleen, 31.8% (7 out of 22) of the most represented miRNAs (CN > 80) were DE. In SLN, the percentage decreased to 9.1% (3 out of 33). Although both spleen and lymph nodes are targets for virus replication, the different vascularization, anatomical organization and function might explain this differential regulation.
In SLN, the four DE miRNAs most represented at different times after the infection with the ASFV virulent strain, ssc-miR-126-3p, ssc-miR-126-5p, ssc-miR-23b and ssc-miR-30d were down-regulated at 7 dpi, coinciding with the massive detection of cytokines and other immune mediators, with a marked leucopenia, hemorrhages and with the death of the infected animals [
10]. The cytokine storm classically described during the acute phase of ASF might be explained at least partially by the marked down-regulation of these specific miRNAs, previously described as key regulators of genes involved in the regulation of the immune response [
42‐
44] including: hematopoietic cell lineage, complement and coagulation cascades, platelet activation, Toll-like receptor signaling pathway, NOD-like receptor signaling pathway, RIG-I-like receptor signaling pathway, cytosolic DNA-sensing pathway, natural killer cell mediated cytotoxicity, antigen processing and presentation, T cell receptor signaling pathway, B cell receptor signaling pathway, Fc epsilon RI signaling pathway, Fc gamma R-mediated phagocytosis, leukocyte transendothelial migration, intestinal immune network for IgA production and chemokine signaling pathway.
miR-451 was the most represented miRNA in spleen of the virulent infected animals and is also DE between 3 and 7 dpi. Thus, miR-451 is down-regulated at 3 dpi and up-regulated at 7 dpi. Interestingly, miR-451, was also DE at 3 dpi in spleen of E75CV1-infected pigs, on this occasion, showing a clear up-regulation. A similar expression profile has been described for some key mediators of the innate immune response [
10], albeit the target prediction analysis showed that this miRNA was potentially able to interact with 37 genes, no significant pathways related to immune response or pathways involved in ASFV-host interaction have been identified.
miR-145-5p was up-regulated at 7 dpi in the spleen and SLN of the animal infected with the virulent strain. As described for miR-451, no pathways related to immune response have been predicted for this miRNA, among the 293 target genes identified. Nevertheless, one pathway related to ASFV-cell interaction, regulation of autophagy through ATG12, has been found. The overexpression of miR-145-5p at the time of maximal virus production (7 dpi) might contribute to an inhibition of autophagy to facilitate ASFV replication and to avoid virus clearance, as has been postulated before [
45].
miR-92a was DE at different times in the spleens of animals infected with the virulent strain E75, with its expression being up-regulated at 3 dpi. When comparing the spleens of animals infected with the virulent or attenuated strains at 3 dpi, there was also DE, where the expression is up-regulated by the virulent strain. Target prediction showed 1192 regulated genes and the pathway analysis revealed that it was involved in five pathways related to immune response, B and T cell receptor signaling pathways, Fc gamma R-mediated phagocytosis, chemokine signaling pathway and leukocyte transendothelial migration. Other associated pathways are endocytosis and regulation of actin cytoskeleton. The participation of actin in the endocytosis process has been proposed as stimulator for the entry of large viruses, like vesicular stomatitis virus [
46] and also for ASFV [
47], therefore, we anticipate that miR-92a could be involved in the entry process of ASFV, corresponding with the peak of virus replication at 7 dpi.
Another miRNA showing a different expression pattern in spleen between 3 dpi and 7 dpi was miR-23a. This miRNA was down-regulated at 7 dpi. A large number of significant pathways were found associated with miR-23a target genes, some of them related to the immune response, such as RIG-I-like receptor signaling pathway leukocyte transendothelial migration, Fc gamma R-mediated phagocytosis, NOD-like receptor signaling pathway, chemokine signaling pathway, T cell receptor signaling pathway, hematopoietic cell lineage and Fc epsilon RI signaling pathway. On the other hand, pathways related to the entry of the virus, like endocytosis and regulation of actin cytoskeleton, as well as other related to the apoptotic machinery were found. Among the target genes predicted, Bcl2 and Caspase 3 were two of the most significant genes involved in apoptosis [
48]. Also, this miRNA regulates eIF2a which is involved in ER stress processes [
49]. Thereby, miR-23a can play a role in the apoptosis induced by ASFV and in the immune response against the virus. Compared with 3 dpi, miR-126-3p was down-regulated in SLN at 7 d after infection with the virulent E75 strain. Only 20 genes were identified as possibly regulated by miR-126-3p, and no related immunological pathways were detected.
miR-126-5p was down-regulated both in SLN and spleen at 7 dpi after infection with the virulent E75 strain, compared with 3 dpi, and could regulate genes related to the immune response and many aspects of the virus-host interaction like entry, apoptosis, regulation of autophagy, ER stress or chemokine receptors. This regulation could be in agreement with our observation about miR-126-5p down-regulation that paralleled the cytokine storm found during the last stages of ASF acute disease.
miR-125b was DE in spleen at 3 dpi between virulent and attenuated strains, being down-regulated by the virulent E75 strain. miR-125b was shown to interact with a large number of genes, thus, some pathways related to immune response have been identified for this miRNA. In addition, it has also been associated with pathways involved in the ASFV-cell interaction such as endocytosis, regulation of actin cytoskeleton and apoptosis [
45]. Target prediction analysis showed that miR-125b can interact with Bcl2. Down-regulation of Bcl2 by miR-125b might contribute to the inhibition of the early apoptosis of E75 infected macrophages thus favoring the success of its replication.
It is worth noting the differential regulation of miR-30e-5p and miR-125a, two miRNAs that follow opposite expression profiles that interestingly were conserved for both tissues studied. Thus, the expression of miR-30e-5p was up-regulated in SLN and spleen at 3 dpi infected with the virulent E75 strain, compared to the attenuated E75CV1, while miR-125a was down-regulated, perhaps playing opposite regulatory roles during the fine equilibrium that separate attenuation from virulence. The fact that all functions associated with miR-125a are shared by miR-30e-5p, allowed us to hypothesize about the relevant roles for pathways specifically regulated by overexpressing miR-30e-5p early after infection with virulent E75 ASFV strain, and contributing to the inhibition of key macrophage innate defensive mechanisms such as mTOR, toll-like receptors, RIG-I-like receptors, cell adhesion molecules, chemokine signaling pathways, Fc gamma R-mediated phagocytosis, leukocyte transendothelial migration or CAMs. Down-regulation of any of these pathways might contribute to successful evasion of the innate immune response.
miR-122 was not expressed in spleen at 3 dpi of animal infected with the virulent E75. Conversely, its expression was notably high at day 7pi. A small number of pathways related with immune response have been observed for this miRNA and it has only been associated with the endocytosis pathway.
Thus, from this analysis we can conclude that ASFV modifies miRNA expression patterns involved in the immune response and might contribute to the course of the disease during the infection.
From the gene network analysis, we found that some miRNAs like miR-451 and miR-145-5p, which are the most represented DE miRNAs in spleen, and are highly up-regulated at 7 dpi, are associated with the viral gene 1242 L. This gene, also regulated by miR-125a and miR-125b, is involved in RNA transcription and processing [
50]. ASFV genome transcription is carried out without using the host RNA polymerase II. The described gene function for this gene is RNA polymerase subunit 2 making it a candidate to explore as a target to regulate viral replication.
miR-125a additionally targets the viral gene P1192R, which codes for a type II DNA topoisomerase that is essential for viral replication and/or transcriptional events [
51]. In addition, this topoisomerase has been proposed as a potential target to control the disease by using poisons and inhibitors against this enzyme [
52]. Down-regulation of miR-125a might contribute to evade blockade of P1192R, thus contributing to the efficient replication of the E75 virulent virus.
Interestingly, the three miRNAs that interact with Bcl2 (miR-23a, miR-125a and miR-125b), do not theoretically regulate A179L, the homologous viral gene of the apoptosis inhibitor gene Bcl2 [
53]. In spite of the high degree of structural similarity to all Bcl-2 proteins [
54], the sequence divergence between both genes might explain this differential regulation and might contribute to the successful apoptosis inhibition of the infected macrophage until the virus cycle is finished.
miR-92a interacts with the three genes identified as being involved in ASFV entry, EEA1, EGFR and PIKFYVE [
55,
56] both in the virulent infection at 3 and 7 dpi and between attenuated and virulent at 3 dpi. Interestingly, this miRNA was found in much higher levels (thirteen times) in spleen of the animal inoculated with the virulent strain at 3 dpi compared to spleens of animals inoculated with the attenuated strain. This difference of expression could be involved in the differences in the dynamics of virus infection depending on its virulence and, is in accordance with the recent study where the diminution of PIKFYVE decreased the ASFV infectivity and viral production [
57].
Target prediction for miR-122 revealed that it interacts with 12 different viral genes classified in multigene family, replication, genes with unknown function and unknown genes. This miRNA interacts with MGF 360-16R and it has recently been described that this viral multigene family component modulates host innate responses by determining the tropism, virulence and suppression of type I IFN response [
57]. In addition, miR-122, together with miR-126-5p, target ATG6, which is activated by ASFV. The virus uses the ER as a site of replication and this process of activation of ATF6 can trigger ER stress and the unfolded protein response (UPR) of the host cell [
58]. miR-122 is the DE miRNA with the highest number of viral target genes and none of these is regulated by other miRNAs. On the other hand, miR-122 is not express in spleen from an infected animal with the virulent E75 strain at 3 dpi, while at 7 dpi its expression was notably increased. In addition, it is well known that miR-122 plays a key role in Hepatitis C virus infection [
59]. Accordingly, this miRNA could be involved in the regulation of the “success of the ASFV infection”.
To our knowledge, this is the first time that a deep sequencing approach has been used to study miRNA gene expression in pigs infected with ASFV. However, a larger number of animals as well as functional analysis would be necessary to support the results obtained in this study. These investigations could confer a more accurate vision of the relation between miRNAs with porcine target genes and ASFV genes to help decipher the role of miRNAs in ASFV infection.