Introduction
African swine fever (ASF) is a hemorrhagic and fatal infectious disease that causes rapid death in domestic pigs (
Sus scrofa), originally described for the first time in Kenya [
1]. It has lately spread across the globe, causing huge economic losses. ASF is caused by the African swine fever virus (ASFV), a large enveloped and structurally complex double-stranded DNA virus. The virus has a linear genome with sizes between 170 and 193 kbp, which encodes 150–167 open reading frames dependent on the strain [
2]. ASF is a major constraint for pig farming in Sub-Saharan Africa [
3].
Previous studies were conducted in different parts of DRC, some have identified the circulation of genotypes I, IX and XIV the causative agent of ASF outbreaks in in the Western, Central and Northern parts [
4]. Other recent studies conducted in the South Kivu province east of DRC showed genotypes IX circulating in asymptomatic pigs while genotype X was found in symptomatic domestic pigs in outbreaks in the region [
5,
6].
African domestic pigs (African indigenous pigs) are more susceptible to ASF, with mortality rates of up to 100% in infected herds compared to wild pigs such as warthogs and bushpigs [
7,
8]. Ticks of the species
Ornithodoros moubata and
Ornithodoros erraticus can transmit the virus [
9]. The difference in resistance to ASFV between domestic and wild pigs is attributed to genetic variation in the hosts and has been suggested to be due to differences in the ability of the virus to modulate the immune responses in these different hosts [
10].
Previous studies have shown evidence that African indigenous domestic pigs in most African countries are less susceptible to ASFV infection than European pigs [
11,
12]. The wild pigs and the African domestic pigs have been exposed to ASFV for many more generations than the European pigs and have had the time to adjust to the surrounding pathogens simply by genetic evolution, which may explain the resistance to ASFV
African swine fever virus primarily infects macrophages, which are major targets for in vivo viral replication [
13]. When infected and activated by ASFV, macrophages drive the immune response by secreting a wide range of mediators, including pro-inflammatory cytokines (e.g., IFN type I and TNF-α) as well as those that facilitate the development of specific immune mechanisms (IL-10, IL-12) through activation of both the Th1 and Th2 responses [
14]. An in vitro study comparing growth curves of ASFV in macrophages from wild pigs and domestic pigs showed a similar capacity of the macrophages in supporting ASFV growth [
7], indicating that other mechanisms may explain differences in virus susceptibility than differences in infection and propagation. Various studies have demonstrated an association between ASFV pathology and overexpression of cytokines such as IFN alpha (IFN-α) and tumor necrosis factor-alpha (TNF-α) in domestic pigs [
15,
16]. Another study has reported that ASFV virulent such as Armenia/07 blocks the synthesis of IFN-β, a key mediator between the innate and adaptive immune response [
17]. Additionally, the nuclear factor kappa-B (NF-κB) transcription factor plays an important role by controlling the transcription of these cytokines (IFN-α and TNF-α) [
18]. The mammalian NF-κB transcription factor comprises RelA (p65), RelB, p50/NF-κB1, p52/NF- κB2, and c-Rel (Supplementary Fig. 1). The heterodimer of RelA (p65) and p50 is the most commonly found heterodimer complex among NF-κB and is the functional component participating in nuclear translocation and activation of NF-κB [
19].
Like other NF-κB sub-units, the
RelA gene product is composed of an N-terminal REL-homology domain (RHD) responsible for DNA binding, dimerization, and NF-κB/REL inhibitor interaction, and a C-terminal transactivation domain (TAD) (Supplementary Fig.
1), which interacts with the basal transcription complex, involving several coactivators of transcription, such as TATA-binding protein (TBP), Transcription Factor IID (TFIIB) and the cAMP-response element binding protein (CREB-CBP) [
20]. The NF-κB gene encodes one of the subunits of the NF-κB heterodimer, which is involved in a variety of intracellular pathways, among others; it invokes overreaction of the host immune system with devastating effects [
19]. Polymorphic variants found in the warthog
RelA gene have been suggested as a potential genetic factor to explain the surviving phenotype upon infection with ASFV [
10], but this was never firmly established. In line with this, variation in the porcine
RelA gene may explain why some domestic pigs are more resilient to ASFV infection than others. Furthermore, several proteins expressed by ASFV have been experimentally validated for their ability to suppress the immune system in vitro by reducing the interferon response and the NF-κB activation [
2,
20]. This creates a favorable environment for viral replication [
21].
The availability of data on immunogenetic features underlying the immune responses in domestic pigs is scarce, and most studies are limited to analyses in domestic pig cells. Among the ASFV genes, A238L substitutes for NF- κB Inhibition Activator (NF-κBIA) by binding to the p65 (
RelA) sub-unit of NF-κB preventing the ability of NF-κB to be activated [
22,
23]. Additionally, the viral protein A238L was shown to act as inhibitor for the activation of the transcriptional co-activator p300/cbp1 [
18,
24]. Similarly, previous studies showed that A238L gene binds and inhibits calcineurin phosphatase and activation of NFAT transcription factor [
25,
26]. Some other recent papers have investigated other ASFV inhibitors of NF-kB including pD345L and I10L by targeting IKK kinase activity [
27,
28]. In addition, the ASFV inhibitor of apoptosis (IAP) has been shown to activate the transcription factor NF-κB [
29].
The present study aimed to identify the polymorphisms in the RelA gene in ASFV-surviving (non-symptomatic but infected pigs) versus susceptible (symptomatic and infected) domestic pigs from the field and to assess the correlation of RelA mRNA expression with cytokine levels. It is envisaged that the findings of this study will promote a better understanding of genetic factors underlying ASF susceptibility.
Discussion
African swine fever virus (ASFV) is a severe hemorrhagic disease associated with huge economic losses in the pig industry worldwide. There is evidence that indigenous pigs in most African countries are asymptomatic carriers of infection from which they rapidly recover [
11,
12]. This study utilized a candidate gene sequencing approach to investigate the genetic basis of host immune response to ASFV infection in pigs that had survived ASFV infection and appeared healthy versus pigs that were symptomatic from ASFV infection. The
RelA gene (expressing p65 protein), one of the key molecules in the NF-kB and nuclear factor of activated T cells (NFAT) host signaling pathways leading to the production of inflammatory cytokines, was sequenced to analyze associations with resistance status in the pigs. In addition, the impact of amino acid substitutions, due to nonsynonymous SNPs, on the stability of the
RelA protein was assessed, as was the correlation of the
RelA expression with cytokine levels in the ASFV non-infected and surviving pigs.
In total, 28 sequence differences within the
RelA gene were found at the genomic level between healthy and surviving domestic pigs. Out of the 28 SNPs identified, 16 were nonsynonymous, resulting in codon change, with the majority being predicted to alter protein function and all of them located in exon 10 within the acidic transactivation domain (TAD2) region, located from the residues 431–554 in porcine p65 protein [
39]. In a previous study [
40], more than 30 potential polymorphisms were reported in several parts of the porcine
RelA (p65) gene sequence from China compared with other mammalian species, where polymorphism sites were identified in both RHD and TAD. Among the polymorphisms from this study, two substitution sites, i.e., P374-S and G489-A in surviving pigs corresponded to SNPs in the
RelA gene at the same positions (P374 and A489), identified in domestic pigs in the study conducted in 2011 [
40]. Additionally, one substitution, S448-T corresponded to an SNP identified in a previous study conducted by Palgrave et al. [
10], which was predicted to change the protein structure by conferring hydrophobicity in the domestic pig sequence compared to the warthog. Some amino acid residues, i.e., serine 536 and serine 276 (in TAD), have been related to known functions of
RelA, such as phosphorylation [
41]. Comparative analysis of the
RelA homology domain (RHL) and the nuclear localization signal in RelA showed complete sequence similarity in symptomatic, surviving, and non-infected (healthy) pigs (data not shown). In the present study, one of the identified nonsynonymous amino acid substitutions was (P374S), which lies in the TAD1 region of p65, while the other 15 were located in the TAD2. In a previous study aiming at revealing signatures for resistance to ASF, genetic variations were identified in both TAD1 and TAD2 of the
RelA subunit of NF-kB of the domestic pig compared with warthog [
10]. TAD1 and TAD2 are involved mainly in activating the transcription of p65-regulated genes, and it operates by participating in a series of protein–protein interactions with various transcriptional regulatory proteins [
39]. Based on this, variations within this region could play a significant role in the pathogenicity of ASFV.
Of the 16 nonsynonymous SNPs found, one (serine at position 448) was predicted to be a phosphorylation site in the surviving group and is located within the TAD2. This contrasts with the previous study that has identified a phosphorylation site at serine-531 within the TAD1 of the domestic pig
RelA gene [
10]. The threonine at position 448 in the symptomatic domestic pig from this study is substituted by a serine residue in the surviving pigs, whereas the threonine was found to be substituted for an alanine residue in the warthog [
10]. The limited availability of porcine
RelA sequence data, however, makes it difficult to make a conclusive comparative analysis of this site. Since up to one-third of eukaryotic proteome functions are controlled by phosphorylation [
42], the present polymorphism at position 448 could play a prominent role in defining the host response to ASFV infection in domestic pigs. This is supported by a previous finding showing that the phosphorylation of RelA induces a structural change, which impacts its ubiquitination and stability, as well as protein-protein interactions [
43].
The damaging/deleterious prediction scores by a combination of sequence and structural homology-based tools, SIFT and PolyPhen-2, revealed respectively 10/16 (62.5%) and 13/16 (81.2%) of nsSNPs as damaging/deleterious. The results of MutPred and PredictSNP predicted 5 high-confidence nsSNPs, i.e. P374-S, T448-S, P462-R, L495-E, and P449-Q, which are associated with the symptomatic phenotype, as damaging. However, the MUPro and I-Mutant3 predictions servers demonstrated seven high-confidence nsSNPs that affect protein stability, and these SNPs were predicted to be damaging/deleterious in both SIFT and Polyphen-2. As these SNPs all occurred in the same sequence resulting in amino acid change, this will likely influence the function of RelA, and as such lead to a difference in RelA function in surviving versus symptomatic pigs.
The qPCR and ELISA showed a significant increase of IFN-α transcripts and cytokine expression levels in ASFV- surviving infected pigs compared to symptomatic and healthy pigs. The difference between susceptible and surviving pigs could be attributed to differences in the host responses to ASFV infection. The stimulated Toll-Like Receptors (TLRs), after infection by a pathogen, induce the activation of signal transduction cascades, which oblige the nuclear factor-κB (NF-κB) to translocate to the nucleus [
44], followed by activation of interferon regulatory factors 3/7 (IRF3/7) or activator protein-1 (AP-1), which cooperate to initiate transcription of different cytokines such as alpha/beta interferon (IFN-α/β) to counteract infection [
45,
46]. The pathways leading to Type 1 interferons and proinflammatory cytokines such as TNF alpha, IL6, and IL12 have traditionally been viewed as relatively discrete pathways. However, it has become evident that there is considerable cross-talk between the NF-kB and the IRF pathways [
47,
48]. Previous investigation on NF-κB-deficient cells has revealed that type I interferon (IFN) response relies mainly on concurrent NF-κB activation [
49]. In addition, an experimental assay has demonstrated that in the absence of NF-κB, the rapid expression of IFNβ is blunted, reducing the propagation of anti-viral signals in the mucosal surface [
50]. Moreover, NF-κB also controls the expression of the downstream IFN auto-amplification loop through STAT1, IRF-1, − 5, and − 7 transcription factors. This points to NF-kB as a central regulator of the combined pro-inflammatory response. Therefore, a distortion of the activity of NF-kB due to SNPs in
RelA may lead to a lower response of type I interferons, as we observed in the susceptible group of pigs. In addition to this, we also observed a higher
RelA expression in the surviving pigs, which can also contribute to the findings of the increased systemic IFN levels seen in these pigs.
Interestingly, there was an increase in both the gene expression and protein levels of IL-10 in the surviving pigs. IL10 is generally an anti-inflamatory cytokine that works to dampen the immune response and avoid tissue damage after initially high levels of proinflammatory cytokines such as TNF-alpha, IL6 and IL12 [
15]. It could be a vital element in coping with ASF infection, to dampen the immune system and thereby damage to the body [
51]. In contrast,
TNF-α gene expression and the corresponding protein levels were moderate in both surviving and non-infected pigs, with a tendency to increased levels in the surviving pigs. This is likely due to the timing of samples as TNF-alpha is one of the very first pro-inflammatory cytokines to appear after infection, but it has also been found that TNF-α mRNA and other pro-inflammatory cytokines were inhibited by ASFV infection, both in vitro and in vivo in wild boars [
52,
53]. Such inhibition could be due to anti-inflammatory signals by, e.g., IL10, as it has previously been shown that IL10 directly inhibits NF-κB activity and, thereby, the expression of pro-inflammatory cytokines [
54].
The association analysis revealed a positive correlation between porcine
RelA (p65) gene mRNA expression and the mRNA expression level of IFN-α and IL-10 cytokines. This supports the important role of NF-kB in the interferon pathway and suggests substantial crosstalk between the pathways. The positive correlation between RelA and IL10 may at first glance appear strange, as IL10 suppresses the NF-kB activity, but IL10 exerts its inhibition in an RelA-independent manner, acting at least partly through inhibition of translocation of RelA to the nucleus but allowing translocation of the p50 homodimer, which has no transactivating domain. The homodimer binds to the DNA and inhibits the binding of functional NF-kB and the transcription of pro-inflammatory cytokines [
54]. Therefore, the positive correlation between NF-kB and IL10 may be initiated by higher activity of NF-KB, leading to increased pro-inflammatory cytokines, then, in turn, an increased IL10 production. The positive correlations between
RelA mRNA and IFN-α and IL-10 suggest a higher overall activity of the
RelA subunit of NF-kB in surviving pigs compared to symptomatic pigs, which may be due to the polymorphisms found in the surviving version of the
RelA gene. The SNPs found in the surviving pigs were all in exon 10, the transactivating domain (TAD2), except one (S9P), which was found in the region between TAD2 and nuclear localization signal (NLS). The TAD2 polymorphisms could affect the efficiency of the viral NF-kBinhibitors, such as the ASFV A238L protein known to have 40% homology to porcine NF-kB inhibitor IkBα, and which binds directly to the
RelA (p65) subunit of NF-kB to inactivate NF-kB [
52]. The A238L protein has also been reported to act as a potential immunosuppressant by inhibition of transcriptional activation from the TNF-α promoter through a mechanism that involves the CREB-binding protein (CBP) transcriptional co-activators [
25,
52].
A previous study found lower polymorphisms in the
RelA subunit of NF-kB from warthogs, which are surviving to ASFV infection [
10] than we found in our pig study. They speculated that the three SNP’ found in the warthog gene (compared to pigs) could be responsible for the surviving status seen in warthogs. A transgenic pig was made using the warthog
RelA gene, but it was found that the pigs were not surviving to ASF. However, a delay in the onset of clinical signs and less viral DNA in blood samples and nasal secretions was observed in some animals [
55]. However, the mechanism behind tolerance in domestic pigs may not be identical to what is found in warthogs.
Nevertheless, it is striking that one distinct amino acid sequence of RelA is present in the surviving pigs in this study, and the two variant RelA proteins seem to be stable versions of the protein as there were not found any other variants at the amino acid level but several variants at the nucleotide level. An effect of the Sanger sequencing in the present study is that pigs that are heterozygous for the RelA gene will have been discarded in the process due to non-conclusive sequences, leaving only the homozygous pigs left for the association, but this is likely to contribute to the highly conclusive finding. We hypothesize that these SNPs (found in the surviving animals) lead to an escape from ASFV suppression, e.g., by the virus gene A238L, which then leads to an increased NF-kB activity, higher cytokine profiles, and resistance of the pigs. One limitation of the study is the lack of knowledge of the ASFV strain(s) in sampling areas. There could have been more than one strain/genotype circulating, and there could have been differences in the virulence of these strains, explaining why some animals survive the infection and others do not. Still, this would not explain the association between the RelA sequence and phenotype found in this study. Further investigation particularly an in silico study is warranted in order to identify which domains are modified and the protein function by the observed amino acid modifications on RelA gene.