4 Discussion

The porcine T-cell response against ASFV infection is largely not understood. Moreover, knowledge about the responses of susceptibleSuidae subspecies outside of Africa, domestic pigs (Sus scrofa domesticus ) and wild boar (Sus scrofa scrofa ), is scarce. Similar to our first approach to analyze these differences upon infection with the highly virulent ASFV strain “Armenia08” , we used a multicolor flow cytometry platform to investigate T-cell responses in domestic pigs and wild boar after infection with the moderately virulent ASFV isolate, “Estonia2014”.
The most fundamental characteristics of an ongoing immune response are local or systemic alterations in the composition of leukocytes in tissues affected by disease. As one of the few known parameters of protection against ASFV infection, the importance of (cytotoxic) CD8α+ lymphocytes has been demonstrated, although to a limited extent only. It has been shown that isolated porcine PBMC, afterin vivo priming with virulent ASFV, were able to specifically lyse ASFV-infected cells in vitro . This was a first indication of virus-specific cytotoxicity. However, PBMCs were not differentiated and protection was not investigated. SLA I- and CD8-dependent lysis of ASFV-infected target cells by PBMCs from ASFV-immune minipigs and specific lysis of isolated CD8α+ but not CD4+ T cells has also been shown. In another study, antibody-dependent depletion of CD8α+ cells in vivo in domestic pigs primed with the low virulent ASFV strain “OUR/T88/3” resulted in loss of protection after homologous challenge with the virulent ASFV strain “OUR/T88/1” . We could show that the CD8α response in wild boar and domestic pigs during infection with moderately virulent ASFV is based primarily on increases of CD4/CD8α+ and to a lesser extent of CD4+/ CD8α+ (DP) T cells. Interestingly, in our previous study with the highly virulent ASFV “Armenia08”, the CD8α response was primarily based on DP T cells . CD4+/ CD8α+ DP T cells are often described to possess memory functions . ASFV-specific memory responses can be excluded during experimental ASFV infection with naïve pigs. However, ASFV-specific responses might play a role in field infections since serological evidence for previous ASFV infections was found in hunted animals . On the other hand, DP memory T cells in the spleen might be activated independent of their cognate antigen by release of IL‑15 or IL‑18 . Porcine DP T cells are also described to exhibit effector functions, like cytotoxic responses or cytokine production . However, in contrast to CD8αα+ or CD8αβ+ αβ T cells, we only found proliferating DP T cells in the spleen but not in other tissues. Moreover, the pronounced loss of perforin in other cytotoxic T-cell populations was not found in DP T cells. This might indicate that DP T cells are orchestrators of systemic responses but do not take part in antiviral responses in disease-affected tissue during moderately virulent ASFV infection.
Pigs belong to a group of mammals with relatively high frequencies of γδ T cells. They can exert effector functions like cytokine production and cytotoxicity, and are even able to present antigens to other lymphocytes . The main effector population is characterized as CD2+/CD8α+ . In the present study, we found pronounced increases of effector γδ T-cell frequencies in spleen, lung, and liver of infected wild boar but not domestic pigs. Moreover, we detected T‑bet-dependent activation of γδ T cells in wild boar only. This is in line with our previous findings during highly virulent ASFV infection, where wild boar were found to have a considerably stronger bias for γδ T-cell responses . This indicates a profound dissimilarity in the antiviral responses of both subspecies and might give an explanation for their different disease severity and survival. Of note, this is in contrast to previous findings, where higher frequencies of circulating γδ T cells correlated with increased survival of infection with moderately virulent ASFV, independent of age or virus dose . There might be some explanations for this discrepancy. First, we detected increased lethality and heightened frequencies of effector γδ T cells in wild boar, while the aforementioned study used domestic pigs. While we were also not able to detect changes of γδ T-cell frequencies in domestic pigs, this might be caused by different ASFV strains used. Finally, the numbers of survivors that showed correlations with γδ T-cell levels in the study by Post et al. were relatively small. This underlines the need for in-depth research not only during ASFV infection in general but also for the differences between wild boar and domestic pigs. Moreover, since infection of professional antigen-presenting cells alter their function , it would be of interest to investigate whether γδ T cells take part in the antigen presentation during ASFV infection.
We found regulatory T cells (Tregs) in both subspecies but higher frequencies in wild boar. The role of Tregs during ASFV infection is largely unexplored. However, previous studies showed that Tregs might present a way of viral immune evasion because they were able to inhibit specifically antiviral responses . Higher percentages of Tregs in wild boar might therefore be an explanation for their higher disease burden in this study and lethality previously . In a parallel study by Sehlet al. using histopathology from tissues of trial 2 of this study, domestic pigs but not wild boar showed lymphohistiocytic interstitial pneumonia even 10 dpi . This might be a sign for prolonged pro-inflammatory responses in domestic pigs in contrast to wild boar. This is in line with our findings of higher Treg frequencies in wild boar. Moreover, this indicates that pro-inflammatory responses are able to counteract ASFV infection, as long as they are not downregulated too early.
A porcine T-cell population that is still not well understood is invariant Natural Killer T (iNKT) cells. We could previously show that iNKT-cell frequencies significantly increased in some tissues during infection with highly virulent ASFV . Although at the time we were not able to investigate effector mechanisms or surface markers on iNKT cellsex vivo , our study provided first evidence that iNKT cells participate in the antiviral response during ASFV infection. The fact, that we were unable to find changes in the general iNKT-cell frequency might be explained by the less virulent ASFV strain in this study. Nevertheless, activation of iNKT cells was shown by significantly increased frequencies of ICOS+ iNKT cells. ICOS is an essential protein for iNKT-cell activation, homeostasis, and survival . Some studies correlated ICOS expression on iNKT cells with pro-inflammatory Th1 responses , while others described it as a marker of effector iNKT cells . Increased expression of CD8α and CD4, as previously established markers of maturation of porcine iNKT cells , underlines these findings. Interestingly, a role for NKT cells has previously been suggested, as CD3+/CD4/CD8α+/CD5±/CD6/CD11b+/CD16+cells expanded after in vitro stimulation of porcine PBMC with ASFV . However, even though we and others could show that the phenotype of iNKT cells differs from that finding , the significant alterations in iNKT-cell frequency in our first study and our findings in this study support the notion that iNKT cells take part in the antiviral response against ASFV.
Besides analysis of the cellular composition of leukocytes in the investigated tissues, effector functions are also pivotal to understand the underlining immune mechanisms. Perforin is one of the major lytic molecules used by cytotoxic lymphocytes to kill target cells . Instead of direct cell lysis, cytotoxic lymphocytes can also induce apoptosis in their target cells by death receptor-mediated pathways using Fas ligand (FasL) or TRAIL . The significant and partially complete loss of perforin 4 to 5 dpi in this study resembled the observed loss of perforin we found during infection with the highly virulent ASFV strain “Armenia08” . However, the perforin decrease was more pronounced during infection with highly virulent ASFV, especially on a systemic level, i.e. in the spleen. There are various explanations for the perforin loss observed in both studies. Perforin-mediated killing is thought to be the major pathway in the early cytotoxic response but can be switched to Fas/FasL-mediated apoptosis induction with a complete loss of perforin expression on RNA and protein level during the course of infection . An effector molecule switch might be an explanation for the observations in our study, however, this is not directly detectable because antibodies against porcine FasL are still missing. Still, there are some lines of evidence suggesting this might be the case. Expression of viral homologues of the mammalian anti-apoptotic protein Bcl-2 and also prevention of apoptosis by these viral homologues has been shown for ASFV strains . Bcl-2 is also known to preferentially inhibit perforin-mediated apoptosis but less Fas/FasL-mediated apoptosis, depending on the cellular target . A switch from perforin-mediated to Fas/FasL-mediated cytotoxic responses might therefore be beneficial and protective and would be in line with the lower disease severity and heightened survival of domestic pigs. Wild boar, in contrast, had higher levels of perforin+ lymphocytes on average, indicating that they might not have switched the cytotoxic pathway or at least not to the extent that domestic pigs did. Given that inhibition of perforin has been shown to protect from tissue damage during viral hepatitis , this might also be an explanation for the more severe inflammation and tissue degradation in the liver of infected wild boar . On the other hand, it cannot be excluded that we missed newly synthesized and immediately secreted perforin, which is not detected by antibody clone dG9 used in this study . However, missing detection due to immediate secretion would still hint to a strong cytotoxic response. In this case, the cytotoxic response would have been higher in domestic pigs because the perforin loss was detected earlier and more pronounced than in wild boar. Therefore, it can be hypothesized that the more pronounced and earlier response in domestic pigs was beneficial and protective at least during infection with moderately virulent ASFV. Wild boar, in contrast, might not have been able to counter the infection because of their impaired response, eventually leading to death as observed in previous studies . Which explanation holds true, switch of cytotoxic pathways or earlier and stronger cytotoxic response, has to be investigated in future studies. It still has to be kept in mind that cytotoxic responses are not only beneficial but might also cause immunopathology and thus, contribute to disease burden .
In summary, we described the first comparative analysis of immune responses of wild boar and domestic pigs during moderately virulent ASFV infection. While more severe in wild boar, domestic pigs showed signs of moderate disease and recovery of all animals. Overall, we found comparable courses of immunity in both subspecies. Both developed a heavily CD8α+-biased response with proliferation of CD8αα+ and CD8αβ+ cells. However, although their αβ T-cell responses were largely similar, wild boar developed a more pronounced effector γδ T-cell response. We also found only small signs of T-bet-dependent activation predominately in lungs and liver of wild boar but none in domestic pigs. Moreover, we found a distinct loss of perforin in cytotoxic T cells in domestic pigs and to a lesser extent also in wild boar, similar to previous results during infection with highly virulent ASFV “Armenia08”. Tregs appeared in higher levels in wild boar. Finally, we were able show the first description of functional iNKT-cell responses during ASFV infection. With this data, our study paves the way for further in-depth analyses of porcine immunity towards ASFV.