1. Introduction
African swine fever (ASF) is currently the main problem facing the pig industry worldwide [
1]. ASF is a devastating haemorrhagic infectious disease caused by African swine fever virus (ASFV), a large and complex enveloped double-stranded DNA virus of the genus Asfirvirus (family Asfarviridae). The disease was endemic in most sub-Saharan African countries and in the island of Sardinia (Italy) until 2007, when highly virulent isolates belonging to genotype II appeared in Eastern Europe. Since then, ASF has become endemic in many European and Asian countries spreading without control into Eastern Europe, China (since 2018) and most Southeast Asian countries, causing a huge economic impact. Outbreaks of uncertain origin have also occurred since 2021 in some Caribbean countries (Dominican Republic and Haiti), posing a threat to the nearby North American pork industry [
2]. The disease, which affects domestic and wild suids of all breeds and ages, presents a variable lethality depending on the virulence of the isolate and the immune status of the infected animals [
3,
4,
5]. Thus, highly virulent isolates often have a lethality close to 100% in naïve animals, which usually die within two weeks of infection [
6,
7].
Vaccines are essential to control viral diseases. The lack of vaccines against ASFV can largely be attributed to gaps in knowledge of the strategies used by the virus to evade host innate and adaptive immunity and the functions of virus proteins responsible for inducing protective immune responses [
8]. Evidence suggests the importance of both arms of adaptive immunity for protection. However, the immunological correlates of protection are not yet understood. On the one hand, antibodies have been suggested to be an essential component of protective immunity against virulent ASFV, although alone they are not sufficient to induce protection [
9]. Conversely, evidence suggests that cellular immune response, even in the absence of specific antibodies against ASFV, may be necessary for effective protection, highlighting the role of CD8+ T lymphocytes [
10,
11,
12,
13,
14,
15,
16].
Vaccines against ASFV based on naturally attenuated live virus (LAV) induce robust immune protection, stimulating both innate and adaptive (cellular and humoral) immunity. LAV immunizations, although characterized by an absence or mild presence of both clinical signs and viremia levels, have traditionally raised safety concerns, making their commercialization unlikely [
8,
17,
18]. Selective deletion of genes involved in virus attenuation and/or induction of protection has also been used as a strategy to produce safe and effective live attenuated vaccines and to differentiate between vaccinated and infected animals (DIVA vaccines). In both cases, viruses attenuated naturally or by gene deletion usually render recovered animals protected from subsequent infections with related viruses, although this does not usually guarantee protection against divergent viruses [
18,
19,
20,
21]. Previous studies have demonstrated that immunization of domestic pigs (DP) and wild boar with a naturally attenuated non-haemadsorbing (non-HAD) genotype II ASFV isolate, obtained from a wild boar hunted in Latvia in 2017 (Lv17/WB/Rie1 strain), conferred high levels of protection against challenges with a virulent ASFV genotype II isolate (Armenia/07) [
22,
23,
24]. It is known that immunizations with this strain induce some mild clinical signs and transient viremia. To improve its safety, the deletion of virulence-associated genes has recently been carried out. Although some deletion mutants generated showed a slight reduction in pathogenicity and lethality during the "in vivo" evaluations in DP, did not show a significant reduction in side effects with respect to the parental virus [
25,
26]. These results highlight the potential use of Lv17/WB/Rie1 strain not just as a vaccine prototype, but also as an excellent platform to elucidate protective mechanisms in immunized animals with especial attention to the role of the cellular immune response. With the objective of characterizing the immunological mechanisms of protection against ASFV, and with special attention to the role played by different subsets of T lymphocytes, we have systematically analyzed these mechanisms in DP immunized with the genotype II Lv17/WB/Rie1 strain and that were protected against virulent challenge with Armenia/07 (Arm07). In addition, several immunomodulatory cytokines were studied to complete the understanding of the protective mechanisms elicited by this vaccine prototype.
4. Discussion
In this study, intradermal immunization of domestic pigs, using a prime/boost regime with the attenuated Lv17/WB/Rie1 strain conferred high levels of protection against virulent challenge with Arm07. Unlike intramuscular inoculation, which may lead to adverse reactions, the intradermal approach not only conferred robust protection but also reduced adverse reactions typically associated with live vaccine candidates in domestic pigs [
24,
25]. Although one of the five immunized pigs developed viraemia and moderate clinical signs after the first intradermal inoculation, leading to its euthanasia prior to the boost, the remaining four animals were fully protected. These four pigs developed robust immunity that was evident after the prime immunization, characterized by the induction of ASFV-specific antibodies and virus-specific IFNγ-T cell responses. This pattern of immune response aligns with previous findings, suggesting that the combination of antibody induction with a potent virus-specific cellular response improves protection against subsequent infections [
39,
40,
41]. Hence, this experimental model provides an excellent opportunity to study in detail the mechanisms defining the protective immune response induced by attenuated strains against virulent ASFV challenge in natural hosts.
The mild and transient increase in body temperature observed after prime immunization with the attenuated strain, coupled with low and transient viremia, is consistent with previous observations in both domestic pigs [
4,
24,
25] and wild boar [
22,
23]. Unlike the findings in wild boar immunized orally with repeated doses of the Lv17/WB/Rie1 strain [
23], the intradermal route did not induce notable viremia or clinical signs after the booster dose. These differences could be attributed to the different routes of administration used (intradermal vs. oral). Previous studies have proven that intradermal vaccination is apparently more effective than other routes of vaccination, such as intramuscular or subcutaneous, even if the latter are administered in repeated doses [
42,
43,
44]. Although vaccination by these routes may be equally immunogenic, the dose may be reduced when the vaccine is administered intradermally [
45]. Dermis is rich in resident dendritic cells (DC), especially Langerhans cells and dermal DC, and although plasmacytoid DC (pDC) are rare in skin, they quickly infiltrate this organ during inflammation. It is known that pDC become stimulated by the virus to produce type I IFN during acute ASFV infections (reviewed in [
46]), cytokine that seems to be crucial in innate protection against some attenuated ASFV strains [
21,
47]. Therefore, as has been suggested in other immunization studies carried out against haemorrhagic viruses such as Ebola [
48], intradermal immunization would trigger enhanced adaptive immune responses by recruiting more dermal DC subsets to the inoculation site, including pDC. It would increase the chances of success against subsequent infections even using reduced or single doses during immunizations, hypothesis that requires further studies with ASFV in domestic pigs and wild boar.
An increase in circulating levels of all cytokines studied was observed after challenge in the non-immunized control pigs. This event, also known as "cytokine storm," is directly associated with severe disease caused by virulent strains when animals lack protection [
49,
50]. Unlike controls after challenge, where IL-8 levels increased significantly in parallel with the development of clinical signs and viremia, IL-8 levels in the immunized animals remained relatively stable after the initial peak following the first immunization (between 3-10 dpi), with minimal fluctuations observed after booster and challenge. Although studies on IL-8 are contradictory, and often fail to observe a significant modulation of this chemokine following infection with virulent or attenuated ASFV isolates, our findings align with previous studies indicating an increase in circulating levels of IL-8 following infection with the virulent Arm07 strain [
51]. The early and controlled increase of IL-8 in immunized pigs suggests an "in vivo" modulation of circulating IL-8 after immunization with the attenuated Lv17/WB/Rie1 strain, which appears to be correlated with protection against ASFV.
Similarly, IL-10 levels remained minimally affected in the immunized/challenged group, except for some animals that exhibited a mild increase shortly after the first immunization (3 dpi). In contrast, in the control group, a substantial increase in IL-10 was observed consistent with the rise in IL-8 after challenge, indicating, in line with other studies (reviewed in [
52]), a direct correlation between elevated levels of both cytokines and an exacerbated and/or uncontrolled inflammatory immune response after Arm07 challenge.
The controlled and early increases of IL-10, cytokine with a potent and broad anti-inflammatory activity, have also been described in wild boar experimentally immunized with Lv17/WB/Rie1 [
53], as well as in domestic pigs experimentally inoculated with some attenuated ASFV vaccine strains [
19,
21,
54]. Thus, our results are consistent with the premise that a controlled and early increase in IL-10 may contribute to controlling viral replication and dampening the exacerbated inflammatory response that often leads to fatal outcomes during ASFV infections. The simultaneous decrease in circulating levels of TNFα and IFNα, two cytokines with important pro-inflammatory functions, which are normally associated with tissue damage and the appearance of clinical signs [
49,
52,
55], showed evidence of a restrained inflammatory response in the immunized animals. In summary, the transient mild to moderate increase in IL-8 and IL-10 observed in pigs immunized with Lv17/WB/Rie1 seems to be directly correlated with survival, by contributing to the control of virus spread and the inflammatory response. Thus, moderate and controlled serum levels of these two cytokines, but especially IL-8, could be good markers of protection as well as of favorable evolution of ASFV infection.
IFNγ plays a pivotal role in inducing and modulating immune responses but it remains unclear whether protection against ASFV is linked to IFNγ production. A study conducted in domestic pigs reported a significant rise in circulating IFNγ, as well as TNFα, seven days after immunization with the attenuated vaccine candidate HLJ/18-7GD [
16]. Similarly, significant increases were observed at 28 dpi in wild boar orally inoculated with Lv17/WB/Rie1 [
53]. However, other studies did not find remarkable changes in circulating levels of IFNγ following immunization of pigs with different attenuated ASFV isolates [
21,
56,
57], including the Lv17/WB/Rie1/d110-11L and Lv17/WB/Rie1 ASFV strains inoculated intramuscularly [
25]. In our study, intradermal immunization with Lv17/WB/Rie1 induced detectable, albeit not significant, increases of circulating IFNγ, as well as TNFα, 7 days after boost (28 dpi), indicating that the presence of these small peaks seems to be directly related to virus recognition and protection against reinfections with the same strain. A second IFNγ peak observed three days after challenge (38 dpi) in some animals would suggest that IFNγ production may also be linked with protection against virulent challenge with Arm07.
Regarding the phenotypic characterization of the T cell responses, the different subpopulations of PBMC in the immunized animals suffered changes in their frequencies throughout the experiment. Without viral stimulation of PBMC, analysis of CD3+ and CD3+ subset frequencies showed that CD3, together with the activated/memory CD4CD8 cells, experienced a moderate and transient increase 7 days after immunization with Lv17/WB/Rie1, while, at the same time, cytotoxic CD8 and helper CD4 T cells experienced a clear drop in their frequencies. Significant increases in CD4CD8 cells have been reported as early as 4 days after immunization with the attenuated genotype I vaccine candidates OURT88/3 and Benin ΔMGF [
54]. Fluctuations in cytotoxic CD8 cells after immunization with both vaccine strains have also been reported [
54]. However, other authors did not observe changes in pigs immunized with OURT88/3 prior to challenge [
41]. In our study, these populations returned to pre-immunization values at 21 dpi, except for the helper CD4 cells, which remained low from 7 dpi and throughout the study. A progressive decrease of circulating CD4 helper T cells has also been described with the same genotype I vaccine strains listed above, in both protected and unprotected pigs [
41,
54]. All subset of CD3 T cells increased again after the challenge, mainly CD4CD8, moderately the CD8 cells and to a lesser extent the CD4 cells. Increases in CD3+ cells, mainly CD4CD8 and CD8, have also been reported following virulent challenge in pigs that became protected after immunization with OURT88/3 [
41]. Frequency analysis of these populations in response to recall ASFV showed the induction of a CD3-specific proliferative response between 21 and 28 dpi, although among the CD3+ subsets studied, only CD8 showed a small increase in response to recall virus at 21 dpi in some of the immunized animals. The increase in these circulating T cell subsets provides evidence of activation of the adaptive cellular immune response following immunization capable of responding to challenge with Arm07.
A peculiarity of the porcine immune system is the high expression of MHC class II (SLA-II) DR in resting lymphocytes, although the use of this marker, in combination with CD8α, has demonstrated to be of great help identifying activation in porcine helper T cells [
58]. Previously to immunizing the animals, circulating CD4 helper T cells lacked CD8 expression and had variable expression of SLA-II, phenotype that matched the normal description of resting helper T cells. CD4CD8 and CD8 T cells, however, expressed high levels of SLA-II, which is in line with the description carried out of these T cell subsets in blood of healthy pigs [
59]. Immunization with Lv17/WB/Rie1 led to the transient upregulation of the SLA-II surface protein in circulating CD8, CD4CD8 and CD4 T cells after prime and boost; however, these increases resulted significant only for the CD4 T cells. The transitory upregulation of SLA-II protein expression induced after immunization with Lv17/WB/Rie1 would indicate an increase in antigen presentation phenomena and would confirm the existence of regulatory mechanisms that activate an adaptive immune response. Due to the fact that co-expression between SLA-II and CD8α appears strongly associated in TCR- αβ T lymphocytes [
58,
59], it was not surprising that SLA-II expression remained high in both CD8 positive T cells during all study.
Simultaneous production of different cytokines or effector molecules on the single T-cell level has been proposed to be a hallmark of protective immune responses. For this purpose, we aimed to identify potential multifunctional virus-specific T cells. The correlation between virus-specific IFNγ-producing cells and protection has been described in some “in vivo” studies by using different techniques such as ELISA, ELISpot assay or flow cytometry [
11,
40,
60,
61]. In other studies, in which the authors carried out a phenotypic characterization by flow cytometry of IFNγ-producing cells in vaccinated pigs, they did not find a clear relationship between protection and induction of IFNγ-specific T cells. However, they could not rule out a possible protective role for these cells [
39,
41,
54]. The combined expression of two cytokines, such as IFNγ and TNFα, is a good indicator of the quality of the responses. Our results showed a high proportion of CD8 and CD4CD8 cells that, along with IFNγ, also co-expressed TNFα, with CD4CD8 cells exhibiting the highest virus-specific IFNγ+TNFα+ response throughout the study. Although a specific CD8 T-cell response was detected in some animals as early as 7 dpi, the induction of both responses became clearer between 21 and 28 dpi, showing the highest level of significance at 28 dpi (7 days after the boost). The induction of elevated percentages of ASFV-specific polyfunctional memory T cells, i.e. IFNγ+TNFα+ CD4CD8 T cells in pigs immunized with the BA71ΔCD2 deleted mutant, a vaccine candidate that conferred protection against virulent challenge with genotype II Georgia2007/1 strain, has also been described recently [
40]. In our study, the high levels of circulating IFNγ+TNFα+ CD4CD8 cells detected 7 days after boosting (28 dpi), in the absence of recall virus antigen, would also suggest that CD4CD8 T cells might be involved of the spontaneous increases of these cytokines detected in sera at this time point. Hence, intradermal immunization with Lv17/WB/Rie1 induced a robust ASFV-specific IFNγ T cell response, which was clearly detectable after the first immunization at 21 dpi, prior to booster, where CD4CD8 and CD8 T cells were identified as the main cellular sources of virus-specific IFNγ and TNFα. These results demonstrate the correlation between the induction of virus-specific CD4CD8 and CD8 T cells and protection against subsequent infections with both attenuated and virulent strains of ASFV genotype II.
One of our aims was to study the role of specific cytotoxic T-lymphocytes (CTL) in protection against subsequent ASFV infection. Although surface expression of CD8 has traditionally been attributed to cytotoxic functions, it is important to note that not all CD8 cells exhibit this capability [
62]. Therefore, the inclusion of markers indicating cytotoxicity, such as perforin or CD107a, may be useful in defining CTL subpopulations. The CD107a assay has been used to study cytotoxic degranulation associated with loss of perforin in porcine T cells following antigenic stimulation in other porcine viral diseases such as classical swine fever [
36,
38], porcine respiratory and reproductive syndrome (PRRS) [
37,
63] or swine influenza A [
64]. Although CD4+CD8+ and CD8 T cell cytotoxic activity, demonstrated by the detection of perforin expression, has been described during experimental ASFV infections [
15,
40,
62,
65], to our knowledge, this is the first study reporting the use of CD107a assay to identify CTL in ASFV-infected pigs. Immunization induced a variable, although progressive, increase of T cells with cytotoxic function in blood, as demonstrated by detection of the CD107a marker. CD8 CTL, CD4CD8 CTL as well as CD4 CTL increased from 21 dpi, and in the case of CD8 CTL they were maintained until 28 dpi. Mainly in the case of CD4CD8 CTL, this increased cytotoxic activity was accompanied by a marked and spontaneous secretion of cytokines and by increased expression of SLA-II. Only in the case of CD4 CTL, their progressive increase in blood was not accompanied by simultaneous cytokine production, although this increase in circulating CD4 CTL did appear to be associated, only after the boost (at 28 dpi), with an up-regulation of SLA-II expression on CD4 helper T cells. Similarly, upon virus stimulation, IFNγ and IFNγ+TNFα+ producing T cells were limited mainly to cytotoxic CD4CD8, and also to cytotoxic CD8. Taken together, these results confirm the important role of CD4CD8 T cells during the early stages of infection with attenuated ASFV strains in stopping viral replication, but also their key role mounting an effective adaptive immune response that induces protection against subsequent infections. The generation of an important subpopulation of antigen-experienced CD4CD8 T cells during the induction of the adaptive immune response was also confirmed. Furthermore, a second subset of memory CD8 T cells (CD4
negCD8
high) was also identified. These cells were able to proliferate quickly after antigen re-encounter, as indicated by the quality of these virus-specific multifunctional CD4CD8 and CD8 T responses elicited upon stimulation with both viruses at day 21 and especially at 28 dpi. Beyond 28 dpi, only CD4CD8 CTL and CD4 CTL remained very significantly elevated. However, after the challenge, while CD4CD8 CTL declined dramatically, CD4 CTL frequencies remained markedly elevated. The strong correlation between CD4 CTL appearance and control of infection with Arm07 suggested an important protective role of this subset in the control of early replication and infection with virulent virus in previously immunized pigs. It is noteworthy that, although the specific cytokine secretion by these T-cell subsets after challenge was weak, all of them (mainly CD4, followed by CD4CD8) were able to mount significant virus-specific CTL responses against both viral stimuli, indicating also the correlation between CD4 and CD4CD8 CTL subsets and protection against subsequent infection with a virulent genotype II isolate. The inclusion of additional markers, such as CD25, might help to better differentiate memory subpopulations among these CTL subsets.
Recent evidence highlights the potential role of CD4 CTL in controlling and protecting against viral diseases in pigs, particularly in the context of porcine respiratory and reproductive syndrome virus (PRRSV) infections [
63,
66]. These studies suggest that, in addition to the high levels of PRRSV-specific CD4CD8 CTL acquired by vaccination or previous infection, elevated levels of PRRSV-specific CD4 CTL are crucial in host defense against subsequent infections, even in the absence of neutralizing antibodies [
63,
66]. While CD4 CTL are unlikely to replace the function of CD8 CTL or CD8CD4 CTL, CD4 cytotoxic activity contributes to immune responses by targeting antigen-presenting cells (APC) via the MHC class II pathway. In mice and humans, CD4 CTL presence is associated with chronic viral infections, autoimmune diseases and cancer, attributing them important antiviral functions and suggesting their potential importance during adaptive cytotoxic immune responses [
67]. The induction of CD4 CTL responses targeting APC may be particularly relevant in scenarios where CD8 CTL responses are insufficient or compromised due to sustained antigenic stimulation. It is possible that the immunization regimen used in our study, involving prime and boost, led to CD8 cytotoxic cell fatigue, potentially explaining the relatively weak specific responses observed just before challenge. Additionally, viruses can evade the host immune system by down-regulating MHC class I expression in infected cells, hindering T-cell recognition of viral antigens. Although this evasion mechanism has not been confirmed during "in vivo" infections with virulent ASFV isolates, it has been demonstrated “in vitro”, implicating the viral protein EP153R [
68]. Thus, the increased frequency of CD4 CTL observed after challenge in our immunized/challenged pigs may play a crucial role in combating "de novo" infection with virulent ASFV, potentially through CD4 CTL-mediated killing of infected APC, such as monocyte/macrophages, which are primary target cells for ASFV. Further elucidation of the mechanisms underlying CD4 CTL differentiation could inform the development of more effective ASFV vaccines.