*2.10. Vertical Transmission and Congenital Viral Persistence Generated by the Moderate Virulence CSFV Strain*

At 114 days of gestation, Sow 5 gave birth to eight live piglets and six stillbirths, the live animals were active and fed normally from the mother immediately after birth. During the seven days after farrowing, two piglets were found dead in the pen, having being crushed by the sow, whereas no clinical signs were registered in the remaining six animals (Piglets 1 to 6). Piglets 1, 3, and 5 remained clinically healthy during the 32 days of the trial. Meanwhile, the other three piglets (Piglets 2, 4, and 6) developed sporadic fever peaks (below 41 ºC) from day 10 until the end of the study. Piglet 6 developed mild polyarthritis from day 10, and Piglets 2 and 4 at days 30 and 23, respectively. Notably, at the time of euthanasia, the piglets weighed around 8.5 kg and continued to show normal feeding behaviour.

On the day of birth, the piglets were positive by RT-qPCR with high CSFV RNA load (Ct values about 23) in the rectal swab samples (Table 4). Despite the absence of CSF specific clinical signs, a high CSFV RNA load (Ct value about 20) was detected in all the serum samples during the study, indicating a permanent viremia in the piglets during the trial (Table 4). In parallel, high and permanent excretion in nasal and rectal swabs was found in all the sampling time points during the 32 days after birth. The Ct values increased in the majority of animals throughout the trial, reaching Ct values around 22 and 24 in nasal and rectal swabs (Table 4).


 **4.** Detection of CSFV RNA in piglets from Sow 5, infected with the CSFV Cat01 strain.

**Table**

#### *2.11. Immune Response Generated by the Moderately CSFV Strain in Sows and Their Litters*

After infection, Sow 6 developed CSFV specific humoral response at 14 dpi, while Sow 5 was positive at 21 dpi, with blocking percentage values of 42% and 60%, respectively, which increased throughout the study. Neutralising antibody response appeared at 21 dpi on Sow 5 (titre 1:120), and at 14 dpi in Sow 6 (titre 1:20), and increased, reaching titres of 1:160 in both sows by 28 dpi. Nevertheless, none of the piglets showed an antibody response either by ELISA or NPLA during the 32 days after birth. Interestingly, IFN-α was detected in the sera from 4 piglets at 8 and 15 days post-birth (dpb) (Figure 5A). On the other hand, alterations in the CD4<sup>+</sup> and CD8<sup>+</sup> T-cell subsets from Cat01-infected piglets were found in the analysed PBMC from persistently infected piglets. While the T-CD4<sup>+</sup> population did not exceed 5% in the uninfected, age-matched piglets, these cells ranged from 4% to 16% in persistently infected animals. On the other hand, the CD8<sup>+</sup> cell subset was increased (about 50%) in the infected animals, being between 9.6 and 23.5% in the uninfected animals (Figure 5B).

**Figure 5.** IFN-α levels and phenotypic profile in peripheral blood mononuclear cells (PBMCs) from CSFV congenital persistently infected piglets. (**A**) Concentration of IFN-α in sera expressed as units/mL from CSFV persistently infected piglets at 7 dpb (grey bars) and 15 dpb (white bars). (**B**) Comparative expression of the T-CD4+ and T-CD8+ cell subsets in PBMCs from uninfected piglets (white dots) and piglets infected with the Cat01strain (grey dots).

#### **3. Discussion**

CSF congenital persistent infection was described several decades ago; however, some aspects regarding the generation of this form of the disease remain to be elucidated, and the available information is not up to date [4,6]. In the present work, three groups of sows were infected with either the Margarita, Cat01, or PdR CSFV strains. Each of these strains hav been previously characterised as of high, moderate, and low virulence, respectively [20–22]. In accordance with previous studies, the infection was carried out at 74 days of gestation, a time-point in which persistent congenital infection can be generated [4]. The capacity for trans-placental transmission and induction of foetal immune response was compared side by side between the high and low virulence strains (Figures 1 and 2, Tables 1 and 2). In accordance with previous data found in piglets, the highly virulent CSFV Margarita strain induced high serum IFN-α levels in sows over a short period of time [23]. By contrast, the IFN-α response induced by the low virulence PdR strain was lower, although it lasted one week longer. This supports the role of high replication rates for the previously described exacerbated innate immune response in the host after infection with highly virulent CSFV strains. This may explain the differences in pathogenesis between the sows from these two groups, with more severe lesions and an inability to clear the virus in the Margarita-infected sows, compared with the clinically healthy status and low replication of the PdR-infected ones. Trans-placental transmission was more efficient with the highly virulent Margarita strain, and high viral RNA load was detected in sera and tissues from the foetuses in this group. Conversely, a small proportion of the foetuses from the PdR infected sows were viraemic with high viral replication in organs, while the majority of them were either non-infected or only showed low viral RNA in tissues. Despite the immune response developed, mainly in Margarita infected sows, CSFV crossed the trans-placental barrier from the sows to their foetuses (Tables 1 and 2, Figure 2). In agreement with previously described data, the high replication rate found in sows infected with a highly virulent CSFV strain may explain the activation of neutralising antibody response in these animals. However, taking into account that the onset of the antibody response in the sows was after two weeks, it is likely that the generation of trans-placental transmission took place during the first week after infection. Considering the previously described data, in order to avoid trans-placental transmission, it is necessary that effective neutralising antibody response be already present at the moment of infection, with titres of at least 1/320 [24].

Mummifications and haemorrhagic lesions were found in the Margarita infected foetuses. Probably, these animals would have died during the perinatal period. On the contrary, neither mummifications nor macroscopic lesions were observed in the PdR infected foetuses, even in those that showed viremia and high levels of viral replication in organs. It is very well known that sows transmit passive immunity to CSFV to the litters via colostrum [25,26]. These maternally derived antibodies (MDA) protect piglets against disease, including CSF, during their firsts weeks of life [15,25,26]. Considering that, in the case that the piglets had been born, the low immunity generated in the sows after infection with the low virulence PdR strain would result in an inefficient transmission of MDA to these litters. There might be major consequences to this situation since the suboptimal level of MDA would favour the infection of the non-infected piglets by their congenital persistently infected littermates and lead to chronic or postnatal persistent infection [10,11]. Recently, it was reported that the lack of maternal immunity led to a high prevalence of CSFV persistently infected piglets in an endemic scenario [10]. Notably, the CSFV persistently infected piglets have been proven to be refractory to vaccination [10,27]. This complex situation may lead to a vicious circle, which greatly impairs control programs of regions where CSF persistent infections are occurring.

In the case of infection carried out with the moderately virulent CSFV Cat01 strain, early labour in one of the infected sows and mummification and stillbirths in both of them were detected. Interestingly, both Cat01 infected sows developed a CSFV neutralising antibody response. However, the viral trans-placental transmission was not impaired, and all the piglets that were born alive in one of the Cat01 infected litters developed persistent congenital infection. These piglets showed normal weight gain, according to standards [28], despite being infected and excreting high viral load with a lack of

CSFV specific antibody response [1,5,6]. Interestingly, the level of viral replication was comparable, or even higher than those found in the foetuses from sows infected with the high virulence strain (Tables 2 and 4). This finding suggests an immunomodulatory capacity of the moderate virulence CSFV strains in the interaction with the host. Previous data showed the efficacy of this type of CSFV strain to also generate persistent postnatal infection [11,27]. Similar to persistent postnatal infection, low levels of IFN-α were found during congenital viral persistence, despite the high viral replication, pointing towards immunosuppressive regulation. Similar mechanisms might be taking place during the establishment of congenital or postnatal viral persistence. Recently, myeloid-derived suppressor cell populations have been determined to play a relevant role in the generation of CSFV postnatal persistence infection [29]. It cannot be discarded that these cell subsets are playing a role during the establishment of CSFV congenital persistent infection, considering that they have been found in cord blood and during neonatal stages in humans [30,31]. On the other hand, a low CD4/CD8 ratio has been reported as a marker for dysregulation of the immune response [32–35]. An increase in the CD8<sup>+</sup> T-cell population, resulting in a low CD4/CD8 ratio, has been reported in CSFV postnatal persistently infected animals [12]. In the present study, an increase in the CD8<sup>+</sup> T-cell subset was observed in the PBMC of infected foetuses and piglets from all the experimental groups. This finding may indicate that immunosuppressive mechanisms are also taking place in animals after trans-placental infection by CSFV.

Activation of innate immunity, evidenced by the IFN-α and IFN-γ levels detected in sera, was found in the foetuses and piglets regardless of the infecting strain and the maturity level of the immune system (Table 3 and Figure 5). Type I interferon response activates the innate immunity after viral infection by playing an antiviral and immunomodulatory role. CSFV has the capacity to induce high levels of IFN-α response in pigs, being associated with disease severity and viral replication in the infected animals [36]. The highest IFN-α response was found in the viraemic foetuses or in those that showed higher viral replication in organs from the group infected with the low virulence PdR strain. Notably, the capacity of the PdR strain for high and prolonged IFN-α activation in piglets has been associated with an uninterrupted 36-uridine sequence found in the 3 untranslated region of the CSFV genome [23]. Activation of IFN-α response in ruminant and human foetuses, following infection with bovine viral diarrhoea and Zika virus, respectively, has been described, and it may support the results obtained in this study [37,38]. Thus, the immunotolerance mechanism that was previously associated with the development of CSF congenital persistent form [1,5] is a complex immunologic phenomenon, and further studies may explain this mechanism and its relation with the establishment of viral persistence.

Previous reports have shown that the levels of sCD163 can be increased as a result of tissue damage during acute infection with highly pathogenic viruses, such as the African swine fever virus (ASFV) [39,40]. In addition, increased IFN-γ levels have also been found as part of the cytokine storm phenomenon responsible for the pathogenesis of ASFV [39–41]. In agreement with the haemorrhagic lesions and levels of viral replication found in foetuses infected with the high virulence CSFV Margarita strain, it is likely that the increase of IFN-γ and sCD163 may be associated with the exacerbated immune response in the host after infection, leading to cellular homeostasis imbalance and tissue damage.

Taken together, our results show that the infecting CSFV strain capacity for viral replication influences its efficacy for trans-placental transmission and the establishment of persistent infection. Likewise, the CSFV strain with a moderate virulence degree proved to be very efficient in generating CSFV congenital persistent infection following trans-placental transmission. Our results indicate that trans-placental infection took place very fast before the neutralising antibody response could be generated in sows. Therefore, vaccines against CSFV indicated for pregnant sows must induce fast and strong immunity to guarantee the viral protection of their offspring against this type of infection.

On the other hand, the foetal immune system is able to recognise the virus and generate immune response after trans-placental infection. Further studies are needed to elucidate the mechanisms by which the specific immune response against CSFV is being impaired, following the initial recognition

of the pathogen. To the best of our knowledge, this is the first report showing the foetal immune response after CSFV infection.

#### **4. Materials and Methods**

#### *4.1. Cells and Viruses*

Production of the viral strains was carried out by infecting susceptible cells with viral suspensions in 2% pestivirus-free foetal bovine serum using the porcine kidney cell line PK-15 (ATCC CCL 33, Middlesex, England), cultured in Eagle's minimum essential medium supplemented with 5% foetal calf serum. Following the infection, cells were incubated at 37 °C in 5% CO2, and after 72 h, the virus was harvested. Peroxidase-linked assay (PLA) [42] was used for viral titration following the statistical methods described by Reed and Muench [43]. The CSFV PdR and the Margarita strains, both belonging to the 1.4 subgenotype [44,45], have been characterised as low and high virulence strains, respectively [20,21]. The Cat01 strain, which belongs to subgenotype 2.3, was selected as a moderate virulence prototype [22].
