Adverse Effects of Classical Swine Fever Virus LOM Vaccine and Jeju LOM Strains in Pregnant Sows and Specific Pathogen-Free Pigs

In Jeju island of South Korea, a classical swine fever (CSF) non-vaccinated region, many pig farmers insisted on abortion and stillbirth in pregnant sows and high mortality of suckling/weaning piglets by circulating CSF virus from 2014 to 2018. We investigated whether CSF viruses isolated from pigs in Jeju Island (Jeju LOM) have recovered their pathogenicity by conducting experiments using pregnant sows and specific pathogen-free (SPF) pigs. The CSF modified live LOM vaccine (MLV-LOM) and Jeju LOM strains induced abortion and stillbirth in pregnant sows. Viral antigens were detected in the organs of fetuses and stillborn piglets in the absence of specific pathological lesions associated with the virulent CSF virus in both groups (MLV-LOM and Jeju LOM strain). However, antigen was detected in one newborn piglet from a sow inoculated with a Jeju LOM strain, suggesting that it may cause persistent infections in pigs. SPF pigs inoculated with the MLV-LOM or Jeju LOM strains were asymptomatic, but virus antigen was detected in several organ and blood samples. Virus shedding in both groups of animals was not detected in the feces or saliva until 21 days post inoculation. The serum concentration of the three major cytokines, IFN-α, TNF-α, and IL-10, known to be related to lymphocytopenia, were similar in both groups when the MLV-LOM or Jeju LOM strains were inoculated into SPF pigs. In conclusion, Jeju LOM strains exhibited most of the characteristics of the MLV-LOM in pigs and resulted in the same adverse effects as the MLV-LOM strain.


Introduction
Classical swine fever virus (CSFV) is a small, enveloped virus with a positive-stranded RNA genome of approximately 12.5 kb in size and contains a single, large, open reading frame that encodes a 3898 amino acid (aa) polyprotein [1]. Classical swine fever (CSF) is a devastating disease, causing substantial economic losses through the death of valuable livestock. Major outbreak of CSF is now rare, but sporadic epizootic incidents still occur frequently, causing chronic, atypical forms of the disease. Virulent CSFV enters the host through the mucous membranes of the oral and nasal cavities, known to infect tonsil

Viral Antigen Detection in Pregnant Pigs 3 Weeks Post Inoculation
One (no. 37-5952) of three pregnant sows in group 1 was autopsied three weeks post inoculation (wpi) with strain JJ16LOM-YJK08 (Jeju LOM), and CSFV RNA was detected in the tonsils by qRT-PCR. In addition, 17/20 (85%) fetuses of the same pregnant sow were positive for CSFV RNA by qRT-PCR (Table 1). CSFV RNA was detected in the heart of a pregnant sow (no. 51-2104) in group 2, which was inoculated with strain JJ16LOM-YYM02 (Jeju LOM), and 9/14 (64%) fetuses were CSFV RNA positive (Table 1). In a pregnant sow (no. LI-5534) in group 3, inoculated with strain 16LOM-KM00 (MLV-LOM), a tonsil sample was CSFV RNA positive, and 7/9 (78%) fetuses were positive for the CSFV RNA (Table 1). No histologic preparations of the three pregnant sows revealed lesions associated with CSF. Pathologic examination of the internal organs of 43 fetuses from the three sows showed no specific lesions, and no viral antigen was detected by the immunohistochemical (IHC) assay. * Lung, spleen, liver, kidney, ileum, brain, bladder.

Pathogenicity in Pregnant Pigs and Their Piglets
In group 1, inoculated with strain JJ16LOM-YJK08 (Jeju LOM), and group 2, inoculated with strain JJ16LOM-YYM02 (Jeju LOM), the total number of fetuses produced was 23 and 32, respectively. In group 1, nine fetuses were mummified, 13 were stillborn, and one was a live birth; in group 2, 11 were mummified, 15 were stillborn, and six were live births. The crown-rump (C-R) length range of fetuses was 5.0-28.5 cm for group 1 and 6.5-30.0 cm for group 2 ( Table 2). The total number of fetuses in group 3 inoculated with strain 16LOM-KM00 (MLV-LOM) were 13, of which 12 were mummified and one was a live birth. The C-R length of fetuses ranged from 11.5 to 28 cm (Table 2). CSFV RNA detection in fetuses was 20/23 (86.9%) for group 1, 23/32 (71.8%) for group 2, and 12/13 (92.3%) for group 3 samples by qRT-PCR (Table 2). When strains JJ16LOM-YJK08 and JJ16LOM-YYM02 were used as inocula, CSFV RNA was detected in all samples (13/13 and 15/15, respectively) from stillborn fetuses, and all 12 mummified fetuses in group 3 contained CSFV RNA (Table 2). In addition, one out of eight live piglets tested from all groups (1-3) was CSFV RNA positive in organ samples ( Table 2).
One weak-born piglet and one still-born fetus delivered from sow no. 34-5053 in group 1 exhibited necrosis, hemorrhage, severe vacuolation, and perivascular cuffing in the white matter of cerebellum, and one fetus also exhibited severe myocardial necrosis in the heart ( Figure 1A,C). Three piglets from sow no. 62-0093 in group 2 were identified as non-suppurative encephalitis with hemorrhage and mineral deposition in the white matter of in the cerebellum. However, no specific lesions were observed in fetuses in the other groups (3 and 4). By IHC staining, viral antigens were detected in the internal organs group 1 weak born piglet and stillborn fetus ( Figure 1B,D) and in the heart of one stillborn group 2 fetus. Seroconversions for CSF neutralizing antibodies in pregnant pigs were on average 7 log Pathogens 2020, 9,18 5 of 13 One weak-born piglet and one still-born fetus delivered from sow no. 34-5053 in group 1 exhibited necrosis, hemorrhage, severe vacuolation, and perivascular cuffing in the white matter of cerebellum, and one fetus also exhibited severe myocardial necrosis in the heart ( Figure 1A, C). Three piglets from sow no. 62-0093 in group 2 were identified as non-suppurative encephalitis with hemorrhage and mineral deposition in the white matter of in the cerebellum. However, no specific lesions were observed in fetuses in the other groups (3 and 4). By IHC staining, viral antigens were detected in the internal organs group 1 weak born piglet and stillborn fetus ( Figure 1B, D) and in the heart of one stillborn group 2 fetus. Seroconversions for CSF neutralizing antibodies in pregnant pigs were on average 7 log 2 for group 1, 6.5 log 2 for group 2, and 7 log 2 for group 3, when measured 3 wpi, and 7 log 2, 9.5 log 2, and 10 log 2 at delivery, respectively.

Discussion
The immunogenicity of the CSF MLV-LOM in pregnant sows was found to be excellent, but it was detected in some organs of their fetuses and sows, thus confirming that the MLV-LOM can pass through the placenta and infect fetuses [14]. In addition, the occurrence of stillborn or mummified fetuses suggests that the MLV-LOM may retain some pathogenicity in gestating pigs [14]. Interestingly, our results showed that the overall prevalence of infected fetuses was higher (92.3%) with the MLV-LOM than with the Jeju LOM strains (71.8-86.9%) isolated from Jeju Island. Histopathological examination revealed degeneration and necrosis of myocytes in the heart. Non-suppurative encephalitis and multifocal mineralizations in the brain only occurred in 9.1% (5/55) of fetuses delivered from sows inoculated with Jeju Island isolates. In addition, viral antigens were detected by the IHC test in 5.5% (3/55) of fetuses. One live piglet delivered from a sow inoculated with strain JJ16LOM-YJK08 exhibited severe vacuolar change with perivascular cuffing in cerebellum and myocardial degeneration, and a large amount of antigen was expressed in internal parenchymal organs. If the piglet had not been autopsied, it would have remained persistently infected and act as a carrier, steadily releasing the antigen. It has been shown that CSF infection occurs between 50 and 70 days of pregnancy, an immunotolerance phenomenon can be induced, and persistently infected offspring are born [3][4][5][6][7]. These piglets were healthy and survive for several months but die due to late onset form of CSF and they also shed high viral loads, which are enough for transmission to other pigs [3][4][5][6][7]. Recent studies have suggested that persistent infection can occur when newborn piglets are infected within the first 8 h of life, or even up to 48 h after birth [18,19].
In a previous report from 1990, safety tests with SPF pigs were conducted for the LOM-850 strain (the original MLV-LOM) and the LOM-suri strain cloned from LOM-850 [20]. Gross findings in SPF pigs inoculated with both strains were not observed, but all the groups showed mild lesions in the lymph nodes and bladder [20]. At necropsy, 10 days after vaccination, the redness of the lymph nodes was decreased, and bladder bleeding ceased [20]. In addition, IHC assay on SPF pigs at 21 days after vaccination were normal. Overall, gross and pathological findings decreased with time after vaccination and returned to normal at 21 days after vaccination [20]. In this study, we observed no specific clinical symptoms related to CSF in the pathogenicity tests on SPF pigs during the experimental period. However, CSFV RNA was detected in lymphoid organs such as tonsils, lymph nodes, spleen, and ileum, and showed a tendency to decrease in tissues with time post inoculation when assessed by qRT-PCR. When samples of SPF pig lymphoid organs were tested using the IHC stain, no lesions of lymphoid depletion, which are known to be characteristic of virulent CSFV infections, were observed. There was no specific high-temperature reaction in SPF pigs observed in this study, but transient leucopenia was observed in SPF pigs inoculated with strain JJ16LOM-JYK08 or JJ16LOM-JYK08-F at 4 dpi. However, the average WBC levels for these groups were 7142 and 6300/µ1, which were not very low, and later, the WBC levels recovered.
Lymphocytopenia was well known as one of the typical characteristics of CSFV infections [8,[21][22][23]. In general, lymphocytopenia, as well as blood coagulation disorder, are considered to be related to the upregulation of inflammatory cytokines, chemokines, and adhesion factors [24,25]. IFN-α is known to be involved in the induction of lymphocytopenia [9,10], and IL-10 is also known to related to general immunosuppression [26,27]. The concentration of IFN-α was highest at 4 dpi after strain LOM16-KO00 and JJ16LOM-YJK08 inoculation, but the difference between the two was not significant. The concentration of IL-10 at 7 dpi was higher in SPF pigs inoculated with strain JJ16LOM-YJK08 or JJ16LOM-YJK08-F than in those inoculated with LOM16-KO00. The recent study showed that TNF-α and IFN-α expressions in the sera of pigs infected with a moderately virulent CSFV increased, whereas concentrations in pigs following inoculation of the vaccine C-strain remained normal [28]. Generally, TNF-α is believed to be involved in the inflammatory response [28]. It has been also shown that an increase in TNF-α or IFN-α may induce the inflammatory response or apoptosis of lymphocytes after virulent CSFV infection [28]. In addition, TNF-α has been considered, alongside IL-1, IL-6, and IL-4, to be responsible for, among others, apoptosis in immune system cells, such as macrophages [10,29,30] and dendritic cells in lymphoid organs [27]. Several previous reports revealed that the increased expression of IL-8 after infection, without elucidation of its direct role in CSFV pathogenesis [24,25,31]. Our results suggest that TNF-α, IL-4, IL-6, and IL-8 showed no difference during the observation period between the three strains (JJ16LOM-YJK08, JJ16LOM-YJK08-F, and 16LOM-KO11).
Cytokine concentrations were at their maximum 2-3 days earlier in pigs infected with CSFV strains of moderate and high virulence when compared with those infected with a strain of low virulence [28]. These findings suggest that the C-strain vaccine virus replicates very slowly in some tissues, including tonsils, and therefore escape recognition by host antiviral responses. This slow virus (CSF vaccine strain) replication may have resulted in avoiding the untimely over-expression of IFNs and proinflammatory cytokines, and so indirectly protect immune cells and organs [28]. Based on previous suggestions, we assume that the MLV-LOM strain may share similarities with the C-strain vaccine virus because they are both live attenuated vaccine strains. In conclusion, the CSFV Jeju LOM strains showed no pathogenicity in SPF pigs but exhibited similarly to the MLV-LOM. The Jeju LOM strains produced similar adverse effects in the fetuses of pregnant sows to those of the MLV-LOM. Therefore, we suggest that the Jeju LOM strains are not pathogenic revertants of the MLV-LOM. Our epidemiologic results for CSFV (Jeju LOM) outbreak in Jeju from 2014 to 2018 indicated five major factors that may cause severe damage by Jeju LOM strain in Jeju pig farms of CSF non-vaccinated area. First, inoculation of MLV-LOM strain in pregnant sow (contraindication to vaccination policy); second, no stamping out for first Jeju LOM inoculated pregnant sow and their offspring suckling piglets; third, failure to clean and disinfect feces containing Jeju LOM strain in the pig farms; fourth, failure of disease prevention between slaughterhouse and pig farm; fifth, quarantine failure between mainland (South Korea) and Jeju island. Therefore, the MLV-LOM strain including Jeju LOM strain should never be inoculated in pregnant pigs without anti-CSF antibody. We suggest that MLV-LOM should be replaced quickly with CSF live marker vaccine with excellent safety and efficacy as a government CSF preventive policy.

Pregnant Pigs Inoculated with MLV-LOM and Jeju LOM Strains
Ten pregnant sows (64-67 days of gestation) were divided into three groups, and were inoculated with strain JJ16LOM-YJK08 (Jeju LOM) in group 1 (n = 3), strain JJ16LOM-YYM02 (Jeju LOM) in group 2 (n = 3), and strain 16LOM-KO00 (MLV-LOM) in group 3 (n = 3) with 10 3.5 TCID 50 /dose. Group 4 (n = 1) served as an uninoculated negative control. For groups 1-3, one pregnant sow from each group was autopsied at 3 wpi, and the remaining sows and their offspring piglets were autopsied after delivery. To investigate the presence of neutralizing antibodies, a NPLA was performed on the serum of sows collected before inoculation, 3 wpi, and at delivery (7 wpi). At 3 wpi, autopsies of pregnant sows and fetuses were tested for signs of CSF. At delivery, all piglets were examined for farrowing and the presence of antigen in their fetuses. The organs of their piglets were subjected to pathologic analysis.

SPF Pigs Inoculated with the MLV-LOM and Jeju LOM Strains
Twenty-two SPF pigs (on average 45-50 days old) were purchased from Optifarm company (Osong, Korea) and divided into four groups to allow comparison of the relative pathogenicity of the Jeju LOM and MLV-LOM. Group 1 animals (n = 10) were inoculated with strain JJ16LOM-YJK08 (Jeju LOM). Those in group 2 (n = 5) were inoculated with strain JJ16LOM-YJK-F (Jeju LOM). Those in group 3 (n = 5) were inoculated with a commercial vaccine strain (16LOM-KO00) (MLV-LOM), and group 4 pigs (n = 2) were mock as a negative uninoculated control. Clinical symptoms in SPF pigs were observed daily. Blood, saliva, and fecal swab samples were taken, and weight and temperature checks were performed before inoculation and at 2, 4, 7, 10, and 14 dpi for virus detection, leukocyte counts, weight and temperature changes, and seroconversion. At 14 dpi, five SPF pigs in group 1, five in group 3, and one in group 4 were autopsied to test for the presence of viruses within organs by qRT-PCR and IHC assays, and to note any lesions in tissues. The remaining SPF pigs were autopsied and collected samples (blood, saliva, and fecal) at 21 dpi.

Multiplex Immunoassay for SPF Pigs
The porcine cytokine and chemokine 9-plex Porcine ProcartaPlexTM Panel 1 (ThermoFisher Scientific, Cat no. EPX090-60829-901) was used to detect nine cytokines and chemokines (IFN-α, IFN-γ, TNF-α, IL-1β, IL-4, IL-6, IL-8, IL-10, and IL-12p40) in the sera from SPF pigs inoculated with MLV-LOM (16LOM-KO11) and Jeju LOM strains (JJ16LOM-YJK08 and JJ16LOM-YJK08-F). ProcartaPlex immunoassays are based on the principles of a sandwich ELISA, using two highly specific antibodies binding to different epitopes of one protein to quantitate all protein targets simultaneously using Luminex ® 200 TM (Luminex Co., TX, USA). In brief, serum fractions from blood collected in EDTA-containing tubes were obtained after centrifugation at 1000× g for 10 min at 20-25 • C. Magnetic beads were vortex-mixed for 30 s, 50 µl of the beads was added to each well, and then pig-specific universal assay buffer and sample were each added in 25 µl volumes to the wells. The plates were shaken at room temperature (RT) for 30 min, overnight at 4 • C in the dark, and then at RT for a further 30 min. Beads were then washed twice. Detection Antibody mix (25 µl) was added to the beads, and the beads were incubated with shaking at RT for 30 min and then washed twice. Streptavidin-PE (50 µl) was then added, and the beads were incubated with shaking at RT for 30 min and washed