*2.4. Viral Loads in Bloods*

Primary viremia of TD/96 strain was first detected in both co-infected\_P1 pigs and co-infected\_P2 pigs at 2 dpi, and in co-infected with Ab pigs at 4 dpi, earlier than that of 94.4 strain, which was first detected in co-infected\_P1 pigs at 4 dpi and in both co-infected\_P2 pigs and co-infected with Ab pigs at 6 dpi (Figures 2A and 3A, Table 1). The viral loads of the TD/96 strain were always significantly higher than those of the 94.4 strain (*p* < 0.05) (Figures 2A and 3A) in all the co-infected pigs. The viral loads of the TD/96 strain were on average 1.89, 2.93, and 1.51 log higher than that of the 94.4 strain at all dpi in co-infected\_P1 pigs, co-infected\_P2 pigs and co-infected with Ab pigs, respectively. Comparing the three groups, the co-infected\_P2 pigs had the highest lead of TD/96 at all dpi; the viral loads were above 3 log higher than that of the 94.4 strain at 4 and 6 dpi. The highest lead of TD/96 was observed at 4 dpi, when it was on average 3.96 log higher than that of the 94.4 strain. No lead of TD/96 was above 3 log in co-infected\_P1 pigs or co-infected with Ab pigs.

**Figure 2.** Comparison of viral loads of co-infected\_P1 (Group 1) and co-infected\_P2 pigs (Group 2). Viral loads in (**A**) blood, (**B**) oral swabs, (**C**) fecal swabs, and (**D**) organs of co-infected\_P1 and co-infected\_P2 pigs were quantified by reverse transcription multiplex real-time polymerase chain reaction. The data represent the mean and standard deviation from three pigs. Values with different superscript letters, a–d, among the four groups of samples at the same dpi (**A–C**) or the same organ (**D**) indicate a statistically significant difference (*p* < 0.05) from each other. The superscript letter "a" indicates the highest viral load and "d" indicates the lowest viral load among the compared groups, while "ab" indicates a viral load in between categories "a" and "b". No significant differences exist between values containing the same letter. The absence of a superscript letter indicates no statistical analysis due to only one sample of a group or only one group within the same dpi.

**Figure 3.** Comparison of viral loads of co-infected\_P1 (Group 1) and co-infected with Ab pigs (Group 3). Viral loads in (**A**) blood, (**B**) oral swabs, (**C**) fecal swabs, and (**D**) organs of co-infected\_P1 and co-infected with Ab pigs were quantified by reverse transcription multiplex real-time polymerase chain reaction. The data represent the mean and standard deviation from three pigs. Values with different superscript letters, a–d, among the four groups of samples at the same dpi (**A–C**) or the same organ (**D**) indicate a statistically significant difference (*p* < 0.05) from each other. The superscript letter "a" indicates the highest viral load and "d" indicates the lowest viral load among the compared groups, while "ab" indicates a viral load in between categories "a" and "b". No significant differences exist between values containing the same letter. The absence of a superscript letter labeled indicates no statistical analysis due to only one sample of a group or only one group within the same dpi.

#### *2.5. Viral Loads in Secretions and Excretions*

In co-infected\_P2 pigs, the TD/96 strain was first detected in oral swabs and fecal swabs at 4 to 6 dpi, whereas in co-infected\_P1 pigs and co-infected with Ab pigs, the TD/96 strain first presented at 6 to 8 dpi (Table 1). The 94.4 strain was first detected in oral swabs and fecal swabs at 8 to 10 dpi in all co-infected pigs, except one co-infected\_P2 pig that died at 9 dpi, in which the virus was not detected (Table 1).

In oral swabs, the viral loads of the TD/96 strain were always significantly higher than those of the 94.4 strain in all the co-infected pigs (*p* < 0.05) (Figures 2B and 3B), with only one exception at 10 dpi of co-infected with Ab pigs. The viral loads of the TD/96 strain were on average 1.88, 3.08 and 1.59 log higher than that of the 94.4 strain at all dpi in co-infected\_P1 pigs, co-infected\_P2 pigs and co-infected with Ab pigs, respectively. Comparing the three groups, the co-infected\_P2 pigs had the highest lead of TD/96 at all dpi, in which the viral loads were above 3 log higher than that of the 94.4 strain at 6 and 8 dpi. The highest lead of TD/96 was observed at 6 dpi, when it was on average 4.74 log higher than that of the 94.4 strain. No lead of TD/96 was above 3 log in co-infected\_P1 pigs or co-infected with Ab pigs.

In fecal swabs, the viral loads of the TD/96 strain were always significantly higher than those of the 94.4 strain in all co-infected pigs (*p* < 0.05) (Figures 2C and 3C), with only one exception at 8 dpi of co-infected\_P2. The viral loads of the TD/96 strain were on average 2.49, 2.7 and 1.84 log higher than that of the 94.4 strain at all dpi in co-infected\_P1 pigs, co-infected\_P2 pigs and co-infected with Ab pigs, respectively. Comparing the three groups, the co-infected\_P2 pigs had the highest lead of TD/96 at all dpi; the highest lead of TD/96, with on average 3.45 log higher than that of the 94.4 strain, was observed at 6 dpi. No lead of TD/96 was above 3 log in co-infected\_P1 pigs or co-infected with Ab pigs.

#### *2.6. Viral Loads in Visceral Organs*

The viral loads of the TD/96 strain were consistently significantly higher than those of the 94.4 strain in most tested organs (*p* < 0.05) (Figures 2D and 3D) in all pigs. The viral loads of the TD/96 strain were on average 2.39, 3.00, and 2.09 log higher than those of the 94.4 strain in co-infected\_P1 pigs, co-infected\_P2 pigs, and co-infected with Ab pigs, respectively. Comparing the three groups, the co-infected\_P2 pigs had the highest lead of TD/96 in most tested organs; the viral loads were above 2.5 log higher than that of the 94.4 strain. The highest lead of TD/96 was observed in lymph nodes, where it was on average 3.33 log higher than that of the 94.4 strain.

#### **3. Discussion**

The shift in CSFV populations in the field, from genotypes 1 and 3 to genotype 2, was observed worldwide [8–12]. However, the mechanisms responsible for the shift remain unclear. A previous study hypothesized that genotype 2 had higher genetic diversity than genotypes 1 and 3 did, which might explain why it is the most prevalent endemic situation [19]. Three hypotheses were proposed: First, virus strains of genotype 2.1 may have higher replication efficiency than the genotype 3.4 strains in pigs; second, the strains of genotype 2.1 may have higher affinity and competitiveness to cellular receptors than those of genotype 3.4; and third, the strains of genotype 2.1 may have better ability than those of genotype 3.4 to escape from antibody neutralization induced by the attenuated lapinized vaccine strain LPC of genotype 1.1, which has been used to protect pigs against the 3.4 strains since the 1950s in Taiwan [12]. To test the first two hypotheses, Huang et al. [18] allowed two viruses belonging to genotypes 2.1 and 3.4, respectively, to compete in vivo and in vitro, and the results revealed that the virus of genotype 2.1 replicated more efficiently than that of genotype 3.4. To further explore and to simulate the field situation, this study analyzed the competitions of the viruses of the two genotypes in co-infected pigs for two passages without neutralizing antibodies (Groups 1 and 2). Moreover, we also examined the dynamics of virus replication and disease development of the infected pigs in the presence of maternally derived neutralizing antibodies induced by LPC vaccine (Group 3). To the best of our knowledge, this study is the first attempt to test the competitions of CSFV in this way.

The new CSFV strain of genotype 2.1 has higher replication efficiency than the historical genotype 3.4 strain in pigs. Indeed, when given equal opportunity to compete in the same animal for the first passage of competition, the genotype 2.1 CSFV (represented by TD/96) had 2.43 log (8.3–5.87) higher TCID50 titer over that of genotype 3.4 (represented by 94.4) in the blood (Group 2 inoculum, Table 1). This advantage of TD/96 in blood was also amply supported by quantifications of viral loads using quantitative reverse transcription multiplex real-time polymerase chain reaction (RT-MRT-PCR) (Table 2). Under various dual infection conditions with or without the presence of maternal antibodies, the TD/96 strain had consistently 1.51−3.08 log higher loads than those of 94.4. The 2.43 log TCID50 replication advantage of TD/96 in the first passage of competition in an animal body falls within the range of 1.51–3.08 log, as estimated by RT-MRT-PCR. Given a second passage of competition, the lead of TD/96 was widened further. The TD/96 strain was first detected in oral swabs and fecal swabs of co-infected\_P2 pigs 2–4 days earlier than in those of co-infected\_P1 pigs. On the other hand, the 94.4 strain was first detected in viremia of co-infected\_P2 pigs 2 days later than in those of co-infected\_P1 pigs, despite the two groups having been inoculated with similar amounts of the 94.4 strain (10<sup>6</sup> TCID50/mL in co-infected\_P1 pigs and 105.87 TCID50/mL in co-infected\_P2 pigs). The results revealed that when the viruses of genotype 2.1 are dominant in the field, the viruses replicate more efficiently and shed earlier than the viruses of genotype 3.4 do. Therefore, the genotype of CSFV in the pig population shifted from genotype 3.4 to 2.1. Given further passages of dual infection in pigs, the competition edge of genotype 2.1 would likely lead to the disappearance of genotype 3.4 in pigs.

Colostrum maternal antibodies offer partial protection for dually infected pigs, despite the genotype difference of the LPC vaccine virus (of genotype 1.1) from those of genotype 2.1 and 3.4 of dually infected viruses (Group 3). The TD/96 and 94.4 strains were first detected in the blood of co-infected with Ab pigs 2 days later than in co-infected\_P1 pigs, although pigs in these two groups were inoculated with the same amounts of the TD/96 and 94.4 strains. The maternal antibodies protection was ineffective when the genotypic heterologous neutralizing titer dropped to below 1:4 during dpi 8−14 (Table 3), for febrile reactions were observed after 10 dpi (Figure 1). This finding suggests that the protection offered by the maternal antibodies is not limited to the initial engagement to neutralize incoming virus but also allows for the host to launch immune responses before the system is overwhelmed.

Antigenic variations among various genotypes of CSFVs certainly render the currently available vaccines more effective in neutralizing historical viruses, while allowing the newly invading virus to escape. This hypothesis has been amply addressed in several previous studies [20–23]. The neutralizing antibodies of co-infected with Ab pigs against the LPC/AHRI strain were higher than those against the TD/96 strain and the 94.4 strain. However, there was no significant difference between the neutralizing antibodies against the TD/96 strain and those against the 94.4 strain. This evidence supported previous studies that antibodies induced by live virus neutralize genotypically homologous strains better than heterologous strains [17,24,25]. In co-infected with Ab pigs, similarly to Group 1, the TD/96 strain was shed by infected pigs earlier than the 94.4 strain was. In addition, the viral loads of the TD/96 strain were significantly higher than those of the 94.4 strain. The results indicated that the presence of maternally derived antibodies induced by modified live virus of genotype 1.1 might not influence the competition between viruses of genotypes 2.1 and 3.4. Previous study demonstrated that the LPC vaccine could offer pigs protection from challenges of field viruses of genotypes 2.1 and 3.4 [26]. This study revealed that when pigs or piglets infected by CSFV field viruses of genotypes 2.1 and 3.4 were vaccinated, the vaccine's efficacy was interfered with, and also that when the maternally derived antibodies declined in piglets before vaccination, the pigs produced lower neutralizing antibodies and had delayed clinical signs. However, the CSFVs were detected in those infected pigs, in which the genotype 2.1 viruses were released earlier and replicated more efficiently than did genotype 3.4 viruses. These results may suggest that the newly invading strains dominated in the field under vaccination.

Contact infection may present another aspect of virus competition. To investigate how the competition occurs between viruses of genotypes 2.1 and 3.4, in this study we employed the intramuscular route for the co-infection. This route can ensure that both viruses enter the hosts with the same amount of virus simultaneously to allow comparison of their replication efficiency and pathogenicity. However, when the virus is introduced through natural routes, the influencing factors could be more numerous and much complicated. To further explore the natural situations and to evaluate the possible effect of the superinfection exclusion phenomenon of CSFV [27] in the field in terms of the competence of these two genotypes, cohabitation infection with 2.1 virus infected pigs and 3.4 virus infected pigs is warranted.

In conclusion, examining the competition of the historical and newly invading genotypes of CSFV in co-infected pigs with different virus populations and with maternally derived neutralizing antibodies revealed that the new CSFV genotype 2.1 replicates more efficiently, at 1.51–3.08 log higher than that of the historical genotype 3.4. The maternally derived antibodies provide partial protection to both wild type CSFVs and correlate with lower clinical scores, febrile reaction, and animal mortality. In the presence of maternal antibodies, pigs could be infected by both wild type CSFVs, with the genotype 2.1 dominating. These results could further our understanding of the prevalence of genotype 2 in the field, which is widely observed in Asia and Europe. This is the first time that the higher replication capacity of genotype 2.1 than that of genotype 3.4 has been demonstrated in vivo with this design.

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

#### *4.1. Cells and Viruses*

Porcine kidney-15 (PK-15) cells were maintained in minimum essential medium supplemented with 10% fetal bovine serum and incubated at 37 ◦C in 5% CO2. The three CSFV strains used in this study, comprised the two representative CSFV strains TD/96 and 94.4 and an attenuated lapinized vaccine strain LPC/AHRI, were propagated in the PK-15 cells [3,12,13].

#### *4.2. mAbs Specific for CSFV*

Three mAbs against CSFVs were used in this study. The mAbs T6 and L71 were produced by the Animal Health Research Institute, Taiwan, and the mAb WH303 by the Animal and Plant Health Agency, the United Kingdom. The mAb T6 recognizes the TD/96 strain of genotype 2.1 but not the 94.4 strain of genotype 3.4. In contrast, the mAb L71 recognizes the 94.4 strain but not the TD/96 strain [18]. The mAb WH303 reacts with most CSFV strains tested [28], including the three strains used in this study.

#### *4.3. Experimental Infections*

Six 4-week-old specific pathogen-free (SPF) pigs were randomly separated into two groups of three pigs: Groups 1 and 2 (Table 1). Pigs in Group 1 (co-infected\_P1) were inoculated intramuscularly with 1 mL of the TD/96 strain and 1 mL of the 94.4 strain simultaneously, each at a virus amount of 106 TCID50/mL, to ensure that both strains could enter the hosts simultaneously. The concept of the Group 1 co-infection experiment design was similar to that described in Group 3 of Huang et al. [18] and was repeated here, in separate pigs, for comparison with Groups 2 and 3 (see below). Pigs in Group 2 (co-infected\_P2) were inoculated intramuscularly with 1 mL of whole blood taken from a pig of Group 1 at 12 dpi. This inoculum contained TD/96 108.3 TCID50 and 94.4 105.87 TCID50 as titrated later (Table 1). Group 3 (co-infected with Ab) included three 4-week-old pigs born from a sow vaccinated with LPC vaccine, in which the maternal antibody response was in decline. These animals were chosen in order to examine the effect of antibody drop on co-infection. These pigs were inoculated intramuscularly with 1 mL of the TD/96 strain and 1 mL of the 94.4 strain simultaneously, as in Group 1, at 106 TCID50/mL. The three groups were housed separately in three negative air-pressure isolation units. For animal welfare reasons, pigs were euthanized when they were moribund and unable to stand up. All surviving pigs were euthanized at 14 dpi, the end of the experimental period. This animal experiment was approved by the Institutional Animal Care and Use Committee of the Animal Health Research Institute (Approval number A02040).

## *4.4. Clinical Signs, Body Temperature, and Sampling Procedures*

Rectal temperature was recorded daily during the experimental period. Fever was defined as a temperature higher than 40 ◦C. For evaluation of clinical signs, the ten parameters described by Mittelholzer et al. [29] were scored from 0 to 3 to represent normal to severe CSF symptoms. The scores of each pig were summed into a total score for each day. Blood, oral swabs, and fecal swabs were collected prior to inoculation at 0 dpi and then at 2-day intervals post infection. Swabs were weighed before and after sampling to normalize the viral loads. Each swab was immersed in 2 mL of phosphate buffered saline (PBS) and centrifuged at 3,000 × *g* for 10 min, and the harvested supernatant was stored at −70 ◦C. Necropsies were performed after euthanasia or death, and tissue samples of tonsil, submandibular and mesenteric lymph nodes, heart muscle, lung, liver, spleen, kidney, bladder, and cerebrum were collected from all pigs.

#### *4.5. Virus Titration*

Ten-fold serial diluted blood (inoculum of Group 2) at 12 dpi from a pig of co-infected\_P1 was added into eight wells each of 96-well plates duplicated and seeded with PK-15 cells. Whether the cells were infected was observed using indirect fluorescent assay (IFA) at 72 hours post infection (hpi). The mAbs T6 and L71 were used for virus titration. One 96-well plate was stained with mAb T6, which recognizes the TD/96 strain but not the 94.4 strain; the other 96-well plate was stained with mAb L71, which recognizes the 94.4 strain but not the TD/96 strain [18]. Virus titers were calculated as TCID50 using the Reed-Muench method [30].
