Plant‐made E2 glycoprotein single‐dose vaccine protects pigs against classical swine fever

Abstract Classical Swine Fever Virus (CSFV) causes classical swine fever, a highly contagious hemorrhagic fever affecting both feral and domesticated pigs. Outbreaks of CSF in Europe, Asia, Africa and South America had significant adverse impacts on animal health, food security and the pig industry. The disease is generally contained by prevention of exposure through import restrictions (e.g. banning import of live pigs and pork products), localized vaccination programmes and culling of infected or at‐risk animals, often at very high cost. Current CSFV‐modified live virus vaccines are protective, but do not allow differentiation of infected from vaccinated animals (DIVA), a critical aspect of disease surveillance programmes. Alternatively, first‐generation subunit vaccines using the viral protein E2 allow for use of DIVA diagnostic tests, but are slow to induce a protective response, provide limited prevention of vertical transmission and may fail to block viral shedding. CSFV E2 subunit vaccines from a baculovirus/insect cell system have been developed for several vaccination campaigns in Europe and Asia. However, this expression system is considered expensive for a veterinary vaccine and is not ideal for wide‐spread deployment. To address the issues of scalability, cost of production and immunogenicity, we have employed an Agrobacterium‐mediated transient expression platform in Nicotiana benthamiana and formulated the purified antigen in novel oil‐in‐water emulsion adjuvants. We report the manufacturing of adjuvanted, plant‐made CSFV E2 subunit vaccine. The vaccine provided complete protection in challenged pigs, even after single‐dose vaccination, which was accompanied by strong virus neutralization antibody responses.

The virus is highly contagious to both feral and domesticated pigs, and is horizontally transmitted by direct contact of healthy pigs with infected animals (Fritzemeier et al., 2000;Laevens et al., 1999;Moennig et al., 2003). Infection severity can vary widely depending on the virulence of the strain. Clinical signs include: (i) fever, (ii) loss of appetite, (iii) weakness, (iv) conjunctivitis and (v) constipation preceding diarrhoea (World Organization for Animal Health (OIE), 2009), often accompanied by death within 10-20 days of infection Zaabel et al., 2012). As such, CSFV has been designated a Containment Group 4 organism by the OIE and as a United States Department of Agriculture (USDA) Select Agent under the U.S. Federal Select Agent Program (OIE, 2013;Zaabel et al., 2012). A neutralizing antibody response against the structural glycoprotein E2 has been detected in pigs that recover from infection and was shown to provide protection against future infection (Hulst et al., 1993;K€ onig et al., 1995;R€ umenapf et al., 1991). Given the susceptibility of infection across the Suidae family, high transmissibility, and lack of palliative care for infected animals, the primary strategies for containing and limiting outbreaks are limiting exposure, culling or vaccination.
Modified live virus (MLV) single-dose vaccines based on attenuated virus strains are safe and elicit rapid, protective immune responses in na€ ıve pigs, and are commonly used in CSFV endemic areas to protect livestock and contain outbreaks (Blome et al., 2006;Van Oirschot, 2003). While providing protection against CSFV infection, animals vaccinated with the most current MLV vaccines are immunologically indistinguishable from wild-type CSFV-infected animals. Consequently, pork-producing countries employing MLV vaccination are negatively impacted by international trade restrictions designed to avoid import of potentially infected pigs into CSFV-free zones. Thus, a crucial disease control and economic aspect in development of next-generation CSFV vaccines is to allow for Differentiation of Infected from Vaccinated Animals (DIVA) to support demonstration of freedom from disease.
As an alternative to MLV vaccines, subunit vaccines only require an immunogenic portion of the target virus, permitting vaccine development in lower containment laboratories, and can be designed to develop a DIVA strategy for CSFV control. In addition to DIVA capability, optimally formulated subunit vaccines, can provide additional advantages such as: (i) single-dose vaccination with long-lasting protection, (ii) enhanced vaccine stability under field conditions and (iii) low-cost production. Indeed, CSFV subunit vaccines generated with purified recombinant E2 (expressed in a baculoviral/insect cell system without a transmembrane domain (European Medicine Agency, 2000) are well established and have been licensed for commercial use. However, several product profile limitations have been recognized for the baculovirus-derived CSFV vaccine including: (i) a slow protective response, (ii) limited protection against vertical transmission and (iii) viral shedding from vaccinates Zaabel et al., 2012). Towards development of an improved CSFV subunit vaccine, Madera et al. recently demonstrated protection against CSFV infection in pigs receiving a single dose of a baculoviral/insect cell-based E2 histidine-tagged subunit vaccine when formulated in the new KNB adjuvant (Galliher-Beckley et al., 2015;Madera et al., 2016). Unfortunately, antigen production using the baculovirus/insect cell expression system remains expensive at manufacturing scale and thus limits widespread deployment of this particular subunit vaccine. With this in mind, we chose to explore an alternative expression platform (Agrobacterium-mediated transient protein production in Nicotiana benthamiana plants) that has the potential to significantly reduce the cost to manufacture a novel CSFV E2 subunit vaccine (Nandi et al., 2016).
Plants have been used as the production host for recombinant vaccines and other therapeutics for over 25 years (Hiatt et al., 1989;Mason et al., 1992). Transient protein expression in N. benthamiana has become one of the main platforms used to produce plant-made therapeutics, with manufacturing facilities available to support clinical development and commercial launch (Holtz et al., 2015). This platform is well-adapted to the manufacture of veterinary products because it: (i) is easy to implement, (ii) provides flexible scales of production, (iii) offers rapid development timelines and (iv) has been demonstrated to produce products at lower cost than traditional cell culture-based expression platforms (Holtz et al., 2015;Kolotilin et al., 2014;Nandi et al., 2016;Shahid and Daniell, 2016). In fact, the first plant-made veterinary product receiving regulatory approval from the USDA Center for Veterinary Biologics in 2006 was a Newcastle disease virus poultry vaccine (Vermij and Waltz, 2006). CSFV E2 antigen has previously been expressed in and purified from plant cells and was shown to trigger a specific immune response in nontarget animal models (Legocki et al., 2005;Marconi et al., 2006;Yiu et al., 2013) and in pigs (Jung et al., 2014). However, no study has been published to date on the protective effect of a plant-made CSFV E2 antigen against CSFV challenge in pigs. Herein, we describe the design and production of a N. benthamiana-produced recombinant CSFV E2 antigen, and present data assessing immunogenicity, safety and efficacy of the antigen against wild-type CSFV challenge in pigs.

Production of E2 antigens
Two CSFV E2 antigen vaccine candidates were initially produced: (i) Transmembrane domain-deleted E2 (GenBank Acc no. ACL98470.1, amino acids 1027-1063), and (ii) Transmembrane domain-deleted E2 fused to a StrepII tag (E2-StrepII). These deletions were made to increase the yield, purity and solubility of the recombinant E2 antigen which are crucial characteristics for easier and economical purification of E2 and vaccine formulation. Also, insect cell-produced E2 antigens from Porcilis â Pesti (MSD Animal Health, Germany) and Kansas State University (European Medicine Agency, 2000; Madera et al., 2016) were also produced as transmembrane domain-deleted antigens. Both CSFV E2 antigens were expressed in plants by vacuum agroinfiltration and purified by affinity chromatography. The expression of both antigens was estimated to be between 120 and 150 mg/kg of whole plant biomass. The integrity of the antigens was first monitored by SDS-PAGE ( Figure 1). Under reducing conditions, signals at~50 kDa were detected for untagged E2, consistent with predicted molecular weights of glycosylated monomeric proteins (Figure 1a-b). Under nonreducing conditions, bands were observed at~100 kDa for both antigens (untagged E2 and E2-strepII), consistent with predicted molecular weights of glycosylated dimeric proteins (Figure 1a-b). Confirmation of E2 dimerization was important as it is a structural attribute of its antigenicity Weiland et al., 1999). Antigen identity was confirmed by western blot (data not shown), Liquid-Chromatography Mass Spectrometry (LC-MS) analysis of intact proteins (data not shown) and tryptic fragmentation of in-gel isolated protein (Figure 1c).

Plant-made E2 induces anti-E2 IgG responses in pigs
To determine the safety and immunogenicity of N. benthamianaproduced CSFV E2 administered intramuscularly (IM), na€ ıve pigs were divided into two treatment groups (n = 6/group). The first group received 200 lg of E2-StrepII antigen plus TS6 adjuvant (Boehringer Ingelheim Animal Health, Lyon, France). The second group served as a placebo control (TS6 adjuvant alone). Initial blood samples were taken from each group 14 days prior to vaccination (D-14) and subsequent blood draws occurred prior to vaccination on the day of vaccination (D0), and after vaccination (D3, D7, D14, D21, D28, D35 and D42). For each group and timepoint, serum generated from whole blood was assayed for E2-specific IgG by ELISA (Figure 2a, b). Pigs from the first group (E2-StrepII + TS6) developed anti-E2 IgG antibodies by 14 days post vaccination (D14) and maintained a high anti-E2 IgG response throughout D42. This anti-E2 IgG immune response was comparable to the anti-E2 IgG response observed in D42 in control serum derived from a pig vaccinated with a commercial baculovirus-derived recombinant E2 subunit vaccine (Porcilis Pesti â , MSD Animal Health). This serum with a characterized anti-E2-neutralizing antibody titre (Friedrich Loeffler Institute, Germany) served as a positive control serum in the anti-E2 IgG ELISA. Serum virus-neutralizing antibody titres (VNT) measured throughout the study mirrored the anti-E2 IgG ELISA results

Plant-made E2 confers full protection against wild-type CSFV challenge in pigs
We examined the ability of N. benthamiana-derived CSFV E2 vaccine to confer protection following CSFV exposure. In anticipation of the possible USDA regulatory concerns on use of the genetically fused StrepII tag in the final product, we removed the StrepII tag to produce an untagged, transmembrane deleted, E2 antigen (pE2). We formulated pE2 with one of two adjuvants: (i) TS6 and (ii) KNB (Kansas State University, Manhattan, KS). In addition, efficacy was determined with these pE2 adjuvant formulations and an insect cell-derived E2 (iE2) KNB-adjuvanted vaccine, KNB-E2 (Madera et al., 2016). The six treatment groups are summarized in Table 1.
Pigs from groups III, IV, V and VI had detectable E2-specific antibodies as early as D14 (post single dose and one week prior to the boost immunization) (Figure 3). Anti-E2-specific antibodies generated in pigs immunized with both E2 (groups III, IV and VI) and iE2 (group V) were captured by plant-made E2 as the coating antigen in the ELISA, confirming the degree of conservation in the immunogenicity profile of both antigens. Interestingly, at D14, anti-E2 IgG levels from both group IV (TS6-pE2-2-Dose) and group VI (KNB-pE2-2-Dose) were higher than the level from group V (KNB-iE2-2-Dose), though this difference was absent by D21.
For the CSFV challenge study, treatment groups were transferred to the Kansas State University Biosecurity Research Institute and acclimated for 7 days. On D35 (or Day Post Challenge 0 [DPC0]), each pig was challenged IM with 5 9 10 5 TCID 50 (1 mL) CSFV Alfort strain. Body temperature and clinical presentation were monitored up to twice daily for the duration of the study. Three of five pigs from group II (nonvaccinated control) required euthanasia due to the severity of CSF clinical signs at DPC8, 11 and 13 (e.g. high fever, loss of body weight, diarrhoea, severe leucopenia [AE, Figure 4a]). In contrast, all pigs from the vaccinated groups (III, IV, V and VI) survived CSFV challenge ( Figure 4a) and continued to gain weight post challenge ( Figure 4b). The mean body temperatures of three vaccinated groups (IV, V and VI) remained in normal range (39-40°C) ( Figure 4c). However, animals in group III (TS6-pE2-1-Dose) exhibited transiently elevated temperatures (D38-D40/DPC3-DPC6), but these resolved by 1 week post challenge and remained in the normal range for the rest of the study. White blood cell (WBC) and lymphocyte counts from vaccinated groups (III, IV, V, VI) remained similar to that of the nonchallenge control group I (Figure 5a-b). Interestingly, group III (TS6-pE2-1-Dose) showed a brief reduction in cell counts around D38/DPC6, but WBC and lymphocyte levels rebounded to that of groups IV, V and VI prior to D43/DPC9. Taken together, these data provide evidence that both plant-and insect cell-derived E2-subunit vaccines protected pigs from CSF clinical signs and prevented weight loss, fever and leucopenia following wild-type CSFV challenge.

Vaccinated pigs developed anti-CSFV IgG and neutralizing antibodies before and after CSFV challenge
To assess the antibody response to CSFV challenge, anti-E2 IgG levels were determined by ELISA. Pigs vaccinated with either plant-or insect cell-derived E2 subunit vaccines (groups III-VI) maintained high levels of E2-specific antibodies through  Figure 5c). Particularly, pigs from group III (TS6-pE2-1-Dose) exhibited a dramatic increase in E2-specific antibodies between 6 and 9 DPC and maintained high levels through 15 DPC, despite low antibody levels for the first 6 DPC. D41/DPC6 corresponds to the time period that this group of pigs no longer exhibited clinical fever and leucopenia.
Since a crucial factor in preventing CSFV infection/persistence is the production of neutralizing antibodies that block viral infection to target cells, we measured CSFV-neutralizing antibodies from treated animals in each group (Table 2). All vaccinated groups (III, IV, V and VI) had CSFV-neutralizing antibodies on the day of challenge (D35) that persisted throughout the postchallenge phase (D50/DPC15). Remarkably, group III (TS6-pE2-1-Dose) displayed the lowest-neutralizing antibody titres at D35/DPC0, but the highest at D50/DPC15 (DPC15).
Lastly, we sought to assess and quantitate the presence of CSFV in serum and mucosal fluid following challenge. Sera or nasal swabs were used to quantify viral RNA by real-time, reverse-transcription PCR. Groups III, IV, V and VI had reduced or undetectable viral loads when compared to Group II (Table 3). In pigs from group III (TS6-pE2-1-Dose) CSFV RNA was detected in the early phase of the challenge (D38/DPC3-D40/DPC6) before becoming undetectable at D50/DPC15, suggesting early CSFV replication followed by viral clearance by the host immune system. Collectively, these results demonstrated the safety and immunogenicity of a N. benthamiana-derived CSFV E2 subunit vaccine to protect pigs from clinical disease after a virulent CSFV challenge, even after a single dose.

Discussion
Most MLV CSFV vaccines used in current vaccination programmes provide full protection against clinical disease as early as 5 days post vaccination (Graham et al., 2012) but lack DIVA capability. As such, these vaccination programmes have been discontinued in most of Europe (European Commission, 2001) and in other countries declared to be CSFV-free because of the economic impact related to import/export restrictions on pork products (Marsh et al., 2005;Morgan and Prakash, 2006). Because one cannot differentiate infected from vaccinated animals, countries that choose to vaccinate using MLVs have difficulty demonstrating freedom from disease and cannot export pigs or pork products to CSFV-free countries while the vaccination campaign is ongoing. Therefore, the use of MLV vaccines is often restricted to emergency situations only. In CSFV-free countries, disease control strategies primarily involve identification of infected animals, slaughter of all pigs within infected areas, movement restrictions and viral surveillance. Outbreaks occasionally reemerge in 'CSFV-free' countries, and the expense of control measures deployed by local and national economies becomes extensive when accounting for the costs associated with: (i) depopulation of infected pigs, (ii) preemptive slaughter of currently uninfected pigs in outbreak and neighbouring farms, (iii) movement restrictions, (iv) limited market access, (v) loss of high quality breeding stocks and (vi) ultimately loss of consumer confidence in pigs and pork products from affected areas (Meuwissen et al., 1999). It is readily apparent that to achieve CSF disease control and eradication, a vaccine must be DIVA capable. In that regard, CSFV E2 subunit vaccines were developed (Lutticken et al., 1998;Moormann et al., 2000) and marketed under the brand names Porcilis â Pesti (MSD Animal Health, Germany) and Bayovac CSFV E2 â (Bayer, Germany). DIVA diagnostic kits using ELISA for the detection of CSFV E rns antigen were developed and used as companion diagnostic tests alongside the E2-based subunit vaccine (Meyer et al., 2017). Such a diagnostic kit has been successfully evaluated with sera from animals vaccinated with the KNB-iE2 antigen (Madera et al., 2016). Although the original E2 subunit vaccine provided a singledose protection, onset of immunity was obtained only after about 10 days , and two doses were recommended to stimulate longer durations of protection. Nevertheless, development of new subunit vaccines remains a major area of research with the goal of fulfilling the following: early onset of protection with sterile immunity, long-lasting immunity (the animal life cycle in the pork industry is about 7-8 months), safety without viral shedding among animals and low-cost production. Importantly, subunit vaccines provide alternatives to protect pig populations against emerging subgenotypes for which current live attenuated CSFV vaccines are not highly protective, such as the emerging subgenotype 2.1d recently reported in India  and south-east Asia .
In this work, we demonstrated that a N. benthamiana-derived CSFV E2 vaccine protected pigs from a lethal CSFV challenge, even after a single-dose vaccination as recently demonstrated in pigs vaccinated with insect-derived E2 formulated with KNB adjuvant (Madera et al., 2016(Madera et al., , 2018 Robust anti-E2 IgG and CSFV-neutralizing antibody responses were generated in vaccinated pigs. Pig groups receiving two doses of E2 vaccine showed no clinical signs, low or no viremia in serum and nasal swabs and continuous weight gain post challenge. Pigs in the single-dose vaccine dose group exhibited transient fever post challenge, which returned to normal levels 3 days later. This transient fever correlated with transient viremia in blood and nasal swabs. Reduction in fever and viremia in the single-dose vaccination group (Group III) also correlated with the induction of high anti-E2 IgG and neutralizing antibodies. In fact, these antibody levels increased dramatically post challenge, beyond levels measured in  groups receiving two vaccine doses. One would expect this to occur whenever the challenge virus replicates and induces a fever response. Future work will closely examine the early onset of protection and viral shedding, typical criteria for evaluating complete protection in vaccinated pigs. Evaluation of those two aspects will be particularly important in an emergency situation where viral replication needs to be fully controlled and eliminated quickly. Full protection was also achieved with commercial E2 subunit vaccines at 14 days post vaccination, but offered only partial protection before 10 days post vaccination Uttenthal et al., 2001). However, these vaccines were offered as  32 lg of antigen per dose at >$2.50/dose (personal communication). Although it is possible that a higher dose could provide better protection, the cost per dose would likely become prohibitive for the end-user. In contrast, the N. benthamiana plant expression system offers extra flexibility and represents a more cost-effective alternative which can provide additional antigen per dose for a fraction of the cost of bioreactor-based manufacturing methods (Nandi et al., 2016). Indeed, the reduced cost of goods manufactured using the N. benthamiana transient platform is primarily attributed to a simple and low-maintenance upstream process, easy scale-up of production and faster development time due to the transient nature of the expression system (Holtz et al., 2015;Nandi et al., 2016). Thus, it is anticipated that the N. benthamiana plant expression system will provide single-dose DIVA-capable CSFV subunit vaccines with a lower manufacturing cost of goods. Alternatively, and to reduce manufacturing costs further, an edible vaccine formulation using non-or semi-processed plant tissue expressing the E2 antigen may be tested (Merlin et al., 2017;ProVacs-Production of Vaccines from Applied Crop Sciences, 2005). In fact, oral bait formulations of E2 antigen have been tested in Europe with some success in controlling CSF in the wild boar population (Rossi et al., 2015) and can be revisited with an improved process and formulation. Additional process evaluation and cost analysis will be needed to define the potential cost savings of using the plant expression platform for new CSFV vaccine candidates. In conclusion, this study demonstrated the safety, immunogenicity, efficacy and potentially low-cost scalability of a single-dose plant-based CSFV E2 vaccine.

Cloning and expression of recombinant E2 antigen
The genetic sequence of the CSFV E2 protein C-strain-ZJ (GenBank Acc # ACL98470.1, amino acids 689-1026 with E782G substitution) fused to the signal peptide of the barley aamylase (GenBank Acc # CAX51374.1) was codon optimized for expression in N. benthamiana plants (ATUM, Newark, CA). A second gene (E2-StrepII) was prepared with the fusion of the StrepII tag sequence WSHPQFEK on the C-terminus of the E2 sequence. All genes were inserted into a Tobacco Mosaic Virusbased proprietary expression vector (iBio Inc., New York, NY) and mobilized in E. coli 10G ELITE electrocompetent cells (Lucigen, Middleton, WI) by electroporation for vector amplification. After sequencing of the insert, expression vectors were mobilized into Agrobacterium tumefaciens strain GV3101 together with the pSOUP helper vector (Hellens et al., 2000) and cultured at 28°C, 225 rpm in Luria-Bertani media supplemented with 50 mg/L kanamycin, 25 mg/L gentamycin and 25 mg/L rifampicin. Transient expression of recombinant E2 proteins was performed by vacuum infiltration as described previously (Holtz et al., 2015). Briefly, N. benthamiana seeds were germinated on rockwool Kiem â plugs (Grodan, Roermond, The Netherlands) placed in Sure To Grow hydroponic media (Sure To Grow, Beachwood, OH) at~26°C under a 16/8 h LED light/dark photoperiod. Five-weekold N. benthamiana plants were vacuum infiltrated for 3 min at 23 in. (584 mm) Hg in 5 mM MES buffer, pH 5.5 containing overnight cultures of Agrobacteria diluted to a final OD 600 nm of 0.05. Infiltrated plants were placed in a growth chamber under constant light at approximately 24°C.

Purification of recombinant E2
The plant-derived E2 antigen was produced in N. benthamiana plants. After 7-8 days post infiltration (DPI), plants were harvested and homogenized in two volumes (w:v) of cold buffer (50 mM sodium phosphate, 150 mM sodium chloride, 70 mM ascorbic acid, 5 mM EDTA, pH 8.0) and centrifuged at 14 000 g for 15 min at 12°C. Major host cell proteins were precipitated by adjusting the pH of the extract to 4.8 with 2 M acetic acid. After a 2-min incubation, the pH of the extract was adjusted to pH 7-8. The extract was centrifuged again at 14 000 g for 15 min at 12°C and clarified through a 0.2 lm Sartopure PP3, size 4 filter (Sartorius, Bohemia, NY). For E2-StrepII antigen purification, extracts were loaded on 5 mL StrepTrap HP column (GE HealthCare Life Sciences, Piscataway, NJ) following the manufacturer's instructions. For untagged E2 antigen purification, clarified extract was loaded on a HiTrap NHS-Activated HP Sepharose column (GE HealthCare Life Sciences, Piscataway, NJ) coupled with the anti-E2 antibody WH303 (APHA, Addlestone, UK). Plant-made E2 was eluted from the immunoaffinity column  with 5 CV of 100 mM citric acid. Elution fractions were then dialysed against 1X Phosphate-buffered saline (PBS), pH 7.4 overnight at 4°C. The insect cell-derived E2 was produced and purified as described previously (Madera et al., 2016).

SDS-PAGE and western blotting
Samples collected during processing and pure E2 proteins were analysed by SDS-PAGE on 4%-10% Bis-Tris gradient NuPAGE â gel (Life Technologies, Carlsbad, CA). Samples were reduced with 2.5% b-mercaptoethanol. Proteins were either stained with SimplyBlue SafeStain (Life Technologies, Carlsbad, CA) or transferred to a nitrocellulose membrane for western blotting. Western blot membranes were first blocked 1 h with 3% bovine serum albumin (BSA) in Tris-buffered Saline (TBS) supplemented with 0.5% Tween-20 (TBST). The membrane was then probed for 1 h with WH303 anti-E2 antibody (APHA Scientific, Addlestone, UK) used at 1 : 10 000 dilution in TBST, 1% BSA. After three successive washes with TBST, the membrane was probed for 1 h with alkaline phosphatase-labelled goat anti-mouse antibody (Sigma-Aldrich, Saint Louis, MO) used at 1 : 5000 in TBST. Membranes were developed with NBT/BCIP substrate (Sigma-Aldrich, Saint Louis, MO). Stained gels and western blot membranes were imaged using a Bio-Rad Gel Doc TM XR+ imager (Bio-Rad, Hercules, CA).

Ethics statement
All animal experiments were conducted following research protocols approved by the Institutional Animal Care and Use Committees (IACUC) of Texas A&M University and Kansas State University.

Challenge study (Manhattan, KS)
All animal studies were performed under the Kansas State University (KSU) IACUC approval (IACUC# 3845). Conventional Large White-Duroc cross-bred weaned specific pathogen-free 3week-old female piglets were obtained from a commercial source. The pigs were kept in a biosafety level II (BSL2) laboratory at the Kansas State University Large Animal Research Center (LARC) prior to CSFV challenge. For the CSFV challenge phase, pigs were transferred to the biosafety level III Agriculture laboratory at the Biosecurity Research Institute (BRI, Kansas State University). Naive pigs were randomly assigned into six groups of five pigs each ( Table 1). The plant-made E2 protein was obtained from iBio CDMO and the transmembrane domaindeleted insect cell-derived E2 was expressed and purified in KSU laboratory as described previously (Madera et al., 2016). The different CSFV E2 subunit vaccine formulations were prepared by mixing the purified antigen with the adjuvant. Antigen formulations and controls were administered IM. Two groups of pigs received either one or two doses of plant-made E2 antigen (pE2) formulated with the TS6 oil-in-water emulsion (200 lg of antigen per dose in 1 mL). Group III (TS6-pE2-1-Dose) received a single dose of pE2 plus TS6 adjuvant on D0, and Group IV (TS6-pE2-2-Dose) received a priming dose of 200 lg pE2 plus TS6 adjuvant on D0, a booster dose on D21. Group V (KNB-iE2-2-Dose) received a priming dose of 75 lg of insect cell-derived E2 antigen (iE2) in the KNB oil-in-water emulsion (75 lg of antigen per dose in 2 mL) on D0, and a booster dose on D21. Group VI (KNB-pE2-2-Dose) received two doses of 75 lg of pE2 plus KNB adjuvant, one on D0 and one on D21. Group I (Adj/À), the placebo control, received 2 mL of TS6 adjuvant mixed in PBS (Phosphate-Buffered Saline) on D0, and were the only group of pigs not challenged with CSFV. Group II (À/+) served as a nonvaccinated control, received 2 mL of PBS alone on D0. On Day 35, pigs in Groups II-VI were challenged with 5 9 10 5 TCID 50 CSFV isolate Alfort (1 mL IM). Pigs clinical signs, weight and rectal temperatures were monitored every day. Sera were sampled on Day 0 (day of first vaccination) and Day 21 (day of second vaccination or 21 DPV). Whole blood, serum and nasal swabs were sampled on Day 35 (day of CSFV challenge or 35DPV/DPC0) and every 3 days after that. Total white blood cell (WBC) and leucocyte differentiation counts were carried out as described before (Madera et al., 2018).

Quantification of anti-E2 antibodies in challenged pigs (challenge study)
Anti-E2 antibodies (IgG) were determined in pig sera by ELISA. 96-well flat-bottomed microtiter plates (Corning, Corning, NY) were coated with 62.5 ng/mL of purified plant-made untagged E2. The untagged E2 protein was used in the ELISA set up to avoid the detection of antibodies raised against the StrepII tag during the pig immunization. Diluted sera (samples run in duplicate) were added and plates were incubated for 1 h at 21°C. Goat anti-porcine IgG labelled with horseradish peroxidase was used as secondary antibody (Southern Biotech, Birmingham, AL). The ELISA plates were developed with 3,3,5,5 tetramethylbenzidine (TMB) substrate (Life Technologies, Carlsbad, CA). The reactions were stopped with 2 N sulphuric acid before plates were read at 450 nm using a SpectraMAX microplate reader. Relative antibody concentrations were then calculated using the Softmax â Pro 6.4 Software (Molecular Devices, Sunnyvale, CA). Virus quantification in sera and nasal swabs by RNA isolation and real-time RT-PCR (challenge study) Viral RNA was isolated from sera and nasal swabs using the IBI Viral Nucleic Acid Extraction Kit II (IBI Scientific, Peosta, IA) following the manufacturer's instructions. Real-time RT-PCR was performed using CSFV-specific primers as previously described (Hoffmann et al., 2005). For quantification, a standard curve was prepared using serial dilutions (10 7 -10 2 ) of a passage 4 CSFV stock (10 7 TCID 50 ) before viral RNA isolation. Viremia was calculated using StepOne TM Software v2.3 (Applied Biosystems, Foster City, CA).

Serum anti-CSFV neutralization assay (challenge study)
A neutralization assay using indirect fluorescent antibody assay (IFA) as the readout method was performed to calculate the anti-CSFV-neutralizing titres in CSFV-challenged pig sera as described previously (Madera et al., 2018). Neutralizing antibody titres in serum samples were expressed as the reciprocal of the highest dilution that cause 50% neutralization.