Classical swine fever (CSF) is a highly infectious and fatal disease that has caused enormous mortality in domestic pigs and wild boars (Blome et al., 2017; Ganges et al., 2020). The disease is caused by CSF virus (CSFV), a positive-sense, enveloped single-stranded RNA virus belonging to the Pestivirus genus within the Flaviviridae family (Postel et al., 2021). The complete genome of CSFV is ~12.3 kb long, with a single large open reading frame encoding four structural proteins (C, E^rns, E1, and E2) and eight nonstructural proteins (N^pro, p7, NS2, NS3, NS4A, NS4B, NS5A, and NS5B) flanked by two untranslated regions (UTRs), 5′-UTR and 3′-UTR (Meyers and Thiel, 1996). After CSFV infection, an antibody response is elicited against envelope glycoproteins (E2 and E^rns) and a nonstructural protein (NS3; König et al., 1995; Rau et al., 2006; Panyasing et al., 2018). Notably, envelope glycoprotein E2 is a prominent, protective antigen that can elicit neutralizing antibodies and is the primary target antigen of molecular and serological diagnostic assays used to assess vaccination effectiveness against CSFV (van Rijn et al., 1999; van Rijn, 2007). Furthermore, E^rns can also induce the neutralizing antibody production, and the antibody response to E^rns can be used as a marker to diagnose CSFV infection in pigs (König et al., 1995; Moormann et al., 2000; Lin et al., 2004).

In CSF-endemic countries, strategies involving biological safety precautions and routine vaccination using traditional live attenuated vaccines (LAVs) and subunit vaccines have been implemented to control and eradicate CSFV (Postel et al., 2018; Coronado et al., 2021). Nevertheless, failure of vaccination programs is an essential factor that impedes CSF control due to immunosuppression in pig herds caused by persistent infection with moderately or low virulent CSFV strains and the interference between vaccine types and maternally derived antibodies (MDA) in piglets (Coronado et al., 2021). These challenges contribute to reduced vaccine efficacy, resulting in sporadic outbreaks and CSF reemergence in vaccinated herds and leading to CSF-endemic disease (Pérez et al., 2012; Muñoz-González et al., 2015; Coronado et al., 2019). Hence, with the enforcement of improved surveillance systems and the advancement of efficient vaccines, developing efficient antibody detection systems is crucial for enhanced CSFV control. Considering this scenario, a suitable serological antibody assessment through extensive screening is needed to monitor the vaccination immunity of pigs against CSFV in practice (Blome et al., 2006; Schroeder et al., 2012).

Currently, the evaluation of antibody responses against CSFV is performed by serum neutralization test (SNT) and the enzyme-linked immunosorbent assay (ELISA) designed to detect antibody responses specific to the E2 or E^rns glycoprotein (Postel et al., 2018; Wang et al., 2020). Notably, detecting E^rns antibodies is a reliable diagnostic test for differentiating infected from vaccinated animals (DIVA) using marker vaccines that induce a protective immune response specific to CSFV E2 while showing an absence of specific antibodies to CSFV E^rns (Langedijk et al., 2001; Lin et al., 2005; Meyer et al., 2017). Given this context, traditional immunoassays, such as SNT or ELISA, can validate DIVA properties and vaccination efficacy of several marker vaccine candidates that have been recently generated, including chimeric pestiviruses CP_E2alf (Reimann et al., 2004) and Flc-LOM-BE^rns (Lim et al., 2019) or two novel chimeric viruses vGPE⁻/PAPeV E^rns and vGPE⁻/PhoPeV E^rns (Huynh et al., 2023) and E2 subunit vaccines (Gong et al., 2019; Park et al., 2021). However, these assays must be conducted in a laboratory setting and involve complex experimental procedures, significant time consumption, and substantial equipment and maintenance costs despite their high sensitivity and accuracy (Postel et al., 2018; Ganges et al., 2020; Wang et al., 2020). Therefore, a reliable and rapid DIVA assay should be considered to successfully implement marker vaccines and validate vaccination efficacy.

An immunochromatographic strip test (ICS) is the most rapid and reliable point-of-care diagnostic tool for prompt policy regarding effective disease control strategies (Manessis et al., 2022). ICS provides benefits compared to traditional immunoassays, such as affordability, rapidness, user-friendliness, and high specificity and sensitivity (Wang et al., 2020; Manessis et al., 2022). Previously, an ICS for CSFV antibody detection was developed based on the chimeric protein CnC2, generated from a specific region encoding E^rns (amino acids 109–160; assigned as Cn) and E2 (amino acids 1–176; assigned as C2) antigen epitopes of CSFV by the Escherichia coli expression system (Li et al., 2012). However, protein expression by the E. coli system can lead to protein misfolding, which may reduce antibody detection efficacy (Bai et al., 2022). Several ICS studies have been performed to improve CSFV E2 antibody detection by enhancing the protein expression system, such as the baculovirus expression system (Bai et al., 2022) or transgenic rice endosperm (Xu et al., 2022). However, ICS for CSFV E2 antibody detection can only be used to monitor vaccination efficacy and early detection of antibody responses, whereas the DIVA test for detecting E^rns antibody is lacking.

Plant-based production systems are promising platforms for recombinant protein production due to their numerous advantages over traditional expression systems (Qiu et al., 2014; Ma et al., 2015). Among these plant systems, Nicotiana benthamiana, a close relative of tobacco, has gained widespread attention for its exceptional suitability for recombinant protein expression. N. benthamiana offers several advantages for recombinant protein production (Ward et al., 2020; Hager et al., 2022; Kallolimath et al., 2023) and is fast-growing, easily cultivated, and amenable to genetic manipulation, making it an attractive host for protein expression studies. Furthermore, N. benthamiana also possesses a robust protein folding and posttranslational modification machinery, allowing for the production of complex, biologically active proteins. A key advantage of N. benthamiana is its ability to produce glycosylated proteins with mammalian-like glycosylation patterns which is particularly advantageous for producing therapeutic proteins because proper glycosylation is often critical for protein stability, functionality, and immunogenicity (Strasser, 2023). Therefore, N. benthamiana-based expression system highlight their advantages, applications, and prospects.

In a previous study, two chimeric pestiviruses, namely vGPE⁻/PAPeV E^rns and vGPE⁻/PhoPeV E^rns, were generated based on the backbone of a CSFV LAV strain combined with the glycoprotein E^rns distantly related to CSFV (Huynh et al., 2023). These two chimeric viruses were proposed as promising marker vaccine candidates, suggesting a feasible approach for establishing an enhanced serological negative marker in DIVA vaccine development for CSF. However, DIVA capability was still not assessed for these chimeric vaccine candidates.

Thus, in this study, a novel ICS was developed to rapidly and reliably detect CSFV E2 and E^rns antibodies. Recombinant E2 and E^rns proteins were transiently expressed in N. benthamiana using Agrobacterium tumefaciens. Subsequently, ICS was constructed to detect CSFV E2 and E^rns antibodies. In addition, ICS performance was compared to commercial ELISA kits for CSFV E2 or E^rns antibody detection. Furthermore, a panel of serum samples derived from pigs vaccinated with chimeric vaccine candidates was utilized, incorporating ICS to evaluate DIVA capability.

Vaccination programs employing conventional CSFV LAVs have been demonstrated to protect pigs against CSFV infection and clinical signs in early-stage vaccination (Blome et al., 2006; Graham et al., 2012). Nevertheless, these conventional vaccines cannot resolve the challenges associated with the DIVA concept and the interference of MDAs. Various CSFV marker vaccines have been developed to address this challenge, such as subunit vaccines or chimeric pestiviruses designed to fulfill the DIVA strategy for controlling and eradicating CSF (Reimann et al., 2004; Gong et al., 2019; Lim et al., 2019; Park et al., 2021; Huynh et al., 2023). With the improvement of vaccine candidates, DIVA diagnostic tools using ELISA kits for detecting CSFV E^rns antibody or monitoring the induction of neutralizing antibodies using some commercial ELISA kits for E2 antibody detection have been developed and employed as adjunctive diagnostic assessments together with the E2 subunit vaccine and chimeric pestiviruses (Meyer et al., 2017; Wang et al., 2020). However, ELISA procedures are time-consuming and require skilled personnel, which might not be suitable for rapid field testing if E2 and E^rns antibody detection needs to be evaluated simultaneously with a large sample number in the field. Colloidal gold ICS have been established and are a valuable diagnostic tool for monitoring antibodies against CSFV in vaccinated pigs (Li et al., 2012; Bai et al., 2022; Xu et al., 2022). Compared to alternative antibody detection methodologies such as SNT or ELISA, ICS is notably user-friendly, rapid, and easily transportable (Li et al., 2012; Wang et al., 2020). Thus far, ICS development has preferably focused on detecting antibodies against the CSFV E2 protein, a powerful tool for the CSFV antibody assessment (Bai et al., 2022; Xu et al., 2022). Meanwhile, ICS development for CSFV E^rns antibody detection for the DIVA diagnostic approach still has no specific innovation (Wang et al., 2020). Furthermore, considering the benefits of transient protein expression in plant-derived systems (Yamamoto et al., 2018; Laughlin et al., 2019; Park et al., 2019), a novel ICS has been developed for the simultaneous detection of dual CSFV E2 and E^rns antibodies. This study established ICS based on colloidal gold recombinant E2 and E^rns proteins derived from N. benthamiana plants to rapidly evaluate CSFV antibodies and integrate DIVA diagnostic features.

In protein manipulation, previous studies have indicated that plant-derived transient expression enhances high protein expression and yield, immunogenicity in animal experiments, and high binding affinity for antibody detection absent in certain conventional expression systems (Sohn et al., 2018; Yamamoto et al., 2018; Xu et al., 2022). Compared to the transgenic approach for CSFV E2 protein expression using baculovirus (Wei et al., 2021; Bai et al., 2022), mammalian cells (Ji et al., 2018; Chen et al., 2023), or the E. coli system (Li et al., 2012; Yi et al., 2022), the CSFV E2-expressing transgenic plant-derived system was preferable in terms of a large amount and stable expression of the target protein and inexpensiveness, as in recent studies (Sohn et al., 2018; Laughlin et al., 2019; Park et al., 2019, 2020; Xu et al., 2022). This approach facilitated the purification of recombinant E2 from N. benthamiana extracts. Although the E2 protein is typically dimeric form, earlier investigations (Park et al., 2019) have confirmed a multimeric form of E2 protein expressed in plant tissue through glycosylation. A few systems have been established for recombinant E^rns protein expression in E. coli (Aebischer et al., 2013; Yi et al., 2022), baculovirus (Wei et al., 2021), or yeast (Huang et al., 2006; Luo et al., 2015), but there are several concerns regarding inappropriate posttranslational modifications, particularly the absence or low proportion of glycosylation in E^rns or E2 protein that may influence the binding affinity of antibodies (Gavrilov et al., 2011). The most important feature of CSFV E2 and E^rns is their heavy glycosylation, particularly E^rns, which contains seven putative N-linked glycosylation sites. N-glycosylation results in a complex mixture of neutral and monosialylated multiantennary N-glycans, accounting for nearly half of the molecular mass of mature glycoproteins (Branza-Nichita et al., 2004; Gavrilov et al., 2011). Interestingly, in terms of N-glycosylation, a previous study (Strasser, 2023) reported that N. benthamiana plant-based expression platforms exhibit remarkable versatility in that they can effectively accommodate transient and stable engineering of the N-glycan processing pathways without inducing cell death or generating adverse growth phenotypes associated with biomass production, protein expression or yield, features lacking in conventional expression systems. In this study, protein expression using a plant-derived system was employed as a first report for E^rns protein expression in which the yielded E^rns protein had a molecular mass of ~45 kDa, comparable to the mature E^rns protein in previous studies (Thiel et al., 1991; Bhattacharya et al., 2019). Although it was speculated that E2 and E^rns proteins would be highly glycosylated in the plant expression system, the purified proteins were successfully conjugated with colloidal gold in this study to develop the ICS, thereby enhancing their availability.

This study collected various sera to assess the sensitivity and specificity of ICS and its correlation with commercial ELISA kits. As mentioned, ICS showed high sensitivity or specificity for detecting CSFV E2 and E^ms antibodies under different vaccination or infection backgrounds, particularly E^rns detection, even at low SNT titers (Tables 1, 5), which would be difficult to confirm through ELISA kits. The performance of ICS exhibited high coincidence rates with ELISA kits detecting E2 and E^ms antibodies. However, the preference for detecting E2 antibodies was evident, particularly later when the E2 subunit vaccine was employed in this study (Supplementary Table S1). Despite ICS exhibiting sensitivity with SNT titers of 16 with serum samples from pigs vaccinated with the modified live LOM strain (Table 1), it surprisingly demonstrated high sensitivity in detecting E2 antibodies in serum samples from pigs vaccinated with commercial conventional CSFV GPE⁻ LAV or chimeric vaccines, even at lower SNT titers of 16 (Tables 5, 6; Supplementary Table S2). Although this study confirmed high sensitivity and specificity across a diverse range of sera, the sample size was small. Therefore, future studies should employ larger sample sizes to assess the reproducibility of ICS. Given the impact of storage temperature, time, and manufacturing batches on the reliability of detection, further research is essential to evaluate the repeatability and stability of ICS. In addition, because the dependence of CSFV-specific antibody responses on several factors, such as CSFV infection with different genotypes, the induction of E2-or E^rns-specific antibody (Popescu et al., 2019), and the failure of vaccination or the delay of antibody responses against CSFV induced by PRRSV (Chen et al., 2019; Li et al., 2020), a comprehensive approach involving a combination of various methods and continuous monitoring is essential to accurately determine the antibody response against CSFV.

ICS was used to evaluate the DIVA profiles of two chimeric pestiviruses, generated based on an infectious cDNA clone derived from the GPE⁻ LAV strain. These chimeric viruses possess the E^rns glycoprotein, which is phylogenetically distant from CSFV, and are potential DIVA vaccine candidates (Huynh et al., 2023). Serum samples obtained from optimal vaccination dose experiments and vaccine-efficient studies were reused to evaluate the DIVA profiles of these chimeric vaccine candidates. In the optimal vaccination dose experiment, successful detection of anti-CSFV E^rns antibodies in serum samples from conventionally vaccinated pigs was confirmed, whereas all chimeric-vaccinated pigs were negative, indicating the effectiveness of this ICS in detecting specific CSFV E^rns antibodies. Meanwhile, the presence of an E2-specific antibody indicated that antibody responses depended on the individual and the time point postvaccination (Supplementary Table S2). Significantly, the DIVA capability of chimeric vaccine candidates was validated in vaccine efficacy studies focusing on CSFV E^rns detection, by which ICS successfully distinguished chimeric-vaccinated animals from infected and conventionally vaccinated animals. The outcomes obtained provide valuable insights into the potential of this ICS to address the limitations of previous ICS that lack DIVA diagnostic capabilities with E^rns antibody detection (Li et al., 2012; Bai et al., 2022; Xu et al., 2022). Although CSFV E^rns is used as a negative marker to evaluate the feasibility of ICS for DIVA in this study, future studies should concentrate on developing a serological antibody assay capable of detecting antibodies against recombinant PAPeV E^rns and PhoPeV E^rns. This development will facilitate the implementation of marker vaccine candidates vGPE⁻/PAPeV E^rns and vGPE⁻/PhoPeV E^rns to control and eradicate CSF in the future. In addition, DIVA application with E^rns antibody detection was not feasible to evaluate CSF-free status individually in animals, a limitation of the sera used in this study. Instead, detection of E^rns CSFV-specific antibodies was suggested for implementation at the herd level, as validated in previous studies (Schroeder et al., 2012; Pannhorst et al., 2015). Therefore, further investigations are necessary to comprehensively validate ICS specificity and sensitivity across larger populations under various field conditions. Because serum samples derived from CSFV genotype 1 were mainly used in this study, evaluating its performance against diverse strains, genotypes, and its potential cross-reactivity with antibodies from other swine pathogens in further studies will enhance its reliability and applicability in practical DIVA strategies.

Until now, the gold standard for determining antibodies against CSFV is SNT. However, existing commercial ELISA kits evaluate E2 or E^rns separately, which incurs additional costs and labor and is more time-consuming. Although ICS development for antibody detection has been established, previous efforts have primarily focused on detecting CSFV E2 protein antibodies. This study is a pioneering effort to develop an ICS capable of detecting E2 and E^rns antibodies simultaneously. This ICS addresses the prior limitation of being unable to differentiate infected from vaccinated animals from previous studies (Bai et al., 2022; Xu et al., 2022). Extensive studies and long-term monitoring are also essential to evaluate the DIVA potential of subunit or chimeric vaccines that contribute significantly to CSF surveillance and control.

E2 and E^rns and antigens derived from the plant showed great potential and can be used to engineer a CSFV E2/E^rns dual-antibody ICS, which is suitable for application in the field to monitor CSFV antibodies in vaccinated pig herds. ICS not only demonstrated high correlation with commercial ELISA kits but also exhibited sensitivity and specificity for detecting E2 and E^rns antibodies. Thus, ICS is a rapid diagnostic tool for detecting antibodies against CSFV. The most vital point of ICS developed in this study is that it can satisfy the DIVA concept using two chimeric vaccine candidates, vGPE⁻/PAPeV E^rns and vGPE⁻/PhoPeV E^rns. Future work should be conducted to evaluate the stability of ICS and verify its repeatability.

The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found in the article/Supplementary material.

The animal study was approved by the Institutional Animal Care and Use Committee of the Faculty of Veterinary Medicine, Hokkaido University (approval no. 18-0038, approved on March 26, 2018, and approval no. 23-0029, approved on March 23, 2023) and performed according to the guidelines of this committee. The genetic modification experiments were performed according to the Cartagena Protocol on Biosafety under the Convention on Biological Diversity and approved by the Institutional Committee on the Safety of Genetic Modification Experiments of Hokkaido University and the Ministry of Education, Culture, Sports, Science and Technology (approval no. 2020-009 and 5-Monkashin-Dai-479, approved on June 8, 2021 and August 31, 2023, respectively). The study was conducted in accordance with the local legislation and institutional requirements.

LH: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Validation, Visualization, Writing – original draft, Writing – review & editing. E-JS: Conceptualization, Data curation, Investigation, Methodology, Project administration, Resources, Supervision, Validation, Writing – review & editing. YP: Conceptualization, Data curation, Investigation, Methodology, Writing – review & editing. JK: Conceptualization, Data curation, Investigation, Methodology, Writing – review & editing. TS: Data curation, Investigation, Writing – review & editing. TH: Data curation, Investigation, Writing – review & editing. NI: Data curation, Investigation, Writing – review & editing. S-HH: Data curation, Investigation, Writing – review & editing. H-NL: Data curation, Investigation, Writing – review & editing. YS: Conceptualization, Data curation, Investigation, Methodology, Project administration, Resources, Supervision, Validation, Writing – review & editing.