Newcastle disease (ND) is a viral avian disease that poses a significant problem to the poultry industry globally and is endemic in many countries (Ganar et al., 2014; Abolnik et al., 2017; Miller and Koch, 2020; World Organization for Animal Health (WOAH), 2022). The impact of ND on poultry production can be particularly severe in rural areas where backyard flocks are a source of income as well as a critical source of food (Martin, 1991; Awan et al., 1994; Alexander, 2000). Significant financial costs are incurred each year globally in the prevention of ND, in maintaining a ND-free status for export purposes, and/or to eradicate ND following an outbreak. Strict biosecurity measures in conjunction with good management and vaccination practices are imperative to prevent the occurrence and spread of this disease (Alexander, 2000; Dimitrov et al., 2016; Absalón et al., 2019; Miller and Koch, 2020). ND is notifiable to the World Organization for Animal Health (WOAH) and is caused by virulent strains of Avian Orthoavulavirus 1 (AOAV-1) (formerly referred to as avian paramyxovirus serotype 1 (APMV-1) and avian avulavirus-1 (AAvV)), genus Orthoavulavirus, subfamily Avulavirinae (International Committee on Taxonomy of Viruses (ICTV), 2022). The ND virus has a non-segmented negative-sense RNA genome of ca. 15,192 bp that encodes six structural proteins: fusion protein (F), hemagglutinin-neuraminidase (HN) protein, Matrix protein (M), nucleoprotein (NP), phosphoprotein (P), and large polymerase (L) protein (Miller and Koch, 2020). The F and HN glycosylated proteins are anchored in and protrude from the surface of the viral envelope, which is derived from the host, with the M protein forming a monolayer on the inside of the viral envelope. The M protein is attached to the inner surface of the viral envelope and plays a role in the assembly of ND viral particles as well as viral budding. NP is the most abundant protein and protects the RNA genome from host nucleases by its encapsidation (“nucleocapsid core”). The NP-RNA nucleocapsid core forms a complex with the P and L proteins to form a ribonucleoprotein (RNP) complex, which is involved in the transcription and replication processes of the NDV. Two additional non-functional proteins, V and W, are generated as a result of polymerase slippage of the P gene during transcription (Steward et al., 1993). The V protein could play a role in viral replication and evading the host’s innate immune system (Chu et al., 2018; Sun et al., 2019), while the function of the W protein remains unclear at this time. Following viral attachment and fusion to the host cell membranes, which is mediated by the F and HN proteins, respectively, the viral nucleocapsid complex is released into the host cell. After viral replication in the cytoplasm, viral particles are assembled and progenies bud from the cell surface (Miller and Koch, 2020). Following infection (or vaccination), the F and HN glycoproteins are the primary targets for neutralizing antibodies (Arora et al., 2010; Kim et al., 2013).

ND has a wide host range, with more than 200 species of wild birds and poultry species susceptible to infection (Alexander, 2000; Miller and Koch, 2020; World Organization for Animal Health (WOAH), 2022). Viral transmission occurs through shedding via the oropharyngeal and cloacal routes, whereafter susceptible birds can be infected either by inhaling or ingesting the virus. Clinical signs of ND vary greatly depending on the strain and host species, and can range from a drop in feed and water intake and/or egg production in layer hens to a mortality rate of up to 100% in unvaccinated birds (Miller and Koch, 2020; World Organization for Animal Health (WOAH), 2022). The onset of symptoms usually occurs between two and fifteen days following natural exposure to NDV, although it may take up to four weeks in some cases (Miller and Koch, 2020). Depending on the clinical signs in specific-pathogen-free (SPF) chickens following experimental inoculation, four pathotypes of disease are used to describe the virulence of ND viral strains. In descending order of virulence, the pathotypes are as follows: 1) velogenic (high morbidity/mortality), 2) mesogenic (high morbidity/low mortality), 3) lentogenic (low morbidity/mortality) and 4) asymptomatic enteric (Miller and Koch, 2020). All NDV isolates are classified into a single serotype, but ND isolates are genetically and antigenically diverse and are continually undergoing evolutionary changes (Dimitrov et al., 2016; Dimitrov et al., 2019; Miller and Koch, 2020). ND isolates are phylogenetically grouped into two classes based on their full genome sequences: class I includes one genotype and isolates are predominantly of low virulence, while class II includes 20 genotypes and isolates are of low and high virulence (Dimitrov et al., 2016; Dimitrov et al., 2019). The host tropism and geographic distribution of the respective genotypes differ, with some exclusively infecting certain avian species in a limited geographic area (Dimitrov et al., 2016; Dimitrov et al., 2019). With the incidence of ND on the rise and the considerable antigenic drift of ND isolates that occurred over time (Dimitrov et al., 2017; Miller and Koch, 2020), the control of ND is critical.

Vaccination against ND is commonly practised around the world and plays a critical role in disease control. Vaccination offers protection against clinical disease and morbidity and can reduce the amount of virus shed upon exposure to a field isolate, consequently limiting the spread to new susceptible hosts. As lentogenic (and asymptomatic enteric) viral strains cause asymptomatic infection in SPF chickens, they are commonly used to develop attenuated-live and inactivated egg-produced ND vaccines for commercial poultry (Collett and Smith, 2020). Live vaccines are suitable for mass application and elicit both cellular and humoral responses, although the immunity can be short-lived (Dimitrov et al., 2017; Collett and Smith, 2020; Nurzijah et al., 2022; World Organization for Animal Health (WOAH), 2022). Inactivated vaccines are considered to be safer than live vaccines and induce good humoral responses but require direct immunization (intramuscularly or subcutaneously), as well as formulation with an adjuvant and/or multiple vaccine doses to enhance the immunogenicity. Egg-based vaccine production is a well-established and cost-effective technology, although both live and inactivated vaccine production involve the handling of infectious viral material, which poses biosafety concerns and necessitates biocontainment facilities. Furthermore, due to the genetic variation between the F and/or HN sequences of lentogenic vaccine strains and contemporary ND isolates, the capacity of these live and inactivated vaccines to confer protection and effectively reduce viral shedding in vaccinated birds is likely reduced (Miller et al., 2013; Dimitrov et al., 2017). Recently, recombinant viral vector vaccines expressing the ND HN and/or F proteins have been approved for use. Recombinant viral vector vaccines effectively stimulate both humoral and cellular immune responses, like live vaccines, but with superior safety, a shorter production time, and Differentiating Infected from Vaccinated Animals (DIVA)-potential. The production of these recombinant viral vector vaccines using the established low-cost egg-based production platform has been extensively investigated, although their production in cell culture systems is also possible, albeit at a higher production cost. For poultry vaccines, low manufacturing costs are imperative to ensure production margins are maintained (Dimitrov et al., 2017; Fulber and Kamen, 2022; Nurzijah et al., 2022; World Organization for Animal Health (WOAH), 2022).

As another alternative to traditional live and inactivated vaccines, the use of virus-like particles (VLPs) as vaccines have become increasingly popular in recent years. VLPs are self-assembling protein shells that display target epitopes of the native virus in a dense arrangement on its surface without any viral genetic material at their core, making them highly immunogenic and non-infectious (biocontainment facilities are not required) (Noad and Roy, 2003). VLPs have been produced using various prokaryotic and eukaryotic expression systems and their suitability as vaccine antigens to confer protect against various viral pathogens is well documented (Pushko et al., 2013). The omission of viral structural proteins also facilitates DIVA-compliance following vaccination of animals with VLP vaccines using appropriate serological tests (Park et al., 2014). ND VLPs have been produced using cell culture-based systems by co-expressing NDV F, HN and/or M (Pantua et al., 2006; Qian et al., 2017; Xu et al., 2019) and ND VLPs as vaccines proved to be highly effective against viral challenge in chickens in comparison to a commercial live ND vaccine (Xu et al., 2019). The use of plants to transiently produce VLP vaccines creates a scalable and cost-effective production platform that does not entail the use of viral material, endotoxins, or animal-derived reagents, and has an unprecedented emergency response capacity (D’Aoust et al., 2010; Kolitilin et al., 2014; Walwyn et al., 2015; Nandi et al., 2016; Morris et al., 2017; Tusé et al., 2020). Agroinfiltration refers to the process whereby a laboratory strain of a soil pathogen (Agrobacterium tumefaciens) is used to transport the gene of interest (incorporated into a selected plant expression vector) from the extracellular spaces in plant leaf tissue to the nucleus of the plant cell for transcription and subsequently expression of the heterologous protein, typically within 7 days (Sainsbury and Lomonossoff, 2008). As such, it is considered to be the most suited platform in the face of a pandemic or bioterrorism threat (Morris et al., 2017; Sainsbury, 2020; Tusé et al., 2020), as well as for infectious disease that entail rapidly evolving viruses and/or viruses with a great genetic and antigenic diversity (such as ND) for which the generation of antigen-matched vaccines might be required to match circulating viral strain a specific geographic region and/or a specific species. In addition to the cost-effective scalability and short production time of transient plant-based expression, this platform also facilitates N-glycosylation of the glycoprotein, which is important for the immunogenicity as well as the stability of VLPs (Kim et al., 2014). In this study the production of VLPs displaying NDV F and/or HN glycoproteins using Agrobacterium-mediated transient expression in Nicotiana benthamiana was investigated, the minimal combination of proteins required to obtain intact, functional ND VLPs was established, various parameters were optimized to maximize protein yield, and the immunogenicity of the VLPs as potential vaccines in chickens, as well as in vivo ND virus neutralization ability of the elicited antibodies, was established.

The use of transient plant-based expression of VLP vaccines against economically important infectious diseases like ND holds enormous potential for animal health (Kolitilin et al., 2014; Sainsbury, 2020; Nurzijah et al., 2022). Insect cell-produced VLP vaccines against ND have been reported (Pantua et al., 2006; Qian et al., 2017; Xu et al., 2019), but successful VLP production using transient plant-based expression have not been documented before now (Nurzijah et al., 2022). The multiple advantages of VLPs produced by transient plant-based expression over other systems led to the ND VLPs investigated in the present study.

Abundant Newcastle disease VLPs displaying at least one ND glycoprotein were produced in N. benthamiana leaves using agroinfiltration. Contrary to previous reports involving ND VLPs produced in cell culture expression systems (Pantua et al., 2006; Qian et al., 2017; Xu et al., 2019), the NDV matrix protein was not required for ND VLP formation in N. benthamiana leaves and co-expression of M with the F and/or HN proteins did not lead to a detectable increase in VLP yield. According to Nurzijah et al. (2022), relative low yields and signs of necrosis from 5 dpi were obtained with the transient expression of NDV HN using the pEAQ-HT plant expression vector. In the present study signs of necrosis was observed from 4 days post infiltration with the expression of the NDV F protein, although this was mitigated by reducing the density (OD₆₀₀) of the infiltration mixture to 1 which allowed harvesting of infiltrated leaves at the optimal day (6 dpi). Incorporating both the F and HN into the VLP as a vaccine is desirable, since protection increases when both HN and F are used together compared to the protection induced by vaccines that express F or HN alone (World Organization for Animal Health (WOAH), 2014).

The use of VLPs as vaccine antigens has become popular due to their effectiveness as immunogens, with both humoral and cellular immune responses being elicited (Bright et al., 2007; Quan et al., 2007; Landry et al., 2010). The VLPs are taken up by dendritic cells (DCs), the most potent antigen presenting cells (APCs), for processing and presentation by major histocompatibility complex (MHC) class II or I molecules for activation of CD4+ T cells (Th2 immune response) or CD8+ T cells (Th1 immune response), respectively. An adjuvant might, however, be used to enhance the immunogenicity of the vaccine, to reduce the number of immunizations (i.e., single vaccination vs. primer-boost regime) and/or to reduce the concentration of VLPs used per vaccine dose. With inactivated ND vaccines an adjuvant (commonly mineral oil-based) is employed to enhance the immunogenicity (Collett and Smith, 2020). Previously, ND HN and/or F antigens produced using either transient or stable expression in plants showed variable immunogenicity in animal studies (Nurzijah et al., 2022). In the present study the partially-purified plant-produced ND F/HN VLPs were formulated without and with a commercial adjuvant and tested for immunogenicity in SPF chickens. The addition of an adjuvant was necessary to elicit strong humoral responses, with the adjuvanted ND F/HN VLPs yielding good levels of HN-specific antibodies 14 days post vaccination. Furthermore, these antibodies demonstrated neutralising effects against two NDV isolates belonging to the homologous genotype VII and genotype XIV.

The capacity to differentiate between infected and vaccinated animals (DIVA) is of great importance for disease surveillance, diagnostic and/or export purposes (ND-free status). Due to the absence of internal viral proteins, VLP vaccines have DIVA-potential with the use of a combination of appropriate serological tests (Liu et al., 2013). In the present study, ND HN- and F-specific antibodies were detected using routine tests (HIs and an anti-NDV F commercial ELISAs, respectively) following vaccination with the plant-produced NDV F/HN VLPs. By making use of another commercial ELISA kit that targets antibodies against an internal ND protein, for example the ID Screen^® ELISA (ID.Vet, France) that detects nucleoprotein (NP)-specific antibodies, DIVA could be attained by confirming the absence of NP-specific antibodies in the sera that tested positive for F- and/or HN-specific antibodies. In practice however, antigen-matched VLP vaccines are likely to be applied in combination with live vaccines in the field, for maximum protection.

Transient plant-based expression is a cost-effective and scalable production platform. Based on the yields obtained here, conservatively estimated, one kilogram of infiltrated leaf material would be sufficient for the prime-boost vaccination of 10,000 chickens (1024 HA units (10 log₂) per dose). In the field, however, the ND VLP vaccine would likely be applied as a booster to a live vectored vaccine given in the hatchery, which would mean that up to 20,000 chickens could receive a booster from one kilogram of infiltrated leaf material. A high antigenic dose of 1024 HAU was used in the present immunogenicity study to ensure measurable antibody titres were elicited following vaccination with the ND VLPs, but the prime-boost vaccination of insect cell-produced ND VLPs with an HA titre of 256 HAU (8 log₂) was previously efficacious in chickens against viral challenge, resulting in an increased protection period, a reduction in viral load, and a shortened shedding period in comparison to a commercial live La Sota ND vaccine (Xu et al., 2019). At the latter dose, the potential number of chickens vaccinated per kilogram of infiltrated leaf material could increase up to 40,000. Furthermore, according to Ruiz et al. (2018), minimally processed VLPs could potentially be used as immunogen and would especially be beneficial for veterinary vaccines due to the cost implication of the reduction in downstream processes. Plant-produced F/HN VLPs that were only subjected to clarification yielded an HA titre of 1:512 (9 log₂) i.e., double the effective ND VLP antigenic dose reported by Xu et al. (2019), therefore, a future aim is to investigate the toxicity an immunogenicity of minimally processed plant-produced VLPs in chickens. There is also potential for mass application of VLP vaccines via in ovo (Schädler et al., 2019) and intranasal administration (Earnest, 2009) with suitable adjuvants, which would be an enormous advantage for commercial poultry production.

Some of the biggest advantages of transient plant-based expression are undoubtedly its adaptability and short production time (Morris et al., 2017; Sainsbury, 2020; Tusé et al., 2020), which facilitates the production of antigen-matched vaccines. Infectious viral avian diseases often involve rapidly evolving viruses or there is a great diversity in viruses isolated from different avian species. The capacity to tailor a vaccine for a specific geographical region and or species is a major advantage of transient plant-based expression, as an antigen-matched vaccine confers the most effective protection against clinical signs of disease and reduction in viral shedding. In South Africa, the global leader in ostrich production with at least 75% of the market share (Department of Agriculture, Land Reform and Rural Development (DALRRD), 2020), vaccination of all ostriches against ND is compulsory for export purposes. The registered egg-produced ND vaccine (Struvac^® ND Plus, Deltamune, South Africa) for ostriches is formulated with a proprietary oil adjuvant to enhance its immunogenicity, but the local adjuvant-associated adverse reactions [a well-documented occurrence (Petrovsky, 2015)] devaluate the meat or hide. In addition to the potential application of a plant-produced ND VLP for the local ostrich industry, there is also a need for pigeon-specific vaccine against NDV variants. NDV strains that infect pigeons are referred to as pigeon paramyxoviruses (PPMV) and due to the important role that pigeons may play in spreading the disease as a result of prolonged viral shedding without clinical signs, strict regulatory measures (including compulsory vaccination) are imposed for racing pigeons (Aviomed, 2019; Miller and Koch, 2020). Using transient plant-based expression, antigen-matched ND vaccines can be developed for different geographical areas and/or for individual avian species (i.e., chickens, ostriches, and racing pigeons, respectively), which would provide maximum protection against these genetic sub-lineages. Vaccination of poultry and other avian species against infectious diseases like ND is of the utmost importance to protect export markets but more importantly ensure food security in many countries where poultry provides the primary source of dietary protein.

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 reviewed and approved by Ethics Committees of the University of Pretoria (UP); the UP Animal Ethics Committee, the CSIR Research ethics committee; and the Department of Agriculture, Land Reform, and Rural Development (section 20).

CA: conceptualization. TS, CA, MO’K, NL: methodology. TS, CR, CA: investigation. CA, MO’K, NL: resources. TS: analysis, writing- original draft preparation. All authors contributed to the article and approved the submitted version.