Structure-based antigen design: Targeting transferrin receptors to prevent respiratory and systemic disease of swine

The Pasteurellaceae family and its importance in pig production

The Pasteurellaceae family is comprised of a group of Gram-negative coccobacilli bacteria that inhabit the respiratory, oral or gastrointestinal tracts of different animal species and can cause a wide range of infectious diseases in animals and humans. The family includes at least three important and well-characterized bacterial genera; namely, Pasteurella, Actinobacillus and Glaesserellathat can efficiently colonize the respiratory mucosa of pigs and trigger a chronic respiratory disorder, pneumonia or septicemia in animals of different ages (Figure 1), resulting in important economic losses for the pig industry worldwide.

Figure1. Temporal illustration of the periods for observing the diseases in pigs caused by bacteria from the Pasteurellacea family.

Pasteurella multocida serogroup D and A are the two major capsular types that cause respiratory disease in pigs. Toxigenic strains of P. multocida (mainly from serogroup D) that secrete dermonecrotic toxin (PMT) are responsible for progressive atrophic rhinitis (PAR)(Foged et al., 1987); a pathological condition of the upper respiratory tract that is initiated in young piglets and significantly reduces the growth performance in the fattening phase (weight gain reduction of 6%) (Donkó et al., 2005). In parallel, toxigenic Bordetella bronchiseptica, which is widespread in pig production, cause nonprogressive atrophic rhinitis, a clinical manifestation with minor growth impact in diseased pigs(Jong, 2006).Although some strains of P. multocida serogroup A can also express PMT and induce PAR, this capsular type is more commonly associated with lung pathology (bronchopneumonia) (Choi et al., 2003) as a secondary agent of enzootic pneumonia caused by Mycoplasma hyopneumoniae infection(Hansen et al., 2010; Pijoan and Fuentes, 1987). However,this trend of pathogenesis appears to be variable, since some clinical strains of P. multocida serogroup A are highly virulent and able to produce necrotic and focal bronchopneumonia, diffuse fibrinous pleuritis and pericarditis without any other co-infection(Oliveira Filho et al., 2018).

In the lung milieu, Actinobacillus pleuropneumoniaecan cause porcine pleuropneumonia, an important inflammatory and highly contagious disease frequently diagnosed in intensive pig farms worldwide. A. pleuropneumoniae can be transmitted from the infected sows to their offspring during the second week of life (Vigre et al., 2002), however, the clinical disease during the sucking period or at nursery phase is suppressed by the high levels of maternal antibodies acquired from the colostrum (Krejci et al., 2005) against the homologous serovars transmitted by the sows. As a strategic colonizer of the tonsils, A. pleuropneumoniae can occasionally persist undetected in the host, due to its capacity to circumvent the induction of specific antibodies against toxins (Chiers et al., 2002; Sassu et al., 2018), contributing to maintenance of the pathogen in subclinically infected animals with potential for transmission. The most important virulence factors of A. pleuropneumoniae are the Apx toxins, which havedifferent degrees of cytotoxicity, haemolytic activity and distribution amongserovars (Frey, 1995; Sarkozi et al., 2015; Schaller et al., 2000). In the lung, the toxins cause tissue damage (APXI, II and III) and erythrocyte hemolysis (APXI and II), resulting in severe edema, inflammation, hemorrhage, necrosis, diffuse fibrinous pleuritis and pericarditis (Ajito et al., 1996; Bertram, 1985; Bosse et al., 2002; Rosendal et al., 1985). Animals that survive infection may have complete resolutionof lesions, but most frequently they retain focal necrotic areas and/or well encapsulated abscesses with overlying areas of fibrinous connective tissue (Bosse et al., 2002; Frey, 1995; Rosendal et al., 1985). Animals infected with A. pleuropneumoniae infection commonly develop pleuritis which leads to substantial losses at the slaughterhouse (Fraile et al., 2010). This lesion is very prevalent (between 25 and 50%) in several countries (Jager et al., 2012). In general, A. pleuropneumoniae cause substantial economic losses for the pig industry, mainly due to a decrease in the average daily gain and feed efficiency (Straw et al., 1989), increased mortality during the fattening phase, the cost of medication and veterinary expenses (Hunneman, 1986), as well as carcass condemnation or reduced meat yield in slaughterhouses.

Actinobacillus suis is another member of the genus Actinobacillus and an important opportunistic bacterium that can cause disease (systemic inflammatory disorder) in pigs of all ages (Ojha et al., 2010). A. suis can adhere to the tonsil’s cells through OmpA protein (Ojha et al., 2010) and reside in it as well as on the mucosal surface of the soft palate (Bujold and MacInnes, 2016; Kernaghan et al., 2012). In certain circumstances, which are poorly understood, the bacterium can trigger necrotic glossitis (Sugie et al., 2019), enteritis, abortion, mastitis, metritis, abortion, meningitis, arthritis and sepsis (MacInnes and Desrosiers, 1999). In contrast to A. pleuropneumoniae, where a vast knowledge about pathogenesis is available, little is known about the induction disease mechanisms of A. suis. Nevertheless, A. suisencodes proteins that bind to collagen I and IV, fibronectin and vitronectin (Bujold and MacInnes, 2016)and also secretes two cytotoxic toxins (APXI and APXIII) (Schaller et al., 2000) with similar function to those secreted by A. pleuropneumoniae.

In contrast to other members of the Pasteurellaceae family, Glaesserellaparasuis, formerly known as Haemophilus parasuis (Dickerman et al., 2020)causes Glasser'sdisease (GD) withoutsecreting toxins. GD is an emergent and worldwide disease that mainly affects pigs in the nursery phase and early fattening stages (Costa-Hurtado et al., 2020). G. parasuishas been classified into three groups according its virulence; and strains that are moderatelyor highly virulent can overcome the innate defenses of the host (Cerda-Cuellar and Aragon, 2008; Costa-Hurtado et al., 2013; Frandoloso et al., 2012; Olvera, 2009) and cause a severe systemic inflammatory disease in pigs. Clinically, diseasedpigs develop polyserositis, polyarthritis, meningitis, corneal opacity and optical nerve congestion (Dazzi et al., 2020), which affects the animals performance and causes significant economic losses. The transmission of the bacteria occurs very early (first week) from the infected sows to their offspring (Cerda-Cuellar et al., 2010) and during the lactation period the specific IgGs transferred through the colostrum control the G. parasuis infection and avoid the development of GD. In the nursery phase, two events are strongly associated with development of GD: (a) the maternal antibodies decrease to a level that is not able to control bacterial replication, and (b) farms that house animals from multiple sources may have multiple serovarsofG. parasuis, making it difficult have all the specific IgGs against capsular antigens required for protection. The bacteria initiate infection by adhesion and invasion of epithelial cells (Bouchet et al., 2009) and can degrade secretory IgAs (Mullins et al., 2011) to evade the specific mucosal immune response from the respiratory tract. Virulent strains are also capable of producing biofilms in vitro(Zhang et al., 2014)and forming a biofilm-like structure on the surface of the tracheal epithelium in vivo (Bello-Orti et al., 2014) which can contribute to mucosal colonization and antimicrobial resistance (Zhang et al., 2014). Interesting, we have observed recently that G. parasuis can efficiently migrate to the maxillary sinus membrane from the respiratory mucosal (manuscript in preparation), which is very vascularized and the bacteria can efficiently invade endothelial cells (Frandoloso et al., 2013). Therefore, this pathway can be an alternative door for the bacteria to produce systemic infection, avoiding the contact with the specialized phagocytic cells localized in the lung.

Serology, bacterins, polysaccharide capsules and current vaccines

Established serological schemes for classifying different groups or lineages of bacteria initially simply involved immunizing animals with killed bacteria and using the resulting sera for identifying which group or lineage a particular isolate belongs to. In bacterial species that possess extracellular polysaccharide capsules, such as many members of the Pasteurellaceae, the serological response is predominantly against the polysaccharide capsule, thus is determined by the specific arrangement of biosynthetic genes involve in capsule production. This approach has not only been useful for identifying the overall characteristics of different lineages of bacteria but led to the strategy of developing inactivatedwhole cell vaccines (bacterin), since a robust anti-capsular response is generally very effective at prevention of infection.

The initial efficacy of bacterin vaccines used in the food production industry provided the conceptual basis for developing conjugate-capsular vaccines in humans, where various safety and toxicity issues made simple bacterins an unlikely commercial product. The development of the conjugate capsular vaccine against type B Haemophilus influenzae (Hib)(Peter et al., 1991) dramatically reduced the disease, raising the prospects of virtually eliminating the disease by this pathogen (Heath, 1998). The success of the Hib vaccine led to the development of conjugate capsular vaccines against the other key human pathogens causing meningitis, Neisseria meningitidis and Streptococcus pneumoniae. The development and testing of conjugate capsular vaccines made it very clear that the protection was highly specific, with virtually no cross-protection with other capsular types.

Ongoing epidemiological studies revealed that the conjugate capsular vaccines not only were effective at reducing infection but were also eliminating carriage by the targeted capsular types (Kellner et al., 2008) which resulted in reduced disease in non-immunized individuals. Although this herd immunity phenomenon enhances the efficacy of the vaccines, it also indicates that the selective pressure on the inhabitants of the upper respiratory tract can result in vaccine escape by genetic exchanges in these naturally transformable bacteria, including acquisition of a different capsular type (Croucher et al., 2011). It is also important to consider the potential reservoir of capsular genes from the upper respiratory tract microbiota as illustrated with the more extensive analysis of related commensal species of Neisseria that were considered capsule negative but clearly do possess capsule loci that could ultimately lead to additional capsule types in N. meningitis (Clemence et al., 2018). Thus, in spite of the demonstrated efficacy of the conjugate capsular vaccines in resulting in dramatic reductions in disease incidence, there is a need for continual monitoring the epidemiology of infections and to continue to explore alternate vaccine approaches.

Although not as extensively studied, the issues raised regarding conjugate capsular vaccines likely apply to bacterins that are prepared from Pasteurellaceae species that cause important infections in food production animals. Since the protection is almost exclusively due to the anti-capsular immune response, the challenge is to select the most appropriate strains for preparation of bacterins and to ensure that the resulting bacterins are, or continue to be, appropriate for the strains that are causing disease on the farm. This challenge is often not met, as the development and use of bacterin vaccines against G. parasuis illustrates. There are fifteen serovars that have been classified as high, intermediate or low virulence (Kielstein and Rapp-Gabrielson, 1992) and most bacterins have been prepared against virulent strains that were prevalent in the region originally being targeted. However, a review of clinical disease isolates collected over a 29-year period in Brazil (Espindola et al., 2019) demonstrated that there was a poor match between available vaccines and clinical disease isolates, with nearly 18% due to ‘non-typeable’ strains that represent up to 9 new capsular serotypes. Notably, one also has to be cautious about conclusions regarding the virulence of strains as there can be loss of virulence during growth in the laboratory, that can be restored by passage in animals, especially the native host (Dazzi et al., 2020).

One of the challenges faced when developing vaccines against members of the Pasteurellaceae is the highly efficient natural transformation system in these species(Redfield et al., 2006) that provides a mechanism for vaccine escape by horizontal exchange of target variants from related residents of the upper respiratory tract microbiome. The porcine pathogens P. multocida, A. pleuropneumoniae and G. parasuis contain all the known genes involved in the natural transformation process and have the required uptake signal sequences (USS) distributed throughout their genomes(Redfield et al., 2006). Published studies have demonstrated natural transformation in A. pleuropneumoniae(Bosse et al., 2004)and G. parasuis(Li et al., 2016) which also possess an overlapping repertoire of USSs. Antigenic variants of transferrin binding protein B have been shown to be distributed amongst these two species as well as A. suis(Curran et al., 2015) but the degree to which recent exchanges contribute to this distribution is uncertain as these proteins have undoubtedly been present in ancestors of the current species. Although we are not aware of published studies demonstrating natural transformation of P. multocida, the retention of USS and intact genes required for the process is a strong indication that it has been maintained in this species. In addition to the obvious acquisition of variants in the vaccine target by horizontal exchange, there is also a subtler impact of horizontal exchange that could result in increased virulence of resident strains in the population (Watkins et al., 2015).

Although reduction or prevention of colonization is clearly effective in reducing natural infection and can eventually lead to elimination of the targeted pathogen as illustrated with the use of conjugate capsular vaccines in humans(Kellner et al., 2008), it is not sufficient to prevent disease by toxin-producing bacteria in most models of infection. The direct administration of the bacterial challenge via aerosol or intra-tracheal administration effectively delivers large doses of toxin that can result in substantial tissue pathology before the immune system can effectively deal with the bacteria introduced into the lung. The relative simplicity of direct challenge experiments and the preferred acceptance by regulatory agencies compared to ‘seeder pig’ models (Dee et al., 2018) tends to bias vaccines compositions towards inclusion of toxins, especially in the case of A. pleuropneumoniae.

Of the five recognized P. multocida serogroups (A, B, D, E and F) (Townsend et al., 2001), vaccines have been developed against porcine strains ofP. multocida serogroup A and D. The commercial vaccines contain inactivated P. multocida strains with or without inactivated toxin formulated with classical adjuvants:aluminum hydroxide (AL)-gel, diethylaminoethyl (DEAE)-dextran and alpha-d-tocopheryl acetate (vitamin E).

Since there are 18 known serotypes of A. pleuropneumoniae(Bosse et al., 2018)there is uncertainty regarding the long-term efficacy of three of the major vaccines used globally that target virulent lineages:bacterins with serovar 1, 5 and 7 strains, bacterins with serovar 2, 4 and 5 strains, or bacterin with serovar 1 and 2 strains plus detoxified toxins (APXI, II and III). The fourth major vaccine contains the three inactivated toxins plus an outer membrane protein, for which the loss or replacement by immunological variants may influence the long-term efficacy of the vaccine.

Commercial vaccines are common strategy forcontrollingG. parasuis infection. Most globally available vaccines are based on bacterins from one (SV1 or SV5) or two sorovars (SV1 and 6 or 4 and 5) of G. parasuis. The recent demonstration that there are more than the 15 capsular type of G. parasuis originally identified and that a substantial proportion of disease in Southern Brazil has been due to non-vaccine capsular types(Pires Espíndola et al., 2019), underscores the limitations of bacterin vaccines. More recently, one attenuated whole cell vaccine based on G. parasuis serovar 5 was licensed on North America. No commercial vaccines are available to prevent A. suis infection.

Autogenous vaccines are emerging as the most logical short-term solution to control disease outbreaks in farms that are using ineffective commercial vaccines (when the disease-associated strains are not covered by vaccine antigens). On the other hand, the most effective long-term solution is development of more cross-protective proteinbased vaccines such as one targeting surface transferrin receptors that is predicted to completely prevent disease by G. parasuis(Barasuol et al., 2017; Curran et al., 2015; Guizzo et al., 2018), and could potentially prevent diseases caused by A. pleuropneumoniae and A. suisas well, since these three pathogens share the same receptor to acquire iron from pig transferrin.

Targeting transferrin binding proteins: an evolutionary and historical perspective

Iron is an essential element for nearly all life forms, likely a consequence of life originating in primordial seas rich in iron nearly 4 billion years ago where it was readily available redox catalyst. In response to the continual decrease in the levels of iron in the seas and oceans over time, systems for the capture and transport of iron into the cell became essential for survival. It is likely that simple organic acids were initially sufficient for capture and transport of iron into the cell. However, as the competition grew for the decreasing levels of available iron, more complex molecules with higher binding affinities for the ferric form of iron appeared. The complex iron binding molecules, termed siderophores (Neilands, 1981), were synthesized and secreted from the cell and, after binding iron, the resulting iron-siderophore complex was captured by surface receptors and transported into cell (Figure 2). The diverse molecular structures of siderophores dictate very specific binding of the iron-siderophore complex by cognate surface receptors.

Evidence for early life on earth primarily comes from stromatolites, structured microbial communities that include cyanobacteria (Reid et al., 2000), highlighting the fact that the capacity of prokaryotes to form microbial communities developed early on. Microbial communities on modern day sea sediment were shown to consist of relatively few siderophore-producing bacteria with many ‘cheaters’ or ‘pirates’ (Figure 2) that could use the resulting iron-siderophore complexes (D'Onofrio et al., 2010).

Figure 2. Iron acquisition in microbial communities. Bacteria possessing uptake systems for specific iron-siderophores can utilize (pirate) siderophores produced by other bacteria.

Iron acquisition in microbial communities. Bacteria possessing uptake systems for specific iron-siderophores can utilize (pirate) siderophores produced by other bacteria.

Although there is considerable discussion and debate on the factors responsible for establishing and maintaining this type of relationship between siderophore producer and cheater in different ecological settings (Kramer et al., 2019), it clearly is not uncommon in microbial communities including those present on mucosal surfaces of vertebrate hosts.

The identification of bacterial surface receptors that directly bind and utilize the host glycoproteins transferrin and lactoferrin as a source of iron for growth (Gonzalez et al., 1990; Ogunnariwo and Schryvers, 1990; Schryvers and Morris, 1988a, b), indicated that there was a more efficient mechanism for non-siderophore producing bacteria to acquire iron for growth on the mucosal surfaces of the upper respiratory and genitourinary tracts. The demonstration that these receptors were essential for survival on the mucosal surface and that transferrin was present at higher levels than lactoferrin (Anderson et al., 2003; Cornelissen et al., 1998), underscore the fact that these receptors evolved for survival on the mucosal surfaces where these bacteria exclusively reside. The observation that the transferrin receptors were exquisitely specific for host transferrin (Gray-Owen and Schryvers, 1993; Schryvers and Gonzalez, 1990) has been shown to be due to selective forces on the evolution of transferrin by the receptor proteins which over a 40 million year period has defined the receptor specificity amongst primates (Barber and Elde, 2014). This also suggests that the presence of receptor homologues in pathogens of poultry (Ogunnariwo and Schryvers, 1992) indicates that these receptors have been present for over 320 million years when the ancestors of birds and mammals diverged.

An evolutionary perspective is useful to help understand several features that make transferrin receptors ideal vaccine targets.Since the bacterial species in the Pasteurellaceae, Neisseriaceae and Moraxellaceae families that possess surface transferrin receptors have been adapting to their host over a long period of time they have become dependent upon these receptors for survival (no longer possess a repertoire of siderophore receptors). The combination of efficient mechanisms of genetic exchange through natural transformation (Mell and Redfield, 2014; Redfield et al., 2006) and a pre-existing global diversity of the surface receptor proteins (Curran et al., 2015) is sufficient to counter the immune responses that may impact survival on the mucosal surface of the host. In other words, these bacteria may primarily rely on ‘shuffling the deck’ rather than mechanisms for rapidly generating escape variants. These principles, along with the realization that the receptor proteins will always be expressed in the normal iron-limited environment of the host, provide the foundation for the concept that with a limited number of our engineered vaccine antigens we could potentially eliminate A.pleuropneumoniae, A. suis and G. parasuis from their porcine host.

Although the transferrin receptors clearly arose to provide an advantage for survival in the upper respiratory and/or genitourinary tract of their vertebrate host, it also provides the bacteria access to a continual supply of iron if the mucosal epithelial barrier is breached. This explains why the bacteria from the Pasteurellaceae, Neisseriaceae and Moraxellaceae families can be primarily commensal bacteria under most circumstances yet be important pathogens of humans and food production animals (Morgenthau et al., 2013). It also underlies the rationale for targeting the transferrin receptor for prevention of invasive infection which prompted earlier efforts at development of a vaccine for prevention of meningitis and sepsis byN. meningitidis(Lissolo et al., 1995) that were abandoned after a Phase 1 trial in humans. Recent developments and renewed efforts at development of vaccines derived from the transferrin receptors illustrate that this represents a missed opportunity, as the unique attributes of these receptors make them ideal targets for vaccine development.

Structure designed vaccines for targeting transferrin receptors

The transferrin receptors in the Pasteurellaceae, Neisseriaceae and Moraxellaceae families clearly arose as a modification of the siderophore receptors in Gram-negative bacteria that preceded the transferrin receptors. In the Gram-negative bacteria the iron-siderophore complex is bound by a specific surface receptor embedded in the outer membrane, and the iron-siderophore complex has to be transported across two membranes to enter the cell, a process shown to be dependent upon thetonB gene (Bagg and Neilands, 1987; Kadner and McElhaney, 1978; Postle, 1990) (Figure 3, left section). Energy from ATP hydrolysis is transduced to the outer membrane receptor by the TonB complex to drive the transport of the iron-siderophore across the outer membrane, which is then shuttled to an inner membrane transport complex by a specific periplasmic binding protein. Transport across the inner membrane into the cell is mediated by an inner membrane transport complex.

Figure 3. Iron acquisition from transferrin by Gram-negative bacteria. The integral outer membrane receptor proteins for transporting the iron siderophore complex (left),or extracting iron from transferrin and transporting it across the outer membrane (middle and right), use energy derived from the TonB complex to drive transport. The iron-siderophore complex (left) or ferric ion (middle and right) are subsequently bound by a periplasmic binding protein and shuttled to an inner membrane ABC transport complex that uses ATP hydrolysis to transport the iron-siderophore or ferric ion into the cytoplasm.

It is likely that the ancestral transferrin receptor was comprised of a single TonB-dependent receptor protein similar to the TbpA2 receptor (Figure 3, middle section) present in some strains of P. multocida(Ogunnariwo et al., 1991) and Histophilus somni(Ekins and Niven, 2002), although TbpA2 binds to the C-lobe instead of the N-lobe and is likely a more recent development as it has only been identified in bovine pathogens. The prototypical transferrin receptor (Figure 3, right section) has a surface lipoprotein component (TbpB), with a long (> 40 amino acids) anchor peptide region that enables it to extend far from the outer membrane surface to capture iron loaded transferrin and then transfer it to TbpA, for removal and transport of iron across the outer membrane.

The TbpB protein is an attractive target for vaccine development due to its accessibility to antibodies at the cell surface and the ability to readily produce substantial quantities of soluble and functional recombinant protein in the cytoplasm of E. coli, suitable for commercial production. Once the structures of TbpB became available (Calmettes et al., 2012; Moraes et al., 2009), it became possible to design mutant TbpBs defective in binding transferrin that could potentially be superior antigens by preventing masking of important epitopes by transferrin during immunization. A mutant TbpB with substantially reduced binding of porcine transferrin due to a conservative change in the side chain of a single amino acid, Y167A, provided complete protection against infection by G. parasuis(Frandoloso et al., 2015) whereas the native TbpB and a commercial vaccine product provided vastly inferior protection from infection.

Notably, an analysis of the global diversity of the TbpB present in G. parasuis, A. pleuropneumoniae and A. suis indicated that the diversity was not strongly associated with species, geographical region or time of isolation (Curran et al., 2015) suggesting that the diversity is old (not primarily due to recent mutation) and that there has been extensive exchange between the three targeted species. In addition, there were three main clusters of diversity primarily associated with the transferrin binding N-lobe of the protein, with relatively little diversity within each cluster, but substantial variation between the clusters. These results strongly suggested that a maximum of three mutant TbpBs would be able to induce a protective immune response that would be effective against all strains of the three targeted pathogens (Figure 4). A subsequent study demonstrated an effective cross-protective response within one of the phylogenetic clusters(Guizzo et al., 2018), and that not all mutant TbpBs will provide superior protection (Figure 4). These results do not suggest that it will not be possible to develop a broadly cross-protective vaccine, but that it will require additional effort and testing to determine the composition of engineered proteins that can induce the broadly cross-protective immune response.

Figure 4. Sequence diversity of TbpBs from porcine pathogens. Maximum likelihood tree demonstrating the overall diversity of TbpBs from G. parasuis, A. pleuropneumoniae and A. suis. Leaf labels identify the strains from which TbpB sequences were obtained and indicate their species and serovar, if known (NT = Nontypeable, Unk = Unknown). The sequences are rooted by the three sequences with a white background, which are the secondary TbpB-like genes. The sequences clustered into three main groups (Group I = yellow background, Group II = green background, Group III = blue background) with high confidence. The branch support values are displayed. The efficacy of vaccines based on TbpB^Y167A which belongs to cluster III is highlighted in the right painel. Figure adapted from Guizzo et al. (2018).

Although protection from infection is the primary goal of most vaccine development programs and is the feature that most regulatory agencies are looking for, the ability to prevent or eliminate colonization is a much more effective preventative measure. Since colonization precedes natural infection, a vaccine that prevents colonization will also eliminate disease, even though it may not be effective in direct challenge models where artificially high levels of bacteria are administered to a target organ (i.e. aerosol or intra-tracheal injection of bacteria into the lungs). The fact that the transferrin receptors are required for survival on the mucosal surface, thus would virtually always be expressed unless there was a sudden supply of iron (i.e. nose bleed), makes them an ideal target for prevention of colonization. Recent studies by our group has demonstrated that immunization with a vaccine preparation containing the TbpB^Y167A protein not only prevented infection by G. parasuis but also eliminated the natural colonization by this bacterium in the pig barn (Frandoloso et al, manuscript in preparation).

In order to understand how our vaccine based on the TbpB^Y167A protein could protect pigs challenged with lethal doses of G. parasuis SV5 and SV7, we conducted several in vitro studies(Barasuol et al., 2017; Frandoloso et al., 2015; Guizzo et al., 2018). As illustrated in Figure 5, the protective mechanism provided by the TbpB^Y167A-based vaccine is entirely humoral. Briefly, immunized pigs (two-dose vaccine protocol) produce high titers of systemic IgGs capable of: (a) blocking the interaction between G. parasuis TbpB and porcine transferrin which restricts the iron uptake, and consequently, G. parasuis depletes its iron reserves and cannot replicate(Figure 5A); (b) efficiently activating the classical pathway of the complement system (Figure 5B), a powerful weapon to killing G. parasuis; and (c) efficiently opsonizing G. parasuis, increasing the efficiency and speed of phagocytosis (Figure 5C). Altogether, these results provide strong evidence that a mutant TbpB vaccine preparation could be developed to eliminate the three targeted pathogens from commercial pig barns, even ones with less effective biosecurity features.

Figure 5. Illustration of the Immunobiological mechanisms mediated by anti-TbpB IgGs. A) Neutralization of the iron uptake receptor by specific IgG anti-TbpB. B) Activation of the classical of the complement system pathway. C) Opsonophagocytosis. All these mechanisms have been demonstrated in our previous studies (Barasuol et al., 2017; Guizzo et al., 2018).

The TbpA protein (Figure 3, right segment) is also a logical target for vaccine development due to its essential role in the iron acquisition process and limited variation in sequence relative to TbpB. However, integral outer membrane proteins are not suitable for commercial production of a protein-based vaccine due to the need for lipids or detergents to maintain solubility and native conformation in aqueous solution. To overcome this limitation, we have developed an approach for displaying the surface loops of TbpB on a scaffold derived from the TbpB protein (Fegan et al., 2019) which may ultimately yield hybrid antigens that provide cross-protection against both of the proteins that comprise the transferrin receptor.

A TbpB based vaccine – one vaccine to conquer all

The experimental evidence to date indicates that there are three phylogenetic clusters of TbpB diversity among the three porcine pathogens G. parasuis, A. pleuropneumoniae and A. suis(Curran et al., 2015), and that a single engineered TbpB is sufficient to induce a fully cross-protective response within a cluster (Guizzo et al., 2018), suggesting that vaccine comprised of 3 mutant TbpBs should providecomprehensive cross-protection against all known TbpB variants (Figure 4). Since the mutant TbpBs not only provide protection from infection, but are capable of eliminating natural colonization, it should be possible to implement a properly designed vaccination program to eliminate these three pathogens from commercial pig production facilities. To date, all new sequences of TbpB variants from these three species fall within the three known phylogenetic clusters, suggesting that it is unlikely that vaccine escape can be accomplished by existing TbpB variants within these three species. However, it may be prudent to search the upper respiratory microbiome of various breeds of pigs for other bacterial species that might possess transferrin receptors with diverse TbpB variants that could serve as a potential reservoir for vaccine escape.

Although the TbpB protein is not essential for iron acquisition from transferrin under laboratory conditions, the TbpB protein has been shown to be essential for survival and disease causation in pigs (Baltes et al., 2002), suggesting that TbpA will not be able to compensate for loss of functional TbpB for evasion of the vaccine-induced immune response. Even if that were possible, designing hybrid antigen vaccines targeting TbpA (Fegan et al., 2019) could overcome this possibility, as recent results suggest that comprehensive protection can be achieved by optimizing this approach (Qamsari et al., 2020).

Published in the proceedings of the International Pig Veterinary Society Congress – IPVS2020. For information on the event, past and future editions, check out https://ipvs2022.com/en.