Inactivated vaccines are currently the most widely used against Haemophilus parasuis, and these are mainly based on the combination of bivalent or multivalent vaccines of highly virulent serotypes such as type 1, 4, 5, 12 and 13, or multiple vaccines together with porcine circovirus disease, swine streptococcal disease and other porcine epidemic diseases. The results showed that high levels of neutralizing antibodies were produced about three weeks after vaccination with inactivated strains (39, 40), and that piglets born from sows after two inoculations had higher levels of maternal antibodies, which could produce early immune protection (7, 41, 42). In 2001, Takahashi (43) developed an inactivated vaccine against serotypes 2 and 5 and evaluated its safety and efficacy in laboratory and field experiments, where it showed high immune protection against infection by serotypes 2 and 5 along with significant reduction of clinical symptoms in response to challenge infection. Martin (44) prepared an inactivated vaccine in Spain against serotypes 2, 4 and 5 that provided 100% protection in piglets and induced higher levels of cytokines in vivo than those of the subunit vaccine group containing outer membrane protein and transferrin-binding protein. Inactivated vaccines against HPS serotypes 4 and 5 have been reported to reduce mortality and clinical symptoms in piglets after challenge with strains of serotypes 4, 5, 13 and 14; however, it did not protect piglets from serotype 12 infection (45, 46). Zhao et al. (47) assessed the efficacy of the trivalent inactivated vaccine for HPS serotypes 4, 5 and 12 in piglets, and the results showed that the protection was 100%.

Although inactivated vaccines still dominate the market, they do have some limitations. Inactivated vaccines cannot contain all disease-causing serotypes at the same time, and the cross-protection efficiency is low. According to one study (48), it is possible to provide heterologous protection against different serotypes if differences in antibody specificity and ability to induce antibody fixation of complement are taken into consideration. Cross-protection against different strains can also be achieved by choosing the proper target proteins as components of subunit vaccines.

Compared with traditional vaccines, subunit vaccines only contain part of the antigenic components of pathogenic bacteria or viruses, eliminating the possible side effects caused by irrelevant components and greatly improving the safety of vaccines (49). At present, two subunit vaccines of Haemophilus parasuis have been approved for marketing, namely Streptococcus suis-Haemophilus parasuis Dual Subunit Vaccine jointly developed by Wuhan Keqian Biological Co. Ltd., China and other units, and Porcine Circovirus type 2-Haemophilus parasuis Double Subunit Vaccines. The main antigenic HPS component in these two vaccines is the PalA protein, and the latter also contains HPS 06257 protein. These two antigenic proteins are the outer membrane proteins of Haemophilus parasuis, and the other outer membrane proteins, such as D15, OmpP2, OppA, HPS-0675 and GAPDH, have been confirmed as potential candidate antigens with good immunogenicity (49). Related studies have shown that HPS 06257 and PalA proteins are well-conserved in different HPS serotypes. These two proteins can induce high-level specific antibodies in vivo and provide good immune protection in the piglet challenge models. Importantly, the PalA protein is highly similar to the P6 protein of Haemophilus parasuis, and D15 is highly similar to D15 of Haemophilus parasuis and OMA87 of Pasteurella multocida (50, 51) as revealed by the proteomes of diverse microorganisms (52). The P6 protein is involved in the immune response during Haemophilus influenzae infection (53, 54). These characteristics provide strong support for these two proteins as effective target antigens for an HPS subunit vaccine.

In addition to the subunit vaccines that have been successfully approved for marketing, a number of other recombinant proteins associated with HPS immunogenicity are being studied for potential use in vaccines. Recombinant PilA (rPilA) binds to PK-15 cells, porcine tracheal epithelial cells, and the extracellular matrix components, laminin and fibronectin. The rPilA can react with convalescent and ultra-immune serum of Glässer’s disease patients (55). Purified rPilA elicited a robust immune response and produced strong immune protection against HPS serotype 5 challenge in a murine model. Transferrin-binding protein B (TbpB) also showed outstanding performance in a series of animal immunogenicity challenge and evaluation experiments, and was considered the most promising antigen for the formulation of a subunit vaccine with broad-spectrum protection against HPS (56–58). Jia et al. (59) screened the isolates of serotype 13, cloned the lpxC- and gmhA-related genes, and these proteins induced high levels of IgG antibodies and an immune response with secretion of IL-4, IL-10, and IFN-γ in mice. They found that immunization with GmhA and LpxC together could stimulate the production of both Th1 and Th2 immune responses, while recombinant LpxC and inactivated bacteria could only produce a Th2 immune response. In terms of protection from HPS challenge in mice, immunization with recombinant LpxC or GmhA individually resulted in 50% protection, while the combination, LpxC + GmhA, provided 60% protection against infection with a lethal dose of HPS. Dai et al. (60) tested the immunoprotective effect of vaccination with the recombinant polyamine transporter, PotD, in mice and showed that this protein could effectively stimulate both humoral and cellular immune responses. PotD immunization enhanced lymphocyte proliferation, and triggered a Th1-type immune response, protecting the mice from a lethal HPS infection and possibly conferring resistance to HPS colonization. Álvaro et al. (61) evaluated the protective effect of a vaccine containing the three recombinant proteins, rOmpP2, rOmpP5 and rOmpD15, against infection by the HPS Nagasaki strain in piglets lacking colostrum. All three recombinant proteins were recognized and induced specific antibodies in pig serum, but they were not sufficient to protect pigs from HPS challenge under experimental conditions.

With the development of next-generation sequencing technology, availability of a host of complete genomes has made it possible to use the strategy of reverse vaccinology for development of subunit vaccines on a large scale. The basic strategy of reverse vaccinology begins with the bioinformatics analysis of genome-wide genetic information of the target pathogens, and the prediction of potential antigens. The main steps include screening and analysis of genomic information (62), identification of open reading frames with unknown functions (63, 64), homology alignment, subcellular localization of unknown functional ORF-encoded proteins, and functional annotation to identify surface proteins and toxin proteins (65). Comparative genomics and pan-genomics (66, 67) can help researchers respond more efficiently to genetic mutations and immune evasion due to differences between pathogen strains. The precursor LolA protein which comprises the HPS outer membrane lipoprotein carrier protein, and the two HPS outer membrane proteins, RlpB and VacJ (68, 69), were identified by bioinformatics methods as candidates for immunoprotective studies of Haemophilus parasuis. The results showed that mice could produce high levels of IgG antibodies in response to immunization with recombinant LolA, resulting in 50% protection against challenge with the virulent strain HPS01, serotype 13. LolA was found to induce Th1 and Th2 immune responses in mice as shown by the level of cytokines of IL-4, IL-10 and interferon-γ. The recombinant proteins RlpB and VacJ induced strong antibody responses and high IFN-γ levels in the immune sera of inoculated animals, but they did not afford sufficient protection of the pigs in the challenge test. Li et al. (49) identified three outer membrane proteins TolC, LppC and HAPS_0926 by bioinformatics methods. Mice immunized with these three outer membrane proteins produced humoral and host cell-mediated responses with significantly increased antigen-specific IgG level and lymphoproliferative responses. CD4+ and CD8+ T cells as well as three cytokines (IL-2, IL-4, and IFN-γ) were significantly increased in immunized mice. Antisera against the candidate antigens were effective in preventing HPS from surviving in a whole-blood survival assay. It is generally perceived that multi-component subunit vaccines induce a more pronounced immune responses than single-component vaccines.

In addition to the above-mentioned candidate antigens for subunit vaccines, the antigenic proteins listed in Table 2 have also been shown to provide partial protection in HPS challenge experiments. It should be noted, however, that in the majority of these trials, the immunized animals were challenged with highly virulent strains, requiring the highest degree of protection.

Among the candidate antigens, the OmpA family of OMPs function as adhesins and invasins in the respiratory system that bind to airway epithelial cells, which constitute the principle defensive barrier in the lung. Their detection of pathogens like HPS through Toll-like receptors results in the activation of signaling pathways and release of antimicrobial and pro-inflammatory molecules. VtaA was selected as a candidate antigen for its passenger domains with an extensive mosaic structure and serum cross-reactivity among VtaA from different strains. Most of the remaining candidate antigens that have high immunogenicity were identified through bioinformatics screening.

Subunit vaccines are being widely studied, because they have the advantage over live vaccines of avoiding the risk of HPS dissemination and potential side effects from viral components. Candidate proteins play vital roles in nutrient uptake and virulence, so blocking them can repress HPS infection (70), but there are also problems such as poor immunogenicity and cross-protection against heterologous serotypes. Although the subunit vaccines tested so far have shown promising protective effects in mouse and piglet models, there are other potential protective antigens that remain unexplored. Future development of multi-component subunit vaccines should look for common antigens in different serotypes for cross-protection. Extensive research is still needed, and the choice of proper adjuvant (s) to improve the immunogenicity of subunit antigens is equally important.

Bacterial ‘ghosts’ (BGs) are empty bacterial particles obtained by the evacuation of cytoplasm and nucleic acids by gentle biological or chemical perforation (71). BGs are a new type of inactivated bacterial vaccine, which maintains antigenicity but is safer because of the absence of cytoplasm and nucleic acids (72). BG vaccines can retain a host of antigens and protein epitopes with adjuvant and drug delivery properties (73), and can induce a stronger immune response because of the native conformation of the epitopes on the BG surface. Hu et al. (74) constructed a BG vaccine from the HPS serotype 5 strain using phage bacteriolytic technology which produced stronger antibody responses in immunized piglets, higher levels of IFN-γ and IL-4, and more CD4+ T lymphocytes than the inactivated vaccine. However, the researchers did not conduct an in-depth investigation of cross-protection. There are few studies on BG vaccines of Haemophilus parasuis due to the difficulty of their production and relatively high cost, although the stronger protective effect could greatly reduce livestock losses to pig farmers from Glässer’s disease.

Live attenuated vaccines are primarily based on manipulation of the main virulence factors by live-passaging or genetic deletion, which make them less- or non-pathogenic, but still immunogenic. Compared with inactivated vaccines, stable live attenuated vaccines can induce longer-lasting immune responses in the body and are important vaccine candidates for Haemophilus parasuis (75). In 2020, Eberle (19) demonstrated that the HPS HS069 mutant with deletion of the cap gene (polysaccharide biosynthetic protein and glycosyltransferase protein) had enhanced biofilm formation ability compared with the wildtype, could effectively adhere to 3D4/21 cells, and had reduced resistance to macrophage phagocytosis. This study not only confirmed the important role of capsular polysaccharides in HPS infection in piglets but also demonstrated how a genetic deletion could be used to generate an attenuated vaccine. Cytolethal distending toxin (CDT) is an important virulence determinant of many bacterial pathogens that acts by blocking the cell cycle. Zhang et al. (21) constructed an attenuated strain of HPS SC096 by deleting the CDT gene and found that the mutant exhibited reduced adherence to and invasion of porcine umbilical vein endothelial cells (PUVEC) and a porcine renal epithelial cell line (PK-15). Deletion of the rfaE gene, a core biosynthetic enzyme gene in LOS, an important virulence factor of HPS, reduced adhesion to PUVECs and PK-15 cells by 10-fold and 12-fold, respectively (76). In the study of Lin et al. (77), piglets infected with the wild-type JS0135 strain exhibited more significant tissue damage and pathological changes compared to piglets infected with a ΔCDT mutant. In addition, Lin et al. demonstrated that loss of the CDT gene cluster in JS0135 led to increased susceptibility to phagocytosis by porcine alveolar macrophages (3D4/2) than that of the wild-type strain. The two-component system senses the density modulator QseBC, which plays an important role in the virulence of the Enterobacteriaceae and Pasteurella families (78). Yan et al. (79) constructed a ΔQseBC mutant of HPS SC1401 and infected porcine alveolar macrophages and mouse alveolar epithelial cells (MLE-12) with this mutant. They found that the ability of the mutant to adhere to and invade PAM and MLE-12 was significantly reduced. In the mouse challenge model, the mortality rate of mice inoculated with HPS SC1401 was 87.5%, while that of mice inoculated with ΔQseBC was only 50%. Pathological and histological examination of spleen and lungs showed that the pathological effects of ΔQseBC were milder than those of the wild-type group, indicating that the deletion mutant weakened the virulence of HPS in mice. In addition, deletion mutants of capD, cheY, hfq, wza, and lgtF genes have all been reported to have reduced pathogenicity.

Compared with inactivated vaccines, live attenuated vaccines have the advantage of reducing the risk that inactivated vaccines are not completely inactivated and still partially infective. They can distinguish between infected and vaccinated animals, and have a relatively long immune protection period. The primary limitation of live attenuated vaccines is that the main virulence factors and associated pathways of HPS are still not fully understood. With continued in-depth research, live attenuated vaccines are expected to become more widely accepted as a highly effective vaccination strategy against Haemophilus parasuis.

Compared to subunit vaccines and live attenuated vaccines, DNA vaccines are safer, more stable, and easier to prepare. Fu et al. (80) constructed a novel DNA vaccine (pCgap) encoding GAPDH and explored its immunoprotective efficacy in mice. They showed that the DNA vaccine was highly expressed in mammalian cells and could induce significant humoral immunity as well as adaptive Th1 and Th2 responses in mice. In the mouse model the pCgap DNA vaccine provided 83.3 and 50% protection against challenge with HPS MD0322 and SH0165 strains, respectively. Since the pCgap gene is quite conserved in HPS and is widely present in 15 serotypes, the vaccine has the potential to be protective against all serotypes and is expected to be a vaccine candidate with high cross-protection efficacy. But it also has certain limitations: DNA vaccines may not elicit robust immune responses, requiring additional measures to enhance their effectiveness (81, 82).

With the development of reverse vaccinology and immunoinformatics, different types of vaccines have become more available. In addition to conventional inactivated vaccines and attenuated vaccines, multi-epitope vaccines are also developing rapidly. A methodology called pan-genomics analysis is used to identify the core genome of a pathogen and predict the B cell and T cell epitopes of the outer membrane proteins encoded by the core genome (83). With this approach, the genes encoding a multi-epitope vaccine could be identified, the antigenicity and physicochemical properties of the proteins could be predicted, and the three-dimensional structure, molecular docking, and molecular dynamics could be simulated (84, 85). Pang et al. (86) retrieved the complete genomes of 105 HPS strains and extracted 8 core genes and their protein sequences using the Roary (87) program for pan-genomic analysis. Next, they designed a multi-epitope vaccine with high scores through linker ligation according to the predicted signal peptide (4), subcellular localization (88, 89), T cell epitope prediction (90, 91), the immunogenicity and physicochemical properties of the antigen (92), and its secondary and tertiary structure (93–95). HPS strains of different serotypes or unclassifiable HPS could be combined to produce polyclonal antibodies in mice after immunization, and the serum levels of IgM + IgG and IgG1 + IgG2 in mice were significantly increased after the second and third immunization, and high levels of B cell populations, cytotoxicity, and T helper lymphocytes were observed, indicating that a cell-mediated immune response was activated. Preliminary evidence suggests that the protein is a promising vaccine candidate.

Multi-epitope vaccines are composed of epitopes of several antigens with high immunogenicity, which solves the problem of low immunogenicity caused by a single antigen and is predicted by computer simulation. Multi-epitope vaccines have been widely used in various animal and human vaccine research, and compared with traditional vaccines, the cost is low. In addition to low production cost and high safety, they can also be composed of a variety of immunogenic peptides of different pathogens to generate protective immune responses to these pathogens as well. However, there are certain limitations. Epitope prediction analysis is typically based on linear sequences and does not consider the influence of the spatial conformation of proteins. With the advancing capability for AI-assisted in-depth interdisciplinary approaches of bioinformatics and multi-omics, the application of multi-epitope vaccines for preventing Glässer’s disease has great prospects (96).