Subunit Protein-Based Vaccines

23.3.1 Overview

The identification of immunodominant protein antigens inducing strong Th1-type immune re­sponses during the first asymptomatic stage of the disease and the demonstration of their pro­tective potential in experimental infection mod­els (mouse and target species) is central to the development of subunit-based vaccines.

23.3.2 Purified protein subunit vaccines

MAP comprises a large number of proteins, not all of which are recognized immunologi­cally (Bannantine et al., 2008, Bannantine et al., 2017). This has prompted studies to tar­get soluble or secreted MAP antigens, prefer­ably with immunologically active homologues, involved in protective responses among other pathogenic mycobacteria (Thakur et al., 2013). Those few that have been trialled in animals (Table 23.2) and used as expressed purified pro­tein subunit targets, each have varying identity to homologues in M. tuberculosis and include: MAP2121c (superoxide dismutase, SOD) a soluble exported protein associated with resist­ance to killing by intracellular host mechanisms and anti-apoptotic properties that is highly immunogenic in mice (Mullerad et al., 2002a), stimulates γδ T cells, and is thought to be impor­tant in the early stages of infection and in granu­loma formation in cattle (Shin etal., 2005); three secreted mycolyl-transferases of the Ag85 com­plex, Ag85A (MAP1609c), Ag85B (MAP0126) and Ag85C (MAP3531c), immunodominant in experimentally infected cattle and mice (Mullerad et al., 2002b; Rosseels et al., 2006a), eliciting strong T-cell responses (proliferation, IL-2 and IFN-γ) in low- and medium-shedder animals but not in culture-negative cows (Shin et al., 2005); MAP Ag74F, a fusion protein of MAP1519, a member of the PPE family, able to elicit significant IFN-γ levels in macrophages of experimentally infected calves (Nagata et al., 2005); and MAP352 7, a serine protease whose homologue in M.

Immunization of C57BL/6 mice sc. with 50 μg∕animal Ag74F in a monophosphoryl lipid A adjuvant (MPL), boosted at 3 weeks and chal­lenged with a virulent MAP cattle strain after 3 weeks showed one-log reductions in bacterial loads in spleen and liver at 16 weeks with ap­parent MAP absence in lymph nodes relative to adjuvant alone (Chen et al., 2008). However, immunization of BALB/c mice sc. with Ag74F alone without adjuvant, boosted at 3 weeks and challenged with a virulent MAP cattle strain after 3 weeks fared worse, showing no signifi­cant protection in spleen and lymph node tissue and 0.5-log decreases in liver and ileal tissue 12 weeks post-challenge (Stabel et al., 2012). In the same study when Ag74F was used simi­larly in various combination with MAP 1087, MAP1204, MAP12 72c and MAP2077c also without adjuvant, results were not signifi­cantly improved. Several parameters and dif­ferent animal breeds were used in these studies; thus, conclusions concerning adjuvant usage as a requirement or individual contributions to protection are tentative. The same Ag74F had been used in combination with SOD, and Ag85 complex units A and B in goats, immunizing sc. with 100 μg∕animal using another adjuvant (a

Table 23.2. Summary of protein vaccine candidates tested in animals.

MAP3527 MAPI519 PPE protein family

Comments

Immunized subcutaneously with 100 μg total protein of cocktail variations (three in each) and 50 μg 74F in PBS.

Kathaperumal etal., 2008

Immunized subcutaneously with mix of 100 μg of each protein in MPL or intramuscularly with MPL + 100 μg bovine IL-12 DNA. Some protection induced but no significant differences between any vaccinated groups

Kathaperumal etal., 2009

Immunized subcutaneously with mix of 100 μg of each protein with and without cationic surfactant dimethyl dioctadecyl ammonium bromide (DDA).

Table 23.2. Continued

ABC, ATP-binding cassette; MLN, mesenteric lymph node; MAP, Mycobacterium avium subsp. paratuberculosis; CFU, colony-forming units; PBS, phosphate buffered saline; MPL, monophosphoryl lipid A.

cationic surfactant: dimethyl dioctadecyl am­monium bromide; DDA), boosted at 3 weeks and challenged with a moderate oral dose of a virulent MAP cattle strain after 3 weeks. In this instance protection greater than two logs mu­cosal tissue culture from seven out of eight ani­mals at 38 weeks (Kathaperumal et al., 2009) relative to adjuvant alone was seen. Faecal shed­ding was not monitored. In a previous study (Kathaperumal et al., 2008) the Ag85 complex had been used in combination with SOD admin­istered as 100 LigZanimal sc. in MPL adjuvant using a similar regimen in calves. This combi­nation again showed some protection relative to adjuvant alone in intestinal tissue, although this was only after 18 weeks. Faecal shedding was not reduced and no commercial heat-killed vac­cine group was included as a control, making it difficult to draw accurate conclusions concern­ing the individual contribution towards protec­tion; nor did it establish the likelihood that these vaccines could improve over commercial killed whole-cell combinations, or compare the results with those of similar subunit combinations de­livered as DNA vaccine in a previous study (Chen et al., 2008).

A series of studies has been made examining the potential of a single-target subunit vaccina­tion using MAP3840, a soluble heat shock pro­tein (Hsp70, DnaK) initially chosen for its ability to induce specific immune responses in MAP- infected and MAP-vaccinated cattle and change in detection parameters with disease progression (Koets et al., 2006). Initial trials in cattle immu­nized with 200 μg sc. adjuvanted with DDA and boosted at week 44 and challenged with a series of low oral doses administered over 3 weeks was initially shown to provide a reduced pattern of faecal shedding relative to unvaccinated group over the 2-year study. Subsequent retesting in cattle, however, using a similar regimen but with extra boosts at weeks 12, 24 and 52 showed a more similar pattern of faecal shedding between vaccinated and sham vaccinated groups over a similar 2-year period (Santema et al., 2012). A third cattle experiment given therapeutically with boosting at 4, 16 and 32 weeks to cattle with naturally acquired chronic MAP infection was able to induce a modest 10-20% reduction in shedding (Santema et al., 2013). The authors attribute the differences observed to unusually high humoral component and suboptimal Th1 cell-mediated responses elicited by the subunit deployed in this vaccine, as measured by a lack of relative increases in INF-γ positive cells and INF-γ release response to Mycobacterium avium- purified protein derivative antigen stimulation in vaccinated animals. This underlines the need to better characterize the subset of immune re­sponses that are actively protective against MAP persistence and dissemination, long-term, in ru­minants, and suggests that using a single subu­nit target in vaccines may be a disadvantageous approach. It highlights the diverse and currently rather unpredictable ways in which some MAP proteins are recognized and handled by the host and intriguingly supports lines of thought in tuberculosis vaccine design that advocate the need for a mix of cell-mediated and humoral responses in an effective mycobacterial vaccine (Kawahara et al., 2019).

Overall, attempts to use purified MAP pro­tein preparations as vaccines have not been as successful as hoped. This is not to say however that this approach should be discarded. Only a very few subunit targets and types of adjuvant have been investigated to date. As the pathways of MAP pathogenesis and the immunoreactivite subunits that, when targeted, protect the vast majority of naturally infected subclinical ani­mals are better described, this method of vacci­nation could still provide insights and solutions.

23.3.3 DNA delivery subunit vaccines

This mode of vaccination involves direct intro­duction of cloned or constructed DNA regions encoding target subunit MAP proteins engi­neered to be expressed from functional eukary­otic promoters into host dendritic cells leading to subunit expression, delivery and immune pro­cessing. They are relatively stable ex vivo, easily generated from known sequences and cheap to produce. They induce a variety of humoral and cellular immune responses and have been able to offer various degrees of protection against in­tracellular pathogens with some approved by the US Food and Drug Administration (FDA) and the United States Department of Agriculture (USDA) for veterinary use (Huygen, 2006; Hobernik and Bros, 2018). The protective potential of this ap­proach was first demonstrated by immunization

of BALB/c mice sc. with 2 μg∕animal plasmid mixes encoding clone pools of 75 antigens in each attached to gold gene gun delivery beads, boosted at 3 weeks then challenged with high load wild-type MAP ip. after 2 weeks, which con­ferred up to a two-log increase in protection rel­ative to a sham plasmid control (Huntley et al., 2005). Twenty-six genes in the protective mix encoding transport/binding, membrane and vir­ulence proteins and mycobactin/polyketide syn­thases were suggested as important; however, individual contributions of respective antigens were not examined. Another combined plasmid approach immunized C57BL/6 mice using a mix of five plasmids encoding the Ag85 ABC complex (MAP0216, MAP1609c, MAP3531c) and SOD (MAP0187c) and MAP2121c (membrane pro­tein) given sc. in a total 250 μg∕animal, boosted three times over 3 weeks and challenged iv. with a high dose of a clinical MAP isolate 3 weeks after the last immunization, which gave signifi­cant reductions in spleen and liver relative to ad­juvant alone (Chen et al., 2008). Further testing comparing C57BL∕6 and BALB/c mice using a plasmid expressing MAP0586c (a transglyco- lase) given im. in a total 100 μg∕animal, boosted four times at 3-week intervals and challenged iv. with a medium dose of a MAP ATCC19698 reference strain 6 weeks after the last immuni­zation, also gave significant reductions in load in spleen and liver at 8 weeks post-challenge (Roupie et al., 2008). A more extensive study (Roupie et al., 2012) using histidine-tagged con­structs was made using eight candidate MAP genes (MAP1963c, MAP2677c, MAP163 7c, MAP0388, MAP3743, MAP3198-9,

MAP2151-52c, MAP0863-5) delivered us­ing the pV1.ns-tPA DNA vector in BALB/c and C57BL∕6 and boosted with 50 μg of purified recombinant protein. This showed some anti­gen specific interferon gamma responses but no protection in any of the constructs when challenged 6 weeks post-immunization with a medium dose of a luminescent mutant of MAP strain S23 (Table 23.3). None of the above was tested subsequently in larger animals. A single study using DNA delivery alone has been report­ed in sheep (Sechi et al., 2006). In this instance animals were immunized im. 1 mg/animal with plasmids expressing MAP3966 or BCG or Mycobacterium avium gene constructs, boosted three times at 20-day intervals then challenged 3 months later with a high oral dose of a human isolate of MAP. Culture and shedding were not performed but histopathological acid-fast stain­ing of small intestinal tissue sections was sug­gestive of increased protection in the vaccine group relative to heat-killed (Gudair) vaccine control after 1 year.

23.3.4 Attenuated live vector subunit delivery vaccines

The delivery of pathogen-specific subunit anti­gens by recombinant live heterologous vectors including attenuated bacteria (Ding et al., 2018) and replication-deficient viruses (Lauer et al., 2017) is an alternative vaccination strategy al­ready trialled in other animal and human dis­eases. The rationale for this approach relies on the properties of infectious vectors that are not an aetiological agent of disease, allowing tran­sient intracellular infection without disease and naturally adjuvant antigen processing. Optimal live vector delivery regimens can require multi­ple vaccinations usually as a prime-boost, but complete cycles of intracellular replication are not required for successful antigen delivery and presentation. Repeated exposure to the same vector is potentially an option (Gabitzsch et al., 2009); however, heterologous vectors for prime and boost are preferable. Vector combinations have included naked DNA and bacterial or viral priming, followed by a viral vector boost (Wu et al., 2018). BCG used as a bacterial prime with the recombinant modified viral boost provides some protection against M. tuberculosis/M. bo­vis infections in both humans and cattle (Dean et al., 2014; Kaveh et al., 2016; Voss et al., 2018) suggesting that this type of approach may also be suitable for MAP vaccination. BCG itself ap­pears to give some protection in mice against infection with MAP (Roupie et al., 2008) and this protection increased when BCG expressed a group of genes from an operon in a putative pathogenicity island in MAP (Heinzmann et al., 2008). Using recombinant BCG for the delivery of MAP antigens would produce problems with cross-reactivity to conventional skin tubercu­lin testing in cattle (Roy et al., 2018). Greater specificity in tuberculin testing could overcome this problem (Srinivasan et al., 2019); however,

Table 23.3. Summary of DNAvaccine Candidatestested in animals.

post-mortem tissue sections revealed absence of lesions or bacteria in the groups vaccinated with the three DNA vaccine constructs

394 T. Bull

Table 23.3. Continued

CFU₁ colony-forming units; MAP, Mycobacterium avium subsp. paratuberculosis.

Dose and route of

Testedspecies Ghallengestrain infection Comments

of New Paratubercul

alternative priming vectors, such as attenuated Salmonella (Faisal et al., 2013b), Lactobaccillus (Johnston et al., 2014), recombinant human ad­enovirus (Bull et al., 2014), simian adenovirus (Folegatti et al., 2019) and attenuated vaccinia virus (Bull et al., 2014), have the potential for bovine vaccination against MAP without induc­ing interference (Table 23.4).

A double-gene knockout attenuated strain of Salmonella enterica transformed with a gene cassette expressing MAP antigens includ­ing MAP0216(Ag85A), MAP1609c(Ag85B), MAP0187c(SOD) and MAP352 7-MAP1519 (fusion Ag74F) was used to immunize C57BL/6 mice sc. with 5 ? 10⁸ CFU/animal with a sin­gle boost at 3 weeks followed 6 weeks later by challenge with a high-dose clinical MAP strain given ip., generated a significant three-log re­duction in spleen and liver loads after 16 weeks (Chandra et al., 2012). However, when a similar construct (removing Ag74F) was immunized into goats with a similar inoculation sc. plus a single boost at 3 weeks followed 3 weeks later by oral challenge with a high oral dose of a wild­type MAP66115-98 (Faisal et al., 2013b), there was no significant decrease in tissue loads and no decrease in faecal shedding relative to a 316F control vaccine after 16 weeks. The authors speculate that lack of protection in this study could have been due to suboptimal antigen pro­cessing, low vector expression and low secretion of the subunit mycobacterial proteins able to sufficiently stimulate in a small animal model but not a large one.

To try and address the problem of mycobac­terial subunit toxicity towards the delivery vec­tor, reduction of secretion through insolubility and low subunit expression as a result of rare codon usage, common in mycobacterial tran­scription, studies have attempted to alter codon usage at source. Instead of cloning genes di­rectly from the MAP genome, DNA constructed from synthesized oligonucleotides has been used allowing engineered substitution of rare codons commonly used by mycobacteria, addition of ef­ficient terminal secretion sequences or eukary­otic processing tags and removal of insoluble transmembrane regions. An approach of this type constructed a eukaryotic codon optimized cassette encoding a fusion protein of soluble regions from four MAP antigens (MAP1589c, MAP1234, MAP2444c, MAP1235) with a 5' ubiquitin expression terminus added to opti­mize intracellular processing (AgHAV). This construct was then cloned into both a prime viral vector, a replication-deficient human ad­enovirus serotype 5 (Ad5.HAV) and a boost viral vector, a replication-deficient vaccinia Ankara (MVA.HAV). Prophylactic immunization of C57BL/6 with 10⁸ plaque-forming units (PFU)/ animal sc. Ad5.HAV, boosted with 10⁸ PFU/ animal sc. MVA.HAV at 2 weeks followed by ip. challenge with low- or high-dose MAPK103 weeks later showed a one-log decrease in spleen and liver after 6 weeks (Bull et al., 2007). Therapeutic immunization using the same dos­ing but employing a MAPK10 challenge 3 weeks prior to prime/boost regimen also gave similar protection. When trialled in cattle, animals were immunized with 10⁹ PFU/animal id. prime Ad5. HAV, boosted with 10⁹ PFU/animal id. MVA. HAV at 2 weeks followed 5 weeks later by oral challenge with high-dose clinical isolate MAP R0808. Vaccinated animals showed a signifi­cant two-log decrease in gut mucosal tissue load relative to sham viral vaccinated animals after 38 weeks. No faecal shedding was observed in the vaccinated group (Bull et al., 2014) and this represents the only vaccine thus far to have ab­rogated MAP shedding in a cattle model. Similar constructs have successfully completed Phase I safety trials and are entering Phase II trials in humans (Folegatti et al., 2019).

23.4