Lactobacillus Mucosal Vaccine Vectors: Immune Responses against Bacterial and Viral Antigens

Lactic acid bacteria (LAB) have been utilized since the 1990s for therapeutic heterologous gene expression. The ability of LAB to elicit an immune response against expressed foreign antigens has led to their exploration as potential mucosal vaccine candidates.

L actic acid bacteria (LAB), alongside other food-based platforms, have been utilized since the 1990s for therapeutic heterologous gene expression (1). The ability of LAB to elicit an immune response against expressed foreign antigens has led to their use as potential candidates as mucosal vaccine vectors. As vaccine vectors, they offer several attractive advantages: simple, noninvasive administration (usually oral or intranasal), the acceptance and maintenance of genetic modifications, low cost, and high safety levels. LAB tend to elicit minimal immune responses against themselves, instead inducing high levels of systemic and mucosal antibodies against the expressed foreign antigen following uptake via the mucosal immune system (2).
LAB for use as vaccine vectors generally include Streptococcus gordonii, Lactococcus lactis, or multiple Lactobacillus species. S. gordonii has generally fallen out of use, with a few exceptions (3). L. lactis and Lactobacillus spp. have continued to grow in use, with the number of publications continuing to increase. Several excellent reviews of L. lactis vaccines have been published (4)(5)(6), as well as articles describing how to generate these recombinant bacteria (7). Because of the large number of recent articles detailing lactobacilli as vaccine vectors, this review focuses on those publications and on the resulting immune responses generated in vivo.
Briefly, this review is divided into sections corresponding to the pathogen/disease of interest (virus, bacterium). Pathogen species or families that have been investigated in multiple studies (i.e., human immunodeficiency virus [HIV], Escherichia coli) are then highlighted, focusing on the immune responses resulting from Lactobacillus vaccination. This review covers only research involving Lactobacillus strains with heterologous gene expression. Studies conducted with unmodified Lactobacillus used either as an adjuvant or for intrinsic antibacterial or antiviral properties are excluded (8,9). The text of this review focuses on in vivo immune responses and on selected in vitro studies with checked only for antibodies in the serum and not in the mucosa (30). Because of the observed therapeutic effect seen in several studies, a human trial using cervical cancer (cervical intraepithelial neoplasia grade 3 [CIN3]) patients was conducted and demonstrated the presence of E7-specific lymphocytes in cervical tissues but not in blood, with the majority of patient tumor pathologies being downgraded (31). Taken together, the data show great promise and potential for the development of anti-HPV Lactobacillus vaccines to meet an important public health need. Influenza virus. The unpredictability of the availability of future influenza virus strains, as well as supply problems stemming from slow growth methods (egg and cell based), means that anti-influenza Lactobacillus vaccines could fill a need, particularly for treatment of infections by highly pathogenic strains such as H5N1. Shi et al. showed that oral administration of an L. plantarum strain expressing H9N2 hemagglutinin (HA) induced fecal IgA, bronchiolar IgA, and serum IgG. B cell levels in secondary lymphoid organs were increased, and CD8 ϩ T cell proliferation and IFN-␥ secretion were greatly improved relative to the levels seen with a typical influenza vaccine. Most importantly, vaccinated mice survived lethal challenge (32). These results were seen again in assays using dendritic cell-targeting peptide (DC-pep) adjuvant, which showed improved immune responses and challenge survival in chickens (33). Similar antibody and T cell results were observed in targeting H5N1 hemagglutinin (HA 1 ) in BALB/c mice (34) and chickens (35). Other influenza virus proteins have also been targeted. Chowdhury et al. granted BALB/c mice protection (via oral or intranasal administration) from multiple lethal challenge strains and showed that inclusion of cholera toxin subunit A1 (CTA1) significantly improved antibody levels and protection (36). A follow-up study showed that antibody levels and IFN-␥ secretion and proliferation, as well as protection against lethal challenge, lasted 7 months postvaccination (37).
Coronavirus. Until the recent outbreaks of severe acute respiratory syndrome (SARS) (2003) and Middle East respiratory syndrome (MERS) (2014/2015), coronavirus (CoV) morbidity and mortality were generally worse for domesticated animals rather than for humans, particularly within porcine and poultry farms. Coronaviruses usually   secretion), levels of anti-S1 and anti-N antibodies were significantly increased, even in atypically studied secretions such as ophthalmic and nasal secretions (44). Interestingly, they observed a synergy against the spike protein, but not against the nucleocapsid, in mice vaccinated against both proteins.
To improve the immune response against TGEV's core neutralizing epitope (COE), Ge et al. fused the COE with E. coli enterotoxin B (LTB), with results which showed some statistical significance, particularly with respect to splenocyte IFN-␥ and IL-4 secretion (45). In perhaps the most directly useful study, Hou et al. observed the increased presence of anti-nucleocapsid antibodies in the milk and colostrum of nursing sows, correlating with increased anti-N serum IgG levels in suckling piglets (46). A recent set of experiments by Jiang et al. delved deeper into the immune response generated by L. casei, highlighted by strong mucosa-dependent protection from infection, stimulation of the IL-17 pathway, and an imbalance between the Th1 and Th2 responses, as indicated by variations in numbers of CD4 ϩ T cells containing either intracellular IFN-␥ or IL-4 (47). Interestingly, some Lactobacillus species have been shown to downregulate IL-17 responses (48), but this simply points to the delicate balance that Th17 cells must strike between pathogen-stimulated inflammation and the potential damage of errant autoimmune inflammation (49). It is clear that homeostasis with respect to inflammation, immunity, lactobacilli, and Th17 cells is a complex subject and is dependent on a number of factors, including host genetics, pathogen, Lactobacillus strain, and adjuvants.
Rotavirus. Diarrheal disease is the second leading cause of death in children under the age of 5 worldwide, with rotavirus responsible for 40% of hospitalizations due to diarrheal illness (50). It is estimated that rotavirus killed approximately 215,000 children in 2013. The World Health Organization recommends inclusion of a rotavirus vaccine in all global vaccination protocols, and there are currently two modified live vaccines licensed worldwide (51). The global implementation is ongoing, but in countries where data are available, vaccination has resulted in a 33% reduction in hospitalization due to rotavirus morbidities. Unfortunately, both vaccines have limited (50% to 60%) efficacy in developing countries and are associated with a low-level risk of intussusception (52). A recombinant Lactobacillus-based vaccine could address the need for a subunit rotavirus vaccine that provides the benefits of a probiotic and the appropriate safety profile for use in neonates and infants. Two main avenues of lactobacillus-based rotavirus protection have been attempted in mice. The first avenue used typical oral vaccination with L. casei, inducing mucosal IgA and neutralizing serum IgG against porcine Rotavirus major protective antigen (PA) VP4 in mice (53). The second used antibody fragments to confer protection. Álvarez et al. expressed a protective antirotavirus llama antibody fragment on the surface of L. rhamnosus, protecting against diarrhea in a mouse pup model (54). Another group adapted the use of anti-rotavirus hyperimmune bovine colostrum (HBC) in the same model system, expressing an anti-HBC protein from Streptococcus, which binds HBC antibodies, thus conferring protection when orally dosed (55).
Fish-related viruses. Aquaculture is an important food supply paradigm, and with it comes the typical pathogen problems that large-scale animal farms encounter. Vaccination against fish pathogens can be performed by intraperitoneal administration (which can be cost-prohibitive), by immersion, or orally via feed, with the latter two options suffering from a lack of vaccine persistence in water and from the particularly strong mucosal tolerance observed in fish. For a comprehensive summary of vaccination attempts in fish, see the excellent review by Embregts and Forlenza (56). Lactobacillus vaccine vectors can provide an effective and easily administered system for pisciculture. The first set of studies targeted infectious pancreatic necrosis virus (IPNV), a birnavirus that afflicts rainbow trout. Direct oral administration with L. casei expressing portions of viral capsid generated significant serum IgM and afforded challenge protection in two studies by the same group (57,58). Two viruses that primarily affect carp, Cyprinid herpesvirus 3 (Koi herpesvirus [KHV]) and Rhabdovirus carpio (spring viremia of carp virus [SVCV]), have also been studied. The two antigens (KHV ORF81 and SVCV glycoprotein) were expressed together in L. plantarum and dosed orally in carp and koi. The resulting serum IgM and challenge survival data were promising, particularly for a vaccine that offers dual protection (59). Further Lactobacillus studies must be conducted, looking in particular at cellular mucosal immunity in fish, as well as at the potential for multiple pathogens to be addressed with a single modified Lactobacillus vaccine.
Other viruses. In addition to the categories already addressed, a large and diverse number of viruses have been targeted using Lactobacillus vector systems. A few are highlighted here, with the rest detailed in Table 1. Classical swine fever virus (CSFV), a flavivirus affecting pigs, has been tested in rabbits, mice, and pigs, with all tests resulting in production of serum and mucosal antibodies (60,61). Importantly, addition of thymosin ␣-1, a T cell-stimulating peptide, was able to increase levels of IgG, IgA, IFN-␥, IL-2, and tumor necrosis factor alpha (TNF-␣) in pigs (62). Porcine parvovirus has been studied in BALB/c mice and pigs, with excellent IgG and IgA responses, as well as challenge protection and virus neutralization (60,63,64). A recent study observed strong protective immune responses in chickens against Newcastle disease virus, a paramyxovirus primarily afflicting poultry, which were improved by the addition of DC-pep, which not only boosted mucosal and serum antibody levels but also increased levels of T helper cells in the spleen and peripheral blood versus the results seen with bacteria without DC-pep (65). Foot-and-mouth disease virus, a Picornavirus afflicting cloven-hooved animals, was investigated in a comprehensive dosing study that assessed anticapsid immune responses resulting from administration of recombinant L. acidophilus via the intramuscular, intraperitoneal, intranasal, or oral route. Of note, this vaccine strategy utilized the bacteria as a delivery vehicle for a capsid-expressing DNA vaccine plasmid, in contrast to utilization of expression of heterologous proteins by the bacteria. The resulting antibody responses were thus much higher via intramuscular and intraperitoneal administration than via mucosal delivery (66). As the ease of use and awareness of Lactobacillus expression systems and their abilities to induce excellent mucosal and systemic immune responses increase, the number and variety of pathogens addressed will likely increase in the future.

BACTERIA
Bacillus anthracis. Though infections are relatively rare, the prevalence of natural Bacillus anthracis in soil and its potential as a bioterrorist agent gives antianthrax vaccines some priority. Protective antigen (PA), the only antigen used in Lactobacillus vaccinations, is well studied and has been tested in other vaccine systems with various degrees of success (67). One of the earliest proof-of-concept Lactobacillus experiments involved dosing BALB/c mice with L. casei either orally or intranasally. That early study showed that the antibody responses against heterologous protein exceeded the antibody responses against the bacteria itself (68). Ten years later, Mohamadzadeh et al. combined an L. acidophilus or L. gasseri strain with DC-pep, resulting in neutralizing antibodies and challenge survival in A/J mice (69,70). That same group later observed colonic DC activation, Th17 and regulatory T cell (Treg) upregulation, and upregulation of pattern recognition receptor genes with a single vaccine dose (71).
Escherichia coli. Enteric Escherichia coli bacteria are a major cause of diarrheal morbidity and mortality, particularly for children in developing countries. The most common antigens targeted for E. coli vaccination are fimbrial proteins, which are bacterial adhesins that aid in host cell binding. Most experiments mentioned here, except one, have targeted enterotoxigenic E. coli (ETEC). A prolific group from China utilized several fimbrial protein antigens (F41, K99, K88) over several years and in several models (BALB/c, C57BL/6, BALB/c pups), all using L. casei. Among their many findings, an increase in levels of several subclasses of serum IgG (IgG1, IgG2a, IgG2b) followed oral dosing, along with increased IL-4 levels and a lesser increase of IFN-␥ levels measured by CD4 ϩ T cell enzyme-linked immunosorbent spot (ELISPOT) assays. Intestinal and bronchiolar IgA levels were increased, and challenge with standard ETEC resulted in protection of Ͼ80% of mice challenged with a lethal dose (72). The studies were repeated using intranasal dosing, which resulted in decreased intestinal IgA levels but increased bronchiolar IgA levels compared to oral delivery (73). Dosing in C57BL/6 mice induced similar IgG and IgA responses, as well as T cell proliferation and challenge protection (74). Challenge protection was conferred to mouse pups born to orally or intranasally immunized dams (75). Wu and Chung targeted two enterotoxins (ST and LT-B), rather than fimbrial proteins, with a secreted green fluorescent protein (GFP)/enterotoxin fusion protein. Similar increases in IgG and IgA levels were observed as well as challenge protection in a patent mouse gut assay (76). Ferreira et al. were the only group to target enteropathogenic E. coli (EPEC) and attempted the only sublingual dosing regimen. Experiments using L. casei expressing a portion of bacterial ␤-intimin (a cell surface protein that aids in attachment to the host cell) resulted in serum IgG and fecal IgA responses, though, interestingly, oral dosing did not generate an IgG response. Splenocytes also secreted elevated levels of IL-6 and IFN-␥, though only the results from the sublingual vaccination were reported (77). While Ferreira et al. performed their studies in C57BL/6 mice, they used C3H/HePas mice as their challenge model, due to that strain's susceptibility to Citrobacter rodentium, a commonly used strain that shares some pathology with EPEC (78). Ferreira et al. observed an increase in survival time, though animals eventually succumbed to disease.
Streptococcus pneumoniae. Most Lactobacillus experiments involving Streptococcus pneumoniae have been performed by the Oliveira laboratory and have focused on pneumococcal surface proteins (either PspA or PspC), with immunity studies conducted in C57BL/6 mice. Early work noted significant increases in bronchiolar IgA but not IgG levels following intranasal administration, with some variations due to bacterial strain differences (79). Strategies to increase antigen expression resulted in increased IgG levels (IgA levels were not measured), with enhancement of multiple IgG subsets (IgG1, IgG2a, IgG2b, IgG3). Challenge survival was improved compared to that seen with controls inoculated with saline solution alone, but no differences from the results seen with animals immunized with bacteria expressing the empty vector plasmid were observed (80). Further experiments identified a propensity for responses involving IgG1 versus IgG2a, which, along with increased IFN-␥ levels and low levels of IL-5, indicated Th1 polarization. The levels of IL-17 secretion and neutrophil recruitment in the lungs varied by route of administration, adding to the idea of the importance of the manner in which vaccines are administered and not just of their expression of antigens (81). A final set of experiments failed to induce significant levels of IgA prior to challenge, but the researchers noted that challenge with S. pneumoniae did induce a significant IgA response, which correlated with reduced bacterial loads.
Other bacteria. Very few of the large number of pathogenic bacterial species have been targeted with lactobacilli, and such studies have been reported in only a few research publications. A few are highlighted here, with the rest addressed in Table 2. Borrelia burgdorferi, the causative agent of Lyme disease, was targeted with an L. plantarum system, resulting in 100% protection following a B. burgdorferi-infected tick challenge (82). Those authors also identified what has become an interesting theme with lactobacillus vaccinations, i.e., that of dual Th1 and Th2 induction. In vitro work with human cells resulted in Th1 and Th2 cytokine responses, and oral administration in C3H-HeJ mice resulted in induction of both IgG1 (Th2) and IgG2a (Th1) (83). The same authors also targeted Yersinia pestis with L. plantarum, observing once again both inflammatory (TNF-␣, IL-12, IFN-␥, and IL-6) and anti-inflammatory (IL-10) cytokines, indicating stimulation of both Th1 and Th2 responses (84). Importantly, however, as with the previous experiment, those were human ex vivo cytokine studies whose results were not confirmed in vivo. A vaccine targeting Helicobacter pylori, a common cause of stomach ulcers, would be extremely beneficial. By targeting H. pylori adhesin Hp0410 with an L. acidophilus strain, Hongying et al. generated anti-adhesion serum IgG and intestinal IgA that reduced bacterial load and gastric inflammation following challenge (85). Antibodies against the -toxoid of Clostridium perfringens were identified in BALB/c mice following oral L. casei administration, and though the statistical significance of the antibody levels was unclear, the animals survived challenge (86).

CONCLUSIONS
In order to combat most pathogens at their main point of entry, next-generation vaccines must establish protective mucosal immunity (87). Lactic acid bacteria, particularly species of genus Lactobacillus, have shown great promise as mucosal vectors that are capable of driving both systemic and mucosal responses, especially in combination with adjuvants. The number of studies involving lactobacilli has steadily increased over the last 20 years, and as data accumulate, key concepts regarding the immune responses that these vectors elicit have emerged. Interestingly, coinduction of Th1 and Th2 cytokines points to the complexity of T cell subsets in the mucosa. A growing number of studies have suggested that T cell effector plasticity in the mucosa, especially in the gut, is the norm and that the gut must strike a balance between tolerance and inflammation (88). This appears to be one major factor arising from these Lactobacillus studies, since evidence of Th17 inflammation, as well as of Treg-based tolerance, points to a complex T cell response. In terms of mucosal vaccination, this reiterates the importance of maintaining a balanced and well-characterized approach to immunogenicity. More work must be done to identify the contributing immune pathways within the mucosa, especially the routes of bacterial uptake into immune inductive sites (M cells, DCs).
There are several major takeaways as development of LAB vaccine platforms continues. While the safety of LAB is an important strength, enhancing protective immunogenicity is a key challenge. Several studies have explored strategies to express adjuvants such as cytokines, pathogen-associated molecular patterns, toxins, and targeting molecules for M cells and DCs. A mechanistic understanding of each of these strategies is necessary to design the right combination of immunogens and adjuvants that will result in protection. The route of administration, while typically oral for LAB, can have an effect on the type of response elicited due to differences in mucosal inductive sites. The intrinsic differences between strains of lactobacilli, as well as the location of antigen expression (surface display, intracellular, secreted), can alter the resulting immune response, and the strains must therefore be properly selected for specific antigens (89). Boosting is also clearly a component of successful vaccination, and there is evidence that heterologous prime-boost strategies may improve, or at least alter, the resulting immune response (90). As always, the model system must be taken into consideration, especially in light of new evidence for mucosal immune differences between the two most common mouse models (BALB/c and C57BL/6) (91). On the basis of their safety and efficacy, as well as their overall cost, Lactobacillus vaccine vectors hold great promise as mucosal vaccines. It is anticipated that the use of clustered regularly interspaced short palindromic repeat (CRISPR)/Cas9 analysis will allow a more sophisticated approach to engineering vaccine candidates (92). Ultimately, it is critical for one of these candidates to successfully navigate the regulatory gauntlet and demonstrate efficacy in a target population.