Antibodies and Protection in Systemic Salmonella Infections: Do We Still Have More Questions than Answers?

Salmonella causes grave systemic infections in humans and other animals and provides a paradigm for other diseases in which the bacteria have both intracellular and extracellular lifestyles. New generations of vaccines rely on the essential contribution of the antibody responses for their protection. The quality, antigen specificity, and functions associated with antibody responses to this pathogen have been elusive for a long time. Recent approaches that combine studies in humans and genetically manipulated experimental models and that exploit awareness of the location and within-host life cycle of the pathogen are shedding light on how humoral immunity to Salmonella operates.

ABSTRACT

Salmonella causes grave systemic infections in humans and other animals and provides a paradigm for other diseases in which the bacteria have both intracellular and extracellular lifestyles. New generations of vaccines rely on the essential contribution of the antibody responses for their protection. The quality, antigen specificity, and functions associated with antibody responses to this pathogen have been elusive for a long time. Recent approaches that combine studies in humans and genetically manipulated experimental models and that exploit awareness of the location and within-host life cycle of the pathogen are shedding light on how humoral immunity to Salmonella operates. However, this area of research remains full of controversy and discrepancies. The overall scenario indicates that antibodies are essential for resistance against systemic Salmonella infections and can express the highest protective function when operating in conjunction with cell-mediated immunity. Antigen specificity, isotype profile, Fc-gamma receptor usage, and complement activation are all intertwined factors that still arcanely influence antibody-mediated protection to Salmonella.

INTRODUCTION

Several serovars of Salmonella enterica cause systemic diseases in humans and other animals. The global estimated burden of typhoid fever (Salmonella enterica subsp. enterica serovar Typhi) is over 21 million illnesses and 200,000 deaths, with sustained high incidence in Southeast Asia and endemic/epidemic occurrence increasingly reported in Africa (1–6). Paratyphoid fever (serovars Paratyphi A, B, and C) has an estimated 5.4 million illnesses worldwide (3). Invasive nontyphoidal Salmonella (iNTS) serovars (e.g., Salmonella enterica serovars Typhimuriun and Enteritidis) are a leading cause of lethal sepsis and severe relapsing infections in young children and immunocompromised individuals, especially in countries of the sub-Saharan African region (6–12), with an estimated 3.8 million illnesses leading to 680,000 deaths annually and very high case fatality ratios (20%) (7, 11). Antimicrobial resistance is an increasing problem in tackling many bacteria, including Salmonella (7, 11, 13–15).

WHY IS IT IMPORTANT TO GAIN A BETTER UNDERSTANDING OF ANTIBODY-MEDIATED IMMUNITY TO SALMONELLA?

There is extensive international consensus on the urgent need for better and affordable vaccines against systemic Salmonella infections. Vaccination has the potential for a high economic and health impact in fighting antimicrobial-resistant infections (16–19). Several classes of vaccines against systemic Salmonella disease have been considered in the past decades (20, 21). These differ in their ability to induce protective cell-mediated and humoral immunity with, broadly speaking, live attenuated vaccines being more efficient than nonliving preparations at eliciting Th1 type T-cell immunity known to contribute to host resistance to this bacterium (22, 23).

Despite the superior protective activity shown in animal models, live Salmonella vaccines can cause lethal infections in immunocompromised hosts (24–27). Mutants of Salmonella, including those that have been considered so far as vaccine candidates, retain virulence and can rapidly kill immune-suppressed mice (24–27). Mice that lack T-cell functions (25, 27) and mice coinfected with malaria are very susceptible to infection with live Salmonella, including with mutants that have been considered as vaccine candidates (28, 29). Efforts to identify single-gene mutations for the development of live vaccine candidates that would be completely safe (totally unable to grow to high numbers) in severely immune-deficient animals have been unsuccessful (24). This raises some concerns for the use of live vaccines in areas where Salmonella is endemic and that have a higher incidence of immune-suppressive conditions. For example, HIV and malaria are comorbidities that make humans more susceptible to systemic Salmonella infection, leading to severe, often fatal disease, and which therefore pose dangers to the use of live vaccines (30, 31). These comorbidities are widespread and epidemiologically colocalize with areas of the developing world where there is a high incidence of epidemic or endemic systemic salmonelloses and that are therefore where anti-Salmonella vaccines are most needed (9, 20, 31, 32).

Mainly for safety reasons, nonliving vaccines are currently being considered prime candidates for immunization against Salmonella diseases. The protective ability of these vaccines relies largely on the induction of antibodies. If we are to use nonliving vaccines as tools against systemic salmonelloses, it is therefore essential to rationally optimize the antibody responses induced by these preparations. This knowledge would also be immensely useful to understand how those comorbidities that impair antibody-mediated functions increase susceptibility to disease and to design vaccine strategies that can at least in part reverse these immune-suppressing conditions. This will need to be based on a clearer understanding of the qualitative and functional features of the protective antibody response against Salmonella.

This article will briefly outline the factors that influence the protective efficacy of the antibody response to systemic Salmonella infections and will embed antibody functions in the context of the location and spread of the bacteria during the infection process. This minireview will also highlight the interactions and dual requirement of T cell- and B cell-mediated immunity, both in the engenderment of antibody and T-cell responses and in the expression of in vivo resistance to the pathogen.

HOW CAN ANTIBODIES PROTECT AGAINST A PATHOGEN THAT HAS AN INTRACELLULAR LIFESTYLE?

The relative importance of antibodies and cellular immunity in host resistance to systemic Salmonella infections has been a matter of debate for a long time.

The controversy initially originated from uncertainties in the location of the bacteria within the tissues of an infected host. It is undisputable that in many animal species, including humans, Salmonella resides inside phagocytes and grows within these cells, using an armory of genes and effectors to foster its replication and evade intracellular killing (24, 33–40). Evidence from mouse models on the role of the phagocyte innate resistance gene Ity (now known as Scla11a1) in the control of bacterial division in vitro and in vivo, as well as studies where the bacteria could be seen to grow within cultured phagocytic cells, has corroborated these views (41–44). However, evidence for extracellular growth and lack of intracellular replication had also been produced, albeit based on electron microscopy studies where infections were allowed to progress to very high premortem bacterial numbers in the tissues. In these studies, intracellular bacteria were seen to undergo various stages of degradation inside macrophages and granulocytes and it was therefore inferred that Salmonella was more likely to be an extracellular pathogen and its virulence directly related to its antiphagocytic property (45).

The debate was further fueled by immunological studies in both experimental animals and humans. On the one hand, evidence was provided for the protective role of antibodies being limited to the very early stages of parenteral experimental infections, when the bacteria have not yet reached an intracellular location, with no detectable effect on their later growth in the tissues (46, 47); on the other hand, the protective effect of whole-cell nonliving typhoid vaccines, which induce mainly humoral immunity, and Vi vaccines that do not directly contain Salmonella-specific T-cell antigens was undeniable (48–53). However, protection by passive transfer of antibodies or T cells alone and by nonliving vaccines was seen only in host-pathogen interactions where the infection is naturally mild and sublethal, such as when challenging resistant mice with either virulent or weakly virulent strains or susceptible mice with weakly virulent strains (46, 50).

These discrepant views on the relative importance of humoral and cellular immunity to Salmonella were largely reconciled by studies showing that passive transfer of both antibodies and T cells, but neither alone, could protect recipient innately susceptible animals against fatal systemic Salmonella disease (54).

More recently, the essential role of both antibodies and T cells in host resistance to Salmonella has been corroborated by additional evidence obtained from human studies. For example, clinical observations indicate that lack of antibodies at young age correlates with the incidence of iNTS in African children (55, 56), despite these children acquiring Salmonella-specific CD4 Th1 cell immunity very early in life; only when antibodies are developed in addition to T-cell immunity is full protection achieved (57). This resistance is then abrogated by comorbidities that can affect either cellular or antibody-mediated functions (29, 58–62). For example, HIV, which has a profound effect on CD4-positive (CD4⁺) T-cell immunity, and malaria, which impairs phagocyte functions, increase susceptibility to systemic Salmonella infections such as iNTS. A similar increase in susceptibility is seen in patients with deficiencies in cytokines such as gamma interferon (IFN-γ) and interleukin 12 (IL-12) and in their receptors, leading to impaired activation of phagocytes and likely also an impairment of T-cell immunity (9, 27, 63–67).

SALMONELLA: A BACTERIUM WITH COMPLEX INTRA- AND EXTRACELLULAR PATHOGENESIS

Where do antibodies target Salmonella?

Salmonella infections are normally acquired by the oral route and reach the blood after invading the gastrointestinal tract (68). In the blood, the bacteria can be found either as extracellular bacteria or associated with CD18⁺ cells (69) (Fig. 1A). Extracellular Salmonella in the blood can be targeted by antibodies that enhance their uptake and killing by resident phagocytes of the spleen and liver (46, 70, 71) and can potentially lyse them via activation of the complement classical pathway (56). From the blood, Salmonella reach an intracellular location within phagocytes of the liver, spleen, and bone marrow (70, 72, 73) (Fig. 1B). Early in infection Salmonella reaches mainly splenic red pulp (F4/80⁺, MSR-A^low) and marginal zone macrophages (Macrophage Scavenger Receptor A [MSR-A⁺]) (74). In the liver, Salmonella localizes preferentially in the resident Kupffer cells. During the infection, bacteria can also be found in dendritic cells or B cells (36, 74–77). As the infection progresses and bacteria reside and possibly grow intracellularly (Fig. 1C), inflammatory cells infiltrate the initial unicellular foci of infection to form multicellular pathological lesions surrounded by normal tissue (Fig. 1D). These lesions contain polymorphonuclear phagocytes (PMNs) during the first few days of infection, which are later replaced by inflammatory macrophages (36, 78).

FIG 1.

The complex pathogenesis of Salmonella consists of an intracellular antibody-refractive growth phase and an extracellular antibody-susceptible phase of spread. (A) After invading the gut, Salmonella bacteria reach the blood, where can be targeted by antibodies. Some bacteria in the blood are lysed by antibody- and complement-dependent serum bactericidal activity (SBA). (B) Bacteria opsonized by antibodies are engulfed and killed more efficiently by resident phagocytes as soon as they reach the tissues (e.g., spleen, liver, bone marrow). (C) Those bacteria that resist killing establish unicellular initial infection foci (infected phagocytes). Surviving bacteria grow mainly in an intracellular location. (D) The initial single-cell infection foci become spatially separated multicellular pathological lesions due to the infiltration of polymorphonuclear phagocytes (PMNs) and later mononuclear cells. In this intracellular location, bacteria are inaccessible to antibodies. Bacteria grow intracellularly within the multicellular infection foci, but their numbers within each phagocyte remain low due to the spread of the infection to infection of new host cells. (E) Some bacteria escape the lesions, become vulnerable to antibodies, and eventually spread at distant sites. When the bacteria are released from infected cells, they might undergo three different fates, as follows: (i) being targeted by antibodies and killed via complement-mediated SBA or opsonophagocytosis, (ii) rapidly infecting neighboring cells within the same lesion, or (iii) traveling to distant sites in the body to establish new unicellular infection foci.

Once Salmonella bacteria have homed inside host cells, their intracellular location would render the bacteria inaccessible to antibodies, making it difficult to envisage a role of humoral immunity in protection. Detailed analysis of intracellular bacterial densities over the course of experimental infections in mice has shed light on how and where the bacteria can become vulnerable to antibodies. In fact, microscopic observation of immune-labeled and/or fluorochrome-expressing intracellular bacteria in the spleen and liver revealed that the majority of infected phagocytes contain very low intracellular numbers at any time of the infection, irrespective of the overall net growth of the bacteria in the organs (77). The infection process is underpinned both by intracellular growth/survival and by an increase in the number of infected cells and multicellular pathological lesions, but not by increases in the number of visible intracellular bacterial per phagocytes (24, 77, 79) (Fig. 1D). This is due to the continuous redistribution of the bacteria from infected host cells to uninfected ones and the spread of individual microorganisms to new infection foci (Fig. 1E), mediated by the Salmonella type three secretion system (T3SS) encoded by the Salmonella pathogenicity island 2 (SPI-2) (80, 81). Salmonella infections are therefore dispersive processes in which bacteria have an intracellular phase of growth and an extracellular one of spread.

Is the dispersiveness of the infection a plausible explanation for the role of antibodies in protection against Salmonella?

It would be reasonable to postulate that in a dispersive infection process, cell-mediated immunity enhances the antimicrobial functions of phagocytes, therefore affecting the fate of the intracellular bacteria (82–86); antibodies would opsonize the extracellular bacteria in transit between cells and target them to activating cellular receptors, thus increasing the antimicrobial activity of otherwise naive phagocytes at new infection foci (87, 88). Antibodies would therefore be expected to have a major impact on the process.

However, we know from a large body of data that antibodies alone, in the absence of cell-mediated immunity, are unable to modify the net growth rate of Salmonella in the spleen and liver of an infected experimental animal (21, 46, 50). For example, adoptive transfer of Salmonella-specific immunoglobulins or immunization with nonliving vaccines do not, surprisingly, affect the growth curve of the bacteria in vivo. An effect of antibodies is visible only in the first few hours of the infection, when the initial kill of the inoculum is enhanced (46). This early protective effect of antibodies is more evident when experimental animals are challenged via the intraperitoneal route, where antibodies most likely accelerate capture of the bacteria by peritoneal phagocytes and therefore abort extracellular growth in the peritoneal cavity. The use of molecularly tagged and therefore individually traceable Salmonella populations, combined with mathematical modeling and statistical analysis of data, has further confirmed the inability of antibodies to significantly affect the growth and spread of bacteria in the body. This research approach has confirmed that, in the early stages of the infection, bacteria spread from cell to cell within each organ, later followed by systemic spread of bacterial populations between distant body sites, coinciding with the appearance of bacteria in the blood (89). Using this system, it became clear that only live vaccines (which induce both humoral and cell-mediated immunity) would be able to control bacteremia and restrain growth and spread of the bacteria in the body. Conversely, mice where antibodies were the only vaccine-mediated effector mechanisms (i.e., mice immunized with killed vaccines or depleted of T cells after immunization) would not be able to control the spread of the bacteria or abort bacteremia (71).

ANTIGEN SPECIFICITY AND PROTECTION

Salmonella infections induce antibody responses against a large array of surface, periplasmic, and cytoplasmic antigens in humans and other infected animals. Immunoreactive antigens have been detected using enzyme-limited immunosorbent assay (ELISA), immunoblotting, and protein microarrays (90–94). These include lipopolysaccharide (LPS) antigens, porins, lipoproteins, fimbriae, flagella, heat shock proteins, and in some serovars, the Vi surface polysaccharide antigen (95, 96). Several antigens have been identified as targets of the protective antibody response. The O antigen, which consists of sugar repeats exposed on the bacterial surface, is a prime target of protective antibodies, with some epitopes being more protective than others. Immunodominant serovar-specific O antigens (e.g., O:4 and O:9) determine to a large extent the specificity of protection between serovars and induce antibodies that are more protective than the ones directed against O antigens shared among different Salmonella serogroups (e.g., O:12) (97–100). The Vi antigen is immunogenic and was shown to confer protective immunity in human volunteers and in field trials, both as a native polysaccharide and as a protein conjugate (49, 51, 53, 101–103).

The functional role of the antibody response to porins is still not fully clarified. Antibody responses against some porins, but not others, are protective in animal models (104, 105). For example, antibodies to the trimeric OmpD porin, but not to the monomeric and abundant OmpA protein, protect mice against parenteral challenge. Furthermore, OmpD is a candidate antigen for iNTS vaccines (20, 105, 106). The reasons for the different protective abilities of OmpA and OmpD have been elegantly postulated to be due to different accessibility of antibodies to OmpD at the bacterial surface. Both OmpA and OmpD are located in the outer membrane of Salmonella in a position potentially shielded by the LPS O side chain. However, OmpD creates a larger footprint than OmpA in the O-antigen layer, sufficient for a single IgG to gain access to the most-exposed surface loop epitopes of OmpD (105). It remains unclear how antibodies to OmpD mediate serum bactericidal activity (SBA), given that at least two Ig need to come together for activation of the classical pathway of the complement. The conclusion that OmpA is poorly accessible by IgG also clashed with data showing its binding by human recombinant IgG specific for a mimotope inserted in OmpA (107).

Antibodies to flagellin are detectable in infected individuals and targeting this antigen can mediate opsonophagocytosis (87, 93).

WHICH IS THE MAIN PROTECTIVE MECHANISM OF ANTI-SALMONELLA ANTIBODIES?

Following their binding to the bacterial surface, antibodies can exert antibacterial efficacy mainly in two ways: either by increasing phagocytosis of bacteria via targeting the bacteria to specific receptors on the surface of immune cells, or by direct bactericidal activity, mediated by the activation of the classical complement pathway (serum bactericidal activity [SBA]). Both of these mechanisms are likely to operate in Salmonella infections. Early studies showed that clearance of Salmonella from the blood of mice is dramatically accelerated by immune serum, suggesting a role for antibodies in the enhancement of phagocytosis (70). More recently, evidence for an enhancement of opsonophagocytosis by Salmonella antibodies has been corroborated by in vitro systems. Sera from humans vaccinated with live attenuated Salmonella enterica subsp. enterica serovar Typhi enhance opsonophagocytosis in vitro, with IgG playing a major role (108). Engagement of opsonized bacteria with the activating Fc-gamma receptors on the surface of murine or human phagocytes results in increased uptake of the bacteria, increased production of reactive oxygen intermediates (ROI), and enhanced antibacterial functions of the infected cells (87, 88, 107). Opsonization drives increases in both the number of phagocytes that ingest bacteria and in the efficiency of each individual cell in ingesting Salmonella, as indicated by the higher intracellular numbers per phagocyte of opsonized versus nonopsonized bacteria (88, 107). Antibodies have also shown to be essential for optimal phagocytosis of iNTS strains by peripheral blood cells from Africans (55).

Antibody-dependent, complement-mediated SBA correlates with susceptibility to iNTS in African children (56), and therefore SBA has been often taken as an indication of the functional activity of antibodies when testing new vaccines against iNTS (109). However, the role of SBA as a true mechanism of antibody-mediated protection must be evaluated with caution. Studies that used human blood and sera to compare the kinetics of SBA and the phagocytosis of iNTS clinical isolates show that phagocytosis allows bacteria to escape the blood and establish intracellular infection before they are killed by the complement membrane attack complex (110). This would indicate that opsonophagocytosis is likely to be the prevailing antimicrobial mechanism mediated by anti-Salmonella antibodies. Furthermore, no correlation was found between protection afforded by live vaccination and by SBA in a controlled human typhoid challenge model (111).

COMPLEMENT OR Fc-GAMMA RECEPTORS?

The complement system and Fc-gamma receptors (FcγRs) can potentially play a crucial role in antibody-mediated immunity against Salmonella diseases. However, their relative importance is unclear.

Complement-mediated SBA can result in bacterial killing, but, as discussed above, it is unclear whether this mechanism is relevant for immunity to Salmonella. Complement appears to be essential for antibody-dependent opsonophagocytosis of iNTS strains by human blood phagocytes and for the production of ROI and bacterial killing (55). However, binding of antibody-opsonized Salmonella to FcγR on human and murine cells in the absence of complement also enhances phagocytosis and ROI-mediated bacterial killing, with the activating FcγRI playing a major role (88, 107).

In vivo studies also provide evidence for a role for both complement and FcγR, with their relevance depending on the experimental model used. When mice lacking FcγRI, -II, -III, and -IV (FcγR knockout [KO]), or mice lacking the complement C3 component (C3 KO), or lacking all four FcγR and C3 (FcγR/C3 KO), were passively immunized with anti‐LPS O4 IgG2a monoclonal antibodies and subsequently infected parenterally with Salmonella Typhimurium, only FcγR KO animals showed a significant reduction in the bacterial loads in the liver, spleen, and mesenteric lymph nodes (112), at a level similar to those observed in wild-type control animals. In this model, therefore, the role of complement prevails on that of FcγRs. However, when mice lacking FcγRI, -II, and -III and mice lacking C3 were immunized with a live attenuated vaccine and later challenged orally with virulent Salmonella, only the FcγR-deficient mice succumbed to the infection (113), indicating a prevailing role of FcγR.

In summary, evidence for a role of both FcγR and complement has been provided, but firm conclusions on their relative importance are difficult to draw due to discrepancies between experimental conditions, models, and host species.

DOES QUALITY OF THE ANTIBODY RESPONSE MATTER?

The qualitative traits of the antibody response have a great effect on its function and potency. The isotype profile has effects on the binding of antibodies to FcγR receptors and on efficiency of complement activation. This in turn has effects on SBA, opsonophagocytosis, and on the enhancement of the intracellular antibacterial mechanisms of phagocytes.

Virtually all classes of Ig are produced in response to infection with Salmonella. As expected, IgM appear early after infection and are usually followed by IgG and IgA (92, 95, 114–117). Different isotype profiles are seen following natural infection or vaccination with live or nonliving vaccines (22, 118). Some nonliving vaccines can induce isotype-switched responses to Salmonella polysaccharide antigens, indicating that the T-cell responses induced by these preparations may not be able to mediate the suppression of bacterial growth in the tissues, but are sufficient to support isotype switching of the Ig response. For example, IgG responses are detected following vaccination with Vi-conjugate vaccines (51, 119, 120); interestingly, outer membrane vesicle vaccines (OMV) are capable of inducing highly effective IgG2a and IgG2b in mice (109); live vaccines induce high titers of all IgG subclasses, including higher levels of IgG2a (22).

The isotype profile of the antibody response impacts on its protective function. The efficacy of individual subclasses has been studied in murine and human systems, and this area of research is not devoid of controversy. In murine studies, polyclonal and monoclonal IgM, as well as IgA to Salmonella polysaccharide antigens, were found to be highly protective and in some cases more protective than IgG (100, 121, 122). Conversely, opsonization of Salmonella with O4-specific IgA, IgG1, IgG2a, or IgG2b, but not with IgM monoclonal antibodies, resulted in cell-dependent bacterial killing in vitro. In in vivo passive immunization studies, IgG2a and IgG2b O4-specific monoclonal antibodies provided higher functional activity than IgA, IgM, and IgG1 by decreasing the bacterial load in the blood and tissues (123).

A role for IgA in protection against Salmonella is shown by the increased susceptibility to infection of polymeric immunoglobulin receptor (pIgR^−/−) knockout mice, which are unable to bind and actively transport dimeric IgA to the mucosae (124), and by the protective ability of IgA monoclonal antibodies in vivo (125).

The relative potency of individual isotypes can be dependent on the antigen that is targeted. In fact, unlike what is seen in the case of antibodies against the LPS-O antigen, it has been shown that mice lacking IgG1, but not lacking IgG2a, are substantially less protected after immunization with the OmpD porin than wild-type controls. Immunization with OmpD was maintained in T bet-deficient mice that do not produce IgG2a. This is consistent with IgG1 having an important role for protection after immunization with OmpD, but IgG2a being less important (126).

The relative potency of human IgG was studied in an in vitro system in which OmpA and flagella were tagged with a foreign CD52 mimotope (TSSPSAD). The bacteria were opsonized with a panel of humanized recombinant CD52 antibodies that share the same antigen-binding V region but have constant regions of different subclasses. This work revealed that, although opsonization with all IgG subclasses increases Salmonella uptake by human phagocytes, differences in potency can easily be revealed. IgG3 resulted in the highest level of bacterial uptake and the highest average bacterial load per infected cell, which was closely followed by those of IgG1, then IgG4, and lastly IgG2. Phagocytosis mediated by IgG1, IgG3, and IgG4 had a higher dependency on FcγRI than on FcγRIIA, whereas IgG2-mediated phagocytosis required FcγRIIA more than FcγRI (87, 107). Therefore, both the subclass of human IgG and the type of FcγR that is available for antibody binding affects the function of anti-Salmonella antibodies.

Some IgG subclasses can be detrimental to the antimicrobial function of the antibody response. In fact, the inability of blood from HIV patients to kill Salmonella is due to an inherent inhibitory effect of anti-LPS antibodies. This inhibition is dependent on high concentrations of antibodies and is strongly associated with IgA and IgG2 anti-LPS antibodies, possibly related to the poor ability of IgA and IgG2 to activate complement and to deposition of complement at sites where it cannot insert in the bacterial membrane (127).

CROSSTALK BETWEEN B CELLS AND T CELLS AND QUALITY TRAITS OF THE ANTIBODY RESPONSE

T-cell responses can be detected in animals and humans following infection with live Salmonella (128–130). Isotype switching and production of antiprotein antibodies require the presence of T cells (27, 118); for example, T cell-deficient athymic mice produce mainly IgG3 and IgM antibodies to lipopolysaccharide, whereas euthymic mice can produce IgM, IgG1, IgG2a, IgG2b, and IgG3 anti-LPS and anti-Salmonella protein antibodies (27). Similarly, some classes of nonliving vaccines only elicit a restricted isotype repertoire (22). The reciprocal interaction between T cells and B cells is essential for the development of both humoral and cell-mediated immunity to Salmonella (Fig. 2). In the absence of functional B cells, the onset of T-cell immunity is impaired, albeit not abrogated, in mice (131–134). Cytokine production and antigen presentation via major histocompatibility complex (MHC) class II molecules from B cells are essential for activation of Th1 and Th17 responses to Salmonella, as shown by studies in bone marrow chimera mice where only B cells are deficient in selected immunological functions (135). Interestingly, B cells can engender T-cell responses via engagement of innate immune receptors early in infection and via specific activation of the antigen-specific B-cell receptor later in the disease. These functions instigate a loop where T-cell activation in turn contributes to the isotype profile of the response. In fact, antigen-specific IgG2c primary responses are dependent on MyD88 signaling to B cells, while other Ig classes are not (IgG1 and IgG3) or are much less so (IgG2b and IgA). Lack of MyD88 signaling in B cells of chimeric mice results in impairment of development of IFN-γ effector T cells, a likely contributory factor in the lack of IgG2c (136).

FIG 2.

B-cell and T-cell cross talk impacts quality of the antibody response. MyD88 signaling and recognition of bacterial antigens via the B-cell receptor (BcR) leads to B-cell cell activation and antigen presentation to naive T-cells via MHC class II molecules. Interleukin 6 (IL-6) and gamma interferon (IFN-γ) from B cells mediate the development of Th immunity. Th immunity in turn triggers activation and maturation of B cells and induces isotype switching of the antibody response.

KEY CONSIDERATIONS FOR THE DEVELOPMENT OF SALMONELLA VACCINES

Antibodies are an essential component of the protective immune response to Salmonella. An ideal vaccine would therefore be the one able to induce both antibody responses and protective cellular immunity.

The induction of some level of T-cell immunity is essential whichever the choice of vaccine. T-cell immunity induced by natural infection or live vaccines is sufficient to support the isotype switching and affinity maturation of the antibody response, but also to mediate the suppression of the growth of virulent bacteria, when either resistant or susceptible animals are reinfected. In contrast, T-cell immunity induced by some nonliving vaccines is sufficient to support isotype switching and antiprotein responses (109, 118), but not protective cellular responses. Why this level or type of T-cell immunity is unable to activate phagocytes and curtail bacterial net growth in the tissues still remains one of the main unanswered questions in bacterial vaccinology.

Antibodies alone can have an impact on the very early stages of the infection, when they enhance bacterial killing before Salmonella bacteria have reached an intracellular location within phagocytes (46, 71); however, they have no effect on the net growth of bacteria in the tissues, or surprisingly, on the control of bacteremia (9, 46, 60, 71, 127, 137). Therefore, it is likely that in individuals who do not have specific Th1 memory, vaccine-induced antibody responses protect by preventing the establishment of the infection following transmission of the pathogen. The situation would be different in populations where background cellular immunity is present. For example, those vaccines that induce mainly antibody responses would be more effective in areas where disease is endemic, where a background of cellular immunity is already present in the population, likely due to low grade preexposure to the pathogen and/or to cross-reactive antigens of other microorganisms. Interestingly, young African children develop Th1 immunity to Salmonella very early in life, but remain susceptible to iNTS until they acquire Salmonella-specific antibodies (57); the aim of an effective iNTS vaccine for children would therefore be the induction of antibodies early in life. A lower protective efficacy of Vi vaccines against typhoid fever has been detected in volunteers in a controlled human challenge study compared to that in field studies, further suggesting the possibility that a different immunological exposure background due to the geographical area may affect vaccine efficacy (102, 103, 138).

The development of both safer live attenuated vaccines and nonliving vaccines would be desirable; the former would be more suitable for travelers, where elicitation de novo of both T-cell and B-cell immunity is required; the latter vaccines would be suitable for use in areas where disease is endemic, where, as discussed earlier, increased safety is a prerequisite and induction of antibodies would suffice.

Optimization of the immune response is important especially for nonliving vaccines, the efficacy of which is likely to be entirely based on antibodies. The importance of the isotype profile in relation to antigen specificity and function has been touched upon above with some representative examples.

Greater efforts are needed to optimize antibody responses induced by vaccines to be used in areas where immune suppressive comorbidities colocalize geographically with the target disease. An example is provided by malaria. This disease has a dual effect on humoral immunity. First, malaria can suppress the acquisition of anti-Salmonella antibodies (59, 139). Second, malaria can impair complement levels (61) and suppress the antimicrobial functions of phagocytes (28, 58, 140), thereby potentially not allowing the expression of antibody-mediated resistance. More research is needed to identify vaccine solutions that can overcome these problems. In fact, iNTS vaccines for Africa must induce immune responses optimized to confer resistance in the presence of multiple and various underlying comorbidities.

The optimal choice of antigens is still debatable, and several single-antigen vaccines are being developed and/or are in use. For example, the vaccines based on the Vi polysaccharide are immunogenic and induce protective serum responses. However, these vaccines include a single antigen that is not essential for the virulence of the bacterium (Vi-negative variants of S. Typhi are capable of causing disease [141]) and is downregulated within the tissues soon after infection (142). Multiantigen vaccines would probably be a wiser choice than single-antigen ones and would also offer the possibility of inducing antibody responses that can potentially protect against a large number of Salmonella serovars. For example, low-reactogenicity outer membrane vesicle-based vaccines that contain multiple structural antigens are very immunogenic and elicit highly functional isotype-switched antibody responses against a variety of polysaccharide and protein determinants (109, 143).

CONCLUSIONS

The overall scenario that emerges from decades of research in experimental animals and in humans indicates that antibodies are certainly essential for resistance against systemic Salmonella infections and can express the highest level of protective functions when operating in conjunction with T cell-mediated immunity. Antigen specificity, isotype profile, FcγR receptor usage, and complement activation (Fig. 3) are intertwined factors that have great influence on antibody-mediated protection to Salmonella.

FIG 3.

Antibody-mediated protection to Salmonella is underpinned by a complex interplay between qualitative traits and effector mechanisms. Antigen specificity, isotype profile, FcγR receptor usage, complement activation, and different effector mechanisms (opsonophagocytosis, killing via reactive oxygen intermediates [ROI] and serum bactericidal activity [SBA]) are intertwined factors that influence antibody-mediated protection to Salmonella.

There is still a great deal of discrepancy between findings and conclusions from different studies over several decades, and this makes rational vaccine design very difficult.

To improve current vaccines and design new ones in the future, it will be necessary to shed light on the mechanisms that underpin both the development of antibody responses and their effector functions. It will also be necessary to learn how to tailor antibody responses to situations in which other components of the immune system might be impaired, as often seen in areas where disease is endemic and where immune-suppressive comorbidities geographically coincide with systemic Salmonella infections. Different vaccines and antibody responses may be needed for travelers and residents in areas where salmonellosis is endemic.

Biographies

Pietro Mastroeni obtained a degree in medicine and surgery from the University of Messina, Italy, in 1990, and a Ph.D. in immunology from the University of Cambridge, United Kingdom, in 1994, where he was also awarded a doctor of science degree in 2017. He has worked at the University of Newcastle and Imperial College, London, and is now based at the University of Cambridge. His work has contributed to the fields of microbiology, immunology, and vaccine development for the last three decades. He is a Fellow of the Royal Society of Biology and of the American Academy of Microbiology.

Omar Rossi holds a master of science degree in molecular biology and received a Ph.D. in biotechnology at the Novartis Vaccine Institute for Global Health (NVGH) in 2014. After postdoctoral studies at NVGH and being a visiting scientist at the Wellcome Trust Sanger Institute, he was appointed as research associate at the University of Cambridge, working on pathogenesis of Salmonella disease. He is currently a senior scientist in the immunoassay group at the GSK Vaccines Institute for Global Health, working on development of vaccines, including ones against iNTS, for low- and middle-income countries. With over ten years of experience in the field of vaccinology, Dr. Rossi's interest is focused on the better understanding of mechanisms underpinning infections and on development of vaccines and functional assays to understand the action of antibodies in disease.