Research Article: The immunological mechanisms that control pneumococcal carriage

Date Published: December 21, 2017

Publisher: Public Library of Science

Author(s): Simon P. Jochems, Jeffrey N. Weiser, Richard Malley, Daniela M. Ferreira, Christoph Dehio.


Colonization of the human nasopharynx by pneumococcus is extremely common and is both the primary reservoir for transmission and a prerequisite for disease. Current vaccines targeting the polysaccharide capsule effectively prevent colonization, conferring herd protection within vaccinated communities. However, these vaccines cover only a subset of all circulating pneumococcal strains, and serotype replacement has been observed. Given the success of pneumococcal conjugate vaccine (PCV) in preventing colonization in unvaccinated adults within vaccinated communities, reducing nasopharyngeal colonization has become an outcome of interest for novel vaccines. Here, we discuss the immunological mechanisms that control nasopharyngeal colonization, with an emphasis on findings from human studies. Increased understanding of these immunological mechanisms is required to identify correlates of protection against colonization that will facilitate the early testing and design of novel vaccines.

Partial Text

Streptococcus pneumoniae (the pneumococcus) is the most common bacterial cause of pneumonia, meningitis, and otitis media in children [1]. Pneumococcal pneumonia is also associated with significant morbidity and mortality in the elderly [2]. Such increased pneumococcal disease rates in the elderly could be associated with an altered colonization niche, increased oropharyngeal carriage, as well as with alterations in immunity (reviewed by [3]). Moreover, pneumococcal pneumonia is a leading cause of death during seasonal and pandemic influenza infections [4]. In addition, pneumococcus is the most common cause of pneumonia, sepsis, and meningitis among those infected with HIV [5]. Pneumococcal disease is also increased with exposure to cigarette smoke and air pollution [6, 7]. A further concern of pneumococcal infection is high rates of resistance to multiple classes of frequently used antibiotics [8].

There is a large amount of evidence supporting the role of capsule-specific antibodies in preventing colonization. Pneumococcal colonization in adults confers virtually complete protection against acquisition of the same strain up to one year later in a controlled human infection model [19]. In contrast, healthy adults who were recently naturally colonized with a different serotype are not less susceptible to acquisition of experimental carriage [32]. This suggests that serotype-specific adaptive immune responses can control the establishment of carriage, although it cannot be excluded that strain-specific, serotype-independent memory responses protect against homologous reacquisition. Similarly, serotype-specific protection against carriage is present in toddlers for at least several serotypes, demonstrating that this plays a role not just in adults [33]. While capsule can be expressed at lower levels during colonization, its expression is required for colonization [34]. Importantly, increased amounts of capsule are associated with transmission in an infant mouse model of transmission [35]. Serotype-specific protection from acquisition of carriage is likely conferred by antibody-mediated bacterial agglutination on the mucosal surface that may aid mucociliary clearance and, as a result, prevent stable establishment along the epithelium (Fig 1) [36]. Agglutination occurs though the bi- or multivalency of immunoglobulin and has been shown to be independent of the immunoglobulin fragment crystallizable (Fc) region and interactions with complement [37]. A secreted pneumococcal protease can specifically cleave human IgA1, the most abundant antibody subtype on the mucosal surface of the nasopharynx, preventing IgA1-dependent agglutination [37]. This protease, however, does not target IgG. Antibodies can also act by facilitating complement-mediated opsonophagocytosis by effector cells and thus prevent acquisition or mediate clearance (Fig 1) [38]. Therefore, capsule-specific immunity can effectively prevent establishment of colonization.

Longitudinal follow-up of children from birth to one year has demonstrated that carriage protects against subsequent carriage events also in a serotype-independent manner [48]. Indeed, this protection is observed prior to the maturation of capsule-specific antibodies, which are not observed until two years of age. Moreover, the duration and density of carriage episodes decrease with age and previous pneumococcal exposures [49, 50]. Epidemiological data and mathematical modeling have attributed such decreases predominantly to the gradual accumulation of serotype-independent immunity [51]. Models taking both serotype-independent and serotype-dependent immunity into account were able to accurately predict serotype frequency, diversity, and carriage duration. In contrast, models that only included serotype-specific immunity predicted a lower serotype diversity, as well as lower carriage rates in adults, compared with reported data. Several immunological mechanisms could explain such a serotype-independent mode of control, including antibodies directed against proteins, T helper type 17 (Th17) memory cells, and trained innate immune responses.

The role of antibodies against pneumococcal proteins in protection from carriage acquisition has been studied repeatedly with conflicting results, particularly in regards to human testing. In a study using a 23F-type human challenge model, increased baseline antibody levels against pneumococcal surface protein A (PspA) were found in individuals protected against acquisition of carriage [18]. Moreover, a recent study in Native American communities found that decreased antibody titers against PspC group 3 were associated with increased colonization by pneumococci expressing this variant [52]. However, because this study did not assess the relationship between antibody levels and carriage in a prospective manner, no conclusions on the protective potential of variant-specific antibodies could be drawn. In contrast, in studies using a 6B-type human challenge model, serum IgG levels against 18 pneumococcal proteins, including PspA and PspC, did not correlate with protection against carriage [19, 32]. However, the immunodominant N-terminus of PspA is highly variable, and it is possible that variant-specific immunity could be necessary to confer protection. A similar lack of correlation between protein-specific antibodies and carriage acquisition was observed in an observational study in young children [53]. Antibodies and mature B cells were found to be dispensable for the clearance of carriage in adult mice [54]. Recently, a vaccine inducing protein-specific antibodies was tested for its effect on carriage in a phase-2 clinical trial in infants in The Gambia [55]. This vaccine contained PCV10 combined with pneumolysin and pneumococcal histidine triad protein D (PhtD), and antibodies against both proteins were potently induced. However, this vaccine only conferred a 0.5% to 2.1% protection against non–vaccine-type (NVT) carriage, depending on dosage [55]. Because this study was designed to look at NVT acquisition, there was no cohort that did not receive any pneumococcal vaccine. Consequently, the effect of vaccination on VT acquisition could not be compared to natural colonization. In all groups, VT acquisition was around 10% in each 3- to 4-month interval in the year post vaccination (2–5 months old, 5–9 months old, 9–12 months old).

Murine models have indicated a role for CD4+ Th17 cells, which produce interleukin-17A (IL-17A), in controlling carriage density and duration following either vaccination with pneumococcal whole-cell vaccine (WCV) or repeated homologous colonization [56–58]. However, these findings have not yet been conclusively corroborated in humans.

In mice, the control of established colonization and clearance by Th17 cells depends on both neutrophils and macrophages (Fig 1) [56, 57]. Immunization with WCV leads to neutrophil recruitment to the nasopharynx following colonization, and depletion of these cells partially abrogates the protective effect of vaccination [56]. Such neutrophil recruitment is necessary to control carriage because neutrophils are not detected in the nasal mucosa of naïve mice [57]. Therefore, it is not surprising that the depletion of neutrophils did not increase pneumococcal acquisition in an infant mouse model of pneumococcal transmission [68]. In contrast to the case in mice, neutrophils are present in the nasal lumen of human adults and children [69, 70]. Therefore, neutrophils might be able to prevent the establishment of colonization in humans, while the effect of additional neutrophil recruitment is unclear.

Recolonization of mice by the same pneumococcal strain also increases levels of macrophages in the nasopharynx, albeit with delayed recruitment kinetics compared with neutrophils [57]. In mice, this recruitment of mucosal macrophages correlates with dynamics of clearance—a process that requires weeks to months. Macrophage depletion increases carriage density and delays carriage clearance in this model [57]. Molecular studies of this macrophage recruitment have identified innate and adaptive receptors and cytokines in a two-phase recruitment (Fig 1). Toll-like receptor 2 (TLR2) and IL-17A but not neutrophils or CD8+ T cells are required for this macrophage recruitment [57]. Indeed, TLR2 but not TLR4 was previously found to be required for pneumococcal clearance in mice [74]. Early recruited macrophages sense lysozyme-digested pneumococcal peptidoglycan through nucleotide-binding oligomerization domain-containing protein 2 (Nod2), produce C-C motif chemokine ligand 2 (CCL2), and induce recruitment of additional macrophages in a positive feedback loop [75]. Deletion of both TLR2 and Nod2 has a more profound effect on pneumococcal clearance than either alone, suggesting that these mechanisms are complementary [75].

The contribution of other immunological modalities to the control of pneumococcal colonization has not been extensively studied and remains unclear. The presence of local inflammation is likely to predispose to carriage acquisition. Induction of inflammation by treatment with bacterial lipopolysaccharides (LPS) increases carriage acquisition in infant mice [68]. Furthermore, inflammation of the mucosa is associated with increased carriage levels in children [86]. Similarly, asymptomatic upper respiratory tract virus infection predisposes to experimental carriage acquisition in humans [87]. It is not clear to what extent respiratory viral infection and inflammation promote increased carriage by impairing innate immune responses, including neutrophil and monocyte function [84, 88]. Coinfection with virus may act through other mechanisms, such as (i) impairment of mucociliary clearance, (ii) up-regulation of pneumococcal receptors such as platelet-activating factor receptor (PAFr) and polymeric immunoglobulin receptor (pIgR) on epithelial cells, (iii) epithelial denudation and adherence to the exposed basal layer, and (iv) increased availability of nutrients such as sialic acid [89–93].

A better understanding of the immunological factors that govern pneumococcal acquisition, control density, and mediate clearance will guide the informed development of novel antipneumococcal interventions.

The immunological mechanisms that mediate serotype-dependent control of carriage have been well described, with capsule-specific memory B cells and IgG being able to prevent colonization through antibody-mediated agglutination and disease through opsonophagocytosis (Fig 1). However, the serotype-independent immune responses that are able to control pneumococcal carriage remain uncertain, especially in humans. It is currently hypothesized that Th17-mediated protein-specific responses play a role in controlling established carriage density and duration through recruitment and activation of neutrophils and macrophages. However, these data are largely derived from studies in adult mice with little human validation to date. Moreover, the array of pneumococcal proteins that would optimally confer such protection are still unidentified. There is also conflicting evidence from human studies for the capacity of protein-specific antibodies to confer protection. Therefore, further studies assessing these mechanisms in humans are necessary to foster the development of more broadly acting, serotype-independent vaccines.