Sugar-Coated Killer: Serotype 3 Pneumococcal Disease

1Streptococcus pneumoniae (Spn), also known as the pneumococcus, is a Gram-positive encapsulated bacterium commonly identified by its lancet-shaped diplococcal morphology. An opportunistic inhabitant of the nasopharynx, humans are the only natural host for Spn. Problems arise in colonized individuals when pneumococci ascend the Eustachian tubes to the middle ear where they can cause otitis media. Alternatively, when pneumococci are aspirated into the lower respiratory tract and induce pneumonia. Notably, the likelihood of either of these events is starkly increased among very young children, those who are immunocompromised, the elderly, and those who are experiencing or have recently experienced viral infection. Critically, and once established in the lower respiratory tract, the pneumococcus can cause bacteremia, i.e. invasive disease. This occurs in ~30% of hospitalized individuals. As result Spn is a leading cause of bacterial sepsis following pneumonia. Bloodborne S. pneumoniae are not restricted to the vasculature and able to invade other organs including the central nervous system to cause meningitis and myocardium to cause cardiac complications. Not well appreciated, but a critical aspect of pneumococcal disease, is that survivors of severe infection, most often the elderly, often experience considerable sequelae. These include onset of frailty, cognitive declines, loss of independence, increased risk for adverse-cardiac events, and reduced lifespan. In the United States, it is estimated there were 31,400 (9.6 cases per 100,000 individuals) cases of invasive pneumococcal disease in 2018 with 3,480 directly attributable deaths. Worldwide the number of serious pneumococcal infections is thought to exceed 1–1.5 million per year. Thus, this bacterium, which has garnered the alias of “The Old Man’s Friend”, is a serious cause of morbidity and mortality worldwide and a leading health problem.

2Capsule is a gelatinous external layer produced by bacteria to protect against phagocytosis and other external hazards. In addition to being a common feature of almost all clinical isolates of Spn, capsule is a common feature of most extracellular bacterial pathogens. Bacterial capsules are polymers, with most versions being composed of a polysaccharide. Yet, capsule can be made of other molecules such as in Bacillus anthracis where it is a polypeptide. Polysaccharide capsules are oligomers synthesized in sequential manner by a group of enzymes that link monosaccharides and other moieties (e.g., acetyl groups) together, and then assemble them as a repeating chain. In most Spn, capsule strands are attached to the bacterial surface via covalent bonds to peptidoglycan. One exception to this is serotype 3 pneumococcal capsule, which instead uses non-covalent interactions with phosphatidylglycerol. Along such lines and given the extensive genomic variability in the enzymes responsible for construction of the oligosaccharide unit, the biochemical variability that exists between bacterial capsule types including within Spn is tremendous. As of 2020, 100 biochemically and serologically distinct versions of S. pneumoniae capsule have been identified each of which is encoded by distinct sets of enzymes arranged together in the capsule operon. Importantly, some serotypes, such as serotype 3, have historically been associated with a much higher attack rate and/or morbidity than other serotypes. Along such lines, observations by Sandgren et al. as well Brueggemann et al. suggest that the biochemical properties of the capsule directly contribute to virulence, as isogenic capsule switch mutants with the same accessory genome content were shown to have differences in virulence relative to the capsule type that was produced.

3Unencapsulated pneumococci are capable of nasopharyngeal colonization and due to selective pressure from the current vaccines (see below) are thought to be an emerging population. Unencapsulated pneumococci are capable of causing localized infections such as sinusitis, conjunctivitis and keratitis, as well as otitis media. Importantly, in some instances, unencapsulated isolates have had the capsule cassette replaced with either aliC, aliD, or pspK. AliC and AliD are oligopeptide binding proteins and their deletion has pleiotropic effects on the bacterium including modulation of surface adhesins. PspK is a homologue of the pneumococcal adhesin Choline binding protein A. In other instances where Spn is unencapsulated, there are mutations in the capsule operon that preclude capsule synthesis. As is detailed below the absence of capsule profoundly changes the interactions the bacteria have with its host.

4During nasopharyngeal colonization, individual pneumococci must avoid entrapment in mucus. It is now recognized that capsule, specifically its electronegativity, acts to electrostatically repel mucus, which is also negatively charged, and thereby avoid entrapment and subsequent expulsion. Nelson et al. showed that isogenic strains carrying capsule with a net negative charge avoided mucous entrapment better than versions carrying neutral-charged capsule or an unencapsulated control. Similarly, Li et al. showed that capsule electronegativity influenced the serotype’s nasopharyngeal carriage prevalence. A requirement for capsule becomes starkly apparent during pneumonia and invasive disease. In a mouse model of intraperitoneal challenge and sepsis, the 50% lethal dose of an unencapsulated derivative of a serotype 3 isolate was 5.0 × 107 colony forming units (CFU), as opposed to the 50% lethal dose of its parent wildtype strain of 1 CFU. The reason for this is that unencapsulated pneumococci are exquisitely susceptible to opsonization and phagocytosis by host factors and immune cells, respectively. C-reactive protein, components of the alternative complement cascade, ficolin, surfactant, and pre-existing antibodies against conserved host proteins generated as result of past colonization events or infection, individually and in complementary fashion opsonize pneumococci for phagocytosis. Capsule is inhibitory of phagocytosis as it modulates recognition by the alternative complement pathway. Additionally, capsule prevents the receptors on immune cells from binding to these molecules even though they are bound to the bacterial surface (e.g., Fc receptor with cell wall bound antibody). Thus, the generation of antibody against the capsule itself, is critical for clearance of this pathogen during disease, though the amount of antibody required for clearance varies dependent on the serotype.

5From the bacterium’s perspective, encasement within capsule is not always optimal, as capsule also inhibits interactions between bacterial adhesins and host epithelial cells, a step required for colonization and disease. One solution that the bacteria has is phase-variation where the bacterium stochastically and at low frequency switches back and forth between a transparent (low capsule) and opaque (high capsule) phase. The transparent version is selected for in the upper airway during colonization, where cell attachment is necessary. Whereas the opaque phase is favored in the lower respiratory tract and bloodstream, during which risk for opsonophagocytosis is greater. It is also now known that pneumococci can shed their capsule and quickly modulate their binding affinity. Using scanning electron microscopy, Hammerschmidt et al. showed that the binding of individual pneumococci to epithelial cells results in the loss of capsule from its surface. One explanation for this are findings by Kietzman et al. which showed that peptidoglycan-bound capsule is released by a hydrolase in response to the bacterium’s exposure to cationic antimicrobial peptides. In this instance, capsule shedding serves dual purpose, protecting the bacterium from these positive-charged membrane-damaging products and exposing otherwise capsule-covered bacterial adhesins which in turn allow for intimate interactions with host cells. This concept is in agreement with past findings by our group which showed not only does capsule influence the relative exposure of pneumococcal virulence determinants, but this also occurs in a capsule type dependent manner. As such, surface proteins must be compatible with the biochemical properties of capsule to be effective. Evidence supporting this includes work by Kelly et al. which showed conversion of a strain from one serotype to the next does not consistently increase or decrease virulence, and instead this was genome dependent. Thus, the interplay between capsule and other factors on the surface are crucial.

6Whereas capsular polysaccharide (CPS) from the K1 serotype of Porphyromonas gingivalis, Neisseria meningitidis, and numerous other bacterial pathogens have been demonstrated to be pro-inflammatory, work by Tuomanen et al. found that purified CPS from Spn is not inflammatory on its own. Pneumococcal capsule’s poor immunogenicity stood in stark contrast to the robust immune response elicited by purified pneumococcal cell wall, which is known to contain Toll-like receptor (TLR)-1/2 and TLR-2/6 binding tri- and di-acetylated teichoic and lipoteichoic acids, respectively. In fact, Spn capsule has been demonstrated to have a general anti-inflammatory role as it hinders the interaction of bacterial components, such as cell wall, with pathogen-associated molecular pattern (PAMP) receptors, such as TLRs. Evidence for this includes work by Kung et al. which showed that encapsulated pneumococci elicited less CXCL8 IL-8, a potent chemokine, from epithelial cells than isogenic unencapsulated mutants. These differences in capsule immunogenicity between bacterial species are likely the result of each capsule type’s distinct biochemical properties and in turn their ability to be recognized by the host.

7Because capsule is highly abundant and surface exposed, antibody against capsule is typically highly opsonic and protective against invasive disease. Capsule is currently utilized as the principal antigen in multiple present-day vaccines against pathogens such as Haemophilus influenzae type B and Salmonella enterica serovar Typhi. CPS is also the primary antigen in the three currently licensed vaccines against Spn. These vaccines include purified polysaccharides from the 10 (serotypes 1, 4, 5, 6B, 7F, 9V, 14, 18C, 19F, and 23F), 13 (plus serotypes 3, 6A, and 19A), and 23 (plus serotypes 2, 8, 9N, 10A, 11A, 12F, 15B, 17F, 20, 22F, and 33F) most virulent serotypes of Spn, respectively. The 10- and 13- valent versions composed of capsule conjugated to a protein carrier, the 23-valent being composed of purified CPS alone. Of note, the 10-valent conjugate vaccine is not approved for use in the USA but is extensively used in many other countries.

8It is at this point important to note that CPS, not being a protein, is not presented by dendritic cells to CD4+ cells in context of MHC and is therefore a T-cell independent antigen. Key implications of this include that children under 5 fail to develop protective immunity against capsule following immunization. Moreover, immunization with purified polysaccharide elicits only a modest and short-lived, <5 years, protective response among adults. It is for this reason that the 10- and 13-valent versions of the pneumococcal vaccine consist of CPS conjugated to a protein carrier. Processing of the protein/polysaccharide by antigen presenting cells now occurs in context of MHC-II presentation resulting in CD4+ T cell involvement, generation of long-lived memory B-cells, and robust immunogenicity in children. The resultant antibody titers against these conjugated capsule types are not only typically sufficient to protect against invasive disease caused by these serotypes but also against nasopharyngeal colonization and thereby confer herd immunity by disrupting transmission. Unfortunately, and as is discussed below, serotype 3 is an important exception as the amount of antibody required for protection is not elicited by the current conjugate vaccine. Other serotypes that require higher antibody concentrations for efficient opsonization include 1 and 5; albeit this is most likely the result of a combination of factors, not just the biochemical properties of capsule, such as the carriage of virulence determinants that block amplifying complement sensitivity.

9Finally, and in contrast to capsule-based vaccines, naturally acquired immunity against Spn is the result of antibody generated against the bacterium’s surface proteins rather than the capsule. For this reason, and given prior exposures as children, healthy adults do not typically develop pneumococcal disease. Ongoing efforts to improve on the existing vaccines are focused on expanding the coverage of the 10- and 13-valent vaccine to include more serotypes, identification of conserved pneumococcal proteins that might be co-administered to provide protection against the serotypes not included in the vaccine, and even development of whole cell pneumococcal vaccines that would generate antibody and a T-cell response against conserved surface proteins.

10Pneumococcal serotypes were named in numerical fashion based on the order they were identified. Thus, and in general, the lower number serotypes were initially more frequent causes of serious pneumococcal disease. This has changed as result of the conjugate vaccines which have reduced incidence of disease caused by the lower numbered serotypes, with now higher numbered serotypes becoming more prevalent as replacement strains. Clinical isolates producing serotype 3 capsule are distinct from most other pneumococcal isolates as they have a highly mucoid appearance and a wet phenotype when grown on plates ( Figure 1 ). In contrast, other serotypes form more discrete colonies. Given their considerable difference in colony appearance, serotype 3 isolates were not initially recognized as being pneumococci. In 1901, pathologists William Howard and Roger Perkins published a paper describing their isolation of an unfamiliar streptococcus from the abdomen of a young woman who had died as result of peritonitis. The investigators observed the emblematic mucoid colonies, describing their appearance as “dewdrops”. In their study, they recognized a series of others who had isolated bacteria with properties similar to their own and offered the name Streptococcus mucosus.

11CPS production in pneumococci occurs by one of two mechanisms: Wzy-mediated or synthase-mediated synthesis. So far, Wzy-mediated capsule production has been observed in all but two pneumococcal serotypes; the two exceptions being serotypes 3 and 37 which instead rely on the synthase-mediated pathway. The loci responsible for either mechanism is located at the same region within the Spn genome, between the aliA and dexB genes, except for serotype 37 which operates outside of the canonical CPS locus. Within the Wzy-mediated locus are the four highly conserved genes cpsA, B, C, and D. Genes encoding the serotype specific glucosyltransferases can be found next along with those encoding the conserved Wzy polymerase and flippase. In Wzy-mediated synthesis, serotype-specific UDP-glycosyltransferases assemble a short polysaccharide chain with an enzyme-specific pattern inside the bacterial cytoplasm. The chain is then transferred to the Wzx flippase which reorients the chain into the periplasm ( Figure 2 ). The short polysaccharide is then added on to the previously synthesized chain by the Wzy polymerase. Once the chain has reached a designated length, the polysaccharide is released and bound covalently to the cell wall. Considerable diversity in the UDP-glycosyltransferases encoded within the capsule operon results in Wzy-mediated capsules having multitudes of monosaccharides ordered in specific patterns, the basis of serotypes. Serotype 3 CPS production utilizes synthase-mediated processes. Though the locus of serotype 3 is organizationally similar to that seen in strains that instead utilize Wzy-mediated synthesis, the CPS locus for synthase models contains multiple mutations and truncations that effectively silence many of the genes. The only functional genes within a serotype 3 CPS locus are cps3D, which encodes UDP-Glc dehydrogenase; and cps3S, which encodes for the actual synthase.

12Synthase-mediated production of CPS is therefore much simpler than Wzy-mediated synthesis. For example, serotype 3 contains only two monosaccharides, glucuronic acid and glucose, arranged in an alternating pattern, compared to the Wzy-mediated synthesis pathways which typically contain four to six monosaccharides per pattern. The sugars of synthase-mediated production are added directly onto the growing (non-reducing) end of the polysaccharide chain which is additionally exported and adhered to the external peptidoglycan via glycosidic bond ( Figure 3 ). CPS synthesis is continuous in serotype 3 strains and the polysaccharide chains only dissociate from the synthase when component concentrations are running low. Possible reasons for why capsule production is considerably greater for serotype 3 than other types include fewer steps in synthesis and therefore fewer checkpoints. Alternatively, this method of CPS production may not be as metabolically taxing on the bacterium.

13Even though serotype 3 is included in the both the 13- and 23-valent vaccines against Spn, it remains a major causes of serious human disease. Notably, rates of disease attributable to serotype 3 have not declined since inclusion of serotype 3 in PCV13. Serotype 3 disease often has severe clinical manifestations: most commonly bacteremia-induced septic shock, meningitis, and pneumonia. In some countries serotype 3 is also been linked to complicated pneumonia, or empyema, where the bacterium is found in the pleural cavity. For all these reasons, serotype 3–influenced invasive disease boasts an ~30% mortality rate. This number climbs to 47% for individuals with multiple comorbidities (Charlson Index of +3), being most often the elderly. Notably in a recent study by Africano et al., serotype 3 was independently associated with development of adverse cardiac events in hospitalized individuals. This was due to the bacterium’s propensity for bacteremia, an aspect of disease which confers the bacterium the opportunity to invade the myocardium. Thus, and despite exhibiting relatively low carriage rates among the population, serotype 3 associated opportunistic disease is consistently a major cause of IPD within young and old age groups with devastating consequences.

14In a study by Choi et al., the amount of protection granted by the PCV-13 vaccine versus serotype 3 capsule production was investigated. Results from this experiment suggest that the profuse production and release of serotype 3 capsule overwhelms the protective capacity of antibody that is elicited by the vaccine. A measured 0.2 μl of serotype 3 culture supernatant containing the respective capsule was sufficient to abolish the antibody-mediated protection provided by the vaccine. In comparison, only 25 μl of supernatant from a serotype 4 isolate was required to reach that same point of neutralization. From these results, it was estimated that approximately eight times more antibody was required to confer protection against serotype 3 invasive pneumococcal disease. In addition to inhibiting interactions of the bacteria with phagocytes, serotype 3 is able to escape from capsular antibody due to the fact that its capsule is not covalently attached to the bacterial surface, allowing it to be released in copious amounts to the extracellular milieu. Capsule antibody therefore instead binds to shed capsule and is neutralized in its capacity to opsonize the bacteria itself. Antibody bound to the capsule on the bacterium’s surface would also be eventually released.

15Bacterial surface components with negative electric charge can repel the like-charged immune cells. This aids the bacterium in avoiding NETs and phagocytosis, as well as complement particles released by phagocytic cells. Many physical qualities of bacteria have been linked to the determination of cellular surface charge; one being capsule type. With 100 biochemically distinct pneumococcal serotypes in existence, the level of surface charge can vary considerably. Surface charge is typically determined by measuring a cell’s zeta potential. Neutral charges exist between −10 and +10 mV. While charges of −30 and +30 mV make up the lower and upper extremes, respectively. The lower the zeta potential, the higher the surface electronegativity and vice versa. Bacteria with lower surface charges (lower zeta potentials) generally perform better at avoiding phagocytosis and complement deposition than their neutral counterparts. Among such lines, Li et al. related lower zeta potential to higher rates of carriage in the nasopharynx. In his study, serotype 3 stood as the capsule type with the lowest average zeta potential.

16As indicated, capsule is inhibitory of the adhesin and host ligand interactions that promote bacterial attachment. Studies with a capsule deficient version of serotype 3 strain A66.1 found its unencapsulated derivative to be 105-fold more adhesive to host cells. Furthermore, wildtype A66.1 recovered from interaction with HEp-2 cells showcased a decreased capsule expression level similar to that seen in the capsule deficient variant. Both observations support the notion that capsule shedding is necessary for interaction with host epithelia. Capsule has been also shown to be inhibitory of bacterial biofilm formation. Presumably, capsule inhibits initial attachment of bacteria to a surface or the epithelia, it likely also interferes with bacteria to bacteria interactions. Interestingly, in vitro, serotype 3 has been found to rely on the formation of unencapsulated small colony variants (SCVs) for the formation of biofilms. These unencapsulated mutants have alterations in the genes encoding the capsule operon as result of oxidative stress imposed by the bacteria itself. These SCVs bind to the abiotic surfaces, and in turn are bound by fully encapsulated Spn. While this has not been shown to occur in vivo, it demonstrates that type 3 capsule most likely imposes challenges that the bacteria must overcome. Importantly, capsule is one component, along with other polysaccharides, of both in vitro and in vivo biofilm extracellular matrices. Thus, and although it is initially inhibitory of biofilm formation, it has an important role in the process.

17Capsule, of course, is not the only determining factor of bacterial virulence. Investigations into variations of pneumococcal surface proteins such as PspA and CbpA (PspC) also showcased different levels of virulence and surface exposure when compared using isogenic WU2 (serotype 3) and TIGR4 (serotype 4) strains, respectively. We have shown, with the pneumococcal serine-rich protein PsrP, that the surface proteins which contain adhesive domains must project these through the capsule in order to function. Thus serotype 3, being thickly encapsulated, puts in place unique parameters for the surface proteins it relies on for host-pathogen interactions.

18More than 20,000 whole genome sequences from isolates of S. pneumoniae are currently available in databases. A significant number of genomes were recently contributed by the Global Pneumococcal Sequencing (GPS) project (https://www.pneumogen.net/gps/). The analysis of 20,027 genomes by Gladstone et al. was performed in the context of issues with current conjugate vaccines, including the concern about the emergence of non-vaccine serotypes. They aimed to study S. pneumoniae serotype, antibiotic resistance, and invasiveness in association with the overall genetic background of the isolates. They partitioned lineages into Global Pneumococcal Sequence Clusters (GPSCs) that were then associated with the features of interest listed above, as well as clonal complex (CC) and of course serotype. A CC groups related bacterial strains through the identification of allele variations or sequence types (STs) within seven highly conserved housekeeping genes (aroE, gki, gdh, xpt, spi, ddl, and recP). For associations with GPSCs, CCs were defined as STs with single locus variant differences within the GPS dataset.

19Serotype 3 has been the focus of recent genome-based studies that used GPSCs as the foundation (at least in part) and focused on the epidemiology and evolution of serotype 3 isolates in the US, as well their prevalence in Spn carriage and in PCV13 evasion in the UK. In particular, CC180, an S. pneumoniae clone also known as Netherlands3-31 or PMEN31, is the major complex of serotype 3, containing a majority of clinical isolates and laboratory strains that have been sequenced and analyzed to date. CC180, which is included in GPS lineage GPSC12, can be divided into 3 clades: clade I, which contains subclades Iα and Iβ, and clade II. When the prevalence of isolates from each clade was investigated over a set timeframe, incidences of Iβ remained relatively unchanged while a switch in preference was observed between clades Iα and II. Clade Iα, the predominant group in terms of isolates showed a slight decrease in population shortly after PCV13 introduction. Meanwhile, the population of clade II expanded, overtaking Clade Iα. By 2014, clade II made up ~41% of CC180-related pneumococcal isolates in the US. Up from ~20% just 4 years earlier in 2010. Interestingly, despite the noticeable increase in Clade II isolates and decrease in Iα isolates, no correlation was found between the introduction of PCV13 and Clade II expansion. This suggested other means for clade II selection. Further investigation into the variation of several protein antigens gave insight into this matter. Genes encoding key virulence determinants such as NanA, PspA, and CbpA showed high levels of variability between the three clades. Variations in each of these genes has been shown to impact pneumococcal virulence in a variety of experimental conditions. Several variations in genes responsible for increased drug resistance, such as Tet32, were also observed within clades Iβ and II, but not within Iα isolates. Azarian et al. hypothesized that these and other untested loci gave clade II the necessary competitive advantage to overtake clade Iα, though this has yet to be tested.

20The inability of PCV13 to elicit protective antibody against serotype 3 capsule suggests that a new alternative approach is necessary. Along such lines, immunization with conserved pneumococcal proteins has shown to confer protection against invasive disease caused by serotype 3. Importantly the selected protein antigens must be conserved across CC180’s three distinct clades, be antigenic, be constitutively expressed at high level in vivo, and be demonstrated to confer protective immunity. Candidate proteins for this include PspA and pneumolysin among others. Ideally, the selected proteins should also confer protection against the serotypes not included in PCV13 or other expanded vaccine formulations, thereby conferring protection against all Spn. In similar fashion, a whole-cell based vaccine using killed Spn that expresses conserved proteins and elicits protective antibody would confer similar serotype-independent protection against Spn. In both instances, these new vaccines would expand protection against serotypes that, although individually infrequent, together inflict a significant burden of human disease. It is noteworthy that efforts are ongoing to expand on the current conjugate vaccines. A 15-valent conjugate vaccine, now including serotypes 22F and 33F, is currently undergoing phase 3 clinical trials and appears to generate more antibody against serotype 3 than the 13-valent version. Whether this will be sufficient to reduce overall serotype 3 disease remains to be determined.

21Throughout this article, we have discussed the vital role of capsule on pneumococcal pathogenesis and how unique feature of serotype 3 provides the bacterium with enhanced virulence and an intrinsic ability to resist vaccine induced antibody ( Figure 4 ). Whereas the biochemical properties of serotype 3 indeed influence its interactions with host cells, more simplistically, the copious amount of capsule that are produced by serotype 3 pneumococci overwhelms the 0.35 µg/ml antibody threshold provided by the current conjugate vaccine formulation. Physiological properties of the capsule itself, such as surface electronegativity, and its ability to confer protection against host factors are also an explanation for why disease caused by serotype 3 has such severe disease manifestations. The bountiful production of serotype 3 capsule both benefits and challenges the bacterium, preventing agglutination by the nasal mucin and phagocytosis by host immune cells while simultaneously deterring adherence to the host epithelium. As such, obtaining a general understanding of the variability and basis for selection within serotype 3 regarding its complementary virulence determinants becomes an important goal. A new approach seems to be necessary in order to prevent serotype 3 disease. Vaccines that generate antibody against conserved pneumococcal proteins have been tested in the past with reasonable efficacy and may hold the key to a stark decrease in pneumococcal-related incidents. Moving forward serotype 3 is likely to continue being a major medical problem for some time until a new alternative prophylactic approach is approved.

22JL wrote the first draft of the paper. CJO, JL, and HT contributed to the writing, editing, and direction of the paper. All authors contributed to the article and approved the submitted version.

23HT and CJO received grant support from NIH grants AI114800 and AI146149.

24The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.