Pneumonia remains the second biggest killer of children under five years of age and is responsible for one-sixth of child deaths in the African region [2], and disproportionately affects children from poor and disadvantaged households [3]. One important strategy to help reduce under-five mortality to 25 per 1,000 live births by 2030 (Sustainable Development Goal 3) is the reduction in the burden of pneumococcal disease through the use of pneumococcal glycoconjugate vaccines (PCVs).
For S. pneumoniae, the most commonly used vaccines globally target a fraction of the more than 93 recognized capsular serotypes [12]. The bacteria’s capsule (CPS) is the most important determinant of virulence and the strongest predictor of prevalence [13], as well as the target of PCVs; thus, changes in CPS serotype frequency have been the focus of many analyses of vaccine effect. However, selection acts on genes outside the operon determining CPS serotype
The pneumococcal conjugate vaccine (PCV) couples capsular polysaccharides to a carrier protein, which elicits T-cell help and results in improved memory±B cell formation, affinity maturation, class switching, and levels of IgG. Two formulations that target 10 or 13 serotypes have been licensed to date [24]. Serotypes included in PCV10 are 1, 4, 5, 6B, 7F, 9V, 14, 18C, 19F, and 23F. The PCV13 formulation includes the additional serotypes 3, 6A, and 19A.
Pneumococcal conjugate vaccines (PCVs) have been used to combat pneumococcal disease in infants. A protein carrier is linked to an otherwise T cell-independent polysaccharide antigen to evoke an immunological response and enhance long-term memory to serotypes included in the vaccines. The introduction and widespread use of PCVs in first world countries has resulted in a significant decline in nasopharyngeal carriage and IPDs caused by vaccine serotypes in children under the age of 2 years.
PCVs also indirectly protect unvaccinated people through herd immunity [10–12]. Encouraging results have also been seen in third world settings where PCVs have been trialled [13,14].PCV given at ages 6, 10 and 14 weeks in third world countries and 2, 4 and 6 months with a booster in the second year of life in first world countries are immunogenic and reduce nasopharyngeal carriage of pneumococcal serotypes included in PCVs [10,15]. These vaccines can reduce pneumococcal disease burden among vaccine recipients (i.e. direct protection) and in unimmunized community members (i.e. indirect effect). These effects are mediated in part, or entirely, by the ability of PCVs to reduce acquisition and density of vaccine serotype nasopharyngeal colonization, which is the biological precondition for progression to pneumococcal disease.
In Uganda, PCV10 was introduced officially in April 2013 and rolled-out countrywide during 2014 [12]. PCV was introduced in Uganda in January 2014, and is given to all children under 1 year at 6, 10 and 14 weeks.
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.
Prior studies have demonstrated hyporesponsiveness to pneumococcal conjugate vaccines (PCVs) when administered in the presence of homologous carriage. An estimated 70% of children under 5 years old carried S. pneumoniae, which is slightly higher than the 56% pneumococcal carriers reported among 1723 children under 5 years old in a rural community of Eastern Uganda, mostly during the dry season from 2008 to 2011 [16,17].
In a study at Mulago National Referral Hospital (Kampala) in 2005-2006 prior to vaccine roll-out, 43% of the serotypes identified in 30 S. pneumoniae isolates from cerebrospinal fluid were included in PCV7 and 70% in the PCV13. The most common serotypes were 6A/6B (40%) (Kisakye et al., 2009). Around the same period, a study on IPD between 2003 and 2007 in Uganda estimated the potential vaccine serotypes coverage to be 55% for PCV7, 58% for PCV10 and 79% for PCV13 (Mudhune et al., 2009).
In the 3-year survey conducted in Eastern Uganda among children under 5 years, it was also estimated that potential vaccine serotype coverage was 42% for PCV10 and 54% for PCV13 (Lindstrand et al., 2016). In the first population-based survey across all population age groups prior to PCV10 introduction in 2014 in Sheema district, south-western Uganda, the carriage prevalence for 10-valent vaccine serotypes among less than 5 year old children was 27.4% with the 13-valent vaccine serotype carriage higher than the 10-valent vaccine serotype carriage in the less than 2 year-old children due to carriage of serotype 6A (Cohuet et al., 2014). In Sheema district, the proportions of serotypes included in PCV10 or PCV13 were low. We found that only 38.1% of all isolates were included in PCV10 among children under 2 years old and 32.8% among children aged 2–4 years.
An important benefit of PCV-induced immunity is a decreased incidence of vaccine-type pneumococcal carriage in the population (herd protection). Since the release and widespread use of PCVs among children, a decrease in the circulation of the targeted serotypes in immunized populations has been observed. As a result, the incidence of invasive pneumococcal disease and pneumonia in both vaccinated children and unvaccinated adults has been reduced dramatically [27]. In fact, most of the overall efficacy of PCV has been attributed to herd immunity for populations at risk. Problems that remain with PCV are the high cost of this complex vaccine, evidence of gradual replacement by serotypes not covered by the vaccine, and poor matching to serotypes circulating in developing countries, which suffer the largest burden of disease [28]
Rodenburg and colleagues showed that, at 24 months of age, children’s responses to PCV against serotypes 6B, 19F or 23F were reduced if they had carried them at any point in the 2 years prior to vaccination [5]. This may be substantially more important in Africa where carriage prevalence is high. The individual and community level impact of PCV on nasopharyngeal colonization is therefore an important endpoint in PCV impact assessments. The overall vaccine impact on both disease and carriage, however, has been dampened by an increase in serotypes not included in the vaccine formulations, that is, non-vaccine serotypes (NVTs). This phenomenon is known as serotype replacement [10, 11]. In addition to serotype replacement, increased capsular switching (genetic recombination of the proteins on the bacterial capsule) in response to vaccine pressure has been reported [14].
Furthermore, the assessment of PCV on nasopharyngeal colonization provides critical data on the degree to which replacement carriage (i.e., increased carriage prevalence by non-vaccine serotypes) occurs, a key issue in determining overall impact on pneumococcal disease burden. it is vital to closely monitor carriage and invasive pneumococcal disease serotypes to observe direct and population herd effects of vaccine on serotype distribution. Inferences about changes in NP carriage due to PCVs are also limited by differences in pre-introduction carriage prevalence between strains and PCV products used.
We aim to assess the impact of PCV vaccination in the presence of such high colonization in a rural setting in South Western Uganda.
An increased carriage density was confirmed as a risk factor for pneumococcal pneumonia in a second patient cohort of adult patients with radiologically confirmed community-acquired pneumonia [16]. While causality cannot be determined from these epidemiological studies, a high pneumococcal density in the nasopharynx may facilitate bacterial invasion and microaspiration into the lungs and thereby increase the likelihood of progression of infection to pneumonia [17].
In addition to being a prerequisite of disease and a reservoir for transmission, carriage also causes an increase in antibody levels against immunodominant pneumococcal surface antigens, including capsular polysaccharides and proteins, potentially immunizing against future colonization and infection [18, 19]. Furthermore, pneumococcal carriage was dense in a high proportion of pneumococcus-positive samples in our study regardless of their PCV vaccination status and high density is associated with increased risk of disease [24]. Studies investigating the effect of PCV on density of carriage for total bacterial load and species-specific load are needed in areas of high colonisation.
There is no established correlate for protection against pneumococcal carriage although 5 lg/mL was found by Goldblatt et al. to be protective against serotype 14 carriage in adults in the United Kingdom [31].Numerous assumptions were made during the development of the common serological CoP and there is equipoise in the scientific community about the relevance of the CoP to carriage and mucosal disease [19]. For some serotypes, greater concentrations of serum IgG were likely to be required to protect at mucosal surfaces (e.g. in the nasopharynx) than in blood [20]. Subsequent analysis of vaccine-induced antibody and the prevention of carriage reinforced the notion that if circulating IgG is indeed a relevant correlate for carriage, remarkably high concentrations are required to reduce carriage acquisition [21]. Deriving CoP for carriage would guide the future use of extended PCVs, as population control of pneumococcal disease by vaccination is now focused principally on its indirect effect mediated through carriage [22]. We aim to explore the complex relationship between existing carriage and vaccine responsiveness.
Pneumococcal conjugate vaccines have been shown to induce indirect protection through reduction in NP carriage [23]. It is of some interest, therefore, to understand what level of antibody might be required to protect against NP carriage, since this might be used as an indirect correlate when assessing the immunogenicity of new formulations of pneumococcal conjugate vaccines. A titer of 5 mg/mL at the beginning of the study was found to correlate with protection against carriage of serotype14. This titer is much higher than the putative serum level of protection against invasive disease that has been estimated by a WHO working group (in the general range of 0.2– 0.4 mg/mL) [1]. It is possible that absolute levels required to prevent carriage may differ between the serotypes, but it is interesting that this putative protective titer (5 mg/mL) is identical to the serum concentration of anti-Hib capsular antibody that was shown to correlate with protection against colonization after administration of a Hib conjugate vaccine [24] The role that antiprotein responses, compared with anticapsular responses, play in the overall protection against the pneumococcus thus remains unclear. Little is known about the effect of pneumococcal carriage on subsequent antibody responses to pneumococcal conjugate vaccine.
Nasopharyngeal carriage is an endpoint for several clinical trials testing novel vaccines because it is a fast, easy-to-measure, and cost-efficient surrogate for disease endpoints [30]. However, the immunological correlates of protection against carriage have not yet been identified in humans, which hinders the development of effective novel vaccines and does not provide a clear licensure pathway for protein-based vaccines.
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
For antibodies to confer direct protection against acquisition, higher anticapsular antibody levels (as induced by PCV) may be required than to protect against invasive disease [43]. Antibody titers greater than 0.35 ug/mL following PCV vaccination are associated with effective protection against invasive disease in infants [44, 45]. However, to confer protection against colonization, antibody titers greater than 4.0 ug/mL might be required, depending on serotype [43, 46]. A recent study assessing anticapsular levels in PCV-vaccinated toddlers in Kenya was unable, however, to identify protective cutoff levels against carriage [47]. This could explain the limited correlation found between serum levels of capsule-specific antibodies and acquisition in children and adults [9, 33
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.
Vaccination also disrupts the composition of the pneumococcal pangenome, which includes mobile genetic elements and polymorphic non-capsular antigens important for virulence, transmission, and pneumococcal ecology. The pneumococcal pangenome consists of “core genes” shared by ≥99% of strains and “accessory genes” present at frequencies ≤99%. Accessory genes may include polymorphic antigens, phage and plasmid-related chromosomal islands, and integrative and conjugative elements (ICE) harboring antimicrobial resistance genes. The latter are mobile genetic elements (MGE), which are often acquired through horizontal gene transfer (HGT) and may remain stable in pneumococcal lineages [21]. Variations in gene content among lineages of a bacterial species are associated with ecological niche specialization and are important for adaptation to changing environments, including selection by vaccine-induced and natural host immunity [21–23]. For the pneumococcus, MGE affect the bacteria’s ability to recombine (i.e., competence) [24], antimicrobial susceptibility [25], and carriage duration [26]. Accessory loci may also be acted upon by negative frequency dependent selection (NFDS), hinting at their underlying role in non-serotype-specific immunity and S. pneumoniae ecology [27]. Taken together, gene variation beyond the capsular polysaccharide loci may significantly impact virulence, fitness, transmission, and, in turn, the overall epidemiology and ecology of pneumococcal strains. Few comparative genomics studies have been carried out on S. pneumoniae, and these studies were based mainly on strains from developed countries and none from Sub-Saharan Africa [19, 20, 21, 22, 23], where the organism exacts its greatest toll. Though these studies have contributed significantly to our understanding of S. pneumoniae, several aspects of the organism particularly, its pathogenicity, evolution and population structure in the Sub-Saharan Africa is still inadequately understood.
Since the introduction of routine pediatric immunization with PCV, several studies have detailed pneumococcal genomic epidemiology in pediatric carriage and otitis media under vaccine pressure. They all demonstrated that the pneumococcal genomic structure of pediatric carriage remained fairly stable, and that serotype replacement occurred mainly through expansion of previously existent clones20,21,22,23,24,25. The single study that examined whole pneumococcal genomes, including virulence factors, reported little effect on the accessory genome at the overall pneumococcal carriage population level despite massive serotype replacement26. However, the effect of pneumococcal vaccination on whole genome epidemiology of invasive pneumococcal disease remains unexplored, although it may hold invaluable information on understanding the long term effects of mass vaccination, especially with regard to changes in clinical manifestation of disease due to the changing prevalence of virulence factors in the pneumococcal population. To understand the evolutionary impact of vaccination and characterize the shift from VT to NVT, we assessed the recombination, evolution, and pneumococcal population history, classified by serotype and by whole-genome sequencing data, across vaccine introduction periods. Furthermore, we investigated metabolic loci variation and pangenome composition over time, with a focus on pneumococcal antigens.
These data improve our understanding of pneumococcal evolution and emphasize the need to consider genome composition when inferring the impact of vaccination.
Infectious diseases including Pneumonia both in children and adults is a research priority for the Faculty of Medicine of Mbarara University of Science and Technology. This project aligns with the research priorities of the University aimed at building research capacity at the medical school to tackle infectious disease. As part of these efforts the University seconded me to undertake Post-doc fellowship sponsored by the Swedish international Development Agency, study to understand the nasopharyngeal colonization of undefives in Mbarara. In furtherance of the capacity building agenda at the University this Ugandan project will be conducted with a significant training and research component for researchers from the Infectious diseases Institute and the University of St Andrews in the United Kingdom.