The post-vaccine microevolution of invasive Streptococcus pneumoniae

Abstract

The 7-valent pneumococcal conjugated vaccine (PCV7) has affected the genetic population of Streptococcus pneumoniae in pediatric carriage. Little is known however about pneumococcal population genomics in adult invasive pneumococcal disease (IPD) under vaccine pressure. We sequenced and serotyped 349 strains of S. pneumoniae isolated from IPD patients in Nijmegen between 2001 and 2011. Introduction of PCV7 in the Dutch National Immunization Program in 2006 preluded substantial alterations in the IPD population structure caused by serotype replacement. No evidence could be found for vaccine induced capsular switches. We observed that after a temporary bottleneck in gene diversity after the introduction of PCV7, the accessory gene pool re-expanded mainly by genes already circulating pre-PCV7. In the post-vaccine genomic population a number of genes changed frequency, certain genes became overrepresented in vaccine serotypes, while others shifted towards non-vaccine serotypes. Whether these dynamics in the invasive pneumococcal population have truly contributed to invasiveness and manifestations of disease remains to be further elucidated. We suggest the use of whole genome sequencing for surveillance of pneumococcal population dynamics that could give a prospect on the course of disease, facilitating effective prevention and management of IPD.

Figure 1. Structure of the invasive pneumococcal population.

The maximum likelihood phylogeny was generated using 59,682 polymorphic sites within a 970,559 bp codon alignment of 755 core OGs with similar phylogenetic profiles as ribosomal protein encoding genes in order to exclude genes that were acquired by, for example, horizontal gene transfer. The inner bars each represent a pneumococcal strain (I) and are colored by capsular serotype. The second circle displays the sequence clusters (II). The year of collection is marked by the color (red: pre-PCV7, blue: post-PCV7) and intensity of coloration per year, for PCV7 vaccine serotypes (III) and non-vaccine serotypes (IV).

Figure 2. Annual diversity among accessory genomes.

The diversity was calculated between strains collected within the same year. Using a binary measure of the presence (1) or absence (0) of a gene in each accessory OG, diversity among strains was calculated by applying the distance measure described by Dutilh et al.. Core OG diversity was determined by comparing the pairwise distances between isolates from the phylogenetic tree per year. Circles: means accessory distance; diamonds: means core OG distance; whiskers: 95% confidence intervals.

Figure 3. Temporal stability of accessory gene pool.

Odds ratios for prevalence of individual OGs in the accessory genome whose frequency significantly altered at bottleneck shortly post PCV7 (2007) or during re-expansion (2008–2011), compared to pre PCV7 (2001–2006) The X-axis indicates the proportion of isolates carrying each OG in 2001 while the Y-axis indicates the log transformed odds ratio of strains carrying each OG in 2007 (green dots), and post PCV7 (blue dots), relative to pre PCV7.