Serotype replacement and vaccine effects in the pneumococcus

We are lucky to have effective vaccines against the pneumococcus, a pathogen causing diseases from ear infections to meningitis. However while these vaccines cover the variants most commonly found in disease, they do not cover all pneumococcal strains. And as a result vaccination has provided an opportunity to observe evolution in real time as the pneumococcal population responded to a new selective pressure.

The first generation of these vaccines (termed ‘conjugate’ because they are a polysaccharide linked or conjugated to a protein carrier) covered 7 or more than 90 serotypes. With Dr Jon Finkelstein of Children’s Hospital and alongside a team including Dr Marc Lipsitch, we studied the carried population of pneumococci in children living in Massachusetts communities. By ‘carried’ I mean the isolates that were in the nasopharynx – not those causing invasive disease. This distinction is important because the strains causing disease may not reflect those found in carriage, but the carried population is that which gets transmitted (it is hard to transmit to a new host from the blood, and even more so from the CSF). This allowed us to track the emergence of serotypes of pneumococcus including 19A, which became an important cause of disease that was moreover associated with drug resistance. At the start of this study, the state of the art method for the analysis of such data was MLST. But the state of the art has changed, such that now it is economical to obtain whole genome sequence. We have published the sequences of more than 600 pneumococci collected between 2001 (as the vaccine was introduced) and 2007.

Among other things this allowed us to find genes associated with pneumococci colonizing younger children, and we hypothesize this reflects a situation where the products of these genes, which are mostly cell surface structures, offer an advantage in terms of attachment but only when the host has not had time to develop immunity. In other words the niche of the pediatric host is further stratified by age.

Logistic regression of COGs (not quite exactly the same as genes, but similar) found in pneumococci colonizing younger hosts. With the exception of the mel/mef macrolide resistance genes, all are cell surface structures. Figure from Pubmed Central. Find the original here.
Logistic regression of COGs (not quite exactly the same as genes, but similar) found in pneumococci colonizing younger hosts. With the exception of the mel/mef macrolide resistance genes, all are cell surface structures. Figure from Pubmed Central. Find the original here.

The genomes also allowed us to visualize spatial structure in the population using a simple and intuitive plot. The sequenced samples were collected from a subset of Massachusetts communities, and if there is spatial structure you would expect that things more closely related will be more likely to come from the same community. This is indeed the case, as shown below.

'Croucher' plot showing spatial structure. Points represent the proportion of pairs of isolates in a sample of >600 genomes originating from the same location for a series of maximum genetic distance thresholds. The red blur is the expectation of no structure, obtained by permutation.
‘Croucher’ plot showing spatial structure. Points represent the proportion of pairs of isolates in a sample of >600 genomes originating from the same location for a series of maximum genetic distance thresholds. The red blur is the expectation of no structure, obtained by permutation. Figure from Pubmed Central. Find the original here.

We continued to collect pneumococci in 2009, 2011 and 2013, and the genomes of these isolates are presently under analysis in the lab. In 2010 a new vaccine was introduced, covering 13 serotypes (the original seven plus six more selected to address replacement in the US and major invasive serotypes often located elsewhere). The initial results of this appear to show the population was relatively stable as vaccine was introduced, but this is unlikely to remain the case.

Changes among the STs found in carriage for three important serotypes in the MA carried population of pneumococci, in samples collected in 2009 and 2011 either side of the introduction of a 13 valent vaccine. Few changes are evident. For more information and original figure, see Chang et al 2015
Changes among the STs found in carriage for three important serotypes in the MA carried population of pneumococci, in samples collected in 2009 and 2011 either side of the introduction of a 13 valent vaccine. Few changes are evident. For more information and original figure, see Chang et al 2015