Population genomics of pneumococcal carriage in Massachusetts children following PCV-13 introduction

Background The 13-valent pneumococcal conjugate vaccine (PCV-13) was introduced in the United States in 2010. Using a large pediatric carriage sample collected from shortly after the introduction of PCV-7 to several years after the introduction of PCV-13, we investigate alterations in the composition of the pneumococcal population following the introduction of PCV-13, evaluating the extent to which the post-vaccination non-vaccine type (NVT) population mirrors that from prior to vaccine introduction and the effect of PCV-13 on vaccine type lineages. Methods and Findings Draft genome assemblies from 736 newly sequenced and 616 previously published pneumococcal carriages isolates from children in Massachusetts between 2001 and 2014 were analyzed. Isolates were classified into one of 22 sequence clusters (SCs) on the basis of their core genome sequence. We calculated the SC diversity for each sampling period as the probability that any two randomly drawn isolates from that period belong to different SCs. The sampling period immediately after the introduction of PCV-13 (2011) was found to have higher diversity than preceding (2007) or subsequent (2014) sampling periods (Simpson’s D 2007: 0.915 95% CI [0.901, 0.929]; 2011: 0.935 [0.927, 0.942]; 2014: 0.912 [0.901, 0.923]). Amongst NVT isolates, we found the distribution of SCs in 2011 to be significantly different from that in 2007 or 2014 (Fisher’s Exact Test p=0.018, 0.0078), but did not find a difference comparing 2007 to 2014 (Fisher’s Exact Test p=0.24), indicating greater similarity between samples separated by a longer time period than between samples from closer time periods. We also found changes in the accessory gene content of the NVT population between 2007 and 2011 to have been reduced by 2014. Amongst the new serotypes targeted by PCV-13, four were present in our sample. The proportion of our sample composed of PCV-13-only vaccine serotypes 19A, 6C, and 7F decreased between 2007 and 2014, but no such reduction was seen for serotype 3. We did, however, observe differences in the genetic composition of the pre- and post-PCV-13 serotype 3 population. Our isolates were collected during discrete sampling periods from a small geographic area, which may limit the generalizability our findings. Conclusion Pneumococcal diversity increased immediately following the introduction of PCV-13, but subsequently returned to pre-vaccination levels. This is reflected in the distribution of NVT lineages, and, to a lesser extent, their accessory gene frequencies. As such, there may be a period during which the population is particularly disrupted by vaccination before returning to a more stable distribution. The persistence and shifting genetic composition of serotype 3 is a concern and warrants further investigation.

and 7F decreased between 2007 and 2014, but no such reduction was seen for serotype 3. We 50 did, however, observe differences in the genetic composition of the pre-and post-PCV-13 51 serotype 3 population. Our isolates were collected during discrete sampling periods from a small 52 geographic area, which may limit the generalizability our findings.

54
Pneumococcal diversity increased immediately following the introduction of PCV-13, but 55 subsequently returned to pre-vaccination levels. This is reflected in the distribution of NVT 56 lineages, and, to a lesser extent, their accessory gene frequencies. As such, there may be a period 57 during which the population is particularly disrupted by vaccination before returning to a more 58 stable distribution. The persistence and shifting genetic composition of serotype 3 is a concern 59 and warrants further investigation.

67
Conjugate vaccination has been a major advance in the reducing pneumococcal disease.

68
The seven-valent pneumococcal conjugate vaccine (PCV-7), introduced in the United States in 69 2000, was highly effective in reducing overall rates of pneumococcal disease, as vaccine type 70 (VT) pneumococci were responsible for the vast majority of cases 4-6 . Carriage of vaccine 71 serotypes also declined, though overall carriage prevalence remained roughly constant due to 72 serotype replacement 2,7,8 .

73
Despite lower overall rates of pneumococcal disease, increases were seen in the incidence 74 of disease due to the replacement non-vaccine type (NVT) population. Serotype 19A in 75 particular became a significant cause of invasive disease 5,9,10 . The thirteen-valent vaccine (PCV-76 13), introduced in 2010, extended coverage to six additional serotypes, including 19A, beyond 77 those included in PCV-7, and has resulted in further reductions in pneumococcal disease 11 . As 78 with PCV-7, however, overall carriage prevalence has not changed substantially 2 . Worryingly, 79 serotype 3, a highly invasive serotype included in PCV-13, appears to have not declined as the    The core genome alignment generated by Roary was used to construct a phylogeny using 124 FastTree v2.1.10 25 . In order to identify clusters of related sequences (Sequence Clusters -SCs), 125 three iterations of hierBAPS were run on the core genome alignment, setting the maximum 126 cluster depth to 1 and maximum number of clusters to 30, 40, and 50 26 .

128
In order to determine the potential effect of PCV-13 on diversity in this population, we 129 calculated Simpson's D for each sampling period, for sequence clusters. This value, which 130 represents the probability that two randomly drawn isolates from a given sampling period belong 131 to different SCs, was calculated as = in that year belonging to sequence cluster and ! !!! is a correction for finite sample size 27 .

133
Following an earlier analysis of serotype diversity in this population, Welch's t-test was used to  The proportion of the population belonging to each SC and their rank order in the 144 population were determined. As diversity increases, the shape of this distribution would be     Previous studies have indicated that PCV-13 may not be as effective against serotype 3 as 281 it is against other included serotypes 2,11,13,14 . The shift we observed in the serotype 3 CC180 282 population following the introduction of PCV-13 may reflect a similar phenomenon to that 283 leading to the recognition of serotype 6C as distinct from 6A following the introduction of PCV-284 7 36,37 . The dominant lineage pre-PCV-13 was also more homogenous (i.e., less diverse) than the 285 post-vaccination population, so it is possible that the immunity generated against serotype 3 by 286 PCV-13 is narrowly tailored to that subset of the population. At present, little genetic variation 287 among the CPS loci was observed, suggesting an alternative explanation for the recent post-288 PCV-13 emergent subclade. Given its propensity for causing disease, the persistence of serotype 289 3 despite its inclusion in PCV-13 warrants further investigation.

290
The response of the pneumococcal population to serotype-targeting conjugate vaccines 291 may also provide insights for other pathogens for which vaccines have been targeted at or 292 differentially affect a subset of their population. The efficacy of the RTS,S malaria vaccine 293 appears to be partially dependent on how well the circumsporozoite protein of a given 294 Plasmodium type matches that in the vaccine 38 . There has also been interest in understanding 295 how the strain dynamics and epidemiology of meningococcal disease caused by the bacteria 296 Neisseria meningitidis will be affected by the rollout of vaccinations against a variety of 297 serogroups 39,40 . While each of these disease systems is different, there is some potential for 298 findings in one to inform hypotheses for how others will behave.

299
Pneumococcal epidemiology has changed substantially as a result of conjugate 300 vaccination. While PCVs have been highly effective in reducing the incidence of pneumococcal 301 disease 4,5,11 , continued vigilance is necessary to monitor for, and respond to, the emergence of 302 potentially dangerous lineages not protected against by current vaccine formulations.