Impact of pneumococcal vaccination on carriage in a Fijian population
Laura Kate Boelsen ORCID: 0000-0002-2480-0753
Submitted in total fulfilment of the requirements of the degree of Doctor of Philosophy
March 2018
Murdoch Children’s Research Institute
and
Department of Paediatrics Faculty of Medicine, Dentistry, and Health Sciences The University of Melbourne Abstract
Streptococcus pneumoniae (the pneumococcus) causes a range of diseases including otitis media, pneumonia and meningitis, and is a major cause of morbidity and mortality in children under five years of age worldwide. Colonisation of the nasopharynx by S. pneumoniae (carriage) is generally asymptomatic and is considered a precursor for pneumococcal disease. S. pneumoniae has several important virulence factors which contribute to colonisation and disease. The most significant of these is the polysaccharide capsule which forms the outer protective layer of the bacterium, with over 90 known capsule types or serotypes.
Current pneumococcal conjugate vaccine formulations target 10-13 of the most common serotypes. Pneumococcal vaccines have been effective at reducing pneumococcal disease both directly, by protecting vaccinated individuals, and indirectly through reducing carriage, which lowers transmission of S. pneumoniae to unvaccinated individuals. However, pneumococcal vaccine introduction has also led to serotype replacement whereby serotypes not included in pneumococcal vaccines have become more prominent in carriage and disease. There is some evidence to suggest that species replacement (whereby vaccine introduction leads to increases in other bacterial species sharing the same ecological niche) can also occur.
In Fiji, pneumonia and meningitis are a significant burden and S. pneumoniae is a major cause of both diseases. There are two main ethnic groups in Fiji, indigenous Fijians (iTaukei) and Fijians of Indian descent (Indo-Fijian). Significant differences in the burden of disease between ethnic groups have been observed for a number of diseases including invasive pneumococcal disease, which is higher in iTaukei compared with Indo-Fijians. An earlier vaccine trial in Fiji (FiPP) examined reduced-dose schedules and the use of the polysaccharide vaccine (23vPPV) to reduce cost and provide greater serotype coverage. However, the use of 23vPPV in children at 12 months of age was associated with poorer immune responses when challenged at 17 months of age.
ii In this thesis we examined whether 23vPPV use had any long-term effects on nasopharyngeal carriage in the same children now aged 5-7 years. We found no overall effect of 23vPPV on pneumococcal carriage or on carriage of respiratory pathogens (H. influenzae, S. aureus and M. catarrhalis). We found that the carriage rates of all four bacterial species were significantly different between the two ethnic groups. When data were stratified by ethnicity, we found that 23vPPV use in iTaukei children was associated with higher carriage of S. aureus but had no effect on S. pneumoniae carriage. In contrast, 23vPPV use in Indo-Fijian children was associated with lower carriage of S. pneumoniae and had no effect on S. aureus carriage. Given the previous immune hyporesponsiveness observed with the use of 23vPPV and these findings, we conclude that the current WHO recommendations against the use of 23vPPV in children less than 2 years of age are appropriate.
Next, we examined whether the 7-valent pneumococcal vaccine (PCV7) had an effect on the microbiome of children of iTaukei and Indo-Fijian descent. No overall effect of PCV7 on the microbiome at 12 months of age was observed, however there were significant differences between iTaukei and Indo-Fijian children in the composition of bacteria within the nasopharynx. After stratifying by ethnicity, PCV7 was associated with lower relative abundance of Streptococcus in iTaukei children and higher relative abundance of Dolosigranulum in Indo-Fijian children. Given Dolosigranulum has been associated with a healthy microbiome, pneumococcal vaccination elicited generally positive effects in both ethnic groups.
With the introduction of PCV10 into the infant immunisation program in Fiji, we next examined whether PCV10 had an indirect effect on carriage in adult caregivers. Firstly, we examined whether there was added benefit to sampling the oropharynx in addition to the nasopharynx when using molecular methods for identification and serotyping. Oropharyngeal (OP) samples were more complex than nasopharyngeal (NP) samples, and contained commensal streptococci which caused spurious identification and serotyping results. We found that the inclusion of OP samples did not significantly improve detection of pneumococci compared with NP sampling alone. We next examined the indirect effect of PCV10 introduction in adults using NP samples, finding evidence to suggest that PCV10 reduced vaccine-type carriage three years post-vaccine introduction, particularly
iii
in iTaukei adults. Fewer adults had low-density carriage of pneumococci following vaccine introduction, which underpinned an observed increase in pneumococcal density in adults over this time. We found that Indo-Fijian adults had lower carriage of pneumococci than iTaukei adults and carried very little vaccine-type pneumococci. Indirect effects of PCV10 introduction may not have been apparent 3 year after PCV10 introduction and an additional year or years of data would strengthen the evidence for an indirect effect of PCV10 in adults.
Taken together, the studies included in this thesis examine direct and indirect effects of several pneumococcal vaccines on carriage in a middle-income setting.
iv Declaration
This is to certify that
i) the thesis comprises only my original work towards the degree of Doctor of Philosophy except where indicated in the Preface,
ii) due acknowledgement has been made in the text to all other material used,
iii) the thesis is fewer than 100,000 words in length, exclusive of tables, maps, bibliographies and appendices
Laura Boelsen
V
Preface
The main body of work presented in this thesis was carried out at the Pneumococcal Research laboratory, Murdoch Children’s Research Institute, Melbourne, Australia under the supervision of Catherine Satzke, Eileen Dunne, Kim Mulholland, Paul Licciardi and Fiona Russell.
The following work was completed by me in the laboratories of collaborators: multiplex PCR and species identification following whole genome sequencing described was completed at the Institute for Infection and Immunity, St. George’s University of London, London, UK under the guidance of Kate Gould and Jason Hinds; initial analysis of the microbiome sequencing was done at the Institute for Infectious Diseases, University of Bern, Bern, Switzerland with technical guidance from Moana Mika and Markus Hilty; MALDI-TOF MS used for species identification of isolates was carried out at the Microbiological Diagnostic Unit – Public health laboratory, Department of Microbiology and Immunology, The University of Melbourne at the Peter Doherty Institute for Infection and Immunity, Melbourne, Australia with assistance and guidance from Jenny Davis, Kerrie Stevens and Janet Strachan.
Processing of the adult NVEP samples collected in 2013 and 2015 was completed by members of the Pneumococcal Research laboratory team (Belinda Ortika, Casey Pell, Monica Nation, Maha Habib, Regina Oakley and Jenna Smyth). Processing of a subset (n=100) of adult NVEP samples collected in 2012 was completed by Ahmed Alamrousi with guidance from myself and the lab team. I processed the remaining samples collected in 2012 (n=918) and all samples collected in 2014.
Sequencing of samples on MiSeq was completed by Stephanie Eggers and Maelle La Moing from the Translational Genomics Unit at Murdoch Children’s Research Institute, Melbourne, Australia.
vi
The following publication is included in this thesis: Boelsen LK, Dunne EM, Lamb KE, Bright K, Cheung YB, Tikoduadua L, Russell FM, Mulholland EK, Licciardi PV, Satzke C. 2015. Long-term impact of pneumococcal polysaccharide vaccination on nasopharyngeal carriage in children previously vaccinated with various pneumococcal conjugate vaccine regimes. Vaccine 33:5708-5714.
For this publication I undertook all the laboratory work, conducted data analysis and interpretation, and drafted the manuscript. EMD guided laboratory work and was involved in data analysis and interpretation. KEL and YBC undertook and assisted with the statistical analysis. KB was the study coordinator and LT was the Fiji study paediatrician. FMR was involved in the design of the study and lead the clinical arm of the Fiji Follow-up study. EKM was responsible for the study design and co-led the overall study. PVL co-led the overall study, assisted with sample collection and was involved in data analysis. CS oversaw the laboratory work, and was involved in data analysis and interpretation. All authors contributed to the manuscript.
The data in the remaining chapters are unpublished material not submitted for publication.
The PhD candidature was supported by the Fay Marles Scholarship provided through The University of Melbourne.
Parts of the work described in this thesis appear in the following publications:
(i) Papers:
Licciardi PV, Toh ZQ, Clutterbuck EA, Balloch A, Marimla RA, Tikkanen L, Lamb KE, Bright KJ, Rabuatoka U, Tikoduadua L, Boelsen LK, Dunne EM, Satzke C, Cheung YB, Pollard AJ, Russell FM, Mulholland EK. 2016. No long-term evidence of hyporesponsiveness after use of pneumococcal conjugate vaccine in children previously immunized with pneumococcal polysaccharide vaccine. J Allergy Clin Immunol 137:1772-1779.e11
vii
Boelsen LK, Dunne EM, Lamb KE, Bright K, Cheung YB, Tikoduadua L, Russell FM, Mulholland EK, Licciardi PV, Satzke C. 2015. Long-term impact of pneumococcal polysaccharide vaccination on nasopharyngeal carriage in children previously vaccinated with various pneumococcal conjugate vaccine regimes. Vaccine 33:5708-5714.
Hoe E, Boelsen LK, Toh ZQ, Sun GW, Koo GC, Balloch A, Marimla R, Dunne EM, Tikoduadua L, Russell FM, Satzke C, Mulholland EK, Licciardi PV. 2015. Reduced IL-17A secretion is associated with high levels of pneumococcal nasopharyngeal carriage in Fijian children. PLoS One 10:e0129199.
(ii) Abstracts:
Boelsen LK, Dunne EM, Mika M, Ratu T, Russell FM, Mulholland EK, Hilty M, Satzke C. 2016. Effects of pneumococcal vaccination on the nasopharyngeal microbiome of children in Fiji. Victorian Infection and Immunity Network Young Investigator Symposium, Melbourne, Australia.
Boelsen LK, Dunne EM, Gould KA, Ratu T, Russell FM, Mulholland EK, Hinds J, Satzke C. 2016. Improving molecular detection, identification and serotyping of Streptococcus pneumoniae in oropharyngeal samples. 10th International Symposium for Pneumococci and Pneumococcal Diseases (ISPPD10), Glasgow, Scotland.
Boelsen LK, Dunne EM, Mika M, Ratu T, Russell FM, Mulholland EK, Hilty M, Satzke C. 2016. Effects of pneumococcal vaccination on the nasopharyngeal microbiome of children in Fiji. 10th International Symposium for Pneumococci and Pneumococcal Diseases (ISPPD10), Glasgow, Scotland.
Boelsen LK, Dunne EM, Mika M, Russell FM, Mulholland EK, Hilty M, Satzke C. 2016. Effects of pneumococcal vaccination on the nasopharyngeal microbiome of children in Fiji. Lorne Infection and Immunity Conference, Lorne, Australia.
viii
Boelsen LK, Dunne EM, Gould KA, Russell FM, Mulholland EK, Hinds J, Satzke C. 2015. Improving molecular detection, identification and serotyping of Streptococcus pneumoniae in complex samples. Victorian Infection and Immunity Network Young Investigator Symposium, Melbourne, Australia.
Boelsen LK, Dunne EM, Gould KA, Russell FM, Mulholland EK, Hinds J, Satzke C. 2015. Improving molecular detection, identification and serotyping of Streptococcus pneumoniae in complex samples. 12th European Meeting for the Molecular Biology of the Pneumococcus (Europneumo 2015), Oxford, UK.
Boelsen LK, Dunne EM, Bright K, Lamb K, Tikoduadua L, Russell FM, Mulholland EK, Licciardi PV, Satzke C. 2014. Long-term impact of 23vPPS on carriage in children following a reduced dose PCV primary series in infancy. Poster presented at 9th International Symposium for Pneumococci and Pneumococcal Diseases, Hyderabad, India.
Boelsen LK, Dunne EM, Bright K, Lamb K, Tikoduadua L, Russell FM, Mulholland EK, Licciardi PV, Satzke C. 2014. Long-term impact of 23vPPS on carriage in children following a reduced dose PCV primary series in infancy. Lorne Infection and Immunity Conference, Lorne, Australia.
Boelsen LK, Dunne EM, Russell FM, Licciardi PV, Satzke C, Mulholland EK. 2013. Investigation of the relationship between pneumococcal vaccination, immunity and carriage in Fijian children. Lorne Infection and Immunity Conference, Lorne, Australia.
ix Acknowledgements
None of the work in this thesis would have been possible without the study funders, the staff in Fiji, the participants and their families. I also appreciate the student support I have received from both MCRI and Department of Paediatrics, The University of Melbourne as well as the funding in the form of a Fay Marles Scholarship, The University of Melbourne and Harold Mitchell Travelling Fellowship.
I would like to express my gratitude to collaborators and all those that have helped me along the way. Particularly those who warmly welcomed me into their labs for a few weeks: Jason Hinds, Kate Gould and the rest of the BμG@S team at the Institute for Infection and Immunity, St. George’s University of London, as well as Markus Hilty and Moana Mika and the rest of the group at the Institute for Infectious Diseases, University of Bern. It was lovely to work with such friendly, enthusiastic and motivated people who were willing to take the time to teach me and make me feel welcome (and a special mention for Jason for doing my washing that time). I’d also like to thank Jenny Davis, Kerrie Stevens and Janet Strachan from the Microbiological Diagnostic Unit at the Peter Doherty Institute for Infection and Immunity who welcomed me to their lab for a very long afternoon of MALDI-TOF MS.
Thank you to my advisory committee over the years – Prof Andy Giraud, A/Prof Phil Sutton, Dr Pierre Smeesters and Prof Amanda Fosang – for all their efforts and for trying to be as accommodating as possible when organising seven (or eight) very busy people that travel regularly to meet.
It’s inevitable that life will get in the way of a PhD at times, so at times this has been an incredibly difficult journey. I love science and discovering new things, and it’s never hard to work late or weekends when you really want to finish some experiments so you can analyse some data. However, it was the times when other aspects of life, whether ongoing illness or some overwhelming situations, cropped up that I felt like giving up. I finally made it (Phew! Took me long enough!) and it really is the support of supervisors, co- workers, friends and family that make all the difference.
x So, I’d like to thank my supervisors, all five of them! Paul, Kim and Fiona have all been there for me when I needed them, and I have valued their advice and support. A very special thanks to Eileen and Catherine, who really have been the first and usually last stop for all my questions, problems, advice and occasional drinking buddies (especially at conferences). I don’t know how many times I asked Eileen a ‘quick question’ which, if I’m honest was rarely that quick, but she was always willing to take the time to go through and discuss any problem I might be having. And without Catherine’s support (both academic and emotional) and motivation I wouldn’t have come this far. Although I always dreaded seeing pages of red pen corrections from Catherine on my drafts, I will always appreciate the extra effort Catherine put in to getting drafts back quickly and my thesis is better for all the feedback – I think I’ll always have a Catherine in my head telling me not to write certain things.
Thank you to the very large group of wonderful people in the Pneumococcal Research group at MCRI for all the support over the years with an extra special thanks to the micro lab team, as well as all the help in the lab, I thank them for all the morning teas, lunches and occasional afternoon teas they’ve spent with me – sometimes it’s nice to have a break. Special thanks to Belinda (even though she’s a Melbourne City fan), Casey (thanks for all the delicious baking over the years), Monica (especially for all her patience showing me the microarray and QIAcube!) and my fellow PhD student Jayne (it seems like we’ve been doing this forever).
A big thanks to my friends, many of whom have only known me as a PhD student – Jake, Amy and Dom, I also look forward to meeting a Laura that isn’t a PhD student. To my family, which seems to have grown a bit since I started, they’ve been incredible and supportive, and many of them have felt like they’ve been doing this PhD for all long time too. So, thanks to Julie (aka Mum), Mark, Melissa, Lachlan, Gayle, Hilbertus, Scott, Nicolette, Luna, Christian, Gemma, Keira, Talia and Ethan. Finally, my husband Mat, on a purely superficial level, I think I would have spent the last few years eating fairy bread for dinner if it wasn’t for him (I tell him this frequently, but he doesn’t believe me). On a more serious level, sometimes in life you find a person that will help you get through anything, regardless of what life throws at you, if and when it all goes horribly wrong, I can deal with it because of Mat.
xi Table of Contents
Abstract ...... ii Declaration...... v Preface ...... vi Acknowledgements ...... x Table of Contents ...... xii List of Figures ...... xvii List of Tables ...... xxi Abbreviations ...... xxiv 1 Introduction ...... 1 1.1 Pneumococcal disease ...... 2 1.2 Streptococcus pneumoniae ...... 3 1.2.1 Identification ...... 3 1.2.2 Virulence factors ...... 4 1.3 Pneumococcal carriage ...... 7 1.4 Treatment and prevention ...... 8 1.5 Early development of a pneumococcal vaccine ...... 8 1.6 The 23-valent pneumococcal polysaccharide vaccine ...... 9 1.7 The 7-valent pneumococcal conjugate vaccine ...... 10 1.7.1 Direct effects of PCV7 ...... 10 1.7.2 Indirect effects of PCV7 ...... 10 1.7.3 Serotype replacement ...... 11 1.8 The post-PCV7 era ...... 11 1.9 Pneumococcal vaccination and carriage ...... 12 1.10 Effect of pneumococcal vaccination on other pathogens...... 14 1.11 Challenges for implementing pneumococcal vaccines ...... 17 1.12 Fiji ...... 18 1.13 Aims ...... 19 2 Materials and methods ...... 20 2.1 Introduction ...... 21 2.2 Study designs and swab collection...... 21 2.2.1 Fiji Pneumococcal Project ...... 21
xii 2.2.2 Fiji Follow-up study ...... 22 2.2.3 New Vaccine Evaluation Project ...... 23 2.3 Bacterial culture ...... 24 2.3.1 FFS culture ...... 24 2.3.2 NVEP culture ...... 25 2.3.3 Culture of OP samples for isolation of individual colonies ...... 25 2.4 Identification ...... 26 2.4.1 S. pneumoniae identification in FFS ...... 26 2.4.2 S. pneumoniae identification in NVEP ...... 26 2.4.1 Optochin sensitivity ...... 29 2.4.2 Bile solubility ...... 29 2.4.3 Phadebact pneumococcus test ...... 30 2.4.4 MALD-TOF MS ...... 30 2.5 Serotyping ...... 30 2.5.1 Latex agglutination serotyping ...... 30 2.5.2 Latex ‘sweep’ agglutination serotyping ...... 31 2.5.3 Quellung serotyping ...... 31 2.5.4 Microarray serotyping ...... 32 2.5.5 Multiplex PCR serotyping ...... 33 2.5.6 Non-typeable pneumococci and pneumococci from unencapsulated lineages ...... 34 2.6 DNA extractions ...... 34 2.6.1 DNA extractions from swabs using spin columns ...... 34 2.6.2 DNA extractions from swabs using MagNA Pure LC instrument ...... 35 2.6.3 DNA extraction for microbiome analysis ...... 35 2.6.4 Extractions from pure culture ...... 36 2.6.5 DNA extraction from plate culture ...... 37 2.6.6 DNA extraction of OP isolates ...... 38 2.7 Quantitative real-time PCR (qPCR) ...... 38 2.7.1 Duplex assays for S. pneumoniae, H. influenzae, S. aureus and M. catarrhalis ...... 38 2.7.2 lytA qPCR ...... 38 2.8 Real-time PCR ...... 39
xiii 2.9 Alternative targets for S. pneumoniae ...... 39 2.9.1 Reference isolates ...... 40 2.9.2 PCRs ...... 41 2.9.3 bguR and piaB qPCR and real-time PCR ...... 42 2.10 Serotype 14-specific PCR ...... 42 2.11 16S rRNA gene PCR ...... 43 2.12 PCR purification ...... 44 2.13 Multilocus sequence typing (MLST) ...... 44 2.13.1 PCR ...... 45 2.13.2 In vitro transcription, base-specific RNA cleavage and MALDI-TOF MS ...... 47 2.13.3 Sequencing of PCR products ...... 47 2.14 Sequencing and sequence processing ...... 48 2.14.1 16S sequencing ...... 48 2.14.2 Whole genome sequencing ...... 51 2.15 Data analysis ...... 52 2.15.1 Comparing categorical data ...... 52 2.15.2 Comparing continuous data ...... 52 2.15.3 α- and β-diversity ...... 52 2.15.4 Relative abundance and prevalence ...... 53 2.15.5 Heatmap and clusters ...... 53 2.15.6 Correlation network ...... 53 2.15.7 FiPP data and species-specific qPCR ...... 54 2.15.8 Random forest models ...... 54 2.15.9 Multivariate linear regression ...... 54 2.15.10 Correlation between relative abundance and α-diversity ...... 55 2.15.11 Multiple comparison correction ...... 55 2.15.12 Adjusted prevalence ratios ...... 55 3 Long-term effect of 23-valent pneumococcal polysaccharide vaccine on nasopharyngeal carriage in Fijian children ...... 56 3.1 Introduction and aims ...... 57 3.1.1 Fiji Pneumococcal Project (FiPP) ...... 57 3.1.2 Immunological findings from FiPP ...... 59
xiv 3.1.3 Microbiological findings from FiPP ...... 59 3.1.4 Fiji follow-up study (FFS) and chapter aims ...... 60 3.2 Results from pilot FFS (additional to Boelsen et al. (231)) ...... 61 3.3 Publication: Boelsen et al. (2015) Long-term impact of pneumococcal polysaccharide vaccination on nasopharyngeal carriage in children previously vaccinated with various pneumococcal conjugate vaccine regimes. Vaccine 33: 5708-5714...... 64 3.3.1 Supplemental content ...... 71 3.4 Additional Analyses ...... 75 3.5 Further discussion ...... 76 3.6 Conclusions ...... 77 4 The effects of pneumococcal vaccination on the nasopharyngeal microbiome of children in Fiji ...... 79 4.1 Introduction and aims ...... 80 4.2 Results ...... 83 4.2.1 Participant characteristics and risk factor analysis ...... 84 4.2.2 Alpha diversity ...... 87 4.2.3 Beta diversity ...... 90 4.2.4 Microbial composition ...... 94 4.2.5 Heatmap and sample clusters ...... 105 4.2.6 Correlation Network ...... 109 4.2.7 Species-specific qPCR ...... 112 4.2.8 S. pneumoniae overall and vaccine type carriage from FiPP ...... 117 4.2.9 Association between risk factors and microbial composition ...... 119 4.2.10 Multivariate Linear Regression...... 128 4.2.11 Relative abundance and α-diversity ...... 133 4.3 Discussion ...... 135 4.4 Conclusions ...... 145 5 Adult pneumococcal carriage in Fiji ...... 146 5.1 Introduction and aims ...... 147 5.1.1 Indirect effects of pneumococcal vaccination in adults ...... 147 5.1.2 New Vaccine Evaluation Project (NVEP) in Fiji ...... 149 5.1.3 Aims ...... 150
xv 5.2 Results ...... 152 5.2.1 Is there an added benefit to sampling the oropharynx in addition to the nasopharynx for adults? ...... 152 5.2.2 Indirect effect of infant pneumococcal vaccination on pneumococcal carriage in adult caregivers ...... 203 5.3 Discussion ...... 226 5.4 Conclusions ...... 243 6 Overview ...... 244 7 Appendices ...... 251 7.1 Appendix A. 16S sequence processing and analysis scripts ...... 252 7.1.1 Sequence processing script using MOTHUR ...... 252 7.1.2 Alpha and Beta diversity ...... 255 7.1.3 Heatmap script in R ...... 256 7.1.4 Correlation network script ...... 260 7.2 Appendix B. Additional data for chapter 4 ...... 263 7.2.1 Clusters and risk factors ...... 263 7.2.2 Additional nMDS plots ...... 267 7.2.3 Additional graphs ...... 276 7.3 Appendix C. Additional data for chapter 5 ...... 278 7.3.1 Characterisation of OP isolates ...... 278 7.3.2 Full serotyping data for adult NP samples ...... 286 References ...... 288
xvi List of Figures
Figure 2.1. Identification algorithm for pure isolates of S. pneumoniae in FFS ...... 27 Figure 2.2. V4 sequencing approach ...... 50 Figure 3.1. Pneumococcal serotypes (identified by culture-based methods) in pilot FFS participants ...... 62 Figure 3.2. Density (genomic equivalents/ml) of S. pneumoniae, H. influenzae, M. catarrhalis and S. aureus carriage in the pilot FFS participants (n=21) ...... 63 Figure 4.1. Richness (A) and Shannon diversity (B) by vaccination status and ethnicity ...... 88 Figure 4.2. Richness and Shannon diversity by vaccination status (A and B) and by ethnicity (C and D) ...... 89 Figure 4.3. nMDS of the dissimilarity in microbial composition by vaccination status and ethnicity ...... 91 Figure 4.4. nMDS of the dissimilarity in microbial composition by vaccination status in iTaukei (A) and Indo-Fijian children (B) ...... 92 Figure 4.5. nMDS of the dissimilarity in microbial composition by vaccination status (A) and by ethnicity (B) ...... 93 Figure 4.6. Relative abundance (%) of OTUs in all samples pooled to family level ..... 96 Figure 4.7. Median relative abundance (%) for the seven most common families (A-G) by vaccination status and ethnicity ...... 97 Figure 4.8. Median relative abundance (%) for the seven most common families by vaccination status (A) and by ethnicity (B) ...... 99 Figure 4.9. Relative abundance (%) of OTUs in all samples pooled to genus level .... 101 Figure 4.10. Median relative abundance (%) for the seven most common genera (A- G) by vaccination status and ethnicity ...... 102 Figure 4.11. Median relative abundance (%) for the seven most common genera by vaccination status (A) and by ethnicity (B) ...... 104 Figure 4.12. Relative abundance (%) heatmap showing all 132 samples ...... 107 Figure 4.13. nMDS of the dissimilarity in microbial composition by clusters ...... 108 Figure 4.14. Network correlation map based on the 100 most common OTUs considering samples from all children ...... 111
xvii Figure 4.15. Carriage densities (genomic equivalents/ml) of S. pneumoniae (A), H. influenzae (B), M. catarrhalis (C) and S. aureus (D) by ethnicity and vaccination status ...... 115 Figure 4.16. Carriage densities (genome equivalents/ml) of S. pneumoniae (A), H. influenzae (B), M. catarrhalis (C) and S. aureus (D) by vaccination status (i) and ethnicity (ii) ...... 116 Figure 4.17. Overall (A) and vaccine type (B) pneumococcal carriage rate (%) for the children included in this study at 6, 9 and 12 months of age, analysed using a subset of data from FiPP (229) ...... 118 Figure 4.18. nMDS of the dissimilarity in microbial composition by symptoms of an upper respiratory tract infection ...... 120 Figure 4.19. nMDS of the dissimilarity in microbial composition by presence of a runny nose (A) and a cough (B) ...... 121 Figure 4.20. Median relative abundance (%) for the seven most common genera by symptoms of an upper respiratory tract infection (URTI) ...... 122 Figure 4.21. Top twenty most important OTUs in random forest models for ethnicity (A) and symptoms of an URTI (B)...... 125 Figure 4.22. nMDS of the dissimilarity in microbial composition by ethnicity and symptoms of an upper respiratory tract infection ...... 127 Figure 5.1. Example of bacterial growth on a gHBA plate ...... 154 Figure 5.2. Number of Streptococcus spp. (excluding S. pneumoniae) detected by the StrepID component of the microarray in 30 oropharyngeal samples ...... 156 Figure 5.3. Comparison of the number of ‘serotypes’ detected by microarray in nasopharyngeal (NP) and oropharyngeal (OP) samples that were lytA positive or equivocal ...... 158 Figure 5.4. An example of serotyping results using microarray for a NP sample containing one (A) or two (B) serotypes, compared with an OP sample (C) ...... 159 Figure 5.5. Percentage of OP samples (n=42) in which each ‘serotype’ was detected using microarray ...... 160
xviii
Figure 5.6. Serotyping result agreement between latex sweep agglutination (green), microarray (blue) and mPCR (red) serotyping for a subset of OP samples (n=11) ...... 166 Figure 5.7. MALDI-TOF MS identification results for 88 non-pneumococcal isolates subcultured from 11 OP samples ...... 170 Figure 5.8. Latex agglutination (A) and microarray serotyping (B) results for 88 non-pneumococcal isolates subcultured from 11 OP samples ...... 171 Figure 5.9 Subtrees (A-D) from phylogenetic analysis of concatenated sequences from gene fragments of seven housekeeping genes...... 180 Figure 5.10. Comparison of the capsule gene locus for isolate FVEP-002-002-03 with serotype 33F ...... 184 Figure 5.11. Comparison of the capsule gene locus for isolate FVEP-002-080-02 with serotype 19B ...... 186 Figure 5.12. Site of isolation (nasopharynx, oropharynx or both) for overall (A), vaccine type (B) and non-vaccine type (C) adult pneumococcal carriage in 2012 ...... 199 Figure 5.13. Pneumococcal carriage density (genomic equivalents/ml) in the nasopharynx (red) compared with the oropharynx (blue) ...... 200 Figure 5.14. Serotype carriage in the nasopharynx (red) and oropharynx (blue) as percentage of total pneumococcal serotypes in each sample type ...... 201 Figure 5.15. Overall (A), vaccine type (B) and non-vaccine type (C) pneumococcal carriage rates in adult caregivers before (2012) and after (2013, 2014 and 2015) infant PCV10 introduction ...... 212 Figure 5.16. Vaccine and non-vaccine types as a percentage of total pneumococcal serotypes identified in the NP before PCV10 introduction in 2012 and after PCV10 introduction in 2013, 2014 and 2015 ...... 216 Figure 5.17. Pneumococcal carriage density in adult caregivers before (2012) and after (2013, 2014, 2015) PCV10 introduction ...... 218 Figure 5.18. Overall pneumococcal carriage in adult caregivers by ethnicity before (2012) and after (2013, 2014 and 2015) PCV10 introduction ...... 221 Figure 5.19. Overall, vaccine type and non-vaccine type pneumococcal carriage in iTaukei (A) and Indo-Fijian (B) adult caregivers before (2012) and after (2013, 2014 and 2015) infant PCV10 introduction ...... 222
xix
Figure 5.20. Pneumococcal carriage density in adult caregivers by ethnicity before (2012) and after (2013, 2014 and 2015) PCV10 introduction ...... 223 Figure 5.21. Percentage of samples with non-pneumococcal bacterial species and serotypes from non-pneumococcal bacterial species of the total samples run on microarray before (2012) and after (2013, 2014, 2015) PCV10 introduction ...... 225 Figure 7.1. Three dimensional nMDS of the dissimilarity in microbial composition by clusters ...... 264 Figure 7.2. First and third dimensional nMDS of the dissimilarity in microbial composition by clusters ...... 265 Figure 7.3. Second and third dimensional nMDS of the dissimilarity in microbial composition by clusters ...... 266 Figure 7.4. nMDS of the dissimilarity in microbial composition by gender ...... 267 Figure 7.5. nMDS of the dissimilarity in microbial composition by season ...... 268 Figure 7.6. nMDS of the dissimilarity in microbial composition by swab collection year ...... 269 Figure 7.7. nMDS of the dissimilarity in microbial composition by breastfeeding status ...... 270 Figure 7.8. nMDS of the dissimilarity in microbial composition based on whether breastfeeding was continued past the first 6 weeks of life ...... 271 Figure 7.9. nMDS of the dissimilarity in microbial composition based on whether breastfeeding was continued past the first 6 months of life ...... 272 Figure 7.10. nMDS of the dissimilarity in microbial composition by exposure to household cigarette smoke ...... 273 Figure 7.11. nMDS of the dissimilarity in microbial composition by antibiotics use in the previous two weeks ...... 274 Figure 7.12. nMDS of the dissimilarity in microbial composition by ethnicity and vaccination status for a subset of 24 samples plus controls ...... 275 Figure 7.13. Top twenty most important OTUs in random forest models by vaccination status for iTaukei (A) and Indo-Fijian (B) children...... 276 Figure 7.14. Theoretical situation where S. pneumoniae density would not differ between vaccinated and unvaccinated children but the relative abundance of S. pneumoniae would differ by vaccination status ...... 277
xx
List of Tables
Table 1.1. Key virulence factors of S. pneumoniae ...... 6 Table 2.1. Demographic characteristics of children in the pilot FFS by vaccine group ...... 22 Table 2.2. Identification algorithm for S. pneumoniae positive samples in NVEP...... 28 Table 2.3. Primer sequences for vanZ, SP_0137, bguR and fucK ...... 40 Table 2.4. Streptococcal reference isolates used to screen potential pneumococcal targets ...... 40 Table 2.5. 16S rRNA PCR primers for V4 region ...... 43 Table 2.6. Primers for MALDI-TOF MS MLST of S. pneumoniae ...... 46 Table 3.1. Timing of vaccination, nasopharyngeal swabs, and blood draws for each group in FiPP ...... 58 Table 3.2. Demographic characteristics of children in FFS stratified by PCV7 dose and 23vPPV status ...... 71 Table 3.3. Demographic characteristics of children in FFS stratified by 23vPPV status or PCV7 dose ...... 72 Table 3.4. Demographic characteristics of children in FFS stratified by ethnicity ...... 73 Table 3.5. Primer and probe sequences and concentrations for quantitative real-time PCR duplex assays ...... 74 Table 3.6. Cumulative carriage of S. pneumoniae in 23vPPV vaccinated and 23vPPV unvaccinated children ...... 75 Table 4.1. Effect of sequencing depth on richness, before and after rarefaction ...... 83 Table 4.2. Participant characteristics for children included analysis by vaccination status and ethnicity ...... 85 Table 4.3. Participant characteristics for children included analysis by vaccination status ...... 86 Table 4.4. Participant characteristics for children included analysis stratified by ethnicity ...... 87 Table 4.5. Summary of the 10 most abundant OTUs ...... 95 Table 4.6. Percentage of the dominant genus in each of the most common families .. 100 Table 4.7. Comparison of carriage rates in unvaccinated and vaccinated children after stratifying by ethnicity using species-specific qPCR data ...... 112
xxi
Table 4.8. Comparison of carriage rates in unvaccinated and vaccinated children using species-specific qPCR ...... 113 Table 4.9. Comparison of carriage rates in iTaukei and Indo-Fijian children using species-specific qPCR data ...... 113 Table 4.10. Significance of random forest models exploring association between risk factors and microbial composition ...... 123 Table 4.11. Comparison of the top seven genera in unvaccinated and vaccinated iTaukei children following multivariate linear regression analysis ...... 128 Table 4.12. Comparison of the top seven genera in unvaccinated and vaccinated Indo-Fijian children following multivariate linear regression analysis ... 129 Table 4.13. Comparison of the top seven genera in unvaccinated and vaccinated children following multivariate linear regression analysis ...... 130 Table 4.14. Comparison of the top seven genera in iTaukei and Indo-Fijian children following multivariate linear regression analysis ...... 131 Table 4.15. Comparison of the top seven genera in children with and without symptoms of an upper respiratory tract infection following multivariate linear regression analysis ...... 132 Table 4.16. Correlation between relative abundance of the most common genera and OTU richness ...... 133 Table 4.17. Correlation between relative abundance of the most common genera and Shannon diversity ...... 134 Table 5.1. Paired NP and OP sample (n=250) processing results following lytA screen, culture and microarray ...... 152 Table 5.2. Growth on gHBA of NP and OP samples that were positive or equivocal for the lytA gene ...... 153 Table 5.3. Any bacterial growth on gHBA dilution plates for NP and OP samples .... 155 Table 5.4. Number of oropharyngeal samples (n=30) containing various streptococcal species detected by microarray ...... 157 Table 5.5. lytA qPCR identification and serotyping results using microarray, latex sweep and multiplex PCR for a subset of OP samples ...... 163 Table 5.6. Characterisation of isolates (n=91) from a subset of OP samples by various identification and serotyping tests ...... 167
xxii
Table 5.7. Identification and serotyping tests for 10 individual isolates derived from sample FVEP-002-460 ...... 173 Table 5.8. Six OP isolates chosen for whole genome sequencing ...... 176 Table 5.9. Identification of six OP isolates following whole genome sequencing using different several different methods ...... 178 Table 5.10. Comparison of N-acetylmuramoyl-L-alanine amidase genes in a subset of OP isolates ...... 182 Table 5.11. Potential S. pneumoniae-specific targets ...... 189 Table 5.12. Results for potential pneumococcal-specific PCR assays tested against a panel of reference isolates ...... 190 Table 5.13. Results for potential pneumococcal-specific real-time PCR assays tested against a panel of reference isolates ...... 192 Table 5.14. Comparison of results for 108 OP samples for pneumococcal targets alone and in combination ...... 194 Table 5.15. Detection of pneumococci in adult caregiver samples (n=508) by isolation site ...... 196 Table 5.16. Comparison of pneumococcal detection in adult caregivers (n=508) following nasopharyngeal sampling alone or with the addition of oropharyngeal sampling ...... 197 Table 5.17. Microarray serotyping results for the eight participants carrying pneumococci in both the nasopharynx and oropharynx ...... 202 Table 5.18. Participant characteristics of adult caregivers in NVEP for each study year ...... 205 Table 5.19. Participant characteristics of adult caregivers in NVEP stratified by ethnicity (iTaukei and Indo-Fijian) for each study year ...... 209 Table 5.20. Prevalence and unadjusted and adjusted ratios for overall, vaccine type and non-vaccine type pneumococcal carriage in adult caregivers ...... 213 Table 7.1. Prevalence of clusters by risk factors ...... 263 Table 7.2. Individual OP isolate results for identification and serotyping ...... 279 Table 7.3. Full serotyping data for NVEP adult NP samples 2012 to 2015 ...... 286
xxiii
Abbreviations
23vPPV – 23-valent pneumococcal polysaccharide vaccine AIC – Akaike information criterion ANOVA – analysis of variance B – blood BHI – blood heart infusion BLAST – basic local alignment search tool C – class CDC – Centers for Disease Control and Prevention CI – confidence interval Ct – cycle threshold eMLSA – electronic multilocus sequence analysis F – family FFS – Fiji Follow-up Study FiPP – Fiji pneumococcal project G – genus gDNA – genomic DNA gHBA – gentamicin horse blood agar HBA – horse blood agar HPV – human papilloma virus Indo-Fijian – Fijian of Indian descent IPD – invasive pneumococcal disease IQR – interquartile range iTaukei – indigenous Fijian MALDI-TOF MS – matrix-assisted laser desorption/ionisation – time of flight mass spectrometry MCRI – Murdoch Children’s Research Institute MLST – multilocus sequence typing mPCR – multiplex PCR mPPV – micro dose of 23vPPV nMDS – non-metric multidimensional scaling NOD – nucleotide oligomerisation domain
xxiv
NP – nasopharyngeal NSD – no serotype detected NT – non-typeable or unencapsulated lineage NTHi – non-typeable Haemophilus influenzae NVEP – New Vaccine Evaluation Project NVT – non-vaccine type O – order OP – oropharyngeal OPA – opsonophagocytic assay OR – odds ratio OTU – operational taxonomic unit P - Phylum PBS – Phosphate buffered saline PCV – pneumococcal conjugate vaccine PCV7 – 7-valent pneumococcal conjugate vaccine PCV10 – 10-valent pneumococcal conjugate vaccine PCV13 – 13-valent pneumococcal conjugate vaccine PERMANOVA – permutational multivariate analysis of variance PPV – positive predictive value qPCR – quantitative polymerase chain reaction RAST – rapid annotation using subsystem technology rRNA – ribosomal RNA RSV – respiratory syncytial virus SDI – Shannon diversity index SDS – sodium dodecyl sulfate SINA – SILVA incremental aligner Spn – Streptococcus pneumoniae ST – sequence type St dev – standard deviation STGG – skim milk tryptone glucose glycerol TLR – toll-like receptor URTI – upper respiratory tract infection VT – vaccine type
xxv
WHO – World Health Organization
xxvi
Chapter 1
Introduction
1
1 Introduction
1.1 Pneumococcal disease
The bacterium Streptococcus pneumoniae (the pneumococcus) causes a range of diseases. These include common infections such as middle ear infections (otitis media) and pneumonia, as well as more severe and invasive pneumococcal diseases (IPDs) such as sepsis and meningitis. The pneumococcus is also responsible for a range of other diseases such as conjunctivitis (1), bronchitis (2), sinusitis (3) and urinary tract infections (4). More rarely, pneumococci can also infect skin and soft tissue (5), cause osteomyelitis (6), septic arthritis (7), peritonitis (8), endocarditis (9) and pericarditis (10).
S. pneumoniae is a significant cause of morbidity and mortality globally, particularly in children under five years of age. In 2000, it was estimated that there were 14.5 million cases of pneumococcal disease and an estimated 826,000 deaths in children under five years of age worldwide each year (11). Many of these deaths are due to pneumonia, for which S. pneumoniae is a leading cause. In 2011, it was estimated that nearly a third of the 1.3 million pneumonia deaths in children under five years of age were caused by S. pneumoniae (12). Although these estimates are based on modelling and may underestimate the true incidences of morbidity and mortality, they highlight the significance of S. pneumoniae as a global pathogen.
It is generally considered that colonisation of the mucosal membrane in the nasopharynx by S. pneumoniae is a precursor for disease (13, 14). From the nasopharynx, pneumococci can cause diseases such otitis media or sinusitis locally, or move more remotely to cause disease in other sites in the body such as pneumonia and meningitis. However, the initial colonisation (or carriage) of S. pneumoniae is generally asymptomatic. Carriage is common in children under 5 years of age. Progression from carriage to disease is dependent on a range of bacterial and host factors and usually occurs within a month of acquiring a new strain (15).
2 The most susceptible populations to pneumococcal disease include the very young, the very old and those with underlying diseases or comorbidities. A number of other factors can contribute to risk and severity of disease, and include attendance at day-care centres, poor nutrition, lack of exclusive breastfeeding in the first six months of life, socio- economic status, poor air quality (e.g. indoor cooking smoke) and prior infection with influenza particularly during peak influenza season (16-18). In adults, there is also evidence that both cigarette smoking and alcohol intake contribute to an increased risk of IPD (19).
1.2 Streptococcus pneumoniae
The bacterium S. pneumoniae is a Gram positive diplococcus, and is a facultative anaerobe. Discovery of the pneumococcus in 1880 is attributed to both Sternberg (20) who recovered it first, and Pasteur (21) who published his findings first. S. pneumoniae was originally known as Diplococcus pneumoniae due to its diplococcoid appearance, however it was reclassified in 1974 (22).
1.2.1 Identification
Classically S. pneumoniae is identified by phenotypic traits such as the ability to produce green discolouration when grown on blood agar plates (alpha-haemolysis), by sensitivity to the antibiotic optochin, and by solubility in bile salts (bile solubility). The bile solubility test is based on the activation of S. pneumoniae autolysins in the presence of bile salts (i.e. sodium deoxycholate), which results in the ‘soluble’ phenotype of S. pneumoniae. However, these traits do not necessarily provide definitive identification (23, 24). As early as 1912 resistance to optochin in pneumococci had been experimentally induced (25) and by 1917 was found in a clinical isolate (26). Additionally, other closely related streptococcal species are optochin sensitive (27, 28). Pneumococci can also be bile insoluble, and bile solubility can be found in non-pneumococcal species (29, 30). However, the majority of isolated pneumococci are optochin sensitive and bile soluble and these tests are therefore recommended by the WHO (31, 32).
More recently, molecular methods have been used to supplement or replace culture-based methodology for pneumococcal identification. A range of genes have been described as specific to S. pneumoniae including lytA (33), ply (34), psaA (35), Spn9802 DNA
3
fragment (36), rpoA (37), sodA (38), tuf (39), recA (40) and piaB (41, 42). Like the culture-based methods, these pneumococcal genes may be absent in true pneumococci or present in non-pneumococcal species. To overcome the reliance on one single assay for molecular identification multi-target approaches, including whole genome sequencing, matrix-assisted laser desorption ionisation-time of flight mass spectrometry (MALDI- TOF MS) (43) and multilocus sequence typing (MLST) (44, 45) have been developed. Currently, the lytA gene, which encodes an autolysin, is the most widely used molecular assay and is recommended by the WHO for molecular identification of isolates (31).
1.2.2 Virulence factors
The pneumococcus has several important virulence factors which contribute to colonisation and disease. The most significant of these virulence factors is the polysaccharide capsule, which forms the outer protective layer of the bacterium. The polysaccharide capsule prevents complement-mediated opsonophagocytosis of S. pneumoniae by the host immune system. The capsule also inhibits mucus binding through electrostatic repulsion and thus allows S. pneumoniae to bind to epithelial cells rather than succumbing to mucociliary clearance (46). The structure of the capsular polysaccharide is highly diverse within the species with over 90 known capsule types or serotypes (47). The genes encoding pneumococcal polysaccharide capsule synthesis have the same chromosomal organisation for nearly all serotypes and form the cps locus flanked by the genes dexB and aliA (48, 49). Serotypes 3 and 37 are an exception, with the genes encoding capsule biosynthesis of these serotypes located elsewhere in the genome.
The serotype of S. pneumoniae plays an important role in determining invasive potential of a pneumococcal strain. Specific serotypes can differ in their likelihood to be isolated from carriage or from disease (i.e. invasive potential) and in their associations with particular diseases (50, 51). It was estimated that between 1980 and 2007, only six to eleven serotypes accounted for over 70% of the IPD in children under five globally, although there were limited data from populous countries such as China and India (52). Capsular type is associated with resistance to complement mediated immunity (53, 54) and serotypes with thicker capsules tend to be more prevalent in carriage and have a greater ability to avoid neutrophil-mediated killing (55).
4
Many studies have suggested that serotype may be more important than genotype when considering invasive potential (56). However, the genetic background (57), and factors other than capsular serotype, also play an important role in determining the invasiveness and disease potential of an isolate (15, 58). It was originally believed that capsule was necessary for disease and unencapsulated pneumococci were therefore avirulent. However, unencapsulated bacteria have also been shown to cause disease, though far less frequently and usually not systemic disease (59-61). The terms ‘non-typeable’ and ‘unencapsulated’ are sometimes used interchangeably, however in some cases an isolate may be non-typeable due to other reasons (such as the serotyping method used or changes in capsule expression) rather than true absence of capsule. Some isolates may also be ‘typed’ when they are actually unencapsulated if the typing method does not capture mutations that cause termination of capsule biosynthesis. In place of typical cps genes, many true unencapsulated pneumococci have novel genes which have been used to classify lineages of unencapsulated pneumococci (62, 63). The persistence of these lineages suggests that these genes play an important role in compensating for the lack of pneumococcal capsule. This is evidenced in one of these lineages which contains a gene that encodes the virulence factor Pneumococcal surface protein K (PspK, also referred to as NspA) (61, 64). The pneumococcus also has a range of other virulence factors, which contribute to colonisation and disease (65-67) (Table 1.1).
Pneumococci also possess other attributes that contribute to its persistence as a significant global pathogen. Pneumococci have an ability to naturally acquire genetic material from the environment and incorporate it into their genomes through recombination (natural transformation) (68). This can play a significant role in increasing antibiotic resistance and adaption to other environmental stressors (69). In a study where genomes of 240 isolates from the lineage of a multi-drug resistant clone were sequenced, over 700 recombination events were detected, mainly affecting genes encoding major antigens (70). S. pneumoniae is also able to form biofilms. These biofilms are highly structured and enable increased resistance to antibiotics (71).
5
Table 1.1. Key virulence factors of S. pneumoniae. Virulence factor Main roles Capsule, CPS Inhibits complement-mediated opsonophagocytosis; inhibits mucus
binding allowing epithelial binding Pneumolysin, Ply A pore-forming protein with cytolytic activity Pneumococcal surface protein A, PspA Interferes with complement fixation on cell surface Pneumococcal surface protein C, PspC Binds to epithelium and sialic acid. Also known as Choline binding protein A, CbpA Pneumococcal surface adhesin A, PsaA Metal-ion-binding protein with specificity for manganese Autolysin A, LytA Autolytic amidase that can allow the release of pneumolysin, peptidoglycan and techoic acids from pneumococcal cells Neuraminidase A, NanA Important for epithelial adherence and plays a role in colonisation and otitis media infections Phosphorylcholine, ChoP Mediates adherence to rPAF on nasopharyngeal epithelial cells and anchors choline-binding proteins to cell well Hyaluronate lyase, Hyl Breaks down hyaluronan-containing extracellular matrix components Pneumococcal adhesion and virulence Binds fibronectin, a component of the A, PavA extracellular matrix Enolase, Eno Binds plasminogen, a component of the extracellular matrix Table adapted from Kadioglu, et al. (65). rPAF, receptor for platelet-activating factor
6
1.3 Pneumococcal carriage
S. pneumoniae colonises the mucosal membrane of the upper airways. In children, pneumococcal colonisation predominantly occurs in the nasopharynx, however pneumococci can also be found in the nares and the oropharynx (72). In contrast, carriage of S. pneumoniae in adults is thought to be more evenly distributed between the nasopharynx and the oropharynx (72). Multiple strains and/or serotypes of pneumococci can be carried simultaneously.
Pneumococci are spread from colonised individuals to uncolonised individuals through respiratory secretions. Transmission has mainly been studied in murine models, and these suggest that transmission of pneumococci is influenced by the bacterial load of pneumococci, the pneumococcal serotype, the inflammatory response in the uncolonised mice and the presence of other pathogens such as influenza A (73-76). In humans, the microbial composition of the upper airways influences the acquisition of pneumococci (77). Acquisition and carriage of pneumococci is also influenced by season with greater transmission in cooler and drier climates, although it is unclear whether this may be linked to viral coinfection during cooler seasons (78).
Pneumococcal carriage rates vary with age, generally peaking at 1-2 years of age (79-81), with the first episode of pneumococcal colonisation occurring as early as the first weeks of life (82, 83). Co-colonisation with multiple serotypes is more frequent at a younger age (84). Pneumococcal carriage rates and serotype distributions vary between countries, regions and within different populations living in the same geographic area. While differing laboratory methods can affect estimates of pneumococcal carriage, generally, the rates of carriage are higher in developing countries and lower in developed countries (85). For example, in 2006 carriage rates in The Gambia were as high as 72% overall and 93% in infants less than a month old (80). In contrast, in Swedish infants one to three months of age, the carriage rate was 12% (86). Environmental and socioeconomic factors such as crowding (particularly attendance at day care centres), family size (particularly having a younger sibling), exposure to tobacco smoke, family income and lack of antibiotic use prior to screening are all risk factors for pneumococcal carriage (13, 87- 89).
7
1.4 Treatment and prevention
Treatment of pneumococcal diseases is mainly through the use of antimicrobial therapies but can also involve supportive therapies such as oxygen supplementation for pneumonia patients (90). Guidelines for treatment vary geographically and are dependent on setting and the specific infection (91-93). As with many other microbes (94), resistance to antimicrobials is a significant barrier to successful treatment of pneumococcal disease (95, 96). Even when treatment is successful for infections there can be significant morbidity. For example, Goetghebuer et al. found that 58% of children surviving pneumococcal meningitis were permanently affected with problems such as hearing loss, mental retardation, motor abnormalities and seizures (97).
Pneumococcal disease can be reduced, to some degree, by controlling risk factors such as improving nutrition, air quality and hygiene. Exclusive breastfeeding in the first months of life can reduce morbidity due to acute lower respiratory tract infection (98), and reduce otitis media episodes (99). Reducing indoor air pollution (100) and regular handwashing with soap (101) can also reduce pneumonia infections. However, the key to preventing pneumococcal disease worldwide is through vaccination.
1.5 Early development of a pneumococcal vaccine
Klemperer and Klemperer (102) first demonstrated the protective ability of patient serum against pneumococcal infection in 1893. The first vaccine was introduced by Wright, et al. (103) in 1911 using whole heat-killed pneumococci, to reduce the large burden of pneumococcal disease in South African gold miners. However, the vaccine composition and dosage were inadequate to provide protection. A few years later Lister (104) determined an adequate dose of heat-killed pneumococci and experimented with different polyvalent vaccines. Owing to flaws in his study design whereby he vaccinated one group of miners and used another groups as a control (even though different groups were known to have different attack rates), the efficacy of his vaccine wasn’t established. In 1926, Felton and Bailey were able to isolate capsular polysaccharide allowing pneumococcal capsule to provide the basis for future vaccines (105). After several vaccine trials by multiple groups (106), the efficacy of capsular polysaccharide vaccines was demonstrated in 1945 by MacLeod et al. (107). Two hexavalent polysaccharide vaccines (108) were licensed in 1948. However, after the development of successful pneumococcal antibiotic
8
therapies these vaccines were later withdrawn from the market due to low demand (106). With the emergence of multiply resistant pneumococci (109) and continuing high morbidity and mortality, a 14-valent polysaccharide vaccine was licensed in 1977 and introduced in 1978 (110). Increasing knowledge of incidences of disease caused by capsular types led to a new 23-valent polysaccharide vaccine in 1983 (111).
1.6 The 23-valent pneumococcal polysaccharide vaccine
The 23-valent pneumococcal polysaccharide vaccine (23vPPV) contains 25 µg/serotype of purified capsular polysaccharides 1, 2, 3, 4, 5, 6B, 7F, 8, 9N, 10A, 11A, 12F, 14, 15B, 17F, 18C, 19A, 19F, 20, 22F, 23F and 33F. Polysaccharide vaccines produce unreliable and inconsistent antibody levels in children less than two years old (112-114). Pneumococcal polysaccharides elicit a T-cell independent immune response and thus are poorly immunogenic in children less than two years due to their immature immune systems (115, 116). Immune hyporesponsiveness, where exposure to an antigen (such as polysaccharides in a vaccine) results in a diminished immune response on re-exposure, was observed for some serotypes following the use of pneumococcal polysaccharide vaccines (116). It is theorised that memory B cells are stimulated in a T-cell independent response but not replenished, and this depletion of memory cells results in diminished responses on subsequent exposure to the polysaccharide antigen resulting in immune hyporesponsiveness.
As well as inconsistent antibody levels, several studies found that polysaccharide vaccination had little effect on carriage in young children (117) or an effect on diseases such as otitis media (118). While vaccine-related reductions in otitis media (119, 120) and acute lower respiratory tract infections (121) have been observed, there were inconsistencies between studies for children vaccinated at less than two years of age (120).
By 1996 a pentavalent protein conjugate vaccine, which would elicit a T-cell dependent immune response and the development of memory B-cells, was being assessed for safety and immunogenicity for use in children less than two years of age (122). Positive results in this study - and in several others (123-125) over the next few years - led to licensure of a heptavalent protein conjugate vaccine in the US in 2000. The 7-valent pneumococcal
9
conjugate vaccine contained purified polysaccharides for serotypes 4, 6B, 9V, 14, 18C, 19F and 23F (2 µg/serotype except 6B at 4 µg) conjugated to a non-toxic variant of the diphtheria toxoid, CRM197 (CRM, cross-reactive material).
Following introduction of the protein conjugate vaccine, the polysaccharide vaccine was no longer recommended for children under the age of two (126), and now is only recommended for individuals at higher risk of pneumococcal infection including those aged over 65 (115, 127). There has also been interest in immunising mothers during pregnancy with 23vPPV to provide protection for infants, however the effectiveness of this approach has not been demonstrated (128).
1.7 The 7-valent pneumococcal conjugate vaccine
1.7.1 Direct effects of PCV7
The introduction of PCV7 into the infant immunisation schedule in the USA in 2000 proved highly successful with reductions in the incidence of IPD (129), pneumonia (130) and otitis media (131). By 2007, the incidence of IPD in the USA was 76% lower than pre-vaccine levels in children <5 years of age (129). Following early reports of the success of PCV7 in the USA (132), a few years later it was introduced into the immunisation schedules of the UK and other European countries where reductions in IPD were also observed (133-136).
1.7.2 Indirect effects of PCV7
In addition to reducing IPD in vaccinated children, reductions in IPD were observed in other unvaccinated age groups as an indirect effect of pneumococcal vaccination (129). In England and Wales, four years after PCV7 was introduced, vaccine type IPD in adults ≥ 65 years of age was 81% lower than pre-vaccine introduction levels (133). The introduction of PCV7 also led to a reduction in the rate of IPD caused by antibiotic- resistant pneumococci (137), due to the most of the serotypes with high-level antibiotic resistances being included in the vaccine (138).
10
1.7.3 Serotype replacement
Within several years of PCV7 introduction in the USA, it was evident that the reductions in IPD incidences caused by serotypes included in PCV7 were, to some extent, being offset by increases in incidences of IPD caused by serotypes not included in the vaccine (129, 139). This pattern of serotype replacement was observed in many other settings where PCV7 had been introduced. In England and Wales, vaccine type IPD in children <2 years of age was 98% lower following PCV7 introduction, however non-vaccine type IPD increased by 68% over the same period (133). Following PCV7 introduction serotype 19A, in particular, became a predominant cause of IPD in many settings (140). However, while the degree of serotype replacement varied between settings, the increase in non- vaccine type IPD was generally smaller than the reductions in vaccine type IPD resulting in a net benefit of PCV7 (141), although that benefit was sometimes only minor in some populations such as Alaskan Native children (142).
1.8 The post-PCV7 era
The issue of serotype replacement led to widespread replacement of PCV7 in immunisation schedules with pneumococcal conjugate vaccines of higher valency (containing 10 or 13 serotypes). The 10-valent pneumococcal conjugate vaccine (PCV10), Synflorix® (GlaxoSmithKline), is also sometimes referred as PHiD-CV (pneumococcal and non-typeable Haemophilus influenzae protein D – conjugate vaccine) and was licensed for use in the European Union in 2009. PCV10 includes serotypes 1, 4, 5, 6B, 7F, 9V, 14, 18C, 19F and 23F (1 µg/serotype, except 4, 18C and 19F at 3 µg/serotype). Serotypes 1, 4, 5, 6B, 7F, 9V, 14 and 23F are conjugated to protein D, an outer membrane protein from non-typeable Haemophilus influenzae, serotype 19F is conjugated to CRM197 and serotype 18C is conjugated to a tetanus toxoid variant. The 13- valent pneumococcal conjugate vaccine (PCV13), Prevenar 13®(Pfizer), was licensed in 2010 for use in the US (143) and contains serotypes 1, 3, 4, 5, 6A, 6B, 7F, 9V, 14, 18C,
19A, 19F and 23F (2.2 µg/serotype except 6B at 4.4 µg/serotype) conjugated to CRM197. Since 2012, the WHO recommends the use of PCVs (10- and 13-valent) in national immunisation programs, particularly in countries where there are ≥50 deaths per 1,000 births (126). Different schedules affect the level of antibody produced (144), with a 3 dose primary schedule recommended (or a 2 dose primary plus booster dose as an alternative) recommended by the WHO.
11
After five years of PCV13 use in the USA, vaccine type IPD was 88% lower in children under 5 years of age compared with pre-PCV13 introduction and no significant increases in non-vaccine type IPD were observed (145). In England and Wales, where PCV13 also replaced PCV7, IPD reductions were observed for vaccine type IPD in all age groups, however there were also significant increases in IPD caused by non-PCV13 types (146). Reductions in IPD in both settings were largely driven by reductions in IPD caused by serotypes 19A and 7F. Reductions in IPD have also been observed in countries where PCV10 replaced PCV7 (147, 148) and in countries where PCV10 and/or PCV13 were introduced without prior PCV7 vaccination (140).
PCV10 or PCV13 has been introduced into the national immunisation programs in over 135 countries. Overall, PCVs are effective at reducing pneumococcal disease. However, issues with current vaccine strategies remain, particularly serotype replacement and the poor effectiveness of PCV against specific serotypes such as serotype 3 (149).
To combat these issues, new vaccines with higher valency, such as a PCV15 vaccine, or new formulations of conjugate vaccines are in clinical development (150). Serotype- independent approaches are also being investigated with clinical trials underway for an inactivated whole cell vaccine derived from an unencapsulated strain, and for a trivalent vaccine comprising pneumococcal histidine triad protein D (PhtD), pneumococcal choline-binding protein A (PcpA) and pneumolysoid (150).
1.9 Pneumococcal vaccination and carriage
The indirect effects of pneumococcal vaccination on disease and serotype replacement are both driven by changes in pneumococcal carriage and transmission following pneumococcal vaccination. By reducing carriage of vaccine type pneumococci in vaccinated children, transmission of vaccine type pneumococci to unvaccinated individuals is reduced and therefore results in a reduction in vaccine type pneumococcal disease. Changes in pneumococcal serotypes circulating in a population following pneumococcal vaccination also allow the opportunity for new serotypes to colonise and cause pneumococcal disease, replacing the vaccine types.
12
While measurement of serotype-specific IPD within a population is currently the gold standard for assessing the impact of PCV, carriage studies also have an important role to play, particularly in developing countries where obtaining IPD data can be challenging. The infrastructure and diagnostic requirements can be prohibitive to measuring IPD in a developing country, requiring the collection and culture of sterile site specimens (which may be affected by prior antimicrobial use). IPD is also a relatively rare event and therefore countries with small populations may struggle to estimate declines in IPD following vaccine introduction.
Other metrics such as pneumonia hospitalisations (which are generally much more common than IPD) are also used to measure vaccine impact. However, there are challenges in proving the pneumococcus as the cause of the pneumonia, as diagnosis of pneumococcal pneumonia is complex (151). Therefore, pneumonia hospitalisations are often reported as all-cause pneumonia. As such, this reporting does not reflect the impact of vaccination on vaccine type pneumococcal pneumonia but rather the net effect of pneumococcal vaccination on all pneumonia (which may include serotype replacement).
Given that carriage of pneumococci precedes pneumococcal disease, measurement of pneumococcal carriage is a biologically relevant approach to measure the impact of pneumococcal vaccination in a population. Carriage samples are relatively easily obtained and can be collected for large cross-sectional or longitudinal studies. Improvements have also been made over the last two decades towards standardising carriage study methodology and methods for detecting and serotyping multiple strains present in individuals (31, 152-154). In addition to measuring changes in circulating serotypes following vaccine implementation, carriage studies have a role in demonstrating indirect effects, predicting serotype replacement and the impact of PCV on cross-reactive serotypes (155, 156), as well as identifying new serotypes (157, 158). Carriage can be used also measure antimicrobial resistances which can then be used to inform treatment strategies.
Reduction in carriage of vaccine types post-introduction of PCV has been evident in many populations (159-161). Reductions in vaccine type carriage have also been observed in
13
several studies exploring reduced-dose schedules (162) with reductions in carriage with as little as one dose (163). As with pneumococcal disease, pneumococcal vaccination results in reduction of vaccine-type carriage for unvaccinated individuals, which has been seen in siblings and adults within the same household (164, 165), and at a population level (166, 167). Serotype replacement is also evident in carriage studies with reductions in vaccine type pneumococcal carriage after vaccination completely offset by increases in carriage of non-vaccine types in many populations (160, 161, 168). However, complete serotype replacement in carriage does not necessarily equate to complete serotype replacement in pneumococcal disease due to differing invasive potential of pneumococcal serotypes (169, 170). For example, Flashe et al. observed that changes in vaccine type pneumococcal carriage were predictive of the incidence of vaccine type pneumococcal otitis media, but non-vaccine type replacement in the nasopharynx did not directly correlate with non-vaccine type otitis media (171). However, pre-PCV carriage data, ideally combined with IPD data, can be used to predict the potential impact of PCVs on disease burden in a population (172, 173).
1.10 Effect of pneumococcal vaccination on other pathogens
Carriage studies also provide an opportunity to examine whether the reduction or elimination of vaccine type pneumococci from the nasopharynx (or oropharynx) through the introduction of vaccines gives rise to increases in other bacterial species sharing the same niche. Species replacement is of particular concern if the replacement species are potential pathogens such as Staphylococcus aureus, Haemophilus influenzae and Moraxella catarrhalis.
To date, there is still uncertainty as to whether pneumococcal vaccination has a wider impact on other bacterial species (174-176). Early last decade it was observed in two independent studies (177, 178) that there was a negative relationship between carriage of S. pneumoniae and S. aureus, in particular vaccine type S. pneumoniae. Subsequently there was concern that this negative association could theoretically result in species replacement by S. aureus following pneumococcal vaccine introduction. Around the same time, as part of a randomised study in children with recurrent acute otitis media, Veenhoven et al. found S. aureus was cultured more frequently from middle ear fluid in
14
children that were vaccinated with PCV7 compared with children in the control group (179).
The negative association between S. pneumoniae and S. aureus has since been observed in many other studies and settings (180-184) but the evidence for species replacement by S. aureus following pneumococcal vaccine introduction is less clear. Two other studies in the Netherlands observed increases in S. aureus nasopharyngeal (NP) carriage associated with pneumococcal vaccination (185, 186), although in one study this increase was only observed in children aged 12 months and was not sustained (186). In contrast, several studies have found no effect of PCV on S. aureus colonisation (184, 187-189). While pre-vaccine introduction data was not available, Lee et al. found that S. aureus (and methicillin-resistant S. aureus) colonisation in children <7 years old was stable from 2003-04 to 2006-07 as part of a study of NP colonisation in children (190).
Our understanding of why species replacement may occur in some circumstances, but not in others, is limited by our understanding of what drives the negative relationship between the two bacterial species. The bactericidal effect of hydrogen peroxide production by S. pneumoniae and catalase production by S. aureus have been implicated (191) but these fail to fully explain the observed patterns (192, 193). S. aureus genotype does not appear to be a significant factor (194), however the pneumococcal pilus may play a role (195). As pilus genes are more likely to be present in vaccine type pneumococci than non- vaccine type pneumococci (196), this may explain why the negative association between S. aureus and the pneumococcus is limited to vaccine type pneumococci in some studies. There also seems to be an immune component which influences the negative interaction between the two bacterial species, as Madhi et al. found this relationship between S. aureus and S. pneumoniae was observed in HIV-uninfected children but not HIV-infected children (184).
It should also be noted that the carriage epidemiology of S. pneumoniae and S. aureus are quite different. S. pneumoniae carriage rates peak in children 1-2 years of age, whereas S. aureus carriage peaks in children <6 months of age and in older children and adolescents (190, 197). In addition, S. aureus is more commonly found in the anterior nares of children and adults rather than the nasopharynx (197, 198).
15
H. influenzae and M. catarrhalis are key respiratory pathogens that colonise the nasopharynx, and positive associations with pneumococci have generally been reported (89, 199-201). However, Xu et al. found S. pneumoniae had a positive relationship with M. catarrhalis and H. influenzae in healthy children but a negative relationship during acute otitis media infection (202). As with S. aureus, the underlying mechanisms that influence co-colonisation of H. influenzae and M. catarrhalis with S. pneumoniae are unclear and the evidence surrounding species replacement by these potential pathogens following pneumococcal vaccination is also uncertain. Spijkerman et al. found higher rates of H. influenzae carriage in children and their parents following PCV7 introduction, with M. catarrhalis also higher at 4.5 years post-PCV introduction but only in children at 24 months of age (185). van Gils et al. found PCV7 vaccination had no overall effect on either H. influenzae or M. catarrhalis despite slight reductions in overall pneumococcal carriage, although there was some evidence M. catarrhalis carriage was lower in children at 12 months that had received a 2+1 schedule compared with no vaccine (203).
Unlike PCV7 and PCV13, PCV10 includes non-typeable H. influenzae (NTHi) protein D as a carrier protein and therefore might be expected to reduce carriage of NTHi, or H. influenzae more broadly. Although PCV10 elicits an antibody response to protein D (204, 205), there inconsistent evidence that PCV10 reduces carriage of H. influenzae. Hammitt et al. found lower carriage of NTHi in Kenyan children following PCV10 introduction, however it was unclear as to whether PCV10 caused the reduction in H. influenzae carriage as H. influenzae carriage seemed to recover in year 2 of the vaccine period and was not associated with a child’s vaccine status (189). Prymula et al. found reduced NTHi carriage in the PCV10 vaccinated children aged 24-27 months as part of a randomised control trial in the Czech Republic, although no consistent effect on NTHi carriage was observed over the study period (206). Leach et al. did not observe a difference in NTHi carriage rates between PCV7 and PCV10 vaccinated children but did note that the prevalence of ear discharge infected with NTHi was lower in children vaccinated with PCV10 compared with those vaccinated with PCV7 (207). Two studies have found no effect of PCV10 on NTHi carriage (208, 209), whereas Brandileone et al. noted an increase in NTHi carriage in Brazilian children following PCV10 introduction. A recent study by Andrade et al. suggested that while PCV10 may not have an effect on carriage
16
rates, PCV10 may reduce the metabolic activity of H. influenzae carried in vaccinated children (210).
With improvements in microbiome research over the last decade, researchers have been able to explore the question of species replacement and the wider impact of pneumococcal vaccination on the NP microbiota. There is some evidence that PCV may alter NP bacterial composition and diversity, however findings have not been consistent across all studies (211-213). In a study by Biesbroek et al. PCV7 had an effect on microbial composition and diversity in Dutch children at 12 months but not at 24 months, which suggests that PCV7 may have a short-term impact on the microbiome (211). However, Feazel et al. found PCV10 had no effect on the microbiome of Kenyan children 180 days after vaccination which suggests that perhaps other factors such as setting are also important (213). The study by Biesbroek et al. also highlighted the complex network of bacterial interactions within the NP microbiome and how vaccination can cause shifts in these interactions, suggesting that perhaps indirect interactions with other bacterial species may influence the interactions between pneumococci and respiratory pathogens such as S. aureus.
1.11 Challenges for implementing pneumococcal vaccines
Despite the general effectiveness of vaccination, there are several factors which can limit the implementation of vaccines, particularly in the resource-limited settings. Vaccine costs, program costs, little or incomplete knowledge of disease burden, and the ability to target populations with the vaccine, are all important issues (214). Other challenges include available infrastructure (215), serotype coverage (85) and the expertise to evaluate cost effectiveness and health benefits (216). Although modelling can be used to assess cost effectiveness and compare health outcomes (217, 218), challenges remain in estimating the relative impact of serotype replacement, indirect effects and efficacy against otitis media amongst other things (219).
Since 2000, Gavi, the Vaccine Alliance, a public-private health organisation, have been financially supporting low- and some middle-income countries introduce vaccines, such as PCV, into national immunisation programs. By the end of 2015 the proportion of countries eligible for Gavi support that had introduced PCV was similar to the proportion
17
of high-income countries that had introduced PCV (74% Gavi countries vs. 76% of high- income countries) (220). However, the rate of PCV introduction in middle-income countries is much lower and is reflective of challenges of competing health priorities and the difficulties faced when a country may ‘graduate’ from Gavi support (216). It is now estimated that two thirds of the world’s poor live in middle income countries and that the majority of vaccine preventable deaths also occur in middle income countries (such as China and India) (221). A recent review of pneumococcal vaccine implementation in middle-income countries highlighted some of the key barriers to implementation, including the lack of knowledge on country-specific disease burden and surveillance (216). It was noted that this could be improved by post-introduction surveillance to monitor the impact of PCV and changes in serotype distribution over time.
1.12 Fiji
Fiji is a middle-income country located in the South Pacific Ocean and is not eligible for Gavi support. Fiji is comprised of more than 330 islands and has an estimated population of over 850,000 people (http://www.statsfiji.gov.fj/). Pneumonia is a significant burden on Fijian society and is the most common reason for visiting outpatient facilities for children under five (222). Like other settings around the world, S. pneumoniae is a major cause of bacterial pneumonia and bacterial meningitis in Fiji (223, 224).
There are two major ethnic groups in Fiji, indigenous Fijians (iTaukei) and Fijians of Indian descent (Indo-Fijians or FID). Significant differences in the burden of disease between ethnic groups have been observed for a number of diseases including lower respiratory tract infections, diabetes, cervical cancer, acute rheumatic fever and rheumatic heart disease (223, 225-227). The two ethnic groups also have differences in pneumococcal carriage rates in children with significantly higher pneumococcal carriage in iTaukei children compared with Indo-Fijian children (176, 228).
Following an earlier vaccine trial in Fiji which examined reduced-dose schedules of PCV7 and the use of 23vPPV to improve serotype coverage (228-230), PCV10 was introduced in the national immunisation schedule in late 2012. As part of this
18
introduction, monitoring of disease and carriage pre- and post-PCV10 introduction was conducted to measure direct and indirect effects of vaccine introduction.
1.13 Aims
This thesis focuses on pneumococcal vaccination and carriage in the context of Fiji to address three aims
Aim 1. To determine whether 23-valent pneumococcal polysaccharide vaccine given at 12 months of age had a long-term impact on carriage of pneumococci and other important respiratory pathogens.
Aim 2. To investigate whether 7-valent pneumococcal conjugate vaccine had an effect on other microbiota and whether this effect differed by ethnicity.
Aim 3. To examine whether the introduction of the 10-valent pneumococcal conjugate vaccine into the immunisation schedules also provided an indirect benefit for unvaccinated individuals, assessed by examining changes in pneumococcal carriage in adults.
19
Chapter 2
Material and Methods
20
2 Materials and methods
2.1 Introduction This chapter details the materials and methods used in this thesis. Note that an abbreviated methods section is also included in Boelsen et al. (231), the publication within chapter 3.
2.2 Study designs and swab collection
2.2.1 Fiji Pneumococcal Project
The swabs used in work described in chapter 4 were collected as part of the Fiji pneumococcal project (FiPP), a phase II vaccine trial in Suva, Fiji (228-230). 554 infants were enrolled into the FiPP and randomised into eight groups receiving zero, one, two or three doses of PCV7, with or without a booster dose of 23vPPV at 12 months of age (see Table 3.1 for timing of vaccine doses and NP swab collection).
In chapter 4, we use NP swabs collected from children aged 12 months who received either three doses of PCV7 or no PCV7 (229). Thirty-six samples were randomly selected from each of the four groups – vaccinated iTaukei, unvaccinated iTaukei, vaccinated Indo-Fijian and unvaccinated Indo-Fijian. A sample size of 36 was chosen as this was maximum number of swabs available from Indo-Fijian vaccinated and unvaccinated children at 12 months of age. At the time of enrolment in FiPP, consent was given to use swabs for future pneumococcal research and ethics approval for the work described in this thesis was given by Health Sciences Human Ethics Sub-Committee at The University of Melbourne (Ethics ID #1543881).
As well as the participant information collected at enrolment (date of birth, sex, ethnicity), the following participant information was collected by the nurse and from the parent/guardian at the time of swab collection: weight, exposure to household cigarette smoke, breastfeeding status and/or age when breastfeeding was stopped, whether the child had received antibiotics in the preceding 2 weeks, whether the child had any current symptoms of upper respiratory tract infection such as cough or rhinitis (including allergic rhinitis), whether the child had hay fever, whether the child had an ear infection, whether there was any ear discharge and if so, how many days of ear discharge.
21
2.2.1.1 Swab collection and storage Swabs were collected from study participants at 12 months of age between 2005 and 2007. Details of the swab collection, storage and transport are given in Russell et al. 2010 (229). In brief, buffered cotton swabs (aluminium shaft-buffered; Sarstedt, Australia) were inserted into the nasopharynx for 5 s then rotated before removing and placing the swab into a sterile cyrovial tube (Simport, Canada) containing 1 ml of the swab storage media STGG (152). After chilled transport to the Colonial War Memorial Hospital in Suva, Fiji, the swabs were stored at -70°C. Swabs were transported on dry ice to the Pneumococcal Research Laboratory at MCRI in Melbourne, Australia and stored at - 80°C.
2.2.2 Fiji Follow-up study
The work described in chapter 3 uses samples collected as part of the Fiji Follow-up study (FFS) as well as those collected as part of the pilot study (pilot FFS). In 2010, the pilot FFS was conducted to optimise methods, where 21 NP swabs were collected from previous FiPP participants (Table 2.1). Between 2011 and 2012, 194 swabs were collected from previous FiPP participants as part of the FFS (participant demographics shown in 3.3.1 Supplemental content, Table 3.2, Table 3.3 and Table 3.4).
Table 2.1. Demographic characteristics of children in the pilot FFS by vaccine group. Parameter at Groups swab collection A B C D E F G H No. of children 4 3 3 3 1 4 2 1 Vaccine (dose) PCV7 3 3 2 2 2 2 1 1 23vPPV 0 1 0 1 0 1 1 0
Age (y) Median 5.1 4.6 4.7 4.3 5.6 4.6 4.4 5.9 IQR 4.4-5.8 4.5-5.1 4.5-5.2 4.3-4.5 n/a 4.6-4.7 4.3-4.5 n/a
Ethnicity, n (%) iTaukei 2 (50) 3 (100) 2 (67) 3 (100) 1 (100) 3 (75) 1 (50) 0 (0) Indo-Fijian 2 (50) 0 (0) 1 (33) 0 (0) 0 (0) 1 (25) 1 (50) 1 (100) Male, n (%) 1 (25) 1 (33) 1 (33) 1 (33) 0 (0) 2 (50) 0 (0) 0 (0)
22
2.2.2.1 Nasopharyngeal swab processing and storage Swab collection and storage was performed in line with WHO guidelines (31, 153), although both FiPP and FFS used cotton swabs. Following NP swab collection (as described above, 2.2.1.1), swabs were then transported in a chilled carrier to the National Centre for Communicable Disease Reference Laboratory, Mataika House, Suva, Fiji. Swabs were vortexed and an aliquot of 200 µl was taken from the original STGG swab culture. The swab and the aliquot were then stored at -80°C. The swab and aliquot were transported to the Pneumococcal Research laboratory at MCRI on dry ice and stored at - 80°C until analysis.
It should be noted that both FFS and FiPP used cotton swabs for nasopharyngeal swab collection, which is not recommended by the WHO as it may include inhibitors to pneumococcal growth (31). However, there is no evidence that cotton swabs interfere with molecular methods for detection of pneumococci which were used in this study.
2.2.3 New Vaccine Evaluation Project
The swabs used in chapter 5 were collected as part of the Fiji New Vaccine Evaluation Project (NVEP). NVEP evaluated the introduction of three vaccines (PCV10, rotavirus vaccine and human papillomavirus vaccine) into the immunisation program in Fiji, with annual cross-sectional carriage surveys to evaluate the effect of PCV10 introduction on carriage. As part of NVEP, swabs were collected between 2012 and 2015 from the following age cohorts: 5-8 weeks of age, 12-23 months of age, 24-83 months of age and the adult caregivers of the infants and children asked to participate. NP swabs were collected from all age groups and OP swabs were also collected from adults. The sample size for NVEP was based on an estimated baseline vaccine type carriage rate of 16% in children 12-23 months of age. With a sample size of 281, NVEP would have 90% power (at a 5% significance level) to show a ~50% reduction in vaccine type carriage (from 16% to 7%). However, in the unvaccinated age groups, as well as the first few years after PCV10 introduction, the effect size was likely to be lower, therefore the sample size was increased to 500 for all groups (Dunne, et al. manuscript under review). Participants were recruited to ensure adequate representation of ethnic and geographic demographics in Fiji, with approximately half of participants being recruited from rural areas and 60% being
23
of the indigenous (iTaukei) ethnicity. Study participant demographics were also collected at the time of swab collection.
2.2.3.1 Swab collection and storage Paired NP and OP swabs were collected from adult caregivers one year pre-PCV10 vaccination, and up to three years post-vaccine introduction (in 2012, 2013, 2014 and 2015 (n=508, 511, 516 and 509, respectively)). Swabs were collected following WHO recommendations (31) using nylon-tipped flocked swabs. NP swab were collected as described above (2.2.1.1). For OP swabs, the swab was swept across the back of the oropharynx including both tonsils and any exudate present, rotating while swabbing. Swabs were placed in 1 ml of STGG and transported in an icebox to the Mataika House laboratory at the Fiji Centre for Communicable Disease Control in Suva, Fiji. On arrival at the laboratory, swabs were vortexed for 20 s before 275 μl of the STGG was aliquoted into two fresh cryovials each. The original swab and aliquots were then stored at -80°C. Samples were transported on dry ice to the Pneumococcal Research Laboratory at MCRI in Melbourne, Australia and stored at -80°C
2.3 Bacterial culture
2.3.1 FFS culture
For both the pilot FFS and FFS, 50 µl from the aliquot of the NP swab culture was diluted 1:2 in STGG. Four ten-fold serial dilutions were then prepared in saline, and 50 μl of the 1:2 and 100μl of all other dilutions were plated onto Columbia Horse Blood Agar plates supplemented with 5 µg/ml of gentamicin (gHBA; Oxoid brand, Thermo Fisher
Scientific, Australia) in duplicate. Plates were incubated for 36-44 h (37°C, 5 % CO2) before being examined for alpha-haemolytic colonies. Viable counts were conducted, and plates were divided into 8 segments. A segment was then selected from a prepared random list of numbers to choose which segment from which to select colonies. The two outermost well separated α-haemolytic colonies in the randomly selected segment, plus any additional morphologically distinct α-haemolytic colonies, were then subcultured onto Columbia Horse Blood Agar plates (HBA; Oxoid brand, Thermo Fisher Scientific,
Australia) and incubated for 24 h (37°C, 5 % CO2).
24
2.3.2 NVEP culture
Initially, 50 μl was plated on Horse Blood Agar plates containing 5 μg/ml of gentamicin (gHBA, Oxoid brand, Thermo Fisher Scientific, Australia), followed by 50 μl of a 1:10 and 1:100 serial dilution in Brain Heart Infusion (BHI) broth. gHBA plates were incubated overnight at 37°C, 5% CO2, and plates were examined for α-haemolytic growth and growth of other bacterial morphologies. Bacterial growth was harvested in 1 ml phosphate buffered saline (PBS) from the dilution plate that was one dilution higher than the plate with confluent or near confluent growth. The resultant scraping was spun at 11,300 x g for 5 min, the supernatant was removed, and the pellet stored at -30°C until DNA extractions were performed.
Where >25 α-haemolytic colonies were present on gHBA plates, a single colony was also subcultured overnight at 37°C, 5% CO2 on HBA plates with an optochin disc (see 0). Pure cultures were then stored in 1 ml of STGG at -80°C. For OP samples, where multiple colony morphologies were present, a colony was selected based having a pneumococcal- like appearance. If the isolated colony was subsequently found to optochin sensitive, serotyping was performed using latex ‘sweep’ agglutination (see 2.5.2).
2.3.3 Culture of OP samples for isolation of individual colonies
The complexity of the OP samples was explored through the thorough examination of a subset of samples in chapter 5. Eleven samples were chosen based on lytA qPCR and microarray serotyping results. The sample culture method described above (2.3.2 NVEP culture) was repeated, and a representative of each colony morphology on any of the dilution plates (neat, 1:10 and 1:100) for each sample was then subcultured onto HBA and incubated overnight (37°C, 5% CO2). Colonies were characterised by haemolysis, shape, size, elevation, surface texture and colour.
25 2.4 Identification
2.4.1 S. pneumoniae identification in FFS
Identification of all isolates in FFS was performed following overnight culture on HBA plates (37°C, 5% CO2). Identification of S. pneumoniae was performed following the algorithm shown in Figure 2.1 using pure cultures of individual α-haemolytic colonies.
2.4.2 S. pneumoniae identification in NVEP
2.4.2.1 General identification algorithm NVEP samples positive for S. pneumoniae were identified based on results from lytA qPCR (see 2.7.2), culture on gHBA plates (see 2.3.2) and microarray testing (see 2.5.4). For the entire NVEP study, including the other age cohorts not examined in this thesis, a standard identification algorithm (Table 2.2) was applied (including section 5.2.2 in this thesis).
26 α-haemolytic colony
optochin susceptibility
susceptible non-susceptible
serotyping
typeable non-typeable
Phadebact® Pneumococcus test
Positive Negative Inconclusive
Spn or Spn or Soluble Spn non-Spn non-Spn Spn or Spn or Insoluble Spn non-Spn non-Spn Spn or Spn or Bile solubility Inconclusive Spn non-Spn non-Spn
all non-typeable presumptive Spn and non-Spn
lytA qPCR
positive negative
S. pneumoniae non - S. pneumoniae
Figure 2.1. Identification algorithm for pure isolates of S. pneumoniae in FFS. Isolates are identified as either S. pneumoniae (Spn) or non-S. pneumoniae (non- Spn). Adapted from Satzke et al. (31).
27 Table 2.2. Identification algorithm for S. pneumoniae positive samples in NVEP. Test results Interpretation lytA qPCR Culture on Microarray2 Pneumococci Serotyping (see 2.7.2) gHBA1 (see 2.5.4) present in results (see 2.3.2) sample? available Positive Spn Yes Yes α-haemolytic (Ct <35) Non-Spn No N/A growth Negative Yes No2 No α-haemolytic N/A Yes No3 growth Equivocal Spn Yes Yes α-haemolytic (Ct 35-<40) Non-Spn No N/A growth Negative No N/A No α-haemolytic N/A No N/A growth Negative N/A N/A No N/A (Ct ≥40) N/A, not applicable; 1Culture of swabs only performed for lytA positive or equivocal samples; 2Microarray results: Spn = S. pneumoniae detected, Non-Spn = Only non-S. pneumoniae bacterial species detected, Negative = No bacterial species detected; 3Samples excluded from vaccine type and non-vaccine type analyses.
2.4.2.2 Comparison of NP and OP samples There were concerns about the validity of the lytA qPCR assay when applied to the OP samples. As such, for the comparison of NP and OP samples in section 5.2.1, we altered the algorithm in Table 2.2 to consider samples positive for S. pneumoniae only where the presence of S. pneumoniae could be confirmed either by isolation of S. pneumoniae at the culture step or detection using microarray testing.
2.4.2.3 Identification of OP isolates as S. pneumoniae For the 91 isolates derived from 11 OP samples (see 2.3.3), identification of an isolate as S. pneumoniae was based on an isolate being optochin sensitive (see 2.7.2), bile soluble (see 2.4.2), lytA real-time PCR positive (2.8) and identified by MALDI-TOF MS as S. pneumoniae (see 0). A result of ‘not readable’ using the bile solubility test, ‘equivocal’ using lytA real-time PCR or ‘no identification’ using MALDI-TOF MS were all considered to be negative for S. pneumoniae. If one of four tests were inconsistent with the typical S. pneumoniae results, an isolate was still considered to be S. pneumoniae.
28 Where two of the four tests were indicative of S. pneumoniae, but the other two tests were not, the StrepID component of microarray was used to discriminate between S. pneumoniae and non-S. pneumoniae. Where only one or none of the tests were indicative of S. pneumoniae, an isolate was considered non-S. pneumoniae.
2.4.1 Optochin sensitivity
Optochin sensitivity was determined using an optochin disc (5 µg; 6 mm; Oxoid brand, Thermo Fisher Scientific, Australia) on the HBA plate. Sensitivity to optochin was defined as a zone of inhibition greater than 14 mm (including 6 mm disc). Isolates with a zone of inhibition less than or equal 14 mm were considered non-susceptible. In FFS, non-susceptible isolates were tested further with bile solubility and the Phadebact® Pneumococcus test (Boule Diagnostics AB, Huddinge, Sweden). These two additional tests were also done for optochin susceptible isolates which were later found to be non- typeable (during serotyping procedure).
2.4.2 Bile solubility
In FFS, bile solubility was determined by emulsifying the bacterial culture of an isolate grown on HBA plates in 2 ml of sterile saline to give a turbid suspension of approximately 3 McFarland. The suspension was divided between two glass test tubes; two drops of 10% (w/v) sodium deoxycholate were added to the first tube and two drops of distilled water were added as a control to the second tube. Tubes were vortexed and incubated at room temperature for 10-15 min. Solubility was defined as some degree of clearing of turbidity in the tube containing deoxycholate compared to the control. Where the isolate formed aggregates and was unable to be emulsified, the test was deemed inconclusive.
For the OP isolates in chapter 5, bile solubility testing was performed using a variation of the method described above and was in line with current WHO recommendations (31, 232). Bacterial growth was emulsified in 1 ml of 0.85% (w/v) saline to 0.5 - 1 McFarland standard. The bacterial suspension was divided into two tubes, with 0.5 ml of 2% (w/v) deoxycholate added to the ‘test’ tube and 0.5 ml of 0.85% (w/v) saline added to the
‘control’ tube. Both tubes were incubated at 37°C, 5% CO2 for 10 min and checked for any clearing of turbidity. If no clearing was observed after 10 min, tubes were incubated for an additional 2 h before turbidity was observed again. As with the previous method
29 used, where there was any clearing of turbidity in the ‘test’ tube, the isolate was deemed bile soluble and where the isolate was unable to be emulsified, the test was deemed ‘not readable’.
2.4.3 Phadebact pneumococcus test
The Phadebact pneumococcus test was performed according to the manufacturer’s instructions. In brief, a drop of the Pneumococcal Control reagent was added to the ‘control’ circle on the slide provided in the kit and a drop of the Pneumococcal Reagent added to the ‘test’ circle on the slide. Colonies were added separately to both the control and test reagents on the slide using fresh disposable loops. The slide was then rocked gently, and the result read after 1 min. The test was defined as positive when the agglutination in the test reagent was stronger than the control. Where aggregation or agglutination was equally strong in both the control and test reagents, the test was deemed inconclusive. All non-typeable presumptive S. pneumoniae and presumptive non-S. pneumoniae in FFS were further tested by lytA real-time PCR (see 2.8) to confirm identification.
2.4.4 MALD-TOF MS
All OP isolates in chapter 5 (n=91) were identified by MALDI-TOF MS using the Vitek® MS (bioMérieux) system. Following an overnight culture of the isolates on HBA (37°C,
5% CO2), approximately four colonies (a visible amount of bacterial growth) were applied to the Vitek MS-DS target slide (bioMérieux) using a sterile toothpick. Immediately after application to the slide, 1 μl of Vitek MS-CHCA matrix solution (bioMérieux) was added to the bacteria on the slide. After allowing the matrix solution to air-dry, the slide was loaded in to the Vitek MS machine. Following run completion, results were analysed using the MYLA® software (version 4, bioMérieux).
2.5 Serotyping
2.5.1 Latex agglutination serotyping
In FFS, S. pneumoniae were serotyped using latex agglutination with commercial S. pneumoniae reagents (Denka-Seiken Co., Ltd., Japan) and reagents produced in-house (233). Serotyping was performed systematically (234), starting with polyvalent reagents, then testing appropriate monovalent and factor reagents. For each reaction 15 μl of either
30 Denka reagents or in-house produced reagents were added to a clean glass slide. Using a 1 μl disposable loop a small but visible scraping of pneumococcal growth from an overnight culture on a HBA plate was added to the serotyping reagent and briefly emulsified. The glass slide was rocked back and forth for 1 min before examining the slide for the level of agglutination and clearing. Only reactions with strong agglutination and strong clearing were considered positive.
2.5.2 Latex ‘sweep’ agglutination serotyping
The subset of OP samples (n=11) that underwent further investigation in chapter 5, as well as the OP isolates (n=91) generated in that chapter, were serotyped using latex sweep agglutination. Latex sweep agglutination serotyping was performed using the method described in Turner et al. (235) using reagents produced in-house (233) with Statens Serum Institute (SSI) antisera. Using a 10 μl disposable loop, all colonies from an overnight culture (on either gHBA or HBA plates, at 37°C with 5% CO2) were emulsified in 0.5 ml of sterile saline. The solution was diluted, if necessary, to a turbidity of 4-5 McFarland. 10 μl of test suspension was added to 10 μl of latex reagent (either pool, group/type or factor) on a clean glass slide. The slide was rotated for 2 min before reading the reaction. For each isolate, all pools were tested, and if positive by any pool the corresponding group/type or factor was tested. A positive reaction was defined as any degree of agglutination, with or without clearing. Suspensions which did not emulsify in saline were deemed ‘not readable’.
2.5.3 Quellung serotyping
Where the serotype was unable to be resolved with latex agglutination in FFS or an OP isolate in chapter 5 had a positive serotyping result using the latex sweep agglutination method, isolates were serotyped with the Quellung reaction using antisera from the Statens Serum Institute (Copenhagen, Denmark) (236). Using a 1 μl disposable loop, a small sweep of pneumococcal growth from an overnight culture on a HBA plate was inoculated into 100 μl of Heart Infusion broth and emulsified. Using a 1 μl disposable loop a drop of the inoculum was added to a microscope slide. An equal volume of antiserum then was added using a fresh 1 μl disposable loop and gently mixed. A small circular coverslip was placed on the suspension and the reaction was observed under
31 phase-contrast at 40X magnification. A positive reaction was defined as an appearance of swollen or enlarged cells compared to a negative control (where no antiserum was added).
FFS isolates that were non-typeable by Quellung (negative for all pools and the omni serum) were tested further for presence of lytA gene by qPCR (237) and lytA positive isolates were tested by multilocus sequence typing.
2.5.4 Microarray serotyping
Serotyping of NVEP samples and isolates in chapter 5 was done using the BUGS Bioscience pneumococcal molecular serotyping platform (154) using the Senti-SPv1.5 S. pneumoniae molecular serotyping microarray (BUGS Bioscience, UK). Microarray serotyping was performed using DNA from the plate culture. DNA was normalised to a concentration of 22.7 ng/μl in a final volume of 20 μl, before being heat fragmented at 95°C for 60 min. The fragmented DNA was incubated at 4°C for 3 min before 8.8 μl (~200 ng) of DNA was transferred to a fresh tube and 8 μl of either a Cy3 or Cy5 labelling dye was added (containing either ULS-Cy3 or ULS-Cy5 with 10X Labelling solution provided in the Genomic DNA High-throughput ULS Labelling Kit; Agilent Technologies). The fragmented DNA with labelling dye was incubated at 85°C for 30 min. The labelled DNA was then purified by adding 10 μl of each sample (combining Cy3 and Cy5 labelled DNA in appropriate wells) to 96-well Agilent-KREApure™ purification plate (Agilent Technologies) and centrifuging at 3,000 x g for 3 min. 12 μl of purified labelled DNA was then added to 32 μl of hybridisation solution (containing 22 μl of Agilent 2X Hybridization Buffer and 10 μl Agilent-CGHblock, both provided in Oligo aCGH/ChIP-on-chip Hybridization Kit; Agilent Technologies) and incubated at 70°C for 15 min. 40 μl of the hybridisation solution was dispensed onto a 8x Hybridization gasket slide (Agilent Technologies). The gasket slide and the microarray slide were then assembled in a hybridisation chamber. Slides were hybridised overnight (for a minimum of 16 h) in a Hybridisation Oven (Shel Lab) at 65°C, rotating at 20 rpm. Microarray slides were then washed in Oligo aCGH Wash Buffer 1 at room temperature for 5 min and in Oligo aCGH Wash Buffer 2 at 35°C for 1 min (both provided in the Oligo aCGH/ChIP-on-chip Wash Buffer Kit; Agilent Technologies). Slides were scanned in the SureScan Microarray Scanner G2600D instrument (Agilent Technologies) using
32 Agilent Microarray Scan Control software (version 9.1.7.1; Agilent Technologies) with a scan protocol developed by the Bacterial Microarray Group at St. George’s (BμG@S, London, UK). Feature extraction was done using Agilent Feature Extraction software (version 12.0.0.7; Agilent Technologies) and files were uploaded to the Senti-Net system (BUGS Bioscience) via the Senti-Net portal (http://senti.bugsbio.org) for analysis and curation of microarray results. Prior to establishment of microarray in our laboratory, nine NP samples and 22 OP samples collected in 2012 were sent to the Bacterial Microarray Group at St. George’s for analysis using an earlier version of the microarray (v1.4).
2.5.5 Multiplex PCR serotyping
The subset of OP samples (n=11) that were examined for complexity in chapter 5 as well as all 10 isolates from sample FVEP-002-460 were serotyped using multiplex PCR. Multiplex PCR was performed using the method described by Carvalho, et al. (33) and on Centers for Disease Control and Prevention (CDC), Streptococcus Laboratory PCR Deduction of Pneumococcal Serotypes website (https://www.cdc.gov/streplab/pcr.html, accessed 14th July 2015). DNA concentrations from swab extractions were measured using a NanoDrop 1000 spectrophotometer (Thermo Scientific) and DNA was diluted to a concentration of 5 ng/μl. All 8 reactions plus the 6C reaction were performed for each sample, with each 25 μl reaction containing 1X Qiagen Multiplex PCR Master Mix (Qiagen) and 5 μl of template DNA. Primer concentrations varied within each reaction as described in the multiplex PCR protocol. PCRs were performed with an initial denaturation step at 95°C for 15 min, followed by 35 cycles of denaturation at 94°C for 30 s, annealing at 54°C for 90 s and elongation at 72°C for 60 s. A final elongation step at 72°C for 10 min was also included.
PCR products were sized and analysed using the Agilent DNA 1000 kit (Agilent Technologies), Agilent 2100 Bioanalyzer instrument (Agilent Technologies) and Agilent 2100 Expert software (version B.02.08.SI648 (SR3), Agilent Technologies). The DNA chip in the Agilent DNA kit was prepared according to manufacturer’s instructions with 1 μl of PCR product added to sample wells containing 5 μl of marker and 1 μl of DNA ladder added to the appropriate well. Only peaks over 20 fluorescence units were considered PCR products.
33 PCR products were also run on an E-Gel® 48 2% Agarose Gel (Invitrogen) using the Mother E-Base™ Device (Invitrogen). PCR products were diluted 1:3 in nuclease-free water and 15 μl was loaded into the E-Gel. 15 μl of the E-Gel® 1 kb Plus DNA Ladder (Invitrogen) and 15 μl of the E-Gel® 50 bp DNA Ladder (Invitrogen) were also loaded into the E-Gel. The E-Gel was inserted into the Mother E-Base, the EG program was selected and the gel was run for 10 min. Gels were visualised using the G:Box gel doc system (Syngene) with GeneSys software (Syngene). PCR products were sized and analysed using GelAnalyzer software (version 2010a, GelAnalyzer.com).
Results from both E-Gel and Bioanalyser methods were compared and only products detected by both methods were considered true products. Product sizes determined using the Bioanalyser method were then compared with expected product sizes for serotypes in the PCR reaction. Products that were within 30 bp of an expected serotype product size were considered positive for the serotype.
2.5.6 Non-typeable pneumococci and pneumococci from unencapsulated lineages
In this thesis we use two terms to define pneumococci without identifiable polysaccharide capsule serotypes. For phenotypic serotyping methods such as latex/latex ‘sweep’ agglutination and multiplex PCR the term ‘non-typeable’ has been used as such methods lack the ability to distinguish between true unencapsulated lineages and other factors such as downregulated capsule or methodological limitations. For microarray serotyping we use the more precise term ‘unencapsulated lineages’ as this method is able to distinguish between serotypeable pneumococci, pneumococci from unencapsulated lineages and pneumococci with serotypes that may be divergent or incomplete.
2.6 DNA extractions
2.6.1 DNA extractions from swabs using spin columns
In FFS, whole genomic DNA was extracted from a 100 µl aliquot of the NP swab storage media (STGG). Aliquots were pelleted at 7380 x g for 10 min. Pellets were then resuspended in a fresh enzymatic lysis buffer (final concentrations of 36.25 mM Phosphate buffer [pH 6.7], 1 mg/ml lysozyme, 0.075 mg/ml mutanolysin, 2 mg/ml
34 proteinase K) and incubated at 56°C for 45 min. 10 µl of 20% v/v SDS was then added to complete cell lysis and extraction was continued using the QIAamp DNA Mini Kit (Qiagen); following the ‘DNA purification from tissues’ protocol including the proteinase K and RNase A steps. Two elution steps of 50 µl Buffer AE were performed in the one tube; the extracted DNA was then divided into two tubes and stored at -30°C.
2.6.2 DNA extractions from swabs using MagNA Pure LC instrument
For NVEP swabs (chapter 5), DNA extractions were performed on 100 μl of STGG storage medium using the MagNA Pure LC instrument (Roche) and the MagNA Pure LC DNA Isolation Kit III (Bacteria, Fungi) (Roche). After pelleting the 100 µl sample aliquot at 7380 x g for 10 min, the pellet was resuspended in 130 μl of Bacterial Lysis Buffer (provided in the MagNA Pure LC DNA Isolation Kit III (Bacteria, Fungi)) and 5 μl of enzymatic cocktail II (containing 5 mg/ml of lysostaphin and 100 mg/ml lysozyme) as described in the Kit instructions. Samples were then incubated at 37 °C for 30 min. 4 μl of RNase A (Qiagen) was added to each sample and mixed gently at room temperature for 2 min. 125 μl of sample lysate was then loaded into the sample cartridge for running on the MagNA Pure LC. The sample cartridges were loaded into the MagNA Pure LC and DNA extractions were performed following the MagNA Pure LC DNA Isolation Kit III (Bacteria, Fungi) protocol. The MagNA Pure LC instrument was configured to add a Lysis/Binding Buffer to complete cell lysis and allow the binding of DNA to the silica surface of Magnetic Glass Particles. The DNA bound to the Magnetic Glass Particles was separated magnetically from the Lysis/Binding Buffer and the bound DNA was washed in subsequent steps using three different Wash Buffers (Wash Buffers I-III). The purified DNA bound to the Magnetic Glass Particles was eluted in 100 μl of Elution Buffer. Extracted DNA was transferred to sterile labelled microfuge tubes and stored at -30°C.
2.6.3 DNA extraction for microbiome analysis
For extracting genomic DNA from swab samples for microbiome analysis, extractions were performed in batches of 23 samples plus a negative control (empty sterile 1.5 ml microfuge tube) using the QIamp DNA Mini Kit (Qiagen). In a biohazard class II cabinet, 200 μl of swab storage media (STGG) was aliquoted into a sterile 1.5 ml microfuge tube
35 for each sample. These aliquots were then spun in a centrifuge for 5 min at 6,000 x g to create a pellet. The supernatant was then discarded, and the pellets were placed at – 80°C for 15 min to promote cell lysis. Following the freeze/thaw step, 200 μl of enzymatic lysis buffer (containing final concentrations of 36.25 mM Phosphate Buffer, 1 mg/ml lysozyme, 0.075 mg/ml mutanolysin and 2 mg/ml Proteinase K) was added, before incubating at 56°C for 45 min. An additional 20 μl of Proteinase K (20 mg/ml, Qiagen) was added before further incubation at 56°C for 10 min. 4 μl of RNase A (at 100 μg/ml, Qiagen) was added and mixed gently at room temperature for 2 min before adding 200 μl of Buffer AL. Samples were then incubated at 70°C for a further 10 min to complete cell lysis. 200 μl of absolute ethanol was added before applying the lysate to a QIAamp mini spin column (Qiagen). The spin column was spun in a centrifuge at 6,000 x g for 1 min before placing into a clean collection tube, discarding the old collection tube and supernatant. 500 μl of wash Buffer AW1 was added to the spin column before centrifugation at 6,000 x g for 1 min. The spin column was placed into a clean collection tube (and supernatant discarded) before adding 500 μl of Buffer AW2. The spin column was spun in a centrifuge at 20,000 x g for 3 min before placing the spin column into a clean collection tube and centrifuging at 20,000 x g for an additional min. The spin column was then placed into a sterile 1.5 ml microfuge tube and 100 μl Buffer AE was incubated on the spin column membrane at room temperature for 5 min to elute the DNA before a final centrifugation step at 6,000 x g for 1 min. Extracted DNA was stored at - 30°C. A reagent-only extraction control was included for each batch of extractions performed.
2.6.4 Extractions from pure culture
For DNA extractions from pure cultures in FFS, bacteria were incubated (37°C, 5 % CO2) overnight on HBA plates or on chocolate agar plates (Oxoid brand, Thermo Fisher Scientific, Australia) for H. influenzae, and were then inoculated into 1 ml Todd-Hewitt Broth (THB). Extractions were performed using DNeasy Blood and Tissue Kit (Qiagen) including steps for Gram-positive bacteria, specifically the inclusion of 20 µl of 200 mg/ml lysozyme with the enzymatic lysis buffer (described in DNeasy Blood and Tissue Handbook). Two elution steps of 100 µl Buffer AE were performed and the resultant 200 μl of extracted DNA was stored at -30°C.
36 2.6.5 DNA extraction from plate culture
Pellets from the plate culture in NVEP (2.3.2 NVEP culture) were resuspended in 180 μl of a freshly prepared lysis buffer (containing final concentrations of 20 mM Tris-HCl, 2 mM Na-EDTA, 1% v/v Triton X-100, 2 mg/ml RNase A (Qiagen) and 20 mg/ml lysozyme (Sigma-Aldrich). Samples were then incubated at 37°C for 60 min before adding 20 μl of 20 mg/ml Proteinase K (Qiagen) and 200 μl of Buffer AL (Qiagen) and incubating at 56°C for a further 30 min.
DNA extraction was generally performed using QIAcube method on the QIAcube HT instrument using the QIAamp 96 DNA QIAcube HT kit and QIAcube HT plasticware. In brief, 400 μl of the sample lysate was loaded onto the S-Block which was then loaded into the instrument. The QIAcube was configured to have an initial step where 200 μl of 100% v/v ethanol was added to each sample and incubated at room temperature for 5 min. 600 μl of sample lysate was loaded into the QIAmp 96 plate and a vacuum step at 35 kPa for 5 min was applied. For the first wash step, 500 μl of Buffer AW1 was added to each sample followed by a 2 min vacuum step at 35 kPa. For the second wash step, 500 μl of 80% v/v ethanol was added to each sample, followed by a 1 min vacuum step at 35 kPa. Prior to the elution step, two vacuum steps (at 55 kPa for 1 min and 35 kPa for 2 min) were used to dry the QIAamp 96 plate. DNA for each sample was eluted in 100 μl of nuclease-free H2O (Ambion) using a vacuum step at 70 kPa for 8 min.
In cases where samples were likely to be problematic, such as samples with high growth of other bacterial species on gHBA plates, the spin column method was performed using the QIAamp DNA Mini Kit (Qiagen). 200 μl of 100% v/v ethanol was added to the sample lysate and the entire volume was transferred to the spin column in the QIAamp DNA Mini Kit. The spin column was spun at 5,000 x g for 1 min before 500 μl of Buffer AW1 was added and spun again at 5,000 x g for 1 min. 500 μl freshly prepared 80% v/v ethanol was then added to the spin column and spun at 18,000 x g for 3 min, followed by an additional centrifuge step at 18,000 x g for 1 min. DNA was eluted in 50 μl of nuclease-free H2O (Ambion) using a centrifugation step at 18,000 x g for 1 min.
For both methods DNA concentrations were measured using the NanoDrop 2000 spectrophotometer (Thermo Scientific) and DNA was stored at -30°C.
37 2.6.6 DNA extraction of OP isolates
Initially, whole genomic DNA was extracted from all OP isolates generated in chapter 5 using the QiaCube HT instrument (Qiagen) following the method described above (2.6.5). However, for many of the isolates (n=52) there were issues during the extraction process resulting in poor quality DNA. For these isolates, a modified version of the spin column method described in 2.6.1 was used. Prior to the addition of the enzymatic lysis buffer, samples underwent a freeze-thaw step at -80°C for 15 min, and DNA was eluted in 100
μl of nuclease-free H2O (Ambion), with two elution steps of 50 μl each.
2.7 Quantitative real-time PCR (qPCR)
2.7.1 Duplex assays for S. pneumoniae, H. influenzae, S. aureus and M. catarrhalis
In chapter 3 and 4, a quantitative real-time PCR (qPCR) assay was used to detect, and determine the carriage densities, of the respiratory pathogens S. pneumoniae, Staphylococcus aureus, Moraxella catarrhalis and Haemophilus influenzae (176). Two duplex assays were performed using primers (Sigma-Aldrich), probes (Eurogentec) and concentrations shown in the supplemental table S4 (Table 3.5 of this thesis) of Boelsen et al. (231). To conduct the assays 2 µl of DNA was used, together with the Brilliant III Ultra-Fast qPCR Master Mix (Agilent Technologies).
Standard curves for each species were prepared using a dilution series of DNA extracted from S. pneumoniae ATCC 6305, H. influenzae F412, M. catarrhalis ATCC 8176 and S. aureus ATCC 29213. Duplex qPCRs were performed on a Stratagene Mx3005 instrument with 40 cycles of 95°C for 20 s and 60°C for 20 s after an initial activation of 95°C for 3 min. Results were analysed using MxPro™ software (Stratagene) with genomic equivalents per ml were determined for each sample.
2.7.2 lytA qPCR
In chapter 5, following DNA extraction from swabs using the MagNA Pure LC instrument, samples were screened for the presence of S. pneumoniae using qPCR
38 targeting the lytA gene (237). A standard curve was prepared using a dilution series of DNA extracted from S. pneumoniae ATCC 6305. qPCRs were performed using 96-well optical plates (Axygen) with 25 μl reactions containing 100 nM of both forward (5’- ACGCAATCTAGCAGATGAAGCA-3’; Sigma-Aldrich) and reverse (5’- TCGTGCGTTTTAATTCCAGCT -3’; Sigma-Aldrich) primers, 200 nM of probe (Cy5- TGCCGAAAACGCTTGATACAGGGAG -BHQ3; Eurogentec), 12.5 μl Brilliant III Ultra-Fast qPCR Master Mix (Agilent Technologies) and 2 μl template DNA or standard curve DNA. Samples were run in duplicate on a Stratagene Mx3005 instrument with 40 cycles of 95°C for 20 s and 60°C for 20 s after an initial activation of 95°C for 3 min. Results were analysed using MxPro™ software (Stratagene) with density in genomic equivalents per ml determined for each sample. Samples were considered positive where the Ct<35, equivocal where the Ct was between 35 and 40, and negative where the Ct>40. Samples that were either positive or equivocal for the lytA gene were cultured for microarray analysis (2.3.2 NVEP culture).
2.8 Real-time PCR For non-typeable presumptive pneumococci (in the FFS) and non-pneumococci, the lytA qPCR above (see 2.7.2) was performed as a singleplex real-time PCR assay. The DNA concentrations for extracted isolates was measured using the NanoDrop 2000 (Thermo Scientific) and isolate DNA was diluted to a concentration of 1 ng/μl. As a positive control, only the 1 ng/μl concentration of the standard curve was used. Isolates with a cycle threshold (Ct) below 30 were considered positive. In FFS, isolates that were negative for lytA were deemed not to be S. pneumoniae and were excluded from further testing and analysis.
2.9 Alternative targets for S. pneumoniae Alternative targets to the lytA gene for S. pneumoniae identification were identified in the literature. Only targets which were reported to discriminate between S. pneumoniae and other streptococcal species (including S. pseudopneumoniae), and were suitable for qPCR screening of samples, were further considered. Genes piaA, piaB and ulaA were selected for further testing along with the four additional genes used for S. pneumoniae identification in the StrepID component of microarray analysis (Table 5.11).
39 Primers (and potential probes) were designed for vanZ, SP_0137, bguR and fucK for use in qPCR (Table 2.3). For each gene, a BLAST search was done for S. pneumoniae (taxid:1313) to find conserved regions for primer design. Primer and probes were designed using the PrimerQuest online tool (Integrated DNA Technologies, http://eu.idtdna.com/PrimerQuest/Home/Index). Primer and probe sequences were checked against other streptococcal species using BLAST.
Table 2.3. Primer sequences for vanZ, SP_0137, bguR and fucK. Gene Product Primer Sequence (5’->3’) (TIGR4 ID) size vanZ Forward GAATAGTCGGGACAGGTTTCT 167 bp (SP_0049) Reverse TCTAGTCTGTGATTTGAACACTCT Forward AGGCGCTATTAGCTTCTTTCTC (SP_0137) 153 bp Reverse CCTAATTCACCTAGCGCTGTAA bguR Forward AGTTTGCCTGTAGTCGAATGA 115 bp (SP_2020) Reverse TTTGAGCTGCCACGAGAG fucK Forward CGCGATTGTCTCAGTACCTAA 201 bp (SP_2167) Reverse CGTCTTAGTTCCTCTATGATCCAC
2.9.1 Reference isolates
The seven gene targets were initially tested against a set of streptococcal reference isolates (Table 2.4). DNA was extracted from the streptococcal reference isolates using spin column method described above (2.6.6). The DNA from the reference set of isolates was then normalised to a concentration of 1 ng/μl for use in PCRs for gene targets.
Table 2.4. Streptococcal reference isolates used to screen potential pneumococcal targets. Species Strain Serotype Origin S. pneumoniae PMP6 (ATCC 6305) 5 Germany S. pneumoniae PMP7 (ATCC 49619) 19F USA S. pneumoniae PMP85 23F Fiji
40 Species Strain Serotype Origin S. pneumoniae PMP849 1 USA S. pneumoniae PMP1064 22A Australia S. pneumoniae 0173-01 NT Fiji S. pseudopneumoniae PMP1300 N/A USA S. pseudopneumoniae 001-009-01 N/A Fiji S. mitis PMP16 N/A Australia S. mitis PMP933 N/A Unknown S. mitis PMP934 N/A Unknown S. salivarius PMP1010 N/A New Zealand S. mutans PMP935 (NCTC 10449) N/A UK S. oralis PMP1056 N/A New Zealand S. infantis PMP1301 N/A USA S. pyogenes PMP1000 N/A Australia S. gordonii PMP1057 N/A New Zealand S. bovis PMP1035 (NCTC 8177) N/A Unknown S. dysgalactiae PMP1051 N/A Unknown S. agalactiae PMP994 N/A Australia S. sobrinus PMP1053 N/A New Zealand S. sanguinis PMP936 (NCTC 7864) N/A Unknown S. anginosus PMP1049 N/A Unknown S. vestibularis PMP1059 N/A New Zealand S. peroris PMP1302 N/A USA S. australis PMP1303 N/A USA S. oligofermantans PMP1305 N/A USA S. cristatus PMP1306 N/A USA S. sinensis PMP1307 N/A USA
2.9.2 PCRs
For each gene (except piaB which was used as a qPCR/real-time PCR target only), each 25 μl PCR reaction contained final concentrations of 1X PCR Gold Buffer (Life
Technologies), 1.5 mM MgCl2 (Life Technologies), 1 μM of each the forward and reverse
41 primers, 1.25 U of AmpliTaq Gold DNA Polymerase (Life Technologies) and 200 μM of each dNTP (Qiagen), with 5 μl of template DNA. Each PCR was run under the following conditions: initial denaturation at 95°C for 10 min; 30 cycles of denaturation at 95°C for 15 s, annealing at 56°C for 30 s and elongation at 72°C for 20 s; and a final elongation step at 72°C for 5 min. PCR products were visualised by running 5 μl of each product and a 100 bp ladder (Life Technologies) on a 1% w/v agarose gel with 1X GelRed (Biotium) at 110 V for 25 min.
2.9.3 bguR and piaB qPCR and real-time PCR
For bguR, a probe sequence (5’-FAM-AAACGTGGGCAGGGAACCTTTGTT-BHQ1- 3’; Sigma-Aldrich) was added for use as a qPCR assay. bguR qPCR and real-time PCR conditions were optimised to give forward primer, reverse primer and probe concentrations of 300 nM, 100 nM and 200 nM, respectively. The piaB qPCR and real- time PCR assays were performed using the primers (Sigma-Aldrich), probe (Sigma- Aldrich) and concentrations described in Trzciński et al. (41). For both the piaB and bguR assays 2 µl of template DNA was used, together with the Brilliant III Ultra-Fast qPCR Master Mix (Agilent Technologies). The piaB and bguR assays were also run on the Stratagene Mx3005 instrument under the same conditions as the lytA qPCR (40 cycles of 95°C for 20 s and 60°C for 20 s after an initial activation of 95°C for 3 min). Results for all qPCR and real-time PCR assays were analysed using MxPro™ software (Stratagene) Samples were considered positive where the Ct<35, equivocal where the Ct was between 35 and 40, and negative where the Ct>40 and isolates were considered positive where the Ct value less than 30.
2.10 Serotype 14-specific PCR Four non-typeable pneumococci had STs related to serotype 14 in FFS. As serotype 14 is a vaccine type we wanted to examine whether these isolates potentially possessed serotype 14 capsule genes. We performed a serotype 14-specific PCR using the following primers in Dias et al. (238): forward 5’- GAAATGTTACTTGGCGCAGGTGTCAGAATT-3’ and reverse 5’-GCCAATACTTCTTAGTCTCTCAGATGAAT-3’. Each 25 μl reaction contained final concentrations of 1X PCR Gold Buffer (Life Technologies), 3.5 mM
42 MgCl2 (Life Technologies), 1 μM of each the forward and reverse primers, 1.25 U of AmpliTaq Gold DNA Polymerase (Life Technologies) and 200 μM of each dNTP (Qiagen), with 5 μl of template DNA. Three serotype 14 isolates from our culture collection (PMP130, PMP131 and PMP1114) were included as positive controls, with the PCR conditions of an initial denaturation at 95°C for 10 min, followed by 35 cycles of denaturation at 94°C for 15 s, annealing at 54°C for 30 s and elongation at 72°C for 60 s. This was followed by a final elongation step at 72°C for 10 min. PCR products were visualised by running 5 μl of each product and a 100 bp ladder (Life Technologies) on a 1% w/v agarose gel with 1X GelRed (Biotium) at 110 V for 25 min. Isolates with a band consistent with a 189 bp product were considered positive for serotype 14.
2.11 16S rRNA gene PCR The V4 region of the 16S rRNA gene was amplified by PCR using the primer pair which included the Illumina-specific adapter sequences (Table 2.5).
Table 2.5. 16S rRNA PCR primers for V4 region plus illumina adapter Product Region Primer Sequence (5’>3’) size (bp)* TCGTCGGCAGCGTCAGATGTGTATAAGAGACAG 515F GTGCCAGCMGCCGCGGTAA V4 ~250 GTCTCGTGGGCTCGGAGATGTGTATAAGAGACAG 806R GGACTACHVGGGTWTCTAAT *Exact product size varies between bacterial species. Illumina adapter sequence is underlined.
The PCR was performed using 1 μM of each primer (HPLC purified in liquid at 100 μM concentration, Sigma-Aldrich, Australia), 1X Phusion Green Hot Start II High-Fidelity PCR Master Mix (Thermo Fisher Scientific) and 10 μl of extracted gDNA in a 50 μl reaction. Following an initial denaturation step at 98°C for 30 s, there were 35 cycles of denaturation at 98°C for 5 s, annealing at 65°C for 20 s and elongation at 72°C for 15 s, followed by a final elongation step at 72°C for 5 min. PCR products were then stored at -20°C. Each PCR run included a no template control.
43 2.12 PCR purification PCR products from the V4 PCR were purified using AMPure XP beads and following the Illumina ‘16S Metagenomic Sequencing Library Preparation’ protocol (Part # 15044223 Rev. B). In brief, PCR reaction tubes underwent centrifugation at 1,000 x g for 1 min to collect liquid. AMPure XP beads (at room temperature) were vortexed for 30 s to evenly disperse the beads. 20 μl of AMPure XP beads were added to each tube and the entire volume in each tube was then gently pipetted up and down 10 times, before incubating at room temperature for 5 min. Tubes were then placed on a magnetic stand for 2 min ensuring that the supernatant had cleared. While tubes were still on the magnetic stand, the supernatant was removed and discarded. The beads were washed by adding 200 μl of freshly prepared 80% v/v ethanol to each tube, incubating the plate for 30 s and discarding the supernatant. A second wash step was performed, repeating the previous wash steps. Excess ethanol was removed, and beads were then air-dried for 10 min. After removing the PCR tubes from the magnetic stand, 52.5 μl of 10 mM Tris (pH 8.5) was added to each tube, the mix was gently pipetted up and down 10 times before incubating at room temperature for 2 min. Tubes were then placed back on the magnetic stand for 2 min, before 50 μl of the supernatant was transferred to a new 96-well PCR plate or PCR tubes (as appropriate).
PCR product sizes and concentrations were measured by running samples on a TapeStation System (Agilent Technologies) using the D1000 ScreenTape System (Agilent Technologies). After allowing reagents to equilibrate at room temperature for 30 min, 3 μl of D1000 Sample Buffer was added to 1 μl of DNA (or D1000 Ladder). PCR products were then loaded into the TapeStation and were analysed using TapeStation software.
2.13 Multilocus sequence typing (MLST) For chapter 3, non-typeable S. pneumoniae were examined by multilocus sequence typing (MLST) using matrix-assisted laser desorption/ionisation – time of flight mass spectrometry (MALDI-TOF MS) (239). Where MALDI-TOF MS was unable to be performed (see below), MLST as conducted by sequencing of PCR products. The allelic profiles of all isolates examined by MLST were submitted to S. pneumoniae MLST
44 database (http://pubmlst.org/spneumoniae/) sited at the University of Oxford (240). Sequence types (STs) were determined based on allelic profiles, with new sequence types assigned for those isolates that did not match any existing ST.
2.13.1 PCR
Extracted genomic DNA from pure culture was used in the PCR amplification of the seven housekeeping genes: aroE, gdh, gki, recP (recA), spi and ddl. Primer pairs for each gene are described in Dunne et al. (239) and are shown in Table 2.6, including the promoter sequence tags T7 (for forward transcription) and SP6 (for reverse transcription) for in vitro transcription and later RNA cleavage. For each gene except spi, PCRs were performed in 25 μl reactions with final concentrations of 0.3 μM for both the forward and reverse primers (Sigma-Aldrich, Sydney, Australia), 1 X PCR Buffer (Qiagen), 2.0 mM
MgCl2, 1.25 U of HotStarTaq DNA Polymerase (Qiagen) and 200 μM of each dNTP (Qiagen) with 3 μl of genomic template DNA (at approximately 1 ng/μl DNA concentration). For spi, reactions were performed as above with the following exceptions: 4 μl of template DNA was used, and the concentration of both the forward and reverse primers was increased to 0.4 μM. After an initial activation step at 95°C for 15 min, there were 35 cycles of denaturation at 94°C for 30 s, annealing at 62°C for 45 s and elongation at 72°C for 1 min with a final elongation step at 72°C for 10 min. For spi, the annealing temperature was lowered to 54°C with all other PCR conditions unchanged.
45 Table 2.6. Primers for MALDI-TOF MS MLST of S. pneumoniae, from Dunne et al. (239) Primer Sequence (5’-3’)a Source aroE_T7F CAGTAATACGACTCACTATAGGGAGAAGGCTTCCTATTAAGCATTCTATTTCTCCCTTC CDC aroE_SP6R CGATTTAGGTGACACTATAGAAGAGAGGCTACAGGAGAGGATTGGCCATCCATGCCCACACTG CDC gdh_T7F CAGTAATACGACTCACTATAGGGAGAAGGCTATGGACAAACCAGCNAGYTT mlst.net gdh_SP6R CGATTTAGGTGACACTATAGAAGAGAGGCTGCTTGAGGTCCCATRCTNCC mlst.net gki_T7F CAGTAATACGACTCACTATAGGGAGAAGGCTGGCATTGGAATGGGATCACC mlst.net gki_SP6R CGATTTAGGTGACACTATAGAAGAGAGGCTTCTCCCGCAGCTGACAC mlst.net recA_T7F CAGTAATACGACTCACTATAGGGAGAAGGCTGAATGTGTGATTCAATAATCACCTCAAATAGAAGG CDC recA_SP6R CGATTTAGGTGACACTATAGAAGAGAGGCTTGCTGTTTCGATAGCAGCATGGATGGCTTCC CDC spi_T7F CAGTAATACGACTCACTATAGGGAGAAGGCTCGCTTAGAAAGGTAAGTTATGAATTT CDC spi_SP6R CGATTTAGGTGACACTATAGAAGAGAGGCTGTGATTGGCCAGAAGCGGAA mlst.net xpt_T7F CAGTAATACGACTCACTATAGGGAGAAGGCTTTAACTTTTAGACTTTAGGAGGTCTTATG CDC xpt_SP6R CGATTTAGGTGACACTATAGAAGAGAGGCTCGGCTGCTTGCGAGTGTTTTTCTTGAG CDC ddl_T7F CAGTAATACGACTCACTATAGGGAGAAGGCTTAAAATCACGACTAAGCGTGTTCTGG CDC ddl_SP6R CGATTTAGGTGACACTATAGAAGAGAGGCTAAGTAGTGGGTAATAGACCACTGGG CDC a The underlined portion of the primer indicates the tagged T7 or SP6 promoter region
46 2.13.2 In vitro transcription, base-specific RNA cleavage and MALDI-TOF MS
To determine the MLST sequence type (ST) we performed MALDI-TOF MS as described in Dunne et al. (239). To digest the T7 and SP6 promoter sequences, 4 μl of 0.15 U/μl shrimp alkaline phosophatase was added to 10 μl of PCR amplicon. This was then incubated at 37°C for 20 min followed by 85°C for 10 min. Four different transcription/cleavage mixes using the massCLEAVE T7/SP6 kit (Sequenom, Australia) were then prepared for each sample: T-Forward with T7 enzyme (TF), T-Reverse with SP6 enzyme (TR), C-Forward with T7 enzyme (CF), and C-Reverse with SP6 enzyme (CR).
Each reaction contained a final concentration of 1.8X polymerase buffer, 5.6 mM of dithiothreitol (DTT), 0.16 μg/μl RNase A, 4.4 U of either T7 or SP6 RNA polymerase, and 0.12 μl of T- or C- mixes from the massCLEAVE T7/SP6 kit. To 2.5 μl of this cleavage mix, 2 μl of the PCR/shrimp alkaline phosphatase mix was added in a 384 well plate. This plate was then incubated at 37°C for 3 h before an additional 21.5 μl of nuclease free water was added to each well. Clean resin was added to a dimple plate and allowed to dry for 10 min. The resin in the dimple plate was transferred to the 384 well plate, followed by rotation of the 384 well plate for 15 min. The plate was then centrifuged at 3,200 x g for 5 min. Each reaction was then immediately dispensed onto a spectroCHIP® using the MassARRAY Nanodispenser using a 4-point calibrant. The spectroCHIP® was then run on the MassARRAY analyser (Sequenom, Australia). The results were then analysed using the iSEQ software (Sequenom) using a library created using allelic data from the S. pneumoniae MLST database.
2.13.3 Sequencing of PCR products
For isolate 0082-01 we were unable to generate an allelic profile using MALDI-TOF MS. In this case, PCR products for each MLST housekeeping gene were sequenced. PCR products were purified using the QIAquick PCR Purification Kit (Qiagen) following the ‘QIAquick PCR Purification Kit using a Microcentrifuge’ protocol. To 20 μl of PCR product, 100 μl of Buffer PB was added (without the additional 10 μl of 3M sodium
47 acetate). Purified DNA was eluted in 30 μl of water before being sent for Sanger sequencing at the Australian Genome Research Facility (AGRF). Sequences were checked using FinchTV (version 1.4.0, Geospiza, Inc.) and used to determine the MLST allelic profile using the S. pneumoniae MLST database.
2.14 Sequencing and sequence processing
2.14.1 16S sequencing
MiSeq sequencing was performed by the Translational Genomics Unit at MCRI, Melbourne, Australia. Initially, a sequencing run of 22 samples (plus the extraction control and the no template PCR control) was performed, with the remaining samples and controls then sequenced in two batches. As the sequencing depth varied between the initial sequencing run and the two batches performed later, four of the initial 22 samples were also included in the later sequencing batches to check results did not differ by sequencing depth.
The V4 purified DNA was prepared for sequencing using the Illumina ‘16S Metagenomic Sequencing Library Preparation’ protocol (Document # 15044223 Rev. B) (see Figure 2.2). As illumina adapters were already attached in the 16S rRNA gene PCR, one PCR step was performed to add the indices and sequencing adapters. The libraries were then cleaned-up and normalised before pooling and sequencing using Illumina MiSeq V3 reagent kits (2 x 300 bp) on the MiSeq platform (Illumina).
Raw sequences were processed and classified using MOTHUR (version 1.35.1) (241) following the MiSeq SOP (http://www.mothur.org/wiki/MiSeq_SOP; accessed 27th July 2015) and silva nr (v119) classification database (242) (see Appendix A. 16S sequence processing and analysis scripts: 7.1.1 Sequence processing script using MOTHUR). Sequences with read lengths falling outside a 225-325 base pairs (bp) range, any ambiguous bases and >8 bp homopolymers were removed. Sequence were aligned across the region spanning the forward and reverse primers, and any overhang was trimmed. Chimeric sequences were identified and removed using UCHIME (243) and sequences were clustered into operational taxonomic units (OTUs) at 97% similarity to the genus level. OTUs for which >33% of sequence reads were from controls were removed to reduce background signal. Samples which had less than 50,000 sequence reads were
48 removed from further analysis, as were any samples that were greater than 20% similar to the controls in downstream analysis as these were considered to have amplified little, if any, bacterial DNA from the swab.
49 Figure 2.2. V4 sequencing approach. Adapted from Illumina ‘16S Metagenomic Sequencing Library Preparation’ protocol (Document # 15044223 Rev. B)
50 2.14.2 Whole genome sequencing
DNA from six isolates (FVEP-002-002-03, FVEP-002-078-01, FVEP-002-080-01, FVEP-002-080-02, FVEP-002-084-03 and FVEP-002-084-07) was sent to the Translational Genomics Unit at MCRI, Melbourne, Australia for library preparation and sequencing. Sequencing was done using Illumina MiSeq V2 reagent kits (2 x 150 bp) on the MiSeq platform (Illumina). Following sequencing, the sequence reads were down- sampled so that the assemblies would have approximately 150x coverage. Sequence assembly was performed using SPAdes version 3.5.0 (244). Assembled contigs were then submitted to the RAST server for annotation (245). Species identification was done using a combination of MetaPhlAN v2.0 (246) and Kraken (247), plus using 16S sequences from the isolates with 16S SILVA Incremental Aligner (SINA) version 1.2.11 (248), Greengenes BLAST (249) and NCBI BLAST (250). All six isolates were examined for the lytA gene (using the S. pneumoniae R6 gene as a reference; NCBI reference sequence: NC_003098.1) and for lytA primer and probes sequences using BLAST+ version 2.6.0 (251) and the -blastn function. For the primer and probe sequence BLAST, default parameters were adjusted to account for smaller sequences (-dust no -evalue 1000 - word_size 7). Any gene annotated as N-acetylmuramoyl-L-alanine amidase (EC 3.5.1.28) by RAST was also aligned against the lytA gene and the lytA primer and probes sequences using Serial Cloner version 2.6.1. Capsule genes were identified following RAST annotation and were aligned against 33F and 19B reference sequences (Genbank accession numbers AJ006986.1 and CR931676.1, respectively), sequences that aligned poorly were checked by NCBI BLAST. Figures for aligned sequences were generated using easyfig version 2.2.2 (252).
51 2.15 Data analysis
Analyses were conducted and graphs were generated using GraphPad Prism version 7.02 and R version 3.2.4. Unless otherwise stated p<0.05 was considered statistically significant. For all scripts used in MOTHUR and in R studio for α- and β-diversity analysis, heatmaps and correlation networks see Appendix A. 16S sequence processing and analysis scripts.
2.15.1 Comparing categorical data
For differences in categorical data, Fisher’s exact test was used, with the Chi-Square test for trend used for differences over time.
2.15.2 Comparing continuous data
Continuous data was checked for normality using the D'Agostino & Pearson normality test on GraphPad Prism version 7.02. For normally distributed data, an unpaired t-test was used. Where data was not normally distributed, for the Mann-Whitney test was used, with the Kruskal-Wallis test used for trends over time.
2.15.3 α- and β-diversity
Alpha-diversity (i.e. within sample diversity) was determined through OTU richness (number of OTUs) and Shannon diversity index (SDI). Alpha-diversity measures were calculated using R Studio (version 3.2.4) and the ‘diversity’ function of the vegan package (version 2.3-5) (253). Richness calculations were rarefied to account for differences in sequencing depth, as greater sequencing depth increases the number of OTUs or richness. Rarefaction was done by subsampling 50,000 sequence reads (based on the lowest number of reads – 53,616 reads) using the ‘rarefy’ function in vegan package. Beta- diversity (i.e. between sample diversity) was calculated using a Jaccard distance matrix (‘vegdist’ function, R vegan package) and permutational multivariate analysis of variance (PERMANOVA, ‘adonis’ function, R vegan package), as well using non-metric multidimensional scaling (nMDS) plots (‘metaMDS’ function, vegan package).
52 2.15.4 Relative abundance and prevalence
Relative abundance was defined as the percentage of total sequence reads from an individual that were from the specific OTU, genus or family being considered. Relative abundance plots were generated in GraphPad Prism (software version 6.07) and relative abundances of the most common families and genera between groups were compared using the Mann-Whitney test (as relative abundances were not normally distributed).
Prevalence was calculated as the total number of individuals with at least one sequence read from the specific OTU, genus or family being considered. In addition, prevalence was also calculated for the total number of individuals having greater than 1% relative abundance of the specific OTU, genus or family being considered.
2.15.5 Heatmap and clusters
Heatmaps were generated in RStudio using the following packages: vegan, RColorBrewer (version 1.1-2) and gplots (version 3.0.1) (254). OTUs were filtered to only those OTUs with a maximum relative abundance above 30%. Samples and OTUs were clustered using the ‘hclust’ function and the average linkage method based on a Jaccard distance matrix. Clusters were characterised using the ‘cascadeKM’ function (vegan package) with the default ‘calinski’ criterion (Calinski-Harabasz (255)) and coloured manually.
2.15.6 Correlation network
To examine co-presence or mutual-exclusion of the most abundant OTUs, correlation networks were generated using the default SPARCC (256) parameters in MOTHUR after filtering to the top 100 most abundant OTUs. The SPARCC data was then formatted and analysed in RStudio with reference to the following R script (https://colab.mpi- bremen.de/micro-b3/svn/EBI2014/Thursday/coocurrences/EBI_co-ocurrence.R, accessed 25-May-2016). Only correlations with a significance of p<0.05 and a correlation r>0.3 were included in networks. Networks were visualised and edited using Cytoscape software version 3.3.0 (257).
53 2.15.7 FiPP data and species-specific qPCR
The culture-based serotyping data from the original FiPP study (229) from the subset of participants in this study at 6, 9 and 12 months of age was used to calculate overall and vaccine-type carriage rates. For both the FiPP data and species-specific qPCR data, GraphPad Prism was used to generate plots and compare carriage rates (and densities) between groups.
2.15.8 Random forest models
The random forest analysis was done in RStudio using R packages ‘randomForest’ version 4.6-14, ‘pylr’ version 1.8.4 and ‘rfUtilities’ version 2.1-3. Rare OTUs were removed, and data was normalised and scaled. Each risk factor was analysed in separate models the ‘randomForest’ and ‘rf.significance’ functions. Each model used the paramters ‘ntree=501’, ‘importance=TRUE’ and ‘proximities=TRUE’, and significance was calculated using 500 permutations. The mean decrease in accuracy or variable importance was examined in each model for the top 20 OTUs.
2.15.9 Multivariate linear regression
Multivariate linear regression was done in RStudio using the ‘lm’ function (base ‘stats’ package in R, version 3.2.4) looking at the effects of vaccination and ethnicity on the relative abundance of the seven most common genera. For each model (or genus) an interaction term between vaccination and ethnicity was considered and Akaike information criterion (AIC) was used to determine whether inclusion of the interaction term was appropriate. The ‘step’ function (base ‘stats’ package) was used on the model containing all relevant risk factors for each genus to determine which risk factors should be included in each model. As well as vaccination status and ethnicity, final models included the following risk factors: symptoms of an upper respiratory tract infection, breastfeeding status, household exposure to cigarette smoke and the year the swab was collected. Each model also included the sequence count for each sample to account for potential differences in sequencing depth between samples – although this had little impact on the model output.
54 2.15.10Correlation between relative abundance and α-diversity
Spearman correlations between relative abundance for the seven most common genera and α-diversity (richness and Shannon diversity index) were analysed using the ‘cor.test’ function (method = “Spearman”) in the base ‘stats’ package in RStudio.
2.15.11 Multiple comparison correction
Due to high number of analyses conducted in chapter 4, results were only considered statistically significant after adjustment using the false discovery rate (Benjamini- Hochberg method) correction (258) with a q-value of 0.2 (based on the p-value distribution).
2.15.12 Adjusted prevalence ratios
In chapter 5, adjusted prevalence ratios and 95% confidence intervals were calculated using log binomial regression models, adjusting for age, sex, weekly household income, presence of a cough, presence of allergic rhinitis and antimicrobial use in the previous two weeks. Potential confounders were included in the models based on differences pre- vaccine introduction (2012) to post-vaccine introduction (2013-2015). Models were calculated using R packages ‘Epi’ version 2.10 and ‘Foreign’ version 0.8-67.
55 Chapter 3
Long-term effect of 23-valent pneumococcal polysaccharide vaccine on nasopharyngeal carriage in Fijian children
56 3 Long-term effect of 23-valent pneumococcal polysaccharide vaccine on nasopharyngeal carriage in Fijian children
3.1 Introduction and aims This chapter explores whether immune hyporesponsiveness observed following 23vPPV vaccination earlier in life has a long-term effect on carriage of S. pneumoniae. The publication of this work (231) forms the main body of the chapter.
3.1.1 Fiji Pneumococcal Project (FiPP)
As outlined previously, developing countries face challenges implementing pneumococcal vaccination programs, including issues with adequate serotype coverage by current PCVs and the significant costs of the vaccines. These issues underpinned the Fiji Pneumococcal Project (FiPP), a phase II vaccine trial in Suva, Fiji. FiPP was designed to investigate whether fewer doses of PCV could still provide adequate protection against pneumococci, and whether the inclusion of 23vPPV as a booster at 12 months could improve serotype protection.
To conduct FiPP, 554 infants were enrolled and randomised into eight groups to receive zero, one, two or three doses of PCV7, with or without a booster dose of 23vPPV at 12 months of age (Table 3.1). Prior to randomisation, children were stratified by ethnicity into indigenous Fijian (iTaukei), Fijians of Indian descent (Indo-Fijians) and ‘other’, ensuring the composition of the study participants was similar to the Fijian population (259). At 17 months, all groups were challenged with a micro-dose (mPPV) of the 23vPPV vaccine (20%) in order to mimic natural infection, and evaluate whether hyporesponsiveness was associated with the use of 23vPPV. Throughout the study, bloods and NP swabs were collected for evaluation. At the end of the FiPP study, participants 2 years of age were given a catch-up dose of PCV7 if they had initially received either no PCV7 or 1 dose of PCV7.
57
Table 3.1. Timing of vaccination, nasopharyngeal swabs, and blood draws for each group in FiPP, adapted from Russell et al. 2010 (229) Timing and procedure Group 6 wk 10 wk 14 wk 18 wk 6 mo 9 mo 12 mo 12.5 mo 17 mo 18 mo 2 y* A PCV7 PCV7 PCV7 B NP NP, B NP, B NP, B, mPPV B B PCV7 PCV7 PCV7 B NP NP NP, B, 23vPPV B NP, B, mPPV B C PCV7 PCV7 B NP NP, B NP, B NP, B, mPPV B D PCV7 PCV7 B NP NP NP, B, 23vPPV B NP, B, mPPV B E PCV7 B NP NP, B NP, B NP, B, mPPV B PCV7 F PCV7 B NP NP NP, B, 23vPPV B NP, B, mPPV B PCV7 G B NP NP NP, B, 23vPPV B NP, B, mPPV B PCV7 H B NP NP NP, B NP, B, mPPV B PCV7 PCV7, 7-valent pneumococcal conjugate vaccine; 23vPPV, 23-valent pneumococcal polysaccharide vaccine; mPPV, microdose (20%) of 23vPPV; NP, nasopharyngeal swab; B, blood draw; *PCV7 administered as a catch-up dose at the end of the FiPP study.
58
3.1.2 Immunological findings from FiPP
The evaluation of PCV7 vaccination in FiPP (259) suggested that a single dose may offer some protection for most PCV7 serotypes and that the serotype-specific antibody levels were similar in the children given 2 or 3 doses of PCV7, although the levels in the children given 3 doses of PCV7 were generally higher. Serotype-specific antibody levels were significantly higher at 17 months in children that received 23vPPV at 12 months compared with children who did not receive 23vPPV (230), and a greater proportion of children that received 23vPPV had serotype-specific IgG antibody levels above 0.35 μg/ml, the threshold of vaccine efficacy defined by the WHO (126). However, after challenge with mPPV at 17 months of age, an immune hyporesponsiveness was observed in children who previously received 23vPPV at 12 months of age, compared with children who did not receive 23vPPV (260). After adjusting for pre-mPPV antibody levels, the serotype-specific IgG responses were lower for all 23vPPV serotypes. Upon mPPV challenge, opsonophagocytic activity (OPA) was also negatively affected by 23vPPV vaccination for all serotypes except 1 and 5 (261).
3.1.3 Microbiological findings from FiPP
Similar to the immunological findings, assessment of carriage post-vaccination (229) showed that children given 2 or 3 doses of PCV7 had similar proportions or rates of vaccine type carriage, which was significantly lower than the rate of vaccine type carriage in unvaccinated individuals. There was some evidence that a single dose of PCV7 was able to reduce pneumococcal carriage at least initially, although this was not significant after adjusting for multiple comparisons. There was a dose-dependent effect on carriage, as children given more doses had correspondingly lower vaccine type carriage. There were no significant differences in carriage rates when comparing those receiving 23vPPV to those that did not.
At 17 months of age, the carriage rate and densities of S. pneumoniae and other respiratory pathogens (M. catarrhalis, H. influenzae and S. aureus) did not differ by vaccination status. However, differences were observed when the groups were separated by ethnicity (176). iTaukei children had higher carriage rates of S. pneumoniae, H. influenzae and M.
59
catarrhalis, and higher densities for S. pneumoniae and M. catarrhalis, than Indo-Fijian children. In the iTaukei children, those receiving PCV7 plus the 23vPPV had higher carriage densities of M. catarrhalis than those receiving no vaccine or PCV7 alone. No other significant differences were observed, although it is important to note S. aureus carriage is very low in this age group.
3.1.4 Fiji follow-up study (FFS) and chapter aims
The significance of the hyporesponsiveness observed in FiPP was unclear given that the serotype-specific antibody levels in 23vPPV-vaccinated children were at a level generally considered to be protective against invasive pneumococcal disease (i.e. >0.35 μg/ml). It was also unclear what impact this may have in the longer-term; in particular, would children given 23vPPV be more susceptible to pneumococcal colonisation or disease?
To investigate the long-term effects of polysaccharide vaccination in these children, the Fiji follow-up study (FFS) was designed to examine the microbiological, immunological and clinical implications of 23vPPV immune hyporesponsiveness. This study involved 194 children, previously enrolled in FiPP and now aged 5-7 years, with an approximately equal number of children in each of the groups. An additional PCV13 dose was administered, with NP swabs and blood samples collected prior to the PCV13 dose, and blood samples collected 28 days post-vaccination. As part of the immunological analysis, measurement of specific IgG to all 23 serotypes in 23vPPV, examination of the memory B cell response to 18 serotypes and opsonophagocytic assays to eight serotypes were examined.
The aim of this chapter is to examine the long-term effects of 23vPPV vaccination on NP carriage in children. To do this, we examined NP swabs from children in the FFS to determine the overall and vaccine type carriage rates as well as the density of pneumococcal carriage, with carriage rates and densities of other respiratory pathogens also determined.
60
3.2 Results from pilot FFS (additional to Boelsen et al. (231)) Twenty-one children that had previously participated in FiPP were recruited in 2010 as part of the pilot FFS, with blood samples and NP swabs used to validate the methods used in the follow-up study. Of the 21 individuals in the pilot FFS, 13 (62%) were carrying S. pneumoniae, 2 (10%) were carrying PCV7 serotypes and 3 (14%) were carrying 23vPPV serotypes. In total, 16 pneumococcal isolates were identified; non-typeable S. pneumoniae were the most common (25%), followed by serotypes 14, 16F and 31 (all 12.5%) (Figure 3.1).
Using qPCR, we found that 13 (62%) samples were positive for S. pneumoniae, 10 (48%) for H. influenzae, 15 (71%) for M. catarrhalis and 6 (29%) for S. aureus. Bacterial densities (Figure 3.2) were consistent with our previous work in Fiji (176). No differences were observed between the carriage densities of all the organisms tested (p>0.05, Mann- Whitney test).
Given the relatively low numbers of individuals in the pilot study, no further analyses were conducted. However, these approaches were then applied to the samples collected from the FFS participants to determine the long-term impact of pneumococcal polysaccharide vaccination on NP carriage (231) (below).
61
30
25 PCV7 types non-PCV7 23vPPV types 20 non-vaccine types
15
10
Percentage of total(%) 5
0 NT 14 31 16F 6C 13 19A 35B 35F 6A
Figure 3.1. Pneumococcal serotypes (identified by culture-based methods) in pilot FFS participants, as percentage of total isolates identified (n=16). NT = non-typeable.
62
10 9 l m / 8 s 10 t n e
l 7
a 10 v i
u 6
q 10 e
e 10 5 m o
n 4 e 10 g 10 3 s s e e li u a a a e z i h r n n rr u e o a a u m t . fl u a S n e c i n . . p M H . S
Figure 3.2. Density (genomic equivalents/ml) of S. pneumoniae, H. influenzae, M. catarrhalis and S. aureus carriage in the pilot FFS participants (n=21). Data are presented as the median and interquartile range. No significant differences were observed between the four bacterial species (p>0.05) using the Mann-Whitney test.
63 3.3 Publication: Boelsen et al. (2015) Long-term impact of pneumococcal polysaccharide vaccination on nasopharyngeal carriage in children previously vaccinated with various pneumococcal conjugate vaccine regimes. Vaccine 33: 5708-5714.
Contents lists available at ScienceDirect Vaccine
journal homepage: www.elsevier.com/locate/vaccine
Long-term impact of pneumococcal polysaccharide vaccination on nasopharyngeal carriage in children previously vaccinated with various pneumococcal conjugate vaccine regimes
Laura K. Boelsen a,b, Eileen M. Dunne a, Karen E. Lamb b,c, Kathryn Bright d, Yin Bun Cheung e,f, Lisi Tikoduadua g, Fiona M. Russell a,h, E. Kim Mulholland a,i, Paul V. Licciardi a,b,j, Catherine Satzke a,b,k,∗ a Pneumococcal Research, Murdoch Childrens Research Institute, Royal Children’s Hospital, Parkville, VIC, Australia b Department of Paediatrics, The University of Melbourne, Parkville, VIC, Australia c Clinical Epidemiology & Biostatistics Unit, Murdoch Childrens Research Institute, Royal Children’s Hospital, Parkville, VIC, Australia d Menzies School of Health Research, Darwin, NT, Australia e Center for Quantitative Medicine, Duke-NUS Graduate Medical School, Singapore, Singapore f Department of International Health, University of Tampere, Tampere, Finland g Ministry of Health, Suva, Fiji h Centre for International Child Health, Department of Paediatrics, The University of Melbourne, Parkville, VIC, Australia i London School of Hygiene and Tropical Medicine, London, UK j Allergy and Immune Disorders, Murdoch Childrens Research Institute, Royal Children’s Hospital, Parkville, VIC, Australia k Department of Microbiology and Immunology, Peter Doherty Institute for Infection and Immunity, The University of Melbourne, Parkville, VIC, Australia article info abstract
Article history: Previously, the Fiji Pneumococcal Project (FiPP) evaluated reduced dose immunization schedules that Received 30 March 2015 incorporated pneumococcal protein conjugate and/or polysaccharide vaccine (PCV7 and 23vPPV, respec- Received in revised form 3 July 2015 tively). Immune hyporesponsiveness was observed in children vaccinated with 23vPPV at 12 months of Accepted 20 July 2015 age compared with children who did not receive 23vPPV. Available online 29 July 2015 Here we assess the long-term impact of 23vPPV vaccination on nasopharyngeal carriage rates and densities of Streptococcus pneumoniae, Haemophilus influenzae, Staphylococcus aureus and Moraxella Keywords: catarrhalis. Nasopharyngeal swabs (n = 194) were obtained from healthy children who participated in Pneumococcal polysaccharide vaccine Pneumococcal conjugate vaccine FiPP (now aged 5–7 years). S. pneumoniae were isolated and identified by standard culture-based meth- Nasopharyngeal carriage ods, and serotyped using latex agglutination and the Quellung reaction. Carriage rates and densities of S. Streptococcus pneumoniae pneumoniae, H. influenzae, S. aureus and M. catarrhalis were determined using real-time quantitative PCR. Staphylococcus aureus There were no differences in the rate or density of S. pneumoniae, H. influenzae or M. catarrhalis carriage Ethnicity by PCV7 dose or 23vPPV vaccination in the vaccinated participants overall. However, differences were observed between the two main ethnic groups: Fijian children of Indian descent (Indo-Fijian) were less likely to carry S. pneumoniae, H. influenzae and M. catarrhalis, and there was evidence of a higher carriage rate of S. aureus compared with indigenous Fijian (iTaukei) children. Polysaccharide vaccination appeared to have effects that varied between ethnic groups, with 23vPPV vaccination associated with a higher carriage rate of S. aureus in iTaukei children, while there was a lower carriage rate of S. pneumoniae associated with 23vPPV vaccination in Indo-Fijian children. Overall, polysaccharide vaccination had no long-term impact on pneumococcal carriage, but may have impacted on S. aureus carriage and have varying effects in ethnic groups, suggesting current WHO vaccine schedule recommendations against the use of 23vPPV in children under two years of age are appropriate. © 2015 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
∗ Corresponding author at: Pneumococcal Research, Murdoch Childrens Research Institute, The Royal Children’s Hospital, Flemington Road, Parkville 3052, VIC, Australia. Tel.: +61 (0)383416438. E-mail address: [email protected] (C. Satzke). http://dx.doi.org/10.1016/j.vaccine.2015.07.059 0264-410X/© 2015 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4. 0/).
61 L.K. Boelsen et al. / Vaccine 33 (2015) 5708–5714 5709
1. Introduction enrolling all FiPP participants that were contactable (n = 195), now aged 5–7 years old. One hundred and ninety five nasopharyngeal Streptococcus pneumoniae (the pneumococcus) is the most com- swabs were collected, of which 194 were available for microbio- mon cause of pneumonia, which is responsible for an estimated logical analysis (Table 1). The characteristics of this subset, such as 1.3 million deaths annually in children under five years of age ethnicity, gender and number of children in each group, were simi- [1]. There are over 90 different serotypes of S. pneumoniae [2,3]. lar to that of FiPP (see Supplemental Table S1 and Table 2 in Russell Prior to the introduction of pneumococcal conjugate vaccines et al., 2010 [17]), and were similar when considering each vaccine (PCVs), the vast majority of pneumococcal disease was caused alone in this study (see Supplemental Table S2)—although children by a limited number of serotypes [4]. Nasopharyngeal carriage who received 3 doses of PCV7 were slightly older than children who of S. pneumoniae is considered a pre-requisite for the develop- received 2 doses of PCV7 (median age 6.21 years vs. 5.92 years, ment of pneumococcal disease [5]. The introduction of PCVs has p = 0.017). The geometric mean antibody levels in this study for led to a dramatic reduction in invasive pneumococcal disease 23vPPV vaccinated and 23vPPV unvaccinated children were simi- caused by vaccine-type S. pneumoniae [6–8]. However, in many lar to those observed in these children at 18 months of age in FiPP settings, there has been little change in overall S. pneumoniae (data not shown). carriage due to replacement of vaccine type S. pneumoniae with This study was approved by the Human Research Ethics Com- non-vaccine types (serotype replacement) [9,10]. The potential mittee, Royal Children’s Hospital, Melbourne and the Fiji National for species replacement following pneumococcal vaccination is Research Ethics Review Committee. also of concern, particularly given that some studies have found a negative relationship between vaccine-type S. pneumoniae and 2.2. Nasopharyngeal swabs Staphylococcus aureus carriage [11,12]. In some settings, pneumo- coccal vaccination has also affected colonization of the respiratory Buffered cotton nasopharyngeal swabs (Sarstedt, Australia) pathogens Moraxella catarrhalis and Haemophilus influenzae [13,14]. were collected and transported as described previously [17], in line The Fiji Pneumococcal Project (FiPP) was a single-blind, open- with World Health Organization guidelines [20,21]. Swabs were labelled randomized Phase II vaccine trial conducted in Suva, stored frozen in 1 ml of skim milk tryptone glucose glycerol (STGG) Fiji, designed to identify a pneumococcal vaccination schedule medium, and later transported on dry ice to the Pneumococcal more suited to resource-poor countries. Specifically, FiPP eval- Research laboratory at the Murdoch Childrens Research Institute uated reduced dose 7-valent pneumococcal conjugate vaccine in Melbourne, where they were stored at −80 ◦C until processing. (PCV7) primary series regimes in infancy, followed by the 23- valent pneumococcal polysaccharide vaccine (23vPPV) booster at 12 months of age [15]. One of the key findings was immune 2.3. Culture, identification and serotyping hyporesponsiveness in 23vPPV-vaccinated children compared with children not vaccinated with 23vPPV, observed following Samples were cultured on Columbia horse blood agar plates con- challenge with a micro-dose (20%) of 23vPPV at 17 months taining 5 g/ml of gentamicin (gHBA; Oxoid brand, Thermo Fisher ◦ ∼ of age [16]. Nasopharyngeal carriage rates of S. pneumoniae Scientific, Australia) and incubated for 36–44 h (37 C, 5% CO2) were unaffected by 23vPPV vaccination at the same time-point [22]. Two ␣-hemolytic colonies were randomly selected for subcul- [17]. ture using a method previously described [22]. These colonies, plus FiPP also highlighted the differences in carriage rates between any additional morphologically distinct ␣-hemolytic colonies, were the two main ethnicities of Fiji, indigenous Fijians (iTaukei) and subcultured onto Columbia horse blood agar plates (HBA; Oxoid Fijians of Indian descent (Indo-Fijians) [18], consistent with an brand, Thermo Fisher Scientific, Australia) and incubated for 24 h ◦ ∼ earlier carriage survey in the same setting [19]. In FiPP, carriage (37 C, 5% CO2). of S. pneumoniae, H. influenzae and M. catarrhalis was higher in Colonies that were optochin sensitive were presumptively iden- iTaukei children than Indo-Fijian children, and higher M. catarrhalis tified as S. pneumoniae. Other ␣-hemolytic colonies that were densities were found in iTaukei children vaccinated with 23vPPV non-susceptible to optochin (intermediate or resistant) were tested ® compared with iTaukei children who were not vaccinated with for bile solubility and with the Phadebact Pneumococcus test 23vPPV [18]. (Boule Diagnostics AB, Huddinge, Sweden) to enable identification. In the follow-up study to FiPP, a subset of children were enrolled Presumptive pneumococci were serotyped by latex agglutina- to investigate the long-term impact of 23vPPV vaccine on carriage, tion using a combination of commercial reagents (Denka-Seiken immunity and clinical outcomes. Here we report the long-term Co., Ltd., Japan) and reagents produced in-house [23,24] using effect of pneumococcal vaccination on nasopharyngeal carriage Statens Serum Institute antisera (SSI, Copenhagen, Denmark). of S. pneumoniae and other respiratory pathogens, and examine Equivocal reactions were confirmed using the Quellung reac- differences in carriage between the two main ethnicities of Fijian tion with antisera from SSI [25]. Isolates that were non-typeable children. were tested for the presence of the lytA gene by real-time PCR [26]. lytA-negative non-typeable isolates were excluded from 2. Materials and methods further analysis. lytA-positive non-typeable isolates were fur- ther examined by multilocus sequencing typing (MLST) [27]. 2.1. Study design MLST allelic profiles were submitted to the S. pneumoniae MLST database (http://pubmlst.org/spneumoniae/) sited at the Univer- The original FiPP study design is described elsewhere [15].In sity of Oxford [28] where new sequence types (STs) were assigned brief, healthy Fijian infants (n = 552) were randomized to receive for those isolates with allelic profiles not matching any existing either 0, 1, 2 or 3 doses of PCV7 (Prevnar®, Pfizer Inc., USA) at 6, ST. A serotype 14-specific PCR [29] was used to test non-typeable 10 and/or 14 weeks, with or without a 12 month dose of 23vPPV isolates that had MLST STs associated with serotype 14. (Pneumovax®, Merck & Co., Inc., USA). Children who were random- ized to receive 0 or 1 dose of PCV7 were given a catch-up dose of 2.4. Quantitative PCR this vaccine at 2 years of age. The long-term implications of the immune hyporesponsiveness observed at 18 months in children Genomic DNA was extracted from 100 l STGG using the QIAmp receiving 23vPPV were investigated in the Fiji follow-up study by DNA Minikit (Qiagen) as previously described [18]. S. pneumoniae,
62 5710 L.K. Boelsen et al. / Vaccine 33 (2015) 5708–5714
Table 1 Previous pneumococcal vaccinations received by the 194 participants in this study.
Group Initial PCV7 doses Initial PCV7 dose 23vPPV (12 mo) End of FiPP PCV7 Total PCV7 doses NP swabs collected timing (wks) dose* in this study (n)
A 3 6, 10, 14 − 0326 B 3 6, 10, 14 + 0 3 28 C 2 6 and 14 − 0225 D 2 6 and 14 + 0 2 30 E1 14 − 1215 F 1 14 + 1 2 25 G 0 None + 1 1 19 H 0 None − 1126
* PCV7 dose given at 2 years of age.
Table 2 MLST allelic profiles for the non-typeable (NT) isolates.
Isolate aroE gdh gki recP spi xpt ddl ST Associated serotypes**
0534–01 8 37 9 29 2 12 14 1619 19A, NT 0075–01 1 5 4 5 5 1 229 10183* Closest match ST9 (14) 0119–04 8 29 4 15 17 12 31 1106 14, NT 0541–01 8 29 4 15 17 12 31 1106 14, NT 0051–01 8 29 4 15 17 12 31 1106 14, NT 0173–01 7 139 4 8 13 38 216 10184* 0586–02 8 37 9 29 2 12 53 344 NT 0586–01 8 37 9 29 2 12 31 10185* Closest match ST344 (NT) 0918–03 8 37 9 29 2 12 31 10185* Closest match ST344 (NT) 0598–03 8 37 9 29 2 12 31 10185* Closest match ST344 (NT) 0009–01 8 37 9 29 2 12 31 10185* Closest match ST344 (NT) 0500–01 213 5 223 1 17 1 14 10182* Closest match ST3516 (19F) 0082–03 16 9 193 226 6 20 5 10181*
* Newly identified in this study; ST = sequence type; ** Serotypes associated with ST or closest match for newly identified STs.
S. aureus, M. catarrhalis and H. influenzae were examined by real- cross-reactive serotypes were considered non-vaccine type). As time quantitative PCR (qPCR) using two duplex assays [18], see serotype 15B is present in 23vPPV, 15B and 15C were considered Supplemental Table S4. Bacterial densities were determined using separately despite the potential of these serotypes to interconvert a standard curve from a dilution series prepared with genomic DNA [32]. Analyses were conducted using the Stata 12 software (release from reference isolates [18]. 12; StataCorp, College Station, TX, USA).
3. Results 2.5. Statistical analysis Of the 194 swabs tested, 85 (44%) were positive for S. pneumo- Logistic regression was used to examine the impact of PCV7 (1, niae, 99 (51%) for H. influenzae, 142 (73%) for M. catarrhalis and 53 2 or 3 doses) and 23vPPV (0 or 1 dose) on carriage of S. pneumoniae, (27%) for S. aureus by qPCR. Using culture, 88 (45%) swabs contained H. influenzae, S. aureus and M. catarrhalis, with each species exam- pneumococci, of which 13 (15%) contained PCV7 vaccine types, 37 ined in separate models. To determine if the effect of 23vPPV was (42%) contained 23vPPV types (including PCV7 types), and 51 (58%) dependent on the number of doses of PCV7 received, an interaction only contained serotypes absent from either vaccine. Overall, 103 between 23vPPV and the number of doses of PCV7 was also con- isolates were identified; non-typeable S. pneumoniae were most sidered in each model. Regression models were used to examine common (n = 13, 12.6%), followed by serotypes 11A (n = 10, 9.7%), the effect of PCV7 and 23vPPV on density, with bacterial density 6A (n = 8, 7.8%) and 16F (n = 7, 6.8%) (Fig. 1). data log transformed prior to analysis. To examine bacterial associ- ations, logistic regression models were fitted to determine the odds 3.1. Non-typeable S. pneumoniae of carrying one bacterial species given carriage of another species. For analyses of vaccine-type carriage culture data were used, while All 19 ␣-hemolytic isolates that were presumptively identified qPCR data were used for all other analyses. as non-typeable S. pneumoniae were tested by lytA real-time PCR. For ethnicity based analyses, due to the small number of indi- Six were negative for the lytA gene and were designated as non- viduals who were of ‘other’ ethnicity (n = 6), only those of iTaukei S. pneumoniae. The remaining 13 isolates were examined by MLST (n = 139) and Indo-Fijian (n = 49) ethnicity were included. Where (Table 2) and represented eight STs, including five new STs. Four zero counts were observed, exact logistic regression was used isolates had STs commonly associated with serotype 14; however, to obtain the median unbiased estimate of the odds ratio [30]. they were negative when tested using a serotype 14-specific PCR Characteristics of study participants by ethnic group are shown and thus designated non-typeable pneumococci. in Supplemental Table S3. An adjusted analyses for potential con- founders for comparisons between iTaukei and Indo-Fijian children 3.2. Impact of pneumococcal vaccination on nasopharyngeal was considered, however as these variables are considered inter- carriage rates and densities mediary steps in the relationship between ethnicity and carriage, and not true confounders, this analysis was not conducted [31]. None of the models considering an interaction term between Where analyses examine vaccine-type serotypes, only those 23vPPV and PCV7 doses showed significant interaction (each serotypes included in each vaccine were considered (potentially p > 0.05; details not shown). Adjustment for PCV7 doses and
63 L.K. Boelsen et al. / Vaccine 33 (2015) 5708–5714 5711
Fig. 1. Pneumococcal serotypes (identified by culture-based methods) carried in study participants as a percentage of total isolates identified (n = 102). NT = non-typeable, other (each serotype represents ∼1% of total isolates) = 1*, 3*, 9A, 10A*, 10B, 15A, 18C**, 19B, 19 C, 35B, 35 C, 38 (**PCV7 types, *non-PCV7 23vPPV types).
70 S. pneumoniae PCV7 vaccine type 60 23vPPV vaccine type (%)
50
Carriage 40
oniae 30
20 S. pneu m 10
0 23vPPV (-) 23vPPV (+) 23vPPV (-) 23vPPV(+) 23vPPV (-) 23vPPV (+) 23vPPV (-) 23vPPV (+) 3 PCV7 doses (n=54)2 PCV7 doses (n=95) 1 PCV7 dose (n=45) Overall
Fig. 2. S. pneumoniae carriage (overall, PCV7 vaccine type and 23vPPV vaccine type carriage identified by culture) by vaccination status in children in Fiji aged 5–7 years. No significant difference by vaccination status in overall S. pneumoniae, PCV7 vaccine type or 23vPPV vaccine type carriage.
23vPPV/PCV7 dose interaction gave results that were similar to the 3.3. Bacterial associations results from unadjusted analysis of the effect of 23vPPV. As such unadjusted results are presented unless specified. Carriage of S. pneumoniae was positively associated with The number of PCV7 doses received had no statistically sig- carriage of H. influenzae and M. catarrhalis (Table 3). Positive associ- nificant impact on pneumococcal carriage rate or density (each ations were also observed between H. influenzae and M. catarrhalis. p > 0.05). Similarly, there were no differences in the rate or den- A negative association, where carriage of one species is less likely to sity of pneumococcal carriage between 23vPPV recipients and occur given the presence of another species, was observed between non-recipients, regardless of adjustment for the number of PCV7 S. aureus and M. catarrhalis, and between S. aureus and H. influenzae. doses received (each p > 0.05). There were no statistically signifi- cant differences in the carriage rates (Fig. 2) or densities of vaccine 3.4. Effect of ethnicity type (PCV7 and 23vPPV) carriage by doses of either vaccine (each p > 0.05). Overall, 1 (2%) Indo-Fijian child and 35 (25%) iTaukei children We found no long-term impact of 23vPPV or PCV7 vaccination were carrying 23vPPV vaccine types, with 1 (2%) Indo-Fijian child on H. influenzae carriage rates or densities. S. aureus carriage rates and 11 (8%) Indo-Fijian children carrying PCV7 vaccine types. Sig- were higher in 23vPPV recipients (OR = 2.41, 95% CI (1.24, 4.68); nificant differences in overall carriage rates were found between p = 0.010) compared with children that did not receive 23vPPV, the two main ethnicities. Indo-Fijian children had lower odds of with 36 (35%) 23vPPV recipients carrying S. aureus compared with carrying S. pneumoniae (OR = 0.20, 95% CI (0.09, 0.44); p < 0.001), 17 (18%) children that did not receive 23vPPV. M. catarrhalis den- H. influenzae (OR = 0.18, 95% CI (0.08, 0.38); p < 0.001) and M. sity was higher (p = 0.036) in children that received 1 dose of PCV7 catarrhalis (OR = 0.12, 95% CI (0.06, 0.24); p < 0.001), and higher 5 (median density 1.51 × 10 cfu/ml) compared with children who odds of carrying S. aureus (OR = 2.31, 95% CI (1.15, 4.61); p = 0.018) 6 received 3 doses of PCV7 (median density 1.38 × 10 cfu/ml). compared with iTaukei children. Carriage densities between the
64 5712 L.K. Boelsen et al. / Vaccine 33 (2015) 5708–5714
Table 3 Bacterial associations: odds ratio (OR) of carriage of one species given carriage of another as determined by qPCR.
Overall H. influenzae M. catarrhalis S. aureus
n (%) n (%) OR (95% CI); p-value n (%) OR (95% CI); p-value n (%) OR (95% CI); p-value
S. pneumoniae (all serotypes) 85 (44) 65 (34) 7.17 (3.76, 13.66); p < 0.001* 72 (37) 3.09 (1.52, 6.27); p = 0.002* 19 (10) 0.64 (0.33, 1.22); p = 0.172 S. pneumoniae (PCV7 VT) 13 (7) 9 (5) 2.28 (0.68, 7.65); p = 0.184 10 (5) 1.24 (0.33, 4.68); p = 0.754 3 (2) 0.79 (0.21, 2.97); p = 0.723 S. pneumoniae (23vPPV VT) 37 (19) 26 (13) 2.72 (1.26, 5.88); p = 0.011* 33 (17) 3.63 (1.22, 10.83); p = 0.021* 8 (4) 0.69 (0.29, 1.62); p = 0.389 H. influenzae 99 (51) 84 (43) 3.57 (1.80, 7.10); p < 0.001* 20 (10) 0.48 (0.25, 0.91); p = 0.024* M. catarrhalis 142 (73) 29 (15) 0.30 (0.15, 0.59); p = 0.001* S. aureus 53 (27)
* Significant (p < 0.05); VT = vaccine type; percentage is out of the 194 participants.
Table 4 Odds ratios for carriage of S. pneumoniae, H. influenzae, M. catarrhalis and S. aureus in iTaukei and Indo-Fijian 23vPPV recipients compared with 23vPPV non-recipients as determined by qPCR.
All no 23vPPV recipients 23vPPV recipients OR 95% CI p-Value
iTaukei (n = 139***) S. pneumoniae n (%) 74 (53) 35 (55) 39 (52) 0.90 0.46, 1.75 0.752 H. influenzae n (%) 86 (62) 41 (64) 45 (60) 0.84 0.42, 1.68 0.623 M. catarrhalis n (%) 119 (86) 59 (92) 60 (79) 0.34 0.12, 0.99 0.048** S. aureus n (%) 32 (23) 6 (9) 26 (35) 5.13 1.95, 13.47 0.001*
Indo-Fijian (n = 49****) S. pneumoniae n (%) 9 (18) 9 (39) 0 (0) 0.06***** 0.00, 0.45 0.003* H. influenzae n (%) 11 (22) 5 (22) 6 (23) 1.48 0.39, 5.71 0.567 M. catarrhalis n (%) 20 (41) 13 (57) 7 (27) 0.44 0.14, 1.42 0.168 S. aureus n (%) 20 (41) 11 (48) 9 (35) 0.88 0.28, 2.75 0.821
* Significant (p < 0.05). ** Not significant after adjusting for PCV7 dose, with (p = 0.310) or without (p = 0.053) including interaction between PCV7 and 23vPPV. *** Of the 139 iTaukei children, there were 75 with 23vPPV and 64 with no 23vPPV. **** Of the 49 Indo-Fijian children, there were 26 with 23vPPV and 23 with no 23vPPV. ***** Exact logistic regression used to obtain the median unbiased estimate due to the zero count. two ethnic groups differed for M. catarrhalis (median density of Of the 88 children in our study carrying pneumococci, 37 (42%) 7.00 × 105 cfu/ml for iTaukei vs 9.49 × 104 cfu/ml for Indo-Fijians; were carrying serotypes included in 23vPPV, a coverage rate which p = 0.008) but not for the other three bacterial species. may suggest a justification for the use of 23vPPV in this popula- No Indo-Fijian 23vPPV recipients were found to carry S. pneumo- tion. However, 21 of the 41 (51%) children vaccinated with 23vPPV niae, while 39% of Indo-Fijian children who did not receive 23vPPV were carrying 23vPPV types compared with 16 of the 47 (34%) chil- were carriers (Table 4). There was no evidence of a difference in the dren not vaccinated with 23vPPV, suggesting that 23vPPV does not odds of carrying S. pneumoniae among iTaukei who did and did not reduce carriage of 23vPPV serotypes in the long-term. receive 23vPPV; however, the odds of carrying S. aureus were signif- In contrast to the original FiPP study, which found a dose- icantly higher in 23vPPV recipients compared with non-recipients dependent effect of PCV7 on pneumococcal carriage [17], we found (Table 4). There was no effect of 23vPPV vaccination on carriage no long-term differences in pneumococcal carriage in children who of H. influenzae or M. catarrhalis (after adjustment for PCV7 doses received 1, 2 or 3 PCV7 doses. However, due to the catch-up dose at received) for either ethnic group. There was no evidence that PCV7 the end of the previous study, there was no PCV7-unvaccinated had an effect on carriage for any of the bacterial species by ethnic group in this study to serve as a comparator. PCV7 vaccination group. appears to have eliminated carriage for three of the seven PCV7 serotypes (serotypes 4, 6B and 9V). Consistent with previous studies, we found differences in car- 4. Discussion riage between the two main ethnic groups in Fiji [18,19]. These differences were most striking when examining the effect of Immune hyporesponsiveness following pneumococcal polysac- 23vPPV vaccination after stratifying by ethnicity, as the association charide vaccination at 12 months of age was described previously between 23vPPV receipt and increased S. aureus carriage appears for the children in this study [16]. The long-term implications to only occur in iTaukei children. Additionally, there was evidence of hyporesponsiveness on nasopharyngeal carriage are unknown, to suggest that 23vPPV may have reduced S. pneumoniae carriage despite concerns that this may lead to increased pneumococcal dis- in Indo-Fijians. In FiPP, there was no evidence that 23vPPV vaccina- ease susceptibility in these children. Currently the World Health tion had varying effects on carriage rates in the two ethnic groups Organization does not recommend 23vPPV to children under the but iTaukei 23vPPV recipients had higher carriage densities of M. age of two years based on immunological findings [33], how- catarrhalis [18], findings not replicated in our study. However, the ever there are limited data on the effects of 23vPPV vaccination carriage rate of S. aureus was very low in FiPP (3.3%, typical for 17- on nasopharyngeal carriage in children and no published studies month old children) compared with our study (27%). Additionally, looking at long-term effects. In this study, we found no long- analyses of carriage for FiPP were limited to the groups given 3 term impact of 23vPPV on S. pneumoniae carriage. Early studies of doses of PCV7 with 23vPPV, 3 doses of PCV7 without 23vPPV and other pneumococcal polysaccharide vaccines reported no effect of the unvaccinated control group, whereas our study used data from polysaccharide vaccination on nasopharyngeal carriage of S. pneu- all groups. moniae [34–36]. More recently, two short-term studies, the original Ethnic differences in susceptibilities to infectious diseases such FiPP study [17] and another study in children aged 1–7 years with as tuberculosis, dengue fever and malaria [38–40], as well as the acute otitis media [37], also found that 23vPPV vaccination had no immune response to rubella vaccination [41], have been observed effect on carriage. in other populations. There was no difference between iTaukei and
65 L.K. Boelsen et al. / Vaccine 33 (2015) 5708–5714 5713
Indo-Fijian children in this study for many pneumococcal carriage numbers in this study, larger studies would be required to fully risk factors such as antibiotic use and exposure to cigarette smoke explore some observations and relationships, such as differences [42], although household income was higher for Indo-Fijian chil- in ethnicity and the impact of 23vPPV on S. aureus carriage. dren (median weekly income ($US) $82.50 vs. $57.75, p = 0.001). We found a relatively large number of non-typeable pneumo- While socio-economic status is a risk factor for pneumococcal cocci in this study and identified four new MLST STs. Other studies carriage [43], the differences in carriage between iTaukei and Indo- with a similar methodology have also found a relatively large pro- Fijian children in our study are consistent with previous studies in portion of non-typeables [55]. Most of our non-typeable isolates Fiji [18,19], where household income levels were not different. It is had STs (or close matches to STs) commonly associated with non- therefore unlikely that the differences observed in the two ethnic typeables, such as ST344. However, three were ST1106 which is groups in our study are attributable to economic status. The rea- associated with serotype 14. Although previous studies have found sons underpinning the differences between the two main groups non-typeable isolates genetically related to serotype 14 [56,57], in Fiji are unknown but maybe due to unknown household and further investigation of our isolates using a serotype 14-specific social factors, or genetic susceptibility. PCR suggest that these isolates may lack part of the capsule locus Some, but not all, studies have reported increased S. aureus or have capsule sequences that are sufficiently divergent that they carriage following PCV7 vaccination [44,45], consistent with obser- are not detected by the serotype 14-specific PCR. vations of a negative relationship between carriage of S. aureus and vaccine-type S. pneumoniae [13,14,46]. In this study, there was no 5. Conclusions clear evidence that the association between 23vPPV receipt and increased S. aureus carriage in iTaukei children was due to bac- Although 23vPPV vaccination was associated with immune terial interactions or species replacement given that there was hyporesponsiveness in children at 17 months, it had no impact on not a corresponding reduction in vaccine-type S. pneumoniae car- S. pneumoniae carriage rates or densities at 17 months and no long- riage. Most of the bacterial interactions we found are consistent term impact on S. pneumoniae carriage rates or densities in children with other studies; such as the positive relationships between S. now aged 5–7 years. However, our data suggest that 23vPPV vac- pneumoniae and M. catarrhalis, and S. pneumoniae and H. influenzae cination may be associated with higher S. aureus carriage, and may [43,47,48]. We found no association between S. aureus and S. pneu- have varying effects in different ethnic groups. Despite observing moniae, vaccine-type or otherwise. However, as all children were no long-term impact on S. pneumoniae carriage rates or density, vaccinated with PCV7, PCV7 vaccine-type carriage was uncom- 23vPPV vaccination results in immune hyporesponsiveness, has mon. While bacterial interactions or species replacement do not potential long-term effects on S. aureus carriage, and there is a lack appear to have an effect at this time point, we have no carriage of evidence showing a clear benefit of using 23vPPV in children data between 17 months and 5–7 years of age, and it is possible under two years of age. As such, current guidelines recommen- that the changing carriage dynamics during this period may have ding against the use of 23vPPV in children under 2 years of age played a role that is no longer apparent. are appropriate. There are limited data on the impact of 23vPPV vaccination on S. aureus carriage and so the mechanism behind this observation Acknowledgements is unknown. One study in children with a history of otitis media noted higher acute otitis media attributable to S. aureus in chil- We thank all the participants, families and staff involved in dren vaccinated with PCV7 and 23vPPV compared with children FiPP and this study. This work was funded by National Insti- that received no pneumococcal vaccination, however carriage of tutes of Health (NIH), USA (NIH Grant 1R01AI085198-01A1) and S. aureus was not examined [49]. The presence of pilus in S. pneu- was supported by the Victorian Government’s Operational Infra- moniae [50] and, to a lesser degree, hydrogen peroxide-mediated structure Support Program. Laura Boelsen is supported by the killing of S. aureus by S. pneumoniae [51] in vitro have both been Fay Marles Scholarship (The University of Melbourne). Paul Lic- implicated in the negative association between S. pneumoniae and S. ciardi and Fiona Russell are recipients of the Australian National aureus. However, it is clear the immune system also plays an impor- Health and Medical Research Council (NHMRC) Early Career Fel- tant role as studies examining the effect of PCV in HIV-infected lowship (APP1013820 and APP1035341), and Catherine Satzke is a and HIV-uninfected children have shown no negative association recipient of the Australian NHMRC Career Development Fellowship between S. pneumoniae and S. aureus in the immunocompromised (APP1087957). Karen Lamb was supported under a National Health children [52,53]. Additionally, prior S. pneumoniae colonization in and Medical Research Council Centre of Research Excellence grant a mouse model was found to inhibit S. aureus acquisition due to (ID31035261) to the Victorian Centre for Biostatistics (ViCBiostat). cross-reactive antibodies to a pneumococcal dehydrogenase [54]. The clinical implication of higher S. aureus carriage in iTaukei Appendix A. Supplementary data 23vPPV vaccinated individuals (compared with iTaukei children not vaccinated with 23vPPV) is also unclear, although our prelim- Supplementary data associated with this article can be found, in inary analyses show no change in hospital admissions between the online version, at http://dx.doi.org/10.1016/j.vaccine.2015.07. groups (Licciardi et al., unpublished results). However, it is pos- 059. sible this may affect diseases for which hospitalization would be unlikely, such as otitis media and skin infections. It should also Conflict of interest statement be noted that although 23vPPV was associated with higher car- riage of S. aureus in iTaukei children, the carriage rate of S. aureus None. in iTaukei 23vPPV recipients was similar to that of Indo-Fijian children who did not receive 23vPPV (35% vs. 48%, respectively). References Immunologically, preliminary analyses show that there were no differences in antibody levels (geometric mean antibody concen- [1] Walker CL, Rudan I, Liu L, Nair H, Theodoratou E, Bhutta ZA, et al. Global burden trations) between 23vPPV vaccinated and 23vPPV unvaccinated of childhood pneumonia and diarrhoea. Lancet 2013;381:1405–16. children, and that the antibody levels seen in each ethnic group [2] Henrichsen J. Six newly recognized types of Streptococcus pneumoniae. J Clin Microbiol 1995;33:2759–62. were similar, consistent with findings from the original FiPP study [3] Park IH, Geno KA, Yu J, Oliver MB, Kim K-H, Nahm MH. Genetic, biochemi- (Licciardi et al., unpublished results). Given the relatively small cal, and serological characterization of a new pneumococcal serotype, 6H, and
66 5714 L.K. Boelsen et al. / Vaccine 33 (2015) 5708–5714
generation of a pneumococcal strain producing three different capsular repeat [30] Mehta CR, Patel NR. Exact logistic regression: theory and examples. Stat Med units. Clin Vaccine Immunol 2015;22:313–8. 1995;14:2143–60. [4] Johnson HL, Deloria-Knoll M, Levine OS, Stoszek SK, Freimanis Hance L, [31] Schisterman EF, Cole SR, Platt RW. Overadjustment bias and unnecessary Reithinger R, et al. Systematic evaluation of serotypes causing invasive adjustment in epidemiologic studies. Epidemiology 2009;20:488–95. pneumococcal disease among children under five: the pneumococcal global [32] van Selm S, van Cann LM, Kolkman MAB, van der Zeijst BAM, van Put- serotype project. PLoS Med 2010;7:e1000348. ten JPM. Genetic basis for the structural difference between Streptococcus [5] Simell B, Auranen K, Kayhty H, Goldblatt D, Dagan R, O’Brien KL. The funda- pneumoniae serotype 15B and 15C capsular polysaccharides. Infect Immun mental link between pneumococcal carriage and disease. Expert Rev Vaccines 2003;71:6192–8. 2012;11:841–55. [33] World Health Organization. 23-valent pneumococcal polysaccharide vaccine. [6] Whitney CG, Farley MM, Hadler J, Harrison LH, Bennett NM, Lynfield R, WHO position paper. Wkly Epidemiol Rec 2008;83:373–84. et al. Decline in invasive pneumococcal disease after the introduction of [34] Douglas RM, Hansman D, Miles HB, Paton JC. Pneumococcal carriage and type- protein–polysaccharide conjugate vaccine. N Engl J Med 2003;348:1737–46. specific antibody: failure of a 14-valent vaccine to reduce carriage in healthy [7] Miller E, Andrews NJ, Waight PA, Slack MP, George RC. Herd immunity and children. Am J Dis Child 1986;140:1183–5. serotype replacement 4 years after seven-valent pneumococcal conjugate vac- [35] Rosén C, Christensen P, Hovelius B, Prellner K. A longitudinal study of the cination in England and Wales: an observational cohort study. Lancet Infect Dis nasopharyngeal carriage of pneumococci as related to pneumococcal vaccina- 2011;11:760–8. tion in children attending day-care centres. Acta Otolaryngol 1984;98:524–32. [8] Williams SR, Mernagh PJ, Lee MH, Tan JT. Changing epidemiology of invasive [36] Wright PF, Sell SH, Vaughn WK, Andrews C, McConnell KB, Schiffman G. Clinical pneumococcal disease in Australian children after introduction of a 7-valent studies of pneumococcal vaccines in infants. II. Efficacy and effect on nasopha- pneumococcal conjugate vaccine. Med J Aust 2011;194:116–20. ryngeal carriage. Rev Infect Dis 1981;3:S108–12. [9] Weinberger DM, Malley R, Lipsitch M. Serotype replacement in disease after [37] Veenhoven RH, Bogaert D, Schilder AGM, Rijkers GT, Uiterwaal CSPM, Kieze- pneumococcal vaccination. Lancet 2011;378:1962–73. brink HH, et al. Nasopharyngeal pneumococcal carriage after combined [10] Vestrheim DF, Høiby EA, Aaberge IS, Caugant DA. Impact of a pneumococcal pneumococcal conjugate and polysaccharide vaccination in children with a conjugate vaccination program on carriage among children in Norway. Clin history of recurrent acute otitis media. Clin Infect Dis 2004;39:911–9. Vaccine Immunol 2010;17:325–34. [38] Modiano D, Petrarca V, Sirima BS, Nebié I, Diallo D, Esposito F, et al. Different [11] Bogaert D, van Belkum A, Sluijter M, Luijendijk A, de Groot R, Rümke HC, et al. response to Plasmodium falciparum malaria in West African sympatric ethnic Colonisation by Streptococcus pneumoniae and Staphylococcus aureus in healthy groups. Proc Natl Acad Sci USA 1996;93:13206–11. children. Lancet 2004;363:1871–2. [39] Sierra BdlC, Kourí G, Guzmán MG. Race: a risk factor for dengue hemorrhagic [12] Regev-Yochay G, Dagan R, Raz M, Carmeli Y, Shainberg B, Derazne E, et al. fever. Arch Virol 2007;152:533–42. Association between carriage of Streptococcus pneumoniae and Staphylococcus [40] Stead WW, Senner JW, Reddick WT, Lofgren JP. Racial differences in suscepti- aureus in children. JAMA 2004;292:716–20. bility to infection by Mycobacterium tuberculosis.N Engl J Med 1990;322:422–7. [13] Dunne EM, Smith-Vaughan HC, Robins-Browne RM, Mulholland EK, Satzke C. [41] Haralambieva IH, Salk HM, Lambert ND, Ovsyannikova IG, Kennedy RB, Warner Nasopharyngeal microbial interactions in the era of pneumococcal conjugate ND, et al. Associations between race, sex and immune response variations to vaccination. Vaccine 2013;31:2333–42. rubella vaccination in two independent cohorts. Vaccine 2014;32:1946–53. [14] Shak JR, Vidal JE, Klugman KP. Influence of bacterial interactions on pneumo- [42] Ghaffar F, Friedland IR, McCracken Jr GH. Dynamics of nasopharyngeal colo- coccal colonization of the nasopharynx. Trends Microbiol 2013;21:129–35. nization by Streptococcus pneumoniae. Pediatr Infect Dis J 1999;18:638–46. [15] Russell FM, Balloch A, Tang MLK, Carapetis JR, Licciardi P, Nelson J, et al. [43] Jourdain S, Smeesters PR, Denis O, Dramaix M, Sputael V, Malaviolle X, et al. Immunogenicity following one, two, or three doses of the 7-valent pneumo- Differences in nasopharyngeal bacterial carriage in preschool children from coccal conjugate vaccine. Vaccine 2009;27:5685–91. different socio-economic origins. Clin Microbiol Infect 2011;17:907–14. [16] Russell FM, Carapetis JR, Balloch A, Licciardi PV, Jenney AWJ, Tikoduadua L, et al. [44] Spijkerman J, Prevaes SM, van Gils EJ, Veenhoven RH, Bruin JP, Bogaert D, et al. Hyporesponsiveness to re-challenge dose following pneumococcal polysac- Long-term effects of pneumococcal conjugate vaccine on nasopharyngeal car- charide vaccine at 12 months of age, a randomized controlled trial. Vaccine riage of S. pneumoniae, S. aureus. H. influenzae and M. catarrhalis. PLoS ONE 2010;28:3341–9. 2012;7:e39730. [17] Russell FM, Carapetis JR, Satzke C, Tikoduadua L, Waqatakirewa L, Chandra [45] van Gils EJ, Hak E, Veenhoven RH, Rodenburg GD, Bogaert D, Bruin JP, R, et al. Pneumococcal nasopharyngeal carriage following reduced doses of et al. Effect of seven-valent pneumococcal conjugate vaccine on Staphylococ- a 7-valent pneumococcal conjugate vaccine and a 23-valent pneumococcal cus aureus colonisation in a randomised controlled trial. PLoS ONE 2011;6: polysaccharide vaccine booster. Clin Vaccine Immunol 2010;17:1970–6. e20229. [18] Dunne EM, Manning J, Russell FM, Robins-Browne RM, Mulholland EK, [46] Lijek RS, Weiser JN. Co-infection subverts mucosal immunity in the upper respi- Satzke C. Effect of pneumococcal vaccination on nasopharyngeal carriage of ratory tract. Curr Opin Immunol 2012;24:417–23. Streptococcus pneumoniae, Haemophilus influenzae, Moraxella catarrhalis, and [47] Xu Q, Almudervar A, Casey JR, Pichichero ME. Nasopharyngeal bacterial inter- Staphylococcus aureus in Fijian children. J Clin Microbiol 2012;50:1034–8. actions in children. Emerg Infect Dis 2012;18:1738–45. [19] Russell FM, Carapetis JR, Ketaiwai S, Kunabuli V, Taoi M, Biribo S, et al. Pneumo- [48] Bae S, Yu J-Y, Lee K, Lee S, Park B, Kang Y. Nasal colonization by four potential coccal nasopharyngeal carriage and patterns of penicillin resistance in young respiratory bacteria in healthy children attending kindergarten or elementary children in Fiji. Ann Trop Paediatr Int Child Health 2006;26:187–97. school in Seoul, Korea. J Med Microbiol 2012;61:678–85. [20] O’Brien KL, Nohynek H. The WHO pneumococcal vaccine trials carriage work- [49] Veenhoven R, Bogaert D, Uiterwaal C, Brouwer C, Kiezebrink H, Bruin J, et al. ing group. Report from a WHO Working Group: standard method for detecting Effect of conjugate pneumococcal vaccine followed by polysaccharide pneu- upper respiratory carriage of Streptococcus pneumoniae. Pediatr Infect Dis J mococcal vaccine on recurrent acute otitis media: a randomised study. Lancet 2003;22:e1–11. 2003;361:2189–95. [21] Satzke C, Turner P, Virolainen-Julkunen A, Adrian PV, Antonio M, Hare KM, [50] Regev-Yochay G, Lipsitch M, Basset A, Rubinstein E, Dagan R, Raz M, et al. et al. Standard method for detecting upper respiratory carriage of Streptococcus The pneumococcal pilus predicts the absence of Staphylococcus aureus co- pneumoniae: updated recommendations from the World Health Organization colonization in pneumococcal carriers. Clin Infect Dis 2009;48:760–3. Pneumococcal Carriage Working Group. Vaccine 2013;32:165–79. [51] Regev-Yochay G, Trzcinski´ K, Thompson CM, Malley R, Lipsitch M. Interfer- [22] Satzke C, Dunne EM, Porter BD, Antonio M, O’Brien KL, Robins-Browne RM, ence between Streptococcus pneumoniae and Staphylococcus aureus: in vitro et al. PneuCarriage Project: identifying the optimum pneumococcal serotyping hydrogen peroxide-mediated killing by Streptococcus pneumoniae. J Bacteriol method(s), including detection of multiple serotype carriage. In: Ninth Interna- 2006;188:4996–5001. tional Symposium on Pneumococci and Pneumococcal Disease (ISPPD9). 2014. [52] Madhi SA, Adrian P, Kuwanda L, Cutland C, Albrich WC, Klugman KP. Long-term [23] Ortika BD, Habib M, Dunne EM, Porter BD, Satzke C. Production of latex agglu- effect of pneumococcal conjugate vaccine on nasopharyngeal colonization by tination reagents for pneumococcal serotyping. BMC Res Notes 2013;6:49. Streptococcus pneumoniae – and associated interactions with Staphylococcus [24] Porter BD, Ortika BD, Satzke C. Capsular serotyping of Streptococ- aureus and Haemophilus influenzae colonization – in HIV-infected and HIV- cus pneumoniae by latex agglutination. J Vis Exp 2014;91:e51747, uninfected children. J Infect Dis 2007;196:1662–6. http://dx.doi.org/10.3791/51747. [53] McNally LM, Jeena PM, Gajee K, Sturm AW, Tomkins AM, Coovadia HM, et al. [25] Habib M, Porter BD, Satzke C. Capsular serotyping of Streptococcus Lack of association between the nasopharyngeal carriage of Streptococcus pneu- pneumoniae using the Quellung reaction. J Vis Exp 2014;84:e51208, moniae and Staphylococcus aureus in HIV-1-infected South African children. J http://dx.doi.org/10.3791/51208. Infect Dis 2006;194:385–90. [26] Carvalho MdGS, Tondella ML, McCaustland K, Weidlich L, McGee L, Mayer [54] Lijek RS, Luque SL, Liu Q, Parker D, Bae T, Weiser JN. Protection from the acqui- LW, et al. Evaluation and improvement of real-time PCR assays targeting sition of Staphylococcus aureus nasal carriage by cross-reactive antibody to a lytA, ply, and psaA genes for detection of pneumococcal DNA. J Clin Microbiol pneumococcal dehydrogenase. Proc Natl Acad Sci USA 2012;109:13823–8. 2007;45:2460–6. [55] Turner P, Turner C, Jankhot A, Helen N, Lee SJ, Day NP, et al. A longitudinal study [27] Dunne EM, Ong EK, Moser RJ, Siba PM, Phuanukoonnon S, Greenhill AR, et al. of Streptococcus pneumoniae carriage in a cohort of infants and their mothers Multilocus sequence typing of Streptococcus pneumoniae by use of mass spec- on the Thailand–Myanmar border. PLoS ONE 2012;7:e38271. trometry. J Clin Microbiol 2011;49:3756–60. [56] Chewapreecha C, Harris SR. Dense genomic sampling identifies highways of [28] Jolley K, Maiden M. BIGSdb Scalable analysis of bacterial genome variation at pneumococcal recombination. Nat Genet 2014;46:305–9. the population level. BMC Bioinf 2010;11:595. [57] Andrade AL, Franco CM, Lamaro-Cardoso J, Andre MC, Oliveira LL, Kipnis A, et al. [29] Dias CA, Teixeira LM, Carvalho MdGS, Beall B. Sequential multiplex PCR for Non-typeable Streptococcus pneumoniae carriage isolates genetically similar to determining capsular serotypes of pneumococci recovered from Brazilian chil- invasive and carriage isolates expressing capsular type 14 in Brazilian infants. dren. J Med Microbiol 2007;56:1185–8. J Infect 2010;61:314–22.
67
3.3.1 Supplemental content
Table 3.2. Demographic characteristics of children in FFS stratified by PCV7 dose and 23vPPV status. 3 PCV7 doses 2 PCV7 doses 1 PCV7 doses Parameter at swab collection No 23vPPV 23vPPV No 23vPPV 23vPPV No 23vPPV 23vPPV No. of children 26 28 40 55 26 19 Male 13 (50) 15 (53) 18 (45) 21 (38) 12 (46) 7 (36) Median age (y) 6.29 6.13 6.08 5.58 5.92 6.17 Ethnicity iTaukei 16 (62) 21 (75) 32 (80) 44 (80) 16 (62) 10 (53) Indo-Fijian 8 (31) 6 (21) 8 (20) 11 (20) 10 (38) 6 (32) Other 2 (8) 1 (4) 0 (0) 0 (0) 0 (0) 3 (16) Median no. of children in house hold (IQR) any 2 (2-3) 3 (2-4) 3 (2-4) 3 (2-4) 3 (2-4) 3 (2.5-4) ≤5 y old 0 (0-1) 0.5 (0-1.25) 0 (0-1) 1 (0-1) 1 (0-1) 1 (0-1) Median weekly income* (IQR) 58 (55-103) 91 (55-144) 83 (55-110) 83 (55-110) 61 (33-83) 69 (35-110) Exposure to cigarette smoking 10 (38) 8 (29) 15 (38) 25 (45) 16 (62) 9 (47) Antimicrobial use in prior two weeks# 2 (8) 3 (11) 5 (13) 8 (15) 4 (15) 3(16) Current URTI symptoms Rhinorrhoea 7 (27) 5 (18) 11 (28) 15 (27) 10 (38) 6 (32) Cough 8 (31) 10 (36) 14 (35) 20 (36) 4 (15) 6 (32) Values are numbers (percentages) unless otherwise indicated; IQR, interquartile range; URTI, upper respiratory tract infection; *Median weekly income is in U.S. dollars, with an exchange rate of 0.55 U.S. dollar for 1 Fijian dollar; #Antimicrobial use in the prior two weeks as reported by parent or guardian; No significant differences between groups, p>0.05 as calculated by Fisher’s Exact test.
71
Table 3.3. Demographic characteristics of children in FFS stratified by 23vPPV status or PCV7 dose. Parameter at swab 3 PCV7 2 PCV7 1 PCV7 collection No 23vPPV 23vPPV doses doses doses No. of children 92 102 54 95 45 Male 43 (47) 43 (42) 28 (52) 39 (41) 19 (42) Median age (y) 6.08 5.92 6.21^ 5.92 6.00 Ethnicity iTaukei 64 (70) 75 (74) 37 (69) 76 (80) 26 (58) Indo-Fijian 23 (25) 26 (25) 14 (26) 19 (20) 16 (36) Other 4 (4) 2 (2) 3 (6) 0 (0) 3 (7) Median no. of children in house hold (IQR) any 3 (2-4) 3 (2-4) 3 (2-3.75) 3 (2-4) 3 (2-4) ≤5 y old 0 (0-1) 1 (0-1) 0 (0-1) 1 (0-1) 1 (0-1) Median weekly income* (IQR) 63 (44-110) 83 (55-110) 69 (55-110) 83 (55-110) 61 (33-110) Exposure to cigarette smoking 41 (45) 42 (41) 18 (33) 40 (42) 25 (56) Antimicrobial use in prior two weeks# 11 (12) 14 (14) 5 (9) 13 (14) 7 (16) Current URTI symptoms Rhinorrhoea 28 (30) 26 (25) 12 (22) 26 (27) 16 (36) Cough 26 (28) 36 (35) 18 (33) 34 (36) 10 (22) Values are numbers (percentages) unless otherwise indicated; IQR, interquartile range; URTI, upper respiratory tract infection; *Median weekly income is in U.S. dollars, with an exchange rate of 0.55 U.S. dollar for 1 Fijian dollar; #Antimicrobial use in the prior two weeks as reported by parent or guardian. ^Children that received 3 doses of PCV7 were slightly older than children that received 2 doses of PCV7 (p=0.017) as calculated by Fisher’s Exact test.
72
Table 3.4. Demographic characteristics of children in FFS stratified by ethnicity.
Parameter at swab collection iTaukei Indo-Fijian Other p-value~ No. of children 139 49 6 Male 59 (42) 43 (51) 2 (33) 0.299 Median age (y) 6.25 5.92 6.79 0.204 Median no. of children in house hold (IQR) any 3 (2-4) 2 (2-3) 2 (1.25-2.75) <0.001^ ≤5 y old 1 (0-2) 0 (0-1) 0 (0-0.75) 0.004^ Median weekly income* (IQR) 58 (44-110) 82 (69-118) 193 (55-220) 0.001^ Exposure to cigarette smoking 60 (43) 20 (41) 3 (50) 0.746 Antimicrobial use in prior two weeks 14 (10) 10 (20) 1 (17) 0.062 Current URTI symptoms Rhinorrhoea 39 (26) 13 (27) 3 (50) 0.837 Cough 43 (31) 16 (33) 2 (33) 0.824 Values are numbers (percentages) unless otherwise indicated; IQR, interquartile range; URTI, upper respiratory tract infection; *Median weekly income is in U.S. dollars, with an exchange rate of 0.55 U.S. dollar for 1 Fijian dollar; #Antimicrobial use in the prior two weeks as reported by parent or guardian. ~Comparing iTaukei and Indo-Fijian children only; ^Statistically significant as calculated by Fisher’s Exact test.
73
Table 3.5. Primer and probe sequences and concentrations for quantitative real-time PCR duplex assays for S. pneumoniae and H. influenzae (reaction 1) and M. catarrhalis and S. aureus (reaction 2). Duplex Species, target Primer and probes [5’->3’], (concentration) Reaction gene [ref] F, ACGCAATCTAGCAGATGAAGCA (100 nM); S. pneumoniae, R, TCGTGCGTTTTAATTCCAGCT (100 nM); lytA (237) P, Cy5-TGCCGAAAACGCTTGATACAGGGAG-BHQ3 (200 nM) 1 F, GGTTAAATATGCCGATGGTGTTG (100 nM); H. influenzae, R, TGCATCTTTACGCACGGTGTA (300 nM); hpd (262) P, 6FAM-TTGTGTACACTCCGT”T”GGTAAAAGAACTTGCAC-SpacerC3 (100 nM) F, CGTGTTGACCGTTTTGACTTT (100 nM); M. catarrhalis, R, TAGATTAGGTTACCGCTGACG (100 nM); copB (176) P, Cy5-ACCGACATCAACCCAAGCTTTGG-BHQ3 (100 nM) 2 F, AGTTGCTTAGTGTTAACTTTAGTTGTA (100 nM); S. aureus, nuc R, GCACTATATACTGTTGGATCTTCAGAA (100 nM); (176) P, TxR-TGCATCACAAACAGATAACGGCGTAAATAGAAG-BHQ2 (100 nM) F, forward; R, reverse; P, probe FAM, 6-carboxylfluorescein; BHQ2, black hole quencher 2; TxR, Texas Red; “T”, BHQ-1 (dT)
74
3.4 Additional Analyses After publication of Boelsen et al. (231), we conducted additional analyses to determine the cumulative effect of the pneumococcal polysaccharide vaccine (23vPPV) on NP carriage. To do this, we determined the effect of 23vPPV on pneumococcal carriage at any time point after 23vPPV was administered at 12 months of age (including data from FiPP, pilot FFS, and FFS).
The results of this cumulative analysis show the only significant differences between 23vPPV vaccinated and unvaccinated children were in the Indo-Fijian children, where 23vPPV vaccinated Indo-Fijian children were less likely to carry pneumococci than Indo- Fijian children that were not vaccinated with 23vPPV (Table 3.6). This was consistent with the findings of the FFS presented in Boelsen et al. (231).
Table 3.6. Cumulative carriage of S. pneumoniae in 23vPPV vaccinated and 23vPPV unvaccinated children from any swabs collected after 12 months of age from FiPP, the pilot study and/or the follow-up study participants no 23vPPV 23vPPV OR 95% CI p-value n (%) n (%)
All 240 234 Any pneumococci 135 (56) 124 (53) 0.877 0.611 to 1.259 0.519 PCV7 serotypes 22 (9) 27 (12) 1.292 0.713 to 2.342 0.452 23vPPV serotypes 61 (25) 55 (24) 0.902 0.593 to 1.371 0.670 non-PCV7 23vPPV serotypes 41 (17) 31 (13) 1.349 0.814 to 2.238 0.253
iTaukei 149 152 Any pneumococci 107 (72) 108 (71) 0.964 0.584 to 1.589 0.899 PCV7 serotypes 16 (11) 23 (15) 1.482 0.749 to 2.933 0.304 23vPPV serotypes 50 (34) 48 (32) 0.914 0.564 to 1.480 0.806 non-PCV7 23vPPV serotypes 34 (23) 25 (16) 0.666 0.375 to 1.183 0.192
Indo-Fijian 79 69 Any pneumococci 22 (28) 8 (12) 0.340 0.140 to 0.824 0.015* PCV7 serotypes 5 (6) 2 (3) 0.442 0.083 to 2.354 0.450 23vPPV serotypes 9 (11) 5 (7) 0.608 0.193 to 1.909 0.417 non-PCV7 23vPPV serotypes 4 (5) 3 (4) 0.852 0.184 to 3.950 1.000 p-values calculated using Fisher’s exact test; *, statistically significant; OR, odds ratio; CI, confidence interval.
75
3.5 Further discussion
Immune hyporesponsiveness has been observed in several studies to differing degrees after vaccination with polysaccharide vaccines, most notably with a group C meningococcal polysaccharide vaccine (263, 264). Prior to our FiPP study, there was debate about whether this applied to pneumococcal polysaccharide vaccines. The FiPP study demonstrated that immune hyporesponsiveness was associated with the use of 23vPPV, however the long-term implications on immunity, carriage and health were unknown. In work conducted as part of this thesis, we have shown that 23vPPV does not have a long-term effect on pneumococcal carriage. These findings are consistent with the results from the immunological and clinical arms of the FFS recently published (265).
In the clinical arm of FFS, no differences were observed in pneumococcal-related hospitalisations between 23vPPV vaccinated and unvaccinated children in the 5 years since the conclusion of the FiPP study (265). This is in contrast to observational evidence in Australian children where 23vPPV increased the risk of hospitalisation due to pneumonia (266). The FFS was limited in the number of children enrolled and it is possible that there were too few to observe an increase in hospitalisations, however the clinical findings are supported by the immunological findings (265), and the carriage data presented in this chapter.
For the immune analysis, PCV13 was used to probe immune responses (in a similar way that mPPV had been used in FiPP). The responses to PCV13 were examined by measuring pneumococcal antibodies through ELISA, as well as examining the quality of antibodies (by OPA) and pneumococcal memory B-cell responses. Immune responses were similar between children that who received 23vPPV at 12 months, compared with children who did not, indicating that the immune hyporesponsiveness was not sustained (267).
There are several factors thought to influence the occurrence and duration of immune hyporesponsiveness following polysaccharide vaccination, particularly when used as a booster dose. In general a 23vPPV booster increases antibody levels, however as noted by O’Brien et al. the quality of the antibodies are inconsistent between studies (268). These differences in antibody quality could be due to natural exposure to some serotypes,
76
as well as the age at vaccination. Early studies of meningococcal polysaccharide vaccination showed antibody levels were lower after the administration of a booster than after an initial dose without a booster dose, except for the 10 µg dose (263). This provides some evidence that the size of the dose, the specific serotypes involved, the level of serotype-specific antibodies present prior to vaccination and the time between doses may all affect whether there is immune hyporesponsiveness (268).
The depletion of memory B-cells is thought to play a key role in immune hyporesponsiveness. Clutterbuck et al. (269), found in adults there was a depletion of memory B-cells after vaccination with 23vPPV (compared to pre-vaccination levels) and postulated that this may be a mechanism for hyporesponsiveness in adults. Another study in infants, suggested that there may have been depletion of the memory B-cell pool following 23vPPV and recommended that PCV be used as a booster in the second year of life rather than 23vPPV (270).
However, hyporesponsiveness mediated by prior vaccine type carriage has been observed for PCV7 vaccination, where several studies have reported that carriage of a vaccine serotype prior to PCV vaccination led to serotype-specific hyporesponsiveness for the carried serotype (271, 272). In one study, this effect was only partially overcome at 12 months when a booster dose was administered (271). There was both lower geometric mean concentration and lower proportion of children with concentration ≥ 0.35 µg/ml (considered protective) for the carried serotypes. Serotype-specific hyporesponsiveness was also observed in the FiPP study for children carrying serotypes 19F and 23F prior to 23vPPV vaccination (273).
3.6 Conclusions Here, we determined whether or not 23vPPV vaccination, and the subsequent immune hyporesponsiveness observed in children that received it, would increase risk of pneumococcal NP carriage in the long-term. Overall, we found no differences in the pneumococcal carriage rate in children aged 5-7 years who did or did not receive 23vPPV at 12 months of age. However, after stratifying by ethnicity in vaccinated and unvaccinated children, it was apparent that 23vPPV reduced carriage of pneumococci in Indo-Fijian children, and was also associated with higher carriage of S. aureus in iTaukei
77
children. While the reasons underpinning the differences between ethnic groups are unclear, our results do not support the use of 23vPPV as a booster in young children.
78
Chapter 4
The effects of pneumococcal vaccination on the nasopharyngeal
microbiome of children in Fiji
79
4 The effects of pneumococcal vaccination on the nasopharyngeal microbiome of children in Fiji
4.1 Introduction and aims In chapter 3 we found that the effects of the 23-valent polysaccharide vaccine on carriage of potential pathogens within the nasopharynx varied between the two main ethnic groups in Fiji. This chapter further investigates findings from chapter 3, taking a broader approach to investigating the impact of pneumococcal vaccination Fiji by looking at NP microbiota, and instead focussing on the pneumococcal conjugate vaccine.
Study of the bacterial ecosystems within our body (i.e. the human microbiome) has increased significantly over the past decade. Improvements in high-throughput sequencing, data analysis tools and lower costs have enabled researchers to move beyond culture-dependent techniques, which have biased our understanding of bacterial communities. Much of the human microbiome research has focused on gastrointestinal tract (274) where gut microbiota have been linked with diseases such as cancer and health issues such as obesity. Researchers have also been able to define a ‘healthy’ gut microbiome versus one in dysbiosis, which has led to development of therapeutic treatments to improve gut health.
One such treatment has been developed for Clostridium difficile infection, where treatment with antibiotics is not always effective and can in turn lead to recurrent infection. Transplantation of faecal matter from individuals with healthy gut microbiota to patients with C. difficile infections was more effective at resolving the infection than the standard treatment with vancomycin (275). This study highlights the importance and the value of microbiome research to improve health outcomes.
To date, the study of bacterial communities in the nasopharynx has mostly been limited to a small number of common respiratory pathogens. However, there are a growing number of studies focused on the NP microbiome examining factors that affect the microbiome (such as age, seasonality, nutrition and exposure to cigarette smoke) (276- 281), associations with diseases or viral infections (212, 282-286) and the effects of vaccination (211, 213) on the microbiome.
80
Pneumococcal vaccines have been effective at reducing pneumococcal disease and carriage but several studies, including our findings in chapter 3, have also highlighted a potential for pneumococcal vaccination to effect disease and carriage of other respiratory pathogens in the nasopharynx. A microbiome approach has been used in a small number of studies to explore whether pneumococcal vaccination could have a broader impact than previously understood. Biesbroek et al. (211) found that PCV7 altered the bacterial composition and increased the bacterial diversity in the nasopharynx of children in the Netherlands. Vaccinated children had increased absolute abundance of Haemophilus and Staphylococcus spp., as well as Actinomyces, Rothia, Neisseria and Veillonella spp., which have been associated with an increased risk of otitis media. Hilty et al. (212) examined the effects of PCV7 in Swiss children with otitis media and found that PCV7 reduced commensals such as Streptococcaeae (excluding pneumococci) and Corynebacteriaeae, suggesting that vaccination may have a negative impact on health particularly as alpha-haemolytic streptococci can have protective effects against otitis media. In contrast with those two studies, Feazel et al. found that PCV10 had no effect on the NP microbiome in Kenyan children. (213). The study examined both unvaccinated and vaccinated children before, and 180 days after, vaccination and found limited or no species replacement had occurred in these children, with time having a much greater impact on microbial composition.
Carriage rates of bacterial species can vary by geographic region and socio-economic status (85, 89). Low-middle income countries often have higher rates of pneumococcal colonisation, suggesting that the impact of vaccination (on pneumococcal carriage and on other species within the nasopharynx) could be different in these settings compared with developed countries. Hence, there is a need for more data from a range of geographic and economic regions.
In Fiji, carriage rates of S. pneumoniae, S. aureus, H. influenzae and M. catarrhalis in children vary between the two main ethnicities, indigenous Fijians and Fijians of Indian descent, suggesting they have distinct microbial profiles (chapter 3 and (176)). In addition, the 23-valent polysaccharide vaccine (23vPPV) had a different effect on carriage of S. pneumoniae and S. aureus between the two ethnic groups (chapter 3). This
81
suggests that there may be a potential for pneumococcal vaccination to have a different effect on the microbiome depending on the ethnicity of the child.
As such, the aim of this chapter is to investigate the wider effects of PCV7 in Fijian children, including for each ethnic group, using samples collected as part of the Fiji Pneumococcal Project (FiPP) described in chapter 3.
82
4.2 Results For the microbiome analysis, we randomly selected 36 NP swabs from children aged 12 months from each of the four groups (unvaccinated iTaukei children, vaccinated [3 doses of PCV7] iTaukei children, unvaccinated Indo-Fijian children and vaccinated Indo-Fijian children) from the FiPP study. Overall, 144 samples, 4 sample repeats, 7 extraction controls and 5 PCR (no template) controls were sequenced resulting in a total of 62,038,435 sequence reads and an average of 387,740 reads per sample. Following sequence processing 59% (36,344,642 total; 227,154 per sample) of these sequence reads remained for analysis. Twelve of the 144 samples were excluded from further analysis following quality control, because they shared greater than 20% sequence read similarity with controls (n=11) or had fewer than 50,000 sequence reads (n=1), leaving 132 samples for analysis.
We identified 4,036 OTUs in total and 847 OTUs with unique taxonomic names, though many (n=1037) could only be classified to the kingdom or phylum level. Before calculating richness, rarefaction was performed to account for differences in sequencing depth. The effect of rarefaction on four samples which were repeated in different sequencing runs was examined (Table 4.1). Rarefaction appeared to result in greater consistency in richness for samples that had been sequenced at different depths.
Table 4.1. Effect of sequencing depth on richness, before and after rarefaction. Sequence Richness Richness Sample reads (n) (raw) (rarefied) FM001a 610365 135 56 FM001b 199164 72 42 FM002a 556999 321 193 FM002b 189225 253 190 FM003a 580294 176 105 FM003b 205993 145 103 FM004a 622693 52 23 FM004b 219908 29 22
83
After rarefaction, OTU richness in the 132 samples ranged from 15 OTUs to 204 OTUs with an average richness of 79 OTUs. The Shannon diversity index ranged from 0.06 to 2.41 with an average of 1.15.
4.2.1 Participant characteristics and risk factor analysis
4.2.1.1 Participant characteristics Overall, between the four groups – unvaccinated and vaccinated, iTaukei and Indo-Fijian children – there were few differences in participant characteristics (Table 4.2). There was a significant difference in mean weight between the four groups (p<0.001, one-way ANOVA). Individually, there were also differences in gender between Indo-Fijian children that were unvaccinated (38% male) and Indo-Fijian children that were vaccinated (55% male) (p=0.026); differences in the median age breastfeeding was stopped in unvaccinated iTaukei (median 30 weeks) and unvaccinated Indo-Fijian children (median 9 weeks) (p=0.045); and differences in exposure to cigarette smoke between unvaccinated iTaukei children (64%) and Indo-Fijian children (38% for unvaccinated and 36% for vaccinated children) (p=0.048). Pooled participant characteristics by both vaccination status (Table 4.3) and ethnicity (Table 4.4) were also considered. The only significant difference between vaccinated and unvaccinated children was the proportion of children breastfeeding: 74% of unvaccinated children were breastfeeding compared with 54% of vaccinated children (p=0.019). Both mean weight (p<0.001) and exposure to cigarette smoke (p=0.038) differed between iTaukei and Indo- Fijian children. Mean weight was higher in iTaukei children (5023 g) compared with Indo-Fijian children (4385 g) and more iTaukei children were exposed to cigarette smoke (55%) than Indo-Fijian children (37%).
84
Table 4.2. Participant characteristics for children included analysis by vaccination status and ethnicity. Unvaccinated Vaccinated Parameter at swab collection iTaukei Indo-Fijian iTaukei Indo-Fijian No. of children 33 32 34 33
Male (%) 18 (55) 12 (38) 19 (56) 22 (67)
Swabs collected in wet season [vs. dry] (%) 25 (76) 25 (78) 28 (82) 20 (61)
Median age [mo] (IQR) 12 (0) 12 (0) 12 (0) 12 (0)
Mean weight [g]# (st dev) 4998 (584) 4248 (547) 5084 (579) 4566 (444)
Breastfeeding (%) 25 (76) 22 (69) 19 (56) 17 (52)
Median age breastfeeding stopped [wks]^ (IQR) 30 (21) 9 (32) 28 (20) 27 (37)
Exposure to cigarette smoking (%) 21 (64) 12 (38) 16 (47) 12 (36)
Antimicrobial in prior 2 wks (%) 4 (12) 4 (13) 5 (15) 0 (0)
Symptoms of current URTI [any] (%) 11 (33) 8 (25) 14 (41) 6 (18)
Runny nose (%) 10 (30) 6 (19) 9 (26) 5 (15)
Cough (%) 7 (21) 4 (13) 10 (29) 5 (15)
allergic rhinitis (%) 0 (0) 0 (0) 1 (3) 0 (0)
ear infection (%) 0 (0) 0 (0) 0 (0) 0 (0)
ear discharge (%) 0 (0) 0 (0) 0 (0) 0 (0) #Weight data incomplete for vaccinated children – 13/34 available for iTaukei children 24/33 available for Indo-Fijian children; ^Data only included for children that had stopped breastfeeding. Significant differences (p<0.05) are shown in bold.
85
Table 4.3. Participant characteristics for children included analysis by vaccination status.
Parameter at swab collection Unvaccinated Vaccinated p-value
No. of children 65 67 Male (%) 30 (46) 41 (61) 0.116 swabs collected in wet season [vs. dry] (%) 50 (77) 48 (72) 0.553 Median age [mo] (IQR) 12 (0) 12 (0) 1.000 Median weight [g] (st dev) 4627 (679) 4748 (548) 0.356 Breastfeeding (%) 48 (74) 36 (54) 0.019* Median age breastfeeding stopped [wks] (IQR) 26 (27) 28 (23) 0.406 Exposure to cigarette smoking (%) 33 (51) 28 (42) 0.383 Antimicrobial in prior 2 wks (%) 8 (12) 5 (7) 0.394 Current URTI [any] (%) 19 (29) 20 (30) 1.000 Runny nose (%) 16 (25) 14 (21) 0.680
Cough (%) 11 (17) 15 (22) 0.514
allergic rhinitis (%) 0 (0) 1 (1) -
ear infection 0 (0) 0 (0) -
ear discharge 0 (0) 0 (0) - *Statistically significant, p<0.05. Significant differences are shown in bold.
86
Table 4.4. Participant characteristics for children included analysis stratified by ethnicity.
Parameter at swab collection iTaukei Indo-Fijian p-value
No. of children 67 65 Male (%) 37 (55) 34 (52) 0.862 swabs collected in wet season [vs. dry] (%) 53 (79) 45 (69) 0.234 Median age [mo] (%) 12 (0) 12 (0) 1.000 Mean weight [g] (st dev) 5023 (578) 4385 (526) <0.001* Breastfeeding (%) 45 (67) 39 (60) 0.470 Median age breastfeeding stopped [wks](IQR) 28 (19) 23 (27.75) 0.105 Exposure to cigarette smoking (%) 37 (55) 24 (37) 0.038* Antimicrobial in prior 2 wks (%) 9 (13) 4 (6) 0.243 Current URTI [any] (%) 25 (37) 14 (22) 0.057 Runny nose (%) 19 (28) 11 (17) 0.147 Cough (%) 12 (18) 9 (14) 0.636 Allergic rhinitis (%) 1 (1) 0 (0) - Ear infection (%) 0 (0) 0 (0) - Ear discharge (%) 0 (0) 0 (0) - *Statistically significant, p<0.05. Significant differences are shown in bold.
4.2.2 Alpha diversity
OTU richness and Shannon diversity were measured for unvaccinated and vaccinated iTaukei and Indo-Fijian children (Figure 4.1). There were no significant differences in OTU richness between the four groups (Figure 4.1.A). However unvaccinated iTaukei children had higher Shannon diversity (Figure 4.1.B), and therefore evenness of OTU distribution, than Indo-Fijian children regardless of whether they were unvaccinated (median Shannon diversity index of 1.36 vs. 1.01, p=0.007) or vaccinated (median Shannon diversity index of 1.36 vs. 1.00, p=0.003). Richness and Shannon diversity were also measured pooling by vaccination status and ethnicity (Figure 4.2). Vaccination did not appear to have an effect on OTU richness (Figure 4.2.A) or Shannon diversity (Figure 4.2.B). While the ethnicity of a child did not influence OTU richness (Figure 4.2.C), it did have an influence on Shannon diversity (Figure 4.2.D). OTUs were more evenly distributed in iTaukei children compared with Indo-Fijian children (median Shannon diversity index of 1.33 vs. 1.07, p=0.001).
87
A
) 250 s U T
O 200 f o
r e
b 150 m u ( n
100 s s e n
h 50 c i R 0
ed ed ed ed at at at at n n n n
vacci vacci vacci vacci n n n u u a kei n ji a i kei au ji -F iT i o au -F d iT o n d I In p = 0.007 B 3 p = 0.003 x e d n i
y t i 2 s r e v i d
n o
n 1 n a h S
0 d d d d te te te te a a a a in in in in c c c c c c c c a a a a v v v v n i n n u e u a i k n ji e u a i k a ji -F u iT i o a -F d iT o n d I n I
Figure 4.1. Richness (A) and Shannon diversity (B) by vaccination status and ethnicity. Error bars are median and interquartile range; there were significant differences in Shannon diversity between iTaukei unvaccinated children and Indo-Fijian children (unvaccinated, p=0.007; vaccinated, p=0.003) as calculated using the Mann-Whitney test.
88
A B ) s 250 x 3 e U d T n O i
f 200 y t o i r 2 e r s b 150 e v m i u d
n n
( 100
s 1 n o s n e a n 50 h h S i c
R 0 0 d d d d te te te te a a in in ina ina c c c c c c c c a a a a v v v n nv u u C D p = 0.001 )
x 3 s
250 e U d T i n O
f 200 y o i t
s r 2 r e e b 150 v i m u d n n (
100 o 1 s n s n e a
n 50 h h S c i
R 0 0 i i n e n e a k ia k ji u ij u i a a -F -F iT o iT o d d In In Figure 4.2. Richness and Shannon diversity by vaccination status (A and B) and by ethnicity (C and D). Error bars are median and interquartile range; there was a significant different in Shannon diversity between iTaukei and Indo-Fijian children (p=0.001).
89
4.2.3 Beta diversity
Beta-diversity, or between sample diversity, was examined to determine the dissimilarity in microbial composition using nMDS plots and PERMANOVA. Differences in microbial composition were examined by vaccination status and ethnicity (Figure 4.3). There were differences in the microbial composition between unvaccinated and vaccinated, iTaukei and Indo-Fijian children (p<0.001). To determine whether the effect of vaccination may vary by ethnic group, samples were stratified by ethnicity and the effect of vaccination was considered for both iTaukei (Figure 4.4.A) and Indo-Fijian (Figure 4.4.B) children. For iTaukei children, vaccination did not appear to have a significant impact on microbial composition (p=0.375), though there was some separation of the means on the nMDS plot suggesting that there might be minor differences in microbial composition. For Indo-Fijian children, vaccination appeared to have some impact on microbial composition (p=0.050). Overall vaccination appeared to have little impact on the microbial composition (Figure 4.5.A) of the nasopharynx (p=0.668). The microbial composition of the nasopharynx was significantly different (p<0.001) between iTaukei and Indo-Fijian children (Figure 4.5.B).
90
p<0.001
Figure 4.3. nMDS of the dissimilarity in microbial composition by vaccination status and ethnicity. Shown are iTaukei unvaccinated (green) and vaccinated (blue) children, and Indo-Fijian unvaccinated (orange) and vaccinated (red) children. The centre points represent the mean of each group and the lines represent distance from the mean for each sample. There was significant difference between the four groups (p<0.001), p-value calculated using PERMANOVA.
91
A
p=0.375
B
p=0.050
Figure 4.4. nMDS of the dissimilarity in microbial composition by vaccination status in iTaukei (A) and Indo-Fijian children (B). Shown are iTaukei unvaccinated (green) and vaccinated (blue) children, and Indo-Fijian unvaccinated (orange) and vaccinated (red) children. The centre points represent the mean of each group and the lines represent distance from the mean for each sample. There was some evidence of a significant difference by vaccination status for Indo-Fijian children (p=0.050) but not iTaukei children (p=0.375), p-value calculated using PERMANOVA.
92
A
p=0.668
B
p<0.001
Figure 4.5. nMDS of the dissimilarity in microbial composition by vaccination status (A) and by ethnicity (B). Shown are unvaccinated children (blue) and vaccinated children (red); and iTaukei children (blue) and Indo-Fijian children (red). The centre points represent the mean of each group and the lines represent distance from the mean for each sample. There was a significant difference by ethnicity (p<0.001) but not by vaccination status, p-value calculated using PERMANOVA.
93
4.2.4 Microbial composition
Having determined that there were some differences in microbial composition, we next looked at what those differences were by examining the relative abundance of individual OTUs and the data pooled by family or genus.
The most abundant OTUs (Table 4.5) were of the genera Dolosigranulum (family Carnobacteriaceae, mean relative abundance 28%), Moraxella (family Moraxellaceae, mean relative abundance 18%) and Corynebacterium (family Corynebacteriaceae, mean relative abundance 16%). Within the 10 most abundant OTUs there were two Moraxella OTUs and two Corynebacterium OTUs and single OTUs from Dolosigranulum, Pseudomonas, Haemophilus, Streptococcus, Staphylococcus and one Flavobacteriaceae OTU that could only be classified to the family level.
Only the Dolosigranulum OTU and Streptococcus OTU (Table 4.5) were present in every sample, although the Streptococcus OTU was only present at a relative abundance higher than 1% in 36% of samples whereas the Dolosigranulum was present at a relative abundance higher than 1% in 90% of the samples.
4.2.4.1 Relative abundance pooled by Family We pooled all OTUs to the family level and examined the relative abundance of the seven most common families (Figure 4.6), which accounted for at least 90% of all sequence reads in 119 out of the 132 samples. For Staphylococcaceae and Streptococcaceae, each family had only one sample with greater than 50% relative abundance, as such Staphylococcaceae and Streptococcaceae were rarely the dominant family. Corynebacteriaceae and Pasteurellaceae were slightly more likely to dominate a sample whereas Pseudomonadaceae, Moraxellaceae and Carnobacteriaceae were often the dominant family with over 20 samples having greater than 50% relative abundance for each family.
We next examined the pooled family data by ethnicity and vaccination status (Figure 4.7), specifically analysing differences between unvaccinated and vaccinated children within each ethnic group. As the data were skewed we used the Mann-Whitney test to compare the relative abundances of each of the seven most common families.
94
Table 4.5. Summary of the 10 most abundant OTUs.
Mean (median) relative Prevalence >1% relative OTU Taxonomy abundance (%)* n (%) abundance n (%) 0001 Firmicutes (P) 28.05 (26.87) 132 (100) 119 (90) Bacilli (C) Lactobacillales (O) Carnobacteriaceae (F) Dolosigranulum (G) 0002 Proteobacteria (P) 17.66 (7.56) 130 (98) 82 (62) Gammaproteobacteria (C) Pseudomonadales (O) Moraxellaceae (F) Moraxella (G) 0003 Proteobacteria (P) 13.03 (0.04) 129 (98) 23 (17) Gammaproteobacteria (C) Pseudomonadales (O) Pseudomonadaceae (F) Pseudomonas (G) 0004 Actinobacteria (P) 15.57 (9.27) 131 (99) 106 (80) Actinobacteria (C) Corynebacteriales (O) Corynebacteriaceae (F) Corynebacterium (G) 0005 Proteobacteria (P) 8.07 (0.09) 130 (98) 45 (34) Gammaproteobacteria (C) Pasteurellales (O) Pasteurellaceae (F) Haemophilus (G) 0006 Proteobacteria (P) 5.80 (0.02) 128 (97) 47 (36) Gammaproteobacteria (C) Pseudomonadales (O) Moraxellaceae (F) Moraxella (G) 0007 Firmicutes (P) 4.14 (0.39) 132 (100) 48 (36) Bacilli (C) Lactobacillales (O) Streptococcaceae (F) Streptococcus (G) 0009 Firmicutes (P) 1.06 (0.04) 131 (99) 10 (8) Bacilli (C) Bacillales (O) Staphylococcaceae (F) Staphylococcus (G) 0010 Actinobacteria (P) 1.31 (0.01) 125 (95) 22 (17) Actinobacteria (C) Corynebacteriales (O) Corynebacteriaceae (F) Corynebacterium (G) 0012 Bacteroidetes (P) 1.33 (0.002) 88 (67) 17 (13) Flavobacteriia (C) Flavobacteriales (O) Flavobacteriaceae (F) *Mean and median relative abundance across all samples; P, phylum; C, class; O, order; F, family; G, genus
95
) 100 % (
e 80 Other c
n Pasteurellaceae a
d 60 Corynebacteriaceae n
u Streptococcaceae b
a 40 Carnobacteriaceae
e Staphylococcaceae v i t 20 Moraxellaceae a l
e Pseudomonadaceae r 0
Samples
Figure 4.6. Relative abundance (%) of OTUs in all samples pooled to family level. Samples are sorted by clusters shown in Figure 4.12 (page 107).
96
A Pseudomonadaceae B Moraxellaceae
20 60 ) 15 ) % % ( ( 50
10 e 5 e c c 40 n n a 0.10 a d d 30 n 0.08 n u u b b 20 a a
0.06 e e 10 v v i i t 0.04 t a a l l 1.0 e e r 0.02 r 0.5 0.00 0.0 C Staphylococcaceae D Carnobacteriaceae p = 0.0449 0.3 60 ) ) % % ( (
50 e e c c n n a 0.2 a 40 d d n n u u 30 b b a a
e e v 0.1 v 20 i i t t a a l l e e 10 r r 0.0 0 E Streptococcaceae F Corynebacteriaceae
10 p = 0.0245 50 ) ) % % ( (
e 8 e 40 c c n n a a d
6 d
n 30 n u u b b a
a
4
e 20 e v i v t i t a l 2 a l e 10 r e r 0 0 G Pasteurellaceae Pseudomonadaceae ) 10
% Moraxellaceae ( 8 e Staphylococcaceae c
n 6 a Carnobacteriaceae d
n 4 Streptococcaceae u
b 2 a Corynebacteriaceae
e v
i 0.10 Pasteurellaceae t a l
e 0.05 r 0.00 Figure 4.7. Median relative abundance (%) for the seven most common families (A-G) by vaccination status and ethnicity. Error bars are interquartile range; p-values comparing unvaccinated and vaccinated children within each ethnic group calculated using Mann-Whitney test.
97
Within both ethnic groups there were no differences between unvaccinated and vaccinated children for the families Pseudomonadaceae, Moraxellaceae, Staphylococcaceae, Corynebacteriaceae and Pasteurellaceae. For iTaukei children, unvaccinated children had higher relative abundance of Streptococcaceae compared with vaccinated children (median relative abundance of 2.58% vs. 0.51%, respectively; p=0.025). There was no difference in Streptococcaceae by vaccination status for Indo- Fijian children, however unvaccinated Indo-Fijian children had lower relative abundance of Carnobacteriaceae compared with vaccinated Indo-Fijian children (median relative abundance of 29.25% vs. 44.59%, respectively; p=0.045). The difference between unvaccinated and vaccinated children for Carnobacteriaceae was not present in iTaukei children.
We next analysed the pooled family data examining overall differences by vaccination status (Figure 4.8.A) and looking at differences between the two ethnic groups (Figure 4.8.B). Considering an overall effect of PCV7 vaccination, there were no differences in any of the seven most common families between unvaccinated and vaccinated children. There were several families that were significantly different between the two ethnic groups: compared with Indo-Fijian children, iTaukei children had higher relative abundance of Moraxellaceae (median relative abundance of 0.32% vs. 29.35%, respectively; p<0.001) and Pasteurellaceae (median relative abundance of 0.03% vs. 0.59%, respectively; p=0.005); and lower relative abundance of Staphylococcaceae (median relative abundance of 0.06% vs. 0.02%, respectively; p=0.002), Carnobacteriaceae (median relative abundance of 36.42% vs. 18.18%, respectively; p=0.001) and Corynebacteriaceae (median relative abundance of 17.92% vs. 4.83%, respectively; p= 0.001).
98
A
) 60 Pseudomonadaceae % (
Moraxellaceae
e 50
c Staphylococcaceae
n 40 a Carnobacteriaceae d 30 n Streptococcaceae u
b 20 Corynebacteriaceae a
e 10 Pasteurellaceae v i t
a 0.10 l
e 0.05 r 0.00
B ) 60 p=0.001 Pseudomonadaceae %
( P<0.001 p=0.001 Moraxellaceae
e 50
c Staphylococcaceae n 40 a Carnobacteriaceae d 30 n Streptococcaceae u p=0.005 b 20 Corynebacteriaceae a p=0.002
e 10 Pasteurellaceae v i t
a 0.10 l
e 0.05 r 0.00
Figure 4.8. Median relative abundance (%) for the seven most common families by vaccination status (A) and by ethnicity (B). Error bars are interquartile range, p-values calculated using Mann-Whitney test.
99
4.2.4.2 Relative abundance pooled by Genus Due to the confidence level with which OTUs are classified and potential bias in classification to the genus level, often analyses are only done to the family level. However, for the most common families there was one predominate genus (Table 4.6) and therefore a genus level analysis was done to provide greater detail on composition without losing confidence in the results.
Table 4.6. Percentage of the dominant genus in each of the most common families. Mean (median) % Family Genus genus in family Pseudomonadaceae Pseudomonas 99.2 (100) Moraxellaceae Moraxella 85.3 (99.9) Staphylococcaceae Staphylococcus 98.6 (100) Carnobacteriaceae Dolosigranulum 100 (100) Streptococcaceae Streptococcus 95.3 (99.6) Corynebacteriaceae Corynebacterium 99.9 (100) Pasteurellaceae Haemophilus 99.2 (100)
We pooled all OTUs to the genus level and examined the relative abundance of the seven most common genera (Figure 4.6). Similar to the family level analysis, the most common genera accounted for at least 90% of all sequence reads in 118 out of the 132 (89%) of samples.
Staphylococcus and Streptococcus were rarely the dominant genus in these samples, despite the latter being present in every sample. Instead, the samples were dominated by Pseudomonas, Moraxella and Dolosigranulum.
We next examined the pooled genus data by ethnicity and vaccination status (Figure 4.10), specifically analysing differences between unvaccinated and vaccinated children within each ethnic group. As the data were skewed we used the Mann-Whitney test to compare the relative abundances of each of the seven most common genera.
100
) 100 % (
Other e 80
c Haemophilus
n Corynebacterium a Streptococcus d 60
n Dolosigranulum
u Staphylococcus b 40 a Moraxella
e Pseudomonas v i
t 20 a l e r 0 Samples
Figure 4.9. Relative abundance (%) of OTUs in all samples pooled to genus level. Samples are sorted by clusters shown in Figure 4.12 (page 107).
101
A Pseudomonas B Moraxella 20 60 ) ) % 15 % ( ( 50
e e c 10 c 40 n n a 5 a d d 30 n n u u b 0.10 b 20 a a
0.08 e e 10 v v i i t 0.06 t a a l 0.04 l 1.0 e e r 0.02 r 0.5 0.00 0.0 C Staphylococcus D Dolosigranulum 0.3 60 p = 0.045 ) ) % % ( (
50 e e c c n n a 0.2 a 40 d d n n u u 30 b b a a
e e v 0.1 v 20 i i t t a a l l e e 10 r r 0.0 0 E Streptococcus F Corynebacterium p = 0.025 8 50 ) ) % % ( (
e e 40 c c 6 n n a a d d 30 n n u u 4 b b a a
20 e e v v i i t t a
a 2 l l 10 e e r r 0 0 G Haemophilus
) 10 % ( 8 e c
n 6 a d
n 4 u
b 2 a
e v
i 0.10 t a l
e 0.05 r 0.00 Figure 4.10. Median relative abundance (%) for the seven most common genera (A-G) by vaccination status and ethnicity. Error bars are interquartile range; p-values comparing unvaccinated and vaccinated children within each ethnic group calculated using Mann- Whitney test.
102
Similar to the family level analysis, within both ethnic groups there were no differences between unvaccinated and vaccinated children for the genera Pseudomonas, Moraxella, Staphylococcus, Corynebacterium and Haemophilus. For iTaukei children, unvaccinated children had higher relative abundance of Streptococcus compared with vaccinated children (median relative abundance of 2.02% vs. 0.52%, respectively; p=0.025). There was no difference in Streptococcus by vaccination status for Indo-Fijian children, however unvaccinated Indo-Fijian children had lower relative abundance of Dolosigranulum compared with vaccinated Indo-Fijian children (median relative abundance of 29.25% vs. 44.59%, respectively; p=0.045). The difference between unvaccinated and vaccinated children for Dolosigranulum was not present in iTaukei children.
We next analysed the pooled genus data examining overall differences by vaccination status (Figure 4.11.A) and looking at differences between the two ethnic groups (Figure 4.11.B). Overall, pneumococcal vaccination did not appear to have an impact on the relative abundance of the most common genera. There were several genera that were present in different relative abundances between the two ethnic groups. Compared with Indo-Fijian children, iTaukei children had higher relative abundance of Moraxella (p<0.001) and Haemophilus (p=0.004), and lower relative abundance of Staphylococcus (p=0.002), Dolosigranulum (p=0.001) and Corynebacterium (p=0.001).
103
A
60 ) %
( 50
e
c 40 Pseudomonas n
a Moraxella
d 30
n Staphylococcus u
b 20 Dolosigranulum a
e 10 Streptococcus v i
t Corynebacterium a l 0.10 Haemophilus e r 0.05 unvaccinated 0.00 vaccinated
B
60 p<0.001 p=0.001
) 50 p=0.001 % (
e Pseudomonas
c 40 n Moraxella a
d 30 Staphylococcus n
u p=0.004 b 20 Dolosigranulum a p=0.002 Streptococcus e
v 10 i Corynebacterium t a l Haemophilus e
r 0.10 iTaukei 0.05 Indo-Fijian 0.00 Figure 4.11. Median relative abundance (%) for the seven most common genera by vaccination status (A) and by ethnicity (B). Error bars are interquartile range, p-values calculated using Mann-Whitney test.
104
4.2.5 Heatmap and sample clusters
We next looked at microbial composition in the samples by generating a heatmap and clustering samples based on their composition using a Jaccard distance matrix (Figure 4.12). Most samples were dominated by one or two taxa with the following clusters identified by their dominant taxa (Figure 4.12): Cluster 1 (shown in orange) dominated by Dolosigranulum and Corynebacterium; Cluster 2 (shown in yellow) dominated by Dolosigranulum and Moraxella; Cluster 3 (shown in purple) dominated by Moraxella and Dolosigranulum; Cluster 4 (shown in green) dominated by Haemophilus; and Cluster 6 (shown in grey) dominated by Pseudomonas. Cluster 5 (shown in blue) contained several samples that did not fit into any of the other clusters, with two samples dominated by Streptococcus, two samples dominated by Staphylococcus and one sample dominated by another Corynebacterium. The most prevalent clusters were Cluster 3 (n=35 samples) and Cluster 1 (n=34 samples) with the least common clusters being Cluster 5 (n=5 samples) and Cluster 4 (n=16 samples).
The clusters were also examined by nMDS plot (Figure 4.13) as another way to visualise the differences in clusters by microbial composition. Of note, the mixed nature of Cluster 5 was shown by the spread of the samples within the cluster, and an effect of Pseudomonas on overall composition was evident by the distance of this cluster to the other clusters.
Associations between clusters and risk factors were considered (see Appendix B. Additional data for chapter, Table 7.1), with several clusters associated with ethnicity, symptoms of an upper respiratory tract infection and household exposure to cigarette smoke. Compared with iTaukei children, Indo-Fijian children more likely to be associated with Cluster 1 (38% Indo-Fijian children vs. 13% iTaukei children, p=0.001) and less likely to be associated with Cluster 3 (17% Indo-Fijian children vs. 36% iTaukei children, p=0.018). Cluster 1 was associated with a higher proportion of children without symptoms of an upper respiratory tract infection (31% no URTI children vs. 13% URTI children, p=0.030) and Cluster 3 was associated with a higher proportion of children exposed to household cigarette smoke (36% children exposed to household cigarette smoke vs. 18% children not exposed to household cigarette smoke, p=0.029).
105
106
Figure 4.12. Relative abundance (%) heatmap showing all 132 samples. Sample clustering is shown on the left-hand side. Taxa with a maximum relative abundance of 30% (across all samples) are shown at the bottom. Clusters for samples have been coloured to show clustered groups. The vaccination status (unvaccinated – green; vaccinated – orange), ethnicity (iTaukei – dark blue; Indo-Fijian – red) and whether the child had any symptoms of an upper respiratory tract infection (no URTI – purple; URTI – light blue) are shown in bars on the left of the heatmap.
107
Figure 4.13. nMDS of the dissimilarity in microbial composition by clusters. Shown are Clusters 1-6. The centre points represent the mean of each group and the lines represent distance from the mean for each sample. There was a significant difference between the clusters (p<0.001), p-value calculated using PERMANOVA.
108
4.2.6 Correlation Network
We examined the relationships between the 100 most common OTUs using the network inference tool, SparCC (Figure 4.14). Of particular note for the correlation network, the predominant Streptococcus OTU had positive relationships with two Moraxella OTUs (r=0.32, p<0.001 an r=0.34, P<0.001) and the predominant Haemophilus OTU (r=0.36, p<0.001); and a negative relationship with Pseudomonas (r=0.31, p<0.001). Also of interest, given the difference between unvaccinated and vaccinated Indo-Fijian children for Dolosigranulum, there was no significant interaction between Streptococcus and Dolosigranulum (r=0.03, p=0.379).
109
110
Figure 4.14. Network correlation map based on the 100 most common OTUs considering samples from all children. Shown are SparCC correlations with p-value<0.05 and r>0.3. The main OTUs from the top seven genera are shown by larger node size. Colours of nodes are grey except for the seven most common genera. The edge (connecting line) colour represents a positive (green solid lines) or negative (red dashed lines) correlation between two OTUs, with more transparent lines representing weaker correlations. Nodes are labelled with the OTU taxonomic name
111
4.2.7 Species-specific qPCR
As the 16S sequencing analysis is limited to genus level, we examined carriage rates and densities by conducting species-specific qPCR of the common respiratory pathogens, including S. pneumoniae, H. influenzae, M. catarrhalis and S. aureus.
4.2.7.1 Carriage rates Overall, of the 132 children tested, 63 (48%) were carrying S. pneumoniae, 64 (48%) were carrying H. influenzae, 85 (64%) were carrying M. catarrhalis and 7 (5%) were carrying S. aureus. We examined carriage rates within each ethnic group for S. pneumoniae, H. influenzae, M. catarrhalis and S. aureus comparing unvaccinated and vaccinated children (Table 4.7).
Table 4.7. Comparison of carriage rates in unvaccinated and vaccinated children after stratifying by ethnicity using species-specific qPCR data. iTaukei Indo-Fijian Unvaccinated Vaccinated Unvaccinated Vaccinated p-value p-value (n=33) (n=34) (n=32) (n=33) S. pneumoniae 23 (70) 23 (68) 1.000 8 (25) 9 (27) 1.000 H. influenzae 24 (73) 22 (65) 0.600 9 (28) 9 (27) 1.000 M. catarrhalis 32 (97) 29 (85) 0.197 11 (34) 13 (39) 0.798 S. aureus 1 (3) 2 (6) 1.000 2 (6) 2 (6) 1.000 Values are numbers (percentages); p-values calculated using Fisher’s exact test.
Within each ethnic group there were no significant differences by vaccination status for any of the four species. This included carriage rates of S. pneumoniae in iTaukei children, where there was evidence that the relative abundance of Streptococcus was higher in unvaccinated children compared with vaccinated children (page 102). We also examined carriage rates by vaccination status (Table 4.8) and ethnicity (Table 4.9) individually.
112
Table 4.8. Comparison of carriage rates in unvaccinated and vaccinated children using species-specific qPCR. Unvaccinated Vaccinated p-value (n=65) (n=67) S. pneumoniae 31 (48) 32 (48) 1.000 H. influenzae 33 (51) 31 (46) 0.723 M. catarrhalis 43 (66) 42 (63) 0.719 S. aureus 3 (5) 4 (6) 0.680 Values are numbers (percentages); p-values calculated using Fisher’s exact test.
Overall, there were no differences in carriage rates of the four species by vaccination status but there were differences between the two ethnic groups. Specifically, iTaukei children had significantly higher carriage rates of S. pneumoniae, H. influenzae and M. catarrhalis.
Table 4.9. Comparison of carriage rates in iTaukei and Indo-Fijian children using species- specific qPCR data. iTaukei Indo-Fijian p-value (n=67) (n=65) S. pneumoniae 46 (69) 17 (26) <0.001 H. influenzae 46 (69) 18 (28) <0.001 M. catarrhalis 61 (91) 24 (37) <0.001 S. aureus 3 (4) 4 (6) 1.000 Values are numbers (percentages); p-values calculated using Fisher’s exact test; Significant differences (p<0.05) are shown in bold.
113
4.2.7.2 Carriage densities We also examined carriage densities of S. pneumoniae, H. influenzae, M. catarrhalis and S. aureus. Within each ethnic group there was no difference in carriage density of any of the four bacterial species between unvaccinated and vaccinated children (Figure 4.15). Vaccinated iTaukei children had slightly higher M. catarrhalis (Figure 4.15) carriage density compared with vaccinated Indo-Fijian children (median density of 9.60 x106 vs. 2.76 x 106, p=0.043). Interestingly, given that density and relative abundance are related, there was no difference in carriage density of S. pneumoniae in iTaukei children by vaccination status (p=0.189). Carriage densities were also considered by overall vaccination status (Figure 4.16) and by ethnicity (Figure 4.16).
There were no differences in carriage densities of the four bacterial species between unvaccinated and vaccinated children nor between iTaukei and Indo-Fijian children. However, despite low carriage of S. aureus, there was some evidence (p = 0.057) that the carriage density of S. aureus (Figure 4.16.D.ii) was higher in iTaukei children (median 1.29 x 106) compared with Indo-Fijian children (1.02 x 105).
114
A S. pneumoniae B H. influenzae 10 9 10 9 l l m / m 8 8 /
10 s 10 s t t n n e e l 7
l 7
10 a 10 a v v i i u u 6 6 q q 10 10 e e
e e 5 5 m 10 m 10 o o n n e 4 e 4 g 10 g 10 10 3 10 3 C M. catarrhalis D S. aureus 9 10 9 p = 0.043 10 l l m m / 8 8 / s 10 s
t 10 t n n e e l 7 l 10 7 a 10 a v v i i u u 6 6 q q 10 10 e e
e e 5 5 m
m 10 10 o o n n 4 e 4 e 10 g 10 g 10 3 10 3
Figure 4.15. Carriage densities (genomic equivalents/ml) of S. pneumoniae (A), H. influenzae (B), M. catarrhalis (C) and S. aureus (D) by ethnicity and vaccination status. Shown are unvaccinated (green) and vaccinated (blue) iTaukei children, and unvaccinated (orange) and vaccinated (red) Indo-Fijian children. Error bars are median and interquartile range, p-values calculated using the Mann-Whitney test. Statistical analysis was not possible for S. aureus density as the number of positive samples was too small.
115
i) Vaccination status ii) Ethnicity A 10 8 10 8 i ii l l m
m / 7 / 7
s 10
s 10 t t n n e e l l a 6 a 6 v v
i 10
i 10 u u q q e e
e 5 10 e 10 5 m m o o n n e e S. pneumoniae 4 4 g 10 g 10
10 3 10 3 B 10 9 i 10 9 ii l 8 l m 8 m / 10 / 10 s s t t n n e
7 e 7 l 10 l 10 a a v v i i u 6 u 6 q 10 q 10 e e
e e m 5 m 5
o 10 o 10 n n H. influenzae H. e e g 10 4 g 10 4
10 3 10 3 C 10 9 i 10 9 ii l 8 l 8 m m
/ / 10 10 s s t t n n e e 7 7 l l 10 10 a a v v i i u u 6
6 q q 10
10 e e
e e
m 5 m 5 o 10 o 10 n n e e M. catarrhalis M. g
g 4 10 4 10
3 10 3 10 D 10 8 i 10 8 ii l l m m /
7 / 7
s 10 s 10 t t
n n e e l l a 6 a 6 v v
i 10 i 10 u u q q e e
e 10 5 e 10 5 m m S. aureus o o n n e e 4 4 g g 10 10
10 3 10 3 unvaccinated iTaukei
vaccinated Indo-Fijian Figure 4.16. Carriage densities (genome equivalents/ml) of S. pneumoniae (A), H. influenzae (B), M. catarrhalis (C) and S. aureus (D) by vaccination status (i) and ethnicity (ii). Error bars are median and interquartile range, no significant differences observed as calculated by Mann-Whitney test.
116 4.2.8 S. pneumoniae overall and vaccine type carriage from FiPP
Vaccination appeared to have an effect on Dolosigranulum in Indo-Fijian children (Figure 4.10), however PCV7 did not affect Streptococcus at this timepoint, nor was there evidence of a direct interaction between Streptococcus and Dolosigranulum. We hypothesised that vaccination may have affected Streptococcus carriage at an earlier timepoint and the effect we observed on Dolosigranulum may be an artefact of earlier interactions. As such we analysed previously collected pneumococcal culture and serotyping data from the FiPP study, for the children included in our current microbiome study. We examined both the overall pneumococcal carriage rate (Figure 4.17.A) and the vaccine type carriage rate (Figure 4.17.B).
For the subset of children in our current study, vaccination did not appear to have an effect on the overall carriage rate at any of the time points for both iTaukei and Indo-Fijian children. This is consistent with the qPCR data (Table 4.7) for the 12 months of age time point for iTaukei (67% unvaccinated, 56% vaccinated, p=0.455) and Indo-Fijian children (22% unvaccinated, 24% vaccinated, p=1.000). However, while the patterns were consistent, the carriage rates of S. pneumoniae were slightly higher by qPCR compared with culture.
For iTaukei children at 6 months of age there was no significant difference in vaccine type carriage rate by vaccination status (15% unvaccinated vs. 6% vaccinated, p=0.259). However, at 9 months of age vaccinated iTaukei children had lower carriage of vaccine type pneumococci compared with unvaccinated iTaukei children (0% vs 24%, p=0.002), with some evidence this was maintained at 12 months of age (12% vs. 30%, p = 0.077).
For Indo-Fijian children the greatest difference between unvaccinated and vaccinated individuals was seen at 9 months of age (9% vs. 3%) however this was not statistically significant (p=0.355). At 12 months of age, the vaccine type carriage rate was very low in Indo-Fijian children with 1 out of the 32 (3%) unvaccinated children and 2 out of the 33 (6%) vaccinated children carrying a vaccine type pneumococci, a difference that was not statistically significant.
117
A
100 unvaccinated iTaukei
) 80 vaccinated iTaukei % ( unvaccinated Indo-Fijian e t
a 60
r vaccinated Indo-Fijian
e g
a 40 i r r a
c 20
0 6 months 9 months 12 months
B
50 unvaccinated iTaukei p=0.002 vaccinated iTaukei ) 40 %
( unvaccinated Indo-Fijian
e t vaccinated Indo-Fijian a 30 r
e g
a 20 i r r a
c 10
0 6 months 9 months 12 months
Figure 4.17. Overall (A) and vaccine type (B) pneumococcal carriage rate (%) for the children included in this study at 6, 9 and 12 months of age, analysed using a subset of data from FiPP (229). Error bars are 95% confidence intervals. Only in iTaukei children at 9 months of age (p=0.002) were differences observed between unvaccinated and vaccinated individuals within each ethnic group, calculated using Fisher’s exact test.
118
4.2.9 Association between risk factors and microbial composition
4.2.9.1 nMDS and PERMANOVA We considered the association of a range of risk factors with microbial composition, looking at both nMDS plots and PERMANOVA results. There were no significant differences in overall microbial composition by gender (p=0.838), season (p=0.106), year of swab collection (p=0.336), breastfeeding status (p=0.383), household exposure to cigarette smoke (p=0.271) or antibiotic use in the past two weeks (p=0.667) (see Appendix B. 7.2.2 Additional nMDS plots). There was a significant difference in microbial composition in children with- and without symptoms of an upper respiratory tract infection (p=0.003) defined as either a runny nose or cough (Figure 4.18). Individually, microbial composition differed in children for both runny nose (p=0.002, Figure 4.19.A) and cough (p=0.026, Figure 4.19.B).
Looking at the seven most common genera (Figure 4.20), children with symptoms of an upper respiratory tract infection had higher relative abundance of Moraxella (p=0.049) and Haemophilus (p=0.001); and lower relative abundance of Dolosigranulum (p=0.003) and Corynebacterium (p=0.001) compared with children without symptoms.
119
p=0.003
Figure 4.18. nMDS of the dissimilarity in microbial composition by symptoms of an upper respiratory tract infection. Shown are samples collected from children that have symptoms of an upper respiratory tract infection (blue) and children without symptoms of an upper respiratory tract infection (red). The centre points represent the mean of each group and the lines represent distance from the mean for each sample. There was a significant difference between the two groups (p=0.003), p-value calculated using PERMANOVA.
120
A
p=0.002
B
p=0.026
Figure 4.19. nMDS of the dissimilarity in microbial composition by presence of a runny nose (A) and a cough (B). Shown in blue are samples collected from children that have a runny nose or a cough and in red are children that do not have a runny nose or a cough. The centre points represent the mean of each group and the lines represent distance from the mean for each sample. There was a significant difference between the children with and without symptoms of a runny nose (p=0.002) and cough (p=0.026), p-value calculated using PERMANOVA.
121
p=0.003 ) 60 p=0.049 Pseudomonas
% Moraxella ( 50 e p=0.001 Staphylococcus c Dolosigranulum
n 40
a Streptococcus d 30 p=0.001 Corynebacterium n
u 20 Haemophilus b a 10 e v i t a
l 0.10 e
r 0.05 0.00 Figure 4.20. Median relative abundance (%) for the seven most common genera by symptoms of an upper respiratory tract infection (URTI). Error bars are interquartile range, p-values calculated using Mann-Whitney test.
122
4.2.9.2 Random forest models We also used random forest models to examine the association between risk factors and microbial composition (Table 4.10). These models supported findings using nMDS and PERMANOVA, with both ethnicity (p<0.001) and symptoms of an URTI (p=0.012) associated with microbial composition. Of URTI symptoms, presence of a runny nose was significant (p=0.004) and there was some evidence that presence of a cough was also significant (p=0.058).
Table 4.10. Significance of random forest models exploring association between risk factors and microbial composition Risk Factor OOB Error (%) p-value PCV7 58 0.836 iTaukei only 66 0.952 Indo-Fijian only 54 0.512 Ethnicity 33 <0.001 URTI symptoms (any) 29 0.012 Runny nose 20 0.004 Cough 20 0.058 Sex 55 0.886 Swab collection By year 60 0.504 By season 32 0.480 Antibiotic use 11 0.358 Cigarette smoke exposure 45 0.238 Child breastfeeding 45 0.570 Stopped by 6 weeks 10 0.400 Stopped by 6 months 20 0.306 OOB, out-of-bag; URTI, upper respiratory tract infection; Significant models (p<0.05) are shown in bold
Looking at the most important OTUs in these models (Figure 4.21), the importance of OTUs such as Corynebacterium, Staphylococcus and Moraxella in the ethnicity model (Figure 4.21.A) were consistent with earlier findings. However, the random forest model
123
highlighted the importance of less abundant OTUs such as Helcococcus, Acinetobacter, Prevotella and Paracoccus in relation to differences in microbial composition between ethnic groups. For symptoms of an URTI (Figure 4.21.B), the random forest model highlighted Helcococcus, Veillonella and Actinomyces are also important in addition to OTUs from the genera Haemophilus and Moraxella.
124
A
B
Figure 4.21. Top twenty most important OTUs in random forest models for ethnicity (A) and symptoms of an URTI (B).
125
4.2.9.3 Ethnicity and symptoms of upper respiratory tract infection Given microbial composition differed by both ethnicity and symptoms of an upper respiratory tract infection, we examined the impact of upper respiratory tract infection symptoms within each ethnic group (Figure 4.22). There was a significant difference in microbial composition by ethnicity and symptoms of an upper respiratory tract infection (p<0.001). Within each ethnic group, the presence of symptoms of an upper respiratory tract infection caused a shift in the microbial composition. Interestingly, the microbial composition in iTaukei children without symptoms of an upper respiratory tract infection was similar to the microbial composition in Indo-Fijian children with symptoms of an upper respiratory tract infection.
126
p<0.001
Figure 4.22. nMDS of the dissimilarity in microbial composition by ethnicity and symptoms of an upper respiratory tract infection. Shown are iTaukei children, with (purple) and without (blue), symptoms of an upper respiratory tract infection, and Indo- Fijian children, with (pink) or without (red), symptoms of an upper respiratory tract infection. The centre points represent the mean of each group, the lines represent distance from the mean for each sample and ellipses represent the standard error with a 95% confidence limit. There was a significant difference between the four groups (p<0.001), p-value calculated using PERMANOVA.
127
4.2.10 Multivariate Linear Regression
Multivariate linear regression analysis was performed to determine whether the results presented in the previous sections held true even after adjusting for multiple potential confounders. Based on applying the AIC to the models, each linear regression model included vaccination status, ethnicity, symptoms of an upper respiratory tract infection, breastfeeding status, household exposure to cigarette smoke, year of swab collection, season of swab collection, antibiotic use in the previous two weeks and sex of the child. The sequence count or number of reads for each sample was also included in the models to account for differences in sequencing depth. Age was excluded from all models as almost all children were aged 12 months (two children aged 13 months).
Table 4.11. Comparison of the top seven genera in unvaccinated and vaccinated iTaukei children following multivariate linear regression analysis. Median Relative Abundance (%) Genus level unvaccinated vaccinated Coefficient 95% CI p-value1 Pseudomonas2 0.04 0.03 1.20 -0.69, 3.09 0.208 Moraxella2 28.73 33.19 -1.11 -2.45, 0.23 0.103 Staphylococcus2 0.02 0.02 -0.19 -1.35, 0.96 0.738 Dolosigranulum 21.26 15.94 -2.24 -11.54, 7.07 0.632 Streptococcus2 2.02 0.52 -1.21 -2.24, -0.18 0.022 Corynebacterium 6.95 3.71 -1.26 -8.66, 6.14 0.735 Haemophilus2 1.02 0.11 -1.52 -2.98, -0.05 0.043 CI, confidence interval; significant differences shown in bold; 1p-value calculated following multivariate linear regression adjusting for ethnicity, symptoms of an upper respiratory tract infection, exposure to household cigarette smoke, breastfeeding status, year of swab collection, season of swab collection, antibiotic use in the previous two weeks and sex of the child; 2Log transformation of relative abundance was used in the linear regression models.
Results were similar from the unadjusted analysis (using the Mann-Whitney test), finding that Streptococcus differed (p=0.022) between unvaccinated and vaccinated iTaukei children. Vaccination was associated with a lower relative abundance of Streptococcus in iTaukei children. Vaccination was also associated with a lower relative abundance of Haemophilus in iTaukei children (p=0.043). Although vaccination in iTaukei children
128
wasn’t significantly associated with microbial composition in the random forest model, the two most important OTUs in the random forest model were Haemophilus and Streptococcus (see Appendix B. Figure 7.13.A) which is consistent with results from the linear regression model.
Table 4.12. Comparison of the top seven genera in unvaccinated and vaccinated Indo- Fijian children following multivariate linear regression analysis. Median Relative Abundance (%) Genus level unvaccinated vaccinated Coefficient 95% CI p-value1 Pseudomonas2 0.06 0.03 -1.34 -2.93, 0.23 0.089 Moraxella2 0.11 1.44 0.62 -1.34, 2.58 0.531 Staphylococcus2 0.06 0.06 0.48 -0.66, 1.61 0.403 Dolosigranulum 29.25 44.59 13.14 0.85, 25.43 0.037 Streptococcus2 0.56 0.62 0.50 -0.79, 1.79 0.438 Corynebacterium 10.39 24.05 5.67 -5.51, 16.85 0.314 Haemophilus2 0.03 0.04 0.71 -1.21, 2.62 0.462 CI, confidence interval; significant differences shown in bold; 1p-value calculated following multivariate linear regression adjusting for ethnicity, symptoms of an upper respiratory tract infection, exposure to household cigarette smoke, breastfeeding status, year of swab collection, season of swab collection, antibiotic use in the previous two weeks and sex of the child; 2Log transformation of relative abundance was used in the linear regression models.
In Indo-Fijian children, the results were also similar to the previous analysis with a higher relative abundance of Dolosigranulum in vaccinated Indo-Fijian compared with unvaccinated Indo-Fijian children (p=0.037) and no difference by vaccination status in Streptococcus (p=0.438). This was also consistent with the random forest model which also suggested that vaccination in Indo-Fijian children was not significantly associated with microbial composition but did suggest that Dolosigranulum was the most important OTU when considering the difference in microbial composition between unvaccinated and vaccinated Indo-Fijian children (see Appendix B. Figure 7.13.B).
129
Table 4.13. Comparison of the top seven genera in unvaccinated and vaccinated children following multivariate linear regression analysis. Median Relative Abundance (%) Genus level unvaccinated vaccinated Coefficient 95% CI p-value1 Pseudomonas2 0.04 0.03 1.03 -0.60, 2.66 0.212 Moraxella2 20.15 19.41 0.00 -1.12, 1.13 0.993 Staphylococcus2 0.04 0.04 0.05 -0.69, 0.80 0.887 Dolosigranulum 23.86 30.62 -1.43 -11.36, 8.51 0.777 Streptococcus2 1.02 0.59 -1.16 -2.21, -0.11 0.030 Corynebacterium 8.57 10.38 0.34 -5.98, 3.19 0.916 Haemophilus2 0.20 0.06 -1.50 -3.03, 0.02 0.053 1p-value calculated following multivariate linear regression adjusting for ethnicity, symptoms of an upper respiratory tract infection, exposure to household cigarette smoke, breastfeeding status, year of swab collection, season of swab collection, antibiotic use in the previous two weeks and sex of the child. An interaction between vaccination status and ethnicity was included in all models except those for Moraxella and Staphylococcus; 2Log transformation of relative abundance was used in the linear regression models.
Unlike the earlier unadjusted analysis, overall vaccination resulted in lower relative abundance of Streptococcus (p=0.030). An interaction term between ethnicity and vaccination status was considered for each model and included in all models except those for Moraxella and Staphylococcus as AIC analysis supported the use of the interaction term for these models. There was also some evidence that vaccination had an impact on Haemophilus (p=0.053), which was not observed in the earlier unadjusted analysis.
130
Table 4.14. Comparison of the top seven genera in iTaukei and Indo-Fijian children following multivariate linear regression analysis. Median Relative Abundance (%) Genus level iTaukei Indo-Fijian Coefficient 95% CI p-value1 Pseudomonas2 0.04 0.04 1.32 -0.36, 2.99 0.122 Moraxella2 29.33 0.20 -2.30 -3.44, -1.16 <0.001 Staphylococcus2 0.02 0.06 0.86 0.10, 1.61 0.026 Dolosigranulum 18.18 36.42 4.84 -5.43, 15.12 0.353 Streptococcus2 0.88 0.62 -0.95 -2.04, 0.13 0.085 Corynebacterium 4.83 17.92 9.89 3.52, 16.27 0.003 Haemophilus2 0.59 0.03 -2.38 -3.97, -0.78 0.004 1p-value calculated following multivariate linear regression adjusting for ethnicity, symptoms of an upper respiratory tract infection, exposure to household cigarette smoke, breastfeeding status, year of swab collection, season of swab collection, antibiotic use in the previous two weeks and sex of the child. An interaction between vaccination status and ethnicity was included in all models except those for Moraxella and Staphylococcus; 2Log transformation of relative abundance was used in the linear regression models.
After adjusting for other risk factors Moraxella, Staphylococcus, Corynebacterium and Haemophilus were significantly different between the two ethnic groups (p<0.001, p=0.026, p=0.003 and p=0.004, respectively). Without the inclusion of the interaction term, there was also a significant difference in Dolosigranulum (p=0.002) between the two ethnic groups.
131
Table 4.15. Comparison of the top seven genera in children with and without symptoms of an upper respiratory tract infection following multivariate linear regression analysis. Median Relative Abundance (%) Genus level no URTI URTI Coefficient 95% CI p-value Pseudomonas* 0.04 0.04 -0.63 -1.96, 0.70 0.348 Moraxella 7.51 30.55 0.56 -0.69, 1.82 0.373 Staphylococcus 0.04 0.02 -0.31 -1.14, 0.52 0.458 Dolosigranulum* 33.07 15.08 -2.42 -17.21, -1.73 0.017 Streptococcus 0.55 1.40 0.90 0.04, 1.75 0.040 Corynebacterium 13.54 2.61 -7.41 -14.44, -0.39 0.039 Haemophilus* 0.04 0.81 2.02 0.76, 3.27 0.002 CI, confidence interval; significant differences shown in bold; *Interaction between vaccination and ethnicity included in model.
Like the previous analysis, Dolosigranulum (p=0.017) and Corynebacterium (p=0.039) were significantly higher in relative abundance in children without symptoms of an upper respiratory tract infection compared with children with symptoms. The relative abundance of Haemophilus (p=0.002) was lower in children without symptoms and there was some evidence that Streptococcus (p=0.040) relative abundance was also lower but unlike the previous analysis, the relative abundance of Moraxella (p=0.373) was not lower in children without symptoms compared with children with symptoms of an upper respiratory tract infection.
The only other risk factor which affected these genera was household exposure to cigarette smoke, where Moraxella was lower in children exposed to household cigarette smoke compared with children not exposed to household cigarette smoke (Median relative abundance 10.47% (exposed) vs. 23.47% (not exposed), p=0.034).
132
4.2.11 Relative abundance and α-diversity
In some microbial communities, the presence or abundance of particular species or genera can have an effect on the richness or diversity of the community, here we have examined the correlation between relative abundance of the most common genera and both richness (Table 4.16) and Shannon diversity (Table 4.17).
Table 4.16. Correlation between relative abundance of the most common genera and OTU richness. Genus Spearman correlation p-value Pseudomonas -0.04 0.631 Moraxella -0.06 0.524 Staphylococcus 0.72 <0.001 Dolosigranulum 0.33 <0.001 Streptococcus 0.37 <0.001 Corynebacterium 0.47 <0.001 Haemophilus 0.04 0.658 Correlation calculated using Spearman’s correlation; significant differences shown in bold.
The relative abundance of Staphylococcus, Dolosigranulum, Streptococcus and Corynebacterium were all associated with greater richness (all p<0.001), this correlation was particularly strong for Staphylococcus (Spearman correlation 0.72).
133
Table 4.17. Correlation between relative abundance of the most common genera and Shannon diversity. Genus Spearman correlation p-value Pseudomonas -0.26 0.003 Moraxella 0.48 <0.001 Staphylococcus 0.09 0.284 Dolosigranulum -0.02 0.859 Streptococcus 0.57 <0.001 Corynebacterium 0.01 0.923 Haemophilus 0.37 <0.001 Correlation calculated using Spearman’s correlation; significant differences shown in bold.
Given previous results showing the dominance of Pseudomonas, it was unsurprising that there was a negative correlation between the relative abundance of Pseudomonas and Shannon diversity (p=0.003). However, there were also positive correlations between Shannon diversity and the relative abundance of Moraxella (p<0.001), Streptococcus (p<0.001) and Haemophilus (p<0.001).
134
4.3 Discussion Despite the widespread introduction of PCVs, there has been conflicting evidence from studies examining a limited number of species, that pneumococcal vaccination may be associated with species replacement. Research looking at the whole microbiome provides a broader insight into potential species replacement following pneumococcal vaccination. With limited data on the impact of pneumococcal vaccination on respiratory microbiota, particularly from children in developing countries, our study aimed to determine the effects of pneumococcal vaccination on the microbiome. Furthermore, given previously observed differences between the two ethnic groups in Fiji, it was of interest to determine whether pneumococcal vaccination has the same effect on each ethnic group.
Similar to the study of Kenyan children (213), and in contrast with the studies in the Netherlands (211) and Switzerland (212), we found little evidence for an overall effect of PCV on NP microbiota. At 24 months in the Netherlands study, there was no significant difference in microbial composition between unvaccinated and vaccinated individuals but there were differences at 12 months. However, it is difficult to compare our results with those in the Netherlands study as the third dose of PCV in our study was given to children at 14 weeks of age compared to 11 months of age in Netherlands study. In the Kenyan study no differences were observed 4 months after vaccination. Given the differences in the Netherlands study were observed 1 month after the last dose, this suggests that any vaccine effects on microbiota occur soon after vaccination and are transient.
We also found no difference in unvaccinated and vaccinated children in terms of the overall pneumococcal carriage rate or density, or vaccine-type carriage rates. In the original FiPP study there was some evidence vaccine-type carriage was lower in vaccinated children at 12 months of age (229), so perhaps as a consequence of subsampling and testing a reduced number of children this effect was not observed. This is unsurprising given our current study was not designed to assess S. pneumoniae carriage in particular, but rather to explore the effect of PCV on the microbiome more broadly.
Of all the factors affecting the microbial composition, the ethnicity of the child had the greatest impact. Ethnicity influenced both community structure, with higher Shannon diversity in iTaukei children, and community composition with differences in relative
135
abundances for Moraxella, Staphylococcus, Dolosigranulum, Corynebacterium and Haemophilus. Consistent with the previous chapter, carriage rates of S. pneumoniae, H. influenzae and M. catarrhalis were higher in iTaukei children. The clear differences between ethnic groups suggested that there was potential for pneumococcal vaccination to have varying effects between ethnic groups.
In iTaukei children, there was evidence pneumococcal vaccination reduced the relative abundance of Streptococcus (both at a genus level and the specific Streptococcus OTU likely to be S. pneumoniae). In addition, there was no evidence that pneumococcal vaccination had an effect on the overall microbial composition or any other of the most common genera. This suggests that pneumococcal vaccination had a very specific and targeted effect in iTaukei children without perturbing the microbiota and would therefore be unlikely to result in species replacement.
However, there was no evidence that pneumococcal vaccination affected the overall carriage rate, the vaccine-type carriage or the overall density of S. pneumoniae in iTaukei children. There could be several reasons why only the relative abundance of Streptococcus was affected by pneumococcal vaccination and not S. pneumoniae carriage rate or density. Firstly, as mentioned with the overall effect of vaccination, because of the low numbers in each group and the variance in density, we may lack the statistical power to detect a difference. However, for the relative abundance data, Streptococcus was present in every sample albeit in very low relative abundance in many samples, and so we would anticipate the power is not lacking. There was some evidence the vaccine type carriage rate was lower in vaccinated iTaukei children compared with unvaccinated iTaukei children at 12 months of age, although it was not statistically significant. Given that pneumococcal serotypes can colonise the respiratory tract at different densities (287), if the vaccine types that were eliminated were replaced by serotypes that colonise at a lower density we may also see a reduction in the relative abundance of Streptococcus, although if this were the case we would expect the overall density to be lower. Another explanation for the apparent contradiction in our results may relate to the total bacterial abundance in each group. It is possible that the absolute abundance differed between vaccinated and unvaccinated individuals – a phenomenon seen in the study by Bisbroek et al. (211) where at 24 months the absolute abundance was much higher in vaccinated
136
children compared with controls. Although, the mechanism for this finding is unclear, it is possible that the increased diversity seen in children shortly after vaccination led to an increase in total abundance over time in the vaccinated group compared to controls. If the S. pneumoniae density was the same between vaccinated and unvaccinated children but the absolute abundance differed, the relative abundance would also differ between vaccinated and unvaccinated children (see Appendix B., Figure 7.14 for graphical representation). If the total abundance was higher in vaccinated children as seen in the study by Bisbroek et al., then we would see results similar to those in our study where the relative abundance of Streptococcus was lower in vaccinated children. This hypothesis could be examined by further testing of our samples for total bacterial abundance.
Unlike iTaukei children, in Indo-Fijian children pneumococcal vaccination had little effect on Streptococcus relative abundance, nor did it affect S. pneumoniae carriage rate (overall and vaccine type) or density. However, the overall carriage rate and vaccine type carriage rate was much lower in Indo-Fijian children. For example, only 3% of unvaccinated Indo-Fijian children carried vaccine type pneumococci suggesting that PCV7 would be unlikely to have a statistically significant effect in these children.
Interestingly, the β-diversity analysis suggested that pneumococcal vaccination may have had some impact on the microbial composition. Looking specifically at the most common genera, the relative abundance of Dolosigranulum was higher in vaccinated Indo-Fijian children. This would not be unusual if the relative abundance of Streptococcus was also impacted by pneumococcal vaccination as Dolosigranulum has been reported to have a negative relationship with S. pneumoniae (288, 289). However, as mentioned earlier there was no difference in Streptococcus relative abundance by vaccination status and furthermore, there was no direct interaction between Dolosigranulum and Streptococcus. It is possible that indirect interactions may have played a role, but these are more complex to explore, and it is uncertain whether the effect on one bacterial species would be proportionate to the effects on other bacterial species. Weiss et al. showed that SparCC performed well at detecting mutual exclusion when there was a competitive relationship for engineered ecological relationships (290), however other tools based on different algorithms, such as CoNet, LSA or MENA (291-293), may provide additional insight into microbial interactions in our data.
137
It is possible that the higher relative abundance of Dolosigranulum in vaccinated Indo- Fijian children may be an artefact of earlier interactions, and while we analysed the data from the FiPP study at 6 months and 9 months of age and found no difference in overall or vaccine-type carriage, at 6 months of age it would still be over 3 months after the last dose of the vaccine (at 14 weeks of age). Potentially, pneumococcal vaccination may have had a transient effect on microbiota directly after administration, which may have impacted Dolosigranulum. Bisbroek, et al. (294) showed that microbiota profiles containing Dolosigranulum were more stable than profiles containing Streptococcus and therefore it is possible that any transient effect may have been maintained for Dolosigranulum but not Streptococcus. Also, as discussed regarding the iTaukei children, the fact that there was no difference in carriage rates at these earlier timepoints does not necessarily mean that there was no difference in the relative abundance of Streptococcus, something we are unable to measure with the data currently available.
Alternatively, the higher Dolosigranulum may not be an effect of vaccination at all, but instead reflect underlying differences in risk factors between unvaccinated and vaccinated Indo-Fijian children. Despite being part of a randomised control trial, random subsampling of children from the FiPP study did result in some differences in risk factors not seen in the overall study participants (229). In Indo-Fijian children, there were fewer male children in the unvaccinated group compared with the vaccinated group, and while gender did not appear to have a significant impact on microbial composition in this study, gender has been shown to effect OTU richness after a freeze-thaw (295) and influence NP microbial composition in adults (296), particularly the genus Dolosigranulum. Given the link between a healthy microbiome and Dolosigranulum, if pneumococcal vaccination does in fact result in increased Dolosigranulum, this may actually be considered a beneficial effect of vaccination.
There are several social, environmental and genetic factors which may drive the differences between the two ethnic groups. We noted in the previous chapter (231) that unlike other populations where socio-economic status is a significant factor influencing the underlying differences between ethnic groups, in Fiji the two ethnic groups have similar household income and therefore does not contribute to the differences between
138
iTaukei and Indo-Fijian children. In this study, there were two differences in risk factors between the two ethnic groups. iTaukei children were slightly heavier than Indo-Fijian children (although the weight dataset was incomplete), and household cigarette smoke exposure was greater in iTaukei children. In our study, overall exposure to cigarette smoke had no impact on microbial composition overall, though there were some specific effects related to Moraxella. Whether this has a meaningful impact is unknown, analysis of clusters and linear regression suggested that Moraxella (and the Moraxella dominated cluster) were associated with lower exposure to household cigarette smoke, and yet iTaukei children with a greater percentage exposed to household cigarette smoke had a higher relative abundance of Moraxella. In other studies there is some evidence that exposure to household cigarette smoke affects the microbiome of the nasopharynx (277) and other sites in the body (297, 298). Although our risk factor analysis was detailed, one limitation of our study is the uncertainty around the levels of exposure. A quantitative and non-subjective measure of exposure may be more relevant. Perhaps the lack of an effect on the microbiome in our study may mean that exposure levels were lower, or the effect may not be strong enough to detect with the number of children in the study.
One key risk factor missing from our study was number of children in household. This information was collected at the 6 months of age swab visit but not at 12 months of age swab visit. Available data from this subset of children at 6 months of age shows some evidence that the number of children in the household overall, but not children under 5 years of age, may be slightly higher for iTaukei children compared with Indo-Fijian children. However, some data was missing from these children and therefore this was not included in the analysis. There may be a difference between the two ethnic groups for this measure as in the FFS (chapter 3) we noted that number of children in the household (both overall and under 5 years of age) was significantly higher for iTaukei children compared with Indo-Fijian children. Hasegawa et al. reported an association between infants with cohabiting siblings and the nasal microbiome, in particular higher abundances of Moraxella (299). Therefore, the number of siblings could also be a factor in influencing the underlying differences in microbiota between the two ethnic groups, particularly as we found there was evidence that Moraxella was higher in iTaukei children. We also lacked data about social contact patterns and household crowding which may also be an important difference between the two ethnic groups.
139
Genetic factors may also play a role in determining microbial composition, however the evidence is mixed. A recent study in monozygotic and dizygotic twins showed that overall host genetics had no influence on microbial composition of the nasal passages, however both genetics and gender did play a role in determining total bacterial density (300). There are specific genetic factors that influence host interaction with microbes and these may be more relevant than overall host genetics. One such genetic factor is the β-defensin gene cluster on chromosome 8p23.1, where one of the genes encodes antimicrobial peptides on the epithelial cell surface of airways. The copy number of the β-defensin gene cluster has been shown to influence the microbial composition of the nasopharynx (301). Lower copy numbers were associated with increased risk of concurrent non-typeable H. influenzae, M. catarrhalis and S. pneumoniae colonisation.
The immune system plays an important role in shaping our microbiota (302, 303). Genetic factors such as polymorphisms in toll-like receptors (TLRs) and nucleotide oligomerisation domain (NOD) proteins can have a profound impact on the immune response and subsequent susceptibilities to diseases such as meningococcal meningitis and tuberculosis (304, 305). Polymorphisms have been linked to microbial community composition, particularly in immunity-related genes, and patterns of polymorphisms have shown to differ by geographic and ethnic origin (306, 307).
We considered the associations of a range risk factors and participant characteristics with the NP microbiome. One of the risk factors we examined was season of swab collection, which impacts on microbiota in temperate climates where there are distinct microbial profiles in spring/summer and autumn/winter (276, 279). However, the climate of Fiji is tropical with little temperature variation throughout the year (308), with a wet season (between November and April) and a dry season (between May and October). While we did not observe a significant difference in microbial composition between samples collected in the wet season and samples collected in the dry season (p=0.106), the nMDS plot did show some evidence of a separation between the samples collected in the two different seasons. With only a relatively small number of swabs collected in the dry season, it is possible that the effect of season on microbiota was not strong enough to detect with our sample size.
140
One risk factor which did have an impact on microbial composition was symptoms of an upper respiratory tract infection (URTI). The relative abundance in children with URTI symptoms was higher for Moraxella and Haemophilus, and lower for Dolosigranulum and Corynebacterium. The association with URTI symptoms and Moraxella and Haemophilus is perhaps unsurprising given they are known respiratory pathogens, however the link between health, Dolosigranulum and Corynebacterium is only just starting to be recognised. Presence of Dolosigranulum and Corynebacterium have been associated with a decreased risk of otitis media (289) as well as decreased episodes of respiratory tract infections (284, 294).
Given symptoms of an upper respiratory tract infection are often caused by viruses rather than bacteria, differences in bacterial microbiota observed in children with and without URTI symptoms suggests an interplay between viruses and bacteria, although this was not specifically examined in our study. In adults, rhinovirus has been shown to decrease α-diversity in individuals with infection and alter the relative abundance of the genera Neisseria and Propionibacterium (286). A recent study looking at respiratory syncytial virus (RSV) infection and NP microbiota found that RSV infection and hospitalisation was associated with carriage of Streptococcus and Haemophilus (309).
Interestingly, when we examined both ethnicity and symptoms of an URTI together, we saw that the microbiome of Indo-Fijian children with URTI symptoms was similar to the microbiome of iTaukei children without URTI symptoms. This suggests that the idea of a general ‘healthy’ NP microbiome may be a somewhat simplistic concept. Further, this study raises the question of whether the effects of pneumococcal vaccination on microbiota can be broadly applied to other populations or whether findings should be considered specific to the population or sub-population being studied.
The most abundant OTU in this study was from the genus Dolosigranulum and family Carnobacteriaceae. Interestingly, the taxonomic assignment differed when the sequence was classified using the Greengenes database (249) rather than the SILVA database (242) used in this study. Greengenes assigned this OTU to the closely related genus Alloiococcus also from the Carnobacteriaceae family (310). Specific BLAST (311) of the
141
OTU sequence for Dolosigranulum and Alloiococcus suggests this is more likely to be Dolosigranulum as the top hits for Dolosigranulum had 100% identity, whereas Alloiococcus had only 93% identity, across the whole sequence (data not shown). However, the true taxonomic classification of this OTU in our study would require additional investigation such as culture-based identification (312). This issue has also been noted in other studies (284, 309) and is likely due both genera being closely related.
Despite Dolosigranulum (and Alloiococcus) being identified as a frequent coloniser of the nasopharynx in children (211, 286), it was rarely described prior to the era of microbiome research and thus the link between health and Dolosigranulum is relatively recent. The paucity of historical data on Dolosigranulum carriage is possibly due to its specific growth requirements (313) and the fact that much of microbiology has been focused on pathogenic organisms and Dolosigranulum is a rare cause of disease. Dolosigranulum pigrum has been isolated as a potential cause of several diseases such as sinusitis, sepsis and pneumonia (314, 315), though the incidence of D. pigrum infections is unknown.
Of the most abundant genera in our study, most have been reported as predominant genera in other studies (213, 276), however Pseudomonas has not previously been commonly described as a predominant genus of the nasopharynx. We noted that Pseudomonas was generally present at a high relative abundance, this is highlighted by the negative correlation between Pseudomonas relative abundance and Shannon diversity. The Pseudomonas dominant cluster also had a very different microbial composition as seen in the nMDS plot (Figure 4.13), including when looking at other nMDS dimensions (see Appendix B., 7.2.1 Clusters and risk factors). In terms of bacterial interactions, Pseudomonas seemed to have sub-network of bacteria with either no interaction or negative interactions with the rest of the network. This again suggests a distinct microbial profile for samples containing significant levels of Pseudomonas.
Pseudomonas aeruginosa is the most common human pathogen of the genus and rarely colonises the nasopharynx although there is limited data on NP colonisation of P. aeruginosa (316, 317). P. aeruginosa has been shown (in combination with antibiotics) to contribute to a reduction in richness in the respiratory tract in children with cystic
142
fibrosis (318). In our study, the presence of Pseudomonas was not correlated with decreased richness (Spearman correlation -0.04, p=0.631). Long term persistence of P. aeruginosa in the nasopharynx has also been shown to result in migration to and persistence in the mouse lung (319). The health implications of the presence of Pseudomonas, particularly the dominant cluster, is unknown. However, P. aeruginosa is an opportunistic pathogen and generally only causes disease in individuals with compromised immune systems and underlying medical conditions such as cystic fibrosis (320). Climate and season affect acquisition of P. aeruginosa and disease severity in several conditions (321, 322), with increases in warmer seasons and tropical climates. As such, carriage of Pseudomonas in the nasopharynx may be more common in Fiji due to the tropical climate. This may also represent an area worth exploring by species-specific qPCR in our study or in future vaccine surveillance studies in Fiji (and other tropical environments) as there was some evidence in the linear regression analysis of an effect of PCV on Pseudomonas (p=0.0697). While Pseudomonas has been identified as a potential contaminant in several studies (323, 324), it was present at a low relative abundance in the ‘typical’ control profile (<1% in 10 out 12 controls). With samples containing Pseudomonas distributed across all seven rounds of DNA extraction, we believe it is unlikely that Pseudomonas is a contaminant in our study.
One key limitation of our research was that the samples used for analysis were not collected for the purposes of our current study. We therefore lacked certain controls such as STGG controls and swab controls which may be important for considering potential microbial contaminants but were not as relevant for assessing pneumococcal carriage. While Pseudomonas did not appear to be a contaminant from the DNA extraction kit or PCR reagents it is possible that it was a contaminant of STGG, which we are unable to determine without an appropriate control. Furthermore, the cotton swabs used may also contain bacterial DNA contaminants (31). The other aspect of our study which was limited by the original FiPP study design was the number of samples that were available for each group. The original study was designed to have proportions of samples collected from each ethnic group which were representative of the general Fijian population where the majority of the population are iTaukei. Our study was designed to have relatively even number of children in each ethnic group by vaccination status, limiting the number of samples available to 36 from each group.
143
A key strength of this study in that samples were obtained from a randomised control trial in a PCV naïve population, to examine the effects of pneumococcal vaccination on NP microbiota. In particular, herd immunity would not have an impact on unvaccinated individuals and risk factors should not differ significantly between the unvaccinated and vaccinated groups. However, as samples were collected between 2005 and 2007 (approximately 10 years of long-term storage at -80ºC), the quality of the samples and the potential for contamination were potential issues. Long-term frozen storage may impact on the quality and quantity of DNA recovered from swab samples and has the potential to have an effect on the microbial community structure (295). However, we were able to obtain good quality DNA and sequencing coverage for most samples. 8% of samples were excluded due either low sequence reads or because they had similar microbial profiles to the controls, consistent with other NP microbiome studies (279). Due to differences in sequencing depth and methodology, measures of richness can be difficult to compare between studies. However, the overall Shannon diversity in our samples was comparable to other studies in the Netherlands (294) and Switzerland (279) taken from children of a similar age. While frozen storage may have had some impact on the microbiota profiles, any effects are likely to be consistent between groups and therefore not affect overall conclusions, emphasising the importance of the randomised control trial study design.
We also examined the issue of reagent contamination which can be a significant problem in microbiome studies, particularly for low biomass samples (323). We sequenced all the PCR products from the extraction and PCR controls, as despite taking several measures to reduce contamination and repeating PCRs with no template added, a PCR product was present in all our controls. Contamination is difficult to eliminate in microbiome studies with potential contaminants in purified water (325, 326), DNA extraction kits (323, 327, 328) and PCR reagents (329, 330). Fortunately, the composition of the controls was different to the samples with clear separation of the control samples in the nMDS plot (see Appendix B., Figure 7.12) and in taxonomy of the OTUs present. We initially examined the first 24 samples sequenced and found approximately 50% of both controls comprised of sequence reads from the families Halomonadaceae and Bacillaceae. In contrast, none of the samples had a relative abundance greater than 1% for Bacillaceae,
144
and only two samples had a relative abundance greater than 1% (1.05% and 1.11%) for Halomonadaceae. While any contamination is undesirable, the level and impact of potential contamination was deemed minor.
4.4 Conclusions Overall, we found no significant impact of pneumococcal vaccination on the NP microbiota. However, we found distinct microbial profiles by ethnic group, and evidence that the effects of pneumococcal vaccination varied between ethnic groups. After stratifying by ethnicity, it was apparent that pneumococcal vaccination had positive effects, either reducing Streptococcus in iTaukei children, or increasing Dolosigranulum in Indo-Fijian, a bacterial genus that has been linked with a healthy microbiome. We also found that children with symptoms of an upper respiratory tract infection had a distinct microbial composition compared with children without symptoms.
145
Chapter 5
Adult Pneumococcal Carriage in
Fiji
146
5 Adult pneumococcal carriage in Fiji
5.1 Introduction and aims In the previous chapters we explored the effects of pneumococcal vaccination in vaccine trials in Fiji, prior to the introduction of a pneumococcal conjugate vaccine into the immunisation schedule. Here, we examine the indirect effects of vaccine introduction in unvaccinated adult caregivers in Fiji, following the introduction of the 10-valent pneumococcal conjugate vaccine (PCV10) into the national immunisation program in 2012.
5.1.1 Indirect effects of pneumococcal vaccination in adults
5.1.1.1 Pneumococcal disease in adults While a significant amount of the burden of pneumococcal disease occurs in children less than five years of age, pneumococcal disease is also prevalent in the elderly due to their waning immune system. Waning of the immune system with age, also known as immunosenescence, is characterised by a loss of a variety of immune cells and an increase in pro-inflammatory cytokines (331, 332). Immunosenescence can lead to a change in the immune response to S. pneumoniae (333) and as a result, the risk of death due to diseases like pneumonia (for which S. pneumoniae is a major cause (334)) increases with age (335).
5.1.1.2 Pneumococcal vaccination in adults Currently, in several developed countries such as Australia, pneumococcal vaccination is recommended for the adults over the age of 65 years, and in individuals of younger age with comorbidities such as chronic heart disease, chronic lung disease and the immunocompromised. Previously, 23vPPV alone has been has been recommended for adults, though some countries now recommend either PCV13 or a combination of PCV13 and 23vPPV (127). There is mixed evidence for the effectiveness of pneumococcal vaccination in adults, particularly for 23vPPV (336-338). Vila-Corcoles et al. found that 23vPPV reduced IPD including in the very elderly (>80 years of age) and high-risk groups. A meta-analysis found that 23vPPV vaccination in adults had efficacy against IPD as well as pneumococcal pneumonia, but not all-cause mortality (337). Consistent
147
with this finding, Baldo et al. found that PCV13, but not 23vPPV, was effective at reducing mortality due to all-cause community acquired pneumonia in adults (335). In a large study of >80,000 adults aged over 65 years in the Netherlands (where infant PCV7 was introduced in 2006, and PCV10 in 2011 with no catch-up), PCV13 was effective at preventing pneumococcal pneumonia and IPD, but had no effect on overall mortality or all-cause pneumonia (338). Given the mixed evidence for pneumococcal vaccination in adults, and the fact that children are an important reservoir for pneumococcal transmission to adults (339), reduction in pneumococcal colonisation through infant vaccination plays an important role in the protection of adults from pneumococcal disease.
5.1.1.3 Indirect effects of infant vaccination on invasive pneumococcal disease in adults The indirect effects of vaccination refer to the protection of unvaccinated members of a community when others within the community are vaccinated (sometimes also referred to as herd immunity, herd protection or community protection). Indirect effects confer some protection to unvaccinated members through reduced transmission (340). Many studies have demonstrated indirect effects of infant pneumococcal vaccination to other population groups including adults, with most examining the impact on IPD (341). Following PCV7 introduction in England and Wales, the incidence of vaccine type IPD in adults ≥ 65 years of age was 81% lower than pre-vaccine introduction levels, although with an increase in non-vaccine type IPD the net benefit was a 19% reduction in IPD overall (133). Reductions in cases of adult IPD caused by vaccine type pneumococci continued in England and Wales after PCV7 was superseded by PCV13 (146). Reductions in adult IPD incidences following PCV introduction were also observed in many other countries including the USA (342), Israel (343), South Africa (344), Canada (345) and Sweden (346). However, it should be noted that in Sweden IPD was reduced in adults under the age of 65 years but not the elderly (>65 years), and in Israel the decrease in IPD in the elderly was smaller and took longer to occur than in adults <65 years of age. In Canada, reductions in vaccine type IPD were offset by increases in non-vaccine type IPD, and in Uruguay there was little change in IPD following PCV7 introduction (347).
Infant pneumococcal vaccination can also have an indirect effect on pneumonia in adults. In the UK, adults in contact with a vaccinated child were less likely to have community acquired pneumonia caused by a vaccine type, compared with adults in contact with
148
unvaccinated children (348). In Scotland, the introduction of PCV7 (and later PCV13) had no impact on hospitalisation due to all-cause pneumonia in adults over 75 years of age, however the length of hospital stays and deaths due to pneumonia decreased in the post-vaccine era in this age group (349).
5.1.1.4 Indirect effects of infant vaccination on pneumococcal carriage in adults There are less data on the indirect effects of infant pneumococcal vaccination on carriage (compared with IPD) but available data suggest that infant pneumococcal vaccination can reduce carriage of vaccine type pneumococci in adults. In Alaska, the proportion of adults carrying vaccine type pneumococci significantly decreased following PCV7 introduction (from 28% of adult carriers pre-vaccine introduction to 5% post introduction) (167). In the US, a study of households in Navajo and White Mountain Apache reservations found that adults living in households with vaccinated children had significantly lower vaccine type carriage compared with adults living with unvaccinated children (165). At a population level, following 8 years of routine PCV7 use in Navajo and White Mountain Apache communities, there was also a general reduction in vaccine type carriage in adults (350). In a PCV7 trial in The Gambia, vaccine type colonisation decreased in adults two years after children were vaccinated (155). In Kenya, following the introduction of PCV10 into the immunisation schedule, there was a reduction in vaccine type carriage in unvaccinated individuals including adults (189).
5.1.2 New Vaccine Evaluation Project (NVEP) in Fiji
In 2012, PCV10 (along with rotavirus and HPV vaccines) was introduced into the immunisation schedule in Fiji. The New Vaccine Evaluation Project (NVEP) was designed to evaluate the impact of these vaccines. As part of assessing the impact of PCV10, annual cross-sectional carriage surveys were conducted from 2012 to 2015 at the same time each year. The 2012 survey was conducted prior to vaccine introduction. In each carriage survey, approximately 500 NP swabs were collected from children 5-8 weeks of age, 12-23 months of age and 2-6 years of age each year. In addition, paired NP and OP swabs were collected (approximately 500 each) from adult caregivers. Caregivers were chosen as they are likely to have frequent contact with children and subsequently may have higher rates of pneumococcal carriage and disease (348, 351). Therefore,
149
indirect effects are more likely to be evident in adult caregivers and caregivers are likely to benefit the most from indirect effects.
5.1.2.1 Detection and serotyping of pneumococci in adults using nasopharyngeal and oropharyngeal samples Current WHO recommendations for detecting pneumococcal carriage in adults advises sampling both the oropharynx and the nasopharynx (31). These recommendations were based on five studies in adults where the sensitivity for detection of S. pneumoniae in the nasopharynx ranged from 58% to 81% (excluding the study by Hendley et al. (339), which involved only 24 adults) when compared with overall detection in both the nasopharynx and the oropharynx (72, 352-354). Compared with children, where the sensitivity for pneumococcal detection in the nasopharynx was over 90% in the majority of studies, in adults there is some evidence that there is added benefit in sampling the oropharynx in addition to the nasopharynx. All these studies used culture-based methodology, and the use of molecular-based techniques was identified as an area where further research was needed (31).
Since the recommendations were published, two studies have compared NP and OP sampling for the detection of pneumococci in adults (41) and adolescents (355) using molecular methods. Both studies concluded that sampling the oropharynx was superior to sampling the nasopharynx, leading to a study in adolescents where only OP samples were used to assess carriage (356). However, there is some evidence that the use of molecular methods for serotyping of S. pneumoniae in OP samples may yield false positive serotyping results owing to pneumococcal capsule loci present in non- pneumococcal species (357).
5.1.3 Aims
The primary aim of this chapter was to examine the indirect effect of PCV10 introduction on pneumococcal carriage in adult caregivers in Fiji. We also aimed to re-evaluate the need for OP sampling in addition to NP sampling, utilising the improved methods for detecting and serotyping pneumococci. We hypothesised that the added benefit to sampling the oropharynx in addition to the nasopharynx in adults may be diminished
150
when utilising these methods. Given the evidence that OP samples may yield false positive serotyping results, we also examined whether sample type (NP or OP) affects identification and serotyping results when using molecular methods.
151
5.2 Results
5.2.1 Is there an added benefit to sampling the oropharynx in addition to the nasopharynx for adults?
5.2.1.1 Examination of NP and OP samples using molecular methods All swabs collected as part of NVEP were processed using standard molecular methods (see chapter 2). There are concerns regarding false positive pneumococcal identification and serotyping results for testing adult OP samples and with molecular methods. As such, we initially assessed a subset of NP and OP samples using molecular-based methods. We processed 250 paired NP and OP samples from adult caregivers collected in 2012. As part of S. pneumoniae identification (2.4.2) and serotyping, following DNA extraction (2.6.2), each sample was screened by lytA qPCR (2.7.2). Samples that were positive (Ct<35) or equivocal (Ct 35-<40) for the lytA gene were then cultured on gHBA plates (neat, 1:10 and 1:100 dilutions) and resultant growth was harvested (2.3.2). DNA was extracted from the plate harvest (2.6.5) and underwent microarray analysis (2.5.4). Following processing of these samples there were several differences between NP and OP samples.
Table 5.1. Paired NP and OP sample (n=250) processing results following lytA screen, culture and microarray. lytA qPCR Result after NP samples OP samples result culture and (n=250) (n=250) microarray positive 14 9 positive 13 (93%) 7 (78%) negative 1^ (7%) 2 (22%) equivocal 19 34 positive 7 (37%) 3 (9%) negative 12^ (63%) 31# (91%) negative 217 207 Percentages are out of samples that were either equivocal or positive for lytA ^all 12 samples had no growth on gHBA plate; #one sample out of the 31 with no growth on gHBA plate, remaining 30 samples were negative by microarray.
After the initial lytA qPCR screen of the samples, we found that OP samples had a higher proportion of samples that were equivocal for the lytA gene compared with NP samples (34/250, 13.6% vs. 19/250, 7.6%; p=0.041). Following the sample culture and microarray
152
steps, we found that a smaller proportion of OP samples that had originally tested as lytA equivocal contained S. pneumoniae compared with NP samples (3/34, 9% vs. 7/19, 37%; p=0.025). Of the 42 OP samples analysed by microarray, only 10 (24%) were positive for S. pneumoniae, compared with NP samples where all 20 (100%) samples were positive for S. pneumoniae (p<0.001).
Table 5.2. Growth on gHBA of NP and OP samples that were positive or equivocal for the lytA gene. NP OP Growth on gHBA n=33 n=43 Pneumococcal growth only 18 (55%) 0 (0%) Pneumococcal and non-pneumococcal growth 2 (6%) 10 (23%) Non-pneumococcal growth only 2 (6%) 32 (74%) No growth 11 (33%) 1 (2%) Values are numbers (percentages); Pneumococcal growth defined by detection of pneumococci in culture or in subsequent microarray testing.
When lytA positive and equivocal samples were cultured on gHBA plates, NP and OP samples had different patterns of growth (Table 5.2). NP samples generally either had pneumococcal growth (55%) or no growth (33%), with only 4 of 33 (12%) samples tested having non-pneumococcal growth. In contrast, almost all OP samples (42/43; 98%) cultured had a complex mix of bacterial growth on gHBA plates. A representative example of the difference in colonial morphology is shown in Figure 5.1.
153
A B
Figure 5.1. Example of bacterial growth on a gHBA plate for a nasopharyngeal sample (A) compared with an oropharyngeal sample (B).
154
The density of bacterial growth on gHBA plates also differed between the two sample types with higher density in OP samples. Specifically, there was growth on all three dilution plates (neat plate, 1:10 dilution plate and 1:100 dilution plate) in 41 (98%) out of the 42 OP samples with bacterial growth compared with 13 (59%) of the 22 NP samples with bacterial growth (Table 5.3).
Table 5.3. Any bacterial growth on gHBA dilution plates for NP and OP samples. NP OP Dilution plate p-value n=22 n=42 Neat 22 (100%) 42 (100%) 1.000 1:10 18 (82%) 41 (98%) 0.044 1:100 13 (59%) 41 (98%) <0.001 Values are numbers (percentages); p-values calculated using Fisher’s Exact test.
Initially, 12 OP samples were sent to the Bacterial Microarray Group at St. George’s University of London and analysed with an earlier version of the microarray. Following the establishment of microarray at MCRI, we analysed a further 30 OP samples using the newer version of microarray which included a StrepID component to identify S. pneumoniae and other common streptococcal species. These 30 OP samples contained multiple Streptococcus species (Figure 5.2), with an average of seven Streptococcus spp. in each sample. The streptococcal species detected were mainly S. salivarius, S. mitis, S. oralis, S. infantis and S. parasanguinis (Table 5.4).
155
14
12
10 y c
n 8 e u q
e 6 r F 4
2
0 5 6 7 8 9 No. of non-pneumococcal Streptococcus spp.
Figure 5.2. Number of Streptococcus spp. (excluding S. pneumoniae) detected by the StrepID component of the microarray in 30 oropharyngeal samples. The StrepID component can detect up to 12 streptococcal species (in addition to S. pneumoniae).
156
Table 5.4. Number of oropharyngeal samples (n=30) containing various streptococcal species detected by microarray. No. (%) of OP samples containing Streptococcus spp. various Streptococcus spp. n=30 S. salivarius 30 (100%) S. mitis 30 (100%) S. oralis 30 (100%) S. infantis 30 (100%) S. parasanguinis 29 (97%) S. anginosus 23 (77%) S. sanguinis 22 (73%) S. constellatus 7 (23%) S. intermedius 1 (3%) Values are numbers (percentages)
We also examined the serotyping results (from microarray testing) for 33 NP and 43 OP samples (Figure 5.3). NP samples typically had either one serotype (n=17) or no serotypes detected per sample (n=13). The OP samples had unusual serotyping results with a median of nine ‘serotypes’ per sample, however in most cases these ‘serotypes’ actually contained either a partial or divergent set of capsule genes (example in Figure 5.4) which corresponded to non-pneumococcal bacterial species. These ‘serotypes’ are therefore unlikely to represent true pneumococcal capsule. Consistent with this hypothesis, there were 29 samples which contained no pneumococci but nevertheless had these unusual serotyping results. Only three OP samples had no pneumococci and also no ‘serotype’ detected. Microarray assigns ‘-like’ to serotypes for which one or more genes are divergent or absent from the typical set of genes for that serotype. The most common of these partial or divergent ‘serotypes’ (Figure 5.5) was serotype 36 or 36-like (detected in 28 out of 42 OP samples) followed by serotype 21/21-like (detected in 19 (45%) OP samples), serotypes 16A/16A-like (detected in 17 (40%) OP samples) and 19B/19B-like (detected in 14 (33%) OP samples). Acapsular lineages NT3b/NT3b-like (detected in 22 (52%) OP samples) and NT2 (detected in 14 (33%) OP samples) were also common.
157
Figure 5.3. Comparison of the number of ‘serotypes’ detected by microarray in nasopharyngeal (NP) and oropharyngeal (OP) samples that were lytA positive or equivocal. The serotypes detected in NP and OP samples include those for which there were an incomplete or divergent set of capsule genes, or reactions from samples with non- pneumococcal bacterial species.
158
A
B
C
Figure 5.4. An example of serotyping results using microarray for a NP sample containing one (A) or two (B) serotypes, compared with an OP sample (C). For the OP sample, none of the ‘serotypes’ had all the cps genes expected for the corresponding serotype as shown by the wide range of signal intensities for each gene.
159
d 80 e t c
e 70 t e d
' 60 e p
y 50 t o r
e 40 s '
s 30 e l p
m 20 a s
f 10 o
% 0 * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * e e * * e * e * * 1 F * F * F e * * * * 4 * * 9 e 5 * b e e e e e 2 e b e e e e e e e e e e e e e e e e e e e e e e e e e k k e e k e k e a C 6 e 7 e A B 3 k B e e e e 3 B 7 3 A k e 3 k k k k k T k 4 k k k k k k k k k k k k k k k k k k k k k k k k k li li k k li k li k 4 1 1 k 1 k 9 9 2 li 4 k k k k 5 3 1 li k li li li li li li T li li li li li li li li li li li li li li li li li li li li li li li li li - - li li - li - li T 1 li li 1 1 - 2 li li li li 3 4 - li T - - - - - N ------]------]- - - - F A - - A - 5 - - - A - - - - 5 - N 6 b 1 A B A N C C 8 7 A L 1 F A 3 5 5 B B A 3 4 V F C 9 9 6 F 4 7 9 0 F 5 A A N F C 4 F A F B 4 N 3 3 2 6 9 0 0 9 4 /2 4 /9 2 8 3 4 0 1 0 /4 /9 0 5 3 3 /4 /7 1 1 1 2 8 3 5 /9 6 8 2 4 8 2 3 9 T /1 1 1 1 1 * /2 * / 1 /3 1 /1 2 * * 1 3 / * * 2 3 * 1 1 2 2 3 3 N * 7 * L C * * 3 V 3 6 F A */ A 2 A 9 0 A B 4 9 [4 4 7 9 b 6 4 1 3 1 3 1 2 [ 3 1 T N
Figure 5.5. Percentage of OP samples (n=42) in which each ‘serotype’ was detected using microarray. ‘Serotypes’ include those from both pneumococcal and non-pneumococcal species, with ‘serotypes’ marked by ‘*’ indicating that the ‘serotype’ is likely from a non-pneumococcal bacterial species and ‘-like’ indicates the ‘serotype’ contains a partial or divergent set of capsule genes.
160
In summary, we found that OP samples were more complex than NP samples, particularly with the presence of other bacterial growth (particularly streptococci) on culture plates and the presence of partial or divergent sets of pneumococcal capsule genes.
161
5.2.1.2 The contribution of non-pneumococcal species to identification and serotyping results in OP samples Based on the findings by Carvalho et al. (357, 358) and Johnston et al. (359), we hypothesised that the non-pneumococcal streptococci in the OP samples may affect both pneumococcal detection (resulting in equivocal lytA results) and serotyping results (as the source of the ‘serotypes’ with partial or divergent capsule genes). Pneumococci were only detected in a few OP samples that were lytA equivocal (3/34, 9%). However, it is unclear the extent to which pneumococci were present in the other 31 OP samples (e.g. pneumococci not detected due to low density) or whether there was cross-reactivity of the lytA primers and probes with closely related streptococci carrying lytA gene homologs, or a combination of the two factors. It is also unclear the extent to which other serotyping methods could be confounded by pneumococcal capsule genes (or partial gene sequences) that are potentially present in non-pneumococcal species.
Here, we aimed to determine the contribution of non-pneumococcal species to the OP identification and serotyping results and explore impact of OP complexity on serotyping methods. We selected 11 OP samples that reflected the spectrum of microarray and lytA qPCR results to characterise in detail.
5.2.1.2.1 Characterisation of OP samples To examine whether potentially false positive results were being obtained by other serotyping methods, we chose two commonly used serotyping methods in addition to the lytA qPCR and microarray already performed on these 11 OP samples. Therefore, the 11 OP samples selected for further investigation were serotyped using latex sweep agglutination and multiplex PCR (Table 5.5)
162
Table 5.5. lytA qPCR identification and serotyping results using microarray, latex sweep and multiplex PCR for a subset of OP samples. lytA qPCR Microarray Latex sweep Sample ID Density1 Microarray serotyping3 mPCR serotyping4,5 result PathID2 agglutination (Ct) FVEP-002- Equivocal 9A-like* (34%) + 7F-like* (26%) + 36-like* (17%) 4 + 9L + 36 + 7F/7A + 24F/24A/24B + 1.35x103 SP-3/5 004 (39.52) + NT4b* (17%) + 27-like* (6%) 15 10F/10C/33C + 9N/9L 33F/33A/37 + 35B + NT4a* (53%) + 24A-like* (21%) + NT4b* (11%) FVEP-002- Equivocal 24F/24A/24B + 4 + 6.35x102 SP-3/5 + 43/39-like* (9%) + 16A-like* (3%) + 28A-like* 24A 418 (38.48) 10F/10C/33C + (2%) + 4-like* (1%) 35A/35C/42 + 9N/9L 10C/21-like* (41%) + NT4a* (20%) + 36-like* 22F/22A + 24F/24A/24B FVEP-002- Equivocal (12%) + 39-like* (8%) + 16A-like* (8%) + 48- 6.60x102 SP-3/5 NSD + 10A + 10F/10C/33C + 496 (37.93) like* (4%) + 24A-like* (4%) + 45-like* (2%) + 20 + 13 19B-like* (1%) 6A/6B/6C/6D + 6C/6D + 22F/22A + 33F/33A/37 + NT4b* (35%) + 10C/21-like* (29%) + 48-like* FVEP-002- Equivocal SP-5/5 35B + 4 + 23F + 10A + 2.89x103 (11%) + 41A (7%) + 20-like* (5%) + 23F (4%) + NSD 460 (35.82) GAS-5/5 10F/10C/33C + 5 + 35A-like* (4%) + 5-like* (3%) + 16A-like* (2%) 35A/35C/42 + 34 + 21 + 20 6A/6B/6C/6D + 19A + 33F/33A/37 + FVEP-002- Equivocal NT4b* (46%) + 33A-like* (40%) + 9V (5%) + 12 + 33 + 1.10x103 SP-3/5 38/25F/25A + 35B + 10A 002 (37.79) 18F-like (4%) + 35A-like (4%) + 19B-like (1%) 35B + 10F/10C/33C + 35A/35C/42 FVEP-002- Equivocal NT4b* (62%) + 36-like* (25%) + 5-like* (7%) + 6.35x102 SP-3/5 19B 4 + 5 + 9N/9L 084 (37.57) 19B-like* (5%) + 9N-like* (1%) 38/25F/25A + 4 + 23F + FVEP-002- Equivocal NT4b* (45%) + 48-like* (36%) + 19B-like* (11%) 39 + 10F/10C/33C + 34 2.62x102 SP-3/5 NSD 488 (39.82) + 43/39-like* (8%) + 21 + 20 + 6A/6B/6C/6D
163
lytA qPCR Microarray Latex sweep Sample ID Density1 Microarray serotyping3 mPCR serotyping4,5 result PathID2 agglutination (Ct) 15A/15F + 7F/7A + NT4b* (26%) + 10C/21-like* (26%) + 7F-like* FVEP-002- Positive 24F/24A/24 + 4 + 3.30x105 SP-3/5 (26%) + 16A-like* (11%) + 45-like* (6%) + 9L- 35B 078 (28.44) 10F/10C/33C + 9N/9L + like* (5%) 21 + 20 FVEP-002- Positive NT4b* (32%) + 48-like* (27%) + 24B (20%) + 24F/24A/24B + 1 + 5.52x103 SP-5/5 19B 080 (34.41) 19B-like* (13%) + 1 (8%) 9N/9L + 13 FVEP-002- Positive NT4b* (42%) + 10C/21-like* (42%) + 19C-like* 15B/15C + 10F/10C/33C 6.35x103 SP-3/5 NSD 422 (34.70) (12%) + 16A-like* (4%) + 21 + 17F 15A/15F + 10F/10C/33C FVEP-002- Equivocal NT4b* (71%) + 10C/21-like* (17%) + 11B-like* 1.00x104 SP-3/5 NSD + 35A/35C/42 + 9N/9L + 486 (35.29) (5%) + 16A-like* (5%) + 19C-like* (2%) 31 + 20 Ct, cycle threshold result for lytA qPCR; NSD, no serotype detected; A positive lytA qPCR result is defined as Ct<35, an equivocal lytA qPCR result are defined as a Ct between 35 and 40, and a negative lytA qPCR result is defined as Ct>40. 1Density in genomic equivalents/ml as determined by lytA qPCR; 2SP-5/5 result for the PathID component on microarray indicates a sample positive for S. pneumoniae, SP-4/5 or less indicates S. pneumoniae is not present and GAS-5/5 indicated the present of group A streptococcus; 3Serotypes marked by ‘*’ indicates that the serotype is likely from a non-pneumococcal bacterial species and ‘-like’ indicates the ‘serotype’ contains a partial or divergent set of capsule genes; 4Each sample also had additional mPCR products that did not correlate with an expected mPCR serotype product size that are not listed here; 5Two samples negative for the cpsA pneumococcal control gene are shown in italics.
164
A serotyping result was obtained in six of the 11 samples serotyped by latex sweep agglutination, with 19B and 35B detected in two samples each. In five of the samples with a latex sweep result, microarray suggested these samples did not contain pneumococci. Serotyping results were obtained in all 11 samples by multiplex PCR, however for two of these samples the cpsA pneumococcal control gene was not positive. For microarray and mPCR less than half (microarray, 28/57, 49%; mPCR, 25/73, 34%) of the serotypes detected were also detected by another method (Figure 5.6). Of the 108 serotypes detected by any method, only four (4/108, 4%) were detected by all three serotyping methods. For example, sample FVEP-002-418 had a latex sweep agglutination result of 24A, and microarray identified a 24A-like serotype which appeared to come from a non-pneumococcal species, whilst mPCR also identified a 24F/24A/24B serotype. Both mPCR and microarray indicated that this sample did not contain pneumococci suggesting that all three serotyping methods can detect pneumococcal ‘serotypes’ when no pneumococci are present, however only microarray is able to designate when a serotype is likely to belong to a non-pneumococcal species.
165
Figure 5.6. Serotyping result agreement between latex sweep agglutination (green), microarray (blue) and mPCR (red) serotyping for a subset of OP samples (n=11). Numbers are serotypes detected by each method, including partial or divergent capsule genes detected by microarray (excluding pneumococci from unencapsulated lineages in the microarray results). A total of 108 serotypes were detected by any of the three methods.
166
5.2.1.2.2 Characterisation of OP isolates We next hypothesised that the non-pneumococcal species within these samples are the main contributors to the false positive results, rather than pneumococci being present at low densities or pneumococci containing a novel set of capsule genes. The 11 samples were re-cultured onto gHBA plates and a representative of each colony morphology present was then subcultured onto HBA plates. This resulted in 91 isolates which were then characterised by lytA real-time PCR (at a standardised concentration of DNA), optochin susceptibility, bile solubility, MALDI-TOF/MS and latex sweep agglutination serotyping (Table 5.6) for identification and serotyping.
Table 5.6. Characterisation of isolates (n=91) from a subset of OP samples by various identification and serotyping tests. All isolates S. pneumoniae non-S. pneumoniae Test n=91 isolates isolates only n=3 n=88 lytA real-time PCR
Positive 3 (3%) 3 (100%) 0 (0 %) Equivocal 10 (11%) 0 (0%) 10 (11%) Negative 78 (86%) 0 (0%) 78 (89%) Optochin
Sensitive 3 (3%) 3 (100%) 0 (0%) Intermediate 1 (1%) 0 (0%) 1 (1%) Resistant 87 (96%) 0 (0%) 87 (99%) MALDI-TOF MS ID obtained 80 (88%) 3 (100%) 77 (88%) No ID obtained 11 (12%) 0 (0%) 11 (13%) Bile solubility
Soluble 5 (5%) 3 (100%) 2 (2%) Insoluble 58 (64%) 0 (0%) 58 (66%) Not readable^ 28 (31%) 0 (0%) 28 (32%) Latex Agglutination
Serotyping result 13 (14%) 3 (100%) 10 (11%) Negative* 45 (49%) 0 (0%) 45 (51%) Not readable^ 33 (36%) 0 (0%) 33 (38%)
167
All isolates S. pneumoniae non-S. pneumoniae Test n=91 isolates isolates only n=3 n=88 Microarray
Serotyping result 39 (43%) 1 (33%) 38 (43%) Negative* 3 (3%) 0 (0%) 3 (3%) Not tested 49 (54%) 2 (67%) 47 (53%) ‘Serotyping result’ indicates a serotyping result was obtained, this includes partial or divergent ‘serotypes’ and those from non-pneumococcal species; *Indicates no serotyping result was obtained; ^Bacterial culture did not emulsify in saline or only partially emulsified in saline rendering the sample ‘not-readable’.
Three of these isolates were identified as S. pneumoniae, being optochin sensitive, bile soluble, lytA real-time PCR positive and were identified by MALDI-TOF MS as S. pneumoniae. These isolates were serotyped by latex agglutination as serotype 1, 15B and 35B. There were 88 isolates that were deemed to be non-pneumococci based on a combination of lytA qPCR results, optochin, bile solubility and MALDI-TOF MS results (see 2.4.2.3 for identification algorithm). Of these, 87 isolates were optochin resistant and one isolate had intermediate optochin susceptibility. Two of the 88 isolates were bile soluble, 58 isolates were insoluble and the remaining 28 isolates were unable to emulsify in saline and were therefore not readable. For lytA real-time PCR, 10 of the 88 isolates were equivocal and 78 isolates were negative, however it should be noted that the equivocal result (Ct 35 to 40) at a DNA concentration of 1 ng/μl is considered negative for S. pneumoniae. For 11 of the 88 isolates no identification was obtained by MALDI- TOF MS despite re-testing. In most cases, lack of identification with MALDI-TOF MS was due to ambiguous results. Nearly half (41/88, 47%) of the isolates were identified as S. mitis/S. oralis, with S. parasanguinis (12/88, 14% of isolates) and S. salivarius (11/88, 13%) also common (Figure 5.7).
Latex agglutination serotyping was attempted on all isolates, however many of the isolates (n=33) were unable to emulsify (or were only partially emulsified) in saline, so latex agglutination serotyping was unable to be performed. A serotyping result was obtained for 10 isolates out of the 55 non-pneumococcal isolates that were able to be tested by latex agglutination. The most common result was positivity for both serotype
168
19B and serogroup 12 (4/55, 7% of isolates tested). Positivity for serotype 19B alone (3/55, 5% of isolates tested), serogroup 35 (2/55, 4%) and both serotype 33F sand serogroup 35 (1/55, 2%) were also observed (Figure 5.8.A).
Quellung was attempted for all 10 isolates with a positive result in latex agglutination, however the typical ‘swelling’ or cell enlargement observed for S. pneumoniae when positive for a serogroup or serotype was absent. Some slight cell enlargement was observed for some isolates (such as for FVEP-002-002-04 which serotyped as 19B in latex agglutination), however this observation was dependent on the operator and therefore was clarified as a negative result.
As microarray is an expensive technique to apply, a subset of isolates (n=41) were chosen for serotyping by microarray to further elucidate identification and serotyping results. All isolates that were equivocal for lytA or positive for a serotype/serogroup in latex agglutination were chosen. Isolates covering range of species identification results were also chosen for serotyping by microarray, with at least one isolate chosen for each unique MALDI-TOF MS result. Serotyping results were obtained for 38 of the 41 (93%) non- pneumococcal isolates tested, with all ‘serotypes’ identified as coming from non- pneumococcal species and all had a partial or divergent set of capsule genes. The most common of these serotypes were 7F-like (5/41, 12% of isolates tested), 36-like (4/41, 10%), 10C/21-like (3/41, 7%), 7F/36-like (3/41, 7%) and 19B-like (2/41, 5%), with the unencapsulated lineage NT4b also common (3/41, 7%) (Figure 5.8.B).
169
50 45 40 s
e 35 t a 30 o l s
i 25
o f 20
% 15 10 5 0 s s s s e e s i s li i u a sa n e u n a in ri su i ia t io r u o n o ct ye a t o g va n o in a sce s st . n li i g l e i fica S g m la a r cr ti / sa sa n u o ccu . n tis a . a e ci lu e i r S . n u ysg f co S d m a S p m d s a i . p . . a r o S . S ia S n a n S th o P o m o R d u se P
Figure 5.7. MALDI-TOF MS identification results for 88 non-pneumococcal isolates subcultured from 11 OP samples.
170
A
90 s
e 80 t
a 70
o l 10 s i
8
o f 6 4 % 2 0 2 5 5 e 1 B 3 3 v 9 ti + 1 + a B F g 9 3 e 3 n 1 B
14 12 s e t 10 a
o l 8 s i 6 o f 4 % 2 0 e e e e b e e e e e e e e e e b e e a e e e e e e ik ik ik ik 4 ik ik ik ik ik ik ik ik ik ik 4 ik ik 4 ik ik ik ik ik iv -l -l -l -l T -l -l -l -l -l -l -l -l -l -l T -l -l T -l -l -l -l -l t F 6 1 6 N F 4 0 F F 9 8 N 1 N 1 3 6 9 a 3 2 3 B 1 A 2 A 3 4 2 A 2 4 3 F 2 g 7 / / 9 /7 6 6 0 3 / + / 3 + 7 / e C F 1 1 1 /1 /1 3 A e F /3 e + + + + 3 n 0 7 2 F 1 3 7 k 4 1 / 4 ik F i b a a a C 2 /2 /3 -l 7 -l 4 4 4 4 + 0 A 6 6 T T T T b 1 3 3 A 3 N N N N 4 3 3 / /3 F + T 1 7 e N /2 ik F -l 7 F 7 Figure 5.8. Latex agglutination (A) and microarray serotyping (B) results for 88 non- pneumococcal isolates subcultured from 11 OP samples. These data exclude results from isolates that were unreadable by latex agglutination (n=33), and isolates were not tested by microarray (n=47).
171
Of interest, during the selection of morphologically distinct colonies from the 11 OP samples, we isolated S. pneumoniae from OP sample FVEP-002-422, which originally did not appear to have pneumococci by microarray. However, at the time of isolation we noted that this colony morphology was only evident on the most dilute agar plate where the density of bacterial growth was low.
As OP sample FVEP-002-460 was the most complex by mPCR (Table 5.5), all 10 isolates derived from this sample were serotyped using mPCR as well as by microarray (Table 5.7). Sample FVEP-002-460 contained one S. pneumoniae isolate (which was positive for lytA real-time PCR, serotyped as 35B by all three serotyping methods and was identified as S. pneumoniae by both microarray and MALDI-TOF MS). For the other nine isolates, the microarray and mPCR serotyping results were generally not concordant, however both showed the presence of pneumococcal capsule genes in non-pneumococcal species. Four of the nine non-pneumococcal isolates were positive for the cpsA pneumococcal control gene in mPCR.
172
Table 5.7. Identification and serotyping tests for 10 individual isolates derived from sample FVEP-002-460. lytA real- Microarray MALDI- Microarray Latex Isolate time PCR PathID1 Microarray serotyping2 mPCR serotyping3,4 TOF/MS ID streptococcal ID Agglutination (Ct)
1 Negative S. mitis/ S. oralis SP-3/5 S. oralis 33A/21/10A-like* Negative 10A + 33F/33A/37 + 5 + 23B 2 Equivocal S. salivarius SP-1/5 S. salivarius 7F-like* ND ^10A + 13 (37.97) 3 Positive S. pneumoniae SP-5/5 S. pneumoniae 35B 35B 35B (17.61) 4 Negative S. mitis/ S. oralis SP-2/5 S. infantis 20-like* Negative 15A/15F + 31 + 20
5 Equivocal S. mitis/ S. oralis SP-1/5 S. infantis 36-like* ND 20 (39.61) 43/29-like* (45%) 6 Negative S. mitis/ S. oralis SP-3/5 S. mitis Negative 13 NT4b* (55%)
7 Equivocal S. mitis/ S. oralis SP-3/5 S. mitis NT4b* Negative 6A/6B/6C/6D + 15A/15F (36.90) +10A 8 Negative S. anginosus SP-1/5 S. anginosus 7F/21-like* Negative 6A/6B/6C/6D + 22F/22A + 35F/47F + 6C/6D 9 Negative No ID obtained SP-2/5 S. oralis 10C/21-like* Negative 7F/7A + 9V/9A + 1 + 35A/35C/42 + 2 10 Negative S. parasanguinis SP-1/5 S. parasanguinis 7F/36-like* Negative 22F/22A + 7F/7A + 5 Ct, cycle threshold result for lytA real-time PCR; A positive lytA real-time PCR result is defined as Ct<35, an equivocal result is defined as a Ct between 35 and 40, and a negative result is defined as Ct>40; Not determined (ND) represents an isolate which was not readable by latex agglutination; 1SP-5/5 result for the PathID component on microarray indicates an isolate positive for S. pneumoniae, SP-4/5 or less indicates
173
the isolate is not S. pneumoniae; 2Serotypes marked by ‘*’ indicates that the serotype is likely from a non-pneumococcal bacterial species and ‘-like’ indicates the ‘serotype’ contains a partial or divergent set of capsule genes; 3Each isolate also had additional mPCR products that did not correlate with an expected mPCR serotype product size that are not listed here; 4Isolates negative for the cpsA pneumococcal control gene are shown in italics.
174
5.2.1.2.2.1 Whole genome sequencing We chose to further explore six isolates by whole genome sequencing, focussing on confirming the species identification of the isolate, investigating the presence of lytA homologs and examining capsule genes, particularly in the isolates which were serotypeable using phenotypic methods. The six isolates were chosen for whole genome sequencing (Table 5.8) based on latex agglutination and microarray serotyping results (FVEP-002-002-03, FVEP-002-084-07, FVEP-002-078-01, FVEP-002-080-02), lytA qPCR result (FVEP-002-080-01) and MALDI-TOF MS result (FVEP-002-084-03). Following whole genome sequencing, sequences were assembled into contigs and the contigs were submitted to the Rapid Annotation using Subsystem Technology (RAST) server for annotation.
175
Table 5.8. Six OP isolates chosen for whole genome sequencing. lytA real-time Microarray Latex Isolate MALDI-TOF MS PCR PathID StrepID Serotyping Agglutination S. oralis (2/5) FVEP-002-002-03 Negative no ID SP-3/5 33F-like* 33F; 35 S. mitis (3/5) S. mitis/ S. oralis (3/5) FVEP-002-078-01 Negative SP-3/5 7F/33A-like* 35 S. oralis S. mitis (3/5) Equivocal FVEP-002-080-01 S. salivarius SP-1/5 S. salivarius (5/5) 7F-like* ND (38.28) S. mitis/ S. mitis (3/5) FVEP-002-080-02 Negative SP-2/5 19B-like* 19B S. oralis S. infantis (2/5) NT4a*(55%) + FVEP-002-084-03 Negative Paracoccus yeei SP-3/5 S. mitis (3/5) negative 36-like*(45%) S. salivarius (5/5) FVEP-002-084-07 Negative S. salivarius SP-1/5 S. mitis (1/5) 36-like* 19B; 12 S. pneumoniae (1/5) Ct, cycle threshold result for lytA real-time PCR; A positive lytA real-time PCR result is defined as Ct<35, an equivocal result is defined as a Ct between 35 and 40, and a negative result is defined as Ct>40. *indicates that the serotype is likely from a non-pneumococcal bacterial species; ‘-like’ indicates the ‘serotype’ contains a partial or divergent set of capsule genes; Not determined (ND) represents an isolate which was not readable by latex agglutination; 1SP-5/5 result for the PathID component on microarray indicates an isolate positive for S. pneumoniae, SP-4/5 or less indicates the isolate is not S. pneumoniae.
176
5.2.1.2.2.1.1 Isolate identification Using the whole genome sequencing data for each isolate, identification was performed using Kraken, MetaPhlAn and compared with the nearest neighbour genome in RAST (Table 5.9). The 16S rRNA gene sequence was also used for identification by comparing with the SILVA and Greengenes databases as well as BLAST. We also did a phylogenetic analysis based on the concatenated sequences of gene fragments from seven housekeeping genes, comparing references sequences from the viridans group streptococci electronic multilocus sequence analysis (eMLSA) database with those from our isolates (Figure 5.9).
Isolates FVEP-002-080-01 and FVEP-002-084-07 were identified as S. salivarius by all methods. This included by eMLSA with FVEP-002-080-01 and FVEP-002-084-07 sharing 99.4% and 99.9% sequence identity over the 3063 bp concatenated sequence with S. salivarius SK56 (strain 391) and S. salivarius SK729 (strain 394), respectively. Definitive identification for the remaining four isolates was more challenging. Isolate FVEP-002-084-03 is likely S. mitis as this was the top result for all methods (except MALDI-TOF MS which identified the isolate as Paracoccus yeei). FVEP-002-084-03 also clustered with S. mitis isolates in eMLSA and shared 96.4% identity with S. mitis SK1126 (strain 145). Isolates FVEP-002-002-03 and FVEP-002-078-01 shared 99.68% identity with each other over the 16S rRNA sequence and were likely to be S. oralis subsp. tigurinus (basonym Streptococcus tigurinus). The eMLSA results for FVEP-002-002-03 support this finding as FVEP-002-002-03 clustered with S. oralis IgA protease negative strains (consistent with S. oralis subsp. tigurinus (360)) and also shared 96.6% identity with strain S. oralis SK664 (strain 175). However, FVEP-002-078-01 clustered with S. oligofermentans SK1136 (strain 170) in eMLSA and shared 96.1% identity over the concatenated sequence. The ambiguity of results for these two isolates likely arises from the poor representation of the species in the databases used for identification and to some extent misclassification of mitis group streptococci in databases like GenBank (360). For isolate FVEP-002-080-02 a definitive identification was not possible based on these results, although the MetaPhlAn results would suggest this isolate is S. infantis with 87% of sequence reads mapping to S. infantis. FVEP-002-080-02 also clustered with S. infantis isolates in eMLSA and shared 96.9% identity with S. infantis SK350 (strain 61).
177
Table 5.9. Identification of six OP isolates following whole genome sequencing using different several different methods. 16S rRNA sequence based ID RAST MALDI-TOF Microarray closest MS StrepID Kraken MetaPhlAn BLAST Greengenes neighbour Probable Isolate (% (number of (% of reads (% of reads (% sequence SILVA (% sequence genome species confidence positive genes/ covered) covered) coverage/ identity) (genome level) five tested) identity) ID) FVEP- No ID S. mitis S. oralis S. tigurinus* S. tigurinus* Streptococcus S. dentisani S. oralis S. (3/5) AZ_3a sp. oral taxon *strain 7747 SK255 tigurinus* 002-002- (38.96) (89.26) S. oralis 071 (98/99) (1005704) 03 (2/5) (99.48)
FVEP- S. mitis/ S. oralis S. oralis No ID S. tigurinus* Streptococcus S. dentisani* S. oralis S. S. oralis AZ_3a sp. oral taxon strain 7747 SK255 tigurinus* 002-078- (3/5) (40.97) (97.8) S. mitis 071 (98/99) (1005704) 01 (99.35) (3/5) FVEP- S. salivarius S. salivarius S. salivarius S. salivarius S. salivarius S. salivarius str. S. salivarius S. salivarius S. (99.9) (5/5) SK126 strain ATCC SK126 salivarius 002-080- (84.37) (100) (100) 7073 (596322.3) 01 (99/100)
FVEP- S. mitis/ S. mitis S. pneumoniae S. infantis S. mitis S. mitis str. S. dentisani* S. infantis S. infantis S. oralis (3/5) 127R strain 7747 SK1076 002-080- (4.00) (87.27) (84.8) S. infantis (99.22) (98/99) (1005705.3) 02 (2/5)
178
16S rRNA sequence based ID RAST MALDI-TOF Microarray closest MS StrepID Kraken MetaPhlAn BLAST Greengenes neighbour Probable Isolate (% (number of (% of reads (% of reads (% sequence SILVA (% sequence genome species confidence positive genes/ covered) covered) coverage/ identity) (genome level) five tested) identity) ID) FVEP- Paracoccus S. mitis S. mitis S. mitis/ S. mitis S. mitis str. S. mitis strain S. mitis S. mitis yeei (3/5) NCTC 127R/str. 209 NS51 SK575 002-084- (17.52) S. oralis/ (96.2) S. pneumoniae 12261 (99.22) (96/99) (1095736) 03 (100) FVEP- S. salivarius S. salivarius S. salivarius S. salivarius S. salivarius S. salivarius S. salivarius S. salivarius S. (99.9) (5/5) (99.22) strain ATCC SK126 salivarius 002-084- (86.80) (100) S. mitis 7073 (596322.3) 07 (1/5) (99/99) S. pneumoniae (1/5) *S. tigurinus and S. dentisani have been reclassified as subspecies of S. oralis (360).
179
A B
D
C
Figure 5.9 Subtrees (A-D) from phylogenetic analysis of concatenated sequences from gene fragments of seven housekeeping genes. Sequences (3063 bp) from the six OP isolates were compared with those from reference isolates in the viridans group streptococci eMLSA database. Subtrees were generated from a minimum-evolution phylogenetic tree, with the evolutionary distances computed using the Maximum Composite Likelihood method and are in the units of the number of base substitutions per site. The six OP isolates are shown with open circles (labelled without the ‘FVEP-002-’ prefix). In subtree A, S. oralis isolates are shown dark blue, S. oralis-mitis biovar 2 isolates are shown medium blue, S. oligofermentans is shown in grey, S. oralis IgA protease negative isolates are shown in light blue and isolates with uncertain species identification are shown in black. In subtree B, S. salivarius isolates are shown in orange, S. thermophilus isolates are shown in brown and S. vestibularis isolates are shown in off- yellow. In subtree C, S. infantis isolates are shown in green and the S. peroris isolate is shown in teal. All isolates in subtree D are S. mitis (shown in red).
180
5.2.1.2.2.1.2 lytA genes BLAST search of any genes initially annotated by RAST to be N-acetylmuramoyl-L- alanine amidase (EC 3.5.1.28), was done to confirm the annotation. Isolate FVEP-002- 078-01 had two sequences annotated as N-acetylmuramoyl-L-alanine amidase, however BLAST revealed that one of these sequences corresponded to lytB, an endo-beta-N- acetlyglucosaminidase. Isolate FVEP-002-080-01, which was lytA equivocal, had three genes which were annotated as N-acetylmuramoyl-L-alanine amidase but only two of these clearly corresponded to N-acetylmuramoyl-L-alanine amidase and the product from the third gene was either described as a hypothetical protein or from the glycosyl hydrolase family 25. This was also excluded from further analysis. No gene encoding N- acetylmuramoyl-L-alanine amidase was found in isolate FVEP-002-084-03.
Using BLAST+ with default Blastn parameters, a representative pneumococcal lytA gene from S. pneumoniae strain R6 (NCBI reference sequence: NC_003098.1) was aligned against the whole genome sequencing contigs. No hits were found for any of the six isolates consistent with their negative and equivocal lytA real-time PCR results. The contigs were also searched using the primer and probes sequences which also did not result in any hits.
All N-acetylmuramoyl-L-alanine amidase genes from the isolates were then aligned against the R6 lytA sequence as well as the primer and probe sequences (Table 5.10). The results show poor alignment (<50% identity) with the pneumococcal lytA gene and match relatively poorly with primer and probe sequences for the lytA qPCR assay. Additionally, for several of the genes the order of primers and probes would not result in an amplified signal. Given the low sequence identity with primers and the probe, these results indicate that the weak positive/equivocal result in lytA qPCR for isolate FVEP-002-080-01 is likely to be a result of non-specific amplification (or other technical issues) rather than homologs of the lytA gene being present.
181
Table 5.10. Comparison of N-acetylmuramoyl-L-alanine amidase genes in a subset of OP isolates with lytA qPCR primer and probe sequences and a representative S. pneumoniae lytA gene from strain R6. Nucleotide identity (%) Order of primers lytA real- Size S. pneumoniae strain R6 Forward Reverse and probe Isolate time PCR Copy Probe (bp) lytA sequence primer primer consistent with (Ct) 25 nt 957 nt 22 nt 21 nt amplification FVEP-002-002-03 Negative 1 1800 259 (27) 7 (32) 14 (56) 13 (62) yes FVEP-002-078-01 Negative 1 1800 160 (17) 15 (68) 15 (60) 15 (71) no
Equivocal 1 2127 269 (28) 8 (36) 14 (56) 12 (57) yes FVEP-002-080-01 (38.28) 2 2931 267 (28) 9 (41) 15 (60) 13 (62) yes 1 2013 252 (26) 15 (68) 13 (52) 14 (67) yes FVEP-002-080-02 Negative 2 1659 250 (26) 13 (59) 14 (56) 12 (57) no 3 888* 326 (34) 11 (50) 15 (60) 13 (62) no 1 2127 264 (28) 13 (59) 13 (52) 12 (57) yes FVEP-002-084-07 Negative 2 2808 249 (26) 13 (59) 16 (64) 11 (52) yes nt, nucleotides; *Phage copy of N-acetylmuramoyl-L-alanine amidase; Note no N-acetylmuramoyl-L-alanine amidase gene was annotated for isolate FVEP-002-084-03.
182
5.2.1.2.2.1.3 Capsule genes Pneumococcal capsule genes (or partial genes) were identified in all six isolates. Two of the isolates were selected for whole genome sequencing, FVEP-002-002-03 and FVEP- 002-080-02, because they were serotypeable by latex agglutination and had a corresponding microarray result indicating partial or divergent capsule genes for that serotype. We explored the similarity of the capsule locus in these two isolates to their corresponding serotype.
Isolate FVEP-002-002-03 serotyped as 33F-like by microarray and as 33F using latex agglutination, therefore we compared the capsule genes present in this isolate with the capsule genes present for serotype 33F (Figure 5.10). The first six genes (wzg, wzh, wzd, wze, wchA and wciB) shared <80% identity to those present in serotype 33F. The first five of these six genes were most similar to those in S. mitis (GenBank: AB181235.2) sharing 93-97% identity to FVEP-002-002-03. With the wciB gene in FVEP-002-002-03 sharing 77% identity to multiple S. pneumoniae strains (which were the top hits in the BLAST search). Whereas the next five genes (wciC, wciD, wciE, wciF and wzy) shared >90% sequence identity with serotype 33F genes. There was a truncated copy of wzx which shared 94% sequence identity with this portion of the wzx 33F gene, followed by wcrG which is not present in serotype 33F; this gene shared 93% sequence identity with a wcrG from serotype 10A. The full length wzx gene from our isolate shared only 76% identity with the 33F wzx gene, and our isolate shared only 73% and 86% identity with the next two genes, wciG and glf, respectively. Unlike 33F, our isolate did not possess a wcjE gene but instead had a wcyO gene which was somewhat similar to the wcyO gene found in serotypes 21 (88% identity) and 33C (75% identity).
183
Gene present by microarray
Figure 5.10. Comparison of the capsule gene locus for isolate FVEP-002-002-03 with serotype 33F (GenBank Accession no. AJ006986.1). Sequence identity (%) between the two sets of genes is shown. Genes are categorised according to characteristics used in Bentley et al. 2006 (49). ‘Gene present by microarray’ indicates whether a gene was categorised as present (grey) or absent (white) based on the signal intensity for the set of probes for each gene.
184
As 33F is included in the mPCR assay we also checked the wzy gene from our isolate with the serotypes 33F/33A/37 primers from the mPCR assay.
FVEP-002-002-03 GAAGGCAATCCATATGATTATGTCGCG |||||||||| || ||||| ||||||| Forward primer GAAGGCAATCAATGTGATTGTGTCGCG
FVEP-002-002-03 GTAGAAGGGTACTATAATCTTCATTTTGAAG ||||||||||||||||||||||||||||||| Reverse primer GTAGAAGGGTACTATAATCTTCATTTTGAAG
There were three mismatches with the forward primer and perfect alignment with the reverse primer, therefore it is possible that this may have contributed to the positive mPCR result for 33F/33A/37 in sample FVEP-002-002.
Isolate FVEP-002-080-02 serotyped as 19B-like by microarray and 19B by latex agglutination, therefore we compared the capsule genes present in this isolate with the capsule genes present for serotype 19B (Figure 5.11). Similar to above, the first six capsule genes (wzg, wzh, wzd, wze, wchA, wchO) of FVEP-002-080-02 shared 56-78% identity with the capsule genes present in serotype 19B. These genes were most similar to Streptococcus sp. oral taxon (Genbank: CP014264.1) with FVEP-002-080-02 sharing between 87-92% identity. However, all other genes had >89% sequence identity with serotype 19B. This isolate did not possess the typical aliA flanking gene downstream of the capsule genes but rather immediately downstream of rmlD was ftsA, a gene encoding cell-division protein FtsA.
Microarray assigns ‘-like’ to serotypes for which one or more genes are divergent or absent from the typical set of genes for that serotype. For both isolates, the microarray results correlate with the sequence identity for each serotype. For FVEP-002-002-03 the genes which were deemed present by microarray shared >90% identity with the 33F reference isolate. For FVEP-002-080-02, those deemed present by microarray shared around 80% identity or higher with the 19B reference isolate.
185
Gene present by microarray
Figure 5.11. Comparison of the capsule gene locus for isolate FVEP-002-080-02 with serotype 19B (GenBank Accession no. CR931676.1). Sequence identity (%) between the two sets of genes is shown. Genes are categorised according to characteristics used in Bentley et al. 2006 (49). ‘Gene present by microarray’ indicates whether a gene was categorised as present (grey) or absent (white) based on the signal intensity for the set of probes for each gene.
186
In summary, we identified non-pneumococcal streptococci that possessed pneumococcal capsule genes, and this caused spurious serotyping results using both traditional and molecular based methods. For each serotype, microarray uses multiple targets for each of the cps genes which is complimented by a multi-locus streptococcal species component, and is therefore able to distinguish these from true pneumococci. We also found that non- pneumococcal streptococci may also contribute to a weak lytA qPCR signal for OP samples.
187
5.2.1.3 Devising a more specific pneumococcal identification approach for screening OP samples In the previous section we showed that the presence of non-pneumococcal streptococci may result in a lytA signal when no pneumococci are present. Although microarray was effective at species discrimination, it is an expensive and analytically challenging method to apply to a set of OP samples, particularly when less than a quarter of the samples contained pneumococci. As such, we next aimed to find a more specific approach to pneumococcal identification in OP samples by using an additional or alternative target to lytA qPCR.
We considered a range of potential gene targets from the literature and those used in the S. pneumoniae StrepID component of the microarray. Several potential targets (rpoA (37), sodA (38), tuf (39), recA (40)) were not considered as they require sequencing of the PCR product for S. pneumoniae identification and were therefore impractical for a high- throughput identification or screening approaches. Several other gene targets (ply (34), psaA (35), Spn9802 DNA fragment (36)) have previously been reported to be positive for other streptococcal species or not positive for some S. pneumoniae strains (237, 361, 362). A PCR method to amplify a 217 bp product of 16S rRNA gene (363) was also not considered as the addition of a probe for adapting the PCR into a qPCR assay may have reduced specificity for S. pneumoniae as it would likely cover a sequence which was not specific to S. pneumoniae.
We selected three pneumococcal targets from the literature, as well as four pneumococcal targets from the StrepID component of microarray, to investigate their use as an additional or alternative to pneumococcal screening using lytA (Table 5.11).
188
Table 5.11. Potential S. pneumoniae-specific targets. Gene ID* Gene Product Reference SP_1032 iron-compound ABC transporter, iron piaA (364) (SP_RS05120) compound binding protein SP_1033 iron-compound ABC transporter, permease piaB (41)^ (SP_RS05125) protein SP_2038 ulaA PTS ascorbate transporter subunit IIC (365) (SP_RS10305) SP_0049 vanZ teicoplanin resistance protein Microarray (SP_RS00290) SP_0137 - ABC transporter ATP-binding protein Microarray (SP_RS00700) SP_2020 bguR GntR family transcriptional regulator Microarray (SP_RS10220) SP_2167 fucK L-fuculose kinase Microarray (SP_RS11055) *Gene ID refers to TIGR4 genome (NCBI Reference Sequence: NC_003028), updated gene ID shown in brackets; ^initially referred to as piaA, later amended to piaB (42).
5.2.1.3.1 Testing targets on a reference set of isolates by PCR Firstly, we designed primers for the pneumococcal species identification targets used in microarray (vanZ, SP_0.137, bguR, fucK). These new targets were then tested along with piaA and ulaA (using previous published PCR protocols) against a set of 30 reference isolates (Table 5.12). Note that as piaB has been used in a qPCR assay concurrently with lytA in other studies, we chose to include this for further testing as a real-time PCR and qPCR assay, and not as a PCR assay. bguR appeared to perform slightly better than the other targets with all non-target organisms negative and all six S. pneumoniae clearly positive. vanZ also performed well, but while the PCR against S. dysgalactiae was negative, a PCR product of an incorrect size was detected for this PCR. uluA performed poorly, failing to detect one of six pneumococci and having an ambiguous result for two of the six.
189
Table 5.12. Results for potential pneumococcal-specific PCR assays tested against a panel of reference isolates. ID Species, serotype^ vanZ SP_0137 piaA bguR fucK ulaA PMP6 S. pneumoniae, 5 + + + + + + PMP7 S. pneumoniae, 19F + + + + + + PMP85 S. pneumoniae, 23F + + + + + +/- PMP849 S. pneumoniae, 1 + + + + + + PMP1064 S. pneumoniae, 22A + + + + + - 0173-01 S. pneumoniae, NT + + + + + +/- 001-009-01 S. pseudopneumoniae ------PMP1300 S. pseudopneumoniae ------PMP16 S. mitis - +/- +/- - +/- - PMP933 S. mitis ------PMP934 S. mitis - -* + - - - PMP1010 S. salivarius - - -* - - - PMP935 S. mutans - -* - - - - PMP1056 S. oralis - -* - - - - PMP1301 S. infantis ------PMP1000 S. pyogenes ------PMP1057 S. gordonii ------PMP1035 S. bovis - -* - - - - PMP1051 S. dysgalactiae -* - +/- - - -
190
ID Species, serotype^ vanZ SP_0137 piaA bguR fucK ulaA PMP994 S. agalactiae - +/- -* - - - PMP1053 S. sobrinus ------PMP936 S. sanguinis ------PMP1049 S. anginosus - -* - - - - PMP1059 S. vestibularis ------PMP1302 S. peroris ------PMP1303 S. australis ------PMP1304 S. australis ------PMP1305 S. oligofermantans ------PMP1306 S. cristatus ------PMP1307 S. sinensis ------+, presence of product by PCR of the correct size; –, no PCR product; +/-, faint PCR product detected; -*, negative but PCR product of incorrect size detected; ^if applicable; PMP1036, S. salivarius DNA was initially contaminated and was not included in PCR testing.
191
5.2.1.3.2 Testing targets on isolates by real-time PCR As bguR performed slightly better than the other targets, it was selected for further testing as a screen for S. pneumoniae and developed into a qPCR assay with the addition of a probe sequence. We then tested all three qPCR targets (lytA, piaB and bguR) against our 30 reference isolates (Table 5.13). All three gene targets performed well, with all three targets were positive for all six S. pneumoniae and negative for all 24 non-pneumococcal species. S. oligofermantans gave a Ct between 35 and 40 for all three assays, however, given that DNA was used at a concentration of 1 ng/μl, this result is considered negative. A Ct above 40 was observed for all other non-pneumococcal species for all targets.
Table 5.13. Results for potential pneumococcal-specific real-time PCR assays tested against a panel of reference isolates. ID Species, serotype^ lytA real-time piaB real-time bguR real-time PCR PCR PCR PMP6 S. pneumoniae, 5 + + + PMP7 S. pneumoniae, 19F + + + PMP85 S. pneumoniae, 23F + + + PMP849 S. pneumoniae, 1 + + + PMP1064 S. pneumoniae, 22A + + + 0173-01 S. pneumoniae, NT + + + 001-009-01 S. pseudopneumoniae - - - PMP1300 S. pseudopneumoniae - - - PMP16 S. mitis - - - PMP933 S. mitis - - - PMP934 S. mitis - - - PMP1010 S. salivarius - - - PMP935 S. mutans - - - PMP1056 S. oralis - - - PMP1301 S. infantis - - - PMP1000 S. pyogenes - - - PMP1057 S. gordonii - - - PMP1035 S. bovis - - - PMP1051 S. dysgalactiae - - - PMP994 S. agalactiae - - -
192
ID Species, serotype^ lytA real-time piaB real-time bguR real-time PCR PCR PCR PMP1053 S. sobrinus - - - PMP936 S. sanguinis - - - PMP1049 S. anginosus - - - PMP1059 S. vestibularis - - - PMP1302 S. peroris - - - PMP1303 S. australis - - - PMP1304 S. australis - - - PMP1305 S. oligofermantans - - - PMP1306 S. cristatus - - - PMP1307 S. sinensis - - - +, positive Ct<35; –, negative Ct>30.
We then tested how well bguR and piaB performed using our set of isolates (n=91) from OP samples (see section 5.2.1.2.2). Both targets were positive for the three pneumococcal isolates, bguR was equivocal for 8/88 (9%) non-pneumococcal isolates and piaB was equivocal for 10/88 (11%) non-pneumococcal isolates. As the DNA concentration was 1 ng/μl, these equivocal results were interpreted as negative reactions.
5.2.1.3.3 Testing targets on OP samples by qPCR The bguR and piaB assays were then used on the 250 OP samples that had been previously screened using the lytA qPCR assay (see section 5.2.1.1). There were 108 samples with a Ct<40 for either bguR, piaB or lytA and 142 samples that were negative (Ct≥40) for all three targets. The 108 samples were cultured on gHBA and analysed by microarray. During the culture step, colonies that had a pneumococcal appearance were subcultured on HBA and optochin sensitive isolates were serotyped. Following microarray and culture, 11 samples out of the 108 were deemed positive for S. pneumoniae.
We compared the performance of the three pneumococcal targets alone, and in combination (calculated based on results from the targets alone), for the 108 samples which were tested by microarray (Table 5.14). We also considered the results if we used a sequential approach, where all samples might be tested by one of the targets in the first
193
round and then any positive or equivocal samples be subsequently tested with the other two targets in a second round of testing.
Table 5.14. Comparison of results for 108 OP samples for pneumococcal targets alone and in combination. True False False True PPV Specificity Target gene/s1 positives positives negatives negatives (%)2 (%)3 (n) (n) (n) (n) lytA 10 29 1 68 26 70 bguR 10 68 1 29 13 30 piaB 10 20 1 77 33 79 lytA + bguR 10 14 1 83 42 86 lytA + piaB 9 11 2 86 45 89 bguR + piaB 9 10 2 87 47 90 lytA + bguR + piaB 9 8 2 89 53 92 4 lytA + bguR or piaB 10 16 1 81 38 84 4 bguR + lytA or piaB 10 23 1 74 30 76 4 piaB + lytA or bguR 9 17 2 80 35 82 A positive is defined as any Ct <40; 1Results for each target gene were determined separately, with results for combinations of targets calculated based on these results; 2Positive predictive value (PPV) is true positives / (true positives + false positives); 3Specificity is true negatives / (true negatives + false positives); 4target combination where target 1 was tested first then any sample positive by target 1 further tested by target 2 and target 3, with a final positive sample defined by Ct<40 for target 1 plus target 2 or target 3.
Although piaB performed the best out of the three targets, individually none of the targets performed well as there were a relatively large number of false positives. piaB still had a poor positive predictive value (PPV) with only a third of piaB ‘positive’ samples actually positive for pneumococci. With a PPV of 26%, the results for lytA were consistent with previous observations of a relatively high number of false positives using this assay for OP samples. All three targets had one false negative: one pneumococci positive sample
194
was positive for lytA and bguR but negative for piaB; and one pneumococci positive sample was positive for piaB but was negative for lytA and bguR.
When two targets were combined the PPV and specificity improved, however the PPV was still below 50%. When piaB was used with lytA or bguR there were two false negatives. Combining all three targets again improved both PPV and specificity. This was the only approach that resulted in a PPV above 50%, but was still relatively poor at 53%.
We also examined whether using an approach where all samples were initially screened with only one target and then any positive or equivocal samples were tested with the other two targets. If a sample was positive or equivocal by at least two targets, it was then considered positive. Screening initially with lytA or piaB were similar, with this approach performing better than using only a single target, but slightly worse than when using two targets. However, these approaches represented greater efficiency than using two targets, and represent a balance between specificity and the introduction of false negatives.
We screened the remaining 258 OP samples from the 2012 collection year with lytA and then tested only positive or equivocal samples with both bguR and piaB. We then analysed by microarray any samples that were positive or equivocal for at least two of the targets. Using this approach, 21 samples were analysed by microarray, of which 11 were positive (giving a positive predictive value of 52%). We found one sample that was equivocal for lytA and negative for both bguR and piaB, however pneumococci were detected in culture. Therefore, this approach also resulted in one false negative. If only those that were positive or equivocal by all three assays were considered positive, we would have deemed 13 samples positive of which nine were true positives, four were false positives (PPV of 69%) and three were false negatives.
In summary, our results show that screening OP samples can be improved by the addition of another pneumococcal target, although this is still only likely to give around 50% chance that pneumococci are present in the sample.
195
5.2.1.4 Examination of whether the inclusion of OP samples improves pneumococcal detection We next determined whether OP sampling had added benefits compared with NP sampling alone, in improving overall pneumococcal detection or detection of vaccine- type pneumococci. To do this we used the paired NP and OP samples (n=508 each) collected as part of NVEP in 2012, prior to vaccine introduction. For NP samples, we used our standard approach and screened with lytA qPCR before culturing positive and equivocal samples, and then performing microarray serotyping on culture positive samples. For OP samples, we screened by the three targets (lytA, bguR, piaB) described in the previous section (5.2.1.3.3, page 193) using the lytA + bguR or piaB approach to select samples for culture and microarray testing.
Table 5.15. Detection of pneumococci in adult caregiver samples (n=508) by isolation site.
NP
Positive Negative Total Positive 8 14 22 OP Negative 45 441 486 Total 53 455 508 NP, nasopharyngeal samples; OP, oropharyngeal samples. Samples defined as positive if pneumococci detected by culture or microarray.
Overall there were 67 adults carrying pneumococci (Table 5.15); for 45 of the 67 (67%) carriers, pneumococci were only detected in the nasopharynx (and not the oropharynx), 14 (21%) were only carrying pneumococci in the oropharynx (and not the nasopharynx) and for eight (12%) of the adults, pneumococci were detected in both sites (Figure 5.12).
Sampling the nasopharynx alone resulted in a carriage rate of 10.4%. In comparison, sampling both the nasopharynx and oropharynx resulted in a carriage rate of 13.2% (Table 5.16) (p=0.206). Therefore, OP sampling (in addition to NP sampling) did not significantly improve detection of pneumococci; just over three-quarters of all pneumococci detected could be found in the nasopharynx alone.
196
The density of pneumococcal carriage was higher in the nasopharynx compared with the oropharynx (median density 2.9 x 104 vs. 5.3 x 103 genome equivalents/ml; p=0.008; Figure 5.13).
Table 5.16. Comparison of pneumococcal detection in adult caregivers (n=508) following nasopharyngeal sampling alone or with the addition of oropharyngeal sampling. NP and OP Pneumococci NP sampling p-value* sampling Overall 53 (10.4) 67 (13.2) 0.206 Vaccine type 12 (2.4) 17 (3.4) 0.452 Non-vaccine type 39 (7.7) 48 (9.4) 0.370 Values are numbers (percentages); NP, nasopharyngeal samples; OP, oropharyngeal samples; *p-values calculated using Fisher’s exact test.
OP sampling also did not improve detection of vaccine type carriage (p=0.452), with NP sampling alone resulting in a vaccine type carriage rate of 2.4% compared with 3.4% when sampling the both the nasopharynx and the oropharynx (Table 5.16). Approximately 70% of vaccine type carriage could be detected by sampling the nasopharynx alone (Figure 5.12). The results were similar for non-vaccine type carriage with no significant improvement in detection with the addition of OP sampling (p=0.370).
We examined the serotypes from each site (Figure 5.14), for OP samples only those serotypes that could be confirmed as coming from pneumococci were used in our analysis. Owing to the complexity of capsule genes present in OP samples, distinguishing between pneumococci from some unencapsulated lineages (those containing aliB genes) and non-pneumococci containing similar genes (366) was generally not possible, which meant that the true carriage of pneumococci from unencapsulated lineages in the oropharynx was unable to be determined. Serotype 16F was one of the most common serotypes in both the nasopharynx (8.2%) and oropharynx (6.7%). In the OP samples serotype 5 was also common (6.7%) but was not present in NP samples. In NP samples serotypes 15B (6.6%), 34 (4.9%) and 8 (4.9%) were also common, with the latter only present in NP samples. We did not detect significant differences in serotype composition
197
between NP and OP serotypes for any individual serotype, however there was some evidence that serotype 5 carriage was higher in the OP (p=0.074).
Overall, 44 serotypes were detected, with 13 (29.5%) serotypes only detected in the nasopharynx, 13 (29.5%) serotypes only detected in the oropharynx and 18 (40.1%) serotypes detected in both sites. When pneumococci were detected in a sample, multiple serotype carriage was more common for OP samples than NP samples (15/22 (68%) vs. 9/51 (18%); p<0.001).
198
A OP only 21%
Both NP only 12% 67%
Overall pneumococci n = 67
B OP only 29%
NP only 47%
Both 24%
Vaccine type pneumococci n = 17
OP only C 19%
Both NP only 8% 73%
Non-vaccine type pneumococci n = 48 Figure 5.12. Site of isolation (nasopharynx, oropharynx or both) for overall (A), vaccine type (B) and non-vaccine type (C) adult pneumococcal carriage in 2012. NP only, percentage of pneumococcal carriers following isolation only from the nasopharynx; OP only, percentage of pneumococcal carriers following isolation only from the oropharynx; Both sites, percentage of pneumococci isolated from both the nasopharynx and oropharynx.
199
p=0.008 10 8 l 7 m
/ 10 s t 6 n 10 e l
a 5
v 10 i
u 4 q 10 e
e 10 3 m
o 2 n 10 e g 10 1 10 0 NP OP
Figure 5.13. Pneumococcal carriage density (genomic equivalents/ml) in the nasopharynx (red) compared with the oropharynx (blue). Error bars are median and interquartile range. There was a significant difference in carriage density between the nasopharynx and the oropharynx (p=0.008) as calculated by the Mann-Whitney test.
200
Figure 5.14. Serotype carriage in the nasopharynx (red) and oropharynx (blue) as percentage of total pneumococcal serotypes in each sample type. NT indicates pneumococci from unencapsulated lineages.
201
We also compared the serotyping results for the eight participants that were carrying pneumococci in both the nasopharynx and oropharynx (Table 5.17). In all eight participants, the same serotype was present in both sites, however in five OP samples and one NP sample additional serotypes were detected. In one participant, 7C was present in the oropharynx whereas 7B was present in the nasopharynx. Where it was possible to isolate pneumococci from OP samples during the culture step, latex agglutination serotyping was performed which was able to confirm several of the serotyping results obtained from microarray.
Table 5.17. Microarray serotyping results for the eight participants carrying pneumococci in both the nasopharynx and oropharynx. Participant NP serotypes OP serotypes 1 17F, 35F 17F 2 6B 6B, 16A 3 3, 7B 3, 7C, 16A, 41A, 48 4 18C 16F, 18C, 24A, 5 5 1 1, 20B 6 15B 15B 7 17F 17F, 35A 8 19F 19F NP, nasopharyngeal; OP, oropharyngeal; Serotypes that were present in both sites are shown in bold; OP serotypes that were confirmed by latex agglutination are underlined.
In summary, the use of OP samples did not significantly improve detection of pneumococci. Although there were some differences in serotype composition, the majority of the serotypes detected were present in the NP (~70%). Furthermore, OP samples are also extremely challenging to process and analyse. As such, only NP samples were used to examine the indirect effect of pneumococcal vaccination in adult caregivers in Fiji.
202
5.2.2 Indirect effect of infant pneumococcal vaccination on pneumococcal carriage in adult caregivers
Annual cross-sectional surveys were conducted before the introduction of PCV10 into the infant immunisation program (2012) and continuing to three years after PCV10 introduction (2013, 2014 and 2015) in Fiji. Each year, approximately 500 NP swabs were collected from adult caregivers. Here, we examined whether PCV10 vaccination in infants had an indirect effect on NP pneumococcal carriage in adult caregivers up to 3 years post-vaccine introduction.
To give context for the potential for indirect effects of vaccination in adult caregivers, among NVEP participants the proportion of children vaccinated with two doses or more of PCV10 in the 12-23 month age group was 2.2% in 2013, 94.6% in 2014 and 99.9% in 2015. In the general age-eligible population, PCV10 coverage for 2 doses or more was 89.9% in 2014 and 90.3% in 2015 (no data available for 2013).
5.2.2.1 Participant characteristics Participant characteristics were collected at the time of swab collection. There were several characteristics that differed significantly between pre-vaccine introduction (2012) and post-vaccine introduction (2013, 2014 or 2015) years (Table 5.18): the proportion of female participants; the age of the participants; the number of children under 18 years of age in the household; the household income and the proportion of participants with low income; the proportion of participants with antimicrobial use in the previous two weeks; and the proportion of participants with allergic rhinitis, cough and diabetes.
In 2014 and 2015 there were significantly fewer female participants compared with 2012 (86.6% and 86.6% compared with 91.3%, p=0.017 and p=0.021, respectively). In 2015, participants were slightly older than in 2012 (p=0.018) with a median age of participants in 2012 of 30 years old (IQR 25-38) compared with 32 years old (IQR 25-44) in 2015. There was a significant difference in the number of children under 18 years of age in the household between 2012 and 2013 (p=0.002), 2014 (p<0.001) and 2015 (p=0.025), however the median number of children was the same between 2012 and 2015 (3 children; IQR 2-4) and only slightly different in 2013 and 2014 (2 children; IQR 1-3). Compared with the proportion of participants with a low income in 2012 (70.7%) there were
203
significantly fewer participants with a low income in 2014 (54.9%; p<0.001) and in 2015 (50.5%; p<0.001). This corresponded to higher household income particularly in 2015 (p<0.001) where the median was F$170 per week (IQR $120-$200) compared with F$150 per week (IQR $100-$185) in 2012. There were no participants with allergic rhinitis in 2014 and 2015 compared with 2012 where there were five participants (1.0%) and although statistically significant (p=0.030 and p=0.031, respectively), this difference is unlikely to have a significant impact on carriage overall. The proportion of participants with a cough at the time of swab collection was lower in 2013 (10.4%; p=0.012) and in 2014 (10.1%, p=0.007) compared with 2012 (15.8%). Compared with 7.5% of participants in 2012, the proportion of participants that had used antimicrobials in the previous two weeks differed significantly in subsequent years (2013, 3.7%, p=0.010; 2014, 2.1%, p<0.001; 2015, 3.1%, p=0.002). The proportion of participants diagnosed with diabetes was lower in 2013 compared with 2012 (2.0% vs. 4.5%, p=0.022).
It should also be noted that the youngest ‘adult’ participant in our study was 13 years of age and the oldest was 112 years of age. However, these age extremes were exceptions, and in each study year over 96% of participants were between the ages of 18 and 65 years of age.
204
Table 5.18. Participant characteristics of adult caregivers in NVEP for each study year. Parameter at swab collection 2012 2013 2014 2015 (n=508) (n=511) (n=516) (n=509) Female 464 (91.3) 449 (87.9) 447 (86.6)* 441 (86.6)* Median age [y] (IQR) 30 (25 - 38) 32 (25 - 40) 31 (25 - 39) 32 (25 - 44)* Ethnicity iTaukei 310 (61.0) 309 (60.5) 296 (57.3) 304 (59.7) Indo-Fijian 196 (38.6) 202 (39.5) 215 (41.7) 203 (39.9) Other 2 (0.4) 0 (0.0) 5 (1.0) 2 (0.4) Location of residence Rural 239 (47.0) 257 (50.3) 216 (41.9) 242 (47.5) Urban 159 (31.3) 165 (32.3) 174 (33.7) 178 (35.0) Peri-urban 110 (21.7) 89 (17.4) 126 (24.4) 89 (17.5) Median number of children living in the home (IQR) <18 y 3 (2 - 4) 2 (1 - 3)* 2 (1 - 3)* 3 (2 - 4)* < 5 y 1 (1 - 2) 1 (1 - 2) 1 (1 - 2)4 1 (1 - 2) Low family income [ 205 Parameter at swab collection 2012 2013 2014 2015 (n=508) (n=511) (n=516) (n=509) Antimicrobial use in past two weeks2 38 (7.5) 19 (3.7)* 11 (2.1)* 16 (3.1)* Exposure to household cigarette smoke 276 (54.3) 276 (54.0) 269 (52.1) 296 (58.2) Diagnosed with diabetes 23 (4.5) 10 (2.0)* 19 (3.7)6 24 (4.7) Diagnosed with heart disease 3 (0.6) 2 (0.4) 1 (0.2)6 6 (1.2) Diagnosed with cancer 1 (0.2) 1 (0.2) 0 (0.0)6 0 (0.0) Diagnosed with lung disease 0 (0.0) 0 (0.0) 0 (0.0)6 1 (0.2) Values are numbers (percentages) unless otherwise indicated; IQR, interquartile range; URTI, upper respiratory tract infection; *Significantly different from 2012 (p<0.05) as calculated by Fisher’s Exact test for categorical parameters or the Mann-Whitney test for continuous parameters 1Preliminary report: poverty and household incomes in Fiji in 2008 – 2009, Suva, Fiji; 2Self-reported; some data was unavailable for all participants and the updated number of participants are 3n=433; 4n=513; 5n=510; 6n=511. 206 We also considered any changes in participant characteristics that occurred in the post- vaccination years (2013, 2014 and 2015) compared to pre-vaccination (2012) after stratifying by ethnicity (Table 5.19). In iTaukei participants there were significant differences in: the proportion of participants that were female; the number of children <18 years of age and <5 years of age in the household; the household income and subsequently the proportion of participants with low income; the proportion of participants with antimicrobial use in the previous two weeks; and the proportion of participants with a cough. In 2014 and 2015 there were significantly fewer female iTaukei participants compared with 2012 (89.5% and 89.5% compared with 95.8%, p=0.004 and p=0.003). There was a significant difference in the number of children under 18 years of age in iTaukei households between 2012 and post-vaccine years 2014 (p<0.001) and 2015 (p=0.002), however the median number of children was the same between 2012 and 2015 (3 children; IQR 2-4 in 2012 and IQR 2-5) and only slightly different in 2014 (2 children; IQR 2-3). The median number of children under 5 years of age in iTaukei households was the same between 2012 and 2015 (2 children; IQR 1-2 children), however using the Mann-Whitney test there was a significant difference between these two years (p=0.015). Compared with the proportion of iTaukei participants with a low income in 2012 (73.0%) there were significantly fewer participants with a low income in 2014 (50.9%; p<0.001) and in 2015 (53.3%; p<0.001). This corresponded to higher iTaukei household income in 2014 and 2015 (p<0.001) where the median was F$140 per week (IQR $100-$180) in 2012 compared with F$170 per week (IQR F$100-200) in 2014 and F$160 (IQR F$100-200) in 2015. The proportion of iTaukei participants with a cough at the time of swab collection was lower in 2013 (10.0%; p=0.005) and in 2014 (11.8%, p=0.040) compared with 2012 (18.1%). Compared with 8.7% of participants in 2012, the proportion of iTaukei participants that had used antimicrobials in the previous two weeks differed significantly in subsequent years (2013, 3.9%, p=0.020; 2014, 2.4%, p=0.001; 2015, 3.6%, p=0.011). In Indo-Fijian participants there were significant differences between 2012 and subsequent years in: the age of the participants; the proportion of participants from a peri- urban area; the number of children <18 years of age in the household; and the household income and subsequently the proportion of participants with low income. 207 In 2013 and 2015 Indo-Fijian participants were slightly older than in 2012 (p=0.040 and p=0.001, respectively) with a median age of Indo-Fijian participants in 2012 of 29 years of age (IQR 24-39 years of age) compared with 33 years of age (IQR 25-44 years of age) in 2013 and 33 years of age (IQR 26-49 years of age) in 2015. There were fewer Indo- Fijian participants living in a peri-urban location in 2013 compared with 2012 (17.3% vs. 26.5%; p=0.029). There was a significant difference in the number of children under 18 years of age in Indo-Fijian households between 2012 and 2013 (p=0.017) and 2014 (p=0.024), however the median number of children was the same between 2012 and 2013 (2 children; IQR 1-3 children in 2012 and IQR 1-2 children in 2013) and between 2012 and 2014 (2 children; IQR 1-2 children). Compared with the proportion of Indo-Fijian participants with a low income in 2012 (67.6%) there were significantly fewer participants with a low income in 2013 (50.0%; p=0.001) and in 2015 (46.8%; p<0.001). The median Indo-Fijian household income was higher in 2015 (F$180 per week, IQR F$105-200 per week) compared with 2012 (F$150 per week, IQR F$142-200; p=0.001 per week). 208 Table 5.19. Participant characteristics of adult caregivers in NVEP stratified by ethnicity (iTaukei and Indo-Fijian) for each study year. 2012 2013 2014 2015 Parameter at swab collection iTaukei Indo-Fijian iTaukei Indo-Fijian iTaukei Indo-Fijian iTaukei Indo-Fijian (n=310) (n=196) (n=309) (n=202) (n=296) (n=215) (n=304) (n=203) Female 297 (95.8) 165 (84.2) 285 (92.2) 164 (81.2) 265 (89.5)* 177 (82.3) 272 (89.5)* 168 (82.8) 31 29 30 33 31 32 31 33 Median age [y] (IQR) (26-38) (24-39) (25-37) (25-44)* (26-37) (25-47) (25-40) (26-49)* Location of residence Rural 153 (49.4) 85 (43.4) 156 (50.5) 101 (50.0) 134 (45.3) 82 (38.1) 140 (46.1) 102 (50.2) Urban 99 (31.9) 59 (30.1) 99 (32.0) 66 (32.7) 96 (32.4) 75 (34.9) 112 (36.8) 64 (31.5) Peri-urban 58 (18.7) 52 (26.5) 54 (17.5) 35 (17.3)* 66 (22.3) 58 (27.0) 52 (17.1) 37 (18.2) Median number of children living in the home (IQR) <18 y 3 (2-4) 2 (1-3) 3 (2-4) 2 (1-2)* 2 (2-3)* 2 (1-2)* 3 (2-5)* 2 (1-3) < 5 y 2 (1-2) 1 (1-2) 2(1-2) 1 (1-1) 2 (1-2) 1 (1-2) 2 (1-2)* 1 (1-2) Low family income 184 (73.0)3 121 (67.6)4 244 (79.0) 101 (50.0)* 149 (50.9)5* 131 (61.8)7 162 (53.3)* 95 (46.8)* [ (IQR) (100-180)3 (105-200)4 (100-160) (100-200) (100-200)5* (100-200)7 (100-200)* (142-200)* Current URTI symptoms Rhinorrhoea 47 (15.2) 9 (2.9) 30 (9.7) 14 (6.9) 37 (12.5) 11 (5.1) 38 (12.5) 18 (8.9) 209 2012 2013 2014 2015 Parameter at swab collection iTaukei Indo-Fijian iTaukei Indo-Fijian iTaukei Indo-Fijian iTaukei Indo-Fijian (n=310) (n=196) (n=309) (n=202) (n=296) (n=215) (n=304) (n=203) Allergic rhinitis 5 (1.6) 0 (0.0) 5 (1.6) 1 (0.5) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) Cough 56 (18.1) 23 (7.4) 31 (10.0)* 22 (10.9) 35 (11.8)* 17 (7.9) 55 (18.1) 27 (13.3) Antimicrobial use in past 2 27 (8.7) 11 (3.5) 12 (3.9)* 7 (3.5) 7 (2.4)* 4 (1.9) 11 (3.6)* 5 (2.5) weeks2 Exposure to household cigarette 174 (56.1) 102 (52.0) 169 (54.7) 107 (53.0) 164 (55.4) 101 (47.0) 177 (58.2) 117 (57.6) smoke Diagnosed with diabetes 8 (2.6) 15 (7.7) 2 (0.6) 8 (4.0) 10 (3.4)6 9 (4.2)7 4 (1.3) 20 (9.9) Diagnosed with heart disease 1 (0.3) 2 (1.0) 1 (0.3) 1 (0.5) 0 (0.0)6 1 (0.5)7 0 (0.0) 6 (3.0) Diagnosed with cancer 1 (0.3) 0 (0.0) 1 (0.3) 0 (0.0) 0 (0.0)6 0 (0.0)7 0 (0.0) 0 (0.0) Diagnosed with lung disease 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0)6 0 (0.0)7 1 (0.3) 0 (0.0) Values are numbers (percentages) unless otherwise indicated; IQR, interquartile range; URTI, upper respiratory tract infection; *Significantly different from 2012 (p<0.05) as calculated by Fisher’s Exact test for categorical parameters or the Mann-Whitney test for continuous parameters 1Preliminary report: poverty and household incomes in Fiji in 2008 – 2009, Suva, Fiji; 2Self-reported; Some data was unavailable for all participants and the updated number of participants are 3n=252; 4n=179; 5n=293; 6n=294; 7n=212. 210 5.2.2.2 Carriage rates and density To assess any indirect effect of infant PCV10 vaccination on NP pneumococcal carriage, we examined overall pneumococcal carriage rates as well as vaccine type and non- vaccine type carriage rates (Figure 5.15). Overall pneumococcal carriage rates were 10.4%, 13.2%, 6.6% and 7.8% in 2012, 2013, 2014 and 2015, respectively. Vaccine type carriage rates were 2.4%, 3.3%, 1.6% and 0.8% and non-vaccine type carriage rates were 7.7%, 9.4%, 5.0% and 6.7% in the same years, The pneumococcal carriage rate in adults declined post-vaccination with a significant overall trend for declining carriage rates (Chi-square test for trend, p=0.013). For comparisons between individual years, there were significant differences between 2012 and 2014 (p=0.033), between 2013 and 2014 (p=0.001), and between 2013 and 2015 (p=0.006). However, there was no significant difference between carriage rates in 2012 and three years postvaccine introduction in 2015 (p=0.159). There was a significant trend for a reduced vaccine type carriage rate between 2012 and 2015 (Chi-squared test for trend p=0.018), with a significantly lower vaccine type carriage rate in 2015 compared with 2012 (p=0.047). There was also a significant difference in vaccine type carriage rates observed between 2013 (which had a slightly higher carriage rate compared with 2012) and 2015 (p=0.004). There were no significant changes in non-vaccine type carriage rates following vaccine introduction with no overall trend nor individual differences from baseline non-vaccine type carriage. There was a lower non-vaccine type carriage rate in 2014 compared with 2013 (p=0.008), and there was some evidence that the non-vaccine type carriage rate was slightly lower in 2014 compared with 2012 (p=0.095) but this was not significant. 211 A ) % p=0.033 ( 20 e a t r 15 g e a r i 10 a r c l a l 5 e r v O 0 2012 2013 2014 2015 B ) 20 % ( e 15 r a t e p=0.047 a g 10 c a r i 5 T V 0 2012 2013 2014 2015 C ) 20 % ( e t a 15 r e g a i 10 r r a c 5 T V N 0 2012 2013 2014 2015 Figure 5.15. Overall (A), vaccine type (B) and non-vaccine type (C) pneumococcal carriage rates in adult caregivers before (2012) and after (2013, 2014 and 2015) infant PCV10 introduction. Shown are 95% confidence intervals. VT, vaccine type; NVT, non- vaccine type. Significant differences were observed in overall carriage rates over time (p=0.013, Chi-square test for trend), with differences observed between 2012 and 2014 (p=0.033) as calculated by Fisher’s exact test. There were significant differences in vaccine type carriage rates over time (p=0.018, Chi-square test for trend), with differences observed between 2012 and 2015 (p=0.047) as calculated by Fisher’s exact test. No significant differences in non-vaccine type carriage rates were observed post-vaccine introduction. 212 Given several participant characteristics were different from 2012, we looked at both unadjusted prevalence ratios and prevalence ratios adjusting for sex, age, weekly household income, presence of a cough, presence of allergic rhinitis and antimicrobial use in the previous two weeks (Table 5.20). Table 5.20. Prevalence and unadjusted and adjusted ratios for overall, vaccine type and non-vaccine type pneumococcal carriage in adult caregivers comparing pre- (2012) with post-PCV10 introduction (2013, 2014 and 2015). Unadjusted Adjusted^ Prevalence prevalence ratio prevalence ratio (%) (95% CI) (95% CI) All pneumococci 2012 10.43 2013 13.16 1.26 (0.90, 1.76) 1.41 (0.98, 2.03) 2014 6.59 0.63 (0.42, 0.95)* 0.76 (0.49, 1.18) 2015 7.83 0.75 (0.51, 1.11) 0.87 (0.58, 1.32) VT pneumococci 2012 2.37 2013 3.35 1.41 (0.68, 2.92) 1.74 (0.76, 4.01) 2014 1.55 0.65 (0.27, 1.59) 0.82 (0.31, 2.18) 2015 0.79 0.33 (0.11, 1.02) 0.42 (0.13, 1.39) NVT pneumococci 2012 7.71 2013 9.43 1.22 (0.82, 1.83) 1.38 (0.90, 2.11) 2014 5.04 0.66 (0.41, 1.06) 0.81 (0.49, 1.34) 2015 6.68 0.87 (0.56, 1.36) 1.01 (0.63, 1.61) CI, confidence interval; VT, vaccine type; NVT, non-vaccine type; Prevalence ratios and 95% CI calculated using log binomial regression models ^ratio adjusted for age, sex, weekly household income, presence of a cough, presence of allergic rhinitis and antimicrobial use in the previous two weeks, 81 observations were omitted due to no income data; *statistically significant, p=0.029. Prior to adjustment, the only significant difference from 2012 was in 2014 where the overall carriage of pneumococci was lower with a prevalence ratio of 0.63 (95% CI 0.42, 0.95; p=0.029). Following adjustment this was no longer significant with an adjusted prevalence ratio of 0.76 (0.49, 1.18; p=0.222). There was some evidence vaccine type 213 carriage was lower in 2015 (compared with 2012) with a prevalence ratio of 0.33 (95% CI 0.11, 1.02; p=0.055), however after adjustment this was also no longer the case with a prevalence ratio of 0.42 (95% CI 0.13, 1.39; p=0.157). The inclusion of household income in the model resulted in the omission of 81 observations which may have affected results. Without the inclusion of income in the adjusted model, the prevalence ratio of vaccine type pneumococci was similar to the unadjusted figure in 2015 (prevalence ratio of 0.33; 95% CI 0.11, 1.02; p=0.054). We examined serotype-specific carriage in the NP before and after vaccine introduction (Figure 5.16). Overall, we identified 57 pneumococcal serotypes in 2012, 71 in 2013, 37 in 2014 and 40 in 2015. Pneumococci from unencapsulated lineages (marked as NT) were the most common pneumococci identified in every year between 2012 and 2015 (19.3%, 15.5%, 21.6% and 12.5%, respectively). Apart from pneumococci with unencapsulated lineages there was little consistency in serotype composition over the study period. Serotype 16F was one of the most common in 2012 (8.8%) and 2013 (9.9%) but was less common in 2014 (2.7%) and not present in our participants in 2015. Similarly, 15B was common in 2012 (7.0%), 2013 (5.6%) and 2014 (5.4%) but was not present in 2015. Serotype 7C was absent in 2012, less common in 2013 (1.4%) and one of the more commonly identified serotypes in 2014 (5.4%) and 2015 (7.5%). From 2012 to 2015, only serotypes 3 (1.8%, 4.2%, 5.4% and 5.0%, respectively), 6C (1.8%, 1.4%, 2.7% and 5.0%, respectively), 11F-like (3.5%, 1.4%, 5.4% and 5.0%, respectively), 13 (1.8%, 8.5%, 2.7% and 7.5%, respectively) and 34 (5.3%, 2.8%, 2.7% and 2.5%, respectively) were present in every year in our participants. The ‘11F-like’ serotype is genetically similar to 11F, however some of the 11F genes are absent or divergent from a typical 11F capsule gene locus. We have found these typically type as serotype 11A using latex agglutination and quellung (367). There was also little consistency in the composition of the vaccine serotypes over the study period. However, the percentage of total serotypes that were vaccine type and the number of vaccine serotypes did appear to decrease. In 2012, 12 out of 57 (21.1%) of serotypes detected were vaccine type compared with four out of 40 (10.0%) in 2015, although this was not significant (p=0.175). In 2012 eight (1, 4, 6B, 9V, 14, 18C, 19F and 23F) of the 10 PCV10 serotypes were present in our participants, in 2013 six (1, 6B, 7F, 214 9V, 19F and 23F) were present, in 2014 three (9V, 14 and 19F) were present and by 2015 only two (14 and 23F) of the ten PCV10 serotypes were present. We also examined the rates of multiple serotype carriage which were 1.4%, 1.2%, 0.6% and 0.4% in 2012, 2013, 2014 and 2015, respectively. There was some evidence of a trend toward declining multiple serotype carriage over the study period (Chi-square test for trend, p=0.055), however there were no significant differences between 2012 and later years, including between 2012 and 2015 (p=0.108). 215 Figure 5.16. Vaccine and non-vaccine types as a percentage of total pneumococcal serotypes identified in the NP before PCV10 introduction in 2012 and after PCV10 introduction in 2013, 2014 and 2015. Vaccine types are highlighted in red and non- vaccine types are shown in black. NT indicates pneumococci from unencapsulated lineages. Note there was no carriage of serotype 5 (included in PCV10) in the participants over the study period. 216 We also examined whether there were any changes in overall pneumococcal density following PCV10 introduction (Figure 5.17.A). There was a significant difference in pneumococcal density between 2012 and 2015 (p=0.003), with a slightly higher density in 2015 (median of 8.53 x 104 genomic equivalents/ml, IQR 2.33 x 104 – 4.87 x 105 genomic equivalents/ml) compared with 2012 (median of 2.94 x 104 genomic equivalents/ml, IQR of 5.89 x 103 – 1.25 x 105 genomic equivalents/ml). This difference is largely due to a reduction in pneumococci carried at a low density (Figure 5.17.B), as the proportion of pneumococcal carriers with a density below 104 genomic equivalents/ml was significantly lower in 2015 compared with 2012 (5.0% vs. 32.1%, p=0.001). The proportion of pneumococcal carriers with a high carriage density (>106 genomic equivalents/ml) almost doubled between 2012 and 2015 (7.6% vs. 15.0%), however this was not significant (p=0.318; Chi-square test for trend, p=0.194). 217 p=0.003 A 9 l 10 m / 10 8 s t n 7 e 10 l a v 6 i 10 u q 5 e 10 e 4 m 10 o n 3 e 10 g 10 2 2012 2013 2014 2015 B low density carriage (<104 genome equivalents/ml) high density carriage (>106 genome equivalents/ml) s e r p=0.001 r i 50 c a r 40 c a l 30 c o o m 20 u e n 10 p f o 0 % 2012 2013 2014 2015 Figure 5.17. Pneumococcal carriage density in adult caregivers before (2012) and after (2013, 2014, 2015) PCV10 introduction. (A) overall pneumococcal density. Error bars are median and interquartile range. There was a difference in density over time (p=0.019, Kruskal-Wallis test) and a significant difference in overall density between 2012 and 2015 (p=0.008) as calculated by the Mann-Whitney test. (B) Proportion (%) of pneumococcal carriers with low (<104 genomic equivalents/ml) and high (>106 genomic equivalents/ml) pneumococcal density. Error bars are 95% confidence intervals. There was a significant difference in the proportion of pneumococcal carriers with low carriage density between 2012 and 2015 (p=0.001) as calculated by Fisher’s exact test. 218 5.2.2.3 Ethnicity We also examined the carriage rates and densities by ethnic group (iTaukei and Indo- Fijian) as differences in pneumococcal carriage rates between these groups have been observed in children (chapter 3, chapter 4 (176, 228)). We found that iTaukei adults had significantly higher carriage rates of pneumococci compared with Indo-Fijian adults (Figure 5.18) in 2012 (15.5% vs. 2.6%, p<0.001), 2013 (18.8% vs. 4.5%, p<0.001), 2014 (9.5% vs. 2.9%, p=0.004) and 2015 (11.2% vs. 2.5%, p=0.003). Considering carriage in iTaukei adults alone (Figure 5.19.A), there was a significant trend for reduced overall pneumococcal carriage over the study period (Chi-square test for trend, p=0.012). In 2014, pneumococcal carriage in iTaukei adults was significantly lower than in 2012 (9.5% vs. 15.5%; p=0.027), however the difference between 2012 and 2015 was smaller and not significant (15.5% vs. 11.2%; p=0.124). Vaccine type carriage also declined in iTaukei adults over the study period (Chi-square test for trend, p=0.003) and vaccine type carriage was significantly lower in 2015 compared with 2012 (0.7% vs. 3.9%; p=0.012). We also calculated unadjusted and adjusted prevalence ratios for iTaukei carriage and found a significant difference in the unadjusted prevalence ratio for 2014 compared with 2012 (prevalence ratio 0.61; 95% CI 0.39, 0.95; p=0.028). However, after adjustment this was no longer significant with (0.69; 95% CI 0.44, 1.10; p=0.119) or without (0.67; 95% CI 0.43, 1.04; p=0.072) the inclusion of household income in the model. Compared with 2012, vaccine type carriage was lower in 2015 with an unadjusted prevalence ratio of 0.17 (95% CI 0.04, 0.75; p=0.020), this was also true after adjustment with (0.20; 95% CI 0.04, 0.92; p=0.039) and without (0.16; 95% CI 0.04, 0.73; p=0.017) the inclusion of household income in the model. Considering carriage in Indo-Fijian adults alone (Figure 5.19.B), there were no significant differences in overall, vaccine type and non-vaccine type pneumococcal carriage rates from pre- to post-vaccine introduction. Vaccine type carriage in Indo-Fijian adults was low with no vaccine type carriage in 2012 and 2013, and 0.9% and 1.0% vaccine type carriage in 2014 and 2015, respectively. These results were consistent with the unadjusted 219 and adjusted prevalence ratio analysis for Indo-Fijian adults, where no changes were observed in overall, vaccine type and non-vaccine type prevalence from pre- to post- vaccine introduction. We also compared the density of pneumococcal carriage between iTaukei and Indo-Fijian adults and found no significant difference in any of the study years (Figure 5.20) (all p>0.05). For iTaukei adults, there was evidence that pneumococcal density increased over time (Kruskal-Wallis test, p=0.026) with density significantly higher (p=0.003) in 2015 (median density of 1.37 x 105 genomic equivalents/ml, IQR 2.50 x 104 – 7.90 x 105 genomic equivalents/ml) compared with 2012 (median density of 1.23 x 104 genomic equivalents/ml, IQR 5.87 x 103 – 1.08 x 105 genomic equivalents/ml). As with the overall carriage density, there were significantly fewer iTaukei adults with low carriage density in 2015 compared with 2012 (5.9% vs. 31.3%, p=0.006). 220 p<0.001 25 p<0.001 iTaukei Indo-Fijian 20 p<0.001 ) p=0.004 % ( e t 15 a r e g i a 10 r c a r 5 0 012 013 014 015 2 2 2 2 Figure 5.18. Overall pneumococcal carriage in adult caregivers by ethnicity before (2012) and after (2013, 2014 and 2015) PCV10 introduction. Significant differences were observed between ethnic groups in 2012 (p<0.001), 2013 (p<0.001), 2014 (p=0.003) and 2015 (p<0.001) as calculated by Fisher’s exact test. 221 A p=0.027 2012 25 2013 2014 ) 20 % 2015 ( e t 15 a r e p=0.012 g a i 10 r r a c 5 0 Overall vaccine type non-vaccine type B 25 2012 2013 ) 20 % ( 2014 e t 15 a 2015 r e g i a 10 r a r c 5 0 Overall vaccine type non-vaccine type Figure 5.19. Overall, vaccine type and non-vaccine type pneumococcal carriage in iTaukei (A) and Indo-Fijian (B) adult caregivers before (2012) and after (2013, 2014 and 2015) infant PCV10 introduction. Significant differences were observed for overall pneumococcal carriage in iTaukei adults over time (p=0.013, Chi-square test for trend), with differences observed between 2012 and 2014 (p=0.027) as calculated by Fisher’s exact test. For vaccine type carriage, significant differences were observed in iTaukei adults over time (p=0.003, Chi-square test for trend), with differences observed between 2012 and 2015 (p=0.012). No significant differences were observed in Indo-Fijian adults over the study period. 222 2012 p=0.003 2013 l 10 9 m 2014 / s 8 t 10 2015 n iTaukei e 7 l 10 Indo-Fijian a v i 6 u 10 q e 5 10 e m 10 4 o n e 10 3 g 10 2 2012 2013 2014 2015 Figure 5.20. Pneumococcal carriage density in adult caregivers by ethnicity before (2012) and after (2013, 2014 and 2015) PCV10 introduction. Error bars are median and interquartile range; no significant differences were observed between the two ethnic groups, as calculated by the Mann-Whitney test. There was a difference in density for iTaukei adults over time (p=0.026, Kruskal-Wallis test), with a difference observed between 2012 and 2015 (p=0.003). 223 5.2.2.4 Non-pneumococcal species and serotypes in the nasopharynx During sample testing, we noticed an apparent increase in the number of NP samples that were more typical of our OP samples in later years. These samples had growth of non- pneumococcal species on gHBA plates and non-pneumococcal serotypes detected in microarray. We analysed this data to determine whether there were any significant changes in the biology of carriage in the nasopharynx of adults pre- (2012) and post- vaccine introduction (2013, 2014 and 2015) (Figure 5.21). Of the NP samples analysed by microarray, there was a significant change in the proportion of samples containing non-pneumococcal ‘serotypes’ over the study period (p=0.045, Chi-square test for trend); this was also significant when the proportion of samples with non-pneumococcal ‘serotypes’ in 2012 was compared with 2015 (4.1% vs. 19.5%, p=0.039). We also examined the proportion of samples with non-pneumococcal species, however some of the samples in 2012 were processed using an older version of microarray serotyping which lacked the StrepID component. These samples were excluded, so 2012 has fewer samples and is therefore more difficult to compare with subsequent years. However, we found some evidence that the proportion of samples analysed by microarray containing non-pneumococcal species increased between 2012 and 2015 (6.25% vs. 24.39%, p=0.056). 224 'Serotypes' from non-pneumococcal species 30 y Non-pneumococcal species a r s r p=0.039 a e 25 l o p r c m i 20 a m s l y a b 15 t o d t e f s 10 o y l a % n 5 a 0 2012 2013 2014 2015 Figure 5.21. Percentage of samples with non-pneumococcal bacterial species and serotypes from non-pneumococcal bacterial species of the total samples run on microarray before (2012) and after (2013, 2014, 2015) PCV10 introduction. There was a significant difference over time in samples containing serotypes from non-pneumococcal species (p=0.045, Chi-square test for trend), with a significant difference observed between 2012 and 2015 (p=0.039), as calculated by Fisher’s exact test. No significant differences were observed in the detection of non-pneumococcal bacterial species (p>0.05, as calculated by Fisher’s Exact test). 225 5.3 Discussion In this study, we examined the indirect effects of infant PCV10 vaccination on carriage of pneumococci in adult caregivers in Fiji as part of the NVEP. Given that there are mixed findings regarding the utility of including OP samples for adult carriage studies, this carriage study also provided an opportunity to address the need for OP samples in addition to NP samples using the best molecular identification and serotyping methods available (154). Firstly, we examined whether NP and OP samples yielded different results when using molecular methods for the identification and serotyping of S. pneumoniae. We found that OP samples were generally more complex than NP samples. This complexity was evident in both the identification (i.e. lytA qPCR) results and in the molecular serotyping (i.e. DNA microarray) results. Compared with NP samples, a larger proportion of OP samples that were positive or weakly positive (equivocal) for the lytA gene using qPCR were subsequently negative for S. pneumoniae. Samples where pneumococci are detected by molecular methods may not have growth when cultured for a variety of reasons, for example some S. pneumoniae may be non-viable after freezing and transport, or due to other factors like antibiotic exposure. However, these factors would be similar for NP and OP samples. In contrast, we found that in OP samples only 78% of lytA positive samples and 9% lytA equivocal samples contained pneumococci, compared to 93% and 37% in NP samples, respectively. In some samples S. pneumoniae may also be present at a low abundance making isolation on gHBA plates difficult. When the OP samples were cultured on selective agar plates, the resultant growth (where present) was a complex mix of different bacterial species. Trzcinski et al. noted similar growth for trans-oral samples and found that the total bacterial load growing on gentamicin plates was significantly higher for trans-oral than trans-nasal plates (41). We did not measure the total bacterial load on gentamicin plates, but we did find that more OP samples had bacterial growth on all three gHBA dilution plates compared with NP samples (98% vs. 59%) consistent with this finding. In contrast, the total pneumococcal carriage density (from swabs) was significantly higher in NP samples compared with OP samples (median density 2.9 x 104 vs. 5.3 x 103 genome equivalents/ml, p=0.008). The contrast between pneumococcal load in swabs and the growth of bacteria on gHBA plates 226 suggests that a smaller proportion of the growth on the gHBA plates for OP samples is pneumococci compared with NP samples. We next explored the contribution of non-pneumococcal species to identification and serotyping results in 11 OP samples by isolating colonies from a small subset of OP samples. When we isolated 91 colonies from 11 OP samples on gHBA plates, we found that the non-pneumococcal growth was mainly closely related streptococcal species. This finding was consistent with findings from the StrepID component of the microarray. Previous studies have noted that the oropharynx of both children and adults is dominated by the genus Streptococcus (368), and growth of non-pneumococcal streptococci on gentamicin plates has been reported for throat swabs (369). When the OP samples were serotyped by microarray we identified a number of partial or divergent serotypes in OP samples that appeared to come from these non-pneumococcal species. We next examined whether these closely related streptococcal species in OP samples may lead to spurious pneumococcal identification and serotyping results. Firstly, we tested whether these non-pneumococcal isolates contained lytA homologs. S. pneumoniae has distinct lytA alleles compared with mitis group streptococci (370), a finding which formed the basis for the lytA qPCR assay used in our study (237). Even from NP and clinical samples there is evidence that lytA qPCR can both misidentify commensal streptococci as pneumococci, and fail to identify pneumococci (371). As such, it is possible that the lytA qPCR assay may give false positives due to lytA homologs in these other streptococcal species present in the OP samples. However, when we tested the 88 non- pneumococcal isolates obtained from the OP samples none were positive for lytA, although 10 isolates gave ‘equivocal’ results. When we examined one of the isolates that was ‘equivocal’ for lytA by whole genome sequencing, we found no clear evidence to suggest that a lytA homolog was causing the lytA signal we observed. However, predicting PCR amplification based on sequence alignments can be difficult. While primer and probe mismatches with template DNA can reduce amplification efficiency or prevent amplification altogether, the effects of mismatches depend on a number of factors including the relative position of the mismatch. Mismatches which are located internally are less deleterious than mismatches at the terminal ends of primer sequences (372, 373). It is possible that there is enough sequence agreement with the lytA homologs in 227 commensal streptococci to result in some amplification albeit with poor efficiency, therefore giving in these ‘equivocal’ results. By isolating and retesting 91 individual colonies from a subset of 11 OP samples, we confirmed that the partial and divergent capsule genes identified in microarray analysis were from non-pneumococcal species, a finding that was consistent with microarray results for the OP samples. Several other studies have also detected pneumococcal capsular biosynthesis (cps) loci and pneumococcal capsule genes in commensal streptococci (359, 374, 375) and in some strains, genes for multiple serotypes have been detected (375). These findings raise the question whether the presence of pneumococcal capsule genes may lead to spurious results using other molecular and culture-based serotyping methods. Therefore, we explored how another molecular serotyping method, multiplex PCR (mPCR), would perform on the subset of 11 OP samples. We used the mPCR method recommended by the CDC which is commonly used for carriage samples. The CDC has expressed some caution using this method for carriage specimens as they have found amplification of pneumococcal-specific serotyping targets in samples that were lytA negative (358). Consistent with their findings in combined NP/OP samples (357), we found that mPCR identified serotypes that, based on the isolates results, came from non-pneumococcal species. mPCR uses the gene cpsA as a control to detect the presence of pneumococci in the sample, however in samples there is no way to distinguish whether the individual serotypes detected by mPCR correspond to pneumococci or non- pneumococci. This is particularly problematic in OP samples where pneumococci are present, as it could be assumed that all ‘serotypes’ detected in a cpsA positive sample would be true pneumococcal serotypes. In sample FVEP-002-460 (which was positive for cpsA) we detected 13 serotypes by mPCR, from our isolate results we found that seven of these ‘serotypes’ were from non-pneumococcal species and one of these serotypes was from S. pneumoniae. We also noted that for six out of the 11 samples tested by mPCR, cpsA was amplified but no S. pneumoniae could be detected (by culture or microarray). Our findings are consistent with findings by Carvalho et al. who reported that 26 of 28 lytA negative combined NP/OP samples from adults were positive for mPCR serotypes or serogroups (357). They found an average of 6.5 serotypes detected by mPCR per combined NP/OP sample (in HIV negative adults), which was similar to the average of 6 serotypes per sample detected in our 11 OP samples by mPCR. 228 Wyllie et al., also found false positive serotyping result in oral samples not containing pneumococci when using molecular qPCR serotyping assays (376). However, they also report that some false positive serotyping results can be identified by flattened amplification curves which are uncharacteristic of amplification curves present for genuine serotypes and S. pneumoniae qPCR assays. Although not within the scope of the present work, it would be interesting to determine whether isolates such as FVEP-002- 002-03 would be distinguishable from true S. pneumoniae 33F if a qPCR serotyping assay was used. There was 97% identity for the wzy 33F gene between this isolate and S. pneumoniae, however there were three mismatches in the mPCR 33F/33A/37 forward primer and one mismatch in the associated probe sequence. It is possible that differences in amplification efficiency due to these mismatches may translate to changes in qPCR amplification curves that could be distinguishable from true pneumococcal serotypes. We also explored the performance of culture-based serotyping methods for the 11 OP samples. We applied the latex sweep agglutination method, which has been shown to be sensitive and specific with NP samples and is also able to detect multiple serotypes (154, 377). Issues were also identified using the latex sweep method on the OP samples: two of the samples contained pneumococci (detected in microarray and confirmed by isolate results) but latex agglutination failed to detect the pneumococcal serotypes; and five of the samples had a latex agglutination serotyping result but no pneumococci detected in the sample (by microarray or culture). Importantly, we demonstrated that these serotypes were not from pneumococcal species by performing latex agglutination on the OP isolates, finding 10 non-pneumococcal isolates that were serotypeable by latex agglutination. Commensal streptococci can cross-react with omniserum in latex agglutination (378), however reactions with specific serotype latex agglutination reagents are not widely reported as researchers are generally do not serotype non-pneumococcal isolates. We did attempt to confirm the latex agglutination results by quellung but no obvious positive reactions were observed. This may indicate that the capsule polysaccharide in these closely related species may be thinner than pneumococcal capsule polysaccharide or, similar to serotype 3, not covalently linked to the cell wall peptidoglycan (379), therefore making it difficult to observe the typical ‘swelling’ in a quellung reaction. Consistent with our findings, Skov Sørensen et al. demonstrated 229 capsule production in several commensal streptococcal species by double immunodiffusion using serogroup and serotype specific antisera (375). They found at least eight capsule polysaccharides from commensal species were antigenically similar or identical to pneumococcal capsule serotypes. These results highlight that spurious serotyping results are not limited to molecular based methods. We followed-up two of the non-pneumococcal isolates (derived from the 11 OP samples) that had latex agglutination serotyping results by whole genome sequencing, and compared the cps locus of these isolates with the pneumococcal cps locus of the corresponding serotype identified using latex agglutination. One of the isolates, FVEP- 002-080-02, serotyped as 19B. It is also interesting to note that 19B was a common result by latex agglutination, with 19B detected in two of the 11 samples and in seven of the 10 isolates that were serotypeable by latex agglutination. As well as the phenotypic similarity with serotype 19B, we found that FVEP-002-080-02 had a full set of 19B capsule genes, although the first six genes shared <78% identity with the 19B reference isolate. Our results suggest this isolate is S. infantis. Interestingly, Skov Sorensen et al. showed that the downstream flanking gene of cps loci in several S. infantis strains was ftsA (rather than aliA as for pneumococci) (375). Our isolate also had ftsA rather than an aliA gene downstream of the cps genes, providing further evidence that our isolates is likely to be S. infantis. We also examined the cps locus of isolate FVEP-002-002-03 which serotyped as 33F using latex agglutination. The cps region of this isolate was similar to the 33F however two copies of wzx were present (one of which was truncated). A wcrG gene was also present which is normally found in serotype 10A. Excluding the truncated wzx gene and the wcrG gene, this cps locus was highly similar to the S. tigurinus strain Az_3a cps locus (375), including the presence of the wcyO gene (which is most similar to that of serotype 21 and is not present in serotype 33F pneumococci). Although we isolated and tested 91 colonies, our exploration of the contribution of non- pneumococcal species to the complexity of OP samples was limited as we only examined a representative of each colony morphology from 11 samples. These represent only a small subsample of the total bacterial growth found on gHBA plates from the OP samples. 230 For example, it is entirely possible that one of the colonies not selected could have a lytA gene more closely resembling that found in S. pneumoniae, and would therefore help explain why some samples are positive or equivocal for lytA yet contain no pneumococci. However, one of the strengths of the study is that we selected a representative of any colony morphology on gHBA plates, rather than just those with pneumococcal-like appearance or displaying alpha-haemolysis, which allowed us to gain a greater understanding of the contributions from other bacterial species within a sample. While identifying pneumococci positive samples using culture and microarray was the best available to us, we may have falsely deemed some samples negative when they did in fact contain S. pneumoniae. While the culture step is important part of the microarray method (when microarray is used directly on samples it can result in loss of sensitivity (154)), by including the culture step we may have missed non-viable S. pneumoniae, or S. pneumoniae that was present in relatively low abundance on the gHBA plates. It is also possible that there could have been competitive inhibition of S. pneumoniae growth on the gHBA plates. S. salivarius was one of the more common streptococcal species present and some strains can inhibit S. pneumoniae growth on blood agar plates (380). One of the OP samples from which isolates were selected was originally deemed negative, but a pneumococcal colony was detected when the sample was re-cultured; interestingly the colony morphology was only found on the most dilute plate (and therefore the plate with the greatest colony separation). As the density of S. pneumoniae was relatively low in many of the OP samples, the growth other streptococci could significantly impact our detection of S. pneumoniae. The wider significance of pneumococcal capsule genes and pneumococcal-like capsule in non-pneumococci is a matter for debate. It could be argued that, although these capsule genes may not be present in S. pneumoniae, they still represent the total genetic potential for capsule production in a child and may be relevant when considering vaccine impact. Research into the evolution and genetics of S. pneumoniae by Kilian et al. has suggested that the diversity of the cps locus emerged from unidirectional interspecies gene transfer from commensal streptococci to S. pneumoniae (381, 382). Therefore the genes present in non-pneumococcal streptococci may be part of generating new pneumococcal capsule 231 types in the future or future capsule switching events. While this information may have relevance, it is unlikely that it could become data that is routinely reported in vaccine impact or carriage evaluation studies. To truly capture this information some form of molecular serotyping would need to be done on all samples in a study and not just those that were positive for S. pneumoniae which is likely to be an impractical and expensive undertaking for most research groups or study centres. It may be more relevant to focus on samples where non-pneumococci are producing capsule that is antigenically similar to that produced by pneumococci, as these could in theory be directly affected by vaccine introduction. Although we did not perform any biochemical structural analysis of the capsule produced by our isolates, the positive reactions for 19B and 33F by latex agglutination suggest that even if this capsule is not exactly the same as the S. pneumoniae capsule for these serotypes, it is at least antigenically similar. Whether these similarities are relevant to vaccine introduction and the immune response is unclear. Given that the lytA qPCR did not appear to be specific for pneumococci in OP samples, we sought to find an identification approach for OP samples that was more specific. We initially identified seven potential targets that could be used in a high-throughput qPCR assay and did not require sequencing of PCR/qPCR products for identification. After screening these potential targets against a set of reference isolates we selected bguR to develop as a qPCR assay. We also selected the piaB qPCR assay (as it has been used for a secondary target for OP samples previously (41, 376)) and compared the performance of bguR, piaB and lytA qPCR assays for 250 OP samples. We found that no single pneumococcal target performed well for OP samples (positive predictive value ranged from 13 to 33% and specificity 30-79%), however we could get some improvement in specificity and positive predictive value by combining two (positive predictive value 42- 47% and specificity 86-90%) or three (positive predictive value 53% and specificity 92%) targets. One major limitation of this analysis was the low number of samples in which S. pneumoniae was detected. Despite testing 250 OP samples we could only confirm the presence of S. pneumoniae in 11 samples which did not allow for greater discrimination 232 between targets and target combinations, particularly in relation to false negatives. We were also reliant on confirming S. pneumoniae by either culture or microarray following culture amplification, therefore it is likely a small proportion of samples may have been non-viable S. pneumoniae and deemed negative. However, there was only one sample out of the 250 OP samples tested which was positive by all three targets and no pneumococci were detected. This is similar to the results for NP samples where one out of the 250 lytA positive samples had no growth (Table 5.1) suggesting that the number of non-viable S. pneumoniae positive samples excluded using this approach is likely to be minimal. Overall, our results suggest that for greater specificity in the detection of S. pneumoniae, an additional pneumococcal target is beneficial for OP samples. Taken together our results highlight the challenging nature of conducting a pneumococcal carriage study utilising OP samples. The presence of commensal streptococci in the oropharynx interferes with identification and serotyping of true pneumococci within a sample regardless of whether culture- or molecular-based serotyping methods are used. The oropharynx of both children and adults is dominated more frequently by other streptococci than the nasopharynx is (368), providing greater opportunity for genetic exchange between streptococci. Donati et al. showed that genetic exchange with streptococci played a major role in evolution of S. pneumoniae and that many S. pneumoniae genes encoding pathogenicity were also present in S. mitis, S. oralis and S. infantis (383). Jensen et al. found that this genetic exchange, particularly within members of the mitis group streptococci, meant the reliance on one single loci may result in erroneous species identification (360). As such, consistent with our results, approaches which utilise multiple targets or loci, such as microarray, or involve the isolation of pneumococci (which may be impractical for a large study) are necessary if OP samples are to be included in an adult carriage study. Given the current WHO recommendations for the use of OP samples in adult carriage studies, we considered whether the inclusion of OP samples could provide added benefit compared with NP sampling alone if molecular methods were used (irrespective of the complexity of OP samples). We found that, although OP sampling did identify additional adult pneumococcal carriers, this did not significantly improve detection of pneumococci compared with NP sampling alone. However, given there was a 30% increase in 233 pneumococcal detection with the inclusion of OP samples, the lack of significance may be due to the low carriage rate in this population. This issue could be addressed with the addition of data from later collection years in our study or could be explored in settings with higher carriage rates in adults. Our findings, that NP sampling (compared with combined NP and OP sampling) had a sensitivity of 79%, were consistent with findings by Watt et al. (352) and Lieberman et al. (353) where the sensitivity of NP sampling was 73% and 81%, respectively. This contrasts with studies by Boersma et al. (384), Greenberg et al. (72) and Levine et al. (354) where NP sampling sensitivity was 24%, 58% and 63%, respectively. However, the findings by Levine et al. were part of a larger longitudinal study (385) where the overall NP sensitivity for all samples collected (rather than the first timepoint only) was 76%, and was therefore more similar to our findings. We found an OP sampling sensitivity of 33%, again consistent with Watt et al. and Lieberman et al. that reported 38% and 36%, respectively. The study by Hendley et al. (339), found sampling of the oropharynx superior to the nasopharynx although this study consisted of only 24 adults and the use of wooden swabs may have affected isolation of pneumococci. Between all these studies there were differences in swabbing method, swab type, swab storage media, criteria used for S. pneumoniae identification and whether multiple pneumococcal morphologies were considered. Clearly study methodology plays a role in the lack of agreement between studies, however different population characteristics (such as age group, country, ethnicity, etc.) are also likely to be contributing factors and the question of whether OP samples should be utilised for adult carriage studies may be dependent on the population in question. Our findings are also in contrast to the two studies where molecular methods were used to compare sampling of the two sites for the detection of pneumococci in adults. Principi et al., found the sensitivity of NP sampling to be just 9% compared with OP sampling which had 100% sensitivity (355). We believe the large discrepancy between this study and our study may be due to the methodology used for identification and serotyping. Principi et al. used assays for real-time lytA and wzg (cpsA) genes to detect pneumococci and used qPCR assays for serotyping PCV13 vaccine types (386). Our mPCR (and whole genome sequencing) results would suggest cpsA is not an ideal target for identification of pneumococci, as several of our lytA positive and equivocal OP samples were positive for cpsA when no pneumococci could be detected. cpsA is also associated with false-negative 234 pneumococcal identification results, particularly in pneumococci with unencapsulated lineages (62, 387). We also note that the real-time PCR serotyping assay for 19F/B, would likely have been positive for our 19B isolate based on the primer and probe sequences in this assay. Our results were also in contrast to those of Trzcinski et al. who found greater pneumococcal detection in trans-oral samples from adults (41). Their method used a combined lytA and piaB approach for identification of S. pneumoniae, and included a culture enrichment step, whereby the initial bacterial growth on gentamicin blood agar plates was harvested and later cultured again on fresh gentamicin plates. This approach identified an additional 62 adult carriers to the initial 10 identified by a single culture step in a study of 268 participants. We used a culture amplification step prior to microarray for our OP samples, which performed a similar function. We did attempt to identify pneumococci at the culture step for OP samples and found just over half of the OP samples that were positive for pneumococci were able to be detected at the culture step. This suggests that pneumococci may be hard to detect in the initial culture of OP samples and culture enrichment may improve pneumococcal detection. The utility of OP samples may be dependent on the population being evaluated. As such future carriage studies in adults may benefit from conducting a pilot study to assess the benefits of conducting both NP and OP sampling, however in populations with low carriage rates this may be impractical. In considering potential use of OP samples the unreliability of many molecular methods should be an important consideration, with serotyping and identification of OP samples requiring either culture isolation of pneumococci or a multi-locus molecular approach. The improved detection of pneumococci in the study by Trzcinski et al. and in our study following some form of culture enrichment or culture amplification step also suggests that detection of pneumococci direct from culture for OP samples may be insufficient to capture true pneumococcal carriage in adults, particularly in the presence of other bacterial growth on culture plates. Given our finding that the use of molecular methods did not significantly improve detection of pneumococci over using NP samples, we did not include OP samples in our analysis examining the indirect effect of PCV10 on carriage in adult caregivers in Fiji. 235 We next examined whether the introduction of PCV10 into the infant immunisation schedule in Fiji had an indirect effect on pneumococcal carriage in adults. We examined carriage in adult caregivers before (2012) and after (2013, 2014 and 2015) the introduction of PCV10. We found that the overall carriage rate prior to vaccine introduction was low at 10.4 %. Carriage rates in adults can vary widely from 2% in adults over 60 years of age in Portugal to nearly 60% in adults 15-39 years of age in The Gambia, and are dependent on a number of factors including vaccine coverage, relative carriage rates in children, age and the methods used for the carriage study (87, 155, 167, 351, 388- 390). Three years post-PCV10 introduction (in 2015) pneumococcal carriage was 25% lower than 2012 (7.8%), however this did not represent a significant difference (p=0.159), suggesting that PCV10 had no effect on overall carriage rates. Although vaccine type carriage was also low in adults prior to vaccine introduction at just 2.4%, three years post-vaccine introduction there was a 67% reduction in vaccine type carriage to 0.8% (p=0.047). This finding is consistent with findings in Kenya with PCV10 where vaccine type carriage declined from 8% to 4% in unvaccinated individuals (aged five years or older). The reductions in vaccine type carriage in unvaccinated adults following PCV introduction are also consistent with studies in The Gambia (155) and the US (165), as well as in England following both PCV7 and subsequent PCV13 introduction (391). Our data are also consistent with preliminary data from other age groups in the NVEP study (Dunne and Satzke et al., unpublished data), which shows a reduction in vaccine type carriage across all age groups. As expected this was particularly true in the 12-23 months of age (the targeted age group) where vaccine type carriage decreased from 21.9% in 2012 to 6.9% in 2015. Increases in non-vaccine type carriage were observed in the 12- 23 months of age group and the 5-8 weeks of age group, with the greatest increase observed in 5-8 weeks of age where non-vaccine type carriage increased from 21.0% in 2012 to 30.3% in 2015. Although we found no changes in non-vaccine type carriage in adults, non-vaccine type carriage may increase in adults in later years. 236 For adults, there were several participant characteristics that differed from pre-vaccine introduction to post-vaccine introduction, such as age, sex, household income, antimicrobial use in the prior two weeks, presence of cough and presence of allergic rhinitis which could potential confound effects observed due to vaccination. Given the increasing household income over the study period could be attributed to inflation, and given that we were missing income data from some participants (which may have a bias in why it was not collected), we also considered a model that excluded household income. When we adjusted for these potential confounders, there was some evidence of reduced vaccine-type carriage in the model excluding income (p=0.0535). Overall, this suggests that there is some evidence that PCV10 had an indirect effect on vaccine type carriage. Non-vaccine type carriage was unchanged over this period which suggests that serotype replacement is not a factor in adults in this population at this time. We also examined whether the introduction of PCV10 had an indirect effect on pneumococcal carriage density in adults. While much of the research into pneumococcal carriage focuses on reduction of carriage rates following vaccine introduction, changes in density may also be important. Higher pneumococcal carriage densities have been associated with disease states such as otitis media (392), pneumonia (393) and invasive pneumococcal disease (394, 395). High density carriage is also associated with disease severity (396, 397) and can also facilitate the transmission of pneumococci (73). We found that although there were only minor changes in overall pneumococcal carriage rates post-vaccine introduction, there was a slight but significant increase in pneumococcal density during this period (2012 median of 2.94 x 104 genome equivalents/ml vs. 2015 median of 8.53 x 104 genome equivalents/ml, p=0.003). Although this may initially be concerning, the increase was driven by a reduction in low density carriage between 2012 and 2015 with just 5% of pneumococcal carriers with a carriage density below 104 genome equivalents/ml in 2015, compared with 32% in 2012 (p=0.001). Over the same period, the percentage of pneumococcal carriers with high density carriage (greater than 106 genome equivalents/ml) increased from 8% in 2012 to 15% in 2015, which did not represent a significant increase (p=0.318). Although the thresholds for high density carriage associated with disease vary between studies, given that the proportion of pneumococcal carriers with high density carriage did not significantly increase, our 237 findings suggest that the increase in pneumococcal density overall is unlikely to represent an increased risk of pneumococcal disease. Finding an indirect effect of pneumococcal vaccination on adult carriage density was an interesting finding and with few other studies reporting carriage density in adults following infant vaccine introduction, we have little data for comparison. Roca et al, examined semiquantitative NP density in unvaccinated individuals aged >5 years as part of a PCV7 vaccine trial in the Gambia and found overall density decreased in the vaccinated villages (398). However, as the density measure was semiquantitative (based on categories 1-4 reflecting growth on gentamicin blood agar plates) rather than cfu/ml (or the molecular equivalent), it is difficult to draw comparisons between data. As the relative difference in median carriage density between 2015 and 2012 is small, it is possible this finding may not reflect a true indirect effect of pneumococcal vaccination but rather secular trends or other differences in participant characteristics from year to year. It is also possible that technical aspects of the study, such as potential improvements in swabbing techniques over time, may have contributed to a perceived higher density. As such, we hope future studies (utilising quantitative methods for determining density) further will shed further light on these findings. Although there is also limited data in children, pneumococcal vaccination has been associated with changes in density more directly. O’Brien et al. and Roca et al. both found that vaccine type carriage density in children vaccinated with PCV7 was lower than controls (398, 399). In contrast, using PCV7 vaccinated children as controls, Dagan et al. found that there was no difference in the carriage density of the six additional serotypes included in PCV13 between PCV7 and PCV13 vaccinated children (400). All three of those studies used a semiquantitative method for determining density which can only resolve differences in density over a very limited range compared with methods such as the lytA qPCR method used in our study or viable counts. Future studies utilising these methods are needed to support these findings. However, in support of vaccination lowering vaccine type density, reduced pneumococcal density has been shown in mice following vaccination which was associated with a Th17 immune response (401), and an association between density and Th17-mediated protection has been demonstrated in humans (402). 238 While a direct effect of pneumococcal vaccination on pneumococcal density may be explained by the immune response following vaccination, an indirect effect on density in unvaccinated individuals is more likely to be driven by specific serotypes. Olwagen et al. found that serotypes 6A/B, 19B/F and 23F were the highest density colonisers in a cohort of PCV naïve children (aged <17 months) (403). As these are primarily vaccine types, this suggests that overall carriage density in a population could decrease over time owing to removal of these high-density colonisers. Although this does not explain why the overall density may increase as an indirect effect of vaccination. The changing non- vaccine serotype composition may have driven the increase in overall density between 2012 and 2015. Rodrigues et al. examined the serotype-specific density of over 400 isolates from children and found specific serotypes colonise the nasopharynx at different densities (287). In particular, non-typeable pneumococci were associated with lower carriage densities in the study by Rodrigues et al. (287) and in our study non-typeable pneumococci (or pneumococci from unencapsulated lineages) were slightly more common in 2012 compared with 2015 (19.3% vs. 12.5%). Whereas serotypes 21, 35B and 23A were associated with higher carriage densities in Rodrigues et al. and in our study these serotypes were present in 2015 but not 2012. Whether similar patterns are observed in adults and whether this is sufficient to result in the changes in carriage density we observed is unclear. Unfortunately, there were not enough serotypes detected in adults over the course of the study to perform a similar analysis, however this data can be combined with data from the other cohorts in NVEP to examine whether density differs by serotype and what this means in the context of vaccination. Consistent with our findings in children in the previous two chapters, we found significant differences in carriage rates of pneumococci between iTaukei and Indo-Fijian adults. Across each of the study years iTaukei adults had higher carriage rates of pneumococci compared with Indo-Fijian adults. There were no obvious risk factors or participant characteristics that contributed to these differences in carriage rates, as any differences between ethnic groups were not consistent over the entire study period. While more iTaukei participants had rhinorrhoea in 2012 and 2014, there were no significant differences in 2013 and 2015. In 2012 and 2013, Indo-Fijian participants had higher household income than iTaukei participants, however in 2014 iTaukei participants had 239 higher household income than Indo-Fijian participants, with no difference in household income in 2015. Stratifying by ethnicity strengthened the evidence for an indirect effect of PCV10, particularly for vaccine type carriage, with significantly lower vaccine type carriage in 2015 compared with 2012 in iTaukei adults (0.7% vs. 3.9%, p=0.012). This was true even after adjusting for risk factors (with or without the inclusion of household income in the model). In contrast, prior to vaccine introduction Indo-Fijian adults had no vaccine type carriage, therefore there was little scope for vaccine-driven effects. However, with only 2.6% of Indo-Fijian adults carrying pneumococci pre-vaccine introduction, the sample size of approximately 200 caregivers in each study year, may have been insufficient to detect vaccine effects. The introduction of pneumococcal vaccine in Fiji has therefore reduced the disparity in vaccine-type carriage between the iTaukei and Indo-Fijian populations. The role of pneumococcal vaccination in reducing racial and ethnic disparities for IPD and pneumonia has been highlighted in other studies (404, 405), usually in settings where this disparity is driven by differences in socio-economic status. There were a few limitations to our study, as previously mentioned the overall carriage rate prior to vaccine introduction was low and even with a 25% reduction in pneumococcal carriage this was not significant. Similarly, given the very low rate of vaccine-type carriage, it is unsurprising that the evidence for a reduction following vaccine introduction was not stronger. A combination of the low carriage rate and the relatively small proportion of carriage that was vaccine type (23%) in 2012 may also explain why we did not observe any significant change in overall carriage or an increase in non-vaccine type carriage despite reductions in vaccine type carriage. However, with serotype replacement, PCVs generally do not have an overall effect on carriage rates and therefore this finding was consistent with other studies (391, 406, 407). As this study was an observational study, many factors may have affected the results and some observations may be due to secular trends. For instance, we lacked data on year-to- year variations in carriage rates pre-vaccine introduction. Interestingly, in 2014 (two years post-PCV10 introduction) overall pneumococcal carriage was significantly lower than in 2012 (pre-PCV10 introduction). Given there was also a slight increase in carriage 240 rate from pre-vaccine introduction to one year post-vaccine introduction, this likely represents natural variation. As such, the test for trend is possibly more reflective of changes in carriage rates due to vaccine introduction rather than measuring the absolute difference between 2012 and 2015, which suggested declining carriage rates following vaccine introduction (p=0.013). We also lacked data on whether the swabs collected from adult caregivers were from the same household. Data on households and relationships were only collected in 2015, and this data suggested that households were mainly comprised of participants of different ages. This suggests that, although we lacked the data to include clustering effects in our models, the effects would likely be minimal. Unlike some other settings, PCV10 was introduced into Fiji without a catch-up campaign. As such, the proportion of children vaccinated with PCV10 in the 12-23 month age group in 2013 (one year post-PCV10 introduction) was 2.2%, compared with 94.6% in 2014 and 99.9% in 2015. This means that indirect effects from PCV10 introduction may not have reached stability and an additional year of data would have been beneficial. Despite this, we have been able to demonstrate a reduction in vaccine type carriage, and an additional year of data would only be likely to increase the magnitude of the effect (341). Some of the strengths of this study are the use of some of the best of the currently available technology to assess carriage. This allowed us to examine additional aspects of pneumococcal carriage such as multiple serotype carriage and the presence of commensal streptococci and their pneumococcal-like serotypes. This also would allow for analysing changes in serotype-specific density, however due to the low numbers and the lack of serotype consistency over the study years, we could not perform this analysis on this dataset. Use of this technology also allowed us to effectively analyse data from a greater number of samples. The relatively large sample sizes in this study meant that even with a low vaccine type carriage rate, we were able to detect changes post-vaccine introduction. We also examined the density of pneumococcal carriage which few studies have examined in the context of vaccine introduction. This data will also be able to serve as a reference point for any future carriage surveys in Fijian adults. 241 Interestingly, we also identified some NP samples that were more characteristic of the OP samples, where there was non-pneumococcal growth on gHBA plates and the presence of non-pneumococcal capsule genes when analysed by microarray. Although NP samples such as this may be uncommon, they have been identified in other studies (377, 408). The proportion of NP samples analysed by microarray that had these OP-like characteristics, seemed to increase post-vaccine introduction. This is consistent with two studies in children where shifts in NP composition to include more Streptococcus spp. were shown to occur in vaccinated children. Valente et al. noted a higher prevalence of Streptococcus spp. in the nasopharynx of PCV13 vaccinated children compared with unvaccinated children (409). Bisbroek et al. found the bacterial composition of the nasopharynx shifted to include more OP flora, including an increase in other Streptococcus spp., following PCV7 vaccination (211). If the increase in NP samples with OP-like characteristics is vaccine driven, this highlights the importance of using technology like microarray for examining carriage post-vaccine introduction. We are curious about whether there is some link between changes in NP composition and increases in pneumococcal carriage density. There is some evidence of a negative interaction between commensal α-haemolytic streptococci and S. pneumoniae, where it has been suggested that α-haemolytic streptococci may have a protective effect against S. pneumoniae (410-412). Some potential contributors to the inhibition of S. pneumoniae by α-haemolytic streptococci include the production of hydrogen peroxide and production of bacteriocin-like inhibitory substances by commensal streptococci (413, 414). Therefore, it is possible that the vaccine serotypes are more competitive with commensal streptococci and the reduction in circulating vaccine types allows increases in commensal streptococci in the nasopharynx. This raises several possibilities such as whether the increases in non-pneumococcal streptococci are inhibiting or masking colonisation of S. pneumoniae at lower densities; or whether the increased presence of commensal streptococci could be contributing to the lytA signal, as indicated in some OP samples, and therefore causing increases in the apparent density of S. pneumoniae in these NP samples. With greater use of technology such as microarray for carriage studies which can detect the presence of non-pneumococcal species, future studies are likely to shed more light on this area. 242 5.4 Conclusions Overall, we found that the introduction of PCV10 into the infant immunisation schedule in Fiji significantly reduced the carriage of vaccine type pneumococci in adult caregivers three years post-vaccine introduction. We did not note any subsequent increase in non- vaccine type carriage over the study period but did observe an increase in overall carriage density. When we examined the use of OP samples for adult carriage studies, we found that both conventional culture-based and molecular methods have the potential to misidentify and/or incorrectly serotype closely related streptococcal species. This is due to some streptococcal species having the potential to produce capsule that is phenotypically and/or genotypically similar to pneumococci. We found that when screening complex samples such as OP samples for pneumococci using qPCR assays, an additional pneumococcal target should be used to reduce the number of false positive results. In general, OP samples were problematic and did not significantly improve detection of pneumococci in Fijian adults over the use of NP samples alone. 243 Chapter 6 Overview 244 6 Overview The overall objective of this research was to evaluate the effect of pneumococcal vaccination on NP microbiology, ideally to provide additional evidence in support of the pneumococcal vaccine program in Fiji and other pacific countries. This thesis examined the use of three different vaccines in Fiji and the effects on NP carriage of pneumococci in different age groups. We examined three issues surrounding pneumococcal vaccine use that addressed current knowledge gaps in the field: hyporesponsiveness following the use of 23vPPV (chapter 3); species replacement and the impact of PCV7 on the NP microbiome (chapter 4); and herd protection following vaccine introduction in a middle- income setting (chapter 5). In all three chapters we examined the effects of pneumococcal vaccination on NP carriage of S. pneumoniae, mediated by direct or indirect effects. In chapter 3 we examined the direct, long-term effect of pneumococcal vaccination in a follow-up study to an earlier vaccine trial in Fiji. We found that 23vPPV given at 12 months of age did not have a long- term effect on pneumococcal carriage in children 5-7 years of age overall. As our primary concern was that the immune hyporesponsiveness observed at 18 months of age may have had a long-term effect in these children leading to increased carriage of serotypes in the 23vPPV vaccine, this was a positive finding. However, when we examined the results after stratifying by ethnicity we found that while 23vPPV may have reduced pneumococcal carriage in Indo-Fijian children, there was evidence that 23vPPV was associated with higher carriage of S. aureus in iTaukei children. Countries which included 23vPPV in their infant vaccine programs, such as in indigenous Australian populations, have now abandoned 23vPPV due to limited evidence of efficacy and concerns about hyporesponsiveness. Similarly, our findings do not support the use of 23vPPV in children. We examined the issue of species replacement in chapter 4, examining the effect of PCV7 given in early life on the NP microbiome in children at 12 months of age. Although there was little overall effect, we did find some potentially beneficial differences between vaccinated and unvaccinated individuals in each ethnic group. In particular, vaccinated iTaukei children had a lower relative abundance of Streptococcus compared with 245 unvaccinated iTaukei children, and vaccinated Indo-Fijian children had a higher relative abundance of Dolosigranulum compared with unvaccinated Indo-Fijian children. As Dolosigranulum has been associated with a healthy microbiome, this alleviates some potential safety concerns regarding species replacement and the introduction of PCV into the infant immunisation schedule in Fiji. Where we found an effect of pneumococcal vaccination in bacterial species or genera not targeted by the vaccine, we examined the role for a direct bacterial association with pneumococci. However, in our study we did not find any evidence of a direct association with pneumococci. There is a complex web of interactions that occur within the microbiome, and it is likely that indirect interactions and host immunity also play a role in determining the emergence of particular species or genera following vaccination. The importance of other microbiota was also highlighted when we revisited pneumococcal carriage sampling in adults using molecular methods, to determine whether the use of OP samples represented added benefit (chapter 5). We found that testing OP samples resulted in a minor, but non-significant, improvement in detection of pneumococci, and OP samples were more complex than NP samples due to the presence of streptococcal species in the oropharynx. These streptococcal species possessed genes similar to pneumococci which interfered with pneumococcal identification and serotyping. As such, OP samples were difficult to serotype accurately except by microarray. While streptococci with capsule genes in samples from the oropharynx have been previously reported (357), reports are not widespread. Our research was novel in that we considered the contribution of a wide range of bacterial strains cultured on blood agar plates rather than only α-haemolytic bacteria or those with a pneumococcal-like appearance. We also demonstrated the effects the presence of these bacteria have on serotyping results using multiple methods, highlighting a significant issue in the field. Some investigators have recommended that OP samples be used instead of NP samples for assessing pneumococcal carriage in adults (41, 355). Based on our findings, sampling the nasopharynx is sufficient for assessing adult carriage and we would not recommend sampling only the oropharynx for examining carriage in adults given the potential for misidentification of pneumococci and false-positive serotyping results using OP samples. We found that the proportion of adult NP samples with ‘OP-like’ characteristics increased 246 following vaccine introduction in Fiji. As such, the issues relating to identification and serotyping encountered with OP samples may have the potential to become issues with NP samples in vaccinated settings in the future. It is possible that the complexity we observed in OP samples from Fijian adults may not occur in OP samples from all populations. However, while the microbiota in the nasopharynx may have been significantly different between ethnic groups, the complexity of the OP samples did not differ between the ethnic groups. Therefore, it is plausible that our findings are applicable to other populations, and that the lack of recognition of the problems with OP samples is due to differences in study methodology. We examined the indirect effect of PCV10 introduction on adult carriage of S. pneumoniae in Fiji (chapter 5) and found some evidence that pneumococcal vaccination in infants reduced vaccine type carriage in adults. We did not observe an overall reduction in carriage and nor did we find an increase in non-vaccine serotypes. However, our power to detect a reduction in pneumococcal carriage was limited by the low carriage rate in adults, with only around 10% of adults sampled carrying pneumococci pre-vaccine introduction, and only 2.4% carrying vaccine serotypes. Additionally, in Fiji there was no catch-up campaign following vaccine introduction and the proportion of 12-23 month old children that were vaccinated one year post-vaccine introduction was low. As indirect effects (including serotype replacement) take approximately two to three years to become evident (341), continued surveillance will be needed to identify serotype replacement. As such, an additional year or years of carriage data would be invaluable in determining the changing carriage and transmission dynamics following PCV10 introduction, and demonstrating the added benefit of infant vaccination for adults. When we stratified by ethnicity, we found less than 3% of Indo-Fijian adults carried pneumococci prior to vaccine introduction, of which all were non-vaccine type pneumococci. Interestingly, in the subset of children that were part of the microbiome study, we also found comparatively low rates of vaccine type carriage in Indo-Fijian children, with only 14% (one out of seven) of pneumococci present in the unvaccinated 247 Indo-Fijian children being vaccine type compared with 45% (10 out of 22) in unvaccinated iTaukei children. Results were the same when PCV13 serotypes were considered in the adult study, suggesting that current pneumococcal vaccine compositions are unlikely to have much impact in the Indo-Fijian population due to low carriage of vaccine types, although fortunately the overall carriage rates in Indo-Fijians are relatively low. In iTaukei adults we found a reduction in vaccine type carriage following PCV10 introduction (from 3.9% to 0.7%). We did not see any reduction in overall pneumococcal carriage nor an increase in non-vaccine type carriage, which was similar to the unstratified results. With the slightly higher baseline carriage rate of 16% overall carriage and 4% vaccine type carriage in iTaukei adults compared to adults overall, we were able to clearly detect a decrease in vaccine type carriage following vaccine introduction. Across this research it was evident that the two main ethnic groups in Fiji had significantly different microbiota, particularly in the carriage of S. pneumoniae. A higher proportion of the iTaukei population had carried pneumococci compared to the Indo-Fijian population, a difference that extended from early life into adulthood and could not be explained by any of the demographic or risk factor data we collected. In addition to the differences in carriage rates of S. pneumoniae, we found that the overall microbial composition of the nasopharynx was significantly different between ethnic groups. When symptoms of an URTI were present, the ‘healthy’ and ‘sick’ microbiota were also different between ethnic groups. In other populations, differences in disease incidences between ethnic groups within a population or country are often attributable to differences related to socioeconomic status (404). We do not believe this is a major underlying factor in Fiji, although in the FFS we did find that iTaukei children had lower weekly household income than the Indo-Fijian children. However, data from the original FiPP study also found differences in carriage rates where the household incomes were not different. Furthermore, the higher carriage of S. pneumoniae in the iTaukei population persisted in adults in each year of the NVEP despite the Indo-Fijian adults sampled having lower household income than the iTaukei adults in 2014. 248 There is conflicting evidence about which factors play the greatest role in shaping our microbiota (including S. pneumoniae), and whether behavioural and cultural factors like diet are more important than host genetics and immunity (415). While initial development of microbial communities in infants is influenced by a number of factors including prenatal exposures (e.g. antibiotics, maternal nutrition), mode of delivery (and exchange of microbiota from mothers) and breastfeeding (280, 416-418), many factors can shape the microbiome throughout life. In some studies, host genetics have been found to be important in shaping microbiota (419-423). However, other studies have shown that mode of subsistence (e.g. farming or fishing), diet and lifestyle factors to be more important determinants of microbial communities than genetic ancestry (424, 425). Host genetics can also influence the immune response, while at the same time the immune response can also be influenced by gut microbiota and diet (426, 427). Environmental exposures (such as house dust) can also mediate microbial composition and immune response (428). The presence of specific pathogens or parasites also shape microbiota (424). Given the complexity and interconnected nature of factors that can determine microbiota, it is likely multiple factors are important in the underlying differences in microbial composition between ethnic groups in Fiji. The 2004 Fiji National Nutrition survey revealed several dietary and lifestyle factors that differed between the two ethnic groups including differences in consumption of white meat (e.g. chicken) and preserved foods, as well as levels of light physical activity and heavy cigarette use (>10 cigarettes day) (429). There is some evidence that diet and composition of intestinal microbiota may be determinants of respiratory microbiota (281, 430), and therefore future studies exploring respiratory microbiota could include collection of additional dietary and lifestyle factors and possibly include study of microbial communities in other niches such as the gastrointestinal tract. Other environmental factors like social contact may also be important in Fiji. Preliminary research has suggested, on average, there were a significantly greater number of social contacts in iTaukei individuals compared with Indo-Fijians (431). There is some evidence that social contact can influence microbiota (432-434) and transmission but with limited data in this area, the relative importance of social contact compared with other factors as 249 a determinant of microbiota is unknown. The role social contact plays in S. pneumoniae carriage is an area of ongoing research in our group. While further studies will be needed to elucidate the underlying differences between the two ethnic groups, from our research it is clear that ethnic-specific variations affect microbiome composition in Fiji. As noted in a study of life expectancy and adult mortality in Fiji, considering the population of Fiji as a whole can mask divergent trends in the ethnic groups (435). As such, we believe that the effects of pneumococcal vaccination should also be considered separately for each ethnic group. Pneumococcal vaccination reduces pneumococcal disease caused by vaccine types worldwide. Pneumococcal vaccines also impact carriage, which in turn leads to indirect effects in populations where vaccines are widely used in paediatric age groups. In Fiji, we have shown that the introduction of PCV10 into the infant immunisation schedule had an indirect effect at reducing carriage of vaccine type pneumococci in adults. We have shown that it may be inadvisable to use 23vPPV as a booster to improve vaccine coverage owing to 23vPPV having no long-term impact on pneumococcal carriage and was previously shown to result in hyporesponsiveness. Species replacement does not appear to be a significant problem in PCV7 vaccinated children. We also note that the microbiota of the two main ethnic groups in Fiji are distinct and that these differences continue from early life to adulthood. 250 Chapter 7 Appendices 251 7 Appendices 7.1 Appendix A. 16S sequence processing and analysis scripts 7.1.1 Sequence processing script using MOTHUR mothur > set.dir(input=“working directory”) mothur > make.contigs(file=stability.files) mothur > summary.seqs(fasta=stability.trim.contigs.fasta) mother > screen.seqs(fasta=stability.trim.contigs.fasta, group=stability.contigs.groups, maxambig=0, minlength=225, maxlength=275, maxhomop=8) mothur > summary.seqs(fasta=stability.trim.contigs.good.fasta) mothur > unique.seqs(fasta=stability.trim.contigs.good.fasta) mothur > count.seqs(name=stability.trim.contigs.good.names, group=stability.contigs.good.groups) mothur > summary.seqs(count=stability.trim.contigs.good.count_table) mothur > pcr.seqs(fasta=silva.nr_v119.align, oligos=V4_oligos.oligos) mothur > summary.seqs(fasta=silva.nr_v119.pcr.align) mothur > pcr.seqs(fasta=silva.nr_v119.align, start=13862, end=23444, keepdots=F) mothur > summary.seqs(fasta=silva.nr_v119.pcr.align) 252 mothur > align.seqs(fasta=stability.trim.contigs.good.unique.fasta, reference=silva.nr_v119.pcr.align) mothur > summary.seqs(fasta=stability.trim.contigs.good.unique.align, count=stability.trim.contigs.good.count_table) mothur > screen.seqs(fasta=stability.trim.contigs.good.unique.align, count=stability.trim.contigs.good.count_table, summary=stability.trim.contigs.good.unique.summary, start=8, end=9582, maxlength=255, minlength=250) mothur > summary.seqs(fasta=current, count=current) mothur > filter.seqs(fasta=stability.trim.contigs.good.unique.good.align, vertical=T, trump=.) mothur > unique.seqs(fasta=stability.trim.contigs.good.unique.good.filter.fasta, count=stability.trim.contigs.good.good.count_table) mothur > pre.cluster(fasta=stability.trim.contigs.good.unique.good.filter.unique.fasta, count=stability.trim.contigs.good.unique.good.filter.count_table, diffs=3) mothur > split.abund(fasta=stability.trim.contigs.good.unique.good.filter.unique.precluster. fasta, count=stability.trim.contigs.good.unique.good.filter.unique.precluster.count_tabl e, cutoff=2, accnos=true) mothur > chimera.uchime(fasta=stability.trim.contigs.good.unique.good.filter.unique.precl 253 uster.abund.fasta, count=stability.trim.contigs.good.unique.good.filter.unique.precluster.abund.cou nt_table, dereplicate=t) mothur > remove.seqs(fasta=stability.trim.contigs.good.unique.good.filter.unique.preclust er.abund.fasta, accnos=stability.trim.contigs.good.unique.good.filter.unique.precluster.abund.uc hime.accnos) mothur > summary.seqs(fasta=current, count=current) mothur > classify.seqs(fasta=stability.trim.contigs.good.unique.good.filter.unique.preclust er.abund.pick.fasta, count=stability.trim.contigs.good.unique.good.filter.unique.precluster.abund.uchi me.pick.count_table, reference=silva.nr_v119.ng.fasta, taxonomy=silva.nr_v119.tax, cutoff=80) mothur > remove.lineage(fasta=stability.trim.contigs.good.unique.good.filter.unique.preclu ster.abund.pick.fasta, count=stability.trim.contigs.good.unique.good.filter.unique.precluster.abund.uchi me.pick.count_table, taxonomy=stability.trim.contigs.good.unique.good.filter.unique.precluster.abund. pick.nr_v119.wang.taxonomy, taxon=Chloroplast-Mitochondria-unknown- Archaea-Eukaryota) mothur > summary.seqs(fasta=current, count=current) mothur > cluster.split(fasta=stability.trim.contigs.good.unique.good.filter.unique.precluster .abund.pick.pick.fasta, 254 count=stability.trim.contigs.good.unique.good.filter.unique.precluster.abund.uchi me.pick.pick.count_table, taxonomy=stability.trim.contigs.good.unique.good.filter.unique.precluster.abund. pick.nr_v119.wang.pick.taxonomy, splitmethod=classify, taxlevel=4, cutoff=0.15) mothur > make.shared(list=stability.trim.contigs.good.unique.good.filter.unique.precluster. abund.pick.pick.an.unique_list.list, count=stability.trim.contigs.good.unique.good.filter.unique.precluster.abund.uchi me.pick.pick.count_table, label=0.03) mothur > classify.otu(list=stability.trim.contigs.good.unique.good.filter.unique.precluster.a bund.pick.pick.an.unique_list.list, count=stability.trim.contigs.good.unique.good.filter.unique.precluster.abund.uchi me.pick.pick.count_table, taxonomy=stability.trim.contigs.good.unique.good.filter.unique.precluster.abund. pick.nr_v119.wang.pick.taxonomy, label=0.03) mothur > count.groups(shared=stability.trim.contigs.good.unique.good.filter.unique.preclu ster.abund.pick.pick.an.unique_list.shared) 7.1.2 Alpha and Beta diversity library(vegan) OTU_file <- otufile1 div<-(OTU_file) div SDI <- diversity (div) #calculate Shannon diversity index# 255 SDI simp <- diversity(div, "simpson") #calculate cimpson diversity index# write.csv(SDI,"C:/Users/laura/Google Drive/RHD/Microbiome/Analysis files/All sequences/SDI.csv") write.csv(simp,"C:/Users/laura/Google Drive/RHD/Microbiome/Analysis files/All sequences/simpson.csv") library(vegan) library(vegan3d) library(RColorBrewer) mapdata <- all_risk_map_data Jdistrisk<-vegdist(OTU_file, method="jaccard", binary=FALSE)#calculate the weighted distance matrix (otherwise make binary=TRUE)# #calculate the P-Values for the different groups# adonis(formula = OTU_file ~ eth + PCV7, data = mapdata, method = "jaccard", permutations = 1999) nmdsrisk <- metaMDS(Jdistrisk) #calculate nMDS# nmdsrisk stressplot(nmdsrisk) 7.1.3 Heatmap script in R library(vegan) library(gplots) library(RColorBrewer) #import data - 'for_heatmap'in my case# #make sure the row names are not considered data# hdata <- for_heatmap 256 #Decide on colour palette for heatmap# scale1 <- colorRampPalette(c("royalblue", "red"), space = "rgb")(100) heatmap(as.matrix(hdata), Rowv = NA, Colv = NA, col = scale1) #determine the maximum relative abundance for each column# maxab <- apply(hdata, 2, max) head(maxab) #Then set a threshold for which families to display i.e. 1%# n1 <- names(which(maxab < 0.01)) n2 <- names(which(maxab <1)) n3 <- names(which(maxab <30)) hdata.1 <- hdata[, -which(names(hdata) %in% n1)] hdata.2 <- hdata[, -which(names(hdata) %in% n2)] hdata.3 <- hdata[, -which(names(hdata) %in% n3)] heatmap(as.matrix(hdata.1), Rowv = NA, Colv = NA, col = scale1, margins = c(10, 2)) heatmap(as.matrix(hdata.2), Rowv = NA, Colv = NA, col = scale1, margins = c(10, 2)) heatmap(as.matrix(hdata.3), Rowv = NA, Colv = NA, col = scale1, margins = c(10, 2)) #heatmap 3 is the most appropriate #calculate the jaccard dissimilarity# data.dist <- vegdist(hdata.3, method = "jaccard") row.clus <- hclust(data.dist, "aver") #Now to make the heatmap with dendrogram and clustering data# heatmap(as.matrix(hdata.3), Rowv = as.dendrogram(row.clus), Colv = NA, col = scale1, margins = c(10, 3)) #to cluster the families together# data.dist.g <- vegdist(t(hdata.3), method = "jaccard") col.clus <- hclust(data.dist.g, "aver") 257 heatmap(as.matrix(hdata.3), Rowv = as.dendrogram(row.clus), Colv = as.dendrogram(col.clus), col = scale1, margins = c(10, 3)) #To make a column highlighting the different groups you need to make a vector and have the groups coded, you can then rename the code to the colour# ethnic <- c(2,2,2,1,1,2,1,1,2,1,2,2,2,1,1,1,1,2,2,1,1,1,2,2,1,2,2,2,2,2,2,1,1,1,1,1,2,1,1,1,2, 2,1,1,2,2,2,2,1,2,1,1,2,1,2,1,2,2,1,2,2,1,2,1,1,1,1,1,1,1,2,2,2,1,2,1,2,2,1,1,2,2,1, 1,1,2,2,2,1,2,2,1,2,1,2,1,1,1,1,1,1,2,2,1,2,1,2,2,1,2,2,2,2,1,1,2,1,1,1,2,1,2,1,1,1, 2,2,1,2,2,2,1) var1 <- ethnic var1 <- replace(var1, which(var1==1), "blue") var1 <- replace(var1, which(var1==2), "red") vaccination <- c(0,1,1,0,1,0,1,1,1,0,1,0,0,1,0,0,1,1,0,1,0,0,0,1,1,0,0,1,0,0,1,0,1,0,0,0,1,1,0,1,1, 0,1,1,0,1,1,1,0,1,1,0,1,0,1,1,0,0,1,0,1,1,0,1,0,1,0,1,0,1,0,0,1,1,0,1,1,1,1,0,1,0,0, 1,0,0,0,1,0,1,1,0,1,1,0,0,0,1,1,1,0,0,1,0,0,1,0,0,0,1,0,1,0,0,1,0,0,0,1,1,0,0,1,0,1, 1,1,1,1,0,1,0) var2 <- vaccination var2 <- replace(var2, which(var2 == 0), "green") #0d PCV7# var2 <- replace(var2, which(var2 == 1), "orange") #3d PCV7# URTI <- c(0,0,1,0,1,1,1,0,0,0,0,1,0,1,0,1,0,0,0,0,1,0,0,0,1,0,0,0,0,1,0,0,0,0,0,1,0,0,1,0,0, 0,1,0,0,0,0,0,0,0,1,1,1,0,0,0,1,0,0,1,0,0,0,1,1,1,1,1,1,0,0,0,0,0,0,0,1,0,1,1,1,0,1, 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,1,0,1,1,0,0,0,0,0,0,0,0,0,1,0,0,0,0,1,1,0,0,1,1,0,1, 1,0,0,0,0,0,0) var3 <-URTI var3 <- replace(var3, which(var3 == 0), "purple3") #no URTI# var3 <- replace(var3, which(var3 == 1), "turquoise3") #URTI# #You then need to combined the vector with the row names/sample ID so the samples will be coloured the right colour while they're clustered# cbind(row.names(hdata), var1) cbind(row.names(hdata), var2) cbind(row.names(hdata), var3) 258 var1 var2 var3 #for using heatmap.2# heatmap.2(as.matrix(hdata.3), Rowv = as.dendrogram(row.clus), Colv = as.dendrogram(col.clus), col = scale1, RowSideColors = var3, margins = c(11, 8), trace = "none", density.info = "none", xlab = "taxonomy", main = "Heatmap", lhei = c(2, 8)) #To get the heatmap without the family clustering heatmap.2(as.matrix(hdata.3), Rowv = as.dendrogram(row.clus), Colv = FALSE, col = scale1, RowSideColors = var3, margins = c(11, 8), dendrogram = "row", trace = "none", density.info = "none", xlab = "taxonomy", main = "Heatmap", lhei = c(2, 8)) heatmap.2(as.matrix(hdata.3), Rowv = as.dendrogram(row.clus), Colv = FALSE, col = scale1, margins = c(11, 8), dendrogram = "row", trace = "none", density.info = "none", xlab = "taxonomy", main = "Heatmap", lhei = c(2, 8)) legend("topright", legend = c("iTaukei", "Indo-Fijian"), fill=c("blue", "red"), border=FALSE, bty="n", y.intersp = 1.0, cex=0.8) legend("topright", legend = c("unvaccinated", "vaccinated"), fill=c("green", "orange"), border=FALSE, bty="n", y.intersp = 1.0, cex=0.8) legend("topright", legend = c("no URTI", "URTI"), fill=c("purple3", "turquoise3"), border=FALSE, bty="n", y.intersp = 1.0, cex=0.8) # Calinski clustering require(vegan) fit <- cascadeKM(scale(data.dist, center = TRUE, scale = TRUE), 1, 15, iter = 1000) plot(fit, sortg = TRUE, grpmts.plot = TRUE) calinski.best <- as.numeric(which.max(fit$results[2,])) cat("Calinski criterion optimal number of clusters:", calinski.best, "\n") # 5 clusters! plot(row.clus) 259 rect.hclust(calinski.best, border = "red") 7.1.4 Correlation network script 7.1.4.1 In MOTHUR mothur > set.dir(input=C:\Users\Laura\Google Drive\RHD\Microbiome\sequence preparation) mothur> get.oturep(list=stability.trim.contigs.good.unique.good.filter.unique.precluster.ab und.pick.pick.an.unique_list.list, fasta=stability.trim.contigs.good.unique.good.filter.unique.precluster.abund.pick. pick.fasta, count=stability.trim.contigs.good.unique.good.filter.unique.precluster.abund.uchi me.pick.pick.count_table, method=abundance) #Phylogenetic - for methods dependent on phylogenetic tree or the unifrac commands# mothur > dist.seqs(fasta=stability.trim.contigs.good.unique.good.filter.unique.precluster.a bund.pick.pick.an.unique_list.0.03.rep.fasta, output=lt) mothur > clearcut(phylip=stability.trim.contigs.good.unique.good.filter.unique.precluster.a bund.pick.pick.an.unique_list.0.03.rep.phylip.dist) mothur > sparcc(shared=new_refined_100.shared) 260 7.1.4.2 In R library(igraph) library(ggplot2) # Build the co-ocurrence networks # First make some functions makeNet <- function(X){ a<-matrix(nrow=1,ncol=4) a[1,1] <- rownames(sparcc$r)[X[1]] a[1,2] <- colnames(sparcc$r)[X[2]] a[1,3] <- sparcc$r[X[1], X[2]] a[1,4] <- sparcc$P[X[1],X[2]] return(a) } graph.transform.weights <- function (X) { require(igraph) data.tmp <- matrix(0, nrow=nrow(X), ncol=2) dimnames(data.tmp)[[2]] <- c("i", "j") data.tmp[,1] <- X[,1] data.tmp[,2] <- X[,2] data <- data.frame(data.tmp) graph.data <- graph.data.frame(data, directed=F) E(graph.data)$weight <- abs(as.numeric(X[,3])) E(graph.data)$cor <- as.numeric(X[,3]) E(graph.data)$pvalue <- as.numeric(X[,3]) summary(graph.data) cat("Average degree:",ecount(graph.data)/vcount(graph.data)*2) return(graph.data) } # Read the sparCC correlation matrix 261 sparcc.cor <- read.table(file = "new_refined_100_sparcc_correlation.txt", sep = "\t", header = T, row.names = 1) dim(sparcc.cor) # Read the sparCC pseudo p-values sparcc.pval <- read.table(file = "new_refined_100_sparcc_pvalue.txt", sep = "\t", header = T, row.names = 1) dim(sparcc.pval) # Create a new data structure that contains a list of data frames with the sparcc results sparcc <- structure(list(r = sparcc.cor, P = sparcc.pval)) # Filter the correlations magnitudes > 0.3 and with a pseudo p-value < 0.05 and will get the # coordinates in the dataframe pval <- 0.05 r <- 0.3 selrp<-which((abs(sparcc$r) > r & sparcc$P < pval) & lower.tri(sparcc$r == TRUE), arr.ind = TRUE) #Use the coordinates to build a dataframe containing the OTU pairs and the correlation coef # and the pseudo pvalues sparcc.graph.df <- t(apply(selrp,1, makeNet)) # Transform the dataframe to an igraph object sparcc.graph <- graph.transform.weights(sparcc.graph.df) # Write the network to a file write.graph(sparcc.graph, "sparcc_graph_100_new_0.3.graphml", format="graphml") 262 7.2 Appendix B. Additional data for chapter 4 7.2.1 Clusters and risk factors Table 7.1. Prevalence of clusters by risk factors. Vaccinated Ethnicity No Yes p-value iTaukei Indo-Fijian p-value Cluster (n=65) (n=67) (n=67) (n=65) 1 15 (23) 19 (28) 0.5529 9 (13) 25 (38) 0.0013 2 10 (15) 12 (18) 0.8164 11 (16) 11 (17) 1.0000 3 17 (26) 18 (27) 1.0000 24 (36) 11 (17) 0.0179 4 9 (14) 7 (10) 0.6020 10 (15) 6 (9) 0.4255 5 3 (5) 2 (3) 0.6779 1 (1) 4 (6) 0.2046 6 11 (17) 9 (13) 0.6328 12 (18) 8 (12) 0.4684 Gender Season collected Female Male p-value Wet Dry Cluster p-value (n=61) (n=71) (n=98) (n=34) 1 17 (29) 17 (24) 0.6910 22 (22) 12 (35) 0.1728 2 9 (15) 13 (18) 0.6448 15 (15) 7 (21) 0.5934 3 15 (25) 20 (28) 0.6955 28 (29) 7 (21) 0.4993 4 9 (15) 7 (10) 0.4321 15 (15) 1 (3) 0.0692 5 3 (5) 2 (3) 0.6619 4 (4) 1 (3) 1.0000 6 8 (13) 12 (17) 0.6304 14 (14) 6 (18) 0.5928 Breastfeeding Exposure to cigarette smoke No Yes No Yes Cluster p-value p-value (n=48) (n=84) (n=71) (n=61) 1 14 (29) 20 (24) 0.5384 16 (23) 18 (30) 0.4263 2 5 (10) 17 (20) 0.2241 15 (21) 7 (11) 0.1643 3 13 (27) 22 (26) 1.0000 22 (31) 13 (21) 0.0292 4 7 (15) 9 (11) 0.5830 6 (8) 10 (16) 0.1889 5 4 (8) 1 (1) 0.0584 4 (6) 1 (2) 0.3727 6 5 (10) 15 (18) 0.3178 8 (11) 12 (20) 0.2258 Antibiotic use in past 2 wks URTI symptoms No Yes No Yes Cluster p-value p-value (n=119) (n=13) (n=93) (n=39) 1 32 (27) 2 (15) 0.5133 29 (31) 5 (13) 0.0303 2 21 (18) 1 (8) 0.6943 17 (18) 5 (13) 0.6098 3 29 (24) 6 (46) 0.1054 21 (23) 14 (36) 0.1328 4 14 (12) 2 (15) 0.6584 8 (9) 8 (21) 0.0780 5 5 (4) 0 (0) 1.0000 4 (4) 1 (3) 1.0000 6 18 (15) 2 (15) 1.0000 14 (15) 6 (15) 1.0000 Values are number (percentage); Significant values are shown in bold, calculated using Fisher’s Exact test 263 7.2.1.1 nMDS plots of clusters including the third dimension nMDS plots in this thesis show only the first and second dimensions. The first two dimension are the dimension that account for the greatest variability in the data, adding the third dimension provides greater detail but is harder to visualise. Figure 7.1. Three dimensional nMDS of the dissimilarity in microbial composition by clusters. Shown are Clusters 1-6. 264 Figure 7.2. First and third dimensional nMDS of the dissimilarity in microbial composition by clusters . Shown are Clusters 1-6. 265 Figure 7.3. Second and third dimensional nMDS of the dissimilarity in microbial composition by clusters. Shown are Clusters 1-6. 266 7.2.2 Additional nMDS plots In section 4.2.9 (Association between risk factors and microbial composition) only risk factors for which the was a significant difference in microbial composition were included, the other nMDS plot are shown below. p=0.838 Figure 7.4. nMDS of the dissimilarity in microbial composition by gender. Shown are male (blue) and female (red) children. The centre points represent the mean of each group and the lines represent distance from the mean for each sample. No significant differences were observed between the two groups (p=0.838), p-value calculated using PERMANOVA. 267 p=0.106 Figure 7.5. nMDS of the dissimilarity in microbial composition by season. Shown are samples collected from children in wet (blue) and dry (red) season. The centre points represent the mean of each group and the lines represent distance from the mean for each sample. No significant differences were observed between the two groups (p=0.106), p- value calculated using PERMANOVA. 268 p=0.336 Figure 7.6. nMDS of the dissimilarity in microbial composition by swab collection year. Shown are samples collected from children in 2005 (blue), 2006 (red) and 2007 (green). The centre points represent the mean of each group and the lines represent distance from the mean for each sample. No significant differences were observed between the two groups (p=0.106), p-value calculated using PERMANOVA. 269 p=0.383 Figure 7.7. nMDS of the dissimilarity in microbial composition by breastfeeding status. Shown are samples collected from children currently being breastfed (blue) and children not being breastfed (red). The centre points represent the mean of each group and the lines represent distance from the mean for each sample. No significant differences were observed between the two groups (p=0.383), p-value calculated using PERMANOVA. 270 p=0.335 Figure 7.8. nMDS of the dissimilarity in microbial composition based on whether breastfeeding was continued past the first 6 weeks of life. Shown are samples collected from children where breastfeeding was stopped within the first 6 weeks of life (blue) and children that were breastfed for more than 6 weeks (red). The centre points represent the mean of each group and the lines represent distance from the mean for each sample. No significant differences were observed between the two groups (p=0.335), p-value calculated using PERMANOVA. 271 p=0.376 Figure 7.9. nMDS of the dissimilarity in microbial composition based on whether breastfeeding was continued past the first 6 months of life. Shown are samples collected from children where breastfeeding was stopped within the first 6 months of life (blue) and children that were breastfed for more than 6 months (red). The centre points represent the mean of each group and the lines represent distance from the mean for each sample. No significant differences were observed between the two groups (p=0.376), p-value calculated using PERMANOVA. 272 p=0.271 Figure 7.10. nMDS of the dissimilarity in microbial composition by exposure to household cigarette smoke. Shown are samples collected from children that have exposure to household cigarette smoke (blue) and children that are not exposed to household cigarette smoke (red). The centre points represent the mean of each group and the lines represent distance from the mean for each sample. No significant differences were observed between the two groups (p=0.271), p-value calculated using PERMANOVA. 273 p=0.667 Figure 7.11. nMDS of the dissimilarity in microbial composition by antibiotics use in the previous two weeks. Shown are samples collected from children that have used antibiotics in the previous two weeks (blue) and children that have not used antibiotics in the previous two weeks (red). The centre points represent the mean of each group and the lines represent distance from the mean for each sample. No significant differences were observed between the two groups (p=0.667), p-value calculated using PERMANOVA. 274 Figure 7.12. nMDS of the dissimilarity in microbial composition by ethnicity and vaccination status for a subset of 24 samples plus controls. Shown are samples collected from unvaccinated iTaukei children (green), vaccinated iTaukei children (blue), unvaccinated Indo-Fijian children (orange), vaccinated Indo-Fijian children (red) and controls (black). The centre points represent the mean of each group and the lines represent distance from the mean for each sample. 275 7.2.3 Additional graphs A B Figure 7.13. Top twenty most important OTUs in random forest models by vaccination status for iTaukei (A) and Indo-Fijian (B) children. 276 lower total abundance higher total abundance = = higher relative abundance lower relative abundance of S. pneumoniae of S. pneumoniae S. pneumoniae e c density/abundance n a d other microbiota n u b a l a t o T unvaccinated vaccinated Figure 7.14. Theoretical situation where S. pneumoniae density would not differ between vaccinated and unvaccinated children but the relative abundance of S. pneumoniae would differ by vaccination status. 277 7.3 Appendix C. Additional data for chapter 5 7.3.1 Characterisation of OP isolates 278 Table 7.2. Individual OP isolate results for identification and serotyping. Microarray lytA qPCR MALDI-TOF Latex Multiplex Sample Isolate Optochin Bile test Ct MS agglutination PCR Result serotyping StrepID PathID value FVEP- S. salivarius ssp. 01 Resistant insoluble No Ct negative 19B; 12 36/7F-like* saliva-5/5 SP-1/5 not tested 002-002 salivarius FVEP- 02 Resistant insoluble No Ct negative S. dysgalactiae negative 7F-like* negative SP-1/5 not tested 002-002 FVEP- oralis-2/5; 03 Resistant insoluble No Ct negative no identification 33F; 35 33F-like* SP-3/5 not tested 002-002 mitis-3/5 infant-1/5; FVEP- not 04 Resistant No Ct negative S. pneumoniae 19B 19B-like* mitis-4/5; SP-2/5 not tested 002-002 readable oralis-1/5 FVEP- not 05 Resistant insoluble No Ct negative S. mitis/S. oralis negative not tested not tested not tested 002-002 tested FVEP- not not 06 Resistant No Ct negative S. mitis/S. oralis not tested not tested not tested not tested 002-002 readable tested oralis-3/5; FVEP- Pseudomonas 07 Resistant insoluble No Ct negative negative 14-like* mitis-4/5; SP-3/5 not tested 002-002 fluorescens infant-1/5 FVEP- not not 08 Resistant No Ct negative S. mitis/S. oralis not tested not tested not tested not tested 002-002 readable tested FVEP- not not 09 Resistant No Ct negative S. mitis/S. oralis not tested not tested not tested not tested 002-002 readable tested infant-3/5; FVEP- 36/33A/39- 01 Resistant insoluble No Ct negative no identification negative mitis-2/5; SP-2/5 not tested 002-488 like* oralis-1/5 FVEP- not infant-3/5; 02 Resistant No Ct negative S. mitis/S. oralis not tested 48-like* SP-2/5 not tested 002-488 readable mitis-2/5 FVEP- not NT4a*(54%)+ mitis-5/5; 03 Resistant No Ct negative S. mitis/S. oralis not tested SP-3/5 not tested 002-488 readable 43-like*(46%) oralis-3/5 FVEP- 04 Resistant insoluble No Ct negative S. parasanguinis negative 10C/21-like* negative SP-1/5 not tested 002-488 279 Microarray lytA qPCR MALDI-TOF Latex Multiplex Sample Isolate Optochin Bile test Ct MS agglutination PCR Result serotyping StrepID PathID value FVEP- Rothia 05 Resistant insoluble No Ct negative negative negative negative SP-0/5 not tested 002-488 mucilaginosa FVEP- S. salivarius ssp. 06 Resistant insoluble No Ct negative 19B; 12 7F/36-like* saliva-5/5 SP-1/5 not tested 002-488 salivarius FVEP- not 01 Resistant insoluble No Ct negative S. mitis/S. oralis negative not tested not tested not tested 002-004 tested FVEP- not 10C/21/7F- 02 Resistant No Ct negative no identification negative saliva-5/5 SP-1/5 not tested 002-004 readable like* FVEP- not 03 Resistant insoluble No Ct negative S. mitis/S. oralis negative not tested not tested not tested 002-004 tested FVEP- not 04 Resistant insoluble No Ct negative S. mitis/S. oralis negative not tested not tested not tested 002-004 tested FVEP- not 05 Resistant insoluble No Ct negative S. mitis/S. oralis negative not tested not tested not tested 002-004 tested FVEP- not 01 Resistant insoluble No Ct negative S. mitis/S. oralis negative not tested not tested not tested 002-084 tested FVEP- not 02 Resistant insoluble No Ct negative S. mitis/S. oralis negative not tested not tested not tested 002-084 tested FVEP- NT4a*(55%)+ 03 Resistant insoluble No Ct negative Paracoccus yeei negative mitis-3/5 SP-3/5 not tested 002-084 36-like*(45%) FVEP- not 04 Resistant insoluble No Ct negative S. mitis/S. oralis negative not tested not tested not tested 002-084 tested FVEP- not not 05 Intermediate No Ct negative S. mitis/S. oralis not tested not tested not tested not tested 002-084 readable tested FVEP- not 06 Resistant insoluble No Ct negative S. mitis/S. oralis negative not tested not tested not tested 002-084 tested saliva-5/5; FVEP- S. salivarius ssp. 07 Resistant insoluble No Ct negative 19B; 12 36-like* mitis-1/5; SP-1/5 not tested 002-084 salivarius pneumo-1/5 FVEP- not 08 Resistant insoluble No Ct negative S. mitis/S. oralis negative not tested not tested not tested 002-084 tested 280 Microarray lytA qPCR MALDI-TOF Latex Multiplex Sample Isolate Optochin Bile test Ct MS agglutination PCR Result serotyping StrepID PathID value FVEP- oralis-3/5; 01 Resistant insoluble No Ct negative S. mitis/S. oralis 35 - 7F/33A-like* SP-3/5 not tested 002-078 mitis-3/5 FVEP- not 02 Resistant insoluble No Ct negative S. mitis/S. oralis negative not tested not tested not tested 002-078 tested FVEP- not 03 Resistant insoluble No Ct negative no identification negative not tested not tested not tested 002-078 tested FVEP- not S. salivarius ssp. not 04 Resistant No Ct negative not tested not tested not tested not tested 002-078 readable salivarius tested FVEP- not 05 Resistant insoluble No Ct negative S. parasanguinis negative not tested not tested not tested 002-078 tested FVEP- not 06 Resistant insoluble No Ct negative S. mitis/S. oralis negative not tested not tested not tested 002-078 tested FVEP- not not 07 Resistant No Ct negative no identification not tested not tested not tested not tested 002-078 readable tested FVEP- not 08 Resistant insoluble No Ct negative S. anginosus negative not tested not tested not tested 002-078 tested 7F/21/33A- FVEP- oralis-3/5; 09 Resistant insoluble No Ct negative S. mitis/S. oralis 35 - like*(65%)+N SP-3/5 not tested 002-078 mitis-3/5 T4b*(35%) FVEP- not 01 Resistant insoluble No Ct negative S. mitis/S. oralis negative not tested not tested not tested 002-496 tested FVEP- not S. salivarius ssp. not 02 Resistant No Ct negative not tested not tested not tested not tested 002-496 readable salivarius tested FVEP- not 03 Resistant insoluble No Ct negative S. mitis/S. oralis negative not tested not tested not tested 002-496 tested FVEP- not parasa-5/5; 04 Resistant No Ct negative no identification not tested 36-like* SP-1/5 not tested 002-496 readable pseudo-1/5 FVEP- parasa-5/5; 05 Resistant insoluble No Ct negative S. parasanguinis negative 10C/21-like* SP-1/5 not tested 002-496 mitis-2/5 FVEP- not 06 Resistant insoluble No Ct negative no identification negative not tested not tested not tested 002-496 tested 281 Microarray lytA qPCR MALDI-TOF Latex Multiplex Sample Isolate Optochin Bile test Ct MS agglutination PCR Result serotyping StrepID PathID value FVEP- not 07 Resistant insoluble No Ct negative S. mitis/S. oralis negative not tested not tested not tested 002-496 tested FVEP- sangui-1/5; 08 Resistant insoluble No Ct negative S. cristatus 19B 24F/16F-like* SP-1/5 not tested 002-496 mitis-1/5 FVEP- S. salivarius ssp. 09 Resistant insoluble No Ct negative 19B; 12 36-like* saliva-5/5 SP-1/5 not tested 002-496 salivarius FVEP- not S. salivarius ssp. 01 Resistant 38.28 equivocal not tested 7F-like* saliva-5/5 SP-1/5 not tested 002-080 readable salivarius FVEP- mitis-3/5; 02 Resistant insoluble No Ct negative S. mitis/S. oralis 19B 19B-like* SP-2/5 not tested 002-080 infant-2/5 FVEP- not not 03 Resistant No Ct negative S. mitis/S. oralis not tested not tested not tested not tested 002-080 readable tested FVEP- not 04 Resistant insoluble No Ct negative S. mitis/S. oralis negative not tested not tested not tested 002-080 tested FVEP- mitis-3/5; 05 Resistant insoluble No Ct negative S. pneumoniae negative NT4b* SP-3/5 not tested 002-080 pneumo-1/5 FVEP- not not 06 Resistant No Ct negative S. mitis/S. oralis not tested not tested not tested not tested 002-080 readable tested FVEP- not 07 Resistant insoluble No Ct negative S. parasanguinis negative not tested not tested not tested 002-080 tested FVEP- not not 08 Resistant No Ct negative S. parasanguinis not tested not tested not tested not tested 002-080 readable tested SP-0/5; BP-4/5; FVEP- not Rothia 09 Resistant No Ct negative not tested negative negative BC-4/5; not tested 002-080 readable mucilaginosa CD-3/5; PA-4/5 FVEP- not 10 Resistant insoluble No Ct negative S. parasanguinis negative not tested not tested not tested 002-080 tested FVEP- not 11 Sensitive soluble 18.31 positive S. pneumoniae 1 not tested not tested not tested 002-080 tested 282 Microarray lytA qPCR MALDI-TOF Latex Multiplex Sample Isolate Optochin Bile test Ct MS agglutination PCR Result serotyping StrepID PathID value FVEP- not not 01 Resistant No Ct negative no identification not tested not tested not tested not tested 002-418 readable tested FVEP- not S. salivarius ssp. 02 Resistant 35.23 equivocal not tested 7F-like* saliva-5/5 SP-1/5 not tested 002-418 readable salivarius FVEP- oralis-3/5; 03 Resistant insoluble 37.55 equivocal S. mitis/S. oralis negative 16A-like* SP-3/5 not tested 002-418 mitis-4/5 FVEP- not not 04 Resistant No Ct negative no identification not tested not tested not tested not tested 002-418 readable tested FVEP- not 05 Resistant insoluble No Ct negative S. mitis/S. oralis negative not tested not tested not tested 002-418 tested 7F/36- FVEP- not oralis-1/5; 06 Resistant 37.22 equivocal S. mitis/S. oralis not tested like*(53%)+N SP-3/5 not tested 002-418 readable mitis-4/5 T4a*(47%) FVEP- not 07 Resistant insoluble No Ct negative S. mitis/S. oralis negative not tested not tested not tested 002-418 tested FVEP- not 08 Resistant insoluble No Ct negative S. mitis/S. oralis negative not tested not tested not tested 002-418 tested FVEP- not 09 Resistant insoluble No Ct negative S. anginosus negative not tested not tested not tested 002-418 tested 10A + FVEP- 33A/21/10A- mitis-3/5; 01 Resistant insoluble No Ct negative S. mitis/S. oralis negative SP-3/5 33F/33A/37 002-460 like* oralis-2/5 + 5 + 23B FVEP- not S. salivarius ssp. 02 Resistant 37.97 equivocal not tested 7F-like* saliva-5/5 SP-1/5 10A +13 002-460 readable salivarius FVEP- 03 Sensitive soluble 17.61 positive S. pneumoniae 35B 35B pneumo-5/5 SP-5/5 35B 002-460 infant-3/5; FVEP- 15A/15A + 04 Resistant insoluble No Ct negative S. mitis/S. oralis negative 20-like* mitis-2/5; SP-2/5 002-460 31 + 20 oralis-1/5 infant-1/5; FVEP- not 05 Resistant 39.61 equivocal S. mitis/S. oralis not tested 36-like* pneumo-1/5 SP-1/5 20 002-460 readable mitis-2/5 283 Microarray lytA qPCR MALDI-TOF Latex Multiplex Sample Isolate Optochin Bile test Ct MS agglutination PCR Result serotyping StrepID PathID value NT4b*(55%)+ FVEP- mitis-4/5; 06 Resistant insoluble No Ct negative S. mitis/S. oralis negative 43/29- SP-3/5 13 002-460 pneumo-1/5 like*(45%) 6A/6B/6C/6 FVEP- D + 07 Resistant insoluble 36.9 equivocal S. mitis/S. oralis negative NT4b* mitis-3/5 SP-3/5 002-460 15A/15F +10A 6A/6B/6C/6 D + FVEP- 08 Resistant insoluble No Ct negative S. anginosus negative 7F/21-like* angino-2/5 SP-1/5 22F/22A + 002-460 35F/47F + 6C/6D 7F/7A + 9V/9A + 1 FVEP- oralis-5/5; 09 Resistant insoluble No Ct negative no identification negative 10C/21-like* SP-2/5 + 002-460 mitis-3/5 35A/35C/42 + 2 FVEP- 22F/22A + 10 Resistant insoluble No Ct negative S. parasanguinis negative 7F/36-like* parasa-1/5 SP-1/5 002-460 7F/7A + 5 FVEP- not 01 Resistant insoluble No Ct negative S. mitis/S. oralis not tested not tested not tested not tested 002-422 tested FVEP- not 02 Sensitive soluble 18.26 positive S. pneumoniae 15B not tested not tested not tested 002-422 tested FVEP- not S. salivarius ssp. 03 Resistant 39.57 equivocal not tested 7F-like* saliva-5/5 SP-1/5 not tested 002-422 readable salivarius FVEP- not 04 Resistant insoluble No Ct negative S. parasanguinis not tested not tested not tested not tested 002-422 tested FVEP- not NT4a*(65%)+ 05 Resistant No Ct negative S. pneumoniae not tested mitis-4/5 SP-3/5 not tested 002-422 readable 7F-like*(35%) FVEP- not 06 Resistant insoluble No Ct negative S. parasanguinis not tested not tested not tested not tested 002-422 tested 284 Microarray lytA qPCR MALDI-TOF Latex Multiplex Sample Isolate Optochin Bile test Ct MS agglutination PCR Result serotyping StrepID PathID value FVEP- not not 07 Resistant No Ct negative S. mitis/S. oralis not tested not tested not tested not tested 002-422 readable tested 7F- FVEP- not S. salivarius ssp. like*(67%)+N saliva-5/5; 01 Resistant 38.43 equivocal not tested SP-3/5 not tested 002-486 readable salivarius T4b*(20%)+21 mitis-5/5 -like*(13%) SP-3/5; mitis-2/5; BP-3/5; FVEP- not Rothia 02 Resistant 39.49 equivocal not tested NT4b* pseudo-1/5; BC-3/5; not tested 002-486 readable mucilaginosa pneumo-2/5 CD-3/5; PA-3/5 FVEP- not 03 Resistant insoluble No Ct negative S. parasanguinis not tested not tested not tested not tested 002-486 tested FVEP- not 04 Resistant insoluble No Ct negative S. parasanguinis not tested not tested not tested not tested 002-486 tested FVEP- not 05 Resistant soluble No Ct negative S. mitis/S. oralis negative not tested not tested not tested 002-486 tested FVEP- not 06 Resistant insoluble No Ct negative S. mitis/S. oralis not tested not tested not tested not tested 002-486 tested FVEP- not 07 Resistant insoluble No Ct negative S. parasanguinis not tested not tested not tested not tested 002-486 tested FVEP- 08 Resistant soluble No Ct negative no identification negative negative negative SP-0/5 not tested 002-486 285 7.3.2 Full serotyping data for adult NP samples Table 7.3. Full serotyping data for NVEP adult NP samples 2012 to 2015. 2012 2013 2014 2015 n % n % n % n % 1 1 1.8 2 2.8 0 0.0 0 0.0 2 0 0.0 1 1.4 0 0.0 0 0.0 3 1 1.8 3 4.2 2 5.4 2 5.0 4 1 1.8 0 0.0 0 0.0 0 0.0 6A 1 1.8 3 4.2 3 8.1 0 0.0 6B 1 1.8 3 4.2 0 0.0 0 0.0 6C 1 1.8 1 1.4 1 2.7 2 5.0 6D 0 0.0 3 4.2 0 0.0 0 0.0 7F 0 0.0 1 1.4 0 0.0 0 0.0 7B 1 1.8 0 0.0 0 0.0 0 0.0 7C 0 0.0 1 1.4 2 5.4 3 7.5 8 3 5.3 1 1.4 0 0.0 0 0.0 9V 2 3.5 1 1.4 1 2.7 0 0.0 10F 0 0.0 0 0.0 0 0.0 1 2.5 10A 0 0.0 1 1.4 0 0.0 0 0.0 10B 0 0.0 0 0.0 0 0.0 1 2.5 11F-like 2 3.5 1 1.4 2 5.4 2 5.0 12F 0 0.0 1 1.4 0 0.0 0 0.0 13 1 1.8 6 8.5 1 2.7 3 7.5 14 2 3.5 0 0.0 3 8.1 2 5.0 15B 4 7.0 4 5.6 2 5.4 0 0.0 15C 0 0.0 1 1.4 0 0.0 0 0.0 16F 5 8.8 7 9.9 1 2.7 0 0.0 16A 1 1.8 0 0.0 0 0.0 0 0.0 17F 2 3.5 0 0.0 0 0.0 0 0.0 18A 1 1.8 0 0.0 1 2.7 0 0.0 18C 2 3.5 0 0.0 0 0.0 0 0.0 19F 1 1.8 2 2.8 4 10.8 0 0.0 19A 0 0.0 1 1.4 0 0.0 3 7.5 20B 0 0.0 0 0.0 1 2.7 1 2.5 21 0 0.0 1 1.4 0 0.0 2 5.0 22F 2 3.5 2 2.8 0 0.0 0 0.0 23F 2 3.5 8 11.3 0 0.0 2 5.0 23A 0 0.0 0 0.0 0 0.0 2 5.0 23B 1 1.8 0 0.0 0 0.0 1 2.5 24B 0 0.0 0 0.0 0 0.0 1 2.5 28A 1 1.8 2 2.8 0 0.0 0 0.0 31 1 1.8 0 0.0 1 2.7 0 0.0 33F 1 1.8 1 1.4 0 0.0 2 5.0 34 3 5.3 2 2.8 1 2.7 1 2.5 286 2012 2013 2014 2015 n % n % n % n % 35F 1 1.8 0 0.0 0 0.0 0 0.0 35A 1 1.8 0 0.0 1 2.7 1 2.5 35B 0 0.0 0 0.0 0 0.0 2 5.0 39 3 5.3 0 0.0 2 5.4 0 0.0 42 0 0.0 0 0.0 0 0.0 1 2.5 NT 10 17.5 11 15.5 8 21.6 5 12.5 Total 57 100 71 100 37 100 40 100 287 References 288 References 1. Marimon JM, Ercibengoa M, Garcia-Arenzana JM, Alonso M, Perez- Trallero E. 2013. Streptococcus pneumoniae ocular infections, prominent role of unencapsulated isolates in conjunctivitis. Clin Microbiol Infect 19:E298-305. 2. Priftis KN, Litt D, Manglani S, Anthracopoulos MB, Thickett K, Tzanakaki G, Fenton P, Syrogiannopoulos GA, Vogiatzi A, Douros K, Slack M, Everard ML. 2013. Bacterial bronchitis caused by Streptococcus pneumoniae and nontypable Haemophilus influenzae in children: the impact of vaccination. Chest 143:152-157. 3. McNeil JC, Hulten KG, Mason EO, Jr., Kaplan SL. 2009. Serotype 19A is the most common Streptococcus pneumoniae isolate in children with chronic sinusitis. Pediatr Infect Dis J 28:766-768. 4. Krishna S, Sanjeevan KV, Sudheer A, Dinesh KR, Kumar A, Karim S. 2012. Pneumococcusuria: from bench to bedside. Indian J Med Microbiol 30:96-98. 5. Garcia-Lechuz JM, Cuevas O, Castellares C, Perez-Fernandez C, Cercenado E, Bouza E. 2007. Streptococcus pneumoniae skin and soft tissue infections: characterization of causative strains and clinical illness. Eur J Clin Microbiol Infect Dis 26:247-253. 6. Pääkkönen M, Peltola H. 2013. Bone and Joint Infections. Pediatr Clin North Am 60:425-436. 7. Baraboutis I, Skoutelis A. 2004. Streptococcus pneumoniae septic arthritis in adults. Clin Microbiol Infect 10:1037-1039. 8. Waisman DC, Tyrrell GJ, Kellner JD, Garg S, Marrie TJ. 2010. Pneumococcal peritonitis: still with us and likely to increase in importance. Can J Infect Dis Med Microbiol 21:e23-27. 289 9. Kan B, Ries J, Normark BH, Chang FY, Feldman C, Ko WC, Rello J, Snydman DR, Yu VL, Ortqvist A. 2006. Endocarditis and pericarditis complicating pneumococcal bacteraemia, with special reference to the adhesive abilities of pneumococci: results from a prospective study. Clin Microbiol Infect 12:338-344. 10. Feinstein Y, Falup-Pecurariu O, Mitrică M, Berezin EN, Sini R, Krimko H, Greenberg D. 2010. Acute pericarditis caused by Streptococcus pneumoniae in young infants and children: Three case reports and a literature review. Int J Infect Dis 14:e175-e178. 11. O'Brien KL, Wolfson LJ, Watt JP, Henkle E, Deloria-Knoll M, McCall N, Lee E, Mulholland K, Levine OS, Cherian T. 2009. Burden of disease caused by Streptococcus pneumoniae in children younger than 5 years: global estimates. Lancet 374:893-902. 12. Walker CL, Rudan I, Liu L, Nair H, Theodoratou E, Bhutta ZA, O'Brien KL, Campbell H, Black RE. 2013. Global burden of childhood pneumonia and diarrhoea. Lancet 381:1405-1416. 13. Bogaert D, de Groot R, Hermans PWM. 2004. Streptococcus pneumoniae colonisation: the key to pneumococcal disease. Lancet Infect Dis 4:144-154. 14. Simell B, Auranen K, Kayhty H, Goldblatt D, Dagan R, O'Brien KL. 2012. The fundamental link between pneumococcal carriage and disease. Expert Rev Vaccines 11:841-855. 15. Gray BM, Converse GM, 3rd, Dillon HC, Jr. 1980. Epidemiologic studies of Streptococcus pneumoniae in infants: acquisition, carriage, and infection during the first 24 months of life. J Infect Dis 142:923-933. 16. Klugman KP, Chien YW, Madhi SA. 2009. Pneumococcal pneumonia and influenza: a deadly combination. Vaccine 27 Suppl 3:C9-C14. 17. Backhaus E, Berg S, Andersson R, Ockborn G, Malmström P, Dahl M, Nasic S, Trollfors B. 2016. Epidemiology of invasive pneumococcal infections: 290 manifestations, incidence and case fatality rate correlated to age, gender and risk factors. BMC Infect Dis 16:367. 18. Lynch JP, III, Zhanel GG. 2010. Streptococcus pneumoniae: epidemiology and risk factors, evolution of antimicrobial resistance, and impact of vaccines. Curr Opin Pulm Med 16:217-225. 19. Cruickshank HC, Jefferies JM, Clarke SC. 2014. Lifestyle risk factors for invasive pneumococcal disease: a systematic review. BMJ Open 4:e005224. 20. Sternberg GM. 1881. A fatal form of septicaemia in the rabbit, produced by subcutaneous injection of human saliva. Natl Board Health Bull 2:781-783. 21. Pasteur L. 1881. Note sur la maladie nouvelle provoquée par la salive d'un enfant mort de la rage. Bull Acad Méd (Paris) 10:94-103. 22. Buchanan RE, Gibbons NR. 1974. Bergey's manual of determinative bacteriolgy, 8th ed. Williams & Wilkins Co., Baltimore. 23. Kellogg JA, Bankert DA, Elder CJ, Gibbs JL, Smith MC. 2001. Identification of Streptococcus pneumoniae revisited. J Clin Microbiol 39:3373- 3375. 24. Mundy LS, Janoff EN, Schwebke KE, Shanholtzer CJ, Willard KE. 1998. Ambiguity in the identification of Streptococcus pneumoniae. Optochin, bile solubility, quellung, and the AccuProbe DNA probe tests. Am J Clin Pathol 109:55-61. 25. Morgenroth J, Kaufmann M. 1912. Arzneifestigkeit bei bakterien (Pneumokokken). Ztschr f Immunitätsforsch 15:610-624. 26. Moore HF, Chesney AM. 1917. A study of ethylhydrocuprein (optochin) in the treatment of acute lobar pneumonia. Arch Int Med 19:611-682. 27. Balsalobre L, Hernández-Madrid A, Llull D, Martín-Galiano AJ, García E, Fenoll A, de la Campa AG. 2006. Molecular characterization of disease- 291 associated streptococci of the mitis group that are optochin susceptible. J Clin Microbiol 44:4163-4171. 28. Borek AP, Dressel DC, Hussong J, Peterson LR. 1997. Evolving clinical problems with Streptococcus pneumoniae: increasing resistance to antimicrobial agents, and failure of traditional optochin identification in Chicago, Illinois, between 1993 and 1996. Diagn Microbiol Infect Dis 29:209-214. 29. Obregón V, García P, García E, Fenoll A, López R, García JL. 2002. Molecular peculiarities of the lytA gene isolated from clinical pneumococcal strains that are bile insoluble. J Clin Microbiol 40:2545-2554. 30. Whatmore AM, Efstratiou A, Pickerill AP, Broughton K, Woodard G, Sturgeon D, George R, Dowson CG. 2000. Genetic relationships between clinical isolates of Streptococcus pneumoniae, Streptococcus oralis, and Streptococcus mitis: characterization of “atypical” pneumococci and organisms allied to S. mitis harboring S. pneumoniae virulence factor-encoding genes. Infect Immun 68:1374-1382. 31. Satzke C, Turner P, Virolainen-Julkunen A, Adrian PV, Antonio M, Hare KM, Henao-Restrepo AM, Leach AJ, Klugman KP, Porter BD, Sa-Leao R, Scott JA, Nohynek H, O'Brien KL. 2013. Standard method for detecting upper respiratory carriage of Streptococcus pneumoniae: updated recommendations from the World Health Organization Pneumococcal Carriage Working Group. Vaccine 32:165-179. 32. Perilla MJ, Ajello G, Bopp C, Elliott J, Facklam R, Knapp JS, Popovic T, Wells J, Dowell SF, World Health Organization. Dept. of Epidemic and Pandemic Alert and Response, National Center for Infectious Diseases (U. S.). 2003. Manual for the laboratory identification and antimicrobial susceptibility testing of bacterial pathogens of public health importance in the developing world. World Health Organization, Geneva, 33. Carvalho MdG, Pimenta FC, Jackson D, Roundtree A, Ahmad Y, Millar EV, O'Brien KL, Whitney CG, Cohen AL, Beall BW. 2010. Revisiting 292 pneumococcal carriage by use of broth enrichment and PCR techniques for enhanced detection of carriage and serotypes. J Clin Microbiol 48:1611-1618. 34. Toikka P, Nikkari S, Ruuskanen O, Leinonen M, Mertsola J. 1999. Pneumolysin PCR-based diagnosis of invasive pneumococcal infection in children. J Clin Microbiol 37:633-637. 35. Morrison KE, Lake D, Crook J, Carlone GM, Ades E, Facklam R, Sampson JS. 2000. Confirmation of psaA in all 90 serotypes of Streptococcus pneumoniae by PCR and potential of this assay for identification and diagnosis. J Clin Microbiol 38:434-437. 36. Suzuki N, Yuyama M, Maeda S, Ogawa H, Mashiko K, Kiyoura Y. 2006. Genotypic identification of presumptive Streptococcus pneumoniae by PCR using four genes highly specific for S. pneumoniae. J Med Microbiol 55:709- 714. 37. Park HK, Yoon JW, Shin JW, Kim JY, Kim W. 2010. rpoA is a useful gene for identification and classification of Streptococcus pneumoniae from the closely related viridans group streptococci. FEMS Microbiol Lett 305:58-64. 38. Hoshino T, Fujiwara T, Kilian M. 2005. Use of phylogenetic and phenotypic analyses to identify nonhemolytic streptococci isolated from bacteremic patients. J Clin Microbiol 43:6073-6085. 39. Picard FJ, Ke D, Boudreau DK, Boissinot M, Huletsky A, Richard D, Ouellette M, Roy PH, Bergeron MG. 2004. Use of tuf sequences for genus- specific PCR detection and phylogenetic analysis of 28 streptococcal species. J Clin Microbiol 42:3686-3695. 40. Zbinden A, Köhler N, Bloemberg GV. 2011. recA-based PCR assay for accurate differentiation of Streptococcus pneumoniae from other viridans streptococci. J Clin Microbiol 49:523-527. 41. Trzciński K, Bogaert D, Wyllie A, Chu MLJN, van der Ende A, Bruin JP, van den Dobbelsteen G, Veenhoven RH, Sanders EAM. 2013. Superiority of 293 trans-oral over trans-nasal sampling in detecting Streptococcus pneumoniae colonization in adults. PLoS One 8:e60520. 42. Wyllie AL, Wijmenga-Monsuur AJ, van Houten MA, Bosch AATM, Groot JA, van Engelsdorp Gastelaars J, Bruin JP, Bogaert D, Rots NY, Sanders EAM, Trzciński K. 2016. Molecular surveillance of nasopharyngeal carriage of Streptococcus pneumoniae in children vaccinated with conjugated polysaccharide pneumococcal vaccines. Sci Rep 6:23809. 43. Dubois D, Segonds C, Prere M-F, Marty N, Oswald E. 2013. Identification of clinical Streptococcus pneumoniae isolates among other alpha and nonhemolytic streptococci by use of the Vitek MS matrix-assisted laser desorption ionization– time of flight mass spectrometry system. J Clin Microbiol 51:1861-1867. 44. Aanensen DM, Spratt BG. 2005. The multilocus sequence typing network: mlst.net. Nucleic Acids Res 33:W728-W733. 45. Enright MC, Spratt BG. 1998. A multilocus sequence typing scheme for Streptococcus pneumoniae: identification of clones associated with serious invasive disease. Microbiology 144:3049-3060. 46. Nelson AL, Roche AM, Gould JM, Chim K, Ratner AJ, Weiser JN. 2007. Capsule enhances pneumococcal colonization by limiting mucus-mediated clearance. Infect Immun 75:83-90. 47. Geno KA, Gilbert GL, Song JY, Skovsted IC, Klugman KP, Jones C, Konradsen HB, Nahm MH. 2015. Pneumococcal Ccapsules and their types: past, present, and future. Clin Microbiol Rev 28:871-899. 48. Mavroidi A, Aanensen DM, Godoy D, Skovsted IC, Kaltoft MS, Reeves PR, Bentley SD, Spratt BG. 2007. Genetic relatedness of the Streptococcus pneumoniae capsular biosynthetic loci. J Bacteriol 189:7841-7855. 49. Bentley SD, Aanensen DM, Mavroidi A, Saunders D, Rabbinowitsch E, Collins M, Donohoe K, Harris D, Murphy L, Quail MA, Samuel G, Skovsted IC, Kaltoft MS, Barrell B, Reeves PR, Parkhill J, Spratt BG. 2006. 294 Genetic analysis of the capsular biosynthetic locus from all 90 pneumococcal serotypes. PLoS Genet 2:e31. 50. Hausdorff WP, Feikin DR, Klugman KP. 2005. Epidemiological differences among pneumococcal serotypes. Lancet Infect Dis 5:83-93. 51. Sleeman KL, Griffiths D, Shackley F, Diggle L, Gupta S, Maiden MC, Moxon ER, Crook DW, Peto TEA. 2006. Capsular serotype-specific attack rates and duration of carriage of Streptococcus pneumoniae in a population of children. J Infect Dis 194:682-688. 52. Johnson HL, Deloria-Knoll M, Levine OS, Stoszek SK, Freimanis Hance L, Reithinger R, Muenz LR, O'Brien KL. 2010. Systematic evaluation of serotypes causing invasive pneumococcal disease among children under five: the pneumococcal global serotype project. PLoS Med 7:e1000348. 53. Hyams C, Yuste J, Bax K, Camberlein E, Weiser JN, Brown JS. 2010. Streptococcus pneumoniae resistance to complement-mediated immunity is dependent on the capsular serotype. Infect Immun 78:716-725. 54. Melin M, Trzciński K, Meri S, Käyhty H, Väkeväinen M. 2010. The capsular serotype of Streptococcus pneumoniae is more important than the genetic background for resistance to complement. Infect Immun 78:5262-5270. 55. Weinberger DM, Trzciński K, Lu Y-J, Bogaert D, Brandes A, Galagan J, Anderson PW, Malley R, Lipsitch M. 2009. Pneumococcal capsular polysaccharide structure predicts serotype prevalence. PLoS Pathog 5:e1000476. 56. Brueggemann AB, Griffiths DT, Meats E, Peto T, Crook DW, Spratt BG. 2003. Clonal relationships between invasive and carriage Streptococcus pneumoniae and serotype- and clone-specific differences in invasive disease potential. J Infect Dis 187:1424-1432. 57. Lipsitch M, O'Neill K, Cordy D, Bugalter B, Trzcinski K, Thompson CM, Goldstein R, Pelton S, Huot H, Bouchet V, Reid R, Santosham M, O'Brien KL. 2007. Strain characteristics of Streptococcus pneumoniae carriage and 295 invasive disease isolates during a cluster-randomized clinical trial of the 7-valent pneumococcal conjugate vaccine. J Infect Dis 196:1221-1227. 58. Mizrachi Nebenzahl Y, Porat N, Lifshitz S, Novick S, Levi A, Ling E, Liron O, Mordechai S, Sahu RK, Dagan R. 2004. Virulence of Streptococcus pneumoniae may be determined independently of capsular polysaccharide. FEMS Microbiol Lett 233:147-152. 59. Xu Q, Kaur R, Casey JR, Sabharwal V, Pelton S, Pichichero ME. 2011. Nontypeable Streptococcus pneumoniae as an otopathogen. Diagn Microbiol Infect Dis 69:200-204. 60. Haas W, Hesje CK, Sanfilippo CM, Morris TW. 2011. High proportion of nontypeable Streptococcus pneumoniae isolates among sporadic, nonoutbreak cases of bacterial conjunctivitis. Curr Eye Res 36:1078-1085. 61. Keller LE, Robinson DA, McDaniel LS. 2016. Nonencapsulated Streptococcus pneumoniae: emergence and pathogenesis. mBio 7:e01792-15. 62. Park IH, Kim KH, Andrade AL, Briles DE, McDaniel LS, Nahm MH. 2012. Nontypeable pneumococci can be divided into multiple cps types, including one type expressing the novel gene pspK. mBio 3:e00035-12. 63. Hathaway LJ, Meier PS, Bättig P, Aebi S, Mühlemann K. 2004. A homologue of aliB is found in the capsule region of nonencapsulated Streptococcus pneumoniae. J Bacteriol 186:3721-3729. 64. Keller LE, Bradshaw JL, Pipkins H, McDaniel LS. 2016. Surface proteins and pneumolysin of encapsulated and nonencapsulated Streptococcus pneumoniae mediate virulence in a chinchilla model of otitis media. Front Cell Infect Microbiol 6:55. 65. Kadioglu A, Weiser JN, Paton JC, Andrew PW. 2008. The role of Streptococcus pneumoniae virulence factors in host respiratory colonization and disease. Nat Rev Microbiol 6:288-301. 296 66. Anderson R, Feldman C. 2011. Key virulence factors of Streptococcus pneumoniae and non-typeable Haemophilus influenzae: roles in host defence and immunisation. South Afr J Epidemiol Infect 26:6-12. 67. Wren JT, Blevins LK, Pang B, Basu Roy A, Oliver MB, Reimche JL, Wozniak JE, Alexander-Miller MA, Swords WE. 2017. Pneumococcal neuraminidase A (NanA) promotes biofilm formation and synergizes with influenza A virus in nasal colonization and middle ear infection. Infect Immun 85:e01044-01016. 68. Johnsborg O, Havarstein LS. 2009. Regulation of natural genetic transformation and acquisition of transforming DNA in Streptococcus pneumoniae. FEMS Microbiol Rev 33:627-642. 69. Straume D, Stamsas GA, Havarstein LS. 2015. Natural transformation and genome evolution in Streptococcus pneumoniae. Infect Genet Evol 33:371-380. 70. Croucher NJ, Harris SR, Fraser C, Quail MA, Burton J, van der Linden M, McGee L, von Gottberg A, Song JH, Ko KS, Pichon B, Baker S, Parry CM, Lambertsen LM, Shahinas D, Pillai DR, Mitchell TJ, Dougan G, Tomasz A, Klugman KP, Parkhill J, Hanage WP, Bentley SD. 2011. Rapid pneumococcal evolution in response to clinical interventions. Science 331:430-434. 71. Marks LR, Parameswaran GI, Hakansson AP. 2012. Pneumococcal interactions with epithelial cells are crucial for optimal biofilm formation and colonization in vitro and in vivo. Infect Immun 80:2744-2760. 72. Greenberg D, Broides A, Blancovich I, Peled N, Givon-Lavi N, Dagan R. 2004. Relative importance of nasopharyngeal versus oropharyngeal sampling for isolation of Streptococcus pneumoniae and Haemophilus influenzae from healthy and sick individuals varies with age. J Clin Microbiol 42:4604-4609. 73. Short KR, Reading PC, Wang N, Diavatopoulos DA, Wijburg OL. 2012. Increased nasopharyngeal bacterial titers and local inflammation facilitate transmission of Streptococcus pneumoniae. mBio 3: e00255-12. 297 74. Diavatopoulos DA, Short KR, Price JT, Wilksch JJ, Brown LE, Briles DE, Strugnell RA, Wijburg OL. 2010. Influenza A virus facilitates Streptococcus pneumoniae transmission and disease. FASEB J 24:1789-1798. 75. Zafar MA, Kono M, Wang Y, Zangari T, Weiser JN. 2016. Infant mouse model for the study of shedding and transmission during Streptococcus pneumoniae monoinfection. Infect Immun 84:2714-2722. 76. Kono M, Zafar MA, Zuniga M, Roche AM, Hamaguchi S, Weiser JN. 2016. Single cell bottlenecks in the pathogenesis of Streptococcus pneumoniae. PLoS Pathog 12:e1005887. 77. Cremers A, Zomer A, Gritzfeld J, Ferwerda G, van Hijum S, Ferreira D, Shak J, Klugman K, Boekhorst J, Timmerman H, de Jonge M, Gordon S, Hermans P. 2014. The adult nasopharyngeal microbiome as a determinant of pneumococcal acquisition. Microbiome 2:44. 78. Weinberger DM, Grant LR, Steiner CA, Weatherholtz R, Santosham M, Viboud C, O'Brien KL. 2014. Seasonal drivers of pneumococcal disease incidence: impact of bacterial carriage and viral activity. Clin Infect Dis 58:188- 194. 79. Adetifa IM, Antonio M, Okoromah CA, Ebruke C, Inem V, Nsekpong D, Bojang A, Adegbola RA. 2012. Pre-vaccination nasopharyngeal pneumococcal carriage in a Nigerian population: epidemiology and population biology. PLoS One 7:e30548. 80. Hill PC, Akisanya A, Sankareh K, Cheung YB, Saaka M, Lahai G, Greenwood BM, Adegbola RA. 2006. Nasopharyngeal carriage of Streptococcus pneumoniae in Gambian villagers. Clin Infect Dis 43:673-679. 81. Gray BM, Turner ME, Dillon HC, Jr. 1982. Epidemiologic studies of Streptococcus pneumoniae in infants: the effects of season and age on pneumococcal acquisition and carriage in the first 24 months of life. Am J Epidemiol 116:692-703. 298 82. Hill PC, Cheung YB, Akisanya A, Sankareh K, Lahai G, Greenwood BM, Adegbola RA. 2008. Nasopharyngeal carriage of Streptococcus pneumoniae in Gambian infants: a longitudinal study. Clin Infect Dis 46:807-814. 83. Granat SM, Mia Z, Ollgren J, Herva E, Das M, Piirainen L, Auranen K, Makela PH. 2007. Longitudinal study on pneumococcal carriage during the first year of life in Bangladesh. Pediatr Infect Dis J 26:319-324. 84. Brugger SD, Frey P, Aebi S, Hinds J, Muhlemann K. 2010. Multiple colonization with S. pneumoniae before and after introduction of the seven- valent conjugated pneumococcal polysaccharide vaccine. PLoS One 5:e11638. 85. Adegbola RA, DeAntonio R, Hill PC, Roca A, Usuf E, Hoet B, Greenwood BM. 2014. Carriage of Streptococcus pneumoniae and other respiratory bacterial pathogens in low and lower-middle income countries: a systematic review and meta-analysis. PLoS One 9:e103293. 86. Aniansson G, Alm B, Andersson B, Larsson P, Nylen O, Peterson H, Rigner P, Svanborg M, Svanborg C. 1992. Nasopharyngeal colonization during the first year of life. J Infect Dis 165 Suppl 1:S38-42. 87. Regev-Yochay G, Raz M, Dagan R, Porat N, Shainberg B, Pinco E, Keller N, Rubinstein E. 2004. Nasopharyngeal carriage of Streptococcus pneumoniae by adults and children in community and family settings. Clin Infect Dis 38:632- 639. 88. Greenberg D, Givon-Lavi N, Broides A, Blancovich I, Peled N, Dagan R. 2006. The contribution of smoking and exposure to tobacco smoke to Streptococcus pneumoniae and Haemophilus influenzae carriage in children and their mothers. Clin Infect Dis 42:897-903. 89. Jourdain S, Smeesters PR, Denis O, Dramaix M, Sputael V, Malaviolle X, Van Melderen L, Vergison A. 2011. Differences in nasopharyngeal bacterial carriage in preschool children from different socio-economic origins. Clin Microbiol Infect 17:907-914. 299 90. van der Poll T, Opal SM. 2009. Pathogenesis, treatment, and prevention of pneumococcal pneumonia. Lancet 374:1543-1556. 91. British Thoracic Society Standards of Care Committee. 2001. BTS guidelines for the management of community acquired pneumonia in adults. Thorax 56 (suppl 4):1-64. 92. Mandell LA, Wunderink RG, Anzueto A, Bartlett JG, Campbell GD, Dean NC, Dowell SF, File TM, Musher DM, Niederman MS, Torres A, Whitney CG. 2007. Infectious Diseases Society of America/American Thoracic Society consensus guidelines on the management of community-acquired pneumonia in adults. Clin Infect Dis 44:S27-S72. 93. Woodhead M, Blasi F, Ewig S, Huchon G, Leven M, Ortqvist A, Schaberg T, Torres A, van der Heijden G, Verheij TJM. 2005. Guidelines for the management of adult lower respiratory tract infections. Eur Respir J 26:1138- 1180. 94. Arias CA, Murray BE. 2009. Antibiotic-resistant bugs in the 21st century — a clinical super-challenge. N Engl J Med 360:439-443. 95. Reinert RR. 2009. The antimicrobial resistance profile of Streptococcus pneumoniae. Clin Microbiol Infect 15:7-11. 96. Jacobs MR. 2008. Antimicrobial-resistant Streptococcus pneumoniae: trends and management. Expert Rev Anti Infect Ther 6:619-635. 97. Goetghebuer T, West TE, Wermenbol V, Cadbury AL, Milligan P, Lloyd- Evans N, Adegbola RA, Mulholland EK, Greenwood BM, Weber MW. 2000. Outcome of meningitis caused by Streptococcus pneumoniae and Haemophilus influenzae type b in children in The Gambia. Trop Med Int Health 5:207-213. 98. Roth DE, Caulfield LE, Ezzati M, Black RE. 2008. Acute lower respiratory infections in childhood: opportunities for reducing the global burden through nutritional interventions. Bull World Health Organ 86:356-364. 300 99. Duffy LC, Faden H, Wasielewski R, Wolf J, Krystofik D. 1997. Exclusive breastfeeding protects against bacterial colonization and day care exposure to otitis media. Pediatrics 100:E7. 100. Smith KR, Samet JM, Romieu I, Bruce N. 2000. Indoor air pollution in developing countries and acute lower respiratory infections in children. Thorax 55:518-532. 101. Luby SP, Agboatwalla M, Feikin DR, Painter J, Billhimer W, Altaf A, Hoekstra RM. 2005. Effect of handwashing on child health: a randomised controlled trial. Lancet 366:225-233. 102. Klemperer G, Klemperer F. 1891. Versuche über Immunisierung und Heilung bei der Pneumokokkeninfection. Berl Klin Wochenstr 28:833-835. 103. Wright AE, Parry Morgan W, Colebrook L, Dodgson RW. 1914. Observations on prophylactic inoculation against pneumococcus infections, and on the results which have been achieved by it. Lancet 183:1-10. 104. Lister FS. 1916. An experimental study of prophylactic inoculation against pneumococcal infection in the rabbit and in man. Publ South Afr Inst Med Res 1:231-287. 105. Felton LD, Bailey GH. 1926. Biologic significance of the soluble specific substances of pneumococci. J Infect Dis 38:131-144. 106. Austrian R. 1981. The development of pneumococcal vaccine. Proc Am Philos Soc 125:46-51. 107. MacLeod CM, Hodges RG, Heidelberger M, Bernhard WG. 1945. Prevention of pneumococcal pneumonia by immunization with specific capsular polysaccharides. J Exp Med 82:445-465. 108. Heidelberger M, MacLeod CM, Di Lapi MM. 1948. The human antibody response to simultaneous injection of six specific polysaccharides of pneumococcus. J Exp Med 88:369-372. 301 109. Jacobs MR, Koornhof HJ, Robins-Browne RM, Stevenson CM, Vermaak ZA, Freiman I, Miller GB, Witcomb MA, Isaäcson M, Ward JI, Austrian R. 1978. Emergence of multiply resistant pneumococci. N Engl J Med 299:735- 740. 110. Austrian R, Douglas RM, Schiffman G, Coetzee AM, Koornhof HJ, Hayden-Smith S, Reid RD. 1976. Prevention of pneumococcal pneumonia by vaccination. Trans Assoc Am Physicians 89:184-194. 111. Robbins JB, Austrian R, Lee C-J, Rastogi SC, Schiffman G, Henrichsen J, Makela PH, Broome CV, Facklam RR, Tiesjema RH, Parke JC. 1983. Considerations for formulating the second-generation pneumococcal capsular polysaccharide vaccine with emphasis on the cross-reactive types within groups. J Infect Dis 148:1136-1159. 112. Borgoño JM, McLean AA, Vella PP, Woodhour AF, Canepa I, Davidson WL, Hilleman MR. 1978. Vaccination and revaccination with polyvalent pneumococcal polysaccharide vaccines in adults and infants. Proc Soc Exp Biol Med 157:148-154. 113. Leinonen M, Säkkinen A, Kalliokoski R, Luotonen J, Timonen M, Mäkelä PH. 1986. Antibody response to 14-valent pneumococcal capsular polysaccharide vaccine in pre-school age children. Pediatr Infect Dis 5:39-44. 114. Temple K, Greenwood B, Inskip H, Hall A, Koskela M, Leinonen M. 1991. Antibody response to pneumococcal capsular polysaccharide vaccine in African children. Pediatr Infect Dis J 10:386-390. 115. World Health Organization. 2008. 23-valent pneumococcal polysaccharide vaccine: WHO position paper. Wkly Epidemiol Rec 83:373-384. 116. Poolman J, Borrow R. 2011. Hyporesponsiveness and its clinical implications after vaccination with polysaccharide or glycoconjugate vaccines. Expert Rev Vaccines 10:307-322. 302 117. Douglas RM, Paton JC, Duncan SJ, Hansman DJ. 1983. Antibody response to Ppneumococcal vaccination in children younger than five years of age. J Infect Dis 148:131-137. 118. Wright PF, Sell SH, Vaughn WK, Andrews C, McConnell KB, Schiffman G. 1981. Clinical studies of pneumococcal vaccines in Infants. II. Efficacy and effect on nasopharyngeal carriage. Rev Infect Dis 3:S108-S112. 119. Sloyer JL, Ploussard JH, Howie VM. 1981. Efficacy of pneumococcal polysaccharide vaccine in preventing acute otitis media in infants in Huntsville, Alabama. Rev Infect Dis 3:S119-S123. 120. Rosén C, Christensen P, Henrichsen J, Hovelius B, Prellner K. 1984. Beneficial effect of pneumococcal vaccination on otitis media in children over two years old. Int J Pediatr Otorhinolaryngol 7:239-246. 121. Riley ID, Alpers MP, Gratten H, Lehmann D, Marshall TFD, Smith D. 1986. Pneumococcal vaccine prevents death from acute lower-respiratory-tract infections in Papua New Guinean children. Lancet 328:877-881. 122. Leach AJ, Ceesay SJ, Banya WAS, Greenwood BM. 1996. Pilot trial of a pentavalent pneumococcal polysaccharide/protein conjugate vaccine in Gambian infants. Pediatr Infect Dis J 15:333-339. 123. O'Brien KL, Steinhoff MC, Edwards K, Keyserling H, Thoms ML, Madore D. 1996. Immunologic priming of young children by pneumococcal glycoprotein conjugate, but not polysaccharide, vaccines. Pediatr Infect Dis J 15:425-430. 124. Rennels MB, Edwards KM, Keyserling HL, Reisinger KS, Hogerman DA, Madore DV, Chang I, Paradiso PR, Malinoski FJ, Kimura A. 1998. Safety and immunogenicity of heptavalent pneumococcal vaccine conjugated to CRM197 in United States infants. Pediatrics 101:604-611. 125. Shinefield HR, Black S, Ray P, Chang I, Lewis N, Fireman B, Hackell J, Paradiso PR, Siber G, Kohberger R, Madore DV, Malinowski FJ, Kimura 303 A, Le C, Landaw I, Aguilar J, Hansen J. 1999. Safety and immunogenicity of heptavalent pneumococcal CRM197 conjugate vaccine in infants and toddlers. Pediatr Infect Dis J 18:757-763. 126. World Health Organization. 2012. Pneumococcal vaccines WHO position paper - 2012 - Recommendations. Vaccine 30:4717-4718. 127. Tomczyk S, Bennett NM, Stoecker C, Gierke R, Moore MR, Whitney CG, Hadler S, Pilishvili T. 2014. Use of 13-valent pneumococcal conjugate vaccine and 23-valent pneumococcal polysaccharide vaccine among adults aged ≥65 years: recommendations of the Advisory Committee on Immunization Practices (ACIP). MMWR Morb Mortal Wkly Rep 63:822-825. 128. Clarke E, Kampmann B, Goldblatt D. 2016. Maternal and neonatal pneumococcal vaccination - where are we now? Expert Rev Vaccines 15:1305- 1317. 129. Pilishvili T, Lexau C, Farley MM, Hadler J, Harrison LH, Bennett NM, Reingold A, Thomas A, Schaffner W, Craig AS, Smith PJ, Beall BW, Whitney CG, Moore MR. 2010. Sustained reductions in invasive pneumococcal disease in the era of conjugate vaccine. J Infect Dis 201:32-41. 130. Grijalva CG, Nuorti JP, Arbogast PG, Martin SW, Edwards KM, Griffin MR. 2007. Decline in pneumonia admissions after routine childhood immunisation with pneumococcal conjugate vaccine in the USA: a time-series analysis. Lancet 369:1179-1186. 131. Grijalva CG, Poehling KA, Nuorti JP, Zhu Y, Martin SW, Edwards KM, Griffin MR. 2006. National impact of universal childhood immunization with pneumococcal conjugate vaccine on outpatient medical care visits in the United States. Pediatrics 118:865-873. 132. Whitney CG, Farley MM, Hadler J, Harrison LH, Bennett NM, Lynfield R, Reingold A, Cieslak PR, Pilishvili T, Jackson D, Facklam RR, Jorgensen JH, Schuchat A. 2003. Decline in invasive pneumococcal disease after the 304 introduction of protein–polysaccharide conjugate vaccine. N Engl J Med 348:1737-1746. 133. Miller E, Andrews NJ, Waight PA, Slack MP, George RC. 2011. Herd immunity and serotype replacement 4 years after seven-valent pneumococcal conjugate vaccination in England and Wales: an observational cohort study. Lancet Infect Dis 11:760-768. 134. Rückinger S, van der Linden M, Reinert RR, von Kries R. 2010. Efficacy of 7-valent pneumococcal conjugate vaccination in Germany: an analysis using the indirect cohort method. Vaccine 28:5012-5016. 135. Rodenburg GD, de Greeff SC, Jansen AG, de Melker HE, Schouls LM, Hak E, Spanjaard L, Sanders EA, van der Ende A. 2010. Effects of pneumococcal conjugate vaccine 2 years after its introduction, the Netherlands. Emerg Infect Dis 16:816-823. 136. Harboe ZB, Valentiner-Branth P, Benfield TL, Christensen JJ, Andersen PH, Howitz M, Krogfelt KA, Lambertsen L, Konradsen HB. 2010. Early effectiveness of heptavalent conjugate pneumococcal vaccination on invasive pneumococcal disease after the introduction in the Danish Childhood Immunization Programme. Vaccine 28:2642-2647. 137. Kyaw MH, Lynfield R, Schaffner W, Craig AS, Hadler J, Reingold A, Thomas AR, Harrison LH, Bennett NM, Farley MM, Facklam RR, Jorgensen JH, Besser J, Zell ER, Schuchat A, Whitney CG. 2006. Effect of introduction of the pneumococcal conjugate vaccine on drug-resistant Streptococcus pneumoniae. N Engl J Med 354:1455-1463. 138. Schrag SJ, Beall B, Dowell SF. 2000. Limiting the spread of resistant pneumococci: biological and epidemiologic evidence for the effectiveness of alternative interventions. Clin Microbiol Rev 13:588-601. 139. Kaplan SL, Barson WJ, Lin PL, Stovall SH, Bradley JS, Tan TQ, Hoffman JA, Givner LB, Mason EO, Jr. 2010. Serotype 19A Is the most common 305 serotype causing invasive pneumococcal infections in children. Pediatrics 125:429-436. 140. Balsells E, Guillot L, Nair H, Kyaw MH. 2017. Serotype distribution of Streptococcus pneumoniae causing invasive disease in children in the post-PCV era: a systematic review and meta-analysis. PLoS One 12:e0177113. 141. Feikin DR, Kagucia EW, Loo JD, Link-Gelles R, Puhan MA, Cherian T, Levine OS, Whitney CG, O'Brien KL, Moore MR. 2013. Serotype-specific changes in invasive pneumococcal disease after pneumococcal conjugate vaccine introduction: a pooled analysis of multiple surveillance sites. PLoS Med 10:e1001517. 142. Wenger JD, Zulz T, Bruden D, Singleton R, Bruce MG, Bulkow L, Parks D, Rudolph K, Hurlburt D, Ritter T, Klejka J, Hennessy T. 2010. Invasive pneumococcal disease in alaskan children: impact of the seven-valent pneumococcal conjugate vaccine and the role of water supply. Pediatr Infect Dis J 29:251-256. 143. Nuorti JP, Whitney CG. 2010. Prevention of pneumococcal disease among infants and children - use of 13-valent pneumococcal conjugate vaccine and 23- valent pneumococcal polysaccharide vaccine - recommendations of the Advisory Committee on Immunization Practices (ACIP). MMWR Recomm Rep 59:1-18. 144. Givon-Lavi N, Greenberg D, Dagan R. 2010. Immunogenicity of alternative regimens of the conjugated 7-valent pneumococcal vaccine: a randomized controlled trial. Pediatr Infect Dis J 29:756-762. 145. Pilishvili T, Gierke R, Farley M, Schaffner W, Thomas A, Reingold A, Harrison L, Lynfield R, Zansky S, Petit S, Barnes M, Bareta J, Beall B, Moore M, Whitney C. 2016. Changes in invasive pneumococcal disease (IPD) following 5 years of 13-valent pneumococcal conjugate vaccine use in the U.S. Tenth International Symposium on Pneumococci and Pneumococcal Disease (ISPPD10), Glasgow, Scotland. 306 146. Waight PA, Andrews NJ, Ladhani SN, Sheppard CL, Slack MP, Miller E. 2015. Effect of the 13-valent pneumococcal conjugate vaccine on invasive pneumococcal disease in England and Wales 4 years after its introduction: an observational cohort study. Lancet Infect Dis 15:535-543. 147. Knol MJ, Wagenvoort GHJ, Sanders EAM, Elberse K, Vlaminckx BJ, de Melker HE, van der Ende A. 2015. Invasive pneumococcal disease 3 years after introduction of 10-valent pneumococcal conjugate vaccine, the Netherlands. Emerg Infect Dis 21:2040-2044. 148. Jokinen J, Rinta-Kokko H, Siira L, Palmu AA, Virtanen MJ, Nohynek H, Virolainen-Julkunen A, Toropainen M, Nuorti JP. 2015. Impact of ten-valent pneumococcal conjugate vaccination on invasive pneumococcal disease in Finnish children – a population-based Study. PLoS One 10:e0120290. 149. Slotved HC, Dalby T, Harboe ZB, Valentiner-Branth P, Casadevante VF, Espenhain L, Fuursted K, Konradsen HB. 2016. The incidence of invasive pneumococcal serotype 3 disease in the Danish population is not reduced by PCV-13 vaccination. Heliyon 2:e00198. 150. Alderson MR. 2016. Status of research and development of pediatric vaccines for Streptococcus pneumoniae. Vaccine 34:2959-2961. 151. Song JY, Eun BW, Nahm MH. 2013. Diagnosis of pneumococcal pneumonia: current pitfalls and the way forward. Infect Chemother 45:351-366. 152. O'Brien KL, Bronsdon MA, Dagan R, Yagupsky P, Janco J, Elliott J, Whitney CG, Yang Y-H, Robinson L-GE, Schwartz B, Carlone GM. 2001. Evaluation of a medium (STGG) for transport and optimal recovery of Streptococcus pneumoniae from nasopharyngeal secretions collected during field studies. J Clin Microbiol 39:1021-1024. 153. O'Brien KL, Nohynek H, The WHO pneumococcal vaccine trials carriage working group. 2003. Report from a WHO Working Group: standard method for detecting upper respiratory carriage of Streptococcus pneumoniae. Pediatr Infect Dis J 22:e1-e11. 307 154. Satzke C, Dunne EM, Porter BD, Klugman KP, Mulholland EK, PneuCarriage project group. 2015. The PneuCarriage Project: a multi-centre comparative study to identify the best serotyping methods for examining pneumococcal carriage in vaccine evaluation studies. PLoS Med 12:e1001903. 155. Roca A, Hill PC, Townend J, Egere U, Antonio M, Bojang A, Akisanya A, Litchfield T, Nsekpong DE, Oluwalana C, Howie SR, Greenwood B, Adegbola RA. 2011. Effects of community-wide vaccination with PCV-7 on pneumococcal nasopharyngeal carriage in the Gambia: a cluster-randomized trial. PLoS Med 8:e1001107. 156. Huang SS, Platt R, Rifas-Shiman SL, Pelton SI, Goldmann D, Finkelstein JA. 2005. Post-PCV7 changes in colonizing pneumococcal serotypes in 16 Massachusetts communities, 2001 and 2004. Pediatrics 116:e408-e413. 157. Jin P, Kong F, Xiao M, Oftadeh S, Zhou F, Liu C, Russell F, Gilbert GL. 2009. First report of putative Streptococcus pneumoniae serotype 6D among nasopharyngeal isolates from Fijian children. J Infect Dis 200:1375-1380. 158. Park IH, Pritchard DG, Cartee R, Brandao A, Brandileone MCC, Nahm MH. 2007. Discovery of a new capsular serotype (6C) within serogroup 6 of Streptococcus pneumoniae. J Clin Microbiol 45:1225-1233. 159. Ghaffar F, Barton T, Lozano J, Muniz LS, Hicks P, Gan V, Ahmad N, McCracken GH. 2004. Effect of the 7-valent pneumococcal conjugate vaccine on nasopharyngeal colonization by Streptococcus pneumoniae in the first 2 years of life. Clin Infect Dis 39:930-938. 160. Vestrheim DF, Høiby EA, Aaberge IS, Caugant DA. 2010. Impact of a pneumococcal conjugate vaccination program on carriage among children in Norway. Clin Vaccine Immunol 17:325-334. 161. Veenhoven RH, Bogaert D, Schilder AGM, Rijkers GT, Uiterwaal CSPM, Kiezebrink HH, van Kempen MJP, Dhooge IJ, Bruin J, IJzerman EPF, de Groot R, Kuis W, Hermans PWM, Sanders EAM. 2004. Nasopharyngeal pneumococcal carriage after combined pneumococcal conjugate and 308 polysaccharide vaccination in children with a history of recurrent acute otitis media. Clin Infect Dis 39:911-919. 162. van Gils EJM, Veenhoven RH, Hak E, Rodenburg GD, Bogaert D, IJzerman EPF, Bruin JP, van Alphen L, Sanders EAM. 2009. Effect of reduced-dose schedules with 7-valent pneumococcal conjugate vaccine on nasopharyngeal pneumococcal carriage in children. JAMA 302:159-167. 163. Frazao N, Sa-Leao R, de Lencastre H. 2010. Impact of a single dose of the 7- valent pneumococcal conjugate vaccine on colonization. Vaccine 28:3445-3452. 164. Givon-Lavi N, Fraser D, Dagan R. 2003. Vaccination of day-care center attendees reduces carriage of Streptococcus pneumoniae among their younger siblings. Pediatr Infect Dis J 22:524-531. 165. Millar EV, Watt JP, Bronsdon MA, Dallas J, Reid R, Santosham M, O'Brien KL. 2008. Indirect effect of 7-valent pneumococcal conjugate vaccine on pneumococcal colonization among unvaccinated household members. Clin Infect Dis 47:989-996. 166. Grivea IN, Panagiotou M, Tsantouli AG, Syrogiannopoulos GA. 2008. Impact of heptavalent pneumococcal conjugate vaccine on nasopharyngeal carriage of penicillin-resistant Streptococcus pneumoniae among day-care center attendees in Central Greece. Pediatr Infect Dis J 27:519-525. 167. Hammitt LL, Bruden DL, Butler JC, Baggett HC, Hurlburt DA, Reasonover A, Hennessy TW. 2006. Indirect effect of conjugate vaccine on adult carriage of Streptococcus pneumoniae: an explanation of trends in invasive pneumococcal disease. J Infect Dis 193:1487-1494. 168. Millar EV, O'Brien KL, Watt JP, Bronsdon MA, Dallas J, Whitney CG, Reid R, Santosham M. 2006. Effect of community-wide conjugate pneumococcal vaccine use in infancy on nasopharyngeal carriage through 3 years of age: a cross-sectional study in a high-risk population. Clin Infect Dis 43:8-15. 309 169. Weinberger DM, Malley R, Lipsitch M. 2011. Serotype replacement in disease after pneumococcal vaccination. Lancet 378:1962-1973. 170. Weinberger DM, Harboe ZB, Sanders EAM, Ndiritu M, Klugman KP, Rückinger S, Dagan R, Adegbola R, Cutts F, Johnson HL, O'Brien KL, Scott JA, Lipsitch M. 2010. Association of serotype with risk of death due to pneumococcal pneumonia: a meta-analysis. Clin Infect Dis 51:692-699. 171. Flasche S, Givon-Lavi N, Dagan R. 2016. Using pneumococcal carriage data to monitor postvaccination changes in the incidence of pneumococcal otitis media. Am J Epidemiol 184:652-659. 172. Flasche S, Le Polain de Waroux O, O'Brien KL, Edmunds WJ. 2015. The serotype distribution among healthy carriers before vaccination is essential for predicting the impact of pneumococcal conjugate vaccine on invasive disease. PLoS Comput Biol 11:e1004173. 173. Weinberger DM, Bruden DT, Grant LR, Lipsitch M, O'Brien KL, Pelton SI, Sanders EA, Feikin DR. 2013. Using pneumococcal carriage data to monitor postvaccination changes in invasive disease. Am J Epidemiol 178:1488- 1495. 174. Reiss-Mandel A, Regev-Yochay G. 2016. Staphylococcus aureus and Streptococcus pneumoniae interaction and response to pneumococcal vaccination: myth or reality? Hum Vaccin Immunother 12:351-357. 175. Shiri T, Nunes MC, Adrian PV, Van Niekerk N, Klugman KP, Madhi SA. 2013. Interrelationship of Streptococcus pneumoniae, Haemophilus influenzae and Staphylococcus aureus colonization within and between pneumococcal- vaccine naive mother-child dyads. BMC Infect Dis 13:483. 176. Dunne EM, Manning J, Russell FM, Robins-Browne RM, Mulholland EK, Satzke C. 2012. Effect of pneumococcal vaccination on nasopharyngeal carriage of Streptococcus pneumoniae, Haemophilus influenzae, Moraxella catarrhalis, and Staphylococcus aureus in Fijian children. J Clin Microbiol 50:1034-1038. 310 177. Regev-Yochay G, Dagan R, Raz M, Carmeli Y, Shainberg B, Derazne E, Rahav G, Rubinstein E. 2004. Association between carriage of Streptococcus pneumoniae and Staphylococcus aureus in children. JAMA 292:716-720. 178. Bogaert D, van Belkum A, Sluijter M, Luijendijk A, de Groot R, Rümke HC, Verbrugh HA, Hermans PWM. 2004. Colonisation by Streptococcus pneumoniae and Staphylococcus aureus in healthy children. Lancet 363:1871- 1872. 179. Veenhoven R, Bogaert D, Uiterwaal C, Brouwer C, Kiezebrink H, Bruin J, IJzerman E, Hermans P, de Groot R, Zegers B, Kuis W, Rijkers G, Schilder A, Sanders E. 2003. Effect of conjugate pneumococcal vaccine followed by polysaccharide pneumococcal vaccine on recurrent acute otitis media: a randomised study. Lancet 361:2189-2195. 180. Chien YW, Vidal JE, Grijalva CG, Bozio C, Edwards KM, Williams JV, Griffin MR, Verastegui H, Hartinger SM, Gil AI, Lanata CF, Klugman KP. 2013. Density interactions among Streptococcus pneumoniae, Haemophilus influenzae and Staphylococcus aureus in the nasopharynx of young Peruvian children. Pediatr Infect Dis J 32:72-77. 181. Ebruke C, Dione MM, Walter B, Worwui A, Adegbola RA, Roca A, Antonio M. 2016. High genetic diversity of Staphylococcus aureus strains colonising the nasopharynx of Gambian villagers before widespread use of pneumococcal conjugate vaccines. BMC Microbiol 16:38. 182. Grijalva CG, Griffin MR, Edwards KM, Williams JV, Gil AI, Verastegui H, Hartinger SM, Vidal JE, Klugman KP, Lanata CF. 2014. Cohort profile: the study of respiratory pathogens in Andean children. Int J Epidemiol 43:1021- 1030. 183. Kwambana B, Barer M, Bottomley C, Adegbola R, Antonio M. 2011. Early acquisition and high nasopharyngeal co-colonisation by Streptococcus pneumoniae and three respiratory pathogens amongst Gambian new-borns and infants. BMC Infect Dis 11:175. 311 184. Madhi SA, Adrian P, Kuwanda L, Cutland C, Albrich WC, Klugman KP. 2007. Long-term effect of pneumococcal conjugate vaccine on nasopharyngeal colonization by Streptococcus pneumoniae—and associated interactions with Staphylococcus aureus and Haemophilus influenzae colonization—in HIV- infected and HIV-uninfected children. J Infect Dis 196:1662-1666. 185. Spijkerman J, Prevaes SM, van Gils EJ, Veenhoven RH, Bruin JP, Bogaert D, Wijmenga-Monsuur AJ, van den Dobbelsteen GP, Sanders EA. 2012. Long-term effects of pneumococcal conjugate vaccine on nasopharyngeal carriage of S. pneumoniae, S. aureus, H. influenzae and M. catarrhalis. PLoS One 7:e39730. 186. van Gils EJ, Hak E, Veenhoven RH, Rodenburg GD, Bogaert D, Bruin JP, van Alphen L, Sanders EA. 2011. Effect of seven-valent pneumococcal conjugate vaccine on Staphylococcus aureus colonisation in a randomised controlled trial. PLoS One 6:e20229. 187. Dukers-Muijrers NH, Stobberingh E, Beisser P, Boesten RC, Jacobs P, Hoebe CJ. 2012. Nasal carriage of Streptococcus pneumoniae serotypes and Staphylococcus aureus in Streptococcus pneumoniae-vaccinated and non- vaccinated young children. Epidemiol Infect 141:631-638. 188. Esposito S, Terranova L, Ruggiero L, Ascolese B, Montinaro V, Rios WP, Galeone C, Principi N. 2015. Streptococcus pneumoniae and Staphylococcus aureus carriage in healthy school-age children and adolescents. J Med Microbiol 64:427-431. 189. Hammitt LL, Akech DO, Morpeth SC, Karani A, Kihuha N, Nyongesa S, Bwanaali T, Mumbo E, Kamau T, Sharif SK, Scott JA. 2014. Population effect of 10-valent pneumococcal conjugate vaccine on nasopharyngeal carriage of Streptococcus pneumoniae and non-typeable Haemophilus influenzae in Kilifi, Kenya: findings from cross-sectional carriage studies. Lancet Glob Health 2:e397-405. 312 190. Lee GM, Huang SS, Rifas-Shiman SL, Hinrichsen VL, Pelton SI, Kleinman K, Hanage WP, Lipsitch M, McAdam AJ, Finkelstein JA. 2009. Epidemiology and risk factors for Staphylococcus aureus colonization in children in the post-PCV7 era. BMC Infect Dis 9:110. 191. Park B, Nizet V, Liu GY. 2008. Role of Staphylococcus aureus catalase in niche competition against Streptococcus pneumoniae. J Bacteriol 190:2275- 2278. 192. Margolis E, Yates A, Levin B. 2010. The ecology of nasal colonization of Streptococcus pneumoniae, Haemophilus influenzae and Staphylococcus aureus: the role of competition and interactions with host's immune response. BMC Microbiol 10:59. 193. Regev-Yochay G, Malley R, Rubinstein E, Raz M, Dagan R, Lipsitch M. 2008. In vitro bactericidal activity of Streptococcus pneumoniae and bactericidal susceptibility of Staphylococcus aureus strains isolated from cocolonized versus noncocolonized children. J Clin Microbiol 46:747-749. 194. Melles DC, Bogaert D, Gorkink RF, Peeters JK, Moorhouse MJ, Ott A, van Leeuwen WB, Simons G, Verbrugh HA, Hermans PW, van Belkum A. 2007. Nasopharyngeal co-colonization with Staphylococcus aureus and Streptococcus pneumoniae in children is bacterial genotype independent. Microbiology 153:686-692. 195. Regev-Yochay G, Lipsitch M, Basset A, Rubinstein E, Dagan R, Raz M, Malley R. 2009. The pneumococcal pilus predicts the absence of Staphylococcus aureus co-colonization in pneumococcal carriers. Clin Infect Dis 48:760-763. 196. Basset A, Trzciνski K, Hermos C, O'Brien KL, Reid R, Santosham M, McAdam AJ, Lipsitch M, Malley R. 2007. Association of the pneumococcal pilus with certain capsular serotypes but not with increased virulence. J Clin Microbiol 45:1684-1689. 313 197. Wertheim HFL, Melles DC, Vos MC, van Leeuwen W, van Belkum A, Verbrugh HA, Nouwen JL. 2005. The role of nasal carriage in Staphylococcus aureus infections. Lancet Infect Dis 5:751-762. 198. Williams RE. 1963. Healthy carriage of Staphylococcus aureus: its prevalence and importance. Bacteriol Rev 27:56-71. 199. Jacoby P, Watson K, Bowman J, Taylor A, Riley TV, Smith DW, Lehmann D. 2007. Modelling the co-occurrence of Streptococcus pneumoniae with other bacterial and viral pathogens in the upper respiratory tract. Vaccine 25:2458- 2464. 200. Dahlblom V, Soderstrom M. 2012. Bacterial interactions in the nasopharynx - effects of host factors in children attending day-care centers. J Infect Public Health 5:133-139. 201. Pettigrew MM, Gent JF, Revai K, Patel JA, Chonmaitree T. 2008. Microbial interactions during upper respiratory tract infections. Emerg Infect Dis 14:1584- 1591. 202. Xu Q, Wischmeyer J, Gonzalez E, Pichichero ME. 2017. Nasopharyngeal polymicrobial colonization during health, viral upper respiratory infection and upper respiratory bacterial infection. J Infect 75:26-34. 203. van Gils EJ, Veenhoven RH, Rodenburg GD, Hak E, Sanders EA. 2011. Effect of 7-valent pneumococcal conjugate vaccine on nasopharyngeal carriage with Haemophilus influenzae and Moraxella catarrhalis in a randomized controlled trial. Vaccine 29:7595-7598. 204. Kim CH, Kim JS, Cha SH, Kim KN, Kim JD, Lee KY, Kim HM, Kim JH, Hyuk S, Hong JY, Park SE, Kim YK, Kim NH, Fanic A, Borys D, Ruiz- Guinazu J, Moreira M, Schuerman L, Kim KH. 2011. Response to primary and booster vaccination with 10-valent pneumococcal nontypeable Haemophilus influenzae protein D conjugate vaccine in Korean infants. Pediatr Infect Dis J 30:e235-243. 314 205. Dicko A, Odusanya OO, Diallo AI, Santara G, Barry A, Dolo A, Diallo A, Kuyinu YA, Kehinde OA, Francois N, Borys D, Yarzabal JP, Moreira M, Schuerman L. 2011. Primary vaccination with the 10-valent pneumococcal non-typeable Haemophilus influenzae protein D conjugate vaccine (PHiD-CV) in infants in Mali and Nigeria: a randomized controlled trial. BMC Public Health 11:882. 206. Prymula R, Hanovcova I, Splino M, Kriz P, Motlova J, Lebedova V, Lommel P, Kaliskova E, Pascal T, Borys D, Schuerman L. 2011. Impact of the 10-valent pneumococcal non-typeable Haemophilus influenzae Protein D conjugate vaccine (PHiD-CV) on bacterial nasopharyngeal carriage. Vaccine 29:1959-1967. 207. Leach AJ, Wigger C, Hare K, Hampton V, Beissbarth J, Andrews R, Chatfield M, Smith-Vaughan H, Morris PS. 2015. Reduced middle ear infection with non-typeable Haemophilus influenzae, but not Streptococcus pneumoniae, after transition to 10-valent pneumococcal non-typeable H. influenzae protein D conjugate vaccine. BMC Pediatr 15:162. 208. van den Bergh MR, Spijkerman J, Swinnen KM, Francois NA, Pascal TG, Borys D, Schuerman L, Ijzerman EP, Bruin JP, van der Ende A, Veenhoven RH, Sanders EA. 2013. Effects of the 10-valent pneumococcal nontypeable Haemophilus influenzae protein D-conjugate vaccine on nasopharyngeal bacterial colonization in young children: a randomized controlled trial. Clin Infect Dis 56:e30-39. 209. Vesikari T, Forsten A, Seppa I, Kaijalainen T, Puumalainen T, Soininen A, Traskine M, Lommel P, Schoonbroodt S, Hezareh M, Moreira M, Borys D, Schuerman L. 2016. Effectiveness of the 10-valent pneumococcal nontypeable Haemophilus influenzae protein D-conjugated vaccine (PHiD-CV) against carriage and acute otitis media - a double-blind randomized clinical trial in Finland. J Pediatric Infect Dis Soc 5:237-248. 210. Andrade DC, Borges IC, Bouzas ML, Oliveira JR, Fukutani KF, Queiroz AT, de Oliveira CI, Barral A, Van Weyenbergh J, Nascimento-Carvalho C. 315 2017. 10-valent pneumococcal conjugate vaccine (PCV10) decreases metabolic activity but not nasopharyngeal carriage of Streptococcus pneumoniae and Haemophilus influenzae. Vaccine 35:4105-4111. 211. Biesbroek G, Wang X, Keijser BJ, Eijkemans RM, Trzcinski K, Rots NY, Veenhoven RH, Sanders EA, Bogaert D. 2014. Seven-valent pneumococcal conjugate vaccine and nasopharyngeal microbiota in healthy children. Emerg Infect Dis 20:201-210. 212. Hilty M, Qi W, Brugger SD, Frei L, Agyeman P, Frey PM, Aebi S, Muhlemann K. 2012. Nasopharyngeal microbiota in infants with acute otitis media. J Infect Dis 205:1048-1055. 213. Feazel LM, Santorico SA, Robertson CE, Bashraheil M, Scott JA, Frank DN, Hammitt LL. 2015. Effects of vaccination with 10-valent pneumococcal non-typeable Haemophilus influenza protein D conjugate vaccine (PHiD-CV) on the nasopharyngeal microbiome of Kenyan toddlers. PLoS One 10:e0128064. 214. Nair M. 2012. Protein conjugate polysaccharide vaccines: challenges in development and global implementation. Indian J Community Med 37:79-82. 215. Lee BY, Assi TM, Rookkapan K, Wateska AR, Rajgopal J, Sornsrivichai V, Chen SI, Brown ST, Welling J, Norman BA, Connor DL, Bailey RR, Jana A, Van Panhuis WG, Burke DS. 2011. Maintaining vaccine delivery following the introduction of the rotavirus and pneumococcal vaccines in Thailand. PLoS One 6:e24673. 216. Tricarico S, McNeil HC, Cleary DW, Head MG, Lim V, Yap IKS, Wie CC, Tan CS, Norazmi MN, Aziah I, Cheah ESG, Faust SN, Jefferies JMC, Roderick PJ, Moore M, Yuen HM, Newell ML, McGrath N, Doncaster CP, Kraaijeveld AR, Webb JS, Clarke SC. 2017. Pneumococcal conjugate vaccine implementation in middle-income countries. Pneumonia 9:6. 217. Hausdorff WP, Dagan R, Beckers F, Schuerman L. 2009. Estimating the direct impact of new conjugate vaccines against invasive pneumococcal disease. Vaccine 27:7257-7269. 316 218. By A, Sobocki P, Forsgren A, Silfverdal SA. 2012. Comparing health outcomes and costs of general vaccination with pneumococcal conjugate vaccines in Sweden: a Markov model. Clin Ther 34:177-189. 219. Newall AT, Creighton P, Philp DJ, Wood JG, MacIntyre CR. 2011. The potential cost-effectiveness of infant pneumococcal vaccines in Australia. Vaccine 29:8077-8085. 220. International Vaccine Access Center (IVAC). March, 2016. State of PCV use and impact evaluations. Johns Hopkins Bloomberg School of Public Health, 221. World Health Organization. 2015. Sustainable access to vaccines in middle- income countries (MICs): a shared partner strategy. Report of the WHO- Convened MIC Task Force, 222. Temple B, Griffiths UK, Mulholland EK, Ratu FT, Tikoduadua L, Russell FM. 2011. The cost of outpatient pneumonia in children <5 years of age in Fiji. Trop Med Int Health 17:197–203. 223. Magree HC, Russell FM, Sa'aga R, Greenwood P, Tikoduadua L, Pryor J, Waqatakirewa L, Carapetis JR, Mulholland EK. 2005. Chest X-ray- confirmed pneumonia in children in Fiji. Bull World Health Organ 83:427-433. 224. Biaukula VL, Tikoduadua L, Azzopardi K, Seduadua A, Temple B, Richmond P, Robins-Browne R, Mulholland EK, Russell FM. 2012. Meningitis in children in Fiji: etiology, epidemiology, and neurological sequelae. Int J Infect Dis 16:e289-e295. 225. Brian G, Ramke J, Maher L, Page A, Szetu J. 2010. The prevalence of diabetes among adults aged 40 years and over in Fiji. N Z Med J 123:68-75. 226. Kuehn R, Fong J, Taylor R, Gyaneshwar R, Carter K. 2012. Cervical cancer incidence and mortality in Fiji 2003–2009. Aust N Z J Obstet Gynaecol 52:380- 386. 317 227. Steer AC, Kado J, Jenney AW, Batzloff M, Waqatakirewa L, Mulholland EK, Carapetis JR. 2009. Acute rheumatic fever and rheumatic heart disease in Fiji: prospective surveillance, 2005-2007. Med J Aust 190:133-135. 228. Russell FM, Carapetis JR, Ketaiwai S, Kunabuli V, Taoi M, Biribo S, Seduadua A, Mulholland EK. 2006. Pneumococcal nasopharyngeal carriage and patterns of penicillin resistance in young children in Fiji. Ann Trop Paediatr Int Child Health 26:187-197. 229. Russell FM, Carapetis JR, Satzke C, Tikoduadua L, Waqatakirewa L, Chandra R, Seduadua A, Oftadeh S, Cheung YB, Gilbert GL, Mulholland EK. 2010. Pneumococcal nasopharyngeal carriage following reduced doses of a 7-valent pneumococcal conjugate vaccine and a 23-valent pneumococcal polysaccharide vaccine booster. Clin Vaccine Immunol 17:1970-1976. 230. Russell FM, Licciardi PV, Balloch A, Biaukula V, Tikoduadua L, Carapetis JR, Nelson J, Jenney AWJ, Waqatakirewa L, Colquhoun S, Cheung YB, Tang MLK, Mulholland EK. 2010. Safety and immunogenicity of the 23- valent pneumococcal polysaccharide vaccine at 12 months of age, following one, two, or three doses of the 7-valent pneumococcal conjugate vaccine in infancy. Vaccine 28:3086-3094. 231. Boelsen LK, Dunne EM, Lamb KE, Bright K, Cheung YB, Tikoduadua L, Russell FM, Mulholland EK, Licciardi PV, Satzke C. 2015. Long-term impact of pneumococcal polysaccharide vaccination on nasopharyngeal carriage in children previously vaccinated with various pneumococcal conjugate vaccine regimes. Vaccine 33:5708-5714. 232. Castillo D, Harcourt B, Hatcher C, Jackson M, Katz L, Mair R, Mayer L, Novak R, Rahalison L, Schmink S, Theodore MJ, Thomas J, Vuong J, Wang X, McGee L. 2011. Laboratory methods for the diagnosis of meningitis caused by Neisseria meningitidis, Streptococcus pneumoniae, and Haemophilus influenzae, 2nd Edition. World Health Organization. 318 233. Ortika BD, Habib M, Dunne EM, Porter BD, Satzke C. 2013. Production of latex agglutination reagents for pneumococcal serotyping. BMC Res Notes 6:49. 234. Porter BD, Ortika BD, Satzke C. 2014. Capsular serotyping of Streptococcus pneumoniae by latex agglutination. J Vis Exp 91:e51747. 235. Turner P, Turner C, Jankhot A, Phakaudom K, Nosten F, Goldblatt D. 2013. Field evaluation of culture plus latex sweep serotyping for detection of multiple pneumococcal serotype colonisation in infants and young children. PLoS One 8:e67933. 236. Habib M, Porter BD, Satzke C. 2014. Capsular serotyping of Streptococcus pneumoniae using the quellung reaction. J Vis Exp 84:e51208. 237. Carvalho MdGS, Tondella ML, McCaustland K, Weidlich L, McGee L, Mayer LW, Steigerwalt A, Whaley M, Facklam RR, Fields B, Carlone G, Ades EW, Dagan R, Sampson JS. 2007. Evaluation and improvement of real- time PCR assays targeting lytA, ply, and psaA genes for detection of pneumococcal DNA. J Clin Microbiol 45:2460-2466. 238. Dias CA, Teixeira LM, Carvalho MdGS, Beall B. 2007. Sequential multiplex PCR for determining capsular serotypes of pneumococci recovered from Brazilian children. J Med Microbiol 56:1185-1188. 239. Dunne EM, Ong EK, Moser RJ, Siba PM, Phuanukoonnon S, Greenhill AR, Robins-Browne RM, Mulholland EK, Satzke C. 2011. Multilocus sequence typing of Streptococcus pneumoniae by use of mass spectrometry. J Clin Microbiol 49:3756-3760. 240. Jolley K, Maiden M. 2010. BIGSdb: Scalable analysis of bacterial genome variation at the population level. BMC Bioinformatics 11:595. 241. Schloss PD, Westcott SL, Ryabin T, Hall JR, Hartmann M, Hollister EB, Lesniewski RA, Oakley BB, Parks DH, Robinson CJ, Sahl JW, Stres B, Thallinger GG, Van Horn DJ, Weber CF. 2009. Introducing mothur: open- 319 source, platform-independent, community-supported software for describing and comparing microbial communities. Appl Environ Microbiol 75:7537-7541. 242. Quast C, Pruesse E, Yilmaz P, Gerken J, Schweer T, Yarza P, Peplies J, Glöckner FO. 2013. The SILVA ribosomal RNA gene database project: improved data processing and web-based tools. Nucleic Acids Res 41:D590- D596. 243. Edgar RC, Haas BJ, Clemente JC, Quince C, Knight R. 2011. UCHIME improves sensitivity and speed of chimera detection. Bioinformatics 27:2194- 2200. 244. Bankevich A, Nurk S, Antipov D, Gurevich AA, Dvorkin M, Kulikov AS, Lesin VM, Nikolenko SI, Pham S, Prjibelski AD, Pyshkin AV, Sirotkin AV, Vyahhi N, Tesler G, Alekseyev MA, Pevzner PA. 2012. SPAdes: a new genome assembly algorithm and its applications to single-cell sequencing. J Comput Biol 19:455-477. 245. Aziz RK, Bartels D, Best AA, DeJongh M, Disz T, Edwards RA, Formsma K, Gerdes S, Glass EM, Kubal M, Meyer F, Olsen GJ, Olson R, Osterman AL, Overbeek RA, McNeil LK, Paarmann D, Paczian T, Parrello B, Pusch GD, Reich C, Stevens R, Vassieva O, Vonstein V, Wilke A, Zagnitko O. 2008. The RAST Server: rapid annotations using subsystems technology. BMC Genomics 9:75. 246. Segata N, Waldron L, Ballarini A, Narasimhan V, Jousson O, Huttenhower C. 2012. Metagenomic microbial community profiling using unique clade- specific marker genes. Nat Meth 9:811-814. 247. Wood DE, Salzberg SL. 2014. Kraken: ultrafast metagenomic sequence classification using exact alignments. Genome Biol 15:R46. 248. Pruesse E, Peplies J, Glöckner FO. 2012. SINA: Accurate high-throughput multiple sequence alignment of ribosomal RNA genes. Bioinformatics 28:1823- 1829. 320 249. DeSantis TZ, Hugenholtz P, Larsen N, Rojas M, Brodie EL, Keller K, Huber T, Dalevi D, Hu P, Andersen GL. 2006. Greengenes, a chimera- checked 16S rRNA gene database and workbench compatible with ARB. Appl Environ Microbiol 72:5069-5072. 250. Johnson M, Zaretskaya I, Raytselis Y, Merezhuk Y, McGinnis S, Madden TL. 2008. NCBI BLAST: a better web interface. Nucleic Acids Res 36:W5-9. 251. Camacho C, Coulouris G, Avagyan V, Ma N, Papadopoulos J, Bealer K, Madden TL. 2009. BLAST+: architecture and applications. BMC Bioinformatics 10:421. 252. Sullivan MJ, Petty NK, Beatson SA. 2011. Easyfig: a genome comparison visualizer. Bioinformatics 27:1009-1010. 253. Oksanen J, Blanchet FG, Kindt R, Legendre P, Minchin PR, O'Hara RB, Simpson GL, Solymos P, Stevens MHH, Wagner H. 2016. vegan: community ecology package, R package version 2.3-5. ( https://CRAN.R- project.org/package=vegan). 254. Warnes GR, Bolker B, Bonebakker L, Gentleman R, Liaw WHA, Lumley T, Maechler M, Magnusson A, Moeller S, Schwartz M, Venables B. 2016. gplots: various R programming tools for plotting data, R package version 3.0.1. (https://cran.r-project.org/web/packages/gplots/index.html). 255. Caliński T, Harabasz J. 1974. A dendrite method for cluster analysis. Commun Stat 3:1-27. 256. Friedman J, Alm EJ. 2012. Inferring correlation networks from genomic survey data. PLoS Comput Biol 8:e1002687. 257. Shannon P, Markiel A, Ozier O, Baliga NS, Wang JT, Ramage D. 2003. Cytoscape: a software environment for integrated models of biomolecular interaction networks. Genome Res 13:2498–2504. 321 258. Benjamini Y, Hochberg Y. 1995. Controlling the false discovery rate: a practical and powerful approach to multiple testing. J R Stat Soc Series B Stat Methodol 57:289-300. 259. Russell FM, Balloch A, Tang MLK, Carapetis JR, Licciardi P, Nelson J, Jenney AWJ, Tikoduadua L, Waqatakirewa L, Pryor J, Byrnes GB, Cheung YB, Mulholland EK. 2009. Immunogenicity following one, two, or three doses of the 7-valent pneumococcal conjugate vaccine. Vaccine 27:5685- 5691. 260. Russell FM, Carapetis JR, Balloch A, Licciardi PV, Jenney AWJ, Tikoduadua L, Waqatakirewa L, Pryor J, Nelson J, Byrnes GB, Cheung YB, Tang MLK, Mulholland EK. 2010. Hyporesponsiveness to re-challenge dose following pneumococcal polysaccharide vaccine at 12 months of age, a randomized controlled trial. Vaccine 28:3341-3349. 261. Russell FM, Carapetis JR, Burton RL, Lin J, Licciardi PV, Balloch A, Tikoduadua L, Waqatakirewa L, Cheung YB, Tang ML, Nahm MH, Mulholland EK. 2011. Opsonophagocytic activity following a reduced dose 7- valent pneumococcal conjugate vaccine infant primary series and 23-valent pneumococcal polysaccharide vaccine at 12 months of age. Vaccine 29:535-544. 262. Wang X, Mair R, Hatcher C, Theodore MJ, Edmond K, Wu HM, Harcourt BH, Carvalho Mda G, Pimenta F, Nymadawa P, Altantsetseg D, Kirsch M, Satola SW, Cohn A, Messonnier NE, Mayer LW. 2011. Detection of bacterial pathogens in Mongolia meningitis surveillance with a new real-time PCR assay to detect Haemophilus influenzae. Int J Med Microbiol 301:303-309. 263. Gold R, Lepow ML, Goldschneider I, Draper TL, Gotschlich EC. 1975. Clinical evaluation of group A and group C meningococcal polysaccharide vaccines in infants. J Clin Invest 56:1536-1547. 264. Leach A, Twumasi PA, Kumah S, Banya WS, Jaffar S, Forrest BD, Granoff DM, LiButti DE, Carlone GM, Pais LB, Broome CV, Greenwood BM. 1997. Induction of immunologic memory in Gambian children by vaccination in 322 infancy with a Group A plus Group C Meningococcal polysaccharide-protein conjugate vaccine. J Infect Dis 175:200-204. 265. Licciardi PV, Toh ZQ, Clutterbuck EA, Balloch A, Marimla RA, Tikkanen L, Lamb KE, Bright KJ, Rabuatoka U, Tikoduadua L, Boelsen LK, Dunne EM, Satzke C, Cheung YB, Pollard AJ, Russell FM, Mulholland EK. 2016. No long-term evidence of hyporesponsiveness after use of pneumococcal conjugate vaccine in children previously immunized with pneumococcal polysaccharide vaccine. J Allergy Clin Immunol 137:1772-1779.e1711. 266. O'Grady KA, Lee KJ, Carlin JB, Torzillo PJ, Chang AB, Mulholland EK, Lambert SB, Andrews RM. 2010. Increased risk of hospitalization for acute lower respiratory tract infection among Australian indigenous infants 5-23 months of age following pneumococcal vaccination: a cohort study. Clin Infect Dis 50:970-978. 267. Moore MR, Whitney CG. 2008. Emergence of nonvaccine serotypes following introduction of pneumococcal conjugate vaccine: cause and effect? Clin Infect Dis 46:183-185. 268. O'Brien KL, Hochman M, Goldblatt D. 2007. Combined schedules of pneumococcal conjugate and polysaccharide vaccines: is hyporesponsiveness an issue? Lancet Infect Dis 7:597-606. 269. Clutterbuck EA, Lazarus R, Yu LM, Bowman J, Bateman EA, Diggle L, Angus B, Peto TE, Beverley PC, Mant D, Pollard AJ. 2012. Pneumococcal conjugate and plain polysaccharide vaccines have divergent effects on antigen- specific B cells. J Infect Dis 205:1408-1416. 270. Knuf M, Pankow-Culot H, Grunert D, Rapp M, Panzer F, Kollges R, Fanic A, Habib A, Borys D, Dieussaert I, Schuerman L. 2012. Induction of immunologic memory following primary vaccination with the 10-valent pneumococcal nontypeable Haemophilus influenzae protein D conjugate vaccine in infants. Pediatr Infect Dis J 31:e31-36. 323 271. Dagan R, Givon-Lavi N, Greenberg D, Fritzell B, Siegrist C-A. 2010. Nasopharyngeal carriage of Streptococcus pneumoniae shortly before vaccination with a pneumococcal conjugate vaccine causes serotype-specific hyporesponsiveness in early infancy. J Infect Dis 201:1570-1579. 272. Madhi SA, Violari A, Klugman KP, Lin G, McIntyre JA, von Gottberg A, Jean-Philippe P, Cotton MF, Adrian P. 2011. Inferior quantitative and qualitative immune responses to pneumococcal conjugate vaccine in infants with nasopharyngeal colonization by Streptococcus pneumoniae during primary series of immunization. Vaccine 29:6994-7001. 273. Licciardi PV, Russell FM, Balloch A, Burton RL, Nahm MH, Gilbert G, Tang ML, Mulholland EK. 2014. Impaired serotype-specific immune function following pneumococcal vaccination in infants with prior carriage. Vaccine 32:2321-2327. 274. Lynch SV, Pedersen O. 2016. The human intestinal microbiome in health and disease. N Engl J Med 375:2369-2379. 275. van Nood E, Vrieze A, Nieuwdorp M, Fuentes S, Zoetendal EG, de Vos WM, Visser CE, Kuijper EJ, Bartelsman JFWM, Tijssen JGP, Speelman P, Dijkgraaf MGW, Keller JJ. 2013. Duodenal infusion of donor feces for recurrent Clostridium difficile. N Engl J Med 368:407-415. 276. Bogaert D, Keijser B, Huse S, Rossen J, Veenhoven R, van Gils E, Bruin J, Montijn R, Bonten M, Sanders E. 2011. Variability and diversity of nasopharyngeal microbiota in children: a metagenomic analysis. PLoS One 6:e17035. 277. Charlson ES, Chen J, Custers-Allen R, Bittinger K, Li H, Sinha R, Hwang J, Bushman FD, Collman RG. 2010. Disordered microbial communities in the upper respiratory tract of cigarette smokers. PLoS One 5:e15216. 278. Kirsten MK, Vishal KS, David N, Sarath CJ, Stephanie DD. 2015. Newborn upper airway microbiome differs from maternal airway microbiome. American Thoracic Society 2015 International Conference, Denver, United States. 324 279. Mika M, Mack I, Korten I, Qi W, Aebi S, Frey U, Latzin P, Hilty M. 2015. Dynamics of the nasal microbiota in infancy: a prospective cohort study. J Allergy Clin Immunol 135:905-912.e911. 280. Biesbroek G, Bosch AA, Wang X, Keijser BJ, Veenhoven RH, Sanders EA, Bogaert D. 2014. The impact of breastfeeding on nasopharyngeal microbial communities in infants. Am J Respir Crit Care Med 190:298-308. 281. Madan JC, Koestler DC, Stanton BA, Davidson L, Moulton LA, Housman ML, Moore JH, Guill MF, Morrison HG, Sogin ML, Hampton TH, Karagas MR, Palumbo PE, Foster JA, Hibberd PL, O’Toole GA. 2012. Serial analysis of the gut and respiratory microbiome in cystic fibrosis in infancy: interaction between intestinal and respiratory tracts and impact of nutritional exposures. mBio 3:e00251-12. 282. Rosas-Salazar C, Shilts MH, Tovchigrechko A, Chappell JD, Larkin EK, Nelson KE, Moore ML, Anderson LJ, Das SR, Hartert TV. 2016. Nasopharyngeal microbiome in respiratory syncytial virus resembles profile associated with increased childhood asthma risk. Am J Respir Crit Care Med 193:1180-1183. 283. Sakwinska O, Bastic Schmid V, Berger B, Bruttin A, Keitel K, Lepage M, Moine D, Ngom Bru C, Brussow H, Gervaix A. 2014. Nasopharyngeal microbiota in healthy children and pneumonia patients. J Clin Microbiol 52:1590-1594. 284. Teo SM, Mok D, Pham K, Kusel M, Serralha M, Troy N, Holt BJ, Hales BJ, Walker ML, Hollams E, Bochkov YA, Grindle K, Johnston SL, Gern JE, Sly PD, Holt PG, Holt KE, Inouye M. 2015. The infant nasopharyngeal microbiome impacts severity of lower respiratory infection and risk of asthma development. Cell Host Microbe 17:704–715. 285. Prevaes SM, de Winter-de Groot KM, Janssens HM, de Steenhuijsen Piters WA, Tramper-Stranders GA, Wyllie AL, Hasrat R, Tiddens HA, van Westreenen M, van der Ent CK, Sanders EA, Bogaert D. 2016. Development 325 of the nasopharyngeal microbiota in infants with cystic fibrosis. Am J Respir Crit Care Med 193:504-515. 286. Allen EK, Koeppel AF, Hendley JO, Turner SD, Winther B, Sale MM. 2014. Characterization of the nasopharyngeal microbiota in health and during rhinovirus challenge. Microbiome 2:22. 287. Rodrigues F, Danon L, Morales-Aza B, Sikora P, Thors V, Ferreira M, Gould K, Hinds J, Finn A. 2016. Pneumococcal serotypes colonise the nasopharynx in children at different densities. PLoS One 11:e0163435. 288. Bomar L, Brugger SD, Yost BH, Davies SS, Lemon KP. 2015. Corynebacterium accolens releases antipneumococcal free fatty acids from human nostril and skin surface triacylglycerols. mBio 7:e01725-15. 289. Laufer AS, Metlay JP, Gent JF, Fennie KP, Kong Y, Pettigrew MM. 2011. Microbial communities of the upper respiratory tract and otitis media in children. mBio 2:e00245-10. 290. Weiss S, Van Treuren W, Lozupone C, Faust K, Friedman J, Deng Y, Xia LC, Xu ZZ, Ursell L, Alm EJ, Birmingham A, Cram JA, Fuhrman JA, Raes J, Sun F, Zhou J, Knight R. 2016. Correlation detection strategies in microbial data sets vary widely in sensitivity and precision. ISME J 10:1669. 291. Faust K, Sathirapongsasuti JF, Izard J, Segata N, Gevers D, Raes J, Huttenhower C. 2012. Microbial co-occurrence relationships in the human microbiome. PLoS Comput Biol 8:e1002606. 292. Ruan Q, Dutta D, Schwalbach MS, Steele JA, Fuhrman JA, Sun F. 2006. Local similarity analysis reveals unique associations among marine bacterioplankton species and environmental factors. Bioinformatics 22:2532-2538. 293. Deng Y, Jiang YH, Yang Y, He Z, Luo F, Zhou J. 2012. Molecular ecological network analyses. BMC Bioinformatics 13:113. 326 294. Biesbroek G, Tsivtsivadze E, Sanders EA, Montijn R, Veenhoven RH, Keijser BJ, Bogaert D. 2014. Early respiratory microbiota composition determines bacterial succession patterns and respiratory health in children. Am J Respir Crit Care Med 190:1283-1292. 295. Kwambana B, Mohammed N, Jeffries D, Barer M, Adegbola R, Antonio M. 2011. Differential effects of frozen storage on the molecular detection of bacterial taxa that inhabit the nasopharynx. BMC Clin Pathol 11:2. 296. Ling Z, Liu X, Luo Y, Yuan L, Nelson KE, Wang Y, Xiang C, Li L. 2013. Pyrosequencing analysis of the human microbiota of healthy Chinese undergraduates. BMC Genomics 14:390. 297. Kumar PS, Matthews CR, Joshi V, de Jager M, Aspiras M. 2011. Tobacco smoking affects bacterial acquisition and colonization in oral biofilms. Infect Immun 79:4730-4738. 298. Biedermann L, Zeitz J, Mwinyi J, Sutter-Minder E, Rehman A, Ott SJ, Steurer-Stey C, Frei A, Frei P, Scharl M, Loessner MJ, Vavricka SR, Fried M, Schreiber S, Schuppler M, Rogler G. 2013. Smoking cessation induces profound changes in the composition of the intestinal microbiota in humans. PLoS One 8:e59260. 299. Hasegawa K, Linnemann RW, Mansbach JM, Ajami NJ, Espinola JA, Fiechtner LG, Petrosino JF, Camargo CA. 2016. Association of household siblings with nasal and fecal microbiota in infants. Pediatr Int 59:473–481. 300. Liu CM, Price LB, Hungate BA, Abraham AG, Larsen LA, Christensen K, Stegger M, Skov R, Andersen PS. 2015. Staphylococcus aureus and the ecology of the nasal microbiome. Sci Adv 1:e1400216. 301. Jones EA, Kananurak A, Bevins CL, Hollox EJ, Bakaletz LO. 2014. Copy number variation of the beta defensin gene cluster on chromosome 8p influences the bacterial microbiota within the nasopharynx of otitis-prone children. PLoS One 9:e98269. 327 302. Hooper LV, Littman DR, Macpherson AJ. 2012. Interactions between the microbiota and the immune system. Science 336:1268-1273. 303. Brestoff JR, Artis D. 2013. Commensal bacteria at the interface of host metabolism and the immune system. Nat Immunol 14:676-684. 304. van Well GTJ, Sanders MS, Ouburg S, Kumar V, van Furth AM, Morré SA. 2013. Single nucleotide polymorphisms in pathogen recognition receptor genes are associated with susceptibility to meningococcal meningitis in a pediatric cohort. PLoS One 8:e64252. 305. Wang C, Chen Z-L, Pan Z-F, Wei L-L, Xu D-D, Jiang T-T, Zhang X, Ping Z-P, Li Z-J, Li J-C. 2014. NOD2 polymorphisms and pulmonary tuberculosis susceptibility: a systematic review and meta-analysis. Int J Biol Sci 10:103-108. 306. Ioana M, Ferwerda B, Plantinga TS, Stappers M, Oosting M, McCall M, Cimpoeru A, Burada F, Panduru N, Sauerwein R, Doumbo O, van der Meer JWM, van Crevel R, Joosten LAB, Netea MG. 2012. Different patterns of Toll-like receptor 2 polymorphisms in populations of various ethnic and geographic origins. Infect Immun 80:1917-1922. 307. Blekhman R, Goodrich JK, Huang K, Sun Q, Bukowski R, Bell JT, Spector TD, Keinan A, Ley RE, Gevers D, Clark AG. 2015. Host genetic variation impacts microbiome composition across human body sites. Genome Biol 16:191. 308. Fiji Meteorological Service. 28 April 2006. The Climate of Fiji, Fiji Meteorological Service, Information Sheet No. 35. (http://www.met.gov.fj/ClimateofFiji.pdf). Accessed 2017. 309. de Steenhuijsen Piters WA, Heinonen S, Hasrat R, Bunsow E, Smith B, Suarez-Arrabal MC, Chaussabel D, Cohen DM, Sanders EA, Ramilo O, Bogaert D, Mejias A. 2016. Nasopharyngeal microbiota, host transcriptome and disease severity in children with respiratory syncytial virus infection. Am J Respir Crit Care Med 194:1104–1115. 328 310. Schillinger U, Endo A. 2014. Minor genera of the Carnobacteriaceae: Allofustis, Alloiococcus, Atopobacter, Atopococcus, Atopostipes, Bavariicoccus, Desemzia, Dolosigranulum, Granulicatella, Isobaculum and Lacticigenium, p 159-170, Lactic Acid Bacteria, Chapter 14. John Wiley & Sons, Ltd. 311. Altschul SF, Madden TL, Schaffer AA, Zhang J, Zhang Z, Miller W, Lipman DJ. 1997. Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Res 25:3389-3402. 312. LaClaire LL, Facklam RR. 2000. Comparison of three commercial rapid identification systems for the unusual gram-positive cocci Dolosigranulum pigrum, Ignavigranum ruoffiae, and Facklamia species. J Clin Microbiol 38:2037-2042. 313. Miller PH, Facklam RR, Miller JM. 1996. Atmospheric growth requirements for Alloiococcus species and related gram-positive cocci. J Clin Microbiol 34:1027-1028. 314. LaClaire L, Facklam R. 2000. Antimicrobial susceptibility and clinical sources of Dolosigranulum pigrum cultures. Antimicrob Agents Chemother 44:2001- 2003. 315. Lécuyer H, Audibert J, Bobigny A, Eckert C, Jannière-Nartey C, Buu-Hoï A, Mainardi J-L, Podglajen I. 2007. Dolosigranulum pigrum causing nosocomial pneumonia and septicemia. J Clin Microbiol 45:3474-3475. 316. Faden H. 1998. Monthly prevalence of group A, B and G Streptococcus, Haemophilus influenzae types E and F and Pseudomonas aeruginosa nasopharyngeal colonization in the first two years of life. Pediatr Infect Dis J 17:255-256. 317. Coughtrie AL, Whittaker RN, Begum N, Anderson R, Tuck A, Faust SN, Jefferies JM, Yuen HM, Roderick PJ, Mullee MA, Moore MV, Clarke SC. 2014. Evaluation of swabbing methods for estimating the prevalence of bacterial carriage in the upper respiratory tract: a cross sectional study. BMJ Open 4:e005341. 329 318. Klepac-Ceraj V, Lemon KP, Martin TR, Allgaier M, Kembel SW, Knapp AA, Lory S, Brodie EL, Lynch SV, Bohannan BJM, Green JL, Maurer BA, Kolter R. 2010. Relationship between cystic fibrosis respiratory tract bacterial communities and age, genotype, antibiotics and Pseudomonas aeruginosa. Environ Microbiol 12:1293-1303. 319. Fothergill JL, Neill DR, Loman N, Winstanley C, Kadioglu A. 2014. Pseudomonas aeruginosa adaptation in the nasopharyngeal reservoir leads to migration and persistence in the lungs. Nat Commun 5:4780. 320. Driscoll JA, Brody SL, Kollef MH. 2007. The epidemiology, pathogenesis and treatment of Pseudomonas aeruginosa infections. Drugs 67:351-368. 321. Stapleton F, Keay LJ, Sanfilippo PG, Katiyar S, Edwards KP, Naduvilath T. 2007. Relationship between climate, disease severity, and causative organism for contact lens-associated microbial keratitis in Australia. Am J Ophthalmol 144:690-698. 322. Psoter KJ, De Roos AJ, Wakefield J, Mayer J, Rosenfeld M. 2013. Season is associated with Pseudomonas aeruginosa acquisition in young children with cystic fibrosis. Clin Microbiol Infect 19:E483-E489. 323. Salter SJ, Cox MJ, Turek EM, Calus ST, Cookson WO, Moffatt MF, Turner P, Parkhill J, Loman NJ, Walker AW. 2014. Reagent and laboratory contamination can critically impact sequence-based microbiome analyses. BMC Biol 12:87. 324. Jervis-Bardy J, Leong LEX, Marri S, Smith RJ, Choo JM, Smith-Vaughan HC, Nosworthy E, Morris PS, O’Leary S, Rogers GB, Marsh RL. 2015. Deriving accurate microbiota profiles from human samples with low bacterial content through post-sequencing processing of Illumina MiSeq data. Microbiome 3. 325. Kulakov LA, McAlister MB, Ogden KL, Larkin MJ, O'Hanlon JF. 2002. Analysis of bacteria contaminating ultrapure water in industrial systems. Appl Environ Microbiol 68:1548-1555. 330 326. McFeters GA, Broadaway SC, Pyle BH, Egozy Y. 1993. Distribution of bacteria within operating laboratory water purification systems. Appl Environ Microbiol 59:1410-1415. 327. Mohammadi T, Reesink HW, Vandenbroucke-Grauls CMJE, Savelkoul PHM. 2005. Removal of contaminating DNA from commercial nucleic acid extraction kit reagents. J Microbiol Methods 61:285-288. 328. Wen Y, Xiao F, Wang C, Wang Z. 2016. The impact of different methods of DNA extraction on microbial community measures of BALF samples based on metagenomic data. Am J Transl Res 8:1412-1425. 329. Rand KH, Houck H. 1990. Taq polymerase contains bacterial DNA of unknown origin. Mol Cell Probes 4:445-450. 330. Corless CE, Guiver M, Borrow R, Edwards-Jones V, Kaczmarski EB, Fox AJ. 2000. Contamination and sensitivity issues with a real-time universal 16S rRNA PCR. J Clin Microbiol 38:1747-1752. 331. Dorrington MG, Bowdish DM. 2013. Immunosenescence and novel vaccination strategies for the elderly. Front Immunol 4:171. 332. Franceschi C, Capri M, Monti D, Giunta S, Olivieri F, Sevini F, Panourgia MP, Invidia L, Celani L, Scurti M, Cevenini E, Castellani GC, Salvioli S. 2007. Inflammaging and anti-inflammaging: a systemic perspective on aging and longevity emerged from studies in humans. Mech Ageing Dev 128:92-105. 333. Goncalves MT, Mitchell TJ, Lord JM. 2016. Immune ageing and susceptibility to Streptococcus pneumoniae. Biogerontology 17:449-465. 334. Simonetti AF, Viasus D, Garcia-Vidal C, Carratala J. 2014. Management of community-acquired pneumonia in older adults. Ther Adv Infect Dis 2:3-16. 335. Baldo V, Cocchio S, Gallo T, Furlan P, Romor P, Bertoncello C, Buja A, Baldovin T. 2016. Pneumococcal conjugated vaccine reduces the high mortality 331 for community-acquired pneumonia in the elderly: an Italian regional experience. PLoS One 11:e0166637. 336. Jose RJ, Brown JS. 2017. Adult pneumococcal vaccination: advances, impact, and unmet needs. Curr Opin Pulm Med 23:225-230. 337. Moberley S, Holden J, Tatham DP, Andrews RM. 2013. Vaccines for preventing pneumococcal infection in adults. Cochrane Database Syst Rev 1:Cd000422. 338. Bonten MJ, Huijts SM, Bolkenbaas M, Webber C, Patterson S, Gault S, van Werkhoven CH, van Deursen AM, Sanders EA, Verheij TJ, Patton M, McDonough A, Moradoghli-Haftvani A, Smith H, Mellelieu T, Pride MW, Crowther G, Schmoele-Thoma B, Scott DA, Jansen KU, Lobatto R, Oosterman B, Visser N, Caspers E, Smorenburg A, Emini EA, Gruber WC, Grobbee DE. 2015. Polysaccharide conjugate vaccine against pneumococcal pneumonia in adults. N Engl J Med 372:1114-1125. 339. Hendley JO, Sande MA, Stewart PM, Gwaltney JM, Jr. 1975. Spread of Streptococcus pneumoniae in families. I. Carriage rates and distribution of types. J Infect Dis 132:55-61. 340. Fine PE. 1993. Herd immunity: history, theory, practice. Epidemiol Rev 15:265-302. 341. Davis SM, Deloria-Knoll M, Kassa HT, O'Brien KL. 2013. Impact of pneumococcal conjugate vaccines on nasopharyngeal carriage and invasive disease among unvaccinated people: review of evidence on indirect effects. Vaccine 32:133–145. 342. Centers for Disease Control and Prevention. 2005. Direct and indirect effects of routine vaccination of children with 7-valent pneumococcal conjugate vaccine on incidence of invasive pneumococcal disease--United States, 1998-2003. MMWR Morb Mortal Wkly Rep 54:893-897. 332 343. Regev-Yochay G, Katzir M, Strahilevitz J, Rahav G, Finn T, Miron D, Maor Y, Chazan B, Schindler Y, Dagan R. 2017. The herd effects of infant PCV7/PCV13 sequential implementation on adult invasive pneumococcal disease, six years post implementation; a nationwide study in Israel. Vaccine 35:2449–2456. 344. von Gottberg A, de Gouveia L, Tempia S, Quan V, Meiring S, von Mollendorf C, Madhi SA, Zell ER, Verani JR, O'Brien KL, Whitney CG, Klugman KP, Cohen C. 2014. Effects of vaccination on invasive pneumococcal disease in South Africa. N Engl J Med 371:1889-1899. 345. Sahni V, Naus M, Hoang L, Tyrrell GJ, Martin I, Patrick DM. 2012. The epidemiology of invasive pneumococcal disease in British Columbia following implementation of an infant immunization program: increases in herd immunity and replacement disease. Can J Public Health 103:29-33. 346. Galanis I, Lindstrand A, Darenberg J, Browall S, Nannapaneni P, Sjostrom K, Morfeldt E, Naucler P, Blennow M, Ortqvist A, Henriques-Normark B. 2016. Effects of PCV7 and PCV13 on invasive pneumococcal disease and carriage in Stockholm, Sweden. Eur Respir J 47:1208-1218. 347. Garcia Gabarrot G, Lopez Vega M, Perez Giffoni G, Hernandez S, Cardinal P, Felix V, Gabastou JM, Camou T. 2014. Effect of pneumococcal conjugate vaccination in Uruguay, a middle-income country. PLoS One 9:e112337. 348. Rodrigo C, Bewick T, Sheppard C, Greenwood S, Macgregor V, Trotter C, Slack M, George R, Lim WS. 2014. Pneumococcal serotypes in adult non- invasive and invasive pneumonia in relation to child contact and child vaccination status. Thorax 69:168-173. 349. Nair H, Watts AT, Williams LJ, Omer SB, Simpson CR, Willocks LJ, Cameron JC, Campbell H. 2016. Pneumonia hospitalisations in Scotland following the introduction of pneumococcal conjugate vaccination in young children. BMC Infect Dis 16:390. 333 350. Scott JR, Millar EV, Lipsitch M, Moulton LH, Weatherholtz R, Perilla MJ, Jackson DM, Beall B, Craig MJ, Reid R, Santosham M, O’Brien KL. 2012. Impact of more than a decade of pneumococcal conjugate vaccine use on carriage and invasive potential in Native American communities. J Infect Dis 205:280-288. 351. Ansaldi F, de Florentiis D, Canepa P, Ceravolo A, Rappazzo E, Iudici R, Martini M, Botti G, Orsi A, Icardi G, Durando P. 2013. Carriage of Streptoccoccus pneumoniae in healthy adults aged 60 years or over in a population with very high and long-lasting pneumococcal conjugate vaccine coverage in children: rationale and perspectives for PCV13 implementation. Hum Vaccin Immunother 9. 352. Watt JP, O'Brien KL, Katz S, Bronsdon MA, Elliott J, Dallas J, Perilla MJ, Reid R, Murrow L, Facklam R, Santosham M, Whitney CG. 2004. Nasopharyngeal versus oropharyngeal sampling for detection of pneumococcal carriage in adults. J Clin Microbiol 42:4974-4976. 353. Lieberman D, Shleyfer E, Castel H, Terry A, Harman-Boehm I, Delgado J, Peled N, Lieberman D. 2006. Nasopharyngeal versus oropharyngeal sampling for isolation of potential respiratory pathogens in adults. J Clin Microbiol 44:525-528. 354. Levine H, Zarka S, Dagan R, Sela T, Rozhavski V, Cohen DI, Balicer RD. 2012. Transmission of Streptococcus pneumoniae in adults may occur through saliva. Epidemiol Infect 140:561-565. 355. Principi N, Terranova L, Zampiero A, Manzoni F, Senatore L, Rios WP, Esposito S. 2014. Oropharyngeal and nasopharyngeal sampling for the detection of adolescent Streptococcus pneumoniae carriers. J Med Microbiol 63:393-398. 356. Principi N, Terranova L, Zampiero A, Montinaro V, Ierardi V, Peves Rios W, Pelucchi C, Esposito S. 2015. Pharyngeal colonization by Streptococcus pneumoniae in older children and adolescents in a geographical area 334 characterized by relatively limited pneumococcal vaccination coverage. Pediatr Infect Dis J 34:426-432. 357. Carvalho Mda G, Pimenta FC, Moura I, Roundtree A, Gertz RE, Jr., Li Z, Jagero G, Bigogo G, Junghae M, Conklin L, Feikin DR, Breiman RF, Whitney CG, Beall BW. 2013. Non-pneumococcal mitis-group streptococci confound detection of pneumococcal capsular serotype-specific loci in upper respiratory tract. PeerJ 1:e97. 358. Carvalho MdG, Bigogo GM, Junghae M, Pimenta FC, Moura I, Roundtree A, Li Z, Conklin L, Feikin DR, Breiman RF, Whitney CG, Beall B. 2012. Potential nonpneumococcal confounding of PCR-based determination of serotype in carriage. J Clin Microbiol 50:3146-3147. 359. Johnston C, Hinds J, Smith A, van der Linden M, Van Eldere J, Mitchell TJ. 2010. Detection of large numbers of pneumococcal virulence genes in streptococci of the mitis group. J Clin Microbiol 48:2762-2769. 360. Jensen A, Scholz CFP, Kilian M. 2016. Re-evaluation of the taxonomy of the Mitis group of the genus Streptococcus based on whole genome phylogenetic analyses, and proposed reclassification of Streptococcus dentisani as Streptococcus oralis subsp. dentisani comb. nov., Streptococcus tigurinus as Streptococcus oralis subsp. tigurinus comb. nov., and Streptococcus oligofermentans as a later synonym of Streptococcus cristatus. Int J Syst Evol Microbiol 66:4803-4820. 361. Wessels E, Schelfaut JJG, Bernards AT, Claas ECJ. 2012. Evaluation of several biochemical and molecular techniques for identification of Streptococcus pneumoniae and Streptococcus pseudopneumoniae and their detection in respiratory samples. J Clin Microbiol 50:1171-1177. 362. Abdeldaim G, Herrmann B, Korsgaard J, Olcén P, Blomberg J, Strålin K. 2009. Is quantitative PCR for the pneumolysin (ply) gene useful for detection of pneumococcal lower respiratory tract infection? Clin Microbiol Infect 15:565- 570. 335 363. El Aila N, Emler S, Kaijalainen T, De Baere T, Saerens B, Alkan E, Deschaght P, Verhelst R, Vaneechoutte M. 2010. The development of a 16S rRNA gene based PCR for the identification of Streptococcus pneumoniae and comparison with four other species specific PCR assays. BMC Infect Dis 10:104. 364. Whalan RH, Funnell SGP, Bowler LD, Hudson MJ, Robinson A, Dowson CG. 2006. Distribution and genetic diversity of the ABC transporter lipoproteins PiuA and PiaA within Streptococcus pneumoniae and related streptococci. J Bacteriol 188:1031-1038. 365. Prère M-F, Fayet OA. 2011. A specific polymerase chain reaction test for the identification of Streptococcus pneumoniae. Diagn Microbiol Infect Dis 70:45- 53. 366. Salter SJ, Hinds J, Gould KA, Lambertsen L, Hanage WP, Antonio M, Turner P, Hermans PW, Bootsma HJ, O'Brien KL, Bentley SD. 2012. Variation at the capsule locus, cps, of mistyped and non-typeable Streptococcus pneumoniae isolates. Microbiology 158:1560-1569. 367. Manna S, Ortika B, Dunne EM, Holt KE, Kama M, Russell FM, Hinds J, Satzke C. 2017. A novel genetic variant of Streptococcus pneumoniae serotype 11A discovered in Fiji. Clin Microbiol Infect. 368. Stearns JC, Davidson CJ, McKeon S, Whelan FJ, Fontes ME, Schryvers AB, Bowdish DM, Kellner JD, Surette MG. 2015. Culture and molecular- based profiles show shifts in bacterial communities of the upper respiratory tract that occur with age. ISME J 9:1246–1259. 369. Robins-Browne RM, Kharsany AB, Ramsaroop UG. 1982. Detection of pneumococci in the upper respiratory tract: comparison of media and culture techniques. Journal of Clinical Microbiology 16:1-3. 370. Llull D, López R, García E. 2006. Characteristic signatures of the lytA gene provide a basis for rapid and reliable diagnosis of Streptococcus pneumoniae infections. J Clin Microbiol 44:1250-1256. 336 371. Simoes AS, Tavares DA, Rolo D, Ardanuy C, Goossens H, Henriques- Normark B, Linares J, de Lencastre H, Sa-Leao R. 2016. lytA-based identification methods can misidentify Streptococcus pneumoniae. Diagn Microbiol Infect Dis 85:141-148. 372. Ayyadevara S, Thaden JJ, Shmookler Reis RJ. 2000. Discrimination of primer 3'-nucleotide mismatch by taq DNA polymerase during polymerase chain reaction. Anal Biochem 284:11-18. 373. Bru D, Martin-Laurent F, Philippot L. 2008. Quantification of the Detrimental Effect of a Single Primer-Template Mismatch by Real-Time PCR Using the 16S rRNA Gene as an Example. Applied and Environmental Microbiology 74:1660-1663. 374. Rukke HV, Hegna IK, Petersen FC. 2012. Identification of a functional capsule locus in Streptococcus mitis. Mol Oral Microbiol 27:95-108. 375. Skov Sørensen UB, Yao K, Yang Y, Tettelin H, Kilian M. 2016. Capsular polysaccharide expression in commensal Streptococcus species: genetic and antigenic similarities to Streptococcus pneumoniae. mBio 7:e01844-16. 376. Wyllie AL, Rumke LW, Arp K, Bosch AA, Bruin JP, Rots NY, Wijmenga- Monsuur AJ, Sanders EA, Trzciński K. 2016. Molecular surveillance on Streptococcus pneumoniae carriage in non-elderly adults; little evidence for pneumococcal circulation independent from the reservoir in children. Sci Rep 6:34888. 377. Turner P, Hinds J, Turner C, Jankhot A, Gould K, Bentley SD, Nosten F, Goldblatt D. 2011. Improved detection of nasopharyngeal cocolonization by multiple pneumococcal serotypes by use of latex agglutination or molecular serotyping by microarray. J Clin Microbiol 49:1784-1789. 378. Holmberg H, Danielsson D, Hardie J, Krook A, Whiley R. 1985. Cross- reactions between alpha-streptococci and Omniserum, a polyvalent pneumococcal serum, demonstrated by direct immunofluorescence, 337 immunoelectroosmophoresis, and latex agglutination. J Clin Microbiol 21:745- 748. 379. Sorensen UB, Henrichsen J, Chen HC, Szu SC. 1990. Covalent linkage between the capsular polysaccharide and the cell wall peptidoglycan of Streptococcus pneumoniae revealed by immunochemical methods. Microb Pathog 8:325-334. 380. Manning J, Dunne EM, Wescombe PA, Hale JD, Mulholland EK, Tagg JR, Robins-Browne RM, Satzke C. 2016. Investigation of Streptococcus salivarius-mediated inhibition of pneumococcal adherence to pharyngeal epithelial cells. BMC Microbiol 16:225. 381. Kilian M, Riley DR, Jensen A, Bruggemann H, Tettelin H. 2014. Parallel evolution of Streptococcus pneumoniae and Streptococcus mitis to pathogenic and mutualistic lifestyles. mBio 5:e01490-14. 382. Kilian M, Poulsen K, Blomqvist T, Håvarstein LS, Bek-Thomsen M, Tettelin H, Sørensen UBS. 2008. Evolution of Streptococcus pneumoniae and its close commensal relatives. PLoS One 3:e2683. 383. Donati C, Hiller NL, Tettelin H, Muzzi A, Croucher NJ, Angiuoli SV, Oggioni M, Dunning Hotopp JC, Hu FZ, Riley DR, Covacci A, Mitchell TJ, Bentley SD, Kilian M, Ehrlich GD, Rappuoli R, Moxon ER, Masignani V. 2010. Structure and dynamics of the pan-genome of Streptococcus pneumoniae and closely related species. Genome Biol 11:R107. 384. Boersma WG, Lowenberg A, Holloway Y, Kuttschrutter H, Snijder JA, Koeter H. 1993. The role of antigen detection in pneumococcal carriers: a comparison between cultures and capsular antigen detection in upper respiratory tract secretions. Scand J Infect Dis 25:51-56. 385. Levine H, Balicer RD, Zarka S, Sela T, Rozhavski V, Cohen D, Kayouf R, Ambar R, Porat N, Dagan R. 2012. Dynamics of pneumococcal acquisition and carriage in young adults during training in confined settings in Israel. PLoS One 7:e46491. 338 386. Marchese A, Esposito S, Coppo E, Rossi GA, Tozzi A, Romano M, Da Dalt L, Schito GC, Principi N. 2011. Detection of Streptococcus pneumoniae and identification of pneumococcal serotypes by real-time polymerase chain reaction using blood samples from Italian children ≤ 5 years of age with community- acquired pneumonia. Microb Drug Resist 17:419-424. 387. Rolo D, A SS, Domenech A, Fenoll A, Linares J, de Lencastre H, Ardanuy C, Sa-Leao R. 2013. Disease isolates of Streptococcus pseudopneumoniae and non-typeable S. pneumoniae presumptively identified as atypical S. pneumoniae in Spain. PLoS One 8:e57047. 388. Almeida ST, Nunes S, Santos Paulo AC, Valadares I, Martins S, Breia F, Brito-Avo A, Morais A, de Lencastre H, Sa-Leao R. 2014. Low prevalence of pneumococcal carriage and high serotype and genotype diversity among adults over 60 years of age living in Portugal. PLoS One 9:e90974. 389. Mackenzie GA, Leach AJ, Carapetis JR, Fisher J, Morris PS. 2010. Epidemiology of nasopharyngeal carriage of respiratory bacterial pathogens in children and adults: cross-sectional surveys in a population with high rates of pneumococcal disease. BMC Infect Dis 10:304. 390. Mueller JE, Yaro S, Ouedraogo MS, Levina N, Njanpop-Lafourcade BM, Tall H, Idohou RS, Sanou O, Kroman SS, Drabo A, Nacro B, Millogo A, van der Linden M, Gessner BD. 2012. Pneumococci in the African meningitis belt: meningitis incidence and carriage prevalence in children and adults. PLoS One 7:e52464. 391. van Hoek AJ, Sheppard CL, Andrews NJ, Waight PA, Slack MP, Harrison TG, Ladhani SN, Miller E. 2014. Pneumococcal carriage in children and adults two years after introduction of the thirteen valent pneumococcal conjugate vaccine in England. Vaccine 32:4349-4355. 392. Smith-Vaughan H, Byun R, Nadkarni M, Jacques NA, Hunter N, Halpin S, Morris PS, Leach AJ. 2006. Measuring nasal bacterial load and its association with otitis media. BMC Ear Nose Throat Disord 6:10. 339 393. Albrich WC, Madhi SA, Adrian PV, van Niekerk N, Mareletsi T, Cutland C, Wong M, Khoosal M, Karstaedt A, Zhao P, Deatly A, Sidhu M, Jansen KU, Klugman KP. 2012. Use of a rapid test of pneumococcal colonization density to diagnose pneumococcal pneumonia. Clin Infect Dis 54:601-609. 394. Brotons P, Bassat Q, Lanaspa M, Henares D, Perez-Arguello A, Madrid L, Balcells R, Acacio S, Andres-Franch M, Marcos MA, Valero-Rello A, Munoz-Almagro C. 2017. Nasopharyngeal bacterial load as a marker for rapid and easy diagnosis of invasive pneumococcal disease in children from Mozambique. PLoS One 12:e0184762. 395. Wolter N, Tempia S, Cohen C, Madhi SA, Venter M, Moyes J, Walaza S, Malope-Kgokong B, Groome M, du Plessis M, Magomani V, Pretorius M, Hellferscee O, Dawood H, Kahn K, Variava E, Klugman KP, von Gottberg A. 2014. High nasopharyngeal pneumococcal density, increased by viral coinfection, is associated with invasive pneumococcal pneumonia. J Infect Dis 210:1649-1657. 396. Alpkvist H, Athlin S, Naucler P, Herrmann B, Abdeldaim G, Slotved HC, Hedlund J, Stralin K. 2015. Clinical and microbiological factors associated with high nasopharyngeal pneumococcal density in patients with pneumococcal pneumonia. PLoS One 10:e0140112. 397. Albrich WC, Madhi SA, Adrian PV, van Niekerk N, Telles JN, Ebrahim N, Messaoudi M, Paranhos-Baccala G, Giersdorf S, Vernet G, Mueller B, Klugman KP. 2014. Pneumococcal colonisation density: a new marker for disease severity in HIV-infected adults with pneumonia. BMJ Open 4:e005953. 398. Roca A, Bottomley C, Hill PC, Bojang A, Egere U, Antonio M, Darboe O, Greenwood BM, Adegbola RA. 2012. Effect of age and vaccination with a pneumococcal conjugate vaccine on the density of pneumococcal nasopharyngeal carriage. Clin Infect Dis 55:816-824. 399. O'Brien KL, Millar EV, Zell ER, Bronsdon M, Weatherholtz R, Reid R, Becenti J, Kvamme S, Whitney CG, Santosham M. 2007. Effect of 340 pneumococcal conjugate vaccine on nasopharyngeal colonization among immunized and unimmunized children in a community-randomized trial. J Infect Dis 196:1211-1220. 400. Dagan R, Juergens C, Trammel J, Patterson S, Greenberg D, Givon-Lavi N, Porat N, Gruber WC, Scott DA. 2017. PCV13-vaccinated children still carrying PCV13 additional serotypes show similar carriage density to a control group of PCV7-vaccinated children. Vaccine 35:945-950. 401. Kuipers K, van Selm S, van Opzeeland F, Langereis JD, Verhagen LM, Diavatopoulos DA, de Jonge MI. 2017. Genetic background impacts vaccine- induced reduction of pneumococcal colonization. Vaccine 35:5235-5241. 402. Hoe E, Boelsen LK, Toh ZQ, Sun GW, Koo GC, Balloch A, Marimla R, Dunne EM, Tikoduadua L, Russell FM, Satzke C, Mulholland EK, Licciardi PV. 2015. Reduced IL-17A secretion is associated with high levels of pneumococcal nasopharyngeal carriage in Fijian children. PLoS One 10:e0129199. 403. Olwagen CP, Adrian PV, Madhi SA. 2017. Comparison of traditional culture and molecular qPCR for detection of simultaneous carriage of multiple pneumococcal serotypes in African children. Sci Rep 7:4628. 404. Segal N, Greenberg D, Dagan R, Ben-Shimol S. 2016. Disparities in PCV impact between different ethnic populations cohabiting in the same region: a systematic review of the literature. Vaccine 34:4371-4377. 405. Flannery B, Schrag S, Bennett NM, Lynfield R, Harrison LH, Reingold A, Cieslak PR, Hadler J, Farley MM, Facklam RR, Zell ER, Whitney CG. 2004. Impact of childhood vaccination on racial disparities in invasive Streptococcus pneumoniae infections. JAMA 291:2197-2203. 406. Tocheva AS, Jefferies JMC, Rubery H, Bennett J, Afimeke G, Garland J, Christodoulides M, Faust SN, Clarke SC. 2011. Declining serotype coverage or new pneumococcal conjugate vaccines relating to the carriage of Streptococcus pneumoniae in young children. Vaccine 29:4400-4404. 341 407. Pelton SI, Loughlin AM, Marchant CD. 2004. Seven valent pneumococcal conjugate vaccine immunization in two Boston communities: changes in serotypes and antimicrobial susceptibility among Streptococcus pneumoniae isolates. Pediatr Infect Dis J 23:1015-1022. 408. Kandasamy R, Gurung M, Thapa A, Ndimah S, Adhikari N, Murdoch DR, Kelly DF, Waldron DE, Gould KA, Thorson S, Shrestha S, Hinds J, Pollard AJ. 2015. Multi-serotype pneumococcal nasopharyngeal carriage prevalence in vaccine naïve Nepalese children, assessed using molecular serotyping. PLoS One 10:e0114286. 409. Valente C, Hinds J, Gould KA, Pinto FR, de Lencastre H, Sá-Leão R. 2016. Impact of the 13-valent pneumococcal conjugate vaccine on Streptococcus pneumoniae multiple serotype carriage. Vaccine 34:4072-4078. 410. Friedel V, Chang A, Wills J, Vargas R, Xu Q, Pichichero ME. 2013. Impact of respiratory viral infections on alpha-hemolytic streptococci and otopathogens in the nasopharynx of young children. Pediatr Infect Dis J 32:27-31. 411. Tano K, Grahn-Hakansson E, Holm SE, Hellstrom S. 2000. Inhibition of OM pathogens by alpha-hemolytic streptococci from healthy children, children with SOM and children with rAOM. Int J Pediatr Otorhinolaryngol 56:185-190. 412. Roos K, Hakansson EG, Holm S. 2001. Effect of recolonisation with "interfering" alpha streptococci on recurrences of acute and secretory otitis media in children: randomised placebo controlled trial. BMJ 322:210-212. 413. Walls T, Power D, Tagg J. 2003. Bacteriocin-like inhibitory substance (BLIS) production by the normal flora of the nasopharynx: potential to protect against otitis media? J Med Microbiol 52:829-833. 414. Tano K, Grahn Hakansson E, Wallbrandt P, Ronnqvist D, Holm SE, Hellstrom S. 2003. Is hydrogen peroxide responsible for the inhibitory activity of alpha-haemolytic streptococci sampled from the nasopharynx? Acta Otolaryngol 123:724-729. 342 415. Gupta VK, Paul S, Dutta C. 2017. Geography, ethnicity or subsistence-specific variations in human microbiome composition and diversity. Front Microbiol 8:1162. 416. Mueller NT, Bakacs E, Combellick J, Grigoryan Z, Dominguez-Bello MG. 2015. The infant microbiome development: mom matters. Trends Mol Med 21:109-117. 417. Bosch AATM, Levin E, van Houten MA, Hasrat R, Kalkman G, Biesbroek G, de Steenhuijsen Piters WAA, de Groot P-KCM, Pernet P, Keijser BJF, Sanders EAM, Bogaert D. 2016. Development of upper respiratory tract microbiota in infancy is affected by mode of delivery. EBioMedicine 9:336-345. 418. Li M, Wang M, Donovan SM. 2014. Early development of the gut microbiome and immune-mediated childhood disorders. Semin Reprod Med 32:74-86. 419. Goodrich JK, Davenport ER, Beaumont M, Jackson MA, Knight R, Ober C, Spector TD, Bell JT, Clark AG, Ley RE. 2016. Genetic determinants of the gut microbiome in UK twins. Cell Host Microbe 19:731-743. 420. Knights D, Lassen KG, Xavier RJ. 2013. Advances in inflammatory bowel disease pathogenesis: linking host genetics and the microbiome. Gut 62:1505- 1510. 421. Spor A, Koren O, Ley R. 2011. Unravelling the effects of the environment and host genotype on the gut microbiome. Nat Rev Micro 9:279–290. 422. Kurilshikov A, Wijmenga C, Fu J, Zhernakova A. 2017. Host genetics and gut microbiome: challenges and perspectives. Trends Immunol 38:633-647. 423. Bonder MJ, Kurilshikov A, Tigchelaar EF, Mujagic Z, Imhann F, Vila AV, Deelen P, Vatanen T, Schirmer M, Smeekens SP, Zhernakova DV, Jankipersadsing SA, Jaeger M, Oosting M, Cenit MC, Masclee AAM, Swertz MA, Li Y, Kumar V, Joosten L, Harmsen H, Weersma RK, Franke L, Hofker MH, Xavier RJ, Jonkers D, Netea MG, Wijmenga C, Fu J, 343 Zhernakova A. 2016. The effect of host genetics on the gut microbiome. Nat Genet 48:1407-1412. 424. Morton ER, Lynch J, Froment A, Lafosse S, Heyer E, Przeworski M, Blekhman R, Segurel L. 2015. Variation in rural African gut microbiota is strongly correlated with colonization by Entamoeba and subsistence. PLoS Genet 11:e1005658. 425. Schnorr SL, Candela M, Rampelli S, Centanni M, Consolandi C, Basaglia G, Turroni S, Biagi E, Peano C, Severgnini M, Fiori J, Gotti R, De Bellis G, Luiselli D, Brigidi P, Mabulla A, Marlowe F, Henry AG, Crittenden AN. 2014. Gut microbiome of the Hadza hunter-gatherers. Nat Commun 5:3654. 426. Kau AL, Ahern PP, Griffin NW, Goodman AL, Gordon JI. 2011. Human nutrition, the gut microbiome, and immune system: envisioning the future. Nature 474:327-336. 427. Round JL, Mazmanian SK. 2009. The gut microbiome shapes intestinal immune responses during health and disease. Nat Rev Immunol 9:313-323. 428. Fujimura KE, Demoor T, Rauch M, Faruqi AA, Jang S, Johnson CC, Boushey HA, Zoratti E, Ownby D, Lukacs NW, Lynch SV. 2014. House dust exposure mediates gut microbiome Lactobacillus enrichment and airway immune defense against allergens and virus infection. Proc Natl Acad Sci U S A 111:805-810. 429. Schultz JT, Vatucawaqa P, Tuivaga J. 2005. 2004 Fiji National Nutrition Survey. Ministry of Health, Suva, Fiji. 430. Hoen AG, Li J, Moulton LA, O'Toole GA, Housman ML, Koestler DC, Guill MF, Moore JH, Hibberd PL, Morrison HG, Sogin ML, Karagas MR, Madan JC. 2015. Associations between gut microbial colonization in early life and respiratory outcomes in cystic fibrosis. J Pediatr 167:138-147.e133. 431. Neal EFG, Flasche S, Rafai E, Reyburn RC, Kama M, Koyamaibole L, Dunne EM, Satzke C, Ortika BD, Ratu T, Bright KJ, Kado J, Tikoduadua 344 L, Devi R, Tuivaga E, Mulholland EK, Edmunds J, Russell FM. 2017. Pneumococcal carriage varies by contact, age, and ethnicity in Fiji. Australian Epidemiological Association Annual Scientific Meeting, Sydney, Australia. 432. Tung J, Barreiro LB, Burns MB, Grenier JC, Lynch J, Grieneisen LE, Altmann J, Alberts SC, Blekhman R, Archie EA. 2015. Social networks predict gut microbiome composition in wild baboons. Elife 4:e05224. 433. Meadow JF, Bateman AC, Herkert KM, O’Connor TK, Green JL. 2013. Significant changes in the skin microbiome mediated by the sport of roller derby. PeerJ 1:e53. 434. Lax S, Smith DP, Hampton-Marcell J, Owens SM, Handley KM, Scott NM, Gibbons SM, Larsen P, Shogan BD, Weiss S, Metcalf JL, Ursell LK, Vazquez-Baeza Y, Van Treuren W, Hasan NA, Gibson MK, Colwell R, Dantas G, Knight R, Gilbert JA. 2014. Longitudinal analysis of microbial interaction between humans and the indoor environment. Science 345:1048- 1052. 435. Taylor R, Carter K, Naidu S, Linhart C, Azim S, Rao C, Lopez AD. 2013. Divergent mortality trends by ethnicity in Fiji. Aust N Z J Public Health 37:509- 515. 345 Minerva Access is the Institutional Repository of The University of Melbourne Author/s: Boelsen, Laura Kate Title: Impact of pneumococcal vaccination on carriage in a Fijian population Date: 2018 Persistent Link: http://hdl.handle.net/11343/214548 File Description: Impact of pneumococcal vaccination on carriage in a Fijian population Terms and Conditions: Terms and Conditions: Copyright in works deposited in Minerva Access is retained by the copyright owner. The work may not be altered without permission from the copyright owner. Readers may only download, print and save electronic copies of whole works for their own personal non-commercial use. Any use that exceeds these limits requires permission from the copyright owner. Attribution is essential when quoting or paraphrasing from these works.