The nose is the best niche for detection of pneumococcal colonisation following experimental challenge in adults of all ages


 Background

Previous studies have suggested that the pneumococcal niche changes from the nose to the oropharynx with age. We use an Experimental Human Pneumococcal Challenge model to investigate pneumococcal colonisation in each anatomical niche with age.
Methods

Healthy adults (n = 112) were intranasally inoculated with Streptococcus pneumoniae serotype 6B (Spn6B) and were categorised as young 18-55yrs (n = 57) or older > 55yrs (n = 55). Colonisation status (frequency and density) was determined by multiplex qPCR targeting the lytA and cpsA-6A/B genes in both raw and culture-enriched nasal wash and oropharyngeal swab samples collected at 2-, 7- and 14-days post-exposure. For older adults, raw and culture-enriched saliva samples were also assessed.
Results

64% of NW samples and 54% of OPS samples were positive for Spn6B in young adults, compared to 35% of NW samples, 24% of OPS samples and 6% of saliva samples in older adults. Many colonisation events were only detected in culture-enriched samples. Experimental colonisation was detected in 72% of young adults by NW and 63% by OPS. In older adults, this was 51% by NW, 36% by OPS and 9% by saliva.
Conclusions

The nose is the best niche for detection of experimental pneumococcal colonisation in both young and older adults.


Introduction
The natural ora of the human upper respiratory tract (URT) is in constant interaction with the external environment, resulting in a diverse ecology of microorganisms colonising epithelial surfaces in the oral and nasal cavities and oropharynx 1 . Streptococcus pneumoniae (Spn, pneumococcus) can proliferate and establish colonisation as a non-pathogenic symbiont, which is asymptomatic for the host, particularly in adults 2,3 . Pneumococcus can however migrate to other organs, resulting in pneumonia, meningitis, and bacteraemia, causing signi cant morbidity and mortality worldwide, especially in the very young and very old.
Colonisation of the URT has been shown to be the source of, and pre-requisite for, pneumococcal disease 4 and its transmission throughout the community 5 . Accurate detection of pneumococcal colonisation is crucial to assessing disease potential, as well as direct and indirect impact of vaccines. The World Health Organisation (WHO) recommends the use of nasopharyngeal swabs for pneumococcal colonisation detection in children and both nasopharyngeal and oropharyngeal swabs (OPS) in adults 6,7 . WHO recommendations were based on culture-based methods and more recent studies using sensitive molecular methods have encouraged other sampling methods such as saliva or nasal wash 8, 9 , due to improved comfort and pneumococcal detection.
Many studies have investigated pneumococcal colonisation prevalence in the rst few years of life, showing that the nasopharynx of 40-95% of young children 10,11 is naturally colonised with pneumococcus. With the onset of adulthood, falling pneumococcal disease rates are accompanied by decreased nasopharyngeal colonisation rates at 10-25% [12][13][14][15] . In older adults, a population at increased risk of pneumococcal disease, colonisation prevalence and density are reported to be lower (19), but study results are heterogeneous with some reporting very low colonisation rates of 1.9-4.2% [16][17][18] . Our recent systematic review analysed 29 studies, 18 with participant-level data (representing 6290 participants), and reported prevalence of detected pneumococcal colonisation of 0-39% by conventional culture methods and 3-23% by molecular methods 19 .
Various theories have been put forward to explain the paradox of increased disease risk with decreased colonisation prevalence, including a more transient colonisation dynamic in older adults 17 , the poor sensitivity of conventional culture methods to detect low density colonisation in polymicrobial samples 20,21 and a transition of the pneumococcal niche from the nasopharynx to the oral cavity with ageing 20 .
The Experimental Human Pneumococcal Colonisation (EHPC) model is a safe, reproducible, and controlled method of studying colonisation dynamics in adult human participants, as the challenge dose and time of exposure is known 22 . We have recently expanded this model to older adults and reported that experimental colonisation was established in 39% of participants (25/64) with no adverse events. Colonisation occurred in 47% (9/19) of participants aged 50-59 and in 42% (13/31) of those aged 60-69 compared to 21% (3/14) in those aged ≥ 70 years. Colonisation density was similar between old and younger adults 23 .
We used 780 nasal, oropharyngeal and saliva samples to de ne precisely whether the niche of experimental pneumococcal colonisation changes in adults with increasing age. Association with bacterial density was also analysed.

Materials And Methods
Clinical trial design, participant cohorts and sample analysis The methodology and inclusion/exclusion criteria for EHPC studies have been previously described 24 .
Brie y, participants were healthy adults aged ≥ 18 years with no major risk factors for pneumococcal disease, colonisation, or transmission (such as: cigarette smoking; close contact with children aged < 5 years; healthcare work or caring responsibilities; steroid therapy and respiratory or immunosuppressive comorbidities). Studies were conducted in accordance with the Declaration of Helsinki and Good Clinical Practice procedures. All participants provided written informed consent and underwent a safety screening. Studies were submitted to and approved by a local NHS Research Ethics Committee. In this study, samples have been sourced from two separate EHPC clinical studies. The young vaccine study was a double-blind, randomised controlled trial conducted between September 2013 -April 2014 investigating the impact of PCV-13 vaccination on experimental pneumococcal colonisation in adults aged 18-50 (NHS REC number 12/NW/0873, ISRCTN number 45340436 registered 18/11/2013) 25 . In the present analysis, we include participants from the control arm only (vaccinated with Hepatitis A vaccine rather than PCV-13). The older adults study was an observational study of the effect of age on experimental pneumococcal colonisation in adults aged 50-80 conducted between June 2016 -February 2018 (NHS REC number 16/NW/0031, ISRCTN number 10948363 registered 08/11/2016) 23 . In total, 112 participants were analysed. The sample sets can be combined because the inoculation dose and strain, as well as the methods for inoculation, nasal wash collection and processing remained identical.
Furthermore, pneumococcal colonisation rates and serotype distribution in healthy adults in Liverpool appears to be relatively stable, based on data collected between 2010-2017 26 . The low prevalence of naturally occurring Spn6 provides reassurance that detected Spn6B is the inoculated strain. Using the full sample set, participants were split into young 18-55 years (n = 57) and older adults > 55 years (n = 55), in keeping with the cut-off used by Marrie et al. 27 .
In both studies, participants were inoculated with 80,000 colony-forming units (CFU) per nostril of live Spn6B pneumococcus (BHN418, GenBank accession number ASHP00000000.1) 25 . Nasal wash (NW) samples were collected before inoculation to screen for pneumococcal colonisation acquired from community. Depending on the study, NW, oropharyngeal swabs (OPS) and saliva samples were then collected post pneumococcal exposure. In the young vaccine study, NW and OPS samples were collected on days 2, 7, 14 (only culture-positives) and 21 post exposure. In the older adults study, NW, OPS and saliva samples were collected on days 2, 7, 9, 14, 22 (only culture-positives) and 29 post exposure. Pneumococcal colonisation status in both studies was determined by NW culture (not discussed here) and by multiplex real-time polymerase chain reaction (qPCR) targeting the lytA and cpsA-6A/B genes in raw and culture-enriched NW, OPS, and saliva samples. The raw and culture-enriched NW, OPS, and saliva qPCR data from days 2, 7 and 14 (covered in both studies) were included in this study.
NW collection was performed as previously described 24 . Brie y, 20mL of 0.9% sodium chloride solution in total (10mL saline per nostril) was introduced using a syringe and held for a few seconds in the participant's nose before being expelled into a sterile container. NW was centrifuged at 4000rpm for 10 minutes. Supernatant was collected and pellet was resuspended in STGG (Skimmed milk-Tryptone-Glucose-Glycerine) 20% glycerol before storage at -80°C. OPS samples were collected in 1 mL STGG and stored at -80°C, until further use. Saliva samples were collected using the salivette device (Sarstedt, UK).
Following collection, each salivette was centrifuged at 4000rpm for 3minutes at 4 o C and after measuring the volume of saliva liquid retrieved, both pellet and supernatant were re-suspended in equal volume of STGG 50% glycerol and stored at -80°C, until further use.
Preparation of raw and culture-enriched NW, OPS, and saliva samples Before DNA extraction, samples were thawed for 30 minutes at room temperature and vigorously vortexed for 20 seconds. 300µL raw NW pellet, 200µL raw OPS and 300µL raw saliva aliquots were prepared. For culture-enrichment, 50µL of NW pellet, OPS or saliva samples were plated on Columbia blood agar supplemented with 5% horse blood and 80µL gentamicin 1mg/mL. Plates were incubated overnight at 37°C in 5%CO 2 . Following incubation, 2mL of STGG 20% glycerol was added onto each plate and microbial growth scraped off. 300µL culture-enriched NW, 200µL culture-enriched OPS and 300µL culture-enriched saliva aliquots were prepared. Both raw and culture-enriched NW, OPS and saliva aliquots were stored at -20 o C until DNA extraction took place.
DNA extraction on raw and culture-enriched NW, OPS, and saliva samples Bacterial genomic DNA was extracted from raw and culture-enriched NW, OPS, and saliva samples. On the day of the extraction, prepared aliquots were thawed at room temperature and vigorously vortexed for 20 seconds. Samples were pelleted at 20,238xg for 10 minutes. Following centrifugation, 300µL of lysis buffer with protease (Agowa Mag mini-DNA extraction kit; LGC Genomics, Germany), 100µL of zirconium beads (diameter of 0.1 mm), and 300µL of phenol pH 8.0 (toxic, performed in a cabinet with charcoal lter) were added to the pellets. Samples were mechanically disrupted at 50 Hz for 3 minutes in a tissue homogenizer followed by 3 minutes on ice, twice. The samples were then centrifuged for 10 minutes at 9,391xg, and the upper aqueous phase was transferred to a sterile 1.5mL Eppendorf tube pre-lled with 600µL binding buffer and 10µL magnetic beads. The samples were incubated in a mixing machine (~ 265 rpm) for 1 hour at room temperature, then washed twice with 200µL of wash buffers 1 and 2. Magnetic beads were dried at 55°C for 10 minutes, eluted in 63µL of elution buffer and stored at -20°C until further use.
Quanti cation of pneumococcal DNA by multiplex qPCR in NW, OPS, and saliva pellet samples We used a multiplex qPCR targeting the lytA 28 and cpsA-6A/B 29 genes as previously described 30 . The reaction mixture of 25µL contained 0.6µM of each lytA primer, 0.3µM of lytA probe, 0.4µM of each cpsA-6A/B primer, 0.2µM of cpsA-6A/B probe, 12.5µM of Taqman Gene Expression Master Mix (Applied Biosystems, USA) and 2.5µL of extracted DNA. The qPCR reaction was run on a Mx3005P machine (Agilent Technologies, USA) on the following programme: 10 minutes at 95°C followed by 40 cycles of 15 seconds at 95°C and 1 minute at 60°C. For standard curve, Spn6B DNA was extracted using the QIAamp DNA mini kit (Qiagen, Germany) and serially diluted 1:10 from 4.14x10 6 copies in 2.5 µL. A sample was considered positive if duplicates had a CT value less than 40.

Quanti cation and Statistical analysis
Statistical analysis was performed by GraphPad Prism version 5.0. Contingency tables were used to assess the differences in the pneumococcal frequency in raw and culture-enriched extracted samples in both niches in each age group and in the colonisation frequency between the two niches in each age group or between the two age groups on all study days. The association was tested using Fisher's exact test and considered signi cant if P < 0.05 (two-sided). Unpaired t-tests were used to compare colonisation densities in experimentally colonised participants, calculated from cpsA-6A/B gene copies in raw samples, between niches and age groups. Samples that were negative in the raw sample but positive in the corresponding culture-enriched sample were imputed a density of 1 copy/ml. A Generalized Linear Model (GLM) with binomial distribution was also used to explore the relationship between age and pneumococcal frequency in both niches.

Results
Culture-enrichment increased pneumococcal detection in all sample types in both age groups.
To detect pneumococcal presence, DNA was extracted from both raw and culture-enriched samples. Samples positive for both lytA and cpsA-6A/B genes were de ned as Spn6B+, whereas those positive only for the lytA gene as lytA+.
For young adults (n = 57, Supplementary There was greater additional bene t in culture-enriching oropharyngeal samples (OPS and saliva) than NW in both age groups. In young adults, 21% (20/94) and 57% (45/79) of Spn6B + samples were detected by culture-enrichment only in NW and OPS respectively, indicating that culture-enrichment increased pneumococcal detection 2.7 times more in OPS samples than NW. In older adults, 35% (20/57), 64% (25/39) and 56% (5/9) of Spn6B + samples were detected by culture-enrichment only in NW, OPS and saliva respectively, indicating that culture-enrichment increased pneumococcal detection 1.1-1.8 times more in oropharyngeal (OPS and saliva) samples than in NW. Moreover, in case of the oropharyngeal niche, pneumococcal detection rates in OPS were 1.6 times higher than in saliva.
Pneumococcal colonisation frequency and density with ageing in both nose and oropharynx.
To investigate the kinetics of experimental colonisation in both niches, we assessed the colonisation frequency and density of pneumococcal DNA in both age groups on days 2, 7 and 14 post pneumococcal exposure as shown in Fig. 1. For D14, only data from culture-positive participants in the older adults study was analysed to ensure comparability with the young vaccine study data. Pneumococcal colonisation frequency was signi cantly higher in young than older adults at all study days post exposure ( Comparing the two niches in both age groups separately, pneumococcal frequency was higher in the nose than the oropharynx at days 2 and 7 and similar at day 14 post pneumococcal exposure in both age groups (Fig. 1A, older adults NW 22/55 (40%) vs OPS 10/55 (18.2%) D2, P = 0.020).
Pneumococcal colonisation density was higher in young adults than in older adults for OPS (P = 0.008) but not for NW (P = 0.30) (Fig. 1B). Pneumococcal colonisation density was higher in NW than in OPS for older adults (P = 0.016) but not for young adults (P = 0.09) (Fig. 1B).
Pneumococcal presence is higher in the nose (NW) than oropharynx (OPS) within and between young and older adults.
In order to compare overall pneumococcal detection rates between the nasal and oropharyngeal niches, an overall nasal (combined raw and culture-enriched NW) and oropharyngeal (combined raw and cultureenriched OPS) pro le were created for each participant and plotted on a heat map as shown in Fig. 2. When extraction from raw and CE samples yielded different results, the positive result was retained regardless of the method. Participants with qPCR-negative samples on all study days were de ned as negative (shown in white). Those with a Spn6B + sample on any study day were classi ed as experimentally colonised (shown in black). Participants with a lytA + but cpsA-6A/B -sample on any study day were classi ed as colonised with a lytA-carrying streptococcus (shown in grey). Participants with Spn6B + samples and lytA+ (cpsA-6A/B -) samples on different study days were classi ed as cocolonised (shown with hatched shading).
Using combined raw and culture-enrichment methods, higher pneumococcal presence was detected in the nose than the oropharynx in both age groups with statistical signi cance in older adults (Supplementary   Table S4, P = 0.016). Pneumococcal presence was signi cantly different between young and older adults in both NW (Supplementary Table S4, P = 0.026) and OPS (Supplementary Table S4, P = 0.004).
OPS is more sensitive than saliva for pneumococcal detection in older adults.
Both OPS (described above, combined raw and culture-enriched OPS) and saliva (combined raw and culture-enriched saliva) samples were used to assess pneumococcal colonisation in the oropharynx of older adults. In saliva, 36/55 (65%) participants were negative, 5/55 (9%) experimentally colonised, 14/55 (25%) colonised with a lytA-carrying streptococcus and no co-colonised were detected. Therefore, overall, the oropharynx of 20/55 (36%) older adults, as assessed by both OPS and saliva, were colonised with Spn6B (Fig. 2B). Only 5 of these (25%) were detected in saliva compared to 19 (95%) in OPS, (Fig. 2B), indicating that in the older age group, saliva is a less sensitive method of assessing pneumococcal colonisation than OPS.

Discussion
This study investigated whether the pneumococcal colonisation niche alters with increasing age following experimental human challenge. Our ndings show that regardless of age, the nasal niche had the highest percentage of experimentally colonised participants.
Experimental colonisation involves the direct inoculation of pneumococcus into the nose and therefore may not precisely imitate natural colonisation dynamics. Nevertheless, we have described in a series of independent studies that participants who become colonised following inoculation develop a consistent colonisation episode of 1-3 weeks of similar density to natural colonisation 24,31 . Our ndings agree with studies of natural colonisation where a higher incidence of pneumococcal growth is found in individuals' nasopharyngeal samples compared with their oropharyngeal samples 15,32 .
Young adults showed a higher percentage of colonised participants than their older counterparts at each study day following inoculation, in both the nose and the oropharynx. The relationship between age and prevalence of colonisation has been well documented [33][34][35] , in agreement with our ndings that prevalence of colonisation decreases with age. This decrease did not reach statistical signi cance when samples were analysed using classical microbiology methods (P = 0.19), in keeping with the increased sensitivity of molecular methods.
Culture-enrichment of samples increased pneumococcal detection in both niches in both age groups. This extra step has been shown previously to increase pneumococcal detection in saliva 20,36 and our group now uses it routinely when analysing clinical trial samples. Although labour-intensive, we believe that its added value justi es recommending it as standard practice in combination with analysing raw samples 37 .
The strength of our study is the collection of paired longitudinal nasal and oropharyngeal samples before and after pneumococcal inoculation of a known strain. Coupled with culture enrichment and molecular methods for capsular polysaccharide-speci c detection, this allows the precise determination of the frequency and density of bacteria for each study day and niche, according to age. A weakness of our work is the lack of saliva in the young cohort and that collection of samples in the older and young cohorts was conducted during different studies. However, the strain used for inoculation as well as the methods for inoculation, nasal wash collection and processing remained identical, allowing for direct comparison of cohorts as done previously 23 . A further limitation could extend to the methods used for nasal sampling and saliva collection. Nasal wash is more comfortable and more sensitive than nasopharyngeal swab for pneumococcal detection in adults 9 , however it is not always feasible outside of a clinic setting. It may be that we would not have seen such a difference between the nose and oropharynx if we had used nasopharyngeal swabs. Our saliva detection levels were also much lower than those reported elsewhere when a spitting method was used for sample collection 20,37 . In older adults, with drier mouths, low sample volumes could be obscured by the use of the Salivette device.
Our results indicate age-related host factors could affect colonisation prevalence in these age groups.
The percentage of participants showing experimental carriage fell between days 2 and day 14 in the nose of older age group, as assessed by NW. This could be evidence of pneumococcal colonisation clearance, which is in uenced by host immune responses such as local phagocytic function and acute mucosal in ammatory responses 37 as well as mechanisms known to be affected by immunosenescence (toll like receptors and reduction in the function of host signalling pathways) 38 . Unlike pneumococcal colonisation prevalence, pneumococcal density in NW was unaffected by sampling age. Several studies have found an opposite relationship between nasopharyngeal colonisation density and age indicating higher pneumococcal densities in younger subjects [39][40][41] .
It has been established that the complexity and diversity of the microbiome increases from the nasopharynx to the oropharynx to saliva 42 . Investigation of NW, OPS and saliva samples in this study showed a higher percentage of participants colonised with a lytA-carrying streptococcus in OPS (and saliva in older adults) in contrast to NW in both age groups. Due to higher levels of species diversity within the oral cavity, and the capacity of pneumococci to exchange genes with other streptococci, the use of lytA as a PCR target needs to be treated with caution. Positive lytA PCR results may indicate the presence of non-pneumococcal species in addition to pneumococcal species, leading to false positives, which was indeed the case here. In addition, in our experience the addition of saliva sampling in older adults was not bene cial as it is less sensitive when compared with OPS.
In summary, this study has shown that the optimal sampling niche to detect experimental pneumococcal colonisation is the nose regardless of age, as assessed by NW. However, individuals show different colonised niches, so reducing sampling to only the nose would exclude detection of pneumococcal colonisation in some patients (both young and older adults). Future studies could investigate the site of pneumococcal colonisation over a longer time-period following experimental inoculation. The current study examined a short period of time, and studies in the literature of natural carriage are snapshots in time.  Experimental pneumococcal colonization in young (18-55yrs, n=57) and older adults (>55yrs, n=55). A.