Nasal Tissue Extraction Is Essential for Characterization of the Murine Upper Respiratory Tract Microbiota

The nasal microbiota is composed of species that play a role in the colonization success of pathogens, including Streptococcus pneumoniae and Staphylococcus aureus. Murine models provide the ability to explore disease pathogenesis, but little is known about the natural murine nasal microbiota.

within the host or being transmitted to other carriers via respiratory droplets (1). Pathogens could inhibit colonization by other pathogens, as epidemiological studies have identified a negative correlation between S. pneumoniae and S. aureus in children (2,3). Consistent with this, studies of pneumococcal vaccination have shown that the removal or reduction of S. pneumoniae is associated with increased S. aureus colonization (4). The contribution of the resident microbiota is relatively unknown, although a recent study demonstrated that a nasopharyngeal bacterial composition with high levels of Corynebacterium or Dolosigranulum species was correlated with a reduced risk of lower respiratory infections (5). The antipathogenic activities of Corynebacterium species have been identified in vitro (6,7). A further understanding of the agonistic and antagonistic interactions between microbial species may explain why certain populations are more susceptible to colonization and infection.
Mechanistic studies involving pneumococcal colonization and the URT microbiota are difficult due to high interindividual variability, low bacterial biomass, and diverse topography. An experimental human pneumococcal colonization model identified increased a-diversity (higher microbial community diversity within a subject) in subjects with successful pneumococcal colonization (8). Identification of the contribution of individual bacterial species was challenging due to distinct microbial compositions between subjects. Mouse models allow for greater control over microbial composition and have been used to assess the alterations of the nasal microbiota during pneumococcal infection (9,10); however, little is known about the naive URT microbiota composition in mice.
Assessment of murine nasal bacteria is predominated by nasal washes. Cannulation and flushing from trachea to nares, often called a nasal wash, are the most frequently used methods to assess nasal colonization, although tissue extraction is sometimes used to assess adherent or invasive pathogens (11)(12)(13)(14). Differences in the duration of colonization and bacterial interactions have been identified using different tissue extraction methods (14)(15)(16)(17). No studies have compared these two sampling methods for their efficiency at extracting the native microbiota and their impact on pathogenmicrobiota associations. We used culture-dependent and -independent methods to assess the composition of the URT microbiota using nasal washes and complete tissue collection. Nasal washes were sufficient for sampling the nasopharynx but did not disrupt the turbinates. Nasal tissue collection yielded a distinct community with increased bacterial loads compared to those in nasal wash samples. We show that nasal wash may underestimate the presence and interactions of Streptococcus pneumoniae within the URT. Furthermore, the URT microbiota composition differs by the source of the mice. This approach to microbial analysis of the murine URT will improve the study design of host-pathogen and pathogen-commensal interactions during colonization.

RESULTS
Nasal wash does not completely disrupt nasal surfaces. The biogeography affected by a nasal wash is unknown. We performed nasal washes using a gentle buffer (phosphate-buffered saline [PBS]) or a harsh buffer (buffer RLT), followed by histological assessment of the nasal tissue via hematoxylin-and-eosin staining ( Fig. 1A and B). PBS does not affect the epithelial architecture of the nasal cavity ( Fig. 1C to E), whereas buffer RLT disrupts the epithelial layer within the nasopharyngeal space and septum as well as the nasal-associated lymphoid tissue ( Fig. 1G and H) but not the nasal turbinates (Fig. 1F). Overall, nasal washes with harsh buffers were unable to disrupt a large majority of the epithelial layer in the nasal cavity, implying that nasal washes do not accurately sample the complete biogeography of the nasal tissue ( Fig. 1A and B).
Nasal tissue extraction recovers more bacteria than nasal washes. Nasal wash and nasal tissue samples were homogenized, plated on brain heart infusion (BHI) agar and fastidious anaerobic agar (FAA), and incubated aerobically and anaerobically, respectively. Overall, the bacterial load was low in both complete nasal tissue (cNT) and nasal wash samples (;10 3 CFU/mouse). Nasal tissue had significantly higher bacterial loads and diversity of colony morphotypes than nasal wash samples ( Fig. 2A). Complementary to culture-based analysis, 16S rRNA gene sequencing revealed that cNT and nasal wash samples clustered separately from each other ( Fig. 2B) (P , 0.05 by permutational multivariate analysis of variance [PERMANOVA]; R 2 = 0.181). Streptococcus species and Staphylococcus species were dominant in nasal tissue and wash samples (Fig. 2C). This difference in microbial composition was driven by nasal tissue containing significantly more Neisseriaceae, Actinomyces, and Bifidobacterium species, while nasal wash samples contained more Erysipelotrichaceae and Cyanobacteria (Fig. 2D) (linear discriminate analysis [LDA] effect size [LEfSe]). No difference was seen in Shannon diversity (P = 0.26) or observed species (P = 0.4181) between nasal wash and nasal tissue samples (see Fig. S1 in the supplemental material).
The mouse nasal microbiota is distinct from the gut microbiota. It has been reported that a significant amount of the bacterial DNA found in the URT is from environmental contamination rather than being due to resident microbes (18). In mice, contaminating DNA can come from exposure to fecal matter and/or reagent controls. To determine the degree to which microbial profiles are influenced by fecal exposure, we compared the nasal tissue and gut microbiota within a mouse. The nasal microbiota was distinct from the gut microbiota ( Fig. S2A) (P , 0.001 by PERMANOVA; R 2 = 0.383). The gut microbiota has greatly increased a-diversity compared to the nasal microbiota (Fig. S2B). Additionally, the dominant families in the gut (Muribaculaceae, Lactobacillaceae, Lachnospiraceae, and Erysipelotrichaceae) are different from the dominant families in the nasal tissue (Streptococcaceae, Staphylococcaceae, and Enterococcaceae) microbiota (Fig. S2C).
The mouse nasal microbiota is distinct from extraction and sequencing controls. Sequencing and extraction controls are essential for low-biomass microbiota analysis as reagents and tissue handling have been shown to influence the community composition of low-biomass communities (19,20). In this study, negative extraction FIG 2 Nasal tissue extraction is essential for complete microbiota assessment. Nasal wash samples were collected using PBS, followed by tissue collection from the same mice and V3 16S rRNA gene high-throughput sequencing. (A) Nasal tissue had significantly more bacteria than nasal wash (NW) samples. (B) Nasal tissue microbial communities clustered separately from nasal wash communities (P , 0.05 by Bray-Curtis PERMANOVA). (C and D) Nasal tissue microbiota were enriched in Neisseriaceae, Bifidobacterium, and Actinomyces, while nasal wash samples were enriched in Erysipelotrichaceae and Cyanobacteria (LEfSe).
(surgical tools dipped in PBS and then exposed to the same extraction process as tissues, cage bedding, and drinking water) and PCR negative (elution water used instead of the DNA template) samples were included to determine the impact of contaminating DNA. The negative samples were distinct from the nasal tissue microbiota ( Fig. S3A) (P , 0.05 by PERMANOVA; R 2 = 0.138). An unweighted pair group method with arithmetic mean (UPGMA) tree based on Bray-Curtis distances demonstrated the separation between negative controls and samples (Fig. S3B). The dominant taxa in the nasal tissue, namely, Streptococcus and Staphylococcus, are not present in the negative samples (Fig. S3C). The inclusion of negative samples with every extraction is still worthwhile to distinguish low-abundance communities from reagent contamination.
Mouse source affects the composition of the nasal microbiota. Differences in the murine gut microbiota between breeding sites have been identified as a source of experimental variability (21). We compared the nasal tissue microbiota from mice bred at McMaster University (n = 32) to those of mice ordered from Jackson Laboratories (JAX) (n = 15). JAX mice have a significantly different nasal microbiota composition compared to in-house-bred mice ( Fig. 3A  Streptococcus pneumoniae colonization leads to domination of the nasal microbiota. Streptococcus pneumoniae colonizes the upper respiratory tract, a process which disrupts preexisting microbial communities prior to causing respiratory and invasive infections (9,10). Investigation into microbial interactions between S. pneumoniae and other nasal microbiota members has been primarily performed by analyzing nasal wash samples, which may overlook microbe-microbe interactions occurring deeper in the nasal tissue. Mice were intranasally colonized with Streptococcus pneumoniae, and nasal wash and tissue samples were collected after 3 days (n = 9 mice). The nasal tissue had higher levels of S. pneumoniae than the paired nasal wash sample (Fig. 4A). The cNT and nasal wash microbiota of colonized mice were distinct (Fig. 3B). S. pneumoniae dominated the nasal tissue ( Fig. 4B and C). LEfSe analysis revealed 24 significantly different genera between the two sampling methods (Fig. S4). Spearman correlations identified that the relative abundance of Streptococcus sequence reads strongly correlated with cultured S. pneumoniae in the cNT (R = 0.9) and nasal wash (R = 0.95) samples (Fig. S5). Furthermore, nasal tissue S. pneumoniae CFU correlated with nasal wash CFU (R = 0.95), and Streptococcus amplicon sequence variant (ASV) relative abundances were correlated between nasal tissue and wash samples (R = 0.88).
Nasal microbiota correlations are dependent on the tissue extraction methodology. Staphylococcus species decrease during Streptococcus pneumoniae colonization when sampled by nasal wash (9,10). Staphylococcus and Streptococcus species were negatively correlated in the nasal wash samples of S. pneumoniae-colonized mice (R = 20.72; P = 0.03) (Fig. 5A); however, there was no significant correlation for the nasal tissue microbiota (R = 20.32; P = 0.41) (Fig. 5B). Corynebacterium species are overrepresented in the nasal microbiota of children and adults negative for pneumococcal colonization and have recently been identified to inhibit the growth of S. pneumoniae in vitro (7). No correlation was found between Corynebacterium and Streptococcus species in the nasal wash microbiota of pneumococcus-colonized mice (R = 0.36; P = 0.35) (Fig. 5C); however, a strong negative correlation (R = 20.95; P = 8.8 Â 10 25 ) between Corynebacterium species and Streptococcus species existed in the nasal tissue (Fig. 5D), suggesting that interactions between these species could occur within regions not affected by nasal washes. These correlations imply that both sampling techniques may be necessary to uncover microbial interactions that are site specific. Furthermore, niche-specific interactions may be driving antagonism or cooperation between S. pneumoniae and other bacterial species in the URT.

DISCUSSION
The upper respiratory tract plays an essential role in breathing, trapping inhaled microbes and particles while heating and humidifying air prior to entering the lungs. The topography provides different niches for bacteria to colonize, resulting in protective or deleterious interactions with other microbial species. This study identified that nasal washes do not completely sample the murine nasal microbiota and may overlook some interactions occurring within the turbinates or submucosa of the nasal tissue. Differences in humidity, mucus secretion, and epithelial cell type could contribute to the colonization success of various pathogens and commensals (22,23). In humans, sampling of different regions of the nasal cavity distinguished the microbial composition of the anterior nares from that of the middle meatus and sphenoethmoidal recess (24). The middle meatus and sphenoethmoidal recess are dominated by Corynebacterium and Staphylococcus species, while anterior nares have a greater abundance of Propionibacterium/Cutibacterium species. In this study, Actinomyces and Neisseria species were increased in the nasal tissue, whereas the nasal wash samples had greater abundances of Erysipelotrichaceae family members. Actinomyces and Neisseria species in the human nasal microbiota have been associated with an increased risk of acute otitis media.
Select bacterial groups, including Corynebacterium, Staphylococcus, Streptococcus, Dolosigranulum, and Moraxella species, are commonly found in the human nose (25,26). The murine nasal microbiota has a dominant Staphylococcus and Streptococcus population but small Corynebacterium, Dolosigranulum, and Moraxella species populations, similar to mice obtained from Jackson Laboratories and other facilities (9). Mice can be experimentally colonized by Corynebacterium and Moraxella species, indicating that the low abundance in specific-pathogen-free (SPF) mice could be due to a lack of exposure (27,28). A major determinant of colonization is adherence, requiring selective interactions between host and bacterial factors. Neisseria meningitidis and Streptococcus pyogenes colonization in mice requires the expression of human-specific adherence factors in the olfactory epithelium (12,29). The composition of the microbiota differed between mice bred in-house and those delivered by Jackson Laboratories. Vendor-specific microbiota differences have been identified in gut and lung studies and have been implicated in altering the outcomes of disease models (18,30,31); however, our study is the first to demonstrate vendorspecific microbiome differences within the nasal tissue. Housing under SPF conditions at different facilities restricts access to many bacteria beyond pathogens, and SPF mice have drastically different respiratory microbiota compositions compared to wild mice (32). Indeed, mice captured from the wild have more, as well as different, bacteria in their lungs, which influences alveolar structure (32). The impact of differential colonization between facilities and sources may impact the baseline physiology or expression of some genes. Krone et al. previously assessed the murine nasal microbiota by nasal wash and found patterns similar to those in our mice, with high levels of Staphylococcaceae, Streptococcaceae, and Erysipelotrichaceae, but failed to detect Actinomycetaceae (9), which were detected in the nasal tissue samples in this study. Interestingly, Weyrich et al. found high levels of Actinomycetaceae (as well as Streptococcaceae and Staphylococcaceae) in nasal tissue samples from their facility (33). Whether these differences are due to different breeding facilities, housing conditions, or sampling techniques is unclear.
Sequencing-based analysis of the URT microbiota is complicated due to the low microbial biomass compared to host DNA (34). Extraction and amplification methods will affect the outcome of sequencing results, and the inclusion of template negative controls is essential for determination of the microbial composition. We demonstrated that the nasal tissue microbiota is distinct from the gut microbiota and negative controls. Furthermore, there was a strong correlation between the cultivable amount of inoculated S. pneumoniae and the ability to detect Streptococcus ASVs in the nasal tissue and wash samples. This strongly suggests that culture-independent analysis of the nasal tissue microbiota is representative of the cultivable microbiota and not contaminating DNA sequences from environmental or extraction sources.
Colonization is essential during bacterial pathogenesis, including adherence prior to invasion or spread to new hosts. Capsule expression by Streptococcus pneumoniae varies depending on nasal location, which alters its ability to adhere, evade killing, or be transmitted to a new host (35)(36)(37)(38). As such, the detection of S. pneumoniae in nasal wash or tissue extraction samples has different implications for transmission versus invasive disease. Multiple studies have reported a negative correlation between Staphylococcus and Streptococcus species, potentially due to hydrogen peroxide production (39,40). We have shown that this negative correlation exists in the nasopharynx of colonized mice but not in the nasal tissue. Conversely, we have shown a negative correlation between Streptococcus and Corynebacterium species in the nasal tissue but not in the nasal wash samples. Corynebacterium species have been demonstrated to liberate host triacylglycerols that kill S. pneumoniae (7). Previous studies have also implicated differences in Haemophilus influenzae and S. pneumoniae antagonism depending on the sampling method (15,16). Together, these data suggest that multiple sampling methods may be necessary to determine bacterial interactions within the URT. Human experimental colonization models assess pneumococcal colonization, as well as microbial communities, via nasal wash (41) and identify many antagonistic interactions that have also been seen in murine models of experimental colonization (e.g., negative correlation of Corynebacterium and Streptococcus [41,42]); however, whether there are microbial interactions missed by this sampling method is unclear. The structure of murine nasal tissue is much more complicated than that of human nasal tissue, so species-specific sampling techniques may be needed to study microbial interactions.
Overall, mechanistic investigation of the URT microbiota in infection and immunity requires animal models. Current studies have used mixed methodologies, including different sampling techniques and mouse vendors, to investigate development and disease phenotypes. Proper extraction and assessment of the nasal tissue of mice are essential to reveal reproducible, mechanistic interactions between host cells and microbial members. Our findings demonstrate that assessment of the nasal microbiota is dependent on biogeography and needs to be integrated into research design for evaluation of pathogen and commensal colonization and interactions.

MATERIALS AND METHODS
Animals. C57Bl/6J mice were bred within the McMaster Central Animal Facility, except for the experiments in which we used female C57Bl/6J mice from Jackson Laboratories (Bar Harbor, ME). Mice from Jackson Labs were 6 to 8 weeks old and acclimated to specific pathogen-free conditions for 2 weeks prior to experiments. All mice had access to food and water ad libitum. All mice used in this study were female mice aged 8 to 12 weeks. Mice were anesthetized using isoflurane and euthanized by exsanguination. All experiments were approved by McMaster University's Animal Research Ethics Board according to the recommendations of the Canadian Council for Animal Care.
Tissue collection. Nasal wash was completed as previously described (43). Briefly, a PE-20 polyethylene tube attached to a 26-gauge needle was inserted through a small incision in the trachea. Lavages were performed with 300 ml of sterilized phosphate-buffered saline (PBS) or buffer RLT (Qiagen), which was flushed through the trachea and collected through the nares in a 1.7-ml microtube. Complete nasal tissue (cNT) was collected from lavaged or naive mice via bisection of the skull with sterilized surgical tools (44). Excised tissues were homogenized in 300 ml PBS in 2-ml screw-top tubes using 2.8-mm ceramic beads for 1 min at 2,000 rpm (MoBio).
Histological analysis. After nasal washes, mouse heads were placed into formalin for 24 h before being placed in a Shandon TBD-2 decalcifier (Thermo Scientific, Kalamazoo, MI) for 4 days. Decalcified heads were placed in formalin for 2 days, followed by twice-daily washes with PBS for an additional 4 days. Samples were washed with and placed in 70% ethanol prior to standard histological processing and embedded in paraffin wax. Cross-sectional slices (5 mm) were mounted on slides and stained with hematoxylin and eosin according to standard protocols. Slides were randomized and scored in a blind fashion. The integrity of the epithelial lining was measured by quantifying the area of intact/disrupted epithelium using ImageJ.
DNA extraction, amplification, and analysis of the 16S rRNA gene. DNA was extracted from PBS nasal wash and cNT homogenates as previously described (10). Samples were mechanically homogenized with 0.2 g of 0.1-mm glass beads and 0.2 g of 2.8-mm glass beads in 800 ml of 200 mM NaPO 4 (pH 8) and 100 ml of guanidine thiocyanate-EDTA-N-lauroyl sarcosine. After homogenization, the sample was incubated with 50 ml of lysozyme (100 mg/ml) and 10 ml of RNase A (10 mg/ml) for 1 h at 37°C, followed by incubation with a solution containing 25 ml sodium dodecyl sulfate (25%), 62.5 ml NaCl (5 M), and 25 ml proteinase K (20 mg/ml) for 1 h at 65°C. Samples were centrifuged for 5 min at maximum speed, and the supernatant was transferred to 900 ml of buffered phenol-chloroform-isoamyl alcohol (25:24:1). Samples were vortexed and centrifuged for 10 min at maximum speed prior to transferring the aqueous phase to DNA Clean and Concentrator-25 columns (Zymo) according to the manufacturer's instructions, except that samples were eluted with 50 ml ultrapure water. Negative controls included PBS-exposed surgical tools, cage bedding, drinking water, and no-template PCR amplicons.
PCR amplification of the 16S rRNA gene (V3 region) involved a two-step, nested PCR, which improves efficiency in the presence of high host DNA levels (45). The first step involved the amplification of the 16S rRNA gene region spanning V1 to V5 using universal primers 8F (AGAGTTTGATCCTGGCTCAG) and 926R (CCGTCAATTCCTTTRAGTTT) for 15 cycles (94°C for 30 s, 56°C for 30 s, and 72°C for 60 s). This product was used as the template in the second reaction for 25 cycles (94°C for 30 s, 47°C for 30 s, and 72°C for 40 s) to amplify the V3 region of the 16S rRNA gene in preparation for MiSeq (Illumina) sequencing. Barcoded primer sequences were adapted similarly to previous work (46). All reactions, including extraction negative and no-template controls, were performed in triplicate to reduce PCR bias, and reaction mixtures were pooled prior to sequencing using the MiSeq sequencing platform (Illumina, Inc., San Diego, CA) at the Farncombe Genomics Facility at McMaster University. Sequences were trimmed using CutAdapt (47) prior to analysis with DADA2 (48) to organize sequences into amplicon sequence variants (ASVs). Taxonomy was assigned using the Silva database (49). Relative abundance, aand b-diversity, and rarefactions were completed using the Phyloseq package (50) in R version 3.5.2 (51). A dendrogram was constructed based on unweighted pair group method with arithmetic mean (UPGMA) hierarchical clustering using Bray-Curtis distances using the phangorn package (52) and plotted with iTOL (53).
Differences between bacterial communities were tested using permutational multivariate analysis of variance (PERMANOVA) within the vegan package (54). Differences in taxon abundances were assessed using linear discriminate analysis (LDA) effect size (LEfSe) (55).
Streptococcus pneumoniae colonization. Mice were intranasally colonized with S. pneumoniae strain P1547 (serotype 6A), obtained from Jeff Weiser (NYU School of Medicine), as previously described (43). S. pneumoniae was grown in tryptic soy broth in a 5% CO 2 incubator at 37°C until cultures were in late log phase (optical density at 600 nm [OD 600 ] = 0.5). Cultures were spun down at 15,000 Â g for 1 min and resuspended at a concentration of 10 9 CFU/ml in PBS. Mice were colonized by depositing 10 ml containing 10 7 CFU of S. pneumoniae directly in the nares. Mice were sacrificed at 3 days postinoculation prior to PBS nasal wash and cNT sample collection.
Bacterial culture. PBS nasal wash and tissue homogenates were incubated overnight on brain heart infusion agar or tryptic soy agar supplemented with 5% sheep's blood and neomycin (10mg/ml) or on prereduced fastidious anaerobic agar in an anaerobic chamber. Colonies were counted and collected via the addition of 1 ml BHI medium and scraping the plate surface with a cell scraper. A portion of the collected colonies was frozen at 280°C in 10% glycerol, while the remainder was prepared for genomic extraction.
Statistical analysis. GraphPad Prism 6 and R were used for statistical analysis. Nonparametric tests were used for comparison of histology scoring and bacterial plate counts. Differences with P values of ,0.05 were considered statistically significant. Microbial community composition differences were determined using the Adonis function (PERMANOVA) with 10,000 permutations.
Data availability. Data related to this study have been deposited in the NCBI database under BioProject accession number PRJNA679949.

SUPPLEMENTAL MATERIAL
Supplemental material is available online only.