Airway response to respiratory syncytial virus has incidental antibacterial effects

Abstract

RSV infection is typically associated with secondary bacterial infection. We hypothesise that the local airway immune response to RSV has incidental antibacterial effects. Using coordinated proteomics and metagenomics analysis we simultaneously analysed the microbiota and proteomes of the upper airway and determined direct antibacterial activity in airway secretions of RSV-infected children. Here, we report that the airway abundance of Streptococcus was higher in samples collected at the time of RSV infection compared with samples collected one month later. RSV infection is associated with neutrophil influx into the airway and degranulation and is marked by overexpression of proteins with known antibacterial activity including BPI, EPX, MPO and AZU1. Airway secretions of children infected with RSV, have significantly greater antibacterial activity compared to RSV-negative controls. This RSV-associated, neutrophil-mediated antibacterial response in the airway appears to act as a regulatory mechanism that modulates bacterial growth in the airways of RSV-infected children.

Fig. 1

The composition of the airway microbiota. a Dendrogram visualising the hierarchical clustering of individual microbiomes of 84 individual children on the basis of the Bray–Curtis dissimilarity matrix. The relative abundance of the 15 highest-ranked operational taxonomic units are presented for each study participant as stacked bar charts. operational taxonomic units under the ‘others’ heading are an aggregate of bacteria that occurred at low abundances (<1%). The taxonomic identities of these low-abundance operational taxonomic units are presented in Supplementary Data 1. b The proteome of the upper airway microbiota (metaproteome) was characterised by mass spectrometric analysis using airway samples from 84 children and the results are presented in a circular bar chart. Each bar represents the mean expression level of one bacterial protein. Values on the y-axis correspond to MS protein expression levels (log₁₀ reporter corrected intensity). Most of the bacterial operational taxonomic units that were identified in metagenomic analyses of the airway microbiota were also identified in the proteome. The highest number of protein identifications belonged to the genus Streptococcus and comprised mainly of ribosomal proteins and proteins involved in energy metabolism pathways. c The abundance of Streptococcus in the airway was compared between RSV-positive and RSV-negative children. RSV-infected children had a higher abundance of airway Streptococcus compared to RSV-negative children. d The effect of RSV infection on Streptococcal colonisation of the upper airway was characterised using a subset of 10 children from whom nasopharyngeal and oropharyngeal swabs were obtained during acute RSV infection and at convalescence, approximately one month later. Seven out of 10 children (70%) exhibited declines in the airway abundance of Streptococcus upon recovery from RSV, while for three children, convalescent-stage Streptococcus abundance was either unchanged from the acute time point or was higher. On the box and whisker plot, the bottom line on the box denotes the 25th data percentile (quartile 1), the middle line denotes the median and the upper line denotes the 75th data percentile (quartile 3). The bottom whisker represents data in the lower extremity of the distribution (quartile 1–1.5 × interquartile range) and upper whisker denotes the upper extremity of the distribution (quartile 3 + 1.5 × interquartile range). Source data are provided as a Source Data file

Fig. 2

Analysis of the airway proteome by tandem mass spectrometry. Analysis of airway proteomes was undertaken using mass spectrometry in a population of 40 children with RSV infection and 44 RSV-negative children. a Differences in the mean expression levels of local airway proteins between the RSV-positive and RSV-negative children was analysed using two-sided Student's t-tests and presented in a volcano plot. The y-axis represents −log₁₀ fdr-adjusted P values (q values) computed using two-sided Student's t-tests and the x-axis represents −log₁₀ fold change in protein expression. Red circles represent differentially expressed proteins (defined by q values: ≤0.05 and ≥1.5-fold change in expression between groups). Red circles in the top left and top right quadrants represent proteins that were significantly elevated in RSV-positive and RSV-negative children, respectively. b, c Gene ontology (GO)-based biological pathways analysis was used to identify biological processes that were significantly overrepresented in RSV-positive and RSV-negative children. Those that were significantly overrepresented in RSV-positive children are shown in b while those that were overrepresented in RSV-negative children are shown in c. Adjusted P-values in b and c were calculated on the enrichR platform and are based on the Fishers extract test. The most significantly enriched pathway in RSV-positive children was neutrophil degranulation and in RSV-negative children was SRP-dependent co-translational protein targeting to membrane, signal sequence recognition. d, e The specific proteins that were significantly differentially expressed within those pathways are shown in d and e, respectively. The proteins highlighted with the red boxes in d are neutrophil granule proteins. The elements of the box and whisker plots are defined in Fig. 1. Source data are provided as a Source Data file

Fig. 3

Analysis of bacterial inhibition activity in airway secretions. a Example plots of children with different bacterial inhibition activities. The y-axis represents the optical density which is directly related to the amount of bacterial (Streptococcus pneumoniae) growth: bacterial growth increases turbidity of the culture media resulting in a proportional increase in optical density. The x-axis represents a doubling dilution of nasal samples from 1:20 to 1:40,960. The dashed horizontal red line denotes an optical density cutoff above which substantial turbidity of the culture media was observed. Nasal secretions from infant 1 were unable to inhibit bacterial growth at any dilution (bacterial inhibition index [BII] = 0). Nasal secretions from infant 2 completely inhibited pneumococcal growth at dilutions ≤ 1:160 (BII = 160). b Children were categorised into five BII strata and differences in the proportional abundance of different bacterial taxa in each stratum presented in pie charts. The airway abundance of Streptococcus decreased with increasing bacterial inhibition activity. c Bacterial inhibition activity in nasal samples (N = 72) was correlated with the expression levels of neutrophil-granule proteins (stratified by functional subset). Representative proteins from all three subsets of neutrophil granules (azurophilic, specific and gelatinase) were positively correlated with bacterial inhibition activity (complete analysis of the correlation between bacterial inhibition activity and all the neutrophil granule proteins can be found in Supplementary Fig. 2). In contrast, there was no correlation between the expression levels of housekeeping proteins (ACTB, B2M and PPIA) and bacterial inhibition activity in nasal samples. Correlation analysis was done using Spearman’s rank order correlation. d Bacterial inhibition activity was compared between RSV-positive (N = 36) and RSV-negative children (N = 36). The mean BII level in RSV-infected children (geometric mean: 52) was significantly greater than the mean BII in RSV-negative children (geometric mean: 10.9). Error bars indicate 95% confidence intervals. e Mean airway IgG levels to two Streptococcus pneumoniae proteins (choline-binding protein D and Histidine Triad protein E) were compared between RSV-positive (N = 36) and RSV-negative children (N = 36). Error bars indicate 95% confidence intervals. P-values in d and e were based on Mann–Whitney U-test. Source data are provided as a Source Data file

Fig. 4

Analysis of the cellular sources of the airway proteome by flow cytometry. a Cells obtained from nasopharyngeal and oropharyngeal swabs (both eluted in universal transport media) from a child with an RSV infection was visualised at ×200 magnification. Multiple cell morphologies including epithelial cells were observed. b Upper airway sample obtained from an infant admitted to hospital with RSV-pneumonia was analysed using a direct immunofluorescent antibody test (Light Diagnostics, Merc). RSV-infected cells are shown emitting an apple-green fluorescent signal, while uninfected cells emit a red fluorescent signal from a counterstain (Evans blue). Image taken using Nikon Eclipse fluorescent microscope image acquisition software. c Flow cytometry analysis was used to identify neutrophils in the sample described in b above (details of full gating strategy can be found in the “Methods” section and in Fig. S5). Cells that were double positive for CD66b and CD16 were gated as neutrophils. d Phagocytic activity of airway-resident neutrophils shown in c above was analysed using flow cytometry. Cells were co-incubated with an E. coli strain that was labelled with a dye (pHrodo) whose florescence was only activated upon phagocytosis. In this example, 17.8% of the neutrophils in the sample exhibited bacterial phagocytosis. e, f A similar example is shown from a different child who had fewer neutrophils and also had reduced phagocytic activity. g The frequency of airway-resident neutrophils in RSV-positive and RSV-negative children is shown as a proportion of all airway cells. The proportion of airway cells that were identified as neutrophils was greater in RSV-positive children (N = 10) compared to RSV-negative children (N = 10) although the difference did not reach statistical significance. Error bars indicate 95% confidence intervals about the geometric mean. h The capacity of airway neutrophils from RSV-positive (N = 10) and RSV-negative (N = 10) children to phagocytose bacteria was compared. A greater proportion of neutrophils from RSV-positive children exhibited phagocytic activity compared to RSV-negative children, although this difference did not achieve statistical significance. Error bars indicate 95% confidence intervals about the geometric mean. Comparison between the RSV-positive and RSV-negative groups in g and h was conducted using the Mann–Whitney U-test. Source data for g and h are provided as a Source Data file

Fig. 5

Association between airway neutrophil granule expression and disease severity. In order to characterise the clinical implications of the increased expression of neutrophil-associated proteins on disease pathology, we assessed the association between the level of expression of these proteins and a cardinal feature of RSV pathology, blood oxygen saturation. a Using proteome data obtained from all admitted children (n = 84), we stratified the expression levels of nine neutrophil granule proteins (ITGAM, BPI, LCN2, MGAM, ITGB2, AZU1, ANPEP, MMP9 and MPO) that were significantly over-expressed in RSV-infected children by the median. Children whose protein expression level was above the population median were classified as high-expressers while those below it were classified as low expressers (see legend for graphical illustration of the stratification strategy). For all proteins analysed, children in the high expression group had significantly lower blood oxygen saturation levels (measured using fingertip pulse oximetry) compared to children in the low expression group. b For comparison, we used the same median stratification strategy to dichotomise children into high and low expression groups for two housekeeping control proteins, beta-actin (ACTB) and beta-2-microglobulin (B2M). For both proteins there was no statistically significant difference in blood oxygen saturation levels between high and low expressers. The dashed horizontal line indicates an oxygen saturation threshold of 90%, below which children would typically be characterised as functionally hypoxic. Statistical comparison between high and low strata was done using a two-sided Student t-test. The elements of the box and whisker plots are defined in Fig. 1. Source data are provided as a Source Data file