Age-dependent nasal immune responses in non-hospitalized bronchiolitis children

Abstract

Bronchiolitis in children is associated with significant rates of morbidity and mortality. Many studies have been performed using samples from hospitalized bronchiolitis patients, but little is known about the immunological responses from infants suffering from mild/moderate bronchiolitis that do not require hospitalization. We have studied a collection of nasal lavage fluid (NLF) samples from outpatient bronchiolitis children as a novel strategy to unravel local humoral and cellular responses, which are not fully characterized. The children were age-stratified in three groups, two of them (GI under 2-months, GII between 2-4 months) presenting a first episode of bronchiolitis, and GIII (between 4 months and 2 years) with recurrent respiratory infections. Here we show that elevated levels of pro-inflammatory cytokines (IL1β, IL6, TNFα, IL18, IL23), regulatory cytokines (IL10, IL17A) and IFNγ were found in the three bronchiolitis cohorts. However, little or no change was observed for IL33 and MCP1, at difference to previous results from bronchiolitis hospitalized patients. Furthermore, our results show a tendency to IL1β, IL6, IL18 and TNFα increased levels in children with mild pattern of symptom severity and in those in which non RSV respiratory virus were detected compared to RSV+ samples. By contrast, no such differences were found based on gender distribution. Bronchiolitis NLFs contained more IgM, IgG1, IgG3 IgG4 and IgA than NLF from their age-matched healthy controls. NLF from bronchiolitis children predominantly contained neutrophils, and also low frequency of monocytes and few CD4⁺ and CD8⁺ T cells. NLF from infants older than 4-months contained more intermediate monocytes and B cell subsets, including naïve and memory cells. BCR repertoire analysis of NLF samples showed a biased VH1 usage in IgM repertoires, with low levels of somatic hypermutation. Strikingly, algorithmic studies of the mutation profiles, denoted antigenic selection on IgA-NLF repertoires. Our results support the use of NLF samples to analyze immune responses and may have therapeutic implications.

Keywords: B lymphocytes; RSV; bronchiolitis; cytokines; immunoglobulins; monocytes; nasal lavage fluid (NLF); neutrophils.

Figure 1

Scheme summarizing the study design. NLF samples were obtained from the indicated groups. Supernatants were used for RT-PCR, CBA and ELISA analysis to determine the viral content, cytokine profile and antibody composition, respectively. The cells recovered from NLF from bronchiolitis patients were used for immunoprofiling by using flow cytometry and for BCR IgH repertoire analyses after RNA extraction, cDNA preparation, amplification, barcoding and NGS sequencing.

Figure 2

Bronchiolitis NLF samples contain elevated levels of cytokines. Cytometric bead array (CBA) determination of a panel of cytokines in the NLF supernatants from controls and bronchiolitis patients: GI (controls n = 12, bronchiolitis n = 16), GII (controls n = 16, bronchiolitis n = 18), and GIII (controls n = 18 bronchiolitis n =19). The scatter dot plots represent the individual values, and the mean ± SEM is depicted for each group. Statistical analyses were performed using non-parametric Mann-Whitney sum rank test or the unpaired t-test with Welch´s correction. (A) Quantification of the pro-inflammatory cytokines IL1β, IL6, TNFα, IL18, IL23 and IL33. (B) Determination of IL10 and IL17A, and (C) IFNγ and IFNα2. (D) Analysis of IL1β, IL6, TNFα, IL18 and IL10 in samples according to the severity of bronchiolitis (based on the Tal modified score), mild (n = 37) or moderate (n = 16) and the presence/absence of respiratory viruses: absence of virus, (Vir-, n = 24), presence of RSV (RSV+, 15 RSV alone and 3 coinfections, n = 18), presence of other virus (OtherV+, n = 11, including 9 single infections and 2 coinfections without RSV). *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001.

Figure 3

Bronchiolitis NLF samples contain elevated levels of Immunoglobulins. IgS were quantified in NLF supernatants from controls and bronchiolitis patients. The scatter dot plots represent individual values, and the mean ± SEM is depicted for each group. Statistical analyses were performed using non-parametric Mann-Whitney sum rank test. (A) IgM, IgA and IgG were determined by ELISA GI (controls n = 12, bronchiolitis n = 14), GII (controls n = 15, bronchiolitis n = 12), and GIII (controls n = 15 bronchiolitis n =14). (B) Cytometric bead array determination (CBA) for IgG1, IgG2, IgG3 and IgG4. GI (controls n = 10, bronchiolitis n = 11), GII (controls n = 14, bronchiolitis n = 9), and GIII (controls n = 14 bronchiolitis n = 10). (C) Ig isotype quantification based on mild (n = 29) and moderate (n = 11) bronchiolitis severity, and on the presence/absence of respiratory viruses (absence of viruses, n = 17; presence of viruses, n = 19). *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001.

Figure 4

Presence of myeloid cells in NLF from bronchiolitis patients. Flow cytometry was performed on NLF bronchiolitis and control samples (n = 39 and n = 46, respectively) stained with anti-CD45, anti-CD16 and anti-CD14 MonAbs. Quantification of the myeloid cell subpopulations was performed for each group analyzed and is shown in the figure as relative (% of CD45⁺ cells). Each dot represents an individual NLF sample, also showing the mean ± SEM. (A) Left, representative dot plots showing the FSC-A and SSC-A in a patient sample (up) and a control (down) NLF sample after excluding the cell dump channels, doublets and dead cells (see Figure S2A ). Middle, representative histograms of the CD45 staining to identify hematopoietic cells: Shown inside are the relative number (%, mean ± SEM, n =39) of CD45⁺ cells in bronchiolitis NLF. Control samples contained less than 10³ CD45^- cells and therefore they are not shown in the following panels. Bronchiolitis samples contained 2 x 10⁶ ± 0.6 live cells (GI), 2.73 x 10⁶ ± 0.7 live cells (GII) and 11.4 x 10⁶ ± 2.6 live cells (GIII). Right, the plot shows the quantification of CD45⁺ cells in the NLF from GI (n = 15), GII (n = 11) and GIII (n = 13). (B) Upper left, representative dot-plot of gated CD45⁺ cells analyzed with anti-CD14 and anti-CD16 to discriminate neutrophils (Nϕ), classical monocytes (cMon, CD14⁺⁺CD16^-), intermediate monocytes (inMon, CD14⁺⁺CD16⁺) and lymphocytes (Lymph). Upper right, representative overlaid contour plot of gated CD45⁺ cells depicting the SSC-A and FSC-A analysis for neutrophils (Nϕ, black), monocytes (Mon, brown) and Lymphocytes (Lymph, blue). Bottom, determination of the SSC-A of electronically gated Nϕ and Mon in NLF samples. (C) Relative numbers (to CD45+ cells)of Nϕ, Mon and Lymph in each group as in panel (B). (D) Frequency of cMon and inMon among CD45⁺ monocytes, and the cMon/inMon ratio (the latter as GI+II and GIII, n = 26 and n = 13, respectively). Comparisons were performed using non-parametric Mann-Whitney sum rank test. *p < 0.05; **p < 0.01.

Figure 5

Age-dependent increase of T and B lymphocytes in NLF samples from bronchiolitis patients. Flow cytometry was performed on NLF bronchiolitis samples (n = 38) stained with anti-CD4, anti-CD8, anti-CD19 and anti-CD20 MonAbs. Quantifications were done on the lymphoid gate described in Figure 4 . Frequency of subpopulations relative to CD45⁺ cells in this gate are shown. (A) Representative contour plots from GI-GIII NLF samples displaying the staining with anti-CD8 and anti-CD4 by flow cytometry. The dotted rectangles inside the plots indicate CD4⁺ and CD8⁺ T lymphocytes, and the numbers are the %, mean ± SEM of CD4⁺ and CD8⁺ cells. (B) Frequency of CD4⁺ and CD8⁺ cells in all NLF samples analyzed. (C) Frequency of CD4⁺ and CD8⁺ cells in NLF samples from GI (n = 15), GII (n = 10) and GIII (n = 13) children. (D) Representative contour plots displaying the staining with anti-CD20 and anti-CD19 to identify CD19⁺CD20⁺ B lymphocytes by flow cytometry. Relative numbers of B cells (%, mean ± SEM) are indicated inside the plots. (E) Frequency of B cell numbers in NLF samples from GI (n = 12), GII (n = 11) and GIII (n = 13). (F) Relative expression of PAX5 determined by RT-qPCR in NLF samples from GI (n = 9), GII (n = 8), GIII (n = 12). The amount of Pax5 transcripts was calculated as the 2^−ΔΔCT relative to that of GAPDH and relative to adult PBMC values, performed in duplicates. (G) Linear regression of Pax5 2^−ΔΔCT data and the number of CD20⁺ cells in NLF samples. In the graphs from panels (B, C and E–G) each dot represents an individual sample and the mean ± SEM is also shown. Comparisons among groups were done using non-parametric Mann-Whitney sum rank test. *p < 0.05; **p < 0.01; ****p < 0.0001.

Figure 6

B cell phenotype in NLF from bronchiolitis affected infants. Cell suspensions were stained with the following MonAbs: anti-CD20, anti-CD19, anti-CD27, anti-IgD, anti-CD70, anti-CD43 and anti-CD5. Shown are contour plots of the indicated staining performed by using concatenated files from adult PBMC (n = 3) and from NLF samples (n = 5). (A) Contour plots of the CD27 and IgD staining on gated CD20⁺ B cells. The quadrants discriminate the B cell populations of naïve(IgD⁺CD27^- Bn), DN (IgD^-CD27^-) and memory (CD27⁺ Bm, IgD^- and IgD⁺). (B) Frequency of each B cell subset relative (%) to the total CD20⁺ cells on individual NLF and PBMC samples. The data are presented as scatter dot plots with the means ± SEM shown; GI (n = 7), GII (n = 7), GIII (n = 10), adult PBMC (n = 3). For comparison, the frequency obtained for adult PBMCs (Sanz et al., filled triangle) and that of PBMCs from 1-24 mo-old children (Berrón-Ruiz et al., empty triangle) are shown. (C) Contour plots representing CD5 versus FSC-A on gated CD19⁺CD27⁺CD70^-CD43⁺ B1 cells. (D) Frequency of CD5^- and CD5⁺ B1 cell subsets as the % of the total CD19⁺ cells. Filled square shows the frequency obtained on adult PBMCs (Rodriguez-Zhurbenko et al.). Data are represented as in (B); GI (n = 6), GII (n = 7), GIII (n = 11), PBMC (n = 3). Data were compared using non-parametric Mann-Whitney sum rank test. *p < 0.05; **p < 0.01; ***p < 0.001.

Figure 7

BCR repertoire analysis in NLF. The NLF repertoires of VDJ- IgM (n = 9), -IgG (n = 4) and IgA (n = 4) rearrangements were determined after RT-PCR-specific amplifications and sequencing by NGS. The obtained sequences were processed, cleaned and analyzed as indicated in Materials and Methods (“Bioinformatics” section). (A) VH usage frequency on sequences from each individual repertoire of NLF samples. Shown are the means ± SEM. See Table S2 for sequence numbers. (B) The diversity of the repertoires was calculated as the ratio between unique sequences and functional sequences. Each dot represents the value for the repertoire of all sequences obtained in the same sample, and overlaid are box-and-whisker plots showing the median, the first quartile to the third quartile and the minimum and maximum values. (C) VH mutations in IgS repertoires. Data are represented as in (B). (D) The ratio of replacement (R) and silent (S) mutations located in the CDRs and FWRs is shown as mean ± SEM. (E) Frequency of SHM hotspots in the AID motifs WRCY and RGYW. Data are shown as mean ± SEM. (F) Antigen selection strength (quantified using the BASELINe algorithm). Data are represented as in (B). (G) Clonal tree obtained for the VH3-23-JH4 family clone identifying IgM (n = 198), IgG (n = 113) and IgA (n = 46) sequences, from a single NLF sample. The analysis was performed based on a minimal substitution model using MUSCLE software, with alignment curation using Gblocks and tree interference with PhyML. Data were compared using an unpaired two tailed Student´s t-test. *p < 0.05; **p < 0.01.

Figure 8

Scheme of the proposed age-dependent nasopharynx-associated lymphoid tissue (NALT) responses in children with mild/moderate bronchiolitis pathology. The cytokine, immunoglobulin and cellular content in NLF samples varies with the age of the children affected with bronchiolitis. (A) Pro-inflammatory and regulatory cytokines are present in all NLF bronchiolitis samples at higher levels than in these from controls. (B) Heterogeneous combinations of CD45⁺ innate and adaptive cell types are detected in NLF from patients and not in controls. Neutrophils (Nϕs) are predominant in all NLF samples groups, which also contain classical and intermediate monocytes (cMon and inMon) and lymphocytes, these two latter increasing in children older than 2 months (GII/GIII), in which activated Nϕs (actNϕs) with increased complexity and cMon and inMon, T cells (CD4⁺ and CD8⁺) and CD19⁺ cells were found. (C) IgM, IgGs and IgA levels are present at higher amounts than in age-matched controls. In addition, IgG levels highly augment in GIII infants. (D) Local B cell repertoires were determined by analyzing the presence of VH rearrangements in NLF samples, which displayed IgA- sequences with low mutations but antigen selection in GII/GIII infants. Created with BioRender.