Single Cell Analysis Identifies Distinct Profiles in Pediatric Acute Respiratory Distress Syndrome

doi:10.21203/rs.3.rs-1992367/v1

Download PDF

Article

Single Cell Analysis Identifies Distinct Profiles in Pediatric Acute Respiratory Distress Syndrome

https://doi.org/10.21203/rs.3.rs-1992367/v1

This work is licensed under a CC BY 4.0 License

Journal Publication

published 30 Jun, 2023

Read the published version in Nature Communications →

Version 1

posted

You are reading this latest preprint version

Acute respiratory distress syndrome (ARDS), termed pediatric ARDS (pARDS) in children, is a severe form of acute respiratory failure (ARF) that is associated with significant morbidity and mortality. Pathologic immune responses have been implicated in the development of pARDS. Here, we present a unique description of microbial sequencing and single cell gene expression in tracheal aspirates (TAs) obtained longitudinally from infants admitted to the pediatric intensive care unit (PICU) with ARF. Important alterations in TA epithelial cell, mononuclear phagocyte (MNP), and neutrophil transcription were associated with subjects’ illness severity, etiology (of ARF), and sampling time point. Specifically, in patients with moderate to severe pARDS compared to those with no to mild pARDS we identified reduced interferon stimulated gene (ISG) and cytokine expression in epithelial cells, reduced development over time of a regulatory IL-10 expressing macrophage population, and progressive airway neutrophilia associated with a unique transcriptional profile in aged neutrophils. In addition, we incorporated viral capture sequencing (ViroCap) with single cell transcriptomic analysis to explore interactions between respiratory syncytial virus (RSV) and host cells in infected and bystander epithelia. Our findings indicate that pARDS is defined by distinct inflammatory cell profiles that are etiology- and severity-dependent and implicate inadequate induction of ISGs, cytokines/chemokines, and repair-associated macrophage transcriptional programs in the pathogenesis of moderate to severe pARDS.

There is NO Competing Interest.

TableS1.xlsx
Table S1
TableS2.xlsx
Table S2
TableS3.csv
Table S3
S1.jpg
Supplementary Figure 1: Additional single cell analyses of aggregate data set. a, UMAP of all cells included in the final data set after initial unsupervised clustering analysis, labeled by cluster number. b, Dot plot of expression levels of RSV transcripts and IAV gene segments captured by single cell transcriptomics. The size of the dot indicates percent of cells within an individual cluster expressing a particular transcript, and the color indicates the average expression of the transcript within the cell population. c, Feature plots of percent of cellular transcriptome occupied by RSV (top) and mitochondrial transcripts (bottom). d, Pie charts of proportion of cells at each sample time point from subjects in each group. e, UMAPs of all cells in data set split by subject group and labeled by cell type as in Figure 2a. Cell numbers indicate total cell numbers from the patient group contributing to the aggregate data set (n=1 Sample for Control, n=2 Samples for RSVposL, and n=3 Samples for RSVposP and RSVneg). f, stacked violin plots of expression of select transcripts differentially upregulated by cells from a particular subject group (green (VCAN, RETN) indicates upregulated in RSVneg cells, red (SLPI, PI3) indicates upregulated in RSVposP cells, and yellow (IFI6, CCL3, and IL6) indicates upregulated in RSVposL cells) grouped by cell type and split by subject group. Statistical comparisons between groups within cell types were made using Wilcoxon rank sum testing with p-value adjustment using the Bonferroni correction. Asterisk indicate significant (threshold of padj < 0.05) upregulation of a transcript relative to other groups amongst cells of a particular type.
S2.jpg
Supplementary Figure 2: Additional analyses of host-pathogen interactions in single cell and ViroCap data. a, Dot plot of differentially expressed genes that define individual cell populations within subclustered epithelial cells. Cell populations are labeled as in Figure 3a. The size of the dot indicates percent of cells within an individual cluster expressing a particular transcript, and the color indicates the average expression of the transcript within the cell population. b, UMAP of subclustered epithelial cells colored according to inferred differentiation trajectory by pseudotime (Monocle326). Differentiation trajectories are overlaid on the UMAP and labeled as discussed in the text. c. Bar plot of proportional abundance of individual subclustered epithelial cell populations within RSV infected and bystander epithelial cell populations. d. UMAPs split by PTPRC expression of subsetted epithelial cell populations that were a mixed population with PTPRC expressing cells (“Immune”, “Immune_HES1”, and “Epithelial_MT_high”) (top) and volcano plot demonstrating DEGs between RSVposP and RSVposL group cells. DGE testing was performed using Seurat’s FindAllMarkers in default settings (Wilcoxon rank sum test with Bonferroni p-value adjustment) comparing RSVposP and RSVposL PTPRC positive (bottom, left) and RSVposP and RSVposL PTPRC negative (bottom, right) cell populations. In both plots a positive avg_log2FC indicates increased expression by RSVposP cells, and a negative avg_log2FC indicates increased expression by RSVposL cells. e. Box plot of RSV read counts in RSVposP and RSVposL TA supernatant. f, Violin plot of ISG module usage21 (with the addition of IFI6) split by presence or absence of H. flu reads in ViroCap sequencing of TA DNA. The numbers in the plot indicate average module usage by the group. Statistical testing for differences between groups was performed using a pairwise Wilcoxon rank sum test with continuity correction with p-value adjustment using the Benjamini & Hochberg correction. Significant differences (threshold of padj <0.05) between key groups is noted by a bar.
S3.jpg
Supplementary Figure 3: Additional analyses of subclustered MNP populations. a,.Violin plots grouped by MNP cell cluster (labeled as in Figure 4a) of select MNP transcripts. b, Feature plots of module usage of previously published 28,30,31 gene sets in subclustered MNPs. c, Violin plots grouped by subject group of module usage of previously published 28,30,31 gene sets. Testing for statistical differences between groups was performed using a pairwise Wilcoxon rank sum test with continuity correction and p-value adjustment using the Benjamini & Hochberg correction. Significant differences (threshold padj <0.05) between groups are labeled with a bar. d. Bar plots of proportional abundance of subject group MNPs within individual MNP cell clusters (labeled as in Figure 4a) at each Sample time point. Testing for significant differences between groups was performed using pairwise proportion test with continuity correction and p-value adjustment using Bonferroni correction (more than two groups compared) and a proportion test (two groups compared). Significant differences (threshold padj <0.05) between groups are labeled with a bar. e. Pie chart of proportion of MNPs that had RSV reads detected (RSVpos) and those that did not (RSVneg) (top, left) and bar chart of proportion of subject group MNPs that had RSV detected and those that did not (top, right). Testing for differences between groups was performed using a proportion test. Significant differences (threshold padj <0.05) between groups are noted with a bar. Box plots of percentage of cellular transcriptome occupied by RSV transcripts in RSV “infected” MNPs in Sample 1 (bottom, left) and Sample 2 (bottom, right). Testing for significant differences between groups was performed using a two-sided unpaired t test. f, Volcano plot demonstrating DEGs between MNP clusters 2,3, and 10 and all other MNPs (top). DGE testing was performed using Seurat’s FindAllMarkers in default settings (Wilcoxon rank sum test with Bonferroni p-value adjustment) comparing clusters 2,3, and 10 and all other MNPs (a positive avg_log2FC indicates increased expression by clusters 2,3, and 10 cells, and a negative avg_log2FC indicates increased expression by all other MNPs). Violin plots (bottom) grouped by Sample time point and split by subject group of Cluster 2,3,10 module usage. Testing for differences between groups was performed using a Kruskall-Wallis test as well as pairwise Wilcoxon rank sum test with p-value adjustment using the Benjamini & Hochberg correction (for comparisons amongst more than two groups) and an unpaired two-sided t test (for comparisons between two groups). Significant differences (threshold padj <0.05) between groups are labeled with a bar. g, Violin plots of regulons32 with significant differences in usage between groups in Sample 1 MNPs. Testing for differences between groups was performed using a Wilcoxon rank sum test with p-value adjustment using the Bonferroni correction. Significant differences (threshold padj <0.05) between groups are labeled with a bar. h, Feature plot of GADD45A expression in subclustered MNPs (top) and dot plots of GADD45A expression by cells within each MNP cluster (labeled as in Figure 4a) at each Sample time point. The size of the dot indicates percent of cells within an individual cluster expressing a particular transcript, and the color indicates the average expression of the transcript within the cell population.
S4.jpg
Supplementary Figure 4: Additional analyses of subsetted neutrophil populations. a, Violin plots of module usage of previously published neutrophil gene sets21 (top). Testing for differences between groups was performed using a Kruskall-Wallis test as well as pairwise Wilcoxon rank sum test with p-value adjustment using the Benjamini & Hochberg correction (for comparisons amongst more than two groups) and an unpaired two-sided t test (for comparisons between two groups). Significant differences (threshold padj <0.05) between groups are labeled with a bar. Feature plots of module usage of previously published neutrophil gene sets21 (bottom). b, Pie charts of proportion of subject group cells within the subsetted neutrophil population at each Sample time point. c, Bar charts of proportional abundance of subject group cells within each neutrophil transcriptional cluster (labeled as in Figure 5a). Testing for significant differences between groups was performed using pairwise proportion test with continuity correction and p-value adjustment using Bonferroni correction (more than two groups compared) and a proportion test (two groups compared). Significant differences (threshold padj <0.05) between groups are labeled with a bar. d, Pie chart of the proportion of RSV “infected” (RSVpos) neutrophils (left) and violin plot of RSV expression levels in neutrophil transcriptional clusters (labeled as in Figure 5a) (right). e, Bar chart of proportion of subject group neutrophils within RSV “infected” (RSVpos) and not RSV “infected” neutrophils (RSVneg) (left). Testing for differences between groups was performed using a proportion test and significant differences (threshold padj < 0.05) are noted with a bar. Box plots of percentage occupied by cellular transcriptome within Sample 1 and Sample 2 RSV “infected” neutrophils. Testing for differences between groups was performed using a two-sided unpaired t test. f, Volcano plot demonstrating DEGs between subject groups in mature and aged neutrophil populations. DGE testing between groups within each neutrophil population was performed using Seurat’s FindAllMarkers in default settings (Wilcoxon rank sum test with Bonferroni p-value adjustment). Volcano plots are labeled by group according to directionality of avglog2FC for each individual comparison within neutrophil clusters.
S5.jpg
Supplementary Figure 5: Concentrations of additional analytes measured in tracheal aspirate fluid. Plots of additional analytes measured in tracheal aspirate samples from Subjects enrolled on a different study at another center. The bar in the plot represents the median concentration of the analyte in each Subject group.
NCOMMS2234358rs.pdf
Reporting Summary

Download PDF

Journal Publication

published 30 Jun, 2023

Read the published version in Nature Communications →

Version 1

posted

You are reading this latest preprint version

Single Cell Analysis Identifies Distinct Profiles in Pediatric Acute Respiratory Distress Syndrome

Status:

Journal Publication

Version 1

Abstract

Figures

Full Text

Additional Declarations

Supplementary Files

Status:

Journal Publication

Version 1