Transcriptome Profiling of Induced Sputum Identified Upregulated TNF-α/NF-κB Signalling and Downregulated Mitochondrial Respiratory Chain Function in Post-Infectious Bronchiolitis Obliterans

Figures

• 

  Figure 1. Deregulated genes in sputum of patients with post-infectious Bronchiolitis obliterans. (A) Volcano plot showing all differentially expressed genes (DEGs; padj ≤ 0.05, base mean > 1), of which 629 were upregulated (red dots) and 565 were downregulated (blue dots) in the sputum of the PiBO patients compared to the controls. KEGG pathway analysis of (B) all DEGs with padj ≤ 0.05 and (C) full gene interacting network was performed using STRING (v12.0, medium confidence score > 0.4). The red and blue numbers above the clusters indicate the number of genes that are upregulated or downregulated, respectively. (D) Heat map and (E) enriched terms of gene ontology (biological process) of the top DEGs (padj < 0.001). (F) Network analysis shows hub genes that play a role in several of these processes.

• 

  Figure 3. Signal pathways affected by the downregulated DEGs. (A) KEGG pathway enrichment of DEGs with padj ≤ 0.05 and two-fold downregulation (base mean > 10). (B) Network analysis showing genes involved in metabolic pathways and oxidative phosphorylation. Violin plots depict normalised gene expression (log2(tpm + 1)) for (C) hub genes—MT-CO3, MTND2, NDUFAB1, NDUFB7, NDUFB8, NDUFS3, NDUFS5, and NDUFS—that are involved in both pathways and (D) genes associated with oxidative stress (CYCS, SOD2) and energy metabolism (HIF-1α, PFKB3). tpm, transcripts per million. * p < 0.05, ** p < 0.01, *** p < 0.001.

• 

  Figure 2. Signal pathways affected by the upregulated DEGs. (A) KEGG pathway enrichment of DEGs with padj ≤ 0.05 and two-fold upregulation (base mean > 10). (B) Network analysis showing genes involved in NF-κB signalling, TNFα-signalling, and apoptosis. Violin plots depict normalised gene expression (log2(TPM + 1)) for (C) hub genes—TNFRSF1, TRAF1, CFLAR, RELA, NFKBIA, and BIRC3—that are involved in all three pathways, (D) genes specific for neurophils—S100A8 and S100A9—and (E) genes specific for macrophages—MARCO, Il-1β, and MRC1. tpm, transcripts per million. * p < 0.05, ** p < 0.01, *** p < 0.001.

• 

  Figure 4. Shift in macrophage polarisation. Sputum slides were stained by multiplex immunohistochemistry (mIHC) using CD15 (cyan) and MPO (red) for neutrophils, CD68 for macrophages, MHCII (magenta) for M1 macrophages, and CD163 (yellow) and CD206 (green) for M2 macrophages. Cell nuclei were counterstained with DAPI (white). The cellular phenotypes of non-polarised alveolar macrophages (AM) and M1 and M2 macrophages were distinguished based on their intensity and parallel co-expression using a trained machine learning algorithm in the Form software. (A) Representative slides from a PiBO patient (left) and a healthy control (right). Scatter plot of absolute cell counts for (B) total cells, neutrophils, and macrophages and (C) macrophage phenotypes, calculated based on the relative frequency of mIHC data. (D) Ratio of M1 and M2 cells to AMs. * p < 0.05, ** p < 0.01.

• 

  Figure 5. TNF-α/NF-κB pathway and oxidative phosphorylation in M1 macrophages and lung epithelial cells. The graphs show expression of CFLAR and RELA as well as NDUFB8 and NDUFS5 in (A) neutrophil precursor cells (HL60), M0 and M1 macrophages (THP-1), and (B) human alveolar epithelial cells (A549) after stimulation with a cytokine mixture (CM, IL-1β, IFN-γ, and TNF-α). Expression of the genes was performed by qPCR, and data are presented as 2−ΔΔCT. Significance between unstimulated and stimulated samples is indicated with # p < 0.05, ## p < 0.01 and between different cell types with * p < 0.05.