Polymyxin Induces Significant Transcriptomic Perturbations of Cellular Signalling Networks in Human Lung Epithelial Cells

Abstract

Inhaled polymyxins are increasingly used to treat pulmonary infections caused by multidrug-resistant Gram-negative pathogens. We have previously shown that apoptotic pathways, autophagy and oxidative stress are involved in polymyxin-induced toxicity in human lung epithelial cells. In the present study, we employed human lung epithelial cells A549 treated with polymyxin B as a model to elucidate the complex interplay of multiple signalling networks underpinning cellular responses to polymyxin toxicity. Polymyxin B induced toxicity (1.0 mM, 24 h) in A549 cells was assessed by flow cytometry and transcriptomics was performed using microarray. Polymyxin B induced cell death was 19.0 ± 4.2% at 24 h. Differentially expressed genes (DEGs) between the control and polymyxin B treated cells were identified with Student’s t-test. Pathway analysis was conducted with KEGG and Reactome and key hub genes related to polymyxin B induced toxicity were examined using the STRING database. In total we identified 899 DEGs (FDR < 0.01), KEGG and Reactome pathway analyses revealed significantly up-regulated genes related to cell cycle, DNA repair and DNA replication. NF-κB and nucleotide-binding oligomerization domain-like receptor (NOD) signalling pathways were identified as markedly down-regulated genes. Network analysis revealed the top 5 hub genes (i.e., degree) affected by polymyxin B treatment were PLK1(48), CDK20 (46), CCNA2 (42), BUB1 (40) and BUB1B (37). Overall, perturbations of cell cycle, DNA damage and pro-inflammatory NF-κB and NOD-like receptor signalling pathways play key roles in polymyxin-induced toxicity in human lung epithelial cells. Noting that NOD-like receptor signalling represents a group of key sensors for microorganisms and damage in the lung, understanding the mechanism of polymyxin-induced pulmonary toxicity will facilitate the optimisation of polymyxin inhalation therapy in patients.

Keywords: lung epithelial cells; polymyxin; pulmonary toxicity; transcriptomics.

Figure 1

The top 15 genes affected by polymyxin B treatment as identified using variable influence on projection (VIP) scores. Key genes were identified using partial least squares-discriminant analysis (PLS-DA). VIP score and heatmap indicate the relative expression level of the corresponding gene in each group. PMB, polymyxin B.

Figure 2

Cell cycle perturbations by polymyxin B. Red boxes present the proteins encoded by up-regulated genes; green boxes present the proteins encoded by down-regulated genes.

Figure 3

Common DEGs involved in the cell cycle pathways due to polymyxin B treatment.

Figure 4

Common up-regulated genes in response to polymyxin treatment shared by Reactome and KEGG. (A) Shared up-regulated DEGs in the DNA replication pathway; (B) shared up-regulated DEGs in the DNA repair pathway.

Figure 5

Perturbations of the signalling networks in A549 cells induced by polymyxin B treatment. (A) Cell cycle is the largest connected component containing 142 nodes and 804 edges; (B) Death receptor signalling component is consisted of 5 down-regulated genes; (C) Cholesterol biosynthesis component contains 19 up-regulated genes and 1 down-regulated genes; (D) Component D shows 5 up-regulated and 5 down-regulated genes which are related to plasma lipoprotein clearance and signalling by TGF-beta receptor complex; (E) mRNA splicing component contains 3 up-regulated and 3 down-regulated genes; both (F) GPCR ligand binding and (G) Glycolysis components are comprised of 2 up-regulated and 3 down-regulated genes; (H) Intra-flagellar transport component contains 4 down-regulated genes, while (I) Fatty acyl-COA biosynthesis component contains 4 up-regulated genes. Red and green nodes are up-regulated and down-regulated genes, respectively, following polymyxin B treatment. Interactions between proteins are indicated by lines.