High-Throughput Screen of Natural Compounds and Biomarkers for NSCLC Treatment by Differential Expression and Weighted Gene Coexpression Network Analysis (WGCNA)

Abstract

Lung cancer is known as the leading cause which presents the highest fatality rate worldwide; non-small-cell lung cancer (NSCLC) is the most prevalent type of lung carcinoma with high severity and affects 80% of patients with lung malignancies. Up to now, the general treatment for NSCLC includes surgery, chemotherapy, and radiotherapy; however, some therapeutic drugs and approaches could cause side effects and weaken the immune system. The combination of conventional therapies and traditional Chinese medicine (TCM) significantly improves treatment efficacy in lung cancer. Therefore, it is necessary to investigate the chemical composition and underlying antitumor mechanisms of TCM, so as to get a better understanding of the potential natural ingredient for lung cancer treatment. In this study, we selected 78 TCM to treat NSCLC cell line (A549) and obtained 92 transcriptome data; differential expression and WGCNA were applied to screen the potential natural ingredient and target genes. The sample which was treated with A. pierreana generated the most significant DEG set, including 6130 DEGs, 2479 upregulated, and 3651 downregulated. KEGG pathway analyses found that four pathways (MAPK, NF-kappa B, p53, and TGF-beta signaling pathway) were significantly enriched; 16 genes were significantly regulated in these four pathways. Interestingly, some of them such as EGFR, DUSP4, IL1R1, IL1B, MDM2, CDKNIA, and IDs have been used as the target biomarkers for cancer diagnosis and therapy. In addition, classified samples into 14 groups based on their pharmaceutical effects, WGCNA was used to identify 27 modules. Among them, green and darkgrey were the most relevant modules. Eight genes in the green module and four in darkgrey were identified as hub genes. In conclusion, we screened out three new TCM (B. fruticose, A. pierreana, and S. scandens) that have the potential to develop natural anticancer drugs and obtained the therapeutic targets for NSCLC therapy. Our study provides unique insights to screen the natural components for NSCLC therapy using high-throughput transcriptome analysis.

Figure 1

Plant morphology, differentially expressed genes (DEGs), and KEGG pathways of samples treated with three candidate TCM. (a–c) Breynia fruticosa, (d–f) Argyreia pierreana, and (g–i) Senecio scandens. Plant morphology pictures cited from Flora of China (source: http://www.iplant.cn). In volcano plots, the horizontal line at ∣log2FC | = 1; vertical line at false discovery rate (FDR) = 0.05. In bar plots of KEGG pathways, red block represented the key pathways.

Figure 2

Expression and regulation of top genes enriched on the key pathways. (a) MAPK signaling pathway; (b) NF-kappa B signaling pathway; (c) p53 signaling pathway; (d) TGF-beta signaling pathway. CK: the negative control.

Figure 3

Estimation of soft-thresholding values, network heat map plot, and clustered correlation. (a) Scale independence and mean connectivity, power of β = 7 (scale free R² = 0.9). (b) The topological overlap heat map of coexpression genes summarized the interaction relationship of selected genes. Each module is represented by different colors. The brightness of yellow and red represented the high connectivity of modules. (c) Module eigengene dendrogram and eigengene adjacency heat map presented the relationship of the modules generated by the clustering analysis.

Figure 4

The gene set enrichment analysis and the identification of modules. (a) Dendrogram of dissimilarity between all filtered gene sets, which enriched according to a dissimilarity measure (1-TOM), and the cluster modules were marked as different colors. Dynamic tree cut algorithm generated the first set, then merged the correlated modules together. Each gene was represented by a branch, and each module was represented by one color. (b) Heat map of the correlation between the pharmaceutical effect of TCM fractions and MEs in NSCLC. The dark degree of the module color represented the significance of their relationship.

Figure 5

The coexpression network. (a) The coexpression network of the significant genes in the green module, including 132 nodes. (b) The coexpression network of the significant genes in the darkgrey module, including 81 nodes. The hub genes were represented in red color.