Determination of tipping point in course of PM_2.5 organic extracts-induced malignant transformation by dynamic network biomarkers

The dynamic network biomarkers (DNBs) are designed to identify the tipping point and specific molecules in initiation of PM2.5-induced lung cancers. To discover early-warning signals, we analyzed time-series gene expression datasets over a course of PM(2.5 )organic extraction-induced human bronchial epithelial (HBE) cell transformation (0th similar to 16th week). A composition index of DNB (CIDNB) was calculated to determine correlations and fluctuations in molecule clusters at each timepoint. We identified a group of genes with the highest CIDNB at the 10th week, implicating a tipping point and corresponding DNBs. Functional experiments revealed that manipulating respective DNB genes at the tipping point led to remarkable changes in malignant phenotypes, including four promoters (GAB2, NCF1, MMP25, LAPTM5) and three suppressors (BATF2, DOK3, DAP3). Notably, co-altered expression of seven core DNB genes resulted in an enhanced activity of malignant transformation compared to effects of single-gene manipulation. Perturbation of pathways (EMT, HMGB1, STAT3, NE-kappa B, PTEN) appeared in HBE cells at the tipping point. The core DNB genes were involved in regulating lung cancer cell growth and associated with poor survival, indicating their synergistic effects in initiation and development of lung cancers. These findings provided novel insights into the mechanism of dynamic networks attributable to PM2.5-induced cell transformation.