Network-based approach identifies SP1 as a common node in emphysema and lung cancer

Abstract

Introduction: Cigarette smoking is the major culprit behind emphysema and lung cancer development, yet their pathogeneses seem to have opposing molecular mechanisms. For example, VEGFR2 blockade is effective against lung cancer yet spurs emphysema in mice. Objectives: Using protein interaction networks, we sought to identify a common node in the frequently comorbid conditions. Methods: A global interactome of 12,760 nodes and 77,005 edges was constructed from data sets including protein-protein, transcription factor-target, kinase-substrate, and protease-substrate interactions. 105 emphysema proteins were identified from the literature, while 989 lung cancer proteins were taken from a list curated by the National Cancer Institute. A hypergeometric enrichment score based upon the number of edges connecting lung cancer proteins and an emphysema interactome was used to derive p-values. Results: An emphysema interactome was constructed within the global network based upon the 105 emphysema proteins identified in the literature. Compared to 100,000 randomly generated networks of 105 nodes, the emphysema interactome contained 89 edges (random mean = 5.16, p < 10^-7) and a largest connected component of 58 nodes (random mean = 3.54, p < 10^-7) indicating a robust, connected network. Among lung cancer proteins, SP1 was the most connected with 16 edges to the emphysema interactome (p = 1.14 x 10^-12). Conclusions: Our network-based analysis predicted SP1 to be a common node in lung cancer and emphysema. The SP1 inhibitor mithramycin is known to impede lung cancer progression in mice, and we are evaluating the effect of mithramycin on emphysema development in mice with results to be presented at the meeting.