Prioritizing Therapeutics for Lung Cancer: An Integrative Meta-analysis of Cancer Gene Signatures and Chemogenomic Data

Abstract

Repurposing FDA-approved drugs with the aid of gene signatures of disease can accelerate the development of new therapeutics. A major challenge to developing reliable drug predictions is heterogeneity. Different gene signatures of the same disease or drug treatment often show poor overlap across studies, as a consequence of both biological and technical variability, and this can affect the quality and reproducibility of computational drug predictions. Existing algorithms for signature-based drug repurposing use only individual signatures as input. But for many diseases, there are dozens of signatures in the public domain. Methods that exploit all available transcriptional knowledge on a disease should produce improved drug predictions. Here, we adapt an established meta-analysis framework to address the problem of drug repurposing using an ensemble of disease signatures. Our computational pipeline takes as input a collection of disease signatures, and outputs a list of drugs predicted to consistently reverse pathological gene changes. We apply our method to conduct the largest and most systematic repurposing study on lung cancer transcriptomes, using 21 signatures. We show that scaling up transcriptional knowledge significantly increases the reproducibility of top drug hits, from 44% to 78%. We extensively characterize drug hits in silico, demonstrating that they slow growth significantly in nine lung cancer cell lines from the NCI-60 collection, and identify CALM1 and PLA2G4A as promising drug targets for lung cancer. Our meta-analysis pipeline is general, and applicable to any disease context; it can be applied to improve the results of signature-based drug repurposing by leveraging the large number of disease signatures in the public domain.

A review of connectivity map and computational approaches in pharmacogenomics

“…Three of the expression sets were profiled on MCF7 breast cancer cell lines (GSE9936-three compounds, GSE5149 and GSE28662-two compounds), and two on MDA-MB-231 metastatic breast cancer lines (GSE33552-two compounds). The rest of the expression sets were profiled in a B-cell lymphoma cell lines, which are chronic lymphocytic leukemia patient-derived cell lines(GSE14973), K422 non-Hodgkin's lymphoma cell lines (GSE7292), lytic-permissive lymphoblastoid cell lines (GSE31447), diffuse large B-cell lymphoma patient-derived cell lines (GSE40003) and mantle cell lymphoma cell lines (GSE34602 [24][25][26][27]. Notably, Cheng et al [28] presented a systematic approach to quantitatively assess the performance of such methods.…”

“…CMapBatch: a meta-analysis of drug response Fortney et al [27] have recently adapted a parallel CMap approach across multiple gene signatures of a disease, and named the method 'CMapBatch'. Specifically, instead of applying CMap to one individual gene signature, the authors apply it to multiple gene signatures for the same disease and then combine the resulting outcomes.…”

“…The Rank Product method was originally developed to identify DEG for replicated experiments based on the ranking of the individual experiments. Fortney et al analyzed 21 signatures (s ¼ 21) for lung cancer obtained from Oncomine [27,39]. The results reveal that CMapBatch produces indeed a more stable list of drugs when compared with the individual gene signatures.…”

“…Thus, scaling up transcriptional knowledge increases the hit percentage significantly from 44 to 78% of the top-ranked drugs. Moreover, the resultant drug hits were characterized in silico and showed slow growth significantly in nine lung cancer cell lines from the NCI-60 collection [27]. In total, 247 candidate therapeutics were identified for which two genes, CALM1 and PLA2G4A, are found to be markers for drug targets in lung cancer [40].…”

A Next Generation Connectivity Map: L1000 Platform and the First 1,000,000 Profiles

“…Examples of CMap use include the anthelmintic drug parbendazole as an inducer of osteoclast differentiation (Brum et al, 2015), celastrol as a leptin sensitizer (Liu et al, 2015), compounds targeting COX2 and ADRA2A as potential diabetes treatments (Zhang et al, 2015), small molecules that mitigate skeletal muscular atrophy (Dyle et al, 2014) and spinal muscular atrophy (Farooq et al, 2009), and new therapeutic hypotheses for the treatment of inflammatory bowel disease (Dudley et al, 2011) and cancer (Singh et al, 2016); (Muthuswami et al, 2013; Wang et al, 2008); (Schnell et al, 2015); (Fortney et al, 2015; Wang et al, 2011); (Churchman et al, 2015); (Rosenbluth et al, 2008); (Saito et al, 2009); (Stockwell et al, 2012). …”