Network-based approach to prediction and population-based validation of in silico drug repurposing

Feixiong Cheng,Albert-LáSzló BarabáSi,Diane E Handy,Rishi J Desai,Ruisheng Wang,Joseph Loscalzo,Sebastian Schneeweiss

doi:10.1038/s41467-018-05116-5

Abstract

Here we identify hundreds of new drug-disease associations for over 900 FDA-approved drugs by quantifying the network proximity of disease genes and drug targets in the human (protein–protein) interactome. We select four network-predicted associations to test their causal relationship using large healthcare databases with over 220 million patients and state-of-the-art pharmacoepidemiologic analyses. Using propensity score matching, two of four network-based predictions are validated in patient-level data: carbamazepine is associated with an increased risk of coronary artery disease (CAD) [hazard ratio (HR) 1.56, 95% confidence interval (CI) 1.12–2.18], and hydroxychloroquine is associated with a decreased risk of CAD (HR 0.76, 95% CI 0.59–0.97). In vitro experiments show that hydroxychloroquine attenuates pro-inflammatory cytokine-mediated activation in human aortic endothelial cells, supporting mechanistically its potential beneficial effect in CAD. In summary, we demonstrate that a unique integration of protein-protein interaction network proximity and large-scale patient-level longitudinal data complemented by mechanistic in vitro studies can facilitate drug repurposing.

Highlights

We identify hundreds of new drug-disease associations for over 900 Food and Drug Administration (FDA)-approved drugs by quantifying the network proximity of disease genes and drug targets in the human interactome
To examine drug effects on CV diseases, we used a network proximity measure that quantifies the relationship between CV-specific disease modules and drug targets in the human protein–protein interaction (PPI) network (Supplementary Fig. 1b)
In the primary analytical approach of censoring patient follow-up time at discontinuation of the initial treatment (“astreated” approach), we observed that carbamazepine was associated with a 56% increased risk [hazard ratio (HR) 1.56, 95% confidence interval (CI) 1.12–2.18] of coronary artery disease (CAD) compared with levetiracetam (Fig. 3a), and hydroxychloroquine (Fig. 3d) was associated with a 24% reduced risk of CAD compared to leflunomide (HR 0.76, 95% CI 0.59–0.97)

Summary

Introduction

We identify hundreds of new drug-disease associations for over 900 FDA-approved drugs by quantifying the network proximity of disease genes and drug targets in the human (protein–protein) interactome. Without full knowledge of the broader network context of the molecular determinants of disease and drug targets in the protein–protein interaction network (human interactome), investigators cannot develop meaningful approaches for efficacious treatment of complex diseases[5] Novel approaches, such as network-based drug-disease proximity, that shed light on the relationship between drugs (drug targets) and diseases [molecular (protein) determinants in disease modules]6–8 can serve as a useful tool for efficient screening of potentially new indications for approved drugs with wellestablished pharmacokinetics/pharmacodynamics, safety and tolerability profiles, or previously unidentified adverse events[9,10,11,12]. We developed a systems pharmacology-based platform that quantifies the interplay between disease proteins and drug targets in the human protein–protein interactome with state-of-the-art pharmacoepidemiologic methods for hypothesis validation using longitudinal data with over 220 million patients We followed this analysis with in vitro assays to test potential drug mechanisms. These results suggest that this integrative approach can be generalized to other drugs/disease combinations

Methods

Results

Conclusion