Harnessing synthetic lethality to predict the response to cancer treatment

Joo Sang Lee,Daniel James,Seung Gu Park,Matthew Davidson,Paul S Meltzer,Sridhar Hannenhalli,Welles Robinson,Joshua J Waterfall,Kuoyuan Cheng,Cyril H Benes,Chani Stossel,Talia Golan,Emma Shanks,Rand Arafeh,Lynn Mcgarry,Yardena Samuels,Arabinda Das ,Ella Buzhor,Dikla Atias,Gidi Stein,Eyal Gottlieb,Arnaud Amzallag,Noam Auslander,Livnat Jerby‐Arnon ,Eytan Ruppin

doi:10.1038/s41467-018-04647-1

Abstract

While synthetic lethality (SL) holds promise in developing effective cancer therapies, SL candidates found via experimental screens often have limited translational value. Here we present a data-driven approach, ISLE (identification of clinically relevant synthetic lethality), that mines TCGA cohort to identify the most likely clinically relevant SL interactions (cSLi) from a given candidate set of lab-screened SLi. We first validate ISLE via a benchmark of large-scale drug response screens and by predicting drug efficacy in mouse xenograft models. We then experimentally test a select set of predicted cSLi via new screening experiments, validating their predicted context-specific sensitivity in hypoxic vs normoxic conditions and demonstrating cSLi’s utility in predicting synergistic drug combinations. We show that cSLi can successfully predict patients’ drug treatment response and provide patient stratification signatures. ISLE thus complements existing actionable mutation-based methods for precision cancer therapy, offering an opportunity to expand its scope to the whole genome.

Highlights

While synthetic lethality (SL) holds promise in developing effective cancer therapies, SL candidates found via experimental screens often have limited translational value
It applies them in a specific sequential manner that minimizes the computational cost of their identification (Fig. 1a, see Methods for full details): First, identification of clinically relevant synthetic lethality (ISLE) mines gene expression and SCNA data of the input patient tumor samples to identify under-represented candidate gene pairs, whose co-inactivation is significantly less frequent than expected by their individual inactivation frequencies, testifying that their co-inactivation is under negative selection
Results are shown for ISLE cSL interactions and compared with the DAISY SL-network, ncSL network, and randomly permuted networks double-drug response, we hypothesized that a pair of drugs will be synergistic if there exists a strongly predicted SL between their gene targets (Methods)

Summary

Introduction

While synthetic lethality (SL) holds promise in developing effective cancer therapies, SL candidates found via experimental screens often have limited translational value. We present a data-driven approach, ISLE (identification of clinically relevant synthetic lethality), that mines TCGA cohort to identify the most likely clinically relevant SL interactions (cSLi) from a given candidate set of lab-screened SLi. We first validate ISLE via a benchmark of large-scale drug response screens and by predicting drug efficacy in mouse xenograft models. We show that cSLi, which are inferred from mining untreated patients’ data, can successfully predict treatment outcomes in cancer patients without any need for training on specific patient cohorts of drug response data Taken together, these results offer a novel approach for precision-based cancer therapy from the patients’ tumor data

Methods

Results

Conclusion