Medium levels of transcription and replication related chromosomal instability are associated with poor clinical outcome

Ataaillah Benhaddou,Sophie Piperno‐Neumann ,Laura Leroy ,Gonzague De Pinieux,Binh Bui,Philippe Rochaix,Axel Le Cesne,Nelly Firmin,Laëtitia Gaston ,Gaëlle Pérot ,Eberhard Stoeckle,Nelly Desplat,Françoise Collin ,Christine Chevreau,Jean‐Yves Blay ,Thibaud Valentin,Gwénaël Ferron ,Sophie Le Guellec,Jean‐Michel Coindre ,Mohamed Ben Haddou,Frédéric Chibon

doi:10.1038/s41598-021-02787-x

Abstract

Genomic instability (GI) influences treatment efficacy and resistance, and an accurate measure of it is lacking. Current measures of GI are based on counts of specific structural variation (SV) and mutational signatures. Here, we present a holistic approach to measuring GI based on the quantification of the steady-state equilibrium between DNA damage and repair as assessed by the residual breakpoints (BP) remaining after repair, irrespective of SV type. We use the notion of Hscore, a BP “hotspotness” magnitude scale, to measure the propensity of genomic structural or functional DNA elements to break more than expected by chance. We then derived new measures of transcription- and replication-associated GI that we call iTRAC (transcription-associated chromosomal instability index) and iRACIN (replication-associated chromosomal instability index). We show that iTRAC and iRACIN are predictive of metastatic relapse in Leiomyosarcoma (LMS) and that they may be combined to form a new classifier called MAGIC (mixed transcription- and replication-associated genomic instability classifier). MAGIC outperforms the gold standards FNCLCC and CINSARC in stratifying metastatic risk in LMS. Furthermore, iTRAC stratifies chemotherapeutic response in LMS. We finally show that this approach is applicable to other cancers.

Highlights

Genomic instability (GI) influences treatment efficacy and resistance, and an accurate measure of it is lacking
We define the characteristics of hotspots as follows: (1) DNA elements are significantly more broken than expected by chance (Hscore ≥ 3) while the immediate surrounding regions are not (Hscore < 3); (2) if the DNA elements and the immediate surrounding regions have a Hscore ≥ 3, a DNA element is considered as a hotspot if its Hscore is at least 1.5 times the Hscore of both surrounding regions
This study describes new tools, measures and insights that address the question of the clinical outcome and its relationship with genomic rearrangement

Summary

Introduction

Genomic instability (GI) influences treatment efficacy and resistance, and an accurate measure of it is lacking. LMS develops due to frequent p53 and RB1 pathway a lterations[31, 32] and a highly rearranged genome with a high number of chromosomal rearrangements This leads to many copy number variations (CNV) and BP that are associated with poor o utcome[33]. Current approaches in cancer genomics often use exome-seq to build a catalogue of mutations in multiple cancer types by sequencing hundreds of tumor samples in order to find diagnostic, prognostic, and therapeutic targets[36]. Hscore measures the propensity of a given DNA element to break more than expected with a random breakage model By using this approach, we have developed a combination of both transcription-associated and replication-associated markers of GI and tested whether both measures are prognostic of metastatic risk in LMS. We assessed whether the transcription- and replication-associated markers are predictive of chemotherapeutic response in LMS

Methods

Results

Conclusion