Abstract 3939: Evolution and drug resistance of primary and relapse tumors in diffuse large B-cell lymphoma

Abstract

Abstract Our integrated multi-omics analysis of paired primary and relapse DLBCL tumors provides comprehensive view to tumor evolution, drug resistance and disease heterogeneity. DLBCL is an aggressive systemic disease in which tumors can arise in virtually any part of the body. Activated B-cell (ABC) and germinal center B-cell (GCB)-like subtypes differ molecularly and clinically. Three-year progression-free survival (PFS) in ABC is around 40% vs over 80% in GCB. Whole genome (WGS, 7 patients), transcriptome (total RNA, 7 patients) and methylome sequencing data produced using the Illumina HiSeq platform was complemented with Illumina TruSeq amplicon sequencing covering 102,571 bp of the genome on 788 custom selected targets. Altogether 23 fresh frozen and 28 paraffin embedded tumor samples from 24 patients with an advanced disease and at least one relapse event were included in the study. To investigate the clonal hierarchy of the primary and relapsed tumors we employed dirichlet process clustering using cancer cell fraction (CCF) estimates or variant allele frequencies from somatic point mutations. Divergent evolution was prominent in the GCB subtype whereas in ABC subtype subclonal mutations are rare and only few genomic differences were detected between primary and relapse tumors (p-value=0.047, Pearson's Chi-squared test on 7 patients with WGS data) suggesting pre-treatment drug resistance which is probably triggered by global hypo- or hypermethylation at relapse stage. In the GCB subtype primary and relapse branches have diverged much before diagnosis and subsequently developed independently. Number of emerging primary or relapse specific clonal and subclonal mutations correlate with PSF and overall survival (OS). Mutational signatures were concordant with the inferred evolutionary models in four out of five GCB cases. Contribution of the somatic hypermutation signature was higher in resistant tumors and correlated with time to progression of the disease. We identified 72 relapse tumor related genes including many known B-cell lymphoma genes such as CD79B, IGLL5, PIM1, SGK1. These were used in survival analysis in the CGCI RNA data set derived from 92 diagnostic DLBCL tumors. Patients with subclonal mutations in these genes had shorter PSF and OS (Kaplan-Meier p-values 0.0122 and 0.0071, respectively). To validate and further define our hypothesis we are simulating DLBCL evolution as a discrete time branching process model (Galton-Watson branching processes). The goal is to reconstruct phylogenetic trees from a simulated tumor and explain different types of evolutionary models inferred from the patient data. Despite of remarkable heterogeneity in DLBCL we observed two main evolutionary models which are concordant with disease subtype, patient’s survival, treatment response at diagnosis and relapse, and molecular features of a tumor. Citation Format: Rainer Lehtonen, Amjad Alkodsi, Kaiyang Zhang, Mikko Kivikoski, Emilia Kozlowska, Alejandra Cervera, Annika Pasanen, Harald Holte, Sirpa Leppä, Sampsa Hautaniemi. Evolution and drug resistance of primary and relapse tumors in diffuse large B-cell lymphoma [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2017; 2017 Apr 1-5; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2017;77(13 Suppl):Abstract nr 3939. doi:10.1158/1538-7445.AM2017-3939