Resolving tumour heterogeneity in solid FFPE samples by DEPArray imagebased, digital sorting

Introduction

Genomic profiling in solid tumour is notoriously challenging due to the presence of contaminating stromal and other kinds of cells. Microdissection is often attempted to enrich for the tumour cell fraction, but is inherently labour intensive and purity is limited especially when the tumour cells are highly intermingled with stromal cells, e.g. inflammatory cells. Also alternative approaches like the fluorescencebased cytometric analysis using FACS sorters to enrich tumour cells from disaggregated FFPE tumour tissues yield low efficiency and purity requiring a large input of cells material for instrument set-up and colour control. However the small sample size, e.g. biopsies, generally available in modern pathology impedes the use of conventional FACS for sorting. We present the results of genetic analysis for cell populations sorted by dieletrophoretic technology with our DEPArray system. This innovative technology solves two pressing problems in preparation of FFPE samples for genomic analysis: small sample size and unwanted admixture of stromal cells and other cell types. The presented procedure leads to robust and reproducible results, making it possible to unambiguously detect the true-positive variants and reliably analyze quantitative trait such as loss of heterozygosity (LoH) and other copy number variations (CNVs), which cannot be evaluated precisely in unsorted FFPE samples. On several loci, we detected by exome sequencing of pure pools of 300 sorted tumour cells somatic mutations with 100% variant frequency clearly indicating LoH and confirming 100% purity of sorted cells. By appropriate analysis of exome data sets as well as by low pass next generation sequencing we were also able to unambiguously identify genetic alterations in sorted tumour cells (e.g. LoH and mono-allelic amplifications) which are difficult to detect in bulk DNA extracted from non sorted tissue with low tumour content.