PS1147 SPATIAL HETEROGENEITY IN CHRONIC LYMPHOCYTIC LEUKAEMIA ANALYSED BY WHOLE‐GENOME/EXOME SEQUENCING AT DIAGNOSIS

Abstract

Background:Chronic lymphocytic leukaemia (CLL) is characterised by the accumulation of neoplastic B cells in peripheral blood (PB), lymph nodes (LN) and bone marrow. CLL cells recirculate from PB to LN where the crosstalk with the microenvironment favours their maintenance and proliferation. Whole‐genome/exome sequencing (WGS/WES) studies focusing on samples from PB have described the mutational landscape of CLL uncovering a significant intra‐ and inter‐patient heterogeneity. However, the degree of spatial heterogeneity in CLL patients at diagnosis is largely unknown.Aims:We set up to dissect the spatial heterogeneity in fifteen CLL patients at diagnosis.Methods:Synchronous PB and LN DNA samples from fifteen CLL cases were analysed by WGS (n = 2) or WES (n = 13). Eight cases were IGHV mutated. Single nucleotide variants (SNVs) and short insertions and deletions (indels) were extracted from WGS/WES. Copy number alterations (CNA) were analysed from sequencing data and compared to SNP6.0 arrays. Structural variants (SV) were assessed in the two cases with WGS. The subclonal architecture of the tumour in PB and LN was reconstructed to asses for differences in the topographic representation of clones.Results:The two cases studied by WGS had 3539 and 2087 mutations, and 98.7% and 95.1% of them were shared by the PB and LN in each patient, respectively. Both cases also shared all CNA and SV in both sites. More than 75% of the mutations identified were clonal and found in both compartments. However, we identified the presence of 3 subclonal populations in each patient. In the first patient, the three subclones were equally represented in PB and LN with a clear pattern of linear evolution. The second patient showed a branching evolution with differences in the representation of two subclones (115 and 287 specific mutations, respectively) found at 7% and 47% of cancer cell fraction (CCF) in PB and at 33% and 19% of CCF in LN, respectively. The WES of the 13 cases identified a total of 49 CNA and 789 coding, splice site or UTR mutations. CNA were always shared at similar CCF between PB and LN in all cases. Regarding gene mutations, 3 cases shared 100% of alterations between PB and LN, 9 cases shared 80–95% (mean 90.5%) mutations, and only one patient had a major diversity in PB and LN (only 57% shared in both sites). Location‐specific mutations were virtually always present in small subclones (CCF < 30%), none was recurrent, and only SETD2 was previously recognised as a CLL driver. Considering the subclonal representation of the shared mutations, 5 (38%) patients had marked differences between PB and LN. The main pattern observed in 4 cases was the presence of a larger subclone in the PB only representing a small subfraction in the LN. Mutated genes in these PB predominant subclones were RICTOR (involved in PI3K‐AKT‐mTOR pathway), DIAPH3 (correlated with cell growth and migration), NPAT (involved in cell cycle and related to CLL pathogenesis), STAT1 (related to CLL differentiation and growth) or EPHA7 (Ras/MAPK signalling). No topographic differences were observed for mutations in known driver genes such as NOTCH1, TP53, ATM, SF3B1, IRF4 and DDX3X.Summary/Conclusion:Most mutations are shared between PB and LN of CLL patients at diagnosis. However, a significant fraction of patients showed marked differences in the abundance of small subclones carrying mutations in genes related to cell growth, migration and BCR signalling between PB and LN. No major spatial diversification seems to occur in known CLL driver genes and CNA between PB and LN at diagnosis.