The 3D Chromatin Landscape of B-Cell Precursor Childhood Acute Lymphoblastic Leukemia

Whilst the molecular pathogenesis of childhood B-cell precursor (BCP) acute lymphoblastic leukemia (ALL) has been studied extensively, its 3D chromatin landscape remains poorly explored. Genome-wide chromosome conformation capture methods have provided the tools to investigate the different units of chromatin organization, such as transcriptionally active (A) and inactive (B) compartments, topologically associating domains (TADs), and fine-scale chromatin loops and enhancer-promoter interactions. The aim of this study is to elucidate the chromatin architecture and topological gene regulation in childhood BCP ALL. To date, 29 primary patient samples were included, comprising the high hyperdiploid (HeH) (n=11), ETV6::RUNX1-positive (n=8), BCR::ABL1-positive (n=2), TCF3::PBX1-positive (n=2), DUX4-rearranged (n=2), intrachromosomal amplification of chromosome 21 (iAMP21) (n=1), KMT2A-rearranged (n=1), near-haploid (n=1) and near-triploid (n=1) genetic subtypes. Leukemic blast cells obtained at diagnosis were analyzed using Micro-C, a high-resolution variation of Hi-C (average number of total reads = 1.4 billion, highest resolution = 5 kb) combined with pair-end sequencing. Chromatin contact heatmaps were generated for each case using Juicer and Cooler. A/B compartments were identified using FanC at 500 kb resolution and visualized in the software IGV, while TAD calling was performed by Juicer, Domaincaller and Insulation Score. Differential chromatin interaction loop calling was made using Pareidolia and Mustache, and structural variant (SV) calling was carried out by EagleC. Preliminary principal component analysis of the 29 cases, based on the first two eigenvectors of the contact matrix (500 kb resolution), showed that HeH, ETV6::RUNX1-positive, TCF3::PBX1-positive and DUX4-rearranged cases each clustered based on their chromatin 3D organization. Furthermore, the A/B compartments of 11 HeH and 8 ETV6::RUNX1-positive cases were analyzed at 500 kb resolution and compartment shifts among the two subtypes were annotated. A total of 390 shifts were detected, where activating shifts (from B to A compartment) happened more often in HeH (263 shifts) than in ETV6::RUNX1-positive cases (127 shifts). Analysis of TAD boundary strength at 25 kb resolution revealed that HeH cases displayed significantly weaker boundaries compared to ETV6::RUNX1-positive cases. TAD boundary strength showed no bias towards the frequently gained or non-gained chromosomes in HeH ALL. By merging individual heatmaps of all HeH and ETV6::RUNX1-positive cases using Cooler, we created subtype-specific profiles and compared the intensity of chromatin interactions between the two genetic subtypes. Chromatin interaction intensity analysis was then combined with previously published RNA-sequencing data to identify transcriptional dysregulation events that could be associated with chromatin interaction changes. Preliminary results show that there was a chromatin loop missing close to the well-known leukemia-related gene IKZF1 in HeH compared to ETV6::RUNX1-positive cases; this gene also showed lower expression in the RNA-sequencing data. FLT3 was associated with weakened chromatin interactions and down-regulated in ETV6::RUNX1-positive cases compared to HeH, in agreement with its known high expression in HeH. Finally, we performed screening of SVs using EagleC and Micro-C heatmaps in HeH and ETV6::RUNX1-positive samples. Out of the 19 included cases, previous whole-genome sequencing (WGS) data were available for 16. We detected 75 SVs, of which 50 were intrachromosomal rearrangements and 25 were translocations. Micro-C heatmaps allowed visual detection of SVs smaller than 1 Mb and permitted identification of the type of SVs. WGS detected 61% of the SVs found in the HeH samples with Micro-C and 51% of those in the ETV6::RUNX1-positive cases. In summary, we present the first high-resolution genome-wide map of chromatin 3D organization in pediatric ALL. Our results indicate that different subtypes of childhood BCP ALL have distinct 3D chromatin landscapes and that abnormal chromatin architectures affect the regulation of leukemia-related genes.