Integrative Analysis of 3D Chromatin Organization Revealed Regulation Mechanisms in Angioimmunoblastic T-Cell Lymphoma

Molecular studies have highlighted recurrent alterations in epigenetic modifier genes, including TET2, DNMT3A, and IDH2, in angioimmunoblastic T-cell lymphoma (AITL). However, the direct impact of gene regulatory elements on gene expression changes leading to T-cell lymphomagenesis remains poorly understood.

We focus on 3D chromatin organization using Hi-C and RNA sequencing (RNA-seq). Hi-C and RNA sequencing was performed on diagnostic FFPE tumor samples from seventeen patients with AITL. Normal CD4+ T cell Hi-C data (Kloetgen et al., 2020) were used as controls. To annotate AITL-specific loops in association with enhancer/ silencer status, we utilized H3K27ac, H3K4me1 and H3K27me3 histone marks in Chromatin Immunoprecipitation Sequencing (ChIP-seq) data of follicular helper T cells (Tfh) (Weinstein et al., 2014). We defined promoter-enhancer (P-E) loops where the non-promoter anchor had H3K27ac without H3K27me3, and promoter-silencer (P-S) loops where the non-promoter anchor had H3K27me3 without H3K27ac.

In comparison to normal CD4+ T-cell controls, recurrent AITL-specific alterations in A/B (active/inactive) compartments, topologically associating domains (TADs), and chromatin loops were identified. Principal-component analysis (PCA) of genome-wide compartment scores indicated AITL samples and normal CD4+ T cells clearly separated by the first two components. On average, there was 3.5% (range: 2.0-5.3) A-to-B compartment switch and 1.9% (range: 0.8-4.8) B-to-A switch when comparing AITL samples with normal CD4+ T cell controls (p=5.17 × 10⁻⁵, two-sided paired t-test). Alterations in TADs were recurrently observed in 434 of 743 cancer census genes in COSMIC, including MYC family genes like MYC, MYCN, and MYCL. Recurrent gains of TADs (AITL-specific TADs absent in control samples) were associated with the MYC (n=15) and MYCL (n=17) genes, while TAD shrinkage (reduction in the range of TADs from control to AITL samples) was observed in the MYCN region (n=15). On average, 60.1% of loops include promoters in at least one anchor. Genes within AITL-specific loops exhibited significantly higher expression than those without loop regions (p=7.39 × 10⁻⁸⁸, Kruskal-Wallis test). Genes within P-E loops showed significantly higher expression than those within P-S loops in the whole transcriptome (p = 6.45 × 10⁻⁵, Kruskal-Wallis test), although this difference was not significant in protein-coding genes (p = 1.00, Kruskal-Wallis test).

Based on the differential loop analysis using Peakachu, significant chromatin loops were observed in the MYC region and genes associated with pathways related to MYC, such as the Wnt signaling pathway and ErbB signaling pathway. AITL-specific loops in the MYC region were observed in eight of seventeen patients. The other ends of the anchors in recurrent AITL-specific MYC-associated loops were located in two intermediate enhancer (H3K4me1 peaks without H3K27ac or H3K27me3 marks) loci of Tfh, chr8: 128180000-128190000 (n=4) surrounding ME2 proximal MYC enhancer in acute myeloid leukemia (Pulikkan et al., 2018), and chr8: 128310000-128330000 (n=6) approximately 60-80 kb towards 3’ from the MYC-LASE enhancer in lung adenocarcinoma (Zhang et al., 2016).

Our findings indicate that AITL exhibits unique chromatin conformations, suggesting the formation of a lymphoma-specific gene regulatory landscape. Recurrent changes in chromatin conformation involve MYC family genes and associated pathways. Distinct gene regulatory elements associated with enhancer regions in AITL is implicated. Correlation of the regulatory candidates with their mutational signature is underway and will be reported at the meeting.