Abstract

Mature T-cell neoplasms (MTCN) are heterogeneous diseases with dismal prognosis. Differentiating between the many entities requires specialized pathology expertise, and studies show up to 30% of minor or major diagnostic reclassifications following expert review of cases. Assay for transposase-accessible chromatin sequencing (ATAC-seq) is a simple technique to profile open chromatin regions, which has been shown to be highly discriminative for clustering solid tumors and acute myeloid leukemias. We applied ATAC-seq to MTCN to explore the epigenetic landscape of these different entities, and built a predictive model to aid in diagnosis. FACS-sorted tumor cells from primary MTCN samples and 50µm sections of frozen tumor tissue from the French TENOMIC T-cell lymphoma consortium were processed according to previously published FAST- and OMNI-ATAC protocols, respectively. In parallel, we applied FAST-ATAC to several normal T and NK cell subsets sorted from PBMC or lymph node suspensions of healthy donors. Sequencing data were processed by an adapted version of the ENCODE ATAC-seq pipeline using a custom Hg38 genome version including HTLV1 sequence. A total of 678 normal and tumor samples were sequenced to provide a comprehensive landscape of chromatin accessibility in MTCN. Unsupervised clustering of normal NK and T cell subtypes (N = 49) and sorted tumoral lymphoma cells (N = 104) confirmed that AITL are derived from TFH cells, HSTL and LGL are closely segregated with NK and gd-T cells. We also found that T-PLL likely derive from the transformation of naïve T cells. Epigenetic profiling by ATAC-seq of FACS-sorted tumor samples resulted in a complete segregation of the known MTCN entities (TFH, ALK+ and ALK- ALCL, HSTL, CTCL, ATLL, LGL and T-PLL). Most PTCL-NOS (13/17) clustered with a predefined MTCN subtype (mainly AITL/TFH-phenotype PTCL, CTCL and lymphomas exhibiting cytotoxic features), showing that this waste basket subgroup is merely virtual, at least from an epigenetic point of view. Using unsupervised deconvolution approaches, we were able to discriminate different MTCN subtypes from 223 processed frozen bulk samples. All known MTCN subtypes were identified by ATAC-seq but a novel subset of PTCL-NOS representing ∼5% of cases was pinpointed, showing high GATA3 transcription factor activity. Subsequent exome sequencing revealed numerous copy number alterations and TP53 (8/12) and NCOR1 mutations (7/12). A support vector machine model was trained to predict diagnosis and showed accurate classification performance by cross-validation and on external validation cohort collected from 5 tertiary care centers. ATAC-seq is a rapid and cost-effective technique with high classification accuracy for PTCL subtypes. Among GATA3-expressing PTCL that spread across multiple epigenetic subgroups, we identified a specific entity with recurrent NCOR1 mutations. Keywords: aggressive T-cell non-Hodgkin lymphoma, genomics, epigenomics, and other -omics, pathology and classification of lymphomas No conflicts of interests pertinent to the abstract.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call