Ph.D. defense 05-July-2021 Epigenetic blueprint of human thymopoiesis and adult T-cell Acute Lymphoblastic Leukemia Anand Mayakonda Division of Cancer Epigenomics (B370) Advisor: Prof. Dr. Christoph Plass Thymopoiesis - development of ɑβ T-cells Author Division 10/4/21 | • Systematic process involving differentiation of immature thymocytes into mature functional T-cells NOTCH1 high • CD34+ early thymic progenitors (ETP) arrive from bone marrow • NOTCH1 expression leads to lineage commitment • T-Cell Receptor+ (TCR), CD4&CD8 double positive cells differentiate into CD4+ or CD8+ cells • Mature CD4+ or CD8+ cells leave thymus and enter periphery 10/4/21 | Page 2 Introduction Hypothesis M&M Results-1 Results-2 Summary Transcription factors and gene expression governing thymopoiesis Author Division 10/4/21 | CD4+ CD4+ CD4+ CD34+ CD34+ CD8 + CD1A- CD1A+ CD8 - CD8 + CD3- CD3+ CD8+ CD34 CD34 TCR NOTCH1 E2A RAG1 ZBTB7B CD4+ GATA3 RUNX MYB TCF1 NOTCH1 MYB RAG1 BCL11B ZBTB7B 10/4/21 | Page 3 Introduction Hypothesis M&M Results-1 Results-2 Summary NOTCH1: Towards T cell lineage commitment Author Division 10/4/21 | NOTCH1 loss causes maturation arrest CD4+ CD4+ CD4+ CD34+ CD34+ CD8 + CD1A- CD1A+ CD8 - CD8 + CD3- CD3+ CD8+ CD34 CD34 TCR CD4+ NOTCH1 Immature and partially differentiated thymocytes T-cell Acute Lymphoblastic Leukemia (T-ALL) Belver L, Ferrando A. The genetics and mechanisms of T cell acute lymphoblastic leukaemia. Nat Rev Cancer. 2016 10/4/21 | Page 4 Introduction Hypothesis M&M Results-1 Results-2 Summary DNA methylation (DNAm) dynamics during development and disease Author Division 10/4/21 | Active gene expression Loss of DNAm during (Mouse) Hematopoiesis Aberrant DNAm changes leads to loss of leads to cellular identity cellular identity and tumorigenesis Reduced gene expression Sina Stäble Ph.D. thesis 10/4/21 | Page 5 Introduction Hypothesis M&M Results-1 Results-2 Summary Rationale: Role of DNAm in thymopoiesis and T-ALL Author Division 10/4/21 | Aim-1: DNA methylation changes during thymopoiesis 1. Developmental associated genomic regions 2. Comparative analysis with Hematopoiesis 3. Role of DNAm in Thymopoiesis to leukemic transformation Aim-2: DNA methylation based classification of T-ALL 1. Identification of epigenetically distinct T-ALL subtypes 2. DNAm as a biomarker: Clinically actionable subtypes 3. Role of DNAm changes in leukemogenesis 10/4/21 | Page 6 Introduction Hypothesis M&M Results-1 Results-2 Summary Materials and Methods Author Division 10/4/21 | Thymic number of thymi subpopulation T-ALL = 143 (Age > 15) CD34+ CD1A- 4 CD34+ CD1A+ 4 NGS panel CD4+ Immature Single Positive (4ISP) 4 (~150 leukemic genes) Early commitment (EC) CD4+CD8+ CD3- 2 TCR- Late commitment (LC) CD4+CD8+ CD3+ 2 DNA methylation TCR+ (850K CpG sites) Single CD4+ 2 Single CD8+ 2 • 7 distinct FACS sorted intra-thymic cell types (by Aurore Touzart) • French GRAAL clinical trials • WGBS by SWIFT protocol (by Dieter Weichenhan and Aurore Touzart) • Median age: 29.9 years • median 25X coverage 10/4/21 | Page 7 Introduction Hypothesis M&M Results-1 Results-2 Summary Methrix: Fast and efficient summarization of generic bedGraph files from Author Division Bisufite 10/4/21 | sequencing Problem: • WGBS data is large (e.g., 28M CpG sites in human genome) • Computational requirements (memory intese) • Input file format discrepancies Solution • Methrix – An R package for WGBS data analysis • Allows format free reading and summarization of WGBS datasets • In memory and hdf5 arrays are supported • Built upon SummarizedExperiment data structure Contributors: R. Toth, J. Hey, M. Schönung, R. Batra, Feuerstein-Akgoz C 10/4/21 | Page 8 Introduction Hypothesis M&M Results-1 Results-2 Summary Author Division 10/4/21 | Rsults-1 DNAm dynamics during T-cell development 10/4/21 | Page 9 Introduction Hypothesis M&M Results-1 Results-2 Summary WGBS of intrathymic cell types Author Division 10/4/21 | Loss of methylation is a common feature across hematopoiesis and thymopoiesis despite being subjected distinct tissue microenvironment (Bone marrow v/s Thymus) Farlik M, Halbritter F, Müller F, et al. DNA Methylation Dynamics of Human Hematopoietic Stem Cell Differentiation. Cell Stem Cell. 2016 10/4/21 | Page 10 Introduction Hypothesis M&M Results-1 Results-2 Summary Progressive loss of DNAm during thymopoiesis Author Division 10/4/21 | Thymic differentially methylated regions (tDMR) Irreversible and unidirectional loss of DNAm |meth| > 20% and P < 0.01 10/4/21 | Page 11 Introduction Hypothesis M&M Results-1 Results-2 Summary Loss of methylation occurs at lineage specific TF binding sites Author Division 10/4/21 | • Locus Overlap Analysis (LOLA) with TFBS (FDR < 0.05) • tDMRs occurs at TFBS necessary for T-cell lineage commitment • NOTCH1 and MYB – master TFs • GATA1 a CD4+ and CD8+ specific TFs Loss of methylation facilitates binding of core TFs necessary of lineage commitment and T-cell development 10/4/21 | Page 12 Introduction Hypothesis M&M Results-1 Results-2 Summary DNA methylation changes reconstruct lymphatic hierarchy Author Division 10/4/21 | Farlik M, Halbritter F, Müller F, et al. DNA Methylation Dynamics of Human Hematopoietic Stem Cell Differentiation. Cell Stem Cell. 2016 10/4/21 | Page 13 Introduction Hypothesis M&M Results-1 Results-2 Summary tDMRs are hypermethylated in T-ALL Author Division 10/4/21 | 187 105 100 164 101 143 22 42 19 42 42 20 19 10 23 20 28 3 3 3 5 0.8 Avg. DNAm of tDMRs T-ALL 0.7 B-ALL AML 0.6 DNAm mediated loss of cellular identity 0.5 0.4 M6:AML M5:AML M7:AML M4:AML M3:AML M2:AML M1:AML M0:AML HeH:BCP t(9;22):BCP iAMP21:BCP t(1;19):BCP >67chr:BCP undefined:BCP non−recurrent:BCP dic(9;20):BCP <45chr:BCP t(12;21):BCP 11q23/MLL:BCP T−ALL:TCP TALL Disease subtypes Cancer Genome Atlas Research Network, et al. Genomic and epigenomic landscapes of adult de novo acute myeloid leukemia. N Engl J Med. 2013 Nordlund J, et al. Genome-wide signatures of differential DNA methylation in pediatric acute lymphoblastic leukemia. Genome Biol. 2013 10/4/21 | Page 14 Introduction Hypothesis M&M Results-1 Results-2 Summary Summary: Author Division 10/4/21 | • Unidirectional and irreversible loss of methylation is a feature of thymopoiesis • Loss of DNAm occurs at the TFBS critical for T-cell lineage commitment • tDMRs reconstructs the hematopoietic developmental system and conserved across T-lineage cell types • tDMRs are significantly hypermethylated in T-ALL postulating the loss of cellular identity 10/4/21 | Page 15 Introduction Hypothesis M&M Results-1 Results-2 Summary Author Division 10/4/21 | Rsults-2 DNAm landscape of adult T-ALL 10/4/21 | Page 16 Introduction Hypothesis M&M Results-1 Results-2 Summary T-ALL origin and subtypes Author Division 10/4/21 | 10/4/21 | Page 17 Introduction Hypothesis M&M Results-1 Results-2 Summary Pediatric T-ALL (<15 yo) v/s Young adult T-ALL (>15 yo) Author Division 10/4/21 | Oncogenic events Adult T-ALL Pediatric T-ALL TLX1 overexpression 30% 5-10% TLX3 overexpression 5% 20-25% DNMT3A mutations 10% 0% IDH1/IDH2 mutations 5-7% 0% JAK1/JAK3 mutations 18-30% <3% 10/4/21 | Page 18 Introduction Hypothesis M&M Results-1 Results-2 Summary Rationale: Epigenetic characterization of adult T-ALL Author Division 10/4/21 | • Adult T-ALL is molecularly distinct and understudied • DNAm as prognostic bio-marker • Identification of clinically actionable subgroups Cohort: French GRAALL 2003–2005 clinical trials 10/4/21 | Page 19 Introduction Hypothesis M&M Results-1 Results-2 Summary DNA methylation identifies T-ALL subtypes Author Division 10/4/21 | 10/4/21 | Page 20 Introduction Hypothesis M&M Results-1 Results-2 Summary Maturation arrest stages and oncogenic events Author Division 10/4/21 | 10/4/21 | Page 21 Introduction Hypothesis M&M Results-1 Results-2 Summary Cluster specific somatic variants Author Division 10/4/21 | Fishers’ exact test (FDR < 0.05) C1: DNMT3A, IDH2 C2: PTEN (LoF) C3: STAT5B C4: BCL11B C5: SUZ12 and EZH2 (PRC2 complex), NF1 Adult T-ALL is epigenetically and genetically heterogenous 10/4/21 | Page 22 Introduction Hypothesis M&M Results-1 Results-2 Summary Unexpected identification hyper methylated T-ALL subgroup (C5) Author Division associated 10/4/21 | with poor outcome C5 treatment characteristics: DNMT3A/ HOXA9 Ø Poor prednisone response TAL1 Ø Poor bone marrow response Ø High MRD 10/4/21 | Page 23 Introduction Hypothesis M&M Results-1 Results-2 Summary DNAm as a biomarker: Hypomethylating agent responsive T-ALL Author Division subgroups 10/4/21 | C3 C4 C5 C2 C2 C3 Treated v/s untreated 1.0 Avg. β value 0.8 0.6 0.4 0.2 0.0 UPNT_670_1304_CTR UPNT_670_1308_CTR UPNT_670_1303_AZA UPMT_670_1309_AZA Untreated Treated FDR < 0.05 & |beta| > 20% Hyper-methylated Hypo-methylated Mice work performed by Aurore Touzart 10/4/21 | Page 24 Introduction Hypothesis M&M Results-1 Results-2 Summary DNA methylation predicts maturation arrest stages of epigenetic Author Division clusters 10/4/21 | 2 x 6 intra-thymic cell types (EPIC-arrays) CD34: CD34+ Early thymic progenitors ISP: Immature Single CD4+ DP_TCR-: CD+ CD8+ TCR- CD3- DP_TCR+: CD+ CD8+ TCR+ CD3+ SP4: CD4+ SP8: CD8+ 10/4/21 | Page 25 Introduction Hypothesis M&M Results-1 Results-2 Summary DNA methylation predicts maturation arrest stages of epigenetic Author Division clusters 10/4/21 | 10/4/21 | Page 26 Introduction Hypothesis M&M Results-1 Results-2 Summary Author Division 10/4/21 | 10/4/21 | Page 27 Introduction Hypothesis M&M Results-1 Results-2 Summary Machine learning models for de-novo classification of T-ALL Author Division 10/4/21 | Test cohort Validation cohort Overview of steps (N = 29) (N = 57) Training: N = 86 Test: N = 57 79 CpG classifier EPIC arrays as a diagnostic tool 10/4/21 | Page 28 Introduction Hypothesis M&M Results-1 Results-2 Summary Summary Author Division 10/4/21 | • Five stable molecular subgroups • Novel DNMT3A/IDH2 subgroup • Maturation arrest stages • Stepwise acquisition of DNAm based classifier for genetic drivers de-novo cluster prediction • Prognostic subgroup 10/4/21 | Page 29 Introduction Hypothesis M&M Results-1 Results-2 Summary Author Significance Division 10/4/21 | • Identification of conserved tDMRs (epigentic signature) • DNAm as a biomarker for risk-based stratification • Identification of hypermethylated T-ALL for targeted therapy Limitations • Lack of data from epigenetic marks (ATAC-seq, active and repressive histone signals) • Limited resolution of EPIC arrays (~3% of the methylome) • Limited representation of C1 samples in T-ALL cohort • Smaller training data-set for machine learning models 10/4/21 | Page 30 Introduction Hypothesis M&M Results-1 Results-2 Summary Publications Author Division 10/4/21 | Primary results Touzart A*, Mayakonda A*, Smith C, et al. Epigenetic analysis of patients with T-ALL identifies poor outcomes and a hypomethylating agent-responsive subgroup. Sci Transl Med. 2021. PMID: 34039737 *Equal contribution Method development Mayakonda A, Schönung M, Hey J, et al. Methrix: an R/bioconductor package for systematic aggregation and analysis of bisulfite sequencing data. Bioinformatics. 2020. PMID: 33346800 Outlook DNAm landscape of human thymopoiesis 10/4/21 | Page 31 Introduction Hypothesis M&M Results-1 Results-2 Summary Acknowledgments Author Division 10/4/21 | Vahid Asnafi, Christoph Plass, Aurore Touzart, Dept. of Immunology, Cancer Epigenomics, Cancer Epigenomics, INSERM, Paris. Dkfz, Heidelberg. Dkfz, Heidelberg. Dept. of Immunology, INSERM, Paris. All members of, • Division of cancer epigenomics, DKFZ • Hospital Neckar (INSERM), Paris • Core facility – DKFZ, INSERM 10/4/21 | Page 32 Author Division 10/4/21 | 10/4/21 | Page 33
Enter the password to open this PDF file:
-
-
-
-
-
-
-
-
-
-
-
-