HMGA1 Chromatin Regulators Disrupt 3D Genome Architecture and Chromatin Looping to Repress MHC II Genes and Drive Immune Evasion during MPN Progression

Introduction: Myeloproliferative neoplasms (MPN) are clonal blood disorders with hyperactive JAK/STAT signaling and an increased risk of transformation to acute myeloid leukemia (AML). Importantly, AML in MPN is almost universally fatal and targetable mechanisms remain elusive. We discovered that High Mobility Group A1 (HMGA1) chromatin regulators are required for MPN progression and leukemic transformation in JAK2^V617F mouse models (Li et al, Blood 2022). Immune escape is a fundamental mechanism of tumor progression in solid tumors, although its role in MPN is unknown. We therefore sought to test the hypotheses that: 1) HMGA1 drives MPN progression by activating gene networks involved in immune evasion, 2) Targeting HMGA1 networks with therapy in advanced MPN will trigger an anti-tumor immune attack.

Methods: To dissect mechanisms underlying HMGA1, we performed multiomic sequencing (seq) approaches, including RNAseq, chromatin immunoprecipitation (ChIPseq), and assays of chromatin accessibility (ATACseq) in JAK2^V617F mutant AML cell lines (DAMI, SET-2, UKE-1). HMGA1 was inactivated via CRISPR or short hairpin RNA (shRNA). To examine three dimensional (3D) genome architecture, we employed chromatin conformation capture (3C). ChIP was used to detect chromatin binding of repressive histones and CCCTC-binding factor (CTCF) complexes. To define pathways relevant to MPN patients, we assessed bulk and single cell (sc) RNAseq of peripheral blood mononuclear cells (PBMCs) and CD34⁺ cells. Artificial intelligence (AI) was used to identify drugs targeting HMGA1.

Results: To elucidate transcriptional networks governed by HMGA1 in MPN, we integrated multiomic approaches (RNA/ChIP/ATACseq) in JAK2^V617F mutant AML cells, which revealed that HMGA1 represses gene networks involved in interferon γ signaling and antigen presentation (AP), including genes encoding major histocompatibility complex (MHC) class I and II proteins. HMGA1 also represses the CD74 gene, which encodes the MHC II invariant chain required for MHC II assembly at the cell surface. Silencing HMGA1 (via CRISPR or shRNA) results in up-regulation of CD74 and MHC genes, with greatest induction of MHC II (HLA-DRA, DRB1, DPA1, DPB1). Similarly, HMGA1 depletion increases MHC II and CD74 protein levels (by flow cytometry or immunoblot). To determine how HMGA1 represses AP genes, we examined 3D genome architecture at the MHC II locus. HMGA1 occupies the MHC II super-enhancer and promoter regions which control MHC II gene expression. Further, HMGA1 occupancy associates with chromatin compaction (by ATACseq) and repressive histone marks (by ChIPseq). Surprisingly, HMGA1 silencing “opens” chromatin at the MHC II super-enhancer and promoter regions, thus allowing for binding of activating histones and CTCF. Importantly, CTCF is a genome organizer that forms complexes to orchestrate chromatin looping to activate MHC II gene expression. Using 3C and ChIP, we discovered that HMGA1 chromatin occupancy disrupts CTCF binding and chromatin looping at MHC II. By contrast, HMGA1 depletion rescues CTCF-mediated chromatin looping and AP gene activation. Using AI, we identified the histone acetylase inhibitor, entinostat, as a drug predicted to target HMGA1 gene networks. Entinostat up-regulates AP gene expression in MPN cells, recapitulating effects of HMGA1 silencing. To test the functional significance of HMGA1 in immune evasion, we performed co-culture experiments with T cells and MPN AML cells. HMGA1 silencing results in T cell activation and cytotoxic T cell killing of JAK2^V617F AML cells. To examine the relevance of this pathway in patients, we interrogated bulk and scRNAseq from PBMC and CD34⁺ cells, respectively, which demonstrate that HMGA1 is up-regulated and associated with repression of MHC II genes in MPN AML, particularly in AML with biallelic TP53 mutations.

Conclusions: We discovered a novel epigenetic program whereby HMGA1 disrupts 3D genome architecture at the MHC II locus to repress AP genes and drive immune evasion during MPN progression. We also define a new mechanism of immune evasion by HMGA1 through chromatin compaction and repressive histone marks that disrupt CTCF-mediated loop formation required to activate MHC II genes. Both HMGA1 depletion or entinostat up-regulate MHC II to activate cytotoxic T cells, thus illuminating HMGA1 as a therapeutic target to stimulate an immune attack and prevent MPN progression.