Learning the histone codes of gene regulation with large genomic windows and three-dimensional chromatin interactions using transformer

The quantitative characterization of the transcriptional control by histone modifications (HMs) has been challenged by many computational studies, but still most of them exploit only partial aspects of intricate mechanisms involved in gene regulation, leaving a room for improvement. We present Chromoformer, a new transformer-based deep learning architecture that achieves the state-of-the-art performance in the quantitative deciphering of the histone codes of gene regulation. The core essence of Chromoformer architecture lies in the three variants of attention operation, each specialized to model individual hierarchy of three-dimensional (3D) transcriptional regulation including (1) histone codes at core promoters, (2) pairwise interaction between a core promoter and a distal cis-regulatory element mediated by 3D chromatin interactions, and (3) the collective effect of the pairwise cis-regulations. In-depth interpretation of the trained model behavior based on attention scores suggests that Chromoformer adaptively exploits the distant dependencies between HMs associated with transcription initiation and elongation. We also demonstrate that the quantitative kinetics of transcription factories and polycomb group bodies, in which the coordinated gene regulation occurs through spatial sequestration of genes with regulatory elements, can be captured by Chromoformer. Together, our study shows the great power of attention-based deep learning as a versatile modeling approach for the complex epigenetic landscape of gene regulation and highlights its potential as an effective toolkit that facilitates scientific discoveries in computational epigenetics.

Polycomb Group Ring Finger Protein 6 suppresses Myc-induced lymphomagenesis

Max is an obligate dimerization partner for the Myc transcription factors and for several repressors, such as Mnt, Mxd1-4 and Mga, collectively thought to antagonize Myc function in transcription and oncogenesis. Mga, in particular, is part of the variant Polycomb group repressive complex PRC1.6. Here, we show that ablation of the distinct PRC1.6 subunit Pcgf6 – but not Mga – accelerates Myc-induced lymphomagenesis in Eµ-myc transgenic mice. Unexpectedly, however, Pcgf6 loss shows no significant impact on transcriptional profiles, in neither pre-tumoral B-cells, nor lymphomas. Altogether, these data unravel an unforeseen, Mga- and PRC1.6-independent tumor suppressor activity of Pcgf6.

Lin28b-Let-7-PRC1 Axis Guides Developmental Maturation of the Hematopoietic System

Polycomb group (PcG) proteins are a well-studied group of chromatin modifiers belonging to one of two distinct multi-protein complexes: Polycomb repressive complex 1 (PRC1) and PRC2. With definitive hematopoiesis, PRCs contribute to many aspects of fetal and adult blood formation. However, it is largely unknown how many of the age-specific effects of PRCs in hematopoiesis are regulated. Here, we show that the definitive hematopoietic stem and progenitor cell (HSPC) compartment is remodeled from the fetus to the neonate and into young adulthood coordinated with changes in mature blood cell output. This process is in part dependent on the PRC1 component Cbx2, which is regulated by the heterochronic Lin28b/let-7 axis. First, we quantified various population of definitive hematopoietic stem cells (HSCs) and multipotent progenitors (MPPs) using midgestation fetal liver (FL, embryonic day 14.5 (E14.5)), newborn bone marrow (BM, postnatal day 0-1), or young adult (postnatal age 6 to 8 weeks) BM. The lymphoid biased multipotent progenitor 4 (MPP4, ~0.9-fold) declined as the mice matured and aged. We also found erythroid-biased MPP2 diminished (~0.7-fold) while myeloid-biased MPP3 increased (~1.7-fold) with maturation. Using isolated long-term (LT) HSCs from these three stages, we found that E14.5 FL (~8.0-fold) and neonatal LT-HSC (~4.0-fold) showed more rapid B-cell reconstitution compared to young adult LT-HSCs upon transplantation. We found that many of these effects were regulated by Lin28b/let-7. Next, we aimed to determine the downstream mediators of Lin28/let-7's effect on HSPCs maturation. By interrogating gene regulatory subnetworks differentially active across mouse HSPC maturation and mining these subnetworks for predicted let-7 target transcripts, we found Cbx2 enriched in E14.5 FL (P=0.003) and adult HSPCs ectopically expressing LIN28B relative to wild-type adult HSPCs. In cell-based assays, we confirmed that let-7 microRNAs directly regulated CBX2 protein levels. Thus, the Lin28b/let-7 axis governs CBX2 protein levels, leading us to hypothesize that this axis exerts its wide-ranging effects on hematopoietic maturation by regulating PRC1 by controlling Cbx2 levels. As CBX2's developmental stage-specific functions have not been investigated, we generated Cbx2-/-embryos and investigated definitive FL hematopoiesis. We observed skewing of myeloerythorid progenitors to an adult-like myeloid-predominant distribution in Cbx2-/- embryos (P=0.0002), and B-cells in Cbx2-/- neonatal spleens were diminished (P=0.04). We further examined this effect using transplanted Cbx2-/- MPP4 from E14.5 FL which resulted in a decreased donor derived B-lymphoid output compared to wild-type littermates (~0.7-fold). To understand the functional role of Cbx2/PRC1 in juvenile hematopoiesis, we next investigated the role of Cbx2 in maintaining histone H2A monoubiquitinylation (H2AK119Ub) - the histone modification placed by PRC1 - in FL HSPCs. In Cbx2-/- FL HSPCs, the global distribution of H2AK119Ub localization did not change, but several specific H2AK119Ub peaks were altered. We observed differential H2AK119Ub abundance associated with a candidate enhancer within the Erg gene, suggestive of control of Erg expression by Lin28b/let-7/Cbx2. We confirmed that this enhancer activated transcription from a minimal promoter (~8-fold). Erg expression was increased in perinatal spleens of Cbx2-/- mice compared to Cbx2+/+ littermates (~4-fold). Moreover, we found that Cbx2 could repress ERG expression as well as other master HSPC transcription factors. Overall, our findings show that the Lin28b/let-7-axis controls developmental stage-specific hematopoietic output through PRC1-mediated chromatin remodeling. These findings demonstrate a key mechanism by which HSPCs alter their properties during developmental maturation with relevance to age-skewed blood disorders. Disclosures No relevant conflicts of interest to declare.

The Pivotal Immunomodulatory and Anti-Inflammatory Effect of Histone-Lysine N-Methyltransferase in the Glioma Microenvironment: Its Biomarker and Therapy Potentials

Enhancer of zeste homolog 2 (EZH2) is a histone-lysine N-methyltransferase that encrypts a member of the Polycomb group (PcG) family. EZH2 forms a repressive chromatin structure which eventually participates in regulating the development as well as lineage propagation of stem cells and glioma progression. Posttranslational modifications are distinct approaches for the adjusted modification of EZH2 in the development of cancer. The amino acid succession of EZH2 protein makes it appropriate for covalent modifications, like phosphorylation, acetylation, O-GlcNAcylation, methylation, ubiquitination, and sumoylation. The glioma microenvironment is a dynamic component that comprises, besides glioma cells and glioma stem cells, a complex network that comprises diverse cell types like endothelial cells, astrocytes, and microglia as well as stromal components, soluble factors, and the extracellular membrane. EZH2 is well recognized as an essential modulator of cell invasion as well as metastasis in glioma. EZH2 oversecretion was implicated in the malfunction of several fundamental signaling pathways like Wnt/β-catenin signaling, Ras and NF-κB signaling, PI3K/AKT signaling, β-adrenergic receptor signaling, and bone morphogenetic protein as well as NOTCH signaling pathways. EZH2 was more secreted in glioblastoma multiforme than in low-grade gliomas as well as extremely secreted in U251 and U87 human glioma cells. Thus, the blockade of EZH2 expression in glioma could be of therapeutic value for patients with glioma. The suppression of EZH2 gene secretion was capable of reversing temozolomide resistance in patients with glioma. EZH2 is a promising therapeutic as well as prognostic biomarker for the treatment of glioma.

Intrinsically disordered sequences in the Polycomb protein Polyhomeotic regulate condensate formation

The Polycomb group (PcG) complex PRC1 localizes in the nucleus in the form of condensed structures called Polycomb bodies. The PRC1 subunit Polyhomeotic (Ph) contains a polymerizing sterile alpha motif (SAM) that is implicated in both PcG body formation and chromatin organization in Drosophila and mammalian cells. A truncated version of Ph containing the SAM (mini-Ph), forms phase separated condensates with DNA or chromatin in vitro, suggesting PcG bodies may form by phase separation. In cells, Ph forms multiple condensates, while mini-Ph forms a single large nuclear condensate. We therefore hypothesize that sequences outside of mini-Ph are required for proper condensate formation. We identified three distinct Intrinsically Disordered Regions (IDRs) in Ph based on sequence composition and complexity. We tested the role of each IDR in Ph condensates using live imaging of transfected Drosophila S2 cells. We find that each IDR uniquely affects Ph SAM-dependent condensate size, number, and morphology.

Polycomb group proteins in cancer: multifaceted functions and strategies for modulation

Polycomb repressive complexes (PRCs) are a heterogenous collection of dozens, if not hundreds, of protein complexes composed of various combinations of subunits. PRCs are transcriptional repressors important for cell-type specificity during development, and as such, are commonly mis-regulated in cancer. PRCs are broadly characterized as PRC1 with histone ubiquitin ligase activity, or PRC2 with histone methyltransferase activity; however, the mechanism by which individual PRCs, particularly the highly diverse set of PRC1s, alter gene expression has not always been clear. Here we review the current understanding of how PRCs act, both individually and together, to establish and maintain gene repression, the biochemical contribution of individual PRC subunits, the mis-regulation of PRC function in different cancers, and the current strategies for modulating PRC activity. Increased mechanistic understanding of PRC function, as well as cancer-specific roles for individual PRC subunits, will uncover better targets and strategies for cancer therapies.

Intrinsically disordered sequences in the Polycomb protein Polyhomeotic regulate condensate formation

The Polycomb group (PcG) complex PRC1 localizes in the nucleus in the form of condensed structures called Polycomb bodies. The PRC1 subunit Polyhomeotic (Ph) contains a polymerizing sterile alpha motif (SAM) that is implicated in both PcG body formation and chromatin organization in Drosophila and mammalian cells. A truncated version of Ph containing the SAM (mini-Ph), forms phase separated condensates with DNA or chromatin in vitro, suggesting PcG bodies may form by phase separation. In cells, Ph forms multiple condensates, while mini-Ph forms a single large nuclear condensate. We therefore hypothesize that sequences outside of mini-Ph are required for proper condensate formation. We identified three distinct Intrinsically Disordered Regions (IDRs) in Ph based on sequence composition and complexity. We tested the role of each IDR in Ph condensates using live imaging of transfected Drosophila S2 cells. We find that each IDR uniquely affects Ph SAM-dependent condensate size, number, and morphology.

Polycomb Requires Chaperonin Containing TCP-1 Subunit 7 for Maintaining Gene Silencing in Drosophila

In metazoans, heritable states of cell type-specific gene expression patterns linked with specialization of various cell types constitute transcriptional cellular memory. Evolutionarily conserved Polycomb group (PcG) and trithorax group (trxG) proteins contribute to the transcriptional cellular memory by maintaining heritable patterns of repressed and active expression states, respectively. Although chromatin structure and modifications appear to play a fundamental role in maintenance of repression by PcG, the precise targeting mechanism and the specificity factors that bind PcG complexes to defined regions in chromosomes remain elusive. Here, we report a serendipitous discovery that uncovers an interplay between Polycomb (Pc) and chaperonin containing T-complex protein 1 (TCP-1) subunit 7 (CCT7) of TCP-1 ring complex (TRiC) chaperonin in Drosophila. CCT7 interacts with Pc at chromatin to maintain repressed states of homeotic and non-homeotic targets of PcG, which supports a strong genetic interaction observed between Pc and CCT7 mutants. Depletion of CCT7 results in dissociation of Pc from chromatin and redistribution of an abundant amount of Pc in cytoplasm. We propose that CCT7 is an important modulator of Pc, which helps Pc recruitment at chromatin, and compromising CCT7 can directly influence an evolutionary conserved epigenetic network that supervises the appropriate cellular identities during development and homeostasis of an organism.

Polycomb Repressive Complex(es) and Their Role in Adult Stem Cells

Populations of resident stem cells (SCs) are responsible for maintaining, repairing, and regenerating adult tissues. In addition to having the capacity to generate all the differentiated cell types of the tissue, adult SCs undergo long periods of quiescence within the niche to maintain themselves. The process of SC renewal and differentiation is tightly regulated for proper tissue regeneration throughout an organisms’ lifetime. Epigenetic regulators, such as the polycomb group (PcG) of proteins have been implicated in modulating gene expression in adult SCs to maintain homeostatic and regenerative balances in adult tissues. In this review, we summarize the recent findings that elucidate the composition and function of the polycomb repressive complex machinery and highlight their role in diverse adult stem cell compartments.

PCGF2 expression associates with survival in triple negative breast cancer.

We mined published microarray data (1) to understand the most significant gene expression differences in the tumors of triple negative breast cancer patients based on survival at time of analysis: dead or alive. We observed significant transcriptome-wide differential expression of polycomb group ring finger 2, encoded by PCGF2 when comparing the primary tumors of triple negative breast cancer patients dead or alive. Importantly, PCGF2 expression was significantly correlated with distant metastasis-free survival in basal subtype breast cancer, a molecular subtype sharing significant overlap with triple negative breast cancer. PCGF2 may be of relevance as a biomarker or as a molecule of interest in understanding the etiology or progression of triple negative breast cancer.