PROJECT SUMMARY

Myelodysplastic syndromes (MDS) are clonal disorders of hematopoietic stem and progenitor cells. Nuclear morphology dysplasia in one or multiple blood lineages, such as nuclear hypersegmentation or hypolobulation in neutrophils, are hallmarks of MDS, indicating that nuclear structure alteration is pervasive in these diseases. We hypothesize that: (i) the genome is dynamically organized during normal hematopoietic differentiation, and (ii) nuclear morphology changes in MDS reflect underlying derangements in genome spatial organization, which contribute to malignant transformation.

The goal of this project is to dissect the relationship between the nuclear structure of MDS cells and the function of their genomes, and ultimately to reveal a causal connection between MDS pathogenesis and disruptions in three-dimensional (3D) genome organization. Toward this goal, we are using high- throughput and single-cell multi-omics approaches to construct epigenetic and transcriptomic molecular atlas of blood cells during normal and malignant hematopoietic differentiation. During the past three years, we have made the following progresses in this project.

1. We have generated high-resolution reference chromatin interaction maps for the two frequently dysplastic lineages (erythroid and granulocytic) in MDS These lineage- and stage-specific maps of healthy human primary cells have revealed dynamic re-wiring of the multiscale 3D genome features during terminal erythropoiesis and granulopoiesis. In addition, we found that the highly expressed cell type-specific genes form lineage- and developmental stage-specific local chromatin domains during normal hematopoiesis. These findings indicated that the temporal- and lineage-specific organization of transcription programs during lineage maturation in the hematopoietic tree is encoded in the multi-scale genome architectural features.
2. To determine the 3D genome derangements in MDS cells at the single-cell level, we have adopted our single-cell Hi-C method, sci-Hi-C, to human primary blood cells, and generated single-cell chromatin interaction profiles in thousands of neutrophils from four healthy donors and three MDS patients. We have found that the 3D genome organization in neutrophils is unique from other cell types and is featured with inter-chromosomal interaction hubs that are enriched for transcriptionally active nuclear sub-compartments close to nuclear However, we have also found that these sci-Hi-C datasets were too sparse and were unable to identify single-cell chromatin domains from them.
3. We have developed a new scHi-C method and successfully adapted it to human PBMCs.
4. We have successfully adapted our newly developed CARE-seq method to primary blood CARE- seq is a multimodal omics tool that enables us to concurrently mapping the 3D genome organization and gene expression in the same single cell in a massively multiplex fashion. The methodological resolution of the Hi-C part of CARE-seq is more than 20 times higher than that of our sci-Hi-C method and thereby enabling us to identify multiscale single-cell 3D genome features (e.g., chromatin compartments, domains and loops). Empowered by the cutting-edge machine learning- based computational tools developed by our collaborator, Dr. Jian Ma. at the Carnegie Mellon University, CARE-seq enables us to link 3D genome structure to RNA expression at the single-cell level. To this end, we have successfully used CARE-seq to dissect cellular components and to identify cell type-specific 3D features at the single-cell level in the complex tissue, mouse brain.

We have recruited six healthy donors and one MDS patient, and isolated neutrophils and PBMCs from these individuals. With these samples, after several months’ extensive optimization, we have made significant progress in adapting CARE-seq to primary blood cells. To this end, we have generated thousands of single-cell 3D genome and matched gene expression datasets with normal human bone marrow cells for studying genome structure and function relationships in hematopoiesis, and integrative data analysis is undergoing. In the next step, we plan to use these single-cell multi- to interrogate the relationship between the nuclear structure and the function of the genomes in MDS cells.