Single-cell characterisation of the bone marrow microenvironment and its contribution to acute myeloid leukemia
Ennis, Sarah
Ennis, Sarah
Loading...
Publication Date
2024-04-09
Type
Thesis
Downloads
Citation
Abstract
The human bone marrow is a complex tissue, responsible for replenishing the entire blood system. It comprises many types of hematopoietic as well as non hematopoietic cells that together coordinate the production of billions of blood cells everyday. In the case of acute myeloid leukemia (AML), this process is damaged and abnormal myeloid-lineage blood cell progenitors accumulate in the bone marrow, impairing the production of healthy blood cells and leading to disease. AML is an aggressive cancer with a poor survival rate owing to the propensity of leukemic cells to develop resistance to chemotherapy and a lack of alternative treatment strategies. It is well-known that cell-cell interactions between AML cells and other cell types in the bone marrow are a major driver of drug resistance and these interactions could represent viable therapeutic targets. However, the exact signalling molecules and cell types involved in these interactions are not well-characterised. Thus, the central aim of this thesis was to characterise the roles played by different bone marrow cell types in the initiation and progression of AML. As the bone marrow is an intricate and heterogenous microenvironment, efforts to profile this tissue using traditional bulk-sequencing methods that require the dissociation of tissues, have left gaps in our understanding. Newer technologies such as single-cell RNA-sequencing (scRNA-seq) allow gene expression to be traced back to individual cells and provide a higher-resolution and more granular view of complex tissues. Therefore, we used scRNA-seq to achieve our aim of characterising the human bone marrow. To begin, the study presented in Chapter 2 describes how we compiled a large single-cell gene expression atlas comprising almost 340,000 bone marrow cells from both healthy and AML donors. This atlas allowed to determine the cellular com position of healthy bone marrow and how it changes at key timepoints in AML progression, namely at diagnosis, post-treatment and relapse. It also allowed us to determine how gene expression is dysregulated in each cell type in AML com pared to healthy individuals. Most importantly, we used this dataset to predict ligand-receptor interactions between hematopoietic stem and progenitor cells and other bone marrow cell types and compared the interactome of these cells in healthy and AML donors. This revealed a significant expansion of interactions related to cytokine signalling and cell adhesion in AML and implicated transforming growth factor beta signalling as a potential driver of quiescence in AML cells. Next, in Chapter 3 we aimed to identify the malignant cell population within the scRNA-seq dataset and compared these cells at diagnosis and relapse. We high lighted several pathways, including inflammatory signalling and metabolic pathways that are altered in most patients during disease progression and may explain how these cells are transformed from being drug-sensitive at diagnosis to drug-resistant at relapse. We also used cell-cell interaction analysis to investigate the incoming sig nals to malignant cells and hypothesised how these signals may drive drug resistance. This analysis revealed marked heterogeneity between patients, indicating that there are multiple mechanisms by which malignant cells may acquire drug resistance. Finally, during the course of performing these analyses we developed and pub lished tools to aid in visualising and exploring cell-cell interaction data, which are described in Chapter 4. These tools include an interactive R Shiny application to browse and mine the interactome data that was generated as part of Chapter 2, and an R package called CCPlotR for generating multiple types of publication-ready visualisations of cell-cell interaction data.
Funder
Publisher
NUI Galway