Integrating Spatial Transcriptome, Single-Cell Sequencing and Histopathology Reveals Bone Marrow Spatial Heterogenous in AML Patients

Acute myeloid leukemia (AML) is a highly heterogeneous hematologic malignancy characterized by a poor prognosis, primarily due to resistance to primary chemotherapy and a high relapse rate among patients in remission. This underscores the limitations of conventional chemotherapy in achieving a cure for the majority of AML patients. Spatial transcriptome (ST) technology has emerged as a powerful tool for studying the spatial gene expression and cell-cell interaction patterns, yet its application to calcium-rich tissues like bone marrow has been challenging. In this study, we firstly introduced a novel approach for the paraffin embedding and decalcification of bone marrow tissue, which is fully compatible with Visium glass slides. To validate our method, we collected bone marrow tissues form AML patients and performed spatial transcriptomic sequencing alongside single-cell RNA sequencing. We totally got 5364 ± 2049 (AML-b8) and 4892 ± 2096 (AML-b9) unique molecular identifiers (UMIs) per spot, which is significantly higher than the published studies of bone marrow, and equal to those of soft tissues. This demonstrates that our approach overcomes the limitations of spatial transcriptome data collection in calcium-rich tissues. Utilizing marker genes, we accurately mapped the spatial distribution of megakaryocytes (MKs) and bone trabecular in bone marrow, confirmed through HE staining. For the first time, we integrated spatial transcriptome data, single-cell sequencing and histological analysis to provide a comprehensive characterization of bone marrow in AML patients. After confirming the cell type of each spot, we present the maps of bone marrow tissues in AML patients that could provide both spatial information and gene expression level. Furthermore, we found that the proportion of blasts in our ST results more accurately reflects the tumor burden in AML patients compared to bone marrow smear and flow cytometry, due to the preservation of tissue architecture. Our analysis distinguished seven subpopulations within the tumor cells of bone marrow, including leukemia stem cells (LSKs), highlighting the potential of our approach in elucidating the distribution and transcriptomic profiles of LSKs in AML patients. In our ST models, we observed a significant correlation between genes’ expression level and their distance to endosteal niche, such as S100A8, S100A9, CYBB, CHMP3, PACS2, UNC93B1, CMKLR1 and TFRC. Functional analysis of intra-tumor hubs revealed key pathways, including Toll-like receptor cascades, apoptosis process and so on, which are pivotal in bone marrow samples. In conclusion, we present a pioneering method for bone marrow spatial transcriptome analysis, offering a complete and detailed profile of bone marrow in AML patients. This work paves the way for a deeper understanding of AML pathophysiology and may lead to the development of more effective therapeutic strategies.