Spatial transcriptomic dissection of the myeloproliferative neoplasm niche uncovers mutant clone interactions and microenvironmental rewiring

Abstract The development of myeloid neoplasms involves complex cell-cell communication and functional changes within the bone marrow niche. Current NGS techniques, such as scRNA-seq, capture transcriptomic changes at the single-cell level but fail to reveal cell-cell spatial context and physical interactions. To address this, we employed spatial transcriptomics approaches to analyze bone marrow samples from patients with myeloproliferative neoplasms (MPNs), including polycythemia vera (PV), essential thrombocythemia (ET), primary myelofibrosis (PMF), and chronic myeloid leukemia (CML).The decalcification process involved in preparing bone marrow core biopsies significantly compromised the quality of the samples for spatial transcriptomic studies. To mitigate this, we conducted Xenium high-plex panel and Visium HD analyses on matched bone marrow core and clot specimens from the same patients. In both platforms, core biopsies showed significantly lower quality than clot sections. In Visium HD, the mean UMI per 8 µm bin was 9.3 for PV core and 48.9 for PV clot; 6.1 for PMF core and 58.7 for clot; 2.9 for ET core and 82.7 for clot; and 6.1 for CML core versus 77.7 for clot. Xenium analysis showed a similar trend, with core samples yielding fewer cells and transcripts. We next performed Robust Cell Type Decomposition (RCTD) cell type annotation using a published scRNA-seq dataset. Comparison of the annotation with the corresponding H&E staining confirmed the accuracy.The JAK2V617F (JAK2VF) mutation is a well-characterized driver of PV. To investigate the spatial behavior of JAK2-mutant cells within the bone marrow niche, we designed custom probes to detect patient-specific mutations. We then performed analysis on a patient sample carrying both JAK2 and TET2 mutations using Xenium with customized probes. Cell population analysis revealed that JAK2-mutant transcripts were enriched in megakaryocytes, erythroid progenitors, and early-stage erythroblasts. We next examined the microenvironment by identifying cells located within 50 µm of JAK2-mutant cells. This revealed an enrichment of myeloid cells and early-stage erythroblasts, with a relative depletion of erythroid progenitors among the neighboring populations. Furthermore, we found that JAK2-mutant cells are also enriched within 50 µm of other JAK2-mutant cells. These data suggest that JAK2-mutant cells preferentially cluster together, and mutant erythroid progenitor cells tend to form separate colonies. Subsequent niche analysis demonstrated that JAK2-mutant cells are preferentially located in focal niches enriched with erythroid cells, macrophages, and HSCs, while being relatively underrepresented in niches enriched with vascular smooth muscle cells (VSMCs), B cells, and osteoblasts.To further investigate gene expression differences within the microenvironment, we compared transcriptomic profiles among JAK2-mutant cells, their neighboring cells, and all other cells in the sample. This analysis revealed distinct patterns of gene expression across the groups. Notably, the neighboring cells around JAK2-mutant cells exhibited significantly elevated expression of genes such as MAEA, PLEK2, ADAM10, and KCNA3, whereas they showed relatively reduced expression of genes including TGFB1, ZUP1, ARRB1, BTK, and others, compared to other populations. Gene Ontology (GO) analysis of the differentially expressed genes revealed enrichment in biological processes related to cell proliferation, phosphatase binding, and protein tyrosine kinase activity.Since the patient also carries a TET2 mutation, we performed a parallel analysis to investigate the spatial distribution of TET2-mutant cells. Cell proximity and neighbor cell enrichment analyses reveal that TET2-mutant cells are enriched in proximity to JAK2-mutant cells. However, TET2-mutant cells did not show self-clustering. Further analysis shows distinct expression patterns between TET2 mutant cells, TET2 mutant neighbor cells, and other cells. GO enrichment analysis showed significant associations with leukocyte immune responses and phosphatase binding activity.Together, our studies reveal that JAK2 and TET2 mutant clones exhibit distinct spatial clustering, preferential niche localization, and induce transcriptional reprogramming of neighboring cells, uncovering a new layer of interaction between mutant clones and their microenvironment in human hematopoiesis.