S246: A HUMAN BONE MARROW ORGANOID FOR DISEASE MODELLING AND DRUG SCREENING IN BLOOD CANCERS

Background: Blood cancers are the most common cancer affecting children, and among the top 10 most common cancers in adults, in whom they remain largely incurable. In the era of advanced “omics”, a major bottleneck to the development of novel therapies is target prioritization and validation. There is a pressing need for experimental platforms that model the cellular and molecular complexity of the bone marrow niche, to enable mechanistic studies into the role of the bone marrow microenvironment in the initiation and propagation of blood cancers. Aims: Organoid technologies have been transformative in other settings, but an organoid with adequate homology to native haematopoietic tissues has not yet been reported. We therefore sought to develop organoids that emulate the cellular, molecular and spatial architecture of human bone marrow. Methods: A directed-differentiation protocol was devised in which human iPSCs formed mesodermal aggregates before hydrogel embedding and commitment to vascular and hematopoietic lineages. Multi-modal imaging and single cell RNA-sequencing were used to confirm cellular, molecular and architectural homology to native human hematopoietic tissues. Results: We optimized hydrogel compositions and cytokine cocktails to support the differentiation of a 3D bone marrow perivascular niche. The resulting organoids contained distinct haematopoietic stem/progenitor cells (HSPCs), stromal and myeloid cellular subtypes including pro-platelet forming megakaryocytes and lumen-forming vasculature. 3D whole-organoid imaging revealed a vascular network with sinusoid-like vessels invested with MSCs/fibroblasts (Fig1A), and the extravasation of haematopoietic cells into vessel lumens. Megakaryocytes closely associated with vessels, and extended pro-platelet protrusions into the vasculature with remarkable similarity to native bone marrow. scRNAseq analysis confirmed transcriptional homology to human bone marrow cellular subtypes, and highlighted strong cell-cell communication networks and haematopoietic support from organoid stroma. We first explored the utility of the organoids to model myelofibrosis, an exemplar malignancy that involves cancer-induced bone marrow remodelling and where there is a need for improved models for target prioritization and novel therapies. Treatment of organoids with TGFβ resulted in a dose-dependent increase in hallmarks of fibrosis (Fig 1B), which was effectively reversed using pharmacological agents, enabling drug screening. Given their homology to native bone marrow, we hypothesized that the organoids may support engraftment of primary human cells. Indeed, cells from healthy donors and patients with blood cancers efficiently engrafted throughout the organoids. After 14 days, organoids engrafted with myelofibrosis but not healthy cells showed dramatic fibrotic remodelling (Fig 1C). The organoids also supported primary cells from other myeloid and lymphoid blood cancers, including those notoriously difficult to maintain in vitro – e.g. myeloma cells engrafted into organoids were >90% viable at 10 days following seeding, while rapidly dying within 48 hrs without organoid support. The ability to study primary cells ex vivo overcomes a huge hurdle to translational research in the myeloma field. Image:Summary/Conclusion: This platform is an enabling technology for the interrogation of disease mechanisms in haematological disorders. Target identification and screening using a species-specific, 3D model that can incorporate primary patient cells will reduce dependence on animal models and may accelerate translational research.