Characterization of Multicellular Niches Supporting Hematopoietic Stem Cells within Distinct Zones

Hematopoietic stem cells (HSCs) possess remarkable self-renewal abilities and multilineage potential, ensuring a lifelong supply of both HSCs and their progeny. In the adult bone marrow (BM), it has been proposed that HSCs reside within a physically confined microenvironment or niche. However, previous studies primarily focused on single-cell-based niche models, yielding valuable yet sometimes conflicting findings. In this report, we identify two distinct niches in the fetal liver (FL): the portal-vessel (PV) niche and the sinusoidal niche. Each of these niches has unique features that regulate the quiescence or differentiation of FL-HSCs. We also present evidence suggesting that distinct niches exist in the bone marrow, playing a role in maintaining BM-HSCs.

We first introduced two spatial transcriptomic methods, Slide-seq and 10x Visium, in our study of mouse FL, as FL is the primary hematopoietic site during the fetal stage. The data revealed the detailed distribution and transcriptomics of HSCs and potential niche cells, including hepatoblasts, endothelial cells, macrophages, megakaryocytes, and mesenchymal stromal cells (MSCs) in the mouse FL. Interestingly, we found that MSCs and hepatoblasts were characterized by enriched N-cadherin expression (N-cad+), with N-cad+ MSCs being particularly abundant in the portal vessel area. Importantly, the majority of FL-HSCs were found in close proximity to N-cad+ MSCs and endothelial cells, indicating that N-cad+ MSCs and endothelial cells play a supportive role in HSC maintenance. Subsequent CellPhoneDB (CPDB) analysis revealed an enriched Cxcl12-Cxcr4 interaction between N-cad+ MSCs and FL-HSCs. We then generated an N-cad^CreER; Cxcl12 mouse model to conditionally knock out the well-studied niche factor, Cxcl12, in N-cad+ cells. Remarkably, deletion of Cxcl12 through N-cad^CreERT induction resulted in the expansion of FL-HSCs with a myeloid bias. Slide-seq further showed that Cxcl12 deletion via N-cad^CreERT induction led to the repositioning of FL-HSCs from the PV region, which is enriched with MSCs, to the hepatic sinusoidal region, which lacks MSCs. These findings suggest the existence of two distinct niches in the FL: the PV niche, which maintains quiescent and multipotential FL-HSCs, and the sinusoidal niche, which supports proliferative FL-HSCs biased towards myeloid lineages.

We then investigated adult bone marrow niches. We transplanted mouse bone marrow HSCs from RFP+ mice into recipient mice with either a wild-type phenotype (WT) or with Cxcl12 deleted from N-cad+ cells (cKO). After 8 weeks, femurs from both groups were collected for immunofluorescence (IF) staining of RFP to determine the distribution of transplanted RFP+ donor cells. Intriguingly, the distribution of donor-derived RFP+ HSPCs was significantly enriched in the trabecular bone area (TBA) in WT recipients. In contrast, in cKO recipients, the RFP+ cells were significantly more prevalent in the central marrow (CM). Cell cycle analysis of donor-derived RFP+ LSK cells showed significantly increased proliferation in cKO recipients compared to WT recipients. Furthermore, in another assay, when the serine protease dipeptidyl peptidase 4 (Dpp4), which truncates Cxcl12 and inhibits its chemotactic activity, was deleted from donor HSPCs, these HSPCs tended to home to the CM in WT recipient mice, suggesting a loss of response to chemotactic signals from the TBA. These findings indicate that distinct niches exist in the bone marrow, with the TBA niche maintaining quiescent HSPCs and the CM niche supporting proliferative HSPCs. To further study the regulatory network in these distinct niches, we are employing spatial transcriptomic techniques, including both 10x Xenium and 10x Visium HD, to reveal the details of HSC and niche cell interactions in each niche.

Taken together, our study reports a paradigm-shifting discovery, moving from a single-cell-based niche model to a zonal regulation model that modulates the cell cycle status and other properties of HSPCs. The key distinction lies not in which cell type functions as a niche, but in the different signaling emanating from these multi-cellular niches across various zones. Additionally, we demonstrated that Cxcl12, a reported quiescence inducer of HSPCs, does not directly affect HSC properties. Instead, Cxcl12 serves as a chemoattractant that regulates HSC localization, which in turn determines HSC activities.

No relevant conflicts of interest to declare.