In Vivo Reprogramming of Adult Liver Sinusoidal Vascular Endothelial Cells into a Hematopoietic Stem and Progenitor Cell Niche

Abstract

Hematopoietic stem and progenitor cells (HSPCs) reside in specialized microenvironments known as the HSPC niche. Vascular endothelial cells (ECs) are one of the most abundant and significant components of the niche, providing known supportive factors such as E-selectin, CXCL12, and SCF. During marrow stress, extramedullary hematopoiesis can occur in the liver or spleen, but the process is inefficient. The sinusoidal vessels of the liver share many similarities with marrow sinusoids. Here, we aim to identify the transcription factor (TF) code that specifies the vascular EC fate in HSPC marrow niche by conducting differential gene expression analysis with the non-niche supportive ECs of the liver. We performed scRNAseq analysis on FACS-isolated ECs from adult zebrafish kidney marrow and liver. We found that these ECs express conserved sinusoidal genes, such as sele, flt4 and gpr182. However, marrow ECs have significantly higher expression of HSPC niche supportive genes, including mrc1a, lyve1b and cxcl12a. We further identified putative TFs that are uniquely upregulated in the marrow ECs, namely tfec, mafbb, foxp4, irf8 and hoxb8a. The expression of these candidate TFs was conserved in the mammalian bone marrow ECs, as shown by relevant mouse bone marrow scRNA-seq and human bone marrow RNA-seq analyses. To determine whether these candidate TFs can functionally specify a HSPC niche EC fate, we selectively overexpressed them using a zebrafish liver sinusoidal EC-specific enhancer in vivo. Upon the overexpression of tfec and mafbb together, we found adult zebrafish liver sinusoidal ECs significantly upregulated key genes known for HSPC niche supportive functions in both embryonic and adult zebrafish HSPC niches, including mrc1a (log2FC=4.9, p<0.005) and lyve1b (log2FC=3.6, p<0.005) and dab2 (log2FC=5.0, p<0.005). Overexpression of TFEC and MAFB in human iPSC-derived ECs showed upregulation of HPSC niche supportive genes that we previously identified from zebrafish in vivo studies, including LYVE1 (log2FC=5.6, p<0.005) and DAB2 (log2FC=1.1, p<0.005) . Notably, previously published HSPC niche genes, including CXCL12(log2FC=5.1 p< 0.005) , PTN (log2FC=2.6 p<0.005) and JAG1 (log2FC=1.7 p<0.005), are also upregulated upon TFEC and MAFB overexpression. Furthermore, gene ontology analysis revealed an enrichment of genes related to cell adhesion (p<0.005), migration (p<0.005), and extravasation pathways (p=0.014). A colony-forming unit (CFU) assay of CD34+CD45+ HSPCs co-cultured with TFEC and MAFB overexpressing hiPSC-derived ECs had significantly greater numbers of CFUs compared to those co-cultured with hiPSC-derived ECs without the TF overexpression: CFU-GEMM (23.3±2.4 vs. 10.0±2.1, p=0.004), CFU-G (89.3±13.4 vs. 46.0±4.2, p=0.027), and CFU-E (42.3±3.3 vs. 26.7±0.9, p=0.021). Our findings revealed a TF code that specifies the HSPC niche supportive fate of vascular ECs and distinguishes vascular ECs in the hematopoietic niche vs. non-niche organs. These results may translate into novel therapies targeting the sinusoidal vascular niches to better stimulate and support human HSPCs.