Fetal liver macrophages contribute to the hematopoietic stem cell niche by controlling granulopoiesis

During embryogenesis, the fetal liver becomes the main hematopoietic organ, where stem and progenitor cells as well as immature and mature immune cells form an intricate cellular network. Hematopoietic stem cells (HSCs) reside in a specialized niche, which is essential for their proliferation and differentiation. However, the cellular and molecular determinants contributing to this fetal HSC niche remain largely unknown. Macrophages are the first differentiated hematopoietic cells found in the developing liver, where they are important for fetal erythropoiesis by promoting erythrocyte maturation and phagocytosing expelled nuclei. Yet, whether macrophages play a role in fetal hematopoiesis beyond serving as a niche for maturing erythroblasts remains elusive. Here, we investigate the heterogeneity of macrophage populations in the fetal liver to define their specific roles during hematopoiesis. Using a single-cell omics approach combined with spatial proteomics and genetic fate-mapping models, we found that fetal liver macrophages cluster into distinct yolk sac-derived subpopulations and that long-term HSCs are interacting preferentially with one of the macrophage subpopulations. Fetal livers lacking macrophages show a delay in erythropoiesis and have an increased number of granulocytes, which can be attributed to transcriptional reprogramming and altered differentiation potential of long-term HSCs. Together, our data provide a detailed map of fetal liver macrophage subpopulations and implicate macrophages as part of the fetal HSC niche. 

Methods
5µm slices of fetal liver from E14.5 wildtype embryos were prepared and used for CODEX staining following the manufacturer’s instructions. Briefly, sections were retrieved from the freezer, let dry on drierite beads, and fixed for 10 min in ice-cold acetone (Sigma Aldrich, St. Louis, MO, USA). After fixation, samples were rehydrated and photobleached twice as described in (Du, Lin et al. 2019). Following photobleaching, sections were blocked and stained with a 20-plex CODEX antibody panel (Table S5 and S6) overnight at 4 °C. After staining, samples were washed, fixed with ice-cold methanol, washed with 1x PBS, and fixed for 20 min with BS3 fixative (Sigma Aldrich, St. Louis, MO, USA). A final washing step with 1x PBS was performed.

A multicycle CODEX experiment was performed following the manufacturer’s instructions. Images were acquired with a Zeiss Axio Observer widefield fluorescence microscope using a 20x objective (NA 0.85) and z-spacing of 1.5µm. The 405, 488, 568, and 647 nm channels were used. After imaging, raw files were exported using the CODEX Instrument Manager (Akoya Biosciences, Marlborough, MA, USA) and processed with CODEX Processor v1.7 (Akoya Biosciences). Image processing included background subtraction using the DAPI signals of the first and last empty cycles of the acquisition, deconvolution, shading correction, and stitching. For cell segmentation, DAPI counterstain was used for object detection, whereas sodium-potassium ATPase antibody staining was used as a membrane marker for delineating the cell shape.

A manual cell classification was performed in CODEX MAV 1.5 (Akoya Biosciences). Annotation of the macrophage clusters was done using the same gating strategy as in flow cytometry, with the difference that F4/80+CD11b+ cells were not gated but F4/80+ Iba1+ cells. HSCs were gated as CD150+ c-Kit+ cells, erythrocytes as CD45-Ter119+ cells, and blood vessels as CD45-CD31+SMA+ cells. After cell classification, Voronoi diagrams were generated in CODEX MAV using the four macrophage clusters, blood vessels, and HSCs as seeds.

LogOddRatio analysis for spatial interactions was performed in CODEX MAV. For this, after cell classification, the four macrophage populations, HSCs, and blood vessels were selected. The selected minimum and maximum distances of interaction were 5 and 50 µm, respectively.

For cellular neighborhood analyses, the .csv files generated with CODEX MAV were exported to CytoMAP (Stoltzfus, Filipek et al. 2020) and the same cell classification was used to annotate the cells. A raster scan with a radius of 50µm was performed to spatially segment the image. To define the cellular neighborhoods based on local composition, a self-organizing map (SOM) clustering algorithm was used, considering only the macrophage populations. Heatmaps were generated to determine the cell composition of each neighborhood. To measure the distances between macrophage clusters and erythrocytes, images were exported to QuPath v0.3. and cells were detected using DAPI signals. Single object classifiers for each marker were trained, and these were used to generate composite classifiers to identify macrophage populations, erythrocytes, and HSCs as before. The distance between the cells of each macrophage cluster and their closest erythrocyte was measured and plotted.

To validate the proximity of macrophage clusters to HSCs, images were exported to QuPath v0.3, cells were segmented, and HSCs were identified, as described above. A circle with a fixed radius of 50µm was drawn and centered on 20 randomly selected HSCs. Next, the same composite classifiers to identify macrophage clusters were applied to the annotated circles, and the number of cells of each macrophage cluster within the defined radius was counted.