Decoding functional hematopoietic progenitor cells in the adult human lung

Abstract

Although the bone marrow is the main site of blood cell production in adults, rare pools of hematopoietic stem and progenitor cells have been found in extramedullary organs. In mice, we have previously shown that the lung contains hematopoietic progenitor cells and is a site of platelet production. Here, in the adult human lung, we show that functional hematopoietic precursors reside in the extravascular spaces with a frequency similar to the bone marrow and are capable of proliferation and engraftment in mice. The gene signature of pulmonary and medullary CD34+ hematopoietic progenitors indicates greater baseline activation of immune-, megakaryocyte/platelet-, and erythroid-related pathways in lung progenitors. Spatial transcriptomics mapped blood progenitors in the lung to an alveolar interstitium niche with only a few cells identified in an intravascular location. In human blood samples collected for stem cell transplantation, CD34+ cells with a lung signature enriched the mobilized pool of hematopoietic stem cells. These results identify the lung as a pool for uniquely programmed blood stem and progenitor cells with the potential to support hematopoiesis in humans.

Figure 1. The human lung contains phenotypic hematopoietic progenitors with in vitro proliferation and differentiation capacity.

(A) Pipeline for flow-cytometric immunophenotyping and evaluation of in vitro colony-forming capacity of hematopoietic progenitor cells from BM, lung and PB of organ donors. (B) Normalized flow cytometry plots of BM, PB in the Live/Lin^- gate from a representative donor showing stem cell subsets within the multipotent (MP [CD34⁺CD38^-], light purple) and the hematopoietic progenitor cell (HPC [CD34⁺CD38⁺], light blue) pool. HSC, hematopoietic stem cell; MPP, multipotent progenitor; MLP, multilymphoid progenitor; CMP, common myeloid progenitor; MEP, megakaryocyte-erythroid progenitor; GMP, granulocyte-macrophage progenitor; CFU-Mk, colony-forming-unit megakaryocyte. (C) Composition of hematopoietic progenitor subsets in the BM, PB and lung (n=8). (D) Numbers of hematopoietic progenitor cell subsets in the BM (white), PB (red) and lung (grey) per 10³ Lin^-CD34⁺ cells. N=8 donors, bars indicate mean number of cells ± SD, colors of the dots represent individual donors. ANOVA followed by Sidak’s multiple comparison test, *p<0.03; **p<0.002; ***p<0.0002; ****p<0.0001; ns, not significant. (E) Frequency of HSCs/MPPs and HPCs as a percentage of total nucleated cells in the lung or BM, respectively. Dot colors represent individual donors, bars indicate mean ± SD. Student’s t-test, **p<0.01; ns, not significant. (F) Culture initiating capacity of lung and BM progenitors in MethoCult™ (n=8): Representative colonies (scale bar, 500μm), colony composition and colony quantity for progenitors derived from the BM and lung. Student’s t-test, ****p<0.0001, ^#ANOVA followed by Sidak’s multiple comparison test. CFU, colony-forming unit; BFU-E (purple), burst-forming unit-erythroid; G (orange), granulocyte; M (red), macrophage; GM (pink), granulocyte macrophage; GEMM (black), granulocyte, erythroid, macrophage, megakaryocyte. (G) Culture initiating capacity of lung and BM progenitors in MegaCult™ (n=6): Representative colonies (scale bar, 100μm), colony quantity and colony size for progenitors from the BM and lung. Bar graph represents mean number of colonies ± SD, Student’s t-test, ****p<0.0001. Stacked bars represent mean proportion ± SD, Kruskal-Wallis test, ****p<0.0001. (H) Proportions of cycling (S-G2-M phase (blue), Ki-67⁺DAPI⁺), preparing/growing (G1 (grey), Ki-67+DAPI^-) and resting cells (G0 (black), Ki-67^-DAPI^-) in the HSC/MPP and HPC pool from BM, PB and lung (n=7). Stacked bars represent mean proportion ± SD, ANOVA followed by Sidak’s multiple comparison test, **p<0.01, *p<0.05. For comparisons not indicated, no statistically significant differences were observed.

Figure 2. Human lung-derived hematopoietic progenitors have in vivo engraftment potential.

(A) Experimental procedure to compare the in vivo engraftment efficiency of lung and BM hematopoietic progenitors: After sublethal irradiation, NSG-SGM3 mice were injected i.v. with either 1.5 x10⁶ Live/Lin^- human lung or BM cells. Ten weeks post transplantation, the BM, PB and lung of recipient mice were collected and investigated for human cell engraftment. The graft properties are summarized in the table below. (B) Representative flow plots of human myeloid (hCD45⁺⁺, hCD33⁺) and lymphoid (hCD45⁺⁺, hCD19⁺) cell engraftment in the BM of a recipient mouse (+, upper panel) and non-transplanted control (-, lower panel). (C) Engraftment efficiency of human cells after xenotransplantation measured by flow cytometry. Bar graphs represent the percentage of hCD45⁺⁺ cell engraftment in the BM, lung, and PB of recipient mice after transplantation of HSPCs from human BM (black) or lung (white). Mean ± SD; Student’s t-test values are given, individual data points for each animal are plotted as gray dots. Blue dotted line indicates threshold for positive engraftment. (D) Detection of human cells in the BM (left panel) and lung (right panel) of recipient mice by immunostaining against human CD45 (hCD45). Scale bar, 50μm. (E) Representative flow cytometry plots of human erythroid (CD45^-, hGlyA⁺ or hCD71⁺) cell engraftment in the BM of a recipient mouse (+, upper panel) and non-transplanted control (-, lower panel). (F) Human erythroid cell expansion (CD45^-, hGlyA⁺ or hCD71⁺) in the BM, lung and PB of recipient mice measured by flow cytometry. Bar graphs representing the percentage of human CD45^-GlyA⁺CD71⁺ cells in BM, lung, and PB after transplantation of HSPCs from human BM (black) or lung (white). Mean ± SD; Student’s t-test values are given, individual data points for each animal are plotted as gray dots. Blue dotted line indicates threshold for positive engraftment. (G) Detection of human erythroid cells in the BM (left panel) and lung (right panel) of recipient mice by immunostaining against human GlyA (hGlyA). Scale bar, 50μm. (H) Proportion of lineage expansion across all human cells detected in the BM, lung and PB, respectively. Stacked bars represent mean proportion ± SD, ANOVA followed by Sidak’s multiple comparison test, ns, not significant.

Figure 3. Comparative transcriptomic analysis of lung and BM HSCs reveals shared and unique gene expression profiles.

(A) UMAP projection of BM and lung Lin^-CD34⁺ progenitor hierarchy from 8 human donors highlighting the HSC/MPP cluster (purple). The pie graph indicates the proportion of cells from the BM (blue) and lung (red) within the multipotent progenitor subset. HSC/MPP, hematopoietic stem cell/multipotent progenitor; My, myeloid cell; Eo/Ba/Ma, eosinophil/basophil/mast cell progenitor; MultiLin, multi-lineage; EMP, erythroid megakaryocytic progenitor; earlyEry, early erythroid progenitor; lateEry, late erythroid progenitor; prog/stroma mix, progenitor stroma cell mix; nd, not determined. (B) Grouping of gene expression patterns into modules using Monocle3. Aggregate expression values of genes in the module highly specific for HSCs (Supplemental Figure 7) are shown individually for the BM and lung. (C) Pseudotime calculation for each cell within the BM and lung using Monocle3 to infer progression through different cellular differentiation to provide insights into the developmental trajectory. (D) Scatter plot of median gene expression of cells in the HSC/MPP cluster from the lung (red) and BM (blue) to visualize consistent (grey) and differentially (highlighted) expressed genes. (E) Venn diagram and top 10 differentially expressed genes. The number in each circle represents the amount of differentially expressed genes between lung (red) and BM (blue), the overlapping number indicates mutual differentially expressed genes based on the Wilcoxon ran-sum test in Seurat’s ‘FindMarkers’ function. (F) Box and violin plots showing the distribution of selected genes upregulated in pulmonary hematopoietic progenitor cells. Wilcoxon adjusted p-value <0.001. (G) Selection of marker genes shared between lung and BM as box and violin plots, respectively. ns, not significant. (H) Box and violin plots showing the distribution of markers genes upregulated in BM HSCs, Wilcoxon adjusted p-value <0.001. (I) T.statistic of ssGSEA scores for selected gene sets (Hallmark, Reactome, Biocarta, KEGG) enriched in pulmonary HSCs categorized by recurring functions. EPO, erythropoietin; ECM, extracellular matrix, FDR, false discovery rate; GFR, growth factor receptor; ssGSEA, single-sample Gene Set Enrichment Analysis. (J) Enrichment ridge plots comparing the distribution of enrichment scores in HSCs from lung (red) and BM (blue) of selected Reactome pathways. Rug plots indicate the scores of individual cells along the ridge plot. P-values are given in the figure, FDR R-HSA-9027277 = 2.38*10^-4; FDR R-HSA-9006335 = 0.09; FDR R-HSA-8936459 = 0.03; R-HSA-76002 = 2.03*10^-10. (K) Enrichment ridge plots showing the distribution of enrichment scores in lung (red) and BM (blue) with individual cell placement on the rug plot to compare selected GOBP (Gene Ontology Biological Process) gene set enrichments. P-values are given in the figure, FDR GO:00025 = 1.77*10^-6; FDR GO:0001816 = 2.42*10^-7; FDR GO:0006955 = 6.70*10^-7; GO:0050729 = 2.02*10^-8.

Figure 4. Spatial mapping of phenotypic CD34⁺ HSPCs in the lung.

(A) Immunofluorescence imaging of putative HSCs (Lin^-/CD34⁺/CD90⁺) in the human lung and BM. Left panel: Representative section of lung showing a Lin^-/CD34⁺/CD90⁺ in the interstitial space. Right panel: Representative section of BM showing two Lin^-/CD34⁺/CD90⁺ cells. (B) Spatial transcriptomics analysis workflow. smFISH was performed to visualize gene expression in human lung tissue. Transcripts were assigned to individual cells after cell segmentation and cells were annotated based on marker gene expression (Supplemental Figure 10A-D). HSPC candidate cells were computationally identified based on their gene signature and visually validated (Supplemental Figure 10E-F, Supplemental Figure 11A-B). (C) Representative image of a putative HSPC in its pulmonary niche. Upper panel (left to right): DAPI staining, QuPath segmentation, zoom on putative HSC (arrow). Selected transcripts are shown. Lower panel (left to right): all transcripts, pseudo-coloring of cell types in the lung tissue based on marker clustering (Supplemental Figure 10). Zoom on putative HSPC in niche. Scale bar, 250 μm. (D) Anatomic location of candidate cells in the lung. Representative images of phenotypic HSPCs in four major locations (alveolar interstitium, peribronchial, perivascular or intravascular) and proportion of cells in each location. Alv, alveolar space; br, bronchus; vasc, vasculature. Scale bar, 150μm. (E) Squidpy co-occurrence score computed every 2 μm between putative HSPCs and the rest of the clusters across lung tissue sections from 4 organ donors. High score values indicate greater co-occurrence probability; endothelial cells (red) co-occur with the HSPCs at short distances. (F) Pie graphs showing the proportion of neighboring cells within a radius of 20 μm from the putative HSPCs in the major anatomic locations.

Figure 5. HSCs with pulmonary signatures are mobilized during apheresis collections for transplantation.

(A) Peripheral blood stem cells of 8 healthy donors given G-CSF for mobilization were collected via apheresis and cryopreserved (Sampling). Live/Lin^-/CD34⁺ cells were flow sorted and encapsulated, 10x Chromium™ Single 3’ v2 libraries were prepared, pooled and sequenced (scRNAseq). For donor demultiplexing via SNPs, bulkRNAseq was performed on Live/Lin⁺ cells (bulkRNAseq). Following Louvain clustering and annotation, phenotypic HSCs were subsetted from the mobilized pool and examined for their expression of canonical, lung and BM HSC signature genes using UCell. (B) Basic demographics of the donor population. (C) Batch corrected UMAP representation highlighting the HSC/MPP cluster (red), arrows indicate developmental trajectory into more committed progenies (erythroid, myeloid, lymphoid). (D) Total number of progenitor cells and number of HSCs per donor. Fraction of HSCs among all cells is given in red. (E) Pie graph showing the proportions of medullary (blue) and extramedullary (lung, red; other, grey) signatures in the HSC fraction of apheresis samples. (F) Bar graph showing the absolute numbers of HSCs across all donor that had a unique BM (blue) or lung (red) signature, cells that exhibited features of both BM and lung (violet) and cells that could not be assigned to either of these categories (grey). (G) Box and Whisker plot representing the percentage of medullary and extramedullary signatures identified the HSC population. Dots represent the individual allogeneic donors.