CD116+ fetal precursors migrate to the perinatal lung and give rise to human alveolar macrophages

Abstract

Despite their importance in lung health and disease, it remains unknown how human alveolar macrophages develop early in life. Here we define the ontogeny of human alveolar macrophages from embryonic progenitors in vivo, using a humanized mouse model expressing human cytokines (MISTRG mice). We identified alveolar macrophage progenitors in human fetal liver that expressed the GM-CSF receptor CD116 and the transcription factor MYB. Transplantation experiments in MISTRG mice established a precursor-product relationship between CD34-CD116+ fetal liver cells and human alveolar macrophages in vivo. Moreover, we discovered circulating CD116+CD64-CD115+ macrophage precursors that migrated from the liver to the lung. Similar precursors were present in human fetal lung and expressed the chemokine receptor CX3CR1. Fetal CD116+CD64- macrophage precursors had a proliferative gene signature, outcompeted adult precursors in occupying the perinatal alveolar niche, and developed into functional alveolar macrophages. The discovery of the fetal alveolar macrophage progenitor advances our understanding of human macrophage origin and ontogeny.

Graphical abstract

Figure 1.

Candidate alveolar macrophage progenitors are present in human fetal liver. (A) Flow cytometry analysis of CD116-expressing cells within the CD34⁻ fraction of human fetal liver (17 wk of gestation). See Fig. S1 for gating strategy. After excluding cells expressing Lin markers (CD3, TCRαβ, TCRγδ, CD19, CD20, CD56, CD94, NKp46, and CD66abce), CD45⁺CD116⁺ cells were divided into CD64⁺ (population 1) and CD64⁻ (population 2) subsets. (B) Flow cytometry of CD116⁺CD64⁺ cells as gated in A. CD116⁺CD64⁺ cells were separated into CD88⁻CD1c/CD141^hi (population 1) and CD88⁺CD1c/CD141^lo-mid (population 2) subsets, corresponding to dendritic cells (CD11c⁺HLA-DR^hi) and monocytes (CD11b⁺HLA-DR⁺CD206⁻CD169⁻) as well as macrophages (CD11b⁺HLA-DR⁺CD206⁺CD169^mid-hi), respectively. (C) Flow cytometry of CD116⁺CD64⁻ cells as gated in A. Population 1 (CD88⁻CD1c/CD141⁺) contained CD11c⁺HLA-DR^hi dendritic cells. Population 2 (CD88⁻CD1c/CD141⁻) consisted of CD11c⁻HLA-DR^mid potential precursors. (D) Frequencies of the indicated populations in human fetal liver. DCs, dendritic cells; hi, high; lo-mid, low-mid; MØ, macrophages. Data are represented as mean ± SEM. Data (A–C) show one fetal liver sample (17 wk of gestation) representative of 12 samples from two independent experiments. Data (D) are pooled from two independent experiments with 12 fetal liver samples at 15–23 wk of gestation.

Figure S1.

Gating strategy to identify candidate macrophage precursors in human fetal liver. Single-cell suspensions from fetal liver (17 wk gestation) were first gated on live CD45⁺Lin⁻CD34⁻ single cells before gating on CD116⁺ cells as shown in Fig. 1 A. Lin markers were CD3, TCRαβ, TCRγδ, CD19, CD20, CD56, CD94, NKp46, and CD66abce. FSC-A, forward scatter area; FSC-H, forward scatter height; SSC-A, side scatter area; SSC-H, side scatter height. Data are representative of two independent experiments with 12 fetal liver samples at 15–23 wk of gestation. The red numbers indicate the order of the flow cytometry plots (sequential gating of cells).

Figure 2.

Gene signatures of CD116⁺CD64⁻ fetal precursor-like cells, CD116⁺CD64⁺ fetal monocytes, and adult CD14⁺CD16⁻ blood monocytes. (A) Experimental outline to define the gene expression profiles of the indicated human cell populations. Cartoon was adapted from Servier Medical Art. (B) Volcano plots of differentially expressed genes (DEGs) between the indicated cell populations. Fold change is plotted versus -log₁₀ P value (not corrected for multiple testing). DEGs with a fold change ≥2 and an FDR-corrected P value ≤0.05 are highlighted in blue and red. (C) Bar graphs showing expression of selected genes in CD116⁺CD64⁻ fetal precursor-like cells, CD116⁺CD64⁺ fetal monocytes, and adult CD14⁺CD16⁻ blood monocytes. Data are represented as mean ± SEM. (D) Gene ontology over-representation analysis of DEGs up-regulated in CD116⁺CD64⁻ fetal precursor-like cells compared with adult CD14⁺CD16⁻ blood monocytes and CD116⁺CD64⁺ fetal monocytes. AM, alveolar macrophages; GCAS, GeneChip Array Station; HTA, Human Transcriptome Array; lo-mid, low-mid. Data (B–D) are from a single microarray experiment with three replicates per cell population (isolated from individual fetal liver or blood samples) obtained from four independent cell sorting experiments.

Figure 3.

CD116-expressing fetal liver cells generate human alveolar macrophages in vivo. (A) Intranasal transfer of purified human Lin⁻CD116⁻, Lin⁺CD116⁻, and Lin⁻CD116⁺ populations from CD34⁻ fetal liver cells into newborn MISTRG mice. Lin markers were CD3, CD19, CD56, NKp46, and CD66abce. Control mice received PBS only. Cartoon was adapted from Servier Medical Art. (B and C) Flow cytometry of human CD45⁺CD11b⁺HLA-DR⁺CD206⁺CD169⁺ macrophages in lung tissue of MISTRG mice at 10 wk (B) and 24 wk (C) after transfer of the different fetal liver cell subsets. See Fig. S2 A for gating strategy. (D) Number of human alveolar macrophages (AM) in MISTRG mice at 10 and 24 wk after transplantation. (E) Immunohistochemistry of lung sections from MISTRG mice 24 wk after transplantation with human Lin⁻CD116⁺ fetal liver cells. Lung sections were stained with anti-human CD68 antibody (brown). Scale bar is 20 µm. (F) Amounts of total protein in the BAL fluid of MISTRG mice 24 wk after transplantation with human fetal liver cells. Control mice received PBS only or were transplanted with human CD34⁺ HSPCs from cord blood by intrahepatic injection (i.h.). aCD68, anti-CD68. Data are represented as mean ± SEM. *, P < 0.05; **, P < 0.01 (one-way ANOVA with Tukey’s multiple comparison post hoc test). Data (B and C) show one lung sample representative of three to five samples (individual mice) per group and time point from three independent experiments. Data (E) show one lung sample representative of four samples (individual mice) from three independent experiments. Data (D and F) are pooled from three independent experiments with n = 3–5 per group and time point (D) or n = 5–9 per group (F).

Figure S2.

Gating strategy to identify human hematopoietic cells in MISTRG mice. (A) Single-cell suspensions from the lung of MISTRG mice were gated on human CD45⁺ single cells before gating on human lung macrophages in Fig. 3, B and C, Fig. 4, B and C, and Fig. 5 C. hCD45, human CD45; mCD45, mouse CD45. (B) Blood cells from MISTRG mice were gated on human CD45⁺ single cells before gating on circulating human macrophage precursors in Fig. 4 G and Fig. 6 A. FSC-A, forward scatter area; FSC-H, forward scatter height; FSC-W, forward scatter width; hCD45, human CD45; mCD45, mouse CD45; SSC-A, side scatter area. Data (A) show one lung sample representative of three to five samples (individual mice) from three independent experiments. Data (B) show one blood sample representative of seven to nine samples (individual mice) from two independent experiments. The red numbers indicate the order of the flow cytometry plots (sequential gating of cells).

Figure S3.

Controls for immunohistochemistry of lung sections. (A and B) Immunohistochemistry of lung sections from MISTRG mice 24 wk after intranasal (A) or intrahepatic (B) transplantation with the indicated cell populations (as in Fig. 3 E and Fig. 4 E). Control mice received PBS or no cells (naive). Lung sections were stained with anti-human CD68 antibody (brown). Scale bars are 20 µm. aCD68, anti-CD68. Data (A and B) show one lung sample representative of three to four samples (individual mice) from three independent experiments (A) or representative of two samples (individual mice) from two independent experiments (B).

Figure S4.

Human interstitial lung macrophages do not develop in MISTRG mice transplanted with CD116⁺ fetal liver cells. (A and B) Flow cytometry of human interstitial macrophages in lung tissue from MISTRG mice 24 wk after transplantation with human CD116⁺ fetal liver cells (A) or CD34⁺ HSPCs (B). Interstitial lung macrophages were gated as CD45⁺CD11b⁺CD14⁺CD16⁻CD64⁺CD206⁺HLA-DR⁺ cells. FSC-A, forward scatter area; hCD45, human CD45; MØ, macrophages. Data (A and B) show one lung sample representative of three to five samples (individual mice) per group from three independent experiments.

Figure 4.

CD116⁺ fetal liver cells are able to migrate to the lung and differentiate into human alveolar macrophages. (A) Intrahepatic injection of purified human Lin^+/−CD116⁻ and Lin⁻CD116⁺ populations from CD34⁻ fetal liver cells into newborn MISTRG mice. Lin markers were CD3, CD19, CD56, NKp46, and CD66abce. Cartoon was adapted from Servier Medical Art. (B and C) Flow cytometry of MISTRG lungs 7 wk (B) and 24 wk (C) after injection of human cells. Control mice were not injected with cells (naive). See Fig. S2 A for gating strategy. (D) Number of human alveolar macrophages (AM) in MISTRG mice at 10 and 24 wk after transplantation. (E) Immunohistochemistry of lung sections from MISTRG mice 24 wk after transplantation with human Lin⁻CD116⁺ fetal liver cells. Lung sections were stained with anti-human CD68 antibody (brown). Scale bar is 20 µm. (F) Amounts of total protein in the BAL fluid of MISTRG mice 24 wk after transplantation with human Lin^-/+CD116⁻ or Lin⁻CD116⁺ fetal liver cells. Control mice were transplanted with human CD34⁺ HSPCs from cord blood or were not injected with cells (naive). Data are represented as mean ± SEM. Picture on the right shows BAL fluid obtained from MISTRG mice transplanted with Lin⁻CD116⁺ fetal liver cells or not transplanted with any human cells (naive). (G) Flow cytometry of MISTRG blood 7 wk after transplantation with the indicated human cells. Control mice were not injected with cells (naive). See Fig. S2 B for gating strategy. aCD68, anti-CD68; hCD45, human CD45. Data are represented as mean ± SEM. **, P < 0.01; ****, P < 0.0001 (one-way ANOVA with Tukey’s multiple comparison post hoc test). Data (B and C) show one lung sample representative of two to eight samples (individual mice) per group and time point from three independent experiments. Data (E) show one lung sample representative of five samples (individual mice) from three independent experiments. Data (D and F) are pooled from three independent experiments with n = 2–8 per group and time point (D) or n = 3–5 per group (F). Data (G) show one blood sample representative of seven to nine samples (individual mice) from two independent experiments.

Figure 5.

CD116⁺CD64⁻ fetal liver cells give rise to human alveolar macrophages in vivo. (A) Intrahepatic injection of purified human Lin⁻CD116⁺CD64⁻ and Lin⁻CD116⁺CD64⁺ populations from CD34⁻ fetal liver cells into newborn MISTRG mice. Lin markers were CD3, CD19, CD56, NKp46, and CD66abce. Cartoon was adapted from Servier Medical Art. (B) Immunohistochemistry of lung sections from MISTRG mice 24 wk after transplantation with either human Lin⁻CD116⁺CD64⁻ or Lin⁻CD116⁺CD64⁺ fetal liver cells. Lung sections were stained with anti-human CD68 antibody (brown). Scale bars are 20 µm. (C) Flow cytometry of human CD45⁺CD11b⁺HLA-DR⁺CD206⁺CD169⁺ lung macrophages in MISTRG mice 20 wk after transplantation. See Fig. S2 A for gating strategy. (D) Number of human alveolar macrophages (AM) derived from Lin⁻CD116⁺CD64⁻ and Lin⁻CD116⁺CD64⁺ fetal liver cells in MISTRG mice at 20 wk after transplantation. (E) Amounts of total protein in the BAL fluid of MISTRG 20–24 wk after transplantation with human Lin⁻CD116⁺CD64⁻ or Lin⁻CD116⁺CD64⁺ fetal liver cells. Newborn control MISTRG mice were either not transplanted or transplanted with human CD34⁺ HSPCs from cord blood. Picture on the right shows BAL fluid obtained from MISTRG mice transplanted with either Lin⁻CD116⁺CD64⁻ or Lin⁻CD116⁺CD64⁺ fetal liver cells. aCD68, anti-CD68; hCD45, human CD45; hi, high. Data are represented as mean ± SEM. *, P < 0.05 (unpaired Student’s t test), or **, P < 0.01; ***, P < 0.001 (one-way ANOVA with Tukey’s multiple comparison post hoc test). Data (B) show one lung sample representative of six to seven samples (individual mice) per group from two independent experiments. Data (C) show one lung sample representative of six to seven samples (individual mice) per group from two independent experiments. Data (C–E) are pooled from two independent experiments with n = 6 or 7 (C and D) or n = 7 or 8 (E) per group.

Figure 6.

Circulating CD116⁺CD64⁻CD14⁻CD115⁺ macrophage precursors migrate to the lung. (A) Flow cytometry of MISTRG blood 20 wk after transplantation with human Lin⁻CD116⁺CD64⁻ or Lin⁻CD116⁺CD64⁺ fetal liver cells. See Fig. S2 B for gating strategy. Frequencies of circulating CD45⁺CD116⁺CD64⁻CD14⁻ cells after transplantation are shown on the right. (B) Fate-mapping of circulating cells in MISTRG mice transplanted with human Lin⁻CD116⁺ fetal liver cells. Blood cells were labeled by the i.v. injection of PE-conjugated fluorescent beads. Cartoon was adapted from Servier Medical Art. (C) Flow-cytometric analysis and frequency of bead⁺ cells in the lung and blood of MISTRG mice at day 7 after bead injection. hCD45, human CD45. Data are represented as mean ± SEM. Data (A) show one blood sample representative of six or seven samples (individual mice) per group from two independent experiments. Data (C) show one blood and lung sample representative of two to six samples per group from two independent experiments. Bar graphs (A and C) show data pooled from two independent experiments with n = 6–7 (A) or n = 2–6 (C) per group.

Figure 7.

Human fetal liver and lung contain CD116⁺CD64⁻CD115⁺CX3CR1⁺ lung macrophage precursors. (A and B) Flow cytometry analysis of CD116-expressing macrophage precursors in human fetal liver (A) and fetal lung (B) at wk 21 and 22 of gestation, respectively. After pregating on CD45⁺CD34⁻ cells (see Fig. S5 for gating strategy), macrophage precursors were gated as shown. Lin markers were CD3, TCRαβ, TCRγδ, CD19, CD20, CD56, CD94, NKp46, and CD66abce. Data (A and B) are representative of fetal liver and fetal lung samples at 15–23 wk of gestation from two independent experiments (n = 12).

Figure S5.

Gating strategy to identify macrophage precursors in human fetal liver and lung. (A and B) Single-cell suspensions from human fetal liver (A) and lung (B) at wk 21 and 22 of gestation, respectively, were first gated on live CD45⁺CD34⁻ single cells before gating on macrophage precursors in Fig. 7, A and B. FSC-A, forward scatter area; FSC-H, forward scatter height; SSC-A, side scatter area; SSC-H, side scatter height. Data are representative of two independent experiments. Data (A and B) are representative of two independent experiments with 12 fetal liver and lung samples at 15–23 wk of gestation. The red numbers indicate the order of the flow cytometry plots (sequential gating of cells).

Figure 8.

Human alveolar macrophages derived from fetal and adult precursors have a similar turnover. (A) Overview of BrdU pulse-chase experiment. MISTRG transplanted with either Lin⁻CD116⁺ fetal liver cells or CD34⁺ HSPCs by intrahepatic injection were pulsed with BrdU for 10 d, followed by a chase period without BrdU for 4 wk. Cartoon was adapted from Servier Medical Art. (B) Flow cytometry of BrdU incorporation in human alveolar macrophages of fetal versus adult origin. Numbers indicate the frequency of BrdU⁺ lung macrophages after the pulse and at the end of the chase period. Human lung macrophages were gated as CD45⁺CD11b⁺HLA-DR⁺CD206⁺CD169⁺ cells. Human lung macrophages from mice without BrdU administration were used as a staining control (No BrdU control). (C) Frequencies of BrdU⁺ lung macrophages of fetal or adult origin after the pulse and at the end of the chase period. hCD45, human CD45. Data are represented as mean ± SEM. **, P < 0.01 (one-way ANOVA with Tukey’s multiple comparison post hoc test). Data (B) show one lung sample representative of two to five samples (individual mice) per group from two independent experiments. Data (C) are pooled from two independent experiments with n = 2–5 per group.

Figure 9.

Fetal macrophage precursors outcompete adult monocytes in occupying the perinatal alveolar niche in MISTRG mice. (A) Experimental setup of competitive adoptive transfer experiment. Purified fetal and adult macrophage precursors (Lin⁻CD116⁺CD64⁻CD14⁻ fetal liver cells and CD14⁺CD16⁻ blood monocytes) were mixed 1:1 and administered intranasally into newborn MISTRG mice. Cartoon was adapted from Servier Medical Art. (B) Flow cytometry of cell mixture before transfer into MISTRG mice to verify the input ratio (left) and to determine surface expression of distinguishing HLA alleles (HLA-A9 versus HLA-B12). (C) Flow cytometry of human alveolar macrophages after competitive transfer of fetal and adult macrophage precursors into MISTRG mice. Human lung macrophages were gated as CD45⁺CD11b⁺HLA-DR⁺CD206⁺CD169⁺ cells. Fetal- versus adult-derived lung macrophages were distinguished by surface expression of HLA-A9 and HLA-B12 as in B. (D) Frequency of fetal- versus adult-derived human alveolar macrophages in MISTRG mice. hCD45, human CD45. Data are represented as mean ± SEM. ****, P < 0.0001 (unpaired Student’s t test). Data (B) show one representative sample from two independent experiments. Data (C) show one lung sample representative of six samples (individual mice) from two independent experiments. Data (D) are pooled from two independent experiments with n = 6.

Figure 10.

Gene signatures of human alveolar macrophages of fetal versus origin. (A) Experimental outline to define the gene expression profiles of human alveolar macrophages (AMs) derived from the indicated fetal and adult precursor populations. Cartoon was adapted from Servier Medical Art. (B) Volcano plots of differentially expressed genes (DEGs) between the indicated cell populations. Log2 fold change is plotted versus -log₁₀ P value (not corrected for multiple testing). DEGs with a log2 fold change ≥1 and an FDR-corrected P value ≤0.05 are highlighted in blue and red. (C) Bar graphs showing expression of selected genes in human lung macrophages derived from fetal precursors (CD116⁺CD64⁻ and CD116⁺CD64⁺ fetal cells) and adult precursors (CD34⁺ HSPCs). Data are represented as mean ± SEM. (D) Gene ontology analysis of DEGs up-regulated in human lung macrophages derived from adult precursors compared with fetal precursors.GCAS, GeneChip Array Station; HTA, Human Transcriptome Array. Data (B–D) are from a single microarray experiment with two to three replicates (individual mice) per cell population obtained from three independent cell sorting experiments.