Identification of a distinct cluster of GDF15high macrophages induced by in vitro differentiation exhibiting anti-inflammatory activities

Abstract

Introduction: Macrophage-mediated inflammatory response may have crucial roles in the pathogenesis of a variety of human diseases. Growth differentiation factor 15 (GDF15) is a cytokine of the transforming growth factor-β superfamily, with potential anti-inflammatory activities. Previous studies observed in human lungs some macrophages which expressed a high level of GDF15.

Methods: In the present study, we employed multiple techniques, including immunofluorescence, flow cytometry, and single-cell RNA sequencing, in order to further clarify the identity of such GDF15^high macrophages.

Results: We demonstrated that macrophages derived from human peripheral blood mononuclear cells and rat bone marrow mononuclear cells by in vitro differentiation with granulocyte-macrophage colony stimulating factor contained a minor population (~1%) of GDF15^high cells. GDF15^high macrophages did not exhibit a typical M1 or M2 phenotype, but had a unique molecular signature as revealed by single-cell RNA sequencing. Functionally, the in vitro derived GDF15^high macrophages were associated with reduced responsiveness to pro-inflammatory activation; furthermore, these GDF15^high macrophages could inhibit the pro-inflammatory functions of other macrophages via a paracrine mechanism. We further confirmed that GDF15 per se was a key mediator of the anti-inflammatory effects of GDF15^high macrophage. Also, we provided evidence showing that GDF15^high macrophages were present in other macrophage-residing human tissues in addition to the lungs. Further scRNA-seq analysis in rat lung macrophages confirmed the presence of a GDF15^high sub-population. However, these data indicated that GDF15^high macrophages in the body were not a uniform population based on their molecular signatures. More importantly, as compared to the in vitro derived GDF15^high macrophage, whether the tissue resident GDF15^high counterpart is also associated with anti-inflammatory functions remains to be determined. We cannot exclude the possibility that the in vitro priming/induction protocol used in our study has a determinant role in inducing the anti-inflammatory phenotype in the resulting GDF15^high macrophage cells.

Conclusion: In summary, our results suggest that the GDF15^high macrophage cells obtained by in vitro induction may represent a distinct cluster with intrinsic anti-inflammatory functions. The (patho)physiological importance of these cells in vivo warrants further investigation.

Keywords: GDF15; anti-inflammatory; cell population; growth differentiation factor 15; human; macrophage; single-cell RNA sequencing.

Figure 1

Immunofluorescence double labeling showing the existence of GDF15^high macrophages in lung tissues. (A) Results obtained in healthy human lung tissues from 3 independent subjects. Two representative microscopic fields were shown. Arrowheads indicated the CD68⁺GDF15^high macrophages. (B) Results obtained in human lung tissues with COPD from 4 independent subjects. Arrowheads indicated the CD68⁺GDF15^high macrophages. (C) Results obtained in rat lung tissues without and with experimental PAH. The nuclei were counterstained with DAPI (blue). White arrowheads indicated CD68⁺GDF15^high macrophages. The red box highlighted the presence of CD68⁺GDF15^high macrophages (white arrowheads); the green box highlighted the presence of CD68⁺GDF15^low macrophages (red arrowheads). The bar graphs showed the absolute and relative abundances of CD68⁺GDF15^high macrophages in normal and PAH lungs. Data were expressed as mean ± SEM. * P < 0.05, unpaired t-test.

Figure 2

GDF15^high macrophages could be derived by in vitro differentiation of mononuclear cells. (A, B) Immunofluorescence staining and flow cytometry results confirming that in vitro differentiation of human peripheral blood mononuclear cells (PBMNCs) with GM-CSF for 7 days yielded CD68⁺ macrophages. (C, D) Flow cytometry and immunofluorescence double labeling results showing that the PBMNC-derived macrophages contained a minor population of GDF15^high cells (arrowheads in D) (example from 3 independent experiments). (E, F) Fluorescence microscopy and flow cytometry data showing that GM-CSF differentiation of rat bone marrow mononuclear cells (BMMNCs) in vitro yielded macrophages (GFP expressing) of a high purity (~90%). CD68pro-GFP rats had a GFP transgene under the control of CD68 promoter. Cells from normal rats showing no GFP fluorescence served as a negative control (left panel in E). (G, H) Flow cytometry and immunofluorescence double labeling data (from 3 independent experiments) showing that the BMMNC-derived macrophages (from CD68pro-GFP rats) contained a minor population of GDF15^high cells (arrowheads in H). The flow cytometry data in panels (C, G) were from cells gated for GFP⁺. The nuclei were counterstained with DAPI (blue).

Figure 3

Flow cytometry results showing that GDF15^high macrophages did not exhibit a typical M1 or M2 phenotype. Experiments were performed in human PBMNC-derived macrophages, using CD86, CD80 and IL-1β as the M1 markers, and CD206, CD163 and IL-4 as the M2 markers. Data were from a single test using pooled samples from 4 healthy volunteers.

Figure 4

Molecular characterization of human PBMNC-derived GDF15^high macrophages with scRNA-seq. (A) Graphical outline of the experimental procedure. (B) UMAP plots showing the identified cell sub-populations (C1 to C7) based on the scRNA-seq data from total 73,768 cells pooled from samples of 3 healthy volunteers, 3 PAH patients harboring mutations in BMPR2 gene, and 3 PAH patients without BMPR2 mutations. The putative nomenclatures for C1 to C7 were given below the graph. The numbers 0 to 13 demarcated the initial cell clusters obtained with the default clustering process of Seurat. (C) Violin plots showing expression patterns of the identified marker genes for C1 to C7. The horizontal bars represented median values. (D) UMAP plots showing expression patterns of the top 10 genes that were overexpressed in GDF15^high macrophages (C5) as compared to GDF15^low cells. (E) UMAP plots showing expression patterns of the top 9 genes encoding secreted proteins which were overexpressed in GDF15^high macrophages as compared to GDF15^low cells. (F) Cell-cell communication network map created using CellChat showing the possible effector cells of the GDF15^high macrophage. (G, H) Predicted ligand-receptor pairs potentially involved in the signaling of reciprocal communications between GDF15^high macrophage and other cell types as listed in (F).

Figure 5

GDF15^high macrophages exhibited reduced inflammatory activation in vitro. (A) Expression patterns of potential substitute cell surface markers for GDF15 based on the scRNA-seq data. (B) Flow cytometry results showing that rat BMMNC-derived macrophages contained a minor fraction of TNFSF9^high cells, whose expression level was correlated with that of GDF15 (from 3 independent experiments). (C) Flow cytometry verification of the correlation between TNFSF9 and GDF15 expressions in human PBMNC-derived macrophages (from 2 independent experiments). (D) Real-time PCR results showing that GDF15^high macrophages (H) exhibited reduced expressions of TNF-α, IL-1β and IL-6 in response to LPS stimulation (1 μg/mL for 6 hr), as compared to GDF15^low cells (L). Rat BMMNC-derived macrophages were FACS purified using TNFSF9 as a substitute marker for GDF15, and primed with IFN-γ (10 ng/mL for 12 hr). (E) Boyden chamber cell migration assay showing that GDF15^high macrophages (H) exhibited reduced migratory activity as compared to GDF15^low cells (L) in the absence and presence of LPS stimulation. (F) Representative fluorescent microscopic images and quantitative data showing that GDF15^high macrophages exhibited reduced phagocytic activity in the presence of LPS stimulation as compared to GDF15^low cells. Phagocytosis was assessed by internalization of fluorochrome-labeled latex beads (orange color). The macrophages were from CD68pro-GFP rats. Data were mean ± SEM. * P < 0.05, one-way ANOVA. NS, no significance.

Figure 6

GDF15^high macrophages exerted anti-inflammatory effects via paracrine mechanisms. (A-C) RAW264.7 cells co-cultured with rat BMMNC-derived GDF15^high (H) or GDF15^low (L) macrophages were left untreated or stimulated with LPS for 4 hr. Results for the expression of pro-inflammatory cytokines (A, real-time PCR), cell migratory activity (B), and phagocytic activity (C) were shown. (D-F) Unsorted rat BMMNC-derived macrophages co-cultured with rat GDF15^high (H) or GDF15^low (L) macrophages were left untreated or stimulated with LPS for 4 hr. Results for the expression of pro-inflammatory cytokines (D, real-time PCR), cell migratory activity (E), and phagocytic activity (F) were shown. Data were mean ± SEM. * P < 0.05, one-way ANOVA. NS, no significance.

Figure 7

GDF15 might be a macrophage-derived anti-inflammatory factor. (A) Real-time PCR results showing that treatment with exogenous GDF15 (20 ng/mL) inhibited LPS-induced expression of pro-inflammatory cytokines in RAW264.7 cells. (B) Flow cytometry results showing that GDF15 treatment had no effects on phagocytosis in RAW264.7 cells without or with LPS stimulation. (C) Representative images and quantitative data of Boyden chamber assay showing that exogenous GDF15 inhibited migration of LPS-challenged RAW264.7 cells. Cells on the membrane were stained with Giemsa. (D) Effects of conditioned medium from GDF15^high macrophages (H), as compared to the medium from GDF15^low cells (L), on the expression of pro-inflammatory cytokines in RAW264.7 cells. All experiments were performed in the presence of LPS stimulation. α-GDF15, GDF15neutralizing antibody; IgG, non-specific immunoglobulin control. (E) The same experiments as those in D carried out in rat BMMNC-derived macrophages. Data were mean ± SEM. * P < 0.05, one-way ANOVA; † P < 0.05, unpaired t-test. NS, no significance.

Figure 8

Detection of GDF15^high macrophages in various human tissues. GDF15^high macrophages (arrowheads) were identified using immunofluorescence double labeling with anti-CD68 (green color) and anti-GDF15 (red color) antibodies in (A) colon tissues from both healthy subjects and patients with ulcerative colitis, (B) kidneys (the normal peri-tumor tissue) (tested in one sample only) and (C) atherosclerotic plaques in the carotid artery (representative data from 6 independent samples showing similar results). The nuclei were counterstained with DAPI (blue). Data were mean ± SEM. NS, no significance (unpaired t-test).

Figure 9

Results of scRNA-seq analysis showing the existence of GDF15^high clusters in rat lung macrophages. (A) Top, UMAP plots showing scRNA-seq clustering of freshly isolated lung alveolar macrophages and interstitial macrophages from rats (3 independent samples from Control group and 3 from experimental PAH group). Total 18977 alveolar macrophage cells (11842 from Control and 7135 from PAH) and 15357 interstitial macrophage cells (6768 from Control and 8589 from PAH) were analyzed. Bottom, patterns of GDF15 expression showing that clusters 2/4 in alveolar and clusters 1/3 in interstitial macrophages were GDF15^high. (B) Comparison of the percentage prevalence of GDF15^high alveolar and interstitial macrophages between control and PAH animals. (C) Venn diagrams showing the numbers of differentially expressed genes between GDF15^high versus GDF15^low macrophages from rat lungs (Control group only) and from human PBMNCs.