HUMANIN produced by human efferocytic macrophages promotes the resolution of inflammation

Abstract

Elimination of apoptotic neutrophils by macrophages, a process called efferocytosis, is a critical step in the resolution of inflammation. Efferocytosis induces the reprogramming of macrophages towards a pro-resolving phenotype and triggers the secretion of pro-resolving factors. While mouse efferocytic macrophages are well-described, less is known about human efferocytic macrophages. Here, using RNA sequencing analysis of three different types of in vitro-derived human efferocytic macrophages, we observed a common modulation of mitochondrial metabolism-related genes in human M0, M1, and M2a-like macrophages, thus correlating with some previous results obtained in other non-human models. These results led us to identify for the first time some particular genes regulated in humans like PLIN5 and MTLN. We also shed light on a mitochondrial gene (MT-RNR2) coding a secreted factor called HUMANIN. Mainly known for its antioxidant and neuroprotective effects, we found that HUMANIN was also associated with pro-resolving properties in human and mouse models. Indeed, HUMANIN was produced early during the resolution of inflammation in an acute peritonitis mouse model. Preventive HUMANIN administration in this model reduced leukocyte infiltration and pro-inflammatory cytokine secretion. These anti-inflammatory properties were accompanied by the early acquisition of a CD11b^low non-efferocytic phenotype by mouse macrophages and by an enhanced expression of pro-resolving genes including Alox15 and Retnla. The ability of HUMANIN to dampen pro-inflammatory cytokine secretion was also confirmed in primary human neutrophils. Finally, HUMANIN was also detected in gingival crevicular fluids of patients suffering from periodontitis after the onset of inflammation, suggesting a role of HUMANIN in the control of inflammation. Overall, our data shed light on new aspects of efferocytosis in humans and identify the pro-resolving potential of HUMANIN. This illustrates its prospective therapeutic interest in inflammatory disorders.

Fig. 1. Efferocytosis induces a transcriptomic program related to mitochondrial metabolism in distinct human macrophage subsets.

A Schematic view of a full human in vitro efferocytosis model. Human M1-, M0- and M2a-like monocyte-derived macrophages (MDM) were co-cultured with CFSE-stained human apoptotic neutrophils (PMN) for 24 h. After the removal of non-engulfed neutrophils, CFSE⁺ efferocytic macrophages were sorted by flow cytometry and used to perform bulk-RNA-seq. B Venn diagram representing shared and cell type-specific differentially regulated genes (DEG) in M1-, M0- and M2a-like MDM after efferocytosis. Eighteen highly modulated genes (log Fold change ± 1.2 p-value ≤ 0.05) were found commonly regulated in all three types of macrophages. C Volcano plots of DEGs in M1-, M0- and M2a-like efferocytic macrophages with a cut-off value of log2(FC = 1.2). Up- and down-regulated genes are colored in purple and green, respectively. D GSEA analysis of DEG after efferocytosis in M1-, M0- and M2a-like efferocytic macrophages using Reactome database. Pathways of interest were highlighted in bold. Up- and down-regulated genes are colored in purple and green, respectively. E Heatmaps of DEG focusing on pathways related to oxidative phosphorylation (OXPHOS) and cholesterol biosynthesis according to their Log2(FC). Compared gene lists of Reactome pathways were adapted according to information acquired by manual data mining. F RNA-seq validation of key genes by RT-qPCR analysis. Examples of up- and down-regulated genes in M1-, M0- and M2a-like macrophages belonging to OXPHOS, cholesterol biosynthesis or common gene signature. mRNA expression was normalized using GAPDH gene and relative expression was based on non-efferocytic cells for each type of MDM. Statistics: Two-tailed Student’s t-test was performed, *P  < 0.05, **P  < 0.01.

Fig. 2. Efferocytosis reduces mitochondrial content and membrane potential, while altering respiratory activity in human pro-inflammatory M1-like macrophages.

Quantification of mitochondrial density (Mito. Density) (A) and membrane potential (Mito. Ψm) (B) were assessed by flow-cytometry using MitoTracker Green FM^TM and TMRM, respectively. M1-like macrophages were co-incubated with CFSE-labeled apoptotic neutrophils (PMN) for 24 h, and then stained with mitochondrial dyes before analyzing by flow cytometry. Efferocytic M1-like macrophages were gated as CFSE⁺ cells. Results are expressed as Relative Fluorescent Intensity (RFI) to each non-phagocytic macrophage (M1 PMN^–, at least n = 7). C Measurement of oxygen consumption rate (OCR) of M1-like macrophages 24 h after apoptotic neutrophil efferocytosis using the Seahorse analyzer. After efferocytosis, macrophages were submitted to sequential inhibitors treatment: oligomycin (Olig., ATP synthase inhibitor), FCCP (decoupling agent) and rotenone/antimycin A (Rot.AA, complex I and complex III inhibitors, respectively) to measure the basal and maximal respiratory capacity of the cells (at least n = 3). Non-efferocytic M1-like macrophages (M1) were compared to efferocytic M1-like macrophages (M1 + PMN). D To promote fatty acid oxidation, M1-like macrophages were incubated (or not) for 24 h with apoptotic neutrophils (PMN) and were then stimulated with palmitate (PA, 200 µM). The Seahorse assay was performed using the same inhibitors as described previously in (C). OCR was measured in the different groups using the Seahorse analyzer. Statistics: Two-tailed paired t-test and two-way ANOVA were performed according to test requirements, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Fig. 3. HUMANIN production is regulated by efferocytosis in humans.

A OXPHOS-related genes found up-regulated in efferocytic M1-like macrophages in our RNA-seq analysis were mapped on the graphic according to their electron chain transport complex origin. Mitochondrial genes are written in red whereas nuclear genes are written in black. MTRNR2L12 MTRNR2L6 genes are encoding HUMANIN isoforms. B Quantification by RT-qPCR of MT-RNR2 mRNA in efferocytic (PMN⁺) and non-efferocytic (PMN^–) M1-like human macrophages (n = 4). C Detection of HUMANIN protein in efferocytic and non-efferocytic M1-like macrophages (M1 + PMN versus M1, respectively) by confocal microscopy. HUMANIN is represented in red and the nucleus in blue (n = 3). D HUMANIN protein expression was assessed in efferocytic and non-efferocytic M1-like macrophages (M1 + PMN versus M1, respectively) by Western blotting (n = 3). Total protein obtained with stain-free imaging was used as a control. T-tests were used according to test requirements. *P < 0.05.

Fig. 4. HUMANIN regulates inflammatory activities but not the viability of human neutrophils, and its level correlates with the inflammatory status of patients suffering from periodontal disease.

A Assessment of apoptosis of PMN cultured with HUMANIN (HN, 100 nM or 1 µM) or GM-CSF (50 ng/ml). Apoptosis at the indicated time points was assessed by flow cytometry using Annexin-V (AnnV) and 7-AAD staining. Apoptotic cells were defined as Annexin-V⁺ and 7AAD^– cells (n = 4). B Effect of HUMANIN on TNFα secretion by human fresh PMN stimulated for 4 h by LPS (100 ng/ml). TNFα level in the supernatants was quantified by ELISA (n = 3). C, D Gingival crevicular fluid (GCF) was collected using strip papers on 1 or 2 sites from 8 patients suffering from periodontitis, during inflammatory (T1) and healing phase induced by mechanical treatments (T2). Proteins contained on the strip were eluted in 0.1 ml of protective buffer. IL-1β (C) and HUMANIN levels (D) (expressed as pg/ml) were measured, in the eluted GCF at T1 and T2, by microfluidic ELISA (ELLA technology) and ELISA respectively. Statistics: Two-tailed paired t-tests or two-way ANOVA were used according to test requirements. *P < 0.05, **P < 0.01, ***P < 0.001.

Fig. 5. Humanin decreases leukocyte recruitment and activation in a murine peritonitis model.

A Peritonitis was induced in C57/Bl6 mice by intraperitoneal (i.p.) injection of 1 mg zymosan-A (Zy.) resuspended in PBS (open dark blue circles) or vehicle PBS (Veh., black circles). HUMANIN (Hu, 20 µg in PBS) or vehicle (Veh) was injected i.p. 30 min before zymosan-A (Zy.+Hu., open light blue circles vs. Hu.+Veh., open circles, respectively). Peritoneal washes were performed at different time points (6 h,12 h 24 and 72 h) and cells were used for different analyses. B Western blot analysis of mouse Humanin (HN) in cell-free peritoneal fluid at 6 h, 12 h, 24 h and 72 h after zymosan-A (Zy.) injection. Total protein obtained with stain-free imaging was used as a control. (n = 3). C Total cell count in peritoneal fluid was performed manually using Trypan blue staining. D Flow cytometry was used to assess number of neutrophils Ly6G⁺, F4/80⁻) and myeloid cells (Ly6G⁻, F4/80^low/High) contained in the peritoneal fluid at the different indicated time points. E Neutrophil viability in the peritoneal fluid at the indicated time points was assessed by flow cytometry using Annexin V and 7-AAD. F After zymosan-A injection, the peritoneal cavity of the mice, treated with HUMANIN or not, was flushed at the indicated time points with PBS and TNFα, IL-6 and IL-1β levels in the recovered fluids were measured by ELISA. Statistics: Two-way ANOVA and two-tailed paired t-test were used according to test requirements. Results are displayed as mean ± SEM (n = 10–14 mice), *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Fig. 6. HUMANIN administration increases pro-resolving CD11b^low macrophages.

A Peritoneal macrophages from mice that received zymosan-A alone (dark blue bars) or with HUMANIN (light blue bars) were collected from mice and isolated by adherence 12 h after zymosan-A injection. For the macrophage restimulation assay, zymosan-A (100 µg/ml) or LPS (100 ng/ml) were added during 16 h. TNFα protein level in cell supernatant was measured by ELISA (n = 4). B Efferocytosis of CD11b^low pro-resolving macrophages at different time points were assessed ex-vivo. Macrophages were collected at 6 h, 12 h and 24 h from mice with peritonitis and mixed with CFSE⁺ apoptotic thymocytes for 45 min. Non-engulfed thymocytes were removed before immuno-staining and flow cytometry analysis. Efferocytosis rates of F4/80⁺, CD11b^low macrophages (green square) from mice that received zymosan-A alone (open dark blue circles) or with HUMANIN (open light blue circles), were determined according to the percentage of CFSE⁺ macrophages. C mRNA expression of genes considered as pro-resolving genes (Alox15, Retlna), or involved in efferocytosis (Mertk, Axl and Pparγ) as well as antioxidant genes (Mt-rnr2, Sod2 and Cat) in phagocytes from peritoneal cavity 12 h after injection of zymosan-A ± Humanin (Zy., open dark blue vs. Zy.+Hu, light blue). Of note, the Mt-rnr2 gene encodes Humanin. Statistics: Two-way ANOVA and two-tailed paired t-test were used according to test requirements. Results are displayed as mean ± SEM (n = 10–14 mice), *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.