Chemerin regulates normal angiogenesis and hypoxia-driven neovascularization

Abstract

Chemerin is a multifunctional protein initially characterized in our laboratory as a chemoattractant factor for leukocyte populations. Its main functional receptor is CMKLR1. We identified previously chemerin as an anti-tumoral factor inhibiting the vascularization of tumor grafts. We show here that overexpression of bioactive chemerin in mice results in a reduction of the density of the retinal vascular network during its development and in adults. Chemerin did not affect vascular sprouting during the post-natal development of the network, but rather promoted endothelial cell apoptosis and vessel pruning. This phenotype was reversed to normal in CMKLR1-deficient mice, demonstrating the role of this receptor. Chemerin inhibited also neoangiogenesis in a model of pathological proliferative retinopathy, and in response to hind-limb ischemia. Mechanistically, PTEN and FOXO1 antagonists could almost completely restore the density of the retinal vasculature, suggesting the involvement of the PI3-kinase/AKT pathway in the chemerin-induced vessel regression process.

Keywords: CMKLR1; ChemR23; Hind-limb ischemia model; Oxygen-induced retinopathy; Retinal angiogenesis; Tumoral neoangiogenesis.

Fig. 1

Overexpression of chemerin decreases the density of the developing vessel retinal network, without affecting sprouting. a, b Representative images of retinas from mice overexpressing chemerin (K5-Chem) and their wild-type controls at post-natal days 4 (P4) and 6 (P6), stained with isolectin B4 (IsoB4). Scale bars: 500 μm in left panels, 50 μm in right panels. c–e Quantification of the vessels area, total vessels length, and junctions density, normalized for the surface, in the peripheral and central parts of the plexus (n = 11 for P4 controls, n = 10 for P4 K5-Chem, n = 13 for P6 controls, n = 14 for P6 K5-Chem). f, g Representative images of whole mount retinas from K5-Chem and control mice at P4 and P6, stained with IsoB4. Scale bars: 500 μm. h Representative confocal images of the angiogenic front of the retinal network at P6, stained with IsoB4 and for ESM1. White arrows point to endothelial tip cells expressing ESM1. Scale bars: 50 μm. i Radial expansion of the vascular network relative to the retina radius in K5-Chem and control mice (n = 7 for P4 controls, n = 4 for P4 K5-Chem, n = 16 for P6 controls, n = 18 for P6 K5-Chem). j Quantification of the number of ESM1⁺ tip cells per mm of network front in K5-Chem (n = 8) and control mice (n = 6). k Representative confocal images of the remodeling zone of the retinal network from WT and K5-Chem mice at P6, stained with IsoB4 and for ERG. Scale bars: 50 μm. l Representative confocal images of the periphery of the retinal network from control and K5-Chem mice at P6, stained with isoB4 and for EdU and ERG. Scale bars: 50 μm. m Density of ERG⁺ ECs in the remodeling plexus of K5-Chem (n = 7) and control mice (n = 6). n Quantification of EdU⁺ ECs among ERG⁺ cells at the network periphery in K5-Chem (n = 6) and control mice (n = 4). P values versus controls by two-tailed unpaired Student’s t-test. All data are shown as mean ± SEM. Each point represents 1 animal

Fig. 2

Chemerin does not affect vessel sprouting nor pericyte coverage. a Representative images of the capillary network sprouting from aortic ring explants collected from WT mice and treated or not with 10, 20, or 50 nM chemerin. Scale bars: 500 μm. b Quantification of the area of capillary outgrowth (n = 15 for controls, n = 9, 15, and 14 for, respectively, 10, 20, and 50 nM chemerin). P values versus control by one-way ANOVA with Tukey’s post hoc test. NS not significant. c, d Representative images and quantification of BrDU⁺ cells in a culture of HUVECs treated or not with 10 nM chemerin and stained with DAPI (n = 7). e, f Representative images and quantification of the migration of HUVEcs treated or not with VEGF and/or 10 nM chemerin and stained with DAPI in a scratch assay (n = 3). Scale bars: 200 μm. P values by one-way ANOVA with Tukey’s post hoc test. NS not significant. g–i Representative confocal images of the pericyte coverage on vessels from P6 retinas. The retinas were co-stained with IsoB4 and for either α-SMA, PDGFRβ, or desmin as pericyte markers. Scale bars: 50 μm. j–l Quantification of the area stained for α-SMA (n = 4), PDGFRβ (n = 6), and desmin (n = 4) relative to the vascular area (%). m Weight curves of control and K5-Chem mice between 1 and 12 weeks after birth (n = 12). P values versus control by 2-tailed unpaired Student’s t-test. All data are shown as mean ± SEM. Each point represents 1 animal or one well

Fig. 3

Overexpression of chemerin promotes vessel regression and endothelial cell apoptosis. a Representative confocal images of the remodeling plexus of P6 retinas collected from control or K5-Chem mice, stained with IsoB4 and for collagen IV (ColIV). White arrows point to ColIV⁺ IsoB4⁻ empty sleeves. Scale bars: 20 μm. b Representative confocal images of the remodeling plexus of P6 retinas collected from control or K5-Chem mice, stained with IsoB4 and for cleaved caspase-3 (cCasp3). Apoptotic cells are indicated by white arrows. Scale bars: 20 μm. c Quantification of empty sleeves in the remodeling plexus and the angiogenic zone, in the vicinity of arteries or veins (n = 5). d Quantification of cCasp3⁺ IsoB4⁺ cells in the remodeling plexus of K5-Chem (n = 6) and control mice (n = 7). e Diagram depicting the experimental timeline of the tube assay using HUVECs, allowing formation of tubes for 5 h in the presence of VEGF, and measuring the evolution of the network after 3 or 9 h in the presence or absence of 30 ng/ml VEGF and/or 10 nM chemerin. f Representative images of HUVEC tubes at 3 and 9 h in control conditions (n = 9) or after treatment with 10 nM chemerin (n = 19), 50 ng/ml VEGF (n = 11), or the combination of both (n = 25). Scale bars: 200 μm. g, h Quantification of the number of tubes 3 and 9 h after chemerin and/or VEGF treatment, normalized to the tube number before treatment (pool of 6 independent experiments). For c, d, P values versus controls by 2-tailed unpaired Student’s t-test. For g, h, P values versus controls by one-way ANOVA with Tukey’s post hoc test. All data are shown as mean ± SEM. Each point represents one animal or one well

Fig. 4

CMKLR1 mediates the anti-angiogenic effects of chemerin. a, b Representative images of the retinas from P6 control mice, stained with IsoB4 (red) and for CMKLR1 (green) in the sprouting zone and remodeling plexus. Scale bars: 50 μm. c–e Quantification of the vessels area, total vessels length, and junction density in the remodeling plexus of P6 retinas of control mice and mice overexpressing chemerin and/or invalidated for CMKLR1 (n = 10 for controls, n = 14 for K5-Chem, n = 11 for CMKLR1^−/−, n = 6 for K5-Chem/CMKLR1^−/−). P values versus control by one-way ANOVA with Tukey’s post hoc test. f–h Quantification of the vessels area, total vessels length, and junctions density in the remodeling and proliferative zones of the P6 retina from control (n = 4) and chemerin^−/− mice (n = 5). P values versus control by 2-tailed unpaired Student’s t-test. All data are shown as mean ± SEM. Each point represents 1 animal

Fig. 5

The lower vascular density in the retina is maintained in adult mice overexpressing chemerin. a Representative 3D images of the vessel architecture in the retina of control and K5-Chem adult mice. Scale bars: 20 μm. b–d Quantification of the vessels area, total vessels length, and junctions density in the deep (n = 27 for controls, n = 37 for K5-Chem), intermediate (n = 15 for controls, n = 19 for K5-Chem), and superficial layer (n = 3 for controls, n = 5 for K5-Chem) of adult mice. Each point represents 1 field of view. e Representative images of H&E-stained sagittal ocular sections from 8-week-old control and K5-Chem mice. Scale bars: 500 μm in left panels, 50 μm in right panels. f ONL thickness was quantified in K5-Chem (n = 8) and control mice (n = 5). GCL ganglion cell layer, INL inner nuclear layer, IS inner segment, ONL outer nuclear layer, OS outer segment. P values versus controls by 2-tailed unpaired Student’s t-test. All data are shown as mean ± SEM

Fig. 6

Chemerin reduces neovascularization in a model of oxygen-induced retinopathy (OIR). a Diagram depicting the experimental timeline of the OIR protocol. Pups were placed in 75 % oxygen (hyperoxia) from P7 to P12, then placed back under normal oxygen conditions (normoxia) until P17. b, c Representative images of the retinal vasculature stained with IsoB4 in control and K5-Chem mice at P12 and P17. Scale bars: 500 μm in left panels, 50 μm in right panels. The red line marks the limits of the avascular area. d Quantification of the avascular area in control and K5-Chem mice at P12 and P17 (n = 8 for P12 controls, n = 5 for P12 K5-Chem, n = 6 for P17 controls, n = 11 for P17 K5-Chem). e Quantification of the tufts area in control (n = 10) and K5-Chem (n = 9) mice. P values versus controls by 2-tailed unpaired Student’s t-test. All data are shown as mean ± SEM. Each point represents 1 animal

Fig. 7

Chemerin inhibits ischemic tissue recovery and angiogenesis in a model of hind-limb ischemia. a Representative laser Doppler images showing delayed reperfusion in the right hind limb of K5-Chem mice at days 4, 7, 14, and 21 following femoral artery ligation, as compared to control mice. b Representative images of gastrocnemius muscle sections stained for CD31. Scale bars: 1 mm. c Quantification of the distal flow in the gastrocnemius muscle by Doppler measurement in K5-Chem and control mice (n = 15). P values versus control by 2-tailed unpaired Student’s t-test. d Quantification of CD31⁺ area in the ischemic and contralateral limb of control and K5-Chem mice at day 21 (n = 12 for controls, n = 13 for K5-Chem, pool of 3 independent experiments). P values by one-way ANOVA with Tukey’s post hoc test. All data are shown as mean ± SEM. Each point represents 1 animal

Fig. 8

Chemerin favors tumor vessel regression. a 1.10⁶ B16-F0 cells were implanted into the left flank of control (n = 15) and K5-Chem (n = 12) mice (in 3 independent experiments). The tumor size was monitored with a caliper until day 9. b The weight of the tumors was measured following sacrifice at day 9. c 1.10⁶ LLC cells were implanted into the left flank of mice and the weight measured after sacrifice at day 13 (n = 12 for controls, n = 10 for K5-Chem). d Immunofluorescence staining for CD31 and ColIV were performed on cryosections of B16 tumors collected from control and K5-Chem mice at day 9. Scale bars: 500 μm in left panels, 100 μm in right panels. e–g Quantification of the CD31⁺ area, vessels number, and ColIV⁺ area relative to the CD31⁺ vascular area (%) in B16 tumors (n = 11 for controls, n = 7 for K5-Chem). h Control and chemerin^−/− mice were grafted with B16 cells and the tumor size was measured. i The weight of the tumors was measured following sacrifice at day 9 (n = 14 for controls, n = 16 for chemerin^−/−). P values versus control by 2-tailed unpaired Student’s t-test. All data are shown as mean ± SEM. Each point represents 1 animal

Fig. 9

Chemerin activity involves the PTEN-AKT-FOXO1 axis. a Representative images of retinas of P6 control and K5-Chem mice, injected intraperitoneally with DAPT or vehicle, after staining with IsoB4. Scale bars: 200 μm. b–d Quantification of the vessels area, total vessels length, and junctions density in the remodeling plexus of these retinas (n = 6 for controls vehicle, K5-Chem vehicle and K5-Chem DAPT, n = 5 for controls DAPT). e–g Quantification of the vessels area, total vessels length, and junctions density in the remodeling plexus of retinas from control and K5-Chem mice injected with bpV(HOpic) or vehicle (n = 7 for controls vehicle, n = 6 for K5-Chem vehicle, n = 4 for controls bpV, n = 5 for K5-Chem bpV). h–j Quantification of the vessels area, total vessels length, and junctions density in the remodeling plexus of retinas from control and K5-Chem mice injected with AS1842856 (AS) or vehicle (n = 4 for controls vehicle and K5-Chem AS, n = 6 for K5-Chem vehicle and controls AS). k Representative image of HUVECs stained for CMKLR1. Scale bar: 20 μm. l Representative images of HUVECs treated or not with CoCl₂ and stained by DAPI and for HIF1a. Scale bars: 20 μm. m Representative Western blot for HIF1a detection in HUVECs treated or not with CoCl₂ and/or 10 nM chemerin. n Quantification of HIF1a normalized to actin. o Representative Western blots of HUVECs treated or not with CoCl₂ and/or 10 nM chemerin. p–r quantification of the phosphorylation ratios of PTEN, AKT, and FOXO1. P values versus controls by 2-tailed unpaired Student’s t-test. All data are shown as mean ± SEM. Each point represents 1 animal