Endothelial basement membrane topography and assembly

a: Topographical analysis of basement membrane deposition using AFM imaging of CDM prepared 6, 24 or 72 h after seeding confluent endothelial cells. Height maps representative of at least 3 independent experiments are shown. White-dotted lines indicate positions of the scan lines of height. Scale bar: 5µm. b: ECM fiber diameter (µm), matrix coverage and mean fibers height (nm) were quantified in 2 (6 h) and 3 (24 and 72 h) independent experiments with a minimum of 4 maps/experiment. Mann-Whitney test for non-Gaussian distribution was performed for fiber diameters; Kruskal-Wallis test was performed for non-Gaussian distribution for matrix coverage; one-way ANOVA was performed for mean height. c-e: Topographical distribution of ECM components using correlative AFM-fluorescence analysis of CDM prepared from confluent cells cultured for 6 or 24 h and immunostained for LOXL2 (green) and fibronectin (magenta) (c), or LOXL2 and collagen IV (cyan) (d) or fibronectin and laminin (yellow) (e). Scale bar: 5µm. Quantification of the proportion of AFM-detected structures containing fluorescent signal was performed in 3 distinct experiments and 6 maps of 5x5 or 10x10 µm from each experiment. Statistical analysis was done using paired two-way ANOVAs.

LOXL2 depletion affects spatio-temporal distribution of basement membrane components

a: LOXL2 is required for fibronectin remodeling at 24 h. Immunostaining of fibronectin (cyan) and collagen IV (red) in CDM generated by control or LOXL2-depleted endothelial cells over 24 h. Scale bar: 10 µm. Alignment, total length and fractal dimension were quantified for fibronectin staining at 24 h in n = 32 images per condition, from 3 independent experiments. Statistical analysis was performed using Mann-Whitney tests. b: Topographical AFM analysis of CDM generated by control or LOXL2-depleted endothelial cells over a period of 72 h. White dotted line indicates position of the scan lines of height shown on the right. Scale bar: 5 µm. Height of individual fibers (nm) was extracted from scan-lines in 2-5 images from each of 4 distinct experiments (total of 60-100 fibers) and unpaired Student t-test was performed. Data is represented as mean + SD. c-d: LOXL2-depletion affects overall basement membrane organization at 72 h. Correlative AFM-fluorescence analysis of CDM generated by control or LOXL2-depleted endothelial cells over 72 h. CDM were immunostained for fibronectin (c) or collagen IV (d). Yellow arrows point structures detected by AFM and lacking collagen IV staining. Scale bar: 5µm. e: Schematic representation of basement membrane assembly by cells expressing (top) or not (bottom) LOXL2 at 6, 24 and 72 h.

LOXL2-mediated basement membrane assembly drives cytoskeleton organization and cell-cell junctions

a: Detection of F-actin with phalloidin-Alexa Fluor 488 was performed at 72 h. Scale bar: 20 µm. Cytoskeleton orientation was assessed in 3 independent experiments (n = 10 fields of view of 60x60 µm per condition). B: Immunostaining of b-catenin 72 h after seeding control or LOXL2-depleted confluent endothelial cells. Cell circularity was calculated in n = 144 (shControl) and n = 122 (shLOXL2) cells, from two distinct experiments. Statistical analysis was done using unpaired t-tests. c: Evaluation of cell-cell junction morphology was performed on VE-Cadherin-stained endothelial monolayers after sorting in straight, reticulate and serrated categories. Data are mean ±SD of 3 independent experiments each producing 4-7 fields of view per condition. p values reported were calculated for straight junctions using unpaired t-tests between control and LOXL2-depleted cells in control conditions (*** p < 0.0005). d: Barrier properties of endothelial monolayers were measured after 72 h in culture. Data are means of 4 experiments performed in duplicate or quadruplicate culture wells (n = 12). Statistical analysis was done using unpaired t-test.

Defective basement membrane scaffolding alters maturation of cell-ECM adhesions and cell-cell junctions

a: Immunostaining of b1 integrin (green), fibronectin (red) and VE-cadherin (white) in control or LOXL2- depleted endothelial cells 24 or 72 h after seeding confluent cells. b: Immunostaining of pY397-FAK (pFAK – cyan), vinculin (green), pY118-paxillin (pPAX – cyan) and b-catenin (red) in control or LOXL2- depleted endothelial cells 72 h after seeding confluent cells. Nuclei were detected with DAPI (blue). Scale bar: 10 µm. c: Control (shCt) or LOXL2-depleted (shLOXL2) endothelial cells were cultured 72 h before cell lysis and western blotting for the indicated proteins. The level of expression (left column), basal phosphorylation (center column) and VEGF-induced phosphorylation (right column) were quantified in 4 independent experiments performed in duplicate culture wells (n = 8). Statistical analysis was done using Kruskall-Wallis tests. d: Schematic of cytoskeleton remodeling and distribution of adhesion proteins in cell-ECM and cell-cell junctions in control and LOXL2-depleted cells.

Defective basement membrane scaffolding affects cell response to topography

a: Schematic of the microgrooved culture substrate. Two µm wide (w) and 1 µm deep (d) grooves were spaced (S) by 2 µm. b: Immunostaining of VE-cadherin (magenta) and staining of F-actin with phalloidin-Alexa Fluor488 (green) 24, 48 or 72 h after seeding control or LOXL2-depleted confluent endothelial cells on microgrooved substrates. Scale bar: 50 µm. Cell orientation maps are shown for each condition. Scale bar: 100 µm. c: Cell shape and orientation were measured after segmentation and image analysis of VE-cadherin staining in 3 independent experiments. d: Immunostaining of b1 integrin (red) and VE-cadherin (white) in cells cultured for 24 and 72 h on micro-grooved substrates. F-actin was stained with phalloidin-Alexa fluor488 (green). Scale bar: 20 µm. e: Immunostaining of fibronectin (green), collagen IV (red) and VE-cadherin (white) in cells cultured for 72 h on micro-grooved substrate. Nuclei were detected with DAPI (blue). Scale bar: 20 µm. Red double arrows indicate the orientation of the microgrooves.

Basement membrane deposition and 3D vascular morphogenesis

a-b: Endothelial cells were cultured in 3D collagen I hydrogels for 24 h before F-actin staining with phalloidin (red). Scale bar: 50 µm. Cell volume (µm2) was calculated in n = 59-111 cells/condition. Proportion of dead cells was assessed by counting fragmented nuclei over total cell nuclei in n = 12-14 images per condition. All parameters were quantified in 3 independent experiments each performed in duplicate hydrogels. Statistical analysis was performed using unpaired t-tests and Mann-Whitney test. Collagen I was visualized using SHG in 2-P microscopy (white-right panel) (b). Scan lines of intensity are shown to illustrate collagen I compaction at the cell surface. Scale bars: 10 µm. c-d: Endothelial cells were cultured in 3D collagen I hydrogels for 72 h before F-actin staining with phalloidin (red). Scale bars: 50 µm. Diameters and number of connections of capillary-like structures were calculated in n = 30-65 cells per condition. Number of dead cells was assessed by counting fragmented nuclei in approximately 15 fields of view/condition. All parameters were quantified in 3 independent experiments each performed in duplicate hydrogels. Statistical analysis was performed using unpaired t-tests. Collagen IV (cyan) was detected using confocal imaging and Z-stack projections are displayed (left panel) (d). Squared box corresponds to the zoomed area displayed as single z plane in the right panel. Scale bars: 50 and 10 µm, respectively. For all graphs, data is represented as mean + SD and p-values are indicated.

Antibodies and probes