Endophilin-A3 and Galectin-8 control the clathrin-independent endocytosis of CD166

While several clathrin-independent endocytic processes have been described so far, their biological relevance often remains elusive, especially in pathophysiological contexts such as cancer. In this study, we find that the tumor marker CD166/ALCAM (Activated Leukocyte Cell Adhesion Molecule) is a clathrin-independent cargo. We show that endophilin-A3—but neither A1 nor A2 isoforms—functionally associates with CD166-containing early endocytic carriers and physically interacts with the cargo. Our data further demonstrates that the three endophilin-A isoforms control the uptake of distinct subsets of cargoes. In addition, we provide strong evidence that the construction of endocytic sites from which CD166 is taken up in an endophilin-A3-dependent manner is driven by extracellular galectin-8. Taken together, our data reveal the existence of a previously uncharacterized clathrin-independent endocytic modality, that modulates the abundance of CD166 at the cell surface, and regulates adhesive and migratory properties of cancer cells.


Description of Additional Supplementary Files
Supplementary Data 1: Data from quantitative proteomic analysis of cell surface proteins in conditions of inhibition of clathrin-mediated endocytosis (CME). HeLa cells were treated for 72h with control or µ2-adaptin siRNAs, and quantitative mass spectrometry analysis (iTraq) was performed on isolated cell surface proteins. 489 proteins (with Unused score  2) were identified, 378 (77.3%) of which were containing 'membrane' in their GO-annotations. Among these hits, 262 (53.6%) were containing the following exact GO-annotations: 'cell surface', 'plasma membrane', 'intracellular space/matrix', 'integral component of plasma membrane' (hits highlighted in green in the table). The iTraq analysis was performed on 4plexes, where reporters 114 and 115 were assigned to two technical replicates of negative control siRNA-treated cells, and reporters 116 and 117 were assigned to two technical replicates of µ2-adaptin siRNA-treated cells. In the table, fold-change ratios are calculated relatively to the reporter 114 (negative control condition). Ratios highlighted in red indicate hits showing an increased abundance in the µ2-adaptin-depleted condition compared to the negative control, while in blue, they indicate a decreased abundance. The color intensity depicts the significance of the ratios: no color, not significant; light, *P < 0.05; medium, **P < 0.01; dark, ***P < 0.001. Genome-edited U2OS expressing CLTA-mRFP were incubated for 1 min with 5 µg ml -1 anti-CD166-ATTO647N antibody and transferred to the Lattice light-sheet microscope for imaging at 30°C. Full 3D volume of 60 planes per U2OS cell was acquired within 3 sec. Spot color indicates the degree of clathrin co-localization, from blue (clathrin negative) to green (clathrin positive). Tracks are depicted in red. Same conditions as Movie 1: genome-edited U2OS expressing CLTA-mRFP were imaged after a 1 min incubation with 5 µg ml -1 anti-CD166-ATTO647N antibody and transferred to the Lattice light-sheet microscope for imaging at 30°C. Full 3D volume of 60 planes per U2OS cell was acquired within 3 sec. This 2D movie shows a side view of a cell slice perpendicular to the detection objective. Insets: note CD166 endocytic events (red) that are negative (left) or positive (right) for clathrin (green). HeLa cells expressing endoA3-GFP were incubated continuously at 37°C with 5 µg ml -1 anti-CD166-A555, and observed by live-cell imaging using a TIRF microscope at 1 sec intervals (exposure time: 50 msec for each channel). Scale bar, 5 µm.

Supplementary Movie 9: Fluid-FM -confocal experiment with Gal8-coated gold nanoparticles on endoA3-GFP expressing cells, related to Figure 3e.
HeLa cells expressing endoA3-GFP (green channel) were exposed locally to single Gal8functionalized nanoparticles (red channel) using a Fluid-FM probe. The recruitment of endoA3-GFP to the site of Gal8 exposure was observed by fast scanning confocal microscopy, recording consecutive images for a total duration of 60 s. Scale bar, 5 µm.

Supplementary Movie 10: Fluid-FM -confocal experiment with Gal1-coated gold nanoparticles on endoA3-GFP expressing cells, related to Supplementary Figure 9i.
HeLa cells expressing endoA3-GFP (green channel) were exposed locally to single Gal1functionalized nanoparticles (red channel) using a Fluid-FM probe. The movements of endoA3-GFP around the site of Gal1 exposure were monitored by fast scanning confocal microscopy, recording consecutive images for a total duration of 60 s. No significant recruitment of endoA3-GFP was observed. Scale bar, 5 µm.

Supplementary Movie 11: Fluid-FM -confocal experiment with bare probe on endoA3-GFP expressing cells, related to Supplementary Figure 9k.
HeLa cells expressing endoA3-GFP (green channel) were exposed locally to bare Fluid-FM probes, as negative control. The transmitted light channel was used to show the exact positioning of the Fluid-FM probe on the cell. The movements of endoA3-GFP around the Fluid-FM probe were monitored by fast scanning confocal microscopy, recording consecutive images for a total duration of 60 s. No significant recruitment of endoA3-GFP was observed. Scale bar, 5 µm. HeLa cells expressing endoA3-GFP (green channel) were exposed locally to single fluorescently labeled gold nanoparticles with no protein attached to it (red channel) using a Fluid-FM probe, as a negative control. The movements of endoA3-GFP around the nanoparticles were monitored by fast scanning confocal microscopy, recording consecutive images for a total duration of 60 s. No significant recruitment of endoA3-GFP was observed. Scale bar, 5 µm.

Supplementary Movie 13: Fluid-FM -confocal experiment with Gal8-coated gold nanoparticles on endoA2-GFP expressing cells, related to Supplementary Figure 9l.
HeLa cells expressing endoA2-GFP (green channel) were exposed locally to single Gal8functionalized nanoparticles (red channel) using a Fluid-FM probe. The movements of endoA2-GFP around the nanoparticles were monitored by fast scanning confocal microscopy, recording consecutive images for a total duration of 60 s. No significant recruitment of endoA2-GFP was observed. Scale bar, 5 µm. SUM159 AP2-GFP cells transfected with control siRNAs or siRNAs targeting endoA3 and/or CD166 were followed during 46h for wound healing in bright field microscopy. Images were acquired every 10 min. Quantifications are shown in Supplementary Figure 10c. Scale bar, 100 µm.