Endocytic trafficking of laminin is controlled by dystroglycan and is disrupted in cancers

Abstract

The dynamic interactions between cells and basement membranes serve as essential regulators of tissue architecture and function in metazoans, and perturbation of these interactions contributes to the progression of a wide range of human diseases, including cancers. Here, we reveal the pathway and mechanism for the endocytic trafficking of a prominent basement membrane protein, laminin-111 (referred to here as laminin), and their disruption in disease. Live-cell imaging of epithelial cells revealed pronounced internalization of laminin into endocytic vesicles. Laminin internalization was receptor mediated and dynamin dependent, and laminin proceeded to the lysosome through the late endosome. Manipulation of laminin receptor expression revealed that the dominant regulator of laminin internalization is dystroglycan, a laminin receptor that is functionally perturbed in muscular dystrophies and in many cancers. Correspondingly, laminin internalization was found to be deficient in aggressive cancer cells displaying non-functional dystroglycan, and restoration of dystroglycan function strongly enhanced the endocytosis of laminin in both breast cancer and glioblastoma cells. These results establish previously unrecognized mechanisms for the modulation of cell-basement-membrane communication in normal cells and identify a profound disruption of endocytic laminin trafficking in aggressive cancer subtypes.

Keywords: Cancer; Dystroglycan; Endocytosis; Laminin.

Fig. 1.

Laminin is internalized into intracellular acidified vesicles. (A) MECs were incubated with CypHer-5-labeled laminin for 18 h, at which point (t = 0 s) live cells were imaged by time-lapse fluorescence microscopy over a 108-s period. Imaging time is indicated at the upper right of each panel. The dashed line in the left panel outlines the cell boundary. Live cell imaging of CypHer–Ln shows its accumulation in acidic and mobile endocytic vesicles, most of which moved rapidly within the cytoplasm. Individual CypHer–Ln-containing vesicles are circled and color coded for tracking. See supplementary material Movie 2 for the associated video. Scale bar: 5 µm. (B) MEC cells were treated with Rhod–Ln for 18 h, trypsinized to remove surface laminin, washed and fixed. The cells were labeled with concanavalin A (ConA) to reveal the cell plasma membrane. A single plane confocal scan clearly shows laminin-containing vesicles within the cell (Merge). Scale bar: 20 µm.

Fig. 2.

Synchronous measures of laminin internalization reveal lysosome-mediated laminin degradation. (A) E3D1 MECs were pulse-labeled with Rhod–Ln at 4°C for 20 min, unbound Rhod–Ln was washed away, and the cells were returned to 37°C. Samples were stripped of cell-surface-bound laminin by trypsinization and analyzed by flow cytometry at 0, 1, 2, 4, 8, 16 and 24 h post laminin labeling (n = 4). (B) Pulse-labeled E3D1 MECs as in A were incubated in the absence (control) or presence of MG-132, leupeptin or DMSO (vehicle) and analyzed by flow cytometry at 24 h post laminin labeling. Data show the mean±s.e.m. (n = 4); *P<0.001.

Fig. 3.

Endocytosed laminin traffics through Rab7-positive vesicles and multivesicular bodies to the lysosome. (A) Following an 18-h incubation with Rhod–Ln (Ln), human breast epithelial UACC893 cells expressing Rab11–GFP showed no colocalization of Ln with Rab11-expressing vesicles, whereas accumulation of laminin was clearly observed within Rab7–GFP- and LAMP1–GFP-expressing vesicles (arrows). n, nucleus. Scale bar: 10 µm. (B) Calculation of the Pearson correlation coefficient from at least four separate images showed colocalization of laminin with Rab7 and LAMP1 markers, but not the Rab11 marker. The data show the mean±s.e.m. (C) Deconvolution imaging of Rhod–Ln (red) in cells expressing the GTPase-deficient mutant Rab5Q79L–GFP (Rab5Q79L) (green) showed accumulation of multiple individual laminin-containing vesicles in multivesicular bodies of the late endosome. The boxed XY image is shown at higher magnification to the right. The vertical line indicates the position of the XZ scan shown at the far right. Scale bar: 5 µm.

Fig. 4.

Laminin endocytosis is receptor mediated. (A) E3D1 MECs were incubated with either 10 µg/ml Rhod–Ln or 40 µg/ml FITC–dextran (500S) in the presence or absence of DMSO (control), 100 µg/ml heparin, 320 mM sucrose or 40 µM dynasore for 18 h, and processed for flow cytometry. All drug treatments resulted in significantly less laminin endocytosis (P<0.01, n = 5), whereas no significant effect was observed with FITC–dextran (P>0.2, n = 4). Data show the mean±s.e.m. (B) E3D1 MECs were incubated continuously with Rhod–Ln or FITC–dextran and internalization was quantified after 2, 4, 8, 16 and 24 h. Note that the rate of FITC–dextran internalization is more rapid than and is distinct from that of Rhod–Ln internalization.

Fig. 5.

The laminin receptor dystroglycan controls laminin internalization. (A) MEpG MECs that lack dystroglycan expression were either infected with empty vector (DG^−/−) or wild-type dystroglycan (DG⁺), incubated with no laminin or Rhod–Ln for 18 h, trypsinized to remove surface laminin, washed and fixed. Cells were labeled with concanavalin A (ConA) to reveal the cell plasma membrane. A single-plane confocal scan shows abundant internal vesicles filled with laminin within dystroglycan-expressing cells, but these are greatly reduced in DG^−/− cells. Scale bar: 20 µm. (B) Laminin (Ln) internalization in cells treated as in A was quantified by flow cytometry. The histogram demonstrates that dystroglycan-expressing MECs (DG⁺) internalize significantly more laminin than dystroglycan-lacking MECs (DG^−/−). (C) Graph of compiled MFI flow cytometry data. DG⁺ MECs internalize 380% more laminin than do DG^−/− MECs. n = 6; *P<0.01. (D) E3D1cre19 MECs that lack β1 integrin expression (see supplementary material Fig. S3) were either infected with empty vector (β1 Int^−/−) or wild-type β1 integrin (β1 Int+). Laminin internalization was assayed by flow cytometry. No significant difference (N.S.) was found between β1-integrin-lacking cells and wild-type β1-integrin-expressing cells. n = 6; P = 0.17. Data in C,D show the mean±s.e.m.

Fig. 6.

Dystroglycan is required for both assembly and internalization of laminin, and it traffics with laminin to the late endosome. (A) MEC cells lacking dystroglycan (DG^−/−) were re-infected to express dystroglycan fused to GFP (dim green cells, DG⁺) or GFP alone (bright green cells, DG⁻) (shown separately in supplementary material Fig. S4A). Rhod–Ln was added to the co-cultured cells, and the cells were imaged live at 37°C. The still image is taken from a 20-h movie. Laminin (red) assembled only on the DG⁺ cells (dim green cells, arrows), and not on the DG^−/− cells (bright green cells, arrowheads). Laminin-positive endocytic vesicles are also seen only within DG⁺ cells. These vesicles can be seen rapidly moving within most of the DG⁺ cells (see supplementary material Movie 6). Scale bar: 50 µm. (B) Cells coexpressing dystroglycan–RFP (DG) and Rab7–GFP (Rab7) constructs were treated with Alexa-Fluor-647-labeled laminin (Ln). Live-cell imaging permitted the simultaneous tracking all three molecules, and revealed strong colocalization of dystroglycan and laminin within Rab7-positive vesicles of the late endosome (see supplementary material Movie 7). The Pearsons's correlation coefficients were 0.62±0.17 for dystroglycan–Ln colocalization (n = 5) and 0.48±0.22 for dystroglycan–Rab7 (n = 4). Arrows are included for positional reference. Scale bar: 10 µm.

Fig. 7.

Functionally glycosylated dystroglycan is required for endocytosis of laminin. Either MDA-MB-231 (MDA231) human breast carcinoma or human LN18 glioblastoma cells were infected with empty vector (vector) or LARGE-glycosyltransferase-expressing retrovirus. (A) Immunoblotting with the glycosylation-specific anti-α-dystroglycan antibody IIH6 demonstrates the absence of glycosylated dystroglycan in vector-infected cells and the presence of glycosylated dystroglycan in LARGE-infected cells. HA-tagged LARGE expression was detected using anti-HA antibodies, β-dystroglycan levels remain unchanged, demonstrating equal protein loading. Numbers on the left indicate the locations of molecular mass markers (in kDa). (B) The laminin assembly assay shows that only LARGE-expressing cells assemble laminin. Scale bar: 50 µm. Insets of the regions highlighted by dashed boxes show details of assembled laminin. Scale bar: 5 µm. (C) Flow cytometry histogram of MDA-MB-231 cells infected with empty vector (green) or LARGE (red), incubated with Rhod–Ln for 18 h and trypsinized for flow cytometry. A shift in fluorescence intensity to the right demonstrates much greater accumulation of laminin in LARGE-expressing cells compared with that of vector-infected cells. The no-laminin control histogram (No Ln) overlaps closely with that of the vector control. (D) Flow cytometry data as in C were compiled from three separate experiments. The MFI values were as follows: MDA-MB-231 vec, 0.1±0.014; LARGE, 6.82±0.241 (P<0.001); LN18 vec, 0.145±0.155; LARGE, 1.49±0.049 (P<0.05), n = 3 (mean±s.e.m.). (E) MDA-MB-231 cells were transfected with the Rab7–GFP fusion protein (green) and incubated with Rhod–Ln for 18 h. Fluorescence imaging of Rhod–Ln (red) in cells without (vector) and with expression of LARGE demonstrates strong accumulation of laminin in Rab7-expressing vesicles only in LARGE-expressing cells (arrows in merged image). Scale bar: 10 µm.