Nck adapter proteins promote podosome biogenesis facilitating extracellular matrix degradation and cancer invasion

Abstract

Background: Podosomes are membrane-bound adhesive structures formed by actin remodeling. They are capable of extracellular matrix (ECM) degradation, which is a prerequisite for cancer cell invasion and metastasis. The signaling mechanism of podosome formation is still unknown in cancer. We previously reported that Nck adaptors regulate directional cell migration and endothelial lumen formation by actin remodeling, while deficiency of Nck reduces cancer metastasis. This study evaluated the role of Nck adaptors in podosome biogenesis and cancer invasion.

Methods: This study was conducted in vitro using both healthy cells (Human Umbilical Vein Endothelial Cell, 3T3 fibroblasts) and cancer cells (prostate cancer cell line; PC3, breast cancer cell line; MDA-MB-231). Confocal and TIRF imaging of cells expressing Green Fluorescence Protein (GFP) mutant under altered levels of Nck or downstream of kinase 1 (Dok1) was used to evaluate the podosome formation and fluorescent gelatin matrix degradation. Levels of Nck in human breast carcinoma tissue sections were detected by immune histochemistry using Nck polyclonal antibody. Biochemical interaction of Nck/Dok1 was detected in podosome forming cells using immune precipitation and far-western blotting.

Results: This study demonstrates that ectopic expression of Nck1 and Nck2 can induce the endothelial podosome formation in vitro. Nck silencing by short-hairpin RNA blocked podosome biogenesis and ECM degradation in cSrc-Y530F transformed endothelial cells in this study. Immunohistochemical analysis revealed the Nck overexpression in human breast carcinoma tissue sections. Immunoprecipitation and far-western blotting revealed the biochemical interaction of Nck/p62Dok in podosome forming cells.

Conclusions: Nck adaptors in interaction with Dok1 induce podosome biogenesis and ECM degradation facilitating cancer cell invasion, and therefore a bona fide target of cancer therapy.

Keywords: Dok1; Nck; c-Src; cancer; extracellular matrix; podosome.

Figure 1

Nck localizes to actin‐rich podosome rosettes of Src‐transformed endothelial cells. A, Representative TIRF images of cell co‐expressing hNck2‐YFP (green) and actin‐mCherry (red) fusion proteins show colocalization of Nck with actin‐rich podosomes (yellow). The right panel of images corresponds to the square selection in the left image. Line intensity scan (bottom left) corresponding to the white line in the merged image as well as scatter plot (bottom right) corresponding to the whole‐cell image (top left) illustrating the relative colocalization of Nck2 and F‐actin in podosome rosettes (Pearson's correlation coefficient Rr = 0.973). B, Representative TIRF images of cell co‐expressing hNck1‐YFP (green) and hNck2‐mCherry (red) fusion proteins show colocalization of Nck1 and Nck2 with podosomal rosettes (yellow). The right panel of images corresponds to the square selection in the left image. Line scan (bottom left) corresponding to the white line and scatter plot (bottom right) corresponding to the whole‐cell image illustrating the relative colocalization of Nck1 and Nck2 in podosome rosettes (Pearson's correlation coefficient Rr = 0.929). Scale bars: 10 µm

Figure 2

Silencing of Nck disrupted Src‐induced podosome formation in endothelial cells. A, Representative TIRF images of cSrc‐Y530F transformed endothelial cells shows very few podosomes in Nck silenced cells compared with control and rescued cells. B, Quantitative analysis indicates a significant decrease in podosome number in Nck‐depleted (shNck 1 and 2) cells compared with control and rescued cells. P < .05. The top and bottom lines of the box plot indicate the third and first quartiles, respectively, while the bold central lines indicate the median values.

Figure 3

Nck adaptors play a critical role in podosome‐mediated extracellular matrix remodeling by endothelial cells. A, Representative confocal images are showing the degradation of a fluorescent matrix (Alexa Fluor 488‐conjugated gelatin) by podosomes induced by Src expression in HUVEC. Control, Nck‐rescued, and Nck‐overexpressing cells (hNck2) exhibit increased matrix degradation relative to the cells with simultaneous knockdown of Nck1 and Nck2 (shNck 1 and 2). Scale bar represents 20 µm. B, Quantitative image analysis showing decreased (P < .001) matrix degradation/cell (left panel) and several actively degrading cells (right panel) in Nck1/Nck2 knockdown (shNck 1 and 2) cells vs control, rescued, and Nck‐overexpressing cells. Bars represent mean ± SEM (n = 3 independent experiments). ***P < .001

Figure 4

Abrogation of Nck signaling impairs endothelial cell invasion. A, Representative images of HUVEC invading through a porous membrane (8 μm pore diameter) precoated with a thin layer of Matrigel^® (125 μg/mL). Serum‐starved cells (4 × 10⁴ cells) were seeded and incubated for 5 h. The top and bottom compartments of Boyden chambers in 24 well plates are filled with starvation medium or starvation medium supplemented with 50 ng/mL of VEGF, respectively. After incubation, nonmigrating cells were removed, and migrating cells were fixed and stained. Brightfield images from multiple fields/treatment were collected. Scale bar, 200 µm. B, Representative western blot showing Nck and GAPDH (loading control) protein levels. C, Quantitative analysis of endothelial cell invasion. Images were processed using the EMBL ImageJ software, and cell counts were expressed as a percentage of controls. Bars represent mean ± SEM (n = 3 independent experiments). ***P < .001 vs control

Figure 5

Nck expression and metalloprotease activity in cancer. A, Immunohistochemical analysis using anti‐Nck polyclonal antibody demonstrates Nck overexpression (right panel) in human ductal breast carcinoma tissues as detected by brightfield imaging of tissue micro array paraffin sections (5 μm). The left panel shows nonspecific background staining of corresponding IgG control sections.  Quantitative image analysis showing significant increase in Nck stained area (***P<.001) compared with normal tissue area (bottom panel). B, Semi‐quantitative RT_PCR of MT1‐MMP in normal and cancer cells. MT1MMP expression was higher in MDA MB‐231 cancer cells compared with HUVEC or MCF10 normal cells. RNA expression levels were normalized to the expression of the reference gene GAPDH, which did not show differential expression between the groups. C, Total MMP and ADAM activities in PC3 cells in vitro. Time‐lapse imaging of PC3 cells, incubated with MMP or ADAM specific fluorogenic substrate, demonstrate higher activities in control vs Nck‐silenced cells. D, Wound healing capacity pf PC3 cells. Nck silencing prevented PC3 cell 2D migration and wound healing. ADAM, a disintegrin and metalloprotease; MMP, matrix metalloproteinases; PC3, prostate cancer cell line; RT, reverse transcriptase

Figure 6

Determination of Nck and Dok1 interaction in NIH3T3 fibroblasts and Human Umbilical Vein Endothelial Cells. A, SH2 profiling of Nck by far‐western blotting in normal NIH3T3 cells and transformed 3T3 cells. A 62 kD band is consistently observed in the transformed cells forming the actin‐rich invasive structures. Although there are other bands in this blot, this particular brand is very distinctly observed in transformed cells. B, Determination of 62 kD band as p‐62 Dok. Protein analysis of normal and transformed 3T3 cells after immunoprecipitation with α‐Dok and SH2‐profiling by far‐western blotting (top); Efficiency of the IP is checked by immunoblotting with α‐Dok (bottom). C, Biochemical interaction of Nck and Dok1 in human endothelial cells. Immunoprecipitation and far‐western blotting revealed biochemical interaction between Nck and Dok1 in human endothelial cells cultured in vitro. D, Representative TIRF images of HUVEC coexpressing LifeAct/mCherry and pLenti/YFP (empty vector). Boxed area is shown on the right at higher magnification. E, Representative TIRF images of HUVEC co‐expressing LifeAct/mCherry (actin/podosomal marker) and hDok1/YFP. Boxed area is shown on the right at higher magnification. F, Line‐scan, corresponding to the white line in image (E), reveals a tight correlation of LifeAct/mCherry and hDok1/YFP fluorescence intensity at podosomes. G, Color scatter plot corresponding to the whole‐cell image shown in (D). Dok1, downstream of kinase 1

Figure 7

Silencing of Dok1 disrupts Src‐induced podosome formation and gelatin matrix degradation. A, Representative confocal images are showing the degradation of a fluorescent matrix (Alexa Fluor 488‐conjugated gelatin) by podosomes induced by Src expression in HUVEC. Control cell exhibit increased matrix degradation relative to Dok1‐silenced cells. Scale bar represents 20 µm. B, Quantitative image analysis showing decreased (***P < .001) matrix degradation area/cell area in Dok1‐silenced cells vs control cells. Bars represent mean ± SD (n = 3 independent experiments). C, Western blot of control and Dok1‐silenced endothelial lysates. Dok1, downstream of kinase 1