Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2014 Jan 20;204(2):247-63.
doi: 10.1083/jcb.201307067.

Inhibition of endothelial FAK activity prevents tumor metastasis by enhancing barrier function

Christine Jean  1 , Xiao Lei Chen, Ju-Ock Nam, Isabelle Tancioni, Sean Uryu, Christine Lawson, Kristy K Ward, Colin T Walsh, Nichol L G Miller, Majid Ghassemian, Patric Turowski, Elisabetta Dejana, Sara Weis, David A Cheresh, David D Schlaepfer
Affiliations

Inhibition of endothelial FAK activity prevents tumor metastasis by enhancing barrier function

Christine Jean et al. J Cell Biol. .

Abstract

Pharmacological focal adhesion kinase (FAK) inhibition prevents tumor growth and metastasis, via actions on both tumor and stromal cells. In this paper, we show that vascular endothelial cadherin (VEC) tyrosine (Y) 658 is a target of FAK in tumor-associated endothelial cells (ECs). Conditional kinase-dead FAK knockin within ECs inhibited recombinant vascular endothelial growth factor (VEGF-A) and tumor-induced VEC-Y658 phosphorylation in vivo. Adherence of VEGF-expressing tumor cells to ECs triggered FAK-dependent VEC-Y658 phosphorylation. Both FAK inhibition and VEC-Y658F mutation within ECs prevented VEGF-initiated paracellular permeability and tumor cell transmigration across EC barriers. In mice, EC FAK inhibition prevented VEGF-dependent tumor cell extravasation and melanoma dermal to lung metastasis without affecting primary tumor growth. As pharmacological c-Src or FAK inhibition prevents VEGF-stimulated c-Src and FAK translocation to EC adherens junctions, but FAK inhibition does not alter c-Src activation, our experiments identify EC FAK as a key intermediate between c-Src and the regulation of EC barrier function controlling tumor metastasis.

PubMed Disclaimer

Figures

Figure 1.
Figure 1.
Pharmacological FAK inhibition prevents tumor-associated VEC-Y658 phosphorylation. (A) Phosphospecific antibody to FAK-Y397 (pY397, brown) staining of paraffin-embedded normal human breast and invasive ductal carcinoma samples. Sections were counterstained with hematoxylin (blue). Indicated are ductal cells, tumor, stroma, and blood vessels (BV). The boxed regions were enlarged (bottom images), and the arrow indicates a tumor–stromal blood vessel with strong pY397 FAK staining. (B) 4T1-L breast carcinoma cells were implanted in the mammary fat pad of BALB/c mice. After 48 h, mice were provided with 5% sucrose (control) or 0.5 mg/kg PND-1186 in 5% sucrose ad libitum in the drinking water. After 18 d, 4T1-L tumor protein lysates (two independent controls and two treated with PND-1186) were analyzed by VEC pY658 and FAK pY397 immunoblotting. Blots were reprobed for total VEC, FAK, and actin expression. IP, immunoprecipitation; WCL, whole-cell lysate. (C) Densitometry of VEC pY658 in 4T1-L tumors from control and PND-1186–treated mice. Values are mean percentages (pY658/total VEC) ± SEM from four independent tumors per experimental point (*, P < 0.05). The mean of vehicle-treated tumors was set to 100. (D) ID8-IP tumor cells were microinjected into the ovarian space, and after 7 d, vehicle or 30 mg/kg PF-271 was administered by oral gavage twice daily. After 28 d, ID8-IP tumor lysates (three independent controls and three treated with PF-271) were analyzed by VEC pY658 and FAK pY397 immunoblotting. Blots were reprobed for total VEC, FAK, and actin expression. (E) Densitometry of VEC pY658 in ID8-IP tumors from vehicle and PF-271–treated mice. Values are mean percentages (pY658/total VEC) ± SEM from three independent tumors per experimental point (*, P < 0.05). The mean of vehicle-treated tumors was set to 100. (F) Control or PF-271–treated ID8-IP tumors were analyzed by combined staining for ECs (CD31), VEC pY658, or activated FAK (pY397 FAK) as indicated. Merged images (DAPI nuclear stain, blue) show VEC pY658 staining in association with ECs (yellow) only in vehicle control–treated mice. Bars, 50 µm.
Figure 2.
Figure 2.
FAK-dependent VEGF-A stimulation of VEC-Y658 phosphorylation in lung tissue, primary mHLECs, and human ECs. (A, top) Lung lysates from FAK-WT or FAK-KD mice were analyzed 2 min after PBS or VEGF-A tail vein injections. Equal VEGFR-2 Y1175 phosphorylation occurs in FAK-WT and FAK-KD in response to VEGF. VEC pY658, FAK pY397, and c-Src pY416 were increased in VEGF-stimulated FAK-WT mice. Increased c-Src pY416, but not VEC pY658 or FAK pY397, were detected in VEGF-stimulated FAK-KD mice. Blots were reprobed for total VEGFR-2, VEC, FAK, and c-Src. Actin was used as a loading control. (bottom) Densitometry of VEC pY658 and FAK pY397 in lung lysates from FAK-WT or FAK-KD mice. Mean values are the ratio of pY658 VEC or pY397 FAK (normalized to total VEC or FAK) between VEGF-treated and control mice ± SEM from three independent experiments. Ratio of VEGF/control in FAK-WT mice was set to 100. (B) 50 ng/ml VEGF-A (10 min) equally activates VEGFR-2 but does not increase VEC pY658 in FAK-KD compared with FAK-WT mHLECs. Blotting and densitometry analyses were performed as described in A. (C) 1 µM PF-271 blocks 50 ng/ml VEGF-A–stimulated (10 min) VEC pY658 and FAK pY397 but not VEGFR-2 tyrosine phosphorylation or c-Src pY416 in HUVECs. Densitometry of VEC pY658 and FAK pY397 (normalized to total VEC or FAK) in VEGF/nontreated or VEGF/PF-271–treated HUVEC is obtained from three independent experiments. The ratio of VEGF/nontreated was set to 100. (A–C) Error bars are means ± SEM. *, P < 0.05. IP, immunoprecipitation; WCL, whole-cell lysate. (D) Lungs from VEGF-treated (2 min) FAK-WT or FAK-KD mice analyzed by combined staining for ECs (CD31) and VEC pY658. Merged images (DAPI nuclear stain, blue) show VEC pY658 costaining in association with ECs (yellow) in small vessels (SV) and large vessels (LV) in VEGF-treated FAK-WT but not FAK-KD mice. Bars, 30 µm.
Figure 3.
Figure 3.
Combined c-Src and FAK activity are required for VEGF-A–induced c-Src–FAK adherens junction localization. (A–C) HPAEC immunofluorescent staining. Serum-starved cells were pretreated or not treated with 1 µM PF-271 or 100 nM Dasatinib for 1 h, fixed, or 50 ng/ml VEGF-A stimulated (10 min) and then fixed. (A) Combined staining for FAK and pY416 c-Src. (B) Combined staining for pY416 c-Src and VEC. (C) Combined staining for FAK and VEC. (A–C) Merged images (DAPI nuclear stain, blue) show staining colocalization (yellow). Arrows denote points of colocalization. Bars, 10 µm. (D) Dasatinib pretreatment inhibits VEGF-A–induced Y416 c-Src, FAK-Y576/577, and Y658 VEC phosphorylation but does not impact Y397 FAK phosphorylation as shown by immunoblotting whole-cell lysates (WCL). Blots were reprobed for total VEC, FAK, and c-Src. Actin is the loading control.
Figure 4.
Figure 4.
FAK directly phosphorylates VEC-Y658. (A) In vitro kinase assays with recombinant FAK kinase domain, full-length c-Src, and GST fusion protein of cytoplasmic VEC (GST-VEC 621–784) or GST alone. Assays were performed in the presence or absence of 0.5 µM PF-271. Immunoblotting was used to monitor c-Src pY416, total c-Src, VEC pY658, GST-VEC fusion, and phosphotyrosine incorporation. GST was not detectably phosphorylated by FAK, and Coomassie blue staining shows recombinant protein levels. (B) MS coverage of GST-VEC phosphorylation by FAK. Green, >95% confidence in the detected sequence; yellow, between 50 and 90% confidence; red, <50% confidence; gray, no detection. Blue, >99% paragon algorithm confidence score with a phosphorylated tyrosine present at the indicated position. (C) Analysis of the product scan for the ion 891.8829 (+2 charge state). The fragment ion masses correspond to fragmentation of peptide MDTTSY[P]DVSVLNSVR with a phosphorylated tyrosine at position Y658 (paragon algorithm confidence >99%). m/z, mass per charge.
Figure 5.
Figure 5.
Tumor-associated VEGF-A triggers VEC pY658 and VEC internalization dependent on EC FAK activity. (A) VEGF-A ELISA from the indicated tumor cell conditioned media after 48 h. Values are means ± SEM from four independent experiments and normalized to tumor cell number (***, P < 0.001). (B) No cells (medium only is used as a control), ID8, or ID8-VEGF cells were added quiescently to a HUVEC monolayer (1:5 tumor cell to HUVEC ratio) for 6 h in the presence of 1 µM PF-271 or DMSO as indicated. Anti–VEGFR-2 immunoprecipitations (IP) were analyzed by phosphotyrosine (pY) immunoblotting and reprobed for total VEGFR-2. Whole-cell lysates (WCL) were analyzed by pY658 VEC, pY397 FAK, and pY416 c-Src and total VEC, FAK, and c-Src. CD31 is EC specific, and actin is the total loading control. (C) FAK activity regulates VEC internalization in response to VEGF. Starved HPAECs were incubated with an antibody to the VEC extracellular domain, treated with 1 µM PF-271, and/or stimulated with VEGF-A (30 min). Acid wash–resistant anti-VEC antibody internalization was visualized by fluorescence microscopy, quantified using ImageJ, and expressed as points of anti-VEC uptake per cell (n ≥ 400 cells, three independent experiments; means ± SEM; *, P < 0.05; **, P < 0.01 by analysis of variance). NT, not treated. (right) Representative images with anti-VEC and nuclei stained with DAPI (blue). Bar, 10 µm.
Figure 6.
Figure 6.
Tumor-associated VEGF-A triggers transcellular migration dependent on EC FAK activity. (A and B) Transmigration of fluorescently labeled ID8 and ID8-VEGF ovarian (A) or B16F10 melanoma (B) tumor cells across the HUVEC monolayer in the presence of 1 µM PF-271 as indicated and analyzed at 3, 6, and 24 h. (C and D) Transmigration of fluorescently labeled ID8 and ID8-VEGF ovarian (C) or B16F10 melanoma (D) tumor cells across FAK-WT or FAK-KD mHLECs as indicated and analyzed at 3, 6, and 24 h. (A–D) Values are means ± SEM from three independent experiments of triplicates (*, P < 0.05; **, P < 0.01; ***, P < 0.001).
Figure 7.
Figure 7.
Selective importance of VEC-Y658 phosphorylation in VEGF-stimulated paracellular permeability and tumor cell transmigration. (A) VEC schematic showing extracellular, transmembrane (TM), and cytoplasmic domain with location of Y658F, Y685F, and Y733F point mutations fused to GFP. (B) Localization of the indicated VEC-GFP constructs to cell–cell junctions in VEC-null ECs. Bar, 10 µm. (C) Flow cytometry shows equivalent VEC-GFP WT, Y658F, Y685F, and Y733F expression in ∼90% of VEC-null ECs. 10,000 cells were analyzed for each condition. Data shown are from a single representative experiment out of three repeats. (bottom) Table showing enumeration of GFP-positive (expressed as percentages) cells and mean fluorescence intensity (MFI). (D) Basal (NT, no treatment) and 100 ng/ml VEGF-A (15 min)–stimulated paracellular permeability measured in confluent cultures of VEC-null ECs reexpressing VEC-WT, VEC-Y658F, VEC-Y685F, or VEC-Y733F. Values are means ± SEM from four independent experiments of triplicates (*, P < 0.05; **, P < 0.01). (E and F) Transmigration of fluorescently labeled ID8-VEGF ovarian (E) or B16F10 melanoma (F) tumor cells across confluent cultures of VEC-null ECs reexpressing VEC-WT, VEC-Y658F, VEC-Y685F, and VEC-Y733F and analyzed at 3, 6, and 24 h. Values are means ± SEM from three independent experiments of triplicates (*, P < 0.05; **, P < 0.01; ***, P < 0.001).
Figure 8.
Figure 8.
Genetic FAK inhibition in ECs prevents VEGF-enhanced tumor cell extravasation and VEC-Y658 phosphorylation in vivo. (A and B) VEGF-A enhances ID8 tumor cell extravasation. Fluorescently labeled ID8 or ID8-VEGF cells were tail vein injected into C57BL/6 mice, and after 6 h (A) or 16 h (B), tumor cell foci were enumerated in the lung tissue. Values (red bars) are means ± SEM (***, P < 0.001). (C and D) Inhibition of ovarian tumor cell extravasation in FAK-KD mice. Fluorescently labeled ID8-VEGF cells were tail vein injected into FAK-WT or FAK-KD mice, and after 6 h (C) or 16 h (D), tumor cell foci were enumerated in lung tissue. Values (red bars) are means ± SEM (***, P < 0.001). (E) Representative lung section images from FAK-WT and FAK-KD mice from analyses in D. ID8-VEGF ovarian carcinoma cells were prelabeled with DiIC12(3), and blood vessels were in situ labeled by FITC-lectin. Bar, 100 µm. (F and G) Inhibition of melanoma cell extravasation in FAK-KD mice. Fluorescently labeled B16F10 cells were tail vein injected into FAK-WT or FAK-KD mice, and after 6 h (F) or 16 h (G), tumor cell foci were enumerated in lung tissue. Values (red bars) are means ± SEM (*, P < 0.05). (H) In vivo signaling. C57BL/6 mice were tail vein injected with ID8 or ID8-VEGF tumor cells, and after 6 h, lung lysates were evaluated by VEGFR-2 immunoprecipitations (IP), phosphotyrosine (pY), and VEGFR-2 blotting. VEC pY658 and FAK pY397 immunoblotting were performed on lung lysates, and blots were sequentially reprobed for total VEC and FAK. GAPDH is the loading control. WCL, whole-cell lysate. (I) Inhibition of VEC pY658 in vivo. FAK-WT or FAK-KD mice were tail vein injected with ID8-VEGF cells or PBS (control), and lung protein lysates were made after 6 h. Independent samples (WT1–3) and (KD1–3) from different mice were evaluated by VEGFR-2 immunoprecipitations, pY, and VEGFR-2 blotting. VEC pY658 and FAK pY397 immunoblottings were performed on lung lysates, and blots were sequentially reprobed for total VEC and FAK. GAPDH is the loading control. (J) Densitometry of VEC pY658 lung lysate immunoblots from FAK-WT and FAK-KD injected with PBS, ID8-VEGF, or B16F10 tumor cells. Values are a mean ratio (pY658/total VEC) ± SEM from four independent lungs per experimental point. FAK-WT1–PBS was set to 100. (*, P < 0.05).
Figure 9.
Figure 9.
Inhibition of B16F10 spontaneous tumor metastasis but not primary tumor growth in FAK-KD mice. (A–D) B16F10 cells were tail vein injected into FAK-WT or FAK-KD mice, and lungs were evaluated after 14 d. (A) Representative Bouin’s-stained lung images. B16F10 cells are dark from melanin production. (B) Metastatic B16F10 tumor sites were enumerated in H&E-stained lung sections from FAK-WT and FAK-KD mice. Box–whisker plots show the distribution of the data: black square, mean; bottom line, 25th percentile; middle line, median; top line, 75th percentile; and whiskers, fifth and 95th percentiles (***, P < 0.001). (C) Mean total lung mass ± SD from FAK-WT (n = 11) and FAK-KD (n = 10) mice injected with B16F10 cells (**, P < 0.01; black bars) and lung mass from nontumor (NT)-bearing mice. (D) Mean area of individual B16F10 lung metastases from FAK-WT and FAK-KD mice were measured using ImageJ from H&E-stained lung sections. Means ± SEM. (E) B16F10 subcutaneous tumor growth over 20 d in FAK-WT (n = 11) and FAK-KD (n = 16 mice). Values are means ± SD. (F–I) DsRed-labeled B16F10 melanoma cells were injected between the skin and the cartilage on the dorsal side of FAK-WT and FAK-KD mouse ears. Primary tumor growth and lung metastasis were evaluated after 21 d. (F) Representative fluorescent images of tumors on the ears of FAK-WT and FAK-KD mice. Arrows indicate the approximate B16F10 injection sites, and ears are encircled (dotted lines). Bars, 3 mm. (G) B16F10 primary tumor size in FAK-WT (n = 10) and FAK-KD (n = 9) mice. Values are means ± SEM. (H) Representative lung section images from FAK-WT and FAK-KD mice. Blood vessels were in situ labeled by FITC-lectin, and B16F10 melanoma cells were visualized by intrinsic fluorescence (DsRed). Bars, 100 µm. (I) Spontaneous B16F10 lung micrometastasis in FAK-WT and FAK-KD mice. 10 lung sections per mouse were enumerated, and values (red bars) are means ± SEM (***, P < 0.001).
See this image and copyright information in PMC

Similar articles

See all similar articles

Cited by

See all "Cited by" articles

References

    1. Abu-Ghazaleh R., Kabir J., Jia H., Lobo M., Zachary I. 2001. Src mediates stimulation by vascular endothelial growth factor of the phosphorylation of focal adhesion kinase at tyrosine 861, and migration and anti-apoptosis in endothelial cells. Biochem. J. 360:255–264 10.1042/0264-6021:3600255 - DOI - PMC - PubMed
    1. Allingham M.J., van Buul J.D., Burridge K. 2007. ICAM-1-mediated, Src- and Pyk2-dependent vascular endothelial cadherin tyrosine phosphorylation is required for leukocyte transendothelial migration. J. Immunol. 179:4053–4064 - PubMed
    1. Arias-Pulido H., Chaher N., Gong Y., Qualls C., Vargas J., Royce M. 2012. Tumor stromal vascular endothelial growth factor A is predictive of poor outcome in inflammatory breast cancer. BMC Cancer. 12:298 10.1186/1471-2407-12-298 - DOI - PMC - PubMed
    1. Baumeister U., Funke R., Ebnet K., Vorschmitt H., Koch S., Vestweber D. 2005. Association of Csk to VE-cadherin and inhibition of cell proliferation. EMBO J. 24:1686–1695 10.1038/sj.emboj.7600647 - DOI - PMC - PubMed
    1. Bobek V., Kolostova K., Pinterova D., Kacprzak G., Adamiak J., Kolodziej J., Boubelik M., Kubecova M., Hoffman R.M. 2010. A clinically relevant, syngeneic model of spontaneous, highly metastatic B16 mouse melanoma. Anticancer Res. 30:4799–4803 - PubMed

Publication types

MeSH terms

Substances