Activated tumor cell integrin αvβ3 cooperates with platelets to promote extravasation and metastasis from the blood stream.

Cell lines, labeling, and in vivo model MDA-MB 435 human tumor cells were from Dr. Janet Price (MD Anderson). Variants of the parental cells lacking αvβ3 and reconstituted with αvβ3WT or αvβ3D723R were previously described ( 13 ). Clone E9 cells were isolated by limiting dilution of the parental cells. Parent Combo cells represent a pool of 20 clones, each expressing non-activated αvβ3 based on in vitro migration and blood perfusion as detailed below. BCM2 were isolated from MDA-MB 435 parental cells, after adding these to human blood, perfusing them across immobilized collagen I at a venous wall shear rate, and recovering the adherent tumor cells. M21 human melanoma cells express activated αvβ3 as reported ( 16 ). For in vivo tracking, the tumor cells were stably transduced with red fluorescent protein ( dsRed2 ) or firefly luciferase ( F-luc ) using a self-inactivating lentiviral vector ( 17 ). For hematogenous metastasis, 1×10 5 or 1×10 6 tumor cells were injected into the tail vein of 5- to 8-week old female C.B17/lcTac- Prkdc scid mice (Taconic) and metastatic activity followed by non-invasive bioluminescence imaging (IVIS200) ( 18 ), fluorescence microscopy of lung whole mounts, or by real time PCR of human alu sequences in lung extracts using sense primer 5’ ACG CCT GTA ATC CCA GCA CTT 3’ and antisense primer 5’ TCG CCC AGG CTG GAG TGC A 3’ ( 19 ).

Induction of thrombocytopenia and in vivo tumor cell extravasation

Thrombocytopenia was induced by i.v. injecting SCID mice with anti-murine GPIbα (CD42b) (Emfret) (3 µg/g bodyweight). Platelet counts were measured using an automated blood cell counter adjusted to murine blood. Tumor cells were injected 4 h after induction of thrombocytopenia. Tumor cell extravasation from the pulmonary microvasculature was analyzed 3 days later. The lungs were inflated with 10% neutral buffered formalin through the trachea, post fixed, and cryosectioned (35µm sections). The endothelium was stained with anti-CD31 (PECAM) and tumor cells with anti-human CD44. Quantification was performed using a Zeiss Axio Imager M1m microscope, which allowed for analysis of a large number of events. Localization of cancer cells inside or outside the vasculature was confirmed by confocal microscopy with a 40× water immersion objective using an Olympus IX81 equipped with UltraVIEW VoX Confocal Imaging System (Perkin Elmer). Images were acquired and analyzed with Velocity software.

Flow cytometry

Integrin expression and αvβ3 activation state were analyzed by flow cytometry (FACS Calibur, Becton Dickinson) with anti αvβ3 mAb VNR1-27.1 ( 20 ), or activation-dependent human ligand-mimetic scFv antibodies Bc-12 and Bc-15 ( 16 ). ScFv binding was detected with M2 mouse anti-FLAG and anti-mouse-APC using TBS with or without 1mM EDTA, 1mM Ca 2+ , or 0.1 mM Mn 2+ as binding and washing buffer. In vitro migration and arrest during blood flow Haptotactic migration toward human extracellular matrix proteins was detailed earlier ( 15 ). Cancer cell arrest during blood flow was measured as described ( 13 ). Briefly, dsRed tagged tumor cells (red fluorescence) were suspended in human blood anticoagulated with 50 nM -Pro-Arg-chloro methyl ketone (PPAK), spiked with 10 µM mepacrine (green fluorescence) and perfused over collagen I or subendothelial matrix at a venous wall shear rate of 50 s −1 (4 dynes/cm 2 ). Adhesive events and cell interactions were recorded by fluorescence video microscopy and quantified at 50 predefined positions by image analysis using MetaMorph software. Tumor cell-platelet interaction during blood flow was further analyzed with a dual laser system (486 nm and 580 nm) and differentiating band pass filters (510 nm and 580 nm), using calcein-green labeled tumor cells and calcein-red labeled mouse platelets, reconstituted into platelet-poor murine plasma anticoagulated with heparin. Tumor cell-platelet interactions were recorded by digital video microscopy and analyzed with MetaMorph software.

Dorsal skin fold chamber and intravital microscopy

Tumor cell interaction with the pulmonary microvasculature was observed by intravital microscopy in lung allografts grown in dorsal skin fold chambers on SCID mice as described ( 21 ). Briefly, titanium frames were implanted into a dorsal skin fold, transplanted with SCID lung allografts 3 days later, and allowed to vascularize for 14 days. The allografts were superfused with TNFα (50µl at 1µg/ml) 2.5 hours before injecting 2.5×10 6 dsRed tagged tumor cells i.v. together with FITC-dextran (500 kDa) to visualize the vasculature. Adhesive events were observed with an intravital microscope (Leitz Biomed) ( 21 ).

Transendothelial migration and platelet preparation

Human pulmonary microvascular endothelial cells were grown for 5 days on transwell filters (8 µm pores) coated with human fibronectin (10µg/ml). Monolayer integrity was measured by transendothelial resistance. Conditioned medium from IMR-90 fibroblasts in the lower chamber was used to attract tumor cells (4×10 4 ) seeded atop the endothelial layer during a 16 hr migration period. Where indicated, washed human platelets (433×10 3 /µl) were added to the tumor cell suspension, with or without stimulation with thrombin receptor activating peptide TRAP-6 (12.5 µM). Non-migrated tumor cells were removed from the filter tops with cotton swaps, filters fixed in 4% paraformaldehyde, and transmigrated tumor cells counted by fluorescence microscopy on the filter underside in 5 optical fields. Effects of platelets and their releasates on the integrity of human pulmonary endothelial monolayers in the presence of tumor cells was measured by incubating intact endothelial layers (transendothelial resistance >30Ω/cm 2 ) with BCM2 cells and washed human platelets stimulated with ADP, or with platelet releasates for increasing time periods at 37 °C and 5% CO 2 .

Supplementary Material Video_Fig.1C Dynamic interaction of M21 melanoma cells with platelets during blood flow DsRed-expressing tumor cells (red fluorescence) were suspended in human blood anticoagulated with 50 nM D-phenylalanyl-L-prolyl-L-arginine chloromethyl ketone (PPAK), spiked with 10 µM mepacrine to visualize all cells and platelets (green fluorescence, note: tumor cells are now red and green) and perfused over collagen I, at a venous wall shear rate of 4 dynes/cm 2 . Adhesive events and cell interactions were recorded by fluorescence video microscopy, and quantified at 50 predefined positions by image analysis using MetaMorph software. The view field shows a group of tumor cells already attached in a platelet dependent manner. Additional tumor cells, entering the view field, arrest on adherent platelets and existing platelet-tumor heteroaggregates. Note that tumor cells can sometimes carry platelets at their surface and extend pseudopods when in contact with platelets (as shown by dual-laser confocal analysis in Figure 1D ). Pseudopod extension further facilitates tumor cell attachment during blood flow. 01