The fibronectin III-1 domain activates a PI3-Kinase/Akt signaling pathway leading to αvβ5 integrin activation and TRAIL resistance in human lung cancer cells

All reagents were purchased from Sigma (St. Louis, MO) unless indicated otherwise. Fetal bovine serum (FBS) was from Hyclone (Logan, UT). Recombinant fibronectin type III domains FnIII-1c, FnIII-10n and FnIII-13 were generated and purified as previously described [17]. Recombinant human TRAIL/TNFSF10 was purchased from R&D systems (Minneapolis, MN). The PI3K inhibitor LY294002 was purchased from Biomol (Plymouth Meeting, PA) and the Akt inhibitor VIII (AG730) was purchased from Sigma. Polyclonal rabbit antibody against phospho-Akt (Ser473), Hoechst 33342 stain for visualization of nuclei, and the Click-iT® TUNEL AlexaFluor® Imaging assay were purchased from Invitrogen (Carlsbad, CA). The rabbit monoclonal antibodies against cleaved caspase 8 and GAPDH were both purchased from Cell Signaling Technology (Danvers, MA). Mouse monoclonal antibodies against integrin αvβ5 (15 F11), integrin β3 (LM609), integrin αv (MAB1980), α5β1, and the monoclonal blocking antibody against integrin αvβ5 (P1F6) were purchased from Millipore (Billerica, MA). The purified rat anti-mouse CD29 (clone 9EG7) which recognizes the ligand bound conformation of the β1 integrin [18] was purchased from BD Biosciences. The rabbit polyclonal antibody to vitronectin, AC7, has been previously described [19]. Pre-immune normal rabbit IgG was used as a control. Alexafluor488-conjugated secondary anti-mouse IgG (H + L) antibody was purchased from Molecular Probes (Eugene, OR). Horseradish peroxidase conjugated secondary antibodies to mouse IgG (H + L) and rabbit IgG (H + L) were purchased from BioRad (Berkeley, CA).

The human tumor cell line NCI-H460 was purchased from the American Type Culture Collection (Manassas, VA). NCI-H460 cells were grown in monolayer culture in complete medium (RPMI 1640 with streptomycin-penicillin and glutamax supplemented with 10 % FBS) at 37 °C in a humidified atmosphere containing 5 % CO₂. Prior to treatment, cells were serum-starved in RPMI-1640 with 0.1 % BSA for 2 h. For the collection of whole cell lysates, monolayers were washed twice in ice-cold PBS and lysed in whole cell lysis buffer (100 mM Tris–HCl, pH  6.8, 2 % SDS, 10 % glycerol, 100 mM DTT).

NCI-H460 cells were cultured in complete medium 48 h or until cells reached ~85 % confluency, rinsed once with PBS and serum-starved in RPMI-1640 with 1 % BSA for 2 h at 37 °C in a humidified atmosphere containing 5 % CO₂. Cells were stimulated with recombinant human TRAIL/TNFSF10 (TRAIL) (R&D systems) or PBS as a control treatment as indicated in the figure legends. Apoptosis was assessed by cleaved caspase 8 protein levels via western blot analysis or by TUNEL assay. The TUNEL assay was performed using Click-It® TUNEL AlexaFluor488® Imaging Kit (Invitrogen) according to the manufacturer’s protocol. In brief, cells were fixed with 4 % paraformaldehyde in PBS at room temperature for 20 min and permeabilized with Triton X-100 (0.25 % in PBS) for an additional 15 min. The cells were then washed twice and incubated with 50 μL of terminal deoxynucleotidyl transferase reaction buffer (Component A) for 10 min at room temperature. The buffer was removed and the cells were incubated with TUNEL reaction mixture containing terminal deoxynucleotidyl transferase for 1 h in a humidified chamber at 37 °C for 1 h. Post treatment, cells were washed three times with 3 % BSA in PBS for 2 min each and then incubated with 50 μL of Click-iT reaction mixture (containing Alexa 488) for 30 min at room temperature, protected from light. The cells were then washed with 3%BSA in PBS and the nuclei were stained with Hoechst 33342 for 1 min at room temperature, protected from light. The coverslips were washed twice with PBS before mounting onto a slide with ProLong® Gold Antifade Mount (Life technologies). Cell monolayers were examined using an Olympus BMX-60 microscope equipped with a cooled CCD sensi-camera (Cooke, Auburn Hills, MI), and images were acquired using Slidebook software (Intelligent Imaging Innovation, Denver, CO). Fluorescence images were processed with ImageJ analysis software for the quantification of TUNEL-positive nuclei and total number of nuclei. The number of TUNEL-positive cells in three random (20×) fields was counted and divided by the total number of nuclei to determine the percentage of TUNEL-positive nuclei.

Whole cell lysates were collected from treated cells and subjected to SDS-PAGE under reducing conditions and transferred onto nitrocellulose membranes (Schleicher and Schuell Bioscience, Keene, NH). Membranes were blocked for either 2 h at room temperature or overnight at 4 °C with 5 % BSA (w/v) in Tris-buffered saline containing 0.1 % Tween-20 then incubated with primary antibodies for 2 h at room temperature or overnight at 4 °C. Blots were washed with Tween-20 and incubated with horseradish peroxidase-linked secondary antibodies (1:10,000) for 1 h at room temperature. Immunoreactive bands were detected using Clarity^TM Western ECL substrate (BioRad). Blots were reprobed after stripping in 62.5 mM Tris–HCl, pH 6.8, 2 % SDS and 10 mM β-mercaptoethanol for 20 min at 60 °C. Western blots were quantified either by ImageJ analysis software (for blots developed on film) or by ChemiDoc™ MP Imaging System with Image Lab (BioRad).

Polystyrene non-tissue culture-treated 48-well plates (Greiner bio-one; Monroe, NC) were coated with vitronectin in PBS overnight at 4 °C. Wells were blocked with 3 % BSA in PBS for 2 h in room temperature. Adherent H460 cells were lifted with Cellstripper® (Cellgro) and resuspended in RPMI-1640 containing 0.1 % BSA. Suspended cells (~10⁵) were serum-starved for 1 h then treated with FnIII-1c or PBS, as a control, for an hour at 37 °C. When pharmacological inhibitors to PI3K (Wortmannin, LY294002) and Akt1/2 (VIII) were used, cells were pre-treated with inhibitors for 30 min prior to treatment with FnIII-1c for an additional hour. Treated cells were seeded onto vitronectin-coated wells at a density of approximately 4x10⁴cells/wells and allowed to adhere for 1 h at 37 °C. Adherent cells were quantified by staining with 0.05 % toluidine blue for 1 h. Each well was washed four times and dye was extracted with 10 % acetic acid. Absorbance was measured at 650 nm and corrected for light scattering by subtracting the absorbance at 405 nm.

NCI-H460 cells (~10⁴ cells/mL) were cultured in complete medium for 48 h on glass coverslips, then serum-starved for 1 h. Serum-starved cells were washed once with PBS, fixed for 20 min in 4 % paraformaldehyde, permeabilized in 0.5 % TritonX-100 for 10 min, blocked in 1 % BSA and immunostained with monoclonal antibodies against integrins α5β1 (1:200 dilution), αvβ5 (1:200) αvβ3 (1:100) and 9EG7, a monoclonal antibody that detects ligated β1 integrins (1:100) for 1 h at room temperature or overnight at 4 °C. Slides were then incubated with Alexafluor488-conjugated secondary anti-mouse IgG (H + L) antibody for an additional hour at room temperature. Nuclei were visualized with Hoechst 33342 dye. After staining, slides were mounted with Prolong Antifade according to the manufacturer’s instructions (Molecular Probes) and examined using an Olympus BMX-60 microscope equipped with a cooled CCD sensi-camera (Cooke, Auburn Hills, MI). Images were acquired using Slidebook software (Intelligent Imaging Innovation, Denver, CO).

Non-small cell human lung carcinoma tissue LC241 NSCLC microarray panels (US Biomax Inc., Rockville, MD), were immunostained using the peroxidase-based ABC system (Vector Laboratories, Burlingame, CA). Vitronectin was detected using a polyclonal antibody to vitronectin and for negative controls, the primary antibody was replaced with pre-immune normal rabbit IgG. Color was developed by reaction with 3,3’-Diaminobenzidine. Tissue sections were counterstained with hematoxylin.

The NCI-H460 cell line was used to evaluate the effect of the unfolded fibronectin type III-1 domain on TRAIL-induced apoptosis. To determine the sensitivity of NCI-H460 cells to TRAIL we performed dose and kinetics studies using recombinant human TRAIL and evaluated two indicators of apoptosis: caspase 8 cleavage and TUNEL Assay. The TUNEL assay detects the DNA fragmentation which occurs in cells undergoing apoptosis. The binding of TRAIL to DR4/DR5 stimulates receptor aggregation and recruitment of caspase 8 to the receptor cytoplasmic domain resulting in the cleavage of caspase 8 and the initiation of the apoptotic program. Cell monolayers were treated with increasing concentrations of TRAIL for 1 h and assessed for cleaved caspase 8 by western blotting. Caspase 8 cleavage was detected at 1 h using 100 ng/ml TRAIL with higher amounts seen in cells treated with 200 and 500 ng/ml TRAIL (Fig. 1a). Kinetic studies using a fixed amount of TRAIL (100 ng/ml) indicated that caspase 8 cleavage occurred after 60 min reaching maximal levels within 90 min of treatment (Fig. 1b). Similar observations were made with the apoptosis assay, in which TUNEL staining was seen within 1 h following treatment with 50 ng/ml TRAIL (Fig. 1c). As evident from the nuclear staining (Hoechst), TRAIL treatment also resulted in the loss of apoptotic cells from the substrate.

To examine the effect of fibronectin Type III domain unfolding on TRAIL signaling, NCI-H460 cells were pre-treated with the FnIII-1c peptide for 1 h, and then incubated with TRAIL for an additional 2.5 h. Whole cell lysates were collected and analyzed for cleaved caspase 8 by western blotting. NCI-H460 cells treated with TRAIL alone expressed cleaved caspase 8 (Fig. 2a), while control cells treated with PBS or FnIII-1c did not. Pre-treatment of cells with FnIII-1c significantly decreased the amount of caspase 8 cleavage in response to TRAIL (Fig. 2a–b). This inhibition was specific to FnIII-1c, as pre-treatment with other fibronectin type III domains (FnIII-10n and FnIII-13) had no significant effect on TRAIL-mediated caspase 8 cleavage (Fig. 2a–b). FnIII-10n and FnIII-13 were selected as controls due to shared characteristics with the FnIII-1c domain, i.e., heparin binding activity (FnIII-13) [20], mechanically unfolded stable intermediate structure (FnIII-10n) [6]. We also examined the amount of TUNEL staining in cells treated with TRAIL in the absence or presence of FnIII-1c (Fig. 2c–d). As shown in panel C, pre-incubation of cells with FnIII-1c greatly decreased TUNEL staining in TRAIL treated cells (Fig. 2c–d). Pretreatment of cells with control type III domains, III-10n and III-13, had no effect on TUNEL staining in response to TRAIL (Fig. 2c–d).

Integrins are a family of cell adhesion receptors known to be involved in tumor progression [21] and aberrant integrin activation is associated with increased metastasis, enhanced cancer cell survival, and resistance to chemotherapy [22]. Earlier studies have shown that integrins negatively regulate Fas induced apoptosis [23]; therefore, we investigated whether integrins were involved in FnIII-1c-mediated inhibition of TRAIL-induced apoptosis. To identify the integrins mediating adhesion of NCI-H460 cells to the substrate, cells cultured for 48 h were stained with monoclonal antibodies to integrins α5β1, αvβ5, αvβ3, and also with a β1 antibody (9EG7) directed against ligand-occupied β1 integrins (β1*) [18]. Only the vitronectin binding integrin, αvβ5, was localized in clusters typical of adhesion complexes (Fig. 3a). In agreement with earlier studies showing low levels of αvβ3 on NCI-H460 cells [24], there was no staining for integrin αvβ3. There was also no staining for the α5β1 integrin consistent with earlier reports showing low levels of fibronectin synthesis by NCI-H460 cells [25]. Staining with the 9EG7 antibody was also negative. Both the αvβ3 and β1 integrins could be stained in control experiments using fibroblasts (data not shown). These results indicate that adhesion of H460 cells was dependent on the interaction of integrin αvβ5 with vitronectin deposited on the substrate from the serum (Fig. 3a).

To test whether FnIII-1c was affecting the activation state of integrin αvβ5, we evaluated cell adhesion over a range of vitronectin concentrations. FnIII-1c or PBS control treated cells were seeded onto vitronectin-coated plates and allowed to adhere for 1 h. As displayed in panel B, FnIII-1c significantly increased cell adhesion to vitronectin across a range of coating concentrations, consistent with an increase in integrin activation (Fig. 3b). Experiments were then done to determine whether the change in integrin activation might be required for FnIII-1c’s inhibition of TRAIL-induced apoptosis. Treatment of cells with a blocking antibody to αvβ5 (P1F6) partially restored TRAIL-induced caspase 8 cleavage in FnIII-1c treated cells (Fig. 3c). Analysis of data from several experiments indicated that this restoration was statistically significant (Fig. 3d). Control experiments using LM609, the integrin αvβ3 blocking antibody, or a non-immune mouse IgG, had no effect on FnIII-1c-mediated inhibition of TRAIL-induced caspase 8 cleavage (Fig. 3c–d). These data suggest that the increase in activation and ligation of integrin αvβ5 was contributing to FnIII-1c-mediated inhibition of TRAIL-induced apoptosis.

Previous studies have shown that the PI3K-Akt pathway can regulate the activation of integrin receptors through inside-out signaling [26]; therefore, we investigated whether Akt was involved in FnIII-1c-mediated activation of integrin αvβ5. Two inhibitors of the PI3K-Akt pathway, VIII and LY294002 (Akt1/2 inhibitor [27] and PI3K inhibitor, respectively) were tested for their effects on FnIII-1c-mediated activation of the αvβ5 integrin. As shown in Fig. 3e, both LY294002 and VIII significantly attenuated FnIII-1c’s ability to increase adhesion to vitronectin. A similar result was obtained using the PI3K inhibitor, Wortmannin (data not shown). These data suggest that FnIII-1c-mediated activation of integrin αvβ5 is dependent on the PI3K-Akt pathway. To evaluate the effect of FnIII-1c on Akt activation, NCI-H460 cells were treated with various concentrations of FnIII-1c for 1 h and the activation of Akt, as indicated by the phosphorylation of Akt at serine 473 (pS473Akt), was assessed by western blotting. Akt phosphorylation in response to FnIII-1c was observed at 5–10 μM (Fig. 4a) within 1 h and remained elevated for several hours (Fig. 4b). The activation of Akt was specific to FnIII-1c, as treatment with the control modules, FnIII-10n (Fig. 4c) and FnIII-13 (Fig. 4d), did not result in an increase in pS743Akt.

In order to evaluate the role of Akt in regulating the inhibition of TRAIL signaling by FnIII-1c, NCI-H460 cells were pre-treated with PI3K and Akt inhibitors prior to incubation with FnIII-1c and TRAIL and assessed for Akt activation and caspase 8 cleavage by western blot. Treatment of cells with PI3K and Akt inhibitors alone reduced baseline levels of pS473Akt, but did not induce caspase 8 cleavage (Fig. 5a) or TUNEL staining (Fig. 5b, c) suggesting that inhibition of the PI3K-Akt pathway alone was not sufficient to initiate apoptotic pathways. TRAIL treatment resulted in caspase 8 cleavage, which was attenuated by the addition of FnIII-1c (Fig. 5a). Both inhibitors of the PI3K-Akt pathway, VII and LY294002, restored TRAIL-dependent caspase 8 cleavage, indicating that FnIII-1c-mediated inhibition of TRAIL-signaling is dependent on the PI3K-Akt pathway. Similar results were seen using TUNEL assays. As shown in Fig. 5b and quantitated in Fig. 5c, incubation of cells with TRAIL resulted in an increase in TUNEL staining, which was completely prevented by pretreatment with FnIII-1c. Pre-incubation of cells with the inhibitors of the PI3-Akt pathway, VIII and LY294002, restored TRAIL-induced apoptosis in the presence of FnIII-1c. These data were quantified across three independent experiments (Fig. 5c). Taken together, these data suggest that the unfolded III-1 domain of fibronectin mediates TRAIL resistance of NCI-H460 cells by activation of the αvβ5 integrin through the PI3K/Akt signaling pathway.

Our data suggest that unfolding of Type III domains within the fibronectin matrix may promote TRAIL resistance which depends on increased activation and ligation of αvβ5 integrin receptors. The primary ligand for the αvβ5 integrin is the plasma protein vitronectin [28]. High levels of ECM proteins are found in the NSCLC stroma [29] and the overexpression of fibronectin has been well documented to be associated with metastatic progression and drug resistance [30]. Much less is known about the distribution and function of vitronectin in NSCLC. Using a tissue array, we examined the localization of vitronectin in human NSCLC lung tumor sections. NSCLC human tumor samples were immunostained for vitronectin (Fig. 6a). VN expression was observed (Fig. 6a) surrounding blood vessels (filled arrows), in the interstitial space between the stroma and tumor (open arrows), and in areas of necrosis (arrow heads). In addition to being quite prominent in areas of necrosis, vitronectin staining was evident consistently around the arterioles (Fig. 6b, black arrow on left) and small arteries (Fig. 6b, black arrow on right). Positive staining was also seen in the tumor-stromal interface. This staining was not present throughout the entire interface but was restricted to small localized areas (Fig. 6c). The source of the vitronectin in these tissues is not known but may arise from the leakage of plasma proteins around blood vessels and localized synthesis by cells within the stroma. Although vitronectin is not typically synthesized by cells other than hepatocytes, reports have documented vitronectin synthesis by monocytes/macrophages [31] and lung epithelial cells [32]. Taken together, the data are consistent with a model whereby tumor cells proximal to both fibronectin and vitronectin within the stromal matrix represent a cohort of TRAIL-resistant cells.