Published OnlineFirst May 30, 2013; DOI: 10.1158/0008-5472.CAN-13-0218

Cancer Tumor and Stem Cell Biology Research

Vitronectin–avb3 Engagement Directs Hypoxia- Resistant mTOR Activity and Sustained Synthesis Linked to Invasion by Breast Cancer Cells

Carolina Pola1, Silvia C. Formenti2,3, and Robert J. Schneider1,2,3

Abstract The tumor microenvironment is a crucial player in the ability of cancer cells to acquire the ability to survive under the hypoxic environment and promote migration and invasion. Translational regulation is an essential part of cancer development and progression. Protein synthesis consumes considerable cellular metabolic energy and is therefore highly regulated, in turn controlling tumor cell proliferation and survival in extreme tumor–host conditions. Protein synthesis is typically downregulated by hypoxia, impairing cell proliferation and migration. Here, we show that breast cancer cells expressing integrin avb3, when engaging the (ECM) protein vitronectin, strongly upregulate both mTOR activity and cap- dependent mRNA translation, which overrides their inhibition by hypoxia and facilitates tumor cell invasion. Interaction of vitronectin with integrin avb3 results in the continued activation of the kinase mTOR despite hypoxia through a mechanism that is dependent on integrin-linked kinase but is independent of focal adhesion kinase. Continuous activation of mTOR despite hypoxia involves release of translation initiation factor eIF4E from its repressor protein 4E-BP1, which is required for vitronectin-mediated tumor cell invasion. As integrin avb3 is associated with breast cancer cell invasion and to , we propose that the interaction with specific ECM can influence cancer cell invasion, in part, by hyperactivation of mTOR, thereby promoting and sustaining protein synthesis under hypoxic conditions. Cancer Res; 73(14); 1–8. 2013 AACR.

Introduction mRNA translation is a checkpoint that is blocked in response to Tumor cells and their microenvironment maintain a dynam- certain physiologic stresses, including low levels of oxygen ic interaction, exchanging growth factors and cytokines and during hypoxia (6). eIF4F consists of the cap-binding protein transforming a local extracellular matrix (ECM) into an acti- eIF4E, scaffold protein eIF4G, and the ATP-dependent RNA vated stroma (1, 2). The active host microenvironment can helicase, eIF4A (7). Assembly of eIF4F is controlled by the modify expression programs in the tumor cell and trigger availability of eIF4E through regulation by the 4E-binding signaling pathways that control cancer cell proliferation, proteins (4E-BP). Hyperphosphorylation of 4E-BP1 by the migration and invasion (3). Components of the tumor stroma mTOR kinase complex 1 (mTORC1) maintains the inactive include a highly structured and tissue-specific ECM, stromal state of 4E-BP1, so that it cannot bind eIF4E (8). Hypoxia cells, infiltrating inflammatory cells, activated fibroblasts, and downregulates mTORC1, activating (dephosphorylating) 4E- endothelial cells (4, 5). BP1, in turn sequestering eIF4E and downregulating cap- Protein synthesis is highly regulated at the initiation phase dependent mRNA translation (6, 9). Both the noncellular – and involves recruitment of ribosomes with the eukaryotic structural components of the ECM (10 12) and hypoxia (6) initiation factor 4F (eIF4F), a complex of proteins, to the m7 can regulate protein synthesis, but their interaction has never GTP capped structure at the 50-end of mRNAs. Initiation of been extensively investigated. We therefore asked whether external signals in the breast ECM play a role in uncoupling breast cancer cell mTOR inhibition during hypoxia to sustain Authors' Affiliations: Department of 1Microbiology and 2Radiation Oncol- protein synthesis, a requirement for cancer cell invasion and ogy, and 3NYU Cancer Institute, New York University School of Medicine, New York, New York disease progression (8). To date, ECM regulators of mRNA translation include Note: Supplementary data for this article are available at Cancer Research fi Online (http://cancerres.aacrjournals.org/). bronectin (10, 11) and integrin receptors, several of which are overexpressed on breast cancers and mediate cell–ECM Corresponding Author: Robert J. Schneider, NYU School of Medicine, Alexandria Center for Life Sciences, 450 East 29th Street, New York, NY interactions (13). The regulatory effect of different 10016. Phone: 212-263-6006; Fax: 646-501-4541; E-mail: on protein synthesis has also not been extensively studied. In [email protected] breast tumors, malignant progression and bone metastasis doi: 10.1158/0008-5472.CAN-13-0218 are strongly associated with the expression of integrin avb3 2013 American Association for Cancer Research. (13–15).

www.aacrjournals.org OF1

Downloaded from cancerres.aacrjournals.org on September 30, 2021. © 2013 American Association for Cancer Research. Published OnlineFirst May 30, 2013; DOI: 10.1158/0008-5472.CAN-13-0218

Pola et al.

In this study, we show that vitronectin engagement of minutes in Dulbecco's Modified Eagle's Media (DMEM). After integrin avb3, typically present on breast cancer cells meta- incubation, cells were added to dishes coated with either 25 static to bone, hyperactivates mTOR, even during hypoxia, mg/mL IV native or denatured, 10 mg/mL 1, or thereby overriding the downregulation of cap-dependent 2.5 mg/mL vitronectin. Cells were allowed to attach for 12 to 16 mRNA translation. Vitronectin is an ECM component localized hours. Plating efficiency was determined as the ratio of plated primarily in the stroma and is generally only abundantly in cells to adhered cells at 12 to 16 hours using the MTT assay for contact with tumor cells after breach of the basement mem- cell number. brane during cancer cell invasion (3, 16). We show that stimulation of breast cancer cell mTOR by integrin avb3- SDS-PAGE, immunoblot analysis vitronectin interaction is linked to strongly increased tumor Cells were washed with ice-cold 1 PBS and lysed in 0.5% cell invasion during hypoxia, through a mechanism that lysis buffer [50 mmol/L HEPES, pH 7.0, 150 mmol/L NaCl, 2 involves integrin-linked kinase (ILK) but not focal adhesion mmol/L EDTA, 0.5% NP-40, 25 mmol/L NaF, 2 mmol/L sodium kinase (FAK), and mTOR activation of eIF4E-dependent mRNA orthovanadate, 25 mmol/L glycerophosphate, protease inhib- translation. Thus, integrin-mediated stimulation of protein itor (Roche)], clarified by centrifugation at 4C for 15 minutes synthesis directed by specific ECM components plays an at 13,000 rpm, protein concentration determined, solubilized important role in promoting hypoxia resistance and tumor and SDS-PAGE carried out, and protein immunoblotting con- cell invasion. ducted. Immunoblotting and protein detection by chemilumi- nescence were carried out (Amersham). Material and Methods Cell lines, cell culture, plasmids, and antibodies Protein synthesis m 35 Mouse monoclonal anti-eIF4A antibody was provided by W. Cells were labeled with 50 Ci of [ S]- per mL Merrick (Case Western Reserve University, Cleveland, OH). (Easytag Express Protein Labeling Mix, Dupont/NEN) in Other antibodies were from commercial sources and included DMEM without cold methionine for 1 hour, washed twice fi rabbit polyclonal antibodies to eIF4E, eIF4G, 4E-BP1-P (Thr 37/ with ice-chilled 1 PBS, lysed, and clari ed as above and fi 46), p70S6K, p70S6K-P (Thr 389), S6rp, S6rp-P (Ser 235/236), speci c activity determined by trichloroacetic acid (TCA) fi AMPKa-P (Thr 172), FAK, FAK-P (Tyr 576/577), mTOR, mTOR- precipitation onto GF/C lters and liquid scintillation count- P (Ser 2448), Akt and Akt-P (Ser 473), ILK, eIF2a and a-P (Ser ing. In hypoxic cells, labeling was conducted using pre-equil- 51; Cell Signaling Technology); integrin b3 (BD Biosciences ibrated media within the hypoxic chamber. Pharmingen), horseradish peroxidase (HRP)-conjugated don- key anti-rabbit or sheep anti-mouse secondary antibodies Lentivirus short hairpin RNA expression vectors (Amersham). Breast cancer cell lines MDA-MB-435, HTB-20, Short hairpin RNAs (shRNA) and scrambled nonsilencing MCF10A, and MDA-MB-231 were obtained from the American (NS) shRNAs were delivered by transduction of cells with lentivirus shRNA expression vectors. Double-stranded shRNAs Type Culture Collection. Cell lines were subjected to IMPACT 0 testing, used within 6 months for this study and grown as for cloning into lentivirus vectors were directed to the 3 - recommended. Human plasma collagen IV, laminin 1 (from untranslated region of mRNAs. Viruses were produced and of Engelbreth-Holm-Swarm mouse sar- cells infected/selected with polybrene (4 mg/mL) and puro- m coma), and vitronectin from human plasma were from Sigma. mycin at 0.8 g/mL for 48 hours. MAb 1976/LM609 (anti-human integrin avb3) was from Che- micon International. Transfection of plasmid DNA was carried MTT proliferation assay m Ninety-six–well plates coated with vitronectin (2.5 mg/mL) out using standard conditions (Fugene) with 1 to 2 g of DNA. Transfection of siRNA to eIF4E was carried out on cells at 50% at 37 C for 1 hour were blocked with 1% bovine serum albumin 3 confluency using 5.6 mLof20mmol/L siRNA per 5 10 cells by (BSA) in PBS. MB-435 cells (2 10 per well) were plated in Oligofectamine (Invitrogen), according to manufacturer's triplicate in 1% serum in DMEM and time-lapse analyzed for 7 instructions, in the absence of serum and antibiotics. Normal days. MTT dye (Sigma) was added to wells and incubated for 4 growth media was replaced after 4 to 6 hours. siRNA to hours, cells were solubilized with MTT stripping buffer (10% eIF4E (Ambion): AAGCAAACCUGCGGCUGAUCU. Nonsilen- SDS, 45% isopropanol, 0.04 N HCl) per well, and optical density cing RNA was purchased from Ambion. determined at 570 nm.

Hypoxia treatments Cell invasion assay Hypoxia experiments were carried out using a BioSphere Filter undersides of Transwells were coated with vitronectin hypoxia incubator at 0.5% O2, 95% N2,5%CO2 with growing (2.5 mg/mL) at 37 C for 1 hour, plates blocked as above and cells subjected to hypoxia for 24 hours unless otherwise noted. 1 102 cells per well previously treated with control or LM609 Normoxic cells were grown and treated at atmospheric oxygen. Ab suspended in migration buffer (RPMI-1640 containing 1 mmol/L MgCl2, 0.2 mmol/L MnCl2, 0.5% BSA), added to the Adhesion procedures upper chamber of Transwell migration chambers (Membranes MDA-MB-435 cells were collected by mild trypsinization and 8.0 Am) and allowed to invade for 16 hours at 37C. Cells incubated with rotation at 37C with control Ab (IgG-isotype remaining on the topside of the membrane were removed and matched control) or anti-avb3 (LM609) at 75 mg/mL for 45 cells that had invaded were stained with crystal violet and

OF2 Cancer Res; 73(14) July 15, 2013 Cancer Research

Downloaded from cancerres.aacrjournals.org on September 30, 2021. © 2013 American Association for Cancer Research. Published OnlineFirst May 30, 2013; DOI: 10.1158/0008-5472.CAN-13-0218

Vitronectin Maintains Protein Synthesis in Hypoxia

A Cell lines B * 120 * 100

HTB20 MCF10A MB-435 MB-231 80 ITG αv 60 40 ITG β3 20 Plating efficiency 0 - + - + - + - + LM609 Ab MB-435 MB-231 HTB-20 MCF10A

Figure 1. Plating efficiency of MB-435 and MB-231 cells to vitronectin occurs primarily through avb3 integrin. A, equal amounts of cytoplasmic protein lysates from HTB-20, MCF10A, MB-435, and MB-231 cells were examined by SDS-PAGE and immunoblot analysis for expression of integrin (ITG) av and b3. B, cell plating efficiency on vitronectin-coated plates was determined from 3 independent studies. Plating efficiency was determined as the average of the ratio of retained to plated cells determined by MTT assay at 12 to 16 hours postplating. Results are normalized to MB-435 cells on vitronectin with control antibody (actual plating efficiency of control was 85%). Control antibody (Ab) or anti-avb3 LM609 Ab was added at 75 mg/mL for 45 minutes before cell plating. , P < 0.05 by the Student t test. quantified at 600 nm. Data are presented as the mean optical fold lesser extent, MD-MB-231 cells (Fig. 1A; ref. 17). Most density and SD from triplicate cultures. breast cancer cell lines express av integrin, as shown here. Adhesion of MB-435 and MB-231 cells to vitronectin-coated Results plates was blocked by an antagonizing antibody to avb3 Adhesion of breast cancer cells to vitronectin overrides (LM609), whereas plating efficiency of HTB-20 or MCF10A hypoxia inhibition of protein synthesis cells, which do not express b3 integrin, was not impaired (Fig. The most common site of breast cancer metastasis is to 1B). Therefore, the major means for interaction of MB-435 and bone, which requires cancer cell expression of avb3 integrin, a MB-231 cells with vitronectin is through avb3 integrin. binding partner of the ECM protein vitronectin. While most We therefore determined the effect of vitronectin and other breast cancer cell lines do not express b3 integrin, 2 of the more components of the ECM on protein synthesis under conditions highly metastatic lines do, including MDA-MB-435 and to a 3- of normoxia or hypoxia (0.5% O2)inavb3-expressing cells.

A MDA-MB-435 cells B MDA-MB-231 cells 250 * 250 200 200 * 150 150

100 100 synthesis synthesis 50 50 Overall protein Overall Overall protein Overall 0 0 Ctrl NT LN VN Ctrl NT LN VN Ctrl NT LN VN Ctrl NT LN VN Normoxia Hypoxia Normoxia Hypoxia C MDA-MB-435 cells D MDA-MB-231 cells Normoxia Hypoxia Normoxia Hypoxia Ctrl NT LN VN Ctrl NT LN VN Ctrl NT LN VN Ctrl NT LN VN eIF4E eIF4E 4E-BP1 4E-BP1

4E-BP1-P S65 4E-BP1-P S65 eIF4G eIF4G eIF4A eIF4A β-Tubulin β-Tubulin

Figure 2. Vitronectin stimulates cap-dependent protein synthesis activity. MDA-MB-435 cells (A) and MDA-MB-231 (B) cells were cultivated on different ECM components, subjected to normoxia or hypoxia for 24 hours, protein synthesis activity measured by 35S-methionine incorporation of equal amounts of protein, and gels exposed identically. Data are presented normalized to control (ctrl) cells grown on plastic in normoxic conditions. C and D, abundance and activation status of proteins were determined by immunoblot of equal amounts of cytoplasmic protein lysates derived from each condition with protein and phosphospecific antibodies as shown. b-Tubulin was used as a loading control. Values are the mean SEM (n ¼ 3). , P< 0.05 by the Student t test (ctrl vs. VN). Ctrl, control (no ECM); NT, native collagen type IV; LN, laminin; VN, vitronectin.

www.aacrjournals.org Cancer Res; 73(14) July 15, 2013 OF3

Downloaded from cancerres.aacrjournals.org on September 30, 2021. © 2013 American Association for Cancer Research. Published OnlineFirst May 30, 2013; DOI: 10.1158/0008-5472.CAN-13-0218

Pola et al.

Total protein synthesis on a per cell basis was measured by metabolic labeling of cells with 35S-methionine, comparing A B cells cultured on plastic (ctrl) with native collagen IV (NT), MDA-MB-435 cells MDA-MB-435 cells m7GTP-chromatography Nx Hx laminin (LN), or vitronectin (VN) after 24 hours. The plating Nx Hx Ctrl VN Ctrl VN efficiency of MB-435 and MB-231 cells on the different ECM Ctrl VN Ctrl VN mTOR components was largely the same (Supplementary Fig. S1). eIF4E mTOR-P S2448 S6 4E-BP1 Only cells on vitronectin strongly stimulated protein synthesis, S6-P T240/244 which persisted even under hypoxia. In contrast, hypoxia eIF4G 4E-BP1 4E-BP1-P S65 reduced protein synthesis by 60% in control cells plated on β-Tubulin plastic and by 30% to 40% on cells plated on laminin or collagen C D MDA-MB-231 cells IV (Fig. 2A and B). MDA-MB-231 cells m7GTP-chromatography We next examined the steady-state levels of initiation fac- Nx Hx Nx Hx Ctrl VN Ctrl VN tors eIF4A, eIF4E, and eIF4G, as well as translation inhibitor 4E- Ctrl VN Ctrl VN mTOR BP1, and its activity by surrogate phosphorylation, in cells eIF4E mTOR-P S2448 4E-BP1 maintained on the different ECM components under normoxia S6 eIF4G or hypoxia (Fig. 2C and D). eIF4E, eIF4A, and eIF4G were all S6-P T240/244 unchanged in abundance under all conditions. 4E-BP1 levels 4E-BP1 E were also unchanged under all conditions, but its phosphor- 4E-BP1-P S65 MDA-MB-435 cells ylation was significantly increased under both normoxia and β-Tubulin Ctrl VN hypoxia in cells grown on vitronectin (Fig. 2C and D), consis- Nx Hx Nx Hx α tent with the increased levels of protein synthesis observed. We eIF2 -P S51 eIF2α therefore pursued the mechanism by which vitronectin upre- β-Tubulin gulates protein synthesis and maintains 4E-BP1 inactivation (hyperphosphorylation) despite hypoxia. Figure 3. Mechanism for vitronectin-mediated increased protein Interaction of cells with vitronectin stimulates mTOR synthesis. A and D, equal amounts of protein lysates from cells plated on plastic or vitronectin under normoxic (Nx) or hypoxic (Hx) conditions and formation of eIF4F complexes even under hypoxia were subjected to m7GTP-Sepharose chromatography, resolved by Equal amounts of eIF4E were recovered using 7-methyl SDS-12% PAGE, and identified by immunoblot for eIF4E, eIF4G, and 4E- GTP-resin from cells adhered to plastic or vitronectin under BP1. Results are representative of 3 independent experiments. B and C, normoxia or hypoxia. Hypoxic MB-435 and MB-231 cells levels and phosphorylation state of translation regulatory proteins. Equal fi attached to plastic showed a 4-fold increase in the inter- amounts of protein lysates were analyzed by immunoblot using speci c antibodies as shown. E, eIF2a and a-Ser51 phosphorylation were action of 4E-BP1 with eIF4E and a reciprocal decrease in analyzed by immunoblot as shown. association with eIF4G (Fig. 3A and D). Cells on vitronectin, however, showed a marked disassociation of eIF4E from 4E- BP1 in both normoxic and hypoxic conditions, which cor- factor eIF2 (18). While analysis of eIF2a showed only a related with the strongly increased translation activity modest increase in its inactivating phosphorylation at ser- shown earlier under these conditions. Cell contact with ine-51 by hypoxia, this was unchanged by adherence of cells vitronectin is therefore associated with increased overall to vitronectin (Fig. 2D). Vitronectin therefore acts primarily protein synthesis and increased formation of eIF4F com- to sustain protein synthesis during hypoxia by maintaining plexes and uncoupling of mTOR and translational repression or increasing mTOR activity. from hypoxia. mTOR is a major regulator of cap-dependent mRNA Integrin avb3 mediates vitronectin inactivation of 4E- translation and the main kinase for maintaining eIF4E BP1 in breast cancer cells availability through 4E-BP1 phosphorylation (inactivation). Vimentin is a well-recognized mediator of cell reprogram- We investigated the activity of mTOR by surrogate phos- ming required for breast cancer cell mobility, invasion, metas- phorylation, as well as that of 2 bona fide downstream tasis, and a variety of processes involved in breast cancer targets of mTOR, ribosomal protein S6 and 4E-BP1 (Fig. progression, including the epithelial-to-mesenychmal transi- 3B and C). Normoxic and hypoxic MB-435 or MB-231 cells on tion (EMT; ref. 19). Vimentin can serve as a direct substrate for vitronectin showed increased activating phosphorylation of regulation of breast cancer cell gene activity associated with mTOR and S6 and inactivating phosphorylation of 4E-BP1, invasion, probably through protein kinase modification, and as consistent with increased formation of eIF4F and protein a liganded partner of integrins (19). Integrins are heterodimeric synthesis shown earlier. Under hypoxia, as expected in the transmembrane proteins that mediate adhesion to ECM com- absence of vitronectin, mTOR phosphorylation was strongly ponents and activate intracellular signaling pathways (13). reduced in control cells, as was S6 and 4E-BP1 phosphor- Given the association of integrin avb3 with invasion and ylation, but was maintained by plating on vitronectin. metastasis of breast cancer cells, we asked whether vitronectin Adherence of cells to fibronectin was shown previously to upregulation of protein synthesis by activation of mTOR and its promote increased protein synthesis by downregulation of sustained activity during hypoxia are mediated by the integrin inactivating phosphorylation of the a-subunit of initiation pair avb3.

OF4 Cancer Res; 73(14) July 15, 2013 Cancer Research

Downloaded from cancerres.aacrjournals.org on September 30, 2021. © 2013 American Association for Cancer Research. Published OnlineFirst May 30, 2013; DOI: 10.1158/0008-5472.CAN-13-0218

Vitronectin Maintains Protein Synthesis in Hypoxia

A B 300 Normoxia Hypoxia IgG LM609 IgG LM609 * Ctrl VN Ctrl VN Ctrl VN Ctrl VN 200 4E-BP1 4E-BP1-P S65 β-Tubulin 100 C Normoxia Hypoxia 0 Ctrl VN Ctrl VN Ctrl VN Ctrl VN IgG LM609 IgG LM609 Overall protein synthesis rate Ctrl IgG LM609 Ctrl IgG LM609 Ctrl VN Ctrl VN Ctrl VN Ctrl VN (anti-αvβ3) (anti-αvβ3) AMPK Normoxia Hypoxia AMPK-P β-Tubulin

Figure 4. avb3 integrin increases protein synthesis and sustains it during hypoxia. A, MB-435 cells treated with control or LM609 (anti-avb3) antibodies were grown on plastic or vitronectin and subjected to normoxia or hypoxia for 24 hours. Translation rates were measured by 35S-methionine incorporation. Data are presented normalized to control cells in normoxic conditions, SEM (n ¼ 3). B, abundance and activation status of 4E-BP1 protein were determined by immunoblot equal amounts of protein lysates with protein and phosphospecific antibodies as shown. b-Tubulin was used as a loading control. C, immunoblot analysis of phosphorylation of AMPK in cells pretreated with or without LM609 antibodies in control and vitronectin-adhered cells. Immunoblots are representative of at least 3 independent experiments. , P < 0.05 (VN IgG vs. LM609).

MB-435 cells were pretreated with control matched IgG or tent with the results shown earlier (Supplementary Fig. S1). LM609 antibody to block integrin avb3, plated on plastic or The inactivating phosphorylation of 4E-BP1 stimulated by vitronectin, subjected to normoxia or hypoxia, specific activity vitronectin was largely blocked by silencing integrin b3, con- of protein synthesis determined, and normalized to total cell sistent with blocking avb3 integrin signaling using LM609 protein (Fig. 4A). Blocking with the anti-avb3 LM609 antibody antibody. Similarly, integrin b3–silenced cells on vitronectin, prevented the increase in protein synthesis typically observed when subjected to hypoxia, failed to upregulate phosphoryla- in tumor cells attached to vitronectin under normoxia and tion of mTOR target p70S6K (Fig. 5A). While silencing of sustained hypoxia. Treatment of cells with anti-avb3 antibody integrin b3 had no impact on protein synthesis activity in also resulted in strong inactivating phosphorylation of 4E-BP1 cells on control plates under normoxia or hypoxia, it strongly in cells on vitronectin under hypoxia or normoxia (Fig. 4B), blocked the stimulation of protein synthesis by vitronectin consistent with the requirement for avb3 vitronectin engage- during normoxia and its maintenance during hypoxia (Fig. 5B ment to stimulate protein synthesis and sustain it under and C). hypoxia. Levels of activated (phosphorylated) AMP kinase (AMPK), which can be stimulated by hypoxia and inhibit Integrin av b3 acts through ILK, PI3K, and AKT, but not mTOR (20), were unchanged under hypoxia, regardless of FAK, to stimulate mTOR activity and protein synthesis adherence of cells to vitronectin or the presence of avb3 Integrin avb3 interaction with the ECM has been shown blocking antibody (Fig 4C). Thus, the ability of vitronectin to to propagate intercellular signals through the activation override hypoxia inhibition of cap-dependent mRNA transla- of FAK and/or ILK, both of which are overexpressed in a tion involves avb3 engagement of vitronectin but does not act number of human cancers and which stimulate, directly through the suppression of AMPK. or indirectly, the activation of phosphoinositide 3-kinase We therefore more directly examined how integrin avb3 (PI3K) and AKT (21–23). Integrin avb1, avb2, and avb3 inactivates 4E-BP1 by engagement to vitronectin, by silencing interaction with vitronectin leads to increased ILK inter- expression of the b3 integrin in MB-435 breast cancer cells action, which binds directly to the cytoplasmic domain of using stable integrating lentivirus vectors. Integrin subunit b3 integrins and can increase FAK clustering and activation at was chosen for silencing as it only heterodimerizes with focal adhesions (24). Moreover, there is evidence that integ- integrin subunits avoraIIb, and aIIb is only expressed in rin avb3 itself may stimulate the increased transcription of and not in breast cancer cells. The integrin subunit av ILK, in a feed-forward type mechanism (22), and hypoxia forms heterodimers with several different b integrins and can mobilize the recruitment of avb3 to the plasma mem- therefore was not silenced. Cells were seeded onto control or brane to activate FAK (23). We therefore investigated vitronectin plates, subjected to hypoxia for 24 hours, and whether FAK, ILK, or both are required to maintain mTOR expression levels of integrin b3 determined by immunoblot activity and protein synthesis during hypoxia in cells adher- analysis, achieving 80% reduction (Fig. 5A). The plating ed to vitronectin. efficiency of cells on vitronectin under hypoxia was unchanged Nonsilencing or integrin b3 shRNAs were expressed in despite integrin b3 silencing (Supplementary Fig. S2), consis- control and vitronectin-plated MB-435 cells maintained under

www.aacrjournals.org Cancer Res; 73(14) July 15, 2013 OF5

Downloaded from cancerres.aacrjournals.org on September 30, 2021. © 2013 American Association for Cancer Research. Published OnlineFirst May 30, 2013; DOI: 10.1158/0008-5472.CAN-13-0218

Pola et al.

Vitronectin promotes integrin avb3–mediated breast ABHypoxia cancer cell invasion in an eIF4E-dependent manner β a b NS ITG 3 shRNA Vitronectin-plated It was previously shown that integrin v 3isakeyfactor Ctrl VN Ctrl VN Nx Hx in promoting invasion and metastasis in breast and other β ITG 3 β β cancers, particularly to bone (15, 23, 25). Moreover, hypoxia eIF4A NS 3 NS 3 shRNA and vitronectin both have been shown to promote integrin 4E-BP1-P S65 ITGβ3 avb3–mediated cell migration and expression of pro-migra- 4E-BP1 eIF4A p70S6K-P T389 tory molecules (26, 27). We therefore studied the effect of p70S6K integrin avb3 and hypoxia on vitronectin-mediated invasion of MB-435 cells. Cells were treated with control IgG or C Plastic-plated Vitronectin-plated LM609 anti-avb3 antibodies and cell membrane invasion 200 * activities measured. MB-435 cells displayed an approximate- ly 6-fold increase in vitronectin-directed invasion under normoxia and hypoxia that was blocked by anti-avb3com- 100 pared with normoxic control cells (Fig. 7A). Vitronectin- mediated invasion studies were also conducted under hyp- 0 oxia using cells that were silenced for eIF4E to test the

Protein synthesis rate β β β β NS 3 NS 3 NS 3 NS 3 shRNA importance of eIF4E (cap)-dependent translation in promot- Normoxia Hypoxia Normoxia Hypoxia ing avb3–vitronectin–mediated invasion (Fig. 7B). Com- pared with eIF4E-expressing cells on vitronectin, silencing Figure 5. Integrin b3 silencing abolishes increased mTOR activity and of eIF4E almost abolished invasion activity stimulated by protein synthesis in MB-435 cells adhered to vitronectin (VN). A, shRNA vitronectin. These data therefore show that vitronectin– silencing of integrin b3 and activation status of 4E-BP1 and activating avb3 stimulation of mTOR activity and eIF4E (cap)-depen- S6K phosphorylation of mTOR protein target p70 were determined by dent protein synthesis are important in facilitating high immunoblot of equal amounts of protein lysates as shown. Results shown are representative of 3 independent experiments. eIF4A was used rates of breast cancer cell invasion under hypoxia. as a loading control. B, level of integrin b3 in cells under normoxia (Nx) or hypoxia (Hx) plated on vitronectin, used for studies shown in C. C, Discussion protein synthesis activity in integrin b3–silenced or control cells, plated on Vitronectin constitutes a major component of the vascular- 35 plastic or vitronectin, measured by S-methionine incorporation, and ized breast cancer microenvironment; however, it is typically normalized to total cellular protein amount. Data are presented as percentage of control (ctrl) in normoxic conditions, SEM (n ¼ 3). only exposed to breast cancer cells in abundance with disrup- , P < 0.05 by the Student t test. tion of the basement membrane during the process of cancer

hypoxia, and the interaction of ILK with integrin b3 was determined by immunoprecipitation and immunoblot analysis (Fig. 6A). There was significant binding of ILK to b3 only in the ABHypoxia Hypoxia ITG β3 Immunop. ILK Immunop. vitronectin-plated control cells, which was strongly reduced by β b b NS ITG β3 shRNA NS ITG 3 shRNA 3 silencing. Reciprocally, silencing of 3 also reduced ILK Ctrl VN Ctrl VN Ctrl VN Ctrl VN levels, possibly by eliminating its transcriptional induction ITG β3 ILK Blot (22), and of course eliminated coprecipitating b3 protein Blot ILK ITG β3 (Fig. 6B). Cell lysates were then examined for surrogate acti- vating phosphorylation of FAK (Y397), AKT (S473), and inac- C Hypoxia D Normoxia NS ILK shRNA tivation of 4E-BP1 (S65) as a measure of mTOR activity in Mock PI-103 Ctrl VN Ctrl VN Ctrl VN Ctrl VN hypoxic cells, with or without silencing of ILK (Fig. 6C). There FAK-P Y397 AKT was no evidence for alteration of FAK protein levels or phos- FAK AKT-P S473 phorylation with plating of cells on plastic or vitronectin under AKT-P S473 p70S6K hypoxia, with or without ILK silencing. In contrast, cells AKT p70S6K-P T389 plated on vitronectin under hypoxia showed a strong increase 4E-BP1-P S65 in both AKT and 4E-BP1 phosphorylation, as well as increased ILK hyperphosphorylated 4E-BP1, the latter a measure of mTOR activity, which was lost with silencing of ILK. Because ILK a b signaling can be dependent on PI3K and AKT in some cell Figure 6. v 3 integrin engagement to vitronectin stimulates AKT and mTOR through ILK. MB-435 cells were silenced with shRNA to integrin b3 systems and not others, MB-435 cells were plated on plastic (A) or to ILK (B), or a nonsilencing control (NS); equal amounts of protein or vitronectin and treated with the pan-PI3K and mTOR lysates were subjected to immunoprecipitation with b3 integrin antibody inhibitor PI-103 to measure the effect on AKT activity (Fig. or IL antibody and immunoblotted for b3 integrin and ILK. C, equal 6D). Vitronectin stimulation of AKT phosphorylation was amounts of protein lysates from integrin b3–silenced MB-435 cells were subjected to immunoblot with antibodies as shown. D, MB-435 cells were inhibited by PI-103. These data therefore indicate that vitro- treated with vehicle or 1 mmol/L of pan-PI3K inhibitor PI-103 for 6 hours. nectin stimulates integrin avb3 signaling through ILK and Equal amounts of protein lysates were subjected to immunoblot as PI3K to AKT to maintain mTOR activity during hypoxia. shown. Results are representative of 3 studies.

OF6 Cancer Res; 73(14) July 15, 2013 Cancer Research

Downloaded from cancerres.aacrjournals.org on September 30, 2021. © 2013 American Association for Cancer Research. Published OnlineFirst May 30, 2013; DOI: 10.1158/0008-5472.CAN-13-0218

Vitronectin Maintains Protein Synthesis in Hypoxia

adhered breast cancer cells is a result of upregulation of A cap-dependent protein synthesis in this setting, a function 10 stimulated by high levels of eIF4E and low levels of 4E-BP1 * 8 * activity. This is consistent with the recent findings of Alain and 6 colleagues, in which the increased ratio of eIF4E to 4E-BP1 was found to be tumor-promoting (29). This is also consistent with 4 the strong correlation between elevated expression of eIF4E, 2 cancer progression, and decreased survival in breast, colon, Fold increase in cell invasion and other cancers (30). 0 Ctrl VN Ctrl VN Ctrl VN Ctrl VN Several previous reports described adhesion-dependent reg- Ctrl IgG LM609 Ctrl IgG LM609 ulators of protein synthesis, but a mechanistic understanding α β α β (anti- v 3) (anti- v 3) and the impact of hypoxia has not been examined in depth Normoxia Hypoxia (10, 31). The strongest ability to stimulate mTOR activity and eIF4E-dependent mRNA translation and to maintain these B 10 Hypoxia activities during hypoxia was found for vitronectin, acting through integrin avb3 (Figs. 4, 5). In contrast to integrin avb3 8 * action in glioma, where FAK activation was found as a primary 6 signaling event to AKT/mTOR (32), in breast cancer cells, 4 ILK signaling and its interaction with integrin b3 were 2 found to predominate (Fig. 6). Collectively, these studies show Fold increase in cell invasion the importance of specific microenvironment/ECM interac- 0 Ctrl VN Ctrl VN tions in determining the multiple pathways for stimulation of mTOR and protein synthesis and their importance in main- taining eIF4E-dependent mRNA translation in hypoxic cancer NS eIF4E siRNA cells. Vitronectin is predominantly a plasma protein that abun- Figure 7. Vitronectin stimulates breast cancer cell migration via integrin dantly infiltrates the tumor stroma by leakage from the avb3 in hypoxic conditions. A, invasion of MDA-MB-435 human breast vasculature (16). It accumulates in local areas around tumor cancer cells was tested through membrane toward vitronectin (VN) using cells, particularly in the subendothelial regions of capillaries Transwell chambers for 12 hours at 37 C during normoxic or hypoxic that are a gateway to intravasating cells involved in metas- conditions. Cells were pretreated with control antibody or anti-avb3Ab (LM609). Invaded cells were fixed, stained with crystal violet, and tasis (16). Our data suggest that by increasing or maintaining quantified. Data are presented as percentage of control cells on plastic protein synthesis, tumor cells in contact with vitronectin under normoxic conditions SEM (n ¼ 3). B, MB-435 cells were under hypoxia are stimulated in invasion activity. Only transfected twice with an siRNA to eIF4E or NS control, then 24 hours vitronectin, and not laminin or native collagen IV, strongly later subjected to invasion assay as described above, under hypoxia, normalized to NS control cells. Data are SEM of n ¼ 3 studies. enhanced protein synthesis concomitant with hyperpho- Immunoblot of equal numbers of cells transfected with NS or eIF4E siRNA sphorylation (inactivation) of 4E-BP1 and maintained it before invasion assay. , P < 0.05 (VN control vs. VN test sample). under hypoxia (Fig. 2). Interestingly, integrin b3 silencing revealed 2 previously unknown aspects of integrin function that impact on translational control. First, b3 silencing cell invasion (3). Integrin avb3 is a key factor that drives tumor studies showed that integrin b3–vitronectin engagement cell invasion and is associated with increased potential for promotes inactivating phosphorylation of 4E-BP1 and acti- metastasis in a variety of cancers, particularly to bone in breast vation of mTOR through stimulation of ILK, which persists cancer (28). Here, we extend the understanding of the function in hypoxic tumor cells. Second, with unliganded integrin b3 of vitronectin in human cancer, showing that it reprograms the (or reduced b3 expression) during hypoxia, vitronectin actu- breast cancer cell to upregulate mTOR and cap-dependent ally partially downregulates cap-dependent protein synthe- mRNA translation through integrin avb3 engagement. Vitro- sis by increasing the activation (dephosphorylation) of 4E- nectin–avb3 interaction sustains both mTOR activation and BP1 (Figs. 4, 5). Unoccupied integrins are known to reduce protein synthesis, overriding the downregulation that typically cell survival through a process known as integrin-mediated occurs under hypoxia, thereby promoting cancer cell invasion cell death (IMD) that is thought to be an alternative method under hypoxic conditions. The mechanism of action involves of inducing apoptosis through the recruitment of caspase-8 integrin avb3–ILK interaction and ILK stimulation of mTOR to the plasma membrane, independent of death receptors or activity, thereby inactivating 4E-BP1 and maintaining a high FADD (33). Although our studies were not of duration to level of eIF4E-dependent mRNA translation despite hypoxia impact on cell survival, they do suggest that vitronectin (Fig. 6). In so doing, integrin avb3 is a mediator of invasion engagement avb3–bearing cancer cells will have increased stimulatory signaling promoted by vitronectin. These data protein synthesis, survival, and invasive properties in a underscore the importance of preservation of canonical pro- hypoxic environment. In summary, our results support the tein synthesis in cancer cell invasion in the hypoxic microen- hypothesis that tumor cells can rapidly modulate their vironment. The increased invasion activity of vitronectin- translation programs in response to particular components

www.aacrjournals.org Cancer Res; 73(14) July 15, 2013 OF7

Downloaded from cancerres.aacrjournals.org on September 30, 2021. © 2013 American Association for Cancer Research. Published OnlineFirst May 30, 2013; DOI: 10.1158/0008-5472.CAN-13-0218

Pola et al.

of the ECM microenvironment that impacts on their invasive Administrative, technical, or material support (i.e., reporting or orga- nizing data, constructing databases): R.J. Schneider properties. Study supervision: R.J. Schneider

Disclosure of Potential Conflicts of Interest No potential conflicts of interest were disclosed. Grant Support This work was supported by grants from the Breast Cancer Research Foundation, the Department of Defense Breast Cancer Research Program, and Authors' Contributions the NIH to S.C. Formenti and R.J. Schneider. Conception and design: C. Pola, S.C. Formenti, R.J. Schneider The costs of publication of this article were defrayed in part by the Development of methodology: C. Pola payment of page charges. This article must therefore be hereby marked Acquisition of data (provided animals, acquired and managed patients, advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this provided facilities, etc.): C. Pola fact. Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): C. Pola, S.C. Formenti, R.J. Schneider Writing, review, and/or revision of the manuscript: C. Pola, S.C. Formenti, Received January 28, 2013; revised April 2, 2013; accepted May 7, 2013; R.J. Schneider published OnlineFirst May 30, 2013.

References 1. Liotta LA, Kohn EC. The microenvironment of the tumour-host inter- 18. Gandin V, Brina D, Marchisio PC, Biffo S. JNK inhibition arrests face. Nature 2001;411:375–79. cotranslational degradation. Bioch Biophys Acta 2010;1803:826–31. 2. Bissell MJ, Radisky D. Putting tumours in context. Nat Rev Cancer 19. Satelli A, Li S. Vimentin in cancer and its potential as a molecular target 2001;1:46–54. for cancer therapy. Cellu Mol Life Sci 2011;68:3033–46. 3. Nelson CM, Bissell MJ. Of extracellular matrix, scaffolds, and signal- 20. Hay N, Sonenberg N. Upstream and downstream of mTOR. Dev ing: tissue architecture regulates development, homeostasis, and 2004;18:1926–45. cancer. Annu Rev Cell Dev Biol 2006;22:287–309. 21. Wickstrom SA, Lange A, Montanez E, Fassler R. The ILK/PINCH/parvin 4. Mueller MM, Fusenig NE. Friends or foes - bipolar effects of the tumour complex: the kinase is dead, long live the pseudokinase! EMBO J stroma in cancer. Nat Rev Cancer 2004;4:839–49. 2010;29:281–91. 5. Bristow RG, Hill RP. Hypoxia and metabolism. Hypoxia, DNA repair 22. Lossner D, Abou-Ajram C, Benge A, Aumercier M, Schmitt M, Reuning and genetic instability. Nat Rev Cancer 2008;8:180–92. U. Integrin alphavbeta3 upregulates integrin-linked kinase expression 6. Connolly E, Braunstein S, Formenti S, Schneider RJ. Hypoxia inhibits in human ovarian cancer cells via enhancement of ILK gene transcrip- protein synthesis through a 4E-BP1 and elongation factor 2 kinase tion. J Cell Physiol 2009;220:367–75. pathway controlled by mTOR and uncoupled in breast cancer cells. 23. Skuli N, Monferran S, Delmas C, Favre G, Bonnet J, Toulas C, et al. Mol Cell Biol 2006;26:3955–65. Alphavbeta3/alphavbeta5 integrins-FAK-RhoB: a novel pathway for 7. Gingras AC, Raught B, Sonenberg N. eIF4 initiation factors: effectors of hypoxia regulation in glioblastoma. Cancer Res 2009;69:3308–16. mRNA recruitment to ribosomes and regulators of translation. Annu 24. Hood JD, Cheresh DA. Role of integrins in cell invasion and migration. Rev Biochem 1999;68:913–63. Nat Rev Cancer 2002;2:91–100. 8. Silvera D, Formenti SC, Schneider RJ. Translational control in cancer. 25. Pecheur I, Peyruchaud O, Serre CM, Guglielmi J, Voland C, Bourre F, Nat Rev Cancer 2010;10:254–66. et al. Integrin alpha(v)beta3 expression confers on tumor cells a greater 9. Holcik M, Sonenberg N. Translational control in stress and apoptosis. propensity to metastasize to bone. FASEB J 2002;16:1266–68. Nat Rev Mol Cell Biol 2005;6:318–27. 26. Cowden Dahl KD, Robertson SE, Weaver VM, Simon MC. Hypoxia- 10. Gorrini C, Loreni F, Gandin V, Sala LA, Sonenberg N, Marchisio PC, inducible factor regulates alphavbeta3 integrin cell surface expression. et al. Fibronectin controls cap-dependent translation through beta1 Mol Biol Cell 2005;16:1901–12. integrin and eukaryotic initiation factors 4 and 2 coordinated pathways. 27. Rolli M, Fransvea E, Pilch J, Saven A, Felding-Habermann B. Activated Proc Natl Acad Sci USA 2005;102:9200–05. integrin alphavbeta3 cooperates with metalloproteinase MMP-9 in 11. Han S, Khuri FR, Roman J. Fibronectin stimulates non-small cell lung regulating migration of metastatic breast cancer cells. Proc Natl Acad carcinoma cell growth through activation of Akt/mammalian target of Sci U S A 2003;100:9482–87. rapamycin/S6 kinase and inactivation of LKB1/AMP-activated protein 28. Felding-Habermann B, O'Toole TE, Smith JW, Fransvea E, Ruggeri kinase signal pathways. Cancer Res 2006;66:315–23. ZM, Ginsberg MH, et al. Integrin activation controls metastasis in 12. Maeshima Y, Sudhakar A, Lively JC, Ueki K, Kharbanda S, Kahn CR, human breast cancer. Proc Natl Acad Sci USA 2001;98:1853–58. et al. Tumstatin, an endothelial cell-specific inhibitor of protein syn- 29. Alain T, Morita M, Fonseca BD, Yanagiya A, Siddiqui N, Bhat M, et al. thesis. Science 2002;295:140–43. eIF4E/4E-BP ratio predicts the efficacy of mTOR targeted therapies. 13. Cabodi S, del Pilar Camacho-Leal M, Di Stefano P, Defilippi P. Integrin Cancer Res 2012;72:6468–76. signalling adaptors: not only figurants in the cancer story. Nat Rev 30. Armengol G, Rojo F, Castellvi J, Iglesias C, Cuatrecasas M, Pons B, Cancer 2010;10:858–70. et al. 4E-binding protein 1: a key molecular "funnel factor" in human 14. Takayama S, Ishii S, Ikeda T, Masamura S, Doi M, Kitajima M. The cancer with clinical implications. Cancer Res 2007;67:7551–55. relationship between bone metastasis from human breast cancer and 31. Chung J, Bachelder RE, Lipscomb EA, Shaw LM, Mercurio AM. integrin alpha(v)beta3 expression. Anticancer Res 2005;25:79–83. Integrin (alpha 6 beta 4) regulation of eIF-4E activity and VEGF trans- 15. Zhao Y, Bachelier R, Treilleux I, Pujuguet P, Peyruchaud O, Baron R, lation: a survival mechanism for carcinoma cells. J Cell Biol 2002;158: et al. Tumor alphavbeta3 integrin is a therapeutic target for breast 165–74. cancer bone metastases. Cancer Res 2007;67:5821–30. 32. Lorger M, Krueger JS, O'Neal M, Staflin K, Felding-Habermann B. 16. Aaboe M, Offersen BV, Christensen A, Andreasen PA. Vitronectin in Activation of tumor cell integrin alphavbeta3 controls angiogenesis human breast carcinomas. Biochim Biophys Acta 2003;1638:72–82. and metastatic growth in the brain. Proc Natl Acad Sci USA 2009;106: 17. Meyer T, Marshall JF, Hart IR. Expression of alphav integrins and 10666–71. vitronectin receptor identity in breast cancer cells. Br J Cancer 1998; 33. Stupack DG, Cheresh DA. Integrins and angiogenesis. Curr Top Dev 77:530–36. Biol 2004;64:207–38.

OF8 Cancer Res; 73(14) July 15, 2013 Cancer Research

Downloaded from cancerres.aacrjournals.org on September 30, 2021. © 2013 American Association for Cancer Research. Published OnlineFirst May 30, 2013; DOI: 10.1158/0008-5472.CAN-13-0218

Vitronectin α− vβ3 Integrin Engagement Directs Hypoxia-Resistant mTOR Activity and Sustained Protein Synthesis Linked to Invasion by Breast Cancer Cells

Carolina Pola, Silvia C. Formenti and Robert J. Schneider

Cancer Res Published OnlineFirst May 30, 2013.

Updated version Access the most recent version of this article at: doi:10.1158/0008-5472.CAN-13-0218

Supplementary Access the most recent supplemental material at: Material http://cancerres.aacrjournals.org/content/suppl/2013/05/30/0008-5472.CAN-13-0218.DC1

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/early/2013/07/05/0008-5472.CAN-13-0218. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.

Downloaded from cancerres.aacrjournals.org on September 30, 2021. © 2013 American Association for Cancer Research.