Introduction
Metastasis is a multi-step process in which cancer cells disseminate from the primary site to distant tissues or organs [
1]. Breast tumors are commonly epithelial in origin, and their ability to invade is enhanced by modulators that stimulate epithelial-mesenchymal transition (EMT), such as transforming growth factor-β (TGF-β) and transcriptional repressors Snail, Slug, and Twist that are induced by TGF-β [
2-
4]. During the metastasis cascade, (epi)genetic changes in cancer cells and signals from the microenvironment promote EMT of the tumor cells
in situ, which facilitates local invasion and intravasation into nearby tissues and circulation. Subsequently, circulating tumor cells with a mesenchymal morphology may extravasate out of the blood stream and invade secondary sites, which involves cell-matrix interactions [
5]. Breast carcinoma cells are able to infiltrate into specific tissues, including bone, lung and brain. Within the new microenvironment the tumor cells start to proliferate, and develop into a macrometastatic lesion [
6].
Integrins are cell-surface adhesion receptors consisting of α and β transmembrane protein subunits, which directly interact with extracellular matrix (ECM) components when regulating cell migration, proliferation, and cell survival via outside-inside and/or inside-outside signaling mechanisms [
7]. In cancer, integrins contribute to tumor growth, invasion and metastasis [
8]. One of the α integrins, αv, dimerizes with the β integrin subunits β1, β3, β5, β6 and β8, and has been implicated in the pathophysiology of malignant tumors [
9]. Integrins αvβ3, αvβ5 and αvβ6 have been reported to be crucial for tumor cell adhesion, migration, survival, maintenance of stem cell phenotype and angiogenesis and for crosstalk with growth factors in the activation of oncogenes and inhibition of tumor suppressors [
10-
13].
αv integrin can be involved in activation of latent TGF-β by binding latency-associated peptide (LAP) [
14], can interact with the TGF-β (type II) receptor and thereby promote TGF-β-induced responses in lung fibroblasts and mammary epithelial cells [
15,
16], and can interact with the TGF-β type III receptor endoglin and stimulate TGF-β/Smad signaling in endothelial cells [
17]. Vice versa the TGF-β receptor can also mediate phosphorylation of certain β-chains of integrins and modulate their function in hepatocellular carcinoma [
18]. Moreover, TGF-β can regulate αv integrin expression in breast epithelial cells and αv integrin can modulate TGF-β receptor expression in dermal fibroblasts [
19,
20]. Thus, αv integrin and TGF-β signaling show extensive interplay and αv integrin may be an effector and mediator of TGF-β signaling responses [
21,
22].
Human metastatic breast cancer cells residing in bone showed high αvβ3 integrin expression. The MDA-MB-231 subclone B02, established from bone metastases, was found to constitutively overexpress αvβ3 integrin compared to the parental MDA-MB-231 cells [
23]. Although αv integrin seems to be an important pharmacological target to inhibit breast cancer metastasis, the mechanism by which it regulates metastatic breast cancer progression is largely unknown. In this study, selective knockdown of αv integrin expression or pharmacological inhibition of αv integrin function was found to potently mitigate the invasion and metastasis of breast cancer cells in zebrafish and mouse xenograft models. In line with previous studies in other cell types, mechanistic
in vitro studies revealed an integrated interplay between αv integrin and TGF-β, a strong driver of invasion and metastasis of breast cancer. Moreover, maximum efficacy of bone metastasis inhibition in mice was accomplished when therapeutic targeting of αv integrin was combined with standard-of-care metastatic breast cancer treatments.
Methods
Cell culture and reagents
Human MDA-MB-231-luc cells [
24] were obtained from Dr Clemens Löwik (Department of Radiology, Leiden University Medical Center, Leiden, The Netherlands) and Dr Gabri van der Pluijm (Department of Urology, Leiden University Medical Center, Leiden, The Netherlands). The MDA-231/B02-luc line was previously published [
23] and used for mouse xenograft experiments. These MDA-MB-231 cell lines were maintained at 37°C in DMEM high glucose containing L-glutamine, 10% FBS and 100 U/ml Pen/Strep (Gibco, Invitrogen, Blijswijk, Netherlands). MCF10A-M4 cells were kindly provided by Dr Fred Miller (Barbara Ann Karmanos Cancer Institute, Detroit, USA) and maintained at 37°C in DMEM/F12 (Gibco, Invitrogen, Blijswijk, Netherlands) containing 5% horse serum (Gibco, Invitrogen, Blijswijk, Netherlands), 0.1 μg/ml cholera toxin (Merck Millipore, Amsterdam, Netherlands), 0.02 μg/ml epidermal growth factor (EGF), 0.5 μg/ml hydrocortisone (Sigma, Zwijndrecht, Netherlands), 10 μg/ml insulin (Sigma, Zwijndrecht, Netherlands), 50 μg/ml streptomycin, and 100 U/ml Pen/Strep (Gibco, Invitrogen, Blijswijk, Netherlands).
Zebrafish embryo production and tumor cell injection
The transgenic zebrafish line Tg(fli1:GFP) was raised, staged and maintained according to standard procedures in compliance with the local Institutional Committee for Animal Welfare of the Leiden University Medical Center (LUMC). Tumor cell injection into zebrafish embryos was conducted as we previously described [
24]. Briefly, approximately 400 fluorescent-labeled mammalian cells were injected into the duct of Cuvier (DoC). After implantation, zebrafish embryos (including non-implanted controls) were maintained at 33°C [
25]. For each cell line or condition, data are representative of at least two independent experiments with at least 50 embryos per group.
In vivo toxicity test of chemical compounds
GLP0187 (Galapagos NV, Mechelen, Belgium) was added to the zebrafish egg water 2 days post fertilization (dpf) for toxicity tests, or 2 days post implantation (dpi) for treatment, and refreshed every second day. For toxicity tests, embryo survival or malformation was scored daily. For GLPG0187 treatment, after 5 days embryos were fixed overnight in 4% buffered paraformaldehyde at 4°C. Embryos were placed in a glass-bottom 96-well plate (Greiner Bio One GmbH, Frickenhausen, Germany), and imaged as described.
Lentiviral transduction
Lentivirus was produced by co-transfecting pLKO-1 (shRNA-knockdown) plasmids and helper plasmids pCMV-VSVG, pMDLg-RRE (gag/pol), and pRSV-REV into HEK293T cells. Cell supernatants were harvested 48 h after transfection and used to infect cells or stored at −80°C. pLKO-1 plasmids with specific shRNAs were obtained from Sigma, Zwijndrecht, Netherlands (MISSION® shRNA). We used TRCN-0000003240 and TRCN-0000010769 for αv integrin knockdown. For stable cell lines, cells were infected at 20% confluence for 24 h with lentiviral supernatants diluted 1:1 with normal culture medium in the presence of 8 μg/ml polybrene (Sigma, Zwijndrecht, Netherlands): 24 h after infection, cells were placed under puromycin (1 μg/ml) selection for 3 days or more.
Fluorescence-activated cell sorting (FACS) analysis
Cells were washed once with PBS and harvested with 0.05% trypsin/0.025% EDTA. Detached cells were washed with EDTA, re-suspended in EDTA (4 × 104/ml), and Fc-blocked for 15 minutes at room temperature. Anti-human αv integrin conjugated to Phycoerythrin (PE) (R&D systems, Abingdon, UK) was added and cells were incubated for 45 minutes at 4°C. PE-conjugated mouse IgG anti body (R&D systems, Abingdon, UK) was used as a negative control. Cells were washed three times in PBS and PE fluorescence was measured on a BD LSR II Cytometer. Results were analyzed with BD FACS Diva 6.1 software.
Apoptosis assay
The apoptotic effects of αv integrin were examined using propidium iodide (PI) flow cytometry as previously described [
26]. Detached and adherent cells were collected and labeled for 15 minutes at room temperature with PI (40 μg/ml) and immediately analyzed on a BD LSRII flow cytometer (BD Biosciences, Breda, Netherlands) using BD FACS Diva6.1 software.
RNA isolation and real-time quantitative PCR (RT-PCR)
RNA was extracted with a NucleoSpin RNA II kit (BIOKE, Leiden, Netherlands) according to the supplier’s manual. For RT-PCR a RevertAid First Strand cDNA Synthesis Kit (Thermo Scientific, Leusden, Netherlands) was used. RT-PCR was performed on a CFX connect real-time PCR system (Bio-Rad, Veenendaal, Netherlands) and analyzed with CFX Manager software version 2.0 (Bio-Rad, Veenendaal, Netherlands). The sequences of the primers are given in the supplemental information. All samples were analyzed in triplicate and normalized to GAPDH.
The sequences of the primers were as follows: Integrin αv forward: 5′-CTTCTTGGTGGTCCTGGTAGG-3′; Integrin αv reverse: 5′-TTTCTGCCACTTGATCCGAAA-3′; GAPDH forward: 5′-AGCCACATCGCTCAGACA C-3′; GAPDH reverse: 5′-GCCCAATACGACCAAATC C-3′; N-cadherin forward: 5′-CAGACCGACCCAAACAGCAAC-3′; N-cadherin reverse: 5′-GCAGCAACAGTAAGGACAAACATC-3′;Snail forward: 5′-ACCACTATGCCGCGCTCTT-3′; Snail reverse: 5′-GGTCGTAGGGCTGCTGGAA-3′; Slug forward: 5′-ATGAGGAATCTGGCTGCTGT-3′; Slug reverse: 5′-GAGGAGAAAATGCCTTTGGA-3′; Vimentin forward: 5′-CCAAACTTTTCCTCCCTGAACC-3′; Vimentin reverse: 5′-CGTGATGCTGAGAAGTTTCGTTGA-3′; CTGF forward: 5′-TGCGAAGCTGACCTGGAAGAGAA-3′; CTGF reverse: 5′-AGCTCGGTATGTCTTCATGCTGGT-3′; IL-11 forward: 5′-ACTGCTGCTGCTGAAGACTC-3′; IL-11 reverse: 5′-CCACCCCTGCTCCTGAAATA-3′; PAI-1 forward: 5′- GCAGGACATCCGGGAGAGA-3′; PAI-1 reverse: 5′-CCTGAGAACCTCCCTTGACCTT-3′.
Western blot analysis
Western blotting was carried out as previously described [
27]. The primary antibodies used were anti-N-Cadherin (BD Biosciences, Breda, Netherlands #610920), anti-α-Smooth Muscle Actin (Sigma, Zwijndrecht, Netherlands #A2547), anti-Snail (Cell Signaling, Leiden, Netherlands #3879), anti-Smad2 (BD Biosciences, Breda, Netherlands #610842), anti-p-Smad2 (Cell Signaling, Leiden, Netherlands #3108), anti-Smad4 (Santa Cruz #sc7966), anti-TGFβRI (Santa Cruz, Heidelberg, Germany #sc 398), anti-TGFβRII (Santa Cruz, Heidelberg, Germany #sc-400), anti-Smad3 (Epitomics, Duiven, Netherlands #1735–1), anti-p-Smad3 (a kind gift from Dr Edward B Leof, Mayo Clinic, Rochester, Minnesota) and anti-β-actin (Sigma, Zwijndrecht, Netherlands #A5441). All the secondary antibodies were from Sigma, Zwijndrecht, Netherlands. Western quantification was performed using image J software.
Transwell migration assay
Migration assays were performed in 24-well polyethylene terephthalate inserts (Corning Life Sciences, Amsterdam, Netherlands, 8.0- μm pore size): 1 × 104 MDA-MB-231 cells were cultured in DMEM with 0.5% FBS and seeded in the upper compartment (replicas for each sample). Then the cells were treated with or without TGF-β3 (5 ng/ml) for 16 h, which allows cells to migrate to the lower side of the insert filter. Cells in the upper side of the filter membrane were removed with a cotton swab. Cells on the lower side of the filter were fixed in 4% paraformaldehyde, stained with crystal violet 0.5% and then counted and photographed by randomly choosing different views under the microscope.
Animals and surgical procedures
Four- to five-week-old female Balb-c nu/nu mice (Charles River, les Oncins, France) were anesthetized with isofluorane and 5 × 10
5 freshly harvested MDA-MB-231/B02 luc cells in 100 μl PBS were inoculated into the tail vein [
28]. All mouse experiments were performed according to ethical guidelines edited by the Animal Institutional Care and Use Committee of Galapagos controlled by French Authorities (agreement number B 93 063 06, DDPP, Seine Saint Denis, France). The
Comité National de Rélexion Ethique sur l’Expérimentaion Animale approved all the mice experiments for this study on 5 July 2012. All research using zebrafish, including housing and experiments, was carried out according to the international guidelines and approved by the local Institutional Committee for Animal Welfare (
Dier Ethische Commissie (DEC) of the LUMC.
Proliferation assay in vitro
MDA-MB-231 or MCF10A-M4 cells were seeded at 2,000 cells per well in a 24-well plate. GLPG 0187 treatment was started 12 h after seeding. The numbers of cells were counted each 36 or 24 h until 108 or 96 h after seeding. Each experiment was performed in triplicates and numbers were calculated with a T20 cell counter (Brio-Rad, Veenendaal, Netherlands).
Proliferation assay in zebrafish
Approximately 70 mCherry-labeled MDA-MB-231 cells were injected in to the yolk sac of 2-dpf fli1:GFP Casper zebrafish embryos. Embryos were sorted 1 day post injection (dpi) by confocal microscopy to assess the fluorescent mass at the yolk sac and the images were scanned by z-stacks. Injected embryos were kept in 96-well plates at 33°C and scanned at 6 dpi. The relative volume of tumor cells was calculated by Stacks software.
Immunoflurescence staining in zebrafish
Xenografted zebrafish embryos were fixed with 4% paraformaldehyde. Samples were first dehydrated with methanol followed by a rehydration step and treated with 10 ug/ml proteinase K for 10 minutes at 37°C. Then cells were blocked and permeabilized with 1% BSA and 0.5% Triton X-100 in PBS before incubation with primary antibody Ki 67 (Merck Millipore, Amsterdam, Netherlands #AB9260; 1:200 in blocking buffer) in 4°C for 12 h. Samples then were washed with 5% Triton X-100-PBS, incubated with donkey anti rabbit IgG Alexa Fluor 647 (Invitrogene #A31572 1:200 in blocking buffer) at room temperature for 2 h. After washing with PBS, the embryos were analyzed the using confocal microscope SP5 STED (Leica, Rijswijk, Netherlands).
In vivo imaging and radiography
Metastatic tumor growth was followed weekly by bioluminescence imaging (BLI) with the NightOwl, (Berthold, Bad Wildbad, Germany). The BLI signal intensity was quantified as the sum of photons within a region of interest given as the total flux (photons/second).
Statistical analysis
Statistical analysis was performed using Prism 4 software (GraphPad La Jolla, USA). Results are expressed as the mean ± SD. Two-way analysis of variance (ANOVA) followed by the two-tailed Student t-test were used. P <0.05 was considered to be statistically significant (*0.01 < P <0.05; **0.001 < P <0.01; *** P <0.001).
Discussion
Studies correlating integrin expression levels in human tumors with pathological outcome, such as patient survival and metastasis have identified several integrins that might have an important role in cancer progression. Here we investigated the role of αv integrin in breast cancer invasion and metastasis by using preclinical models for human breast cancer. Selective shRNA-mediated knockdown of αv integrin expression was found to potently mitigate invasion and metastasis of breast cancer cells in zebrafish and mouse xenograft models. This study further established the zebrafish embryo xenograft model as a robust and dependable animal model for cancer research. Transplanted fluorescently labeled mammalian tumor cells in zebrafish can survive, invade and metastasize, and thus display similar behavior to cells transplanted in the traditional mammalian models [
41]. Moreover, the zebrafish embryo model also enables us to monitor the metastasis cascade at the single-cell level.
Metastasis is associated with acquisition of mesenchymal characteristics. Here we explored the impact of αv integrin on maintenance of mesenchymal morphology in mammary carcinoma. Reduced expression of αv integrin in MDA-MB-231 cells was associated with downregulation of mesenchymal effectors. Moreover, αv integrin was upregulated by TGF-β and found to participate in stimulation of TGF-β-induced cell migration, TGF-β-target gene activation, and TGF-β/Smad signaling. This is in line with previous studies showing that various integrins can stimulate TGF-β-induced signaling at multiple levels [
14,
21,
22,
42,
43].
Whether the reduction in basal migration, basal phosphorylation of Smad2 and basal expression of
PAI1 and
CTGF after αv integrin knockdown is caused by reduction of integrin-mediated activation of latent TGF-β, or by interference with autocrine active TGF-β-like signaling remains to be established. Irrespective of the exact mechanism, our results indicate that αv integrin at least in part mediates TGF-β/Smad signaling, which has been shown to be critical for the bone metastasis signature of MDA-MB-231 cells [
44,
45]. This notion is consistent with previous observations that interfering with TGF-β signaling inhibits integrin expression and TGF-β-induced metastasis of breast cancer cells [
46], and that expression of αvβ3 integrin is a key determinant for homing of breast cancer cells to bone [
47]. In addition, αv integrin will also have functions that are independent of TGF-β, such as outside-in and inside-out signaling, which may also be important for invasive and metastatic properties of breast cancer cells [
21]. In the bone metastatic lesions, integrins are essential for the interaction between tumor cells and ECM, and also play a role in osteoclast-bone binding [
48]. For example, αvβ3 has high affinity binding with the bone marrow stroma ligands, osteopontin and vitronectin, to promote prostate bone metastasis [
49].
Interestingly, analysis of a previously published dataset, in which 52 breast cancer cell lines were transcriptionally profiled [
31], revealed significant correlation between expression of α
v integrin and the mesenchymal markers
N-Cadherin,
Slug and
Vimentin. Moreover,
αv integrin was found to be highly expressed in mesenchymal breast cancer, with an invasive and metastatic phenotype. A previous study revealed that BMP7 may inhibit TGF-β-induced EMT and bone metastasis in both breast and prostate cancer by decreasing αvβ3 integrin expression [
19,
50]. Moreover, antibody-mediated blockade of αv integrin function, in particular of αvβ6, downregulates TGF-β-induced EMT and inflammation associated with fibrosis, metastasis and cancer [
51-
54].
Integrin-mediated signaling can enhance cell survival through several mechanisms. These include regulation of the expression of BCL-2 [
55], FLIP [
56], or survival-promoting pathways such as PI3K-AKT [
57] or nuclear factor κB (NF-κB) signaling [
58]. Our findings showed that αv integrin knockdown does not significantly enhance apoptosis in the two breast cancer cell lines studied. This might be explained by the fact that we were targeting a single integrin, and did not fully disrupt integrin-mediated cell survival.
Integrins are appealing therapeutic targets because they are expressed in various cell types involved in tumor progression, and interact with growth factor receptors. Several preclinical studies have shown that integrin antagonists inhibit tumor growth by affecting both tumor cells and tumor-associated host cells (that is, the angiogenic endothelium). Here we assessed the effect of a new non-peptide integrin-specific antagonist, GLPG0187, on breast cancer progression. Administration of GLPG0187 resulted in a decrease of tumor invasion in the zebrafish embryo model. Furthermore, GLPG0187 effectively inhibited the progression of established bone metastasis in a mouse model of breast cancer and showed superior activity when combined with antiresorptive and chemotherapeutic standard-of-care agents. It should be noted that genetic depletion and pharmacological inhibition represent two different approaches to inhibit αv integrin function.
Demonstrating that GLPG0187 inhibits bone metastasis in mice is not sufficient to conclude that the αv integrins on tumor cells are targeted and responsible for the observed effect. This inhibitor acts in a non cell-autonomous manner and also selectively inhibits the interaction of αv integrins of non-tumor cells, for example, stromal cells with extracellular matrix components [
19]. Additional studies in which, for example, the effect of genetic depletion of αv integrin in tumor cells on bone metastasis is analyzed, are needed to directly implicate αv integrin effects on tumor cells in the bone metastatic response. The (dose-dependent) selectivity spectra of shRNA-mediated depletion of the total protein and pharmacological inhibition with GLPG0187 can also be different. For instance, responses that are initiated upon αv integrin activation might be subject to different thresholds of signaling intensity/duration. Functions of αv integrin that are not dependent on its interaction with the ECM may not be affected by GLPG0187, but will be affected by depletion of the protein. Therefore, these two different approaches to address the functional role of αv integrin in breast cancer cells complement and strengthen each other.
Conclusions
Here we demonstrate that αv integrin is required for breast cancer cell invasion and metastasis by regulating mesenchymal markers expression and crosstalk with TGF-β signaling. We translate findings obtained in cell culture and an innovative cost effective and rapid zebrafish xenograft model to an
in vivo mouse model. More specifically, the αv integrin small molecule antagonist GLPG0187 inhibited bone metastasis, and maximum efficacy was achieved when combined with antiresorptive zoledronate or chemotherapeutic paclitaxel. An important issue concerning the interpretation of efficacy of GLPG0187 is that with this approach we cannot exclude possible effects through targeting of integrins on other cell types than the tumor cells, including cells in the metastatic niche or in the vasculature [
19]. In bone metastasis the activity of osteoclasts is important for bone resorption, a process that can be inhibited by targeting integrins. Various integrins, including αvβ3, have been implicated in tumor angiogenesis. Taken together, our data suggest that breast cancer patients with high levels of αv integrin will most likely benefit from a combinatorial pharmacological approach that includes αv integrin targeting.
Competing interests
Philippe Clément-Lacroix and Philippe Pujuguet are employed by Galapagos. All other authors do not have any competing interests.
Authors’ contributions
All authors have been involved in the drafting of the manuscript and the critical analysis of the data, and have read and given final approval of the version to be published. They also agree to be accountable for all aspects of the work. More specifically, YL participated in conception and design, development of methodology, in vivo xenograft assay, clinical data analysis and manuscript writing. YD participated in development of methodology, in vivo xenograft assay and manuscript writing. PP participated in conception, design, and carried out the in vivo mouse xenograft assay. JR carried out immunostaining. TvL carried out FACS analysis. LZ participated in clinical data analysis. PtD participated in conception and design, financial support and manuscript writing. PC participated conception and design, and revision of the manuscript. HvD participated in data analysis and revision of the manuscript.