Background
Despite improvements in conventional anti-cancer therapies such as chemotherapy, radiotherapy and surgery, the five-year survival for patients with non-small cell lung cancer (NSCLC) remains poor. Vascular endothelial growth factor (VEGF) is produced by most tumor types and stimulates the growth of new blood vessels within a tumor where it plays a pivotal role in the process of angiogenesis [
1]. The biological effects of VEGF are mediated via binding to specific tyrosine kinase receptors including VEGFR-1 (Flt-1) and VEGFR-2 (KDR) in addition to non-tyrosine kinase receptors such as Neuropilin-1 (NP1) and Neuropilin-2 (NP2). Co-expression of NP1 and NP2 in NSCLC tissue is significantly correlated with tumor progression and poor prognosis [
2]. NP1 has also been shown to be an independent predictor of cancer relapse and poor survival in NSCLC patients [
3].
In Phase III trials, blocking VEGF using the recombinant humanized VEGF monoclonal antibody Bevacizumab (Avastin®) has proven successful in increasing the objective tumor response rate and in prolonging progression-free and overall survival in patients with NSCLC [
4,
5]. The survival benefit is however relatively short and the majority of patients eventually relapse. The current use of tyrosine kinase inhibitors alone and in combination with chemotherapy has been underwhelming [
6] and the precise effects of removing VEGF from the circulation remains unclear. In a recent study [
7], it was demonstrated that circulating and tumor VEGF-A and NP1 tumor protein expression could select for patients most likely to benefit from the addition of Bevacizumab to chemotherapy in advanced or metastatic gastric cancer patients. Patients with low baseline expression of NP1 showed a trend towards improved overall survival compared to patients with high NP1 expression. These studies suggest that NP1 may play an important role in VEGF-mediated signaling in the tumor cells themselves.
In this study we demonstrate that VEGF is an autocrine growth and cell survival factor for NSCLC cells, acting principally through the NP1 receptor, promoting lung tumor growth. The results indicate that NP1, in particular, should be evaluated as a predictive biomarker with levels of expression potentially defining those patients most likely to benefit from VEGF targeted therapies. Furthermore, NP1 may be a target for therapy in NSCLC and other tumors.
Discussion
At present, drugs targeting angiogenic growth factors are postulated as mediating their anti-tumor effects by inhibiting new blood vessel formation. Experimental models have demonstrated that members of the VEGF family promote tumor growth
in vivo by inducing angiogenesis [
8]. When co-expressed in cells expressing VEGFR-2, NP1 enhances the binding of VEGF
165 to VEGFR-2 and subsequent VEGF
165-mediated chemotaxis [
9,
10]. Although the biological role of VEGFR-1 has remained unclear, cross-linking experiments have shown that VEGF
121 is able to bind both NP1 and NP2 in cells that co-express VEGFR-1, suggesting an interaction between VEGFR-1 and the NPs [
11]. Although experimental evidence indicates that endothelial migration and sprouting that is mediated by VEGF
121 (which binds to both NP1 and VEGFR-2, but cannot form bridges between them) may be inhibited by anti-NP1 antibodies [
12], it is possible that NP1 may have functions that are independent of VEGFR-2, potentially through the NP1 interacting protein (NIP) [
13]. In xenograft experiments, anti-NP1 antibodies have a modest suppressive effect on tumor growth, but significant additive suppressive effects on tumor growth when combined with anti-VEGF therapies [
14]. This is accompanied by reductions in tumor vascular density and maturity, suggesting that targeting NP1 is a valid anti-angiogenic strategy and may help overcome resistance to anti-VEGF therapies.
This anti-angiogenic hypothesis however fails to take into consideration that in patients, tumor cells may proliferate in the absence of neo-angiogenesis by co-opting and modifying the existing vasculature. A role for VEGF in preventing tumor cell apoptosis is supported by reports demonstrating that over-expression of the soluble VEGF receptor NP1, which prevents VEGF binding to the cell surface receptors in tumor cells, is associated with tumor cell apoptosis [
15]. NP1 is expressed on many tumor cell types and increased expression of both NP1 and NP2 has been found to correlate with tumor aggressiveness, advanced disease and poor prognosis [
16,
17].
To address the hypothesis that VEGF is a growth and cell survival factor for NSCLC, cells were treated with VEGF
165 that binds to all four VEGF receptors, VEGFR-1, VEGFR-2, NP1 and NP2. These data demonstrated that VEGF stimulated growth of lung tumor cells expressing NP1, but had no effect on cells that did not express the NP1 receptor. Of interest was our finding that H460 cells, in which NP1 receptor expression is absent, failed to respond to VEGF despite its expression of VEGFR-2. We believe that a critical element behind VEGF-mediated cell survival involves the Neuropilin receptors acting either as hetereodimers or homodimers. It was previously hypothesized by Soker
et al [
10] that in endothelial cells expressing both NP1 and VEGFR-2, NP1 mediates VEGFR-2 activity by serving as a co-receptor, thereby enhancing VEGF binding to the VEGFR-2 receptor resulting in VEGF-mediated downstream signaling events, chemotaxis and angiogenesis. As we currently know, Neuropilins are unable to form homodimers, and as such, must mediate their effects through heterodimeric interactions with other receptors such as KDR or Plexins, highlighting NP1, or NP2, as critical elements involved in mediating VEGF signaling and supporting cell survival in NSCLC.
In cells responding to recombinant VEGF protein, neutralizing antibodies to VEGF inhibited lung tumor cell growth and resulted in the arrest of cells in the G0/G1 phase of the cell cycle, suggesting an important role for VEGF signaling in lung tumor cells. Knockdown of VEGF expression in cancer cells which are responsive to VEGF, reduced cell proliferation further supporting a role for VEGF as a cell growth and survival factor in NSCLC. Such findings are in agreement with those previously reported [
18] demonstrating a role for VEGF in the survival of H1299 lung cancer cells expressing VEGFR-1, VEGFR-2 and NP1. The effects of VEGF on downstream proliferation and survival signaling in the NP1 expressing cell lines, A549 and SKMES1, were clearly demonstrated in this study with induction of phosphorylation of the PI3K mediator Akt and, to a lesser extent, the MAPK signaling proteins Erk1/2, respectively.
These results and other accumulating evidence suggest that the function of VEGF in tumor progression may not be limited to angiogenesis and that the more important role of this pathway is in epithelial cell survival and proliferation [
19]. VEGF autocrine signaling via NP1 has been demonstrated in breast cancer cells [
20,
21]. NP1 complexing with plexin-A1 involving both VEGF and SEMA3a has been implicated in the chemotaxis of breast cancer cell lines [
22]. Preclinical data also support a role for tumor cell NP1 in mediating lung and renal cancer cell migration, proliferation and invasion [
3,
23]. In human FG pancreatic cells, over-expression of NP1 induced both Erk1/2 and JNK signaling pathways [
24]. Down-regulation of the NP1 receptor using siRNA sensitized PANC-1 cells to the cytotoxic effects of the chemotherapeutic agent Gemcitabine, compared to NP1-over-expressing cells. Such findings further implicate NP1 as a cell survival factor in epithelial tumors. In prostate cancer, NP1 was found to be highly expressed by prostate cancer cell lines and displayed a positive association with invasiveness, suggesting that it may be one of the primary receptors responsible for VEGF autocrine effects in prostate cancer cells [
25]. A positive association between NP1 expression and
in vivo bone metastatic potential was found in ARCaPM xenografts and was further confirmed in clinical prostate cancer specimens. Hamerlik
et al showed that VEGF-VEGFR2-NP1-mediated signaling in glioma stem-like cells is maintained in an autocrine manner through the continuous secretion of VEGF, thereby allowing constitutive activation of downstream pro-survival pathways and growth of glioblastomas, tumor invasion and increased resistance to treatment [
26]. In other studies, VEGF and NP1 expression by tumor epithelial cells also regulates the stemness of cutaneous tumors and the expansion of the cancer stem cell (CSC) pool, contributing to enhanced tumor growth [
27]. Conditional deletion of VEGF in tumor epithelial cells caused tumors to regress, whereas VEGF over-expression by tumor epithelial cells accelerated tumor growth. In addition, VEGF affected skin tumor growth by promoting cancer stemness and symmetric CSC division leading to CSC expansion. When expressed as a co-receptor in cutaneous CSCs, deletion of NP1 blocked the ability of VEGF to promote cancer stemness and renewal.
Our data support the observations that NPs play a central role in epithelial cancer cell survival. Stable transfection of NP1 in NP1 non-expressing cells induced NSCLC cell growth in vitro and in vivo. The in vitro proliferation was augmented by addition of recombinant VEGF. In contrast, stable knockdown of NP1, NP2 and anti-VEGFR-2 antibody treatment induced inhibition of tumor cell growth in NP1 positive lung adenocarcinoma and squamous cell lines. This work cannot discount an independent role for NP2 as a transmitter of cell survival signaling for VEGF and this will be explored in future studies. What is clear however, is that expression of NP1 together with VEGFR-2 may be critical for the autocrine survival and growth effects of VEGF in NSCLC.
The combined targeting of ligand and co-receptor may help to overcome resistance to targeted agents such as bevacizumab in a subset of patients demonstrated to over-express NP1. This is supported by preclinical data demonstrating that anti-NP1 antibodies have additive anti-cancer activity in combination with anti-VEGF therapy [
14]. A more comprehensive analysis of the expression of NP1 by epithelial tumor cells may help to inform prospectively planned biomarker driven studies of the clinical benefit of bevacizumab, VEGFR- and NP1-targeted agents. In this regard, recent studies showing NP1 immuno-positivity by tumor cells in 6% of primary and 14% of metastatic breast cancers, and 36% of primary and 50% of metastatic NSCLC provides a framework for testing this combined approach in patients [
28,
29].
Methods
Cell lines
A panel of non-small cell lung cancer cells, H460 (large cell carcinoma), H647 (adenosquamous carcinoma), A549 (adenocarcinoma) and SKMES1 (squamous cell carcinoma) were used. H460 and H647 cells were purchased from the American Tissue Culture Collection (ATCC), while A549 and SKMES1 cells were purchased from the European Cell and Culture Collection (ECACC). H460 and H647 cells were maintained in Roswell Park Memorial Institute (RPMI-1640) medium in a humidified atmosphere of 5% CO2 in air at 37°C. A549 cells were maintained in Ham’s F12 supplemented with 4 mM L-glutamine, while SKMES1 cells were cultured in EMEM media supplemented with 2 mM L-glutamine and 1% non-essential amino acids (NEAA). All media were supplemented with 10% heat-inactivated fetal bovine serum (FBS), penicillin (100 U/ml) and streptomycin (100 μg/ml) (Lonza, UK). All cells were maintained as monolayer cultures and exponentially growing cultures were used in all experiments. All cell lines were tested and authenticated six months prior to this study using the PowerPlex® 16 HS System (Source BioScience, UK), a multiplex STR system.
Analysis of mRNA expression by RT-PCR
Total RNA was isolated using Tri-reagent (MRC Inc, OH, USA). First-strand cDNA was prepared from 1 μg of total RNA using Superscript III reverse transcriptase (Invitrogen, UK) according to manufacturer’s instructions. PCR reactions were carried out for the VEGF receptors, VEGFR-1, VEGFR-2, NP1 and NP2. The endothelial cell line, EAhy926, was used as a positive control. Primer sequences used were as follows:
VEGFR-1 Forward: 5′CAAGTGGCCAGAGGCATGGAGTT3′
Reverse: 5′GATGTAGTCTTTACCATCCTGTTG3′
VEGFR-2 Forward: 5′GAGGGCCTCTCATGGTGATTGT3′
Reverse: 5′TGCCAGCAGTCCAGCATGGTCTG3′
NP1 Forward: 5′ATGGAGAGGGGGCTGCCG3′
Reverse: 5′CTATCGCGCTGTCGGTGTA3′
NP2 Forward: 5′CCCCGAACCCAACCAGAAGA3′
Reverse: 5′GAATGCCATCCCAGATGTCCA3′
VEGF Forward: 5′CGCAAGCTTAGGAGTACCCTGATGAG3′
Reverse: 5′CCGTCTAGAACATTTGTTGTGCTGT′
β-actin amplification was carried out in parallel to account for loading differences between samples:
β-actin Forward: 5′TGTTTGAGACCTTCAACACCC3′
Reverse: 5′AGCACTGTGTTGGCGTACAG3′
Specificity of all primers was confirmed by comparing the primer sequence for each gene against the Genbank database. PCR products were analyzed on a 1% agarose gel and images acquired using the BioSpectrum® Imaging System (UVP, CA, USA).
Western blot analysis
Total protein was extracted from cells using ice-cold RIPA buffer (50 mM Tris HCl, pH 7.4, 150 mM NaCl, 1 mM EDTA, 1% (v/v) Triton-X 100, 0.1% (w/v) SDS) supplemented with phenylmethylsulfonyl fluoride (PMSF) and protease inhibitor cocktail (2 mM AEBSF, 1 mM EDTA, 130 μM Bestatin, 14 μM E-64, 1 μM Leupeptin, 0.3 μM Aprotinin). Protein concentrations were determined using the bicinchoninic acid assay (BCA) as per manufacturer’s instructions. Protein (40 μg) from whole cell lysates was fractionated on 8% or 12% SDS-PAGE gels and transferred to a PVDF membrane (PALL Corporation, FL, USA). Transfer efficiency and loading was confirmed by reversible staining of the membrane with Ponceau S solution (Sigma-Aldrich, UK) following protein transfer. Membranes were blocked at room temperature with 5% non-fat dry milk in Tris-buffered saline (TBS) containing 0.1% Tween-20 (TBS-T), followed by incubation with the appropriate primary antibodies at room temperature or otherwise stated: rabbit anti-VEGF (Millipore, CA, USA), 1:2000 at 4°C in 5% BSA TBS-T (0.05%); mouse anti-Flt-1 (Millipore, CA, USA), 1:500 at 4°C in 3% Marvel TBS-T (0.05%); rabbit anti-KDR (Upstate, USA), 1:5000 at room temperature in 5% Marvel TBST-T (0.05%); goat anti-NP1 and rabbit anti-NP2 (Santa Cruz Biotech, CA, USA), 1:400 at room temperature in 5% Marvel TBS-T (0.1%); anti β-actin (Merck Biosciences, UK), 1:20000 at room temperature in 5% Marvel TBS-T (0.1%). Membranes were washed in TBST and incubated with a secondary horseradish peroxidase (HRP)-labeled antibody for 1 h at room temperature (1:2000 at room temperature in 0.1% TBS-T). Membranes were washed in TBST following incubation with secondary antibody. Bound antibody complexes were detected and visualized using SuperSignal™ West Pico enhanced chemiluminescence substrate (Pierce, IL, USA).
Cell proliferation
Cell survival/proliferation was measured using the bromodeoxyuridine (BrdU) cell proliferation ELISA according to manufacturer’s instructions (Roche Diagnostics, Germany). BrdU labeling solution was added at a final concentration of 10 μM. Cells were fixed for 60 min followed by incubation for 90 min anti-BrdU antibody (1:100). Wells were washed and incubated in substrate solution. Absorbance was measured at 450 nm using a reference wavelength at 690 nm.
High content imaging & confocal microscopy
NSCLC cells were seeded (1 × 104) in MatriPlate™ 96-well glass bottomed micro-well plates (Matrical Bioscience, WA, USA) and allowed to adhere overnight. Following serum depletion (0.5% FBS), cells were treated with recombinant human VEGF (100 ng/ml), VEGF neutralizing antibodies (1 μg/ml) or in combination, for 6 h. Cells were fixed in 3% paraformaldehyde and washed in PBS. After washing, cells were blocked in 5% normal goat serum for 1 h followed by incubation with primary rabbit phospho-Akt (1:400) (Millipore) and phospho-p44/p42 MAPK (1:400) (Cell Signaling Technology) primary antibodies in 4% BSA overnight at 4°C. Cells were washed in blocking buffer and incubated with a secondary Alexa Fluor® 488 (Invitrogen) goat anti-rabbit antibody (1:1000), red phalloidin (1:1,000) and Hoechst 33342 (1:500) at room temperature for 1 h. After washing in PBS, localization and expression levels of phospho-Akt and phospho-Erk1/2 were examined on the In Cell 1000 analyzer (GE Healthcare, UK) using IN Cell Investigator high-content image analysis software (version 1.5). For confocal microscopy analysis, NSCLC cells were seeded in glass chamber slides and allowed to adhere overnight. Following serum depletion (0.5% FBS), cells were treated with recombinant human VEGF (100 ng/ml) or VEGF neutralizing antibodies (1 μg/ml) for 6 hrs. Cells were fixed in 3% paraformaldehyde and washed in PBS. After washing, cells were incubated in blocking buffer containing 5% bovine serum albumin (BSA) for 1 h and incubated with rabbit phospho-Akt (1:200) and p44/p42 MAPK (Erk1/2) (1:50) primary antibodies (Cell Signaling Technology) at 4°C overnight. Cells were then washed in PBS and incubated with a secondary Alexa Fluor® 488 (Invitrogen) goat anti-rabbit antibody (1:1000) and Hoechst 33342 for nuclear staining at room temperature for 1 h. After washing in PBS, localization and expression levels of phospho-Akt and phospho-Erk1/2 were examined using a Zeiss LSM 510 laser scanning confocal microscope (Carl Zeiss International, Germany).
Cell cycle analysis
Cells were detached and pelleted by centrifugation at 1300 rpm for 3 min. Supernatants were discarded and cells were suspended in 1 ml phosphate-buffered saline (PBS) and fixed in 90% ice-cold ethanol. Following incubation at room temperature for 30 min, cells were resuspended in 1 ml PBS containing propidium iodide (25 μg/ml) and DNase-free RNase A (100 μg/ml) and left at 37°C for 30 min. DNA synthesis and cell cycle distribution was measured by FACS (Becton Dickinson, UK).
siRNA transient transfections
siRNA ON-TARGETplus SMART pool siRNA for NP1, NP2, VEGF and VEGFR-2 (KDR) were designed and synthesized (Dharmacon Inc, USA). Each siRNA pool contains four individual sequences to silence target gene expression at the mRNA level by at least 75%. A non-targeting scrambled control was also included for each target gene of interest. Cells at 60% confluence were transfected in penicillin/streptomycin-free media with each siRNA (100 nM) using DharmaFect1 transfection reagent (Dharmacon Inc, USA) according to manufacturer’s instructions. After 6 h, siRNAs were removed and cells were maintained in complete media for 24, 48 and 72 h. At each time point, total protein was extracted from A549 and SKMES1 cells for Western blot analysis to determine knockdown of each gene at the protein level. As an alternative to siRNA, due to low levels of knockdown of VEGFR-2, a blocking antibody to VEGFR-2 (sc-19530) (Santa Cruz Biotech, Germany) was also used.
Generation of NP1 stable transfected NSCLC cells
NP1 plasmid DNA was inserted into the site of the mammalian vector pcDNA3.1(-) (Invitrogen Corporation, CA, USA) to generate pcDNA3.1(-)-NP1 plasmid constructs. The NP1 plasmid constructs, including a pcDNA3.1(-) empty vector control, were individually transfected into the NP1 negative cell line, H460. Stable transfections were carried out using FuGENE HD™ transfection reagent (Roche Diagnostics Ltd., UK). Cells (3 × 105) were cultured in their respective supplement-free medium and transfected with either 1 μg pcDNA3.1(-)-NP1 or pcDNA-3.1(-) (control vector) in antibiotic-free media containing 3 μL/mL FuGENE HD™ according to manufacturers’ instructions. Following transfection, cells were further incubated for 24 h at 37°C. Antibiotic selection was then carried out by treating the cells with Geneticin G418 (800 μg/mL). Following several rounds of antibiotic selection, clones were selected and characterized at the mRNA and protein levels in order to examine relative NP1 expression levels.
In vivo tumor growth studies
Nude mice on a BALB/c background (CBy.CG-Foxn1
nu) were purchased from Jackson Laboratories (Bar Harbor, MD, USA). Female mice, 10 weeks of age were utilized. Animals were housed under specific pathogen-free conditions in individually ventilated and filtered cages under positive pressure. All animal experiments were performed in compliance with Irish Department of Health and Children regulations (Licence B100/3250) and approved by the Trinity College Dublin BioResource Ethical Review Board. Mice were anaesthetized with isofluorane and injected subcutaneously on the left-hand side dorsal flank with 3 × 106 H460 empty vector control cells (n = 8) or 3 × 106 NP1 stable transfectant cells (n = 8). Mice were monitored and weighed weekly. Final tumor volume was recorded using digital callipers and calculated based on the equation (D1)2 × D2 × 0.524, where D1 is the smaller of the two diameters of the tumor measured in both directions. Experiments were terminated when the tumor volume reached 2 cm3. Tumors were excised and retained for further analyses. H460 cells (NP1-negative) were transfected with a NP1 plasmid to over-express this receptor for the in vivo component of this study.
Statistical analysis
Statistical analysis was carried out using analysis of variance (ANOVA) with post-hoc analysis using Bonferroni multiple comparisons test, unless otherwise stated. Where the means of two data sets were compared, an unpaired Students t-test was used. Data is graphically represented as mean ± standard error of the mean (SEM) following three independent experiments, where p < 0.05 was considered statistically significant. All data were analyzed using GraphPad InStat™ (version 3.0) statistical software.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
MPB and KJO’B contributed to the conception, design and coordination of the study. MPB carried out all experiments, participated in statistical analysis, data interpretation and was responsible for preparing the manuscript. SGG participated in the study design and provided the NP1 plasmid construct. KG and AMD helped with image analysis and confocal microscopy. EH and PJF advised and helped with the in vivo mouse study. DJR and GPP revised the manuscript. All authors read and approved the final manuscript prior to submission.