The PDGF (platelet-derived growth factor) family members are potent mitogens for cells of mesenchymal origin and serve as important regulators of cell migration, survival, apoptosis and transformation. Tumour-derived PDGF ligands are thought to function in both autocrine and paracrine manners, activating receptors on tumour and surrounding stromal cells. PDGF-C and -D are secreted as latent dimers, unlike PDGF-A and -B. Cleavage of the CUB domain from the PDGF-C and -D dimers is required for their biological activity. At present, little is known about the proteolytic processing of PDGF-C, the rate-limiting step in the regulation of PDGF-C activity. In the present study we show that the breast carcinoma cell line MCF7, engineered to overexpress PDGF-C, produces proteases capable of cleaving PDGF-C to its active form. Increased PDGF-C expression enhances cell proliferation, anchorage-independent cell growth and tumour cell motility by autocrine signalling. In addition, MCF7-produced PDGF-C induces fibroblast cell migration in a paracrine manner. Interestingly, PDGF-C enhances tumour cell invasion in the presence of fibroblasts, suggesting a role for tumour-derived PDGF-C in tumour–stromal interactions. In the present study, we identify tPA (tissue plasminogen activator) and matriptase as major proteases for processing of PDGF-C in MCF7 cells. In in vitro studies, we also show that uPA (urokinase-type plasminogen activator) is able to process PDGF-C. Furthermore, by site-directed mutagenesis, we identify the cleavage site for these proteases in PDGF-C. Lastly, we provide evidence suggesting a two-step proteolytic processing of PDGF-C involving creation of a hemidimer, followed by GFD-D (growth factor domain dimer) generation.

INTRODUCTION

PDGF (platelet-derived growth factor) regulates a diverse array of cellular processes, such as cell proliferation and motility, under normal physiological conditions as well as during the pathogenesis of a number of human diseases [15]. Immunohistochemical analysis of human breast cancer tissues showed localized membranous PDGFR (PDGF receptor) expression/activation in periepithelial stromal cell populations, suggesting a paracrine stimulation of adjacent stromal tissue by breast tumour cells [6]. Moreover, a previous study of invasive ductal breast carcinomas demonstrated an association between α-PDGFR staining and lymph node metastasis [7]. In experimental models of breast cancer, PDGF initiates a human breast carcinoma desmoplastic response via paracrine signalling [8,9]. PDGF autocrine signalling was shown to be essential for tumour formation and metastasis in Ras-mediated mammary tumours, as well as in MMTV (murine mammary tumour virus)-Neu/TGF-β (transforming growth factor-β) transgenic mice [10]. Importantly, our recent tissue microarray analysis of 216 patients with invasive breast cancer showed that increased PDGF-C expression correlates with lymph node metastases, increased Ki-67 proliferation staining, and lower rates of 7-year disease-free survival (Y. H. Meng, N. Hurst Jr, A. Najy, A. Bottrell, J. Won, F. Miller, C. J. Kim, E.-S. Kim, A. Moon, E. J. Kim, S. Y. Park and H. -R. C. Kim, unpublished work). Furthermore, a role for PDGF-C in malignancy was suggested by the observation that Ewing sarcoma cell lines showed increased expression and secretion of PDGF-C [11,12]. When a dominant-negative PDGF-C construct was expressed in these cell lines, they showed a reduction of anchorage-independent growth, but not a full reversion of the phenotype [11,12]. PDGF-C autocrine signalling has also been suggested for the initiation and progression of brain tumours, such as glioblastoma and medulloblastoma [13,14]. However, little is known at present about the role of PDGF-C in breast cancer.

Although the classical PDGF ligands A and B are secreted as active heterodimers or homodimers, newly identified PDGF ligands C and D are secreted as latent homodimers containing an N-terminal CUB domain and a C-terminal GFD (growth factor domain) [15,16]. Previous reports have demonstrated that plasmin and a component of FBS (fetal bovine serum) were capable of processing latent PDGF-C into the GFD, and identified the PDGF-C cleavage site to be between Lys225 and Ala226 by N-terminal sequencing of the GFD of PDGF-C isolated from BHK-570 cells [15,17]. A later study reported that PDGF-C is a substrate of tPA (tissue plasminogen activator) in in vitro biochemical assays, and Arg231 in the hinge region of PDGF-C is essential for its cleavage by tPA [18]. Proteolytically activated PDGF-C stimulates its cognate receptor, α-PDGFR, but can also activate β-PDGFR via αβ-PDGFR heterodimerization [15,17].

The goals of the present study were to identify breast carcinoma-produced proteases responsible for extracellular proteolytic cleavage of PDGF-C, a key step to initiate PDGF-C/PDGFR signalling, and to investigate the potential oncogenic activities of PDGF-C in breast cancer. In the present study we identified tPA, uPA (urokinase-type plasminogen activator) and matriptase as potential activators of PDGF-C in breast cancer. We show that the FL-PDGF-C (full-length PDGF-C) dimer undergoes proteolytic cleavage in a two-step process, creating a hemidimer containing one chain of FL-PDGF-C monomer and one chain of GFD-PDGF-C monomer followed by GFD dimer (GFD-D). Interestingly, while Lys225 is the putative proteolytic cleavage site, the LLGK (amino acids 222–225) motif appears to be critical for the first cleavage for the generation of hemidimer; both the LLGK and RKSR (amino acids 231–234) motifs in the hinge region between the CUB and GFD domains of PDGF-C are essential for the second cleavage for the generation of the GFD dimer. Importantly, increased PDGF-C expression in MCF7 cells increased cell proliferation, anchorage-independent cell growth and tumour cell motility, demonstrating a potential oncogenic activity of PDGF-C in breast cancer. Importantly, we also provide evidence that PDGF-C potentiates tumour cell invasion through paracrine signalling in fibroblasts.

MATERIALS AND METHODS

Cell culture and reagents

Cell lines used in the present study were purchased from American Type Culture Collection (A.T.C.C.) and maintained as recommended. MCF7 human breast carcinoma cells, the resultant MCF7 transfectant cell lines and murine NIH 3T3 fibroblasts were cultured in a humidified 5% CO2 incubator with DMEM (Dulbecco's modified Eagle's medium)/F12 medium supplemented with 10% FBS. CV-1 green monkey kidney cells were cultured in DMEM supplemented with 10% FBS. All cell lines were supplemented with 100 units/ml penicillin, 100 mg/ml streptomycin, 2 mM glutamine and fungizone. The proteases tPA and uPA, and the protease inhibitors PAI-1 (plasminogen-activator inhibitor-1), TAPI [TNFα (tumour necrosis factor α) protease inhibitor] and α1-PDX (α1-antitrypsin Portland) were purchased from EMD Biosciences. HAI-1 [HGF (heptocyte growth factor) activator inhibitor type 1] was obtained from R&D Systems. Aprotinin and leupeptin were purchased from Sigma–Aldrich.

Construction of viral PDGF-C expression vectors

An RT (reverse transcription)–PCR approach was taken to clone PDGF-C into a vaccinia expression vector. Total RNA was isolated from the prostate cancer cell lines DU145 and PC3 using TRIzol® reagent and used in cDNA synthesis reactions using SuperScript RT-III (Invitrogen). The resultant cDNAs were then used in PCRs that yielded a 1071-bp product containing the 1035-bp ORF (open reading frame) of PDGF-C encoding FL-PDGF-C. Additionally, this product contained the restriction sites for NcoI and BamHI at the 5′ and 3′ ends respectively of the PDGF-C ORF. Furthermore, to the C-terminus of the PDGF-C product we added a His6 epitope tag (forward, 5′-CATGCCATGGGGAGCCTCTTCGGGCTTCTC-3′; reverse, 5′-CGGGATCCCTAATGGTGATGGTGATGATGTCCTCCTGTGCTCCCTCT-3′; the His6 tag is underlined). This product was then digested with NcoI and BamHI and inserted into the NcoI/BamHI site of the vaccinia virus expression vector pTF7-ECM1 (a gift from Dr R. Fridman, Wayne State University, Detroit, MI, U.S.A.). Fidelity of the in-frame sequence encoding the FL-PDGF-C–His fusion protein was confirmed by DNA sequencing (Elim Biopharmaceuticals). This plasmid is referred to as pTF7-PDGF-C-FL–His. The pTF7-PDGF-C-FL–His construct was then used as a template in site-directed mutagenesis to create point mutants. Mutations were made at Lys225 to alanine (K225A), Arg231 and Arg234 to alanine (R231A/R234A), and Lys225, Arg231 and Arg234 to alanine (K225A/R231A/R234A). Following the manufacturer's protocol for the QuikChange® Site-directed Mutagenesis Kit (Stratagene), primers were designed to create a single mutant at K225A (forward, 5′-CCAACTTGGCAACTTCTTGGCGCGGCTTTTGTTTTTGG-3′; reverse, 5′-CCAAAAACAAAAGCCGCGCCAAGAAGTTGCCAAGTTGG-3′) and a double mutant at R231A/R234A (forward, 5′-GGCTTTTGTTTTTGGAGCAAAATCCGCAGTGGTGGATCTG-3′; reverse, 5′-CAGATCCACCACTGCGGATTTTGCTCCAAAAACAAAAGCC-3′). After confirmation of sequencing, the pTF7-PDGF-C-FL–His R231A/234A construct was used as a template for creation of a triple mutant at K225A/R231A/R234A using the same primers that generated the K225A mutant. The sequencing of this construct was then confirmed.

Production of rPDGF-C (recombinant PDGF-C) protein

Established vaccinia virus protocols [19] were followed to generate rPDGF-C ligand. Briefly, CV-1 cells were first infected with the recombinant vaccinia virus vTF7-3, which expresses the T7 RNA polymerase. After 30 min of infection, the cells were washed with PBS and transfected with the plasmid pTF7-PDGF-C-FL–His, using Effectene® reagent (Qiagen). Expression of the FL-PDGF-C protein inserted into the pTF7 plasmid is reliant on infection of the cells by vTF7-3. Cell-host machinery then transcribes the gene of interest. At 48 h after co-infection/transfection with vaccinia virus and pTF-7 PDGF-C-FL–His, the serum-free CM (conditioned medium) was collected and cleared of cellular debris by a 5 min centrifugation at 2000 g. The resultant CM containing FL-PDGF-C protein was used in subsequent PDGF-C processing experiments.

Construction of a mammalian PDGF-C expression vector

To insert FL-PDGF-C into the mammalian expression vector pcDNA3.1 (+) (Invitrogen), a double restriction enzyme digest was performed. The wild-type FL-PDGF-C–His sequence was first cut from pIND-PDGF-C-FL–His plasmid with the enzymes AflII and BamHI. The resulting product was then gel-purified and ligated into the pcDNA3.1 (+) plasmid. This plasmid is referred to as pcDNA3.1-PDGF-C-FL–His.

Establishment of a FL-PDGF-C-overexpressing MCF7 breast carcinoma cell line

Human breast carcinoma MCF7 cells were transfected with pcDNA3.1-PDGF-C-FL–His and pcDNA3.1 (+) empty vector using Effectene® reagent and then selected using 400 μg/ml Geneticin (G418) in complete culture medium for 12 days. The resulting resistant cells were pooled together and are referred to as MCF7-PDGF-C and MCF7-Neo respectively. Overexpression of FL-PDGF-C was confirmed by RT–PCR as follows. Total RNA was extracted from cells using TRIzol® reagent (Invitrogen). Total RNA (5 μg) from each cell line was used to synthesize cDNA using SuperScript RT-III (Invitrogen). Resultant cDNA was then used as a template in a PCR, using Taq polymerase (Promega) and the following primers: forward, 5′-TCCAGCAACAAGGAACAGAA-3′ and reverse, 5′-GGGTCTTCAAGCCCAAATCT-3′. These primers amplify a 200-bp product that represents part of the CUB domain in the FL-PDGF-C protein. This primer pair does not allow for discrimination of endogenously and exogenously expressed FL-PDGF-C. GAPDH (glyceradehyde-3-phosphate dehydrogenase) expression was used as a positive control with the following primers: forward, 5′-ATCACCATCTTCCAGGAGCGA-3′ and reverse, 5′-GCCAGTGAGCTTCCCGTTCA-3′.

Custom Ab (antibody) raised against the PDGF-C GFD

An Ab was raised against the PDGF-C protein using a synthetic peptide (N-CGRKSRVVDLNLLTEEVRLYSC-C) representing a portion of the PDGF-C GFD (amino acids 230–250). The resultant Ab was affinity-purified (Zymed Biomedical) and is referred to as anti-PDGF-C GFD Ab.

PDGF-C-mediated paracrine activation of α-PDGFR

MCF7-PDGF-C and MCF7-Neo cells were cultured in serum-free media for 48 h. The resultant CM was collected and centrifuged for 5 min at 2000 g to remove cellular debris. Serum-starved NIH 3T3 cells were treated with this CM from the MCF7 transfectants for 10 min. Lysates were collected using RIPA lysis buffer [0.5% sodium deoxycholate, 1% Nonidet P40, 50 mM Tris/HCl (pH 7.6), 2 mM EGTA, 2 mM EDTA, 150 mM NaCl, 2 mM sodium orthovanadate, 1 mM PMSF, 1 mM sodium fluoride and a protein inhibitor cocktail tablet (Roche)]. The lysates were centrifuged for 20 min at 4°C at 12000 g to remove debris, and the supernatant was collected. Total protein concentrations were determined using the BCA (bicinchoninic acid) Protein Assay Kit (Pierce Biotechnology). Lysates (500 μg) were then used for immunoprecipitation with an anti-α-PDGFR Ab (Santa Cruz Biotechnology) and Protein G–agarose beads (Pierce Biotechnology). Immunoprecipitates were washed three times with RIPA lysis buffer and resolved by reducing SDS/PAGE. Tyrosine-phosphorylated α-PDGFR was detected by immunoblot using an anti-phosphotyrosine Ab (Upstate Biotechnology). Total levels of α-PDGFR were detected using the same Ab used for immunoprecipitation. Activation assays were repeated three times.

Inhibition experiments with MCF7-PDGF-C cells

MCF7-PDGF-C cells were seeded in six-well plates, washed three times with PBS, and incubated in the presence of serum-free medium with the protease inhibitors leupeptin, aprotinin, E64, E64c, TAPI, PAI-1, HAI-1, a tPA-blocking Ab or a uPA-blocking Ab for 48 h. After incubation, the resultant CM was collected and analysed by SDS/PAGE under reducing conditions, immunoblotted and probed with anti-PDGF-C GFD Ab. Inhibitory analysis was observed in three separate experimental runs.

Ni-NTA (Ni2+-nitrilotriacetate) concentration of rPDGF-C

rPDGF-C expressed by vaccinia virus contains a His6 tag at the C-terminal end of FL-PDGF-C. Therefore a Ni-affinity protocol was used to immunoprecipitate the FL-PDGF-C and/or GFD forms. Briefly, 10× binding buffer (500 mM sodium phosphate, 10 mM imidazole and 0.5% Tween 20) was added to CM collected from vaccinia virus co-infected/transfected FL-PDGF-C-expressing cells in the presence of a 50% Ni-NTA agarose bead slurry (Qiagen). The media and beads were then rocked overnight at 4°C, after which time the beads are spun down by centrifugation and washed twice in wash buffer (1× binding buffer and 300 mM NaCl). The rPDGF-C was then eluted from the beads by incubating them overnight at 4°C in the presence of 10 mM EDTA. This PDGF-C eluate was then aliquoted and stored at −80°C until needed for subsequent experiments.

In vitro cleavage of rPDGF-C by tPA, uPA or matriptase

Latent rPDGF-C was incubated with various concentrations of human tPA [in 50 mM Tris/HCl (pH 7.5) and 50 mM NaCl] or human uPA {in 50 mM Tris/HCl (pH 8.8), 50 mM NaCl and 0.1% PEG [poly(ethylene) glycol]-4000} for 16–18 h at 37°C. For matriptase experiments, latent rPDGF-C was incubated for 2 h with various concentrations of the matriptase catalytic domain (a gift from Dr C.-Y. Lin, University of Maryland, College Park, MD, U.S.A.) at 37°C [in 50 mM Tris/HCl (pH 7.5) and 100 mM NaCl]. After incubation, the products were analysed by SDS/PAGE under reducing or non-reducing conditions and immunoblotted with anti-PDGF-C GFD Ab. PDGF-C protease cleavage experiments were verified in three independent experiments.

Plasminogen-casein or casein zymography

MCF7-PDGF-C and MCF7-Neo CM were resolved on an 11% plasminogen-casein SDS/PAGE for 1h at 35 mA on ice. After electrophoresis, the gel was washed twice with 2.5% Triton X-100, before incubation overnight at 37°C in 0.1 M glycine (pH 8.0). The following day the gel was stained for 2 h at room temperature (20°C) with gentle agitation in 0.1% Amido Black. After incubation, the gel was destained in 30% methanol/10% acetic acid. Finally, the gel was placed in a softening solution of 5% glycerol/5% acetic acid for 30 min before drying overnight. The gel was subsequently imaged on a Microtek Scanmaker i900 using Photoshop Elements software. Zymographic analysis was repeated three different times.

Scratch migration assay

NIH 3T3 fibroblasts were seeded in a six-well plate and allowed to attach overnight at 37°C. The next day, these cells were washed once with PBS and incubated for 30 min at 37°C with 1.0 ml of serum-free DMEM/F12 medium+25 μg/ml mitomycin C. After incubation, the medium was aspirated, and an injury line was scraped with a 1–200 μl yellow pipette tip. The cells were carefully washed once with PBS. Next, 2 ml of either MCF7-PDGF-C or MCF7-Neo CM was added to each of three wells. The cells were then placed in the incubator at 37°C for 24 h. Pictures were taken of the cells every 8 h to monitor closure of the gap. Using ImageJ (NIH), closure was assessed as the percentage of cleared area remaining at time 0, 8 and 16 h. Migration assays were repeated three times.

Soft agar analysis

MCF7-PDGF-C and MCF7-Neo cells were embedded in a 0.35% Bacto agarose (BD Biosciences) gel at 5000 cells/ml and laid on a 0.6% agarose bottom layer. Experiments were performed in triplicate. Cells were allowed to grow for 2 weeks, and were then stained with 4% Geimsa (Riedel-de Haën). Colonies ranging from 200 to 500 μm in size were counted using the Optronix GelCount. Soft agar analysis was observed in three separate experimental runs.

WST-1 (water-soluble tetrazolium salt 1) proliferation assay

MCF7-PDGF-C and MCF7-Neo cells were plated in a 96-well plate at 2000 cells per well. Experiments were carried out in sets of six wells per cell line. Cells were then treated with WST-1 reagent (Roche) according to the manufacturer's recommendations for 3 h, and then samples were read on a Benchmark microplate reader (Bio-Rad Laboratories) at 450 nm. PDGF-C-mediated cell proliferation was verified in three independent experiments.

Matrigel™ cell invasion

MCF7-PDGF-C and MCF7-Neo cells were plated in serum-free medium at a density of 3.75×105 cells/ml into Matrigel™-coated transwells (BD Biosciences). Transwells were then placed in 24-wells containing either serum-free medium or a confluent layer of serum-starved NIH 3T3 cells (co-culture). Invasion was permitted to occur over 48 h, and then transwells were cleaned and stained with 0.9% Crystal Violet. The number of invading cells was counted in five different high-powered fields using an inverted Nikon TMS microscope. Cell invasion analysis was repeated three different times.

RESULTS

MCF7 cells process PDGF-C into its active GFD

To examine whether breast cancer cells contain the protease(s) responsible for PDGF-C activation and to examine the role of PDGF-C in breast cancer progression, we established MCF7 cells engineered to express PDGF-C (MCF7-PDGF-C) or a control vector (MCF7-Neo) as described in the Materials and methods section. PDGF-C overexpression in MCF7-PDGF-C cells was confirmed by RT–PCR analysis (Figure 1A). Importantly, immunoblot analysis under reducing conditions using CM collected from these cells detected both the full-length monomer of PDGF-C (FL-M) with a molecular mass of ~48 kDa, as well as the processed GFD monomer (GFD-M) of PDGF-C with a molecular mass of ~17 kDa (Figure 1B, left-hand panel). Consistently, immunoblot analysis under non-reducing conditions confirmed the full-length dimer (FL-D) of PDGF-C (~85 kDa) and GFD dimer (GFD-D) of PDGF-C (~26 kDa) (Figure 1B, right-hand panel). These results demonstrate that breast cancer cells express proteases capable of processing PDGF-C independent of a component in serum, which was previously reported to be a major enzyme to process PDGF-C to its active form [17]. Importantly, we found that breast carcinoma-processed GFD–PDGF-C dimer was biologically active for induction of α-PDGFR phosphorylation, as demonstrated by a receptor activation assay in NIH 3T3 fibroblasts (Figure 1C).

Expression and processing of PDGF-C in MCF7 cells

Figure 1
Expression and processing of PDGF-C in MCF7 cells

(A) PDGF-C levels in MCF7-PDGF-C (lane 1) and MCF7-Neo (lane 2) were examined by RT–PCR. (B) Serum-free CM collected from MCF7-PDGF-C or MCF7-Neo cells were analysed by immunoblot analysis using anti-PDGF-C GFD Ab under reducing (left-hand panel) and non-reducing (right-hand panel) conditions. FL-M, full-length monomer; GFD-M, GFD monomer; FL-D, full-length dimer; GFD-D, GFD dimer. The molecular mass in kDa is indicated on the left-hand side. (C) Serum-starved NIH 3T3 fibroblasts were stimulated for 10 min with CM collected from MCF7-PDGF-C (lane 1) or MCF7-Neo (lane 2) cells, serum-free (SF) medium with 20 ng/ml PDGF-AA (lane 3, positive control) or serum-free medium (lane 4, negative control). Cell lysates were immunoprecipitated (IP) using an anti-PDGF-C Ab and activated α-PDGFR was detected by immunoblot (IB) using a an anti-phosphotyrosine (pTyr) Ab.

Figure 1
Expression and processing of PDGF-C in MCF7 cells

(A) PDGF-C levels in MCF7-PDGF-C (lane 1) and MCF7-Neo (lane 2) were examined by RT–PCR. (B) Serum-free CM collected from MCF7-PDGF-C or MCF7-Neo cells were analysed by immunoblot analysis using anti-PDGF-C GFD Ab under reducing (left-hand panel) and non-reducing (right-hand panel) conditions. FL-M, full-length monomer; GFD-M, GFD monomer; FL-D, full-length dimer; GFD-D, GFD dimer. The molecular mass in kDa is indicated on the left-hand side. (C) Serum-starved NIH 3T3 fibroblasts were stimulated for 10 min with CM collected from MCF7-PDGF-C (lane 1) or MCF7-Neo (lane 2) cells, serum-free (SF) medium with 20 ng/ml PDGF-AA (lane 3, positive control) or serum-free medium (lane 4, negative control). Cell lysates were immunoprecipitated (IP) using an anti-PDGF-C Ab and activated α-PDGFR was detected by immunoblot (IB) using a an anti-phosphotyrosine (pTyr) Ab.

Tumour-derived PDGF-C mediates both paracrine and autocrine signalling

Since PDGF was shown to induce a desmoplastic response in breast cancer [9], we sought to determine whether tumour-derived PDGF-C induces fibroblast cell migration. To this end, a scratch migration assay was performed, whereby NIH 3T3 fibroblasts treated with CM collected from MCF7-PDGF-C cells demonstrated an increased ability to migrate compared with the Neo-treated control (Figure 2A). Next we examined the autocrine effects of tumour-derived PDGF-C on induction of transformed phenotype and tumour cell growth. As shown in Figures 2(B) and 2(C), increased PDGF-C expression resulted in an increase in anchorage-independent cell growth and an increased cell proliferation as assessed using a soft agar assay and a WST-1 proliferation assay respectively. Lastly, to assess the effect of PDGF-C on tumour cell invasion, Matrigel™ invasion assays were performed. MCF7 cells were barely invasive in a Matrigel™ invasion assay, regardless of PDGF-C expression (Figure 2D). However, when MCF7-PDGF-C and MCF7-Neo cells were seeded on top of transwell filters coated with growth factor reduced-Matrigel™ in the presence of NIH 3T3 fibroblasts in the lower chamber, MCF7-PDGF-C cells were more invasive than the Neo control. This suggests that fibroblasts respond to tumour-derived PDGF-C, resulting in secretion of factors that potentiate the invasive phenotype in tumour cells. Taken together, these results suggest the potential oncogenic activities of PDGF-C in breast tumour growth, its mediation of tumour–stromal interactions, and tumour cell invasion via both autocrine and paracrine manners.

In vitro transformative properties of PDGF-C in MCF7 cells

Figure 2
In vitro transformative properties of PDGF-C in MCF7 cells

(A) Paracrine signalling. A scratch migration assay of NIH 3T3 fibroblasts was performed in the presence of CM collected from MCF7-neo or MCF7-PDGF-C cells. Using ImageJ (NIH), closure of the gap was quantified as the percentage of cleared area remaining at time 0, 8 and 16 h from three independent experiments (top panel). Representative 40× images of time 0 and 16 h are shown (bottom panel). (B and C) Autocrine signalling. Anchorage-independent growth and proliferation of MCF7-neo or MCF7-PDGF-C cells were assessed by a soft agar colony formation assay (B) and WST-1 cell proliferation assay (C) respectively. Positive colonies were quantified from three separate experiments using Optronix GelCount in (B) and cell proliferation was quantified from three independent WST-1 assays in (C). (D) Autocrine and double paracrine signalling. Matrigel™ invasion assay of MCF7-neo or MCF7-PDGF-C cells through a modified Boyden chamber was performed in the absence or presence of NIH 3T3 fibroblasts in the bottom chamber. Values are means±S.D. (n=3). *P<0.05.

Figure 2
In vitro transformative properties of PDGF-C in MCF7 cells

(A) Paracrine signalling. A scratch migration assay of NIH 3T3 fibroblasts was performed in the presence of CM collected from MCF7-neo or MCF7-PDGF-C cells. Using ImageJ (NIH), closure of the gap was quantified as the percentage of cleared area remaining at time 0, 8 and 16 h from three independent experiments (top panel). Representative 40× images of time 0 and 16 h are shown (bottom panel). (B and C) Autocrine signalling. Anchorage-independent growth and proliferation of MCF7-neo or MCF7-PDGF-C cells were assessed by a soft agar colony formation assay (B) and WST-1 cell proliferation assay (C) respectively. Positive colonies were quantified from three separate experiments using Optronix GelCount in (B) and cell proliferation was quantified from three independent WST-1 assays in (C). (D) Autocrine and double paracrine signalling. Matrigel™ invasion assay of MCF7-neo or MCF7-PDGF-C cells through a modified Boyden chamber was performed in the absence or presence of NIH 3T3 fibroblasts in the bottom chamber. Values are means±S.D. (n=3). *P<0.05.

A serine protease processes FL-PDGF-C in MCF7 cells

The above results indicate that the MCF7 cell line provided a good model to elucidate which protease(s) were capable of activating this growth factor in the context of breast carcinoma cells. To identify the class of protease responsible for the proteolytic processing of FL-PDGF-C, MCF7 cells were cultured in serum-free conditions in the presence of several class-specific protease inhibitors. As shown in Figure 3(A), aprotinin, a serine protease inhibitor, effectively inhibited FL-PDGF-C processing in these cells. Unlike aprotinin, the cysteine protease inhibitor leupeptin, the cathepsin inhibitors E64 and E64c, the MMP (matrix metalloproteinase) inhibitor TAPI, and the furin inhibitor PDX-Portland showed no significant inhibition of FL-PDGF-C processing, indicating a specific role of serine proteases in the activation of FL-PDGF-C in breast carcinoma. This finding is consistent with a previous report that the serine protease tPA is responsible for the processing of PDGF-C in human fibroblasts [20].

PDGF-C is processed by serine proteases, specifically tPA and uPA

Figure 3
PDGF-C is processed by serine proteases, specifically tPA and uPA

(A) MCF7-PDGF-C cells were incubated in serum-free media with the various class-specific inhibitors for 48 h; the collected CM was resolved under reducing SDS/PAGE and immunoblotted with an anti-PDGF-C GFD Ab. (B) Left-hand panel: MCF7-PDGF-C cells were incubated with serum-free media without or with PAI-1. Right-hand panel: MCF7-PDGF-C (lanes 1 and 3) or MCF7-neo (lanes 2 and 4) cells were incubated with serum-free media containing tPA- or uPA-specific inhibitors for 48 h; the collected CM was resolved under reducing SDS/PAGE and immunoblotted with an anti-PDGF-C GFD Ab. (C) MCF7-PDGF-C (lane 3) and MCF7-Neo (lane 4) CM was run in a plasminogen-dependent zymogram with recombinant tPA (lane 1) and recombinant uPA (lane 2) serving as positive activity controls. (D) rPDGF-C was generated by co-infecting/transfecting CV-1 cells with vaccinia virus and the pTF7-PDGF-C–His construct. After 48 h of serum-starvation the CM was collected. This CM was concentrated using Ni-NTA agarose beads overnight. After washing of the beads, rPDGF-C was eluted with 10 mM EDTA and then rPDGF-C was incubated with tPA or uPA overnight. Finally, the resultant products were resolved by SDS/PAGE and immunoblotted with an anti-PDGF-C GFD Ab. (E) rPDGF-C was incubated with tPA, uPA, streptokinase and streptokinase+glu-plasminogen (native form of plasminogen with an N-terminal glutamate residue) overnight, and the resultant products were resolved by SDS/PAGE and immunoblotted with an anti-PDGF-C GFD Ab. The molecular mass in kDa is indicated on the left-hand side. Ctrl, control; FL-M, full-length monomer; GFD-M, GFD monomer; SK, streptokinase; Undig., undigested.

Figure 3
PDGF-C is processed by serine proteases, specifically tPA and uPA

(A) MCF7-PDGF-C cells were incubated in serum-free media with the various class-specific inhibitors for 48 h; the collected CM was resolved under reducing SDS/PAGE and immunoblotted with an anti-PDGF-C GFD Ab. (B) Left-hand panel: MCF7-PDGF-C cells were incubated with serum-free media without or with PAI-1. Right-hand panel: MCF7-PDGF-C (lanes 1 and 3) or MCF7-neo (lanes 2 and 4) cells were incubated with serum-free media containing tPA- or uPA-specific inhibitors for 48 h; the collected CM was resolved under reducing SDS/PAGE and immunoblotted with an anti-PDGF-C GFD Ab. (C) MCF7-PDGF-C (lane 3) and MCF7-Neo (lane 4) CM was run in a plasminogen-dependent zymogram with recombinant tPA (lane 1) and recombinant uPA (lane 2) serving as positive activity controls. (D) rPDGF-C was generated by co-infecting/transfecting CV-1 cells with vaccinia virus and the pTF7-PDGF-C–His construct. After 48 h of serum-starvation the CM was collected. This CM was concentrated using Ni-NTA agarose beads overnight. After washing of the beads, rPDGF-C was eluted with 10 mM EDTA and then rPDGF-C was incubated with tPA or uPA overnight. Finally, the resultant products were resolved by SDS/PAGE and immunoblotted with an anti-PDGF-C GFD Ab. (E) rPDGF-C was incubated with tPA, uPA, streptokinase and streptokinase+glu-plasminogen (native form of plasminogen with an N-terminal glutamate residue) overnight, and the resultant products were resolved by SDS/PAGE and immunoblotted with an anti-PDGF-C GFD Ab. The molecular mass in kDa is indicated on the left-hand side. Ctrl, control; FL-M, full-length monomer; GFD-M, GFD monomer; SK, streptokinase; Undig., undigested.

To determine whether the plasminogen activator system of proteases plays a role in FL-PDGF-C processing in MCF7 cells, CM was collected from MCF7-PDGF-C cells in the presence of the tPA/uPA-specific inhibitor PAI-1, or in the presence of blocking antibodies against tPA and uPA. As shown in Figure 3(B), PAI-1 effectively, although not completely, inhibited proteolytic cleavage of FL-PDGF-C into GFD-PDGF-C. Similarly, tPA-blocking antibodies, and to a lesser degree uPA-blocking antibodies, were able to inhibit the processing of PDGF-C. The tPA or uPA activities in CM from MCF7-PDGF-C and MCF7-Neo were confirmed by plasminogen-casein zymography, as shown in Figure 3(C). These findings suggest that MCF7-produced tPA and, to a lesser extent, uPA activated PDGF-C.

Both tPA and uPA are able to process PDGF-C in vitro

To determine whether recombinant tPA and/or uPA can directly process PDGF-C, we first generated rPDGF-C proteins using the vaccinia virus system, followed by purification and concentration using Ni-NTA agarose beads. When rPDGF-C was incubated with increasing concentrations of tPA or uPA overnight at 37°C, both tPA and uPA were able to process rPDGF-C into the GFD-PDGF-C (Figure 3D), suggesting that PDGF-C is a substrate for tPA and uPA. However, if plasminogen contamination during purification of rPDGF-C exists, uPA/tPA-mediated rPDGF-C processing may be indirect, as tPA and uPA are potent activators of plasminogen to plasmin, an enzyme known to cleave PDGF-C [15,21]. To address this, rPDGF-C was incubated with 100 nM streptokinase, a bacterial plasminogen activator. Streptokinase activates plasminogen in a non-proteolytic manner, and streptokinase itself has no proteolytic activity of its own [22]. As shown in Figure 3(E), in the presence of streptokinase, there was no processing of PDGF-C to its growth domain form, indicating no significant amount of plasminogen to be activated for the processing of PDGF-C under our experimental conditions. As a positive control to ensure that streptokinase is functional, rPDGF-C was incubated with streptokinase and 0.3 nM plasminogen. As shown in Figure 3E (lane 5), streptokinase-activated plasminogen completely processed rPDGF-C into a smaller size of GFD–PDGF-C. These results demonstrate that both tPA and uPA are able to directly process PDGF-C to its GFD form in vitro, independent of plasmin.

Identification of tPA/uPA cleavage sites within the hinge domain of PDGF-C

Gilbertson et al. [17] predicted the serum-sensitive cleavage site of PDGF-C to be RKSR (amino acid residues 231–234), but identified the PDGF-C cleavage site to be between Lys225 and Ala226 by N-terminal sequencing of the GFD of PDGF-C isolated from BHK-570 cells [17]. A later study reported through mutagenic analysis that Arg231 in the hinge region of PDGF-C is essential for cleavage by tPA [18]. Taking these results, together with our amino acid sequence analysis, we focused on two putative sites, LLGK (amino acids 222–225) and RKSR (amino acids 231–234), for uPA- and/or tPA-mediated proteolytic cleavage of PDGF-C, as depicted in Figure 4(A). By site-directed mutagenesis and the vaccinia virus expression system, rPDGF-C mutants with Lys225 mutated to alanine (K225A), Arg231 and Arg234 mutated to alanine (R231A/R234A), and Lys225, Arg231 and Arg234 mutated to alanine (K225A/R231A/R234A) were generated. It should be noted that since the present study regarding tPA processing of PDGF-C parallels two studies that have been performed by Fredriksson et al. [18,20], we followed the protocols contained within these studies for comparison of tPA processing of PDGF-C. The greatest inhibition of tPA-mediated proteolytic cleavage was observed in the K225A/R231A/R234A mutant, whereas the R231A/R234A mutant showed the least inhibition of processing by tPA when the cleavage products were analysed under reducing conditions (Figure 4B, top panel). Interestingly, when these samples were analysed under non-reducing conditions, GFD dimer was generated from wild-type rPDGF-C proteins only (Figure 4B, bottom panel, lane 2). With the K225A or R231A/R234A mutants, generation of GFD dimer was almost completely inhibited with a noticeable accumulation of a product of a molecular mass of ~50 kDa (Figure 4B, bottom panel, lanes 4 and 6). This product has been termed a hemidimer, containing one FL-PDGF-C monomer and one GFD-PDGF-C monomer [20]. When the uPA cleavage site was examined in PDGF-C, Lys225 in the LLGK motif appeared to be more critical than the arginine residues (Arg231 and Arg234) in the RKSR motif (Figure 4C, top panel). Similarly to the tPA-mediated PDGF-C processing, GFD dimer formation was drastically inhibited in these mutants (Figure 4C, bottom panel). These results suggest that Lys225 in the LLGK motif within the hinge region between the CUB domain and GFD is critical for tPA- and uPA-mediated first cleavage of PDGF-C for the generation of the hemidimer. However, both LLGK and RKSR motifs are important for the second cleavage for the generation of the GFD-PDGF-C dimer.

Mutational analysis identifies tPA and uPA cleavage sites

Figure 4
Mutational analysis identifies tPA and uPA cleavage sites

(A) Comparison of sequence alignments between human PDGF-C, murine PDGF-C and human PDGF-D. Mutagenesis sites are marked with asterisks and the tPA/uPA substrate specificity in the P1–P4 residues are shown. (B) Western blot analysis of PDGF-C wild-type (WT), PDGF-C K225A, PDGF-C R231A/R234A and PDGF-C K225A/R231A/R234A mutants under reducing (top panel) and non-reducing (bottom panel) conditions when incubated with 100 nM tPA for 4 h at 37°C in the presence of fibrinogen (fibr.) fragments. (C) Western blot analysis of PDGF-C wild-type (WT), PDGF-C K225A, PDGF-C R231A/R234A and PDGF-C K225A/R231A/R234A under reducing (top panel) and non-reducing (bottom panel) conditions when incubated with 100 nM uPA overnight at 37°C. The molecular mass in kDa is indicated on the left-hand side. FL-D, full-length dimer; GFD-D, GFD dimer; HD, hemidimer; FL-M, full-length monomer; GFD-M, GFD monomer.

Figure 4
Mutational analysis identifies tPA and uPA cleavage sites

(A) Comparison of sequence alignments between human PDGF-C, murine PDGF-C and human PDGF-D. Mutagenesis sites are marked with asterisks and the tPA/uPA substrate specificity in the P1–P4 residues are shown. (B) Western blot analysis of PDGF-C wild-type (WT), PDGF-C K225A, PDGF-C R231A/R234A and PDGF-C K225A/R231A/R234A mutants under reducing (top panel) and non-reducing (bottom panel) conditions when incubated with 100 nM tPA for 4 h at 37°C in the presence of fibrinogen (fibr.) fragments. (C) Western blot analysis of PDGF-C wild-type (WT), PDGF-C K225A, PDGF-C R231A/R234A and PDGF-C K225A/R231A/R234A under reducing (top panel) and non-reducing (bottom panel) conditions when incubated with 100 nM uPA overnight at 37°C. The molecular mass in kDa is indicated on the left-hand side. FL-D, full-length dimer; GFD-D, GFD dimer; HD, hemidimer; FL-M, full-length monomer; GFD-M, GFD monomer.

Matriptase is able to process PDGF-C in vitro

Incomplete inhibition of PDGF-C processing in the presence of the tPA/uPA inhibitor PAI in MCF7 cells (Figure 3B) suggested the presence of additional serine proteases with the capacity to process PDGF-C. Interestingly, the serine protease matriptase was shown to have substrate specificity for the amino acid motif RXSR by phage display [23]. Since this putative matriptase cleavage site was found in the RKSR motif in the hinge region of PDGF-C, we sought to investigate whether matriptase can process PDGF-C. First, we confirmed matriptase expression in MCF7 cells by immunoblot analysis, as shown in Figure 5(A). When rPDGF-C was incubated with various concentrations of the purified catalytic domain of matriptase for 2 h, FL-PDGF-C was processed into GFD-PDGF-C by 1–5 nM matriptase, as shown in Figure 5(B). To assess whether the cleavage site for matriptase was the same as for tPA and uPA, rPDGF-C R231A/R234A, K225A and K225/R231A/R234A mutants were also incubated with 1 nM matriptase for 2 h at 37°C. Matriptase-treated rPDGF-C samples were then analysed by immunoblot analysis under reducing and non-reducing conditions. Similarly to uPA and tPA, Lys225 was critical for matriptase cleavage of FL-PDGF-C into GFD-PDGF-C, as detected under reducing conditions (Figure 5C, top panel). However, none of these PDGF-C mutant dimers underwent matriptase-mediated proteolytic cleavage in both chains for the generation of the GFD dimer as analysed under non-reducing conditions (Figure 5C, bottom panel), suggesting that both LLGK and GRSR motifs are required for matriptase generation of the GFD-PDGF-C dimer.

Matriptase is capable of processing PDGF-C

Figure 5
Matriptase is capable of processing PDGF-C

(A) Western blot analysis of MCF7-PDGF-C and MCF7-Neo CM and lysates for matriptase expression. The blot was probed with anti-matriptase Ab under non-reducing conditions. (B) rPDGF-C wild-type (WT) incubated with various concentrations of matriptase for 2 h at 37°C and then resolved by SDS/PAGE under reducing conditions. (C) rPDGF-C wild-type (WT) and mutants were incubated with 1 nM matriptase for 2 h at 37°C and then resolved by SDS/PAGE under reducing conditions (top panel) or non-reducing conditions (bottom panel). (D) rPDGF-C wild-type (WT) or R231A/R234A were incubated 2 h at 37°C with 1 nM matriptase before being suspended in serum-free DMEM/F12 media and then added to NIH 3T3 cells for 10 min. Subsequently, the lysates from these cells were analysed by Western blot for the presence of phospho-β-PDGFR. Total β-PDGFR was used as a loading control. MCF7-PDGF-C cells were treated with the matriptase inhibitor HAI-1 and PDGF-C processing (E) and biological activity (F) were monitored. In the same gel, an unnecessary lane separating SFM and NT treatment was removed in (E). A white line is drawn to demarcate this change. (G) CM from (E) was utilized to assess NIH 3T3 cell migration. Values are means±S.D. (n=3). * and **P<0.05. The molecular mass in kDa is indicated on the left-hand side. Ctrl., control; FL-D, full-length dimer; GFD-D, GFD dimer; HD, hemidimer, FL-M, full-length monomer; GFD-M, GFD monomer; IB, immunoblot; IP, immunoprecipitate; Mat., matriptase; NT, no treatment; p-, phospho; SF, serum-free; SFM, serum-free medium; t-, total; Undig., undigested.

Figure 5
Matriptase is capable of processing PDGF-C

(A) Western blot analysis of MCF7-PDGF-C and MCF7-Neo CM and lysates for matriptase expression. The blot was probed with anti-matriptase Ab under non-reducing conditions. (B) rPDGF-C wild-type (WT) incubated with various concentrations of matriptase for 2 h at 37°C and then resolved by SDS/PAGE under reducing conditions. (C) rPDGF-C wild-type (WT) and mutants were incubated with 1 nM matriptase for 2 h at 37°C and then resolved by SDS/PAGE under reducing conditions (top panel) or non-reducing conditions (bottom panel). (D) rPDGF-C wild-type (WT) or R231A/R234A were incubated 2 h at 37°C with 1 nM matriptase before being suspended in serum-free DMEM/F12 media and then added to NIH 3T3 cells for 10 min. Subsequently, the lysates from these cells were analysed by Western blot for the presence of phospho-β-PDGFR. Total β-PDGFR was used as a loading control. MCF7-PDGF-C cells were treated with the matriptase inhibitor HAI-1 and PDGF-C processing (E) and biological activity (F) were monitored. In the same gel, an unnecessary lane separating SFM and NT treatment was removed in (E). A white line is drawn to demarcate this change. (G) CM from (E) was utilized to assess NIH 3T3 cell migration. Values are means±S.D. (n=3). * and **P<0.05. The molecular mass in kDa is indicated on the left-hand side. Ctrl., control; FL-D, full-length dimer; GFD-D, GFD dimer; HD, hemidimer, FL-M, full-length monomer; GFD-M, GFD monomer; IB, immunoblot; IP, immunoprecipitate; Mat., matriptase; NT, no treatment; p-, phospho; SF, serum-free; SFM, serum-free medium; t-, total; Undig., undigested.

To determine whether the GFD dimer and/or hemidimer generated by matriptase have biological activity, we performed the following experiment. Wild-type rPDGF-C and R231A/R234A mutant were incubated with 1 nM matriptase for 2 h at 37°C. The processed PDGF-C was then suspended in serum-free media and utilized for the treatment of serum-starved NIH 3T3 cells for 10 min, followed by immunoblot analysis using an anti-phospho-β-PDGFR Ab. β-PDGFR activation in NIH 3T3 cells was chosen as the readout for PDGF-C activity for the following reasons. NIH 3T3 cells express a high level of β-PDGFR, and PDGF-C was shown to activate β-PDGFR via αβ-PDGFR heterodimerization [17]. As shown in Figure 5(D), wild-type rPDGF-C processed by matriptase was biologically active and induced phosphorylation of PDGFR, whereas the hemidimer or GFD monomer generated by matriptase cleavage of the R231A/R234A PDGF-C mutant exhibited no biological activity.

To determine whether MCF7-produced matriptase is critical for PDGF-C processing and signalling, MCF7-PDGF-C cells were treated with HAI-1, an endogenous inhibitor of matriptase. Generation of the GFD of PDGF-C was reduced, accompanied with accumulation of FL-PDGF-C (Figure 5E), supporting a role for matriptase in the proteolytic processing of PDGF-C in breast carcinoma. Next, we examined the biological consequence of reduced PDGF-C processing on its ability to activate PDGFR and induce cell motility in a paracrine manner. To this end, serum-starved fibroblasts were treated with CM collected from MCF7-PDGF-C cells in the absence or presence of HAI-I and analysed for the status of β-PDGFR activation and for their motility. As shown in Figure 5(F), β-PDGFR activation was less apparent by treatment with CM from HAI-1-treated cells compared with the non-treated cells. Accordingly, NIH 3T3 cell migration was greatly inhibited following treatment with CM from HAI-1-treated MCF7-PDGF-C cells compared with the control (Figure 5G). Taken together, these results demonstrated a critical role for matriptase in the activation of PDGF-C signalling.

DISCUSSION

The present study demonstrates that both tPA and uPA can process latent FL-PDGF-C into the active form, whereas our previous study found that uPA, but not tPA, is capable of processing latent PDGF-D into the active form [24]. tPA and uPA belong to the chymotrypsin family of serine proteases, and share a high degree of structural similarity and primary physiological substrates, such as plasminogen [2527]. Unlike plasmin, they display highly restricted substrate specificity and play unique roles under physiological and pathological conditions. It is thought that tPA is involved in vascular fibrinolysis, whereas uPA has been implicated in the immune response, tissue remodelling, angiogenesis, cancer growth and metastasis [25,28]. Since tPA and uPA have differing physiological roles, yet are similar structurally and exhibit a high degree of substrate specificity, several studies have been performed to understand their basis for recognizing substrate sequences [23,2933]. These studies utilized substrate phage display, phage substrate subtraction library and PS-SCL (positional scanning-synthetic combinatorial library) techniques to determine specificity. Briefly, substrate phage display identifies the most labile substrates, and substrate subtraction identifies the most selective substrates for a given enzyme, whereas PS-SCL allows a rapid and minimally difficult technique for the determination of proteolytic substrate specificity. Both tPA and uPA prefer arginine residues over lysine in the P1 position; however, uPA has less preference for arginine than for lysine, indicating that uPA is a less specific protease than tPA [32]. Both enzymes show similar specificity for serine/glycine/alanine residues at the P2 position [23,33]. Furthermore, the P3 and P4 residues have been shown to be the primary determinants in substrate recognition between tPA and uPA, with a P3 arginine and P4 large hydrophobic residue being the most selective for tPA over uPA [30]. In our mutagenesis assays, PDGF-C processing still occurs with the R231A mutation. This may result from an introduction of a new cleavage site as a result of mutation. Indeed, the R231A mutation introduces a potential cleavage site at Lys232 with an alanine residue in the P2 site (R231A) [33]. However, although these studies provide insights into the ability of tPA and uPA to cleave certain peptide sequences, it was suggested that the P1–P4 peptide sequences alone are not enough for recognition and cleavage in vivo; there is also a requirement for protease–protein interactions at nearby or distant sites [29]. Consistent with the notion that uPA is a less specific protease than tPA, our previous study [24] and the present study found that tPA can only cleave PDGF-C, whereas uPA can process both PDGF-C and -D.

A previous report by Fredriksson et al. [18] has demonstrated that tPA processing of PDGF-C requires Arg231 in the RKSR (amino acids 231–234) motif in the hinge region to mediate proteolytic activation of PDGF-C [18]. This conclusion was reached by site-directed mutagenesis of PDGF-C followed by an in vitro cleavage assay with tPA and a PDGFR activation assay. Although this conclusion is valid, the present study provided an alternative Lys225-processing site and further insight into PDGF-C processing by tPA, as well uPA and matriptase. The results of the present study indicate that Lys225 in the LLGK (amino acid 222–225) motif is more critical for the first cleavage generation of the PDGF-C hemidimer. In line with this finding, the Lys225 site was the cleavage site of PDGF-C sequenced by Gilbertson et al. [17] in BHK-570 cells. More importantly, the present study showed that both the LLGK and RKSR motifs are essential for generation of the biologically active GFD-PDGF-C dimer. These results suggest differences in the structure of the hinge regions of the two PDGF-C chains for the recognition/access of serine proteases to latent PDGF-C for proteolytic activation. The results of the present study also provide important information that the presence of GFD-PDGF-C detected under reducing conditions may not be a good predictor for the biological activity of PDGF-C.

The present study also identified matriptase as a potential activator of PDGF-C in breast cancer. Matriptase was first identified as the major gelatinase in hormone-dependent human breast cancer cells [34] and, since its initial discovery, matriptase has been cloned by several laboratories and has shown a restricted epithelial expression profile in both human and murine tissues [3540]. At present, only a handful of substrates have been identified, including PAR-2 (proteinase-activated receptor-2), single chain uPA, the pro-form of MT-SP1 (membrane-type serine protease 1) and HGF [23,41]. Importantly, increasing evidence suggested a critical role for matriptase in human cancers, as it is highly expressed in carcinomas of the head and neck, and mesothelium, breast, ovary, cervix, prostate, lung and gastrointestinal cancers [4245]. Dysregulated matriptase activity is thought to directly affect the extracellular microenvironment during cancer progression through altering processing of matrix components, could act through downstream effector molecules such as growth factors and receptors, chemokines or protease zymogens, leading to their activation or deactivation. Our finding suggests the possibility of matriptase's oncogenic activity involving activation of PDGF-C signalling.

The relative biological significance of tPA, uPA and matriptase for the regulation of PDGF-C activity in vivo is likely to depend on the bioavailability of these enzymes in a given microenvironment. The expression and role of the plasminogen system has been extensively studied during tumour progression and showed increased expression/activity of uPA and tPA in several tumour types, including breast cancer, gliomas, non-small cell lung cancer, pancreatic cancer and prostate cancer [4650]. In a recent study of breast cancer patient tissue samples, uPA and uPAR (uPA receptor) expression was correlated with increasing tumour stage and in breast cancer bone metastases by in situ hybridization analyses [51]. However, a role for tPA in breast cancer is unclear. One study showed that tPA expression in breast cancer is not associated with the malignant or benign state [52]. Another report states that low intratumoral tPA expression levels correlate with aggressiveness and poor prognosis for breast cancer patients [53]. Irrespective of these reports, increased PDGF-C expression and serine protease-mediated proteolytic activation of PDGF-C in human breast carcinoma MCF7 cells resulted in fibroblast migration in a paracrine manner, as well as increased MCF7 cell proliferation and anchorage-independent growth in an autocrine manner. In addition, serine protease-activated PDGF-C signalling induced an invasive phenotype in MCF7 cells in the presence of fibroblasts. These results suggest the potential for regulatory cross-talk between extracellular serine proteases and PDGF signalling as being critical for carcinoma growth, as well as for tumour–stromal interactions leading to tumour invasion.

Furthermore, we utilized the Oncomine database (Compendium Biosciences), which is a compendium of published cDNA arrays from over 100 institutions that allows for gene comparison across multiple array studies, in an effort to obtain information as to which serine protease(s) is (are) likely to be available for PDGF-C activation during breast cancer progression. Using this tool, we found that among the PDGF family members, PDGF-C and its cognate receptor α-PDGFR were more likely to be up-regulated during breast cancer progression. Interestingly, although only one out of 19 studies reported an increase in tPA expression in breast carcinoma, almost all studies found increases in matriptase and uPA in breast cancer tissues (Supplementary Table S1 at http://www.BiochemJ.org/bj/441/bj4410909add.htm). Lastly, our laboratory (Y. H. Meng, N. Hurst Jr, A. Najy, A. Bottrell, J. Won, F. Miller, C. J. Kim, E.-S. Kim, A. Moon, E. J. Kim, S. Y. Park and H.-R.C. Kim, unpublished work) demonstrated in a tissue microarray analysis of 216 patients with invasive breast cancer that increased PDGF-C expression correlates with lymph node metastasis, increased Ki-67 proliferation staining and lower rates of 7-year disease-free survival, providing compelling evidence for PDGF-C in human breast cancer progression. Taken together, matriptase and uPA are more likely than tPA to be the relevant enzymes for PDGF-C activation, contributing to breast cancer progression.

Abbreviations

     
  • Ab

    antibody

  •  
  • CM

    conditioned medium/media

  •  
  • DMEM

    Dulbecco's modified Eagle's medium

  •  
  • FBS

    fetal bovine serum

  •  
  • GFD

    growth factor domain

  •  
  • HAI-1

    HGF (heptocyte growth factor) activator inhibitor type 1

  •  
  • HGF

    heptocyte growth factor

  •  
  • Ni-NTA

    Ni2+-nitrilotriacetate

  •  
  • ORF

    open reading frame

  •  
  • PAI-1

    plasminogen-activator inhibitor-1

  •  
  • PDGF

    platelet-derived growth factor

  •  
  • FL-PDGF-C

    full-length PDGF-C

  •  
  • PDGFR

    PDGF receptor

  •  
  • rPDGF-C

    recombinant PDGF-C

  •  
  • PS-SCL

    positional scanning-synthetic combinatorial library

  •  
  • RT

    reverse transcription

  •  
  • TAPI

    TNFα (tumour necrosis factor α) protease inhibitor

  •  
  • tPA

    tissue plasminogen activator

  •  
  • uPA

    urokinase-type plasminogen activator

  •  
  • WST-1

    water-soluble tetrazolium salt 1

AUTHOR CONTRIBUTION

Newton Hurst, Abdo Najy, Carolyn Ustach and Hyeong-Reh Choi Kim conceived and designed the experiments. Newton Hurst, Abdo Najy and Lisa Movilla performed the experiments and analysed the data. Newton Hurst, Abdo Najy, Carolyn Ustach and Hyeong-Reh Choi Kim wrote the paper.

We thank Dr Shijie Sheng for thoughtful discussions and Dr M. Katie Conley-LaComb with her assistance in the preparation of this paper prior to submission, and the Translational Research Core Facility of the Karmanos Cancer Institute for their assistance with analysis of the soft agar colony assay.

FUNDING

This work was supported by the National Institutes of Health/National Cancer Institute [RO1 grant numbers CA064139 and CA123362 (to H.-R.C.K.)]; and the Ruth L. Kirschstein National Research Service Award [grant number T32-CA009531 (to N.H.)].

References

References
1
Deuel
T. F.
Polypeptide growth factors: roles in normal and abnormal cell growth
Annu. Rev. Cell Biol.
1987
, vol. 
3
 (pg. 
443
-
492
)
2
Ross
R.
Raines
E. W.
Bowen-Pope
D. F.
The biology of platelet-derived growth factor
Cell
1986
, vol. 
46
 (pg. 
155
-
169
)
3
Rosenkranz
S.
Kazlauskas
A.
Evidence for distinct signaling properties and biological responses induced by the PDGF receptor α and β subtypes
Growth Factors
1999
, vol. 
16
 (pg. 
201
-
216
)
4
Anderberg
C.
Li
H.
Fredriksson
L.
Andrae
J.
Betsholtz
C.
Li
X.
Eriksson
U.
Pietras
K.
Paracrine signaling by platelet-derived growth factor-CC promotes tumor growth by recruitment of cancer-associated fibroblasts
Cancer Res.
2009
, vol. 
69
 (pg. 
369
-
378
)
5
Crawford
Y.
Kasman
I.
Yu
L.
Zhong
C.
Wu
X.
Modrusan
Z.
Kaminker
J.
Ferrara
N.
PDGF-C mediates the angiogenic and tumorigenic properties of fibroblasts associated with tumors refractory to anti-VEGF treatment
Cancer Cell
2009
, vol. 
15
 (pg. 
21
-
34
)
6
Coltrera
M. D.
Wang
J.
Porter
P. L.
Gown
A. M.
Expression of plateletderived growth factor B-chain and the platelet-derived growth factor receptor β subunit in human breast tissue and breast carcinoma
Cancer Res.
1995
, vol. 
55
 (pg. 
2703
-
2708
)
7
Carvalho
I.
Milanezi
F.
Martins
A.
Reis
R. M.
Schmitt
F.
Overexpression of platelet-derived growth factor receptor α in breast cancer is associated with tumour progression
Breast Cancer Res.
2005
, vol. 
7
 (pg. 
R788
-
R795
)
8
Walker
R. A.
The complexities of breast cancer desmoplasia
Breast Cancer Res.
2001
, vol. 
3
 (pg. 
143
-
145
)
9
Shao
Z. M.
Nguyen
M.
Barsky
S. H.
Human breast carcinoma desmoplasia is PDGF initiated
Oncogene
2000
, vol. 
19
 (pg. 
4337
-
4345
)
10
Jechlinger
M.
Sommer
A.
Moriggl
R.
Seither
P.
Kraut
N.
Capodiecci
P.
Donovan
M.
Cordon-Cardo
C.
Beug
H.
Grunert
S.
Autocrine PDGFR signaling promotes mammary cancer metastasis
J. Clin. Invest.
2006
, vol. 
116
 (pg. 
1561
-
1570
)
11
Zwerner
J. P.
May
W. A.
PDGF-C is an EWS/FLI induced transforming growth factor in Ewing family tumors
Oncogene
2001
, vol. 
20
 (pg. 
626
-
633
)
12
Zwerner
J. P.
May
W. A.
Dominant negative PDGF-C inhibits growth of Ewing family tumor cell lines
Oncogene
2002
, vol. 
21
 (pg. 
3847
-
3854
)
13
Lokker
N. A.
Sullivan
C. M.
Hollenbach
S. J.
Israel
M. A.
Giese
N. A.
Platelet-derived growth factor (PDGF) autocrine signaling regulates survival and mitogenic pathways in glioblastoma cells: evidence that the novel PDGF-C and PDGF-D ligands may play a role in the development of brain tumors
Cancer Res.
2002
, vol. 
62
 (pg. 
3729
-
3735
)
14
Andrae
J.
Molander
C.
Smits
A.
Funa
K.
Nister
M.
Platelet-derived growth factor-B and -C and active α-receptors in medulloblastoma cells
Biochem. Biophys. Res. Commun.
2002
, vol. 
296
 (pg. 
604
-
611
)
15
Li
X.
Ponten
A.
Aase
K.
Karlsson
L.
Abramsson
A.
Uutela
M.
Backstrom
G.
Hellstrom
M.
Bostrom
H.
Li
H.
, et al. 
PDGF-C is a new protease-activated ligand for the PDGF α-receptor
Nat. Cell Biol.
2000
, vol. 
2
 (pg. 
302
-
309
)
16
Bergsten
E.
Uutela
M.
Li
X.
Pietras
K.
Ostman
A.
Heldin
C. H.
Alitalo
K.
Eriksson
U.
PDGF-D is a specific, protease-activated ligand for the PDGF β-receptor
Nat. Cell Biol.
2001
, vol. 
3
 (pg. 
512
-
516
)
17
Gilbertson
D. G.
Duff
M. E.
West
J. W.
Kelly
J. D.
Sheppard
P. O.
Hofstrand
P. D.
Gao
Z.
Shoemaker
K.
Bukowski
T. R.
Moore
M.
, et al. 
Platelet-derived growth factor C (PDGF-C), a novel growth factor that binds to PDGF α and β receptor
J. Biol. Chem.
2001
, vol. 
276
 (pg. 
27406
-
27414
)
18
Fredriksson
L.
Ehnman
M.
Fieber
C.
Eriksson
U.
Structural requirements for activation of latent platelet-derived growth factor CC by tissue plasminogen activator
J. Biol. Chem.
2005
, vol. 
280
 (pg. 
26856
-
26862
)
19
Fuerst
T. R.
Earl
P. L.
Moss
B.
Use of a hybrid vaccinia virus-T7 RNA polymerase system for expression of target genes
Mol. Cell. Biol.
1987
, vol. 
7
 (pg. 
2538
-
2544
)
20
Fredriksson
L.
Li
H.
Fieber
C.
Li
X.
Eriksson
U.
Tissue plasminogen activator is a potent activator of PDGF-CC
EMBO J.
2004
, vol. 
23
 (pg. 
3793
-
3802
)
21
Lei
H.
Velez
G.
Hovland
P.
Hirose
T.
Kazlauskas
A.
Plasmin is the major protease responsible for processing PDGF-C in the vitreous of patients with proliferative vitreoretinopathy
Invest. Ophthalmol. Vis. Sci.
2008
, vol. 
49
 (pg. 
42
-
48
)
22
Kunamneni
A.
Abdelghani
T. T.
Ellaiah
P.
Streptokinase: the drug of choice for thrombolytic therapy
J. Thromb. Thrombolysis
2007
, vol. 
23
 (pg. 
9
-
23
)
23
Takeuchi
T.
Harris
J. L.
Huang
W.
Yan
K. W.
Coughlin
S. R.
Craik
C. S.
Cellular localization of membrane-type serine protease 1 and identification of proteaseactivated receptor-2 and single-chain urokinase-type plasminogen activator as substrates
J. Biol. Chem.
2000
, vol. 
275
 (pg. 
26333
-
26342
)
24
Ustach
C. V.
Kim
H. R.
Platelet-derived growth factor D is activated by urokinase plasminogen activator in prostate carcinoma cells
Mol. Cell. Biol.
2005
, vol. 
25
 (pg. 
6279
-
6288
)
25
Collen
D.
Lijnen
H. R.
Basic and clinical aspects of fibrinolysis and thrombolysis
Blood
1991
, vol. 
78
 (pg. 
3114
-
3124
)
26
Lamba
D.
Bauer
M.
Huber
R.
Fischer
S.
Rudolph
R.
Kohnert
U.
Bode
W.
The 2.3 Å crystal structure of the catalytic domain of recombinant two-chain human tissue-type plasminogen activator
J. Mol. Biol.
1996
, vol. 
258
 (pg. 
117
-
135
)
27
Spraggon
G.
Phillips
C.
Nowak
U. K.
Ponting
C. P.
Saunders
D.
Dobson
C. M.
Stuart
D. I.
Jones
E. Y.
The crystal structure of the catalytic domain of human urokinase-type plasminogen activator
Structure
1995
, vol. 
3
 (pg. 
681
-
691
)
28
Blasi
F.
Carmeliet
P.
uPAR: a versatile signalling orchestrator
Nat. Rev. Mol. Cell Biol.
2002
, vol. 
3
 (pg. 
932
-
943
)
29
Coombs
G. S.
Dang
A. T.
Madison
E. L.
Corey
D. R.
Distinct mechanisms contribute to stringent substrate specificity of tissue-type plasminogen activator
J. Biol. Chem.
1996
, vol. 
271
 (pg. 
4461
-
4467
)
30
Ke
S. H.
Coombs
G. S.
Tachias
K.
Navre
M.
Corey
D. R.
Madison
E. L.
Distinguishing the specificities of closely related proteases. Role of P3 in substrate and inhibitor discrimination between tissue-type plasminogen activator and urokinase
J. Biol. Chem.
1997
, vol. 
272
 (pg. 
16603
-
16609
)
31
Ke
S. H.
Coombs
G. S.
Tachias
K.
Corey
D. R.
Madison
E. L.
Optimal subsite occupancy and design of a selective inhibitor of urokinase
J. Biol. Chem.
1997
, vol. 
272
 (pg. 
20456
-
20462
)
32
Rijken
D. C.
Groeneveld
E.
Substrate specificity of tissue-type and urokinase-type plasminogen activators
Biochem. Biophys. Res. Commun.
1991
, vol. 
174
 (pg. 
432
-
438
)
33
Harris
J. L.
Backes
B. J.
Leonetti
F.
Mahrus
S.
Ellman
J. A.
Craik
C. S.
Rapid and general profiling of protease specificity by using combinatorial fluorogenic substrate libraries
Proc. Natl. Acad. Sci. U.S.A.
2000
, vol. 
97
 (pg. 
7754
-
7759
)
34
Shi
Y. E.
Torri
J.
Yieh
L.
Wellstein
A.
Lippman
M. E.
Dickson
R. B.
Identification and characterization of a novel matrix-degrading protease from hormone-dependent human breast cancer cells
Cancer Res.
1993
, vol. 
53
 (pg. 
1409
-
1415
)
35
Lin
C. Y.
Anders
J.
Johnson
M.
Sang
Q. A.
Dickson
R. B.
Molecular cloning of cDNA for matriptase, a matrix-degrading serine protease with trypsin-like activity
J. Biol. Chem.
1999
, vol. 
274
 (pg. 
18231
-
18236
)
36
Takeuchi
T.
Shuman
M. A.
Craik
C. S.
Reverse biochemistry: use of macromolecular protease inhibitors to dissect complex biological processes and identify a membrane-type serine protease in epithelial cancer and normal tissue
Proc. Natl. Acad. Sci. U.S.A.
1999
, vol. 
96
 (pg. 
11054
-
11061
)
37
Kim
M. G.
Chen
C.
Lyu
M. S.
Cho
E. G.
Park
D.
Kozak
C.
Schwartz
R. H.
Cloning and chromosomal mapping of a gene isolated from thymic stromal cells encoding a new mouse type II membrane serine protease, epithin, containing four LDL receptor modules and two CUB domains
Immunogenetics
1999
, vol. 
49
 (pg. 
420
-
428
)
38
List
K.
Szabo
R.
Molinolo
A.
Nielsen
B. S.
Bugge
T. H.
Delineation of matriptase protein expression by enzymatic gene trapping suggests diverging roles in barrier function, hair formation, and squamous cell carcinogenesis
Am. J. Pathol.
2006
, vol. 
168
 (pg. 
1513
-
1525
)
39
Oberst
M. D.
Singh
B.
Ozdemirli
M.
Dickson
R. B.
Johnson
M. D.
Lin
C. Y.
Characterization of matriptase expression in normal human tissues
J. Histochem. Cytochem.
2003
, vol. 
51
 (pg. 
1017
-
1025
)
40
Tanimoto
H.
Underwood
L. J.
Wang
Y.
Shigemasa
K.
Parmley
T. H.
O'Brien
T. J.
Ovarian tumor cells express a transmembrane serine protease: a potential candidate for early diagnosis and therapeutic intervention
Tumour Biol.
2001
, vol. 
22
 (pg. 
104
-
114
)
41
Lee
S. L.
Dickson
R. B.
Lin
C. Y.
Activation of hepatocyte growth factor and urokinase/plasminogen activator by matriptase, an epithelial membrane serine protease
J. Biol. Chem.
2000
, vol. 
275
 (pg. 
36720
-
36725
)
42
Oberst
M.
Anders
J.
Xie
B.
Singh
B.
Ossandon
M.
Johnson
M.
Dickson
R. B.
Lin
C. Y.
Matriptase and HAI-1 are expressed by normal and malignant epithelial cells in vitro and in vivo
Am. J. Pathol.
2001
, vol. 
158
 (pg. 
1301
-
1311
)
43
Riddick
A. C.
Shukla
C. J.
Pennington
C. J.
Bass
R.
Nuttall
R. K.
Hogan
A.
Sethia
K. K.
Ellis
V.
Collins
A. T.
Maitland
N. J.
, et al. 
Identification of degradome components associated with prostate cancer progression by expression analysis of human prostatic tissues
Br. J. Cancer
2005
, vol. 
92
 (pg. 
2171
-
2180
)
44
Zeng
L.
Cao
J.
Zhang
X.
Expression of serine protease SNC19/matriptase and its inhibitor hepatocyte growth factor activator inhibitor type 1 in normal and malignant tissues of gastrointestinal tract
World J. Gastroenterol.
2005
, vol. 
11
 (pg. 
6202
-
6207
)
45
Tanimoto
H.
Shigemasa
K.
Tian
X.
Gu
L.
Beard
J. B.
Sawasaki
T.
O'Brien
T. J.
Transmembrane serine protease TADG-15 (ST14/Matriptase/MT-SP1): expression and prognostic value in ovarian cancer
Br. J. Cancer
2005
, vol. 
92
 (pg. 
278
-
283
)
46
Chernicky
C. L.
Yi
L.
Tan
H.
Ilan
J.
Tissue-type plasminogen activator is upregulated in metastatic breast cancer cells exposed to insulin-like growth factor-I
Clin. Breast Cancer
2005
, vol. 
6
 (pg. 
340
-
348
)
47
Goh
K. Y.
Poon
W. S.
Chan
D. T.
Ip
C. P.
Tissue plasminogen activator expression in meningiomas and glioblastomas
Clin. Neurol. Neurosurg.
2005
, vol. 
107
 (pg. 
296
-
300
)
48
Laurell
H.
Bouisson
M.
Berthelemy
P.
Rochaix
P.
Dejean
S.
Besse
P.
Susini
C.
Pradayrol
L.
Vaysse
N.
Buscail
L.
Identification of biomarkers of human pancreatic adenocarcinomas by expression profiling and validation with gene expression analysis in endoscopic ultrasound-guided fine needle aspiration samples
World J. Gastroenterol.
2006
, vol. 
12
 (pg. 
3344
-
3351
)
49
Offersen
B. V.
Pfeiffer
P.
Andreasen
P.
Overgaard
J.
Urokinase plasminogen activator and plasminogen activator inhibitor type-1 in nonsmall-cell lung cancer: relation to prognosis and angiogenesis
Lung Cancer
2007
, vol. 
56
 (pg. 
43
-
50
)
50
Pulukuri
S. M.
Estes
N.
Patel
J.
Rao
J. S.
Demethylation-linked activation of urokinase plasminogen activator is involved in progression of prostate cancer
Cancer Res.
2007
, vol. 
67
 (pg. 
930
-
939
)
51
Fisher
J. L.
Field
C. L.
Zhou
H.
Harris
T. L.
Henderson
M. A.
Choong
P. F.
Urokinase plasminogen activator system gene expression is increased in human breast carcinoma and its bone metastases: a comparison of normal breast tissue, non-invasive and invasive carcinoma and osseous metastases
Breast Cancer Res. Treat.
2000
, vol. 
61
 (pg. 
1
-
12
)
52
Jankun
J.
Merrick
H. W.
Goldblatt
P. J.
Expression and localization of elements of the plasminogen activation system in benign breast disease and breast cancers
J. Cell. Biochem.
1993
, vol. 
53
 (pg. 
135
-
144
)
53
Corte
M. D.
Verez
P.
Rodriguez
J. C.
Roibas
A.
Dominguez
M. L.
Lamelas
M. L.
Vazquez
J.
Garcia Muniz
J. L.
Allende
M. T.
Gonzalez
L. O.
, et al. 
Tissue-type plasminogen activator (tPA) in breast cancer: relationship with clinicopathological parameters and prognostic significance
Breast Cancer Res. Treat.
2005
, vol. 
90
 (pg. 
33
-
40
)

Supplementary data