Introduction
Inflammatory breast cancer (IBC) is the most lethal type of breast cancer with a three-year survival rate of 42% as compared with 85% for non-IBC [
1]. In 2010, an international panel of leading experts recommended the clinical consensus for a minimal standard diagnosis of IBC include erythema occupying at least one-third of the breast, hardening and retraction of the nipple, dimpling (peau d'orange) of the skin, and no response to antibiotic treatment [
2]. These clinical characteristics are accompanied by extensive dermal lymphovascular invasion in which tumor emboli are present within dermal lymphatics [
3].
Proteases such as the cysteine protease cathepsin B have been implicated in the initiation, promotion and dissemination of cancers including IBC [
4‐
6]. In IBC, high levels of cathepsin B are found to correlate with increases in numbers of metastatic lymph nodes [
7]. In tumor cells, cathepsin B redistributes into exocytic vesicles at the cell periphery leading to its secretion and association with the tumor cell surface by binding to the light chain of the annexin II heterotetramer [
8,
9]. More specifically, we have shown that in colon cancer cells cathepsin B localizes in caveolae [
10], a membrane microdomain in which the annexin II heterotetramer is also localized [
11]. Downregulation of caveolin-1, the structural protein of caveolae, reduces the cell surface association of cathepsin B and decreases degradation of type IV collagen and invasion by the colon cancer cells, consistent with a functional role for caveolae-associated cathepsin B in invasion [
12].
Caveolin-1 was initially hypothesized to be a tumor suppressor in breast cancer [
13]. More recent data suggest that high expression of caveolin-1 is a characteristic of triple-negative and other basal-like breast cancers [
14], including IBC of a basal phenotype. Indeed, caveolin-1 is highly expressed in both IBC cells and tissues [
7,
15,
16]. We previously hypothesized that the high levels of caveolin-1 expression in bladder, colon, esophageal and prostate cancers promote cell surface proteolytic events that lead to extracellular matrix (ECM) degradation and tumor invasion [
17]. For example, proteases of the plasminogen cascade, specifically pro-urokinase plasminogen activator (pro-uPA) and its receptor uPAR have been localized to caveolae [
12,
18]. These findings may be of functional significance as cathepsin B is capable of processing the zymogen pro-uPA to its active derivative uPA [
19] and is found upstream of plasminogen in a proteolytic pathway on the surface of a number of cell lines [
20‐
22]. Moreover, uPAR complexes with caveolin-1 via β1-integrin, an association that has been shown to mediate uPAR-dependent adhesion and β1-integrin-induced signal transduction [
23,
24]. This suggests that caveolae may serve as sites on the cell surface linking proteolytic and signaling pathways that are involved in tumor invasion.
We hypothesize that participation of cathepsin B in IBC invasion is facilitated by its colocalization at the cell surface with members of the plasminogen cascade and the expression of caveolin-1. Here we demonstrate that cathepsin B as well as uPA and uPAR are associated with caveolar fractions in IBC cells and that cathepsin B is active within these fractions. We confirmed that cathepsin B and caveolin-1 are coexpressed in tissues from IBC patients.
Materials and methods
Cell lines
SUM149 and SUM190 human IBC cell lines [
25] (a kind gift from Dr. Stephen Ethier, Wayne State University, Detroit, MI, USA) were cultured in Hams F-12 media (Mediatech, Manassas, VA, USA) containing 5 μg/ml insulin, 1 μg/ml hydrocortisone, antibiotics (penicillin/streptomycin), and 5% (SUM149) and 2% (SUM190) fetal bovine serum (HyClone, Logan, UT, USA). Media for SUM190 cells were further supplemented with 5 mM ethanolamine, 10 mM HEPES, 5 μg/ml transferrin, 6.6 ng/ml 3,3',5-triiodo-L-thyronine sodium salt, 8.7 ng/ml sodium selenite, and 1 mg/ml bovine serum albumin. All cells were maintained in 5% CO
2/humidified atmosphere at 37°C. Unless otherwise stated all tissue culture reagents were from Sigma Aldrich (St. Louis, MO, USA).
Preparation of cell lysates and conditioned media
IBC cells were grown on plastic to 70% confluency and then serum-starved overnight. Conditioned media were collected and centrifuged at 100 × g at 4°C to remove whole cells, and then re-centrifuged at 800 × g at 4°C to remove cell debris. All conditioned media samples were concentrated to equal volumes in Ultrafree-0.5 PBGC Centrifugal Filter Units with 5 kDa molecular weight cut off Biomax Membranes (Millipore, Billerica, MA, USA). Cells were harvested in lysis buffer (10 mM Tris, pH 7.5, 150 mM NaCl, 1% Triton X-100, 60 mM octylglucoside) and then passed 10 times through a syringe with a 20-gauge needle and centrifuged for five minutes at 10,000 × g at 4°C. Supernatants were recovered and protein concentrations were quantified using micro-BCA reagents according to the manufacturer's instructions (Pierce, Rockford, IL, USA).
Preparation of 3D reconstituted basement membrane overlay cultures
Sixty mm dishes were coated with 300 μl of reconstituted basement membrane (rBM; Cultrex, Trevigen, Gaithersburg, MD, USA) and allowed to solidify for 15 minutes at 37°C. Cells (1 × 106) were seeded on top of solidified rBM and grown in complete media containing 2% rBM. Within 24 hours, cells formed 3D spheroid structures.
Isolation of caveolae-enriched fractions
Non-detergent and detergent-based protocols were used to prepare caveolae-enriched membrane fractions of IBC cells [
26]. In the non-detergent based method, cells grown as a two-dimensional (2D) monolayer on plastic (4 × 100 mm dishes) and three-dimensional (3D) on rBM (4 × 60 mm dish) for two and five days, respectively, were washed with PBS, collected into 500 mM sodium carbonate buffer, pH 11.0, homogenized in a Dounce homogenizer on ice and sonicated three times for 10 seconds each. A discontinuous sucrose gradient was prepared as previously detailed [
26]. Briefly, the cell homogenate was mixed thoroughly with an equal volume of 90% (w/v) sucrose, then overlaid with 35% (w/v) sucrose and 5% (w/v) sucrose and subsequently ultracentrifuged (185,000 ×
g) for 19 hours at 4°C. Fractions of 1 mL were collected and equal-volume aliquots of fractions 3-11 were analyzed by SDS-PAGE and immunoblotting.
A successive detergent-based method of cell fractionation was also employed to separate Triton X-100-soluble (TS) and Triton X-100-insoluble (TI) membrane fractions [
27]. 3D cultures were prepared and grown, as described above, for two days and thereafter washed with cold PBS, incubated with 300 μl of lysis buffer containing 1% Triton X-100 for 20 minutes on ice, collected and centrifuged at 14,000 ×
g for 10 minutes at 4°C. The supernatant (TS fraction) was collected and the pellet was resuspended in 300 μl of lysis buffer containing 1% Triton X-100 plus 60 mM octylglucoside, incubated on ice for 20 minutes, passed through a syringe with a 21.5-gauge needle and centrifuged at 14,000 ×
g for 10 minutes at 4°C. The supernatant (TI fraction) was recovered. Equal-volume aliquots of each fraction were analyzed by SDS-PAGE and immunoblotting.
SDS-PAGE and immunoblotting
Cell lysates and conditioned media samples were equally loaded on the basis of the protein concentration of the respective cell lysates, separated by SDS-PAGE (10 or 12%) under either reducing or non-reducing conditions, transferred to nitrocellulose membranes and immunoblotted with primary antibodies against human cathepsin B (1:4000; [
28]), β1-integrin (1:3000; a kind gift from Dr. Kenneth Yamada, National Institute of Dental and Craniofacial Research, National Institutes of Health, Bethesda, MD, USA), uPA (1:2000; Abcam, Cambridge, MA, USA), uPAR (1:2000; Abcam, Cambridge, MA, USA), caveolin-1 (1:4000; BD Biosciences, Bedford, MA, USA), or β-tubulin (1:1500; prepared from hybridoma E7 cells, Developmental Studies Hybridoma Bank, National Institute of Child Health and Human Development, University of Iowa, Iowa City, IA, USA) and secondary antibodies conjugated with horseradish peroxidase (HRP; 1:10,000; Pierce) in TBS buffer (20 mM Tris, pH 7.5, 0.5 M NaCl) containing 0.5% Tween 20 and 5% (w/v) non-fat dry milk. After washing, bound antibodies were detected by enhanced chemiluminescence according to the manufacturer's instructions (PerkinElmer, Waltham, MA, USA).
Cathepsin B activity assay
Cathepsin B activity was measured using the synthetic fluorometric substrate Z-Arg-Arg-NHMec as previously described [
10]. Briefly, equal-volume aliquots of TS and TI fractions were incubated with activator buffer for 15 minutes at 37°C. Following this activation step, 150 μM Z-Arg-Arg-NHMec (pH 6.0) was added to the assay buffer and fluorescence was measured in triplicate, at one-minute intervals for 30 minutes, at an excitation of 360 nm and an emission of 465 nm. Data are represented as relative fluorescent units and presented as mean ± standard deviation (SD) of three independent experiments. Statistical significance was determined by a two-tailed t-test with assumed equal variance.
Live-cell proteolysis assay
Proteolytic cleavage of DQ-collagen IV substrate (Invitrogen, Carlsbad, CA, USA) by live IBC cells was imaged in real time and quantified as previously described [
29,
30]. Briefly, IBC cells (2.5 × 10
4) were seeded on round glass coverslips coated with rBM containing 25 μg/ml DQ-collagen IV substrate and incubated at 37°C for 40 minutes to allow for cell attachment before adding complete media containing 2% rBM. Following 24 to 48 hours of incubation at 37°C and 5% CO
2, live cells were imaged (37°C and 5% CO
2) with a Zeiss LSM 510 META NLO confocal microscope using a 40× Plan neofluar (N.A., 0.7) objective. DQ-collagen IV cleavage products were observed as green fluorescence. Where specified, the assay was performed in the presence of 10 μM CA074 (Peptides International, Louisville, KY, USA). Statistical significance was determined by a two-tailed t-test with assumed equal variance.
Invasion assay
Cell culture inserts (8.0 μm transparent PET membranes (BD Biosciences, Franklin Lakes, NJ, USA)), were coated with 2 mg/ml rBM and incubated in a 24-well plate at room temperature to permit rBM solidification. SUM149 cells (5.0 × 104) in serum-free media were seeded onto the rBM-coated inserts and incubated for 24 hours at 37°C in the presence of dimethyl sulfoxide (DMSO, vehicle control) or 10 μM CA074. The stimulant for invasion was complete media. Cells that had not invaded were removed with a cotton swab. Cells that had invaded were fixed with 3.7% formaldehyde (Polysciences, Inc., Warrington, PA, USA), stained with 4',6-diamidino-2-phenylindole (Invitrogen, Carlsbad, CA, USA) and imaged at 20× magnification with a Zeiss Axiophot conventional epifluorescent microscope. Ten random microscopic fields per filter were analyzed. The number of cells that invaded was assessed by counting nuclei with MetaMorph™ image analysis software (Molecular Devices, Sunnyvile, CA, USA). Statistical significance was determined by a two-tailed t-test with assumed equal variance.
Immunocytochemical staining
Intracellular co-staining was performed at room temperature on 2D (permeabilized with 0.1% saponin) and 3D cultures (permeabilized with 0.2% Triton X-100). The cultures were incubated with the following primary antibodies diluted in 0.2% Triton X-100/PBS: rabbit anti-caveolin-1 (1:50), mouse anti-uPA (1:25), and mouse anti-uPAR (1:12.5). Secondary antibodies (donkey anti-rabbit Alexa Fluor 488 and donkey anti-mouse Alexa Fluor 555 (Invitrogen, Carlsbad, CA, USA)) were diluted (1:5000) in 0.2% Triton X-100/PBS plus 5% normal donkey serum. All immunocytochemical staining was imaged on a Zeiss LSM 510 META NLO confocal microscope using either a 100× oil Plan neofluar (N.A., 1.3) objective (2D cultures) or 40X Plan apofluar (N.A., 0.7) objective (3D cultures).
Immunohistochemical staining of patient samples
Institutional review board approval from Ain Shams University ethics committee and the National Cancer Institute, Cairo University, along with patient consent forms, were obtained for the purpose of patient enrollment in this study. Inclusion criteria for patients and collection of tissue samples were as previously described [
7]. Pre-treated formalin fixed paraffin embedded IBC (
n = 23) and non-IBC (
n = 27) patient tissue samples were subjected to immunohistochemical (IHC) analysis [
7]. Briefly, tissue sections were incubated for one hour at room temperature with either monoclonal anti-caveolin-1 (1:150) or polyclonal anti-cathepsin B (1:500) primary antibodies. A second incubation was performed with HRP rabbit/mouse (EnVision + Dual Link System-HRP (DAB +)) for 45 minutes. Nuclei were counterstained with hematoxylin, sections were mounted with Permount
® and imaged.
P values for the relation between cathepsin B expressing breast carcinoma cells, of both IBC and non-IBC, and caveolin-1 protein expression were assessed with a chi-square test.
Discussion
IBC is a rare but highly aggressive form of breast cancer with symptoms that develop rapidly (i.e., weeks or months) after initial diagnosis [
38]. Current IBC-specific treatments are very limited and development of new therapeutics is needed, specifically therapies against pathways that mediate the aggressive IBC phenotype. IBC is, however, not amenable to laboratory investigations as there are only two commercially available IBC cell lines (SUM149 and SUM190) for
in vitro investigations and one human xenograft model (MARY-X) [
25,
39]. Our approach in this study was to investigate which proteases expressed by SUM149 and SUM190 IBC cells are associated with caveolin-1, which is highly expressed in IBC
in vivo [
7,
16], participate in ECM degradation and invasion and confirm the presence of these proteases in IBC patient samples. Our findings implicate cathepsin B as one contributor to the aggressive IBC phenotype.
Cathepsin B had previously been shown to activate pro-uPA, a serine protease and member of the plasminogen cascade involved in ECM degradation, matrix metalloproteinase (MMP) activation, and tumor cell invasion [
19]. In SUM149 cells, uPA and its receptor uPAR colocalize with active cathepsin B in caveolae. The presence of active cathepsin B in caveolae of IBC cells suggests a potential role for this enzyme in pericellular proteolysis as was previously shown in colon carcinoma cells. Downregulation of caveolin-1 in the colon carcinoma cells decreases cathepsin B localization to caveolae in parallel with decreases in ECM degradation and cell invasion [
12]. In the SUM149 cells in this study, which co-express cathepsin B and caveolin-1, we determined that degradation of type IV collagen was predominantly pericellular and that a cell impermeant cathepsin B inhibitor reduced their degradation of type IV collagen and invasion. Although significant, the lack of complete inhibition suggests that cathepsin B was only one of several proteases in the SUM149 cells that degrade type IV collagen and mediate invasion. Rao and colleagues have demonstrated that downregulation of cathepsin B and MMP9 more effectively reduces invasion of prostate tumor cells
in vitro and tumor growth
in vivo than downregulation of either cathepsin B or MMP9 [
40]. On the other hand, downregulation of cathepsin B and uPAR more effectively reduces invasion of human glioma cells
in vitro and
in vivo in an intracranial xenograft model than downregulation of either cathepsin B or uPAR [
41]. The above studies indicate that proteases can compensate for one another. There is in addition functional redundancy, including among cysteine cathepsins [
6,
42]. A striking example is the redistribution of active cathepsin × to the surface of mammary tumor cells isolated from mice deficient in cathepsin B that had been crossed with mice predisposed to develop mammary cancer, in this case MMTV-PyMT transgenic female mice. Moreover, cathepsin × neutralizing antibodies reduced invasion of the cathepsin B-deficient mammary tumor cells, a result that is consistent with cathepsin × compensating for the absence of cathepsin B [
43].
The studies above illustrate that
in vitro assays such as invasion and ECM degradation assays are meaningful surrogates for
in vivo tumor endpoints. In a recent collaborative study (N Withana, BF Sloane and BS Parker, unpublished observations), we demonstrated that either knockdown or inhibition of cathepsin B in 4T1 mammary carcinoma cells reduced collagen degradation
in vitro, as assessed by our live-cell proteolysis assay, and bone metastasis
in vivo. Here using this live-cell proteolysis assay, we observed differences in the degradation of type IV collagen by the two IBC cell lines. These differences were further supported by variations in expression and secretion of proteases and known caveolae-associated proteins. We propose that these differences may be due, in part, to the differences in receptor status of the two IBC cell lines. Although both lack estrogen and progesterone receptors, only SUM149 is HER2 negative, classifying it as a triple-negative breast cancer cell line. Triple-negative breast cancer is a subtype of breast cancer characterized as extremely aggressive, having a poor prognosis and difficult to treat with high risks of both recurrence and death [
44,
45]. A role for cathepsin B has previously been reported in several triple-negative breast cancer cell lines that are not IBC (e.g., BT20, BT549, MCF-10AneoT, and MDA-MB-231) [
31,
46‐
48], and inhibition of active cathepsin B in these cells reduced their invasion
in vitro.
In vivo studies also provide evidence for an association between cathepsin B and breast cancers that are triple negative [
49] (N Withana, BF Sloane and BS Parker, unpublished observations). In addition, a transgenic mouse model of mammary cancer characterized by loss of hormone receptors with progression of disease [
50] exhibited both reduced primary tumor growth and lung metastases when deficient in cathepsin B [
51], suggesting a link between cathepsin B expression and/or activity and invasive and metastatic breast disease. This may be especially true in IBC as expression of cathepsin B was found to be positively correlated with lymph node metastasis in IBC tissues, a correlation not observed in non-IBC tissues [
7]. As such, cathepsin B has been proposed to be a prognostic marker for IBC and potentially a component of a proposed molecular signature for IBC that already includes caveolin-1 [
16]. Our current findings show that cathepsin B and caveolin-1 were co-expressed in tumor cells of IBC patient samples and not in those of non-IBC patients. Ongoing work will demonstrate if this co-expression is elevated in triple-negative breast cancer patients other than those diagnosed with IBC.
Another enzyme implicated in the aggressive IBC phenotype is RhoC GTPase, which is increased in expression and activity [
52,
53]. We speculate that there may be a network that links cathepsin B, caveolin-1, and Rho signaling pathways in IBC. In colon, prostate, and non-IBC tumors, phosphorylated caveolin-1 has been shown to promote migration and invasion via a Rho signaling pathway [
54]; this has not yet been assessed in IBC. A P132L mutation in caveolin-1 confers a dominant-negative effect on invasiveness of human schirrhous breast cancers [
55] and upregulates genes involved in invasiveness and metastasis, including Rho-related signaling molecules and genes expressed by stem cells [
56]. Studies in MDA-MB-231 breast carcinoma cells by Bourguignon and colleagues [
57] connect Rho, caveolin-1, and cathepsin B. Rho kinase signaling events, mediated upstream by CD44-NHE1 interactions localized to lipid microdomains containing caveolin-1, result in acidification of the microenvironment surrounding breast cancer cells, activate secretion of cathepsin B and promote cellular invasiveness. We have previously shown that slight acidification of the microenvironment of a variety of tumors (melanoma, colon, and breast) increases secretion and activity of cathepsin B and proteolysis of type IV collagen [
8,
58,
59]. Whether there is a universal link between Rho, caveolin-1, cathepsin B, and acidification of the tumor microenvironment has not yet been evaluated.
Acknowledgements
We thank Dr. Stephen P. Ethier for the kind gift and morphological authentication of the SUM149 and 190 cell lines. We also thank our Egyptian colleagues: breast surgeon, Dr. Mohamed El-Shinawi, Faculty of Medicine, Ain Shams University, Cairo, Egypt, and pathologist, Dr. Mohamed A. Nouh, National Cancer Institute, Cairo University, Giza, Egypt for patient recruitment and clinical diagnoses.
This work was supported by the National Institutes of Health [R01 CA131990 (BFS), R03 TW008624 (BFS & MMM)], AVON Foundation [02-2007-049 (BFS & MMM)] and Department of Defense [BC083400 (BCV)].
Imaging was performed in the Microscopy, Imaging and Cytometry Resources Core supported, in part, by NIH Center grant P30CA22453 to the Karmanos Cancer Institute, Wayne State University and by the Perinatology Research Branch of the National Institutes of Child Health and Development, Wayne State University.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
BCV, AA, MMM and DCM carried out the experiments. BCV, MMM, BFS and DCM made substantial contributions to concept and design of experiments as well as drafting and/or revising the manuscript. All authors have read and approved the manuscript.