SEARCH

SEARCH BY CITATION

Keywords:

  • proton pump;
  • vesicle transport;
  • exocytosis;
  • Rab27;
  • heat shock protein;
  • invasion

Abstract

  1. Top of page
  2. Abstract
  3. Material and Methods
  4. Results
  5. Discussion
  6. Conclusions
  7. Acknowledgements
  8. Author contributions
  9. References
  10. Supporting Information

The secretory Rab27B small GTPase promotes invasive growth and metastasis in estrogen receptor (ER) α-positive breast cancer cells by orchestrating the peripheral targeting of vesicles secreting proinvasive growth regulators. Increased Rab27B expression is associated with poor prognosis in breast cancer patients. The molecular mechanisms of peripheral Rab27B secretory vesicle distribution are poorly understood. Mass spectrometry analysis on green fluorescent protein (GFP)-Rab27B vesicles prepared from GFP-Rab27B transfected MCF-7 human breast cancer cells detected eight subunits of the vacuolar H(+)-ATPase (V-ATPase) and the presence of V0a1 and V0d1 subunits was confirmed by Western blot analysis. Reversible inhibition of V-ATPase activity by bafilomycin A1 or transient silencing of V0a1 or V0d1 subunits demonstrated that V-ATPase controls peripheral localization and size of Rab27B vesicles. V-ATPase expression and activity further controls Rab27B-induced collagen type I invasion, cell-cycle progression and invasive growth in the chorioallantoic membrane assay. In agreement, Rab27B-dependent extracellular heat shock protein90α release and matrix metalloprotease-2 activation is markedly reduced by bafilomycin A1 and transient silencing of V0a1 and V0d1 subunits. Poor prognosis ERα-positive primary breast tumors expressing high levels of Rab27B also expressed multiple V-ATPase subunits and showed a strong cytoplasmic and peripheral V-ATPase V1E expression. In conclusion, inhibiting V-ATPase activity by interfering agents and drugs might be an effective strategy for blocking Rab27B-dependent proinvasive secretory vesicle trafficking in ERα-positive breast cancer patients.

Abbreviations
CAM

chorioallantoic membrane

CM

conditioned medium

ER

estrogen receptor

GFP

green fluorescent protein

H&E

hematoxylin and eosin

HSP

heat shock protein

MMP

matrix metalloprotease

PBS

phosphate-buffered saline

PPI

proton pump inhibitor

V-ATPase

vacuolar H+ ATPase

Regulated exocytosis of secretory vesicles may offer cancer cells a selective advantage for invasive growth. Secretory vesicle trafficking is regulated by the combined actions of actin filaments and microtubules using the molecular motor proteins dyneins, kinesins and/or members of the myosin family and Rab GTPases as critical regulators of vesicle transport.[1] While researching the functional significance of the Rab27 family in the regulation of the exocytotic pathway, we demonstrated that overexpression of Rab27B in estrogen receptor (ER) α-positive breast cancer cell lines such as MCF-7-promoted invasive growth and metastasis through secretion of the molecular chaperone heat shock protein (HSP)90α.[2] Extracellular HSP90α regulates a clientele of receptors, growth factors and proteases favoring the balance toward invasive growth.[3] One reported client protein is matrix metalloprotease (MMP)−2, a protease that requires its association with extracellular HSP90α for activity.[4] Clinical significance of our findings demonstrated that upregulation of Rab27B transcripts and protein correlated with lymph node metastasis, tumor grade and positive ERα-status in breast cancer.[2] Currently, the molecular mechanisms of targeting and regulation of Rab27B- and HSP90α-containing vesicles remain poorly understood.

Vacuolar H+ ATPase (V-ATPase) is a multisubunit ATP-dependent proton pump composed of a membrane sector V0 and a cytosolic catalytic sector V1.[5] The integral V0 component functions in proton translocation, whereas the peripheral V1 component hydrolyzes ATP. The V1 domain consist of a A3B3 subunit hexamer flanked by peripheral stalks consisting of E and G subunits (three in most eukaryotic cells); and one of each C, D, F and H subunits.[6] The V0 domain consists of six different subunits (a, c, c′, c″, d and e). The V1 and V0 domains are connected by a central stalk composed of subunits V1D, V1F and V0d. V-ATPases are mainly located in membranes of various subcellular compartments, such as endosomes and lysosomes, and at the plasma membrane of metastatic MDA-MB-231 breast cancer cells.[7] A recent report demonstrated that V-ATPases play an important role in secretory-granule maturation.[8] Genetic evidence demonstrated V-ATPase V0a-subunit-dependent exocytosis of synaptic vesicles in Drosophila[9] and exosomes in Caenorhabditis elegans.[10] Functional studies indicate that Matrigel invasion by MDA-MB-231 breast cancer cells is disrupted by molecular targeting and inhibition of V-ATPases by siRNA, bafilomycin A1 and concanamycin A.[7, 11]

In our study, we use MCF-7 cells expressing green fluorescent protein (GFP)-Rab27 fusion protein as a proper model of Rab27-induced breast cancer invasion.[2] We identified multiple V-ATPase V1 and V0 subunits in purified Rab27B-positive vesicles. Pharmacologic inhibition of V-ATPase activity and knockdown of V-ATPase subunit V0a1 or V0d1 expression reversibly inhibited Rab27B-induced cancer cell invasion and cell-cycle progression. This was accomplished through dysregulation of the peripheral localization of secretory Rab27B vesicles, HSP90α secretion and activation of MMP-2. We further examined the expression of V-ATPase subunit V1E in primary breast tumors previously analyzed for Rab27B expression. Our data provide the first insights into the peripheral trafficking of Rab27B vesicles and determination of the invasive phenotype by V-ATPases in ERα-positive breast cancer.

Material and Methods

  1. Top of page
  2. Abstract
  3. Material and Methods
  4. Results
  5. Discussion
  6. Conclusions
  7. Acknowledgements
  8. Author contributions
  9. References
  10. Supporting Information

Cell lines, expression vectors, conditioned medium and transfections

The ERα-positive, noninvasive and nonmetastatic human breast cancer cell lines MCF-7, T47D and ZR75.1 (ATCC, Manassas, VA) were maintained in culture medium containing Dulbecco's Minimal Essential Medium supplemented with 10% fetal bovine serum, 100 U/mL penicillin and 100 µg/mL streptomycin (Invitrogen, Carlsbad, CA). MCF-7 cells expressing GFP and GFP-Rab fusion proteins, and the preparation of serum-free conditioned medium (CM) were described previously.[2] MCF-7 cells expressing Snail-6SA are the constitutive epithelial-to-mesenchymal transition (EMT) model established by constitutive activation of the Snail transcription factor, with replacement of six Ser to Ala in the GSK3β-phosphorylation consensus motif.[12] V-ATPase-specific HiPerformance guaranteed siRNAs (V-ATPase V0a1 siRNA: SI04325937: CTGGTTCATCAAAGTGCGGAA and SI00307279: AAGCTTCGATTTGTTGAGAAA; V-ATPase V0d1 siRNA: SI04189871: CTGGCTCGGGCTGACGACTAT and SI04229302: TTACAGTTACATGATCGACAA) and a scrambled siRNA-negative control were purchased from Qiagen (Venlo, The Netherlands). siRNA transfections were performed by electroporation using the Cell Line Nucleofector Kit V (Lonza, Cologne, Germany).

Antibodies and reagents

The following primary antibodies were used: mouse monoclonal anti-GFP (1:1,000) (MAB3580; Millipore, Billerica, MA), mouse monoclonal antitubulin (1:1,000) (T5168; Sigma-Aldrich, St. Louis, MO), rabbit polyclonal anti-Rab27B (1:1,000),[13] mouse monoclonal anticyclin E (1:500) (AHF0312; Invitrogen), mouse monoclonal anticyclin A (1:250) (33–4,900; Zymed Laboratories, San Francisco, CA), rabbit polyclonal anti-p27 (1:1,000) (sc-527; Santa Cruz Biotechnology, Santa Cruz, CA), rabbit polyclonal anti-HSP90α and β (1:1,000) (PA3-012, PA3-013; Affinity Bioreagents, Golden, CO), mouse monoclonal anti-V-ATPase V0d1 (1:1,000) (sc-81887, Santa Cruz Biotechnology), rabbit polyclonal anti-ATP6E (1:100) (sc-20946, Santa Cruz Biotechnology), rabbit polyclonal anti-V-ATPase V0a1 (1:1,000) (HPA022144, Atlas Antibodies, Stockholm, Sweden). Secondary antibodies coupled to horseradish peroxidase, Alexa-444, Alexa-555 or biotin were obtained from Amersham Pharmacia Biotech (Diegem, Belgium) or Invitrogen. The nuclear stain, 4′,6-diamidino-2-phenylindole, and a filamentous actin stain, phalloidin–tetramethyl rhodamine isothiocyanate (TRITC), were purchased from Sigma-Aldrich. V-ATPase inhibitors bafilomycin A1 and concanamycin A were purchased from Sigma-Aldrich. Recombinant proMMP-2 protein was obtained from R&D systems (Minneapolis, MN).

Invasion assay

Type I collagen gel was prepared as described previously.[14] Single cells (2 × 105) suspended in 1-mL culture medium were seeded on top of the type I collagen gel under control, V-ATPase inhibitor (bafilomycin A1 or concanamycin A treatment: 0.1, 1 or 5 nM for 3 or 24-hr), V-ATPase inhibitor rescue (1 nM V-ATPase inhibitor treatment for 24-hr followed by three washing steps with culture medium and 24-hr recuperation period) or siRNA targeting (7-hr after transfection) conditions. The collagen gels were incubated on a flat surface in a humidified atmosphere of 10% CO2 in air at 37°C. Cell morphology was studied and invasion was scored after 24-hr.[14] The number of invasive and noninvasive cells was counted in ten randomly selected microscopic fields with a 20× objective and a 10× eye piece by two blinded observers using an inverted phase contrast microscope (DMI 3000B, Leica, Wetzlar, Germany). The invasion index was calculated as the ratio of the number of cells that invaded the gel divided by the number of noninvasive cells counted in each field. Trypan blue exclusion was used to determine possible toxic effects of bafilomycin A1 or concanamycin A treatment.

Factor shape and GFP-Rab27B vesicle quantification

Collagen matrices were fixed in 3% paraformaldehyde for 10 min, phalloidin–TRITC stained and imaged by confocal microscopy as described previously.[14] Factor shape and GFP-Rab27B fluorescence intensity were analyzed with Adobe Photoshop. The factor shape refers to a value that is affected by an object's shape but is independent of its dimensions. It was calculated as (perimeter)2/(4π × area), which describes the deviation of an object from a geometric circle. It gives a minimal value of 1 for a perfect circle and larger values for shapes having a higher ratio of perimeter to area. To analyze the GFP-Rab27B vesicle distribution, the total area was outlined with the “magnetic lasso” tool using Adobe Photoshop software. The cell periphery was defined as the outermost 25% of the cell area of cells spread on collagen type I matrix using the “refine edges” tool. The total pixels in 15 single cells and the signals in the peripheral part of the cell were measured, and the percentage of peripheral GFP-Rab27B vesicles, that is, vesicles present in the outermost 25% of the cell area was calculated. The surface area of GFP-Rab27B vesicles was analyzed in 15 single cells using Image J software.

Immunoelectron microscopy

MCF-7 GFP-Rab27B cells under control, bafilomycin A1 (1 nM, 24-hr treatment) or rescue (1 nM treatment for 24-hr followed by three washing steps with culture medium and 24-hr recuperation period) conditions were fixed with 4% paraformaldehyde in 0.1 M phosphate buffer at room temperature for 1-hr, scraped in 1% gelatin, centrifuged, cut in 1 mm2 blocks and immersed in 2.3 M sucrose in Dulbecco's phosphate-buffered saline (PBS) at 4°C. Thin cryosections (70 nm) were cut with a Leica EM UCT cryo ultramicrotome. For immunolabeling, the sections were first incubated in a blocking serum containing 5% BSA with 0.1% gelatin in PBS. Antibodies and gold conjugates were diluted in 0.1% BSA-C in PBS (Aurion, Wageningen, The Netherlands). After blocking as described above, the sections were exposed to the primary antibody for 60 min, followed by exposure to a rabbit antimouse IgG (Zymed) for 30 min and protein A-gold complex (size, 10 nm) for 30 min.[15] The efficiency of blocking was controlled by performing the labeling procedure in the absence of the second primary antibody. The sections were counterstained with methylcellulose/uranyl acetate and examined in a Philips CM100 transmission electron microscope (FEI, Eindhoven, The Netherlands). Images were captured by morada CCD camera (Olympus Soft Imaging Solutions GMBH, Munster, Germany). The average distance of GFP-Rab27B vesicles from the periphery was measured using Adobe Photoshop software.

Protein analysis

Western blot analysis was performed as described previously.[2] For gelatin zymography, CM was prepared from MCF-7 GFP-Rab27B cells under control, bafilomycin A1 (5 nM, 24-hr treatment) or rescue (1 nM treatment for 24-hr followed by three washing steps with culture medium and 24-hr recuperation period) conditions. CM (20 µL) was resuspended in 10 µL nonreducing sample buffer (0.5 M Tris-HCl, pH = 6.8, 20% glycerol, 4% SDS and 0.005% bromophenol blue) without boiling. Samples were loaded on Novex 10% zymogram gelatin substrate gels (Invitrogen). After electrophoresis, gels were washed twice for 30 min in a 2% Triton X-100 (Bio-Rad, Hercules, CA) water solution at room temperature and incubated for overnight at 37°C in MMP substrate buffer (50 mM Tris-HCl, pH 7.5, 10 mM CaCl2). Gels were rinsed again in distilled water and stained with Coomassie Brilliant Blue. Proteolytic activities appeared as clear bands of lysis against a dark background of stained gelatin.

Flow cytometric cell-cycle analysis and cell proliferation assay

The analysis of cell-cycle distribution is described in detail in Ref. [2]. To examine whether Rab27B affects cell proliferation in a V-ATPase-dependent manner, proliferation rates of MCF-7 GFP or MCF-7 GFP-Rab27B cells transiently targeted with control or V-ATPase V0a1 or V0d1 siRNAs were studied. The total number of viable cells in each well was counted with a Countess automated cell counter (Invitrogen) every day for 4 days. Trypan blue staining was used for exclusion of dead cells.

Chick embryo studies

Fertilized eggs from a local commercial hatchery were incubated in an automatic incubator at a temperature of 37.8°C and 50% humidity. At day 3, the chorioallantoic membrane (CAM) was dropped by removal of 2–3 mL albumen and the eggs were sealed with a semi-permeable film. Normal development was verified and embryos with malformations or dead embryos were excluded. At day 9, 106 MCF-7 GFP-Rab27B cells transfected with control siRNA or siV-ATPase V0d1 were suspended in 100 µL Matrigel and deposited (in five drops of 20 µL) after gentle laceration on the CAM surface. The shell window was resealed. Five CAMs were inoculated for each condition. The CAM was harvested at day 14 of incubation. The membrane was fixed by application of 300 µL buffered formalin on the intact CAM and photographed with a Leica M205 FA20 fluorescence stereomicroscope. The total tumor area and the number of metastatic nodi were calculated using Image J software. Tumors were extracted and lysed to analyze cyclin A by Western blotting, or fixed in 4% buffered formalin for 12-hr, followed by a wash with PBS and transfer to 70% ethanol, and then embedded in paraffin, sectioned and stained with hematoxylin and eosin (H&E) or TUNEL (Roche Applied Science, Indianapolis, IN).

GFP-Rab27B vesicle isolation

MCF-7 GFP or GFP-Rab27B cells (2 × 108 cells) were trypsinized and resuspended in culture medium. The cell suspension was centrifuged for 10 min at 500g, followed by three washes with 5-mL PBSD+. The cell pellet was resuspended in 1 mL homogenization solution (250 mM sucrose in PBSD+ supplemented with protease inhibitor cocktail [Roche Applied Science, Indianapolis, IN]). Cells were homogenized on ice via sonication. Different centrifugations were performed using a 70.1 Ti rotor Beckman Coulter centrifuge (Beckman Coulter, Fullerton, CA): low speed centrifugation at 3,000g for 10 min at 4°C, followed by high-speed centrifugation at 30,000g for 60 min at 4°C. Next, the supernatant was incubated at a 1:1 ratio (v/v) with anti-GFP-labeled magnetic microbeads suspended in homogenization solution (50 µL microbeads/10 × 106 cells) (MACS MicroBeads, Miltenyi Biotec, Auburn, CA) for 30 min on ice. Total samples (2 mL) were loaded on the automated MACS separator (Miltenyi Biotec). Vesicles were eluted in elution buffer (Miltenyi Biotec). After elution, homogenization buffer was added in a ratio of 1:1 v/v. The purity of the vesicle fraction was checked before and after magnetic separation via flow cytometry (Calibur, Becton Dickinson, Franklin Lakes, NJ). Vesicles were pelleted by centrifugation at 140,000g for 1-hr.[2]

Patient samples and immunohistochemistry

Clinical data and primary breast carcinoma samples were collected as described previously.[2] Lysates were prepared from microdissected tumor tissue. Briefly, one frozen section stained with H&E was mounted with a coverslip, and the remaining adjacent serial sections were left without a coverslip for tissue removal. Using the covered slide stained with H&E as the template, areas containing stroma and accumulated collagen were removed. The remaining epithelial tissue, obtained from a minimum of ten sections, was lysed and analyzed by Western blotting. Sixty-three primary breast tumors characterized for Rab27B protein signal (described in Ref. [2]) were evaluated for V-ATPase V1E expression. Sections (4 µm) were cut from paraffin-embedded tissue blocks. Sections were fixed in neutral buffered 4% formaldehyde and deparaffinized with EZ PrepTM Concentrate (Ventana Medical Systems, Tucson, AZ). Slides were loaded in the Ventana autostainer (Benchmark XT, Ventana Medical Systems) and stained with the Ultraview Universal DAB detection system (Ventana Medical Systems) according to the manufacturer's instructions. Antigen retrieval was performed using EDTA pretreatment (Ventana Medical Systems). Sections were then incubated with ATP6E polyclonal antibody diluted at 1/100, during 32 min at 37°C. Nine samples showed no primary breast cancer tissue and were therefore excluded for analysis. Two observers quantified independently.

Statistical analysis

Comparisons were performed using a two-sided unpaired Student's t-test after D'Agostino–Pearson testing for normal distribution, χ2-test, Mann–Whitney test or two-way repeated measures analysis of variance test. Frequency tables of immunohistochemistry data were analyzed by the χ2 test. All data presented are representative of at least three independent experiments. All statistical tests were two-sided. P-values of <0.05 were considered to be statistically significant.

Results

  1. Top of page
  2. Abstract
  3. Material and Methods
  4. Results
  5. Discussion
  6. Conclusions
  7. Acknowledgements
  8. Author contributions
  9. References
  10. Supporting Information

Identification of V-ATPase proton pump subunits in purified Rab27B-positive vesicles

We previously reported mass spectrometry analysis on 97% pure GFP-Rab27B vesicles isolated from MCF-7 cells stably expressing GFP-Rab27B.[2] These data revealed the presence of eight subunits of the V-ATPase catalytic V1 domain (B2, E1, G1 and H) and membrane-bound V0 domain (a1, a4, c and d1) (Fig. 1a). Western blot analysis confirmed the presence of V0a1 and V0d1 subunits in purified GFP-Rab27B vesicles (Fig. 1b). Similar observations were made in MCF-7 cells overexpressing a constitutively active form of this small GTPase (GFP-Rab27B Q78L) (data not shown). Overexpression of GFP-Rab27B stimulated V-ATPase V0a1 and V0d1 expression (Supporting Information Fig. 1), suggesting that increased secretory GTPase activity induces expression of components necessary to support and steer GTPase-regulated vesicle transport.

image

Figure 1. Identification of V-ATPase proton pump subunits in purified GFP-Rab27B vesicles. (a) Proteomic profiling of GFP-Rab27B vesicles purified from invasive MCF-7 GFP-Rab27B breast cancer cells identified eight V-ATPase subunits.[2] The number of peptides identified and percentage sequence coverage are indicated for each subunit. (b) Western blot analysis for GFP (30 kDa is GFP and 57 kDa is GFP-Rab27B) and V-ATPase V0a1 (116 kDa) and V0d1 (40 kDa) after GFP immunoprecipitation of vesicle preparations of MCF-7 GFP and GFP-Rab27B cells (detailed experimental conditions in Ref. [2]). Asterisks indicate heavy- and light-antibody chain.

Download figure to PowerPoint

The size and peripheral distribution of Rab27B vesicles requires V-ATPase expression and activity

Pharmacologic inhibition of the catalytic activity and RNA interference were used to perform loss-of-function experiments. MCF-7 cells in which GFP-Rab27B was stably expressed exhibited cellular extensions and spread morphology on a type I collagen matrix, as shown in the control condition (Fig. 2a). In these cells, GFP-Rab27B displayed a punctuate and vesicular distribution with accumulation in the cellular extensions (Fig. 2a). More than 70% of MCF-7 GFP-Rab27B cells had a peripheral distribution of GFP-Rab27B vesicles, that is vesicles present in the outermost 25% of the cell area (Fig. 2c). After 3-hr bafilomycin A1 treatment (1 nM), GFP-Rab27B vesicles were redistributed from the cell periphery toward the cytoplasm without a major impact on cell morphology and vesicle surface area (Figs. 2a2c). Upon 24-hr bafilomycin A1 treatment (1 nM), the shape of MCF-7 Rab27B cells was markedly modified to a round morphology (Fig. 2a). Concomitantly, redistributed GFP-Rab27B vesicles clustered with a time-dependent increase in vesicle surface area (Figs. 2b and 2c). A rescue experiment, in which bafilomycin A1 (1 nM) was rinsed out after a 24-hr treatment followed by a 24-hr recuperation period, showed a reorganization of the vesicle surface area and restoration of the number of MCF-7 GFP-Rab27B cells with a peripheral distribution of the GFP-Rab27B vesicles (Figs. 2a2c). Immunoelectron microscopy was used to further explore the cellular distribution of GFP-Rab27B vesicles (Fig. 2d). Staining for GFP-Rab27B (10 nm gold) was prevalent at the plasma membrane under control and rescue conditions. After 24-hr treatment of bafilomycin A1 (1 nM), GFP-Rab27B was almost absent from the plasma membrane and showed a cytoplasmic accumulation. Transient and selective targeting of V0a1 or V0d1 subunits by siRNA depleted these subunits by 70–80%, respectively, but did not affect GFP-Rab27B expression as assessed by Western blotting. It is worth noting that V0d1 subunit knockdown markedly affected protein expression levels of V0a1 subunit, suggesting an essential role for V0d1 central stalk subunit in the stability and assembly of the V-ATPase multimolecular complex (Fig. 2e). Although silencing of the V0a1 and V0d1 subunits did not affect total GFP-Rab27B expression levels, it markedly affected the number of MCF-7 GFP-Rab27B cells with peripheral GFP-Rab27B distribution (Fig. 2f).

image

Figure 2. Effect of pharmacological inhibition and siRNA targeting of V-ATPase on surface area, number and peripheral localization of GFP-Rab27B vesicles. (a–c) MCF-7 GFP-Rab27B cells were seeded on collagen type I matrix under control, V-ATPase inhibitor (1 nM for 3 or 24 hr) or rescue (1 nM bafilomycin A1 treatment for 24-hr followed by three washing steps and 24-hr recuperation period) conditions. (a) Laser scanning confocal images of GFP-Rab27B vesicle size and distribution in MCF-7 cells (as exemplified by arrows). Scale bar, 20 µm. (b and c) Box-and-whisker plots of changes in surface area (b) and % of MCF-7 GFP-Rab27B cells with peripheral localization (c). Fifteen individual cells (typically from one to two cells per image) from two independent stainings per treatment were analyzed using Image J and Adobe Photoshop software. (d) Electron micrographs of ultrathin cryosections of MCF-7 GFP-Rab27B cells, grown on plastic substrate under control, bafilomycin A1 (1 nM, 24-hr) or rescue conditions. Rab27B vesicles were immunogold-labeled with anti-GFP antibodies (10 nm of gold). Insets zoom in on the cell periphery with the plasma membrane at the right position. Scale bars, 1 µm. Right corner, measurement of the distance from the plasma membrane of individual GFP-Rab27B-positive vesicles on randomly taken electron micrographs. (e) Impact of V-ATPase siRNAs on V-ATPase V0a1 and V0d1 subunit expression in MCF-7 GFP-Rab27B cells by immunoblotting. (f) The percentage of MCF-7 GFP-Rab27B cells exhibiting a peripheral localization of GFP-Rab27B in which V-ATPase V0a1 or V0d1 subunits were depleted. In (b), (c), (d) and (f) the median of the distribution is indicated by the horizontal line in the box. p-Values are calculated using Mann–Whitney test. Statistically significant p-values are indicated. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

Download figure to PowerPoint

Functional implication of the V-ATPase proton pump in Rab27B-induced breast cancer cell invasion and cell-cycle progression

To evaluate the role of V-ATPase-dependent Rab27B trafficking in collagen type I invasion, we analyzed matrix invasion in the presence and absence of bafilomycin A1. A 24-hr bafilomycin A1 (0.1, 1 and 5 nM) treatment, abrogated matrix invasion of the highly invasive MCF-7 cells overexpressing wild type or constitutively active (Q78L) GFP-Rab27B (Fig. 3a and Supporting Information Fig. 2) in a dose-dependent manner. Similar data were obtained with concanamycin A (a macrolide antibiotic with similar structure and binding specificity as bafilomycin A1) or with other ERα-positive breast cancer cell lines overexpressing wild-type GFP-Rab27B such as T47D and ZR75.1 (Supporting Information Fig. 3 and data not shown). This inhibitory effect showed a specificity for Rab27B-stimulated invasion as bafilomycin A1 (1 and 5 nM for 24-hr) had no effect on the EMT-dependent invasion of MCF-7 Snail-6SA cells (Supporting Information Fig. 4). The factor shape, calculated using (perimeter)2/(4π × area) was 4.1 ± 0.3. Bafilomycin A1 treatment induced a rounded appearance in MCF-7 GFP-Rab27B cells (factor shape, 1.9 ± 0.2). Importantly, despite the major changes in cellular invasion and morphology, drug removal after 24-hr bafilomycin A1 treatment (1 nM) rescued the elongated cell morphology (factor shape, 3.6 ± 0.4) and allowed invasion into native collagen type I matrices (Fig. 3a). Trypan blue exclusion assay revealed no increased levels of dead cells after 24-hr treatment with bafilomycin A1 or concanamycin A. Transient and selective silencing of V0a1 or V0d1 subunits by siRNA was accompanied by loss of the elongated cell morphology (factor shape values, 2.5 ± 0.3 and 2.1 ± 0.4, respectively) and invasive potential (Fig. 3b). Transient silencing of V0a1 or V0d1 subunits showed no effect on invasion of EMT-induced MCF-7 Snail-6SA breast cancer cells (Supporting Information Fig. 4).

image

Figure 3. Involvement of V-ATPase in matrix invasion and cell-cycle progression of GFP-Rab27B expressing breast cancer cells. (a) MCF-7 GFP-Rab27B cells were seeded on collagen type I gels and reversibly treated with V-ATPase inhibitor bafilomycin A1 (0.1, 1 or 5 nM for 24-hr). Upper part, phase contrast images showing morphology on collagen type I gels of MCF-7 GFP-Rab27B cells (bafilomycin A1: 1 nM, 24-hr). Scale bar, 50 µm. Bottom part, quantification of collagen type I invasion. (b–f) MCF-7 GFP-Rab27B cells were transiently transfected with control siRNA (siCON) or siV-ATPase V0a1 or V0d1 siRNAs. (b) Upper part, phase contrast images showing morphology of MCF-7 GFP-Rab27B cells on collagen type I. Scale bar, 50 µm. Bottom part, quantification of collagen type I invasion. (c) MCF-7 GFP-Rab27B cells were grown to 50% confluence, followed by 24-hr serum starvation, and 24-hr serum-induced (0.5%) cell-cycle progression. Percentages of cells in G1, S and G2 stage of the cell cycle, as measured by flow cytometry, are represented as the means with upper 95% confidence intervals of two independent experiments. (d) Western blot analysis of MCF-7 GFP-Rab27B cells for the positive (cyclin A and E) and negative (p27) G1 to S-phase cell-cycle regulators. Protein levels were quantified as immunostaining intensity relative to tubulin using Image J software. (e and f) Measurement of cell proliferation rates of MCF-7 GFP-Rab27B (e) or GFP (f) cells. A total of 104 cells were plated into each well of a total of 12 wells on day 1 to establish one growth curve under each condition in triplicate. The total number of cells per well was automatically counted every day until day 4. In (a) and (b), arrows indicate cellular extensions and local spreading. Data are presented as the mean ± s.e.m. of three independent experiments performed in triplicate. All data are compared to the control (CON or siCON). Statistically significant p-values, calculated using the χ2-test, are indicated (*p < 0.0001). In (e) and (f), mean number of cells is plotted with upper 95% confidence intervals. Data are compared to the control (siCON). Statistically significant p-values, calculated using the two-way repeated measures analysis of variance test, are indicated.

Download figure to PowerPoint

image

Figure 4. Effect of siRNA targeting of V-ATPase on invasive growth of GFP-Rab27B expressing breast cancer cells. (a) GFP-fluorescence images of MCF-7 GFP-Rab27B cells transfected by siCON or siV-ATPase V0d1 on the chorioallantoic membrane (CAM) 5 days after inoculation. Arrows show secondary nodi spreading on the CAM surface. Scale bar is 1 mm. (b and c) Box-and-whisker plots showing quantification of tumor size and secondary nodule formation of MCF-7 GFP-Rab27B cells transfected by siCON or siV-ATPase V0d1. The median of the distribution is indicated by the horizontal line in the box. Statistically significant p-values, calculated using Mann–Whitney test, are indicated. (d) Histological hematoxylin and eosin-analysis of tumors from MCF-7 GFP-Rab27B cells transfected by siCON or siV-ATPase V0d1. Arrows indicate the presence of secondary nodi at a distance from the primary tumor. The presence of nucleated chick erythrocytes capillaries is indicated by arrowheads. Scale bar, 200 µm. (e and f) TUNEL staining on paraffin-embedded MCF-7 GFP-Rab27B siCON or siV-ATPase V0d1 primary tumors (e) and cyclin A Western blot detection on tumor lysates (f). [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

Download figure to PowerPoint

Next, we investigated the effect of V-ATPase siRNA on Rab27B-induced cell-cycle progression by flow cytometry and Western blotting. Cell-cycle progression was analyzed after serum starvation followed by readdition of 0.5% serum. Silencing of V0d1 subunit blocked G1 to S-phase transition in MCF-7 GFP-Rab27B cells (Fig. 3c). This was not observed in control MCF-7 GFP cells (data not shown). In addition, expression of the positive cell-cycle regulators cyclin A and cyclin E decreased with 90 and 40%, respectively, whereas expression of the negative cell-cycle regulator p27 increased twofold after silencing of V0d1 subunit in MCF-7 GFP-Rab27B cells (Fig. 3d). Consistently, MCF-7 GFP-Rab27B cells demonstrated much lower levels of cell proliferation after transient silencing of V0a1 or V0d1 subunits (Fig. 3e). It is worth noting that V-ATPaseV0a1 knockdown induced a milder reduction in invasion, proliferation and cyclin/p27 expression compared to knockdown of the central stalk V0d1 subunit. Silencing of V0a1 or V0d1 subunits had minimal effects on proliferation of control MCF-7 GFP cells and MCF-7 Snail-6SA cells, demonstrating the dependence to V-ATPase activity of Rab27B-stimulated proliferation (Fig. 3f and Supporting Information Fig. 5).

image

Figure 5. V-ATPase-dependent HSP90α secretion and proMMP-2 activation by GFP-Rab27B expressing breast cancer cells. (a and b) Impact of reversible bafilomycin A1 treatment (5 nM, 24-hr) (a) and transient transfection of siCON versus siV-ATPase V0a1 or V0d1 (b) on the secretion of HSP90α and β in the CM of MCF-7 GFP-Rab27B cells. CM samples were denaturated and separated by SDS-PAGE. (c) Gelatin zymography of MMP-2 activity in CM from MCF-7 GFP-Rab27B cells reversibly treated by bafilomycin A1 (5 nM, 24-hr) and preincubated with exogenously added proMMP-2 (100 ng/mL) in serum-free culture medium for 24-hr. Arrowhead indicates 72-kDa proMMP-2 and arrow indicates 68-kDa active protease. In (a, b and c), intensity was quantified relative to control condition by Image J analysis.

Download figure to PowerPoint

To further investigate whether V-ATPase activity is implicated in Rab27B-mediated invasive growth in vivo, we implanted 106 MCF-7 GFP-Rab27B cells onto the CAM of 9-day-old fertilized chicken eggs. Invasive growth was monitored after 5 days by fluorescence stereomicroscopy. Both MCF-7 GFP-Rab27B control and V-ATPase-silenced conditions were tumorigenic in five out of five inoculated CAMs (Fig. 4a). However, the total area of the CAM occupied by MCF-7 GFP-Rab27B tumor nodes (Fig. 4b) or number of secondary nodes (Fig. 4c) was significantly decreased after transient siRNA targeting of V0d1 subunit. Histological analysis revealed that MCF-7 GFP-Rab27B cells adopt an invasive behavior and infiltrate the adjacent connective tissue to form secondary tumor nodes (Fig. 4d). In contrast, no such infiltration was observed with MCF-7 GFP-Rab27B cells transiently transfected with V0d1 subunit siRNA. In both conditions, nucleated erythrocytes were present in the capillaries infiltrating the tumor mass (Fig. 4d). TUNEL staining demonstrated 2.4% of MCF-7 GFP-Rab27B cancer cells in apoptotic state compared to 10.6% of MCF-7 GFP-Rab27B cells transiently transfected with V0d1 subunit siRNA (p < 0.001) (Fig. 4e). Tumor node lysates of MCF-7 GFP-Rab27B cells exhibited decreased levels of positive cell cycle regulator cyclin A after transient silencing of V0d1 subunits (Fig. 4f).

V-ATPase-dependent Rab27B vesicle size and localization affects HSP90α secretion and MMP-2 activation

We previously demonstrated MMP-2-dependent activation by extracellular secretion of HSP90α as an important regulator of Rab27B-mediated invasion.[2] HSP90α levels were fourfold higher in the CM of MCF-7 GFP-Rab27B cells compared to MCF-7 GFP-Rab27B cells treated with bafilomycin A1 (1 nM, 24-hr) (Fig. 5a). Removal of bafilomycin A1 restored HSP90α release from MCF-7 GFP-Rab27B cells. HSP90β secretion was not affected (Fig. 5a). In agreement, transient silencing of V0a1 or V0d1 subunits resulted in decreased release of HSP90α, but not HSP90β (Fig. 5b). V-ATPase targeting by siRNA or bafilomycin A1 (1 nM, 24-hr) revealed no difference in intracellular levels of HSP90α or HSP90β in MCF-7 GFP-Rab27B cells (Supporting Information Fig. 6). MMP-2 is a protease that degrades extracellular matrix; the 68-kDa active form is produced by cleavage of a 72-kDa precursor. When recombinant proMMP-2 was exogenously added to MCF-7 GFP-Rab27B cells, this MMP was activated in a V-ATPase-dependent manner as demonstrated by the reversible inhibitory effects on proMMP-2 processing of bafilomycin A1 (5 nM, 24-hr) (Fig. 5c).

image

Figure 6. V-ATPase expression in clinical breast cancer specimens. (a) Representative photographs of normal breast (left) and ERα-positive breast cancer tissues with moderate (i.e., score 1) (middle) or high (i.e., score 2) (right) Rab27B expression, stained for V-ATPase V1E subunit. Arrow indicates peripheral V-ATPase V1E localization. Scale bar, 50 µm. (b) Associations of V-ATPase status (negative or positive) with Rab27B immunohistochemical scores for 54 primary breast tumors. p-Values were calculated using the χ2-test. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

Download figure to PowerPoint

Presence of V-ATPase subunits in primary human breast tumors

We next analyzed the expression of V-ATPase subunit proteins in primary ERα-positive breast tumors. We previously showed a significant association between Rab27B expression levels and lymph node-positive ERα-expressing breast tumors.[2] Lysates of epithelial tissues microdissected from fresh-frozen primary human breast cancer tissue were subjected to Western blotting using V-ATPase V0a1-, V-ATPase V0d1- and Rab27B-specific antibodies. Both V-ATPase subunits showed the presence in breast tumor samples characterized by high Rab27B expression (Supporting Information Fig. 7). Immunohistochemical analysis of 54 primary breast cancer samples for V-ATPase V1E expression demonstrated no to faint cytoplasmic or plasma membrane staining in normal breast tissue or breast tumors with no or weak Rab27B expression (i.e., score 0 and score 1 as described in Ref. [2]) (Figs. 6a and 6b) (Supporting Table 1). This relative low expression level may indicate a basal vesicle trafficking activity under physiological conditions or good prognosis breast tumors. In agreement, cell-culture experiments revealed upregulation of V-ATPase V0a1 and V0d1 expression upon overexpression of GFP-Rab27B; suggesting that increased secretory GTPase activity induces expression of components necessary to support and steer GTPase-regulated vesicle traffick (Supporting Information Fig. 1). Intense staining of V-ATPase V1E in the cytoplasm and plasma membrane was observed in 11 out of 16 samples (68.75%) with high Rab27B staining (i.e., score 2) (Figs. 6a and 6b). Our data support the notion that these V-ATPase subunits and Rab27B signatures are relevant for the identification of poor-prognosis breast cancer patients.

Discussion

  1. Top of page
  2. Abstract
  3. Material and Methods
  4. Results
  5. Discussion
  6. Conclusions
  7. Acknowledgements
  8. Author contributions
  9. References
  10. Supporting Information

The Rab27B-regulated secretory pathway in ERα-positive human breast cancer cells delivers proinvasive regulators to the extracellular environment.[2] However, the mechanisms of Rab27B vesicle transport in breast cancer remain poorly documented. Proteomic analysis of purified GFP-Rab27B vesicles identified V-ATPase as a potential regulator of Rab27B-dependent secretory vesicle trafficking. In agreement, quantitative proteomics revealed V-ATPase subunits V1B and V0a1 in the secretory organelles of rat liver homogenates.[16] Biochemical analysis confirmed the presence of V-ATPase in Rab27B vesicles in our study. It is known that V-ATPases create proton gradients required for efficient vesicle trafficking in the secretory pathway as well as migration and invasiveness of breast cancer cells.[7, 8, 11] It is conceivable that V-ATPase activity is required for correct vesicular targeting at the cell periphery and extrusion of their molecular contents within the extracellular space. V-ATPases are pH-sensors that couple intravesicular pH to trafficking. Indeed, Hurtado-Lorenzo et al.[17] demonstrated that through its pH-dependent interactions with the small GTP-binding protein ARF6 and its GDP/GTP exchange factor ARNO, V-ATPase regulates transport from early to late endosomes. Impaired recruitment of ARNO and Arf6 from the cytosol to endosomes resulted in the inability of early endosomes to form carrier vesicles, leading to a blockage of endocytic trafficking between early and late endosomes. A similar mechanism may account for Rab27B-positive vesicles. Our results are consistent with other studies that have evaluated the morphology of secretory pathway compartments. Treatment of professional secretory cells with bafilomycin A1 or concanamycin A eliminated regulated secretion. Drug treatment caused the formation of large, mixed organelles, with components of immature granules and lysosomes. Markers of the trans-Golgi network and earlier secretory pathway were unaffected.[8] Monitoring Rab27B vesicles by confocal microscopy and immunoelectron microscopy demonstrated the functional role of V-ATPase activity and expression levels in the size and peripheral localization of Rab27B vesicles. As a consequence, loss of peripheral Rab27B vesicle accumulation reduced the invasive potential of MCF-7 breast cancer cells in native collagen type I matrices and the CAM tissue of the chick embryo. These responses were associated with inhibition of cellular proliferation and cell-cycle progression at the G1 to S transitions in MCF-7 cells. Further studies are therefore expected to identify the intracellular and extracellular mechanisms associated with acidification and trafficking of Rab27B vesicles and their dual roles in cancer cell proliferation and invasion. Bafilomycin A1 treatment in cell culture does not lead to toxicity based on trypan blue exclusion assay. However, in the in vivo CAM experiment, at 5 days increased apoptosis is observed after siRNA targeting of V-ATPase V0d1. Multiple reasons may be responsible for the observed cell death effects: (i) specific V-ATPase V0d1 targeting seems to affect the integrity of the complete V-ATPase molecule as indicated by concomitant loss of V-ATPase V0a1 after V0d1 silencing; (ii) TUNEL labeling was performed 5 days after siRNA targeting of V-ATPase. We found at least a 60% decrease in V-ATPase V0a1 and V0d1 expression from 2 days till 5 days and (iii) the disintegration of the V-ATPase complex combined with the longer duration times where functional V-ATPase concentrations are decreased may make Rab27B cells more sensitive to apoptosis compared to the short-term (max. 24-hr) drug treatments.

At the molecular level, loss of peripheral Rab27B localization resulted in reduced HSP90α secretion and reduced pro-MMP-2 activation in the extracellular space. In contrast, the V-ATPase inhibitor bafilomycin A1 was previously shown to promote pro-MMP-2 activation in A2058 melanoma cells, HT1080 fibrosarcoma cells and Panc-1 pancreatic cancer cells.[18, 19] Consistently, blockade of V-ATPase activity results in the accumulation of TIMP-free active MT1-MMP at the cell surface and the consequent pro-MMP-2 activation. Taken together, these data are in conflict with two additional reports on the reduction of MMP-2 activity in hepatocarcinoma cells and A549 lung carcinoma cells after V-ATPase knockdown.[20, 21] However, the status of the proinvasive secretory vesicle pathway versus Rab27B was not explored in these reports. Multiple mechanisms of MMP and pro-MMP-2 activation are described to be cell-context dependent and are mediated by several molecular protagonists, such as membrane-associated MT1-MMP, integrin clustering, HSP90α, some of which are V-ATPase-independent, or critically regulated by V-ATPases, in a positive or a negative manner. In agreement with this complexity, V-ATPase inhibitors or siRNAs had no effect on the Rab27-independent invasion model induced by a dominant epithelial to mesenchymal transitory phenotype after ectopic expression of the SNAIL-6SA transcription factor. On the other hand, MT1-MMP is not expressed in MCF-7 cells, and pro-MMP-2 activation is specifically dependent on the Rab27B-mediated extracellular delivery of HSP90α, in a direct connection with our clinical studies, establishing a link between the Rab27B overexpression and the aggressiveness of human breast cancers.[2] Thus, according to the mechanisms involved in pro-MMP-2 activation and MMP equipment, V-ATPases may either stimulate or restrict the metastatic cascade.

Few studies reported V-ATPase expression in clinical cancer specimens. An increased V-ATPase V1E subunit staining intensity was previously described during the neoplastic progression of pancreatic normal ducts to pancreatic ductal adenocarcinoma.[18] A more intense V-ATPase V1C1 subunit expression was also observed at the periphery of tumor nests in oral squamous cell carcinoma samples.[22] Human breast cancer specimens express the V-ATPase V0a1, V0d1 and V1E subunits. Critically, in human breast cancer specimens, the V-ATPase V1E protein was identified in more than two-thirds of the poor prognosis Rab27B-expressing tumors. Thus, the molecular and biological relevance of our data is strongly supported by a correlation between V-ATPase V1E subunit and Rab27B expression level and localization pattern in breast cancer cell lines expressing ectopic GFP-Rab27B, and patient-derived breast cancer tissue. Based on this body of evidence, V-ATPase should be considered as a critical molecular switch in Rab27B-induced invasion. Therefore, our study identifies a new therapeutic target in ERα-positive breast cancers with poor prognosis. Clinically, proton pump inhibitors (PPIs), such as esomeprazole, have been largely and successfully used for the treatment of peptic ulcer diseases.[23] PPIs target (H+/K+)-ATPases localized within the intracellular secretory lumen of gastric parietal cells. Most importantly, PPIs also inhibit the activity of V-ATPases, and thus blocking proton transport across vesicle membranes and potentially blocking secretory vesicle secretion.[24] Furthermore, one can postulate that PPIs can circumvent some aspects of anticancer drug resistances in breast cancer patients. This assumption is supported by recent reports on the pharmacologic and genetic targeting of V-ATPases in sensitization of cancer cells to cytotoxic drugs by PPIs.[7, 11, 25]

Conclusions

  1. Top of page
  2. Abstract
  3. Material and Methods
  4. Results
  5. Discussion
  6. Conclusions
  7. Acknowledgements
  8. Author contributions
  9. References
  10. Supporting Information

Although multiple mechanisms likely underlie and contribute to peripheral Rab27B vesicle distribution, our data suggest that V-ATPase is a promising therapeutic target to prevent the release of proinvasive growth regulators and thereby inhibit Rab27B-dependent invasive growth of ERα-positive breast cancer.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Material and Methods
  4. Results
  5. Discussion
  6. Conclusions
  7. Acknowledgements
  8. Author contributions
  9. References
  10. Supporting Information

The authors thank G. De Bruyne, J. Mestach, M. Demeulemeester and S. Decloedt for excellent technical assistance. Dr. Ir. W. Hendrix assisted with image analysis by Adobe Photoshop. This study was supported by Fund for Scientific Spearheads of the Ghent University Hospital, Stichting tegen Kanker, the Intramural Research Program of the National Human Genome Research Institute, a postdoctoral grant (O.D.W. and A.H.) from Fund for Scientific Research-Flanders, and a postdoctoral grant (A.H.) from Research Council of Ghent University.

Author contributions

  1. Top of page
  2. Abstract
  3. Material and Methods
  4. Results
  5. Discussion
  6. Conclusions
  7. Acknowledgements
  8. Author contributions
  9. References
  10. Supporting Information

A.H. performed laboratory experiments including Western blot analysis, proliferation and invasion assays, data analysis and mass spectrometry. R.S. and W.W. performed immunoelectron and confocal microscopy, respectively. G.S. assisted in the CAM assay. K.L. assisted in the immunohistochemical stainings. A.H. and O.D.W. drafted the manuscript. W.W., H.D., G.B., R.V.d.B., V.C., C.G., M.B. and O.D.W. critically revised the manuscript. A.H., W.W., C.G., M.B. and O.D.W. designed the study. All authors read and approved the final manuscript.

References

  1. Top of page
  2. Abstract
  3. Material and Methods
  4. Results
  5. Discussion
  6. Conclusions
  7. Acknowledgements
  8. Author contributions
  9. References
  10. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Material and Methods
  4. Results
  5. Discussion
  6. Conclusions
  7. Acknowledgements
  8. Author contributions
  9. References
  10. Supporting Information

Additional Supporting Information may be found in the online version of this article.

FilenameFormatSizeDescription
ijc28079-sup-0001-suppinfo.pdf209K

Supplementary Figure 1. Effect of Rab27B overexpression in MCF-7 breast cancer cells on the expression of V-ATPase subunits V0a1 and V0d1. Lysates were subjected to Western blotting with V-ATPase V0a1, V0d1 and Rab27B-specific antibodies. Tubulin was used as loading control.

Supplementary Figure 2. Effect of V-ATPase inhibitor bafilomycin A1 on matrix invasion of MCF-7 breast cancer cells expressing GFP-Rab27B Q78L, a constitutive active mutant defective in GTP hydrolysis. Upper part, phase contrast images showing morphology of MCF-7 GFP-Rab27B Q78L cells on collagen type I gels under control condition or reversibly treated with V-ATPase inhibitor bafilomycin A1 (1 nM, 24h). Scale bar, 50 µm. Bottom part, quantification of collagen type I invasion by MCF-7 GFP-Rab27B Q78L cells untreated (con) or reversibly treated with bafilomycin A1 (0.1, 1 or 5 nM) for 24h. Data are presented as the mean ± s.e.m. of three independent experiments performed in triplicate.

Supplementary Figure 3. Effect of V-ATPase inhibitor concanamycin A on matrix invasion of GFP-Rab27B wild type and Q78L expressing MCF-7 breast cancer cells. Quantification of collagen type I invasion by MCF-7 GFP-Rab27B (wild type and Q78L) cells untreated or reversibly treated with concanamycin A (0.1, 1 or 5 nM) for 24h. Data are presented as the mean ± s.e.m. of three independent experiments performed in triplicate.

Supplementary Figure 4. Effect of pharmacological inhibition and siRNA targeting of V-ATPase on EMT-dependent matrix invasion by MCF-7 Snail-6SA cells. Upper part, phase contrast images showing morphology of MCF-7 Snail-6SA cells on collagen type I gels under control condition or treated with V-ATPase inhibitor bafilomycin A1 (1 nM, 24h). Scale bar, 50 µm. Bottom part, quantification of collagen type I invasion by MCF-7 Snail-6SA cells untreated (con), reversibly treated with bafilomycin A1 (1 or 5 nM) for 24h, or transiently transfected with control siRNA (siCON) or siV-ATPase V0a1 or V0d1 siRNAs. Data are presented as the mean ± s.e.m. of three independent experiments performed in triplicate.

Supplementary Figure 5. Effect of siRNA targeting of V-ATPase on proliferation of MCF-7 Snail-6SA cells. A total of 104 cells were plated into each well of a total of 12 wells on day 1 to establish one growth curve under each condition in triplicate. The total number of cells per well was automatically counted every day until day 4.

Supplementary Figure 6. Effect of V-ATPase targeting by siRNA and bafilomycin A1 inhibitor (1 nM, 24h) on HSP90α and β expression in MCF-7 GFP-Rab27B breast cancer cells. Lysates were subjected to Western blotting with HSP90α and β antibodies. Tubulin was used as loading control.

Supplementary Figure 7. Epithelial tissue was microdissected from frozen fresh primary human breast cancer tissues (patient samples 1 and 2 are poor prognosis ERα-breast cancers with high Rab27B expression). Lysates were subjected to Western blotting with V-ATPase V0a1, V0d1 and Rab27B-specific antibodies. Tubulin was used as loading control.

Supplementary Table 1. Immunohistochemical analysis of V-ATPase in primary breast tumors

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.