A critical role of platelet TGF-β release in podoplanin-mediated tumour invasion and metastasis

Abstract

The tumour microenvironment is critical for various characteristics of tumour malignancies. Platelets, as part of the tumour microenvironment, are associated with metastasis formation via increasing the rate of tumour embolus formation in microvasculature. However, the mechanisms underlying the ability of tumour cells to acquire invasiveness and extravasate into target organs at the site of embolization remain unclear. In this study, we reported that platelet aggregation-inducing factor podoplanin expressed on tumour cell surfaces were found to not only promote the formation of tumour-platelet aggregates via interaction with platelets, but also induced the epithelial-mesenchymal transition (EMT) of tumour cells by enhancing transforming growth factor-β (TGF-β) release from platelets. In vitro and in vivo analyses revealed that podoplanin-mediated EMT resulted in increased invasiveness and extravasation of tumour cells. Treatment of mice with a TGF-β-neutralizing antibody statistically suppressed podoplanin-mediated distant metastasis in vivo, suggesting that podoplanin promoted haematogenous metastasis in part by releasing TGF-β from platelets that was essential for EMT of tumour cells. Therefore, our findings suggested that blocking the TGF-β signalling pathway might be a promising strategy for suppressing podoplanin-mediated haematogenous metastasis in vivo.

Figure 1. Podoplanin-mediated platelet aggregation promotes epithelial-mesenchymal transition in tumour cells.

(a) Immunoblot analysis showing podoplanin expression in UM-UC-5 and H226 but not in A549 cells. TopoIIβ was used as a loading control. (b) UM-UC-5, H226 and A549 cells (5 × 10⁴ cells) were incubated with washed mouse platelets (5 × 10⁷ platelets/200 μl assay) suspended in Tyrode’s buffer containing 2% platelet-poor plasma and 250 μM CaCl₂. Light transmittance of samples was measured to determine the aggregation rate using an aggregometer. (c) Schematic representation of collection of tumour-platelet reactant supernatants. (d,e) Morphological and physiological changes in cells after treatment with or without supernatants of tumor-platelet reactants for 48 h. (d) Images were captured using phase-contrast microscopy (left panels) or immunofluorescence microscopy (right panels) after staining for E-cadherin (green), F-actin (red; phalloidin) and nuclear DNA (blue; Hoechst 33342). Scale bars represent 50 μm. (e) Cellular lysates were immunoblotted with antibodies to N-cadherin, claudin-1, podoplanin (PDPN, clone D2–40) and TopoIIβ.

Figure 2. TGF-β released from platelets is critical for podoplanin-induced epithelial-mesenchymal transition.

(a) Measurement of TGF-β1 or PDGF-BB concentrations in tumour-platelet reactants using Bio-Plex suspension array system. Error bars indicate standard deviation (SD). **P < 0.01 by Student’s t test. (b,c) Morphological and physiological changes in cells after treatment with or without 3 ng/ml recombinant TGF-β1 for 48 h. (b) Cells were stained for E-cadherin (green), F-actin (red; phalloidin) and nuclear DNA (blue; Hoechst 33342). Scale bars represent 50 μm. (c) Cell lysates were immunoblotted with antibodies to N-cadherin, claudin-1, podoplanin (PDPN, clone D2-40) and TopoIIβ. (d) Cells were either left untreated or treated with supernatants of platelets alone (platelets), supernatants of platelet–cell reactants (platelets + cells), or 3 ng/ml of recombinant TGF-β1 for 0.5 h. The cell lysates were immunoblotted with antibodies against phospho-Smad2/3 (pSmad2/3), Smad3, and TopoIIβ.

Figure 3. TGF-β/TGFβR signaling is involved in podoplanin-induced epithelial-mesenchymal transition in UM-UC-5 cells.

(a–c) UM-UC-5 cells were treated with or without TGF-β1 neutralizing mAb (1D11 mAb) or TGFβR inhibitors (LY2157299 or SB431542) for 2 h, followed by incubation with supernatants of UM-UC-5-platelet reactants for 48 h. Morphological and physiological changes in treated cells were examined by immunoblotting (a), immunofluorescence staining (b) and invasion assay using a matrigel-coated transwell chambers (c). (a) Cell lysates were immunoblotted with antibodies to N-cadherin, claudin-1, podoplanin (PDPN, clone D2-40) and TopoIIβ. (b) Cells were stained for anti-E-cadherin (green), F-actin (red; phalloidin) and nuclear DNA (blue; Hoechst 33342). Scale bars represent 50 μm. (c) Cells were either left untreated or treated with supernatants of UM-UC-5-platelet reactants for 48 h. Next, 5 × 10⁴ UM-UC-5 cells were added to the upper chambers of matrigel-overlaid membranes. After incubation for an additional 48 h at 37 °C, cells migrating through the membranes were fixed and stained with crystal violet (lower panels; scale bars represent 200 μm). Optical density (OD) of crystal violet extracted from cells was measured at 540 nm and presented as a percentage of the OD values of control cells. All data are shown as means ± standard deviation (SD, n = 8). **P < 0.01 by the Mann-Whitney U test (upper panel).

Figure 4. Podoplanin is necessary for TGF-β release from platelets and epithelial-mesenchymal transition.

UM-UC-5 cells were infected with lentivirus containing shRNA targeting human podoplanin (shPDPN_23 and shPDPN_26) or control (shControl). Cells with stable knockdown of podoplanin were used in the experiments. (a) Immunoblot analysis of podoplanin expression in shPDPN_23, shPDPN_26 and shControl cells. TopoIIβ was used as a loading control. (b) ShPDPN_23, shPDPN_26 and shControl cells (5 × 10⁴ cells) were incubated with washed platelets (4 × 10⁷ platelets/200 μl) suspended in Tyrode’s buffer containing 2% platelet-poor plasma and 250 μM CaCl_2. Light transmittance of samples was measured to determine the aggregation rate using an aggregometer. (c) TGF-β1 concentrations in tumour-platelet reactants were determined by enzyme-linked immunosorbent assay. All data are shown as means ± standard deviation (SD, n = 3). **P < 0.01 by the Mann-Whitney U test. (d,e) Morphological and physiological changes in shPDPN_23, shPDPN_26 and shControl cells after treatment with or without supernatants of tumour-platelet reactants for 48 h. (d) Cells were stained for E-cadherin (green), F-actin (red; phalloidin) and nuclear DNA (blue; Hoechst 33342). Scale bars represent 50 μm. (e) Cell lysates were immunoblotted with antibodies to N-cadherin, claudin-1, podoplanin (PDPN, clone D2-40) and TopoIIβ (e). (f) Cells were either left untreated or treated with supernatants of stable transfectant-platelet reactants for 48 h. Next, treated transfectants (5 × 10⁴/well) were added to the upper chambers of matrigel-overlaid membranes. After incubation for an additional 48 h at 37 °C, cells migrating through matrigel-overlaid membranes were fixed and stained with crystal violet (lower panels; scale bars represent 200 μm). Optical density (OD) of crystal violet extracted from cells was measured at 540 nm and presented as percentages of the OD values of supernatant-untreated shcontrol cells. All data are shown as means ± SD (n = 8). N.S., not significant. **P < 0.01 by the Mann-Whitney U test (upper panel).

Figure 5. Neutralization of TGF-β attenuates tumour extravasation and pulmonary metastasis.

(a) Mouse control IgG (Mouse IgG) or TGFβ neutralizing mAb (clone 1D11) (100 μg/mouse) was administrated by the intravenous (i.v.) route to 5-week-old male CB17/Icr-Prkdc^scid/CrlCrlj mice 1 h before i.v. inoculation of UM-UC-5 cell suspensions (1.0 × 10⁶/mouse). In some experiments, 1D11 mAb was administrated 2 and 4 days after tumor cell inoculation (1D11 mAb (x3)). Mice were euthanized 30 days after cell inoculation and metastatic foci on the lung surface were counted. Bars represent mean values. **P < 0.01 by the Mann-Whitney U test. (b) Mouse control IgG (Mouse IgG) or TGFβ neutralizing mAb (clone 1D11) (100 μg/mouse) was administrated to 5-week-old male CB17/Icr-Prkdc^scid/CrlCrlj mice 1 h before tumour inoculation. Calcein-AM-labeled UM-UC-5 cell suspensions were then inoculated (1.0 × 10⁶/mouse) into mice. After 30 min or 48 h after i.v. tumour inoculation, mice were euthanized and frozen lung sections were fixed and stained by Hoechst 33342. Representative merged images of calcein-AM-labeled UM-UC-5 cells (green) stained for nuclear DNA (blue) are shown. (c) The fluorescence intensity of calcein-AM was measured using BioRevo BZ-9000. N.S., not significant. *P < 0.05 by the Mann-Whitney U test.