Specific interaction of tissue-type plasminogen activator (t-PA) with annexin II on the membrane of pancreatic cancer cells activates plasminogen and promotes invasion in vitro

Abstract

Background: Overexpression of tissue plasminogen activator (t-PA) in pancreatic cancer cells promotes invasion and proliferation in vitro and tumour growth and angiogenesis in vivo.

Aims: To understand the mechanisms by which t-PA favours cancer progression, we analysed the surface membrane proteins responsible for binding specifically t-PA and studied the contribution of this interaction to the t-PA promoted invasion of pancreatic cancer cells.

Methods: The ability of t-PA to activate plasmin and a fluorogenic plasmin substrate was used to analyse the nature of the binding of active t-PA to cell surfaces. Specific binding was determined in two pancreatic cancer cell lines (SK-PC-1 and PANC-1), and complex formation analysed by co-immunoprecipitation experiments and co-immunolocalisation in tumours. The functional role of the interaction was studied in Matrigel invasion assays.

Results: t-PA bound to PANC-1 and SK-PC-1 cells in a specific and saturable manner while maintaining its activity. This binding was competitively inhibited by specific peptides interfering with the interaction of t-PA with annexin II. The t-PA/annexin II interaction on pancreatic cancer cells was also supported by co-immunoprecipitation assays using anti-t-PA antibodies and, reciprocally, with antiannexin II antibodies. In addition, confocal microscopy showed t-PA and annexin II colocalisation in tumour tissues. Finally, disruption of the t-PA/annexin II interaction by a specific hexapeptide significantly decreased the invasive capacity of SK-PC-1 cells in vitro.

Conclusion: t-PA specifically binds to annexin II on the extracellular membrane of pancreatic cancer cells where it activates local plasmin production and tumour cell invasion. These findings may be clinically relevant for future therapeutic strategies based on specific drugs that counteract the activity of t-PA or its receptor annexin II, or their interaction at the surface level.

Figure 1

Binding of tissue-type plasminogen activator (t-PA) to the surface of pancreatic cancer cells determined by a functional assay. (A) t-PA catalysed plasminogen activation on PANC-1 cells. Cells were grown until confluence and preincubated with 15 nM or 250 nM of recombinant t-PA over 20 minutes. Cells were then washed to remove unbound t-PA, and plasmin generation was determined at different times after addition of plasminogen (0.2 μM) and H-D-Val-Leu-Lys-7-amido-4-methylcoumarin (AMC 0.2 mM), as described in experimental procedures. Data shown for relative fluorescence units (rfu) represent generation of active plasmin by active t-PA. Inset shows the linearity of plasmin generation during the first 20 minutes of the reaction. (B) Time course of binding of t-PA to PANC-1 cells. Cells were preincubated with 15 nM t-PA at 37°C for the times indicated and plasmin generation was determined after 15 minutes. Equilibrium binding was reached at 20 minutes. Data are mean (SEM) of values from two experiments performed in triplicate.

Figure 2

Binding of tissue-type plasminogen activator (t-PA) to PANC-1 and SK-PC-1 cancer cells. PANC-1 (A) and SK-PC-1 (B) cells were grown until confluence and incubated with increasing concentrations of recombinant t-PA for 20 minutes. Bound t-PA was determined 15 minutes after addition of substrate. One representative experiment of four performed in triplicate is indicated. Shown is the best fit of binding models to the data. The binding constants, as revealed by non-linear regression analyses of the data, are indicated in the text.

Figure 3

Involvement of annexin II in the binding of tissue-type plasminogen activator (t-PA) to pancreatic cancer cells. Experiments were performed with confluent PANC-1 monolayers. (A) Specificity of t-PA binding was tested using as a competitor of recombinant t-PA (rt-PA 10 nM) t-PA inactivated with diisopropyl fluorophosphate (DFP) at increasing concentrations. (B) Specific contribution of annexin II to binding of t-PA to the surface of PANC-1 cells was studied by disrupting the interaction with a blocking peptide (LCKLSL, filled symbols) or a control peptide (LGKLSL, open symbols). Data are mean (SEM) (n = 3). Inhibition of t-PA binding by LCKLSL was statistically significant (p = 0.007 compared with rt-PA only; p = 0.009 compared with control peptide). (C) Annexin II was released from the cell surface by treatment with EGTA (50 mM) prior to incubation with rt-PA with or without addition of peptide LCKLSL (5 mM). Data are expressed as the percentage of binding of control samples without competitor and represent the mean (SEM) of triplicate samples from a single experiment. Three experiments were performed with very similar results. Inset shows a western blot performed with supernatants concentrated from control cells and EGTA treated cells blotted with antiannexin II antibody. Statistical significance was determined by the Student’s t test: *p = 0.009, **p = 0.0004, ***p = 0.0001.

Figure 4

Co-immunoprecipitation of tissue-type plasminogen activator (t-PA) and annexin II. SK-PC-1 cells were maintained in complete medium until 70% confluence and processed as described in experimental procedures. Left panel: top, cell lysates were subjected to immunoprecipitation with an antibody to annexin II (lane P represents the pelletted fraction and lane S the supernatant or non-immunoprecipitated fraction) or with isotype matching antiannexin I antibody as a control (lane Ct). The electrophoresed proteins were analysed by immunoblotting with anti-t-PA antibody; bottom, the efficacy of the antibody in immunoprecipitating annexin II was checked by blotting the pellet and supernatant fractions with antiannexin II antibody. Right panel: top, cell lysates were subjected to immunoprecipitation with a control goat serum (lane Ct) or with anti-t-PA antibody (lane P represents the pelletted fraction and lane S the supernatant or non-immunoprecipitated fraction) and proteins were analysed by immunoblotting with a mouse antiannexin II antibody; bottom, the efficacy of the antibody in immunoprecipitating t-PA was checked by blotting the pellet and supernatant fractions with anti-t-PA antibody.

Figure 5

Immunocolocalisation of tissue-type plasminogen activator (t-PA) and annexin II in pancreatic tumours. (A) and (B) show representative images of different tumour samples processed for double immunofluorescence staining with mouse antiannexin II and goat anti-t-PA antibodies. Immunostaining for annexin II (Ann-II) is shown in green, immunostaining for t-PA in red, and colocalisation is shown in yellow (Merge). Original magnification ×200.

Figure 6

Expression of annexin II in pancreatic cancer cell lines. Western blotting with antiannexin II antibody of cellular protein extracts (50 μg) from the cell lines: lane 1, SK-PC-1; lane 2, CAPAN-1; lane 3, RWP-1; lane 4, BxPC-3; lane 5, PANC-1; and lane 6, Hs766T. Blots were normalised by subsequent incubation with antibodies to β-actin. The histogram at the bottom shows the ratio of annexin II/β-actin values obtained by laser scanning densitometry of autoradiograms.

Figure 7

Role of the interaction of tissue-type plasminogen activator (t-PA) with annexin II in invasion. SK-PC-1 cells were cultured on Matrigel coated Transwell filters for 24, 48, and 72 hours, with or without the peptide LCKLSL (5 mM) or the control peptide LGKLSL (5 mM) added to the upper chamber. Left panel, quantitative determination of invading cells representing the mean (SEM) of two experiments performed in triplicate. Inhibition of invasion by peptide LCKLSL was significant at 72 hours (p = 0.084 at 24 hours, p = 0.056 at 48 hours, and p = 0.009 at 72 hours). Right panel, cells at the bottom surface of Transwell membranes stained with crystal violet in a representative experiment.