Quantification of heparin's antimetastatic effect by single‐cell force spectroscopy

In circulation, cancer cells induce platelet activation, leading to the formation of a cancer cell‐encircling platelet cloak which facilitates each step of the metastatic cascade. Since cancer patients treated with the anticoagulant heparin showed reduced metastasis rates and improved survival, it is supposed that heparin suppresses the cloak's formation by inhibiting the interaction between platelet's adhesion molecule P‐selectin with its ligands on cancer cells. To quantify this heparin effect, we developed a single‐cell force spectroscopy approach and quantified the adhesion (maximum adhesion force [FA] and detachment work [WD]) between platelets and human non‐small cell lung cancer cells (A549). A configuration was used in which A549 cells were glued to tipless cantilevers and force‐distance (F‐D) curves were recorded on a layer of activated platelets. The concentration‐response relationship was determined for heparin at concentrations between 1 and 100 U/mL. Sigmoid dose‐response fit revealed half‐maximal inhibitory concentration (IC50) values of 8.01 U/mL (FA) and 6.46 U/mL (WD) and a maximum decrease of the adhesion by 37.5% (FA) and 38.42% (WD). The effect of heparin on P‐selectin was tested using anti‐P‐selectin antibodies alone and in combination with heparin. Adding heparin after antibody treatment resulted in an additional reduction of 9.52% (FA) and 7.12% (WD). Together, we quantified heparin's antimetastatic effect and proved that it predominantly is related to the blockage of P‐selectin. Our approach represents a valuable method to investigate the adhesion of platelets to cancer cells and the efficiency of substances to block this interaction.

thrombophlebitis with concealed cancer. 7 From this time on this connection has been reported for various types of cancer in several case reports and cohort studies. 5 The platelet-cancer relationship is known for their deleterious action in all steps of cancer progression but there are also observations that indicate protective functions of platelets in cancer. 8 However, it was shown that the metastatic activity correlates to the platelet count: The depletion of platelets inhibits metastasis, 9 whereas platelet reconstitution restores metastatic abilities. 10 It is undoubted that platelets contribute to metastasis in various ways. 11 One of the first steps of the metastatic cascade is the intravasation and a cancer cell that enters the blood stream is termed a circulating tumor cell (CTC). Only a very small fraction of CTCs survives the first minutes in blood vessels 12 due to cell-degrading high shear forces and immune surveillance. 13 Platelets protect CTCs by forming a CTCencircling protective cloak. This platelet cloak then facilitates the shielded cancer cells in each following step of the metastatic cascade, starting with immune evasion and the tethering and arrest to the vessel wall, as well as the extravasation, colonization, angiogenesis, and tumor growth which, in the end, leads to the formation of a metastatic focus. 4 Formation of this protective platelet cloak needs a stable adhesion between platelets and CTCs. Under physiological conditions, platelet surface receptors are stored in their granules and platelet activation leads to the redistribution of various adhesion molecules from membranes of secretory granules to the cell surface of platelets. 11,14 CTCs release a variety of platelet-activating mediators such as thromboxan A 2 , thrombin, ADP, CD97, and high-mobility group box 1 (HMGB1). 3,5 The ability of cancer cells to release factors inducing platelet activation is called tumor cell-induced platelet activation (TCIPA) and one of the reasons of a prothrombotic state in cancer. 15 Platelets in close vicinity to CTCs became activated, redistribute their adhesion molecules from secretory granules to the cell surface and release platelet-activating molecules like thromboxan A 2 , thrombin and ADP. These mediators, among others, in turn serve as amplifying signal in the activation of additional platelets. 16 Upon activation platelets become "sticky" due to the membrane expression and activation of several adhesion molecules like six different integrins (α IIb β III , α 2 β 1 , α 5 β 1 , α 6 β 1 , α L β 2 , α v β 3 ), GPIb-IX-V, GPVI, CLEC-2, and P-selectin. 17 Binding of platelets to CTCs and the formation of a protective cloak is mediated by these platelet surface receptors. In the first step of this physical interaction, P-selectin forms a weak bond with sialyl-Lewis X carbohydrates like P-selectin glycoprotein ligand-1 (PSGL-1, CD162) 18 as well as with its other ligands like tethered mucins 1, 3, 4, 10 to 18 3 on cancer cells. Binding of P-selectin triggers binding of the cytoplasmic tail of P-selectin to talin1 which in turn activate α IIb β 3. 19 It follows the firm adhesion of platelets to CTCs established by α IIb β 3 and other platelet integrins. 20 However, the P-selectin induced integrin-mediated platelet adhesion to tumor cells is an important step of metastasis. 21,22 The average rupture force (F rupt ) of a P-selectin PSGL-1 ligand receptor bond was quantified on the single molecule level and found to be F rupt = 171 ± 56 pN. 23 The present work aims on quantification of the platelet to cancer cell adhesion with a particular emphasis on the P-selectin-dependent adhesion using atomic force microscopy (AFM)-based single-cell force spectroscopy (SCFS). SCFS is the favorable technique for this purpose since it allows measuring the adhesion between two cells under physiological conditions with a high force resolution. 24 Furthermore, SCFS allows adding inhibitors or antibodies during an experiment thus enhancing the validity of receptor-blocking maneuvers. Heparin was used to quantify the P-selectin dependent adhesion since it is a potent inhibitor of P-selectin 25 and known for its antimetastatic potential. 26 The ability of tumor cells to activate platelets (TCIPA) causes a sevenfold increased risk of venous thromboembolism (VTE) 27 and need therefore an anticoagulant therapy. Heparins are the standard anticoagulation treatment for cancer-related VTE. 28 It was shown that heparins not only reduce mortality and morbidity related to VTE but also improve the overall survival of cancer patients. 27,29 It is supposed that heparin suppresses the adhesion of platelets to CTCs by inhibition of platelets P-selectin 30,31 and thereby prevent the formation of the protective cloak. Following this logic, antiadhesive features of heparin would be the equivalent of its antimetastatic effect.
Anti-P-selectin antibodies were used alone and in combination with heparin in order to test this hypothesis.

| Platelet isolation from peripheral blood
Venous blood was drawn from healthy volunteers, who gave their informed consent and did not take any anti-coagulative drug within the last 2 weeks, using a 5 mL citrate S-Monovette (Sarstedt, Nümbrecht, Germany). In order to prevent platelet activation 500 μL

| Antibody experiments
To address the effects of heparin to a specific target molecule, anti-P-   The effect of anti-P-selectin antibodies was tested in paired experiments: the A549-platelet adhesion was measured and followed by adding antibodies in-situ. The application of antibodies known to block P-selectin activity reduced the F A significantly to 68.05%

| Statistics
(Mann-Whitney U test, α: .05; Figure 8). Hereinafter heparin was added in-situ and caused a significant further reduction of F A by 9.52% (median Antibody − median Heparin ).
W D was also significantly reduced to 48.63% upon application of the antibodies (Figure 9). Heparin application after antibody treatment additionally reduced W D significantly by 7.12% (median Antibody − median Heparin ).   Table 1 detachment work completely. However, since heparin does not influence the adhesion of platelets to collagen, the observed effects of heparin reflect changes between the force/work of platelets and A549 cells.
The activation of a platelet leads not only to shape change but also to redistribution of adhesion proteins to the platelet surface, for example, glycoprotein IIb/IIIa (GPIIb/IIIa, also known as integrin αIIbβ3) and P-selectin (CD62P). 44 P-selectin is a transmembrane glycoprotein involved in cell-cell interactions under different pathological conditions, including cancer metastasis and inflammation. 22,30,45 Binding between platelets and other cell types is mediated primarily by P-selectin. The density of P-selectin is approximately 350 sites/ μm 2 or 10 000 P-selectin molecules on each activated platelet. 46 P-Selectin forms weak bonds with its ligand P-selectin glycoprotein ligand-1 (PSGL-1). PSGL-1 is a transmembrane glycoprotein expressed on nearly all types of cancer cells and it was shown that platelets interact with A549 lung cancer cells via P-selectin-PSGL-1 bonds. 18 Heparin, a highly sulfated glycosaminoglycan, is a potent inhibitor of P-selectin activity 21 and inhibits the binding of tumor cells to immobilized P-selectin. 47 The inhibitory effect of heparin on A549-platelet interaction was tested using SCFS and the dose-response relationship was determined in the range of 1 to 100 U/mL heparin. acquisition of the F-D curves under heparin free condition. F-D curves were recorded in presents of P-selectin blocking antibodies and hereinafter heparin was added to a final concentration of 5 U/mL. Figures 8 and 9 show that anti-P-selectin antibodies reduce F A to 68.05% and W D to 48.63% (median values). Application of heparin (5 U/mL) reduced adhesion further by additional 9.52% (F A ) and 7.12% (W D ). The remaining adhesion upon antibodies and heparin treatment is 58.53% (F A ) and 41.51% (W D ) and therefore below the bottom asymptote values obtained from the fitting curve of the heparin experiments. The small but significant difference of heparin and anti-P-selectin antibodies on adhesion indicated that heparin blocks not exclusively the P-selectin-PSGL-1 bond. It is likely that heparin also inhibits the adhesion by blocking other, yet unidentified, receptors since heparin binds to a variety of membrane-anchored and extracellular proteins. 48,49 It was reported that binding of heparin and its derivates to heparan sulfate (HS) binding proteins modulate the function of these HS-binding proteins with a pivotal role in cancer growth and progression. 50 However, the antibody approach demonstrates that the P-selectin-PSGL-1 bond contributes largely to the adhesion between A549 cancer cells and platelets.
The meaning of platelets for metastasis is an undoubted case and the binding of platelets to circulating cancer cells a crucial step. 4 It seems reasonable to conclude that blocking the binding of platelets to circulating cancer cells will cause a substantial reduction of metastasis and may explain the antimetastatic effect of heparin. 51 Our approach aimed to quantify the inhibitory effect of unfractionated heparin by means of SCFS. Therapeutic levels of unfractionated heparin normally range from 0.3 to 0.7 U/mL (by factor Xa inhibition). 52

| CONCLUSION
With our method, we quantified the antimetastatic effect of heparin on single-cell level. For both parameters of the platelet-A549 cell interaction, F A and W D , we were able to determine the IC 50 values of the anticoagulant. In our antibody experiments we showed that heparin's antimetastatic effect predominantly is related to P-selectin but also to yet unidentified adhesion molecules. At long sight, this may support the development of new antimetastatic strategies.