Tumor cell–induced platelet aggregation. (A) Representative traces showing a dose-dependent platelet aggregation response to increasing concentrations (as indicated) of Caco-2 (left trace) or PC3M-luc cells (right trace). TRAP (10 μM) was used as control of the platelet response. These tracings represent fragments of the aggregometry output focusing on the time where most changes in platelet responses are observed. (B) Corresponding dose-response curve displaying the concentration response of WP (300 × 10³ platelets per μL) to increasing concentration of Caco-2 (n = 6), PC3M-luc (n = 6), or PC3 (n = 3) cells. TRAP (10 μM) was used as positive control (data are expressed as mean % Agg ± SEM). (C) Representative curves comparing the lag time in aggregation of Caco-2 and PC3M-luc (n = 4). PC3M-luc–triggered WP aggregation was monitored for 30 minutes per experiment. If aggregation failed to occur during this time period, a lag time of 31 minutes was recorded (data are expressed as mean ± SEM; **P < .01; ***P < .0001). (D) The direct interaction of washed platelets with cancer cells was measured by flow cytometry based on CD41a (platelet-specific surface marker) expression on cancer cells. Cells were gated by forward/side scatter characteristic to exclude cellular debris and platelets. The percentage of CD41a binding was determined in the cell gate as percentage gated, and the fold increase in cancer cells preincubated with platelets compared with cancer cells alone was recorded. Flow cytometry histograms are representative of 4 (PC3M/Caco-2) or 3 (PC3) independent experiments that yielded similar results (*P < .05, n = 4/3, mean ± SEM).