br Statistical analysis br Data was exported from FACSDiva s
2.7. Statistical analysis
Data was exported from FACSDiva software and managed in Microsoft Excel. The index of platelet activation (IPA) was calculated such that: IPA = gMFI(CD62P) X n(CD41a+CD62P+ events) for each interval, and combined intervals (Fig. 1) [10,36]. IPA allows for de-scription of P-selectin expression relative to the number of platelets expressing the marker. Statistical analyses were performed using PAST 3.04 . A Shapiro-Wilk test of normality determined that data were not normally distributed thus the Kruskal-Wallis test was conducted followed by a post-hoc Dunn's test with significance at p < 0.05.
Fig. 1. Representative scatter plot showing the spread of platelet activation across the interval gates. Quadrant 1 is the inactive platelet population (CD41a+CD62P−) population, while the activated population of interest (CD41a+CD62P+) was within the 2nd quadrant. Interval gates (I1-I5) were drawn within the quadrant, indicative of increasing CD62P expression. Thrombosis Research 177 (2019) 51–58
The interval gating strategy employed allows mapping of the spread of P-selectin, in addition to ascertaining total P-selectin expression (exposure) with a right shift along the X-axis reflecting higher levels of P-selectin expression (Fig. 1). Interval gates beyond I5 showed no P-selectin expression for all samples and were disregarded in the IPA calculation. Interval 1 and 2 (I1 and I2) represent the bulk of CD41a+ platelets with low P-selectin expression (CD41a+CD62P+), interval gates 3 and 4 (I3 and I4) represent platelets with moderate P-selectin expression (CD4a1+CD62P++) and interval 5 (I5) presents the smallest proportion of platelets with high P-selectin expression (CD41a+CD62P+++) (Fig. 1).
Baseline level of activation showing levels of IPA was determined with the negative control, whole blood unexposed to 6NBDG (Fig. 2). This was corroborated by platelet ultrastructural assessment showing that the majority were round, with a smooth membrane and multiple sur-face folds (Fig. 3A). This contrasted as expected, with the positive control that ensured firstly, that platelets were responsive to activation via agonist action and had not been rendered fully active by the platelet preparation protocol  and secondly, as a technical flow cytometry control. Thrombin induced a greater spread of P-selectin expression and an increase (p > 0.05) in overall IPA (I1–I5) (Fig. 2). This was reflected ultrastructurally (Fig. 3B), where platelets showed greater membrane folds, with filipodia extending from the platelet body.
Breast cancer cells (diluent and matched media controls) induced significantly higher overall IPA levels compared to untreated whole blood, comparable with the positive control. While not significantly greater to the IPA induced by cells treated with normal media, the di-luent elicited a slightly higher IPA and a greater spread of data (Fig. 2A, B). Ultrastructurally, platelets in the diluent control for both cell lines (Fig. 4A, 5A), displayed a relatively smooth but folded platelet mem-brane with lamellipodia extending and attaching to the substratum. This was more apparent in media controls, which also showed micro-vesicle presence (Fig. 4B, 5B). Microvesicles were characterised as small rounded particles of varying sizes with a clear border, adhered to the coverslip. However, some debris was interspersed throughout the sample making the distinction between the specific particles diﬃcult .
Cells treated with hormone-therapy prior to incubation with whole blood induced a greater spread and overall IPA than the negative control (Fig. 2A, B). Compared to matched media and diluent controls, Anastrozole treatment reduced, albeit not significantly, breast cancer cell-induction of platelet activation as defined by median IPA (Fig. 2A, B). Ultrastructurally, Anastrozole-treated MCF7 cells (Fig. 4C), induced a smooth yet folded platelet membrane with extending filipodia. In stark contrast, Anastrozole-treated T47D cells although inducing a si-milarly reduced IPA (albeit p > 0.05) compared to diluent and media controls, altered morphology substantially (Fig. 5C). Platelets had a rougher, folded, membrane with pores associated with the OCS evident. Platelet aggregations were present, in addition to thick fibrin fibres forming dense plaques with fibrin pores present. The reduction in IPA induced by Anastrozole thus reflects a loss of P-selectin expression as-sociated with later stages of coagulation.
Tamoxifen enhanced the ability of MCF7 cells to induce platelet activation, although not significantly so (p > 0.05) compared to the media control (Fig. 2A). Platelets were spread with lamellipodia and filipodia extension, and pores of the OCS and microvesicles evident (Fig. 4D). Tamoxifen-treated T47D cells induced a slightly lower level of IPA (Fig. 2B), compared to media and diluent controls. Platelets exhibited a smoother yet folded membrane with lamellipodia and fili-podia extending outward from the platelet body towards nearby pla-telets suggesting early stages of platelet aggregation. Microvesicles were also visible (Fig. 5D).