br Fig Scanning electron microscopy images

Fig. 5. Scanning electron microscopy images of platelets in lysed whole blood (WB) samples co-incubated with T47D Liproxstatin-1 and its corresponding treatments. A: Platelet exposed to diluent-treated T47D cells displaying a smooth membrane with extending filipodia. B: Platelet exposed to media-treated T47D cells displaying membrane folds (*) and multiple filipodia extending outward from the platelet body (white arrows), with the presence of microvesicles (black arrows). C: Platelets incubated with Anastrozole-treated T47D cells displaying multiple membrane folds (*) giving a rougher appearance, with thick fibrin fibres forming fibrin sheets and fibrin pores present (encircled). Platelet aggregation evident. D: Platelets incubated with Tamoxifen-treated T47D cells displaying membrane folds and lamellipodia and filipodia extending outward from the platelet body towards neighbouring platelets, showing early aggregation (white arrows).
which is not present within the MCF7 cell line [47]. PAF which can be secreted by tumour cells themselves, is an inflammatory phospholipid which mediates platelet aggregation and degranulation [48,49]. Breast cancer cells, including MCF7 and T47D cells, are able to produce thrombin, as well as synthesise inflammatory cytokines mediating platelet recruitment and activation [50]. Our laboratory had previously shown that MCF7 cells induce platelet activation, as evinced by pseu-dopodia extension and aggregation in as little as 5 min [30]. In this study, we thus halved the exposure time to facilitate identifying early activation [10].
Clinical studies indicate Tamoxifen is associated with a 4% higher risk of VTE development [20], while Anastrozole is associated with a lower but significant risk [51]. In this study, hormone-therapy differ-entially affected the ability of the cell lines to induce platelet activation. While our results support other studies assessing Tamoxifen induction of platelet activation [35], our results contradict recent laboratory-based studies which show that Tamoxifen is able to inhibit platelet activation and consequently platelet aggregation [26,27]. This reflects not only differences in dosage, 2 μM Tamoxifen used in our study compared to a range of 3 μM–20 μM [26,27], but also in experimental methodology. In this study platelets were neither washed nor treated directly with Tamoxifen (or Anastrozole); rather cells were pre-treated with hormone-therapy prior to incubation with whole blood reflecting 
the tumour microenvironment and complementing clinical studies. Notably by not washing platelets, plasma was retained, a component which has recently been shown to be essential to platelet activation [13].
Tamoxifen pre-treatment caused T47D cells to induce platelet acti-vation greater than that of the MCF7 cell line, in the P-selectin high (CD62P+++) interval gate. Anastrozole pre-treatment of both cell lines reduced overall IPA. This reduction in IPA does not however, reflect a protective effect against platelet activation. Qualitative assessment in-dicated platelets in a highly active state with pseudopodia extension and the OCS present. This reduction in IPA is thus due to the subsequent loss of P-selectin, which is released rapidly initially, allowing for high amounts to be detected, with levels diminishing as platelet activation progresses [52]. This reduction could also be attributed to a lack of sensitivity of our flow cytometry protocol and thus non-detection of P-selectin positive microvesicles, which were evident under scanning electron microscopy. This agrees with clinical studies showing that Tamoxifen and Anastrozole are associated clinically with greater numbers of P-selectin positive microparticles [53].
Additionally, Anastrozole-treated T47D samples displayed thick fi-brin fibre and plaque formation, indicative of a hypercoagulatory state [30,54,55]. The visualisation of both fibrin plaque and microvesicles further indicates that lysing whole blood maintained the effects of
plasma in the coagulatory process [10,50,56]. Plaque formation is in-dicative of potentially impaired fibrinolytic activity, which while es-sential for the release of tumour cells from their primary site, also re-flects the potential for thromboembolic complications [57].
Our study highlights the importance of the potential effects of dif-ferent cancer subphenotypes on platelet activation. Limitations of this study include small sample size and would benefit from additional ex-ploration of platelet activation. In addition, our study shows the effects of DMSO pre-treatment was not significantly different from respective media controls. Previous studies indicate a converse inhibitory effect of DMSO on platelet aggregation [58,59]. However, the effects of DMSO on platelet activation, has not been well characterised in the literature [58,59].