https://doi.org/10.1007/s00392-025-02625-4
1Universitätsklinikum Tübingen Innere Medizin III, Kardiologie und Angiologie Tübingen, Deutschland; 2Deutsches Herzzentrum München Klinik für Herz- und Kreislauferkrankungen München, Deutschland; 3LMU Klinikum der Universität München Medizinische Klinik und Poliklinik I München, Deutschland
Background
Patients with coronary artery disease (CAD) are at increased risk of developing ischemic events, and contemporary antiplatelet therapy often leads to bleeding events following percutaneous coronary intervention (PCI). Glycoprotein VI (GPVI) is a key receptor in collagen-dependent thrombus formation and is crucial for platelet homeostasis.
Methods
We analyzed the influence of GPVI inhibition with revacept in a randomized, double-blinded trial enrolling 334 patients with CAD undergoing elective PCI. Ex vivo platelet function analyses were assessed alongside plasma chemokine concentrations. We then elucidated changes in GPVI-dependent chemokine concentrations in patients with bleeding events, categorized by Bleeding Academic Research Consortium (BARC) classification during the 30-day clinical follow-up.
Results
Changes in platelet function were observed in patients with revacept treatment and were associated with a characteristic alteration of circulating chemokine concentrations. We found that collagen- and ADP-induced platelet aggregation correlated with plasmatic concentrations of platelet-derived chemotactic cytokines (Figure 1A&B). Notably, the strongest correlation was observed between chemokine expression and sGPVI levels, indicating a major impact of these mediators on collagen-dependent platelet aggregation (Figure 1C). Additionally, extensive correlation analyses revealed a compelling interrelationship between distinct plasma chemokines (Figure 1D).
Further, patients with adverse bleeding events exhibited a distinct fingerprint of chemokines that was associated with modulation of in vitro platelet functions (Figure 2). Specifically, patients with BARC 2-5 bleeding events showed reduced concentrations of Eotaxin (Figure 2A&B). This chemoattractant exhibited a significant drop following PCI in patients with relevant hemorrhage, suggesting an enhanced bleeding risk in patients with CCS (Figure 2C&D).
Moreover, assessment of GPVI-associated changes in chemokine signaling and platelet functions demonstrated an increased diagnostic value in predicting bleeding events in patients with CAD. By training machine learning models on cytokine concentrations, we achieved high diagnostic accuracy in estimating the individual post-interventional bleeding risk of patients with CCS (Figure 3A). Eotaxin emerged as the most significant cytokine contributing to precise bleeding risk estimation (Figure 3B), potentially enhancing early risk discrimination for bleeding events. Additionally, incorporating Eotaxin into machine learning models improved the 30-day risk stratification among all patients with CCS (Figure 3C). Thus, the assessment of plasmatic cytokines, including Eotaxin, was critically associated with the prognosis of post-interventional CAD patients receiving antiplatelet therapy (Figure 3D).
Conclusions
The composition of platelet-derived chemokines correlated with platelet functions following antiplatelet treatment. Thus, assessing chemokines may offer a perspective to identify patients at increased risk of bleeding events. Likewise, modulation of platelet chemokines in patients receiving revacept treatment may contribute to the efficacy of antiplatelet therapy and potentially attenuate pathophysiological cascades leading to hemorrhagic diathesis in patients with CAD.