1Universitätsklinikum Tübingen Innere Medizin III, Kardiologie und Angiologie Tübingen, Deutschland; 2University Tübingen, Institute of Pharmaceutical Sciences Tübingen, Deutschland; 3University Tübingen, Department for Preclinical Imaging and Radiopharmacy Werner Siemens Imaging Center Tübingen, Deutschland; 4University Tübingen, Interfaculty Institute of Biochemistry Tübingen, Deutschland
Ischemic myocardial tissue injury is associated with enhanced pericellular liberation of heme containing proteins such as myoglobin and hemoglobin. Degradation of heme proteins results in free hemin levels which contribute to enhanced platelet activation and thrombus formation. Conventional antithrombotic therapies are ineffective to inhibit hemin-dependent thrombus formation. cGMP-dependent signaling plays a major role in regulation of platelet activation. Thus, we asked whether pharmacological elevation of cGMP levels with DEA/NO-riociguat interferes with hemin-dependent platelet activation.
We found that in contrast to COX-1 (10 µM indometacin) and P2Y12 (10 µM cangrelor) inhibition (Figure A), hemin-induced platelet aggregation was significantly reduced in the presence of DEA/NO-riociguat (max. aggregation, mean±SD: 72.9±6.74 hemin vs 39.8±4.10 hemin + 0.5 µM DEA/NO & 220 nM riociguat, p=0.0009) (Figure B). Further, DEA/NO-riociguat attenuated platelet-dependent thrombus formation on immobilized collagen under flow (thrombus area, mean±SD: 29.3±15.2 hemin vs 0.75±0.75 hemin + 0.5 µM DEA/NO & 220 nM riociguat, p=0.044) which was not found in experiments when indomethacin or cangrelor was present (Figure C). This indicates that cGMP modulation has a specific and significant impact on hemin-dependent platelet function. Further, hemin-induced a concentration-dependent formation of aggregatory (CD42b+, PAC1+ AnnexinV-), procoagulant (CD42b+, PAC1-, AnnexinV+) and ferroptotic bodies / microvesicle (CD42b-, PAC1-, AnnexinV+) subpopulations as defined by flow cytometry. The formation of ferroptotic bodies / microvesicles was significantly decreased in platelets pretreated with DEA/NO-riociguat (0.5 µM & 220 nM), indicating that enhanced levels of cGMP suppress hemin-dependent destruction and dissolution of the plasma membrane (ferroptotic bodies / microvesicle, mean±SD: 53.5±9.3 hemin vs 17.1±7.7 hemin + 0.5 µM DEA/NO & 220 nM riociguat, p=0.0016). Hemin-mediated iron overload of cells leads to significant changes of the lipid composition of plasma membranes a marker of ferroptosis. To further disclose the significance of cGMP for platelet ferroptosis we performed untargeted lipidomics mass spectrometry (UHPLC–ESI-QTOF-MS/MS). Hemin induced significant changes to the platelet lipidome. Especially arachidonic acid derivates were increased. In the presence of DEA/NO-riociguat (0.5 µM & 220 nM) formation of ferroptotic lipids (e.g. 12-HHT, TXB2, 15-HETE) was reduced (Figure D).
We conclude that cGMP modulates hemin-induced platelet activation and thrombus formation in vitro and cGMP effects hemin-induced platelet ferroptosis and changes in the platelet lipidome. Thus, it is tempting to speculate that modulating platelet cGMP levels may be a novel strategy to control thrombosis and myocardial ischemia (Figure E).