Thrombotic thrombocytopenia associated with COVID-19 infection or vaccination: Possible paths to platelet factor 4 autoimmunity

Michel Goldman and Cédric Hermans discuss thrombotic mechanisms in COVID-19 and rare adverse reactions to SARS-CoV-2 vaccinations.

summarize the key features of heparin-induced thrombocytopenia before proposing that COVID-19 and adenovirus-vectored COVID-19 vaccines can on rare occasions cause autoimmune thrombotic thrombocytopenia mimicking heparin-induced thrombocytopenia.

PF4 autoimmunity in heparin-induced thrombocytopenia
Heparin-induced thrombocytopenia is a severe prothrombotic condition that occurs in less than 5% of patients receiving heparin. Anti-PF4 antibodies are key biomarkers of heparininduced thrombocytopenia [25]. They recognize an epitope exposed on PF4 tetramers upon conformational changes induced by their interaction with heparin or other long polyanions [26]. Indeed, injection of heparin has been shown to induce the release of PF4 [27], resulting in the assembly of PF4/heparin complexes, which activate complement and bind circulating B lymphocytes in a complement-dependent manner [28]. B cells responsible for the synthesis of PF4 autoantibodies display unique characteristics that enable them to rapidly mount an IgG response following a first exposure to heparin [29]. Indeed, B cells, whichAU : Pleaseconfirmthattheedit are able to produce anti-PF4 antibodies, are present in healthy individuals but in an anergic state that normally prevents their activation. This B cell tolerance might be broken upon heparin exposure and under some inflammatory conditions [30]. In these situations, anti-PF4 IgG antibodies elicit thrombus formation and thrombocytopenia via multiple mechanisms. Immune complexes assembled with PF4 bound to heparin induce platelet activation and aggregation by cross-linking FcγRIIA receptors [25]. Anti-PF4 antibodies also activate the procoagulant activity of monocytes by cross-linking their FcγRI receptors and of endothelial cells via the recognition of PF4 firmly attached to surface proteoglycans (PGs) [31]. Thrombocytopenia results from enhanced apoptosis and clearance of antibody-coated platelets in addition to consumption in the coagulation process [8].
A prothrombotic syndrome with all the features of heparin-induced thrombocytopenia has been reported in the absence of heparin exposure [32]. These observations led to the definition of a so-called "spontaneous heparin-induced thrombocytopenia" caused by anti-PF4 autoantibodies elicited by polyanions reproducing the conformational changes induced in PF4 tetramers by heparin [33]. Potential polyanions triggering "spontaneous heparin-induced thrombocytopenia" include bacterial wall components, nucleic acid materials, or endogenous PGs released by damaged cells.

Thrombotic thrombocytopenia during COVID-19: An autoimmune reaction induced by SARS-CoV-2?
The high incidence of thrombotic and thromboembolic events during severe COVID-19 results in the frequent administration of heparin in affected patients [34]. Thrombosis can develop in unusual locations such as cerebral venous sinuses [35]. When thrombocytopenia develops in this setting, heparin-induced thrombocytopenia must be considered as a possible cause [18]. Indeed, several studies report the presence of anti-PF4/heparin antibodies in COVID-19 patients. However, these antibodies sometimes occur in absence of heparin administration [18]. Furthermore, they do not always activate platelets in presence of heparin/PF4 complexes [36], although they do so in presence of PF4 alone [14], suggesting that they were induced by another mechanism than classical heparin-induced thrombocytopenia [26]. There is indeed clinico-biological evidence that infection with SARS-CoV-2 by itself can elicit antibody-mediated thrombotic thrombocytopenia. IgG antibodies present in the serum of severe COVID-19 patients were found to induce platelet apoptosis and procoagulant activity via FcγRIIA receptor-dependent mechanisms [13]. The antigenic specificity of these antibodies was not defined, but one can speculate that at least some of them are directed against PF4.

Thrombotic thrombocytopenia following COVID-19 vaccination
Several observations of prothrombotic thrombocytopenic events following vaccination with the adenovirus-vectored vaccine ChAdOx1 nCoV-19 vaccine (Vaxzevria, Oxford/AstraZeneca) were reported in European countries [44][45][46]. The incidence of these events is very low (around 1 in 100,000 recipients) but still significant by comparison with the background rate. As the clinical presentation is often reminiscent of heparin-induced thrombocytopenia, the hypothesis of a vaccine-induced autoimmune response to PF4 was put forward. Indeed, Greinacher and colleagues, Schultz and colleagues, and Scully and colleagues reported the detection of platelet-activating anti-PF4 antibodies in sera of patients suffering from unusual thrombotic events associated with thrombocytopenia within 4 to 16 days after injection of the ChAdOx1 nCoV-19 vaccine [44][45][46]. Shortly after these observations, 17 cases of  [47]. Strikingly, serum anti-PF4 antibodies were present in the 11 patients in whom they were searched for [47]. So far, there is no evidence for an increased incidence of similar events after administration of mRNA vaccines, suggesting a role for the adenoviral vectors in the induction of the anti-PF4 autoimmune response.
Indeed, Greinacher and colleagues recently reported that ChAdOx1 nCoV-19 vaccineinduced anti-PF4 antibodies do not cross-react with the SARS-CoV-2 spike protein, excluding a phenomenon of molecular mimicry between the viral protein and PF4 [48]. The same group formulated several hypotheses about the vaccine components that could be involved, including adenovirus-derived substances [49]. As adenoviruses are known to activate platelets [50], it is plausible that the replication-deficient adenoviral vector could be directly responsible for the release of platelet-derived PF4. However, this hypothesis implies that significant amounts of vaccine particles would reach the bloodstream after intramuscular injection, which seems unlikely. An alternative scenario depicted in Fig 2 would involve endothelial cells. Indeed, endothelial cells are efficiently transduced upon intramuscular injection [51]. Transduced endothelial cells might be directly damaged by the spike protein that they synthesize, as suggested by in vitro and in vivo observations [52,53]. Furthermore, endothelial cells might expose the spike protein on their luminal side, possibly bound to PG of the glycocalyx as heparan sulfate PGs were shown to be attachment factors for the spike protein [54]. Platelets might then be recruited and activated by the spike protein bound to endothelial cells [9]. PF4 released by activated platelets could combine with anionic PGs shed from endothelial cells. In such a scenario, both the adenovirus and the spike protein would contribute to the formation of immunogenic PF4 following vaccination with adenoviral vector-based COVID-19 vaccines.

Concluding remarks
Autoantibodies to PF4 contribute to thrombotic thrombocytopenia, which occasionally occurs during COVID-19 or after vaccination with adenoviral vector-based vaccines against SARS--CoV-2. We propose that heparan sulfate PG shed from damaged endothelial cells contribute to making PF4 immunogenic. As far as postvaccine thrombotic events are concerned, it will be important to specify the role of the adenoviral vector in view of the current developments of other vaccines based on the same technology. Finally, further research is needed to identify the risk factors, which predispose rare individuals to these severe complications.