SARS-CoV-2 spike protein-induced platelet activation: Mechanism for virus and vaccine-induced thrombotic thrombocytopenia


 Covid-19 pandemic stimulated an extremely fast development of effective vaccines. Recent studies found platelet-activating antibodies against platelet factor 4 (PF4) in both clinically ill Covid-19 patients and vaccine-induced thrombotic thrombocytopenia (VITT) patients. Here, we use various tools to identify the binding reaction of the SARS-CoV-2 spike glycoprotein (SP) with PF4 that results in immunogenic platelet-activating PF4/SP complexes. This binding is evidenced by an increase in mass, optical intensity, and stable binding force observed by quartz crystal microbalance, enzyme immune assay, and force spectroscopy, respectively. The SP induced an increase in the size of PF4 and switched the surface zeta potential of the PF4 from positive to negative values as evaluated by dynamic light scattering. The SP-induced platelet aggregation was identified by functional assay and flow cytometry but in a concentration-dependent manner. Our results indicated that the formed PF4/SP complexes can, on one hand, trigger the formation of PF4-antibodies and on the other hand mediate/activate platelets followed by inducing thrombotic events, which is the mechanism for excessive procoagulant activity observed in Covid-19 patients. With vector-based vaccines, we suggest that soluble SP are produced during the transcription process, forming antigenic PF4/SP complexes that result in a high rate of clotting effects in vaccinated individuals with Ad26.COV2.S and ChAdOx1nCoV-19 vaccines. An additional consideration of PF4/SP complexes in the current guidelines for the diagnosis of VITT will improve the treatment in patients. Our results serve a high demand to develop an effective method to treat Covid-19 patients and improve the safety for Covid-19 vaccination.


Introduction
The covid-19 pandemic caused to date (July 24, 2021) ~195 million infected cases and ~4.2 million deaths globally. 1,2 Hospitalized Covid-19 patients are at risk for developing thrombotic complications such as stroke, limb ischemia, or venous thromboembolism. 3 The incidence of thromboembolism in patients with severe pneumonia, who have a high risk of inhospital mortality, 4 is about 25%. 5 Abnormal coagulation profiles were observed in ~71.4% of the deaths. 6 Heparins are recommended for the management of coagulopathy even at higher/double doses, 7 as they appear to be associated with a better prognosis. 8 However, heparin-induced thrombocytopenia (HIT) antibodies were confirmed in severe Covid-19 patients. [9][10][11][12] HIT antibodies are developed due to immune response (especially by B lymphocytes) to antigenic complexes, formed between the positively charged chemokine platelet factor 4 (PF4, CXCL4) and negatively charged heparin (H). 13 These human antibodies cross-link and activate platelets, monocytes, and neutrophils, and bind to endothelial cells (EC) through PF4/H complexes, causing life-threatening complications. 14 These antibodies contain a platelet-activating FcγRIIa receptor that induced procoagulant platelets. 15 The typical HIT antibodies recognize only PF4/H complexes while PF4-autoantibodies interact with both PF4 alone and PF4/H complexes, causing platelet activations. 15 The incidence of HIT in Covid-19 patients at the intensive care unit is higher (1.16%) than for critically ill patients without Covid-19 (0.18%). 16 To fight against the pandemic, different types of Covid-19 vaccines have been widely administered. [17][18] However, individuals exposed to the incorporated replication-incompetent adenoviral vector vaccine (ChAdOx1 nCoV-19, AstraZeneca) developed PF4-antibodies, which activated platelets, causing thrombotic thrombocytopenia. [19][20][21][22][23] Exposed entities to the mRNA vaccines (BNT162b2, mRNA-1273) also suffered from this side effect, [24][25] but with a lower rate than the above vaccines. 26 Although platelet-activating PF4-antibodies have been detected in both Covid-19 patients and vaccinated individuals, the mechanism of the development of these antibodies has not yet been elucidated. In general, the inflammatory state and the activation of the coagulation cascade play a role in the development of plateletactivating antibodies.
The SARS-CoV-2 spike protein plays a key role in the receptor recognition and cell membrane fusion process, composing of S1 and S2 subunits. 27 The virus binds initially to the cellular receptor angiotensin-converting enzyme 2 (ACE2) through the S1 subunit to get entry into cells. [28][29] After binding, the virus promotes the fusion of its remaining S2 subunit to undergo large-scale conformational rearrangements. This process leads to an exposure of the hydrophobic fusion peptide on the S2 subunit to fuse the viral into cellular membranes. 30 It is proposed that the positively charged amino acids of the S1 binding region promote the virus fusion with the host. In contrast, the S2 subunit includes amino acids that exhibit a mixture of not only negatively and positively charged residues but also hydrophobic and hydrophilic regions, in which hydrophobic and hydrophilic surfaces are dominant. 31 In the meanwhile, the plasma-derived purified PF4 protein has a ring of positively charged residues which enable it to bind electrostatically to different molecules, resulting in clear pro-inflammatory properties in vivo. 32 Healthy donor's sera contain less than 0.5 µg/ml PF4 but PF4 plasma levels are strongly increased up to 30 µg/ml in both severe and non-severe Covid-19 patients. 33 High concentration of PF4 enhances binding of HIT antibodies. 34 PF4 binds strongly to the negatively charged polyanions, 35 extracellular glycosaminoglycans (GAGs) on platelets, 14 and heparan sulfate on endothelial cells, [36][37] inhibiting local antithrombin (AT) activity and thus promoting coagulation.
Based on protein characteristics, we hypothesize that SARS-CoV-2 spike protein subunit 2 (SP) clusters the PF4, forming immunogenic complexes that trigger the immune cells, especially B-cells to secrete platelet-activating PF4-antibodies in some patients who had rare biological characteristics, leading to thrombotic thrombocytopenia.
Here, we proved that the SP clusters PF4 by using multiple nanobiophysical and analytical methods. Dynamic light scattering (DLS) was used to track the change in the size and zeta potential of PF4 caused by SP whereas Quartz crystal microbalance (QCM) was applied to track mass change due to the binding of SP to PF4. Single-molecule force spectroscopy (SMFS) based atomic force microscopy was used to identify the binding force of the interactions while Enzyme-linked immunosorbent assay (ELISA) was used to confirm binding on a 96-well plate. We found that the SP clustered PF4 and switched the surface zeta potential of the resulting complexes to negative values and the maximal size of PF4/SP complexes was gained only at a narrow range of PF4:SP concentration. Importantly, SP strongly induced platelet aggregation as tested in functional HIT assays and visualized by confocal laser microscopy (CLSM). SP-induced platelet activation was identified by flow cytometry. Based on these results and recent findings, we propose a mechanism in which the resulting antigenic PF4/SP complexes promote immune B-cells to producing plateletactivating PF4-antibodies that can massively induce platelet aggregation/activation. This explains the unusual clotting disorders in both Covid-19 patients and vaccinated individuals.

Ethics
The use of blood obtained from healthy volunteers was approved by the ethics board of Thüringen.

Quartz crystal microbalance (QCM)
All preparations and experiments were done at room temperature (RT). The quartz sensor QSX 301 with a resonance frequency of 4.95 MHz ± 50 kHz (Biolin Scientific Darmstadt, Germany) was cleaned in a 5:1:1 mix of H2O:NH3:H2O2 solution in an ultrasonic bath for 10 min. 38 After rinsing with water and ethanol and drying with nitrogen, a self-assembled monolayer (SAM) of cysteamine and glutaraldehyde which contains functional aldehyde (-CHO) groups for binding of protein, was formed on the sensors. PF4 of 20 µg/ml was covalently immobilized on the SAM layer for 15 min before blocking free aldehyde (-CHO) groups with 1M ethanolamine for 1 hour. After rinsing with PBS, 2.5 µg/ml SP was added for binding with PF4 on the chip at a pumping speed of 100 µL/min with an incubation time of 10 minutes. The real-time resonant frequency change was recorded at the third overtone due to stability constraints at a higher order. QCM real-time resonant frequency changes were observed on Qsoft software (version 2.5.22.707, Q sense, Biolin Scientific, Europe) and analyzed using the Sauerbrey equation 39 through Qtools software (version 3, Quantum Design, Darmstadt, Germany).

Dynamic light scattering (DLS)
DLS experiments were performed using our previous protocol testing with HIT antibodies. 15,40,41 To test if SP cluster PF4, SP up to 40 µg/ml was titrated in a cuvette containing PF4 of 20 µg/ml in water. The changes in the hydrodynamic diameter of proteins were measured in water at 20ºC and light scattering was detected at 173º using 15 repetitions using the Zetasizer Nano-S system (Malvern Instruments Ltd., Malvern, UK). The change in the surface zeta potential of PF4 was determined by titrating SP concentration up to 10 µg/ml in a folded capillary zeta cell containing PF4 of 20 µg/mL in water (pH7.4, conductivity of 0.318 mS/cm). The migration speed in an electric field was assessed with DLS for 10 repetitions.
Data analysis was performed using the Zetasizer software, version 7.11 (Malvern Instruments Ltd., Malvern, UK) and Origin software, version 7.5.

Single-molecule force spectroscopy (SMFS)
We applied our standard protocol to functionalize cantilevers and solid surfaces. 15,40,42 Briefly, both cantilevers with spring constants of 30 pN and 6 pN (Biolevers, Olympus, Japan) and glass substrates were first cleaned and coated with 1 mg/mL alpha-thio-omega-carboxy poly(ethylene glycol) (HS-PEG-COOH, PEG, MW= 3400 Da, Iris Biotech GmbH, Marktredwitz, Germany) before the activation of -COOH groups for 1h in an amine coupling kit solution (EDC:NHS, 1:1 volume, Biocore, Uppsala, Sweden). Subsequently, the cantilevers were incubated at RT for 1h in SP of 70 µg/ml (Biozol, Munich, Germany). The solid surfaces were incubated overnight at 4 o C with PF4 (20 µg/ml) or PF4/H complexes formed by 20 µg/mL PF4 and 0.5 U/ml UFH (Iduron Ltd., Manchester, UK). SMFS measurements were carried out in PBS with the JPK NanoWizard 3 (Berlin, Germany) with 900 force curves for each condition. The binding forces were collected using the JPK data processing software (version 4.4.18+). The mean rupture force values and their corresponding errors were determined by applying Gaussian fits to the data using Origin software (version 8.6).

Enzyme-linked immunosorbent assay (ELISA)
PF4/SP ELISA was performed by coating PF4 of 20 µg/ml in DPBS overnight at 4 o C on a 96well plate and then blocked with 7.5% goat serum as described for HIT ELISA. 43 After rinsing five times with DPBS, 100 µl SP of 10 µg/ml was incubated on PF4 coated plates for 1h at RT for binding. After removal of unbound SP, anti-SP antibody conjugate (Biozol, Germany), 1:500 dilution in DPBS was added for 1 h at RT. Wells were washed with DPBS and 100 μl anti-mouse IgG HRP (1:50000 dilution) were incubated for 1 h at RT. Wells were washed five times with DPBS and 100 TMB solution was added for 5 min before stopping the reaction by 100 μl H2SO4. Plate was ready to be measured at 450 nm absorbance for optical density (OD).

Isolation of human platelets
Human blood from healthy donors who were drug-free within the previous 10 days was collected into a tube of ACD-A 1.5 mL BD-Vacutainer (Fresenius Kabi, Bad Homburg, Germany) as previously described. [44][45] Platelet-rich plasma (PRP) was first obtained by centrifugation at 120 g for 20 min at room temperature. To isolate platelets, PRP in the presence of 15% acidcitrate dextrose (ACD-A, Fresenius Kabi, Germany) and 2.5 U/ml Apyrase (grade IV SIGMA, Munich, Germany) was centrifuged at 650g for 7 min. The platelet pellet was resuspended in buffer pH 6.3 composed of 137 mM NaCl, 2.7 mM KCl, 11.9 mM NaHCO3, 0.4 mM Na2HPO4, 2.5 U/ml Hirudin, and incubated for 15 min at 37°C before centrifuging at 650g for 7 min.
Platelet pellets were again carefully resuspended in suspension buffer and adjusted to a concentration of 300 x 10 9 /L using a blood counter (pocH-100i, SYSMEX, Germany) and then incubated for 45 min at 37°C before use.

Confocal laser scanning microscopy (CLSM) imaging of platelet morphology
The heparin-induced platelet aggregation (HIPA) tests were performed as previously described for HIT antibodies. 15 Briefly, 75 µL of washed platelets (300,000 platelets/µl) was incubated with 20 µL SP of 10 and 50 µg/ml concentrations (final) either without, with low (0.2 U/ml) or high (100 U/ml) unfractionated heparin concentration. Samples were added to microtiter plates and incubated on a magnetic stirrer (1000 rpm) for 45 minutes at RT.
For imaging with CLSM, samples were transferred to Petri dishes and kept at RT for 15 min before adding 4% PFA for fixation for 20 min at RT. 45 Anti-CD42a Alexa 647 antibodies (1:1000 dilution) were added to the immobilized platelets and incubated overnight at 4°C. Unbound antibodies were removed by rinsing with PBS before imaging by the CLSM (LSM710, Carl Zeiss, Gottingen, Germany) at RT in the dark. ImageJ and Origin (version 8.6) software were used to further process the images and to quantify the fluorescence intensity detected on each sample.
To minimize PF4 release, platelet aggregation tests were incubated at 37 o C for 1h in static condition without stirring in the presence of SP concentration up to 50 µg/ml. The PF4:SP concentration ratio that induces platelet aggregation was determined by titrating PF4 up to 30 µg/ml into a platelet sample (300,000 platelets/µl) containing 10 µg/ml SP.

SARS-CoV-2-SP clusters PF4
As SARS-CoV-2 spike protein 46 composes of S1 and S2 subunits (Fig. 1A, top), we first compared characteristics of these two subunits by tracking their hydrodynamic sizes and surface zeta potential (Fig. 1A, bottom). The S2 subunit (red box) has a zeta potential around zero and hydrodynamic size of ~250 nm while these are about -5 mV for the S1 subunit, indicating a high tendency of forming an aggregation of S2 subunits. The S1 subunit shows a negative zeta potential of ~5 mV and the hydrodynamic size of ~50 nm, indicating a low tendency of aggregation of these proteins. We have previously shown that most HIT antibodies which have surface potential around zero are active and induce platelet aggregation. 40 Thus, the S2, which is simplified by SP, was selected for further investigation.
We then mixed SP with PF4 and identified variation in hydrodynamic size and zeta potential of the mixture using Dynamic light scattering (DLS) (Fig. B). The theoretical size of S2 subunit (MW = 115 kDa, <10 nm) and PF4 tetramer (MW = 32 kDa, <5nm) are smaller than the experimental ones (Fig. 1C). This is because hydrodynamic size does not reflect the real size of proteins. The low ionic concentration medium produced an extended double layer of ions 47 around the molecules, which reduced the diffusion speed resulting in a larger hydrodynamic diameter. Furthermore, owing to the highly hydrophobic characteristics of PF4 molecules, they tend to come closer to each other in many buffers as previously described. 41 When titrating SP into a cuvette containing PF4 (20 µg/ml), an increase in the size of the PF4/SP complex up to 2.5 µg/ml SP was observed, and at 40 µg/ml SP the size of particles was significantly reduced. This indicates that the maximal size of PF4/SP complexes was formed at PF4:SP molar ratio of 28.8 (20 µg/ml PF4: 2.5 µg/ml SP). The binding of SP with PF4 was also observed by changes in surface zeta potential (Fig. 1D). The SP has a surface zeta potential ~zero while PF4 has a ring of positively charged lysine 48 that exhibits a zeta potential around +6 mV. When adding SP of various concentrations to PF4 (20 µg/ml), the zeta potential of the mixture switched to negative values. This indicated that the PF4/SP complexes are not formed by a simple electrostatic interaction like in the case of the interaction between PF4 and heparin, in which heparin neutralizes the positive charges of PF4. 49 We have previously shown that heparin brought two hydrophobic PF4 molecules together due to the neutralization of surface charges on PF4 by heparin, causing a merge of multiple hydrophobic PF4 molecules and changed the conformation of PF4. 50 As SP protein exhibits a mixture of complex surfaces with different properties, 31 it is plausible that PF4 first binds to the SP region of negatively charged residues and then undergoes conformational changes by interacting with the hydrophobic or hydrophilic residues of SP that switched the surface zeta potential of PF4 from positive to negative values. Thus, the SP might also bring PF4 closely to each other which causes a merge of multiple PF4 molecules in a similar way heparin does, resulting in large conformational changes in PF4s. When heparin was added, the size of the complexes was strongly reduced, indicating the ability of heparin in inhibition of the binding between PF4 and SP. These binding features of SARS-CoV-2-SP are somewhat similar to those of HIT autoantibodies with PF4 molecules. 27 We further utilized quartz crystal microbalance (QCM), which is a label-free and highly sensitive technique, to detect binding of SP (S2) with PF4 based on additional mass or frequency changes on a real-time basis (Fig. 1E). QCM spectra showed a strong frequency shift (Fig. 1F) or high mass change (Fig. 1G) when SP of 2.5 µg/ml was added to PF4 coated sensor. These results suggested that there was a binding reaction between SP and PF4.

Binding force between SARS-CoV-2 SP and PF4
We next determined the binding force of interactions between PF4 and SP using singlemolecule force spectroscopy (SMFS) based atomic force microscopy. A single SARS-CoV-2-SP was immobilized on a cantilever while PF4 molecules were coated onto a glass substrate ( Fig. 2A) (Fig. 2A). The magnitude of the binding force of SP with PF4 is similar to that with PF4/H (Fig. 2B, C).
Even though PF4/H complexes are larger and may offer more binding sites than PF4 alone, the contact area of the immobilized SP to these molecules may not significantly different, and therefore, did not induce a strong increase in force. However, PF4 slightly showed higher variation in both force (32.4  7.4 pN) and frequency of interaction (274.7  189.1 counts) than PF4/H complexes (32.9  5.5 pN and 247.8  136.2 counts) (Fig. 2C, D). Additionally, at >50 pN, SP showed higher interaction when interacting with PF4/H complexes than with PF4 alone. This difference is caused by an optimal exposure of binding epitope in PF4/H complexes while various forms of PF4 including monomer, dimer, trimer, and tetramer lead to the large variation of binding force with SP. Consistent with the above results obtained by biophysical methods, SP also showed a strong binding when interacting with PF4 immobilized on a 96-well plate as tested in ELISA (Fig. 2E).
Quantification of all particle sizes (Fig. 3H) or particles with size >20 µm 2 (Fig. 3I) showed the largest size and highest frequency (%) of formation of large particles at 50 µg/ml SP.
In HIPA experiments, PF4 molecules were released from platelet -granules while stirring the platelet samples. Thus, the results reflect that upon binding, the resulting PF4/SP complexes induced platelet aggregation. To clarify if SP induces platelet aggregation only in the presence of PF4, we next minimized PF4 release by incubating washed platelets with SP at static conditions and quantified the size of platelets (Supplementary Fig. 1). Some platelets with a larger size than their native size were observed at 10 µg/ml SP whereas higher SP concentration reduced platelet extent (Fig. 4A). Quantification of particle sizes showed the highest frequency of particles >9 µm 2 at 10 µg/ml, but it reduced gradually as SP concentration increases (Fig. 4B). In this condition, the unknown concentration of PF4 released from washed platelets did not allow us to determine the relationship between PF4:SP concentration and platelet morphology. To further understand if PF4 together with SP at a defined concentration affects platelets, we titrated PF4 up to 30 µg/ml into washed platelets in the presence of 10 µg/ml SP. We observed that the frequency of platelets (%) at size >5µm 2 significantly increased at PF4 concentration between 10 and 20 µg/ml (Fig. 4C). In these experiments, we observed a clear effect of SP on platelet's state and PF4 played an important role in inducing platelet aggregation, especially at a certain PF4:SP molar concentration ratio (between 3.6 and 7.2). We further used a P-lectin maker (CD62P) to track if SP induces platelet activation by flow cytometry (Supplementary Fig. 2). SP without PF4 induced an increase of platelet activation at higher concentrations and further enhanced activation in the presence of PF4 but also followed a concentration-dependent manner. This indicates a synergy effect between SP and PF4 in activating platelets. It is, therefore, necessary to detect not only PF4-antibodies but also PF4:SP concentration ratio to identify a better treatment strategy for patients suspected of thrombosis thrombocytopenia.

Discussion
By utilizing multiple analytical techniques, we found that SARS-CoV-2 SP clusters PF4. It induced conformational changes in PF4 and switched the surface zeta potential of PF4 from positive to negative values. The formed PF4/SP complexes induced platelet aggregation in a concentration-dependent manner.
From our observation, we propose a mechanism for SARS-CoV-2-induced thrombocytopenia in Covid-19 patients (Fig. 5). The binding of the SARS-CoV-2 virus to positively charged PF4 molecules forms antigenic PF4/virus complexes (Fig. 5A). In the HIT system, HIT antibodies are produced due to immune response to the ultra-large antigenic PF4/Heparin complexes which can be up to several micrometers in size. 51,41 As the SARS-CoV-2 with a size of 60-140 nm, 27 the cross-linking among PF4 molecules and viruses can develop large enough immune complexes that on one hand mediate/activate platelets (Fig.   5B) and on the other hand promote B-cells to producing PF4-antibodies in a similar way ultra-large antigenic PF4/H complexes do (Fig. 5C). Consistently, recent reports showed that platelet-activating PF4-antibodies are detected in Covid-19 infected patients. 9,22 Previously, we reported that only a low concentration of PF4-antibodies or HIT autoantibodies (~5 µg/ml) induced massive platelet activation even in the absence of heparin. 27 These antibodies activate human platelets, monocytes, and neutrophils. They also bind to endothelial cells through PF4/polyanion complexes, 14,35 inducing a frequent immune-mediated adverse drug reaction in form of the life-threatening autoimmune HIT. 52 The synergic connection between the PF4/Virus in Covid-19 patients and PF4-antibodies can cause unusual thrombotic thrombocytopenia (Fig. 5D). This can be the mechanism for excessive procoagulant activity observed in Covid-19 patients. However, the degree of complication may occur differently depending on the amount of PF4 and SP in individuals with unique biological backgrounds.
In addition, we suggest that the interaction of SP and PF4 offers a possible underlying mechanism for a high rate of clotting effects in vaccinated patients. The Ad26.COV2.S and  53 These two latter vaccines require an additional step to produce mRNA from DNA. 54 In this process, transcription of wildtype and codon-optimized spike open reading frames enables alternative splice events that lead to C-terminal truncated, soluble/partial spike proteins that do not contain a necessary part for adhering to cells as they should. Therefore, these soluble spike proteins can not trigger immune cells to produce anti-SP antibodies but they can enter the blood circulation, forming complexes with PF4 that trigger the formation of PF4-antibodies ( Fig. 6). When testing VITT patients sera in HIPA, Greinacher et al found that these sera increased platelet activation overtime counted from the first day of being vaccinated with AstraZeneca vaccine. 21 These results suggest that the concentration of the soluble SP increase overtime. However, the authors also found that VITT occurs only from day 5 after vaccination which is consistent with our finding that the platelet activation occurs only when a critical PF4:SP concentration ratio reaches. In addition to soluble SPs, cell fragments carrying SPs that are developed by cell disrupting/dying during the immune response to vaccines. These SP carriers can also cluster PF4, forming antigenic PF4/SP-cell-fragment complexes that also adhere and activate platelets as well as trigger the development of PF4antibodies.
Furthermore, the adenoviruses have an isoelectric point at ~pH 6.2, 55-56 meaning that they have a neutrally charged surface at this pH. However, by entering blood circulation (pH between 7.35 and 7.45), the neutral surfaces of adenoviruses can be switched to negative net charges. We speculate that in some vaccinated individuals with rare biological characteristics, a high amount of PF4 is secreted before and further enhanced after vaccination. High density together with positively charged residues of PF4 can facilitate binding with the negative net charges of adenovirus vaccine, forming antigenic PF4/vaccine complexes. Consistently, the binding of PF4 to the Astrazeneca vaccine has been observed in vitro. 57 If this vaccine enters the blood circulation, the formed PF4/vaccine can also trigger the production of PF4antibodies.
Together, these factors strongly and quickly induce VITT. As different factors including PF4/SP, PF4/vaccine complexes, and PF4/SP-cell-fragment can trigger the production of PF4-antibodies after vaccination, a high concentration of these antibodies will be developed. Consistently, it has been shown that high OD ~3 was obtained in PF4/Heparin ELISA when tested with VITT patients sera 21 while HIT sera show OD~2 units 58 .
The soluble SPs are an add-in complication to the basic side effect induced by mRNA-based vaccines. The mRNA-based vaccine works simply by entering the cell and produces SP that diffuses and adheres on cells membrane following by the production of antispike protein antibodies (Fig. 6). This explains the higher frequency of thrombotic thrombocytopenia after vaccination with vector-based vaccine as compared with mRNAbased vaccines. The frequency of side effects, however, depends on the efficiency of the vaccines in producing spike proteins as we observed a reduction of platelet aggregation at too high or too low SP concentration.
Although the formation of PF4/SP complexes induces platelet aggregation, the outcome of the Covid-19 patients differs from vaccinated individuals. In Covid-patients, the uncontrollable amount of SP on viruses mainly dominates in the lung, and therefore, the formation of blood clots in this organ blocks the airway that induces severe acute respiratory syndrome leading to death. In vaccinated individuals, the blood clots are mainly formed at the site of injection and are rarely found in other organs. Thus, severe side effects in vaccinated individuals are rare depending on the location of the blood clots. This also explains why the severely ill rate in vaccinated individuals is much lower than in Covid-19 patients.
As PF4/SP complexes also induce platelet aggregation, we suggest an additional consideration to the guidance from the GTH 59 and ISTH 60 for diagnostic and therapeutic algorithm in patients with the vaccine (AstraZeneca)-induced thrombocytopenia (VITT). If serum is negative in the screening test for HIT antibodies but positive in HIPA/SRA, the critical PF4:SP concentration ratio that forms platelet-activating PF4/SP complexes is reached. In this case, heparin or even high-dose heparin treatment is possible (Fig. 7).    After being vaccinated with mRNA-based vaccine (blue circle) or vector-based vaccine (orange), spike proteins translated from cells will be transported to the cell plasma membrane for the production of anti-spike protein antibodies. However, an unexpect process during transcription of wildtype and codon-optimized spike open reading frames enables alternative splice events that lead to C-terminal truncated, soluble/partial spike proteins that do not contain nessessary part for adhering to cells as they should (gray rectangular). These soluble spike proteins can not trigger immune cells to produce anti-SP antibodies but they can then enter the blood circulation and form immunogenic complexes with PF4.  Figure 7. Suggested additional consideration (blue circle) for diagnostic and therapeutic algorithm in patients with VITT to the guidance from the GTH (gray box). 59 (blue) We recommend if serum is negative in screening test for HIT antibodies but positive in HIPA/SRA, the risk ratio of PF4:SP is reached, and therefore, heparin (even high dose heparin) treatment is possible.