Variant-derived SARS-CoV-2 spike protein does not directly cause platelet activation or hypercoagulability

Abstract

Thrombosis has been associated with severity and mortality in COVID-19. SARS-CoV-2 infects the host via its spike protein. However, direct effects of spike proteins from SARS-CoV-2 variants on platelet activity and coagulability have not been examined. An ethically approved ex vivo study was performed under a preplanned power analysis. Venous blood was collected from 6 healthy subjects who gave prior written consent. The samples were divided into 5 groups: without spike proteins (group N) and with spike proteins derived from alpha, beta, gamma, and delta SARS-CoV-2 variants (groups A, B, C, and D, respectively). Platelet aggregability, P-selectin expression, platelet-associated complement-1 (PAC-1) binding, platelet count, and mean platelet volume (MPV) were measured in all 5 groups, and thromboelastography (TEG) parameters were measured in groups N and D. The % change in each parameter in groups A to D was calculated relative to the value in group N. Data were analyzed by Friedman test, except for TEG parameters, which were evaluated by Wilcoxon matched pairs test. P < 0.05 was considered significant. This study included 6 participants based on a power analysis. There were no significant differences in platelet aggregability under stimulation with adenosine diphosphate 5 µg/ml, collagen 0.2 or 0.5 µg/ml, and Ser-Phe-Leu-Leu-Arg-Asn-amide trifluoroacetate salt (SFLLRN) 0.5 or 1 µM in groups A–D compared to group N. There were also no significant differences in P-selectin expression and PAC-1 binding under basal conditions or SFLLRN stimulation, and no significant differences in platelet count, MPV and TEG parameters. Platelet hyperactivity and blood hypercoagulability have been reported in COVID-19 patients, but spike proteins at 5 µg/ml from SARS-CoV-2 variants (alpha, beta, gamma, delta) did not directly cause these effects in an ex vivo study. This study was approved by the Ethics Committee of Kyoto University Hospital (R0978-1) on March 06, 2020.

Supplementary Information

The online version contains supplementary material available at 10.1007/s10238-023-01091-4.

Background

COVID-19 has caused many infections and deaths worldwide [1]. New infections and deaths still occur due to the appearance of viral variants and particularly variants of concern (VOCs). The viral spike protein is the main difference among COVID-19 variants. SARS-CoV-2 uses its spike protein to bind to proteins on human cells and invade these cells [2].

COVID-19 is a systemic disease that affects multiple organs, including the hematopoietic system, which causes blood hypercoagulability [3–12]. Thrombosis in COVID-19 patients is common [11–16] and is associated with disease severity and risk of death [17]. COVID-19 patients showed a high incidence of thrombotic complications despite anticoagulation [12–14, 16], and thus hypercoagulation in COVID-19 and COVID-19 thrombosis seem to be related.

Zhang et al. [18] conducted an ex vivo study and showed that the spike protein from a wild-type strain directly promotes platelet activation. Nowadays, wild strain SARS-CoV-2 has been replaced by variant viruses [19]; however, there is no information on the direct effects of spike proteins in SARS-CoV-2 variants on platelet activity and coagulability, which have been reported for the wild-type strain [18]. Therefore, we assessed effects on platelet activation and coagulability of spike proteins from four SARS-CoV-2 variants, alpha (B.1.1.7), beta (B.1.351), gamma (P.1), and delta (B.1.617.2), which were all designated as VOCs by the World Health Organization; at the time, the study was planned.

Methods

Subject selection

This study was approved by the Ethics Committee of Kyoto University Hospital (approval number R0978-1) and was conducted according to the guidelines of the Declaration of Helsinki. We included healthy volunteers who met the following three criteria: (1) healthy adults aged 20 to 65 years, (2) not taking blood coagulation-related medication, and (3) no blood coagulation-related diseases. Prior written informed consent was obtained from all subjects.

Measurement groups

Subjects with spike proteins derived from alpha, beta, gamma, and delta variants were classified as groups A, B, C, and D, respectively. The group without spike protein is group N.

Materials and processing

Drugs and equipment, washed platelet preparation, and platelet-stimulating agent processing are described in Electronic supplementary material 1, and the details of the structure and stability of the spike proteins used in the experiments are included in Electronic supplementary material 2.

In determining the concentration of spike protein to use in experiments, we first referred to the findings in Zhang et al. [18] of concentration-dependent enhancement of platelet aggregation using 0–2 µg/ml of spike protein from a wild-type strain and of enhancement of P-selectin expression by 2 µg/ml spike protein (the maximum concentration used in an aggregation assay). We hypothesized that a higher concentration of 5 µg/ml would further enhance platelet function, and we conducted a preliminary study to examine platelet aggregability and P-selectin expression using spike protein from a wild-type strain at 0, 1, 2 and 5 µg/ml (Table S1, Figs. S1 and S2 [Electronic supplementary material 3]). In this preliminary study, platelet aggregability tended to increase with the concentration of spike protein and appeared to be maximized at 5 µg/ml. Thus, we chose to use 5 µg/ml of spike proteins in the main study.

Spike proteins from SARS-CoV-2 variants (alpha, beta, gamma, delta) were diluted to 100 µg/ml in phosphate-buffered saline (PBS), divided into small portions, combined with 0.1% w/v bovine serum albumin, and refrozen at − 80 °C. Just before the start of experiments, the spike protein was thawed and excess was discarded without refreezing.

Measured parameters

In the current study, the same experimental techniques were used as those in our studies of the effects of drugs on enhancing and inhibiting platelet function [20, 21].

Platelet aggregability

Spike proteins were added to washed platelets and left to stand at 37 °C for 5 min. Light transmittance rates were then measured for 10 min under adenosine diphosphate (ADP) 5 µM, collagen 0.2 or 0.5 µg/ml, and Ser-Phe-Leu-Leu-Arg-Asn-amide trifluoroacetate salt (SFLLRN) 0.5 or 1 µM stimulation. The control was purified water (100% light transmittance). Two consecutive measurements were made, and the average of the maximum values was used as the result (details in Electronic supplementary material 1).

P-selectin expression

Spike proteins were added to washed platelets and left to stand at 37 °C for 5 min. The mixture was then added to PBS ± SFLLRN (final concentration 1 µM), left to stand at 22 °C for 15 min, and then fixed at 4 °C (details in Electronic supplementary material 1). Centrifugation at 4 °C and 1,600 g for 15 min and washing with PBS were repeated twice, followed by addition of 1 µl each of anti-CD61 antibody and anti-CD62P antibody. Flow cytometry was performed as previously reported [20] after standing in the dark at 22 °C for 1 h. Mean fluorescent intensity (MFI) of CD62P-positive platelets among 10,000 CD61-positive platelets was calculated under basal conditions (no platelet stimulant) or SFLLRN stimulation.

Platelet-associated complement-1 (PAC-1) binding

Spike proteins were added to whole blood after collection, and the sample was left to stand at 37 °C for 5 min. A total of 2.5 µl of the blood plus spike proteins were added to 5 µl of anti-CD61 antibody, 5 µl of anti-PAC-1 antibody, and 37.5 µl of PBS with or without SFLLRN (final concentration 1 µM), left to stand in the dark at 22 °C for 15 min, and then fixed at 4 °C (details in Electronic supplementary material 1). MFI of PAC-1-positive platelets among 10,000 CD61-positive platelets was calculated under basal conditions or SFLLRN 1 µM stimulation.

Platelet count, mean platelet volume (MPV) and thromboelastography (TEG) parameters

Spike proteins were added to whole blood and left to stand at 37 °C for 5 min before testing. Due to the large and costly amount of spike protein required for measurements with TEG 6s, we compared TEG parameters in groups N and D only. TEG 6s represents platelet and fibrinogen function using four main parameters: R, K, MA, and LY30. A global hemostasis cartridge contains four reagents: CK, CKH, CRT, and CFF. The following five parameters were measured: R-CK, K-CK, MA-CK, MA-CFF, and MA-CRT minus MA-CFF. R-CK, K-CK, and MA-CK are the R, K, and MA values measured using CK; MA-CFF is the MA value measured using CFF; and MA-CRT minus MA-CFF is the difference between MA values measured using CRT and CFF. R is the time until clot formation starts; K is the rate of clot formation; and MA is the maximum clot strength. CK contains kaolin, a coagulation promoter; CRT contains kaolin and tissue factor, which further enhances the clotting reaction; and CFF contains kaolin, tissue factor, and glycoprotein IIb/IIIa receptor antagonist, which inhibits platelet participation in coagulation and allows assessment of fibrinogen-only clot strength. CRT minus CFF indicates the platelets-only clot strength.

Statistical analysis

For each test in each subject, the percentage changes of values in groups A, B, C and D were calculated using group N as a reference: value in groups A, B, C, or D/value in group N × 100 (%). Results are presented as medians and quartiles. Platelet aggregation rate, P-selectin expression, PAC-1 binding, and blood count were analyzed by Friedman test. TEG parameters were analyzed by Wilcoxon matched pairs test. Both tests were performed with Prism 9 for macOS, ver. 9.3.1 (GraphPad Software, San Diego, CA, USA). To ensure sufficient power, we performed a power analysis. With a difference of 1.5 times, the standard deviation as clinically significant, power of 0.80, and an α level of 0.05, 6 subjects were estimated to be necessary. A two-sided α < 0.05 was considered to be significant. All data are shown in Tables S2-5 [Electronic supplementary material 4].

Results

The current study included 6 participants based on a power analysis. The subjects were voluntarily asked to provide the date of their last COVID-19 vaccination prior to blood collection, and all were found to have been vaccinated 5 to 7 months after their second Pfizer mRNA vaccination. Platelet aggregability and P-selectin expression were measured using washed platelets without leukocytes to eliminate the effects of the vaccine. PAC-1 binding, platelet count, MPV, and TEG were measured using whole blood with leukocytes. The % granulocytes after adding spike protein was not low, which suggests that the neutrophil activity of the collected blood was sufficient (Table S6 [Electronic supplementary material 4]). Medians and quartiles for % changes in groups A to D relative to group N are shown in Figs. 1, 2, 3, 4, 5.

Fig. 1.

Maximum platelet aggregation rates shown as % changes relative to group N. Medians and quartiles are shown on the graphs. There were no significant differences among the variants with any of the platelet stimulants. a Aggregation rates under ADP 5 µM stimulation (n = 6). b Aggregation rates under collagen 0.2 or 0.5 µg/ml stimulation (n = 5). c Aggregation rates under SFLLRN 0.5 or 1 µM stimulation (n = 5). ADP: adenosine diphosphate; SFLLRN: Ser-Phe-Leu-Leu-Arg-Asn-amide trifluoroacetate salt

Fig. 2.

MFI of P-selectin shown as % changes relative to group N under basal conditions. SFLLRN 1 µM was used as the platelet stimulant. Medians and quartiles are shown on the graphs. There were no significant differences among the variants with no stimulation or SFLLRN 1 µM stimulation (n = 6). MFI: mean fluorescent intensity; SFLLRN: Ser-Phe-Leu-Leu-Arg-Asn-amide trifluoroacetate salt

Fig. 3.

MFI of PAC-1 shown as % changes relative to group N under basal conditions. SFLLRN 1 µM was used as the platelet stimulant. Medians and quartiles are shown on the graphs. There were no significant differences among the variants with no stimulation or SFLLRN 1 µM stimulation (n = 6). PAC-1: platelet-associated complement-1; MFI: mean fluorescent intensity; SFLLRN: Ser-Phe-Leu-Leu-Arg-Asn-amide trifluoroacetate salt

Fig. 4.

Blood counts shown as % changes relative to group N. Medians and quartiles are shown on the graphs. There were no significant differences in platelet count and MPV among the variants (n = 6). MPV: mean platelet volume

Fig. 5.

TEG parameters shown as % changes in group D relative to group N. Medians and quartiles are shown on the graphs. The differences between groups N and D were not significant (n = 6). a R is the time until clot formation starts and CK is a reagent containing kaolin. R-CK indicates R values measured using CK and refers to the rate of coagulation initiation. b K is the rate of clot formation. K-CK indicates K values measured using CK and refers to the rate of coagulation. c MA is the maximum clot strength. MA-CK indicates MA values measured using CK and refers to the blood clot strength. d CFF is a reagent containing kaolin, tissue factor, and glycoprotein IIb/IIIa receptor antagonist. MA-CFF shows MA values measured using CFF and indicates the fibrinogen contribution to MA-CK. e CRT is a reagent containing kaolin and tissue factor. MA-CRT minus MA-CFF is the difference between MA values measured using CRT and CFF, and indicates the platelet contribution to MA-CK. TEG: thromboelastography

Platelet aggregability

Platelet aggregability was measured with ADP (5 µM), collagen (0.2 or 0.5 µg/ml), or SFLLRN (0.5 or 1 µM) stimulation (Fig. 1). The aggregation rate for each group with the aggregation rate of group N set at 100% did not change significantly with any of the platelet-stimulating agents (p = 0.69 for ADP 5 µM, p = 0.37 for collagen 0.2 µg/ml, p = 0.31 for collagen 0.5 µg/ml, p = 0.28 for SFLLRN 0.5 µM, p = 0.63 for SFLLRN 1 µM).

P-selectin expression, PAC-1 binding, platelet count, and MPV

For P-selectin and PAC-1, the MFI of each group did not change significantly under either condition with the MFI of group N under basal conditions set at 100% (p = 0.78 for P-selectin under basal conditions, p = 0.50 for P-selectin under SFLLRN 1 µM stimulation, p = 0.92 for PAC-1 binding under either condition) (Figs. 2, 3). There were also no significant differences in platelet count (p = 0.56) and MPV (p = 0.11) (Fig. 4).

TEG

Differences between groups N and D were not significant (p = 0.44 for R-CK, p = 0.63 for K-CK, p > 0.99 for MA-CK, p = 0.84 for MA-CFF, p = 0.44 for MA-CRT minus MA-CFF) (Fig. 5).

Discussion

This is the first study to assess the direct effects of spike proteins from four SARS-CoV-2 variants on human platelets and coagulability. Contrary to results for the spike protein from a wild-type strain [18], addition of spike proteins from the SARS-CoV-2 variants had no effects on platelet aggregability, platelet activity, platelet count, MPV, or TEG parameters in this ex vivo study.

Most studies of platelet function and/or blood coagulability in COVID-19 have used blood from COVID-19 patients [3–12, 22–32]. Hence, it has been unclear whether the increased platelet function and coagulability in these patients are due to SARS-CoV-2 or the spike protein itself, or to indirect effects triggered by SARS-CoV-2 infection, such as systemic inflammation. Increased platelet activity and coagulability have been found upon reaction of spike proteins from COVID-19 wild-type strain with human platelets [18], but similar effects in variants have not been examined, which indicates the novelty and relevance of the current study.

COVID-19 is known to cause blood hypercoagulability [3–12], which considered to be influenced with a high incidence of thrombotic complications despite anticoagulation in COVID-19 patients [12–14, 16]. However, there is inconsistency among studies on platelet function and coagulability in COVID-19. Several reports [22–24] have suggested that COVID-19 patients have significantly increased aggregability compared to healthy subjects. However, in contrast, Bertolin et al. [6] found no difference in ADP-stimulated aggregability in patients compared to healthy subjects, and Heinz et al. [7] found rather low aggregability in patients. Herrmann et al. [11] measured aggregability in critically ill patients over a 2-week period and found this to be well below baseline levels under ADP stimulation.

There have also been several reports of significantly increased P-selectin expression in COVID-19 patients compared to healthy controls [3, 18, 22, 25, 26], but some studies have found similar P-selectin expression in patients and healthy controls [6, 27]. Regarding PAC-1 binding, one report showed a significant increase in COVID-19 patients, and especially in severe COVID-19, compared to healthy controls [18], whereas another found a significant decrease [27]. For agonist-stimulated PAC-1 binding, it has been suggested that patients have increased expression compared to healthy subjects [18], but another study indicated that patients have decreased expression independent of severity [22].

For MPV, the conclusions among studies of COVID-19 patients are more consistent. Patients have higher MPV than healthy subjects [6, 18, 23] and critically ill patients have higher MPV than non-critically ill patients [18, 28–30], but rarely vice versa. The increase in MPV may be a hallmark of severity of COVID-19 and the increase in platelet count [23, 28–31]. However, there are many methods for MPV measurement, which makes it difficult to determine whether the MPV of an individual patient is normal or slightly elevated [32].

Blood viscoelasticity tests, regardless of the measuring device, have shown that COVID-19 patients have hypercoagulability (shortened clotting times, increased clot strength, and shortened fibrinolysis times) [5–11, 27]. There also seem to be greater effects in critically ill patients than in moderately ill patients [27, 33]. However, in these studies, the findings for hypercoagulation were seen for selected parameters [5–9, 27, 33] and specific patients [9, 10].

One reason for the wide variation in findings among studies is that platelet hyperactivation and blood hypercoagulation may be due to multiple factors and mechanisms, rather than to SARS-CoV-2 itself [34, 35]. Our results support this view, since neither platelet activity nor blood coagulability were altered by a change of the SARS-CoV-2 spike protein alone.

Since SARS-CoV-2 establishes infection by binding to the host angiotensin-converting enzyme 2 (ACE2) receptor [36] and thrombosis is common in COVID-19 [11–16], a relationship between SARS-CoV-2 binding to the ACE2 receptor and COVID-19 thrombosis is likely. In fact, many studies have shown that increased inflammatory cytokines in highly ACE2-positive organs and vascular endothelial cells can lead to a thrombogenic response and COVID-19 thrombosis [3, 10, 27, 34]. There are two major pathways for this mechanism: direct infection of the vascular endothelium via ACE2 on vascular endothelial cells, resulting in endothelial damage [34], and systemic ACE2 downregulation [37, 38]. SARS-CoV-2 infection of the vascular endothelium via ACE2 on endothelial cells damages the endothelium and causes platelet aggregation [34, 39], allowing thrombosis to occur [34, 40]. Systemic ACE2 receptor downregulation occurs in the renin-angiotensin system pathway, an important mechanism that maintains systemic sodium levels and causes vasoconstriction. This pathway is regulated by ACE2 to prevent excessive inflammation and thrombus formation [37]. Internalization [34] and downregulation [37, 38] of ACE2 by SARS-CoV-2 spike protein leads to elevation of inflammatory cytokines such as interleukin-1α, interleukin-6, and tumor necrosis factor-α, and endothelial biomarkers such as von Willebrand factor [41]. Hence, endothelial function is impaired, thrombus formation becomes uncontrolled, and tissue factors are released that amplify the platelet thrombotic response [40]. Thus, COVID-19 thrombosis can be explained by these two mechanisms of direct or indirect SARS-CoV-2 effects on ACE2 in organs or the endothelium.

This study has several limitations. First, the number of subjects was small, which may be one of the reasons for the lack of significant differences. However, the study was conducted under a power analysis using paired comparisons. Second, it is unclear whether the spike protein concentration used was appropriate: 5 µg/ml of spike protein may be too much or too little, and the concentration of spike protein in patients with severe COVID-19 is uncertain. Therefore, our findings should be confirmed or refuted using different concentrations of spike protein. However, we note that the concentration of spike protein was chosen with reference to a previous study [18] and our preliminary study (Electronic supplementary material 2). Third, the concentrations of platelet stimulants might have been too high in measurements of aggregation rate and platelet activation markers. Collagen- or SFLLRN-stimulated aggregability was measured at two concentrations, low and high, but there may have been other concentrations at which differences in results between the spike protein and control groups may have been observed. Fourth, extra platelet activation or consumption could have occurred during preparation of washed platelets, which may have affected the results. Fifth, the spike proteins used were not glycosylated, which could have increased protein instability during dilution and thawing, and prevented accurate responses. Sixth, platelet aggregability and P-selectin expression were measured using washed platelets without leukocytes. Therefore, it is unclear how cell-mediated immunity due to neutrophils, nitric oxide release, or production of related molecules, which may contribute to platelet function and coagulopathy, affects platelet aggregability and P-selectin expression. Seventh, the study was performed ex vivo and the effects of blood flow, blood vessels, shear stress on platelets, and interactions between platelets and other blood cells could not be evaluated. Eighth, we did not confirm platelet binding of spike protein, as shown in a previous study [18]. However, there is still little evidence for ACE2 on platelets [22, 23, 34]. In this regard, our main focus was whether spike proteins from variants have a direct impact on platelet function.

Within these limitations, our results add evidence that spike proteins derived from SARS-CoV-2 variants do not contribute to development of COVID-19 thrombosis.

Conclusions

Spike proteins (5 µg/ml) derived from SARS-CoV-2 variants (alpha, beta, gamma, delta) did not directly cause platelet hyperactivity or blood hypercoagulability in an ex vivo study.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

EK performed all experiments, analyzed all data, wrote the initial draft, prepared all figures and electronic supplementary materials, and edited the manuscript. YM edited the manuscript. SK designed the study and edited the manuscript. ME supervised the research and revised the manuscript.

Funding

This work was supported by a Grant-in-Aid for Scientific Research (KAKENHI) (21K09044).

Data availability

All data generated or analyzed during this study are included in the published article and electronic supplementary materials.

Declarations

Conflict of interest

The authors declare that they have no competing interests.

Ethical approval

The research protocol, written instructions to subjects, and consent form were approved by the Ethics Committee of Kyoto University Hospital (R0978-1).

Consent to participate

The study was carried out according to the guidelines of the Declaration of Helsinki. Prior written informed consent was obtained from all subjects.

Consent for publication

Not applicable.