The severe coagulation syndrome in numerous organs is the major life-threatening conditions characterizing the acute infection by SARS-CoV-2. Endothelial inflammation/dysfunction, platelet hyper-reactivity, generation of neutrophil extracellular traps, promote the activation of the coagulation cascade in an infection-dependent manner.1 The outcome of the COVID-19 vaccination campaign has also revealed a potential risk of vaccine-induced immune thrombotic thrombocytopenia (VITT), or thrombosis with thrombocytopenia syndrome (TTS).1 Although this risk is extremely low, the propensity of the Spike protein to induce inflammatory and coagulation factors in placental and endothelial cells (ECs)2,3 has called for caution during the SARS-CoV-2 vaccination campaign.
We recently described a potential function of Spike (S) as a co-factor for estrogen receptor alpha (ERα) nuclear signaling,4 mediated by interaction of a nuclear receptor co-regulator (NRC) LXD-like motif present on the viral protein S2 subunit, with the activation function 2 (AF-2) region on ERα. To clarify the relevance of this interaction for the expression/release of pro-coagulation factors by human ECs, we exposed the immortalized Ea.Hy926 EC line to increasing concentrations of wild-type (wt) S, 17β-Estradiol (ES; a natural ERα agonist) Tumor Necrosis Factor-alpha (TNFα; a pro-inflammatory cytokine involved in SARS-CoV-2 coagulopathy), Raloxifene and Fulvestrant (RAL, FS; two ERα inhibitors with a function as Selective Estrogen Receptor Modulator, SERM, and Degrader SERD, respectively). Incremental amounts of the S-protein increased the pro-coagulant activity (PCA) of the cells (Supplementary Fig. 1a, c) and the expression of Tissue factor (TF) mRNA (Supplementary Fig. 1b, d). The ability of the two ERα inhibitors to reduce cellular TF mRNA and viral secretion (measured by RT-PCR amplification of the ORF1ab transcript in the supernatant) was confirmed in real infection experiments performed on Ea.Hy926 cells with SARS-CoV-2 (Fig. 1a, b, Supplmentary Fig. S2).
The knowledge based on the EXSCALATE supercomputing platform4 guided us in introducing point mutations in the putative LXD-like motifs present in the S2 domain of Spike mimicking Nuclear Receptor Coactivator (NCOA) function. These regions—not related to known amino acid mutations reported in SARS-CoV-2 variants (Fig. 1c, d), were mutated into sequences (Sp5 and Sp7) with a potentially lower affinity for ERα, but maintaining a wild-type conformation. We first assessed the effect of the trimeric wt-S, Sp5, and Sp7 on the PCA and TF expression in ECs. The results showed a significant reduction of the PCA, and a decreased TF transcript and activation levels (Fig. 1e–h). Interestingly, TF expression and activity after treating cells with the two mutants were similar to the levels observed in RAL and FS treatments, thus validating the selectivity of the wt-S intracellular function via the interaction with ERα. These results were also validated in a line derived from human lung ECs (HULEC-5a), confirming the clinical relevance of our findings (Fig. 1i).
To exclude that interaction of other viral proteins with ERα accounts for the wt-S-induced TF upregulation, we generated a recombinant Vesicular Stomatitis Virus, in which the viral envelope glycoprotein was replaced by the wt-S or the Sp5/Sp7 mutants. Characterization of these pseudo-viruses (PSVs) by electron microscopy and preliminary infection in Ea.Hy926 cells indicated a similar content of the three variants (Supplementary Fig. 3), while TF mRNA was clearly expressed at lower levels in cells infected with PSVs containing the mutated Spike variants compared to wt-S (Fig. 1j). The wt-S and the Sp5/Sp7 coding sequences were also introduced into an expression vector designed to elicit immune responses similarly to expression systems employed in SARS-CoV-2 vaccination.5 Vectors were amplified and electroporated in the lower limb adductor muscles in mice. As shown in Fig. 1k, the three Spike variants were expressed at comparable levels, and this was accompanied by a comparable immunization of the mice against the S1 domain of the protein as assessed by ELISpot performed with splenocytes derived from the mice receiving the wt and the Sp5/Sp7 variants (Fig. 1l). Immunization was also confirmed by detection of anti-Spike antibodies in the plasma of the injected mice (Fig. 1m). To verify the coagulation response in vivo, the recalcification time as well as the levels of D-Dimer, von Willebrand Factor (vWF) and TF were measured in the plasma of control, wt-S, Sp5, and Sp7 electroporated mice. Results showed a transient increase in blood clotting at 48 h followed by a return to basal levels at 96 h (Fig. 1n, S4) in mice treated with wt-S, but not Sp5 and Sp7. The analysis of circulating coagulation markers characterizing the COVID-19 infection6 produced similar results (Fig. 1o–q). These data were finally validated by employing a disseminated thrombosis in vivo model, where thromboembolic death/paralysis after collagen/epinephrine injection was much lower in Sp5/Sp7-treated than in wt-S-treated mice (Fig. 1r).
Although vasculopathy consequent to COVID-19 depends on the damage/inflammation of the endothelium mediated principally by ACE-2, other mechanisms are emerging even not directly related to the infective activity of the virus. For example, sensing of the E protein by inflammatory cells causes a strong innate immunity reaction mediated by interaction with TLR2,7 and vasculopathy in the heart could be determined by disruption of the vascular homeostatic function of pericytes due to the activation of CD147 receptor signaling by the Spike S1 domain.8 By providing evidences that the activation of coagulative cascade might derive from the interaction of Spike with ERα, the present study establishes, for the first time, an important function of the viral protein as a co-factor for induction of TF, and of other coagulation effectors (D-Dimer, vWF) in COVID-19.6 Various mechanisms could account for this elevation according to various possibilities. For example, it is possible that after internalization of the viral protein, the Spike/ER complex translocates into the nucleus to directly activate transcription of ER-dependent targets. A second possibility is that the pro-coagulation effects of Spike occur mostly via the membrane-bound ER, and thus via a second-messenger mechanism. While this extranuclear function of Spike on activation of the coagulation cascade via ERα could be reconciled with what found recently on the release of vWF by endothelial cells, mediated by an ACE-2 independent interaction with the cytoskeleton-associated protein-4,9 it does not explain the TF transcriptional upregulation observed in our experiments.
In summary, also corroborated by mounting evidences from clinical and in vitro studies showing the potential effectiveness of SERMs (e.g., Raloxifene) as an anti-viral agent in COVID-19,10 our results suggest a new non-infective pathologic action of the S-protein at the vascular level enhancing the endothelial pro-coagulation activity. Given the residual risk of coagulopathy observed in subjects treated with COVID-19 vaccines, our study indicates two variants of the original Spike sequence that could be employed to design new versions of COVID-19 vaccines lacking any residual risk of VITT in the still ongoing vaccination and boosting campaign.
The raw data introduced in the present manuscript are available upon reasonable request to the corresponding Authors.
Iba, T. & Levy, J. H. Thrombosis and thrombocytopenia in COVID-19 and after COVID-19 vaccination. Trends Cardiovasc. Med. 32, 249–256 (2022).
Guo, X. et al. Regulation of proinflammatory molecules and tissue factor by SARS-CoV-2 spike protein in human placental cells: implications for SARS-CoV-2 pathogenesis in pregnant women. Front. Immunol. 13, 876555 (2022).
Bhargavan, B. & Kanmogne, G. D. SARS-CoV-2 spike proteins and cell-cell communication inhibits TFPI and induces thrombogenic factors in human lung microvascular endothelial cells and neutrophils: implications for COVID-19 coagulopathy pathogenesis. Int. J. Mol. Sci. 23, 10436 (2022).
Solis, O. et al. The SARS-CoV-2 spike protein binds and modulates estrogen receptors. Sci. Adv. 8, eadd4150 (2022).
Conforti, A. et al. COVID-eVax, an electroporated DNA vaccine candidate encoding the SARS-CoV-2 RBD, elicits protective responses in animal models. Mol. Ther. 30, 311–326 (2022).
Velavan, T. P. & Meyer, C. G. Mild versus severe COVID-19: laboratory markers. Int. J. Infect. Dis. 95, 304–307 (2020).
Zheng, M. et al. TLR2 senses the SARS-CoV-2 envelope protein to produce inflammatory cytokines. Nat. Immunol. 22, 829–838 (2021).
Avolio, E. et al. The SARS-CoV-2 Spike protein disrupts human cardiac pericytes function through CD147 receptor-mediated signalling: a potential non-infective mechanism of COVID-19 microvascular disease. Clin. Sci. 135, 2667–2689 (2021).
Li, K. et al. SARS-CoV-2 Spike protein promotes vWF secretion and thrombosis via endothelial cytoskeleton-associated protein 4 (CKAP4). Signal Transduct. Target. Ther. 7, 332 (2022).
Iaconis, D. et al. Characterization of raloxifene as a potential pharmacological agent against SARS-CoV-2 and its variants. Cell Death Dis. 13, 498 (2022).
We acknowledge Dr. Michele Santoro at TIGEM, Pozzuoli, Italy, for generating sp5/sp7 mutants.
F.C., M.Ma., C.T., D.I., F.G., A.R.B., and M.A. are employees of Dompé Farmaceutici S.p.A that owns IP rights related to the study findings reported in this paper.
The protocol for the in vivo DNA transfer was approved by the local animal ethics committee and by the Italian Ministry of Health (authorization # 586/2019-PR).
About this article
Cite this article
Barbieri, S.S., Cattani, F., Sandrini, L. et al. Relevance of Spike/Estrogen Receptor-α interaction for endothelial-based coagulopathy induced by SARS-CoV-2. Sig Transduct Target Ther 8, 203 (2023). https://doi.org/10.1038/s41392-023-01488-3