Spike protein up-regulates inammatory axis of both thromboinammation and leukotriene in severe COVID-19

Hypercytokinemia is a critically fatal factor in COVID-19. However, underlying pathogenic mechanisms are unknown. Here we show that brinogen and leukotriene-A4 hydrolase (LTA4H), two of the most potent inammatory contributors, are elevated by 67.7 and astonishing 227.7% in the plasma of patients infected by SARS-CoV-2 and admitted to intensive care unit in comparison with healthy control, respectively. Conversely, transferrin identied as a brinogen immobilizer in our recent work and Spink6 are down-regulated by 40.3 and 25.9%, respectively. Furthermore, we identify Spink6 as the rst endogenous inhibitor of LTA4H, a pro-inammatory enzyme catalyzing nal and rating limited step in biosynthesis of leukotriene-B4 that is an extremely inammatory mediator and a target to design superior anti-inammatory drugs. Additionally, virus Spike protein is found to evoke LTA4H and brinogen expression in vivo. Collectively, these ndings identify the imbalance between inammatory drivers and antagonists, which likely contributes to hypercytokinemia in COVID-19. Spink6 may have superior anti-inammatory function because it specically targets epoxide hydrolase of LTA4H to inhibit leukotriene-B4 biosynthesis without effecting LTA4H’s aminopeptidase activity.


Introduction
In December 2019, the rst pneumonia case caused by the infection of Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2 or COVID-2019) was reported [1][2][3] . COVID-2019 outbreak has since rapidly spread to other places in China 4 , as well as other countries around the world. As of April 6, 2020, 1 288 592 cases of COVID-19 have been con rmed and 70 446 cases died worldwide, carrying a mortality of approximately 5.467%, compared with a mortality rate of less than 1% from in uenza.
Accumulating evidence indicates that a subgroup of patients with severe COVID-19 have a cytokine storm syndrome (hypercytokinemia) to cause hyperin ammation. On 11 March, 2020, world health organization (WHO) characterized SARS-CoV-2 as a pandemic. This is the rst pandemic caused by a coronavirus 6 . Unfortunately, to date, there is no vaccine or antiviral treatment for this coronavirus.
As other pathogenic human coronavirus such as severe acute respiratory syndrome CoV (SARS-CoV) and Middle East respiratory syndrome CoV (MERS-CoV) 7 , SARS-CoV-2 infects lower respiratory tract to cause lung immunopathology and lead to deleterious clinical manifestations such as severe pneumonia, fatal acute lung injury and acute respiratory distress syndrome, consequently resulting in high morbidity and mortality 8 9 . The severity of clinical manifestations is often associated with dysregulated immune response and elevated proin ammatory cytokine/chemokine responses that possibly progress to a hypercytokinemia. All those observations strongly suggest that immunopathological events play crucial roles to cause fatal pneumonia after SARS-CoV-2 infections. However, the underlying molecular events causing the hypercytokinemia are unknown.
In this study, we investigate the immunopathology mechanisms after SARS-CoV-2 infection. We nd that: 1) the virus Spike protein promotes LTA4H expression; 2) Spink6 is the rst identi ed endogenous inhibitor of LTA4H, one of the most potent in ammatory driver; 3) signi cant imbalance between two of the most potent in ammatory drivers ( brinogen and LTA4H) and their antagonists (transferrin and Spink6) possibly causes hypercytokinemia in COVID-19.

Results
Fibrinogen and its endogenous immobilizer transferrin are up-and down-regulated in the plasma of patients infected by SARS-CoV-2, respectively As one of the most potent in ammatory mediators, brinogen has powerful role in driving in ammatory responses and can be up-regulated by virus infection [10][11][12] . As illustrated in Figure 1A and Table S1 and 2, blood routine examination showed that mean plasma brinogen concentration in ICU (4.868 mg/ml) (S-SC, n = 22; male 12; female 10; age: 36-77 years) and in non-ICU (4.109 mg/ml) patients (MM, n = 40; male 20; female 20; age: 10-73 years) was much higher than that in control healthy (2.903 mg/ml) (HC, n = 37; male 19; female 18; age: 28-56 years). Recently, we have demonstrated that the iron transporter transferrin is an immobilizer of brinogen by binding to brinogen with a molecular ratio of 4:1 13 . We next checked transferrin concentration in the plasma of the patients. Mean transferrin concentration (1.443 mg/ml) in the plasma of patients infected by SARS-CoV-2 who had been admitted to ICU (n = 20; male 12; female 8; age: 36-77 years) was signi cantly lower than that in non-ICU patients (1.854 mg/ml) (MM, n = 43; male 23; female 20; age: 10-73 years) and that in healthy control (2.417 mg/ml) (HC, n = 22; male 14; female 8; age: 28-56 years) ( Figure 1B and Table S1 and 3).
After a 24-48 h in vivo administration (40 g/kg), the virus Spike protein signi cantly up-regulated brinogen in mice ( Figure 1C), further con rming that SARS-CoV-2 infection is able to evoke brinogen expression. The increase of in ammatory mediator brinogen and the decrease of its endogenous immobilizer transferrin in the plasma of patients infected by SARS-CoV-2 likely produce the increased tendency for in ammation.
The serine protease inhibitor-like protein Spink6 is identi ed as the rst endogenous inhibitor of LTA4H As a ubiquitously expressed proin ammatory enzyme, LTA4H has two opposing roles in in ammation regulation including the biosynthesis of in ammatory mediator leukotriene-B4 (LTB4) and antiin ammation by degrading and inactivating the chemotactic tripeptide PGP [13][14][15] . None of LTA4H's endogenous inhibitor has been identi ed. Surface plasmon resonance (SPR) analysis revealed that Spink6 directly interacted with LTA4H (Figure 2A), and the association rate constant (Ka),, dissociation rate constant (Kd),, and equilibrium dissociation constantKD) values for the interaction between Spink6 and LTA4H were 2.8× 10 5 M −1 s −1 , 1.3× 10 −4 s −1 ,and 0.45 nM (Figure 2A). We found here that Spink6, which is as a serine protease inhibitor with 57 amino acid residues 17 , selectively inhibited the epoxide hydrolase function of LTA4H ( Figure 2B-D) to inhibit LTA4 hydrolysis and nally block the formation LTB4, an important pro-in ammatory lipid mediator. However, Spink6 had no effect on the aminopeptidase function of LTA4H ( Figure 2E), suggesting that Spink6 does not impair the anti-in ammatory function of LTA4H.
LTA4H and Spink6 are up-and down-regulated in the plasma of COVID-19 patients, respectively and Spike protein induces LTA4H expression LTA4H plays a key role in catalyzing nal and rating limited step in biosynthesis of leukotriene-B4, which has been implicated in many acute and chronic in ammatory diseases 14,15,[17][18][19][20][21] . The concentration of LTA4H was investigated in the plasma of the patients. As illustrated in Figure 3A and  Figure 3B and Table S1 and 5). LTA4H's up-regulation and its endogenous inhibitor Spink6's down-reregulationin the plasma of patients infected by SARS-CoV-2 likely increase the production of leukotriene-B4, an extremely in ammatory mediator.
In order to investigate the mechanism to induce LTA4H's up-regulation in the patients, we intravenously administrated the virus Spike protein to mice and test it's effect on LTA4H expression in vivo. As illustrated in Figure 3C, compared with the basic LTA4H concentration of 2.2 g/ml in mouse plasma, the LTA4H concentration was increased to 3.2 and 4.3 g/ml at day 1 and day 2 after the Spike protein administration (40 g/kg), respectively, further con rming that SARS-CoV-2 infection is able to evoke LTA4H expression.
Transferrin and Spink6 inhibit TNF-secretion induced by brinogen and LTA4H, respectively As two of the most potent in ammatory drivers [10][11][12][13][14][15][16] , brinogen and LTA4H have ability to drive in ammation by directly or indirectly promoting in ammatory cytokines secretion. Considering that signi cantly elevated concentrations of brinogen and LTA4H were found in the plasma of patients infected by SARS-CoV-2 and in the mouse plasma treated by the virus Spike protein, we tested if transferrin and Spink6 have ability to inhibit in ammatory cytokine secretion induced by brinogen and LTA4H. As illustrated in Figure 4A, brinogen (1 mg/ml, 3 M) signi cantly induced TNF-secretion in mouse macrophage cells of RAW264.7, while transferrin inhibited the cytokine secretion in a dosedependent manner. Unexpectedly, TNF-secretion induced by brinogen was also inhibited by Spink6 ( Figure 4B), suggesting that thromboin ammation induced by brinogen is associated with leukotriene in ammatory signaling. In addition, as expectedly, the virus Spike protein (40 g/ml, 0.5 M) up-regulated TNF-expression in the RAW264.7 cells while Spink6 inhibited the up-regulation ( Figure 4C). Given that LTB4 was associated with pulmonary disorders and neutrophil migration in lipopolysaccharide (LPS)induced acute lung injury 22,23 , we investigated if Spink6 has effect on TNF-expression promoted by LPS in human monocytic cell (THP-1). As illustrated in Figure 4D, Spink6 (40 nM) inhibited TNF-secretion induced by both virus Spike protein (40 g/ml, 0.5 M) and LPS (1 g/ml) in the THP-1 cells.

Stimulation assays
Mice were intravenously injected with Spike (40 μg/kg; 40069-V08B, Sino Biological, China), and the plasma from different groups of mice after stimulation of 0, 12, 24, and 48 h, respectively, was extracted and the concentrations of brinogen and LTA4H were tested using a mouse brinogen ELISA kit and a LTA4H ELISA kit, respectively, as described above.

Statistical analysis
The data obtained from independent experiments are presented as means ± SD. All statistical analyses were two-tailed with 95% con dence intervals (CI). Kolmogorov-Smirnov test (K-S test) was used in the analysis of normal distribution, and data were then analyzed using one-way ANOVA with post hoc Dunnett adjustment for p values. If only 2 groups were compared, unpaired t-test was applied. In all statistical analyses, signi cance was set at p<0.05. Analyses were done with SPSS software (v22.0) and GraphPad Prism (v5.0).

Discussion
This study showed that two of the most potent in ammatory drivers ( brinogen and LTA4H) were signi cantly up-regulated, while their antagonists (transferrin and Spink6) were signi cantly downregulated in the plasma of patients infected by SARS-CoV-2. The drivers' up-regulation and agonists' down-regulation are positively and negatively associated with the severity of clinical manifestations induced by SARS-CoV-2 infection and likely contribute to the hypercytokinemia, respectively, and likely contribute to the hypercytokinemia ( Figure 4E). Especially, the concentration of LTA4H, which is a proin ammatory enzyme which generates the known in ammatory mediator LTB4 and a target to design superior anti-in ammatory drugs, was extremely elevated by 227.7% in the plasma of ICU patients. Furthermore the virus Spike protein was found to stimulate LTA4H expression in vivo. In addition, the endogenous inhibitor of LTA4H, Spink6 was identi ed.
Respiratory failure resulted from acute respiratory distress syndrome (ARDS) is the leading cause of mortality in COVID-19 24,25 . Secondary haemophagocytic lymphohistiocytosis (sHLH) is a hyperin ammatory syndrome that is characterized by a fulminant and fatal hypercytokinaemia likely causing multiorgan failure. sHLH is most commonly triggered by viral infections in adults. A pro le of increased cytokines such as interleukin (IL)-2, IL-7, granulocytecolony stimulating factor, monocyte chemoattractant protein 1, and TNF-α resembling sHLH is associated with the severity of COVID-19 disease 9, 24, 26-29 . Recently, a recent retrospective, multicentre study of 150 COVID-19 cases suggests that the mortality is due to virally driven hyperin ammation 25 . All the data indicate hypercytokinaemia is a critically fatal factor in COVID-19 and thus addressing the mechanism causing hypercytokinaemia would provide strategy for the disease treatment.
As a plasma factor with high concentration of ~6-12 M, brinogen is a known in ammatory mediator containing powerful roles in driving in ammatory responses by in uencing multiple aspects of leukocyte and platelet biology to drive thromboin ammatory axis through direct and indirect mechanisms 10 . Recently, we have demonstrated that another plasma factor transferrin with high concentration of ~26-52 M acts as an immobilizer of brinogen by normally binding to brinogen with a molar ratio of 4:1 13 , suggesting that transferrin likely blocks the in ammatory functions of brinogen. In this report, we did nd that transferrin inhibited the in ammatory response induced by brinogen, further con rming transferrin's function as an anti-in ammatory factor. Signi cant transferrin decrease (-40.3%, 19 M with ability to immobilize only 4.75 M brinogen) and brinogen increase (+67.7%, 15 M) observed in the plasma of the ICU patients likely cause hyperin ammatory responses due to the extreme imbalance between transferrin and brinogen.
More importantly, another extreme imbalance between LTA4H (+227.7%, from 73.24 up to 240 ng/ml) and Spink6 (-25.9%, from 20.93 down to 15.51 ng/ml) identi ed as inhibitor of LTA4H here was observed in the plasma of the ICU patients. LTA4H is a pro-in ammatory epoxide hydrolase which catalyzes the conversion of LTA4 to the in ammatory mediator LTB4, which is an in ammatory axis of LTA4-LTA4H-LTB4 14 . As an extremely in ammatory lipid mediator, LTB4 is implicated in many acute and chronic in ammatory diseases in humans 15,[18][19][20] . Leukotrienes are thought to contribute to the airway in ammatory response in bronchiolitis caused by some virus infection [30][31][32] . LTA4H represents an attractive therapeutic target because it triggers the nal critical and rate limited step for LTB4 biosynthesis 16 . These extreme imbalances of brinogen and LTA4H, which are two of the most potent in ammatory contributors, and their antagonists of transferrin and Spink6 likely cause excessive in ammatory responses and even hypercytokinemia as observed in the plasma of patients infected by SARS-CoV-2. Interference of these imbalances may provide safe and effective anti-hypercytokinemia strategies.
In this report, we can not completely address the mechanisms to cause the up-regulation of the in ammatory contributors including brinogen and LTA4H and to cause the down-regulation of their antagonists including transferrin and Spink6, but at least the partial mechanism has been found here. As illustrated in Figure 1C & 3C, the virus Spike protein showed ability to signi cantly promote both brinogen and LTA4H expression in mice. Given the extreme up-regulation of plasma LTA4H concentration observed in the ICU patients, we suggest that LTA4H-dominantly in ammatory responses induced by Spike protein in severe COVID-19 are critical to cause hypercytokinemia.
Given LTB4 as an extremely in ammatory mediator and LTA4H's key role in LTB4 production, LTA4H inhibition is a potent strategy for the development of anti-in ammatory drugs. Many trials have been actively pursued but very few candidates show clinic e cacy because LTA4H is a bifunctional enzyme containing both epoxide hydrolase and aminopeptidase activity, exerted by two overlapping catalytic sites and those candidates show limited selectivity 14 . LTA4H inhibitors sparing the aminopeptidase activity of the enzyme are speculated as superior and safer drugs. In this report, Spink6 was identi ed as the rst endogenous inhibitor of LTA4H. It selectively inhibited the function epoxide hydrolase without effecting the aminopeptidase activity of the enzyme, providing an excellent candidate and/or template for the development of superior and safer drugs. In addition, Spink6 showed ability to inhibit cytokine secretion induced by products of virus (Spike) and bacteria (LPS) and endogenous in ammatory mediator ( brinogen and LTA4H), indicating it is also a wide spectrum anti-in ammatory reagent. Figure 1 Increased brinogen and decreased transferrin level in patients infected by SARS-CoV-2. Concentrations of brinogen (A) and transferrin (B) in plasma from patients infected by SARS-CoV-2 who had been admitted to intensive care unit (ICU) (S-SC), non-ICU patients (MM), and healthy control (HC) were determined. Data represent mean ± SD (n = 20 -43), **p < 0.01 by (A and B) unpaired t-test. (C) Concentration of brinogen in the plasma of mice after Spike (40 μg/kg, intravenous injection) stimulation was determined by ELISA. Data represent mean ± SD (n = 6), **p < 0.01 by (C) unpaired t-test. Tf: transferrin; FIB: brinogen.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download. Supplementaryinformation.pdf