Critical role of oxidized LDL receptor-1 in intravascular thrombosis in a severe influenza mouse model

Although coagulation abnormalities, including microvascular thrombosis, are thought to contribute to tissue injury and single- or multiple-organ dysfunction in severe influenza, the detailed mechanisms have yet been clarified. This study evaluated influenza-associated abnormal blood coagulation utilizing a severe influenza mouse model. After infecting C57BL/6 male mice with intranasal applications of 500 plaque-forming units of influenza virus A/Puerto Rico/8/34 (H1N1; PR8), an elevated serum level of prothrombin fragment 1 + 2, an indicator for activated thrombin generation, was observed. Also, an increased gene expression of oxidized low-density lipoprotein (LDL) receptor-1 (Olr1), a key molecule in endothelial dysfunction in the progression of atherosclerosis, was detected in the aorta of infected mice. Body weight decrease, serum levels of cytokines and chemokines, viral load, and inflammation in the lungs of infected animals were similar between wild-type and Olr1 knockout (KO) mice. In contrast, the elevation of prothrombin fragment 1 + 2 levels in the sera and intravascular thrombosis in the lungs by PR8 virus infection were not induced in KO mice. Collectively, the results indicated that OLR1 is a critical host factor in intravascular thrombosis as a pathogeny of severe influenza. Thus, OLR1 is a promising novel therapeutic target for thrombosis during severe influenza.

The elevation of expression of Il6, Icam1, and Olr1 genes in the aorta and lungs of mice during severe influenza. Endothelial dysfunction was considered involved in the induction of abnormal blood coagulation during severe influenza. Thoracic aorta and lung samples from control and PR8 virus-infected mice were collected at 1, 3, and 6 dpi, and gene expression related to inflammation and endothelial functions was a b www.nature.com/scientificreports/ investigated (Fig. 2). Interleukin-6 (Il6), a proinflammatory cytokine, was significantly increased by 15.8-and 3.6-fold in the aorta (Fig. 2a) and by 228.46-and 63.78-fold in the lung (Fig. 2d) of infected mice at 3 and 6 dpi, respectively (p < 0.0001, two-way ANOVA). Intercellular adhesion molecule-1 (Icam1), which encodes adhesion molecules for leukocytes, was expressed at slightly but significantly increased levels in the infected mouse aorta (1.69-and 1.49-fold at 3 and 6 dpi, respectively, p < 0.05, two-way ANOVA; Fig. 2b), whereas its expression in the lungs was significantly increased only at 3 dpi (2.53-fold, p < 0.0001, two-way ANOVA; Fig. 2e). OLR1, an endothelial receptor for LDL, is a key player in oxidized LDL-induced atherogenesis and endotoxin-induced inflammation 14,19 . Interestingly, aortic Olr1 expression was significantly increased in PR8 virus-infected mice by 5.6-and 3.0-fold at 3 and 6 dpi, respectively (p < 0.0001, two-way ANOVA; Fig. 2c). The lung of infected mice also showed a significant level of induction of Olr1 at 3 dpi (1.33-fold, p < 0.05, two-way ANOVA; Fig. 2f). Also, expression levels of Il6 and Olr1 were significantly correlated in samples collected at 3 dpi (aorta, R 2 = 0.8585, p < 0.0001; lung, R 2 = 0.6488, p < 0.05; linear regression analysis). Aortic samples collected at 6 dpi also showed a weaker but significant correlation between the levels of these genes (R 2 = 0.4516, p < 0.0005, linear regression analysis). Given its critical role in endotoxin-induced inflammation and endothelial dysfunction 14,17 , OLR1 was hypothesized to be involved in local and systemic inflammation as well as abnormal blood coagulation observed in mice with severe influenza. This hypothesis led to conduct influenza virus infection experiments in Olr1 KO (KO) mice, which was previously established 19 , to confirm whether this host factor is involved in those pathological events.  (Fig. 4b). These results indicated that the host factor OLR1 plays an important role in pathologi-a b The results further suggested that OLR1 is dispensable for local cytokine production. Collectively, these results revealed that OLR1 plays a role in the activation of thrombin generation in severe influenza without affecting the production of inflammatory cytokines.

Suppression of the influenza-induced thrombosis in the lungs of Olr1 KO mice. Pulmonary
inflammation and thrombus formation in the lungs at the lethal phase of influenza were further investigated utilizing the severe influenza mouse model, and the results from WT mice were compared to those from KO mice (Fig. 6). No apparent difference was found in the microscopic observation of the lungs between WT and KO mice (Fig. 6a, b). The degree of lung inflammation caused by PR8 virus infection was examined in hematoxylin and eosin (HE)-stained sections from WT and KO mice sacrificed at 6 dpi when mice showed severe body weight losses. Lungs from both WT and KO mice demonstrated obvious peribronchial inflammation, inflammatory cells in alveoli, thickened alveolar walls, and alveolar hemorrhage after virus infection (Fig. 6c,  d). Also, leukocytes in the vascular intima and perivascular spaces were observed in infected mice (Fig. 6c,   www.nature.com/scientificreports/ To confirm thrombus formation associated with severe influenza, phosphotungstic acid hematoxylin (PTAH) staining was performed, in which fibrin is stained in blue and thrombi can be visualized as well as fibrin deposition. As shown in Fig. 6e,f, increased intravascular thrombus formation in the lungs of WT mice infected with PR8 virus was indicated by multiple fibrin clots in blood vessels (Fig. 6e, arrows). In contrast to WT-infected mice, only sporadic thrombi were found in blood vessels of KO-infected mice similar to those in the uninfected groups, although HE staining indicated severe pulmonary inflammation in KO mice (Fig. 6f). Intravascular fibrin deposition observed in the lungs of infected KO mice seemed due to the clotting of residual blood in the blood vessels of the lungs after euthanasia, as they were found mainly in small veins, even in areas of less severe inflammation as in uninfected mice. The average numbers of clots [± standard error of the mean (SEM)] in the lung sections were 6.5 ± 2.7 in WT-control mice (n = 4), 53.8 ± 4.5 in WT-infected mice (n = 4), 7.7 ± 2.9 in KOcontrol mice (n = 3), and 9.8 ± 3.3 in KO-infected mice (n = 4). The number was significantly larger only in the WT infection group, compared to those in other groups (p < 0.0001, two-way ANOVA). These results demonstrated a critical role of OLR1 in severe influenza-induced intravascular thrombus formation in the lungs.

Discussion
This study demonstrated prolonged PT and increased thrombin generation from prothrombin at the lethal phase of severe influenza in a mouse model. The results were consistent with previous findings in mouse and ferret influenza models 6,7 . Also, this study revealed a significant induction of aortic Olr1 and its critical contribution to the thrombin generation and intravascular thrombosis in the lungs of mice with severe influenza. The important physiological roles of OLR1 have already been demonstrated in platelet activation, endothelial dysfunction, leukocyte migration, plaque formation, and atherosclerosis as a consequence of these pathological events 12,14,16,20 . In the context of acute coronary syndrome as a result of the progression of atherosclerosis, OLR1 has been thought www.nature.com/scientificreports/ to be involved in prothrombotic pathways induced by oxidized LDL 21 . Also, this study shows that OLR1 is critically involved in the first step of thrombus formation by promoting thrombin production in severe influenza. Although the detailed molecular mechanism needs to be elucidated, OLR1 induced in the vascular cells of mice infected with influenza virus may have promoted adhesion between platelets and the endothelial surface, as reported previously 16 , and increased thrombin generation by activated platelet. The contribution of OLR1 to PT prolongation in severe influenza was also demonstrated in this study. Suppose consumptive coagulopathy is the cause of PT prolongation as already suggested in a ferret influenza model 7 , the suppression of thrombin formation in KO mice would have reduced the consumption of coagulation factors, resulting in only mild PT prolongation. Interestingly, increased intravascular fibrin clotting was not evident in other tissues of infected mice, e.g., the liver (data not shown), despite the elevation of circulating thrombin. Therefore, not only circulating thrombin but also factors associated with virus infection and/or a severe inflammatory response in the lung during influenza appear involved in a stable fibrin clot formation. For example, integrating previous reports on influenza and thrombus formation, increased expression of tissue factor, externalization of phosphatidylserine, and decreased blood flow velocity induced in the lungs by influenza virus infection are considered to promote thrombus formation preferably in the tissue [22][23][24][25] .
OLR1 expression is very low under normal conditions and upregulated in response not only to its ligand oxidized LDL 26 but also various stimuli, such as lipopolysaccharide (LPS) [27][28][29] . Several transcription factors have been reported to regulate Olr1 transcription. Given the genetic regulation of Olr1 by inflammatory cascades through the activation of the transcription factor NF-κB 30,31 , influenza virus infection-induced systemic cytokine secretion could have induced Olr1 in this study. A strong positive correlation between gene expression levels of Olr1 and Il6 suggested a link between OLR1 and inflammation. Because inflammation in the lungs and elevated blood cytokine levels in KO mice were similar to those in WT mice after virus infection, OLR1 does not regulate inflammatory responses but is a downstream factor induced by inflammation in the present experimental condition. The detailed mechanisms of Olr1 induction by influenza virus infection remain elucidated. In addition to systemically secreted cytokines, of course, the possibility that the virus directly infects the cells of blood vessels 32 and induces Olr1 needs to be taken into account. However, given that Olr1 is induced by various factors, its expression may have been induced by others aside from inflammatory molecules, such as oxidized LDL, angiotensin II, and metabolic abnormalities 26,33,34 , which have been previously reported to be induced during acute influenza 18,35,36 . Further studies on the mechanisms of Olr1 induction by a viral infection will provide insights into biological responses to and pathogenesis of infectious diseases far beyond just cytokine induction.
Influenza virus infection-induced lung inflammation and systemic cytokine secretion were not affected by the absence of OLR1 in this study. However, OLR1 was involved in endotoxin-induced acute lung inflammation in a previous study in which an anti-OLR1 antibody pretreatment completely blocked immune cell activation and infiltration into the lungs after intraperitoneal injection with endotoxin 15,37 . This difference may reflect a pathophysiological difference between the host response to endotoxin and that to virus infection. In septic models, the administered LPS binds to Toll-like receptor (TLR) 2/4 on the cell membrane and causes an inflammatory response by activating NF-κB signaling in each cell in the first step. OLR1 has been reported to colocalize and cooperate with TLR2 to activate inflammatory responses by the outer membrane protein A of Gram-negative bacteria 38 . Furthermore, OLR1 functions as a bacterial receptor that enhances the adhesion of Gram-negative and Gram-positive bacteria to cells 39 . Therefore, because OLR1 is involved in the very early steps of TLR-mediated signaling on the cell membrane, the absence of OLR1 and its blockade may have strongly suppressed the inflammatory responses in the septic model. In contrast, during viral infection, virus entry into cells occurs first, and various viral molecules, such as viral membrane glycoproteins, viral constituent proteins, and nucleic acids, activate inflammation-related transducing cascades inside the infected cells, leading to the activation of NF-κB and other transcription factors to promote cytokine production 40 . Especially in the case of the influenza virus, at least two pathways thought to be independent of TLR2/4 have been reported to activate NF-κB: (1) endoplasmic reticulum stress induced by the overload of viral protein hemagglutinin 41 and (2) double-stranded RNA-activated protein kinase 42 . Therefore, cytokine production in viral infection could be activated independently on OLR1. WT and KO mice showed a similar degree of weight loss after virus infection. This may be due to anorexia caused by increased circulating cytokines 43 . When focusing on biological responses after cytokine induction, weight loss does not seem to be a good indicator of the severity of the disease.
In summary, the findings indicated that influenza virus infection induces Olr1 gene expression in the vascular system to promote thrombin generation and resultant intravascular clotting in the lungs. Thus, OLR1 is a promising novel therapeutic target to suppress the prothrombotic state during severe influenza. In addition, thrombosis has been observed to occur in many viral infections and is thought to be involved in the symptoms and severity of the diseases, including coronavirus disease 2019 44 . The importance of OLR1 in thrombosis should be considered in a wide range of infectious diseases. Mice. The Olr1 KO mice B6.129P2-Olr1 tm1Saw (KO mice in this manuscript) were generated as previously reported 19 . Male C57BL/6 mice purchased from Hokudo (Sapporo, Japan) and KO mice kindly given by Dr. Sawamura were kept in a BSL-2 laboratory and a clean room, respectively, at the International Institute for Zoonosis Control, Hokkaido University, under standard laboratory conditions (room temperature 22 °C ± 2 °C, relative humidity 50% ± 10%) and a 12 /

Measurement of the serum levels of cytokines and chemokines. Measurement of cytokines and
chemokines in serum samples was carried out as previously reported 18 . The serum levels of IL-6, IP-10, MCP-1, and MIP-1β were determined using a MAGPIX Milliplex kit (Merck, Darmstadt, Germany) according to the manufacturer's instructions. Briefly, 25 μL serum samples, standards, and controls were added to a 96-well plate comprising an equal amount of assay buffer for serum samples or serum matrix for standards and controls. Next, magnetic beads coated with antibodies against the target cytokines were added to each well, and the plates were incubated on a plate shaker overnight at 4 °C. After washing with washing buffer in the kit, the samples were reacted with biotinylated detection antibodies for 1 h and then with streptavidin-phycoerythrin for 30 min. After washing and the addition of loading buffer from the kit, the samples were analyzed by the MAGPIX system (Luminex, Austin, TX, USA).

Measurement of lung viral titers.
Mice were euthanized at 3 and 6 dpi, and their lung samples were collected and homogenized in 1 mL RPMI-anti medium [RPMI-1640 (Thermo Fisher Scientific) with 100 U/ mL penicillin (Sigma-Aldrich), 100 µg/mL streptomycin (Sigma-Aldrich), and 20 µg/mL gentamicin (Thermo Fisher Scientific)]. After centrifugation at 3,000 rpm for 10 min, the supernatants were collected and stored at − 80 °C until further analysis. For plaque assays, monolayers of MDCK cells were prepared by seeding 1.2 × 10 6 cells in 3 mL RP10 medium (RPMI-1640) supplemented with 10% inactivated fetal bovine serum (GE Healthcare UK Ltd., Little Chalfont, Buckinghamshire, UK), 1 mM sodium pyruvate (Thermo Fisher Scientific), 50 µM 2-mercaptoethanol (Merck), 100 U/mL penicillin, 100 µg/mL streptomycin, and 20 µg/mL gentamicin in each well of the tissue culture six-well plate and incubated overnight at 37 °C in 5% CO 2 . The monolayers were washed with RPMI-anti, and 125 µL tenfold serially diluted lung lysates were added to each well. The viruses were allowed to adsorb to the monolayers for 45 min, with shaking of the plates at 15 min intervals. Then, 3 mL prewarmed overlay medium consisting of Leibovitz L-15 with glutamine at pH 6.8 (Thermo Fisher Scientific) supplemented with 0.028% (w/v) NaHCO 3 (Merck), 100 IU/mL penicillin, 100 mg/mL streptomycin, 0.1% (w/v) TPCK-treated trypsin (Merck), and 0.9% (w/v) agarose (BD Biosciences, Franklin Lakes, NJ, USA) were added to each well. The plates were then incubated at 37 °C in 5% CO 2 for 3 days. Plaques on the monolayers were then counted without staining.