Sickle cell disease (SCD) is caused by a single point mutation (Glu->Val) at position 6 in the β-globin gene that leads to polymerization of deoxy-hemoglobin S (HbS) and the characteristic sickling of red blood cells. The ongoing polymerization of HbS promotes hemolysis, inflammation, and vaso-occlusive pain crises (1). During hemolysis, HbS is released into the vasculature and readily oxidized to methemoglobin, which can release free heme (2, 3). Normally, free hemoglobin and heme in plasma are safely cleared by haptoglobin and hemopexin (4, 5). However in SCD, chronic hemolysis depletes circulating haptoglobin and hemopexin (6, 7), allowing free heme to activate toll-like receptor 4 (TLR4) signaling on endothelial and inflammatory cells (3, 8–10). TLR4 signaling on endothelium leads to NF-ĸB activation, rapid release of Weibel-Palade bodies, discharge of P-selectin and von Willebrand factor, and vaso-occlusion in murine models of SCD (3).

TLR4 signaling induced by heme or its canonical ligand, lipopolysaccharide (LPS), requires cofactors CD14 and MD-2 (3, 8, 11–15). The transduction mechanism for TLR4 activation is well characterized for LPS from gram negative bacteria. CD14 delivers LPS to MD-2, which forms a stable heterodimer with the extracellular domain of TLR4 (16), leading to dimerization of the TLR4-MD-2 complex (17). This brings the toll/interleukin-1 receptor (TIR) domains on the cytoplasmic tails of TLR4 together allowing recruitment of cytoplasmic adapter molecules that promote oxidant production by NADPH oxidase and activation of downstream signaling pathways, leading to activation of pro-inflammatory transcription factors, such as NF-κB and AP-1, as well as the production of type 1 interferons (18).

MD-2 is co-expressed with TLR4 on the cell membrane of various cell types including leukocytes and endothelium (11, 19, 20), but soluble MD-2 (sMD-2) is elevated in plasma from patients with inflammatory conditions, such as sepsis, HIV infection, and endotoxemia (21–23). Epithelial cells in the gut and the airways express TLR4, but not MD-2, and are therefore entirely reliant on sMD-2 for TLR4 signaling (24–26). sMD-2 circulates in plasma of healthy individuals primarily as disulfide-linked oligomers (27). During experimental human endotoxemia, sMD-2 in septic plasma increases rapidly like an acute phase reactant and contains both sMD-2 oligomers and monomers. The monomeric form of sMD-2 represents the biologically active form of MD-2 (28).

In addition to being a required co-factor for TLR4 responses to LPS, MD-2 is also a required cofactor for TLR4 responses to heme (15). Heme binds to MD-2 at amino acids W23 and Y34 on MD-2, at a site independent of the LPS-binding site, to activate TLR4 pro-inflammatory signaling (15). Since SCD is a pro-inflammatory condition with an activated endothelium and elevated plasma heme, we examined sMD-2 levels in SCD plasma and its potential role in endothelial cell activation. We tested the hypothesis that sMD-2 is increased in SCD plasma and, by binding heme, contributes to pro-inflammatory signaling by endothelial cells.

These data demonstrate that SCD plasma has increased levels of sMD-2 that can bind plasma heme and activate pro-inflammatory IL-8 secretion by endothelial cells through TLR4 signaling. The combination of heme-induced oxidative stress, inflammation, and adhesion of circulating blood cells to vascular endothelium is a key driver of the pro-inflammatory and prothrombotic vasculature that promotes sludging and stasis of blood flow in the post-capillary venules and ongoing ischemia-reperfusion physiology (44–46). Nature provides complicated systems to respond to heme overload, such as heme oxygenase-1 (HO-1) induction to degrade heme and heme-binding proteins to neutralize and transport heme (43, 47). Among the known heme-binding proteins, plasma HPX sequesters and transports heme to the liver for catabolism and detoxification via induction of HO-1 and ferritin (48, 49). Plasma HPX levels are depleted in SCD patients and mice because of chronic intravascular hemolysis (6, 7, 31). We have previously shown that increasing HPX through supplementation (3, 5) or gene transfer (31) prevents heme-induced inflammation and vaso-occlusion in SS mice. In this study, HPX-treated SS plasma lost its enhanced ability to induce IL-8 secretion in HUVEC, while adding HPX to AA plasma had no effect on basal IL-8 secretion. These findings demonstrate that heme can bind to sMD-2 and suggest that the addition of HPX to SS plasma sequesters heme away from sMD-2, thereby preventing TLR4-mediated IL-8 production by endothelial cells.

MD-2 expression can be regulated in specific tissue responses to sterile inflammatory stimuli and bacteria. In human hepatocytes, IL-6 induces the expression of MD-2 (34). In monocytes, corneal and intestinal epithelial cells, IL-10 or INF-γ can induce MD-2 expression (23, 50–52). Using immunohistochemical analysis, Wolfs et al. found MD-2 is expressed by endothelial cells and inflammatory cells in the livers and lungs of septic patients and suggested these cells are a source of the enhanced circulating sMD-2 levels during acute systemic inflammatory diseases such as endotoxemia and sepsis (22). In a recently published study, we found both MD-2 and MIP-2α (a murine IL-8 homologue) mRNA levels were increased in the livers of Townes SS mice compared to control AA mice, these differences are consistent with markedly increased hemolysis in Townes SS mice (32). In this study, we showed that HUVEC secreted sMD-2 in response to heme or LPS. We speculate that endothelial cells might be a potential source of some of the sMD-2 in SCD plasma, which will need further studies to validate.

MD-2 is a required accessory molecule for TLR4 signaling, indispensable for LPS recognition and signaling (25, 53). MD-2 knockout mice are hypo-responsive to LPS and are able to survive endotoxic shock, supporting a significant role of MD-2 in TLR4-dependent inflammatory responses in vivo (54). As a pattern recognition protein, MD-2 has been shown to bind to various ligands besides LPS to activate TLR4 signaling and promote inflammation, including heme (15), palmitic acid (55), angiotensin (56), and some synthetic compounds with no similarity to LPS (57). In our study, we found that heme agarose pulls-down sMD-2 from SCD plasma and recombinant sMD-2 has a Soret band at 414 nm, which indicates that sMD-2 binds heme.

Heme-induced microvascular stasis in SS mice requires endothelial TLR4 signaling (3). Heme activates TLR4 signaling on endothelial cells, leading to delivery of Weibel-Palade body constituents P-selectin and VWF to the surface of the vessel wall and the activation of the pro-inflammatory transcription factor NF-ĸB (3). Our finding that sMD-2-heme in SS plasma induces TLR4-dependent IL-8 secretion is in congruence with these previous observations. Thus, sMD-2 in plasma provides a potential pathway for heme-MD-2-mediated activation of TLR4 signaling to induce activation of the endothelium and vaso-occlusion in SCD.

The low level of IL-8, s-VCAM-1 and IL-6 production in HUVEC and HMVEC-L incubated with AA plasma and the incomplete inhibition of IL-8 secretion by HUVEC incubated with SS plasma depleted of sMD-2 or SS plasma plus TAK-242 indicates that other pro-inflammatory molecules in SS plasma might also induce HUVEC IL-8 secretion (35, 37, 39). A basal level TLR4 receptor activity induced by AA plasma has been reported before. Using a TLR4 reporter cell line that specifically recognizes ligands that bind and activate TLR4 (TLR4/NF-kB/SEAP Stable Reporter Cells), Xu et al. (39) found plasma from healthy AA control subjects induced low levels of TLR4 receptor activity; about 10% of the total TLR4 receptor activity induced by AA plasma was from HMGB1-dependent TLR4 receptor activity. In our study we found the TLR4 inhibitor TAK-242 had no significant effects on AA plasma-induced IL-8 secretion, suggesting other non-TLR4 pathways are likely contributing to the low level of IL-8 secretion induced by AA plasma. In SS plasma, 85% of IL-8 production by HUVEC was inhibited by TAK-242. The remaining IL-8 production induced by SS plasma could have been induced by other pro-inflammatory molecules that do not require TLR4 such as IL-1β, IL-18, and IL-6 (35–39).

The function of sMD-2 has been extensively studied in LPS-TLR4 signaling. Following synthesis, MD-2 is either secreted directly into the medium as a soluble, active protein, or binds directly to TLR4 in the endoplasmic reticulum before migrating to the cell surface (12). Heme activates TLR4 signaling by binding to a site on MD-2 that is independent and distinct from the LPS binding site on MD-2 (15). As shown in other studies (3, 8), heme-MD-2/TLR4-mediated pro-inflammatory effects are not due to LPS contamination. In this study, endothelial cell IL-8 secretion in response to SS plasma was not blocked by the LPS antagonist, polymyxin B.

SCD patients are known to be at higher risk for contracting bacterial infections compared to healthy humans (58, 59). The molecular pathophysiology that contributes to this infection susceptibility is incompletely understood. Because lung epithelial cells express TLR4 without MD-2, they remain resistant to endotoxin (25). sMD-2, produced either by neighboring cells or brought to the epithelial cells by plasma exudate, is required for LPS activation of lung epithelial cells (60). Increased circulating sMD-2 in SCD plasma could make SCD patients more susceptible to pulmonary infections and acute chest syndrome, especially in response to enhanced hemolysis (9, 61).

There are several limitations of this study. Although we found SS plasma increased IL-8, sVCAM-1, and IL-6 secretion in both HUVEC and HMVEC-L, endothelial cells vary widely among vascular beds such as lung and brain. The functions of sMD-2-heme in different endothelial cell beds and non-endothelial cells with TLR4 receptors, such as platelets and macrophages need be defined by separate studies.

In conclusion, SCD is considered a chronic inflammatory disease with hemolysis, vaso-occlusion, and ischemia-reperfusion, resulting in activation of the innate immune system and persistent activation of leukocytes, platelets, and endothelial cells. Therapeutic approaches for SCD include targeting HbS polymerization and the inflammatory processes that trigger vaso-occlusion (1, 35, 62). Clinical studies targeting TLR4 signaling as a therapeutic approach for sepsis and septic shock have brought several compounds and antibodies to clinical trials, with unsuccessful results. One such unsuccessful therapeutic is TAK-242, a small-molecule inhibitor of TLR4 signaling, that failed to suppress cytokine levels in patients with sepsis or respiratory failure in a phase III study (63). The LPS-MD-2/TLR4 antagonist, eritoran, also failed to reduce mortality among patients with severe sepsis (64). Nevertheless, targeting MD-2 to interfere with MD-2–TLR4 signaling has been extensively explored, and the results support the concept that MD-2 is an effective target to treat inflammatory disorders such as sepsis and acute lung injury (53, 65–68). We provide evidence that sMD-2 is increased in SCD, contributes to pro-inflammatory signaling in endothelial cells, and therefore, might be a potential therapeutic target for SCD and other hemolytic conditions.

The original contributions presented in the study are included in the article/ Supplementary Material . Further inquiries can be directed to the corresponding author.

The studies involving human participants were reviewed and approved by University of Minnesota’s Institutional Review Board. The patients/participants provided their written informed consent to participate in this study. The animal study was reviewed and approved by University of Minnesota’s Institutional Animal Care and Use Committee.

PZ, GMV, and JDBel designed the study, wrote, and edited the manuscript. PZ and JDBel prepared the figures. PZ, JN, and FA performed the experiments. JDBec and ATN collected data, analyzed results, and reviewed the manuscript. All authors contributed to the article and approved the submitted version.

This work was funded by NIH grant R01 HL114567.

JDBec receives funding from Bayer not related to work herein. JDBel and GMV receive research funding from CSL Behring and Mitobridge (Astellas).

The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.