Toxic effects of cell-free hemoglobin on the microvascular endothelium: implications for pulmonary and nonpulmonary organ dysfunction

Abstract

Levels of circulating cell-free hemoglobin are elevated during hemolytic and inflammatory diseases and contribute to organ dysfunction and severity of illness. Though several studies have investigated the contribution of hemoglobin to tissue injury, the precise signaling mechanisms of hemoglobin-mediated endothelial dysfunction in the lung and other organs are not yet completely understood. The purpose of this review is to highlight the knowledge gained thus far and the need for further investigation regarding hemoglobin-mediated endothelial inflammation and injury to develop novel therapeutic strategies targeting the damaging effects of cell-free hemoglobin.

BACKGROUND

Increased circulating cell-free hemoglobin (CFH) is a characteristic of several pathologic entities, including hemolytic diseases, sickle cell anemia, and inflammatory conditions such as sepsis (Table 1). In sepsis, higher levels of plasma CFH are associated with disease severity and organ dysfunction (23), including injury to the lung (20). Red blood cell fragility and damage during these pathologies leads to release of hemoglobin into the intravascular space. Although there have been many reports implicating CFH in endothelial activation and injury over the past 30 years, the precise signaling mechanisms involved in CFH-mediated endothelial dysfunction are still unknown. Our current knowledge of CFH-induced endothelial injury is limited to mechanisms involving nitric oxide scavenging (1, 2, 5–7), oxidative stress (29–31), general endothelial cell activation (3, 8, 26, 32), and more recently, endothelial barrier dysfunction (25, 33, 34) (Fig. 1). Hemoglobin is composed of two α-globin and two β-globin subunits, each containing a heme group with a central iron atom. The iron can redox cycle from its oxygen-carrying ferrous (2+) form to ferric (3+, methemoglobin) or the highly reactive ferryl (4+) form. The multiple oxidation states and the potential for heme and iron to be released from hemoglobin adds to the difficulty of identifying specific mechanisms of cellular and tissue injury, especially in vivo since the oxidation state can change rapidly and heme and iron release are difficult to quantify at the tissue or cellular level. Complicating things further, free heme can itself induce hemolysis, releasing more CFH into the circulation (35). Considering this, the purpose of this review is to provide a comprehensive overview of the studies to date that have investigated CFH-mediated endothelial injury in the lung and nonpulmonary organs and to highlight the knowledge gaps and need for more in-depth studies of the underlying pathological processes of CFH-induced organ dysfunction. The potential molecular mechanisms related to CFH-mediated endothelial injury that are discussed in this review are summarized in Fig. 2.

Table 1.

Disease states associated with elevated circulating CFH

Disorder	References
Hemolytic disorders
Inherited (sickle cell disease, thalassemias)	(1–4)
Acquired [blood transfusion, Rh incompatibility, hemodialysis, cardiac surgery, preeclampsia, extracorporeal circulation (ECMO), malaria]	(5–16)
Pulmonary hypertension	(17–19)
Sepsis	(20–22)
Acute respiratory distress syndrome	(23)
Trauma/hemorrhage/transfusion-related acute lung injury (TRALI)	(24)
Primary graft dysfunction (PGD) after lung transplantation	(25)
Vasculopathies (atherosclerosis, intraventricular hemorrhage, subarachnoid hemorrhage)	(26–28)

Figure 1.

Circulating cell-free hemoglobin (CFH) and its released components contribute to endothelial injury via NO scavenging, oxidative stress, endothelial activation, and endothelial barrier dysfunction. CFH is released from damaged red blood cells during several hemolytic and inflammatory pathologies. Elevated circulating levels of CFH are associated with higher risk of organ dysfunction and mortality. Hemoglobin is composed of two α-globin and two β-globin subunits, each consisting of an iron-containing heme group. The iron can redox cycle from its oxygen-carrying ferrous (2+) form to ferric (3+, methemoglobin) or the highly reactive ferryl (4+) form, which can oxidize lipids and other substrates, increasing oxidative stress. Moreover, there is potential for the release of heme and/or iron from circulating CFH, adding complexity to the contributions of CFH to endothelial injury. Mechanisms demonstrated to be involved in CFH-mediated endothelial injury include NO scavenging, oxidative stress, endothelial activation, and endothelial barrier dysfunction. NO, nitric oxide; ROS, reactive oxygen species.

Figure 2.

Potential signaling mechanisms involved in CFH-mediated endothelial injury. Though the entirety of the precise signaling mechanisms contributing to endothelial injury is yet to be elucidated, multiple pathways have been implicated. A: CFH causes increased ROS that lead to activation of NLRP3 signaling and activation of IL-1β, as well as degranulation of WPBs. Protective cellular mechanisms activated by CFH include HO-1, HIF, and FHC to counter oxidative stress. B: endothelial activation in response to CFH signals through the TLR4/MyD88 pathway leading to activation of MAPKs, NF-κB and complement signaling, upregulation of adhesion molecules, and release of pro-inflammatory cytokines; PAR-1 activation by CFH-mediated upregulation of the TF signaling cascade leads to increased coagulation signaling (including increased VWF and P-selectin). C: endothelial barrier dysfunction induced by CFH or its components is characterized by disruption of cell-cell junctions (adherens junctions and tight junctions) and reorganization of the actin cytoskeleton (stress fiber formation); signaling mechanisms that may be involved include upregulation of MAPKs, Src, Rho, MMPs, and pMLC signaling. Many of these signaling pathways are complex and overlapping. Of note, some studies show that TLR4/MAPK/NLRP3 signaling was not required in CFH-mediated endothelial injury; the precise signaling mechanisms involved most likely depend on context and require further investigation. CFH, circulating cell-free hemoglobin; C3, complement component 3; C5a, complement component 5a; E-sel, E-selectin; FHC, ferritin heavy chain; HIF, hypoxia-inducible factor; HO-1, heme oxygenase-1; ICAM-1, intercellular adhesion molecule 1; IL-1β, interleukin 1 beta; IL-6, interleukin 6; IL-8, interleukin 8; JNK, c-Jun N-terminal kinases; MAPK, mitogen-activated protein kinase; MMPs, matrix metalloproteinases; MyD88, myeloid differentiation primary response 88; NF-κB, nuclear factor kappa-light-chain-enhancer of activated B cells; NLRP3, NOD-, LRR- and pyrin domain-containing protein 3; PAR-1, protease-activated receptor-1; pMLC, phosphorylated myosin light chain; P-sel, P-selectin; ROCK, Rho-associated protein kinase; ROS, reactive oxygen species; TF, tissue factor; TLR4, Toll-like receptor 4; VCAM-1, vascular cell adhesion protein 1; VWF, von Willebrand factor; WPB, Weibel-Palade bodies.

CELLULAR AND MOLECULAR MECHANISMS OF CFH-MEDIATED TOXICITY

Nitric Oxide Scavenging

It has been recognized for over 150 years that hemoglobin has the capacity to bind nitric oxide (NO) (36). A number of studies have demonstrated the ability of intravascular CFH to scavenge NO and impair vasodilation, especially during conditions such as sickle cell disease-induced hemolysis (1, 2, 4), blood transfusion with storage lesion (5), hemodialysis (6), and hemolysis during cardiac surgery (7). CFH-mediated NO scavenging leading to vasoconstriction triggers a downstream cascade of endothelial events including decreased blood flow, increased inflammation, platelet activation, and organ injury (37). The effects are amplified when the endothelium is already dysfunctional, as is seen in patients with hyperlipidemia or diabetes (38). In fact, infusion of hemoglobin in db/db (diabetic) mice or mice fed a high-fat diet induced severe systemic vasoconstriction, which was prevented with inhalation of NO (39). In mice with hemorrhagic shock, transfusion of red blood cells (RBCs) stored for an extended amount of time, i.e., “stored” or “old” RBCs (14 days for porcine/mouse and 35–40 days for human), which are more fragile and more prone to CFH release, led to increased levels of CFH and augmented inflammatory markers, oxidative stress, and death; the effects were exacerbated in mice fed a high-fat diet and improved with NO inhalation (9).

Oxidative Stress

Hemoglobin can exist in multiple oxidation states with differing pro-oxidant capacity depending on the oxidation state of the heme iron moiety. An oxidative environment can convert ferrous hemoglobin (Hb²⁺) to the ferric (Hb³⁺, methemoglobin) or a more highly reactive ferryl (Hb⁴⁺) form, both of which induce greater toxicity through oxidation of lipids (17, 40, 41) and lipoproteins (42). Highly oxidized forms of hemoglobin have been detected in plasma and urine samples of mice with intravascular hemolysis and in human samples of cerebrospinal fluid following intraventricular hemorrhage (43). In a clinical study of preeclampsia, which is characterized by high levels of circulating CFH, patients had higher plasma membrane peroxidation capability and higher plasma levels of protein carbonyl groups, markers of oxidation (10), suggesting that CFH was driving oxidative injury. Circulating oxidants, including reactive oxygen species (ROS), oxidized hemoglobin, or oxidized lipids, have the capability to activate and/or damage other cells such as leukocytes and red blood cells that can further increase the oxidative environment creating a feed-forward mechanism that can amplify endothelial injury (40, 44). Exposure of endothelial cells to oxidized forms of hemoglobin also increases pathways that protect against hemoglobin-induced oxidative injury including intracellular heme oxygenase and ferritin (29). When taken up by cells, heme is broken down by heme oxygenase to labile iron (which leads to ferritin production), carbon monoxide, and biliverdin (which can be converted to bilirubin). Deletion of the heme oxygenase-1 gene has been shown to exacerbate many pathologies in mouse models of disease (45, 46). Each of these breakdown products has demonstrated cytoprotective effects in response to toxic stimuli (47). For example, exogenous CO administration protected against hypoxic lung injury in rats (48), and CO exposure in vitro inhibited proinflammatory cytokine production in response to LPS (49). Bilirubin (50) and ferritin (51) both exert antioxidant properties to protect endothelial cells from oxidative stress. Together, these antioxidant and anti-inflammatory products of heme catabolism maintain cellular homeostasis by protecting against oxidative stress and death (45, 46). However, in the setting of hemoglobin or heme exposure, these protective responses can be overwhelmed resulting in, for example, susceptibility to the toxic effects of hydrogen peroxide-mediated cellular injury (29, 31).

Endothelial Activation

Endothelial cell (EC) activation is characterized by upregulation of inflammatory cytokines, adhesion molecules, and procoagulant mediators. Increased circulating CFH has been implicated in general EC activation in patients with hemolytic disease (3), malaria-induced intravascular hemolysis (8), ruptured atherosclerotic lesions, and intraventricular hemorrhage (26). Several studies in animal models have demonstrated hemolysis-mediated EC activation as well. For instance, in a phenylhydrazine (PHZ) model of intravascular hemolysis, several inflammatory and EC injury markers were elevated in kidneys and hemolysis led to multiorgan injury (52). In the same model, intravascular hemolysis led to complement-dependent liver damage involving Toll-like receptor 4 (TLR4) signaling and upregulation of P-selectin on the endothelial surface (53). A study by Wozniak et al. (11) lends further evidence to the importance of studying the inflammatory effects of increased levels of circulating CFH. In a swine model of blood transfusion, stored RBCs contributed to pulmonary endothelial dysfunction as indicated by increased pulmonary vascular resistance index, and lung injury as measured by functional, histological, and biochemical parameters of acute lung injury recommended by the American Thoracic Society (54). When the stored RBCs were washed mechanically, levels of CFH increased and were associated with increased endothelial activation and tissue injury; alternatively, RBC rejuvenation before transfusion using a Food and Drug Administration-approved method to increase cellular adenosine triphosphate (ATP) and 2,3-diphosphoglycerate (2,3-DPG), reduced plasma CFH and was associated with a reduction in endothelial dysfunction, leukocyte sequestration, and lung and kidney injury (11). Direct evidence of hemoglobin-mediated endothelial activation was reported by Liu and Spolarics who showed that stimulation of endothelial cells with ferric hemoglobin induced nuclear factor κ-light-chain-enhancer of activated B cells (NF-κB)-mediated upregulation of adhesion molecules and release of chemokines and cytokines (32). Together, these studies indicate that CFH-mediated endothelial activation can contribute to lung damage in several pathophysiological contexts.

Barrier Dysfunction and Hyperpermeability

Endothelial barrier dysfunction leading to tissue edema can be triggered by inflammation and is especially deleterious in the lung, where disruption of the alveolar-capillary barrier causes pulmonary edema formation and acute respiratory distress syndrome (ARDS). It is increasingly apparent that high levels of circulating CFH during hemolytic and inflammatory pathologies disrupt the endothelial barrier, leading to microvascular hyperpermeability, edema, and organ dysfunction. Our group and others have shown that increased levels of CFH in patients during sepsis (20, 21), hemolytic anemia-associated pulmonary hypertension (55), and primary graft dysfunction after lung transplantation (25) are associated with increased lung injury and poor outcomes. Elevated plasma levels of CFH also lead to lung endothelial barrier dysfunction and lung microvascular hyperpermeability in rodent models of sickle cell disease (56), pulmonary hypertension (55), transfusion-induced acute lung injury (24), and sepsis (57). Direct hemoglobin infusion in an ex vivo-perfused human lung model resulted in increased lung weight and Evans blue-labeled albumin leakage from the circulation into the airspace, indicating increased vascular permeability (25). In vivo, hemoglobin injection caused blood-brain barrier disruption demonstrated by protein leak into brain tissue and loss of tight junction proteins; potential mechanisms include oxidative injury, indicated by increased staining of 4-hydroxynonenal (4-HNE, marker of lipid peroxidation) and 8-hydroxy-2′-deoxyguanosine (8-OHdG, marker of oxidative DNA damage) (58), apoptosis, determined by increased protein levels of cleaved caspase-3 (58), Rho kinase activation (59), and matrix metalloproteinase signaling (59). Though not lung specific, these studies offer clues to the potential mechanisms leading to alveolar-capillary barrier disruption due to increased circulating CFH.

In lung microvascular endothelial cells, direct hemoglobin stimulation has been shown to cause barrier dysfunction assessed by several measures, including decreased transendothelial electrical resistance (25, 33, 34), increased flux of protein across a monolayer (33, 34), cytoskeletal rearrangement and stress fiber formation (33), and formation of intercellular gaps (33). These effects have also been seen in other endothelial cell types, including pulmonary artery (60, 61), dermal (62), and aortic (63) endothelial cells. Many of these studies provide strong evidence for the role of oxidative stress in hemoglobin-mediated endothelial barrier dysfunction. For example, hemoglobin increases intracellular oxidative signaling as evidenced by induction of pathways that protect against oxidative injury such as heme oxygenase-1 (60) and hypoxia inducible factor (62). Hemoglobin-induced hyperpermeability can be reduced by antioxidant treatments such as n-acetylcysteine (NAC) (60, 61), ascorbate (34, 60), acetaminophen (a hemoprotein reductant) (25, 64), superoxide dismutase, or catalase (62). Furthermore, exposure to hyperoxia enhances hemoglobin-mediated lung endothelial permeability (25), and the ferryl (4⁺) form of hemoglobin is more potent than the ferric or ferrous forms (60, 65). In addition, although ferrous hemoglobin (Hb²⁺) alone did not cause endothelial barrier dysfunction in several endothelial cell types, including human lung microvascular endothelial cells, addition of glucose oxidase, a hydrogen peroxide-producing enzyme, stimulated a significant drop in transendothelial electrical resistance and loss of the adherens junction protein β-catenin at cell-cell contacts; the effects were prevented by haptoglobin, the endogenous scavenger of hemoglobin (66). Other signaling pathways involved in hemoglobin-induced endothelial barrier dysfunction include myeloid differentiation primary response 88 (MYD88) and NF-κB (62), as well as Src (61) or mitogen-activated protein kinases (MAPKs) (65). Conversely, others reported that blocking MAPKs, TLR4, NOD-, LRR-, and pyrin domain-containing protein 3 (NLRP3), or receptor for advanced glycation end products (RAGE) (61) did not reduce the barrier disrupting effects of RBC supernatants (containing increased levels of hemoglobin). The differences in reported signaling mechanisms could be explained by experimental use of lysed RBC supernatants rather than purified hemoglobin or differences in type of endothelial cell used. Of note, Lisk et al. (62) also reported that blocking TLR4 was unable to prevent barrier dysfunction of human dermal microvascular endothelial cells in response to direct stimulation with hemoglobin. Thus, although there is ample evidence for hemoglobin-induced endothelial barrier dysfunction and subsequent microvascular hyperpermeability across a variety of models and species, the precise signaling mechanisms may be tissue specific and are still incompletely elucidated.

CONTRIBUTION OF HEMOGLOBIN COMPONENTS TO CFH-MEDIATED TOXICITY

Toxic Effects of Heme

Although there is still much to parse out, several studies have advanced our understanding of the contribution of specific CFH components to hemoglobin-mediated injury and it is clear that free heme alone can induce many of the injurious effects observed for cell-free hemoglobin. The toxic effects of heme (or “hemin”—the chloride salt of heme used in many studies) have primarily been studied in the context of sickle cell disease (SCD), characterized by systemic hemolysis, inflammation, and coagulation. Infusion of hemoglobin or heme in SCD mice for 1 h induced vaso-occlusion and ischemic injury due to obstruction of the microcirculation by sickled RBCs; the authors concluded that heme was the major active molecule causing the effect since ferric hemoglobin, which was capable of releasing heme, caused much greater vaso-occlusion (measured by percent stasis) than cyanomethemoglobin (a form of hemoglobin in which heme cannot be released), and infusion of heme alone matched the percent stasis induced by hemoglobin (67). The authors also concluded that the intact iron was required since an iron-free heme mimic protoporphyrin IX (PPIX) had little effect on stasis. Further investigation into heme-induced vaso-occlusion revealed that hemoglobin or heme infusion in vivo or direct stimulation in vitro increased endothelial adhesion molecule surface expression, Weibel-Palade body (WPB) degranulation, oxidant production, and TLR4 signaling in mouse microvessels, human umbilical vein endothelial cells (HUVECs), or mouse pulmonary vein endothelial cells (mPVECs); antibodies blocking individual adhesion molecules, hemopexin (the primary endogenous scavenger of heme), NADPH oxidase (NOX) and protein kinase C (PKC) inhibitors, iron chelation, antioxidant NAC, and TLR4 inhibition or knockout all blocked the stasis induced by heme. In addition, inhibition of TLR4 blocked leukocyte rolling and adhesion, activation of NF-κB, and mortality induced by heme in SCD mice (67). Building upon this study, a recent investigation reported that inhibition of tissue factor, factor Xa, thrombin, or protease-activated receptor-1 (PAR-1) reduced stasis in a hemoglobin-infusion model of vaso-occlusive crisis in SCD mice (68). Another study showed that intact heme (but not ferrous or ferric hemoglobin, iron-free heme, or iron alone) stimulated NLRP3 inflammasome activation and production of interleukin 1 beta (IL-1β) in HUVECs; the effect was enhanced by LPS priming and partially dependent on ROS production (69). A recent study also showed heme-induced ROS production in HUVECs; the toxic effects could be prevented with hydroxyurea-induced upregulation of intracellular antioxidant systems (70). Though ferric hemoglobin stimulated heme oxygenase-1 (HO-1), ROS production, and IL-1β, much higher concentrations were needed than those required with heme. Heme-mediated activation of caspase-1 and cleavage of IL-1β was prevented in NLRP3 knockout mice (69). These results are supported by a report of red blood cell-derived microparticles containing high levels of heme, especially from patients with SCD, that transfer heme to HUVECs and induce apoptosis through activation of TLR4/ROS signaling (71). Mechanisms of heme- and iron-mediated endothelial cytotoxicity have been reviewed in more detail previously (72, 73).

Lung-Specific Heme-Mediated Endothelial Toxicity

Specific to the lung, direct stimulation of human lung microvascular endothelial cells with heme decreased transendothelial electrical resistance, increased barrier dysfunction as measured by flux of FITC-dextran, and induced necroptosis; barrier dysfunction and cell death were prevented by TLR4 inhibitor TAK-242, antioxidant NAC, iron chelator deferoxamine, and necroptosis inhibitor necrostatin-1 (74). A recent study also reported that heme-induced barrier dysfunction of human lung microvascular endothelial cells is characterized by disruption of tight junction proteins zonula occludens-1 (ZO-1), claudin-1, and claudin-5, activation of p22 and p38/MAPK pathways, and promotion of stress fiber formation (75). Heme injection (intravenous) into mice resulted in p38 activation [phosphorylation of p38 and upstream MAPK-activated protein kinase 2 (MK2) and dual specificity mitogen-activated protein kinase kinase 3 (MKK3)], loss of tight junction proteins, and disruption of the lung endothelial barrier; the effects were blocked in MKK3^−/− mice (75).

Complexity of Heme-Mediated Endothelial Responses

Though the above studies support an injurious effect of heme on the endothelium, not all studies support a role for heme alone. For instance, even though endothelial activation and vascular permeability were exacerbated in a model of intravascular hemolysis, particularly in the liver, there was no difference in alveolar-capillary barrier permeability in mice with genetic deletion of hemopexin, the primary endogenous scavenger of heme, compared to wild-type (76). One study even showed that pretreatment of human microvascular endothelial cells with hemin, the chloride salt of heme, inhibited permeability induced by hydrogen peroxide; however, hyperpermeability was exacerbated in bovine aortic endothelial cells (77). Conversely, heme administration to wild-type mice was not lethal, but greatly enhanced mortality in a cecal ligation and puncture (CLP) model of sepsis (22). The results of these studies indicate that the endothelial response to heme is complex, and it will be important to examine the differences in the inflammatory, pro-thrombotic, and barrier function responses in different vessels (conduit vs. microvascular) and specific tissue beds during distinct disease states.

Role for Iron in CFH-Mediated Endothelial Dysfunction

The role of free iron in hemoglobin-related endothelial dysfunction is even less well understood. As mentioned above, several studies point to a critical role for the iron moiety in hemoglobin- or heme-mediated endothelial injury (67, 69), and iron chelation has shown protection against hemoglobin- or heme-mediated endothelial dysfunction (67, 74). On one hand, one study showed free iron was not sufficient to activate endothelium (69). However, another study demonstrated that iron loading of endothelial cells contributes to cytotoxicity and lipid peroxidation (78). The mechanisms involved in iron-mediated endothelial dysfunction, particularly in the context of hemoglobin-mediated injury, require additional investigation.

THERAPEUTIC TARGETING OF CFH TO PRESERVE ENDOTHELIAL FUNCTION

Haptoglobin and Hemopexin

Haptoglobin and hemopexin are endogenous scavengers of hemoglobin and heme, respectively, and have been investigated for their usefulness in preventing hemoglobin-mediated injury and inflammation. Comprehensive reviews describing the role of endogenous hemoglobin scavengers and defense pathways during inflammatory processes, including sickle cell disease, pulmonary hypertension, vasculopathies, atherosclerosis, sepsis, and blood transfusion, have been published previously (79–81). Here, we will explore the potential benefit of exogenous administration of haptoglobin or hemopexin in pathologies with increased circulating levels of hemoglobin or heme.

Elevated plasma levels of cell-free hemoglobin are associated with poor outcomes in sepsis (20, 21), primary graft dysfunction (25), pulmonary arterial hypertension (18), acute respiratory distress syndrome (23), and blood transfusion (12), indicating hemoglobin-or heme-binding scavengers have potential in a number of acute and chronic pathological conditions. In a retrospective observational study of critically ill patients with sepsis, those with higher levels of haptoglobin or hemopexin were significantly less likely to die in the hospital using univariate analysis; using multivariate analysis, however, the significant association remained only with haptoglobin levels (21). In addition, Larsen et al. (22) observed an association of low hemopexin concentration with more organ dysfunction and increased mortality in patients with septic shock. In patients with SCD, low levels of haptoglobin or hemopexin were associated with increased lipid peroxidation; postmortem analysis of pulmonary artery and lung parenchyma tissue from patients with SCD and pulmonary hypertension (PH), in which haptoglobin and hemopexin levels were depleted, showed evidence of oxidized LDL deposits in the pulmonary artery, right ventricular fibrosis, and markers of endothelial dysfunction (82). Furthermore, in children with sickle cell anemia, the HP2-2 genotype, a common variant of the haptoglobin gene which has a lower affinity for binding plasma CFH, was associated with an increased risk for the development of acute severe vaso‐occlusive pain episodes (83). Interestingly, removing hemoglobin from cerebral spinal fluid derived from patients with subarachnoid hemorrhage via haptoglobin sequestration restored vasodilatory responses to NO in isolated arteries ex vivo (27).

In several animal models of increased levels of CFH, haptoglobin and hemopexin have been shown to protect against severity of tissue damage, organ dysfunction, and death. In mice with septic shock and RBC transfusion, infusion of haptoglobin or hemopexin improved survival, attenuated inflammation, and prevented hemoglobinuria and kidney injury (84). Haptoglobin infusion also attenuated organ damage and disease severity in a canine model of pneumonia (85), a guinea pig model of blood transfusion with storage lesion (86), a rat model of pulmonary hypertension (17), a rat model of hemolysis (87), and a sheep model of hemoglobin-induced cerebral vasospasm (27). Glucocorticoid stimulation of endogenous haptoglobin synthesis reduced the hemoglobin-induced hypertensive response in canines, and haptoglobin infusion in guinea pigs prevented hypertension, hemoglobinuria, peroxidative activity, and oxidative tissue damage in response to CFH infusion (88). In mice deficient in heme oxygenase-1, sepsis induced by low-grade cecal ligation and puncture resulted in increased plasma levels of hemoglobin and heme, decreased levels of haptoglobin and hemopexin, organ injury indicated by histological detection of necrosis in liver, kidney, and heart tissue with corresponding increased plasma levels of aspartate aminotransferase (AST), blood urea nitrogen (BUN), and creatinine phosphokinase (CPK), and mortality; restoration of hemopexin prevented the negative outcomes (22). Hemopexin was also able to prevent heme loading of the vascular endothelium and subsequent endothelial activation, oxidative stress, and cardiac function in SCD mice (89). Notably, purified human haptoglobin is approved in Japan for clinical use during several hemolytic processes including cardiopulmonary bypass surgery, massive transfusion, and extracorporeal circulation, among others (90).

Acetaminophen and Ascorbate

In 2010, Boutaud et al. (91) discovered that the commonly used analgesic acetaminophen was able to inhibit hemoprotein-induced lipid peroxidation by reducing ferryl to ferric heme and quenching globin radicals. Acetaminophen administration decreased oxidant injury and tissue damage and improved kidney function in a rat rhabdomyolysis model. This study revealed the potential therapeutic application for acetaminophen in diseases involving hemoprotein-mediated injury. Specific to CFH-mediated lung injury, acetaminophen was shown to prevent vascular permeability in an ex vivo perfused human lung model and in human lung microvascular endothelial cells in response to direct CFH stimulation (25). In a retrospective study, Janz et al. (21) observed that patients with elevated circulating levels of CFH who received acetaminophen during ICU admission had decreased plasma levels of oxidative injury markers and lower risk of death in the hospital independent of other risk factors for poor outcomes. Building on this, in a single center, randomized, double-blind, placebo-controlled phase 2 clinical trial in patients with sepsis with detectable plasma levels of CFH, acetaminophen (1 g every 6 h for 3 days, enteral administration) reduced oxidative injury and improved renal function (92) compared to placebo treatment. Furthermore, in patients with severe falciparum malaria with prominent intravascular hemolysis, a phase 2, open-label, randomized controlled trial (NCT01641289) demonstrated that acetaminophen (1 g every 6 h for 3 days, enteral administration) reduced creatinine levels, indicating renoprotection; the therapy was most effective in patients with higher plasma hemoglobin levels (>45,000 ng/mL) (13). These studies provide compelling preclinical and clinical evidence for potential utility of acetaminophen for treatment of pathologic entities associated with elevations in circulating cell-free hemoglobin. A prospective multicenter phase 2 b randomized placebo-controlled double-blinded interventional platform trial, Acetaminophen and Ascorbate in Sepsis: Targeted Therapy to Enhance Recovery (ASTER), has been proposed (NCT04291508) to investigate the clinical efficacy of acetaminophen for patients with evidence of sepsis-induced hemodynamic or respiratory failure. Interestingly, the other arm of the proposed ASTER trial investigating ascorbate in critically ill patients could also be effective for hemoglobin-mediated pathologies. Ascorbate, a potent antioxidant, has been shown to attenuate CFH-mediated endothelial barrier dysfunction (34, 60).

Other Potential Therapeutic Options

Therapeutic strategies for hemoglobin-mediated endothelial injury (Table 2) could target several facets of the pathophysiologic process, including prevention of hemolysis, neutralization of hemoglobin or its released products, inhibition of downstream signaling mechanisms, or repair of tissue injury (Fig. 3). A combination of these strategies may provide greater and/or longer lasting benefits. One such therapy tested during hemolysis-induced inflammation in normal and SCD mice is nitrite. Nitrite is able to reduce hemolysis, deliver NO to areas of hypoxia, and reduce platelet and leukocyte adhesion to endothelium (94). Nitrite therapy has also been shown to decrease transfusion-related acute lung injury in mice (24). A recent study confirmed the ability of nitrite, along with additional NO donors, to protect RBCs from hemolysis; the most effective was S-nitrosoglutathione (GSNO) (95). A recent study in SCD mice also showed that inhibition of xanthine oxidase, a potent inducer of oxidant production, with febuxostat decreased hemolysis, indicated by a decrease of circulating CFH and increase in circulating haptoglobin levels, and improved pulmonary vasoreactivity, measured by wire myography of isolated arteries (96). The endogenous scavengers of hemoglobin and heme have already been discussed, but other options for neutralizing hemoglobin are possible. In fact, plasma proteins such as albumin are known to bind hemoglobin and heme, and it has been shown that hemoproteins must be unbound in the plasma to stimulate inflammatory responses in the endothelium (69, 93). However, albumin was not sufficient to prevent inflammation or improve survival during RBC transfusion in mice with septic shock (84). Targeting hemolysis or circulating hemoglobin may be beneficial in early stages of critical illness or for chronic illness but targeting downstream mechanisms may be required when damage is already done. Especially when it comes to disruption of the alveolar-capillary barrier, it will be important to investigate strategies to enhance barrier function or repair dysfunctional endothelium to improve outcomes in inflammatory injury; the many strategies currently being investigated are beyond the scope of this review.

Table 2.

Potential therapies targeting CFH-mediated endothelial injury

Therapy	Method of Action	References
Haptoglobin	Endogenous scavenger of hemoglobin	(17, 21, 27, 66, 80, 82, 84–86, 88, 90)
Hemopexin	Endogenous scavenger of heme	(22, 80, 82, 84, 89, 90)
Plasma proteins (e.g. albumin)	Binds hemoglobin/heme	(69, 93)
Acetaminophen	Hemoprotein reductant	NCT04291508 (21, 25, 91, 92)
Ascorbate	Antioxidant	(34, 89)
NO donors [nitrite, nitrate, spermine NONOate (SPNO), S-nitrosoglutathione (GSNO)]	Delivers NO (increases bioavailability in hypoxic areas and reduces hemolysis)	(24, 94, 95)
Febuxostat	Xanthine oxidase inhibitor	(96)

Figure 3.

Potential therapeutic strategies that target CFH-mediated endothelial injury. [Color key: Red = Injurious, Green = Cytoprotective, Blue = Potential therapies] Hemolysis (lysis of red blood cells) in the circulation releases CFH and subsequent free heme to affect endothelial function via oxidative stress and inflammation. Intracellular anti-inflammatory, anti-oxidant, and anti-apoptotic pathways driven by upregulation of HO-1 break down heme into labile iron (which leads to production of cytoprotective ferritin), CO, and biliverdin (which is converted to bilirubin by biliverdin reductase). Potential therapies against CFH-mediated endothelial injury are aimed at targeting hemolysis (NO donors such as nitrate, nitrite, GSNO, or Febuxostat), CFH (haptoglobin, plasma proteins), heme (hemopexin, plasma proteins), or oxidative stress/inflammation (acetaminophen, ascorbate, NO donors). BR, bilirubin; BV, biliverdin; BVR, biliverdin reductase; CFH, cell-free hemoglobin; CO, carbon monoxide; Fe, iron; GSNO, S-nitrosoglutathione; HO-1, heme oxygenase-1; RBC, red blood cells; XO, xanthine oxidase; NO, nitric oxide.

Ongoing Research Priorities

Further investigation into the pathophysiology and therapeutic targeting of CFH-mediated endothelial dysfunction holds promise in several inflammatory and hemolytic conditions. Questions remain regarding how different disease states induce red cell fragility leading to release of hemoglobin and its components, the cellular mechanisms by which CFH and heme interact with the endothelium, what downstream signaling mechanisms play a role in CFH-mediated endothelial injury, and how each step in this process can be efficiently targeted. Therapeutic strategies to neutralize CFH or heme with haptoglobin or hemopexin, respectively, or downstream effectors, have been mostly limited to preclinical studies (97); moving these investigations into clinical trials is necessary to evaluate their effectiveness in a variety of inflammatory syndromes. Though clinical trials using acetaminophen during conditions in which CFH is elevated show promise in reducing organ injury, more investigation into its potential benefits is required. Additionally, more work should be done to investigate the potential for pharmacologic interventions already being used to target vascular dysfunction, especially those targeting oxidative pathways such as vitamin C, xanthine oxidase, or polyphenols [recently reviewed by Daiber and Chlopicki (98)], in the context of CFH- and heme-mediated pathologies. Additionally, circulating levels of CFH, heme, haptoglobin, or hemopexin may have value as biomarkers. Preclinical and clinical evidence shows associations of changes in the circulating levels of these proteins with poor outcomes in sepsis (21, 99, 100), pulmonary hypertension (18), subarachnoid hemorrhage (101), preeclampsia (102), and cardiopulmonary bypass (103). In addition, measurement of CFH was used to identify patients with elevated circulating CFH for enrollment in a pilot clinical trial of acetaminophen to target CFH-mediated oxidative injury in patients with sepsis (92); a similar approach could be used to target other therapies such as scavengers of CFH and heme to patients with elevated CFH who would be more likely to benefit from these strategies. Evaluating and combining strategies to prevent hemolysis, measure and neutralize hemoglobin and its released components, and interrupt downstream signaling events involved in CFH-mediated endothelial injury in preclinical and clinical studies in which circulating CFH is elevated has potential to uncover novel therapeutic strategies for a number of inflammatory and hemolytic conditions.

CLOSING REMARKS

Evidence continues to grow for the contributions of elevated circulating levels of CFH and its components to endothelial injury in disease states characterized by red cell fragility and damage. The complex pathophysiological processes involved in hemolytic and inflammatory diseases in which circulating CFH contributes to endothelial injury will most likely require a multifaceted therapeutic approach to improve patient outcomes. Further investigation into the precise signaling mechanisms involved in CFH-mediated endothelial injury utilizing complex physiological disease models with translational potential is warranted to reach this goal.

GRANTS

This work was supported by National Institutes of Health National Heart, Lung, and Blood Institute (NHLBI) Grant HL094296 (to J.E.M.), NHLBI Grant HL135849 (to J.A.B. and L.B.W.), and NHLBI Grant HL103836 (to L.B.W.).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

J.E.M., J.A.B., and L.B.W. prepared figures; J.E.M. drafted manuscript; J.E.M., J.A.B., and L.B.W. edited and revised manuscript; J.E.M., J.A.B., and L.B.W. approved final version of manuscript.