The role of circulating cell-free hemoglobin in sepsis-associated acute kidney injury

Abstract

Sepsis is a heterogeneous clinical syndrome that is commonly complicated by acute kidney injury (sepsis-AKI). Currently, no approved pharmacological therapies exist to either prevent sepsis-AKI or to treat sepsis-AKI once it occurs. A growing body of evidence supports a connection between red blood cell (RBC) biology and sepsis-AKI. Elevated levels of circulating cell-free hemoglobin (CFH) released from RBCs during hemolysis are common during sepsis and can contribute to sepsis-AKI through several mechanisms including tubular obstruction, nitric oxide depletion, oxidative injury, and pro-inflammatory signaling. A number of potential pharmacological therapies targeting CFH in sepsis have been identified including haptoglobin, hemopexin, and acetaminophen, and early phase clinical trials suggest that acetaminophen may have beneficial effects on lipid peroxidation and kidney function in patients with sepsis. Bedside measurement of CFH levels may facilitate predictive enrichment for future clinical trials of CFH-targeted therapeutics. However, rapid and reliable bedside tests for plasma CFH will be required for such trials to move forward.

1. Introduction

Sepsis, the clinical syndrome of organ dysfunction and immune dysregulation in response to infection,¹ remains a leading cause of both hospitalization and inpatient mortality in the United States,² and is associated with high hospital costs, morbidity, and mortality.³ Acute kidney injury (AKI) is a common complication of sepsis and septic shock, with a reported incidence of 10 to 65% among patients admitted to an intensive care unit (ICU) with sepsis.^4,5 The development of AKI during sepsis (sepsis-AKI) carries both prognostic and therapeutic implications. Sepsis-AKI is associated with increased hospital costs, prolonged hospitalization, and increased mortality both in the ICU and at 90 days.^4,5 Furthermore, as many antibiotics and other common pharmaceuticals are either cleared from the body via the kidneys or are directly nephrotoxic, AKI often complicates the treatment of sepsis by forcing clinicians to either modify or discontinue important therapeutic drugs.

The pathophysiology of sepsis-AKI is incompletely understood.⁶ Hemodynamic effects cannot fully explain sepsis-AKI as it can occur in the absence of overt hypotension,^4,7 and only a minority of sepsis-AKI patients have acute tubular necrosis (ATN) on histopathology.⁸ Currently, no approved pharmacological therapies exist to either prevent sepsis-associated AKI or to treat AKI once it occurs.⁹ Over the last decade, multiple lines of evidence have converged to support the hypothesis that plasma cell free hemoglobin (CFH) may contribute to the pathogenesis of organ dysfunction including AKI during sepsis.^10–14 In this review, we discuss the biology of CFH as it pertains to sepsis and other critical illness states, the renal effects of CFH both in animal models and in humans, and potential therapies to mitigate the nephrotoxic effects of CFH during sepsis.

2. Red blood cell biology during sepsis

2.1. Overview

In mammals, nearly all hemoglobin (Hb) is contained within circulating red blood cells (RBCs). Hb that escapes the RBC cytoplasm because of hemolysis, termed cell-free hemoglobin (CFH), is tightly regulated. Mammals have several mechanisms to rapidly scavenge and remove CFH and its byproducts from circulation, including the plasma proteins haptoglobin¹⁵ and hemopexin.¹⁶ Circulating CFH is not detectable in any significant quantities under normal physiologic conditions. However, many acute and chronic medical states can increase circulating CFH levels. These include genetic conditions such as sickle cell disease and the thalassemias, parasitic infections from the Plasmodium and Babesia genera, auto-immune disorders, hemodialysis, and mechanical circulatory support devices such as left ventricular assist devices, cardiopulmonary bypass, and extracorporeal membrane oxygenation. Sepsis also exposes the RBC to multiple physiologic stressors that increase risk of hemolysis and release of CFH into circulation.¹⁷

2.2. CFH release during disseminated intravascular coagulation

Perhaps the most readily recognized cause of hemolysis during sepsis is disseminated intravascular coagulation (DIC), which commonly complicates sepsis.¹⁸ DIC is an acquired microangiopathy characterized by systemic activation of both the coagulation and fibrinolytic cascades leading to formation of microvascular thrombi and consumption of circulating platelets and coagulation factors.¹⁹ The formation of fibrin strands results in mechanical shearing of RBCs, resulting in intravascular hemolysis and release of CFH into the circulation.^18,19 The degree of hemolysis can be profound in septic patients with DIC, often necessitating multiple RBC transfusions to compensate for ongoing losses both from hemolysis and bleeding.¹⁹ Although septic patients with DIC can develop clinically significant elevations in CFH levels, there are other mechanisms that contribute to hemolysis during sepsis, and the absence of DIC does not preclude high levels of circulating CFH.¹¹

2.3. RBC deformability

During sepsis, RBCs undergo changes in cellular geometry and deformability.^20,21 Deformability, a cellular property that characterizes a cell’s ability to change shape during flow,^20,21 is an important property of RBCs that facilitates laminar flow in high-flow vessels with high shear rate and allows RBCs to navigate through small end-organ capillary networks and the interendothelial slits of the spleen.²⁰ Deformability of RBCs is decreased during sepsis²¹ resulting in more rigid RBCs that are prone to injury during transit through small vessels and to destruction by the reticuloendothelial system. These reductions in RBC deformability can be induced by changes in 2,3 diphosphoglycerate levels,²² alterations in glucose metabolism,²³ depletion of nitric oxide,²⁴ exposure to reactive oxygen species,²⁵ and interactions with white blood cells,²⁴ all of which occur during sepsis.

2.4. RBC suicidal death during sepsis

Cytokine signaling during sepsis can trigger suicidal RBC death, termed eryptosis. In an experimental model, Kempe et al. reported that exposure of RBCs from healthy human donors to plasma from septic adult patients resulted in increased RBC lysis compared to exposure of RBCs to plasma from non-septic healthy adults.²⁶ The increased hemolysis was accompanied by increases in RBC cytosolic Ca²⁺ activity and ceramide formation, leading to increased exposure of phosphatidylserine on the RBC plasma membrane, which triggered eryptosis and release of CFH.²⁶ Furthermore, exposure of donor RBCs to sphingomyelinase derived from Staphylococcus aureus also triggered RBC lysis along the same pathway.²⁶ Under normal physiologic conditions, phosphatidylserine exposure during eryptosis promotes binding of eryptotic RBCs to receptors expressed by splenic macrophages.²⁷ This event facilitates phagocytosis of RBCs for controlled degradation,²⁷ which prevents release of CFH into circulation.¹⁷ However, this pathway may be impaired during sepsis. Phosphatidylserine also triggers RBC binding to endothelial cells,²⁸ activation of the coagulation cascade,²⁹ and increases RBC aggregation,²⁸ mechanisms that contribute to uncontrolled hemolysis during acute sickle cell crises and to the thrombophilia of inherited hemoglobin disorders.^28,29 Uncontrolled activation of these eryptosis signaling pathways during sepsis may overwhelm the endogenous capacity to remove RBCs with phosphatidylserine exposure, thus promoting uncontrolled hemolysis and release of CFH into circulation.

2.5. Pathogen-specific contributors to CFH release

Pathogen-specific factors may also contribute to hemolysis in sepsis. Parasitic infections by Plasmodium spp. and Babesia spp., the causative organisms of malaria and babesiosis, directly cause hemolysis as a result of their intraerythrocytic life cycle.^30,31 These organisms can induce both acute hemolysis with large elevations in plasma CFH levels^32,33 as well as chronic low-grade hemolysis leading to chronic CFH exposure.³¹ Many gram-negative and gram-positive bacterial species can produce pore-forming toxins (PFTs) during infection.³⁴ PFTs are a structurally diverse class of toxin proteins that are secreted by the invading pathogens, bind to target host cells, then perforate the host cell plasma membrane, which disrupts cell homeostasis and causes subsequent cell death.³⁴ Among these, the hemolysins produced by Staphylococcus aureus, Enterococcus spp., and Escherichia coli are notable for their ability to induce RBC hemolysis in vitro.³⁴ Although it is unclear if these hemolysins always lead to clinically significant hemolysis in vivo, these toxins commonly serve as critical virulence factors in invasive bacterial infections.

3. Cell free hemoglobin is elevated during sepsis

Two recent studies have reported on CFH levels in septic adults.^11,12 In a study of 391 adult ICU patients with sepsis, Janz et al. observed that plasma CFH was detectable in 81% of patients even in the absence of overt hemolytic disease or DIC.¹¹ Higher levels of CFH were independently associated with increased risk of in-hospital mortality when controlling for age, comorbid medical conditions, and severity of illness.¹¹ Similar findings were reported by Adamzik et al. who measured CFH levels in 161 critically ill adults with sepsis.¹² A higher CFH level (defined as plasma CFH concentration above the cohort median) was associated with increased risk of 30-day mortality, and the CFH level had better predictive performance for mortality than plasma procalcitonin, sequential organ failure assessment (SOFA) score, or simplified acute physiology scale II (SAPS-II) score.¹² In both studies, even relatively low levels of plasma CFH were associated with increased risk of death.

Similar findings have been observed in experimental animal models. Larsen et al. examined the role of CFH using a cecal ligation and puncture (CLP) murine polymicrobial sepsis model.¹⁰ Elevated plasma CFH levels only occurred following “high grade” CLP (using a cecal ligature resulting in 80–90% lumen occlusion and multiple cecal punctures), whereas elevated CFH levels did not occur following “low grade” CLP (with 50–60% lumen occlusion and a single cecal puncture).¹⁰ Notably, the administration of exogenous heme, the iron-containing porphyrin subunit of hemoglobin, to mice treated with “low grade” CLP resulted in a marked increase in markers of organ dysfunction and increased mortality,¹⁰ indicating that heme can potentiate the toxic systemic effects of sepsis. Furthermore, mice with genetic knockout of heme-oxygenase 1 (Hmox-1), a rate-limiting enzyme in the metabolic degradation of heme and an important antioxidant enzyme, also had increased organ dysfunction and death.¹⁰ Taken together with the data from human studies, it is clear that sepsis on its own can induce release of CFH, and high CFH levels can serve as a biomarker of increased risk of organ dysfunction and death.

4. Mechanisms of renal toxicity by cell-free hemoglobin

The release of CFH into the circulation can have multiple toxic effects on the kidney (Table 1).³⁵

Table 1.

Potential mechanisms of hemoglobin renal toxicity in sepsis-AKI

Tubular obstruction
Extracellular hemoglobin forms small (32kD) αβ-chain heterodimers that are freely filtered by the glomerulus
Breakdown of filtered hemoglobin in urinary space releases heme
Heme accumulation forms pigment casts in urinary space
Tubular obstruction promotes proximal tubular endocytosis of heme and tubular epithelial injury
Extravascular translocation of hemoglobin
Hemoglobin αβ-heterodimers can translocate through the vascular walls of peritubular capillaries
Breakdown of hemoglobin releases free heme and causes iron deposition in the renal interstitium
Reaction with nitric oxide (NO)
Consumes NO leading to vasoconstriction and endothelial dysfunction
Reaction of oxyhemoglobin with NO generates ferric (Fe3+) hemoglobin
Generation of reactive oxygen species
Ferric (Fe3+) and ferryl (Fe4+) iron hemoglobin outside of
Globin radicals can drive lipid and protein peroxidation
Free heme oxidizes LDL, promoting inflammation and cytotoxicity
Increased oxidative stress promotes hemolysis and more CFH release
Heme-mediated immune signaling
Heme potentiates neutrophilic inflammation and inhibits neutrophil apoptosis
Heme can bind TLR4 triggering pro-inflammatory signaling

4.1. Tubular obstruction

Much of the CFH released during hemolysis dissociates into αβ-heterodimers than can pass through the glomerular filtration barrier and are excreted unchanged in the urine.³⁶ Breakdown of CFH in the urinary space releases heme, which in sufficient quantities can accumulate into pigment casts that can cause tubular obstruction.^37,38 Similar tubular obstruction with pigment casts is also observed during rhabdomyolysis, when release of large quantities of the hemoprotein myoglobin into the circulation leads to heme accumulation in the urinary space.^37,38 Rhabdomyolysis is discussed in more detail later in this review. Although tubular obstruction from pigment casts is often transient and its effects on filtration pressures are unclear,³⁷ even transient periods of stasis may promote increased endocytosis of toxic metabolites by the proximal tubular epithelium, resulting in hemosiderin deposition in the surrounding interstitium, tubular epithelial injury, and tubular necrosis.^37,38

4.2. Extravascular translocation and release of free heme

CFH αβ-heterodimers can also directly translocate through the peritubular capillaries and accumulate in the renal interstitium, resulting in detectable CFH levels in renal lymphatics in experimental animal models.³⁹ CFH translocation into the renal interstitium may result in prolonged retention of CFH in the kidneys even after CFH has cleared from the plasma. Breakdown of the globin protein outside of the RBC releases the iron-complexed porphyrin heme into the circulation or extravascular space (free heme). The complexed iron in free heme can react with cell membranes, plasma proteins, and lipids, and the heme ring can serve as a ligand for other molecular signaling pathways as described below.

4.3. Oxidative stress and reactive oxygen species generation

CFH represents a significant source of oxidative stress due to the chemical reactivity of the iron moiety of heme. The RBC cytoplasm normally serves as a reducing environment that maintains the heme-complexed iron in a reduced ferrous (Fe²⁺) state.⁴⁰ Once released outside of the RBC cytoplasm, the iron moiety can undergo oxidation to form the more reactive ferric (Fe³⁺) and ferryl (Fe⁴⁺) states.^40,41 These highly reactive oxidized iron species can in turn drive formation of other reactive oxygen species including lipid peroxides^40,41 and F₂-isoprostanes,^40,41 leading to increased oxidative stress and cellular injury. Furthermore, since the RBC itself is sensitive to oxidative stress, CFH-mediated formation of reactive oxygen species can induce a positive feedback loop that promotes RBC plasma membrane instability, leading to increased hemolysis and release of more CFH.²⁵

4.4. Nitric oxide consumption

CFH also consumes nitric oxide in regions of both high and low oxygen tension,^42,43 with the resultant NO depletion leading to vasoconstriction and endothelial dysfunction.^42,43 This mechanism is thought to drive the acute hypertensive response observed in massive hemolysis as well as chronic vascular disease in patients with chronic hemolytic disorders like sickle cell anemia.⁴² Sepsis has variable effects on macrovascular and microvascular renal blood flow (RBF)^44,45,46 and both vasoconstriction and vasodilation have been observed simultaneously in different vascular beds. Inducible nitric oxide synthetases (iNOS) and NO signaling are markedly upregulated in the kidney during sepsis and have critical roles in renal control of microvascular function and local blood flow.^47,48 Although pharmacological inhibition of iNOS (decreasing local NO in the kidney) may improve renal capillary perfusion in experimental animal models,^47,48 the diffuse nature of CFH-mediated NO depletion may also overwhelm the kidney’s ability to regulate RBF and glomerular filtration at the level of the glomerular afferent and efferent arterioles, leading to impaired control of microvascular perfusion and worsened renal injury.

4.5. CFH-mediated immune dysregulation

Lastly, the byproducts CFH can interact with the immune system through multiple pathways that contribute to immune dysregulation. Heme can potentiate neutrophilic inflammation by stimulating neutrophil chemotaxis,⁴⁹ elaboration of the reactive oxidative burst,⁴⁹ and inhibiting neutrophil apoptosis.⁵⁰ CFH also can interact with the innate immune system through Toll-like receptors (TLR).^51,52 In particular, CFH can bind to pathogen-associated molecular patterns (PAMPs) such as lipopolysaccharide (LPS) and lipoteichoic acid (LTA) which can then modulate the binding of these PAMPs to TLR-4 and TLR-2, respectively, altering expression of downstream cytokines.⁵² Free heme can also directly bind to TLR4.^51,53 Although it is not clear if these pathways are relevant to CFH-mediated AKI in humans, TLR4 is constitutively expressed in renal vascular walls and may also be present on the apical membrane of renal tubular epithelial cells,⁵⁴ and pharmacological blockade or genetic deletion of TLR4 signaling can protect against renal injury in various experimental models of sepsis-AKI.⁵⁴ Therefore, heme-TLR4 interactions may play an important role in sepsis-AKI when plasma CFH is elevated.

5. Renal toxicity of circulating hemoproteins in experimental models

Several lines of evidence from experimental animal models suggest that CFH and free heme can exert nephrotoxic effects and are reviewed here.

5.1. Myoglobin and rhabdomyolysis

There is clear evidence that elevated levels of circulating hemoproteins other than hemoglobin can induce AKI. Conditions such as major crush injury, extreme physical exertion, malignant hyperthermia, or severe electrolyte disturbances can trigger massive skeletal muscle myocyte death, termed rhabdomyolysis, resulting in release of large quantities of the hemoprotein myoglobin into circulation.^37,41 Myoglobin is a monomeric heme-containing globin protein found in skeletal and heart muscle that is structurally related to hemoglobin and serves as an oxygen storage protein for myocytes.⁵⁵ Because of its structural and chemical similarities to CFH, elevated circulating myoglobin exerts nephrotoxic effects though mechanisms similar to CFH including tubular obstruction by heme casts, deposition of iron in the renal interstitium, and generation of lipid peroxidation products.^41,56,57 In rodent models, both direct venous perfusion with exogenous myoglobin⁵⁶ and intramuscular injection with 50% glycerol to induce muscle necrosis^41,57 consistently induce AKI, particularly in conjunction with other predisposing conditions including volume depletion, renal ischemia, or metabolic acidosis.⁵⁶ Although myoglobin shares may common mechanistic effects with CFH, some features, particularly interactions with TLRs, appear to be more relevant to CFH and have not been described for myoglobin.

5.2. CFH infusion and transfusion-associated hemolysis

Several studies have examined the effects of CFH in isolation on renal function in canine, guinea pig, and rodent animal models. Direct intravenous infusion of exogenous CFH results in hemoglobinuria and an acute hypertensive response in dogs and guinea pigs^43,58 as does exposure to CFH from transfusion of stored blood in guinea pigs.⁵⁹ These hypertensive changes are accompanied by vascular endothelial dysfunction and subendothelial vascular injury.⁵⁹ Transfusion with older stored RBCs reliably leads to increased release of CFH and heme compared to transfusion with younger fresh RBCs both in guinea pig and canine models using massive exchange transfusion^59,60 and in a mouse model of hemorrhagic shock with RBC resuscitation.⁶¹ In both settings, the release of CFH results in renal histopathologic changes including increased deposition of free iron,^58,59,61, lipid peroxidation,⁵⁸ and renal tubular dilation and necrosis.^58,59,61 It is also accompanied by increased kidney expression of biochemical markers of renal tubular injury such as neutrophil gelatinase-associated lipocalin (NGAL) and kidney injury molecule-1 (KIM-1).⁶¹ The impact on glomerular filtration is less clear, as elevated CFH levels decreased glomerular filtration as measured by serum creatinine in some models,⁵⁹ while in other models CFH did not affect glomerular filtration.^60,61 Interestingly, several groups have shown that these renal changes could be abrogated by augmenting plasma haptoglobin (Hp) levels to capture and sequester CFH before it could reach the kidneys, either by pre-treating dogs with glucocorticoids to induce hepatic expression of endogenous Hp⁵⁸ or by directly infusing additional exogenous Hp.^58,59,61

5.3. CFH during sepsis

The effects of free heme and CFH on renal function during sepsis have been specifically tested in two experimental animal models.^10,13 As described above, Larsen et al. reported the effects of free heme in a murine CLP model. Transgenic mice deficient in Hmox1 (Hmox1^−/−) were less able to metabolize circulating free heme and CFH compared to wild type (Hmox1^+/+) mice. Compared to Hmox1^+/+ mice, Hmox1^−/− mice had increased evidence of renal injury as indicated by increased tubular epithelial necrosis on renal histology and decreased glomerular filtration as indicated by increased blood urea nitrogen (BUN) levels following CLP-induced sepsis.¹⁰ Furthermore, intraperitoneal administration of exogenous heme to septic Hmox1^+/+ mice resulted in increased markers of renal injury, and blocking free heme with exogenous hemopexin was protective against renal injury.¹⁰

Shaver et al. tested the effects of CFH on kidney injury in a murine polymicrobial peritonitis sepsis model.¹³ Polymicrobial peritonitis was induced using an intraperitoneal injection of cecal slurry (IP CS) from a donor mouse, a method which recapitulates most features of the more common CLP model,⁶² while facilitating more reproducible control over the inflammatory response and degree of lethality as the dose of cecal slurry is more easily titrated than CLP.⁶² Treatment with a sub-lethal dose of IP CS resulted in modest elevations in CFH levels at 24 hours¹³ which were accompanied by similarly modest kidney mRNA expression of renal epithelial injury markers NGAL and KIM-1, but without substantial changes in GFR at 24 or 48 hours.¹³ In contrast, simultaneous treatment with both IP CS and intravenous CFH to augment CFH levels resulted in markedly increased NGAL and KIM-1 mRNA expression. This was accompanied by histologic evidence of renal tubular injury with tubular distortion and detachment of tubular epithelial cells from the basement membrane, findings not observed in mice treated with IP CS alone. The IP CS + IV CFH mice were also the only treatment group to experience decreases in GFR over the 48-hour period following treatment.¹³

Taken together, the data from model animal studies using myoglobin, CFH in isolation, and CFH during sepsis strongly support the hypothesis that heme and CFH are nephrotoxic, and that CFH potentiates acute kidney injury during sepsis.

6. Clinical evidence for renal toxicity of cell-free hemoglobin in conditions other than sepsis

The relationship between CFH and AKI has been studied extensively in conditions other than sepsis, which may provide additional insight into the potential impact of CFH in sepsis-AKI.

6.1. Cardiopulmonary Bypass Surgery

The clinical effects of CFH on renal function are perhaps best characterized in the setting of cardiopulmonary bypass (CPB) to facilitate cardiac surgery.^63–65 Hemolysis is a well-recognized consequence of CPB,⁶⁶ and is caused by several factors including direct mechanical RBC injury from contact with non-endothelial surfaces such as tubing and pumps, interaction with the gas-permeable membrane of the oxygenator, activation of the coagulation cascade, and RBC transfusion to replace blood loss during the procedure.⁶⁶ Plasma CFH levels above 50 mg/dL are routinely encountered during CPB,^63,64 and associations between CFH levels during or immediately following surgery and risk of subsequent postoperative AKI have been noted in both adults^63,67 and children.⁶⁴ A study by Billings et al. merits particular note as the investigators examined the relationship between CFH levels, oxidative stress, and AKI.⁶⁷ CPB patients who experienced post-operative AKI had higher peak CFH levels as well as higher plasma levels of the oxidative stress markers F₂-isoprostanes and isofurans during the surgical procedure.⁶⁷ Moreover, peak CFH levels were correlated with higher peak plasma isofuran levels during the surgery, both in patients with and without AKI. Lastly, patients with AKI had sustained elevations in urine isofuran levels for up to two days postoperatively, suggesting the AKI patients experienced sustained periods of increased oxidative stress.⁶⁷ Taken together, this data supports the contention that CFH-mediated oxidative stress can play a significant role in driving AKI in these patients.

6.2. Transfusion-associated hemolysis

Although clinical studies of the renal effects of RBC transfusion in critically ill adults have reported variable results,^68,69 hemolysis during RBC transfusion has clear effects on systemic and pulmonary vascular endothelial function. In a study of 18 healthy volunteers comparing autologous transfusion of 5-day-old RBCs versus 42-day-old RBCs, Risbano et al. reported that transfusion with older blood increased post-transfusion hemolysis, with resultant depletion of plasma NO and impaired vascular endothelial function as measured by acetylcholine-dependent forearm blood flow.⁷⁰ Berra et al. reported that transfusion with 40-day-old RBCs similarly led to greater increases in mean pulmonary artery pressure compared with 3-day-old RBCs in 14 obese but otherwise healthy volunteers.⁷¹ These observations may have clinical relevance in sepsis-AKI as endothelial dysfunction is a prominent feature of sepsis even in the absence of RBC transfusions,⁷² and normal renal homeostasis depends on tight regulation of endothelial barrier function and microvascular perfusion.⁷³ Therefore, it is reasonable to hypothesize that CFH-mediated disruption of normal endothelial function during sepsis could affect the renal microvasculature. However, this mechanism has not yet been directly evaluated in sepsis-AKI.

6.3. Sickle Cell Anemia

The inborn genetic hemoglobinopathies including sickle cell disease (SCD) and the thalassemias are among the most common monogenetic disorders in the world.^74,75 These disorders are frequently characterized by chronic hemolysis and punctuated by episodes of acute hemolysis that quickly overwhelms endogenous CFH processing mechanisms.⁷⁵ The kidney is particularly vulnerable to injury in SCD because of the tendency for HbS to polymerize in the renal medulla, where the partial pressure of oxygen is low, leading to vascular occlusion by sickled RBCs and subsequent renal infarction with papillary necrosis.⁷⁵ Although mechanisms including endothelial dysfunction, volume depletion, medications, and acute infection all contribute to AKI in SCD,⁷⁵ the presence and degree of an SCD patient’s “hemolysis phenotype” also appears to influence renal outcomes. Degree of hemolysis as measured either by biochemical markers or drop in hemoglobin concentration are associated with increased risk of AKI in both children and adults presenting with acute vaso-occlusive crises.^76,77 Furthermore, hemoglobinuria, a marker of chronic ongoing hemolysis, was associated with increased incidence of chronic kidney disease (CKD) as well as progression of CKD in a large longitudinal study of African-American adults with SCD.⁷⁸

6.4. Malaria

Malaria is another condition where CFH release can exert substantial effects on the kidney.³³ Blackwater fever, a syndrome of malaria-induced massive intravascular hemolysis, hemoglobinuria, and acute renal failure, has been recognized for over a century.⁷⁹ Increased hemolysis in malaria has been associated with abnormal renal function,³³ although most studies have not specifically examine the correlation between CFH and AKI. Most recently, there has been interest in understanding the mechanistic relationship between parasite burden, CFH release, and renal function. In an observational study of 107 patients with severe malaria in Bangladesh, Plewes et al. found that CFH levels were higher in patients with AKI compared to those without AKI.³² Unsurprisingly, higher CFH levels were also correlated with higher parasite burdens and other biochemical markers of hemolysis. Perhaps more interesting, the investigators also noted that among those patients admitted to the hospital initially without AKI, increased oxidative stress markers (plasma F₂-isoprostanes and isofurans) and decreased RBC deformability predicted subsequent creatinine elevation and need for hemodialysis.³² This observed relationship between the mechanistic factors that contribute to abnormal CFH biology and AKI risk provides further evidence that CFH is driving the increased risk of AKI in severe malaria patients.

7. Targeting cell-free hemoglobin in sepsis-induced acute kidney injury

Since the release of CFH into circulation is a common feature of sepsis and can cause renal dysfunction in experimental models, CFH represents an obvious therapeutic target for sepsis-AKI, with several potential mechanistic approaches. Potential strategies to mitigate the effects of CFH during sepsis-AKI include limiting release of CFH, increasing scavenging of CFH, and decreasing CFH-mediated oxidative injury.

7.1. Limiting CFH release

Prevention of hemolysis, and thus prevention of the release of CFH from the RBC, is an attractive target. However as discussed previously, multiple physiologic processes influence CFH release during sepsis, and many of these factors are not clearly amenable to pharmacological intervention.¹⁷ Therefore, few studies specifically targeting the mechanisms of hemolysis during sepsis have been reported.

Limiting RBC transfusions to when hemoglobin concentration falls below 7.0 g/dL (restrictive transfusion) may reduce exposure to transfusion-associated hemolysis, and thus may indirectly mitigate CFH-mediated renal injury during sepsis. However, data on the renal effects of transfusion thresholds in sepsis are limited.⁶⁹ A recent meta-analysis of transfusion threshold in sepsis or septic shock identified only two randomized trials that reported on renal failure at 28 days following admission.⁶⁹ Neither study reported on AKI, change in creatinine, or other renal biomarkers.⁶⁹ No significant differences were noted between patients assigned to restrictive or liberal strategies, although these conclusions are limited by the small samples and low event rates, with a total of only 83 events in 1277 patients studied.⁶⁹ Nevertheless, since a restrictive transfusion threshold both conserves a scarce resource and does not increase risk of other adverse outcomes,⁶⁹ limiting RBC transfusion is a reasonable strategy to mitigate risk of CFH exposure.

7.2. Haptoglobin

Release of sufficient CFH into the vascular compartment during sepsis can overwhelm endogenous scavenger mechanisms. Therefore, supplementation with exogenous scavenger proteins could prevent CFH from causing downstream effects. Haptoglobin (Hp) is the primary endogenous scavenger of CFH in mammals.¹⁵ A dimeric plasma protein synthesized by the liver, Hp binds irreversibly to CFH.³⁵ The resultant CFH-Hp complex binds to the CD163 receptor present on monocytes and macrophages,⁸⁰ triggering endocytosis and clearance of CFH from the circulation by the reticuloendothelial system.³⁵ Hp is an acute phase reactant protein in humans, leading to variable Hp levels in sepsis.^81,82 In a study of 387 patients with severe sepsis, lower plasma Hp levels were independently associated with increased hospital mortality.⁸² Notably, the relationship between plasma Hp level and mortality was only observed in patients with elevated CFH levels, whereas Hp level did not affect mortality in patients without elevated CFH,⁸² further suggesting that circulating Hp is critical for limiting the toxicity of CFH during sepsis.

Several experimental studies support a potential therapeutic role for supplemental Hp to mitigate CFH-mediated renal injury. As noted previously, augmenting plasma Hp levels with either glucocorticoids⁵⁸ or exogenous Hp^59,61 abrogates the renal toxicity of CFH in animal models of transfusion-associated hemolysis. Using a canine model of severe S. aureus pneumonia, Remy et al. showed that purified human Hp improved survival, and reduced vasodilatory shock, lung injury scores and circulating levels of non-transferrin bound iron compared to infusion of human albumin.⁶⁰ The beneficial effects of supplemental Hp were accentuated in animals also treated with large-volume exchange transfusion to increase circulating CFH levels,⁶⁰ suggesting that the benefit of haptoglobin supplementation was related primarily to improved scavenging of CFH. Mean creatinine levels were lower among the Hp-treated group at several time-points, although the differences were limited by small sample size and not statistically significant.⁶⁰

Purified human Hp is commercially available for clinical use in Japan and is approved for treatment of severe hemolysis during extracorporeal cardiopulmonary bypass, severe burn injuries, or following massing transfusion. Several uncontrolled trials in Japan have suggested that intraoperative haptoglobin supplementation during CPB can reduce risk of subsequent postoperative AKI,^83,84 but prospective trials have been limited by small sample sizes. To date, no clinical trials have tested haptoglobin supplementation in patients with sepsis.

7.3. Hemopexin

As noted previously, breakdown of CFH can release the iron-complexed porphyrin heme (free heme) into circulation. As free heme can cause toxicity in its own right, it necessitates its own clearance mechanisms. Hemopexin (Hpx) serves as the primary endogenous scavenger of free heme in humans.^16,85 Hpx is a glycoprotein synthesized by the liver and is normally present in the circulation in high concentrations.^16,85 Although other plasma proteins including albumin and lipoproteins of LDL or HDL particles can also bind porphyrins, Hpx has the highest binding affinity for heme of any protein (K < 10⁻¹³M).^35,85 Once bound to Hpx, heme is delivered primarily to the liver, preventing the pro-oxidant and pro-inflammatory effects of heme.⁸⁶ Heme binding by Hpx also prevents heme from intercalating in the plasma cell membrane where it can drive lipid peroxidation.⁸⁷ Binding of Hpx to the CD91 receptor on hepatocytes and phagocytic cells triggers endocytosis of the Hpx-heme complex, subsequent degradation of heme by HO-1, and degradation or recycling of Hpx.⁸⁸

In contrast to Hp, the role of Hpx in the acute phase response remains controversial.^85,89 In several observational studies of critically ill septic adults, non-survivors of sepsis had lower plasma Hpx levels than survivors,^10,82 although many sepsis patients still have Hpx levels within the reference range of 0.5 to 1.5 mg/dL.^16,85 Hpx supplementation has been explored in experimental animal models. As noted previously, Larsen et al. found that Hpx supplementation mitigated renal injury in mice with intraperitoneal sepsis, whereas treatment with equivalent amounts of non-heme-binding IgG did not improve renal injury.¹⁰ In contrast, Graw et al. found that Hpx supplementation did not prevent renal injury in a mouse model of hemorrhagic shock with massive RBC resuscitation.⁶¹ Currently, Hpx has not been tested in clinical trials of sepsis and is not available for clinical use.

7.4. Acetaminophen

The common analgesic and antipyretic acetaminophen (APAP) has also been identified as a potential therapy for CFH-mediated organ injury. Although traditionally prescribed for its effects on cyclooxygenase-prostaglandin pathways, at clinically relevant doses, APAP can reduce ferryl (Fe⁴⁺) complexed iron in CFH to the less reactive ferric (Fe³⁺) species.^41,90 The specificity of APAP for Ferryl-CFH reduction is due to structural similarity between the heme moiety of CFH and the peroxidase moiety of cyclooxygenase.⁴¹ This was experimentally demonstrated in a rat rhabdomyolysis model of AKI. Boutaud et al. observed that treatment with APAP before or after induction of rhabdomyolysis reduced markers of oxidative stress, reduced heme-to-protein cross-linking from peroxide reactions, and mitigated AKI (Figure 1).⁴¹ APAP’s unique hemoprotein reductant activity is not present in non-specific anti-oxidants such as vitamin E and N-acetyl cysteine at doses that have been studied in clinical sepsis.^91,92 Thus, APAP has the potential to target CFH-mediated organ dysfunction in sepsis where other antioxidants have failed.

Figure 1.

Renoprotective effects of acetaminophen in rats with rhabdomyolysis. (A) Gross examination of a representative kidney from a normal rat (Left), a rat with rhabdomyolysis (Center), and a rat with rhabdomyolysis treated with acetaminophen (APAP) (Right). Effect of treatment with APAP on reactive oxygen species (B) and on kidney function (C) in normal rats (controls), in normal rats treated with APAP (control + ApAP), in rats with rhabdomyolysis (Rhabdo), and in rats with rhabdomyolysis treated with APAP (Rhabdo + ApAP). (B) Urinary excretion of F₂-isoprostanes. **, P < 0.0001 for Rhabdo vs. controls; *, P < 0.005 for Rhabdo + APAP vs. controls. (C) Plasma creatinine levels. **, P < 0.0001 for Rhabdo vs. controls; *, P = 0.0002 for Rhabdo + APAP vs. controls. For both figures, the brackets indicate that APAP significantly reduced urinary F₂-isoprostanes and plasma creatinine compared with Rhabdo alone. (n ≥ 6 in each group). Abbreviations: ApAP – Acetaminophen. F₂-IsoPs − F₂-isoprostanes. Reproduced with permission.⁴¹

Several small clinical trials have tested APAP specifically for mitigating CFH-mediated organ injury.^65,93–95 Collectively, these trials suggest a beneficial effect of APAP on hemoglobin-mediated oxidative injury and AKI. APAP demonstrated favorable effects on plasma levels of lipid peroxidation products in two clinical trials in children or adults undergoing cardiac surgery with CPB.^65,95 Although no differences in post-operative AKI were noted in these trials, they were limited by small sample sizes and low event rates.^65,95 Janz et al. reported a phase 2a randomized placebo-controlled clinical trial of APAP in 40 patients with severe sepsis and elevated plasma CFH measured at enrollment. Patients who received enteral APAP at 1 gram every 6 hours for 3 days had significantly reduced oxidative stress as measured by plasma F -isoprostanes.⁹³ APAP-treated patients also had lower serum creatinine levels on days 3 and 4 after randomization as well as lower peak serum creatinine through the remainder of the hospitalization (Figure 2).⁹³ Finally, in a randomized clinical trial in 61 adults with severe falciparum malaria in Bangladesh and Thailand, Plewes et al. showed a clear beneficial effect of APAP on both lipid peroxidation and AKI as measured by relative change in serum creatinine over the first 72 hours after enrollment (Figure 3).⁹⁴ The renal benefits of APAP were confined to patients with higher plasma CFH levels (> 4.5 mg/dL) as well as those with higher levels of lipid peroxidation products on enrollment (Figure 3).⁹⁴ Also interestingly, a secondary analysis using pharmacokinetic-pharmacodynamic (PK–PD) mixture model showed that higher total APAP exposure over 72 hours correlated with a lower probability of declining renal function.⁹⁴

Figure 2.

Creatinine measurements over study period in the Acetaminophen for the Reduction of Oxidative injury in Severe Sepsis (ACROSS) Trial. In patients never receiving any type of renal replacement therapy during or after the study (n = 36), creatinine levels were similar at baseline (study day 0) and were significantly reduced in the acetaminophen group on study day 3, the day following study completion (day 4), through the remainder of hospitalization, and at discharge or death. P values represent between-group comparisons at each time point. Circles and squares represent medians of placebo and acetaminophen groups, respectively. Whiskers represent interquartile ranges (between 25th and 75th percentiles). Table under figure represents number of patients (n) at each time point. Reproduced with permission.⁹³

Figure 3.

Effect of acetaminophen on serum creatinine in randomized trial of severe falciparum malaria. Points represent mean percent change in creatinine from baseline at 12, 24, 36, 60, 48, and 72 hours in the entire cohort (A) and in patients stratified by level of intravascular hemolysis (B and C). Plasma cell-free hemoglobin (CFH) ≥45000 ng/mL (B); plasma CFH <45000 ng/mL (C). Frequencies in rows below figures represent number of patients (n) at each time point. P value represents overall treatment effect. Abbreviations: Cr, creatinine; CFH, cell-free hemoglobin. Reproduced under the terms of the Creative Commons CC BY license from reference,⁹⁴ which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Notably, all of the aforementioned trials specifically targeted patients with either predicted or measured elevations of plasma CFH. This stands in contrast to the HEAT Trial, a large multi-center randomized clinical trial of APAP for ICU patients with fever and suspected infection, which did not observe any differences in renal outcomes or days free from renal failure.⁹⁶ The HEAT Trial specifically tested the hypothesis that allowing “permissive hyperthermia” (by avoiding APAP for fever) would improve outcomes compared to standard treatment of fever with APAP. The trial did not report CFH or myoglobin levels, and the incidence of conditions associated with elevated circulating hemoproteins such as clinical rhabdomyolysis or DIC were either low or not reported.⁹⁶ Furthermore, the total APAP exposure of the treatment group was lower (median 8 grams over 2 days) than in the trials by Janz et al. and Plewes et al. (12 grams over 3 days).^93,94 Therefore, although APAP may not be beneficial in an unselected ICU population, the HEAT trial does not exclude the potential value of APAP for CFH-mediated organ dysfunction. Collectively, these data highlight both the potential use for plasma CFH levels as a predictive biomarker for sepsis-AKI, and that any future clinical trials of a pharmacological agent targeting CFH during sepsis will need to consider specific enrichment for patients with elevated CFH levels.

7.5. Other strategies

Several other therapeutic agents have the potential to target CFH-mediated injury in sepsis but are less well studied. Vitamin C (ascorbic acid) has hemoprotein reductant activity and can prevent the increase in endothelial permeability induced by exposure of human endothelial cells to CFH.⁹⁷ Vitamin C has also shown some promise in observational studies of sepsis,⁹⁸ but whether this is related to effects on circulating CFH is not known. Notably, the recently published CITRIS-ALI study, a randomized trial of intravenous vitamin C in patients with sepsis-induced acute respiratory distress syndrome, demonstrated no differences in renal outcomes between patients treated with vitamin C or placebo despite significant differences in mortality.⁹⁹ Therapies that target free iron might also be beneficial in the setting of elevated CFH levels in sepsis,¹⁰⁰ however no studies have been reported in humans to date.

7.6. Current barriers to therapeutics targeting CFH

In order to use circulating CFH levels as a tool for predictive enrichment in clinical trials of CFH-targeting therapies, a rapid and reliable point-of-care measurement of circulating CFH is needed. Currently, there is no rapid bedside test available for CFH. Although HemoCue® America markets a point-of-care device that can measure low levels of CFH (HemoCue® Plasma/Low Hb System), this device is optimized for blood banking rather than measurement of the relatively low levels of plasma cell-free hemoglobin that have been documented in sepsis patients.^11,94 Although other methods are available for measurement of plasma cell-free hemoglobin such as an enzyme-linked immunosorbent assay,⁹⁴ none are yet rapidly available at the bedside for clinical trial enrollment.

8. Conclusions

Sepsis is a heterogeneous clinical syndrome that has as yet defied all attempts to identify effective pharmacological therapies. Sepsis-AKI is similarly in need of effective therapies as it remains associated with higher costs, longer hospital stays, and worse clinical outcomes. A growing body of evidence suggests that an elevated level of circulating CFH is a feature of some, but not all patients with sepsis, and CFH may play a role in the pathophysiology of sepsis-AKI through a variety of injurious mechanisms. Several potential therapies to target CFH in sepsis have been identified including haptoglobin, hemopexin and acetaminophen. Bedside measurement of CFH levels may facilitate predictive enrichment for future therapeutic trials of these CFH-targeted therapeutics such that only patients with elevated CFH, who would be most likely to benefit, would be enrolled. However, rapid, accurate bedside tests for plasma CFH will need to be developed in order for such trials to move forward.

Abbreviations

CFH

cell-free hemoglobin