Research ArticleKidney Disease

The protective role of macrophage migration inhibitory factor in acute kidney injury after cardiac surgery

See allHide authors and affiliations

Science Translational Medicine  16 May 2018:
Vol. 10, Issue 441, eaan4886
DOI: 10.1126/scitranslmed.aan4886

MIF muffles kidney injury

Patients undergoing open-heart surgery are susceptible to complications, including acute kidney injury (AKI). Stoppe et al. observed that patients who had high serum concentrations of macrophage migration inhibitory factor (MIF) after cardiac surgery had lower risk of AKI. Renal ischemia-reperfusion injury induced more inflammation and kidney cell death in mice lacking Mif than in wild-type mice. Treating isolated kidney cells with MIF protected against hypoxia-induced cell death. Mice treated with MIF or a soluble form of CD74 (MIF receptor) showed reduced kidney injury after ischemia-reperfusion. This study suggests that MIF may protect the kidney from ischemia-reperfusion injury.

Abstract

Acute kidney injury (AKI) represents the most frequent complication after cardiac surgery. Macrophage migration inhibitory factor (MIF) is a stress-regulating cytokine that was shown to protect the heart from myocardial ischemia-reperfusion injury, but its role in the pathogenesis of AKI remains unknown. In an observational study, serum and urinary MIF was quantified in 60 patients scheduled for elective conventional cardiac surgery with the use of cardiopulmonary bypass. Cardiac surgery triggered an increase in MIF serum concentrations, and patients with high circulating MIF (>median) 12 hours after surgery had a significantly reduced risk of developing AKI (relative risk reduction, 72.7%; 95% confidence interval, 12 to 91.5%; P = 0.03). Experimental AKI was induced in wild-type and Mif−/− mice by 30 min of ischemia followed by 6 or 24 hours of reperfusion, or by rhabdomyolysis. Mif-deficient mice exhibited increased tubular cell injury, increased regulated cell death (necroptosis and ferroptosis), and enhanced oxidative stress. Therapeutic administration of recombinant MIF after ischemia-reperfusion in mice ameliorated AKI. In vitro treatment of tubular epithelial cells with recombinant MIF reduced cell death and oxidative stress as measured by glutathione and thiobarbituric acid reactive substances in the setting of hypoxia. Our data provide evidence of a renoprotective role of MIF in experimental ischemia-reperfusion injury by protecting renal tubular epithelial cells, consistent with our observation that high MIF in cardiac surgery patients is associated with a reduced incidence of AKI.

INTRODUCTION

Open-heart surgery is associated with considerable morbidity, with acute kidney injury (AKI) being one of the most frequently observed postoperative complications (1). Postoperative AKI determines the outcome and lethality of patients, resulting in considerable costs for the health care system (24). Hemodynamic instability, an excessive inflammatory response, and perioperative release of reactive oxygen species (ROS) are involved in the pathogenesis of cardiac surgery–related AKI (1, 5, 6). Prolonged anesthesia, perioperative drug treatment, operative trauma, and patient positioning in the setting of cardiac surgery can contribute to rhabdomyolysis (skeletal muscle breakdown) and further increase the risk of postoperative AKI (79).

Macrophage migration inhibitory factor (MIF) is an inflammatory and stress-regulating cytokine with chemokine-like functions that is rapidly released in response to various stimuli from preformed, intracellular pools, in contrast to other innate mediators that require transcriptional activation and synthesis before release (10, 11). MIF exerts its effects via binding and activating its receptors, CXCR2, CXCR4, and CD74, with subsequent recruitment and phosphorylation of its co-receptor CD44. CXCR2 and CXCR4 are chemokine receptors involved in inflammatory actions of MIF and in the pathogenesis of various diseases, including myocardial infarction, atherosclerosis, and sepsis (12, 13). In contrast, CD74 was suggested to mediate protective effects of MIF, particularly in cardiac ischemia-reperfusion (I/R) injury (1416). Recently, a soluble form of the CD74 ectodomain, sCD74, was identified (17). In observational studies of sepsis and liver transplantation, increased MIF in the serum was associated with severe AKI (18, 19). Comparable to sepsis, cardiac surgery with use of cardiopulmonary bypass (CPB) is associated with systemic oxidative stress and hemodynamic changes, resulting in a transiently decreased perfusion of organs, including the kidney. We and others have demonstrated that MIF is elevated during the perioperative inflammatory response after cardiac surgery (20, 21). Several studies have revealed cardioprotective effects of MIF in the setting of acute myocardial I/R injury (20, 22, 23), but the functional role and clinical relevance of MIF in the development of postoperative AKI remain unknown.

Renal tubular epithelial cell injury and cell death contribute to the pathophysiology of AKI (24). The role of apoptosis as the key mechanism of tubular injury and AKI has recently been challenged (2527). We and others have previously demonstrated alternative pathways of regulated cell death, in particular necroptosis and ferroptosis, as crucial pathophysiological pathways of AKI (25, 26). Necroptosis is a form of regulated necrosis and is mediated by receptor-interacting protein kinase-3–dependent phosphorylation of mixed lineage kinase domain-like (MLKL) (24). When MLKL is phosphorylated, it leads to permeabilization of cellular membranes and necrotic cellular demise (28). Oxidative stress is considered a relevant stimulus for the development of cardiac surgery–associated AKI by driving cell death, in particular ferroptosis (5). Ferroptosis is another form of regulated necrotic cell death. It is characterized by iron-dependent lipid peroxidation and is initiated by intracellular glutathione (GSH) depletion (24). A currently favored hypothesis deems ferroptosis as a trigger of subsequent synchronized renal tubular necrosis (29).

MIF was shown to have antioxidant properties (2024), potentially by its thiol-protein oxidoreductase (TPOR) activity. Here, we tested the hypothesis that MIF might be a protective factor limiting AKI through cytoprotective and antioxidative effects, using a comprehensive combined clinical and experimental approach including the state-of-the-art assessment of apoptosis, necroptosis, and ferroptosis.

RESULTS

Elevated MIF is associated with reduced incidence of AKI in cardiac surgery patients

Baseline characteristics of the patients included in this prospective, observational study are provided in table S1 and fig. S1. Concurrent with previous findings (30), MIF concentration in serum peaked directly after cardiac surgery, followed by a gradual decline in the postoperative period until day 1 (fig. S2A). Compared to patients with low postoperative serum MIF (below the median value), patients with high serum MIF showed a reduced incidence of AKI 12 hours after surgery (relative risk reduction, 72.7%; 95% confidence interval, 12 to 91.5%; P = 0.030; Table 1 and Fig. 1A). Patients with low serum MIF had a significantly higher incidence of more severe AKI [Acute Kidney Injury Network (AKIN) stage II or stage III] immediately and 12 hours after surgery (10% of patients with low MIF versus 0% patient with high MIF, P = 0.024). Comparing circulating MIF of AKI patients versus non-AKI patients, those with AKI had significantly reduced MIF during the perioperative time course (6 hours, P = 0.021; 12 hours, P = 0.016; Fig. 1B). Likewise, higher urinary MIF was associated with reduced risk of AKI, and urinary MIF in the AKI group was significantly lower during the postoperative time course (P = 0.046; Fig. 1, C and D). Urinary MIF did not correlate with serum MIF (24 hours after surgery, P = 0.628, r = −0.068; fig. S2, B and C).

Table 1 Association of serum MIF and incidence of AKI after cardiac surgery.

AKIN, Acute Kidney Injury Network. P values were calculated using the Fisher’s exact test.

View this table:
Fig. 1 Elevated MIF is associated with reduced incidence of AKI and enhanced antioxidant capacity in cardiac surgery patients.

(A) Percentage of patients with AKI within 72 hours after cardiac surgery, stratified by median serum concentration (n = 60). (B) Perioperative kinetics of serum MIF in patients without postoperative AKI compared to patients with AKI. (C) Incidence of AKI stratified by median urinary MIF concentration. (D) Perioperative kinetics of urinary MIF. (E) Inverse correlation between postoperative serum MIF and postoperative urinary NGAL 24 hours after cardiac surgery. (F and G) Antioxidant capacity of serum samples measured using the RedoxSYS Diagnostic System (Luoxis Diagnostics). Comparison of the antioxidant capacity in serum samples with high serum MIF (>median) with low serum MIF (≤median) at the corresponding time points. (H and I) Comparison of the antioxidant capacity in serum samples of patients with AKI with patients without AKI. NGAL served as a kidney injury marker. Data are means ± SEM. r, Pearson coefficient; R, goodness of fit. (A and C) P < 0.05 analyzed by Fisher’s exact test. (B and D) P < 0.05 versus other groups at the corresponding time point (difference between groups) analyzed by Mann-Whitney U test.

MIF is inversely correlated with postoperative urinary NGAL

The urinary tubular injury marker neutrophil gelatinase-associated lipocalin (NGAL) (31) was significantly (P = 0.03) increased after surgery (fig. S2D). Postoperative serum MIF inversely correlated with postoperative urinary NGAL (0 hours after surgery, P = 0.038; r = −0.321; r = 0.103; 24 hours after surgery, P = 0.009; r = −0.408; r = 0.167) (Fig. 1E).

MIF is associated with the total antioxidant capacity after cardiac surgery

The total antioxidant capacity in the patients’ serum was measured using the RedoxSYS Diagnostic System (Luoxis Diagnostics). After cardiac surgery, the total antioxidant capacity rapidly decreased and remained low during the postoperative time course (Fig. 1F). Patients with high postoperative serum MIF (higher than the median value) immediately (0 hours) after surgery had a significantly elevated total antioxidant capacity when compared to low postoperative serum MIF (lower than the median value) at the corresponding time point (P = 0.04; Fig. 1G). Although the antioxidant capacity in patients with low serum MIF started to decline early between pre-OP and 0 hours after surgery, the total antioxidant capacity remained stable in patients with high postoperative MIF (>median value) directly (0 hours) after surgery and showed a delayed decrease within the first 6 hours after surgery (P = 0.004). No significant differences in the antioxidant capacity were detected at the other perioperative time points. Patients who developed postoperative AKI had significantly reduced total antioxidant capacity before surgery (P = 0.04) and immediately (0 hours) after surgery (P = 0.03) when compared to patients without AKI (Fig. 1, H and I).

Renal ischemia induces MIF release into the plasma

To investigate whether renal ischemia might be associated with MIF release into the bloodstream, we assessed kinetics of circulating MIF in patients with surgery-induced renal ischemia during kidney tumor enucleation. Patients with temporary renal ischemia by clamping of the kidney artery showed a significant increase in MIF during and after surgery (P = 0.01), whereas there was no significant increase in MIF in patients undergoing tumor enucleation without renal ischemia (P = 0.25; Fig. 2A). The duration of renal ischemia correlated positively with postoperative circulating MIF shortly after surgery, in the recovery room, and on the first postoperative day (POD1) (recovery room, P = 0.045; r = 0.676; POD1, P = 0.0005; r = 0.738; r = 0.545; Fig. 2B).

Fig. 2 Renal ischemia as a stimulus for MIF release into the bloodstream.

(A) Kinetics of serum MIF in patients undergoing kidney tumor enucleation. Comparison of patients exposed to renal hypoxia (n = 18) by cross-clamping of the renal artery versus no hypoxia (n = 28). Blood samples were drawn 1 day before surgery (pre-OP), 5 min after tumor enucleation (intra-OP), in the recovery room shortly after the termination of surgery, and on POD1. (B) Association between the duration of renal ischemia and postoperative serum MIF on the first day after surgery. Data are means ± SEM; §§P < 0.01 versus other groups at the corresponding time point (difference between groups) and *P < 0.05 versus baseline analyzed by Mann-Whitney U test.

Mif deficiency aggravates tubular injury and regulates cell death after I/R injury and in rhabdomyolysis-induced AKI

To assess the functional role of MIF in AKI in vivo and to identify potential mechanisms, we compared the degree of acute tubular injury and regulated cell death in wild-type (WT) and Mif-deficient (Mif−/−) mice after uni- or bilateral renal ischemia followed by 6 or 24 hours of reperfusion or 24 hours after induction of rhabdomyolysis by intramuscular glycerol injection. Compared to WT mice, Mif−/− mice showed significantly aggravated AKI as shown by higher serum creatinine (P = 0.04; Fig. 3B). This was reflected histologically by more prominent tubular cell injury in Mif−/− mice when compared to WT, as measured by the tubular injury score (Figs. 3, C and D, and 4, B, C, E, and F, and fig. S3, B and C). The number of apoptotic tubular cells, analyzed by quantification of cleaved caspase-3–positive tubular cells using immunohistochemistry, was significantly higher in Mif−/− mice compared to WT mice in all the experiments (24 hours I/R: cortex P = 0.003, outer medulla P = 0.02; Figs. 3, C and D, 4, B, C, E, and F, and fig. S3, B and C), but the overall number of apoptotic cells remained low. In analyses of the downstream kinase of necroptosis, pMLKL (phosphorylated MLKL), we found a significantly higher number of pMLKL-positive tubules in Mif−/− mice compared to WT mice (24 hours I/R: cortex P = 0.001, outer medulla P = 0.01; Figs. 3, C and D, 4, B, C, E, and F, and fig. S3, B and C). Besides, we detected a significantly higher MLKL expression in the kidneys of Mif−/− compared to WT mice after ischemia by immunohistochemical staining of MLKL and Western blot analysis (24 hours I/R: cortex P < 0.001, outer medulla P = 0.008; fig. S4, A and B).

Fig. 3 Comparison of AKI and tubular injury in WT and Mif−/− mice in an experimental model of AKI induced by unilateral ischemia and 24 hours of reperfusion.

(A) Schematic depicting the induction of acute kidney damage after 35 min of ischemia by cross-clamping of the renal artery and reperfusion for 24 hours in WT mice and Mif−/− mice (n = 5 both groups). (B) Serum creatinine [normalized to body weight (BW)] in WT and Mif−/− mice 24 hours after I/R injury. (C and D) Tubular necrosis was evaluated applying the tubular injury score. Immunohistochemical staining of apoptotic cells, cleaved caspase-3–positive cells, and necroptotic, pMLKL-positive tubules. +/+, WT mice; −/−, Mif−/− mice; hpf, high-power field. Scale bars, 100 μm. Data are means ± SD; *P < 0.05, **P < 0.01 analyzed by Student’s t test.

Fig. 4 Comparison of tubular injury in WT and Mif−/− mice in an experimental model of AKI induced by unilateral ischemia and 6 hours of reperfusion or by rhabdomyolysis.

(A) Schematic depicting the induction of acute kidney damage after 35 min of ischemia by cross-clamping of the renal artery and reperfusion for 6 hours in WT mice and Mif−/− mice (n = 10 both groups). (B and C) Tubular necrosis was evaluated applying the tubular injury score. Immunohistochemical analysis of apoptotic cells, cleaved caspase-3–positive cells, and necroptotic, pMLKL-positive tubules. (D) Schematic depicting the induction of rhabdomyolysis by intramuscular (i.m.) glycerol injection in WT mice and Mif−/− mice (n = 10 both groups). (E and F) Histological analysis of tubular damage by tubular injury score and immunohistochemical staining of apoptotic cells, cleaved caspase-3–positive cells, and necroptotic, pMLKL-positive tubules in WT mice and Mif−/− mice. Scale bars, 100 μm. Data are means ± SD; *P < 0.05, **P < 0.01 analyzed by Student’s t test.

Mif deficiency increases immune cell infiltration

In line with the more severe tubular injury, Mif−/− kidneys showed significantly aggravated inflammatory cell infiltration of Erhr3+ macrophages, F4/80+ monocytes and dendritic cells, and Ly6G+ granulocytes at 24 hours after I/R injury and after rhabdomyolysis (Fig. 5, A and C). At 6 hours after I/R injury, the immune cell infiltration of Ly6G+ granulocytes was significantly increased in Mif−/− mice (cortex, P = 0.001; outer medulla, P = 0.03), whereas the number of F4/80+ monocytes was only significantly increased in the cortex (P < 0.001) but not in the outer medulla (P = 0.13; Fig. 5B). In accordance with the observation of increased immune cell infiltration, the relative mRNA expression of Cxcl-1, a chemotactic cytokine for granulocytes and monocytes, was significantly higher in Mif−/− mice compared to WT mice 6 hours after hypoxia (P = 0.04; Fig. 5D).

Fig. 5 Renal inflammatory cell infiltration after I/R injury and rhabdomyolysis.

Quantification of inflammatory cells in renal tissue after injury. The immune cell infiltration of granulocytes (A), monocytes and dendritic cells (B), and macrophages (C) in Mif−/− mice was analyzed after 35 min of ischemia and 6 hours of reperfusion (n = 10 both groups), 35 min of ischemia and 24 hours of reperfusion (n = 5 both groups), and 24 hours after the induction of rhabdomyolysis (n = 10 both groups). (D) Relative mRNA expression of Cxc1 in kidney tissue of Mif-deficient and WT mice after 6 hours of I/R injury. Cxcl1, chemokine (C-X-C motif) ligand 1; ErHr3, Er-Hematopoiesis related-3; F4/80, monocytes/macrophages marker; Ly6G, lymphocyte antigen 6 complex locus G6D. Scale bars, 100 μm. Data are means ± SD; *P < 0.05, **P < 0.01 analyzed by Student’s t test.

MIF reduces oxidative stress and lipid peroxidation in vitro and in vivo

Oxidative stress was induced by incubation with hydrogen peroxide (H2O2) or hypoxia in primary murine tubular epithelial cells (pmTECs), which were left untreated or were pretreated with recombinant MIF (rMIF). MIF treatment increased intracellular GSH as shown by a higher intracellular GSH/GSSG (reduced glutathione/glutathione disulfide) ratio (H2O2, P = 0.02; hypoxia, P = 0.04; Fig. 6A). Similarly, thiobarbituric acid reactive substances (TBARS), which is a widely used marker of lipid peroxidation (32), was significantly reduced in pmTECs incubated with rMIF and exposed to hypoxia, compared to untreated cells (P = 0.005; Fig. 6B). Consistent with these in vitro findings, we found significantly lower GSH and higher TBARS in tissue lysates of Mif−/− mice after I/R and rhabdomyolysis compared to WT mice (GSH: 6 hours I/R, P = 0.03; 24 hours I/R, P = 0.04; rhabdomyolysis, P = 0.005; TBARS: 6 hours I/R, P = 0.04; 24 hours I/R, P = 0.03; rhabdomyolysis, P = 0.045; Fig. 6, C and D).

Fig. 6 Effects of rMIF on oxidative stress in vitro and oxidative stress in kidney tissue of WT and Mif−/− mice in vivo.

(A and B) GSH in cell lysates of pmTECs after normoxia, H2O2, or hypoxia treatment (A) or in kidney tissue lysates (B) measured with the GSH/GSSG assay kit. (C and D) TBARS in cell lysates after hypoxia for 24 hours (C) or in kidney tissue lysates (D). TBARS was normalized to protein and to control [phosphate-buffered saline (PBS) or WT in normoxia conditions]. In vitro biological n = 3; 6 hours I/R n = 10; 24 hours I/R n = 5, rhabdomyolysis n = 10. Data are means ± SD; *P < 0.05, **P < 0.01 analyzed by Student’s t test.

MIF reduces tubular cell death in vitro

In vitro, using both full or starvation medium, treatment of WT pmTECs with rMIF (100 ng/ml) during hypoxia for 24 hours (<1% O2) significantly reduced cytotoxicity as shown morphologically (Fig. 7A) and by the release of lactate dehydrogenase (LDH) (full medium, P = 0.003; starving medium, P ≤ 0.001; Fig. 7B). Similarly, pmTECs challenged with hypoxia for 18 hours followed by 6 hours of normoxia (~20% O2) and treated with rMIF displayed significantly reduced LDH release when compared to PBS treatment (full medium, P = 0.001; starving medium, P = 0.02; fig. S5). In addition, incubation with rMIF (100 and 500 ng/ml) reduced the phosphorylation of MLKL after hypoxia (fig. S6B). In vitro treatment of Mif−/− pmTECs with rMIF in hypoxia did not show significant cytoprotective effects (P = 0.09; Fig. 7C).

Fig. 7 Effects of rMIF administration on cytotoxicity and oxidative stress in vitro and effects of rMIF treatment on AKI in vivo.

(A to C) Morphology (A) and LDH release of WT (B) and Mif-deficient (C) pmTECs challenged with hypoxia (<1% O2) for >24 hours in full or starvation medium. Comparison of pretreatment with PBS and rMIF (100 ng/ml). LDH was normalized to 100% cell death/toxicity. (D and H) Schematic depicting the rMIF administration and acute kidney damage by cross-clamping of the renal artery for 35 min, contralateral nephrectomy, and reperfusion for 24 hours in WT mice and Mif−/− mice (n = 5 each group). Mice were injected with 16 μg of rMIF twice (30 min before cross-clamping and 6 hours after cross-clamping). (E and I) Serum creatinine levels (normalized to BW) 24 hours after I/R injury. (F, G, J, and K) Tubular necrosis was evaluated applying the tubular injury score. Immunohistochemical staining of apoptotic cells, cleaved caspase-3–positive cells, and pMLKL-positive tubules. i.p., intraperitoneally. Scale bars, 100 μm. Data are means ± SD; *P < 0.05, **P < 0.01 analyzed by Student’s t test.

rMIF administration in vivo reduces AKI

In WT mice, treatment with rMIF 30 min before and 6 hours after I/R injury significantly mitigated AKI after 24 hours of reperfusion. Compared to vehicle-treated mice, rMIF treatment significantly reduced serum creatinine (P = 0.03; Fig. 7E) and the extent of tubular injury, apoptosis, and necroptosis (Fig. 7, F and G). However, in Mif−/− mice, treatment with rMIF did not significantly ameliorate AKI (creatinine, P = 0.73; Fig. 7, I to K).

sCD74 receptor ectodomain augments MIF to reduce cytotoxicity in vitro and reduces tubular injury in the cortex after I/R injury

To investigate whether the sCD74 receptor ectodomain modulates MIF’s beneficial effects, we treated pmTECs with sCD74 in the setting of hypoxia (and H2O2) and analyzed cell death and oxidative stress. In vitro, cytotoxicity was ameliorated by pretreatment with rMIF plus sCD74 (P = 0.002) or with sCD74 alone (P = 0.03; Fig. 8A). Intracellular GSH in the setting of oxidative stress was more potently restored by co-incubation with rMIF and sCD74 compared to rMIF alone (Fig. 8B). In the setting of I/R injury in vivo, sCD74 administration 30 min before and 6 hours after ischemia did not significantly reduce serum creatinine (Fig. 8, C and D). However, mice treated with sCD74 had significantly reduced tubular injury (P = 0.03) and a significantly reduced number of pMLKL-positive tubules in the renal cortex (P = 0.045; Fig. 8E).

Fig. 8 Effects of recombinant sCD74 on cytotoxicity and oxidative stress in vitro and on tubular cell injury in vivo.

(A) LDH release of WT pmTECs challenged with hypoxia for >24 hours in starvation medium. Comparison of pretreatment with PBS, sCD74 alone (40 nM), or sCD74 (40 nM) plus rMIF (100 ng/ml). LDH was normalized to 100% cell death/toxicity. (B) GSH in cell lysates of pmTECs after normoxia, H2O2, or hypoxia treatment measured with the GSH/GSSG assay kit. (C) Schematic depicting recombinant sCD74 administration and acute kidney damage induced by cross-clamping of the renal artery for 35 min, contralateral nephretomy, and reperfusion for 24 hours in WT mice and Mif−/− mice (n = 5 each group). Mice were injected with 20 μg of sCD74 twice (30 min before cross-clamping and 6 hours after cross-clamping). (D) Serum creatinine normalized to BW 24 hours after I/R injury. (E and F) Tubular necrosis was evaluated applying the tubular injury score. Immunohistochemical staining of apoptotic cells, cleaved caspase-3–positive cells, and pMLKL-positive tubules. Data are means ± SD; *P < 0.05, **P < 0.01 analyzed by Student’s t test. ns, not significant.

DISCUSSION

Postoperative AKI in patients after cardiac surgery is associated with a complicated postoperative course and an excess in mortality (4). Despite substantial improvements in surgical techniques, perfusion, and anesthesiological management, the incidence and morbidity of AKI of patients undergoing cardiac surgery remain high (4). This necessitates a better understanding of the pathophysiology of AKI, which could lead to better risk stratification or new therapeutic approaches.

Our findings provide preliminary evidence for a potentially beneficial role of increased circulating MIF in the prevention of AKI in patients undergoing cardiac surgery. Serum MIF increased rapidly in the perioperative setting, and in a small cohort of patients with kidney tumor enucleation, we identified renal ischemia as another stimulus for MIF release. In cardiac surgery patients, serum MIF was inversely correlated with the tubular injury marker NGAL. The increase in postoperative MIF was not only associated with reduced risk of AKI in general but was also associated with a markedly decreased occurrence of more severe grades of AKI (AKIN II and AKIN III).

These clinical studies do not establish a causative relationship between elevated serum MIF and postoperative AKI on a mechanistic level. Therefore, we extended the clinical findings by performing experimental studies in mouse models to clarify the role of MIF in noninfectious AKI. Using a model of renal I/R injury by cross-clamping of the renal artery, we mimicked the clinical setting of hemodynamic instability, inflammatory response, and release of ROS seen in cardiac surgery patients with postoperative development of AKI. In extension, we included a model of rhabdomyolysis, a commonly observed perioperative condition that contributes to the risk of postoperative AKI in cardiac surgery patients (79). In these models, Mif deficiency significantly aggravated the extent of kidney injury, indicating a fundamental role of MIF in kidney protection. Augmentation of serum MIF by perioperative rMIF injection in WT mice mitigated the severity of ischemic AKI. This result suggests that circulating MIF from sources other than the kidney, such as injured endothelium in the heart, might also be involved in protection against renal injury. It further supports the hypothesis that MIF might act as a beneficial, systemic stress-regulating mediator within the inflammatory response and thus may open future therapeutic approaches to reduce the risk of postoperative AKI by perioperative application of rMIF or small–molecular weight MIF agonists (15).

In the heart, the protective role of MIF in I/R injury has been attributed to signaling through the CD74/CD44/AMPK pathway, which was shown to increase glucose uptake into the cells (14). However, previous studies in kidneys have demonstrated that there is no difference in the severity of AKI between WT and Ampk-deficient mice, suggesting that the MIF-dependent CD74/AMPK signaling pathway might not be critically involved in AKI (33).

AKI is characterized by tubular epithelial injury and cell death, the latter appearing morphologically as necrosis and therefore termed “tubular necrosis” (34). Although morphologically indistinguishable from necrosis, recently, alternative regulated cell death pathways, such as necroptosis and ferroptosis, were shown to be crucially involved in tubular injury in AKI, potentially being more important than apoptosis (25, 27, 28). Previous studies have reported that MIF may mediate antiapoptotic effects under specific conditions, such as hypoxic neuronal loss and cancer (3538). In line with this, we found aggravated apoptosis in Mif−/− mice after I/R injury and rhabdomyolysis—although this remained an uncommon event. Mif-deficient mice also displayed significantly more pMLKL-positive, necroptotic tubules and a higher amount of lipid peroxidation, an indicator of ferroptosis. In this context, the increased nonphosphorylated MLKL expression detected by immunohistochemistry may indicate an additional independent factor that lowers the threshold for necroptosis. Our study demonstrates that MIF may have an important role in limiting necroptosis and ferroptosis in AKI.

Ferroptosis was demonstrated to occur in a synchronized manner in renal tubules, resulting in the breakdown of entire functional units (26), but the initial trigger that ignites this cascade remains controversial. Our data on whole, pMLKL-positive, necroptotic tubule segments are in line with this concept of synchronized tubular necroptosis or ferroptosis, in which MIF might be involved as a counterregulatory factor. In the setting of Mif deficiency, this is potentially explained by sensitization to ferroptosis and necroptosis by an imbalance in redox systems. In vivo, this imbalance was mirrored by reduced GSH in Mif−/− kidney tissues. Besides, cardiac surgery patients with high serum MIF directly after surgery displayed a higher total antioxidant capacity compared to patients with low serum MIF. Thus, the protective effect of MIF might include the reduction of oxidative stress by its intrinsic TPOR activity (23, 3942). The proposed action of MIF in the setting of cardiac surgery–related AKI is illustrated in fig. S7. The potentially protective action of MIF on these newly identified cell death pathways might also play a role in other organs where MIF has been shown to exhibit protection via the reduction of regulated necrosis and apoptosis, in tissues such as the heart and lung (15, 43).

The soluble form of the MIF receptor CD74, sCD74, is implicated in multiple inflammatory diseases (17, 30, 44) and was shown to form circulating complexes with MIF. In vitro, low amounts of sCD74 inhibited MIF signaling, whereas high amounts of sCD74 further promoted MIF’s signaling activity (17, 44). In a cohort of cardiac surgery patients, preliminary data indicated a reduced incidence of AKI in patients with detectable sCD74-MIF complexes compared to patients without detectable sCD74-MIF complexes in their blood (30). It was shown that sCD74-MIF complex formation increases MIF catalytic TPOR activity, indicating a potentially protective mechanism that increases MIF’s antioxidant activity (30). We found that pretreatment of tubular epithelial cells with sCD74 and MIF before stimulation with H2O2 (mimicking reperfusion injury) reduced oxidative stress compared to PBS or MIF treatment alone. In the setting of I/R injury in mice, the additional administration of sCD74 resulted in a nonsignificant trend toward reduced tubular cell injury and alleviated AKI. However, we are aware that these preliminary findings should be considered cautiously and require further in-depth analysis of the exact underling molecular mechanisms. Studies of dosing and timing of sCD74 injection will be required to confirm the protective role of sCD74 in AKI.

In autoimmune kidney diseases that affect glomeruli (crescentic glomerulonephritis or lupus nephritis), MIF was shown to aggravate inflammation via its chemokine-like functions (45). During ischemic AKI, we found that immune cell infiltration was enhanced in Mif−/− mice compared to WT mice. This finding is in line with the increased extent of cell death, which is a major driver of inflammation due to the release of highly immunogenic damage-associated molecular patterns and the production of proinflammatory cytokines (46) referred to as necroinflammation (28, 47). These data also indicate that the protective effects of MIF against cell death might be more potent and supervene the potentially proinflammatory and chemokine-like effects of MIF in AKI.

There are limitations to our analyses that need to be considered. The results obtained from our observational, clinical study remain correlative and do not necessarily show a causative relationship. Follow-up large-scale clinical studies to validate the clinical significance of circulating MIF as an effective tool for perioperative risk stratification will be required. For therapeutic applications of MIF in humans, clinical validation of rMIF orthologs is necessary. A further limitation of our in vivo and in vitro experimental studies is that renal I/R injury or rhabdomyolysis are models that only partly mimic the pathophysiology of cardiac surgery–associated AKI. Although ischemia, oxidative stress, inflammation, and rhabdomyolysis are all recognized mechanisms, additional and complex individual clinical factors are implicated in AKI in cardiac surgery patients, which cannot be fully addressed with in vivo or in vitro models. In I/R injury, MIF treatment in WT mice protected against AKI and tubular injury, whereas MIF administration in Mif−/− mice did not show significant effects. A potential explanation of the divergent effects is the altered MIF co-receptor expression of CD44 in Mif−/− mice (48, 49). Previous studies demonstrated that the CD44 receptor expression is MIF-dependent and that Mif−/− mice display a suppressed CD44 receptor expression in the kidney (48, 50). Cd44 deficiency prevents MIF signaling via CD74 (49). Therefore, a down-regulated CD44 receptor expression could be an explanation for the lack of effectiveness of the investigated MIF treatment in Mif−/− mice. Because WT mice are more comparable to the investigated patients (without Mif deficiency), we suggest that the obtained results of MIF treatment in WT mice better reflect the clinical setting.

The presented findings highlight a potential renoprotective role of MIF in AKI, which may be of particular clinical significance after cardiac surgery. Mechanistically, we found that MIF potently limits necroptosis and mitigates oxidative stress by restoring intracellular GSH and reducing lipid peroxidation, both of which are considered hallmarks of ferroptosis (51, 52). The identified protective effects of MIF may open future perspectives for perioperative risk stratification and potential therapeutic options to reduce the incidence of AKI in cardiac surgery patients by application of MIF homologs to augment MIF’s cytoprotective actions intraoperatively.

MATERIALS AND METHODS

Study design

This study was designed to investigate the role of MIF in AKI. The clinical study was designed as a prospective, observational study (ClinicalTrials.gov: NCT 02488876), and after screening for eligibility, 60 patients were successfully enrolled after receiving informed consent in the period from September 2015 to March 2016 (fig. S1). The study was approved by the institutional review board [ethics committee, University Hospital, Rheinisch-Westfälische Technische Hochschule (RWTH) Aachen, Germany] and performed in adherence to the Declaration of Helsinki. All experiments involving animals were carried out in compliance with the local review boards and the Yale Animal Care and Use Committee (No. 2013-11583). Adult congenic WT and Mif−/− mice on a C57/Bl6 background were used in this study. Mif−/− mice were generated as previously reported (45). Two noninfectious models were used to mimic postoperative AKI: I/R injury induced by clamping the renal artery and glycerol injection into the muscle to induce rhabdomyolysis. A subset of mice that received I/R injury was subjected to contralateral nephrectomy. Mice (at least n = 5 per group) were treated with rMIF or sCD74 before or after I/R, and blood and tissue were analyzed. pmTECs were used at passage 3 for in vitro analysis under hypoxic and normoxic conditions. Additional details can be found in Supplementary Materials and Methods, including a list of antibodies used for immunohistochemical staining (table S2).

Patients undergoing cardiac surgery

All patients were scheduled for elective, cardiac surgery with the use of aortic cross-clamping, cardioplegic myocardial arrest, and CPB. The patients were consecutively enrolled after written informed consent was obtained. Exclusion criteria were emergency operations, pregnancy, lack of informed consent, and age less than 18 years. Serum samples were drawn 1 day before surgery, immediately postoperatively (0 hours), and 6, 12, and 24 hours after admission to the intensive care unit (ICU). After blood collection, the samples were centrifuged (3000 rpm for 10 min), and the supernatants were transferred to cryotubes for storage at −80°C until final analysis. Urine samples were collected preoperatively, immediately postoperatively, and 24 hours after admission to the ICU and were transferred to cryotubes for storage at −80°C. The development of AKI was classified in accordance with the AKIN criteria within 72 hours after cardiac surgery (fig. S1B) (20).

Patients undergoing kidney tumor enucleation

Because CPB is associated with a transient decrease in renal perfusion, we investigated whether renal ischemia itself induces MIF secretion into the bloodstream. Serum MIF of patients undergoing kidney tumor enucleation was analyzed as a predefined secondary analysis of a recently completed study (ethical committee vote 012/13, registered on ClinicalTrials.gov as NCT: NCT01839084). Of 46 patients, 18 patients were exposed to renal I/R by cross-clamping of the renal artery during the surgical procedure for 6 to 35 min. Patients who received an unexpected total nephrectomy intraoperatively were excluded from subsequent analysis. Anesthesia, perioperative medication, surgical procedure, and treatment in the ICU were performed in accordance with the local standards, as previously described (30). Serum samples were drawn 1 day before surgery (pre-OP), 5 min after tumor enucleation (intra-OP), in the recovery room shortly after the termination of surgery, and on POD1. Subsequently, the samples were transferred to cryotubes for storage at −80°C until final analysis.

Statistical analysis

Experimental and clinical data were statistically analyzed using a commercially available software package, SPSS 19.0 (SPSS Inc.). GraphPad Prism (GraphPad Software Inc.) was used to generate figures. All metric data were tested for normal distribution using the Shapiro-Wilk W test. Given the explorative nature of this study, normally distributed results of single measurements were compared between the groups at different time points after surgery using Student’s t test. Post hoc testing was performed using the Bonferroni test. Non-normally distributed single measurements were compared using the Mann-Whitney U test. Nonparametric data were compared using the Fisher’s exact test. For correlation studies, linear regression analysis was performed using the Pearson correlation coefficient. In all cases, P < 0.05 was considered statistically significant, and two-sided testing was used.

SUPPLEMENTARY MATERIALS

www.sciencetranslationalmedicine.org/cgi/content/full/10/441/eaan4886/DC1

Materials and Methods

Fig. S1. Flowchart.

Fig. S2. MIF and NGAL concentrations in cardiac surgery patients.

Fig. S3. Protective effects of MIF in an experimental model of AKI induced by bilateral ischemia and 24 hours of reperfusion.

Fig. S4. Up-regulated MLKL expression in Mif−/− mice.

Fig. S5. Cytotoxicity of 18 hours of hypoxia followed by 6 hours of normoxia in pmTECs.

Fig. S6. MIF receptors and phosphorylation of MLKL in tubular epithelial cells.

Fig. S7. Scheme of the proposed action of MIF in the setting of cardiac surgery–related AKI.

Table S1. Baseline characteristics of patients undergoing cardiac surgery.

Table S2. List of primary and secondary antibodies for Western blot, immunohistochemistry, and immunofluorescence staining.

References (5357)

REFERENCES AND NOTES

Acknowledgments: We are thankful for the valuable suggestions and technical support received from A. Fahlenkamp as well as the technical assistance of C. Beckers, K. Jülicher, S. Otten, M. C. Timm, and M. Tatarek-Nossol. Funding: This study was supported by the Deutsche Forschungsgemeinschaft (STO 1099/-2 to C.S., BO 3755/3-1 and BO 3755/6-1 to P.B., BE 1977/9-1 and SFB1123/A03 to J.B., FA 1048/2-1 to A. Fahlenkamp, SFB TRR 57 to P.B. and J.B., and SFB TRR 219 to P.B.), by the Interdisciplinary Centre for Clinical Research (IZKF) at the RWTH Aachen (SP5 to CS and K7-3 to P.B.), by German Center for Cardiovascular Diseases (DZHK), partner site Munich Heart Alliance (to J.B.), by the German Ministry of Education and Research (BMBF Consortium STOP-FSGS number 01GM1518A to P.B.), and by the NIH (AR049610 and AR050498 to R.B. and L.L.). In addition, this work has been funded by a grant from the Department of Defense to G. Moeckel (DM160460) and from the German Research Foundation, Else-Kröner Fresenius Stiftung, Fresenius, Astute Medical, and Astellas to A.Z. Author contributions: L.A., C.S., A.G., J.S., S.K., G. Marx, S.R., A.O., L.L., G. Moeckel, A.L., J.V., O.E.B., A.Z., J.B., S.D., R.B., and P.B. substantially contributed to the design and conception of the presented experimental and clinical studies. L.A., A.M., S.D., and J.V. performed the experimental studies. L.A. drafted the manuscript. C.S., S.D., R.B., P.B., and J.B. critically reviewed and edited the draft. C.S., L.A., M.C., A.K., and B.-S.K. conducted the clinical study, acquired the clinical relevant data, and performed the following analysis. L.A., C.S., A.G., J.S., A.M., S.K., M.C., G. Marx, A.K., B.-S.K., S.R., A.O., L.L., G. Moeckel, A.L., O.E.B., A.Z., J.B., S.D., R.B., and P.B. contributed to the presented analysis of experimental and clinical data and to the interpretation of present findings. All authors read and approved the final manuscript. Competing interests: L.L., R.B., and J.B. are co-inventors on issued patents or patent applications describing the therapeutic utility of MIF modulation (USPTO patent numbers 6030615, 6080407, 20030013122, 20110262386, 9643922, and 9540322). The other authors declare that they have no competing financial interests. Data and materials availability: All data necessary for the interpretation of results were included in the manuscript. Requests for information or materials should be addressed to the corresponding authors.
View Abstract

Navigate This Article