Introduction
IL-6 is a proinflammatory cytokine involved in the acute phase response to a wide variety of physiologic insults. For example, serum IL-6 is elevated in patients with sepsis, acute lung injury (ALI), congestive heart failure, acute myocardial infarction, and acute kidney injury (AKI) and predicts increased morbidity and mortality in these conditions [
1‐
8]. We have recently demonstrated that serum IL-6 is increased at two hours in patients with AKI and predicts prolonged mechanical ventilation in children undergoing cardiac surgery [
9]. A pathogenic role of IL-6 in AKI, ALI, and multiple-organ dysfunction syndrome has been suggested.
Increased serum IL-6 in patients with critical illness may be due to multiple factors; for example, increased IL-6 production by stimulated macrophages in injured organs is well described [
10]. In addition to increased production, it is possible that certain co-existing conditions, such as AKI, may reduce serum cytokine clearance. In this regard, data is accumulating that the kidney plays a key role in cytokine clearance and metabolism. Because most cytokines are between 10 to 30 kd, filtration of circulating serum cytokines by the glomerulus is expected. Although filtration and excretion of the intact cytokine occurs, this is not the major mechanism of renal cytokine elimination. Instead, cytokines, like other proteins, are filtered at the glomerulus and then endocytosed and metabolized by the proximal tubule [
11‐
16]. Since proximal tubule injury and dysfunction is the hallmark of AKI, reduced renal IL-6 metabolism might contribute to increased serum IL-6 in patients with AKI. Paradoxically, impaired proximal tubule metabolism of IL-6 would also result in increased urine IL-6; in this case, filtered IL-6 would not be metabolized by the proximal tubule and would therefore appear intact in the urine.
In the present study, therefore, we hypothesized that urine IL-6 would increase in AKI associated with proximal tubule injury. To test this hypothesis, we measured urine IL-6 and other cytokines in pediatric patients undergoing cardiac surgery who did and did not develop AKI. To examine the role of the kidney and proximal tubule in cytokine handling, mouse models of ischemic AKI (renal failure with proximal tubular injury), cisplatin-induced AKI (renal failure with proximal tubular injury), and pre-renal azotemia (renal failure without proximal tubular injury) were studied. To directly test the role of proximal tubules in IL-6 metabolism, we utilized freshly isolated proximal tubules exposed to normoxic and hypoxic conditions.
Materials and methods
Patients
After obtaining approval from both the Colorado Institutional Review Board (COMIRB) and Clinical and Translational Research Center (CTRC) all children undergoing scheduled first time cardiopulmonary bypass (CPB) for repair of congenital heart disease at The Children's Hospital in Denver, Colorado were screened for inclusion in the study. Patients were excluded if they had known underlying chronic kidney disease (preoperative estimated Schwartz clearance < 80 ml/min/1.73 m2), exposure to nephrotoxins within one week of surgery (intravenous contrast, aminoglycosides), proteinuria (dipstick 1+ or greater), urinary tract infection, diabetes, baseline serum creatinine that was unavailable, or inability to obtain consent. Twenty-five patients (aged 8 days to 14 years; median age 4.4 months) were enrolled between February 2007 and March 2008. Written informed consent was obtained for all patients enrolled in the study prior to any sample collection. Two patients were subsequently excluded due to gross hemolysis of the urine samples. Of the 23 patients included in the analysis, 10 met pre-specified criteria for AKI and 13 did not.
The primary outcome assessed was the development of AKI post-CPB. AKI was defined, according to RIFLE criteria R, as a 50% or greater increase in pre-operative serum creatinine at 24 hours. Other clinical variables collected and analyzed included duration of cardiopulmonary bypass (minutes), age, gender, and length of stay (ICU and hospital). There was no management component of this study; patients were managed according to standard of care.
Patient urine collections
Fresh urine was collected from a Foley catheter at three time points: pre-operatively and at two and six hours after coming off CPB. Samples were centrifuged for five minutes at 2,000 RPM and the supernatant was aliquoted and immediately placed in -80°C freezer until analysis. All samples were analyzed within 15 months of initial collection.
Urine creatinine and cytokine measurement in patients
Urine creatinine was determined using a quantitative colorimetric creatinine determination assay (QuantiChrom™ creatinine assay kit-DICT-500) (BioAssay Systems, Hayward, CA, USA) as described below for mice. Urine IL-6, IL-8, IL-10, IL-1β, and TNF-α were measured in duplicate using human ELISA kits according to assay instructions (R&D Systems, Minneapolis, MN, USA). The detection limits are as follows: 1) IL-6 is 0.7 pg/mL, 2) TNF-α is 1.6 pg/mL, 3) IL-1β is 1 pg/mL, 4) IL-8 is 3.5 pg/mL (average of 53 assays), and 5) IL-10 is 3.9 pg/mL.
Statistical analysis of patient data
Data was analyzed using SAS version 8.1 (SAS Institute, Inc, Cary, NC, USA) and SPSS 11.5. Given the small sample size and non-normal distributions, a Wilcoxon Rank Sum test was used to test for statistically significant differences in continuous subject demographics as well as urine IL-6 at baseline, IL-6 at two hours, and IL-6 at six hours between subjects with and without AKI. A chi-square test was used to compare categorical subject demographic variables. In addition, a receiver operating characteristic (ROC) curve was used to assess the relationship between urine IL-6 at six hours and AKI.
Animals
Eight- to ten-week-old male, wild-type, C57BL/6 mice weighing 20 to 25 g were used (Jackson Labs, Bar Harbor, ME, USA). Mice were maintained on a standard diet and water was made freely available. All experiments were conducted with adherence to the NIH Guide for the Care and Use of Laboratory Animals. The animal protocol was approved by the Animal Care and Use Committee of the University of Colorado (Protocol numbers 81102007(06)1D and 81110(02)1E).
Ischemic AKI and bilateral nephrectomy in mice
Three surgical procedures were performed: (1) sham operation, (2) ischemic AKI, and (3) bilateral nephrectomy, as previously described by our laboratory [
17,
18]. Briefly, adult male C57B/6 mice were anesthetized with IP Avertin (2,2,2-tribromoethanol: Aldrich, Milwaukee, WI, USA), a midline incision was made, the bladder was emptied of urine by gentle pressure, and the renal pedicles identified. For ischemic AKI, pedicles were clamped for 22 minutes. After clamp removal, kidneys were observed for restoration of blood flow by the return to their original color. Sham surgery consisted of the same procedure except that clamps were not applied. For bilateral nephrectomy, both renal pedicles were tied off with suture, and the kidneys were removed. The abdomen was closed in one layer.
Cisplatin model of AKI in mice
Six hours before cisplatin administration, food and water were withheld. Cisplatin (Aldrich) was freshly prepared the day of administration in normal saline at a concentration of 1 mg/ml. Mice were given either 30 mg/kg body weight of cisplatin or an equivalent volume of vehicle (saline), after which the mice again had free access to food and water. The cisplatin model of AKI is well established in our laboratory [
19,
20].
Pre-renal azotemia (that is, volume depletion) model in mice
Mice received either 0.5 mg of furosemide (in 100 μL saline) or vehicle (100 μL saline) intraperitoneally and food and water were withheld for six hours. At three hours, vehicle-treated mice received an IP dose of saline to maintain pre-injection body weight (600 to 1,000 μL) while furosemide-treated mice received sham injection (50 μL saline).
Collection and preparation of mouse urine and serum samples
Immediately post-procedure, mice were placed in urine collection containers and spontaneously voided urine was collected. Blood was obtained at sacrifice via cardiac puncture. To assure uniformity, all samples were processed identically. Blood was allowed to clot at room temperature for two hours then centrifuged at 3,000 g for 10 minutes. Serum was collected and centrifuged a second time at 3,000 g for one minute to ensure elimination of red blood cells. Samples with notable hemolysis were discarded.
Hematocrit
Blood was collected in a capillary tube and spun in a micro capillary centrifuge (International Equipment Company, Needham Heights, MA, USA) for three minutes. Hematocrit was determined using a micro-hematocrit capillary tube reader (Monoject Scientific, St Louis, MO, USA).
Renal histology
Kidney halves were fixed in 3.78% formaldehyde which was paraffin embedded, sectioned at 4 μm and stained with periodic acid-Schiff (PAS) by standard methods.
Creatinine and blood urea nitrogen (BUN) measurement in mice
Serum and urine creatinine were determined using a quantitative colorimetric creatinine determination assay (QuantiChrom™ creatinine assay kit-DICT-500) (BioAssay Systems). BUN was measured using a QuantiChrom assay kit (QuantiChrom™ urea assay kit BIUR-500 (BioAssay Systems)).
Urine, serum, and renal IL-6 measurement
Urine, serum, and renal IL-6 were measured by ELISA using a species specific (that is, mouse or human) kit according to assay instructions (R&D Systems). Renal IL-6 was determined on whole kidney homogenates and corrected for total protein content using a Bio-Rad DC protein assay kit (Hercules, CA, USA). The detection limit of the human IL-6 assay is 0.7 pg/mL; the detection limit of the mouse IL-6 assay is 1.6 pg/mL.
Injection of recombinant human IL-6
A total of 200 ng of recombinant human IL-6 (hIL-6) (R&D Systems) or vehicle (PBS with 1% albumin) was injected via tail vein five hours after 100 μL saline injection (vehicle), 0.5 mg furosemide injection (pre-renal azotemia), sham operation, or ischemic AKI. Urine was collected for one hour after IL-6 injection. At one hour post-injection, the mice were sacrificed and blood was obtained.
Addition of recombinant human IL-6 to freshly isolated mouse proximal tubules
Proximal tubules were isolated from the kidney cortex using the collagenase digestion and percoll centrifugation as we have previously described in detail [
20]. At 20 minutes of either normoxia or hypoxia, 16.6 ng of recombinant human IL-6 (hIL-6) was added to media with and without proximal tubules. At 25 minutes, samples were centrifuged and washed at 800 g × 2, and the media and pellet were snap frozen for future analysis.
Statistical analysis of murine data
Data were analyzed by one-way analysis of variance at each time point; if significant F-statistic from analysis of variance existed, this test was followed by Dunnett post hoc multiple comparison procedure with sham operation as the control group. For all other comparisons Student's t-test was used. A P-value of ≤ 0.05 was considered statistically significant.
Discussion
Herein, we demonstrate that urine IL-6 increased by six hours in pediatric patients with AKI after cardiopulmonary bypass (CPB) and is thus a potential early biomarker of AKI. The development of biomarkers that can identify AKI early is a translational research priority [
23] as failure of therapeutic trials in AKI is widely believed to be due the dependence on serum creatinine, a late marker of kidney injury [
24], to diagnose AKI. Multiple serum and urine biomarkers are currently being tested for their ability to diagnose AKI. It is unlikely, however, that one biomarker will be able to accurately diagnose AKI;panels of biomarkers will be required [
25]. Thus, the identification of new biomarkers that can enhance the diagnostic potential of currently studied biomarkers is still needed.
To examine the diagnostic utility of increased urine IL-6 in patients with AKI, we studied animal models of ischemic AKI, cisplatin-induced AKI, and pre-renal azotemia. We found that urine, serum, and renal IL-6 were all increased in mice with ischemic AKI and cisplatin-induced AKI, but not pre-renal azotemia. Ischemic AKI and cisplatin-induced AKI are both associated with proximal tubule injury and acute tubular necrosis (ATN), while proximal tubule injury and necrosis are absent in our model of pre-renal azotemia. ATN from ischemia and nephrotoxins are the most common causes of AKI in hospitalized patients and distinguishing pre-renal azotemia from ATN remains a challenging clinical dilemma [
26], thus, increased urine IL-6 may have clinical utility for this purpose. It is important to note, however, that urine IL-6 was not zero with pre-renal azotemia and certain controls; therefore, small amounts of IL-6 may appear in the urine in the absence of structural renal injury. Thus, as with most biomarkers, it will be important to establish what level of urine IL-6 is clinically significant in regard to the identification of ATN or AKI. The increase in renal and serum IL-6 confirm previous studies [
10,
17,
18] and highlight the early pro-inflammatory nature of AKI. The timing of increased serum IL-6 relative to increased urine IL-6 is consistent with our hypothesis that serum/circulating IL-6 appears in the urine in AKI with proximal tubular injury.
To test our hypothesis that serum/circulating IL-6 is filtered and remains intact in the urine in AKI with proximal tubule injury, hIL-6 was given intravenously to mice with pre-renal azotemia (renal failure without proximal tubule injury) or ischemic AKI (renal failure with proximal tubule injury) and urine was collected for one hour. The use of hIL-6 in this experiment is advantageous because it is homologous to murine IL-6 and expected to be handled in a similar manner, but it does not cross react with murine IL-6 on the ELISA test used to measure it; thus, potential confounding effects of endogenous murine IL-6 production are avoided. If our hypothesis that circulating IL-6 is filtered by the glomerulus and then resorbed and metabolized by the proximal tubule were correct, then urine hIL-6 would be low in mice with functioning kidneys and high in mice with AKI. Indeed, urine hIL-6 was dramatically increased in ischemic AKI and was reduced in pre-renal azotemia and controls with normal renal function. Finally, to directly examine IL-6 metabolism by proximal tubules, hIL-6 was added to the media of normal and hypoxic isolated proximal tubules and hIL-6 was reduced in the media of normal versus injured hypoxic proximal tubules. Together these data suggest that renal filtration coupled with impaired proximal tubule metabolism of IL-6 contributes to the increase in urine IL-6 observed in AKI.
Although previous studies have not examined the effect of AKI on renal IL-6 handling, our data demonstrating a role of the kidney in the filtration and metabolism of IL-6 are consistent with the known role of the proximal tubule in protein metabolism. Other proteins that are filtered at the glomerulus and then endocytosed and metabolized by the proximal tubule include light chains [
27], hormones (for example, insulin, parathyroid hormone), small peptides, and β
2-microglobulin [
28]. The role of the kidney in the clearance and metabolism of IL-6 [
29] and other cytokines such as IL-1, GCSF, and IL-10 has also been described [
11‐
16]. In fact, data suggest that renal proximal tubule metabolism is responsible for at least 10% of cytokine elimination [
16,
29].
How does impaired renal elimination/metabolism of IL-6 affect serum levels? We found that serum levels of hIL-6 were markedly elevated in mice with ischemic AKI or bilateral nephrectomy one hour after IV injection, but were reduced in mice with normal renal function or pre-renal azotemia. Thus, AKI was associated with sustained levels of serum IL-6. The clinical relevance of these findings is notable since proinflammatory cytokines such as IL-6 are known to mediate organ dysfunction. Since numerous insults causing IL-6 production may occur in patients with AKI (for example, hemorrhage, infection), impaired metabolism with systemic accumulation of IL-6 may contribute to the adverse clinical outcomes associated with AKI, particularly in the setting of the systemic inflammatory response syndrome and multiple organ dysfunction syndrome.
Our pilot study in patients, although promising, has a few important limitations. First, although elevated serum IL-6 after CPB is well described, serum IL-6 was not measured in our patients. Because a key source of urine IL-6 is circulating serum IL-6, the potential role of urine IL-6 as a biomarker of AKI may depend on the availability of tandem serum and urine IL-6 values. Second, results were obtained in a small number of homogenous pediatric patients from a single center. Third, the cause of AKI was not specifically assessed, although it is presumed to be due to ATN. Finally, although other urine cytokines (for example, IL-8, IL-10, IL-1β, TNF-α) were not predictive of AKI in this population, it is possible that these cytokines may have diagnostic utility in other forms of AKI. For example, urine IL-6 [
30], IL-8 [
30], and Gro-α [
31] were increased post-transplant in patients with delayed graft function. Ours is the first study to examine the utility of these urine cytokines to diagnose AKI in patients with native kidneys; thus, further studies will need to be performed to determine whether urine IL-6 or other urine cytokines are useful to diagnose AKI in conditions other than post-cardiopulmonary bypass.
Conclusions
The identification of early biomarkers of AKI is critical for the development of successful treatments to improve kidney function and reduce systemic complications. We demonstrate that increased urine IL-6, using a cut point of 75 pg/mg, can diagnose AKI post-cardiopulmonary bypass with 88% sensitivity within six hours of CPB. In animal models, we demonstrate that 1) renal IL-6 production and serum IL-6 increase early in AKI, 2) urine IL-6 increases in AKI associated with ATN, 3) renal elimination of IL-6 is impaired in AKI. Thus, AKI may be a unique clinical scenario where increased production and impaired elimination of cytokines occurs; the resultant increase in systemic cytokine burden may contribute to the increased morbidity and mortality of patients with AKI. Since IL-6 is known to be a mediator of both AKI and ALI, IL-6 may be both a diagnostic marker of AKI as well as a therapeutic target.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
PD designed the study and prepared the manuscript. CA, CK, AAH and NA carried out biomarker measurements. JK designed the study and collected human samples. MC designed the study. AK performed statistical analyses. CLE assisted with drafting of the manuscript. SF performed animal surgeries, prepared and performed the proximal tubule experiments, developed the pre-renal azotemia model, conceived of the study, and drafted the manuscript. All authors read and approved the final manuscript.