Urine interleukin-6 is an early biomarker of acute kidney injury in children undergoing cardiac surgery

Abstract

Introduction: Interleukin-6 (IL-6) is a proinflammatory cytokine that increases early in the serum of patients with acute kidney injury (AKI). The aim of this study was to determine whether urine IL-6 is an early biomarker of AKI and determine the source of urine IL-6. Numerous proteins, including cytokines, are filtered by the glomerulus and then endocytosed and metabolized by the proximal tubule. Since proximal tubule injury is a hallmark of AKI, we hypothesized that urine IL-6 would increase in AKI due to impaired proximal tubule metabolism of filtered IL-6.

Methods: Urine was collected in 25 consecutive pediatric patients undergoing cardiac bypass surgery (CPB). AKI was defined as a 50% increase in serum creatinine at 24 hours (RIFLE (Risk, Injury, Failure, Loss, End stage), R). Mouse models of AKI and freshly isolated proximal tubules were also studied.

Results: Urine IL-6 increased at six hours in patients with AKI versus no AKI (X2 = 8.1750; P < 0.0042). Urine IL-6 > 75 pg/mg identified AKI with a sensitivity of 88%. To assess whether increased urine IL-6 occurs in functional versus structural renal failure, mouse models of pre-renal azotemia after furosemide injection (no tubular injury), ischemic AKI (tubular injury) and cisplatin AKI (tubular injury) were studied. Urine IL-6 did not significantly increase in pre-renal azotemia but did increase in ischemic and cisplatin AKI. To determine if circulating IL-6 appears in the urine in AKI, recombinant human (h)IL-6 was injected intravenously and urine collected for one hour; urine hIL-6 increased in ischemic AKI, but not pre-renal azotemia. To determine the effect of AKI on circulating IL-6, serum hIL-6 was determined one hour post-intravenous injection and was increased in ischemic AKI, but not pre-renal azotemia. To directly examine IL-6 metabolism, hIL-6 was added to the media of normal and hypoxic isolated proximal tubules; hIL-6 was reduced in the media of normal versus injured hypoxic proximal tubules.

Conclusions: Urine IL-6 increases early in patients with AKI. Animal studies demonstrate that failure of proximal tubule metabolism of IL-6 results in increased serum and urine IL-6. Impaired IL-6 metabolism leading to increased serum IL-6 may contribute to the deleterious systemic effects and increased mortality associated with AKI.

Figures

Figure 1

Urine IL-6 is increased after cardiopulmonary bypass in pediatric patients. Urine was collected at baseline and two and six hours after cardiopulmonary bypass and IL-6 was determined. Box and whisker plots indicate the 10th, 25th, 50th (median), and 90th percentile values of urinary IL-6. At six hours post-cardiopulmonary bypass, the median urine IL-6 was significantly increased in patients with AKI versus those without AKI. No difference was observed between pre-operative urine IL-6 values in patients with AKI versus no AKI (P = 0.65). * Denotes statistical significance, P < 0.002.

Figure 2

Clinical utility of urine IL-6 to diagnose early acute kidney injury. (A) A urine IL-6 of ≥75 pg/mg predicts acute kidney injury with 88% sensitivity. (B) Receiver operating characteristic (ROC) curve for urine IL-6 at six hours after cardiopulmonary bypass.

Figure 3

Mouse model of pre-renal azotemia and ischemic AKI. Prerenal azotemia after furosemide injection is characterized by increased urine output (A), increased total body weight loss (B), increased hematocrit (C), and normal creatinine (D) compared to vehicle injection, sham operation, and ischemic AKI (urine output was determined at two hours; total body weight loss, hematocrit, and creatinine were determined at six hours). BUN (E) is increased in both pre-renal azotemia and ischemic AKI (BUN was determined at six hours). Two hours post-ischemic AKI (F), patchy necrosis with areas of normal proximal tubules in intact brush borders (arrows) is present; at six hours post ischemic AKI (G), proximal tubule necrosis is wide spread. Renal histology is normal after furosemide injection (H).

Figure 4

Urine, serum, and renal IL-6 in pre-renal azotemia and ischemic AKI. (A) Urine IL-6 increases in mice with ischemic AKI, but not pre-renal azotemia. Spontaneously voided urine was collected at baseline and from zero to two hours, two to four hours, and four to six hours after vehicle-injection (Veh), furosemide injection/pre-renal azotemia (Pre), sham operation (Sham) and ischemic AKI (AKI). Urine IL-6 was increased at four to six hours after ischemic AKI; median and SD (*P < 0.01 vs. Veh, Pre, Sham, n = 5 to 7). (B) Serum IL-6 increases in mice with ischemic AKI prior to the increase in urine IL-6. Serum IL-6 was determined at baseline, and two, four and six hours after vehicle-injection (Veh), furosemide injection/pre-renal azotemia (Pre), sham operation (Sham) and ischemic AKI (AKI) and was significantly increased at two, four and six hours after AKI (P < 0.01 vs. Veh, Pre, Sham at all time points; n = 4 to 11). (C) Kidney IL-6 increases in mice with ischemic AKI prior to the increase in urine IL-6. Kidney IL-6 was determined at baseline, and two, four and six hours after vehicle-injection (Veh), furosemide injection/pre-renal azotemia (Pre), sham operation (Sham) and ischemic AKI (AKI) and was significantly increased at two, four and six hours after AKI (P < 0.01 vs. Veh, Pre, Sham at all time points; n = 3 to 7).

Figure 5

Renal function and urine, serum, and renal IL-6 in cisplatin-induced AKI. (A) Serum creatinine and (B) BUN increase on Day 3 in mice with cisplatin-induced AKI. Serum creatinine and BUN were determined on Days 1, 2 and 3 after vehicle or cisplatin injection and was significantly increased on Day 3 (P < 0.05 for creatinine and P = 0.0001 for BUN vs. Veh; n = 5 to 12). (C) Urine IL-6 increases on Days 2 and 3 in mice with cisplatin-induced AKI. Urine was collected at the time of sacrifice on Days 1, 2 and 3 after vehicle or cisplatin injection. Urine IL-6 did not increase significantly until Days 2 and 3 after cisplatin injection, when proximal tubule injury is present (P < 0.002 vs. Veh, n = 5 to 12). (D) Serum IL-6 increases on Days 2 and 3 in mice with cisplatin-induced AKI. Serum IL-6 was determined on Days 1, 2 and 3 after vehicle or cisplatin injection and was significantly increased on Days 2 and 3 (P < 0.05 on Day 2 and P < 0.0001 on Day 3 vs. Veh; n = 5 to 12). (E) Kidney IL-6 increases on Day 3 in mice with cisplatin-induced AKI. Kidney IL-6 was determined on Days 1, 2 and 3 after vehicle or cisplatin injection and was significantly increased on Day 3 (P < 0.01 on Day 3 vs. Veh; n = 5 to 12).

Figure 6

Fate of intravenously injected recombinant human IL-6 in pre-renal azotemia, ischemic AKI, and bilateral nephrectomy. A total of 200 ng of recombinant human (h) IL-6 was administered by tail vein injection five hours after vehicle-injection (Veh), furosemide injection/pre-renal azotemia (Pre), sham operation (Sham), ischemic AKI (AKI), or after bilateral nephrectomy. Urine was collected for one hour and serum was collected at one hour. (A) Serum human IL-6 is elevated in mice with ischemic AKI and bilateral nephrectomy versus vehicle injection, pre-renal azotemia, and sham operation (*P < 0.01 versus vehicle injection, pre-renal azotemia, and sham operation, n = 4; **P < 0.05 versus ischemic AKI). (B) Urine human IL-6 is increased in mice with ischemic AKI versus vehicle injection, pre-renal azotemia, and sham operation (***P < 0.01 versus vehicle injection, pre-renal azotemia, and sham operation, n = 4 to 5). (Mice with bilateral nephrectomy are anuric; therefore, no urine values are reported for this group).

Figure 7

Addition of recombinant human IL-6 to freshly isolated proximal tubules. A total of 200 ng of recombinant human (h) IL-6 was added to media with and without freshly isolated proximal tubules after 20 minutes of either normoxic or hypoxic conditions. Media human IL-6 concentration was determined after five minutes of incubation. Human IL-6 was significantly reduced in the media containing normoxic proximal tubules versus normoxic and hypoxic media without proximal tubules (*P < 0.02, n = 5 to 6). Media human IL-6 was higher in hypoxic proximal tubules versus normoxic proximal tubules (†P = 0.05, n = 6).