Postoperative cellular stress in the kidney is associated with an early systemic γδ T-cell immune cell response

The clinical part of the study was performed first. In this observational study, we included 20 consecutive patients undergoing multibranched endovascular thoracoabdominal aortic repair. Occlusion and intraluminal manipulation of the renal artery during multibranched endovascular aortic repair lead to a brief episode of kidney hypoperfusion and reperfusion. Two patients were excluded: one because of intraoperative interruption of aneurysm repair owing to technical difficulties and one because of requirement of urgent reoperation < 24 hours after surgery. Eligible patients fulfilled the following four criteria: (1) adults (aged 18 years or older), (2) elective repair of thoracoabdominal aortic aneurysm, (3) presence of a central venous catheter for repeated blood collection, and (4) provision of written informed consent. Exclusion criteria were (1) preexisting liver and kidney failure, defined as a spontaneous international normalized ratio > 2 and encephalopathy, or serum creatinine (SCr) > 2 mg/dl, or preoperative requirement of renal replacement therapy; and (2) anemia with hemoglobin concentration of < 7 g/dl.

Ultrasound imaging was conducted at the bedside in a standardized manner. A conventional ultrasound B-scan was followed by color-coded Doppler sonography. Renal perfusion was quantified by mean values of renal resistive index measured in the renal artery and in the interlobar arteries. All examinations were supervised and analyzed at the bedside by a board-certified sonographist, according to German Society for Ultrasound in Medicine guidelines. Postoperative AKI was defined according to the 2012 Kidney Disease: Improving Global Outcomes criteria [9].

Blood samples for measuring laboratory parameters such as SCr, creatinine kinase, lactate, hemoglobin, interleukin (IL)-6, C-reactive protein, and central venous oxygen saturation were collected by central venous sampling after the induction of anesthesia, before skin incision, and at prespecified points after surgery (less than 4 hours after the end of surgery and 24 hours after the start of surgery). Urinary [TIMP-2]•[IGFBP7] was measured simultaneously after surgery by means of the immunoassay method integrated in the Astute 140 Meter Kit (Nephrocheck test; Astute Medical Inc., San Diego, CA, USA). A cutoff value of [TIMP-2]•[IGFBP7] ≥ 0.3 was used to identify tubular cellular stress and to predict AKI [10]. All values for [TIMP-2]•[IGFBP7] are reported in units of (ng/ml)²/1000. Blood samples for immune monitoring were collected directly into a Cyto-Chex® BCT tube (Streck, Omaha, NE, USA) by central venous sampling after the induction of anesthesia, before skin incision, and 24 hours after the start of surgery. Immediately after collection, the contents of the Cyto-Chex® tube were mixed by gentle inversion, and the tube was immediately transported to the in-house immune monitoring laboratory. Samples were typically processed within 30 minutes after blood collection; however, standard operating procedures allowed samples to be stored at 4 °C for up to 4 hours before processing. To prepare samples for analysis, 100 μl of whole blood was transferred from the Cyto-Chex® tube into a conical tube (Sarstedt, Nürnbrecht, Germany). Unspecific binding was blocked with 10 μl of human FcR blocking reagent (Miltenyi Biotec, Bergisch Gladbach, Germany) in DPBS without Ca²⁺ or Mg²⁺ (PAA Laboratories, Cölbe, Germany) and incubated on ice for 30 minutes. Primary conjugated antibodies were added in optimized amounts to stain the cells at 4 °C for 1 hour. To remove erythrocytes, samples were incubated in 2 ml of 1 × BD lysing solution (BD Biosciences, Heidelberg, Germany) at 4 °C for 10 minutes before being washed twice with 2 ml of DPBS. Samples were analyzed using a FACSCanto II cytometer (BD Biosciences) within 30 minutes of antibody staining. As an internal stability control for cytometer performance, a BD™ Cytometer Setup and Tracking (CST) bead assay (BD Biosciences) was conducted before each analysis. As many CD45⁺ events as possible were recorded. Data were analyzed by a blinded operator using FlowJo® software (FlowJo, Ashland, OR, USA).

The experimental study was performed after analysis of the clinical data. Six- to 8-week-old C57BL/6 mice (Charles River Laboratories, Sulzfeld, Germany) underwent kidney IRI as described elsewhere [11]. In brief, the renal vessels were exposed bilaterally through a midline incision and occluded with vascular clamps for 30 minutes. During the procedure, animals were kept at a constant temperature (37 °C) and hydrated with warm saline. After 24 hours, the mice were killed, and their organs were harvested for further analysis. Sham-operated animals (controls) were treated identically except for renal artery occlusion. All animal experiments were conducted upon approval by the Government of Upper Palatinate.

In mice, spleens and kidneys were harvested and processed to single-cell suspensions. Renal leukocytes were isolated by mechanically disrupting renal tissue in 5 ml of Paris buffer (20 mM Tris-HCl,125 mM NaCl, 10 mM KCl, 10 mM glucose, pH 7.4). To remove debris, samples were passed through a 100-μm cell strainer (Falcon). The resulting cell suspension was centrifuged at 300 × g for 10 minutes to pellet the cells. The pellet was then suspended in 36% Percoll (Amersham Pharmacia, Stockholm, Sweden), gently overlaid onto 72% Percoll and centrifuged at 1000 × g for 30 minutes at room temperature. Cells were isolated from the Percoll interface and washed twice in magnetic-activated cell sorting (MACS) buffer at 300 × g at 4 °C for 10 minutes. Samples were resuspended in 250 μl of MACS buffer. The number of leukocytes was assessed using trypan blue exclusion on a hemocytometer. Single-cell suspensions were then stained with antibodies against CD3, CD4, CD8, and γδ T-cell receptor (eBiosciences, Frankfurt, Germany), and antigen expression was determined with a FACSCanto flow cytometer (BD Biosciences). Data were analyzed with FlowJo software.

Human HMBG1 serum levels were analyzed with a commercially available HMGB1 enzyme-linked immunosorbent assay kit (IBL International, Hamburg, Germany) according to the manufacturer’s instructions.

Murine kidney pieces were collected in RNAlater buffer (Qiagen, Hilden, Germany) for further analysis. Complementary DNA (cDNA) was isolated from murine kidney samples using the μMACS ONEstep cDNA Kit (Miltenyi Biotec). Relative gene expression changes between GAPDH and HMGB1 were determined by means of the comparative cycle threshold method. The following QuantiTect Primer Sets (Qiagen) were used: QT01658692 for GAPDH and QT02520210 for HMGB1.

Baseline characteristics are shown in Table 1. A schematic overview of our diagnostic trial is given in Additional file 1: Figure S1a. Nine (50%) of 18 patients undergoing endovascular aortic repair showed elevated urinary cell stress biomarkers [TIMP-2]•[IGFBP-7] within 4 hours after surgery. On the day after surgery, patients with elevated biomarkers showed a 1.53-fold (SD 0.33) mean increase in creatinine levels compared with baseline values, whereas patients with physiological biomarker levels only showed a 1.13-fold (SD 0.11) mean elevation (Additional file 1: Figure S1b) (p = 0.021) and thus had a significantly higher rate of AKI (OR, 12.25; 95% CI, 1.33–113.06; p = 0.027).

Postoperative renal perfusion was similar in both biomarker groups, as shown by mean arterial pressure over the first 24 hours and standard Doppler sonography. Patients who developed early tubular cell stress required longer duration of surgery and more blood transfusions and had a more positive intraoperative fluid balance, higher vasopressor dose, and higher SOFA score at admission than patients with urinary [TIMP-2][IGFBP-7] < 0.3 (Table 1, Additional file 1: Figure S1c).

All patients underwent clinical-grade standardized immune monitoring (Fig. 1a). When comparing γδ T-cell frequencies before and after surgery, we found that declines in γδ T-cell frequencies were more pronounced in the patient subset with elevated renal biomarkers (Fig. 1b). Strikingly, the decrease in γδ T cells from the circulation correlated with the value of urinary biomarkers (Fig. 1c) (r = 0.53. p = 0.022). There was no correlation of postoperative changes with other T-cell subsets in the first 24 hours. Similar results were found by evaluation of the absolute cell count (Additional file 2: Figure S2) (r = 0.60, p = 0.01).

C57BL/6 mice showed a markedly increased level of serum creatinine and urea within the first 24 hours after standardized warm ischemia (Fig. 2a). In parallel, murine kidneys showed histopathological features of early kidney injury (Fig. 2b). Lymphocytes infiltrated the renal interstitium, and infiltrates were analyzed by flow cytometry. γδ T cells were increased 6.65-fold compared with sham-operated controls (p = 0.002) in ischemic kidneys within the first 24 hours after injury, whereas levels of other T-cell subsets tested were similar in both groups (Fig. 2c). Similarly to the human setting, peripheral (blood) murine γδ T cells were significantly reduced in IRI mice as compared with controls (Fig. 3a left), showing that γδ T cells home from blood to the kidneys after cellular injury and induction of IRI.

In search of a mechanistic driver of γδ T-cell homing, we investigated the role of HMGB1. HMGB1 has recently been described as a potent proinflammatory cytokine released from ischemia-reperfusion-injured cells. HMGB1 messenger RNA (mRNA) expression was significantly increased in murine kidneys within the first 24 hours after IRI (Fig. 3a, right). In humans, HMGB1 protein was significantly elevated in biomarker-positive patients 1 day after surgery compared with biomarker-negative patients (median [IQR] 0.87 [0.68–1.82] versus 2.63 [1.59–12.20] ng/ml, p = 0.019) (Fig. 3b).