All reagents were purchased from Merck (Madrid, Spain), except where otherwise indicated.
In vivo experimental model
Male Wistar rats (240-260 g) were used in accordance with the Principles of the Declaration of Helsinki and the European Guide for the Care and Use of Laboratory Animals (Directive 2010/63/UE) and Spanish national and regional regulations (Law 32/2007/Spain, RD 1201/2005 and RD 53/2013). The Bioethics Committee approved all procedures for Animal Care and Use of the University of Salamanca. Animals were maintained under controlled environmental conditions, fed on standard chow and allowed to drink water ad libitum.
- Animal model 1: Evaluation of subclinical damage after AKI. Rats were subdivided into two groups (n=6 per group and time point): control group (C): rats receiving saline solution (0.9% NaCl, i.p.); cisplatin group (CP): rats receiving cisplatin (5 mg/kg, i.p.), as in our previous studies. [17] Renal function was evaluated in the following days: Basal (B), immediately prior to cisplatin administration; day of maximum kidney damage (D4; i.e. the day of highest pCr); day of recovery (R0; i.e. the day in which pCr returns to basal levels); and one and two weeks after R0 (R1 and R2 respectively) (Figure 1).
- Animal model 2: Evaluation of predisposition to new AKI episodes. This experimental model was performed to assess if rats were predisposed to suffer new AKI episode when pCr had already returned to basal levels after a previous AKI episode. We used our previously published model, in which predisposition to AKI is unmasked by treating rats with a sub-nephrotoxic dosage regime of gentamicin. [18] [19] [20] In this model, only predisposed rats (and not controls) undergo AKI when subject to sub-nephrotoxic gentamicin.
Rats were subdivided into 4 groups (n=6 each): cisplatin + gentamicin R0 group (R0 + G6): rats receiving cisplatin (5 mg/kg, i.p.) + gentamicin (50 mg/kg/day for 6 days, i.p., starting the day of recovery, R0); cisplatin + gentamicin R1 group (R1 + G6): rats receiving cisplatin (5 mg/kg, i.p.) + gentamicin (50 mg/kg/day for 6 days, i.p., starting 1 week after recovery, R1); cisplatin + gentamicin R2 group (R2 + G6): rats receiving cisplatin (5 mg/kg, i.p.) + gentamicin (50 mg/kg/day for 6 days, i.p., starting two weeks after recovery, R2); gentamicin control group (CT G6): rats receiving saline solution (i.p.) + gentamicin (50 mg/kg/day, for 6 days, i.p., starting in R0, R1 and R2) (Figure 1)
Sample collection
At selected time points, urine and plasma samples were collected to evaluate renal function. For urine collection, rats were individually allocated in metabolic cages. 24-hour urine was collected, centrifuged and stored at -80°C. Blood was drawn from the tail vein, centrifuged and plasma was stored at -80°C. Immediately before sacrifice, kidneys were perfused by the aorta with saline solution and dissected. One half was frozen in liquid nitrogen and subsequently kept at -80°C. The other half was fixed in buffered 3.7% p-formaldehyde for histological studies.
Renal function studies
Plasma (pCr) and urinary creatinine (uCr) and proteinuria were determined with commercial kits based respectively on the Jaffe reaction and the Bradford method, following the manufacturer´s instructions (BioAssay System, Hayward, CA USA). Glomerular filtration rate (GFR) was estimated by the creatinine clearance (ClCr), using the formula: ClCr=uCr x UF/pCr, were UF stands for urine flow.
The furosemide stress test (FST) was used to evaluate tubular integrity, as previously described. [21] Briefly, a single dose of furosemide (20 mg/kg, i.p.) was administered to rats, and they were immediately allocated in metabolic cages to collect individual urine during the following hour. Urine volume and K+ excretion (LAQUAtwin B-731 Compact K+meter, Horiba Scientific, Kyoto, Japan) were measured as indicators of tubular performance. Rats with damaged tubuli show lower furosemide-induced urinary volume and K+ excretion, compared to control rats. [21]
Histological studies
Paraformaldehyde-fixed kidney samples were immersed in paraffin, cut into 5 μm-thick slices and stained with haematoxylin and eosin (HE) and with Periodic acid–Schiff (PAS). Whole-kidney images were obtained with photographs taken using the DotSlide virtual microscopy technique (Olympus BX51, Olympus Iberia, Barcelona, Spain). Images were analysed with the Olyvia Software (Olympus Iberia).
Tissue damage was quantified using a severity score, as previously described. [22] Briefly, 10 external medullar fields [i.e. the area damaged by cisplatin [17, 23] were chosen from dot slide images of each rat, and each field was divided into 6 sections. In each section, four parameters were evaluated: necrosis, tubular dilatation or atrophy, hyaline deposits and epithelial disorganization. A score of 0–3 was assigned to each parameter in a blind manner, according to the following criteria: 0, normal histology; 1, alteration in up to one-third of the section; 2, alteration from one-third to two-thirds of the section; 3, alteration in the whole area of the section. Section scores were added to give a field score (maximal score per field=12). The average score of 10 fields was used for each kidney specimen.
Measurement of urinary biomarkers
ELISA kits were used to measure urinary levels of KIM-1 (Cusabio, Wuhan, China), NGAL (Bioporto, Hellerup, Denmark), albumin (Abcam), transferrin (Abcam, Cambridge, UK), TIMP-2 (Abcam) and IGFBP-7 (Cloud Clone, Houston TX, USA), following the manufacturer instructions. The product of [TIMP-2]*[IGFBP7] urinary concentrations is marketed as NephroCheck®, and improves the diagnostic accuracy and outcomes of AKI patients. We have chosen these biomarkers due to previous studies by our research group [19] and others [24] [25] which show their increased urinary levels in AKI patients and in different experimental models of AKI.
Statistical analysis
Statistical analysis was performed using the GraphPad Prism 7 software (San Diego, CA USA). Data normal distribution was evaluated with the Shapiro-Wilk normality test. Data are shown as mean ± standard error of the mean (SEM) and dot blots. One-way ANOVA with Bonferroni´s test (for data with normal distribution) or Dunn´s test (for data with non-normal distribution) were performed to compare experimental groups; Spearman correlations were used to analyse the ability of biomarkers to predict renal damage. P < 0.05 was considered statistically significant.