We conducted a retrospective clinical study with plasma samples from 28 patients with AKI obtained from the Biobank of the Ramón y Cajal University Hospital-IRYCIS (National Biobank Registry B.0000678, Madrid), part of the ISCIII Biobanks and Biomodels platform (PT20/00045). We studied the course of AKI at the time of admission to the hospital, and after 24 and 72 h. All subjects signed an informed consent form. The inclusion criteria in the study were: 1) being over 18 years of age, and 2) having been diagnosed with AKI at the time of hospital admission. The exclusion criteria were: 1) being under 18 years of age, 2) pregnancy, and 3) not having informed consent. The type of AKI was diagnosed and classified according to the AKIN (Acute Kidney Injury Network) guidelines.16,17 Demographic and clinical data, as well as the presence of comorbidities, were collected in the first 24 h of admission to the hospital to calculate the APACHE (Acute Physiology and Chronic Health Evaluation) II score.18 The control samples corresponded to a total of 26 healthy individuals with no pathologies, also from the Ramón y Cajal University Hospital-IRYCIS Biobank. This study was approved by the 12 de Octubre University Hospital Ethics Committee and conducted in accordance with the principles of the Declaration of Helsinki.

Blood samples from subjects were collected and processed at the time of hospital admission, and 24 and 72 h after admission. Plasma samples were obtained after centrifuging the blood at 2,000 rpm for 10 minutes at 4 °C and were stored at −80 °C in the Ramón y Cajal University Hospital-IRYCIS Biobank. The Department of Biochemistry at the Ramón y Cajal University Hospital analysed the main laboratory parameters related to renal function used in routine clinical practice, such as serum creatinine (sCreat), urea and estimated glomerular filtration rate (eGFR). Correlation analyses were performed between the different renal function variables, mineral metabolism parameters and cytokine IL-6 at the three study time points. In addition, all variables were analysed based on whether the subjects survived (no death) or died (death). Plasma levels of the proinflammatory cytokine IL-6 (ProcartaPlex™ Kit, Invitrogen, Paris, France), C-terminal FGF23 (cFGF23, ELISA, Immutopics, San Clemente, CA), and soluble α-klotho (ELISA, IBL, Fujioka-Shi, Japan) were determined using commercial enzyme-linked immunosorbent assay (ELISA) kits.

For the experimental study there were used, adult male mice C57BL/6J (wild-type strain, WT) from Charles River International Inc. (Wilmington, MA) aged 12–14 weeks. AKI was induced by a single intraperitoneal injection of folic acid (FA, 250 mg/kg, Sigma-Aldrich, Sant Louis) dissolved in 0.3 M NaHCO3; control mice were given 0.3 M NaHCO3 as a vehicle.19 Plasma samples were extracted from the different experimental groups 24 and 72 h after the FA injection. Transgenic mice overexpressing the klotho protein (Tg-Kl mice; EFmKl46), provided by Dr Kuro-O (Jichi Medical University, Japan), were also used. These mice are able to overexpress the transmembrane form of klotho ubiquitously by inserting the human elongation factor 1-α promoter (EFmKl).20Tg-Kl mice underwent the same AKI induction experimental procedure as WT mice and plasma samples were collected at the same time points. Animal studies were conducted in accordance with the European Union Guidelines for the Ethical Care of Experimental Animals (2010/63/EU) and were approved by the Autonomous University of Madrid Bioethics Committee and by the Directorate General of Agriculture and Environment of the Community of Madrid’s Regional Ministry of the Environment (PROEX 186.5/20).

Plasma levels of cFGF23 and IL-6 were analysed using ELISA kits following the manufacturers’ instructions. For the quantification of cFGF23, the ELISA kit from Immunotopics, Inc. (San Clemente, CA) was used, and IL-6 levels were analysed with the ProcartaPlex™ kit (Invitrogen, Paris, France).

To determine whether the data follow a normal distribution, the Kolmogorov-Smirnov test was performed. Differences between parametric variables were analysed using Student’s t-test or one-way ANOVA with the Newman-Keuls test. The probability of mortality in subjects with AKI based on cytokine IL-6 plasma concentration was calculated by calculating the area under the curve (AUC) using the ROC (receiver operating characteristic curve) analysis. The optimal cut-off value from which the IL-6 plasma concentration discriminates between subjects who survived versus those who did not was obtained by calculating the Youden index.21 The different correlations between the variables were calculated with the Pearson correlation coefficient. For statistical analysis of the data, GraphPad Prism v8.0 software (GraphPad Software Inc., San Diego, CA) and SPSS Statistics v22 (IBM, Armonk, NY) were used. Data are represented as mean ± standard deviation in the clinical study, and mean ± standard error in the experimental study. A p-value <0.05 was considered statistically significant.

Of the cohort of 28 subjects with AKI, 75% (21) were males with a mean age of 64.3 ± 16.1 years. According to the AKIN criteria for diagnosis and classification of AKI stages, of the total number of subjects, 8.7% were in stage 1, 17.4% in stage 2 and 73.9% in stage 3. The causes of AKI were as follows: acute tubular necrosis (19), prerenal (7), acute glomerulonephritis (1) and acute tubulointerstitial nephritis (1). Of the total number of subjects, six died (21.4%). Table 1 shows the results of the biochemical parameters by renal function, the concentration of biomarkers related to mineral metabolism in patients with AKI at the three study time points, as well as the values obtained in the control group. All subjects, regardless of the time point, had an eGFR lower than 60 ml/min/1.73 m2 after AKI, and the values of eGFR increased progressively and significantly after 72 h post-AKI as compared to the values at hospital admission and with the 24 h post-AKI (p < 0.001). The highest values for both serum creatinine (sCreat) and urea were recorded at hospital admission, decreasing significantly at 72 h (p < 0.05 vs admission, p < 0.05 vs 24 h post-AKI for serum creatinine). An improvement in creatinine clearance was also observed 72 h after admission compared to admission (p < 0.01). Supplementary Fig. S1 in Appendix A shows the results for the renal function parameters analysed, separating subjects who survived from those who did not. We found that serum creatinine decreased significantly between admission and 72 h in subjects who survived, while no changes were observed in those who died. Regarding urea, a significant decrease occurred in the survivor group after 72 h. In contrast, in the non-survivor group, urea levels remained significantly high compared to the survivor group after 72 h. Finally, eGFR increased after 72 h in survivors, while there were no changes in non-survivors. Regarding the markers of mineral metabolism, it was observed that as compared with the reference values of the control group, there was a significant elevation of cFGF23 at the three study time points (p < 0.001 vs admission, p < 0.01 vs 24 h post-AKI and p < 0.05 vs 72 h post-AKI), as well as a significant decrease in klotho (p < 0.001 vs admission, 24 h post-AKI and 72 h post-AKI). However, among AKI patients no significant differences were found in cFGF23 and klotho plasma levels between the three study time points. These results indicate the improvement in renal function after 72 h, was not associated with changes in cFGF23 and klotho marker levels at admission and at 24 and 72 h post-AKI.

Plasma IL-6 concentration was significantly higher in patients with AKI as compared with the control group at the three study time points, with the highest circulating IL-6 concentration at hospital admission (Fig. 1A, p < 0.001 vs control group). IL-6 levels decreased but remained significantly elevated compared to the reference control group 24 and 72 h after admission (Fig. 1A, p < 0.05 vs control group). A comparison of systemic IL-6 levels and the change over time of the different renal function parameters measured at the post-hospital admission time points (sCreat, urea and eGFR) revealed that, as renal function improved with an increase in eGFR (Fig. 1C), a decrease in sCreat (Fig. 1B) and a decrease in urea (Fig. 1D), IL-6 levels also decreased, indicating a clear relationship between acute kidney injury and the increase in this proinflammatory cytokine at hospital admission. Correlations between the different renal function parameters and IL-6 levels at the three study time points were analysed, without finding significant correlation between them (data not shown), except for a positive, significant correlation between urea and IL-6 levels after 72 h (Fig. 1E).

Figure 1.

Analysis of systemic levels of the proinflammatory cytokine IL-6 in subjects with AKI and its relationship with renal function. (A) Plasma IL-6 levels and the relationship with serum creatinine (sCreat) (B), estimated glomerular filtration rate (eGFR) (C) and urea (D) values in subjects with AKI at hospital admission, and 24 and 72 h after admission. (E) Correlation between systemic IL-6 levels and urea 72 h post-AKI. *p < 0.05, ***p < 0.001 vs control.

(0.34MB).

It was analysed whether plasma IL-6 levels varied in AKI subjects who survived (no death) versus those who ultimately died (death) at the three study time points. Fig. 2A shows that in surviving patients, the concentration of IL-6 decreased from admission to 72 h post-admission, while patients who died had higher IL-6 values compared to those who survived at all time points, particularly at 72 h post-admission (Fig. 2A, p < 0.05, p < 0.01 and p < 0.001 vs survivors after admission, 24 and 72 h, respectively). Furthermore, IL-6 levels after 72 h were significantly higher than admission in those subjects who died (Fig. 2A, p < 0.001 vs admission). For this reason, we analysed whether IL-6 could be a good biomarker for predicting mortality in our cohort of AKI patients by calculating the area under the curve (AUC) and analysing the ROC curves at the three study time points. Whether the AUC of a biomarker is one, the ability to predict an outcome is very high yielding a cut-off value from which the biomarker would indicate a high probability of mortality. Fig. 2B shows the AUCs of IL-6 at the three study time points in AKI patients, and Table 2 details the AUC values, confidence intervals (CI), cut-off points, sensitivity, specificity and p-values of IL-6 at admission and at 24 and 72 h post-admission. As can be observed, IL-6 is able to significantly discriminate between AKI subjects who will survive from those who will not at the three study time points, already achieving good discrimination at the subject’s admission to the hospital (AUC = 0.942; p = 0.0013; Fig. 2B), and reaching a total prediction with an AUC value of one 72 h post-hospital admission (Fig. 2B, Table 2, AUC = 1.000, 95% CI = 1.00–1.00, p = 0.001). The cut-off point for plasma IL-6 concentration values that predict mortality is 78.34 pg/ml at hospital admission and 41.50 pg/ml at 72 h post-admission (Table 2).

Figure 2.

IL-6 levels in subjects who survived AKI versus those who did not and their capacity to predict for mortality. (A) Plasma levels of cytokine IL-6 in subjects with AKI who survived (no death) versus those who died (death) at hospital admission, and 24 and 72 h after AKI diagnosis. (B) ROC curves to determine the predictive value of mortality of cytokine IL-6 at hospital admission (maroon line), 24 (red line) and 72 h after AKI diagnosis (pink line) (B). *p < 0.05, **p < 0.01, ***p < 0.001 vs their respective no death group by study time point.

(0.2MB).

In our cohort of patients, we previously observed (Table 1) a decrease in systemic levels of klotho after the AKI subject’s admission to hospital. For this reason, we studied its relationship with IL-6 levels at each time point analyzed. It was observed that there was a relevant decrease in the plasma klotho concentration compared to the group of healthy individuals at hospital admission, which coincided with the highest concentration of IL-6 observed in AKI subjects (Fig. 3A). In contrast, circulating levels of cFGF23 increased, peaking at admission, coinciding with the maximum values observed for IL-6. Both biomarkers decreased at 24 and 72 h post-FRA (Fig. 3B). The correlations between IL-6 and the markers of mineral metabolism, klotho and cFGF23, measured at the three study time points were analyzed. There was no correlation between IL-6 and klotho levels (Fig. 3C, E and G). However, there was a significant, positive correlation between the concentration of cFGF23 and IL-6 at all time points analysed: admission, and after 24 and 72 h (Fig. 3D, F and H).

Figure 3.

Relationship between IL-6 levels with mineral metabolism markers cFGF23 and klotho in subjects with AKI. Relationship between plasma cytokine IL-6 concentration and systemic klotho concentration (A) and cFGF23 (B). Correlations between systemic IL-6 levels with klotho at admission (C), 24 h (E) and 72 h after AKI (G), and with cFGF23 at admission (D), 24 h (F) and 72 h after AKI (H).

(0.44MB).

Since it was observed a decrease in plasma klotho levels in patients with AKI as compared to the control group, it was decided to study, in an experimental murine model of nephrotoxic AKI, induced by the injection of folic acid, whether klotho overexpression could prevent the increase in IL-6. First, it was analyzed the plasma IL-6 concentration in both WT mice and the klotho-overexpressing transgenic mouse strain (Tg-Kl), at 24 and 72 h after the induction of AKI. Since there was a positive and significant correlation between cFGF23 and IL-6 levels in AKI patients at hospital admission (Fig. 3D), as well as after 24 and 72 h (Fig. 3F and H), we analysed the cFGF23 levels in all experimental groups. It was found that cFGF23 increased significantly 24 h after AKI induction (Fig. 4A, p < 0.001); at 72 h post-AKI there was a in decrease in cFGF23 but it remained significantly elevated in WT mice (Fig. 4A, p < 0.05). However, in Tg-Kl mice with AKI, cFGF23 did not increase, and was significantly lower than in WT mice at 24 h post-AKI (Fig. 4A, p < 0.001). The same pattern was observed for circulating IL-6 levels in mice, with a significant increase in IL-6 in WT mice after 24 h from kidney injury (Fig. 4B, p < 0.001), decreasing after 72 h, while in Tg-Kl mice, IL-6 did not increase significantly, and its levels were significantly lower than those found in WT mice 24 h post-AKI (Fig. 4B, p < 0.001). Fig. 4C shows a comparison between IL-6 and cFGF23 levels in the two strains of mice with AKI with and without klotho overexpression, also it appears that klotho overexpression protects against the systemic increase of both molecules induced by AKI.

Figure 4.

Relationship between IL-6 and cFGF23 levels in mice with AKI induced in WT mice and Tg-Kl mice with klotho gene overexpression. Plasma cFGF23 levels (A) and cytokine IL-6 (B) in WT mice and klotho-overexpressing mice (Tg-Kl) 24 and 72 h after AKI induction. (C) Relationship between cytokine IL-6 and cFGF23 in WT and Tg-Kl mice 24 and 72 h after induction of AKI (C). *p < 0.05, ***p < 0.001 vs WT vehicle; ###p < 0.001 vs WT group 24 h post-AKI.

(0.31MB).