Archives of Toxicology https://doi.org/10.1007/s00204-018-2225-9

ORGAN TOXICITY AND MECHANISMS

Comparison of iohexol and iodixanol induced nephrotoxicity, mitochondrial damage and mitophagy in a new contrast-induced acute kidney injury rat model Wei Cheng1 · Fei Zhao1 · Cheng‑Yuan Tang1 · Xu‑Wei Li1 · Min Luo1 · Shao‑Bin Duan1  Received: 8 January 2018 / Accepted: 17 May 2018 © Springer-Verlag GmbH Germany, part of Springer Nature 2018

Abstract Recent progress in angiography and interventional therapy has revived interest in comparison of nephrotoxicity of low-or iso-osmolar contrast media, but detailed mechanisms and effective treatments of contrast-induced acute kidney injury (CIAKI) remain elusive. We established a new model of CI-AKI and compared the nephrotoxicity of iohexol and iodixanol with a focus on renal oxidative stress, mitochondrial damage and mitophagy. Our results showed that 48-h dehydration plus furosemide injection before iohexol administration successfully induced CI-AKI in rats. Compared with iodixanol, iohexol induced a greater decrease in renal function, more severe morphological damage and mitochondrial ultrastructural changes, an increased number of apoptotic cells, decreased antioxidative enzymes with activation of NLRP3 inflammasome in renal tissue. Renal contrast media kinetics showed the immediate excretion of iohexol and the transient renal accumulation of iodixanol. Plasma mtDNA Tc numbers were positively correlated with markers of renal mitochondrial disruption but negatively correlated with the level of serum creatinine and the score of tubular injury. Of note, iodixanol appeared to induce a stronger activation of mitophagy than iohexol, evidenced by greater protein levels of LC3II and PINK1/Parkin in the renal tissue of iodixanol-treated rats. Taken together, our results indicate that iohexol induced more severe nephrotoxicity than iodixanol in vivo due to apoptosis, destruction of antioxidative defense machinery, activation of NLRP3 inflammasome, mitochondrial damage and mitophagy. Plasma mtDNA may serve as a biological marker for renal mitochondrial disruption and damage in CI-AKI. Antioxidative defense and mitophagy are involved in the process of CI-AKI and may be promising targets of therapies. Keywords  Nephrotoxicity · Contrast media · Acute kidney injury · Oxidative stress · Mitochondrial damage · Mitophagy

Introduction Contrast-induced acute kidney injury (CI-AKI) is defined as the dynamic changes in serum creatinine (SCr) after exposure to iodine contrast media (CM), and has become the third leading cause of hospital-acquired acute kidney injury (Silver et  al. 2015). CI-AKI leads to prolonged Electronic supplementary material  The online version of this article (https​://doi.org/10.1007/s0020​4-018-2225-9) contains supplementary material, which is available to authorized users. * Shao‑Bin Duan [email protected] 1



Department of Nephrology, The Second Xiangya Hospital, Central South University, 139 Renmin Road, Changsha 410011, Hunan, People’s Republic of China

hospitalization and increased costs, and is a strong predictor of poor early and late outcomes (Centola et al. 2016; Luo et al. 2017). Current guidelines recommend intravenous hydration, use of low- (LOCM) or iso-osmolar contrast media (IOCM) and reduced volume of CM as prevention strategies for CI-AKI (Wang et al. 2016). However, detailed mechanisms and effective treatment of CI-AKI remain elusive, and literature reports are controversial of whether IOCM is associated with less risk for CI-AKI than LOCM. Previous reports have proved the vital roles of oxidative stress and inflammation in CI-AKI (Briguori et al. 2011; Liu et al. 2014; Yang et al. 2015). Many studies have reported the protective role of autophagy and mitophagy during renal ischemia/reperfusion (IRI) and cisplatin induced AKI (Kaushal and Shah 2016; Tang et al. 2015). However, it remains unclear whether nephrotoxicity of iohexol and iodixanol is associated with antioxidative

13

Vol.:(0123456789)



Archives of Toxicology

defense machinery, mitochondrial damage and mitophagy in CI-AKI. The current CI-AKI models are defective since highosmolar contrast media (HOCM) is administered, prepared with pharmacological or surgical procedures, which are time consuming and uncommon in clinical practice. It is necessary to establish an appropriate experimental animal model that can be used to understand the detailed mechanisms and explore strategies for prevention of this disease. In the present study, a new, highly efficient and reproducible rat model of CI-AKI was established. Nephrotoxicity induced by iohexol and iodixanol was evaluated, and the nature of antioxidative defense machinery and mitochondrial dysfunction in the development of CI-AKI was further dissected.

Materials and methods The nonionic low-osmolar contrast medium was (1) the LOCM, iohexol (350; 300 mg iodine/mL; GE Healthcare, Shanghai, China) and (2) the IOCM, iodixanol (320 mg iodine/mL; GE Healthcare, Shanghai, China) via tail vein administration; Furosemide (Harvest Pharmaceutical Co., Shanghai, China) was injected intramuscularly. Male Sprague-Dawley rats, weighing approximately 250–300 g, were acclimatized for 7 days before the start of study and handled in accordance with the guidelines on animal care of the Second Xiangya Hospital of Central South University. The definition of CI-AKI in European Society of Urogenital Radiology (ESUR) is a relative increase in SCr by 25% from baseline within 72 h after CM exposure (Thomsen and Morcos 2003), while CI-AKI is defined as an increase in SCr ≥ 50% from the baseline value within 48 h in Kidney Disease Improving Global Outcomes (KDIGO) (Kellum et al. 2013).The experiment was divided into three parts as following.

Confirmation experiment Another 12 rats were enrolled and randomly assigned to two experimental groups (n = 6). Animals were dehydrated for 48 h and underwent furosemide (10 ml/kg) injection 30  min before iohexol (350  mg iodine/mL, 15  ml/kg, group CM) or 0.9% normal saline (15 ml/kg, group NS) administration.

Comparative experiment Twenty-four rats with similar renal function were enrolled and randomly assigned to four experimental groups (n = 6). (1) NS group: dehydration for 48 h and saline (10 ml/kg) injection 30 min before 0.9% saline (15 ml/kg) administration; (2) FM group: dehydration for 48 h and furosemide (10 ml/kg) injection 30 min before 0.9% saline (15 ml/ kg) administration; (3) iodixanol group: dehydration for 48 h and furosemide (10 ml/kg) injection 30 min before iodixanol (15 ml/kg) administration; (4) iohexol group: dehydration for 48 h and furosemide (10 ml/kg) injection 30 min before iohexol (15 ml/kg) administration; iohexol (350 mg iodine /mL) 6 mL/kg and iohexol (300 mg iodine/ mL) 9 mL/kg were mixed to match the iodine concentration of iodixanol (320 mg iodine/mL) and injected; Each rat received 15 ml/kg of iohexol or iodixanol (equivalent to an iodine dose of 4.8 g iodine /kg body weight) in iohexol or iodixanol group. Animals were given saline, iohexol, or iodixanol intravenously at a rate of 1 ml/kg/min. All animals had ad  libitum access to water and food after injection.

Blood biomarkers assay Blood urea nitrogen (BUN) and SCr levels were measured by an automatic biochemical analyzer (Hitachi 7170A, Japan).

Preliminary experiment Thirty-six rats were randomly assigned to six experimental groups (n = 6): dehydration for 24 h and furosemide (10 ml/kg) injection 30 min before iohexol (15 ml/kg) administration (Group 1); dehydration for 48 h and normal saline (10 ml/kg) injection 30 min before iohexol (15 ml/ kg) administration (Group 2); dehydration for 48 h and furosemide (10 ml/kg) injection 30 min before iohexol administration (Group 3: iohexol 10  ml/kg; Group 4: iohexol 15 ml/kg; Group 5: iohexol 20 ml/kg); (Fig. 1a). Iohexol (350  mg iodine/mL) was used in preliminary experiment.

13

Haematoxylin and eosin staining Kidney tissue was fixed in 10% neutral-buffered formalin for 24  h, embedded in paraffin, and tissue sections 4 µm thick were cut using a microtome and stained with hematoxylin eosin for histopathological evaluation. For semiquantitative analysis of the frequency and severity of renal lesions, we selected randomly 10 high-magnification (×200) fields of the cortex and outer stripe of the outer medulla. The specimens were scored according to the extent of foamy degeneration and detachment of tubular

Archives of Toxicology

Fig. 1  A new rat model of CI-AKI. a, b Preliminary experiment. Thirty-six male Sprague-Dawley rats were deprived of water as indicated time interval and followed by furosemide injection or not 30  min before different doses of iohexol (350  mg iodine/mL; 10, 15, 20  ml/kg) administration. Blood samples were drawn before dehydration and 24  h after iohexol injection (*p < 0.01 vs. baseline SCr value, n = 6). Modeling rates of CI-AKI in rats were compared

according to definition of ESUR and KDIGO. c Confirmation experiment. Effect of either normal saline or iohexol (350  mg iodine/mL, 15 ml/kg) on serum creatinine. Another 12 male Sprague-Dawley rats were deprived of water for 48 h and followed by furosemide injection 30  min before normal saline or iohexol administration. Blood samples were drawn before dehydration and 24  h after iohexol injection (*p < 0.05 vs. baseline SCr value, #p < 0.05 vs. group NS, n = 6)

cells on a semiquantitative scale (Racusen and Solez 1986): no injury (0), mild < 25% (1), moderate < 50% (2), severe < 75% (3), very severe > 75% (4).

number in kidney sections were counted under a fluorescence microscope (Motic, BA410E). All cells were counted in five different views per section. TUNEL-positive cells were expressed as a percentage of total cells.

Transmission electron microscope examination (TME) The kidney samples fixed with 2.5% glutaraldehyde from rats of comparative experiment were treated with conventional dehydration, osmosis, embedding, sectioning and staining. The ultrastructure of the renal cells was observed under a Hitachi H7700 electron microscope.

dUTP nick‑end labeling assay Terminal deoxynucleotidyl transferase-mediated dUTP nickend labeling (TUNEL) assay was performed on the paraffin sections of the renal corticomedullary boundary zone with a commercial kit (In situ Cell Death Detection kit; Roche, Basel, Switzerland) according to the manufacturer’s protocol. The number of TUNEL-positive cells and total cell

Renal lipid peroxidation/superoxide dismutase assay The malondialdehyde (MDA), superoxide dismutase (SOD) oxidase activities in the renal tissue from rats of comparative experiment were measured using commercial kits according to the manufacturer’s protocol. (Jiancheng Bioengineering Institute, Nanjing, China).

Gene expression analyses by qPCR Plasma DNA from rats of comparative experiment was prepared using the QIAamp DNA Mini and Blood Mini kit (Qiagen, Germantown, MD). RNA isolated from frozen kidney tissues using Trizol reagent (Invitrogen, America) and RNA (1 µg) was reverse transcribed to cDNA using

13



Archives of Toxicology

Reverse Transcriptase M-MLV (Takara, Japan) according to manufacturer’s instructions. Primers were designed and synthesized by Sangon Biotech (Sangon Biotech, Shanghai, China) (Suppl. Table 1). Briefly, plasma mtDNA (measured by mt-co1/Cox1) (Unuma et al. 2015), renal mRNA expression of cytochrome-c oxidase-1 (COX1), peroxisome proliferator-activated receptor-γ coactivator 1α (PGC-1α), NADH dehydrogenase [ubiquinone] 1 beta subcomplex subunit 8 (NDUFB8), heme oxygenase-1 (HO-1), glutamate cysteine ligase modifier (GCLM) and NAD(P)H quinone oxidoreductase 1 (NQO1) were evaluated by quantitative PCR through a LightCycler480 System with SYBR Premix Ex Taq II (Takara, Japan). The relative abundances of plasma mtDNA were expressed as threshold cycles (Tc). Of note, higher Tc represents lower levels of mtDNA (Hu et al. 2017; Simmons et al. 2013). Other gene expression levels were calculated according to the ­2−ΔCq method. The data obtained were normalized to the expression of the stable reference gene (β-actin, ACTB).

Western blotting Proteins in the supernatants of renal homogenate from rats of comparative experiment were separated by 12% SDSPAGE and transferred to nitrocellulose membranes. The membranes were incubated with primary antibodies against [NOD]-like pyrin domain containing protein 3 (NLRP3) inflammasome and high mobility group box 1 (HMGB1), PTEN-induced putative kinase 1(PINK1), Parkin, p62, LC3I/II (1:1000 Abcam, Cambridge, UK) and β-actin (1:5000, Abcam, Cambridge, UK) followed by incubation with horseradish peroxidase conjugated secondary antibodies (1:5000, Abcam, Cambridge, UK). Protein bands were detected using ECL Western blotting detection reagent (Millipore Corporation, Billerica, USA) and quantified using the Tanon 5200 Multi image analysis software (Tanon, Shanghai, China).

Computed tomographic measurements Six male Sprague-Dawley rats were initially anesthetized using an inhalation of 4% isoflurane (Baxter, Germany) after the administration of iohexol or iodixanol [4.8 g (I)/ kg body weight, n = 3] based on methods of comparative experiment. CT device (SOMATOM Force, Germany) was used to monitor renal iodine concentrations by scanning the lower abdominal region of the rats. To assess the early kinetics of iohexol or iodixanol in the kidney, the first 300 s post injection (tail vein administration) was imaged dynamically with a sampling rate of 20 s. Additional single measurements were performed at 1-, 2-, 4-, 6-, and 24-h after tail vein administration to determine the renal CM retention for a period of 24 h. A calibration factor of 25 Hounsfield unit

13

(HU) per 1.0 mg iodine/g was used to convert the CT signals into iodine concentrations (Dean et al. 1983).

Statistical analysis Statistical analysis was performed using the statistical software SPSS Version 22.0. Data were represented as mean standard deviation (Mean ± SD) for independent experiments. For comparison between two groups, a paired t test was performed. Comparisons among more than two groups were assessed by one-way analysis of variance (ANOVA) followed by LSD test for post-hoc comparisons. The enumeration data were shown using the rate and Chi square test. The Spearman-rank correlation was used for analysis. p < 0.05 was considered statistically significant.

Results A new rat model of CI‑AKI In this study, we constructed a highly efficient, reproducible novel rat CI-AKI model based on water dehydration. There was no difference in SCr levels between the baseline value among each group (p > 0.05). After the interventions of dehydration alone, no difference of SCr levels were detected between the baseline value and 24 h after iohexol administration in Group 2. In other five groups (dehydration plus furosemide before iohexol administration), SCr levels increased significantly 24 h after iohexol administration (p < 0.05). We calculated the change of SCr and modeling rates and found that only in the group 4, SCr elevated more than 50% in all the rats (Fig. 1b). To confirm the change of SCr in CI-AKI model (group 4), 12 rats were dehydrated for 48 h and underwent furosemide injection 30 min before iohexol or saline administration. As shown in Fig. 1c, SCr was significantly increased after iohexol injection (p < 0.05), while no change was observed in NS group. We calculated the change of SCr and found that all rats of the CM group reached the definition of CI-AKI, suggesting that dehydration for 48 h plus furosemide injection 30 min before iohexol (350 mg iodine/mL, 15 ml/kg) administration is optimal to induce CI-AKI.

Iohexol resulted in marked deterioration of renal function compared with iodixanol To study the effects of iohexol and iodixanol on renal function, we compared renal function in four groups (NS, FM, iodixanol, and iohexol group) and found that there was no difference of SCr levels between NS and FM groups, but SCr increased significantly in iodixanol and iohexol group (p < 0.05). Furthermore, iohexol induced significantly higher

Archives of Toxicology

Fig. 2  Iohexol resulted in marked deterioration of renal function, more severe morphological injury and apoptotic death of tubular cells than iodixanol. a Representative photomicrographs of tubular cell injury in rat kidney tissue sections of the control (NS), furosemide (FM), iodixanol (4.8 g iodine/kg body weight) and iohexol (4.8 g iodine/kg body weight) groups. (Original magnifications: ×200. Hematoxylin and eosin stain). b Representative photomicrographs of glomerulus injury (arrow) in rat kidney tissue sections of the four groups (Original magnifications: ×4400. under a Hitachi H7700 electron microscope). c Immunofluorescent labeling for TUNEL in

rat kidney tissue sections of the four groups (×200). Rat kidney tissue was stained for TUNEL (green). Nuclei were stained with DAPI (blue). d, e Changes in the levels of Scr and BUN before dehydration and 24 h after saline or contrast medium injection in the four groups. f Quantitative analysis of histologic scoring. g Quantitative analysis of TUNEL-positive cells. Greater than 200 cells in each group were evaluated to determine the percentage of TUNNEL-positive cells. (#p < 0.05 vs. group NS, ▲p < 0.05vs. group FM, **p < 0.05 vs. group iodixanol n = 6). (Color figure online)

13



Archives of Toxicology

levels of SCr and BUN than iodixanol (p < 0.05), and only the iohexol group reached the definition of both ESUR and KDIGO (Fig. 2d, e).

Iohexol resulted in more severe morphological damage To compare the different effects of iohexol and iodixanol on morphological damage, kidney histopathological changes were examined. NS group were almost normal (histologic scoring 0.10 ± 0.14), a little noticeable detachment but no foamy degeneration of tubular cells in the FM group were observed (histologic scoring 0.93 ± 0.52); severe detachment and foamy degeneration of tubular cells were seen in the iohexol group, but less in the iodixanol group (histologic scoring 3.81 ± 0.18 vs. 2.98 ± 0.51). The differences among the four groups were statistically significant (p < 0.05) (Fig. 2a, f). We further employed TEM to detect the effects of iohexol and iodixanol on glomerulus injury. Obvious podocytes edema and endotheliocyte swelling with membrane rupture were observed in the rats injected iohexol. However, iodixanol only induced endotheliocyte swelling with partial membrane rupture (Fig. 2b).

Iohexol resulted in more apoptotic death of tubular cells compared with Iodixanol In the iohexol [TUNEL-positive cell (%) 33.64 ± 1.82] and iodixanol groups [TUNEL-positive cell (%) 15.76 ± 2.96], apoptotic cells were markedly increased compared with the NS group [TUNEL-positive cell (%) 3.79 ± 0.70] and FM group [TUNEL-positive cell (%) 8.03 ± 1.14]. When compared with the iodixanol group, the number of apoptotic cells in iohexol group was significantly higher (p < 0.05) (Fig. 2c, g).

Iohexol and iodixanol induced inflammation and oxidative stress We examined the expression of NLRP3 inflammasome and HMGB1 by Western blot and found iohexol induced a remarkable increase of NLRP3 compared with the other three groups (p < 0.05) (Fig.  3a) however, the level of HMGB1 was similarly increased in the iohexol and iodixanol groups (p < 0.05) (Fig. 3b). We further explored the expression of oxidant and antioxidant enzymes in CI-AKI model, and found that both iohexol and iodixanol resulted in comparable increases in the levels of MDA but decreases in the levels of SOD and GCLM (p < 0.05). However, the levels of HO-1 were markedly increased in the iodixanol group and iohexol group (p < 0.05), and higher expression of HO-1was induced by iodixanol (p < 0.05). Compared with

13

iodixanol, the levels of NQO1 decreased after iohexol treatment (p < 0.05) (Fig. 3c–e).

Iohexol and iodixanol induced mitochondrial ultrastructural changes in renal tubular epithelial cells To detect the effects of iohexol and iodixanol on mitochondrial ultrastructural changes, we employed TEM and found normal mitochondria, few lysosomes and autophagosomes were presented in NS and FM group. However, karyopyknosis and chromatin set in edge with a marked accumulation of autophagosomes in iohexol and iodixanol groups. In iodixanol group, the number of mitochondria with irregular morphologies decreased, some fragmented mitochondria and somewhat swollen mitochondria were observed. In iohexol group, the mitochondria showed more severe ultrastructural changes, including swollen mitochondrial, loss of cristae, and vacuolization in matrix (Fig. 4a).

Plasma mtDNA Tc numbers were correlated with renal mitochondrial disruption and renal damage To compare the effects of iohexol and iodixanol on mitochondrial integrity, plasma mtDNA (mt-co1/Cox1) and renal mRNA expression of Cox1, PGC-1α and NDUFB8 were assessed via qPCR. Results showed that plasma mtDNA levels in iohexol (Tc = 25.35) and iodixanol (Tc = 25.01) groups are significantly elevated compared with those in NS (Tc = 26.26) and FM (Tc = 26.58) groups (Fig. 4b). Threshold cycles (Tc) help to represent the relative abundances of plasma mtDNA, and of note, lower Tc indicates higher plasma mtDNA (Simmons et al. 2013). Plasma mtDNA Tc numbers were positively correlated with markers of mitochondrial disruption(Whitaker et al. 2015) (COX1: r = 0.56, p = 0.01; PGC-1α: r = 0.638, p = 0.002; NDUFB8: r = 0.635, p = 0.003) but negatively correlated with the level of serum creatinine (r = − 0.501, p = 0.024) and the score of tubular injury (r = − 0.607, p = 0.005), but was not associated with the level of BUN (r = − 0.386, p = 0.093) (Fig. 4c–h).

Iodixanol induced more mitophagy in renal tubular epithelial cells of rats To determine whether iohexol and iodixanol induced mitophagy, we examined the expression of PINK1, Parkin, p62, LC3II by Western blot (Fig. 5a). As shown in Fig. 5d, e, when compared with NS and FM groups, iohexol and iodixanol induced a rapid increase in LC3II and a significant decrease in p62 (p < 0.05), two biochemical hallmarks of autophagy activation. The different changes of LC3II in iohexol and iodixanol were associated with increased

Archives of Toxicology Fig. 3  Iohexol and iodixanol induced inflammation and oxidative stress. a Representative blots and densitometry of NLRP3 and b HMGB1 in rat kidney tissue were quantified by western blot analysis as estimates of inflammation; c MDA and d SOD concentrations in renal tissues. e Transcript levels of antioxidant gene HO-1, GCLM and NQO1 in renal tissues were quantified by qPCR. (#p < 0.05 vs. group NS, ▲p < 0.05 vs. group FM, **p < 0.05 vs. group iodixanol n = 6)

expression of PINK1, Parkin in iodixanol group (p < 0.05) (Fig.  5b, c), suggesting more PINK1/parkin mediated mitophagy induced by iodixanol. To further verify induced mitophagy, we assessed the co-localization of mitochondria and autophagosomes by TEM and evaluated autophagosome and mitophagosome formation by counting autophagosomal structures containing mitochondria per 50 cell sections. The data indicated more mitophagy induced by iodixanol than by iohexol (Fig. 5b, c).

Evaluation of renal CM kinetics After the administration of iohexol or iodixanol [4.8 g (I)/ kg body weight] based on 48-h dehydration plus furosemide injection, we calculated the change of SCr and found that SCr elevated more than 50% after iohexol treatment (18.50 ± 0.61 vs. 29.37 ± 0.97 µmol/L, n = 3, p < 0.05), but not in iodixanol group (17.93 ± 0.78 vs. 21.40 ± 0.26 µmol/L, n = 3, p < 0.05). We also observed rapid CM clearance from the kidneys. We found that HU values were decreasing from

13



Archives of Toxicology

Fig. 4  Plasma mtDNA Tc numbers were correlated with renal mitochondrial disruption and renal damage. a Iohexol and iodixanol induced mitochondrial ultrastructural changes in renal tubular epithelial cells. Representative photomicrographs of the renal morphology in NS group; representative photomicrographs of autophagosomes (black arrow), mitophagosomes (red arrow) mitochondrial swelling, loss of cristae (red triangle arrow), and vacuolization in matrix (black triangle arrow) were observed in FM, iodixanol and iohexol groups (Original magnifications: ×10000); b plasma mitochondrial DNA Tc number were decreases after iohexol and iodixanol treatment; c–e

plasma mtDNA Tc numbers were positively correlated with markers of mitochondrial disruption (COX1: r = 0.56, p = 0.01; PGC-1α: r = 0.638, p = 0.002; NDUFB8: r = 0.635, p = 0.003); f–h plasma mitochondrial DNA Tc number were negatively correlated with the level of Scr (r = − 0.501, p = 0.024) and the score of tubular injury, (r = − 0.607, p = 0.005) but was not associated with the level of BUN (r = − 0.386, p = 0.093). (#p < 0.05 vs. group NS, ▲p < 0.05vs. group FM, **p < 0.05 vs. group iodixanol n = 6). Tc threshold cycle. Higher Tc represent lower levels of Plasma mtDNA. (Color figure online)

220 s after iohexol injection. At this time point, the CT signal detected for iohexol was 1148.6 ± 247.9 HU [equivalent to an iodine dose of 45.96 ± 9.92 mg (I)/g]. For iodixanol, dynamic CT measurements revealed constantly increasing HU values up to 2184.3 ± 522.5 HU [equivalent to an iodine dose of 87.37 ± 20.88 mg (I)/g] at 280 s after iodixanol injection (Fig. 6a). However, renal CM concentrations at 1 h after the injection of iodixanol were on average 2.3-fold higher than the concentrations observed after the administration of iohexol. Renal CM retention was detected but near-constant concentrations during the measuring period from 4 to 24 h after i.v. with iodixanol (401.4 ± 41.7 to 314.7 ± 26.7 HU) and iohexol (213.8 ± 39.5 to 194.3 ± 27.0 HU) (Fig. 6b). In contrast, iodixanol (12.59 ± 1.07 mg iodine/g) resulted in a 1.6-fold higher iodine exposure than iohexol (7.77 ± 1.08 mg

iodine/g) in entire kidney 24 h after the injection of iohexol and iodixanol (Fig. 6c).

13

Discussion In this study, we established a new, highly efficient and reproducible rat model of CI-AKI based on dehydration and furosemide injection, demonstrating that iohexol induced more severe nephrotoxicity than iodixanol in CI-AKI rats. We also provided preliminary evidence of plasma mtDNA as a biomarker of renal mitochondrial disruption and renal damage in CI-AKI. CI-AKI is a serious complication in patients after angiography and interventional therapy and the pathogenesis

Archives of Toxicology

Fig. 5  Iodixanol induced more mitophagy in renal tubular epithelial cells of rats. Renal cortex and outer medulla were collected for immunoblot analyses of  PINK1, Parkin, p62, LC3I/II and transmission electronic microscopy analysis. a Representative blots. b Densitometry of PINK1. c Densitometry of Parkin. d Densitometry of P62. e Densitometry of LC3II. f Quantitation of autophagosome and

mitophagosome (autophagosomal structures containing mitochondria) by TEM. g Representative TEM images of autophagosomes (pointed by the black arrow) and mitophagosomes (pointed by red arrow) in each group. (#p < 0.05 vs. group NS, ▲p < 0.05 vs. group FM, **p < 0.05 vs. group iodixanol n = 6). (Color figure online)

is not fully understood. Dehydration with uninephrectomy (5/6 nephrectomy procedure) (Liu et al. 2014), pharmacological procedures (glycerin, indomethacin) (Duan et al. 2000; Kodama et al. 2013), IRI-based approach (Linkermann et al. 2013), dehydration with HOCM administration (Ozkan et al. 2012) are not optimal ways to induce AKI in animals. Water deprivation is regarded as an appropriate pretreatment method because it is easy to handle with need of no pharmacological or surgical procedures. Another study reported that dehydration for 3 days plus furosemide injection 20 min before contrast medium (10 ml/kg) administration is optimal to induce CI-AKI (Sun et al. 2014). In our study, multiple pretreatment plans were compared, such as dehydration time, with or without furosemide injection, different doses of CM to improve the modeling rate of CI-AKI. Finally, we found that water deprivation alone without furosemide injection was not a sufficient pretreatment to induce

CI-AKI, because dehydration for 48 h hardly influenced the blood volume or resulted in hypoperfusion. Effective blood volume further decreased after furosemide injection, and then the administration of the contrast medium caused direct nephrotoxicity and strong renal vasoconstriction with intensive renal tubular epithelial cell apoptosis and necrosis and renal parenchymal injury. In brief, our results show that dehydration for 48 h plus furosemide injection before contrast medium administration to induce CI-AKI rats model is more accurate, efficient and reproducible for it is common in clinical work and reached definition of both ESUR and KDIGO. Most studies have indicated that iodixanol appears to cause less nephrotoxicity in renal tubular epithelial cells than iohexol does (Reed et al. 2009), but pathophysiologic mechanism remains unclear. Our study found that iohexol induced significantly higher level of SCr and BUN than

13



Archives of Toxicology

Fig. 6  Evaluation of renal CM kinetics. a After the administration of iohexol or iodixanol [4.8 g (I)/kg body weight] based on methods of comparative experiment. The renal tissue iodine accumulation of iohexol and iodixanol were determined using X-ray fluorescence analyses: dynamic CT measurements show Hounsfield units during a time interval of 5 min after the administration of iohexol (n = 3) and iodix-

anol (n = 3); b additional CT measurements show Hounsfield units at baseline and at 1-, 2-, 4-, 6-, and 24-h after tail vein administration in the entire kidney; c pronounced iodine exposure (mg iodine/g) in entire kidney 24 h after the injection of iohexol and iodixanol. A calibration factor of 25 Hounsfield unit (HU) per 1.0  mg iodine/g was used to convert the CT signal into iodine concentrations

iodixanol, and morphological results showed that iohexol induced more severe injury of endothelial cells in glomerulus and mitochondrial damage in renal tubular epithelial cells than iodixanol. The number of apoptotic cells and expression of NLRP3 were significantly higher in the iohexol group than in the iodixanol group, but iodixanol induced higher expression of antioxidative enzymes (HO1, NQO1), more autophagy of renal tissue and mitophagy in renal tubular epithelial cells than iohexol. Our results indicated that iohexol induced more nephrotoxicity than iodixanol as reflected on increased apoptotic tubular cells, destruction of antioxidative defense machinery, activation

of NLRP3 inflammasome, mitochondrial damage and mitophagy. Our previous studies in  vitro have showed that cytotoxicity of CM was related to both osmolality and iodine concentration (Duan et al. 2006). Other scholars (Zhuang et al. 2011; Heglund et al. 1995) provided the main pharmacokinetic parameters (iohexol: Ke = 1.69, t1/2 = 24.60 min; and iodixanol: Ke = 1.675; t1/2 = 25 min) based on an onecompartment model. The analysis of early renal CM kinetics in this study yielded the following two patterns: the immediate excretion of the low-osmolar iohexol and the transient renal accumulation of iso-osmolar iodixanol formulations.

13

Archives of Toxicology

Taken together, iohexol seems to promote faster CM excretion when compared with iodixanol, but iohexol induced more severe renal damage than iodixanol, which indicated that rapid excretion or low renal tissue iodine concentration could not effectively reduce nephrotoxicity of iohexol in our rat model. Previous reports have proved that oxidative stress is an important mechanism of CI-AKI (Briguori et al. 2011; Liu et al. 2014; Yang et al. 2015). HMGB1 is an autophagy sensor under conditions of oxidative stress, as loss of HMGB1 inhibits autophagy and increases oxidative injury in response to reactive oxygen species (ROS) (Tang et al. 2011). In our study, both iohexol and iodixanol resulted in comparable increases in expression of HMGB1, MDA and decreased vitality of SOD and GCLM in renal tissue, but compared with iodixanol, antioxidative enzymes (HO-1 and NQO1) decreased after iohexol treatment. The results indicated that iohexol and iodixanol had different effects on the antioxidative defense machinery. Under the effects of iohexol, the antioxidative defense machinery is compromised leading to increased oxidative stress and mitochondrial dysfunction culminating in apoptosis and necrosis. Compared with iohexol, iodixanol has less effect on the antioxidative defense machinery that leads to stronger antioxidants to counter oxidative stress and reduces or prevents mitochondrial damage. Mitochondria are the main source and organelle target of ROS (Emma et al. 2016). It has been proved that there exists abnormal mitochondrial morphology, mitochondrial DNA mutation, decreased mitochondrial DNA copy number in AKI due to ischemia and toxic drugs (Hall et al. 2013) and a clear involvement of mitochondria in acute defense against cadmium-induced oxidative stress and toxicity of the kidney (Nair et al. 2015). Recent studies show that circulating mtDNA, a damage-associated molecular pattern, is released after mitochondria injured and contributes to cytokine production, splenic apoptosis, and tubular mitochondrial disorders via TLR9 (Tsuji et al. 2016). Urinary mitochondrial DNA may serve as a biomarker of mitochondrial disruption and renal dysfunction in acute kidney injury (Whitaker et al. 2015). Release of mtDNA into the cytosol depends on the NLRP3 inflammasome and mitochondrial ROS (Nakahira et al. 2011). NLRP3 inflammasome exerted primary effects on CI-AKI by modulating the cell apoptosis (Shen et al. 2016). Whether mtDNA participate in the pathogenetic process of CI-AKI still remains unclear. Our previous studies have shown that iohexol and iodixanol induced mitochondrial damage and ROS generation in HK-2 cells (Lei et al. 2018). In this study, we observed that plasma mtDNA Tc numbers were correlated positively with markers of renal mitochondrial disruption, but negatively correlated with the level of Scr and the score of tubular injury, which provide preliminary evidence of plasma mtDNA as a novel biomarker of renal

mitochondrial integrity and renal damage in CI-AKI, but the changes in these indicators may not be truly reflective of mitochondrial damage or disruption. The mechanistic relevance of plasma mtDNA levels in the status of renal mitochondria warrants further investigation. Mitophagy, one type of autophagy, is known as selective removal of damaged and depolarized mitochondria (Eiyama and Okamoto 2015). Mitophagy has been implicated in several kidney disease models, including acute renal IRI, diabetic kidney disease and FSGS mouse model (Tang et al. 2015). The best well-known signaling pathway of mitophagy is regulated by PINK1/Parkin and the PINK1/Parkin pathway of mitophagy is activated to protect against renal IRI (Tang et al. 2017). Our previous study showed that autophagy was activated in response to iohexol-induced cell injury and apoptosis and may exert a renoprotective role in HK-2 cells (Xiao et al. 2017). However, the difference between iohexol and iodixanol induced autophagy and mitophagy level remains unclear. The present study found that iohexol and iodixanol induced upregulation of autophagy in renal tissue and mitophagy in the tubular epithelial cells. Iodixanol induced more increased autophagy, mitophagy and PINK1/Parkin expression than iohexol in CI-AKI. Thus, we speculate that iodine contrast media causes mitochondrial damage of renal tubular epithelial cells. Mitophagy is activated to regulate mitochondrial quality control. If the number of damaged mitochondria exceeds the clearance capacity of mitophagy, mitophagy is compromised, dysfunctional mitochondria accumulate in the cell and produce a large number of ROS, which forms a vicious circle of mitochondria injury. Our study still has some limitations. First, we did not use drugs intervention, gene overexpression or knockout model to elucidate the possible regulatory mechanisms in CI-AKI. More importantly, additional experiments on humans should be done, and more work should be focused in the future on the specific effects of antioxidative defense machinery, mitochondrial damage and mitophagy in CI-AKI. In summary, we established a new model of CI-AKI, which can well mimic CI-AKI in clinical practice and is highly reproducible. Iohexol induced more severe nephrotoxicity than iodixanol in vivo due to apoptosis, destruction of antioxidative defense machinery, activation of NLRP3 inflammasome, mitochondrial damage and mitophagy in CI-AKI rats. Plasma mtDNA may serve as a biological marker for mitochondrial disruption and renal damage in CI-AKI. Our data provide an important theoretical basis for targeting antioxidative defense machinery, autophagy and mitophagy for the therapy of CI-AKI. Acknowledgements  This work was supported by the National Natural Science Foundation of China (No. 81570618) and National Science and Technology Support Program of China (No. 2013BAH05F02).

13



Archives of Toxicology

Compliance with ethical standards  Conflict of interest  All authors declare that they have no conflict of interest.

References Briguori C, Quintavalle C, De Micco F, Condorelli G (2011) Nephrotoxicity of contrast media and protective effects of acetylcysteine. Arch Toxicol 85(3):165–73. https ​ : //doi.org/10.1007/s0020​ 4-010-0626-5 Centola M, Lucreziotti S, Salerno-Uriarte D et al (2016) A comparison between two different definitions of contrast-induced acute kidney injury in patients with ST-segment elevation myocardial infarction undergoing primary percutaneous coronary intervention. Int J Cardiol 210:4–9. https​://doi.org/10.1016/j.ijcar​d.2016.02.086 Dean PB, Kivisaari L, Kormano M (1983) Contrast enhancement pharmacokinetics of six ionic and nonionic contrast media. Invest Radiol 18(4):368–74 Duan SB, Liu FY, Luo JA et al (2000) Nephrotoxicity of high- and low-osmolar contrast media. The protective role of amlodipine in a rat model. Acta Radiol 41(5):503–507 Duan S, Zhou X, Liu F et al (2006) Comparative cytotoxicity of highosmolar and low-osmolar contrast media on HKCs in vitro. J Nephrol 19(6):717–24 Eiyama A, Okamoto K (2015) PINK1/Parkin-mediated mitophagy in mammalian cells. Curr Opin Cell Biol 33:95–101. https​://doi. org/10.1016/j.ceb.2015.01.002 Emma F, Montini G, Parikh SM, Salviati L (2016) Mitochondrial dysfunction in inherited renal disease and acute kidney injury. Nat Rev Nephrol 12(5):267–80. https​://doi.org/10.1038/nrnep​ h.2015.214 Hall AM, Rhodes GJ, Sandoval RM, Corridon PR, Molitoris BA (2013) In vivo multiphoton imaging of mitochondrial structure and function during acute kidney injury. Kidney Int 83(1):72–83. https​://doi.org/10.1038/ki.2012.328 Heglund IF, Michelet AA, Blazak WF, Furuhama K, Holtz E (1995) Preclinical pharmacokinetics and general toxicology of iodixanol. Acta Radiol Suppl 399:69–82 Hu Q, Ren J, Wu J et al (2017) Urinary mitochondrial DNA levels identify acute kidney injury in surgical critical illness patients. Shock 48(1):11–17. https​://doi.org/10.1097/SHK.00000​00000​00083​0 Kaushal GP, Shah SV (2016) Autophagy in acute kidney injury. Kidney Int 89(4):779–91. https​://doi.org/10.1016/j.kint.2015.11.021 Kellum JA, Lameire N, Group KAGW. (2013) Diagnosis, evaluation, and management of acute kidney injury: a KDIGO summary (Part 1). Crit Care 17(1):204. https​://doi.org/10.1186/cc114​54 Kodama A, Watanabe H, Tanaka R et  al (2013) A human serum albumin-thioredoxin fusion protein prevents experimental contrast-induced nephropathy. Kidney Int 83(3):446–54. https​://doi. org/10.1038/ki.2012.429 Lei R, Zhao F, Tang C-Y et al (2018) Mitophagy plays a protective role in iodinated contrast-induced acute renal tubular epithelial cells injury. Cell Physiol Biochem 46(3):975–985. https​://doi. org/10.1159/00048​8827 Linkermann A, Heller JO, Prokai A et al (2013) The RIP1-kinase inhibitor necrostatin-1 prevents osmotic nephrosis and contrast-induced AKI in mice. J Am Soc Nephrol 24(10):1545–1557. https​://doi. org/10.1681/ASN.20121​21169​ Liu TQ, Luo WL, Tan X et  al (2014) A novel contrast-induced acute kidney injury model based on the 5/6-nephrectomy rat and nephrotoxicological evaluation of iohexol and iodixanol

13

in  vivo. Oxid Med Cell Longev 2014:427560. https​: //doi. org/10.1155/2014/42756​0 Luo M, Yang Y, Xu J et al (2017) A new scoring model for the prediction of mortality in patients with acute kidney injury. Sci Rep 7(1):7862. https​://doi.org/10.1038/s4159​8-017-08440​-w Nair AR, Lee WK, Smeets K et al (2015) Glutathione and mitochondria determine acute defense responses and adaptive processes in cadmium-induced oxidative stress and toxicity of the kidney. Arch Toxicol 89(12):2273–2289. https​://doi.org/10.1007/s0020​ 4-014-1401-9 Nakahira K, Haspel JA, Rathinam VA et al (2011) Autophagy proteins regulate innate immune responses by inhibiting the release of mitochondrial DNA mediated by the NALP3 inflammasome. Nat Immunol 12(3):222–30. https​://doi.org/10.1038/ni.1980 Ozkan G, Ulusoy S, Orem A et al (2012) Protective effect of the grape seed proanthocyanidin extract in a rat model of contrastinduced nephropathy. Kidney Blood Press Res 35(6):445–53. https​://doi.org/10.1159/00033​7926 Racusen LC, Solez K (1986) Nephrotoxic tubular and interstitial lesions: morphology and classification. Toxicol Pathol 14(1):45–57. https​://doi.org/10.1177/01926​23386​01400​106 Reed M, Meier P, Tamhane UU, Welch KB, Moscucci M, Gurm HS (2009) The relative renal safety of iodixanol compared with low-osmolar contrast media: a meta-analysis of randomized controlled trials. JACC Cardiovasc Interv 2(7):645–54. https​:// doi.org/10.1016/j.jcin.2009.05.002 Shen J, Wang L, Jiang N et al (2016) NLRP3 inflammasome mediates contrast media-induced acute kidney injury by regulating cell apoptosis. Sci Rep 6:34682. https​://doi.org/10.1038/srep3​ 4682 Silver SA, Shah PM, Chertow GM, Harel S, Wald R, Harel Z (2015) Risk prediction models for contrast induced nephropathy: systematic review. BMJ 351:h4395. https​://doi.org/10.1136/bmj.h4395​ Simmons JD, Lee YL, Mulekar S et al (2013) Elevated levels of plasma mitochondrial DNA DAMPs are linked to clinical outcome in severely injured human subjects. Ann Surg 258(4):591–596. https​ ://doi.org/10.1097/SLA.0b013​e3182​a4ea4​6 (discussion 596–8) Sun S, Zhang T, Nie P et al (2014) A novel rat model of contrastinduced acute kidney injury. Int J Cardiol 172(1):e48–50. https​:// doi.org/10.1016/j.ijcar​d.2013.12.066 Tang D, Kang R, Livesey KM, Zeh HJ 3rd, Lotze MT (2011) High mobility group box 1 (HMGB1) activates an autophagic response to oxidative stress. Antioxid Redox Signal 15(8):2185–2195. https​ ://doi.org/10.1089/ars.2010.3666 Tang C, He L, Liu J, Dong Z (2015) Mitophagy: basic mechanism and potential role in kidney diseases. Kidney Dis (Basel) 1(1):71–9. https​://doi.org/10.1159/00038​1510 Tang C, Han H, Yan M et al. (2017) PINK1-PRKN/PARK2 pathway of mitophagy is activated to protect against renal ischemiareperfusion injury. Autophagy. https​://doi.org/10.1080/15548​ 627.2017.14058​80 Thomsen HS, Morcos SK (2003) Contrast media and the kidney: European Society of Urogenital Radiology (ESUR) guidelines. Br J Radiol 76(908):513–518. https​://doi.org/10.1259/bjr/26964​464 Tsuji N, Tsuji T, Ohashi N, Kato A, Fujigaki Y, Yasuda H (2016) Role of mitochondrial DNA in septic AKI via toll-Like receptor 9. J Am Soc Nephrol 27(7):2009–2020. https​://doi.org/10.1681/ ASN.20150​40376​ Unuma K, Aki T, Funakoshi T, Hashimoto K, Uemura K (2015) Extrusion of mitochondrial contents from lipopolysaccharide-stimulated cells: involvement of autophagy. Autophagy 11(9):1520– 1536. https​://doi.org/10.1080/15548​627.2015.10637​65 Wang N, Qian P, Yan TD, Phan K (2016) Periprocedural effects of statins on the incidence of contrast-induced acute kidney injury: a systematic review and trial sequential analysis. Int J Cardiol 206:143–52. https​://doi.org/10.1016/j.ijcar​d.2016.01.004

Archives of Toxicology Whitaker RM, Stallons LJ, Kneff JE et al (2015) Urinary mitochondrial DNA is a biomarker of mitochondrial disruption and renal dysfunction in acute kidney injury. Kidney Int 88(6):1336–1344. https​://doi.org/10.1038/ki.2015.240 Xiao L, Xu X, Zhang F et al (2017) The mitochondria-targeted antioxidant MitoQ ameliorated tubular injury mediated by mitophagy in diabetic kidney disease via Nrf2/PINK1. Redox Biol 11:297–311. https​://doi.org/10.1016/j.redox​.2016.12.022 Yang SK, Duan SB, Pan P, Xu XQ, Liu N, Xu J (2015) Preventive effect of pentoxifylline on contrast-induced acute kidney injury

in hypercholesterolemic rats. Exp Ther Med 9(2):384–388. https​ ://doi.org/10.3892/etm.2014.2132 Zhuang L, Gao J, Zeng Y et al (2011) HPLC method validation for the quantification of lomustine to study pharmacokinetics of thermosensitive liposome-encapsulated lomustine containing iohexol for CT imaging in C6 glioma rats. Eur J Drug Metab Pharmacokinet 36(2):61–9. https​://doi.org/10.1007/s1331​8-011-0030-4

13