Kidney-based in vivo model for drug-induced nephrotoxicity testing

The need is critical and urgent for a real-time, highly specific, and sensitive acute kidney injury biomarker. This study sought to establish a sensitive and specific Miox-NanoLuc transgenic mouse for early detection of drug-induced nephrotoxicity. We generated Miox-NanoLuc transgenic mice with kidney-specific NanoLuc overexpression. Our data showed that Miox-NanoLuc-produced luminescence was kidney-specific and had good stability at room temperature, 4 °C, − 20 °C, and repeated freeze–thaw cycles. Serum levels of BUN and creatinine were significantly increased at day 2 or 3 in cisplatin-treated mice and at day 5 in aristolochic acid (AAI)-treated mice. Particularly, the serum and urine Miox-NanoLuc luminescence levels were significantly increased at day 1 in cisplatin-treated mice and at day 3 in AAI-treated mice. Renal pathological analysis showed that the kidney sections of cisplatin-treated mice at day 5 and AAI-treated mice at day 13 showed cytolysis and marked vacuolization of tubular cells. In conclusion, we developed a new platform to early quantify drug-induced nephrotoxicity before serum BUN and creatinine levels increased and pathological tubular cell injury occurred. This model may serve as an early detection for drug- and food-induced nephrotoxicity and as an animal model to investigate tubular cell injury.

www.nature.com/scientificreports/ In addition, cadmium-induced toxicity in mice kidney significantly increased Miox levels in the tubular cells 28 , and the Miox concentration in the serum and urine may negatively reflect the rental function 29 . Since Miox is a highly conserved enzyme in mouse, rat, and human kidney, Miox is likely to be an ideal biomarker of AKI for early detection of kidney injury. Since NanoLuc luciferase (Nluc) has higher sensitivity than the commonly used bioluminescent firefly luciferase (FFluc) 30,31 , the purpose of this study was to develop a new Miox-NanoLuc transgenic mice platform to quantify the drug-induced nephrotoxicity by detecting the kidney-specific Miox-NanoLuc produced luminescence in mouse serum and urine.

Materials and methods
Animals. Wild Briefly, the 50 mer homologous arms (HR, black capital letters in Fig. 1A) flanking the start codon on the exon 1 was capped upon a counter selector rpsL-neo by polymerase chain reaction (PCR), and was electroporated into E. coli. hosting the BAC clone to insert the rpsL-neo by homologous recombination. The NanoLuc-polyA (Red capital letters in Fig. 1A) capping the same 50 mer homologous arms was then used to replace the counter selector rpsL-neo again by homologous recombination so as to construct the transgene Miox-NanoLuc (Fig. 1A).
The detailed protocol described in the instruction manual of the counter selection BAC modification kit (Gene Bridges, Heidelberg, Germany). The BAC transgene was purified, Sal I digest, and pulsed-field gel electrophoresis to isolate the 52.1 kb transgene for C57BL/6 mouse pronuclear microinjection. Drug-induced nephrotoxicity. Eight-week-old male WT mice and age-matched Miox-NanoLuc transgenic male mice were used for examining the drug-induced nephrotoxicity. Various doses of cisplatin (10, 20, or 40 mg/kg body weight) or Aristolochic acid I (AAI, Sigma-Aldrich; 2 or 3.5 mg/kg body weight) were injected intraperitoneally in Miox-NanoLuc transgenic mice. The vehicle control mice were injected with normal saline (cisplatin group) or corn oil (AAI group). After collecting the blood and urine samples from each mouse, the levels of nephrotoxic biomarkers in the blood and urine were measured. The kidneys of the mice were used to examine various morphologic, biochemical, and pathologic studies. www.nature.com/scientificreports/ The transgene was constructed using the mouse BAC clone RP23-94D6 as backbone with an inserted Miox gene containing its 5′UTR and 3′UTR. The NanoLuc-PolyA expression cassette was inserted into the start codon of Miox by RED/ET recombination system as described in "Materials and methods". (B) PCR genotyping of wild type and transgenic Miox-NanoLuc mice. RT-PCR was performed to examine the wild-type (WT), heterozygous Miox-NanoLuc mice (Tg/+) using different primer sets as described in "Materials and Methods". A single PCR product of 449-bp in length was amplified from WT mice, while the two 449-bp and 514-bp PCR fragments were amplified from the Tg/+ mice. (C) The expression of Miox protein in serum, urine and kidney were further confirmed by ELISA. There were no significant differences in the Miox expression between B6 mice, WT, and Tg/+ transgenic mice. Quantitative data were represented as mean ± standard error (SE). The n.s. symbol indicated no significance statistically. (D) Localization of Miox-NanoLuc in the proximal renal tubule cells. Before and after cisplatin treatment, the localization of Miox-NanoLuc luminescence signal in the kidney was determined by histological analysis. In order to measure the luminescence intensity of mouse tissue, the incised mouse tissues of liver, heart, spleen, lung and kidney were treated with 0.5 µl of Reporter Lysis Buffer (Promega, WI), ground, and sonicated for 1 min. The sample was centrifuged at 4 °C for 20 min, and the supernatant was collected to analyze the protein concentration by Protein Assay Dye Reagent Concentrate kit (Bio-Red, CA). The following measurement steps were the same as above.

Measurement of the creatinine and BUN levels in serum and urine.
Enzyme-linked immunosorbent assay (ELISA). ELISA was used to detect the levels of urine KIM-1 (R&D system, MN), cystatin C (R&D system, MN), clusterin (R&D system, MN), β2-microglobulin (MyBio-Source, CA), trefoil factor-3 (Cloud-clone corp., TX), and Miox (MyBioSource, CA). Briefly, 100 µl of Assay Diluent was added into 96-well microplate, and 50 µl of standard, control, or sample were then added to each well. After incubation at 4 °C for 2 h, each well was washed 3 times and added detection antibody at 4 °C for 2 h. After washed with PBS, HRP-conjugated antibody was added into each well and further incubated for 1 h. After PBS washing, TMB substrates were added to each well for 20 min, and stop solution was added to stop the reaction. The absorbance of each well was measured by Epoch microplate spectrophotometer (Biotek, VT).

Hematoxylin and Eosin (H&E) staining.
After sacrifice, the mouse kidney tissues were fixed with 10% formalin and embedded in paraffin. The embedded tissues were then sliced into 3 μm thick section, mounted on glass slides, and incubated at 75 °C for 30 min. After deparaffinization using xylene for 10 min, the specimens were rehydrated using a graded ethanol series (95%, 85%, and 70% ethanol). After washing, the specimens were treated with Hematoxylin for 2 min, washed in running tap water for 1 min, and then incubated with acid alcohol for 1 s. Afterwards, the specimens were incubated with ammonia water solution for 1 s and washed in running tap water for 10 min. After counterstaining in Eosin solution for 90 s, the specimens were dehydrated using 70%, 80%, 90%, and 100% ethanol. Finally, the specimens were mounted with mounting medium, and the kidney injury was observed under the microscopy.
Statistical analysis. Statistical data were expressed as mean ± standard error (SE). Difference of the expression intensities between wild type mice and Miox-NanoLuc transgenic mice was compared using two-sample t-test. Difference in serum BUN, serum creatinine, serum luminescence, urine luminescence, urine luminescence/creatinine, and change of body weight after treating with vehicle, cisplatin (10, 20, or 40 mg/kg), or AAI (2 or 3.5 mg/kg) were compared using one-way ANOVA with a post hoc-pairwise, Dunnett's test as comparing with vehicle-treated group for each follow-up time point. A non-parametric method, Kruskall-Wallis test with post hoc-pairwise, Mann-Whitney test when the baseline measurement (at Day 0) didn't follow normal distribution. All data were graphed using GraphPad Prism 6 for Windows (version 6.01) © 1992-2012 GraphPad Software, Inc. All statistical analyses were carried out with IBM SPSS statistical software version 22 for Windows (IBM, NY).
Results establishment of Miox-NanoLuc transgenic mice. According to analysis results using bioinformatic database (BioGPP, https ://biogp s.org; GTX, https ://www.infor matic s.jax.org/expre ssion .shtml ; GENEvisible, https ://genev isibl e.com/searc h), Miox was found to be specifically expressed in epithelium of proximal tubule. Therefore, Miox was selected as the host promoter for renal-specific transgene expression. The NanoLuc-PolyA expression cassette was inserted into the start codon of Miox gene in a BAC clone (RP23-94D6) by RED/ET recombination system, resulting in NanoLuc-PolyA instead of Miox expression driven by Miox promoter. The BAC transgene was purified, Sal I digested, and pulsed-field gel electrophoresis to isolate the 52.1 kb transgene for C57BL/6 mouse pronuclear microinjection (Fig. 1A). We got 26 pups in which two founders (line 6 and 28) carry the transgene, which was verified by genotyping PCR. As shown in Fig. 1B, a single PCR product of 449-bp in length was amplified in wild type mice, while, the two 449-bp and 514-bp PCR fragments were amplified in heterozygous transgenic mice (Tg/+). In addition, the expression levels of Miox in serum, urine, and kidney were determined by ELISA. There was no significant difference in Miox expression between B6  4 weeks storage, the urine was repeated freeze-thaw at various time points, and then the luminescence intensity was measured. Radiance (means ± SEM; n = 6 mice). *< 0.05, **p < 0.01, and ***p < 0.001 were considered statistically significance between Miox-NanoLuc transgenic mice and WT mice. www.nature.com/scientificreports/ mice, wild-type mice and Miox-NanoLuc transgenic heterozygous mice (Fig. 1C). To further confirm whether Miox-NanoLuc was localized in the proximal tubular cells, histological analysis was performed to examine the luminescent NanoLuc signal in the kidney of NanoLuc transgenic mice. Strong luminescence intensity of Nano-Luc was indeed localized in the renal tubule cells (Left panel, Fig. 1D). However, the luminescence intensity of NanoLuc in the renal tubule cells disappeared dramatically after cisplatin treatment of Miox-NanoLuc transgenic mice (Right panel in Fig. 1D). This result confirmed that Miox-NanoLuc was localized in proximal tubular cells, and release during AKI.
Tissue-specific expression of Miox-NanoLuc in kidney. The expression of Miox-NanoLuc in Miox-NanoLuc transgenic mice was further confirmed by IVIS analysis system. A shown in Fig. 2A, a strong luminescence intensity was observed only in the kidney of Miox-NanoLuc transgenic mice, but not in wild-type mice. Furthermore, Miox luminescence in liver, heart, spleen, lung and kidney tissues of Miox-NanoLuc transgenic mice was further examined. The luminescence intensity was significantly increased in the kidney of Miox-Nano-Luc transgenic mice (Fig. 2B,C). Consistent result was also observed in various tissue of Miox-NanoLuc transgenic mice by luciferase activity assay (Supplementary Fig. 1). We then examined the Miox luminescence in the serum and urine of Miox-NanoLuc transgenic mice. Compared to wild-type mice, Miox-NanoLuc transgenic mice have significant elevated serum and urine luminescence levels (Fig. 2D). These results indicated that Miox-NanoLuc is specifically expressed in the serum, urine and kidney tissues of Miox-NanoLuc transgenic mice.

Good stability of Miox-NanoLuc luminescence.
To examine the stability of Miox-NanoLuc luminescence in serum and urine, the serum and urine of Miox-NanoLuc transgenic mice were stored at room temperature, 4 °C or − 20 °C. Afterwards, the luminescence intensity at each storage time point was further evaluated. As shown in Fig. 2E, luminescence intensity can still be detected in serum and urine after storing at room temperature for 60 min. The luminescence in serum and urine can be stably detected even if stored at 4 °C or − 20 °C for 4 weeks (Fig. 2F,G). In addition, the luminescence was still detected in serum and urine after repeated freezethaw (Fig. 2H).

Miox-NanoLuc luminescence is as an early biomarker for drug-induced nephrotoxicity. Next,
we examined whether Miox-NanoLuc luminescence can be detected earlier than BUN and creatinine under drug-induced nephrotoxicity. As shown in Fig. 3A,B, serum BUN and creatinine levels were significantly increased after 2 or 3 days of cisplatin treatment (40 mg/kg). Similarly, serum BUN and creatinine levels were also significantly increased after 5 days of AAI treatment (3.5 mg/kg; Fig. 4A,B). Remarkably, serum Miox-NanoLuc luminescence was significantly increased after 1 day of cisplatin treatment (Fig. 3C) and 3 days of AAI treatment (Fig. 4C). Consistently, urine Miox-NanoLuc luminescence increased significantly after 1 day of cisplatin treatment (Fig. 3D) and 3 day of AAI treatment (Fig. 4D), indicating that changes of Miox-NanoLuc can be detected earlier than creatinine and BUN. On the other hand, cisplatin (Fig. 3E) and AAI treatment (Fig. 4E) resulted in weight loss. Next, we determine whether renal tubular injury was occurred after cisplatin and AAI treatment. The glomeruli of vehicle-treated mice had no collapse of the capillary loops and pathological injury (Fig. 3F,F), while the renal tubular cells of cisplatin-treated mice (10 and 20 mg/kg) showed little renal tubular injury (Fig. 3G,H). However, in the 40 mg/kg cisplatin-treated mice, renal tubular cells showed cytolysis, detachment from the underlying basal lamina, and loss of brush border as well as tubular cell vacuolization (Fig. 3I).
Similarly, AAI treatment also caused tubular cell lysis, detachment from the basal membrane, and vacuolization ( Fig. 4G,H, and Supplementary Fig. 2). In summary, the above results indicated that Miox-NanoLuc luminescence can not only reflect cisplatin-and AAI-induced nephrotoxicity, but also can be detected before the pathological change of renal tubules damage. We further examined the abilities of other biomarkers such as KIM-1, cystatin, clusterin, and trefoil factor 3 in the early detection of drug-induced nephrotoxicity. As shown in Fig. 5, KIM-1, cystatin, clusterin, and urine proteins began to increase significantly after 2-3 days of 40 mg/ kg cisplatin treatment. However, trefoil factor 3 level was not changed during cisplatin treatment (Fig. 5D). Taken together, these results suggested that Miox-NanoLuc luminescence can be used as an early biomarker for drug-induced nephrotoxicity.

Discussion
Since serum creatinine is insensitive and nonspecific for identifying kidney injury 21,32 , creatinine is an unreliable biomarker for acute changes in renal function. Before the level of creatinine increased, the glomerular filtration rate (GFR) may be significantly reduced. Up to 50% of renal function may have already been lost before creatinine may change. Consequently, creatinine is a poor biomarker for AKI due to its inability to help diagnose early acute kidney failure or differentiate among its various causes 16 . The present study further showed that, compared to creatinine and BUN levels, the Miox-NanoLuc intensity was significantly increased at the early stage of cisplatin-or AAI-induced AKI. In particular, a significant change in serum luminescence appeared in drug-treated mice before pathological changes occurred, suggesting that the Miox is an suitable biomarker that reflects drug-induced nephrotoxicity before the pathological changes.
Although urinary and serum proteins have been intensively investigated as potential biomarkers for the diagnosis of kidney injury, there is still a lack of ideal biomarkers for AKI. All of the potential biomarkers identified so far have their own strengths and weaknesses, so that a combination of biomarkers might be the optimal solution. In the present study, urine luminescence significantly increased in cisplatin-treated Miox-NanoLuc transgenic mice on day 1, but the levels of KIM-1, clusterin, cystatin, trefoil factor 3, and urine protein were not significant increased. The results were also supported by Vinken et al. 33 that KIM-1 together with clusterin increased after 3 days of cisplatin exposure, whereas creatinine, BUN, and urinary total proteins increased after 5 days of Scientific RepoRtS | (2020) 10:13640 | https://doi.org/10.1038/s41598-020-70502-3 www.nature.com/scientificreports/ cisplatin exposure. Similarly, in the drug-induced AKI rat models by Sasaki et al 34 , KIM-1 and clusterin levels were dramatically after administration of 100 mg/kg gentamicin for 7 days. Therefore, the release of endogenous renal Miox proteins may be faster than the inducible biomarkers KIM-1, clusterin, cystatin, and trefoil factor 3 as well as the traditional kidney injury indicators creatinine and BUN. These results suggested that Miox is a good biomarker for the early detection of drug-induced nephrotoxicity, since changes are detected at an early stage of proximal tubule cell injury. Although many proteins are considered as potential biomarkers of kidney injury [16][17][18][19] , most of these biomarkers were less constant in blood or unstable in normal urine, and therefore rely more on other extra-renal pathologies. By contrast, our data exhibited that continued luminescence was detected in serum or urine stored at room temperature for 60 min. In addition, whether stored at 4 °C or − 20 °C for weeks, or repeatedly thawed and frozen, the luminescence in serum or urine was still detectable. This suggests that luminescence of Miox-NanoLuc has good stability in both serum and urine. . Miox-NanoLuc luminescence as an early biomarker for cisplatin-induced nephrotoxicity. After intraperitoneal injection of vehicle or cisplatin (10, 20, or 40 mg/kg), the levels of (A) serum BUN, (B) serum creatinine, (C) serum Miox-NanoLuc luminescence, (D) urine Miox-NanoLuc luminescence, and (E) percent of change in mouse body weight were measured at day 0, 1, 2, 3, and 5. The pathological renal damage caused by vehicle (F), 10 mg/kg cisplatin (G), 20 mg/kg cisplatin (H), and 40 mg/kg cisplatin were detected by H&E staining at day 5. Data were presented as mean ± SE and compared using one-way ANOVA with a post hocpairwise, Dunnett's test. *< 0.05, **p < 0.01, and ***p < 0.001 were considered to be statistically significance between vehicle and 40 mg/kg cisplatin. # < 0.05, ##p < 0.01, and ###p < 0.001 were considered to be statistically significance between vehicle and 20 mg/kg cisplatin. & < 0.05 and && p < 0.01 were considered to be statistically significance between vehicle and 10 mg/kg cisplatin. www.nature.com/scientificreports/ The next question was focused on the means by which highly expressed Miox-NanoLuc resulted in druginduced proximal tubule cell injury. Previous studies demonstrated that necrosis/apoptosis and oxidant stress are the factors most at play in enhanced nephrotoxicity 35 . The Miox promoter includes an oxidant response element, and its overexpression will accentuate the generation of ROS in vitro 36,37 . Miox is upregulated during oxidative stress and hyperglycemia in vitro and in vivo, and this upregulation aggravates tubulointerstitial injury 38,39 . Under high glucose conditions, overexpression of Miox has been reported to cause mitochondrial dysfunctions in LLC-PK1 cells, including cytochrome C release, DNA fragmentation, and mitochondrial fission 37 . Moreover, in cisplatin-induced nephrotoxicity, the increased ROS generation in Miox-overexpressed cell could be related to pathology of the NAPDH oxidase system 37,40 . In AAI-induced nephrotoxicity, the effect was associated with a decrease in renal function, massive necrosis, increased inflammation, and oxidative stress 41 . Taken together, these results suggest that cisplatin-or AAI-induced nephrotoxicity appears to result in an increase in Miox-NanoLuc expression in Miox-NanoLuc transgenic mice, most likely via ROS generation.
Food and drug safety are vital to the health of populations. The first and most important food safety assessment is toxicity testing. Acute toxicity tests, based on the confirmation of half of the lethal dose (LD50) in experimental animals, indicating the degree of toxicity. After a repeated dose toxicity test, the animal will be sacrificed and observed for clinical pathology and histopathology to check for lesions; only then can the drug undergo efficacy tests and clinical trials. Despite these strict regulations, 2-3% of new drugs approved for marketing worldwide are later withdrawn due to toxicity problems undetected during the toxicity test. This situation results in not only serious damage to the health of users, but also huge losses to the pharmaceutical industry. It indicates that the original safety assessment method is still incomplete and urgently needs improvement. Therefore, our platform may be applied as a potential method to detect drug-induced nephrotoxicity before compounds are approved for use and marketed.
On the other hand, Miox levels in renal proximal tubular epithelium increased during hyperglycemia 36 . High-glucose ambience provides an environment conductive to the binding of binding of SP-1 to the Miox promoter and thereby regulate its expression 42 . Furthermore, it has been clinically found that diabetes, sepsis and ischemia/reperfusion increase sensitivity to AKI [43][44][45] . Miox expression significantly increased in patients with type 2 diabetes mellitus and correlated with tubulointerstitial lesiony 27 . Importantly, in diabetic patients without early signs of glomerular damage, serum and urine MIOX levels have increased. Thus, this platform can  .0 mg/kg AAI (G), and 3.5 mg/kg AAI (H) were detected by H&E staining at day 13. Data were presented as mean ± SE and compared using one-way ANOVA with a post hoc-pairwise, Dunnett's test. *p < 0.05, **p < 0.01, and ***p < 0.001 were considered to be statistically significance between vehicle and 3.5 mg/kg AAI. # < 0.05, ## p < 0.01, and ### p < 0.001 were considered to be statistically significance between vehicle and 2.0 mg/kg AAI. www.nature.com/scientificreports/ not only be directly applied to a wide range of drug/food testing, but also will be further extended to different disease models, such as diabetes, sepsis or ischemia/reperfusion. In conclusion, we developed a new platform by which to quantify drug-induced nephrotoxicity by detecting the luminescence produced by the kidney-specific protein Miox-NanoLuc in mouse serum and urine. On this platform, luminescence produced by Miox-NanoLuc was increased at an early stage in the serum and urine of drug-treated transgenic mice before levels of serum BUN and creatinine increased and tubular cell injury could otherwise be noted. Moreover, this platform may be used for early detection of drug-and food-induced nephrotoxicity and as an animal model to investigate tubular cell injury. Figure 5. The ability of KIM-1, cystatin, clusterin, and trefoil factor 3 biomarkers in early detection of cisplatininduced nephrotoxicity. After were intraperitoneal injection with vehicle or cisplatin (10, 20, or 40 mg/kg), the levels of (A) urine KIM-1, (B) urine cystatin C, (C) urine clusterin, (D) urine trefoil factor 3, and (E) urine protein were measured at day 0, 1, 2, 3, and 5. Data were presented as mean ± SE and compared using oneway ANOVA with a post hoc-pairwise, Dunnett's test. *< 0.05, **p < 0.01, and ***p < 0.001 were considered to be statistically significance between vehicle and 40 mg/kg cisplatin. # < 0.05, ##p < 0.01, and ###p < 0.001 were considered to be statistically significance between vehicle and 20 mg/kg cisplatin. & < 0.05 and && p < 0.01 were considered to be statistically significance between vehicle and 10 mg/kg cisplatin.