BASIC RESEARCH www.jasn.org

Megalin Blockade with Suppresses Drug-Induced Nephrotoxicity

† ‡ Yoshihisa Hori,* Nobumasa Aoki,* Shoji Kuwahara, Michihiro Hosojima,§ Ryohei Kaseda,§ ‡| ‡| ‡| ‡ Sawako Goto, Tomomichi Iida, Shankhajit De, Hideyuki Kabasawa,§ Reika Kaneko, ‡ | Hiroyuki Aoki, Yoshinari Tanabe,¶ Hiroshi Kagamu,** Ichiei Narita, Toshiaki Kikuchi,* and ‡ Akihiko Saito

Departments of *Respiratory Medicine and Infectious Diseases, ‡Applied Molecular Medicine, §Clinical Nutrition Science, and |Clinical Nephrology and Rheumatology, Niigata University Graduate School of Medical and Dental Sciences, Niigata, Japan; †Advanced Disaster Medical and Emergency Critical Care Center and Intensive Care Units and ¶Division of Infection Control and Prevention, Niigata University Medical and Dental Hospital, Niigata, Japan; and **Division of Respiratory Medicine, Saitama Medical University International Medical Center, Hidaka-shi, Japan

ABSTRACT Nephrotoxicity induced by antimicrobial or anticancer drugs is a serious clinical problem. Megalin, an endocytic receptor expressed at the apical membranes of proximal tubules, mediates the nephrotoxicity of aminoglyco- sides and , key antimicrobials for multidrug-resistant organisms. The mechanisms underlying the neph- rotoxicity induced by , an antimicrobial for -resistant , and cisplatin, an important anticancer drug, are unknown, although the nephrotoxicity of these drugs and genta- micin, an aminoglycoside, is suppressed experimentally with cilastatin. In the clinical setting, cilastatin has been used safely to suppress dehydropeptidase-I–mediated renal metabolism of , a antimi- crobial, and thereby limit tubular injury. Here, we tested the hypothesis that cilastatin also blocks megalin- mediated uptake of vancomycin, cisplatin, colistin, and aminoglycosides, thereby limiting the nephrotoxicity of these drugs. Quartz crystal microbalance analysis showed that megalin also binds vancomycin and cisplatin and that cilastatin competes with megalin for binding to gentamicin, colistin, vancomycin, and cisplatin. In kidney- specific mosaic megalin knockout mice treated with colistin, vancomycin, or cisplatin, the megalin-replete proximal tubule epithelial cells exhibited signs of injury, whereas the megalin-deficient cells did not. Further- more, concomitant cilastatin administration suppressed colistin-induced nephrotoxicity in C57BL/6J mice. Notably, cilastatin did not inhibit the antibacterial activity of gentamicin, colistin, or vancomycin in vitro,just as cilastatin did not affect the anticancer activity of cisplatin in previous studies. In conclusion, megalin blockade with cilastatin efficiently suppresses the nephrotoxicity induced by gentamicin, colistin, vancomycin, or cis- platin. Cilastatin may be a promising agent for inhibiting various forms of drug-induced nephrotoxicity medi- ated via megalin in the clinical setting.

J Am Soc Nephrol 28: ccc–ccc, 2017. doi: 10.1681/ASN.2016060606

Received June 2, 2017. Accepted November 25, 2017. Niigata, Japan. Dr. Reika Kaneko, Department of Frontier Surgery, Graduate School of Medicine, Chiba University, Chiba, Japan. Y.H., N.A. and S.K. contributed equally to this work. Correspondence: Prof. Akihiko Saito, Department of Applied Published online ahead of print. Publication date available at Molecular Medicine, Niigata University Graduate School of www.jasn.org. Medical and Dental Sciences, 1-757 Asahimachi-dori, Chuo-ku, Niigata 951-8510, Japan. Email: [email protected] Present address: Dr. Yoshihisa Hori, Department of Internal Medi- cine, Niigata Tokamachi Hospital, Tokamachi, Japan. Dr. Ryohei Copyright © 2017 by the American Society of Nephrology Kaseda, Department of Clinical Nephrology and Rheumatology, Niigata University Graduate School of Medical and Dental Sciences,

J Am Soc Nephrol 28: ccc–ccc, 2017 ISSN : 1046-6673/2806-ccc 1 BASIC RESEARCH www.jasn.org

Drug-induced nephrotoxicity is an important factor for the causing tubuloglomerular alterations.14 Megalin also mediates development of AKI1 and CKD.2 Irrespective of the risks, the uptake of nephrotoxic drugs, such as aminoglycosides,15 however, the use of nephrotoxic is unavoidable in B,15 and colistin.16 Vancomycin and cisplatin are patients with CKD and infection, and some nephrotoxic an- not known to bind to megalin (Table 1). ticancer agents are usually not used or are used less often in Cilastatin, an inhibitor of renal dehydropeptidase-I (DHP-I), patients with CKD, despite their excellent anticancer effects. was originally introduced to clinical practice in 1987 combined Multidrug-resistant organisms called carbapenem-resistant with imipenem, a carbapenem antimicrobial. It suppresses the Enterobacteriaceae,suchasPseudomonas aeruginosa, Acinetobacter renal metabolism of imipenem and thereby, tubular injury.17 baumannii,andKlebsiella pneumoniae,areontheriseandhave Cilastatin was also found to be an inhibitor of organic anion become problematic worldwide.3 Infections caused by these bac- transporter 3.18 Cilastatin competes with the renal uptake teria lead to poor clinical outcomes and high mortality rates; of vancomycin and cisplatin and their consequent nephro- especially in sepsis, mortality rates reach as high as 50%.4 toxicity (Table 1).19,20 However, there is no evidence indicating Colistin, a cationic polypeptide , has re-emerged as a that the nephrotoxicity of vancomycin and cisplatin is caused by last resort therapy for carbapenem-resistant Enterobacteriaceae DHP-I– or organic anion transporter 3–medicated mechanisms. infections and is frequently the only agent to which these organ- Indeed, these drugs were reported to be taken up by PTECs via isms are susceptible.5 Despite its efficacy, the clinical use of co- basolateral organic cation transporters (OCTs), whereas the listin has been limited owing to its serious nephrotoxicity. Recent effects of OCT inhibitors to suppress their renal uptake and studies reported the nephrotoxicity rate in patients receiving co- nephrotoxicity are partial.9,21 The nephrotoxicity of gentamicin, listin as 40%–45% using the RIFLE criteria.6 an aminoglycoside, also reportedly competed with cilastatin,22 Vancomycin is also recognized as a nephrotoxic agent but the mechanism is also unknown (Table 1). but remains a first-line antibiotic for infections caused by Thus, we hypothesized that cilastatin may be a megalin methicillin-resistant Staphylococcus aureus. In the last several blocker that competes with megalin binding to vancomycin years, there have been growing concerns about the emergence and cisplatin as well as aminoglycosides and colistin, thereby of methicillin-resistant S. aureus strains with reduced suscepti- inhibiting the nephrotoxicity of these drugs. To test this hy- bility to vancomycin,7 and high-dose vancomycin therapy has pothesis, we have evaluated the binding of megalin to such been conducted. These attempts, however, led to a higher agents, particularly vancomycin, cisplatin, and cilastatin. We incidence of renal injury.8 Therefore, nephrotoxicity has have also analyzed the megalin-dependent nephrotoxicity of been one of the most pressing issues in infections caused these drugs using kidney-specific megalin knockout (KO) mice by resistant organisms. and examined the competitive inhibitory effect of cilastatin in Cisplatin is one of the preferred initial chemotherapeutic vitro and in vivo. agents for the treatment of various human malignancies. The most serious and common adverse event of cisplatin admin- istration is kidney injury, which occurs in approximately one RESULTS third of patients.9 Nephrotoxicity, the main dose-limiting factor of the drug, is associated with a history of diabetes Gentamicin, Colistin, Vancomycin, and Cisplatin Are mellitus and cardiovascular disease, and it worsens the Bound by Megalin, Which Is Competitively Inhibited by prognosis.10 Cilastatin Megalin, a large (approximately 600-kD) glycoprotein We firstly performed quartz crystal microbalance (QCM) member of the LDL receptor family,11 is expressed at the apical analysis to assess the direct binding of megalin with these membranes of proximal tubule epithelial cells (PTECs).12 drugs. Toassess their interactions accurately, we used megalin Megalin mediates intracellular signal transduction and protein purified from rat renal microvillar membranes by plays a pivotal role in the reabsorption of glomerular-filtered affinity chromatography using a megalin-specific mAb as de- substances.13 In a high-fat diet–induced mouse model of meta- scribed previously.23 Affinity-purified megalin, immobilized bolic syndrome, megalin mediates the proximal tubular uptake on a sensor chip, was bound directly by gentamicin and colistin of nephrotoxic substances, such as lipid-modified proteins, as reported previously15,16 as well as by vancomycin, cisplatin, and cilastatin (Figure 1, A–E, respectively). However, megalin was not bound by a comparative amount of tenofovir (Figure Table 1. Studies of nephrotoxic drugs and cilastatin on 1F), an anti-HIV drug that is taken up by PTECs via basolateral binding to megalin and renoprotection by cilastatin OCT1 to induce cellular toxicity.24 Recombinant glutathione Drugs Binding to Megalin Renoprotection by Cilastatin S-transferase (GST) protein immobilized on a sensor chip as a Gentamicin Ref. 15 Ref. 22 negative control was bound by an anti-GSTantibody but was not Colistin Ref. 16 Unknown bound by cilastatin (Figure 1G) and the other megalin-bound Vancomycin Unknown Ref. 19 drugs (data not shown). In addition, a sensor chip, on which no Cisplatin Unknown Ref. 20 proteins were immobilized, was also not bound by cilastatin Cilastatin Unknown Not applicable (Figure 1H) and the other megalin-bound drugs (data not shown).

2 Journal of the American Society of Nephrology J Am Soc Nephrol 28: ccc–ccc,2017 www.jasn.org BASIC RESEARCH

Affinity analysis showed that the Kd values for the binding of megalin with colistin and vancomycin were 0.076460.0118 and 1.15960.168 nmol/L, respectively (Supplemental Figure 1). For gentamicin, cisplatin, and cilastain, the Kd values for binding to megalin were not determined due to their limited dissolution in the reaction solution used for saturation analysis. The binding of gentamicin, colistin, vancomycin, and cisplatin to megalin was all competed by cilastatin (Figure 1, I–L, respectively). These findings indicate that all of these drugs bind specifically to neighboring sites of megalin.

Vancomycin, Cisplatin, and Colistin Induce PTEC Injury Dependent on Megalin Wenext investigated whether PTEC injury, the major pathologic feature induced by vancomycin,19 cisplatin,9 and colistin,16 is mediated via megalin using kidney-specificmosaicmegalin KO mice (apoE cre,megalinlox/lox).25 Megalin gene deletion occurs in approximately 60% of PTECs in these mice, thereby allowing a direct comparison between megalin KO and intact PTECs in the same mice. As shown in Figure 2, megalin- expressing PTECs were extensively vacuolized or destroyed in mice administered colistin, vancomycin, or cisplatin, whereas meg- alin KO PTECs were protected from injury in the same mice. In addition, kidney injury molecule-1 (KIM-1), an established kidney injury marker, was expressed concomitantly with megalin in the same PTECs of mice administered colistin, vancomycin, or cis- platin (Figure 3). Comparatively, KIM-1 was not expressed even in megalin-expressing PTECs in mice injected with saline (vehicle), and the expression of megalin and KIM-1 was not always colocal- ized in a unilateral ureteral obligation model of the mice (Figure 3). These results indicate that the colocalized expression of megalin and KIM-1 in colistin-, vancomycin-, or cisplatin-administered mosaic megalin KO mice is not an artifact but verifies the megalin-dependent nephrotoxicity induced by these drugs.

Concomitant Administration of Cilastatin Suppresses Colistin-Induced Nephrotoxicity in C57BL/6J Mice Wealso found that cilastatin suppresses colistin-induced neph- rotoxicity in C57BL/6J mice (Figure 4), as shown previously in renal injury models induced by gentamicin,22 vancomycin,19 and cisplatin.20 Concomitant administration of cilastatin with colistin to the mice suppressed colistin-induced tubulointerstitial injury, including tubular vacuolization, tubular dilation or atro- phy, brush border loss, tubular cell lysis, and cast formation (Fig- ure 4A); urinary excretion of N-acetyl-b-(D)-glucosaminidase (NAG), a PTEC injury marker (Figure 4B); and KIM-1 expres- sion in PTECs (Figure 4C). Figure 1. QCM analysis revealed that megalin is bound specifi- cally by gentamicin, colistin, vancomycin, and cisplatin and that the binding of all drugs is competed by cilastatin. In a buffer were administered at the time points indicated by the arrows containing 2 mM CaCl2, (A and I) gentamicin (final concentration for binding to the sensor chips immobilized with (A–FandI–L) in the chamber: 500 mg/ml or 1.05 mM), (B and J) colistin (30 mg/ml affinity-purified megalin, (G) recombinant GST, or (H) no protein in or 0.0107 mM), (C and K) vancomycin (125 mg/ml or 0.0863 mM), the (A–H) absence or (I–L) presence of cilastatin (500 mg/ml or 1.31 mM (D and L) cisplatin (50 mg/ml or 0.167 mM), (E, G, and H) cilastatin for I–K and 1.25 mg/ml or 3.28 mM for L). Ab, anti-GST antibody (500 mg/ml or 1.31 mM), and (F) tenofovir (150 mg/ml or 0.522 mM) (125 ng/ml).

J Am Soc Nephrol 28: ccc–ccc, 2017 Megalin Blockade with Cilastatin 3 BASIC RESEARCH www.jasn.org

approximated,26 and thus, the extensive renal tubular reabsorp- tion of colistin has been estimated. In contrast, a high plasma protein binding rate, approximately 65%–98%, was reported for cisplatin.27 Although minor amounts of glomerular filtration were speculated, the remarkable excretion of cisplatin in urine was identified due to its uptake via OCTs9 in the basolateral membranes of PTECs followed by tubular secretion.28 Hence, each drug accumulates in significant amounts by glomerular filtration or tubular secretion in the tubular lumens, where they should be encountered by megalin for reabsorption. In this study, we have shown the direct binding of megalin with colistin, gentamicin, vancomycin, cisplatin, and cilastatin using QCM analysis. This real-time analysis also revealed that prior administration of cilastatin competitively inhibited the binding of megalin with the other megalin-bound drugs. We have already shown that the receptor-ligand binding kinetics analyzed by QCM were comparable with those of other meth- – Figure 2. Megalin mediates colistin-, vancomycin-, and cisplatin- ods, such as surface plasmon resonance.29 31 A previous study induced PTEC injury shown by morphologic changes. Colistin, by Suzuki et al.16 indicated that colistin bound to megalin fi vancomycin, and cisplatin were administered to kidney-speci c with a Kd of 130.3679.0 nmol/L; however, the study had mosaic megalin KO mice to induce PTEC injury. In PAS-stained some limitations: they used megalin-containing brush border kidney sections immunostained for megalin, vacuolization or de- membrane vesicles for competitive binding assays, and thus, structive changes were observed exclusively in megalin-expressing the direct binding of megalin with colistin was not clear. In this PTECs (black arrowheads) but were not observed in megalin KO study, we used high-grade quality affinity-purified megalin, PTECs (white arrowheads) in mice administered each of these drugs. fi fi which provided more accurate binding speci city and quan- These histologic ndings were never detected in megalin-expressing fi PTECs of mice given saline (vehicle). Scale bar, 40 mm. ti cation of its binding kinetics. The assay temperature (23°C) in our study was closer to the body environment compared with that in the previous study (4°C),16 which was also likely Cilastatin Has No Influence on the Antibacterial responsible for the determination of the greater affinity of co- Activity of Colistin, Gentamicin, and Vancomycin listin for megalin. Finally, we found that cilastatin does not inhibit the antibacte- In addition, our results suggest that cilastatin suppresses rial activity of colistin, gentamicin, and vancomycin (Figure 5). colistin-induced nephrotoxicity through competition for Discs containing 10 mg cilastatin revealed no growth-inhibitory binding to megalin. Cilastatin did not show a complete reno- zone against ATCC 25922 and S. aureus ATCC protective effect in the mouse model. However, to induce ap- 29213. Discs containing colistin, gentamicin, or vancomycin parent nephrotoxicity including renal morphologic changes exhibited an adequate inhibitory ring, and discs containing and KIM-1 expression in C57BL/6J mice, which are occasion- these drugs plus cilastatin had the same inhibitory zone di- ally resistant to nephrotoxic agents,32 a high dose of colistin ameter as in the absence of cilastatin. These findings indicate (30 mg/kg per day), which is close to its subcutaneous injection– that cilastatin has no synergistic effect on bacteriostatic action mediated median lethal dose (LD50) in mice (48.6 mg/kg; in the disc diffusion method. Cilastatin is also known to not www.caymanchem.com/msdss/17584m.pdf), was needed for inhibit the anticancer effect of cisplatin.20 the experiment. In addition, the dose of cilastatin (intravenous injection–mediated LD50 in mice: 8700 mg/kg; www.merck.ca/ assets/en/pdf/products/PRIMAXIN-PM_E.pdf) used to com- DISCUSSION pete with the nephrotoxic action of colistin was relatively low: the molar ratio between colistin and cilastatin in this experi- The original target of cilastatin, DHP-I, is localized at the brush ment was approximately 1:25. border membranes of PTECs, where megalin is also located and In the clinical setting, colistimethate, the non-nephrotoxic likely to be encountered by the drug in vivo. Because the serum prodrug of colistin, is used at a concentration of 400–800 mg protein binding rates for gentamicin, colistin, and vancomycin per 60 kg/d, and 30% of the administered colistimethate is con- are 3.4%, 66%, and 34%, respectively, according to their drug verted to colistin,33 indicating that much less colistin is used information forms, significant fractions of the administered clinically than in this study. Furthermore, the applicable molar drugs should consist of free forms that are filtered by glomeruli quantity of cilastatin in clinical practice is estimated to be into the proximal tubular lumens and encounter megalin. A low approximately 61- to 123-fold greater than the amount of colis- urinary recovery rate of colistin has been reported; however, tin converted from colistimethate, leading to the expectation significant colistin clearance by glomerular filtration has been that cilastatin should be able to suppress colistin-induced

4 Journal of the American Society of Nephrology J Am Soc Nephrol 28: ccc–ccc,2017 www.jasn.org BASIC RESEARCH

variety of drug-induced nephrotoxicities and expected to be applied widely in clin- ical practice. The structure of the domain of megalin in- volved in its binding with gentamicin has been studied.34 Theauthorssolvedthenuclearmag- netic resonance structure of the tenth comple- ment type repeat in the second ligand binding cluster of human megalin to investigate its in- teraction with gentamicin. However, the over- all mechanisms by which megalin is bound by multiple drugs and how cilastatin competes with the binding of other drugs to megalin re- main unknown. Additional structural studies of the ligand binding domains of megalin are needed to elucidate the mechanisms. Aminoglycosides, colistin, vancomycin, and cisplatin are also ototoxic. Because mega- lin is also expressed in the inner ear35,36 and because single-nucleotide polymorphisms at the megalin gene might be associated with in- dividual susceptibility to cisplatin-induced ototoxicity,37 megalin blockade treatment with cilastatin may also be useful for prevent- ing the ototoxicity induced by these drugs. In conclusion, megalin blockade with cilastatin suppresses efficiently the nephrotox- icity inducedbygentamicin,colistin,vancomy- cin, and cisplatin. The clinical safety of cilastatin hasbeenestablishedby itslong-termcombined use with imipenem. Future pharmacologic development of cilastatin alone as a megalin blocker will be expected for its broad clinical application to inhibit the nephrotoxicity me- diated by these drugs and probably some other compounds, such as cyclosporin.38

Figure 3. Megalin mediates colistin-, vancomycin-, and cisplatin-induced PTEC injury shown fl by KIM-1 expression. Double immuno uorescenceofmegalinandKIM-1inthekidney CONCISE METHODS sections showed that KIM-1 expression is only seen in megalin-expressing PTECs (arrow- heads) and is not seen in megalin KO PTECs (arrows) in the kidney-specificmosaicmegalin KO mice administered colistin, vancomycin, or cisplatin. Vehicle administration did not in- Materials duce KIM-1 expression, even in megalin-expressing PTECs (arrowheads). In the unilateral Colistin, gentamycin, and vancomycinwere pur- ureteral obligation (UUO) model of the mice, KIM-1 was expressed in not only megalin- chased from Sigma-Aldrich. Cilastatin was pur- expressing PTECs but also, megalin KO PTECs (arrows). Scale bar, 40 mm. chased from Wako Pure Chemical Industries, Ltd. and Sigma-Aldrich. Cisplatin was obtained from Wako Pure Chemical Industries, Ltd. and nephrotoxicity in the clinical setting. In addition, the maximum NIPPON KAYAKU Co. Ltd., and tenofovir was from Wako Pure dosage of imipenem/cilastatin in current clinical use is limited by Chemical Industries, Ltd. imipenem-mediated toxicity (intravenous injection–mediated LD50 in mice: 1500 mg/kg; www.merck.ca/assets/en/pdf/products/ QCM PRIMAXIN-PM_E.pdf); therefore, cilastatin would be used Binding of the drugs to megalin was examined using a highly sensitive at a higher dose if its monotherapy was available. Cilastatin 27-MHz QCM instrument (AFFINIX Q, QCM2000; ULVAC) as de- also reportedly inhibits the nephrotoxicity of vancomycin, scribed previously.30 Briefly, QCM sensor chips were activated with a cisplatin, and gentamicin.19,20,22 These findings thus indicate 1:1 mixture of 100 mg/ml N-hydroxysuccinimide and 100 mg/ml that cilastatin is a promising agent for the management of a 1-(3-dimethylaminopropyl)-3-ethylcarbodiimide hydrochloride in

J Am Soc Nephrol 28: ccc–ccc, 2017 Megalin Blockade with Cilastatin 5 BASIC RESEARCH www.jasn.org

water. Immunoaffinity-purified rat megalin23 was then immobilized on the sensor chips at a concentration of 65 mg/ml in buffer B (10 mM HEPES, pH 7.4, 150 mM NaCl, and 2 mM

CaCl2). The sensor chips were soaked in the in- cubation chamber at 23°C in 8 ml buffer B. Each drug dissolved in buffer B was added to the chamber to test binding to megalin. As a nega- tive control, recombinant GST protein was pre- pared by using a prokaryotic expression system with the pGEX-6p-1 vector (GE Healthcare) ac- cording to the manufacturer’s instructions and immobilized on the sensor chip. An anti-GST antibody, used as a ligand of GST, was pur- chased from GE Healthcare. Cilastatin was used for competition with the other drugs for binding to megalin: cilastatin was first added to the chamber and reached the steady state of its binding to megalin, and then, each of the drugs was injected into the chamber to assess compet- itive binding inhibition by cilastatin. The reso- nance frequency of QCM at equilibrium was defined as the zero position. The stability and drift of the 27-MHz QCM frequency in solution were both 63 Hz. Receptor-ligand binding ki- netics were analyzed by the saturation method using AFFINIX Q8 (QCM2012; ULVAC) as de- scribed previously.29,30

Treatment of Kidney-Specific Megalin KO Mice with Nephrotoxic Drugs Male kidney-specific megalin KO mice (apoE cre,megalinlox/lox) established by Thomas Willnow (Max Delbrück Center for Molecular Medicine)25 were used in this study. The mice were housed in a temperature-controlled room and allowed free access to a standard diet and tap water. To induce the kidney injury model, 11-week-old mice received a subcutaneous in- Figure 4. Cilastatin suppresses colistin-induced nephrotoxicity in C57BL/6J mice. (A) jection of colistin (30 mg/kg body wt) once daily Representatively, tubular vacuolization, tubular dilation or atrophy, brush border loss, for 4 days or an intraperitoneal injection of van- tubular cell lysis, and cast formation are observed in PAS-stained kidney sections of comycin (400 mg/kg body wt) once daily for colistin-administered C57BL/6J mice, but these findings were ameliorated drastically 4 days or cisplatin (20 mg/kg body wt) once for by the concomitant administration of cilastatin. Semiquantitative morphometric 1 day. At 24 hours after the final treatment with analysis of tubular damage revealed that cilastatin administration significantly sup- colistin or vancomycin and 72 hours after cis- pressed colistin-induced nephrotoxicity; n=8, n=10, and n=10 for the vehicle, colistin, platin treatment, the animals were euthanized m , and colistin and cilastatin groups, respectively. Scale bar, 40 m. *P 0.05; under intraperitoneal pentobarbital sodium an- ***P,0.001. (B) Colistin-induced increase of urinary N-acetyl-b-(D)-glucosaminidase esthesia (50 mg/kg body wt) or by cervical dis- excretion in C57BL/6J mice was significantly suppressed by the concomitant ad- location. Then, the right kidneys were removed ministration of cilastatin. A significant interaction was found by two-way ANOVA – (P,0.01), and subsequent Bonferroni post hoc analysis identified significant differ- and processed for periodic acid Schiff (PAS) and fl – ences between the indicated points (n=5 each). **P,0.01; ***P,0.001. (C) Repre- immuno uorescence staining. Vehicle saline sentative photomicrographs showing that the concomitant administration of cilastatin injected male kidney-specific megalin KO mice suppresses colistin-induced KIM-1 expression in PTECs of C57BL/6J mice. G, glo- were used as negative controls. All animal ex- merulus. Scale bar, 40 mm. periments in this study were on the basis of the

6 Journal of the American Society of Nephrology J Am Soc Nephrol 28: ccc–ccc,2017 www.jasn.org BASIC RESEARCH

serum in PBS and an antimegalin antiserum (1:2000)14 in PBS containing 0.1% Triton X-100 at 4°C overnight. Immunopositive reactions were visualized by using Texas red anti-rabbit IgG (Vector Laboratories) in the dark. Nonimmune rabbit or goat serum was used as negative con- trol without using a primary antibody. After staining, the samples were mounted with the SlowFade Gold Antifade Reagent (Life Technol- ogies), and images were captured by using a Figure 5. Combined disc tests show that cilastatin has no effect on the antimicrobial fluorescence microscope (BZ-9000; Keyence). activity of gentamicin, colistin, and vancomycin. Discs containing the drugs were placed on Mueller–Hinton agar plates that had been inoculated with (A and B) E. coli ATCC 25922 or (C) S. aureus ATCC 29213. The contents of the numbered discs were Competition of Colistin-Induced as follows: 1, cilastatin (10 mg); 2, gentamicin (10 mg); 3, gentamicin (10 mg) and Nephrotoxicity with Cilastatin in cilastatin (10 mg); 4, colistin (10 mg); 5, colistin (10 mg) and cilastatin (10 mg); 6, C57BL/6J Mice vancomycin (30 mg); and 7, vancomycin (30 mg) and cilastatin (10 mg). The discs Male C57BL/6J mice (Charles Rivers Laborato- containing cilastatin represent no inhibitory zone, and the combined discs revealed no ries International) at 11 weeks of age were di- antagonistic effects. Scale bar, 10 mm. vided into three groups. The first group (n=10) was treated subcutaneously with colistin (30 mg/kg body wt) once daily for 4 days, the National Institutes of Health Guide for the Care and Use of Labora- second group (n=10) was treated subcutaneously with colistin (30 mg/kg tory Animals and approved by the Institutional Animal Care and Use body wt) and cilastatin (100 mg/kg body wt) once daily for 4 days, and the Committee of Niigata University. third group (n=8) was treated subcutaneously with vehicle saline for 4 days. At 24 hours after the final treatment, the mice were euthanized Unilateral Ureteral Obligation in Kidney-Specific under intraperitoneal pentobarbital sodium anesthesia (50 mg/kg body Megalin KO Mice wt). The right kidneys were dissected and processed for semiquantitative Male kidney-specific megalin KO mice at 11 weeks of age were anes- morphometric analysis and KIM-1 immunostaining. Urinary NAG was thetized by an intraperitoneal injection of pentobarbital sodium measured by using the colorimetric method. (50 mg/kg body wt). The right ureter was exposed via flank incisions and ligated with nylon threads, and the incisions were closed. After Semiquantitative Morphometric Analysis of Tubular 5 days, the mice were euthanized by cervical dislocation, and the right Injury kidney was removed and processed for immunofluorescence staining. Semiquantitative renal morphometric analysis was performed using 2-mm-thick paraffin sections stained with PAS. Tubular damage was Preparation of Kidney Sections for PAS and scored by evaluating the percentage of injured tubules showing dila- Immunofluorescence Staining tion or atrophy, brush border loss, cell lysis, and cast formation in the After removal of the kidneys, 3-mm-thick horizontal cross-sections corticomedullary junction and outer medulla. Lesions were graded containing the hila were obtained immediately and fixed in a 4% on a scale from zero to four: zero, none; one, ,10%; two, $10% and paraformaldehyde phosphate buffer for 72 hours at room tempera- ,25%; three, $25% and ,50%; and four, .50% of 15 randomly ture. After fixation, the tissues were dehydrated in a graded ethanol selected viewing fields (original magnification, 3400) each section series from 70% to 100%, cleared in xylene, and embedded in paraffin. per kidney. Morphometry was conducted in a blinded fashion by an From each tissue sample, 2-mm-thick sections were cut using a mi- experienced pathologist. crotome (REM-710; Yamato Kohki Industrial Co., Ltd.) and stained with PAS. In addition, 4-mm-thick sections were used for immuno- Renal Immunohistochemistry for KIM-1 fluorescence staining. For the detection of KIM-1 in kidney tissues, 4-mm-thick paraffin sections were stained with the avidin-biotin-peroxidase complex Immunofluorescence Staining method. A goat anti-mouse KIM-1 antibody (1:800; AF1817; R&D Sys- Kidney tissue sections were deparaffinizedbyimmersioninxyleneand tems) was used as the primary antibody. After deparaffinization, anti- rehydrated through a graded ethanol series. Antigen retrieval was ac- gen retrieval was performed by using citrate buffer, and endogenous complished by heating the slides in a microwave in 10 mM citrate peroxidase activity was eliminated with 3% H2O2. Nonspecificbinding buffer, pH 6.0. After incubation with 5% rabbit serum in PBS to block was blocked with 5% rabbit serum, and endogenous biotin binding nonspecific binding, the sections were incubated overnight at 4°C was blocked with an Avidin/Biotin Blocking Kit (Vector Laboratories). with a goat anti-mouse KIM-1 antibody (1:500; AF1817; R&D Sys- Finally, immunoreactivity was developed with the DAB Substrate- tems) in PBS containing 0.1% Triton X-100, and immunopositive Chromogen System (Dako), and the sections were counterstained with reactions were visualized by using fluorescein anti-goat IgG (Vector Mayer hematoxylin. Negative control staining was always performed in Laboratories). Sequentially, the sections were incubated with 5% goat parallel with PBS instead of the primary antibody.

J Am Soc Nephrol 28: ccc–ccc, 2017 Megalin Blockade with Cilastatin 7 BASIC RESEARCH www.jasn.org

Antibacterial Activity Assay 4. Tam VH, Rogers CA, Chang KT, Weston JS, Caeiro JP, Garey KW: Im- The agar disc diffusion method was performed according to Clinical pact of multidrug-resistant bacteremia on and Laboratory Standards Institute guidelines. Briefly, a test organism patient outcomes. Antimicrob Agents Chemother 54: 3717–3722, 2010 (E. coli ATCC 25922 or S. aureus ATCC 25923) was cultured in Mueller– 5. Li J, Nation RL, Turnidge JD, Milne RW, Coulthard K, Rayner CR, Paterson Hinton broth and adjusted to a McFarland Standard of 0.5. Bacterial DL: Colistin: The re-emerging antibiotic for multidrug-resistant Gram- suspensions were streaked onto Mueller–Hinton agar plates. Paper discs negative bacterial infections. Lancet Infect Dis 6: 589–601, 2006 containing colistin (10 mg), gentamycin (10 mg), vancomycin (30 mg), 6. Pogue JM, Lee J, Marchaim D, Yee V, Zhao JJ, Chopra T, Lephart P, and cilastatin (10 mg) were deposited on the inoculated plates, and the Kaye KS: Incidence of and risk factors for colistin-associated nephro- toxicity in a large academic health system. Clin Infect Dis 53: 879–884, plates were incubated at 37°C overnight. The plates were subsequently 2011 inspected for bacterial growth, and the zone of growth inhibition sur- 7. Sader HS, Fey PD, Limaye AP, Madinger N, Pankey G, Rahal J, Rybak MJ, rounding the paper disc was measured. Snydman DR, Steed LL, Waites K, Jones RN: Evaluation of vancomycin and potency trends (MIC creep) against methicillin-resistant Staphylococcus aureus isolates collected in nine U.S. medical centers Statistical Analyses from 2002 to 2006. Antimicrob Agents Chemother 53: 4127–4132, In the animal experiments, semiquantitative morphometric data 2009 were analyzed by Kruskal–Wallis one-way ANOVA by ranks and 8. Hidayat LK, Hsu DI, Quist R, Shriner KA, Wong-Beringer A: High-dose comparisons between each group were made by Dunn Multiple vancomycin therapy for methicillin-resistant Staphylococcus aureus fi – Comparison Test. Urinary NAG excretion data were analyzed using infections: Ef cacy and toxicity. Arch Intern Med 166: 2138 2144, 2006 9. Pabla N, Dong Z: Cisplatin nephrotoxicity: Mechanisms and renoprotective two-way ANOVA and comparisons between each group were made strategies. Kidney Int 73: 994–1007, 2008 by Bonferroni post hoc analysis. All statistical analyses were carried 10. Mizuno T, Ishikawa K, Sato W, Koike T, Kushida M, Miyagawa Y, out using Predictive Analytics SoftWare Statistics 18 (SPSS, Inc.) and Yamada K, Hirata S, Imai E, Noda Y: The risk factors of severe acute GraphPad Prism5 (MDF). The level of significance was P,0.05. kidney injury induced by cisplatin. Oncology 85: 364–369, 2013 11. Saito A, Pietromonaco S, Loo AK, Farquhar MG: Complete cloning and sequencing of rat gp330/“megalin,” a distinctive member of the low density lipoprotein receptor gene family. Proc Natl Acad Sci USA 91: – ACKNOWLEDGMENTS 9725 9729, 1994 12. Christensen EI, Nielsen S, Moestrup SK, Borre C, Maunsbach AB, de Heer E, Ronco P, Hammond TG, Verroust P: Segmental distribution of We thank Prof. Thomas Willnow for his gift of the experimental the endocytosis receptor gp330 in renal proximal tubules. Eur J Cell material described in Concise Methods and Ms. Nanako Sugita, Biol 66: 349–364, 1995 Mr. Ryota Kobayashi, and Mr. Koichi Komochi for their technical 13. Saito A, Sato H, Iino N, Takeda T: Molecular mechanisms of receptor- mediated endocytosis in the renal proximal tubular epithelium. J Biomed assistance. Biotechnol 2010: 403272, 2010 This work was supported by Japanese Ministry of Education, 14. Kuwahara S, Hosojima M, Kaneko R, Aoki H, Nakano D, Sasagawa T, Culture, Sports, Science and Technology for ScientificResearch Kabasawa H, Kaseda R, Yasukawa R, Ishikawa T, Suzuki A, Sato H, grants-in-aid 25870245 (to N.A.) and 26461216 (to A.S.); a Health and Kageyama S, Tanaka T, Kitamura N, Narita I, Komatsu M, Nishiyama A, Labour Sciences Research Grant from the Ministry of Health, Labour Saito A: Megalin-mediated tubuloglomerular alterations in high-fat diet-induced kidney disease. JAmSocNephrol27: 1996–2008, 2016 and Welfare;and the Practical Research Project for Renal Diseases from 15. Moestrup SK, Cui S, Vorum H, Bregengård C, Bjørn SE, Norris K, Gliemann the Japan Agency for Medical Research and Development. An abstract J, Christensen EI: Evidence that epithelial glycoprotein 330/megalin me- of this researchwas presented at the meeting of the American Society of diates uptake of polybasic drugs. J Clin Invest 96: 1404–1413, 1995 Nephrology in Chicago, November 17–20, 2016. 16. Suzuki T, Yamaguchi H, Ogura J, Kobayashi M, Yamada T, Iseki K: Megalin contributes to kidney accumulation and nephrotoxicity of colistin. Antimicrob Agents Chemother 57: 6319–6324, 2013 17. Kahan FM, Kropp H, Sundelof JG, Birnbaum J: Thienamycin: Devel- DISCLOSURES opment of imipenen-cilastatin. J Antimicrob Chemother 12[Suppl D]: A.S. received a research grant from Ajinomoto Pharmaceuticals Co. Ltd. (EA 1–35, 1983 Pharma Co. Ltd.). 18. Takeda M, Narikawa S, Hosoyamada M, Cha SH, Sekine T, Endou H: Characterization of organic anion transport inhibitors using cells stably expressing human organic anion transporters. Eur J Pharmacol 419: 113–120, 2001 REFERENCES 19. Toyoguchi T, Takahashi S, Hosoya J, Nakagawa Y, Watanabe H: Nephrotoxicity of vancomycin and drug interaction study with cilastatin 1. Nash K, Hafeez A, Hou S: Hospital-acquired renal insufficiency. Am J in rabbits. Antimicrob Agents Chemother 41: 1985–1990, 1997 Kidney Dis 39: 930–936, 2002 20. Humanes B, Lazaro A, Camano S, Moreno-Gordaliza E, Lazaro JA, 2. Venkatachalam MA, Weinberg JM, Kriz W, Bidani AK: Failed tubule Blanco-Codesido M, Lara JM, Ortiz A, Gomez-Gomez MM, Martín- recovery, AKI-CKD transition, and kidney disease progression. JAm Vasallo P, Tejedor A: Cilastatin protects against cisplatin-induced Soc Nephrol 26: 1765–1776, 2015 nephrotoxicity without compromising its anticancer efficiency in rats. 3. Talbot GH, Bradley J, Edwards JE Jr., Gilbert D, Scheld M, Bartlett JG; Kidney Int 82: 652–663, 2012 Antimicrobial Availability Task Force of the Infectious Diseases Society 21. Nakamura T, Takano M, Yasuhara M, Inui K: In-vivo clearance study of of America: Bad bugs need drugs: An update on the development vancomycin in rats. J Pharm Pharmacol 48: 1197–1200, 1996 pipeline from the antimicrobial availability task force of the Infectious 22. Jado JC, Humanes B, Lopez-Parra V, Camano S, Lara JM, Cercenado Diseases Society of America. Clin Infect Dis 42: 657–668, 2006 E, Tejedor A, Lázaro A: Effects of cilastatin on gentamicin-induced

8 Journal of the American Society of Nephrology J Am Soc Nephrol 28: ccc–ccc,2017 www.jasn.org BASIC RESEARCH

renal damage: In vitro and in vivo evidence. NephrolDialTransplant Megalin knockout mice as an animal model of low molecular weight 29: iii90, 2014 proteinuria. Am J Pathol 155: 1361–1370, 1999 23. Orlando RA, Kerjaschki D, Kurihara H, Biemesderfer D, Farquhar MG: 32. Tanaka-Kagawa T, Suzuki M, Naganuma A, Yamanaka N, Imura N: gp330 associates with a 44-kDa protein in the rat kidney to form the Strain difference in sensitivity of mice to renal toxicity of inorganic Heymann nephritis antigenic complex. Proc Natl Acad Sci U S A 89: mercury. J Pharmacol Exp Ther 285: 335–341, 1998 6698–6702, 1992 33. Couet W, Grégoire N, Gobin P, Saulnier PJ, Frasca D, Marchand S, 24. Kohler JJ, Hosseini SH, Green E, Abuin A, Ludaway T, Russ R, Santoianni Mimoz O: Pharmacokinetics of colistin and colistimethate sodium R, Lewis W: Tenofovir renal proximal tubular toxicity is regulated by OAT1 after a single 80-mg intravenous dose of CMS in young healthy vol- and MRP4 transporters. Lab Invest 91: 852–858, 2011 unteers. Clin Pharmacol Ther 89: 875–879, 2011 25. Leheste JR, Melsen F, Wellner M, Jansen P, Schlichting U, Renner- 34. Dagil R, O’Shea C, Nykjær A, Bonvin AM, Kragelund BB: Gentamicin Müller I, Andreassen TT, Wolf E, Bachmann S, Nykjaer A, Willnow TE: binds to the megalin receptor as a competitive inhibitor using the Hypocalcemia and osteopathy in mice with kidney-specificmegalin common ligand binding motif of complement type repeats: Insight gene defect. FASEB J 17: 247–249, 2003 from the nmr structure of the 10th complement type repeat domain 26. Li J, Milne RW, Nation RL, Turnidge JD, Smeaton TC, Coulthard K: Use alone and in complex with gentamicin. JBiolChem288: 4424–4435, of high-performance liquid chromatography to study the pharmacoki- 2013 netics of colistin sulfate in rats following intravenous administration. 35. Mizuta K, Saito A, Watanabe T, Nagura M, Arakawa M, Shimizu F, Antimicrob Agents Chemother 47: 1766–1770, 2003 Hoshino T: Ultrastructural localization of megalin in the rat cochlear 27. Ferraro G, Massai L, Messori L, Merlino A: Cisplatin binding to human serum duct. Hear Res 129: 83–91, 1999 albumin: A structural study. Chem Commun (Camb) 51: 9436–9439, 2015 36. Arai M, Mizuta K, Saito A, Hashimoto Y, Iwasaki S, Watanabe T, Mineta 28. Harland SJ, Newell DR, Siddik ZH, Chadwick R, Calvert AH, Harrap KR: H: Localization of megalin in rat vestibular dark cells and endolymphatic Pharmacokinetics of cis-diammine-1,1-cyclobutane dicarboxylate platinum(II) sac epithelial cells. Acta Otolaryngol 128: 627–633, 2008 in patients with normal and impaired renal function. Cancer Res 44: 37. Riedemann L, Lanvers C, Deuster D, Peters U, Boos J, Jürgens H, am 1693–1697, 1984 Zehnhoff-Dinnesen A: Megalin genetic polymorphisms and individual 29. Hama H, Saito A, Takeda T, Tanuma A, Xie Y, Sato K, Kazama JJ, Gejyo sensitivity to the ototoxic effect of cisplatin. Pharmacogenomics J 8: F: Evidence indicating that renal tubular metabolism of leptin is me- 23–28, 2008 diated by megalin but not by the leptin receptors. Endocrinology 145: 38. Carmellini M, Frosini F, Filipponi F, Boggi U, Mosca F: Effect of cilastatin 3935–3940, 2004 on cyclosporine-induced acute nephrotoxicity in kidney transplant re- 30. Kaseda R, Iino N, Hosojima M, Takeda T, Hosaka K, Kobayashi A, cipients. Transplantation 64: 164–166, 1997 Yamamoto K, Suzuki A, Kasai A, Suzuki Y, Gejyo F, Saito A: Megalin- mediated endocytosis of cystatin C in proximal tubule cells. Biochem Biophys Res Commun 357: 1130–1134, 2007 31. Leheste JR, Rolinski B, Vorum H, Hilpert J, Nykjaer A, Jacobsen C, This article contains supplemental material online at http://jasn.asnjournals. Aucouturier P, Moskaug JO, Otto A, Christensen EI, Willnow TE: org/lookup/suppl/doi:10.1681/ASN.2016060606/-/DCSupplemental.

J Am Soc Nephrol 28: ccc–ccc, 2017 Megalin Blockade with Cilastatin 9