Integrin-Linked Kinase Deficiency in Collecting Duct Principal Cell

BASIC RESEARCH www.jasn.org

Integrin-Linked Kinase Deﬁciency in Collecting Duct Principal Cell Promotes Necroptosis of Principal Cell and Contributes to Kidney Inﬂammation and Fibrosis

Ming Huang,1,2 Shuai Zhu,1,2 Huihui Huang,2 Jinzhao He,1,2 Kenji Tsuji,2 William W. Jin,2 Dongping Xie,3 Onju Ham,2 Diane E. Capen,2 Weining Lu,4 Teodor G. Paunescu, 2 Baoxue Yang,1 and Hua A. Jenny Lu2

1Department of Pharmacology, School of Basic Medical Sciences, Peking University Health Science Center, Beijing, China; 2Center for Systems Biology, Program in Membrane Biology, Division of Nephrology, Department of Medicine, Massachusetts General Hospital and Harvard Medical School, Boston, Massachusetts; 3Department of Physiology, Tongji University School of Medicine, Shanghai, China; and 4Renal Section, Departments of Medicine, and Pathology & Laboratory Medicine, Boston University Medical Center, Boston, Massachusetts

ABSTRACT Background Necroptosis is a newly discovered cell death pathway that plays a critical role in AKI. The involvement of integrin-linked kinase (ILK) in necroptosis has not been studied. BASIC RESEARCH Methods We performed experiments in mice with an Ilk deletion in collecting duct (CD) principal cells (PCs), and cultured tubular epithelial cells treated with an ILK inhibitor or ILK siRNA knockdown. Results Ilk deletion in CD PCs resulted in acute tubular injury and early mortality in mice. Progressive interstitial fibrosis and inflammation associated with the activation of the canonical TGF-b signaling cas- cade were detected in the kidneys of the mice lacking ILK in the CD PCs. In contrast to the minimal apoptosis detected in the animals’ injured CDs, widespread necroptosis was present in ILK-deficient PCs, characterized by cell swelling, deformed mitochondria, and rupture of plasma membrane. In addition, ILK deficiency resulted in increased expression and activation of necroptotic proteins MLKL and RIPK3, and membrane translocation of MLKL in CD PCs. ILK inhibition and siRNA knockdown reduced cell survival in cultured tubular cells, concomitant with increased membrane accumulation of MLKL and/or phospho- MLKL. Administration of a necroptosis inhibitor, necrostatin-1, blocked cell death in vitro and significantly attenuated inflammation, interstitial fibrosis, and renal failure in ILK-deficient mice. Conclusions The study demonstrates the critical involvement of ILK in necroptosis through modulation of the RIPK3 and MLKL pathway and highlights the contribution of CD PC injury to the development of inflammation and interstitial fibrosis of the kidney.

JASN 30: 2073–2090, 2019. doi: https://doi.org/10.1681/ASN.2018111162

Received November 26, 2018. Accepted July 15, 2019. CKD affects approximately 10% of the world’s M. H. and S. Z. contributed equally to this work. population and places a signiﬁcant burden on Published online ahead of print. Publication date available at economy and health care worldwide.1,2 Effective www.jasn.org. therapies to prevent or halt CKD progression are Correspondence: Dr. Hua A. Jenny Lu, Program in Membrane lacking, largely due to limited understanding of Biology, Department of Medicine, Massachusetts General Hos- thepathologicandmolecularbasisofthedisease. pital, 185 Cambridge Street, CPZN 8150, Boston, MA 02114, or Dr. Baoxue Yang, Department of Pharmacology, School of Basic It is increasingly recognized that AKI is a major Medical Sciences, Peking University Health Science Center, 38 contributor to the development of CKD.3 There- Xueyuan Road, Beijing 100191, China. E-mail: Lu.Hua@mgh. fore, it is critical to understand the injury and re- harvard.edu or [email protected] covery mechanisms of renal tubular cells and their Copyright © 2019 by the American Society of Nephrology

JASN 30: 2073–2090, 2019 ISSN : 1046-6673/3011-2073 2073 BASIC RESEARCH www.jasn.org roles in mediating inflammatory response and fibrosis in the Significance Statement kidney. There are two main types of cell death that occur during Necroptosis has emerged as an important cell death pathway that kidney tubular injury: apoptosis and necrosis. Apoptotic cells contributes to inflammation and injury of many organs, including the are frequently detected in the renal proximal tubule, distal kidney. The mechanisms underlying necroptosis are not well un- derstood. The authors have identified a previously unrecognized 3 tubule, and loop of Henle in various AKI models. Tubular important role of integrin-linked kinase (ILK) in mediating nec- cell apoptosis was once considered to be a key form of cell roptosis in collecting duct epithelial cell using genetically en- death leading to CKD.4 However, this point of view was re- gineered mice to lack Ilk in the collecting duct principal cells of the cently challenged because inhibition of apoptosis, e.g., by cas- kidney. These Ilk-knockout mice develop acute tubular injury, in- fi fl pase inhibitors, failed to alleviate renal injury.5 Therefore, an terstitial brosis and in ammation in the kidneys. Treating both the ILK inhibited cultured cells and ILK-deficient mice with a necroptosis alternative pathology has to be sought. In recent years, break- inhibitor, necrostatin-1, reduced the harmful effects associated with through studies show that necrosis can occur in a highly reg- the loss of ILK. The study shows that ILK plays an important role in ulated manner through complex molecular interactions, and regulating necroptosis in kidney tubular cells. it plays a critical role in mediating tissue injury and interstitial fibrosis.6,7 significant necroptosis of CD PCs instead. ILK deficiency in Necroptosis is the most-characterized pathway of regulated PCs activates necroptotic signaling and promotes inflammation necrosis in higher eukaryotic cells.6 Necroptosis is activated via and interstitial fibrosis in the kidney. More importantly, blocking the formation of the necrosome, which includes receptor in- necroptosis using necrostatin-1 (Nec-1), a chemical inhibitor of teracting protein kinase 1 (RIPK1), RIPK3, and mixed lineage RIPK1, attenuates CD injury and fibrosis in Ilk KO mice. There- kinase domain-like pseudokinase (MLKL). The homotypic fore, our data support a previously unrecognized, important func- interaction of RIPK3 and RIPK1 drives RIPK3 phosphorylation of ILK in mediating necroptosis in CD epithelial cells. tion, which in turn recruits and phosphorylates MLKL.8 Phos- phorylated MLKL oligomerizes and translocates to the plasma membrane, eventually resulting in membrane rupture.9 When METHODS cells undergo necroptosis, they present with organelle and cell swelling, permeabilization of the plasma membrane, and spill- See Supplemental Methods for a detailed description. ing of intracellular contents.10 Recent studies indicate necroptosis is an important player in some high-effect diseases, such Experimental Animals as myocardial ischemia and reperfusion injury (IRI), sepsis, All animal experiments were approved by the Massachusetts and intestinal inflammation.11–13 Furthermore, it has been General Hospital (MGH) Subcommittee on Research Animal shown to contribute to the pathogenesis of renal IRI and Care, in compliance with the National Institutes of Health AKIs induced by nephrotoxic agents such as cisplatin and io- (NIH) Guide for the Care and Use of Laboratory Animals.All dinated contrast.14–16 The involvement of many signaling mice were on the C57BL/6J background. B6N;129-Ilktm1Star/J pathways that are critically associated with cell adhesion and mice carrying loxP sites have been previously described and survival, such as integrin and its downstream integrin-linked are available from the Jackson laboratory (stock number kinase (ILK), has not yet been studied in necroptosis. 023310).29 Transgenic mice harboring aquaporin 2 (Aqp2) ILK is a critical scaffold protein located in focal adhesions. Cre recombinase (Aqp2-Cre+) were generated by Dr. Günther Through interacting with the cytoplasmic domain of Schütz’s group (Heidelberg, Germany)30 and obtained from b-integrins, ILK transduces integrin signaling to the interior Dr. Wenzheng Zhang (Albany, NY).31 Homozygous, floxed Ilk fl fl of the cell and mediates diverse cellular processes, including mice (Ilk / ) were crossed with Aqp2-Cre+ mice to generate fl fl cell survival, proliferation, adhesion, differentiation, and migra- PC Ilk KO mice (Ilk / ; Aqp2-Cre+). The littermates without fl fl tion.17 In vivo studies have suggested that ILK plays a critical role Aqp2-Cre (Ilk / ; Aqp2-Cre2) were used as control. in developing and maintaining the structure and function of For the Nec-1 (Sigma-Aldrich, St. Louis, MO) treatment fl fl many organs, including the kidney.18–20 For example, inactivation experiment, 2-week-old wild-type and Ilk / ; Aqp2-Cre+ mice of ILK in mouse kidney podocytes leads to podocyte damage, were given 1.65 mg/kg Nec-1 or PBS containing 1% DMSO progressive proteinuria, glomerulosclerosis, and severe tubuloin- through intraperitoneal injection every 2 days for 2 weeks. terstitial fibrosis.21–23 Aberrant expression of ILK is associated Mice were euthanized at the age of 4 weeks. with renal tubulointerstitial fibrosis, renal cell carcinoma, diabetic glomerulopathy, and congenital nephrotic syndrome.24–28 Kidney Tissue Preparation To further investigate the potential function of ILK in collecting Kidney tissues were collected as previously described.32 Briefly, duct (CD) principal cells (PCs), we generated knockout (KO) mice were anesthetized using isoflurane (3% inhalant). The right mice with Ilk deletion in PCs. We found that deletion of ILK kidneywassnapfrozeninliquidnitrogenandstoredat280°C for in PCs led to profound tubular injury and renal failure, unex- protein and RNA analysis. The left kidney was fixed by perfusion pectedly without significant apoptosis of tubular cells. Further through the left ventricle with paraformaldehyde-lysine- investigation demonstrated that deleting ILK in PCs induces periodate fixative. After postfixation and extensive washing

2074 JASN JASN 30: 2073–2090, 2019 www.jasn.org BASIC RESEARCH

A B C 4 w Body weight Survival rate Figure 1. Renal failure and kidney tubular 25 100 injury are present in PC Ilk KO mice. (A) 20 WT 80 KO WT Representative images of wild-type (WT) and 15 60 KO PC Ilk KO mice at 4 weeks old. (B) Body # 40 10 * ,

% survival weight. *P 0.05 versus 4-week-old wild type; 5 20 #P,0.05 versus 8-week-old wild-type mice. (C) Body weight (g) 0 0 – 2w 4w 8w 0 2 4 6 8 10 12 w Survival rate was shown by a Kaplan Meier curve. n=8–10 mice per group. (D and E) Ele- vated serum creatinine (Scr) and BUN in PC D E F – , , Scr BUN Ilk KO mice. n=6 10. **P 0.01, ***P 0.001 ## && & versus 4-week-old wild-type mice; P,0.01, 1.5 ## 150 ### Urinary NGAL ###P,0.001 versus 8-week-old wild-type mice; 4w 8w &P,0.05, &&P,0.01 versus 4-week-old KO 1.0 100 *** WT KO WT KO mice. (F) Increased urinary excretion of NGAL ** (kDa)

(mg/dL) 0.5 50 in PC Ilk KO mice as shown by immuno- 22 BUN (mg/dL) blotting. (G) Representative images of H&E Serum creatinine 0.0 0 staining of kidney cortex and outer medulla (OM) of wild-type and KO mice. Tubular di- 4w WT 4w KO 8w WT8w KO 4w WT 4w KO 8w WT 8w KO lation and microcysts were present in Ilk KO kidney (arrows). Detached and dead tubular G H 4w WT 4w KO 8w KO cells were observed within tubular lumen (arrowheads). Scale bar, 50 mm. (H) Tubular injury score was calculated based on the & percentage of damaged tubules. The degree ### 4 *** of injury was graded blindly in ten ran-

Cortex 3 domly chosen fields as follows: 0, normal; 1, , – – . 2 10%; 2, 11% 25%; 3, 26% 75%; 4, 75%. ***P,0.001 versus 4-week-old wild-type 1 mice; ###P,0.001 versus 8-week-old wild-type 0 & Tubular injury score mice; P,0.05 versus 4-week-old KO mice. n=4–7. Original magnification, 2003.(I)Immu- OM 4w WT4w KO 8w WT8w KO nofluorescence staining using an Airyscan confocal microscope confirmed deletion of ILK in PCs in KO kidney. ILK (red) was expressed on the basal membrane that was costained with I J wheat germ agglutinin (WGA)–FITC (green) in WT KO WT KO both PCs (stained purple with anti-AQP2 antibody) and ICs (negative for AQP2 staining) in wild-type kidney, whereas membrane expres- fi ILK sion of ILK (red) was signi cantly reduced or ILK even absent in Ilk KO PCs. Arrows indicate ILK expression in WGA-positive basal membranes in wild-type and KO CDs, respectively. Scale bar, 5 mm. (J) Immunofluorescence staining re-

/ vealed reduced ILK expression (red) in PCs in the KO kidney. PC-speciﬁcmarkerAQP2 AQP2 / WGA

AQP2 ﬁ

V-ATPase stained green and IC-speci c marker V-ATPase stained blue. ICs had low level of expression of ILK in general and there was no detectable change in ILK expression in KO kidney compared with the wild type. Arrows indicate ILK

ILK / AQP2 WGA expression in basal membranes in wild-type / AQP2 and KO PCs, respectively. Scale bar, 10 mm. V-ATPase ILK Data are presented as mean6SEM. Statistics were performed using the t test.

JASN 30: 2073–2090, 2019 ILK Deletion Causes PC Necroptosis 2075 BASIC RESEARCH www.jasn.org

A B 4w WT 4w KO 4w WT 4w KO 8w KO α -SMA AQP2 Sirius red-bright light Col l Polarized light AQP2

C 4w WT 4w KO 8w KO FN AQP2 Masson trichrome

D WT KO WT KO MW(kDa) Relative protein expression (Normalized to GAPDH) NGAL 22 03691215

NGAL FN 250 ** FN α-SMA 42 *** α-SMA 75 ** p-Smad3 50 p-Smad3 ** 75 T-Smad2/3 T-Smad2/3 50 *** p-Smad3/ 4w WT GAPDH 37 T-Smad2/3 4w KO

E F G β FN Col l TGF i 25 60 ## 20 # ## 20 15 40 15 10 10 *** 20 expression expression expression 5 5 *** *** Relative mRNA Relative mRNA Relative mRNA 0 0 0

4w WT 4w KO 8w WT 8w KO 4w WT 4w KO 8w WT 8w KO 4w WT 4w KO 8w WT 8w KO

Figure 2. Ilk KO in PCs causes kidney fibrosis and activates the TGF-b/Smad signaling pathway. (A) Immunofluorescence staining revealed the increased expression of a-SMA (red and arrows, top panels), ECM collagen type 1 (Col I; red and arrows, middle panels), and fibronectin (FN; red and arrows, bottom panels) in the cortex of 4-week-old KO mice. CD PCs were highlighted by immunostaining

2076 JASN JASN 30: 2073–2090, 2019 www.jasn.org BASIC RESEARCH with PBS, kidneys were embedded in optimal cutting tem- membranes (MilliporeSigma, Burlington, MA). Membranes perature compound or paraffin, and cut into 4-mm (cryostat) were incubated with primary antibodies (Supplemental Table 1) or 3-mm (paraffin) sections. followed by horseradish peroxidase–conjugated secondary antibodies (Santa Cruz, Dallas, TX), and then developed using the Measurement of Serum Creatinine and BUN ECL Pro Kit (PerkinElmer, Waltham, MA). Relative intensity of Serum creatinine and BUN content were measured using the protein expression was quantified by OD using the Quantity QuantiChrom Creatinine Assay Kit (BioAssay Systems, Hay- One software (Bio-Rad). ward, CA) and Stanbio Urea Nitrogen Kit (Stanbio Laboratory, Boerne, TX), respectively, following the manufacturers’ in- Hematoxylin and Eosin Staining structions. Approximately 60 ml of serum was used for each Hematoxylin and eosin (H&E) staining was conducted using measurement. kidney paraffin sections as previously described.35 The tubular injury score was calculated based on the percentage of tubules Immunofluorescence Staining and Immunoblotting in kidney sections that displayed tubular damage, including Paraffin-embedded kidney sections were deparaffinized. Cryo- tubular dilation or atrophy, tubular necrosis, loss of the brush stat kidney sections were rehydrated in PBS. Routine immu- border, or cast formation.36 The degree of injury was graded nostaining was performed as previously described.33 Briefly, semiquantitatively and blindly by two independent re- sections were blocked with 1% BSA and incubated with pri- searchers from ten randomly chosen fields of each kidney sec- mary antibodies of various dilutions from 1:100 to 1:3000 tion, according to the extent of injury involved in each field (Supplemental Table 1) or corresponding normal IgG of as follows: 0, normal; 1, ,10%; 2, 11%–25%; 3, 26%–75%; 1:100 dilution, followed by the fluorophore-conjugated sec- 4, .75% of the observed tubules. ondary antibody (Jackson ImmunoResearch Laboratories, West Grove, PA). Slides were viewed under a Nikon Eclipse Picrosirius Red Staining and Masson Trichrome 90i epifluorescence (Nikon Instruments; Melville, NY) or a Staining Zeiss LSM800 confocal microscope (Carl Zeiss Microscopy; Paraffin-embedded kidney sections were stained using the Picro- Thornwood, NY). Analysis of fluorescence intensity was per- sirius Red Stain Kit (Polysciences, Warrington, PA) or Trichrome formed using ImageJ software. Immunofluorescence staining Stain Kit (Masson; Sigma-Aldrich) following the manufacturers’ of kidney with MLKL, CD68, and lymphocyte antigen 6 com- instructions, and examined by a Zeiss LSM800 confocal micro- plex locus G (Ly6G) was performed on cryosectioned slides. scope. Using picrosirius red staining, type 1 collagen fibers stain The rest of the immunostaining was performed on paraffin yellow to orange and type 3 collagen fibers stain green under sections. Quantification of the number of F4/80- or Ly6G- polarized light. Using Masson trichrome staining, collagen fibers positive cells was performed manually in ten randomly se- usually stain blue, cytoplasm stains pink, and nuclei stain dark lected fields (magnification, 4003) for each specimen. Data brown. are presented as the number of F4/80- or Ly6G-positive cells per high power field with a magnification of 4003. Transmission Electron Microscopy Immunofluorescence staining of cultured cells was per- Transmission electron microscopy (TEM) was performed as formed as previously described.34 Briefly, cells grown on glass previously described without modification.37 Briefly, parafor- coverslips were fixed with 4% paraformaldehyde for 20 min- maldehyde-lysine-periodate–fixed kidney slices were post- utes, and then permeabilized with 0.01% Triton X-100 for 4 fixed with 2% glutaraldehyde in 0.1 M sodium cacodylate minutes. After blocking with 1% BSA, they were subjected to buffer at 4°C for 24 hours. The slices were incubated with routine immunostaining as mentioned above. 1% osmium tetroxide in cacodylate buffer at room tempera- Immunoblotting was performed as previously described.34 ture for 1 hour, followed by several rinses in 0.1 M sodium Briefly, homogenized kidney tissues or cell lysates were sepa- cacodylate buffer. After dehydration in a graded ethanol series rated by SDS-PAGE and transferred to polyvinylidene difluoride from 50% to 100%, followed by brief dehydration in with anti-AQP2 antibody (green). Nuclei were stained blue with DAPI. Scale bar, 10 mm. (B) Deposition of collagen fibrils in 4-week-old wild- type (WT) and 4- and 8-week-old KO kidney was examined by picrosirius red staining. Collagen type 1 (red) and type 3 fibrils (green) were observed under polarized light (bottom three panels). Top three panels are images obtained under the bright light. Scale bar, 50 mm. (C) Masson trichrome staining revealed the excessive deposition of collagen fibers in 4- and 8-week-old KO kidney compared with the control. Scale bar, 50 mm. (D) Representative immunoblotting revealed increased expression of NGAL, FN, a-SMA, phosphorylated Smad3 (Ser423/425, p-Smad3, arrows), and total Smad2/3 (T-Smad2/3) in 4-week-old Ilk KO kidney (left). Individual protein signal intensity was quantified by densitometry and plotted over the intensity of glyceraldehyde-3-phosphate dehydrogenase (GAPDH). The ratio of p-Smad3 over T-Smad2/3 is also shown in the graph (right). **P,0.01, ***P,0.001 versus wild-type mice. (E–G) Real-time PCR confirmed the augmented mRNA levels of (E) FN, (F) Col I, and (G) TGF-biinIlk KO kidney. Expression of individual genes was normalized to b-actin. ***P,0.001 versus 4-week-old wild-type mice; #P,0.05, ##P,0.01 versus 8-week-old wild-type mice. Data presented as mean6SEM, n=6–7 for (D–G). Statistics were performed using the t test. MW, molecular weight.

JASN 30: 2073–2090, 2019 ILK Deletion Causes PC Necroptosis 2077 BASIC RESEARCH www.jasn.org

A B WT KO 40 ***

F4/80 20

10 AQP2 F4/80 positive cells/field 0 WT KO

C D WT KO 15 ***

10 Ly6G

5 Ly6G positive cells/field AQP2 0 WT KO

E WT KO F TNF- 10 ** 8 6 CD68 4

expression 2 Relative mRNA

AQP2 0 WT KO

G H I J IL-6 IL-1 IL-33 CXCL1 3 5 8 ** 10 ** *** * 6 8 4 2 6 3 4 4 2 1 expression expression expression expression 2 2 1 Relative mRNA Relative mRNA Relative mRNA Relative mRNA 0 0 0 0 WT KO WT KO WT KO WT KO

K WT KO WT KO MW(kDa) 15 WT ** KO 65 p-p65 10 * 65 p65 5

GAPDH 37 (Normalized to GAPDH) 0 Relative protein expression p-p65 p65

2078 JASN JASN 30: 2073–2090, 2019 www.jasn.org BASIC RESEARCH propylene oxide, kidney slices were infiltrated with Eponate as the percentage of treated group over control group was resin (Ted Pella, Redding, CA) and embedded in fresh Eponate calculated as OD treat/OD control3100% for the MTT assay, at 60°C. Ultra-thin sections were cut, stained with uranyl ac- and fluorescencetreat/fluorescencecontrol3100% for the DNA etate and lead citrate, and examined under a JEOL JEM 1011 quantification assay. transmission electron microscope (JEOL, Peabody, MA). Im- ages were taken using an AMT digital imaging system (Ad- Small interfering RNA Transfection vanced Microscopy Techniques, Danvers, MA). ILK small interfering RNA (siRNA) (sense sequence, AAGGA- CACAUUCUGGAAGGGG; antisense sequence, CCUUCCA- RNA Isolation and Quantitative Real-Time PCR GAAUGUGUCCUUGG) was purchased from GE Dharmacon Mouse kidney RNA was extracted using QIAshredder and (Lafayette, CO). mCCDC11 cells were seeded in six-well plates RNeasy purification kits (Qiagen, Valencia, CA). The synthesis and transfected with ILK siRNA (50 nM) or scrambled siRNA of cDNA from mRNAwas performed following the protocol of mixed with 5 ml Lipofectamine 2000 (Invitrogen) in 1 ml the High-Capacity RNA-to-cDNA Kit (Applied Biosystems, DMEM for 6 hours. At 24 hours after transfection, some con- Foster City, CA). Quantitative real-time PCR was carried trol cells (transfected with or without scrambled siRNA) were out using a QuantStudio 3 Real-Time PCR system and the incubated with 2 mM cpd22, and 1 ng/ml recombinant TNF-a PowerUp SYBR Green PCR Master Mix (Life Technology, (300-01A; PeproTech), respectively. The cell viability assay and Carlsbad, CA). The level of specific mRNA was normalized immunofluorescence staining were performed 48 hours after to b-actin gene expression. The relative mRNA expression was ILK siRNA transfection. 2ΔΔ determined by the 2 Ct method. The sequences of primers are summarized in Supplemental Table 2. Statistical Analyses Data are shown as mean6SEM of independent replicates Cell Culture and Cell Viability Assays (n$3). Experimental data were analyzed using the t test for LLC-PK1 cells and mCCDC11 cells were cultured in DMEM two groups or with one-way ANOVA for multiple groups, containing 10% FBS in a 5% carbon dioxide/95% air-humidified using GraphPad Prism version 5.01 (GraphPad Software, atmosphere at 37°C. Cells were trypsinized in 0.25% trypsin- San Diego, CA). A P value ,0.05 was considered statistically EDTA, and plated on dishes. Cells were allowed to grow for 24 significant. hours, 0.5, 1, and 2 mM of the ILK inhibitor cpd22 (407331; Calbiochem, San Diego, CA) was added, and then incubated with cells for 24 hours. Control cells were treated with 0.1% RESULTS DMSO. After treatment for 24 hours, cells were harvested, lysed in lysis buffer, and prepared for electrophoresis (SDS- Renal Failure and Kidney Tubular Injury Occur in Mice PAGE) and immunoblotting. All experiments were repeated with Ilk Deletion in PCs at least three times. To delete ILK in CD PCs, we crossed Aqp2-Cre and floxed Ilk The LLC-PK1 were seeded on 96-well plates and grown transgenic mice. The expression of Cre recombinase is con- to approximately 80% confluence. Cells were then treated trolled by the Aqp2 promoter, which is expressed specifically fl fl with 0.2% DMSO or ILK inhibitor cpd22 in the presence in CD PCs after embryonic day 18.5. The Ilk / ; Aqp2-Cre (Ilk or absence of the necroptotic inhibitor Nec-1 (50 mM) KO) mice appeared similar to wild-type littermates at birth for 24 hours. Both the 3-(4,5-dimethylthiazol-2-yl)-2,5- and within 2–3 weeks after birth. However, at around 4 weeks, diphenyltetrazolium bromide (MTT; Affymetrix, Santa KO mice exhibited significantly lower body weight and were Clara, CA) assay and DNA quantification assay (Invitrogen) less active compared with their littermates (Figure 1, A and B). were applied to assess cell viability by measuring OD values At 8 weeks old, more than half of the KO mice became se- or fluorescence intensity, respectively. Relative cell viability verely distressed, exhibiting hunched posture, ruffled fur, and

Figure 3. PC-specificablationofIlk induces renal inflammation. (A, C, and E) Representative immunofluorescence staining revealed increased F4/80-positive macrophages (red and arrows, A), Ly6G-positive neutrophils (red and arrows, C) and CD68-positive macrophages (red and arrows, E) in 4-week-old KO kidney. AQP2 stained green and nuclei stained blue. Scale bar, 20 mm. (B) F4/80-positive cells and (D) Ly6G-positive cells in 4-week-old kidney was quantified in ten randomly chosen fields for each specimen. Data are presented as the number of positive cells per high power field (mean6SEM/field). n=4–6. Original magnification, 4003.(F–J) Increased gene expression of (F) TNF-a,(G)IL-6,(H)IL-1b, (I) IL-33, and (J) CXCL1 in 4-week-old Ilk KO kidney was revealed by real-time PCR. Expression of individual genes was normalized to the expression of b-actin. (K) Representative immunoblotting images showed increased expression of phosphorylated (Ser536) and total NF-kB p65 subunit in 4-week-old KO kidney (left). The ratio of phosphorylated p65 and total NF-kB p65 over glyceraldehyde-3-phosphate dehydrogenase (GAPDH) is plotted in the graph (right). *P,0.05, **P,0.01, ***P,0.001 versus wild-type (WT) mice. Data presented as mean6SEM, n=5–8 for (F–K). Statistics were performed using the t test. MW, molecular weight.

JASN 30: 2073–2090, 2019 ILK Deletion Causes PC Necroptosis 2079 BASIC RESEARCH www.jasn.org

Figure 4. CD epithelial cell injury and interstitial fibrosis are revealed by TEM in 4- and 8-week-old Ilk KO kidney. (A, a) CD appeared intact in 4-week-old wild-type (WT) kidney. (b and c, arrows) Increased intercellular junction and gaps between PCs and basement membrane were observed in CDs of 4-week-old Ilk KO kidney. (b) Detached tubular cells and cell debris were seen within the tubular lumen. Scale bar, 5 mm. (B) Increased interstitial fibrosis and inflammatory cell infiltration in Ilk KO kidney. (a) Extracellular matrix and fibrils were accumulated around CDs in KO kidney. The CD basement membrane was thickened (arrowhead). (a) Activated fibroblasts (F), (b) monocytes (M), and (c) macrophages (Mf) were detected in the interstitium next to CDs in KO kidney. Scale bar, 2 mm. (C) Ilk KO PCs were undergoing necroptosis as indicated by cell swelling with (b, double arrows) lucent cytoplasm, (c, hollow arrow) apical membrane rupture, and (c) release of abundant amorphous cellular contents that were filled in the lumen of CDs. (a) Wild-type PCs remained intact. Scale bar, 2 mm. (D) Deformed mitochondria (arrows) and swollen endoplasmic reticulum (arrowheads) were present in (b) Ilk KO PCs, compared with (a) wild-type PCs. Scale bar, 500 nm. emaciation. The Kaplan–Meier curve revealed a mortality of lipocalin (NGAL) was also detected in KO mice (Figure 1F). approximately 60% at 8 weeks of age (Figure 1C). The concentration of urinary NGAL in 8-week-old KO mice Serum creatinine andBUNweresignificantly increased in 4- was estimated to be 0.542 mg/ml (Supplemental Figure 1). week-old KO mice and continued to rise at 8 weeks, compared H&E staining revealed increased tubular dilation and tubular with wild-type mice, indicating progressive renal failure in cell atrophy in 4-week-old KO kidney (Figure 1G). Kidney PC Ilk KO mice (Figure 1, D and E). Increased urine secre- architecture was severely disrupted with tubular damage and tion of kidney injury marker neutrophil gelatinase-associated expansion of extracellular matrix (ECM) in 8-week-old KO

2080 JASN JASN 30: 2073–2090, 2019 www.jasn.org BASIC RESEARCH

A B WT KO WT KO MW(kDa) MLKL AQP2 MLKL/AQP2

MLKL 52 75 RIPK3 WT 50 GAPDH 37 KO 10 WT * 8 KO 6 4 *** D

expression 2 p-RIPK1 AQP2 p-RIPK1/AQP2

Relative protein 0 MLKL RIPK3 (Normalized to GAPDH) WT C p-MLKL AQP2 p-MLKL/AQP2 KO WT E p-RIPK3 AQP2 p-RIPK3/AQP2 WT KO KO

Figure 5. Necroptosis signal is induced in CDs of Ilk KO mice. (A) Representative immunoblotting showed the upregulation of expression of MLKL and RIPK3 in 4-week-old KO mouse kidney (top panel). Expression of MLKL and RIPK3 was quantified by densitometry and normalized to the expression of glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (bottom panel). *P,0.05, ***P,0.001 versus wild-type (WT) mice. Data presented as mean6SEM, n=6–8. Statistics were performed using the t test. (B) Im- munofluorescence staining revealed increased apical membrane expression and accumulation of MLKL (red, arrowheads) in AQP2- positive PCs (green) in Ilk KO mice, whereas the staining of MLKL (red, arrows) was more diffuse inside cells and in the basal region in wild-type PCs (green). Scale bar, 10 mm. (C) Increased apical membrane accumulation of phosphorylated MLKL (Ser345, p-MLKL, red) signal was clearly detected in Ilk KO PCs (AQP2 stained green) by Airyscan fluorescence confocal microscopy. Scale bar, 10 mm. (D and E) Increased expression of phosphorylated RIPK1 (Ser166, p-RIPK1, red) and phosphorylated RIPK3 (Thr231/Ser232, p-RIPK3, red) was also detected in Ilk KO PCs (green) by immunofluorescence staining. Scale bar, 10 mm. MW, molecular weight. kidneys. Detached tubular cells were clustered within the in PCs (Figure 1, I and J, Supplemental Figure 2). There was no dilated tubules. Tubular injury score was significantly higher detectable alteration of ILK expression in intercalated cells in KO mice than wild-type mice (Figure 1H). (ICs) in PC Ilk KO and wild-type mice (Figure 1J). Downregulation of ILK expression in PC Ilk KO kidney was examined by immunofluorescence staining. In wild-type kid- PC Ilk KO Induces Interstitial Fibrosis and Activation of ney, ILK expression was clearly detected in the basal mem- TGF-b/Smad Signaling Pathway in Kidney brane in AQP2-expressing PCs, and was costained with wheat Immunofluorescence staining for markers of activated fibro- germ agglutinin–FITC (Figure 1, I and J, Supplemental Fig- blasts, a-smooth muscle actin (a-SMA), and ECM including ure 2). Conversely, in PC Ilk KO mice, ILK expression was collagen type 1 and fibronectin was performed in 4-week- significantly diminished or even absent in the basal membrane old mouse kidney. a-SMA, collagen I, and fibronectin staining

JASN 30: 2073–2090, 2019 ILK Deletion Causes PC Necroptosis 2081 BASIC RESEARCH www.jasn.org

A a b LLC-PK1 mCCDC11 120 120 100 * 100 80 80 *** *** *** 60 60 *** 40 40 20 20 Cell viability as *** Cell viability as 0 0 2 percentage of control (%) 1 4 percentage of control (%) con 0.5 con 0.25 DMSO ILK KD cpd22 TNF-D cpd22 (µM)

B con ILK siRNA cpd22-2µM TNF-D p-MLKL

C con cpd22-0.5µM cpd22-1µM cpd22-2µM MLKL

D E cpd22(µM) con 120 ### 2.5 0.5µM con 0.5 1 2 MW(kDa) * 100 1µM * 2.0 p-Akt 60 2µM * 80 1.5 60 Akt 60 1.0 * * 40 MLKL * * 52 Cell viability as 0.5 20 75 RIPK3 0 50 0 percentage of control (%)

Relative protein expression p-Akt MLKL RIPK3 E-actin 45 con /Akt /E-actin /E-actin cpd22 cpd22 +Nec-1

Figure 6. ILK inhibition and knockdown upregulates necroptotic signal in cultured renal tubular epithelial cells. (A, a) LLC-PK1 relative cell viability was reduced with the treatment of cpd22 in a dose-dependent manner as measured by MTT assay. Cpd22 (0.25–4 mM) was incubated with cells for 24 hours. *P,0.05, ***P,0.001 versus control (con). (b) ILK siRNA knockdown (ILK KD) or treatment with TNF-a or cpd22 (2 mM) caused cell loss as measured by DNA quantification assay in mCCDC11. ***P,0.001 versus control. (B) In- creased plasma membrane accumulation of phospho-MLKL (p-MLKL, red) was clearly detected in mCCDC11 cells transfected with ILK siRNA. Similar membrane accumulation of phospho-MLKL was also present in cells treated with cpd22 or TNF-a as detected by immunofluorescence staining. Scale bar, 10 mm. (C) Immunofluorescence staining revealed that cpd22 treatment caused translocation of MLKL signals (red) from cytosol to the plasma membrane in LLC-PK1, compared with the control. Nuclei were stained blue with DAPI. Scale bar, 20 mm. (D) Representative immunoblotting showed the reduction of phosphorylated Akt (Ser 473, p-Akt) and upregulation of MLKL and RIPK3 in LLC-PK1 with cpd22 treatment for 24 hours (left panel). Bar graph represents the ratio of p-Akt over total Akt, and the ratio of MLKL and RIPK3 over b-actin expression as quantified by densitometry (right panel). *P,0.05, **P,0.01 versus control.

2082 JASN JASN 30: 2073–2090, 2019 www.jasn.org BASIC RESEARCH was barely detectable in wild-type kidney, but markedly in- CD Epithelium Injury in PC Ilk KO Mice Is Revealed by creased in the interstitium of Ilk KO kidney, in proximity to Electron Microscopy or surrounding AQP2-positive CDs (Figure 2A, Supplemen- We next examined the CD epithelial cells using TEM in Ilk KO tal Figure 3). Picrosirius red staining and Masson trichrome kidney. The ultrastructure of the glomerulus, segments of staining further revealed excessive deposition of collagen proximal tubules, and thin and thick ascending and descend- fibrils in 4- and 8-week-old KO kidney compared with the ing loops appeared intact in 4-week-old KO and wild-type wild type (Figure 2, B and C). A significant elevation of kidney (data not shown). CDs appeared intact in wild-type NGAL, fibronectin, and a-SMA in 4-week-old KO kidney mice (Figure 4A, a). However, in KO kidney, CDs were severely was revealed by immunoblotting (Figure 2D). In addition, distorted, with disrupted intercellular junctions between ad- the expression of several ECM genes, including fibronec- jacent CD tubular cells and increased spacing between PCs and tin and collagen I, was significantly increased in 4-week- the basement membrane (Figure 4A, b and c). Detached ep- old KO kidney and even more dramatically at 8 weeks old ithelial cells and cellular debris were frequently observed in the by quantitative real-time PCR, indicating an escalating fi- tubular lumen (Figure 4A, b). brogenic gene expression over time in KO kidney (Figure 2, The CD basement membranes were often thickened, and EandF). there was a large accumulation of ECM and fibrils surrounding The canonical signaling pathway of TGF-b and its down- the CDs in 4-week-old KO kidney, which was more prominent stream effector molecules, Smads, were then investigated in in 8-week-old KO kidney (Figure 4B, a). Activated fibroblasts PC Ilk KO kidney. Total Smad2/3 was mildly increased with enlarged nuclei and expanded long processes were fre- whereas phosphorylated Smad3 was dramatically increased quently observed together with increased infiltration of in KO kidney (Figure 2D). TGF-b–induced protein (TGF- monocytes and macrophages in interstitium surrounding bi) is a target protein induced by active TGF-b and is ex- the CDs (Figure 4B). creted to ECM. TGF-bi mRNA was also increased six times In 4-week-old KO kidney, PCs were swollen with lucent and 13.5 times more in 4- and 8-week-old KO kidney, re- cytoplasm, in contrast to the dense cytoplasm in adjacent spectively, compared with wild type (Figure 2G), ICs and PCs in wild-type kidney (Figure 4C, a and b). Many indicating a robust activation of the canonic TGF-b signal- PCs had lost plasma membrane continuity and were ruptured ing pathway. in KO kidney (Figure 4C, c). Amorphous cellular contents, associated with or without vesicular membrane structures, PC Ilk Deletion Induces Extensive Inflammation in were released and filled the lumen of CDs in KO kidney (Fig- Kidney ure 4C, c). Furthermore, mitochondria in Ilk KO PCs ap- We further examined the presence of inflammatory signals peared deformed and the endoplasmic reticulum was swollen in PC Ilk KO kidney because inflammation contributes to (Figure 4D, b). Despite dramatic morphologic alteration, the the exacerbation of AKI and renal fibrosis.38,39 Immunofluo- nuclear membrane and nucleus of ILK-deficient PCs appeared rescence staining with the macrophage marker F4/80, CD68, intact. Approximately 75% of observed PCs had undergone and neutrophil marker Ly6G revealed a significantly increased necroptosis in PC Ilk KO kidney, based on TEM examination. number of macrophages and neutrophils in KO kidney (Fig- In contrast, ICs appeared grossly intact with seemingly normal ure 3, A–E). Furthermore, KO kidney showed significantly plasma membrane, dense cytosol content, and seemingly nor- augmented gene expression of proinflammatory cytokines, mal mitochondria in PC Ilk KO kidney. including TNF-a,IL-6,IL-1b, IL-33, and chemokine CXCL1 by quantitative real-time PCR (Figure 3, F–J). In ad- Ilk Deletion in PC Causes Upregulation and Activation dition, both the phosphorylated and total NF-kB RelA/p65 of Necroptosis Signal in Kidney subunit were elevated in KO kidney by immunoblotting (Fig- Necroptosis in Ilk KO kidney was first studied by immuno- ure 3K), indicating the activation of the NF-kB pathway, an blotting. Immunoblotting showed that expression of MLKL important regulator of inflammation.40 We also examined in- and its upstream kinase RIPK3 was significantly elevated in flammatory cytokine expression in other tissues. Despite a KO kidney (Figure 5A). Next, necroptosis signaling was exam- significantly increased expression of IL-33 and CXCL1 in KO ined by immunofluorescence staining using antibodies against kidney, mRNA levels of IL-33 and CXCL1 in liver and spleen total and phospho-MLKL, RIPK1, and RIPK3. The total MLKL were not significantly altered in 4-week-old KO mice, sug- signal was located diffusely inside PCs and in the PC basal mem- gesting that PC Ilk KO caused inflammation predominantly brane in wild-type kidney (Figure 5B). However, in KO kidney, in the kidney. immunostaining of MLKL appeared more intense and,

(E) Nec-1 treatment of LLC-PK1 cells reversed the reduction of cell viability induced by cpd22. LLC-PK1 cells were treated with cpd22 (2 mM) in the presence or absence of Nec-1 (50 mM) for 24 hours. Relative cell viability was assessed by measuring the cellular DNA content. *P,0.05 versus control; ###P,0.001 versus 2 mM cpd22 treatment group. Data presented as mean6SEM. Experiments were repeated at least three times. Statistics were performed using one-way ANOVA.

JASN 30: 2073–2090, 2019 ILK Deletion Causes PC Necroptosis 2083 BASIC RESEARCH www.jasn.org

A B C

0.4 ** 80 ** Mice born ## Nec-1 or vehicle 0.3 60 # Genotype every other day 0.2 40

0.1 20 WT or KO mice Scr (mg/dL) BUN (mg/dL) P0 P14 P28 0.0 0

D WT+DWT+NKO+DKO+N WT+DWT+NKO+DKO+N

WT KO *** DMSO Nec-1 DMSO Nec-1 MW(kDa) 25 WT+D WT+N 20 KO+D NGAL 22 KO+N 15 MLKL 52 10 *** 75 ## RIPK3 5 ### 50 *** #

GAPDH 37 (Normalized to GAPDH) 0 Relative protein expression NGAL MLKL RIPK3

E F WT+DMSO WT+Nec-1 KO+DMSO KO+Nec-1 4 *** 3 ### 2

Tubular injury score 0

WT+D WT+N KO+D KO+N

G H WT+DMSO WT+Nec-1 KO+DMSO KO+Nec-1 WT+D 4 WT+N KO+D *** 3 KO+N

MLKL ### 2 **

expression 1 Relative mRNA AQP2 0 MLKL RIPK3

Figure 7. Inhibition of necroptosis attenuates tubular injury and renal function decline in PC Ilk KO mice. (A) Diagram of Nec-1 administration schedule for wild-type (WT) and Ilk KO mice. After genotyping, 2-week-old wild-type and KO mice were treated with 1.65 mg/kg Nec-1 or vehicle (1% DMSO in PBS) by intraperitoneal injection every 2 days for 2 weeks. (B) Nec-1 reduced the elevation of serum creatinine (Scr) and (C) BUN in Ilk KO mice. **P,0.01 versus wild-type with DMSO mice; #P,0.05, ##P,0.01 versus KO with DMSO mice. (D) Representative immunoblotting showed that Nec-1 reduced expression of NGAL, MLKL, and RIPK3 in Ilk KO kidney (left). The expression of NGAL, MLKL, and RIPK3 was quantiﬁed by densitometry, and normalized to glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (right). ***P,0.001 versus wild-type with DMSO mice; #P,0.05, ##P,0.01, ###P,0.001 versus KO with DMSO mice. (E) Representative images of H&E staining in kidney cortex. Nec-1–treated Ilk KO mice showed less tubular dilation and microcyst formation, compared with wild-type kidney treated with or without Nec-1. Scale bar, 50 mm. (F) Nec-1 treatment improved the tubular injury score in Ilk KO kidney. ***P,0.001 versus wild type with DMSO mice; ###P,0.001 versus KO with DMSO mice. (G) Immunoﬂuorescence staining showed that Nec-1 blocked translocation of MLKL signal (red) from cytosol to plasma membrane in Ilk KO PCs (AQP2 stained green). Scale bar, 10 mm. (H) Gene expression of MLKL and RIPK3 was evaluated by real-time PCR. Expression of individual genes was normalized to b-actin. **P,0.01, ***P,0.001 versus wild-type with DMSO mice; ###P,0.001 versus KO with

2084 JASN JASN 30: 2073–2090, 2019 www.jasn.org BASIC RESEARCH interestingly, it clearly accumulated in the apical membrane in C). H&E staining of the kidney revealed that Nec-1 treatment Ilk KO PCs (Figure 5B, Supplemental Figure 4). Subsequent significantly improved tubular atrophy and dilation, and over- immunofluorescence staining using antibody against phosphor- all tubular injury score in KO mice (Figure 7, E and F). ylated MLKL (Ser345) revealed a dramatic apical mem- We further examined the necroptosis pathway in Nec-1– brane accumulation of phospho-MLKL in Ilk KO PCs (Figure treated KO mice. Immunoblotting showed that Nec-1 treat- 5C). Furthermore, immunofluorescence staining with anti– ment significantly reversed the upregulation of MLKL and phospho-RIPK1 (Ser166) and anti–phospho-RIPK3 (Thr231/ RIPK3 in KO kidney compared with vehicle (Figure 7D). Ser232) antibodies revealed increased phospho-RIPK1 and Quantitative real-time PCR showed that RIPK3 mRNA was RIPK3 signals in Ilk KO PCs compared with the wild type (Figure also significantly reduced whereas MLKL mRNAwas slightly 5, D and E). reduced in Nec-1–treated KO mice (Figure 7H). In addition, Nec-1 decreased the membrane translocation of MLKL in ILK- ILK Inhibition and Knockdown Increases Cell Death and deficient PCs compared with wild-type PCs, indicating that Upregulates Necroptotic Signal in Cultured Renal Nec-1 blocked the MLKL activation induced by ILK deficiency Tubular Epithelial Cells (Figure 7G, Supplemental Figure 4). We next explored the role of ILK inhibition and knockdown in cultured kidney tubular cells, LLC-PK1, and a mouse CD cell Nec-1 Ameliorates Inflammation and Kidney Fibrosis in line, mCCDC11, respectively. A well known ILK inhibitor, PC Ilk KO Mice cpd22, was used to block ILKactivity, as indicated by inhibition We investigated further inflammatory responses and kidney of the phosphorylation of Akt, a key downstream target of ILK fibrosis in Ilk KO mice after Nec-1 treatment. Intriguingly, (Figure 6D). Treatment of LLC-PK1 cells with 0.25–4 mM Nec-1 caused a marked reduction in the gene expression of cpd22 for 24 hours reduced cell viability in a dose-dependent TNF-a, IL-6, and IL-1b in KO kidney (Figure 8A). Nec-1 also manner as measured by MTT (half maximal inhibitory con- significantly blocked the upregulation of the phosphorylated centration of 2.855 mM; Figure 6A, a). ILK siRNA knockdown NF-kB p65 and total p65 subunit, as well as the F4/80-positive in mCCDC11 increased cell loss, as measured by the DNA macrophage infiltration in KO mice (Figure 8, B–D). In addition, quantification assay, to the same degree as cells treated with immunofluorescence staining revealed a dramatic reduction of TNF-a or cpd22 (Figure 6A, b). Immunofluorescence staining a-SMA, fibronectin, and collagen type 1 in the interstitium in detected increased plasma membrane accumulation of phos- Nec-1–treated KO kidney compared with vehicle-treated KO pho-MLKL in mCCDC11 cells with ILK siRNA knockdown mice (Figure 8E). Consistently, Nec-1 also blocked the increase and in cells treated with cpd22 or TNF-a (Figure 6B). Immu- of fibronectin and a-SMA protein, as well as the mRNA of fi- nostaining of cpd22-treated LLC-PK1 revealed a similar trans- bronectin, collagen type 1, and TGF-bi in KO kidney (Figure 8, B location of MLKL signal from cytosol to the plasma membrane and F). Therefore, inhibiting necroptosis by Nec-1 effectively (Figure 6C), concomitant with the increased expression of antagonizes inflammation and interstitial fibrosis in the kidney MLKL and RIPK3 detected by immunoblotting (Figure 6D). induced by ILK deficiency in PCs. We then applied a necroptosis inhibitor Nec-1, which is an allosteric inhibitor of RIPK1.41 When LLC-PK1 cells were incubated with Nec-1 in the presence of cpd22, Nec-1 was DISCUSSION able to reverse the reduction of cell survival caused by ILK inhibition (Figure 6E). The goal of this study was to address the specificroleofILKin maintaining tubular structure and function. By deleting ILK in Nec-1 Inhibits Necroptosis and Attenuates AKI in PC Ilk PCs in mouse kidney, we unexpectedly uncovered the critical KO Mice involvement of ILK in mediating necroptosis of CD PCs, which We next examined an essential role of necroptosis in mediating led to profound inflammation and interstitial fibrosis in the kidney injury in PC Ilk KO mice using the necroptosis inhib- kidney. This study highlights the essential function of ILK itor Nec-1. We defined our treatment window of Nec-1 to be signaling in promoting PC survival through modulating nec- between 2 and 4 weeks after birth (Figure 7A). Compared with roptosis, and provides a novel understanding of the crucial vehicle-treated KO mice, Nec-1–treated KO mice showed sig- contribution of the ECM-integrin-ILK signaling in PCs to nificant reduction in serum creatinine and BUN (Figure 7, B tubular injury, inflammation, and interstitial fibrosis of the and C), as well as renal NGAL expression (Figure 7D). In wild- kidney. type mice treated with Nec-1 or vehicle, there was no detect- Apoptosis and necrosis are considered two major mecha- able difference in serum creatinine or BUN (Figure 7, B and nisms of cell death in acute tubular injury. The role of ILKin cell

DMSO mice. Data presented as mean6SEM, n=5–7. Statistics were performed using one-way ANOVA. KO+D, KO mice treated with 1% DMSO in PBS; KO+N, KO mice treated with Nec-1; P, postnatal day; WT+D, wild-type mice treated with 1% DMSO in PBS; WT+N, wild- type mice treated with Nec-1.

JASN 30: 2073–2090, 2019 ILK Deletion Causes PC Necroptosis 2085 idtp W)wt MOmice; DMSO with (WT) wild-type niie h peuaino hshrltdNF- phosphorylated of upregulation the inhibited ersnaieimmuno Representative 2086 IL-1 and 8. Figure yrgns GPH rgt.*** and (right). FN, (GAPDH) p65, hydrogenase total p-p65, of expression AI RESEARCH BASIC b in e- lcsin blocks Nec-1 JASN Ilk Okde srvae yra-iePR xrsino niiulgnswsnraie to normalized was genes individual of Expression PCR. real-time by revealed as kidney KO E C AB AQP2 Col l AQP2 FN AQP2 -SMA AQP2 F4/80 Relative mRNA expression 10 15 0 5 www.jasn.org fl oecnesann hwdrdcdin reduced showed staining uorescence TNF- WT+DMSO WT+DMSO fl maoyrsos and response ammatory *** ### F IL-6 P Relative mRNA *** ### , expression KO+N KO+D WT+N WT+D ### .0 esswl-yewt MOmice; DMSO with wild-type versus 0.001 6 0 2 4 P , IL-1 .0 essK ihDS ie B ersnaieimnbotn hwdta Nec-1 that showed immunoblotting Representative (B) mice. DMSO with KO versus 0.001 a *** SAwsquanti was -SMA WT+Nec-1 WT+Nec-1 ### FN *** ### GAPDH -SMA k p-p65 6 pp5,ttlp65, total (p-p65), p65 B p65 fi FN rssi PC in brosis CoI I MONec-1 DMSO *** TKO WT ### fi KO+DMSO KO+DMSO db estmtyadnraie ogyeadhd--hsht de- glyceraldehyde-3-phosphate to normalized and densitometry by ed Ilk TGF- fi MONc1MW(kDa) Nec-1 DMSO taino 48-oiiemcohgs(e)i Nec-1 in (red) macrophages F4/80-positive of ltration *** Okde.()Nc1sprse eeepeso fTNF- of expression gene suppressed Nec-1 (A) kidney. KO  i ### fi rnci F) and (FN), bronectin KO+N KO+D WT+N WT+D KO+Nec-1 KO+Nec-1 ## 37 43 250 65 65 P , 0.01, -SMA p-p65 p65 FN ### 036912151821 Relative proteinexpression (Normalized toGAPDH) P , D ## ###

.0 essK ihDS ie (C) mice. DMSO with KO versus 0.001 F4/80 positive ## ##

a cells/field *** 10 20 30 40 *** SAin -SMA 0

WT+D ***

WT+N KO+N KO+D WT+N WT+D *** b Ilk KO+D atn *** -actin. ***

Okde lf) Protein (left). kidney KO KO+N JASN ### 30: P , 2073 – .0 versus 0.001 rae KO treated – 00 2019 2090, a ,IL-6, www.jasn.org BASIC RESEARCH apoptosis has been studied in multiple organs.42,43 Hepatocyte- complex signaling pathways and are closely intertwined.56 or cardiomyocyte-specificILKdeficiency induced apoptosis in What determines apoptosis versus necroptosis is highly de- mice.19,44 More recently, deletion of ILK in distal tubules of pendent on the specific context. For example, various seg- kidney was reported to cause kidney failure and tubular injury, ments of kidney tubules may undergo different forms of cell which was also attributed to apoptosis of tubular cells.45 How- death, even in response to similar insults. Unilateral ureteral ever, in our PC Ilk KO mice, we have not observed prevailing obstruction predominantly elicits necrosis in the proximal tu- apoptosis of CD cells (data not shown). Further investigation led bules but apoptosis in CDs, whereas IRI causes apoptosis and to the discovery of necroptosis in CDs induced by ILK deficiency. necroptosis in proximal tubules.57–59 Isolated CD necroptosis We have revealed ultrastructural characteristics of necroptosis in has not been reported in known AKI models. In our study, ILK-deficient PCs by TEM, and demonstrated upregulation and absence of the ILK signal in PCs may alter the balance toward activation of key necroptotic complex proteins, MLKL, RIPK1, necroptosis as opposed to apoptosis. and RIPK3, in KO kidney. These results suggest the involvement Moreover, cell detachment caused by ILK deficiency may of ILK signaling in tubular cell necroptosis. also trigger necroptosis. Anchorage dependence has long been How does ILK affect the necroptotic process? We have recognized as an essential requirement for cell survival.60 The shown that MLKL translocates from the cytosol to the plasma disruption or loss of integrin binding to ECM initiates a form membrane in Ilk KO PCs and cpd22-treated cultured cells. of caspase-dependent apoptosis that is termed anoikis. As a Significant membrane accumulation of phospho-MLKL is focal adhesion adaptor molecule engaged by integrins, ILK present in Ilk KO PC in mice and in ILK knockdown functions as a pivotal effector in the transduction of survival mCCDC11 cells. It is known that MLKL activation (phosphor- signals from ECM and growth factors.61 We did observe cell ylation) and subsequent homo-oligomerization is required for detachment from the basement membrane in CDs of Ilk KO necroptosis to occur.9 However, it remains unclear by which kidney, therefore cannot rule out the possibility that ILK mechanism MLKL translocates to the plasma membrane. A deficiency leads to cell detachment and detached PCs subse- recent study revealed that heat shock protein 90 is capable of quently undergo necroptosis. Whether cell detachment in- mediating necroptosis directly through regulating membrane duces necroptosis directly has not been described. Our in vitro translocation of activated MLKL.46 We hypothesize that, study in attached cultured cells has shown that inhibiting ILK through regulating cytoskeletal dynamics, ILK may directly or ILK knockdown causes upregulation and activation of nec- or indirectly modulate intracellular trafficking of MLKL and/ roptotic proteins such as MLKL, indicating that ILK deficiency or other necroptotic complex factors. ILK is shown to interact alone is capable of activating necroptosis signaling. with scaffold protein IQGAP1 and effector mDia to promote Emerging studies suggest that the CD of the kidney may the trafficking of caveolin-1 to the plasma membrane by stabi- function as an essential regulator of inflammation in response lizing microtubules.47 We have recently reported that ILK to renal injury in addition to transporting water and salt.37,62,63 regulates a water channel AQP2 to re-entry in the exocytotic Why is CD injury “inflammatory?” It could be due to PC nec- pathway through modulating F-actin polymerization.48 There- roptosis, which is known to be highly inflammatory in general. fore, it is likely that ILK may affect necroptosis through regu- During necroptosis, cell membrane disruption releases a num- lating the trafficking of necroptotic complex proteins, such as ber of cellular contents serving as damage-associated molecular MLKL. ILKwas once considered a serine/threonine kinase,49,50 patterns. These damage-associated molecular patterns trigger but more recent studies suggest that ILK is likely a “pseudoki- the production and release of proinflammatory cytokines and nase.”51–55 Our data suggest that ILK signaling is involved in chemokines to amplify the inflammatory response.64,65 modifying the phosphorylation of RIPK1, RIPK3, and MLKL. Therefore, PC necroptosis is capable of provoking massive However, how the deletion of ILK phosphorylates necroptotic kidney inflammation and fibrosis. complex proteins, directly or indirectly, warrants further The process of cell death occurred during kidney injury is investigation. highly complex. Besides apoptosis and necroptosis, ferroptosis ILK may also be involved in mediating the balance between is recently reported to be involved in certain AKIs.66,67 In apoptosis and necroptosis. Multiple previous studies have addition, multiple pathways may crosstalk and work synergisti- revealed that ILK deficiency induces cell apoptosis in many cally to drive AKI.68 Despite being one of the major inhibitors tissues.19,44 Necroptosis and apoptosis are known to share of necroptosis, Nec-1 is also reported to inhibit ferroptosis.67 kidney. CD PCs were indicated by staining of AQP2 (green). Scale bar, 20 mm. (D) The number of F4/80-positive cells was quantified per field (mean6SEM/field). ***P,0.001 versus wild-type with DMSO mice; ###P,0.001 versus KO with DMSO mice. Original magnification, 4003. (E) Representative immunofluorescence images revealed decreased expression of a-SMA (red), FN (red), and collagen type 1 (Col I, red) in Nec-1–treated Ilk KO kidney. Green fluorescence signal indicates the AQP2-positive CDs. Nuclei were stained blue with DAPI. Scale bar, 20 mm. (F) Nec-1 decreased gene expression of FN, Col I, and TGF-biinIlk KO kidney by real-time PCR. Expression of individual genes was normalized to b-actin. ***P,0.001 versus wild-type with DMSO mice; ###P,0.001 versus KO with DMSO mice. Data presented as mean6SEM, n=5–7. Statistics were performed using one-way ANOVA. KO+D, KO mice treated with 1% DMSO in PBS; KO+N, KO mice treated with Nec-1; WT+D, wild-type mice treated with 1% DMSO in PBS; WT+N, wild-type mice treated with Nec-1.

JASN 30: 2073–2090, 2019 ILK Deletion Causes PC Necroptosis 2087 BASIC RESEARCH www.jasn.org

Attenuating PC necroptosis by Nec-1 in Ilk KO mice, or protect- Supplemental Table 2. List of primers used for quantitative real- ing cells from death induced by ILK inhibition in vitro cannot time PCR. rule out the involvement of ferroptosis in our study. Therefore, Supplemental Figure 1. Increased urinary NGAL secretion in 8 whether ILK also modulates ferroptosis in addition to necrop- weeks old Ilk KO mice was revealed by immunoblotting, normalized tosis in kidney CD PCs remains to be elucidated. to NGAL standard. In summary, we have demonstrated a critical novel function Supplemental Figure 2. ILK expression in wild type and PC Ilk KO of ILK in promoting CD PC survival by modulating the RIPK3- kidney was revealed by immunofluorescence staining under low MLKL–dependent necroptosis pathway. In addition, our magnification. study highlighted the critical contribution of CD injury to Supplemental Figure 3. Immunofluorescence staining using isotype- interstitial inflammation and fibrosis of the kidney. matched immunoglobulin G (IgG) revealed minimal, nonspecific background staining in the wild type and fibrotic Ilk KO kidney. Supplemental Figure 4. Necrostatin-1 (Nec-1) treatment blocked membrane accumulation of MLKL in Ilk KO PCs by immunofluo- ACKNOWLEDGMENTS rescence staining viewed under low magnification. We thank Dr. Dennis Brown (MGH and Harvard Medical School, Boston, MA) for valuable advice on image acquisition and analysis. Ms.M.Huang,Dr.H.Huang,andDr.H.A.J.Luconceivedand REFERENCES designed research; Ms. M. Huang, Ms. Zhu, Ms. He, Dr. Tsuji, Mr. Jin, Dr. Xie, Dr. Ham, and Ms. Capen performed experiments; 1. Humphreys BD: Mechanisms of renal fibrosis. Annu Rev Physiol 80: – Ms. M. Huang and Ms. Zhu analyzed data; Ms. M. Huang, Ms. Zhu, 309 326, 2018 2. Wang J, Wang F, Saran R, He Z, Zhao MH, Li Y, et al.: Mortality risk of Dr. H. Huang, Dr. Tsuji, and Ms. Capen interpreted results of ex- chronic kidney disease: A comparison between the adult populations in periments; Ms. M. Huang prepared figures;Ms.M.Huang,Dr.Yang, urban China and the United States. PLoS One 13: e0193734, 2018 and Dr. H. A. J. Lu drafted manuscript; Ms. M. Huang, Dr. W. Lu, 3. Linkermann A, Chen G, Dong G, Kunzendorf U, Krautwald S, Dong Z: Dr. Paunescu, Dr. Yang, and Dr. H. A. J. Lu edited and revised Regulated cell death in AKI. JAmSocNephrol25: 2689–2701, 2014 manuscript; Ms. M. Huang, Dr. Yang, and Dr. H. A. J. Lu approved 4. Sancho-Martínez SM, López-Novoa JM, López-Hernández FJ: Patho- fi physiological role of different tubular epithelial cell death modes in nal version of manuscript. acute kidney injury. Clin Kidney J 8: 548–559, 2015 5. Linkermann A, Bräsen JH, Himmerkus N, Liu S, Huber TB, Kunzendorf U, et al.: Rip1 (receptor-interacting protein kinase 1) mediates nec- DISCLOSURES roptosis and contributes to renal ischemia/reperfusion injury. Kidney Int 81: 751–761, 2012 6. Zhou W, Yuan J: Necroptosis in health and diseases. Semin Cell Dev None. Biol 35: 14–23, 2014 7. Chen H, Fang Y, Wu J, Chen H, Zou Z, Zhang X, et al.: RIPK3-MLKL- mediated necroinflammation contributes to AKI progression to CKD. FUNDING Cell Death Dis 9: 878, 2018 8. Sun L, Wang H, Wang Z, He S, Chen S, Liao D, et al.: Mixed lineage This work was supported by National Institute of Diabetes and Digestive kinase domain-like protein mediates necrosis signaling downstream of – and Kidney Diseases (NIDDK) grants R01-DK-096015 and R21-DK-092619, RIP3 kinase. Cell 148: 213 227, 2012 NephCure Kidney International, a Gottschalk research grant from the Amer- 9. Chen X, Li W, Ren J, Huang D, He WT, Song Y, et al.: Translocation of ican Society of Nephrology, the S&R Foundation Ryuji Ueno award, and MGH mixed lineage kinase domain-like protein to plasma membrane leads to – Executive Committee on Research support to Dr. H. A. J. Lu. The Microscopy necrotic cell death. Cell Res 24: 105 121, 2014 Core facility of the MGH Program in Membrane Biology receives additional 10. Pasparakis M, Vandenabeele P: Necroptosis and its role in in- fl – support from the Boston Area Diabetes and Endocrinology Research Center ammation. Nature 517: 311 320, 2015 (NIDDK grant DK-57521) and from the Center for the Study of Inflammatory 11. Moreno-Gonzalez G, Vandenabeele P, Krysko DV: Necroptosis: A Bowel Disease (NIDDK grant DK43351). Dr. W. Lu is supported by NIH novel cell death modality and its potential relevance for critical care – grants R01-DK078226. Dr. Paunescu is supported by the MGH Executive medicine. Am J Respir Crit Care Med 194: 415 428, 2016 Committee on Research. Dr. Yang is supported by the National Natural Sci- 12. Zhe-Wei S, Li-Sha G, Yue-Chun L: The role of necroptosis in cardio- ence Foundation of China grant 81620108029. vascular disease. Front Pharmacol 9: 721, 2018 13. Pierdomenico M, Negroni A, Stronati L, Vitali R, Prete E, Bertin J, et al.: Necroptosis is active in children with inflammatory bowel disease and contributes to heighten intestinal inflammation. Am J Gastroenterol SUPPLEMENTAL MATERIAL 109: 279–287, 2014 14. Linkermann A, Bräsen JH, Darding M, Jin MK, Sanz AB, Heller JO, et al.: This article contains the following supplemental material online at Two independent pathways of regulated necrosis mediate ischemia- http://jasn.asnjournals.org/lookup/suppl/doi:10.1681/ASN.2018111162/-/ reperfusion injury. Proc Natl Acad Sci U S A 110: 12024–12029, 2013 DCSupplemental. 15. Xu Y, Ma H, Shao J, Wu J, Zhou L, Zhang Z, et al.: A role for tubular necroptosis in cisplatin-induced AKI. J Am Soc Nephrol 26: 2647–2658, Supplemental Methods. 2015 Supplemental Table 1. List of primary antibodies used for im- 16. Linkermann A, Heller JO, Prókai A, Weinberg JM, De Zen F, Himmerkus munoblotting (IB) and immunofluorescence staining (IF). N, et al.: The RIP1-kinase inhibitor necrostatin-1 prevents osmotic

2088 JASN JASN 30: 2073–2090, 2019 www.jasn.org BASIC RESEARCH

nephrosis and contrast-induced AKI in mice [published correction ap- 35. Tanwar PS, Zhang L, Kaneko-Tarui T, Curley MD, Taketo MM, Rani P, pears in J Am Soc Nephrol 25: 2942, 2014]. J Am Soc Nephrol 24: et al.: Mammalian target of rapamycin is a therapeutic target for murine 1545–1557, 2013 ovarian endometrioid adenocarcinomas with dysregulated Wnt/b-catenin 17. Yen CF, Wang HS, Lee CL, Liao SK: Roles of integrin-linked kinase in cell and PTEN. PLoS One 6: e20715, 2011 signaling and its perspectives as a therapeutic target. Gynecol Minim 36. Wu H, Chen G, Wyburn KR, Yin J, Bertolino P, Eris JM, et al.: TLR4 Invasive Ther 3: 67–72, 2014 activation mediates kidney ischemia/reperfusion injury. JClinInvest 18. Cano-Peñalver JL, Griera M, Serrano I, Rodríguez-Puyol D, Dedhar S, 117: 2847–2859, 2007 de Frutos S, et al.: Integrin-linked kinase regulates tubular aquaporin-2 37. Mamuya FA, Xie D, Lei L, Huang M, Tsuji K, Capen DE, et al.: Deletion of content and intracellular location: A link between the extracellular b1-integrin in collecting duct principal cells leads to tubular injury and matrix and water reabsorption. FASEB J 28: 3645–3659, 2014 renal medullary fibrosis. Am J Physiol Renal Physiol 313: F1026–F1037, 19. Gkretsi V, Mars WM, Bowen WC, Barua L, Yang Y, Guo L, et al.: Loss of 2017 integrin linked kinase from mouse hepatocytes in vitro and in vivo re- 38. Bonventre JV, Yang L: Cellular pathophysiology of ischemic acute sults in apoptosis and hepatitis. Hepatology 45: 1025–1034, 2007 kidney injury. J Clin Invest 121: 4210–4221, 2011 20. Smeeton J, Zhang X, Bulus N, Mernaugh G, Lange A, Karner CM, et al.: 39. Meng XM, Nikolic-Paterson DJ, Lan HY: Inflammatory processes in Integrin-linked kinase regulates p38 MAPK-dependent cell cycle arrest renal fibrosis. Nat Rev Nephrol 10: 493–503, 2014 in ureteric bud development. Development 137: 3233–3243, 2010 40. Christian F, Smith EL, Carmody RJ: The regulation of NF-kB subunits by 21. El-Aouni C, Herbach N, Blattner SM, Henger A, Rastaldi MP, Jarad G, phosphorylation. Cells 5: 12, 2016 et al.: Podocyte-specific deletion of integrin-linked kinase results in 41. Vandenabeele P, Grootjans S, Callewaert N, Takahashi N: Necrostatin- severe glomerular basement membrane alterations and progressive 1 blocks both RIPK1 and IDO: Consequences for the study of cell death glomerulosclerosis. JAmSocNephrol17: 1334–1344, 2006 in experimental disease models. Cell Death Differ 20: 185–187, 2013 22. Kanasaki K, Kanda Y, Palmsten K, Tanjore H, Lee SB, Lebleu VS, et al.: 42. Edwards LA, Thiessen B, Dragowska WH, Daynard T, Bally MB, Dedhar Integrin beta1-mediated matrix assembly and signaling are critical for S: Inhibition of ILK in PTEN-mutant human glioblastomas inhibits the normal development and function of the kidney glomerulus. Dev PKB/Akt activation, induces apoptosis, and delays tumor growth. Biol 313: 584–593, 2008 Oncogene 24: 3596–3605, 2005 23. Dai C, Stolz DB, Bastacky SI, St-Arnaud R, Wu C, Dedhar S, et al.: 43. Friedrich EB, Liu E, Sinha S, Cook S, Milstone DS, MacRae CA, et al.: Essential role of integrin-linked kinase in podocyte biology: Bridging Integrin-linked kinase regulates endothelial cell survival and vascular the integrin and slit diaphragm signaling. JAmSocNephrol17: development. Mol Cell Biol 24: 8134–8144, 2004 2164–2175, 2006 44. Dai J, Matsui T, Abel ED, Dedhar S, Gerszten RE, Seidman CE, et al.: 24. Blattner SM, Kretzler M: Integrin-linked kinase in renal disease: Con- Deep sequence analysis of gene expression identifies osteopontin as a necting cell-matrix interaction to the cytoskeleton. Curr Opin Nephrol downstream effector of integrin-linked kinase (ILK) in cardiac-specific Hypertens 14: 404–410, 2005 ILK knockout mice. Circ Heart Fail 7: 184–193, 2014 25. Li Y, Tan X, Dai C, Stolz DB, Wang D, Liu Y: Inhibition of integrin-linked 45. Raman A, Reif GA, Dai Y, Khanna A, Li X, Astleford L, et al.: Integrin- kinase attenuates renal interstitial fibrosis. JAmSocNephrol20: linked kinase signaling promotes cyst growth and fibrosis in polycystic 1907–1918, 2009 kidney disease. JAmSocNephrol28: 2708–2719, 2017 26. Kretzler M, Teixeira VP, Unschuld PG, Cohen CD, Wanke R, Edenhofer I, 46. Jacobsen AV, Lowes KN, Tanzer MC, Lucet IS, Hildebrand JM, Petrie et al.: Integrin-linked kinase as a candidate downstream effector in EJ, et al.: HSP90 activity is required for MLKL oligomerisation and proteinuria. FASEB J 15: 1843–1845, 2001 membrane translocation and the induction of necroptotic cell death. 27. Guo L, Sanders PW, Woods A, Wu C: The distribution and regulation of Cell Death Dis 7: e2051, 2016 integrin-linked kinase in normal and diabetic kidneys. Am J Pathol 159: 47. Wickström SA, Lange A, Hess MW, Polleux J, Spatz JP, Krüger M, et al.: 1735–1742, 2001 Integrin-linked kinase controls microtubule dynamics required for 28. Engelman Mde F, Grande RM, Naves MA, de Franco MF, de Paulo plasma membrane targeting of caveolae. Dev Cell 19: 574–588, 2010 Castro Teixeira V: Integrin-linked kinase (ILK) expression correlates 48. Mamuya FA, Cano-Peñalver JL, Li W, Rodriguez Puyol D, Rodriguez with tumor severity in clear cell renal carcinoma. Pathol Oncol Res 19: Puyol M, Brown D, et al.: ILK and cytoskeletal architecture: An impor- 27 –33, 2013 tant determinant of AQP2 recycling and subsequent entry into the 29. Terpstra L, Prud’homme J, Arabian A, Takeda S, Karsenty G, Dedhar S, exocytotic pathway. Am J Physiol Renal Physiol 311: F1346–F1357, et al.: Reduced chondrocyte proliferation and chondrodysplasia in 2016 mice lacking the integrin-linked kinase in chondrocytes. JCellBiol162: 49. Legate KR, Montañez E, Kudlacek O, Fässler R: ILK, PINCH and parvin: 139–148, 2003 The tIPP of integrin signalling. Nat Rev Mol Cell Biol 7: 20–31, 2006 30. Ronzaud C, Loffing J, Bleich M, Gretz N, Gröne HJ, Schütz G, et al.: 50. McDonald PC, Fielding AB, Dedhar S: Integrin-linked kinase--essential Impairment of sodium balance in mice deficient in renal principal cell roles in physiology and cancer biology. JCellSci121: 3121–3132, 2008 mineralocorticoid receptor. J Am Soc Nephrol 18: 1679–1687, 2007 51. Ghatak S, Morgner J, Wickström SA: ILK: A pseudokinase with a unique 31. Wu H, Chen L, Zhou Q, Zhang X, Berger S, Bi J, et al.: Aqp2-expressing function in the integrin-actin linkage. Biochem Soc Trans 41: 995–1001, cells give rise to renal intercalated cells. J Am Soc Nephrol 24: 243–252, 2013 2013 52. Wickström SA, Lange A, Montanez E, Fässler R: The ILK/PINCH/parvin 32. Tsuji K, Paunescu TG, Suleiman H, Xie D, Mamuya FA, Miner JH, et al.: complex: The kinase is dead, long live the pseudokinase! EMBO J 29: Re-characterization of the glomerulopathy in CD2AP deficient mice by 281–291, 2010 high-resolution helium ion scanning microscopy. Sci Rep 7: 8321, 53. Vaynberg J, Fukuda K, Lu F, Bialkowska K, Chen Y, Plow EF, et al.: Non- 2017 catalytic signaling by pseudokinase ILK for regulating cell adhesion. 33. Vedovelli L, Rothermel JT, Finberg KE, Wagner CA, Azroyan A, Hill E, Nat Commun 9: 4465, 2018 et al.: Altered V-ATPase expression in renal intercalated cells isolated 54. Fukuda K, Gupta S, Chen K, Wu C, Qin J: The pseudoactive site of ILK is from B1 subunit-deficient mice by fluorescence-activated cell sorting. essential for its binding to alpha-Parvin and localization to focal adhe- Am J Physiol Renal Physiol 304: F522–F532, 2013 sions. Mol Cell 36: 819–830, 2009 34. Lu H, Sun TX, Bouley R, Blackburn K, McLaughlin M, Brown D: Inhibition 55. Fukuda K, Knight JD, Piszczek G, Kothary R, Qin J: Biochemical, pro- of endocytosis causes phosphorylation (S256)-independent plasma teomic, structural, and thermodynamic characterizations of integrin- membrane accumulation of AQP2. Am J Physiol Renal Physiol 286: linked kinase (ILK): Cross-validation of the pseudokinase. J Biol Chem F233–F243, 2004 286: 21886–21895, 2011

JASN 30: 2073–2090, 2019 ILK Deletion Causes PC Necroptosis 2089 BASIC RESEARCH www.jasn.org

56. Lalaoui N, Lindqvist LM, Sandow JJ, Ekert PG: The molecular rela- 63. Yu L, Moriguchi T, Kaneko H, Hayashi M, Hasegawa A, Nezu M, et al.: tionships between apoptosis, autophagy and necroptosis. Semin Cell Reducing inflammatory cytokine production from renal collecting duct Dev Biol 39: 63–69, 2015 cells by inhibiting gata2 ameliorates acute kidney injury. MolCellBiol 57. Cachat F, Lange-Sperandio B, Chang AY, Kiley SC, Thornhill BA, Forbes 37: e00211–e00217, 2017 MS, et al.: Ureteral obstruction in neonatal mice elicits segment-specific 64. Mulay SR, Linkermann A, Anders HJ: Necroinflammation in kidney tubular cell responses leading to nephron loss. Kidney Int 63: 564–575, disease. JAmSocNephrol27: 27–39, 2016 2003 65. Kaczmarek A, Vandenabeele P, Krysko DV: Necroptosis: The release of 58. Liu W, Chen B, Wang Y, Meng C, Huang H, Huang XR, et al.: RGMb damage-associated molecular patterns and its physiological relevance. protects against acute kidney injury by inhibiting tubular cell nec- Immunity 38: 209–223, 2013 roptosis via an MLKL-dependent mechanism. Proc Natl Acad Sci U S A 66. Martin-Sanchez D, Ruiz-Andres O, Poveda J, Carrasco S, Cannata-Ortiz 115: E1475–E1484, 2018 P, Sanchez-Niño MD, et al.: Ferroptosis, but not necroptosis, is im- 59. Havasi A, Borkan SC: Apoptosis and acute kidney injury. Kidney Int 80: portant in nephrotoxic folic acid-induced AKI. J Am Soc Nephrol 28: 29–40, 2011 218–229, 2017 60. Stupack DG, Cheresh DA: Get a ligand, get a life: Integrins, signaling 67. Friedmann Angeli JP, Schneider M, Proneth B, Tyurina YY, Tyurin VA, and cell survival. JCellSci115: 3729–3738, 2002 Hammond VJ, et al.: Inactivation of the ferroptosis regulator Gpx4 61. Dedhar S: Cell-substrate interactions and signaling through ILK. Curr triggers acute renal failure in mice. Nat Cell Biol 16: 1180–1191, 2014 Opin Cell Biol 12: 250–256, 2000 68. Müller T, Dewitz C, Schmitz J, Schröder AS, Bräsen JH, Stockwell BR, 62. Fujiu K, Manabe I, Nagai R: Renal collecting duct epithelial cells reg- et al.: Necroptosis and ferroptosis are alternative cell death pathways ulate inflammation in tubulointerstitial damage in mice. J Clin Invest that operate in acute kidney failure. Cell Mol Life Sci 74: 3631–3645, 121: 3425–3441, 2011 2017

2090 JASN JASN 30: 2073–2090, 2019 Supplemental materials

Supplemental table of contents 1 Supplemental Methods 2 Supplemental Table 1. List of primary antibodies used for immunoblotting (IB) and immunofluorescence staining (IF). 3 Supplemental Table 2. List of primers used for quantitative real-time PCR. 4 Supplemental Figure 1. Increased urinary NGAL secretion in 8 weeks old Ilk KO mice was revealed by immunoblotting, normalized to NGAL standard. 5 Supplemental Figure 2. ILK expression in wild type and PC Ilk KO kidney was revealed by immunofluorescence staining under low magnification. 6 Supplemental Figure 3. Immunofluorescence staining using isotype-matched immunoglobulin G (IgG) revealed minimal, non-specific background staining in the wild type and fibrotic Ilk KO kidney. 7 Supplemental Figure 4. Necrostatin-1 (Nec-1) treatment blocked membrane accumulation of MLKL in Ilk KO PCs by immunofluorescence staining viewed under low magnification.

Supplemental Methods Experimental animals All animal experiments were approved by the Massachusetts General Hospital Subcommittee on Research Animal Care, in compliance with the National Institutes of Health (NIH) Guide for the Care and Use of Laboratory Animals. All mice were on the C57BL/6J background. B6N;129- Ilktm1Star/J mice carrying loxP sites between exons 4 and 5 and after exon 12 were generated by Dr. René St-Arnaud from McGill University (Montréal, Canada) and are available from Jackson laboratory (Stock No: 023310). Transgenic mice harboring aquaporin 2 (Aqp2) Cre recombinase (Aqp2-Cre+) were generated by Dr. Günther Schütz’s group from the German Cancer Research Center (Heidelberg, Germany) and obtained from Dr. Wenzheng Zhang of Albany Medical Center (Albany, NY). To generate PC specific Ilk knockout mice, homozygous floxed Ilk male mice (Ilk fl/fl) were crossed with female Aqp2-Cre+ mice to generate Ilk fl/+; Aqp2-Cre+ mice. Ilk fl/+; Aqp2-Cre+ mice and Ilk fl/fl mice were then crossed to generate homozygous knockout (Ilk fl/fl; Aqp2-Cre+) mice. The littermates without Aqp2-Cre (Ilk fl/fl; Aqp2-Cre−) were used as control. For the necrostatin-1 (Nec-1, N9037, Sigma-Aldrich, St. Louis, MO) treatment experiment, Nec-1 was first dissolved in DMSO and then diluted in phosphate-buffered saline (PBS). 2 weeks old wildtype and Ilk fl/fl; Aqp2-Cre+ mice were given 1.65 mg/kg Nec-1 or PBS containing 1% DMSO through intraperitoneal injection every two days for 2 weeks. Mice were sacrificed at the age of 4 weeks. Blood and urine were collected at that time.

Kidney tissue preparation Kidney tissues were collected as following. Briefly, mice were anesthetized using isoflurane (3% inhalant). The right kidney was snap frozen in liquid nitrogen and stored at -80°C for protein and RNA analysis. The left kidney was fixed by perfusion through the left ventricle with PBS (0.137

M NaCl, 2.7 mM KCl, 10 mM Na2HPO4, 2 mM KH2PO4, pH 7.4) followed by paraformaldehyde lysine-periodate (PLP) fixative (4% paraformaldehyde, 75 mM lysine-HCl, 10 mM sodium periodate and 0.15 M sucrose in 37.5 mM sodium phosphate). Freshly perfusion-fixed kidneys were cut into 5 mm slices and incubated in PLP fixative overnight at 4°C. After washing with PBS for three times, kidney slices were kept in PBS containing 0.01% sodium azide at 4°C until use. For cryosections, kidney slices were immersed in 30% sucrose in PBS overnight at 4°C, and then embedded in Tissue-Tek OCT compound (Sakura Finetek, Torrance, CA). 4 μm cryosections were cut using a Leica CM3050S cryostat (Leica Biosystems, Buffalo Grove, IL), mounted on Superfrost Plus microscope slides (Fisher Scientific, Pittsburgh, PA) and stored at −20°C until use. For paraffin embedding, kidney slices were fixed in formalin overnight at 4°C followed by three washes with PBS. Slices were dehydrated in an ethanol gradient of 50%, 70%, 95% and 100%, then xylene, and finally immersed in paraffin through an automatic Leica TP1020 tissue processor (Leica Biosystems) over 24 hours. After embedded in paraffin, 3 μm sections were cut using a Leica RM 2025 rotary microtome (Leica Biosystems) and stored at room temperature until use.

Measurement of serum creatinine and blood urea nitrogen Blood samples were collected from mice at the time of tissue harvesting. Serum was stored in a −80°C freezer until use. Serum creatinine concentration was measured by modified Jaffe method using the QuantiChrom Creatinine Assay Kit (Cat. No. DICT-500, BioAssay Systems, Hayward, CA). Briefly, 30 μl serum, water (as blank) or standard (2 mg/dL) was added to each well of a 96- well plate in triplicates. Then 200 μl working reagent (mixed by 100 μl reagent A and 100 μl reagent

B) was added to each well. Optical density was measured at 0 min (OD0) and then at 5 min (OD5) using 510 nm of wave length. Serum creatinine concentrations were calculated using the formula provided by the manufacturer. Blood urea nitrogen (BUN) content was analyzed by Stanbio Urea Nitrogen Kit (Stanbio Laboratory, Boerne, TX) according to manufacturer’s instruction. Briefly, 20 μl serum, water or standard solution was added to each tube in triplicates. Then 1 ml BUN color reagent and 2 ml BUN acid reagent were added to each tube. After incubating in boiling water bath for 10 min, tubes were immediately immersed in cold water for 3 min. Optical density was measured at 520 nm. Serum BUN concentrations were calculated according to manufacturer’s instruction.

Immunofluorescence staining and immunoblotting Paraffin embedded kidney sections were deparaffinized in xylene and rehydrated with a graded series of ethanol (100% and 95%, 80%, and 70% ethanol in distilled water) and water. After treatment with Tris-EDTA buffer (10 mM Tris Base, 1 mM EDTA, 0.05% Tween 20, pH 9.0) at 95°C for 20 min, kidney sections were subjected to immunostaining. For immunofluorescence staining of cryosections, slides were rehydrated in PBS and incubated with 1% SDS in PBS for 4 min followed by washing with PBS for three times. Both deparaffinized sections and cryosections were blocked with 1% (w/v) bovine serum albumin in PBS, and incubated with primary antibodies of various dilutions from 1:100 to 1:3000 (Supplemental Table 1). Meanwhile, corresponding normal rabbit IgG (A6154, Sigma-Aldrich) and mouse IgGκ (sc-516176, Santa Cruz, CA) were used to serve as negative control. After washing with PBS for three times, slides were incubated with corresponding fluorophore-conjugated secondary antibodies (Jackson ImmunoResearch Laboratories, West Grove, PA) at room temperature for 1 hour. After three more washes with PBS, slides were mounted and viewed under a Nikon Eclipse 90i epifluorescence (Nikon Instruments, Melville, NY) or a Zeiss LSM800 confocal microscope with Airyscan super-resolution capabilities (Carl Zeiss Microscopy, Thornwood, NY). Analysis of fluorescence intensity was performed using ImageJ software (NIH; https://imagej.nih.gov/). Immunofluorescence staining of kidney with MLKL, CD68 and Ly6G was performed on cryosections, while immunostaining with other primary antibodies was performed on paraffin embedded sections. Quantification of the number of F4/80 or Ly6G positive cells was performed manually in 10 randomly selected fields (magnification: 400 x) for each specimen. Data presented as the number of F4/80 or Ly6G positive cells per high power field (HPF) with a magnification of 400 x. Immunofluorescence staining of cultured cells was performed similarly as kidney sections with

3 minor modification. Cells grown on glass coverslips were fixed with 4% paraformaldehyde in PBS for 20 min at room temperature and washed with PBS for three times. After permeabilization with 0.01% Triton X-100 in PBS for 4 min, they were subjected to routine immunostaining for MLKL or phospho-MLKL as described above. Immunoblotting of kidney tissue lysate or cell lysate was performed as follows. Briefly, kidney or cell lysates were homogenized in lysis buffer (16 mM NaF, 8 mM Na3VO4, 0.5% NP-40, 0.1% Triton-X-100, 0.1% SDS in PBS) containing protease inhibitor cocktail (Bimake, Houston, TX). Lysates were obtained by centrifugation at 12,000 g for 15 min at 4°C. Total protein was measured using the BCA assay following manufacturer’s instruction (Pierce, Thermo Scientific, Waltham, MA). Protein lysates were separated by SDS-polyacrylamide gel electrophoresis, then electrophoretically transferred to polyvinylidene difluoride membranes (MilliporeSigma, Burlington, MA) using a semi-dry transfer cell (Bio-Rad, Hercules, CA). After blocking with 5% non-fat milk in Tris-buffered saline-0.1% Tween (TBS-T), membranes were incubated with primary antibodies of various concentration (Supplemental Table 1) at 4°C overnight. After washing with TBS-T, the horseradish peroxidase (HRP) conjugated secondary antibodies were applied to membranes (Santa Cruz). Finally, membranes were developed using the ECL Pro Kit (PerkinElmer, Waltham, MA) and exposed to autoradiographic film (Santa Cruz). Relative intensity of protein expression was quantified by optical density using the Quantity One software (Bio-Rad). To detect urinary NGAL by immunoblotting, 3 μl spot urine collected from each mouse was mixed with 5 μl SDS-PAGE sample loading buffer, then subjected to routine SDS-PAGE and immunoblotting. 2 g of purified recombinant NGAL protein (230-3002s, Raybiotech, Peachtree Corners, GA) was loaded as a standard for quantification.

Hematoxylin and eosin (H&E) staining H&E staining was conducted as follows. Briefly, paraffin-embedded kidney sections were deparaffinized, rehydrated and nuclei were stained with Weigert’s iron hematoxylin (Sigma-Aldrich). Sections were then rinsed in running tap water and differentiated with 0.3% acid alcohol, followed by 2 min of incubation with Scott’s tap water (238 mM sodium bicarbonate, 29 mM magnesium sulphate in distilled water). The sections were then stained with 1% eosin (Sigma-Aldrich) for cytosol staining. The slides were finally dehydrated and mounted with cytoseal 60 mounting media (Thermo). Nuclei were stained dark brown, while the cytoplasm and extracellular matrix were stained pink. Tubular injury score was calculated based on the percentage of tubules in kidney sections that displayed tubular damage, including tubular dilation or atrophy, tubular necrosis, loss of the brush border, or cast formation. The degree of injury was graded semi quantitatively and blindly by two independent researchers from 10 randomly chosen fields of each kidney section. The score was given according to the extent of injury involved in each field as follows: 0, normal; 1, involvement less than 10%; 2, involvement up to 11 to 25%; 3, involvement up to 26 to 75%; 4, extensive damage involving more than 75% of the observed tubules.

Picrosirius red staining Paraffin-embedded kidney sections were stained using the Picrosirius red stain kit following manufacturer’s instruction (24901, Polysciences Inc, Warrington, PA). Briefly, after deparaffinization and rehydration, paraffin-embedded kidney sections were stained with Weigert’s iron hematoxylin for 8 min, and incubated with solution A for 2 min at room temperature. Sections were then rinsed in water and incubated with solution B for 60 min, followed by 2 min of incubation with solution C. The slides were then immersed in 70 % ethanol for 45 sec, dehydrated and mounted using cytoseal 60 mounting media. The Picrosirius red stained kidney sections were examined by a Zeiss LSM800 confocal microscope (Carl Zeiss Microscopy) under normal and polarized light. The staining for type I collagen fibers stains yellow to orange, and type III collagen stains green under polarized light.

Masson’s trichrome staining Paraffin-embedded kidney sections were stained using the Trichrome stains kit (Masson, Sigma-Aldrich) according to manufacturer’s instruction. Briefly, paraffin-embedded kidney sections were deparaffinized, rehydrated and incubated with preheated Bouin’s solution at 56 °C for 15 min. The sections were then washed in tap water, followed by staining with Weigert’s iron hematoxylin for 5 min. Sections were washed in tap water for 5 min, rinsed in deionized water and then stained in Biebrich Scarlet-Acid Fucshin for 5 min. After incubating with working phosphotungstic/phosphomolybdic acid solution for 5 min, sections were placed in Aniline blue solution for 5 min, followed by 2 min of incubation with 1% acetic acid. Finally, sections were dehydrated and mounted. The Masson’s trichrome stained kidney sections were examined by the Zeiss LSM800 confocal microscope (Carl Zeiss Microscopy). In Masson’s trichrome staining, collagen fibers stained blue, cytoplasm stained pink, and nuclei stained dark brown.

Transmission electron microscopy Transmission electron microscopy (TEM) was performed as follows. Briefly, PLP fixed kidney slices were post-fixed with 2% glutaraldehyde in 0.1 M sodium cacodylate buffer at 4°C for 24 hours. The slices were incubated with 1% osmium tetroxide in cacodylate buffer at room temperature for 1 hour, followed by several rinses in 0.1 M sodium cacodylate buffer. After dehydration in a graded ethanol series from 50% to 100%, followed by brief dehydration in propylene oxide, kidney slices were infiltrated with Eponate resin (Ted Pella, Redding, CA) and embedded in fresh Eponate at 60°C. Ultra-thin sections were cut, stained with uranyl acetate and lead citrate, and examined under a JEOL JEM 1011 transmission electron microscope (JEOL, Peabody, MA). Images were taken using an AMT digital imaging system (Advanced Microscopy Techniques, Danvers, MA).

RNA isolation and Quantitative real time-PCR Mouse kidney RNA was extracted using QIAshredder and RNeasy purification kits (Qiagen, Valencia, CA). The synthesis of cDNA from mRNA was performed following the protocol of the

High-Capacity RNA-to-cDNA™ Kit (Applied Biosystems, Foster City, CA). Quantitative real-time PCR was carried out using a QuantStudio 3 Real-Time PCR system (Thermo Fisher Scientific, Waltham, MA) and the Power up SYBR Green PCR Master Mix (Life Technology, Carlsbad, CA). The level of specific mRNA was normalized to the level of housekeeping gene β-actin expression. The relative mRNA expression was determined by the 2-∆∆Ct method. The sequences of primers that used for the study are summarized in Supplemental Table 2.

Cell culture LLC-PK1 cells and mCCDC11 cells were cultured in Dulbecco’s modified Eagle medium

(DMEM, Corning, Corning, NY) containing 10% fetal bovine serum in a 5% CO2/95% air humidified atmosphere at 37°C. Cells were trypsinized in 0.25% trypsin-EDTA, and plated on standard tissue culture dishes for adherent cell (353002, Corning). Cells were allowed to grow for 24 hours, then an ILK inhibitor, cpd22 of 0.5, 1 and 2 μM was added and incubated with cells for 24 hours (407331, Calbiochem, San Diego, CA). After treatment for 24 hours, cells were harvested and lysed in lysis buffer, and prepared for electrophoresis (SDS-PAGE) and immunoblotting. All experiments were repeated at least three times.

Cell viability assay Two assays were performed to assess cell viability, including MTT (3-(4,5-dimethylthiazol-2- yl)-2,5-diphenyltetrazolium bromide, Affymetrix, Santa Clara, CA) assay and cellular DNA content quantification assay. The LLC-PK1 were seeded on 96 well plates and grown to approximately 80% confluence. Cells were then treated with 0.2% DMSO, ILK inhibitor cpd22 in the presence or absence of the necroptotic inhibitor necrostatin-1 (Nec-1, 50 μM). For MTT assay, 24 hours after treatment, MTT was added and incubated with cells at 37°C for 3 hours. After removal of MTT, DMSO was added to each well to dissolve the formazan crystals. Absorbance was measured at 548 nm on a microplate reader (model DTX880, Beckman-Coulter, Fullerton, CA). Relative cell viability was presented as the percentage of treated group over control group, as OD treat/OD control x100%. Cell viability was also assessed by measuring cellular DNA content using the CyQUANT NF Assay Kit (C35006, Invitrogen). 24 hours after treatment as mentioned above, cells in a 96 well plate were incubated with the DNA binding dye at 37°C for 1 hour. Cellular DNA content was analyzed by measuring the fluorescence intensity by the SpectraMax M5 Multi-Mode microplate reader (Molecular Devices, San Jose, CA) with excitation at 485 nm and emission detection at 530 nm. Relative cell viability was presented as the percentage of treated group over control group, as fluorescence treat / fluorescence control x100%. siRNA transfection ILK siRNA (sense sequence: AAGGACACAUUCUGGAAGGGG, antisense sequence: CCUUCCAGAAUGUGUCCUUGG) was purchased from GE Dharmacon (Lafayette, CO).

6 mCCDC11 cells were seeded in 6-well plate and transfected with ILK siRNA (50 nM) or scrambled siRNA mixed with 5 μl Lipofectamine 2000 (Invitrogen) in 1 ml DMEM for 6 hours. 24 hours after transfection, some control cells (transfected with or without scrambled siRNA) were incubated with 2 μM cpd22, and 1 ng/ml recombinant TNF-α (300-01A, PeproTech) respectively. Cell viability assay and immunofluorescence staining were performed 48 hours after ILK siRNA transfection.

Statistical analyses Data are shown as mean ± standard error, SE, of independent replicates (n ≥ 3). Experimental data were analyzed with Student’s t-test for two groups or with one-way ANOVA for multiple groups using GraphPad Prism version 5.01 (GraphPad Software, San Diego, CA). A P value < 0.05 was considered statistically significant (*P < 0.05, **P < 0.01, ***P < 0.001). Individual P values are specified in figure legends.

Supplemental Table 1. List of primary antibodies used for immunoblotting (IB) and immunofluorescence staining (IF). Antigen Manufacturer Catalog Species Dilution(IB) Dilution(IF) AQP2 Santa Cruz sc-9882 goat - 1:1800 ILK Cell Signaling 3862s rabbit 1:2000 1:1000 NGAL R&D Systems AF-1857 goat 1:1000 - α-SMA Sigma A5228 mouse 1:1500 1:3000 collagen I Sigma C2456 mouse - 1:1000 fibronectin Sigma F3648 rabbit 1:5000 1:2000 p-Smad3(Ser423/425) Abcam ab51451 rabbit 1:1000 - Smad2/3 Santa Cruz sc-133098 mouse 1:2000 - GAPDH Cell Signaling 2118s rabbit 1:10000 - NF-κB p65 Cell Signaling 8242 rabbit 1:1000 - p-NF-κB p65(Ser536) Cell Signaling 3033s rabbit 1:1000 - F4/80 Invitrogen 14-4801- rat - 1:200 85 Ly6G Biolegend 127601 rat - 1:100 MLKL EMD MABC60 rat 1:1000 - Millipore 4 MLKL ABGENT AP14272b rabbit - 1:500 -ev p-MLKL (Ser345) Abcam ab196436 rabbit - 1:1000 RIPK3 Santa Cruz sc-374639 mouse 1:1000 - p-RIPK3 Cell Signaling 57220s rabbit - 1:200 (Thr231/Ser232) p-RIPK1 (Ser166) Cell Signaling 31122s rabbit - 1:400 CD68 Novus NB100- mouse - 1:400 683 V-ATPase Provided by - chicken - 1:800 Dr. Dennis Brown Akt Cell Signaling 4691s rabbit 1:3000 - p-Akt (Ser 473) Cell Signaling 4058s rabbit 1:4000 - β-actin Cell Signaling 4967s rabbit 1:10000 -

Supplemental Table 2. List of primers used for quantitative real-time PCR. Gene Forward Reverse Fibronectin 5’-CTG GAG TCA AGC CAG ACA 5’-CGA GGT GAC AGA GAC CAC CA-3’ AA-3’ Collagen 5’-GCT CCT CTT AGG GGC CAC 5’-CCA CGT CTC ACC ATT GGG G- IA1 T-3’ 3’ TGF-βi 5’-CCT CAC CTC CAT GTA CCA 5’-TGG AAA TGA CCTTGT GAA-3’ CAATGA GAG-3’ TNF-α 5’- CCC TCA CAC TCA GAT CAT 5’- GCT ACG ACG TGG GCT ACA G CTT CT -3’ -3’ IL-6 5’- CCA AGA GGT GAG TGC TTC 5’- CTG TTG TTC AGA CTC TCT CC CC -3’ CT -3’ IL-1β 5’- GCA ACT GTT CCT GAA CTC 5’- ATC TTT TGG GGT CCG TCA AAC T -3’ ACT -3’ MLKL 5’-ACA ATG AGT GTG CGC AGC 5’-CTA CGA GGA AAC TGG AGC CTC-3’ TGC TG-3’ RIPK3 5’-ACA GGC CAT CCT TCC AGG 5’- GCT TGG CTC TCT GGC AGA AC-3’ CAA G -3’ IL-33 5′-TCC TTG CTT GGC AGT ATC 5′-TGC TCA ATG TGT CAA CAG CA-3′ ACG-3′ CXCL1 5′-CTG GGA TTC ACC TCA AGA 5′-CAG GGT CAA GGC AAG CCT C ACA TC-3′ -3′

Supplemental Figure 1. (A) Increased urinary NGAL secretion in 8 weeks old KO mice was revealed by immunoblotting. 2 μg of purified recombinant NGAL (Std rNAGL) was loaded as a standard for calculation. (B) Urinary secretion of NGAL in 8 weeks wild type and KO mice was quantified by densitometry and normalized to the expression of Std rNAGL.

Supplemental Figure 2. ILK expression (red) in wild type (WT) and PC Ilk KO kidney was revealed by immunofluorescence staining under low magnification. AQP2 stained purple in PCs, and membrane stained green by WGA-FITC. Arrowheads indicate loss of basal ILK signal in AQP2 positive PCs. Arrows indicate intact ILK staining in non-CD tubules in Ilk KO kidney. Scale bar=10 μm.

Supplemental Figure 3. Immunofluorescence staining using isotype-matched immunoglobulin G (IgG) revealed minimal, non-specific background staining in the wild type and fibrotic Ilk KO kidney. Immunostaining of α-SMA, collagen I (A) and fibronectin (B) signals (red) was significantly increased in interstitium in Ilk KO kidney compared with wild type kidney. Immunostaining was performed using serial sections of the same kidney. Scale bar=20 μm.

Supplemental Figure 4. Necrostatin-1 (Nec-1) treatment blocked membrane accumulation of MLKL in Ilk KO PCs by immunofluorescence staining viewed under low magnification. AQP2 positive PCs stained green, and MLKL stained red. Arrows highlighted collecting ducts (CDs) that were stained positive for AQP2 in Ilk KO kidney. Scale bar=20 μm.