Abstract

   Cisplatin is one of the major causes of acute kidney injury (AKI) in
   clinical practice, and ferroptosis is an essential form of cell death
   in cisplatin-induced AKI (CP-AKI). WW domain binding protein-2 (WBP2),
   a molecular chaperon, is involved in the progression of various
   malignancies, but its role in renal injuries has not been investigated.
   Our present study employed bioinformatics analysis to identify WBP2 as
   a potential modulator of AKI and ferroptosis. Preliminary laboratory
   investigations showed that WBP2, highly expressed in renal proximal
   tubular cells, was downregulated in CP-AKI. Further studies
   demonstrated that WBP2 decelerated ferroptosis to alleviate CP-AKI.
   Mechanistically, WBP2 interacted with glutathione peroxidase 4 (GPX4, a
   key detoxicating enzyme for ferroptosis) via its PPXY1 motif to inhibit
   ferroptosis. Furthermore, the in-depth investigations revealed that
   WBP2 competed with heat shock cognate protein 70 (HSC70) for the
   binding with the KEFRQ-like motifs of GPX4, leading to the deceleration
   of chaperon-mediated autophagy of GPX4. All in all, this study
   indicated the beneficial effect of WBP2 in CP-AKI and its relevance
   with ferroptosis, thus providing a novel insight into the modulation of
   ferroptosis in cisplatin-related nephropathy.

   Keywords: WBP2, GPX4, Ferroptosis, Cisplatin, Acute kidney injury

Graphical abstract

   [45]Image 1
   [46]Open in a new tab

Highlights

     * •
       WBP2, ferroptosis, GPX4, chaperon-mediated autophagy.

1. Introduction

   Cisplatin is a widely used chemotherapeutic drug for various solid
   malignancies, such as prostate cancer, testicular cancer, and ovarian
   cancer [[47]1]. However, the clinical administration of cisplatin is
   largely hindered by its commonly observed nephrotoxicity. It is
   estimated that about 30% of cisplatin (high dose)-treated patients
   presented different manifestations of renal injuries, and its
   nephrotoxic effects can be accumulated after repeated use [[48]2]. The
   major target of cisplatin-induced AKI (CP-AKI) is renal proximal
   tubules, where most cisplatin in the glomerular ultrafiltrate is
   reabsorbed [[49]3]. Therefore, renal proximal tubular cells were
   constantly used in the laboratory investigations of CP-AKI. Previous
   publications revealed that various cell death pathways were
   incorporated in the progression of CP-AKI [[50]4], but their exact
   regulatory machinery has not been fully elucidated; thus, no effective
   treatments were available in the clinical practice.

   Ferroptosis is a type of regulated cell death identified in 2012, and
   it was initially used to describe erastin-induced cell death [[51]5].
   Later on, it was discovered that ferroptosis was operative in various
   pathophysiological scenarios, including ischemia/reperfusion injury,
   degenerative disorder, and chemotherapy of malignancies [[52][6],
   [53][7], [54][8]]. Unlike other types of regulated cell death,
   ferroptosis is characterized by decreased crista and condensed inner
   membrane in mitochondrial morphology, iron-catalyzed excessive lipid
   peroxidation, and the independency of caspase [[55]9]. When cells were
   under overwhelming ferroptotic pressure, the key terminator of
   ferroptosis GPX4, together with glutathione (GSH), was involved to
   catalyze lethal lipid hydroperoxides to non-toxic lipid alcohols,
   leading to the inhibition of ferroptosis [[56]10]. The role of
   ferroptosis in renal injuries has been investigated in multiple acute
   or chronic pathological states [[57]11,[58]12], and our previous
   investigation presented the initial evidence that ferroptosis was an
   integral process in CP-AKI [[59]13]. However, the modulation of
   ferroptosis was poorly investigated in it.

   WW domain binding protein-2 (WBP2) was initially identified as a
   chaperon of Yes-associated Protein (YAP), and their interaction led to
   the activation of estrogen and progesterone receptors [[60]14,[61]15].
   Recent studies have revealed multiple binding partners of WBP2 and its
   pivotal role in various aspects of signaling transduction, such as
   steroid signaling pathway, Wnt signaling pathway, and Hippo signaling
   pathway [[62][15], [63][16], [64][17], [65][18]]. Notably, three PPXY
   motifs are present in the C-terminal domain of WBP2, which dominates
   the interaction with its targeted proteins [[66]19]. The oncogenic
   property of WBP2 has been well described in various malignancies, and
   recent publications showed that WBP2 was also incorporated in the
   pathogenesis of hearing loss, fertility disorders, and hepatic
   steatosis [[67][20], [68][21], [69][22]]. The role of WBP2 in renal
   diseases has not been investigated. This present work used
   bioinformatics analysis to identify WBP2 as a latent regulator of
   AKI-related ferroptosis. Further investigations showed that WBP2
   interacted with and stabilized GPX4 to inhibit ferroptosis in CP-AKI,
   suggesting that WBP2 might be a novel therapeutic target for it.

2. Methods

   Bioinformatics. Gene expression profiles of AKI were filtrated through
   the Gene Expression Omnibus (GEO) database
   ([70]http://www.ncbi.nlm.nih.gov/geo). Inclusion criteria were as
   follows: [1] Availability of kidneys from AKI patients in the dataset;
   [2] five or more kidney samples in the dataset; [3] Samples of kidney
   transplant patients undergoing acute rejection. Three eligible datasets
   were selected, including [71]GSE1563 and [72]GSE30718, and
   [73]GSE61739.

   All microarray data were submitted to the GEO database
   ([74]http://www.ncbi.nih.gov/geo). The raw data were downloaded as
   MINiML files. It contains the data for all platforms, samples, and GSE
   records. The extracted data were normalized by log2 transformation. The
   microarray data were normalized by the normalized quantiles function of
   the preprocess Core package in R software (version 4.4.2). Probes were
   converted to gene symbols according to the annotation information of
   the normalized data in the platform. Probes matching multiple genes
   were removed from these datasets. As in different datasets or the same
   dataset but in different platforms, extracting multiple data sets with
   common gene symbols, marking different datasets or platforms as
   different batches, used the remove Batch Effect function of the limma
   package in the R software to remove batch effects [[75]23]. The result
   of the data preprocessing was assessed by Density plot, and the UMAP
   plot was drawn to illustrate the samples before and after batch effect.

   The weighted gene co-expression network analysis (WGCNA) was performed
   using the R package WGCNA (version 1.70) [[76]24]. Before analysis, the
   hierarchical clustering analysis was performed using the Hclust
   function in R language to exclude the outlier samples. Subsequently,
   the appropriate soft powers β (ranged from 1 to 20) was selected using
   the function of “pickSoftThreshold” in the WGCNA package according to
   the standard of scale-free network. Next, the soft power value β and
   gene correlations matrix among all gene pairs calculated by Pearson
   analysis were used to build adjacency matrix, which was calculated by
   the formula: aij = |Sij|β (aij: adjacency matrix between gene i and
   gene j, Sij: similarity matrix which is composed of Pearson correlation
   coefficients of all gene pairs, β: soft power value). After choosing
   the power of 6, the adjacency was transformed into a topological
   overlap matrix (TOM), which could measure the network connectivity of a
   Gene defined as the sum of its adjacency with all other Genes for
   network Gene ration, and the corresponding dissimilarity (1-TOM) was
   calculated. To classify Genes with similar expression profiles into
   Gene modules, average linkage hierarchical clustering was conducted
   according to the TOM-based dissimilarity measure with a minimum size
   (Gene group) of 100 for the Genes dendrogram. To further analyze the
   module, we calculated the dissimilarity of module eigen Genes, chose a
   cut line of 0.5 for module dendrogram and merged some module. Finally,
   we obtained 8 co-expression modules. In this study, the soft threshold
   β was 2 in the WGCNA analysis of AKI. The other parameters were the
   following: networkType = “unsigned”, minModuleSize = 100,
   mergeCutHeight = 0.50 and deepSplit = 2.

   For network visualization, the Cytoscape.js library was used as
   previously reported [[77]25]. In order to uncover the biological
   function related to the network nodes, Gene Ontology (GO) analysis and
   Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment
   analysis were performed using clusterprofiler R package [[78]26].

   Cell culture and treatment. BUMPT cells, mouse renal proximal tubular
   cell line, were initially obtained from Drs. William Lieberthal and
   John Shwartz at Boston University. Cell culture media, containing DMEM,
   10% FBS, and antibiotics, were used to culture the cells in a
   humidified environment. The cell culture incubator had 5% CO[2] and
   37 °C in temperature. Lentivirus was used to manipulate the expression
   of WBP2 in BUMPT cells. HK-2 cells (human proximal tubular cell line)
   and HMC cells (human mesangial cell line) were also used in our study.
   HK-2 cell and HMC cells were maintained in DMEM/F12 cell culture media
   and the identical culture environment was used for them. The
   incorporated lentiviruses in this study were: WBP2 overexpression
   lentivirus (Lv-WBP2), WBP2 knockdown lentivirus (Lv-shWBP2), WBP2
   overexpression lentivirus for HK-2 cells (Lv-WBP2-HK-2), WBP2 knockdown
   lentivirus for HK-2 cells (Lv-shWBP2-HK-2), and their empty vector
   viruses (Lv-NC and Lv-shNC). Cells were seeded into different plates
   and treated with 20 μM cisplatin (Sigma-Aldrich, #P4394) for 12–24 h.
   Cisplatin used in vitro studies was initially dissolved in
   dimethylformamide (DMF) before adding into the culture media. For
   several biochemistry experiments, 293t cells were used to increase the
   transfection rate. The same culture media were used for 293t cells.
   Other treatments used in this study were: Fer-1(Sigma-Aldrich,
   #SML0583, 0.4 μM), myoglobin (Sigma-Aldrich, #M1882, 10 mg/ml),
   Chloroquine (MCE, #HY-17589A,20 mM), NH[4]Cl (Sigma-Aldrich,
   #213330,20 μM), and MG132 (MCE, #HY-13259, 15 μM). For
   hypoxia/reoxygenation treatment, cells were maintained in a hypoxia
   chamber for 12 h and they were then cultured in a normal environment
   for 6 h.

   Animals. 8–12 weeks old C57BL/6J male mice were used in vivo studies,
   and the mice were divided into multiple groups accordingly (each group
   had 8-10 mice). For the induction of CP-AKI, cisplatin was dissolved in
   PBS solution at the concentration of 1 mg/ml, and cisplatin solution
   was intraperitoneally injected into the mice at the dose of 25 mg/kg.
   Mice were sacrificed two days later, and the kidneys and blood samples
   were harvested. For the inhibition of ferroptosis, Fer-1 (5 mg/kg) was
   administrated intraperitoneally 45 min before the injection of
   cisplatin. Intraparenchymal injection of adenovirus solution in the
   renal cortex was used to manipulate the expression of WBP2 in mice.
   Briefly, the mouse was anesthetized, and the left kidney was exposed.
   The left renal pedicle was clamped, and three sites were selected for
   adenovirus injection with a 31-gauge needle (30 μL adenovirus solution
   in each site). The needle was maintained for 15 s after each injection
   to reduce the volume of leakage. Besides, the clamp was maintained for
   another 5 min after the removal of the needle to ensure the permeation
   of adenovirus. The duration of renal artery occlusion was about 8 min
   for each kidney. The adenovirus used in this study was at the
   concentration of 1.5–2 ✕ 10^12 particles/ml. The incorporated
   adenoviruses in this study were: WBP2 overexpression adenovirus
   (Ad-WBP2), WBP2 knockdown adenovirus (Ad-shWBP2), and their empty
   vector adenoviruses (Ad-NC and Ad-shNC).

   Preparation of human samples: Patients pathologically diagnosed with
   acute tubular necrosis (ATN) were incorporated into the AKI group, and
   their renal biopsy sections were used for further analysis. Patients
   undergoing radical nephrectomy were classified into the control group.
   Normal renal tissues (at least 4 cm away from the border of the tumor)
   were embedded into the paraffin, and the renal sections were obtained.

   For the preparation of human primary tubular cells (HPT cells), normal
   renal cortex (at least 4 cm away from the border of the tumor) was
   immediately obtained and stored in cold Hanks solution after the
   surgery. The renal samples were cut with a scissor, and they were then
   transferred into a 15 ml tube filled with pre-heated Hanks solution.
   The renal samples in the tube were then subjected to
   centrifuge-mediated purification twice (1000 rpm, 5 min). The
   supernatants were removed, and red ball cells were erased. Collagenase
   solution (1 mg/ml) was used to digest the samples for 1 h at 37 °C, and
   intermittent shanking was used to aid in the digestion. The mixture was
   then transferred into a 50 ml tube, and Hanks solution was used to
   resuspend the mixture. Subsequently, the mixture was purified with a
   70 μm-filter in the centrifuge (1000 rpm, 5 min). The supernatant was
   removed, and the cells were seeded into the collagen-coated plates.
   Cells were maintained in a humidified environment for 48 h to assure
   attachment. When cell fusion was observed (usually 5–7 days), HPT cells
   were identified by CK18 staining, and cisplatin treatment (20 μM, 20 h)
   was applied.

   Morphological studies. For in vitro studies, cells were immediately
   scanned by light microscopy after the treatment. For in vivo studies, H
   & E staining was used. 4-μm paraffin sections were deparaffinized and
   rehydrated with routing protocols. The sections with stained with
   hematoxylin for 3 min and eosin for 30 s. After the rinse by tap water,
   the sections were mounted and evaluated by light microscopy.

   CCK8 assay. CCK8 assay was conducted to evaluate the cell survival
   rate. About 5000 cells were seeded into each well of the 96-well plate,
   and cells were maintained overnight for attachment. Cells were treated
   with cisplatin or RSL3 for an appropriate time before the removal of
   cell culture media. 200 μL CCK solution (TargetMol, #C0005) was added
   in each well, and the plate was maintained at 37 °C for 30 min to
   complete the reaction. The plate was subjected to 5 min’ shank, and OD
   450 nm reading was obtained from the microplate reader.

   Renal function analysis. For the evaluation of renal function, serum
   creatinine and blood urea nitrogen (BUN) levels were detected. A Serum
   Creatinine Colorimetric Assay Kit (Jiancheng, China, #C011-2-1) was
   used to detect serum creatinine levels as requested by the instruction.
   Briefly, 6 μL double distilled water (used as the blank group), serum
   samples, or standards were added in the wells containing 180 μL Enzyme
   Solution A. The 96-well plate was then maintained at 37 °C for 5 min,
   and OD 546 nm readings were obtained. Subsequently, 60 μL Enzyme
   Solution B was added into each well, and the mixture was harbored in
   the environment of 37 °C for 5 min. Then, OD 546 nm readings were
   obtained for another time. The concentration of serum creatinine was
   then calculated accordingly. For the evaluation of serum BUN levels,
   BUN assay kits (Jiancheng, China, #C013-2-1) was used according to its
   instruction. Briefly, 20 μL double distilled water (used as the blank
   group), serum samples or standards were mixed with 250 μL urease
   solution, and the reaction was maintained at 37 °C for 10 min. After
   that, 1 ml phenol solution and 1 ml sodium hypochlorite solution were
   added, and the mixtures were maintained in 37 °C water bath for 10 min.
   Finally, OD 640 nm readings were obtained, and the data was calculated
   accordingly.

   Evaluation of ROS generation. DHE staining was used to evaluate the
   overall ROS generation. BUMPT cells and frozen sections were stained
   with 10 μM DHE solution at 37 °C for 30 min, and the stained sections
   were examined by fluorescence microscopy. For lipid ROS evaluation in
   vitro, C11 BODIPY 581/591 kit (ThermoFisher, #[79]C10445) was used to
   stain the cells accordingly. Briefly, cells were stained with 10 μM
   Image-iT® Lipid Peroxidation Sensor (Component A) for 30 min at 37 °C
   after the treatment. After that, the media were removed, and the cells
   were washed with PBS three times. Finally, the cells were subjected to
   fluorescence microscopic analysis and quantification. For lipid ROS
   evaluation in vivo, immunofluorescence staining of 4-HNE was applied to
   the renal sections, and quantification was applied accordingly.

   Western blot studies. Renal cortex or cells were lysed by RIPA buffer
   containing proteinase inhibitors. BCA assay (Takara, #T9300A) was used
   to evaluate the protein concentration, and equal amounts of protein
   were loaded into the SDS-PAGE gel. Electrophoresis was performed to
   separate the proteins, which were then transferred onto the PVDF
   membranes. The membranes were blocked by 10% milk for 1 h at room
   temperature, and they were then incubated with primary antibody
   dilutions overnight at 4 °C. The membranes were washed and incubated
   with secondary antibody dilutions for 1–2 h at room temperature. After
   three times of washing, the membranes were finally evaluated by ECL
   chemiluminescence system. The incorporated antibodies were:
   WBP2(Proteintech, #12030-1-AP and Santa Cruz, #sc-514247, 1:1000),
   NCOA4 (Bethyl Laboratories, #A302-272A, 1: 1000), GPX4 (Abcam,
   #ab125066, 1: 1000), FTH1 (Abcam, #ab183781, 1: 1000, and Cell
   Signaling Technology, #4393,1:1000), KIM-1 (R&D Systems, #AF1817,
   0.25 μg/mL), NGAL (R&D Systems, #AF1857, 0.25 μg/mL), Lamp1 (Abmart,
   #TD7033S, 1: 1000), Flag (Sigma-Aldrich, #F1804, 1: 5000), HA (Santa
   Cruz, #sc-7392, 1: 1000), Myc (Proteintech, #16286-1-AP, 1: 5000),
   HSC70 (Proteintech, #10654-1-AP, 1: 5000), Lamp2a(Proteintech,
   #66301-1-AP, 1: 1000), GFP (Proteintech, #66002-1-Ig, 1: 5000), β-Actin
   (Proteintech, #66009-1-Ig, 1: 5000), secondary antibody anti-Mouse
   (Proteintech, #SA00001-1, 1:5000), secondary antibody anti-Rabbit
   (Proteintech, #SA00001-2, 1:5000), and secondary antibody anti-Goat
   (Proteintech, #SA00001-4, 1:5000),

   qRT-PCR analysis. Cells were homogenized by TRIzol Reagent (Invitrogen,
   catalog 15-596-026), and the routing protocols were used to exact mRNA
   from the cells. The remaining gDNA was erased, and the cDNA was
   obtained from the mRNA by using a TAKARA kit (TaKaRa, #RR037A). Equal
   amounts of cDNA were subjected to quantitative PCR reaction by using a
   Vazyme kit (Vazyme, #Q111-02). The Cq values of each sample were
   obtained and calculated accordingly. β-actin was used as the internal
   control, and ΔΔCq values were used for the statistical analysis. The
   primers used in our studies were indicated in the [80]Supplemental
   Table 1.

   Immunohistochemistry staining. 4-μm paraffin sections were obtained
   after harvesting the renal samples. The sections were heated and
   deparaffinized before the rehydration. After washing, heat-induced
   antigen retrieval was applied to the sections, and endogenous
   peroxidase was suppressed. Goat serum was used to incubate the sections
   for 1 h at room temperature to block the background. Subsequentially,
   the sections were incubated with primary antibody dilutions overnight
   at 4^oC. After three times of washing, the sections were incubated with
   secondary antibody dilutions for 1 h at room temperature. PBS was used
   to wash the sections three times, and DAB solution was applied to the
   sections for 30–60 s accordingly. Hematoxylin was used to stain the
   nuclei for 30 s. Finally, the sections were dehydrated and mounted
   before the light microscopic examination. The incorporated antibodies
   were: WBP2(Proteintech, #12030-1-AP, 1:200), NCOA4 (Bethyl
   Laboratories, #A302-272A, 1: 200), GPX4 (Abcam, #ab125066, 1: 200),
   FTH1 (Abcam, #ab183781, 1: 200), Anti-Mouse IgG-HRP (Abcam, #ab6789, 1:
   500), and Anti-Rabbit IgG-HRP (Abcam, #ab97051, 1: 500).

   Immunofluorescence staining. BUMPT cells were seeded on the coverslips,
   and cisplatin was used to treat the cells. After the treatment, cells
   were fixed with 4% paraformaldehyde before the permeation. 4-μm
   paraffin sections of renal samples were deparaffinized, rehydrated, and
   permeated before the antigen retrieval. All sections were subjection to
   5% BSA solution incubation (1 h, room temperature) for blocking. The
   sections were then incubated with primary antibody dilution overnight
   at 4^oC. After three times washing, secondary antibody dilutions were
   used to incubate the sections for 1–2 h at room temperature. The
   sections were rinsed, and the nuclei were stained with DAPI solution.
   Finally, the sections were subjected to fluorescence microscopic
   evaluation. The incorporated antibodies were: WBP2(Proteintech,
   #12030-1-AP, 1:200), Megalin (Abcam, # ab184676; 1:500), GPX4(Santa
   Cruz, #sc-166570, 1: 200), Lamp1 (Abmart, #TD7033S, 1: 100), CK18
   (AMSBIO, #A01357-1, 1:100), 4-HNE (R&D system, MAB3249, 1:100), GFP
   (Proteintech, #66002-1-Ig, 1: 200), Anti-Rabbit IgG-Alexa Fluor 488
   (Abcam, #ab150077, 1: 500), Anti-Mouse IgG-Alexa Fluor 488 (Abcam,
   #150113, 1: 500), Anti-Rabbit IgG-Alexa Fluor 594 (Abcam, #ab150080, 1:
   500), and Anti-Mouse IgG-Alexa Fluor 594 (Abcam, #150116, 1: 500).

   Evaluation of protein-protein interaction. The potential interaction
   between WBP2 and GPX4 was initially explored by String analysis, and
   its website is: [81]STRING: functional protein association networks
   (string-db.org). After that, co-immunoprecipitation (CO-IP) was used to
   validate their interaction. Cells were lysed by CO-IP lysis buffer with
   proteinase inhibitors. The protein concentration was initially detected
   by BCA assay, and about 750 μg protein (about 250 μl in volume) was
   incubated with 1–2 μg primary antibody overnight at 4^oC. 20 μl protein
   A/G beading solution (Santa-cruz, #sc-2003) was added into the mixture,
   and they were shanked for 4 h at room temperature to precipitate the
   targeted proteins. The mixtures were centrifuged, and the supernatants
   were removed. The beads were washed six times before boiling with
   2✕loading buffer. After cooling down, the supernatants were loaded into
   the SDS-PAGE gel, and a routing western blot procedure was applied to
   develop the bands. The incorporated antibodies were: Flag
   (Sigma-Aldrich, #F1804), HA (Proteintech, #51064-2-AP), Myc
   (Proteintech, #16286-1-AP), GPX4(Abcam, #ab125066), Rabbit IgG
   (Proteintech, #B900610), and Mouse IgG (Proteintech, #B900620).

   Evaluation of GPX4 activity. Homogenization buffer (pH 7.4) was
   initially prepared, which contains 0.1 M KH[2]PO[4]/K[2]HPO[4], 0.15 M
   KCl, 0.05% [wt/vol] CHAPS, 5 mM β-mercaptoethanol, and protease
   inhibitors. About 2✕10^6 cells or 200 μg renal cortex were homogenized
   in 200 μl homogenization buffer. The homogenates were centrifuged
   (10000 rpm, 15 min) at 4^oC, and the supernatants were harvested.
   Bradford assay was used to measure the protein concentration of the
   supernatants, and their concentrations were normalized with
   homogenization buffer. Subsequently, GPX4 assay buffer (pH 7.8) was
   made, which contains 5 mM EDTA, 5 mM GSH (MCE,#HY-D0187), 0.1%
   [vol/vol] Triton X-100, 180 IU/mL glutathione reductase (Beyotime,
   #P2372S), and 160 mM NADPH/H^+ (Beyotime, #ST360). 50 μL supernatants
   were mixed with 1 ml GPX4 assay buffer, and the mixtures were
   maintained at 22 °C for 5 min. After that, 5ul of 30 mM cumene
   hydroperoxide (Aladdin, #C109598) was added into the mixture to ignite
   the reaction. OD340nm readings were obtained every 10 s until the
   reading was stabilized. The rate of changes in OD340 readings was
   calculated as GPX4 activity.

   TUNEL staining. The state of DNA damage was evaluated by TUNEL staining
   kit (Roche, #11684795910). Briefly, cells were washed with PBS three
   times after the treatment, and they were then fixed with 4%
   paraformaldehyde for 1 h at room temperature. Permeation was applied to
   the cells at 4 °C with 2 min’ incubation of 0.1% Triton X-100
   (dissolved in 0.1% sodium citrate). TUNEL regent was used to stain the
   cells for 60 min at 37 °C. After two times of PBS washing, the sections
   were mounted and evaluated by fluorescence microscopy.

   Preparation of lysosomal and lysosome-free fractionations: The
   isolation of lysosomes was performed as previously indicated in the kit
   (Solarbio, #EX1230). Cells were harvested and washed with PBS solution
   twice. 1 ml Solution A was used to resuspend the cells, and the cells
   were shaken for 10 min at 4 °C. Dounce homogenizer was used to
   facilitate the process. The mixture was sequentially centrifuged at
   1000×g for 5 min and 1000×g for 10 min, and the supernatant was
   collected. Subsequently, the supernatant was centrifuged at 20000×g for
   20 min, and the pellet was harvested. 500 μL solution B was added and
   resuspended, and the mixture was centrifuged at 20000×g for 20 min, and
   the pellet was collected, which was dissolved by 100 μL solution C.
   After 30 min’ shaking, the mixture was centrifuged at 12000×g for
   15 min, and the supernatant with abundant lysosomal proteins was used
   for further studies.

   Evaluation of labile iron levels. FerroOrange (DOJINDO, #F374) staining
   was used to detect the labile iron in BUMPT cells. Briefly, DMSO was
   used to dissolve FerroOrange to make the stock solution (1 mM). Working
   solution of FerroOrange (2 μM) was made by adding stocking solution
   into cell culture media (1:500). Cells were stained with working
   solution for 30 min at 37^oC. Fluorescence microscopy was used to
   examine the cells immediately after the staining.

   Plasmid construction. The cDNA templates of WBP2 and GPX4 were acquired
   from Miaolingbio (Wuhan, China). They were truncated with routing
   protocols. The full-length and truncated forms of WBP2 and GPX4 were
   subcloned in the backbone plasmid, including PCDH-3✕Myc, PCDH-3✕Flag,
   and PCDH-3✕HA. For point mutation, mutated sites were designed in the
   primers, and the cDNA sequence was divided into two parts by the PCR
   procedure. Subsequently, fusion PCR was performed to obtain the mutated
   full-length cDNA sequence. For the expression of GPX4, a
   selenoprotein-expressing lentivirus plasmid, i.e.
   seleno-GFP-3✕flag-PCDH, was generated. The details of the synthesis
   were indicated in the supplemental files.

   Statistics. For bioinformatics analysis, statistical procedures were
   performed by the R statistical software tool version 3.6.1
   ([82]www.r-project.org). For experimental data, Graphpad Prism 9.0 was
   used to calculate the differences between each group. Mean ± SD was
   used to indicate the values of results. 1-way ANOVA with Dunn's
   multiple comparisons was used for statistics.

   Study approval. Animal procedures used in this study were approved by
   the Animal Care and Use Committee of the Second Xiangya Hospital of
   Central South University, China (20220515). Human studies were approved
   by the Second Xiangya Hospital of Central South University
   (LYF2022146). Written informed consent from participants or their
   guardians was obtained.

3. Results

3.1. WBP2, a novel modulator of AKI and ferroptosis, participated in the
occurrence of CP-AKI

   In order to identify novel modulators of AKI & ferroptosis, a series of
   bioinformatics analyses was performed. After quality control and
   removal of the batch effect between batches ([83]Supplemental Figs.
   1A–1E, see “Methods”), a total of 83 AKI patients and 68 controls from
   3 datasets ([84]GSE1563, [85]GSE30718, and [86]GSE61739) were enrolled
   in this study ([87]Supplemental Table 2). Subsequently, WGCNA analysis
   was used to investigate correlation patterns among genes across
   microarray renal samples, which identified 8 modules ([88]Fig. 1A &
   [89]Supplemental Table 3). Notably, the eight modules corresponded to
   6840 genes correlated with AKI ([90]Fig. 1B). Among these, Black and
   Brown modules showed significance in both comparisons, both showing
   upregulation of the eigenvectors in the AKI samples ([91]Fig. 1C and
   [92]Supplemental Figs. 1F–1G). The core-related genes (hub genes) from
   the Brown and Black modules were analyzed, and their interactions were
   demonstrated by a network map ([93]Fig. 1D). In order to explore the
   physiological processes captured by the network, we evaluated the
   cellular and molecular progress significantly enriched in all the
   network nodes ([94]Supplemental Fig. 1H - 1O). Importantly, the
   majority of them belong to canonical pathways involved mainly in
   apoptosis and inflammation, indicative of their essential role in the
   progression of AKI.

Fig. 1.

   [95]Fig. 1
   [96]Open in a new tab

   WBP2, a novel modulator of AKI and ferroptosis, participated in the
   occurrence of CP-AKI. A The identification of WGCNA co-expression
   modules; B The determination of the clinical features of AKI; C The
   correlation analysis of the genes in all modules and the clinical
   features of AKI; D The network map of Brown and black modules; E The
   merging of WGCNA network and ferroptosis-related gene network, and the
   identification of WBP2 as the hub gene; F The downregulation of WBP2 in
   kidneys of AKI patients; G-H Western blot studies revealed that the
   expression of WBP2 was decreased in cisplatin-treated BUMPT cells and
   kidneys. DMF: dimethylformamide, CP: cisplatin. (For interpretation of
   the references to colour in this figure legend, the reader is referred