A neurodegeneration gene, WDR45, links impaired ferritinophagy to iron accumulation

Abstract Neurodegeneration with brain iron accumulation (NBIA) is a clinically and genetically heterogeneous group of neurodegenerative diseases characterized by the abnormal accumulation of brain iron and the progressive degeneration of the nervous system. One of the recently identified subtypes of NBIA is β‐propeller protein‐associated neurodegeneration (BPAN). BPAN is caused by de novo mutations in the WDR45/WIPI4 (WD repeat domain 45) gene. WDR45 is one of the four mammalian homologs of yeast Atg18, a regulator of autophagy. WDR45 deficiency in BPAN patients and animal models may result in defects in autophagic flux. However, how WDR45 deficiency leads to brain iron overload remains unclear. To elucidate the role of WDR45, we generated a WDR45‐knockout (KO) SH‐SY5Y neuroblastoma cell line using CRISPR‐Cas9‐mediated genome editing. Using these cells, we demonstrated that the non‐TF (transferrin)‐bound iron pathway dominantly mediated the accumulation of iron. Moreover, the loss of WDR45 led to defects in ferritinophagy, a form of autophagy that degrades the iron storage protein ferritin. We showed that impaired ferritinophagy contributes to iron accumulation in WDR45‐KO cells. Iron accumulation was also detected in the mitochondria, which was accompanied by impaired mitochondrial respiration, elevated reactive oxygen species, and increased cell death. Thus, our study links WDR45 to specific iron acquisition pathways and ferritinophagy. Cover Image for this issue: https://doi.org/10.1111/jnc.15388


| INTRODUC TI ON
Neurodegeneration with brain iron accumulation (NBIA) is a clinically and genetically heterogeneous group of neurodegenerative diseases characterized by the abnormal accumulation of brain iron and progressive extrapyramidal dysfunction (Levi & Finazzi, 2014). MRI analyses of the brains of patients with NBIA have indicated that the most prominent iron accumulation is in the basal ganglia, particularly in the globus pallidus and the substantia nigra (Haack et al., 2012;Hayflick et al., 2013). Thus far, 15 genes have been identified to be accountable for NBIA (Levi et al., 2019). Only two of the causative genes have been directly associated with mutations in the known iron metabolism proteins, including CP (ceruloplasmin) and FTL (ferritin light chain) (Hayflick et al., 2018). The roles of the remaining genes that participate in iron metabolism have yet to be determined for a variety of cellular processes, such as lipid metabolism, lysosomal activity, and autophagic processes (Arber et al., 2016;Lu et al., 2011;Wang et al., 2019).
Currently, no effective treatment is available for NBIA.
Despite the location of the gene on the X chromosome, BPAN does not follow the classic X-linked dominant inheritance. Females have a higher incidence of BPAN than males (e.g., 20 females vs. 3 males), likely because of the lethality of hemizygous mutation in males . Female patients are all heterozygous. The phenotypes of the affected male patients carrying BPAN mutations are clinically indistinguishable from females . Patients of both genders display a similar phenotype because of somatic or germline mosaicism . BPAN patients display global developmental delays in infancy and early childhood, followed by neurological deterioration in early adulthood with progressive dystonia, parkinsonism, cognitive decline, and seizures (Haack et al., 2012;Hayflick et al., 2013;Saitsu et al., 2013;Schneider et al., 2013). Recent studies on exome sequencing have identified de novo mutations in the WDR45 (WD repeat domain 45) gene in BPAN patients (Haack et al., 2012;Saitsu et al., 2013).
These heterozygous mutations involve both missense and truncation mutations that lead to the loss of WDR45 function (Haack et al., 2012;Saitsu et al., 2013). The upregulation of toxic iron in the fibroblasts of two BPAN-affected patients (Ingrassia et al., 2017) suggests that alterations in iron homeostasis are associated with WDR45 deficiency.
Recent data on fibroblasts and pluripotent stem cell-derived neurons from a female patient with BPAN have shown elevated iron levels and mitochondrial and lysosomal dysfunction (Seibler, 2018). However, these studies did not directly assess how WDR45 deficiency leads to disruptions in brain iron homeostasis.
WDR45 regulates autophagy, an evolutionarily conserved mechanism of the degradation and recycling of dysfunctional cellular components (Klionsky & Emr, 2000). Structured as a beta-propeller scaffold protein, WDR45 is one of the four mammalian homologs of yeast Atg18, which plays important roles in regulating autophagosome formation (Bakula et al., 2017). WDR45 has attracted considerable attention recently because of its interaction with ATG2A, a lipid transfer protein that is important for expansion of the phagophore, the precursor to the double-membrane autophagosome that transports cargo during autophagy (Chowdhury et al., 2018;Maeda et al., 2019). Lymphoblast cells derived from BPAN patients exhibit lower autophagic activity and the accumulation of early autophagic structures (Saitsu et al., 2013), which supports the loss-of-function mechanisms underlying BPAN. Rodent models further reveal the critical roles of WDR45 in neuronal development and survival. Neuronspecific wdr45-knockout (KO) mice display poor motor coordination, learning and memory defects, and extensive axon swelling (Zhao et al., 2015). The primary neuronal cells in these mice show defects in autophagic flux, with the accumulation of ubiquitin-positive aggregates in both neurons and axons (Zhao et al., 2015). In addition, constitutive wdr45-KO mice display cognitive impairments, abnormal synaptic transmission, and lesions in several brain regions (Wan et al., 2019). The defect in autophagy caused by WDR45 deficiency leads to elevated ER stress and neuronal apoptosis (Seibler, 2018), which is consistent with neurodegeneration found in BPAN patients.
However, how the roles of WDR45 in autophagy are related to iron accumulation in BPAN patients remains unknown.
Iron is an essential trace element required by almost all living organisms (Ganz, 2013). As the most metabolically active organ in the body, the brain requires iron for oxidative metabolism, myelination, mitochondrial energy generation, and the biosynthesis of neurotransmitters (Madsen & Gitlin, 2007;Salvador, 2010;Todorich et al., 2008).
However, excess iron becomes toxic because of the generation of highly reactive free radicals through Fenton chemistry, resulting in oxidative stress and directly damaging DNA, lipids, and proteins (Uttara et al., 2009). Therefore, the brain requires precise iron homeostasis to guard against iron accumulation while providing the optimal level of this nutrient, which is essential for its development and functions.
Ferritinophagy is one of the mechanisms that could link the roles of WDR45 in autophagy to iron homeostasis. Ferritinophagy is the process through which the intracellular iron storage protein ferritin is sequestered within autophagosomes and delivered to lysosomes for degradation (Mancias et al., 2014). This process is crucial in liberating iron from ferritin and thus in maintaining cellular iron homeostasis (Mancias et al., 2014). Because of the role of WDR45 in autophagy, we hypothesize that WDR45 deficiency promotes defects in ferritinophagy, which may underlie iron accumulation. The molecular mechanisms that link alterations in the autophagy function to changes in iron levels by WDR45 deficiency remain unknown. Thus, the goal of the present study is to define the contribution of WDR45 to iron homeostasis, including iron uptake and ferritinophagy, and to determine the functional consequences of iron accumulation.

| Generation of WDR45 KO SH-SY5Y cell lines
The WDR45 CRISPR-edited SH-SY5Y cell lines were generated by electroporation of CRISPR-Cas9 ctRNPs using the Neon Transfection System (ThermoFisher) (Jacobi et al., 2017). Briefly, tracrRNA and crRNA (target sequence: 5'-ATGACTCAACAGCCACTTCG-3') were annealed at equimolar amounts by heating at 95°C and slow cooling at room temperature to form ctRNA complexes. Recombinant spCas9 protein (IDT) and annealed ctRNA were separately diluted in Buffer R (supplied with the Neon Transfection System) to a concentration of 40 μM each. Equal amounts of spCAS9 protein and ctRNA were added together and incubated at room temperature for 20 min to form ctRNPs. SH-SY5Y cells were trypsinized, washed 3 times with 1X PBS, and 2 × 10 5 cells were resuspended in 18 μL of Buffer R. Two microliters of ctRNP were added to the 18-μL aliquot of cells and incubated at room temperature for 5 min. Using a 10μl Neon tip, each 10μl aliquot of the cell suspension was electroporated using 2 pulses at 1200V and 20-ms pulse width. Cells were dispensed in separate wells of a 96-well plate with pre-warmed media. Cells were allowed to recover for 48 h. Cells were then trypsinized and washed 2 times with 1× PBS. Cells were resuspended in 500 μL of 1× PBS and single-cell sorted on a MoFlo Astrios Cell Sorter (Beckman Coulter, University of Michigan Flow Core). Genomic DNAs from clonal cell lines were isolated using the Puregene Cell and Tissue Culture Kit (Qiagen). PCR products containing the predicted CRISPR cut site were amplified (WDR45-F1: 5'-GCAGTCACTCCAGAGTCAGTAGTTAG; WDR45-R1: 5'-GTAGATGCGCACACCTGTCTCCATG) and sequenced (University of Michigan DNA Sequencing Core).

| Cell culture and reagents
All culture media and supplements were purchased from Invitrogen.
Heat-inactivated fetal bovine serum was purchased from Sigma-Aldrich. SH-SY5Y human neuroblastoma cells (RRID:CVCL_0019) were grown in DMEM containing 10% fetal bovine serum, penicillin (100 IU/ml), and streptomycin (100 mg/ml) at 37°C in a humidified, 5% CO 2 incubator. This cell line is not listed in the International Cell Line Authentication Committee (ICLAC) database of cross-contaminated or misidentified cell lines. No further authentication was performed in the laboratory. All the cells we used in this study were within 20 passages. The MYC-FLAG-tagged human WDR45 plasmid was purchased from Origene (RC209654). The dual-tagged mCherry-GFP-LC3 plasmid was procured from Addgene (123230; deposited by Robin Ketteler). For small interfering RNA (siRNA)-mediated gene suppression of ferritin, SH-SY5Y cells were plated in six-well plates overnight and then transfected with 50 nM of control siRNA (SIC001-10NMOL; MilliporeSigma) or FTH1 siRNA (SASI_Hs01_00112824) using Lipofectamine 3000 (Thermo Fisher Scientific) according to the manufacturer's specifications. Pyrimidinone 8 was purchased from Sigma-Aldrich (ENA432539789).

| TF-bound and Non-TF-bound 59 Fe uptake assays
For TF-bound 59 Fe uptake assays, the loading of 59 Fe onto TF was performed as previously described (Burdo et al., 2003). Briefly, 59 FeCl 3 (PerkinElmer Life Sciences) was incubated with TF-loading buffer (0.1 M HEPES, pH 7.5, 0.15 M NaCl) containing 20 mM NaHCO 3 and 88 μM nitrilotriacetic acid for 5 min at room temperature. Then, 20 μM apo-TF was added and incubated for 2 h at RT. Non-TF-bound 59 Fe was removed from the 59 Fe-TF solution by buffer exchange using a Nanosept 10K molecular weight cutoff Omega spin column (PALL Corporation) and spinning 3 × 10 min at 2841 g. This procedure yielded a 59 Fe-TF purity of >95%. Cells were washed three times with prewarmed PBS and incubated for 90 min at 37°C with serum-free medium containing 59 Fe-TF. Cells were chilled on ice and washed twice with ice-cold PBS. Radioactivity was determined with a gamma counter and normalized to the cell protein measured in lysates using the Bradford assay. A non-TF-bound 59 Fe uptake assay was performed as described previously (Choi et al., 2019). Cells were washed twice with serum-free medium (SFM) and incubated for 1 h in SFM containing 2% bovine serum albumin to bind residual TF and prevent iron uptake via TF-bound iron endocytosis. After incubation, cells were washed with serum-free medium and incubated with 1 μM 59 Fe with 150 μM ascorbate in pH 6.0 or 7.4 uptake buffer at 37°C. Cells were chilled on ice for 5 min, washed three times with ice-cold PBS containing iron chelator solution (1 mM bathophenanthroline sulfonate and 1 mM diethylenetriaminepentaacetic acid) to remove any residual 59 Fe, and then were directly harvested for intracellular radioactivity. Radioactivity was determined with a gamma counter and normalized to the cell protein measured in lysates using the Bradford assay.

| Trace element analysis
Cells were grown on 100-mm tissue culture dishes and analyzed for metals by inductively coupled plasma mass spectrometry (ICP-MS) as we described previously (Choi et al., 2018(Choi et al., , 2019(Choi et al., , 2020. Briefly, the cell samples were digested with 2 ml/g total wet weight nitric acid (BDH ARISTAR ® ULTRA) for 24 h and then digested with 1 ml/g total wet weight hydrogen peroxide (BDH Aristar ® ULTRA) for 24 h at room temperature. The samples were stored at 4°C until metals were quantified. Ultrapure water was used for final sample dilution.
For mitochondrial iron levels, mitochondria were isolated through differential centrifugation as we previously described (Choi et al., 2018), and then mitochondrial iron levels were measured via ICP-MS.

| RNA isolation and qPCR
Total RNA was isolated from cells using TRIzol reagent (Invitrogen) following the manufacturer's instructions. Purified RNA was then reverse-transcribed with SuperScript ® III First-Strand Synthesis System (Invitrogen). The qRT-PCR was performed using Power SYBR-Green PCR Master Mix (Applied Biosystems). 18S RNA was used for normalization of the mRNA. The primers used for qPCR are listed in Table S1 and were all purchased from Integrated DNA Technologies.

| MtDNA copy number
Total DNA was extracted from cell samples via TRIzol (Invitrogen) extraction as described previously (Choi et al., 2018). Following complete removal of the RNA-containing aqueous phase, DNA extraction buffer [Tris base (1 M), sodium citrate dibasic trihydrate (50 mM), and guanidine thiocyanate (4 M)] were added to the tubes containing the remaining Trizol-separated interphase and infranatant. The tubes were shaken vigorously and centrifuged at 12 000 g at room temperature for 30 min. The aqueous phase was collected, and the genomic and mitochondrial DNA was precipitated in isopropanol. Samples were respun at 12 000 g at 4°C to pellet the DNA.
The DNA pellet was then washed in 70% ethanol, respun, and, after careful ethanol removal, resuspended in TE buffer. To quantify the mtDNA copy number, qPCR was performed as described above against external standards for mtDNA and β-globin using primers listed in Table S1.

| Immunoblot analysis
The total lysates were prepared in RIPA buffer plus protease inhibitors (Roche, 11836153001). Protein concentrations were determined by Bradford assay. Samples were separated by electrophoresis and transferred to a nitrocellulose membrane (Bio-Rad, 1620115). Protein expression was quantified using densitometry (Image Studio Lite; LI-COR).

| Immunofluorescence and microscopy
For confocal studies, cells were plated onto glass coverslips and fixed with 4% paraformaldehyde in PBS for 10 min, and immunofluorescence staining was performed as previously described (Choi et al., 2018(Choi et al., , 2019(Choi et al., , 2020. To permeabilize the cells, cells were incubated with 0.2% Triton X-100 in PBS for 5 min. Nonspecific binding was blocked with 4% bovine serum albumin in PBS for 30 min, and cells were detected after incubation with anti-mouse ferritin (Santa Cruz Biotechnology, sc-74513) for 1 h. Detection of ferritin was performed by using an anti-mouse IgG antibody conjugated to Alexa 488 (Thermo Fisher Scientific, A11029) for 20 min. Rabbit anti-LC3 (Cell Signaling Technology, 2775) or rabbit anti-LAMP1 (Proteintech, 21997-1-AP) was used as a marker of autophagosomes or lysosomes, respectively. Detection of LC3 or LAMP1 was performed by using an anti-rabbit IgG antibody conjugated to Alexa Fluor 568 (Thermo Fisher Scientific, A11036) for 20 min. Coverslips were drained, mounted in ProLong Gold (Thermo Fisher Scientific), and sealed with nail polish. Immunofluorescence imaging was performed by using a Nikon Eclipse A-1 confocal microscope (Nikon Instruments) with an X60 oil immersion lens. For the dual fluorescent LC3 assay, cells were transfected with dual tagged mCherry-GFP-LC3 and were then fixed 48h post transfection. Green vs. green/red (yellow) vesicles were quantified by analyzing 20 different images by two individuals, one blind to the experimental design and data analysis.
Exposure settings were unchanged throughout acquisition.

| Calcein-AM assays for labile iron pool
Cells were grown in a black walled 96-well plate and incubated overnight. Medium was removed, and cells were washed three times with PBS and then incubated with 1 µM calcein-AM (Life Technologies) in PBS for 20 min at 37°C in 5% CO 2 . Cell-associated fluorescence (excitation, 488 nm; emission, 517 nm) was measured at 25°C using a plate reader. Reduced fluorescence intensity reflects quenching as a result of free iron binding to calcein.

| Measurement of ROS formation
ROS levels were measured using a fluorescent indicator specific for H 2 O 2 production (H 2 DCFDA) as described previously (Choi et al., 2018(Choi et al., , 2019. H 2 DCFDA, a cell-permeant ROS indicator, is nonfluorescent and can be converted to highly fluorescent 2',7'-dichlorofluorescein (DCF) following the removal of the acetate groups by intracellular esterases and ROS-induced oxidation. Cells were incubated with 3 μM DCFH-DA (Invitrogen) for 30 min at 37°C.
Fluorescence of DCF (excitation 495 nm, emission 520 nm) was measured at 25°C using a plate reader.

| Seahorse XFe96 extracellular flux analysis
Extracellular flux analyses were performed using a Seahorse XFe-96 analyzer (Agilent Technologies), according to the manufacturer's instructions. Twenty-four hour before the assay, cells were cultured on Seahorse XF-96 plates at a density of 3 × 10 4 cells per well. Cells were washed and incubated with XF assay Medium (Seahorse Bioscience), supplemented with 25 mM glucose, 1 mM sodium pyruvate, 2 mM l-glutamine at 37°C and 0% CO 2 for 1 h. The baseline oxygen consumption rates (OCR) were measured at 37°C four times before sequentially injecting the following: Oligomycin (2 μM) to measure the ATP-linked OCR, the oxidative phosphorylation uncoupler FCCP (1.5 μM) to determine maximal respiration, and rotenone (1 µM) and antimycin A (1 µM) to determine the non-mitochondrial respiration.
The baseline extracellular acidification rates (ECAR) were measured at 37°C four times before sequentially injecting the following: dglucose (10 mM) to measure the glycolytic rate, oligomycin (2 μM) to inhibits the mitochondrial ATP synthase, and 2-deoxyd-glucose (2-DG; 50 mM) to inhibit glycolysis. OCR and ECAR were automatically calculated by the Seahorse XFe-96 software.

| Isoprostane analysis
Isoprostane was chosen as a marker of oxidative damage and measured by a competitive enzyme-linked immunosorbent assay (ELISA) for one of the isoprostanes 8-iso Prostaglandin F2α (8-iso-PGF) with a commercial kit (Cayman Chemical). The assay was based on the competition between 8-iso-PGF and 8-isoprostane-acetylcholinest erase (AChE) conjugate for a limited number of binding sites in each ELISA plate well. The concentration of 8-iso-PGF is inversely proportional to the number of binding sites available, whereas AChE is held constant. The assays were performed as we described previously (Choi et al., 2018(Choi et al., , 2020).

| CASP3 activity
The CASP3 activity in cells was measured by the Caspase-3 Assay Kit (Abcam, ab39401). Briefly, the cells were harvested and lysed on ice. After centrifugation, protein was measured and then adjusted to the concentration as suggested, then the CASP3 activity was examined by following the manufacturer's instructions.

| Proliferation assay
Cells were plated on 96-well plates. After 2 days of cell growth, 10 µL/well of WST-8 reagent (cell counting kit 8; Sigma-Aldrich, 96992) was added and absorbance at 450 nm was measured by using a BioTek Synergy microplate reader (BioTek Instruments). Medium without cells was used as background, and the A 450 of background was subtracted from the samples.

| Study design and statistical analysis
This study was not pre-registered. Institutional ethics approval was not required for our study. No randomization was performed to allocate treatments in the study. No exclusion criteria were predetermined. Statistical analysis was performed using GraphPad Prism 8 (GraphPhad Software).
Data are presented as individual values and represent the means ± SEM. To compare two groups, two-tailed p values were calculated using an unpaired t-test. To compare more than two groups, p values were calculated using 1-way ANOVA with Tukey's multiple comparisons test or two-way ANOVA with Bonferroni's multiple comparisons test as specified in each figure legend. Normality of the data was not assessed. Sample sizes were not statistically pre-determined but were estimated based on previous studies (Choi et al., 2018(Choi et al., , 2019 and similar to those generally used in the field. Outliers were identified using the GraphPad ROUT (robust regression and outlier removal) method (Q = 1%). Values of p < 0.05 were considered statistically significant. Asterisks in graphs, wherever present, denote statistically significant differences.

| Generation of a WDR45-KO SH-SY5Y cell line
To assess the role of WDR45 in neuron-related cells, we inactivated the WDR45 gene in SH-SY5Y neuroblastoma cells (SH-SY5Y ΔWDR45 ).
The SH-SY5Y cell line maintains many neuronal properties and has been used to investigate other neurodegenerative diseases (Xicoy et al., 2017). We used CRISPR-Cas9 to target editing near the start codon in exon 3 of WDR45 ( Figure S1a). A CRISPR-Cas9-mediated cut introduced a homozygous 64-bp deletion, which spanned exon 3 and intron 3 ( Figure S1b,c). The resulting RNA is predicted to encode a 58 amino acid polypeptide with a translation start in exon 3 and a premature stop in intron 3 (data not shown). Quantitative PCR (qPCR) analysis, using primer pairs that amplified a downstream segment of WDR45 mRNA (exons 8-9), showed an ~84% reduction compared with the wild type (WT) ( Figure S1d), indicating that the mutant WDR45 mRNA underwent nonsense-mediated decay.
These results confirmed the inactivation of the WDR45 gene in SH-SY5Y ΔWDR45 cells.

| Loss of WDR45 enhances intracellular iron levels
Brain iron accumulation is a key hallmark among the heterogeneous group of NBIA, including BPAN (Levi et al., 2019). Thus, we used inductively coupled plasma mass spectrometry (ICP-MS) to measure the total content of metals in SH-SY5Y ΔWDR45 cells. The total iron content was significantly increased ~2.0-fold (t-test, t = 9.778, Figure 1a). We also measured the levels of other trace elements, including zinc, copper, and manganese.
Unexpectedly, a significant increase in the total zinc and copper Thus, these data demonstrate a direct link between the expression of WDR45 and intracellular iron levels.

| Loss of WDR45 alters cellular iron acquisition pathways
To identify the mechanism by which the loss of WDR45 increases total iron levels, we examined the cellular iron uptake pathways.
Cellular iron acquisition involves two distinct pathways: (1) the canonical TF (transferrin)-bound iron (TBI) uptake pathway and (2) the nontransferrin-bound iron uptake (NTBI) pathway. We first examined the kinetics of the uptake of 59 Fe by the TBI pathway in SH-SY5Y WT and SH-SY5Y ΔWDR45 cells. The cells were incubated with 59 Fe-TF at 37°C, and then cell-associated 59 Fe was monitored over time. We found that 59 Fe-TBI uptake kinetics did not differ between the two groups ( Figure 2a). We then examined the kinetics of 59 Fe-NTBI uptake by SH-SY5Y ΔWDR45 cells. The major NTBI transporters, SLC11A2/DMT1 (solute carrier family 11 member 2) and SLC39A8/ ZIP8-SLC39A14/ZIP14, showed distinct pH activities. SLC11A2/ DMT1 was the most efficient at pH 5.5-6.0, whereas SLC39A8/ZIP8 and SLC39A14/ZIP14 showed maximal iron transport near or higher than the physiological pH of 7.0 (Knutson, 2017;Wang et al., 2012).
Thus, we examined 59 Fe-NTBI uptake under different pH conditions.
These results suggest that the NTBI uptake driven by SLC11A2/ DMT1 was aberrantly enhanced in SH-SY5Y ΔWDR45 cells.

| Altered expression of iron-related proteins in SH-SY5Y ΔWDR45 cells
To further corroborate the contribution of a specific iron transporter, we examined the mRNA expression of known iron transporters involved with iron uptake and efflux within the brain. The peptide that inhibits iron transport through the internalization and degradation of SLC40A1/FPN (Nemeth et al., 2004). The HAMP transcript level was significantly higher in SH-SY5Y ΔWDR45 cells than in SH-SY5Y WT cells (Figure 2d). We further measured the mRNA expression of iron-related proteins, including CP (ceruloplasmin) and APP (amyloid beta precursor protein). CP is the iron ferroxidase that facilitates SLC40A1/ferroportin export activity and is widely expressed throughout the brain (Hadziahmetovic et al., 2008). Two molecular isoforms of CP exist: secreted CP (sCP) and a membrane glycosylphosphatidylinositol (GPI)-anchored form of CP (GPI-CP). In the brain, sCP is detectable in the cerebrospinal fluid (Paradowski et al., 1995), whereas GPI-CP is expressed in mammalian astrocytes (Patel & David, 1997). No significant difference was detected in the level of the transcript encoding sCP between SH-SY5Y WT and SH-SY5Y ΔWDR45 cells (Figure 2d). In our hands, the transcript encoding GPI-CP was not detected in both groups. APP is another iron-related protein known to have a functional role in iron homeostasis through stabilizing the iron efflux protein SLC40A1/FPN (Duce et al., 2010;McCarthy et al., 2014;Rogers et al., 2016;Wong et al., 2014). We also did not detect any significant difference in the APP transcript level be-

| Autophagic flux in SH-SY5Y ΔWDR45 cells
Previous studies on WDR45 mutant patient-derived lymphoblast cells and wdr45-KO mice have shown defective autophagic flux (i.e., the rate of autophagy) (Saitsu et al., 2013;Wan et al., 2019;Zhao et al., 2015). Thus, we sought to determine whether our SH-SY5Y ΔWDR45 cells also exhibited altered autophagic flux. During autophagy, a cytosolic proteolytically processed form of MAP1LC3/ LC3 (microtubule-associated protein 1 light chain 3), referred to as LC3-I, is conjugated to phosphatidylethanolamine to form LC3-II, which in turn is recruited to phagophore membranes (Tanida et al., 2008). When autophagosomes fuse with lysosomes to form autolysosomes, the population of LC3-II that was present inside the autophagosomes is consequently degraded by lysosomal hydrolases (Mizushima & Yoshimori, 2007;Tanida et al., 2005). Lysosomal dysfunction induced by lysosomal inhibitors leads to the marked accumulation of LC3-II in autophagic structures (Tanida et al., 2005).
Thus, we evaluated the autophagic flux by measuring LC3-II levels in the absence and presence of the lysosomal inhibitor chloroquine (CQ) (Mauthe et al., 2018;Redmann et al., 2017). Immunoblot analysis revealed no difference in LC3-II levels between SH-SY5Y WT and SH-SY5Y ΔWDR45 cells under basal conditions (i.e., without Torin 1 treatment), although the total amount of LC3 (LC3-I + LC3-II) was lower in the SH-SY5Y ΔWDR45 cells ( Figure S2a,b). The induction of autophagy with the MTORC1 inhibitor Torin 1 led to an elevation of LC3-II levels in SH-SY5Y WT cells and SH-SY5Y ΔWDR45 cells ( Figure S2a,b). The concomitant treatment with CQ to block autophagosome-lysosome fusion led to a further elevation of the LC3-II levels in both SH-SY5Y WT and SH-SY5Y ΔWDR45 cells, indicating that the Torin 1-dependent increase in LC3-II was not the result of a block in fusion ( Figure S2a,b). We further assessed the autophagic flux by measuring the endogenous autophagic substrate SQSTM1/ p62 (sequestosome 1) levels. SQSTM1/p62, the first autophagic cargo receptor identified in mammalian cells, mediates the formation and autophagic clearance of intracellular protein aggregates (Bjorkoy et al., 2005;Komatsu et al., 2007;Pankiv et al., 2007). Consistent with LC3-II levels in SH-SY5Y WT cells, the SQSTM1/p62 levels were reduced by Torin 1 but elevated by the concomitant treatment with CQ ( Figure S2a,c). However, in SH-SY5Y ΔWDR45 cells, the SQSTM1/ p62 levels were slightly reduced by Torin 1, but not further changed by the concomitant treatment with CQ; the latter result suggests a partial impairment of degradation in the SH-SY5Y ΔWDR45 cells. Taken together, these data suggest no significant difference in autophagic flux between SH-SY5Y WT and SH-SY5Y ΔWDR45 cells. Instead, the basal levels of autophagy machinery are reduced in SH-SY5Y ΔWDR45 cells, suggesting the impaired biogenesis of autophagosomes.
The increased ferritin levels in SH-SY5Y ΔWDR45 cells prompted us to assess the ferritinophagy process, by which ferritin is sequestered within autophagosomes and delivered to lysosomes for degradation (Mancias et al., 2014;Masaldan et al., 2018). This process is important in liberating iron from ferritin and thus in maintaining cellular iron homeostasis (Mancias et al., 2014). We hypothesized that impaired ferritin degradation explains the iron

| Loss of WDR45 leads to lysosomal dysfunction
To further identify which step of ferritinophagy is defective in SH-SY5Y ΔWDR45 cells, we monitored the autophagic flux using a chimeric LC3 fused to both GFP (acid sensitive) and mCherry (acid stable) (Kimura et al., 2007;Pankiv et al., 2007). In autophagosomes, the physiological pH allows both GFP (green) and mCherry (red) fluorescence. However, upon fusion with a lysosome, forming an autolysosome, the low pH quenches the GFP signal, permitting only mCherry to fluoresce (Figure 4a). The GFP-mCherry-LC3 was expressed in SH-SY5Y WT and SH-SY5Y ΔWDR45 cells, and the cells were then analyzed using confocal microscopy.
As expected, SH-SY5Y WT cells exhibited most of the vesicles as red, indicating normal autolysosome formation (Figure 4b). By contrast, SH-SY5Y ΔWDR45 cells showed predominantly yellow vesicles (Figure 4b), indicating either impaired fusion between autophagosomes and lysosomes, or lysosomal dysfunction including elevated pH. To distinguish between these possibilities, we assessed whether ferritin is accumulated in either autophagosomes or lysosomes in SH-SY5Y ΔWDR45 cells. We found that SH-SY5Y ΔWDR45 cells exhibited a substantial overlap of ferritin with the LAMP1 marker corresponding to endosomes and lysosomes but not with LC3, the autophagosome marker ( Figure 4c). These data suggest that loss of WDR45 specifically interferes with lysosomal function to degrade ferritin.

| Loss of WDR45 increases mitochondrial iron levels and alters the expression of genes involved in oxidative phosphorylation
When iron is inside cells, it is either stored in ferritin or transferred to organelles such as mitochondria. Mitochondria are responsible for heme synthesis and iron-sulfur cluster (Fe-S) biogenesis, which are mainly present in the electron transport chain (Napier et al., 2005;Ponka, 1997). Thus, we hypothesized that WDR45 defi-

| WDR45 deficiency impairs mitochondrial respiration
The accumulation of mitochondrial iron accompanied by reduced levels of oxidative phosphorylation machinery prompted us to test whether WDR45 deficiency influences the major function of mitochondria. Using Seahorse XF extracellular flux assays, we measured mitochondrial respiratory chain function and oxygen consumption rate (OCR). The results showed that mitochondrial respiratory chain function was significantly decreased in basal respiration, ATP-linked respiration, and maximal respiration in SH-SY5Y ΔWDR45 cells compared with SH-SY5Y WT cells (Figure 5c, These results suggest that WDR45 deficiency promotes a general decline in cellular metabolism rather than a shift to glycolysis to compensate for the impaired mitochondrial function.

F I G U R E 4
Lysosomal dysfunction in SH-SY5Y ΔWDR45 cells. (a) Schematic illustration of dual-fluorescent LC3 assay. The tandem tagged mCherry-GFP-LC3 is a pH-sensitive sensor that is used to monitor the autophagic flux. The GFP tag (green) is acid-sensitive, whereas the mCherry tag (red) is acid-insensitive. In autophagosomes, both GFP and mCherry tags emit fluorescent light resulting in yellow fluorescence. By contrast, the fusion of autophagosomes to lysosomes results in acidic autolysosomes where GFP is quenched, and mCherry continues to emit red fluorescence. (b) Representative confocal images of SH-SY5Y WT or SH-SY5Y ΔWDR45 cells expressing mCherry-GFP-LC3 are shown. SH-SY5Y WT cells show mostly red fluorescence in lysosomes because of quenching of GFP at low pH, whereas SH-SY5Y ΔWDR45 cells display a yellow fluorescence in vesicular structures representing autophagosomes. The blue color indicates nuclei counterstained with DAPI. Scale bar: 10 µm. (c) Representative confocal images of ferritin and LC3 or LAMP1 in SH-SY5Y ΔWDR45 cells. To detect ferritin, the cells were incubated with an anti-mouse ferritin antibody followed by Alexa Fluor 488-conjugated secondary antibody (green), indicated by arrows.
To detect LC3 or LAMP1, the cells were incubated with anti-LC3 antibody or anti-LAMP1 antibody followed by Alexa Fluor 568-conjugated secondary antibody (red). Blue colors indicate nuclei counterstained with DAPI. Scale bar: 10 µm

| Loss of WDR45 enhances oxidative damage and induces cell death
The excess of iron promotes the generation of highly reactive free radicals through Fenton chemistry, resulting in oxidative stress and directly damaging DNA or oxidizing lipids and proteins (Uttara et al., 2009). To test whether WDR45 deficiency led to increased oxidative stress, we measured the levels of 2′,7′-dichlorofluorescein (DCF) as a proxy of oxidative stress (LeBel et al., 1992). The DCF fluorescence was slightly but significantly increased ~1.2 fold (t-test, t = 4.185, df = 14, p = 0.0009) in SH-SY5Y ΔWDR45 cells compared with SH-SY5Y WT cells (Figure 6a). We next measured the oxidative damage  et al., 2015). The levels of 8-isoprostane were significantly higher by ~3.7 fold (t-test, t = 5.352, df = 4, p = 0.0059) in SH-SY5Y ΔWDR45 cells than in SH-SY5Y WT cells (Figure 6b).

| Knockdown of ferritin and treatment with a SLC11A2/DMT1 inhibitor reduce iron overload in SH-SY5Y ΔWDR45 cells
Having established the role of WDR45 in ferritinophagy (Figures 3   and 4), we sought to determine the roles of ferritin in iron accumulation observed in SH-SY5Y ΔWDR45 cells. To this end, we knocked down FTH1 (ferritin heavy chain 1) in SH-SY5Y ΔWDR45 cells using siRNA and measured iron levels. RT-qPCR analysis showed that indicating reversed cell death (Figure 7e,f, One-way ANOVA, F(2, 6) = 43.65, p = 0.0003). In addition, to confirm a role for SLC11A2/ DMT1 in iron accumulation observed in SH-SY5Y ΔWDR45 cells, pyrimidinone 8 was used as a pharmacological inhibitor of this transporter (Montalbetti et al., 2015). Incubation with pyrimidinone 8 for 10 min prior to the start of and during the 59 Fe-NTBI uptake assay reduced 59 Fe-NTBI uptake by SH-SY5Y ΔWDR45 cells (Figure 7g, Oneway ANOVA, F(2, 6) = 17.35, p = 0.0032). We also confirmed that pyrimidinone 8 treatment reduced the iron concentration to a normal level in SH-SY5Y ΔWDR45 cells (Figure 7h, One-way ANOVA, F(2, 6) = 19.13, p = 0.0025). These combined data indicate that elevated ferritin level and SLC11A2/DMT1-mediated NTBI uptake pathway contribute to iron accumulation in SH-SY5Y ΔWDR45 cells.

| DISCUSS ION
The goal of this study was to identify the major proteins and pathways involved in iron accumulation as a result of WDR45 deficiency and how the altered iron uptake and metabolism contribute to neurodegeneration. We provided several lines of evidence for the major pathways leading to brain iron overload in WDR45 deficiency, and our results shed light on the mechanisms by which WDR45 deficiency affects iron-induced neurodegeneration ( Figure 8). First, our 59 Fe flux study revealed that WDR45 deficiency induced significant changes in cellular iron uptake by acquiring NTBI in the more acidic environment rather than TBI. Second, we found that WDR45 deficiency led to the upregulation of the NTBI uptake transporter SLC11A2/DMT1−IRE and the downregulation of iron exporter SLC40A1/FPN levels, which may underlie neuronal iron accumulation. Third, we showed that cellular iron overload in WDR45 deficiency was associated with impaired ferritinophagy. Fourth, WDR45 deficiency leads to mitochondrial iron accumulation, altered mitochondrial metabolism, and enhanced ROS production, thus potentially contributing to neuronal cell death and neurodegeneration.
The effects of WDR45 deficiency on iron transport have not been directly examined. We found that the SH-SY5Y WT cells in our study took up iron presented in both the TBI and NTBI forms.
Although TBI is a constituent of brain interstitial fluid, our 59 Fe-NTBI uptake results revealed that the levels of NTBI uptake were much higher in SH-SY5Y ΔWDR45 cells compared with SH-SY5Y WT cells, suggesting that NTBI acts as a major substrate for iron uptake by SH-SY5Y ΔWDR45 cells. Correspondingly, the entry of NTBI into the cells was associated with increased levels of ferritin expression. When iron is inside cells, it is either deposited into ferritin or incorporated into the mitochondria; otherwise, iron should be subsequently exported by SLC40A1/FPN, the only known cellular iron exporter in mammals (Abboud & Haile, 2000;McKie et al., 2000). We found that SH-SY5Y ΔWDR45 cells displayed significantly reduced SLC40A1/ FPN expression along with increased HAMP/hepcidin expression, suggesting that the reduced iron export function in response to elevated expression of HAMP is also responsible for cellular iron sequestration as a result of WDR45 deficiency. In contrast to the increased levels of ferritin and HAMP that we observed, a study on patient-derived WDR45 mutant fibroblast cells reported reduced levels of L-ferritin and HAMP following elevated intracellular iron levels (Seibler, 2018). These inconsistent observations may be explained by the different cell types used in the studies (fibroblast vs. neuroblastoma cells) or the nature of WDR45 mutations (hypomorphic mutation in the patient cells vs. the null mutations in our study).
One striking result of our study was that in addition to iron, the Further studies are necessary to define the role of WDR45 in the transport of zinc or copper and whether and how they contribute to neurological issues in human patients.
However, lymphoblast cells from patients with static encephalopathy of childhood with neurodegeneration in adulthood (SENDA), now renamed BPAN, showed a reduction, rather than a block, in autophagic flux (Saitsu et al., 2013). Furthermore, the study by Zhao et al. (2015) did not analyze flux using standard methodologies (Klionsky et al., 2008); the authors compared WT and mutant cells with the lysosomal inhibitor bafilomycin A 1 , but they did not examine either cell line in the same experiment in the absence and presence of inhibitor, which is necessary to determine flux. Finally, the study by Wan et al. (2019) Figure S2a). Moreover, LC3-II levels were similar in SH-SY5Y WT and SH-SY5Y ΔWDR45 cells even after Torin 1 treatment and the concomitant treatment with CQ. Similarly, we did not detect any substantial differences in SQSTM1/p62 levels between the two groups. These data suggest that WDR45 may not be essential for autophagosome formation. Similar to our results, a recent study using the mouse N2a/Neuro2a neuroblastoma cell line failed to detect a defect in autophagy in wdr45 KO cells (in contrast to the previous study from some of the same authors (Zhao et al., 2015)), but did observe dysfunctional autophagy in wdr45 wdr45b double KO cells (Ji et al., 2021). These data suggest a partial overlap in the function of the two homologous gene products, similar to the result seen with the homologous yeast proteins Atg18 and Atg21 (Nair et al., 2010). Notably, we observed that overall levels of LC3-I and LC-II were reduced in SH-SY5Y ΔWDR45 cells compared with SH-SY5Y WT cells, raising the possibility that WDR45 may play some role in the biogenesis of autophagosomes. Importantly, we think that SH-SY5Y ΔWDR45 cells are partially defective in lysosomal function, which may account for the phenotypes observed in all of these studies.
Mitochondria are one of the major cellular storage organelles for iron (Arber et al., 2016). Accordingly, mitochondrial iron levels should be tightly regulated to provide sufficient iron for numerous cellular processes that use iron as a co-factor, including the electron transport chain, while guarding against the generation of radicals induced by an excess of redox-active iron (Urrutia et al., 2014). We found Three slides per group were used for quantification. One-way ANOVA followed by post hoc Tukey's multiple comparisons test (**p < 0.01). (g) Cells were treated with 10 µM inhibitor for 10 min prior to the start of and during the 59 Fe-NTBI uptake assay for 1 h in pH 6.0 uptake buffer at 37°C. Whole cell 59 Fe content was determined using a γ-counter. Each sample was normalized for protein concentration. n = 3 independent cell culture preparations. One-way ANOVA followed by post hoc Tukey's multiple comparisons test (**p < 0.01). (h) Cells were treated with 10 µM inhibitor for 16 h, and total iron levels were measured by ICP-MS. n = 3 independent cell culture preparations. One-way ANOVA followed by post hoc Tukey's multiple comparisons test (**p < 0.01) that WDR45 deficiency leads to mitochondrial iron accumulation, impaired mitochondrial respiration, elevated ROS, and increased neuronal cell death. Consistent with our findings, mitochondrial iron accumulation has been implicated in many neurodegenerative diseases (Mena et al., 2015). In addition, we found markedly reduced levels of nDNA-encoded SDH subunits without affecting the levels of mtDNA-encoded genes in SH-SY5Y ΔWDR45 cells (Figure 5b).
Inhibition of SDH leads to neuronal death in caudate and putamen nuclei and symptoms of Huntington disease (Túnez et al., 2010).
Another competitive SDH inhibitor malonate can cause neuronal injuries and degeneration (Beal et al., 1993)  Our study is the first to link WDR45 to ferritinophagy. The SH-SY5Y ΔWDR45 cells showed impaired release of 59 Fe from ferritin and the sequestration of iron-loaded ferritin in lysosomes ( Figure 4).
What are the mechanisms by which WDR45 regulates ferritinophagy? NCOA4 is a receptor protein that links ferritin-containing cargo to phagophores (Mancias et al., 2014). However, we found that WDR45 deficiency did not alter NCOA4 levels. The worm ortholog of WDR45, EPG-6, is essential for an early step of autophagosome formation (Lu et al., 2011). However, our confocal microscopy data demonstrated that WDR45 deficiency did not affect the formation of autophagosomes but instead inhibited amphisome-lysosome fusion or lysosomal function. Thus, these data suggest that WDR45 plays an essential role in a late step of ferritinophagy involving lysosomal function. Another key question concerns how impaired ferritinophagy is linked to the increased iron uptake by SLC11A2/DMT1 in the WDR45-deficient cells. We speculate that the accumulation of ferritin-bound iron induces cytoplasmic iron deficiency (Figure 3h), which would upregulate SLC11A2/DMT1 expression, thereby leading to further iron incorporation into SH-SY5Y ΔWDR45 cells. In support of this idea, ferritin knockdown and SLC11A2/DMT1 inhibitor alleviated the iron accumulation in WDR45-deficient cells (Figure 7). However, the present study does not rule out the possibility of alternative mechanisms such as upregulation of ferritin expression in response to oxidative stress. The expression of ferritin is primarily regulated through post-transcriptional, iron-dependent machinery based on the interaction by the IRPs and IRE located on the target mRNAs (Leibold & Guo, 1992). Oxidative stress has been shown to increase the synthesis of both subunits of ferritin as well as the activity of IRP in HeLa cells (Orino et al., 2001). Conversely, overexpression of either ferritin subunit reduces ROS formation in response to oxidative stress (Orino et al., 2001). Moreover, ER stress activation was reported to induce the expression of ferritin in HepG2 cells (Oliveira et al., 2009). These studies led us to postulate the alternative hypothesis that excess iron-induced oxidative stress may be involved with the upregulation of ferritin in WDR45-deficient F I G U R E 8 Schematic model for the impact of WDR45 deficiency on iron accumulation, ferritinophagy, and cell death. WDR45-deficient cells accumulate iron via the nontransferrin-bound iron pathway (NTBI). The increased total iron levels were accompanied by an increase in the iron storage protein ferritin. WDR45 deficiency impairs ferritinophagy, the autophagic degradation of ferritin. Elevated iron levels were also detected in the mitochondria, which may impair mitochondrial function, elevate reactive oxygen species, and increase cell death cells. Iron is needed for cell survival via many cellular processes such as oxidative metabolism. Meanwhile, iron is also needed for programed cell death mediated by ferroptosis. In our study, impaired ferritinophagy in WDR45 mutant cells was accompanied by increased cell death. These data suggest that reduced labile iron as a result of the impaired ferritinophagy had a larger impact on the pro-survival role of iron compared with ferroptosis. All experiments were conducted in compliance with the ARRIVE guidelines.

CO N FLI C T O F I NTE R E S T
The authors declare no competing interest.

O PE N R E S E A RCH BA D G E S
This article has received a badge for Open Materials because it provided all relevant information to reproduce the study in the manuscript. More information about the Open Science badges can be found at https://urldefense.com/v3/__https://cos.io/ourservices/open-science-badges/__;!!N11eV2iwtfs!5ezBrT68HKDaTc KFB_tgQSbQctmf3Js9zk_hsOVYVIezTUXnccegybbzifUxMQ$

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.