ADCK4 Deficiency Destabilizes the Coenzyme Q Complex, Which Is Rescued by 2,4-Dihydroxybenzoic Acid Treatment

Significance Statement

ADCK4 mutations generally manifest as steroid-resistant nephrotic syndrome, and cause coenzyme Q₁₀ (CoQ₁₀) deficiency. However, ADCK4’s function remains obscure. Using mouse and cell models, the authors demonstrated that podocyte-specific Adck4 deletion in mice significantly reduced survival and caused severe FSGS, effects that were prevented by treatment with 2,4-dihydroxybenzoic acid (2,4-diHB), a CoQ₁₀ precursor analogue. ADCK4-knockout podocytes exhibited a significantly reduced CoQ₁₀ level and defects in mitochondrial function that were rescued by 2,4-diHB treatment, thus these phenotypes were attributed to decreased CoQ₁₀ levels. The authors also found that ADCK4 interacted with mitochondrial proteins, including COQ5, and that ADCK4 knockout decreased COQ complex levels. These findings reveal that ADCK4 is required for CoQ₁₀ biosynthesis and mitochondrial function in podocytes, and suggests a treatment strategy for nephrotic syndrome caused by ADCK4 mutations.

Visual Abstract

Abstract

Background

Mutations in ADCK4 (aarF domain containing kinase 4) generally manifest as steroid-resistant nephrotic syndrome and induce coenzyme Q₁₀ (CoQ₁₀) deficiency. However, the molecular mechanisms underlying steroid-resistant nephrotic syndrome resulting from ADCK4 mutations are not well understood, largely because the function of ADCK4 remains unknown.

Methods

To elucidate the ADCK4’s function in podocytes, we generated a podocyte-specific, Adck4-knockout mouse model and a human podocyte cell line featuring knockout of ADCK4. These knockout mice and podocytes were then treated with 2,4-dihydroxybenzoic acid (2,4-diHB), a CoQ₁₀ precursor analogue, or with a vehicle only. We also performed proteomic mass spectrometry analysis to further elucidate ADCK4’s function.

Results

Absence of Adck4 in mouse podocytes caused FSGS and albuminuria, recapitulating features of nephrotic syndrome caused by ADCK4 mutations. In vitro studies revealed that ADCK4-knockout podocytes had significantly reduced CoQ₁₀ concentration, respiratory chain activity, and mitochondrial potential, and subsequently displayed an increase in the number of dysmorphic mitochondria. However, treatment of 3-month-old knockout mice or ADCK4-knockout cells with 2,4-diHB prevented the development of renal dysfunction and reversed mitochondrial dysfunction in podocytes. Moreover, ADCK4 interacted with mitochondrial proteins such as COQ5, as well as cytoplasmic proteins such as myosin and heat shock proteins. Thus, ADCK4 knockout decreased the COQ complex level, but overexpression of ADCK4 in ADCK4-knockout podocytes transfected with wild-type ADCK4 rescued the COQ5 level.

Conclusions

Our study shows that ADCK4 is required for CoQ₁₀ biosynthesis and mitochondrial function in podocytes, and suggests that ADCK4 in podocytes stabilizes proteins in complex Q in podocytes. Our study also suggests a potential treatment strategy for nephrotic syndrome resulting from ADCK4 mutations.

Coenzyme Q (CoQ, ubiquinone)—a lipophilic component located in the inner mitochondrial membrane, Golgi apparatus, and cell membranes—plays a pivotal role in oxidative phosphorylation.¹ CoQ shuttles electrons from complexes I and II to complex III in the mitochondrial respiratory chain.² It also has a critical function in antioxidant defense because of its redox potential.³ The CoQ biosynthesis pathway has been extensively studied in Saccharomyces cerevisiae.⁴ A minimum of 12 proteins, encoded by the Coq genes, form a complex that serves to stabilize each other and is involved in coenzyme synthesis.⁵ Based on protein homology, approximately 15 homologous COQ genes have been identified in humans.¹

Primary CoQ deficiencies due to mutations in ubiquinone biosynthetic genes (COQ2, COQ4, COQ6, COQ7, COQ9, PDSS1, PDSS2, aarF domain containing kinase 3 [ADCK3], and ADCK4) have been identified.^6–12 Clinical manifestations of CoQ₁₀ deficiency vary depending on the genes involved, and mutations in the same gene can result in diverse phenotypes depending on the mutated allele.¹³ COQ2,¹¹ COQ6,¹⁰ PDSS2,⁹ and ADCK4¹² have also been implicated in steroid-resistant nephrotic syndrome (SRNS). Although no effective therapy has been described for SRNS, supplementation of CoQ₁₀ in cases of SRNS resulting from CoQ deficiency acts to alleviate the associated clinical symptoms.^10,12,14 This is partially true for ADCK4-related glomerulopathy and several cases have been reported accordingly.^12,14 Recently, 2,4-dihydroxybenzoic acid (2,4-diHB) has been shown to ameliorate disparate phenotypes in yeast, mouse, and Caenorhabditis elegans models harboring a mutation in the respective COQ7, Mclk1, or clk-1 genes.^15–17 In these studies, 2,4-diHB was shown to act as an alternate or “bypass” ring precursor, able to restore endogenous CoQ biosynthesis. In addition, 2,4-diHB prevents renal disease of podocyte-specific Coq6^−/− mice.¹⁸

ADCK3 (also known as COQ8A) and ADCK4 are two mammalian orthologs of yeast Coq8p/Abc1, which belong to the microbial UbiB family; they appear to result from gene duplication in vertebrates.^6,19 UbiB and Coq8p are required for CoQ biosynthesis in prokaryotes and yeast, respectively, and are speculated to activate an unknown monooxygenase in the CoQ biosynthesis pathway.¹⁹ Coq8p, ADCK3, and ADCK4 are present on the matrix side of the inner mitochondrial membrane.^20–22 Coq8p is essential for the organization of high molecular mass Coq polypeptide complexes and for phosphorylated forms of the Coq3, Coq5, and Coq7 polypeptides that are involved in methylation and hydroxylation steps in CoQ biosynthesis.^21,23 Similarly, it has been shown that ADCK3 interacts with CoQ biosynthesis enzymes in a protein complex (complex Q).²⁴ Moreover, ADCK3 lacks protein kinase activity in the trans form; exhibits ATPase activity; and has highly conserved, unorthodox protein kinase–like domains, including the KxGQ motif, present in UbiB and eukaryotic COQ8 homologs, including ADCK4.²⁵

However, it is not clear whether ADCK4 functions in a manner similar to that of ADCK3. Mutations in the ADCK4 (also known as COQ8B) gene generally manifest as adolescence-onset SRNS, sometimes accompanied with medullary nephrocalcinosis or extrarenal symptoms, including seizures.^12,26 The molecular mechanisms underlying SRNS resulting from ADCK4 mutations are not well understood, largely because the function of ADCK4 is unclear. Therefore, in this study, we investigated the function of ADCK4 using mouse and cell models.

Methods

Mouse Breeding and Maintenance

The animal experimental protocols were reviewed and approved by the Institutional Animal Care and Use Committee at the University of Michigan (#08619), Boston Children’s Hospital (#13-01-2283), and Yonsei University College of Medicine (#2015-0179). All mice were handled in accordance with the Guidelines for the Care and Use of Laboratory Animals. Mice were housed under pathogen-free conditions with a light period from 7:00 am to 7:00 pm and had ad libitum access to water and irradiated rodent chow (catalog #0006972; LabDiet, St. Louis, MO). Targeted Adck4^{tm1a(EUCOMM)Hmgu} (Adck4^tm1a) embryonic stem cells were obtained from EUCOMM and injected into the blastocysts of mice. Chimeric mice were bred with C57BL/6J mice to establish germline transmission. Nphs2.Cre⁺ (stock #008205) and Pgk1.Flpo⁺ (#011065) mice were obtained from Jackson Laboratory. Mouse lines were bred onto the C57BL/6J genetic background. Genotyping was performed by PCR using the following primers: #1, GGA​TAG​GGG​GCT​GGA​GAG​ATG; #2, GCC​CGC​CTC​CCT​GTA​TCT​TAG; #3, TCG​GAG​AGG​AAA​GGA​CTG​GAG; #4, CCC​TTT​CCC​TTG​AGT​TCA​CAG​C; and #5, TGG​CCT​CAA​ACT​CAT​GAA​AAT​ACT​CC. Mice were maintained in mixed sex and were randomly assigned to the different experimental groups. For mouse studies, experimental results were validated over multiple litters, across several generations of the mouse colony with n>6. Data collection of urine, whole blood, and plasma analysis data were blinded to genotype so that the operator did not know the genotype during performing the measurements. Histologic and ultrastructural analysis was not blinded to genotype but all samples were processed using the same protocol.

Supplementation of 2,4-diHB to the Mice in Drinking Water

2,4-diHB at a concentration of 25 mM was administered to the mice via drinking water and changed twice a week. The treatment was started at 3 months of age and continued up to 18 months of age.

Urine Analysis

Urine was collected by housing the mice overnight (12 hours) in metabolic cages. All samples were immediately frozen and stored at −80°C. The samples were thawed on ice before urine albumin and creatinine measurements. Urinary albumin was measured using the Albumin Blue Fluorescent Assay Kit (Active Motif), as per the manufacturer’s instructions. Urine creatinine was measured using the liquid chromatography with tandem mass spectrometry (LC-MS/MS) method as described previously.²⁷ Proteinuria was expressed as milligram of albumin per milligram of creatinine.

Whole Blood and Plasma Analysis

Blood was collected from mice via the facial vein bleeding method and collected in citrate tubes. The blood sample was subsequently analyzed using the Vetscan VS2 Chemistry Analyzer, as per the manufacturer’s instructions. Plasma samples obtained by centrifugation of whole blood were immediately frozen at −80°C. Plasma creatinine was measured using the LC-MS/MS method as described previously.²⁷

Immunoblotting and Immunofluorescence Staining

These experiments were performed as described previously.²⁸ Anti-podocin (P0372) and anti-FLAG M2 (F3165; Sigma-Aldrich); anti-nidogen (NBP1-97701; Novus); anti-nephrin (GP-N2; Progen); anti-synaptopodin (PA5-56997; ThermoFisher); anti-COQ3 (28051-1-AP), anti-COQ5 (17453-1-AP), and anti-COQ9 (14874-1-AP; Proteintech, Rosemont, IL); anti–α smooth muscle actin (anti-αSMA; #19245), anti–p-mammalian target of rapamycin (anti–p-mTOR; #5536P), anti-mTOR (#2983P), anti–p-p38 (#4511S), anti–p-extracellular signal–regulated kinase (anti–p-ERK; #4377S), anti-pJNK (#9251S), anti-p38 (#9212S), and anti-LC3 (#4108S; Cell Signaling Technology, Danvers, MA); anti-collagen IV (ab6586), anti-COXIV (ab33985), and anti-actin (ab49900; Abcam, Cambridge, UK); and ADCK4 (LS-C119206; LSBio, Seattle, WA) were purchased from the indicated commercial sources. Alexa Fluor 488 Phalloidin and secondary antibodies were purchased from Invitrogen. Fluorescent images were obtained using an SP5× or SP8-U-FLIM laser scanning microscopes (Leica) using excitation wavelengths of 405, 488, and 594 nm and 100×/63× oil immersion objectives (HCX PL APO CS 100×/1.44 OIL and HC PL APO CS2 63×/1.40 OIL) or an LSM 700 microscope (Carl Zeiss) using excitation wavelengths of 405, 488, 555, and 639 nm and a 40× water immersion objective (C-Apochromat, 1.2 numerical aperture, Zeiss). Images were processed and analyzed using Leica AF, Zeiss LSM software, ImageJ, and Adobe Photoshop CS6 software.

Histologic Analysis

The kidney tissues were fixed in 4% paraformaldehyde, sectioned (5 µm thickness), and stained with hematoxylin and eosin, Periodic acid–Schiff, Masson trichrome, SFOG, and Jones silver stain following the standard protocols for histologic examination.

Ultrastructural Analysis

The kidney tissues and cells were fixed in 2.5% glutaraldehyde, 1.25% paraformaldehyde, and 0.03% picric acid in 0.1 M sodium cacodylate buffer (pH 7.4) overnight at 4°C. They were then washed with 0.1 M phosphate buffer, postfixed with 1% osmium tetroxide dissolved in 0.1 M PBS for 2 hours, dehydrated in ascending gradual series (50%‒100%) of ethanol, and propylene oxide was used for infiltration. Samples were embedded using the Poly/Bed 812 Kit (Polysciences) according to manufacturer’s instructions. After pure fresh resin embedding and polymerization in a 65°C oven (TD-700; DOSAKA, Kyoto, Japan) for 24 hours, sections of approximately 200–250 nm thickness were cut and stained with toluidine blue for light microscopy. Sections of 70 nm thickness were double stained with 6% uranyl acetate (22400; EMS) for 20 minutes and lead citrate (Fisher) for 10 minutes for contrast staining. The sections were cut using a Leica EM UC-7 with a diamond knife (Diatome) and transferred onto copper and nickel grids. All the sections were observed by transmission electron microscopy (TEM; JEM-1011, JEOL, and Zeiss 912) at an acceleration voltage of 80 kV.

Plasmids, Cell Culture, Transfection, and Lentivirus Transduction

Single guide RNAs (sgRNAs) targeting human ADCK4 (sgRNA1, GCT​GCA​CAA​TCC​GCT​CGG​CAT; sgRNA2, GTA​AGG​TCT​GCA​CAA​TCC​GCT; and sgRNA3, GAC​CTT​ATG​TAC​AGT​TCG​AG) were cloned into BsmBI-digested lentiCRISPR v2 (plasmid #52961; Addgene). ADCK4 cDNA was cloned into the p3xFLAG CMV26 (C-terminal) vector (Sigma-Aldrich). Bacterial alkaline phosphatase (BAP) cDNA cloned into the p3xFLAG CMV7 vector was digested using Kpn1 and EcoR1 restriction enzymes (New England BioLabs) and ligated into the p3xFLAG CMV24 vector.

Immortalized human podocytes²⁹ were maintained in RPMI+GlutaMAX-I (Gibco) supplemented with 10% FBS, penicillin (50 IU/ml)/streptomycin (50 μg/ml), and insulin-transferrin-selenium-X. Human proximal tubule (HK-2) and HEK293 cells were maintained in DMEM supplemented with 10% FBS and 1% penicillin/streptomycin. Plasmids were transfected into podocytes or HEK293 cells using Lipofectamine 2000 (Invitrogen). HEK293 cells stably expressing p3xFLAG-ADCK4 or BAP were selected and maintained with 1 mg/ml G418.

To establish ADCK4 knockout (KO) cells, lentiCRISPR v2, pMD2.G, and psPAX2 were transfected into Lenti-X 298T cells (Clontech). Supernatants containing lentivirus were collected 48 hours after transfection and passed through a 0.2 µM filter. Cultured podocytes and HK-2 cells were transduced with lentivirus, selected, and maintained with 4 μg/ml puromycin.

Cell Viability Assay

Cell viability assays were performed using the Cell Counting Kit-8 (Dong-In Biotech). Cell suspensions (100 µl; 1×10⁵/ml) with culture medium were added to a 96-well plate and incubated for 24 hours in a carbon dioxide incubator. The medium was replaced with phenol-free fresh medium with or without 30 µM AA and/or 500 µM 2,4-diHB for 15 hours. Four replicate wells were included for each condition. CCK-8 reagent (10 µl) was added to each well and incubated for 1 hour; OD of the sample at 450 nm was measured.

Isolation of Glomeruli

Mice were anesthetized with isoflurane. The kidneys were perfused with Dynabeads (M450-epoxy, 4.5 µm; Invitrogen). Dynabeads were washed with ice-cold PBS containing 0.1% BSA before use, according to the manufacturer’s instructions. Then, 20 ml of Dynabeads in HBSS at a concentration of 8×10⁷ beads/ml was injected into the left cardiac ventricle. Following perfusion, the kidneys were removed, minced, and digested with 1 mg/ml collagenase A and 0.1 mg/ml deoxyribonuclease type I at 37°C for 30 minutes with gentle agitation. The digested tissue was then gently passed in series through 180-, 90-, and 75-µm sieves, followed by intermittent rinsing with ice-cold sterile HBSS. To eliminate adherence of the glomeruli, each sieve was rinsed with HBSS containing 1% BSA before use. The mixture passed through the sieves was then centrifuged at 200 × g for 5 minutes. The supernatant was discarded and the pellet was dissolved in 2 ml of ice-cold HBSS and then transferred into a 1.7 ml tube. Glomeruli that contained Dynabeads were isolated using a magnetic particle concentrator and washed at least three times with ice-cold sterile HBSS. The entire procedure was performed on ice, except for the collagenase digestion, which was performed at 37°C.

Cellular Lipid Extraction and CoQ Quantification via HPLC-MS/MS

Cells (approximately 0.1 g) and isolated glomeruli were thawed on ice and resuspended in 1.5 ml of PBS (0.14 M sodium chloride, 12.0 mM monosodium phosphate, and 8.1 mM disodium phosphate; pH 7.4), followed by homogenization using a polytron (PT 2500E; Kinematica) for 1 minute at 10,000 rpm on ice. Lipid extracts were prepared as previously described with minor modifications.³⁰ Briefly, dipropoxy-CoQ₁₀ was used as the internal standard and was added at a constant volume to all of the cell pellets as well as to a set of five CoQ₉ and CoQ₁₀ standards of known concentrations ranging linearly from 7.2 pmol to 400 pmol (to obtain a typical standard curve for CoQ quantification in the cell pellets). The samples were vortexed in 2 ml of methanol for 30 seconds, followed by addition of 2 ml petroleum ether. After vortexing for an additional 30 seconds, the organic upper layer was transferred to a new tube. Another 2 ml of petroleum ether was added to the original methanol layer and samples were vortexed again for 30 seconds. The organic phase was removed and the combined organic phase was dried under a stream of nitrogen gas. The samples were resuspended in 200 µl of ethanol containing 1 mg/ml benzoquinone to oxidize all of the lipids. HPLC-MS/MS analysis of the samples was performed using a 4000 QTRAP linear MS/MS spectrometer from Applied Biosystems (Foster City, CA). Applied Biosystem software, Analyst version 1.4.2, was used for data acquisition and processing. Chromatographic separation was achieved through a reverse phase Luna 5 µM PFP(2) column (Phenomenex) with a mobile phase comprising 90% solvent A (95:5 mixture of methanol/isopropanol containing 2.5 mM ammonium formate) and 10% solvent B (isopropanol containing 2.5 mM ammonium formate) at a constant flow rate of 1 ml/min. All samples were analyzed in multiple-reaction-monitoring mode. The precursor and product ion transitions monitored were as follows: mass/charge (m/z), 795.6/197.08 for CoQ₉; m/z, 812.6/197.08 for CoQ₉ with ammonium adduct; m/z, 863.6/197.08 for CoQ₁₀; m/z, 880.6/197.08 for CoQ₁₀ with ammonium adduct; m/z, 919.7/253.1 for dipropoxy-CoQ₁₀; and m/z, 936.7/253.1 for dipropoxy-CoQ₁₀ with ammonium adduct. The method used to quantify the signals corresponding to CoQ₉ and CoQ₁₀ was as described.³¹

Mitochondrial Respiratory Enzyme Activity Measurement

Cell lysates (15–50 µg) were diluted in phosphate buffer (50 mM monopotassium phosphate, pH 7.5) and subjected to spectrophotometric analysis for isolated respiratory chain complex activities at 37°C using a spectrophotometer (PerkinElmer). Complex II activity was measured at 600 nm (ε=19.1 mmol⁻¹cm⁻¹) after the addition of 20 mM succinate, 80 µM dichlorophenolindophenol, 300 µM potassium cyanide, and 50 µM decylubiquinone. Complex II activity was defined as the flux difference with or without 10 mM malonate. Complex II and III activity was also determined at 550 nm (ε=18.5 mmol⁻¹cm⁻¹) in the presence of 10 mM succinate, 50 µM cytochrome c, and 300 µM potassium cyanide. Complex II-III activity was defined as the flux difference before and after the addition of 10 mM thenoyltrifluoroacetone. All chemicals were obtained from Sigma-Aldrich.

Determination of Reactive Oxygen Species

Intracellular reactive oxygen species (ROS) production was measured using the CellROX Deep Red reagent (Invitrogen) following the manufacturer’s instructions. Control and ADCK4 KO podocytes were seeded at 1×10⁴ cells/well in a 96-well plate and incubated for 24 hours. The CellROX reagent was added to the medium at a final concentration of 5 µM, and the plate was then incubated for 30 minutes. After washing twice with HBSS containing calcium and magnesium, the cells were treated with hydrogen peroxide (H₂O₂; 100 or 300 µM), tert-butyl hydroperoxide (tBHP; 30 or 100 µM), or AA (30, 100, or 500 µM) in HBSS. The fluorescence at 640/665 nm was measured using a Cytation 5 cell imaging multimode reader (Biotek, Winooski, VT). Mitochondrial ROS generation was assessed using the MitoSOX Red reagent (Invitrogen), a red mitochondrial superoxide indicator. The MitoSOX reagent was added at final concentration of 5 µM, and the cells were incubated for 15 minutes and then washed twice with HBSS. Subsequently, the cells were treated with 500 µM H₂O₂, 100 µM tBHP, 30 µM AA, or 100 µM AA. The fluorescence at 510/580 nm was measured using Cytation 5.

Isobaric Tag Labeling for Relative and Absolute Quantification

Isobaric tag labeling for relative and absolute quantification (iTRAQ) was performed by Poochon Scientific as described previously.³² Proteins (100 µg) were extracted from control and ADCK4 KO podocytes, digested with trypsin, and labeled using the 8-plex iTRAQ Labeling Kit (AB Sciex). Fractionation of the iTRAQ–multiplex-labeled peptide mixture was carried out using an Agilent AdvanceBio column (2.7 µm, 2.1×250 mm) on an Agilent UHPLC 1290f system (Agilent, Santa Clara, CA). The HPLC-MS/MS analysis was performed using a Thermo Scientific Q-Exactive hybrid quadrupole-orbitrap mass spectrometer and a Dionex UltiMate 3000 RSLCnano system (Thermo Fisher Scientific, San Jose, CA). MS raw data files were searched against human protein sequence databases obtained from the National Center for Biotechnology Information website using Proteome Discoverer 1.4 software (Thermo Fisher Scientific) based on the SEQUEST and percolator algorithms. The false discovery rate was set to 1%. The resulting Proteome Discoverer report contains all assembled proteins, with peptide sequences, peptide spectrum match counts, and the iTRAQ tag–based quantification ratio. The iTRAQ tag–based quantification was used to determine the relative abundance of proteins identified in the iTRAQ data set. One-way ANOVA was used to identify proteins differentially expressed between control and ADCK4 KO podocytes, corrected using the Bonferroni multiple comparison test. Quantitative ratios were log2 normalized for final quantitative testing. The relative abundance of the proteins was varied 1.5-fold between control and ADCK4 KO podocytes (n=4) or between AA-treated control and AA-treated ADCK4 KO podocytes (n=2). The raw data files of the iTRAQ data set were submitted to the ProteomeXchange Consortium via the PRIDE partner repository⁶ with the data set identifier PXD016725.

Identification of ADCK4 Interactors

Proteins (75 mg) from HEK293 cells stably expressing p3XFLAG-ADCK4 or -BAP were incubated with 80 µl of FLAG M2 agarose beads (Sigma-Aldrich) for 48 hours at 4°C in an orbital shaker. The agarose beads were washed four times with lysis buffer to restrict nonspecific binding. Subsequently, 200 µl of elution buffer containing 150 ng/µl 3xFLAG peptide was added and the samples were incubated overnight. The eluates were analyzed by immunoblotting, Coomassie blue staining, and silver staining. The eluates were digested and subjected to nanoflow LC-MS/MS analysis. Peptides were separated on a C18 precolumn (75 μm×2 cm, nanoViper, Acclaim PepMap100; Thermo Fisher Scientific) and analytic C18 column (75 μm×50 cm, PepMap RSLC; Thermo Fisher Scientific). Peptides were analyzed using an LC-MS/MS system consisting of Easy nLC 1000 (Thermo Fisher Scientific) and an Orbitrap Fusion Lumos mass spectrometer (Thermo Fisher Scientific) equipped with a nanoelectrospray source. MS/MS spectra were analyzed against the Uniprot human database using the following software analysis protocol. The reversed sequences of all proteins were appended into the database for the calculation of false discovery rate. ProLucid was used to identify the peptides, with a precursor mass error of 5 ppm and a fragment ion mass error of 200 ppm.³³ The output data files were filtered and sorted to compose the protein list using DTASelect (The Scripps Research Institute, La Jolla, CA), with two and more peptide assignments for a protein identification and a false positive rate of <0.01.³⁴ The MS proteomics data were deposited to the ProteomeXchange Consortium via the PRIDE partner repository with the data set identifier PXD016147.

Gene Ontology Analysis

ADCK4 interactors were analyzed using the Database for Annotation, Visualization, and Integrated Discovery (DAVID) for functional annotation. The functional annotation tool in the online version of DAVID (version 6.8) was run (http://david.abcc.ncifcrf.gov/) using the default parameters and Gene Ontology (GO) categories representing molecular function, cellular component, and biologic process were separately analyzed for enrichment. A P value of <0.05 was considered significant.

Statistical Analyses

Statistical analyses were performed using Graph Pad Prism 7 software. The results are presented as mean±SE or SD for the indicated number of experiments. Statistical analysis of continuous data was performed with the two-tailed t test or multiple comparison analysis, as appropriate. Specific tests performed in the experiments are indicated in the figure legends. The results with P<0.05 were considered statistically significant.

Results

Podocyte-Specific Adck4 KO Mice Developed Progressive Proteinuria, Severe FSGS, and Increased Adult Mortality

To evaluate the role of Adck4 in kidney function, we generated a transgenic Adck4 (Adck4^tm1a) mouse line, using embryonic stem cells obtained from EUCOMM (Supplemental Figure 1A). Efficient targeting of the Adck4 gene was confirmed by genotyping (Supplemental Figure 1B). Whole body loss of Adck4 in Adck4^tm1a mice proved lethal, which is consistent with the report of the International Mouse Phenotyping Consortium (www.mousephenotype.org). To circumvent Adck4^tm1a embryonic lethality, we generated podocyte-specific Adck4 KO mice Adck4^tm1d or Nphs2.Cre⁺;Adck4^flox/flox (hereafter referred to as Adck4^ΔPodocyte) by crossing the Nphs2-Cre⁺ mouse with the Adck4^flox/flox mouse in which two loxP sites surround exons 5 and 6 in the Adck4 gene. Although young Adck4^ΔPodocyte mice appeared grossly normal, including their kidneys (Supplemental Figure 2), increased morbidity (hunched posture and seedy fur) (Supplemental Figure 1C), a significantly increased mortality (Figure 1A), and weight loss (Supplemental Figure 3) were observed in older (>9 months old) Adck4^ΔPodocyte mice compared with littermate controls. Necropsy of 10-month-old Adck4^ΔPodocyte mice revealed pale and significantly small kidneys compared with those in littermate controls (Supplemental Figure 1C), indicating that podocyte-specific deletion of Adck4 causes structural and functional kidney defects in Adck4^ΔPodocyte mice. To examine the renal function of Adck4^ΔPodocyte mice, we performed serial urine and plasma analyses for 18 consecutive months (Supplemental Figure 1D). Adck4^ΔPodocyte mice displayed the first significant decrease in plasma albumin level at 5 months of age (Supplemental Figure 1E) and the increase in albumin-creatinine ratio (18.81-fold, P=0.0005) remained significant throughout the study period compared with those in littermate controls (Figure 1B). The increase over time in albuminuria was the maximum, up to 31.2-fold, in Adck4^ΔPodocyte mice compared with that in littermate controls (Figure 1B). The onset of kidney function decline in Adck4^ΔPodocyte mice was associated with a significant increase in plasma creatinine and plasma BUN levels at 7 months of age, progressing to CKD, followed by renal failure, and consequently death (Figure 1A, Supplemental Figure 1, F–H). Histologic analysis of kidneys from Adck4^ΔPodocyte mice at 10 months of age demonstrated severe global and FSGS with extensive interstitial fibrosis and tubular atrophy (Figure 1C). To characterize the glomerular phenotype of Adck4^ΔPodocyte mice, we quantified the number of sclerotic glomeruli in Adck4^ΔPodocyte mice at 10 months of age and found that Adck4^ΔPodocyte mice had a significantly increased number of sclerotic glomeruli (mean 96.03%) compared with that in littermate controls (Figure 1D). To characterize the molecular abnormalities in the glomeruli of Adck4^ΔPodocyte mice, we analyzed the expression pattern of the slit diaphragm proteins podocin (Figure 2A) and nephrin (Supplemental Figure 4A), basement membrane marker nidogen (Figure 2A, Supplemental Figure 4B), and primary process marker synaptopodin (Supplemental Figure 4B) in the kidneys of 10-month-old mice. Staining of podocyte markers was significantly reduced in the glomeruli of Adck4^ΔPodocyte mice compared with that of the control glomeruli (Figure 2B, Supplemental Figure 4, D and F), demonstrating that Adck4 function is required for podocyte maintenance and homeostasis. In addition, we analyzed the expression of the fibrotic markers collagen IV and αSMA (Supplemental Figure 4, A and C) in the kidneys of Adck4^ΔPodocyte mice. Indeed, the kidneys of Adck4^ΔPodocyte mice presented significantly increased expression of collagen IV and αSMA (Supplemental Figure 4, E and G) in the glomeruli, characteristic of glomerular fibrosis. To study the structural changes in the glomeruli of Adck4^ΔPodocyte mice at the ultrastructural level, we performed TEM using the kidney of 10-month-old Adck4^ΔPodocyte mice. The results revealed the abnormal structure of glomeruli, severe foot process effacement, and disturbed podocyte morphology in Adck4^ΔPodocyte mice (Figure 2C). The number of filtration-slit units per micrometer of basement membrane was significantly reduced in Adck4^ΔPodocyte mice compared with that in wild-type mice (Figure 2D). In addition, the podocytes of Adck4^ΔPodocyte mice appeared to contain abnormal mitochondria characterized by hyperproliferation and increased size as mice aged (Supplemental Figure 5). Overall, the glomerular phenotype of podocyte-specific Adck4 KO mice recapitulates the pathology of FSGS in humans resulting from ADCK4 mutations.

Figure 1.

Nphs2.Cre⁺;Adck4^flox/flox mice developed FSGS. (A) Nphs2.Cre⁺;Adck4^flox/flox mutant mice exhibited reduced life span with a median survival period of 316 days and hazard ratio of 17.52 compared with that of littermate controls (log-rank [Mantel–Cox] test, P=0.0001; hazard ratio [log rank]). (B) Urinary albumin-creatinine ratio serial analysis at indicated ages and genotypes revealed progressive proteinuria in Nphs2.Cre⁺;Adck4^flox/flox mutant mice (red hexagon), but not in littermate controls (black diamond) (n=15–17 animals per group). Dotted line displays the onset of renal failure. Note that once chronic renal failure ensues, urinary albumin excretion reduced as observed in SRNS. P values were calculated using an unpaired t test: *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001; each data point represents the mean value of technical duplicates; the error bars represent SEM. (C) Kidney serial sections and representative images of 10-month-old mice. The Nphs2.Cre⁺;Adck4^flox/flox mutant mice exhibited severe FSGS (arrows) with severe interstitial fibrosis and tubular atrophy (arrow heads). In contrast, wild-type littermate control mice displayed normal histologic kidney morphology. Scale bars, upper row 500 μm, middle row 100 μm, and lower row 20 μm. (D) Glomerular sclerosis by analyzing Masson trichrome staining. n=3–4 mice at 10 months of age in each group; each graph bar indicates a single animal and >70 glomeruli were counted per an animal; P values were calculated using an unpaired t test; ****P<0.0001. H&E, hematoxylin and eosin; PAS, Periodic acid–Schiff.

Figure 2.

Nphs2.Cre⁺;Adck4^flox/flox mice exhibited glomerulopathy. (A) Immunofluorescence staining for the slit diaphragm protein podocin (green) and the basement membrane marker nidogen (red). A normal expression pattern of podocin was observed in 10-month-old littermate control mice. Nphs2.Cre⁺;Adck4^flox/flox mice showed mostly omitted podocin staining (arrows), appearing only on a few capillary loops (arrowhead). (B) Quantification of antibody staining in (A) demonstrated a significantly decreased expression of podocin in glomeruli of Nphs2.Cre⁺;Adck4^flox/flox mice compared with glomeruli of control mice. (C) TEM representative images of mice at the age of 10 months. Nphs2.Cre⁺;Adck4^flox/flox mice revealed severe podocyte foot process effacement (arrows) and an increased amount of dysmorphic mitochondria (*). Glomerular basement membrane (GBM) is highlighted by a dotted line. Scale bars, 10 μm left panel, 2 μm middle and lower right panels, 1 µm upper right panel, and 250 nm high magnification panel. (D) Nphs2.Cre⁺;Adck4^flox/flox mutant mice showed significant loss of filtration slits per micron of glomerular basement membrane. n=3–4 mice in each group, two to three glomeruli per animal were analyzed. P values were calculated using an unpaired t test; ***P<0.001, ****P<0.0001, error bars represent SD. FIU, fluorescence intensity units.

Treatment with 2,4-diHB Prevented the Development of Renal Pathology in Podocyte-Specific Adck4 KO Mice

Given that albuminuria started at around 4 months of age and renal structural abnormalities and functional decline manifested relatively late in Adck4^ΔPodocyte mice, we decided to initiate the treatment with 2,4-diHB at a concentration of 25 mM in drinking water when the mice were 3 months old, to prevent the disease onset and to mitigate disease progression. Supplementation of 2,4-diHB did not have any effect on the survival rate, albumin-creatinine ratio, and kidney function and histology of the control mice (Figure 3, Supplemental Figure 6). Adck4^ΔPodocyte mice treated with 2,4-diHB showed a normal survival rate despite maintaining proteinuria compared with that of healthy treated littermate controls (Figure 3, A and B) and a significantly improved survival rate (P=0.0078) compared with that of nontreated Adck4^ΔPodocyte mice, which displayed an increased mortality rate progressing to ESKD with a median survival period of 316 days and hazard ratio of 9.36 (Supplemental Figure 6B). The mortality rate reduction in Adck4^ΔPodocyte mice treated with 2,4-diHB was associated with significantly improved plasma albumin level and renal function (Supplemental Figure 6, B–E). This revealed normal glomerular histology and a significantly reduced rate of sclerotic glomeruli (mean 14.76%) up to 18 months of age in treated Adck4^ΔPodocyte mice (Figure 3, C and D, Supplemental Figure 7, A and B). The improvement in functional, histologic, and ultrastructural findings in Adck4^ΔPodocyte mice treated with 2,4-diHB was also associated with significantly improved expression of nephrin (Figure 4, A and B) and synaptopodin (Supplemental Figures 8 and 9). Moreover, Adck4^ΔPodocyte mice treated with 2,4-diHB showed significantly reduced expression of the fibrotic marker αSMA (Figure 4C). Treatment with 2,4-diHB served to maintain the normal podocyte morphology and configuration at the ultrastructural level in the glomeruli of Adck4^ΔPodocyte mice (Figure 4D) by preserving normal slit morphology. However, the filtration-slit frequency was found to decrease compared with that of littermate controls (Figure 4E). In summary, the treatment of Adck4^ΔPodocyte mice with 2,4-diHB significantly prevented the development of FSGS and foot process effacement, while maintaining normal renal function in treated mice at 18 months of age. Therefore, our findings demonstrate that 2,4-diHB is effective in protecting against renal disease progression and in improving survival in Adck4^ΔPodocyte mice.

Figure 3.

Treatment of Nphs2.Cre⁺;Adck4^flox/flox mutant mice with 2,4-diHB prevented FSGS progression, resulting in normal survival rate. (A) Urinary albumin-creatinine ratio serial analysis at indicated ages and genotypes (n=9–11 animals per group). Nphs2.Cre⁺;Adck4^flox/flox mutant mice treated with 2,4-diHB (green square) were protected from developing severe, progressive proteinuria, although proteinuria was significantly increased compared with that in healthy treated littermate controls (black circle). Green arrow indicates the start of treatment. Each data point represents the mean value of technical duplicates; the error bars represent SEM. (B) Nphs2.Cre⁺;Adck4^flox/flox mutant mice treated with 2,4-diHB presented similar survival rate as that of healthy treated littermate controls (log-rank [Mantel–Cox] test, P=0.797). (C) Kidney serial sections and representative images of 18-month-old mice. The wild-type littermate control mice and Nphs2.Cre⁺;Adck4^flox/flox mutant mice treated with 2,4-diHB displayed normal histologic kidney morphology. Scale bars, upper row 500 μm, middle row 100 μm, and lower row 20 μm. (D) Glomerular sclerosis by analyzing Masson trichrome staining. n=3–4 mice at 18 months of age in each group; each graph bar indicates a single animal and >69 glomeruli were counted per an animal. P values were calculated using an unpaired t test; *P<0.05, **P<0.01, ***P<0.001. H&E, hematoxylin and eosin; PAS, Periodic acid–Schiff.

Figure 4.

Treatment with 2,4-diHB prevented loss of podocytes in Nphs2.Cre⁺;Adck4^flox/flox mice. (A) Immunofluorescence staining of 18-month-old mice for the slit diaphragm protein nephrin (green) and the glomerular fibrosis marker αSMA (red). Scale bars, 20 μm. (B and C) Quantification of antibody staining in (A). The (B) nephrin and (C) αSMA expression of Nphs2.Cre⁺;Adck4^flox/flox mice treated with 2,4-diHB was comparable with that of treated littermate control mice. n=3 images in each group were analyzed, P values were calculated using ordinary one-way ANOVA/Tukey multiple comparisons test: *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001, error bars represent SD. (D) Representative TEM images of mice at the age of 18 months. Nphs2.Cre⁺;Adck4^flox/flox mice treated with 2,4-diHB displayed mild foot process morphology changes with infrequent regions of effacement (arrows). The mitochondrial morphology remained normal. Scale bars, 10 μm left panel and 2 μm for middle and right panels. (E) Nphs2.Cre⁺;Adck4^flox/flox mice treated with 2,4-diHB at the age of 18 months displayed reduced frequency of filtration slits per micrometer of glomerular basement membrane (GBM) compared with littermate controls under treatment; however, they still had significant remaining filtration slits. n=3–4 mice in each group, two to three glomeruli per animal were analyzed; P values were calculated using an unpaired t test; *P<0.05, ****P<0.0001, error bars represent SD. FIU, fluorescence intensity units.

Loss of ADCK4 Caused Mitochondrial Defects in Podocytes

To investigate the function of ADCK4 at the cellular level, we generated ADCK4 KO cells of human podocytes and HK-2 cells, which originated from human proximal tubule epithelial cells. Each cell line was subjected to deletion of exon 6 of the ADCK4 gene using the CRISPR/Cas9 system, and the absence of ADCK4 expression was confirmed by immunoblotting (Supplemental Figure 10, A–D). KO of ADCK4 did not affect the viability of both cell lines (Supplemental Figure 10E). We have previously shown that the level of CoQ₁₀ decreased in fibroblasts and lymphoblasts derived from patients with ADCK4 mutations,¹² but not in podocytes. Therefore, in this study, we verified this finding using the established KO cells and found that ADCK4 KO resulted in decreased CoQ₉ in both cultured podocytes and HK-2 cells compared with that in control cells (Figure 5A). However, CoQ₁₀ was reduced only in cultured podocytes, but not in HK-2 cells (Figure 5B). The basal CoQ₁₀ level in podocytes was threefold higher than that in HK-2 cells (Figure 5B). In addition, we measured the CoQ content of the glomeruli isolated from control and Adck4^ΔPodocyte mice and found a decreasing trend for both CoQ₉ and CoQ₁₀ in Adck4^ΔPodocyte mice; however, this trend was not statistically significant (Supplemental Figure 11). Because CoQ shuttles electrons from complexes I and II to complex III in the mitochondrial respiratory chain,² the activity of complex II-III is dependent on the CoQ₁₀ level in the mitochondria. Therefore, we measured the activity of complexes II and II-III and found that the activity of both was significantly reduced in ADCK4 KO podocytes (Figure 5C) compared with that in control cells, but not in ADCK4 KO HK-2 cells (Figure 5D). Decreased complex II-III activity observed in ADCK4 podocytes was partially rescued by the addition of 2,4-diHB to culture media (Figure 5E). The reduced form of CoQ (QH2) plays a role as a potent lipid-soluble antioxidant, scavenging free radicals and preventing lipid peroxidative damage.¹ Although ADCK4 KO in itself did not affect the viability of cultured podocytes (Supplemental Figure 10E), we examined cell viability upon AA treatment because CoQ-deficient yeast mutants were found to be more sensitive to polyunsaturated fatty acids, such as AA, which are prone to autoxidation and break down into toxic products.³⁵ AA treatment reduced cell viability in both the control and ADCK4 KO podocytes, and ADCK4 KO podocytes were relatively more affected (Figure 5F). Decreased cell viability by AA treatment was rescued by supplementation of 2,4-diHB (Figure 5F). Overall, these findings suggested that the loss of ADCK4 caused CoQ deficiency and that podocytes were more susceptible than HK-2 cells.

Figure 5.

Coenzyme Q was deficient in ADCK4 KO podocytes. (A and B) Coenzyme Q contents of cultured podocytes and HK-2 cells. The CoQ₉ level was decreased in (A) both cultured podocytes and HK-2 cells and the CoQ₁₀ level was severely deficient in (B) ADCK4 KO podocytes. (C and D) Respiratory chain complex II and succinate-cytochrome c reductase (complex II-III) enzyme activities were measured in (C) podocytes and (D) HK-2 cells. Complex II-III activities were decreased only in podocytes, whereas complex II activities were affected in both cell lines. (E) Succinate-cytochrome c reductase (complex II-III) enzyme activities were measured in podocytes. Decreased activities in ADCK4 KO podocytes were partially restored by the addition of 500 μM 2,4-diHB. (F) Cell viability was measured using the Cell Count Kit-8 assay. ADCK4 KO podocytes exhibited susceptibility to 30 μM AA treatment. Decreased cell viability was reversed by the addition of 500 μM 2,4-diHB. P values were calculated using an unpaired t test; *P<0.05, **P<0.005; error bars represent mean±SD. TTFA, thenoyltrifluoroacetone.

To understand the molecular changes induced by ADCK4 KO, we performed proteomic analysis and quantified changes in protein abundance by MS-based proteomics using iTRAQ³² in podocytes with or without AA treatment. Proteomic characterization of the control and ADCK4 KO podocytes revealed >2500 proteins, 421 (16%) of which were mitochondrial proteins. By GO analysis, using the DAVID functional annotation tool, differentially expressed proteins in the control and ADCK4 KO cells were annotated based on biologic process. The results indicated that proteins related to cellular defense response were upregulated and those associated with cytokine production pathway were downregulated in ADCK4 KO podocytes compared with those in the control cells (Supplemental Figure 10F). In an injury situation, with the use of AA treatment, coenzyme metabolism–related proteins and intermediate filament–related proteins were downregulated, whereas DNA regulation proteins were upregulated in ADCK4 KO podocytes (Supplemental Figure 10G).

Disrupted Mitochondrial Morphology and Mitochondrial Membrane Potential Were Observed in ADCK4 KO Podocytes

Abnormal proliferation of polymorphous mitochondria in the cytoplasm of podocytes is one of the characteristic ultrastructural findings of CoQ₁₀-related diseases.¹¹ Similarly, patients with ADCK4 mutations showed mitochondrial abnormalities in podocytes and proximal tubules.²⁶ We examined the ultrastructure of mitochondria in control and ADCK4 KO cells by TEM. The formation of cristae was disrupted and the shape of mitochondria was disintegrated in ADCK4 KO podocytes (Figure 6, A and B), whereas ADCK4 KO HK-2 cells showed normal features of the mitochondria (Supplemental Figure 12A). We also observed the effect of AA treatment on the ultrastructure of mitochondria by TEM. Although the mitochondria of control podocytes were less affected by AA, AA-treated ADCK4 KO podocytes showed more severe mitochondrial defects such as swollen and shortened cristae and fewer inner membranes (Figure 6, A and B). These results indicate that ADCK4 KO confers susceptibility to cellular stress, such as autoxidation of polyunsaturated fatty acids. To examine functional defects of mitochondria in ADCK4 KO cells, we measured the mitochondrial membrane potential using JC-10 dye, which is concentrated in the mitochondrial matrix based on membrane polarization.³⁶ The JC-10 fluorescence intensity was significantly reduced in the ADCK4 KO cells (Figure 6C). This finding was confirmed by another mitochondrial membrane potential assay using the potentiometric probe tetramethylrhodamine methyl ester–based fluorimetric assay. Tetramethylrhodamine methyl ester fluorescence intensity was significantly reduced in ADCK4 KO cells; however, supplementation of 2,4-diHB partially restored the lowered mitochondrial membrane potential (Figure 6D). In contrast, mitochondrial membrane potential was not different between the control and KO in HK-2 cells (Supplemental Figure 12, B and C). In addition, we examined the proteins of the mitochondrial oxidative phosphorylation system and found that UQCRC2 of complex III was significantly decreased in ADCK4 KO podocytes compared with control podocytes, whereas proteins in other complexes were not different (Supplemental Figure 13, A and B). However, in glomerular lysates, there was no difference in oxidative-phosphorylation-system complexes between control and Adck4^ΔPodocyte mice (Supplemental Figure 13, C and D). Therefore, the loss of ADCK4 caused morphologic and functional defects of mitochondria in cultured podocytes by disrupting CoQ₁₀ biosynthesis.

Figure 6.

ADCK4 KO podocytes showed mitochondrial defects. (A) TEM of podocytes showing mitochondrial morphology. Black and white boxed areas are enlarged mitochondria in ADCK4 KO podocytes showing abnormal fission and disrupted cristae (black arrows). Mitochondria after AA treatment were severely disrupted in ADCK4 KO podocytes. Scale bars, 0.5 μm in first column, 0.1 μm in second column, and 0.05 μm in third and fourth columns. (B) Morphometric analyses showed that ADCK4 KO podocytes had similar number of mitochondria, but cristae number was decreased compared with control cells. AA treatment further decreased mitochondrial volume density in ADCK4 podocytes. (C and D) Mitochondrial membrane potential (ΔΨ) was measured using (C) JC-10 and (D) tetramethylrhodamine methyl ester (TMRM). ADCK4 KO podocytes showed reduced ΔΨ compared with that of control podocytes. Reduced ΔΨ of ADCK4 KO podocytes was partially rescued by the addition of 500 μM 2,4-diHB. (E) The mitochondrial superoxide level measured using the MitoSOX reagent increased in ADCK4 KO cells. ADCK4 KO cells exhibited higher sensitivity to treatment with H₂O₂, tBHP, and especially AA. Addition of 2,4-diHB decreased mitochondrial superoxide levels in both control and ADCK4 KO podocytes. P values in (B–E) were calculated using an unpaired t test; *P<0.05, **P<0.005, ***P<0.0001; the error bars represent mean±SD.

CoQ acts as an antioxidant against ROS³⁷; therefore, we investigated ROS production in ADCK4 KO podocytes. Mitochondrial superoxide, assessed by the MitoSOX reagent, increased in ADCK4 podocytes compared with control podocytes in the basal state, and the difference became more significant upon stimulation with 500 μM H₂O₂, 100 μM tBHP, 30 μM AA, or 100 μM AA (Figure 6E). In addition, cellular ROS production, measured using the CellROX reagent, also increased in ADCK4 podocytes compared with control podocytes in the basal state, and treatment with H₂O₂, tBHP, or AA increased ROS production overall. However, no significant difference was observed between control and ADCK4 KO podocytes (Supplemental Figure 14).

Because we previously reported that ADCK4 knockdown reduced podocyte migration,¹² in this study we examined the cytoskeleton of ADCK4 KO cells. Actin phalloidin staining revealed shrunk cellular area in ADCK4 KO podocytes (Figure 7), whereas ADCK4 KO HK-2 cells showed a surface area similar to that of control HK-2 cells (Supplemental Figure 12D). The mitochondrial marker COX IV was not affected by ADCK4 KO in podocytes or HK-2 cells (Figure 7, Supplemental Figure 12D). Interestingly, shrunk cellular area in ADCK4 KO podocytes was not restored by 2,4-diHB supplementation, suggesting that this cellular phenotype is not related to decreased CoQ₁₀ level and that ADCK4 might have other cellular functions in addition to its role in the CoQ biosynthesis pathway. AA treatment significantly reduced cellular area in both the control and ADCK4 KO podocytes (Figure 7). Shrunk cellular area by AA treatment was not rescued by 2,4-diHB, further confirming that this phenotype is not related to CoQ₁₀ deficiency (Figure 7).

Figure 7.

ADCK4 KO resulted in decreased cellular area. (A) Immunofluorescence of COXIV and phalloidin staining in podocytes. COXIV marks mitochondria, whereas phalloidin marks actin cytoskeleton. Staining intensity of COXIV or phalloidin was not different between control and ADCK4 KO podocytes. Scale bars, 50 μm. (B) Quantification of cellular area. Phalloidin-stained area was shrunk in ADCK4 KO podocytes, and it became more prominent upon AA treatment. Decreased cellular area was not reversed by the addition of 2,4-diHB. A total of 100 cells from three independent experiments; P values were calculated using an unpaired t test; *P<0.05, **P<0.005; the error bars represent mean±SD.

ADCK4 Interacted with and Stabilized COQ Proteins

We performed proteomic analysis to understand the function of ADCK4 via the identification of its interactome. We generated HEK293 cells that stably overexpressed C-terminal, FLAG-tagged BAP (BAP-3xFLAG) and ADCK4 (ADCK4-3xFLAG) (Supplemental Figure 15A). We confirmed that ADCK4-3xFLAG mostly localized to the mitochondria (Supplemental Figure 15B). Following affinity purification using anti-FLAG beads (Supplemental Figure 15, C and D), protein eluates were analyzed using a liquid chromatograph coupled to a high-resolution mass spectrometer (LC-MS/MS). In total, 612 proteins were identified as interactors of ADCK4. Among them, the cytoplasmic proteins, including myosin (MYH10, MYH11, MYO1B, and MYO1C), filamin (FLNC), and kinase proteins (STK24, STK25, STK38, and ROCK1) were detected. In addition, the mitochondrial proteins, including ATP synthase subunit (ATP5L), cytochrome c oxidase subunit (COX6A1 and UQCRQ), and COQ5 were also identified as interactors (Figure 8A). GO analysis of mitochondrial interactors showed that these proteins are involved in transferase activity, oxidoreductase activity, nucleotide binding, and ATPase activity (Figure 8B). Because COQ5, functioning as a C-methyltransferase in the CoQ biosynthesis pathway,³⁸ was identified as an interactor of ADCK4, we confirmed the interaction in podocytes by coimmunoprecipitation (Supplemental Figure 16A) and examined COQ proteins in ADCK4 KO podocytes. Although mRNAs of several COQ genes were not affected by ADCK4 KO (Supplemental Figure 16B), the protein levels of COQ3, COQ5, and COQ9 significantly decreased in ADCK4 KO podocytes compared with control podocytes, indicating these complex Q proteins were destabilized in the absence of ADCK4 (Figure 8C). The decrease in the level of these proteins was further confirmed in glomerular lysates of Adck4^ΔPodocyte mice (Supplemental Figure 16C). Decreased COQ5 was restored not only by transfection of wild-type ADCK4, but also by 2,4-diHB (Figure 8D). In addition, we examined the effect of ADCK4 mutations, which were previously identified in individuals with nephrotic syndrome.¹² COQ5 was also rescued by ADCK4 mutant proteins; however, the extent of the rescue was less than that by wild-type ADCK4 (Figure 8E). In addition, because STK38 is a negative regulator of MAPKKK1/2 kinase signaling,³⁹ we examined the MAPK pathway by Western blotting and found that phosphorylated ERK1/2 was significantly increased in ADCK4 KO podocytes (Figure 8F). Moreover, we observed considerably enhanced MAPK signaling including p-p38, p-ERK1/2, and p-JNK under lipid peroxidation injury induced by AA (Figure 8F). Because Coq2 silencing resulted in increased autophagy and mitophagy in Drosophila nephrocytes,⁴⁰ we examined mTOR and LC3 in ADCK4 KO podocytes; however, we could not observe increased LC3-II expression (Supplemental Figure 17). Taken together, ADCK4 contributes to stabilizing the Q complex, elucidating CoQ deficiency in the absence of ADCK4, and other signaling pathways were also affected upon ADCK4 KO in podocytes.

Figure 8.

ADCK4 interacted with COQ5 and stabilized complex Q in podocytes. (A) ADCK4-interacting proteins isolated form HEK293 cells overexpressing ADCK4-3XFLAG. Both cytoplasmic and mitochondrial proteins were detected as interactors by nanoflow LC-MS/MS. (B) GO analysis showed that ADCK4-interacting mitochondrial proteins are associated with transferase activity, oxidoreductase activity, nucleotide binding, and ATPase activity. (C) Immunoblot showed that proteins in complex Q, COQ3, COQ5, and COQ9 were significantly reduced in ADCK4 KO podocytes. (D) Decreased COQ5 protein level was rescued by heterologous ADCK4 expression and 2,4-diHB treatment in ADCK4 KO podocytes. (E) Effect of ADCK4 mutations on COQ5 rescue in ADCK4 KO podocytes. COQ5 was rescued to a lower extent by truncated ADCK4 mutant proteins, whereas ADCK4 mutant proteins harboring missense variations were not different from wild-type ADCK4. (F) Immunoblot analysis of the MAPK pathway. p-p38 and p-pERK were increased in ADCK4 KO podocytes and p-p38 was further induced by AA. Unphosphorylated p38 was used as the loading control. Densitometry analyses in (C–E) are representative of at least three independent experiments and band intensities were normalized to that of β-actin. *P<0.05, **P<0.005; t test.

Discussion

In this study, we demonstrated that podocyte-specific deletion of Adck4 in mice resulted in proteinuria and foot process effacement, recapitulating the features of nephrotic syndrome caused by ADCK4 mutations. These defects were efficiently ameliorated by treatment with 2,4-diHB, an unnatural precursor analogue of the CoQ biosynthesis pathway. ADCK4 KO podocytes exhibited reduced activity of complex II and III, mitochondrial defects, and sensitivity to AA, which resulted from CoQ deficiency. ADCK4 mutations cause adolescence-onset nephrotic syndrome, which often progresses to ESKD in the second decade of life.¹⁴ The late onset of renal disease is differentiated from nephropathy resulting from mutations in WT1, NPHS1, or NPHS2, which generally manifests in the first year of life. Similarly, in this study, Adck4^ΔPodocyte mice exhibited renal disease, beginning at approximately 4 months, and progressed to ESKD by 12 months. ADCK4-associated glomerulopathy can be partially treated by CoQ₁₀ supplementation; however, it is not always successful.^12,14,41 Failure to respond to CoQ supplementation can be attributed to several possible causes, one of which is related to the progression of renal disease to an irreversible stage. Therefore, early genetic diagnosis is necessary to recognize ADCK4 mutations. In this regard, because the onset of ADCK4-associated glomerulopathy occurs relatively late in life, if properly diagnosed, its therapy can be initiated before the disease becomes fulminant. Another reason for the failure of CoQ supplementation might be the poor oral availability of CoQ, which makes its therapeutic efficacy variable and limited. In this study, we demonstrated that 2,4-diHB efficiently ameliorated proteinuria and prevented FSGS in Adck4^ΔPodocyte mice. Treatment with 2,4-diHB has been shown to bypass defects in the penultimate step of CoQ biosynthesis, which is mediated by Coq7(Clk1) hydroxylase (Figure 9).^16,17 Further, 2,4-diHB has also been shown to be more effective in tamoxifen-inducible Mclk1/Coq7 KO mice than CoQ.¹⁵ It seems more readily absorbed than CoQ and is safe, because it has been used as a food flavor modifier due to its sweet taste. Therefore, translational studies are required to investigate whether 2,4-diHB can be effective in individuals with ADCK4 mutations.

Figure 9.

Proposed pathway for the synthesis of CoQ₁₀ via alternative ring precursors. 4-Hydroxybenzoic acid (4-HB) is the canonic aromatic ring precursor used in CoQ₁₀ biosynthesis. The alternative ring precursors 2,4-diHB and vanillic acid also serve as substrates for COQ2. After attachment of the decaprenyl tail, the other COQ polypeptides that participate in the canonic biosynthetic pathway serve to modify the ring so that CoQ₁₀ may be produced. This allows bypassing of certain steps. Some of the late-stage CoQ intermediates are needed to stabilize the Q complex (CoQ synthome). Cells with deficiencies in certain steps of the pathway show stabilization of certain COQ polypeptides in the presence of the alternate ring precursors, which can bypass the deficient steps. 4HB, 4-hydroxybenzoic acid; DMQ, demethoxy-coenzyme Q; DMeQ, demethyl-coenzyme Q; decaprenyl PP, decarprenyl pyrophosphate; DMAPP, dimethylallyl pyrophosphate; IPP, isopentenyl pyrophosphate; 2,4-diHB, 2,4-dihydroxybenzoic acid.

2,4-diHB is expected to be beneficial for enzymatic deficiency in the CoQ biosynthetic pathway because it bypasses the defect in COQ7.^1,15 It is, therefore, interesting that 2,4-diHB rescued disease phenotypes in Adck4^ΔPodocyte mice because ADCK4, although an uncharacterized mitochondrial protein, is not an enzyme directly involved in the CoQ biosynthetic pathway. This finding suggests that ADCK4 supports an enzymatic component in CoQ biosynthesis. Via proteomic analysis, we found that ADCK4 interacts with COQ5 and the expression of proteins in complex Q—namely, COQ3, COQ5, and COQ9—significantly decreased in ADCK4 KO podocytes. It has been previously suggested that a physical and functional interaction between ADCK3 and COQ5 is important.^24,25

In this study, phalloidin staining showed that the cytoskeleton of only podocytes was defective, but not that of HK-2 cells. This observation suggests that mitochondrial dynamics may also play important roles in maintaining the shape and function of podocytes because podocytes, like neurons,⁴² require a proper supply and high amount of energy to maintain their foot processes. In addition, because the foot process has rich microfilaments, these interactions may also participate in podocyte homeostasis.^43,44 Moreover, these cytoskeletal defects in ADCK4 KO podocytes were also consistent with the iTRAQ analysis data, which elucidated that cytoskeleton-related proteins were downregulated in ADCK4 KO podocytes compared with those in control podocytes. In contrast to other cellular defects, the shrunken cytoskeleton of ADCK4 KO podocytes was not rescued by 2,4-diHB, suggesting that this cellular phenotype may not be related to CoQ₁₀ deficiency. This may also explain why CoQ₁₀ supplementation is not always effective. In addition, it has recently been shown that podocytes predominantly rely on glycolysis and that mitochondrial defects are not sufficient to cause FSGS,⁴⁵ suggesting the pathogenic mechanism of ADCK4 mutations is complex.

CoQ₁₀ is well known for its antioxidant activity, thereby protecting cells from oxidative stress.³ In this study, we treated the cells with AA, one of the polyunsaturated fatty acids, to induce lipid peroxidation stress and found that the MAPK pathway signaling was activated in ADCK4 KO podocytes. The MAPK signaling pathway is essential in regulating several cellular processes including inflammation, cell stress response, and cell proliferation.⁴⁶ In addition, AA treatment more significantly reduced cell viability in ADCK4 podocytes than in control podocytes and the reduced viability was rescued by the supplementation of 2,4-diHB. The reduced form of CoQ₁₀ may act to scavenge lipid peroxyl radicals and function as an antioxidant, thereby preventing the initiation of lipid peroxidation, because it has been reported to eliminate perferryl radicals.^1,47 In this regard, ADCK4 KO may confer hypersensitivity to lipid peroxidation stress. This suggests that cellular stress may be necessary to develop renal diseases in addition to loss of ADCK4, partially explaining the relative late onset of nephrotic syndrome resulting from ADCK4 mutations.

Recent studies have revealed that ADCK3 lacks canonic protein kinase activity in the trans form; instead, it binds to lipid CoQ₁₀ intermediates and exhibits ATPase activity.^25,48 In this study, GO analysis also revealed that interactors of ADCK4, especially in the mitochondria, are significantly associated with oxidoreductase activity, which can be related to the antioxidant property of CoQ₁₀. Furthermore, proteomic analysis revealed the ATP-binding proteins as ADCK4 interactors, suggesting that ADCK4 may have the ATPase activity, like ADCK3. Yet, the precise role of ADCK4 is not clear and further studies are required to verify whether ADCK4 has ATPase or kinase activity toward an undiscovered substrate.

In conclusion, our study suggests that ADCK4 in podocytes stabilizes proteins in complex Q in podocytes, and thereby contributes to CoQ synthesis and plays a role in maintaining the cytoskeleton structure. Cellular defects and renal phenotypes by ADCK4 deficiency were mostly rescued by 2,4-diHB supplementation, an unnatural precursor analogue of CoQ₁₀, demonstrating the important role of ADCK4 in the CoQ biosynthesis pathway. Our study provides insights into the functions of ADCK4 in CoQ biosynthesis and pathogenesis of nephrotic syndrome.

Disclosures

Dr. Hildebrandt is a cofounder of Goldfinch-Bio. The authors have nothing to disclose.

Funding

This study was supported by National Institutes of Health grant DK076683 (to Dr. Hildebrandt) and National Science Foundation grant MCB-1330803 (to Dr. Clarke). Dr. Hildebrandt is the William E. Harmon Professor. Dr. Gee was supported by the Chung-Am (TJ Park) Science Fellowship and the Research Program through the National Research Foundation of Korea funded by the Korean Government (Ministry of Science and ICT, South Korea), grant 2018R1A5A2025079. Dr. Widmeier was supported by the Leopoldina Fellowship Program, Deutsche Akademie der Naturforscher Leopoldina – Nationale Akademie der Wissenschaften (German National Academy of Sciences Leopoldina) grant LPDS 2015-07.

Supplemental Material

This article contains the following supplemental material online at http://jasn.asnjournals.org/lookup/suppl/doi:10.1681/ASN.2019070756/-/DCSupplemental.

Supplemental Figure 1. Generation and validation of the Nphs2.Cre⁺;Adck4^flox/flox mouse model.

Supplemental Figure 2. Renal histology of 2 month-old Nphs2.Cre⁺;Adck4^flox/flox mice.

Supplemental Figure 3. Body weight of non-treated and treated Nphs2.Cre⁺;Adck4^flox/flox mice.

Supplemental Figure 4. The expression of podocyte markers was decreased in Nphs2.Cre⁺;Adck4^flox/flox mice.

Supplemental Figure 5. Deletion of Adck4 led to increased mitochondria aspect ratio as mice age.

Supplemental Figure 6. Treatment with 2,4-diHB protected from the renal failure of Nphs2.Cre⁺;Adck4^flox/flox mice.

Supplemental Figure 7. Treatment with 2,4-diHB prevented focal segmental glomerular sclerosis in Nphs2.Cre⁺;Adck4^flox/flox mice.

Supplemental Figure 8. Treatment with 2,4-diHB prevented loss of expression pattern of podocyte markers and glomerular fibrosis in Nphs2.Cre⁺;Adck4^flox/flox mice.

Supplemental Figure 9. Treatment with 2,4-diHB prevented loss of synaptopodin expression in Nphs2.Cre⁺;Adck4^flox/flox mice.

Supplemental Figure 10. Generation of ADCK4 knockout cells and characterization of ADCK4 deficient podocyte using iTRAQ.

Supplemental Figure 11. Coenzyme Q contents of Nphs2.Cre+;Adck4^flox/flox mice.

Supplemental Figure 12. ADCK4 knockout (KO) did not cause mitochondrial defects in HK-2 cells.

Supplemental Figure 13. The effect of loss of ADCK4 on OXPHOS complexes.

Supplemental Figure 14. Quantitation and statistical analysis of oxidative stress based on staining with CellROX oxidative stress regents.

Supplemental Figure 15. Purification of ADCK4-binding proteins for LC/MS-MS.

Supplemental Figure 16. ADCK4 stabilized COQ proteins at the protein level.

Supplemental Figure 17. AA did not induce autophagic marker in ADCK4 KO podocytes.