Dynein-Mediated Trafficking: A New Mechanism of Diabetic Podocytopathy

Abstract

Key Points

• The expression of dynein is increased in human and rodent models of diabetic nephropathy (DN), eliciting a new dynein-driven pathogenesis.

• Uncontrolled dynein impairs the molecular sieve of kidney by remodeling the postendocytic triage and homeostasis of nephrin.

• The delineation of the dynein-driven pathogenesis promises a broad spectrum of new therapeutic targets for human DN.

Background

Diabetic nephropathy (DN) is characterized by increased endocytosis and degradation of nephrin, a protein that comprises the molecular sieve of the glomerular filtration barrier. While nephrin internalization has been found activated in diabetes-stressed podocytes, the postinternalization trafficking steps that lead to the eventual depletion of nephrin and the development of DN are unclear. Our work on an inherited podocytopathy uncovered that dysregulated dynein could compromise nephrin trafficking, leading us to test whether and how dynein mediates the pathogenesis of DN.

Methods

We analyzed the transcription of dynein components in public DN databases, using the Nephroseq platform. We verified altered dynein transcription in diabetic podocytopathy by quantitative PCR. Dynein-mediated trafficking and degradation of nephrin was investigated using an in vitro nephrin trafficking model and was demonstrated in a mouse model with streptozotocin (STZ)-induced DN and in human kidney biopsy sections.

Results

Our transcription analysis revealed increased expression of dynein in human DN and diabetic mouse kidney, correlated significantly with the severity of hyperglycemia and DN. In diabetic podocytopathy, we observed that dynein-mediated postendocytic sorting of nephrin was upregulated, resulting in accelerated nephrin degradation and disrupted nephrin recycling. In hyperglycemia-stressed podocytes, Dynll1, one of the most upregulated dynein components, is required for the recruitment of dynein complex that mediates the postendocytic sorting of nephrin. This was corroborated by observing enhanced Dynll1-nephrin colocalization in podocytes of diabetic patients, as well as dynein-mediated trafficking and degradation of nephrin in STZ-induced diabetic mice with hyperglycemia. Knockdown of Dynll1 attenuated lysosomal degradation of nephrin and promoted its recycling, suggesting the essential role of Dynll1 in dynein-mediated mistrafficking.

Conclusions

Our studies show that hyperglycemia stimulates dynein-mediated trafficking of nephrin to lysosomes by inducing its expression. The decoding of dynein-driven pathogenesis of diabetic podocytopathy offers a spectrum of new dynein-related therapeutic targets for DN.

Introduction

Diabetic nephropathy (DN) is the most prevalently acquired podocytopathy in humans and contributes to more than 50% of the cases of end-stage glomerulopathy.¹ Hyperglycemia and associated glomerular hyperperfusion are considered the major causes of DN; tight glucose control has clearly been shown to reduce the incidence of microalbuminuria and postpone the development of DN.^2–4 However, we lack a clear understanding of how hyperglycemia directly causes kidney injury, which in turn limits our ability to develop effective therapies.

Podocytes are highly differentiated, epithelial cells that encapsulate the glomerular capillary loops. The leading edge of podocytes branches into foot processes that interdigitate with those of neighboring podocytes to form slit diaphragms (SDs), creating the molecular sieve of the glomerular filtration barrier (GFB). The SD consists of the extracellular domain of nephrin anchored in the cell membrane of the foot processes.^5–7 Microalbuminuria, an early presentation of DN, reflects an impaired GFB due to podocyte injuries that range from foot process effacement to more severe situation of podocyte detachment as illustrated by electron microscopy.^8,9 In diabetic mouse models and in human DN, loss of nephrin and disruption of the normal SD architecture precedes the onset of microalbuminuria.^10,11 This observation suggests that altered nephrin homeostasis influences the nanostructure of SDs during the initial phase of diabetic podocyte injury.^12,13 Although studies have found that high glucose (HG) stimulates the internalization of nephrin, there is a critical gap between the initially observed nephrin internalization from the surface membrane and the final depletion of nephrin. Our study extends the current understanding of mistrafficking in diabetic podocyte injury by focusing on the understudied “postendocytic sorting,” a process that is independent on the transient and short distanced actin-based trafficking and the early diabetes signaling pathways (protein kinase C α/β-arrestin,¹⁴ regulator for ubiquitous kinase/c-Cbl–interacting protein of 85 kDa,¹⁵ protein kinase C, and casein kinase substrate in neurons 2¹⁶ or IQ (Isoleucine and Glutamine) motif-containing GTPase (3-phospho-D-glycerate 1-phosphotransferase)-activating proteins¹⁷), as well as regulation of nephrin phosphorylation (Src Homology-Containing transforming protein A [ShcA]).¹⁸

Among the motor proteins that mediate postendocytic sorting, kinesin and cytoplasmic dynein guide reciprocal anterograde and retrograde transport of cargo proteins along the microtubule, between the surface membrane and the cytosol.¹⁴ The dynein transport complex consists of heavy chains, the Adenosine 5′-Triphosphatase (ATPase) subunit to generate energy for the sliding of the complex along microtubule¹⁹; light intermediate chains and intermediate chains that anchor cargo proteins or recruit adapter proteins²⁰; light chains that assemble and activate the whole complex^21,22; and dynactin 1, a required activator for the retrograde trafficking of dynein.²⁰ Dynein plays a unique role in modulating protein metabolism by mediating postendocytic sorting of proteins and targeting them for ubiquitin-dependent degradation.^19,23–26 This study was initially motivated by our recent finding that dysregulated dynein-mediated trafficking in inverted formin 2 (INF2)–related podocytopathy was responsible (Supplementary Figures 1,2) for nephrin depletion.²⁵ Furthermore, when we analyzed the dynein expression profile in various human podocytopathies, we found the most prominent changes in DN. These findings suggested that a dynein-related mechanism may drive the development of podocytopathy in DN. Here, we find that hyperglycemia in diabetic mice and in cultured podocytes increases expression of the dynein complex, which in turn alters trafficking and degradation of nephrin, explaining how diabetic stresses potentially drive podocytopathy.

Materials and Methods

The primers for real-time quantitative PCR as well as the antibodies, small interfering RNA (siRNA) duplex sequences, chemical compounds, and plasmids used in this study are listed in Tables 1–5, respectively.

Table 1.

Primers for real-time quantitative PCR

Gene	Primers for Quantitative PCR (Forward Primer/Reverse Primer)
msDCTN1	ATGAGTACGGAGGCAAGCG/AGAATCACGCCCACCCATTTG
msNPHS1	GATGCGGAGTACGAGTGCC/GGGGAACTAGGACGGAGAGG
msDynll1	ATTGCGGCCCATATCAAGAAG/GTGCCACATAACTACCGAAGTTT
msDync1li1	GGGAAGACAAGCCTCATAAGAAG/AGTAGCCCTTTGTGGTACAGA
msDync1h1	AAGCACCTGCGTAAGCTGG/GCGGGTCTGACAGGAACTTG
msDync1i1	TAGTCCCAACCCCTATGTCTCC/TGCAGTCGTCTCCTTGTTAATG
msINF2	CTCCTCGTACTCAGCTTGTCC/GGTCACCAAAATCCTGGTTGTC
msGAPDH	AGGTCGGTGTGAACGGATTTG/TGTAGACCATGTAGTTGAGGTCA

INF2, inverted formin 2.

Table 5.

Plasmids

Plasmids	Company	Catalog No.
pCDNA3 mouse nephrin	Gift from Dr. Lawrence B. Holzman
GFP-Rab7	Addgene	12605
GFP-Rab11	Addgene	12674
pEGFP-DCTN1	Addgene	36154
pCMV3-C-GFPSpark-Dynll1	SinoBiochemical	MG51246-ACG
EB3-tdTomato	Addgene	50708

Table 2.

Antibodies and fluorescent probes (immunofluorescent, Western blotting, immunoprecipitation]

Antibody	Dilution	Company	Cat#
Mouse antinephrin (G-8)	IF: 1:100	Santa Cruz	sc-376522
Mouse antinephrin conjugated to agarose	IP: 2 µg/400 µl lysate	Santa Cruz	sc-376522 AC
Normal mouse IgG conjugated to agarose	IP: 2 µg/400 µl lysate	Santa Cruz	sc-2343
Rabbit antinephrin	WB: 1:1000	Invitrogen	PA5-91907
Sheep anti-DCTN1	WB: 1:1000 IF: 1:100	R&D Systems	AF6657
Mouse anti-Dynll1	IF: 1:100	Santa Cruz	sc-136287
Rabbit anti-Dynll1	WB: 1:1000 IF: 1:100	Thermo Fisher	PA5-97920
Rabbit anti-INF2	IF: 1:100	Bethyl Lab	A303-427A
Rabbit anti-HDAC6	WB: 1:1000	Novus	NBP1-78981
Donkey anti-mouse (Alexa Fluor 488)	IF: 1:200	Thermo Fisher	A32766
Donkey anti-rabbit (Alexa Fluor 488)	IF: 1:200	Thermo Fisher	A32790
Donkey anti-mouse (Alexa Fluor 594)	IF: 1:200	Thermo Fisher	A32744
Donkey anti-sheep (Alexa Fluor 488)	IF: 1:200	Thermo Fisher	A-11015
Mouse anti-rabbit IgG-HRP	WB: 1:5000	Santa Cruz	sc-2357
Rabbit anti-sheep IgG-HRP	WB: 1:5000	Thermo Fisher	31480
Biotin mouse anti-ubiquitin	WB: 1:1000	Invitrogen	13-6078-82
Mouse anti β-actin-HRP	WB: 1:5000	Santa Cruz	sc-47778
Mouse anti-WT1	IF: 1:100	Novus	NBP24460700
Mouse anti-acetylated α-tubulin	IF: 1:200	Santa Cruz	sc-23950
Acti-stain 488 phalloidin		Cytoskeleton	PHDG1-A
ViaFluor microtubule stains (488)		Biotium	70062

IF, immunofluorescent; IP, immunoprecipitation; WB, Western blotting; INF2, inverted formin 2; HDAC6, histone deacetylase 6; HRP, horseradish peroxidase.

Table 3.

siRNA duplex sequences

Target	Catalog No.	Sense (5′-3′)	Antisense (5′-3′)
Mouse Dynll1	sc-36229A	GGCCAUUCUUCUGUUCAAAtt	UUUGAACAGAAGAAUGGCCtt
Mouse Dynll1	sc-36229B	CACCUCGUUUGAAUCUGUUtt	AACAGAUUCAAACGAGGUGtt
Mouse Dynll1	sc-36229C	GGCUUCAUUCUCUGUACAAtt	UUGUACAGAGAAUGAAGCCtt
Control	sc-37007	UUCUCCGAACGUGUCACGUtt	ACGUGACACGUUCGGAGAAtt

Table 4.

Chemical compounds

Chemical Compounds	Concentration	Company	Catalog No.	CAS#
Ciliobrevin D	50 μM	Sigma	250401	1370554-01-0
Leupeptin	200 µg/ml	MedChemExpress	HY-18234	103476-89-7
Bortezomib	100 nm	Sigma	5.04314	179324-69-7
Cycloheximide	10 µM	Sigma	01810	66-81-9
STZ	50 mg/kg, i.p.a daily × 5	MedChemExpress	HY-13753	18883-66-4

STZ, streptozotocin.

^aIntraperitoneal injection. Additional doses of STZ were provided to female mice to induce significant hyperglycemia.

Transcriptome Analyses of Dynein Genes

Transcription analyses of cytoplasmic dynein genes were performed using Nephroseq kidney transcriptome databases and data mining platform (www.nephroseq.org, University of Michigan, Ann Arbor, MI). Specifically, the transcription data for dynein components were retrieved from the Hodgin Diabetes Mouse Glom dataset (2013, Affymetrix Mouse Genome 430 2.0 Array platform).²⁷ A cluster of overexpressed dynein genes were enriched in diabetic mice with a fasting glucose level >300 mg/dl (n=5), compared with mice with a fasting glucose level<300 mg/dl (n=16). In the analysis of the Woroniecka Human Diabetic Kidney Disease dataset (2011, Affymetrix Human Genome U133A 2.0 Array platform)²⁸ and the European renal cDNA bank (ERCB) nephrotic syndrome dataset (2018, RNA-Seq technique), a cluster of upregulated dynein genes were enriched in DN (n=10), compared with kidneys of healthy living donors (n=12). In the analysis of the Nakagawa CKD Kidney Dataset (2015, Agilent Whole Human Genome Microarray 4×44 K Array),²⁹ a cluster of upregulated dynein genes were enriched in CKD (n=48), compared with kidneys of healthy living donors (n=5).

In the Nephroseq data analysis platform, the two-sided t test was used to compare the differential expression profiles of two groups. The Log2 median-centered expression profile of the genes was expressed as a heatmap; the Z scores were normalized to depict relative values and color-coded. The Log2 [fold change (FC)] in expression of each individual gene in the kidney was compared between mice with fasting glucose >300 mg/dl and those with fasting glucose <300 mg/dl and between human DN and healthy living donors. Furthermore, the Pearson correlation was used to analyze the relationship of individual dynein gene expression to the fasting glucose levels and albuminuria of diabetic mice in the Hodgin Diabetes Mouse Glom dataset (n=5 for the db/db mouse model, n=21 for all diabetic mouse models included in the study) or to the GFR of human subjects in the Woroniecka Human Diabetic Kidney Disease dataset (n=9) and the ERCB Chronic Kidney Disease dataset (n=8).

Cell Culture and High Glucose–Induced Podocytopathy

An immortalized mouse podocyte cell line,²⁵ which had been generated by expression of a temperature-sensitive mutant of the SV40 large-T antigen per Saleem's protocol,³⁰ was maintained in RPMI 1640 media with 10% FBS, 1% insulin–transferrin–selenium supplement (Gibco, Grand Island, NY), and 50 IU/ml penicillin/streptomycin. Podocytes were grown at 33°C and differentiated at 37°C for 2 weeks. The medium was replaced with RPMI 1640 containing different concentrations of glucose, and mannitol was added to maintain the same osmolality among groups: (1) normal glucose (NG), glucose 5.5 mM+mannitol 24.5 mM; (2) prediabetic level of glucose (PDB), glucose 11 mM+mannitol 19 mM; and (3) HG, glucose 30 mM alone. After 72 hours of treatment, cells were collected for total RNA extraction, cell biology assays, or biochemical experiments.

Knockdown of Dynll1 in Cultured Podocytes

As previously described,²⁵ podocytes growing in a six-well plate were transfected with 0.75 µg (or 60 pmols) msDynll1 siRNA Oligo Duplex (Santa Cruz #sc-36229), using the siRNA transfection reagent (Santa Cruz #sc-29528) according to the manufacturer's instructions. Cells transfected with a nontargeting control siRNA duplex served as a negative control. Cells were used for experiments 72 hours after the transfection.

Quantitative PCR

Total RNA was extracted from the podocytes or mouse glomeruli using the RNeasy Plus Mini Kit (Qiagen, #74134). 1 μg of RNA was reverse-transcribed using Oligo dT primers and a high-capacity cDNA Synthesis Kit according to the manufacturer's instructions (Applied Biosystems, #4368814). Quantitative PCR amplifications were performed using the Power SYBR Green PCR Master Mix (Applied Biosystems, #4368577) and an Applied Biosystem QuantStudio 7 Pro instrument. The primer sequences for real-time quantitative PCR were obtained from PrimeBank of Harvard University (https://pga.mgh.harvard.edu/primerbank/) and synthesized by Integrated DNA Technologies. The specificity of the primers was validated by the corresponding melting curves. Relative quantification of dynein gene transcription was performed using GlycerAldehyde-3-Phosphate DeHydrogenase (GAPDH) as an endogenous control gene (constant expression across NG and HG conditions in cultured podocytes)³¹ and cells treated with NG as references. The Log2 FC of each gene was calculated for comparison. For the in vitro study, the relative quantification of mRNA was compared among cells treated with different concentrations of glucose (5.5, 11, or 30 mM) and the assay was repeated in three independent experiments. For in vivo study, the relative quantification was compared between streptozotocin (STZ)-induced diabetic mice and vehicle-only controls (n=6 in each group).

Co-Immunoprecipitation and Immunoblotting

Cultured podocytes were lysed on ice for 1 hour in Nonidet P40 lysis buffer with 0.5% deoxycholate, cOmplete, Mini Protease Inhibitor Cocktail (Roche, #11836153001), and PhosSTOP phosphatase inhibitor cocktail (Sigma, #4906845001). For co-immunoprecipitation (Co-IP), cell lysates were immunoprecipitated with mouse antinephrin conjugated agarose (Santa Cruz, sc-376522 AC, 2 µg for 400 μl cell lysate of 1×10⁶ cells), using normal mouse IgG-conjugated agarose as control (Santa Cruz Biotechnology, sc-2343). For immunoblotting, the immunoprecipitated proteins or proteins in cell lysates were separated by SDS-PAGE and transferred to polyvinylidene fluoride or polyvinylidene difluoride membranes. After blocking, the blot was incubated with the primary antibody at 4°C overnight. After washes, the blot was incubated with the appropriate peroxidase-labeled secondary antibody at room temperature for 30 minutes. Signals were detected using the Clean-Blot Immunoprecipitation Detection Reagent (HRP) kit (Thermo Fisher, #21230) and SuperSignal West Dura Extended Duration Substrate (Thermo Fisher, #34075). For quantification, densitometry was performed using ImageJ software [National Institutes of Health (NIH)].

Immunofluorescence Labeling of Cultured Podocytes

Podocytes grown on collagen I coated coverslips (Thermo fisher # A1142801) were fixed with 4% paraformaldehyde (Electron Microscopy Sciences, #30525-89-4) for 5 minutes, followed by standard labeling with primary and secondary antibodies as described in our previous study.²⁵ Cells were co-stained with fluorophore-conjugated phalloidin or ViaFluor reagent to show the colabeling of nephrin with actin or microtubule (Supplementary Figure 3).

Immunofluorescent Labeling of Kidney Sections

Paraffin-embedded kidney biopsy sections were obtained from the Pathology Department of the University of Iowa Hospitals and Clinics under an approved Institutional Review Board protocol. Based on an established protocol,²⁵ the sections were deparaffinized and rehydrated. Antigen retrieval was performed by incubating the slides in IHC-Tek Epitope Retrieving buffer (IHC world, #IW-1000) in an IHC-Tek Epitope Retrieval Steamer (IHC world, #IW1102) for 45 minutes, followed by standard immunofluorescent (IF) staining. Dynll1 and nephrin were colabelled by rabbit anti-Dynll1+Alexa Fluor 488-conjugated donkey anti-rabbit IgG and mouse antinephrin+Alexa Fluor 594-conjugated Donkey anti-mouse IgG. Fluorescence signals were captured from three nonsclerotic glomeruli of each section under an SP8 Leica confocal microscope. Dynll1-nephrin colocalization was mapped by using the colocalization threshold plugin of Fiji software and quantified with the Colo2 plugin as the Manders overlap coefficient.^32,33 These parameters were compared among DN (n=30 glomeruli in ten specimens), FSGS (n=26 glomeruli in ten specimens), minimal change disease (MCD, n=17 glomeruli in seven specimens), and normal kidney (n=9 glomeruli in three specimens). Mouse kidney sections were stained with fluorophore-conjugated phalloidin to show the actin cytoskeleton; and co-stained with WT1 to show podocyte-specific expression of Dynll1 (Supplementary Figure 4).

Nephrin Trafficking Model and Time-Lapse Imaging

As shown in the schematics in Figure 1A, cells transfected with the pcDNA3-nephrin plasmid (gift from Dr. Holzman⁵) were incubated with a mouse antinephrin antibody recognizing the extracellular domain (Santa Cruz #sc-376522), followed by an Alexa Fluor 594-conjugated donkey anti-mouse IgG (Thermo fisher #A32744) to trigger cross-linking and internalization of nephrin. As described in our previous study,²⁵ live cell imaging was performed by scanning cells every 5 seconds under a Leica SP8 Confocal microscope for a total of 20 minutes. Single molecule tracking of internalized nephrin with the coexpressed green fluorescent protein (GFP)-Dynll1 fusion protein (Sino Biological, pCMV3-C-GFPSpark-Dynll1, # MG51246-ACG), enhanced green fluorescent protein (EGFP)-DCTN1 fusion protein (Addgene pEGFP-p150Glued, #36154), or tdTomato-EB3 fusion protein (Addgene, EB3-tdTomato, #50708) was performed as 5 repeated measurements×5 cells×3 independent experiments, and the data analysis was performed using the Fiji plugin KymographClear coupled with KymographDirect. The fraction and the average velocity of the retrograde (red), anterograde (green), and static (blue) trafficking events were calculated and compared in cells with different treatments.

Figure 1.

Enhanced dynein-mediated postendocytic trafficking of nephrin in hyperglycemic conditions. (A) In vitro nephrin trafficking assay in mouse podocytes using antibody-mediated cross-linking of nephrin: Nephrin molecules expressed on the podocyte surface were cross-linked by an antinephrin primary antibody and Alexa Fluro 594–labeled secondary antibody, which triggered nephrin internalization and recruitment of trafficking adapters. As demonstrated by Co-IP ((B) by measuring dynein components in nephrin pulldown) and an immunofluorescence colocalization assay ((C) cross-linked nephrin labeled in red by Alexa Fluor 594, Dynll1, and DCTN1 labeled in green by Alexa Fluor 488), HG treatment (HG, 30 mM) increased the recruitment of Dynll1, DCTN1, and HDAC6 to cross-linked nephrin, compared with NG (5.5 mM). These changes could be diminished by inhibiting dynein activity (using Ciliobrevin D, 50 µM) (n=3, *P<0.05 versus HG+0.3% DMSO) or by knocking down Dynll1 in podocytes (n=3, ^P<0.05 versus HG+control siRNA).

Surface Biotinylation–Based Recycling Assay

As described in our previous work²⁵ and as outlined in schematics in Figure 2C, podocytes transfected with the pcDNA3-nephrin plasmid were incubated on ice in EZ link sulfo-NHS-SS-biotin (Pierce, #21331) for 45 minutes. Cells were then returned to 37°C and underwent internalization, followed by surface stripping by incubating with 100 mM Mesna (Santa Cruz, CAS 19767-45-4) at 4°C to remove biotin from uninternalized nephrin. Cells were returned to 37°C to allow for recycling of the internalized nephrin for 2 hours, followed by a second surface stripping with Mesna to remove biotin from the recycled nephrin. The biotinylated nephrin in cell lysates was pulled down by incubation with streptavidin-conjugated beads (Pierce, #20353) at 4°C overnight with continuous rotation, and the biotinylated nephrin was measured by immunoblotting using antinephrin. The biotinylated nephrin in specimen 4 and specimen 5 reflect initially internalized nephrin and residual nonrecycled nephrin, respectively. The percentages of recycled nephrin were determined with the following formula: % recycled=[(initially internalized nephrin)−(nonrecycled nephrin)]/(initially internalized nephrin)]×100%. P<0.05 versus HG, n=3.

Figure 2.

Dynein-mediated nephrin degradation and recycling in HG treatment podocytes. (A) Tracking of nephrin and the coexpressed GFP-Rab7 or GFP-Rab11. Endocytosed nephrin proteins in HG-treated cells were enriched in Rab7-positive lysosomes with reduced Rab11 coating compared with NG-treated cells. These changes were much less pronounced in HG-treated cells in which dynein activity was inhibited with Ciliobrevin D (versus HG+0.3% DMSO) or siRNA-mediated Dynll1 knockdown (versus cells transfected with control siRNA). (B) Cells were pretreated with cycloheximide (to reduce the background of newly synthesized nephrin) and underwent nephrin cross-linking and internalization. The extent of postendocytic degradation of nephrin in these cells was reflected by the reduction in nephrin protein levels as measured by Western blotting. Nephrin cross-linking in HG-treated cells resulted in reduced nephrin, which was rescued by incubation with Ciliobrevin D (50 µM), Leupeptin (200 µg/ml), or bortezomib (100 nm). 0.3% DMSO was added to control cells without exposure to specific chemical compounds. Compared with cells transfected with control siRNA, Dynll1 knockdown reduced the cross-linking–induced depletion of nephrin protein in HG-grown cells. n=3, *P<0.05 versus HG+DMSO after nephrin cross-linking. ^P<0.05 versus HG+Control siRNA after nephrin cross-linking. (C) Nephrin recycling was examined by using a surface biotinylation-based recycling assay as outlined in the schematics. Cells went through the following steps: (1) biotinylation of all surface nephrin; (2) antibody-mediated internalization of nephrin; (3) first surface stripping with Mesna to remove biotin from uninternalized nephrin, i.e., to leave only the internalized nephrin biotinylated; (4) recycling of the biotinylated nephrin; and (5) second surface stripping with Mesna to remove biotin from the recycled nephrin. Cell lysates collected at steps (4) and (5) were subjected to streptavidin IP to pull-down biotinylated nephrin. The biotinylated nephrin measured in specimen 4 and specimen 5 reflects initially internalized nephrin and residual nonrecycled nephrin, respectively. The percentages of recycled nephrin were determined with the following formula: % recycled=[(initially internalized nephrin)−(nonrecycled nephrin)]/(initially internalized nephrin)]×100%. n=3, *P<0.05 versus HG+DMSO; ^P<0.05 versus HG+control siRNA. IP, immunoprecipitation.

STZ-Induced DN in C57BL/6J Mice

As described in a published protocol,^34,35 8-week-old, fasted C57BL/6J wild-type mice (male:female ratio=1:1) received intraperitoneal (i.p.) injection of STZ (MedChemExpress LLC, # HY-13753) dissolved in 0.1  M sodium citrate buffer (pH=4.5), at the dose of 50  mg/kg body weight, daily for 5 consecutive days. Age- and sex-matched vehicle control mice received i.p. injections of sodium citrate buffer of the same volume and regimen. Whole blood glucose was measured in tail vein blood specimens using the Germaine Laboratories AimStrip Plus Blood Glucose Testing System. Urine albumin and creatinine were measured in specimens collected before and at different months after the injections, using the Albumin ELISA Quantitation kit (Bethyl Laboratories Inc.) and a colorimetric creatinine quantification kit (Bioassay Systems). The urine albumin-to-creatinine ratio was calculated for comparison.

Kidney Specimen Handling, Histology, and Quantification

Mice were anesthetized by isoflurane inhalation and then perfused with EZ link sulfo-NHS-SS-biotin per an in vivo surface biotinylation protocol before sacrifice.³⁶ The mouse kidneys were then harvested for light microscopy, transmission electron microscopy (TEM), immunogold labeling (IGL), and for molecular biology studies at the Comparative Pathology Laboratory and the Central Microscopy Research Facility at the University of Iowa. The glomerular hypertrophy was analyzed by quantifying the area of each individual glomerulus. Podocytopathy was quantified by measuring the percentages of capillary loops covered by injured podocytes with effaced foot processes or disassembled SDs. For IGL,²⁵ 8 nm sections were placed on carbon-coated and glow-discharged, formvar-coated nickel slot grids; blocked grids were incubated in a mouse antipodocin antibody (Sigma Aldrich, #SAB4200810) and a rabbit antinephrin antibody (Invitrogen, #PA5-91907) at room temperature, followed by goat anti-rabbit Au 12 nm conjugated (Jackson ImmunoResearch Laboratories Inc. #111-205-144) and goat anti-mouse Au 6 nm conjugated (Jackson ImmunoResearch Laboratories Inc. #115-195-146). Mouse glomeruli were isolated based on the standard protocol as described in reference³⁷ and then lysed in RLT buffer (RNeasy Plus Universal Mini Kit, Qiagen, #74034) or Nonidet P40 lysis buffer for mRNA or protein analysis, respectively.

Statistical Analyses

Data analyses were performed using SPSS version 10.0. IF signal analysis and densitometric analysis of immunoblots were performed using NIH Fiji/ImageJ software. Data were expressed as mean±SEM. An independent sample t test was used to compare the difference between two groups. One-way ANOVA was used for comparisons among multiple groups, and a post hoc q test was used to compare the difference between groups. In a two-tailed test, P<0.05 was considered significant.

Study Approval

Paraffin-embedded kidney biopsy sections were obtained from the Pathology Department of the University of Iowa, and the University of Iowa Institutional Review Boards approved the secondary use of archived kidney biopsy specimens. The animal protocol was approved by the Institutional Animal Care and Use Committee of the University of Iowa and is in accordance with the NIH guidelines for use of live animals. All work was performed in accordance with the principles and procedures outlined in the NIH guidelines.

Results

Dynein Expression Is Upregulated in DN and Podocytes with Hyperglycemic Injury

We reanalyzed existing transcriptome datasets of DN in human and mouse models of diabetes. In the Hodgin Diabetes Mouse Glom dataset,²⁷ we found a cluster of dynein genes with significantly increased expression in diabetic mice with fasting glucose >300 mg/dl, compared with those with fasting glucose <300 mg/dl (P<0.05, Figure 3A), including genes that encode dynein heavy chain (DYNC1H1), intermediate chain (DYNC1I1), and dynactin 1 (DCTN1). In db/db mice, the kidney expression of DYNC1LI1 (r²=0.856) and DYNC1LI2 (r²=0.870) correlated with their fasting glucose levels. In all diabetic mice that were analyzed, the kidney expressions of DCTN1 (r²=0.355), DYNC1H1 (r²=0.676), DYNC2LI1 (r²=0.401), and DYNC1I1 (r²=0.549) correlated positively with fasting glucose levels (Supplementary Figure 1B). In Endothelial Nitric Oxide Synthase (eNOS)-/- db/db mice and STZ-induced DN models, the expression of DYNC1LI2 (r²=0.612), DYNC1I1 (r²=0.898), and DYNLT3 (Dynein Light Chain Tctex-Type 3) (r2=0.599) correlated with the severity of albuminuria (Supplementary Figure 1A).

Figure 3.

Increased dynein expression in DN. (A) Transcription analyses of Hodgin Diabetes Mouse Glom dataset in Nephroseq revealed upregulated dynein expression in diabetic mice with significant fasting hyperglycemia. Heatmap shows a cluster of dynein genes with increased expression in glomeruli of diabetic mice with fasting glucose>300 mg/dl. The Log2 median–centered expression level of each individual dynein gene is color-coded, reflecting Z scores normalized to depict relative value. (B) Transcription analyses in Woroniecka Human Diabetic Kidney Disease dataset revealed upregulated dynein expression in human DN versus normal controls. Heatmap shows a cluster of dynein genes with increased expression in human DN. (C and D) Quantification of mRNA levels of dynein components in podocytes growing in media containing different glucose concentrations (C: HG, 30 mM; PDB, 11 mM; and NG, 5.5 mM; n=3 independent experiments; *P<0.05 versus NG) or in glomeruli isolated from STZ-induced diabetic mice (D) were measured by relative quantitative PCR (normalized to GAPDH) (D: n=6 mice each group; *P<0.05 versus vehicle control).

Coordinate expression of dynein genes was also found in the Woroniecka Human Diabetic Kidney Disease dataset²⁸ (Figure 3B), including the significantly upregulated DYNC1H1, DCTN1, and DYNLL1 (encoding dynein light chain 1). The expression level of DYNC1I1 (r²=0.535) and DYNC2LI1 (r²=0.654) correlated inversely with GFR in human DN. In the ERCB Chronic Kidney Disease dataset, expression levels of Dynll1 (r²=0.919) and DCTN1 (r²=0.649) correlated with the decline of GFR in human DN (Supplementary Figure 1C).

We next verified the mRNA levels of dynein components in cultured podocytes growing in media containing different glucose concentrations and in glomeruli isolated from the STZ-induced diabetic mice. Compared with cells growing in NG (glucose 5.5 mM with 24.5 mM of mannitol to maintain the same osmolality), cells growing in PDB (glucose 11 mM with 19 mM of mannitol) or HG (glucose 30 Mm alone) showed increased expression of dynein components (Figure 3C). Among them, Dynll1 and DCTN1 had the most elevated expression in response to HG exposure, suggesting they are potential direct responders to hyperglycemia. Expression of the dynein components was increased and peaked in glomeruli isolated from diabetic mice 3 months after STZ injection (Figure 3D), when the mice have developed significant hyperglycemia >300 mg/dl (Figure 4D). The expression of INF2, which encodes a sequestrator for Dynll1, was only slightly increased in HG-treated cells and in the early stage of STZ-induced DN (3 months) but was far less prominent than that of Dynll1. The expression of NPHS1, the gene encoding nephrin, was not reduced in diabetic podocytopathy in vitro or in vivo; instead, its expression increased in the early stage of STZ-induced DN.

Figure 4.

Podocytopathy in STZ-induced diabetic mice. The histological and ultrastructural features of diabetic glomerulopathy and podocytopathy developed in mice 3 months after receiving i.p. injection of STZ, as shown by periodic acid–Schiff staining and Masson trichrome staining (A: scale bars=50 μm) and TEM (B: ×12,000, scale bars=1 μm; ×30,000, scale bars=200 nm). In mice with STZ-mediated DN, podocytopathy presented with effacement (black arrows) of foot processes or cortical dislocation of SDs (white arrows). Podocytes with STZ-induced diabetic injury are characterized by the dislocation of nephrin and podocin from the SD to the cytosol (arrows) as detected by IGL (C: ×40,000, scale bars=200 nm). Hyperglycemia (D) and albuminuria (E), quantified as the albumin/creatinine ratio, in mice with STZ-mediated DN (*P<0.05 versus preinjection, n=6). Histology and ultrastructural features of diabetic glomerulopathy were expressed as the average glomerular area (F: n=6 mice×5 glomeruli per mouse) and the percentage of glomerular capillaries covered by effaced foot processes (G: n=6). *P<0.05 versus vehicle control.

Dynein-Mediated Postendocytic Sorting of Nephrin Is Enhanced in High Glucose–Induced Podocyte Injury

We found that HG treatment of podocytes increased the dynein-dependent trafficking of nephrin to the lysosome. Here, we used an in vitro model of antibody-triggered nephrin internalization in cultured podocytes²⁵ (Figure 1A). The recruitment of dynein components to internalized nephrin was measured by a Co-IP assay (Figure 1B) and visualized with an immunofluorescence colocalization assay (Figure 1C). Growing podocytes under HG conditions (HG, 30 mM) resulted in higher levels of Dynll1 and DCTN1 recruitment to the internalized nephrin than observed in podocytes grown under NG concentrations (NG, 5.5 mM). In addition, we found increased recruitment of histone deacetylase 6 (HDAC6) to internalized nephrin in HG-treated cells. HDAC6 is the protein that bridges the dynein transport complex with ubiquitinated protein that is designated for degradation.^38,39 The recruitment of both dynein components and HDAC6 was diminished by antagonizing dynein with the addition of Ciliobrevin D (50 µM, a direct dynein ATPase inhibitor)⁴⁰ to the media or by siRNA-mediated knockdown of Dynll1, a component essential for the integrity of the dynein transport complex⁴¹ (cells transfected with Dynll1 siRNA versus control siRNA). These data suggest HDAC6 is recruited using dynein, consistent with previous studies showing the association of Dynein with HDAC6.⁴²

Dynein Enhanced the Retrograde Trafficking of Nephrin in High Glucose–Treated Podocytes

We imaged live cells to track internalized nephrin with coexpressed GFP-Dynll1, EGFP-DCTN1, or tdTomato end-binding protein (EB3). As shown in Figure 5A, kymograph analysis demonstrates that the internalized nephrin shared the same trajectory with tagged Dynll1 and DCTN1, and the nephrin moved in the opposite direction along the microtubule to tdTomato-EB3, a marker for the plus ends of microtubules.⁴³ These findings indicate a dynein-mediated retrograde trafficking of endocytosed nephrin away from the plus ends. In cells with Dynll1 knockdown, we found that the internalized nephrin is disassociated from the coexpressed DCTN1, nephrin seems to be more static at the peripheral membrane with significantly abrogated retrograde trafficking.

Figure 5.

Live cell imaging of dynein-mediated trafficking. (A) Trajectory analysis of nephrin with coexpressed GFP-Dynll1, GFP-DCTN1, or tdTomato-EB3 by using live cell imaging and kymograph analysis. Nephrin and dynein components (Dynll1 or DCTN1) shared the same trajectory of retrograde trafficking to the minus end of microtubules. The track of nephrin was parallel to but moving in the opposite direction to that of EB3, which labels the plus ends of microtubules. (B)The postendocytic trafficking of nephrin was compared in podocytes growing in NG (+0.3% DMSO), HG (+0.3% DMSO), or HG in the presence of Ciliobrevin D (50 µM). The tracks of nephrin particles were illustrated by using TrackMate of Fiji software, the warm color reflecting a higher velocity and the cooler color a lower velocity. The fractions (100% stacked bar chart, upper panel) and the average velocities (lower panel graph) of anterograde (green, from cytosol to surface membrane), retrograde (red, from surface membrane to cytosol), and static (blue) trafficking events of nephrin in five repeated measurements×5 cells×3 independent experiments were collected and quantified by using KymographClear coupled with KymographDirect software. *P<0.05 versus HG+DMSO; ^P<0.05 versus HG+Control siRNA. (C) Cotracking of nephrin with GFP-DCTN1 in cells with siRNA-mediated Dynll1 knockdown showed disassociated DCTN1 (green) and nephrin (red) trajectories.

We then compared the nephrin trafficking parameters in podocytes growing in NG or HG. Compared with cells growing in NG, cells growing HG showed significantly more and faster retrograde events (from surface membrane to cytosol), as reflected by increased fractions and velocities of retrograde events and reduced anterograde events, i.e., recycling from cytosol to surface membrane. These differences were suppressed by incubating cells with Ciliobrevin D, or with siRNA-mediated Dynll1 knockdown, indicating an enhanced dynein-dependent postendocytic trafficking in our in vitro model of diabetic podocytopathy, where Dynll1 plays a key role (Figure 5B).

Dynein-Mediated Trafficking Alters Nephrin Degradation and Recycling in High Glucose–Treated Podocytes

We observed increased sorting of nephrin to the lysosomes, which is marked by Rab7 in HG-treated podocytes (Figure 2A). This is consistent with increased dynein recruitment that would be expected to direct endosomes toward lysosomes for degradation. To test this idea, we followed the degradation of internalized nephrin in HG-treated podocytes using a cycloheximide chase to inhibit synthesis of new nephrin. Over this period, there was a dramatic loss of nephrin in HG-treated podocytes after cross-linking–induced nephrin internalization, compared with NG-grown podocytes. The stabilization of nephrin in cells treated with lysosome inhibitor Leupeptin suggests a lysosome-mediated nephrin degradation in HG-treated cells. Finally, the depletion of nephrin was also blocked by Ciliobrevin D, or by Dynll1 knockdown, suggesting that HG-induced degradation of nephrin was dynein-dependent, where Dynll1 plays a key role (Figure 2B).

Next, we tested whether the enhanced dynein-mediated trafficking and degradation seen in HG-treated cells is associated with decreased recycling of nephrin. As shown in Figure 2A, by tracking internalized nephrin with coexpressed GFP-Rab11, a marker for recycling endosomes, we found a reduced degree of nephrin recycling in cells with exposure to HG. Furthermore, using a surface biotinylation-based recycling assay (described in Figure 2C), we showed that internalized nephrin primarily recycled back to the cell surface in podocytes growing in NG but that the recycling was greatly diminished under hyperglycemic conditions. Nephrin recycling in the HG-treated cells was restored with the addition of Ciliobrevin D or knockdown of Dynll1, supporting the essential role for dynein, especially Dynll1, in mediating the altered trafficking of nephrin in podocytes under hyperglycemic stress.

Dynein-Mediated Mistrafficking of Nephrin Related to Its Homeostasis In Vivo

We used the STZ-induced diabetic mice as an in vivo model to determine whether dynein-mediated triage and homeostasis of nephrin observed in HG-treated podocytes in vitro correlate with the onset of nephropathy. STZ-induced diabetic mice developed significant hyperglycemia after 2 months (Figure 4D) and albuminuria after 3 months (Figure 4E). By month 3, these diabetic mice developed histologic features of DN, including mesangial expansion and glomerular hypertrophy (reflected by increased size of glomeruli, Figure 4F). The TEM demonstrated features of podocytopathy reflected by effacement of foot process and cortical relocation of SDs (Figure 4B), as well as an increased percentage of capillaries covered by abnormal podocytes (Figure 4G). By IGL, we showed the “trapping” of nephrin and podocin in the cytosol of podocytes in the STZ-induced diabetic mice, instead of localizing to the sites of SDs, as seen in the control mice (Figure 4C).

Immunofluorescence labeling of kidney sections from STZ-induced diabetic mice revealed increased expression of Dynll1 along the glomerular capillaries that colocalized with nephrin (Figure 6A). As depicted by Manders overlap coefficients, the colocalization of Dynll1 with nephrin was significantly higher in podocytes of diabetic mice. This was corroborated biochemically with Co-IP experiments showing increased recruitment of DCTN1 and Dynll1 to nephrin from glomeruli isolated from STZ-induced diabetic mice (Figure 6B), as we had seen in cultured podocytes (Figure 1B). This corresponded to the reduced amount of surface nephrin—as measured in streptavidin pulldown—after the in vivo surface biotinylation assay.³⁶ We also observed reduced nephrin protein but not mRNA levels (Figure 3D) as well as nephrin ubiquitination in diabetic kidneys (Figure 6B), suggesting a dynein-mediated trafficking and ubiquitin-mediated degradation of nephrin in lysosomes, resulting in a rapid turnover of the protein that increases the podocytes' susceptibility to diabetic injury.

Figure 6.

In vivo evidence of dynein-related mistrafficking and degradation of nephrin. (A) IF staining showed that overexpressed Dynll1 colocalizes with nephrin in STZ-induced mouse diabetic podocytopathy. The amount of Dynll1 colocalized with nephrin in nonsclerotic glomeruli was quantified using the Colo2 plugin of Fiji software using the Manders overlap coefficient for comparison. *P<0.05 versus vehicle control, n=6 mice×3 glomeruli per mouse. (B) The recruitments of Dynll1 and DCTN1 were examined in nephrin pulldown and were expressed as the ratio to total nephrin, reflecting the involvement of dynein in nephrin trafficking in vivo. Nephrin protein targeted for ubiquitin-mediated degradation was examined by measuring ubiquitinated nephrin (IP with antinephrin and IB with anti-ubiquitin). Nephrin localized on the podocyte surface was measured by streptavidin IP after the in vivo surface biotinylation assay, expressed as the percentage of total nephrin for comparison. *P<0.05 versus vehicle control, n=6. IP, immunoprecipitation.

Evidence of Dynein-Mediated Trafficking of Nephrin in Human DN

Upregulated Dynll1 expression colocalizing with nephrin was recently recognized as a feature of enhanced dynein-mediated trafficking of nephrin in vivo in a form of FSGS caused by pathogenic mutations in INF2.²⁵ In podocytes harboring INF2-R218Q mutation, the R218 mutant of INF2 was found to dissociate Dynll1, release it for nephrin recruitment, and subsequently facilitate the trafficking of nephrin to the lysosomal degradation pathway. Thus, the nephrin-Dynll1 colocalization is used as an in vivo biomarker for dynein-mediated trafficking of nephrin.

Our immunofluorescence labeling in kidney biopsy sections obtained from patients with DN revealed an enriched expression of Dynll1 along the glomerular capillaries that colocalized with nephrin (Figure 7). Such a pattern was also observed in patients with FSGS but is absent from the normal kidney or in MCD, a reversible podocytopathy with a benign outcome.

Figure 7.

In situ biomarker for dynein-mediated trafficking of nephrin in human glomerulopathies. Colocalization of Dynll1 and nephrin under conditions of increased Dynll1 expression was examined by co-IF staining and was compared in progressive podocytopathies (FSGS and DN) versus normal kidney and benign podocytopathy (MCD). Scale bar=20 μm. The amount of Dynll1 that colocalized with nephrin in nonsclerotic glomeruli was quantified using the Colo2 plugin of Fiji software as Manders coefficients, and these parameters were compared among DN (n=30 glomeruli in ten specimens), FSGS (n=26 glomeruli in ten specimens), MCD (n=17 glomeruli in seven specimens), and normal kidney (n=9 glomeruli in three specimens). *P<0.05 versus normal kidney.

We further tested whether the significant Dynll1-nephrin colocalization in human DN is related to an altered sequestration of INF2. Compared with normal control kidney specimens, there is a significant colocalization of Dynll1 with INF2 along the edges of podocytes in patients with DN (Figure 8). Although the expression of INF2 was increased in DN, this level of upregulation is far less dramatic than that of Dynll1 (Figure 3, C and D), suggesting inadequate sequestration of Dynll1 by INF2 in diabetic podocytopathy.

Figure 8.

Dynll1 protein expression is increased and colocalizes with INF2 in human DN.

Discussion

DN is a nearly ubiquitous end organ complication of diabetes, reflective of the fact that chronic hyperglycemia is a major cause of podocyte injury. This injury results in increased GFB permeability evidenced by the development of albuminuria, followed by the progression to glomerulosclerosis^8,44–46. The ability to identify the earliest pathophysiological events of DN and intervene in these before the development of diabetic kidney injury holds the greatest promise for improving disease outcome. Increased nephrin endocytosis has been described in diabetic podocyte injury,^14,15 and loss of the precise cell surface membrane targeting of nephrin results in podocyte foot process effacement and SD disassembly; however, the current understanding of nephrin mistrafficking is limited to the initial step of internalization immediately from the surface membrane. Because the default pathway for internalized proteins is their recycling back to the plasma membrane,⁴⁷ the postinternalization steps that would ultimately control nephrin degradation and loss from the plasma membrane in diabetic podocytopathies remains to be determined. Our study reveals a new mechanism that the turnover of nephrin in DN is mediated by dynein-driven, microtubule-dependent trafficking that diverts internalized nephrin toward lysosomal degradation and thus opens a door to various novel therapeutic approaches.

In this study, we started by analyzing multiple DN transcriptome datasets collected in Nephroseq kidney transcriptome databases. We discovered elevated expression of genes encoding dynein components that correlated positively with hyperglycemia and the severity of DN, providing in vivo evidence to support a dynein-related pathogenesis of DN. In vitro studies found podocytes grown in hyperglycemic conditions recapitulate a number of signaling mechanisms leading to podocyte injury in in vivo models of diabetes, including evoking 5′ adenosine monophosphate-activated protein kinase (AMPK)-related insulin resistance,⁴⁸ reactive oxygen species,⁴⁹ and rapamycin-induced endoplasmic reticulum stress.⁵⁰ We verified that expression of dynein components is elevated in podocytes growing in hyperglycemic conditions and in an STZ-induced type I diabetic mouse model; these results support the role for dynein in mediating the early pathophysiology of DN induced by hyperglycemia. Our data highlight DynII1 and DCTN1 as key responders to hyperglycemia, which can be explained by the fact that their promoters harbor binding motifs for hyperglycemia-responsive transcription factors such as specificity proteins and Krüppel-like factors,⁵¹ which will be further investigated in our future studies. DCTN1 encodes dynactin 1, a key activator for dynein-mediated retrograde trafficking²⁰; Dynll1 encodes the light chains that maintain the integrity of the entire dynein complex and its activity, as evidenced by the observation that dynein-mediated processes can be limited by molecular traps for Dynll1 protein.^21,41 The overall activity of the dynein complex can be inhibited by regulating the availability of Dynll1.^21,41 For instance, INF2, a Dynll1-binding partner, is highly expressed in podocytes and functions as a sequestrator for Dynll1 to inhibit dynein-mediated processes. In the case of inherited FSGS, mutations in INF2 (e.g., R218Q) that disrupt its association with Dynll1 essentially activate dynein and potentiate dynein-dependent processes.²⁵ In the case of DN, activation of dynein is accomplished at least in part by transcription. We observed colocalization of Dynll1 with INF2 in podocytes of diabetic patients, suggesting that while INF2 works in sequestrating Dynll1, its levels are likely insufficient to compensate for the increased production of dynein under diabetic conditions.

To test whether the hyperglycemia-induced dynein overexpression results in increased dynein-mediated trafficking of nephrin, we observed increased recruitment DCTN1 and Dynll1 to the internalized nephrin in HG-treated podocytes, suggesting an enhanced involvement of dynein in the postendocytic sorting of nephrin. This mechanism was also evidenced in vivo by the observed enrichment of Dynll1 and DCTN1 in nephrin pulldowns in STZ-induced mouse DN and reflected by increased expression of Dynll1 that colocalized with nephrin in human DN. Furthermore, the recruitment of HDAC6 to internalized nephrin in HG-treated cells indicates the sorting of these nephrin to lysosomal degradation pathways because HDAC6 functions as an adapter that connects dynein with ubiquitinated cargo proteins and diverts them to lysosomes.^38,39 This was supported by the increased flux of internalized nephrin to Rab7-positive lysosomes in HG-treated cells and further validated by the finding of increased nephrin degradation that is salvaged by inhibiting lysosomal proteases using Leupeptin. Meanwhile, under hyperglycemic conditions, we found decreased sorting of nephrin to Rab11-positive recycling endosomes and impaired surface recycling of nephrin, which proved to be a dynein-dependent change. The fact that these changes could be rescued by inhibiting dynein ATPase with Ciliobrevin D or by knocking down Dynll1 demonstrated the key role of dynein, especially Dynll1, in hyperglycemia-induced mistrafficking of nephrin. These findings support the hypothesis that diabetic podocytopathy occurs when hyperglycemia activates the expression of dynein, which in turn facilitates the depletion of nephrin by triaging it away from recycling and into degradation pathways through adapters such as HDAC6.³⁸ Because nephrin is an integral membrane protein, these findings suggest that dynein-mediated postendocytic sorting of nephrin results in its degradation, primarily by lysosomes by sorting through multivesicular bodies.

In STZ-induced DN, we observed increased ubiquitinated nephrin and reduced nephrin protein but preserved transcription of NPHS1 mRNA, suggesting an increased nephrin degradation through ubiquitin-mediated sorting to the lysosome. These changes also seem to correlate with a change in the overall proteolytic capacity of podocytes as observed in the work by Meyer-Schwesinger et al. that revealed an increased ubiquitin proteolysis system in DN and FSGS, but not in to MCD, which is a benign and reversible podocytopathy.⁵² Thus, forward transport of nephrin to the lysosome and shunting it away from recycling in a dynein-dependent manner is likely a process that also involves increased ubiquitination of nephrin. This not only implies that the dynein-associated components might be valuable therapeutic targets but also that a specific ubiquitin ligase that could target nephrin under DN conditions might also qualify as a therapeutic target.

Taken together, our work revealed a new mechanism for dynein-mediated diabetic podocytopathy, in which dynein components key to its activity are overexpressed and promote the postendocytic sorting of nephrin from functional recycling to be degraded. Our finding not only fills the long-existing knowledge gap between the initial internalization and the final loss of nephrin witnessed during the progression of DN, by focusing a newly identified dynein, a microtubule motor protein complex that has unique communications with proteolysis systems. Furthermore, it sparks new therapeutic opportunities for human DN including modifiers targeting specific proteins or relevant cellular events, including dynein activity, ubiquitin ligases or deubiquitinating enzymes, and HDAC6 activity, as well as therapeutics for the modulation of microtubule polymerization and the signal transduction pathway(s) that connects hyperglycemia with dynein expression/activation.

Disclosures

R.C. Piper reports the following: Research Funding: Supported by R01 GM58202 (to R.C.P. PI). C. Nester reports the following: Consultancy: Advisory Board—Alexion, Apellis, Biocryst, and Novartis; Research Funding: Achillion—Site PI–C3G Trial; Alexion; Avacopan - Site PI - C3G Trial; Apellis—Site PI–C3G Trial; Novartis—Site PI–C3G Trial; Retrophin–Pediatric Recruiting Site for the Duet Trial; and Honoraria: Advisory Board—Apellis, Alexion, Biocryst, and Novartis. All remaining authors have nothing to disclose.

Funding

This work was supported by the University of Iowa Child Health Research Career Development Award (PI: Alexander Bassuk, NIH5K12HD027748-30) and University of Iowa Stead Family Children's Hospital Children's Miracle Network grant 2022 (PI: H.S., CMN86023539).

Author Contributions

C. Allamargot was responsible for the confocal microscopy and contributed to the design, protocol development, and quantitative imaging data analysis; J. Misurac contributed to the writing and critical editing of the manuscript; C. Nester contributed to the design of the human pathology study and result interpretation, as well as the writing of the manuscript; R.C. Piper contributed to the research design, data review, and interpretation, as well as writing of the manuscript; H. Sun and J. Weidner were responsible for the design and execution of experiments, statistical analysis, and drafting of the manuscript; H. Sun was responsible for the organization, correspondence and revision of the manuscript; and all authors approved the final version of the manuscript.

Data Sharing Statement

The data used to support the findings of this study are available from the corresponding author upon request.

Supplemental Materials

This article contains the following supplemental material online at http://links.lww.com/KN9/A249.

Supplementary Figure 1. Correlation of dynein expression with diabetic nephropathy phenotypes.

Supplementary Figure 2. Dynein genes transcription analysis in Nakagawa CKD Kidney dataset and their correlations with disease phenotype.

Supplementary Figure 3. HG-induced cytoskeletal organization and dynamic changes.

Supplementary Figure 4. Immunofluorescent staining of F-actin, acetylated α-microtubules and WT1 (costained with Dynll1) in mice glomeruli with STZ-induced diabetic nephropathy.