Suppressed ER‐associated degradation by intraglomerular cross talk between mesangial cells and podocytes causes podocyte injury in diabetic kidney disease

Abstract

Mesangial lesions and podocyte injury are essential manifestations of the progression of diabetic kidney disease (DKD). Although cross‐communication between mesangial cells (MCs) and podocytes has recently been suggested by the results of single‐nucleus RNA sequencing analyses, the molecular mechanisms and role in disease progression remain elusive. Our cDNA microarray data of diabetic mouse glomeruli suggested the involvement of endoplasmic reticulum (ER) stress in DKD pathophysiology. In vitro experiments revealed the suppression of the ER‐associated degradation (ERAD) pathway and induction of apoptosis in podocytes that were stimulated with the supernatant of MCs cultured in high glucose conditions. In diabetic mice, ERAD inhibition resulted in exacerbated albuminuria, increased apoptosis in podocytes, and reduced nephrin expression associated with the downregulation of ERAD‐related biomolecules. Flow cytometry analysis of podocytes isolated from MafB (a transcription factor known to be expressed in macrophages and podocytes)‐GFP knock‐in mice revealed that ERAD inhibition resulted in decreased nephrin phosphorylation. These findings suggest that an intraglomerular cross talk between MCs and podocytes can inhibit physiological ERAD processes and suppress the phosphorylation of nephrin in podocytes, which thereby lead to podocyte injury under diabetic conditions. Therapeutic intervention of the ERAD pathway through the cross talk between these cells is potentially a novel strategy for DKD.

Abbreviations

ATF6

activating transcription factor 6

Bax

Bcl2‐associated X protein

Bcl‐2

B‐cell lymphoma 2

BSA

bovine serum albumin

CHOP

C/EBP homologous protein

DAPI

4′,6‐diamidino‐2‐phenylindole

DKD

diabetic kidney disease

EC

endothelial cell

EerI

Eeyarestatin I

ER stress

endoplasmic reticulum stress

ERAD

endoplasmic reticulum‐associated degradation

ESRD

end‐stage renal disease

FITC

fluorescein isothiocyanate

HRP

horseradish peroxidase

IRE1α

inositol requiring enzyme‐1

Kif

kifunensine

MafB

V‐maf musculoaponeurotic fibrosarcoma oncogene homolog B

MC

mesangial cell

MPC5

mouse podocyte clone 5

PDGF‐B

platelet‐derived growth factor‐B

PVDF

poly‐vinylidene difluoride

SDS

sodium dodecyl sulfate

STZ

streptozotocin

sup

supernatant

TNF‐α

tumor necrosis factor‐α

TUNEL

TdT‐mediated dUTP Nick End Labeling

UPR

unfolded protein response

VEGF

vascular endothelial growth factor

WT‐1

Wilms tumor 1

XBP1

X‐box binding protein 1

1 INTRODUCTION

Diabetic kidney disease (DKD) is the leading cause of chronic kidney disease (CKD) and end‐stage renal disease (ESRD), which leads to a high risk of mortality and is a major healthcare burden worldwide. Despite advancements in managing glycemic control, hypertension, dyslipidemia, and related metabolic abnormalities, the number of patients that progress to ESRD due to DKD remains high. The clear pathophysiology and mechanisms of this disease progression are yet to be fully elucidated, and very few effective therapies are available for preventing or reversing the progression of DKD.

One of the main clinical features of DKD is albuminuria, and podocyte injury is inevitable upon the appearance of macroalbuminuria.¹ Mesangial lesions such as diffuse mesangial expansion or nodular lesions, also known as Kimmelstiel‐Wilson nodules, are also a pathological hallmark of the progression of DKD. In recent years, the cross‐communication between mesangial cells (MCs) and podocytes has been suggested in studies wherein ligand‐receptor intercellular signaling pathways were analyzed by evaluating the samples from diabetic kidneys of humans through single‐nucleus RNA sequencing.² However, the detailed molecular mechanisms and the significance of DKD progression remain elusive. In previous studies, the roles of cellular cross talk among other glomerular component cells, including those of platelet‐derived growth factor‐B (PDGF‐B) between MCs and endothelial cells (ECs) or vascular endothelial growth factor (VEGF) between podocytes and ECs, are well‐established^{3, 4}; however, the molecular cross talk between MCs and podocytes has been only scarcely investigated.⁵ Only a few studies have reported MC‐podocyte‐tubular cross talk in IgA nephropathy.⁶

Recently, endoplasmic reticulum (ER) stress has been shown to play important roles in the progression of various diseases, including glomerular disorders.^7-9 We previously reported that ER stress induced by glucolipotoxicity and homeostatic inflammation through intraglomerular cross talk may play a role in the progression of DKD.^{10, 11} Moreover, several recent studies have focused on the role of ER‐associated protein degradation (ERAD) processes, which is one of the three major branches of the ER stress response in podocyte dysfunction.^{12, 13} ERAD is a protein quality control mechanism against the accumulation of improperly folded proteins in the ER; therefore, dysfunction of the ERAD may lead to cellular damage. Recent studies have suggested the importance of this pathway in maintaining cellular homeostasis in podocytes.¹⁴ Thus, we hypothesized that the cross‐communication between MCs and podocytes may affect the ER stress responses of these cells and thereby participate in the progression of DKD, especially in podocyte injury. In the present study, we aimed to clarify the role of this cellular cross talk in podocyte injury under diabetic conditions.

2 MATERIALS AND METHODS

2.1 Cell culture

Immortalized mouse podocyte cell line (MPC5) was kindly provided as a gift from Prof. Katsuhiko Asanuma (Department of Nephrology, Chiba University Graduate School of Medicine, Chiba, Japan). To propagate cells, podocytes were incubated at 33°C in RPMI 1640 medium (25 mmol/L glucose) supplemented with 10% fetal bovine serum (FBS) and 10 units interferon‐γ (Sigma‐Aldrich, St. Louis, MO, USA) per 1 mL medium. For cell differentiation, the cells were incubated at 37°C without interferon‐γ for 10 to 14 days. Rat MCs were kindly provided by Dr Daisuke Nakano and Prof. Akira Nishiyama (Department of Pharmacology, Kagawa University Medical School, Japan). The cells were isolated from male Sprague Dawley rats and maintained according to previous publications.^{15, 16} The MCs were cultured in Dulbecco's Modified Eagle Medium/Nutrient Mixture F‐12 (17.5 mmol/L glucose), and the medium was replaced with RPMI 1640 (high glucose (HG): 25 mmol/L or normal glucose (NG): 5.6 mmol/L) and incubated for 24 to 48 hours before podocyte stimulation. To equalize osmolarity, we added D‐mannitol (Nacalai Tesque, Kyoto, Japan) to the NG group. Subsequently, the medium was replaced, and the supernatant of MCs (MC‐sup) was incubated for 24 to 48 hours, collected and centrifuged at 1600 g for 5 min, and added directly onto podocytes for stimulation. The concentration of the supernatant was adjusted according to the number of MCs counted using an automated cell counter (Bio‐Rad Laboratories, Inc Hercules, CA, USA).

To evaluate the responses of the cultured podocytes against ER stress and ERAD inhibition, we administered Eeyarestatin I (EerI; R&D systems, Minneapolis, MN, USA), which is an ERAD inhibitor, at 1‐5 µg/mL to the cultured podocytes. EerI specifically targets the p97‐associated deubiquitinating process and inhibits ataxin‐3‐dependent deubiquitination. The cells were incubated for 24 hours at 37°C. Moreover, we administered kifunensine (Kif; Sigma‐Aldrich), which has ER‐associated mannosidase inhibitory activity, at a final concentration of 100 µM to suppress the ERAD function.

2.2 Western blot analysis

Protein samples in cultured podocytes, homogenized mouse kidney tissue, and mouse glomeruli were extracted and sorted using sieving methods,^{10, 17} electrophoresed in SDS‐polyacrylamide gel, and transferred onto polyvinylidene fluoride (PVDF) membranes. The membranes were blocked with 3% skim milk in PBS Tween 20 (Sigma‐Aldrich) and incubated with primary antibodies at 4°C overnight. Then, they were washed and incubated with the secondary antibody conjugated to horseradish peroxidase (HRP) for 2 hours. Bands were detected using enhanced chemiluminescence (Amersham ECL Prime and ECL Select; GE Healthcare, Pittsburgh, PA, USA).¹⁸ Quantification of the bands was conducted using ImageJ Ver. 1.45 (National Institutes of Health, US). All primary and secondary antibodies used in this study are listed in Table S1.

2.3 Cluster analyses

To evaluate the critical pathways involved in diabetic glomerular lesion, we conducted cluster analyses using DAVID bioinformatics resources.¹⁹ The cDNA microarray dataset, with commonly increased (more than double) or decreased (less than half) number of genes in the kidneys of two diabetic mice, was enrolled in the cluster analyses, as we previously performed on diabetic glomeruli that were isolated from two distinct diabetic mouse models: streptozotocin (STZ)‐induced diabetic mice and A‐ZIP/F‐1 diabetic mice.¹⁰ As for the STZ‐induced mice, diabetes was induced at 9 weeks of age by a single intraperitoneal injection of STZ (180 mg/kg body weight; Sigma‐Aldrich), and the mice were analyzed 8 weeks later. A‐ZIP/F‐1 mice are the model for lipoatrophic diabetes; this model, whose characteristics are in contrast to those of STZ‐induced insulin‐dependent diabetic model, exhibits severe insulin resistance, hyperinsulinemia, massive proteinuria, and mesangial expansion such as human type 2 diabetic nephropathy.²⁰ Enrichment scores were determined, and the indices and significance of each cluster were calculated using geometric mean as previously described.^{19, 21}

2.4 Quantitative PCR array

Total RNA was isolated from MPC5 using RNeasy Mini Kit and RNeasy Micro Kit (Qiagen, Hilden, Germany) according to the manufacturer's instructions. The cDNA in each sample was synthesized using PrimeScript RT Master Mix (Takara Bio, Otsu, Japan). We used a quantitative PCR array kit specific for ER stress‐associated genes (RT² Profiler PCR Arrays Unfolded Protein Response; Qiagen); this PCR array kit can establish profiles of the expression of 84 key genes that recognize and respond to the accumulation of misfolded proteins in the ER. The relative expression of each gene was calculated using the ΔΔCT method and normalized against the expression of designated housekeeping genes.

2.5 Evaluation of apoptosis by TUNEL staining

For TUNEL staining, we used In Situ Cell Death Detection Kit, Fluorescein (Roche Life Science, Indianapolis, IN, USA) according to the manufacturer's instructions. Samples were fixed in 4% paraformaldehyde in phosphate buffered saline (PBS) for 10 minutes at room temperature (25 ± 2°C, the same applies hereafter), washed three times with PBS, and then, treated with 0.1% Triton X‐100 in 0.1% sodium citrate for 4 min at 4°C. The samples were washed with PBS and stained for 60 minutes at 37°C in the absence of light. The staining reaction mix (enzyme solution; terminal deoxynucleotidyl transferase and labeling solution; fluorescein‐dUTP) was prepared according to the manufacturer's instructions. Subsequently, the samples were washed three times with PBS and stained with DAPI solution (2 µg/mL) for 30 minutes at 37°C. To establish the negative control, samples were incubated with labeling solution only. To establish the positive control, samples were treated with DNase I (10 units/µL) for 10 minutes at room temperature.²²

2.6 Immunofluorescence

Renal tissue was frozen in optimal cutting temperature (OCT) compound and cryosectioned with 4 µm thickness. The sections were fixed in acetone, blocked in 1% BSA, and incubated with primary antibodies at 4°C overnight. Afterward, the samples were washed with PBS and incubated with Alexa Fluor‐conjugated secondary antibodies (Alexa Fluor 488 goat anti‐guinea pig, Alexa Fluor 488 goat anti‐rabbit, and Thermo Fisher Scientific) for 1 hour at room temperature. Immunofluorescence images were captured using BZ‐X700 All‐in‐one Fluorescence Microscope (Keyence, Osaka, Japan). At least seven glomeruli were randomly selected from the samples, and fluorescence intensity was quantified; the mean intensity was calculated using BZ‐H3A software (Keyence).

2.7 Flow cytometry

We used db/db diabetic mice and MafB‐GFP knock‐in mice. Podocytes were sorted according to the expression of nephrin in db/db mice for the subsequent evaluation of apoptotic responses and Derlin‐2 expression. We employed MafB‐GFP knock‐in mice²³ for the evaluation of nephrin phosphorylation in podocytes. MafB is a member of the Maf family of transcription factors and has been reported to be expressed in podocytes and macrophages. The GFP gene was inserted into the chromosomal mafB locus in MafB‐GFP knock‐in mice. Glomeruli were isolated using a sieving method, and the isolated glomeruli were dissociated using MACS Cell Separation (Miltenyi Biotec, Bergisch Gladbach, Germany) according to the manufacturer's instructions. The cultured podocytes were incubated in 1 mM EDTA for 30 minutes at 37°C for dissociation, centrifuged at 410 g for 5 minutes, and resuspended in cold (4°C) PBS at 1.0‐1.5 × 10⁵ cells/mL. The cell suspension was blocked with 1% BSA in PBS for 10 minutes at 4°C and incubated with primary antibodies for 40 minutes at 4°C. The cells were washed and incubated with a secondary antibody (Alexa Fluor 405‐, 555‐conjugated; Thermo Fisher Scientific) for 40 minutes at 4°C in the absence of light. Apoptosis was evaluated using Annexin V‐FITC Apoptosis Detection Kit (Nacalai Tesque) according to the manufacturer's instructions. The cell suspensions were incubated with secondary antibodies and then, with Annexin V‐FITC and propidium iodide solutions for 15 minutes at room temperature. The samples were washed with PBS, resuspended in PBS, and analyzed using flow cytometry (SH800S Cell Sorter, Sony Biotechnology).²⁴

2.8 Animal models

Type 2 obese diabetic db/db (BKS.Cg‐+ Lepr^db/+ Lepr^db/Jcl) mice and lean control db/m (BKS.Cg‐m +/+ Lepr^db/Jcl) mice were purchased from CLEA Japan (Tokyo, Japan). Eight‐week‐old male mice were subcutaneously administered 1 mg/kg/day EerI or vehicle for 14 consecutive days using ALZET Osmotic Pumps (Durect, Cupertino, CA, USA). Urine samples were collected within 24 hours periods prior to and 10 days after drug or vehicle administration using metabolic cages (Shinano, Tokyo, Japan). Urine albumin and creatinine levels were measured using an immunoturbidimetric method (Oriental Yeast, Shiga, Japan). Blood was obtained from the tail vein, and blood glucose levels were measured using StatStrip Xpress (Nova Biomedical, Waltham, MA, USA) ad libitum. For the evaluation of podocytes using flow cytometry, glomeruli were isolated from 8‐ to 10‐week‐old male db/db mice and MafB‐GFP knock‐in mice that were treated with EerI for three consecutive days. All animal experiments were conducted in accordance with the guidelines for care and use of laboratory animals, and the experimental designs were approved by Kumamoto University (No. A30‐013).

2.9 Statistical analysis

Statistical data are expressed as mean ± standard error of the mean (SEM). Statistically significant differences among multiple groups were assessed using two‐way factorial ANOVA and Bonferroni's post hoc test. Comparisons between two groups were performed using unpaired Student's t test. A P < .05 was considered statistically significant.

3 RESULTS

3.1 ER stress is a potential factor in the pathogenesis of DKD

To evaluate the pathways involved in the pathophysiological changes of glomeruli in diabetes, we performed cluster analyses using DAVID Bioinformatic Resources. We analyzed the cDNA microarray dataset of glomeruli from two diabetic mouse models,¹⁰ as described in the methods section (Figure 1). We selected these diabetic models to minimize the interferences from the renal toxicity of STZ, genetic background,²⁵ and the direct target molecules of insulin or leptin.²⁶ Cluster analyses revealed that the gene clusters associated with apoptosis/cell death and inflammatory response, which were commonly increased in the two diabetic mouse models (Table 1A), exhibited the highest enrichment scores (both 3.03) in the glomeruli. In contrast, as shown in Table 1B, the gene clusters associated with redox and ATP energy metabolism were included under the genes that had their numbers decreased in both diabetic models, and relatively high enrichment scores (2.08 and 1.24, respectively), were assigned to these genes. Relation and cross talk between inflammation and ER stress in human diseases have been frequently reported^{27, 28}; moreover, apoptosis and ER stress also have been frequently reported.^29-31 Therefore, we presumed that ER stress might be involved in the progression of diabetic glomerular lesions. ATP depletion and energy metabolism may be associated with various kinds of pathophysiology; however, the ER and mitochondria are closely related.³² From these results, we speculate that ER stress is linked to the progression of diabetic glomerular lesions.

FIGURE 1

Schematic flowchart of the cluster analysis of genes in diabetic glomeruli. Glomeruli were isolated from insulin‐dependent and insulin‐resistant diabetic mouse models using a sieving method. Cluster analyses were performed using the dataset obtained from the glomerular cDNA microarray analysis

TABLE 1. Cluster analyses of the genes from diabetic glomeruli in STZ and A‐ZIP diabetic mouse models

Annotation cluster	Genes	Enrichment score
(A) STZ/B6 cont. > 2.0, A‐ZIP/FVB cont. > 2.0 (270 genes)
Classification stringency medium
Cluster: 84, Enrichment Score > 1.5
Cluster 1	Apoptosis	3.03
	Cell death
Cluster 2	Inflammatory response	3.03
	Defense response
Cluster 3	Single peptide	2.61
	Glycoprotein
Cluster 4	Response to wounding	2.38
	Complement and coagulation
	Acute inflammatory response
Cluster 5	Guanylate‐binding protein	2.09
	GTPase activity
Cluster 6	Negative regulation of single transduction	1.91
	Negative regulation of BMP signaling pathway
	Regulation of BMP signaling pathway
Cluster 7	Signal	1.85
	Secreted
	Extracellular space
	Extracellular region
	Extracellular region part
(B) STZ/B6 cont. < 0.5, A‐ZIP/FVB cont. < 0.5 (75 genes)
Classification stringency medium
Cluster: 15, Enrichment Score > 1.0
Cluster 1	Oxide reductase	2.08
	Oxidation reduction
	Steroid metabolic process
	NAD(P)‐binding domain
	Steroid dehydrogenase activity
Cluster 2	ATP biosynthetic process	1.24
	ATP metabolic process
	Ribonucleoside triphosphate
	Biosynthetic process
	Purine ribonucleoside triphosphate biosynthetic process
	Ribonucleoside metabolic process
	Purine ribonucleotide metabolic process
	Purine nucleotide metabolic process
	Nucleobase, nucleoside, and nucleotide biosynthetic process
	Adenyl nucleoside binding
	Purine nucleoside binding
	Nucleotide binding
	Nitrogen compound biosynthetic process

Note

• Data on the commonly increased (more than double, panel A) or decreased (less than half, panel B) genes in the two distinct diabetic mouse models were obtained using DAVID Bioinformatics Resources. Enrichment scores were calculated using the P‐values showing the significance of relevance of each annotation cluster; enrichment score greater than 1.3 was considered statistically significant.

3.2 Mesangial cell‐derived humoral factors suppress ER‐associated degradation and promote apoptosis in podocytes under HG conditions

We evaluated podocyte ER stress response induced by a cross talk in MCs that were cultured under HG conditions (glucose concentration: 25 mmol/L). To capture the overall trends of podocyte responses against MC‐supernatant stimulation, we first conducted a quantitative PCR array focused on ER stress‐associated genes.

Based on the results of the PCR array, a heatmap was generated indicating that the genes related to apoptosis, protein folding, and ER chaperones tended to be upregulated, whereas the genes related to the ERAD pathway were downregulated by HG MC‐sup (supernatant of mesangial cells) stimulation (Figure S1A‐C). These results suggest that the cross talk in MCs under HG conditions can suppressively affect the ERAD pathway in podocytes and induce apoptotic responses. Western blot results indicated that the protein expression of CHOP, which is a molecule related to apoptotic response against ER stress load, increased, whereas the expression of ERAD‐related proteins such as phosphorylated IRE1α, Derlin‐1, and Derlin‐2 decreased upon stimulation with HG MC‐sup; this is in reference to the protein expression in podocytes that were stimulated with MC‐sup and cultured under NG conditions (Figure 2A,B). Phosphorylated IRE1α promotes the expression of ERAD‐related molecules. We considered that the increased ratio of phosphorylated IRE1α is consistent with the increased expression of Derlin‐1 and ‐2. The MPC5 treated with EerI, which is an ERAD inhibitor, exhibited nearly strengthened responses against HG MC‐sup stimulation (Figure 2A,B). These results are generally consistent with the microarray data on diabetic mouse glomeruli. Some of the representative apoptosis‐related genes tended to be upregulated, whereas ERAD‐related genes tended to be downregulated in two diabetic mouse models (Figure S2).

FIGURE 2

Western blot analysis of ER stress markers of podocytes stimulated with NG MC‐sup, HG MC‐sup, and Eeyarestatin I (EerI). A, The protein expression levels of CHOP, IRE1α (both total; t‐ and phosphorylated; p‐ forms), Derlin‐1, and Derlin‐2 in MPC5 stimulated with the supernatant of mesangial cells cultured under normal glucose conditions (NG [+D‐mannitol)]) MC‐sup, the supernatant of mesangial cells cultured under high glucose conditions (HG MC‐sup) and Eeyarestatin I for 24 hours are shown. B, Quantification of the bands in panel A. n = 4

We conducted western blot analyses of Bax (pro‐apoptotic) and Bcl‐2 (anti‐apoptotic) in order to evaluate the apoptosis. The ratio of Bax/Bcl‐2 was elevated (Figure 3A,B) by the stimulation of podocytes with HG MC‐sup and ERAD inhibitor. Apoptotic evaluation by TUNEL staining of MPC5 also showed that the number of TUNEL‐positive cells significantly increased with stimulation by HG MC‐sup and ERAD inhibitor (Figure 3C,D). From the results thus far, we speculate that the humoral factors derived from MCs cultured under HG conditions induce ERAD suppression and apoptosis in podocytes; these alterations were not observed in MPC5 cultured in HG or NG (with or without D‐mannitol) medium for 24 to 48 hours without MC‐sup treatment (Figure S3). Therefore, we suggest that ERAD‐inhibitory and apoptosis‐inducible responses are mainly mediated by MC‐sup under HG conditions, irrespective of the glucose concentration in podocyte cultures.

FIGURE 3

Evaluation of apoptosis of podocytes stimulated with HG MC‐sup and Eeyarestatin I. A, Expression levels of Bax and Bcl‐2 determined using western blot analysis. B, Bax/Bcl‐2 protein expression ratio determined using western blot analysis. *P < .05 vs Veh. n = 4. C, TUNEL staining of MPC5 treated with the supernatant of mesangial cells cultured under normal glucose conditions (NG MC‐sup), the supernatant of mesangial cells cultured under high glucose conditions (HG MC‐sup) and Eeyarestatin I. Blue, DAPI; Green, TUNEL. D, The quantification of the percentage for TUNEL‐positive cells. *P < .05, n = 4

To further investigate the mechanism of ERAD inhibition, we used kifunensine (Kif), an ER‐associated mannosidase inhibitor. EerI and Kif block the ERAD pathway; however, EerI prevents the retrotranslocation of irremediably misfolded substrates, whereas Kif prevents the early recognition of misfolded intermediates.³³ Western blot results are shown in Figure S4; EerI or HG MC‐sup downregulated Derlin‐1 and ‐2 expression in podocytes, whereas Kif did not.

3.3 Administration of an ERAD inhibitor exacerbates podocyte injury in db/db mice

To evaluate the functional role of ERAD inhibition in podocyte injury under HG conditions, we conducted an in vivo experiment using type 2 diabetic db/db mice. The administration of EerI to mice for 2 weeks did not affect their body weight, urine volume, or blood glucose levels (Figure 4A‐C).However, the urine albumin levels were significantly increased in the db/db + EerI group than in the db/db + Veh group (Figure 4D).

FIGURE 4

Effects of continuous administration of an endoplasmic reticulum‐associated degradation (ERAD) inhibitor on db/m mice and diabetic db/db mice. Body weight (A) and blood glucose (B) were measured at days 0 and 8 after Eeyarestatin I (EerI) administration. Urine collection (C, urinary volume; D, urinary albumin excretion) was performed from day 1 to 0 and day 10 to 11 after EerI administration started. n = 5‐8; *P < .05

Apoptosis was induced in diabetic mice and EerI‐administered nondiabetic mice. The combination of diabetes and ERAD inhibition significantly increased the apoptosis of podocytes (Figure 5A,B). The expression of Derlin‐2, which is an ERAD‐related molecule, was significantly decreased, in contrast to the apoptotic responses (Figure 5C). The immunofluorescence analysis of glomeruli indicated that the expression of Derlin‐2 and nephrin was significantly reduced in the mice that were administered an ERAD inhibitor (Figure 6). Derlin‐2 is also expressed in the interstitium of kidneys^{14, 34}; however, our double immunostaining of Derlin‐2 and WT‐1, a podocyte nucleus marker, indicated that Derlin‐2 was expressed predominantly in glomeruli, and their expression generally co‐localized. (Figure S5).

FIGURE 5

Flow cytometric analyses of apoptosis and Derlin‐2 expression of podocytes isolated from the kidney of db/m and db/db mice treated with endoplasmic reticulum‐associated degradation (ERAD) inhibitor. A, Annexin V and propidium iodide staining of podocytes followed by flow cytometry analysis. Podocytes were sorted by gating nephrin‐positive cell population. Early phase apoptosis (surrounded by solid circle) and late phase apoptosis (surrounded by dotted circle) were evaluated. B, The proportion of early phase apoptosis and late phase apoptosis of each group. n = 4‐5. *P < .05. C, Derlin‐2 fluorescence intensity analyzed by flow cytometry. The values shown are the ratios with reference to the db/m + Veh group. *, * *; vs db/m + Veh * P < .05, * * < .01, †; vs db/db + Veh P < .05; n = 4‐5

FIGURE 6

Derlin‐2 and nephrin expression in the glomeruli of db/m and db/db mice treated with an endoplasmic reticulum‐associated degradation (ERAD) inhibitor. Immunofluorescence of Derlin‐2 and nephrin (A), and their expression levels (B,C). n = 4‐7. *P < .05 vs db/m + Veh

3.4 ERAD inhibition suppresses nephrin phosphorylation in podocytes

We examined the effects of ERAD inhibition on the phosphorylation of nephrin in podocytes. Tyrosine phosphorylation has been reported to be important for normal podocyte morphology and function; this phosphorylation is detected using antibodies that are specific for Y1217 and Y1176/Y1193 phosphorylation.³⁵ To effectively sort podocytes, we employed MafB‐GFP knock‐in mice.^{23, 36} GFP‐positive cells sorted from isolated glomeruli were analyzed. The ERAD inhibition by EerI in mice resulted in a significant decrease in nephrin phosphorylation of podocytes (Figure 7A,B). Furthermore, in vitro experiments revealed that the phosphorylation of nephrin was significantly suppressed in podocytes treated with EerI and HG MC‐sup than that in podocytes treated with NG MC‐sup (Figure 8A,B).

FIGURE 7

Effects of endoplasmic reticulum‐associated degradation (ERAD) inhibition upon nephrin phosphorylation in podocytes of MafB‐GFP knock‐in mice. Podocytes were sorted by gating GFP‐positive and anti‐nephrin Ab‐Alexa Fluor 555‐positive cell population (a total of 100 000 cells were flowed). A, Nephrin phosphorylation of podocytes was evaluated using Alexa Fluor 405 (Y‐axis). Cell populations with nephrin phosphorylation were gated and counted. B, The percentages of podocytes with nephrin phosphorylation were quantified. n = 5. *P < .05

FIGURE 8

Effects of endoplasmic reticulum‐associated degradation (ERAD) inhibition upon nephrin phosphorylation in cultured podocytes. Phosphorylated nephrin in cultured podocytes MPC5 treated with the supernatant of mesangial cells cultured under high glucose conditions (HG MC‐sup) and EerI was evaluated and quantified by western blot analyses. n = 4. *P < .05

4 DISCUSSION

ER stress has been reported to be involved in various kinds of kidney diseases, especially proteinuric kidney diseases including DKD, in both rodents and humans.^{7, 8, 37} Recent studies have particularly focused on the role of ERAD in podocyte dysfunction.^{12, 13} In a previous study, we also observed that ERAD‐associated genes were downregulated in diabetic glomeruli.¹⁰ The mechanisms and role of ERAD suppression, however, have not yet been fully elucidated. In the present study, we have shown that a cross talk among MCs under HG conditions intervenes the normal ERAD pathway of podocytes and results in podocyte dysfunction.

There have been reports that hyperglycemia and proteinuria induces ER stress in the kidney.³⁷ Therefore, we speculated that diabetic kidney is exposed to ER stress load and that unfolded protein responses (UPR) need to be activated. IRE1α, an ER stress sensor, and ATF6 upregulate ER chaperones that promote the proper folding and ERAD of proteins³⁸; hence, ER stress and ERAD pathway are closely related. (Figure 9A). Furthermore, IRE1α is reportedly auto‐phosphorylated, catalyzing the splicing of mRNA encoding Xbp1, and driving the expression of ERAD‐associated factors such as Derlin‐1 and ‐2.^{34, 39} This pathway is critical for the degradation of misfolded proteins; a process that would attenuate ER stress load. Our comprehensive analysis of ER stress responses revealed that ERAD‐suppressive and pro‐apoptotic responses were promoted by HG‐conditioned MC medium in podocytes. Moreover, these responses were enhanced in a concentration‐dependent manner in MCs cultured in glucose. Recently, Derlin‐2, which is the downstream target protein of IRE1/XBP1, was reported to be critically involved in the susceptibility of podocytes to injury in mouse and human proteinuric nephropathies.¹⁴ Hence, our data suggest that the cross talk between MCs and podocytes can reduce the Derlin‐2 expression in podocytes through ERAD suppression, which thereby leads to podocyte injury. The mechanism of ERAD inhibition by the crosstalk remains unclear. We used Kif and EerI, which are ERAD inhibitors of distinct mechanisms. Kif inhibits the early (or preliminary) stage of ERAD, whereas EerI prevents the retrotranslocation of irremediably misfolded substrates (later phase). Our results suggest that the effects and mechanisms of HG MC‐sup were similar to those of EerI rather than those of Kif. Misfolded substrates are no longer remediable during retrotranslocation, which EerI inhibits. However, they still are recoverable upon recognition as misfolded substrates; this recognition is inhibited by Kif. The cytotoxicity of EerI is more intense than that of Kif, presumably owing to their mechanistic differences.³³ Altogether, our results indicated that the effects of HG MC‐sup was similar to those of EerI, supporting our hypothesis on the mechanism of podocyte damage.

FIGURE 9

Proposed schemas of unfolded protein responses (UPR) and mesangial cell (MC)‐podocyte cross talk under diabetic conditions. A, A schema of ER stress responses. B, Cross talk in MCs under high glucose (HG) conditions inhibits normal endoplasmic reticulum‐associated degradation (ERAD) processes in podocytes; this may suppress the nephrin phosphorylation and result in podocyte dysfunction

Although there have been reports that suggested the existence of the intercellular cross talk between MCs and podocytes,⁵ we found the elucidation of the existence of such cross talk in vivo challenging in the present study. With regard to IgA nephropathy, which is a typical mesangial proliferative kidney disease, cytokines such as TNF‐α derived from immune‐complex‐deposited MCs can affect podocyte function and induce cell damage.⁶ Furthermore, a recent study, wherein single‐nucleus RNA sequencing analyses of diabetic glomeruli² were performed, revealed that some ligand genes are upregulated in diabetic MCs and that their cognate receptors exhibit increased expression in podocytes; the presence of possible pathways of local effector molecules of MCs and podocytes was also proposed. Indeed, podocyte‐specific genetic manipulation can alter the MC phenotype through certain humoral factors in mice.⁴⁰ Furthermore, exosomal microRNAs derived from MCs were increased in the urine of diabetic patients.⁴¹ These findings may support the hypothesis of in situ communication between MCs and podocytes.

The consequences of ERAD inhibition on podocytes appeared to be cell damaging; however, the detailed mechanisms of signaling pathways that lead to cellular dysfunction remain unclear. Nephrin, which is one of the major components of the podocyte slit diaphragm, is important for the proper function of the glomerular filtration barrier. Nephrin tyrosine phosphorylation was reported to be important for stabilizing podocyte foot processes.³⁵ Our results suggest that ERAD inhibition can affect the phosphorylation of nephrin in podocytes and can cause podocyte dysfunction in the diabetic milieu. Furthermore, mutated or improperly folded nephrin proteins cannot achieve physiological trafficking to the plasma membrane and are, therefore, retained in the ER.^42-45 Such misfolded proteins are normally degraded by the ERAD system; however, when the ERAD pathway fails to work, the process is stagnated and may hamper the proper trafficking and phosphorylation of nephrin. This hypothesis is supported by some in vitro experiments wherein ERAD inhibition or HG MC‐sup stimulation of podocytes resulted in decreased nephrin phosphorylation. Unfortunately, a limitation of our study is that we could not show the direct functional analyses of ERAD due to technical problems. Illustrated in Figure 9B is a proposed scheme for the possible interaction of MC‐podocyte cross talk under diabetic conditions. From the results of this study, we consider any agents or manipulations that can accelerate the ERAD pathway to be potential therapeutic strategies. We are considering a plan to search for potential agents that function as ERAD activators by drug screening or drug repositioning.

In conclusion, our results suggest that an intraglomerular cross talk between MCs and podocytes can inhibit normal ERAD processes and suppress nephrin phosphorylation in podocytes, thereby leading to podocyte injury under diabetic conditions. Interventions in the ERAD pathway or the cross‐communication between MCs and podocytes may be novel therapeutic strategies for DKD.

ACKNOWLEDGMENTS

We thank Ms Hikari Shibuta, Kazumi Saito, and Naoko Hirano for their technical assistance and Ms Noriko Nakagawa and Miki Horikiri for their secretarial assistance.

CONFLICT OF INTEREST

The authors have stated explicitly that there are no conflicts of interest in connection with this article.

AUTHOR CONTRIBUTIONS

T. Kuwabara and D. Fujimoto designed the study; D. Fujimoto, Y. Hata, S. Umemoto, T. Kanki, and T. Kuwabara performed the experiments, and M. Mukoyama supervised the entire study; S. Takahashi generated the MafB‐GFP knock‐in mice; D. Fujimoto, T. Kuwabara, and M. Mukoyama drafted the manuscript; T. Kuwabara, T. Mizumoto, M. Hayata, Y. Kakizoe, Y. Izumi, and M. Mukoyama interpreted the results. All authors approved the final version of the manuscript.