Podocyte injury is important in development of diabetic nephropathy (DN). Although several studies have reported single-cell-based RNA sequencing (RNA-seq) of podocytes in type 1 DN (T1DN), the podocyte translating mRNA profile in type 2 DN (T2DN) has not previously been compared with that of T1DN. We analyzed the podocyte translatome in T2DN in podocin-Cre; Rosa26fsTRAP; eNOS−/−; db/db mice and compared it with that of streptozotocin-induced T1DN in podocin-Cre; Rosa26fsTRAP; eNOS−/− mice using translating ribosome affinity purification (TRAP) and RNA-seq. More than 125 genes were highly enriched in the podocyte ribosome. More podocyte TRAP genes were differentially expressed in T2DN than in T1DN. TGF-β signaling pathway genes were upregulated, while MAPK pathway genes were downregulated only in T2DN, while ATP binding and cAMP-mediated signaling genes were downregulated only in T1DN. Genes regulating actin filament organization and apoptosis increased, while genes regulating VEGFR signaling and glomerular basement membrane components decreased in both type 1 and type 2 diabetic podocytes. A number of diabetes-induced genes not previously linked to podocyte injury were confirmed in both mouse and human DN. On the basis of differences and similarities in the podocyte translatome in T2DN and T1DN, investigators can identify factors underlying the pathophysiology of DN and novel therapeutic targets to treat diabetes-induced podocyte injury.

The increasing prevalence of diabetes in the world has become a major and unavoidable threat to public health, health care systems, and the economy (1). Diabetic nephropathy (DN) is one of the most prominent microvascular complications of diabetes and continues to be the leading cause of end-stage renal disease and a major source of morbidity and mortality (2). Podocytes play a fundamental role in maintaining glomerular filtration function, and numerous studies suggest that podocyte injury plays a critical role in development and progression of DN (3,4). Understanding gene expression changes in podocytes in vivo is necessary to better understand the role of podocytes in DN. There have been previous broad-based approach studies profiling gene expression of glomeruli or cortex of diabetic kidneys to uncover the pathogenic mechanisms of DN (57). However, these previous studies provided somewhat limited information on podocyte injury per se due to averaging of information across different cell types.

Using transcriptomic analyses of isolated glomeruli from streptozotocin (STZ)-induced diabetic endothelial nitric oxide synthase knockout (eNOS−/−) mice, He and colleagues showed alteration of genes and pathways in endothelial injury, proliferation, and angiogenesis in isolated glomerular endothelial cells (8) and alterations of genes and signaling pathways relating to actin cytoskeleton and cell adhesion in isolated podocytes (9). Using single-cell RNA-sequencing (scRNA-seq) analyses, several groups have investigated gene expression in isolated glomerular cells in normal kidneys as well as diabetic glomeruli from experimental models of STZ-induced type 1 diabetes and kidney (813). However, the findings in relation to STZ-induced type 1 DN (T1DN) may not always be identical to those related to podocyte injury resulting from type 2 diabetes and, to date, no studies have identified podocyte-specific gene expression alterations in experimental models of T2DN.

Translating ribosome affinity purification (TRAP) is an effective method to identify active translation in specific cells with a heterogeneous tissue in vivo. The TRAP approach was developed to detect expression profile changes in rare cell populations, such as adult neuronal cells, where recovery of intact cells is almost impossible (14,15). TRAP strategy depends on the Cre recombinase–dependent activation of a GFP-tagged L10a ribosomal protein subunit (GFP-L10a) within the cell type of interest. With use of anti-GFP beads, ribosome-bound mRNA populations within the cell type of interest are purified from a whole-organ lysate. Unlike microdissection, cell panning, cell sorting, or other single-cell isolation techniques, TRAP does not require extensive fixation, mechanical dissociation, or enzymatic digestion and thereby can maintain the original intact mRNA profile with minimal in vitro stress responses. In addition, the most important advantage of the TRAP approach is that it only identifies the subset of mRNAs being actively translated from the specific cell population at the time of study. As a key regulatory node in gene expression, the profile of mRNA undergoing translation correlates better with actual protein expression than total mRNA and is more relevant to the cell phenotype (16,17). To date, the use of the TRAP approach in kidney cells has only been reported in acute kidney injury, focal segmental glomerulosclerosis and fibrosis, and unilateral ureteral obstruction (1821). For better understanding of the role of podocytes in DN, both type 1 and type 2 diabetic mouse models were used to investigate the translating mRNA profiling of podocytes with use of the TRAP approach.

Animal Studies

All animal experiments were performed in accordance with the guidelines of the Institutional Animal Care and Use Committee of Vanderbilt University. C57BLKS/J (BKS) eNOS−/− mice were crossed first with podocin-Cre mice and then with Rosa26fsTRAP mice (22,23). The resultant male podocin-Cre; Rosa26fsTRAP; eNOS−/− mice (10 weeks old, BKS) were injected with STZ at 50 mg/kg i.p. for five consecutive days for induction of type 1 diabetes. Onset of diabetes was evaluated through measurement of fasting blood glucose with a β-Glucose analyzer (HemoCue, Lake Forest, CA). Spot urine albumin-to-creatinine ratio (ACR) was evaluated with Albuwell M kits (Exocell Inc., Philadelphia, PA). Type 2 diabetes BKS podocin-Cre; Rosa26fsTRAP; eNOS−/−; db/db mice were also generated. Male age-matched BKS eNOS−/− mice were used as control for both type 1 and type 2 diabetes models.

TRAP mRNA Isolation

We purified TRAP mRNAs purified from mouse kidneys following approaches described by Heiman et al. (24). Briefly, after mice were anesthetized, kidneys were removed, washed with ice-cold tissue lysis buffer, minced into small pieces, and transferred to a prechilled homogenizer with tissue lysis buffer. A postnuclear supernatant was prepared by centrifugation of the lysate at 4°C for 10 min at 2,000g. After addition of NP-40 and DHPC for 5 min on ice, the postmitochondrial supernatant was prepared by centrifugation at 4°C for 10 min at 20,000g. The affinity matrix was prepared with coating of Streptavidin MyOne T1 Dynabeads (Invitrogen/Thermo Fisher Scientific) with GFP antibodies 19C8 and 19F7 (HtzGFP-19F7 and HtzGFP-19C8; Antibody and Bioresource Core Facility, Memorial Sloan-Kettering Cancer Center). Freshly resuspended beads were added to the postmitochondrial supernatant and incubated at 4°C for 18 h with gentle end-over-end mixing. The beads were collected with a well-chilled magnet on ice. After four washes with high-salt buffer, the beads were resuspended in 100 μL Nanoprep lysis buffer with β-mercaptoethanol. The RNA was then isolated with a QIAGEN RNeasy Micro Kit with on column DNase digestion according to the manufacturer’s instructions. RNA quantity and quality were determined with an Agilent 2100 Bioanalyzer. Total kidney RNAs were extracted using TRI Reagent (Molecular Research Center, Cincinnati, OH) and chloroform extraction followed by purification with an RNeasy kit (QIAGEN, Valencia, CA).

RNA-seq

Global RNA sequencing was performed on all TRAP samples using the Ovation RNA-Seq Systems 1–16 for Model Organisms—Mouse (NuGEN Technologies, Redwood City, CA). The sample sizes were as follows: n = 3 in the 6-month-old T2DN group, n = 4 in control for 3-week-old T1DN and control for 6-month-old T2DN groups, and n = 5 in all other groups. Library preparation was performed according to the manufacturer’s instructions. RNA underwent DNase treatment, and cDNA was fragmented (200 base pairs) on a Covaris LE220. Prior to sequencing, cDNA libraries were checked for quality (High Sensitivity DNA Kit; Agilent Technologies, Santa Clara, CA) and quantity (Invitrogen Qubit dsDNA HS Assay kit; Thermo Fisher Scientific, Waltham, MA), and equal molar concentrations were pooled and concentrated (DNA Clean & Concentrator kit; Zymo Research, Irvine, CA). Multiplexed samples were run across three lanes, and RNA-seq was performed with the NovaSeq 6000 (Illumina, San Diego, CA) by the Vanderbilt Technologies for Advanced Genomics (VANTAGE) core laboratory. Reads were trimmed to remove adapter sequences using cutadapt v1.16 (25). GENCODE vM16 gene annotations were provided to STAR for improvement of the accuracy of mapping. Quality control on both raw reads and adaptor-trimmed reads was performed with FastQC (https://www.bioinformatics.babraham.ac.uk/projects/fastqc/). featureCounts v1.15.2 was used to count the number of reads mapped to each gene (26). Fragments per kilobase of transcript per million mapped reads (FPKM) values for all genes in each sample were calculated based on gene length in GENCODE vM16 gene annotations.

RNA-seq Data Analysis

R package DESeq2 was used to analyze differential gene expression between TRAP mRNA samples. DESeq2 estimates variance-mean dependence in raw count data and tests for differential expression based on a model using negative binomial distribution (27). Genes with fold change >1.5 and P value <0.05 were considered to have significant differences of expression. The Database for Annotation, Visualization and Integrated Discovery (DAVID 6.8) was used to identify functional categories for the genes enriched in podocytes and genes with significant differences of expression (28,29). Gene Ontology Biological Process (GO_BP), Gene Ontology Molecular Function (GO_MF), and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichments were analyzed with DAVID, with use of all 735 podocyte-specific protein-coding genes as background for genes with significant differences of expression. The RNA-seq data were submitted to the National Center for Biotechnology Information Gene Expression Omnibus (GEO) database.

Antibodies

Rabbit anti-neuroglobin was from Proteintech (13499-1-AP), rabbit anti-TSPAN2 was from Bioss (bs-9411R), rabbit anti-claudin 5 (34-1600) and rabbit anti-nephrin were from Thermo Fisher Scientific, rabbit anti–Wilms tumor protein (WT1) (a marker of podocytes) was from Abcam (ab89901), and rabbit anti-MAGI2 was from Life Technologies Corporation (PA547901).

Immunostaining and Immunofluorescent Staining

The kidney was fixed in 3.7% formaldehyde, 10 mmol/L sodium m-periodate, 40 mmol/L phosphate buffer, and 1% acetic acid and then dehydrated through a graded series of ethanols, embedded in paraffin, sectioned (5 µm), and mounted on glass slides. Immunostaining and immunofluorescent staining were performed as previously reported (30). Immunofluorescent staining of relative normal human kidney and human T2DN was obtained from de-identified nephrectomy samples, approved by Vanderbilt University Medical Center Institutional Review Board. There were seven nephrectomy samples from patients with no evidence of diabetes and four kidney biopsy samples from patients with moderate to severe DN (three men and one woman). Samples were from individuals 55–72 years of age.

Isolation of Glomeruli Using Dynabeads M-450

Mouse glomeruli were isolated as previously described (31). Briefly, mice were anesthetized with an injection of nembutal sodium solution (50 mg/kg i.p.) and perfused with 4 × 107 Dynabeads diluted in 20 mL PBS through the heart. The kidneys were removed, minced, and digested in collagenase II (1 mg/mL, cat. no. C6885; Sigma-Aldrich, St. Louis, MO) and DNase I (50 units/mL, D5025; Sigma-Aldrich) at 37°C for 5 min with agitation. The tissue suspension was pressed through a 100-μm cell strainer. Glomeruli containing Dynabeads were gathered by a magnetic particle concentrator and washed with ice-cold PBS. The isolated glomeruli were lysed and used for immunoblot analysis.

Real-time PCR and Periodic Acid Schiff Staining

Total RNAs of glomeruli were isolated with TRIzol reagent (Invitrogen). SuperScript III First-Strand Synthesis System (Invitrogen) was used to synthesize cDNA from equal amounts of total RNA from each sample. Quantitative RT-PCR was performed with TaqMan real-time PCR (7900HT; Applied Biosystems). The master mix and all gene probes were also purchased from Applied Biosystems. The probes used in the experiments included mouse GAPDH (Mm99999915), Ngb (Mm00452101), Wt1 (Mm01337048), Mgat5b (Mm01252571), Cldn5 (Mm00727012), and Magi2 (Mm01159076). Periodic acid Schiff staining was performed according to the protocol provided by the manufacturer (Sigma-Aldrich).

Statistical Analysis

Statistical analyses were performed with GraphPad Prism7 (GraphPad Software, La Jolla, CA). Data are presented as means ± SEM. Comparisons between groups were made with Fisher exact test, two-tailed unpaired Student t test, or ANOVA with Bonferroni post hoc test.

Data and Resource Availability

All RNA-seq data were deposited in the GEO database (GSE173989).

Generation of Podocyte-Specific TRAP Mice With T1DN and T2DN

We have reported that eNOS−/−; db/db mice develop accelerated T2DN, characterized by significant albuminuria, decreased glomerular filtration rate, increased glomerular basement membrane thickness, mesangial expansion, mesangiolysis, and focal segmental and early nodular glomerulosclerosis (32). We generated podocin-Cre; Rosa26fsTRAP; eNOS−/−; db/db mice to investigate the podocyte translatome profile in 3-month-old (early stage) and 6-month-old (advanced stage) mice during the development of T2DN, using TRAP plus RNA-seq (Fig. 1A). Both fasting blood sugars and albuminuria were already higher than those of control eNOS−/− mice at age 3 months (Fig. 1B and Supplementary Fig. 1A). We and others have also reported that STZ-induced type 1 diabetes leads to robust albuminuria, mesangial expansion, mesangiolysis, and glomerular sclerosis in eNOS−/− mice compared with mild albuminuria and mesangial expansion in wild-type mice (9,10,33,34). Therefore, we also generated podocin-Cre; Rosa26fsTRAP; eNOS−/− mice and induced type 1 diabetes with STZ to investigate the podocyte translatome during the development of T1DN (Fig. 1A). As expected, STZ-treated podocin-Cre; Rosa26fsTRAP; eNOS−/− mice quickly developed hyperglycemia (Supplementary Fig. 1B) and albuminuria as early as 3 weeks after STZ injection (Fig. 1C). Of note, control podocin-Cre; Rosa26fsTRAP; eNOS−/− mice also had mild albuminuria due to vascular permeability in microvascular beds (32). For determination of podocyte responses to type 1 diabetes over time, podocyte-specific TRAP mRNAs were isolated from mouse kidneys at 3 and 24 weeks after STZ injection, representing an early stage and a later, more advanced stage of T1DN, respectively. Moderate or severe mesangial expansion and progressive glomerulosclerosis were observed in comparisons of 3- and 6-month-old podocin-Cre; Rosa26fsTRAP; eNOS−/−; db/db mice (Supplementary Fig. 1C). Mesangial expansion was observed in STZ-podocin-Cre; Rosa26fsTRAP; eNOS−/− mice at 3 weeks and persistent mesangial expansion and severe glomerulosclerosis were observed at 24 weeks after STZ (Supplementary Fig. 1D). In these two mouse lines, podocin-Cre drives Cre recombinase expression in podocytes, labeling the ribosomes of podocytes by expression of GFP fused to L10a (GFP-L10a), an integral component of the 60S ribosomal subunit (22,23). As indicated in Fig. 1D, fluorescent GFP-L10a was specifically expressed in the podocytes in podocin-Cre; Rosa26fsTRAP; eNOS−/− mice.

Generation of the Podocyte-Enriched Translational Profile

RNA sequencing was performed for each TRAP mRNA sample. A total of 13,501 RNAs were identified with reliable expression with FPKM >1 in diabetic or control podocytes. The vast majority (11,979) were protein-coding gene mRNAs, with processed pseudogene mRNAs (552) being the second most abundant species (Fig. 2A). Small amounts of other pseudogene RNAs, long intergenic noncoding RNAs (lincRNAs), antisense RNAs, and other noncoding RNAs were also detected in the TRAP mRNA samples, probably due to background RNA binding (Fig. 2A). Since the TRAP approach mainly isolates the mRNAs undergoing translation, we focused on the protein-coding mRNAs only.

Many genes expressed in podocytes are also expressed in other kidney cells. To focus on the mouse podocyte-enriched translational profile during diabetes, we sequenced podocyte TRAP mRNAs from five nondiabetic mice and one mRNA sample pooled from the whole kidneys of the same five mice to identify a set of genes with enriched expression in podocytes. A total of 735 protein-coding genes were enriched more than threefold with FPKM >1.5 in podocyte TRAP mRNA compared with whole kidney mRNA, and 126 of these genes were enriched >15-fold with FPKM >15 (Fig. 2B and Supplementary Table 1). Well-known podocyte-specific genes, including Wt1, Nphs1, Nphs2, Synpo, and Ptpro, were among the 126 highly enriched and expressed genes in podocytes (Fig. 2C and D). However, housekeeping genes and genes typically expressed in other kidney cell types, such as Slc34a1, Ggt1, Pecam1, Tie1, Cdh1, Aqp1, and Aqp4, were not enriched at all (Fig. 2D), indicating that this gene set only includes genes specifically enriched in podocytes.

We also detected translation of some potential novel podocyte-specific marker genes. As examples, Mgat5b, the fourth most enriched gene in our podocyte TRAP gene set, encodes α-1,6-mannosylglycoprotein 6-β-N-acetylglucosaminyltransferase B, and its expression and function in podocytes are unknown (Table 1). Cldn5 is a tight-junction protein, and Ngb is a hypoxia-induced globin. Nr4a1, Nr4a3, and Mafa are also on the list of the 126 highly enriched and expressed podocyte genes, and their specific functions in podocytes have not been investigated (Supplementary Table 1).

Comparison Between Our Podocyte TRAP mRNA Profile and Published Podocyte Gene Profiles

We compared our podocyte TRAP gene set with previously published data sets. Using FACS sorting plus microarray analysis, Brunskill et al. (35) identified 141 protein-coding genes highly enriched and expressed in podocytes, and 135 of these genes were included in our podocyte TRAP gene set, with only 6 of them excluded (Fig. 2E). Some of the exclusion of genes could be caused by response to stress during the processing of FACS isolation. For example, the Hspa8 gene, a member of the heat shock protein 70 family, which protects the proteome from stress family 70 (36), was not enriched in our podocyte TRAP gene set (fold enrichment 0.56). In another podocyte gene set obtained by FACS sorting, 184 podocyte genes identified by Boerries et al. (37) were present in our podocyte TRAP gene set. Most podocyte genes (205 of 277) identified by Park et al. (38) using single-cell transcriptomic analysis were enriched more than threefold in our podocyte TRAP gene set (Fig. 2E). Most genes indicated to be podocyte specific (186 of 345) identified by Wu et al. (13) using scRNA-seq were found in our podocyte TRAP gene set. Recently, Fu et al. (10) identified 11 genes specifically expressed in podocytes using scRNA-seq, and all of them were strongly enriched in our podocyte TRAP gene set (Supplementary Table 1). Three of these genes (Enpep, Dpp4, and Ildr2), which were also expressed in kidney tubular cells according to scRNA-seq data, had less enrichment in our podocyte TRAP gene data (6.0- to 11.1-fold) than the other eight genes, with enrichment of 16.7- to 83.8-fold (Supplementary Table 1). Grgic et al. (20) used TRAP to identify new podocyte genes in focal segmental glomerulosclerosis, and most of our podocyte TRAP-enriched genes (502 of 735) overlapped with their podocyte translational profile genes (Supplementary Table 1). These comparisons indicate that our podocyte TRAP gene set is a reliable gene set for podocyte-enriched translational profile analysis.

The 126 highly enriched actively translated genes were further analyzed with GO_BP, GO_MF, and KEGG. As indicated in Fig. 2F, genes related to glomerulus development, glomerular visceral epithelial cell development and differentiation, kidney development, and ureteric bud development were highly enriched. Genes related to axon guidance and nervous system development, actin binding, and Wnt protein binding were also included among these 126 genes (Fig. 2F and Supplementary Table 2). Genes related to actin cytoskeleton organization, response to hypoxia, the TGF-β receptor signaling pathway, regulation of cell shape, cell differentiation, and cell adhesion were also highly enriched in the 735 enriched podocyte genes (Supplementary Fig. 2 and Table 2).

Differentially Expressed Genes in Podocytes During Development of T2DN

The heat map of expression of all 735 podocyte-enriched genes, ranked by fold enrichment in both early (3 months) and advanced (6 months) type 2 diabetic podocytes, is shown in Fig. 3A and B. In general, genes with more enrichment were more likely to be downregulated, while genes with less enrichment were more likely to be upregulated, especially in the advanced stage of disease.

In early type 2 diabetes podocytes, 162 genes were significantly upregulated, and 80 of these genes were upregulated >1.5-fold, including genes related to responses to glucocorticoid action and wounding healing, protein transport, and positive regulation of apoptotic processes. At the same time, 145 genes were significantly downregulated, and 64 of these genes were downregulated >1.5-fold, including genes related to cell adhesion, hydrolase activity, transporter activity, and axon guidance (Fig. 3C and Supplementary Table 3).

In advanced type 2 diabetes podocytes, 172 genes were significantly upregulated, and 144 of these genes were upregulated >1.5-fold, including genes related to transcription, MAPK signaling pathway, R-SMAD binding, aging, actin filament organization, and positive regulation of neuron differentiation. At the same time, 168 genes were significantly downregulated, and 138 of these genes were downregulated >1.5-fold, including genes related to glomerular development, axon guidance, and cell adhesion (Fig. 3C and Supplementary Table 3).

The scatterplot in Fig. 3D indicates expression of representative podocyte genes in the early stage and advanced stages of T2DN. For example, podocyte Tspan2 was upregulated only in early T2DN, while Ngb, Rarres1, Mgat5b, Cldn5, Npr1, and Npr3 were all downregulated in both early and advanced T2DN. The heat map of the top 20 podocyte genes with >1.5-fold change in expression in early and advanced stages of T2DN is shown in Fig. 3E. The volcano plot distribution of all 735 podocyte genes with fold change and P value for the early and advanced type 2 diabetes podocytes is shown in Supplementary Figs. 3 and 4. A full list of differentially expressed genes (DEGs) in podocytes in type 2 diabetes is available in Supplementary Table 4.

There were 71 podocyte TRAP genes consistently upregulated from the early stage to the advanced stage of T2DN, including genes related to wound healing (Sparc, Sdc4, Mustn1, and Tpm1), angiogenesis (Tnfrsf12a, Emp2, and Vash1), and Ras signaling (Rras, Rras2, Gab2, and Arf6) (Fig. 3F and Supplementary Table 3). Genes involved in calcium-dependent protein binding and positive regulation of cytosolic calcium ion concentration were more enriched in upregulated genes in the early stage of T2DN, while transcription-related genes were more enriched in upregulated genes in the advanced stage of T2DN (Supplementary Table 3). At the same time, 81 podocyte genes were consistently downregulated from the early stage to the advanced stage of T2DN, including genes involved in cell adhesion (Ntng2 and Cldn5) (Fig. 3F and Supplementary Table 3). Genes related to negative regulation of angiogenesis were more enriched in the case of downregulated genes for the early stage of T2DN, while genes related to negative regulation of apoptotic processes were more enriched in the case of downregulated genes for advanced stages of T2DN (Supplementary Table 3).

DEGs in Podocytes During Development of T1DN

The heat map of expression of all 735 podocyte-enriched genes, ranked by fold enrichment in both early (3 weeks) and late (24 weeks) type 1 diabetes podocytes, is shown in Fig. 4A. Similar to what was seen in T2DN, genes with more enrichment were more likely to be downregulated, while genes with less enrichment were more likely to be upregulated, especially in the later stage of T1DN.

In early type 1 diabetes podocytes, 74 genes were significantly upregulated, and 35 of these genes were upregulated >1.5-fold, including genes related to tight-junction proteins, cellular response to insulin stimulus, and GTP binding. At the same time, 81 genes were significantly downregulated, and 32 of these genes were downregulated >1.5-fold, including genes related to serine-type endopeptidase activity (Fig. 4B and Supplementary Table 5).

In later type 1 diabetes podocytes, 73 genes were significantly upregulated, and 29 of these genes were upregulated >1.5-fold. At the same time, 62 genes were significantly downregulated, and 24 of these genes were downregulated >1.5-fold, including genes related to regulation of transcription and the cAMP signaling pathway (Fig. 4B and Supplementary Table 5).

The scatterplot in Fig. 4C shows expression of representative genes in the early and later type 1 diabetes podocytes. The heat map of the top 20 podocyte genes with >1.5-fold expression change in the early stage and the later stage of T1DN are shown in Fig. 4D. The volcano plot distribution of all 735 podocyte genes with fold change and P value for the early and later type 1 diabetes podocytes is shown in Supplementary Fig. 5 and 6. A full list of DEGs in podocytes in type 1 diabetes is available in Supplementary Table 6.

Most of the podocyte-enriched genes were only significantly changed in either the early stage (54 upregulated vs. 63 downregulated) or in the later stage (55 upregulated vs. 44 downregulated) of T1DN (Fig. 4E). For example, Mafa was only downregulated in later T1DN. However, there were 20 podocyte genes that were upregulated and 18 podocyte genes that were downregulated at both early and later stages of T1DN (Fig. 4E and Supplementary Table 5). For example, podocyte Pros1 was upregulated, while Ngb, Cldn5, Npr1, and Npr3 were downregulated, in both early and later T1DN. Pros1 was previously reported to play a role in protection against podocyte injury in type 1 diabetic mice (39).

Comparison of DEGs in Podocytes Between Type 1 and Type 2 Diabetic Mice

In comparisons with T1DN mice, more podocyte TRAP genes were differentially expressed in T2DN mice at the early stage (162 vs. 74 upregulated for T2DN and T1DN, respectively, and 145 vs. 81 downregulated) and the later stage (172 vs. 73 upregulated and 168 vs. 62 downregulated) in the 735-podocyte TRAP gene set; a clearer pattern was observed in the 126-podocyte TRAP gene set (Fig. 5A and Supplementary Table 7). Of interest, the expression of some podocyte TRAP genes only changed in either T1DN or T2DN, and a small number even changed in opposite directions in T1DN vs. T2DN. For example, both Rab13 and Stat1 are related to cellular responses to insulin stimulus. However, podocyte Rab13 was upregulated only in T1DN, while podocyte Stat1 was upregulated in T1DN but downregulated in T2DN. Podocyte ATP binding–related genes (Epha6, Ksr1, Tyro3, Ube2ql1, Speg, Nod1, Trib2, and Adcy5) and cAMP-mediated signaling-related genes (Glp1r, Ksr1, and Adcy5) were only downregulated in T1DN. R-SMAD binding–related genes (Jun, Pmepa1, Fos, and Smad6) and TGF-β signaling pathway genes (Inhbb, Smad6, Nbl1, and Bmpr1a) were upregulated, and MAPK activity–related genes (Ntrk3, Plce1, Tgfa, and Arrb1), ion transport–related genes (Clic5, Gabrb1, Steap3, Clic3, Kcnj12, Slco2a1, Itpr3, Atp1b2, and Slc38a4), and genes related to response to hydrogen peroxide (Gnao1, Nr4a3, Stat1, Stk26, and Cryab) were downregulated only in type 2 diabetic podocytes (Supplementary Table 8).

The Venn diagram in Fig. 5B shows that 82 podocyte TRAP-enriched genes were upregulated in both T1DN and T2DN. These genes included actin filament organization genes (Prkci, Srf, Actn1, Nck2, and Emp2) and positive regulation of apoptotic process genes (Unc13b, Mtch1, Gadd45b, Prmt2, Tnfrsf12a, Ankrd1, Rarb, and Cyr61). A total of 71 podocyte genes were downregulated in both T1DN and T2DN (Fig. 5B), including vascular endothelial growth factor receptor signaling pathway genes (Foxc2, Sulf1, and Vegfa), serine-type endopeptidase activity genes (Htra1, Plat, Pcsk6, and Rhbdl3), and genes important for the glomerular basement membrane (Nphs1, Wt1, Col4a3, and Sulf1) (Supplementary Table 8). The full list of the unique and shared upregulated and downregulated genes in podocytes in T2DN and T1DN can be found in Supplementary Table 9.

Validation of the Expression of Selected DEGs in Podocytes in Mouse and Human Diabetic Kidney Disease

Some podocyte actively translating genes that were highly enriched and differentially expressed in T1DN or T2DN were further validated by their localization in podocytes. Ngb was highly expressed in the 126-podocyte TRAP gene set, and its translating mRNA levels were decreased at the early and later stages of both T1DN and T2DN (Fig. 6A). Immunostaining confirmed decreased Ngb expression in eNOS−/−; db/db mice compared with control eNOS−/− mice (Fig. 6B). Immunofluorescent staining and immunostaining also confirmed that Ngb was primarily expressed in the nuclei of podocytes (colocalization with WT1) in normal wild-type mice (Fig. 6C and Supplementary Fig. 7A).

Cldn5 is another gene highly enriched and expressed in podocytes, and its translating mRNA levels were decreased at the early and later stages of both T1DN and T2DN (Supplementary Fig. 8A). Immunostaining confirmed decreased claudin 5 in eNOS−/−; db/db mice compared with control eNOS−/− mice (Supplementary Fig. 8A). Immunofluorescent staining and immunostaining determined that claudin 5 was primarily expressed in the plasma membrane of podocytes (colocalization with nephrin) in normal wild-type mice (Supplementary Figs. 7B and 8B). Translating Tspan2 mRNA levels were increased in the early stage of type 2 diabetic podocytes, and immunostaining confirmed that Tspan2 protein levels were increased in early eNOS−/−; db/db mice (Supplementary Fig. 8C).

To validate further the expression of selected DEGs in podocytes in T2DN, we isolated glomeruli from 3-month-old male eNOS−/− mice and eNOS−/−; db/db mice and performed quantitative PCR analysis. As indicated in Fig. 7A, the mRNA levels of glomerulus Wt1, Ngb, Mgat5b, Cldn5, Nr4a3, Nr4a1, and Magi2 were significantly lower in eNOS−/−; db/db mice than in eNOS−/− mice. We also performed double immunofluorescent staining to evaluate NGB and CLDN5 protein expression in podocytes in relatively normal human kidneys and T2DN. In normal human kidney, NGB was expressed in most podocytes (NGB and WT1 double-positive cells) but in only a small portion of podocytes in human T2DN (Fig. 7B). In contrast, the percentage of TSPAN2-expressing podocytes was markedly increased in human T2DN (Fig. 7C).

In the current study, we have generated podocyte-specific TRAP models of both accelerated T2DN (podocin-Cre; Rosa26fsTRAP; eNOS−/−; db/db) and accelerated T1DN (STZ-induced podocin-Cre; Rosa26fsTRAP; eNOS−/−) and report the following major findings. 1) We generated a translating mRNA gene set enriched in podocytes using a TRAP plus RNA-seq technique that was compared with other data sets. 2) We evaluated DEGs in podocytes during the development of T2DN and T1DN and found that there were many more DEGs in podocytes during the development of T2DN than during development of T1DN. 3) Genes with more enrichment were more likely to be downregulated, while genes with less enrichment were more likely to be upregulated. 4) A number of genes were differentially expressed only in either T2DN or T1DN, but genes related to actin filament organization and positive regulation of apoptotic process genes were upregulated, while genes related to vascular endothelial growth factor receptor signaling pathway, glomerular basement membrane, and serine-type endopeptidase activity were downregulated, in both T2DN and T1DN. And 5) a subset of novel candidate genes, particularly Ngb and Tspan2, were validated by their protein expression in podocytes and alteration during the development of DN in both mouse and human.

Determination of mRNA abundance of a tissue or cell only partially correlates with protein abundance (40). Although bulk RNA-seq or scRNA-seq provides invaluable information, there is not always a one-to-one correlation between levels of any single mRNA species and its protein level, with some estimates having correlations as low as R = 0.21 for some genes (41,42). In general, a better indication of the status of a cell is derived from investigation at the proteome level rather than the transcriptome level. However, it is technically difficult to investigate the proteome of many cell types. The podocyte is a case in point, since isolated glomeruli are a mixture of different cell populations: glomerular endothelial cells, mesangial cells, podocytes, parietal epithelial cells, and a mixture of immune cell types (10). The translatome is composed of all mRNAs that are translated in a single cell type at the time of study. Since the translatome is the active intermediary step between the transcriptome and the proteome, it represents a more accurate approximation for estimating the actual protein expression (16,17,43). With the development of RNA-seq techniques, ribosome profiling has become a powerful tool for globally monitoring in vivo gene expression at the translatome level (44). In addition, this approach can capture all podocyte-specific mRNA without introducing any expression alteration due to the stress of isolating the cells. Podocin-Cre is an ideal podocyte-specific Cre recombinase to be used for crossing with TRAP mice to study podocyte-specific mRNA translation. In podocin-Cre; Rosa26fsTRAP mice, GFP-L10a was consistently expressed in all the podocytes in whole kidney even when podocin expression was changed during diabetes because once the podocin gene promoter was activated in a cell, at the same time the stop cassette in the Rosa26 locus was removed forever in the genomic DNA of this cell by Cre recombinase. GFP-L10a was always expressed in this cell and its offspring cells consistently, no matter the alteration of podocin gene expression. Therefore, TRAP technology is especially valuable for studying podocyte gene expression in different disease states.

Many of the alterations in translating genes in podocytes observed in both models of DN relate to alterations of podocyte structure and function, including a number of genes that have not been previously studied in DN. One of the most interesting findings is the podocyte expression of neuroglobin (Ngb). Neuroglobin was first reported in brain as a hypoxia-inducible globin (45) and, surprisingly, was the most enriched gene in our podocyte TRAP data. There is only one other report of neuroglobin expression in podocytes, with upregulation in a model of HIF stabilization with podocyte VHL deletion (46). Although neuronal localization is most predominant in the cytoplasm and is associated with mitochondria, we found abundant expression of neuroglobin in podocytes of normal wild-type mice with colocalization with WT1, indicating expression in the podocyte nucleus. We found similar podocyte nuclear localization in normal human kidneys, and NGB expression in podocytes was decreased in human diabetic kidney. Ngb mRNA translating levels in podocytes were decreased at both early and later stages of both T2DN and T1DN, and immunostaining confirmed that its expression in podocytes was decreased in T2DN. The potential role of neuroglobin in podocyte physiology and pathophysiology warrants further investigation.

The tight-junction protein claudin 5, encoded by Cldn5, was previously reported to be highly expressed in podocytes without any alterations in expression in puromycin aminonucleoside nephrosis (47). In contrast, Cldn5 was reported to be downregulated in podocytes in a mouse model of inducible podocyte injury (NEP25) that progressively develops glomerulosclerosis after immunotoxin injection (48), and Susztak’s group reported that CLDN5 was downregulated in glomeruli from human diabetic kidneys (6). We confirmed that claudin 5 was localized to podocytes and found that podocyte-translating Cldn5 mRNA levels were decreased in both T2DN and T1DN, which was confirmed with immunohistochemical analysis. These findings suggest that further studies to determine the potential role of decreased claudin 5 expression in mediating detachment of podocytes in DN may be fruitful.

Tetraspanin 2 Tspan2, another top podocyte-enriched gene (87.5-fold enriched), was also identified as a podocyte-specific marker by Brunskill et al. (35) using FACS-sorting podocytes. In response to high glucose, Tspan2 has been reported to increase apoptosis of human pancreatic β-cells by regulating the JNK/β-catenin signaling pathway (49). In our TRAP data, Tspan2 was upregulated in early type 2 diabetes podocytes and immunostaining confirmed increased expression in type 2 diabetic podocytes in mouse as well as human diabetic kidney.

Recently, using single-cell RNA profiling of glomerular cells, He and colleagues (10) showed that Magi2 was primarily expressed in podocytes in the glomeruli of STZ eNOS−/− mice. We also determined that Magi2 was highly enriched and actively translated in podocytes and decreased in both T1DN and T2DN. Immunofluorescent staining confirmed its localization in podocytes and decreased expression in both T1DN and T2DN (Supplementary Fig. 9).

Members of the orphan nuclear receptor family Nr4a1 and Nr4a3 were among the 126 highly enriched and expressed podocyte genes. Of interest, global deletion of Nr4a1 was previously reported to be involved in mediating DN (50). In contrast, we detected decreased expression of Nr4a1 as well as Nr4a3 in advanced T2DN. Recent studies have suggested that members of this nuclear receptor family bind unsaturated fatty acids (51).

Rarres1 encodes a retinoic acid receptor responder protein. It was the third most enriched gene in podocytes in our TRAP mRNA data set. Using the Nephrotic Syndrome Study Network (NEPTUNE) consortium data set and other publicly available transcriptomic data sets, He’s group reported that podocyte RARRES1 is positively related to renal function decline in human glomerular disease. Its expression is increased in focal segmental glomerulosclerosis and DN, and it induces podocyte cell death. Podocyte-specific overexpression of RARRES1 led to glomerular injury and albuminuria, while podocyte-specific deletion of Rarres1 protected against adriamycin-induced nephropathy (52).

Recently, Koehler et al. (53) reported the proteome profiles from isolated podocytes from mouse models in the early disease states of focal segmental glomerulosclerosis after chemical induction of glomerular diseases with doxorubicin or lipopolysaccharide. We plan to repeat this experiment for our TRAP mice to directly compare the correlation between proteome profiles and translatome profile with what we see in DN.

The current study does have some limitations. Previous studies have shown that density of podocytes decreased in DN due to podocyte apoptosis, detachment, or epithelial-mesenchymal transition (EMT) potential mechanisms (5456). In our TRAP data, we did detect enrichment of positive regulation of apoptotic process genes, upregulated in both type 1 and type 2 diabetic podocytes. Although TRAP can determine changes in translating mRNA per podocyte, it does not detect changes in podocyte number. Second, although three to five mice from each group were sequenced, the sample size per group was still not large enough to completely eliminate the effect of variation among different mice. Finally, although we selected two representative time points in each type of diabetes model, studies from more time points will provide more information regarding podocyte injury in the development of DN. In addition, further studies will be necessary to focus on the functional and mechanistic roles of identified candidate genes in development or exacerbation of podocyte injury in DN.

In conclusion, podocyte-specific translated mRNA profiling provides important clues concerning mechanisms of podocyte injury in T2DN and T1DN. Our podocyte TRAP plus RNA-seq strategy allows comparison of translatome similarities and differences between type 2 diabetic podocytes and type 1 diabetic podocytes and may be valuable in identifying factors underlying the pathophysiology of DN and novel therapeutic targets to treat diabetes-induced podocyte injury.

This article contains supplementary material online at https://doi.org/10.2337/figshare.14888034.

Funding. These studies were supported by National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK), National Institutes of Health, grants DK51265, DK95785, and DK62794 (to R.C.H. and M.-Z.Z.); NIDDK grant DK103067 (R.C.H., A.B.F., K.C.V. and M.-Z.Z.); the Vanderbilt O’Brien Center (NIDDK grant P30DK114809) (R.C.H. and M.-Z.Z.); U.S. Department of Veterans Affairs VA Merit Award 00507969 (R.C.H.); and the Vanderbilt Center for Kidney Disease.

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

Author Contributions. M.-Z.Z. and R.C.H. designed the study. Y.W., A.N., Y.P., S.C., A.S.T., S.W., X.F., C.L.T., M.A.R.S., D.L.M., D.C., R.M.A., W.Z., Q.S., and K.C.V. performed the experiments. Y.W., A.B.F., and M.-Z.Z. participated in figure preparation. Y.W., M.-Z.Z., and R.C.H. wrote the manuscript. All authors read and approved the final version of manuscript. R.C.H. and M.-Z.Z. are the guarantors of this work and, as such, had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

1.
Duarte
AA
,
Mohsin
S
,
Golubnitschaja
O
.
Diabetes care in figures: current pitfalls and future scenario
.
EPMA J
2018
;
9
:
125
131
2.
Dronavalli
S
,
Duka
I
,
Bakris
GL
.
The pathogenesis of diabetic nephropathy
.
Nat Clin Pract Endocrinol Metab
2008
;
4
:
444
452
3.
Li
JJ
,
Kwak
SJ
,
Jung
DS
, et al
.
Podocyte biology in diabetic nephropathy
.
Kidney Int Suppl
2007
106
:
S36
S42
4.
Reddy
GR
,
Kotlyarevska
K
,
Ransom
RF
,
Menon
RK
.
The podocyte and diabetes mellitus: is the podocyte the key to the origins of diabetic nephropathy?
Curr Opin Nephrol Hypertens
2008
;
17
:
32
36
5.
Hodgin
JB
,
Nair
V
,
Zhang
H
, et al
.
Identification of cross-species shared transcriptional networks of diabetic nephropathy in human and mouse glomeruli
.
Diabetes
2013
;
62
:
299
308
6.
Woroniecka
KI
,
Park
AS
,
Mohtat
D
,
Thomas
DB
,
Pullman
JM
,
Susztak
K
.
Transcriptome analysis of human diabetic kidney disease
.
Diabetes
2011
;
60
:
2354
2369
7.
Rubin
A
,
Salzberg
AC
,
Imamura
Y
,
Grivitishvilli
A
,
Tombran-Tink
J
.
Identification of novel targets of diabetic nephropathy and PEDF peptide treatment using RNA-seq
.
BMC Genomics
2016
;
17
:
936
8.
Fu
J
,
Wei
C
,
Zhang
W
, et al
.
Gene expression profiles of glomerular endothelial cells support their role in the glomerulopathy of diabetic mice
.
Kidney Int
2018
;
94
:
326
345
9.
Fu
J
,
Wei
C
,
Lee
K
, et al
.
Comparison of glomerular and podocyte mRNA profiles in streptozotocin-induced diabetes
.
J Am Soc Nephrol
2016
;
27
:
1006
1014
10.
Fu
J
,
Akat
KM
,
Sun
Z
, et al
.
Single-cell RNA profiling of glomerular cells shows dynamic changes in experimental diabetic kidney disease
.
J Am Soc Nephrol
2019
;
30
:
533
545
11.
Karaiskos
N
,
Rahmatollahi
M
,
Boltengagen
A
, et al
.
A single-cell transcriptome atlas of the mouse glomerulus
.
J Am Soc Nephrol
2018
;
29
:
2060
2068
12.
Lu
Y
,
Ye
Y
,
Yang
Q
,
Shi
S
.
Single-cell RNA-sequence analysis of mouse glomerular mesangial cells uncovers mesangial cell essential genes
.
Kidney Int
2017
;
92
:
504
513
13.
Wu
H
,
Kirita
Y
,
Donnelly
EL
,
Humphreys
BD
.
Advantages of single-nucleus over single-cell RNA sequencing of adult kidney: rare cell types and novel cell states revealed in fibrosis
.
J Am Soc Nephrol
2019
;
30
:
23
32
14.
Doyle
JP
,
Dougherty
JD
,
Heiman
M
, et al
.
Application of a translational profiling approach for the comparative analysis of CNS cell types
.
Cell
2008
;
135
:
749
762
15.
Heiman
M
,
Schaefer
A
,
Gong
S
, et al
.
A translational profiling approach for the molecular characterization of CNS cell types
.
Cell
2008
;
135
:
738
748
16.
Ingolia
NT
,
Ghaemmaghami
S
,
Newman
JR
,
Weissman
JS
.
Genome-wide analysis in vivo of translation with nucleotide resolution using ribosome profiling
.
Science
2009
;
324
:
218
223
17.
King
HA
,
Gerber
AP
.
Translatome profiling: methods for genome-scale analysis of mRNA translation
.
Brief Funct Genomics
2016
;
15
:
22
31
18.
Liu
J
,
Krautzberger
AM
,
Sui
SH
, et al
.
Cell-specific translational profiling in acute kidney injury
.
J Clin Invest
2014
;
124
:
1242
1254
19.
Grgic
I
,
Krautzberger
AM
,
Hofmeister
A
, et al
.
Translational profiles of medullary myofibroblasts during kidney fibrosis
.
J Am Soc Nephrol
2014
;
25
:
1979
1990
20.
Grgic
I
,
Hofmeister
AF
,
Genovese
G
, et al
.
Discovery of new glomerular disease-relevant genes by translational profiling of podocytes in vivo
.
Kidney Int
2014
;
86
:
1116
1129
21.
Hofmeister
A
,
Thomaßen
MC
,
Markert
S
, et al
.
Development of a new macrophage-specific TRAP mouse (MacTRAP) and definition of the renal macrophage translational signature
.
Sci Rep
2020
;
10
:
7519
22.
Zhou
P
,
Zhang
Y
,
Ma
Q
, et al
.
Interrogating translational efficiency and lineage-specific transcriptomes using ribosome affinity purification
.
Proc Natl Acad Sci U S A
2013
;
110
:
15395
15400
23.
Moeller
MJ
,
Sanden
SK
,
Soofi
A
,
Wiggins
RC
,
Holzman
LB
.
Podocyte-specific expression of cre recombinase in transgenic mice
.
Genesis
2003
;
35
:
39
42
24.
Heiman
M
,
Kulicke
R
,
Fenster
RJ
,
Greengard
P
,
Heintz
N
.
Cell type-specific mRNA purification by translating ribosome affinity purification (TRAP)
.
Nat Protoc
2014
;
9
:
1282
1291
25.
Dobin
A
,
Davis
CA
,
Schlesinger
F
, et al
.
STAR: ultrafast universal RNA-seq aligner
.
Bioinformatics
2013
;
29
:
15
21
26.
Liao
Y
,
Smyth
GK
,
Shi
W
.
featureCounts: an efficient general purpose program for assigning sequence reads to genomic features
.
Bioinformatics
2014
;
30
:
923
930
27.
Love
MI
,
Huber
W
,
Anders
S
.
Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2
.
Genome Biol
2014
;
15
:
550
28.
Huang
W
,
Sherman
BT
,
Lempicki
RA
.
Systematic and integrative analysis of large gene lists using DAVID bioinformatics resources
.
Nat Protoc
2009
;
4
:
44
57
29.
Huang
W
,
Sherman
BT
,
Lempicki
RA
.
Bioinformatics enrichment tools: paths toward the comprehensive functional analysis of large gene lists
.
Nucleic Acids Res
2009
;
37
:
1
13
30.
Zhang
MZ
,
Yao
B
,
Wang
Y
, et al
.
Inhibition of cyclooxygenase-2 in hematopoietic cells results in salt-sensitive hypertension
.
J Clin Invest
2015
;
125
:
4281
4294
31.
Li
Y
,
Pan
Y
,
Cao
S
, et al
.
Podocyte EGFR inhibits autophagy through upregulation of rubicon in type 2 diabetic nephropathy
.
Diabetes
2021
;
70
:
562
576
32.
Zhao
HJ
,
Wang
S
,
Cheng
H
, et al
.
Endothelial nitric oxide synthase deficiency produces accelerated nephropathy in diabetic mice
.
J Am Soc Nephrol
2006
;
17
:
2664
2669
33.
Yuen
DA
,
Stead
BE
,
Zhang
Y
, et al
.
eNOS deficiency predisposes podocytes to injury in diabetes
.
J Am Soc Nephrol
2012
;
23
:
1810
1823
34.
Zhang
MZ
,
Wang
Y
,
Paueksakon
P
,
Harris
RC
.
Epidermal growth factor receptor inhibition slows progression of diabetic nephropathy in association with a decrease in endoplasmic reticulum stress and an increase in autophagy
.
Diabetes
2014
;
63
:
2063
2072
35.
Brunskill
EW
,
Georgas
K
,
Rumballe
B
,
Little
MH
,
Potter
SS
.
Defining the molecular character of the developing and adult kidney podocyte
.
PLoS One
2011
;
6
:
e24640
36.
Saibil
H
.
Chaperone machines for protein folding, unfolding and disaggregation
.
Nat Rev Mol Cell Biol
2013
;
14
:
630
642
37.
Boerries
M
,
Grahammer
F
,
Eiselein
S
, et al
.
Molecular fingerprinting of the podocyte reveals novel gene and protein regulatory networks
.
Kidney Int
2013
;
83
:
1052
1064
38.
Park
J
,
Shrestha
R
,
Qiu
C
, et al
.
Single-cell transcriptomics of the mouse kidney reveals potential cellular targets of kidney disease
.
Science
2018
;
360
:
758
763
39.
Zhong
F
,
Chen
H
,
Xie
Y
, et al
.
Protein S protects against podocyte injury in diabetic nephropathy
.
J Am Soc Nephrol
2018
;
29
:
1397
1410
40.
Vogel
C
,
Marcotte
EM
.
Insights into the regulation of protein abundance from proteomic and transcriptomic analyses
.
Nat Rev Genet
2012
;
13
:
227
232
41.
Fortelny
N
,
Overall
CM
,
Pavlidis
P
,
Freue
GVC
.
Can we predict protein from mRNA levels?
Nature
2017
;
547
:
E19
E20
42.
Liu
Y
,
Beyer
A
,
Aebersold
R
.
On the dependency of cellular protein levels on mRNA abundance
.
Cell
2016
;
165
:
535
550
43.
Smircich
P
,
Eastman
G
,
Bispo
S
, et al
.
Ribosome profiling reveals translation control as a key mechanism generating differential gene expression in Trypanosoma cruzi
.
BMC Genomics
2015
;
16
:
443
44.
Brar
GA
,
Weissman
JS
.
Ribosome profiling reveals the what, when, where and how of protein synthesis
.
Nat Rev Mol Cell Biol
2015
;
16
:
651
664
45.
Burmester
T
,
Weich
B
,
Reinhardt
S
,
Hankeln
T
.
A vertebrate globin expressed in the brain
.
Nature
2000
;
407
:
520
523
46.
Steenhard
BM
,
Isom
K
,
Stroganova
L
, et al
.
Deletion of von Hippel-Lindau in glomerular podocytes results in glomerular basement membrane thickening, ectopic subepithelial deposition of collagen alpha1alpha2alpha1(IV), expression of neuroglobin, and proteinuria
.
Am J Pathol
2010
;
177
:
84
96
47.
Koda
R
,
Zhao
L
,
Yaoita
E
, et al
.
Novel expression of claudin-5 in glomerular podocytes
.
Cell Tissue Res
2011
;
343
:
637
648
48.
Okabe
M
,
Motojima
M
,
Miyazaki
Y
,
Pastan
I
,
Yokoo
T
,
Matsusaka
T
.
Global polysome analysis of normal and injured podocytes
.
Am J Physiol Renal Physiol
2019
;
316
:
F241
F252
49.
Hwang
IH
,
Park
J
,
Kim
JM
, et al
.
Tetraspanin-2 promotes glucotoxic apoptosis by regulating the JNK/β-catenin signaling pathway in human pancreatic β cells
.
FASEB J
2016
;
30
:
3107
3116
50.
Sheng
J
,
Li
H
,
Dai
Q
, et al
.
NR4A1 promotes diabetic nephropathy by activating Mff-mediated mitochondrial fission and suppressing Parkin-mediated mitophagy
.
Cell Physiol Biochem
2018
;
48
:
1675
1693
51.
Li
XX
,
Wang
ZJ
,
Zheng
Y
, et al
.
Nuclear receptor Nur77 facilitates melanoma cell survival under metabolic stress by protecting fatty acid oxidation
.
Mol Cell
2018
;
69
:
480
492.e7
52.
Chen
A
,
Feng
Y
,
Lai
H
, et al
.
Soluble RARRES1 induces podocyte apoptosis to promote glomerular disease progression
.
J Clin Invest
2020
;
130
:
5523
5535
53.
Koehler
S
,
Kuczkowski
A
,
Kuehne
L
, et al
.
Proteome analysis of isolated podocytes reveals stress responses in glomerular sclerosis
.
J Am Soc Nephrol
2020
;
31
:
544
559
54.
Susztak
K
,
Raff
AC
,
Schiffer
M
,
Böttinger
EP
.
Glucose-induced reactive oxygen species cause apoptosis of podocytes and podocyte depletion at the onset of diabetic nephropathy
.
Diabetes
2006
;
55
:
225
233
55.
Petermann
AT
,
Pippin
J
,
Krofft
R
, et al
.
Viable podocytes detach in experimental diabetic nephropathy: potential mechanism underlying glomerulo sclerosis
.
Nephron Exp Nephrol
2004
;
98
:
e114
e123
56.
Kang
YS
,
Li
Y
,
Dai
C
,
Kiss
LP
,
Wu
C
,
Liu
Y
.
Inhibition of integrin-linked kinase blocks podocyte epithelial-mesenchymal transition and ameliorates proteinuria
.
Kidney Int
2010
;
78
:
363
373
Readers may use this article as long as the work is properly cited, the use is educational and not for profit, and the work is not altered. More information is available at https://www.diabetesjournals.org/content/license.