Inhibition of p38 MAPK decreases hyperglycemia-induced nephrin endocytosis and attenuates albuminuria

Abstract

Chronic hyperglycemia, as in diabetes mellitus, may cause glomerular damage with microalbuminuria as an early sign. Noteworthy, even acute hyperglycemia can increase glomerular permeability before structural damage of the glomerular filter can be detected. Despite intensive research, specific antiproteinuric therapy is not available so far. Thus, a deeper understanding of the molecular mechanisms of albuminuria is desirable. P38 MAPK signaling is involved in the development of hyperglycemia-induced albuminuria. However, the mechanism of increased p38 MAPK activity leading to increased permeability and albuminuria remained unclear. Recently, we demonstrated that acute hyperglycemia triggers endocytosis of nephrin, the key molecule of the slit diaphragm, and induces albuminuria. Here, we identify p38 MAPK as a pivotal regulator of hyperglycemia-induced nephrin endocytosis. Activated p38 MAPK phosphorylates the nephrin c-terminus at serine 1146, facilitating the interaction of PKCα with nephrin. PKCα phosphorylates nephrin at threonine residues 1120 and 1125, mediating the binding of β-arrestin2 to nephrin. β-arrestin2 triggers endocytosis of nephrin by coupling it to the endocytic machinery, leading to increased glomerular permeability. Pharmacological inhibition of p38 MAPK preserves nephrin surface expression and significantly attenuates albuminuria. KEY MESSAGES: Acute hyperglycemia triggers endocytosis of nephrin. Activated p38 MAPK phosphorylates the nephrin c-terminus at serine 1146, facilitating the interaction of PKCα with nephrin. PKCα phosphorylates nephrin at threonine residues 1120 and 1125, mediating the binding of β-arrestin2 to nephrin. β-arrestin2 triggers endocytosis of nephrin by coupling it to the endocytic machinery, leading to a leaky glomerular filter. Pharmacological inhibition of p38 MAPK preserves nephrin surface expression and significantly attenuates albuminuria under hyperglycemic conditions.

Keywords: Albuminuria; Diabetes; Endocytosis; Nephrin; Podocyte.

Fig. 1

Hyperglycemia induces phosphorylation of p38 MAPK in glomeruli and podocytes. a Western blot analysis of phospho-p38 (WB: p-p38) and total p38 (WB: p38) expression in glomerular lysates of normoglycemic (control) and hyperglycemic (STZ) mice (day 5). β-actin (WB: actin) served as a loading control. The results of five independent experiments were quantified by densitometry and graphed as phospho-p38 (p-p38) to total p38 (p38) ratio – the control group was normalized to 100%. Control (n = 5) vs. STZ (n = 6): 100.0 ± 10.8 vs. 181.5 ± 11.2% (***p < 0.001); b Western blot analysis of phospho-p38 and total p38 expression in immortalized murine podocytes under normoglycemic (5.5 mM) and high (30 mM) glucose conditions at the given time points. The results of five independent experiments were quantified by densitometry graphed as phospho-p38 (p-p38) to total p38 (p38) ratio – normoglycemic (5.5 mM) group was normalized to 100%. A total of 5.5 mM glucose (n = 6) vs. 30 mM glucose 1 h (n = 6)/30 mM glucose 2 h (n = 6): 100.0 ± 9.9 vs. 147.3 ± 10.3% / 187.6 ± 18.5% (**p < 0.01 vs. control)

Fig. 2

Inhibition of p38 MAPK does not alter development of hyperglycemia but prevents hyperglycemia-induced albuminuria. a Blood glucose levels of normoglycemic (control), hyperglycemic (STZ), and hyperglycemic mice treated with p38 MAPK inhibitor (STZ + SB202190) at day 1 and day 5. Glucose levels day 5: control (n = 11) vs. STZ (n = 14): 161.0 ± 5.6 vs. 341.6 ± 25.3 mg/dL (****p < 0.0001); STZ (n = 14) vs. STZ + SB202190 (n = 16): 341.6 ± 25.3 vs. 302.1 ± 18.8 mg/dl (p = ns). b Molecular mass markers are indicated in kilodalton (kDa), SDS-PAGE/Coomassie gel staining of urine (day 5) in normoglycemic (control), hyperglycemic (STZ), and hyperglycemic mice treated with p38 MAPK inhibitor (STZ + SB202190). BSA at 1, 5, and 10 µg/µL served both as a control and standard. c Quantitative analysis of the urinary albumin to creatinine ratio (UACR) (day 5) in normoglycemic (control), hyperglycemic (STZ), and hyperglycemic mice treated with p38 MAPK inhibitor (STZ + SB202190). UACR day 5: control (n = 18) vs. STZ (n = 15) 15.6 ± 3.0 vs. 152.9 ± 14.3 mg/g (****p < 0.0001); STZ (n = 15) vs. STZ + SB202190 (n = 8): 152.9 ± 14.3 vs. 53.5 ± 4.7 mg/g (****p < 0.0001)

Fig. 3

P38 MAPK interacts with nephrin and phosphorylates nephrin (molecular mass markers are indicated in kilodalton (kDa)). a Coimmunoprecipitation. HEK293T cells were transiently transfected with either vector or nephrin. Cells were either kept under normal (5.5 mM) or high (30 mM) glucose for 24 h. Nephrin was immobilized, and interaction with p38 MAPK was determined by p38-antibody staining. Compared to normal glucose conditions (5.5 mM), high glucose (30 mM) showed a significant increase in nephrin-p38 interaction (interaction was normalized to 100% for normal glucose conditions (5.5 mM)) ratio of IP nephrin: p38 / lysate: p38 [%]: 5.5 mM vs. 30 mM (n = 3): 100.0 ± 11.9 vs. 169.9 ± 7.3% (p < 0.05). b In vitro phosphorylation assay. Aliquots of recombinant wild-type nephrin cytoplasmic domain (aa 1087–1241) and controls (GST.ATF2, GST.β-arrestin2, and GST) were expressed in E. coli. Subsequently, recombinant p38 MAPK and [γ-32P] were added. GST.ATF2 was used as the positive control. Phosphorylation was visualized by autoradiography. Coomassie staining of an SDS-PAGE gel showed equal protein input. c, d In vitro phosphorylation assay of different truncated nephrin cytoplasmic domains. Aliquots of recombinant wild-type nephrin cytoplasmic domain (amino acids (aa) 1087–1241) and different truncated nephrin cytoplasmic domains were expressed in E. coli. Following, recombinant p38 MAPK and [γ-32P] were added. Phosphorylation was visualized by autoradiography. Coomassie staining of an SDS-PAGE gel showed equal protein input. e Graphical summary of truncated proteins and phosphorylation status. f In vitro phosphorylation assay of nephrin cytoplasmic domain and mutated nephrin cytoplasmic domain S1146A. Aliquots of recombinant wild-type nephrin cytoplasmic domain (GST.nephrin; aa 1087–1241) and mutated nephrin cytoplasmic domain (GST.nephrinS1146A) and GST were expressed in E. coli. Subsequently, recombinant p38 MAPK and [γ-32P] were added. Phosphorylation was visualized by autoradiography. Coomassie staining of an SDS-PAGE gel showed equal protein input. Western blot analysis of phosphorylation of nephrin serine 1146 and p38 MAPK with phospho-specific antibodies. Staining of total nephrin and GST served as loading controls

Fig. 4

Phosphorylation of nephrin serine 1146 in murine podocytes and nephrin serine 1146 and threonine 1120/1125 in diabetic mice (molecular mass markers are indicated in kilodalton (kDa)). a Western blot analysis of phosphorylation of nephrin serine 1146 in immortalized murine podocytes under low (5.5 mM) and high glucose (30 mM) conditions at given time points (n = 4 each). The results were quantified by densitometry and graphed as phospho-nephrin (p-nephrin) to total nephrin (nephrin) ratio–normoglycemic (5.5 mM) group was normalized to 100%. A total of 5.5 mM glucose vs. 30 mM glucose 4 h: 100.0 ± 8.4 vs. 216.4 ± 26.4% (*p < 0.05). b Western blot analysis of phosphorylation of nephrin serine 1146 and threonine 1120/1125 in nondiabetic (control) and diabetic (STZ) mice (day 5)–control was normalized to 100%: S1146 control (n = 3) vs. STZ (n = 4): 100.0 ± 6.0 vs. 162.2 ± 13.8% (*p < 0.05); T1120/1125 control (n = 3) vs. STZ (n = 4): 100.0 ± 2.9 vs. 191.9 ± 3.2% (****p < 0.0001)

Fig. 5

Interaction of PKCα with wild-type nephrin or mutated nephrin S1146A and biotinylation assay. a Western blot analysis of coimmunoprecipitation: HEK293T cells overexpressing untagged PKCα (PKCα), flag-tagged PICK1, and wild-type nephrin or mutated nephrin S1146A. Nephrin was immobilized, and the level of interaction was determined by staining of PKCα and PICK1. Staining of lysates of PCKα, PICK1, and nephrin served as loading controls. The results were quantified by densitometry and graphed as either the ratio of the PICK1 or PKCα coimmunoprecipitation signal intensity (IP nephrin: PICK1/lysate PICK1 or IP nephrin: PKCα/ lysate PKCα) to the lysate signal intensity (ratio lysate: PICK1 or lysate PKCα) – nephrin was normalized to 100%. %PKCα: nephrin (n = 6) vs. nephrin S1146A (n = 6): 100.0 ± 6.9 vs. 35.0 ± 2.9% (****p < 0.0001). %PICK: nephrin (n = 5) vs. nephrin S1146A (n = 5): 100.0 ± 8.8 vs. 27.7 ± 3.9% (***p < 0.001). b Biotinylation assay in HEK293T under normal (5.5 mM) and high (30 mM) glucose conditions. Immunoprecipitation of nephrin and the nephrin mutant S1146A (A). The biotinylated fraction of nephrin and its mutant was analyzed (WB: streptavidin). Staining of nephrin in the lysate and immunoprecipitation (WB nephrin) was performed. Actin (WB: actin) in the lysate served as loading controls. The results were quantified by densitometry and graphed as the ratio of biotinylated signal intensity to total nephrin signal intensity – 5.5 mM glucose group was normalized to 100%. A total of 5.5 mM glucose nephrin WT (n = 5) vs. 30 mM glucose nephrin WT (n = 5): 100.0 ± 9.3 vs. 57.2 ± 2.5% (*p < 0.05). A total of 5.5 mM glucose nephrin S1146A (n = 5) vs. 30 mM glucose nephrin S1146A (n = 5): 100.0 ± 9.4 vs. 93.6 ± 14.2% (p = ns)

Fig. 6

Inhibition of p38 MAPK decreases β-arrestin2 mediated nephrin endocytosis in vitro and in vivo. a Biotinylation assay in HEK293T under normal (5.5 mM) and high (30 mM) glucose conditions. Immunoprecipitation of nephrin was performed, and the biotinylated fraction of nephrin was analyzed (WB: streptavidin). Staining of nephrin in the lysate and immunoprecipitation (WB nephrin) and actin (WB: actin) in the lysate served as loading controls. The results were quantified by densitometry and graphed as the ratio of biotinylated signal intensity to total nephrin signal intensity – 5.5 mM glucose group was normalized to 100%. A total of 30 mM glucose (n = 3) vs. 30 mM glucose + SB202190 (n = 3): 62.4 ± 5.0 vs. 109.3 ± 5.2% (*p < 0.05). A total of 5.5 mM glucose (n = 3) vs. 30 mM glucose + SB202190 (n = 3):100.0 ± 14.1 vs. 109.3 ± 5.2% (p = ns). Data represent means ± SEM. Statistical analysis: unpaired t-test with Welch’s correction. b Representative immunofluorescence staining of murine kidney sections of normoglycemic (control), hyperglycemic (STZ), and hyperglycemic mice treated with p38 MAPK inhibitor (STZ + SB202190) (day 5, n = 3). Staining was performed with an anti-nephrin antibody (red) and nuclear DNA with DAPI (blue). c Representative immunofluorescence images of colocalization of nephrin with early endosomal antigen (EEA1) in immunofluorescence staining of murine kidney sections of healthy mice (control), untreated hyperglycemic mice (STZ), and hyperglycemic mice treated with p38 MAPK inhibitor (STZ + SB202190) (day 5, n = 3). Staining was performed with an anti-nephrin antibody (red), an anti-EEA1-antibody (green), and nuclear DNA with DRAQ5 (blue). White arrows indicate colocalization of nephrin with EEA1-positive vesicles. d Inhibition of p38 MAPK decreases β-arrestin2 mediated nephrin endocytosis in diabetic mice in a biotinylation assay of murine glomerular lysates of non-diabetic (control), diabetic (STZ), and diabetic mice treated with p38 MAPK inhibitor (STZ + SB202190) (day 5). Immunoprecipitation of nephrin was performed, and the biotinylated fraction of nephrin was analyzed (WB: streptavidin). Staining of nephrin in the lysate and immunoprecipitation (WB: nephrin) and actin (WB: actin) in the lysate served as loading controls. The results were quantified by densitometry and graphed as the ratio of biotinylated signal intensity to total nephrin signal intensity – the control group was normalized to 100%. Control (n = 4) vs. STZ (n = 4): 100.0 ± 0.7 vs. 74.2 ± 3.2% (**p < 0.01). STZ (n = 4) vs. STZ + SB202190 (n = 4): 74.2 ± 3.2 vs. 88.0 ± 4.5% (*p < 0.05). Control (n = 4) vs. STZ + SB202190 (n = 4): 100.0 ± 0.7 vs. 88.0 ± 4.5% (p = NS). Data represent means ± SEM. Statistical analysis: unpaired t-test with Welch’s correction

Fig. 7

Molecular mechanism of hyperglycemia-induced nephrin endocytosis. High glucose levels activate p38 MAPK. p38 MAPK then phosphorylates nephrin at S1146, facilitating the interaction of PKCα with nephrin. PKCα phosphorylates nephrin at Thr1120/Thr1125, creating a β-arrestin2 biding site. β-arrestin2 couples nephrin to the endocytic machinery and triggers its internalization