Alteration of forkhead box O (foxo4) acetylation mediates apoptosis of podocytes in diabetes mellitus

Abstract

The number of kidney podocytes is reduced in diabetic nephropathy. Advanced glycation end products (AGEs) accumulate in patients with diabetes and promote the apoptosis of podocyte by activating the forkhead box O4 (Foxo4) transcription factor to increase the expression of a pro-apoptosis gene, Bcl2l11. Using chromatin immunoprecipitation we demonstrate that AGE-modified bovine serum albumin (AGE-BSA) enhances Foxo4 binding to a forkhead binding element in the promoter of Bcl2lll. AGE-BSA also increases the acetylation of Foxo4. Lysine acetylation of Foxo4 is required for Foxo4 binding and transcription of Bcl2l11 in podocytes treated with AGE-BSA. The expression of a protein deacetylase that targets Foxo4 for deacetylation, sirtuin (Sirt1), is down regulated in cultured podocytes by AGE-BSA treatment and in glomeruli of diabetic patients. SIRT1 over expression in cultured murine podocytes prevents AGE-induced apoptosis. Glomeruli isolated from diabetic db/db mice have increased acetylation of Foxo4, suppressed expression of Sirt1, and increased expression of Bcl2l11 compared to non-diabetic littermates. Together, our data provide evidence that alteration of Foxo4 acetylation and down regulation of Sirt1 expression in diabetes promote podocyte apoptosis. Strategies to preserve Sirt1 expression or reduce Foxo4 acetylation could be used to prevent podocyte loss in diabetes.

Figure 1. Binding of Foxo4 to the Bcl2l11 promoter.

Conditionally immortalized podocytes were treated with either AGE-BSA (100 µg/ml) or BSA (100 µg/ml) as control and 30 mM of hyperglycemia (HG) or 30 mM of mannitol (Man) as osmotic control for 2 h. Binding of Foxo4 to a predicted forkhead binding elementer (FBE) in the promoter region of Bcl2l11 was characterized by chromatin immunoprecipiation (ChIP) using either an anti-Foxo4 antibody (Foxo4) or normal goal IgG (IgG). (A) Binding was assessed using primers flanking FBE (FBE_for and FBE_rev) and immunoprecipitated DNA samples as PCR templates. (B) Primers flanking a region −2241 to −2043 upstream of the translational start site were used in a PCR reaction with Foxo4 immunoprecipitated DNA as templates to assess non-specific immunoprecipitation. Representative images of products of endpoint PCR resolved on agarose gels are shown. (C) Data of quantitative realtime PCR of ChIP DNA using primers flanking the FBE. Results from 3 independent sets of ChIPs. * p<0.05.

Figure 2. Acetylation of Foxo4 is required for AGE-BSA-induced Bcl2l11 acetylation.

(A) Acetylation of endogenous Foxo4 in cultured podocytes treated with 500 µM of H₂O₂, 100 µg/ml of AGE-BSA or 100 µg/ml of BSA for 2 h was assessed by gel electrophoresis of immunoprecipitating Foxo4, followed by immunoblotted with an anti-acetyl-lysine antibody to detect acetylated Foxo4. Immunoblotting with an anti-Foxo4 antibody confirms the immunoprecipitation of Foxo4. (B) H₂O₂- and AGE-BSA-induced acetylation of a FLAG-Foxo4 fusion protein was confirmed using a fluorescence-based in vitro acetylation assay. 293 cells transfected with a FLAG-Foxo4 fusion construct and serum starved for 16–18 hrs were treated with either H₂O₂ (500 µM) or AGE-BSA (50 µg/ml) at 0, 0.5, 1, 2, 4, or 16 hr. (Relative Foxo4 acetylation with AGE-BSA at 1 h is 1.36±0.12 and at 2 h is 1.32±0.11, *p<0.05, n = 4; with 500 µM of H₂O₂ at 2 hr is 2.80±0.22 *p<0.05, n = 3). The expression of Bcl2l11 in podocytes with stable over expression of the Foxo4 KA mutant (Foxo4 KA Mut) was compared to control podocytes (BB) after AGE-BSA (100 µg/ml) treatment for either 24 h (in RT PCR experiments) or 72 h (for Western Blotting experiments). Foxo4 KA Mut over expression abrogated AGE-BSA induced Bcl2l11 expression as characterized by Western blotting (C) and RT PCR (D). Binding of Foxo4 to the FBE in the Bcl2l11 promoter was reduced in podocytes with Foxo4 KA Mut over expression as characterized by ChIP of Foxo4 followed by RTPCR using primers flanking FBE (FBE_for and FBE_rev).

Figure 3. Levels of Sirt1 mRNA and protein.

A: Total mRNA extracted from podocytes treated with AGE-BSA (10 µg/ml, 50 µg/ml, and 100 µg/ml) or BSA (100 µg/ml) for 4 h was used to quantify the Sirt1 mRNA level by realtime PCR. Sirt1 expression was significantly reduced with 50 µg/ml and 100 µg/ml of AGE-BSA compared to 100 µg/ml of BSA. (n = 3, * p<0.05) B: Sirt1 is reduced by AGE-BSA (50 and 100 µg/ml) compared to BSA (50 and 100 µg/ml) treatment as assessed by Western blotting. A representative blot is shown here.

Figure 4. Sirt1 immunostaining.

Archival renal biopsy samples from patients with no apparent renal disease (normal), minimal change disease (MCD), mild diabetic nephropathy (mild DN) and nodular DN were used for Sirt1 immunostaining. Representative staining are shown for Normal (A, B, C) minimal change disease (D, E, and F), mild DN (G and H) and nodular DN (I). Normal rabbit IgG labeling of MCD (J) displayed minimal nonspecific staining. A higher magnification view (600×) of cells near the periphery of a glomerulus shows the nuclear pattern of Sirt1 staining (arrows in C and F).

Figure 5. Sirt1 overexpression and knockdown on podocyte apoptosis.

The expression of Sirt1 in murine podocyte without infection (no infection), infected with the empty lentivirus VVE/BBW (BB), or infected with the VVE/BBW-Sirt1 lentivirus with stable overexpression of Sirt1 (Sirt1) was confirmed by Western blotting (A). Apoptosis of BB (□) and SIRT (▪) podocytes treated with BSA (50 µg/ml) or AGE-BSA (50 and 100 µg/ml) for 16 h were quantified by flow cytometry after annexin-V-FITC labeling (B). *p<0.05, n = 3. Murine podocytes tranduced with either a scrambled shRNA or Sirt1-targeting shRNA were treated with either 30 mM of mannitol (MAN) or glucose (HG). Knockdown of Sirt1 expression was confirmed by Western blotting (C). Fold change in active caspase compared to scrambled shRNA podocytes treated with MAN is shown. *p<0.05, n = 3.

Figure 6. Acetylation of Foxo4 and expression of Sirt1 and Bcl2l11 in isolated glomeruli of db/db and db/m mice.

A representative Western blot of glomerular protein lysates from db/db and db/m mice at 7–8weeks of age were immunoprecipitated with an anti-Foxo4 antibody and immunoblotted with an anti-acetyl-lysine antibody (A). Immunoblotting for total Foxo4 was used to normalize for sample loading and IP efficiency. The ratio of band density for acetyl-lysine Foxo4 and total Foxo4 (acetylated/total Foxo4) is presented as relative Foxo4 acetylation (B). Relative band intensity is significantly higher in db/db than db/m (1.811±0.143 vs 1.123±0.04, n = 4 per group, p<0.05). Glomeruli were isolated from diabetic db/db and non-diabetic db/m mice at 11–12 weeks of age. mRNA expression of Sirt1 (C) and Bcl2l11 (D) were quantified by real time PCR. Data is presented as fold change in gene expression relative to db/m. (n = 7 in db/m group; n = 5 in db/db group). *p<0.05.