Nuclear Factor Erythroid 2-Related Factor 2 Drives Podocyte-Specific Expression of Peroxisome Proliferator-Activated Receptor γ Essential for Resistance to Crescentic GN

Abstract

Necrotizing and crescentic rapidly progressive GN (RPGN) is a life-threatening syndrome characterized by a rapid loss of renal function. Evidence suggests that podocyte expression of the transcription factor peroxisome proliferator-activated receptor γ (PPARγ) may prevent podocyte injury, but the function of glomerular PPARγ in acute, severe inflammatory GN is unknown. Here, we observed marked loss of PPARγ abundance and transcriptional activity in glomerular podocytes in experimental RPGN. Blunted expression of PPARγ in podocyte nuclei was also found in kidneys from patients diagnosed with crescentic GN. Podocyte-specific Pparγ gene targeting accentuated glomerular damage, with increased urinary loss of albumin and severe kidney failure. Furthermore, a PPARγ gain-of-function approach achieved by systemic administration of thiazolidinedione (TZD) failed to prevent severe RPGN in mice with podocyte-specific Pparγ gene deficiency. In nuclear factor erythroid 2-related factor 2 (NRF2)-deficient mice, loss of podocyte PPARγ was observed at baseline. NRF2 deficiency markedly aggravated the course of RPGN, an effect that was partially prevented by TZD administration. Furthermore, delayed administration of TZD, initiated after the onset of RPGN, still alleviated the severity of experimental RPGN. These findings establish a requirement for the NRF2-PPARγ cascade in podocytes, and we suggest that these transcription factors have a role in augmenting the tolerance of glomeruli to severe immune-complex mediated injury. The NRF2-PPARγ pathway may be a therapeutic target for RPGN.

Keywords: focal segmental glomerulosclerosis; glomerular disease; glomerulonephritis; metabolism; podocyte; renal protection.

Figure 1.

Glomerular PPARγ expression is decreased in NTN. (A) Western blot analysis of PPARγ expression in glomeruli isolated from NTS-challenged mice (NTS) and non-injected mice (control). Tubulin is used as loading control. (B) Quantification of Western blot bands for PPARγ normalized to tubulin band intensity (means of six mice per group, of two independent experiments). (C) PPARγ activity was determined by RT-PCR analysis of the relative abundance of Cd36 as PPARγ target gene. Relative Cd36 mRNA expression in glomerular extracts from control or NTS-challenged mice (means of five mice per group, of two independent experiments). (D) Representative images of PPARγ expression (red) and podocalyxin (green) by immunofluorescence on kidney sections from control mice (control) or NTS-challenged mice (NTS) 10 days after the first injection of nephrotoxic serum. Images are representative of at least six mice per condition. Scale bar, 50 μm. **P<0.01 versus control mice.

Figure 2.

Podocyte-specific deletion of PPARγ does not modify kidney structure and function. (A) RT-PCR analysis of PPARγ mRNA expression in primary culture of podocytes from Pod-PPARγ WT and Pod-PPARγ lox mice (means of seven mice per group, of two independent experiments). (B) Immunofluorescence staining for PPARγ (green) and nidogen (red). Nidogen served as a marker for the basement membrane. Scale bar, 10 μm. (C) Representative images of Masson trichrome-stained kidney sections from Pod-PPARγ WT and Pod-PPARγ lox mice at 10 weeks old. Images are representative of at least five mice. Scale bar, 20 μm. (D) Urinary albumin excretion rates (means of 20 mice per group, of four independent experiments) and (E) BUN concentration at day 10 after NTS injection in groups of mice as in (A) (means of six mice per group, of two independent experiments). **P<0.005 versus Pod-PPARγ WT mice.

Figure 3.

Selective deletion of PPARγ from podocytes accelerates renal destruction in NTN. (A) Masson trichrome-stained kidney sections from Pod-PPARγ WT and Pod-PPARγ lox mice—crescent outlined in yellow—and (B) proportion of crescentic glomeruli (day 10 after NTS injection). Scale bar, 20µm. (C, D) Albumin urinary excretion rate (C) and BUN concentration (D) at day 10 after NTS injection in groups of mice as in A. (Means of 17 mice per group, of three independent experiments). *P<0.05; **P<0.01; ***P<0.005 versus Pod-PPARγ WT mice. (E) Ultrastructural analysis of podocytes by transmission electron microscopy from NTS-injected Pod-PPARγ WT and Pod-PPARγ lox mice.

Figure 4.

Podocyte-specific deletion of PPARγ accentuates inflammatory cells infiltration in NTN. (A) Representative pictures showing immunostaining for sheep IgG and mouse IgG in renal cortex from Pod-PPARγ WT and Pod-PPARγ lox mice after NTS injection on day 10 and from untreated normal Pod-PPARγ WT mice (control) Scale bar, 20µm. (B, C) Quantitative image analysis of immunofluorescent staining for glomerular mouse IgG (B) and sheep IgG deposition (C) 10 days after NTS injection. (D) Titers of mouse IgG to sheep IgG measured in serial dilutions of plasma from Pod-PPARγ WT or Pod-PPARγ lox mice immunized with sheep NTS and from non-immunized control mice (means of five mice per group, of two independent experiments). (E) Immunostaining for CD3⁺ and F4/80⁺ cells (scale bar, 20µm) and (F, G) quantification by image analysis of CD3⁺ and F4/80⁺ infiltrates in renal cortex from nonimmunized controls and NTS-injected Pod-PPARγ WT or Pod-PPARγ lox mice at day 10 after NTS injection (means of eight mice per group, of two independent experiments). (H and I) mRNA expression of Mcp1 and Il6 was determined by RT-PCR analysis in renal cortex tissue from groups of mice as in (A) (means of eight mice per group, of two independent experiments). *P<0.05; **P<0.01; ***P<0.005 versus control mice.

Figure 5.

Loss-of-function approach by podocyte-specific deletion of Pparγ or gain-of-function with rosiglitazone treatment do not modify podocyte proliferation and migration in vitro. (A, B) Representative pictures and quantification of podocyte proliferation involving decapsulated glomeruli from Pod-PPARγ lox mice or podocytes treated with rosiglitazone (means of 11 mice per group, of three independent experiments). Podocyte outgrowth area was assessed after 4 days, Scale bars, 20 µm. (C) RT-PCR analysis of the relative abundance of ki67 in primary podocyte cultures treated either with or without rosiglitazone (10 µM) for 16 hours, or from Pod-PPARγ lox mice (means of five mice per group, of two independent experiments). (D, E) Representative images and quantification of wound assay showing the migration of podocytes incubated either with or without rosiglitazone for 12 hours, or from Pod-PPARγ lox mice (means of 12 mice per group, of three independent experiments). Migration was assessed over a period of 12 hours. Scale bars, 100 µm.

Figure 6.

Podocyte-specific deletion of PPARγ aggravates sclerotic lesions in DXR model. (A) Albumin urinary excretion rate at day 9 after DXR administration in Pod-PPARγ WT and Pod-PPARγ lox mice. (B) Representative photomicrographs of Masson trichrome-stained glomerular sections from three Pod-PPARγ WT and Pod-PPARγ lox mice, 9 days after DXR administration. The latter group exhibited more prevalent capillary loop denudation, parietal epithelial cells to podocyte or capillary loop bridges and segmental sclerosis. Scale bar, 10 µm. (C) Proportion of sclerotic glomeruli at day 9 after DXR administration in Pod-PPARγ WT and Pod-PPARγ lox mice (means of five mice per group, of one experiment). *P<0.05 versus control Pod-PPARγ WT mice.

Figure 7.

PPARγ abundance and activity is decreased in NRF2 KO podocytes in vitro and in vivo. (A) Representative transmission electron micrographs of glomeruli from 10-week-old NRF2 WT and NRF2 KO male mice at baseline. (B) Masson trichrome-stained kidney sections of NRF2 WT and NRF2 KO 10-week-old mice at baseline. Scale bars, 20 µm. (C, D) Urinary albumin excretion rate (C) and BUN concentration (D) at baseline of NRF2 WT and NRF2 KO 10-week-old male mice. (E) PPARγ activity determined by RT-PCR analysis of the relative abundance of Cd36 mRNA in primary podocyte cultures from NRF2 WT or NRF2 KO mice. (F) PPARγ protein expression in primary podocyte cultures (F) from NRF2 WT or NRF2 KO mice. (G) Representative image of NRF2 (red) and podocalyxin (green) expression by immunofluorescence on kidney sections from NRF2 WT and NRF2 KO mice at baseline. Scale bars, 20 μm. Values are means of six mice per group, of two independent experiments. *P<0.05; **P<0.01; ***P<0.001 versus NRF2 WT mice.

Figure 8.

Decrease in NRF2 pathway upon selective deletion of PPARγ in podocytes in vitro and in vivo. (A) Western blots analysis of NRF2 in freshly isolated glomeruli from Pod-PPARγ WT and Pod-PPARγ lox mice at baseline (means of six mice per group, of two independent experiments). (B) NRF2 activity determined by RT-PCR analysis of the relative abundance of Nqo1 and Gstm1, as NRF2 target genes, in primary podocyte cultures from Pod-PPARγ WT and Pod-PPARγ lox mice. Values are means±SEM from five mice. **P<0.01 versus Pod-PPARγ WT mice. (C) Representative images of NRF2 and WT1 co-staining in cultured podocytes (means of eight mice per group, of three independent culture experiments). Note the distinct pattern of NRF2 expression with frequent nuclear localization in Pod-PPARγ WT cells whereas Pod-PPARγ lox podocytes display no NRF2 staining in nuclei, consistent with blunted transcriptional activity. Scale bars, 20 µm.

Figure 9.

NRF2-deficient mice develop more severe glomerulonephritis than normal littermates although with similar anti-sheep IgG humoral response. (A) Representative pictures of Masson trichrome-stained kidney sections—crescent outlined in yellow. Scale bars, 20 µm (B) Albuminuria and (C) BUN concentrations in NTS-challenged NRF2 KO and NRF2 WT mice (means of 15–17 mice per group, of four independent experiments). (D) Representative transmission electron micrographs of glomeruli from NRF2 WT mice at day 10 after NTS injection. Nephritic NRF2 KO mice display glomerular capillary endotheliosis and more widespread podocyte foot process effacement than their NRF2 WT counterparts. **P<0.01; ***P<0.001 versus NTS-injected NRF2 WT mice. (E) Representative photomicrographs after immunofluorescent staining for mouse IgG and sheep IgG in renal cortex from nonimmunized (control) mice and from NTS-injected NRF2 WT and NRF2 KO mice. (F, G) Quantitative image analysis of immunofluorescence staining for glomerular mouse IgG (F) and sheep IgG deposition (G) 10 days after nephrotoxic serum injection. (H) Titers of mouse IgG to sheep IgG measured in plasma from controls and NTS-injected NRF2 WT or NRF2 KO (means of five mice per group, of two independent experiments).

Figure 10.

PPARγ agonism attenuates the deleterious consequences of NRF2 deficiency in RPGN. (A) Representative pictures of Masson trichrome-stained kidney sections from NTS-challenged NRF2 KO and NRF2 WT mice treated with pioglitazone (Pio) or vehicle—crescent outlined in yellow—and (B) proportion of crescentic glomeruli. (C) BUN concentrations in groups of mice as in (A) (means of four to ten mice per group, one experiment). *P<0.05; **P<0.01; ***P<0.001 versus NTS- and vehicle-treated NRF2 WT mice; ^##P<0.01 versus vehicle-treated NRF2 KO nephritic group.

Figure 11.

Limited effectiveness of pioglitazone administration in protecting from NTN when the Pparγ gene is absent in podocytes. (A) Masson trichrome-stained kidney sections of Pod-PPARγ WT and Pod-PPARγ lox mice with or without pioglitazone treatment at day 10 after NTS injection—crescent outlined in yellow. Scale bar, 10 µm. (B) Proportion of crescentic glomeruli in kidneys from groups of mice as in (A). (C) Albumin urinary excretion rate and BUN concentration (D) in groups of mice as in (A) (Means of 17 mice per group, of three independent experiments). *P<0.05; **P<0.01; ***P<0.005 versus untreated NTS-challenged Pod-PPARγ WT mice; ^#P<0.05 and ^##P<0.01 versus Pod-PPARγ lox mice treated with pioglitazone (Pod-PPARγ lox NTS+Pio).

Figure 12.

Pioglitazone treatment improves glomerular structure and function in NTN. (A) Urinary albumin excretion rates in noninjected mice (control) and at day 4 and 10 after NTS in NTS-challenged mice (NTS) or NTS-challenged mice treated with pioglitazone started either at the same time as NTS (NTS+Pio) or in a curative protocol, started 4 days later (NTS+Pio delayed). (B) BUN concentration in groups of mice as in (A). (C) Representative images of Masson trichrome-stained kidney sections from groups of mice as in (A)—crescent outlined in yellow. Images are representative of at least eight mice per group. Scale bar, 20 μm. (D) Proportion of crescentic glomeruli in groups of mice as in (B). Values are means±SEM of 8–12 mice per group, of two independent experiments. *P<0.05 versus unchallenged control mice; ^#P<0.05 versus NTS-challenged mice (NTS); ^ΦP<0.05 versus NTS+Pio. (E) Titers of mouse IgG to sheep IgG measured in serial dilutions of plasma from mice immunized with sheep NTS treated with pioglitazone (Pio) as in (A) and from nonimmunized control mice. (F) Representative pictures showing immunostaining for mouse IgG and sheep IgG in renal cortex in noninjected mice (control) or 10 days after NTS injection in groups as in (A). (G, H) Quantitative image analysis of immunofluorescence staining for glomerular mouse IgG and sheep IgG deposition in groups as in (A). Values are means of five mice per group, of two independent experiments. ***P<0.001 versus unchallenged control mice.

Figure 13.

Pioglitazone treatment decreases renal leukocytes infiltration in NTN. (A) Immunostaining for CD3⁺ and F4/80⁺ cells in renal cortex from nontreated mice (control), NTS-challenged mice (NTS) and NTS-challenged mice with pioglitazone treatment at same time (NTS+Pio). Scale bar, 20µm. (B, C) Quantification of CD3⁺ area or F4/80⁺ area to total area ratio (means of eight mice per group, of two independent experiments). (D, E) mRNA expression of mcp1 and il6 determined by RT-PCR analysis in renal cortex tissue from groups of mice as in (A). *P<0.05, **P<0.01; ***P<0.001 versus unchallenged control mice and #P<0.05 versus NTS-challenged mice.

Figure 14.

Loss of nuclear PPARγ in podocytes and crescents in patients diagnosed for RPGN. Representative images of immunostaining for PPARγ in sections of kidney biopsies from random control patient (CT) or 3 subjects diagnosed with RPGN (MPA and GPA). Scale bars, 100 μm. (A, B, C and D) higher magnification of left panel (red square). Selective nuclear PPARγ (black arrows) is seen in glomerular endocapillary cells as well as with an extracapillary pattern in glomerular epithelial cells: podocytes and parietal epithelial cells. PPARγ expression is abolished in the nuclei of cells forming the crescent in favor of cytoplasmic expression (yellow arrows), whereas endocapillary staining is retained.