REDD1 expression in podocytes facilitates renal inflammation and pyroptosis in streptozotocin-induced diabetic nephropathy

Abstract

Sterile inflammation resulting in an altered immune response is a key determinant of renal injury in diabetic nephropathy (DN). In this investigation, we evaluated the hypothesis that hyperglycemic conditions augment the pro-inflammatory immune response in the kidney by promoting podocyte-specific expression of the stress response protein regulated in development and DNA damage response 1 (REDD1). In support of the hypothesis, streptozotocin (STZ)-induced diabetes increased REDD1 protein abundance in the kidney concomitant with renal immune cell infiltration. In diabetic mice, administration of the SGLT2 inhibitor dapagliflozin was followed by reductions in blood glucose concentration, renal REDD1 protein abundance, and immune cell infiltration. In contrast with diabetic REDD1^+/+ mice, diabetic REDD1^-/- mice did not exhibit albuminuria, increased pro-inflammatory factors, or renal macrophage infiltration. In cultured human podocytes, exposure to hyperglycemic conditions promoted REDD1-dependent activation of NF-κB signaling. REDD1 deletion in podocytes attenuated both the increase in chemokine expression and macrophage chemotaxis under hyperglycemic conditions. Notably, podocyte-specific REDD1 deletion prevented the pro-inflammatory immune cell infiltration in the kidneys of diabetic mice. Furthermore, exposure of podocytes to hyperglycemic conditions promoted REDD1-dependent pyroptotic cell death, evidenced by an NLRP3-mediated increase in caspase-1 activity and LDH release. REDD1 expression in podocytes was also required for an increase in pyroptosis markers in the glomeruli of diabetic mice. The data support that podocyte-specific REDD1 is necessary for chronic NF-κB activation in the context of diabetes and raises the prospect that therapies targeting podocyte-specific REDD1 may be helpful in DN.

Fig. 1. Hyperglycemia-induced REDD1 expression was associated with increased renal immune cell infiltration in diabetic mice.

Diabetes was induced in mice by administration of streptozotocin (STZ). Analyses were performed 16 weeks after administration of STZ or vehicle (Veh). Hyperglycemia was controlled by daily administration of dapagliflozin (DG) beginning 14 weeks after diabetes induction. As a control for DG intervention, diabetic mice were administered a vehicle containing 0.1% DMSO. A Fasting blood glucose concentrations were measured. Arrow indicates the initiation of DG/DMSO intervention. B REDD1 and actin protein abundance was assessed in kidney cortical tissue homogenates by western blotting. Representative blots are shown. Protein molecular mass is indicated at right of each blot. Individual data points are plotted with values presented as means ± SD (n = 3–4). C F4/80 positive immune cells were identified in renal sections by immunohistochemistry (brown). Nuclei were counterstained with hematoxylin (HT, blue). Representative micrographs are shown (scale bar 50 µm). D F4/80 positive cells were quantified in 20 fields per section. Data distribution is represented by violin plot. Differences between groups were identified by one-way ANOVA. *p < 0.05 versus Veh; #p < 0.05 versus DMSO.

Fig. 2. REDD1 was necessary for increased expression of inflammatory factors in the kidney of diabetic mice.

Diabetes was induced in REDD1^+/+ and REDD1^−/− mice by streptozotocin (STZ) administration. Non-diabetic control mice received vehicle (Veh). A REDD1 protein was evaluated in kidney cortical tissue homogenates by western blotting. Representative blots are shown. Molecular mass in kDa is indicated at right of each blot. B Correlation between fasting blood glucose and urine ACR is shown for REDD1^+/+ mice (blue; Pearson r = 0.72; p < 0.0001) and REDD1^−/− mice (red; Pearson r = 0.55; p = 0.029). C Ccl2 mRNA expression was quantified in kidney homogenates by qPCR. D CCL2 protein abundance was quantified in kidney homogenates by western blotting. E Il1b mRNA expression was quantified in kidney homogenates by qPCR. F IL-1β protein levels were determined in kidney homogenates by western blotting and quantified by ELISA. Individual data points are plotted with values presented as means ± SD (n = 4–6). Differences between groups were identified by two-way ANOVA. *p < 0.05 versus Veh; #p < 0.05 versus REDD1^+/+. n.d., not detected.

Fig. 3. REDD1 was required for renal infiltration of pro-inflammatory macrophages in diabetic mice.

A Kidney sections from diabetic (STZ) and non-diabetic (Veh) mice were immunolabeled for F4/80 (brown) and counterstained with hematoxylin (HT, blue). Representative micrographs are shown (scale bar 50 µm). B F4/80 positive cells in A were quantified in 20 fields per renal section. Data distribution is represented by violin plot. C–G Flow cytometry was used to identify renal immune cell populations. Gating strategy to determine immune cell populations in the kidney is shown (C). Populations of CD45+ cells (D), CD11b + F4/80+ macrophages (E), CD86 + M1 macrophages (F) and CD206 + M2 macrophages (G) were determined. Individual data points are plotted with values presented as means ± SD (n = 3–4). Differences between groups were identified by two-way ANOVA. *p < 0.05 versus Veh; #p < 0.05 versus REDD1^+/+.

Fig. 4. REDD1 deletion prevented diabetes-induced NF-κB activation in the kidney.

A, B Diabetes was induced in REDD1^+/+ and REDD1^−/− mice by administration of streptozotocin (STZ). Non-diabetic control mice received vehicle (Veh). A Nuclear isolates were prepared from kidney homogenates. NF-κB and Lamin B were examined in nuclear isolates by western blotting and NF-κB activity was quantified by DNA-binding ELISA. Representative blots are shown with protein molecular mass in kDa indicated at right of each blot. B REDD1 (red) and Nephrin (green) were visualized in kidneys by immunofluorescence microscopy. White box indicates area shown at increased magnification. Representative micrographs are shown (scale bar 50 μm). C–I Wild-type (WT) and REDD1 knockout (KO) CIHP-1 were exposed to culture media containing either 30 mM glucose (HG) or 5 mM glucose plus 25 mM mannitol (OC) for 48 h. NF-κB phosphorylation at S536 and REDD1 protein abundance was determined in cell lysates by western blotting (C). Nuclear localization of NF-κB p65 (white arrowheads) was evaluated by immunofluorescence (D). Nuclei were visualized with DAPI (scale bar 25 μm). NF-κB activity was measured in lysates from cells expressing NF-κB firefly luciferase/Renilla luciferase reporter plasmids by dual luciferase assay (E). Relative expression of IL1B and CCL2 mRNA were determined by qPCR (F). IL-1β secreted into culture media was determined by ELISA (G). Chromatin immunoprecipitation (ChIP)-PCR analysis was carried out in WT and REDD1 KO podocytes to determine binding of p65 NF-κB to the promoter region of the CCL2 gene (H). CCL2 protein levels were determined in cell lysates by western blotting (I). J NF-κB p65 phosphorylation and NF-κB luciferase reporter activity was evaluated in REDD1 KO cells expressing either an empty vector control (EV) or hemagglutinin (HA)-tagged REDD1. Individual data points are presented as means ± SD (n = 4–6). Differences between groups were identified by two-way ANOVA. *p < 0.05 versus Veh or NG; #p < 0.05 versus REDD1^+/+, WT, or EV.

Fig. 5. Podocyte-specific expression of REDD1 was required for increased renal macrophage infiltration in the kidney of diabetic mice.

A, B Differentiated wild-type (WT) and REDD1 knockout (KO) CIHP-1 were exposed to culture media containing either 30 mM glucose (HG) or 5 mM glucose plus 25 mM mannitol as an osmotic control (OC) for 48 h. Transwell migration assay was used to evaluate chemotaxis in a co-culture model with CIHP-1 and THP-1 macrophages (A). Macrophages were stained with crystal violet and cells that migrated across the Transwell were counted (B). C Cre-lox recombination was used to achieve conditional podocyte-specific REDD1 knockout (REDD1 PodKO). D–H Diabetes was induced in REDD1^fl/fl and REDD1 PodKO mice by streptozotocin (STZ) administration. Non-diabetic groups were administered a vehicle (Veh) control. All assessments were performed after 16 weeks of diabetes. Urine albumin to creatinine ratio (ACR) was determined (D). Kidney sections from diabetic and non-diabetic mice were immunolabeled for REDD1 (red) and the podocyte marker Nephrin (green) (E). Protein abundance of CCL2 was determined in renal homogenates by western blotting (F). Representative blots are shown with protein molecular mass in kDa indicated at right of each blot. Kidney sections were immunolabelled for F4/80 (red) and nuclei were counterstained with Hoechst 33342 (blue) (G). Representative micrographs (scale bar 50 µm) are shown. Immune cell populations of CD11b + F4/80+ macrophages (H) and CD86 + M1 macrophages (I) were determined by flow cytometry. Individual data points are plotted. Significance was analyzed by two-way ANOVA and pairwise comparisons were made using the Tukey’s test for multiple comparisons. *p < 0.05 versus OC or Veh; #, p < 0.05 versus WT or REDD1^fl/fl.

Fig. 6. REDD1 expression in podocytes was required for NLRP3 inflammasome activation in diabetic mice.

A–E Differentiated wild-type (WT) and REDD1 knockout (KO) CIHP-1 cells were exposed to culture media containing either 30 mM glucose (HG) or 5 mM glucose plus 25 mM mannitol as an osmotic control (OC) for 48 h. Relative expression of NLRP3 mRNA was determined in cell lysates by qPCR (A). NLRP3 protein content relative to GAPDH was estimated by western blotting (B). Representative blots are shown with molecular mass in kDa indicated to the right of the blot. Active caspase-1 (green) was visualized by immunofluorescence microscopy using FAM-YVAD-FMK FLICA (fluorescently labeled inhibitor of caspase) probe (C; scale bar 25 µm). Gasdermin D (GSDMD) N-terminal cleavage (GSDMD-N) was determined in cell lysates by western blotting (D). LDH released into cell culture supernatant was quantified (E). F-I, Diabetes was induced in REDD1^fl/fl and REDD1 PodKO mice by streptozotocin (STZ) administration. Non-diabetic mice were administered a vehicle (Veh) control. Nlrp3 mRNA expression in glomerular isolates was quantified by qPCR (F). NLRP3 and GSDMD protein in glomerular isolates were determined by western blotting (G). Immunofluorescence microscopy was used to determine colocalization of NLRP3, GSDMD, and WT-1 with the podocyte marker nephrin (H). IL-1β protein content in renal homogenates was quantified by ELISA (I). Representative micrographs (scale bar 50 µm) are shown. Individual data points are plotted. Significance was analyzed by two-way ANOVA and pairwise comparisons were made using the Tukey’s test for multiple comparisons. *p < 0.05 versus OC or Veh; #p < 0.05 versus WT or REDD1^fl/fl.

Fig. 7. Working model for the role of podocyte-specific REDD1 expression in facilitating renal inflammation and pyroptosis in diabetic nephropathy.

Diabetes-induced hyperglycemia enhances REDD1-dependent activation of NF-κB signaling in podocytes, resulting in increased cytokine and chemokine production. Podocyte-specific expression of REDD1 is required for kidney immune cell infiltration, macrophage polarization, and NLRP3 inflammasome-associated pyroptosis in diabetes.