High Basolateral Glucose Increases Sodium-Glucose Cotransporter 2 and Reduces Sirtuin-1 in Renal Tubules through Glucose Transporter-2 Detection

Abstract

Under diabetic conditions, sodium-glucose cotransporter 2 (SGLT2) for glucose uptake in proximal tubules (PTs) increases, whereas NAD⁺-dependent protein deacetylase silent mating type information regulation 2 homolog 1 (Sirtuin-1; SIRT1) for PT survival decreases. Therefore, we hypothesized that increased glucose influx by SGLT2 reduces SIRT1 expression. To test this hypothesis, db/db mice with diabetes and high-glucose (HG)-cultured porcine PT LLC-PK1 cells in a two-chamber system were treated with the SGLT2 inhibitor canagliflozin. We also examined SIRT1 and SGLT2 expression in human kidney biopsies. In db/db mice, SGLT2 expression increased with concomitant decreases in SIRT1, but was inhibited by canagliflozin. For determination of the polarity of SGLT2 and SIRT1 expression, LLC-PK1 cells were seeded into Transwell chambers (pore size, 0.4 µm; Becton Dickinson, Oxford, UK). HG medium was added to either or to both of the upper and lower chambers, which corresponded to the apical and basolateral sides of the cells, respectively. In this system, the lower chamber with HG showed increased SGLT2 and decreased SIRT1 expression. Canagliflozin reversed HG-induced SIRT1 downregulation. Gene silencing and inhibitors for glucose transporter 2 (GLUT2) blocked HG-induced SGLT2 expression upregulation. Gene silencing for the hepatic nuclear factor-1α (HNF-1α), whose nuclear translocation was enhanced by HG, blocked HG-induced SGLT2 expression upregulation. Similarly, gene silencing for importin-α1, a chaperone protein bound to GLUT2, blocked HG-induced HNF-1α nuclear translocation and SGLT2 expression upregulation. In human kidney, SIRT1 immunostaining was negatively correlated with SGLT2 immunostaining. Thus, under diabetic conditions, SIRT1 expression in PTs was downregulated by an increase in SGLT2 expression, which was stimulated by basolateral HG through activation of the GLUT2/importin-α1/HNF-1α pathway.

Figure 1

Effect of canagliflozin on parameters of glucose metabolism. (A) Schedule for canagliflozin (Cana) treatment and experimental groups. Effects of Cana on body weight (B), food intake (C), fasting plasma glucose level (D) and glycated Hb levels (E) in db/db mice. HbA1c, hemoglobin A1c. Data represent the mean ± SEM (n = 8 mice/group). *P < 0.05 vs. db/m and ^§P < 0.05 vs. db/db.

Figure 2

Effect of Cana on SGLT2 and SIRT1 expression in db/db mice. (A) Immunohistochemical analysis using a SGLT2-specific antibody. Representative kidney sections are shown for each group of mice. Quantitation of relative density is shown in the bar graph in the right panel. Scale bar, 100 µm. *P < 0.05 vs. db/m control and ^§P < 0.05 vs. db/db control (n = 8 mice/group) (B) The upper panel shows representative immunoblotting analysis of SGLT2 expression. The Bar graph in the lower panel indicates quantification of SGLT2 levels. Protein expression was normalized to that of α-tubulin. Relative protein levels are shown as the fold-change to the db/m (control) group. *P < 0.05 vs. db/m control and ^§P < 0.05 vs. db/db control. The results are representative of four independent experiments. (C) Immunohistochemical analysis using a SIRT1-specific antibody. Representative kidney sections are shown for each group of mice. Quantitation of relative density is shown in the bar graph. Scale bar, 50 µm. *P < 0.05 vs. db/m control and ^§P < 0.05 vs. db/db control (n = 8 mice/group). (D) The upper panel shows representative immunoblotting analysis of SIRT1 expression. The bar graph in the lower panel indicates quantification of SIRT1 levels. Protein expression was normalized to that of α-tubulin. Relative protein levels are shown as the fold-change to the db/m (control) group. *P < 0.05 db/m control and ^§P < 0.05 vs. db/db control. Results are representative of four independent experiments.

Figure 3

Effect of high glucose on SGLT2 expression in a two-chamber culture system. (A) Confluent growth-arrested cell monolayers were stimulated on the apical or basolateral side for up to 24 h, as described in the Methods section. Representative SGLT2 fluorescence of the four groups of culture conditions is shown. The bar graph represents the fluorescence intensity of each group (*P < 0.05 vs. the group with 5.5 mM D-glucose in the lower chamber and 5.5 mM D-glucose in the upper chamber, n = 4 independent experiments. Scale bar, 50 µm. (B) Effects of inhibitors for candidate signaling pathways on SGLT2 expression. LLC-PK1 cells were treated with a Na-K ATPase channel inhibitor (glibenclamide, 50 µM), GLUT2 inhibitor (phloretin, 100 µM), angiotensin II type 1 receptor (AT1R) inhibitor (losartan, 1 µM), and AT2R inhibitor (PD123319, 100 µM). Results are representative of four independent experiments. The bar graph represents the band intensity of each group (*P < 0.05 vs. NG group and ^§P < 0.05 vs. HG group, n = 4 independent experiments). (C) Effects of Glut2-siRNA (100 nmol/L) or non-targeting control siRNA (100 nmol/L) on SGLT2 expression. Results are representative of four independent experiments. The bar graph in the right panel indicates the quantification of SGLT2 levels. Protein expression was normalized to that of α-tubulin. Relative protein levels are shown as the fold-change to the NG group. *P < 0.05 vs. NG group, ^§P < 0.05 vs. HG group, and ^¶P < 0.05 vs. HG with control siRNA, n = 4. (D) Subcellular fractionation and immunoblotting demonstrate HG-induced redistribution from the cytoplasm to the nucleus of HNF-1α. HG, high D-glucose (22.5 mM) condition; NG, normal D-glucose (5.5 mM) condition; N, nuclear fraction; C, cytosolic fraction. HNF-1α protein levels were determined by densitometry and are indicated for each fraction. The percent of HNF-1α in the nucleus relative to the total HNF-1α was calculated using values determined using a densitometry. *P < 0.05 vs. NG group. Results are representative of four independent experiments. (E) Effects of HNF-1α-siRNA (100 nmol/L) or non-targeting control siRNA (100 nmol/L) on SGLT2 expression. The bar graph in the right panel indicates quantification of SGLT2 levels. Relative protein levels are shown as the fold-change to the NG (control) group. *P < 0.05 vs. NG group, ^§P < 0.05 vs. HG group, and ^¶P < 0.05 vs. HG with control siRNA, n = 4 independent experiments.

Figure 4

The GLUT2/importin-α1/HNF-1α pathway mediates high-glucose-induced SGLT2 expression. (A) Quantitative real-time PCR analysis of the relative abundance of mRNAs encoding importins in LLC-PK1 cells. Real-time PCR data were normalized to those of the GAPDH mRNA. Relative fold-differences were calculated using the mean value (n = 6) of importin-α6. (B) Immunoblotting confirmed the efficiency of siRNA knockdown of each of the representative importin-α isoform that was abundantly expressed in PTs. Results are representative of four independent experiments. (C) Subcellular fractionation and immunoblotting analysis of the localization of HNF-1α. LLC-PK1 cells were transfected with the indicated siRNA duplexes and 48 h later, cells were treated with media containing high glucose levels. Cytoplasmic and nuclear lysates were collected and analyzed using immunoblotting with an anti-HNF-1α antibody. Results are representative of four independent experiments. (D) LLC-PK1 cells were transiently transfected with siRNAs targeting importin-α1 or a non-targeting control siRNA. Whole cell lysates were prepared 48 h post-transfection and analyzed using immunoblotting. Results are representative of four independent experiments. The bar graph represents the band intensity of each group (*P < 0.05 vs. NG group and ^§P < 0.05 vs. HG group, n = 4). (E,F) Immunoprecipitation of nuclear extracts from LLC-PK1 cells that were treated with NG or HG. Importin-α1 formed complexes with HNF-1α (E), whereas it dissociated from GLUT2 (F) under HG conditions. Results are representative of four independent experiments. (G) Model for the molecular mechanism by which the GLUT2/importin-α1/HNF-1α pathway is involved in HG-induced SGLT2 expression upregulation.

Figure 5

Effect of Cana on the expression of SIRT1 and its downstream signaling components. Confluent growth-arrested LLC-PK1 cell monolayers were stimulated with HG medium on the basolateral side for up to 24 h with or without pretreatment with Cana in the apical side. Immunofluorescence analysis (A) and immunoblotting (B) for SGLT2 expression in LLC-PK1 cells. The relative quantification of the SGLT2 immunofluorescence was measured and is indicated in the bar graphs. Scale bar, 50 µm. *P < 0.05 vs. NG group, n = 4 independent experiments. (C) The effect of SGLT2 inhibitors on cellular glucose entry in LLC-PK1 cells. LLC-PK1 cells were incubated in Dulbecco’s modified minimal essential medium containing 100 μM 2-NBDG for 15 min from the apical side of the cell. Cellular glucose entry was assessed as 2-NBDG entry into the cell, as described in the Methods section. Scale bar = 20 µm. *P < 0.05 vs. NG group and ^§P < 0.05 vs. HG without Cana, n = 4 independent experiments. NG, normal glucose (5.5 mM); HG, high glucose (22.5 mM); HG + Cana, HG with Cana treatment. (D) Immunoblotting analysis of SIRT1 expression in LLC-PK1 cell monolayers stimulated with HG medium on the basolateral side for up to 24 h with or without pretreatment with low or high doses of Cana. Quantification of immunoblotting images was normalized to α-tubulin. *P < 0.05 vs. NG group and ^§P < 0.05 vs. HG without Cana, n = 4 independent experiments.

Figure 6

Immunohistochemical analysis of SIRT1 and SGLT2 expression in human renal biopsy specimens. (A) Representative photomicrographs of hematoxylin and eosin staining or immunostaining of SIRT1 and SGLT2 in renal needle-biopsy specimens of patients with diabetic nephropathy (DN) (DN-1 and -5, Table 1 for patient details). Scale bar, 50 nm. (B) The relationship between the intensity of immunostaining of SIRT1 and that of SGLT2 in the proximal tubular region in renal biopsy specimens from patients with DN, n = 11 subjects.