SGLT2 mediates glucose reabsorption in the early proximal tubule

Abstract

Mutations in the gene encoding for the Na(+)-glucose co-transporter SGLT2 (SLC5A2) associate with familial renal glucosuria, but the role of SGLT2 in the kidney is incompletely understood. Here, we determined the localization of SGLT2 in the mouse kidney and generated and characterized SGLT2-deficient mice. In wild-type (WT) mice, immunohistochemistry localized SGLT2 to the brush border membrane of the early proximal tubule. Sglt2(-/-) mice had glucosuria, polyuria, and increased food and fluid intake without differences in plasma glucose concentrations, GFR, or urinary excretion of other proximal tubular substrates (including amino acids) compared with WT mice. SGLT2 deficiency did not associate with volume depletion, suggested by similar body weight, BP, and hematocrit; however, plasma renin concentrations were modestly higher and plasma aldosterone levels were lower in Sglt2(-/-) mice. Whole-kidney clearance studies showed that fractional glucose reabsorption was significantly lower in Sglt2(-/-) mice compared with WT mice and varied in Sglt2(-/-) mice between 10 and 60%, inversely with the amount of filtered glucose. Free-flow micropuncture revealed that for early proximal collections, 78 ± 6% of the filtered glucose was reabsorbed in WT mice compared with no reabsorption in Sglt2(-/-) mice. For late proximal collections, fractional glucose reabsorption was 93 ± 1% in WT and 21 ± 6% in Sglt2(-/-) mice, respectively. These results demonstrate that SGLT2 mediates glucose reabsorption in the early proximal tubule and most of the glucose reabsorption by the kidney, overall. This mouse model mimics and explains the glucosuric phenotype of individuals carrying SLC5A2 mutations.

Figure 1.

Generation of transgenic Sglt2^−/− mice. (A) Targeting strategy used to disrupt the Slc5a2 locus. Homologous recombination (represented by X) between the targeting vector and the Slc5a2 gene results in the replacement of exons 1 through 5 with the selection cassette. (B) Southern hybridization indicating proper gene targeting in four ES cell clones. Clone 4 is transmitted through the germline; control represents untransfected ES cell DNA. (C) PCR amplification of tail-snip DNA from Sglt2^−/−, WT (+/+), and heterozygous (+/−) mice. Separate PCRs are performed for the WT allele (left lane) and the targeted allele (right lane) for each sample.

Figure 2.

Renal expression of Glut1, Glut2, and Glut12 is unaltered but the expression of Sglt1 is reduced in Sglt2^−/− mice. (A) Renal mRNA expression of Glut1, Glut2, and Glut12 are unchanged whereas Sglt1 is reduced in Sglt2^−/− relative to WT mice. Real-time PCR confirms the knockout of Sglt2 in kidney of Sglt2^−/− mice. ND, not detectable. n = 5 per genotype. *P < 0.01 versus WT mice. (B) Western blotting shows reduced renal SGLT1 protein expression (related to β-actin) in Sglt2^−/− versus WT mice (n = 4 per genotype). *P < 0.05 versus WT mice. Gene knockout of Sglt1 in mice (Sglt1^−/−) confirms the specificity of the signal.

Figure 3.

Expression of SGLT2 protein in the apical brush border membrane of the early proximal convoluted tubule. (A through D) Renal cortical sections (A and B) and outer medullary sections (C and D) are shown for WT and Sglt2^−/− mice. Left pictures show staining with SGLT2 Ab (red fluorescence); right pictures show additional co-staining with phalloidin (green fluorescence) to label filamentous actin of the proximal tubular brush border. Nuclei are stained with the marker DAPI (blue fluorescence). (A and B) The SGLT2 Ab provides specific and selective staining of the apical brush border membrane of the early proximal convoluted tubules. The staining for SGLT2 begins where the proximal tubule initiates from Bowman space (*) and ends along the proximal convoluted tubule (arrowhead). Further downstream sections of the proximal convoluted tubule (identified by reduced height of brush border) do not stain for SGLT2 (arrows). (C and D) No specific SGLT2 signal is detected in the proximal straight tubules of the outer medulla. Arrows mark the distal ends of proximal straight tubules.

Figure 4.

Urinary glucose concentration is enhanced in Sglt2^−/− mice despite normal blood glucose levels and GFR: Studies in awake mice. (A) Glucose measured in spontaneous urine collections and subsequent blood collections from tail vein. (B) GFR determined by plasma FITC-inulin kinetics. (C) Food and fluid intake assessed in regular cages. (D) BP measurements using a tail-cuff system in trained awake mice (n = 9 to 10 per group). *P < 0.05 versus WT mice.

Figure 5.

Renal excretion of glucose in Sglt2^−/− is a function of the amounts of glucose filtered: Clearance studies under anesthesia (n = 9 to 10 per group).

Figure 6.

Reabsorption of glucose is absent in the early proximal tubule of Sglt2^−/− mice: Micropuncture studies under anesthesia. (A) Free-flow collections of tubular fluid are performed along accessible proximal tubules at the kidney surface to establish a profile for FR-glucose versus FR-fluid. (B and C) Mean FR-glucose (B) and fractional reabsorption of chloride (C) for early (FR-fluid <40%) and late (FR-fluid ≥40%) proximal tubular collections and up to the urine (n = 18 to 23 nephrons in four to five mice). *P < 0.001 versus WT mice.