Intestinal absorption of glucose in mice as determined by positron emission tomography

Abstract

Key points: The goal was to determine the importance of the sodium-glucose cotransporter SGLT1 and the glucose uniporter GLUT2 in intestinal glucose absorption during oral glucose tolerance tests (OGTTs) in mice. Glucose absorption was determined in mice using positron emission tomography and three non-metabolizable glucose probes: one specific for SGLTs, one specific for GLUTs, and one a substrate for both SGLTs and GLUTs. Absorption was determined in wild-type, Sglt1^-/- and Glut2^-/- mice. Gastric emptying was a rate-limiting step in absorption. SGLT1, but not GLUT2, was important in fast glucose absorption. In the absence of SGLT1 or GLUT2, the oral glucose load delivered to the small intestine was slowly absorbed. Oral phlorizin only inhibited the fast component of glucose absorption, but it contributed to decreasing blood glucose levels by inhibiting renal reabsorption.

Abstract: The current model of intestinal absorption is that SGLT1 is responsible for transport of glucose from the lumen into enterocytes across the brush border membrane, and GLUT2 for the downhill transport from the epithelium into blood across the basolateral membrane. Nevertheless, questions remain about the importance of these transporters in vivo. To address these questions, we have developed a non-invasive imaging method, positron emission tomography (PET), to monitor intestinal absorption of three non-metabolized glucose tracers during standard oral glucose tolerance tests (OGTTs) in mice. One tracer is specific for SGLTs (α-methyl-4-[¹⁸ F]fluoro-4-deoxy-d-glucopyranoside; Me-4FDG), one is specific for GLUTs (2-deoxy-2-[¹⁸ F]fluoro-d-glucose; 2-FDG), and one is a substrate for both SGLTs and GLUTs (4-deoxy-4-[¹⁸ F]fluoro-d-glucose; 4-FDG). OGTTs were conducted on adult wild-type, Sglt1^-/- and Glut2^-/- mice. In conscious mice, OGTTs resulted in the predictable increase in blood glucose that was blocked by phlorizin in both wild-type and Glut2^-/- animals. The blood activity of both Me-4FDG and 4-FDG, but not 2-FDG, accompanied the changes in glucose concentration. PET imaging during OGTTs further shows that: (i) intestinal absorption of the glucose load depends on gastric emptying; (ii) SGLT1 is important for the fast absorption; (iii) GLUT2 is not important in absorption; and (iv) oral phlorizin reduces absorption by SGLT1, but is absorbed and blocks glucose reabsorption in the kidney. We conclude that in standard OGTTs in mice, SGLT1 is essential in fast absorption, GLUT2 does not play a significant role, and in the absence of SGLT1 the total load of glucose is slowly absorbed.

Keywords: PET; absorption; intestine.

Figure 1. Distribution of 2‐FDG in representative wild‐type and Glut2^–/– mice 60 min after gavage to illustrate the determination of gastric emptying and intestinal absorption

For each mouse, coronal (c) and sagittal (s) sections (0.2 mm thick) are shown, and the areas corresponding to the three‐dimensional volumes of interest (VOIs) for intestine (in), stomach (st), brain (br), heart (h) and urinary bladder (bl) are delineated. Images are displayed according to the NIH intensity scale for tracer activity, from red (highest), through green (intermediate) to purple (lowest). VOI analysis was performed using Amide software. Gastric emptying was estimated from the difference between the 2‐FDG measured in the whole‐body VOI and the amount remaining in the stomach, and intestinal absorption was estimated from the amount of 2‐FDG in the whole body minus the sum of the amounts detected in the stomach and the small intestine.

Figure 2. Standard OGTTs in wild‐type and in Glut2^–/– mice

Time course of blood glucose levels after orogastric gavage of 2 mg g⁻¹ of glucose. Data are means ± SEM from 10 (wild‐type) or 4 (Glut2^–/–) mice. Groups that share letters are statistically similar (‘a’ and ‘ab’, or ‘b’ and ‘ab’), whereas those not sharing any letters (‘a’ and ‘b’) are significantly different (P < 0.05, one‐way ANOVA and Tukey's test).

Figure 3. Time course of blood glucose concentration and activity of glucose PET tracers Me‐4FDG, 2‐FDG and 4‐FDG following oral delivery in mice

Following pre‐treatment with SGLT1 inhibitor phlorizin (1 mg g⁻¹) or with phlorizin‐free excipient (control), fasted wild‐type mice were administered an oral dose of glucose (2 mg g⁻¹) with 300 μCi of Me‐4FDG (A), 2‐FDG (B) or 4‐FDG (C). Blood glucose concentration (circles) and radiotracer activity (squares) were simultaneously monitored over the next 2 h in control (filled symbols) and phlorizin treated (open symbols) mice. The fasting blood samples (‘0 min’) were taken immediately before phlorizin or vehicle pre‐dosing. Data are from individual experiments where phlorizin and control mice were tested side by side, and results were confirmed in at least one additional set of experiments with different mice for each tracer.

Figure 4. Absorption of orally delivered glucose PET tracers Me‐4FDG, 2‐FDG and 4‐FDG in mice

Three hundred microcuries of Me‐4FDG, 2‐FDG or 4‐FDG was administered by oral gavage into the stomach of conscious control or Glut2^–/– mice, and 10 min PET scans were performed after 60 min. A, volumetric renderings of representative co‐registered microPET and CT scans; PET images in this and subsequent figures are displayed according to the NIH intensity scale for tracer activity, from red (highest), through green (intermediate) to purple (lowest). Stomach (st), intestine (in), brain (br), urinary bladder (bl), left kidney (lk) and right kidney (rk) are indicated where visible. B and C, gastric emptying (B) and intestinal tracer absorption (C) 60 min after oral bolus administration in experiments as in A. Symbols represent data from individual experiments; red symbols are for the mice shown in A. Except for Me‐4FDG in Glut2^–/– mice, where only two experiments were performed, bars indicate the mean ± SEM of 6–14 experiments. No significant differences in gastric emptying (P = 0.4319) or in tracer absorption (P = 0.766) were found between tracers or genotypes (two‐way ANOVA).

Figure 5. Me‐4FDG absorption is reduced in Sglt1^−/− mice, but not in phlorizin‐treated wild‐type mice

Glucose (2 mg g⁻¹) in presence of 300 μCi Me‐4FDG was administered by oral gavage into the stomach of anaesthetized wild‐type or Sglt1^–/– mice, and continuous PET data were acquired over the following 62 min; a CT scan was performed at the end of the dynamic PET scan. Wild‐type mice were pre‐treated with 1 mg g⁻¹ phlorizin, or with phlorizin‐free vehicle, as described above. Volumetric renderings of end‐point co‐registered images for representative mice are shown, and their values for gastric emptying, Me‐4FDG absorption and excretion are given at the bottom. PET images are displayed as in Fig. 4 A; stomach (st), intestine (in) and urinary bladder (bl) are indicated where visible. PET images are displayed according to the NIH intensity scale for tracer activity, from red (highest), through green (intermediate) to purple (lowest).

Figure 6. Phlorizin delays absorption and increases excretion of 4‐FDG in mice

Following pre‐treatment with phlorizin, or with phlorizin‐free vehicle (control), 2 mg g⁻¹ glucose in presence of 300 μCi 4‐FDG was administered by oral gavage into the stomach of conscious wild‐type mice, and brief 10 min PET scans were performed 10, 25 or 55 min later. A, volumetric renderings of co‐registered images in three control (top) and three phlorizin‐treated (bottom) mice treated side‐by‐side and scanned sequentially as appropriate; results were confirmed in at least one additional trial including all time points. Stomach (st), intestine (in) and urinary bladder (bl) are indicated where visible. B and C, 4‐FDG absorption (B) and excretion (C) for the mice shown in A. Results from each mouse are represented by a unique symbol, which is also shown to the side of each panel in A. Data points are joined by dashed lines for illustrative purposes only. Insets, mean ± SEM of 4‐FDG absorption (B) and excretion (C) at 65 min, in three control (C) and three phlorizin‐treated (P) mice. *P < 0.05, Student's t test. PET images are displayed according to the NIH intensity scale for tracer activity, from red (highest), through green (intermediate) to purple (lowest).

Figure 7. Dynamic Me‐4FDG microPET scans of control and Sglt1^−/− mice undergoing OGTTs

Glucose (2 mg g⁻¹) in presence of 300 μCi Me‐4FDG was administered by oral gavage into the stomach of anaesthetized wild‐type or Sglt1^–/– mice, and continuous PET data were acquired over the following 62 min; a CT scan was performed at the end of the dynamic PET scan. A, volumetric renderings of co‐registered images at selected times after bolus administration, for one representative wild‐type mouse (top) and one representative Sglt1^–/– mouse (bottom). PET images are displayed as in Fig. 4 A; stomach (st) and intestine (in) are indicated where visible. B and C, Me‐4FDG absorption (B) and excretion into urinary bladder (C) as a function of time. Data are for the mice shown in A; gastric emptying within the first 2 min was 52% and 60%, and at the end of the experiments was 60% and 80% in the wild‐type mouse and the Sglt1^–/– mouse, respectively. Similar results were obtained in two additional experiments each. These two experiments are shown in Videos S1 and S2.

Figure 8. Time course of 2‐FDG absorption and distribution in mice

2‐FDG (300 μCi) was administered by oral gavage into the stomach of anaesthetized wild‐type or Glut2^–/– mice, and continuous PET data were acquired over the following 60 min; a CT scan was performed at the end of the dynamic PET scan. A, volumetric renderings of co‐registered images at selected times after bolus administration, for one representative wild‐type mouse (top) and one representative Glut2^–/– mouse (bottom). PET images are displayed as in Fig. 4 A; stomach (st), intestine (in) and urinary bladder (bl) are indicated where visible. B and C, 2‐FDG absorption (B) and excretion into urinary bladder (C) as a function of time. Data are the mean ± SEM for five wild‐type and three Glut2^–/– mice. Gastric emptying at the end of the experiment was 40 ± 10% in wild‐type and 22 ± 10% in Glut2^{–/ –} mice (P = 0.28, Student's t test). PET images are displayed according to the NIH intensity scale for tracer activity, from red (highest), through green (intermediate) to purple (lowest).

Figure 9. Dynamic Me‐4FDG, 2‐FDG and 4‐FDG microPET scans of control mice following direct sugar delivery into the duodenum

Two hundred microcuries of Me‐4FDG (in 5 mm α‐MDG), 2‐FDG (in 5 mm 2‐DG) or 4‐FDG (in 5 mm glucose) was infused into the duodenum of anaesthetized mice, and PET data were continuously acquired for 60 min. A, coronal (c) and sagittal (s) PET images at selected times after bolus administration for three representative mice. Images are displayed as in Fig. 4 A, and intestine (in), brain (br), heart (h) and urinary bladder (bl) are indicated where visible. B and C, time course of PET tracer absorption (B) and excretion (C) for the mice shown in A. Results were confirmed in at least one additional experiment for each tracer.