The Journal of Nutrition Vol. 128 No. 10 October 1998, pp. 1786-1793

Fermentable Dietary Fiber Increases GLP-1 Secretion and Improves Glucose Homeostasis Despite Increased Intestinal Glucose Transport Capacity in Healthy Dogs^1,2,3

Stefan P. Massimino^*, Michael I. McBurney^{*, , 4}, Catherine J. Field^*, , Alan B. R. Thomson, Monika Keelan, Michael G. Hayek, and Gregory D. Sunvold

Nutrition & Metabolism Research Group, ^* Departments of Agricultural, Food and Nutritional Science and  Medicine, University of Alberta, Edmonton, Alberta, Canada T6G 2P5, and ^** Research and Development, The IAMS Co., Lewisburg, OH 45338

	ABSTRACT

Abstract Introduction Methods Results Discussion References

Ileal proglucagon gene expression and postprandial plasma concentrations of proglucagon-derived peptides are reported to change with the type and quantity of dietary fiber ingested by rats. Within the intestine, proglucagon encodes several proglucagon-derived peptides known to modulate intestinal absorption capacity and pancreatic insulin secretion. To determine whether the chronic ingestion of fermentable dietary fiber regulates the expression and synthesis of proglucagon-derived peptides in the distal intestine to modulate glucose homeostasis, the following study was conducted: 16 adult dogs (23 ± 2 kg) were fed isoenergetic, isonitrogenous diets containing a mixture of high fermentable dietary fibers (HFF) or low fermentable (LFF) wood cellulose for 14 d in a randomized cross-over design. Food was withheld for 16 h before an oral glucose tolerance test was conducted supplying 2 g of glucose/kg body wt, and peripheral blood was collected via a hind-leg catheter at 0, 15, 30, 45, 60, 90 and 120 min for plasma glucose, insulin and glucagon-like peptide-1(7-36)NH₂ (GLP-1) analyses. Intestinal samples were collected after the second dietary treatment. Ileal proglucagon mRNA, intestinal (GLP-1) concentrations and the integrated area under the curves (AUC) for plasma GLP-1 and insulin were greater and plasma glucose AUC was reduced when dogs were fed the HFF diet compared to the LFF diet (P < 0.05). Intestinal villi heights, brush border and basolateral glucose transporter protein abundance and jejunal transport capacities were significantly greater when dogs were fed the HFF diet than when fed the LFF diet. In conclusion, improvements in glucose homeostasis are observed in healthy dogs when they ingest fermentable fibers.

	INTRODUCTION

Abstract Introduction Methods Results Discussion References

The addition of dietary fibers to a meal results in decreased postprandial hyperglycemia and lower insulin requirements (Anderson et al. 1995, Jenkins et al. 1980). However, the chronic ingestion of diets enriched with dietary fiber has been associated with increased plasma insulin concentrations and reduced hyperglycemia to oral glucose tolerance tests (Pastors et al. 1991, Lovejoy and DiGirolama 1992, Groop et al. 1993).

Proglucagon is a 160-amino-acid polypeptide encoded by the glucagon gene (Bell et al. 1983) of L cells found in greatest concentrations in the mucosa of the distal ileum and large intestine (Holst 1997). Several highly conserved proglucagon-derived peptides have been identified, including glucagon-like peptide-1 (7-36)NH₂ (GLP-1)⁵ and glucagon-like peptide-2 (GLP-2) (Mojsov et al. 1987, Orskov et al. 1986). GLP-1 is considered an antidiabetogenic agent because it stimulates insulin secretion, inhibits glucagon secretion and delays gastric emptying (Holst, 1997). GLP-2 appears to be co-secreted with GLP-1 (Orskov et al. 1986) and modulates small bowel epithelial proliferation (Drucker et al. 1996) and basolateral glucose transporter (GLUT2) abundance (Cheeseman and Tsang 1996).

Ileal proglucagon gene expression in rats increases with the amount of fiber ingested (Reimer and McBurney 1996). Gee et al. (1996) reported that plasma concentrations of proglucagon-derived peptides increased with the ingestion of fermentable fibers and decreased with the removal of fermentable fiber from the diet. These observations are consistent with a positive effect of microbial fermentation of dietary fiber and subsequent production of short-chain fatty acids (SCFA) in the large intestine. The addition of SCFA to total parenteral nutrition formulations increased intestinal proglucagon mRNA, GLUT2 mRNA, mucosal mass and ileal uptakes of D-glucose in rats following 80% small bowel resection (Tappenden et al. 1997). Recently, Reimer et al. (1997) reported that proglucagon mRNA abundance was increased in rats fed diets containing fermentable fiber sources.

Based on the abovementioned work, we hypothesized that the consumption of a high fermentable fiber (HFF) diet would increase intestinal proglucagon mRNA abundance. This would be associated with an increase in postprandial secretion of GLP-1 and insulin, as well as an increase in intestinal glucose transport capacity, but the net effect would be improved glucose homeostasis.

	SUBJECTS AND METHODS

Abstract Introduction Methods Results Discussion References

Diets.  Experimental diets were formulated to be isonitrogenous and isoenergetic and to provide ~19.5 MJ/kg of diet with 35% of the energy from carbohydrate, 30% from fat and 35% from protein (Table 1). The low fermentable fiber (LFF) diet contained wood cellulose as the fiber source, and the HFF diet contained a mixture of more fermentable plant fibers (beet pulp, Michigan Sugar, Saginaw MI; gum arabic, TIC Gums, Belcamp, MD; fructooligosaccharides, Golden Technologies Corporation, Golden, CO). These diets were chosen because cellulose is relatively inert within the intestine and has an in vitro fermentability of 9 mmol of SCFA/kg of fiber organic sources, whereas the mixture of soluble and insoluble fiber sources found within the HFF diet has an in vitro fermentability of 229 mmol of SCFA/kg fiber organic matter (Sunvold et al. 1995a-c).

Animals.  All procedures received ethical approval from the Health Sciences Animal Welfare Committee of the University of Alberta and are consistent with the guidelines of the Canadian Council on Animal Care.

Adult mongrel dogs (n = 16) were used. Upon arrival, they were acclimated for 7 d and fed a nutritionally complete diet (Can-Pro, Beaumont AB, Canada). All dogs were weighed daily and individually fed to meet energy requirements using the formula: energy intake (MJ) = 0.553 × kg kg body weight (BW)^0.67 (NRC 1985). Food was offered once daily between 0900-1000 h and water was consumed ad libitum. A crossover experimental design was used whereby dogs were randomly assigned to receive the HFF or LFF diet for 14 d followed by the alternate diet for an additional 14 d. Because the 16 dogs could not be accommodated at one time, dogs were paired throughout the experiment.

Oral glucose tolerance test (OGTT).  Food was removed at 1600 h on d 13 and 27. At 0845-0900 h on d 14 and 28, the dogs were loosely restrained in a table sling, and an OGTT was conducted using 700 g/L of glucose solution in water to provide 2 g of glucose/kg of BW. Peripheral blood was sampled at 0, 15, 30, 45, 60, 90 and 120 min via a catheter in the saphenous vein.

Peripheral blood samples.  Blood samples for general chemistry screen and complete blood counts were stored on ice until analysis. Hematological analyses were conducted using a Coulter STKS instrument (Coulter Electronics Inc., Hialeah, FL), and manual differential counts were performed by the staff at the Veterinary Pathology Laboratory (Edmonton AB, Canada). Blood samples for insulin and GLP-1 analyses were collected in EDTA-heparinized tubes with aprotinin [5 × 10⁵ kallikrein inhibitor units (KIU/L of blood; Sigma Chemicals, St. Louis, MO)] and stored at 70 (GLP-1) or 35°C (insulin). Blood samples for serum glucose determinations were placed in 250-µL microcentrifuge tubes and centrifuged at 2900 × g for 10 min at room temperature; the serum was removed by pipet and stored at 35°C.

Intestinal samples.  On d 28, the dogs were anesthetized by intravenous injection of somnitol (MTC Pharmaceuticals, Cambridge ON, Canada) using 1 mL/2.27 kg of BW via the saphenous catheter subsequent to the OGTT. Intestinal samples were taken for northern blot analysis and immediately placed in liquid nitrogen. Jejunal and ileal samples for nutrient uptake assays were placed in ice-cold saline, and assays were performed within 30 min of sampling. Jejunal and ileal segments were scraped to obtain mucosal samples for western blot analyses. Histological samples were placed directly into formalin and slides were prepared by staff at the Veterinary Pathology Laboratory.

Glucose.  Serum glucose was determined using the Sigma Diagnostics Glucose (Trinder) Reagent for the enzymatic determination of glucose at 505 nm (Sigma Chemicals).

Insulin.  Serum insulin concentrations were determined using the Coat-A-Count® I¹²⁵ diagnostic radioimmunoassay (RIA, Diagnostics Products Corporation, Los Angeles, CA).

Plasma GLP-1(7-36)NH₂ extraction.  GLP-1 immunoreactive peptides were extracted as described by Reimer and McBurney (1996). Samples were lyophilized overnight and stored at 70°C.

Intestinal GLP-1(7-36)NH₂ extraction.  Extraction of GLP-1(7-36)NH₂ from intestinal segments has been described by Xiaoyan (1996) and was carried out with modifications. Briefly, intestinal segments (jejunum, ileum and colon) were boiled with 2 mol/L of acetic acid for 1 h and then centrifuged at 4500 × g for 10 min. The supernatant was transferred and neutralized with 1 mol/L of NaOH. For RIA purposes, the sample of supernatant was diluted 1:10 with RIA buffer (100 mmol/L of Tris, 50 mmol/L of NaCl, 200 mmol/L of Na₂-EDTA, 0.2 g/L of Na azide, pH 8.5).

GLP-1(7-36)NH₂ RIA.  Concentrations of GLP-1(7-36)NH₂ were measured using a competitive binding RIA described by Xiaoyan (1996) with modifications. Briefly, the lyophilized plasma samples were reconstituted in 250 µL of RIA assay buffer (100 mmol/L of Tris, 50 mmol/L of NaCl, 20 mmol/L of Na₂-EDTA, 0.2 g/L of Na azide, pH 8.5). Polypropylene tubes (12 × 75 mm) were used for control, standard and sample, and the entire procedure was carried out on ice. GLP-1 (7-36)NH₂ standards (Peninsula Laboratories, Belmont, CA) made from serial dilutions ranged from 4000 to 15 ng/L. Tubes were mixed and incubated 24 h at 4°C. Following incubation, 50 Bq of ¹²⁵I-GLP-1(7-36)NH₂ tracer was added to the tubes; the tubes were mixed by vortexing and incubated for 48 h at 4°C. Dextran-charcoal suspension (4 g/L of dextran T70, 80 g/L of charcoal in assay buffer) was added to all tubes (100 µL) except total count (TC) tubes. Tubes were mixed by vortexing and left on ice for 15 min, centrifuged at 2200 × g for 30 min, and 600 µL of supernatant was transferred to new tubes that were counted using a Cobra^TM Auto-Gamma counter (Packard Instrument Company, Downers Grove, IL).

GLP-1(7-36)NH₂ iodination.  GLP-1 (7-36)NH₂ was iodinated using the chloramine-T method as described by Xiaoyan (1996). Briefly, 30-40 µg of GLP-1 (7-36) NH₂ was dissolved in 30-40 µL of double-distilled H₂O, and 10 µL of aliquots was added to 10 µL of 0.5 mol/L PO₄ (pH 7.0) followed by 10 µL of 18.5 kBq ¹²⁵I Chloramine-T. The tube was tapped for exactly 30 s, and the reaction was stopped by the addition of 10 µL of sodium metabisulfite (5 g/L) followed by 1 mL of 1 mL/L trifluoroacetic acid (TFA). A Sep-Pak C-18 cartridge (Waters, Milford, MA) was used to separate the ¹²⁵I-GLP-1(7-36)NH₂. The cartridge was primed with 10 mL of acetonitrile with 1 mL/L of TFA followed by 10 mL of 0.1 mL/L of TFA, and finally a 10-mL bolus of air. The ¹²⁵I-GLP-1 (7-36)NH₂ was eluted with five washes of 100-600 mL/L acetonitrile mixed with 1 mL/L of TFA using the following volumes and acetonitrile concentrations: i) 5 mL of 100 mL/L of acetonitrile, ii) 5 mL of 200 mL/L of acetonitrile, iii) four washes with 1 mL of 300 mL/L of acetonitrile, iv) one wash of 380 mL/L of acetonitrile, and finally v) five washes with 1 mL of 400 mL/L of acetonitrile). Each eluted fraction was mixed well, and a 10-µL aliquot was counted using a Cobra^TM Auto-Gamma counter. The label was usually eluted in fraction 1, 2 and/or 3 of the 400 mL/L of acetonitrile. Fractions containing the labeled GLP-1(7-36)NH₂ were pooled and stored at 35°C. The ¹²⁵I-GLP-1(7-36)NH₂ has a storage life of approximately 2 wk.

Isolation of total RNA.  Total RNA was isolated from each intestinal segment using Trizol^TM (Gibco BRL, Burlington, ON, Canada) according to the protocol provided by the manufacturer. Quantity and purity of RNA were determined by ultraviolet spectrophotometry at 260, 280 and 230 nm.

Northern blot analysis.  Messenger RNA was measured by northern blot analysis as described by Reimer et al. (1997) with the following exceptions. After electrophoresis, the gels were soaked in two changes of 10× standard saline citrate (SSC) (1.5 mmol/L of NaCl, 0.15 mmol/L of trisodium citrate, pH 7.0) and blotted onto a zeta-probe GT Genomi tested blotting membrane (BioRad, Mississauga, ON, Canada). The RNA was fixed onto membranes by baking in vacuum at 80°C for 2 h. Membranes were prehybridized for 2 h at 50°C and hybridized for 12-16 h at 50°C with the addition of 16.7 kBq (1 × 10⁶ cpm) of [³²P] CTP-labeled. First, the membranes were washed with 2× SSC at room temperature for 5 min and then in 2× SSC with 1 g of sodium dodecyl sulfate (SDS)/L for either 10 min (GLUT2) or 15 min (proglucagon, SGLT-1). The membranes were transferred to a bath of 0.2× SSC with 10 g of SDS/L as follows: proglucagon (70°C for 10 min), SGLT-1 (70°C for 20 min) and GLUT2 (60°C for 2-3 min). Lastly, the membranes were washed in 0.2× SSC at room temperature for 2-3 min. Membranes were heat sealed in plastic bags and exposed to Kodak XRA5 film (Eastman Kodak, Rochester, NY) at 70°C using an intensifying screen (Dupont Canada, Mississauga, ON, Canada). For statistical analysis, the signals were quantified using laser densitometry (Model GS-670 Imaging Densitometer, BioRad. The 28 and 18S ribosomal bands were quantified from negatives of photographs of the membranes and were used to confirm the integrity of the RNA and compensate for minor-loading discrepancies.

Riboprobes.  A 3.8-kb radiolabeled GLUT2 antisense riboprobe was generated from Xba I-linearized plasmid DNA [pGEM4Z-HTL-3] and T7 polymerase. The 350-bp proglucagon sense riboprobe was generated from Rsa I-linearized plasmid DNA [pGEM4Z-HTL-3] and Sp6 polymerase. Lastly, the 2.1-kb SGLT-1 antisense riboprobe was generated from a 1.4-kb fragment of lamb intestinal SGLT-1 clone (aa 207-664) (Wood et al. 1994).

Brush border membrane (BBM) and basolateral membrane (BLM) isolation, preparation and enrichment.  All procedures were performed on ice using previously described procedures for the simultaneous isolation of BLM and BBM (Tappenden and McBurney, in press). Briefly, mucosal scrapings were homogenized in mucosal membrane suspension solution (MSS buffer, 125 mmol/L of sucrose, 1 mmol/L of Tris-HCl, 0.05 mmol/L of phenyl methyl sulfonyl fluoride buffer, pH 7.4) with a Polytron homogenizer (Brinkmann Instruments Canada, Mississauga, ON, Canada) for 30 s at setting 8. Aliquots of this homogenate were taken for enrichment assays, whereas the remainder was homogenized twice more at setting 8 for 30 s and centrifuged for 15 min at 2,400 × g with no brake. The pellet (unbroken cells, nuclear material) was discarded, and the supernatant was centrifuged at 43,700 × g for 20 min with no brake. The resulting supernatant was discarded. The upper white fluffy pellet (P1) comprised the BLM and the inner dark brown pellet (P2) comprised the BBM. Both P1 and P2 were gently resuspended in MSS buffer. The lower pellet (P2) was resuspended in MSS buffer and centrifuged for 20 min at 43,700 × g with no brake. The supernatant was discarded; the fluffy white pellet was resuspended in MSS buffer and combined with P1. The lower dark pellet was resuspended in MSS and combined with P2. The combined P1 suspension was homogenized for 15 s at setting 8, layered onto a 20-g/L Percoll® (Sigma, Toronto, ON, Canada) gradient and centrifuged for 30 min at 46,000 × g with maximal brake. The resulting white fluffy BLM layer was removed, resuspended in MSS buffer and centrifuged at 115,000 × g for 30 min with maximal brake. The membrane layer was removed, resuspended in MSS buffer and homogenized for 15 s with at setting 8. One mole of CaCl₂ was added to a final concentration of 10 mmol/L and the homogenate was gently stirred on ice for 10 min. The mixture was centrifuged for 10 min at 7,700 × g. The supernatant was discarded; the pellet resuspended in MSS buffer and homogenized for 15 s at setting 8. Samples were centrifuged for 20 min at 46,000 × g. The supernatant was discarded and the final BLM pellet was resuspended in MSS buffer. Aliquots were then taken for marker enrichment assays.

The combined P2 suspension (BBM) was homogenized for 15 s at setting 8 and centrifuged for 10 min at 1,900 × g. The pellet was discarded and the supernatant was centrifuged 15 min at 14,600 × g. The pellet was discarded. To the supernatant, sufficient 1 mmol/L of CaCl₂ was added to a final concentration of 10 mmol/L of CaCl₂ and the homogenate was stirred gently on ice for 20 min. The mixture was centrifuged for 30 min at 3,000 × g with maximal brake. The supernatant was then centrifuged for 30 min at 46,000 × g and the resulting BBM pellet was resuspended in distilled, deionized water.

Membrane purity was confirmed through a 10-20-fold enrichment of BBM and BLM markers. The enrichment assay described by Esmann (1988) was used for the BLM enzyme Na⁺K⁺-ATPase. The enrichment assay for the BBM enzyme alkaline phosphatase was measured using an alkaline phosphatase kit (Sigma Diagnostics, St. Louis, MO).

Western blot analysis.  BBM and BLM proteins were resolved by 100 g/L of SDS-polyacrylamide gel electrophoresis at 100-200 V for 1-2 h as described by Tappenden and McBurney (in press). Briefly, proteins were transferred onto a nitrocellulose membrane (MSI Laboratories, Houston, TX). Nonspecific binding was blocked by immersing membranes in 50 g/L of nonfat dry milk in Tris-buffered saline with Tween 20 (TBST: composed of 20 mmol/L of Tris, 137 mmol/L of NaCl, 0.5 g/L of Tween-20, pH 7.5) for 1 h with gentle agitation, and then incubated with primary antibodies to SGLT-1 (Chemicon International, Temecula, CA) at a dilution of 1:1000 or GLUT2 (Chemicon International) at a dilution of 1:500 overnight at 4°C. Membranes were washed 3 × 10 min in TBST with gentle agitation, and then incubated with the secondary antibody (anti-rabbit IgG HRP-conjugate, Signal Transduction, PDI Bioscience, Aurora, ON, Canada) at a dilution of 1:4000 for at least 2 h with gentle agitation. Blots were covered completely with Supersignal CL-HRP^TM (Pierce, Rockford, IL) working solution and incubated for 5 min before being exposed to Kodak XRA5 film. Loading consistency and protein transfer were confirmed by staining the blots with Ponceau S [1 g/L of Ponceau S (BDH), 837 mmol/L of acetic acid]. Statistical analysis was performed on the relative intensities of the bands. For statistical analysis, the signals were quantified using laser densitometry (Model GS-670 Imaging Densitometer, BioRad).

Measurement of transport kinetics.  Transport kinetics were measured as previously described by Thomson and Rajotte (1983). Briefly, a 12-cm segment of intestine was removed from each animal, opened along the mesenteric border and carefully washed with ice-cold saline to remove visible mucus and debris. Pieces of intestine (1 cm²) were cut out and the tissue was mounted as flat sheets in incubation chambers containing oxygenated Kreb's bicarbonate buffer (pH 7.4) at 37°C. Tissue discs were preincubated in this buffer for 15 min to allow equilibration at this temperature. After preincubation, the chambers were transferred to beakers containing [³H]-inulin and various [¹⁴C]-probe molecules in oxygenated Kreb's bicarbonate buffer (pH 7.4) at 37°C. The concentration of solutes was 4, 8, 16, 32 and 64 mmol/L for D-glucose and D-fructose, 16 mmol/L for L-glucose and 0.1 mmol/L for lauric acid. The preincubation and incubation solutions were mixed using circular magnetic bars which were adjusted with a strobe light to achieve a stirring rate of 600 rpm and a slightly effective resistance of the intestinal unstired water layer (Thomson and Dietschy 1980). The experiment was terminated by removing the chambers, quickly rinsing the tissue in cold saline for approximately 5 s and cutting the exposed mucosal tissue from the chamber with a circular steel punch. The tissue was dried overnight in an oven at 55°C to determine the dry weight of the tissue and then saponified with 0.75 mmol/L of NaOH. Scintillation fluid (Beckman Ready Solv HP, Toronto, ON, Canada) was added to the sample and radioactivity determined using an external standardization technique to correct for variable quenching of the two isotopes (Beckman Beta LS-5801, Beckman Instruments, Mountain View, CA). The uptake of nutrients was expressed as nmol·100 mg dry tissue¹·min.

Villi height and crypt depth measurements.  Intestinal segments were cut into sections at the Veterinary Pathology Laboratory. Intestinal villi height and crypt depths were measured under a light microscope using Northern Exposure Image Analysis software (Empix Imaging, Mississauga, ON, Canada). Ten recordings were made for each animal and each segment, with the average used for statistical analysis.

Statistical analyses.  All statistical analyses were performed using the Statistical Analysis System (SAS) statistical package (version 6.10, SAS Institute, Cary, NC). For proglucagon and SGLT-1 mRNA abundance, and SGLT-1 and GLUT2 transporter abundance, data were analyzed using the general linear models procedure (proc GLM), and the model included diet, gel, period, pair and diet × period. Both period and diet × period were found to be nonsignificant and subsequently excluded. Villi height, crypt depth and intestinal GLP-1 concentrations were analyzed using proc GLM and the model included diet, pair, diet × period and period. Again both period and diet × period were nonsignificant and were excluded. Plasma area under the curves (AUC) for GLP-1, insulin and glucose were analyzed using paired t-tests within proc GLM. Repeated measures of analysis of variance were used to test for differences between animal weights. Intestinal transport rates for D-glucose, L-glucose, D-fructose and lauric acid were analyzed using paired t-tests within proc GLM. Data presented are means ± SEM. Significant differences were identified at P < 0.05.

	RESULTS

Abstract Introduction Methods Results Discussion References

Effect of diet on body weight.  Energy requirements were individually calculated, and dietary portions were adjusted accordingly so that dog weights did not differ by experimental diet (23.4 ± 1.8, 22.9 ± 1.8 and 23.5 ± 1.8 kg for pre-experimental, HFF and LFF, respectively) or by period [23.4 ± 1.8, 23.4 ± 1.8 and 23.4 ± 1.8 kg for pre-experimental (d 7), period 1 (d 21), and period 3 (d 35), respectively].

Effect of OGTT on plasma GLP-1, insulin and glucose.  Plasma GLP-1 concentrations were greater at 30 and 90 min when dogs ingested the HFF diet compared to the LFF diet (Fig. 1A). Insulin concentrations were greater at 90 min when dogs consumed the HFF diet rather than the LFF diet (Fig. 1B). Dietary fiber type did not influence blood glucose concentrations at any time points during the OGTT (Fig. 1C). The incremental AUC was significantly higher for GLP-1 (Fig. 2A) and insulin (Fig. 2B) when dogs ate the HFF diet. The AUC for glucose was significantly lower when dogs consumed the HFF diet (Fig. 2C).

View larger version (22K): [in this window] [in a new window]   Fig 1. Effect of fermentable fiber on plasma glucagon-like peptide-1 (GLP-1) (A), insulin (B) and glucose (C) concentrations at 0, 15, 30, 45, 60, 90 and 120 min after an oral glucose load (2 g/kg body weight) to dogs which consumed a high fermentable fiber diet (HFF) and a low fermentable fiber diet (LFF) for 14 d. Values are means ± SEM, n = 13 (glucose), n = 14 (insulin and GLP-1). * Significant difference between diets at specific sampling times (P < 0.05).

View larger version (28K): [in this window] [in a new window]   Fig 2. Incremental area under the curve for plasma (A) glucagon-like peptide-1 (GLP-1), insulin (B) and glucose (C) at 0, 15, 30, 45, 60, 90 and 120 min after an oral glucose load (2 g/kg body weight) to dogs which consumed a high fermentable fiber diet (HFF) and a low fermentable fiber diet (LFF) for 14 d. Values are means ± SEM. n = 13 (glucose) and n = 14 diet (insulin and GLP-1). Bars with different letters are significantly different (P < 0.05).

Effect of diet on intestinal proglucagon and GLP-1 concentration.  The ingestion of HFF diets was associated with greater proglucagon mRNA abundance in the ileum and the colon of dogs (Fig. 3). Proglucagon mRNA expression was not detected in the duodenum. GLP-1 concentrations were significantly greater in mucosal scrapings from dogs consuming the HFF diet compared to those consuming the LFF diet (41 ± 4 pmol GLP-1/mg of protein vs. 25 ± 4 pmol GLP-1/mg of protein, P < 0.05).

View larger version (22K): [in this window] [in a new window]   Fig 3. Effect of fermentable fiber on intestinal proglucagon mRNA of dogs which consumed a high fermentable fiber diet (HFF) and a low fermentable fiber diet (LFF) for 14 d. Values are means ± SEM, n = 16. Bars with different letters are significantly different (P < 0.05). Bars are not comparable between intestinal segments. Each lane was loaded with 15 µg of total RNA.

Histology.  Duodenal villi height tended (P = 0.1) to be higher in dogs consuming the HFF diet compared to those eating the LFF diet (1505 ± 83 vs. 1294 ± 83 µm), but there were no differences in duodenal crypt depth (289 ± 28 vs. 262 ± 28 µm). Jejunal villi height was significantly higher in dogs consuming the HFF diet (1517 ± 43 µm) rather than those consuming the LFF diet (1343 ± 43 µm), but no significant differences were found in crypt depth (277 ± 19 vs. 234 ± 19 µm). Ileal villi height and crypt depth were not significantly different between dogs consuming the HFF diet compared to those eating the LFF diet (1035 ± 45 vs. 993 ± 45 µm and 251 ± 46 vs. 357 ± 46 µm, respectively). Mean colonic crypt depth (726 ± 33 µm) was not affected by diet.

Nutrient uptake.  Consumption of the HFF diet was associated with a significantly greater maximal glucose uptake capacity (V_max) for D-glucose in the jejunum (Fig. 4 and Table 2). A significant diet effect was also noted in lauric acid uptake in the jejunum (Table 2). The Michaelis affinity constant (K_m) was unaffected by diet (Table 2). The estimation of paracellular D-glucose uptake and the K_d for D-glucose, as determined by extrapolation of L-glucose uptakes at 16 mmol/L through the origin and normalizing to 1 mmol/L, were not significantly affected by diet. K_d for D-fructose was unaffected by diet.

View larger version (17K): [in this window] [in a new window]   Fig 4. Effect of fermentable fiber on the in vitro uptake of D-glucose into the jejunum (A) and ileum (B) of dogs which consumed a high fermentable fiber diet (HFF) and a low fermentable fiber diet (LFF) for 14 d. Values are means ± SEM, n = 8. * Significant difference between diets at a specific D-glucose concentration (P < 0.05).

Glucose transporters.  Diet did not affect SGLT-1 mRNA abundance in any of the intestinal segments measured. The consumption of the HFF diet was associated with greater jejunal SGLT-1 transporter abundance which only tended (P = 0.09) to be greater in the ileum of dogs when consuming the HFF diet (Fig. 5). GLUT2 transporter abundance was greater in both the jejunum and ileum of dogs when they consumed the HFF diet (Fig. 6).

View larger version (24K): [in this window] [in a new window]   Fig 5. Effect of fermentable fiber on jejunal (A) and ileal (B) SGLT-1 transporter abundance of dogs which consumed a high fermentable fiber diet (HFF) and a low fermentable fiber diet (LFF) for 14 d. Values are means ± SEM. n = 7 (ileum) and n = 6 (jejunum). Bars with different letters are significantly different (P < 0.05). Each well was loaded with 60 µg of total protein and membranes were stained with ponceau S to confirm equal protein loading.

View larger version (24K): [in this window] [in a new window]   Fig 6. Effect of fermentable fiber on intestinal basolateral glucose transporter abundance in jejunum (A) and ileum (B) of dogs which consumed a high fermentable fiber diet (HFF) and a low fermentable fiber diet (LFF) for 14 d. Values are means ± SEM, n = 6 (jejunum, ileum HFF) and n = 7 (ileum LFF). Bars with different letters are significantly different (P < 0.05). Each well was loaded with 60 µg of total protein and membranes were stained with ponceau S to confirm equal protein loading.

	DISCUSSION

Abstract Introduction Methods Results Discussion References

The ingestion of high-fiber diets modulates proglucagon gene expression and plasma GLP-1 concentration 30 min after an OGTT in rats (Reimer and McBurney 1996). Recently, Reimer et al. (1997) reported that proglucagon mRNA abundance was increased in rats fed diets containing fermentable fiber sources. We report that the ingestion of fermentable fiber per se, in diets containing similar amounts of total dietary fiber, is associated with increased ileal proglucagon gene expression, increased intestinal glucose transport capacity, increased secretion of GLP-1 and insulin and improved glucose homeostasis following an oral glucose meal in dogs.

The GLP-1 antibody used in the RIA in this study is specific for the c-terminally amidated GLP-1 isomers but does not distinguish between intestinal GLP-1(7-36)NH₂, the enzymatically-degraded form GLP-1(9-36)NH₂ or pancreatic GLP-1(1-36)NH₂. However, of these GLP-1 isomers, only GLP-1(7-36)NH₂ is insulinotropic (Suzuki et al. 1989) so we can assume that GLP-1(7-36)NH₂ secretion was increased. The canine plasma GLP-1 concentrations measured in this study are similar to those previously reported in dogs (Wen et al. 1995) and humans (Vaag et al. 1996).

GLP-1 has been termed an antidiabetogenic agent because it is a potent insulin secretogogue (Holst, 1997). However, GLP-1 also inhibits gastric emptying (Willms et al. 1996). In this study, we demonstrate that fiber fermentability, independent of changes in fiber intake, modulates intestinal proglucagon mRNA expression and GLP-1 secretion. Postprandial GLP-1 secretion was significantly greater when dogs consumed the HFF diet rather than the LFF diet.

The mechanism whereby diet modulates proglucagon expression and GLP-1 secretion remains unknown. The L-cell has a pyramid shape with the apical process found in the microvilli of the intestinal lumen, and the base, rich in endocrine granules, is located near the basal lamina. This led to the conclusion that the cell responds to luminal contents with a basal discharge of granular contents (Holst 1997). However, it is unknown if the L-cell responds directly to nutrients and absorption or if other signals are involved. Colonic infusions of various fibers and SCFA do not affect GLP-1 release in food-deprived rats (Plaisancie et al. 1995), but increased plasma GLP-1 concentrations were observed within 6 h of systemic SCFA administration in rats receiving parenteral nutrition (Tappenden 1997). In this study, we did not determine whether the absolute number of L-cells was increased or if proglucagon expression and GLP-1 production per L-cell changed. However, Hoyt et al. (1996) reported that proglucagon mRNA per cell, as determined by in situ hybridization, increased in food-restricted rats after refeeding. This observation suggests increased proglucagon expression and GLP-1 secretion per L-cell.

Long-term consumption of fiber-supplemented diets is associated with changes in intestinal mass, length and villus appearance as well as cell proliferation rates and enterocyte migration along the crypt-villus axis (see review by McBurney 1994). When dogs consumed the HFF diet, they had significantly longer jejunal villi and increased D-glucose transport capacity. Passive diffusion, as measured by L-glucose, requires nonprotein-mediated movement across the unstirred water layer and the BBM. The passive diffusion of L-glucose was unaffected by diet, although the unstirred water layer resistance, as measured by lauric acid uptake, was significantly greater in the jejunum of dogs when they consumed the HFF diet. This suggests that the biological importance of the changes in unstirred water layer resistance is marginal in terms of carbohydrate diffusion to the vicinity of enterocyte brush border transporters. Protein-mediated transport is predominantly altered by changes in V_max, and alterations in the glucose absorption rate are usually the result of changes in the number of transport sites per enterocyte, arising from changes in the rates of synthesis or degradation (see review by Ferraris and Diamond 1997). Changes in V_max may also be achieved by increasing the number of cells capable of transporting D-glucose, as may occur when there is an increase in the height of the villi (Fig. 4). The longer villi of the jejunum and the greater rate of in vitro D-glucose transport per 100 mg tissue suggest a synergistic effect on the in vivo uptake of D-glucose. Indeed, consumption of the HFF diet was associated with increased jejunal SGLT-1 mRNA abundance and increased GLUT2 glucose transporter concentrations in the jejunum and ileum. Thus, both transporter quantity and activity were greater when dogs consumed the HFF diet compared to when they ate the LFF diet.

Increased availability of SCFA facilitates functional adaptation by increasing total intestinal mass, mucosal and submucosal weight, and increases ileal DNA, RNA and protein concentrations (Koruda et al. 1988, Tappenden et al. 1996). Intravenous supplementation of SCFA significantly increases ileal proglucagon expression and D-glucose uptake rates in parenterally-fed rats (Tappenden et al. 1997). GLP-2 is synthesized from the proglucagon gene along with GLP-1 (Orskov et al. 1986) and regulates small bowel epithelial proliferation and basolateral GLUT2 glucose transporter abundance (Cheeseman and Tsang 1996, Drucker et al. 1996). Since postprandial serum GLP-1 concentrations were significantly increased when dogs consumed the HFF diet, concomitant changes in GLP-2 secretion, although not measured, may explain the greater glucose transporter abundance and V_max measured in dogs adapted to the HFF diet.

Medical nutrition therapies for individuals with type 2 diabetes mellitus are often augmented with oral medications to stimulate pancreatic insulin secretion or the use of exogenous insulin to supplement residual pancreatic function. This study suggests that the ingestion of fermentable fiber is associated with increased synthesis and secretion of intestinal GLP-1, a known insulin secretogogue, in response to a standardized glucose load. The observed improvements in glucose homeostasis in healthy dogs consuming a HFF diet suggest that insulin secretion and/or tissue sensitivity to insulin was increased to compensate for the greater glucose transport capacity of the intestine. D'Alessio et al. (1995) reported that GLP-1 increased both insulin-dependent and independent glucose disposition, but the most accepted interpretation would favor changes in insulin secretion over increased insulin sensitivity (Holst 1997). Indeed, we did see evidence of increased insulin secretion in the dogs when eating HFF diets. However, GLP-1 inhibits gastric emptying (Willms et al. 1996) which might slow glucose delivery to the small intestine and subsequent transport across the gut into the systemic circulation. Thus, GLP-1-mediated actions on gastric emptying, and therefore glucose presentation to intestinal transporters and ultimately glucose homeostasis cannot be excluded from this study. Despite this ambiguity, the ingestion of fermentable fibers by healthy dogs for 14 d is associated with greater GLP-1 secretion and smaller oscillations in postprandial blood glucose concentrations. The relative importance of GLP-1 action on gastric emptying, pancreatic insulin secretion, and peripheral insulin sensitivity in terms of glucose homeostasis remains to be determined.

	FOOTNOTES

Manuscript received 16 January 1998. Initial reviews completed 6 March 1998. Revision accepted 10 June 1998.

	ACKNOWLEDGMENT

The GLP-1(7-36)NH₂Ab (KMJ-03) (1:20000) was a generous gift from Dr. Chris McIntosh (University of British Columbia, Vancouver, British Columbia, Canada).

	LITERATURE CITED

Abstract Introduction Methods Results Discussion References

• Anderson J. W., O'Neal D. S., Riddell-Mason S., Floore T. L., Dillon D. W., Oeltgen P. R. Postprandial serum glucose, insulin, and lipoprotein responses to high- and low-fiber diets. Metab. Clin. Exp. 1995; 44:848-854
• Bell G. I., Sanchez-Pescador R., Laybourn P. J., Najarian R. C. Exon duplication and divergence in the human preproglucagon gene. Nature 1983; 304:368-371[Medline]
• Cheeseman C. I., Tsang R. The effect of GIP and glucagon-like peptides on intestinal basolateral membrane hexose transport. Am. J. Physiol. 1996; 271:G477-482[Abstract/Free Full Text]
• Chiou P. W., Yu B., Lin C. Effect of different components of dietary fiber on the intestinal morphology of domestic rabbits. Comp. Biochem. Physiol. 1994; 108:629-638
• D'Alessio D. A., Prigeon R. L., Ensinck J. W. Enteral enhancement of glucose disposition by both insulin-dependent and insulin-independent processes. A physiological role of glucagon-like peptide I. Diabetes 1995; 44:1433-1437[Abstract]
• Drucker D. J., Erlich P., Asa S. L., Brubaker P. L. Induction of intestinal epithelial proliferation by glucagon-like peptide 2. Proc. Nat. Acad. Sci. USA 1996; 93:7911-7916[Abstract/Free Full Text]
• Esmann M. Solubilization of Na⁺, K⁺-ATPase. Methods Enzymol. 1988; 156:72-79[Medline]
• Ferraris R. P., Diamond J. Regulation of intestinal sugar transport. Physiol. Rev. 1997; 77:257-302[Abstract/Free Full Text]
• Gee J. M., Lee-Finglas W., Wortley G. W., Johnson I. T. Fermentable carbohydrates elevate plasma enteroglucagon but high viscosity is also necessary to stimulate small bowel mucosal cell proliferation in rats. J. Nutr. 1996; 126:373-379
• Groop P. H., Aro A., Stenman S., Groop L. Long-term effects of guar gum in subjects with non-insulin-dependent diabetes mellitus. Am. J. Clin. Nutr. 1993; 58:513-518[Abstract/Free Full Text]
• Holst J. J., Orskov C., Nielsen O. V., Schwartz T. W. Truncated glucagon-like peptide I, an insulin-releasing hormone from the distal gut. FEBS Lett. 1987; 211:169-174[Medline]
• Holst J. J. Enteroglucagon. Annu. Rev. Physiol. 1997; 59:257-271[Medline]
• Hoyt E. C., Lund P. K., Winesett D. E., Fuller C. R., Ghatei M. A., Bloom S. R., Ulshen M. H. Effects of fasting, refeeding, and intraluminal triglyceride on proglucagon expression in jejunum and ileum. Diabetes 1996; 45:434-439[Abstract]
• Jenkins D. J., Wolever T. M., Taylor R. H., Ghafari H., Jenkins A. L., Barker H., Jenkins M. J. Rate of digestion of foods and postprandial glycaemia in normal and diabetic subjects. Br. Med. J. 1980; 281:14-17
• Koruda M. J., Rolandelli R. H., Settle R. G., Zimmaro D. M., Rombeau J. L. Effect of parenteral nutrition supplemented with short-chain fatty acids on adaptation to massive small bowel resection. Gastroenterology 1988; 95:715-720[Medline]
• Lovejoy J., DiGirolamo M. Habitual dietary intake and insulin sensitivity in lean and obese adults. Am. J. Clin. Nutr. 1992; 55:1174-1179[Abstract/Free Full Text]
• McBurney M. I. The gut: central organ in nutrient requirements and metabolism. Can. J. Physiol. Pharmacol. 1994; 72:260-265[Medline]
• Mojsov S., Weir G. C., Habener J. F. Insulinotropin: glucagon-like peptide I (7-37) co-encoded in the glucagon gene is a potent stimulator of insulin release in the perfused rat pancreas. J. Clin. Invest. 1987; 79:616-619
• National Research Council (1985) Nutrient requirements for dogs. National Academy Press. Washington, D.C.
• Orskov C., Holst J. J., Knuhtsen S., Baldissera F. G., Poulsen S. S., Nielsen O. V. Glucagon-like peptides GLP-1 and GLP-2, predicted products of the glucagon gene, are secreted separately from pig small intestine but not pancreas. Endocrinology 1986; 119:1467-1475[Abstract]
• Pastors J. G., Blaisdell P. W., Balm T. K., Asplin C. M., Pohl S. L. Psyllium fiber reduces rise in postprandial glucose and insulin concentrations in patients with non-insulin-dependent diabetes. Am. J. Clin. Nutr. 1991; 53:1431-1435[Abstract/Free Full Text]
• Plaisancie P., Dumoulin V., Chayvialle J. A., Cuber J. C. Luminal glucagon-like peptide-1(7-36) amide-releasing factors in the isolated vascularly perfused rat colon. J. Endocrinol. 1995; 145:521-526[Abstract]
• Reimer R. A., McBurney M. I. Dietary fiber modulates intestinal proglucagon messenger ribonucleic acid and postprandial secretion of glucagon-like peptide-1 and insulin in rats. Endocrinology 1996; 137:3948-3956[Abstract]
• Reimer R. A., Thomson A.B.R., Rajotte R., Basu T. K., Oorakuil B., McBurney M. I. A physiological level of rhubarb fiber affects proglucagon gene expression and intestinal glucose uptake in rats. J. Nutr. 1997; 127:1923-1928[Abstract/Free Full Text]
• Sunvold G. D., Fahey G. C. Jr., Merchen N. R., Reinhart G. A. In vitro fermentation of selected fibrous substrates by dog and cat fecal inoculum: influence of diet composition on substrate organic matter disappearance and short-chain fatty acid production. J. Anim. Sci. 1995a; 73:1110-1122[Abstract]
• Sunvold G. D., Fahey G. C. Jr., Merchen N. R., Titgemeyer E. C., Bourquin L. D., Bauer L. L., Reinhart G. A. Dietary fiber for dogs: IV. In vitro fermentation of selected fiber sources by dog fecal inoculum and in vivo digestion and metabolism of fiber-supplemented diets. J. Anim. Sci. 1995b; 73:1099-1109[Abstract]
• Sunvold G. D., Hussein H. S., Fahey G. C. Jr., Merchen N. R., Reinhart G. A. In vitro fermentation of cellulose, beet pulp, citrus pulp, and citrus pectin using fecal inoculum from cats, dogs, horses, humans, and pigs and ruminal fluid from cattle. J. Anim. Sci. 1995c; 73:3639-3648[Abstract]
• Suzuki S., Kawai K., Ohashi S., Mukai H., Yamashita K. Comparison of the effects of various C-terminal and N-terminal fragment peptides of glucagon-like peptide-1 on insulin and glucagon release from the isolated perfused rat pancreas. Endocrinology 1989; 125:3109-3114[Abstract]
• Tappenden, K. A. (1997) Short-Chain Fatty Acids Enhance Intestinal Adaptation in Rats Receiving Total Parenteral Nutrition: A Multiorgan Analysis. Ph.D. Thesis. University of Alberta, Edmonton, Canada.
• Tappenden K. A., McBurney M. I. Systemic short-chain fatty acids rapidly alter gastrointestinal structure, function and the expression of early response genes. Dig. Dis. Sci. 1998; 43:1526-1536[Medline]
• Tappenden K. A., Thomson A.B.R., Wild G. E., McBurney M. I. Short-chain fatty acid-supplemented total parenteral nutrition enhances functional adaptation to intestinal resection in rats. Gastroenterology 1997; 112:792-802[Medline]
• Tappenden K. A., Thomson A.B.R., Wild G. E., McBurney M. I. Short-chain fatty acids increase proglucagon and ornithine decarboxylase messenger RNAs after intestinal resection in rats. J. Parenter. Enteral. Nutr. 1996; 20:357-362[Abstract]
• Thomson A.B.R., Dietschy J. M. Intestinal kinetic parameters: effects of unstirred layers and transport preparation. Am. J. Physiol. 1980; 239:G372-377[Abstract/Free Full Text]
• Thomson A.B.R., Rajotte R. Effect of dietary modification on the uptake of glucose, fatty acids, and alcohols in diabetic rats. Am. J. Clin. Nutr. 1983; 38:394-403[Abstract/Free Full Text]
• Vaag A. A., Holst J. J., Volund A., Beck-Nielsen H. B. Gut incretin hormones in identical twins discordant for non-insulin-dependent diabetes mellitus (NIDDM)evidence for decreased glucagon-like peptide 1 secretion during oral glucose ingestion in NIDDM twins. Eur. J. Endocrinol. 1996; 135:425-432[Abstract]
• Wen J., Phillips S. F., Sarr M. G., Kost L. J., Holst J. J. PYY and GLP-1 contribute to feedback inhibition from the canine ileum and colon. Am. J. Physiol. 1995; 269:G945-952[Abstract/Free Full Text]
• Willms B., Werner J., Holst J. J., Orskov C., Creutzfeldt W., Nauck M. A. Gastric emptying, glucose responses, and insulin secretion after a liquid test meal: effects of exogenous glucagon-like peptide-1 (GLP-1)-(7-36) amide in type 2 (noninsulin-dependent) diabetic patients. J. Clin. Endocrinol. Metab. 1996; 81:327-332[Abstract]
• Wood S., Scott D., Beechy R., Shirazi-Beechy S. Cloning and sequencing of the ovine intestinal Na+/glucose transporter (SGLT1). Biochem. Soc. Trans. 1994; 22:266S[Medline]
• Xiaoyan, J. (1996) Regulation of pancreatic and gastric endocrine secretion by GIP and GLP-1(7-36)amide. Ph.D. Thesis. University of British Columbia, Vancouver.
• Zhao F. Q., Glimm D. R., Kennelly J. J. Distribution of mammalian facilitative glucose transporter messenger RNA in bovine tissues. Int. J. Biochem. 1993; 25:1897-1903[Medline]

0022-3166/98 $3.00 ©1998 American Society for Nutritional Sciences