Isomaltodextrin, a highly branched α-glucan, increases rat colonic H₂ production as well as indigestible dextrin

Abstract

Colonic hydrogen (H₂) protects against inflammation-induced oxidative stress. We examined the effect of a new highly branched α-glucan, isomaltodextrin (IMD), on colonic H₂ production in rats. Rats were fed a 16.7% IMD, 8.8% indigestible dextrin (ID), or 10.4% high amylose cornstarch diet (Expt. 1), were fed diets containing 3.3–16.7% IMD (Expt. 2), or were fed diets containing 16.7% IMD or 5.2% fructooligosaccharide (FOS) (Expt. 3), for 14 days. Compared with the control group, feeding IMD or other α-glucans dose dependently and significantly increased H₂ excretion and portal H₂ concentration. The ability of IMD to increase H₂ production was not inferior to that of FOS. The cecal Firmicutes/Bacteroidetes ratio in the IMD group was 5–14% of that in the control group. The cecal abundance of bifidobacteria was significantly greater in the IMD group than in the control group. Taken together, IMD, as well as other α-glucans, significantly increased colonic H₂ production in a dose-dependent manner.

Graphical abstract

Isomaltodextrin(IMD), a highly branched α-glucan, dose dependently increases colonic H₂ generation and decrease cecal Firmicutes(F)/Bacteroidetes(B) ratio.

Key words:

• isomaltodextrin
• colonic fermentation
• hydrogen gas
• large intestine
• rats

Nondigestible saccharides (NDSs), such as oligosaccharides, dietary fiber, and resistant starch promote colonic fermentation, which then generates large amounts of gases and short-chain fatty acids (SCFA).^1,2) Many investigators have studied various physiological effects of fermentation products. Colonic fermentation with low gas production has been considered to be useful because excessive gas production in the large intestine by administration of NDSs, such as oligosaccharides, dietary fiber, and resistant starch, causes flatulence and bloating.^3–5) However, we previously found that molecular hydrogen (H₂) generated by colonic fermentation of high amylose cornstarch (HAS) and pectin increased portal H₂ concentration and then reduced hepatic oxidative stress.⁶⁾ Also, we demonstrated the localization of colonic H₂ to the adipose tissue due to diffusion into the abdominal cavity and suggested the reduced mRNA abundance of adipocytokine due to colonic H₂.⁷⁾ Therefore, NDSs that can enhance colonic H₂ generation would be of significant benefit because excessive oxidative stress can lead to the onset and progression of various metabolic syndromes.

The physiological effect of only a few α-glucans, such as indigestible dextrin (ID) and resistant starch through colonic fermentation, has been examined because most α-glucans are generally almost completely digested by α-amylase in saliva and pancreatic juice and digestive enzymes in brush border membranes such as maltase, isomaltase, trehalase, glucoamylase, and limit dextrinase. The digestibility of saccharides is influenced by its structure including the degree of branch linkage and crystallinity. Therefore, α-glucans that have a unique structure could be partially nondigestible. Also, the utilization of NDSs in the large intestine is affected by bacterial enzyme activity and composition of microbiota as well as solubility of NDSs.

Isomaltodextrin (IMD) is a novel, highly branched, soluble α-glucan with a relatively low degree of polymerization (DP); it is enzymatically produced from maltodextrin and is resistant to amylase (Supplemental Fig. 1).⁸⁾ This glucan partially escapes digestion in the small intestine and then could undergo fermentation by bacteria in the large intestine, and therefore, it may play a role as a prebiotic. However, no study has demonstrated its ability as a fermentation substrate in the large intestine. ID, which is chemically made from starch and is also a soluble glucan with low DP, promotes fermentation in the large intestine. However, strictly speaking, ID is not a pure α-glucan because ID has a small proportion of β-glycosidic linkage.⁹⁾ It has been reported that breath H₂ excretion increased upon oral ingestion of ID in healthy subjects, although breath H₂ excretion increased by a greater extent upon ingestion of fructooligosaccharide (FOS), which is one of the fructans.¹⁰⁾ Unlike IMD and ID, HAS is an insoluble α-glucan with a high DP.

The fermentation rate of NDSs in the large intestine is likely influenced by the monosaccharide composition, glycosidic linkages between the monosaccharide residues, and the DP of the particular NDS.^11,12) Most bacteria preferentially utilize glucose as a source of carbohydrates; therefore, we assumed that a large supply of α-glucan to the large intestine could induce high production of H₂ in the large intestine. In the present study, we examined whether the pure, soluble, branched α-glucan, IMD, enhances colonic H₂ production and whether IMD feeding dose dependently increases colonic H₂ production.

Materials and methods

Samples

Isomaltodextrin, indigestible dextrin (Fibersol-2), HAS (Hi-maize 1043), and FOS (Meioligo-P) were kindly supplied by Hayashibara Co., Ltd. (Okayama, Japan), Matsutani Chemical Industry Co., Ltd. (Hyogo, Japan), Nippon NSC Ltd. (Tokyo, Japan) and Meiji Food Materia Co., Ltd. (Tokyo, Japan), respectively.

Measurement of properties of isomaltodextrin and indigestible dextrin

The number average molecular weight (Mn), weight average molecular weight (Mw), molecular weight distribution (Mw/Mn) and average DP in IMD and ID were determined using an HPLC apparatus equipped with a TSKgel α-M (13 μm, 7.8 × 300 mm, Tosoh, Tokyo, Japan) column with a refractive index detector.

Animals and diets

The study was approved by Nayoro City University Animal Care and Use Committee, and the animals were maintained in accordance with guidelines for the care and use of laboratory animals, Nayoro City University (approval numbers: 13–02 and 14–02). Male Sprague-Dawley rats weighing 260–280 g (8-week old) were obtained from Japan SLC (Haruno colony; Shizuoka, Japan). They were housed in individual cages made of screen bottoms of stainless steel in a room maintained at 23 ± 1 °C with humidity ranging from 50 to 70%, under lighting conditions of 12-h light/12-h darkness (07.00–19.00 h) daily. The rats were acclimatized by feeding a 25% casein control (C) diet for 3 days. The C diet was composed of 250 g/kg casein, 482.5 g/kg cornstarch, 100 g/kg sucrose, 70 g/kg soybean oil, 35 g/kg AIN-93G mineral mix,¹³⁾ 10 g/kg AIN-93 vitamin mix,¹³⁾ 2.5 g/kg choline bitartrate, and 50 g/kg cellulose.

Experiment 1

To determine the amount of hydrogen generated from various glucans in the large intestine, we examined the changes in H₂ production in rats fed a diet containing glucans. After the acclimation period, (breath + flatus) H₂ excretion over a 5-min period was measured by placing the rat inside a sealed polypropylene chamber for 5 min. After 5 min, an appropriate volume sample of the gaseous phase was withdrawn using a gas-tight syringe, and gas chromatography (GC) (Biogas analyzer BAS-1000; Mitleben, Osaka, Japan) was used to determine the amount of (breath + flatus) H₂ excretion. The H₂ excretion per 5 min was calculated by using the standard curve produced from peak height of standard H₂ gas. These rats were further assigned into four groups (n = 8 per group) based on (breath + flatus) H₂ excretion and body weight. Each group was administered the C diet supplemented with or without NDS for 10 d. Considering the digestibility of NDSs in vivo, which is 70%, 43%, and 52% for IMD, ID, and HAS, respectively (unpublished data), NDS was added to the C diet in such a way as to be able to deliver equal amounts to the large intestine; 16.7% IMD, 8.8% ID, or 10.4% HAS was supplemented to the C diet. It was not confirmed that equal amounts of NDSs were delivered to the large intestine in this study because rats with the large intestine were used. The supplementation of NDS was achieved by replacement of an equal weight of cornstarch in the C diet. The determination of the H₂ excretion was carried out on day 3, 7, 10 and 14.

Experiment 2

To determine the dose response of IMD on H₂ production in the large intestine, we examined the changes in (breath + flatus) H₂ excretion and portal H₂ concentration in rats fed a diet containing different amounts of IMD. After the acclimation period, (breath + flatus) H₂ excretion over a 5-min period was measured using the same method as in Experiment 1. The rats were assigned into five groups (n = 8 per group) based on (breath + flatus) H₂ excretion and body weight. Each group was administered the C diet supplemented with or without 3.3%, 6.7%, 10.0%, or 16.7% IMD, which is equal to 1%, 2%, 3%, or 5% as amounts of the nondigestible fraction to be supplied to the large intestine, respectively, for 14 days. IMD supplementation was achieved by the replacement of an equal weight of cornstarch in the C diet. The determination of the H₂ excretion was carried out on day 3, 7, 10, and 14.

Experiment 3

To compare H₂ generation in the large intestine by IMD feeding with that by FOS feeding, the latter of which leads to high generation of H₂, we examined the differences in (breath + flatus) H₂ excretion and portal H₂ concentration in rats administered IMD or FOS. After the acclimation period, (breath + flatus) H₂ excretion over a 5-min period was measured using the same method as earlier. The rats were assigned into three groups (n = 8 per group) based on (breath + flatus) H₂ excretion and body weight. Each group was administered the C diet supplemented with or without 16.7% IMD or 5.2% FOS for 10 days. As Experiment 1, IMD and FOS were added to the C diet in such a way as to be able to deliver equal amounts to the large intestine. IMD or FOS supplementation was achieved by the replacement of an equal weight of cornstarch in the C diet. The determination of the H₂ excretion was carried out on day 3, 7, 10, and 14.

Sampling

On the end of the experiment, rats were anesthetized with sodium penthobarbital (50 mg/kg body weight). After immediate laparotomy, 1 ml of blood was removed from the portal vein into sealed heparin vials for H₂ analysis. Also, ca. 1 g of perirenal adipose tissue was immediately removed into sealed vials for H₂ analysis. An appropriate volume sample of the gaseous phase was withdrawn using a gas-tight syringe, and H₂ concentration was determined using GC (Biogas analyzer BAS-1000; Mitleben). The cecum was removed and weighed, and the cecal content was collected and stored at −80 °C in airtight tubes until SCFA, bacterial DNA and bacterial enzyme analysis. The cecal wall was blotted and weighed after rinsing in an ice-cold saline solution.

Quantitative PCR of microbial cecal content

Bacterial DNA was isolated from frozen cecal contents using a DNA extraction kit (ISOFECAL for Beads Beating kit, Nippon Gene Co., Ltd., Tokyo, Japan) according to the manufacturer’s protocol, except for sample disruption. Frozen cecal contents (200 mg) were disrupted at 2700 rpm for 90 s using a Multi-Beads shocker (MB601U, Yasui Kikai Co, Osaka, Japan). Extracted DNA samples were dissolved in DNase-free, TE buffer (pH 8.0; 10 mM Tris-HCl, 1 mM EDTA) and stored at −30 °C until quantitative PCR analyses. The PCR was performed in a total volume of 20 μl. All bacteria were detected using the SYBR Premix Ex Taq II (Takara), with 200 nmol/l each of respective primers and <1 ng DNA for each reaction. The primers used to detect the bacteria were based on 16S rRNA gene sequences.^7,14) The PCR amplification reactions were carried out as follows: 2 min at 50 °C, 3 min at 95 °C, followed by 40 cycles of 15 s at 95 °C and 1 min at 60 °C. Melting curve analysis was performed after amplification. Detection was carried out on an Agilent Mx3000P (Agilent Technologies). Tenfold serial dilutions of known quantities of TA cloning vector (pCR2.1-TOPO, Life Technologies) with inserted 16S rDNA gene sequences from Lactobacillus acidophilus (JCM1132) and Bacteroides vulgatus (JCM5826) were used to generate standard curves in order to count copies of total bacteria and Firmicutes, and Bacteroidetes, respectively. Absolute rDNA abundance was calculated based on the standard curves. Each assay was performed in duplicate in the same analysis.

SCFA analysis

After homogenization of cecal contents, cecal organic acids (acetate, propionate, and n-butyrate) were measured using an HPLC system (LC-10A, Shimadzu, Kyoto, Japan) equipped with a Shim-pack SCR-102H column (8 mm i.d. × 30 cm long, Shimadzu) and an electroconductivity detector (CDD-6A, Shimadzu).¹⁵⁾ Briefly, ~300 mg cecal contents was homogenized in 2 ml of 10 mmol/L NaOH and then centrifuged at 10,000 × g for 15 min. The supernatant obtained was applied to HPLC analysis.

Bacterial enzyme activity

Bacterial enzyme activity in the cecal content was measured by the release rate of 4-nitrophenol from nitrophenylglucosides according to the method of Komiyama et al.¹⁶⁾ The following substrates were used: for α-glucosidase, 4-nitrophenyl-α-D-glucopyranoside; for β-glucosidase, p-nitrophenyl-β-D-glucopyranoside; for α-galactosidase, p-nitrophenyl-α-D-galactopyranoside; and for β-galactosidase, o-nitrophenyl-β-D-galactopyranoside. The cecal content (0.15 g) was diluted with 0.5 mL of ice-chilled PBS (pH 7.4) and then was homogenized at 2700 rpm at 4 °C for 90 s using a Multi-Beads shocker (MB601U, Yasui Kikai Co, Osaka, Japan). The homogenate was centrifuged (21,880 × g, 10 min, 4 °C), and the supernatant was immediately filtered using an Ultra free-MC (Merck Millipore, Tokyo, Japan). The reaction mixture contained 0.05 mL of substrate solution (5 mM) and 0.1 mL of a 1:4 (v/v) dilution of the above filtrate in 0.1 mol/L phosphate buffer (pH 6.4). Incubation was carried out at 37 °C for 30 min, and after the addition of 0.8 mL of 0.25 mol/L cold sodium carbonate, 4-nitrophenol was quantified at 400 nm. Enzymatic activity was expressed as micromoles of product formed per minute (unit) per gram of cecal content.

Statistical analysis

Values obtained from the experiments were expressed as mean ± SEM. Data were subjected to Bartlett’s test for the homogeneity of variances, and unequal variances were stabilized by log transformation. For samples with equal variances, one-way analysis of variance (ANOVA) was used, followed by Tukey–Kramer post hoc test for multiple comparisons between individual group means. If sample variances were still unequal after log transformation, we used the Steel–Dwass test. Repeated-measures two-way ANOVA was used to analyze the change in H₂ excretion among the NDS groups across time. H₂ excretion within an individual group was compared with repeated-measures one-way ANOVA. The Tukey–Kramer test and Steel–Dwass test were performed using SAS JMP software (version 9.0.2; Tokyo, Japan). Significance was defined as a p < 0.05.

Results

The molecular weight and DP of IMD were higher than those of ID. That is, the Mn, Mw, Mw/Mn and DP of IMD were 2330, 4710 g/mol, 2.0 and 29, respectively, while those of ID were 1420, 2900 g/mol, 2.0 and 18, respectively.

Experiment 1

Food intake was significantly lower in the IMD and ID groups than in the C group, but a significant difference in body weight gain was not observed (ANOVA p = 0.1091; Table 1). The AUC for 0–14 days in (breath + flatus) H₂ excretion (AUC_0–14d) and portal and perirenal H₂ concentrations were significantly greater in all groups administered glucan, especially the ID group, than in the C group (Fig. 1, Table 1). However, (breath + flatus) H₂ excretion in the IMD and ID groups on d 14 was 11% and 9% of that on the day that it peaked in the respective group (Fig. 1). The weights of cecal contents and tissue were significantly higher in the three glucan groups than in the C group, and the pH of cecal contents was significantly lower in the three glucan groups (Table 2). Only the cecal succinate concentration, but not the cecal concentration of short-chain fatty acids, was significantly higher in the IMD group than in the C group, while n-butyrate and succinate concentrations were significantly higher in the ID and HAS groups than in the C group. Compared with the C group, cecal α-D-glucosidase activity was significantly higher in the ID group and was significantly lower in the HAS group, and β-D-glucosidase activity was significantly weaker in the IMD and HAS groups and was stronger in the ID group. The total number of bacteria was significantly higher in the three glucan groups compared with the C group. The abundance of Firmicutes was significantly greater in the ID and HAS groups than in the C group. In addition, the abundance of Bacteroidetes was significantly higher in the IMD group compared with the C group. Compared with the C group, the abundance of Bacteroides–Prevotella was significantly greater in the three glucan groups, while that of Clostridium cluster XIVa was higher in the glucan groups except for the IMD group. The numbers of Bacteroides–Prevotella and Bifidobacterium were higher in the IMD and HAS groups, but not the ID group, compared with those in the C group. No significant difference in the abundance of Lactobacillus was observed.

Table 1. Food intake, bodyweight gain, breath and flatus H₂ excretion, portal and abdominal H₂ concentrations in rats fed nondigestible α-glucan for 14 days (Expt. 1).

	C	IMD	ID	HAS
Food intake, g	309 ± 10^a	278 ± 8^bc	273 ± 9^c	307 ± 4^ab
Body weight gain, g	64 ± 5	57 ± 4	57 ± 4	68 ± 3
Breath and flatus H2 excretion AUCd0–14, mmol	0.413 (0.202–1.50)^b	18.1 (7.17–54.2)^a	36.8 (18.6–75.7)^a	17.5 (7.03–35.5)^a
Portal H2, μmol/L	0.50 (0.009–2.79)^b	3.11 (1.69–14.1)^a	3.71 (1.69–8.6)^a	9.10 (3.53–18.9)^a
Perinetal fat
Weight, g	6.12 ± 0.53	5.03 ± 0.52	4.67 ± 0.31	6.39 ± 0.47
Weight, g/100 g body weight	1.55 ± 0.12	1.30 ± 0.12	1.22 ± 0.08	1.61 ± 0.12
H2, μmol/kg	0.109 (0.082–0.416)^c	0.478 (0.288–2.84)^b	2.39 (0.189–7.60)^ab	10.8 (3.03–19.1)^a

Notes: Values are expressed as means ± SEM or medians (minimum value-maximum value), n = 8. Values in a row with different superscript letters indicate a statistically significant difference, p < 0.05.

AUC_d0–14, area under the curve for 0–14 days; C, group fed the control diet; HAS, group fed the control diet supplemented with 10.4% high amylose cornstarch; ID, group fed the control diet supplemented with 8.8% indigestible dextrin; IMD, group fed the control diet supplemented with 16.7% isomaltodextrin.

Fig. 1. Change in (breath + flatus) H₂ excretion in rats fed the (C), IMD, ID and HAS diets for 14 days (Expt. 1).

Notes: Values are means, with their standard errors represented by vertical bars, n = 8. Labeled means at a specific time point without a common letter differ, p < 0.05. Repeated-measures 2-way ANOVA was used to analyze H₂ excretion among the three high-fat groups across time (NDS, p < 0.0001; time, p < 0.0001; interaction, p = 0.0052). (C), th group fed the control diet; HAS, group fed the high amylose cornstarch diet (104 g/kg); H₂, hydrogen; ID, group fed the 8.8% indigestible dextrin diet; IMD, groups fed the 16.7% isomaltodextrin diet.

Table 2. Weight of cecal tissue and content, cecal organic acid concentration, and bacterial enzyme activity, counts and ratio in rats fed nondigestible α-glucan for 14 days (Expt. 1).

	C	IMD	ID	HAS
Tissue, g	0.749 (0.641–0.935)^c	1.25 (1.15–1.29)^b	1.52 (1.28–1.63)^a	1.22 (0.959–1.51)^ab
Contents
Weight, g	3.21 (2.50–3.62)^c	7.22 (6.11–8.99)^b	12.0 (8.23–14.7)^a	6.47 (5.55–10.7)^b
pH	6.47 ± 0.05^b	5.77 ± 0.06^a	5.66 ± 0.03^a	5.70 ± 0.05^a
Organic acid, μmol/g
Acetate	28.8 (24.0–41.2)^a	22.3 (14.9–32.3)^ab	23.4 (19.8–25.5)^b	34.6 (25.9–55.4)^a
Propionate	10.6 (9.18–13.2)^a	11.3 (8.76–13.0)^a	8.48 (3.97–17.4)^ab	7.24 (5.24–9.56)^b
n-Butyrate	2.32 (0.728–3.48)^b	2.47 (2.08–3.41)^b	5.07 (2.35–8.81)^a	7.44 (4.20–9.78)^a
Succinate	21.0 (13.0–29.0)^b	54.1 (49.3–62.7)^a	43.4 (38.1–55.8)^a	51.9 (21.4–65.7)^a
Bacterial enzyme, units/g cecal content
α-D-glucosidase	14.8 (11.5–21.2)^b	27.7 (9.86–50.0)^ab	37.5 (18.5–98.4)^a	8.73 (3.44–31.2)^b
β-D-glucosidase	1.42 (1.21–1.49)^b	0.707 (0.568–1.33)^c	3.53 (1.87–8.29)^a	0.789 (0.668–1.28)^c
α-D-galactosidase	4.98 (4.42–5.45)^a	1.29 (0.977–1.67)^c	3.06 (1.64–4.66)^b	1.10 (0.876–1.50)^c
β-D-galactosidase	8.12 ± 0.55^a	2.99 ± 0.63^c	5.30 ± 0.66^b	2.51 ± 0.42^c
Bacteria counts, log10copies/g
Total bacteria	12.2 ± 0.1^b	12.6 ± 0.1^a	12.7 ± 0.0^a	12.8 ± 0.0^a
Phylum
Firmicutes (F)	12.0 ± 0.1^b	12.0 ± 0.0^b	12.4 ± 0.0^a	12.5 ± 0.1^a
Bacteroidetes (B)	11.5 ± 0.1^b	12.3 ± 0.1^a	12.2 ± 0.0^b	12.2 ± 0.1^b
F:totalbacteria	0.716 ± 0.027^a	0.229 ± 0.032^c	0.522 ± 0.027^b	0.571 ±0.05^b
B:total bacteria	0.213 ± 0.016^b	0.508 ± 0.030^a	0.315 ± 0.015^b	0.279 ± 0.042^b
F/B	3.54 ± 0.34^a	0.50 ± 0.11^c	1.71 ± 0.17^b	2.57 ± 0.53^ab
Genus
Bacteroides–Prevotella	10.5 ± 0.2^c	11.8 ± 0.1^ab	11.6 ± 0.1^b	12.1 ± 0.1^a
Clostridium cluster XIVa	11.3 ± 0.1^b	11.5 ± 0.0^b	11.9 ± 0.0^a	11.8 ± 0.1^a
Lactobacillus	10.5 ± 0.1	10.8 ± 0.2	11.1 ± 0.2	10.9 ± 0.1
Bifidobacterium	8.23 ± 0.11^b	10.3 ± 0.1^a	8.48 ± 0.38^b	10.5 ± 0.3^a

Notes: Values are expressed as means ± SEM or medians (minimum value–maximum value), n = 8. Values in a row with different superscript letters indicate a statistically significant difference, p < 0.05.

Enzymatic activity was expressed as micromoles of product formed per minute (unit) per gram of cecal content.

C, group fed the control diet; HAS, group fed the control diet supplemented with 10.4% high amylose cornstarch; ID, group fed the control diet supplemented with 8.8% indigestible dextrin; IMD, group fed the control diet supplemented with 16.7% isomaltodextrin.

Experiment 2

Food intake and body weight gain were dose dependently lower in rats receiving dietary IMD supplementation than in the C group (Table 3). The AUC_0–14d and portal and perirenal H₂ concentrations were dose dependently increased by IMD feeding compared with the C group and those in the 16.7% IMD group were 66-, 26-, and 28-fold of the respective value in the C group (Fig. 2 and Table 3). IMD feeding dose dependently reduced the weight of perirenal adipose tissue. IMD feeding dose dependently increased the weights of cecal contents and tissue (Table 4). The pH of cecal contents was dose dependently reduced by IMD feeding. As in Experiment 1, the cecal acetate concentration did not differ among the five groups. The succinate concentration dose-dependently increased by IMD feeding and that in the 16.7% IMD group was 14-fold of the value in the C group. Cecal α-D-glucosidase activity in all IMD groups was about 3-fold of the value in the C group, while IMD feeding dose dependently decreased β-D-glucosidase activity. IMD feeding dose dependently increased the total number of bacteria. The abundance of Bacteroidetes was dose dependently increased by IMD feeding compared with that in the C group. In addition, the Firmicutes/Bacteroidetes ratio in the 3.3%, 6.7%, and 16.7% IMD groups was 48%, 12%, and 5% of the value in the C group, respectively. The abundance of Bacteroides–Prevotella and Bifidobacterium increased and that of Clostridium cluster XIVa decreased in a dose-dependent manner.

Table 3. Dose response of isomaltodextrin on food intake, bodyweight gain, breath and flatus H₂ excretion, portal and abdominal H₂ concentrations in rats.

	C	3.3IMD	6.7IMD	10IMD	16.7IMD
Food intake, g	323 ± 14	324 ± 10	311 ± 14	300 ± 5	278 ± 12
Body weight gain, g	77 ± 5^a	82 ± 4^a	78 ± 4^a	74 ± 3^ab	59 ± 5^b
Breath and flatus H2 excretion AUCd0–14, mmol	0.319 (0.183–0.944)^d	1.77 (0.506–2.55)^c	7.51 (2.99–22.9)^ab	9.59 (4.13–17.4)^b	20.9 (11.5–27.6)^a
Portal H2, μmol/L	0.230 (0.120–2.79)^b	1.11 (0.39–5.98)^ab	4.93 (1.76–8.66)^a	7.70 (2.01–9.53)^a	5.99 (3.75–13.5)^a
Perirenal fat
Weight, g	7.37 ± 0.58^a	7.25 ± 0.44^a	6.62 ± 0.64^ab	6.38 ± 0.38^ab	5.08 ± 0.38^b
Weight, g/100 g body weight	1.90 ± 0.12^a	1.85 ± 0.10^a	1.68 ± 0.14^ab	1.65 ± 0.10^a	1.38 ± 0.0^b
H2, μmol/kg	0.105 (0.011–0.219)^c	0.104 (0.053–1.31)^bc	0.880 (0.193–2.21)^ab	1.19 (0.250–10.8)^ab	2.92 (0.474–15.2)^a

Notes: Values are expressed as means ± SEM or medians (minimum value-maximum value), n = 8. Values in a row with different superscript letters indicate a statistically significant difference, p < 0.05.

AUC_d0–14, area under the curve for 0–14 days; C, group fed the control diet; 3.3IMD-16.7IMD, group fed the control diet supplemented with 3.3–16.7% isomaltodextrin.

Fig. 2. Dose response of IMD on (breath + flatus) H₂ excretion in rats (Expt. 2).

Notes: Values are means, with their standard errors represented by vertical bars, n = 8. Labeled means at a specific time point without a common letter differ, p < 0.05. Repeated-measures 2-way ANOVA was used to analyze H₂ excretion among the three high-fat groups across time (dose, p < 0.0001; time, p < 0.0001; interaction, p < 0.0001). C, th group fed the control diet; H₂, hydrogen; 3.3IMD-16.7IMD, groups fed the 3.3–16.7% isomaltodextrin diet.

Table 4. Dose response of isomaltodextrin on weight of cecal tissue and content, cecal organic acid concentration, and bacterial enzyme activity, counts and ratio in rats.

	C	3.3IMD	6.7IMD	10IMD	16.7IMD
Tissue, g	0.711 (0.609–0.774)^d	0.853 (0.777–0.934)^c	0.911 (0.830–1.04)^bc	1.02 (0.970–1.23)^b	1.25 (1.09–1.66)^a
Contents
Weight, g	2.95 (2.58–3.33)^d	3.71 (3.22–4.61)^c	4.77 (2.67–6.18)^bcd	7.01 (5.56–8.10)^b	8.89 (6.96–10.9)^a
pH	6.72 ± 0.09^a	6.46 ± 0.05^a	5.82 ± 0.07^ab	5.51 ± 0.06^ab	5.45 ± 0.10^b
Organic acid, μmol/g
Acetate	43.9 (36.4–48.3)	44.1 (35.2–53.4)	41.6 (26.5–55.2)	38.0 (30.8–52.8)	35.3 (28.3–52.7)
Propionate	13.8 (10.3–15.3)^b	16.2 (12.9–18.6)^ab	14.1 (9.94–16.3)^ab	15.6 (14.8–17.5)^a	19.5 (12.5–21.3)^ab
n-Butyrate	4.56 (3.71–5.25)^b	4.58 (0.00–5.95)^ab	5.28 (4.68–8.72)^a	5.54 (4.02–8.25)^ab	6.06 (5.33–16.2)^a
Succinate	3.90 (1.35–15.4)^b	15.2 (6.83–19.8)^b	43.6 (35.7–54.2)^a	49.5 (33.0–63.4)^a	53.4 (41.4–65.5)^a
Bacterial enzyme, units/g cecal content
α-D-glucosidase	18.7 (16.5–26.3)^b	57.8 (38.9–78.9)^a	63.5 (36.5–73.1)^a	55.0 (31.2–92.9)^a	54.2 (26.8–77.1)^a
β-D-glucosidase	1.35 (1.06–1.50)^ab	1.87 (1.09–2.74)^a	1.61 (1.15–2.09)^ab	1.42 (1.32–1.64)^ab	0.99 (0.68–1.83)^b
α-D-galactosidase	5.14 (3.79–7.66)^a	4.85 (3.04–7.15)^ab	3.63 (2.74–5.35)^abc	3.27 (2.48–3.72)^bc	2.44 (1.45–3.74)^c
β-D-galactosidase	5.86 (5.04–8.18)^a	7.22 (4.19–9.79)^a	4.68 (3.64–7.21)^ab	4.25 (3.58–4.79)^b	3.70 (2.93–5.59)^b
Bacteria counts, log10copies/g
Total bacteria	12.0 ± 0.1^bc	11.7 ± 0.1^c	12.0 ± 0.1^bc	12.4 ± 0.1^ab	12.4 ± 0.1^a
Pylum
Firmicutes (F)	11.8 ± 0.1^a	11.5 ± 0.1^ab	11.4 ± 0.0^b	11.6 ± 0.1^ab	11.4 ± 0.2^b
Bacteroidetes (B)	11.3 ± 0.1^c	11.3 ± 0.1^c	11.8 ± 0.1^b	12.1 ± 0.1^ab	12.2 ± 0.1^a
F:total bacteria	0.640 ± 0.043^a	0.603 ± 0.053^a	0.251 ± 0.047^b	0.202 ± 0.044^b	0.123 ± 0.034^b
B:total bacteria	0.201 ± 0.028^b	0.357 ± 0.027^b	0.606 ± 0.041^a	0.600 ± 0.073^a	0.726 ± 0.043^a
F/B	3.82 ± 0.72^a	1.82 ± 0.27^b	0.47 ± 0.12^bc	0.65 ± 0.39^bc	0.19 ± 0.07^c
Genus
Bacteroides–Prevotella	11.1 ± 0.1^c	11.1 ± 0.1^c	11.7 ± 0.1^b	12.0 ± 0.1^ab	12.2 ± 0.1^a
Clostridium cluster XIVa	11.5 ± 0.1^a	11.1 ± 0.1^ab	11.1 ± 0.0^ab	11.3 ± 0.1^ab	11.1 ± 0.1^b
Lactobacillus	10.7 ± 0.1	9.89 ± 0.1	10.1 ± 0.1	10.5±0.3	10.3 ± 0.3
Bifidobacterium	7.83 ± 0.11^b	7.48 ± 0.22^b	9.71 ± 0.29^a	10.2 ± 0.3^a	10.0 ± 0.3^a

Notes: Values are expressed as means ± SEM or medians (minimum value–maximum value), n = 8. Values in a row with different superscript letters indicate a statistically significant difference, p < 0.05.

Enzymatic activity was expressed as micromoles of product formed per minute (unit) per gram of cecal content.

C, group fed the control diet; 3.3IMD-16.7IMD, group fed the control diet supplemented with 3.3–16.7% isomaltodextrin.

Experiment 3

Food intake and body weight gain were significantly lower in the 16.7% IMD group and tended to be lower in the 5.2% FOS group than in the C group (Table 5). The AUC_0–14d was significantly greater to the same degree in the 16.7% IMD and 5.2% FOS groups than in the C group (Fig. 3 and Table 5). Portal and perirenal H₂ concentrations were also significantly greater in both groups than in the C group, and the portal H₂ concentration in the 5.2% FOS group was 2.7-fold of that in the 16.7% IMD group. Feeding IMD or FOS significantly reduced the weight of the perirenal adipose tissue compared with that in the C group. The weights of cecal contents and tissue were significantly higher in the 16.7% IMD and 5.2% FOS groups than in the C group, and the pH of cecal contents was significantly lower in both groups (Table 6). Cecal acetate concentration was significantly higher in the 5.2% FOS group, but not in the 16.7% IMD, compared with that in the C group. The propionate concentration was significantly higher or lower in the 16.7% IMD and 5.2% FOS groups, respectively, compared with that in the C group. The butyrate concentration was significantly higher in both groups, and the butyrate concentration in the 5.2% FOS group was 3.9-fold of that in the 16.7% IMD group. Cecal α-D-glucosidase activity was significantly higher in the 16.7% IMD group, but was significantly lower in the 5.2% FOS group compared with that in the C group. Compared with the C group, the total number of bacteria was significantly higher in both groups. The abundance of Firmicutes was significantly greater in the 5.2% FOS group and was significantly lower in the 16.7% IMD group than in the C group. In addition, the abundance of Bacteroidetes was significantly higher in the IMD group only. The Firmicutes/Bacteroidetes ratio in the 16.7% IMD and 5.2% FOS groups was 5% and 424% of the value in the C group, respectively. Compared with the C group, the abundance of Bacteroides–Prevotella was significantly greater in both groups, and that in the 16.7% IMD and 5.2% FOS groups was 1400 and 230% of the value in the C group, respectively. The abundance of Bifidobacterium in the 16.7% IMD and 5.2% FOS groups was 50,200 and 81,2000% of the value in the C group. The abundance of Clostridium cluster XIVa was significantly lower or higher in the 16.7% IMD and 5.2% FOS groups, respectively, compared with that in the C group.

Table 5. Comparison between isomaltodextrin and fructooligosaccharide on food intake, bodyweight gain, breath and flatus H₂ excretion, portal and abdominal H₂ concentrations in rats.

	C	16.7IMD	5.2FOS
Body weight gain, g	77 ± 5^a	59 ± 5^b	68 ± 4^ab
Food intake, g	323 ± 14^a	278 ± 12^b	292 ± 9^ab
Breath and flatus H2 excretion AUCd0–14, mmol	0.319 (0.183–0.944)^b	20.9 (11.5–27.6)^a	19.0 (14.6–46.5)^a
Portal H2, μmol/L	0.230 (0.120–4.71)^b	5.99 (3.75–13.5)^a	16.4 (5.43–22.4)^a
Perirenal fat
Weight, g	7.37 ± 0.58^a	5.08 ± 0.38^b	5.41 ± 0.41^b
Weight, g/100 g body weight	1.90 ± 0.12^a	1.38 ± 0.08^b	1.41 ± 0.10^b
H2, μmol/kg	0.105 (0.011–0.219)^b	2.92 (0.474–15.2)^a	7.65 (1.12–28.5)^a

Notes: Values are expressed as means ± SEM or medians (minimum value–maximum value), n = 8. Values in a row with different superscript letters indicate a statistically significant difference, p < 0.05.

AUC_d0–14, area under the curve for 0–14 days; C, group fed the control diet; 5.2FOS, group fed the control diet supplemented with 5.2% fructooligosaccharide; 16.7IMD, group fed the control diet supplemented with 16.7% isomaltodextrin.

Fig. 3. Comparison between IMD and FOS on (breath + flatus) H₂ excretion in rats (Expt. 3).

Notes: Values are means, with their standard errors represented by vertical bars, n = 8. Labeled means at a specific time point without a common letter differ, p < 0.05. Repeated-measures 2-way ANOVA was used to analyze H₂ excretion among the three high fat groups across time (NDF, p < 0.0001; time, p < 0.0001; interaction, p < 0.0046). C, th group fed the control diet; FOS, groups fed the 5.2% fructooligosaccharide diet; H₂, hydrogen; IMD, groups fed the 16.7% isomaltodextrin diet.

Table 6. Comparison between isomaltodextrin and fructooligosaccharide on weight of cecal tissue and content, cecal organic acid concentration, and bacterial enzyme activity, counts and ratio in rats.

	C	16.7IMD	5.2FOS
Tissue, g	0.711 (0.609–0.774)^b	1.25 (1.09–1.66)^a	1.26 (1.04–1.52)^a
Contents
Weight, g	2.95 (2.58–3.33)^c	8.89 (6.96–10.9)^a	5.98 (4.00–8.92)^b
pH	6.72 ± 0.09^a	5.45 ± 0.10^b	5.49 ± 0.12^b
Organic acid, μmol/g
Acetate	43.9 (36.4–48.3)^b	35.3 (28.3–52.7)^b	65.7 (47.0–91.8)^a
Propionate	13.8 (10.3–15.3)^b	19.5 (12.5–21.3)^a	5.92 (3.88–13.6)^c
n-Butyrate	4.56 (3.71–5.25)^c	6.06 (5.33–16.2)^b	23.6 (22.0–35.5)^a
Succinate	3.90 (1.35–15.4)^c	53.4 (41.4–65.5)^a	16.0 (8.17–25.1)^b
Bacterial enzyme, units/g cecal content
α-D-glucosidase	18.7 (16.5–26.3)^b	54.2 (26.8–77.1)^a	15.5 (12.5–18.9)^c
β-D-glucosidase	1.35 (1.06–1.50)^b	0.99 (0.68–1.83)^b	1.60 (1.34–2.98)^a
α-D-galactosidase	5.14 (3.79–7.66)^a	2.44 (1.45–3.74)^b	3.37 (2.47–6.26)^b
β-D-galactosidase	5.86 (5.04–8.18)^a	3.70 (2.93–5.59)^b	8.49 (4.16–12.8)^a
Bacteria counts, log10copies/g
Total bacteria	12.0 ± 0.1^c	12.4 ± 0.1^b	12.7 ± 0.1^a
Pylum
Firmicutes (F)	11.8 ± 0.1^b	11.4 ± 0.2^c	12.7 ± 0.1^a
Bacteroidetes (B)	11.3 ± 0.1^b	12.2 ± 0.1^a	11.5 ± 0.1^b
F:total bacteria	0.640 ± 0.043^b	0.123 ± 0.034^c	0.913 ± 0.027^a
B:total bacteria	0.201 ± 0.028^b	0.726 ± 0.043^a	0.069 ± 0.09^c
F/B	3.82 ± 0.72^b	0.19 ± 0.07^b	16.2 ± 3.2^a
Genus
Bacteroides–Prevotella	11.1 ± 0.1^c	12.2 ± 0.1^a	11.5 ± 0.1^b
Clostridium cluster XIVa	11.5 ± 0.1^b	11.1 ± 0.1^c	12.1 ± 0.1^a
Lactobacillus	10.7 ± 0.1	10.3 ± 0.3	10.8 ± 0.1
Bifidobacterium	7.83 ± 0.11^c	10.0 ± 0.3^b	11.7 ± 0.2^a

Notes: Values are expressed as means ± SEM or medians (minimum value-maximum value), n = 8. Values in a row with different superscript letters indicate a statistically significant difference, p < 0.05.

Enzymatic activity was expressed as micromoles of product formed per minute (unit) per gram of cecal content.

C, group fed the control diet; 5.2FOS, group fed the control diet supplemented with 5.2% fructooligosaccharide; 16.7IMD, group fed the control diet supplemented with 16.7% isomaltodextrin.

Discussion

The fermentation rate of NDSs which includes the gas production rate and SCFA production rate is likely influenced by the monosaccharide composition, the glycosidic linkages between the monosaccharide residues, and the DP of the particular NDS.^11,12) In the case of fructans, the lower the DP, the more rapid its fermentation rate, resulting in increased production of gases.¹⁷⁾ NDSs with low DP would be utilized easily by bacteria because solubility increases as the DP decreases. Breath H₂ excretion can be used as an indicator of colonic H₂ production.¹⁸⁾ In the present study, considering the AUC_0–14d, although the extent of colonic H₂ production in the ID group (DP, 18) was 200% of the value in the IMD group (DP, 29), this difference was not statistically significant. Also, increased H₂ excretion by HAS feeding was almost as much as that by IMD feeding, although HAS is an insoluble α-glucan with a high DP unlike IMD. Resistant starches, such as HAS, have been reported to be relatively rapidly fermented in the large intestine.^19,20) These findings could support our results. Therefore, HAS, as well as IMD and ID, is also suggested to increase colonic H₂ production. Unlike fructan, α-glucans that reach the large intestine would be easily utilized by colonic bacteria and the fermentation rate would be immune to the differences in DP and solubility. The present study demonstrated that IMD, a new indigestible α-glucan that is enzymatically made from maltodextrin, as well as ID and HAS, partially at least reached the large intestine and induced colonic H₂ production. However, further investigation is required to verify that this α-glucan could increase colonic H₂ production independent of its DP and solubility.

In the present study, high α-glucosidase activity in cecal contents was observed in rats fed NDSs that have α-glycoside linkage, except for HAS, and the increase in the activity reached a plateau by administration of the 3.3% IMD diet. Also, only ID which has a small proportion of β-glycoside linkage (<3% of all glycosidic linkage) as well as α-glycoside linkage, amplified β-glucosidase activity in cecal contents. These results are understandable because bacteria are exposed to a limited carbon source. However, the effect of ID on β-glucosidase activity should be compared with that of β-glucan by further study. The shift in glycosidic enzyme activity in cecal contents would be attributed to dynamic changes in the composition of colonic microbiota as well as altered gene expression of the enzyme in colonic bacteria.

NDSs potentially shift the microbiota composition in the large intestine because the preferred fermentation substrate differs depending on the genus of bacteria. Most (90%) of the phylotypes of gut bacteria are members of two phyla: the Bacteroidetes (Gram-negative) and the Firmicutes (Gram-positive).²¹⁾ The Firmicutes account for about 70% of the total bacterial population within the human colonic microbiota²¹⁾ and the population corresponded to the Firmicutes/Bacteroidetes (F/B) ratio of the C group (3.5) in the present study. The F/B ratio after IMD feeding decreased to 0.5 and Bacteroidetes were predominant, although the F/B ratio after ID and HAS feedings was 1.7 and 2.6, respectively (Table 2). Also, the effect of IMD was dose dependent. Unlike the bacterial phylum Firmicutes, Bacteroides spp., which belong to the bacterial phylum Bacteroidetes, can take soluble polysaccharides into the periplasmic space and degrade them.^22,23) Bacteria belonging to the genus Bacteroides may preferentially utilize soluble NDSs compared with other bacteria. Holscher et al.²⁴⁾ reported that polydextrose and soluble corn fiber decreased the F/B ratio in feces in humans. Therefore, these reports support the present results in that the abundance of Bacteroidetes increased and the F/B ratio decreased by administration of IMD and ID, which are soluble. A high F/B ratio has been reported to be related to increased adipose tissue and the development of obesity.²⁵⁾ Our results that adipose tissue decreased in the IMD group, which showed a low F/B ratio (0.19 vs. 3.82 in the C group), appear to be significant. However, the weight of adipose tissue decreased in the FOS group even though the F/B ratio (16.2) was higher than that in the C group. It would be simplistic to assume that the shift in the composition of colonic microbiota directly contributes to the change in adipose tissue weight.

Bifidobacteria are thought to confer health benefits to the host.²⁶⁾ The bifidogenic effect of FOS, inulin, galactooligosaccharide, and resistant starch has been well established in various rat and human studies.^27,28) Isomaltooligosaccharide was reported not to be bifidogenic in humans in a controlled human study.²⁸⁾ To the best of our knowledge, there have been few studies demonstrating the bifidogenic effect of α-glucan except resistant starch. In this study, IMD feeding increased the number of cecal bifidobacteria in a dose-dependent manner, suggesting that IMD is a bifidogenic α-glucan.

In our previous study, oxidative stress induced by hepatic ischemia-reperfusion was suppressed when H₂ excretion and portal H₂ concentration in rats fed NDSs were more than 5 μmol/5 min and in the range of 3–9 μmol/L, respectively.^6,29) FOS feeding generates considerable amounts of colonic H₂ in rats and humans compared with feeding of other NDSs.^7,10) Therefore, FOS would be a useful NDS that could alleviate oxidative stress and oxidative damage. In the present study, the H₂ excretion and portal H₂ concentration in rats fed ID or IMD on day 3–10 crossed these benchmarks. In the present study, IMD dose dependently increased colonic H₂ production and its ability to increase the H₂ production was not inferior to that of FOS. However, compared with rats fed FOS, adipose H₂ concentration in rats fed IMD was lower. In our previous study, high adipose H₂ concentration (9 μmol/kg) in dietary-induced obese rats fed FOS decreased adipose mRNA abundance of inflammatory cytokine. Therefore, unlike FOS, IMD may not increase adipose H₂ concentration even though (breath + flatus) H₂ excretion and portal H₂ concentration are high. H₂ production rate in the large intestine could change the movement of colonic H₂. Further study is required to determine why adipose H₂ concentration is different in rats with high colonic H₂ generation.

Taken together, IMD, which is a highly branched α-glucan and partially escapes digestion in the small intestine, significantly increased colonic H₂ production as well as ID and HAS. IMD feeding dose dependently increased colonic H₂ production. Furthermore, the ability of IMD to increase colonic H₂ production was almost equal to that of FOS. Therefore, IMD would be a new potential food source for colonic H₂ production. Moreover, IMD shifts the composition of colonic microbiota to a low F/B ratio and a high level of bifidobacteria.

Author contributions

NN designed the study; NN, HT, and TY conducted the study; NN analyzed the dataset; NN wrote the manuscript; and NN had primary responsibility for the final content. All authors were involved in designing the study, reviewing and interpreting the results, and drafting the manuscript. All authors read and approved the final manuscript.

Disclosure statement

The study was proposed, designed and all analysis performed solely by the authors.

Funding

This work was supported by Hayashibara Co., Ltd. (Okayama, Japan).

Supplemental material

The supplemental material for this paper is available at http://dx.doi.org/10.1080/09168451.2015.1104237.