1-Kestose supplementation mitigates the progressive deterioration of glucose metabolism in type 2 diabetes OLETF rats

The fructooligosaccharide 1-kestose cannot be hydrolyzed by gastrointestinal enzymes, and is instead fermented by the gut microbiota. Previous studies suggest that 1-kestose promotes increases in butyrate concentrations in vitro and in the ceca of rats. Low levels of butyrate-producing microbiota are frequently observed in the gut of patients and experimental animals with type 2 diabetes (T2D). However, little is known about the role of 1-kestose in increasing the butyrate-producing microbiota and improving the metabolic conditions in type 2 diabetic animals. Here, we demonstrate that supplementation with 1-kestose suppressed the development of diabetes in Otsuka Long-Evans Tokushima Fatty (OLETF) rats, possibly through improved glucose tolerance. We showed that the cecal contents of rats fed 1-kestose were high in butyrate and harbored a higher proportion of the butyrate-producing genus Anaerostipes compared to rats fed a control diet. These findings illustrate how 1-kestose modifications to the gut microbiota impact glucose metabolism of T2D, and provide a potential preventative strategy to control glucose metabolism associated with dysregulated insulin secretion.


Fasting plasma glucose and insulin concentrations.
To assess the effects of 1-kestose supplementation on glucose metabolism in the T2D model, we measured fasting plasma glucose and insulin concentrations over the course of the experimental period in OLETF rats. The fasting plasma glucose concentrations in LETO and OLETF rats were not different up to 18 weeks of age (Fig. 1A). At 22 weeks of age, the fasting plasma glucose concentration was elevated only in OLETF/CON, and that in OLETF/KES was comparable to that in LETO rats; thus, values were significantly different between OLETF/CON and OLETF/KES (Fig. 1A).
At 14 weeks of age, fasting plasma insulin concentrations tended to be higher in OLETF rats than that in LETO rats; however, only that in OLETF/CON was significantly higher than that in LETO rats (Fig. 1B). At 18 weeks of age, the insulin concentration in OLETF/CON was further elevated, but that in OLETF/KES remained almost constant; thus, the insulin concentration was significantly higher only in OLETF/CON than the other three groups (Fig. 1B). At 22 weeks of age, the insulin concentration in OLETF/CON declined and was not significantly different from that in LETO rats (Fig. 1B). On the other hand, the insulin concentration in OLETF/ KES at the same age was elevated and was significantly higher than that in LETO/KES (Fig. 1B).
C-peptide concentration. Since the decrease in insulin concentration in OLETF/CON at 22 weeks of age suggests a decline in insulin secretion from pancreatic β-cells 12 , we measured the C-peptide concentration at 18 and 22 weeks of age. The C-peptide concentrations of the four groups at both 18 and 22 weeks of age and the observed changes from 18 and 22 weeks of age showed almost the same pattern as for insulin concentrations (Fig. 2), suggesting that insulin secretion declined only in OLETF/CON. Table 1. Rat body weight, and food and water intakes. Values represent the mean ± SEM, n = 5-8. LETO Long-Evans Tokushima Otsuka; OLETF Otsuka Long-Evans Tokushima Fatty; CON control diet; KES 1-kestose diet. # P < 0.05 compared with LETO/CON. $ P < 0.05 compared with LETO/KES. *P < 0.05 compared with OLETF/CON. The data were compared using Tukey-Kramer's test. a Body weight is a value at the termination of the experiment. b Food intake is the average of the 16 weeks of the experimental period. c Water intake is the average of one week of the recorded period from age 22 to 23-week-old (from 15 to 16th week of the experiment).

Scientific RepoRtS
| (2020) 10:15674 | https://doi.org/10.1038/s41598-020-72773-2 www.nature.com/scientificreports/ Glycemic response. We evaluated the effect of 1-kestose supplementation on glucose tolerance in OLETF rats. In the OGTT, plasma glucose concentrations were higher in OLETF rats than in LETO rats (Fig. 3A). Although 1-kestose supplementation had no effect on the patterns of plasma glucose concentrations in LETO rats, the plasma glucose concentrations in OLETF rats tended to be lower in the 1-kestose group, and the concentration at 120 min in the OGTT was significantly lower in OLETF/KES than in OLETF/CON (Fig. 3A).
Values of the area under the curve (AUC) for plasma glucose were significantly higher in OLETF rats than in LETO rats (Fig. 3B). Although there were no significant differences in the AUC between control and 1-kestose diet groups of either rat type, plasma glucose tended to be lower in OLETF/KES than in OLETF/CON (P = 0.27) (Fig. 3B).

Tissue weights and concentrations of blood components.
To examine whether 1-kestose supplementation imparts physiological effects on the study groups, we compared tissue weights and concentrations of blood components. We found that the liver and perirenal adipose tissue weight were higher in OLETF rats than in LETO rats after 16 weeks of the experimental diet (Table 2). Yet, there were no significant differences between the two dietary groups in either OLETF or LETO rats ( Table 2).
Blood samples were collected under a fed condition at the end of the experiment. The concentrations of all blood components measured, except for insulin, were significantly higher in OLETF rats than in LETO rats; however, they were not different between the two dietary groups in either OLETF or LETO rats (Table 3). Consistent with the values of OLETF/CON and OLETF/KES in Figs. 1B and 2, plasma insulin levels tended to be lower in OLETF/CON than in OLETF/KES, although the difference was not statistically significant (P = 0.88) ( Table 3).
Weights of ceca and cecal contents, and cecal pH. At the end of the experiment, we assessed the effect of 1-kestose on the intestinal environment. 1-Kestose supplementation induced an about two-fold increase in the cecal weights of LETO and OLETF rats (Fig. 4A). Increases in the weights of cecal contents were also  www.nature.com/scientificreports/ observed with 1-kestose supplementation in both rats (Fig. 4B). The pH of cecal contents tended to be lower in the KES groups than in the CON groups of both rats, indicating the slight acidification of cecal contents in the KES groups of both rats (OLETF/KES versus OLETF/CON, P = 0.19) (Fig. 4C). These results suggest that 1-kestose supplementation alters the gut environment, presumably via microbial metabolites such as SCFAs and changes in microbiota communities.

SCFA composition in cecal contents.
Supplementation with 1-kestose may exert its effects on glucose metabolism through increases in SCFA production in the cecum. Concentrations of acetate in the cecal con-   www.nature.com/scientificreports/ tents were comparable among all groups (Fig. 5A), but the total amount of acetate was increased in the cecum of LETO/KES and OLETF/KES (P < 0.05) (Fig. 5D). Concentrations of butyrate in the cecal contents and total amounts in the cecum were greatly enhanced in LETO/KES and OLETF/KES compared to LETO/CON and OLETF/CON, respectively (total butyrate amount, P < 0.05) (Fig. 5B,D). Although cecal concentrations of propionate were reduced by 1-kestose supplementation in both rats (Fig. 5C), the total amounts were unchanged (Fig. 5D). Cecal contents of valerate, isobutyrate, and isovalerate tended to be lower in the KES groups than in the CON groups of both rats; however, those trends were inverted due to increased cecal content weights in the KES groups (Table 4).
Gut microbial composition of the ceca. Finally, to determine the effect of gut microbiota composition on alterations of SCFA components, 16S rRNA gene sequencing was performed to study the gut microbiota of cecal samples collected from the experimental groups. The most abundant phylum was Firmicutes in all groups (Fig. 6A). Cecal samples of LETO/KES and OLETF/KES showed markedly increased relative abundance (mean) of Actinobacteria from 2.3 to 26.3% of OTUs and from 2.2 to 31.7% of OTUs, respectively, compared to their counterparts (Fig. 6A,B). At a family level, a relatively lower abundance of Lachnospiraceae, belonging to the phylum Firmicutes, was observed in OLETF/CON compared to LETO/CON (P = 0.06) (Fig. 6C). The abundance of Lactobacillaceae was higher in OLETF/CON compared to LETO/CON (P < 0.05) (Fig. 6C).
Focusing on specific genera within Lachnospiraceae, Anaerostipes with some represented by butyrate-producing bacteria, was present at significantly higher levels in LETO/KES and OLETF/KES compared to LETO/CON and OLETF/CON (Fig. 6D). In accordance with the higher abundance of the family Lactobacillaceae in OLETF/ CON, Lactobacillus was observed at a significantly higher level compared to LETO/CON, whereas OLETF/ KES showed no significant difference in the genus from LETO/KES (Fig. 6E). Consistent with the higher level of the family Bifidobacteriaceae in LETO/KES and OLETF/KES compared to their counterparts (P < 0.05 and P = 0.30, respectively), the level of the genus Bifidobacterium was higher in the cecal contents of LETO/KES, and showed a higher trend in OLETF/KES (P = 0.30) compared to its counterpart (Fig. 6C,F). At a species level, A. caccae was found at significantly higher levels in LETO/KES and OLETF/KES compared to their counterparts ( Table 5). 1-Kestose supplementation in OLETF rats tended to lower the abundance of Akkermansia muciniphila www.nature.com/scientificreports/ compared to OLETF/CON (P = 0.06) ( Table 5). Taken together, 1-kestose induced the accumulation of butyrate by increasing the level of the butyrate-producing microbe A. caccae.

Discussion
In the present study, we investigated the effects of 1-kestose supplementation on glucose metabolism and gut microbiota in a T2D animal model. We used OLETF rats as the T2D model, and they displayed relatively slow development of insulin resistance and hyperglycemia compared to other genetically obese and diabetic model animals, such as Zucker fatty rats and ob/ob mice. As such, the effects of prebiotics on the gut microbiota may be less acute 13 . OLETF rats fed a control diet (OLETF/CON) in this study showed hyperinsulinemia at 14 and 18 weeks of age without hyperglycemia (Fig. 1B). At 22 weeks of age, they then became hyperglycemia with  www.nature.com/scientificreports/ www.nature.com/scientificreports/ decreased insulin level due to the decline in insulin secretion from pancreatic β-cells, resulting in increased water intake, a typical symptom of diabetes (Figs. 1A and 2; Table 1). In contrast, 1-kestose supplementation resulted in later onset of these glycemic symptoms in OLETF rats; the insulin level in OLETF/KES was significantly higher than that in LETO/KES only at 22 weeks of age (Fig. 1B), but the blood glucose level in OLETF/KES remained constant at a level similar to that observed in the LETO rats (Fig. 1A). Following 16-week 1-kestose supplementation in both LETO and OLETF rats, the ceca were enlarged (Fig. 4A), and cecal butyrate concentrations were markedly increased (Fig. 5B). Also, cecal valerate and isovalerate levels were increased in LETO/KES and OLETF/ KES, respectively (Table 4). These results suggest that the alteration of SCFAs, specifically butyrate, in the ceca may be responsible for a suppressive effect on glucose intolerance in OLETF/KES. The results of the present study indicate the potential of 1-kestose to delay T2D onset and reduce glucose intolerance severity, which is supported by the previous reports using FOSs 8,13 . Glucose tolerance was improved by either 13 weeks of supplementation with 10% FOSs to a high-fat diet 8 or six weeks of FOS-added water intake 13 in diet-induced diabetic mice with showing the reduced body weight. Our study did not affect body weight (Table 1), which may be attributable to differences in experimental conditions, particularly, the duration of the experiment investigating the chronic effects of 1-kestose on T2D onset. One possible mechanism behind the suppressive effect of 1-kestose on glucose intolerance may be the involvement of incretin hormone glucagonlike peptide-1 (GLP-1). Plasma GLP-1 concentrations were increased by indigestible carbohydrate intake in rodents and T2D subjects, co-occurring with improved glucose tolerance or postprandial hyperglycemia [13][14][15] , and the increased GLP-1 secretion was due to SCFAs produced by the gut microbiota through metabolism of the indigestible carbohydrates 16 . These findings imply that GLP-1 secretion participated in the reduction of glucose intolerance severity observed in OLETF/KES through an increased SCFA production by 1-kestose metabolism of gut microbiota (Fig. 5). This possibility may be clarified by a further detailed study examining the acute and chronic effects of 1-kestose supplementation on plasma incretin levels.
In this study, 1-kestose supplementation increased the cecal weight and lowered pH levels and was accompanied by a remarkable increase in total SCFA concentrations (Figs. 4 and 5), as described by others 7,17 . In our previous study, supplementation with 1-kestose also raised the levels of acetate and butyrate in the cecal contents of Sprague-Dawley (SD) rats, while there was no significant difference in the level of propionate between the control and 1-kestose supplemented groups 9 . The conflicting results of 1-kestose supplementation on SCFA production between the previous and present works might be attributable to differences in animal species (SD versus LETO and OLETF) and duration of feeding period (4 versus 16 weeks), which might change the gut environment of the animals, especially pH; the acidic pH 5.5 preferentially promotes butyrate production and limits propionate formation in vitro 18,19 . Thus, our findings suggest that the high production of butyrate, in response to 1-kestose supplementation, may lead to increasing gut acidity, and provide a more favorable gut environment for butyrate producers.
We observed a higher proportion of butyrate-producing A. caccae in rats fed KES than in those fed CON ( Fig. 6D; Table 5). Consistent with our previous in vitro study 10 , the present study indicates that feeding 1-kestose stimulates the growth of A. caccae in vivo, resulting in a marked accumulation of butyrate. The microorganism is also known to generate ATP from lactate as a carbon source, which leads to the accumulation of butyrate 20 . In addition, we demonstrated that 1-kestose increased the proportion of bifidobacteria in the rat ceca (Fig. 6F). Bifidobacteria produce lactate and acetate via carbohydrate metabolism, although the ratio of the two acids produced varies in accordance with the available carbohydrates 10 . This might suggest that bifidobacteria metabolized 1-kestose and produced lactate and acetate. The lactate produced was presumably a source of butyrate production by A. caccae in the ceca of LETO/KES and OLETF/KES. Similar findings were reported by feeding galactooligosaccharides and A. caccae in rats, and by administration of germinated barley foodstuff in humans [20][21][22] . Previous studies revealed that this species grows well on degrees of polymerization (DP) 3 FOS, 1-kestose, but shows poor growth on DP4 FOS, nystose. This difference in growth stimulation activities is due to substrate specificities of FOS degradation enzymes in the organism and gene induction activity attributable to the two FOSs 10,23 . These findings indicate that beneficial physiological outcomes such as suppressing the progress of T2D observed in KES-fed rats, are specific to certain oligosaccharides including 1-kestose. A recent study revealed that A. caccae was protective against an allergic response to food 24 , which might be an unrevealed outcome of 1-kestose administration.
Our findings raise questions that should be addressed in the future, such as determining a causative relationship between 1-kestose-induced changes in the gut microbiota and mitigation of glucose metabolism by adopting fecal microbiota transplantation in germ-free or antibiotic-treated mice. Thereby, we may identify mechanisms Table 5. Relative abundance of each species. Values represent the median value plus the interquartile range, n = 5-8. LETO Long-Evans Tokushima Otsuka; OLETF Otsuka Long-Evans Tokushima Fatty; CON control diet; KES 1-kestose diet. † P < 0.05 compared with LETO/CON. *P < 0.05 compared with OLETOF/CON. Data were compared using Kruskal-Wallis test followed by Dunn's multiple comparison test.

Methods
Experimental animals. All procedures for animal experiments in the present study were approved by the